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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
36b9bd02-2e62-44f9-a705-c95b07df9b72 | deepsportradar-v1-computer-vision-dataset-for | 2208.08190 | null | https://arxiv.org/abs/2208.08190v1 | https://arxiv.org/pdf/2208.08190v1.pdf | DeepSportradar-v1: Computer Vision Dataset for Sports Understanding with High Quality Annotations | With the recent development of Deep Learning applied to Computer Vision, sport video understanding has gained a lot of attention, providing much richer information for both sport consumers and leagues. This paper introduces DeepSportradar-v1, a suite of computer vision tasks, datasets and benchmarks for automated sport understanding. The main purpose of this framework is to close the gap between academic research and real world settings. To this end, the datasets provide high-resolution raw images, camera parameters and high quality annotations. DeepSportradar currently supports four challenging tasks related to basketball: ball 3D localization, camera calibration, player instance segmentation and player re-identification. For each of the four tasks, a detailed description of the dataset, objective, performance metrics, and the proposed baseline method are provided. To encourage further research on advanced methods for sport understanding, a competition is organized as part of the MMSports workshop from the ACM Multimedia 2022 conference, where participants have to develop state-of-the-art methods to solve the above tasks. The four datasets, development kits and baselines are publicly available. | ['Davide Zambrano', 'Carlo Del Don', 'Maxime Istasse', 'Vladimir Somers', 'Gabriel Van Zandycke'] | 2022-08-17 | null | null | null | null | ['sports-understanding'] | ['miscellaneous'] | [-1.27490550e-01 -4.37812090e-01 -5.19059956e-01 -2.95797527e-01
-8.53100717e-01 -7.55065799e-01 1.18831657e-01 1.34067997e-01
-6.45268142e-01 1.95122957e-01 1.46315888e-01 2.94885278e-01
3.46191078e-01 -4.61232901e-01 -9.77747142e-01 -2.14506090e-01
3.62817198e-02 4.91352051e-01 6.92347169e-01 -4.29832727e-01
3.61313164e-01 2.63333350e-01 -1.95231128e+00 8.51523697e-01
7.70128891e-03 8.81219208e-01 1.49140418e-01 1.01011539e+00
3.09851617e-01 7.75338352e-01 -5.27321100e-01 -9.68309045e-01
5.91488898e-01 -2.85125703e-01 -1.00134289e+00 2.15048790e-01
9.61461067e-01 -2.99850792e-01 -7.27774858e-01 9.39250350e-01
7.48030841e-01 4.06017572e-01 1.36032835e-01 -1.31955457e+00
-1.00600176e-01 7.12341011e-01 -8.07628870e-01 8.65330696e-01
4.39409524e-01 5.54502547e-01 9.15424407e-01 -7.77233660e-01
5.38789988e-01 1.06469512e+00 9.35293853e-01 4.45777118e-01
-8.68979692e-01 -7.96135128e-01 1.78272367e-01 7.27969348e-01
-1.31304896e+00 -3.37979048e-01 4.39181328e-01 -4.85121369e-01
7.73305953e-01 2.35300869e-01 1.41191387e+00 1.23065186e+00
-2.36193776e-01 1.43385231e+00 9.06057119e-01 -3.37851077e-01
-2.09518060e-01 1.43579952e-02 4.74690527e-01 4.03944641e-01
1.97606638e-01 2.47857362e-01 -1.02409005e+00 6.50602639e-01
9.64355588e-01 -3.14901799e-01 1.40275806e-01 -3.92558187e-01
-1.01956654e+00 7.91953683e-01 1.90859422e-01 8.17747936e-02
1.61039054e-01 4.28094029e-01 8.29640865e-01 -3.50801088e-02
2.55345821e-01 2.59000152e-01 -5.14339805e-02 -8.59514117e-01
-9.13216829e-01 9.92438853e-01 5.22886395e-01 1.01786888e+00
2.47467175e-01 7.25392550e-02 -2.53054887e-01 9.62453902e-01
-2.93990877e-02 5.16963042e-02 1.60862967e-01 -1.25590384e+00
7.55039215e-01 3.59808862e-01 -1.84675530e-02 -1.09989893e+00
-4.34171408e-01 -5.82459331e-01 -2.33195171e-01 8.82532448e-02
6.85941339e-01 1.89406022e-01 -6.58327520e-01 1.17524743e+00
9.79883894e-02 6.51624620e-01 -9.36149731e-02 1.60426271e+00
1.51441646e+00 5.21875203e-01 -3.40320095e-02 5.02389669e-01
1.50069201e+00 -1.33851087e+00 -4.56039280e-01 -6.27205670e-01
4.46145684e-01 -7.65585303e-01 1.11064625e+00 6.27872407e-01
-1.46964049e+00 -1.09962738e+00 -1.13188612e+00 -2.79013216e-01
-2.40349144e-01 1.23181045e-01 6.78360164e-01 7.23939657e-01
-6.13488078e-01 4.33887362e-01 -8.07163239e-01 -2.79665530e-01
3.99469405e-01 6.77191913e-02 -3.41892093e-01 -2.18078747e-01
-1.24884713e+00 7.09480226e-01 6.67865098e-01 2.97039181e-01
-1.07860458e+00 -8.59918356e-01 -9.42053497e-01 -2.49967620e-01
6.23984218e-01 -4.68391329e-01 1.67023551e+00 -1.00425029e+00
-1.27372468e+00 1.66149938e+00 5.28343141e-01 -7.15574563e-01
6.53600693e-01 -6.22760296e-01 -2.10238278e-01 -6.44678324e-02
3.72661769e-01 6.76861048e-01 2.99565077e-01 -1.01200318e+00
-1.32474446e+00 -4.04705435e-01 3.59485745e-01 5.30277848e-01
6.05580360e-02 2.98856139e-01 -1.46607327e+00 -6.47934318e-01
-2.91437775e-01 -1.04979920e+00 -2.07387164e-01 -3.91565174e-01
-2.32501045e-01 -6.66158274e-02 4.17750776e-01 -5.54985464e-01
1.08823574e+00 -2.08521175e+00 2.24964157e-01 -1.83368668e-01
9.13290232e-02 4.86532718e-01 -4.51424718e-02 1.20844632e-01
-2.13742286e-01 -3.39268267e-01 2.20094919e-01 -2.00039372e-01
3.83391492e-02 2.36240506e-01 1.28990769e-01 5.04352093e-01
-2.91970730e-01 8.06993663e-01 -8.34793568e-01 -5.04568636e-01
3.97366762e-01 -8.16964265e-03 -5.50330341e-01 1.29504390e-02
1.53818697e-01 3.35049152e-01 -2.72978246e-02 6.37768447e-01
6.00322962e-01 3.04777652e-01 -2.30274960e-01 -3.96504104e-01
-2.39090502e-01 -1.90631747e-01 -1.83441651e+00 2.19043684e+00
1.28790066e-01 9.98590231e-01 1.65537521e-01 -1.37353694e+00
6.52340770e-01 -1.20718099e-01 6.02281213e-01 -7.10609078e-01
1.26481503e-01 -1.05689257e-01 -2.23259330e-01 -6.77796781e-01
9.18969870e-01 1.37965813e-01 -2.83854783e-01 1.85907241e-02
1.74580038e-01 -3.85164768e-02 9.22870934e-01 1.50716618e-01
7.18950331e-01 3.34869444e-01 -1.54540524e-01 -3.30496058e-02
3.30220401e-01 5.17887294e-01 7.16326892e-01 1.03970337e+00
-4.53645170e-01 8.59896719e-01 2.47082934e-01 -9.68243420e-01
-1.04487801e+00 -9.42142487e-01 2.36096144e-01 1.26580548e+00
6.33055270e-01 -6.66171610e-01 -8.48964155e-01 -3.52289796e-01
-1.13936691e-02 1.74570724e-01 -5.58005214e-01 -2.09082216e-01
-8.62778902e-01 -5.97580791e-01 8.47810924e-01 1.07918787e+00
8.05851102e-01 -8.75266790e-01 -6.89634144e-01 1.82428584e-01
-6.33003831e-01 -1.49847138e+00 -6.10354006e-01 -3.17186803e-01
-6.30175173e-01 -1.46303201e+00 -7.07806408e-01 -9.28121626e-01
-3.18205267e-01 5.32398403e-01 1.65314579e+00 5.17367991e-03
-7.58889437e-01 6.96046710e-01 -6.10851526e-01 -7.51652300e-01
-8.25931318e-03 1.49019778e-01 -1.79672211e-01 -1.30167857e-01
7.64146745e-01 4.41870727e-02 -7.62313902e-01 5.03526151e-01
-6.62705064e-01 1.37729287e-01 3.49418193e-01 4.61856484e-01
9.30561543e-01 -9.80082378e-02 -4.70902212e-02 -6.88246131e-01
3.85779500e-01 3.38976905e-02 -6.55078113e-01 -9.51324105e-02
-3.54757980e-02 -5.44209659e-01 -2.49996290e-01 -2.72052169e-01
-8.03716421e-01 3.40664387e-01 -3.29045206e-01 -4.07134920e-01
-3.56544256e-01 1.88034907e-01 -2.16497496e-01 -3.97486612e-02
8.16556394e-01 7.48307481e-02 -1.70622572e-01 -6.05582297e-01
4.01647210e-01 5.88062406e-01 1.27940321e+00 -6.64550185e-01
4.97748673e-01 4.33268785e-01 -3.09922159e-01 -7.60442138e-01
-1.51459634e+00 -1.03379452e+00 -9.09145772e-01 -9.47469354e-01
1.21102595e+00 -1.40654993e+00 -8.06313455e-01 6.97507203e-01
-8.80334377e-01 -2.32124969e-01 -6.05285645e-01 7.58340120e-01
-7.08361268e-01 3.49927694e-01 -7.78033435e-01 -3.90978992e-01
-3.71222869e-02 -1.36102283e+00 1.12074220e+00 4.60348368e-01
7.31719658e-02 -5.86583674e-01 2.68170595e-01 1.29831755e+00
-3.78926128e-01 4.01604086e-01 -8.29040557e-02 -4.30355966e-01
-5.13515949e-01 -5.11854649e-01 -6.76715747e-02 4.59031522e-01
-7.14328229e-01 -3.55009675e-01 -8.42567384e-01 -1.78738281e-01
-5.15097022e-01 -4.30175036e-01 1.06388450e+00 7.78746486e-01
1.37741840e+00 5.72295845e-01 -1.37362912e-01 9.26913917e-01
9.60922062e-01 -1.35284334e-01 6.58335149e-01 1.04939830e+00
8.90953243e-01 5.18678784e-01 1.05521631e+00 4.85145479e-01
6.15756273e-01 1.15161216e+00 5.64303398e-01 -3.00276071e-01
-3.80213439e-01 -2.16664329e-01 2.75901049e-01 4.94106501e-01
-7.43053794e-01 9.56745818e-02 -8.51597488e-01 5.71335733e-01
-2.02251029e+00 -1.38965237e+00 -6.74456537e-01 1.98107100e+00
4.19636786e-01 2.94903606e-01 1.08113825e+00 2.54154325e-01
1.02691865e+00 4.17142883e-02 -4.77215201e-01 -3.20046842e-01
-2.98515290e-01 5.93290031e-02 7.99371719e-01 1.55297192e-02
-1.68127954e+00 1.24181771e+00 6.55789804e+00 9.15168583e-01
-4.85428452e-01 1.07668839e-01 4.31664258e-01 -4.29327399e-01
6.32454157e-01 -1.57296896e-01 -1.25025582e+00 2.17888147e-01
6.64773405e-01 2.38290187e-02 1.64237633e-01 9.90645468e-01
2.54521430e-01 -1.62342891e-01 -1.09475529e+00 1.50607932e+00
4.48472291e-01 -1.74081492e+00 -2.41576821e-01 -7.53202066e-02
5.39938748e-01 -2.40565613e-02 3.18906009e-02 5.93631744e-01
1.95407361e-01 -9.27552342e-01 1.19429719e+00 4.86359596e-01
4.83920276e-01 -1.06738055e+00 1.00297427e+00 1.98520392e-01
-1.41681826e+00 -3.37219149e-01 -4.52424824e-01 -2.95010388e-01
1.99129626e-01 -3.90932383e-03 -1.05445288e-01 3.54811609e-01
1.60237062e+00 1.24323165e+00 -7.66812623e-01 1.45481694e+00
4.39318344e-02 5.76011002e-01 -3.96861844e-02 3.40564162e-01
2.74731278e-01 -8.82257074e-02 6.16135120e-01 1.62996161e+00
-1.37736008e-01 2.21339643e-01 6.35740578e-01 4.17417318e-01
2.71657202e-02 2.46409290e-02 -1.58430547e-01 1.48461729e-01
2.39316169e-02 1.26240754e+00 -8.88569057e-01 -2.67702579e-01
-5.20363033e-01 7.23182738e-01 -2.21750960e-02 -1.11361519e-02
-1.10659933e+00 -1.83473855e-01 8.43390882e-01 5.62271833e-01
2.46482193e-01 -2.56411463e-01 -2.09964961e-01 -1.12451458e+00
-5.88279814e-02 -1.24422300e+00 8.19238961e-01 -6.37585402e-01
-1.17924619e+00 9.67083499e-02 4.76235896e-01 -1.24879122e+00
1.13645785e-01 -7.49670863e-01 -3.30359101e-01 3.45996082e-01
-1.17744112e+00 -1.14971328e+00 -7.19370604e-01 7.86875129e-01
1.18331218e+00 -4.59842950e-01 1.51747063e-01 7.06741691e-01
-7.66360521e-01 7.61781275e-01 -1.32350028e-01 6.67801678e-01
7.76084125e-01 -1.22352266e+00 4.13769543e-01 8.56642604e-01
5.65166056e-01 -3.79018247e-01 9.00120199e-01 -2.56958336e-01
-1.34792125e+00 -8.97445619e-01 9.73750129e-02 -5.68468690e-01
5.77321053e-01 -4.23470676e-01 -4.30561960e-01 7.19589651e-01
3.65480557e-02 -6.04995489e-02 8.16089869e-01 6.72525913e-02
5.12578115e-02 -4.79601651e-01 -3.47036570e-01 3.04766595e-01
1.23959720e+00 -1.62913706e-02 -4.97937411e-01 5.57040572e-01
1.08831227e-01 -1.15121984e+00 -8.28318536e-01 2.18251035e-01
5.44840932e-01 -1.17768395e+00 1.32373238e+00 -7.83544123e-01
4.26216513e-01 -2.33412877e-01 5.10225631e-02 -8.67398500e-01
-2.54182130e-01 -3.40482801e-01 1.74973384e-01 9.82178688e-01
6.25557378e-02 4.37373817e-01 1.42070830e+00 4.51181263e-01
-5.51303446e-01 -2.92241752e-01 -8.43739748e-01 -6.12484932e-01
-8.76953884e-04 -9.55184579e-01 1.57323003e-01 5.80677152e-01
-1.76783010e-01 3.14358205e-01 -6.67820573e-01 -3.16197122e-03
8.48677695e-01 1.68102413e-01 1.47441339e+00 -1.19962168e+00
-5.61220229e-01 -4.47937936e-01 -1.22586524e+00 -1.34520698e+00
-7.94274732e-02 -7.29943395e-01 8.43232721e-02 -1.44679546e+00
5.02355158e-01 1.98726326e-01 5.90042099e-02 2.48098612e-01
-9.60208476e-02 8.77314746e-01 6.03627682e-01 1.32861495e-01
-1.06532764e+00 -8.03191215e-03 1.30940211e+00 -3.45933825e-01
-1.41624793e-01 4.16707486e-01 -6.15329087e-01 8.80892038e-01
5.80194354e-01 -3.44644576e-01 -3.93317826e-02 -6.53638422e-01
4.41847332e-02 -4.83898744e-02 7.08179533e-01 -1.46281302e+00
2.53950596e-01 -2.49366257e-02 5.08504152e-01 -8.70511353e-01
7.75106490e-01 -4.24921066e-01 7.91445747e-02 2.74300426e-01
-4.15887922e-01 2.24034548e-01 4.74216521e-01 6.49831474e-01
-6.05638683e-01 -3.07574928e-01 8.22916746e-01 -5.89352369e-01
-1.54938364e+00 3.64813983e-01 -4.93230462e-01 5.18240750e-01
1.38730812e+00 -8.82833004e-01 3.90422618e-04 -3.56257677e-01
-1.26642358e+00 5.43148577e-01 -6.18960187e-02 7.76627004e-01
6.07146382e-01 -1.31131506e+00 -1.16333389e+00 -1.52303129e-01
2.81134188e-01 -1.43414466e-02 6.73206747e-01 6.77793503e-01
-9.81832504e-01 2.31846377e-01 -5.34454942e-01 -9.79645729e-01
-1.68033922e+00 2.36848384e-01 5.52689552e-01 -1.92463115e-01
-9.35754955e-01 1.09651291e+00 -9.48808491e-02 -9.85718593e-02
5.06896138e-01 -1.88215449e-02 -5.28870702e-01 4.26588617e-02
7.22301304e-01 5.58832824e-01 7.46950433e-02 -9.44786131e-01
-1.29888520e-01 6.89071596e-01 -1.71023190e-01 -4.49285582e-02
1.37679243e+00 -1.30366981e-01 5.94138205e-01 3.74534905e-01
7.72437871e-01 -5.70196629e-01 -1.43779325e+00 -9.37761888e-02
-5.98381199e-02 -8.54286134e-01 3.08795512e-01 -4.58867311e-01
-1.45801973e+00 1.07411230e+00 9.67343390e-01 -9.28590149e-02
9.02022719e-01 1.92084312e-01 8.29474449e-01 3.51499319e-02
3.00381146e-02 -1.71069109e+00 3.60880226e-01 5.25248885e-01
6.44407153e-01 -1.31325114e+00 2.37202048e-02 -4.21161354e-01
-1.09910679e+00 1.15699136e+00 9.89469826e-01 -2.21971035e-01
2.22360477e-01 2.34271675e-01 1.39835507e-01 -3.77625197e-01
-1.40018821e-01 -5.38955748e-01 4.46117342e-01 8.70479524e-01
3.17814201e-01 -2.37900361e-01 -9.77222472e-02 1.00674844e+00
-4.91378844e-01 3.09498888e-03 5.95267653e-01 9.20535803e-01
-4.60771352e-01 -1.06843317e+00 -5.71008980e-01 5.97711280e-02
-6.30162418e-01 2.54484504e-01 -5.27653039e-01 1.20828855e+00
4.22072083e-01 8.70529234e-01 2.00158045e-01 -3.82918805e-01
1.08529902e+00 -4.60144460e-01 6.75437510e-01 -6.68238699e-01
-8.42621803e-01 1.78173453e-01 3.10968488e-01 -8.17088306e-01
-7.29696155e-01 -8.84302437e-01 -7.73470044e-01 -4.21204180e-01
-2.03008696e-01 2.13084280e-01 5.95751703e-01 8.10275257e-01
-9.74716842e-02 5.79784870e-01 5.55940531e-02 -1.24259412e+00
-5.91088198e-02 -6.55345678e-01 -8.82421613e-01 6.01845145e-01
-3.89426410e-01 -8.54394257e-01 2.23012611e-01 4.33453411e-01] | [7.804211616516113, 0.16752324998378754] |
bb9987e3-7f5f-471e-b4f2-656dc51f46a8 | discogem-a-crowdsourced-corpus-of-genre-mixed | null | null | https://aclanthology.org/2022.lrec-1.351 | https://aclanthology.org/2022.lrec-1.351.pdf | DiscoGeM: A Crowdsourced Corpus of Genre-Mixed Implicit Discourse Relations | We present DiscoGeM, a crowdsourced corpus of 6,505 implicit discourse relations from three genres: political speech, literature, and encyclopedic texts. Each instance was annotated by 10 crowd workers. Various label aggregation methods were explored to evaluate how to obtain a label that best captures the meaning inferred by the crowd annotators. The results show that a significant proportion of discourse relations in DiscoGeM are ambiguous and can express multiple relation senses. Probability distribution labels better capture these interpretations than single labels. Further, the results emphasize that text genre crucially affects the distribution of discourse relations, suggesting that genre should be included as a factor in automatic relation classification. We make available the newly created DiscoGeM corpus, as well as the dataset with all annotator-level labels. Both the corpus and the dataset can facilitate a multitude of applications and research purposes, for example to function as training data to improve the performance of automatic discourse relation parsers, as well as facilitate research into non-connective signals of discourse relations. | ['Vera Demberg', 'Frances Yung', 'Tianai Dong', 'Merel Scholman'] | null | null | null | null | lrec-2022-6 | ['relation-classification'] | ['natural-language-processing'] | [ 8.57257172e-02 7.99396813e-01 -5.22247434e-01 -4.09791559e-01
-8.12298656e-01 -1.00034177e+00 1.12519956e+00 7.71762729e-01
-4.21006113e-01 1.04315555e+00 9.98188972e-01 -3.96825463e-01
1.34649456e-01 -5.66930771e-01 -4.41159964e-01 -3.84243727e-01
3.17278266e-01 8.70571077e-01 3.85382473e-01 -3.89609337e-01
3.43143731e-01 -2.63449520e-01 -1.13299716e+00 7.25675166e-01
9.81273770e-01 7.00195909e-01 1.48418933e-01 4.16087002e-01
-4.53850865e-01 1.35318983e+00 -1.26500499e+00 -7.79411912e-01
-3.69487405e-01 -5.17899454e-01 -1.41288853e+00 2.07281746e-02
-3.58782187e-02 3.02053899e-01 1.04485437e-01 8.98647606e-01
3.89542311e-01 -5.55317737e-02 8.17172885e-01 -8.61931443e-01
-7.60892868e-01 1.31658494e+00 -2.81748354e-01 5.26606381e-01
6.35560930e-01 -1.80600613e-01 1.40264809e+00 -6.48681581e-01
1.29131377e+00 1.47990489e+00 3.16779286e-01 1.72755435e-01
-1.20848966e+00 -2.78532267e-01 -5.03476448e-02 2.59499729e-01
-1.03732038e+00 -5.50764024e-01 6.29478574e-01 -9.24117088e-01
9.97575521e-01 4.38593090e-01 4.59185690e-01 1.31253290e+00
-3.43782902e-01 5.67106605e-01 1.24791873e+00 -8.23607326e-01
1.53350338e-01 1.80587590e-01 2.91242898e-01 3.29239041e-01
2.44944870e-01 -6.03480875e-01 -7.12826908e-01 -4.85191554e-01
2.86581874e-01 -1.01922011e+00 -3.50352347e-01 4.26020384e-01
-1.33579326e+00 9.67152417e-01 2.08407208e-01 7.14743018e-01
-1.38413534e-01 -2.28964791e-01 7.06865966e-01 1.63711801e-01
9.73861933e-01 9.40093577e-01 -2.61231482e-01 -4.51670021e-01
-8.89231414e-02 3.58722121e-01 1.19694936e+00 8.55233252e-01
4.74416912e-01 -5.30112028e-01 -4.30295378e-01 1.03814387e+00
2.74848074e-01 2.05180734e-01 3.63284558e-01 -1.26818991e+00
9.62952673e-01 1.04276633e+00 2.23499790e-01 -1.07285285e+00
-4.21250314e-01 1.93232596e-01 -2.67662495e-01 -3.38864744e-01
8.69962037e-01 -2.95213163e-01 -2.20719483e-02 1.49132252e+00
5.69247246e-01 -4.72630769e-01 8.67298469e-02 9.37431931e-01
1.28406990e+00 4.13893491e-01 3.55700046e-01 -3.20028573e-01
1.87238050e+00 -5.54078877e-01 -1.04690015e+00 -1.99497357e-01
1.09359717e+00 -1.19539714e+00 9.87504005e-01 -2.66730622e-03
-8.86057675e-01 -8.88437331e-02 -8.93810213e-01 -4.21296328e-01
-2.04012781e-01 9.96710509e-02 6.32196009e-01 3.11756551e-01
-3.45478684e-01 1.82720214e-01 -3.89021069e-01 -2.08622873e-01
5.94798386e-01 -1.09293595e-01 -1.94141820e-01 3.50405902e-01
-1.47254443e+00 1.24126637e+00 5.08896768e-01 -1.11092269e-01
-1.55388534e-01 -6.87854961e-02 -8.93705666e-01 -3.04981560e-01
5.54582536e-01 -1.63371280e-01 1.30448663e+00 -7.55875647e-01
-1.22390640e+00 1.57290184e+00 -2.72805333e-01 -4.42942500e-01
4.93174821e-01 -2.88017809e-01 -2.39192501e-01 2.90315710e-02
5.14863372e-01 1.24837808e-01 2.68136740e-01 -1.10889208e+00
-7.69188464e-01 -1.52938664e-01 2.45236158e-01 3.45421642e-01
-1.15409270e-01 4.24874842e-01 1.04034059e-01 -5.63302934e-01
-4.47228961e-02 -8.67481530e-01 3.28156322e-01 -5.49869180e-01
-6.82698548e-01 -8.92445683e-01 6.28069341e-01 -6.77558780e-01
1.16669762e+00 -1.92882216e+00 2.89026082e-01 -7.94894919e-02
4.63830680e-01 1.32433891e-01 3.00640404e-01 5.98009408e-01
3.82590979e-01 4.18699652e-01 2.56545171e-02 -2.50364929e-01
-5.08516803e-02 4.76355851e-01 -2.09825173e-01 5.40413141e-01
2.57203221e-01 9.59042430e-01 -1.12170291e+00 -8.16190958e-01
-1.93992972e-01 1.92319974e-02 -6.36745170e-02 4.78807949e-02
-7.38912821e-01 7.60080218e-01 -6.66854918e-01 2.60723054e-01
-1.36553884e-01 -5.48862100e-01 6.10318959e-01 5.95487095e-02
-1.20733045e-01 1.09243786e+00 -5.49328327e-01 1.18985033e+00
-2.07313210e-01 1.22311330e+00 1.27489746e-01 -7.19817996e-01
1.01004195e+00 6.09435797e-01 3.52284051e-02 -4.46579427e-01
5.59764385e-01 3.66642505e-01 3.46311420e-01 -8.43226671e-01
8.40021729e-01 -9.94736701e-02 -3.00197542e-01 6.66261971e-01
-1.80285916e-01 -7.75811970e-02 5.46654880e-01 1.73325092e-01
9.89192426e-01 -2.25589171e-01 5.91051221e-01 -4.72227722e-01
3.75236839e-01 6.11782372e-01 5.70990741e-01 4.03413266e-01
-1.74138602e-02 2.81374544e-01 1.13116276e+00 -3.07342231e-01
-8.00295889e-01 -4.29823428e-01 -3.81502420e-01 1.36248517e+00
2.72269338e-01 -4.30198371e-01 -5.92196167e-01 -5.40645599e-01
-2.31504425e-01 7.03482330e-01 -7.11861789e-01 5.02280533e-01
-9.15004015e-01 -6.71344399e-01 7.51481175e-01 2.81275034e-01
3.67528290e-01 -1.14451456e+00 -5.49861252e-01 2.29775861e-01
-9.73452091e-01 -1.45364332e+00 -6.47830516e-02 2.11410880e-01
-9.93417054e-02 -1.58646584e+00 -8.11253861e-02 -9.27019835e-01
2.79192209e-01 -1.49218947e-01 1.37629342e+00 2.93536127e-01
3.01241606e-01 -1.29483059e-01 -7.65963912e-01 -4.64282870e-01
-1.07746232e+00 4.62276340e-01 -3.15892518e-01 -4.18988198e-01
6.01260662e-01 -2.76745349e-01 -2.65934207e-02 1.54653892e-01
-4.17350888e-01 4.91019674e-02 -2.23237798e-01 7.62971878e-01
1.14161216e-01 -2.91000873e-01 7.01587319e-01 -1.41139209e+00
1.26085627e+00 -6.72597110e-01 -2.95460790e-01 2.97632217e-01
-1.29304111e-01 -7.56689236e-02 2.61670977e-01 -5.45743048e-01
-1.25718963e+00 -5.45183957e-01 3.23518179e-02 4.88095582e-01
-1.45824114e-02 6.80401504e-01 -7.70363808e-02 6.35879457e-01
1.26292205e+00 -6.96538389e-01 -1.14915967e-01 -2.26904452e-01
4.18839008e-01 1.03518796e+00 4.44274426e-01 -1.02241385e+00
1.86597511e-01 1.51581675e-01 -2.66785026e-01 -8.97045791e-01
-1.27495217e+00 -4.40793484e-01 -5.20884097e-01 -2.18735233e-01
1.11741102e+00 -9.35689688e-01 -6.99142396e-01 -4.74491380e-02
-1.64545894e+00 -6.52611196e-01 -3.07957202e-01 2.34939650e-01
-2.69338846e-01 8.30732360e-02 -9.70443785e-01 -7.68818974e-01
-2.35300753e-02 -1.08171105e+00 9.41336453e-01 1.16015755e-01
-1.29933655e+00 -1.14420617e+00 1.84840962e-01 8.45930696e-01
-1.40517607e-01 6.13190711e-01 1.11039174e+00 -1.11750901e+00
-2.01260492e-01 1.08314753e-01 -1.25615254e-01 -3.09646666e-01
1.90104619e-01 2.87617613e-02 -1.13529694e+00 5.04599452e-01
-1.38863117e-01 -7.82972276e-01 5.02582371e-01 1.44310826e-02
3.68582904e-01 -4.61081922e-01 -3.73502314e-01 -2.45336398e-01
6.75490260e-01 1.06110111e-01 4.11655009e-01 5.33133924e-01
6.09528780e-01 1.08370388e+00 5.56747794e-01 2.91995287e-01
6.10979378e-01 6.36025310e-01 4.80992682e-02 3.93958658e-01
-1.26716807e-01 -3.81957889e-02 5.67447357e-02 9.49522316e-01
-2.60748565e-01 -4.88128215e-01 -1.21827853e+00 6.15789413e-01
-2.10503936e+00 -8.98474872e-01 -6.08893275e-01 1.42460191e+00
1.32108533e+00 2.14528203e-01 9.04901400e-02 3.19614291e-01
9.17941451e-01 4.76150215e-01 8.75628516e-02 -3.56945336e-01
-4.75149274e-01 -1.16493583e-01 -1.00365551e-02 7.76881158e-01
-1.10981095e+00 1.00392699e+00 5.91418743e+00 7.54780591e-01
-6.02794647e-01 4.40013021e-01 6.52784109e-01 3.19826752e-01
-5.16147912e-01 1.74598873e-01 -5.41663051e-01 4.31418777e-01
6.98188603e-01 -1.59941167e-01 8.70319009e-02 6.24432147e-01
1.57058761e-02 -5.09029567e-01 -1.00888252e+00 3.77333254e-01
-2.00046733e-01 -1.57346511e+00 -6.09033525e-01 9.94122773e-02
7.62391448e-01 -5.14688231e-02 -5.82770228e-01 1.09331161e-01
6.45759225e-01 -9.38010991e-01 9.73080277e-01 2.98266020e-02
7.62829661e-01 -1.39864892e-01 1.04947424e+00 4.91299450e-01
-7.56827712e-01 3.53045017e-01 5.75610176e-02 -6.49474502e-01
4.36422467e-01 6.53490543e-01 -1.26719153e+00 1.80388689e-01
2.96262741e-01 3.75522435e-01 -5.35722971e-01 3.05955887e-01
-7.63256669e-01 9.45849240e-01 -3.18350315e-01 -6.18595421e-01
-6.19581304e-02 1.39279157e-01 7.83412516e-01 1.27988970e+00
-3.72191280e-01 4.42636877e-01 3.60387176e-01 7.33383834e-01
-3.46387923e-01 1.84880674e-01 -4.12931412e-01 -3.24574232e-01
1.15722275e+00 1.06135333e+00 -1.19829488e+00 -2.76741475e-01
-3.03525120e-01 3.84733707e-01 8.08559597e-01 3.14718872e-01
-5.45113087e-01 8.07799175e-02 3.54560524e-01 1.79833904e-01
-9.74649563e-02 -8.11108574e-02 -6.90970361e-01 -8.55546057e-01
2.44950107e-03 -8.50572050e-01 2.41353124e-01 -5.51739216e-01
-1.49278760e+00 7.85663307e-01 1.21578582e-01 -7.80768096e-01
-5.22283316e-01 -4.78739113e-01 -4.19165373e-01 7.38324404e-01
-8.46475244e-01 -9.89738345e-01 -2.96655595e-01 -8.40742663e-02
4.32874680e-01 -1.71854123e-01 9.50679958e-01 2.19436996e-02
-3.32086444e-01 1.08631298e-01 -4.04278040e-01 6.29093885e-01
7.90221155e-01 -1.25057745e+00 2.13178992e-01 4.00225997e-01
3.94775182e-01 4.14257497e-01 9.95068729e-01 -8.15362096e-01
-5.27583063e-01 -6.46255255e-01 1.30616450e+00 -9.24644828e-01
9.84206796e-01 -1.88977733e-01 -9.75093186e-01 6.80842340e-01
5.82910001e-01 -2.50705838e-01 9.73909378e-01 7.06012249e-01
-2.76666015e-01 6.95477784e-01 -8.67296338e-01 5.71049869e-01
1.03181732e+00 -7.13495553e-01 -1.11916268e+00 5.59837580e-01
8.67249966e-01 -1.05198765e+00 -1.01005447e+00 7.92570263e-02
1.15934364e-01 -6.38111293e-01 4.05345500e-01 -4.52213198e-01
9.17801261e-01 -1.68278456e-01 -2.02902704e-01 -1.05909443e+00
-1.21796345e-02 -6.77614629e-01 -3.54039341e-01 1.73068774e+00
8.08198929e-01 -6.78207338e-01 3.25726241e-01 8.10481966e-01
-6.19224831e-02 -7.24275589e-01 -1.10322416e+00 -4.10701692e-01
1.22779183e-01 -4.92665559e-01 4.36500251e-01 1.26286972e+00
5.94406188e-01 1.08941042e+00 1.54515132e-01 -3.36835355e-01
-7.48309791e-02 1.29564345e-01 5.99601328e-01 -1.27883852e+00
-5.59500903e-02 -4.76291746e-01 -1.69705838e-01 -7.66243517e-01
5.93921304e-01 -9.15879607e-01 6.15229830e-02 -1.52437210e+00
-6.15283847e-03 -7.69632757e-01 7.08914280e-01 3.81056786e-01
-3.89266521e-01 1.90942645e-01 8.39116797e-02 5.30368567e-01
-6.45495236e-01 4.17634189e-01 1.32277703e+00 -1.96935385e-02
-3.23914379e-01 -2.77836591e-01 -8.70690227e-01 1.11989391e+00
6.52582645e-01 -6.20421827e-01 -1.65365756e-01 -4.33779418e-01
5.66261828e-01 7.04103112e-02 1.05732709e-01 -1.99653447e-01
1.11867785e-01 -3.35704029e-01 1.52894706e-02 -1.46895364e-01
2.82455415e-01 -2.52780318e-01 -1.21510886e-01 -8.15154687e-02
-8.40098262e-01 -1.42084509e-01 -1.60466433e-01 3.96434218e-01
-3.23487461e-01 -5.10153830e-01 2.70892799e-01 -6.82830885e-02
-3.71954679e-01 -6.88351035e-01 -5.77941656e-01 1.02083063e+00
8.77864778e-01 6.76784366e-02 -1.10170197e+00 -3.91827881e-01
-6.88078523e-01 1.35674968e-01 4.12259549e-01 4.15515155e-01
-8.35824534e-02 -1.21162009e+00 -9.08895254e-01 -4.58860219e-01
9.22878906e-02 4.34484482e-01 -4.65538710e-01 6.14147544e-01
-4.51974452e-01 9.71818194e-02 3.58560771e-01 -5.07494271e-01
-1.24673533e+00 6.55145794e-02 -2.24006981e-01 -4.67947185e-01
-6.41034305e-01 8.71745408e-01 -2.43482366e-01 -1.21857822e-01
-3.46360505e-02 -2.83841521e-01 -5.83880246e-01 5.52877009e-01
5.33764005e-01 3.03040236e-01 -2.18887478e-01 -1.04004872e+00
-2.97868788e-01 -8.25830922e-02 2.63221532e-01 -4.01939511e-01
1.14139450e+00 -4.18179154e-01 -5.21864653e-01 9.16238308e-01
7.50426531e-01 4.67243314e-01 -7.39599407e-01 -3.04052562e-01
6.21953964e-01 -2.01083347e-01 -4.62602347e-01 -6.45594895e-01
-2.40230903e-01 2.70092398e-01 -3.78112346e-01 1.00445890e+00
3.95484954e-01 7.03743815e-01 4.91047859e-01 2.46589959e-01
3.63665193e-01 -1.19203162e+00 3.65489423e-02 9.09445465e-01
1.03349125e+00 -1.39480841e+00 1.50984108e-01 -9.09305990e-01
-9.79760230e-01 9.24133837e-01 6.04620874e-01 3.72907482e-02
3.05128127e-01 5.05957544e-01 3.38304728e-01 -6.09818518e-01
-6.52477086e-01 -2.75214523e-01 2.93337852e-01 5.74128270e-01
1.14065170e+00 3.12332928e-01 -8.02060366e-01 8.25093806e-01
-6.27541304e-01 -5.93129694e-01 4.51535344e-01 6.41961336e-01
-3.05644929e-01 -1.19144452e+00 -6.35298014e-01 4.75994498e-01
-5.81197798e-01 1.24202050e-01 -8.91301632e-01 7.61758745e-01
2.10307479e-01 1.27587688e+00 1.63417250e-01 -1.86440945e-01
-6.72494248e-02 8.80298316e-02 3.61904711e-01 -9.09500360e-01
-7.42673099e-01 -2.33694389e-01 1.34491444e+00 -7.86135159e-03
-1.02541208e+00 -6.17479801e-01 -1.45319879e+00 -2.86574721e-01
-5.67275286e-01 3.56275856e-01 8.11036006e-02 1.55551612e+00
7.02642370e-03 4.64106262e-01 -8.53492394e-02 -5.69848835e-01
-3.77057493e-03 -1.51855719e+00 -1.57300919e-01 6.84166014e-01
4.76876125e-02 -8.82748842e-01 -2.37257153e-01 2.19730109e-01] | [10.701557159423828, 9.317848205566406] |
e16dbd9c-97ed-425a-b4dc-5e91f8569f57 | one-shot-face-reenactment-using-appearance | 2102.03984 | null | https://arxiv.org/abs/2102.03984v3 | https://arxiv.org/pdf/2102.03984v3.pdf | One-shot Face Reenactment Using Appearance Adaptive Normalization | The paper proposes a novel generative adversarial network for one-shot face reenactment, which can animate a single face image to a different pose-and-expression (provided by a driving image) while keeping its original appearance. The core of our network is a novel mechanism called appearance adaptive normalization, which can effectively integrate the appearance information from the input image into our face generator by modulating the feature maps of the generator using the learned adaptive parameters. Furthermore, we specially design a local net to reenact the local facial components (i.e., eyes, nose and mouth) first, which is a much easier task for the network to learn and can in turn provide explicit anchors to guide our face generator to learn the global appearance and pose-and-expression. Extensive quantitative and qualitative experiments demonstrate the significant efficacy of our model compared with prior one-shot methods. | ['Kun Zhou', 'Mengmeng Wang', 'Yong liu', 'Shanqi Liu', 'Shuang Li', 'Tianjia Shao', 'Yi Yuan', 'Guangming Yao'] | 2021-02-08 | null | null | null | null | ['face-reenactment'] | ['computer-vision'] | [ 1.31530449e-01 5.60084164e-01 3.43558975e-02 -5.61160743e-01
-3.04987103e-01 -6.17253423e-01 5.21175444e-01 -1.08574843e+00
1.91640332e-01 4.28072214e-01 -1.82301570e-02 3.67618322e-01
3.79292399e-01 -8.15574408e-01 -7.33136058e-01 -9.07107890e-01
3.28835011e-01 1.64653525e-01 -3.63435537e-01 -5.42871654e-01
-2.53915757e-01 9.05744076e-01 -1.52836907e+00 -4.80539538e-02
5.43629289e-01 9.94452178e-01 -3.57308120e-01 5.82384706e-01
2.09402904e-01 7.11812377e-01 -7.31800377e-01 -7.09345877e-01
5.07609427e-01 -7.69676685e-01 -3.81335467e-01 3.73252839e-01
7.58485734e-01 -6.72599316e-01 -5.87463915e-01 1.13996840e+00
4.64533746e-01 9.79847386e-02 6.25889182e-01 -1.61903358e+00
-1.02138567e+00 2.28213862e-01 -6.96197331e-01 -4.74719971e-01
3.71322155e-01 4.88436490e-01 5.55503845e-01 -9.60049748e-01
7.69187868e-01 1.65033078e+00 5.92694581e-01 1.08896351e+00
-1.29733276e+00 -1.22586167e+00 2.25475773e-01 -2.01280728e-01
-1.54023075e+00 -8.90273452e-01 1.19060171e+00 -2.35054027e-02
-1.45567253e-01 1.47041857e-01 7.32695341e-01 1.42759693e+00
1.92839637e-01 3.78230780e-01 9.39104497e-01 -1.64657608e-01
-4.71861549e-02 6.39522122e-03 -7.07480669e-01 9.03085530e-01
-2.65618593e-01 3.00648302e-01 -2.12892413e-01 1.17582502e-02
1.25767982e+00 2.06797272e-01 -1.29011691e-01 -4.64345664e-01
-5.11027753e-01 7.09205270e-01 8.17156255e-01 -3.88750695e-02
-2.17570856e-01 3.20724607e-01 -1.69620365e-01 4.17380571e-01
3.94167691e-01 2.87306130e-01 -9.96541157e-02 4.14581984e-01
-6.03751540e-01 3.95896435e-01 6.27463579e-01 9.92097676e-01
1.00621343e+00 3.99635106e-01 -3.67700666e-01 8.73040974e-01
4.50434685e-01 7.61231720e-01 2.01873139e-01 -1.17271876e+00
-7.89419860e-02 4.89088088e-01 -1.40880764e-01 -1.06195843e+00
-2.95154471e-02 -1.96850806e-01 -9.17245090e-01 6.33073151e-01
2.33312935e-01 -4.85370785e-01 -1.32031691e+00 2.36374521e+00
5.25825441e-01 3.26914757e-01 -1.52607560e-01 7.09522069e-01
9.49837387e-01 4.81538326e-01 2.32225563e-02 -3.96668948e-02
1.14152598e+00 -8.81257117e-01 -7.81053782e-01 -4.09417421e-01
-1.91890553e-01 -6.85679436e-01 8.73054028e-01 -2.74009109e-01
-1.13971710e+00 -6.73221409e-01 -1.14752746e+00 -5.35941198e-02
-2.26277426e-01 -4.69357520e-02 4.76225585e-01 6.08259559e-01
-1.29282165e+00 5.08187294e-01 -5.90689719e-01 -1.00653365e-01
6.05587304e-01 4.97159779e-01 -5.86578906e-01 -1.76121909e-02
-1.14466131e+00 5.17136693e-01 -1.85063109e-01 1.97399676e-01
-1.08848989e+00 -6.55860007e-01 -1.14149761e+00 -4.32695933e-02
2.07134113e-01 -7.53973961e-01 1.27806115e+00 -1.63477004e+00
-2.15213847e+00 1.12240362e+00 -8.81104916e-02 2.17589781e-01
5.72393119e-01 3.42086524e-01 -3.52419287e-01 1.05602369e-01
1.34991771e-02 1.14475906e+00 1.55219460e+00 -1.52925563e+00
-4.10817228e-02 -4.50217634e-01 1.12128891e-01 8.41140449e-02
-2.65742928e-01 1.35111630e-01 -7.16717601e-01 -9.30187643e-01
9.44657344e-03 -1.09734631e+00 -1.78358540e-01 6.33502483e-01
-3.87243032e-01 2.40396857e-01 1.08863306e+00 -5.38897038e-01
6.24014258e-01 -2.29955792e+00 2.40149453e-01 3.28588784e-01
4.32139874e-01 6.70942888e-02 -6.61941707e-01 6.52149096e-02
-4.97176856e-01 3.37121151e-02 -1.55303210e-01 -6.14566922e-01
-1.91560209e-01 1.24097921e-01 -3.88998419e-01 5.00159681e-01
4.81591284e-01 1.33485699e+00 -7.75066972e-01 -3.54707092e-01
3.27427387e-02 8.38925004e-01 -6.53100014e-01 5.86987615e-01
-1.19049795e-01 7.28795886e-01 -3.44969183e-01 6.37304366e-01
9.85702872e-01 1.04416125e-01 2.22384203e-02 -2.65743911e-01
3.39846909e-01 -3.48817080e-01 -8.45456421e-01 1.41317141e+00
-4.11235929e-01 3.41461569e-01 5.30830085e-01 -3.81244630e-01
1.04301405e+00 3.27634305e-01 4.08579916e-01 -6.27622187e-01
5.07465720e-01 -2.53121346e-01 -3.62051763e-02 -6.77598864e-02
2.23498908e-03 -3.44795853e-01 -3.41951661e-02 4.74568665e-01
2.18107551e-01 -3.51164132e-01 -4.09534723e-01 5.14469482e-02
5.99700689e-01 1.86596662e-01 -5.06610870e-02 -1.05205238e-01
4.57591742e-01 -7.34933674e-01 7.01467872e-01 2.64024913e-01
-1.59139246e-01 9.19610322e-01 4.71479565e-01 -4.36209232e-01
-1.10711217e+00 -1.21630776e+00 1.17934123e-01 9.79093850e-01
1.59758613e-01 -1.13664955e-01 -1.29471147e+00 -6.57142878e-01
-1.35685369e-01 2.64174789e-01 -1.18336952e+00 -6.65947676e-01
-6.29429519e-01 -4.70501602e-01 5.74579239e-01 4.28669155e-01
6.37520194e-01 -1.21072340e+00 4.98385280e-02 -9.59644765e-02
6.72915652e-02 -9.85080659e-01 -1.00111520e+00 -3.73086005e-01
-4.02450681e-01 -8.44357967e-01 -7.23568022e-01 -9.86545861e-01
1.22923338e+00 -1.03275515e-01 7.76059985e-01 2.31615633e-01
-3.02876055e-01 1.20320663e-01 -5.42751886e-02 -5.00007927e-01
-6.72460377e-01 -2.14902312e-01 6.24616519e-02 6.07719123e-01
-3.21852058e-01 -9.53290224e-01 -7.91091502e-01 4.72295225e-01
-1.03270841e+00 1.26289263e-01 3.84439349e-01 6.98674321e-01
4.20637578e-01 -1.85932174e-01 6.57918274e-01 -9.76162195e-01
4.85894829e-01 -1.75783098e-01 -4.35373306e-01 8.64488110e-02
-2.73378074e-01 -9.90218446e-02 8.23530972e-01 -7.30147719e-01
-1.09918857e+00 2.89281368e-01 -2.78840989e-01 -9.12721336e-01
4.46786992e-02 -3.52607608e-01 -7.84272611e-01 -6.31981254e-01
4.58587468e-01 4.41716835e-02 5.91187179e-01 -2.35488936e-01
8.62501025e-01 2.74805754e-01 8.78099978e-01 -4.62316722e-01
1.58613396e+00 6.66919887e-01 3.18666212e-02 -4.22035038e-01
-7.72319198e-01 3.51775944e-01 -8.39076817e-01 -1.53457344e-01
6.80058002e-01 -8.45452845e-01 -7.27869630e-01 9.95673776e-01
-1.09199870e+00 -3.53082359e-01 -4.45905536e-01 -3.74038517e-01
-5.96709073e-01 -1.05789229e-01 -5.77540100e-01 -3.46513987e-01
-3.65250915e-01 -1.01443326e+00 1.15896165e+00 6.60583138e-01
-3.70231122e-02 -8.37905347e-01 2.49816533e-02 -2.26598550e-02
5.92790306e-01 5.90729713e-01 7.56496131e-01 -6.80959448e-02
-5.04067659e-01 -3.20024282e-01 -1.33037418e-01 4.64408487e-01
5.60984671e-01 4.61323500e-01 -1.10968995e+00 -5.86562216e-01
2.58719660e-02 -2.87167698e-01 4.58774835e-01 1.14402644e-01
1.13347435e+00 -6.57248020e-01 -2.52941579e-01 1.22309685e+00
1.02422750e+00 1.67762518e-01 9.09493983e-01 -2.65092999e-01
8.32688749e-01 5.37507534e-01 2.53207088e-01 2.02266589e-01
8.04091319e-02 7.72686839e-01 5.19595206e-01 -5.94454348e-01
-5.57849348e-01 -7.34521866e-01 4.51960295e-01 3.62623632e-01
6.70723617e-02 -9.22305137e-02 -2.10980162e-01 1.31447569e-01
-1.45278096e+00 -1.01365685e+00 8.21172953e-01 1.86436749e+00
9.85514045e-01 -3.22621465e-01 5.30051664e-02 -3.35365027e-01
8.70473862e-01 2.74188221e-01 -8.03897619e-01 -2.49842599e-01
7.47120753e-02 6.11689568e-01 8.97435993e-02 4.82081831e-01
-9.15747285e-01 1.28485680e+00 7.17951679e+00 6.53084934e-01
-1.44031680e+00 -1.45831227e-01 8.94930184e-01 -1.34832978e-01
-3.10915232e-01 -9.70029160e-02 -5.78187346e-01 4.87785459e-01
4.90958273e-01 -3.04503828e-01 6.99882865e-01 9.29727077e-01
8.08029203e-04 5.92513084e-01 -1.07447362e+00 8.85852993e-01
3.63154113e-01 -1.13199651e+00 3.48808855e-01 1.52078122e-01
8.88944983e-01 -5.54310620e-01 4.67626274e-01 2.82323927e-01
4.14146453e-01 -1.37919736e+00 7.28662491e-01 6.79270804e-01
1.40580082e+00 -9.62347806e-01 2.46886924e-01 2.29655001e-02
-1.19661653e+00 -3.85775231e-02 -1.88441947e-01 1.64072320e-01
6.25024661e-02 -7.50375390e-02 -5.55184543e-01 1.43879935e-01
4.50820655e-01 5.18891573e-01 -6.01609826e-01 5.42616248e-01
-5.98305106e-01 1.38573304e-01 -9.57377553e-02 4.28113103e-01
-8.70193318e-02 -3.07493716e-01 5.43865263e-01 5.52752316e-01
1.64382741e-01 1.58526272e-01 -8.04335054e-04 1.31302118e+00
-8.00924122e-01 -7.92920664e-02 -7.23178983e-01 1.70099795e-01
5.19430339e-01 1.58736420e+00 -3.30164492e-01 -1.11239143e-01
-8.95789359e-03 1.11160433e+00 2.69623280e-01 4.79719102e-01
-1.00893521e+00 -5.52923203e-01 1.19404531e+00 3.28022122e-01
1.56173125e-01 2.58535683e-01 2.33084530e-01 -9.57832396e-01
-7.79367685e-02 -1.12657797e+00 -1.39515966e-01 -8.69918585e-01
-1.35347879e+00 8.71143878e-01 -1.71687722e-01 -1.02194643e+00
-3.74261856e-01 -4.74023998e-01 -1.11118388e+00 9.39865708e-01
-1.13293564e+00 -1.69971859e+00 -6.54769242e-01 8.12399924e-01
1.13561235e-01 -2.14361921e-01 8.33403528e-01 1.93077490e-01
-8.47686589e-01 1.21951842e+00 -3.18166196e-01 5.70061684e-01
9.09737170e-01 -7.11789250e-01 6.32629037e-01 7.40607917e-01
-1.59028172e-02 6.72092974e-01 4.72809523e-01 -3.97390693e-01
-1.34721875e+00 -1.40698481e+00 1.05490744e-01 -4.66746211e-01
3.93655330e-01 -6.65137053e-01 -7.41305411e-01 8.38755965e-01
1.65087640e-01 3.17682296e-01 3.30979466e-01 -3.65228444e-01
-4.85806435e-01 -4.24566597e-01 -1.52400076e+00 9.44897354e-01
1.13705516e+00 -7.30538309e-01 -2.48442575e-01 7.53760934e-02
8.47545445e-01 -5.81219196e-01 -5.33505559e-01 4.14857060e-01
7.70424187e-01 -7.03592420e-01 1.08913207e+00 -6.76563859e-01
3.93792093e-01 -2.44413063e-01 1.84701249e-01 -1.49928617e+00
-5.05314171e-01 -1.09730005e+00 1.57800496e-01 1.50908399e+00
1.52512580e-01 -7.45113552e-01 8.27480316e-01 7.04471111e-01
1.47194639e-01 -7.58826375e-01 -7.82139421e-01 -5.56389213e-01
2.37544507e-01 1.01637043e-01 1.19847083e+00 7.98614681e-01
-5.67951977e-01 3.31567645e-01 -5.62088609e-01 1.31500974e-01
6.46802962e-01 2.57731769e-02 1.04878247e+00 -8.99061382e-01
-4.61158864e-02 -3.47588241e-01 -4.73304629e-01 -7.76659071e-01
5.63436508e-01 -9.01364863e-01 1.49021998e-01 -8.95956516e-01
1.52397960e-01 -2.89154708e-01 -7.15579540e-02 8.06235015e-01
-1.64450601e-01 6.89418375e-01 2.89903790e-01 -4.07932736e-02
-1.75645389e-02 9.26703751e-01 1.88819158e+00 -2.29926612e-02
-1.03997886e-01 6.08836599e-02 -1.10054135e+00 7.37816513e-01
4.86904293e-01 -3.88487488e-01 -4.90022898e-01 -2.14686960e-01
-1.75159693e-01 -6.43776432e-02 5.92187405e-01 -7.72266388e-01
2.61703786e-02 -3.33184302e-01 8.51144016e-01 1.46103889e-01
5.56843877e-01 -7.35184252e-01 4.18007404e-01 1.99686080e-01
-1.92731351e-01 -3.71445939e-02 1.49353907e-01 3.16993117e-01
-7.95789510e-02 1.47711292e-01 1.34720731e+00 5.88757498e-03
-3.07062536e-01 1.05548704e+00 1.19707868e-01 5.49915358e-02
1.17646849e+00 -2.05728665e-01 -1.79138169e-01 -8.03565502e-01
-7.77404606e-01 6.87818527e-02 9.04370844e-01 7.18089163e-01
5.83113611e-01 -1.77807307e+00 -7.11473584e-01 8.08078289e-01
2.33912654e-02 1.01139173e-02 1.76842630e-01 2.70924240e-01
-3.17788988e-01 -3.68120253e-01 -5.76743901e-01 -3.26199412e-01
-1.19376612e+00 4.21304405e-01 7.82632053e-01 2.52915740e-01
-5.01825154e-01 9.23625350e-01 7.31087565e-01 -5.22729397e-01
-7.71015659e-02 3.18390012e-01 4.39589769e-02 -1.45420685e-01
6.12462044e-01 -1.13966912e-01 -3.50804299e-01 -8.76266718e-01
-2.14195400e-01 8.52483809e-01 -5.95967583e-02 -1.35406092e-01
1.26708841e+00 1.40739875e-02 -3.15045297e-01 -1.15678929e-01
1.41091156e+00 1.77217707e-01 -1.74945271e+00 -6.97788820e-02
-8.61321986e-01 -6.76444173e-01 -2.48041287e-01 -5.78865409e-01
-1.73299646e+00 6.62583053e-01 5.53264141e-01 -3.80144209e-01
1.20883429e+00 6.93548843e-02 6.94794893e-01 5.58139477e-03
2.41497993e-01 -8.12273562e-01 4.70796883e-01 2.15086147e-01
1.22640312e+00 -9.56156313e-01 -3.03300768e-01 -5.38769960e-01
-4.07892764e-01 9.14652824e-01 9.18028951e-01 -2.90019840e-01
7.28538275e-01 4.91593093e-01 4.01765436e-01 -1.50033936e-01
-4.60853457e-01 2.32564270e-01 3.27541649e-01 8.76311719e-01
8.13477784e-02 -6.11033514e-02 5.07448256e-01 5.20153165e-01
-4.42754894e-01 -2.89090425e-01 2.86013246e-01 5.15725672e-01
-3.65448743e-02 -1.29239905e+00 -2.24640578e-01 9.10147205e-02
-1.69776410e-01 1.41984880e-01 -7.17747450e-01 7.59046733e-01
3.03435504e-01 5.52005708e-01 1.57412827e-01 -4.37342286e-01
4.27373886e-01 1.85017847e-02 6.93096042e-01 -7.31284142e-01
-3.66712213e-01 -1.01912983e-01 -5.06153762e-01 -7.37637758e-01
-2.83033252e-02 -4.67878640e-01 -1.10138047e+00 -4.90197748e-01
-2.39973199e-02 -2.98626333e-01 4.18954134e-01 6.78086758e-01
4.70629275e-01 4.77144599e-01 1.37565172e+00 -1.12304318e+00
-4.73315865e-01 -7.86255121e-01 -5.65745831e-01 6.93061948e-01
5.23504257e-01 -6.54113829e-01 -4.37060148e-01 2.08023608e-01] | [12.669669151306152, -0.1180625930428505] |
b48f083e-ae0c-4b2d-a83a-73365cd09d37 | context-aware-classification-of-legal | 2304.02787 | null | https://arxiv.org/abs/2304.02787v2 | https://arxiv.org/pdf/2304.02787v2.pdf | Context-Aware Classification of Legal Document Pages | For many business applications that require the processing, indexing, and retrieval of professional documents such as legal briefs (in PDF format etc.), it is often essential to classify the pages of any given document into their corresponding types beforehand. Most existing studies in the field of document image classification either focus on single-page documents or treat multiple pages in a document independently. Although in recent years a few techniques have been proposed to exploit the context information from neighboring pages to enhance document page classification, they typically cannot be utilized with large pre-trained language models due to the constraint on input length. In this paper, we present a simple but effective approach that overcomes the above limitation. Specifically, we enhance the input with extra tokens carrying sequential information about previous pages - introducing recurrence - which enables the usage of pre-trained Transformer models like BERT for context-aware page classification. Our experiments conducted on two legal datasets in English and Portuguese respectively show that the proposed approach can significantly improve the performance of document page classification compared to the non-recurrent setup as well as the other context-aware baselines. | ['Dell Zhang', 'Grace E. Lee', 'Martina Forster', 'Pavlos Fragkogiannis'] | 2023-04-05 | null | null | null | null | ['document-image-classification'] | ['computer-vision'] | [ 7.69668698e-01 -1.62141085e-01 -2.80361980e-01 -2.56785721e-01
-1.23980653e+00 -6.61206722e-01 9.59416687e-01 3.84747595e-01
-3.92212152e-01 6.78913116e-01 3.35899740e-01 -6.73804283e-01
-1.86971456e-01 -5.15857399e-01 -5.65607488e-01 -7.92476714e-01
2.85230249e-01 2.72565782e-01 5.60906947e-01 1.31780908e-01
9.26192224e-01 2.64082938e-01 -1.56090748e+00 7.64557302e-01
8.78949106e-01 9.80102718e-01 6.56672060e-01 7.53166378e-01
-8.60647500e-01 8.53055060e-01 -7.56741345e-01 -3.33045065e-01
-8.31686705e-02 -4.24133986e-01 -9.51565802e-01 3.39083642e-01
3.77228081e-01 -8.23511109e-02 -1.41983345e-01 8.49668860e-01
4.02497768e-01 -9.81483310e-02 8.34784746e-01 -6.04790568e-01
-6.18735850e-01 6.99647248e-01 -9.61725295e-01 5.25951087e-01
2.90189236e-01 -6.69707775e-01 9.67876375e-01 -8.22927654e-01
6.79241061e-01 9.43310261e-01 5.68021536e-01 7.74172088e-03
-6.83994234e-01 -3.05545777e-01 3.60165119e-01 4.90289867e-01
-1.02601278e+00 -3.20401251e-01 8.15334141e-01 -2.18534082e-01
1.06568766e+00 2.49373019e-01 3.32442760e-01 1.15684569e+00
3.17677647e-01 9.67761636e-01 1.30681598e+00 -9.32400644e-01
-9.23234448e-02 3.64211768e-01 5.80347955e-01 5.28977275e-01
4.50193994e-02 -6.74353540e-01 -4.16002214e-01 3.27266231e-02
3.25575173e-01 9.30374023e-03 -2.56195277e-01 -2.81792879e-01
-9.71757352e-01 5.22733510e-01 -4.97681387e-02 7.25287437e-01
-3.71498764e-01 -4.39447880e-01 8.25726330e-01 -2.27969997e-02
4.27396625e-01 1.29172891e-01 -2.23987579e-01 -3.41811150e-01
-1.23572624e+00 2.77964119e-02 7.43498623e-01 8.47050428e-01
4.16977346e-01 -3.35079372e-01 -3.81424427e-01 1.12759316e+00
9.28133875e-02 1.88667849e-01 8.71510863e-01 -2.83920318e-01
1.03858042e+00 5.46233416e-01 4.72430885e-02 -8.67730021e-01
-8.33250582e-02 -4.64132965e-01 -5.23804963e-01 -3.24178666e-01
3.48040573e-02 2.98664480e-01 -1.26348317e+00 9.94349420e-01
1.60065722e-02 -2.10883021e-01 2.52390891e-01 3.59619975e-01
6.11102283e-01 1.00474787e+00 9.74648595e-02 -5.11374295e-01
1.66255462e+00 -1.33708012e+00 -1.04063678e+00 -2.07508892e-01
5.65364420e-01 -1.18491328e+00 1.26866317e+00 3.69035065e-01
-9.37927663e-01 -3.47611159e-01 -1.03297818e+00 -4.74256314e-02
-6.79908216e-01 4.65458900e-01 4.53625649e-01 7.81022131e-01
-7.64900327e-01 1.50330782e-01 -5.40030837e-01 -6.63712800e-01
3.20065439e-01 -7.43502676e-02 -1.27056435e-01 -1.24671265e-01
-8.83448660e-01 7.21749485e-01 3.33543599e-01 1.53575435e-01
-5.21631479e-01 -2.12765843e-01 -4.60932255e-01 5.58387451e-02
6.12566352e-01 -5.39090633e-02 1.35341704e+00 -9.22710538e-01
-1.20637000e+00 6.84436321e-01 -4.51578736e-01 -5.06361067e-01
5.04317939e-01 -3.55168521e-01 -4.77783591e-01 2.93555379e-01
1.60897553e-01 1.84118778e-01 8.84510875e-01 -1.22687948e+00
-8.09567094e-01 -4.22845751e-01 1.12521954e-01 5.68060160e-01
-7.46704400e-01 2.00147659e-01 -1.01414526e+00 -8.31041634e-01
3.63408551e-02 -7.95973361e-01 1.35034308e-01 -4.97138619e-01
-5.27249694e-01 -2.81246483e-01 1.40095329e+00 -9.58878875e-01
1.41719306e+00 -2.08456612e+00 -3.06944460e-01 2.22524345e-01
-4.90565419e-01 4.09635186e-01 1.64404567e-02 6.82990730e-01
-5.59946671e-02 2.06874728e-01 -1.31783530e-01 -2.39944249e-01
-1.70079947e-01 2.19804525e-01 -6.71730757e-01 6.01974577e-02
-1.28074035e-01 6.37593031e-01 -4.75597650e-01 -9.57730055e-01
-1.95336509e-02 6.36773944e-01 -4.65932488e-02 -5.64967692e-02
-7.73035213e-02 -7.09079728e-02 -5.95324874e-01 4.97036815e-01
4.35697645e-01 -3.01298589e-01 3.88298512e-01 -1.71281889e-01
-1.08399279e-01 8.35635364e-01 -1.08211613e+00 1.47201788e+00
-5.28376877e-01 7.20890164e-01 -2.48479918e-01 -1.26839352e+00
6.52027726e-01 5.46957195e-01 2.18444109e-01 -9.87961948e-01
-9.47083384e-02 2.34038472e-01 -2.75700092e-01 -5.66398919e-01
8.61657739e-01 2.74693698e-01 8.38892013e-02 5.05510449e-01
-4.37628210e-01 3.52998823e-01 5.98871827e-01 1.61894828e-01
9.90571737e-01 1.91419363e-01 1.67994574e-01 3.92105151e-03
8.68784726e-01 3.30509916e-02 1.96956351e-01 9.02048230e-01
1.45370916e-01 5.54872572e-01 3.85286719e-01 3.46258394e-02
-1.09253252e+00 -6.92881942e-01 -2.99945474e-01 9.73035932e-01
-9.78508145e-02 -4.44231123e-01 -9.12690938e-01 -9.78274286e-01
-2.44937763e-01 5.50936520e-01 -4.38251048e-01 2.75394440e-01
-7.68238246e-01 -8.14469099e-01 4.20370758e-01 6.94022179e-01
8.20337415e-01 -9.71249998e-01 -5.22872865e-01 2.08336055e-01
-2.62354434e-01 -1.21892738e+00 -4.56345648e-01 3.88455003e-01
-8.49551260e-01 -9.25295413e-01 -1.14099038e+00 -1.11654723e+00
6.94776773e-01 4.65509772e-01 6.38779998e-01 -4.58847955e-02
-1.45457014e-01 3.79664719e-01 -5.62637210e-01 -2.00581849e-01
-3.15185815e-01 5.34703910e-01 -4.42351729e-01 3.25333513e-02
3.38345230e-01 -2.42647931e-01 -3.11464697e-01 -3.34001891e-03
-1.01059043e+00 1.44977644e-01 9.77187157e-01 8.30581546e-01
4.23455894e-01 5.31747162e-01 5.00751972e-01 -1.41480017e+00
9.02000546e-01 -1.06410436e-01 -3.74435157e-01 8.15503895e-01
-9.84058440e-01 2.03442693e-01 5.32088995e-01 -3.14585239e-01
-1.46189630e+00 -2.66269088e-01 1.10327050e-01 2.17518285e-02
8.61890707e-03 8.23220849e-01 -7.96966553e-02 2.21030965e-01
1.81725487e-01 6.17297769e-01 -4.44555551e-01 -7.47436345e-01
6.25136271e-02 1.26381075e+00 5.15273452e-01 -4.52867597e-01
5.67679465e-01 1.47431806e-01 -2.48386234e-01 -7.43220925e-01
-7.85465062e-01 -7.66176403e-01 -7.12983727e-01 -1.29168749e-01
7.72625446e-01 -6.37360334e-01 -2.11053357e-01 4.16620165e-01
-9.98502076e-01 1.80910334e-01 1.74804091e-01 2.37198099e-01
-1.87044039e-01 9.03023064e-01 -6.66806638e-01 -9.73176241e-01
-4.75731283e-01 -1.06243730e+00 1.05622756e+00 5.05877249e-02
-7.17103779e-02 -8.51007342e-01 -1.37327999e-01 6.61057472e-01
3.03899676e-01 -3.82210970e-01 1.32828867e+00 -6.42684221e-01
-3.81741405e-01 -4.59907055e-01 -3.14076960e-01 3.66159827e-01
4.82667685e-01 -1.35528585e-02 -8.51976991e-01 -7.58694410e-02
-7.58422539e-02 -1.54852301e-01 9.34966624e-01 5.45003787e-02
1.24384594e+00 -4.77080256e-01 -5.78283548e-01 -3.70957144e-02
1.35623932e+00 6.85781658e-01 7.67210245e-01 7.94253528e-01
5.36299467e-01 5.48556387e-01 7.97432482e-01 2.66431659e-01
2.86460668e-01 6.80902600e-01 -1.94320604e-01 1.30989194e-01
-3.25040445e-02 -1.77227676e-01 3.74358088e-01 1.08722281e+00
1.08277336e-01 -5.10218918e-01 -9.72681105e-01 6.55482054e-01
-1.63573825e+00 -9.79003608e-01 5.01158983e-02 2.03885388e+00
9.80366647e-01 5.19140065e-01 -2.39686131e-01 6.09124541e-01
8.08922172e-01 4.05006647e-01 -2.57910401e-01 -6.35924459e-01
-1.78332344e-01 -9.89175960e-02 6.13849640e-01 2.97739625e-01
-1.30004382e+00 7.18243062e-01 6.28431606e+00 9.50502574e-01
-1.17220759e+00 -3.38509399e-03 5.16137123e-01 1.68801963e-01
-1.42450780e-01 -2.58297417e-02 -1.17851532e+00 5.44056237e-01
1.05753064e+00 -7.70430937e-02 5.10885939e-03 8.67706418e-01
1.09997526e-01 -3.23650301e-01 -6.95397139e-01 7.57563531e-01
4.99476492e-01 -1.21692562e+00 4.26134825e-01 -1.62889045e-02
6.32068217e-01 -4.31352377e-01 2.63577282e-01 2.03611091e-01
-2.96938241e-01 -6.42147124e-01 7.51261055e-01 2.83918768e-01
3.82911503e-01 -6.44461691e-01 8.95751834e-01 3.27273846e-01
-1.05813408e+00 -5.81143089e-02 -1.65314123e-01 3.89072835e-01
-4.09371639e-03 5.32528281e-01 -9.23452914e-01 7.42650926e-01
8.12541485e-01 6.85095966e-01 -8.03798854e-01 8.95453691e-01
1.73750315e-02 6.92678154e-01 -8.46679136e-02 -1.46877944e-01
5.02989173e-01 -9.16363448e-02 1.38886616e-01 1.44843113e+00
4.19086874e-01 -2.38537982e-01 -1.47735989e-02 8.99797529e-02
-1.64373219e-02 4.54711735e-01 -4.39414203e-01 -3.60897124e-01
2.01294586e-01 1.15084898e+00 -1.11153412e+00 -6.58254027e-01
-7.66525805e-01 1.13200819e+00 -1.87661517e-02 4.01536554e-01
-5.84777713e-01 -7.31939435e-01 -1.47182077e-01 2.34063163e-01
7.36758709e-01 -1.68057978e-01 -1.95121124e-01 -9.97491896e-01
6.09456539e-01 -9.70463634e-01 4.97057050e-01 -8.85405242e-01
-7.96176732e-01 7.84336686e-01 -4.07179408e-02 -1.16374183e+00
-4.82005179e-01 -5.00832319e-01 -3.33508521e-01 6.78674459e-01
-1.86244810e+00 -1.17178905e+00 -3.77290435e-02 4.51215357e-01
9.69514549e-01 -1.14092208e-01 6.52888775e-01 4.08264309e-01
-6.52710676e-01 3.62362355e-01 4.59645629e-01 1.96532995e-01
8.26464176e-01 -1.15840769e+00 9.27887857e-02 9.45948601e-01
3.15876186e-01 7.52101839e-01 6.33957684e-01 -7.78243244e-01
-1.28954518e+00 -7.68525302e-01 1.30094683e+00 -1.80512428e-01
5.04579425e-01 -3.02167952e-01 -1.11070299e+00 6.67197764e-01
6.52444184e-01 -3.94819021e-01 5.86180687e-01 2.98448317e-02
-3.85001332e-01 -3.29366118e-01 -7.48903990e-01 5.82914591e-01
6.64127767e-01 -7.87238777e-01 -9.11066949e-01 3.69737446e-01
1.87607959e-01 -2.60512859e-01 -6.42647207e-01 1.81907266e-01
4.51664805e-01 -6.06756985e-01 7.53730893e-01 -2.11024716e-01
3.07794094e-01 -2.61904895e-01 3.63722555e-02 -7.01695800e-01
1.18142545e-01 -1.73798099e-01 7.28801414e-02 1.70922995e+00
5.04152000e-01 -4.65068668e-01 7.57838726e-01 2.73380339e-01
-4.50289994e-02 -5.90470493e-01 -8.14261258e-01 -6.42717838e-01
-3.10087293e-01 -2.83576369e-01 3.75135005e-01 6.58224404e-01
8.97518918e-02 4.50880855e-01 -1.87144190e-01 -1.21785346e-02
2.82436758e-01 2.74499923e-01 3.82144570e-01 -9.14011955e-01
2.88416259e-03 -3.97422552e-01 -2.00012863e-01 -1.06843257e+00
7.65260532e-02 -7.99697280e-01 1.14475824e-01 -1.93166196e+00
5.91879129e-01 -3.19207728e-01 -4.32186842e-01 5.82645774e-01
-1.05293602e-01 4.58746515e-02 4.08017114e-02 5.75224698e-01
-7.06798911e-01 2.02355042e-01 9.06784475e-01 -3.05027276e-01
3.81480753e-02 -8.63830224e-02 -4.80885625e-01 4.31074530e-01
5.51313579e-01 -4.03986841e-01 -5.84801137e-01 -3.18172455e-01
8.63216743e-02 -1.09118775e-01 -2.19446067e-02 -9.43896651e-01
2.63934791e-01 2.00805590e-01 3.79640400e-01 -1.10541415e+00
1.17669456e-01 -8.25907886e-01 -2.96883762e-01 2.00144947e-01
-5.51470697e-01 3.58505517e-01 1.72487825e-01 6.09266877e-01
-5.02395332e-01 -6.80398583e-01 3.28848541e-01 -1.54621512e-01
-7.48913288e-01 -2.46970028e-01 -8.35449398e-01 -2.38307595e-01
9.63640928e-01 -2.65626222e-01 -3.24704051e-01 -2.65907437e-01
-3.19391817e-01 -1.53716892e-01 3.97270143e-01 6.99465036e-01
4.23311055e-01 -1.00545359e+00 -1.99622571e-01 4.77211103e-02
8.56962055e-02 -3.89185965e-01 1.15662850e-01 7.14693367e-01
-4.88671392e-01 1.12894022e+00 1.56241041e-02 -4.35066998e-01
-1.72543395e+00 4.98457372e-01 -1.10604927e-01 -7.18492627e-01
-7.71233737e-01 2.53093332e-01 -4.89305109e-02 -3.52524035e-02
5.02940774e-01 -3.37817252e-01 -6.99574649e-01 3.69746983e-01
5.90186357e-01 1.47342943e-02 4.17698562e-01 -5.27061343e-01
-2.16669649e-01 7.87281394e-01 -6.94634914e-01 -2.13271782e-01
1.15114236e+00 -3.71969491e-01 -1.12704895e-01 7.17918396e-01
1.23842108e+00 4.07496765e-02 -7.13180661e-01 -3.84759873e-01
5.57669699e-01 -3.07809025e-01 1.38144597e-01 -7.77485430e-01
-9.23655093e-01 9.60033238e-01 5.41295111e-01 3.18343639e-01
1.24733078e+00 -2.46650457e-01 7.01312482e-01 5.93464315e-01
1.87374815e-01 -1.43651068e+00 -1.63657125e-02 4.21225667e-01
7.18707740e-01 -9.43044603e-01 1.11661337e-01 -3.76155704e-01
-6.26179039e-01 1.14880025e+00 1.72629878e-01 2.96246916e-01
4.67985064e-01 3.07929575e-01 1.73607409e-01 7.08080009e-02
-7.36269534e-01 -9.28075984e-03 3.31536055e-01 2.36641169e-01
6.84355140e-01 -5.57179511e-01 -7.78265238e-01 3.61841559e-01
-5.84110357e-02 2.00581513e-02 5.60933590e-01 1.49688387e+00
-3.14789861e-01 -1.38942456e+00 -4.40565377e-01 6.46412730e-01
-1.07315183e+00 -3.12801003e-01 -2.71880478e-01 7.48600781e-01
-3.78071487e-01 7.08877683e-01 7.34852552e-02 1.12195209e-01
1.41327366e-01 5.69458663e-01 2.68989444e-01 -5.77134073e-01
-5.82435131e-01 3.73242021e-01 3.14081818e-01 -1.31458685e-01
-7.49856293e-01 -1.04882240e+00 -1.12099218e+00 1.38440624e-01
-3.45905960e-01 4.47364718e-01 9.26278710e-01 1.03030014e+00
2.55178720e-01 6.05131447e-01 4.28596944e-01 -4.28033799e-01
-4.25933570e-01 -1.08523071e+00 -5.34344137e-01 2.10626990e-01
1.46431178e-01 -4.25790757e-01 -1.85071066e-01 3.07137102e-01] | [11.66153621673584, 2.778660297393799] |
109b58d5-50d1-48f0-9f72-52c50b7c1573 | temporal-data-meets-llm-explainable-financial | 2306.11025 | null | https://arxiv.org/abs/2306.11025v1 | https://arxiv.org/pdf/2306.11025v1.pdf | Temporal Data Meets LLM -- Explainable Financial Time Series Forecasting | This paper presents a novel study on harnessing Large Language Models' (LLMs) outstanding knowledge and reasoning abilities for explainable financial time series forecasting. The application of machine learning models to financial time series comes with several challenges, including the difficulty in cross-sequence reasoning and inference, the hurdle of incorporating multi-modal signals from historical news, financial knowledge graphs, etc., and the issue of interpreting and explaining the model results. In this paper, we focus on NASDAQ-100 stocks, making use of publicly accessible historical stock price data, company metadata, and historical economic/financial news. We conduct experiments to illustrate the potential of LLMs in offering a unified solution to the aforementioned challenges. Our experiments include trying zero-shot/few-shot inference with GPT-4 and instruction-based fine-tuning with a public LLM model Open LLaMA. We demonstrate our approach outperforms a few baselines, including the widely applied classic ARMA-GARCH model and a gradient-boosting tree model. Through the performance comparison results and a few examples, we find LLMs can make a well-thought decision by reasoning over information from both textual news and price time series and extracting insights, leveraging cross-sequence information, and utilizing the inherent knowledge embedded within the LLM. Additionally, we show that a publicly available LLM such as Open-LLaMA, after fine-tuning, can comprehend the instruction to generate explainable forecasts and achieve reasonable performance, albeit relatively inferior in comparison to GPT-4. | ['Yanbin Lu', 'Zongyi Liu', 'Shujing Dong', 'Yuan Ling', 'Zheng Chen', 'Xinli Yu'] | 2023-06-19 | null | null | null | null | ['knowledge-graphs'] | ['knowledge-base'] | [-5.47439933e-01 1.32839605e-01 -1.75235763e-01 -3.59367192e-01
-1.10960543e+00 -8.51828396e-01 7.07812667e-01 -5.42028202e-03
3.19786608e-01 6.36981487e-01 3.94852340e-01 -1.08570004e+00
-9.97528136e-02 -9.36695993e-01 -1.03352356e+00 -1.71937302e-01
-3.09993804e-01 5.65823257e-01 -1.41358063e-01 -3.99287075e-01
4.25991118e-01 1.54073685e-01 -1.28836203e+00 5.13316453e-01
6.81564450e-01 1.25409579e+00 -7.00485185e-02 6.62727058e-01
-6.25590920e-01 1.71634459e+00 -5.45608580e-01 -9.46955979e-01
2.70154029e-01 -1.21196885e-04 -6.61862373e-01 -2.91730434e-01
8.21323097e-02 -3.17963749e-01 8.27168375e-02 7.77466536e-01
1.46081001e-01 -1.66636974e-01 5.41464329e-01 -1.38880503e+00
-1.01346946e+00 1.20017779e+00 -4.65195388e-01 5.90672851e-01
1.19883947e-01 3.82038802e-01 1.23268890e+00 -7.45760322e-01
3.18926901e-01 1.41564739e+00 1.05922902e+00 2.29576290e-01
-9.67364669e-01 -6.03342831e-01 3.21578175e-01 1.32426590e-01
-8.07583451e-01 -1.14300698e-01 5.39970279e-01 -5.87270975e-01
1.65872681e+00 3.83464634e-01 4.91832942e-01 1.19444227e+00
6.99629247e-01 7.10195124e-01 1.17834079e+00 -4.26607102e-01
4.53099579e-01 7.78232440e-02 4.53135163e-01 5.46623826e-01
1.74635202e-02 3.68256450e-01 -6.01936638e-01 -6.02597296e-01
5.25021791e-01 2.67082691e-01 1.88379019e-01 4.07408148e-01
-1.17233622e+00 1.07621551e+00 2.01761335e-01 1.11354955e-01
-5.97361684e-01 2.89904982e-01 4.64545697e-01 4.60239202e-01
8.92213106e-01 5.29143572e-01 -1.19906282e+00 -2.81557500e-01
-9.97461796e-01 4.73091185e-01 1.38894987e+00 8.57121348e-01
3.34410727e-01 4.72485483e-01 -6.82059750e-02 1.83062270e-01
3.47495407e-01 6.52204037e-01 8.37994337e-01 -7.56936550e-01
7.13136613e-01 1.52779102e-01 1.62454203e-01 -8.54016125e-01
-3.58172894e-01 -5.87699711e-01 -5.36167264e-01 3.26674571e-03
3.15574527e-01 -5.73281884e-01 -6.19226873e-01 1.48953092e+00
-2.52427123e-02 5.93243062e-01 4.28424418e-01 5.94622672e-01
4.74605381e-01 8.61370027e-01 9.37568918e-02 -2.95449764e-01
1.55684984e+00 -1.16862357e+00 -6.29868090e-01 -4.58225876e-01
6.10477507e-01 -6.41412079e-01 8.33926916e-01 4.02307451e-01
-1.02104843e+00 -5.02530217e-01 -8.28956366e-01 2.99731195e-02
-5.63463748e-01 -2.93305278e-01 9.86681700e-01 2.50474632e-01
-7.60145664e-01 6.33326530e-01 -8.82442534e-01 2.00823784e-01
-7.00154752e-02 -1.40496880e-01 3.58609259e-01 2.99786031e-01
-1.54879642e+00 9.95138109e-01 3.94088954e-01 -3.91488448e-02
-5.25461853e-01 -1.17917705e+00 -6.15898073e-01 3.52099240e-01
3.17711323e-01 -9.19047356e-01 1.72691059e+00 -6.81369424e-01
-1.54263163e+00 1.93575546e-01 -1.26302987e-01 -1.21646130e+00
6.04499578e-01 -4.26581025e-01 -7.69354761e-01 -4.26977962e-01
-1.02578655e-01 9.17635038e-02 6.33499980e-01 -5.55408776e-01
-5.65648437e-01 -2.58026779e-01 -2.57585980e-02 -5.29297888e-01
3.00058778e-02 1.05248630e-01 8.72795358e-02 -1.02532351e+00
-2.72298485e-01 -7.40377963e-01 -3.32104921e-01 -9.68515515e-01
-4.96070504e-01 -2.48438761e-01 4.02178049e-01 -1.02148926e+00
1.43497407e+00 -1.78273380e+00 -3.64380956e-01 2.62719810e-01
-3.43262739e-02 -1.16371721e-01 4.44887653e-02 6.46623731e-01
-2.12466955e-01 3.77891093e-01 6.65207580e-02 -2.40365341e-01
4.62037653e-01 1.82198435e-01 -1.57000756e+00 -2.05710873e-01
2.41328239e-01 1.33888984e+00 -6.45898938e-01 -4.88719866e-02
-5.85235283e-02 2.09493473e-01 -3.56052488e-01 2.94044204e-02
-8.20237339e-01 2.56982213e-03 -4.61039513e-01 6.71921313e-01
2.31643334e-01 -7.05172718e-01 1.26186952e-01 1.39431983e-01
-1.24848403e-01 4.97753888e-01 -9.45684850e-01 1.14869416e+00
-5.77784538e-01 3.92966568e-01 -6.50015116e-01 -7.92660832e-01
1.00514197e+00 3.01798522e-01 -1.34336457e-01 -6.41720891e-01
-4.73907851e-02 3.84960681e-01 -3.49818796e-01 -4.26791698e-01
4.23125148e-01 -2.79802978e-01 -2.04190850e-01 7.81365156e-01
-2.20283926e-01 9.70310066e-03 9.10894722e-02 6.22286424e-02
9.27420795e-01 2.09086254e-01 3.55341852e-01 -1.76256165e-01
1.88778609e-01 2.66018599e-01 4.36312139e-01 1.02743971e+00
4.01694745e-01 1.90586969e-01 6.80646598e-01 -9.96326208e-01
-9.92544353e-01 -7.20548451e-01 6.61799014e-02 1.12412429e+00
-6.88182771e-01 -4.41322297e-01 -5.50408959e-01 -3.85247707e-01
4.24865127e-01 1.53442919e+00 -5.74976563e-01 3.16468656e-01
-3.15659106e-01 -1.02629983e+00 4.73325312e-01 7.69356608e-01
2.75465637e-01 -8.61274421e-01 -5.90142310e-01 6.31547928e-01
-5.19679971e-02 -1.12462199e+00 -2.01194569e-01 8.83200690e-02
-9.35888767e-01 -7.61349499e-01 -2.89606869e-01 -7.32790232e-02
4.14040945e-02 -1.77570015e-01 1.57430005e+00 -2.55748510e-01
-2.04810929e-02 1.63503766e-01 -4.37583961e-02 -1.04694057e+00
-6.72973454e-01 8.57896805e-02 -1.52408659e-01 -2.72410184e-01
5.88569641e-01 -5.63315749e-01 -1.92867950e-01 -1.00902893e-01
-7.33932078e-01 2.36495137e-01 4.97491717e-01 7.25701272e-01
3.73901963e-01 5.95051162e-02 7.63777673e-01 -9.81816649e-01
7.16712058e-01 -9.30677295e-01 -1.03514099e+00 6.00730598e-01
-9.73937988e-01 4.31695491e-01 6.43141448e-01 -3.30203146e-01
-1.09938848e+00 -5.43036222e-01 1.08999848e-01 -3.84852260e-01
2.91880190e-01 1.26737750e+00 5.36471248e-01 5.45480788e-01
4.61195856e-01 3.67578387e-01 6.08327687e-02 -5.28880000e-01
3.96977037e-01 3.94173712e-01 7.03818619e-01 -6.43424988e-01
7.48362124e-01 2.73029506e-01 -1.31094262e-01 -7.24043623e-02
-1.34371150e+00 1.43076450e-01 -2.09742382e-01 1.53101474e-01
6.08985960e-01 -1.21774316e+00 -7.75843561e-01 3.11917782e-01
-1.14014935e+00 -2.63487399e-01 -1.52955145e-01 5.54773211e-01
-5.51919401e-01 8.45106766e-02 -1.16199112e+00 -1.00488937e+00
-7.45793819e-01 -7.98194587e-01 9.04772997e-01 -1.56203890e-02
-5.76904595e-01 -1.54879057e+00 1.44751966e-01 4.06885475e-01
7.50060856e-01 3.83033365e-01 1.24510705e+00 -1.43093038e+00
-7.25826263e-01 -3.69897246e-01 6.85228128e-03 1.63349554e-01
-2.28682429e-01 -2.57995389e-02 -1.06170321e+00 1.43769644e-02
2.27195665e-01 -2.92773902e-01 7.76544392e-01 3.90119761e-01
1.01385164e+00 -7.43001580e-01 1.05755679e-01 4.38240260e-01
1.27006567e+00 6.57303333e-02 2.80865669e-01 6.16648078e-01
3.89967024e-01 4.87848222e-01 4.42784220e-01 7.88556159e-01
7.79234231e-01 1.32240444e-01 1.18629530e-01 2.66945392e-01
5.52299798e-01 -5.16452968e-01 6.48441017e-01 1.32820070e+00
-2.20055953e-01 1.06097357e-02 -1.12041485e+00 1.94572359e-01
-2.13756990e+00 -1.23804140e+00 -8.92127231e-02 1.84973025e+00
9.58624065e-01 4.35392201e-01 -1.23522073e-01 -2.71494865e-01
1.64561391e-01 8.86456668e-02 -6.90387905e-01 -5.12916327e-01
-3.04035395e-01 -1.38430260e-02 5.72294772e-01 6.06724977e-01
-9.18553889e-01 6.66681886e-01 6.93341827e+00 8.93704891e-01
-1.10079646e+00 2.95395833e-02 9.09049988e-01 -1.25012994e-01
-8.02339315e-01 8.86275396e-02 -1.10521257e+00 5.85757911e-01
1.92579985e+00 -6.49754643e-01 5.39336443e-01 1.08634579e+00
1.59078956e-01 4.14696991e-01 -1.19164431e+00 6.21369004e-01
-1.33274481e-01 -1.83556676e+00 1.43230468e-01 -1.81189664e-02
7.21739829e-01 2.32548296e-01 1.78091694e-02 8.57722700e-01
9.13466156e-01 -1.03525245e+00 1.10482812e+00 9.33556437e-01
4.11121473e-02 -7.69558132e-01 8.77315581e-01 6.67692244e-01
-1.02425206e+00 -4.76122707e-01 -3.08426261e-01 -5.37140369e-01
1.32564783e-01 8.02733898e-01 -8.51544201e-01 8.63099992e-01
5.77698171e-01 7.71208823e-01 -4.32379663e-01 3.47416312e-01
-1.02738850e-01 7.86555111e-01 -1.12517916e-01 -6.12717979e-02
3.57047915e-01 -2.22904701e-03 1.31274045e-01 1.13502312e+00
7.13873386e-01 1.86354652e-01 1.34729058e-01 1.11918151e+00
2.98886061e-01 -6.06378168e-02 -3.73198271e-01 -2.42570400e-01
3.90906364e-01 8.08957219e-01 -3.14016730e-01 -9.66787994e-01
-7.91579485e-01 3.45357925e-01 2.46491522e-01 4.47042167e-01
-1.03633630e+00 1.38718620e-01 5.05321324e-01 -2.39252031e-01
4.97175097e-01 -5.90615198e-02 -4.17958736e-01 -1.51490033e+00
-8.89832303e-02 -1.20314395e+00 6.93376184e-01 -1.25790417e+00
-1.78892386e+00 6.72615588e-01 1.20668925e-01 -9.86783504e-01
-1.06944454e+00 -6.60959303e-01 -6.61859512e-01 1.17387617e+00
-1.69711447e+00 -1.18572915e+00 3.03257793e-01 2.67350376e-01
8.19438100e-01 -4.99550939e-01 8.00216198e-01 -1.51543349e-01
-4.74148452e-01 8.38926956e-02 2.74091840e-01 5.94953038e-02
3.25845510e-01 -1.39705098e+00 1.20507741e+00 6.96668446e-01
2.23971218e-01 7.66945362e-01 8.90915930e-01 -8.15963387e-01
-1.63280404e+00 -1.12043202e+00 1.17809510e+00 -8.31964493e-01
1.49203861e+00 -1.15614243e-01 -1.24525476e+00 1.23164308e+00
3.12078595e-01 -2.60008454e-01 7.86493897e-01 1.63698524e-01
-7.25783944e-01 -1.29317511e-02 -6.78513110e-01 4.17759925e-01
4.62282926e-01 -6.74471080e-01 -1.13678300e+00 4.92807716e-01
1.11761403e+00 -5.52605808e-01 -1.12818468e+00 1.98019281e-01
5.37812293e-01 -8.90246928e-01 1.00965762e+00 -1.15605569e+00
8.24565589e-01 9.07416269e-02 -1.47263467e-01 -1.30859494e+00
-2.58984536e-01 -8.53321254e-01 -5.15775442e-01 1.32329023e+00
6.88677669e-01 -1.06965756e+00 1.50391227e-02 1.07968533e+00
1.09757200e-01 -6.64968908e-01 -5.60085654e-01 -8.75846505e-01
3.39506090e-01 -9.31225836e-01 1.32589352e+00 1.00996804e+00
2.37953160e-02 1.10369980e-01 -5.31317174e-01 3.91121984e-01
5.95178306e-01 8.08486402e-01 7.33998001e-01 -1.20365334e+00
-8.25464666e-01 -3.13143730e-01 2.02024892e-01 -6.81691706e-01
4.21799541e-01 -8.11989486e-01 -4.03050721e-01 -1.04793859e+00
1.33619040e-01 4.70836572e-02 -4.97169673e-01 5.28727233e-01
-1.11458033e-01 -4.26962852e-01 3.80239159e-01 4.73074168e-01
-2.31592000e-01 1.64337501e-01 7.72380650e-01 -9.96054783e-02
1.01943187e-01 5.55628203e-02 -7.48071134e-01 8.78489256e-01
8.07337761e-01 -2.40541235e-01 -3.11159641e-01 -3.21567267e-01
6.09622538e-01 6.27573431e-01 6.37376785e-01 -4.10707772e-01
3.99713278e-01 -2.18057781e-01 3.13688308e-01 -7.77324617e-01
-8.13943073e-02 -5.02246857e-01 5.87795138e-01 6.32904172e-01
-5.01909316e-01 6.93712592e-01 6.01585150e-01 6.92021489e-01
-4.45872635e-01 -2.67793592e-02 2.94916462e-02 -4.70431119e-01
-9.18016970e-01 -2.74295285e-02 -1.92971081e-01 1.70128569e-01
6.58285797e-01 4.07670557e-01 -7.01692641e-01 -4.91414487e-01
-6.66743100e-01 3.43892008e-01 -1.97602049e-01 6.26403570e-01
1.68442667e-01 -1.21931827e+00 -8.88097286e-01 2.80946702e-01
-1.77749202e-01 -4.85574573e-01 2.84403890e-01 7.82070041e-01
-2.93412596e-01 9.49379742e-01 1.26398027e-01 -3.10439140e-01
-5.93669295e-01 8.26931655e-01 1.21735252e-01 -8.07445765e-01
-6.00373089e-01 5.24578035e-01 2.03172490e-02 -4.39907432e-01
-3.45582254e-02 -1.00496697e+00 9.97353345e-02 1.68785468e-01
8.09021473e-01 1.86011031e-01 8.09038281e-02 1.54597566e-01
-9.62728560e-02 3.87492418e-01 1.01680122e-01 -1.13038860e-01
1.55080283e+00 -1.73094764e-01 -1.75422862e-01 9.49072778e-01
7.57033765e-01 -5.24936197e-03 -1.14107752e+00 -4.79254067e-01
5.27258039e-01 -9.16063488e-02 -1.39687821e-01 -1.01554680e+00
-8.88249934e-01 8.06140006e-01 -7.45433420e-02 7.30818212e-01
7.28866816e-01 -9.83174443e-02 1.03930211e+00 4.86882865e-01
4.40589786e-01 -7.67811120e-01 -3.47040057e-01 7.08245039e-01
9.74639475e-01 -1.15971851e+00 -1.66816369e-01 1.12000994e-01
-9.94543552e-01 1.38866150e+00 1.52514279e-01 1.13520995e-01
7.86296189e-01 5.88107586e-01 4.52744454e-01 -2.63447851e-01
-1.67399371e+00 3.47311616e-01 2.63800859e-01 -1.08564183e-01
3.51524323e-01 1.36570409e-01 2.61123598e-01 1.23269570e+00
-8.61488223e-01 2.72075236e-01 4.43986714e-01 6.84739828e-01
-3.46669286e-01 -7.06594706e-01 -5.46665013e-01 4.09560204e-01
-8.85879636e-01 -5.00446975e-01 5.15507646e-02 8.76408458e-01
-4.47620183e-01 7.49392390e-01 3.19773614e-01 -1.15435086e-01
1.35149911e-01 7.28008211e-01 -2.90225565e-01 -2.53228486e-01
-7.49798298e-01 1.14036903e-01 2.38888472e-01 -3.99116307e-01
-1.23255253e-01 -7.58094788e-01 -1.19354343e+00 -6.83871686e-01
2.09011182e-01 1.58665091e-01 5.18990219e-01 1.18060601e+00
6.64005399e-01 6.55033648e-01 5.25242209e-01 -5.98096728e-01
-1.48344135e+00 -8.85691702e-01 -6.01738572e-01 1.16261318e-01
3.58468413e-01 -1.72939762e-01 -5.17150998e-01 3.86599481e-01] | [6.756169319152832, 3.2318689823150635] |
c9558ce4-8600-41da-9526-74c375b897ff | deep-cross-modal-projection-learning-for | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Ying_Zhang_Deep_Cross-Modal_Projection_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Ying_Zhang_Deep_Cross-Modal_Projection_ECCV_2018_paper.pdf | Deep Cross-Modal Projection Learning for Image-Text Matching | The key point of image-text matching is how to accurately measure the similarity between visual and textual inputs. Despite the great progress of associating the deep cross-modal embeddings with the bi-directional ranking loss, developing the strategies for mining useful triplets and selecting appropriate margins remains a challenge in real applications. In this paper, we propose a cross-modal projection matching (CMPM) loss and a cross-modal projection classification (CMPC) loss for learning discriminative image-text embeddings. The CMPM loss minimizes the KL divergence between the projection compatibility distributions and the normalized matching distributions defined with all the positive and negative samples in a mini-batch. The CMPC loss attempts to categorize the vector projection of representations from one modality onto another with the improved norm-softmax loss, for further enhancing the feature compactness of each class. Extensive analysis and experiments on multiple datasets demonstrate the superiority of the proposed approach. | ['Huchuan Lu ', 'Ying Zhang'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['nlp-based-person-retrival'] | ['computer-vision'] | [ 2.91366637e-01 -2.91548908e-01 -3.34885478e-01 -4.99465197e-01
-1.01626682e+00 -3.40413749e-01 8.91980648e-01 8.88771787e-02
-4.88449335e-01 2.32324183e-01 3.57757151e-01 5.65519854e-02
-3.69976342e-01 -3.31466407e-01 -5.05032837e-01 -8.92957091e-01
1.87111899e-01 3.68242353e-01 -1.28971487e-01 4.63463008e-01
3.03876847e-01 3.34079087e-01 -1.55460286e+00 5.71591794e-01
8.56794715e-01 1.54754639e+00 2.10137159e-01 2.32835680e-01
-2.50862956e-01 4.27874684e-01 -4.68984284e-02 -6.95944190e-01
4.82569307e-01 -2.59620965e-01 -4.20423299e-01 1.40723392e-01
7.33228683e-01 1.79758653e-01 -4.52256322e-01 1.16063988e+00
6.79408371e-01 2.28251070e-01 1.05047262e+00 -1.66029191e+00
-8.39529276e-01 2.43566394e-01 -9.09725606e-01 -6.59849793e-02
1.78755224e-01 -1.12159373e-02 1.24210870e+00 -1.38188708e+00
4.20483559e-01 1.29764307e+00 5.48685193e-01 2.91525841e-01
-1.42953122e+00 -6.92373753e-01 9.45578739e-02 3.65298927e-01
-1.53830433e+00 -2.27584675e-01 7.90050983e-01 -7.73367703e-01
5.94746530e-01 3.21594924e-01 2.77611345e-01 9.96901333e-01
6.45952374e-02 9.51145172e-01 9.49290633e-01 -3.76827866e-01
-2.63141781e-01 7.74892986e-01 -1.16472565e-01 6.72899008e-01
-2.86275297e-02 -7.64557347e-02 -7.19955623e-01 -1.98745310e-01
3.89849782e-01 2.72314161e-01 -3.71179044e-01 -9.75650787e-01
-1.22563088e+00 8.89060318e-01 2.94209450e-01 1.76423386e-01
-1.50748566e-01 -8.05521309e-02 4.96574640e-01 -4.25433405e-02
3.26555490e-01 2.47712195e-01 -1.18741326e-01 8.74842610e-03
-9.70945179e-01 -9.02306810e-02 2.41748124e-01 8.60517740e-01
7.25548208e-01 -3.58078182e-01 -6.30200982e-01 1.27533877e+00
5.12282729e-01 5.32613456e-01 6.51749909e-01 -5.87535679e-01
8.90610218e-01 1.01270795e+00 -2.40460813e-01 -1.31517422e+00
-9.75267775e-03 -2.12220311e-01 -8.00202429e-01 7.88847506e-02
1.06533006e-01 1.57905459e-01 -5.62437654e-01 1.83264756e+00
3.40887517e-01 -1.94118768e-01 -1.42533243e-01 9.14709866e-01
5.75798988e-01 7.76173532e-01 2.18612835e-01 7.20853591e-03
1.20651984e+00 -8.71717513e-01 -5.20368814e-01 -2.50893712e-01
4.17723864e-01 -9.90886867e-01 1.36378169e+00 -7.03295693e-02
-9.62453842e-01 -5.30924261e-01 -1.02840483e+00 -1.93450615e-01
-1.74573615e-01 6.43236399e-01 4.21661556e-01 2.61007875e-01
-5.18152058e-01 3.62062275e-01 -5.71170986e-01 -2.30888620e-01
4.41393077e-01 3.15217674e-01 -7.48567760e-01 -2.63834983e-01
-9.99450028e-01 7.51213312e-01 3.88569444e-01 -2.27285046e-02
-4.20216233e-01 -1.03168249e+00 -8.29921722e-01 2.02307656e-01
5.68454750e-02 -4.39575970e-01 5.38332999e-01 -1.08770788e+00
-1.12214065e+00 1.20351839e+00 6.80179521e-02 -1.09423243e-01
7.41221309e-01 -6.95578679e-02 -3.49176317e-01 5.02229407e-02
2.66106069e-01 8.18314910e-01 9.51989055e-01 -1.25413978e+00
-7.06929922e-01 -6.57785475e-01 -2.83973873e-01 6.41607106e-01
-9.69921589e-01 -1.69935912e-01 -9.05710638e-01 -6.92276180e-01
2.56989866e-01 -8.50791931e-01 1.43561855e-01 4.90457237e-01
-4.17439669e-01 -3.93915385e-01 7.78868854e-01 -6.76981747e-01
1.27911413e+00 -2.43051291e+00 3.69214684e-01 1.65037036e-01
-1.29739702e-01 -7.93655927e-04 -2.09235683e-01 4.09568876e-01
-2.23098010e-01 -2.47327864e-01 -2.57350296e-01 -5.38370192e-01
3.37208092e-01 -7.79661536e-02 -3.61114115e-01 6.06611669e-01
2.59923548e-01 7.04089463e-01 -5.13009310e-01 -7.52297342e-01
2.09593862e-01 5.95060945e-01 -4.04086649e-01 3.16102266e-01
2.06653163e-01 -6.49735034e-02 -1.25063509e-01 3.51536721e-01
6.94149077e-01 -3.22086990e-01 1.06598571e-01 -6.64467096e-01
3.19367021e-01 -1.43157944e-01 -1.18677831e+00 1.65770888e+00
-4.53141719e-01 6.77946866e-01 -1.84226125e-01 -1.13653088e+00
8.26339364e-01 8.68296772e-02 6.00612581e-01 -8.43655467e-01
7.09560215e-02 5.30536994e-02 -3.84509474e-01 -5.29965758e-01
2.73750484e-01 -3.39647323e-01 -7.80783445e-02 5.18345773e-01
7.72340894e-02 3.82369071e-01 7.95118362e-02 2.62435138e-01
3.68928134e-01 3.85178551e-02 -1.03332117e-01 -1.33350849e-01
6.55727983e-01 -3.03557307e-01 4.15766597e-01 2.42926598e-01
-1.84421852e-01 8.07836890e-01 5.40071964e-01 -9.00187939e-02
-9.76642966e-01 -1.30402815e+00 -3.10659170e-01 1.10427570e+00
2.77958453e-01 -6.67161718e-02 -4.10998851e-01 -9.29793477e-01
2.85908878e-01 5.91254830e-01 -7.04532802e-01 -5.02467871e-01
-2.71395057e-01 -6.99620306e-01 2.68248767e-01 6.22504711e-01
2.35942096e-01 -7.00694978e-01 -2.77473312e-02 -2.84087479e-01
-3.20232153e-01 -1.04178405e+00 -1.00104511e+00 7.55036101e-02
-5.46902537e-01 -9.36690688e-01 -8.34932983e-01 -1.16544771e+00
9.22064900e-01 2.11559176e-01 7.03310370e-01 -3.96029055e-01
-3.61678690e-01 3.76423955e-01 -6.25069514e-02 4.16317284e-02
-2.31664293e-02 -1.79154888e-01 -5.42809116e-03 5.42077124e-01
5.83279073e-01 -3.23326766e-01 -6.34232581e-01 3.03664297e-01
-9.84878600e-01 1.55764908e-01 5.55516362e-01 9.61819470e-01
8.22381198e-01 -1.23004295e-01 1.08351290e-01 -4.97953743e-01
6.82850003e-01 -4.11014378e-01 -4.53522235e-01 6.14613950e-01
-8.48373830e-01 2.49116361e-01 4.74359423e-01 -7.22445309e-01
-8.55469406e-01 1.22190289e-01 2.81478196e-01 -7.99633920e-01
3.41932327e-01 5.09737015e-01 -4.08267468e-01 2.02193633e-01
3.03344578e-01 4.60048586e-01 6.71229735e-02 -2.51444429e-01
5.21073103e-01 7.45059609e-01 3.64611953e-01 -5.44882298e-01
8.45202267e-01 4.16772753e-01 -8.99173990e-02 -6.64007962e-01
-7.51497328e-01 -7.14795828e-01 -5.29856801e-01 -4.04810794e-02
8.10671985e-01 -7.83262730e-01 -4.89755511e-01 2.40273938e-01
-8.09259892e-01 2.88720518e-01 -2.46937379e-01 6.17711842e-01
-4.30890471e-01 7.09865093e-01 -5.04158773e-02 -6.89538479e-01
-4.22511548e-01 -1.17257619e+00 1.03881931e+00 1.57618999e-01
-3.23231332e-02 -9.23756003e-01 1.92145735e-01 4.57038164e-01
8.30865130e-02 -1.78240556e-02 1.08578074e+00 -5.81013620e-01
-3.10748905e-01 -6.51755691e-01 -6.92021906e-01 5.53838670e-01
2.80091874e-02 -4.78923880e-02 -8.18418562e-01 -4.87474114e-01
-2.05823317e-01 -4.57497150e-01 1.00708723e+00 3.93579483e-01
1.26876104e+00 -2.55488843e-01 -3.09035689e-01 7.18163311e-01
1.57266867e+00 1.42019754e-02 5.06707549e-01 2.37050533e-01
7.37533092e-01 6.76094115e-01 7.09548593e-01 3.67578745e-01
2.47242182e-01 8.48554790e-01 2.57478654e-01 6.80693462e-02
1.28121208e-02 -4.91767049e-01 3.99962187e-01 6.85511470e-01
6.25364900e-01 -1.16509125e-01 -7.53926337e-01 6.80079818e-01
-1.89602125e+00 -9.58660543e-01 2.88377732e-01 2.42119837e+00
7.59592652e-01 -6.48084655e-02 -6.65684193e-02 9.28378627e-02
1.00215864e+00 1.98164776e-01 -4.72857684e-01 -2.30635270e-01
-3.47236812e-01 -2.02639863e-01 2.17979595e-01 3.69301021e-01
-1.37499166e+00 4.94512975e-01 4.96905470e+00 1.22207820e+00
-1.20863867e+00 -1.67358760e-02 6.18547380e-01 -2.47806221e-01
-3.94624799e-01 -2.31564254e-01 -8.00539017e-01 6.67099357e-01
5.41477561e-01 -7.73962513e-02 2.27215886e-01 8.19064617e-01
-1.10499665e-01 1.62291571e-01 -1.22160506e+00 1.31589520e+00
3.19616973e-01 -1.21017206e+00 4.69747782e-01 -3.89365628e-02
7.68184602e-01 -1.19980328e-01 6.94619954e-01 1.71167836e-01
-2.46978715e-01 -9.74436283e-01 7.51103580e-01 4.90196198e-01
1.08778310e+00 -7.62587726e-01 4.68245924e-01 3.17788981e-02
-1.15775251e+00 -2.20397681e-01 -5.04209578e-01 6.14169300e-01
-1.32842049e-01 4.10410017e-01 -6.25744641e-01 3.85136217e-01
6.72954202e-01 7.25515723e-01 -5.79957187e-01 9.78441894e-01
1.91516146e-01 4.72920761e-02 -2.07289413e-01 1.48761347e-02
2.85968393e-01 -6.52492106e-01 4.42765951e-01 1.28270555e+00
2.46196106e-01 -5.90303957e-01 1.26475975e-01 8.17856491e-01
-1.88772202e-01 4.31082010e-01 -3.08411062e-01 -3.40907127e-01
4.24284518e-01 1.35900068e+00 -3.55793983e-01 -5.40761538e-02
-4.31051165e-01 1.15174854e+00 5.11065245e-01 3.13595384e-01
-8.61653149e-01 -4.58826572e-01 7.78968632e-01 3.70516777e-02
3.52481633e-01 2.36513868e-01 -5.79697967e-01 -1.02982509e+00
4.50639278e-01 -5.96673787e-01 6.17686749e-01 -5.61020076e-01
-1.81604588e+00 3.90577197e-01 -1.90844700e-01 -1.54556704e+00
2.17976689e-01 -6.43693149e-01 -5.10718703e-01 1.11197221e+00
-1.40177882e+00 -1.39887607e+00 -1.86863154e-01 5.62890410e-01
1.71455890e-01 -3.56112778e-01 5.38872838e-01 6.68158054e-01
-5.94012618e-01 1.17474377e+00 5.30403972e-01 2.52597034e-01
8.53275061e-01 -1.12576377e+00 -4.04623449e-01 6.65192306e-01
1.19320601e-01 4.71509904e-01 5.09471834e-01 -1.34743437e-01
-1.24181056e+00 -1.27991390e+00 8.90764236e-01 -1.77407652e-01
6.04028761e-01 -3.14584881e-01 -8.31388295e-01 4.36519593e-01
1.01374462e-01 4.14315499e-02 8.83388042e-01 -3.22989821e-02
-7.42589355e-01 -4.60898101e-01 -9.28053737e-01 6.98166251e-01
6.06953800e-01 -8.54046285e-01 -2.79002279e-01 3.73653173e-01
4.33905840e-01 2.40649609e-03 -9.10836160e-01 3.34188312e-01
8.20008159e-01 -6.91908777e-01 1.10204697e+00 -7.88900316e-01
7.11409390e-01 -2.69120246e-01 -5.76315522e-01 -1.00591290e+00
-3.59598368e-01 -4.80471030e-02 1.27821282e-01 1.49508369e+00
5.47233999e-01 -2.28720382e-01 7.59909332e-01 5.92011511e-01
1.25186935e-01 -1.04523015e+00 -1.10791850e+00 -6.03813112e-01
2.05442488e-01 -3.77738535e-01 1.43086955e-01 1.00135994e+00
1.37289733e-01 4.69850749e-01 -5.28570056e-01 7.10082278e-02
7.01385319e-01 3.08861524e-01 5.74847400e-01 -9.82089341e-01
-2.49984220e-01 -6.48704290e-01 -6.45331442e-01 -9.59800661e-01
2.09003404e-01 -1.28048801e+00 4.22894731e-02 -1.43709469e+00
8.32598805e-01 -4.18821514e-01 -6.57130837e-01 9.00898427e-02
-3.71767879e-01 1.95741564e-01 2.34767601e-01 2.40371019e-01
-5.84088385e-01 1.01517856e+00 9.93740082e-01 -4.93456632e-01
4.36268225e-02 -1.01160049e-01 -5.73426604e-01 5.23273528e-01
5.63754857e-01 -5.68085074e-01 -5.25759101e-01 -4.89316016e-01
1.62282154e-01 -1.82939067e-01 1.25889301e-01 -7.55716264e-01
2.97737122e-01 -1.56871125e-01 4.80480433e-01 -8.57893586e-01
5.52321315e-01 -1.07037425e+00 2.50908248e-02 2.72661150e-01
-7.45056808e-01 -4.11447063e-02 -4.13244441e-02 8.28477800e-01
-3.67694527e-01 -7.90705085e-02 1.15042293e+00 3.79830331e-01
-4.32523727e-01 5.11504054e-01 1.64959952e-01 3.53757620e-01
1.20723629e+00 -4.89906728e-01 -1.05921350e-01 -1.33915469e-01
-3.87085348e-01 3.86764437e-01 4.29602504e-01 7.25944519e-01
8.75908732e-01 -1.90898561e+00 -6.43402159e-01 3.81194144e-01
6.75774217e-01 -5.10688841e-01 3.90283078e-01 9.82904434e-01
-8.63089114e-02 4.90918577e-01 -1.97136670e-01 -7.53185272e-01
-1.41908729e+00 5.04173756e-01 4.45632100e-01 -3.46868575e-01
-4.20253187e-01 9.84881103e-01 6.49115384e-01 -6.31925583e-01
5.01590431e-01 2.17748359e-01 -2.15359092e-01 2.44591385e-01
4.94515181e-01 4.19613332e-01 -2.10814327e-02 -8.46325517e-01
-3.92590731e-01 5.60618460e-01 -9.48715284e-02 -2.37536341e-01
1.16121578e+00 -1.45980552e-01 -1.25470638e-01 4.33027685e-01
1.90935719e+00 -3.58105779e-01 -1.27612460e+00 -6.09287620e-01
-4.47371714e-02 -7.76028335e-01 8.64064470e-02 -6.34237409e-01
-1.16907597e+00 1.09245062e+00 9.96357620e-01 -1.96092933e-01
1.00812411e+00 5.54610742e-03 7.13093579e-01 -5.81379645e-02
-3.62420529e-01 -1.20607173e+00 3.42370838e-01 2.29523599e-01
1.02584553e+00 -1.30545557e+00 -4.71245870e-02 -1.18467808e-01
-1.02843571e+00 9.08890843e-01 6.16842389e-01 6.08534701e-02
8.50264490e-01 -1.88032568e-01 -2.95341527e-03 -1.11192591e-01
-5.41805148e-01 9.47326794e-02 7.69652426e-01 5.71210384e-01
3.30720931e-01 8.37565660e-02 -2.89238691e-01 6.02635205e-01
2.51463503e-01 -4.06954318e-01 -1.79091454e-01 5.03359258e-01
-2.36463666e-01 -9.66260672e-01 -2.44074449e-01 7.23064065e-01
-3.09815228e-01 -1.09289000e-02 -3.86013538e-01 4.49349284e-01
3.53351049e-02 5.82202435e-01 2.11715981e-01 -5.78641295e-01
4.03808087e-01 1.59420818e-01 4.88431066e-01 -3.56081218e-01
-1.97214037e-01 5.28623648e-02 -3.74839246e-01 -1.99159607e-01
-2.58860618e-01 -7.19546258e-01 -9.40600038e-01 -1.32635117e-01
-4.05434072e-01 3.75620350e-02 5.63821614e-01 6.15024626e-01
4.04076606e-01 1.84242669e-02 8.96765530e-01 -6.50420785e-01
-9.43874002e-01 -7.70696044e-01 -7.36416698e-01 9.89138246e-01
-4.96716276e-02 -7.14386642e-01 -4.59855914e-01 -7.08742589e-02] | [10.934236526489258, 1.2289223670959473] |
a7cde74e-626a-41db-9ef7-f984933e36e5 | logic-lm-empowering-large-language-models | 2305.12295 | null | https://arxiv.org/abs/2305.12295v1 | https://arxiv.org/pdf/2305.12295v1.pdf | Logic-LM: Empowering Large Language Models with Symbolic Solvers for Faithful Logical Reasoning | Large Language Models (LLMs) have shown human-like reasoning abilities but still struggle with complex logical problems. This paper introduces a novel framework, Logic-LM, which integrates LLMs with symbolic reasoning to improve logical problem-solving. Our method first utilizes LLMs to translate a natural language problem into a symbolic formulation. Afterward, a deterministic symbolic solver performs inference on the formulated problem. We also introduce a self-refinement stage, which utilizes the symbolic solver's error messages to revise symbolic formalizations. We demonstrate Logic-LM's effectiveness on four logical reasoning datasets: ProofWriter, PrOntoQA, FOLIO, and LogicalDeduction. Our results show significant improvement compared to LLMs alone, with an average performance boost of 62.6% over standard prompting and 23.5% over chain-of-thought prompting. Our findings suggest that Logic-LM, by combining LLMs with symbolic logic, offers a promising avenue for faithful logical reasoning. Code and data are publicly available at https://github.com/teacherpeterpan/Logic-LLM. | ['William Yang Wang', 'Xinyi Wang', 'Alon Albalak', 'Liangming Pan'] | 2023-05-20 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [ 2.38374025e-02 7.76105106e-01 -5.18844903e-01 -3.78467560e-01
-8.91135693e-01 -5.43762088e-01 4.59983349e-01 2.35051677e-01
3.32471400e-01 7.75053978e-01 1.47396117e-01 -1.03541553e+00
-2.14524254e-01 -7.92437911e-01 -9.12361383e-01 2.30638683e-01
2.76089013e-01 5.15652955e-01 3.04314673e-01 -3.49067628e-01
3.40192348e-01 3.05659145e-01 -1.04272199e+00 7.90528297e-01
1.31132102e+00 4.25365567e-01 -1.24245033e-01 8.08352530e-01
-4.14587200e-01 2.11902666e+00 -4.15651202e-01 -4.56984460e-01
-3.21410060e-01 -4.61293727e-01 -1.46503949e+00 -3.56433272e-01
4.62922662e-01 -5.59824824e-01 -2.60279477e-01 1.18071294e+00
-2.38379166e-01 -4.58000526e-02 1.02834769e-01 -1.51497102e+00
-4.79600281e-01 1.31067300e+00 -1.04598731e-01 -2.83478051e-01
1.18377984e+00 5.10348022e-01 1.11716282e+00 -4.84888792e-01
5.93118072e-01 1.78175855e+00 7.34857380e-01 6.42316997e-01
-1.44114864e+00 -6.38699651e-01 1.84472933e-01 7.57735252e-01
-1.47185481e+00 -3.22387993e-01 5.35266221e-01 -2.71147609e-01
1.47534478e+00 3.60600382e-01 6.14392221e-01 5.15726209e-01
3.74487996e-01 9.33520377e-01 1.35535538e+00 -9.17048812e-01
1.70486182e-01 1.70891747e-01 5.33778429e-01 1.21074104e+00
4.90096509e-02 -2.32689172e-01 -7.65803635e-01 -2.85464078e-01
3.57453585e-01 -5.25068521e-01 -4.42572795e-02 3.82998943e-01
-8.68366599e-01 6.73410535e-01 7.35562816e-02 5.06603494e-02
-1.89294010e-01 6.26575649e-01 1.97119206e-01 3.07060301e-01
-2.28348240e-01 7.21061826e-01 -1.89331457e-01 -3.43280405e-01
-1.07525814e+00 7.53959537e-01 1.31263685e+00 9.80123580e-01
2.46603549e-01 -1.21313781e-01 -1.63713560e-01 2.74863839e-01
5.76624691e-01 6.99892700e-01 1.33239683e-02 -1.62049067e+00
2.81160295e-01 8.39903474e-01 7.28720427e-02 -1.04127514e+00
-2.43459240e-01 1.67344525e-01 -2.39054292e-01 8.15068930e-02
3.42720807e-01 1.51799470e-01 -3.21183980e-01 1.69789577e+00
2.00639263e-01 3.73570532e-01 3.49417657e-01 6.35250688e-01
8.32177758e-01 7.50403404e-01 5.10217734e-02 -6.04244620e-02
1.31777155e+00 -1.08891904e+00 -6.83485389e-01 2.10542958e-02
8.13777328e-01 -4.82203186e-01 1.19996297e+00 9.68116939e-01
-1.57358623e+00 4.09671590e-02 -8.16089392e-01 -2.71759599e-01
2.89034341e-02 -1.40426069e-01 9.21275377e-01 -1.52756192e-03
-1.20271266e+00 2.90566087e-01 -8.69080007e-01 -2.09351834e-02
3.36300641e-01 2.39321202e-01 4.45016697e-02 -4.18262154e-01
-1.31090987e+00 9.83014166e-01 4.75294203e-01 -1.00817159e-01
-7.93500781e-01 -8.19674730e-01 -1.00136590e+00 1.48758665e-01
8.08729589e-01 -4.73843366e-01 2.03010941e+00 -4.73573178e-01
-1.78176403e+00 5.46529651e-01 -5.37977099e-01 -7.37208426e-01
4.56201792e-01 -2.88196683e-01 -3.42910260e-01 4.35764045e-01
3.20302665e-01 5.32328784e-01 1.81330055e-01 -8.59281838e-01
-2.49496371e-01 4.58719701e-01 7.68810630e-01 -2.37195700e-01
3.72517735e-01 4.30282950e-01 -2.26166666e-01 -2.01250970e-01
1.79860219e-01 -8.12525213e-01 1.21362386e-02 -7.93491751e-02
-6.07557058e-01 -6.78834140e-01 2.31776208e-01 -7.26195514e-01
1.57067299e+00 -1.68617809e+00 1.89018890e-01 2.62005866e-01
3.48842382e-01 3.70606899e-01 -5.51600456e-02 7.25773454e-01
1.53465299e-02 1.98212475e-01 -3.45764533e-02 2.20702410e-01
4.66685444e-01 3.17443311e-01 -9.46399093e-01 -1.04781248e-01
4.23446089e-01 1.33981729e+00 -1.23551619e+00 -9.54368830e-01
1.29248202e-01 -8.36200714e-02 -1.02004290e+00 8.49232897e-02
-1.35178053e+00 1.03607111e-01 -2.92801619e-01 8.43687892e-01
5.59295237e-01 -7.81413019e-01 6.07259095e-01 3.55995655e-01
1.07801616e-01 8.67389321e-01 -1.22597039e+00 1.57224095e+00
-4.09771562e-01 5.85302532e-01 -2.41461471e-01 -4.65807080e-01
4.45741326e-01 3.24751675e-01 -2.41625816e-01 -5.68280756e-01
-1.51563480e-01 2.49145955e-01 -3.57766561e-02 -7.19120741e-01
2.52817482e-01 -2.36132771e-01 -1.33221209e-01 7.45968878e-01
-2.85162449e-01 -7.02601254e-01 4.44067299e-01 7.89285779e-01
1.30076420e+00 4.35227841e-01 4.71838713e-01 -4.56337370e-02
8.91362488e-01 6.92462027e-01 5.00054121e-01 8.47069442e-01
1.76999509e-01 -2.53896117e-01 1.16623902e+00 -2.58511573e-01
-3.87084246e-01 -9.02189851e-01 1.72618210e-01 8.37395072e-01
5.39592048e-03 -1.11926889e+00 -6.66365623e-01 -4.16401744e-01
2.42978722e-01 1.51440322e+00 9.17714611e-02 -2.40886703e-01
-7.16401994e-01 2.66683251e-01 1.26400876e+00 6.12284303e-01
5.01893997e-01 -1.10445940e+00 -5.80757856e-01 1.47434235e-01
-3.73764157e-01 -1.23494279e+00 2.53410429e-01 -1.62160814e-01
-6.69952631e-01 -1.21811020e+00 4.80553240e-01 -4.61841106e-01
7.02508152e-01 -4.13453132e-02 9.76319075e-01 7.68083394e-01
1.50326282e-01 2.65458196e-01 -3.24652284e-01 -2.29133114e-01
-9.13303018e-01 -3.79110694e-01 -5.26944958e-02 -8.05292308e-01
3.22916567e-01 -4.04223382e-01 2.63831347e-01 -1.18126474e-01
-6.88571513e-01 7.95631468e-01 1.69766888e-01 6.78925335e-01
3.58657032e-01 2.91577458e-01 1.32029235e-01 -7.87506878e-01
7.67884493e-01 -2.21796811e-01 -9.35367346e-01 6.57101274e-01
-7.09571898e-01 4.21779871e-01 9.40904737e-01 -3.26188505e-01
-1.03068626e+00 -5.30275643e-01 1.59600109e-01 -2.24106163e-01
-2.14576684e-02 8.69749486e-01 2.82429993e-01 -6.47755116e-02
6.52888119e-01 1.68597952e-01 -1.54560013e-02 -4.46052887e-02
4.02893692e-01 4.25875902e-01 8.13853979e-01 -1.51699293e+00
9.94349301e-01 1.73954591e-02 3.24281454e-01 -2.25152865e-01
-1.11923409e+00 2.46689975e-01 -1.86484918e-01 -4.90733273e-02
4.32609677e-01 -8.96584868e-01 -1.40882850e+00 1.70712218e-01
-1.53710139e+00 -1.01424170e+00 -1.85133174e-01 2.75684625e-01
-5.87792695e-01 3.26970428e-01 -9.00258601e-01 -9.88953710e-01
-1.96431950e-01 -1.04526114e+00 7.85555780e-01 3.31763953e-01
-9.18726265e-01 -8.97473812e-01 -1.19012170e-01 7.50128806e-01
9.43057686e-02 9.92927253e-02 1.35551047e+00 -5.29863834e-01
-9.03154254e-01 -2.18336564e-02 -3.92896980e-01 2.88671881e-01
-4.12493944e-01 1.86930969e-01 -4.89904702e-01 2.34411478e-01
-4.70265299e-01 -9.16117609e-01 7.36381933e-02 -4.90301289e-02
1.19400644e+00 -8.02234530e-01 -9.05410796e-02 3.52201164e-01
1.06249416e+00 -1.71863977e-02 4.35083747e-01 3.19155365e-01
3.39047521e-01 1.26637027e-01 8.02040875e-01 2.91221023e-01
9.46393788e-01 3.58605087e-01 -9.44320932e-02 5.26613474e-01
-1.80407241e-01 -7.69445181e-01 5.30902147e-01 5.95196009e-01
9.93915126e-02 2.27760226e-01 -1.71852434e+00 1.95770279e-01
-2.09408736e+00 -1.06245983e+00 -2.15589896e-01 1.29036963e+00
1.54991579e+00 2.65956014e-01 -2.45007470e-01 2.30441734e-01
1.21059194e-01 -2.71912694e-01 -4.15055364e-01 -9.21952248e-01
1.32434785e-01 6.36848986e-01 -1.95320025e-01 1.28401041e+00
-3.38312179e-01 1.53215873e+00 6.08169365e+00 8.41804087e-01
-1.08420169e+00 -1.68228477e-01 -1.86893314e-01 -1.24290474e-01
-5.86809933e-01 5.61039627e-01 -8.53287697e-01 -3.50143202e-03
1.12608552e+00 -4.06905264e-01 1.02915919e+00 6.58137023e-01
2.98573971e-01 -4.98310834e-01 -1.29311419e+00 4.63429362e-01
-3.31712551e-02 -1.65460205e+00 9.39401165e-02 -5.34738719e-01
6.44579947e-01 -4.84958649e-01 -3.86442900e-01 6.47091806e-01
6.89781845e-01 -1.36354482e+00 1.12207484e+00 8.60518992e-01
6.21003509e-01 -5.30586660e-01 4.79905844e-01 8.47141445e-01
-8.34146321e-01 -4.97115366e-02 2.01855496e-01 -7.93633401e-01
7.15983361e-02 1.72245353e-01 -1.27549589e+00 4.39861059e-01
1.94495916e-01 5.37367284e-01 -6.50336683e-01 4.61481422e-01
-1.33480084e+00 8.81541848e-01 -3.48856002e-01 -3.25911820e-01
1.74127594e-01 1.72570243e-01 4.84546334e-01 1.28780448e+00
-3.44157785e-01 6.78686321e-01 2.35832095e-01 1.49501717e+00
2.25962281e-01 -4.28165197e-01 -1.76189050e-01 -5.04064143e-01
6.84480786e-01 8.66038680e-01 -3.35017622e-01 -7.64070570e-01
-1.89202026e-01 4.44798768e-01 3.26105416e-01 4.12392169e-01
-1.23684216e+00 -3.85367274e-01 2.02872649e-01 1.13074139e-01
-2.75554329e-01 -3.88838738e-01 -5.18491805e-01 -1.27939248e+00
3.79081331e-02 -1.51059437e+00 1.77130044e-01 -1.48668051e+00
-6.13703430e-01 6.97706118e-02 6.71029866e-01 -4.22617137e-01
-5.00207603e-01 -5.30952692e-01 -6.26591206e-01 8.65839064e-01
-1.47496688e+00 -1.27882600e+00 -2.06904903e-01 2.47279152e-01
2.84546018e-01 1.09486587e-01 9.20179248e-01 -2.26200685e-01
-4.64255661e-01 5.32072484e-01 -7.73957789e-01 2.17411295e-03
3.64811540e-01 -1.15495229e+00 8.14535096e-02 9.63050067e-01
-2.56136298e-01 1.21855032e+00 7.01393545e-01 -8.24786663e-01
-1.91086113e+00 -8.99731815e-01 1.55758107e+00 -4.55684751e-01
1.13597190e+00 -7.98975229e-02 -1.01421857e+00 1.19055974e+00
1.84321538e-01 -3.06306064e-01 4.34566289e-01 -7.28785023e-02
-6.83206260e-01 1.71788201e-01 -1.14830124e+00 9.10910428e-01
9.40511644e-01 -9.03350353e-01 -1.20253074e+00 6.57743514e-01
1.02021432e+00 -8.85281742e-01 -7.88300633e-01 2.14432091e-01
5.15160859e-01 -6.36286676e-01 6.76707864e-01 -9.35712576e-01
1.01762915e+00 -6.41137123e-01 -3.43232274e-01 -6.07452393e-01
-1.91661030e-01 -9.90027070e-01 -7.10193515e-01 1.06570113e+00
2.13427424e-01 -8.30446064e-01 2.61235327e-01 1.09847414e+00
8.26063454e-02 -1.06361580e+00 -5.26136458e-01 -6.72504187e-01
1.65047228e-01 -1.03905213e+00 6.95563078e-01 7.76523948e-01
9.39360142e-01 2.12148681e-01 2.27769703e-01 3.88694584e-01
4.28244740e-01 5.75061142e-01 8.84997189e-01 -7.03902721e-01
-7.33885646e-01 -6.23052359e-01 3.91518697e-02 -8.96492779e-01
1.00885379e+00 -1.46303225e+00 -6.32765517e-02 -1.76326716e+00
1.64284438e-01 -1.34356663e-01 1.77943543e-01 1.26445055e+00
1.31960690e-01 -3.61521482e-01 4.21452284e-01 -1.17189385e-01
-9.99861717e-01 1.80375036e-02 1.22894764e+00 -2.55834341e-01
-1.44581839e-01 -4.28075433e-01 -7.79690564e-01 1.03683555e+00
7.52785504e-01 -3.64687592e-01 -3.08151037e-01 -3.40812743e-01
7.22024798e-01 5.83332539e-01 8.52620065e-01 -9.93324399e-01
6.81317925e-01 -1.06522131e+00 -2.55120516e-01 -4.19998795e-01
-8.49148929e-02 -2.87954092e-01 1.15338862e-01 8.16250205e-01
-7.15906262e-01 -9.80537310e-02 6.73228562e-01 -2.06406638e-01
-2.05852538e-01 -3.56062502e-01 4.63804603e-01 -1.04765305e-02
-7.39849031e-01 -6.47716463e-01 -5.59247613e-01 2.81695187e-01
9.72387969e-01 3.61175746e-01 -6.03761554e-01 -3.96905988e-01
-4.83808637e-01 7.04943061e-01 4.04780686e-01 -1.17506310e-01
7.72515893e-01 -1.03732836e+00 -4.57217932e-01 -3.57870273e-02
-2.06715599e-01 3.21357995e-01 -1.67696863e-01 1.33286619e+00
-1.00869501e+00 8.66973758e-01 1.11381121e-01 -4.19932783e-01
-1.36816955e+00 2.91141689e-01 3.20337236e-01 -2.91647047e-01
-7.31123388e-01 8.96671355e-01 -4.78712887e-01 -5.21543205e-01
2.69188762e-01 -1.13325548e+00 4.56203848e-01 -6.95459485e-01
5.80410182e-01 3.86817992e-01 -4.70545053e-01 -3.49718407e-02
-7.56279945e-01 4.01460648e-01 1.80136383e-01 -2.77438074e-01
1.18996418e+00 3.48716497e-01 -8.46922755e-01 3.72311920e-01
3.93354774e-01 1.67321727e-01 -4.62854981e-01 -4.40029413e-01
2.15387642e-01 -1.47789508e-01 -1.95772529e-01 -1.29508996e+00
-8.15768447e-03 5.07098377e-01 -7.26235092e-01 -5.27703501e-02
9.59866047e-01 1.69612318e-01 7.59774446e-01 9.99373257e-01
5.85667431e-01 -6.28457427e-01 5.89799173e-02 1.01307893e+00
1.00609457e+00 -9.47921872e-01 3.61123532e-01 -4.41626847e-01
-4.43828970e-01 1.25851357e+00 7.10610449e-01 -9.93220583e-02
5.26343808e-02 6.72582984e-01 -2.25812882e-01 -2.34826609e-01
-1.34990191e+00 2.80840427e-01 7.91447237e-03 1.14516038e-02
4.04367954e-01 3.47326905e-01 -1.62168175e-01 8.76522541e-01
-6.28595471e-01 8.02197754e-01 6.57370329e-01 1.20982909e+00
-3.55629981e-01 -1.11234546e+00 -8.27999294e-01 2.36867648e-02
-1.63952172e-01 -2.01477721e-01 -4.85074639e-01 7.54989445e-01
-1.01290211e-01 1.12498558e+00 -4.27262336e-01 -3.27228487e-01
7.73942322e-02 5.40643930e-01 9.61153209e-01 -8.35857868e-01
-5.14221072e-01 -4.39626038e-01 5.24058521e-01 -9.00098860e-01
-6.95189610e-02 -5.00529230e-01 -2.23962951e+00 -9.88432229e-01
1.32818222e-02 2.45815545e-01 7.65784010e-02 1.17414856e+00
1.35031477e-01 5.86828828e-01 -1.19173601e-01 -1.50091261e-01
-9.12803829e-01 -4.09269869e-01 9.03115049e-02 -7.23389611e-02
3.79602700e-01 -3.12866598e-01 -1.33411556e-01 1.82478905e-01] | [9.299925804138184, 7.274547100067139] |
da8a6415-588b-4acd-b786-0251dab30192 | deepvqe-real-time-deep-voice-quality | 2306.03177 | null | https://arxiv.org/abs/2306.03177v1 | https://arxiv.org/pdf/2306.03177v1.pdf | DeepVQE: Real Time Deep Voice Quality Enhancement for Joint Acoustic Echo Cancellation, Noise Suppression and Dereverberation | Acoustic echo cancellation (AEC), noise suppression (NS) and dereverberation (DR) are an integral part of modern full-duplex communication systems. As the demand for teleconferencing systems increases, addressing these tasks is required for an effective and efficient online meeting experience. Most prior research proposes solutions for these tasks separately, combining them with digital signal processing (DSP) based components, resulting in complex pipelines that are often impractical to deploy in real-world applications. This paper proposes a real-time cross-attention deep model, named DeepVQE, based on residual convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to simultaneously address AEC, NS, and DR. We conduct several ablation studies to analyze the contributions of different components of our model to the overall performance. DeepVQE achieves state-of-the-art performance on non-personalized tracks from the ICASSP 2023 Acoustic Echo Cancellation Challenge and ICASSP 2023 Deep Noise Suppression Challenge test sets, showing that a single model can handle multiple tasks with excellent performance. Moreover, the model runs in real-time and has been successfully tested for the Microsoft Teams platform. | ['Ross Cutler', 'Jegor Guzvin', 'Tanel Parnamaa', 'Ando Saabas', 'Nicolae-Catalin Ristea', 'Evgenii Indenbom'] | 2023-06-05 | null | null | null | null | ['acoustic-echo-cancellation', 'acoustic-echo-cancellation'] | ['medical', 'speech'] | [ 7.35338032e-02 -2.66789287e-01 6.11221492e-01 -3.19433749e-01
-1.04431164e+00 -4.46985841e-01 3.61625463e-01 -4.61730808e-01
-5.25651395e-01 2.37302855e-01 5.64454019e-01 -5.89407027e-01
-1.95048705e-01 -1.05878353e-01 -4.77825195e-01 -4.24241096e-01
-6.53259158e-02 8.23307931e-02 9.21326950e-02 -7.36971736e-01
-1.76035315e-01 5.18673897e-01 -1.60817909e+00 1.84621915e-01
9.62690651e-01 8.97309363e-01 3.26009244e-02 1.11159861e+00
1.01476803e-01 6.93258703e-01 -9.41637456e-01 -4.11296427e-01
2.49247193e-01 -9.39227566e-02 -4.31648791e-02 -6.07993662e-01
6.62033916e-01 -2.95649052e-01 -8.12634230e-01 9.36361969e-01
1.44479668e+00 2.39274502e-01 1.98849440e-01 -1.14459372e+00
-1.36870161e-01 6.63196504e-01 -3.56658965e-01 1.97194472e-01
1.75717086e-01 8.80878344e-02 8.35151732e-01 -8.74540925e-01
-7.85162300e-02 1.06013119e+00 1.15129268e+00 4.68782216e-01
-1.07779372e+00 -1.21091890e+00 4.55035307e-02 3.06457758e-01
-1.23202848e+00 -1.21698499e+00 4.30184335e-01 1.27271041e-01
1.30203092e+00 4.40211445e-01 5.88063240e-01 1.07352531e+00
9.75257829e-02 1.06279147e+00 5.09824216e-01 -2.60992199e-01
7.45990723e-02 -1.58986434e-01 2.07674593e-01 7.82057047e-02
-1.52684122e-01 2.73887068e-01 -9.38554227e-01 8.67378637e-02
4.07354534e-01 -4.53271687e-01 -6.73341870e-01 2.68873394e-01
-8.54754567e-01 3.32499206e-01 2.90560573e-01 2.37189367e-01
-4.37334865e-01 4.74605262e-01 6.11059785e-01 6.22184277e-01
3.78931344e-01 4.03219491e-01 -4.94444966e-01 -5.66642165e-01
-1.07083166e+00 2.92405188e-01 1.10805714e+00 1.01632631e+00
6.67460486e-02 5.68172336e-01 -3.56616199e-01 1.16718960e+00
1.29305422e-01 7.46505082e-01 3.46051157e-01 -8.85520518e-01
4.27329808e-01 -3.55020583e-01 7.41088390e-02 -8.24225843e-01
-7.61930585e-01 -1.27255428e+00 -9.57835615e-01 -1.59378842e-01
6.33284971e-02 -5.55359006e-01 -9.00241017e-01 1.78417516e+00
-1.71762452e-01 6.46833241e-01 -2.27445289e-02 1.02886689e+00
1.00029755e+00 7.29559243e-01 -1.27982631e-01 2.62646466e-01
1.07893407e+00 -1.10630548e+00 -1.14494216e+00 -6.10557199e-01
4.15153146e-01 -9.74756658e-01 5.44978857e-01 7.07096815e-01
-1.10197759e+00 -5.57247818e-01 -1.23360884e+00 -8.55203047e-02
-3.26605439e-02 1.30652025e-01 4.85927492e-01 9.83311415e-01
-1.51983106e+00 3.54288310e-01 -6.26011014e-01 -1.39102200e-03
-1.71823829e-01 6.51507258e-01 -9.53265354e-02 5.26320860e-02
-1.51649904e+00 7.99347699e-01 -5.10312140e-01 8.55441391e-01
-7.48806477e-01 -1.01489091e+00 -7.12100983e-01 6.42563283e-01
1.48648724e-01 -2.62202650e-01 1.69168711e+00 -7.08503187e-01
-1.68291140e+00 3.07324499e-01 -1.36773288e-01 -5.47830462e-01
3.96629870e-01 -8.24723303e-01 -9.66686904e-01 -3.06023955e-01
-3.48254025e-01 3.96012217e-01 8.94977093e-01 -8.27480495e-01
-7.40541339e-01 -9.77295935e-02 -4.33885843e-01 2.38899827e-01
-3.91918927e-01 9.11877751e-02 -8.90416503e-01 -6.83856010e-01
2.11483866e-01 -1.03277469e+00 -1.12226084e-01 -4.24633116e-01
-3.03554565e-01 8.02568495e-02 5.92215657e-01 -1.15252018e+00
1.34025502e+00 -2.45799661e+00 5.54716773e-02 3.47184300e-01
6.61917776e-02 8.05620372e-01 -5.12617350e-01 2.47744992e-01
-1.46240830e-01 -1.73678219e-01 7.61121586e-02 -7.23436773e-01
3.19388002e-01 -2.32062697e-01 -3.45691293e-01 3.81862611e-01
-2.14013532e-01 7.16428876e-01 -5.22497535e-01 3.74990642e-01
2.29035363e-01 7.31500268e-01 -6.46273673e-01 1.03281550e-01
3.86869758e-01 1.75060302e-01 1.35809809e-01 3.99194241e-01
1.06487143e+00 3.04785967e-01 2.08642632e-01 -3.60388845e-01
-2.13226140e-01 5.43219626e-01 -1.31265676e+00 1.62523246e+00
-8.06859791e-01 1.28491175e+00 9.83674347e-01 -6.53093159e-01
7.82408357e-01 4.83467191e-01 1.48088157e-01 -1.27569675e+00
8.59219953e-02 5.83484054e-01 3.12873185e-01 -6.33805469e-02
8.24942052e-01 2.86084972e-03 -3.14282365e-02 2.14468136e-01
1.87322468e-01 1.43872611e-02 -1.51019171e-01 3.23105425e-01
1.47515464e+00 -4.36996579e-01 -3.07052791e-01 -3.37582260e-01
4.35267866e-01 -7.14652181e-01 6.75977886e-01 1.15717387e+00
-3.95219356e-01 7.46813118e-01 2.03233987e-01 4.40820009e-02
-7.73350894e-01 -8.62220585e-01 8.49850848e-02 1.34150362e+00
-9.49558094e-02 -3.60405505e-01 -6.98980749e-01 -3.21357474e-02
3.84314507e-02 7.44970560e-01 -8.56140256e-02 -3.14958423e-01
-7.22317040e-01 -5.23158729e-01 9.78621542e-01 4.36355442e-01
5.00475109e-01 -9.15304303e-01 -1.35124549e-01 5.55625498e-01
-4.32448089e-01 -1.33526826e+00 -8.43890309e-01 4.02254045e-01
-2.59828031e-01 -4.21936989e-01 -9.71197903e-01 -6.99233592e-01
-9.74450782e-02 6.79514289e-01 1.25782073e+00 -4.12020460e-02
-1.52585268e-01 5.07524908e-01 -1.90386191e-01 -5.83106577e-01
-1.69441327e-01 2.64903426e-01 3.55455071e-01 5.11770174e-02
4.49384958e-01 -6.14879429e-01 -7.69261837e-01 3.87345046e-01
-5.44914305e-01 -3.11995596e-01 8.14086258e-01 6.92203104e-01
-1.59297436e-01 -7.36346021e-02 8.53430867e-01 -4.25694644e-01
1.01806509e+00 -2.07856998e-01 -5.98337114e-01 8.42049196e-02
-1.54570401e-01 -3.17909986e-01 4.12410885e-01 -2.03741789e-01
-1.14483798e+00 -1.73646063e-01 -9.18005586e-01 -2.56252468e-01
1.51868448e-01 5.74906111e-01 -1.13956936e-01 -1.34943604e-01
4.17437941e-01 3.06615263e-01 -6.23267069e-02 -4.72832769e-01
6.41028136e-02 1.12132859e+00 7.68548131e-01 1.31350666e-01
6.45898581e-01 5.95885776e-02 -3.92386287e-01 -1.23827994e+00
-6.27996147e-01 -6.38626158e-01 -2.55251713e-02 -2.55065739e-01
3.88787299e-01 -1.42332065e+00 -1.11479914e+00 7.42897391e-01
-1.12408578e+00 -2.71008313e-01 2.79081702e-01 6.22033834e-01
3.26181613e-02 1.38653561e-01 -6.75732374e-01 -1.02432299e+00
-7.50569642e-01 -9.92432356e-01 9.86761212e-01 3.00616235e-01
-3.34529817e-01 -5.43100834e-01 -7.37425825e-03 3.26748848e-01
1.21104467e+00 -6.44363344e-01 3.63181710e-01 -6.03323579e-01
-4.82400656e-01 -4.24567252e-01 -3.94717567e-02 5.65509975e-01
-2.92665631e-01 -3.62810969e-01 -1.44240880e+00 -3.85554701e-01
2.37901025e-02 -8.65597185e-03 1.07367527e+00 8.10509503e-01
9.13263738e-01 2.39694312e-01 -8.45497400e-02 7.98142076e-01
7.09201515e-01 2.92653978e-01 9.06180084e-01 4.22220752e-02
3.45358640e-01 4.31181639e-01 8.72914493e-02 3.85712385e-01
2.74005502e-01 7.03454971e-01 1.16261050e-01 -4.10048246e-01
-1.52976230e-01 6.86556250e-02 5.78544438e-01 1.17528415e+00
4.84717846e-01 -5.36815226e-01 -7.90111482e-01 6.07454240e-01
-1.65217710e+00 -7.18138635e-01 -2.31980100e-01 2.06412816e+00
4.13379580e-01 1.24794714e-01 -3.45049977e-01 2.00742155e-01
5.65410197e-01 1.06620803e-01 -5.40246725e-01 -5.57495832e-01
-1.84114888e-01 4.87133712e-01 5.97156763e-01 5.22172689e-01
-1.05662417e+00 7.72642136e-01 6.43615627e+00 1.00173187e+00
-1.11971116e+00 -1.06813461e-02 3.51063907e-01 -3.35358202e-01
-2.93335050e-01 -6.26556039e-01 -7.76582420e-01 1.45830616e-01
1.41731501e+00 2.04191044e-01 6.10752344e-01 4.06505913e-01
2.90963590e-01 1.99156836e-01 -9.47356999e-01 1.28889799e+00
-4.25798760e-04 -1.09186375e+00 -5.65171719e-01 -1.19312368e-01
3.42674166e-01 5.33566236e-01 2.96670377e-01 9.48679805e-01
2.43851498e-01 -9.73396003e-01 7.74394870e-01 4.51230884e-01
5.97724438e-01 -8.43179643e-01 9.78019118e-01 2.63585709e-02
-1.08108425e+00 -3.14421594e-01 -4.64112945e-02 1.40959620e-01
4.50085908e-01 8.31900597e-01 -4.01213348e-01 4.01825190e-01
8.72417271e-01 4.09396678e-01 -9.09242257e-02 1.29441273e+00
-1.23740792e-01 8.55114937e-01 -5.26345968e-01 5.28728925e-02
1.86006352e-01 6.61636889e-02 7.20418751e-01 1.45077646e+00
5.20238221e-01 1.05227441e-01 -4.47681040e-01 3.58618230e-01
-4.02810603e-01 -2.62042791e-01 -2.02634707e-02 8.17385614e-02
5.86667657e-01 1.27454412e+00 -2.53278941e-01 8.19541365e-02
-4.84530032e-01 9.20106649e-01 -7.40282387e-02 6.26472414e-01
-1.02868748e+00 -9.45393205e-01 1.20647383e+00 -2.42582336e-01
3.17350954e-01 -3.07863533e-01 -2.76945114e-01 -9.76575613e-01
-7.25868791e-02 -1.24060583e+00 -1.27107352e-01 -5.65483630e-01
-1.02566111e+00 6.93169415e-01 -8.65498424e-01 -1.15772831e+00
-1.84137765e-02 -3.75459820e-01 -5.58711529e-01 1.27822304e+00
-1.71079469e+00 -5.96495330e-01 -3.17115426e-01 3.21759731e-01
4.56322581e-01 -2.42667437e-01 7.91225433e-01 1.19726825e+00
-6.50505483e-01 1.14215064e+00 4.23913926e-01 1.17454231e-01
9.87736106e-01 -9.91178155e-01 1.03128982e+00 9.46368277e-01
-1.06562577e-01 7.05623388e-01 7.89561391e-01 -4.48918790e-01
-1.57360160e+00 -8.17929447e-01 9.65182245e-01 1.04289785e-01
4.44007158e-01 -8.24140310e-01 -8.78565133e-01 4.57693577e-01
3.69105101e-01 -3.21369857e-01 6.96141899e-01 4.89472687e-01
1.99847762e-02 -4.94616807e-01 -5.95968306e-01 7.67070234e-01
1.02798438e+00 -6.53625488e-01 6.35608193e-03 -5.62468097e-02
7.55444586e-01 -6.82916522e-01 -3.87878537e-01 3.06746423e-01
8.24294984e-01 -1.00119090e+00 9.85358000e-01 -2.16308847e-01
7.61452988e-02 -8.54970068e-02 -1.56024128e-01 -1.59725463e+00
-4.09024149e-01 -1.20725191e+00 -9.10886005e-02 1.13620424e+00
6.03640139e-01 -6.20780468e-01 5.76946318e-01 8.08914483e-01
-7.23075151e-01 -6.71248510e-02 -1.00832438e+00 -6.01414084e-01
-3.73898953e-01 -8.33116889e-01 4.98376876e-01 4.63613927e-01
-3.59181076e-01 4.34575319e-01 -8.65310550e-01 2.83853382e-01
4.46554154e-01 -4.76740301e-01 7.84738064e-01 -1.02583301e+00
-4.19261068e-01 -5.64129531e-01 -1.81026608e-01 -1.62000251e+00
4.08781916e-02 -5.61114907e-01 4.22381014e-01 -1.44476104e+00
-5.00098228e-01 -3.13358277e-01 -4.28950727e-01 1.07072234e-01
-1.01348959e-01 1.50338277e-01 3.44270647e-01 -2.62369186e-01
-4.53072518e-01 9.07610238e-01 7.05796480e-01 -1.86196089e-01
-3.39869648e-01 3.94165933e-01 -9.40123200e-01 4.49595302e-01
4.70834374e-01 -2.62400120e-01 -1.30505651e-01 -8.28907907e-01
2.81396151e-01 3.06105435e-01 -8.49226341e-02 -1.25290501e+00
8.89097333e-01 7.21668184e-01 1.99057043e-01 -6.82883561e-01
6.49816990e-01 -7.55312562e-01 -8.09959769e-02 3.78536046e-01
-4.61586356e-01 -2.28842080e-01 6.87919438e-01 4.15656298e-01
-4.00274307e-01 7.93804750e-02 5.99351704e-01 4.19012994e-01
-6.33594036e-01 -1.29407709e-02 -9.00965750e-01 -1.13740541e-01
1.93621725e-01 1.35131896e-01 -1.62460998e-01 -1.14519405e+00
-6.87729716e-01 5.48263848e-01 -5.38692355e-01 6.67839229e-01
5.77522457e-01 -1.07511806e+00 -1.02635694e+00 3.63163799e-01
-3.34356993e-01 -5.14370203e-01 8.43474329e-01 9.08740401e-01
-4.92373288e-01 6.41982734e-01 9.77495611e-02 -5.68530381e-01
-1.23862147e+00 -2.99693376e-01 8.15086901e-01 -2.34399736e-02
-6.07219994e-01 1.20926249e+00 3.18131298e-02 -8.89809370e-01
9.78151202e-01 -1.61431789e-01 -1.42190233e-01 -6.08240068e-02
8.74860466e-01 5.71551561e-01 5.48322201e-01 -2.56387651e-01
-4.65472013e-01 2.22284123e-01 -2.59308219e-01 -4.17014211e-01
1.22223234e+00 -2.91109383e-01 9.66344252e-02 -1.96236193e-01
1.19508731e+00 1.32768467e-01 -9.82901156e-01 -4.08294171e-01
-3.32306661e-02 -1.77050278e-01 6.77438498e-01 -1.06670034e+00
-1.27638412e+00 9.54491436e-01 9.29453075e-01 -1.53052732e-01
1.47095525e+00 -6.54529989e-01 1.07913983e+00 4.40262973e-01
-7.04969466e-02 -1.17335331e+00 -2.24042490e-01 9.90337312e-01
9.52322662e-01 -1.05275953e+00 -4.51205224e-01 -2.56338686e-01
-5.55311680e-01 9.09432292e-01 4.88917649e-01 1.88951954e-01
8.66906285e-01 7.93912172e-01 3.63147080e-01 1.49407730e-01
-8.23517084e-01 -1.00529931e-01 3.17303568e-01 3.64308089e-01
6.37187362e-01 -5.63906766e-02 1.22626983e-01 9.31795835e-01
-2.09734336e-01 -2.10087553e-01 4.93850917e-01 6.93913639e-01
-2.61067420e-01 -7.22315192e-01 -2.52078921e-01 4.06879246e-01
-5.22828162e-01 -4.79962736e-01 -2.55273461e-01 3.95525277e-01
-4.28810149e-01 1.37934625e+00 2.17330962e-01 -7.15054691e-01
7.91043460e-01 -1.14225209e-01 -1.91150028e-02 -1.12446107e-01
-1.20616972e+00 5.84916592e-01 3.56497198e-01 -5.86760342e-01
1.17065981e-01 -4.57185805e-01 -9.94634211e-01 -6.04778469e-01
-3.45993727e-01 2.20909849e-01 9.68170941e-01 7.61498570e-01
7.74712741e-01 1.26879203e+00 3.76349807e-01 -8.96989822e-01
-5.99446356e-01 -1.14551532e+00 -7.19368398e-01 7.48662231e-03
6.77224874e-01 -1.51752114e-01 -3.97323430e-01 -6.62200153e-01] | [14.993049621582031, 5.97265625] |
2bf45ced-84c7-4481-a9d4-e0d58f09e413 | enhanced-english-universal-dependencies-an | null | null | https://aclanthology.org/L16-1376 | https://aclanthology.org/L16-1376.pdf | Enhanced English Universal Dependencies: An Improved Representation for Natural Language Understanding Tasks | Many shallow natural language understanding tasks use dependency trees to extract relations between content words. However, strict surface-structure dependency trees tend to follow the linguistic structure of sentences too closely and frequently fail to provide direct relations between content words. To mitigate this problem, the original Stanford Dependencies representation also defines two dependency graph representations which contain additional and augmented relations that explicitly capture otherwise implicit relations between content words. In this paper, we revisit and extend these dependency graph representations in light of the recent Universal Dependencies (UD) initiative and provide a detailed account of an enhanced and an enhanced++ English UD representation. We further present a converter from constituency to basic, i.e., strict surface structure, UD trees, and a converter from basic UD trees to enhanced and enhanced++ English UD graphs. We release both converters as part of Stanford CoreNLP and the Stanford Parser. | ['Sebastian Schuster', 'Christopher D. Manning'] | 2016-05-01 | enhanced-english-universal-dependencies-an-1 | https://aclanthology.org/L16-1376 | https://aclanthology.org/L16-1376.pdf | lrec-2016-5 | ['implicit-relations'] | ['natural-language-processing'] | [-5.35813831e-02 5.32851815e-01 -4.44517523e-01 -9.84864235e-01
-5.48214853e-01 -9.28590953e-01 4.13391113e-01 5.17652631e-01
-2.13072956e-01 8.98736179e-01 7.63401926e-01 -6.73301935e-01
-6.17630258e-02 -9.72385228e-01 -1.40387312e-01 -7.34743327e-02
-1.14752293e-01 5.41177809e-01 4.22770590e-01 -6.43701613e-01
4.75343354e-02 4.51767355e-01 -1.07804322e+00 3.42688292e-01
6.62933409e-01 1.44512728e-01 3.37959170e-01 6.84373617e-01
-7.74538219e-01 9.11344886e-01 -5.98667383e-01 -7.10228622e-01
-2.70103067e-01 -3.60813916e-01 -1.31572330e+00 -3.05864066e-01
3.20001483e-01 -1.26849949e-01 -1.48012385e-01 9.67127025e-01
4.58571091e-02 9.58946645e-02 4.06291217e-01 -9.62135077e-01
-9.61433589e-01 1.27908850e+00 -2.33633295e-01 7.58705556e-01
7.36231387e-01 -6.37638032e-01 1.93450201e+00 -5.85218012e-01
9.01951373e-01 1.43380022e+00 5.86976528e-01 6.71852469e-01
-1.34323502e+00 -4.18394208e-01 6.75385892e-01 -3.24865542e-02
-9.34451699e-01 -3.61099988e-01 6.16893053e-01 -2.26178408e-01
1.97886801e+00 1.13772348e-01 2.49850571e-01 9.66813147e-01
4.32773024e-01 5.61264813e-01 9.51829374e-01 -8.42681706e-01
-3.28599334e-01 -5.35034910e-02 1.30043018e+00 6.50354803e-01
6.66904509e-01 -6.09177090e-02 -4.34130222e-01 -4.60141674e-02
6.79861009e-01 -5.03842354e-01 4.82980981e-02 3.74264233e-02
-6.20554149e-01 1.09560573e+00 2.72107363e-01 9.09440815e-01
1.39587820e-01 4.56530340e-02 5.22863209e-01 4.37156230e-01
4.84495670e-01 2.19885409e-01 -9.05484438e-01 1.76431462e-01
-8.00700188e-02 9.16955993e-02 1.16609526e+00 1.21792269e+00
1.02427757e+00 -8.91256779e-02 1.86324179e-01 9.90198493e-01
4.04441863e-01 6.60049915e-02 2.89787471e-01 -6.55280173e-01
6.07443511e-01 5.86213470e-01 -2.53353000e-01 -7.35400259e-01
-6.16959929e-01 -2.18560353e-01 -1.54652655e-01 -3.51444751e-01
8.21960270e-02 -5.63527197e-02 -8.84224355e-01 1.95263433e+00
2.84856141e-01 -5.07512033e-01 4.52573866e-01 3.83406550e-01
1.42445195e+00 5.60725272e-01 6.49223685e-01 -1.18610986e-01
1.65344632e+00 -6.15689993e-01 -1.15293646e+00 -9.01420832e-01
1.44801056e+00 -5.03225327e-01 1.00294340e+00 -1.77710727e-01
-1.07984495e+00 -3.78533006e-01 -1.09082901e+00 -6.59882307e-01
-6.79918110e-01 -3.84349436e-01 1.16308510e+00 8.01492691e-01
-1.07266545e+00 5.31964600e-01 -7.99572170e-01 -5.20308733e-01
-7.20158517e-02 3.72849554e-01 -6.42704964e-01 -7.99807385e-02
-1.65487492e+00 1.49002910e+00 5.98583400e-01 -3.76547635e-01
-5.08027524e-02 -5.54053724e-01 -1.59504902e+00 -1.99017245e-02
2.25400642e-01 -4.05662268e-01 1.36066508e+00 -6.68774366e-01
-1.05564356e+00 1.36165953e+00 -2.28097007e-01 -1.80889979e-01
-6.17761970e-01 -5.18684983e-01 -5.38227499e-01 -2.09234849e-01
4.33716953e-01 1.64568081e-01 -4.43960279e-02 -9.81946051e-01
-7.42335677e-01 -6.12931967e-01 5.12613356e-01 1.31605923e-01
-3.66700627e-02 7.85258532e-01 -2.81373620e-01 -7.05345750e-01
1.38271675e-01 -6.00615084e-01 -1.59442559e-01 -8.64385962e-01
-2.48054072e-01 -8.56430292e-01 3.96459609e-01 -6.12662137e-01
1.60698593e+00 -1.64738536e+00 2.07874045e-01 4.39805388e-02
1.29306272e-01 5.49921021e-02 -2.36212105e-01 6.21728361e-01
-5.24103940e-01 4.40665841e-01 -3.74433249e-01 -2.52204210e-01
1.31867111e-01 1.35792077e+00 4.26869169e-02 8.01204219e-02
4.27343249e-01 1.00022566e+00 -9.25111353e-01 -6.12491846e-01
1.21069387e-01 1.79616228e-01 -5.87737143e-01 2.03004405e-01
-1.22426607e-01 3.79883945e-02 -6.24184370e-01 4.72501844e-01
5.56782961e-01 9.94218513e-02 8.90352964e-01 -8.01063981e-03
-4.10777003e-01 1.37479448e+00 -8.70483577e-01 1.78516090e+00
-7.49390602e-01 3.82257015e-01 2.72842616e-01 -9.26966667e-01
1.02185321e+00 3.04215938e-01 -1.07985109e-01 -4.90499884e-01
1.61493495e-01 2.53566712e-01 2.22000450e-01 -4.30846572e-01
6.12556994e-01 -5.91011643e-01 -6.31913722e-01 3.95173669e-01
5.19772470e-01 -4.73677814e-01 6.57051623e-01 6.21735156e-01
1.22083890e+00 3.88180077e-01 1.04144800e+00 -6.71412051e-01
6.41885400e-01 -3.32235582e-02 7.52879202e-01 2.54297763e-01
-2.88918726e-02 1.45290792e-01 7.83765435e-01 -3.19018394e-01
-6.33781374e-01 -1.06221974e+00 -4.82416421e-01 1.41805029e+00
-1.28881574e-01 -1.00453651e+00 -4.03907001e-01 -1.04250872e+00
-2.60229677e-01 1.27175891e+00 -7.05443501e-01 2.79893935e-01
-1.00879753e+00 -6.32948816e-01 5.45022070e-01 9.87487555e-01
-3.32136266e-02 -1.35216081e+00 -2.76311964e-01 5.21873653e-01
-2.30218321e-01 -1.51707578e+00 -1.29337370e-01 8.01313579e-01
-6.77104414e-01 -9.76660788e-01 2.95073599e-01 -1.29580104e+00
3.80774587e-01 -2.23756492e-01 1.72720528e+00 5.44881284e-01
-2.18154863e-02 7.76335523e-02 -9.16485250e-01 -3.22323531e-01
-7.00665414e-01 2.25965410e-01 -1.84469804e-01 -8.50062430e-01
8.27166021e-01 -6.92839026e-01 2.42643878e-01 -2.04701751e-01
-8.81082177e-01 -3.28545690e-01 2.52587914e-01 5.56060374e-01
2.84263253e-01 -3.46878529e-01 4.82362479e-01 -1.55602503e+00
5.88140905e-01 -5.12116194e-01 -2.85080075e-01 2.27148622e-01
-3.16028476e-01 4.14517671e-01 4.01927263e-01 3.47763181e-01
-1.44757652e+00 -7.99419582e-02 -9.12281215e-01 5.50318658e-01
-2.45722473e-01 8.80122602e-01 -4.15381253e-01 3.96607280e-01
5.67102671e-01 -4.14638013e-01 -5.99221051e-01 -4.44269150e-01
6.57140672e-01 4.78678018e-01 4.33543593e-01 -1.07185233e+00
4.72945184e-01 2.00857054e-02 -1.99511424e-01 -8.82773340e-01
-1.39065564e+00 -5.31955957e-01 -1.31639397e+00 5.17399251e-01
1.17840755e+00 -6.11672223e-01 8.76225680e-02 -1.26321554e-01
-1.64797354e+00 -2.77368426e-01 -3.39010686e-01 2.71909505e-01
-2.83513904e-01 7.14241982e-01 -1.03719747e+00 -4.61380690e-01
-2.94794917e-01 -5.22847712e-01 8.19766819e-01 3.49665363e-03
-6.99223936e-01 -1.63012278e+00 3.89359832e-01 -1.12334736e-01
-6.10662997e-03 2.36149088e-01 1.44690621e+00 -9.50566351e-01
2.88837314e-01 -1.56835541e-01 -2.71275878e-01 3.31581116e-01
5.07190883e-01 -1.33005409e-02 -4.80778575e-01 1.10377744e-01
-6.04679473e-02 -2.73367792e-01 5.26004016e-01 1.10431097e-01
4.00203951e-02 -6.59097433e-02 -4.77094531e-01 2.54336685e-01
1.50444448e+00 1.52360320e-01 5.99004328e-01 2.39181340e-01
6.46289647e-01 1.17846107e+00 3.04041177e-01 4.78964485e-03
7.37587750e-01 2.40923628e-01 1.32743940e-01 1.25109687e-01
-3.62635404e-01 -2.48836175e-01 4.82900530e-01 1.35958791e+00
7.39716589e-02 -1.77536204e-01 -9.47142184e-01 7.55384266e-01
-1.44681478e+00 -5.34492671e-01 -8.80188704e-01 1.61992300e+00
1.05200946e+00 3.89085680e-01 -3.56821626e-01 -8.61196294e-02
7.39415705e-01 3.26024234e-01 1.35487556e-01 -1.26778865e+00
-2.31141478e-01 9.77038920e-01 1.92489237e-01 1.12879646e+00
-7.85573065e-01 1.71423221e+00 6.94374228e+00 1.44265324e-01
-2.82350361e-01 3.08352470e-01 -1.95212260e-01 6.29487455e-01
-6.62639260e-01 4.31916535e-01 -1.12675750e+00 -3.74026954e-01
1.17714775e+00 -2.35918000e-01 -2.08732352e-01 6.72423244e-01
-3.88794899e-01 2.60617901e-02 -1.32855046e+00 3.72182339e-01
-1.44991219e-01 -1.08383334e+00 1.77201226e-01 -1.93852291e-01
3.65692794e-01 3.44079614e-01 -7.32439399e-01 4.49310809e-01
1.20489967e+00 -9.59848225e-01 4.62827414e-01 -2.88143516e-01
7.20742881e-01 -5.77077389e-01 8.71266305e-01 7.23344162e-02
-1.78946066e+00 2.55166143e-01 -4.39046860e-01 -5.11138678e-01
5.42054534e-01 2.39365160e-01 -1.15164094e-01 9.73261297e-01
7.03266144e-01 8.47907662e-01 -3.99009347e-01 -6.77814931e-02
-1.16095364e+00 3.12309951e-01 -1.32711276e-01 -2.43810743e-01
4.41633314e-01 -3.11121345e-01 4.46210593e-01 1.78225601e+00
-1.93693534e-01 8.01000416e-01 1.26688465e-01 5.80977619e-01
2.09115699e-01 2.38708347e-01 -6.61470354e-01 -8.36645905e-03
4.74344075e-01 1.16149437e+00 -6.92373276e-01 -5.97398698e-01
-1.04212046e+00 7.08147407e-01 8.72640431e-01 4.85188328e-02
-5.28366625e-01 -4.85219955e-01 9.67005312e-01 -9.31041017e-02
3.26424003e-01 -6.30260706e-01 -2.74038106e-01 -1.22086954e+00
-2.78302550e-01 -6.19706690e-01 9.19525862e-01 -5.69863498e-01
-1.53663528e+00 9.85795796e-01 1.86871782e-01 -2.97621131e-01
-2.67381310e-01 -8.94812763e-01 -7.03582168e-01 9.59869564e-01
-1.61372232e+00 -1.06176770e+00 1.99502900e-01 5.78724325e-01
6.51709616e-01 5.69861054e-01 1.34539628e+00 -3.63925546e-02
-5.33463299e-01 2.88758367e-01 -5.72164536e-01 4.82066810e-01
3.31691533e-01 -1.50499260e+00 9.97782946e-01 9.21389163e-01
2.83922791e-01 9.88708138e-01 7.96721816e-01 -8.43285143e-01
-1.02086592e+00 -7.56706178e-01 1.76628602e+00 -8.68049443e-01
1.22106779e+00 -4.95827585e-01 -1.16791821e+00 1.55827534e+00
6.41357839e-01 1.56695992e-02 9.71011341e-01 6.84713781e-01
-6.78854883e-01 3.63757491e-01 -9.01074827e-01 4.27401513e-01
1.53153789e+00 -6.06292784e-01 -1.66185379e+00 9.00098383e-02
9.45505202e-01 -3.78010303e-01 -9.61036921e-01 3.55682254e-01
2.66282171e-01 -9.43115294e-01 6.15549028e-01 -9.37316954e-01
3.70124638e-01 1.05117492e-01 -5.06185770e-01 -1.30822563e+00
-6.27432883e-01 -4.61816967e-01 2.29075864e-01 1.51564097e+00
8.16765368e-01 -8.97252321e-01 4.67633337e-01 7.68851280e-01
-6.91357911e-01 -2.11645082e-01 -8.97258759e-01 -7.95779169e-01
6.08323634e-01 -7.81597853e-01 4.79845494e-01 1.19434893e+00
7.23766565e-01 1.09399271e+00 3.02329570e-01 1.85049191e-01
5.33636212e-01 5.66968583e-02 3.24635863e-01 -1.47793198e+00
-2.78452903e-01 -2.24872842e-01 -2.99817085e-01 -1.13120925e+00
6.96536183e-01 -1.34893513e+00 3.98360156e-02 -2.08591390e+00
-1.60932705e-01 -6.81597948e-01 -6.21096753e-02 6.02176487e-01
-2.06880152e-01 -7.85700232e-02 -1.28561676e-01 -8.52442086e-02
-1.01876780e-01 2.47295097e-01 8.05892944e-01 2.71714866e-01
-1.86599016e-01 -3.62892449e-01 -9.20937955e-01 9.97281849e-01
7.74716139e-01 -7.49825656e-01 -2.96655178e-01 -8.14547598e-01
3.80640686e-01 -1.76504441e-02 -3.35380524e-01 -2.96360344e-01
-1.23158246e-01 -1.14891723e-01 -1.33772925e-01 -3.94410819e-01
1.23303421e-02 -4.78013605e-01 -2.41848335e-01 3.72274399e-01
-2.64985919e-01 2.87497252e-01 3.50099653e-01 5.03414646e-02
-2.57608205e-01 -6.90532506e-01 5.73229134e-01 -4.31191236e-01
-1.00787485e+00 -1.71608683e-02 -4.74590898e-01 3.39221269e-01
7.92014360e-01 -1.04697041e-01 -4.59796399e-01 6.42185658e-02
-1.08240974e+00 2.34893233e-01 7.09438622e-02 5.98295927e-01
3.90332311e-01 -9.81395960e-01 -7.21818388e-01 2.18564700e-02
1.37547776e-01 -7.40566626e-02 -1.38628811e-01 2.37569332e-01
-4.62256074e-01 7.02758670e-01 -1.19255088e-01 3.26625332e-02
-1.35025239e+00 5.27358174e-01 7.95559064e-02 -6.26381934e-01
-6.89685881e-01 9.96580124e-01 4.06445682e-01 -7.10283995e-01
-3.42983216e-01 -6.74575031e-01 -6.23268723e-01 1.28448188e-01
3.98959756e-01 -8.47861469e-02 2.28785321e-01 -9.82295036e-01
-8.00842643e-01 5.63930333e-01 -2.37844989e-01 4.90505807e-02
1.39158034e+00 -3.16321194e-01 -7.12186873e-01 4.26314861e-01
1.07606328e+00 3.90311688e-01 -5.73268354e-01 -3.49242061e-01
6.81192398e-01 -2.33758744e-02 -2.01318398e-01 -4.47823524e-01
-6.78319752e-01 6.38446569e-01 -4.56102371e-01 6.86951101e-01
7.35580444e-01 7.90453732e-01 7.98367083e-01 2.57262945e-01
2.73610562e-01 -7.99552023e-01 -2.45200634e-01 1.12684631e+00
8.98023605e-01 -6.81307197e-01 9.60908905e-02 -1.14127183e+00
-4.86320943e-01 1.18354654e+00 7.44140685e-01 -3.52101535e-01
8.33399236e-01 8.44773471e-01 1.94372788e-01 -6.07952774e-01
-7.93447733e-01 -7.31048346e-01 -6.62165135e-02 8.56367528e-01
1.18532670e+00 1.87733874e-01 -8.10355723e-01 8.98064792e-01
-6.11394286e-01 -6.24736071e-01 6.65915847e-01 1.27920389e+00
-5.56881070e-01 -1.81755435e+00 -5.09851351e-02 1.54405966e-01
-7.20953941e-01 -7.05995142e-01 -5.84045649e-01 1.15326405e+00
-8.45775679e-02 1.25842738e+00 3.16483736e-01 -8.68107751e-02
7.24363685e-01 2.54032820e-01 7.81228840e-01 -1.47243226e+00
-7.89179981e-01 -3.72443557e-01 8.83707941e-01 -3.90660107e-01
-7.04475343e-01 -5.21538317e-01 -2.00350738e+00 -2.22017378e-01
-5.16974509e-01 4.17046994e-01 4.92742181e-01 1.21327484e+00
-7.44952857e-02 4.29373354e-01 1.37663186e-01 -5.16513765e-01
-1.86677471e-01 -1.02667928e+00 -6.68360651e-01 2.12863937e-01
2.94186305e-02 -4.13633645e-01 -4.09057438e-01 -5.28670326e-02] | [10.417211532592773, 9.635456085205078] |
47f72463-ba24-441a-8551-6164b2cf0d18 | ffb-a-fair-fairness-benchmark-for-in | 2306.09468 | null | https://arxiv.org/abs/2306.09468v1 | https://arxiv.org/pdf/2306.09468v1.pdf | FFB: A Fair Fairness Benchmark for In-Processing Group Fairness Methods | This paper introduces the Fair Fairness Benchmark (\textsf{FFB}), a benchmarking framework for in-processing group fairness methods. Ensuring fairness in machine learning is critical for ethical and legal compliance. However, there exist challenges in comparing and developing of fairness methods due to inconsistencies in experimental settings, lack of accessible algorithmic implementations, and limited extensibility of current fairness packages and tools. To address these issues, we introduce an open-source, standardized benchmark for evaluating in-processing group fairness methods and provide a comprehensive analysis of state-of-the-art methods to ensure different notions of group fairness. This work offers the following key contributions: the provision of flexible, extensible, minimalistic, and research-oriented open-source code; the establishment of unified fairness method benchmarking pipelines; and extensive benchmarking, which yields key insights from $\mathbf{45,079}$ experiments. We believe our work will significantly facilitate the growth and development of the fairness research community. The benchmark, including code and running logs, is available at https://github.com/ahxt/fair_fairness_benchmark | ['Xia Hu', 'Na Zou', 'Han Zhao', 'Qifan Wang', 'Yu Chen', 'Jianfeng Chi', 'Xiaotian Han'] | 2023-06-15 | null | null | null | null | ['benchmarking', 'benchmarking'] | ['miscellaneous', 'robots'] | [-2.94606388e-01 -1.25488117e-01 -4.05916363e-01 -1.06720757e+00
-7.27265716e-01 -5.19536614e-01 5.91067791e-01 3.49144578e-01
-7.48570979e-01 9.17678773e-01 3.18277180e-01 -6.00439310e-01
-1.03322463e-02 -5.24212122e-01 3.59525569e-02 -2.38440976e-01
1.65312648e-01 2.25078523e-01 -3.38823467e-01 -1.07256599e-01
7.51929224e-01 4.18515146e-01 -1.41432643e+00 1.25000775e-01
1.11561966e+00 7.33088493e-01 -7.40846157e-01 5.83615422e-01
6.95643947e-03 1.22515750e+00 -4.92157102e-01 -1.15760362e+00
4.12333965e-01 -6.64059997e-01 -1.31077957e+00 -8.34745109e-01
9.03938115e-01 -8.35137010e-01 -2.70374538e-03 1.08078682e+00
9.25889432e-01 4.36813086e-01 4.79439378e-01 -1.59640026e+00
-6.81708217e-01 6.69665158e-01 -6.20657325e-01 3.97080988e-01
5.49837276e-02 4.91336733e-01 1.19985008e+00 -3.55253667e-01
4.69894350e-01 1.29307067e+00 6.63167775e-01 1.07616544e+00
-1.04877090e+00 -1.06166291e+00 -3.23455900e-01 1.25548661e-01
-1.07728696e+00 -9.93599057e-01 1.73285440e-01 -3.89279753e-01
8.24042022e-01 8.46923649e-01 3.92057538e-01 8.11249554e-01
4.23000567e-03 3.65220666e-01 1.50342393e+00 -2.20123023e-01
2.22726926e-01 5.18083908e-02 4.11649019e-01 6.75320327e-01
4.04353321e-01 3.71385247e-01 -5.75248122e-01 -6.07532144e-01
2.22023010e-01 -2.54310995e-01 7.95633793e-02 1.55268162e-01
-7.36538410e-01 1.02763128e+00 2.92354554e-01 -7.45374113e-02
1.40868975e-02 3.36604178e-01 8.64385307e-01 2.55875409e-01
8.08192790e-01 4.27864492e-01 -2.18917187e-02 -5.29464006e-01
-1.23506355e+00 7.45558441e-01 9.61336732e-01 7.60280430e-01
5.37543237e-01 -2.97798395e-01 -5.36494315e-01 8.13246310e-01
2.46169940e-01 2.59081542e-01 -6.80626482e-02 -1.87006152e+00
2.44436279e-01 1.58996940e-01 1.68909803e-01 -9.73039925e-01
-3.51395696e-01 -1.08160764e-01 -6.62832379e-01 4.92491096e-01
6.11604631e-01 5.29640988e-02 -1.16890714e-01 1.87738001e+00
4.19396669e-01 -4.46760148e-01 -5.62221885e-01 1.11797714e+00
1.08218908e+00 -5.01324497e-02 7.23742664e-01 -1.78670466e-01
1.18668652e+00 -9.73051786e-01 -4.78920341e-01 1.65142700e-01
5.01411319e-01 -1.05522823e+00 1.07750309e+00 1.10173382e-01
-1.49498260e+00 -2.05182403e-01 -4.98315841e-01 -7.30876088e-01
-3.14345717e-01 -6.18609250e-01 8.03934753e-01 1.22990429e+00
-1.04566705e+00 8.85091603e-01 -3.91988426e-01 -4.11264211e-01
1.02322710e+00 7.71062300e-02 -3.25124383e-01 -1.37785167e-01
-1.13065350e+00 9.96033072e-01 -1.29271671e-01 -2.61469513e-01
-6.15611970e-01 -1.13084245e+00 -5.95988333e-01 -8.40605348e-02
1.31987378e-01 -8.18825126e-01 1.28423738e+00 -8.91022921e-01
-1.07412851e+00 1.44189322e+00 -3.37795494e-03 -1.42520666e-01
1.19744027e+00 -9.26259533e-02 4.54553887e-02 -2.19065413e-01
3.68127793e-01 5.47236681e-01 1.64634734e-01 -8.27662468e-01
-5.81920326e-01 -2.93438047e-01 -6.50157705e-02 1.88284904e-01
-3.54981452e-01 9.32583988e-01 2.57268269e-02 -3.37884784e-01
-8.68757725e-01 -5.48896670e-01 -2.43682176e-01 3.44713420e-01
-2.79286593e-01 -2.25490332e-01 5.58974221e-03 -5.72754323e-01
1.45710099e+00 -2.12272525e+00 -6.02659047e-01 1.01991728e-01
8.02884042e-01 2.29539394e-01 2.08543446e-02 3.68755370e-01
2.03045815e-01 7.29442954e-01 5.07440977e-03 -6.44887626e-01
5.54987848e-01 -1.03229351e-01 -7.37697631e-02 8.99609268e-01
-2.71043628e-01 7.44796276e-01 -1.17404854e+00 -8.43364477e-01
2.90067732e-01 1.44058898e-01 -7.92354524e-01 3.57048452e-01
3.02081525e-01 3.43320966e-01 -1.26469851e-01 4.48616564e-01
1.03340471e+00 1.47294149e-01 9.14494470e-02 1.37860641e-01
-5.25876939e-01 5.95390499e-01 -8.11959982e-01 1.27666175e+00
6.82320893e-02 2.28144899e-01 2.27165192e-01 -5.18171728e-01
6.38804018e-01 2.57116668e-02 4.44115549e-01 -8.01326692e-01
3.23537409e-01 2.63842881e-01 2.47784350e-02 -8.69849324e-02
7.07468927e-01 -3.92391920e-01 -2.10949838e-01 9.93414283e-01
-9.82730612e-02 -3.94420743e-01 5.42688072e-01 2.39369348e-01
9.98499870e-01 -1.76631197e-01 6.66682482e-01 -7.18983412e-01
2.49648482e-01 -3.07325006e-01 6.93770528e-01 1.09376180e+00
-1.32222581e+00 3.18834782e-01 8.50759983e-01 -8.19330394e-01
-9.25102830e-01 -1.01274395e+00 -3.55906546e-01 1.73554683e+00
-2.63084590e-01 -7.02285230e-01 -7.02207625e-01 -6.25651479e-01
3.78185451e-01 8.02803457e-01 -8.44297886e-01 -2.87253968e-02
-1.39681280e-01 -8.90600204e-01 1.11407471e+00 1.44158095e-01
4.91872936e-01 -8.16847324e-01 -7.84868717e-01 -2.19578847e-01
-3.36969793e-01 -7.56478906e-01 -6.83882535e-01 -3.95655960e-01
-6.57933891e-01 -1.33794796e+00 -1.72942564e-01 -8.40487331e-03
2.37376973e-01 2.17991322e-02 1.67843342e+00 6.89345837e-01
-4.09188896e-01 1.03318289e-01 8.41973796e-02 -5.92640340e-01
-4.76003498e-01 -1.94547668e-01 -8.20020512e-02 -6.77260697e-01
5.89909136e-01 -2.32738927e-01 -9.86830592e-01 3.65463912e-01
-6.07062161e-01 3.40296701e-03 -2.43628591e-01 6.99431360e-01
3.76537144e-01 -3.72432023e-01 5.29140592e-01 -1.31653762e+00
1.06628454e+00 -4.37387615e-01 -5.69594502e-01 -5.28058931e-02
-1.18173599e+00 -3.81024897e-01 5.92034042e-01 5.71498454e-01
-9.44237828e-01 -8.75212252e-01 -1.37902454e-01 1.44533724e-01
-4.51278687e-02 -1.14574539e-03 1.94884494e-01 -1.01878330e-01
9.61950541e-01 -5.81259310e-01 2.61575609e-01 -2.30048433e-01
6.05130613e-01 6.24721825e-01 3.78249854e-01 -1.08886921e+00
2.26657405e-01 3.04010153e-01 -9.24962610e-02 -1.33948237e-01
-8.29086542e-01 -3.57716113e-01 -1.83973357e-01 -3.51667553e-01
5.86799026e-01 -7.33200490e-01 -1.16367054e+00 2.98540682e-01
-7.26883173e-01 -5.93724370e-01 -2.69669771e-01 3.77353244e-02
-4.29164350e-01 5.65612137e-01 -9.53215897e-01 -7.80904233e-01
-1.06214523e+00 -1.01381600e+00 6.54776931e-01 4.22024846e-01
-7.73407757e-01 -9.35566783e-01 3.00574720e-01 8.99675727e-01
8.65190506e-01 3.51252466e-01 6.60296082e-01 -5.95358133e-01
5.19494936e-02 1.43130965e-04 -5.83348811e-01 4.79164988e-01
-6.61139265e-02 5.89912236e-01 -1.19853592e+00 -1.53293163e-01
-4.29804593e-01 -8.24483454e-01 6.85750067e-01 2.54453748e-01
1.43621933e+00 -4.28555518e-01 2.57696331e-01 6.49102092e-01
1.17849290e+00 -5.67550421e-01 5.91204941e-01 3.22193772e-01
4.11718309e-01 7.07613587e-01 5.55840909e-01 9.39027607e-01
7.44144201e-01 3.09974283e-01 2.84218758e-01 -1.94596469e-01
6.16068505e-02 2.62778014e-01 -2.03249589e-01 4.70935315e-01
-6.51084304e-01 1.07097447e-01 -1.04779959e+00 1.34200767e-01
-1.79801130e+00 -1.46882987e+00 -2.45663524e-01 2.40604758e+00
1.20786726e+00 -1.93841383e-01 5.73895156e-01 -1.44046456e-01
5.91074586e-01 1.40656888e-01 -3.70790511e-01 -1.21215129e+00
2.26186037e-01 3.64021629e-01 2.90682703e-01 7.58986712e-01
-1.06805205e+00 8.94493461e-01 6.63673544e+00 8.65512311e-01
-7.70498335e-01 2.12368906e-01 1.15099537e+00 -5.54507673e-01
-4.67633188e-01 -1.34106860e-01 -2.56902426e-01 5.93240917e-01
9.91160512e-01 -8.08573544e-01 7.99148142e-01 7.83947229e-01
5.27795553e-01 -1.62017688e-01 -9.80117559e-01 6.29294097e-01
-1.49796858e-01 -1.13475370e+00 -4.74268764e-01 -1.89325511e-02
6.38683975e-01 1.72099292e-01 1.66951045e-01 3.37761909e-01
8.13376904e-01 -1.33260000e+00 8.23824584e-01 4.31743920e-01
1.00664485e+00 -8.80824387e-01 8.27347755e-01 -6.92747459e-02
-6.53681636e-01 2.10494503e-01 -5.62196970e-01 -5.19800127e-01
-2.22843423e-01 9.72192585e-01 -2.35877395e-01 6.20687723e-01
1.08445382e+00 6.27731979e-01 -5.88565469e-01 8.38321567e-01
-4.34675850e-02 6.20071948e-01 4.02150489e-02 7.12343305e-02
-1.98226318e-01 -6.01911917e-02 -6.33498132e-02 1.46427751e+00
-2.93416500e-01 2.01611489e-01 4.00014147e-02 1.21268392e+00
-4.53917086e-01 4.48448002e-01 -2.48233482e-01 -8.13580528e-02
6.27672434e-01 1.67203891e+00 -5.20192087e-01 -2.22418115e-01
-4.03692365e-01 4.60705638e-01 8.04459155e-01 -1.42551661e-01
-9.75400150e-01 -2.86303401e-01 1.07259738e+00 -1.19155884e-01
-7.62343287e-01 2.14294523e-01 -8.75893474e-01 -9.31877017e-01
-3.66046250e-01 -1.24105203e+00 9.55389798e-01 -3.28154445e-01
-1.78003919e+00 2.19839111e-01 -8.06061551e-03 -6.65533066e-01
-2.89603625e-03 -2.87430763e-01 -6.73098087e-01 1.32805467e+00
-1.45655811e+00 -9.85116601e-01 -6.47765875e-01 4.23862487e-01
-2.59285659e-01 -3.15773413e-02 1.05975664e+00 4.70507622e-01
-6.28680229e-01 1.05067945e+00 -2.19718307e-01 -4.22358215e-02
1.57026696e+00 -1.32309031e+00 4.84470069e-01 5.96433043e-01
-7.50543118e-01 1.08337486e+00 6.63553834e-01 -3.95606935e-01
-8.28206062e-01 -9.34377491e-01 1.19811726e+00 -8.94777238e-01
6.52498782e-01 -2.39603654e-01 -5.29860735e-01 4.73998725e-01
2.46521965e-01 2.23995656e-01 1.50116920e+00 4.59373742e-01
-6.63500011e-01 -1.36169791e-01 -1.63810289e+00 5.78900933e-01
1.12924409e+00 -6.11942768e-01 -9.67859849e-02 2.22505584e-01
2.08555050e-02 -4.63829637e-01 -1.05858815e+00 1.36129349e-01
1.06512284e+00 -1.70557022e+00 7.26069570e-01 -7.77605534e-01
1.07112157e+00 2.03900233e-01 2.43007950e-02 -8.95735919e-01
-5.14362693e-01 -6.93854511e-01 3.50451261e-01 1.39250541e+00
6.74356967e-02 -8.68874371e-01 5.84322214e-01 1.45682430e+00
1.57856688e-01 -7.13938773e-01 -8.42920244e-01 -2.46638656e-01
7.35263586e-01 -4.49354887e-01 8.89478028e-01 1.30207253e+00
2.36912206e-01 -1.84113964e-01 -4.86488134e-01 -5.41027904e-01
1.00762177e+00 1.02344140e-01 9.44814444e-01 -9.40021694e-01
-4.40840758e-02 -1.03829229e+00 -2.56793052e-01 -4.37778719e-02
2.87608474e-01 -1.01544356e+00 -1.93187431e-01 -1.29230547e+00
8.82712960e-01 -7.13616788e-01 -4.45265055e-01 8.68189871e-01
-6.35589540e-01 5.73098481e-01 4.28646207e-01 1.95780143e-01
-9.09289479e-01 3.28746080e-01 1.02065825e+00 2.01082274e-01
3.52828115e-01 -1.08908847e-01 -1.37596297e+00 3.00725073e-01
9.62398887e-01 -6.29287302e-01 -1.04572214e-01 -3.29428494e-01
1.69323325e-01 -4.66721714e-01 4.22446787e-01 -5.19544005e-01
-9.37282515e-04 -6.81699216e-01 2.68704414e-01 1.59832492e-01
-4.17040616e-01 -3.63104641e-01 1.15714157e-02 4.74875927e-01
-6.34021282e-01 1.50367141e-01 2.51287907e-01 -9.80057642e-02
2.14309335e-01 -9.94484425e-02 1.12998176e+00 -3.73306155e-01
-3.44468892e-01 3.89551371e-01 4.93135005e-02 6.10168457e-01
7.14028597e-01 2.52315134e-01 -1.10985231e+00 -3.42805803e-01
3.31876017e-02 5.59043825e-01 8.20373952e-01 8.49133134e-02
-4.50816229e-02 -1.20278800e+00 -9.45276797e-01 -5.06804436e-02
2.43108168e-01 -4.86458898e-01 3.49281251e-01 8.95450890e-01
-8.74748826e-01 3.07086557e-02 -7.72490382e-01 -5.37522789e-03
-1.30707324e+00 2.06577420e-01 5.38824618e-01 -2.02062398e-01
-8.34877416e-03 9.52403843e-01 -1.25672653e-01 -8.84732604e-01
9.01000798e-02 2.45930985e-01 1.35256201e-01 -2.54764289e-01
8.90894890e-01 1.00800407e+00 1.71046127e-02 -5.50098896e-01
-6.02467537e-01 -1.19204000e-01 4.87729684e-02 -5.33171743e-02
1.02592313e+00 -5.53938001e-02 -8.20987880e-01 1.31418660e-01
9.36328709e-01 3.93821836e-01 -8.09745431e-01 2.53998727e-01
-3.29052180e-01 -9.85971868e-01 -5.55761270e-02 -9.50847566e-01
-8.95817816e-01 7.34863102e-01 3.73239875e-01 1.36973127e-03
9.58418190e-01 -5.17775655e-01 2.85317034e-01 -2.03756616e-01
4.77213323e-01 -1.27318275e+00 -5.10806620e-01 3.77683550e-01
5.88077724e-01 -1.36235189e+00 6.24886394e-01 -2.43505120e-01
-5.51943183e-01 9.88864124e-01 7.85942435e-01 3.24267566e-01
5.58408320e-01 6.65423498e-02 5.17453671e-01 -1.51753396e-01
-6.54824317e-01 1.84278369e-01 2.48339221e-01 3.43280584e-01
1.21557486e+00 3.38265598e-01 -9.55661416e-01 7.88385332e-01
-5.96453726e-01 2.23300084e-01 4.16138679e-01 8.74150276e-01
-1.33684322e-01 -1.25248992e+00 -1.08090430e-01 7.33229518e-01
-9.70899761e-01 -3.08041573e-01 -4.88084644e-01 6.22965276e-01
8.01263750e-02 1.25792551e+00 -1.65648848e-01 -1.68454155e-01
1.85039714e-01 -8.84637237e-02 4.37739491e-01 -5.67712128e-01
-1.12608075e+00 -5.01798511e-01 4.65581477e-01 -1.10564482e+00
-4.16053176e-01 -5.91630459e-01 -9.56000924e-01 -1.65598273e+00
1.44242659e-01 2.75264651e-01 3.94756496e-01 6.97133660e-01
3.12963873e-01 1.34864271e-01 4.57432747e-01 -6.02646351e-01
-6.53240144e-01 -5.75729311e-01 -5.18538356e-01 7.24254847e-01
-3.86755876e-02 -2.81333685e-01 -4.83648181e-01 -4.33934480e-01] | [8.923243522644043, 5.29857063293457] |
be3f3b8f-f59d-4798-b8d3-d03da6aef614 | investigation-of-the-challenges-of-underwater | 2306.08738 | null | https://arxiv.org/abs/2306.08738v1 | https://arxiv.org/pdf/2306.08738v1.pdf | Investigation of the Challenges of Underwater-Visual-Monocular-SLAM | In this paper, we present a comprehensive investigation of the challenges of Monocular Visual Simultaneous Localization and Mapping (vSLAM) methods for underwater robots. While significant progress has been made in state estimation methods that utilize visual data in the past decade, most evaluations have been limited to controlled indoor and urban environments, where impressive performance was demonstrated. However, these techniques have not been extensively tested in extremely challenging conditions, such as underwater scenarios where factors such as water and light conditions, robot path, and depth can greatly impact algorithm performance. Hence, our evaluation is conducted in real-world AUV scenarios as well as laboratory settings which provide precise external reference. A focus is laid on understanding the impact of environmental conditions, such as optical properties of the water and illumination scenarios, on the performance of monocular vSLAM methods. To this end, we first show that all methods perform very well in in-air settings and subsequently show the degradation of their performance in challenging underwater environments. The final goal of this study is to identify techniques that can improve accuracy and robustness of SLAM methods in such conditions. To achieve this goal, we investigate the potential of image enhancement techniques to improve the quality of input images used by the SLAM methods, specifically in low visibility and extreme lighting scenarios in scattering media. We present a first evaluation on calibration maneuvers and simple image restoration techniques to determine their ability to enable or enhance the performance of monocular SLAM methods in underwater environments. | ['Kevin Köser', 'Mengkun She', 'David Nakath', 'Michele Grimaldi'] | 2023-06-14 | null | null | null | null | ['simultaneous-localization-and-mapping', 'image-enhancement', 'image-restoration'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.61733717e-01 -3.82406056e-01 7.62862921e-01 -3.24692398e-01
-3.85222495e-01 -7.62726903e-01 4.78608072e-01 8.46814141e-02
-9.50590670e-01 9.11048412e-01 -8.69157836e-02 -1.92509964e-01
-6.86588734e-02 -4.69268680e-01 -7.38146901e-01 -6.78416371e-01
-4.20201451e-01 2.07750142e-01 3.12852740e-01 -3.67482036e-01
4.62084681e-01 7.73239374e-01 -1.68552208e+00 -7.39231586e-01
6.55766547e-01 3.72409791e-01 4.24127430e-01 8.69550884e-01
1.49462715e-01 2.45740652e-01 -6.23754978e-01 1.15283169e-01
4.63604957e-01 -3.93274069e-01 -3.04205984e-01 -1.22784525e-02
6.12369359e-01 -6.38158083e-01 -2.29847774e-01 9.34959471e-01
9.01473403e-01 3.14107329e-01 3.68622422e-01 -8.02552640e-01
1.37534603e-01 -5.99274188e-02 -1.83395490e-01 2.07349166e-01
5.06957412e-01 2.60883927e-01 5.00358462e-01 -8.35617065e-01
4.78118896e-01 9.05376017e-01 8.27169418e-01 2.25881472e-01
-1.03595185e+00 -4.29170996e-01 -1.36341110e-01 6.24314919e-02
-1.22273028e+00 -7.85422146e-01 3.31400007e-01 -3.11411560e-01
1.03516495e+00 -1.47712708e-01 6.49901628e-01 3.41068029e-01
4.15191680e-01 2.13464200e-01 1.17817664e+00 -5.79126894e-01
3.06718439e-01 1.10240526e-01 -1.99029148e-01 6.33955240e-01
7.17789173e-01 2.04297483e-01 -7.07607508e-01 -3.04903183e-02
4.45944458e-01 -3.40015382e-01 -7.13859737e-01 -6.66102231e-01
-7.24609375e-01 7.11969972e-01 3.94374132e-01 3.26644932e-03
-1.30239949e-01 2.11345538e-01 1.80022612e-01 5.28752923e-01
1.29900545e-01 7.23854005e-01 -2.38694102e-01 -2.52145350e-01
-7.20594168e-01 1.42948285e-01 8.59659851e-01 7.51610935e-01
1.07312763e+00 3.13747644e-01 6.81117713e-01 6.44180477e-01
6.67035043e-01 9.56463277e-01 2.24806666e-01 -9.71199155e-01
3.42414528e-01 8.36321786e-02 5.65662920e-01 -8.45465183e-01
-6.74320400e-01 -2.45003775e-01 6.78699911e-02 8.03061724e-01
1.93046913e-01 -4.91760731e-01 -8.56015325e-01 1.36558402e+00
1.86349109e-01 -8.05369485e-03 8.47579658e-01 1.03463507e+00
7.40550697e-01 3.49015802e-01 -4.72007722e-01 -2.81430960e-01
8.55833888e-01 -4.30229902e-01 -6.61848605e-01 -6.77109718e-01
5.43116093e-01 -1.00051785e+00 6.33178473e-01 5.59014007e-02
-8.33438218e-01 -1.47335634e-01 -1.39758956e+00 3.07610482e-01
-1.86458275e-01 -1.42528147e-01 3.96476418e-01 6.63667500e-01
-1.17751265e+00 3.99515301e-01 -1.31819868e+00 -8.23162496e-01
-3.24463725e-01 4.04769301e-01 -6.07963800e-01 -3.05066407e-01
-9.85801876e-01 1.42231333e+00 -2.05403224e-01 3.21482778e-01
-8.71228278e-01 -3.12878758e-01 -1.35883284e+00 -3.78330439e-01
-8.28470290e-02 -1.65734574e-01 1.22598481e+00 -4.41328436e-01
-1.47434032e+00 4.69840229e-01 -3.60266149e-01 -5.84244132e-01
4.56299245e-01 -4.05723035e-01 1.79438889e-01 2.16807231e-01
1.17986277e-01 4.74965245e-01 2.63758212e-01 -1.68786907e+00
-4.25393879e-01 -3.67967844e-01 2.41520256e-01 8.00511837e-01
4.14794832e-02 -1.55563489e-01 -4.15380359e-01 2.87673384e-01
5.11103570e-01 -1.21984684e+00 -2.35316217e-01 2.03637183e-01
2.13932082e-01 6.83201969e-01 7.25301504e-01 -3.11205953e-01
2.73489088e-01 -1.93962717e+00 -2.04691023e-01 9.03593004e-02
-5.68229139e-01 1.24239460e-01 6.91843405e-02 1.07767808e+00
7.06151187e-01 -2.40542203e-01 -8.50310475e-02 -7.59176672e-01
-2.41599143e-01 7.37177730e-01 -3.96064669e-02 1.05154467e+00
-3.75395715e-01 2.16380239e-01 -7.82545924e-01 -2.58789897e-01
6.04006290e-01 5.06561697e-01 -4.52313185e-01 3.59141767e-01
4.03438747e-01 6.67705655e-01 1.21018030e-01 6.58471048e-01
7.22934902e-01 5.23313582e-01 1.99719846e-01 -6.97425008e-02
-8.82640064e-01 1.45710438e-01 -1.23717952e+00 1.45544040e+00
-8.03302824e-01 1.11640561e+00 7.15328038e-01 -3.38997215e-01
1.11477923e+00 3.26780081e-02 2.42786393e-01 -8.74174058e-01
-4.21142264e-04 4.15593475e-01 6.37672395e-02 -6.28194094e-01
9.14826155e-01 -5.91856658e-01 3.38951439e-01 1.32617494e-02
-2.97953635e-01 -4.94041741e-01 5.92542551e-02 1.83721021e-01
7.79510021e-01 2.09442586e-01 1.64532006e-01 -5.21923423e-01
3.05805281e-02 1.87335566e-01 4.41308916e-01 8.83910060e-01
-2.42350519e-01 7.44778633e-01 -1.21305563e-01 -2.30739579e-01
-8.44897687e-01 -8.43359530e-01 -3.67173225e-01 4.76717174e-01
1.15428567e+00 2.29577981e-02 1.70726050e-03 1.39147520e-01
1.26909927e-01 3.38086843e-01 -4.86008674e-01 1.00824535e-01
-4.43233699e-01 -1.02776301e+00 4.84934598e-01 3.64084780e-01
4.43703979e-01 -6.55577660e-01 -1.23776042e+00 3.78836513e-01
7.95944110e-02 -1.33143747e+00 3.74080658e-01 4.30302262e-01
-7.96257496e-01 -1.08967519e+00 -4.11194474e-01 -6.19718492e-01
5.40488958e-01 6.86127603e-01 6.56115472e-01 1.61586165e-01
-2.11133331e-01 6.64228261e-01 -7.41228998e-01 -4.09106851e-01
-3.17819268e-01 -4.89573896e-01 4.35123503e-01 -3.08350712e-01
-8.80068466e-02 -4.64773595e-01 -7.24560857e-01 5.08557618e-01
-6.84738100e-01 -3.99835885e-01 4.05241281e-01 6.10170603e-01
1.04087465e-01 -4.05818105e-01 -8.39390734e-04 -3.28263640e-01
1.62210122e-01 -3.17494303e-01 -8.95890057e-01 -2.59380698e-01
-5.07631481e-01 -1.29659325e-01 2.25970820e-01 4.49014455e-02
-8.26405823e-01 2.53111959e-01 -2.08321512e-01 -1.20899845e-02
1.96866039e-02 8.03849816e-01 2.67482549e-01 -8.99983525e-01
7.10682929e-01 4.31344092e-01 3.53447676e-01 -3.17611337e-01
-3.38157117e-01 7.28717983e-01 4.25208271e-01 -2.57093906e-01
1.01161110e+00 9.91461277e-01 2.10941434e-01 -1.52970672e+00
-3.42361391e-01 -6.45352066e-01 -4.66547698e-01 -3.81659091e-01
5.60895801e-01 -1.14660490e+00 -3.14555049e-01 5.44663489e-01
-7.15858996e-01 -6.51152611e-01 4.23689604e-01 9.07489896e-01
-2.10719675e-01 7.86210835e-01 -3.74290794e-01 -1.13730884e+00
-1.30680934e-01 -1.50725019e+00 8.59290540e-01 5.84931135e-01
1.93191364e-01 -1.16266239e+00 3.48316193e-01 9.27202925e-02
6.36311471e-01 1.58520669e-01 7.93838874e-03 -2.65400708e-02
-5.74269950e-01 -8.14561918e-02 3.10731009e-02 3.95197086e-02
2.73524582e-01 -4.39240746e-02 -8.95938396e-01 -1.02286553e+00
-4.37142849e-01 -4.49119657e-01 6.83644712e-01 1.45871356e-01
-2.85775304e-01 2.47126892e-01 -2.17835784e-01 8.78344476e-01
1.71175182e+00 3.39591473e-01 6.09151125e-01 8.33047986e-01
4.14653122e-01 6.01508260e-01 8.71814251e-01 4.34458792e-01
4.14265692e-01 7.77974665e-01 9.74018395e-01 -1.79258600e-01
5.56934774e-02 1.83181524e-01 2.56749898e-01 3.39392006e-01
-1.23406298e-01 -3.05498272e-01 -7.92959869e-01 7.81892419e-01
-1.34686601e+00 -5.29916883e-01 -2.65392244e-01 2.27131152e+00
1.28274947e-01 -1.65551856e-01 -6.01257622e-01 -4.36747819e-02
3.76769006e-01 3.09999399e-02 -3.52720648e-01 -2.28450760e-01
-2.70534694e-01 -2.45712548e-01 1.05301726e+00 1.05815136e+00
-8.17201018e-01 7.86101103e-01 6.17630339e+00 -5.58698237e-01
-1.34481215e+00 -3.14032197e-01 -4.94856894e-01 2.39237383e-01
-2.20780522e-01 2.99952120e-01 -9.32837844e-01 8.26626793e-02
5.54938853e-01 2.57392079e-01 9.91505831e-02 6.63786292e-01
4.79390889e-01 -6.53581440e-01 -5.63810766e-01 7.52567887e-01
2.48185560e-01 -8.41845095e-01 -4.89033490e-01 2.42467985e-01
7.73985684e-01 6.26199841e-01 -3.41098636e-01 -7.02366233e-02
2.36951798e-01 -4.99850184e-01 7.08763599e-01 1.44536465e-01
5.79980552e-01 -4.43939090e-01 1.21602750e+00 8.29486027e-02
-9.84453321e-01 9.05290768e-02 -5.93648732e-01 -6.69681430e-01
4.67269748e-01 2.05205619e-01 -1.16092181e+00 4.01041090e-01
8.85617077e-01 3.80635828e-01 -2.65422404e-01 1.65385914e+00
-2.52392054e-01 4.38049763e-01 -8.47971380e-01 -2.97013074e-01
5.14689863e-01 -1.24297887e-01 6.81915700e-01 1.15394175e+00
4.51331019e-01 8.58190879e-02 5.23901619e-02 2.23635778e-01
5.06533563e-01 6.46947175e-02 -8.89977336e-01 4.34982300e-01
5.09690642e-01 9.75341558e-01 -6.19005978e-01 1.78567201e-01
-5.66866219e-01 7.35609889e-01 2.68391334e-02 4.02145654e-01
-2.80064493e-01 -4.08419520e-01 1.09094596e+00 -5.64541258e-02
2.47366503e-01 -9.37059462e-01 -1.47002071e-01 -9.75678623e-01
-1.22017927e-01 -4.03959185e-01 -2.14475691e-02 -7.97531366e-01
-5.94736755e-01 3.96768987e-01 -1.15260715e-03 -1.37115049e+00
3.95452566e-02 -5.59154630e-01 -5.61423361e-01 8.31090987e-01
-1.99701405e+00 -7.46278346e-01 -9.13411856e-01 5.62883504e-02
3.83134037e-01 9.85590518e-02 6.19275689e-01 9.99342874e-02
5.14433682e-02 2.32178286e-01 6.38859272e-01 -1.41080245e-01
8.10150504e-01 -1.16292512e+00 1.79786190e-01 1.43405020e+00
-8.99979379e-03 7.16451347e-01 1.41435909e+00 -7.86999941e-01
-1.94503748e+00 -5.47689915e-01 3.11688125e-01 -1.64737999e-01
5.30110598e-01 -1.19106479e-01 -6.77515388e-01 5.64397097e-01
2.56761778e-02 -6.84241578e-02 3.87228727e-01 -2.20692828e-01
3.16366136e-01 -2.13009324e-02 -9.59387004e-01 4.41831231e-01
5.01853287e-01 -1.85348913e-01 -3.54814917e-01 8.04440379e-02
8.34826455e-02 -7.71695137e-01 -5.88149548e-01 8.78900230e-01
8.40164661e-01 -1.21207738e+00 7.02935934e-01 1.42641604e-01
-1.20959491e-01 -7.36988902e-01 -4.89546686e-01 -1.39376128e+00
2.76906461e-01 -3.47105414e-01 6.87164068e-01 8.11010897e-01
3.67642075e-01 -9.45196211e-01 7.42528558e-01 3.28331500e-01
-5.80995739e-01 -1.27909034e-01 -1.01411235e+00 -6.73067927e-01
-1.67812288e-01 -1.31857246e-01 -1.45405516e-01 4.01000977e-01
-2.20618606e-01 -2.48599201e-02 -6.01278901e-01 1.12737572e+00
7.35736728e-01 3.60927761e-01 1.19251537e+00 -9.06318486e-01
-1.63446099e-01 3.42589766e-01 -7.97983050e-01 -9.93008435e-01
-1.29120708e-01 -7.01349825e-02 7.66138852e-01 -1.97366464e+00
-1.92251831e-01 -5.91761231e-01 3.10935348e-01 1.82125896e-01
-7.60201961e-02 7.84305096e-01 1.54096931e-01 3.87023777e-01
-2.97523379e-01 6.52785242e-01 7.76255608e-01 1.95395976e-01
-2.74686456e-01 3.32244337e-02 -5.71262129e-02 4.78888392e-01
6.30160928e-01 -2.07942516e-01 -3.63428950e-01 -1.00076044e+00
9.85804945e-02 1.26461983e-01 1.44982323e-01 -1.38339770e+00
5.52262783e-01 6.37563737e-03 6.99706823e-02 -2.53810644e-01
7.35777795e-01 -7.75147855e-01 2.00407784e-02 6.72589481e-01
2.77084976e-01 8.21372345e-02 3.14414680e-01 5.30967057e-01
-3.48695844e-01 -6.76074684e-01 1.00738227e+00 -2.22771972e-01
-1.19569361e+00 -2.74844795e-01 -5.32240093e-01 -1.96689337e-01
8.66187215e-01 -5.14313400e-01 -3.45344156e-01 -7.89418638e-01
-3.52253139e-01 3.80598456e-01 1.18163896e+00 7.02651171e-03
7.23090649e-01 -3.44159096e-01 -5.71388066e-01 2.05626026e-01
2.63478816e-01 -3.04741785e-02 7.95547143e-02 7.38527775e-01
-1.37027490e+00 -1.12827547e-01 -2.46788293e-01 -6.96265638e-01
-1.51151931e+00 -6.47689477e-02 7.05547452e-01 5.76273263e-01
-6.17627919e-01 7.50569224e-01 -9.10487026e-02 -4.13820356e-01
1.48733050e-01 3.43335159e-02 -1.46882996e-01 -2.09533617e-01
4.37479943e-01 3.07822078e-01 2.71102160e-01 -6.90199077e-01
-5.38343072e-01 9.35939014e-01 3.11642021e-01 -2.87005365e-01
1.22003448e+00 -7.39459395e-01 1.32833973e-01 3.14228237e-01
9.71881449e-01 3.84745091e-01 -1.63429701e+00 8.71031880e-02
-4.47672933e-01 -4.44917023e-01 1.08036362e-01 -3.68252665e-01
-6.32630527e-01 8.10716510e-01 7.28675604e-01 -9.64766443e-02
7.20225990e-01 -6.06338345e-02 3.63906920e-01 7.06006050e-01
7.90099859e-01 -6.40422165e-01 -1.20224625e-01 9.41544890e-01
4.87493366e-01 -1.36967039e+00 3.22078854e-01 -1.90245867e-01
-4.51349199e-01 1.20618391e+00 4.70568061e-01 -1.11885136e-02
2.69222826e-01 6.56666994e-01 1.02579963e+00 -2.12965924e-02
-1.43072173e-01 -5.17079175e-01 -4.95158821e-01 4.42853779e-01
1.82143934e-02 -3.57869864e-01 -2.19720289e-01 -5.54387331e-01
-2.99327523e-01 -5.42102158e-01 1.19928288e+00 1.37160993e+00
-8.83930206e-01 -6.84108078e-01 -6.21486545e-01 -7.67354146e-02
-3.08138728e-01 5.90257272e-02 -4.31088507e-02 1.08507144e+00
-1.65259033e-01 1.01316249e+00 -6.25136793e-02 -2.42973089e-01
3.89969230e-01 -2.85594940e-01 3.76209915e-01 -4.90303814e-01
-8.60069245e-02 -1.46286055e-01 4.03621823e-01 -2.23431408e-01
-8.79632652e-01 -9.84261215e-01 -1.41123855e+00 1.71914980e-01
-4.17238057e-01 5.92340529e-01 1.30749178e+00 1.05821288e+00
-7.11375698e-02 7.52470270e-02 3.96358311e-01 -1.44038999e+00
-2.74231106e-01 -1.05327296e+00 -5.09671867e-01 -1.02828197e-01
5.09290397e-01 -1.04263103e+00 -9.05110776e-01 -2.73236841e-01] | [7.543638706207275, -1.7968518733978271] |
a915ece3-4c0f-44ad-b403-9523f01043c8 | ad-yolo-you-look-only-once-in-training | 2303.15703 | null | https://arxiv.org/abs/2303.15703v3 | https://arxiv.org/pdf/2303.15703v3.pdf | AD-YOLO: You Look Only Once in Training Multiple Sound Event Localization and Detection | Sound event localization and detection (SELD) combines the identification of sound events with the corresponding directions of arrival (DOA). Recently, event-oriented track output formats have been adopted to solve this problem; however, they still have limited generalization toward real-world problems in an unknown polyphony environment. To address the issue, we proposed an angular-distance-based multiple SELD (AD-YOLO), which is an adaptation of the "You Only Look Once" algorithm for SELD. The AD-YOLO format allows the model to learn sound occurrences location-sensitively by assigning class responsibility to DOA predictions. Hence, the format enables the model to handle the polyphony problem, regardless of the number of sound overlaps. We evaluated AD-YOLO on DCASE 2020-2022 challenge Task 3 datasets using four SELD objective metrics. The experimental results show that AD-YOLO achieved outstanding performance overall and also accomplished robustness in class-homogeneous polyphony environments. | ['Sung Won Han', 'WooSeok Shin', 'Hyun Joon Park', 'Jin Sob Kim'] | 2023-03-28 | null | null | null | null | ['sound-event-detection', 'direction-of-arrival-estimation', 'sound-event-localization-and-detection'] | ['audio', 'audio', 'audio'] | [-3.03746134e-01 -7.12218404e-01 2.17693508e-01 -2.81467527e-01
-1.43628693e+00 -4.49509352e-01 1.91188529e-01 6.52153268e-02
-1.52174249e-01 4.05539781e-01 2.57993788e-01 1.59101933e-01
-2.97365546e-01 -4.76583868e-01 -5.61024606e-01 -7.22615063e-01
-3.66251379e-01 4.52117957e-02 8.46962273e-01 8.85566771e-02
1.98609918e-01 2.27833182e-01 -1.77049839e+00 4.53574002e-01
2.89156616e-01 1.18825817e+00 4.85109478e-01 8.41063380e-01
2.28068471e-01 6.89242542e-01 -7.66938329e-01 2.97942907e-01
1.25938013e-01 -5.43339789e-01 -1.50655285e-01 -6.64126277e-01
5.46607733e-01 -1.53390309e-02 -3.12267691e-01 6.41778708e-01
1.09883201e+00 3.35547715e-01 3.40608299e-01 -1.41622615e+00
6.48186505e-02 5.62096238e-01 -1.54570252e-01 9.77464437e-01
4.12719965e-01 -1.52297229e-01 1.30881774e+00 -1.39534760e+00
-1.88943241e-02 9.86323476e-01 1.06154919e+00 6.26996979e-02
-5.49846947e-01 -8.84099722e-01 -5.92436679e-02 5.36633134e-01
-1.62945211e+00 -5.20707965e-01 8.28985572e-01 -3.91287982e-01
1.11302578e+00 5.64793110e-01 3.63651186e-01 9.28066432e-01
1.23865604e-01 7.20882893e-01 6.68689728e-01 -4.85318959e-01
2.89034963e-01 -2.76737660e-01 2.82570701e-02 1.94555134e-01
-2.44262949e-01 3.60622287e-01 -1.43134832e+00 -1.23411089e-01
2.77537495e-01 -3.81378263e-01 -1.60177797e-01 3.25523734e-01
-1.13631260e+00 3.16827208e-01 2.75349885e-01 1.05737686e-01
-1.72605336e-01 2.99467742e-01 4.61717606e-01 1.77787215e-01
6.18329227e-01 3.46413493e-01 -5.46563327e-01 -2.37673268e-01
-1.06863177e+00 3.74596089e-01 5.55229604e-01 8.88786256e-01
3.61386567e-01 3.86724561e-01 -4.44751441e-01 9.37044144e-01
5.21014214e-01 5.93378305e-01 3.55921656e-01 -5.04227281e-01
6.78461015e-01 -1.82594344e-01 2.33253594e-02 -1.11684275e+00
-7.84246743e-01 -9.31389749e-01 -4.69898224e-01 -2.67315209e-01
9.62215066e-02 -1.90491199e-01 -6.38695776e-01 1.64151442e+00
3.41095477e-01 7.64382184e-01 -2.19744340e-01 9.13221538e-01
7.48169541e-01 9.15373027e-01 -4.09007408e-02 -2.27639660e-01
1.32801008e+00 -8.33563030e-01 -9.47345197e-01 -2.04347342e-01
2.63425857e-01 -9.35823619e-01 9.71551538e-01 4.88244832e-01
-7.05513477e-01 -9.12393332e-01 -8.80671620e-01 4.72025901e-01
-7.97900185e-02 3.92554313e-01 3.23628843e-01 7.30834603e-01
-6.06087267e-01 3.48076582e-01 -8.58684242e-01 -1.40857309e-01
8.97275284e-03 -1.13912299e-01 1.28717780e-01 3.58101875e-01
-1.54561126e+00 3.16510648e-01 2.84133196e-01 2.65987396e-01
-1.17891622e+00 -1.00612807e+00 -6.12097263e-01 2.73357958e-01
2.89417148e-01 1.14659080e-02 1.51344025e+00 -8.82570073e-02
-1.25066960e+00 1.69789433e-01 -1.86571419e-01 -5.28333426e-01
2.39066169e-01 -5.88334680e-01 -1.11949146e+00 1.83142886e-01
3.97022158e-01 2.42216319e-01 7.28987455e-01 -7.78327644e-01
-1.17589951e+00 1.46350786e-01 -3.78286451e-01 1.80006996e-01
-2.30137140e-01 3.38307232e-01 -3.74747396e-01 -1.01056325e+00
3.79751354e-01 -7.85753310e-01 1.39983043e-01 -1.02466829e-01
-4.04242367e-01 -3.35860759e-01 8.00934732e-01 -4.51156527e-01
1.80550420e+00 -2.58374977e+00 -5.81276596e-01 1.09394304e-02
-1.97869256e-01 1.14792652e-01 8.87426287e-02 7.88509369e-01
3.28138322e-02 -2.87194550e-01 6.83016405e-02 -2.80833960e-01
1.68858111e-01 -5.93805611e-02 -8.18087637e-01 2.60739088e-01
6.58303723e-02 1.76346526e-01 -8.20026875e-01 -5.23644090e-01
1.51439905e-01 2.26395458e-01 -7.75854468e-01 2.52938479e-01
-4.35652696e-02 1.87406659e-01 -3.51739198e-01 7.22365797e-01
5.61374307e-01 3.98080945e-02 -2.59756386e-01 -2.55024672e-01
-3.67859751e-01 5.99731266e-01 -1.52072418e+00 1.36760950e+00
-5.75312793e-01 9.14093733e-01 -2.73465723e-01 -4.68555540e-01
9.66181576e-01 7.57992685e-01 5.04572153e-01 -6.64001286e-01
-4.50515598e-01 3.77218008e-01 -1.09528504e-01 -5.22103488e-01
4.64208513e-01 -1.13073625e-01 -1.72879994e-01 1.71933606e-01
-1.46323517e-02 5.31828776e-02 1.43343024e-02 -7.10090920e-02
1.30387831e+00 -9.10402834e-02 1.95467725e-01 -1.25019848e-01
2.90704280e-01 -3.54124993e-01 9.32410419e-01 1.13877749e+00
-3.25342089e-01 1.13360250e+00 -2.11055186e-02 -3.00448179e-01
-2.08566114e-01 -1.57279694e+00 -3.68493885e-01 1.41199589e+00
1.56604305e-01 -7.43516386e-01 -3.34535956e-01 -3.99506927e-01
-7.38869980e-02 8.34997118e-01 -3.64957117e-02 -1.51497319e-01
-7.33868837e-01 -8.11003804e-01 1.06184173e+00 6.65079057e-01
4.97989953e-01 -1.16381800e+00 -6.29450977e-01 5.89895070e-01
-4.85838175e-01 -1.22973323e+00 -7.83258200e-01 4.47551042e-01
-2.05866456e-01 -7.04398215e-01 -3.04091990e-01 -5.65909743e-01
6.77726138e-03 4.17923808e-01 8.60471964e-01 -4.27610666e-01
-3.86662930e-01 4.86308068e-01 -7.41099536e-01 -7.74800360e-01
7.62006864e-02 -3.18929732e-01 3.48472238e-01 2.72812992e-01
1.17363669e-01 -7.59373486e-01 -7.19168246e-01 7.25580513e-01
-6.69677258e-01 -4.76489067e-01 3.42683077e-01 5.27841806e-01
6.71913862e-01 3.97739410e-01 1.25638163e+00 -3.77344042e-01
4.83124495e-01 -6.93301558e-01 -3.83115202e-01 3.69176082e-02
-5.88018060e-01 -5.69878340e-01 6.69361293e-01 -3.82773668e-01
-9.05061781e-01 2.13161990e-01 -4.83879089e-01 -3.91042054e-01
-2.74430849e-02 4.31383997e-01 -2.75714755e-01 2.80666113e-01
6.47312641e-01 4.48826671e-01 -8.73837888e-01 -8.23306203e-01
-1.25827014e-01 8.83210838e-01 7.84503341e-01 -4.72060889e-01
5.19926667e-01 2.37162739e-01 -1.38061911e-01 -8.67280662e-01
-9.89293039e-01 -8.70188296e-01 -1.93059653e-01 -4.46495742e-01
7.61087418e-01 -1.22629309e+00 -3.73682857e-01 5.00758946e-01
-1.11999500e+00 1.39018208e-01 -1.54374331e-01 9.17433143e-01
-4.12068397e-01 8.49999040e-02 -3.11478376e-01 -9.81898963e-01
-3.14118147e-01 -8.20255101e-01 1.16546261e+00 2.25441173e-01
-3.19040954e-01 -3.98399740e-01 4.78143245e-01 3.52526344e-02
3.85280907e-01 -9.07846987e-02 6.34772778e-01 -7.12842405e-01
-5.35180509e-01 -3.95836383e-01 2.46192679e-01 1.95921957e-01
8.01128000e-02 -2.56760627e-01 -1.52752197e+00 2.32427884e-02
-6.98590353e-02 1.05425827e-01 6.31868899e-01 4.71136004e-01
1.18212664e+00 -3.96288261e-02 -3.18017095e-01 4.39793587e-01
9.60395813e-01 4.78917956e-01 3.60113978e-01 2.14927197e-01
4.57083136e-01 1.64137721e-01 1.00751066e+00 8.28717947e-01
4.36827570e-01 1.26318681e+00 4.50884640e-01 2.53050417e-01
-5.30469418e-01 -4.99114066e-01 4.70679849e-01 1.07739186e+00
4.00016695e-01 -7.05063343e-01 -8.08338344e-01 7.05236793e-01
-1.66859496e+00 -1.17800415e+00 -2.96448201e-01 1.98971164e+00
7.39176810e-01 2.03454167e-01 5.37660569e-02 4.13848490e-01
7.86348701e-01 3.14310730e-01 -2.87555367e-01 1.01450300e-02
-1.15709230e-01 2.75660962e-01 1.49965435e-01 1.39627948e-01
-1.29573178e+00 5.84421635e-01 6.38268328e+00 1.23388088e+00
-1.04358959e+00 3.61222923e-01 -2.70686150e-02 -2.56151617e-01
1.09089643e-01 -7.93064088e-02 -1.24922562e+00 7.21906066e-01
1.04849231e+00 -2.20194552e-02 -8.85433238e-03 7.18650639e-01
4.77451056e-01 -1.36038363e-01 -9.80058432e-01 9.81601119e-01
-1.32099643e-01 -1.16626096e+00 -3.87317121e-01 -4.63786274e-01
3.88929397e-01 1.67110041e-01 6.48814142e-02 4.39672172e-01
-3.95946890e-01 -4.88843828e-01 1.19672704e+00 4.81309742e-01
5.98581493e-01 -5.62287688e-01 3.21741492e-01 4.60819781e-01
-1.75577962e+00 -3.01546276e-01 -1.69531092e-01 -1.56791899e-02
5.70626140e-01 8.57284367e-01 -1.05411220e+00 7.01115429e-01
1.19574046e+00 4.67809767e-01 -3.13977748e-01 1.73703337e+00
-1.27947137e-01 1.28007567e+00 -6.79683506e-01 4.54564579e-02
-5.39342016e-02 6.98107779e-01 1.03746140e+00 1.47818327e+00
8.56788874e-01 -1.70401275e-01 1.90698698e-01 3.31909716e-01
1.81961477e-01 2.56648120e-02 -2.60970950e-01 2.46863708e-01
1.06081259e+00 8.35538805e-01 -6.41907215e-01 2.21166909e-01
-2.32868999e-01 4.73798662e-01 -2.82410890e-01 9.54544321e-02
-1.07801104e+00 -6.96367860e-01 5.91589391e-01 8.85074437e-02
5.66831410e-01 -1.14889413e-01 1.47733686e-03 -5.48538327e-01
1.73664987e-01 -6.74098194e-01 6.22677922e-01 -9.68369842e-01
-1.12296152e+00 7.84663081e-01 8.52772146e-02 -1.95903385e+00
-8.13729391e-02 -9.11937132e-02 -8.82035077e-01 5.13922751e-01
-1.39674592e+00 -8.09539258e-01 -5.92331029e-02 2.67574996e-01
8.55338037e-01 -2.11542264e-01 8.29116225e-01 1.01279318e+00
-5.50666869e-01 8.22513938e-01 2.11063288e-02 -1.12928540e-01
9.75927889e-01 -1.07472289e+00 3.72714460e-01 9.72777843e-01
7.26698756e-01 9.15104449e-02 9.02445078e-01 -5.84332466e-01
-9.79571223e-01 -1.32598722e+00 1.17476130e+00 -4.09742206e-01
4.80135322e-01 -4.51063544e-01 -7.00068831e-01 3.95096213e-01
-2.95634121e-01 4.43926811e-01 6.77637875e-01 1.77033655e-02
-2.41401762e-01 -5.47125041e-01 -6.28323138e-01 5.01958653e-02
9.38573599e-01 -7.32358575e-01 -4.28382188e-01 4.00069058e-01
8.12039971e-01 -6.06179416e-01 -5.43087602e-01 5.32072544e-01
4.05848145e-01 -8.90808582e-01 1.15869403e+00 -1.52027696e-01
-1.18976653e-01 -8.43869627e-01 -5.61761677e-01 -1.11205697e+00
-1.51119888e-01 -7.17075527e-01 -2.74678200e-01 1.57431936e+00
3.79742324e-01 -2.39653438e-01 1.31053463e-01 -2.67665982e-01
-7.41688907e-01 -4.99071509e-01 -1.33565485e+00 -1.09206867e+00
-6.92141294e-01 -1.08945453e+00 5.61955571e-01 4.52942044e-01
-3.11980009e-01 9.77338105e-02 -8.68047118e-01 9.89196420e-01
3.80648881e-01 8.95832106e-03 3.47369909e-01 -1.10108221e+00
-5.97628534e-01 -2.68902779e-02 -4.30569530e-01 -1.09594297e+00
-3.47702593e-01 -8.27971697e-01 6.49283171e-01 -1.04113162e+00
-4.79002446e-01 -5.04331887e-01 -7.35416770e-01 3.86169970e-01
-1.94625169e-01 3.46271157e-01 1.10120393e-01 3.80743027e-01
-9.14153039e-01 5.51005125e-01 6.79863751e-01 1.87361941e-01
-3.35066229e-01 7.98477650e-01 -3.20272744e-01 6.68226779e-01
5.83391666e-01 -1.06350970e+00 -4.33089226e-01 -4.23679799e-01
3.41275215e-01 3.20806146e-01 5.22404909e-01 -1.66605961e+00
6.21237993e-01 8.29606503e-02 8.49662200e-02 -1.10703087e+00
4.97986764e-01 -6.51156366e-01 3.61627996e-01 1.30993888e-01
-3.97465050e-01 -3.65351811e-02 3.87567401e-01 9.04451549e-01
-4.60565120e-01 1.55258402e-02 5.79146564e-01 4.62266326e-01
-9.25139904e-01 9.95572880e-02 -5.36722243e-01 1.59023613e-01
9.28430796e-01 2.05693364e-01 -2.78749824e-01 -2.57731408e-01
-7.24376142e-01 8.18434432e-02 -7.85491467e-01 5.13947368e-01
5.57956636e-01 -1.49610150e+00 -6.02664828e-01 2.13017181e-01
2.80183822e-01 -1.68645456e-01 5.80883563e-01 7.96126127e-01
-2.73542136e-01 5.76085329e-01 2.42185313e-02 -7.65025318e-01
-1.15944242e+00 -4.68734046e-03 4.63023603e-01 2.42658183e-02
-8.24615419e-01 1.10984337e+00 2.73743480e-01 -3.38592291e-01
6.57588184e-01 -3.26285630e-01 -2.23142713e-01 3.25433202e-02
5.91921866e-01 6.04754448e-01 3.83529365e-01 -4.16446388e-01
-7.46721208e-01 4.04677123e-01 2.54237175e-01 -1.99910417e-01
1.13042271e+00 -4.35726166e-01 4.28784817e-01 6.56481922e-01
9.21078324e-01 4.42264318e-01 -1.15607607e+00 -2.88501829e-01
8.93860981e-02 -4.51233059e-01 1.74340248e-01 -9.55144882e-01
-8.10252428e-01 9.65025008e-01 9.07275856e-01 4.48497117e-01
1.31398046e+00 2.12140456e-02 9.25782621e-01 2.87300080e-01
3.15532863e-01 -1.01188338e+00 1.43046319e-01 6.42114997e-01
7.57780433e-01 -7.81710804e-01 -2.97635078e-01 -1.76878467e-01
-5.29673755e-01 9.83867347e-01 6.63249493e-01 2.77445644e-01
1.02729332e+00 4.58599091e-01 3.45643014e-01 -8.52140412e-02
-7.57324696e-01 -2.20656514e-01 3.24394375e-01 4.22111422e-01
1.28568500e-01 -3.45730316e-03 5.47308959e-02 1.11133790e+00
-3.42566758e-01 -3.97419125e-01 -6.36487529e-02 8.53085399e-01
-7.13946819e-01 -7.53750861e-01 -5.67009330e-01 2.12820351e-01
-5.00554323e-01 -1.80545330e-01 6.65573627e-02 5.48507452e-01
2.40224451e-01 1.22744322e+00 3.85990664e-02 -6.12450540e-01
5.96375823e-01 3.82052846e-02 -3.21555175e-02 -5.86047769e-01
-6.24118209e-01 4.57538843e-01 7.46212304e-02 -4.88857716e-01
5.06087728e-02 -7.12562263e-01 -1.43634212e+00 3.49473566e-01
-5.31275630e-01 4.29987133e-01 5.09388089e-01 9.70964074e-01
5.18824220e-01 1.06937253e+00 1.10168481e+00 -6.86494529e-01
-6.48430169e-01 -9.30533469e-01 -7.43433058e-01 -1.62400737e-01
5.52399695e-01 -7.05218494e-01 -6.11587584e-01 3.54649834e-02] | [15.20653247833252, 5.28936243057251] |
55b8d54e-0de1-4b99-ab69-bb734e29a417 | do-prosody-transfer-models-transfer-prosody | 2303.04289 | null | https://arxiv.org/abs/2303.04289v1 | https://arxiv.org/pdf/2303.04289v1.pdf | Do Prosody Transfer Models Transfer Prosody? | Some recent models for Text-to-Speech synthesis aim to transfer the prosody of a reference utterance to the generated target synthetic speech. This is done by using a learned embedding of the reference utterance, which is used to condition speech generation. During training, the reference utterance is identical to the target utterance. Yet, during synthesis, these models are often used to transfer prosody from a reference that differs from the text or speaker being synthesized. To address this inconsistency, we propose to use a different, but prosodically-related, utterance during training too. We believe this should encourage the model to learn to transfer only those characteristics that the reference and target have in common. If prosody transfer methods do indeed transfer prosody they should be able to be trained in the way we propose. However, results show that a model trained under these conditions performs significantly worse than one trained using the target utterance as a reference. To explain this, we hypothesize that prosody transfer models do not learn a transferable representation of prosody, but rather an utterance-level representation which is highly dependent on both the reference speaker and reference text. | ['Simon King', 'Atli Thor Sigurgeirsson'] | 2023-03-07 | null | null | null | null | ['text-to-speech-synthesis', 'speech-synthesis'] | ['speech', 'speech'] | [ 4.94963318e-01 8.58024776e-01 -2.92600214e-01 -5.04449010e-01
-6.00604951e-01 -5.36210537e-01 7.81856954e-01 -9.67220739e-02
9.35278134e-04 8.38908911e-01 8.76211345e-01 2.53054444e-02
6.30924404e-01 -8.58543038e-01 -9.22128975e-01 -5.31733632e-01
4.94965196e-01 5.36558926e-01 1.91893920e-01 -6.72239065e-01
-1.33838296e-01 1.82513192e-01 -1.38557553e+00 7.03455985e-01
2.98933476e-01 4.92195874e-01 4.14299875e-01 7.09195435e-01
-2.40130067e-01 8.32265556e-01 -1.04990804e+00 -2.15881869e-01
1.15376942e-01 -1.16718745e+00 -1.00495744e+00 2.30558112e-01
1.57999769e-01 -2.68800080e-01 4.49552797e-02 9.45087254e-01
2.07295045e-01 2.01058567e-01 8.15721154e-01 -8.00556540e-01
-6.03826880e-01 1.30383921e+00 2.32403174e-01 -1.96689263e-01
4.24087942e-01 -8.35349858e-02 1.03743231e+00 -6.10975981e-01
9.52318192e-01 1.33126128e+00 2.55635530e-01 1.29457188e+00
-1.56709003e+00 -3.65887046e-01 3.79658006e-02 -3.55769664e-01
-9.27603304e-01 -8.00191104e-01 9.78370488e-01 -3.87328148e-01
6.60439134e-01 2.65767008e-01 3.73277783e-01 1.59103131e+00
4.34650108e-03 4.94664520e-01 8.19915831e-01 -7.46226072e-01
1.83488518e-01 7.81201363e-01 -4.52400595e-02 1.02462143e-01
-4.46249574e-01 3.40845436e-01 -4.40538257e-01 9.50650424e-02
6.00781798e-01 -7.83966422e-01 -5.94325960e-01 -2.47670397e-01
-1.13148320e+00 9.75791633e-01 2.00220510e-01 7.85416186e-01
-4.92191613e-01 9.84164700e-02 5.80244482e-01 6.23900414e-01
3.52084845e-01 5.79232216e-01 -4.53579932e-01 -6.53949901e-02
-8.63574505e-01 1.59135833e-01 1.13232553e+00 8.76378000e-01
6.01687431e-01 4.45729554e-01 -8.32402557e-02 1.06513023e+00
3.76308709e-01 1.79249823e-01 8.00834775e-01 -8.85373294e-01
3.20063382e-01 2.19260827e-01 8.87801722e-02 -5.27038455e-01
7.58113787e-02 -2.68368218e-02 -3.56429249e-01 2.04249263e-01
4.82022852e-01 -2.92023182e-01 -8.09736133e-01 2.16313624e+00
9.28501785e-02 -5.95455170e-02 7.79743016e-01 8.35001469e-01
6.99934363e-01 1.18203223e+00 -4.75876220e-03 -4.38025415e-01
9.83306289e-01 -9.90836024e-01 -8.57937276e-01 -1.26435265e-01
6.97251439e-01 -8.56943786e-01 1.25632501e+00 1.66359469e-01
-1.40848613e+00 -9.03251708e-01 -1.26132476e+00 9.26277712e-02
-2.50409156e-01 -1.26226097e-02 -2.12764502e-01 8.00956666e-01
-9.32109237e-01 5.75412631e-01 -4.27835047e-01 -3.11102450e-01
-4.32054311e-01 2.44503126e-01 -3.28107744e-01 5.18404186e-01
-1.40095663e+00 1.10215259e+00 7.67479599e-01 -3.91251087e-01
-7.30725706e-01 -7.33257234e-01 -1.07414722e+00 2.63517201e-01
4.10475675e-03 -5.36466658e-01 1.61486948e+00 -1.70709467e+00
-2.13670969e+00 8.00316215e-01 -1.13499194e-01 -6.00465477e-01
2.30075702e-01 2.23795772e-02 -4.01477635e-01 1.84174463e-01
-1.94826812e-01 8.87793124e-01 1.03101933e+00 -1.51345122e+00
-3.14675122e-01 1.39249250e-01 -5.19258939e-02 1.30281508e-01
-3.41311060e-02 2.22300738e-01 9.77955982e-02 -6.10980153e-01
-1.79178149e-01 -9.15548861e-01 1.96451053e-01 -4.38950539e-01
-3.57501268e-01 -2.14347571e-01 7.79200435e-01 -5.29201627e-01
9.26367044e-01 -2.25372624e+00 4.20046300e-01 -1.95730217e-02
-4.35790300e-01 2.62467891e-01 -2.85396278e-01 7.24862814e-01
-4.54928041e-01 2.25977331e-01 -3.04373145e-01 -5.31914771e-01
8.58257860e-02 3.19065690e-01 -8.74633253e-01 8.65442529e-02
4.31131750e-01 6.82220459e-01 -7.17895091e-01 -1.67071685e-01
6.76795691e-02 5.51527619e-01 -5.33427954e-01 5.04763126e-01
-5.92641652e-01 7.23119617e-01 -1.35211855e-01 -3.30283821e-01
2.32852459e-01 3.71026754e-01 2.87678540e-01 -9.65975672e-02
-8.41273069e-02 9.52417970e-01 -8.64951551e-01 1.53244829e+00
-6.97550535e-01 6.33922815e-01 2.19720885e-01 -1.02187967e+00
1.34057784e+00 1.02413607e+00 7.05304593e-02 -4.69767600e-01
4.03165445e-02 3.15528274e-01 4.90567505e-01 -2.89537370e-01
4.60199475e-01 -8.76656055e-01 -6.35382831e-02 5.52139044e-01
1.94348395e-01 -6.74815655e-01 -1.64756812e-02 -1.49488583e-01
6.43061876e-01 3.18526983e-01 2.85315543e-01 -2.11793169e-01
8.16143095e-01 -1.73776135e-01 4.80899096e-01 3.19374412e-01
1.37491301e-01 7.79779196e-01 4.18227702e-01 -1.47503570e-01
-1.24050331e+00 -1.06803119e+00 8.96867458e-03 1.09879541e+00
-3.19974065e-01 -3.85214984e-01 -1.00264692e+00 -5.84154189e-01
-5.04456401e-01 1.42722821e+00 -5.96394539e-01 -2.39358932e-01
-1.06529307e+00 6.96247742e-02 6.17505848e-01 2.96659619e-01
-9.39661190e-02 -1.54826415e+00 -5.29887736e-01 5.68857908e-01
-1.50845990e-01 -9.21685934e-01 -6.98712349e-01 2.98110813e-01
-6.44838333e-01 -4.05769765e-01 -7.45583296e-01 -8.81446242e-01
4.34370667e-01 -4.10020500e-01 1.10549486e+00 2.34775841e-02
5.50814211e-01 3.51708800e-01 -5.54307997e-01 -4.37975019e-01
-1.63049829e+00 5.50417081e-02 -5.80576397e-02 1.87266201e-01
-1.08431987e-02 -5.87409019e-01 1.09032013e-01 2.21539855e-01
-1.10436511e+00 1.34465441e-01 2.18577623e-01 9.88075197e-01
2.56664395e-01 -2.61253148e-01 1.01772618e+00 -8.64645600e-01
7.66769409e-01 -3.87667835e-01 -2.55480826e-01 6.30076677e-02
-4.83722053e-03 2.57238209e-01 1.02445459e+00 -6.77221537e-01
-1.14159381e+00 1.94819923e-02 -5.66639781e-01 -4.16047961e-01
-5.67034841e-01 4.93935168e-01 -3.75228733e-01 7.11223781e-01
7.26520658e-01 3.13310385e-01 1.65628582e-01 -4.38441932e-01
4.16937202e-01 7.57065713e-01 3.47518146e-01 -6.20786369e-01
6.91583037e-01 -1.87814787e-01 -3.86595666e-01 -1.10731065e+00
-4.81615275e-01 8.82133958e-04 -6.71571553e-01 2.86658034e-02
1.11887884e+00 -6.08911812e-01 -1.48913488e-01 1.03716936e-03
-1.68807065e+00 -6.19803369e-01 -7.48195946e-01 6.09840930e-01
-1.06025994e+00 -4.09272825e-03 -5.22911191e-01 -9.10478055e-01
-4.63931523e-02 -1.24768555e+00 9.07641590e-01 -5.83208315e-02
-7.16174662e-01 -1.24437559e+00 3.21728706e-01 2.42230296e-02
4.29206580e-01 1.42545864e-01 1.25147593e+00 -1.07990372e+00
-1.74414977e-01 2.20625460e-01 2.91776747e-01 8.18430543e-01
7.56568432e-01 1.53770342e-01 -1.09873867e+00 -1.11709267e-01
4.44100618e-01 -4.91903007e-01 4.91971850e-01 1.41484618e-01
3.37350219e-01 -5.57299674e-01 -3.88002023e-02 1.94956690e-01
9.63831544e-01 4.40051883e-01 6.21330857e-01 -2.11022407e-01
2.38704607e-01 1.26763237e+00 2.96542287e-01 -2.68923402e-01
8.63754302e-02 9.31567311e-01 1.32748947e-01 1.89852923e-01
-3.67858797e-01 -5.19414663e-01 9.67224777e-01 1.28733993e+00
3.01361501e-01 -2.27914229e-01 -6.90120459e-01 6.37569726e-01
-1.47560620e+00 -1.21558189e+00 1.06385544e-01 2.19478226e+00
1.40481973e+00 3.14850897e-01 1.57987386e-01 1.08110026e-01
5.95391452e-01 2.63354003e-01 -1.41500026e-01 -9.42263246e-01
6.01819763e-03 3.13811481e-01 -2.88523257e-01 9.55790579e-01
-6.31903470e-01 1.01234555e+00 6.44693851e+00 4.49739903e-01
-1.72221386e+00 3.04378811e-02 2.54784942e-01 2.93860883e-01
-6.11410022e-01 -1.38514824e-02 -6.44507051e-01 4.94769573e-01
1.73122323e+00 -4.96017545e-01 4.25654173e-01 6.59545898e-01
3.99032563e-01 2.62469083e-01 -1.86330569e+00 4.05453682e-01
1.20499194e-01 -1.41929448e+00 2.91109174e-01 -6.50518611e-02
4.90837038e-01 -4.97798890e-01 -9.39462036e-02 6.10409915e-01
-2.55373772e-02 -1.20881569e+00 8.83031905e-01 3.73493701e-01
4.51506227e-01 -6.42200947e-01 4.97344643e-01 7.20028102e-01
-8.89475882e-01 3.76305073e-01 -2.84585595e-01 -6.80129603e-02
3.56312841e-01 -7.50211552e-02 -1.41166162e+00 4.87687260e-01
-4.15984169e-03 3.05843115e-01 -4.99490388e-02 2.71323830e-01
-3.36892664e-01 8.92457068e-01 -3.26298326e-02 1.58048242e-01
7.29780123e-02 -1.43077478e-01 7.35168397e-01 1.24998379e+00
3.26964259e-01 -9.55981985e-02 2.41881579e-01 1.25703657e+00
-5.36990613e-02 3.65065724e-01 -1.01863241e+00 -2.50866175e-01
3.10225934e-01 6.93777263e-01 -2.51384139e-01 -5.36858022e-01
-3.15299332e-01 8.69126022e-01 -3.93195562e-02 2.46796280e-01
-6.78613603e-01 -1.33472472e-01 6.27538145e-01 1.34069696e-01
2.96029210e-01 -3.09211072e-02 2.61836182e-02 -9.04086113e-01
-1.48096517e-01 -1.03536272e+00 -3.48998606e-01 -8.54703963e-01
-1.12046814e+00 9.54811692e-01 2.39005372e-01 -1.12395692e+00
-1.00800490e+00 -2.82336593e-01 -9.53460157e-01 1.40081668e+00
-1.29798222e+00 -1.07276917e+00 3.79207164e-01 3.81358176e-01
1.02165675e+00 -2.24620014e-01 1.26236856e+00 -1.41276777e-01
-2.43249033e-02 3.65499973e-01 -5.66188514e-01 2.38237977e-01
7.48167753e-01 -1.23380780e+00 3.01787913e-01 4.79705960e-01
2.68673807e-01 7.47531056e-01 1.21969497e+00 -4.79770482e-01
-7.65442491e-01 -9.59157884e-01 9.82515037e-01 -2.89330542e-01
5.90537310e-01 -3.47959042e-01 -1.39862359e+00 8.08908939e-01
8.42294335e-01 -4.01798159e-01 8.14960718e-01 -2.31395558e-01
-2.36270875e-01 9.49059948e-02 -8.92331123e-01 6.26924217e-01
3.96525145e-01 -6.95394337e-01 -1.39167964e+00 1.86359212e-02
1.29237938e+00 -2.60064125e-01 -8.22870195e-01 2.54465997e-01
2.01778993e-01 -9.04274583e-01 5.37084043e-01 -7.48836756e-01
5.52948952e-01 -2.01012194e-01 -4.18082774e-01 -1.72471821e+00
2.74014384e-01 -5.10739684e-01 1.15340494e-01 1.60811758e+00
8.24173450e-01 -5.47632277e-01 6.12167895e-01 4.54781860e-01
-5.19955397e-01 -2.24617645e-01 -8.70873630e-01 -9.55414236e-01
7.09575236e-01 -1.60342336e-01 6.07255757e-01 8.89771998e-01
2.15265498e-01 8.90272021e-01 -3.61363769e-01 1.20726891e-01
-5.59770204e-02 -2.19047014e-02 8.79921317e-01 -1.14179790e+00
-5.55867195e-01 -3.37599218e-01 1.43102352e-02 -1.09288335e+00
6.63566709e-01 -9.03095186e-01 3.78907323e-01 -1.19449091e+00
-4.97902632e-01 -2.50100344e-01 1.99124590e-02 2.50455081e-01
1.26201332e-01 -1.70535028e-01 5.11665463e-01 5.57967275e-02
4.55552399e-01 7.91835010e-01 1.27044976e+00 -1.28392711e-01
-6.03199780e-01 2.50779420e-01 -3.18214893e-01 7.12477684e-01
9.64036763e-01 -5.77164114e-01 -6.47879899e-01 -1.02334015e-01
-2.79225707e-01 5.74052215e-01 5.09403571e-02 -7.43646920e-01
-1.90349780e-02 -3.28508854e-01 -1.03579434e-02 -2.03563899e-01
8.96116257e-01 -8.25662851e-01 2.32657343e-01 3.49022567e-01
-8.25818837e-01 -2.69009978e-01 2.77403980e-01 6.12781830e-02
-6.41556978e-01 -8.78563821e-01 9.48834419e-01 -1.61022902e-01
-1.97472289e-01 -4.57557887e-01 -5.97298980e-01 -2.64608055e-01
8.03062022e-01 -3.23973924e-01 -1.31081983e-01 -6.93813503e-01
-1.05122244e+00 -3.47452700e-01 5.20266294e-01 5.54397941e-01
4.72012758e-01 -1.31606829e+00 -8.86455595e-01 3.19768429e-01
-7.73982406e-02 -2.35862195e-01 -1.79516748e-01 5.51329672e-01
-2.28950731e-03 3.97434920e-01 -3.30916047e-01 -4.25163418e-01
-1.14139688e+00 6.30657256e-01 4.58818793e-01 -6.38214052e-02
-5.41511834e-01 4.07135695e-01 4.16516036e-01 -6.09427094e-01
1.28595501e-01 -5.45105100e-01 -2.24113241e-01 1.92321599e-01
4.48118418e-01 -3.38719726e-01 -1.38269275e-01 -9.36664104e-01
1.22814447e-01 3.19735289e-01 3.22864279e-02 -6.87799633e-01
1.13726103e+00 -4.17959839e-02 1.52101174e-01 1.04381335e+00
1.14801931e+00 6.43793523e-01 -1.25489080e+00 -8.19117203e-02
1.41742721e-01 -1.36248488e-02 -3.89334470e-01 -6.59306228e-01
-6.46255791e-01 9.52360749e-01 -4.07155305e-02 6.91959679e-01
8.45468998e-01 5.73654398e-02 5.82955539e-01 1.40183091e-01
2.26953536e-01 -8.59508693e-01 9.23918262e-02 7.98630416e-01
1.47757232e+00 -8.33878696e-01 -6.30387187e-01 -4.33887720e-01
-8.72010231e-01 1.35234690e+00 4.65546399e-01 -3.15982789e-01
4.86366838e-01 1.35516763e-01 2.97477156e-01 3.73372614e-01
-1.05995381e+00 -9.19230655e-02 3.40565085e-01 7.89071560e-01
8.58853579e-01 -1.15776077e-01 -1.32327855e-01 2.99673975e-01
-7.54448533e-01 -2.31478363e-01 8.66762817e-01 5.35585642e-01
-5.65482438e-01 -1.68788135e+00 -4.82297331e-01 -1.81108564e-01
-3.11882883e-01 -6.75195071e-04 -7.96050191e-01 8.31812143e-01
7.67106377e-03 9.67409849e-01 1.43216804e-01 -1.92297503e-01
3.98247957e-01 7.43627131e-01 3.19022328e-01 -1.34484887e+00
-8.88021827e-01 1.95385844e-01 2.97212064e-01 -1.19550884e-01
-5.10155559e-01 -4.15642977e-01 -1.41510594e+00 1.47497147e-01
-7.97140822e-02 5.63623667e-01 7.08919585e-01 9.02765214e-01
-1.12317078e-01 6.85064912e-01 7.09989071e-01 -9.83120799e-01
-6.74550176e-01 -1.15895939e+00 -4.10589427e-01 4.81073231e-01
6.79501295e-01 -2.81509727e-01 -2.63433248e-01 4.71647739e-01] | [15.024724960327148, 6.72933292388916] |
c760c46f-9a09-4be6-a617-b507fd4449b9 | anomaly-detection-framework-using-rule | 1410.7709 | null | http://arxiv.org/abs/1410.7709v1 | http://arxiv.org/pdf/1410.7709v1.pdf | Anomaly Detection Framework Using Rule Extraction for Efficient Intrusion Detection | Huge datasets in cyber security, such as network traffic logs, can be
analyzed using machine learning and data mining methods. However, the amount of
collected data is increasing, which makes analysis more difficult. Many machine
learning methods have not been designed for big datasets, and consequently are
slow and difficult to understand. We address the issue of efficient network
traffic classification by creating an intrusion detection framework that
applies dimensionality reduction and conjunctive rule extraction. The system
can perform unsupervised anomaly detection and use this information to create
conjunctive rules that classify huge amounts of traffic in real time. We test
the implemented system with the widely used KDD Cup 99 dataset and real-world
network logs to confirm that the performance is satisfactory. This system is
transparent and does not work like a black box, making it intuitive for domain
experts, such as network administrators. | ['Antti Juvonen', 'Tuomo Sipola'] | 2014-10-28 | null | null | null | null | ['traffic-classification'] | ['miscellaneous'] | [-2.91923881e-01 -3.64613652e-01 -1.38908669e-01 -2.57351846e-01
3.08925271e-01 -4.90220308e-01 2.71045238e-01 4.53223020e-01
-3.97294134e-01 5.93568265e-01 -4.59208697e-01 -1.12882221e+00
-1.33291736e-01 -1.23578262e+00 1.28468061e-02 -3.17172170e-01
-2.64049262e-01 6.47242010e-01 5.78779459e-01 -1.38542667e-01
4.56015080e-01 8.79658639e-01 -1.61915147e+00 2.62178034e-01
7.99880147e-01 1.07177281e+00 -5.24885535e-01 5.24971366e-01
-5.90515018e-01 7.98848987e-01 -9.94785428e-01 -2.72267550e-01
5.34983993e-01 -1.46990493e-01 -6.70740068e-01 1.16461813e-01
-2.17611730e-01 -1.32976532e-01 -1.34836480e-01 9.35210347e-01
1.82139724e-01 -1.42834738e-01 1.49390668e-01 -1.70682645e+00
-1.91867441e-01 1.10173166e-01 -5.87324202e-01 7.20480561e-01
1.23436294e-01 1.59551471e-01 5.52286029e-01 -3.09660167e-01
4.50444609e-01 1.16308880e+00 3.82770211e-01 3.88244957e-01
-1.05992234e+00 -8.54720891e-01 8.80891178e-03 4.18358594e-01
-1.08096695e+00 -8.23330600e-03 7.77140796e-01 -3.70987743e-01
1.03970110e+00 4.25641775e-01 4.55901533e-01 8.54341269e-01
1.09062448e-01 2.96802253e-01 9.34938371e-01 -4.71267521e-01
6.49598658e-01 5.92271805e-01 8.03223789e-01 3.47640544e-01
8.89127433e-01 9.16786194e-02 1.92551285e-01 -4.50021625e-01
2.96012044e-01 5.95695138e-01 3.06133032e-01 -7.54789039e-02
-6.00762546e-01 9.46952641e-01 -7.93766417e-03 5.36404848e-01
-5.81402659e-01 -5.87465286e-01 6.74392760e-01 9.68546331e-01
3.14861536e-01 3.60197008e-01 -6.25544012e-01 -4.01550859e-01
-5.07608235e-01 5.95633220e-03 1.25146711e+00 4.53054398e-01
8.13981473e-01 2.39364117e-01 6.02994740e-01 5.60487568e-01
1.46690652e-01 2.92162120e-01 5.59614539e-01 -4.88727391e-01
3.06403786e-01 1.33512545e+00 -3.07837278e-01 -1.31136274e+00
-5.37919641e-01 -8.13717470e-02 -7.11442351e-01 5.86893559e-01
4.93425310e-01 -2.42752701e-01 -7.49307394e-01 1.04933643e+00
4.87402678e-01 4.56228033e-02 -1.31184503e-01 5.78443468e-01
2.82688022e-01 4.54443038e-01 1.69199646e-01 -3.98768276e-01
1.28948772e+00 -2.09362239e-01 -6.96492970e-01 -5.40365130e-02
9.62096512e-01 -6.04111254e-01 1.07130218e+00 6.34568393e-01
-3.31203610e-01 -2.74776131e-01 -8.67270708e-01 5.83827198e-01
-1.03567648e+00 -5.72392762e-01 9.19762671e-01 1.06079245e+00
-3.42196226e-01 4.03481215e-01 -7.05151379e-01 -6.68776810e-01
5.15794396e-01 3.62520337e-01 -4.22702551e-01 1.18959062e-01
-1.03890967e+00 8.65365446e-01 9.51443791e-01 -3.91726673e-01
-2.43662477e-01 -4.80866253e-01 -4.92511183e-01 2.01037247e-02
6.50810242e-01 -7.83473924e-02 7.34139681e-01 -6.61917686e-01
-1.04929733e+00 5.53496718e-01 2.44257227e-01 -5.82715988e-01
2.19862416e-01 5.00420257e-02 -1.20073605e+00 5.76486401e-02
-3.09866160e-01 -3.80316645e-01 6.70658767e-01 -9.73931551e-01
-8.10625672e-01 -7.19565332e-01 3.08795720e-02 -5.11976957e-01
-7.22326815e-01 3.58241946e-01 7.51341879e-02 -3.91470820e-01
-1.64172947e-01 -7.72959352e-01 -3.38049114e-01 -4.64906812e-01
-4.03352380e-01 -2.05282137e-01 1.77342463e+00 -2.98534304e-01
1.78026760e+00 -2.07359624e+00 -7.80093849e-01 9.38248873e-01
2.20740601e-01 8.89900506e-01 5.41602448e-02 2.89281696e-01
-3.65283370e-01 5.17529190e-01 -1.79178119e-02 4.02152270e-01
-3.55902582e-01 5.45465529e-01 -2.49472812e-01 1.10077560e-02
1.98979482e-01 3.30319613e-01 -5.94506443e-01 -4.10865784e-01
6.02194011e-01 -2.11250689e-02 -5.59557796e-01 2.00592637e-01
-1.51978269e-01 1.75191045e-01 -6.55050635e-01 8.36223364e-01
7.12197542e-01 -3.02341819e-01 1.69110775e-01 3.90869975e-02
-1.12384737e-01 2.66852919e-02 -1.36741471e+00 6.85081661e-01
-2.33758122e-01 5.06350338e-01 -1.12602957e-01 -1.36350679e+00
1.08374286e+00 2.33211949e-01 6.17596447e-01 -9.64635253e-01
2.81119943e-01 1.56719729e-01 3.14739466e-01 -6.37404919e-01
-2.29424685e-02 4.61470596e-02 1.81792583e-03 1.03423595e+00
-2.58516729e-01 3.04181844e-01 6.07168198e-01 2.58286536e-01
1.55885589e+00 -7.64473319e-01 4.10766184e-01 1.20379686e-01
7.29754567e-01 4.29362983e-01 6.76911891e-01 4.08136129e-01
-1.67252392e-01 -1.95535019e-01 7.70702600e-01 -1.06016195e+00
-1.04386234e+00 -8.64995480e-01 -4.55063506e-04 7.58018315e-01
-1.82045355e-01 -6.68557286e-01 -4.63173747e-01 -1.11236048e+00
7.51592666e-02 5.34189641e-01 -1.93235785e-01 -2.33835697e-01
-5.73676884e-01 -1.01433063e+00 3.56832057e-01 3.75701264e-02
5.22284150e-01 -1.09400117e+00 -5.72696745e-01 2.87725687e-01
2.57419020e-01 -1.36693513e+00 4.14558232e-01 8.38023517e-03
-1.07076013e+00 -1.54762650e+00 6.32152975e-01 -2.68238574e-01
7.35939562e-01 2.95911640e-01 1.05071962e+00 5.13958573e-01
-7.73828924e-01 -4.25226707e-03 -6.13422513e-01 -9.02904093e-01
-5.53994358e-01 5.67669347e-02 3.60661566e-01 2.52722222e-02
1.12679112e+00 -6.93386674e-01 -1.77523047e-01 5.60198128e-01
-1.13365221e+00 -5.84935129e-01 6.19497359e-01 4.67855692e-01
1.85662016e-01 9.68184710e-01 9.00134563e-01 -1.21748066e+00
1.04291964e+00 -6.68033898e-01 -8.06505859e-01 2.54336558e-02
-1.12223947e+00 -1.23039134e-01 1.00791359e+00 -4.19176728e-01
-7.44266570e-01 -3.22875530e-01 -2.13213284e-02 -2.52196163e-01
-8.02219808e-01 3.61220270e-01 -1.81912050e-01 -6.90255761e-02
8.98043990e-01 -2.16096625e-01 3.77637416e-01 -6.87342286e-01
-1.48211837e-01 1.12812185e+00 -4.00213860e-02 -1.99971229e-01
1.16076386e+00 3.94018143e-01 5.93219995e-02 -1.12253916e+00
-3.99242908e-01 -4.51441556e-01 -6.13964200e-01 -2.57767159e-02
3.75261843e-01 -7.96496347e-02 -1.04220176e+00 3.41148227e-01
-7.45866060e-01 1.67894915e-01 -3.23937416e-01 4.84662056e-01
5.18873297e-02 5.13813317e-01 -5.57862759e-01 -7.85656691e-01
-3.40745807e-01 -6.59754992e-01 1.28556743e-01 1.87642574e-01
-3.07027072e-01 -9.72556055e-01 4.95799892e-02 3.71935189e-01
6.93122089e-01 3.13436478e-01 1.24193144e+00 -1.36011899e+00
-4.26560819e-01 -9.03976560e-01 -3.67522359e-01 4.84001935e-01
3.66288126e-01 3.81481826e-01 -7.33313560e-01 -1.63078278e-01
-4.85264435e-02 9.36500635e-03 2.40900069e-01 -1.71746880e-01
1.58140397e+00 -4.31924015e-01 -2.19219923e-01 8.40707496e-02
1.24558008e+00 5.86324990e-01 8.74432087e-01 6.21037424e-01
5.34667432e-01 6.44986808e-01 5.74659586e-01 5.39092839e-01
2.71111336e-02 3.77471685e-01 4.22401011e-01 -7.88168535e-02
5.16362190e-01 1.72397867e-01 6.90708309e-02 5.35212517e-01
-1.16246566e-01 1.00731514e-01 -1.19358051e+00 5.35783768e-02
-1.79635763e+00 -1.34139884e+00 -2.43273169e-01 2.05859399e+00
4.24337536e-01 6.87451124e-01 5.38185000e-01 8.49880636e-01
5.54390490e-01 -2.93828607e-01 -2.43547738e-01 -8.55156481e-01
3.12981427e-01 3.67539376e-01 2.33666778e-01 8.69866684e-02
-9.52240169e-01 7.89788425e-01 6.12779093e+00 5.94806671e-01
-1.35031569e+00 -1.11129716e-01 3.24852079e-01 7.42644146e-02
1.49758413e-01 1.46929458e-01 -4.62468714e-01 6.48394823e-01
1.46048391e+00 -2.72796720e-01 1.56156152e-01 1.05951869e+00
5.30293643e-01 1.96148362e-02 -7.96401143e-01 8.56473744e-01
-4.07605886e-01 -1.19204295e+00 1.88192740e-01 5.55279493e-01
7.98285678e-02 -2.17411086e-01 -3.34851861e-01 4.58275169e-01
2.94579566e-01 -7.41633952e-01 -3.21466386e-01 2.03189984e-01
-1.39822802e-02 -9.23037529e-01 8.50533247e-01 5.94932377e-01
-7.57150233e-01 -6.53767943e-01 -2.61449337e-01 -4.15082723e-01
2.70751696e-02 9.90547001e-01 -1.02621973e+00 3.07152241e-01
7.83751070e-01 2.99971193e-01 -7.14976549e-01 1.14778137e+00
1.28730804e-01 7.99323022e-01 -5.02494752e-01 8.77340976e-03
7.61253163e-02 -2.14686021e-01 4.56018299e-01 7.50816226e-01
3.29721086e-02 -5.77412099e-02 2.36173794e-01 4.06024486e-01
3.07768255e-01 2.82443762e-01 -8.63667667e-01 -3.85237902e-01
5.49451470e-01 1.32981980e+00 -9.19371426e-01 -2.85327554e-01
-6.89673185e-01 2.75412023e-01 -4.36321162e-02 6.26482442e-02
-5.55044830e-01 -6.47860050e-01 8.54958236e-01 5.92932820e-01
-1.43905461e-01 -2.73777872e-01 -3.74515802e-01 -1.14076853e+00
-2.62649520e-03 -1.09784722e+00 8.82855594e-01 -8.43582973e-02
-1.45084155e+00 6.83322430e-01 -1.09752141e-01 -1.35779381e+00
-3.21519285e-01 -8.56004357e-01 -9.82761443e-01 2.73544639e-01
-1.06344092e+00 -5.76763034e-01 -3.91273499e-01 9.66037333e-01
2.12251142e-01 -5.90358734e-01 9.42960083e-01 7.75956094e-01
-1.11390734e+00 3.98692936e-01 -2.11848527e-01 4.56402183e-01
3.23850483e-01 -9.59228337e-01 2.71995217e-01 9.85342205e-01
1.11866277e-02 6.27396822e-01 6.33067489e-01 -6.80637002e-01
-1.19189537e+00 -9.11368310e-01 4.64050621e-01 -4.90082830e-01
8.33577156e-01 -3.62096906e-01 -1.19574714e+00 2.98178166e-01
-3.00731391e-01 2.87245750e-01 1.14176178e+00 2.92941093e-01
-4.71038640e-01 -4.92201209e-01 -1.53652549e+00 2.99284637e-01
5.56395352e-01 -2.65933335e-01 -5.32834828e-01 3.55416179e-01
3.18699151e-01 4.59936857e-01 -8.25182974e-01 2.87509829e-01
1.76400885e-01 -9.20618236e-01 8.84251118e-01 -1.25909567e+00
-3.71053666e-01 -4.23273236e-01 9.10296589e-02 -9.76495564e-01
-1.32991016e-01 -3.17136735e-01 -4.69970524e-01 1.09011781e+00
5.55537045e-01 -1.10242641e+00 9.66355205e-01 8.91096890e-01
3.49620461e-01 -4.99128044e-01 -7.06282735e-01 -1.07270586e+00
-5.23123443e-01 -6.34089291e-01 7.75838375e-01 1.21898031e+00
2.97992647e-01 5.10595024e-01 -1.71168000e-01 8.52354243e-02
7.76231527e-01 7.81552047e-02 9.55288887e-01 -2.01515794e+00
5.15129007e-02 -4.19029504e-01 -1.12062979e+00 2.01799229e-01
-3.88044827e-02 -4.31178808e-01 -9.68359411e-01 -1.08511722e+00
-2.97499210e-01 -7.72990406e-01 -2.64326662e-01 7.41206348e-01
2.88169831e-01 1.33104548e-01 -1.16881453e-01 2.67486662e-01
-4.44467098e-01 -1.21349037e-01 6.60216272e-01 5.34758344e-02
-4.17186052e-01 2.49656156e-01 -6.88792467e-01 8.60325575e-01
1.14131582e+00 -5.02347469e-01 -5.56627214e-01 2.28678644e-01
7.17927888e-02 -4.36322421e-01 8.79857540e-02 -1.15254045e+00
2.17314541e-01 -4.21813726e-01 4.10823762e-01 -4.61291254e-01
-1.47439331e-01 -1.42738008e+00 2.92801578e-02 6.25370979e-01
3.22838753e-01 3.15551192e-01 1.16083145e-01 4.44080263e-01
-3.93301636e-01 -2.43749134e-02 8.00373316e-01 -2.00761668e-03
-9.25895691e-01 4.95562643e-01 -4.34105694e-01 3.08691040e-02
1.74549258e+00 -4.98778850e-01 -6.16698563e-01 -1.66233569e-01
-6.37030840e-01 2.02919856e-01 4.41702604e-01 6.82863593e-01
6.07690036e-01 -1.18068981e+00 -2.23387748e-01 6.26163125e-01
2.44314775e-01 -3.12036127e-01 -2.74157804e-02 6.09684289e-01
-6.36179566e-01 2.86121011e-01 -5.82884729e-01 -3.41299057e-01
-1.37238157e+00 1.07913435e+00 4.28113043e-02 -3.70934546e-01
-6.73870444e-01 -6.53561726e-02 -5.96998036e-01 -3.95692259e-01
2.15541199e-01 1.53806388e-01 -2.80126333e-01 1.89022884e-01
1.01834095e+00 6.44378662e-01 3.52413356e-01 -1.67495593e-01
-6.20177150e-01 1.30728781e-01 -4.09713089e-01 3.98423165e-01
1.43161154e+00 1.12742916e-01 -2.54231215e-01 1.36507647e-02
1.06765032e+00 -2.14782044e-01 -1.65225089e-01 -1.87925801e-01
5.20234644e-01 -9.49278235e-01 -2.12678149e-01 -5.23721516e-01
-1.19611740e+00 8.87231648e-01 6.09321892e-01 1.12249458e+00
1.21538627e+00 -3.09029847e-01 7.60299623e-01 7.28433430e-01
3.62911344e-01 -1.41643989e+00 1.12033553e-01 6.68262780e-01
1.11853749e-01 -1.30886495e+00 -1.11906990e-01 -3.12330961e-01
-4.87975061e-01 1.20882845e+00 9.10334826e-01 5.43957297e-03
1.19809353e+00 4.82454509e-01 1.79019302e-01 -4.45781767e-01
-9.22638655e-01 -1.52959198e-01 -2.99183041e-01 7.08926320e-01
1.25802442e-01 -1.45669673e-02 -4.48801339e-01 3.37808639e-01
-9.30024311e-02 2.60496419e-02 6.02804780e-01 1.27767432e+00
-7.27241158e-01 -1.42922461e+00 -4.73445326e-01 9.55082238e-01
-6.83634818e-01 3.63393635e-01 -5.03575623e-01 8.20603609e-01
-1.03024743e-01 1.15161896e+00 2.38376573e-01 -7.15371013e-01
5.45578122e-01 1.99752808e-01 -3.29014927e-01 -5.63134968e-01
-2.12810129e-01 -6.13743305e-01 2.05046639e-01 -6.39213681e-01
8.22517052e-02 -9.65116248e-02 -1.22272265e+00 -1.00237298e+00
-3.14963728e-01 5.01095712e-01 7.59309292e-01 9.64499772e-01
6.24474287e-01 2.91706324e-01 1.04882562e+00 -8.27277005e-02
-3.91536474e-01 -8.09640825e-01 -5.40156603e-01 7.20956564e-01
2.47730408e-02 -6.49502397e-01 -4.45939809e-01 -3.72433782e-01] | [5.243856906890869, 7.181462287902832] |
0cb422cf-b784-471b-a34e-c1f0683d14f0 | multimedia-generative-script-learning-for | 2208.12306 | null | https://arxiv.org/abs/2208.12306v3 | https://arxiv.org/pdf/2208.12306v3.pdf | Multimedia Generative Script Learning for Task Planning | Goal-oriented generative script learning aims to generate subsequent steps to reach a particular goal, which is an essential task to assist robots or humans in performing stereotypical activities. An important aspect of this process is the ability to capture historical states visually, which provides detailed information that is not covered by text and will guide subsequent steps. Therefore, we propose a new task, Multimedia Generative Script Learning, to generate subsequent steps by tracking historical states in both text and vision modalities, as well as presenting the first benchmark containing 5,652 tasks and 79,089 multimedia steps. This task is challenging in three aspects: the multimedia challenge of capturing the visual states in images, the induction challenge of performing unseen tasks, and the diversity challenge of covering different information in individual steps. We propose to encode visual state changes through a selective multimedia encoder to address the multimedia challenge, transfer knowledge from previously observed tasks using a retrieval-augmented decoder to overcome the induction challenge, and further present distinct information at each step by optimizing a diversity-oriented contrastive learning objective. We define metrics to evaluate both generation and inductive quality. Experiment results demonstrate that our approach significantly outperforms strong baselines. | ['Heng Ji', 'Girish Chowdhary', 'Julia Hockenmaier', 'Lifu Huang', 'Hou Pong Chan', 'Manling Li', 'Qingyun Wang'] | 2022-08-25 | null | null | null | null | ['multimodal-generation', 'multimedia-generative-script-learning'] | ['natural-language-processing', 'natural-language-processing'] | [ 7.14210808e-01 -2.13973504e-02 2.03799475e-02 -2.95264393e-01
-1.11070919e+00 -5.59585392e-01 1.24894071e+00 -2.92537481e-01
-4.38567042e-01 8.16077650e-01 4.62439358e-01 8.00923631e-02
4.11040097e-01 -5.37556231e-01 -1.17588544e+00 -6.93016648e-01
4.62195985e-02 5.83018303e-01 2.34612510e-01 -2.33074516e-01
3.47044528e-01 2.43419334e-01 -1.83960378e+00 8.19141448e-01
7.83465266e-01 8.15985322e-01 6.56049490e-01 7.85399854e-01
8.43873993e-02 1.38252473e+00 -7.54829645e-01 -2.66361326e-01
-1.46892279e-01 -1.01981652e+00 -8.55023682e-01 3.03930402e-01
4.85089540e-01 -4.25798476e-01 -5.20877421e-01 1.02160347e+00
5.36629915e-01 4.80387390e-01 8.65603983e-01 -1.38550663e+00
-7.73910522e-01 6.80970311e-01 -1.46087110e-01 7.18523562e-02
8.40318799e-01 6.22494161e-01 7.02451468e-01 -8.01741779e-01
1.05245006e+00 1.42192781e+00 1.78369716e-01 1.04209626e+00
-1.03071630e+00 -4.37182784e-01 3.67648602e-01 5.63060284e-01
-1.16792083e+00 -7.01108813e-01 7.16510236e-01 -4.84671295e-01
8.78347397e-01 2.08635088e-02 6.37902737e-01 1.84327698e+00
2.69300193e-01 1.11531734e+00 1.04824948e+00 -1.26634046e-01
3.36368859e-01 -6.55214936e-02 -4.64641422e-01 9.20531452e-01
-9.62670445e-02 3.19078565e-01 -9.18186784e-01 3.40443343e-01
5.93471348e-01 -1.38596833e-01 -2.83422679e-01 -2.62693435e-01
-1.66464269e+00 5.38503528e-01 3.34824026e-01 -6.76415414e-02
-4.28809494e-01 3.86964619e-01 2.46624410e-01 1.54860824e-01
-4.79553506e-04 4.45761830e-01 3.48036550e-02 -1.38773337e-01
-7.10293293e-01 2.54477561e-01 6.01761401e-01 1.30543888e+00
9.31153297e-01 3.36936951e-01 -9.16652203e-01 4.72622067e-01
2.38256857e-01 7.92460442e-01 5.66349506e-01 -1.22099197e+00
5.73397100e-01 4.46636170e-01 4.10536602e-02 -3.49564850e-01
-2.54821509e-01 1.96584594e-02 -8.19247127e-01 1.23637579e-01
8.05056617e-02 -1.46151125e-01 -1.37754679e+00 2.01503491e+00
1.56138003e-01 2.81381577e-01 2.25012422e-01 7.54890263e-01
1.09106386e+00 1.16945684e+00 3.11125278e-01 -2.85402715e-01
1.25533724e+00 -1.15982771e+00 -7.63509691e-01 -6.50816262e-01
2.47401789e-01 -5.05900264e-01 1.40742946e+00 2.38251552e-01
-1.29344273e+00 -8.58705282e-01 -1.12160265e+00 -8.17217082e-02
-3.44632387e-01 7.12882727e-02 3.00578296e-01 -2.70376001e-02
-1.18896985e+00 4.26401347e-01 -9.31999385e-01 -2.91089982e-01
3.08140039e-01 -1.26531556e-01 -5.87275922e-02 -2.69921362e-01
-1.39422214e+00 8.21756005e-01 7.61616886e-01 -1.24490507e-01
-1.72045064e+00 -5.88509977e-01 -1.37165654e+00 -8.47829506e-02
4.23871040e-01 -1.05495858e+00 1.29667437e+00 -7.94781506e-01
-1.40409374e+00 7.06520259e-01 -2.61485428e-01 -4.16880280e-01
5.78287184e-01 -1.45303868e-02 -1.93746418e-01 3.94522518e-01
3.30088258e-01 1.10399342e+00 1.00073814e+00 -1.63421965e+00
-7.11957097e-01 -2.69261897e-01 1.00992262e-01 6.87709987e-01
-4.05135006e-02 -2.75842130e-01 -9.34823871e-01 -5.13002455e-01
-2.27846667e-01 -1.20845091e+00 -2.59254932e-01 -4.71039891e-01
-5.12722254e-01 -1.38550907e-01 6.02600992e-01 -6.61148071e-01
9.59478319e-01 -1.87009645e+00 5.04868686e-01 -3.26649368e-01
-1.82798207e-01 -6.25350475e-02 -3.88594985e-01 4.45200443e-01
2.95218349e-01 -2.37124085e-01 -2.84116864e-01 -6.23910844e-01
1.25120610e-01 2.22992271e-01 -4.38619703e-01 -1.26758948e-01
3.20470899e-01 1.20388067e+00 -1.19432032e+00 -7.35879540e-01
2.15671480e-01 5.24263382e-01 -2.28131413e-01 5.67942381e-01
-7.10323513e-01 7.00100064e-01 -3.29244554e-01 5.84766507e-01
8.93916786e-02 -4.88386810e-01 9.20827314e-02 -2.65370935e-01
-3.86946909e-02 3.53941113e-01 -9.04185116e-01 2.27894187e+00
-4.42662179e-01 5.31474054e-01 -5.59824646e-01 -7.02822804e-01
5.18868446e-01 1.87509909e-01 2.50656486e-01 -1.02582288e+00
-1.32187515e-01 -3.69621277e-01 -5.76486886e-01 -6.14427090e-01
6.98895037e-01 1.46091431e-01 -6.39077544e-01 4.34035093e-01
3.07803959e-01 -4.34780359e-01 6.30147457e-01 6.66728020e-01
1.14570200e+00 6.60162032e-01 2.39251718e-01 3.19390595e-01
3.34131688e-01 1.91042736e-01 3.69124383e-01 1.05525362e+00
-1.69165835e-01 8.36261868e-01 3.62366259e-01 -2.07360938e-01
-8.70250225e-01 -1.27211094e+00 5.48677444e-01 1.28519917e+00
3.84304345e-01 -3.10929954e-01 -6.26513422e-01 -8.06911409e-01
-4.96170193e-01 1.10853612e+00 -9.23996210e-01 -7.95753956e-01
-6.32847011e-01 -5.20698845e-01 3.52976292e-01 5.55672228e-01
7.83503294e-01 -1.69956076e+00 -6.62605464e-01 4.18668985e-02
-8.44210386e-01 -1.38601315e+00 -6.55170560e-01 2.71697640e-02
-5.86839974e-01 -8.72838199e-01 -7.75763512e-01 -9.23328102e-01
5.05825698e-01 1.49848238e-01 1.49743891e+00 -4.23599362e-01
-2.21819997e-01 8.04822981e-01 -3.61107945e-01 -2.91937768e-01
-9.55035746e-01 6.46603778e-02 -2.50609934e-01 7.22054765e-03
-1.09503917e-01 -2.40986079e-01 -5.12830317e-01 1.32230863e-01
-9.49745238e-01 5.84583759e-01 9.46757674e-01 6.48041666e-01
8.16270590e-01 -3.03666145e-01 3.46101344e-01 -7.46882915e-01
4.65793341e-01 -3.81907374e-01 -3.81518692e-01 4.97983098e-01
-4.39563185e-01 4.08345073e-01 5.31845033e-01 -4.18751568e-01
-1.59799373e+00 4.48550522e-01 1.77852958e-02 -4.90621328e-01
-3.53823692e-01 3.84295166e-01 -3.39663208e-01 4.49996650e-01
7.34520078e-01 9.78470564e-01 -2.76250690e-01 8.73103440e-02
5.37090838e-01 2.44015887e-01 9.77868497e-01 -6.13773584e-01
6.29934192e-01 4.34058607e-01 2.06533656e-03 -6.33287907e-01
-1.08268654e+00 -2.78909236e-01 -3.44350845e-01 -6.27514720e-01
1.06678832e+00 -1.12442124e+00 -3.83967936e-01 5.01585007e-01
-1.29482949e+00 -7.50886202e-01 -4.03048068e-01 1.16932310e-01
-1.16445637e+00 3.47593158e-01 -4.96514201e-01 -6.45436168e-01
-3.76370549e-01 -1.24456215e+00 1.50914598e+00 2.26198941e-01
-1.56960990e-02 -7.98804998e-01 2.45651081e-01 3.57747883e-01
2.69012377e-02 4.39548582e-01 6.38674200e-01 -2.95870453e-01
-8.90174568e-01 1.13275126e-01 6.93926066e-02 1.63998097e-01
-5.57753257e-02 -3.59517425e-01 -7.66905189e-01 -3.07606757e-01
-2.52628744e-01 -7.64009416e-01 1.17867589e+00 1.35983706e-01
1.04932225e+00 -2.80764252e-01 -1.97580814e-01 5.30011714e-01
1.17029583e+00 5.15413880e-01 1.09727883e+00 2.52977759e-01
8.05838585e-01 5.61979294e-01 8.10176790e-01 4.61860418e-01
7.24169135e-01 5.34660876e-01 5.84494591e-01 2.14291960e-01
-6.14971042e-01 -5.67143619e-01 9.02468204e-01 5.23531854e-01
-8.29947069e-02 -3.43270302e-01 -8.54149163e-01 6.18843496e-01
-1.95477402e+00 -1.34389269e+00 2.74659127e-01 2.08744121e+00
9.58446383e-01 6.61731139e-02 -8.56606588e-02 -3.13873410e-01
7.50053525e-01 3.30964863e-01 -8.84445667e-01 1.57199338e-01
-7.06044659e-02 4.36625537e-03 -2.36957204e-02 3.11985135e-01
-1.07747293e+00 1.27401149e+00 6.21336794e+00 5.66625297e-01
-8.06803644e-01 -7.33089820e-02 3.67344648e-01 -2.40614235e-01
-1.83441892e-01 -1.82045087e-01 -9.80884373e-01 5.72050035e-01
9.02337790e-01 -2.08575904e-01 4.67761219e-01 6.81983292e-01
1.49716079e-01 -6.26285374e-02 -1.41952503e+00 9.88139212e-01
6.58509016e-01 -1.25693274e+00 6.89418852e-01 -3.99565071e-01
1.07418096e+00 -1.18505284e-01 4.26219046e-01 6.56998873e-01
5.96632421e-01 -7.92168081e-01 9.79979932e-01 7.31714606e-01
9.35856700e-01 -5.78225791e-01 1.28168553e-01 2.89885670e-01
-1.24044812e+00 1.82401612e-02 -7.46225789e-02 4.01624471e-01
3.49292904e-01 8.22037235e-02 -9.65301216e-01 4.49576616e-01
7.22708106e-01 8.30485284e-01 -6.64502382e-01 7.53673732e-01
-6.26099050e-01 4.46193099e-01 2.50661373e-01 -1.98520437e-01
3.58720362e-01 1.20431349e-01 6.30587459e-01 1.27280378e+00
4.34226394e-01 -2.00511992e-01 1.93127587e-01 1.06858981e+00
-1.15421623e-01 -3.96706313e-01 -8.05151403e-01 -1.05096608e-01
4.62118775e-01 1.16604280e+00 -4.44380611e-01 -7.59610116e-01
-8.41700435e-02 1.57489848e+00 1.29066870e-01 5.65831840e-01
-1.11423981e+00 -2.71119654e-01 3.99892241e-01 -2.40498692e-01
2.76390731e-01 -2.16594115e-01 2.30533883e-01 -1.32243812e+00
-8.32924023e-02 -1.08033919e+00 5.27087510e-01 -1.27003348e+00
-7.92047262e-01 6.51411295e-01 1.21103987e-01 -1.30286169e+00
-8.28979075e-01 -1.96322739e-01 -4.30301517e-01 5.14520884e-01
-1.55111802e+00 -1.45032489e+00 -7.81909764e-01 9.02553618e-01
1.18615937e+00 -8.36393088e-02 3.87912899e-01 3.93152162e-02
-4.59542483e-01 3.48356873e-01 -2.47320324e-01 -1.00847259e-01
9.38736737e-01 -1.36037076e+00 5.00605106e-01 1.08567011e+00
2.24384978e-01 3.21741253e-01 7.23657727e-01 -8.20569575e-01
-1.49308789e+00 -1.54059160e+00 6.85154319e-01 -6.36797369e-01
2.26699546e-01 -4.12945867e-01 -6.08620226e-01 9.16115105e-01
3.21874827e-01 -5.00967860e-01 3.60507309e-01 -5.35708725e-01
-1.99870318e-01 2.05850303e-01 -5.60039222e-01 9.91155505e-01
1.55171657e+00 -3.51044148e-01 -6.59263611e-01 4.49504316e-01
9.09800887e-01 -5.77325284e-01 -4.38735574e-01 3.57542872e-01
2.95534641e-01 -8.92961800e-01 1.06890297e+00 -7.08050013e-01
9.33023393e-01 -2.02329770e-01 -9.41627026e-02 -1.55400133e+00
-4.66999024e-01 -7.03232765e-01 -4.68156517e-01 1.22604835e+00
3.85183662e-01 -8.01988915e-02 7.76571691e-01 2.57707506e-01
-5.19098580e-01 -2.11959943e-01 -4.66070145e-01 -8.07460546e-01
-1.65860608e-01 -3.15857977e-01 5.04698098e-01 4.89753962e-01
-4.00650948e-01 6.94878221e-01 -6.43184066e-01 1.97989836e-01
8.64245951e-01 2.44062483e-01 9.10278976e-01 -6.43664658e-01
-1.51115641e-01 -2.21218795e-01 -1.41276658e-01 -1.21299219e+00
3.15526724e-01 -8.14069986e-01 5.64825296e-01 -1.89555740e+00
4.42382753e-01 1.52951404e-01 -2.77427435e-01 3.79313678e-01
-4.54303831e-01 2.58841157e-01 2.99604177e-01 2.95181334e-01
-1.41210234e+00 9.00785804e-01 1.55424356e+00 -5.15123069e-01
-2.39388421e-01 -2.79218554e-01 -7.53864765e-01 5.82488775e-01
4.41299170e-01 -1.98333055e-01 -6.96406901e-01 -5.51798224e-01
4.64855164e-01 2.46695921e-01 6.03327692e-01 -1.41200721e+00
4.40954357e-01 -3.39201391e-01 5.25214434e-01 -8.63976061e-01
6.03238940e-01 -4.48509276e-01 2.49776483e-01 4.39672589e-01
-6.42336786e-01 -1.06908664e-01 3.20797712e-02 8.08445334e-01
-1.00959323e-01 -7.89777637e-02 6.00572348e-01 -4.45096195e-01
-1.51597166e+00 3.79224658e-01 -4.85154837e-01 2.81673282e-01
1.20272589e+00 -3.59978434e-03 -4.17581141e-01 -7.02270627e-01
-8.49308133e-01 4.01518196e-01 4.33099210e-01 6.30540192e-01
9.20059085e-01 -1.55739152e+00 -8.29108000e-01 -6.14680797e-02
5.96700609e-01 -5.95124699e-02 4.32813793e-01 3.69765788e-01
-2.48793527e-01 2.26311982e-01 -2.88997412e-01 -7.74719417e-01
-1.06929433e+00 5.71019709e-01 8.53420347e-02 -3.65517795e-01
-4.32344228e-01 7.87894607e-01 4.78531420e-01 -1.37178794e-01
2.99582273e-01 -1.99509766e-02 -3.65300506e-01 1.42532453e-01
7.01759279e-01 3.42900932e-01 -7.61036649e-02 -5.47051072e-01
1.32515123e-02 3.90530378e-01 -1.65601060e-01 -4.60783660e-01
1.01143110e+00 -3.49662572e-01 1.69987530e-01 5.80057502e-01
1.01604378e+00 -6.83354557e-01 -1.79414594e+00 -3.05999309e-01
-3.85397732e-01 -2.47203950e-02 -3.20492268e-01 -1.05115080e+00
-7.67661333e-01 8.14768732e-01 5.24525344e-01 -1.23741321e-01
1.10988414e+00 2.57211804e-01 8.95517528e-01 7.15653718e-01
4.34903145e-01 -1.21330953e+00 8.82249177e-01 7.05201805e-01
1.21161354e+00 -1.44908440e+00 -2.86099941e-01 -8.75522643e-02
-1.17146838e+00 6.02539718e-01 9.00254607e-01 3.22285324e-01
-2.10757717e-01 2.53820699e-02 -6.92969784e-02 -1.17414698e-01
-1.10277629e+00 -5.63707948e-01 4.26948607e-01 7.71973848e-01
-2.77139395e-02 -3.60099256e-01 1.75899163e-01 3.84165317e-01
-4.42772284e-02 -4.84934002e-02 3.91284913e-01 1.21572137e+00
-3.87828588e-01 -6.74494922e-01 -2.37629071e-01 3.34084928e-01
-1.41113549e-01 -1.50525808e-01 -6.35033488e-01 5.77726960e-01
-1.33424610e-01 9.13719893e-01 6.11160807e-02 -4.15804446e-01
4.47111309e-01 2.97603130e-01 7.38409579e-01 -6.97339833e-01
-2.48239130e-01 -1.83455542e-01 1.32972315e-01 -8.57182562e-01
-4.10465211e-01 -7.92997777e-01 -1.15588975e+00 1.44904956e-01
2.90700674e-01 -7.37581551e-02 7.08150506e-01 6.99768186e-01
5.22924662e-01 9.49155629e-01 5.40165961e-01 -1.18880498e+00
-4.81314778e-01 -8.46342921e-01 -4.64517735e-02 8.69128227e-01
1.98789120e-01 -5.93183100e-01 -3.35453928e-01 7.30794072e-01] | [10.792472839355469, 0.6282715201377869] |
b6e1767c-7aab-4fba-97fe-b9de704115cc | interpretable-multimodal-misinformation | 2305.05964 | null | https://arxiv.org/abs/2305.05964v1 | https://arxiv.org/pdf/2305.05964v1.pdf | Interpretable Multimodal Misinformation Detection with Logic Reasoning | Multimodal misinformation on online social platforms is becoming a critical concern due to increasing credibility and easier dissemination brought by multimedia content, compared to traditional text-only information. While existing multimodal detection approaches have achieved high performance, the lack of interpretability hinders these systems' reliability and practical deployment. Inspired by NeuralSymbolic AI which combines the learning ability of neural networks with the explainability of symbolic learning, we propose a novel logic-based neural model for multimodal misinformation detection which integrates interpretable logic clauses to express the reasoning process of the target task. To make learning effective, we parameterize symbolic logical elements using neural representations, which facilitate the automatic generation and evaluation of meaningful logic clauses. Additionally, to make our framework generalizable across diverse misinformation sources, we introduce five meta-predicates that can be instantiated with different correlations. Results on three public datasets (Twitter, Weibo, and Sarcasm) demonstrate the feasibility and versatility of our model. | ['Haoliang Li', 'Wenya Wang', 'Hui Liu'] | 2023-05-10 | null | null | null | null | ['misinformation'] | ['miscellaneous'] | [ 7.72734210e-02 5.34719169e-01 -5.50251901e-01 -4.94077772e-01
-7.50840485e-01 -7.02152789e-01 8.43898773e-01 4.26365465e-01
-2.00445503e-02 6.97328269e-01 4.70890731e-01 -3.01747739e-01
-2.69008487e-01 -7.61430085e-01 -7.30907440e-01 -1.10648490e-01
-8.06907266e-02 2.45485827e-01 -1.76647771e-02 -3.27469677e-01
-6.01415634e-02 1.28456131e-01 -1.47000372e+00 8.30549479e-01
1.13438511e+00 1.11275554e+00 -4.87076312e-01 2.96847314e-01
-3.18808824e-01 1.25834525e+00 -5.87045729e-01 -9.10619736e-01
-3.08422357e-01 -1.65808216e-01 -6.43485844e-01 -4.94522043e-02
3.04599792e-01 -3.92835230e-01 -3.15256387e-01 1.32432973e+00
-8.52049366e-02 -2.32512429e-01 5.76146305e-01 -1.53501225e+00
-8.36068332e-01 1.34532583e+00 -9.91949961e-02 -1.92591906e-01
6.03422642e-01 8.66814852e-02 1.37947512e+00 -5.67381918e-01
6.28753483e-01 1.64156950e+00 5.91235936e-01 3.42368275e-01
-1.07477367e+00 -5.47759175e-01 2.89572239e-01 2.95979053e-01
-1.10160935e+00 -3.01299065e-01 7.76299953e-01 -2.17804387e-01
6.70015574e-01 4.99952644e-01 5.66444099e-01 1.29623270e+00
-2.84600705e-02 1.17303360e+00 8.45774472e-01 -1.45388976e-01
2.94163495e-01 3.39437395e-01 1.75894573e-01 9.86659050e-01
4.43573862e-01 -3.21840733e-01 -6.89966261e-01 -1.56276211e-01
3.41366440e-01 -1.01665007e-02 -1.53915986e-01 3.16456571e-04
-1.13035691e+00 9.61642742e-01 5.81498623e-01 2.34811619e-01
-3.48607332e-01 4.97479707e-01 4.16273236e-01 8.14060643e-02
3.08222622e-01 7.21956015e-01 -9.20360629e-03 -1.59314170e-01
-6.17513716e-01 2.99852222e-01 8.06475461e-01 8.05373609e-01
5.47135592e-01 -1.23637766e-01 -1.24115437e-01 5.71166515e-01
5.17358005e-01 5.77982426e-01 3.76621723e-01 -1.06704557e+00
5.61150014e-01 1.09184611e+00 8.26959983e-02 -1.57003939e+00
-4.90648121e-01 -3.07849050e-01 -6.68333113e-01 -4.00644869e-01
3.48484367e-01 -2.28793502e-01 -5.08224428e-01 1.62771642e+00
-1.61940139e-02 -1.04183130e-01 1.36111632e-01 1.00142109e+00
8.78956497e-01 5.83001137e-01 2.49578193e-01 1.56812683e-01
1.32739079e+00 -6.12039268e-01 -8.52344573e-01 -4.31639284e-01
5.78517854e-01 -4.01979424e-02 1.04073966e+00 5.20842493e-01
-8.31351280e-01 1.47078156e-01 -1.14984477e+00 -1.28938094e-01
-5.97586751e-01 -4.86081578e-02 9.83711720e-01 4.55607504e-01
-5.71662009e-01 2.06468627e-01 -5.34079731e-01 -6.45978153e-02
5.82875192e-01 1.97172344e-01 -2.58636594e-01 -3.28783877e-02
-1.71305084e+00 7.25445449e-01 6.91165924e-01 1.28799811e-01
-6.05611563e-01 -2.57093102e-01 -8.22004378e-01 2.38363519e-01
6.63621843e-01 -6.33689106e-01 1.07429373e+00 -1.36571372e+00
-1.31939185e+00 4.57552910e-01 1.34760812e-01 -5.55474520e-01
3.97213310e-01 -3.68301958e-01 -4.53754753e-01 5.40360630e-01
-6.73297048e-02 6.77629411e-01 7.22450078e-01 -1.35792100e+00
-3.62125486e-01 -6.53747022e-02 6.69256687e-01 2.11085424e-01
-6.54794753e-01 -6.15255348e-02 -4.61628288e-01 -3.93406004e-01
1.33954629e-01 -6.52996004e-01 8.58389363e-02 -1.01457216e-01
-9.18321371e-01 -2.57564664e-01 5.92991769e-01 -5.59799612e-01
1.46419954e+00 -2.22576237e+00 3.39264154e-01 5.49953878e-01
5.70514083e-01 -1.44458367e-02 -3.40943150e-02 4.09944892e-01
1.77720502e-01 6.13818586e-01 8.65861177e-02 -1.76538661e-01
4.26744699e-01 2.91350126e-01 -4.55495030e-01 6.73748255e-02
5.69570482e-01 1.02065456e+00 -1.05979240e+00 -6.00066900e-01
-1.57507837e-01 2.34434560e-01 -5.30739546e-01 1.92276035e-02
-7.43940413e-01 -5.40193021e-02 -7.50462234e-01 1.01883209e+00
1.22016311e-01 -6.69827044e-01 4.77775484e-01 -2.02718660e-01
2.37015143e-01 2.01624349e-01 -8.56299400e-01 1.11918390e+00
-2.10556045e-01 5.73679864e-01 -3.72251458e-02 -5.19882202e-01
6.03675067e-01 1.44722864e-01 1.45215154e-01 -6.06817544e-01
4.01359409e-01 -2.32720282e-02 -3.35696965e-01 -8.90331924e-01
5.19256294e-01 -8.47449526e-02 -3.99329036e-01 4.43933398e-01
-3.49692732e-01 1.50364473e-01 2.24387571e-01 4.82955128e-01
8.60105693e-01 -1.27523839e-01 9.21001732e-02 2.76189625e-01
4.91185099e-01 1.97910830e-01 2.05759779e-01 7.28606522e-01
4.81663831e-02 1.53975248e-01 1.09737825e+00 -2.27324918e-01
-6.07908487e-01 -9.02404785e-01 1.68667622e-02 1.18066883e+00
3.47529083e-01 -6.06652677e-01 -6.98937476e-01 -8.18836212e-01
1.42298326e-01 8.30894887e-01 -5.82916856e-01 -1.72824979e-01
-3.08243722e-01 -6.93938255e-01 9.26153183e-01 2.93699414e-01
5.42929173e-01 -6.93895996e-01 -3.77753735e-01 -1.09710298e-01
-7.20627129e-01 -1.21262753e+00 1.11168660e-01 -2.93497056e-01
-6.14029646e-01 -1.11210763e+00 1.16990112e-01 -3.19300920e-01
7.09430516e-01 7.20441490e-02 8.69121552e-01 3.29542786e-01
1.88135758e-01 4.30367619e-01 -3.25579822e-01 -3.13946337e-01
-5.95267415e-01 6.75650761e-02 -8.33173841e-03 1.45634949e-01
2.17757225e-01 -6.10076673e-02 -2.04897165e-01 -9.51998215e-03
-1.30774152e+00 2.77876467e-01 6.33395851e-01 8.75242472e-01
-1.28134945e-02 2.10991144e-01 6.27899587e-01 -9.04302776e-01
9.69538212e-01 -9.47496235e-01 -4.14422750e-01 3.15502107e-01
-5.45735419e-01 2.03130022e-01 3.75397235e-01 -4.95807081e-01
-9.94241416e-01 -2.22469032e-01 1.76838532e-01 -2.04711780e-01
1.44665152e-01 1.27795947e+00 -5.01448363e-02 1.33209199e-01
7.23891079e-01 -1.51823744e-01 2.34256759e-01 -7.48708770e-02
7.02492774e-01 6.11103237e-01 3.73436302e-01 -4.99510318e-01
8.46585512e-01 3.34344178e-01 -8.00909847e-02 -6.68609142e-01
-8.48570287e-01 1.36328489e-02 1.82423797e-02 -4.82520103e-01
7.63070643e-01 -8.30274999e-01 -1.22279489e+00 7.72855952e-02
-1.27807653e+00 1.67957664e-01 3.20712417e-01 2.67484576e-01
-2.04529360e-01 4.74406838e-01 -8.00126255e-01 -1.09522414e+00
-3.14264707e-02 -1.00337267e+00 9.36562836e-01 1.54626677e-02
-5.21017015e-01 -1.07213676e+00 -4.91025120e-01 5.95555723e-01
2.75337964e-01 4.08674866e-01 1.41366661e+00 -8.77484381e-01
-8.46922398e-01 -5.29528201e-01 -5.76744020e-01 7.56270206e-03
-3.46387982e-01 2.33811140e-02 -9.47637379e-01 2.93963164e-01
-5.20710051e-01 -7.56243944e-01 5.83767474e-01 6.65491959e-03
1.18490016e+00 -1.05137622e+00 -3.79352659e-01 3.18549275e-02
1.17684948e+00 -2.96931356e-01 5.14855385e-01 5.20192564e-01
6.08022034e-01 9.27132189e-01 4.46934611e-01 4.27168250e-01
7.88682520e-01 4.30811644e-01 9.33985889e-01 2.25910544e-02
3.04993331e-01 -5.82080007e-01 6.23329937e-01 4.55508500e-01
1.57539323e-01 -4.03315157e-01 -1.00068700e+00 1.83498621e-01
-2.01127934e+00 -1.15855074e+00 -8.54271650e-02 1.79167473e+00
9.70190287e-01 2.30750322e-01 1.12825841e-01 3.51830348e-02
5.14815509e-01 3.94712538e-02 -3.53608489e-01 -2.00065538e-01
-3.00965846e-01 -5.83931208e-01 2.13513851e-01 6.56588674e-01
-1.00907612e+00 8.05939853e-01 6.62247229e+00 5.24403632e-01
-1.15026414e+00 -4.46592979e-02 7.11989403e-01 -8.58526230e-02
-9.24169958e-01 -2.58112699e-01 -4.88695234e-01 3.32911760e-01
8.63908112e-01 -7.73709640e-02 5.32632828e-01 7.25624382e-01
9.06028971e-02 -1.19681872e-01 -1.25187778e+00 7.57000864e-01
2.09820077e-01 -1.48236501e+00 2.80012399e-01 -1.73825920e-01
4.73019123e-01 -2.60445982e-01 1.77214369e-01 3.60121727e-01
1.79141536e-01 -1.18770933e+00 9.86032724e-01 6.98723376e-01
5.31386673e-01 -6.46211028e-01 8.64493430e-01 4.20395851e-01
-4.27978754e-01 -5.04802167e-01 2.25311503e-01 -2.04226207e-02
2.88513675e-03 5.27167082e-01 -1.24391913e+00 3.81618887e-01
3.77281666e-01 7.00295269e-01 -9.43986356e-01 5.88674843e-01
-4.68429714e-01 5.00718772e-01 -1.68617055e-01 -5.75014532e-01
2.40681812e-01 9.95979607e-02 5.68480134e-01 1.27302611e+00
4.53000069e-02 -3.20911199e-01 1.01748042e-01 1.21047878e+00
-2.51570493e-01 -1.02161400e-01 -6.49736822e-01 -7.73995817e-01
5.25044620e-01 1.14552784e+00 -4.48000342e-01 -3.09765369e-01
-2.92528450e-01 6.00491047e-01 3.23994458e-01 6.18066847e-01
-1.02468181e+00 -3.28558862e-01 3.47418278e-01 -2.43579298e-02
-1.96504876e-01 -1.12496652e-01 -5.41456997e-01 -1.20361423e+00
-3.62692140e-02 -1.15036714e+00 2.75365919e-01 -7.75707424e-01
-1.41229832e+00 4.60991502e-01 3.01030815e-01 -9.51511383e-01
-4.28258032e-01 -6.98121071e-01 -2.63190717e-01 3.80329460e-01
-1.43866313e+00 -1.42768383e+00 -1.80920377e-01 6.72141090e-02
6.79399148e-02 -5.42442501e-03 6.49921596e-01 1.49050251e-01
-6.90876663e-01 5.23280561e-01 -3.93258959e-01 2.27455005e-01
5.28719246e-01 -1.02935135e+00 -1.00688063e-01 6.10833704e-01
-2.81811446e-01 9.40135956e-01 7.98456609e-01 -7.59632111e-01
-1.79123747e+00 -9.95660782e-01 1.00967848e+00 -6.88828170e-01
1.14130008e+00 -3.13790768e-01 -8.54479969e-01 7.66052067e-01
1.65063083e-01 -6.35973275e-01 7.59225368e-01 3.75334412e-01
-7.51734734e-01 -1.77131593e-02 -1.05173147e+00 9.41276729e-01
7.29043007e-01 -6.60663247e-01 -6.53254926e-01 4.45984274e-01
7.05524921e-01 -2.82068282e-01 -6.35301054e-01 1.69261947e-01
6.02890790e-01 -7.65760422e-01 8.97193015e-01 -8.38756144e-01
8.45890939e-01 -2.67344713e-01 -1.47561282e-01 -9.20001209e-01
-5.45843542e-02 -4.29683834e-01 -4.14738983e-01 1.13707054e+00
7.25370228e-01 -6.10072851e-01 5.74714839e-01 9.96112525e-01
9.78498980e-02 -6.58027768e-01 -7.74110317e-01 -3.20757359e-01
-2.56804436e-01 -6.91633224e-01 7.18010664e-01 9.84950304e-01
7.99242198e-01 3.58345836e-01 -3.03705752e-01 3.97919536e-01
4.15610194e-01 -9.04905796e-02 4.07693267e-01 -1.19503713e+00
-4.33851004e-01 -7.82485962e-01 -3.08677077e-01 -9.14090157e-01
2.66689748e-01 -1.05068398e+00 3.49330567e-02 -1.38150084e+00
2.48080969e-01 -4.11873907e-01 -3.18782311e-03 8.72605085e-01
2.49247551e-02 3.19483846e-01 1.82434961e-01 1.49076238e-01
-1.01784158e+00 4.03219789e-01 1.02609360e+00 -4.02659804e-01
3.27019654e-02 -4.74097818e-01 -1.02938557e+00 8.37107182e-01
5.31063855e-01 -2.32350782e-01 -4.99106437e-01 -5.43155909e-01
1.24860108e+00 1.68951377e-01 6.84816301e-01 -6.97816432e-01
2.29403362e-01 -2.87804514e-01 2.54549384e-01 -2.18815401e-01
4.12285805e-01 -7.85127640e-01 -2.97676530e-02 4.32049483e-01
-8.37568760e-01 2.79459599e-02 4.80436981e-02 7.45158494e-01
-3.55302423e-01 -1.86476663e-01 1.87304839e-01 2.00822636e-01
-5.32226324e-01 -1.04455918e-01 -6.84345424e-01 -1.43899262e-01
8.96750033e-01 1.28864929e-01 -8.12073469e-01 -9.56961215e-01
-4.00701612e-01 3.68310213e-01 2.63649940e-01 6.03393614e-01
8.12978268e-01 -1.33698690e+00 -3.69899541e-01 -1.92357805e-02
3.81835699e-01 -1.83804035e-01 -5.54195344e-02 9.56174076e-01
-3.95615011e-01 3.68998200e-01 1.16330065e-01 -4.68917698e-01
-9.43698466e-01 3.13766003e-01 2.70683795e-01 5.94478361e-02
-1.87891081e-01 5.05250752e-01 -1.23721540e-01 -3.36096406e-01
5.51897109e-01 -5.24356067e-01 -3.69182438e-01 3.86044127e-03
6.74013674e-01 3.05887848e-01 -2.85419464e-01 -3.76113176e-01
-2.00573683e-01 -1.91562012e-01 7.18022436e-02 -1.06625460e-01
9.79887009e-01 -1.31569147e-01 -4.03618097e-01 5.89347184e-01
7.42500007e-01 9.50528458e-02 -8.38648498e-01 -5.46013117e-02
2.63272583e-01 -2.23264053e-01 9.10612848e-03 -1.17720175e+00
-5.49464643e-01 6.61737859e-01 1.08471505e-01 6.53690100e-01
8.02804470e-01 1.86300278e-01 6.07715905e-01 7.77737200e-01
1.17839374e-01 -1.04540873e+00 4.05512422e-01 6.48689330e-01
9.62601900e-01 -1.36818898e+00 -1.25079632e-01 -6.43647492e-01
-8.64021957e-01 1.34146750e+00 5.55427551e-01 4.61028188e-01
1.85524717e-01 -2.64888001e-03 1.97435599e-02 -3.06842923e-01
-7.81504929e-01 8.59915614e-02 4.14809436e-01 2.45151445e-01
4.60196972e-01 3.40715162e-02 -1.84980020e-01 9.50152218e-01
-1.86120998e-02 1.66978333e-02 4.42581892e-01 6.75780952e-01
-5.20293534e-01 -7.62243688e-01 -3.72334808e-01 3.39824647e-01
-4.66416568e-01 -3.10372766e-02 -7.31485069e-01 6.71010971e-01
1.34705810e-03 1.09761727e+00 -9.83573794e-02 -6.23809397e-01
-9.64836851e-02 1.18362688e-01 3.50462086e-02 -4.00538772e-01
-5.31241417e-01 -2.06479430e-01 7.15748250e-01 -6.36555254e-01
-3.32347006e-01 -4.52424914e-01 -1.52029562e+00 -4.73978430e-01
-1.09763443e-02 2.43016332e-03 7.46782124e-01 1.23846495e+00
4.82707262e-01 3.70187134e-01 3.21513385e-01 -4.21359539e-01
-6.69898450e-01 -6.58706188e-01 -7.84158483e-02 6.09890282e-01
4.17443275e-01 -6.29936516e-01 -2.25563049e-01 -2.23664846e-02] | [9.675955772399902, 8.023552894592285] |
c0bbf1f2-4297-48d2-88ac-1db037bab2c1 | relational-self-supervised-learning | 2203.08717 | null | https://arxiv.org/abs/2203.08717v1 | https://arxiv.org/pdf/2203.08717v1.pdf | Relational Self-Supervised Learning | Self-supervised Learning (SSL) including the mainstream contrastive learning has achieved great success in learning visual representations without data annotations. However, most methods mainly focus on the instance level information (\ie, the different augmented images of the same instance should have the same feature or cluster into the same class), but there is a lack of attention on the relationships between different instances. In this paper, we introduce a novel SSL paradigm, which we term as relational self-supervised learning (ReSSL) framework that learns representations by modeling the relationship between different instances. Specifically, our proposed method employs sharpened distribution of pairwise similarities among different instances as \textit{relation} metric, which is thus utilized to match the feature embeddings of different augmentations. To boost the performance, we argue that weak augmentations matter to represent a more reliable relation, and leverage momentum strategy for practical efficiency. The designed asymmetric predictor head and an InfoNCE warm-up strategy enhance the robustness to hyper-parameters and benefit the resulting performance. Experimental results show that our proposed ReSSL substantially outperforms the state-of-the-art methods across different network architectures, including various lightweight networks (\eg, EfficientNet and MobileNet). | ['Chang Xu', 'Xiaogang Wang', 'ChangShui Zhang', 'Chen Qian', 'Fei Wang', 'Shan You', 'Mingkai Zheng'] | 2022-03-16 | null | null | null | null | ['self-supervised-image-classification'] | ['computer-vision'] | [ 1.08887792e-01 6.95387051e-02 -6.01891160e-01 -5.06695807e-01
-2.51507759e-01 -2.26541281e-01 6.18568063e-01 1.30208611e-01
-2.32199073e-01 6.37814879e-01 3.19985747e-01 -2.06583999e-02
-3.87906581e-01 -5.78135848e-01 -8.74417543e-01 -7.85166740e-01
-6.74247444e-02 3.65833305e-02 1.26222894e-01 -1.09066911e-01
3.96164022e-02 8.90688971e-02 -1.15721214e+00 2.54571706e-01
7.53366411e-01 1.49312437e+00 1.89378366e-01 1.72189884e-02
-2.14336231e-01 1.05537999e+00 -2.60717452e-01 -4.40444648e-01
3.80023241e-01 -3.01067710e-01 -7.29672909e-01 2.56805420e-02
5.71188092e-01 -8.63147229e-02 -6.74104273e-01 1.04495537e+00
2.71788120e-01 1.47081763e-01 6.91310287e-01 -1.56619883e+00
-8.75129640e-01 7.42373109e-01 -1.00395775e+00 4.13495094e-01
-1.83690816e-01 6.71574101e-02 1.32144582e+00 -9.79858637e-01
5.21313071e-01 1.18958163e+00 5.54172575e-01 3.12095344e-01
-1.22593772e+00 -9.60231245e-01 6.81084335e-01 5.05099773e-01
-1.51298189e+00 -4.27524388e-01 1.03381991e+00 -2.21698940e-01
4.95103478e-01 -5.70889078e-02 3.92814189e-01 1.14720714e+00
-1.65432394e-01 7.90901124e-01 1.20481145e+00 -2.57177293e-01
1.53646052e-01 4.39020693e-01 2.51187354e-01 8.00121427e-01
2.66141266e-01 -1.44684672e-01 -6.90230727e-01 2.08765164e-01
7.41798460e-01 4.86059278e-01 -2.62291729e-01 -7.36649156e-01
-1.19896245e+00 7.12361276e-01 1.19120574e+00 1.72475487e-01
-3.16147208e-01 5.58759421e-02 3.67014170e-01 3.08788002e-01
5.41612804e-01 1.19117595e-01 -3.58015746e-01 2.28016004e-01
-6.21792853e-01 -1.55110687e-01 3.88288349e-01 1.02558136e+00
1.02787876e+00 3.58955599e-02 -8.14652517e-02 9.22366977e-01
4.03647006e-01 4.38487753e-02 5.05959213e-01 -5.29703557e-01
8.88111591e-01 9.41280484e-01 -4.62290138e-01 -1.30464947e+00
-3.39606315e-01 -8.93253505e-01 -1.24978566e+00 -4.41375794e-03
6.41445965e-02 -1.26666576e-01 -7.30218649e-01 1.90169370e+00
1.34016663e-01 5.90952873e-01 2.59706080e-01 8.59814644e-01
1.14178109e+00 5.53865552e-01 1.51535481e-01 -2.07953826e-01
1.00660288e+00 -1.17323375e+00 -4.58600014e-01 -4.11000520e-01
5.85375428e-01 -4.00628328e-01 8.95433962e-01 -1.16768323e-01
-8.84396911e-01 -7.75624573e-01 -1.21338892e+00 2.00624749e-01
-3.71188223e-01 1.17537320e-01 6.34637773e-01 2.36964285e-01
-8.62771749e-01 5.32790422e-01 -5.54178536e-01 -4.95192975e-01
8.74706924e-01 2.73550212e-01 -5.48196077e-01 -7.92631432e-02
-9.99722004e-01 6.46488726e-01 3.24971527e-01 1.27331898e-01
-6.47355974e-01 -8.00833821e-01 -7.97444522e-01 6.61772192e-02
5.31721115e-01 -4.25698102e-01 5.65917611e-01 -1.14889896e+00
-1.01087844e+00 8.46362352e-01 -9.57942307e-02 -6.51055694e-01
3.10790598e-01 -2.32345968e-01 -4.79253411e-01 7.90276378e-02
1.09270982e-01 6.79916084e-01 9.80303586e-01 -1.62420213e+00
-4.81287688e-01 -3.79748315e-01 1.41411617e-01 4.35331762e-01
-7.50205398e-01 -1.72634438e-01 -4.58745807e-01 -6.29122853e-01
2.83840507e-01 -8.89633596e-01 -1.20628543e-01 2.32962087e-01
-5.43038964e-01 -2.91271925e-01 9.77633953e-01 -1.60505325e-01
1.13492358e+00 -2.36429858e+00 1.26305863e-01 4.49438959e-01
6.40421748e-01 4.03090239e-01 -1.11596487e-01 3.53022248e-01
-3.87954921e-01 2.81433892e-02 -1.06343009e-01 -4.11784679e-01
-2.00234562e-01 4.35173392e-01 -3.57988596e-01 4.06742603e-01
3.30446571e-01 8.45525861e-01 -8.79262924e-01 -7.24935055e-01
7.90768266e-02 5.37100434e-01 -3.57340723e-01 4.09905404e-01
1.71564311e-01 4.76718754e-01 -3.72840703e-01 4.29050118e-01
6.69098437e-01 -7.06589401e-01 2.02326208e-01 -6.53933167e-01
1.44721702e-01 1.85769126e-01 -1.20089626e+00 1.69291556e+00
-4.32296067e-01 4.79781926e-01 -1.81894436e-01 -1.29538858e+00
1.10228038e+00 -1.29448008e-02 3.35133433e-01 -7.72777140e-01
-1.01332098e-01 -9.81368124e-02 8.05754364e-02 -4.48262870e-01
1.80500835e-01 2.18720004e-01 4.17539388e-01 3.28908533e-01
2.44496018e-01 6.07123494e-01 -8.73913243e-02 5.11424661e-01
8.23514044e-01 2.00153172e-01 2.90230989e-01 -1.19837739e-01
3.64182830e-01 -4.55766708e-01 6.76636636e-01 6.87255859e-01
-1.86351776e-01 5.16414344e-01 6.49188876e-01 -2.60806441e-01
-7.13018239e-01 -9.83216941e-01 -9.95623544e-02 1.29532981e+00
5.86744189e-01 -5.31840622e-01 -3.40568542e-01 -9.67340171e-01
3.69170518e-03 1.92392305e-01 -8.55026722e-01 -2.91872978e-01
-4.96516287e-01 -8.23076546e-01 2.28220671e-01 7.87916839e-01
8.23950171e-01 -1.01530516e+00 -1.58391610e-01 -1.05938345e-01
1.74740747e-01 -1.21478522e+00 -2.31906220e-01 3.52778196e-01
-8.33203316e-01 -1.04314017e+00 -4.99658793e-01 -9.92436111e-01
8.15979540e-01 5.88212788e-01 9.63937700e-01 8.50593075e-02
-2.31900699e-02 1.57417297e-01 -5.27193606e-01 7.22602159e-02
1.79748446e-01 1.56054333e-01 1.13132060e-01 5.06851315e-01
1.19315989e-01 -9.22594786e-01 -9.01548326e-01 3.98926437e-01
-5.88799059e-01 1.06681839e-01 8.42060566e-01 8.59986722e-01
6.84997439e-01 -4.18774085e-03 6.79923892e-01 -1.10765541e+00
3.34183604e-01 -9.07664537e-01 -2.83839405e-02 3.42957288e-01
-8.72645438e-01 7.66369253e-02 8.42458308e-01 -4.20284033e-01
-8.73551905e-01 -2.90266514e-01 3.78906935e-01 -7.46870458e-01
-1.30134169e-02 5.70500314e-01 -1.90041140e-01 -2.36377642e-01
4.64702278e-01 3.05211097e-01 8.62080138e-03 -4.49327230e-01
5.02993047e-01 3.73074859e-01 4.35582161e-01 -4.70205247e-01
1.06137359e+00 5.66372514e-01 1.44142777e-01 -5.62059164e-01
-1.22075212e+00 -5.07820070e-01 -5.88337898e-01 -7.32932240e-02
4.92702216e-01 -1.05233240e+00 -5.46360970e-01 4.65143993e-02
-7.40527391e-01 -9.83039364e-02 -3.23129237e-01 4.09840733e-01
-2.15369999e-01 1.79342225e-01 -3.32970530e-01 -6.25441492e-01
-3.24406445e-01 -9.67483580e-01 6.34241760e-01 4.89040405e-01
9.37238708e-02 -9.70120847e-01 -6.07682988e-02 1.94151506e-01
4.84249264e-01 1.70958638e-01 9.73802388e-01 -8.72452021e-01
-5.07633090e-01 9.99852940e-02 -9.46908832e-01 4.70465243e-01
2.37963438e-01 -2.17043221e-01 -9.86209095e-01 -3.76939893e-01
-4.69432622e-01 -4.68756437e-01 1.08000505e+00 1.22554168e-01
1.40906405e+00 -4.40360785e-01 -5.41291296e-01 8.77229393e-01
1.34098613e+00 -1.17676482e-01 5.39198697e-01 3.03927690e-01
1.29628229e+00 6.93602026e-01 3.43202531e-01 3.74478459e-01
6.11110985e-01 6.01140261e-01 6.37317419e-01 -3.63626927e-01
-3.51554126e-01 -4.96677816e-01 1.27808034e-01 9.37510669e-01
-2.11500466e-01 -5.51486537e-02 -6.30571663e-01 4.01858449e-01
-2.14842916e+00 -8.27190042e-01 8.98113325e-02 2.03552032e+00
6.37348711e-01 2.88910657e-01 1.08523875e-01 2.87280586e-02
8.19004595e-01 5.62685668e-01 -6.82941854e-01 1.86757773e-01
-2.75531262e-01 2.41700023e-01 3.35634738e-01 3.75921391e-02
-1.18654954e+00 9.07445669e-01 4.95213461e+00 8.86682212e-01
-1.07329309e+00 4.21328433e-02 8.49497855e-01 8.73500016e-03
2.83788331e-02 8.54335427e-02 -7.10451841e-01 4.15258825e-01
4.60461378e-01 2.79103816e-02 2.08774358e-01 8.54061604e-01
1.75398942e-02 3.51084977e-01 -1.07012272e+00 1.16674614e+00
1.63799584e-01 -1.33977902e+00 1.38589963e-01 -6.74379617e-02
7.82815576e-01 -1.54857012e-02 4.77572650e-01 4.40824568e-01
6.51911050e-02 -1.00905883e+00 4.26657200e-01 5.60481966e-01
7.03877687e-01 -7.75872886e-01 6.76898122e-01 -2.90535465e-02
-1.49904680e+00 -1.50436118e-01 -5.06199419e-01 4.10881117e-02
-2.07351208e-01 3.32621008e-01 -6.82079434e-01 6.32935464e-01
7.83627570e-01 1.40697896e+00 -9.48793650e-01 1.09619009e+00
-1.63314313e-01 7.50605106e-01 4.07114178e-02 1.35928884e-01
3.51337224e-01 -1.54389396e-01 3.50476325e-01 1.08161438e+00
-1.32701650e-01 -2.10934073e-01 3.96875948e-01 7.31084585e-01
-4.25899982e-01 2.40348652e-01 -5.07117987e-01 1.97330043e-01
5.57250142e-01 1.42083967e+00 -5.27932584e-01 -1.87978148e-01
-6.30198956e-01 7.50386298e-01 7.99842179e-01 5.56456506e-01
-6.85428381e-01 -9.98864919e-02 6.70176089e-01 1.44706190e-01
3.50175738e-01 7.71851763e-02 -2.63573140e-01 -1.02616036e+00
1.21074140e-01 -5.56014419e-01 6.21541560e-01 -6.58049703e-01
-1.70693982e+00 8.07747543e-01 -1.01522803e-01 -1.57929265e+00
1.45796612e-01 -4.25922424e-01 -7.85088122e-01 5.95166326e-01
-1.91556025e+00 -1.44111574e+00 -5.35730779e-01 7.49424696e-01
3.09948802e-01 -4.20267016e-01 6.11354053e-01 2.66162843e-01
-9.74475920e-01 9.35995102e-01 4.12858389e-02 2.46362790e-01
9.06500638e-01 -1.11992681e+00 -1.38130609e-03 6.58589721e-01
3.80622029e-01 8.71736586e-01 2.57120579e-01 -4.02232200e-01
-1.11279774e+00 -1.42499650e+00 3.02966028e-01 -4.58368324e-02
8.17763388e-01 -1.64053500e-01 -1.07216096e+00 6.95383787e-01
2.45645627e-01 7.21450925e-01 7.55466104e-01 1.93102151e-01
-8.89175177e-01 -5.51568270e-01 -8.58601749e-01 5.34793615e-01
1.42055476e+00 -6.73635483e-01 -3.23085070e-01 2.53669918e-01
7.75737941e-01 -1.61878578e-02 -8.09059620e-01 4.31024343e-01
4.16033864e-01 -9.75404263e-01 1.03612828e+00 -8.76308918e-01
6.34381056e-01 -4.01089549e-01 -2.28241667e-01 -1.22349930e+00
-4.27801371e-01 -4.78002250e-01 -4.88808662e-01 1.53014421e+00
4.19862449e-01 -6.50683880e-01 8.37490022e-01 3.99448872e-01
7.83156455e-02 -1.10345197e+00 -6.83314025e-01 -7.12707937e-01
-2.32658535e-01 -1.04913488e-01 5.37817717e-01 1.32854080e+00
4.74301763e-02 5.56915164e-01 -5.74499965e-01 2.53379375e-01
7.77381837e-01 2.56578654e-01 6.33413792e-01 -1.28928113e+00
-3.35843772e-01 -4.13201302e-01 -7.19764411e-01 -1.05004406e+00
3.12751025e-01 -1.13475597e+00 -4.26404387e-01 -1.23554552e+00
5.44336438e-01 -8.27747226e-01 -8.73736382e-01 6.24363303e-01
-4.36472595e-01 3.90524387e-01 2.26110294e-01 5.25449455e-01
-1.18363297e+00 8.38509679e-01 1.08122265e+00 -1.00928739e-01
-7.57001564e-02 -6.36216030e-02 -9.52561736e-01 6.82966769e-01
6.30506694e-01 -3.74662966e-01 -6.06227994e-01 -3.89858335e-01
1.86858118e-01 -3.21218759e-01 6.28348172e-01 -1.04257035e+00
4.18984830e-01 1.48696870e-01 4.97818321e-01 -2.33634919e-01
2.18928531e-01 -9.53435481e-01 -2.57598668e-01 3.56131531e-02
-7.24608302e-01 1.56704541e-02 -1.85172975e-01 9.80365515e-01
-2.51120001e-01 1.86664537e-02 7.31004119e-01 5.18156923e-02
-8.93255949e-01 6.33678138e-01 4.57354337e-01 2.59093940e-01
8.90610933e-01 -1.44709468e-01 -4.49166387e-01 -3.51437390e-01
-5.53406775e-01 2.79709220e-01 1.60747960e-01 5.11054039e-01
7.74478197e-01 -1.51169801e+00 -6.04305983e-01 1.08506456e-01
5.50790310e-01 8.51079598e-02 2.07268670e-01 8.22583556e-01
-2.80674379e-02 4.54046056e-02 -1.83725193e-01 -6.69671774e-01
-1.04772997e+00 5.09814858e-01 1.24056026e-01 -3.11219901e-01
-7.95667350e-01 1.06518412e+00 5.34781337e-01 -2.55988210e-01
4.59156811e-01 1.24929167e-01 -5.09815276e-01 1.41002432e-01
5.26784658e-01 1.63494721e-01 -1.43165529e-01 -8.36841881e-01
-4.18428481e-01 5.71850598e-01 -6.57461762e-01 4.38313544e-01
1.48547983e+00 -1.31763190e-01 1.01495817e-01 5.15988708e-01
1.48312151e+00 -2.76884377e-01 -1.59165251e+00 -9.30621147e-01
-1.23875275e-01 -4.77348834e-01 1.42889231e-01 -5.45359373e-01
-1.63269603e+00 6.85899556e-01 7.39663959e-01 1.32130221e-01
9.43422377e-01 1.64903969e-01 5.66789508e-01 3.19678515e-01
2.68840015e-01 -8.12691987e-01 3.03097993e-01 2.92993724e-01
6.34135067e-01 -1.49042284e+00 1.34852707e-01 -5.11788130e-01
-1.00705755e+00 8.91762078e-01 9.82500136e-01 -3.51215661e-01
7.20653117e-01 -6.78619146e-02 -1.30312711e-01 -2.75733709e-01
-7.12421834e-01 -3.87811810e-01 3.99328262e-01 7.42020369e-01
4.32281345e-01 -5.44306897e-02 1.11997694e-01 5.19310832e-01
7.10870922e-02 -3.74449998e-01 1.89369563e-02 7.60307968e-01
-9.91206765e-02 -8.68614674e-01 7.85914212e-02 6.74753070e-01
-1.39848694e-01 -2.90074706e-01 -1.80994824e-01 7.19613612e-01
1.33546248e-01 7.36750662e-01 1.95982799e-01 -6.00826561e-01
2.55705953e-01 -1.89185873e-01 3.18515331e-01 -5.25058627e-01
-5.34413457e-01 -2.74849713e-01 -1.75958842e-01 -5.52675962e-01
-7.37961709e-01 -3.79286706e-01 -1.10311866e+00 -1.33711919e-01
-2.42980033e-01 -7.51630515e-02 1.84786573e-01 1.08484507e+00
4.33024555e-01 5.93116462e-01 1.01069355e+00 -6.34720445e-01
-3.05968374e-01 -9.07430768e-01 -5.37351847e-01 5.76702476e-01
3.04948896e-01 -9.16430950e-01 -3.87380660e-01 -1.67836934e-01] | [9.628482818603516, 2.687084913253784] |
329b4098-72f0-4d8a-95a9-a7c839a02fd7 | guided-speech-enhancement-network | 2303.07486 | null | https://arxiv.org/abs/2303.07486v1 | https://arxiv.org/pdf/2303.07486v1.pdf | Guided Speech Enhancement Network | High quality speech capture has been widely studied for both voice communication and human computer interface reasons. To improve the capture performance, we can often find multi-microphone speech enhancement techniques deployed on various devices. Multi-microphone speech enhancement problem is often decomposed into two decoupled steps: a beamformer that provides spatial filtering and a single-channel speech enhancement model that cleans up the beamformer output. In this work, we propose a speech enhancement solution that takes both the raw microphone and beamformer outputs as the input for an ML model. We devise a simple yet effective training scheme that allows the model to learn from the cues of the beamformer by contrasting the two inputs and greatly boost its capability in spatial rejection, while conducting the general tasks of denoising and dereverberation. The proposed solution takes advantage of classical spatial filtering algorithms instead of competing with them. By design, the beamformer module then could be selected separately and does not require a large amount of data to be optimized for a given form factor, and the network model can be considered as a standalone module which is highly transferable independently from the microphone array. We name the ML module in our solution as GSENet, short for Guided Speech Enhancement Network. We demonstrate its effectiveness on real world data collected on multi-microphone devices in terms of the suppression of noise and interfering speech. | ['Matthias Grundmann', 'George Sung', 'Yunpeng Li', 'Chehung Lee', 'Jamie Menjay Lin', 'Hakan Erdogan', 'Shao-Fu Shih', 'Yang Yang'] | 2023-03-13 | null | null | null | null | ['speech-enhancement'] | ['speech'] | [ 2.79918045e-01 -1.50515556e-01 5.01545429e-01 -7.46907443e-02
-1.06057954e+00 -4.10774767e-01 2.40949959e-01 -2.85300791e-01
-3.97554398e-01 2.88437635e-01 4.06585813e-01 -3.39809150e-01
-2.26680189e-01 -5.87230682e-01 -4.13884640e-01 -1.00786209e+00
3.62409055e-01 -3.85324746e-01 1.29115298e-01 -1.79218456e-01
-2.52337337e-01 2.49734730e-01 -1.55879176e+00 9.77549776e-02
1.09265113e+00 1.03423524e+00 7.06995189e-01 8.69163454e-01
2.75584534e-02 2.62269348e-01 -6.56922400e-01 -4.22897562e-02
2.48026296e-01 -4.56181139e-01 2.12570548e-01 2.35508785e-01
1.53760955e-01 -3.23821783e-01 -2.66897291e-01 1.16528618e+00
1.13047683e+00 2.82720953e-01 2.96825439e-01 -7.42724895e-01
-2.09537055e-02 4.12136734e-01 -4.84876156e-01 1.33264348e-01
1.96109042e-01 1.49969840e-02 7.24736094e-01 -1.25861585e+00
-6.55113533e-02 1.06912971e+00 7.19728172e-01 4.47810471e-01
-9.34998512e-01 -8.04946601e-01 4.99133728e-02 -1.39219806e-01
-1.24212480e+00 -1.08361888e+00 9.68135595e-01 -1.63633734e-01
5.47397554e-01 5.41191518e-01 4.60740030e-01 7.59157538e-01
-3.21198642e-01 7.42763281e-01 9.51587379e-01 -4.58393723e-01
2.38731667e-01 2.33738631e-01 8.36375058e-02 1.86715871e-01
-1.58759430e-01 2.90318251e-01 -5.14922738e-01 -1.65908784e-01
8.64623189e-01 -1.38014466e-01 -8.08685899e-01 7.72309229e-02
-8.09737861e-01 3.32429737e-01 2.72340804e-01 6.07646823e-01
-5.46059847e-01 -1.98289394e-01 -2.93951668e-02 2.54510731e-01
4.28860456e-01 1.69908538e-01 -2.46807069e-01 5.35710528e-02
-1.06182468e+00 3.56215313e-02 6.17040455e-01 7.71822870e-01
3.81647915e-01 2.44282499e-01 -1.67824462e-01 1.34429300e+00
5.25092959e-01 7.40302205e-01 3.17029774e-01 -5.91512620e-01
6.62845612e-01 -7.36456588e-02 1.41389742e-01 -9.98927236e-01
-1.56903654e-01 -9.90380168e-01 -1.10225141e+00 1.83146730e-01
1.75524503e-01 -6.16017640e-01 -5.61911583e-01 1.77290475e+00
4.76934642e-01 4.98404145e-01 2.53875982e-02 1.04048789e+00
7.29293168e-01 1.00307012e+00 -1.97304785e-01 -6.30520165e-01
1.14064717e+00 -9.84929502e-01 -1.06154013e+00 -3.21678936e-01
-4.46734354e-02 -9.35721517e-01 7.86045432e-01 6.85948014e-01
-1.49496412e+00 -8.01190197e-01 -1.13434315e+00 1.98750615e-01
-1.23590222e-02 3.00772697e-01 -4.52696672e-03 7.91061461e-01
-1.25109422e+00 2.51802415e-01 -6.06292605e-01 1.49250716e-01
-1.41285956e-01 4.15964097e-01 -1.48864239e-01 1.10409237e-01
-9.63010192e-01 5.05260468e-01 -1.48970425e-01 5.71064472e-01
-5.45604527e-01 -4.30606604e-01 -6.67439342e-01 2.60063618e-01
3.02230150e-01 -5.98997176e-01 1.23236191e+00 -1.02421451e+00
-1.75797212e+00 1.12111166e-01 -4.53033924e-01 1.11873440e-01
1.63149849e-01 -2.88718820e-01 -7.86702394e-01 5.06974151e-03
-2.40844384e-01 -3.17059867e-02 1.39355433e+00 -1.26758003e+00
-7.44212329e-01 -2.53533006e-01 -2.32914820e-01 4.76692080e-01
-6.97691083e-01 1.78592399e-01 -6.77370489e-01 -1.02371240e+00
4.10055399e-01 -4.79730785e-01 -2.96967745e-01 -1.63793772e-01
-2.20186755e-01 3.08124453e-01 6.42820179e-01 -9.45421219e-01
1.47696733e+00 -2.60849118e+00 3.52991283e-01 2.66239315e-01
1.63696453e-01 6.77586317e-01 -3.17491174e-01 2.56214499e-01
-2.51506954e-01 -2.07198903e-01 -1.34177417e-01 -7.22899139e-01
-1.87031016e-01 -2.70258784e-01 -2.42586821e-01 5.03971100e-01
1.32520925e-02 3.33898246e-01 -7.10756838e-01 -2.51660764e-01
3.25751573e-01 8.41616750e-01 -7.09347785e-01 7.46301115e-01
3.97981584e-01 6.16535306e-01 -3.22458595e-01 2.83229023e-01
9.98901367e-01 2.22057879e-01 7.79894888e-02 -3.14963818e-01
-2.97568738e-01 2.17039645e-01 -1.85517287e+00 1.54293692e+00
-7.88301528e-01 2.99142748e-01 1.27102733e+00 -7.12764919e-01
1.00357604e+00 8.02845418e-01 1.54271722e-01 -4.72526312e-01
5.97926229e-02 4.63149250e-01 3.81275676e-02 -4.92825538e-01
2.57591426e-01 -3.21918964e-01 3.32039267e-01 3.07549745e-01
5.51170036e-02 4.09801900e-02 -3.51015776e-01 -2.40848124e-01
9.95848238e-01 -4.91895765e-01 1.72803849e-01 -1.12823226e-01
8.88066053e-01 -8.21433067e-01 5.55151939e-01 6.39458537e-01
-1.10982661e-03 6.36956215e-01 -3.21997851e-01 4.18806493e-01
-6.65244818e-01 -1.06783319e+00 1.27270564e-01 1.27856350e+00
1.43534407e-01 -3.36432129e-01 -8.43331754e-01 -1.80783570e-01
-4.37040240e-01 2.94975162e-01 1.19999468e-01 3.47504802e-02
-7.79935718e-01 -5.08497775e-01 4.23934281e-01 3.51374805e-01
4.88718778e-01 -8.10659111e-01 -1.51898190e-01 3.26856941e-01
-1.32863536e-01 -7.00940251e-01 -8.43076110e-01 3.64537239e-01
-7.07365036e-01 -6.23274803e-01 -9.39884007e-01 -8.70347738e-01
5.94516575e-01 7.39326894e-01 5.34451306e-01 1.47575006e-01
4.36707556e-01 2.64691859e-01 -2.39623070e-01 -2.95231283e-01
-2.75092483e-01 -1.37309387e-01 1.97464392e-01 5.58960497e-01
-1.65973365e-01 -1.01956236e+00 -6.38075173e-01 3.97995740e-01
-9.19948220e-01 -2.94738561e-02 7.54630923e-01 9.28690493e-01
2.51399815e-01 3.67163867e-01 7.18133092e-01 -2.78634936e-01
9.69363391e-01 -3.50509524e-01 -4.90251780e-01 -4.13973182e-02
-1.13705128e-01 -2.70638704e-01 8.18364024e-01 -4.98854697e-01
-1.47973537e+00 -6.29390255e-02 -7.88966715e-01 -4.33494210e-01
-2.62940645e-01 4.42012191e-01 -8.01880181e-01 -2.91943941e-02
3.54371965e-01 3.13133001e-01 -1.07851401e-01 -9.65012848e-01
2.32781067e-01 1.35107327e+00 6.96200907e-01 -2.31926247e-01
9.26030815e-01 1.00396223e-01 -2.81084627e-01 -1.25318539e+00
-2.18022630e-01 -7.82153845e-01 -2.62656361e-01 -3.49312514e-01
5.34757972e-01 -9.91528094e-01 -5.81393123e-01 7.20703661e-01
-1.28064144e+00 -3.69873531e-02 7.48365894e-02 5.94670713e-01
-8.51535499e-02 2.66985476e-01 -5.69229484e-01 -1.23054826e+00
-4.82214123e-01 -1.27480531e+00 1.07476068e+00 2.60864675e-01
1.05109453e-01 -8.57474387e-01 -3.35018337e-02 5.44515401e-02
7.52614558e-01 -5.03321469e-01 4.60371107e-01 -3.41401339e-01
-5.16762257e-01 -1.02191351e-01 1.29260033e-01 6.28312051e-01
2.10303053e-01 -4.68552023e-01 -1.39884531e+00 -5.70709705e-01
7.86105335e-01 1.93872407e-01 6.84293270e-01 6.17121577e-01
8.99838865e-01 -3.35690767e-01 -1.77121893e-01 7.73570657e-01
1.37070000e+00 4.07592356e-01 6.06115699e-01 -8.11600685e-02
4.59467441e-01 6.46210790e-01 4.25011814e-01 2.82651901e-01
-2.64705926e-01 8.57607245e-01 2.54497051e-01 -4.54379052e-01
-2.60229230e-01 -2.41150349e-01 4.64362919e-01 1.51141798e+00
2.66888708e-01 -5.42667747e-01 -4.71976280e-01 4.59632069e-01
-1.74417663e+00 -7.67737150e-01 -6.84687123e-02 2.30554318e+00
6.97323203e-01 -1.71726719e-01 -3.53474542e-02 4.62413371e-01
9.12692308e-01 2.72526443e-01 -3.57971758e-01 -1.20158583e-01
-1.18163936e-01 4.69513923e-01 1.13851734e-01 8.43392909e-01
-9.77595329e-01 2.48704955e-01 6.02277851e+00 1.14839208e+00
-1.17546797e+00 2.41283581e-01 2.69658566e-01 -1.36443451e-01
-3.00643891e-01 -4.17633832e-01 -5.71523726e-01 3.60087335e-01
7.38526404e-01 1.61579147e-01 5.91128647e-01 5.39147139e-01
6.80735230e-01 7.36581013e-02 -8.65731657e-01 1.30067515e+00
-1.79900289e-01 -8.12428892e-01 -4.02498245e-01 -3.29364650e-02
2.53465712e-01 -3.27646315e-01 2.20960483e-01 2.25981046e-02
-3.19538921e-01 -9.15241599e-01 7.89334118e-01 4.66505200e-01
6.77202463e-01 -5.83834291e-01 4.85895693e-01 8.14604580e-01
-1.42172253e+00 -2.93322504e-01 -4.92987633e-02 -1.56254560e-01
3.80636156e-01 8.40060174e-01 -3.94331127e-01 7.03826427e-01
3.59653145e-01 3.52541625e-01 1.08378768e-01 1.32788467e+00
-2.89326996e-01 6.71191812e-01 -3.85564834e-01 2.63120085e-01
-2.09005222e-01 -2.57450372e-01 1.02125466e+00 1.25598192e+00
6.16220236e-01 3.34519148e-01 8.00419524e-02 6.03940845e-01
1.60352707e-01 1.55143738e-01 -3.94049615e-01 4.43530798e-01
6.98450625e-01 1.28385866e+00 -7.70365968e-02 -1.31155968e-01
-4.69952106e-01 7.29177475e-01 -1.30456790e-01 6.82772279e-01
-7.59696186e-01 -6.30921066e-01 8.37900221e-01 -4.72428370e-03
4.10143614e-01 -1.16797209e-01 -2.55821854e-01 -1.12054110e+00
2.74221003e-01 -1.11227310e+00 -1.67967707e-01 -8.08120847e-01
-1.10730743e+00 8.72525811e-01 -4.00823385e-01 -1.33718419e+00
-1.75565362e-01 -5.68985641e-01 -8.09953809e-01 1.54941881e+00
-1.36173022e+00 -7.89761662e-01 -1.25204474e-01 7.71747649e-01
5.47739446e-01 3.97593901e-02 6.39371157e-01 8.76969934e-01
-7.57464945e-01 6.54790998e-01 2.65263528e-01 -9.00732130e-02
6.68285131e-01 -1.06680954e+00 -6.20241649e-02 1.29004037e+00
-1.68672696e-01 7.69791842e-01 8.12181354e-01 -4.28169519e-01
-1.26939011e+00 -1.03947365e+00 8.29168677e-01 3.12310636e-01
2.99339890e-01 -4.89259839e-01 -9.16847944e-01 1.22775778e-01
3.50016057e-01 -3.04222554e-01 6.20331049e-01 -9.72947627e-02
2.09820986e-01 -4.82830346e-01 -9.07204390e-01 5.08622468e-01
9.22236323e-01 -5.72812974e-01 -5.33312857e-01 -6.23084046e-02
5.10649860e-01 -4.09277558e-01 -4.48659033e-01 2.54517049e-01
3.31771791e-01 -9.63238299e-01 9.62411404e-01 -2.76418198e-02
5.81905209e-02 -6.51769102e-01 -2.30554834e-01 -1.55354047e+00
-3.24402481e-01 -1.02724242e+00 -3.74667570e-02 1.65550721e+00
3.52564186e-01 -5.90870738e-01 4.66396630e-01 3.62855405e-01
-3.92935336e-01 -5.86528540e-01 -7.81144619e-01 -5.38012803e-01
-3.28445882e-01 -6.23086929e-01 5.77204049e-01 5.01752734e-01
4.36600149e-02 5.03600597e-01 -5.80435216e-01 6.41414702e-01
6.79101884e-01 -2.53598839e-01 6.89549863e-01 -8.03723574e-01
-9.13204193e-01 -2.59698778e-01 2.67838746e-01 -1.93847930e+00
-2.23863095e-01 -4.67447609e-01 5.75878024e-01 -1.33900571e+00
-2.21830726e-01 -3.88634294e-01 -2.08148047e-01 -1.68704242e-02
-4.17034537e-01 -1.72672674e-01 2.12909698e-01 3.95774432e-02
-9.16240290e-02 6.89169109e-01 1.06417632e+00 -4.62661944e-02
-5.29324174e-01 4.79447305e-01 -8.37840676e-01 7.86203027e-01
3.91900957e-01 -1.92474157e-01 -5.62152863e-01 -7.21960664e-01
-2.27290004e-01 5.47082067e-01 1.89883590e-01 -1.07100725e+00
7.05521166e-01 2.06316575e-01 2.96599767e-03 -3.76651794e-01
6.47794783e-01 -1.13711250e+00 4.48209614e-01 -5.46136200e-02
-2.26788163e-01 -2.15901434e-01 2.25558832e-01 5.12336910e-01
-4.57124710e-01 -2.21440762e-01 7.72310138e-01 1.13598540e-01
-2.18285710e-01 1.11837022e-01 -4.97422963e-01 -4.52815235e-01
4.80320126e-01 -9.98815224e-02 -5.00410162e-02 -6.53688669e-01
-8.72083068e-01 8.15620646e-02 -2.36123838e-02 1.47503078e-01
6.61046207e-01 -1.27524650e+00 -5.47840714e-01 5.24102747e-01
-5.66517532e-01 -8.16662088e-02 4.85639334e-01 8.59514117e-01
1.55684888e-01 3.29161018e-01 2.74187058e-01 -4.54774082e-01
-1.33425140e+00 4.77739900e-01 6.22220159e-01 -1.54143929e-01
-2.60258496e-01 9.04661000e-01 3.73178482e-01 -1.19519718e-01
5.33054709e-01 -2.15980515e-01 -3.34210038e-01 -1.03703387e-01
8.98850977e-01 4.65913355e-01 2.43543133e-01 -6.34322882e-01
-9.03308988e-02 6.17945850e-01 5.03692448e-01 -5.58344960e-01
1.38164496e+00 -5.03273487e-01 -1.82622701e-01 1.42407984e-01
1.24284136e+00 6.05198741e-01 -1.03624094e+00 -2.80854404e-01
-4.42481995e-01 -5.60514271e-01 4.52216953e-01 -6.99317217e-01
-1.13134575e+00 9.86128151e-01 6.67651117e-01 3.29908967e-01
1.88552344e+00 -3.60195667e-01 8.03070426e-01 1.58256292e-02
2.62794703e-01 -7.71278083e-01 2.45275069e-02 2.43003681e-01
9.78632510e-01 -7.58065999e-01 -5.44296861e-01 -5.62048554e-01
-3.68666053e-02 8.87500584e-01 3.86830300e-01 5.74296676e-02
8.42321575e-01 6.56243086e-01 2.24630117e-01 1.41537517e-01
-3.95636976e-01 -2.50685513e-01 4.64099705e-01 5.68089187e-01
3.84416640e-01 -1.27960145e-01 -2.90246665e-01 1.20933115e+00
2.64160689e-02 -1.27686709e-01 1.84810814e-02 6.53162837e-01
-7.54155219e-01 -1.25616741e+00 -8.65697026e-01 1.83787778e-01
-4.08611894e-01 -3.06125909e-01 -1.64888799e-02 4.00850363e-02
1.73592642e-01 1.54196930e+00 -2.09581763e-01 -6.26769900e-01
7.98389971e-01 -8.07850510e-02 3.47120017e-02 -4.84395862e-01
-8.22703838e-01 9.04630244e-01 5.46574593e-02 -2.69377530e-01
-3.49433422e-01 -5.37503600e-01 -7.44654357e-01 -1.02535285e-01
-7.86622822e-01 3.44760209e-01 5.21939516e-01 8.29910100e-01
3.60550433e-01 7.76759803e-01 8.61261308e-01 -1.06311202e+00
-6.64454699e-01 -1.06430209e+00 -7.68260360e-01 6.89828023e-02
7.76787758e-01 -3.10783505e-01 -4.56403732e-01 -1.21570796e-01] | [15.000247955322266, 5.865296363830566] |
b9dce45a-5f65-48ec-9842-3fcc0146da37 | the-emerging-trends-of-multi-label-learning | 2011.11197 | null | https://arxiv.org/abs/2011.11197v3 | https://arxiv.org/pdf/2011.11197v3.pdf | The Emerging Trends of Multi-Label Learning | Exabytes of data are generated daily by humans, leading to the growing need for new efforts in dealing with the grand challenges for multi-label learning brought by big data. For example, extreme multi-label classification is an active and rapidly growing research area that deals with classification tasks with an extremely large number of classes or labels; utilizing massive data with limited supervision to build a multi-label classification model becomes valuable for practical applications, etc. Besides these, there are tremendous efforts on how to harvest the strong learning capability of deep learning to better capture the label dependencies in multi-label learning, which is the key for deep learning to address real-world classification tasks. However, it is noted that there has been a lack of systemic studies that focus explicitly on analyzing the emerging trends and new challenges of multi-label learning in the era of big data. It is imperative to call for a comprehensive survey to fulfill this mission and delineate future research directions and new applications. | ['Haobo Wang', 'Ivor W. Tsang', 'Xiaobo Shen', 'Weiwei Liu'] | 2020-11-23 | null | null | null | null | ['extreme-multi-label-classification'] | ['methodology'] | [ 1.97485343e-01 -3.30088079e-01 -4.02062178e-01 -7.30975032e-01
-5.81920624e-01 -3.98845792e-01 2.82485336e-01 3.73704016e-01
-3.77020478e-01 4.83533651e-01 -1.41217992e-01 -3.33334580e-02
-1.98788434e-01 -4.78602231e-01 -1.19090326e-01 -7.79315114e-01
2.23933816e-01 5.01163483e-01 -2.79654980e-01 -9.69869271e-02
7.07697943e-02 5.29194772e-01 -1.71353257e+00 5.87884903e-01
4.98193026e-01 1.17370868e+00 -7.24327192e-02 3.38623434e-01
-3.00139546e-01 9.81394768e-01 -3.44199806e-01 -5.22656322e-01
-3.74440402e-02 -2.14472547e-01 -9.65852737e-01 2.51162916e-01
4.75550830e-01 -4.06603441e-02 1.68982521e-01 1.08210945e+00
6.03593230e-01 -1.68202706e-02 5.80776811e-01 -1.28131354e+00
-6.11941993e-01 3.37661386e-01 -7.73924768e-01 6.46215528e-02
-1.71791077e-01 -2.44051397e-01 1.35032225e+00 -7.62033522e-01
2.87299842e-01 1.06168306e+00 6.16578698e-01 5.34368157e-01
-8.01873744e-01 -8.38481605e-01 7.90829137e-02 3.90795350e-01
-1.12293458e+00 -9.15761217e-02 7.32891142e-01 -5.29399633e-01
7.22022295e-01 3.33413780e-01 1.84128121e-01 1.09862733e+00
2.76067436e-01 7.78551340e-01 1.15871549e+00 -5.86037219e-01
-4.57963487e-03 2.35098571e-01 5.15988231e-01 6.64379418e-01
-1.02769338e-01 -7.48586953e-02 -4.45241958e-01 -1.18050247e-01
1.86623096e-01 4.24024194e-01 2.03662768e-01 -2.24397391e-01
-1.01913846e+00 1.07514060e+00 3.42986971e-01 6.87023580e-01
-1.53315797e-01 -7.45825842e-02 6.96477532e-01 6.02973759e-01
8.11107755e-01 4.44919944e-01 -6.92984283e-01 1.44674983e-02
-6.75083756e-01 4.47902409e-03 5.11969626e-01 7.86682785e-01
9.06214833e-01 -5.83080314e-02 2.79123455e-01 1.11843848e+00
2.04563692e-01 8.02233145e-02 5.71263909e-01 -7.25103021e-01
2.44371742e-01 9.45317686e-01 -3.06994379e-01 -7.74878383e-01
-9.94927466e-01 -6.30016148e-01 -1.33038545e+00 7.84208328e-02
2.58768409e-01 -1.42978832e-01 -4.26126033e-01 1.50634313e+00
3.63221854e-01 4.61360998e-02 -1.11240394e-01 5.90558529e-01
8.83506060e-01 4.10031974e-01 4.13201272e-01 -3.48839760e-01
1.41053951e+00 -1.10825896e+00 -7.32396960e-01 -4.82693166e-01
1.21233213e+00 -7.62818813e-01 8.88365448e-01 2.30874449e-01
-4.55360770e-01 -5.57224095e-01 -8.43756974e-01 -1.05874605e-01
-6.16728842e-01 1.91052854e-02 1.01966560e+00 5.17854810e-01
-6.68987632e-01 3.82549942e-01 -3.14871728e-01 -2.91272044e-01
4.74429607e-01 4.45216507e-01 -3.94284278e-01 -4.62532192e-01
-1.42882764e+00 8.52838516e-01 4.30250973e-01 9.86628458e-02
-4.99180436e-01 -4.79750752e-01 -7.89531827e-01 -1.10889869e-02
3.91261697e-01 -2.55433619e-01 1.17321587e+00 -9.40124750e-01
-8.76497269e-01 1.28528607e+00 2.09545195e-01 5.27717061e-02
2.37358212e-01 5.21956682e-02 -6.17287695e-01 -2.88421094e-01
6.44320920e-02 5.52048802e-01 6.15272045e-01 -1.17462742e+00
-1.03017485e+00 -7.16860771e-01 -1.57749280e-01 2.58864224e-01
-8.82251859e-01 4.31592256e-01 8.80611390e-02 -3.35495591e-01
-2.82876659e-02 -1.20280397e+00 -1.68889433e-01 -4.26185519e-01
-1.78492069e-01 -8.73137116e-01 8.40254784e-01 -4.91255671e-02
1.03888249e+00 -2.32854629e+00 -2.00659167e-02 -2.78838754e-01
5.91131330e-01 3.42312068e-01 -2.93387979e-01 5.34863472e-01
-2.55434424e-01 1.57176778e-01 1.74746159e-02 -4.39257860e-01
-3.96197252e-02 8.04647058e-03 -2.61210799e-01 4.08547699e-01
5.75153008e-02 9.72629488e-01 -8.95600438e-01 -4.17857230e-01
1.98898986e-01 4.00888652e-01 9.83854756e-02 2.86512643e-01
-1.32612213e-01 6.34343028e-01 -5.07162988e-01 8.16998959e-01
4.06867325e-01 -9.19226885e-01 2.83464104e-01 -7.10015371e-02
-6.33475855e-02 -3.24653864e-01 -1.06396616e+00 1.26577842e+00
-4.74953324e-01 4.22605991e-01 -7.15218252e-04 -1.38278401e+00
8.56425941e-01 5.22418678e-01 1.01561749e+00 -6.50845349e-01
4.16168600e-01 5.16440392e-01 -8.31159875e-02 -5.52227557e-01
2.30344266e-01 -6.70543373e-01 -3.19502413e-01 7.01953769e-01
6.53026700e-02 1.86236948e-01 2.01873779e-01 -1.37150362e-01
7.37481356e-01 -5.66987276e-01 2.85858691e-01 -6.45156503e-02
4.45066065e-01 -2.77572095e-01 6.28003120e-01 3.21742296e-01
-3.32994401e-01 2.79605150e-01 3.24804217e-01 -1.05752277e+00
-7.60477126e-01 -2.48191357e-01 -3.11995864e-01 1.79508889e+00
-9.48583782e-02 -2.77236670e-01 -2.64807642e-01 -9.09855008e-01
5.29048927e-02 2.36974940e-01 -6.60433114e-01 -1.53825879e-01
-4.17496055e-01 -1.13020909e+00 3.68040502e-01 3.47155809e-01
4.99465615e-01 -1.34929550e+00 -2.41858959e-01 2.16773182e-01
-2.21979931e-01 -1.10599327e+00 -1.07892692e-01 6.99297726e-01
-7.35560536e-01 -1.18606305e+00 -5.61423838e-01 -1.06527019e+00
3.32772076e-01 3.46835732e-01 1.21301460e+00 2.43292391e-01
-2.95817345e-01 5.11246957e-02 -5.80220819e-01 -4.99110967e-01
-3.69956106e-01 5.16525984e-01 2.19881907e-02 2.51684994e-01
8.09365273e-01 -3.89377356e-01 -1.82073340e-01 4.41586882e-01
-1.00047195e+00 1.37296245e-01 5.63779116e-01 7.28455245e-01
5.11458814e-01 3.42562407e-01 1.22209656e+00 -1.26580644e+00
6.83612287e-01 -7.59273231e-01 -3.40231776e-01 4.88192260e-01
-8.41941893e-01 -1.42372698e-01 7.10025847e-01 -2.29388684e-01
-7.91769862e-01 -6.92073703e-02 -3.07964206e-01 2.40959264e-02
-3.87109667e-01 5.43485940e-01 -1.20297551e-01 -1.79838628e-01
4.35537517e-01 -1.40122205e-01 -2.32556134e-01 -4.99120623e-01
4.32545364e-01 1.14123821e+00 -1.17331468e-01 -3.37682039e-01
6.39660656e-02 3.89125794e-01 2.49009207e-01 -5.05292118e-01
-1.66553581e+00 -8.98720324e-01 -9.72571373e-01 -3.24002445e-01
9.72785890e-01 -6.75103247e-01 -7.42784142e-01 8.20404232e-01
-8.50189984e-01 -2.39642605e-01 5.90537153e-02 2.58547127e-01
-3.69006515e-01 3.64402622e-01 -8.50784302e-01 -3.01582485e-01
-4.50184971e-01 -1.30014205e+00 9.76915956e-01 1.76805988e-01
-6.37579784e-02 -1.25250077e+00 1.89009562e-01 8.85340095e-01
3.16161990e-01 3.33261900e-02 1.08521247e+00 -8.63531828e-01
-2.74782002e-01 -4.01944935e-01 -4.99396682e-01 4.02310878e-01
3.89727831e-01 -2.66570389e-01 -1.22135842e+00 -4.80058819e-01
1.83418155e-01 -1.01156878e+00 7.15271950e-01 9.11286920e-02
1.28085589e+00 3.24240386e-01 -2.65064090e-01 3.92722100e-01
1.39615464e+00 6.61612749e-02 -6.07698560e-02 3.42536688e-01
1.01082766e+00 7.65757620e-01 6.09481990e-01 4.16939110e-01
6.17696285e-01 4.99249756e-01 5.69443047e-01 -2.15231627e-01
3.66106723e-03 3.95652562e-01 -3.39222163e-01 1.11478364e+00
1.96750090e-01 -2.76109248e-01 -1.15500140e+00 1.54280007e-01
-1.86292839e+00 -7.85806537e-01 -3.70468527e-01 1.80448675e+00
7.56760061e-01 -1.83344126e-01 -2.00577721e-01 9.94928107e-02
7.69427896e-01 3.25179338e-01 -7.64779150e-01 -1.67999014e-01
-1.65289029e-01 -2.16763034e-01 9.58628878e-02 3.95329185e-02
-1.48608446e+00 8.97211134e-01 6.45463753e+00 7.67100692e-01
-1.36370540e+00 2.79889137e-01 8.07478786e-01 1.72060549e-01
9.02417153e-02 -4.75615352e-01 -1.11156666e+00 3.77875656e-01
1.09790909e+00 1.91778898e-01 7.09752366e-02 8.52226675e-01
-2.36152351e-01 1.06333621e-01 -1.09622872e+00 1.15055358e+00
-1.66244749e-02 -1.15037692e+00 -3.39914501e-01 3.23622644e-01
9.77308154e-01 2.99433768e-01 2.34002560e-01 5.28176248e-01
3.80406082e-01 -9.11155283e-01 9.02217180e-02 7.74393901e-02
8.40878665e-01 -8.60005319e-01 9.44752336e-01 6.57082438e-01
-1.16162753e+00 -4.27081317e-01 -4.33627963e-01 -2.33239025e-01
1.49587825e-01 9.64026213e-01 -3.93964380e-01 3.09566140e-01
4.35937166e-01 9.00061131e-01 -6.27445757e-01 7.87532628e-01
5.14657423e-02 4.88064438e-01 1.20270647e-01 1.54975936e-01
3.73277873e-01 -5.42999953e-02 -4.01220024e-01 1.08318162e+00
-1.55197950e-02 -1.90356612e-01 5.68819404e-01 6.51918501e-02
-4.65345651e-01 4.69097733e-01 -5.98998606e-01 -2.31283084e-01
1.89513370e-01 1.70295799e+00 -7.92795420e-01 -3.05556536e-01
-9.69833434e-01 7.01957405e-01 8.39042127e-01 -1.25117404e-02
-7.24184275e-01 9.62953866e-02 4.51224416e-01 -3.33053946e-01
-2.93991745e-01 -2.75658488e-01 -4.60398108e-01 -1.15274179e+00
-4.23696846e-01 -1.04513347e+00 7.86564708e-01 -3.34534466e-01
-1.72947836e+00 5.21637559e-01 -3.88903737e-01 -9.95895863e-01
-1.48141071e-01 -5.43473363e-01 -1.09101892e-01 6.06272519e-01
-1.70485234e+00 -1.38859296e+00 -3.83987069e-01 5.40972590e-01
6.89346492e-01 -2.80993521e-01 1.11021662e+00 7.57792830e-01
-6.32064581e-01 5.15337467e-01 5.09781241e-01 1.49844944e-01
1.08281600e+00 -1.05065429e+00 -2.16732714e-02 9.76931155e-02
4.40232635e-01 6.71861693e-02 1.11109309e-01 -3.71244907e-01
-9.28278208e-01 -1.14051926e+00 1.17143214e+00 -5.36655068e-01
6.86587930e-01 -2.69735336e-01 -8.70786428e-01 7.32962847e-01
1.22523904e-01 3.73811781e-01 1.59010839e+00 5.56532502e-01
-4.36461687e-01 -2.15284482e-01 -9.04555321e-01 1.28512442e-01
6.52649343e-01 -5.56679547e-01 -1.36716247e-01 7.80833304e-01
5.18537760e-01 -1.20504145e-02 -1.13115656e+00 5.46991646e-01
3.26300591e-01 -8.63940418e-01 6.57523155e-01 -7.12890327e-01
3.45401883e-01 2.17507496e-01 -7.23093003e-02 -1.28028071e+00
-6.24885798e-01 1.23732969e-01 1.03520513e-01 1.14628160e+00
1.99595183e-01 -5.20257354e-01 8.26372445e-01 7.24075794e-01
-9.05991793e-02 -9.55440581e-01 -6.36054814e-01 -5.29186308e-01
2.67671913e-01 -2.52605230e-01 4.21140373e-01 1.57447898e+00
-1.53045744e-01 7.59800255e-01 -7.34166384e-01 -1.88076437e-01
7.51130819e-01 5.17890394e-01 2.93212801e-01 -1.75285053e+00
1.80908278e-01 -3.21013927e-01 -1.60300002e-01 -6.01201952e-01
5.57489038e-01 -1.06917071e+00 -3.11818659e-01 -1.48468804e+00
5.48682630e-01 -7.90104985e-01 -8.22871208e-01 6.29724026e-01
-1.96040615e-01 5.34180701e-01 1.84051991e-01 5.30075312e-01
-1.02989364e+00 4.26061183e-01 1.39905381e+00 -2.27296144e-01
2.38222227e-01 2.95385033e-01 -9.21130776e-01 7.90745080e-01
1.04325223e+00 -7.02431500e-01 -3.29696864e-01 -4.64348167e-01
6.38191283e-01 -8.98191612e-03 -2.49973923e-01 -8.20021272e-01
2.19696850e-01 -5.27817428e-01 1.05031051e-01 -4.26928014e-01
5.60938716e-02 -9.30468380e-01 -4.70835827e-02 1.53611377e-01
-5.54234684e-01 9.96158943e-02 -1.48470491e-01 4.54984009e-01
-4.02851999e-01 -4.67716724e-01 1.01601648e+00 -3.68820101e-01
-1.05060065e+00 5.98036349e-01 -2.24633425e-01 1.71481028e-01
1.25653589e+00 3.25873256e-01 -2.28737429e-01 -1.60454735e-02
-8.43100727e-01 3.53219956e-01 2.11550415e-01 7.04707503e-01
2.20006764e-01 -1.45620537e+00 -6.06359720e-01 1.62646070e-01
3.84625345e-01 6.37345836e-02 4.49415863e-01 5.73653221e-01
-6.53249547e-02 5.20586967e-01 -2.03352883e-01 -3.54227126e-01
-1.46836257e+00 8.38955462e-01 2.24851802e-01 -7.08075225e-01
-3.66427392e-01 9.00467873e-01 1.79619431e-01 -6.78671360e-01
2.33244121e-01 3.11302781e-01 -4.90604192e-01 6.30847275e-01
6.08319521e-01 3.40053201e-01 3.38209182e-01 -6.34765923e-01
-3.40265006e-01 5.62974095e-01 -3.56662393e-01 5.53364098e-01
1.54589438e+00 -9.91563872e-02 -4.20368284e-01 8.00663292e-01
1.48904383e+00 -3.83494020e-01 -9.18453395e-01 -4.01467979e-01
3.72379333e-01 -1.26340821e-01 3.55856568e-02 -8.30389261e-01
-1.24835646e+00 1.13334596e+00 5.60741127e-01 4.90690947e-01
8.75196218e-01 2.92298943e-01 7.78843045e-01 4.57476825e-01
4.96484935e-01 -1.29395878e+00 4.61717933e-01 6.79787695e-01
5.97866178e-01 -1.87046921e+00 -1.97402909e-01 -2.41867498e-01
-5.41723132e-01 1.10965228e+00 7.51796782e-01 4.17124629e-01
9.20442104e-01 1.92886740e-01 4.49984431e-01 -5.16873360e-01
-6.09646380e-01 -1.28613681e-01 3.03154141e-02 2.81736583e-01
7.46260405e-01 2.05602422e-01 -2.48407617e-01 2.30958804e-01
2.86234856e-01 -3.02563719e-02 1.37063488e-01 7.73250461e-01
-6.53618395e-01 -1.53664911e+00 -5.70879802e-02 6.92222178e-01
-5.41251063e-01 -1.37534728e-02 -1.95258826e-01 3.22503388e-01
3.99214059e-01 1.04875040e+00 -1.82043433e-01 -4.58966881e-01
-7.80655164e-03 4.50606346e-01 1.68447629e-01 -9.14758742e-01
-6.15235150e-01 -1.77195683e-01 -1.35389268e-01 -1.87520325e-01
-5.10568023e-01 -3.78315568e-01 -1.17146242e+00 -3.08761001e-01
-4.34717864e-01 9.59514081e-02 8.39462578e-01 1.09825730e+00
3.01490247e-01 4.51960623e-01 8.92240226e-01 -4.87227678e-01
-8.10010910e-01 -1.01997972e+00 -9.07376647e-01 6.72801852e-01
1.52020216e-01 -6.25432491e-01 -4.51092660e-01 -1.25216648e-01] | [9.614852905273438, 4.281433582305908] |
4ad8e4c1-d8c2-4dba-a888-8f7343995bc8 | uatvr-uncertainty-adaptive-text-video | 2301.06309 | null | https://arxiv.org/abs/2301.06309v1 | https://arxiv.org/pdf/2301.06309v1.pdf | UATVR: Uncertainty-Adaptive Text-Video Retrieval | With the explosive growth of web videos and emerging large-scale vision-language pre-training models, e.g., CLIP, retrieving videos of interest with text instructions has attracted increasing attention. A common practice is to transfer text-video pairs to the same embedding space and craft cross-modal interactions with certain entities in specific granularities for semantic correspondence. Unfortunately, the intrinsic uncertainties of optimal entity combinations in appropriate granularities for cross-modal queries are understudied, which is especially critical for modalities with hierarchical semantics, e.g., video, text, etc. In this paper, we propose an Uncertainty-Adaptive Text-Video Retrieval approach, termed UATVR, which models each look-up as a distribution matching procedure. Concretely, we add additional learnable tokens in the encoders to adaptively aggregate multi-grained semantics for flexible high-level reasoning. In the refined embedding space, we represent text-video pairs as probabilistic distributions where prototypes are sampled for matching evaluation. Comprehensive experiments on four benchmarks justify the superiority of our UATVR, which achieves new state-of-the-art results on MSR-VTT (50.8%), VATEX (64.5%), MSVD (49.7%), and DiDeMo (45.8%). The code is available in supplementary materials and will be released publicly soon. | ['Wanli Ouyang', 'Xiangyang Ji', 'Weiping Wang', 'Fu Li', 'Yuxin Song', 'Min Yang', 'Yu Zhou', 'Chang Liu', 'Wenhao Wu', 'Bo Fang'] | 2023-01-16 | null | null | null | null | ['video-retrieval'] | ['computer-vision'] | [-1.39698476e-01 -3.85575324e-01 -4.65895593e-01 -4.40754503e-01
-1.10621870e+00 -5.21149635e-01 7.20144391e-01 1.22422032e-01
-4.77904290e-01 4.63823825e-01 5.18926144e-01 -2.85496525e-02
-3.25520128e-01 -5.32405853e-01 -1.00697219e+00 -4.17422146e-01
9.95378494e-02 6.78758800e-01 2.66499341e-01 -4.36395966e-03
1.46973684e-01 3.54997404e-02 -1.74638844e+00 7.63542175e-01
6.90550804e-01 1.49477863e+00 4.17422861e-01 4.35347110e-01
-3.46528381e-01 8.07928562e-01 -2.60669023e-01 -8.67907286e-01
-6.67203590e-02 6.36190102e-02 -7.61757731e-01 4.41039279e-02
7.50782430e-01 -7.23721385e-01 -9.34837937e-01 1.15887523e+00
3.59217763e-01 4.04733121e-01 8.68930697e-01 -1.42769647e+00
-9.55857575e-01 9.03643787e-01 -4.75933075e-01 1.11013345e-01
5.29427052e-01 1.00818068e-01 1.35952771e+00 -1.29303598e+00
7.86665499e-01 1.60499835e+00 9.56474915e-02 5.17753303e-01
-9.41607356e-01 -5.96712768e-01 1.88864172e-01 8.05776119e-01
-1.58068597e+00 -4.29371119e-01 4.23064172e-01 -3.84095162e-01
9.60900843e-01 1.66017070e-01 5.50363541e-01 1.39731348e+00
2.00306904e-02 1.49284565e+00 6.08688831e-01 1.36488248e-02
1.18776992e-01 6.56636208e-02 -9.00568627e-03 6.44403338e-01
1.10455543e-01 -6.98753595e-02 -9.69657898e-01 -4.62966412e-02
5.85126817e-01 3.78095120e-01 -2.64501303e-01 -5.66113889e-01
-1.43421030e+00 6.79121435e-01 2.86026359e-01 6.13347180e-02
-3.25571150e-01 3.76280010e-01 7.09982216e-01 1.98818594e-01
1.05094373e-01 1.41555354e-01 -2.75143445e-01 -4.21757758e-01
-9.27610338e-01 3.18618566e-01 5.43636739e-01 1.51022673e+00
6.31993473e-01 -2.77542263e-01 -6.32073224e-01 8.38233769e-01
4.80046690e-01 6.41084611e-01 3.88519943e-01 -9.94376481e-01
7.96519816e-01 3.97647649e-01 2.62152646e-02 -9.45176721e-01
3.15403372e-01 -1.38188421e-03 -7.21390367e-01 -6.38930440e-01
5.30108251e-02 3.36942315e-01 -7.57089913e-01 1.50452507e+00
2.21590325e-01 3.11070025e-01 1.78893790e-01 1.09315014e+00
8.84674370e-01 8.77240419e-01 1.90830812e-01 5.93873560e-02
1.49923325e+00 -9.40913856e-01 -7.76312411e-01 2.55684871e-02
3.59804213e-01 -6.96789086e-01 1.26651442e+00 2.48395070e-01
-1.09312582e+00 -4.63179201e-01 -6.67730987e-01 -3.95693779e-01
-3.78197551e-01 1.35221839e-01 4.81724054e-01 4.47038896e-02
-9.05893683e-01 4.52795833e-01 -7.74107456e-01 -2.56635189e-01
5.64742863e-01 8.65566507e-02 -2.74560332e-01 -5.27848780e-01
-1.57699931e+00 4.66608584e-01 6.22399867e-01 -8.86700526e-02
-1.19978714e+00 -7.97882199e-01 -9.62537587e-01 2.02183530e-01
8.91095400e-01 -6.56541765e-01 1.07214355e+00 -4.59865868e-01
-1.15074980e+00 7.53281176e-01 -1.40269861e-01 -3.76592368e-01
5.48505187e-01 -5.55586934e-01 -4.52985018e-01 4.70214695e-01
1.16000347e-01 8.97010505e-01 1.09582043e+00 -1.10776508e+00
-8.12757313e-01 -3.19082052e-01 3.95362705e-01 4.10013735e-01
-6.20928645e-01 3.43915191e-03 -1.07143533e+00 -7.86454797e-01
-8.42187181e-02 -7.60924339e-01 3.77901971e-01 2.63885647e-01
-2.79120654e-01 -6.26198709e-01 8.02477539e-01 -6.55725896e-01
1.35088634e+00 -2.26736069e+00 5.85610092e-01 -1.09738909e-01
2.48044610e-01 -1.04235709e-02 -6.35806620e-02 5.01097143e-01
3.24578434e-01 3.79263535e-02 6.03168346e-02 -4.56976473e-01
6.01001978e-01 1.53120190e-01 -6.65118635e-01 1.90361157e-01
1.01189494e-01 1.07099950e+00 -1.05999124e+00 -8.90605748e-01
3.61969411e-01 5.07763743e-01 -6.72392964e-01 2.88700491e-01
-4.86729711e-01 -1.31003082e-01 -7.72969663e-01 9.10316050e-01
3.63997579e-01 -4.24879700e-01 -5.53487204e-02 -8.20360899e-01
3.51149589e-01 -2.84374636e-02 -1.06815195e+00 2.15470147e+00
-3.31542253e-01 7.50383794e-01 -3.07918966e-01 -7.71709621e-01
4.05904561e-01 1.87900960e-01 3.69678497e-01 -5.46462893e-01
1.03281066e-01 1.02421694e-01 -5.10783792e-01 -6.99974954e-01
9.82672393e-01 3.01214248e-01 -3.48031610e-01 1.17183723e-01
3.08448017e-01 -2.68009417e-02 1.90988123e-01 7.30054021e-01
8.45189154e-01 2.02811658e-01 -2.41915472e-02 3.01245321e-02
3.37528408e-01 -1.48104846e-01 2.01304078e-01 7.47442484e-01
-8.41924697e-02 4.31318909e-01 3.64984185e-01 -3.97468656e-02
-8.24756324e-01 -1.26090789e+00 -1.54922009e-01 1.07176828e+00
6.54841900e-01 -7.10828602e-01 -5.21205485e-01 -6.21385753e-01
2.89467931e-01 7.90260971e-01 -4.30638492e-01 -3.88376892e-01
-2.49539062e-01 -6.90277591e-02 4.27268475e-01 6.75021589e-01
5.00646234e-01 -8.28589737e-01 -3.43655139e-01 -7.16625666e-03
-5.32997906e-01 -1.43984616e+00 -8.86855364e-01 -2.36353815e-01
-6.20490551e-01 -8.03667128e-01 -7.53344655e-01 -5.27134955e-01
3.25346410e-01 3.85554045e-01 1.22028315e+00 -1.60766929e-01
-2.08287627e-01 9.11691487e-01 -6.66738033e-01 9.11938697e-02
-1.05280496e-01 -2.50162035e-01 1.27922684e-01 6.85517639e-02
5.06906033e-01 -4.80378307e-02 -7.05362678e-01 3.60811412e-01
-1.25920010e+00 -1.58873759e-02 4.40350473e-01 1.04893839e+00
8.88888180e-01 3.66713218e-02 2.10800141e-01 -4.99546111e-01
4.20135826e-01 -8.19118559e-01 -5.00194073e-01 7.03615904e-01
-3.22989166e-01 2.17913374e-01 4.67720002e-01 -6.28866673e-01
-1.03060973e+00 -4.35588330e-01 2.37835780e-01 -1.35373974e+00
3.41173522e-02 6.68438494e-01 -3.42474073e-01 3.67818862e-01
1.29606888e-01 3.30212265e-01 -3.92603755e-01 -1.95433199e-01
5.83529413e-01 7.57604539e-01 5.16414464e-01 -1.10789871e+00
5.82673848e-01 2.77867198e-01 -3.48218977e-01 -5.56154549e-01
-7.85518169e-01 -6.03721559e-01 -5.95034957e-02 -3.15712959e-01
9.15457129e-01 -1.21576023e+00 -7.12383866e-01 4.24959689e-01
-1.06264150e+00 -1.74898565e-01 -1.96780905e-01 6.55519068e-01
-7.02966511e-01 4.80942696e-01 -6.44778192e-01 -4.79597747e-01
-7.82640502e-02 -1.42431986e+00 1.51986897e+00 1.88516244e-01
8.27477425e-02 -8.32433343e-01 -4.34205443e-01 6.64848030e-01
2.21068859e-01 -4.11167145e-01 9.30826008e-01 -6.20146573e-01
-1.08125436e+00 -2.37985812e-02 -3.99771571e-01 2.30674118e-01
-6.98492751e-02 3.10960915e-02 -7.27891803e-01 -3.28448027e-01
-3.69919747e-01 -8.44628751e-01 9.74849522e-01 2.99051166e-01
1.80174243e+00 -8.09515417e-02 -4.78318274e-01 4.22233641e-01
1.28264904e+00 4.46464866e-02 5.95561087e-01 7.45204017e-02
7.30245292e-01 5.40367663e-01 1.03943586e+00 6.90669060e-01
6.70594335e-01 7.27488279e-01 4.74308848e-01 7.36338317e-01
-1.20822094e-01 -5.77658176e-01 4.74546731e-01 8.80124629e-01
2.19919369e-01 -6.54536784e-01 -6.66158080e-01 5.97675681e-01
-2.07377481e+00 -1.08272958e+00 5.13898790e-01 2.01176047e+00
1.05220771e+00 9.73124814e-04 -3.02301496e-01 -4.07087654e-01
7.25497305e-01 3.28018337e-01 -8.36903870e-01 2.76951969e-01
-1.05174199e-01 -1.78864762e-01 2.67041236e-01 2.36568794e-01
-1.04223216e+00 1.08310556e+00 4.68602133e+00 1.50490010e+00
-6.62442207e-01 5.72322384e-02 5.35537958e-01 -2.72474319e-01
-6.16491079e-01 -1.55143440e-01 -9.86048400e-01 5.78320146e-01
6.00793600e-01 -2.37162024e-01 6.51984215e-01 7.24610865e-01
-2.54007608e-01 3.44742201e-02 -1.44115222e+00 1.44563234e+00
1.73295140e-01 -1.57021570e+00 5.21286726e-01 -2.30731010e-01
5.60499668e-01 -1.07561968e-01 3.01306844e-01 7.15718508e-01
2.55934715e-01 -6.92357123e-01 9.41906810e-01 7.34294832e-01
1.02513278e+00 -6.12078428e-01 4.12827671e-01 -7.83290248e-03
-1.31299567e+00 -1.23177059e-02 -4.59115714e-01 8.03623438e-01
2.02940762e-01 3.65028799e-01 -3.60237926e-01 7.13046134e-01
1.05743039e+00 1.05136168e+00 -2.51669735e-01 7.20108330e-01
9.21796113e-02 3.07615280e-01 -2.57324964e-01 -1.27412438e-01
2.42454156e-01 6.31183907e-02 6.22778773e-01 1.08460343e+00
4.46716517e-01 4.53057289e-02 9.19039845e-02 8.60153973e-01
-5.11363983e-01 -1.09437823e-01 -3.89184892e-01 -4.51190740e-01
7.82598794e-01 1.03846335e+00 -3.00580889e-01 -5.78929365e-01
-5.34120917e-01 1.10120344e+00 3.33389282e-01 5.75895369e-01
-1.14699161e+00 -1.73984990e-01 8.67929041e-01 -1.84794039e-01
4.77136463e-01 -5.64080514e-02 4.73341703e-01 -1.61972833e+00
2.80621380e-01 -1.00603747e+00 6.34604394e-01 -1.06382692e+00
-1.66802096e+00 4.13223386e-01 5.54276347e-01 -1.40673542e+00
-2.20674276e-01 -6.61847413e-01 1.83269512e-02 4.34696734e-01
-1.46034050e+00 -9.79212999e-01 -3.24915528e-01 8.77580523e-01
8.60704601e-01 -3.80074531e-01 4.40967798e-01 4.91520524e-01
-2.45714143e-01 8.39744031e-01 1.76166460e-01 5.83080575e-02
8.58964145e-01 -1.04027653e+00 2.13713981e-02 4.61314023e-01
2.89078236e-01 6.83165073e-01 6.30362511e-01 -6.53207719e-01
-2.05575418e+00 -1.19527948e+00 6.33612514e-01 -3.84100974e-01
1.07769334e+00 -2.65328228e-01 -9.69880223e-01 6.89514399e-01
2.92221874e-01 2.10763440e-01 4.57642466e-01 -1.46881640e-01
-6.46433592e-01 -2.65518337e-01 -9.62402105e-01 9.74620938e-01
1.21107590e+00 -1.00818670e+00 -6.86169863e-01 4.11670595e-01
1.18292797e+00 -7.36946344e-01 -1.19839060e+00 2.88818687e-01
5.46310604e-01 -6.60081506e-01 1.15090048e+00 -6.72912896e-01
6.06297016e-01 -2.54344940e-01 -8.36890221e-01 -1.08431005e+00
4.55858484e-02 -4.07518327e-01 -6.72588348e-01 1.25890648e+00
3.76873091e-02 -3.61464769e-01 5.62874973e-01 7.45672762e-01
-1.32327259e-01 -8.26616228e-01 -9.69916165e-01 -7.11194277e-01
-2.80775428e-01 -4.87549305e-01 6.28657222e-01 8.04052174e-01
-4.25320677e-02 8.72773752e-02 -3.97146821e-01 2.25996986e-01
8.08830142e-01 2.27683321e-01 3.71427149e-01 -7.61791945e-01
-4.25066411e-01 -3.30506951e-01 -5.60424864e-01 -1.50940001e+00
5.73530316e-01 -8.79463613e-01 -4.59645130e-02 -1.32089913e+00
4.97716963e-01 -2.53512591e-01 -4.60497409e-01 2.54180193e-01
-2.09348798e-01 -1.88302368e-01 2.14709833e-01 2.72191018e-01
-1.12581909e+00 1.07183278e+00 1.28827333e+00 -6.24888361e-01
2.42375895e-01 -4.90942061e-01 -3.72444987e-01 3.10524225e-01
4.35705066e-01 -2.84074873e-01 -8.28982592e-01 -7.81872988e-01
3.60444725e-01 3.34308267e-01 5.73077857e-01 -7.75442779e-01
4.63253587e-01 -2.82493681e-01 1.73328295e-01 -8.18478346e-01
7.34179318e-01 -1.05520511e+00 1.00518085e-01 -6.34551719e-02
-7.44033694e-01 2.73425095e-02 1.57307625e-01 9.91074443e-01
-5.58430851e-01 -1.68826506e-01 2.09575117e-01 1.94634721e-01
-1.23888493e+00 8.53854358e-01 -5.72707839e-02 3.31931502e-01
9.40321088e-01 -1.09442778e-01 -4.40518051e-01 -4.09759581e-01
-5.30378520e-01 6.65889680e-01 4.17697340e-01 7.89556146e-01
1.02735817e+00 -1.64462340e+00 -5.29119790e-01 -1.78550094e-01
6.32737637e-01 1.32545665e-01 6.87197387e-01 6.31913900e-01
5.24157099e-02 4.58126009e-01 3.85634713e-02 -8.46537709e-01
-1.10564113e+00 6.54435873e-01 6.55583069e-02 5.93300126e-02
-4.47984636e-01 8.33616674e-01 2.90333837e-01 5.28271087e-02
5.47082186e-01 -3.76439661e-01 -2.49026064e-02 1.49130315e-01
5.56964159e-01 2.95854777e-01 -2.35118076e-01 -5.76531589e-01
-3.07593048e-01 4.03576463e-01 -4.42424119e-01 4.10157181e-02
9.46853936e-01 -3.06123227e-01 -2.32165586e-02 5.37062347e-01
1.49548316e+00 -3.92727166e-01 -1.36104190e+00 -6.99900806e-01
-1.73542663e-01 -7.35293627e-01 2.10293785e-01 -4.99696434e-01
-1.10891354e+00 8.35536540e-01 4.42342252e-01 -1.45168945e-01
1.00163257e+00 4.25057441e-01 9.11744893e-01 7.37476349e-01
5.66779017e-01 -1.15404880e+00 3.99623454e-01 4.56972361e-01
9.19768095e-01 -1.29677343e+00 -1.52294740e-01 -6.30390570e-02
-9.44278777e-01 9.99547541e-01 6.61271513e-01 2.55239099e-01
5.93960524e-01 -3.91340964e-02 -5.65414131e-01 -2.62661755e-01
-1.14086092e+00 -1.22833671e-02 5.07894337e-01 2.50294983e-01
2.04072937e-01 1.01379536e-01 1.82534471e-01 5.98389328e-01
3.50222260e-01 3.25890817e-02 1.52978852e-01 7.68898666e-01
-2.90282816e-01 -7.87907839e-01 -6.24279678e-03 6.56759322e-01
-1.84819371e-01 -3.52146327e-01 1.11516170e-01 6.05522573e-01
-2.79364198e-01 7.18590319e-01 2.12299302e-01 -5.05447090e-01
2.26883799e-01 2.78613940e-02 4.86835599e-01 -3.60565782e-01
-7.56561756e-02 -3.85595858e-02 1.59661118e-02 -7.90468276e-01
-4.62713122e-01 -8.35288465e-01 -1.32959294e+00 -3.78672183e-01
-1.88221768e-01 -1.27816889e-02 3.94577742e-01 7.67098665e-01
5.91053307e-01 3.99576515e-01 2.93349087e-01 -6.20727420e-01
-9.09831643e-01 -6.44381225e-01 -5.38979888e-01 4.71836954e-01
1.71027288e-01 -9.83868718e-01 -2.48075396e-01 1.08334586e-01] | [10.334847450256348, 0.9442437887191772] |
68984598-3cf5-474f-9b09-3d392674a249 | on-the-equivalence-between-graph-isomorphism | 1905.12560 | null | https://arxiv.org/abs/1905.12560v2 | https://arxiv.org/pdf/1905.12560v2.pdf | On the equivalence between graph isomorphism testing and function approximation with GNNs | Graph Neural Networks (GNNs) have achieved much success on graph-structured data. In light of this, there have been increasing interests in studying their expressive power. One line of work studies the capability of GNNs to approximate permutation-invariant functions on graphs, and another focuses on the their power as tests for graph isomorphism. Our work connects these two perspectives and proves their equivalence. We further develop a framework of the expressive power of GNNs that incorporates both of these viewpoints using the language of sigma-algebra, through which we compare the expressive power of different types of GNNs together with other graph isomorphism tests. In particular, we prove that the second-order Invariant Graph Network fails to distinguish non-isomorphic regular graphs with the same degree. Then, we extend it to a new architecture, Ring-GNN, which succeeds in distinguishing these graphs and achieves good performances on real-world datasets. | ['Lei Chen', 'Joan Bruna', 'Zhengdao Chen', 'Soledad Villar'] | 2019-05-29 | on-the-equivalence-between-graph-isomorphism-1 | http://papers.nips.cc/paper/9718-on-the-equivalence-between-graph-isomorphism-testing-and-function-approximation-with-gnns | http://papers.nips.cc/paper/9718-on-the-equivalence-between-graph-isomorphism-testing-and-function-approximation-with-gnns.pdf | neurips-2019-12 | ['graph-regression'] | ['graphs'] | [ 1.43009081e-01 3.00307781e-01 -2.48844609e-01 -1.01206824e-01
1.51050732e-01 -7.12989748e-01 6.39608622e-01 2.39469796e-01
-6.07423671e-02 4.62938070e-01 6.73909113e-02 -5.88431120e-01
-5.48083842e-01 -1.50612533e+00 -7.25568831e-01 -5.11938214e-01
-6.81368709e-01 6.45671785e-01 2.96976417e-01 -5.39422512e-01
-9.41362679e-02 6.66318357e-01 -1.53594160e+00 -2.07165945e-02
7.40375996e-01 7.63363540e-01 -4.50906187e-01 5.91458440e-01
-2.94015203e-02 7.44386196e-01 -2.21355602e-01 -5.94335198e-01
4.75667477e-01 -4.21002150e-01 -7.53778577e-01 -2.98177749e-01
6.22264028e-01 6.33962676e-02 -8.77645314e-01 1.58400285e+00
2.39897087e-01 1.05949454e-01 4.76320118e-01 -1.76352262e+00
-8.38545263e-01 1.03826237e+00 -6.74183369e-02 5.70001192e-02
2.70814419e-01 -9.28610787e-02 1.50719094e+00 -9.56772715e-02
4.50246364e-01 1.31230068e+00 1.00239515e+00 3.41949791e-01
-1.29718864e+00 -4.38732237e-01 -2.51930475e-01 2.27387309e-01
-1.25094688e+00 6.07404709e-02 6.76262379e-01 -2.52429634e-01
8.43190551e-01 5.01496255e-01 7.52328873e-01 9.11227047e-01
1.94484368e-01 3.98971438e-01 1.25163758e+00 -4.97993290e-01
1.55306578e-01 -3.64926696e-01 4.48731571e-01 1.07172918e+00
6.06247365e-01 2.83539683e-01 -4.09209691e-02 3.73093523e-02
7.67222166e-01 -1.11775301e-01 -4.00827765e-01 -5.90585947e-01
-9.33344722e-01 9.30178702e-01 6.84729695e-01 9.29980040e-01
1.66541681e-01 2.45670021e-01 5.60840487e-01 7.50141680e-01
8.45043659e-02 7.35453069e-01 1.47348838e-02 3.98341447e-01
-4.92282450e-01 -1.11442097e-01 1.20469105e+00 9.36096847e-01
7.45873153e-01 7.68368095e-02 -1.89862192e-01 3.00392538e-01
-1.22525007e-01 3.62425804e-01 3.09332103e-01 -5.98734915e-01
1.05119802e-01 8.68639708e-01 -6.59225702e-01 -1.44745898e+00
-7.12395489e-01 -6.53637826e-01 -1.39759111e+00 -1.93851843e-01
5.12621045e-01 4.38591897e-01 -6.43853843e-01 2.06408453e+00
-1.71205908e-01 2.97860831e-01 -1.65727526e-01 5.06619751e-01
6.29743457e-01 3.14896673e-01 -1.73896953e-01 2.82720681e-02
1.09689605e+00 -6.94729209e-01 -3.99808824e-01 1.49087816e-01
8.29472303e-01 6.68319017e-02 9.89211023e-01 1.28905609e-01
-1.03895986e+00 -4.94824260e-01 -1.26602209e+00 1.16194777e-01
-6.78657413e-01 -4.21660215e-01 1.01219654e+00 7.85488665e-01
-1.58679330e+00 1.14192164e+00 -5.03064215e-01 -3.79635453e-01
1.48639828e-01 5.84191263e-01 -5.95665395e-01 -1.02222346e-01
-1.54484975e+00 6.13278747e-01 7.17443049e-01 3.41050401e-02
-4.34481859e-01 -4.18756127e-01 -1.12035120e+00 4.99280959e-01
4.68954235e-01 -7.17137754e-01 8.05291593e-01 -9.94331777e-01
-1.03597808e+00 9.54053879e-01 4.93889570e-01 -8.32509577e-01
4.75608528e-01 5.39331794e-01 -7.92015493e-01 3.74932773e-02
-1.07174046e-01 1.17940463e-01 2.80258417e-01 -5.79346180e-01
-2.97560543e-01 -5.36613524e-01 5.72253823e-01 -4.27193612e-01
-6.04033351e-01 -1.64398283e-01 -1.56228721e-01 -3.88699740e-01
-1.60841215e-02 -1.00381887e+00 -1.52685642e-01 -4.24279392e-01
-6.81008816e-01 -6.27435982e-01 4.85192686e-01 5.55691049e-02
1.34576917e+00 -2.05311942e+00 1.40571922e-01 8.82696867e-01
8.98003995e-01 3.38274539e-01 -4.24735427e-01 5.97814202e-01
-4.22403902e-01 3.46476406e-01 -1.28987387e-01 3.88860524e-01
6.04408503e-01 3.67657244e-01 -2.60603517e-01 6.02267504e-01
7.16299415e-02 1.31293499e+00 -8.80563736e-01 -2.48843327e-01
1.09325005e-02 9.79456827e-02 -5.21318793e-01 -1.23320065e-01
-2.61950433e-01 -4.66448814e-02 -2.92322963e-01 3.84672970e-01
7.30961978e-01 -6.15859926e-01 8.22041094e-01 -6.88507408e-02
1.49382874e-01 2.14759447e-02 -1.14076531e+00 9.81569290e-01
-7.43348077e-02 6.02904081e-01 -2.80049831e-01 -1.30972755e+00
8.39222074e-01 -4.79511358e-02 2.35162079e-01 -8.64035547e-01
3.16997617e-01 3.10726017e-01 3.95076990e-01 -1.21021606e-01
2.64121681e-01 1.28581664e-02 -1.42911196e-01 4.39319193e-01
1.73865810e-01 2.89478272e-01 7.43277609e-01 1.78878576e-01
1.54755986e+00 -3.61887634e-01 4.39065188e-01 -5.63690007e-01
8.98963690e-01 -5.36490738e-01 2.03793049e-01 1.06567371e+00
3.25085185e-02 -4.55160663e-02 1.08265877e+00 -7.79526591e-01
-1.12290597e+00 -1.33617723e+00 2.18081549e-01 1.11032200e+00
9.26489234e-02 -6.98395848e-01 -7.02960551e-01 -7.05825150e-01
1.54442951e-01 3.03294063e-01 -8.57880831e-01 -4.53483999e-01
-5.98596990e-01 -4.92174685e-01 1.05010164e+00 6.67446017e-01
6.55756354e-01 -8.49549353e-01 -2.47104559e-02 -1.77292794e-01
1.74785629e-01 -1.06253386e+00 -2.60590345e-01 2.73778379e-01
-7.64350474e-01 -1.46446776e+00 -2.67652750e-01 -9.44024563e-01
6.41275883e-01 -6.80353343e-02 1.46266627e+00 3.96857202e-01
2.09808037e-01 4.65708733e-01 -1.45438686e-01 -3.51218693e-02
-7.92878270e-01 2.83693850e-01 2.13996261e-01 -1.14695262e-02
3.74871880e-01 -1.07394588e+00 -1.40748441e-03 4.88134861e-01
-1.23144889e+00 -2.39006713e-01 5.22815049e-01 5.27427018e-01
4.02097553e-01 2.79547304e-01 2.75816500e-01 -9.42842841e-01
8.87071908e-01 -3.70841146e-01 -1.01197326e+00 3.80278170e-01
-6.92900181e-01 6.21798694e-01 1.17911172e+00 -2.37232789e-01
-7.73196816e-02 -4.18697149e-01 -4.51959707e-02 -3.70395780e-01
1.73648149e-01 7.01105952e-01 -2.67741472e-01 -4.55964595e-01
5.73411226e-01 1.83665723e-01 -1.17881909e-01 -5.26808426e-02
2.02187270e-01 8.87792632e-02 7.22765267e-01 -5.64713955e-01
1.05490530e+00 3.48702729e-01 9.38799024e-01 -7.01615632e-01
-5.59886158e-01 -4.59311232e-02 -4.19374108e-01 -8.90007894e-03
4.95983124e-01 -2.37591714e-01 -1.15260804e+00 3.42495412e-01
-7.92155504e-01 -2.85578549e-01 -3.26971173e-01 4.01098579e-01
-5.72435498e-01 9.08529401e-01 -7.16555059e-01 -5.05235732e-01
-2.70740122e-01 -9.33005452e-01 5.27070582e-01 -1.22187123e-01
3.58905457e-02 -1.19627905e+00 2.27385178e-01 -3.78866017e-01
4.46338981e-01 4.72747892e-01 1.49571145e+00 -1.27339828e+00
-6.63071334e-01 -1.85667098e-01 -3.55797648e-01 3.18949342e-01
-3.08874875e-01 -1.89527571e-01 -5.87283731e-01 -5.10824263e-01
-1.82055697e-01 -1.59472480e-01 8.96428943e-01 2.18972638e-01
1.12042928e+00 -3.60892296e-01 -2.60509342e-01 8.24158967e-01
1.59136260e+00 -1.03225067e-01 7.71698833e-01 2.03257963e-01
6.49949491e-01 3.01914811e-01 -2.68062353e-01 -1.24171980e-01
1.34834588e-01 5.03865778e-01 6.27150774e-01 1.47436365e-01
1.20027989e-01 -5.11602938e-01 1.75869405e-01 1.17548978e+00
-3.67115766e-01 -4.95633900e-01 -7.82587767e-01 1.04608431e-01
-1.65988290e+00 -9.49432194e-01 -2.67091334e-01 2.05023551e+00
2.22537890e-01 1.01380885e-01 2.77408063e-01 3.11911076e-01
1.03690934e+00 3.88265043e-01 -3.74705553e-01 -6.01722360e-01
-4.23083723e-01 5.60792267e-01 5.55698097e-01 2.52663970e-01
-1.01641214e+00 6.04746580e-01 6.53419590e+00 7.86338568e-01
-9.15004492e-01 -2.47430116e-01 4.18256074e-01 6.46054387e-01
-6.05780661e-01 -1.59165030e-03 -1.58184975e-01 2.92068511e-01
1.01433206e+00 -3.50748837e-01 6.87985957e-01 7.91250706e-01
-6.54314160e-01 6.05268478e-01 -1.58398139e+00 8.54431808e-01
-9.32142138e-03 -1.30527973e+00 1.90450728e-01 3.52083385e-01
5.01675367e-01 5.44763990e-02 -3.04244161e-02 5.75110316e-01
5.59633493e-01 -1.38519466e+00 2.61165679e-01 3.94870013e-01
8.11470091e-01 -9.51708257e-01 8.55781317e-01 2.17052802e-01
-1.35177231e+00 -1.65074721e-01 -4.64082122e-01 -1.94475368e-01
-4.26579148e-01 3.45362931e-01 -6.44199789e-01 9.48610723e-01
3.35390419e-01 4.34385419e-01 -7.01581776e-01 1.00263453e+00
-3.12262177e-01 3.46533775e-01 -1.93452924e-01 -2.47342169e-01
3.36657226e-01 -3.74688447e-01 5.14399707e-01 9.23145354e-01
3.19066972e-01 -2.26441652e-01 2.35331044e-01 9.85060573e-01
-4.80544955e-01 4.67248484e-02 -1.10400152e+00 -3.46566826e-01
9.36681852e-02 1.16739810e+00 -8.20339322e-01 -1.20864980e-01
-2.40169615e-01 5.87144375e-01 6.64200664e-01 -3.40454765e-02
-7.66471684e-01 -5.20889521e-01 3.14620763e-01 9.19312164e-02
2.48692885e-01 -2.07623661e-01 3.52452368e-01 -1.21929181e+00
-7.23151416e-02 -1.09141004e+00 7.95661867e-01 -5.95879674e-01
-1.44887972e+00 7.95737803e-01 -1.97666511e-02 -9.14474666e-01
-3.63826871e-01 -1.10949337e+00 -5.70432246e-01 4.67642635e-01
-1.15628171e+00 -9.77522135e-01 -3.13472956e-01 7.77258992e-01
-5.83957732e-01 -1.01813011e-01 9.35644209e-01 2.94778943e-01
-1.17282979e-01 8.32752109e-01 -1.80484243e-02 5.61267793e-01
1.60500854e-01 -1.40929067e+00 6.79768622e-01 7.87009299e-01
6.09447241e-01 6.27258360e-01 5.23225307e-01 -4.00285125e-01
-1.72518420e+00 -9.21932995e-01 8.63534093e-01 -2.16611564e-01
9.86025214e-01 -3.91057193e-01 -7.97003090e-01 8.96932423e-01
2.32463665e-02 2.04675362e-01 4.66337949e-01 2.34081119e-01
-8.03941727e-01 -1.51497811e-01 -1.02145636e+00 8.52667987e-01
1.64609754e+00 -6.77242219e-01 -3.65529150e-01 2.78685123e-01
7.49807656e-01 -1.58784777e-01 -9.69168544e-01 6.86661124e-01
4.47063982e-01 -1.37532365e+00 7.74121046e-01 -1.00599599e+00
1.44820720e-01 -1.55849054e-01 -2.58744597e-01 -1.34080470e+00
-6.27965152e-01 -7.75210857e-01 7.82124233e-03 7.57097185e-01
8.05409923e-02 -1.14871514e+00 7.82364964e-01 -3.09882592e-02
-1.79843009e-01 -7.31599271e-01 -8.77913237e-01 -1.26977146e+00
-2.17215326e-02 -3.52081388e-01 9.11324978e-01 1.14400065e+00
1.44868061e-01 3.91451418e-01 -1.94536731e-01 7.81425610e-02
3.81944507e-01 4.45332736e-01 7.34362364e-01 -1.78900349e+00
-4.66307670e-01 -7.97157347e-01 -1.24625301e+00 -6.24995947e-01
5.80143273e-01 -1.69364476e+00 -5.56759834e-01 -1.21740675e+00
1.11982606e-01 1.41303539e-02 -3.98981124e-01 4.80793625e-01
4.89009768e-01 3.30790728e-01 6.14916719e-02 -2.30003625e-01
-7.14617789e-01 1.24151185e-01 1.16939259e+00 -2.97541767e-01
7.53964633e-02 -4.73994836e-02 -8.02841544e-01 7.95463383e-01
8.20179462e-01 -1.27032369e-01 -4.59844321e-01 -6.20505400e-02
7.27661967e-01 -1.37420282e-01 4.49227184e-01 -1.22262168e+00
3.02275151e-01 1.17097028e-01 4.31403555e-02 -2.13324159e-01
-3.16147685e-01 -8.12916517e-01 3.28715831e-01 7.60257244e-01
-3.90827507e-01 4.49954599e-01 2.59457575e-03 5.38294792e-01
-1.35764956e-01 -1.77460089e-01 6.26407683e-01 -7.55707845e-02
-5.57949841e-01 5.86528122e-01 1.73597783e-01 3.96756738e-01
7.62613297e-01 -3.06209564e-01 -5.79713941e-01 -4.91595328e-01
-6.34739041e-01 5.24469055e-02 4.17470813e-01 1.91245317e-01
2.13840485e-01 -1.53847611e+00 -5.18300712e-01 4.34563011e-01
2.20799163e-01 -4.66324031e-01 -3.03199925e-02 9.59264159e-01
-6.02036893e-01 5.46167135e-01 -4.44902986e-01 -4.51172292e-01
-1.19618607e+00 9.12966490e-01 5.71350992e-01 -7.18356431e-01
-4.44019288e-01 3.08138281e-01 2.47103572e-01 -5.64018309e-01
1.51349083e-01 -5.06486475e-01 1.09406570e-02 -2.42825702e-01
2.17424616e-01 6.01498306e-01 7.87604898e-02 -4.95350778e-01
-2.87336498e-01 4.43128139e-01 2.08878785e-01 4.06024694e-01
1.29523766e+00 5.18100083e-01 -6.86970294e-01 1.57614410e-01
1.38817239e+00 2.70605624e-01 -4.46368963e-01 -2.17414349e-01
2.52951443e-01 -5.72001189e-02 -6.55883610e-01 -3.16140443e-01
-1.01942599e+00 6.49500668e-01 1.96494013e-01 1.15010393e+00
1.16942799e+00 1.30669028e-01 5.04590452e-01 9.26466286e-01
3.35279912e-01 -8.47034097e-01 -8.53982270e-02 9.55637991e-01
5.34850359e-01 -6.91570640e-01 -2.37736896e-01 -4.09590960e-01
1.73942938e-01 1.42798400e+00 3.94189477e-01 -4.27561760e-01
4.32694614e-01 1.25721633e-01 -7.54928350e-01 -3.16533357e-01
-4.67888415e-01 -4.04879481e-01 6.07776344e-01 7.99279273e-01
1.64711345e-02 2.20268592e-01 -4.12653983e-01 4.59718227e-01
-4.89232838e-01 -3.16013210e-02 4.34129328e-01 2.91338116e-01
-2.36641794e-01 -1.09279895e+00 7.42110386e-02 4.84323442e-01
-3.31880838e-01 -2.64830202e-01 -7.84147739e-01 1.25880337e+00
-2.61871535e-02 5.03670514e-01 8.58801678e-02 -7.10891724e-01
2.09401771e-01 -5.98770082e-02 7.65691221e-01 -2.50775695e-01
-4.12636489e-01 -6.25498235e-01 2.22970232e-01 -6.66159809e-01
-2.81646997e-01 -1.58826709e-02 -8.19917202e-01 -7.47337759e-01
-1.67429850e-01 3.59988481e-01 1.91913262e-01 6.34129226e-01
2.34210864e-01 4.27969038e-01 6.52260184e-01 -2.51325071e-01
-7.77462959e-01 -6.61552846e-01 -9.96194184e-01 3.97166997e-01
1.54491931e-01 -3.57219934e-01 -5.92766285e-01 -8.77554238e-01] | [6.887083530426025, 6.214320182800293] |
a221e393-e88a-47d1-95ef-e9f1d955be45 | the-tiny-time-series-transformer-low-latency | 2303.08951 | null | https://arxiv.org/abs/2303.08951v1 | https://arxiv.org/pdf/2303.08951v1.pdf | The Tiny Time-series Transformer: Low-latency High-throughput Classification of Astronomical Transients using Deep Model Compression | A new golden age in astronomy is upon us, dominated by data. Large astronomical surveys are broadcasting unprecedented rates of information, demanding machine learning as a critical component in modern scientific pipelines to handle the deluge of data. The upcoming Legacy Survey of Space and Time (LSST) of the Vera C. Rubin Observatory will raise the big-data bar for time-domain astronomy, with an expected 10 million alerts per-night, and generating many petabytes of data over the lifetime of the survey. Fast and efficient classification algorithms that can operate in real-time, yet robustly and accurately, are needed for time-critical events where additional resources can be sought for follow-up analyses. In order to handle such data, state-of-the-art deep learning architectures coupled with tools that leverage modern hardware accelerators are essential. We showcase how the use of modern deep compression methods can achieve a $18\times$ reduction in model size, whilst preserving classification performance. We also show that in addition to the deep compression techniques, careful choice of file formats can improve inference latency, and thereby throughput of alerts, on the order of $8\times$ for local processing, and $5\times$ in a live production setting. To test this in a live setting, we deploy this optimised version of the original time-series transformer, t2, into the community alert broking system of FINK on real Zwicky Transient Facility (ZTF) alert data, and compare throughput performance with other science modules that exist in FINK. The results shown herein emphasise the time-series transformer's suitability for real-time classification at LSST scale, and beyond, and introduce deep model compression as a fundamental tool for improving deploy-ability and scalable inference of deep learning models for transient classification. | ['Jason D. McEwen', 'Julien Peloton', 'Tarek Allam Jr.'] | 2023-03-15 | null | null | null | null | ['model-compression', 'astronomy'] | ['methodology', 'miscellaneous'] | [-1.32548213e-01 -5.33483803e-01 1.82011917e-01 -3.79748017e-01
-7.65068114e-01 -7.18480825e-01 7.84548044e-01 3.69529098e-01
-6.17294908e-01 4.07139778e-01 -2.95324385e-01 -9.84650135e-01
-6.16095841e-01 -1.03372848e+00 -5.50226450e-01 -6.31867290e-01
-3.58732700e-01 8.52943003e-01 4.70543474e-01 -9.54526290e-02
2.87018448e-01 1.05159736e+00 -1.80461264e+00 2.82784015e-01
3.20127517e-01 1.55147755e+00 -1.55582324e-01 1.07773054e+00
7.26459399e-02 6.52901053e-01 -6.88847363e-01 8.15542862e-02
3.64078879e-01 1.49805412e-01 -5.11854827e-01 -7.63610363e-01
4.28701103e-01 -2.95644611e-01 -5.33970237e-01 3.20403099e-01
7.18362927e-01 2.05015168e-01 1.29591584e-01 -1.09870875e+00
5.69401264e-01 7.22111687e-02 -2.96565965e-02 8.86501610e-01
1.05979305e-03 5.29987037e-01 5.40515780e-01 -3.48813623e-01
2.96200216e-01 9.19487715e-01 9.69122112e-01 -2.58314371e-01
-1.04261053e+00 -7.32281864e-01 -5.11793077e-01 4.56536829e-01
-9.37114239e-01 -5.57151616e-01 1.15443259e-01 -4.49939698e-01
1.66388416e+00 6.63917243e-01 7.82902718e-01 9.16558623e-01
4.73740369e-01 -1.38791874e-01 8.72574031e-01 -1.03471845e-01
5.35189450e-01 -2.07484052e-01 1.18828259e-01 4.99017030e-01
1.89172640e-01 4.55020875e-01 -7.77139723e-01 -5.39162099e-01
5.02408624e-01 -4.53407243e-02 4.22102213e-02 1.98162526e-01
-1.23987615e+00 8.78440142e-01 1.44854605e-01 9.73353833e-02
-1.01888821e-01 4.20110881e-01 7.67416835e-01 4.58594203e-01
5.10219932e-01 6.99954152e-01 -7.44971812e-01 -6.85420096e-01
-1.14278257e+00 5.55591285e-01 9.58805561e-01 4.68059987e-01
3.40601474e-01 2.52794445e-01 -8.61046687e-02 2.89196759e-01
-1.24567464e-01 6.09754562e-01 5.95715761e-01 -1.07860434e+00
1.13802189e-02 5.91968894e-01 -1.29456103e-01 -6.44656122e-01
-8.73624682e-01 -8.11776996e-01 -7.96122491e-01 4.21522945e-01
5.05783081e-01 -9.38822851e-02 -6.24691904e-01 1.21541238e+00
6.18639529e-01 4.16144818e-01 -1.54931217e-01 5.96679747e-01
3.50848585e-01 6.11691833e-01 -3.38971883e-01 -3.25721242e-02
1.65340829e+00 -2.17280895e-01 -1.11604065e-01 -2.23061383e-01
7.76962280e-01 -9.87812638e-01 7.43326485e-01 7.37251699e-01
-1.04998243e+00 -5.95332265e-01 -1.12885213e+00 -1.89742938e-01
-5.22773385e-01 -2.81032532e-01 1.14075232e+00 4.52611208e-01
-9.37404335e-01 7.97674358e-01 -1.33849835e+00 -4.37357903e-01
2.18800902e-01 3.40169162e-01 -1.06658190e-02 2.11590856e-01
-1.01069307e+00 6.93760514e-01 1.19960316e-01 -2.40935192e-01
-1.06015325e+00 -1.42960477e+00 -4.38120723e-01 4.25593585e-01
-2.10339278e-02 -5.26267648e-01 1.60178387e+00 8.62601325e-02
-8.10081244e-01 5.56637943e-01 -1.38868636e-03 -1.06781566e+00
3.16969246e-01 1.07672505e-01 -6.69807017e-01 3.13100457e-01
-3.94589156e-01 2.16698244e-01 6.37019753e-01 -7.89024159e-02
-7.87755013e-01 -3.28842074e-01 -1.36113122e-01 -4.48607564e-01
-3.61513704e-01 1.94956079e-01 -5.96750230e-02 -2.53232211e-01
-2.17590444e-02 -8.67661178e-01 -1.55352056e-01 1.80166230e-01
1.80582151e-01 -1.52422607e-01 1.15210056e+00 -5.27247369e-01
9.29291666e-01 -1.98149323e+00 -2.92985022e-01 2.06599578e-01
2.63409436e-01 4.25931841e-01 3.80809903e-01 5.43192208e-01
-1.21068656e-01 -4.55131412e-01 -2.56114393e-01 -1.70091182e-01
-1.56831555e-02 2.71982342e-01 -6.81688726e-01 5.60734570e-01
-5.06490581e-02 4.82815474e-01 -5.48717439e-01 1.14583775e-01
1.72349378e-01 4.65367705e-01 -5.45000076e-01 1.91538662e-01
-4.55615997e-01 4.39333022e-01 -2.29157090e-01 4.66675818e-01
4.00775701e-01 -4.64384258e-01 -1.70986354e-01 -2.08728686e-01
-6.69008493e-01 6.64567173e-01 -8.69024813e-01 1.60456812e+00
-6.07319236e-01 8.83558810e-01 2.09803611e-01 -1.26062238e+00
8.74796391e-01 1.01218246e-01 5.86193919e-01 -9.42367554e-01
1.29201233e-01 2.64917135e-01 1.26056120e-01 -4.12692279e-01
3.85869592e-01 -1.48683628e-02 2.10113768e-02 7.16871679e-01
-1.74394604e-02 -2.15305388e-01 1.61075249e-01 3.70412678e-01
1.77696109e+00 -1.70270905e-01 -2.26669565e-01 -4.59056824e-01
6.50414359e-03 3.19620401e-01 1.11835726e-01 8.48356783e-01
8.25876743e-02 1.40661687e-01 5.58340967e-01 -9.31762040e-01
-1.31744349e+00 -7.44328141e-01 -6.59868062e-01 1.12563992e+00
-8.25453401e-01 -5.28155208e-01 -3.55490744e-01 -6.87919036e-02
1.58324942e-01 6.98457956e-01 -1.50244951e-01 -2.59228617e-01
-3.92210275e-01 -1.14458668e+00 8.72163534e-01 3.12810034e-01
3.48554105e-01 -6.58202708e-01 -1.33212769e+00 3.23655814e-01
2.45338053e-01 -1.09015036e+00 3.71522814e-01 5.45253277e-01
-8.37442398e-01 -1.07569587e+00 1.23886272e-01 -9.60905254e-02
-4.26818617e-02 1.53616503e-01 1.14197493e+00 1.18814625e-01
-9.79748249e-01 1.77984461e-01 -1.04772702e-01 -7.36793339e-01
-2.21270695e-01 2.40295529e-02 2.78085917e-01 -4.82933372e-01
3.74006778e-01 -8.91623199e-01 -6.18736446e-01 -3.39743607e-02
-9.66709673e-01 -1.38121098e-01 3.85739326e-01 5.50561309e-01
3.21239710e-01 3.62138391e-01 4.11968529e-01 -5.90963721e-01
-1.34846076e-01 -6.90180957e-01 -1.42055094e+00 -2.42604375e-01
-7.28320241e-01 2.48836398e-01 8.35736096e-01 -2.21224487e-01
-7.94723153e-01 -1.86871350e-01 -4.12893653e-01 -3.13580215e-01
1.50586754e-01 5.74601829e-01 5.42857528e-01 -5.28960168e-01
9.38690543e-01 2.35551983e-01 2.27138251e-01 -5.79506159e-01
-8.11351314e-02 6.59938872e-01 8.95244956e-01 -6.07141912e-01
6.82318747e-01 8.00363660e-01 4.98881191e-01 -9.78896737e-01
-8.48276854e-01 -4.17069465e-01 -1.65618911e-01 -1.33329898e-01
4.89564389e-01 -1.04609907e+00 -1.12929964e+00 3.74018103e-01
-9.32540238e-01 -3.69340092e-01 -1.21763721e-01 5.97386718e-01
-2.56871134e-01 3.21834646e-02 -3.83155823e-01 -5.98644078e-01
-4.96614277e-01 -9.84493196e-01 1.15680945e+00 1.48098934e-02
-6.17845096e-02 -8.08060825e-01 2.17457980e-01 2.31458187e-01
9.98121798e-01 3.41265887e-01 9.39259768e-01 -8.75059128e-01
-6.93144500e-01 -6.73839748e-01 -3.12205911e-01 2.47148611e-03
-6.83713019e-01 -1.74530014e-01 -1.21188116e+00 -5.20457625e-01
6.55354321e-01 -3.74195218e-01 6.82288110e-01 1.41758010e-01
1.76213562e+00 -1.30119234e-01 -2.96601623e-01 1.03400755e+00
1.28713775e+00 -2.10860390e-02 5.36147296e-01 2.01352090e-01
2.78435171e-01 2.10744962e-01 3.30042571e-01 8.62572610e-01
-1.03438370e-01 5.69950163e-01 3.93907368e-01 1.67641953e-01
1.21054605e-01 4.33726192e-01 8.76470879e-02 8.68710101e-01
2.07955927e-01 2.37008985e-02 -1.09222937e+00 2.56101310e-01
-1.40404522e+00 -1.15322971e+00 -4.11993742e-01 2.61929607e+00
7.50525057e-01 6.65075541e-01 -2.22576648e-01 3.05227429e-01
-7.25956559e-02 -5.33707589e-02 -6.93869650e-01 -4.92550522e-01
2.61288643e-01 8.96246970e-01 7.65030503e-01 3.77671361e-01
-6.24473572e-01 1.53604954e-01 5.82735395e+00 9.57062840e-01
-1.49463117e+00 2.77689934e-01 5.14911532e-01 -7.51232684e-01
-3.00510004e-02 8.19626376e-02 -1.07491791e+00 6.59574687e-01
2.24091625e+00 -2.65409648e-01 6.29173994e-01 8.81864429e-01
2.71827012e-01 -3.43258888e-01 -1.07529831e+00 8.69199216e-01
-2.52986163e-01 -1.86877918e+00 -6.50503039e-01 2.44467005e-01
1.43665463e-01 7.47150242e-01 -3.22767645e-01 6.06873572e-01
1.44637436e-01 -1.00465167e+00 3.02916050e-01 6.12696350e-01
8.21778357e-01 -8.66727591e-01 5.21594048e-01 5.21533787e-01
-1.07094145e+00 -8.00444186e-02 -5.88551819e-01 -4.79513139e-01
3.18569969e-03 1.17808127e+00 -9.54024255e-01 4.56957340e-01
1.01309860e+00 2.49364585e-01 -8.06420624e-01 1.09502816e+00
5.60179830e-01 1.01772296e+00 -1.04172719e+00 2.40508467e-02
1.66391388e-01 2.91539669e-01 4.56255257e-01 1.12030363e+00
6.58502340e-01 8.22393503e-03 -1.01708569e-01 3.17129701e-01
1.21064931e-01 -8.40516448e-01 -5.25390804e-01 -6.93637282e-02
6.81912959e-01 1.61643946e+00 -7.24746346e-01 -6.39989436e-01
-3.24766904e-01 2.05921084e-01 1.47236422e-01 -2.89251685e-01
-9.53818202e-01 -4.34385985e-01 8.32559586e-01 3.21095824e-01
1.74683332e-01 -5.78801751e-01 -3.66812259e-01 -6.72683656e-01
-2.48545825e-01 -9.77609694e-01 6.00256622e-01 -7.71917045e-01
-1.10104764e+00 4.11484867e-01 1.05497815e-01 -9.40162599e-01
-3.24721009e-01 -7.31053233e-01 -7.27850199e-01 8.31493855e-01
-1.09279466e+00 -6.39951766e-01 -4.99148637e-01 3.87624383e-01
2.42436439e-01 -3.41435879e-01 9.68570888e-01 6.24300897e-01
-3.62155288e-01 2.53841400e-01 4.86814976e-01 -4.74680245e-01
5.22408664e-01 -1.05858445e+00 7.03928649e-01 8.31064105e-01
-5.73010147e-02 3.62552553e-01 9.94925559e-01 -5.73135138e-01
-2.07464886e+00 -1.19916630e+00 6.51639163e-01 -5.29495299e-01
1.05748832e+00 -6.98523581e-01 -1.14878142e+00 5.26666641e-01
-1.99209601e-02 8.36509839e-02 5.40323436e-01 1.04882263e-01
-3.60078901e-01 -3.33796859e-01 -9.90424037e-01 1.35845691e-01
7.10976958e-01 -7.08598614e-01 -2.74578452e-01 1.12488949e+00
9.21810031e-01 -5.58231533e-01 -1.11237431e+00 4.12789792e-01
4.28994387e-01 -1.04505801e+00 1.15434027e+00 -4.93015707e-01
2.15706781e-01 -1.01195201e-01 7.04349428e-02 -7.69618809e-01
-1.71453327e-01 -8.22779834e-01 -4.26724136e-01 7.47638881e-01
-2.16193013e-02 -6.44443214e-01 6.85023785e-01 3.38976055e-01
-5.42371452e-01 -5.13025880e-01 -1.24818349e+00 -8.65133524e-01
-1.79193720e-01 -1.00083876e+00 4.64419335e-01 7.36208200e-01
-2.80064344e-01 8.46051052e-02 4.61966470e-02 4.29792762e-01
6.99896395e-01 9.22944248e-02 7.15703487e-01 -1.37929046e+00
-7.98895240e-01 -3.18008423e-01 -5.23464739e-01 -2.09748909e-01
-4.99937803e-01 -8.75525594e-01 -3.01028997e-01 -9.98449743e-01
-4.14671361e-01 -5.73747516e-01 -3.42343897e-02 5.43302536e-01
4.41367090e-01 2.69653112e-01 -3.67596030e-01 9.17861462e-02
-3.06124449e-01 1.57621443e-01 4.29562837e-01 2.40995780e-01
3.87170404e-01 -1.42400652e-01 -1.30318552e-01 4.94554162e-01
7.45716035e-01 -4.71794635e-01 -1.63641632e-01 -5.05629003e-01
5.01311123e-01 8.03798586e-02 7.02031374e-01 -1.83618546e+00
7.02704668e-01 8.68681297e-02 7.17172921e-01 -6.79716110e-01
6.31349444e-01 -4.94071543e-01 4.45737034e-01 8.24480951e-01
-6.68155104e-02 3.85935098e-01 7.72613168e-01 3.01298797e-01
1.08618990e-01 -3.85344774e-02 9.66869056e-01 -7.44153559e-02
-3.67364794e-01 2.59764940e-01 -4.31339920e-01 3.21758352e-02
8.65177751e-01 5.80296576e-01 -7.59149373e-01 -3.94239463e-02
-1.27639174e-01 1.15068741e-01 2.05409169e-01 7.92632774e-02
5.33646904e-02 -8.98676693e-01 -6.27495527e-01 6.68695152e-01
-9.75225270e-02 1.29928648e-01 5.29616833e-01 9.82969820e-01
-9.58204448e-01 6.86897576e-01 -1.99024782e-01 -7.04464018e-01
-8.83600652e-01 4.18600589e-01 3.06889415e-01 -2.40097806e-01
-8.01270902e-01 8.69195938e-01 -3.60127777e-01 -1.95028290e-01
3.07506442e-01 -4.66220319e-01 4.21950102e-01 7.11839423e-02
1.07967043e+00 3.72282624e-01 1.05386066e+00 3.78899723e-01
-2.15456024e-01 -5.37713356e-02 -1.65220443e-02 -8.14868882e-02
1.53193128e+00 5.97341776e-01 -2.03604847e-01 3.52245361e-01
1.05423141e+00 -1.83779433e-01 -1.07149339e+00 -7.39724636e-02
-7.35679939e-02 -3.47727776e-01 4.88904566e-01 -8.18814099e-01
-6.78765833e-01 1.05442429e+00 8.56555223e-01 5.45193672e-01
1.10706866e+00 -3.01610678e-02 1.05719328e+00 7.56257892e-01
3.19514424e-01 -7.73546636e-01 -2.55652130e-01 5.98387599e-01
6.38356924e-01 -8.76142859e-01 2.90736198e-01 3.27404737e-01
1.98457882e-01 1.16877997e+00 2.79426932e-01 1.38638606e-02
5.67433119e-01 8.27326298e-01 -3.27854007e-01 -4.75201160e-01
-1.44762981e+00 2.76654840e-01 -1.61894113e-01 2.97998071e-01
1.39331222e-01 -2.13646874e-01 3.95126566e-02 2.48720020e-01
-5.47158539e-01 -4.27020993e-03 4.08031166e-01 1.14735746e+00
-8.06668103e-01 -9.55010712e-01 -7.10282505e-01 1.02352476e+00
-3.87544781e-01 -1.95896588e-02 2.39840433e-01 5.52122414e-01
-1.18495733e-01 7.81924605e-01 5.10948062e-01 -1.95443764e-01
2.70169340e-02 2.19474778e-01 1.18834391e-01 -2.78589606e-01
-8.17166626e-01 -1.50556549e-01 1.74646646e-01 -8.98235619e-01
3.15509379e-01 -6.44981444e-01 -1.20399475e+00 -9.51994359e-01
3.54311734e-01 1.77920878e-01 1.30191338e+00 1.05043626e+00
6.38261259e-01 7.31724620e-01 3.83393556e-01 -1.06560969e+00
-7.37176538e-01 -8.61460209e-01 -1.34117767e-01 -1.74323753e-01
2.24507570e-01 -4.32536632e-01 -6.07291460e-01 -2.73643583e-01] | [8.099111557006836, 3.155766725540161] |
1ee692d1-fc43-430e-8d0d-547dcc28422a | recent-advances-and-applications-of-machine | 2303.07647 | null | https://arxiv.org/abs/2303.07647v3 | https://arxiv.org/pdf/2303.07647v3.pdf | Recent Advances and Applications of Machine Learning in Experimental Solid Mechanics: A Review | For many decades, experimental solid mechanics has played a crucial role in characterizing and understanding the mechanical properties of natural and novel materials. Recent advances in machine learning (ML) provide new opportunities for the field, including experimental design, data analysis, uncertainty quantification, and inverse problems. As the number of papers published in recent years in this emerging field is exploding, it is timely to conduct a comprehensive and up-to-date review of recent ML applications in experimental solid mechanics. Here, we first provide an overview of common ML algorithms and terminologies that are pertinent to this review, with emphasis placed on physics-informed and physics-based ML methods. Then, we provide thorough coverage of recent ML applications in traditional and emerging areas of experimental mechanics, including fracture mechanics, biomechanics, nano- and micro-mechanics, architected materials, and 2D material. Finally, we highlight some current challenges of applying ML to multi-modality and multi-fidelity experimental datasets and propose several future research directions. This review aims to provide valuable insights into the use of ML methods as well as a variety of examples for researchers in solid mechanics to integrate into their experiments. | ['Horacio D. Espinosa', 'Enrui Zhang', 'Hanxun Jin'] | 2023-03-14 | null | null | null | null | ['experimental-design'] | ['methodology'] | [ 8.11802596e-02 -3.64254922e-01 -1.79395467e-01 2.63706576e-02
-8.54319692e-01 1.43218264e-01 6.72331378e-02 3.10973376e-01
-2.89262772e-01 9.17317808e-01 -1.70991644e-01 8.07702988e-02
-5.94292581e-01 -7.54860342e-01 -8.27737212e-01 -9.22021687e-01
-3.67085040e-01 8.46992254e-01 2.73361862e-01 6.18463121e-02
4.38408881e-01 7.04434872e-01 -1.59776938e+00 1.67052835e-01
4.77576345e-01 8.60406041e-01 3.18257600e-01 4.86892998e-01
2.36148462e-01 1.99180901e-01 1.26487417e-02 3.33881229e-02
-2.70922959e-01 -4.04065736e-02 -9.37582731e-01 -2.67284632e-01
6.16236180e-02 -2.81968653e-01 -4.17754412e-01 4.23787504e-01
7.43446112e-01 3.26482832e-01 1.17869222e+00 -1.07139099e+00
-7.13404119e-01 7.97100961e-01 -6.43294096e-01 1.19029589e-01
4.58592921e-01 2.57733047e-01 1.10894012e+00 -1.37683785e+00
7.02947259e-01 9.95544195e-01 7.37033904e-01 5.44862568e-01
-1.23252261e+00 -2.26392359e-01 -2.62226254e-01 3.36847275e-01
-1.20476401e+00 -3.23824853e-01 1.12178981e+00 -8.63605499e-01
8.23998272e-01 2.35431015e-01 5.95000327e-01 9.90868688e-01
9.18433249e-01 7.57578373e-01 1.25542998e+00 -6.55471683e-01
5.38188279e-01 -3.32728565e-01 8.01503658e-02 4.33832258e-01
4.15758461e-01 5.14541686e-01 -7.46953130e-01 -1.29122198e-01
9.00026977e-01 -3.36243689e-01 -2.31234897e-02 -5.42151868e-01
-1.24017787e+00 5.39267302e-01 1.16116308e-01 4.20703292e-01
-2.36128211e-01 4.34756249e-01 4.60603952e-01 -4.59477939e-02
4.03421849e-01 6.33225143e-01 -5.39119899e-01 -7.61426389e-02
-1.07761991e+00 8.48680854e-01 6.35900676e-01 4.84762102e-01
5.51886022e-01 1.86304525e-01 4.56030488e-01 9.22452807e-01
7.83220232e-01 8.76697719e-01 8.28515273e-03 -9.05891418e-01
5.91065213e-02 -1.90719604e-01 -2.19640777e-01 -8.05669606e-01
-5.67411065e-01 1.05928024e-02 -7.23376095e-01 4.12220776e-01
1.24624826e-01 2.06163704e-01 -7.94673681e-01 1.19627225e+00
4.73736256e-01 2.43519336e-01 -4.54106510e-01 9.53889728e-01
1.19859183e+00 3.56645316e-01 1.65153053e-02 -3.26408923e-01
9.46020722e-01 -3.83332849e-01 -6.02756321e-01 -1.01030715e-01
2.50202417e-01 -1.06321883e+00 9.87278521e-01 1.12672947e-01
-1.36080456e+00 -2.55082458e-01 -9.41881120e-01 7.13753626e-02
-1.90724641e-01 -2.16745213e-01 6.34321988e-01 2.09372744e-01
-2.35243484e-01 1.20901477e+00 -1.49372613e+00 -2.78683901e-01
3.20993543e-01 2.73208559e-01 -1.03268594e-01 3.25483680e-02
-1.22633362e+00 9.80936408e-01 -5.66343181e-02 2.24614158e-01
-7.51319230e-01 -1.25073349e+00 -7.28286088e-01 -5.97810149e-01
3.27973306e-01 -8.33981335e-01 1.06194425e+00 8.97736609e-01
-1.63482308e+00 8.57132375e-01 -3.09671205e-03 3.63572627e-01
3.05351377e-01 -3.87206167e-01 -2.05846757e-01 6.67359754e-02
9.87440795e-02 7.18381852e-02 2.26870999e-01 -1.61520052e+00
8.96645412e-02 -3.65952998e-01 -6.61047041e-01 -3.51497114e-01
3.30188483e-01 -1.57984838e-01 1.22310631e-01 -7.62583792e-01
6.38721943e-01 -9.65164423e-01 -4.83704746e-01 9.77804065e-02
-8.26789975e-01 -5.15575781e-02 6.86999857e-01 -3.65263283e-01
1.14373410e+00 -1.65709972e+00 4.71337259e-01 3.72780442e-01
3.54093224e-01 -3.52241606e-01 3.55079383e-01 9.64475095e-01
1.62213683e-01 -1.31315142e-01 -8.99045348e-01 -1.75290927e-01
-4.22357731e-02 3.02685872e-02 2.27957945e-02 6.32571757e-01
-7.48847425e-02 1.15259683e+00 -8.41114700e-01 -6.07448816e-01
5.57444751e-01 1.93814337e-01 -4.29847002e-01 2.57903095e-02
-3.48529547e-01 7.61345088e-01 -6.59905076e-01 1.21477199e+00
7.27916062e-01 -1.08228333e-01 -2.02568844e-01 -5.01640320e-01
-2.38553956e-01 9.56403837e-03 -1.35693181e+00 1.46628881e+00
-4.00068283e-01 2.49920219e-01 2.60756046e-01 -1.09119129e+00
7.47643590e-01 3.44558746e-01 1.10614753e+00 -3.69261801e-01
8.43299478e-02 9.81338084e-01 -8.54738653e-02 -7.96762228e-01
6.31751567e-02 -7.11683273e-01 -7.75374696e-02 6.94623351e-01
-1.68694958e-01 -1.02960014e+00 6.04639985e-02 -8.82199481e-02
6.86166584e-01 5.15188754e-01 9.29346085e-02 -5.51421046e-01
6.00319743e-01 -1.61898330e-01 1.09564215e-01 4.67002273e-01
-6.99221343e-02 5.27628362e-01 -4.71304536e-01 -4.58580941e-01
-1.27884948e+00 -1.54459071e+00 -7.89646745e-01 5.36590695e-01
2.20689937e-01 -2.28772044e-01 -3.87745112e-01 3.44525546e-01
6.55023098e-01 2.66209066e-01 -5.48955798e-01 -8.75029489e-02
-9.17943776e-01 -1.11427557e+00 8.60899240e-02 8.41821313e-01
-2.60603651e-02 -1.19594836e+00 -1.00187749e-01 3.61438423e-01
-9.44035277e-02 -1.25733614e+00 1.71186253e-01 1.64114624e-01
-1.21446872e+00 -1.11609256e+00 -2.87882298e-01 -6.02017045e-01
2.44668454e-01 -9.46745798e-02 1.07741177e+00 1.83960617e-01
-8.49861920e-01 6.86306536e-01 -4.73184027e-02 -2.61787921e-01
-7.16679990e-01 -3.24113062e-03 5.84128439e-01 -8.38028848e-01
1.16195351e-01 -9.92577732e-01 -4.69664454e-01 1.60106361e-01
-8.90280783e-01 -2.24299163e-01 5.93794405e-01 7.17817366e-01
1.10819829e+00 7.45569915e-02 7.29448140e-01 -7.42449999e-01
5.22482872e-01 -5.54768741e-01 -1.44252673e-01 -1.47983944e-03
-7.98298120e-01 -4.35921177e-03 5.25850594e-01 -6.89730495e-02
-8.97120237e-01 -3.42613488e-01 -6.75737143e-01 4.97570299e-02
-2.04885021e-01 9.10637498e-01 -8.16473439e-02 -3.94349188e-01
6.82442844e-01 -9.72179603e-03 1.82313740e-01 -5.62942088e-01
1.47681579e-01 5.06107986e-01 3.22953910e-01 -1.51153064e+00
8.67316484e-01 4.79331762e-01 7.54173338e-01 -1.17988825e+00
-5.33782780e-01 -2.71603018e-01 -9.69929457e-01 -5.47479749e-01
5.18803418e-01 -1.33869439e-01 -9.07549262e-01 8.12175214e-01
-4.67397243e-01 -3.31338614e-01 -5.47057092e-01 8.09558630e-01
-1.04928124e+00 5.20114481e-01 -1.16585088e+00 -6.71325624e-01
-3.21471781e-01 -1.61031687e+00 1.28224850e+00 3.55786607e-02
-3.33765686e-01 -1.38601863e+00 3.74315828e-01 6.33579910e-01
4.56469715e-01 5.37813008e-01 1.14216566e+00 1.54221465e-03
-5.77206850e-01 -2.61826158e-01 3.90444517e-01 6.30411610e-04
2.44709522e-01 6.31000996e-01 -7.98652768e-01 -1.38372704e-01
-2.13361196e-02 -6.28810883e-01 8.90425861e-01 9.53503191e-01
1.17229092e+00 4.71351922e-01 -7.92969882e-01 2.00777411e-01
1.55056095e+00 -1.09023221e-01 3.38816702e-01 2.11823761e-01
7.04724133e-01 4.60897416e-01 5.18109918e-01 4.28482592e-01
2.07599234e-02 6.05438888e-01 2.02872053e-01 1.77627861e-01
-2.84026831e-01 1.13015715e-02 -2.04960495e-01 1.18651891e+00
-5.52780688e-01 1.57403238e-02 -8.54312181e-01 2.00114518e-01
-1.42017484e+00 -1.01742733e+00 -4.64360774e-01 2.07508898e+00
8.50466251e-01 1.30081192e-01 -1.84989125e-01 4.36667740e-01
6.30151629e-01 -8.43196511e-02 -7.09443986e-01 -8.52232203e-02
-2.94572443e-01 5.48460960e-01 5.11065423e-02 6.24859571e-01
-1.07814622e+00 6.34357929e-01 8.06711483e+00 8.10942173e-01
-1.13050735e+00 -2.75032759e-01 4.69452858e-01 8.94645974e-02
-5.18153608e-01 -2.08659768e-01 -5.64870834e-01 4.54356700e-01
7.87699580e-01 -1.44176915e-01 2.06493288e-01 5.39724171e-01
3.87576669e-01 -5.23311853e-01 -1.21033192e+00 5.78801513e-01
-4.61286217e-01 -1.78989768e+00 -2.25069076e-01 -5.40647209e-02
8.11904371e-01 1.24581441e-01 -2.16030460e-02 -2.22561419e-01
-2.51399010e-01 -9.26423609e-01 5.81727147e-01 6.32898986e-01
8.18916440e-01 -2.19131336e-01 7.47329175e-01 3.04265730e-02
-1.37082493e+00 3.89291048e-01 -3.26441199e-01 -1.34097293e-01
8.31054688e-01 1.37956691e+00 -2.59670228e-01 8.41978431e-01
8.06953609e-01 6.86216414e-01 1.79760933e-01 9.88014460e-01
2.02360064e-01 6.51817858e-01 -6.01412594e-01 -1.09147755e-02
-3.84110838e-01 -3.00519586e-01 7.35814810e-01 8.63494754e-01
1.62073039e-02 2.62967288e-01 2.56143481e-01 9.70306158e-01
1.31749526e-01 -8.23934842e-03 -3.05429131e-01 -8.22052285e-02
6.84954166e-01 8.99554968e-01 -8.19275022e-01 3.47373903e-01
-3.41015071e-01 5.07085696e-02 -1.61040351e-02 1.97956651e-01
-6.65853620e-01 5.61814643e-02 5.53677261e-01 6.14998639e-01
-4.34769869e-01 -7.21646667e-01 -6.77178025e-01 -6.12299979e-01
-3.20731439e-02 -1.89500913e-01 2.64171392e-01 -5.78335643e-01
-1.75251758e+00 -2.23119870e-01 5.99959671e-01 -7.67744958e-01
-7.71160144e-03 -1.00511086e+00 -4.26990241e-01 4.18153167e-01
-1.01725519e+00 -1.02005565e+00 -1.57441366e-02 -1.92645893e-01
1.84503332e-01 5.43078072e-02 6.65417433e-01 3.75443161e-01
-5.08149385e-01 1.76084131e-01 6.22430325e-01 -2.88835645e-01
7.37164080e-01 -1.02060163e+00 3.16322953e-01 2.34651536e-01
-3.62179726e-01 5.17581880e-01 1.14528871e+00 -1.12200689e+00
-1.94718456e+00 -4.87897158e-01 1.38222039e-01 -3.42656732e-01
8.97991061e-01 1.21997437e-02 -1.04391778e+00 4.18944001e-01
-4.75159347e-01 1.90966249e-01 9.12281930e-01 1.53625235e-02
3.29409480e-01 1.82645187e-01 -1.26631749e+00 4.92087781e-01
1.05702031e+00 -4.36111540e-01 -3.89859617e-01 4.23024535e-01
2.79711694e-01 -7.42433012e-01 -1.60910726e+00 1.02252448e+00
9.05643106e-01 -5.96399546e-01 1.35275757e+00 -5.57241380e-01
9.93539870e-01 -1.87471490e-02 -1.79758728e-01 -9.93695140e-01
-5.77675641e-01 -2.42613062e-01 -4.52870905e-01 7.59937704e-01
1.25589371e-01 -5.22112608e-01 1.08236837e+00 7.49999583e-01
-6.54027879e-01 -1.50927067e+00 -1.13503158e+00 -7.15057731e-01
8.01393032e-01 -6.09551311e-01 1.42112702e-01 5.32366455e-01
1.14492074e-01 -1.64049357e-01 -8.82768631e-02 -1.30785421e-01
1.09457874e+00 3.67009848e-01 3.83598775e-01 -1.43863308e+00
1.55073600e-02 -3.42887700e-01 -4.87345994e-01 -7.55591571e-01
1.68555886e-01 -8.10226798e-01 2.42003530e-01 -1.60658002e+00
3.70505303e-01 -7.04369664e-01 -1.31147265e-01 -1.76050350e-01
-1.20567858e-01 3.53447616e-01 -5.12679338e-01 5.17980516e-01
2.62151152e-01 7.23781407e-01 1.86835957e+00 -1.71828493e-01
-4.45050858e-02 9.29322019e-02 -2.37231687e-01 4.86702025e-01
6.02705836e-01 -2.10949540e-01 -6.34689182e-02 -1.23215560e-02
2.43780583e-01 6.08097576e-02 3.19549471e-01 -1.08738232e+00
9.80605260e-02 -8.52998197e-01 4.17311251e-01 -9.59181964e-01
2.55573839e-01 -7.03562498e-01 4.40769643e-01 5.89286566e-01
-3.75550054e-02 -2.93514222e-01 2.42814228e-01 5.16999424e-01
-5.23211509e-02 -3.53661329e-01 1.10398901e+00 -2.08858699e-01
-5.85296094e-01 4.93415177e-01 -5.82741439e-01 7.00160190e-02
1.09463501e+00 -2.79343039e-01 -1.35371312e-01 2.00507000e-01
-1.04363263e+00 3.33492756e-02 4.53772336e-01 4.07126769e-02
8.91699731e-01 -1.29715490e+00 -6.68552399e-01 -3.75499874e-02
-3.08085941e-02 1.47995174e-01 9.04599726e-01 8.86291385e-01
-8.13909173e-01 4.12900835e-01 -8.37082937e-02 -8.74358118e-01
-7.58010924e-01 3.79767716e-01 2.92160064e-01 -4.23286371e-02
-6.00101471e-01 9.82685745e-01 -2.60289758e-01 -5.10529637e-01
-4.43587631e-01 -3.43988329e-01 3.08160514e-01 -5.85526347e-01
-1.07741259e-01 1.02881074e+00 5.43186784e-01 -5.88413119e-01
-6.31909907e-01 1.19175971e+00 2.43198037e-01 -6.74575055e-03
1.73015547e+00 -1.65281713e-01 -4.47055280e-01 1.06074500e+00
1.02000916e+00 7.95726106e-02 -9.47535276e-01 -7.04989210e-02
-2.25382760e-01 -1.81833491e-01 7.07683638e-02 -4.95504975e-01
-1.05014944e+00 8.18064213e-01 2.29952902e-01 -1.86158344e-01
6.97856426e-01 6.47492647e-01 1.08891368e+00 2.32926682e-01
6.58724308e-01 -1.37449384e+00 5.14129102e-01 2.31031120e-01
1.11744571e+00 -1.11806583e+00 8.20799708e-01 -9.48237062e-01
1.35867730e-01 1.43733358e+00 7.24491119e-01 -1.96588054e-01
1.56084573e+00 9.24864590e-01 -3.33385646e-01 -4.93642420e-01
-3.27773392e-01 4.08333689e-01 3.44322860e-01 4.71359253e-01
8.68634701e-01 1.12368673e-01 -2.95295924e-01 3.13008904e-01
-2.97049135e-01 -2.31694002e-02 4.96278375e-01 1.60014117e+00
-6.85295999e-01 -1.36884522e+00 -6.09228492e-01 7.16378987e-01
-1.83821559e-01 1.51168004e-01 1.44994274e-01 7.75576115e-01
-2.94799805e-01 5.24771333e-01 -1.40618682e-01 -2.75598437e-01
1.75517455e-01 4.96910959e-02 9.62759495e-01 -5.72909892e-01
1.95244670e-01 -3.29825021e-02 -1.29964948e-01 -3.75959665e-01
-7.20103145e-01 -8.91573131e-01 -1.68950939e+00 -5.69776595e-01
-6.84502184e-01 7.83734173e-02 6.61593080e-01 1.29155350e+00
-9.67854932e-02 5.78041911e-01 2.73416728e-01 -1.21324527e+00
-2.97213256e-01 -8.16348910e-01 -9.65414643e-01 1.58157662e-01
-6.07844815e-02 -1.52547741e+00 -4.42555338e-01 9.99539196e-02] | [6.362446308135986, 3.4275434017181396] |
b23b7d73-a906-4209-80ae-cb25db86bedb | reinforcement-learning-enhanced-shared | 2206.08088 | null | https://arxiv.org/abs/2206.08088v1 | https://arxiv.org/pdf/2206.08088v1.pdf | Reinforcement Learning-enhanced Shared-account Cross-domain Sequential Recommendation | Shared-account Cross-domain Sequential Recommendation (SCSR) is an emerging yet challenging task that simultaneously considers the shared-account and cross-domain characteristics in the sequential recommendation. Existing works on SCSR are mainly based on Recurrent Neural Network (RNN) and Graph Neural Network (GNN) but they ignore the fact that although multiple users share a single account, it is mainly occupied by one user at a time. This observation motivates us to learn a more accurate user-specific account representation by attentively focusing on its recent behaviors. Furthermore, though existing works endow lower weights to irrelevant interactions, they may still dilute the domain information and impede the cross-domain recommendation. To address the above issues, we propose a reinforcement learning-based solution, namely RL-ISN, which consists of a basic cross-domain recommender and a reinforcement learning-based domain filter. Specifically, to model the account representation in the shared-account scenario, the basic recommender first clusters users' mixed behaviors as latent users, and then leverages an attention model over them to conduct user identification. To reduce the impact of irrelevant domain information, we formulate the domain filter as a hierarchical reinforcement learning task, where a high-level task is utilized to decide whether to revise the whole transferred sequence or not, and if it does, a low-level task is further performed to determine whether to remove each interaction within it or not. To evaluate the performance of our solution, we conduct extensive experiments on two real-world datasets, and the experimental results demonstrate the superiority of our RL-ISN method compared with the state-of-the-art recommendation methods. | ['Hongzhi Yin', 'Xinhua Wang', 'Tong Chen', 'Jinyu Zhang', 'Lei Guo'] | 2022-06-16 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [ 2.65895445e-02 -1.82588056e-01 -6.73534751e-01 -1.40158638e-01
-2.77326167e-01 -5.05608380e-01 4.17921364e-01 -1.61853015e-01
-1.29106030e-01 5.49626291e-01 6.56058431e-01 -4.85334575e-01
-2.36863464e-01 -7.76857138e-01 -4.57841218e-01 -4.03887749e-01
2.10947052e-01 5.57510197e-01 2.71150380e-01 -4.40784037e-01
2.08626598e-01 1.31002188e-01 -1.25564015e+00 2.02590391e-01
1.22921801e+00 6.11117601e-01 3.45116973e-01 1.02243789e-01
-1.18470617e-01 8.64837825e-01 -4.31502730e-01 -2.34243140e-01
1.75431192e-01 -5.68397224e-01 -5.85862756e-01 2.47894049e-01
-3.22807997e-01 -5.99031687e-01 -5.48310935e-01 9.43440080e-01
3.56393397e-01 7.86059916e-01 7.52950907e-01 -8.90566766e-01
-1.23470283e+00 9.36295450e-01 -6.50238693e-01 1.77523538e-01
2.59552211e-01 3.52995694e-02 1.31533635e+00 -6.99373245e-01
4.04956013e-01 1.08472276e+00 2.76031494e-01 5.33659935e-01
-1.06243300e+00 -7.55721509e-01 1.05972433e+00 5.11492938e-02
-1.03613436e+00 -4.07709442e-02 9.66941178e-01 -3.17975312e-01
8.07728887e-01 8.15918669e-02 5.41448712e-01 1.20753753e+00
-3.54602098e-01 9.54592347e-01 5.82747161e-01 -3.88096645e-02
7.03564510e-02 1.52365312e-01 5.03413618e-01 2.66254544e-01
2.93582112e-01 1.30568445e-03 -1.34741724e-01 -2.30382040e-01
9.03622329e-01 5.65288723e-01 -3.15702468e-01 -3.33753467e-01
-7.96275198e-01 1.04227388e+00 3.75557512e-01 4.23582792e-01
-5.77285349e-01 -4.95519578e-01 2.88488418e-01 2.83562899e-01
5.90890884e-01 4.59877491e-01 -5.32220244e-01 3.25416736e-02
-7.24523544e-01 1.49541602e-01 7.76860118e-01 9.21216309e-01
7.02468634e-01 -6.01278134e-02 -3.27413976e-01 1.22894645e+00
3.70374113e-01 2.30924552e-03 7.00078011e-01 -4.86971915e-01
4.12242442e-01 9.44124341e-01 2.76430130e-01 -9.95659649e-01
-3.14434886e-01 -8.41283262e-01 -1.01035535e+00 -4.65240717e-01
1.59648299e-01 -2.81280786e-01 -6.26160979e-01 1.75620115e+00
1.43539771e-01 4.22677219e-01 -6.85429871e-02 1.23230982e+00
7.13094771e-01 6.35868728e-01 -1.45251839e-03 -4.36383367e-01
1.17142653e+00 -1.38852012e+00 -6.09564722e-01 -2.36137003e-01
4.21990007e-01 -5.53227305e-01 1.10901642e+00 4.24780250e-01
-8.35553825e-01 -6.91968322e-01 -9.72106278e-01 1.25802249e-01
-2.42051765e-01 3.79905015e-01 4.58581895e-01 3.42875332e-01
-7.16769457e-01 6.63055539e-01 -4.05396491e-01 -3.84157419e-01
6.31422624e-02 4.67976153e-01 1.71527177e-01 -1.23105176e-01
-1.65292668e+00 5.11021912e-01 2.13691622e-01 -5.79843335e-02
-6.13005102e-01 -4.79589075e-01 -5.18700421e-01 4.06795889e-01
7.72561967e-01 -7.17590332e-01 1.19235253e+00 -1.27555585e+00
-1.70911217e+00 1.42631218e-01 -9.94696841e-02 -2.06033394e-01
2.22745255e-01 -2.53850579e-01 -8.68988514e-01 -3.78387034e-01
-4.15281057e-02 -9.08920169e-02 7.45872080e-01 -1.23812449e+00
-9.10794258e-01 -2.87509680e-01 4.85718697e-01 5.75447917e-01
-6.45964921e-01 -1.56625047e-01 -9.60171521e-01 -9.63287234e-01
-6.91271573e-02 -1.01213980e+00 -2.35817507e-01 -8.26073170e-01
-2.86963969e-01 -4.30538893e-01 4.56567079e-01 -7.12912500e-01
1.79458666e+00 -2.20546317e+00 2.73499519e-01 3.67928237e-01
1.17801785e-01 4.82150912e-01 -3.00282896e-01 4.88641888e-01
-3.26445773e-02 5.18599451e-02 1.88156739e-01 -3.32326025e-01
2.65897699e-02 1.99917540e-01 -3.71834576e-01 2.43110791e-01
-2.60749251e-01 6.04498744e-01 -8.91577065e-01 9.25589129e-02
3.20562795e-02 3.59692454e-01 -1.00677550e+00 4.11206365e-01
-3.49805713e-01 5.02256274e-01 -7.61847794e-01 1.69582233e-01
7.28222072e-01 -7.44739175e-01 7.10236073e-01 -1.84271932e-01
-4.35016379e-02 5.22906184e-01 -1.34340405e+00 1.42293155e+00
-3.23507607e-01 -2.29180127e-01 -9.88557786e-02 -1.04454792e+00
9.29982305e-01 8.38890970e-02 6.09577417e-01 -8.70033979e-01
-3.05703990e-02 2.69930414e-03 8.22005570e-02 -2.34908387e-01
7.59793341e-01 1.83889419e-01 5.51446490e-02 8.57254148e-01
-1.80372402e-01 8.71322691e-01 -1.37757398e-02 3.84660542e-01
9.76536989e-01 2.24385172e-01 3.81751150e-01 -5.14579862e-02
8.38718772e-01 -2.90272295e-01 7.38510668e-01 8.54275346e-01
-3.90475839e-02 3.35132688e-01 3.64665121e-01 -2.99080938e-01
-5.61238348e-01 -5.66970944e-01 5.10143459e-01 1.69336474e+00
6.70501709e-01 -4.96087253e-01 -3.45491707e-01 -1.09488857e+00
6.54519498e-02 7.47655690e-01 -5.77823758e-01 -5.56380868e-01
-5.41400850e-01 -7.88857579e-01 -1.10074773e-01 4.85119432e-01
2.94515729e-01 -1.08476913e+00 2.39214912e-01 4.25402761e-01
-4.16112661e-01 -7.87568688e-01 -1.09708166e+00 -2.96825543e-02
-7.62365580e-01 -9.25769389e-01 -9.23909009e-01 -6.71873093e-01
6.49077058e-01 8.04248273e-01 1.15030456e+00 2.45852113e-01
7.20714927e-01 1.53301045e-01 -7.42548048e-01 2.09417075e-01
-5.63760735e-02 6.38979077e-01 1.92236915e-01 3.07554394e-01
5.14380991e-01 -7.33047426e-01 -1.04086030e+00 7.74742126e-01
-6.92682385e-01 -2.53609478e-01 7.31418669e-01 9.18653786e-01
3.29572171e-01 8.42951015e-02 9.12335873e-01 -1.46843052e+00
1.14656401e+00 -7.77182877e-01 -2.89867550e-01 2.92093158e-01
-7.20081568e-01 -6.08198270e-02 1.05502713e+00 -7.69089818e-01
-1.10972309e+00 -2.08667189e-01 -1.92707613e-01 -4.36494589e-01
4.29327786e-02 7.03962862e-01 -2.13459373e-01 5.21184087e-01
3.36500615e-01 4.64399487e-01 -2.86331832e-01 -7.05241263e-01
2.90519089e-01 8.27170432e-01 9.84120667e-02 -5.16993165e-01
5.98168671e-01 1.23224877e-01 -9.84351039e-01 -4.16456312e-01
-1.05624032e+00 -9.31006312e-01 -2.93940634e-01 -5.34138158e-02
5.44508457e-01 -9.80593324e-01 -8.63072455e-01 1.96028426e-02
-7.91454732e-01 -3.56009245e-01 3.47808491e-05 5.50133824e-01
-2.46548384e-01 4.79939461e-01 -6.85515046e-01 -7.14696705e-01
-4.02511656e-01 -1.06268454e+00 6.20057940e-01 2.74586558e-01
-2.20227823e-01 -8.74927580e-01 1.82412684e-01 4.08432364e-01
3.70387584e-01 -5.87099910e-01 1.11705375e+00 -1.14994633e+00
-4.62793142e-01 8.81711245e-02 -3.36185247e-01 1.95163071e-01
4.08178121e-01 -2.77319551e-01 -4.47196245e-01 -5.47722995e-01
-2.24105746e-01 8.45772699e-02 8.26361954e-01 2.90076226e-01
1.19584870e+00 -3.51760536e-01 -4.04816568e-01 2.39953220e-01
1.17610550e+00 3.59710187e-01 4.36100930e-01 2.62450635e-01
7.84173012e-01 4.35711801e-01 6.65260077e-01 6.15286112e-01
6.99445367e-01 7.29815066e-01 3.08389813e-01 -1.34463608e-01
1.33691192e-01 -5.20853043e-01 4.79812056e-01 1.04188871e+00
-3.12593579e-01 -5.28436661e-01 -2.58558661e-01 3.70720059e-01
-2.33456683e+00 -8.32274735e-01 2.12310404e-01 2.23273063e+00
4.53272313e-01 2.76023299e-01 7.28973567e-01 -2.20114380e-01
9.55590725e-01 1.13243245e-01 -8.68579984e-01 -7.62319714e-02
1.91811368e-01 -3.75294417e-01 2.83429980e-01 2.92021036e-01
-9.90635097e-01 1.05706859e+00 5.23357058e+00 8.15389693e-01
-1.00559270e+00 -9.81155932e-02 3.67584586e-01 6.98617026e-02
-4.40690160e-01 -1.60377935e-01 -7.80625939e-01 7.62542546e-01
6.44704700e-01 -2.85741817e-02 8.45408738e-01 7.89553940e-01
3.83473784e-01 3.05350661e-01 -7.80651689e-01 7.25692630e-01
3.67238037e-02 -9.74753320e-01 3.73433083e-01 2.46143520e-01
6.94892764e-01 -6.37545139e-02 5.59096858e-02 9.13036048e-01
7.93367267e-01 -4.71789062e-01 3.49586964e-01 5.60823143e-01
3.49034607e-01 -1.05025613e+00 5.43221176e-01 6.98946893e-01
-1.29585063e+00 -4.28667307e-01 -6.22383654e-01 -2.33951700e-03
-2.05011852e-03 3.79082888e-01 -3.27376485e-01 9.30487454e-01
4.34521705e-01 1.29300892e+00 -2.72337139e-01 7.38437176e-01
-1.59177840e-01 7.05200374e-01 1.04949012e-01 -4.82819648e-03
2.21437722e-01 -6.09227657e-01 4.49345380e-01 9.57024753e-01
3.58726352e-01 4.32718486e-01 6.38317227e-01 5.85421026e-01
-3.17768455e-01 3.19299161e-01 -2.96604335e-01 1.21418405e-02
4.40303802e-01 1.20249987e+00 -5.83390117e-01 -3.99557918e-01
-7.84734368e-01 1.07986426e+00 5.43762743e-01 6.09789014e-01
-8.39770257e-01 -6.04489632e-02 7.89629042e-01 1.72714829e-01
6.58545911e-01 8.72379839e-02 5.27286530e-03 -1.55326760e+00
-2.78086871e-01 -1.11804903e+00 7.02128887e-01 -3.70746404e-01
-1.60921586e+00 4.77764279e-01 -4.78826046e-01 -1.43845284e+00
-7.93777779e-02 -2.03488722e-01 -6.06777847e-01 7.53675997e-01
-1.71886468e+00 -1.01616228e+00 4.41682748e-02 9.40451086e-01
5.98722398e-01 -3.05885136e-01 6.03098631e-01 7.47247875e-01
-9.77096438e-01 8.73982430e-01 4.33575928e-01 1.35201529e-01
8.72424901e-01 -9.57022548e-01 1.90563187e-01 6.47827864e-01
-1.44099131e-01 1.07036090e+00 4.30142105e-01 -7.79486656e-01
-1.33734691e+00 -1.35303187e+00 5.62801898e-01 -1.91386789e-01
5.02538979e-01 -2.26034150e-01 -1.16417038e+00 7.51255929e-01
-2.75533963e-02 -3.38647157e-01 8.20613801e-01 6.38917983e-01
-2.54720718e-01 -2.99322419e-02 -7.01202095e-01 7.62219071e-01
1.28184688e+00 -3.29481214e-01 -6.72558188e-01 1.54563770e-01
8.31076026e-01 -4.14691865e-02 -7.65590370e-01 2.70965725e-01
5.18098235e-01 -8.72886121e-01 1.05558765e+00 -7.07718670e-01
4.40357625e-01 -2.65122980e-01 1.03270069e-01 -1.61354017e+00
-8.89724493e-01 -6.55184507e-01 -4.17045981e-01 1.43234313e+00
3.05434138e-01 -6.64773047e-01 7.27417767e-01 4.08779889e-01
-1.47552207e-01 -5.72285771e-01 -1.92550823e-01 -6.34003818e-01
-1.02853164e-01 4.22713086e-02 9.29222941e-01 1.11438560e+00
1.73646823e-01 8.15295637e-01 -8.97451639e-01 2.45266870e-01
1.16698660e-01 6.68314993e-01 6.67258561e-01 -1.40094125e+00
-6.35596454e-01 -4.22979474e-01 5.21353900e-01 -1.72040296e+00
2.20222086e-01 -8.17481697e-01 -1.74351007e-01 -1.44655430e+00
2.30074361e-01 -4.28140700e-01 -8.97436082e-01 2.16935903e-01
-3.64532948e-01 -4.56967056e-02 1.02438189e-01 5.49456179e-01
-1.05591679e+00 6.35728121e-01 1.47472870e+00 -4.19787206e-02
-7.63376355e-01 4.20126319e-01 -1.38947892e+00 7.01323330e-01
5.86997867e-01 -3.38142544e-01 -6.39002264e-01 -2.72918403e-01
2.55104214e-01 2.80565649e-01 -2.12874040e-01 -5.35942733e-01
4.27629918e-01 -2.38395736e-01 1.59757733e-01 -6.22898936e-01
5.24871089e-02 -9.15497661e-01 -8.65583196e-02 2.66579181e-01
-4.35140729e-01 -1.03465684e-01 -1.93736017e-01 9.76392269e-01
-7.90072754e-02 1.31592587e-01 4.91074443e-01 -1.86329231e-01
-7.06671417e-01 6.01326108e-01 -5.04653454e-01 -5.30888177e-02
7.10409343e-01 -5.72404154e-02 -1.66350096e-01 -5.52904725e-01
-8.69598508e-01 4.00476515e-01 2.93978602e-01 7.25445509e-01
3.59737158e-01 -1.34861624e+00 -5.30418813e-01 2.61032581e-01
2.49541357e-01 -3.83251965e-01 5.64943731e-01 4.49563116e-01
2.46714100e-01 3.41790557e-01 -8.38037673e-03 -6.44820854e-02
-9.37793314e-01 9.33855355e-01 1.61677510e-01 -8.11299682e-01
-4.73902911e-01 5.18161297e-01 4.70687270e-01 -6.18241072e-01
2.79503137e-01 -5.72879799e-02 -1.00925326e+00 2.47853398e-01
4.42397714e-01 2.97887266e-01 6.89802179e-03 -4.92131472e-01
-1.57568663e-01 3.42658341e-01 -6.81053817e-01 3.43995899e-01
1.38972247e+00 -4.85516965e-01 1.59868121e-01 1.23281255e-01
7.82140374e-01 4.73680086e-02 -1.20723712e+00 -7.14258194e-01
-1.96985349e-01 -3.01901370e-01 -1.27148509e-01 -8.79046142e-01
-1.14389765e+00 5.55727482e-01 1.54805005e-01 5.26218235e-01
1.15121710e+00 -1.96779460e-01 9.97897327e-01 2.25218475e-01
1.05742343e-01 -1.22884655e+00 2.39465550e-01 9.03707325e-01
6.15306556e-01 -1.04546118e+00 -1.02715082e-02 -3.64807785e-01
-9.39087093e-01 7.48718798e-01 8.77056062e-01 -3.48442435e-01
8.14162433e-01 -3.05838883e-01 -2.55386293e-01 1.01207688e-01
-6.99708819e-01 -3.44101280e-01 4.23988789e-01 3.43424708e-01
3.88339013e-01 1.54932449e-02 -4.72749114e-01 1.36906183e+00
1.50206447e-01 1.92903534e-01 3.80723715e-01 4.83272016e-01
-3.15445244e-01 -1.36124730e+00 1.07713141e-01 7.15026259e-01
-3.66270572e-01 -2.63897274e-02 -3.29582185e-01 5.06655335e-01
5.28523605e-03 1.01599777e+00 -2.07200631e-01 -7.17224717e-01
6.41728461e-01 -1.33470923e-01 -2.06594631e-01 -7.59881973e-01
-9.34575140e-01 4.81170237e-01 -1.16201364e-01 -2.88048506e-01
-4.21897531e-01 -5.24089992e-01 -9.74511266e-01 -4.76836503e-01
-4.81769770e-01 3.54395449e-01 -8.48129839e-02 1.01608181e+00
6.15758955e-01 9.89521444e-01 9.05137539e-01 -6.56687438e-01
-6.72157645e-01 -9.19270694e-01 -8.39262903e-01 5.79800606e-01
2.65686959e-01 -8.17150533e-01 -1.55012399e-01 -3.54478598e-01] | [10.169344902038574, 5.5833845138549805] |
ed508051-b0e7-4b95-8bb6-e63219927b7a | graph-meets-llm-a-novel-approach-to | 2305.14449 | null | https://arxiv.org/abs/2305.14449v3 | https://arxiv.org/pdf/2305.14449v3.pdf | Graph Meets LLM: A Novel Approach to Collaborative Filtering for Robust Conversational Understanding | Conversational AI systems such as Alexa need to understand defective queries to ensure robust conversational understanding and reduce user friction. These defective queries often arise from user ambiguities, mistakes, or errors in automatic speech recognition (ASR) and natural language understanding (NLU). Personalized query rewriting is an approach that focuses on reducing defects in queries by taking into account the user's individual behavior and preferences. It typically relies on an index of past successful user interactions with the conversational AI. However, unseen interactions within the user's history present additional challenges for personalized query rewriting. This paper presents our "Collaborative Query Rewriting" approach, which specifically addresses the task of rewriting new user interactions that have not been previously observed in the user's history. This approach builds a "User Feedback Interaction Graph" (FIG) of historical user-entity interactions and leverages multi-hop graph traversal to enrich each user's index to cover future unseen defective queries. The enriched user index is called a Collaborative User Index and contains hundreds of additional entries. To counteract precision degradation from the enlarged index, we add additional transformer layers to the L1 retrieval model and incorporate graph-based and guardrail features into the L2 ranking model. Since the user index can be pre-computed, we further investigate the utilization of a Large Language Model (LLM) to enhance the FIG for user-entity link prediction in the Video/Music domains. Specifically, this paper investigates the Dolly-V2 7B model. We found that the user index augmented by the fine-tuned Dolly-V2 generation significantly enhanced the coverage of future unseen user interactions, thereby boosting QR performance on unseen queries compared with the graph traversal only approach. | ['Aram Galstyan', 'Yanbin Lu', 'Xiaojiang Huang', 'Xing Fan', 'Eunah Cho', 'Fan Yang', 'Ziyan Jiang', 'Zheng Chen'] | 2023-05-23 | null | null | null | null | ['link-prediction', 'collaborative-filtering', 'automatic-speech-recognition'] | ['graphs', 'miscellaneous', 'speech'] | [ 3.90003920e-01 3.44117135e-01 -8.80050659e-02 -2.98011750e-01
-1.14913678e+00 -7.65097857e-01 6.72177911e-01 2.09912986e-01
-3.19631159e-01 3.97461325e-01 6.87536418e-01 -2.60162324e-01
-3.31951499e-01 -6.02750003e-01 -5.83479643e-01 1.98797807e-01
-4.05333452e-02 7.85549760e-01 2.76273638e-01 -9.47672129e-01
3.52811292e-02 2.64461875e-01 -1.59649444e+00 5.51366150e-01
1.02559328e+00 5.56542516e-01 3.43336523e-01 1.24200630e+00
-8.11595768e-02 6.75031722e-01 -8.67069423e-01 -4.15204108e-01
3.58007429e-03 -3.82076770e-01 -1.07210958e+00 -7.14669898e-02
2.86214590e-01 -5.52499235e-01 -7.42044866e-01 5.16143501e-01
5.78368068e-01 6.86425626e-01 2.01818928e-01 -1.19862998e+00
-5.73164523e-01 7.75584042e-01 2.06865713e-01 4.36400443e-01
1.08400309e+00 5.51291779e-02 1.44976008e+00 -7.18869865e-01
7.20162272e-01 1.36680019e+00 2.71149546e-01 6.62430525e-01
-9.27907348e-01 -2.82246321e-01 3.84222537e-01 2.14037403e-01
-1.29105484e+00 -6.53910100e-01 5.84204912e-01 -1.19661137e-01
1.36416578e+00 8.46978247e-01 4.94794309e-01 1.18613982e+00
-3.09800863e-01 1.11245811e+00 -4.25526239e-02 -4.38560754e-01
-1.81728423e-01 5.13179488e-02 3.71295005e-01 7.22185910e-01
-3.30854356e-01 -2.23312043e-02 -8.89575005e-01 -4.98445481e-01
1.75223619e-01 -1.22109666e-01 -5.18344045e-01 3.26942727e-02
-9.48711276e-01 6.15882039e-01 1.42419592e-01 3.82556587e-01
-2.40429029e-01 -1.34591952e-01 2.42028162e-01 5.16787648e-01
2.28265598e-01 1.04215789e+00 -6.00976288e-01 -5.00783741e-01
-4.45557952e-01 3.03072631e-01 1.18548274e+00 1.19907606e+00
8.25022936e-01 -3.59062225e-01 -7.94442773e-01 1.24924278e+00
-1.05474852e-02 4.02960390e-01 3.82138669e-01 -1.05555737e+00
5.60601115e-01 9.66938853e-01 1.32237047e-01 -1.06992030e+00
-1.63974538e-01 -5.64787090e-01 -2.16689631e-01 -9.09558117e-01
-9.99256689e-03 -2.18160097e-02 -7.79211402e-01 1.91418421e+00
2.03779429e-01 -5.17134415e-03 2.56831884e-01 5.01771986e-01
8.13436687e-01 6.40402198e-01 -2.41986319e-01 -3.31407309e-01
1.07767284e+00 -1.05395579e+00 -5.49596369e-01 -2.95764983e-01
9.41193700e-01 -7.13735461e-01 1.34115672e+00 1.68605506e-01
-8.76350820e-01 -4.68433499e-01 -7.68078983e-01 -2.22970322e-01
-2.91277438e-01 -5.50791286e-02 5.31153798e-01 5.31413317e-01
-1.18741059e+00 4.49459404e-01 -4.42134053e-01 -5.60003161e-01
-1.97385848e-01 4.19746369e-01 9.39564873e-03 -2.44770765e-01
-1.62600362e+00 7.27351546e-01 4.46974784e-02 -1.03722885e-01
-8.89407098e-01 -6.88283682e-01 -8.28686357e-01 2.18399271e-01
9.80443835e-01 -7.82397091e-01 1.63108957e+00 -2.50831425e-01
-1.32479656e+00 3.16620469e-01 -7.29607880e-01 -2.78324574e-01
1.05554871e-01 -2.83679903e-01 -5.24501920e-01 6.91239461e-02
-7.19110891e-02 3.85147095e-01 5.35300612e-01 -1.13606930e+00
-8.07720482e-01 -3.01029235e-01 6.12866879e-01 7.09382594e-01
-3.24138075e-01 -3.01145971e-01 -1.16612470e+00 -6.35056913e-01
-5.01245111e-02 -1.09871674e+00 1.96131431e-02 -7.77965248e-01
-3.61287504e-01 -5.36972106e-01 7.66034305e-01 -7.76339650e-01
1.87647355e+00 -1.94048893e+00 2.07473576e-01 5.44729888e-01
2.65005887e-01 2.45058179e-01 -4.55032796e-01 8.76714051e-01
3.91077846e-01 2.24926189e-01 1.28929481e-01 -3.89740765e-01
-8.67327303e-02 4.30643737e-01 -1.80688605e-01 -3.99070770e-01
-2.55641490e-01 1.15996027e+00 -9.69647944e-01 -2.25319564e-01
-1.32951438e-01 2.21160978e-01 -1.05985677e+00 5.53441703e-01
-5.79838634e-01 3.75468671e-01 -5.81374705e-01 6.15587473e-01
-2.29827091e-01 -3.29610467e-01 1.36969164e-01 -1.44545555e-01
3.23706448e-01 5.08805633e-01 -7.71635592e-01 1.68379760e+00
-5.92050672e-01 4.00552809e-01 -3.72634456e-02 -4.68881965e-01
6.03712976e-01 4.03753519e-01 5.86453855e-01 -8.50710332e-01
-1.66600004e-01 5.68157062e-03 -1.06095687e-01 -5.67498446e-01
8.66522014e-01 6.59290254e-01 -3.07722628e-01 4.83159125e-01
-3.97435762e-02 -1.68377031e-02 4.27314103e-01 8.01269352e-01
1.58942175e+00 -4.61901516e-01 9.76089295e-03 3.32449049e-01
7.59124398e-01 -4.70697358e-02 2.49926627e-01 1.26863992e+00
-3.89364325e-02 3.13559234e-01 2.67248899e-01 6.77663535e-02
-5.92769146e-01 -7.55608439e-01 5.30349135e-01 1.78419590e+00
1.75171912e-01 -9.23525393e-01 -6.40795827e-01 -5.97092271e-01
5.86473532e-02 9.96360898e-01 -2.46813461e-01 -6.92856550e-01
-5.61419487e-01 -2.71867812e-01 7.31449485e-01 1.12903357e-01
2.79230267e-01 -9.69709992e-01 6.93409517e-02 3.12857628e-01
-7.54976094e-01 -1.22615695e+00 -1.15777230e+00 -3.50850999e-01
-3.79058480e-01 -1.08557558e+00 -2.82781184e-01 -3.59851599e-01
3.08622509e-01 3.88717204e-01 1.26228082e+00 4.29870069e-01
-2.08352536e-01 1.22963727e+00 -8.06924343e-01 -1.27330720e-01
-7.05537081e-01 3.92987728e-01 2.39257112e-01 1.31310411e-02
2.61267990e-01 -4.41998243e-01 -5.76085031e-01 5.16871333e-01
-7.01522052e-01 -1.92864120e-01 3.49195689e-01 8.31584394e-01
1.69623330e-01 -1.29064903e-01 6.79040670e-01 -9.72982883e-01
1.23218334e+00 -3.76378685e-01 -1.37314707e-01 8.13115418e-01
-8.20553064e-01 2.72199392e-01 3.99295300e-01 -5.21224618e-01
-1.07822156e+00 -2.21057460e-01 -2.80722469e-01 -2.91977078e-01
2.97935128e-01 6.33971393e-01 -1.91199780e-01 3.66539508e-02
7.34820187e-01 2.90846914e-01 -2.46771723e-01 -4.07439798e-01
4.68779296e-01 8.16884220e-01 6.62877023e-01 -6.28475845e-01
4.11941528e-01 -1.80876330e-01 -6.94962740e-01 -9.99883354e-01
-7.71428108e-01 -9.73986268e-01 -9.92720649e-02 -4.41581935e-01
6.28710210e-01 -6.17157578e-01 -1.15115023e+00 -1.11730965e-02
-1.26173341e+00 -1.25021145e-01 -1.20168477e-01 2.62680829e-01
-4.21967000e-01 5.70565641e-01 -5.17194867e-01 -1.05523372e+00
-4.47999835e-01 -1.22175014e+00 1.16558468e+00 1.84898898e-01
-7.10020304e-01 -6.71436429e-01 -4.94463602e-03 8.85703623e-01
2.42440209e-01 -4.64184731e-01 1.11670554e+00 -1.23314261e+00
-7.45027900e-01 -4.82764035e-01 1.70053244e-01 1.95821881e-01
1.46969482e-01 -3.81998897e-01 -9.16252613e-01 -4.53229159e-01
-2.17384338e-01 -3.31066805e-03 4.76287752e-01 -1.69126853e-01
8.53604376e-01 -5.72488546e-01 -4.04343128e-01 9.67179015e-02
6.74265981e-01 5.14995992e-01 2.96335608e-01 -2.32607007e-01
7.11671472e-01 5.52171350e-01 5.00417233e-01 5.47586918e-01
6.54332459e-01 8.58184457e-01 9.93275866e-02 3.62925053e-01
-5.59584051e-02 -7.02008188e-01 3.93704146e-01 9.36143100e-01
2.38278881e-01 -8.96968305e-01 -8.95384550e-01 5.43945253e-01
-1.76246655e+00 -8.13682795e-01 2.74036437e-01 2.19990611e+00
7.53023922e-01 1.59190625e-01 -6.54084012e-02 -1.25947744e-01
5.55170894e-01 2.39598025e-02 -7.02141047e-01 -2.84692436e-01
7.96804354e-02 6.37518391e-02 3.11834395e-01 1.11219645e+00
-4.62035090e-01 1.08023155e+00 5.91951036e+00 6.81655169e-01
-5.19655406e-01 8.19864310e-03 2.92899877e-01 -6.70329854e-02
-6.71068966e-01 -1.17060989e-02 -1.04157424e+00 1.95649505e-01
9.53440666e-01 -4.99616176e-01 9.75301504e-01 5.47779381e-01
1.02023631e-01 -5.63829765e-02 -1.40774977e+00 9.05329406e-01
2.39254907e-01 -1.22619331e+00 4.35552061e-01 -5.59153520e-02
3.26891720e-01 -6.17557578e-03 -1.87959731e-01 9.39675629e-01
3.86850417e-01 -6.50325477e-01 -2.43063215e-02 8.54807138e-01
5.82842827e-01 -6.85129583e-01 3.75473976e-01 7.11144984e-01
-1.16889131e+00 -3.47769737e-01 1.85508341e-01 2.75086373e-01
1.86779529e-01 9.73088741e-02 -1.43622494e+00 5.63981652e-01
5.17170131e-01 2.66397744e-01 -5.93377888e-01 5.66781163e-01
2.77665228e-01 3.64994407e-01 -4.66285676e-01 -1.15661077e-01
7.97039047e-02 1.00701228e-01 1.08742177e+00 9.11119401e-01
1.96958005e-01 5.04248798e-01 4.36407864e-01 4.10776049e-01
-3.46582681e-01 1.89261347e-01 -7.29868531e-01 -4.38712060e-01
8.21419179e-01 7.95597255e-01 1.30466923e-01 -4.21753228e-01
-3.23413342e-01 1.23846996e+00 2.70629108e-01 5.36974132e-01
-4.15132493e-01 -1.89217836e-01 7.74850368e-01 5.38144149e-02
-1.45774573e-01 -8.75998288e-02 3.14885765e-01 -1.11195457e+00
1.28319189e-01 -1.27014661e+00 7.49636829e-01 -6.73089802e-01
-1.00517571e+00 7.05903590e-01 8.25812593e-02 -6.62534237e-01
-8.59722555e-01 5.92924245e-02 -3.73977125e-01 6.91602886e-01
-8.89568985e-01 -8.52597296e-01 -1.52449206e-01 6.28106534e-01
9.68978286e-01 -3.92801166e-01 9.50874925e-01 5.04317999e-01
-4.83738571e-01 1.13997293e+00 -3.08006197e-01 -1.94122702e-01
5.73569536e-01 -8.93252492e-01 4.89441693e-01 7.40587533e-01
3.12435597e-01 8.82387877e-01 6.03492498e-01 -9.74620163e-01
-1.58276284e+00 -8.24855149e-01 1.37898135e+00 -8.86865258e-01
4.05253172e-01 -3.31291556e-01 -1.00303245e+00 7.42525637e-01
-1.25261456e-01 -6.55092120e-01 7.04641223e-01 3.30445379e-01
-9.41244736e-02 -7.52414018e-02 -8.15252841e-01 9.46787298e-01
1.47123289e+00 -9.77810562e-01 -6.67579710e-01 4.50135857e-01
1.35367584e+00 -5.01133263e-01 -8.41432154e-01 5.88192046e-01
3.71336311e-01 -4.78658259e-01 1.13842428e+00 -7.84362733e-01
-2.84748197e-01 5.36681106e-03 -2.78495014e-01 -1.13682735e+00
-9.60533172e-02 -1.25785172e+00 -4.84142989e-01 1.15932918e+00
7.41210580e-01 -3.95431638e-01 7.32714593e-01 1.11600482e+00
-3.72562349e-01 -7.08262563e-01 -6.25173867e-01 -4.13298815e-01
-8.02011549e-01 -5.78537166e-01 6.78375065e-01 5.75258851e-01
1.39530882e-01 7.16497719e-01 -4.26058710e-01 3.93062562e-01
1.24230660e-01 -2.51641870e-01 8.97914052e-01 -1.13321412e+00
-4.61608142e-01 -3.06334853e-01 1.12242378e-01 -1.61672413e+00
7.94417709e-02 -9.18258429e-01 -2.67321393e-02 -1.39745426e+00
-1.41681686e-01 -3.04308951e-01 3.02714240e-02 3.21738839e-01
-2.59258330e-01 -3.35103244e-01 9.19296294e-02 3.05995226e-01
-8.22476447e-01 6.04762137e-01 1.23717237e+00 -1.72839776e-01
-7.18732297e-01 4.33308810e-01 -7.45495319e-01 3.45103830e-01
4.77576286e-01 -1.77238092e-01 -9.58793044e-01 -4.88367647e-01
3.03765565e-01 5.34742653e-01 -9.77916420e-02 -6.92479670e-01
7.76118696e-01 4.48371656e-02 -3.03889453e-01 -6.52243853e-01
6.70771003e-01 -6.51812315e-01 -9.86818820e-02 7.93673918e-02
-7.69972980e-01 2.40279526e-01 9.81796440e-03 9.73311007e-01
-2.42115900e-01 8.98921341e-02 1.27477303e-01 -1.71053484e-01
-7.06445277e-01 3.97187948e-01 -4.11254048e-01 1.99365318e-01
5.71115434e-01 -2.33632140e-02 -1.60753176e-01 -1.08042955e+00
-1.03997588e+00 8.23589921e-01 2.72629652e-02 8.70745480e-01
6.90871358e-01 -1.04655516e+00 -4.33946282e-01 1.62337914e-01
5.60545325e-01 -1.84592322e-01 2.68743068e-01 3.98094296e-01
6.87083825e-02 6.79653883e-01 5.93720853e-01 -4.20856774e-01
-1.43818688e+00 2.57780671e-01 3.08259964e-01 -4.48321700e-01
-3.05807531e-01 1.06246412e+00 -7.66966268e-02 -6.42013907e-01
7.49163091e-01 -1.77458644e-01 -3.30418646e-01 6.15066811e-02
3.53967816e-01 4.42324013e-01 1.87749982e-01 -5.43175161e-01
-2.37374395e-01 -6.68621138e-02 -4.60743487e-01 -2.52402037e-01
8.93281221e-01 -5.15038908e-01 6.06498495e-02 3.39841843e-01
1.32014835e+00 2.09279567e-01 -5.79503775e-01 -5.71014822e-01
2.80302137e-01 -3.89498591e-01 -2.89145917e-01 -9.48066235e-01
-5.25718689e-01 2.27154166e-01 1.43603861e-01 2.93248773e-01
9.23518777e-01 1.82247102e-01 1.25775814e+00 1.12075508e+00
3.97043884e-01 -1.16950870e+00 3.52890491e-01 9.61664259e-01
1.08410621e+00 -1.11390638e+00 -3.82195443e-01 -3.75006527e-01
-8.09065998e-01 5.88975608e-01 7.90514290e-01 6.82783961e-01
3.39006633e-01 -2.96136111e-01 -7.75805628e-03 -2.77935535e-01
-1.09965611e+00 -4.88515615e-01 7.47079730e-01 2.93431580e-01
2.56679803e-01 -1.13953702e-01 -1.31241113e-01 6.87992275e-01
-5.12245059e-01 -2.27457255e-01 1.87920079e-01 8.80214155e-01
-3.04651052e-01 -1.32220566e+00 8.83920267e-02 7.08773792e-01
-2.68030111e-02 -3.35081428e-01 -7.75705814e-01 3.81752938e-01
-2.90722251e-01 1.44092035e+00 -2.01382563e-01 -9.57750261e-01
8.31663132e-01 3.69413227e-01 1.19408131e-01 -9.36848164e-01
-7.99249947e-01 -1.94971621e-01 6.66508377e-01 -7.89087236e-01
1.89700857e-01 -3.74001384e-01 -1.14907110e+00 -2.47035831e-01
-4.47096378e-01 4.56133872e-01 2.94026494e-01 9.51502681e-01
7.72902310e-01 4.81023371e-01 7.04181850e-01 -2.31649563e-01
-5.77173173e-01 -8.88156891e-01 -4.13595932e-03 5.44254363e-01
3.91116410e-01 -4.88428295e-01 -2.58985847e-01 -2.62220025e-01] | [12.071883201599121, 7.778079986572266] |
2af2c270-5427-4786-ad78-fe87c9c85ade | learning-graph-structure-for-multi-label | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Tan_Learning_Graph_Structure_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Tan_Learning_Graph_Structure_2015_CVPR_paper.pdf | Learning Graph Structure for Multi-Label Image Classification via Clique Generation | Exploiting label dependency for multi-label image classification can significantly improve classification performance. Probabilistic Graphical Models are one of the primary methods for representing such dependencies. The structure of graphical models, however, is either determined heuristically or learned from very limited information. Moreover, neither of these approaches scales well to large or complex graphs. We propose a principled way to learn the structure of a graphical model by considering input features and labels, together with loss functions. We formulate this problem into a max-margin framework initially, and then transform it into a convex programming problem. Finally, we propose a highly scalable procedure that activates a set of cliques iteratively. Our approach exhibits both strong theoretical properties and a significant performance improvement over state-of-the-art methods on both synthetic and real-world data sets. | ['Anton Van Den Hengel', 'Zhen Zhang', 'Fuyuan Hu', 'Chunhua Shen', 'Qinfeng Shi', 'Mingkui Tan', 'Junbin Gao'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['multi-label-image-classification'] | ['computer-vision'] | [ 5.25174141e-01 1.34441108e-01 -7.12293148e-01 -6.54182315e-01
-1.07935524e+00 -6.95171118e-01 3.60114872e-01 4.17412221e-01
-2.77519882e-01 7.69395053e-01 -2.74905622e-01 -3.78126770e-01
-2.20072106e-01 -5.21611273e-01 -6.44593716e-01 -7.17664540e-01
-3.28633860e-02 7.15559185e-01 3.32409918e-01 4.01353538e-01
7.05935806e-02 4.63165104e-01 -1.18969107e+00 3.44975144e-01
8.72642636e-01 7.69768298e-01 -7.91968703e-02 5.60488224e-01
-2.38374174e-02 8.26302171e-01 -1.29277274e-01 -4.74543363e-01
7.29272515e-02 -4.50120956e-01 -9.99621093e-01 5.41277945e-01
3.34773660e-01 2.18221694e-01 -1.48815587e-01 9.84858930e-01
3.78217287e-02 -3.36621076e-01 8.45148861e-01 -1.59632385e+00
-2.60862827e-01 6.14320040e-01 -1.04071450e+00 -2.37001285e-01
1.81538686e-01 -3.27304304e-01 1.39902854e+00 -6.51383519e-01
5.32014430e-01 1.15084767e+00 5.07980347e-01 1.94077015e-01
-1.78387451e+00 -5.64596415e-01 4.87774998e-01 1.09732985e-01
-1.29562378e+00 -2.67699808e-01 7.99431086e-01 -5.53199172e-01
4.38015431e-01 3.16337675e-01 4.55644459e-01 7.15161562e-01
-1.54055417e-01 9.52303767e-01 1.44993198e+00 -6.86659992e-01
1.14081271e-01 3.25928897e-01 4.20299262e-01 1.26517212e+00
2.23893017e-01 -2.96513230e-01 -4.22927678e-01 -7.15730667e-01
5.60757518e-01 -2.53941178e-01 -2.19714522e-01 -8.65403235e-01
-8.89676869e-01 1.15666056e+00 3.12456638e-01 -1.31052166e-01
-9.50521231e-03 3.43815863e-01 1.35380387e-01 2.02278450e-01
3.90691638e-01 1.69630453e-01 -4.98331100e-01 4.34909612e-01
-8.18225741e-01 -9.34148952e-02 9.35547054e-01 8.34676445e-01
8.91732991e-01 -7.21155941e-01 6.01490848e-02 9.23607767e-01
8.09010863e-01 1.96784571e-01 -1.77221447e-01 -6.52613103e-01
4.43974644e-01 6.94716811e-01 -6.73242658e-02 -8.91096234e-01
-5.42461753e-01 -3.43530118e-01 -6.39936626e-01 1.75916374e-01
6.86607778e-01 -6.18022308e-02 -9.35498059e-01 1.84001660e+00
4.07043755e-01 1.21078894e-01 -2.68311530e-01 6.20634675e-01
5.00655532e-01 3.58806431e-01 7.14730918e-02 -4.82674062e-01
1.24911880e+00 -1.12892807e+00 -5.16216278e-01 -4.69398975e-01
7.33482778e-01 -4.68723178e-01 6.67184591e-01 4.26052958e-01
-6.96804523e-01 5.36985174e-02 -9.03717041e-01 1.37775704e-01
-2.75235297e-03 2.30851591e-01 9.34941530e-01 9.13567960e-01
-8.75938833e-01 4.26005542e-01 -8.05931926e-01 -9.77506340e-02
5.53486347e-01 6.51286542e-01 -4.93479371e-01 -3.92097801e-01
-7.57517219e-01 5.74678421e-01 4.51101422e-01 -3.85087244e-02
-5.23223281e-01 -2.20682204e-01 -9.30776298e-01 -1.17078602e-01
7.28515506e-01 -6.64817870e-01 1.01149189e+00 -8.55892122e-01
-1.24131918e+00 1.08053148e+00 -2.47101843e-01 -1.73838794e-01
4.41149116e-01 1.00678377e-01 -7.89797753e-02 1.82866603e-01
-8.72430354e-02 5.56449652e-01 8.23781371e-01 -1.57262301e+00
-6.16100609e-01 -3.78232479e-01 2.18570709e-01 1.01822481e-01
-2.83179253e-01 3.46322000e-01 -7.95153022e-01 -3.53568166e-01
3.74333739e-01 -1.31134188e+00 -5.16168118e-01 1.59408972e-01
-8.80205631e-01 -2.08470032e-01 5.33981383e-01 -2.58330643e-01
1.21617794e+00 -1.71947670e+00 2.49397889e-01 6.76713943e-01
4.34146285e-01 -7.83871040e-02 8.50740541e-03 3.06658924e-01
-1.16143398e-01 3.08679521e-01 -4.70517635e-01 -3.32053244e-01
-4.63100933e-02 3.26936632e-01 9.06393453e-02 7.10356295e-01
1.34189054e-01 8.25056016e-01 -8.97589445e-01 -8.27201486e-01
9.98238195e-03 2.54347205e-01 -6.46332145e-01 -1.64867043e-02
-2.79757172e-01 4.63612020e-01 -6.17015481e-01 5.67185044e-01
7.24762261e-01 -9.31882083e-01 8.80637050e-01 -1.15806863e-01
2.98101395e-01 1.87890921e-02 -1.41794240e+00 1.58938110e+00
-2.18238100e-01 3.20991904e-01 4.70814109e-02 -1.15751767e+00
5.43742597e-01 5.24665415e-02 6.94012761e-01 7.83841964e-03
1.22514725e-01 -1.15015525e-02 -2.24439874e-01 -3.22173357e-01
-7.62730464e-02 -2.39777505e-01 -1.39717296e-01 5.21182775e-01
4.02272753e-02 8.57372135e-02 2.64380604e-01 2.14668766e-01
1.22152770e+00 -1.17767259e-01 5.55331945e-01 -1.24766216e-01
3.35634887e-01 -1.79601178e-01 6.54146373e-01 7.97901690e-01
-2.93561686e-02 6.22697651e-01 1.04289544e+00 -2.32586637e-01
-5.44325650e-01 -8.74258935e-01 -2.48001620e-01 1.15565789e+00
1.66460708e-01 -6.02961361e-01 -6.64255440e-01 -1.25397861e+00
8.20731446e-02 3.83256108e-01 -6.00871086e-01 1.32128641e-01
-3.44162852e-01 -1.21209967e+00 2.40815863e-01 3.22634399e-01
4.70514894e-02 -5.26054144e-01 -1.90878779e-01 1.11585416e-01
-1.37760639e-01 -1.32846332e+00 -4.04077649e-01 5.06196558e-01
-8.74136508e-01 -1.52194333e+00 -2.36817122e-01 -8.13519597e-01
1.07844079e+00 2.03420267e-01 1.21449125e+00 2.67259657e-01
-1.99248374e-01 1.52436376e-01 -1.35689378e-01 -1.09322993e-02
-3.63014072e-01 1.99278191e-01 -2.84993291e-01 3.09725255e-01
-3.41665335e-02 -3.95773411e-01 -3.21872860e-01 4.94772881e-01
-8.44666362e-01 3.15752715e-01 8.25769186e-01 9.14255857e-01
9.18048084e-01 1.67203188e-01 4.45433348e-01 -1.75123739e+00
3.51102144e-01 -3.45182776e-01 -6.87168837e-01 6.98599577e-01
-8.49870741e-01 4.25099432e-01 3.70866239e-01 -2.84121454e-01
-7.63049364e-01 8.05264115e-01 4.97661494e-02 -2.45415226e-01
-2.25219697e-01 7.27343917e-01 -2.38806635e-01 -3.99212569e-01
4.75540251e-01 -2.04730302e-01 -2.00538874e-01 -4.31910306e-01
5.37249506e-01 4.87190187e-01 9.33181867e-02 -6.89187050e-01
7.97578394e-01 6.00900114e-01 3.72605264e-01 -4.87063259e-01
-1.24892676e+00 -7.86131442e-01 -9.98883009e-01 -1.17748402e-01
5.34460247e-01 -7.06829071e-01 -6.53410792e-01 3.78855765e-01
-9.94119942e-01 -2.88477063e-01 2.94484407e-01 3.31098914e-01
-5.22735775e-01 4.54866827e-01 -6.41672671e-01 -6.59720540e-01
1.83869109e-01 -1.18059921e+00 1.03426170e+00 8.79210308e-02
9.77708772e-02 -1.14530122e+00 1.32314265e-01 6.74660802e-01
-2.93964684e-01 5.34557462e-01 1.19098198e+00 -6.95833325e-01
-6.39676213e-01 -3.27530086e-01 -4.44558412e-01 -2.67661754e-02
4.99004796e-02 -1.43820010e-02 -9.84090388e-01 -4.13307697e-01
-3.85714144e-01 -6.91792548e-01 1.06282187e+00 2.98174202e-01
1.28879201e+00 -3.23097408e-01 -8.72616947e-01 4.19091731e-01
1.54640687e+00 -2.72609353e-01 3.45877886e-01 -1.97754186e-02
8.57558906e-01 5.16620636e-01 4.74141568e-01 2.45138153e-01
5.61268330e-01 8.31371248e-01 5.31624913e-01 -2.44168445e-01
1.70657039e-01 -2.25596189e-01 -5.40553220e-02 6.79636717e-01
2.63039142e-01 -4.82089728e-01 -1.01713097e+00 2.33770996e-01
-2.22462916e+00 -5.86247444e-01 -4.13894266e-01 2.12759948e+00
1.06602883e+00 2.24058613e-01 1.68104723e-01 1.38398096e-01
8.17224562e-01 5.06458171e-02 -5.10569394e-01 -9.34806243e-02
-7.18490854e-02 -1.61651650e-03 6.05914831e-01 4.01414096e-01
-1.44408190e+00 7.49132514e-01 6.99217319e+00 7.52442658e-01
-8.25131416e-01 5.75045906e-02 8.24871182e-01 7.88959637e-02
-2.18692094e-01 3.34192127e-01 -7.92358279e-01 1.70626551e-01
6.80321276e-01 1.57718316e-01 3.70569706e-01 7.20722198e-01
-1.95385009e-01 -1.24627382e-01 -1.23711848e+00 8.96023095e-01
1.02467649e-01 -1.10557020e+00 -1.14941202e-01 2.57131815e-01
9.68628526e-01 -7.01516867e-02 -1.21133126e-01 -2.87079047e-02
7.21801996e-01 -1.18409395e+00 5.18970668e-01 2.16883376e-01
8.29268932e-01 -6.70141518e-01 3.61016989e-01 4.84748781e-01
-1.18535256e+00 -5.98278977e-02 -1.73779935e-01 2.13039115e-01
8.63424735e-04 8.58580410e-01 -7.06906259e-01 5.71245253e-01
2.03594700e-01 5.46521544e-01 -7.96377599e-01 1.27397466e+00
-6.24908328e-01 9.01375532e-01 -4.21232998e-01 9.65665504e-02
1.66705713e-01 -1.03040211e-01 1.67628482e-01 1.24234033e+00
-3.19442689e-01 -1.46781970e-02 6.88625991e-01 4.75784600e-01
-3.63532543e-01 3.27047735e-01 -4.26060557e-01 3.42618823e-02
2.72375852e-01 1.49538827e+00 -1.19773400e+00 -4.96051423e-02
-7.06149757e-01 7.31536806e-01 9.19963360e-01 2.04430670e-01
-9.18550968e-01 -1.05102159e-01 2.75894284e-01 -6.91260248e-02
2.38492310e-01 -6.11132048e-02 -2.85861760e-01 -1.26188290e+00
-1.43896779e-02 -7.93627501e-01 8.40694845e-01 -3.20061475e-01
-1.33175457e+00 2.86718905e-01 9.06558614e-03 -9.07354593e-01
-3.92940901e-02 -7.07612813e-01 -2.05696955e-01 5.61637878e-01
-1.68881249e+00 -1.36340690e+00 -1.94979325e-01 3.51889461e-01
2.66888976e-01 3.05132806e-01 9.02989566e-01 1.66590080e-01
-7.15348065e-01 6.66712999e-01 2.61843145e-01 1.03479542e-01
6.77013397e-01 -1.52642703e+00 -1.33612640e-02 6.41904891e-01
6.02132738e-01 2.29781181e-01 3.93442661e-01 -4.16726053e-01
-1.24850285e+00 -1.08624876e+00 7.23859966e-01 -3.63022983e-01
6.76734507e-01 -5.52373946e-01 -9.36394453e-01 7.60569990e-01
-1.11629099e-01 3.60689014e-01 1.05910099e+00 5.72070420e-01
-5.86267114e-01 1.57372043e-01 -9.91384387e-01 3.29528987e-01
9.81728137e-01 -4.51635748e-01 -3.01457699e-02 8.02763164e-01
2.66569912e-01 -5.41474223e-01 -8.85197997e-01 2.90996104e-01
5.82491875e-01 -7.20386028e-01 7.62431860e-01 -8.49808335e-01
2.71023422e-01 -3.15863311e-01 -8.15562624e-03 -1.28647304e+00
-4.10008818e-01 -7.23762512e-01 -5.08428887e-02 1.05796015e+00
9.45250869e-01 -5.89073777e-01 1.01597333e+00 7.51010895e-01
3.45656842e-01 -1.00363016e+00 -7.18178809e-01 -5.79423904e-01
-6.08278848e-02 -4.28260148e-01 1.40446261e-01 1.04688942e+00
1.14973545e-01 4.85615402e-01 -3.94932389e-01 4.25439149e-01
1.01548886e+00 5.34362018e-01 6.91643953e-01 -1.47934353e+00
-5.39092183e-01 -2.73231119e-01 -4.75389361e-01 -9.58129883e-01
5.33055544e-01 -1.19738793e+00 5.43070920e-02 -1.57519007e+00
8.20368588e-01 -9.44738090e-01 -5.52825928e-01 8.58827174e-01
-4.31327403e-01 2.43514970e-01 1.53471017e-02 2.25684524e-01
-8.75152409e-01 1.36169806e-01 9.65669811e-01 -1.04299851e-01
-8.16930309e-02 2.55824327e-01 -8.72202456e-01 8.87891114e-01
7.02592075e-01 -9.67600286e-01 -4.29369777e-01 -1.99484929e-01
4.46781904e-01 2.74998128e-01 2.97499329e-01 -5.35720050e-01
3.81977826e-01 -5.53299010e-01 1.85287848e-01 -3.80310148e-01
2.17214361e-01 -7.76324093e-01 1.80556431e-01 2.74765491e-01
-5.70095062e-01 -3.82084876e-01 -1.54051008e-02 9.87700820e-01
-7.86452070e-02 -3.39419216e-01 1.06282341e+00 2.37940904e-02
-2.53199399e-01 3.66732806e-01 6.17077621e-03 1.94705412e-01
1.17677712e+00 1.46091178e-01 -2.96294361e-01 -1.56681508e-01
-8.17467570e-01 4.97676015e-01 4.49041247e-01 1.57369465e-01
3.09076190e-01 -1.25823367e+00 -6.31365180e-01 -1.92616263e-03
3.38305473e-01 -2.96779238e-02 -2.58679926e-01 8.42605293e-01
-2.73207426e-01 3.05622846e-01 2.06111982e-01 -6.89712346e-01
-1.70494258e+00 5.92321157e-01 2.97049761e-01 -6.95040107e-01
-2.83198446e-01 7.74996161e-01 3.11721206e-01 -4.67446566e-01
1.75810084e-01 -1.78219795e-01 -1.83179364e-01 6.91124424e-02
1.57916248e-01 2.96064373e-02 -5.80275506e-02 -6.46764457e-01
-4.81849432e-01 5.76157391e-01 -2.52187729e-01 8.02287459e-02
1.29522645e+00 -7.29546845e-02 -1.70809686e-01 4.19373184e-01
1.29812813e+00 5.24041653e-02 -1.18056118e+00 -5.23840964e-01
3.38379562e-01 -4.69487399e-01 9.69836395e-03 -7.70703971e-01
-1.04858792e+00 5.79787254e-01 1.22215092e-01 3.04590821e-01
1.07586229e+00 4.41068292e-01 2.83764988e-01 5.52353919e-01
2.86780447e-01 -8.74641299e-01 1.21943578e-01 1.67248294e-01
4.48763728e-01 -1.49571145e+00 2.82523811e-01 -1.06332779e+00
-4.75899160e-01 1.12125218e+00 3.23044330e-01 1.96071729e-01
8.31459165e-01 4.42330033e-01 -2.55842898e-02 -3.24278563e-01
-9.63831186e-01 -2.12706998e-01 4.68485296e-01 3.86669099e-01
3.87720883e-01 2.00860769e-01 -1.69843167e-01 4.13230270e-01
2.65926361e-01 -1.03290975e-01 3.07649881e-01 9.20973778e-01
-3.61851156e-01 -1.37981081e+00 -1.74892217e-01 7.60191441e-01
-5.87076247e-01 9.44729801e-03 -5.59033155e-01 5.71066797e-01
-1.37999341e-01 1.10904133e+00 -4.93445218e-01 -1.61529571e-01
1.01493470e-01 5.22574745e-02 6.48866117e-01 -7.73073435e-01
-2.20376477e-01 1.14131562e-01 2.63447016e-01 -4.47223634e-01
-5.18272460e-01 -7.97251999e-01 -1.13497698e+00 1.14587963e-01
-8.71119738e-01 2.58761104e-02 6.97931826e-01 1.03770971e+00
1.12837687e-01 2.70613819e-01 8.61449599e-01 -3.50760221e-01
-4.84375805e-01 -5.15404582e-01 -5.79241991e-01 3.68971795e-01
-3.44315320e-02 -7.13980973e-01 -2.85098404e-01 1.19610369e-01] | [9.18376636505127, 4.191214084625244] |
7fa9a5b4-03fd-4e21-9d96-84153f9fd950 | a-multi-stage-framework-with-context | 1810.07075 | null | http://arxiv.org/abs/1810.07075v1 | http://arxiv.org/pdf/1810.07075v1.pdf | A Multi-stage Framework with Context Information Fusion Structure for Skin Lesion Segmentation | The computer-aided diagnosis (CAD) systems can highly improve the reliability
and efficiency of melanoma recognition. As a crucial step of CAD, skin lesion
segmentation has the unsatisfactory accuracy in existing methods due to large
variability in lesion appearance and artifacts. In this work, we propose a
framework employing multi-stage UNets (MS-UNet) in the auto-context scheme to
segment skin lesion accurately end-to-end. We apply two approaches to boost the
performance of MS-UNet. First, UNet is coupled with a context information
fusion structure (CIFS) to integrate the low-level and context information in
the multi-scale feature space. Second, to alleviate the gradient vanishing
problem, we use deep supervision mechanism through supervising MS-UNet by
minimizing a weighted Jaccard distance loss function. Four out of five commonly
used performance metrics, including Jaccard index and Dice coefficient, show
that our approach outperforms the state-ofthe-art deep learning based methods
on the ISBI 2016 Skin Lesion Challenge dataset. | ["Chang'an Zhan", 'Shaofeng Yuan', 'Feng Yang', 'Yujiao Tang'] | 2018-10-16 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 6.49279952e-01 -2.78607011e-01 -1.16615243e-01 -4.35072094e-01
-1.12471688e+00 -2.29959965e-01 2.83259243e-01 2.46503115e-01
-5.59497535e-01 5.28521657e-01 9.03744996e-02 -1.31264150e-01
-3.62514138e-01 -5.06594896e-01 3.24645787e-02 -1.01593757e+00
3.41318876e-01 -7.87568092e-02 3.83356273e-01 -6.56084195e-02
2.58960515e-01 3.90929997e-01 -1.13528693e+00 5.57141602e-01
1.31172454e+00 1.16456866e+00 1.56037420e-01 7.55937934e-01
-9.30529162e-02 5.42082191e-01 -3.56567532e-01 -5.99661529e-01
1.17093474e-01 -6.23974741e-01 -8.19235206e-01 1.05960056e-01
5.48149168e-01 -1.05779126e-01 -1.58118233e-02 1.19600022e+00
7.94284880e-01 -3.94433230e-01 8.13121915e-01 -7.89601445e-01
-3.85481387e-01 2.12545469e-02 -1.03742993e+00 2.55135834e-01
4.50545736e-02 8.45390931e-03 6.35458648e-01 -7.68525958e-01
7.94794679e-01 9.04303133e-01 1.13080144e+00 5.24749696e-01
-8.47615242e-01 -2.39048168e-01 -1.04830518e-01 3.60742211e-01
-1.29369342e+00 -1.25191078e-01 4.27288800e-01 -3.21626604e-01
5.27437150e-01 6.80400431e-01 7.55104303e-01 8.29493940e-01
4.90025312e-01 8.75616968e-01 1.36355197e+00 -5.42506635e-01
1.12349421e-01 -1.31226122e-01 5.47027253e-02 9.79997873e-01
1.21318087e-01 -1.10265382e-01 -2.17448935e-01 -7.08565786e-02
7.70410001e-01 9.55838487e-02 2.53684027e-03 -1.23367421e-01
-1.09522116e+00 7.64958620e-01 8.98257911e-01 4.18761224e-01
-3.51635039e-01 -7.35485479e-02 5.94106078e-01 7.99311604e-03
4.97836918e-01 7.21236989e-02 -5.17182723e-02 1.40400171e-01
-9.09703493e-01 -8.99849683e-02 3.04971188e-01 2.27803960e-01
8.54018554e-02 -5.40973902e-01 -6.81577384e-01 9.57135856e-01
1.69580609e-01 6.75156191e-02 7.68622637e-01 -6.92205667e-01
1.07854635e-01 9.39499021e-01 -3.92716408e-01 -7.38998413e-01
-4.12893593e-01 -6.38385057e-01 -1.21119618e+00 3.20050269e-01
3.58806282e-01 -1.57267347e-01 -1.36331570e+00 1.34341645e+00
6.48134232e-01 2.88861245e-01 -8.43594819e-02 1.06125844e+00
6.40659094e-01 -1.33776486e-01 3.37463588e-01 -1.29687935e-01
1.42967486e+00 -1.14908683e+00 -8.30668092e-01 1.06821964e-02
9.79510009e-01 -8.31048131e-01 8.14861596e-01 2.96153039e-01
-8.10164511e-01 -2.92502135e-01 -9.98237491e-01 -7.54011199e-02
-4.58314747e-01 6.28118634e-01 7.31460989e-01 6.01235747e-01
-9.88735795e-01 5.04894972e-01 -1.06307542e+00 -6.08607054e-01
9.12437141e-01 2.32867077e-01 -7.42632270e-01 -3.61106068e-01
-9.04502213e-01 1.05128372e+00 1.20485291e-01 3.94324362e-01
-5.78299224e-01 -6.34941638e-01 -6.12088859e-01 -5.51453590e-01
2.56328940e-01 -7.95713663e-01 1.02348089e+00 -1.00030792e+00
-1.31794465e+00 1.04633427e+00 -3.17627162e-01 -4.67086494e-01
8.17089498e-01 -1.06412776e-01 -2.53412098e-01 3.36425722e-01
1.60579175e-01 6.07090890e-01 5.68640530e-01 -7.92493224e-01
-7.90780306e-01 -6.45463347e-01 -3.13459396e-01 4.86466855e-01
-3.40583861e-01 -8.09353888e-02 -4.69961643e-01 -6.89696670e-01
1.21195830e-01 -8.38492334e-01 -6.58797503e-01 5.47383547e-01
-5.49046993e-01 -1.99360058e-01 8.93688083e-01 -1.09343183e+00
1.41444838e+00 -2.00116324e+00 1.90582588e-01 1.73240080e-01
3.15091491e-01 5.62708735e-01 -6.40508384e-02 -7.60388076e-02
-8.76919031e-02 1.47191510e-01 -5.90552866e-01 -3.10367286e-01
-3.43883008e-01 7.92902038e-02 7.23900557e-01 4.84631568e-01
4.75825787e-01 8.77732158e-01 -7.87797809e-01 -8.10846090e-01
3.24166626e-01 6.04152977e-01 -2.37664118e-01 -6.66331947e-02
1.04774661e-01 3.40782166e-01 -2.97523826e-01 8.95419717e-01
7.41576433e-01 -3.90426278e-01 7.46140182e-02 -4.93872613e-01
1.53680220e-01 -3.11111659e-01 -8.70395720e-01 2.21526885e+00
-3.86561185e-01 5.35777032e-01 1.58748478e-01 -6.88775420e-01
3.57600152e-01 2.44002521e-01 6.10356867e-01 -5.36522985e-01
2.18744174e-01 3.10826659e-01 2.34286234e-01 -8.30146074e-01
6.24542081e-05 -2.61169653e-02 3.22064459e-01 5.89666003e-03
-1.11850336e-01 1.05947658e-01 1.05950318e-01 1.12438016e-01
1.24983954e+00 -6.06742576e-02 6.19556725e-01 -9.79433805e-02
7.75443912e-01 -2.35789232e-02 6.74730062e-01 2.25631461e-01
-6.20115876e-01 8.06876302e-01 5.16578138e-01 -1.96503788e-01
-7.58126557e-01 -8.29112947e-01 -3.63547474e-01 6.86244071e-01
5.55638671e-02 -1.57263175e-01 -1.03491080e+00 -1.26812887e+00
3.52773368e-02 1.76385567e-01 -1.16056120e+00 -1.07845388e-01
-2.62667716e-01 -1.11632228e+00 5.54715574e-01 6.20166123e-01
7.77076125e-01 -4.52587277e-01 -4.60257798e-01 3.16909924e-02
-3.94071005e-02 -7.58503854e-01 -6.44589961e-01 -2.50643492e-03
-7.58068621e-01 -1.35589778e+00 -1.10890186e+00 -7.99937189e-01
8.83041441e-01 2.03955576e-01 6.08193636e-01 2.05595344e-01
-1.06073046e+00 -1.40163451e-01 -2.83309430e-01 -2.00255081e-01
-3.16604137e-01 6.11541681e-02 -4.92363006e-01 1.54409513e-01
5.80697477e-01 -1.74714312e-01 -9.62148011e-01 2.14561686e-01
-1.00942791e+00 2.41712600e-01 1.15848422e+00 1.11122835e+00
9.15161848e-01 -2.18305308e-02 5.50513566e-01 -8.69603157e-01
5.65604270e-01 -2.19792038e-01 -6.10475279e-02 6.44732118e-01
-4.87927377e-01 -3.29995126e-01 1.80842113e-02 -2.57238567e-01
-1.05390573e+00 2.20889345e-01 -3.43950123e-01 -1.73358262e-01
-1.39096528e-01 5.45865297e-01 7.80265406e-02 -4.98777807e-01
7.04782724e-01 3.40709612e-02 2.55449027e-01 -2.44908273e-01
6.91531748e-02 8.23511422e-01 4.54746544e-01 1.71017423e-01
2.89751709e-01 5.23410559e-01 4.27559048e-01 -4.98838633e-01
-1.11556423e+00 -5.72774172e-01 -8.17750335e-01 -6.76519424e-02
1.03813386e+00 -8.03115249e-01 -3.18763673e-01 8.92872453e-01
-8.67560029e-01 -4.38678600e-02 -8.11331123e-02 3.36445481e-01
-8.67901370e-02 5.12690306e-01 -6.28574371e-01 -6.47997022e-01
-7.52595246e-01 -1.27965653e+00 1.36398888e+00 6.24297082e-01
-5.47301844e-02 -1.17129552e+00 3.98630602e-03 5.01090407e-01
4.73671705e-01 7.85589278e-01 7.04918206e-01 -4.38458472e-01
9.87318456e-02 -5.53326488e-01 -6.52377248e-01 5.91353953e-01
5.41899323e-01 1.07240453e-01 -8.23845506e-01 -3.88812721e-01
-2.18542978e-01 -1.60139456e-01 1.07806766e+00 6.26897931e-01
1.32881725e+00 1.08999290e-01 -5.70187211e-01 7.14128256e-01
1.72690129e+00 3.93886827e-02 4.57481802e-01 4.68916446e-02
7.01764941e-01 5.62045455e-01 7.78802156e-01 4.83456105e-02
4.27255780e-01 3.78159970e-01 3.98463249e-01 -4.16859120e-01
-6.56551659e-01 -7.88470656e-02 -1.51218712e-01 5.59419572e-01
-4.94974628e-02 1.22336097e-01 -9.14021909e-01 6.86028719e-01
-1.82899833e+00 -6.02367043e-01 -2.14889824e-01 1.98860288e+00
8.91649127e-01 -3.35700661e-02 -1.21192321e-01 1.98229343e-01
8.66946459e-01 -1.88455924e-01 -8.10956359e-01 -1.58667058e-01
-2.18455553e-01 1.23300195e-01 4.63030964e-01 4.26084369e-01
-1.35052431e+00 7.47174203e-01 5.85536480e+00 1.28148854e+00
-1.14470220e+00 2.59378016e-01 7.40964174e-01 1.20988280e-01
8.96798894e-02 -3.10309887e-01 -5.30686498e-01 5.20128608e-01
4.37445641e-01 2.16850102e-01 -2.04412788e-01 4.67877537e-01
-1.78140979e-02 -4.84277517e-01 -8.69938314e-01 9.75736737e-01
2.32680768e-01 -1.35900676e+00 -9.70275402e-02 1.68356895e-01
8.26936126e-01 -1.53099000e-01 2.57680148e-01 -1.37765497e-01
-1.01760022e-01 -1.14170909e+00 -1.23759672e-01 7.74017870e-01
1.17491138e+00 -6.48590207e-01 1.21123242e+00 -7.63527229e-02
-1.02634060e+00 1.87055320e-02 -6.08270168e-02 5.72073519e-01
-1.17903858e-01 7.03166306e-01 -9.86072540e-01 9.21266139e-01
2.97485054e-01 9.36726809e-01 -1.05925131e+00 1.47288752e+00
-9.33917686e-02 3.79855365e-01 -3.06473285e-01 3.75072397e-02
3.47109497e-01 -1.41931504e-01 5.03980577e-01 1.32904446e+00
2.23566115e-01 -6.40283301e-02 -9.90192369e-02 3.71532083e-01
1.88694447e-01 1.19520105e-01 -2.54336178e-01 1.97372764e-01
1.58085346e-01 1.60371840e+00 -6.44034743e-01 -2.02147260e-01
-7.57517815e-02 1.31848407e+00 -4.85723242e-02 -1.90039854e-02
-5.79044938e-01 -5.49636483e-01 5.37796021e-01 -8.15378800e-02
5.44529594e-02 2.68696159e-01 -5.59820533e-01 -1.02541459e+00
1.41458049e-01 -8.20175171e-01 6.74580693e-01 -4.69352722e-01
-1.49910331e+00 5.48925400e-01 -6.21281385e-01 -1.21377885e+00
8.85163695e-02 -6.57606244e-01 -7.92633116e-01 8.54044080e-01
-1.80676091e+00 -1.51627779e+00 -6.56176686e-01 6.80350840e-01
5.82736850e-01 -1.31545648e-01 9.21860635e-01 2.76664674e-01
-9.09703612e-01 9.34854627e-01 2.10983958e-02 2.46307820e-01
9.23640788e-01 -1.45420563e+00 7.26228058e-02 8.53960574e-01
-3.78895402e-01 1.70134574e-01 1.77992344e-01 -6.03126884e-01
-1.04850531e+00 -1.23564672e+00 4.40331429e-01 -4.25489962e-01
5.14836848e-01 1.40422493e-01 -7.42089331e-01 1.32919461e-01
1.26333132e-01 3.57232600e-01 1.15639198e+00 -2.56681293e-01
-9.90300179e-02 -1.56986967e-01 -1.61574233e+00 5.47523737e-01
8.15535247e-01 -3.58521104e-01 -2.31406584e-01 4.01847780e-01
4.13639814e-01 -4.92122859e-01 -1.17405999e+00 6.88041091e-01
6.64740324e-01 -8.63754570e-01 7.49779165e-01 -6.02312326e-01
4.77982521e-01 -2.73800522e-01 -1.90218151e-01 -1.27964950e+00
-2.68975556e-01 -2.49444380e-01 6.83616102e-02 8.90221000e-01
2.10762352e-01 -3.87542516e-01 1.03146613e+00 2.87638545e-01
-8.59306753e-02 -1.32952058e+00 -1.21692646e+00 -1.95902094e-01
2.28476554e-01 1.30639486e-02 2.94793576e-01 7.71435976e-01
-2.03299925e-01 3.66689228e-02 -3.44640575e-02 1.34534746e-01
8.69379044e-01 -2.08129957e-01 1.87903821e-01 -1.16461086e+00
1.68652043e-01 -6.48408771e-01 -7.05189466e-01 -3.86995971e-01
-3.53435844e-01 -9.36848044e-01 -1.80982932e-01 -1.78627956e+00
3.62064719e-01 -3.53931367e-01 -6.92661703e-01 4.77371395e-01
-5.82283676e-01 4.78261143e-01 -1.01944454e-01 -7.23441737e-03
-6.61526918e-01 2.05199614e-01 1.59859991e+00 -1.77895948e-01
8.25241394e-03 -7.64644668e-02 -7.56370306e-01 7.93755889e-01
7.32613623e-01 -2.69146442e-01 -1.01492971e-01 -6.19488060e-01
-1.94796115e-01 -2.94364095e-02 4.65328842e-01 -1.09871054e+00
5.41110873e-01 -5.93417026e-02 8.68223488e-01 -6.17496133e-01
2.42282867e-01 -7.73215294e-01 -2.61450410e-01 6.80099189e-01
-3.08384955e-01 -2.37367779e-01 2.70320475e-02 7.31755614e-01
-4.75324124e-01 -3.85960154e-02 1.09751058e+00 1.07641434e-02
-5.12414038e-01 4.96748447e-01 -1.38213307e-01 -3.42124432e-01
1.36438930e+00 -2.96033978e-01 -3.27692837e-01 3.88676226e-01
-7.11275458e-01 3.85790020e-01 2.64349192e-01 2.08669290e-01
7.42239416e-01 -1.47333014e+00 -9.01993513e-01 5.69158159e-02
1.84213310e-01 1.64973676e-01 7.52489924e-01 1.51918101e+00
-6.24207437e-01 2.93218166e-01 -2.46254280e-01 -7.71962762e-01
-1.65789557e+00 9.22068208e-02 7.65325844e-01 -6.87077820e-01
-3.93241823e-01 1.14518082e+00 -1.09987386e-01 -3.00343901e-01
2.19233617e-01 -1.64639309e-01 -2.92679548e-01 1.46692052e-01
6.52948499e-01 3.54102403e-01 3.23915184e-01 -4.65513527e-01
-4.19787258e-01 8.69014859e-01 -6.86915159e-01 1.23937719e-01
1.07092750e+00 -1.88370440e-02 -2.66559720e-01 -1.27393991e-01
1.28670931e+00 -4.37418431e-01 -1.12089562e+00 -2.23003760e-01
5.98823559e-03 -5.21714628e-01 5.08693814e-01 -1.37064600e+00
-1.17278624e+00 9.84165192e-01 1.45752358e+00 -1.69305027e-01
1.44779027e+00 -5.20699382e-01 9.95849609e-01 8.47643092e-02
-2.06342712e-02 -1.08738625e+00 -5.37427096e-03 -6.73738047e-02
6.35379910e-01 -1.61106825e+00 9.02715102e-02 -4.86227691e-01
-8.89020622e-01 1.09138322e+00 7.25456536e-01 -1.20333947e-01
6.20354593e-01 1.99462414e-01 3.35087031e-01 -4.89667850e-03
-6.95217729e-01 -3.50298941e-01 5.25999904e-01 6.18546188e-01
4.07128543e-01 1.03746139e-01 -6.45515501e-01 4.79779989e-01
3.42256039e-01 1.92099512e-01 1.57651424e-01 9.50678408e-01
-3.07454079e-01 -1.17379892e+00 -8.51151273e-02 7.88552701e-01
-7.19153821e-01 2.74706185e-02 -5.77337921e-01 6.00077748e-01
3.77903253e-01 7.46807694e-01 -2.53234208e-01 -4.32481229e-01
1.23289734e-01 -1.47360593e-01 5.40812135e-01 -3.94669473e-01
-7.23048806e-01 1.96846351e-01 -1.54938251e-01 -5.35932899e-01
-4.65870649e-01 -6.11048162e-01 -8.83160651e-01 1.86072856e-01
-4.72498089e-01 -1.45061880e-01 8.70475352e-01 9.02154803e-01
3.01167786e-01 6.32425010e-01 6.08723223e-01 -3.35854471e-01
-4.95622754e-01 -1.07689297e+00 -4.18151110e-01 2.98522919e-01
4.64069426e-01 -4.77278382e-01 -1.40074626e-01 -1.29814178e-01] | [15.625676155090332, -2.9020097255706787] |
ba6b6569-31df-4931-9b81-93ae864ec753 | diffroll-diffusion-based-generative-music | 2210.05148 | null | https://arxiv.org/abs/2210.05148v2 | https://arxiv.org/pdf/2210.05148v2.pdf | DiffRoll: Diffusion-based Generative Music Transcription with Unsupervised Pretraining Capability | In this paper we propose a novel generative approach, DiffRoll, to tackle automatic music transcription (AMT). Instead of treating AMT as a discriminative task in which the model is trained to convert spectrograms into piano rolls, we think of it as a conditional generative task where we train our model to generate realistic looking piano rolls from pure Gaussian noise conditioned on spectrograms. This new AMT formulation enables DiffRoll to transcribe, generate and even inpaint music. Due to the classifier-free nature, DiffRoll is also able to be trained on unpaired datasets where only piano rolls are available. Our experiments show that DiffRoll outperforms its discriminative counterpart by 19 percentage points (ppt.) and our ablation studies also indicate that it outperforms similar existing methods by 4.8 ppt. Source code and demonstration are available https://sony.github.io/DiffRoll/. | ['Yuki Mitsufuji', 'Dorien Herremans', 'Shusuke Takahashi', 'Naoya Takahashi', 'Naoki Murata', 'Toshimitsu Uesaka', 'Ryosuke Sawata', 'Kin Wai Cheuk'] | 2022-10-11 | null | null | null | null | ['music-transcription'] | ['music'] | [ 3.39831501e-01 1.40278712e-01 2.99072564e-01 1.71439722e-01
-1.48728144e+00 -8.60848367e-01 5.63668370e-01 -5.28467655e-01
-2.14064475e-02 7.54156947e-01 3.38576823e-01 -2.91731711e-02
1.43863603e-01 -4.24821883e-01 -8.73349786e-01 -7.44085133e-01
1.06751248e-01 5.33726335e-01 -3.16851914e-01 6.67809248e-02
-6.65358081e-02 9.20319036e-02 -1.37304914e+00 3.24502289e-01
7.19057500e-01 6.88132584e-01 2.40080819e-01 1.08936906e+00
3.17641169e-01 5.82244039e-01 -8.88919115e-01 -3.62876147e-01
3.78867209e-01 -9.61888075e-01 -5.62016606e-01 -2.14026406e-01
4.82702613e-01 -2.36953184e-01 -1.05110332e-01 8.22277665e-01
7.67855883e-01 7.82320425e-02 8.02752733e-01 -1.05102551e+00
-4.92494375e-01 9.85631764e-01 -5.14276624e-01 -2.40585893e-01
6.63702130e-01 2.08146982e-02 1.06399190e+00 -9.40268993e-01
6.13303661e-01 9.97399449e-01 9.86814201e-01 5.39229512e-01
-1.62767124e+00 -7.34605014e-01 -4.29408103e-01 -1.57355413e-01
-1.46877658e+00 -6.68180168e-01 7.34041631e-01 -3.42674553e-01
5.08975089e-01 6.49705470e-01 6.93846047e-01 1.61193681e+00
1.99412671e-03 1.00443184e+00 1.11854458e+00 -5.45647621e-01
2.45788112e-01 -3.62338871e-01 -4.87246096e-01 1.75399154e-01
-1.78781494e-01 1.12801082e-01 -7.71153390e-01 -1.28631592e-01
8.08472812e-01 -5.69041133e-01 -3.62134337e-01 1.27755284e-01
-1.30617964e+00 5.39998829e-01 -3.80146652e-02 2.69402653e-01
-4.18448806e-01 5.76653957e-01 2.38141134e-01 2.66865820e-01
3.17452133e-01 6.55647337e-01 -8.86576548e-02 -8.05208087e-01
-1.38648641e+00 6.17107749e-01 9.53363240e-01 9.99071896e-01
3.66071135e-01 3.95871252e-01 -3.59861493e-01 8.18902671e-01
-6.52066022e-02 5.97887218e-01 5.22742152e-01 -1.08262336e+00
3.52767140e-01 -3.15317124e-01 6.22154735e-02 -7.15461910e-01
-8.61607492e-02 -4.49168354e-01 -8.51994216e-01 -1.57347713e-02
3.81617010e-01 -4.42690104e-01 -8.77375424e-01 1.65688717e+00
-1.54424727e-01 5.47252595e-01 -1.96385890e-01 8.87150466e-01
5.78887224e-01 8.45272362e-01 -4.39426094e-01 -2.15383992e-01
1.12423170e+00 -9.02514338e-01 -7.26859212e-01 -5.37446924e-02
1.80434898e-01 -1.35158598e+00 1.33664000e+00 9.67442274e-01
-1.16514528e+00 -6.41735137e-01 -9.38597918e-01 -1.02605030e-01
3.26155186e-01 5.39446354e-01 5.13182044e-01 6.78158522e-01
-1.03805912e+00 7.24681258e-01 -8.93984020e-01 6.40351698e-02
3.00915450e-01 -7.66037703e-02 -1.19948938e-01 2.70893991e-01
-9.23364043e-01 3.19918215e-01 2.32317582e-01 -4.37146565e-03
-9.69953597e-01 -6.82657301e-01 -6.77712440e-01 -1.48146357e-02
1.47278741e-01 -7.66747475e-01 1.71397305e+00 -1.01297081e+00
-2.04271817e+00 5.14792204e-01 -3.52311879e-01 -5.63694835e-01
7.23340690e-01 -5.72888970e-01 -2.83405989e-01 9.37082693e-02
1.39117137e-01 4.01153654e-01 1.22413981e+00 -1.22713184e+00
-1.81833923e-01 2.63731658e-01 -3.16756099e-01 8.99536088e-02
1.58478662e-01 -1.21760935e-01 -5.58845043e-01 -1.25757957e+00
6.60538599e-02 -1.34434688e+00 8.59029766e-04 -5.30958951e-01
-8.51959705e-01 2.13042974e-01 6.27926171e-01 -8.66072416e-01
1.32061410e+00 -2.16723180e+00 3.16158146e-01 1.09287567e-01
-1.08634755e-01 1.45833448e-01 -1.32008493e-01 7.82032907e-01
-1.57617807e-01 1.50360931e-02 -4.75791931e-01 -8.79151881e-01
3.46922368e-01 1.79061145e-01 -8.25596571e-01 6.68430403e-02
1.33502975e-01 8.83547246e-01 -8.09071541e-01 -1.29879609e-01
7.22171972e-03 7.27677047e-01 -7.42926478e-01 1.09331638e-01
-2.72076219e-01 9.05857682e-01 -8.79143849e-02 5.23822069e-01
7.15674222e-01 1.69985875e-01 2.39246696e-01 -8.67723376e-02
-1.01067744e-01 4.96327788e-01 -1.20491385e+00 2.16650152e+00
-5.87314904e-01 6.86063051e-01 -1.63317502e-01 -5.82293689e-01
9.32209253e-01 5.39557219e-01 2.32472077e-01 -3.55837524e-01
4.78412732e-02 3.18504035e-01 -2.06492811e-01 -2.11540177e-01
6.22094393e-01 -3.12305093e-01 -3.43574196e-01 6.21966600e-01
2.09571064e-01 -6.51312530e-01 1.93277285e-01 9.42242593e-02
1.17883945e+00 5.72058737e-01 1.06164880e-01 1.29246429e-01
-2.61359159e-02 -3.59576374e-01 5.39317548e-01 8.10562074e-01
3.65001261e-01 1.35842896e+00 4.87493366e-01 1.33668156e-02
-9.99566615e-01 -1.33982229e+00 3.90560851e-02 7.17666924e-01
-3.43831092e-01 -9.92842019e-01 -9.18231368e-01 -3.84148866e-01
-2.37280101e-01 9.69042361e-01 -4.34999704e-01 5.23028970e-02
-6.53192341e-01 -6.68124080e-01 8.61887693e-01 3.40678841e-01
2.47074574e-01 -1.00936353e+00 -4.14123237e-01 2.77420700e-01
-4.90034103e-01 -9.14568484e-01 -7.77534187e-01 1.97459906e-01
-6.20946825e-01 -6.54401839e-01 -9.07522500e-01 -4.92452174e-01
8.47165436e-02 -1.48063198e-01 1.27201247e+00 -5.03585160e-01
-2.69648582e-01 2.59038836e-01 -5.39767504e-01 -5.66940606e-01
-6.25005364e-01 1.08255297e-01 5.33871762e-02 9.38720182e-02
-1.45390019e-01 -1.33480048e+00 -5.85309267e-01 -5.49935410e-03
-9.06657994e-01 4.00168628e-01 6.90653384e-01 7.73441911e-01
8.52517903e-01 -1.36456206e-01 5.62561333e-01 -8.88565958e-01
7.52350986e-01 -2.49148190e-01 -3.20588559e-01 -2.38872036e-01
-3.17611933e-01 -8.08469020e-03 6.79003298e-01 -5.59678912e-01
-9.21035826e-01 1.56877562e-01 -4.72435683e-01 -5.99528670e-01
-1.96016997e-01 3.59197527e-01 -5.40200546e-02 3.77418429e-01
7.45547414e-01 4.44339275e-01 -2.92786807e-01 -8.09340715e-01
5.10716319e-01 6.07115269e-01 7.92008936e-01 -6.68180287e-01
9.20143068e-01 2.72656292e-01 -1.22469515e-01 -6.96421862e-01
-7.26873338e-01 -4.81482483e-02 -2.61518538e-01 -1.19346805e-01
7.64706135e-01 -9.58784401e-01 -6.59913182e-01 3.14936608e-01
-9.48556185e-01 -5.57110488e-01 -6.28893733e-01 4.93704200e-01
-1.14162099e+00 1.87508821e-01 -7.65554965e-01 -8.71565163e-01
-4.68725979e-01 -8.79712284e-01 1.40871787e+00 -6.53146440e-03
-7.70501137e-01 -5.20418108e-01 4.51825976e-01 1.66650370e-01
1.54522240e-01 4.19686526e-01 4.10790980e-01 -1.60661995e-01
-4.41443741e-01 -1.68110490e-01 2.92653918e-01 5.00902832e-01
1.55086562e-01 7.92640150e-02 -1.33428609e+00 -2.13244617e-01
1.74971949e-02 -2.97494292e-01 8.27010155e-01 3.18089068e-01
1.11973464e+00 -5.00657499e-01 6.27885312e-02 8.93251300e-01
1.14919960e+00 6.54740259e-02 8.75055492e-01 7.14629740e-02
5.29832780e-01 -8.47186819e-02 4.39431608e-01 6.64434016e-01
-1.24427991e-03 9.84023988e-01 1.60520792e-01 2.64323533e-01
-5.73052406e-01 -7.77019739e-01 6.22308135e-01 1.15266562e+00
-3.65944952e-01 -2.50022292e-01 -7.40596354e-01 3.83798629e-01
-1.72605956e+00 -1.12135959e+00 -2.33918250e-01 2.18158078e+00
1.07577741e+00 -6.82538971e-02 1.97576746e-01 4.89488572e-01
3.46603513e-01 7.23516494e-02 -1.30120531e-01 -4.77356970e-01
-1.37044817e-01 8.15425634e-01 3.67055386e-02 4.30506617e-01
-9.98421013e-01 7.77489185e-01 6.27034950e+00 1.12462020e+00
-1.20815897e+00 2.40824848e-01 3.59244555e-01 -2.49757200e-01
-3.08541626e-01 1.68532897e-02 -3.58218223e-01 6.90048218e-01
9.58229482e-01 -6.31968528e-02 6.62554741e-01 5.79733253e-01
3.71795893e-01 -2.49668360e-01 -1.06733596e+00 1.23476160e+00
3.77386846e-02 -1.13958335e+00 5.65325134e-02 -1.17674328e-01
8.40465784e-01 -2.80092925e-01 3.23124081e-01 2.95336783e-01
1.31986648e-01 -1.18523204e+00 1.12068236e+00 7.30229318e-01
9.93462920e-01 -7.54962444e-01 3.04300100e-01 3.53641510e-01
-1.08053672e+00 3.53418380e-01 2.33378224e-02 -1.75078616e-01
3.40939015e-01 7.96259284e-01 -1.15640628e+00 9.64540124e-01
5.08016229e-01 6.46385431e-01 -2.04421937e-01 1.07799697e+00
-5.87370753e-01 1.20371914e+00 -4.24559683e-01 4.74342257e-01
-9.16054472e-02 -2.67646700e-01 8.18514049e-01 1.39315271e+00
8.67473125e-01 -2.07237199e-01 -9.96753424e-02 9.79340732e-01
-8.63679498e-02 -1.36026768e-02 -4.07142550e-01 -1.68883786e-01
3.87862921e-01 1.16030240e+00 -5.44395626e-01 -1.45077080e-01
1.76698446e-01 1.47445869e+00 -9.17850658e-02 4.33705479e-01
-9.98432219e-01 -4.21049029e-01 4.54571575e-01 -2.25079060e-02
5.08495986e-01 -3.90671581e-01 -2.90304989e-01 -1.22152317e+00
1.10795662e-01 -1.08425260e+00 -3.05960160e-02 -1.08699632e+00
-1.08337569e+00 8.27970803e-01 -2.30139509e-01 -1.67577899e+00
-7.32330978e-01 -1.80399403e-01 -7.71092176e-01 1.03222203e+00
-8.32664251e-01 -1.16520500e+00 -6.88818991e-02 3.78474295e-01
4.97892410e-01 9.88499224e-02 1.24081838e+00 2.97150642e-01
-3.06164563e-01 6.12237513e-01 2.48730615e-01 1.61327329e-02
8.65905821e-01 -1.46709037e+00 5.77809095e-01 8.21889699e-01
8.38722169e-01 4.83003587e-01 1.11971545e+00 -3.67855191e-01
-1.29238796e+00 -1.01013815e+00 7.94429064e-01 -4.71275985e-01
4.96468723e-01 -5.44405401e-01 -6.12215400e-01 7.43372560e-01
2.41111025e-01 -2.41683558e-01 7.86685526e-01 2.87361164e-02
-4.60763782e-01 -5.60761802e-02 -6.60523713e-01 5.93019605e-01
1.01231134e+00 -6.05228126e-01 -5.04380345e-01 1.32885084e-01
5.02493262e-01 -6.66602731e-01 -7.45401919e-01 1.29924893e-01
6.03132606e-01 -1.12385106e+00 7.66661465e-01 -3.36205550e-02
4.68243390e-01 -4.63488638e-01 -7.01056644e-02 -1.54219103e+00
-1.08376136e-02 -1.43754721e+00 -2.42798418e-01 1.38778007e+00
3.54038239e-01 -3.95482093e-01 5.91942310e-01 -8.31341371e-02
-3.73839766e-01 -4.02902246e-01 -1.01206827e+00 -9.52916801e-01
-7.20409025e-03 -9.47207093e-01 3.99782509e-01 7.18440354e-01
-1.46563143e-01 3.49900901e-01 -7.61404812e-01 -2.20505614e-03
4.33534980e-01 5.27451396e-01 1.01558173e+00 -8.36854637e-01
-1.05174720e+00 -3.12931180e-01 -7.83436224e-02 -1.14651847e+00
-5.86527959e-02 -1.00886345e+00 2.72845358e-01 -1.31135786e+00
-1.06122404e-01 -4.51856494e-01 1.27337232e-01 4.90989208e-01
4.69332980e-03 7.87002563e-01 5.81291020e-01 1.83053315e-01
-1.71934202e-01 7.22984791e-01 1.22204626e+00 5.38198203e-02
-3.02920312e-01 4.13117886e-01 -6.50069535e-01 5.68399012e-01
9.64288831e-01 -5.89529037e-01 -4.47665006e-01 -2.49052733e-01
2.31347546e-01 1.90568522e-01 5.46622336e-01 -1.36100531e+00
-2.18215376e-01 2.73333997e-01 3.06131274e-01 -4.26397115e-01
6.94374263e-01 -2.42263556e-01 7.65941679e-01 2.29020134e-01
-1.59450591e-01 -7.21426308e-02 3.67775559e-01 3.14908355e-01
-3.96686226e-01 -1.00045800e-01 2.88040847e-01 1.32453352e-01
-8.60154908e-03 6.92662075e-02 -5.32852471e-01 1.65667251e-01
5.42443454e-01 -3.12059987e-02 1.05222754e-01 -8.32570970e-01
-1.01505566e+00 -4.99962896e-01 3.23722929e-01 2.19161719e-01
4.78727549e-01 -1.46003282e+00 -7.92823493e-01 1.18915729e-01
-2.15668246e-01 2.15351321e-02 2.81498998e-01 8.63242269e-01
-4.37465012e-01 2.74117827e-01 8.55835527e-02 -4.68761683e-01
-1.07873058e+00 2.21133336e-01 3.67217809e-02 -1.51268288e-01
-8.00987184e-01 9.35358167e-01 1.57702759e-01 -1.69370577e-01
1.27375633e-01 -4.33874369e-01 3.71145219e-01 -1.51503868e-02
4.28878754e-01 9.65514705e-02 1.39230207e-01 -4.57738876e-01
-9.30030923e-03 5.03122926e-01 4.04894322e-01 -6.67137206e-01
1.25253201e+00 1.13735817e-01 7.87617639e-03 8.03789318e-01
9.15171504e-01 7.43611276e-01 -1.23368669e+00 2.92558789e-01
-2.04608768e-01 -3.88145357e-01 -2.31493846e-01 -9.34199214e-01
-7.50541866e-01 8.35019708e-01 3.41440439e-01 1.77756473e-01
1.38630104e+00 -1.31901249e-01 8.56237233e-01 -3.93518247e-02
4.40839291e-01 -6.78349376e-01 3.63499820e-02 5.54564118e-01
1.28094471e+00 -4.66726065e-01 -3.73964548e-01 -3.17538649e-01
-5.96822143e-01 1.10426366e+00 4.61397544e-02 -3.90045315e-01
3.61624628e-01 6.13892257e-01 1.54800624e-01 4.19939607e-01
-6.20203555e-01 -1.26506403e-01 3.35048378e-01 6.05261683e-01
6.12560451e-01 1.74747810e-01 -1.62940279e-01 9.13749993e-01
-1.08809447e+00 1.77309200e-01 6.82039618e-01 6.03390336e-01
1.25508830e-01 -1.46542978e+00 -4.73594129e-01 1.33023873e-01
-4.43085372e-01 -3.58100414e-01 -3.11445385e-01 5.97745061e-01
2.24426061e-01 7.40735888e-01 -4.24983092e-02 -5.74974239e-01
2.55447417e-01 2.74756521e-01 6.45987570e-01 -5.97378492e-01
-5.49903929e-01 5.27348936e-01 2.74427652e-01 -5.78887761e-01
-2.94509202e-01 -7.61852503e-01 -1.05626893e+00 -2.25652322e-01
9.78629366e-02 3.06695253e-01 4.44131821e-01 4.14988577e-01
4.43753392e-01 8.32919002e-01 4.45452899e-01 -1.20300782e+00
-4.14105266e-01 -1.14240432e+00 -6.15525901e-01 2.41644293e-01
3.41183305e-01 -2.24933326e-01 -2.59580821e-01 3.43787253e-01] | [15.709324836730957, 5.703652381896973] |
a99aef55-7692-4692-9db8-eacc11113eeb | improving-automatic-parallel-training-via | 2307.02031 | null | https://arxiv.org/abs/2307.02031v1 | https://arxiv.org/pdf/2307.02031v1.pdf | Improving Automatic Parallel Training via Balanced Memory Workload Optimization | Transformer models have emerged as the leading approach for achieving state-of-the-art performance across various application domains, serving as the foundation for advanced large-scale deep learning (DL) models. However, efficiently training these models across multiple GPUs remains a complex challenge due to the abundance of parallelism options. Existing DL systems either require manual efforts to design distributed training plans or limit parallelism combinations to a constrained search space. In this paper, we present Galvatron-BMW, a novel system framework that integrates multiple prevalent parallelism dimensions and automatically identifies the most efficient hybrid parallelism strategy. To effectively navigate this vast search space, we employ a decision tree approach for decomposition and pruning based on intuitive insights. We further utilize a dynamic programming search algorithm to derive the optimal plan. Moreover, to improve resource utilization and enhance system efficiency, we propose a bi-objective optimization workflow that focuses on workload balance. Our evaluations on different Transformer models demonstrate the capabilities of Galvatron-BMW in automating distributed training under varying GPU memory constraints. Across all tested scenarios, Galvatron-BMW consistently achieves superior system throughput, surpassing previous approaches that rely on limited parallelism strategies. | ['Bin Cui', 'Xiaonan Nie', 'Fangcheng Fu', 'Xupeng Miao', 'Youhe Jiang', 'Yujie Wang'] | 2023-07-05 | null | null | null | null | ['navigate'] | ['reasoning'] | [-1.66305244e-01 -5.51476181e-01 -3.75577301e-01 -3.31485599e-01
-6.64441347e-01 -4.07986850e-01 4.19918716e-01 2.19180644e-01
-4.90894258e-01 3.37714434e-01 -1.09567910e-01 -6.91273570e-01
-3.60589951e-01 -8.56598020e-01 -3.84455740e-01 -5.05478978e-01
5.08884639e-02 7.94370532e-01 3.52995157e-01 -1.84556156e-01
1.97567955e-01 6.81304038e-01 -1.66110432e+00 4.31722611e-01
8.75336528e-01 9.68544960e-01 4.29145187e-01 4.66296941e-01
-3.81531984e-01 6.59865916e-01 -5.18845677e-01 -3.46579731e-01
4.55833137e-01 -1.21023446e-01 -8.04413617e-01 -1.41028807e-01
4.66539234e-01 -3.51704627e-01 9.80529264e-02 6.59099519e-01
8.01142812e-01 2.25688890e-02 6.96504787e-02 -1.09280109e+00
1.79791301e-01 4.11350489e-01 -6.56548440e-01 4.02205527e-01
-1.26447111e-01 9.77806300e-02 1.10208988e+00 -7.40449965e-01
4.41098183e-01 1.02456748e+00 7.80400813e-01 2.41260752e-01
-1.41099095e+00 -6.63311720e-01 2.30686873e-01 4.78260547e-01
-1.39438105e+00 -3.65826428e-01 6.00295901e-01 -2.75111049e-01
1.45313740e+00 3.05826724e-01 7.89887249e-01 6.89041376e-01
2.06140533e-01 9.41828012e-01 9.42912042e-01 -5.73696375e-01
3.86756688e-01 -2.61675894e-01 2.01928735e-01 9.14481103e-01
1.79520607e-01 -1.91193923e-01 -9.71427381e-01 -3.37531060e-01
2.31431410e-01 -5.59419952e-02 1.73662845e-02 -3.90174508e-01
-8.71116638e-01 8.25670660e-01 2.33791113e-01 1.68688595e-01
-2.85235584e-01 5.71066365e-02 7.71174312e-01 1.42092526e-01
3.40833813e-01 3.80115420e-01 -6.34848535e-01 -3.75880092e-01
-1.06326818e+00 3.77908230e-01 9.11931515e-01 7.63136387e-01
8.67364407e-01 -1.99634686e-01 -4.50005472e-01 1.01533782e+00
1.22286536e-01 2.17325538e-01 5.61156690e-01 -7.29553759e-01
5.47367156e-01 9.81427968e-01 -3.70796204e-01 -8.53009641e-01
-5.83330989e-01 -7.94689059e-01 -8.69609833e-01 2.79931398e-03
6.41562045e-02 8.43944959e-03 -4.91928369e-01 1.47067010e+00
6.59684539e-01 -2.41011996e-02 -1.16990246e-01 8.05148423e-01
4.50078636e-01 4.24782217e-01 1.63093671e-01 1.07961833e-01
1.51472521e+00 -1.41869462e+00 -2.82431096e-01 -3.41606736e-01
1.10530972e+00 -7.08221674e-01 1.27675509e+00 5.86013913e-01
-9.49915946e-01 -3.71706635e-01 -1.10686576e+00 -2.62793809e-01
-1.18127130e-01 5.25464654e-01 9.39039826e-01 8.64554822e-01
-1.12687087e+00 5.51028490e-01 -1.31392634e+00 -3.25193554e-01
4.39647675e-01 5.37278593e-01 8.58616158e-02 -1.18260823e-01
-5.43160021e-01 7.17335999e-01 5.43220103e-01 -9.47621614e-02
-5.86002827e-01 -8.53648067e-01 -3.50411355e-01 2.02637225e-01
4.58884984e-01 -8.97883117e-01 1.36282611e+00 -3.21665674e-01
-1.55040681e+00 6.33262515e-01 -4.27702554e-02 -6.40624344e-01
4.18017060e-01 -2.59321243e-01 4.76285256e-02 -1.69550672e-01
-1.68820724e-01 4.49387819e-01 4.20077831e-01 -7.16100931e-01
-7.66663909e-01 -4.29232329e-01 -3.77251320e-02 4.33550984e-01
-9.87401903e-01 7.05315918e-02 -9.95230317e-01 -3.35583180e-01
-1.62094347e-02 -1.05397403e+00 -3.63738298e-01 -2.82695889e-01
-2.37583876e-01 -2.62917966e-01 8.72389793e-01 -1.17911339e-01
1.69004524e+00 -1.90815604e+00 9.84774977e-02 1.47858202e-01
4.88036841e-01 4.92891937e-01 -2.65028514e-02 4.60763723e-01
5.35354316e-01 -2.69905835e-01 1.17652364e-01 -7.13949144e-01
1.77501097e-01 3.67819428e-01 -1.60361692e-01 1.55762672e-01
-1.25697732e-01 7.05807447e-01 -6.08261347e-01 -5.93460977e-01
6.99812919e-02 2.00762540e-01 -9.69512403e-01 1.80726752e-01
-3.44529569e-01 2.40853243e-02 -4.43797350e-01 5.16877413e-01
6.44129276e-01 -6.35173321e-01 4.68296587e-01 -3.75816077e-01
-1.37200609e-01 2.27984622e-01 -1.05908263e+00 1.84580088e+00
-8.89051497e-01 4.23624009e-01 1.14039518e-01 -9.77407694e-01
9.35172200e-01 -1.97862372e-01 4.92622912e-01 -8.84156168e-01
2.19032124e-01 3.21160525e-01 2.05631033e-02 -2.75891274e-01
4.64302301e-01 2.82867163e-01 1.11471646e-01 6.47959888e-01
-1.36927232e-01 1.22346535e-01 5.08911729e-01 -4.75018285e-02
1.29905152e+00 4.36332822e-02 2.64773041e-01 -4.70674664e-01
5.23643315e-01 3.12681824e-01 4.94890451e-01 8.00734818e-01
4.76801917e-02 1.06740646e-01 3.64017367e-01 -9.38507140e-01
-8.32555950e-01 -5.66092134e-01 -1.38937935e-01 1.61964869e+00
-2.63657831e-02 -9.11621511e-01 -7.47447908e-01 -4.49230492e-01
-1.68884844e-01 4.60382938e-01 -1.21112809e-01 8.89912322e-02
-7.44054198e-01 -1.13811457e+00 5.83712816e-01 3.92476946e-01
7.27522016e-01 -7.43328393e-01 -1.40402365e+00 3.37405443e-01
-8.25670362e-02 -1.23614550e+00 -2.59441406e-01 4.06821579e-01
-9.99386013e-01 -8.92758667e-01 -1.57702789e-01 -5.06861567e-01
4.05699521e-01 3.56503308e-01 1.46532941e+00 3.53734583e-01
-6.66409612e-01 5.63415438e-02 -1.44388497e-01 -2.06553534e-01
-2.19201714e-01 6.85375750e-01 -2.06743896e-01 -3.20012689e-01
4.10895795e-01 -5.19182265e-01 -7.53466129e-01 2.98811048e-01
-6.49115741e-01 6.08220994e-01 6.31413519e-01 9.51207161e-01
7.12670028e-01 4.08204570e-02 5.08680753e-02 -9.31419969e-01
9.10087168e-01 -3.25093418e-01 -8.03046346e-01 4.65989053e-01
-9.19180751e-01 3.59442353e-01 8.48473012e-01 -4.00770038e-01
-8.79605949e-01 2.52649300e-02 -3.25758100e-01 -5.50823450e-01
7.27252811e-02 6.02486670e-01 -6.17470145e-02 -1.52449176e-01
7.03476965e-01 1.99257836e-01 -1.63149565e-01 -4.85212624e-01
1.23347782e-01 3.70568752e-01 1.08723216e-01 -1.28161991e+00
1.92170754e-01 2.54963070e-01 1.82440042e-01 -6.24789476e-01
-9.34496105e-01 -5.70501089e-01 -3.84301007e-01 2.38589272e-02
5.02902031e-01 -7.79272437e-01 -9.67588305e-01 3.13585699e-01
-1.06201339e+00 -5.81964195e-01 3.58014219e-02 1.50548235e-01
-2.82040924e-01 1.69220045e-01 -5.58782279e-01 -5.73443830e-01
-8.51774037e-01 -1.54619253e+00 1.37373304e+00 -6.68525547e-02
-1.57735035e-01 -9.32032585e-01 1.68520749e-01 5.17236233e-01
7.04615593e-01 6.24637213e-03 1.06056666e+00 -4.91391212e-01
-8.47423375e-01 -7.31903687e-02 -3.06685954e-01 -1.14223994e-01
-2.37490803e-01 -2.25723967e-01 -7.08891571e-01 -6.28387153e-01
-2.52885371e-01 -4.36875403e-01 5.78415871e-01 -5.48879150e-03
1.64292514e+00 -2.18014210e-01 -4.20012772e-01 1.09324062e+00
1.64530396e+00 -1.04211077e-01 7.40546510e-02 7.63831258e-01
6.29288554e-01 2.34715670e-01 4.60667253e-01 7.98753798e-01
5.04125118e-01 1.03659165e+00 3.39939654e-01 -7.40322024e-02
1.32542867e-02 -8.58309194e-02 1.02095984e-01 8.81650090e-01
1.40119702e-01 -1.79396451e-01 -1.32549095e+00 3.06622803e-01
-1.92721951e+00 -3.96243393e-01 6.45684674e-02 2.13108253e+00
7.60619879e-01 1.75749570e-01 1.58487484e-01 1.21041737e-01
1.08970098e-01 1.34977669e-01 -6.20805681e-01 -5.47377110e-01
3.35601747e-01 4.67012495e-01 6.27443075e-01 8.34039599e-02
-8.25457633e-01 1.09063876e+00 5.97021580e+00 1.18534803e+00
-1.33465087e+00 3.45628500e-01 6.78943098e-01 -4.02088881e-01
-1.34281805e-02 -1.86111033e-01 -1.35346532e+00 2.50426233e-01
9.98928249e-01 -2.46706232e-01 5.11248589e-01 1.11339891e+00
2.33401909e-01 -1.01549022e-01 -1.15382588e+00 1.10748208e+00
-1.76238492e-01 -1.53497362e+00 -2.02036113e-03 1.52543411e-01
6.33761764e-01 4.36316788e-01 -5.80814555e-02 4.10257757e-01
4.48850989e-01 -8.01728189e-01 5.56924999e-01 -1.68377414e-01
7.12804914e-01 -6.84405029e-01 5.42414725e-01 6.95754349e-01
-1.33010185e+00 -2.15432331e-01 -2.89177001e-01 -7.23957121e-02
-6.47645146e-02 5.82598746e-01 -8.12657893e-01 4.43148196e-01
9.19163346e-01 2.57837594e-01 -6.34914279e-01 8.58671010e-01
1.14786468e-01 6.10010326e-01 -6.94452047e-01 -1.46662638e-01
3.45221698e-01 -7.10515678e-02 1.03546724e-01 1.41790497e+00
3.80992442e-01 -8.38979036e-02 4.68125552e-01 5.48376977e-01
-1.94126111e-03 4.71681923e-01 -1.94543272e-01 2.21409097e-01
6.99564636e-01 1.33615553e+00 -9.19900000e-01 -1.70422569e-01
-4.56370533e-01 6.83635175e-01 7.93294668e-01 -8.65792260e-02
-9.23435271e-01 1.04988113e-01 6.77964807e-01 -3.12236045e-02
1.76706418e-01 -5.45215309e-01 -5.97213864e-01 -1.04950130e+00
7.29830712e-02 -1.09238422e+00 5.20476103e-01 -1.89105466e-01
-7.40500093e-01 8.43927562e-01 6.69179931e-02 -1.06885993e+00
-4.86356653e-02 -5.54111898e-01 -5.56136072e-01 7.51404583e-01
-1.54597104e+00 -1.17789698e+00 -4.51596022e-01 4.99608338e-01
6.69127643e-01 -3.12406689e-01 9.18263197e-01 4.32007223e-01
-9.65382516e-01 8.53645921e-01 1.63793072e-01 -5.34516692e-01
3.84671271e-01 -9.47258770e-01 4.21131045e-01 7.73754120e-01
1.42947182e-01 6.22545719e-01 5.52343965e-01 -2.96534061e-01
-1.88909864e+00 -8.83425295e-01 6.12884641e-01 4.16653305e-02
7.23498225e-01 -4.42166537e-01 -7.48145938e-01 2.40596026e-01
3.02749388e-02 1.19221315e-01 8.68348420e-01 5.31324863e-01
-1.91205427e-01 -4.39786553e-01 -7.53517568e-01 7.41357088e-01
1.11471140e+00 -1.01138324e-01 -1.91321038e-02 4.41107154e-01
5.18093288e-01 -8.22505653e-01 -8.84471655e-01 4.42875743e-01
5.24668038e-01 -1.16079247e+00 8.95437360e-01 -3.16838205e-01
3.94085675e-01 -6.71812966e-02 -6.83422312e-02 -9.92439806e-01
-1.96741059e-01 -5.33102036e-01 -4.17113930e-01 7.31188178e-01
1.17190816e-01 -4.93514508e-01 1.24887657e+00 5.71389616e-01
-1.32991731e-01 -1.39683139e+00 -7.84960330e-01 -7.08401322e-01
-2.48882905e-01 -6.42941415e-01 8.42223465e-01 6.45010710e-01
-3.22400153e-01 3.74275357e-01 -1.66211441e-01 9.14415568e-02
6.18988514e-01 4.28701848e-01 1.09542441e+00 -9.04372275e-01
-7.96268165e-01 -7.47125208e-01 -1.59625739e-01 -1.25320041e+00
3.35461274e-02 -9.37522471e-01 -2.45584711e-01 -1.37821352e+00
1.92441583e-01 -7.79607356e-01 -2.03402653e-01 8.68830979e-01
5.61775826e-02 2.58702368e-01 3.52243513e-01 4.57822353e-01
-7.98376441e-01 5.40319026e-01 1.01879299e+00 4.64878567e-02
-2.41517603e-01 -1.13945104e-01 -5.56356907e-01 5.47020912e-01
8.11308861e-01 -2.78353482e-01 -6.18689060e-01 -9.96045172e-01
3.69256347e-01 3.76140624e-02 -5.12118973e-02 -1.22891331e+00
5.31396627e-01 -2.50377804e-01 -8.83076265e-02 -6.31253541e-01
1.76328003e-01 -8.32207918e-01 1.65884048e-01 5.92206419e-01
-1.48054868e-01 4.39902157e-01 5.46007335e-01 2.69362777e-01
-1.24576323e-01 -1.31973341e-01 7.33733892e-01 1.89388990e-01
-8.27845991e-01 4.48312074e-01 -1.19735092e-01 -4.29947628e-03
1.00532699e+00 -8.08610842e-02 -4.23389852e-01 3.92674237e-01
-2.89890885e-01 5.18152297e-01 3.54710728e-01 3.00553858e-01
3.21213514e-01 -1.04725802e+00 -4.66498852e-01 3.25666070e-01
8.22035000e-02 1.31227508e-01 2.97712445e-01 8.56623292e-01
-9.63427663e-01 5.32004833e-01 -1.79495245e-01 -8.47768664e-01
-1.57198763e+00 2.81942666e-01 3.19361746e-01 -9.66946721e-01
-8.28016758e-01 8.41141284e-01 -1.86181460e-02 -3.47299427e-01
1.82751507e-01 -1.18845470e-01 8.48568380e-02 -2.36440212e-01
4.07896698e-01 5.42920470e-01 7.32903838e-01 -1.71567351e-01
-3.89967114e-01 4.00600642e-01 -2.70296216e-01 2.71482050e-01
1.33315492e+00 5.23796342e-02 -2.00337753e-01 -1.52684063e-01
1.07306850e+00 -2.05672920e-01 -1.11322844e+00 -3.19719583e-01
1.82475731e-01 -5.12401044e-01 4.23719257e-01 -6.54529274e-01
-1.14640546e+00 8.24549675e-01 5.83257914e-01 -4.21734042e-02
1.40758801e+00 -3.13968420e-01 9.46310997e-01 6.29872859e-01
6.55440748e-01 -1.02999496e+00 -7.34982342e-02 6.66856647e-01
3.96778196e-01 -1.12365365e+00 3.28836352e-01 -2.52470672e-01
-2.39395529e-01 1.23902583e+00 8.79345000e-01 2.70852238e-01
4.98974681e-01 6.27370775e-01 -1.35821551e-01 -3.35795850e-01
-1.20275509e+00 -5.20803779e-03 1.90673798e-01 -9.99398455e-02
5.33436120e-01 -1.12994285e-02 -5.69281757e-01 3.75791132e-01
-4.19703156e-01 1.12733178e-01 -7.30031654e-02 1.10550451e+00
-2.63749778e-01 -1.53428519e+00 -2.30554909e-01 5.52527189e-01
-3.05398583e-01 -3.71633321e-02 2.37254202e-01 5.46509266e-01
1.92512646e-01 5.94389677e-01 -5.60743660e-02 -3.96757096e-01
2.65700817e-01 -1.14469178e-01 4.93835598e-01 -6.24373138e-01
-8.99646938e-01 8.85257199e-02 3.61893848e-02 -6.83083534e-01
-1.04755707e-01 -4.46591705e-01 -1.07403529e+00 -5.41134536e-01
-1.58500850e-01 3.27462852e-02 9.00559008e-01 9.74832296e-01
8.39033484e-01 5.90250731e-01 2.92441815e-01 -8.39719713e-01
-7.87907600e-01 -4.84760553e-01 -3.66913676e-02 3.82265411e-02
-2.59231538e-01 -5.56828976e-01 9.36146304e-02 -2.17184395e-01] | [8.5361909866333, 3.488621473312378] |
ea4f6e71-74a0-4ea3-8ee0-6e17ee521572 | coap-dos-an-iot-network-intrusion-dataset | 2206.14341 | null | https://arxiv.org/abs/2206.14341v1 | https://arxiv.org/pdf/2206.14341v1.pdf | CoAP-DoS: An IoT Network Intrusion Dataset | The need for secure Internet of Things (IoT) devices is growing as IoT devices are becoming more integrated into vital networks. Many systems rely on these devices to remain available and provide reliable service. Denial of service attacks against IoT devices are a real threat due to the fact these low power devices are very susceptible to denial-of-service attacks. Machine learning enabled network intrusion detection systems are effective at identifying new threats, but they require a large amount of data to work well. There are many network traffic data sets but very few that focus on IoT network traffic. Within the IoT network data sets there is a lack of CoAP denial of service data. We propose a novel data set covering this gap. We develop a new data set by collecting network traffic from real CoAP denial of service attacks and compare the data on multiple different machine learning classifiers. We show that the data set is effective on many classifiers. | ['Shankar Banik', 'Prosenjit Chatterjee', 'Jared Mathews'] | 2022-06-29 | null | null | null | null | ['network-intrusion-detection'] | ['miscellaneous'] | [ 5.65742329e-02 -3.88255030e-01 -3.71291697e-01 -5.06012142e-01
5.67054115e-02 -7.21097112e-01 5.05973876e-01 2.96073984e-02
-4.31146562e-01 4.37837273e-01 -1.54586434e-01 -7.17834651e-01
-4.60151076e-01 -1.36664796e+00 -1.06730096e-01 -5.22943854e-01
9.64379217e-03 7.34703243e-01 7.29348481e-01 -4.09266979e-01
1.97974853e-02 9.74890471e-01 -1.36107278e+00 2.48851106e-01
8.09167512e-03 1.26185358e+00 -6.46005929e-01 4.95146245e-01
-5.01057386e-01 6.79267868e-02 -7.92991877e-01 -3.47957194e-01
6.83416665e-01 3.53351906e-02 -6.78669274e-01 -3.10388327e-01
-2.41207585e-01 -1.08050620e-02 -1.41639754e-01 1.22281027e+00
5.72073281e-01 -2.29193419e-01 1.59698322e-01 -2.26479745e+00
4.89052087e-02 7.53773093e-01 -6.86163455e-02 5.11348605e-01
1.61428154e-01 9.97332204e-03 6.79672062e-01 -9.84037668e-03
9.26300958e-02 1.33703244e+00 3.26420784e-01 8.21832836e-01
-9.00577903e-01 -1.21086788e+00 -1.89343825e-01 2.94459075e-01
-9.80775595e-01 -3.27327132e-01 9.24739957e-01 -9.71964523e-02
9.09558117e-01 3.92481953e-01 3.39798957e-01 1.26873744e+00
3.41951758e-01 -2.05184668e-01 9.09563243e-01 -2.22764984e-01
6.64222956e-01 3.23648006e-01 3.86066139e-01 1.50446385e-01
6.77186906e-01 1.65040851e-01 3.91644895e-01 -3.88025314e-01
4.42974642e-02 8.97760570e-01 3.50803852e-01 7.65632540e-02
-6.74064219e-01 9.49525952e-01 4.61933345e-01 7.39288509e-01
-1.68440551e-01 -1.95262700e-01 6.64132714e-01 6.59629822e-01
-1.42068788e-02 1.79142635e-02 -1.11402106e+00 -7.49535263e-02
-1.30989954e-01 -6.10527694e-01 1.26697767e+00 7.01017916e-01
6.16251767e-01 8.60966593e-02 1.00311482e+00 2.86491245e-01
3.47316116e-01 9.45628643e-01 5.08439660e-01 -4.65884209e-01
8.11139271e-02 9.61275876e-01 -3.58563423e-01 -1.06809914e+00
-7.24308074e-01 8.52272213e-02 -1.12743819e+00 1.14433333e-01
2.04530776e-01 7.02397376e-02 -5.85094392e-01 1.31788254e+00
3.10387969e-01 1.51159197e-01 1.93981573e-01 -1.84060680e-03
4.12254930e-01 5.76487958e-01 3.60150546e-01 -3.19139719e-01
1.41032028e+00 -1.32909670e-01 -5.52742839e-01 -8.44178605e-04
2.93787003e-01 -6.19459450e-01 6.99463785e-01 4.00653809e-01
-3.54812026e-01 -3.10153425e-01 -9.94577289e-01 7.64797151e-01
-1.16398072e+00 -1.01309776e+00 8.33564818e-01 1.53833079e+00
-4.88185763e-01 3.27560872e-01 -6.45301163e-01 -7.75897861e-01
7.72317171e-01 9.21930313e-01 -2.68271327e-01 -6.11879863e-02
-9.61802900e-01 6.86083853e-01 6.04714215e-01 -6.51988804e-01
-6.78272128e-01 -1.78748816e-01 -3.87267053e-01 4.86347266e-02
1.76016256e-01 -3.22729498e-01 6.92412078e-01 -6.71069324e-01
-1.14525819e+00 5.38008451e-01 4.72677112e-01 -5.25339067e-01
2.08921451e-02 4.03219581e-01 -1.52925599e+00 1.72233373e-01
-2.90759236e-01 1.79976955e-01 9.11343932e-01 -8.23019028e-01
-6.38832867e-01 -4.86444354e-01 2.26770435e-02 -1.10195994e+00
-8.64184558e-01 6.34148896e-01 8.34872663e-01 -6.70325831e-02
2.04883978e-01 -9.49983895e-01 -1.95048526e-01 -1.81538284e-01
-4.85356599e-01 -5.00462055e-01 1.88887060e+00 2.51909554e-01
7.67537355e-01 -1.88330984e+00 -8.99711788e-01 4.60556030e-01
5.71273752e-02 6.83076680e-01 -1.37711212e-01 4.73103255e-01
-2.81448781e-01 8.66880298e-01 1.32438436e-01 2.86678851e-01
-5.17255589e-02 7.09860861e-01 -5.35529077e-01 1.84169456e-01
-2.00624779e-01 5.83187997e-01 -6.28012538e-01 -2.43210047e-01
6.73726916e-01 5.20334542e-01 -3.23261946e-01 -1.06154010e-01
-2.53647029e-01 7.03998685e-01 -9.11956429e-01 8.61065209e-01
8.22250009e-01 -1.72632083e-01 1.16069734e-01 -2.90098518e-01
1.58521414e-01 4.07078505e-01 -9.08456385e-01 8.55784297e-01
-5.45712411e-01 1.54300802e-03 -2.66172171e-01 -1.21395016e+00
1.05505443e+00 5.10150194e-01 9.17305589e-01 -6.11795783e-01
9.22966778e-01 2.98133105e-01 3.72937620e-01 -4.16264892e-01
-7.23668039e-01 -2.43592560e-01 -1.76966235e-01 7.90718257e-01
-3.37231934e-01 1.12861507e-01 9.17480811e-02 -5.28702177e-02
1.59090698e+00 -1.08843374e+00 4.38252121e-01 -2.11543664e-01
8.65053296e-01 -2.30561733e-01 6.20759070e-01 7.20672965e-01
-6.64032876e-01 -3.04115146e-01 -5.41229360e-02 -1.02269113e+00
-9.54267204e-01 -1.20858753e+00 -2.25240797e-01 6.59996390e-01
5.03335260e-02 -2.58311093e-01 -3.75859529e-01 -9.96054173e-01
-1.68550193e-01 5.35932064e-01 -1.62089709e-02 -3.53430688e-01
-5.04592299e-01 -9.12642956e-01 6.89625502e-01 2.14015201e-01
1.01044905e+00 -1.20197058e+00 -4.67443794e-01 2.74168670e-01
1.45554379e-01 -1.74460113e+00 2.60414392e-01 4.48031202e-02
-1.00770032e+00 -1.43399787e+00 6.04311407e-01 -5.72474658e-01
4.44786161e-01 4.32996213e-01 1.02090538e+00 3.35599512e-01
-4.65799898e-01 5.06343603e-01 -4.42131281e-01 -1.06152976e+00
-9.92490411e-01 1.19229086e-01 7.90266871e-01 1.07275315e-01
1.18865085e+00 -1.28767550e+00 -3.20170403e-01 9.18667316e-01
-9.54590321e-01 -8.70868444e-01 3.58047485e-01 9.63528156e-02
8.81399289e-02 1.27477133e+00 8.02717745e-01 -3.56653363e-01
9.72963050e-02 -6.90981865e-01 -7.07484663e-01 -2.02676818e-01
-7.06271470e-01 -3.43778968e-01 9.00572717e-01 -8.54663610e-01
-3.82392704e-01 -3.09116811e-01 -3.82598966e-01 -3.29030715e-02
-7.97931194e-01 -4.36441779e-01 -7.48696804e-01 -3.15364122e-01
5.09597957e-01 -4.15466517e-01 -2.11248562e-01 -6.21507227e-01
-1.43636867e-01 1.23180318e+00 -9.45139155e-02 -2.68959492e-01
1.31700587e+00 1.02773714e+00 4.60963160e-01 -1.08339131e+00
-7.84042299e-01 -4.21749830e-01 -3.49156111e-01 1.75270244e-01
7.22501338e-01 -2.37256348e-01 -1.30818927e+00 2.94844776e-01
-8.75991344e-01 3.51016253e-01 -4.83709991e-01 5.10651946e-01
1.00697689e-01 7.17810243e-02 -4.67450529e-01 -9.02691662e-01
-5.23909926e-01 -1.08507729e+00 3.84330034e-01 1.94573164e-01
1.83617941e-03 -8.69878411e-01 -1.44600151e-02 2.94390827e-01
8.24708045e-01 1.65954784e-01 1.03126800e+00 -1.53244054e+00
-4.09975320e-01 -7.28485942e-01 -2.77627617e-01 5.43521345e-01
8.33394051e-01 -7.68814906e-02 -9.11492825e-01 -4.05605286e-01
6.38363123e-01 8.80879313e-02 2.11841673e-01 2.56727897e-02
1.14049506e+00 -6.63264275e-01 -5.35893381e-01 3.32280904e-01
1.49836314e+00 6.84833527e-01 8.52737427e-01 3.52598965e-01
4.92291719e-01 6.24461770e-01 -2.47758478e-02 1.26911506e-01
-7.60737211e-02 2.87670165e-01 9.78479266e-01 3.14844996e-01
4.16786402e-01 5.10997884e-02 2.03687195e-02 5.73336959e-01
2.55190343e-01 -4.03361648e-01 -9.01893795e-01 1.30945414e-01
-1.28434098e+00 -1.19433320e+00 -1.28388211e-01 1.96931672e+00
-3.01338658e-02 5.80090225e-01 4.16687697e-01 9.96211946e-01
6.89920366e-01 -1.44484296e-01 -5.12033761e-01 -5.64857185e-01
2.65789591e-02 4.71253693e-01 6.20806932e-01 1.67028323e-01
-1.15648019e+00 5.62290490e-01 5.67158508e+00 4.45533425e-01
-1.20458877e+00 4.48338002e-01 2.59667993e-01 4.22436714e-01
2.36855417e-01 7.98413754e-02 -6.83851659e-01 7.51041770e-01
1.51828921e+00 9.94080678e-03 2.80303657e-01 1.09297311e+00
2.74451464e-01 3.25180560e-01 -6.52025342e-01 1.11305678e+00
-4.73862737e-01 -8.59546602e-01 1.34251088e-01 3.73519659e-01
2.67350137e-01 4.13869828e-01 -2.89502919e-01 -1.00846209e-01
5.83706200e-01 -8.43070209e-01 -2.97516972e-01 -1.52946845e-01
4.15453970e-01 -8.28535140e-01 8.66101205e-01 2.62988150e-01
-1.03099036e+00 -7.33146608e-01 -3.99797469e-01 -3.36198777e-01
1.07773773e-01 9.22380388e-01 -4.95396048e-01 2.01069593e-01
9.50888753e-01 4.43816632e-01 -3.68986875e-01 8.41599941e-01
-6.34986311e-02 7.33576119e-01 -7.60620892e-01 -2.53373355e-01
-2.25501299e-01 5.02192527e-02 6.08210623e-01 6.49850726e-01
1.92983627e-01 1.04504615e-01 5.58510840e-01 2.58894205e-01
-6.04650155e-02 -1.85750276e-01 -9.62162554e-01 -1.26225457e-01
6.44090950e-01 1.33251774e+00 -1.14107108e+00 -1.08285047e-01
-5.62684476e-01 5.63547611e-01 -8.10473084e-01 -2.32853413e-01
-5.34348428e-01 -3.04303080e-01 1.20175409e+00 1.22863285e-01
-8.87389779e-02 -3.08133990e-01 -1.29817486e-01 -1.06844079e+00
-5.62350571e-01 -9.58336651e-01 6.30121589e-01 -2.93461174e-01
-1.96910834e+00 7.94236422e-01 -5.09768389e-02 -1.31590104e+00
1.96440816e-01 -8.39137316e-01 -6.55157626e-01 1.02717066e-02
-1.30669558e+00 -1.07580388e+00 -2.32602596e-01 1.14391565e+00
3.28186929e-01 -4.73177910e-01 1.20708644e+00 5.75554907e-01
-4.53890502e-01 1.14353895e-01 -4.06331897e-01 4.42838430e-01
2.01233745e-01 -3.26454461e-01 4.04514998e-01 6.57928944e-01
2.80031800e-01 5.25046825e-01 5.37234366e-01 -6.06048226e-01
-1.31413364e+00 -1.02567565e+00 6.09182656e-01 -5.98842263e-01
8.23128760e-01 -4.09297943e-01 -4.05666351e-01 8.07255089e-01
-2.39985064e-01 3.57446194e-01 8.66210043e-01 -3.55283588e-01
-8.20583463e-01 -7.43567228e-01 -2.11768103e+00 2.66080827e-01
9.25197363e-01 -3.20037037e-01 -4.02350038e-01 5.49715877e-01
9.26926672e-01 8.32114697e-01 -6.35279834e-01 5.37538886e-01
3.83636445e-01 -9.85085368e-01 1.20104098e+00 -8.54583740e-01
-6.94310665e-01 -3.35176438e-01 -6.26489401e-01 -6.37845218e-01
2.05460787e-01 -7.92952657e-01 1.02470860e-01 1.51624477e+00
9.83145982e-02 -1.54959440e+00 8.69961560e-01 5.30518472e-01
6.37589931e-01 1.47177756e-01 -1.04632819e+00 -1.32112038e+00
-1.03064865e-01 -8.68019581e-01 1.09217095e+00 1.02476633e+00
-1.65452272e-01 5.49639285e-01 6.29091188e-02 3.12401921e-01
9.75754857e-01 -2.74966389e-01 6.73523188e-01 -2.23718977e+00
1.92032471e-01 -1.26523927e-01 -1.01146388e+00 1.15066757e-02
-1.36855328e-02 -6.71459556e-01 -8.09493840e-01 -1.10031843e+00
-1.97333992e-01 -9.19717491e-01 -5.40671349e-01 8.37850034e-01
1.00045288e+00 6.19324327e-01 1.61781237e-02 3.02865326e-01
-1.80638015e-01 -1.20285392e-01 3.59244883e-01 -5.09925485e-01
-1.84254318e-01 6.02860332e-01 -5.20482659e-01 1.18259072e+00
1.56694579e+00 -9.22770321e-01 -5.28543651e-01 1.45345479e-01
2.11958855e-01 -5.03192186e-01 3.74743283e-01 -1.33726323e+00
-2.54737157e-02 -2.72795439e-01 7.78930038e-02 -3.45431566e-01
2.81113625e-01 -1.87915087e+00 2.33318731e-01 9.99392927e-01
3.93855304e-01 2.25469112e-01 -6.67670295e-02 4.65480119e-01
3.94161940e-01 1.22587234e-01 1.12441885e+00 -2.48812944e-01
-3.47499102e-01 7.02238083e-01 -2.17810825e-01 -1.33626536e-01
1.09816456e+00 -1.55133069e-01 -6.21128023e-01 -2.81630546e-01
-6.63153470e-01 -3.37253362e-01 4.95344937e-01 6.59466505e-01
3.61248106e-01 -9.82195854e-01 -8.79983604e-02 5.10291219e-01
-3.17015685e-02 -5.98631859e-01 -4.33124423e-01 3.52299333e-01
-2.57954806e-01 3.88459086e-01 -4.70468462e-01 -4.69239205e-01
-1.30386937e+00 1.19081831e+00 2.55650580e-01 -2.15915337e-01
-6.18005216e-01 2.90137023e-01 -6.14265442e-01 -4.90395844e-01
2.23300740e-01 -8.36364087e-03 -4.36597429e-02 -2.07949147e-01
9.92808402e-01 6.83518648e-01 2.53165364e-01 -6.20228052e-01
-9.35183525e-01 6.24505877e-01 1.99967757e-01 1.42984971e-01
1.17614782e+00 -1.39402196e-01 -4.78844315e-01 6.24092706e-02
1.42246544e+00 4.00579423e-02 -1.79411292e-01 5.14099002e-02
3.04749161e-01 -4.84276026e-01 -2.63609111e-01 -7.07543194e-01
-1.36965406e+00 7.43078232e-01 1.07113373e+00 1.29935968e+00
1.19704032e+00 -1.24274589e-01 1.29485083e+00 7.51008868e-01
1.03623521e+00 -6.04661524e-01 1.26995429e-01 4.41003740e-01
-1.11889280e-01 -1.41212881e+00 -1.19506814e-01 -8.22224379e-01
2.90216416e-01 1.10274601e+00 3.14189374e-01 -1.77882284e-01
1.55971789e+00 6.95291877e-01 1.91256493e-01 -2.94881433e-01
-2.87686765e-01 -2.06843317e-01 -6.08360052e-01 1.11662245e+00
-5.30519426e-01 -8.13604221e-02 -1.78774893e-01 2.33618557e-01
-3.05194348e-01 -1.02477521e-01 5.29965997e-01 1.03054202e+00
-6.19216621e-01 -1.81212425e+00 -3.48755032e-01 3.66776884e-01
-9.25674617e-01 2.36607790e-01 -2.06539348e-01 5.82620025e-01
1.31034166e-01 1.96811271e+00 1.62653163e-01 -8.31539273e-01
1.64927423e-01 1.70898810e-01 7.62060955e-02 -2.85174310e-01
-3.67216080e-01 -6.98128760e-01 -9.78136882e-02 -6.50737107e-01
-5.12709081e-01 2.82962881e-02 -1.16068745e+00 -8.36880922e-01
-2.30882600e-01 1.74334466e-01 1.32758009e+00 1.12291968e+00
3.25199455e-01 3.54844809e-01 1.28514600e+00 -5.06346464e-01
-2.50783861e-01 -5.23157299e-01 -4.86146182e-01 4.78863299e-01
1.45822182e-01 -5.43258429e-01 -9.58559632e-01 -2.76079953e-01] | [5.1560797691345215, 7.15510368347168] |
c746c7f5-9348-41be-ae96-affd62f22c84 | visual-semantic-parsing-from-images-to | 2210.14862 | null | https://arxiv.org/abs/2210.14862v2 | https://arxiv.org/pdf/2210.14862v2.pdf | Visual Semantic Parsing: From Images to Abstract Meaning Representation | The success of scene graphs for visual scene understanding has brought attention to the benefits of abstracting a visual input (e.g., image) into a structured representation, where entities (people and objects) are nodes connected by edges specifying their relations. Building these representations, however, requires expensive manual annotation in the form of images paired with their scene graphs or frames. These formalisms remain limited in the nature of entities and relations they can capture. In this paper, we propose to leverage a widely-used meaning representation in the field of natural language processing, the Abstract Meaning Representation (AMR), to address these shortcomings. Compared to scene graphs, which largely emphasize spatial relationships, our visual AMR graphs are more linguistically informed, with a focus on higher-level semantic concepts extrapolated from visual input. Moreover, they allow us to generate meta-AMR graphs to unify information contained in multiple image descriptions under one representation. Through extensive experimentation and analysis, we demonstrate that we can re-purpose an existing text-to-AMR parser to parse images into AMRs. Our findings point to important future research directions for improved scene understanding. | ['Afsaneh Fazly', 'Vladimir Pavlovic', 'Dhaivat J. Bhatt', 'Kalliopi Basioti', 'Federico Fancellu', 'Zhan Shi', 'Mohamed Ashraf Abdelsalam'] | 2022-10-26 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 6.09071136e-01 4.99753535e-01 -1.31538257e-01 -5.34974098e-01
-2.76615828e-01 -7.66697466e-01 8.65483820e-01 5.55163205e-01
-7.73438737e-02 2.59919494e-01 6.27658963e-01 -5.50494850e-01
1.55330941e-01 -9.42869127e-01 -6.73781514e-01 1.58316065e-02
3.19351137e-01 -5.43836281e-02 2.38967746e-01 -1.62073702e-01
1.66643485e-01 4.48507190e-01 -1.72707796e+00 6.82527900e-01
4.46230441e-01 6.06589854e-01 2.50491291e-01 5.79402387e-01
-7.12154388e-01 1.20518696e+00 -5.22010803e-01 -6.74210727e-01
-1.30475402e-01 -4.59596336e-01 -1.07777703e+00 4.25786734e-01
5.81635356e-01 -9.25597996e-02 -2.50842005e-01 1.04307282e+00
-2.62198716e-01 1.04879752e-01 4.69374299e-01 -1.43650758e+00
-1.03495085e+00 5.56473792e-01 -3.88028890e-01 8.29498097e-03
7.43298948e-01 1.05322272e-01 1.22975373e+00 -8.15392792e-01
9.23973322e-01 1.64421976e+00 3.19356620e-01 5.51190078e-01
-1.35237420e+00 -9.85271111e-02 6.72374606e-01 -2.42147986e-02
-1.29045022e+00 -5.12947202e-01 6.84697926e-01 -5.97870409e-01
1.22188151e+00 4.01956856e-01 7.44906545e-01 8.71031404e-01
-2.06662774e-01 7.18786716e-01 9.78652775e-01 -7.32327223e-01
1.93102241e-01 1.45243749e-01 4.57821012e-01 6.51592791e-01
6.00459754e-01 -4.81547892e-01 -6.06543481e-01 1.47214070e-01
9.58212316e-01 1.56598329e-03 -1.58337265e-01 -5.45595109e-01
-1.23372185e+00 6.68542504e-01 5.65488279e-01 3.67260426e-01
-3.37612063e-01 3.82629275e-01 3.68637055e-01 -7.75019005e-02
4.22661453e-01 3.75761062e-01 9.59946737e-02 -2.94764601e-02
-4.85315442e-01 1.70803577e-01 5.75894415e-01 1.23296106e+00
9.00461018e-01 -2.37711929e-02 -1.55612826e-01 4.99021471e-01
4.87618387e-01 3.52063000e-01 1.71031784e-02 -1.14885545e+00
4.95659530e-01 1.04180813e+00 2.67962702e-02 -1.38165617e+00
-1.85601741e-01 1.04496675e-02 -3.83887351e-01 1.14880003e-01
2.91339517e-01 3.33579779e-01 -1.07744038e+00 1.76540303e+00
1.50647119e-01 -1.48312986e-01 3.03353906e-01 8.35963607e-01
1.03366995e+00 5.63448429e-01 7.95574546e-01 1.88428134e-01
1.96783745e+00 -6.37808263e-01 -7.49114871e-01 -6.82700276e-01
7.29636431e-01 -3.65211070e-01 1.50466812e+00 -2.30742898e-02
-1.19848394e+00 -4.79795963e-01 -9.34035540e-01 -5.55951178e-01
-7.05855012e-01 -1.62412092e-01 7.43983686e-01 5.09021640e-01
-1.13717115e+00 7.98620060e-02 -7.36361444e-01 -7.34909475e-01
4.70696330e-01 -9.55311581e-02 -6.27866328e-01 -8.77942219e-02
-8.11315775e-01 8.31866920e-01 5.79636157e-01 1.48535371e-02
-4.29256529e-01 -4.72697139e-01 -1.43824840e+00 2.13211805e-01
5.41738331e-01 -1.00619149e+00 1.18932533e+00 -1.25485981e+00
-9.88056064e-01 1.27063024e+00 -4.76967692e-01 -3.98612618e-01
-1.25610068e-01 -1.88100040e-01 -2.80057490e-01 4.49630022e-01
2.19847232e-01 8.89228880e-01 3.91619056e-01 -1.66882920e+00
-5.25847793e-01 -4.12807375e-01 7.56023288e-01 2.75215089e-01
-9.59941149e-02 3.24127167e-01 -6.23778760e-01 -7.20893979e-01
2.69011319e-01 -5.84589183e-01 -2.11124763e-01 1.60800442e-01
-3.13826859e-01 -1.09087052e-02 8.53909552e-01 -5.68860590e-01
1.24285388e+00 -2.22634244e+00 2.19709456e-01 3.64025570e-02
4.55566943e-01 5.66637516e-02 -1.79472759e-01 6.79788649e-01
-1.85879216e-01 6.27056718e-01 -1.70172617e-01 -4.58820552e-01
1.10877715e-01 5.39862931e-01 -5.76764762e-01 1.49194360e-01
3.59801769e-01 1.15129983e+00 -9.54501569e-01 -7.81581163e-01
4.32986766e-01 5.70641220e-01 -4.77969795e-01 1.18220322e-01
-4.84302789e-01 3.22366059e-01 -5.28679669e-01 4.51738387e-01
4.37384665e-01 -4.85087633e-01 3.51686805e-01 -4.75404233e-01
-2.95573268e-02 3.35857302e-01 -1.04461122e+00 1.79645944e+00
-5.57204723e-01 8.94035816e-01 -1.14863344e-01 -8.33776712e-01
7.66533017e-01 1.22012168e-01 1.77823473e-02 -5.64411640e-01
-1.35909572e-01 -3.15431327e-01 -3.19836825e-01 -4.95398372e-01
7.44404316e-01 -1.95004687e-01 -2.98892200e-01 4.16127980e-01
6.75690128e-03 -3.09991419e-01 3.54330361e-01 7.55728722e-01
6.85535491e-01 4.79796022e-01 6.66818619e-01 -2.07056124e-02
5.02969205e-01 3.98619950e-01 3.27949256e-01 7.26815045e-01
3.13742608e-02 6.23553813e-01 6.55117273e-01 -5.33709228e-01
-9.86523390e-01 -1.19290888e+00 9.80682895e-02 9.66585517e-01
3.58531207e-01 -1.09000015e+00 -7.46508181e-01 -4.99489397e-01
-2.52045661e-01 1.11514318e+00 -4.50462162e-01 9.75345299e-02
-7.00989604e-01 -9.34781358e-02 3.83037388e-01 8.02843750e-01
2.78205603e-01 -1.11175084e+00 -1.24550176e+00 -7.43325651e-02
-2.02405289e-01 -1.53495538e+00 -5.39256260e-02 -3.23907435e-01
-6.00877225e-01 -9.20358479e-01 -2.40072414e-01 -5.84854960e-01
8.71853769e-01 4.89084721e-01 1.52298462e+00 3.20185989e-01
-3.28765482e-01 1.07007086e+00 -6.67365789e-01 -4.95070666e-01
-4.92994756e-01 -4.63326484e-01 -3.72992903e-01 1.76659841e-02
4.72443163e-01 -4.02901530e-01 -4.32210267e-01 -6.83607608e-02
-1.30991197e+00 8.03303957e-01 2.82186925e-01 2.49695182e-01
7.07985878e-01 -2.45770618e-01 1.55738473e-01 -1.11424112e+00
5.05599499e-01 -3.25100005e-01 -3.20530653e-01 5.89087725e-01
2.93058567e-02 2.39853919e-01 3.93179536e-01 -1.02266289e-01
-1.33890843e+00 -1.00298189e-01 2.89704621e-01 -1.67186588e-01
-5.93913198e-01 6.12157226e-01 -2.33577982e-01 3.64394605e-01
4.20221746e-01 1.51673391e-01 -2.83775687e-01 -1.46107361e-01
1.15914643e+00 4.82491434e-01 7.40334630e-01 -7.54998744e-01
6.19711518e-01 6.46344721e-01 1.45463005e-01 -9.62094605e-01
-9.63970065e-01 -3.35991234e-01 -8.81373405e-01 -2.18709409e-01
1.27957082e+00 -9.64654148e-01 -2.72387385e-01 -1.68908447e-01
-1.36522877e+00 -4.51494344e-02 -6.58065856e-01 2.17254072e-01
-6.49702847e-01 5.63830078e-01 -3.46958101e-01 -8.71349633e-01
7.77549576e-04 -8.84376109e-01 1.36302996e+00 1.60605103e-01
-5.57077408e-01 -1.09353971e+00 -4.48458493e-01 3.69076937e-01
1.18460342e-01 5.41072786e-01 1.29427838e+00 -2.93072104e-01
-7.54738271e-01 8.28451589e-02 -5.65491557e-01 -5.71908131e-02
3.07412416e-01 1.95177048e-02 -8.56408775e-01 1.23559810e-01
-2.19231874e-01 -1.95518851e-01 4.22895193e-01 9.50716063e-02
1.07809055e+00 -3.47470611e-01 -4.68664259e-01 4.04156536e-01
1.58955967e+00 1.90629378e-01 6.98689401e-01 2.72398770e-01
9.55128610e-01 9.05886054e-01 4.12252188e-01 1.13924213e-01
6.24825835e-01 6.29901767e-01 2.94902533e-01 -3.55767995e-01
-5.05015671e-01 -5.53844512e-01 1.32411823e-01 4.97671008e-01
-7.17749670e-02 -3.84845346e-01 -1.05556679e+00 4.56797928e-01
-2.01643205e+00 -1.04027092e+00 -3.61263633e-01 1.89534295e+00
5.29323936e-01 -7.48426020e-02 -6.79489970e-02 -2.47942179e-01
5.42967021e-01 3.94317299e-01 -2.09153011e-01 -5.51107943e-01
-1.78473040e-01 -2.50456799e-02 2.16738552e-01 4.29909348e-01
-7.91612923e-01 1.31243014e+00 6.41848183e+00 2.69558012e-01
-7.50205636e-01 -1.03540801e-01 3.93825769e-01 3.53753120e-01
-7.27416277e-01 4.84003246e-01 -4.36212808e-01 -5.76399118e-02
7.88800418e-01 -1.80216953e-01 3.80854309e-01 7.16890693e-01
1.17478818e-01 -1.75893188e-01 -1.33735108e+00 1.08915615e+00
2.40854472e-01 -1.45210814e+00 8.18413556e-01 -1.58079505e-01
1.67809889e-01 -5.57090700e-01 -1.76909074e-01 -4.25844155e-02
4.43643153e-01 -1.27987766e+00 1.16127181e+00 5.99218726e-01
8.04171085e-01 -3.00697237e-01 2.64035285e-01 2.74401829e-02
-1.64502954e+00 1.49376050e-01 -3.38625461e-01 -3.43492776e-01
5.02693236e-01 1.46156579e-01 -6.74761236e-01 8.90329599e-01
6.17656052e-01 7.54993379e-01 -8.54037821e-01 4.21559423e-01
-4.92448956e-01 2.66181648e-01 1.88061483e-02 1.41727656e-01
7.48513043e-02 -1.95625409e-01 4.01331216e-01 1.43111157e+00
1.29429549e-02 2.96872675e-01 2.61445224e-01 1.25967264e+00
2.67599612e-01 4.89583164e-02 -1.10356688e+00 -3.12685430e-01
4.13024634e-01 1.14549184e+00 -1.14634347e+00 -6.17771685e-01
-8.91604841e-01 8.66390586e-01 3.70949596e-01 5.93210459e-01
-6.69227779e-01 -3.29370618e-01 6.01872206e-01 3.21749210e-01
4.03361283e-02 -5.67213476e-01 -3.42884868e-01 -1.45169795e+00
4.15469892e-02 -7.59110570e-01 4.51543599e-01 -1.55844498e+00
-9.94473755e-01 6.58084154e-01 6.48446739e-01 -9.70476449e-01
-2.55425662e-01 -8.59642386e-01 -3.70280802e-01 7.36329615e-01
-1.22187650e+00 -1.59380829e+00 -4.71272796e-01 4.55650151e-01
6.87669337e-01 3.74895543e-01 8.75642478e-01 -2.37523198e-01
-3.31412613e-01 2.86601782e-02 -7.82830179e-01 1.96746260e-01
7.97389969e-02 -1.22573388e+00 6.78666890e-01 1.19162917e+00
7.16111064e-01 1.01335311e+00 7.45283544e-01 -6.15338862e-01
-1.42171109e+00 -9.06676292e-01 7.93679416e-01 -7.24407136e-01
5.67265511e-01 -7.26980805e-01 -9.13811624e-01 1.18981683e+00
4.94563311e-01 2.12861281e-02 7.48985112e-01 -1.45847048e-03
-6.89726710e-01 3.40198338e-01 -8.79558504e-01 1.12388384e+00
1.45128858e+00 -8.65331292e-01 -1.08698475e+00 -7.60288537e-02
1.07705522e+00 -3.22027773e-01 -7.20285237e-01 2.67512619e-01
2.77314067e-01 -9.41835403e-01 1.23951566e+00 -8.71787369e-01
3.78318310e-01 -5.27831614e-01 -4.27912295e-01 -9.76777732e-01
-2.12282494e-01 -2.40075558e-01 6.77525997e-02 1.37894070e+00
1.56520560e-01 -4.32823777e-01 4.26357597e-01 9.28340197e-01
-8.21054131e-02 -2.47321129e-01 -5.10956347e-01 -4.73605245e-01
-2.62975365e-01 -8.11137259e-01 6.51825190e-01 7.99151480e-01
1.99064583e-01 6.41862333e-01 1.25354737e-01 2.75979429e-01
4.31749672e-01 3.24938059e-01 8.67435157e-01 -1.25575125e+00
-2.71492023e-02 -2.14709371e-01 -6.73669696e-01 -9.50624406e-01
3.63993913e-01 -8.76121879e-01 -1.91532120e-01 -2.21062160e+00
2.71620005e-01 -2.69074142e-01 9.58444849e-02 7.84847498e-01
-1.02094881e-01 1.14969686e-01 6.58000231e-01 1.52661219e-01
-8.05746198e-01 2.31816709e-01 1.04616868e+00 -8.66948739e-02
-9.69960988e-02 -7.42403686e-01 -9.76181567e-01 1.11350322e+00
5.97914457e-01 -2.59717882e-01 -7.24963546e-01 -7.15100586e-01
4.20262545e-01 2.40139645e-02 7.14019001e-01 -8.18831146e-01
4.50943895e-02 -4.65645134e-01 2.07946584e-01 -3.48827273e-01
4.20793205e-01 -8.92065585e-01 3.81896049e-01 7.20713288e-02
-4.06471938e-01 3.92533481e-01 3.27113688e-01 5.02555370e-01
-3.19980830e-01 -2.33060583e-01 2.84939557e-01 -4.59445983e-01
-1.18697977e+00 -2.15642065e-01 -3.81353587e-01 1.26981333e-01
1.03100312e+00 -7.21531749e-01 -5.70239663e-01 -5.32575727e-01
-7.87874401e-01 -1.51218278e-02 7.70386934e-01 5.21664262e-01
7.40663350e-01 -1.15280771e+00 -2.74343759e-01 5.06066009e-02
4.94222432e-01 1.10515565e-01 1.88240170e-01 2.72488326e-01
-4.95021284e-01 4.38726276e-01 -2.90025085e-01 -6.54744983e-01
-1.37071383e+00 7.95607150e-01 8.68122876e-02 1.23453982e-01
-1.02724540e+00 3.13063443e-01 9.41093266e-01 -6.07949756e-02
-4.84995581e-02 -7.91373789e-01 -2.25318670e-01 -1.78688332e-01
4.98555630e-01 -1.02373712e-01 -4.43124890e-01 -1.14092326e+00
-4.49423850e-01 8.07182968e-01 2.95679361e-01 -3.23283941e-01
1.07902336e+00 -5.00861943e-01 -2.81209439e-01 5.38131475e-01
8.42939973e-01 2.58548170e-01 -1.10600805e+00 -1.42109901e-01
2.81811148e-01 -5.23834825e-01 -2.70883113e-01 -6.09225988e-01
-6.44271910e-01 9.22286630e-01 1.15461521e-01 5.06325126e-01
1.14264512e+00 6.13743722e-01 3.13821077e-01 2.42763892e-01
3.79957676e-01 -5.20508766e-01 1.46230564e-01 2.70791739e-01
1.00090444e+00 -8.60279083e-01 1.10925816e-01 -8.92585576e-01
-9.33640778e-01 1.27084494e+00 6.12383842e-01 2.12165207e-01
1.32176518e-01 2.27246970e-01 8.69642496e-02 -6.08754337e-01
-5.44955909e-01 -6.05193555e-01 3.67477804e-01 8.96865368e-01
5.15868723e-01 3.74144278e-02 7.88106304e-03 5.08960485e-01
-1.82673603e-01 -1.48482636e-01 6.90463781e-01 1.13135707e+00
-3.35820436e-01 -1.02876055e+00 -3.24219882e-01 -3.69186737e-02
-4.69171762e-01 -2.06957445e-01 -5.88186264e-01 1.02989578e+00
-8.15145904e-04 1.01305139e+00 4.94217932e-01 -7.48457434e-03
4.40972269e-01 9.67050195e-02 6.24187410e-01 -1.01784945e+00
-2.22273096e-01 -3.63200456e-01 2.72440284e-01 -6.19391859e-01
-8.89148414e-01 -3.96928042e-01 -1.55508804e+00 1.72099814e-01
1.69669971e-01 -1.08307697e-01 4.01331902e-01 8.96570027e-01
4.90383834e-01 5.04903734e-01 -2.23627329e-01 -6.31458759e-01
2.64349699e-01 -4.08180773e-01 -1.78904697e-01 9.12189066e-01
2.74085969e-01 -5.20467579e-01 5.64329289e-02 6.02347314e-01] | [10.58191967010498, 1.6230825185775757] |
114277cc-2b5a-47db-aa1a-ccd5228c7a4b | real-time-detection-of-false-readings-in | null | null | https://ieeexplore.ieee.org/document/9765458 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9765458 | Real-Time Detection of False Readings in Smart Grid AMI Using Deep and Ensemble Learning research model code github | In the advanced metering infrastructure, smart meters are deployed at the consumers’ side to
regularly transmit fine-grained electricity consumption readings to the system operator (SO) for billing and
real-time load monitoring and energy management. However, fraudulent consumers may compromise their
meters to launch electricity-theft cyberattacks by reporting low-consumption readings to reduce their bills.
These false readings not only cause financial losses but also degrade the grid’s performance because they are
used for energy management and load estimate. The existing solutions in the literature focus only on securing
the billing, so they are not designed to detect the attacks in real time, and thus the SO may use false readings
for a long period of time in load monitoring and energy management until they are identified. In this paper,
we propose a general ensemble-based deep-learning detector that enables the SO to detect false readings in
real time. To do that, we first train several deep learning models on samples generated from a sliding window
of the readings. Then, we use the best-performing model to train several models on different ratios of false
readings and use them in our ensemble-based detector. Extensive experiments are conducted, and the results
indicate that comparing to the literature, our detector can detect the false readings after sending a few false
readings (around 15) comparing to the existing daily and weekly detection approaches that need 144 and
1,008 readings, respectively. | ['AND ABDULLAH M. ABUSORRAH', 'AHMAD H. MILYANI', 'ABDULAH JEZA ALJOHANI', 'JUNAID KHALID', 'MOHAMED M. E. A. MAHMOUD', 'Mohamed I. Ibrahem', 'MOHAMMED J. ABDULAAL'] | 2022-04-29 | null | null | null | ieee-access-2022-4 | ['ensemble-learning', 'ensemble-learning', 'management', 'energy-management'] | ['computer-vision', 'methodology', 'miscellaneous', 'time-series'] | [-1.62727982e-01 -2.01548308e-01 2.41087601e-01 -2.86570847e-01
-6.15547180e-01 -7.17964828e-01 4.55083102e-01 3.14016998e-01
-2.31024712e-01 7.30674505e-01 -2.47016564e-01 -5.60116768e-01
1.77581757e-01 -1.39957952e+00 -2.79576480e-01 -9.51922476e-01
-2.56649703e-01 3.96960288e-01 8.46258253e-02 8.50876719e-02
1.60194457e-01 2.57668614e-01 -1.00338733e+00 3.96946259e-02
9.98991430e-01 1.44070423e+00 -3.88661891e-01 6.87209964e-01
2.97407776e-01 5.59201360e-01 -1.47955763e+00 -1.40155569e-01
3.37168604e-01 -4.04415488e-01 -3.48627448e-01 -2.17378378e-01
-5.25468886e-01 -1.25477695e+00 -4.27171625e-02 1.42375553e+00
6.70229614e-01 -3.03281099e-01 3.46719116e-01 -1.77339399e+00
-3.92569095e-01 1.07055581e+00 -5.50224781e-01 4.67715144e-01
8.36948827e-02 3.72075379e-01 8.14553499e-01 -3.21650207e-01
-6.93289459e-01 5.45362592e-01 5.37306011e-01 1.45797372e-01
-7.50218868e-01 -1.11895680e+00 -1.31455302e-01 6.26551151e-01
-1.31866014e+00 -1.68931693e-01 9.19416487e-01 2.98329815e-02
1.30824673e+00 4.61312115e-01 8.18082273e-01 8.61230910e-01
3.62991899e-01 5.56195557e-01 9.48248625e-01 -1.58283636e-01
7.50217974e-01 -7.88059011e-02 6.85142726e-02 -1.62201673e-01
7.90737808e-01 2.06641421e-01 8.85696858e-02 -4.65067714e-01
2.57146865e-01 1.30942211e-01 -3.16917270e-01 6.99523687e-01
-7.05460608e-01 1.02970243e+00 6.14227653e-02 6.48977518e-01
-7.71576762e-01 -4.01303777e-03 6.42436743e-01 3.04896504e-01
2.92419761e-01 1.63840562e-01 -5.90531945e-01 -1.59139663e-01
-9.94476318e-01 -1.36915013e-01 1.11303961e+00 4.72711474e-01
3.52258265e-01 6.73547089e-01 1.69760302e-01 -2.15839799e-02
1.60889730e-01 6.27750993e-01 9.37804937e-01 -2.01805562e-01
3.19204062e-01 5.47617137e-01 4.86328453e-01 -8.79947841e-01
-5.90969443e-01 -2.94130147e-01 -1.29083502e+00 2.75371015e-01
1.15478644e-02 -5.28769732e-01 -3.13590407e-01 1.01926625e+00
1.73112810e-01 4.61911857e-01 -5.85125051e-02 5.87390006e-01
2.24847019e-01 6.13407016e-01 -5.22776097e-02 -4.54287410e-01
1.29296303e+00 -4.12643909e-01 -9.25350726e-01 2.24224791e-01
4.64395434e-01 -3.63782316e-01 5.19425988e-01 7.41624892e-01
-7.96596229e-01 -3.60978335e-01 -1.64328372e+00 7.28364825e-01
-6.61040187e-01 -3.40426564e-02 1.82039469e-01 9.49031413e-01
-6.03416085e-01 8.31025422e-01 -9.42635536e-01 1.76778615e-01
3.94922614e-01 2.66992629e-01 3.79602939e-01 6.44018233e-01
-1.58862555e+00 1.08510494e+00 7.10470378e-01 2.71878093e-01
-6.94904327e-01 -3.46851975e-01 -5.90122759e-01 2.88139939e-01
-2.03755543e-01 -1.43741310e-01 1.45784271e+00 -5.75414419e-01
-1.31747377e+00 -1.43485516e-02 4.45698142e-01 -1.22208893e+00
7.82599092e-01 -1.68858364e-01 -1.13571155e+00 -2.61913240e-03
-6.87423199e-02 -4.22656357e-01 7.97262967e-01 -8.16252708e-01
-1.02109098e+00 -2.28213593e-01 -2.32281536e-01 -3.75185817e-01
-3.18844467e-01 -4.49471325e-01 1.01062810e+00 -4.61855143e-01
-2.68074065e-01 -2.73995727e-01 1.99597273e-02 -9.88001764e-01
-6.82519376e-01 -6.46628678e-01 1.37013555e+00 -9.30338621e-01
1.29903531e+00 -1.71498144e+00 -1.11577058e+00 4.22457486e-01
6.50732592e-02 7.68187761e-01 5.22860885e-01 6.63926899e-01
-2.65738696e-01 1.90417916e-01 -5.46532534e-02 -1.59965411e-01
3.62236649e-01 3.21092188e-01 -6.44554675e-01 7.61621177e-01
1.59684315e-01 9.55491841e-01 -9.50973868e-01 2.92872280e-01
1.01392984e+00 2.34185666e-01 3.28777909e-01 4.51949060e-01
-1.63713098e-01 2.13165581e-01 -4.59679514e-01 3.41682673e-01
1.04452658e+00 -2.30500296e-01 1.64227128e-01 -4.43487734e-01
7.15832561e-02 6.08186424e-01 -1.31139839e+00 5.08975744e-01
-3.35173517e-01 2.52146900e-01 -3.42367947e-01 -1.46468425e+00
8.73626471e-01 5.08051634e-01 7.38546073e-01 -8.89815509e-01
4.86520201e-01 2.14191377e-01 -1.76783621e-01 -2.55054712e-01
8.50661844e-02 1.15388893e-01 -2.63046026e-01 9.20669079e-01
-6.43397868e-01 1.79348558e-01 3.02734114e-02 -1.12281509e-01
1.39755404e+00 -5.39472878e-01 5.94374180e-01 9.26817283e-02
7.76112080e-01 -3.80883873e-01 5.91825366e-01 6.13589287e-01
-2.37058625e-01 -3.89087349e-01 -8.75921845e-02 -6.72556579e-01
-9.70567763e-01 -1.06733108e+00 -2.74467081e-01 4.87365097e-01
-6.20093942e-02 -3.73985134e-02 -5.99547803e-01 -8.07377279e-01
2.44337514e-01 1.59642518e+00 -2.19501749e-01 -3.33344072e-01
-5.56940854e-01 -1.11469924e+00 6.66311681e-01 7.04551458e-01
1.16751194e+00 -1.09712148e+00 -1.13696754e+00 7.87693024e-01
-4.18221280e-02 -9.45409656e-01 -2.56102204e-01 3.81454587e-01
-2.28593707e-01 -1.42051196e+00 -9.07997787e-02 -5.51114492e-02
4.22593206e-01 2.59948410e-02 1.12405646e+00 2.75967568e-01
-9.44843590e-02 -6.30205050e-02 -4.50915068e-01 -6.45956278e-01
-8.34826708e-01 -2.63254851e-01 2.85267085e-01 -2.39177328e-02
1.03202271e+00 -8.86245012e-01 -7.42644727e-01 1.72106043e-01
-9.72496688e-01 -5.41528404e-01 3.03948760e-01 7.55483031e-01
3.58412578e-03 9.28207457e-01 1.43689096e+00 -5.86090744e-01
8.42709601e-01 -6.01053715e-01 -1.17406571e+00 8.84184018e-02
-1.21956336e+00 -2.32391387e-01 1.28413296e+00 -2.87106097e-01
-3.49733651e-01 -3.32866788e-01 -1.84229895e-01 -7.63803050e-02
-2.82541454e-01 -6.01190217e-02 -3.03215563e-01 4.00363356e-02
1.17137600e-02 5.71571529e-01 -4.44024801e-01 -5.00910103e-01
-5.70525341e-02 9.72020566e-01 5.92075944e-01 2.41078153e-01
1.18924868e+00 1.99205175e-01 -3.30673337e-01 -3.89657944e-01
-5.08100271e-01 -1.79792225e-01 -2.03013018e-01 1.05787227e-02
3.59213740e-01 -8.58960748e-01 -1.43676198e+00 9.91255581e-01
-1.15884936e+00 1.58118710e-01 -2.95083433e-01 3.06214660e-01
1.28263161e-02 5.77759862e-01 -6.77564800e-01 -1.17911363e+00
-9.77591634e-01 -8.33453298e-01 7.27191329e-01 3.88697118e-01
-1.76698118e-01 -9.59074795e-01 1.84855107e-02 -1.26713917e-01
6.76126778e-01 5.19565642e-01 7.11297154e-01 -1.26107287e+00
-3.73605549e-01 -6.80226147e-01 -2.69945413e-02 9.24503446e-01
5.78837454e-01 -3.21175277e-01 -9.76400197e-01 -7.44489908e-01
6.07672155e-01 -1.70867685e-02 1.45426840e-01 7.10724443e-02
1.19693553e+00 -8.93819511e-01 -1.18412219e-01 4.27933097e-01
1.61559176e+00 6.32351339e-01 7.27098703e-01 4.35128748e-01
1.39137715e-01 -3.00602436e-01 2.08109826e-01 1.03137171e+00
4.93594587e-01 1.64130241e-01 8.43595803e-01 1.93161666e-02
8.57998729e-01 -1.72231674e-01 3.63300741e-01 8.11189473e-01
2.77192771e-01 -4.05187488e-01 -2.18435153e-01 3.75565141e-01
-1.56751287e+00 -1.20324576e+00 -1.00824364e-01 2.34461236e+00
6.64866805e-01 4.27749068e-01 2.51451999e-01 8.55065584e-01
7.12209821e-01 8.98630619e-02 -1.12597013e+00 -2.77473062e-01
2.95866337e-02 3.27601969e-01 8.93998384e-01 8.83508250e-02
-1.01940310e+00 6.43547550e-02 5.68272352e+00 5.65650284e-01
-1.01997042e+00 2.33895913e-01 6.35259092e-01 5.49744405e-02
-5.14865667e-03 -3.12519163e-01 -6.64437294e-01 1.23194611e+00
1.34541416e+00 -5.19462824e-01 3.27009439e-01 9.25929725e-01
7.41101086e-01 -2.73866475e-01 -1.06695223e+00 9.47806060e-01
-3.37283403e-01 -1.08323550e+00 -4.32295024e-01 1.88159466e-01
5.71623683e-01 6.27169833e-02 -6.11213446e-01 1.73167318e-01
9.63241816e-01 -9.77440357e-01 4.40503329e-01 2.92488724e-01
1.22943312e-01 -1.26728261e+00 1.34136856e+00 5.71838796e-01
-1.35982823e+00 -2.87037104e-01 -3.09678853e-01 -1.17614046e-01
4.55055267e-01 1.31283617e+00 -1.01666760e+00 7.68018723e-01
8.82107615e-01 2.98913866e-01 -3.09354186e-01 7.77786195e-01
-4.15536612e-01 9.89058852e-01 -7.26972759e-01 -2.50965893e-01
7.90262446e-02 -8.19268525e-02 1.02200106e-01 7.25079656e-01
3.79101008e-01 5.68862297e-02 3.69996846e-01 9.53229547e-01
-6.94615021e-02 -5.65890074e-01 -2.47962981e-01 2.84789622e-01
9.68175530e-01 1.27791929e+00 -5.45945585e-01 -5.66903532e-01
-2.46636465e-01 9.46928799e-01 -2.48227447e-01 1.38903588e-01
-1.07430005e+00 -4.54951823e-01 4.72924143e-01 -2.22438872e-01
3.45098913e-01 -6.67020455e-02 -3.92373800e-01 -1.01459360e+00
2.41382316e-01 -5.65715134e-01 4.64929223e-01 -3.46553385e-01
-1.95122468e+00 1.16715044e-01 -3.33935589e-01 -1.27224720e+00
-8.09864283e-01 -1.34144813e-01 -1.34495044e+00 8.25670660e-01
-1.63357449e+00 -5.21326959e-01 -2.08863631e-01 8.02245021e-01
3.83765817e-01 8.94881487e-02 9.43660021e-01 2.07094476e-01
-6.60672605e-01 6.75474882e-01 4.64782596e-01 7.05276489e-01
-6.84057698e-02 -1.37094629e+00 7.52810061e-01 1.04188764e+00
-6.04046360e-02 -2.47687921e-01 7.01298594e-01 -5.71539819e-01
-1.21393490e+00 -1.10180116e+00 4.71321195e-01 2.87701227e-02
7.69197583e-01 -1.47130996e-01 -8.87714088e-01 6.17002606e-01
5.73889494e-01 -1.03133366e-01 6.20671988e-01 -7.53780484e-01
-6.38939515e-02 -3.19167733e-01 -1.85475063e+00 -1.62819326e-01
6.52683824e-02 -2.22134024e-01 -6.46152496e-01 3.25159669e-01
1.77280113e-01 5.82719482e-02 -1.05003214e+00 2.07547337e-01
2.37974942e-01 -1.22046280e+00 5.45451760e-01 -9.42325592e-02
-5.61650157e-01 -3.63039374e-01 1.12457938e-01 -1.74107397e+00
-1.74988374e-01 -7.48609841e-01 -5.65835476e-01 1.38985848e+00
-1.41671672e-01 -1.48309064e+00 5.29338419e-01 4.48001802e-01
2.63881207e-01 -2.91267782e-01 -1.31720054e+00 -6.75516605e-01
-2.81447247e-02 -5.12047470e-01 1.61406779e+00 9.68310416e-01
8.32467824e-02 3.85722667e-02 -2.29760244e-01 8.93226266e-01
1.00312304e+00 2.02986568e-01 4.74579334e-01 -1.04499221e+00
-1.61046088e-01 -3.64372820e-01 -3.69593829e-01 -6.67467773e-01
-2.77370036e-01 -2.53217608e-01 -7.13397786e-02 -1.20938993e+00
-3.78987283e-01 -7.18302205e-02 -5.78202724e-01 7.10209608e-01
1.30906135e-01 7.07235187e-02 -5.54399453e-02 6.57549649e-02
-1.04013845e-01 5.20886898e-01 1.97914258e-01 -3.09168309e-01
1.38543770e-01 4.57532585e-01 -3.57399076e-01 8.03669810e-01
1.29108763e+00 -2.73948580e-01 -1.87726736e-01 -5.92397004e-02
-1.48874745e-02 -4.26339358e-02 5.52240491e-01 -1.16731596e+00
1.46015182e-01 -1.53776649e-02 9.06690955e-01 -1.24756980e+00
-3.84722859e-01 -1.19640124e+00 2.26550087e-01 9.75767136e-01
2.46824875e-01 4.16161627e-01 -2.49481604e-01 4.03146446e-01
1.48357347e-01 -2.96944380e-01 7.62730956e-01 -1.60318941e-01
-4.53858227e-01 8.95341784e-02 -5.56663215e-01 -3.14548075e-01
1.39761031e+00 1.24287657e-01 -5.23452282e-01 -5.02448916e-01
-4.90040511e-01 6.17075682e-01 9.41852480e-02 3.56958836e-01
5.43126166e-01 -1.43436384e+00 -7.33161390e-01 5.97587109e-01
-3.71152550e-01 -1.22719128e-02 -8.32759589e-02 4.88830239e-01
-7.20075294e-02 3.63201618e-01 1.61917433e-01 -5.63731849e-01
-6.25213742e-01 3.98582339e-01 7.07241535e-01 -5.97963989e-01
-6.83275759e-01 4.84183803e-02 -6.52183354e-01 1.19428061e-01
1.90959096e-01 -5.52617490e-01 -3.89005989e-02 9.54548642e-02
1.05341268e+00 7.67733932e-01 5.62856376e-01 -1.92975491e-01
-5.06624758e-01 -5.60378954e-02 4.52564023e-02 5.36285162e-01
1.32464087e+00 -2.13315144e-01 1.96088273e-02 2.70422816e-01
1.09155405e+00 -3.10645401e-01 -9.76584256e-01 -6.75622746e-03
1.68092385e-01 -3.00316036e-01 1.30690053e-01 -9.34100270e-01
-1.34538758e+00 5.36140263e-01 8.44670832e-01 1.34342599e+00
1.38970864e+00 -4.42270100e-01 1.41433764e+00 1.68110505e-01
7.91704655e-01 -1.19413674e+00 -3.47715229e-01 7.67644122e-03
2.28006780e-01 -9.47771847e-01 -8.78706947e-03 3.60242963e-01
-1.43588021e-01 1.21186185e+00 2.67233074e-01 -3.17966729e-01
7.71204233e-01 7.92045534e-01 -1.57555997e-01 7.74255767e-02
-5.91357887e-01 2.40710407e-01 -5.16673028e-01 8.20531905e-01
-2.32541040e-01 6.17240191e-01 -1.37576535e-01 8.02217185e-01
-5.43038785e-01 1.98444694e-01 7.57387638e-01 7.38893509e-01
-5.18703818e-01 -9.76118982e-01 -4.65756685e-01 8.12823057e-01
-5.71840167e-01 9.87207070e-02 1.36172965e-01 3.64558101e-01
1.50009617e-01 1.54826820e+00 3.22557449e-01 -4.44681972e-01
3.20862085e-01 1.60998777e-01 -3.10281694e-01 8.57911408e-02
-6.17533624e-01 -1.46018505e-01 -2.32434832e-02 -3.80299449e-01
-1.17438294e-01 -5.10058522e-01 -1.44025564e+00 -9.29724574e-01
-6.45308137e-01 3.44560146e-01 4.38694447e-01 1.30330539e+00
-1.43419532e-02 4.27755535e-01 1.65553808e+00 -6.96168780e-01
-1.47287440e+00 -1.29777026e+00 -1.07583821e+00 3.71328533e-01
4.88734305e-01 -2.26450667e-01 -9.59696531e-01 -4.27771062e-01] | [6.027786731719971, 2.595210075378418] |
bf0f0c7a-a79a-4d0d-9c89-5f8358299a6d | unmasking-communication-partners-a-low-cost | 2011.03630 | null | https://arxiv.org/abs/2011.03630v1 | https://arxiv.org/pdf/2011.03630v1.pdf | Unmasking Communication Partners: A Low-Cost AI Solution for Digitally Removing Head-Mounted Displays in VR-Based Telepresence | Face-to-face conversation in Virtual Reality (VR) is a challenge when participants wear head-mounted displays (HMD). A significant portion of a participant's face is hidden and facial expressions are difficult to perceive. Past research has shown that high-fidelity face reconstruction with personal avatars in VR is possible under laboratory conditions with high-cost hardware. In this paper, we propose one of the first low-cost systems for this task which uses only open source, free software and affordable hardware. Our approach is to track the user's face underneath the HMD utilizing a Convolutional Neural Network (CNN) and generate corresponding expressions with Generative Adversarial Networks (GAN) for producing RGBD images of the person's face. We use commodity hardware with low-cost extensions such as 3D-printed mounts and miniature cameras. Our approach learns end-to-end without manual intervention, runs in real time, and can be trained and executed on an ordinary gaming computer. We report evaluation results showing that our low-cost system does not achieve the same fidelity of research prototypes using high-end hardware and closed source software, but it is capable of creating individual facial avatars with person-specific characteristics in movements and expressions. | ['Christian Geiger', 'Ralf Dörner', 'Alexander Pech', 'Philipp Ladwig'] | 2020-11-06 | null | null | null | null | ['face-reconstruction'] | ['computer-vision'] | [ 6.83610216e-02 5.23293018e-01 6.42683148e-01 -3.85180175e-01
-5.67569971e-01 -4.98678535e-01 2.03349918e-01 -1.19434762e+00
-9.76040140e-02 4.64112014e-01 -2.23865852e-01 -2.91150361e-01
8.30628455e-01 -5.58706343e-01 -8.38721395e-01 -1.09215103e-01
1.90258130e-01 2.48004779e-01 -2.09369659e-02 -5.10411918e-01
-4.42055792e-01 7.61440396e-01 -2.00187278e+00 1.10654496e-01
-1.54773593e-02 9.69597220e-01 -2.93038040e-01 1.22822213e+00
4.41931635e-01 8.17480564e-01 -9.14881587e-01 -6.37717307e-01
5.61909676e-01 -3.26128423e-01 -3.54771256e-01 1.81412309e-01
7.38395929e-01 -8.80684376e-01 -5.19472837e-01 4.00061339e-01
9.42828119e-01 -2.54817232e-02 -6.94766045e-02 -1.63408709e+00
-4.08078849e-01 -2.12726206e-01 -4.97945786e-01 -5.08210063e-01
1.15179944e+00 3.89503479e-01 1.62510186e-01 -6.44354224e-01
8.55015635e-01 1.24572825e+00 9.89670813e-01 1.46964133e+00
-1.22673368e+00 -9.92261469e-01 -2.16826841e-01 -6.47361279e-01
-1.85360074e+00 -7.74865031e-01 6.98064804e-01 -4.69009459e-01
9.69677866e-01 4.59387481e-01 9.73134339e-01 1.75901353e+00
1.94792315e-01 6.44965917e-02 1.07421243e+00 -1.63724288e-01
6.46819696e-02 5.42731225e-01 -6.03510916e-01 9.23447311e-01
-3.19022983e-01 2.48562843e-01 -2.73057461e-01 -2.09301665e-01
1.54027116e+00 -1.41670555e-01 -3.64152133e-01 -1.29834265e-01
-6.58268869e-01 4.89736408e-01 6.89385608e-02 -3.13018300e-02
-3.99865955e-01 3.34415793e-01 1.86109200e-01 3.09975445e-01
2.97704071e-01 -9.51543301e-02 -4.10433829e-01 -5.93314230e-01
-6.65290594e-01 3.51682395e-01 9.05269861e-01 1.14127469e+00
3.58499676e-01 2.41715640e-01 2.39940405e-01 4.07348007e-01
3.00164729e-01 2.76192933e-01 2.32144997e-01 -1.25256693e+00
-8.71507078e-02 2.27569923e-01 3.38762522e-01 -1.05126166e+00
-3.80213946e-01 2.50077903e-01 -4.47136313e-01 1.03233063e+00
1.42902628e-01 -7.33778298e-01 -8.73681188e-01 1.72591436e+00
6.82306051e-01 7.34258533e-01 -1.22295424e-01 1.13755846e+00
1.33134532e+00 3.16891491e-01 -1.22083046e-01 1.07475795e-01
1.27372658e+00 -4.15253013e-01 -6.73053443e-01 -5.35773411e-02
1.55221477e-01 -5.91875851e-01 1.50152874e+00 2.98483074e-01
-1.14619625e+00 -6.61873460e-01 -1.10514057e+00 -2.25122035e-01
5.65169007e-03 -1.73154980e-01 8.73660564e-01 1.46161187e+00
-1.51588953e+00 5.28658748e-01 -8.76093864e-01 -4.65888023e-01
1.61916479e-01 8.66106927e-01 -7.46211767e-01 3.34622115e-01
-7.12873876e-01 7.93294668e-01 -4.64429617e-01 1.41930982e-01
-7.30202079e-01 -7.57106423e-01 -8.34679127e-01 -3.73658001e-01
1.53318346e-01 -9.91128445e-01 1.38307595e+00 -1.33105052e+00
-2.54341030e+00 1.20366800e+00 2.45244935e-01 2.22565934e-01
7.95958757e-01 -9.73444209e-02 -6.02978170e-01 1.45039901e-01
-5.03469765e-01 7.18263507e-01 8.80877554e-01 -1.29659545e+00
-5.92058152e-02 -5.03908575e-01 1.84791014e-01 5.18181510e-02
-1.02735884e-01 4.37807590e-01 -3.96911263e-01 -1.33617505e-01
-1.93714812e-01 -1.26364684e+00 1.29395276e-01 5.30577421e-01
1.68282379e-04 4.37379003e-01 1.19965255e+00 -8.22920501e-01
4.28611338e-01 -2.17889214e+00 -1.27887577e-01 1.80517673e-01
2.72875011e-01 2.62434661e-01 2.14421585e-01 -1.64930653e-02
-3.07727940e-02 2.58335546e-02 2.67025083e-01 -5.73122799e-01
-5.19888438e-02 2.06098303e-01 8.32252800e-02 6.21839762e-01
-1.87011063e-01 8.93885970e-01 -3.13477814e-01 -3.28520060e-01
4.77782369e-01 1.29928088e+00 -6.30326629e-01 5.40140510e-01
2.78384477e-01 9.74397421e-01 3.73918675e-02 5.89891136e-01
7.71783710e-01 2.10838214e-01 2.50369757e-01 1.45478234e-01
2.31191248e-01 -7.94368703e-03 -1.31672502e+00 1.67665827e+00
-7.85642028e-01 7.63551235e-01 6.56825781e-01 -5.03896773e-02
9.17735636e-01 6.59644008e-01 2.29501694e-01 -3.60941201e-01
4.69367236e-01 -8.06771964e-02 -5.42460978e-01 -5.78733325e-01
4.93396074e-01 -4.06978637e-01 -7.85156265e-02 2.44450524e-01
-9.16837081e-02 -3.06559950e-01 -9.00499761e-01 -1.71027362e-01
1.06197369e+00 5.84286690e-01 -4.98302244e-02 4.60768580e-01
1.93750095e-02 -5.03202558e-01 3.62070680e-01 2.02921227e-01
-5.04431464e-02 1.12562597e+00 1.98799074e-01 -3.70152891e-01
-1.08996284e+00 -9.71609712e-01 1.90571979e-01 8.51735532e-01
-1.15376137e-01 -3.67185622e-01 -9.77465332e-01 -1.36404812e-01
-7.67732039e-02 5.04837453e-01 -6.13210618e-01 6.80881739e-02
-5.19915760e-01 1.03489354e-01 8.88286769e-01 6.33046150e-01
2.84573793e-01 -8.78405869e-01 -9.31805253e-01 8.18028022e-03
2.93971449e-01 -1.28365231e+00 -2.46928826e-01 -7.00104475e-01
-5.64535320e-01 -7.11486459e-01 -7.77814031e-01 -6.33375704e-01
7.58461595e-01 7.68595412e-02 1.12267780e+00 7.81175941e-02
-4.07909483e-01 7.60687411e-01 -1.31718859e-01 -3.62290055e-01
-1.61563680e-01 -4.97367650e-01 1.58190921e-01 6.03441373e-02
1.00733511e-01 -7.78127491e-01 -5.50544500e-01 4.45407450e-01
-3.48103851e-01 -2.40397099e-02 -3.63533907e-02 4.95566756e-01
2.92536288e-01 -3.20036650e-01 -7.06316456e-02 -8.16024244e-01
6.65448844e-01 -3.62717994e-02 -6.06792152e-01 -2.57261813e-01
9.08162594e-02 -7.77297437e-01 5.39581537e-01 -6.77745640e-01
-1.02321398e+00 4.67431962e-01 -5.78524113e-01 -8.39973629e-01
-3.24099302e-01 -2.76067317e-01 -3.67795408e-01 -6.51841223e-01
6.40041411e-01 -3.29835594e-01 3.34117204e-01 -9.26489159e-02
2.93995142e-01 1.00441730e+00 8.01279187e-01 -3.38071227e-01
7.22186387e-01 5.27101815e-01 -1.87434688e-01 -1.09671390e+00
1.33953124e-01 1.62338242e-01 -4.19984341e-01 -7.06658125e-01
6.66544139e-01 -1.23256600e+00 -1.67514491e+00 5.44142306e-01
-1.01275754e+00 -5.22597492e-01 -8.28253403e-02 3.95623356e-01
-6.16373777e-01 -2.06689864e-01 -6.96129203e-01 -1.09509587e+00
-2.65458465e-01 -1.01064289e+00 1.40129256e+00 5.22736847e-01
-5.40867805e-01 -5.46144307e-01 -7.74148032e-02 5.40846407e-01
3.94792438e-01 9.18113947e-01 -9.43850875e-02 3.07945549e-01
-3.64169538e-01 -5.65017164e-01 2.79064655e-01 1.13586575e-01
-1.37125343e-01 4.14198518e-01 -1.43500447e+00 -3.29395384e-01
3.69352475e-02 -4.43426430e-01 -2.01970160e-01 3.24472457e-01
9.70719218e-01 -2.64440209e-01 -1.72036842e-01 8.89055550e-01
1.15246665e+00 1.68582126e-01 1.14194477e+00 -1.20310791e-01
9.30125177e-01 5.37136316e-01 2.69027740e-01 3.76014858e-01
3.71435523e-01 9.58317220e-01 3.09375197e-01 -3.41492236e-01
6.05396666e-02 -4.28504020e-01 5.40577114e-01 1.76167637e-01
-5.54524839e-01 -4.49372940e-02 -5.10069132e-01 -2.00762078e-01
-1.19045770e+00 -8.80131006e-01 -6.83784410e-02 2.26488447e+00
4.16420579e-01 -2.48728827e-01 5.08514225e-01 -7.20650032e-02
7.11081803e-01 -5.11824071e-01 -3.45323861e-01 -7.16535091e-01
2.32358158e-01 6.94315672e-01 2.93181211e-01 4.74920720e-01
-6.46277249e-01 8.56431246e-01 6.56122732e+00 -1.40473738e-01
-1.45571434e+00 3.46416324e-01 4.60152715e-01 -5.03815889e-01
-1.19124129e-01 -3.15045893e-01 -3.92281920e-01 1.42726585e-01
1.22670841e+00 3.26022923e-01 4.96291786e-01 1.15533686e+00
2.07472175e-01 2.07310632e-01 -1.03569186e+00 1.30646849e+00
2.08859310e-01 -1.03342271e+00 -7.20413387e-01 3.60829860e-01
3.81738752e-01 -3.85704607e-01 2.07964912e-01 3.19698691e-01
1.17081858e-01 -1.33910131e+00 7.95473337e-01 3.96555841e-01
1.46914327e+00 -9.85358417e-01 5.76525390e-01 1.00349352e-01
-9.29519475e-01 3.27749342e-01 -7.75546804e-02 -2.97858208e-01
1.99845657e-01 -2.79067218e-01 -1.02324462e+00 -1.19816750e-01
7.91700780e-01 -8.77175406e-02 -1.41779497e-01 2.85682231e-01
1.29820751e-02 1.55757949e-01 -4.37499136e-01 -1.05881304e-01
-3.97554219e-01 -1.76899154e-02 3.10866505e-01 7.07171321e-01
4.35904652e-01 5.30936718e-01 -1.84020072e-01 7.45860100e-01
-1.55957624e-01 -1.31256342e-01 -9.55664039e-01 1.99958086e-01
3.66385803e-02 1.22135079e+00 -5.33660710e-01 7.66449720e-02
-4.61857229e-01 1.64769459e+00 3.43187340e-02 1.47054577e-02
-1.23989117e+00 -5.18834770e-01 1.18912435e+00 5.45691013e-01
1.27702788e-01 -2.26475880e-01 -6.22320771e-02 -1.09314299e+00
2.30497047e-01 -9.78637576e-01 -2.22549528e-01 -1.25349009e+00
-7.94362187e-01 9.62767005e-01 -5.38942099e-01 -1.25069547e+00
-5.35423100e-01 -5.55708349e-01 -3.56990188e-01 9.94489253e-01
-6.26843929e-01 -1.55272198e+00 -8.72583151e-01 1.06071270e+00
2.20563821e-03 -1.55917108e-01 1.39020920e+00 4.34961826e-01
-5.06293833e-01 1.09447145e+00 -4.34019297e-01 1.73506707e-01
4.83883351e-01 -7.76415646e-01 7.08620250e-01 4.59619790e-01
-1.51782408e-01 6.67318404e-01 7.08725870e-01 -5.07422686e-01
-1.88279974e+00 -7.65466690e-01 2.38816977e-01 -8.22272897e-01
2.09163241e-02 -1.06883013e+00 -5.65939724e-01 1.16595221e+00
1.06853828e-01 4.76961493e-01 8.61275792e-01 1.48753775e-02
-2.98150420e-01 -1.78697497e-01 -1.89238763e+00 7.71063745e-01
1.13531315e+00 -8.53993773e-01 -5.10069653e-02 1.19245477e-01
5.45562088e-01 -1.12265360e+00 -8.66707623e-01 -1.19567245e-01
1.18024230e+00 -1.12020910e+00 8.53402317e-01 -3.25223297e-01
2.13495985e-01 -3.49230260e-01 8.34111422e-02 -9.38264608e-01
9.98470932e-02 -1.34969747e+00 4.84075956e-02 1.26173615e+00
1.38353020e-01 -7.18131602e-01 1.18207896e+00 1.48305905e+00
1.07535750e-01 -3.71997148e-01 -9.48796868e-01 -4.34251428e-01
-4.74020064e-01 -4.86124724e-01 7.98102140e-01 6.81297898e-01
1.74041241e-02 3.28323185e-01 -6.83504462e-01 4.17874068e-01
2.18523100e-01 -5.48158586e-01 1.47717381e+00 -1.05492163e+00
-4.52654004e-01 2.12794021e-01 -1.00500953e+00 -5.44249117e-01
2.36577258e-01 -3.11866909e-01 -1.06002457e-01 -9.05641377e-01
-3.20716053e-01 -1.58159599e-01 6.28618360e-01 4.70617622e-01
3.61589074e-01 6.62686646e-01 1.22256681e-01 -2.05891714e-01
4.67979275e-02 2.88382024e-01 1.05830097e+00 3.69849652e-01
-4.42403644e-01 6.59014285e-02 -6.20701790e-01 7.26275742e-01
4.61491704e-01 -2.36341432e-01 -4.24421012e-01 -9.20853093e-02
9.59397666e-03 4.47811633e-01 8.06804717e-01 -1.18305516e+00
-4.24502641e-02 2.26696491e-01 7.61542797e-01 1.44901574e-01
9.76211369e-01 -1.06606805e+00 1.04802310e+00 3.00354719e-01
2.62328893e-01 8.82389173e-02 4.40024465e-01 -2.57448275e-02
3.29446405e-01 4.36265141e-01 8.11014116e-01 -3.00703287e-01
-3.32638264e-01 2.18551785e-01 -4.52098578e-01 -3.33549291e-01
1.35332227e+00 -6.41085207e-01 -8.50404496e-04 -9.48064804e-01
-7.47325897e-01 -5.03765643e-01 9.22101080e-01 5.29125571e-01
9.24225330e-01 -1.22613716e+00 -5.98185718e-01 6.93648994e-01
-2.08890900e-01 -2.24834010e-01 4.98019099e-01 2.03580528e-01
-1.22465825e+00 3.17220087e-03 -5.87401986e-01 -5.54356337e-01
-1.88595760e+00 3.79938215e-01 6.85217619e-01 6.01640880e-01
-9.99285638e-01 1.01343536e+00 -1.68147847e-01 -6.54514790e-01
2.37666741e-02 1.35606200e-01 1.13548160e-01 -5.17673373e-01
7.84920633e-01 1.62298232e-01 1.87059611e-01 -1.00271511e+00
-4.23313618e-01 5.94493270e-01 3.55494559e-01 -4.99264538e-01
1.21229053e+00 -1.47801265e-02 3.41764688e-01 2.14521959e-01
1.27526546e+00 6.65105656e-02 -1.38865685e+00 5.59335113e-01
-1.07656467e+00 -9.60950792e-01 1.23307265e-01 -4.19980317e-01
-1.34342134e+00 4.97754365e-01 8.76206100e-01 -3.15911859e-01
1.10222566e+00 -1.78676084e-01 8.09793234e-01 -6.84520379e-02
1.00521696e+00 -8.39397371e-01 -5.24951704e-02 -9.97746885e-02
8.90490830e-01 -8.58791351e-01 -2.11591437e-01 -6.84680521e-01
-9.28959370e-01 8.53876829e-01 8.26413393e-01 -1.73682287e-01
5.97961903e-01 8.24348748e-01 4.97451812e-01 -1.78835496e-01
-5.07700682e-01 2.72977024e-01 -1.77516714e-01 1.04776275e+00
3.73904496e-01 2.62505800e-01 2.30082661e-01 5.46152711e-01
-7.80687511e-01 3.51464421e-01 7.66155064e-01 1.18047631e+00
3.99078131e-01 -8.01160514e-01 -5.51030815e-01 3.68219502e-02
-6.86739087e-01 4.04738724e-01 -4.66939390e-01 9.37213480e-01
1.63428605e-01 1.07125497e+00 2.61936009e-01 -9.36694801e-01
6.02897465e-01 -7.28552714e-02 8.31094205e-01 -4.88241225e-01
-1.03822339e+00 -8.71813744e-02 9.84658152e-02 -9.81742799e-01
-1.28669351e-01 -5.65608680e-01 -1.00451159e+00 -9.23083305e-01
4.29736339e-02 -5.39044321e-01 1.07478368e+00 3.50233883e-01
8.19491148e-01 3.94256860e-01 8.02703261e-01 -1.33543789e+00
5.43753617e-02 -6.63741529e-01 -7.44628787e-01 2.74966359e-01
2.27990031e-01 -5.92283070e-01 -2.20666528e-01 4.21102047e-02] | [12.928070068359375, -0.39840927720069885] |
1a408946-7752-4f22-bc19-04b0edb22bb2 | mrclens-an-mrc-dataset-bias-detection-toolkit-1 | 2207.08943 | null | https://arxiv.org/abs/2207.08943v1 | https://arxiv.org/pdf/2207.08943v1.pdf | MRCLens: an MRC Dataset Bias Detection Toolkit | Many recent neural models have shown remarkable empirical results in Machine Reading Comprehension, but evidence suggests sometimes the models take advantage of dataset biases to predict and fail to generalize on out-of-sample data. While many other approaches have been proposed to address this issue from the computation perspective such as new architectures or training procedures, we believe a method that allows researchers to discover biases, and adjust the data or the models in an earlier stage will be beneficial. Thus, we introduce MRCLens, a toolkit that detects whether biases exist before users train the full model. For the convenience of introducing the toolkit, we also provide a categorization of common biases in MRC. | ['Eric P. Xing', 'Haohan Wang', 'Yifan Zhong'] | 2022-07-18 | null | null | null | null | ['machine-reading-comprehension'] | ['natural-language-processing'] | [ 2.46022463e-01 2.52987504e-01 -2.39765927e-01 -9.77369249e-01
-1.92202330e-01 -5.00620067e-01 5.21742225e-01 3.78876358e-01
-5.94108284e-01 5.35694122e-01 4.49948072e-01 -6.86544240e-01
7.46514723e-02 -6.60895407e-01 -6.73972845e-01 -2.62445450e-01
3.69653791e-01 3.42932850e-01 4.51769494e-02 -1.72104299e-01
1.02264798e+00 2.56407529e-01 -1.72832727e+00 2.43632004e-01
1.16795707e+00 4.13304031e-01 2.42107481e-01 5.99978626e-01
-6.96054474e-02 8.68676782e-01 -5.95748901e-01 -4.48806226e-01
1.17361695e-02 -3.21620822e-01 -9.35213625e-01 -5.72684705e-01
6.33975565e-01 -5.25306702e-01 -2.00697511e-01 1.06909561e+00
5.44885278e-01 2.80342817e-01 4.82328892e-01 -8.82959783e-01
-1.12524211e+00 9.37911570e-01 -2.23673791e-01 6.54390156e-01
4.87848550e-01 4.07143056e-01 7.80249417e-01 -8.35878551e-01
1.71277180e-01 1.33177602e+00 6.10928476e-01 7.68745840e-01
-1.21816993e+00 -7.01403439e-01 6.19444966e-01 2.48468444e-01
-9.65461016e-01 -6.49170995e-01 6.69721723e-01 -3.88753831e-01
9.82167959e-01 4.31244701e-01 5.51934659e-01 1.48320436e+00
1.58429384e-01 9.27229702e-01 1.71623337e+00 -6.51583552e-01
1.90397054e-01 2.44534329e-01 9.43413258e-01 4.82245535e-01
5.62343478e-01 1.95404395e-01 -8.76632392e-01 -4.45387438e-02
6.07298732e-01 -1.95604131e-01 -6.06361806e-01 -1.38770506e-01
-1.21363556e+00 7.66661108e-01 5.26182592e-01 2.85183907e-01
-3.02693605e-01 -1.88463584e-01 1.24582015e-01 6.09622240e-01
4.19894308e-01 1.11371887e+00 -9.04494524e-01 -1.89540058e-01
-8.53165865e-01 3.45360518e-01 8.50477159e-01 7.98467755e-01
6.09656036e-01 -6.58451542e-02 -2.54970938e-01 8.10077310e-01
1.19791411e-01 1.15376540e-01 8.77176166e-01 -7.17652977e-01
3.68304640e-01 6.37682498e-01 -8.86221230e-02 -1.00360513e+00
-6.84302986e-01 -5.39880455e-01 -6.04558170e-01 -7.22030327e-02
5.37099600e-01 4.89539281e-02 -9.73900497e-01 1.86815989e+00
-1.73746683e-02 -5.00331759e-01 -9.11435112e-02 9.38281476e-01
7.05906272e-01 2.35004157e-01 2.90190428e-01 1.40977740e-01
1.12567794e+00 -7.39651263e-01 -6.57821178e-01 -7.66334236e-01
6.60211265e-01 -3.64939302e-01 1.56589627e+00 5.13141572e-01
-1.29352522e+00 -6.42593443e-01 -1.20976591e+00 -4.05465454e-01
-5.66551149e-01 -9.26356986e-02 6.86733782e-01 8.31027985e-01
-9.62818384e-01 6.49335682e-01 -5.67617059e-01 -3.86471093e-01
5.21254361e-01 2.37259567e-01 2.08505422e-01 -2.60774583e-01
-1.29828537e+00 1.47340643e+00 4.35563713e-01 2.28459984e-01
-6.84018552e-01 -6.25769317e-01 -6.88584685e-01 1.42261952e-01
2.72511363e-01 -6.61492050e-01 1.44589412e+00 -1.17821932e+00
-1.32046676e+00 8.82017195e-01 -5.19095302e-01 -1.06128640e-01
2.39246860e-01 -4.19833362e-01 -1.53601021e-01 -2.97464162e-01
-1.99205011e-01 6.06512606e-01 7.35467553e-01 -9.18311298e-01
-4.75587636e-01 -4.76453722e-01 2.31521919e-01 2.72926927e-01
-2.61502922e-01 1.41367674e-01 1.40492499e-01 -5.85882604e-01
3.26709688e-01 -6.59939945e-01 -7.47273937e-02 -3.96048039e-01
-2.75032878e-01 -5.32121241e-01 -1.27477974e-01 -7.94111013e-01
1.27689493e+00 -1.73301697e+00 1.94412202e-01 2.25293979e-01
3.43347847e-01 1.10010378e-01 -8.51567462e-02 5.95781431e-02
-9.01187286e-02 5.04257798e-01 -3.62394042e-02 1.43482052e-02
-1.25693576e-02 -7.46062025e-02 -1.35092810e-01 1.71834633e-01
2.10531488e-01 9.32563007e-01 -8.43366981e-01 -4.04783301e-02
-2.45722145e-01 3.13784182e-02 -5.86399853e-01 3.14199597e-01
-6.84550926e-02 9.27043036e-02 -2.06885591e-01 6.17364645e-01
5.99649668e-01 -4.31206882e-01 1.35867000e-01 2.42884040e-01
-7.43493661e-02 1.06697333e+00 -8.81443024e-01 1.38420057e+00
-1.27696082e-01 7.63202488e-01 -2.21223965e-01 -8.51467252e-01
6.82710290e-01 1.57542676e-01 -4.99655575e-01 -8.94704461e-01
1.87358871e-01 1.43469706e-01 6.28424525e-01 -7.38118589e-01
5.91212392e-01 5.92987686e-02 2.95378000e-01 7.20669925e-01
-2.44606044e-02 1.00804441e-01 -5.02034612e-02 1.39685765e-01
9.40082669e-01 -5.26997820e-02 4.10258055e-01 -4.67002690e-01
2.00247407e-01 8.12247023e-02 2.92612493e-01 1.23164868e+00
-1.03851184e-01 4.70182806e-01 2.83342332e-01 -4.80612695e-01
-8.78619611e-01 -7.99678862e-01 -1.86156616e-01 1.69343638e+00
-3.25762391e-01 -1.31784916e-01 -7.62615979e-01 -6.58372700e-01
-1.13195121e-01 1.21996439e+00 -8.43443334e-01 -3.84816498e-01
-4.04434800e-01 -8.98071766e-01 4.11021233e-01 7.04340816e-01
5.04174232e-01 -1.02232718e+00 -8.33986104e-01 1.29304931e-01
-1.14657901e-01 -5.40699482e-01 -5.04737534e-02 3.93144071e-01
-1.27818620e+00 -9.19509292e-01 -3.95884097e-01 -6.98229849e-01
1.00209570e+00 2.38339752e-01 1.51511085e+00 7.92779267e-01
1.81299448e-01 1.42476231e-01 -4.36466783e-01 -8.24074686e-01
-3.03264797e-01 4.73755658e-01 -8.41849484e-03 -6.22681737e-01
1.00787425e+00 -2.51331121e-01 -5.49319208e-01 2.75314748e-01
-6.96126819e-01 2.88921624e-01 6.54752314e-01 7.19601512e-01
-1.42120734e-01 -3.91838372e-01 8.70275438e-01 -1.10422051e+00
1.21820235e+00 -6.31814063e-01 -2.56260484e-01 3.93665791e-01
-1.14799929e+00 1.25250101e-01 2.73237616e-01 -6.48055077e-01
-9.47747350e-01 -5.43776512e-01 -1.26175255e-01 2.51222312e-01
-2.99342781e-01 7.95528054e-01 -8.90010297e-02 -2.77023744e-02
9.87345219e-01 1.91402227e-01 -1.40450433e-01 -5.37692130e-01
1.00407749e-01 6.69837356e-01 1.54438972e-01 -6.31005883e-01
5.07187963e-01 -2.91181087e-01 -7.75605083e-01 -3.43136907e-01
-1.29345393e+00 -6.29158691e-02 -7.44091153e-01 -9.16814283e-02
5.23597658e-01 -9.11952972e-01 -2.85555720e-01 5.72013080e-01
-1.07891750e+00 -5.15869558e-01 8.04937482e-02 4.26105440e-01
-1.14469774e-01 -7.51557276e-02 -4.45192158e-01 -7.36403465e-01
-1.06664762e-01 -1.05398512e+00 2.52369195e-01 6.70182765e-01
-8.44565630e-01 -9.91631448e-01 -1.16689317e-01 4.11706507e-01
9.03168857e-01 -4.54587936e-01 1.31739402e+00 -1.08714414e+00
-4.49718863e-01 -1.07908934e-01 -6.82012662e-02 3.35897177e-01
-1.47908568e-01 -1.91161260e-01 -1.25284994e+00 -2.24358276e-01
1.85176775e-01 -5.43393135e-01 7.91468859e-01 2.85356104e-01
1.65306687e+00 -3.36059868e-01 -1.31008178e-01 3.98247451e-01
1.05404270e+00 4.11020331e-02 6.38802648e-01 5.24754524e-01
5.52131593e-01 8.54592085e-01 2.94522882e-01 -1.02701113e-01
6.84032619e-01 4.19107750e-02 9.15951356e-02 1.23730361e-01
1.94344908e-01 -3.73270988e-01 3.12552601e-01 1.03960943e+00
-6.75384477e-02 -1.65765956e-01 -1.04644012e+00 5.80311656e-01
-1.46714306e+00 -6.77923858e-01 -1.31850332e-01 2.18644118e+00
1.13002241e+00 3.79097670e-01 -2.05841154e-01 -5.47459126e-02
4.24639940e-01 8.41488466e-02 -8.06649268e-01 -8.53493035e-01
-1.58109665e-01 3.77311438e-01 2.57472724e-01 4.60846215e-01
-4.77073729e-01 9.27196860e-01 8.34296608e+00 3.61267626e-02
-9.38600719e-01 -2.12384742e-02 8.11700344e-01 -8.97402167e-02
-5.32426536e-01 9.86835081e-03 -6.32599115e-01 2.76422650e-01
1.08722413e+00 2.07111817e-02 5.47821045e-01 8.07513535e-01
6.49213120e-02 -5.16541839e-01 -1.54135776e+00 6.11712873e-01
2.71167159e-01 -9.13181007e-01 6.81106821e-02 -2.95605749e-01
6.32327497e-01 6.11566529e-02 8.73209164e-03 5.89950204e-01
2.09608838e-01 -1.27937746e+00 6.11332655e-01 7.77114451e-01
1.98002636e-01 -4.26732868e-01 5.94998837e-01 7.08999634e-01
-4.00631204e-02 -9.12466273e-02 -6.66435897e-01 -8.82585347e-01
-4.05246288e-01 4.33580190e-01 -8.01939964e-01 -1.80348605e-01
7.39890814e-01 3.25353682e-01 -1.26087332e+00 1.09900296e+00
-5.44729590e-01 1.09319460e+00 3.58104333e-02 -5.08965671e-01
-2.92679310e-01 8.27714577e-02 1.91126928e-01 1.00566864e+00
5.08476645e-02 2.22039148e-01 -4.20475811e-01 1.00693834e+00
-2.18983009e-01 1.14085391e-01 -4.77842510e-01 3.17191370e-02
6.54773176e-01 8.98963034e-01 -2.79594183e-01 -5.47859550e-01
-4.08048898e-01 8.03062260e-01 6.89915240e-01 4.23607856e-01
-4.09619510e-01 -1.76415160e-01 4.67046916e-01 1.90396875e-01
-1.68406099e-01 -3.16025764e-01 -9.66021597e-01 -1.28434181e+00
4.34186868e-02 -1.56421936e+00 1.25240818e-01 -1.11826932e+00
-1.23434865e+00 6.29197359e-02 6.42887950e-02 -2.18676969e-01
-5.50758690e-02 -7.72759378e-01 -7.05410182e-01 1.09389091e+00
-1.41348624e+00 -4.85125661e-01 -4.88668323e-01 1.81434169e-01
6.85374677e-01 -1.11123063e-01 7.71177709e-01 -2.28903934e-01
-8.08684051e-01 6.42969608e-01 -1.24357633e-01 5.74656427e-02
8.72364759e-01 -1.31172550e+00 5.79257548e-01 8.48698974e-01
-6.91278726e-02 1.34606612e+00 8.50328386e-01 -6.59412086e-01
-1.18358409e+00 -3.29553068e-01 1.21309125e+00 -8.45286131e-01
3.27776015e-01 -2.78334171e-01 -1.48171246e+00 9.73782122e-01
5.84670782e-01 -7.41589129e-01 8.10789824e-01 7.56281734e-01
-3.78980845e-01 1.94979951e-01 -9.47500706e-01 7.63684571e-01
1.10688674e+00 -4.40424114e-01 -1.24771810e+00 1.31524548e-01
3.41701180e-01 -3.93932343e-01 -6.52263820e-01 3.07501465e-01
7.03423083e-01 -1.10217166e+00 5.44733882e-01 -9.11452293e-01
5.40098608e-01 6.22152016e-02 1.51445210e-01 -1.82950556e+00
-5.75362265e-01 -2.09498137e-01 -3.27600837e-02 1.05118835e+00
8.62671196e-01 -9.14814949e-01 5.25361955e-01 1.21986878e+00
2.68637668e-04 -6.59023523e-01 -5.01478493e-01 -1.58543333e-01
4.42416877e-01 -4.31535363e-01 8.43959093e-01 1.26651895e+00
8.80448297e-02 5.08205354e-01 -1.16528563e-01 1.42224148e-01
3.18497449e-01 -2.18721673e-01 6.38307393e-01 -1.42716479e+00
-2.41641402e-02 -5.72326183e-01 2.41508693e-01 -1.34860802e+00
-3.38650793e-02 -6.27698600e-01 1.68183252e-01 -1.44733751e+00
3.49196643e-01 -4.86827523e-01 -4.99383241e-01 4.13265198e-01
-8.25182140e-01 -2.92298913e-01 2.70159915e-03 3.17520082e-01
-3.83926868e-01 1.88237429e-01 1.24081790e+00 7.55535141e-02
-9.20967478e-03 -3.19199695e-04 -1.39636230e+00 7.92307258e-01
1.21627295e+00 -5.20343244e-01 -3.98448586e-01 -8.87839258e-01
7.03279912e-01 -6.48371160e-01 4.25825149e-01 -9.15654302e-01
2.30588377e-01 -1.01980679e-01 8.65982890e-01 -4.72345293e-01
-1.14384413e-01 -3.04674566e-01 -3.76599282e-01 2.94468552e-01
-9.33409929e-01 4.59766537e-01 2.12748542e-01 2.14251816e-01
1.37946114e-01 -7.35985816e-01 5.47292233e-01 -4.00512308e-01
-5.29645503e-01 -3.12591940e-01 -6.44717515e-01 1.71558350e-01
4.01859492e-01 -2.13751793e-01 -5.85268319e-01 -4.01992738e-01
-4.23337281e-01 3.64157647e-01 5.99422753e-01 6.80459380e-01
5.27628660e-01 -8.83605003e-01 -6.55606449e-01 3.15492630e-01
6.21437915e-02 -2.64536180e-02 -9.12547931e-02 7.57545531e-01
-2.45277107e-01 4.53864187e-01 -2.87366301e-01 -2.39527732e-01
-9.03008580e-01 4.27367479e-01 3.94548923e-01 -2.87230145e-02
-2.49650389e-01 1.03543973e+00 4.39940281e-02 -6.82456493e-01
3.76884133e-01 -6.21698022e-01 -3.91216576e-01 2.41406471e-01
6.86264992e-01 3.11740905e-01 1.92777932e-01 -5.61117977e-02
-1.18005581e-01 2.55215466e-02 -4.47436482e-01 7.24588782e-02
1.16417181e+00 -2.57218003e-01 -1.09921753e-01 9.04191256e-01
5.48961580e-01 -9.01906192e-02 -9.59191382e-01 -2.51878411e-01
1.91191614e-01 -5.52475333e-01 1.23153411e-01 -1.41721094e+00
-5.25107205e-01 1.03195024e+00 4.67664301e-01 1.52256086e-01
9.53297496e-01 -2.99539030e-01 1.80305690e-01 5.92303097e-01
1.48496613e-01 -1.34685004e+00 -7.88717568e-02 7.94920146e-01
8.89548600e-01 -1.47527313e+00 3.06435466e-01 8.61801580e-02
-6.48665547e-01 1.03116941e+00 1.07232010e+00 -8.55915844e-02
3.33801687e-01 -4.22865376e-02 1.81615993e-01 -1.95307150e-01
-9.85794425e-01 4.25337292e-02 3.10682595e-01 6.09164476e-01
9.50397611e-01 -5.18739484e-02 -4.39294904e-01 6.05728388e-01
-4.79443401e-01 9.37120151e-03 6.70187950e-01 9.90496457e-01
-6.70703888e-01 -1.02302241e+00 -5.17938852e-01 9.25766826e-01
-4.12282050e-01 -4.24196422e-01 -6.27746105e-01 8.18541229e-01
-9.23994109e-02 9.34015930e-01 2.42151525e-02 -2.08034053e-01
3.77702773e-01 6.88046277e-01 6.15939319e-01 -6.31941557e-01
-1.00834107e+00 -5.37456810e-01 9.62664783e-02 -5.08929610e-01
-3.35671335e-01 -7.82117426e-01 -8.41184378e-01 -3.22067142e-01
-4.87604111e-01 -8.60181823e-02 6.08788311e-01 1.09045410e+00
2.97478855e-01 3.85474563e-01 1.79988652e-01 -4.75261122e-01
-8.55533421e-01 -1.54452884e+00 -1.66875646e-01 2.68602043e-01
2.59350836e-01 -5.22955000e-01 -5.05129993e-01 -3.38945359e-01] | [10.310673713684082, 8.051393508911133] |
5961c0ed-d10b-4c60-93ee-617d2fb3dcf8 | quantized-gan-for-complex-music-generation | 2204.00604 | null | https://arxiv.org/abs/2204.00604v2 | https://arxiv.org/pdf/2204.00604v2.pdf | Quantized GAN for Complex Music Generation from Dance Videos | We present Dance2Music-GAN (D2M-GAN), a novel adversarial multi-modal framework that generates complex musical samples conditioned on dance videos. Our proposed framework takes dance video frames and human body motions as input, and learns to generate music samples that plausibly accompany the corresponding input. Unlike most existing conditional music generation works that generate specific types of mono-instrumental sounds using symbolic audio representations (e.g., MIDI), and that usually rely on pre-defined musical synthesizers, in this work we generate dance music in complex styles (e.g., pop, breaking, etc.) by employing a Vector Quantized (VQ) audio representation, and leverage both its generality and high abstraction capacity of its symbolic and continuous counterparts. By performing an extensive set of experiments on multiple datasets, and following a comprehensive evaluation protocol, we assess the generative qualities of our proposal against alternatives. The attained quantitative results, which measure the music consistency, beats correspondence, and music diversity, demonstrate the effectiveness of our proposed method. Last but not least, we curate a challenging dance-music dataset of in-the-wild TikTok videos, which we use to further demonstrate the efficacy of our approach in real-world applications -- and which we hope to serve as a starting point for relevant future research. | ['Sergey Tulyakov', 'Yan Yan', 'Menglei Chai', 'Panos Achlioptas', 'Yu Wu', 'Kyle Olszewski', 'Ye Zhu'] | 2022-04-01 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 4.84088510e-01 -9.88402143e-02 2.91811139e-03 1.58364803e-01
-1.04012072e+00 -9.48458970e-01 8.27610731e-01 -6.18429601e-01
6.40370101e-02 7.91768074e-01 2.85765469e-01 2.30689973e-01
-1.75963804e-01 -7.11557865e-01 -8.90964091e-01 -8.82560313e-01
-3.14798616e-02 2.68753290e-01 -2.02700078e-01 -4.52393860e-01
1.06841378e-01 1.12358555e-01 -1.65068460e+00 3.15709978e-01
7.48839796e-01 8.76812041e-01 -1.45437434e-01 8.86659503e-01
2.64266402e-01 9.65628445e-01 -9.60023105e-01 -6.86154187e-01
3.45248163e-01 -1.05457556e+00 -4.37751621e-01 -4.83953319e-02
4.24218923e-01 -2.62977511e-01 -1.98917404e-01 8.95196736e-01
5.87270021e-01 1.84909642e-01 7.93343723e-01 -1.25440061e+00
-7.71933496e-01 8.55602682e-01 -3.19742769e-01 -1.88281149e-01
6.31475925e-01 5.36533117e-01 1.15003610e+00 -5.86819470e-01
6.41935706e-01 1.04261172e+00 8.02521408e-01 8.02932858e-01
-1.31999850e+00 -9.32183325e-01 -3.18726182e-01 -1.11267090e-01
-1.42464316e+00 -4.65464026e-01 1.05200756e+00 -3.04832995e-01
4.13919576e-02 5.18372715e-01 8.84851575e-01 1.78829765e+00
-1.41476998e-02 7.80864954e-01 1.06811118e+00 -3.53060931e-01
1.89419329e-01 -3.96669567e-01 -6.19125426e-01 4.08859313e-01
-1.82367891e-01 2.98981488e-01 -8.45224738e-01 -1.69335693e-01
1.06192684e+00 -1.66959152e-01 -4.78299350e-01 -5.37668429e-02
-1.59834778e+00 7.66921401e-01 2.05470145e-01 3.39809418e-01
-3.45461488e-01 7.11523235e-01 3.78115237e-01 3.16142976e-01
-1.11081973e-01 3.12236726e-01 2.82477111e-01 -5.24487793e-01
-1.45732737e+00 4.90851432e-01 6.31878853e-01 9.79757726e-01
2.86755621e-01 5.09262800e-01 -5.67219436e-01 5.98265052e-01
6.64604902e-02 6.24773324e-01 6.76749170e-01 -1.07772708e+00
3.73899996e-01 4.36299406e-02 1.49980292e-01 -9.88878310e-01
1.09960772e-01 -4.52481031e-01 -1.04349232e+00 2.53051877e-01
3.11771512e-01 -1.86037421e-01 -5.16231477e-01 1.98608959e+00
9.72573832e-02 7.65828609e-01 2.21967012e-01 9.90424693e-01
7.64774859e-01 4.88108605e-01 -2.27243990e-01 -1.25565946e-01
1.04221308e+00 -8.47994983e-01 -6.39317811e-01 5.22705436e-01
-1.38506487e-01 -8.83684337e-01 1.49828804e+00 6.99803174e-01
-1.23860121e+00 -9.04061198e-01 -1.09611464e+00 1.48749724e-01
3.09277117e-01 3.47239405e-01 5.24119556e-01 7.34965742e-01
-8.66631627e-01 8.10846329e-01 -7.83337593e-01 4.27634679e-02
2.94517368e-01 8.21763054e-02 -1.68656603e-01 2.93563783e-01
-1.05616224e+00 6.12723157e-02 3.52164805e-01 1.56188250e-01
-1.35888624e+00 -6.79835320e-01 -5.38683772e-01 -8.49674568e-02
2.48508409e-01 -9.40697670e-01 1.24539173e+00 -1.17333710e+00
-2.02752066e+00 4.84388620e-01 2.01743394e-01 -4.00082976e-01
8.17886949e-01 -4.77965295e-01 -4.20524955e-01 3.53592366e-01
-1.68805141e-02 6.63344800e-01 1.15421641e+00 -1.44371295e+00
-2.54046887e-01 1.63677230e-01 2.96839148e-01 8.29616264e-02
-3.31708103e-01 -2.05588311e-01 -4.31291968e-01 -1.33471727e+00
-2.82976836e-01 -1.13447356e+00 1.56845562e-02 -2.27530524e-01
-7.54017889e-01 1.91885367e-01 4.19740677e-01 -3.96280915e-01
1.24586463e+00 -2.41240835e+00 6.31120861e-01 2.43452176e-01
-3.72374691e-02 -2.91416962e-02 -1.96668878e-01 6.52252078e-01
1.27721757e-01 8.59727040e-02 -3.13014567e-01 -3.83426905e-01
3.38106722e-01 1.53474972e-01 -6.06969416e-01 1.91182047e-01
3.08172256e-01 1.00433719e+00 -8.25179935e-01 -4.24548686e-01
3.33912969e-02 7.05443203e-01 -7.67436564e-01 1.00224875e-01
-3.47645879e-01 9.01908040e-01 -5.21081924e-01 7.00909913e-01
5.18239066e-02 2.32615974e-02 6.23169728e-02 -2.17873380e-01
2.14367554e-01 -1.02539904e-01 -1.42322004e+00 2.21305585e+00
-4.45606202e-01 3.84917229e-01 -1.84831589e-01 -6.00338399e-01
9.97576416e-01 5.69981754e-01 2.99964249e-01 -3.28294605e-01
1.49157584e-01 2.46369302e-01 -9.67528597e-02 -2.94493645e-01
5.89827776e-01 -3.70196998e-01 -3.33135962e-01 3.35577309e-01
2.57197261e-01 -3.02333057e-01 1.64810911e-01 7.24972710e-02
9.41150844e-01 6.23400390e-01 2.12434977e-01 -8.60744901e-03
4.88967001e-01 -1.17397293e-01 2.87988245e-01 6.96919799e-01
1.83375031e-01 9.20747340e-01 5.07838845e-01 -5.90687841e-02
-9.69985545e-01 -1.28492320e+00 3.09660286e-01 8.65956128e-01
1.49261961e-02 -4.76657778e-01 -8.58886182e-01 -3.18267196e-01
-3.12090993e-01 6.58879459e-01 -6.47242785e-01 -9.85228643e-02
-7.38259196e-01 -4.56017554e-01 1.34322655e+00 5.17966866e-01
3.36650878e-01 -1.38906467e+00 -5.94302237e-01 2.80896783e-01
-1.76888406e-01 -9.02270317e-01 -6.17977142e-01 -2.40367115e-01
-6.70356989e-01 -1.09827578e+00 -1.03754783e+00 -7.51686215e-01
5.77907730e-03 -2.59811014e-01 1.23852432e+00 -5.06598912e-02
-2.35666528e-01 5.70730507e-01 -5.63779831e-01 -2.27797538e-01
-6.88015759e-01 -2.87609696e-02 1.44227430e-01 3.55575651e-01
-3.80499303e-01 -1.11198306e+00 -6.73190594e-01 2.61651367e-01
-1.20245004e+00 9.35637876e-02 4.51820314e-01 9.92731869e-01
8.78806829e-01 -7.31495395e-02 7.24081993e-01 -7.31115639e-01
7.44833291e-01 -2.84097731e-01 -2.16461405e-01 1.02776349e-01
1.12251043e-01 -1.69274315e-01 9.03851688e-01 -7.40162313e-01
-7.05061018e-01 -2.16631331e-02 -2.61325687e-01 -8.83197486e-01
-9.06819850e-02 3.01208496e-01 -1.02499679e-01 2.22929537e-01
8.07001770e-01 5.02498031e-01 -2.48484597e-01 -3.09086859e-01
7.30342269e-01 3.49074364e-01 1.24079287e+00 -1.01234567e+00
1.10704815e+00 4.41912830e-01 1.80573806e-01 -6.29021227e-01
-3.52672845e-01 2.10925668e-01 -4.66026574e-01 -3.67918760e-01
9.69901323e-01 -9.89742577e-01 -9.64643776e-01 2.73019373e-01
-8.83622050e-01 -5.20201921e-01 -4.94611353e-01 4.35247034e-01
-1.08925700e+00 2.98899114e-01 -7.37948358e-01 -8.10110867e-01
-4.44954336e-01 -1.04499102e+00 1.25900364e+00 9.14058909e-02
-3.99834156e-01 -7.72856891e-01 3.20391715e-01 2.58987159e-01
1.69076160e-01 1.06259525e+00 6.06548548e-01 -1.50531977e-01
-5.91305673e-01 6.65246993e-02 5.09397626e-01 2.63168722e-01
1.52752876e-01 2.65640944e-01 -9.53314304e-01 -2.97658950e-01
-2.59004325e-01 -6.77481890e-01 6.75344408e-01 3.13913114e-02
1.00858521e+00 -3.09099048e-01 2.54222125e-01 7.62947083e-01
1.36099100e+00 5.53195626e-02 6.92978263e-01 6.15148433e-02
6.08167470e-01 1.24912702e-01 5.37665069e-01 6.49310112e-01
-6.02532849e-02 9.71225500e-01 5.40575981e-01 2.03569874e-01
-3.80986333e-01 -6.38880074e-01 6.52688265e-01 1.11723685e+00
-7.54837096e-01 -2.78900921e-01 -4.11444187e-01 3.37081641e-01
-1.50407016e+00 -1.35075486e+00 1.47158459e-01 1.87460363e+00
9.76454854e-01 -1.30774081e-03 5.59403181e-01 5.94597816e-01
5.31418562e-01 8.95087570e-02 -4.64082837e-01 3.46963406e-02
-2.91721761e-01 8.74271572e-01 -2.19421640e-01 9.63833928e-02
-1.01442456e+00 7.85837948e-01 5.52707005e+00 1.18623579e+00
-1.18771553e+00 2.92231087e-02 1.87185988e-01 -4.58044171e-01
-5.48900962e-01 -3.78929675e-01 -2.45836064e-01 5.81147850e-01
7.14597762e-01 1.91101246e-02 9.30154502e-01 5.98207891e-01
1.19265340e-01 6.49994671e-01 -1.14520597e+00 1.15750468e+00
6.65901154e-02 -1.50529516e+00 3.55287373e-01 -1.82776883e-01
7.99613953e-01 -5.40518999e-01 3.55747312e-01 4.00003493e-01
2.54319251e-01 -1.23618627e+00 1.38319337e+00 7.38670409e-01
1.33127785e+00 -8.99024367e-01 3.67535532e-01 2.08269358e-01
-1.29725242e+00 1.16059124e-01 5.52765056e-02 1.17579401e-01
2.15916991e-01 1.02602966e-01 -3.33579242e-01 9.97895300e-01
4.96993899e-01 6.83296263e-01 -3.65898162e-01 7.10183203e-01
-4.26975250e-01 1.00710797e+00 -1.34870917e-01 6.85409978e-02
1.54141694e-01 -2.80797482e-01 5.85558474e-01 1.06263828e+00
6.59512937e-01 8.27638581e-02 -2.01221313e-02 1.06161332e+00
-2.02379987e-01 9.51508954e-02 -5.01435161e-01 -2.62332767e-01
4.38141435e-01 1.10683036e+00 -5.79560220e-01 -2.33840123e-01
-6.78564832e-02 1.21460748e+00 -1.13218360e-01 2.97946572e-01
-1.25133324e+00 -3.16896826e-01 5.56691706e-01 -2.25175932e-01
5.03680825e-01 -2.26827309e-01 -5.60450703e-02 -1.27325964e+00
7.81937167e-02 -1.36212194e+00 1.03646368e-01 -7.61250854e-01
-1.22527623e+00 8.33031356e-01 -2.26194263e-01 -1.79882097e+00
-7.31531262e-01 -1.83936819e-01 -8.32910776e-01 4.95657563e-01
-1.03200984e+00 -1.45406985e+00 -3.79324079e-01 9.09172356e-01
4.10538316e-01 -4.01565343e-01 1.04046154e+00 3.88297319e-01
-1.91780180e-01 9.43159640e-01 5.67332283e-02 2.45937198e-01
5.33852756e-01 -1.07967496e+00 3.13842893e-01 4.76305872e-01
8.60116482e-01 5.66410184e-01 6.45714641e-01 -3.31614614e-01
-1.63223422e+00 -1.07655072e+00 -4.47868966e-02 -4.24808979e-01
6.03816092e-01 -3.45231056e-01 -4.98897552e-01 5.52777886e-01
2.95898885e-01 -1.55636415e-01 8.75208318e-01 -4.21775103e-01
-2.87379742e-01 8.54688287e-02 -9.69475806e-01 7.01053381e-01
1.26633871e+00 -4.56227601e-01 -6.03612065e-01 -4.39419821e-02
7.02711463e-01 -5.49022734e-01 -1.06413782e+00 3.72824937e-01
9.02555466e-01 -1.20992088e+00 1.02970147e+00 -5.43414891e-01
8.30629408e-01 -5.32734096e-01 -4.70253855e-01 -1.21217644e+00
-1.22573502e-01 -1.09794176e+00 -1.06969677e-01 1.60218906e+00
8.52233693e-02 6.87871948e-02 7.03620493e-01 -3.81111652e-01
-2.76557624e-01 -5.26330709e-01 -9.09440339e-01 -9.08027470e-01
7.22129494e-02 -5.63890517e-01 8.77741218e-01 8.97194505e-01
-4.54105049e-01 2.04727873e-01 -8.74072492e-01 4.16966751e-02
5.81797838e-01 3.76060754e-01 1.28843963e+00 -9.07484412e-01
-1.03359377e+00 -4.89339918e-01 -5.62408686e-01 -6.90055549e-01
1.15020081e-01 -8.53524208e-01 -1.47427887e-01 -8.07419658e-01
-1.19098313e-01 -2.58091182e-01 -3.24721098e-01 2.98929900e-01
-1.17067657e-01 1.06228805e+00 7.38005757e-01 3.44966233e-01
-4.67414647e-01 9.49164331e-01 1.50259984e+00 -2.34834790e-01
-2.74965644e-01 3.67988311e-02 -4.49168712e-01 6.50387466e-01
7.42422521e-01 -4.06346142e-01 -4.92724657e-01 -2.50406027e-01
1.43164203e-01 4.37403947e-01 7.07810819e-01 -1.45583379e+00
-2.21265838e-01 -9.45622325e-02 2.81978637e-01 -1.42036542e-01
6.82291567e-01 -5.49366772e-01 7.13330805e-01 5.01648068e-01
-3.85840386e-01 -6.02125600e-02 3.94603051e-02 7.79016793e-01
-4.98514324e-01 -1.96109861e-02 5.05468071e-01 3.25721875e-02
-3.45966697e-01 6.51705712e-02 5.92642650e-02 2.87119359e-01
8.81068945e-01 -6.31646812e-02 4.82337847e-02 -7.09083498e-01
-7.80761898e-01 -3.04418534e-01 5.06905973e-01 3.58548731e-01
4.90312368e-01 -1.87080443e+00 -9.47317421e-01 2.12340862e-01
7.90404454e-02 -2.64271528e-01 1.90572739e-01 5.70315242e-01
-5.77433765e-01 -7.41679147e-02 -4.24280763e-01 -5.81492186e-01
-9.81780887e-01 4.46609229e-01 3.86692933e-03 -1.70420259e-02
-8.09292495e-01 5.37545145e-01 8.95555913e-02 -1.36033714e-01
2.02105641e-01 -3.83500218e-01 -2.68352814e-02 -1.23933993e-01
3.59830678e-01 2.96282381e-01 -3.59862477e-01 -5.58538556e-01
-1.07901320e-01 6.35675550e-01 8.34559500e-01 -5.58543742e-01
1.23205209e+00 4.37297463e-01 2.58177310e-01 7.11658359e-01
6.59334302e-01 6.07318640e-01 -1.26786029e+00 2.33456179e-01
-3.82837266e-01 -3.23114991e-01 -6.45580471e-01 -5.87726116e-01
-1.11335742e+00 7.77384996e-01 2.93393493e-01 1.52049974e-01
1.26295698e+00 -8.35741460e-02 8.67137492e-01 2.21094847e-01
5.86752892e-01 -4.14577693e-01 4.11943227e-01 -1.06107593e-01
1.16147339e+00 -5.84979534e-01 -3.52154851e-01 -1.51675716e-01
-7.84488559e-01 1.15543628e+00 2.38478601e-01 -5.30605793e-01
1.44981220e-01 1.65976048e-01 6.53156936e-02 1.63443759e-01
-5.51890910e-01 -7.55859632e-03 3.26849282e-01 6.01262629e-01
3.90615582e-01 1.52282521e-01 -2.65108913e-01 1.11836314e+00
-9.01570261e-01 2.83461958e-01 5.01914382e-01 5.01426697e-01
1.57753289e-01 -1.17115664e+00 -6.08777583e-01 1.13871552e-01
-5.00881195e-01 -3.51168923e-02 -3.39290261e-01 1.01080835e+00
3.75385463e-01 9.27660584e-01 -1.85970411e-01 -6.32455945e-01
2.97147721e-01 3.05693224e-02 7.72995770e-01 -2.39181355e-01
-9.02204454e-01 3.05342734e-01 -5.91543838e-02 -4.88557994e-01
-8.42562497e-01 -6.17887259e-01 -1.05613279e+00 -2.32289433e-01
4.02094144e-03 1.16283908e-01 1.95385247e-01 4.69447762e-01
1.40589043e-01 8.21381927e-01 5.21223307e-01 -1.20174754e+00
-6.80017591e-01 -1.00616860e+00 -6.54730201e-01 8.42044830e-01
2.73504883e-01 -6.24614418e-01 -1.82728246e-01 4.33456779e-01] | [15.782713890075684, 5.503054618835449] |
a7c74868-30c4-4492-a757-65835da44c63 | part-level-action-parsing-via-a-pose-guided | 2203.04476 | null | https://arxiv.org/abs/2203.04476v2 | https://arxiv.org/pdf/2203.04476v2.pdf | Part-level Action Parsing via a Pose-guided Coarse-to-Fine Framework | Action recognition from videos, i.e., classifying a video into one of the pre-defined action types, has been a popular topic in the communities of artificial intelligence, multimedia, and signal processing. However, existing methods usually consider an input video as a whole and learn models, e.g., Convolutional Neural Networks (CNNs), with coarse video-level class labels. These methods can only output an action class for the video, but cannot provide fine-grained and explainable cues to answer why the video shows a specific action. Therefore, researchers start to focus on a new task, Part-level Action Parsing (PAP), which aims to not only predict the video-level action but also recognize the frame-level fine-grained actions or interactions of body parts for each person in the video. To this end, we propose a coarse-to-fine framework for this challenging task. In particular, our framework first predicts the video-level class of the input video, then localizes the body parts and predicts the part-level action. Moreover, to balance the accuracy and computation in part-level action parsing, we propose to recognize the part-level actions by segment-level features. Furthermore, to overcome the ambiguity of body parts, we propose a pose-guided positional embedding method to accurately localize body parts. Through comprehensive experiments on a large-scale dataset, i.e., Kinetics-TPS, our framework achieves state-of-the-art performance and outperforms existing methods over a 31.10% ROC score. | ['Tao Mei', 'Yongdong Zhang', 'Dong Wu', 'Kun Liu', 'Wu Liu', 'Xinchen Liu', 'Xiaodong Chen'] | 2022-03-09 | null | null | null | null | ['action-parsing'] | ['natural-language-processing'] | [ 4.14712727e-01 -5.64891193e-03 -6.05229259e-01 -3.83714467e-01
-6.15192354e-01 -3.14348489e-01 2.92576730e-01 -2.28949770e-01
-7.71953687e-02 3.02637964e-01 5.07510781e-01 2.27570385e-01
1.64691716e-01 -7.50142872e-01 -8.94772649e-01 -6.54366493e-01
3.20013724e-02 1.84230134e-01 4.63377297e-01 9.33936089e-02
-1.58320248e-01 2.58326054e-01 -1.57893980e+00 6.34081304e-01
5.62634766e-01 1.18530881e+00 -9.13821161e-02 5.98000169e-01
7.15256929e-02 9.17765319e-01 -3.71122658e-01 -3.14744681e-01
8.39667544e-02 -6.48325980e-01 -8.59483302e-01 5.40584207e-01
5.73121965e-01 -5.33126414e-01 -6.15100563e-01 9.86514032e-01
9.90576595e-02 -4.07994166e-02 3.99333149e-01 -1.41003394e+00
-1.51743859e-01 4.31054205e-01 -5.79817057e-01 9.61695537e-02
5.38677812e-01 2.68528789e-01 9.55930591e-01 -6.13875151e-01
6.10739589e-01 1.40177488e+00 4.99210715e-01 6.43269241e-01
-8.22818041e-01 -6.10163391e-01 4.52645928e-01 4.88823265e-01
-1.20063269e+00 -2.80359119e-01 9.05564845e-01 -6.86394155e-01
4.24742967e-01 4.92680743e-02 8.85557234e-01 1.14361227e+00
3.50233138e-01 1.22332180e+00 6.87313378e-01 1.56526770e-02
5.87895662e-02 -6.16201818e-01 -5.38214557e-02 1.08138025e+00
1.41173780e-01 -2.63017595e-01 -5.77849805e-01 3.03773433e-01
1.02349186e+00 4.42114502e-01 -2.24659324e-01 -3.67635220e-01
-1.45761871e+00 5.26003599e-01 2.62387514e-01 1.72130957e-01
-5.70146203e-01 3.62744093e-01 5.84594667e-01 -2.62037724e-01
3.55140418e-01 -6.57049865e-02 -6.75066292e-01 -2.80111223e-01
-8.18637908e-01 1.82684511e-01 5.92224777e-01 7.05983281e-01
6.06346905e-01 -9.47530121e-02 -5.41419744e-01 5.34220874e-01
4.23584372e-01 2.96455950e-01 3.28116357e-01 -1.16364455e+00
6.53982759e-01 9.78715777e-01 -7.27762282e-02 -1.14085186e+00
-3.64845395e-01 -1.09925322e-01 -9.07452583e-01 -8.30666646e-02
5.72067797e-01 -1.28807649e-01 -9.67597485e-01 1.73990071e+00
6.12316251e-01 5.50901175e-01 -1.17741950e-01 1.17173719e+00
9.52418149e-01 7.04249203e-01 2.54489601e-01 -1.78038865e-01
1.68298733e+00 -1.29011118e+00 -5.90920746e-01 -4.94254023e-01
4.91059035e-01 -2.12520704e-01 7.79817402e-01 2.27613583e-01
-9.18959737e-01 -9.45146918e-01 -8.01236033e-01 -1.09626234e-01
-2.11141072e-02 4.57164645e-01 6.10716343e-01 3.07916880e-01
-5.61813593e-01 5.18293500e-01 -1.17022610e+00 -2.06924438e-01
5.86566210e-01 2.59426504e-01 -5.90780914e-01 -2.13127255e-01
-1.11630845e+00 1.75674662e-01 4.10494089e-01 1.17926873e-01
-9.88402545e-01 -3.73856366e-01 -1.20275569e+00 2.14035422e-01
7.60973811e-01 -6.11090004e-01 1.03883827e+00 -1.30783844e+00
-1.39871395e+00 9.14549828e-01 -2.63468385e-01 -3.46226394e-01
2.04223678e-01 -4.08551157e-01 -3.36627811e-01 5.54575384e-01
1.18575059e-01 7.92894602e-01 8.86878431e-01 -8.99226427e-01
-8.72496605e-01 -4.77415144e-01 4.60167378e-01 1.05010062e-01
-1.64235935e-01 1.01431549e-01 -1.08301520e+00 -7.32548058e-01
4.09726679e-01 -8.57251585e-01 -1.60189584e-01 1.48028910e-01
-3.46672118e-01 -3.05751443e-01 6.66208208e-01 -9.31855917e-01
1.28164113e+00 -2.18841887e+00 3.88352066e-01 -3.30132365e-01
3.67427111e-01 9.76928398e-02 -5.95498458e-02 5.82508259e-02
-1.88457649e-02 5.77908792e-02 -2.72869356e-02 -1.45764142e-01
1.60671752e-02 3.72330040e-01 6.68942258e-02 4.83019054e-01
3.26237768e-01 9.71320331e-01 -8.10901165e-01 -8.69353056e-01
2.84240097e-01 4.19266075e-01 -7.08115339e-01 2.64439404e-01
-2.03735620e-01 6.07023180e-01 -9.07668173e-01 1.01633358e+00
2.77172208e-01 -3.23709488e-01 1.72915995e-01 -5.62966645e-01
2.09746197e-01 -9.59268808e-02 -1.16086364e+00 1.77006328e+00
-7.78181255e-02 3.14729720e-01 1.40639499e-01 -1.25288475e+00
4.71089542e-01 4.16141003e-01 9.93513525e-01 -4.05535489e-01
1.28864184e-01 -3.92640084e-02 1.02267154e-01 -8.01238954e-01
1.14231266e-01 9.22423005e-02 -4.67869252e-01 4.87256460e-02
1.54887110e-01 5.51696301e-01 3.11696649e-01 -1.30369693e-01
1.27377212e+00 6.68851793e-01 5.06068528e-01 2.21228480e-01
7.48763263e-01 -1.23092532e-01 1.09368467e+00 3.33974928e-01
-5.76866031e-01 7.21351385e-01 6.24140620e-01 -6.98958635e-01
-4.84811634e-01 -8.21740508e-01 3.08296323e-01 1.23408091e+00
6.25943124e-01 -5.62721252e-01 -1.06874955e+00 -1.03530645e+00
-2.15677872e-01 -4.01209891e-02 -6.76537335e-01 -1.90254897e-01
-9.11095619e-01 -3.76266479e-01 3.93220872e-01 8.62251520e-01
7.93815970e-01 -1.35065699e+00 -6.60482109e-01 2.06479773e-01
-6.16376519e-01 -1.46596372e+00 -6.54089153e-01 -2.51317441e-01
-9.26070809e-01 -1.42062867e+00 -6.14381194e-01 -7.86676228e-01
5.83080649e-01 9.10536051e-02 1.06656516e+00 1.53558210e-01
-1.70520216e-01 3.71996105e-01 -6.77496135e-01 1.09267585e-01
7.08907619e-02 -1.24418370e-01 5.74259423e-02 5.12484252e-01
3.38716388e-01 -2.90351301e-01 -8.37920666e-01 6.36732042e-01
-8.26867580e-01 2.69966185e-01 9.43419993e-01 6.75216436e-01
1.00286269e+00 1.84422851e-01 7.54017234e-02 -5.46459198e-01
-2.59755366e-02 -2.99682707e-01 -3.82517055e-02 4.04456913e-01
1.35183275e-01 -8.75410885e-02 6.83356106e-01 -4.96766120e-01
-8.21284592e-01 6.20654881e-01 -1.39191449e-01 -6.22841895e-01
-5.71826816e-01 3.37195992e-01 -7.64571309e-01 1.56859919e-01
9.05677155e-02 5.38607955e-01 -2.42164537e-01 -3.74705702e-01
2.03652084e-01 3.65940243e-01 6.77203059e-01 -5.81386030e-01
5.45449615e-01 4.63871926e-01 -5.16964402e-03 -5.82125843e-01
-1.12061608e+00 -4.57696378e-01 -9.25914645e-01 -5.18169284e-01
1.57707131e+00 -1.03739285e+00 -7.47870088e-01 6.09683633e-01
-1.10775268e+00 -2.69342989e-01 -6.13862313e-02 6.06224418e-01
-8.21418703e-01 3.92369151e-01 -6.40498340e-01 -4.33858246e-01
-5.53202862e-03 -1.29144764e+00 1.64576161e+00 3.27312082e-01
-1.26916364e-01 -6.28162682e-01 -1.94014162e-01 7.16078341e-01
-4.09300119e-01 4.85943615e-01 5.49108922e-01 -4.07155246e-01
-8.07675660e-01 -1.67516723e-01 -1.60942987e-01 2.42814764e-01
1.28331348e-01 -4.17775213e-04 -5.06962299e-01 1.42311584e-02
-2.29070723e-01 -3.57163966e-01 8.29670131e-01 6.60683095e-01
1.53349030e+00 -3.68241280e-01 -4.38598096e-01 7.68730998e-01
8.95834327e-01 1.85381830e-01 7.19687879e-01 1.15758203e-01
9.46362615e-01 5.71638703e-01 1.03869534e+00 5.13012707e-01
4.16120648e-01 9.42943156e-01 5.02981186e-01 2.90668309e-02
-2.26346985e-01 -5.49913585e-01 5.70307672e-01 5.15244663e-01
-3.39995116e-01 -2.49535382e-01 -5.26993632e-01 2.43533015e-01
-2.08669662e+00 -1.21990943e+00 -4.97494824e-02 1.75863481e+00
5.66221952e-01 1.48449987e-01 2.91933924e-01 9.51117426e-02
8.48506689e-01 5.48026145e-01 -6.79995358e-01 1.22908771e-01
2.08507448e-01 -1.44115090e-01 3.49410802e-01 -2.01411899e-02
-1.65453672e+00 1.02297175e+00 5.32738686e+00 8.44209433e-01
-9.35469568e-01 -3.08880936e-02 7.09116340e-01 1.03326380e-01
3.24363679e-01 -2.36041352e-01 -8.22650909e-01 7.04782248e-01
5.70807874e-01 3.06469858e-01 1.21535912e-01 9.19071734e-01
4.00969744e-01 3.78918648e-02 -1.21622074e+00 1.13778651e+00
2.93099165e-01 -1.10501695e+00 1.21321483e-02 -2.96609793e-02
4.61825669e-01 -4.92684394e-01 -4.10468787e-01 5.78204989e-01
-2.22572207e-01 -9.33008790e-01 8.54786694e-01 6.05402172e-01
7.19835162e-01 -4.40823644e-01 6.71957910e-01 4.84082043e-01
-1.75552535e+00 -7.55948722e-02 -4.80577089e-02 -1.62987828e-01
3.68810654e-01 9.30101946e-02 2.72053154e-03 5.42324126e-01
8.58125627e-01 1.16347766e+00 -4.51835990e-01 8.88080895e-01
-4.64476466e-01 8.05357814e-01 6.84910407e-03 2.98991501e-01
4.44252566e-02 -9.12009403e-02 7.53295273e-02 9.51034963e-01
2.92857170e-01 6.39607787e-01 7.23719299e-01 4.10198331e-01
-1.43406793e-01 -8.56334120e-02 -1.96391210e-01 -2.97237456e-01
1.35794878e-02 1.20321274e+00 -7.59093940e-01 -4.07129645e-01
-5.62751889e-01 1.26482427e+00 1.13195693e-02 1.31389141e-01
-1.10303140e+00 -1.23810351e-01 9.31742132e-01 1.40329912e-01
4.84005690e-01 -1.15055531e-01 1.53317982e-02 -1.35049164e+00
2.92190164e-01 -1.01959014e+00 5.27821600e-01 -7.20845044e-01
-8.58347476e-01 3.27606350e-01 -9.12888646e-02 -1.57207251e+00
-2.94387907e-01 -7.52882242e-01 -5.69485843e-01 3.11332017e-01
-8.98146451e-01 -1.35129380e+00 -5.91881990e-01 6.28615439e-01
1.01401699e+00 1.40546873e-01 4.62620169e-01 4.25971806e-01
-7.54750013e-01 4.83895838e-01 -2.82222122e-01 7.77709782e-01
4.17296678e-01 -9.05727029e-01 9.34942588e-02 9.55226362e-01
4.20344025e-01 2.98140496e-01 4.41705614e-01 -7.24569857e-01
-1.43310356e+00 -1.21379709e+00 6.45410419e-01 -4.32813942e-01
4.48115259e-01 -2.54942060e-01 -6.89260244e-01 7.08986759e-01
-3.62911224e-01 4.35042113e-01 4.50497299e-01 -1.15728430e-01
-2.65423238e-01 -2.81493634e-01 -8.65284085e-01 4.14344132e-01
1.43393397e+00 -3.91710669e-01 -6.65559590e-01 2.84006983e-01
7.88719356e-01 -6.40466213e-01 -8.44846487e-01 6.16076589e-01
8.82803798e-01 -1.06909752e+00 1.16156101e+00 -8.04645717e-01
9.68396187e-01 -4.48665887e-01 -1.75421908e-01 -8.60393226e-01
-4.21907604e-01 -1.37344599e-01 -4.88712847e-01 1.07883644e+00
-1.29891545e-01 -6.40390441e-02 1.04238713e+00 3.83351177e-01
-3.71199965e-01 -1.05169868e+00 -8.06414843e-01 -6.27866149e-01
-4.71101463e-01 -5.01088619e-01 3.92927557e-01 5.59464574e-01
-2.69720167e-01 1.18416511e-01 -7.54413664e-01 3.88433158e-01
4.00023669e-01 4.50714827e-01 9.06330287e-01 -1.09325266e+00
-4.71327752e-01 -2.78303593e-01 -8.75655890e-01 -1.43317199e+00
2.75738269e-01 -4.73244697e-01 2.63957232e-01 -1.70419085e+00
4.09788251e-01 2.29203016e-01 -3.64720166e-01 7.02621996e-01
-2.90374368e-01 4.30139154e-01 1.73367485e-01 2.60778964e-01
-1.04849684e+00 4.49538559e-01 1.46755397e+00 -2.81480879e-01
2.75598187e-02 1.04903638e-01 -4.78165120e-01 1.04231966e+00
5.02771854e-01 -3.83850157e-01 -2.59709626e-01 -2.81618983e-01
-2.61971086e-01 4.55243200e-01 6.28703475e-01 -1.34348512e+00
2.04920303e-02 -4.62572366e-01 5.15243411e-01 -6.12128735e-01
4.77931231e-01 -7.89082944e-01 5.75185381e-02 6.62991881e-01
-3.24129611e-01 -3.14603448e-01 -1.90341502e-01 7.98318386e-01
-5.70931911e-01 1.14505075e-01 5.20732641e-01 -6.91021234e-02
-1.19894540e+00 8.25509429e-01 -1.81673035e-01 1.58087611e-01
1.28422499e+00 -4.03396010e-01 -1.31155565e-01 -2.88578898e-01
-8.14449847e-01 3.01141769e-01 3.17276955e-01 6.38461471e-01
5.10317385e-01 -1.48618710e+00 -5.10909140e-01 1.40186563e-01
2.88050354e-01 -1.58886731e-01 6.50644600e-01 1.00852215e+00
-3.91157120e-01 3.60064507e-01 -3.19573402e-01 -6.62009478e-01
-1.62534738e+00 4.57073599e-01 2.19296455e-01 -2.38352850e-01
-5.84311724e-01 8.08810234e-01 6.08462989e-01 -1.08117104e-01
1.77674800e-01 -5.51340342e-01 -4.32737023e-01 -9.31616127e-02
6.38490200e-01 1.70767158e-01 -5.21815002e-01 -1.11526513e+00
-5.65357983e-01 9.32714939e-01 3.36113960e-01 3.75587285e-01
1.04921329e+00 7.52469897e-02 7.25026131e-02 2.72696406e-01
1.19949424e+00 -2.15337753e-01 -1.70463324e+00 -3.30596715e-02
-3.40596378e-01 -5.32781601e-01 -2.35967100e-01 -5.39555728e-01
-1.36053216e+00 1.10432386e+00 3.99345785e-01 -1.38285249e-01
1.29421616e+00 3.64483029e-01 1.05185378e+00 1.70718729e-01
4.55293000e-01 -9.94221509e-01 2.78489470e-01 3.65779877e-01
5.48843145e-01 -1.27629602e+00 -1.91887952e-02 -6.20357871e-01
-7.17456520e-01 1.10052335e+00 8.97901773e-01 5.29721156e-02
3.86242479e-01 -2.10130289e-01 -1.18541621e-01 -2.16401547e-01
-4.21390206e-01 -3.69643062e-01 6.85997546e-01 4.33932453e-01
4.00829047e-01 1.48003146e-01 -3.72191548e-01 1.04511404e+00
2.57556558e-01 2.81735092e-01 -6.74931984e-03 7.11092234e-01
-6.14775956e-01 -1.00212228e+00 -2.82776684e-01 5.44660270e-01
-5.80529571e-01 2.65101492e-01 -4.55115438e-01 7.57373214e-01
6.85431600e-01 7.98633337e-01 -8.37299377e-02 -7.73699164e-01
2.67400831e-01 5.36484923e-03 3.38583231e-01 -6.20547712e-01
-2.21690491e-01 6.84929118e-02 9.72869713e-03 -1.19681787e+00
-7.43609607e-01 -7.80785322e-01 -1.42511702e+00 9.26000625e-02
1.90783694e-01 -1.07293941e-01 1.03573076e-01 1.23035252e+00
3.15152973e-01 6.36153519e-01 3.82028192e-01 -1.02913487e+00
-2.48426795e-01 -6.95216298e-01 -5.93732178e-01 7.74539649e-01
1.02748331e-02 -8.63839746e-01 -1.00190200e-01 4.33673054e-01] | [8.178271293640137, 0.4891689419746399] |
3cb8ec05-6e44-4e37-9951-aba21e83d192 | very-low-resource-sentence-alignment-luhya-1 | 2211.00046 | null | https://arxiv.org/abs/2211.00046v1 | https://arxiv.org/pdf/2211.00046v1.pdf | Very Low Resource Sentence Alignment: Luhya and Swahili | Language-agnostic sentence embeddings generated by pre-trained models such as LASER and LaBSE are attractive options for mining large datasets to produce parallel corpora for low-resource machine translation. We test LASER and LaBSE in extracting bitext for two related low-resource African languages: Luhya and Swahili. For this work, we created a new parallel set of nearly 8000 Luhya-English sentences which allows a new zero-shot test of LASER and LaBSE. We find that LaBSE significantly outperforms LASER on both languages. Both LASER and LaBSE however perform poorly at zero-shot alignment on Luhya, achieving just 1.5% and 22.0% successful alignments respectively (P@1 score). We fine-tune the embeddings on a small set of parallel Luhya sentences and show significant gains, improving the LaBSE alignment accuracy to 53.3%. Further, restricting the dataset to sentence embedding pairs with cosine similarity above 0.7 yielded alignments with over 85% accuracy. | ['Bruce A. Bassett', 'Everlyn Asiko Chimoto'] | 2022-10-31 | null | null | null | null | ['sentence-embeddings', 'sentence-embeddings'] | ['methodology', 'natural-language-processing'] | [-9.29199014e-05 1.46636199e-02 -2.66291678e-01 -5.31758010e-01
-1.48914576e+00 -6.81490421e-01 6.78826213e-01 2.53067285e-01
-8.63434672e-01 9.30686295e-01 6.73764408e-01 -7.79019833e-01
2.28125229e-01 -6.62029684e-01 -6.96766675e-01 -3.28263134e-01
2.41508111e-02 8.34941506e-01 -3.89990479e-01 -4.91785765e-01
2.00023428e-01 2.61614993e-02 -6.92714751e-01 4.04378653e-01
1.05801940e+00 -1.81719899e-01 2.19440386e-01 9.51718926e-01
-1.74762949e-01 -1.28578335e-01 -4.69481528e-01 -7.77631581e-01
4.40605819e-01 -4.25529897e-01 -9.29682493e-01 -4.94732916e-01
7.35616028e-01 -1.40120059e-01 -6.92464188e-02 6.88593090e-01
8.08513105e-01 -3.19780141e-01 5.82826018e-01 -9.00100052e-01
-1.12009645e+00 9.01816130e-01 -7.77233005e-01 7.10963011e-01
4.72314775e-01 1.61911562e-01 1.50647712e+00 -1.38322628e+00
7.89701879e-01 9.95745897e-01 9.04942095e-01 4.97965008e-01
-1.17973197e+00 -7.05753505e-01 -5.38439870e-01 3.35975513e-02
-1.27173996e+00 -6.85176075e-01 2.25565448e-01 -3.51766169e-01
1.69967401e+00 2.50972509e-01 2.96953768e-01 1.11840510e+00
5.43236673e-01 3.13916713e-01 1.29076636e+00 -8.59701931e-01
-1.62628949e-01 9.19813439e-02 -1.04853734e-01 5.28852522e-01
5.76110840e-01 -3.87732029e-01 -7.20071673e-01 -3.41088861e-01
2.69408584e-01 -4.58651543e-01 2.26541355e-01 2.31707796e-01
-1.51261878e+00 1.14509571e+00 6.26175478e-02 4.93880570e-01
-3.38336825e-01 -1.94052920e-01 4.00477529e-01 6.25768125e-01
6.63322330e-01 9.32559371e-01 -6.39710069e-01 -3.89985114e-01
-7.47430801e-01 3.03666908e-02 6.69290304e-01 8.12338412e-01
6.51244581e-01 -8.94387364e-02 5.91251552e-02 1.21334720e+00
-3.56680192e-02 7.29055345e-01 7.41137207e-01 -3.41157079e-01
1.10829043e+00 1.40715688e-01 1.21604465e-01 -6.80440843e-01
-5.85585497e-02 -1.17382802e-01 -3.78829032e-01 -4.26137000e-01
4.37161088e-01 -3.38635772e-01 -8.92868638e-01 1.87650907e+00
1.62543297e-01 -3.77587199e-01 4.56689537e-01 5.48947036e-01
2.73382455e-01 1.10601580e+00 -2.91317329e-02 3.42762768e-02
1.31260347e+00 -1.11113071e+00 -4.51173097e-01 -6.81424320e-01
1.22965050e+00 -1.62845540e+00 1.48833847e+00 -2.22824350e-01
-1.11493993e+00 -4.54842269e-01 -1.05306876e+00 -9.55064967e-02
-2.60146588e-01 4.86978814e-02 4.01180983e-01 5.65228403e-01
-1.11302912e+00 4.95835096e-01 -7.77701855e-01 -9.05128598e-01
3.94006550e-01 3.66319984e-01 -7.89509058e-01 -3.85608882e-01
-1.20458651e+00 1.45557225e+00 7.55270198e-02 -2.22311556e-01
-1.96920991e-01 -6.75749421e-01 -7.94511020e-01 -2.11359188e-01
-3.73307854e-01 -2.94553936e-01 1.01660907e+00 -4.34886426e-01
-1.26456761e+00 1.27803838e+00 -3.45670640e-01 -5.11697829e-01
1.13940269e-01 -6.55628145e-01 -5.81898630e-01 -1.43704191e-01
6.50350392e-01 6.76984727e-01 1.29991859e-01 -4.70812529e-01
-5.48037231e-01 -1.13710470e-01 -2.92334944e-01 2.23248705e-01
-5.64154565e-01 5.94506621e-01 -1.03381641e-01 -6.53874218e-01
-1.87310725e-01 -9.43191767e-01 -1.57039821e-01 -5.85950196e-01
-6.60804287e-02 -2.53873497e-01 2.42346317e-01 -1.14691842e+00
1.11779559e+00 -1.72650886e+00 4.37836125e-02 2.37218887e-02
-2.91146517e-01 4.71351296e-01 -9.79782641e-01 1.08899546e+00
6.21787608e-02 3.05889398e-01 -1.55964807e-01 -1.48738295e-01
-9.50851589e-02 3.89297277e-01 -1.71409890e-01 4.35306817e-01
7.02547550e-01 1.05305910e+00 -9.11842108e-01 -4.63429570e-01
-2.83979233e-02 2.00692803e-01 -6.87496543e-01 3.47751200e-01
3.29480678e-01 -3.78017984e-02 -8.89209472e-03 6.71905398e-01
4.82959390e-01 2.76622415e-01 3.86327744e-01 2.40545064e-01
-2.44682208e-01 9.84622955e-01 -2.54362047e-01 1.77830148e+00
-9.04637039e-01 8.23114216e-01 -3.20750147e-01 -6.29767537e-01
1.15878367e+00 1.94290638e-01 1.86111376e-01 -9.26544309e-01
-7.72534311e-02 7.70596564e-01 7.05452025e-01 -4.22996014e-01
7.82788277e-01 -2.14609519e-01 -3.93705904e-01 8.83867621e-01
1.71498597e-01 -1.27494082e-01 3.17800343e-01 -9.00696218e-03
1.23765731e+00 3.94194014e-02 3.13981503e-01 -6.52840853e-01
1.82435036e-01 3.32253098e-01 6.01422608e-01 4.70261544e-01
-1.38775945e-01 6.59097195e-01 8.09719637e-02 -5.26291668e-01
-1.87391269e+00 -1.09977293e+00 -2.32897326e-01 1.06688261e+00
-4.91985917e-01 -6.53262198e-01 -6.43884659e-01 -4.92731154e-01
-2.25113437e-01 8.83058727e-01 -4.53010976e-01 4.21965830e-02
-1.15057492e+00 -1.00760746e+00 6.86826944e-01 3.35782856e-01
9.62162837e-02 -8.58401120e-01 -1.73013479e-01 4.86345083e-01
-3.01164269e-01 -1.14112806e+00 -6.31913781e-01 3.46927531e-02
-7.15581894e-01 -7.05754697e-01 -6.65041208e-01 -1.06263757e+00
5.13222873e-01 1.83259249e-01 1.23833609e+00 -1.64330825e-01
-2.92078704e-01 -2.14958027e-01 -4.36410427e-01 -3.06143582e-01
-7.06833124e-01 5.57444751e-01 5.80806136e-01 -5.18488288e-01
1.09187341e+00 -4.98798281e-01 -9.40863937e-02 6.93418980e-02
-3.18066806e-01 -6.84024170e-02 8.37476432e-01 1.04777348e+00
3.10778707e-01 -7.14233994e-01 6.77385569e-01 -8.40775788e-01
7.33860791e-01 -5.70388496e-01 -1.98344454e-01 3.76218528e-01
-5.72457910e-01 3.76955941e-02 7.67608762e-01 -4.33203459e-01
-4.67998058e-01 -4.43623722e-01 -4.72592056e-01 1.07628018e-01
2.05143884e-01 4.00877446e-01 3.93349499e-01 3.27022970e-01
6.85778022e-01 -5.44331595e-02 -1.10542849e-01 -4.65302169e-01
4.53797489e-01 1.21428835e+00 3.65796357e-01 -6.37807310e-01
1.14412212e+00 -1.78349242e-01 -5.28076887e-01 -7.38891900e-01
-6.51266515e-01 -4.60941076e-01 -6.66996896e-01 4.95304137e-01
9.40709531e-01 -9.75932777e-01 1.80269405e-01 -5.23798652e-02
-1.27707744e+00 -2.71845043e-01 -7.06795752e-02 8.87210548e-01
-2.48246089e-01 1.18061639e-01 -6.68339968e-01 -3.42985362e-01
-7.54607737e-01 -1.03040624e+00 8.52882862e-01 -1.25167325e-01
-9.55503464e-01 -8.63834262e-01 8.23324442e-01 4.51935291e-01
5.26101351e-01 -1.08546913e-01 1.10819912e+00 -1.05890203e+00
4.74252850e-02 -8.99071544e-02 -7.31237605e-02 2.53261656e-01
4.93485093e-01 1.18064143e-01 -6.44553423e-01 -5.71382701e-01
-4.23052788e-01 -5.22199154e-01 3.07596147e-01 -1.31935194e-01
2.03015834e-01 -3.60135436e-01 4.00132537e-02 4.09459352e-01
1.39095211e+00 1.23038396e-01 4.41028804e-01 7.66268611e-01
4.50445414e-01 4.85696852e-01 6.96352482e-01 -3.63309085e-02
2.27535650e-01 7.68117189e-01 -4.34008330e-01 -8.06886628e-02
-1.27132639e-01 -4.19655114e-01 7.81412780e-01 1.88224399e+00
2.33786151e-01 -6.95269555e-02 -1.16764760e+00 1.02150977e+00
-1.32266903e+00 -8.29476774e-01 -1.59504279e-01 1.94605112e+00
1.20523632e+00 7.61121884e-02 -1.63686886e-01 -2.68429786e-01
4.29041326e-01 2.15037912e-01 7.12500364e-02 -1.35488534e+00
-1.24587201e-01 1.11194181e+00 6.22893274e-01 7.73838878e-01
-8.34646642e-01 1.21855140e+00 6.42632389e+00 6.22153163e-01
-8.43727529e-01 3.09057623e-01 4.30775225e-01 -2.22671688e-01
-5.91496706e-01 7.89404213e-02 -9.21690345e-01 4.22973067e-01
1.62840092e+00 -2.93546110e-01 2.54042000e-01 5.89103639e-01
9.12610665e-02 1.56252116e-01 -1.01697850e+00 7.01294661e-01
1.71997696e-01 -1.47513556e+00 3.13739404e-02 9.67443734e-02
1.06804633e+00 7.34383702e-01 -1.13285094e-01 3.91746253e-01
5.74729800e-01 -1.21774328e+00 1.72161564e-01 -1.65347189e-01
1.07239604e+00 -8.61496687e-01 1.02178395e+00 1.00597821e-01
-6.99083030e-01 3.09014529e-01 -8.09424341e-01 -1.92418650e-01
3.31549376e-01 2.45796829e-01 -1.10090125e+00 3.93876970e-01
6.33428872e-01 5.06086349e-01 -6.53030097e-01 4.26771075e-01
-2.26439655e-01 9.56382632e-01 -4.40494269e-01 -2.63338506e-01
6.33828580e-01 -5.27568281e-01 3.59068125e-01 1.40888596e+00
6.22753501e-01 -1.83634281e-01 -1.90845579e-01 2.92893589e-01
-2.78602719e-01 4.58603352e-01 -9.00376618e-01 -3.55285466e-01
7.43587375e-01 1.04935181e+00 -1.63304329e-01 -2.15925276e-01
-6.24283731e-01 1.13733697e+00 7.93500960e-01 -6.61624670e-02
-6.04772568e-01 -7.91437328e-01 1.12881649e+00 -1.76451489e-01
2.48872548e-01 -4.66276079e-01 -2.50113517e-01 -1.18796360e+00
1.65191621e-01 -1.15836406e+00 -1.54781388e-02 -5.38092196e-01
-1.41514993e+00 8.93363297e-01 -2.23503143e-01 -9.72277045e-01
-5.00924349e-01 -7.45057046e-01 -8.73819053e-01 1.39397573e+00
-1.27174520e+00 -1.22767055e+00 4.35927391e-01 -4.73028868e-02
8.85554433e-01 -6.34984434e-01 1.38931608e+00 4.03251618e-01
-6.76307023e-01 9.24749196e-01 5.02186775e-01 2.76793301e-01
1.02308643e+00 -9.32210743e-01 1.27206337e+00 1.11361718e+00
6.92212999e-01 1.03211462e+00 6.27572894e-01 -5.32569766e-01
-1.41180611e+00 -1.12435496e+00 2.13059950e+00 -7.38503277e-01
1.02142918e+00 -7.15819299e-01 -7.99016833e-01 7.34732091e-01
7.86027670e-01 -5.77794790e-01 1.20224285e+00 4.74670768e-01
-4.83264357e-01 9.36686769e-02 -9.57325220e-01 9.58547831e-01
9.02651429e-01 -8.18255305e-01 -1.10637486e+00 5.68283498e-01
6.52579784e-01 -1.74448993e-02 -1.07804668e+00 9.66288596e-02
5.78051567e-01 -5.40307760e-01 6.99062645e-01 -9.77410376e-01
8.64617229e-01 9.82601717e-02 -4.59491521e-01 -1.47853637e+00
-3.09504181e-01 -5.96209168e-01 6.01970196e-01 1.33765280e+00
7.76513875e-01 -7.29646683e-01 5.48310935e-01 8.56156647e-02
-1.75238684e-01 -1.00476491e+00 -9.98955190e-01 -9.29912746e-01
7.04758227e-01 -1.29969001e-01 5.00288188e-01 1.11859274e+00
1.48472115e-02 7.86237419e-01 -2.81262875e-01 -9.66355130e-02
1.57451451e-01 -1.63183920e-02 8.72106075e-01 -5.70968688e-01
-2.74264306e-01 -1.63561612e-01 -2.49131024e-01 -6.14410162e-01
1.32959709e-01 -1.23360801e+00 -1.34270042e-01 -1.64408851e+00
3.94884586e-01 -6.11793160e-01 -1.07599329e-02 4.50406492e-01
-2.67877072e-01 7.23079383e-01 -6.91156313e-02 2.91234285e-01
2.01537982e-01 4.08185393e-01 6.39286280e-01 -3.71240862e-02
2.55167373e-02 -5.68330765e-01 -6.10733271e-01 3.69383842e-01
1.25259602e+00 -7.47432530e-01 -6.33459613e-02 -1.06604147e+00
1.15658864e-01 -5.25331080e-01 -4.84702766e-01 -5.62096000e-01
-2.80778646e-01 -2.73192704e-01 1.91720083e-01 -3.64254653e-01
1.86354116e-01 -9.33385566e-02 -1.24471031e-01 4.87355024e-01
-3.48900408e-01 7.95833409e-01 3.10654759e-01 -1.44616470e-01
-1.31628364e-01 -3.14623266e-01 5.93639016e-01 3.97620276e-02
-4.31265712e-01 -7.12203309e-02 -4.03026909e-01 4.02373940e-01
7.21600533e-01 -9.72837657e-02 -2.66482085e-01 -4.35414873e-02
3.35744955e-02 1.14860617e-01 5.30494034e-01 5.52280784e-01
4.21523482e-01 -1.48818946e+00 -1.28599834e+00 4.95132387e-01
3.06684285e-01 -4.27674949e-01 -4.81939137e-01 6.03416324e-01
-8.51244926e-01 5.43995678e-01 -5.67044199e-01 -2.19174042e-01
-1.36479795e+00 -5.90790696e-02 -2.98557162e-01 -3.75598222e-01
-4.09642607e-01 9.48165238e-01 -3.85529399e-01 -9.65260684e-01
-5.50963640e-01 3.74137312e-02 2.64073581e-01 -2.06598267e-01
2.96667814e-01 8.65287632e-02 2.06012074e-02 -9.00239706e-01
-5.22051692e-01 6.36661530e-01 -3.64474505e-01 -5.07751822e-01
1.57324159e+00 8.62305239e-02 -2.64987916e-01 3.47141653e-01
1.48171186e+00 6.27985120e-01 -3.14024538e-01 -4.78656851e-02
2.23445579e-01 -6.46892846e-01 -4.82449800e-01 -5.06026089e-01
-1.99801058e-01 9.96787131e-01 2.11638525e-01 -4.61450905e-01
5.52908242e-01 -1.79635555e-01 1.18097842e+00 6.12703979e-01
3.13703299e-01 -1.06851888e+00 -1.30948856e-01 7.96204209e-01
5.96990824e-01 -1.23030007e+00 2.26159859e-02 2.46207386e-01
-5.02704084e-01 9.31631207e-01 5.14881968e-01 -3.48116159e-01
-4.45106253e-02 1.85041010e-01 4.75945801e-01 1.24748357e-01
-8.33152711e-01 1.48996264e-01 1.09207779e-01 4.81531858e-01
1.00341570e+00 1.92535728e-01 -9.62535203e-01 5.01729995e-02
-8.98964942e-01 -4.40410525e-01 5.44625819e-01 7.86476552e-01
-3.35806727e-01 -1.74332714e+00 -1.59233496e-01 5.63094378e-01
-6.83313847e-01 -6.52618289e-01 -4.57096964e-01 1.00941122e+00
-2.35718876e-01 7.91095555e-01 4.99500662e-01 -5.54477274e-01
3.81904766e-02 4.07784969e-01 4.96839255e-01 -8.62777591e-01
-6.12755835e-01 -4.79658730e-02 6.21616066e-01 -1.52700439e-01
-1.39822647e-01 -5.64750373e-01 -8.85246754e-01 -6.25239134e-01
-2.37852633e-01 4.64532852e-01 6.61175430e-01 7.24716067e-01
3.21588486e-01 -1.39687508e-01 8.21729660e-01 -4.57519084e-01
-8.86865973e-01 -1.48208606e+00 7.83727691e-02 4.69950885e-01
-7.19736814e-02 -1.99600849e-02 -1.36431545e-01 -3.15363467e-01] | [11.309114456176758, 10.215248107910156] |
cb54b6c8-d634-46ae-9868-06fd8b728814 | stereosnakes-contour-based-consistent-object | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Ju_StereoSnakes_Contour_Based_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Ju_StereoSnakes_Contour_Based_ICCV_2015_paper.pdf | StereoSnakes: Contour Based Consistent Object Extraction For Stereo Images | Consistent object extraction plays an essential role for stereo image editing with the population of stereoscopic 3D media. Most previous methods perform segmentation on entire images for both views using dense stereo correspondence constraints. We find that for such kind of methods the computation is highly redundant since the two views are near-duplicate. Besides, the consistency may be violated due to the imperfectness of current stereo matching algorithms. In this paper, we propose a contour based method which searches for consistent object contours instead of regions. It integrates both stereo correspondence and object boundary constraints into an energy minimization framework. The proposed method has several advantages compared to previous works. First, the searching space is restricted in object boundaries thus the efficiency significantly improved. Second, the discriminative power of object contours results in a more consistent segmentation. Furthermore, the proposed method can effortlessly extend existing single-image segmentation methods to work in stereo scenarios. The experiment on the Adobe benchmark shows superior extraction accuracy and significant improvement of efficiency of our method to state-of-the-art. We also demonstrate in a few applications how our method can be used as a basic tool for stereo image editing. | ['Gangshan Wu', 'Tongwei Ren', 'Ran Ju'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['stereo-matching'] | ['computer-vision'] | [ 5.18192828e-01 -6.12586141e-02 2.06357643e-01 -3.24526548e-01
-4.87680167e-01 -5.88423848e-01 3.45080137e-01 -2.86757022e-01
-4.91087794e-01 7.22330272e-01 -2.00834945e-01 2.01618508e-03
1.14588730e-01 -7.85661757e-01 -5.80546916e-01 -4.79798228e-01
6.47168100e-01 4.27529544e-01 1.05225027e+00 -1.09752215e-01
6.43929541e-01 6.69916034e-01 -1.80429077e+00 3.22562382e-02
1.05746353e+00 7.42861867e-01 6.24724865e-01 1.66567430e-01
-4.19369906e-01 2.08767176e-01 -2.53223747e-01 -4.86615628e-01
6.60839796e-01 -2.43473902e-01 -9.28111494e-01 5.73793769e-01
6.96449637e-01 -3.68102551e-01 5.39446995e-02 1.48536611e+00
4.81250107e-01 -5.06291129e-02 5.23307681e-01 -1.01500750e+00
-2.15883374e-01 -3.06077916e-02 -1.14701259e+00 8.11883956e-02
4.05523121e-01 -2.73526698e-01 4.03011352e-01 -9.31962490e-01
9.21697676e-01 1.04341817e+00 4.06709611e-01 5.25588632e-01
-1.11747384e+00 -5.95373213e-01 2.28101090e-01 1.37625366e-01
-1.55356789e+00 -2.62673974e-01 8.98320913e-01 -4.58862036e-01
6.56574667e-01 4.76770937e-01 9.07563210e-01 1.88483641e-01
7.86841661e-02 7.02068865e-01 1.35594726e+00 -5.37195146e-01
-1.02305869e-02 4.27594334e-01 1.69284325e-02 8.00450206e-01
2.39648938e-01 2.18091384e-02 -2.82422215e-01 -1.78222451e-03
1.06593943e+00 1.05244942e-01 -5.42152822e-01 -6.96807146e-01
-1.16492736e+00 4.11252946e-01 4.84880880e-02 3.68604034e-01
6.18509017e-03 -1.78720355e-01 2.43311808e-01 4.11154255e-02
5.46994984e-01 1.14627242e-01 -1.24904320e-01 1.66906968e-01
-1.12682366e+00 1.92922190e-01 4.88226503e-01 1.23685491e+00
8.24413002e-01 -1.77943602e-01 2.95439154e-01 9.86619413e-01
1.50575683e-01 2.74362147e-01 2.40274027e-01 -1.25469279e+00
3.44047040e-01 7.65477836e-01 7.46291596e-03 -1.05313110e+00
-7.18405545e-02 -6.37911707e-02 -6.55101061e-01 5.42989552e-01
3.90963256e-01 2.66893357e-01 -1.00729394e+00 1.37950599e+00
6.58959031e-01 -8.50893781e-02 -2.21711755e-01 9.29108977e-01
6.95688069e-01 5.15745640e-01 -2.82122999e-01 -3.14070702e-01
1.21475124e+00 -8.20162714e-01 -9.63485003e-01 -1.49348900e-01
1.25683814e-01 -1.31751931e+00 7.43996978e-01 4.94210869e-01
-1.47504747e+00 -4.12627220e-01 -9.90913868e-01 -2.03668803e-01
-4.38666493e-01 -1.06906146e-01 6.15691721e-01 6.87891006e-01
-1.10505998e+00 3.64439249e-01 -4.99896914e-01 -4.91105378e-01
2.30911344e-01 5.51120102e-01 -3.80635023e-01 2.21982300e-02
-7.98874140e-01 8.69286716e-01 5.13514519e-01 -6.29323274e-02
-9.43567008e-02 -5.35578668e-01 -6.50916576e-01 -2.17546582e-01
4.60221589e-01 -6.65527821e-01 1.00360131e+00 -8.66932869e-01
-1.24532890e+00 1.37207639e+00 -3.83094847e-01 3.19796316e-02
8.48179579e-01 9.21094976e-03 -1.06771819e-01 2.95623749e-01
2.41268918e-01 8.96577716e-01 6.11049592e-01 -1.52953076e+00
-1.09549224e+00 -5.38481474e-01 9.81270149e-03 5.30927360e-01
-1.77750692e-01 -1.31101191e-01 -1.20617998e+00 -5.97117305e-01
5.71916938e-01 -6.97718203e-01 -9.04449895e-02 3.59197527e-01
-3.22130680e-01 -3.53728831e-02 1.04037273e+00 -5.91765046e-01
1.23969638e+00 -2.12304115e+00 4.75466922e-02 3.06040555e-01
1.04130022e-01 1.03786752e-01 3.95919144e-01 1.33490786e-01
4.01234739e-02 3.19358744e-02 -5.27600586e-01 -1.59488514e-01
-4.10128981e-01 7.34805455e-03 7.68459141e-02 6.02733970e-01
-1.73696727e-01 3.83356005e-01 -4.87845272e-01 -1.23397326e+00
5.27725637e-01 3.51806700e-01 -5.52096903e-01 3.68553661e-02
2.45152846e-01 4.33091074e-01 -3.62860918e-01 7.20713019e-01
1.22854614e+00 8.64590853e-02 -4.39061299e-02 -5.15655458e-01
-5.03238499e-01 -3.15804392e-01 -1.58192062e+00 1.90982127e+00
-2.03383639e-01 4.59116340e-01 2.74235755e-01 -6.99227571e-01
8.37975025e-01 2.05955446e-01 6.52402997e-01 -7.86361933e-01
6.55566603e-02 4.35372531e-01 -3.98724377e-01 -3.99135947e-01
6.22161865e-01 -7.03914315e-02 2.89681077e-01 3.95327359e-01
-3.21795553e-01 -7.46907771e-01 3.89208376e-01 1.27190903e-01
2.88367510e-01 2.80003011e-01 5.27411401e-01 -5.13939977e-01
8.52879286e-01 9.01094154e-02 7.93944895e-01 4.27776247e-01
-1.66988343e-01 9.12741363e-01 -3.52635272e-02 -2.46606410e-01
-1.17593670e+00 -9.28490937e-01 -5.59638143e-01 2.76427329e-01
8.88591051e-01 2.18187850e-02 -1.03775728e+00 -4.26681787e-01
-2.45067462e-01 3.83820295e-01 -2.01410875e-01 4.22289550e-01
-7.41618693e-01 -5.01612365e-01 -1.03663318e-02 3.92321497e-01
9.51206267e-01 -8.13081563e-01 -7.73973763e-01 1.41243829e-04
-2.39177242e-01 -1.18269014e+00 -8.56790543e-01 -2.66564399e-01
-1.25806987e+00 -1.12511075e+00 -9.89391148e-01 -1.19350982e+00
1.01231492e+00 6.92585766e-01 8.87115002e-01 2.37690002e-01
-3.63564610e-01 4.08416092e-01 -7.31363595e-02 -2.74964154e-01
-1.50744766e-01 -2.34452993e-01 -3.59935582e-01 -1.24752536e-01
3.14974457e-01 -5.06864250e-01 -8.08766663e-01 6.03637516e-01
-1.14258826e+00 5.70600033e-01 2.85036355e-01 6.19930029e-01
9.19061959e-01 1.81333348e-01 3.55246142e-02 -1.17825067e+00
3.58090937e-01 7.82629475e-02 -9.43483174e-01 4.13097888e-01
-7.22859919e-01 -1.31917641e-01 2.29587063e-01 6.64535463e-02
-1.61591935e+00 4.82056141e-01 7.89818168e-02 -6.57205507e-02
-1.58891723e-01 -1.07979424e-01 -2.14577213e-01 -3.50350887e-01
1.30484983e-01 3.45695436e-01 9.01366100e-02 -4.94814932e-01
9.23252329e-02 7.70473123e-01 4.41364646e-01 -3.32454562e-01
6.93503678e-01 9.80697870e-01 -3.05637568e-02 -7.53648281e-01
-3.05339783e-01 -8.35227847e-01 -8.59096646e-01 -3.76334935e-01
9.46121335e-01 -6.62033856e-01 -5.56742311e-01 5.62184155e-01
-1.35792291e+00 1.76362440e-01 -3.78466491e-03 4.13161755e-01
-6.16609395e-01 9.80956435e-01 -4.41503167e-01 -8.09019387e-01
-3.14153492e-01 -1.34035659e+00 1.20009923e+00 3.91918689e-01
-3.84807922e-02 -1.02385128e+00 -1.48925036e-01 5.52290142e-01
2.03393027e-01 1.65876642e-01 7.46127963e-01 2.23169383e-02
-8.87280822e-01 1.50041413e-02 -4.66445535e-01 2.22611532e-01
2.37269267e-01 1.47510469e-01 -9.52872097e-01 -3.47314700e-02
2.45335758e-01 1.19879603e-01 6.39072776e-01 3.73316586e-01
1.18356180e+00 1.90381706e-01 -5.36275268e-01 6.76345289e-01
1.87228966e+00 5.98407030e-01 7.39630997e-01 4.32298750e-01
6.37982905e-01 8.83698821e-01 9.74605918e-01 1.78383961e-01
1.04355551e-01 9.49788630e-01 2.72011966e-01 -5.49154460e-01
-7.92111382e-02 3.50489728e-02 -1.18319228e-01 8.55602384e-01
-2.41348103e-01 -2.99551617e-02 -7.36763358e-01 6.19982541e-01
-1.77801323e+00 -9.22153354e-01 -3.90313327e-01 2.32400513e+00
9.24751878e-01 -8.05188045e-02 -1.30597785e-01 2.62959599e-01
1.07446563e+00 -7.25608766e-02 -4.04241234e-01 -4.29594398e-01
-6.02124445e-03 2.31913198e-02 6.84427500e-01 6.99251950e-01
-1.07501423e+00 8.39045644e-01 6.25532961e+00 9.36890066e-01
-9.70352411e-01 7.05388635e-02 4.82184231e-01 6.90490529e-02
-3.57063085e-01 2.06842413e-03 -8.59876454e-01 4.07274336e-01
-1.44722044e-01 -2.19142854e-01 1.65189862e-01 5.83312988e-01
1.44559294e-01 -6.86254561e-01 -8.67896020e-01 1.28100681e+00
1.65093094e-01 -1.20974839e+00 1.11211166e-01 1.95688874e-01
8.77330899e-01 -4.78816569e-01 -1.00225955e-01 -5.77490509e-01
-2.12962806e-01 -6.96855068e-01 9.01587188e-01 3.87665659e-01
8.88298810e-01 -7.65693903e-01 7.40422547e-01 3.49036962e-01
-1.26945639e+00 4.43469584e-01 -4.31147128e-01 2.56645173e-01
5.03124595e-01 5.29060304e-01 -6.00087583e-01 5.32920420e-01
8.11978638e-01 5.51020265e-01 -3.76232237e-01 1.30359936e+00
2.02204496e-01 -2.50356436e-01 -3.78176600e-01 2.29011565e-01
9.74391326e-02 -6.36709988e-01 5.68078101e-01 1.10273194e+00
3.11104476e-01 3.27579200e-01 1.03489007e-03 9.17127728e-01
2.17876002e-01 3.95860106e-01 -8.08822513e-01 4.18488771e-01
3.87510747e-01 1.10142553e+00 -1.19456685e+00 -3.51251930e-01
-6.27884448e-01 1.35719573e+00 -6.80903941e-02 2.71572471e-01
-7.32809246e-01 -4.21428144e-01 -4.26549930e-03 3.76908213e-01
2.24012986e-01 -1.53965980e-01 -4.93963480e-01 -1.15400279e+00
1.34885073e-01 -6.66114628e-01 2.24509358e-01 -7.60215640e-01
-9.92130041e-01 5.60101032e-01 2.52615571e-01 -1.53116083e+00
1.30533464e-02 -3.96989644e-01 -5.03656328e-01 8.75851154e-01
-1.53242326e+00 -9.02980089e-01 -4.21350837e-01 7.84001231e-01
9.08197165e-01 7.69307986e-02 5.38662732e-01 5.21168649e-01
-2.52327293e-01 2.13151082e-01 8.35093558e-02 -1.25107452e-01
6.97225153e-01 -1.19954073e+00 6.44696206e-02 9.93701637e-01
-1.04680866e-01 5.39027989e-01 4.71658319e-01 -7.57983088e-01
-8.94022286e-01 -6.44948423e-01 8.19557667e-01 -7.60300681e-02
-2.71480959e-02 -1.41892374e-01 -7.96271503e-01 3.07369143e-01
4.32330191e-01 -2.83769876e-01 4.56447363e-01 -4.80348736e-01
2.75663942e-01 -1.08799145e-01 -1.40520763e+00 6.99327469e-01
1.11968637e+00 -2.39256456e-01 -5.88906467e-01 1.77451000e-01
3.55326176e-01 -7.29795277e-01 -6.71852827e-01 4.96718556e-01
6.62411392e-01 -1.49922955e+00 9.44804490e-01 1.37628973e-01
1.86554149e-01 -5.44274569e-01 2.89248247e-02 -7.99983144e-01
3.01823411e-02 -4.02278066e-01 5.32906294e-01 1.27570510e+00
2.09901035e-01 -5.63178182e-01 7.90506124e-01 8.02814543e-01
-2.83379816e-02 -6.01390541e-01 -9.78488445e-01 -6.16189957e-01
-2.89711714e-01 -5.77763431e-02 4.27568048e-01 8.84144723e-01
-1.67038262e-01 -4.88003604e-02 -1.69827297e-01 -5.78634031e-02
9.91081178e-01 4.73719925e-01 6.31256461e-01 -1.28022540e+00
-8.04971308e-02 -3.10563564e-01 -4.53752369e-01 -1.13234365e+00
-1.52017802e-01 -6.71820939e-01 -4.66848835e-02 -1.51361811e+00
4.78210926e-01 -5.15455544e-01 2.27579936e-01 -6.79364875e-02
-1.05004899e-01 4.80473548e-01 1.34187073e-01 2.94302553e-01
-2.16234758e-01 2.22259194e-01 1.69700158e+00 -9.76630673e-03
-2.81547934e-01 5.36853587e-03 -4.36563313e-01 1.06351686e+00
7.53566980e-01 -4.08420920e-01 -5.34814000e-01 -4.98579264e-01
9.30712745e-03 -3.09300665e-02 2.01422676e-01 -7.67328680e-01
4.27037209e-01 -2.27154240e-01 2.74907440e-01 -9.25839305e-01
3.72534513e-01 -1.22014606e+00 3.72104883e-01 3.54404032e-01
1.79346099e-01 1.63588524e-02 1.96918756e-01 5.84302902e-01
-5.32337308e-01 -5.72327197e-01 1.10529661e+00 -5.43479502e-01
-1.01140440e+00 1.43008888e-01 -1.25900999e-01 -6.09148629e-02
1.27393854e+00 -1.02862537e+00 5.55365682e-02 -1.64866865e-01
-5.64631104e-01 8.64707381e-02 1.01885259e+00 8.35269019e-02
8.24081182e-01 -1.13602114e+00 -3.18908632e-01 2.54239261e-01
-1.07462496e-01 2.34358802e-01 1.78659350e-01 8.82525504e-01
-1.05864704e+00 4.83611792e-01 -4.26232517e-01 -7.96104252e-01
-1.66610754e+00 2.85729498e-01 3.68018180e-01 3.29900756e-02
-6.44475937e-01 5.76438427e-01 5.89267075e-01 -1.68411434e-01
2.15802372e-01 -1.79435164e-01 -2.27984384e-01 -7.57576004e-02
1.47496924e-01 5.04996300e-01 -7.57857738e-03 -6.00866437e-01
-4.06771660e-01 1.28953588e+00 -1.76639110e-01 -3.69793355e-01
1.19217861e+00 -5.39494812e-01 -2.64337689e-01 2.09964618e-01
1.01706159e+00 3.10934752e-01 -1.14433563e+00 -5.36285937e-02
-2.71784067e-01 -1.03240824e+00 1.76418871e-01 -3.58611465e-01
-1.23722696e+00 8.63051951e-01 8.05022240e-01 9.63163674e-02
1.35494649e+00 -1.59370318e-01 8.15804541e-01 -1.07312575e-01
6.81532383e-01 -1.49041355e+00 -2.76612729e-01 1.59369875e-02
7.70944893e-01 -1.36205447e+00 3.43977779e-01 -1.20832276e+00
-5.92637837e-01 1.25553799e+00 7.35586107e-01 -5.59034292e-04
6.42024815e-01 3.57580096e-01 -6.87848851e-02 -2.11290151e-01
-9.30806026e-02 -2.91288495e-01 3.14763188e-01 5.32205403e-01
5.00116408e-01 -3.04084063e-01 -9.45836663e-01 -3.80857214e-02
1.45552263e-01 1.20539367e-02 6.47404552e-01 1.03530633e+00
-5.18417656e-01 -1.26006424e+00 -5.12179315e-01 1.86115533e-01
-4.79680657e-01 -1.63255557e-02 -3.74504030e-01 1.04655278e+00
2.50201643e-01 8.34441960e-01 1.59566343e-01 1.03311986e-01
5.20532608e-01 -1.26000077e-01 6.59173846e-01 -5.94144225e-01
-3.88875276e-01 5.04320562e-01 -1.65997475e-01 -7.73168683e-01
-9.35112238e-01 -6.85722947e-01 -1.33222520e+00 -1.48910820e-01
-5.63562691e-01 -1.05802290e-01 5.97084463e-01 5.95526636e-01
1.94473758e-01 8.27369615e-02 4.08936948e-01 -7.45159686e-01
-2.65492313e-02 -4.07056957e-01 -8.34735751e-01 5.41128814e-01
-1.80564269e-01 -7.94895470e-01 -3.30501199e-01 4.68675673e-01] | [9.305294036865234, -2.4027369022369385] |
a5c4fb82-333e-4c8a-a001-a9884e38b892 | fine-grained-object-categorization-for | 2210.04613 | null | https://arxiv.org/abs/2210.04613v2 | https://arxiv.org/pdf/2210.04613v2.pdf | Enhancing Fine-Grained 3D Object Recognition using Hybrid Multi-Modal Vision Transformer-CNN Models | Robots operating in human-centered environments, such as retail stores, restaurants, and households, are often required to distinguish between similar objects in different contexts with a high degree of accuracy. However, fine-grained object recognition remains a challenge in robotics due to the high intra-category and low inter-category dissimilarities. In addition, the limited number of fine-grained 3D datasets poses a significant problem in addressing this issue effectively. In this paper, we propose a hybrid multi-modal Vision Transformer (ViT) and Convolutional Neural Networks (CNN) approach to improve the performance of fine-grained visual classification (FGVC). To address the shortage of FGVC 3D datasets, we generated two synthetic datasets. The first dataset consists of 20 categories related to restaurants with a total of 100 instances, while the second dataset contains 120 shoe instances. Our approach was evaluated on both datasets, and the results indicate that it outperforms both CNN-only and ViT-only baselines, achieving a recognition accuracy of 94.50 % and 93.51 % on the restaurant and shoe datasets, respectively. Additionally, we have made our FGVC RGB-D datasets available to the research community to enable further experimentation and advancement. Furthermore, we successfully integrated our proposed method with a robot framework and demonstrated its potential as a fine-grained perception tool in both simulated and real-world robotic scenarios. | ['Georgios Tziafas', 'Hamidreza Kasaei', 'Songsong Xiong'] | 2022-10-03 | null | null | null | null | ['3d-object-recognition', 'object-categorization', 'fine-grained-image-classification'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-2.56563388e-02 -3.38371515e-01 2.34772697e-01 -4.38840121e-01
-4.84903425e-01 -5.72777450e-01 4.29571658e-01 -4.25892125e-04
-3.46311629e-01 4.85170394e-01 -1.21142000e-01 2.08902270e-01
-1.91665322e-01 -7.98534274e-01 -8.92503202e-01 -5.56475699e-01
5.31974547e-02 3.92024368e-01 1.95922032e-01 -2.66422421e-01
1.85494542e-01 6.82861865e-01 -2.07255292e+00 2.02282146e-01
8.54526222e-01 1.52875245e+00 5.62387109e-01 2.51174897e-01
1.18470855e-01 4.90779459e-01 -4.80512321e-01 -1.65728435e-01
5.60781360e-01 2.43718997e-01 -4.28358108e-01 2.80154407e-01
5.78149080e-01 -1.85254529e-01 -1.72923476e-01 1.06503379e+00
4.69019681e-01 1.68083683e-01 7.16213822e-01 -1.57168782e+00
-1.01533449e+00 9.99060348e-02 -5.05847573e-01 -2.04340070e-01
2.47858927e-01 2.09332973e-01 6.85510159e-01 -1.04398167e+00
1.83184028e-01 1.47332895e+00 7.05369055e-01 2.72779286e-01
-9.30136800e-01 -7.45423675e-01 4.45245616e-02 2.49121949e-01
-1.52878642e+00 -1.49998844e-01 6.91962659e-01 -4.95331287e-01
9.67604339e-01 -1.91008419e-01 6.89739645e-01 1.06626678e+00
1.48507565e-01 6.37515128e-01 1.36581683e+00 -2.38732159e-01
3.18814635e-01 -3.93719450e-02 2.21647527e-02 6.81991637e-01
5.23863554e-01 3.11037838e-01 -1.94049597e-01 2.49441639e-01
1.03643501e+00 3.03528368e-01 3.01941521e-02 -7.11679339e-01
-1.64541018e+00 7.02354908e-01 9.08583879e-01 1.26813069e-01
-5.42644083e-01 6.52616397e-02 4.08579797e-01 1.77432615e-02
-4.52109203e-02 3.32763433e-01 -3.60237688e-01 -4.74833511e-02
-2.42069706e-01 3.73134494e-01 7.11222172e-01 1.37053764e+00
8.49882007e-01 -1.44284934e-01 6.59253001e-02 1.15275860e+00
3.97354066e-01 7.77212679e-01 5.07978559e-01 -8.47035170e-01
4.27408278e-01 6.03339911e-01 3.13989639e-01 -1.50194728e+00
-4.97763604e-01 -3.05508047e-01 -1.07022762e+00 2.16553509e-01
3.19838136e-01 3.63679528e-01 -1.14389729e+00 1.33054495e+00
1.71522960e-01 -2.59189755e-01 3.07672843e-02 1.40077949e+00
1.05499136e+00 4.08453256e-01 1.82420596e-01 3.49435210e-01
1.41321361e+00 -1.17454398e+00 -4.17363375e-01 -3.33962113e-01
1.45906195e-01 -7.32558429e-01 1.40888345e+00 2.59197801e-01
-5.66078722e-01 -9.97820616e-01 -1.18230820e+00 2.17887908e-02
-5.45591354e-01 3.90203416e-01 6.90042496e-01 3.23603481e-01
-7.27428794e-01 2.57016838e-01 -6.78780854e-01 -7.08123147e-01
3.15959334e-01 1.19566560e-01 -6.32888556e-01 -4.13441479e-01
-8.86608124e-01 9.66215432e-01 5.45564711e-01 3.02113891e-01
-9.09628987e-01 -1.98669106e-01 -9.98162508e-01 -1.00093260e-01
2.14537919e-01 -5.16777694e-01 1.18831921e+00 -4.45052505e-01
-1.24426138e+00 8.57685506e-01 3.91104490e-01 -1.40045896e-01
3.44399869e-01 -1.68426543e-01 -3.44005495e-01 -3.53383273e-02
4.40691233e-01 8.59032035e-01 5.38177192e-01 -1.43058801e+00
-9.21779931e-01 -6.27611637e-01 3.08796018e-01 1.87157080e-01
5.02935722e-02 -4.53013062e-01 -4.86489266e-01 -5.88948250e-01
4.79407221e-01 -1.09575605e+00 -1.17664777e-01 1.70896217e-01
-2.43264079e-01 -2.28212193e-01 6.85132384e-01 -1.89810604e-01
2.88306624e-01 -2.31302428e+00 -4.00060415e-02 -1.70754448e-01
-1.00634299e-01 2.69276239e-02 -1.61337420e-01 1.67412758e-01
2.47546926e-01 -1.78014338e-01 -1.00846170e-03 -1.56746343e-01
3.51958275e-01 4.76418048e-01 5.48062660e-02 4.24940735e-01
3.08880478e-01 8.82180810e-01 -7.57454395e-01 -4.56441671e-01
6.73798025e-01 2.67074525e-01 -2.45765612e-01 2.92638183e-01
-3.91252153e-03 3.33617628e-01 -4.96596694e-01 1.25881457e+00
7.69995332e-01 -1.52919486e-01 -1.75057366e-01 -7.03674257e-01
-7.55255343e-03 -3.49902660e-01 -1.35062432e+00 1.70073545e+00
-5.56403875e-01 2.81877577e-01 1.17027149e-01 -9.72148776e-01
1.33472919e+00 -1.18391126e-01 1.97059259e-01 -1.13519943e+00
2.46472657e-01 3.69444340e-01 -2.24015266e-01 -5.80872476e-01
8.30067873e-01 -7.62610212e-02 -5.41991711e-01 -2.33714916e-02
-2.03038082e-02 -3.80031437e-01 1.10717803e-01 -4.00902152e-01
8.14537585e-01 2.62294203e-01 2.47369036e-01 -2.03714386e-01
2.13008896e-01 3.69734734e-01 5.16741812e-01 7.66246140e-01
-7.53920972e-01 6.71082497e-01 -3.12019169e-01 -6.01540446e-01
-1.15526855e+00 -1.15766263e+00 -5.74815460e-02 9.08164024e-01
7.86238313e-01 1.24333315e-01 -3.73274714e-01 -4.55578715e-01
2.85281718e-01 3.33741903e-01 -5.40254533e-01 -8.41567814e-02
-4.20936972e-01 -4.21225518e-01 5.20094156e-01 6.61397398e-01
1.11082935e+00 -1.25925100e+00 -9.13222790e-01 8.93999711e-02
-3.89781952e-01 -1.51799428e+00 -1.69223607e-01 3.12471807e-01
-5.26610851e-01 -1.21357894e+00 -6.01249516e-01 -1.21568680e+00
3.74926478e-01 8.19212258e-01 9.51612711e-01 -3.94564301e-01
-3.41013819e-01 2.97642559e-01 -6.18653297e-01 -4.60219532e-01
-8.68111923e-02 -3.48484106e-02 1.76946148e-01 -2.00812966e-01
5.00499070e-01 -3.52175206e-01 -6.45680010e-01 8.01344812e-01
-6.42639697e-01 -1.09080732e-01 9.45727527e-01 8.71769965e-01
8.49575937e-01 1.24882162e-01 5.64152896e-01 -2.01314226e-01
5.81343353e-01 -3.17779571e-01 -4.85989004e-01 1.58702970e-01
-3.29288900e-01 -2.19937533e-01 6.41499519e-01 -5.31060338e-01
-7.45107532e-01 1.98620275e-01 4.05481039e-03 -6.01418495e-01
-7.08575785e-01 3.37115973e-01 -1.94204718e-01 -1.14827365e-01
6.29111052e-01 1.14660539e-01 -1.23313732e-01 -5.54528177e-01
2.78595179e-01 1.04311419e+00 8.82136822e-01 -5.49413025e-01
5.37821531e-01 2.85582811e-01 -1.96781799e-01 -7.20675826e-01
-5.90117574e-01 -6.14121675e-01 -6.60772562e-01 -2.16821581e-01
8.47758949e-01 -1.19861615e+00 -8.56389940e-01 8.41510892e-01
-8.48452926e-01 -2.44721562e-01 -1.18126743e-01 6.89641178e-01
-7.57411361e-01 1.97637692e-01 -4.12901521e-01 -5.79426885e-01
-1.78644970e-01 -1.38173056e+00 1.47303355e+00 2.92344391e-01
6.71491697e-02 -3.81325692e-01 -3.88006121e-01 4.58548874e-01
4.58045125e-01 5.62630415e-01 7.56058991e-01 -1.92141384e-01
-5.24743438e-01 -1.34285063e-01 -7.09498763e-01 8.40903968e-02
4.06706721e-01 -2.52589345e-01 -8.74098957e-01 -2.77217120e-01
-2.88924187e-01 -6.81860685e-01 4.63846952e-01 4.61127497e-02
9.67938900e-01 1.21013738e-01 -2.22888201e-01 4.68860835e-01
1.44912505e+00 3.15248758e-01 4.85801816e-01 7.41344869e-01
8.62107575e-01 4.60264981e-01 1.24423039e+00 3.87584686e-01
7.07001507e-01 8.35880637e-01 8.78014743e-01 2.25374829e-02
-1.83480486e-01 -2.78573602e-01 -1.19088478e-01 6.55005991e-01
3.03327963e-02 -8.92917588e-02 -8.32523882e-01 7.19449282e-01
-1.90641963e+00 -6.06744945e-01 2.97215790e-03 1.77050829e+00
3.84806842e-01 4.13737744e-02 1.76666062e-02 1.92536384e-01
7.93064654e-01 -1.89049914e-01 -6.66745305e-01 -2.73566484e-01
-1.12637654e-01 -8.48776177e-02 6.40919805e-01 -1.85809791e-01
-1.49698055e+00 9.27532852e-01 5.61509562e+00 6.34436071e-01
-1.21516287e+00 -9.78540629e-02 2.62013435e-01 3.56728822e-01
3.05658251e-01 -7.22916305e-01 -6.09966218e-01 3.89370888e-01
5.94914675e-01 4.18216914e-01 4.27873462e-01 1.32975173e+00
-3.25389579e-02 -7.96567425e-02 -1.00598204e+00 1.48827910e+00
5.19472808e-02 -9.47067976e-01 -9.51971337e-02 -9.06860754e-02
6.28764153e-01 1.49797410e-01 -1.82604119e-01 5.67515492e-01
2.83713937e-01 -1.07295418e+00 1.19633245e+00 3.12902004e-01
8.59834850e-01 -6.50682092e-01 8.40821087e-01 4.56580728e-01
-1.40355754e+00 -2.86065251e-01 -5.90843678e-01 -1.64258450e-01
-1.12809189e-01 3.85351717e-01 -9.29131091e-01 4.76071924e-01
1.34963334e+00 7.12360799e-01 -4.49962258e-01 1.06181359e+00
-6.55222014e-02 -1.48102269e-01 -3.15712065e-01 -2.84775972e-01
2.54138559e-01 -3.83834243e-02 3.91654372e-02 9.87316370e-01
3.65782201e-01 6.92424476e-02 4.05827999e-01 7.17017889e-01
1.07573479e-01 -2.42270485e-01 -5.95189869e-01 1.27009183e-01
5.52672744e-01 1.33243370e+00 -7.77146816e-01 8.03877320e-03
-3.31645399e-01 1.02016592e+00 3.98444176e-01 1.84103519e-01
-8.75774920e-01 -5.46703339e-01 7.60566115e-01 -3.51148486e-01
5.16398311e-01 -6.09452963e-01 -2.45152280e-01 -9.96790826e-01
1.84294328e-01 -8.59490871e-01 6.94237128e-02 -1.11271596e+00
-1.56017911e+00 7.23735809e-01 -1.35311574e-01 -1.41723931e+00
-6.81336746e-02 -9.56232667e-01 6.69663846e-02 7.62512326e-01
-1.72403431e+00 -1.53109705e+00 -1.16702604e+00 7.30859756e-01
7.06611753e-01 2.95908805e-02 9.33961689e-01 4.63486671e-01
-2.84982562e-01 5.19831836e-01 1.17875844e-01 1.40572354e-01
6.03434682e-01 -1.05255449e+00 2.97829330e-01 3.91019583e-01
-3.19305271e-01 4.37646657e-01 5.17705321e-01 -4.05556470e-01
-1.75899923e+00 -1.59508705e+00 3.58685315e-01 -2.11335510e-01
3.21590394e-01 -3.83482009e-01 -5.79010248e-01 4.51644093e-01
-2.95910895e-01 2.67858714e-01 2.42633939e-01 -1.75293952e-01
-3.43977481e-01 -1.60702646e-01 -1.54298365e+00 3.74395251e-01
1.22214508e+00 -4.50299770e-01 -7.52294481e-01 6.83879927e-02
7.56583333e-01 -4.51177329e-01 -1.19764745e+00 7.43135750e-01
8.66964817e-01 -9.29521859e-01 1.05228209e+00 9.11800787e-02
4.25765425e-01 -6.36976302e-01 -9.11075830e-01 -1.47881222e+00
-5.70866048e-01 3.77022028e-01 1.49198025e-01 9.87538218e-01
-4.97314110e-02 -6.06822848e-01 6.05633855e-01 3.09719265e-01
-4.73313838e-01 -4.80617166e-01 -6.41294777e-01 -8.80206823e-01
-2.52709627e-01 -4.25548881e-01 8.71938348e-01 7.25568712e-01
-4.07254577e-01 2.14827955e-01 -1.44683167e-01 4.03045058e-01
6.19820893e-01 5.34004390e-01 9.86001015e-01 -1.20646882e+00
-1.09338164e-02 -1.83937207e-01 -8.27930987e-01 -9.87051368e-01
-1.90695733e-01 -5.57205975e-01 5.18657923e-01 -1.70033538e+00
1.30866945e-01 -9.19875383e-01 -2.00903773e-01 5.17250597e-01
1.97269306e-01 7.62054741e-01 3.24518085e-01 3.83034259e-01
-6.12362504e-01 7.18210876e-01 1.42600596e+00 -5.18169284e-01
1.91284437e-03 -1.93033099e-01 -6.00675106e-01 5.97899079e-01
8.15313399e-01 8.67109969e-02 -1.63746029e-01 -5.32507896e-01
-3.61077875e-01 -1.53100759e-01 5.92834234e-01 -1.26129389e+00
9.62687209e-02 -1.35014534e-01 5.48147440e-01 -8.45291197e-01
5.12961805e-01 -1.05993652e+00 2.04035178e-01 3.47423434e-01
4.27572615e-02 -8.20635855e-02 2.35211611e-01 4.54252869e-01
-3.59047920e-01 2.62844086e-01 8.39421511e-01 -2.95961082e-01
-1.47469878e+00 2.35845089e-01 -7.55148530e-02 -2.49799758e-01
1.21307850e+00 -6.02313638e-01 -3.86479169e-01 3.24411057e-02
-5.28993309e-01 2.03933716e-01 6.76755965e-01 8.75330865e-01
7.94278562e-01 -1.60730708e+00 -3.94274950e-01 5.25112033e-01
7.28049874e-01 3.03350657e-01 2.56544828e-01 4.43999916e-01
-6.52476430e-01 6.95636451e-01 -7.90825427e-01 -9.46040630e-01
-9.18545365e-01 6.89447641e-01 2.30300099e-01 2.71332502e-01
-4.71241355e-01 5.06388128e-01 2.31988251e-01 -1.05680168e+00
4.35567051e-01 -7.32470036e-01 -2.33198896e-01 -2.37613857e-01
2.05559552e-01 4.27013695e-01 2.28113830e-01 -9.86705601e-01
-5.01822293e-01 9.08786595e-01 2.98585922e-01 4.20046300e-01
1.27547419e+00 -3.55025381e-01 2.37277970e-01 3.93031389e-01
1.23030090e+00 -5.34842372e-01 -1.29971933e+00 -1.81068659e-01
-1.75281271e-01 -5.78640819e-01 -2.37886235e-01 -9.21423793e-01
-8.65686595e-01 7.63864160e-01 1.00950849e+00 2.01173544e-01
1.07579851e+00 9.66929123e-02 6.62860215e-01 5.20604730e-01
1.19259775e+00 -8.65974486e-01 2.34197348e-01 6.78815782e-01
1.02977550e+00 -1.66159093e+00 -3.70303154e-01 -5.36087275e-01
-5.80667078e-01 9.40215111e-01 8.18684340e-01 -2.72610903e-01
3.48676890e-01 -6.24358915e-02 2.46277764e-01 -2.44883910e-01
-1.92933425e-01 -2.46151999e-01 6.75450787e-02 9.94434893e-01
9.08263475e-02 4.02419090e-01 1.53159454e-01 5.52671969e-01
-3.96336079e-01 1.41069666e-01 2.07988352e-01 1.14714241e+00
-5.89685917e-01 -5.27748227e-01 -5.88929296e-01 2.08843619e-01
-8.40635970e-03 2.55023509e-01 -1.14438511e-01 1.00390363e+00
5.06051481e-01 1.28265274e+00 1.50979966e-01 -7.41924405e-01
8.65634143e-01 -5.25743961e-01 5.19103885e-01 -4.15780842e-01
-3.07171375e-01 -1.41031712e-01 4.17403355e-02 -7.13601589e-01
-5.29589117e-01 -3.47823828e-01 -1.15097034e+00 -2.70987004e-01
-1.72764674e-01 -1.12703033e-01 9.71965075e-01 7.29652286e-01
4.38642740e-01 5.66765904e-01 5.80557466e-01 -1.37085044e+00
-6.11412585e-01 -1.07507646e+00 -8.50550592e-01 6.15219831e-01
1.86978281e-01 -1.27092648e+00 -8.64756554e-02 -1.24675587e-01] | [7.625561237335205, -1.3058170080184937] |
824b8856-784e-41bf-ac63-3984e91f8fd5 | high-throughput-virtual-screening-with-data | 1312.1003 | null | http://arxiv.org/abs/1312.1003v1 | http://arxiv.org/pdf/1312.1003v1.pdf | High Throughput Virtual Screening with Data Level Parallelism in Multi-core Processors | Improving the throughput of molecular docking, a computationally intensive
phase of the virtual screening process, is a highly sought area of research
since it has a significant weight in the drug designing process. With such
improvements, the world might find cures for incurable diseases like HIV
disease and Cancer sooner. Our approach presented in this paper is to utilize a
multi-core environment to introduce Data Level Parallelism (DLP) to the
Autodock Vina software, which is a widely used for molecular docking software.
Autodock Vina already exploits Instruction Level Parallelism (ILP) in
multi-core environments and therefore optimized for such environments. However,
with the results we have obtained, it can be clearly seen that our approach has
enhanced the throughput of the already optimized software by more than six
times. This will dramatically reduce the time consumed for the lead
identification phase in drug designing along with the shift in the processor
technology from multi-core to many-core of the current era. Therefore, we
believe that the contribution of this project will effectively make it possible
to expand the number of small molecules docked against a drug target and
improving the chances to design drugs for incurable diseases. | ['Upul Senanayake', 'Rahal Prabuddha', 'Roshan Ragel'] | 2013-12-04 | null | null | null | null | ['molecular-docking'] | ['medical'] | [-1.67853851e-02 -5.53597033e-01 -1.07113034e-01 -9.60115194e-02
-2.46496603e-01 -5.58418691e-01 2.20367014e-01 4.85257179e-01
-5.84827542e-01 1.05354023e+00 -1.83940366e-01 -8.60936224e-01
1.45704731e-01 -8.87017727e-01 -5.39982378e-01 -7.18328953e-01
2.86034811e-02 6.16571844e-01 2.74985999e-01 -2.80279845e-01
5.55145562e-01 1.13075113e+00 -1.34420061e+00 2.02597052e-01
8.30489099e-01 1.26273036e-01 1.36593387e-01 5.27799726e-01
1.52317345e-01 1.66210592e-01 -3.70141864e-01 -1.89628601e-01
3.35968286e-02 -3.66224647e-01 -8.74444008e-01 -6.35647655e-01
5.87047897e-02 9.32242721e-02 2.30439320e-01 7.31608748e-01
8.82404149e-01 -2.36065146e-02 2.52646804e-01 -5.13626993e-01
4.02015179e-01 -2.73486495e-01 -4.26871151e-01 1.58750340e-01
5.27226925e-01 3.42028052e-01 5.79449654e-01 -8.14724207e-01
8.34307790e-01 1.08305883e+00 4.68410045e-01 3.48841578e-01
-1.16420531e+00 -6.48249865e-01 -5.31937718e-01 4.55529362e-01
-1.55079186e+00 -3.20805430e-01 1.07828580e-01 -3.35565716e-01
1.73891807e+00 5.58246732e-01 7.52863288e-01 3.43140543e-01
8.21232378e-01 -1.01046398e-01 1.00306237e+00 -5.14910758e-01
6.41322374e-01 7.51784490e-03 -9.75591689e-02 6.37455106e-01
5.17980278e-01 2.94937909e-01 -4.43072617e-01 -6.44184053e-01
4.48975801e-01 6.58557490e-02 -1.81889176e-01 -4.07457620e-01
-9.90569770e-01 1.08367777e+00 3.37758392e-01 2.98820943e-01
-4.58060682e-01 -9.19138342e-02 5.56598067e-01 7.03926235e-02
-4.18679900e-02 8.71732950e-01 -7.60269344e-01 -3.94478738e-01
-6.92392886e-01 3.80441695e-01 1.07226610e+00 2.83090055e-01
6.01635516e-01 -3.40543836e-01 3.94782215e-01 1.98246703e-01
4.29702163e-01 2.41533220e-01 2.06620872e-01 -2.67107546e-01
-6.17119968e-02 7.88962364e-01 7.87627846e-02 -6.18872523e-01
-7.93369710e-01 -1.86828300e-01 -3.53520095e-01 6.37220979e-01
2.68019378e-01 -2.18930557e-01 -8.73265922e-01 1.13842964e+00
7.91068137e-01 -1.41228557e-01 1.48795143e-01 6.13811016e-01
4.47829932e-01 5.95228612e-01 3.74802917e-01 -4.58364546e-01
1.56000733e+00 -6.38293147e-01 -3.78020585e-01 4.92840737e-01
8.89441133e-01 -1.39653194e+00 6.84429348e-01 6.12902284e-01
-6.98932946e-01 -3.68944436e-01 -1.38396621e+00 2.26244912e-01
-6.04362369e-01 -4.31945324e-01 9.92039800e-01 1.04635108e+00
-7.66695261e-01 1.00991189e+00 -1.22296667e+00 -3.50266695e-01
3.10657233e-01 1.25142825e+00 -6.92265093e-01 -2.10983530e-01
-8.92109096e-01 1.18708205e+00 5.55261731e-01 -3.34721535e-01
-4.73514438e-01 -8.60967338e-01 -3.24947149e-01 -1.91790238e-01
1.91820472e-01 -7.25099444e-01 6.75086319e-01 -3.16462487e-01
-1.48588717e+00 4.88776624e-01 -3.03801894e-01 -2.81258464e-01
4.22724448e-02 9.66243073e-02 -1.12136386e-01 -1.56188324e-01
-2.54254550e-01 4.23737168e-01 1.13774367e-01 -4.38980222e-01
-4.42749709e-01 -6.60386324e-01 -2.13711277e-01 2.99829453e-01
1.98578596e-01 4.26150233e-01 -1.84895456e-01 -3.70335095e-02
-3.82787079e-01 -1.15546477e+00 -6.91254497e-01 -4.66327786e-01
-3.06643005e-02 -8.52493495e-02 6.96954370e-01 -2.78991938e-01
9.83760357e-01 -1.81695497e+00 3.23173761e-01 5.96768975e-01
1.55005604e-01 7.00254142e-01 2.77316451e-01 9.48286533e-01
-3.53408605e-01 -7.99276307e-02 2.87442416e-01 4.13390517e-01
-4.70871478e-01 1.60181701e-01 1.75891489e-01 6.52419806e-01
-2.55772680e-01 4.92622256e-01 -5.68909585e-01 -5.70081212e-02
2.45219857e-01 9.00785148e-01 -7.41685450e-01 -4.78383489e-02
-2.88233906e-01 5.49049675e-01 -6.62244141e-01 4.49946970e-01
7.72054613e-01 -2.41490781e-01 4.96396542e-01 -8.77554268e-02
-5.86935759e-01 1.91802710e-01 -1.03083861e+00 1.57090318e+00
-2.43331771e-02 5.66710159e-03 -3.29327822e-01 -2.87597150e-01
7.14841247e-01 4.79970843e-01 5.49441278e-01 -7.58265257e-01
1.80558771e-01 5.39910972e-01 6.83888257e-01 -1.51964161e-03
2.66519547e-01 -3.55861723e-01 4.11198944e-01 -7.27392435e-02
-3.49459857e-01 2.49061540e-01 3.09626043e-01 -7.05027254e-03
1.13672531e+00 1.55591726e-01 5.64845204e-01 -4.48735774e-01
8.27770770e-01 6.36251509e-01 3.37828159e-01 -1.09061643e-01
5.24890982e-02 -1.90350220e-01 4.61601794e-01 -9.06116903e-01
-1.24308360e+00 -4.56681341e-01 -5.62124074e-01 7.46691287e-01
-2.88687289e-01 -8.36134195e-01 -7.47160196e-01 -2.49738902e-01
-5.85879907e-02 2.63229489e-01 -2.38057867e-01 8.62836987e-02
-4.83762592e-01 -1.13677514e+00 2.29251117e-01 -1.14571124e-01
-8.39736834e-02 -8.96859109e-01 -7.80314624e-01 7.28897035e-01
7.99901903e-01 -4.76152122e-01 -7.22793490e-02 7.05862284e-01
-9.94165182e-01 -1.33299613e+00 -4.50437874e-01 -4.16631997e-01
4.00893301e-01 5.97445518e-02 7.76869059e-01 2.84585595e-01
-8.27600002e-01 -5.06992638e-01 -3.49973917e-01 -5.49947083e-01
-3.76933277e-01 -6.73023015e-02 9.28675383e-02 -6.77003205e-01
6.86151624e-01 -5.52328825e-01 -7.85649836e-01 1.73794672e-01
-6.17072284e-01 7.09463358e-02 4.25765276e-01 6.60216331e-01
9.33522701e-01 8.89428481e-02 2.59555429e-01 -1.06512558e+00
4.48274076e-01 -8.71175230e-02 -8.82991016e-01 -8.94307420e-02
-7.32487261e-01 3.27105880e-01 7.09721208e-01 -8.10064152e-02
-4.68550980e-01 5.29610097e-01 -7.24014819e-01 7.86373839e-02
-6.37558848e-02 6.16300404e-01 -3.97977173e-01 -8.27059627e-01
5.65990508e-01 -1.38768703e-01 2.09062919e-01 -6.52752757e-01
-1.21418290e-01 4.47153717e-01 -1.57331496e-01 -3.29077065e-01
2.69081861e-01 1.83583468e-01 6.89733624e-01 -8.33067536e-01
1.03162676e-01 -6.96999490e-01 -3.23143512e-01 2.98112899e-01
9.44564581e-01 -8.43822777e-01 -1.37987912e+00 2.04456270e-01
-1.03907049e+00 -7.43392706e-02 3.18006366e-01 6.21908009e-01
-1.25076428e-01 4.26920682e-01 -2.18597397e-01 -1.17376380e-01
-5.68862438e-01 -1.82552481e+00 6.84809029e-01 3.89019132e-01
-3.53708863e-01 -8.92815351e-01 6.60548747e-01 4.05610561e-01
4.70991999e-01 5.75748384e-01 1.14989221e+00 -7.25640357e-01
-6.14312649e-01 -1.90956831e-01 1.24900460e-01 -1.59305125e-01
1.24424100e-01 2.79752016e-02 -5.58159411e-01 -4.62585330e-01
-1.40965611e-01 3.26313153e-02 3.17748964e-01 1.12286024e-01
6.48712397e-01 4.19794917e-02 -6.55138075e-01 5.12358248e-01
1.78862131e+00 7.89118946e-01 1.00145841e+00 5.68970799e-01
6.37106419e-01 1.37429804e-01 9.09025669e-01 4.96122271e-01
-2.05225483e-01 1.21151006e+00 5.34386098e-01 -4.22066361e-01
2.67146021e-01 2.73621887e-01 2.40897685e-01 3.75490546e-01
-4.94641632e-01 7.43790641e-02 -1.09471619e+00 -1.80563390e-01
-1.49494624e+00 -8.67846727e-01 -4.62306470e-01 2.56830311e+00
1.16497386e+00 -1.62255332e-01 1.46865591e-01 3.55869010e-02
1.84590966e-01 -6.22609615e-01 -6.15398586e-01 -1.00911713e+00
3.73262525e-01 1.08641052e+00 7.02255785e-01 6.91215217e-01
-7.77218759e-01 1.00800264e+00 5.82049227e+00 9.25734580e-01
-1.42374527e+00 -1.75675228e-01 4.01496738e-01 1.33178383e-01
2.97999084e-01 1.29320145e-01 -1.00293219e+00 4.38704818e-01
1.56185007e+00 -4.26256061e-01 2.20469549e-01 9.05824363e-01
5.45482934e-01 -5.35369039e-01 -1.16894257e+00 7.67087102e-01
-3.64147186e-01 -1.74776828e+00 5.72728366e-02 7.97528565e-01
4.08428937e-01 1.91198617e-01 -3.54485780e-01 -1.38957620e-01
-1.21897519e-01 -1.49790072e+00 -3.09280664e-01 1.39280692e-01
5.23548841e-01 -1.45058632e+00 9.86752391e-01 2.34654412e-01
-9.82354105e-01 6.34750485e-01 -4.21830088e-01 -2.29228914e-01
2.43806886e-03 2.87387937e-01 -1.22153008e+00 5.72085857e-01
3.24899048e-01 1.18550420e-01 -5.87858140e-01 1.40085053e+00
3.56552243e-01 6.71219826e-02 -3.46863925e-01 -2.29277655e-01
1.99201390e-01 -2.93669432e-01 1.90842584e-01 8.64692390e-01
-9.63407159e-02 1.12177476e-01 3.67876917e-01 -2.16958318e-02
1.56601250e-01 5.59106231e-01 -3.15098673e-01 -1.80523489e-02
1.66628256e-01 1.02690494e+00 -8.63861978e-01 -1.29310459e-01
-3.81174833e-01 8.77998233e-01 -1.63136855e-01 -3.60372066e-01
-8.68475080e-01 -5.86072624e-01 1.06276214e+00 2.54996896e-01
2.14412913e-01 -3.03377390e-01 5.41665684e-03 -6.36340261e-01
-5.85570514e-01 -1.35611641e+00 1.83131099e-01 -3.06600541e-01
-5.00200927e-01 5.30797660e-01 -3.23722452e-01 -7.39218354e-01
1.42425507e-01 -9.14401948e-01 -4.26002473e-01 1.39680374e+00
-1.29473472e+00 -6.97171986e-01 2.20952302e-01 5.65509021e-01
3.13424319e-01 -8.30780566e-02 1.40525270e+00 5.48395813e-01
-4.74165976e-01 4.21904594e-01 4.73915368e-01 -5.99026978e-01
8.10638011e-01 -9.57619905e-01 1.43039018e-01 3.77633780e-01
-1.63728774e-01 1.13744450e+00 9.26866710e-01 -8.50910306e-01
-2.01052761e+00 -6.16118550e-01 7.69757688e-01 -2.72842467e-01
5.71618021e-01 -7.62666808e-03 -7.89974928e-01 4.59339097e-02
2.44267732e-01 -4.09127980e-01 1.40624797e+00 1.85691774e-01
2.95632565e-03 2.86415726e-01 -1.15795076e+00 4.84038204e-01
2.19578966e-01 -6.79859892e-02 -3.89263034e-02 8.02919805e-01
4.86537278e-01 -7.61330187e-01 -1.19059515e+00 1.89592287e-01
4.72902805e-01 -8.71193111e-01 9.80534136e-01 -7.29469717e-01
-9.81294438e-02 -7.47923493e-01 5.96746616e-02 -8.50758314e-01
-2.83607751e-01 -6.55527949e-01 2.20603794e-01 4.41527665e-01
5.41494608e-01 -6.18448138e-01 1.16149259e+00 6.53573751e-01
-5.71352579e-02 -1.01499295e+00 -1.12005103e+00 -5.78123569e-01
1.35443628e-01 1.36532381e-01 4.04042602e-01 5.15674531e-01
1.75254852e-01 7.06832111e-01 -1.86470330e-01 2.37589534e-02
3.31630677e-01 -5.69627732e-02 8.21717322e-01 -1.26504385e+00
-5.58115065e-01 6.52457587e-03 -8.00332069e-01 -3.06709617e-01
-5.93923271e-01 -7.11607456e-01 -6.70480013e-01 -1.18274879e+00
5.20798028e-01 -2.28182897e-01 1.57079305e-02 6.14775062e-01
-3.24283578e-02 2.31780559e-01 7.22554624e-02 6.40196353e-02
-2.43490756e-01 2.21938677e-02 1.28899443e+00 2.30856329e-01
-5.02452075e-01 -2.47577764e-02 -5.25731683e-01 5.76955140e-01
7.81015038e-01 -6.45609975e-01 -1.72154054e-01 3.79188985e-01
4.11632538e-01 -1.80221975e-01 -1.71238497e-01 -9.91030991e-01
1.20944627e-01 -3.12037647e-01 6.04132950e-01 -6.64066494e-01
2.69839644e-01 -1.03170943e+00 7.54675508e-01 1.20315552e+00
6.01379871e-01 1.85854763e-01 6.77904248e-01 2.64021784e-01
6.19207602e-03 -5.41776642e-02 9.89834845e-01 -1.07508659e-01
-5.53185821e-01 2.22804606e-01 -4.45237100e-01 -5.68798125e-01
1.43867147e+00 -4.40868884e-01 -1.31008565e-01 5.28642416e-01
-8.12768459e-01 -2.66959500e-02 9.05564308e-01 -5.72322346e-02
3.76110405e-01 -5.97005010e-01 -3.47913682e-01 1.30279407e-01
-2.05988362e-02 -3.55970323e-01 1.63299233e-01 8.69522989e-01
-1.48645484e+00 9.87015724e-01 -5.02883315e-01 -3.22729975e-01
-2.07077432e+00 7.33764648e-01 1.76119149e-01 -4.42970067e-01
-4.44139898e-01 5.10053754e-01 -2.11428329e-01 2.56109178e-01
-7.19729997e-03 2.25457206e-01 -1.74418092e-01 -2.05318585e-01
9.15597558e-01 2.63578236e-01 5.95596433e-01 -5.47725439e-01
-8.65269423e-01 4.05139476e-01 -6.50809705e-01 3.55946004e-01
1.60836351e+00 6.31479442e-01 -3.85156214e-01 -3.23209524e-01
1.03256798e+00 2.58268565e-01 -6.78068399e-01 6.55024230e-01
1.71328694e-01 -3.83906305e-01 2.12052986e-01 -1.03828752e+00
-3.41200918e-01 6.08565390e-01 8.95755947e-01 -5.28301477e-01
8.59730244e-01 -2.97921509e-01 6.07849956e-01 5.71152151e-01
7.83981740e-01 -6.47041202e-01 -3.69436800e-01 4.46591198e-01
4.55051959e-01 -8.84408474e-01 6.83207870e-01 -5.09587884e-01
-2.94268936e-01 1.32742786e+00 2.07118273e-01 1.03287725e-03
2.83877105e-01 3.62083197e-01 -3.30632538e-01 -4.88054544e-01
-6.52603388e-01 5.78185078e-03 -1.58082217e-01 5.06823003e-01
9.10843432e-01 9.27985758e-02 -9.82008696e-01 -4.45806347e-02
2.18242422e-01 1.17745839e-01 4.58385199e-01 1.29800403e+00
-7.81001925e-01 -2.40510988e+00 -5.77818751e-01 1.41046625e-02
-7.80411422e-01 -3.36099833e-01 -6.45973206e-01 9.98636484e-01
1.95801899e-01 7.86874294e-01 -4.89897102e-01 -1.23140894e-01
3.41441989e-01 -5.67678316e-03 7.56938100e-01 -8.40698361e-01
-1.07955062e+00 1.62986308e-01 6.12709261e-02 -5.89623809e-01
-2.64899671e-01 -2.57161170e-01 -1.52306044e+00 -9.12226319e-01
-4.21991646e-01 6.57925129e-01 1.11869860e+00 6.71817362e-01
8.22078824e-01 5.62527537e-01 3.47019136e-01 -9.36703563e-01
-1.33558959e-01 -3.74830693e-01 -3.85288179e-01 -4.69259143e-01
-1.83914736e-01 -6.92217588e-01 3.57336342e-01 -2.99177375e-02] | [4.909379482269287, 5.49327278137207] |
7c2e74ce-5c3b-45dc-be09-c4e4fc67cb59 | robustifying-multi-hop-qa-through-pseudo | 2107.03242 | null | https://arxiv.org/abs/2107.03242v1 | https://arxiv.org/pdf/2107.03242v1.pdf | Robustifying Multi-hop QA through Pseudo-Evidentiality Training | This paper studies the bias problem of multi-hop question answering models, of answering correctly without correct reasoning. One way to robustify these models is by supervising to not only answer right, but also with right reasoning chains. An existing direction is to annotate reasoning chains to train models, requiring expensive additional annotations. In contrast, we propose a new approach to learn evidentiality, deciding whether the answer prediction is supported by correct evidences, without such annotations. Instead, we compare counterfactual changes in answer confidence with and without evidence sentences, to generate "pseudo-evidentiality" annotations. We validate our proposed model on an original set and challenge set in HotpotQA, showing that our method is accurate and robust in multi-hop reasoning. | ['Dohyeon Lee', 'Sang-eun Han', 'Seung-won Hwang', 'Kyungjae Lee'] | 2021-07-07 | null | https://aclanthology.org/2021.acl-long.476 | https://aclanthology.org/2021.acl-long.476.pdf | acl-2021-5 | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 3.27533901e-01 1.03777933e+00 -2.48188347e-01 -6.64682209e-01
-1.13270223e+00 -8.21547210e-01 6.89907014e-01 3.80696505e-01
-4.17545408e-01 1.38916588e+00 4.08587426e-01 -6.99977636e-01
-2.63401657e-01 -9.58387792e-01 -1.13357341e+00 -1.04478896e-01
4.58535582e-01 9.29847181e-01 6.39139473e-01 -3.92666996e-01
2.45992824e-01 -3.82982790e-02 -1.23641908e+00 8.83436143e-01
1.06022704e+00 7.13840485e-01 -5.71855545e-01 9.53684688e-01
-1.52332053e-01 1.94790530e+00 -7.85557568e-01 -1.51439118e+00
7.40684345e-02 -7.29560018e-01 -1.54292965e+00 -5.09977937e-01
6.64393187e-01 -2.56654143e-01 1.91509664e-01 9.53913867e-01
1.51354924e-01 -1.38047889e-01 6.90058172e-01 -1.32484722e+00
-7.42417276e-01 1.10966289e+00 6.82433462e-03 3.55885923e-01
9.32219446e-01 2.47923091e-01 1.59907031e+00 -2.89173335e-01
7.14923739e-01 1.33637786e+00 8.11228812e-01 7.27947235e-01
-1.30026388e+00 -3.58646572e-01 1.74901307e-01 6.61526322e-01
-7.28497267e-01 -3.93237740e-01 5.96761107e-01 -2.82868981e-01
7.93456793e-01 6.98660970e-01 3.06329280e-01 1.30466557e+00
1.30485550e-01 8.00526679e-01 1.52084529e+00 -6.28337264e-01
3.84722054e-01 3.15072060e-01 4.13868964e-01 6.93722665e-01
3.64414364e-01 -1.31826929e-03 -5.96111417e-01 -4.34006572e-01
-1.22289918e-01 -5.09055614e-01 -2.29506418e-01 -1.29713133e-01
-1.08495700e+00 1.06225371e+00 3.01482290e-01 1.30053654e-01
-4.90856081e-01 3.69400114e-01 1.13311268e-01 5.91026723e-01
5.04501104e-01 8.26367140e-01 -8.77243340e-01 5.89142293e-02
-8.69596303e-01 6.80605173e-01 1.24785757e+00 4.05162096e-01
4.79665369e-01 -6.09056830e-01 -7.06040919e-01 3.00494522e-01
4.87688065e-01 3.76024693e-01 3.43678802e-01 -1.60992098e+00
6.02694273e-01 5.99824071e-01 5.22005439e-01 -7.73508132e-01
-3.55787426e-01 -2.89483100e-01 -3.16221088e-01 1.60831877e-03
7.30662167e-01 -2.36990631e-01 -4.30488288e-01 1.93526721e+00
5.41433096e-01 6.65589049e-02 3.78586084e-01 7.71715343e-01
5.41081727e-01 2.61509657e-01 1.85569003e-01 -2.36271724e-01
1.42807281e+00 -9.54394579e-01 -8.02751005e-01 -7.58303329e-02
8.41253340e-01 -3.59004289e-01 1.13376498e+00 5.04740179e-01
-1.13092315e+00 -2.05807999e-01 -9.83089089e-01 -2.02446163e-01
-8.99744779e-02 -2.78607279e-01 5.31078458e-01 4.62228626e-01
-9.03792918e-01 5.74148178e-01 -1.30017966e-01 1.22284733e-01
1.94555894e-01 7.99179226e-02 -7.94692412e-02 -8.63159075e-02
-1.74826086e+00 1.31669426e+00 5.55578470e-01 -1.27000764e-01
-6.50727332e-01 -7.96003163e-01 -6.69673264e-01 8.30467716e-02
7.11241066e-01 -1.05443513e+00 1.81066418e+00 -8.66447926e-01
-1.34214151e+00 8.94406199e-01 -3.50284696e-01 -1.17672706e+00
1.06410265e+00 -3.72478217e-01 -3.19408536e-01 2.28243709e-01
3.07215840e-01 6.18115425e-01 6.23330593e-01 -1.29339349e+00
-6.02593005e-01 -2.99086481e-01 8.31711292e-01 -1.06325626e-01
9.10853297e-02 -2.12090805e-01 4.30440456e-01 -8.15554336e-02
-6.94423839e-02 -6.99706495e-01 -1.43960640e-01 -1.94752038e-01
-6.80980086e-01 -5.25798619e-01 5.08833230e-01 -7.37671614e-01
1.05151546e+00 -1.35717237e+00 -1.55872792e-01 1.00787304e-01
2.27507949e-01 -3.55271064e-02 7.31010586e-02 7.24336058e-02
-3.83393280e-02 3.43544811e-01 -4.85829473e-01 6.13515936e-02
3.98870558e-01 6.64201677e-01 -8.85030985e-01 -1.24034546e-01
3.95315140e-01 1.16491675e+00 -1.19227755e+00 -7.79525638e-01
-1.66725889e-01 -2.66589105e-01 -8.63770247e-01 2.32963562e-01
-1.00221419e+00 5.37750840e-01 -1.68238789e-01 2.14365393e-01
3.88920903e-01 -5.33898652e-01 4.73484963e-01 -6.89830557e-02
3.23723882e-01 7.57443666e-01 -1.06767619e+00 1.24787533e+00
-6.62674069e-01 3.73184860e-01 -3.83810341e-01 -1.06236851e+00
5.53379238e-01 5.20386279e-01 -1.65957287e-01 -7.38990843e-01
-2.18222126e-01 4.55667526e-01 6.97493851e-02 -9.98262882e-01
3.35435629e-01 -6.44093513e-01 -1.40795603e-01 6.58892572e-01
1.81258813e-01 -4.18203324e-01 4.52688664e-01 3.94459635e-01
1.25687706e+00 2.45037958e-01 3.48605424e-01 -6.63835257e-02
7.85552859e-01 3.11349928e-01 4.57881510e-01 1.04245424e+00
-4.70338091e-02 4.54433620e-01 9.08251107e-01 -4.90823239e-01
-7.69068420e-01 -1.12000680e+00 1.56502262e-01 9.83110428e-01
-7.46595860e-02 -1.98750198e-01 -5.93720257e-01 -1.53143156e+00
-8.80483836e-02 1.64018798e+00 -8.19413066e-01 -1.17797524e-01
-6.69206142e-01 -4.46757436e-01 7.96415865e-01 2.77492404e-01
6.22122705e-01 -9.85168040e-01 -6.98802590e-01 1.30746767e-01
-9.88266885e-01 -9.20619309e-01 2.11401895e-01 3.37865166e-02
-8.34900796e-01 -1.56717002e+00 -2.71551430e-01 -4.44294810e-02
2.79504508e-01 -4.34156269e-01 1.85168540e+00 4.97869909e-01
3.39113951e-01 3.71204585e-01 -3.47050846e-01 -5.39542675e-01
-7.49134421e-01 -1.00694105e-01 -3.94277751e-01 -1.60503492e-01
2.90663838e-01 -3.30572426e-01 -5.14730036e-01 1.65898025e-01
-9.10975039e-01 -1.16115518e-01 4.24141735e-01 8.38844538e-01
3.54137033e-01 -3.94506902e-01 8.64838421e-01 -1.57867467e+00
8.06889176e-01 -6.84163690e-01 -2.27876008e-01 8.48679662e-01
-7.96234846e-01 5.34608603e-01 7.59498417e-01 8.31658207e-03
-1.60466075e+00 -4.41096783e-01 -4.68644291e-01 1.77738458e-01
-1.56151623e-01 3.77653360e-01 2.91825049e-02 5.12486398e-01
1.12186491e+00 -2.62319386e-01 -4.88084644e-01 -2.74415053e-02
8.77821386e-01 2.81590492e-01 4.65332866e-01 -1.00115168e+00
7.90178061e-01 5.30632794e-01 1.31748393e-02 -6.41169623e-02
-1.69619477e+00 -4.28179055e-02 -4.07474875e-01 -5.02236979e-03
8.27950120e-01 -4.68884826e-01 -7.70889699e-01 -3.39393675e-01
-1.46770167e+00 -4.11249459e-01 -6.59933209e-01 3.38308930e-01
-6.07268691e-01 3.88908833e-01 -5.26180267e-01 -1.00569260e+00
-3.07399511e-01 -5.39605975e-01 8.17557156e-01 1.05586082e-01
-7.92730808e-01 -1.12333024e+00 5.22499800e-01 1.01194775e+00
1.29812464e-01 2.01419711e-01 1.18814623e+00 -1.06102431e+00
-4.38734561e-01 -7.26305097e-02 -6.01406489e-03 3.29436183e-01
-3.96450788e-01 -4.76201773e-02 -1.16951299e+00 3.38199198e-01
1.34242743e-01 -7.56896794e-01 7.30439603e-01 6.08411501e-04
1.00889397e+00 -9.02604818e-01 1.28207551e-02 -4.18027014e-01
1.11469674e+00 -2.18233451e-01 6.69795096e-01 1.40185744e-01
4.22286196e-03 9.62070167e-01 7.42658257e-01 2.99616102e-02
7.21523404e-01 4.36637044e-01 3.64596933e-01 3.05659562e-01
-8.86293799e-02 -5.90380251e-01 6.97746798e-02 2.90834159e-01
-8.62688571e-02 -4.22821850e-01 -7.00351775e-01 6.21023417e-01
-2.16887426e+00 -1.41570079e+00 -6.09098613e-01 1.75106275e+00
1.21940386e+00 2.99942046e-01 -1.09125637e-01 2.84841537e-01
3.39844853e-01 -7.23608583e-02 -2.98154831e-01 -5.65808296e-01
-2.29300007e-01 5.29789090e-01 -7.72191957e-02 8.79731536e-01
-6.34570003e-01 5.47946990e-01 6.57048750e+00 5.29579639e-01
-4.12827373e-01 4.30172235e-01 7.24113226e-01 9.64066982e-02
-1.07260692e+00 4.92882818e-01 -3.40434551e-01 3.56660396e-01
1.08370388e+00 1.10995352e-01 2.11031977e-02 7.16960728e-01
-1.48848757e-01 -3.47354054e-01 -1.41292083e+00 1.39298812e-01
6.38177916e-02 -1.69143772e+00 1.83759436e-01 -6.12963855e-01
8.48089159e-01 -6.35183394e-01 -1.96361974e-01 7.34626770e-01
9.32746232e-01 -9.23341036e-01 1.00294399e+00 7.13138044e-01
1.58175930e-01 -4.06264216e-01 1.14395726e+00 7.63770819e-01
-5.28983653e-01 -2.45380431e-01 -5.43726236e-02 -4.04859722e-01
2.80775517e-01 7.34304488e-01 -1.02978563e+00 8.03949714e-01
5.14150023e-01 -2.98583023e-02 -7.03816235e-01 7.29168534e-01
-1.21442258e+00 9.18597043e-01 -5.04739359e-02 -2.48723552e-01
-3.01521961e-02 1.80021554e-01 2.41990536e-01 9.92000222e-01
8.93756002e-02 2.23849729e-01 -3.39862108e-01 1.05931520e+00
-3.63698184e-01 -3.31582248e-01 -5.06274581e-01 3.90090287e-01
3.19716066e-01 9.98903990e-01 -2.53717422e-01 -7.46138930e-01
-1.64120615e-01 8.60688865e-01 5.70080519e-01 2.23026648e-01
-9.35349226e-01 4.10279483e-02 -1.93033665e-01 -3.36781554e-02
-1.12207077e-01 4.08574134e-01 -4.14508641e-01 -1.12619472e+00
2.29880944e-01 -9.72406149e-01 9.47385788e-01 -1.18563259e+00
-1.53757572e+00 2.37841591e-01 1.92151904e-01 -5.50277889e-01
-6.42726779e-01 -5.43567181e-01 -7.81005144e-01 6.25130236e-01
-1.80677032e+00 -1.06849539e+00 1.98069587e-03 3.72785479e-01
3.16780090e-01 4.59119409e-01 8.07451546e-01 5.67304567e-02
-1.69683658e-02 4.47205573e-01 -6.61133289e-01 -9.07205865e-02
7.24562824e-01 -1.67251289e+00 -1.58595853e-02 8.39464188e-01
4.89575446e-01 6.44677758e-01 1.21182203e+00 -4.51989174e-01
-8.14801514e-01 -8.71558189e-01 1.41603732e+00 -1.13100553e+00
7.68131256e-01 1.41639695e-01 -1.01557338e+00 9.55211163e-01
5.46936154e-01 -2.36745551e-01 7.19376624e-01 4.37835455e-01
-7.68291891e-01 3.50408405e-02 -1.43684518e+00 6.59365416e-01
9.10487354e-01 -4.84581202e-01 -1.49570644e+00 6.04696631e-01
1.06663346e+00 -2.02420399e-01 -5.23527682e-01 6.08904958e-01
3.29282433e-01 -1.24942660e+00 7.75246739e-01 -1.04638875e+00
8.82788360e-01 -4.73820657e-01 -2.31851161e-01 -1.16596770e+00
1.53093681e-01 -3.41659188e-01 -5.26428759e-01 1.01973879e+00
1.04040229e+00 -7.77286530e-01 7.60641277e-01 8.36777925e-01
4.59501222e-02 -7.54483879e-01 -1.22223783e+00 -3.89467776e-01
2.12098196e-01 -5.30860126e-01 6.67240143e-01 9.57623065e-01
5.24360547e-03 6.37192309e-01 -2.27716580e-01 1.66557804e-01
4.68668908e-01 3.92008424e-01 7.00074255e-01 -1.25794208e+00
-6.88843548e-01 -1.29524440e-01 2.03007862e-01 -8.49101543e-01
4.47121054e-01 -8.31571937e-01 4.28358838e-02 -1.74884951e+00
2.75696039e-01 -3.35541219e-01 1.85715705e-02 4.56234097e-01
-4.89831269e-01 2.16117188e-01 4.89067808e-02 -1.97948113e-01
-1.14564908e+00 2.83018619e-01 1.11531281e+00 -2.66078115e-01
4.47303712e-01 1.52314901e-01 -9.14523065e-01 9.32128310e-01
7.47957408e-01 -5.27707279e-01 -6.11785293e-01 -3.09166819e-01
1.11627197e+00 1.45381793e-01 8.48730326e-01 -7.17750371e-01
1.81897238e-01 -2.01988339e-01 2.22474113e-01 -3.58016133e-01
2.02080399e-01 -6.49542928e-01 2.97596250e-02 7.72655070e-01
-9.48208094e-01 1.55656859e-01 -1.35429725e-01 6.88625634e-01
-2.49952957e-01 -7.77692497e-01 3.09024692e-01 -3.53185147e-01
-4.77871478e-01 -3.79838496e-01 -9.08827484e-02 6.60372496e-01
8.07550669e-01 2.66144067e-01 -8.50958526e-01 -5.57454050e-01
-7.97919989e-01 3.62238079e-01 4.02533598e-02 5.31831309e-02
5.26325285e-01 -1.01096761e+00 -9.68903482e-01 -5.83367229e-01
2.00626120e-01 -1.15429223e-01 3.73233706e-01 9.07399476e-01
-4.00883555e-01 4.47658360e-01 3.44378471e-01 -3.23356867e-01
-9.53391254e-01 5.53114235e-01 5.12747943e-01 -9.63121772e-01
-8.01345333e-02 8.55547249e-01 -2.92252004e-01 -8.41220438e-01
-1.68389931e-01 -3.89784068e-01 -2.93437928e-01 5.81353009e-02
3.11342925e-01 2.66732275e-01 2.25847930e-01 1.56333432e-01
-2.48126015e-01 1.19029589e-01 6.24902025e-02 -4.56612974e-01
8.19876552e-01 1.04456797e-01 -2.44410202e-01 5.83913863e-01
5.41678309e-01 2.96707004e-01 -7.43540347e-01 -4.07383628e-02
4.97739941e-01 -3.59155029e-01 -6.39867425e-01 -1.44817662e+00
-2.36834601e-01 8.47999811e-01 9.47073027e-02 5.43875694e-01
5.18108070e-01 1.91281900e-01 5.72749913e-01 8.21651161e-01
3.84011358e-01 -1.05290115e+00 3.41601491e-01 4.18340683e-01
1.17903161e+00 -1.38188815e+00 -1.03170566e-01 -9.05493721e-02
-8.19798648e-01 8.43522608e-01 8.03053617e-01 6.10956922e-02
6.04549237e-02 -1.30654275e-01 1.99280336e-01 -4.64389175e-01
-1.16336560e+00 1.09181598e-01 1.20995842e-01 3.56461257e-01
5.28578401e-01 9.95489284e-02 -4.87303734e-01 5.76130152e-01
-3.67887408e-01 -6.10258207e-02 5.39528430e-01 7.14114428e-01
-1.75035715e-01 -9.63433146e-01 -4.47162539e-01 4.68089581e-01
-6.44496620e-01 -1.27708778e-01 -8.03948879e-01 7.24371910e-01
1.79315343e-01 1.49338591e+00 -3.18797231e-01 -1.42416731e-01
3.43906879e-01 5.12318909e-01 6.23206317e-01 -4.97357339e-01
-6.55881464e-01 -1.04077053e+00 7.15784848e-01 -2.99606413e-01
-7.62390852e-01 -1.93761095e-01 -1.17647362e+00 -2.45437145e-01
-2.99203902e-01 7.61020422e-01 2.31388912e-01 1.41296446e+00
1.60730973e-01 6.60822928e-01 2.59761572e-01 1.17517620e-01
-1.08735013e+00 -9.65938568e-01 1.65769920e-01 7.00978398e-01
2.60963500e-01 -4.83702153e-01 -4.43734348e-01 -5.73893525e-02] | [10.819925308227539, 7.871837615966797] |
b22b3031-d287-4ef1-966b-b114d33666c6 | open-access-dataset-for-electromyography | 2201.01051 | null | https://arxiv.org/abs/2201.01051v2 | https://arxiv.org/pdf/2201.01051v2.pdf | Open Access Dataset for Electromyography based Multi-code Biometric Authentication | Recently, surface electromyogram (EMG) has been proposed as a novel biometric trait for addressing some key limitations of current biometrics, such as spoofing and liveness. The EMG signals possess a unique characteristic: they are inherently different for individuals (biometrics), and they can be customized to realize multi-length codes or passwords (for example, by performing different gestures). However, current EMG-based biometric research has two critical limitations: 1) a small subject pool, compared to other more established biometric traits, and 2) limited to single-session or single-day data sets. In this study, forearm and wrist EMG data were collected from 43 participants over three different days with long separation while they performed static hand and wrist gestures. The multi-day biometric authentication resulted in a median EER of 0.017 for the forearm setup and 0.025 for the wrist setup, comparable to well-established biometric traits suggesting consistent performance over multiple days. The presented large-sample multi-day data set and findings could facilitate further research on EMG-based biometrics and other gesture recognition-based applications. | ['Ning Jiang', 'Jiayuan He', 'Ashirbad Pradhan'] | 2022-01-04 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 4.12297636e-01 -4.34275657e-01 -4.24068838e-01 -2.08723575e-01
-7.16235042e-01 -5.59086084e-01 2.58482277e-01 -2.57980168e-01
-7.05884099e-01 6.99469030e-01 -6.93214387e-02 9.42822099e-02
-2.93205619e-01 -1.89558402e-01 -1.31916538e-01 -8.78448129e-01
-2.28894934e-01 -1.56609938e-02 -2.32204244e-01 1.76923722e-01
3.10243607e-01 7.19174922e-01 -1.46721506e+00 -3.11840266e-01
4.94565308e-01 9.49027896e-01 -7.24973902e-02 5.65672874e-01
2.81689167e-01 -7.51395047e-01 -9.15326178e-01 -5.17174542e-01
2.02447593e-01 -4.14159328e-01 -2.88308740e-01 -3.42932522e-01
3.73035312e-01 -4.64654565e-01 -1.94672063e-01 7.36702740e-01
1.10628366e+00 7.96180870e-03 4.80133533e-01 -9.36989427e-01
-6.01770878e-01 2.18432426e-01 -8.21240067e-01 -1.83646560e-01
8.84020209e-01 3.86538059e-01 5.08867919e-01 -5.24888039e-01
3.59526455e-01 8.88280392e-01 9.38896358e-01 1.27940083e+00
-1.33424950e+00 -9.62729931e-01 -4.23954368e-01 2.60812667e-04
-1.59704280e+00 -5.53748608e-01 6.54713154e-01 -3.66239786e-01
5.17412961e-01 4.36122358e-01 8.42165232e-01 1.61999393e+00
3.08055401e-01 3.15835446e-01 1.56176412e+00 -9.89173949e-02
2.13866085e-02 -8.56194943e-02 1.58245433e-02 7.05131963e-02
6.64581120e-01 2.07166210e-01 -7.84833968e-01 -3.34765643e-01
8.80388498e-01 2.14400575e-01 -3.77421677e-01 1.81849763e-01
-1.37496495e+00 2.44682491e-01 -2.11490199e-01 6.25211477e-01
-7.63181746e-01 -1.47205768e-02 3.03390771e-01 3.34776312e-01
-1.75133958e-01 2.78770149e-01 -3.12743932e-01 -1.05703402e+00
-7.84362376e-01 -1.48972899e-01 9.84241068e-01 5.31072021e-01
-5.22647128e-02 6.09256327e-02 -9.36435089e-02 9.25740361e-01
2.96087563e-01 1.04579294e+00 7.03888595e-01 -4.26135540e-01
3.85367215e-01 1.89131692e-01 7.05387071e-02 -8.88736129e-01
-5.87674081e-01 1.08996645e-01 -9.51613307e-01 4.36757654e-02
9.35923576e-01 -2.52785087e-01 -5.47731757e-01 1.67737937e+00
5.24771549e-02 2.61768699e-01 -2.56420851e-01 1.01703334e+00
8.24682415e-01 -3.34552765e-01 1.17661193e-01 -3.49003702e-01
1.64706576e+00 2.79202051e-02 -8.91507626e-01 1.02457449e-01
-1.16882324e-01 -8.03939998e-01 1.27854300e+00 7.49475598e-01
-7.55869806e-01 -3.78508717e-01 -9.17098463e-01 6.61477745e-01
-1.24081671e-01 2.94600099e-01 5.73905706e-01 1.75966334e+00
-5.64513445e-01 6.72633529e-01 -1.03965807e+00 -7.52449751e-01
2.21648365e-01 7.30013013e-01 -7.26959050e-01 5.03424346e-01
-9.41751838e-01 6.10612929e-01 -3.06108654e-01 3.43000859e-01
-2.55320340e-01 -8.01459029e-02 -5.46100438e-01 -3.71239305e-01
-7.83377662e-02 -8.46028235e-03 4.92988050e-01 -1.63978145e-01
-2.07172155e+00 1.02437818e+00 -3.48555982e-01 1.07650541e-01
4.07814115e-01 -4.99506801e-01 -7.86191583e-01 1.34111091e-01
-2.77196020e-01 2.79520024e-02 9.39846277e-01 -7.89205730e-01
1.10088199e-01 -9.17260706e-01 -4.08173382e-01 -1.14766471e-01
-6.31720781e-01 2.54960209e-01 -3.08490312e-03 -8.78634274e-01
4.29541349e-01 -1.17176175e+00 4.33592469e-01 -1.98219195e-01
-3.02371711e-01 2.21782532e-02 4.88666296e-01 -9.63594794e-01
1.25399697e+00 -2.20889592e+00 1.30264595e-01 3.43299776e-01
-3.62832509e-02 4.47901279e-01 -1.44094452e-01 5.69128215e-01
1.50813684e-01 2.72633076e-01 2.36697402e-02 -3.49437557e-02
-2.06612810e-01 4.18787822e-03 2.08573788e-01 9.39528406e-01
-2.05172628e-01 8.41474295e-01 -7.07525015e-01 -2.00858295e-01
1.92727774e-01 4.89437103e-01 2.38602310e-01 5.87749407e-02
9.55550194e-01 9.51588273e-01 -2.57171005e-01 1.13598263e+00
7.86146283e-01 4.33513463e-01 2.85474956e-01 -3.23680133e-01
1.80287823e-01 2.86992211e-02 -1.35236537e+00 1.64263070e+00
-2.32029513e-01 5.97537875e-01 1.00792617e-01 -6.47603750e-01
1.11812639e+00 6.45515978e-01 7.92778194e-01 -2.38243476e-01
3.43950629e-01 3.83526653e-01 2.93935686e-01 -8.24331403e-01
2.05483079e-01 -2.25098893e-01 3.13541293e-02 8.02441359e-01
-9.74798128e-02 1.37538224e-01 -1.59748614e-01 -6.03144586e-01
7.58116186e-01 1.25421137e-01 2.22429499e-01 -7.95536190e-02
5.56943357e-01 -9.00327444e-01 4.42847192e-01 7.23240495e-01
-4.86955941e-01 5.87354064e-01 -2.62865555e-02 2.42450442e-02
-2.65194714e-01 -9.95178580e-01 -5.16246915e-01 5.36071181e-01
3.25045347e-01 -1.83942392e-01 -6.92466319e-01 -4.54706639e-01
4.04236645e-01 -1.16936080e-01 -4.73149091e-01 -4.35185209e-02
-4.82899368e-01 -6.20562792e-01 1.37618041e+00 5.73099077e-01
3.87321651e-01 -9.23478067e-01 -6.89674139e-01 2.14067355e-01
-2.55121589e-01 -1.00527048e+00 -4.51479316e-01 -2.81300873e-01
-1.00070798e+00 -1.20988607e+00 -1.31064677e+00 -4.04866457e-01
4.00891215e-01 1.10040814e-01 2.09450439e-01 6.22523502e-02
-2.98883498e-01 7.85723507e-01 -3.70055765e-01 -2.73821622e-01
-5.79145923e-02 8.09291676e-02 9.23381448e-01 2.43603051e-01
8.20771158e-01 -7.26665258e-01 -7.65713215e-01 8.73186767e-01
-4.52082217e-01 -7.32116938e-01 5.81360638e-01 1.04842019e+00
2.48397157e-01 -4.68487859e-01 8.09214830e-01 -2.18180344e-01
1.08684325e+00 -9.54574272e-02 1.52485520e-01 4.05242026e-01
-7.80628264e-01 -3.35534781e-01 5.52343242e-02 -1.02702594e+00
-7.47846365e-01 -3.35507184e-01 -1.59131005e-01 2.62717754e-01
-4.72345114e-01 2.77513295e-01 -2.64469922e-01 -6.63512766e-01
5.46959817e-01 3.72381032e-01 5.98263562e-01 -6.94962859e-01
-6.28882870e-02 1.52135456e+00 6.23053849e-01 -9.44466591e-01
7.29200542e-01 2.52497196e-01 -4.83210236e-02 -1.14677095e+00
4.07081485e-01 -5.48805654e-01 -8.70591700e-01 -5.38545132e-01
5.79935789e-01 -4.01183307e-01 -1.58952141e+00 1.31300974e+00
-8.34910154e-01 -8.52191299e-02 6.80613704e-03 1.11144674e+00
-2.68908888e-01 8.18257034e-01 -7.49043643e-01 -1.40204775e+00
-6.64648771e-01 -8.86000872e-01 1.10399759e+00 4.58925068e-01
-6.57039762e-01 -4.91808295e-01 -1.66825324e-01 3.36460203e-01
5.42912722e-01 3.86165708e-01 2.21265391e-01 -3.96648794e-01
3.75554413e-01 -8.40012610e-01 1.20733403e-01 2.15209723e-01
8.67453754e-01 -2.21432939e-01 -9.98740494e-01 -7.48666465e-01
2.43287057e-01 -3.82339120e-01 7.45909512e-02 1.51579633e-01
7.66707361e-01 -2.41148677e-02 -2.22276926e-01 4.44765538e-01
9.29368794e-01 6.62630737e-01 8.59191716e-01 1.57036796e-01
4.29683745e-01 5.64744294e-01 5.28258085e-01 5.13335764e-01
-1.25725552e-01 8.51905048e-01 -1.41219899e-01 2.37484142e-01
3.11755352e-02 3.78319472e-02 3.68888885e-01 6.83621466e-01
-9.51379001e-01 2.43562564e-01 -5.88573635e-01 1.98725268e-01
-1.27342641e+00 -9.84110236e-01 -1.55114830e-01 2.75955343e+00
7.76109993e-01 -4.02631044e-01 6.22221291e-01 5.57375550e-01
8.70345414e-01 -1.40962288e-01 -7.45432138e-01 -1.43154725e-01
-3.11837703e-01 5.91945708e-01 5.95461249e-01 -1.38768464e-01
-7.20001101e-01 3.16310406e-01 6.58083916e+00 3.92609686e-01
-1.53670895e+00 6.10832013e-02 -9.65614840e-02 -1.45007551e-01
1.52896106e-01 -4.01634008e-01 -7.33513653e-01 5.96050024e-01
7.37771630e-01 -1.36384312e-02 5.58942795e-01 2.52020836e-01
1.66787118e-01 -2.05085456e-01 -9.43739772e-01 1.48675489e+00
1.72188878e-01 -5.19532740e-01 -4.62331146e-01 5.39439440e-01
2.01997817e-01 -3.65139604e-01 1.73292413e-01 -2.69315094e-01
-7.01754093e-01 -9.00647104e-01 1.94325835e-01 5.62744737e-01
1.52207530e+00 -2.15963930e-01 7.95363128e-01 -1.85594223e-02
-1.11379647e+00 1.43671900e-01 1.54506322e-02 -3.56815420e-02
2.78695911e-01 1.88291930e-02 -3.16379607e-01 5.25654316e-01
6.44549608e-01 4.86759841e-01 -2.58405179e-01 9.18295622e-01
-1.19388223e-01 7.16880918e-01 -6.61500871e-01 -4.59186286e-01
-4.50659364e-01 -2.81450182e-01 6.77514613e-01 1.00313115e+00
5.25058150e-01 7.34220305e-03 -4.92207021e-01 5.48037469e-01
3.67223442e-01 9.62253585e-02 -4.32462335e-01 -3.98996979e-01
7.62931406e-01 1.10149896e+00 -5.67192733e-01 2.74030238e-01
-4.22980338e-01 1.11086309e+00 -5.75056553e-01 5.40294468e-01
-5.27520180e-01 -7.66227484e-01 7.17367887e-01 6.41145036e-02
-2.43239090e-01 -5.67400217e-01 -7.07524776e-01 -1.33535790e+00
3.85343105e-01 -1.07524836e+00 2.43715703e-01 1.50369983e-02
-1.37589335e+00 2.01189041e-01 -1.72580913e-01 -1.56299818e+00
-1.83710530e-01 -6.72048450e-01 -3.96842390e-01 1.25513959e+00
-5.90366900e-01 -1.04091012e+00 -4.77650702e-01 6.34745002e-01
3.46968509e-02 -3.51747483e-01 1.44539678e+00 4.67077374e-01
-4.75648701e-01 1.31046844e+00 2.15956923e-02 2.46259019e-01
9.35428619e-01 -8.85240853e-01 3.95007879e-01 5.99594295e-01
2.84572393e-02 1.59540379e+00 2.64825732e-01 -7.98576772e-01
-1.97801113e+00 1.97259903e-01 4.39063758e-01 -3.05133075e-01
4.02909428e-01 -1.00438774e-01 -6.65587842e-01 2.48691335e-01
-3.28496397e-01 -4.05918777e-01 1.35076344e+00 3.81696492e-01
-1.58795252e-01 -2.06204593e-01 -1.57219803e+00 5.35922885e-01
1.14269888e+00 -7.81939983e-01 -4.09857452e-01 -3.31672728e-01
-3.62130314e-01 -4.46296722e-01 -1.64710999e+00 4.14514095e-01
1.98993456e+00 -4.63182271e-01 8.08481991e-01 -4.40250039e-01
-4.35495198e-01 -5.95207289e-02 5.69960475e-02 -8.47939312e-01
-3.27954590e-02 -1.06864703e+00 -1.75640672e-01 1.54067516e+00
4.07218449e-02 -1.15954709e+00 8.77307177e-01 1.04389489e+00
6.46490037e-01 -5.82659662e-01 -1.24319232e+00 -1.30880892e+00
-3.79311323e-01 -4.44253623e-01 7.79168069e-01 6.99253321e-01
5.63636661e-01 -3.12415689e-01 -8.01794171e-01 -2.78695315e-01
7.21353352e-01 -2.06540421e-01 1.11601484e+00 -1.37440681e+00
-3.70495081e-01 -4.29411322e-01 -8.22755635e-01 -9.34367776e-01
-2.39401132e-01 -5.63836336e-01 -2.96847492e-01 -8.62169325e-01
4.13491316e-02 -4.40514147e-01 -4.26978588e-01 5.29072285e-01
-1.48171991e-01 4.44674134e-01 1.81228276e-02 4.43499386e-01
5.23593903e-01 7.70714581e-02 1.08451593e+00 -8.18547234e-02
-4.86465394e-01 4.50064719e-01 -6.29431725e-01 1.69047609e-01
6.13850296e-01 -2.53443658e-01 -3.72299179e-03 6.33926168e-02
-3.81110072e-01 2.21268430e-01 2.88376033e-01 -9.23711836e-01
1.09545171e-01 -7.63261393e-02 2.75677413e-01 -1.84213072e-01
4.45610106e-01 -7.32912004e-01 4.54967827e-01 7.60731459e-01
1.79921001e-01 -1.79323286e-01 7.58351460e-02 4.67434317e-01
4.63448130e-02 -5.48211560e-02 3.05714518e-01 2.38223433e-01
-3.55393082e-01 1.43707976e-01 -3.71355027e-01 -3.14025074e-01
8.17883253e-01 -1.04612887e+00 -8.18745941e-02 -3.81967515e-01
-9.34233546e-01 -4.73006994e-01 4.39684957e-01 5.63212454e-01
6.95749521e-01 -1.33452272e+00 -4.54820126e-01 5.21036088e-01
1.58655062e-01 -7.76137650e-01 1.93366811e-01 1.39965093e+00
-1.72523931e-01 1.61844909e-01 -6.75718367e-01 -7.73448348e-01
-1.71783257e+00 -2.12243512e-01 3.82927507e-02 4.11942184e-01
-6.57690525e-01 4.02780890e-01 -7.51857638e-01 -8.81259143e-02
3.23388606e-01 1.77501485e-01 -1.35587990e-01 1.59741849e-01
5.71788132e-01 7.44389057e-01 7.02703446e-02 -6.73138380e-01
-7.15709746e-01 1.00008976e+00 3.90678138e-01 -3.83513927e-01
1.14809024e+00 -1.60546347e-01 3.42178456e-02 5.94727874e-01
7.42010176e-01 1.78443506e-01 -8.06856632e-01 2.26480346e-02
4.19571847e-02 -8.94573212e-01 -7.44387984e-01 -7.08425760e-01
-9.89525616e-01 7.38334715e-01 1.05730021e+00 -2.03861389e-02
9.69256997e-01 -4.47780639e-01 9.60423112e-01 2.51849443e-01
9.94217157e-01 -1.01811194e+00 -2.79446989e-01 -8.76275301e-02
8.93548667e-01 -8.48939836e-01 2.20016353e-02 -1.33826494e-01
-3.36897969e-01 1.30599630e+00 2.99057484e-01 1.77614093e-01
6.33163452e-01 1.97207332e-01 2.19869435e-01 -2.34969743e-02
4.23279554e-01 1.28029197e-01 6.11012995e-01 1.16016233e+00
9.15210783e-01 4.23321247e-01 -1.42513394e+00 9.88825858e-01
-2.51772732e-01 2.71438807e-01 2.00755969e-01 9.85824049e-01
2.98735797e-01 -1.49326503e+00 -5.85375667e-01 6.63318992e-01
-8.37493658e-01 3.55976313e-01 -5.07969141e-01 6.94113195e-01
-2.46806398e-01 1.16479039e+00 -5.50458848e-01 -9.21242654e-01
4.13935333e-01 3.29223156e-01 8.60430002e-01 -3.09636086e-01
-6.98207498e-01 2.98689548e-02 -8.31108764e-02 -5.60223043e-01
-6.26509547e-01 -1.07864356e+00 -8.44911933e-01 -3.58142257e-01
-6.39146149e-01 -4.39066961e-02 1.08140600e+00 8.57085228e-01
3.70587856e-01 -1.92758799e-01 5.18055856e-01 -8.89094412e-01
-7.15110004e-01 -1.44775712e+00 -1.18390667e+00 5.96500874e-01
-4.88559417e-02 -8.11391354e-01 -1.51975229e-01 -1.52061731e-01] | [13.90792465209961, 1.6584053039550781] |
5b3ab7a7-b5a2-45f9-99f4-b4ce802ca030 | point-pc-point-cloud-completion-guided-by | 2305.17770 | null | https://arxiv.org/abs/2305.17770v1 | https://arxiv.org/pdf/2305.17770v1.pdf | Point-PC: Point Cloud Completion Guided by Prior Knowledge via Causal Inference | Point cloud completion aims to recover raw point clouds captured by scanners from partial observations caused by occlusion and limited view angles. Many approaches utilize a partial-complete paradigm in which missing parts are directly predicted by a global feature learned from partial inputs. This makes it hard to recover details because the global feature is unlikely to capture the full details of all missing parts. In this paper, we propose a novel approach to point cloud completion called Point-PC, which uses a memory network to retrieve shape priors and designs an effective causal inference model to choose missing shape information as additional geometric information to aid point cloud completion. Specifically, we propose a memory operating mechanism where the complete shape features and the corresponding shapes are stored in the form of ``key-value'' pairs. To retrieve similar shapes from the partial input, we also apply a contrastive learning-based pre-training scheme to transfer features of incomplete shapes into the domain of complete shape features. Moreover, we use backdoor adjustment to get rid of the confounder, which is a part of the shape prior that has the same semantic structure as the partial input. Experimental results on the ShapeNet-55, PCN, and KITTI datasets demonstrate that Point-PC performs favorably against the state-of-the-art methods. | ['AnAn Liu', 'Nicu Sebe', 'Bruno Lepri', 'Weijie Wang', 'Ruidong Chen', 'Chuanqi Jiao', 'Weizhi Nie'] | 2023-05-28 | null | null | null | null | ['point-cloud-completion', 'causal-inference', 'causal-inference'] | ['computer-vision', 'knowledge-base', 'miscellaneous'] | [ 2.19600108e-02 2.09751531e-01 -1.22949362e-01 -4.78395790e-01
-6.66691363e-01 -4.99891728e-01 6.09084308e-01 -2.40127981e-01
9.75732654e-02 4.12243336e-01 1.30493060e-01 6.62509818e-03
-2.44106844e-01 -1.16045105e+00 -1.31319797e+00 -6.28512442e-01
3.53660107e-01 9.34964478e-01 1.73307344e-01 -1.48400040e-02
3.20512503e-01 8.24349284e-01 -1.67230463e+00 3.12766641e-01
6.83366895e-01 8.26490641e-01 6.73273981e-01 5.07714367e-03
-6.53175116e-01 5.57145663e-02 -1.51894495e-01 -1.91439599e-01
3.93904269e-01 3.47561866e-01 -4.43939269e-01 1.21323854e-01
6.27622366e-01 -5.65822780e-01 -4.47679013e-01 8.58387589e-01
1.73883185e-01 -4.91043888e-02 5.67208230e-01 -1.11628187e+00
-6.94476128e-01 1.02824293e-01 -5.47383666e-01 -5.41279197e-01
2.45624766e-01 2.47911662e-01 7.32109725e-01 -1.46266317e+00
7.58879304e-01 1.53903604e+00 5.89142382e-01 3.21144640e-01
-1.16492653e+00 -9.52189088e-01 2.25404277e-01 -8.53964314e-02
-1.55978501e+00 -4.14352089e-01 1.14565635e+00 -3.18222404e-01
7.90738702e-01 1.47619277e-01 5.55029213e-01 8.13689828e-01
-7.74073675e-02 6.04826331e-01 6.73329473e-01 -4.76416126e-02
1.27053678e-01 7.97747634e-03 -1.68349519e-01 6.59994960e-01
2.50104874e-01 4.38656896e-01 -6.17483258e-01 -5.24275243e-01
1.24891925e+00 5.66978335e-01 -2.36211360e-01 -8.27710629e-01
-1.07920229e+00 5.96833348e-01 7.60000169e-01 -2.08977476e-01
-5.67198038e-01 1.80658326e-01 -4.86044973e-01 -9.59114432e-02
3.78234208e-01 1.09654903e-01 -6.68025553e-01 4.42517847e-01
-6.31716430e-01 3.27188492e-01 6.04886115e-01 1.34006906e+00
1.32262290e+00 -1.50735065e-01 -5.39272558e-03 5.30361414e-01
6.60942018e-01 1.01141453e+00 -2.09498584e-01 -1.00031614e+00
8.13189387e-01 8.14842105e-01 1.96220562e-01 -9.54234123e-01
-5.12329005e-02 -2.93102950e-01 -8.12870562e-01 2.37263411e-01
1.45507101e-02 3.01045269e-01 -1.07282281e+00 1.57656288e+00
4.05741334e-01 6.31965637e-01 -2.00499862e-01 8.29644501e-01
8.55377316e-01 6.55567884e-01 -1.47918910e-01 1.40480295e-01
1.17097640e+00 -4.92846400e-01 -3.21039647e-01 -2.28955239e-01
-5.98194599e-02 -6.99404418e-01 9.11155999e-01 1.84330627e-01
-9.45701182e-01 -5.88527501e-01 -1.06814444e+00 -2.24003702e-01
-1.95357859e-01 3.68671827e-02 4.77822095e-01 8.29977691e-02
-7.90857911e-01 7.97859669e-01 -8.73724759e-01 1.49321511e-01
6.59328759e-01 4.20807958e-01 -4.52512085e-01 -4.89268959e-01
-5.45825362e-01 4.90989357e-01 1.84555531e-01 2.66949199e-02
-9.16873276e-01 -1.19843578e+00 -7.92094111e-01 6.15648478e-02
4.24553365e-01 -1.15963721e+00 8.23103845e-01 -3.25953990e-01
-1.23641396e+00 6.11627698e-01 -5.21405637e-01 1.89997777e-01
1.83892623e-01 -3.00804377e-01 5.07246479e-02 1.44225553e-01
1.49241745e-01 9.82787490e-01 1.18732774e+00 -1.81759369e+00
-3.19350511e-01 -6.93553209e-01 -1.59417465e-01 2.35447004e-01
2.60427833e-01 -5.86974561e-01 -6.79181218e-01 -4.99314815e-01
8.66150856e-01 -8.96996558e-01 -1.71018571e-01 4.79495376e-01
-2.30027989e-01 -3.14767301e-01 9.32787716e-01 -5.62401831e-01
5.41434169e-01 -2.15138984e+00 6.01013042e-02 4.51388091e-01
1.60799801e-01 -4.49874960e-02 -3.52941036e-01 3.96675318e-01
-1.27172261e-01 2.84713581e-02 -5.07270396e-01 -7.27132261e-01
-9.08441544e-02 6.73365176e-01 -9.81372893e-01 3.25109720e-01
4.52642858e-01 9.85590637e-01 -8.14382672e-01 -2.84438044e-01
5.01263022e-01 6.84213400e-01 -5.68235397e-01 3.38023186e-01
-5.46979070e-01 6.76386535e-01 -7.37634957e-01 7.04880834e-01
1.25822532e+00 -2.32562408e-01 -2.58089542e-01 -3.10156345e-01
-1.55771524e-01 2.07489446e-01 -1.22048354e+00 2.22236300e+00
-2.75323838e-01 -1.59320518e-01 3.13578509e-02 -6.06492460e-01
1.19825959e+00 1.76485240e-01 5.07437825e-01 -4.55244303e-01
-2.37250805e-01 3.58684920e-03 -5.67562640e-01 -3.17368835e-01
3.97874117e-01 -2.40194768e-01 2.96960890e-01 2.79392183e-01
-9.51858163e-02 -7.20115542e-01 -6.33435488e-01 8.61440003e-02
8.93499017e-01 6.14942968e-01 -4.21788730e-02 1.63192436e-01
3.99736464e-01 -1.57087971e-03 7.09796548e-01 6.26178086e-01
4.06871498e-01 1.08448863e+00 6.55145645e-02 -5.49583972e-01
-1.06493783e+00 -1.51047564e+00 -5.54381013e-02 4.52628940e-01
3.78438085e-01 -2.90192068e-01 -8.90611708e-02 -4.94283736e-01
2.57430971e-01 9.31100726e-01 -3.51250201e-01 -8.40878859e-02
-7.83413231e-01 -2.62479186e-01 3.04423515e-02 6.57266974e-01
3.94412369e-01 -1.06393516e+00 -3.03954214e-01 1.51951000e-01
-2.88058490e-01 -1.02281094e+00 -3.02077234e-01 -2.38441765e-01
-1.41457176e+00 -1.01114202e+00 -2.54976302e-01 -4.64003265e-01
1.09184241e+00 6.68448269e-01 1.12539172e+00 2.52458036e-01
3.68984491e-02 4.10067827e-01 -4.63317521e-02 -3.91960770e-01
-2.77384184e-02 -3.25207114e-01 -2.50293463e-01 4.24748473e-02
2.01661065e-01 -1.07380605e+00 -5.29912531e-01 1.95482209e-01
-8.60977590e-01 2.31108174e-01 9.98221874e-01 5.90229094e-01
1.09630239e+00 -1.39014095e-01 1.20291203e-01 -5.57916105e-01
-1.62031144e-01 -6.01441860e-01 -6.33189499e-01 1.90503243e-02
-4.69230562e-01 3.42572838e-01 3.65321815e-01 -2.93679267e-01
-1.24960208e+00 4.99498039e-01 -3.49801369e-02 -1.23695123e+00
-1.84435278e-01 3.27062517e-01 -5.98229170e-01 1.66863441e-01
2.25029811e-01 4.46017742e-01 1.80606321e-01 -1.02641404e+00
5.38298488e-01 3.58159766e-02 6.28151655e-01 -8.11824322e-01
1.15218949e+00 1.01261830e+00 3.05089265e-01 -7.10398078e-01
-6.39758825e-01 -5.59184253e-01 -7.89913654e-01 1.56098217e-01
4.92762387e-01 -1.11608410e+00 -5.07707179e-01 1.84070453e-01
-1.54288399e+00 2.06132784e-01 -3.95992428e-01 2.15791374e-01
-5.53883553e-01 5.02451479e-01 -1.75726727e-01 -5.46871245e-01
-2.34063327e-01 -8.92827332e-01 1.59667575e+00 3.34684588e-02
3.22098672e-01 -4.30447429e-01 -5.47321513e-02 1.94136724e-01
-2.31924374e-02 1.23040773e-01 1.02395880e+00 -1.42328635e-01
-1.49950540e+00 -1.15254954e-01 -4.25610811e-01 7.30246753e-02
9.74296555e-02 -2.22552270e-01 -9.28444743e-01 -2.10160062e-01
2.41922051e-01 1.73104614e-01 9.94115174e-01 3.43827486e-01
1.26629448e+00 -2.95443177e-01 -6.87633097e-01 7.80668497e-01
1.57131946e+00 -1.65207550e-01 8.27402890e-01 -2.22400784e-01
1.00347853e+00 6.01814985e-01 6.04866862e-01 5.41143954e-01
5.48035502e-01 4.67190534e-01 9.39159214e-01 2.12359130e-01
-2.97693610e-01 -9.42642987e-01 -1.94750912e-02 8.04285765e-01
-4.35345545e-02 2.13001907e-01 -8.58677328e-01 4.67699736e-01
-1.83666050e+00 -6.88263535e-01 -2.25951895e-01 2.43917441e+00
6.41148388e-01 -1.08993061e-01 -6.28943026e-01 -2.21864238e-01
7.01320469e-01 9.35673341e-02 -5.61797976e-01 5.37202179e-01
-3.43191475e-02 2.61106670e-01 3.65665168e-01 5.93358874e-01
-7.51293302e-01 9.25556242e-01 5.32833052e+00 7.09140182e-01
-7.89353907e-01 9.56267565e-02 -1.43550020e-02 1.77334517e-01
-7.91305542e-01 5.32033861e-01 -7.71715105e-01 2.25031853e-01
2.71825999e-01 2.55987227e-01 4.53518122e-01 7.38107085e-01
9.68482047e-02 8.80999432e-04 -1.27504194e+00 1.26391923e+00
-1.33333064e-03 -1.40332246e+00 4.03230488e-01 2.25033358e-01
5.46447754e-01 1.33610085e-01 -6.25653043e-02 7.27670267e-02
2.37961233e-01 -7.70283103e-01 7.36407995e-01 1.15377772e+00
8.95708144e-01 -4.56100106e-01 2.84086615e-01 7.20459580e-01
-1.22259545e+00 1.72569871e-01 -8.64443421e-01 -1.16017893e-01
2.11794943e-01 7.47855723e-01 -9.30872500e-01 9.81234252e-01
6.42925739e-01 8.18009913e-01 -4.03236538e-01 1.07340884e+00
-4.70019728e-01 3.71833593e-01 -6.54213548e-01 5.22704482e-01
-1.92525208e-01 -3.26728582e-01 8.23303640e-01 4.34758693e-01
5.86815178e-01 4.23839957e-01 2.13279396e-01 1.52684760e+00
1.22500556e-02 -2.29615882e-01 -8.89280796e-01 4.43353295e-01
8.10114205e-01 1.12558591e+00 -5.32909393e-01 -3.39284301e-01
-3.32157582e-01 5.69932163e-01 4.02616948e-01 2.97374815e-01
-4.48412627e-01 1.54529735e-01 6.12524569e-01 2.77627259e-01
5.68494320e-01 -3.67433906e-01 -6.60145521e-01 -1.05826867e+00
2.04221085e-01 -1.35594994e-01 -2.99388897e-02 -1.21883464e+00
-1.43430078e+00 1.77196085e-01 2.94565082e-01 -1.48144317e+00
9.52489022e-03 -4.23368365e-01 -7.08247125e-01 1.06150675e+00
-1.59004736e+00 -1.46893799e+00 -5.66424727e-01 6.50851786e-01
4.63205099e-01 1.37705982e-01 7.08566427e-01 3.91833261e-02
6.42486056e-03 -9.73551273e-02 -3.52215707e-01 -6.53875247e-02
5.25458574e-01 -8.25606406e-01 4.37667012e-01 6.02518082e-01
1.47168085e-01 7.01203585e-01 5.01377761e-01 -1.15120184e+00
-1.72851145e+00 -1.19187796e+00 6.75501049e-01 -8.04562628e-01
-7.98902195e-03 -2.13734195e-01 -1.14348614e+00 6.73251450e-01
-3.33381653e-01 1.97109267e-01 1.43766105e-01 -9.99981835e-02
-5.14209509e-01 -1.31421596e-01 -1.04071450e+00 4.15423185e-01
1.30543613e+00 -4.24981773e-01 -1.03786981e+00 3.86463813e-02
1.12345648e+00 -4.53193694e-01 -5.70814669e-01 6.83720410e-01
3.83940876e-01 -6.77465439e-01 1.35155046e+00 -2.95719773e-01
4.97553051e-01 -5.16942203e-01 -4.46078271e-01 -1.12157118e+00
-3.78866255e-01 -3.07913065e-01 -1.86142355e-01 1.28020430e+00
-2.50528958e-02 -5.47607303e-01 9.57050741e-01 7.00221002e-01
-4.80565488e-01 -5.41353464e-01 -8.72792006e-01 -6.11934185e-01
7.86939934e-02 -6.59249365e-01 1.33857048e+00 7.71471679e-01
-7.44402528e-01 2.68780857e-01 -1.66903079e-01 6.32884741e-01
9.74251091e-01 5.87514400e-01 9.36104894e-01 -1.64810157e+00
-2.45200656e-02 -3.49098374e-03 4.18598428e-02 -1.25930333e+00
2.70242572e-01 -8.89369965e-01 1.29557848e-01 -1.52969003e+00
1.04562201e-01 -8.34257543e-01 4.29248177e-02 6.90712392e-01
4.81021823e-03 -1.36498049e-01 3.77046227e-01 6.47250772e-01
-6.65898100e-02 9.72481012e-01 1.50193238e+00 -2.90531933e-01
-2.24154070e-01 1.82045460e-01 -5.75558126e-01 7.39840806e-01
4.32767212e-01 -7.41191506e-01 -4.36514318e-01 -6.99126363e-01
2.86055386e-01 3.03647459e-01 8.97991836e-01 -7.67655551e-01
4.86517012e-01 -2.89263695e-01 5.73938191e-01 -1.50763774e+00
8.02417040e-01 -1.23463285e+00 5.41250467e-01 5.61344028e-02
3.34810674e-01 -2.10329667e-02 5.95634282e-02 8.63397777e-01
8.54450166e-02 -9.52983722e-02 2.01742694e-01 -3.83680493e-01
-4.96797055e-01 7.48443246e-01 5.47007024e-01 -3.84396046e-01
5.43594658e-01 -3.02660376e-01 -1.82450101e-01 -1.72442898e-01
-6.04071081e-01 2.58170515e-01 5.46334803e-01 7.42552757e-01
1.22855902e+00 -1.48703516e+00 -6.62138522e-01 6.77890956e-01
1.56770691e-01 8.30088556e-01 3.30346614e-01 4.67464387e-01
-1.79291293e-01 3.88228416e-01 -8.83570015e-02 -1.04997158e+00
-7.63634920e-01 7.43391335e-01 -9.28775221e-02 1.10159472e-01
-9.97519493e-01 5.40634453e-01 6.85636997e-01 -7.73069561e-01
-6.38772771e-02 -3.82563323e-01 1.02768384e-01 -3.24039578e-01
2.80855983e-01 1.67623430e-01 1.34509608e-01 -5.94705105e-01
-2.28419021e-01 9.76087391e-01 3.10758322e-01 -1.55413002e-01
1.46878874e+00 -2.24552229e-01 -4.05278563e-01 2.03875110e-01
9.51550126e-01 5.20481765e-02 -1.70179868e+00 -6.77290678e-01
-2.62299597e-01 -8.29236984e-01 -5.13285547e-02 -6.91520870e-01
-1.09019041e+00 8.42597306e-01 3.09614837e-01 -4.86409485e-01
8.25642467e-01 2.22053677e-01 5.41333735e-01 3.13825279e-01
7.40369976e-01 -5.42459965e-01 -4.65448014e-02 4.22051221e-01
1.42387223e+00 -1.06902516e+00 9.66456383e-02 -8.62718284e-01
-1.93367511e-01 9.96930540e-01 5.98230302e-01 -3.45445544e-01
9.85249102e-01 5.47408536e-02 -3.73611093e-01 -5.09892166e-01
-7.80343056e-01 -8.75002444e-02 3.41663778e-01 7.03495443e-01
-5.04048824e-01 1.38266429e-01 3.04565281e-01 7.11129367e-01
-2.36360833e-01 3.46077010e-02 1.34200409e-01 7.67882347e-01
-4.00748610e-01 -1.13842595e+00 -7.34501123e-01 4.06037390e-01
2.18866184e-01 -2.03325469e-02 -2.33821705e-01 6.80788338e-01
3.00047815e-01 5.82668006e-01 3.06823105e-01 -3.56221139e-01
4.95178103e-01 1.55939862e-01 5.64828038e-01 -7.88635492e-01
4.52439450e-02 1.33021697e-01 -4.32341844e-01 -8.54951084e-01
-2.83544064e-01 -6.11268044e-01 -1.42581069e+00 -1.84634045e-01
-1.38297364e-01 -2.23550081e-01 8.19831550e-01 1.00800073e+00
7.46997237e-01 3.27179641e-01 5.73397040e-01 -1.38328600e+00
-5.53779542e-01 -7.69576013e-01 -2.59270102e-01 4.26442057e-01
3.17028373e-01 -1.04458463e+00 -1.79546073e-01 -8.57864395e-02] | [8.380796432495117, -3.481930732727051] |
e5096a1c-566d-4a43-b22a-62eb0baa578d | dialect-identification-in-nuanced-arabic | 2102.09749 | null | https://arxiv.org/abs/2102.09749v2 | https://arxiv.org/pdf/2102.09749v2.pdf | Dialect Identification in Nuanced Arabic Tweets Using Farasa Segmentation and AraBERT | This paper presents our approach to address the EACL WANLP-2021 Shared Task 1: Nuanced Arabic Dialect Identification (NADI). The task is aimed at developing a system that identifies the geographical location(country/province) from where an Arabic tweet in the form of modern standard Arabic or dialect comes from. We solve the task in two parts. The first part involves pre-processing the provided dataset by cleaning, adding and segmenting various parts of the text. This is followed by carrying out experiments with different versions of two Transformer based models, AraBERT and AraELECTRA. Our final approach achieved macro F1-scores of 0.216, 0.235, 0.054, and 0.043 in the four subtasks, and we were ranked second in MSA identification subtasks and fourth in DA identification subtasks. | ['Anshul Wadhawan'] | 2021-02-19 | null | https://aclanthology.org/2021.wanlp-1.35 | https://aclanthology.org/2021.wanlp-1.35.pdf | eacl-wanlp-2021-4 | ['dialect-identification'] | ['natural-language-processing'] | [-1.17072448e-01 -1.25209972e-01 1.14043273e-01 -4.20512557e-01
-1.28524208e+00 -9.67591703e-01 9.65767622e-01 -3.85674601e-03
-5.07855654e-01 1.05409026e+00 1.97282821e-01 -5.26565075e-01
-2.63443410e-01 -4.91725773e-01 -4.03245687e-01 -6.00773990e-01
-3.83691043e-02 1.18668580e+00 1.18345410e-01 -7.83646405e-01
3.78248990e-01 4.84713137e-01 -1.03558171e+00 3.65232617e-01
1.33709502e+00 6.87636137e-01 -9.49641094e-02 5.49215078e-01
2.58797612e-02 1.04416621e+00 -5.46002805e-01 -6.31818175e-01
-4.44358662e-02 -3.80584806e-01 -1.36866081e+00 -1.77358687e-01
5.29263854e-01 -3.46136987e-01 -4.50217649e-02 9.46069181e-01
3.48774731e-01 -1.14762411e-01 9.22699153e-01 -1.21015739e+00
-3.76514077e-01 1.29769945e+00 -8.50028157e-01 1.76185325e-01
3.95696759e-01 -2.73622960e-01 1.03694534e+00 -1.42563689e+00
6.37240350e-01 1.25926983e+00 8.49965274e-01 1.67938381e-01
-1.19712198e+00 -5.96970797e-01 -4.92785424e-02 2.63646990e-01
-1.73769414e+00 -9.42403913e-01 4.12206650e-01 -4.81979638e-01
8.36504757e-01 4.85564232e-01 -7.77842924e-02 8.16902399e-01
-2.41167903e-01 8.33549380e-01 1.22059572e+00 -6.96807563e-01
-2.87515640e-01 -3.78100425e-02 7.67819107e-01 6.38625741e-01
-1.94512680e-02 -6.34527147e-01 -3.96346092e-01 -2.72060692e-01
2.50306040e-01 -8.87163758e-01 1.02584034e-01 1.65206164e-01
-1.49767852e+00 7.97976315e-01 1.68783963e-01 5.14703751e-01
-5.09913266e-01 -5.48998058e-01 3.19307983e-01 5.72028756e-01
4.91986603e-01 4.03649300e-01 -5.06985009e-01 -8.94690454e-02
-1.07474613e+00 5.37250102e-01 6.57795906e-01 8.80964637e-01
6.15776956e-01 -4.68719825e-02 -6.70586750e-02 1.37160945e+00
2.38695115e-01 6.35048747e-01 3.68457109e-01 -5.60326219e-01
9.23721492e-01 3.69570315e-01 3.80492717e-01 -6.47007644e-01
-6.24372065e-01 -2.65123621e-02 -6.24426067e-01 -1.50904313e-01
1.13195956e+00 -6.61359131e-01 -1.06211197e+00 1.65871298e+00
1.45703889e-02 -4.77050662e-01 8.39856789e-02 7.46717811e-01
6.37000978e-01 7.03537941e-01 -5.26391268e-02 7.66528994e-02
1.49515378e+00 -9.96810734e-01 -6.23481154e-01 -2.81581283e-01
7.10470676e-01 -1.20120037e+00 8.74825239e-01 4.14309204e-01
-1.14663994e+00 -2.57624507e-01 -9.66457546e-01 -6.80682287e-02
-6.42302096e-01 4.73493338e-01 1.62241653e-01 8.37637961e-01
-1.19946241e+00 3.87119949e-02 -4.69001025e-01 -4.34065193e-01
-1.60445757e-02 3.79839361e-01 -4.23614234e-01 -5.97736947e-02
-1.34570003e+00 1.16498673e+00 2.60192037e-01 1.62698537e-01
-4.94400293e-01 -5.74836671e-01 -7.89695501e-01 -2.17779666e-01
-5.29845320e-02 3.18159685e-02 1.11534142e+00 -1.11760533e+00
-1.36304283e+00 1.22674763e+00 -1.98223293e-01 -4.33382034e-01
3.76293510e-01 -1.94680616e-01 -8.47738564e-01 -3.63923609e-01
5.34065723e-01 3.58041465e-01 7.70882070e-01 -1.14321983e+00
-1.21510947e+00 -7.93741107e-01 -4.80754673e-02 1.72011569e-01
1.04011461e-01 7.44733632e-01 -2.08416134e-01 -8.54270041e-01
2.77028024e-01 -1.14414346e+00 1.54440507e-01 -1.25927877e+00
-5.76194286e-01 -6.31386936e-02 5.87736011e-01 -1.58899486e+00
1.42004085e+00 -1.76600099e+00 2.98951447e-01 7.77313769e-01
1.26915589e-01 2.22661912e-01 -2.01900497e-01 5.04430175e-01
-1.85294628e-01 -1.43076107e-01 -3.00717384e-01 8.50917585e-03
8.63885432e-02 -2.98100531e-01 -4.64394301e-01 2.19733492e-01
2.51018107e-01 7.27538347e-01 -7.25323856e-01 -1.27617285e-01
-3.06797266e-01 7.87410140e-03 -1.30444899e-01 -1.68735236e-01
-6.73717028e-03 3.37126195e-01 1.05895519e-01 9.07956362e-01
8.72819364e-01 2.90262282e-01 4.47906703e-01 1.09327575e-02
-5.15177190e-01 6.08561933e-01 -1.13736129e+00 1.32909691e+00
-1.86405405e-01 7.47332036e-01 4.16900992e-01 -6.74168169e-01
7.35725820e-01 1.36424333e-01 4.40833479e-01 -6.11226618e-01
6.56235814e-02 7.14432478e-01 2.54497468e-01 -1.62303939e-01
7.95380533e-01 4.14779931e-01 -5.27557313e-01 6.67186975e-01
-6.60828175e-03 2.84161717e-01 5.36081612e-01 1.50316671e-01
6.85358524e-01 -3.64435581e-03 2.31026962e-01 -7.77960896e-01
1.06533790e+00 4.24012244e-01 1.88060284e-01 6.81609809e-01
-1.22850247e-01 4.65357244e-01 4.75818604e-01 -3.73952508e-01
-8.50856781e-01 -1.08648527e+00 -2.86745846e-01 1.55495584e+00
-3.85205001e-01 -3.85426074e-01 -9.91907954e-01 -7.34698117e-01
-2.64139414e-01 8.19640338e-01 -4.04523134e-01 2.33632118e-01
-1.11463952e+00 -1.24256718e+00 1.21217477e+00 2.40100563e-01
5.11485696e-01 -7.93766201e-01 1.15021884e-01 3.23424846e-01
-9.70672429e-01 -7.02475786e-01 -7.04310596e-01 -2.64564697e-02
-2.54387170e-01 -8.32650840e-01 -9.46182013e-01 -1.14598060e+00
2.82685608e-01 5.89391813e-02 1.24806166e+00 -2.36345962e-01
3.50863576e-01 -2.22520620e-01 -3.81081939e-01 -3.96118820e-01
-5.05845487e-01 9.60246146e-01 -1.90884583e-02 2.62724966e-01
7.34523058e-01 -4.14129980e-02 2.11809337e-01 3.44272584e-01
-3.80493999e-01 -9.23812017e-02 2.69953668e-01 6.29127741e-01
1.23669818e-01 -3.36751282e-01 8.97042215e-01 -1.04585516e+00
7.09316492e-01 -5.78648686e-01 -5.59726894e-01 5.91237426e-01
-4.11023676e-01 -3.16015407e-02 3.85917395e-01 -7.22728148e-02
-1.06582546e+00 3.19052398e-01 -5.05701125e-01 6.49661839e-01
-2.39871994e-01 8.47905874e-01 -4.48963702e-01 1.34403646e-01
7.07987607e-01 2.48067066e-01 -1.08079035e-02 -3.95588458e-01
1.96515352e-01 1.33289659e+00 5.75520217e-01 -6.10664129e-01
5.61778367e-01 -1.09440371e-01 -5.13936520e-01 -9.57913637e-01
-5.99133193e-01 -3.73664290e-01 -1.04336143e+00 -1.24510303e-01
6.93165779e-01 -1.00891745e+00 -2.55136132e-01 1.20278299e+00
-7.97793686e-01 -4.53153372e-01 4.32862252e-01 6.86973929e-02
-3.57669860e-01 2.92751163e-01 -8.04173946e-01 -4.27664697e-01
-4.05716598e-01 -1.13917029e+00 7.51526833e-01 -3.48729461e-01
-5.10745347e-01 -7.02262104e-01 2.44762748e-01 5.08146167e-01
3.85543078e-01 -2.53887456e-02 1.30057216e+00 -8.67959499e-01
-5.84516972e-02 5.76297529e-02 -3.82647783e-01 -9.78123918e-02
3.00231308e-01 3.71293053e-02 -1.05881131e+00 -2.33757764e-01
-4.66960967e-01 -3.26125890e-01 6.65096045e-01 4.05240297e-01
4.67578202e-01 -3.09844881e-01 -1.05606345e-02 2.70208716e-01
9.61909533e-01 2.66450375e-01 6.54904187e-01 5.61811507e-01
6.94549143e-01 7.27790654e-01 7.67904222e-01 -3.06172073e-02
9.24409330e-01 9.00032759e-01 -1.04720712e-01 2.91188434e-02
1.17976051e-02 1.37450218e-01 5.62525392e-01 8.17589998e-01
-2.14060038e-01 -2.77127028e-02 -1.61538815e+00 1.00886500e+00
-1.89416468e+00 -9.15816069e-01 -6.07222319e-01 2.15544438e+00
1.01409090e+00 -2.13605091e-01 9.87375200e-01 3.59147638e-01
6.65971518e-01 -1.03379875e-01 -2.80016456e-02 -4.80412126e-01
-4.60984766e-01 1.42984316e-01 5.44462979e-01 9.57084596e-01
-1.37526560e+00 1.56253004e+00 5.99586344e+00 8.37014973e-01
-1.01468074e+00 -9.46537498e-03 5.24320006e-01 2.93301433e-01
1.72605127e-01 -3.15760940e-01 -9.21264291e-01 5.17345667e-01
1.39252019e+00 -2.65510995e-02 7.60418355e-01 1.32337973e-01
-7.35629648e-02 -2.46338379e-02 -7.87634611e-01 6.54299080e-01
1.71302795e-01 -1.12539947e+00 2.18878567e-01 5.62975295e-02
6.38485372e-01 2.65621960e-01 6.14735298e-02 3.49203348e-01
6.17612422e-01 -1.10586095e+00 1.21375310e+00 3.32146138e-01
8.95064175e-01 -1.14239025e+00 8.13757300e-01 2.56862491e-01
-1.01593876e+00 -6.70907125e-02 5.29216640e-02 1.08209074e-01
-6.61188588e-02 1.13399580e-01 -9.31945086e-01 6.12863243e-01
5.11553168e-01 5.02143621e-01 -7.02114761e-01 6.25364482e-01
-1.30061423e-02 9.36177611e-01 -1.45749941e-01 2.63112903e-01
5.97450197e-01 -4.37185228e-01 4.85688657e-01 1.50399518e+00
2.84209728e-01 -2.47813717e-01 -3.68891992e-02 1.61095455e-01
-6.82419464e-02 3.09511989e-01 -3.18489432e-01 1.86201215e-01
4.80925471e-01 1.00188684e+00 -4.58928645e-01 -3.19482744e-01
-2.80875385e-01 1.04956841e+00 3.81835252e-01 2.98461616e-01
-8.32579195e-01 -8.82587910e-01 6.92594469e-01 7.94775842e-04
1.45699918e-01 -2.15477005e-01 -4.56753969e-01 -1.13667238e+00
-2.25472450e-01 -1.43474793e+00 7.11887896e-01 -4.52549487e-01
-1.14346814e+00 9.30193245e-01 -1.71770200e-01 -7.22853780e-01
-7.55251884e-01 -7.00389326e-01 -8.89582038e-02 1.37636304e+00
-1.17532301e+00 -1.75192857e+00 4.18767668e-02 8.21329892e-01
5.05333841e-01 -6.80229247e-01 9.83063519e-01 7.23876059e-01
-7.28724718e-01 8.82135689e-01 3.91515553e-01 5.95242143e-01
1.21147823e+00 -1.35034823e+00 5.98270118e-01 1.01156759e+00
9.91818458e-02 5.13226151e-01 4.08755660e-01 -4.19742376e-01
-8.23704422e-01 -1.05464447e+00 1.81045580e+00 -8.17399561e-01
9.90176976e-01 -4.18509871e-01 -7.07189500e-01 1.01048529e+00
2.97337115e-01 -9.29285109e-01 4.26814526e-01 3.89353245e-01
-3.49037290e-01 1.76419560e-02 -1.23238468e+00 6.02510989e-01
7.85151005e-01 -6.09730184e-01 -5.45581639e-01 3.30871075e-01
1.07322931e-01 -4.46416795e-01 -9.30828154e-01 -7.44999126e-02
6.55200005e-01 -4.02001947e-01 1.00112963e+00 -8.08217049e-01
2.91330755e-01 -1.50668874e-01 -2.91543931e-01 -1.64341331e+00
-5.02328396e-01 -5.54098606e-01 3.99588138e-01 1.47011399e+00
1.06821966e+00 -7.06675589e-01 3.63331079e-01 2.99657762e-01
-9.26567093e-02 7.18671009e-02 -7.59669363e-01 -6.02056205e-01
6.20752633e-01 -2.53152698e-01 1.09474552e+00 1.45762265e+00
-2.51715146e-02 7.52753198e-01 -3.75628620e-01 3.75422895e-01
4.92824852e-01 1.27631677e-02 7.04312027e-01 -1.29445589e+00
4.17535007e-01 -6.52743876e-01 8.38253051e-02 -5.59502542e-01
2.26778045e-01 -1.06374443e+00 -9.18560922e-02 -1.19032001e+00
-1.81675836e-01 -6.50541842e-01 -7.16743395e-02 6.64292872e-01
-1.03376813e-01 4.66187775e-01 4.77619842e-02 4.68008518e-01
-2.09224895e-01 -2.08851710e-01 3.43608886e-01 -3.13055038e-01
-4.20412779e-01 2.28785351e-01 -8.13960135e-01 6.17167532e-01
1.26616406e+00 -2.80097365e-01 -1.47667937e-02 -7.54215419e-01
1.54824987e-01 -8.73758569e-02 4.78555188e-02 -7.01509416e-01
7.18760490e-02 -1.07733384e-01 5.26758470e-02 -8.72441411e-01
1.10886589e-01 -6.25320852e-01 -4.09833863e-02 2.38079771e-01
-1.54614940e-01 5.79877853e-01 2.88841277e-01 -3.84622455e-01
-2.04545766e-01 -2.56297737e-01 7.77570963e-01 1.51170954e-01
-7.43215680e-01 -1.09750770e-01 -1.04491746e+00 1.70677170e-01
5.86949825e-01 5.60168512e-02 -2.62029797e-01 -2.58882850e-01
-8.61736536e-01 2.42445350e-01 2.46765971e-01 4.56875741e-01
8.89969766e-02 -1.20844805e+00 -1.43184602e+00 3.60579908e-01
2.17322633e-01 -7.55317390e-01 -2.70439118e-01 9.36697304e-01
-8.00351322e-01 6.97012544e-01 -6.15095258e-01 -2.15652406e-01
-1.46856308e+00 -8.58847946e-02 3.73479873e-01 -4.87936825e-01
1.77429020e-02 6.46614254e-01 -4.39375490e-01 -1.10338402e+00
3.69202308e-02 4.61572148e-02 -4.99942452e-01 5.36044836e-01
6.36817455e-01 8.00743401e-01 4.87480640e-01 -1.41955590e+00
-7.00751722e-01 2.81666845e-01 -4.74408388e-01 -6.17535472e-01
1.32158208e+00 -2.91652203e-01 -6.28113508e-01 3.15496475e-01
7.87320554e-01 5.40297508e-01 -4.13200587e-01 -2.83628821e-01
7.21198797e-01 -4.99542467e-02 -8.29388127e-02 -1.34044206e+00
-5.97976506e-01 4.02776301e-01 5.21553159e-01 2.92308211e-01
8.52818608e-01 -1.94048598e-01 8.62123489e-01 3.61598700e-01
3.10076714e-01 -1.28520298e+00 -8.34106207e-01 1.38640726e+00
9.69365656e-01 -1.17028999e+00 -4.90433186e-01 -2.32793167e-01
-8.04491937e-01 8.65080714e-01 2.99320012e-01 2.54607826e-01
5.97451985e-01 -7.89174903e-03 5.25403678e-01 -1.64861605e-01
-1.08859293e-01 -3.91349882e-01 6.16028309e-01 6.76096678e-01
6.69514358e-01 3.13285619e-01 -4.77154970e-01 8.52672100e-01
-6.64673984e-01 -4.67325360e-01 5.39166987e-01 6.88595474e-01
-2.91069120e-01 -1.08270073e+00 -6.34652257e-01 4.73834664e-01
-6.11053228e-01 -3.25018913e-01 -9.89270031e-01 9.92033958e-01
-4.98145185e-02 1.22530341e+00 1.41136408e-01 -5.70514321e-01
2.96122253e-01 4.55504924e-01 3.02361995e-01 -2.99441785e-01
-9.58517551e-01 -6.42102882e-02 6.57162547e-01 3.76856676e-03
-1.34982914e-01 -1.13186991e+00 -1.15526903e+00 -6.75876439e-01
-1.08593404e-01 1.85498446e-01 4.19426978e-01 8.81306350e-01
1.54560387e-01 -1.48297057e-01 5.99057853e-01 -5.55810630e-01
-3.81217390e-01 -1.19043112e+00 -5.41914284e-01 1.59558624e-01
4.12207723e-01 -2.65573502e-01 1.18043572e-02 1.38254389e-01] | [10.170493125915527, 10.816917419433594] |
64bf5f58-6c39-40fa-b46b-158eb67ecb66 | improving-self-supervised-single-view-depth | 1908.11112 | null | https://arxiv.org/abs/1908.11112v1 | https://arxiv.org/pdf/1908.11112v1.pdf | Improving Self-Supervised Single View Depth Estimation by Masking Occlusion | Single view depth estimation models can be trained from video footage using a self-supervised end-to-end approach with view synthesis as the supervisory signal. This is achieved with a framework that predicts depth and camera motion, with a loss based on reconstructing a target video frame from temporally adjacent frames. In this context, occlusion relates to parts of a scene that can be observed in the target frame but not in a frame used for image reconstruction. Since the image reconstruction is based on sampling from the adjacent frame, and occluded areas by definition cannot be sampled, reconstructed occluded areas corrupt to the supervisory signal. In previous work arXiv:1806.01260 occlusion is handled based on reconstruction error; at each pixel location, only the reconstruction with the lowest error is included in the loss. The current study aims to determine whether performance improvements of depth estimation models can be gained by during training only ignoring those regions that are affected by occlusion. In this work we introduce occlusion mask, a mask that during training can be used to specifically ignore regions that cannot be reconstructed due to occlusions. Occlusion mask is based entirely on predicted depth information. We introduce two novel loss formulations which incorporate the occlusion mask. The method and implementation of arXiv:1806.01260 serves as the foundation for our modifications as well as the baseline in our experiments. We demonstrate that (i) incorporating occlusion mask in the loss function improves the performance of single image depth prediction models on the KITTI benchmark. (ii) loss functions that select from reconstructions based on error are able to ignore some of the reprojection error caused by object motion. | ['Maarten Schellevis'] | 2019-08-29 | null | null | null | null | ['depth-and-camera-motion'] | ['computer-vision'] | [ 5.61389029e-01 4.86478835e-01 -1.56563535e-01 -4.19939190e-01
-6.21287763e-01 -3.53979617e-01 5.61348498e-01 -1.70996830e-01
-3.78582925e-01 8.31082404e-01 3.30963105e-01 1.38708264e-01
1.50754094e-01 -7.92886019e-01 -1.31675577e+00 -6.99894667e-01
8.58717188e-02 3.63763869e-01 3.75184506e-01 2.09066257e-01
1.70063823e-01 4.42449003e-01 -1.70356381e+00 7.05355227e-01
5.11778593e-01 1.00536966e+00 3.16930711e-01 7.17095971e-01
3.99713702e-02 1.06032908e+00 -4.83716160e-01 -1.17716782e-01
6.16964400e-01 -5.47753334e-01 -5.76947331e-01 2.95249850e-01
1.08834279e+00 -8.56273711e-01 -4.55610484e-01 7.46472716e-01
2.26468310e-01 4.05093795e-03 5.30883968e-01 -1.02801704e+00
2.89061159e-01 8.04339051e-02 -5.12990713e-01 4.17691618e-02
4.46794301e-01 1.56224981e-01 8.27118278e-01 -8.69032383e-01
1.14053452e+00 1.19048357e+00 4.22028989e-01 5.66434681e-01
-1.49029565e+00 -2.77940333e-01 3.00436765e-01 1.37023389e-01
-1.02051759e+00 -6.22630715e-01 8.50306153e-01 -5.34775257e-01
9.12400484e-01 1.76015526e-01 4.94725615e-01 1.01168001e+00
3.41718465e-01 7.78183341e-01 9.84019637e-01 -4.71494198e-01
2.61898965e-01 2.82717794e-01 -1.16559595e-01 7.18385160e-01
2.14391248e-03 4.63798761e-01 -6.92820609e-01 1.90402284e-01
9.16158199e-01 -3.14706743e-01 -6.48523808e-01 -7.97323585e-01
-1.20311952e+00 8.35830331e-01 3.51135850e-01 -1.15706511e-01
-2.46511310e-01 3.28854084e-01 2.81280369e-01 3.82615715e-01
7.56838679e-01 1.05975680e-01 -4.47398007e-01 -3.62576768e-02
-1.32949865e+00 6.54510185e-02 6.98538780e-01 6.96971476e-01
1.02691066e+00 3.84546146e-02 -9.40347314e-02 5.72000623e-01
1.63093433e-01 1.65381595e-01 1.50026217e-01 -1.54431808e+00
5.12517273e-01 2.53883749e-01 1.77727103e-01 -6.39301717e-01
-1.69822827e-01 -3.78171533e-01 -3.83160621e-01 8.77221346e-01
7.48900712e-01 -1.73060261e-02 -8.31107020e-01 1.82824945e+00
4.47905034e-01 3.41533095e-01 1.40074253e-01 9.88761365e-01
7.39575624e-01 6.32736981e-01 -3.60409915e-01 -3.81014466e-01
9.09564555e-01 -1.14621699e+00 -5.22213638e-01 -4.18426126e-01
6.24013782e-01 -7.58977592e-01 8.52080345e-01 6.85092092e-01
-1.36968017e+00 -7.14018583e-01 -1.27893817e+00 -3.38919789e-01
2.08988443e-01 -2.45196130e-02 3.04812133e-01 4.07298416e-01
-1.18993771e+00 9.36335862e-01 -8.95684004e-01 -2.41833359e-01
3.20228100e-01 2.29572818e-01 -4.91760433e-01 -2.71486282e-01
-8.02669585e-01 8.76111090e-01 1.83684945e-01 -2.38778561e-01
-1.20238733e+00 -8.85659873e-01 -9.46662605e-01 -1.31476492e-01
3.55148733e-01 -9.16004956e-01 8.61861229e-01 -1.49402571e+00
-1.34918797e+00 9.77388918e-01 -4.94620681e-01 -7.05124080e-01
1.07816184e+00 -2.38856539e-01 2.43168175e-01 6.04736149e-01
1.47364140e-01 1.07636344e+00 1.07759798e+00 -1.49949837e+00
-6.98197126e-01 -2.92785466e-01 2.72536248e-01 4.17752326e-01
1.36126772e-01 -4.50186431e-01 -5.42462945e-01 -5.73479533e-01
2.49172613e-01 -9.78196800e-01 4.33123410e-02 4.51440126e-01
-1.00462630e-01 4.40969467e-01 9.04899895e-01 -8.15928280e-01
6.80848777e-01 -2.09320903e+00 2.86291152e-01 -1.02766812e-01
1.06899135e-01 -1.47781163e-01 -9.03971717e-02 1.54658124e-01
-1.85756564e-01 -1.22905686e-01 -3.54702264e-01 -8.83467197e-01
-4.30781275e-01 3.96384776e-01 -5.24260819e-01 6.42269671e-01
1.82608981e-02 4.87057537e-01 -6.58365726e-01 -3.53689194e-01
6.84379637e-01 6.24672771e-01 -7.85281897e-01 2.96736956e-01
-3.68698597e-01 8.53212118e-01 3.01923640e-02 2.42745966e-01
6.99009418e-01 4.57554087e-02 -1.45746261e-01 -3.08980405e-01
-1.13432445e-01 4.09100384e-01 -1.17105174e+00 1.88805294e+00
-4.88019317e-01 7.65958071e-01 2.57549763e-01 -9.11174357e-01
4.89760309e-01 3.22042137e-01 6.24752581e-01 -5.16740739e-01
-1.30417556e-01 1.12588905e-01 -2.75991648e-01 -3.94051492e-01
3.20728153e-01 -1.65172130e-01 4.13619518e-01 2.30634555e-01
9.00586024e-02 -2.01183408e-01 1.28645167e-01 2.27183715e-01
1.10653973e+00 6.48684144e-01 5.17260022e-02 -3.66231129e-02
5.47196031e-01 -1.08536646e-01 6.36497974e-01 5.46166539e-01
-1.22970678e-01 1.11793971e+00 5.93541145e-01 -3.85005593e-01
-1.23939323e+00 -1.07363343e+00 -3.31067920e-01 5.71679533e-01
1.82027996e-01 -1.80049822e-01 -6.84248209e-01 -7.76028872e-01
-1.95147410e-01 7.31453717e-01 -6.72837913e-01 -6.51122257e-02
-6.68246627e-01 -2.28432730e-01 1.54738232e-01 3.14706534e-01
4.91677552e-01 -9.33604121e-01 -7.48456478e-01 2.13796422e-01
-4.41761702e-01 -1.40482187e+00 -2.88712353e-01 2.54631966e-01
-1.21426797e+00 -1.05545926e+00 -6.23485923e-01 -4.42101270e-01
7.22585201e-01 3.94609183e-01 1.26212466e+00 3.94727886e-02
7.66580999e-02 5.92047036e-01 -1.92119524e-01 -4.62684445e-02
-5.17609417e-01 -2.37493366e-01 -2.30076209e-01 1.35823056e-01
-1.52850091e-01 -6.50891721e-01 -7.95721054e-01 2.64687330e-01
-8.99708569e-01 4.24480796e-01 1.88409284e-01 8.22673023e-01
7.52116084e-01 -1.79418907e-01 4.80330810e-02 -9.92974758e-01
-5.05481303e-01 -3.01218659e-01 -6.55350685e-01 -1.32374212e-01
-4.72693026e-01 9.57177952e-02 4.84690011e-01 -1.57061249e-01
-1.05797052e+00 3.18149477e-01 -1.29522681e-01 -8.73015583e-01
-1.49499595e-01 4.12661768e-02 -3.16606909e-01 -2.18615860e-01
5.10794222e-01 1.40769273e-01 1.69672519e-01 -3.70475292e-01
9.77530703e-02 4.54268940e-02 3.41664225e-01 -3.64001691e-01
4.79031146e-01 1.03551197e+00 2.69265771e-01 -8.51466656e-01
-9.04482126e-01 -4.05244738e-01 -6.48928881e-01 -4.02793139e-01
8.70078623e-01 -1.21162677e+00 -1.32699773e-01 2.23873675e-01
-1.21305013e+00 -4.91135716e-01 -5.19985139e-01 6.22275949e-01
-8.04551661e-01 5.91929317e-01 -4.95470941e-01 -7.38032877e-01
5.58726713e-02 -1.27315629e+00 1.24256897e+00 -1.81039587e-01
-2.29698017e-01 -1.08622432e+00 -7.22731277e-02 5.67260623e-01
-3.50876749e-02 3.73344958e-01 5.71357846e-01 -8.78499374e-02
-1.02616525e+00 3.18826698e-02 5.62954545e-02 7.30256319e-01
-1.87031567e-01 -2.64880154e-02 -1.38117695e+00 -4.41977203e-01
4.74553764e-01 -1.84447050e-01 1.18472052e+00 7.52565980e-01
8.95834029e-01 -3.59771512e-02 -1.91833898e-01 8.98885906e-01
1.72924173e+00 7.41107985e-02 1.03717434e+00 3.64393353e-01
9.17519152e-01 1.01310587e+00 5.62766612e-01 1.15458094e-01
2.78388381e-01 8.30320179e-01 7.42226005e-01 -1.23835765e-01
-4.41691756e-01 -3.76500279e-01 6.50204539e-01 2.48793766e-01
8.20002854e-02 -2.76256412e-01 -4.16424394e-01 5.21475494e-01
-1.70073700e+00 -9.23102379e-01 -3.94026875e-01 2.79300547e+00
6.12139583e-01 4.77109671e-01 -6.75499514e-02 1.60346135e-01
3.93048525e-01 3.36968929e-01 -4.92361665e-01 -2.25679368e-01
-2.97930330e-01 2.39181608e-01 5.80493927e-01 1.07062757e+00
-1.01675510e+00 7.88786411e-01 5.45426369e+00 5.99721253e-01
-9.93354201e-01 2.20953390e-01 7.82515466e-01 -4.37345177e-01
-2.52727032e-01 3.70947748e-01 -7.87957132e-01 4.81707722e-01
7.16678381e-01 3.83905351e-01 2.83132493e-01 6.55111611e-01
5.58223188e-01 -5.89322150e-01 -1.60056913e+00 8.62103820e-01
3.10853451e-01 -1.24027288e+00 -5.72478175e-02 2.62229323e-01
1.00984859e+00 -6.53886944e-02 -1.88908987e-02 -1.09158326e-02
-2.60641307e-01 -8.16640675e-01 1.00261045e+00 4.51289445e-01
6.45546734e-01 -5.18457413e-01 5.96132457e-01 3.21649432e-01
-8.53894293e-01 1.98686257e-01 -3.04496855e-01 -2.14223474e-01
3.20068955e-01 9.01817262e-01 -6.31485939e-01 4.81461167e-01
5.06234229e-01 8.85277033e-01 -3.81079793e-01 8.90169680e-01
-4.01488692e-01 5.76141238e-01 -2.26115555e-01 7.64320493e-01
5.16622812e-02 -3.62374574e-01 8.58777940e-01 8.38290453e-01
1.57525554e-01 -1.94748580e-01 8.69496241e-02 8.54677916e-01
8.13922882e-02 -1.55413151e-01 -8.06014001e-01 6.99183524e-01
4.12064232e-02 8.24580073e-01 -6.06208861e-01 -4.72999305e-01
-5.86019993e-01 1.25409234e+00 1.56618685e-01 4.09493953e-01
-9.27753448e-01 2.73332238e-01 6.67942941e-01 5.68670332e-01
5.24159849e-01 -6.95849806e-02 -4.98229146e-01 -1.29445076e+00
1.64484099e-01 -6.19358003e-01 1.75887018e-01 -9.80796516e-01
-9.01690900e-01 3.36136818e-01 1.71586685e-02 -1.50923312e+00
-3.76977175e-01 -3.63816112e-01 -4.59803820e-01 7.58388519e-01
-1.60124755e+00 -7.01466680e-01 -4.02669340e-01 3.22042376e-01
7.70585716e-01 2.51500756e-01 5.04179060e-01 2.93688715e-01
-2.06915215e-01 3.54764730e-01 -2.97243092e-02 -2.63697505e-01
8.48077595e-01 -1.17987812e+00 -4.85032126e-02 9.16241705e-01
2.26040289e-01 7.69066587e-02 9.04500008e-01 -5.47375858e-01
-8.18679273e-01 -1.12652981e+00 8.61000657e-01 -5.77396691e-01
6.59547746e-02 -3.27883422e-01 -9.57413733e-01 8.41511905e-01
-2.17095353e-02 1.96283072e-01 1.27666339e-01 -4.41025347e-01
-1.89620629e-01 -1.68181181e-01 -1.19527674e+00 4.24812466e-01
9.43403959e-01 -5.05210996e-01 -2.38554806e-01 3.29935879e-01
6.32821560e-01 -5.61962783e-01 -6.76719964e-01 4.86240357e-01
5.01821995e-01 -1.53823292e+00 1.04067731e+00 -6.79498166e-02
9.54224467e-01 -3.68927985e-01 -2.03107268e-01 -9.29985642e-01
2.29883581e-01 -3.10503215e-01 -2.73663044e-01 1.01387513e+00
1.41467497e-01 -3.32721114e-01 1.08826101e+00 5.81142068e-01
-1.70377627e-01 -6.08076334e-01 -1.14390719e+00 -7.05021799e-01
7.03143999e-02 -5.14140546e-01 -1.49530917e-01 7.11600482e-01
-5.93719542e-01 3.24899226e-01 -4.28310126e-01 3.09450924e-02
9.31730092e-01 -9.47774053e-02 9.42340314e-01 -9.32645857e-01
-6.23535633e-01 -1.36938155e-01 -4.85826403e-01 -1.41291153e+00
1.43306166e-01 -5.99919975e-01 -2.49759052e-02 -1.45016146e+00
-3.11068166e-02 -1.96727142e-01 2.06282213e-01 1.24563910e-01
9.93259177e-02 4.13357228e-01 2.46516079e-01 3.77906382e-01
-2.03594297e-01 5.78576207e-01 1.20317924e+00 2.15734523e-02
-6.65276572e-02 8.03874899e-03 -9.10461545e-02 9.46026564e-01
5.04700184e-01 -5.19413590e-01 -5.71699500e-01 -4.10659313e-01
-4.26539853e-02 3.82493049e-01 7.15426564e-01 -1.01189983e+00
-1.13515399e-01 2.45835379e-01 5.24884760e-01 -5.49249351e-01
7.70964742e-01 -9.55164075e-01 3.47104937e-01 4.68601376e-01
-2.85391957e-01 -2.82647967e-01 1.52536288e-01 7.29550302e-01
-2.09422484e-01 -4.30713862e-01 9.65995193e-01 -2.33273506e-01
-5.96164703e-01 6.43851757e-02 -1.82016239e-01 -4.89400811e-02
9.70536709e-01 -6.30542457e-01 -2.66929753e-02 -6.37838662e-01
-8.62483978e-01 -6.42220154e-02 9.08243835e-01 1.91755220e-01
6.04456604e-01 -1.12990177e+00 -7.19665408e-01 1.98782638e-01
-1.28718436e-01 1.71084896e-01 2.53687203e-01 8.52556109e-01
-8.51444542e-01 3.22800398e-01 -5.66441827e-02 -8.02862227e-01
-1.37004828e+00 4.82791513e-01 4.87002075e-01 -4.14328694e-01
-8.75005782e-01 8.48202407e-01 7.69156218e-01 -4.47976924e-02
5.07049441e-01 -3.70868772e-01 -1.92780793e-02 -1.51048720e-01
4.03172880e-01 2.24601254e-01 1.32273078e-01 -7.72130132e-01
-9.36454237e-02 6.41562104e-01 8.40626135e-02 -4.59182352e-01
1.20907700e+00 -3.42045307e-01 1.04421534e-01 6.98830247e-01
1.52442741e+00 1.32206753e-01 -1.90023112e+00 -4.26056795e-02
-4.80257154e-01 -7.55069673e-01 3.53926152e-01 -7.59948850e-01
-1.24443996e+00 8.68141055e-01 7.54835844e-01 -2.77135491e-01
1.07302523e+00 -2.26994291e-01 6.80606902e-01 -1.87329322e-01
5.17903149e-01 -8.94183278e-01 7.76756033e-02 2.56651163e-01
8.45136404e-01 -1.31371009e+00 1.66771546e-01 -6.22099817e-01
-5.12903035e-01 1.02720475e+00 6.61465108e-01 -3.22174966e-01
4.62592661e-01 1.54489160e-01 -2.10259184e-01 -3.38448808e-02
-7.70643651e-01 -6.06698394e-02 1.98625103e-01 5.03925502e-01
2.49131039e-01 -3.86004090e-01 -2.85818875e-01 3.91861014e-02
6.17959024e-03 -5.65423071e-02 7.07595825e-01 7.74932086e-01
-3.17151129e-01 -1.04784369e+00 -4.99699116e-01 3.53132546e-01
-3.19604695e-01 -1.65390372e-01 -1.95660055e-01 6.83590353e-01
3.84711206e-01 8.15587878e-01 2.30974868e-01 -1.20736457e-01
2.19025776e-01 -4.92126010e-02 8.20457697e-01 -6.88582361e-01
-3.47499579e-01 2.49350637e-01 1.98126882e-01 -8.50537002e-01
-5.74324250e-01 -8.37583184e-01 -1.22780526e+00 -1.56691432e-01
-1.39434397e-01 -2.39464328e-01 6.03263140e-01 8.83559525e-01
1.57744125e-01 6.14785135e-01 6.44070566e-01 -1.16437876e+00
-7.53796548e-02 -7.40628183e-01 -4.50455099e-01 4.72563654e-01
5.36614895e-01 -6.93910956e-01 -8.87277424e-01 4.03878063e-01] | [8.773252487182617, -2.3782365322113037] |
b0f2fadb-b885-45e2-ab5a-ae863158ecaa | new-pyramidal-hybrid-textural-and-deep | 2203.15090 | null | https://arxiv.org/abs/2203.15090v1 | https://arxiv.org/pdf/2203.15090v1.pdf | New pyramidal hybrid textural and deep features based automatic skin cancer classification model: Ensemble DarkNet and textural feature extractor | Background: Skin cancer is one of the widely seen cancer worldwide and automatic classification of skin cancer can be benefited dermatology clinics for an accurate diagnosis. Hence, a machine learning-based automatic skin cancer detection model must be developed. Material and Method: This research interests to overcome automatic skin cancer detection problem. A colored skin cancer image dataset is used. This dataset contains 3297 images with two classes. An automatic multilevel textural and deep features-based model is presented. Multilevel fuse feature generation using discrete wavelet transform (DWT), local phase quantization (LPQ), local binary pattern (LBP), pre-trained DarkNet19, and DarkNet53 are utilized to generate features of the skin cancer images, top 1000 features are selected threshold value-based neighborhood component analysis (NCA). The chosen top 1000 features are classified using the 10-fold cross-validation technique. Results: To obtain results, ten-fold cross-validation is used and 91.54% classification accuracy results are obtained by using the recommended pyramidal hybrid feature generator and NCA selector-based model. Further, various training and testing separation ratios (90:10, 80:20, 70:30, 60:40, 50:50) are used and the maximum classification rate is calculated as 95.74% using the 90:10 separation ratio. Conclusions: The findings and accuracies calculated are denoted that this model can be used in dermatology and pathology clinics to simplify the skin cancer detection process and help physicians. | ['Sengul Dogan', 'Turker Tuncer', 'Mehmet Baygin'] | 2022-03-28 | null | null | null | null | ['skin-cancer-classification'] | ['medical'] | [ 5.28891563e-01 -3.52102518e-01 -3.12573105e-01 -2.08407491e-02
-6.39923573e-01 -1.36018828e-01 3.67504448e-01 2.93589711e-01
-4.57434148e-01 9.37790096e-01 -2.47712675e-02 5.41536175e-02
-2.42070943e-01 -9.93975520e-01 2.77550310e-01 -1.26109540e+00
-4.80980277e-02 -2.83162355e-01 3.69490981e-01 -8.05342644e-02
5.91754258e-01 8.58424306e-01 -1.69485784e+00 8.00475836e-01
1.07766867e+00 1.09728158e+00 2.02655327e-02 1.12754989e+00
-2.07171813e-01 5.85726380e-01 -5.42802334e-01 -6.03814535e-02
2.60066718e-01 -6.06960297e-01 -7.59445190e-01 2.23527521e-01
1.60111308e-01 -8.62689093e-02 4.86199886e-01 1.07902420e+00
8.45193148e-01 -1.63122281e-01 1.14720821e+00 -9.39385712e-01
-3.57298136e-01 -2.41644859e-01 -8.46792459e-01 2.31706530e-01
4.18138176e-01 7.94860795e-02 1.52155772e-01 -8.53437722e-01
6.36364579e-01 9.28155780e-01 8.06320548e-01 6.09084606e-01
-9.14638877e-01 -5.81917167e-01 -6.81633174e-01 6.37937784e-01
-1.66372120e+00 6.57929406e-02 5.61308205e-01 -2.24706888e-01
8.43778908e-01 6.39167309e-01 1.02774835e+00 7.96333432e-01
7.33183086e-01 2.69318342e-01 1.95915377e+00 -9.01884139e-01
3.71550143e-01 4.13655072e-01 2.76499838e-02 8.81399989e-01
3.53614718e-01 -1.59797654e-01 -2.95769900e-01 -1.40872881e-01
7.93504238e-01 -1.20582052e-01 7.65798464e-02 4.78921860e-01
-3.67733896e-01 7.47336090e-01 3.54280800e-01 8.50314200e-01
-5.93848109e-01 -1.12440638e-01 4.67127919e-01 1.45570740e-01
6.73341826e-02 -2.63764672e-02 -1.06039522e-02 4.08615656e-02
-8.40946555e-01 -6.61936849e-02 5.13568819e-01 2.74719030e-01
5.27482271e-01 -1.46595374e-01 -3.04469466e-01 1.03696990e+00
2.01167017e-01 4.91329521e-01 8.23413253e-01 -5.19158065e-01
-2.54415542e-01 6.84422791e-01 -3.93994451e-01 -1.16418421e+00
-2.91849047e-01 -8.98435190e-02 -1.10883498e+00 5.56046426e-01
9.03955624e-02 1.12056047e-01 -1.28127038e+00 7.02777088e-01
4.42111194e-01 4.01940234e-02 2.10380301e-01 7.50567615e-01
7.17115521e-01 6.50835812e-01 4.31598991e-01 -2.82222331e-01
1.64582932e+00 -5.59833169e-01 -6.48084939e-01 6.52087152e-01
3.17058146e-01 -1.14737225e+00 5.97913742e-01 7.71177590e-01
-7.13590205e-01 -5.65956771e-01 -9.92612898e-01 2.91216999e-01
-7.02443302e-01 5.55150032e-01 5.24225831e-01 1.05628800e+00
-8.74333560e-01 4.15624559e-01 -9.04581964e-01 -1.03998828e+00
5.07376671e-01 2.56462246e-01 -5.47217309e-01 -3.40769291e-01
-8.95409167e-01 9.10571754e-01 1.69235453e-01 1.16708882e-01
-4.05440539e-01 -2.19209552e-01 -4.15362954e-01 -4.34712797e-01
-5.11324644e-01 -3.50036919e-01 3.76439184e-01 -1.33379960e+00
-1.37200475e+00 1.04147661e+00 -3.19562852e-01 -1.78522304e-01
1.58595294e-01 3.91939551e-01 -5.89140952e-01 9.99466658e-01
-1.32545624e-02 2.60985851e-01 5.63857138e-01 -9.79438126e-01
-8.46674860e-01 -5.58160007e-01 -6.87559783e-01 3.29207122e-01
-5.92103839e-01 2.17724174e-01 1.00173227e-01 -4.32278156e-01
1.17002122e-01 -5.24652004e-01 -3.69117558e-01 2.50632197e-01
-4.63419825e-01 -1.27761781e-01 8.50413024e-01 -1.13712251e+00
1.23101437e+00 -1.92514181e+00 -5.35893559e-01 9.09500837e-01
-2.89206713e-01 5.53758562e-01 -8.85159001e-02 6.27724946e-01
-1.12419732e-01 3.81405711e-01 -1.05499551e-01 4.52274233e-01
-5.51692307e-01 -7.15954825e-02 5.80800653e-01 4.79252100e-01
3.32266241e-01 2.78006047e-01 -5.62594354e-01 -1.05051124e+00
6.33887947e-01 7.80510426e-01 1.38094291e-01 -2.74229944e-01
4.45919514e-01 -1.03376076e-01 -5.25229454e-01 1.09495246e+00
1.10679114e+00 2.92398512e-01 -6.14549704e-02 -6.36562347e-01
-1.32163167e-01 -6.95211947e-01 -1.35484624e+00 9.78793323e-01
-2.29834422e-01 3.86805505e-01 1.26048908e-01 -7.35422730e-01
1.08286655e+00 3.43624562e-01 5.87578654e-01 -8.37313533e-01
1.35352582e-01 2.01848522e-01 -2.83546239e-01 -1.20541012e+00
-1.90216169e-01 -3.64540607e-01 5.78238785e-01 -3.72464478e-01
-3.27649742e-01 -2.27439348e-02 2.86454469e-01 -1.12292662e-01
9.41844404e-01 -1.81501955e-01 6.22730374e-01 -2.88081855e-01
9.91818488e-01 3.89638066e-01 3.76482934e-01 1.42802387e-01
-5.03130853e-01 3.85910034e-01 4.48258460e-01 -2.63678282e-01
-7.63417840e-01 -1.06874716e+00 -5.73211491e-01 4.54168111e-01
-1.32343829e-01 1.86105192e-01 -9.26617563e-01 -4.11322296e-01
-8.80257040e-02 3.76772255e-01 -6.89784586e-01 3.31867896e-02
-1.21162578e-01 -1.09167778e+00 6.33753896e-01 1.98175520e-01
9.00147378e-01 -9.49964285e-01 -4.59272832e-01 -1.83119521e-01
2.16032222e-01 -5.15831351e-01 2.99927205e-01 -7.75968432e-02
-9.08233166e-01 -1.46195400e+00 -9.26687062e-01 -1.21143138e+00
1.05067611e+00 8.90272558e-02 3.95664304e-01 1.21947184e-01
-1.53529751e+00 2.91041642e-01 -5.93113184e-01 -1.95028841e-01
-2.97831774e-01 -3.21211547e-01 -3.34824711e-01 5.67049198e-02
7.68038809e-01 -2.80096561e-01 -8.80678296e-01 -1.68756619e-01
-9.08251643e-01 -3.03253889e-01 1.06450951e+00 8.67544234e-01
7.52681792e-01 6.08750105e-01 4.54495549e-01 -6.64741814e-01
6.35472357e-01 -2.55247146e-01 -2.97994260e-02 3.13398898e-01
-5.10440946e-01 -4.89479482e-01 8.65336657e-01 -1.25260800e-01
-1.07536268e+00 1.88945048e-02 -2.79351801e-01 1.45482123e-01
-3.97399217e-01 2.98960954e-01 2.28732247e-02 -4.65626836e-01
1.08075655e+00 6.10245228e-01 2.82218814e-01 -1.48875579e-01
-2.17357010e-01 9.28052902e-01 4.11374457e-02 -4.39600833e-02
3.98795545e-01 5.76204240e-01 4.27208513e-01 -1.42382181e+00
3.55889536e-02 -6.09158874e-01 -4.63776857e-01 -4.26556200e-01
1.18583179e+00 -5.83355963e-01 -4.52537179e-01 8.56206715e-01
-6.91121399e-01 1.07434407e-01 7.95939565e-02 3.90751451e-01
-2.78326161e-02 5.84555447e-01 -6.48715854e-01 -1.07520366e+00
-8.24044824e-01 -8.02736461e-01 6.01059437e-01 7.89373159e-01
2.90383026e-02 -1.07987356e+00 -1.42189562e-01 3.29065382e-01
6.86232507e-01 9.25609231e-01 9.34245825e-01 -3.16759050e-01
4.46791798e-02 -5.50958037e-01 -2.86748737e-01 6.10672176e-01
6.00572705e-01 7.87040412e-01 -8.55651259e-01 -1.11994721e-01
-3.03909540e-01 -3.00950855e-01 8.35953176e-01 4.98769283e-01
1.24427700e+00 -2.22949609e-01 -4.33060259e-01 4.01220828e-01
2.18184876e+00 4.22138482e-01 1.00185585e+00 1.60052299e-01
1.56993508e-01 5.10570109e-01 6.47581100e-01 3.45420599e-01
7.13372529e-02 -2.05113336e-01 3.09808016e-01 -4.82233256e-01
-6.52238190e-01 3.69049869e-02 4.87536862e-02 4.54852164e-01
-4.35873479e-01 3.14896524e-01 -6.88914776e-01 6.14685059e-01
-8.72994244e-01 -1.15570176e+00 -4.42212641e-01 2.16174555e+00
9.22806680e-01 -5.19539602e-02 1.26724139e-01 7.51394987e-01
1.01727736e+00 -4.19982851e-01 -2.53638059e-01 -6.25240266e-01
-2.37156123e-01 1.00274563e+00 6.14711761e-01 4.12267685e-01
-1.01411843e+00 5.06013036e-01 5.27874231e+00 1.28036904e+00
-1.68699479e+00 -1.54869303e-01 8.84121001e-01 3.74246657e-01
-6.02624263e-04 -1.99070707e-01 -5.90324163e-01 4.77266639e-01
6.51857078e-01 1.73899323e-01 1.10116363e-01 5.18923104e-01
3.59652787e-01 -8.38582575e-01 -9.14583504e-02 1.06974936e+00
1.45112589e-01 -1.03371763e+00 3.12113434e-01 -1.62509258e-03
7.37498879e-01 -5.98679423e-01 4.26616259e-02 -2.62404084e-01
-2.79378384e-01 -1.24610865e+00 -1.65405750e-01 9.91248012e-01
1.04828775e+00 -9.72153425e-01 9.86703813e-01 -2.91156713e-02
-1.18263102e+00 -2.22273260e-01 -4.03252631e-01 3.06584328e-01
-6.06372297e-01 9.30210114e-01 -9.93172169e-01 5.27826667e-01
6.78004801e-01 2.11460233e-01 -8.73183906e-01 1.20726025e+00
1.72900349e-01 7.01098323e-01 -4.91563410e-01 -6.88254297e-01
8.24016258e-02 -1.16959833e-01 1.44125581e-01 1.27123165e+00
3.19563687e-01 -4.49422449e-02 -1.61648139e-01 4.02307004e-01
7.34865010e-01 6.84520304e-01 -4.08606499e-01 -2.14674398e-02
5.59599817e-01 1.56231296e+00 -1.01558650e+00 -1.04451150e-01
-2.33306270e-02 1.03714192e+00 -3.89336884e-01 1.70211062e-01
-5.19138336e-01 -1.13361764e+00 1.78520102e-02 3.69849950e-01
-1.34909013e-02 1.97524920e-01 -4.09302562e-01 -4.49284852e-01
-2.12413713e-01 -7.82957792e-01 4.87897038e-01 -4.92833823e-01
-1.42850566e+00 3.14473450e-01 -1.72329023e-01 -9.77967203e-01
3.23274046e-01 -9.26387668e-01 -1.00197089e+00 1.15255368e+00
-1.56396425e+00 -1.57569027e+00 -8.82846475e-01 6.11636400e-01
2.22761095e-01 -1.43901020e-01 1.11283827e+00 7.21157342e-02
-4.94915098e-01 5.34251034e-01 3.28643918e-01 1.24098293e-01
6.62288845e-01 -1.22971570e+00 -7.29105353e-01 6.08895183e-01
-7.60162592e-01 3.67993593e-01 3.26506257e-01 -6.88428581e-01
-1.21714520e+00 -1.00642145e+00 7.21528947e-01 3.22225362e-01
1.48036584e-01 2.19143197e-01 -2.74690628e-01 -3.85076493e-01
3.06589127e-01 1.31134078e-01 1.20930541e+00 -6.46876276e-01
2.05666661e-01 -5.74797332e-01 -2.09084630e+00 4.13473457e-01
2.01471120e-01 -2.67050326e-01 1.06429562e-01 3.92556608e-01
-3.98060024e-01 1.12199292e-01 -1.23082054e+00 1.81847394e-01
7.73461699e-01 -1.24465930e+00 8.83862853e-01 -2.17733771e-01
4.15579796e-01 -2.78795004e-01 -2.32097194e-01 -8.00583005e-01
-2.68537551e-01 -1.52849676e-02 5.09207666e-01 1.08359349e+00
1.76316723e-01 -6.51790977e-01 1.02325952e+00 1.62045971e-01
3.61457050e-01 -1.21838284e+00 -1.07632709e+00 -3.96694571e-01
7.87753835e-02 2.91939616e-01 -1.77325699e-02 8.40361059e-01
-1.08889394e-01 -1.56448543e-01 3.38187665e-01 -2.16423683e-02
9.05897915e-01 -5.20790875e-01 2.09404692e-01 -9.72992361e-01
3.12759489e-01 -2.33887836e-01 -8.81435037e-01 2.21065551e-01
-6.26944482e-01 -5.38704276e-01 -5.14012814e-01 -1.87911284e+00
2.48344317e-01 -4.36427981e-01 -3.49134117e-01 6.00970805e-01
-8.57973322e-02 4.75711435e-01 -1.57292619e-01 -1.84482783e-01
2.56033927e-01 -3.64889987e-02 1.38283741e+00 -1.15557730e-01
-8.22794512e-02 -4.24249396e-02 -4.34882522e-01 5.40433347e-01
9.08665717e-01 -2.35290565e-02 -2.68328071e-01 4.48863864e-01
-1.37155294e-01 -1.05551332e-01 4.84996557e-01 -1.40637851e+00
3.01733017e-01 -6.40994370e-01 1.07073081e+00 -5.75286150e-01
4.16155547e-01 -7.68839300e-01 -4.83009629e-02 1.19051802e+00
-4.97858413e-02 -3.84953201e-01 -5.95262684e-02 2.20137239e-01
-2.22007126e-01 -5.10745227e-01 1.28255582e+00 -3.52151990e-01
-8.95020545e-01 -6.76469207e-02 -6.07063055e-01 -6.88249230e-01
1.73856819e+00 -9.40234661e-01 -2.39613622e-01 1.25652686e-01
-7.44384706e-01 -3.19442749e-01 2.87419707e-01 -4.40207034e-01
9.34463322e-01 -1.38661957e+00 -7.11672723e-01 2.97621161e-01
2.34058388e-02 -3.59631121e-01 6.53679430e-01 1.07473528e+00
-1.28317356e+00 1.87335163e-02 -7.45589614e-01 -4.09269243e-01
-1.75022471e+00 -1.99875496e-02 4.89245385e-01 -7.67768100e-02
-1.22414164e-01 9.07129526e-01 -7.73297310e-01 1.32023156e-01
-1.64959967e-01 -1.32251948e-01 -6.94106936e-01 -5.97568415e-02
6.44218504e-01 9.34981465e-01 9.21748728e-02 -4.77852613e-01
-4.54953283e-01 9.40575123e-01 1.58272069e-02 6.39344603e-02
1.01649833e+00 3.32755625e-01 -4.51067388e-01 -6.39623543e-03
1.26767600e+00 2.23459557e-01 -4.57915246e-01 5.93047321e-01
-4.43739325e-01 -6.16663456e-01 -7.80823678e-02 -1.19609427e+00
-7.67512262e-01 9.32720065e-01 1.42050469e+00 3.33354324e-01
1.73246717e+00 -6.73512995e-01 7.49070048e-01 4.77552302e-02
-1.96157787e-02 -1.73581922e+00 9.84124541e-02 -1.16183177e-01
7.24108338e-01 -8.44700158e-01 2.75280863e-01 -6.64149404e-01
-5.97543359e-01 1.60650027e+00 6.13616943e-01 -4.96974766e-01
7.92396665e-01 2.52343357e-01 2.46523365e-01 -3.44665325e-03
-5.60813546e-01 -2.41993636e-01 -9.42355767e-02 1.08708417e+00
4.95923996e-01 9.23473164e-02 -1.01622379e+00 4.41379666e-01
1.90227941e-01 2.86438584e-01 2.99917489e-01 1.19838321e+00
-7.61134088e-01 -9.24989820e-01 -9.04530704e-01 9.28237498e-01
-7.40721643e-01 1.66427568e-01 -4.17310804e-01 7.61318862e-01
6.49647892e-01 9.38189447e-01 -1.50762782e-01 -4.24754053e-01
-6.28137887e-02 -8.29075426e-02 6.91100776e-01 -1.49246484e-01
-4.95583236e-01 -5.68934754e-02 9.31388885e-02 -2.09484175e-01
-6.04113877e-01 -3.22351515e-01 -1.44927049e+00 -1.63353652e-01
-4.10838783e-01 1.76548868e-01 1.14055753e+00 5.72764575e-01
1.04636438e-01 3.01394224e-01 8.23744416e-01 -2.53486395e-01
-4.02714849e-01 -1.05938053e+00 -8.07925284e-01 2.76174039e-01
3.30679752e-02 -4.24123764e-01 -2.84839869e-01 3.65104586e-01] | [15.556584358215332, -2.982295513153076] |
266ae77f-c9ae-474a-85c8-7210e5be0a9b | the-defender-s-perspective-on-automatic | 2305.12804 | null | https://arxiv.org/abs/2305.12804v2 | https://arxiv.org/pdf/2305.12804v2.pdf | The defender's perspective on automatic speaker verification: An overview | Automatic speaker verification (ASV) plays a critical role in security-sensitive environments. Regrettably, the reliability of ASV has been undermined by the emergence of spoofing attacks, such as replay and synthetic speech, as well as adversarial attacks and the relatively new partially fake speech. While there are several review papers that cover replay and synthetic speech, and adversarial attacks, there is a notable gap in a comprehensive review that addresses defense against adversarial attacks and the recently emerged partially fake speech. Thus, the aim of this paper is to provide a thorough and systematic overview of the defense methods used against these types of attacks. | ['Hung-Yi Lee', 'Helen Meng', 'Lingwei Meng', 'Jiawen Kang', 'Haibin Wu'] | 2023-05-22 | null | null | null | null | ['speaker-verification'] | ['speech'] | [ 3.50675613e-01 2.65127450e-01 1.65035814e-01 -4.26218331e-01
-7.17145681e-01 -1.02627397e+00 8.64494741e-01 -2.77735591e-01
-8.93504992e-02 7.81050265e-01 3.25077981e-01 -5.63040972e-01
2.85633683e-01 -2.18384355e-01 -3.72512728e-01 -5.81101179e-01
-1.57349646e-01 1.12802424e-01 1.17056236e-01 -7.71602809e-01
1.59921765e-01 7.42914796e-01 -1.39062071e+00 2.32452035e-01
5.61965287e-01 5.79706013e-01 -4.59737033e-01 8.39462221e-01
-1.44126758e-01 5.01066506e-01 -1.38967848e+00 -8.67101252e-01
-2.37967484e-02 -3.86943281e-01 -7.42954195e-01 -1.26325026e-01
2.85102457e-01 -3.99026722e-01 -7.71394789e-01 1.32340181e+00
7.66447604e-01 -4.24389750e-01 4.93726395e-02 -1.83129692e+00
-6.26789927e-01 9.84204412e-01 3.60734574e-02 2.61096925e-01
7.12067187e-01 4.39030416e-02 5.24368584e-01 -6.39793873e-01
2.38868237e-01 1.48933041e+00 5.00694394e-01 1.11214244e+00
-5.78050375e-01 -1.08439958e+00 6.18506819e-02 -5.02269901e-02
-1.34980989e+00 -1.05188453e+00 8.78132582e-01 -1.26239538e-01
5.17746866e-01 6.15476370e-01 1.81151390e-01 1.74680626e+00
9.43940133e-02 6.39001787e-01 1.07618737e+00 -4.13175911e-01
-3.62013001e-03 6.48930252e-01 6.36237785e-02 3.33556503e-01
4.51206505e-01 5.88000178e-01 -7.60308683e-01 -8.78223598e-01
-3.71344797e-02 -5.49742281e-01 -5.12623966e-01 1.66138425e-01
-8.90661538e-01 9.23934877e-01 -1.08854234e-01 3.38609487e-01
-1.01968974e-01 -1.97629750e-01 7.19757795e-01 6.36541367e-01
2.01054543e-01 2.94367820e-01 -1.21558085e-01 -1.42810136e-01
-7.80329823e-01 2.89947957e-01 8.65921676e-01 6.43815219e-01
1.17008179e-01 9.95406568e-01 3.56434166e-01 6.23166144e-01
5.80823064e-01 1.03805280e+00 5.77845275e-01 -4.59004343e-01
4.58471507e-01 -3.40902507e-01 1.49693757e-01 -1.16379571e+00
2.08959617e-02 -1.22888483e-01 -5.86386681e-01 1.85136393e-01
1.16596214e-01 -2.74891555e-01 -6.73062146e-01 1.78784454e+00
3.04060608e-01 7.21196309e-02 4.69711989e-01 6.35589480e-01
9.16102827e-01 3.65918577e-01 -1.81400493e-01 -3.06027830e-01
1.18723834e+00 -7.04746723e-01 -1.34536064e+00 -4.81776536e-01
1.59821242e-01 -1.11877370e+00 5.27003169e-01 1.91954583e-01
-8.27404499e-01 -8.02842304e-02 -1.39185882e+00 5.17426074e-01
-5.89289665e-01 -7.39670455e-01 4.96101379e-01 1.89823091e+00
-1.04021323e+00 6.59127310e-02 -4.92703915e-01 -1.73663408e-01
3.70574594e-01 3.85672122e-01 -3.98563117e-01 9.82362106e-02
-1.86944258e+00 9.15187120e-01 -2.57424936e-02 1.43815979e-01
-9.23551083e-01 -1.37607798e-01 -1.01137829e+00 -4.62762296e-01
1.20282374e-01 -1.22212194e-01 1.27419376e+00 -7.96179652e-01
-1.68871117e+00 8.93985748e-01 -1.32508904e-01 -8.49313498e-01
5.79759479e-01 1.30517781e-01 -1.36414099e+00 2.50082523e-01
-1.80035740e-01 2.18429565e-02 1.43605900e+00 -1.33015561e+00
-8.90710950e-02 -4.84289676e-01 -2.15996072e-01 -2.55771399e-01
-1.13839544e-01 8.90575588e-01 4.22598332e-01 -9.29462850e-01
2.17494383e-01 -9.86967504e-01 1.86380208e-01 -3.07827801e-01
-5.86748123e-01 1.50123224e-01 1.37565076e+00 -8.11298549e-01
1.15547168e+00 -2.42413330e+00 -4.80921537e-01 1.06664345e-01
-1.21341497e-02 7.94195890e-01 1.58639833e-01 7.44020462e-01
-2.27153659e-01 3.64315003e-01 -3.17041308e-01 -3.88963997e-01
4.73695546e-02 2.47802570e-01 -9.53468025e-01 7.29228675e-01
-1.77502379e-01 7.57361054e-01 -7.55342185e-01 -1.56597570e-01
1.18054748e-01 7.29136348e-01 2.78995540e-02 -1.51945464e-02
3.34269017e-01 3.16522986e-01 -3.36589366e-01 1.04491353e+00
1.02415156e+00 3.97704393e-01 1.65648013e-01 4.96536493e-01
8.06157589e-02 6.55459464e-01 -1.05953455e+00 8.28510106e-01
7.79918721e-03 6.81986570e-01 8.07864368e-01 -7.56351709e-01
9.34227407e-01 9.76034522e-01 6.57442138e-02 -8.22268203e-02
2.28037417e-01 5.97188830e-01 -1.77160695e-01 -3.21330786e-01
4.96874541e-01 -2.68539816e-01 -3.02165598e-01 6.04109347e-01
-2.82823771e-01 -2.27447629e-01 -7.74128318e-01 4.32534963e-01
7.95182645e-01 -5.83441436e-01 3.09421271e-01 3.37174744e-03
1.05576086e+00 -4.04494971e-01 2.98787087e-01 8.40002179e-01
-1.04274929e+00 3.85156363e-01 1.68956801e-01 -1.74084485e-01
-5.16212165e-01 -9.05193686e-01 -1.36437565e-01 6.39984787e-01
1.51512355e-01 -3.45040202e-01 -7.76363313e-01 -1.02570450e+00
1.17424645e-01 5.42619407e-01 -4.27882999e-01 -3.42703283e-01
-6.20727658e-01 -5.36197841e-01 1.70522892e+00 -6.80759689e-03
5.77339709e-01 -9.00182426e-01 1.13816269e-01 1.49301007e-01
-4.34983492e-01 -1.52068961e+00 -3.54846418e-01 -2.87427902e-01
-3.78185570e-01 -8.26316655e-01 -4.22874391e-01 -7.43747115e-01
2.97130197e-01 6.52407825e-01 6.17207646e-01 1.91447452e-01
1.19686522e-01 4.08178955e-01 -4.36546713e-01 -8.25433552e-01
-1.35785210e+00 -4.10545021e-01 7.49300599e-01 1.41801149e-01
2.76327699e-01 -2.75590390e-01 -4.87199500e-02 4.73468930e-01
-8.57246101e-01 -8.77296686e-01 -3.03982422e-02 8.03082526e-01
-4.09205288e-01 3.00327297e-02 1.03276968e+00 -6.82716787e-01
7.29790330e-01 -2.55248249e-01 -3.19117904e-01 1.95581198e-01
-6.01039052e-01 -2.59267509e-01 4.16333288e-01 -3.76780301e-01
-9.66492593e-01 -3.19016010e-01 -6.53028011e-01 -1.74823552e-02
-2.73384154e-01 2.31299233e-02 -4.22760278e-01 -8.51320028e-01
7.83643007e-01 5.73483407e-01 5.06737709e-01 -2.68616110e-01
5.90133406e-02 1.46126735e+00 3.39160770e-01 -9.84677374e-02
1.25190902e+00 6.30225599e-01 -3.90741169e-01 -1.37253129e+00
-3.31901997e-01 -1.62408084e-01 -8.94307494e-02 -2.58873731e-01
4.11981583e-01 -6.49075091e-01 -6.94637597e-01 1.20241976e+00
-1.32921159e+00 2.02153236e-01 3.24899137e-01 3.62857401e-01
-1.43851444e-01 1.16363478e+00 -7.29897320e-01 -1.16637421e+00
-5.34953892e-01 -1.33273101e+00 7.82684863e-01 -1.46674141e-01
-1.12269409e-01 -7.11802959e-01 -6.55249879e-02 7.88835585e-01
6.90380991e-01 2.76107460e-01 1.63889349e-01 -1.02297950e+00
-7.37405345e-02 -8.07220817e-01 2.55676389e-01 5.83368540e-01
3.10088217e-01 1.35787874e-01 -1.48311245e+00 -4.23841238e-01
5.91677666e-01 -2.17071474e-01 3.08968812e-01 4.34561372e-02
4.28751171e-01 -6.84004664e-01 -1.80001974e-01 4.01455909e-01
7.65580833e-01 5.73378503e-01 5.46627700e-01 2.66918838e-01
8.63568634e-02 8.54557037e-01 4.40055162e-01 2.57468671e-01
5.56769967e-02 6.93433106e-01 5.52361965e-01 4.60317373e-01
6.49551377e-02 -2.36689791e-01 7.06111789e-01 6.62225842e-01
4.71545160e-01 -3.62921715e-01 -6.99127555e-01 8.27878565e-02
-1.02612841e+00 -1.17395401e+00 -1.94905385e-01 2.34554720e+00
4.66438025e-01 2.20917270e-01 1.81322366e-01 8.16334724e-01
1.28547120e+00 8.65871906e-01 -1.15860000e-01 -7.03123093e-01
-5.80078900e-01 -2.15369072e-02 4.64326411e-01 9.52901900e-01
-1.04910922e+00 1.12841797e+00 7.96971369e+00 3.21906596e-01
-1.31072164e+00 2.85644174e-01 5.09904921e-02 3.46304357e-01
-4.10390571e-02 -1.61480680e-01 -6.92633629e-01 4.97460604e-01
1.14326715e+00 -2.38557830e-01 4.99047130e-01 6.35818243e-01
1.86273187e-01 5.44195592e-01 -2.60016650e-01 1.04022312e+00
5.84565818e-01 -1.08427393e+00 -1.27777522e-02 1.81396887e-01
2.84185320e-01 1.03065751e-01 1.76527068e-01 8.18576217e-02
8.23079124e-02 -7.41563380e-01 7.27230251e-01 -5.03073573e-01
6.45401537e-01 -6.09708965e-01 7.75105596e-01 1.24890238e-01
-8.95946145e-01 -4.84815761e-02 -1.04280114e-02 8.38497654e-03
5.27032375e-01 3.71221542e-01 -7.44935870e-01 5.21571040e-01
4.48600352e-01 3.58625129e-02 -1.50047868e-01 4.60091084e-01
-3.93890113e-01 7.84481466e-01 -1.62975714e-01 -1.12092786e-01
1.74079046e-01 3.42654318e-01 1.10080290e+00 1.04488683e+00
-9.26184505e-02 1.16778323e-02 -3.24109256e-01 2.90963411e-01
1.88764825e-01 -2.77992815e-01 -9.76783812e-01 -3.39726567e-01
8.81937087e-01 5.83258152e-01 -1.59672096e-01 -3.30400765e-01
-2.66071439e-01 1.11473024e+00 -5.43298960e-01 3.96907479e-01
-9.60704446e-01 -4.84959453e-01 1.03757656e+00 -3.71546075e-02
-1.62646815e-01 -5.07673860e-01 -2.35497117e-01 -1.22068143e+00
-4.27680761e-02 -1.45222104e+00 3.59644622e-01 -2.92907834e-01
-9.24301803e-01 9.06413436e-01 -3.07393730e-01 -1.08506739e+00
-2.39940226e-01 -1.16268918e-01 -4.66275066e-01 7.54475951e-01
-1.33163130e+00 -8.59785318e-01 4.94484380e-02 6.77419662e-01
4.88088816e-01 -5.98246574e-01 9.73878086e-01 1.61883116e-01
-4.61976737e-01 1.04022312e+00 -1.09366581e-01 1.70618266e-01
7.41741896e-01 -6.78523660e-01 1.01836801e+00 9.46639001e-01
-1.04598077e-02 7.92235255e-01 1.12152863e+00 -5.28304279e-01
-1.38141930e+00 -6.83103204e-01 1.14565885e+00 -5.01612663e-01
6.37801647e-01 -5.97598970e-01 -8.96544397e-01 7.02522933e-01
1.95099622e-01 -1.45549402e-01 7.40442872e-01 -5.82215607e-01
-8.40502918e-01 9.69173834e-02 -1.79269528e+00 4.03373420e-01
6.55565262e-01 -9.62787628e-01 -6.88411891e-01 4.02107984e-01
9.02977467e-01 -2.48421445e-01 -4.21434760e-01 1.09762758e-01
7.11840987e-01 -9.86741126e-01 1.08325028e+00 -5.07830918e-01
-4.10171688e-01 -1.36454597e-01 -3.11684728e-01 -9.00747359e-01
2.75924176e-01 -1.46515322e+00 -1.50246367e-01 1.30017090e+00
2.47024253e-01 -1.49188662e+00 7.08144903e-01 4.24706995e-01
-1.50770813e-01 -4.90665063e-02 -1.22970152e+00 -1.02230430e+00
-1.81977734e-01 -6.02662921e-01 9.19631600e-01 1.32734644e+00
2.35807672e-01 -1.29158637e-02 -7.81647325e-01 7.39625931e-01
8.48977685e-01 -7.85206199e-01 8.45596373e-01 -8.19242895e-01
6.86552376e-02 -1.49965435e-01 -7.05401659e-01 -7.31398284e-01
2.95863152e-01 -4.81129408e-01 -1.59416005e-01 -7.40088582e-01
-5.42795181e-01 -2.79269274e-02 7.85249323e-02 3.95265035e-02
6.40171114e-03 2.69974411e-01 3.93586718e-02 2.32544675e-01
1.82390839e-01 3.70582759e-01 7.88298547e-01 -2.04266101e-01
-6.71866536e-02 5.33118427e-01 -7.88862586e-01 5.91064930e-01
1.05479395e+00 -7.37305403e-01 -2.30460703e-01 -6.92346767e-02
-3.22098695e-02 3.25062782e-01 1.42193377e-01 -6.56039536e-01
-4.69716452e-03 -2.39485372e-02 -5.46342671e-01 -3.78887653e-01
4.94569957e-01 -6.53159201e-01 -7.53557533e-02 8.23934436e-01
-1.27244636e-01 1.40436962e-01 4.60971408e-02 6.92999184e-01
-4.40159768e-01 -1.59697726e-01 1.12014282e+00 5.42352237e-02
-3.68977368e-01 -4.99922745e-02 -5.75251877e-01 -6.84113503e-02
8.95732105e-01 -3.91469330e-01 -6.57332838e-01 -8.90645027e-01
-6.02425039e-01 -1.74902424e-01 2.78462797e-01 7.69585907e-01
8.87304246e-01 -1.01748383e+00 -8.11174989e-01 4.99945074e-01
1.14348186e-02 -7.82948911e-01 4.82792825e-01 5.41136086e-01
-5.12211978e-01 5.46738446e-01 8.94826651e-02 -1.54695809e-01
-1.82354271e+00 7.35133886e-01 4.58200037e-01 3.18320841e-01
-3.90162140e-01 8.41995001e-01 -2.30650380e-01 -3.87140960e-01
3.56322914e-01 3.34242076e-01 -1.09829865e-02 -4.69038725e-01
1.02528989e+00 4.84622031e-01 1.72212571e-01 -1.24454391e+00
-8.51041555e-01 -9.49497372e-02 -1.20383680e-01 -4.99869466e-01
5.96206844e-01 -4.13568348e-01 -2.53851473e-01 1.12128340e-01
1.15703189e+00 5.28972983e-01 -4.13758844e-01 -1.42454535e-01
-2.16896176e-01 -6.87835932e-01 -1.82504073e-01 -6.85162246e-01
-6.73230290e-01 5.70572555e-01 3.91078115e-01 7.76053965e-01
6.26550078e-01 -2.29743391e-01 1.12206078e+00 8.41942653e-02
5.60603797e-01 -6.91422343e-01 -4.77626845e-02 5.45524895e-01
9.67684031e-01 -1.14024782e+00 -2.29306743e-01 -7.87577629e-01
-7.34810054e-01 6.70927286e-01 1.06534727e-01 1.94766223e-01
8.73868227e-01 4.08082426e-01 6.11640930e-01 2.60834303e-02
-3.25605035e-01 5.18453062e-01 -2.99927026e-01 1.09096026e+00
1.56309471e-01 2.10600138e-01 -2.50985175e-01 3.12623620e-01
-6.33191705e-01 -6.09714508e-01 8.60807717e-01 1.13479257e+00
-2.79168129e-01 -1.35603881e+00 -1.06704760e+00 -1.88878179e-01
-9.41489756e-01 3.62510569e-02 -9.25545812e-01 5.13011098e-01
-4.36665118e-01 1.63860393e+00 -6.44253671e-01 -7.91283190e-01
1.46897465e-01 2.97531039e-01 -1.55085772e-02 -2.75999129e-01
-7.65886128e-01 -4.45562989e-01 3.69251698e-01 -3.62813860e-01
-3.34587216e-01 -6.84421420e-01 -7.84313798e-01 -7.40601242e-01
-5.28145432e-01 3.59560549e-01 1.24055624e+00 8.76181960e-01
2.30299935e-01 1.84305757e-02 1.17986953e+00 -4.08395886e-01
-1.12208235e+00 -8.73939693e-01 -8.32022727e-01 1.45548433e-01
8.76657486e-01 -4.91484225e-01 -1.10526896e+00 -4.55160499e-01] | [14.07193660736084, 5.866696357727051] |
3e868bc8-bdbb-40c1-8fe8-be3d602e91dc | cell-segmentation-by-combining-marker | 2004.01607 | null | https://arxiv.org/abs/2004.01607v1 | https://arxiv.org/pdf/2004.01607v1.pdf | Cell Segmentation by Combining Marker-Controlled Watershed and Deep Learning | We propose a cell segmentation method for analyzing images of densely clustered cells. The method combines the strengths of marker-controlled watershed transformation and a convolutional neural network (CNN). We demonstrate the method universality and high performance on three Cell Tracking Challenge (CTC) datasets of clustered cells captured by different acquisition techniques. For all tested datasets, our method reached the top performance in both cell detection and segmentation. Based on a series of experiments, we observed: (1) Predicting both watershed marker function and segmentation function significantly improves the accuracy of the segmentation. (2) Both functions can be learned independently. (3) Training data augmentation by scaling and rigid geometric transformations is superior to augmentation that involves elastic transformations. Our method is simple to use, and it generalizes well for various data with state-of-the-art performance. | ['Filip Lux', 'Petr Matula'] | 2020-04-03 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 2.97254592e-01 -2.10277334e-01 -3.84290628e-02 -1.91460662e-02
-9.65411246e-01 -7.06966102e-01 7.88365304e-01 5.23008764e-01
-7.20608652e-01 7.84694374e-01 -2.01329544e-01 -1.78113788e-01
3.50356042e-01 -7.37424433e-01 -7.73660481e-01 -1.03823507e+00
-8.43019933e-02 5.72811425e-01 6.16029143e-01 -4.57498692e-02
2.54779518e-01 9.41741049e-01 -9.44309473e-01 -5.93199506e-02
6.04466975e-01 9.89015043e-01 -4.37697805e-02 1.10142279e+00
-1.21332005e-01 7.66005754e-01 -1.50197148e-01 2.87721753e-01
3.30099106e-01 -1.95931584e-01 -8.04827034e-01 -1.92122553e-02
3.98415297e-01 2.68870741e-02 -1.76612571e-01 6.09371901e-01
4.94721025e-01 -2.70200998e-01 8.20725799e-01 -1.02654123e+00
-5.18658459e-01 8.18618387e-02 -5.27268350e-01 3.66809338e-01
-5.24306633e-02 2.44193435e-01 5.18983245e-01 -8.66462350e-01
9.79393184e-01 6.55805528e-01 1.15576398e+00 7.02145398e-01
-1.48148930e+00 -3.67977977e-01 -1.43562287e-01 -7.46591210e-01
-1.30254793e+00 -2.84260422e-01 1.72730222e-01 -7.01690257e-01
8.61407280e-01 1.19601741e-01 8.74953866e-01 5.70717335e-01
3.86038810e-01 7.07470775e-01 1.13335407e+00 -2.78062463e-01
2.67034888e-01 -2.52608627e-01 9.74056721e-02 8.32926273e-01
2.82267481e-01 -3.02725751e-02 -1.35594636e-01 -1.47965744e-01
1.15238714e+00 2.18698978e-01 -3.27364534e-01 -3.99038196e-01
-1.76213861e+00 6.61424518e-01 3.22971791e-01 4.92794394e-01
-1.07187457e-01 6.73055530e-01 3.10666174e-01 -2.07858860e-01
3.40726197e-01 3.88260126e-01 -5.91612041e-01 -1.03631079e-01
-1.02792656e+00 2.41404757e-01 6.88368082e-01 8.24956775e-01
5.66362083e-01 -1.90241069e-01 -2.70561427e-01 4.64276910e-01
1.49049848e-01 5.00835538e-01 4.22559679e-01 -1.03377151e+00
-1.71841040e-01 8.49400163e-01 5.81537858e-02 -6.50897145e-01
-7.67221928e-01 -3.48536283e-01 -8.45407486e-01 3.14142078e-01
1.12852228e+00 -2.02522904e-01 -1.11106050e+00 1.53251350e+00
3.90869141e-01 4.60356504e-01 -9.35496092e-02 5.54982543e-01
7.70465732e-01 3.63223195e-01 1.80045918e-01 -1.11798517e-01
1.25375926e+00 -8.04931283e-01 -6.60110176e-01 1.32646546e-01
9.80754256e-01 -5.75498164e-01 9.00745690e-01 5.73391728e-02
-1.09139633e+00 -2.76288390e-01 -9.94102120e-01 -2.95422435e-01
-7.36833990e-01 -2.97312383e-02 8.21717799e-01 4.33499992e-01
-1.27153301e+00 7.15987027e-01 -1.32241511e+00 -7.74134159e-01
1.04651260e+00 7.42419302e-01 -4.25051779e-01 2.58353651e-01
-3.12396765e-01 4.30384129e-01 1.24470159e-01 -3.56953651e-01
-7.99760938e-01 -1.17259312e+00 -6.77534521e-01 -1.28191903e-01
-1.47757605e-01 -9.52233195e-01 8.96591723e-01 -4.57760543e-01
-1.58998966e+00 1.34163022e+00 -1.65008143e-01 -5.48850834e-01
7.16699183e-01 -1.39188111e-01 1.70259133e-01 3.36167723e-01
6.83834106e-02 1.16201580e+00 7.33096004e-02 -1.34299648e+00
-8.43116462e-01 -2.25499943e-01 -5.75195014e-01 -2.58349121e-01
-2.95457959e-01 -1.31782532e-01 -6.56009078e-01 -7.35091984e-01
1.76029969e-02 -9.75963891e-01 -4.27681059e-01 4.19082284e-01
-3.86156976e-01 2.92896777e-01 1.09903193e+00 -4.46381897e-01
7.15665519e-01 -1.86262202e+00 1.21847689e-02 8.78468156e-02
3.93958300e-01 2.02895433e-01 4.00987118e-02 1.89613655e-01
3.26497376e-01 3.38759214e-01 -4.86989677e-01 -4.92302865e-01
-3.41104746e-01 1.35958448e-01 -7.95844346e-02 6.98486030e-01
2.26281226e-01 1.49510431e+00 -7.84386516e-01 -7.95248270e-01
1.30532280e-01 6.79701626e-01 -3.69142383e-01 9.36440658e-04
-7.75468051e-02 8.77657592e-01 -2.01733187e-01 1.03953862e+00
5.32655835e-01 -7.26287007e-01 -3.89253907e-02 -2.34480053e-02
-1.19774036e-01 -4.61788237e-01 -7.10437357e-01 1.39063370e+00
1.47387967e-01 7.32134879e-01 3.43332700e-02 -7.05941260e-01
8.20144832e-01 1.18637070e-01 9.99994278e-01 -1.65378645e-01
3.92594486e-01 4.96320009e-01 -3.27920169e-01 -1.43627718e-01
5.66343702e-02 -9.45729166e-02 1.16341025e-01 2.81565338e-01
8.74967948e-02 -1.95530072e-01 2.03186810e-01 1.13819376e-01
1.26416111e+00 1.33680865e-01 2.59006947e-01 -7.54146039e-01
6.46736503e-01 2.07807153e-01 5.25886834e-01 5.94030976e-01
-4.50651824e-01 8.48026872e-01 5.78571200e-01 -6.26414776e-01
-1.01718867e+00 -9.94098544e-01 -1.65648013e-01 9.62745368e-01
3.71937484e-01 6.80424795e-02 -9.03768897e-01 -5.24997294e-01
1.86280951e-01 -1.47222787e-01 -9.33377922e-01 3.99305880e-01
-8.15066397e-01 -9.78085279e-01 9.49180424e-01 8.64281893e-01
5.55636048e-01 -8.20531070e-01 -7.36665130e-01 2.44555458e-01
1.35112703e-01 -1.41918886e+00 -5.55805147e-01 3.15278709e-01
-1.18461227e+00 -1.29645205e+00 -8.38109374e-01 -1.19453013e+00
8.38082731e-01 2.65645117e-01 8.71775091e-01 7.37946510e-01
-3.81575704e-01 3.35101783e-01 9.96832103e-02 -3.79110426e-01
-2.73090601e-01 2.43443564e-01 -2.01219395e-01 -1.35993943e-01
6.36347830e-02 -3.65500331e-01 -6.33947194e-01 2.52613336e-01
-8.05027425e-01 -5.24789235e-03 2.43841335e-01 9.17899966e-01
1.31148708e+00 -3.11848372e-01 4.08707172e-01 -1.27692950e+00
3.34682435e-01 -6.32060319e-02 -6.76801264e-01 5.56950457e-02
-5.04525125e-01 -2.64473557e-01 6.19023383e-01 -5.11853755e-01
-6.44631088e-01 5.10309160e-01 -1.85960773e-02 -1.32667065e-01
-3.60830337e-01 1.74928159e-01 2.64845937e-01 -7.43531883e-01
5.10094881e-01 2.11269587e-01 4.58904415e-01 -9.62771922e-02
-4.08358127e-02 1.04516195e-02 9.11841333e-01 -4.66637701e-01
7.08532393e-01 1.36946428e+00 5.18839598e-01 -6.50485277e-01
-4.95716423e-01 -4.99517500e-01 -9.78320301e-01 -1.84651107e-01
1.10439312e+00 -7.43659794e-01 -8.11767042e-01 9.38873827e-01
-8.60233903e-01 -1.03084040e+00 -4.04189140e-01 2.10876673e-01
-6.61009848e-01 2.92708695e-01 -1.20042109e+00 -3.09230655e-01
-3.44353199e-01 -1.04485679e+00 1.32953382e+00 5.05111396e-01
-3.89601700e-02 -1.47540188e+00 2.63269067e-01 1.58888668e-01
6.02429867e-01 9.44177210e-01 7.61722088e-01 -7.31293619e-01
-8.43877852e-01 -2.29511216e-01 -2.25113854e-01 -3.91213119e-01
3.23646694e-01 5.04794180e-01 -9.69982207e-01 -4.77672428e-01
-5.51585257e-01 -1.67130664e-01 1.10265994e+00 8.17533672e-01
1.07478368e+00 1.35034725e-01 -1.06418788e+00 1.11962593e+00
1.54838586e+00 6.82374090e-02 7.87187159e-01 4.51349288e-01
7.40467370e-01 2.06748873e-01 3.02645952e-01 5.13918586e-02
1.80330053e-01 4.70537812e-01 2.53900796e-01 -8.22869480e-01
-3.67493123e-01 3.74705493e-02 -1.09166481e-01 5.63183367e-01
-9.49354321e-02 -1.41817614e-01 -1.15451276e+00 6.62146389e-01
-1.82894969e+00 -5.68561196e-01 -3.88240755e-01 1.86076689e+00
7.09412754e-01 1.27029225e-01 3.84922534e-01 -1.35702968e-01
6.32244587e-01 -3.83782923e-01 -6.92486703e-01 1.84909731e-01
-4.21769053e-01 3.21787536e-01 7.78665185e-01 5.37011921e-01
-1.36717141e+00 1.12709951e+00 7.36885834e+00 7.08443820e-01
-1.22867930e+00 8.77237786e-03 1.09474766e+00 1.66822955e-01
-9.00216680e-03 -1.81342497e-01 -8.71972680e-01 3.26789707e-01
4.10774082e-01 1.07470743e-01 6.43796548e-02 2.89733469e-01
-3.25196050e-03 -2.52950564e-02 -1.10340977e+00 5.92356205e-01
-3.80719393e-01 -2.03723669e+00 -6.41440228e-02 5.06050646e-01
9.15220320e-01 2.11508095e-01 4.66421619e-02 8.19632411e-02
4.80031103e-01 -1.24327409e+00 3.29922974e-01 5.67874014e-01
8.46082389e-01 -5.19291222e-01 1.00202358e+00 2.33334273e-01
-1.32233179e+00 1.92183509e-01 3.65010509e-03 2.22880885e-01
1.93125874e-01 3.20923001e-01 -7.50035584e-01 1.84232146e-01
5.10288537e-01 8.04954648e-01 -7.31646359e-01 1.15946388e+00
3.40285331e-01 6.12539649e-01 -5.59873939e-01 1.37306869e-01
9.83909369e-02 -2.30929643e-01 2.44966671e-01 1.72443509e+00
2.46051505e-01 4.46947291e-02 1.38967350e-01 7.68532336e-01
-2.48138249e-01 5.99857457e-02 -5.31828105e-01 3.60777718e-03
4.23122257e-01 1.51285422e+00 -1.57846475e+00 -3.19072276e-01
-1.48864865e-01 6.03224516e-01 3.76774251e-01 3.09770823e-01
-9.01101887e-01 -1.12501226e-01 4.91854370e-01 2.29553312e-01
5.33174157e-01 -3.09652358e-01 -6.83378458e-01 -6.40312314e-01
-5.84334135e-01 -2.83031285e-01 4.47021544e-01 -4.09161568e-01
-1.15849638e+00 3.20539325e-01 -4.42038029e-01 -1.03754246e+00
1.50177285e-01 -8.88452709e-01 -6.96285844e-01 3.49341393e-01
-1.68000960e+00 -1.38114071e+00 -6.34566545e-01 3.40463191e-01
2.42090136e-01 1.92864779e-02 8.73362243e-01 2.24049211e-01
-5.90038359e-01 3.63499135e-01 -3.19274627e-02 4.47606087e-01
5.00958979e-01 -1.49903727e+00 3.12363088e-01 6.32778108e-01
-3.34230587e-02 5.69174588e-01 3.57473761e-01 -5.99713683e-01
-1.27772701e+00 -1.24622643e+00 5.58303595e-01 -6.21176600e-01
5.78260601e-01 -3.18540007e-01 -9.66582537e-01 6.59878492e-01
2.15464905e-01 6.18881404e-01 8.61872494e-01 -5.08318424e-01
8.23024884e-02 1.05925322e-01 -1.27603698e+00 6.07817471e-01
7.59767652e-01 -1.33856550e-01 1.31366387e-01 3.85637045e-01
5.70261598e-01 -8.19061100e-01 -1.20075381e+00 5.53880632e-01
6.13706946e-01 -8.32722127e-01 1.00019252e+00 -4.90443230e-01
2.57162839e-01 -5.11875391e-01 3.80929634e-02 -7.75872111e-01
-4.40812707e-01 -5.85108578e-01 -3.29255909e-01 1.01100802e+00
4.52836454e-01 -6.57936454e-01 1.10175753e+00 3.52869093e-01
-1.39120772e-01 -1.24852610e+00 -9.33103800e-01 -7.07230270e-01
5.09899199e-01 1.55798972e-01 5.28710067e-01 9.93031144e-01
-1.34407967e-01 6.62025064e-02 2.21238285e-01 1.57713164e-02
5.06425440e-01 1.00216158e-02 1.04022586e+00 -1.27469265e+00
1.08282432e-01 -6.91490710e-01 -4.04865772e-01 -1.15604663e+00
-3.40046324e-02 -8.83976221e-01 -8.82965326e-02 -1.51784718e+00
3.31823349e-01 -4.94603992e-01 -3.29749465e-01 6.20447278e-01
-2.58406043e-01 7.07982183e-01 5.57167083e-02 3.95981014e-01
-6.49885535e-01 1.93279888e-02 1.48430479e+00 -1.10484481e-01
-2.26887345e-01 -2.86960334e-01 -5.38537264e-01 7.36319244e-01
9.46103394e-01 -2.89591074e-01 1.45636901e-01 -1.49054497e-01
-1.23687424e-01 -1.18379332e-01 5.51906109e-01 -1.27981257e+00
6.28177524e-01 -7.78100714e-02 6.81566894e-01 -5.94815433e-01
2.09299341e-01 -7.59712279e-01 1.84377760e-01 7.97747493e-01
-3.08195323e-01 -1.42508090e-01 5.12353539e-01 6.34459972e-01
8.04720223e-02 3.52348238e-01 1.20281339e+00 -1.29646705e-02
-3.20958346e-01 4.45354223e-01 -5.49594402e-01 2.82053351e-01
1.19869125e+00 -7.85880685e-01 -4.74934399e-01 -1.04914680e-01
-3.59307289e-01 2.80696779e-01 9.66262817e-01 -2.67915547e-01
3.51152360e-01 -1.34613037e+00 -5.65510333e-01 1.23395942e-01
8.22243188e-03 2.22328112e-01 -3.15213591e-01 1.30231977e+00
-1.19333959e+00 3.35446894e-01 -1.36811793e-01 -1.16945314e+00
-1.11924803e+00 3.90257359e-01 7.67803311e-01 -6.26142681e-01
-5.11442482e-01 8.96999300e-01 1.64635003e-01 -5.18807590e-01
-2.53412984e-02 -8.17456245e-01 -1.06120072e-01 -3.51895481e-01
2.42142916e-01 3.83260369e-01 -7.85742328e-02 -6.76355302e-01
-4.23004806e-01 1.02868521e+00 -3.02290563e-02 3.11533332e-01
1.38770616e+00 7.47197643e-02 -1.96041569e-01 3.56908917e-01
1.09041941e+00 1.55604072e-02 -1.80158496e+00 1.18873388e-01
-3.40871178e-02 -1.60779476e-01 -4.92016561e-02 -6.08789742e-01
-1.38714218e+00 6.15717709e-01 5.49139559e-01 2.56563246e-01
8.28591764e-01 1.87261477e-02 1.20545387e+00 2.61269957e-01
2.40901738e-01 -1.00952661e+00 1.14284679e-01 5.53839445e-01
1.88335195e-01 -1.21641135e+00 8.51075351e-02 -7.43091464e-01
-2.67737150e-01 1.17544389e+00 5.67853451e-01 -2.12501347e-01
7.50671804e-01 8.23727369e-01 3.89141142e-01 -5.43695092e-01
-6.46228492e-01 -6.60429597e-02 -1.06541120e-01 6.35335088e-01
4.99171555e-01 -1.18593745e-01 5.72721586e-02 2.50028431e-01
2.44407833e-01 2.67610520e-01 3.30049962e-01 1.00989342e+00
-4.34138417e-01 -7.94101000e-01 -9.96530578e-02 4.28691924e-01
-6.44671142e-01 1.88209161e-01 -4.91557211e-01 1.29612672e+00
9.57755819e-02 4.86410588e-01 1.85612991e-01 5.80987930e-02
2.02893931e-03 -2.52012849e-01 5.61503410e-01 -3.53772551e-01
-7.35387862e-01 2.78303355e-01 -5.10297358e-01 -5.36282301e-01
-5.85366964e-01 -6.85746253e-01 -1.92090285e+00 -2.29816109e-01
-3.15360487e-01 -3.62895787e-01 1.51057392e-01 8.75921071e-01
4.36861992e-01 4.85769331e-01 1.71124697e-01 -8.80512655e-01
7.07384795e-02 -5.41699886e-01 -6.73075438e-01 5.80272973e-01
4.29356486e-01 -5.53845406e-01 -2.87423819e-01 7.63208807e-01] | [14.560383796691895, -3.186525344848633] |
6d89ae70-a1b6-47e0-b4aa-fe678cc384ca | pf-net-point-fractal-network-for-3d-point | 2003.00410 | null | https://arxiv.org/abs/2003.00410v1 | https://arxiv.org/pdf/2003.00410v1.pdf | PF-Net: Point Fractal Network for 3D Point Cloud Completion | In this paper, we propose a Point Fractal Network (PF-Net), a novel learning-based approach for precise and high-fidelity point cloud completion. Unlike existing point cloud completion networks, which generate the overall shape of the point cloud from the incomplete point cloud and always change existing points and encounter noise and geometrical loss, PF-Net preserves the spatial arrangements of the incomplete point cloud and can figure out the detailed geometrical structure of the missing region(s) in the prediction. To succeed at this task, PF-Net estimates the missing point cloud hierarchically by utilizing a feature-points-based multi-scale generating network. Further, we add up multi-stage completion loss and adversarial loss to generate more realistic missing region(s). The adversarial loss can better tackle multiple modes in the prediction. Our experiments demonstrate the effectiveness of our method for several challenging point cloud completion tasks. | ['Xinyi Le', 'Jiawen Xu', 'Feng Ni', 'Zitian Huang', 'Yikuan Yu'] | 2020-03-01 | pf-net-point-fractal-network-for-3d-point-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Huang_PF-Net_Point_Fractal_Network_for_3D_Point_Cloud_Completion_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Huang_PF-Net_Point_Fractal_Network_for_3D_Point_Cloud_Completion_CVPR_2020_paper.pdf | cvpr-2020-6 | ['point-cloud-completion'] | ['computer-vision'] | [-9.13225040e-02 2.81757355e-01 3.58237833e-01 3.62956747e-02
-9.52262998e-01 -6.93558753e-01 2.75277436e-01 2.68827416e-02
8.16001594e-02 5.51508546e-01 -3.31782550e-01 3.60225663e-02
-6.07319809e-02 -9.34350967e-01 -1.28355670e+00 -2.37953335e-01
-1.56493187e-01 7.68207192e-01 2.15991110e-01 -3.10612589e-01
2.60399252e-01 1.09528267e+00 -1.19614542e+00 8.62542242e-02
1.08759320e+00 8.83639872e-01 3.09267670e-01 5.77112496e-01
2.99803335e-02 1.15329690e-01 -4.62563753e-01 -3.71918470e-01
6.95565164e-01 4.40253913e-01 -4.39542472e-01 -7.81152025e-02
6.54736578e-01 -5.64791918e-01 -4.48227048e-01 9.02800858e-01
3.59571666e-01 5.24561591e-02 6.58129513e-01 -1.42704749e+00
-3.66371691e-01 3.54698524e-02 -6.72515869e-01 -5.86616874e-01
3.20448428e-01 2.41575584e-01 6.26252234e-01 -1.32856941e+00
6.01675272e-01 1.39364028e+00 1.46161711e+00 2.54136026e-01
-1.14868665e+00 -8.65001023e-01 -8.55073333e-02 -3.26601952e-01
-1.62119639e+00 -6.40133247e-02 1.36455655e+00 -3.43443990e-01
3.91445547e-01 1.69917941e-01 7.65022814e-01 7.05178559e-01
2.95645446e-01 4.89687473e-01 6.59400523e-01 -9.55531001e-02
1.25253648e-01 -3.83546621e-01 -6.28518224e-01 5.78978240e-01
-3.98120657e-02 3.52802604e-01 -3.06296274e-02 -5.56978524e-01
1.47702718e+00 6.61427021e-01 -1.46115303e-01 -7.18517661e-01
-1.37940454e+00 6.93687379e-01 9.48517025e-01 -1.63096458e-01
-5.48623502e-01 3.99336874e-01 -4.72904881e-03 -6.60173967e-02
6.40173733e-01 2.87163109e-01 -5.32869816e-01 1.14352621e-01
-1.11330724e+00 6.48147583e-01 5.82245708e-01 1.35501111e+00
1.07351696e+00 -5.57636563e-03 -1.16624124e-01 4.90110070e-01
3.30928177e-01 6.23781919e-01 -2.04518646e-01 -1.36256933e+00
7.22010672e-01 6.74175262e-01 3.23251009e-01 -1.38036633e+00
-3.15196812e-01 -4.46052343e-01 -1.23535788e+00 6.32108510e-01
8.99982229e-02 -1.20186798e-01 -9.32269931e-01 1.30874848e+00
5.57508707e-01 6.70438588e-01 -1.25064492e-01 8.05862486e-01
6.33897126e-01 7.34035075e-01 -3.18491220e-01 2.47837842e-01
8.73619854e-01 -7.81675220e-01 -1.14700466e-01 1.38948649e-01
9.91829708e-02 -6.78705037e-01 9.66770291e-01 1.37493387e-01
-1.24158800e+00 -7.24123359e-01 -1.01425862e+00 -1.54317155e-01
9.25967470e-02 5.56497425e-02 5.31182170e-01 -1.18809231e-01
-9.68124151e-01 1.05124366e+00 -8.75220299e-01 4.54786777e-01
9.13723767e-01 3.97767305e-01 -3.54729176e-01 -3.46612662e-01
-7.76005685e-01 2.03242481e-01 7.90973380e-02 8.00913796e-02
-9.17576909e-01 -1.48856318e+00 -8.17613184e-01 2.30472356e-01
1.77921891e-01 -1.10320759e+00 9.55732465e-01 -6.75354242e-01
-1.09643483e+00 3.53654891e-01 2.81034131e-03 -2.95055956e-01
9.21984434e-01 -6.13264106e-02 1.10308565e-01 2.38709375e-01
2.78960288e-01 1.02814543e+00 1.13542080e+00 -1.82143474e+00
-5.05749762e-01 -3.79747689e-01 -3.31950784e-02 1.21830158e-01
5.04495680e-01 -7.40895629e-01 -3.85064662e-01 -8.95267129e-01
4.84847188e-01 -8.22895110e-01 -5.64892888e-01 6.53108895e-01
-4.98263031e-01 -1.30304471e-01 1.15547919e+00 -6.87043905e-01
4.72772241e-01 -2.24574280e+00 -2.01297358e-01 3.10957253e-01
5.98293900e-01 1.48893684e-01 -2.54573137e-01 4.18680847e-01
-8.22004005e-02 3.06751192e-01 -4.61381346e-01 -7.70266056e-01
-9.55848917e-02 2.11340621e-01 -6.53666258e-01 2.68471241e-01
3.04962903e-01 1.10313249e+00 -9.36441183e-01 -2.22742066e-01
4.25818264e-01 8.76145661e-01 -6.98980033e-01 -1.42864520e-02
-3.19865048e-01 5.90988576e-01 -5.07775366e-01 7.99168646e-01
1.34563303e+00 -3.20242941e-02 -6.24142110e-01 -2.82166988e-01
1.35436803e-01 -3.64219785e-01 -1.11401844e+00 2.08509994e+00
-5.03340423e-01 2.02323601e-01 1.44719630e-01 -2.61013865e-01
1.09266996e+00 2.06858560e-01 7.33321548e-01 -1.90888762e-01
-3.27773124e-01 1.86231375e-01 -4.96349514e-01 1.91760585e-02
6.38531029e-01 -2.97471792e-01 6.25253692e-02 -1.23271616e-02
-4.18774992e-01 -5.85833371e-01 -7.54543722e-01 2.26917714e-01
1.01836896e+00 2.81444401e-01 -1.49844095e-01 1.83956861e-01
3.47893655e-01 1.95260718e-01 7.74756491e-01 5.90464413e-01
7.43890787e-03 1.38248682e+00 2.88329154e-01 -6.41202867e-01
-1.57518137e+00 -1.36191940e+00 -4.87317108e-02 1.20490931e-01
4.54035461e-01 -2.92426914e-01 -4.37279373e-01 -7.22145736e-01
3.92408311e-01 5.13725221e-01 -3.33119482e-01 -1.51559198e-02
-6.46221459e-01 1.13639832e-01 4.23898995e-01 4.51172769e-01
4.35545981e-01 -9.66050446e-01 -1.23029865e-01 2.56226063e-01
-2.72071689e-01 -9.44957435e-01 -5.78376293e-01 -3.32861006e-01
-1.29181468e+00 -1.07314217e+00 -7.67987072e-01 -7.24937975e-01
1.04318810e+00 2.91910708e-01 1.20392656e+00 2.96695173e-01
3.31262196e-03 2.67960429e-01 -1.32603317e-01 -4.89194572e-01
-4.34503555e-01 4.24342752e-02 -7.04038963e-02 -2.57682770e-01
-2.16439009e-01 -1.12580895e+00 -7.75475621e-01 2.79261351e-01
-9.74969089e-01 1.45586014e-01 6.85779989e-01 6.90982938e-01
1.12824309e+00 4.11119789e-01 3.44307214e-01 -5.63103497e-01
4.39271957e-01 -4.28600848e-01 -6.02085590e-01 -7.10642487e-02
-4.41639051e-02 -1.70223385e-01 7.43092597e-01 -3.46464396e-01
-4.69351828e-01 5.01441658e-01 -4.35498655e-01 -1.43015826e+00
-6.06710836e-02 1.85694117e-02 -2.52919346e-01 -5.48600495e-01
5.38029850e-01 2.91459382e-01 4.00935002e-02 -6.15871787e-01
3.10185313e-01 5.48827983e-02 7.50484049e-01 -5.99645078e-01
1.46467984e+00 8.48024070e-01 4.41206276e-01 -4.50557023e-01
-5.40920854e-01 -4.14624453e-01 -1.03196585e+00 -3.12550887e-02
5.07982075e-01 -1.19227695e+00 -7.98132300e-01 3.46254557e-01
-1.70417774e+00 -2.19813967e-03 -6.78976595e-01 4.43659052e-02
-7.48311579e-01 5.66762984e-01 -3.51462990e-01 -5.59909821e-01
-6.01472437e-01 -9.91336226e-01 1.67348051e+00 -2.17143193e-01
2.82069117e-01 -8.46573234e-01 1.44189596e-01 -6.89874962e-02
6.69344142e-02 8.16821218e-01 6.18438184e-01 -5.84629960e-02
-1.09544516e+00 -5.44656336e-01 -1.55130789e-01 3.73583376e-01
5.69715649e-02 5.56325056e-02 -7.33804584e-01 -3.50344926e-01
1.64056774e-02 4.85579707e-02 5.21328270e-01 3.73672992e-01
1.49032068e+00 -5.28071046e-01 -4.77460325e-01 1.06436312e+00
1.68239594e+00 -3.16800326e-01 8.23149264e-01 -5.71522862e-02
1.19247854e+00 3.27099234e-01 5.89168549e-01 5.90875685e-01
4.38126534e-01 5.31874955e-01 9.94394064e-01 -2.42490500e-01
-1.00363567e-01 -9.90743756e-01 -5.96553423e-02 6.18908525e-01
-2.26914212e-02 1.05150931e-01 -9.51747417e-01 4.10943389e-01
-1.75726068e+00 -8.13993871e-01 -2.46536002e-01 2.08103204e+00
3.18002582e-01 -1.72796566e-02 -1.00418597e-01 -6.40373863e-03
8.63935173e-01 4.41411398e-02 -8.24469268e-01 1.48799241e-01
9.62235853e-02 1.36633784e-01 5.67681551e-01 5.38905025e-01
-8.70628238e-01 9.00036871e-01 6.21548986e+00 1.22155929e+00
-8.52066100e-01 1.48771219e-02 3.83415610e-01 7.28470385e-02
-6.36357963e-01 2.31351525e-01 -5.07921398e-01 5.54318726e-01
9.72140580e-02 3.91066596e-02 3.54938149e-01 1.07169640e+00
1.55898169e-01 3.44630867e-01 -9.00390983e-01 1.32510602e+00
-8.43519643e-02 -1.71108758e+00 2.82299161e-01 2.64923513e-01
8.38000894e-01 1.62488922e-01 -1.04173101e-01 1.92041516e-01
2.68074453e-01 -9.81371999e-01 7.39381075e-01 9.75627601e-01
1.02714133e+00 -1.08620250e+00 4.78899032e-01 7.85211325e-01
-1.29642141e+00 1.48077995e-01 -6.92326486e-01 9.92914196e-03
2.81454235e-01 8.84710908e-01 -7.69558370e-01 8.97205830e-01
5.94039142e-01 8.13134551e-01 -3.77830982e-01 1.28364754e+00
-6.21558987e-02 9.89923775e-02 -6.84426308e-01 6.06201887e-01
1.08453907e-01 -3.64928603e-01 1.09516168e+00 4.68584776e-01
4.74534512e-01 3.39986086e-02 3.18130821e-01 1.31149864e+00
-2.84635216e-01 -2.55156428e-01 -8.29242527e-01 5.26447773e-01
7.03630209e-01 1.31529987e+00 -3.87368143e-01 -1.13489376e-02
1.23683572e-01 8.55327964e-01 3.88870150e-01 2.44361565e-01
-6.69836164e-01 -3.30954671e-01 7.16109276e-01 5.50534844e-01
3.64207745e-01 -5.35696566e-01 -5.34930825e-01 -1.04347277e+00
4.61549044e-01 -3.63766074e-01 -3.07403356e-01 -1.13398194e+00
-1.49836385e+00 4.98123407e-01 -2.64099717e-01 -1.86506665e+00
1.20231926e-01 -1.04712300e-01 -9.85260010e-01 9.62063432e-01
-1.50188148e+00 -1.54189575e+00 -4.51897502e-01 5.78592718e-01
4.05149132e-01 -2.53624134e-02 4.24530327e-01 9.73642543e-02
9.26967636e-02 4.32539642e-01 8.50774720e-02 1.90204922e-02
3.09539914e-01 -1.03591883e+00 8.77502620e-01 5.73035955e-01
-1.47648662e-01 3.86292994e-01 2.20356837e-01 -1.12718213e+00
-1.21435583e+00 -1.64487517e+00 4.13891524e-01 -6.14751756e-01
1.59766003e-01 -4.73716021e-01 -1.00009203e+00 5.43649375e-01
-5.42378068e-01 2.00592801e-01 -7.61376172e-02 -4.00435001e-01
-1.82680920e-01 -1.22720517e-01 -1.63801479e+00 4.90753472e-01
1.09171546e+00 -3.40414584e-01 -4.27222729e-01 5.78662634e-01
1.29691076e+00 -7.99882054e-01 -9.46802974e-01 7.10770428e-01
2.69604772e-01 -8.62292528e-01 1.48652840e+00 -2.52556443e-01
8.00501943e-01 -5.01159906e-01 2.22558863e-02 -1.36516917e+00
-4.05315667e-01 -8.14528823e-01 -9.14395899e-02 1.09353876e+00
-5.01412116e-02 -4.57935691e-01 1.20703423e+00 3.50785494e-01
-4.99778777e-01 -9.71143007e-01 -1.08964610e+00 -7.57766962e-01
3.98115456e-01 -6.07056320e-01 1.27799404e+00 8.72263968e-01
-7.30028450e-01 -1.47366524e-01 -3.62876862e-01 5.79843402e-01
8.73161018e-01 9.63279419e-03 1.07459998e+00 -1.42881310e+00
1.63865816e-02 -1.96910407e-02 -5.48313677e-01 -1.21670544e+00
1.95985377e-01 -8.11713398e-01 -7.95965791e-02 -1.64613497e+00
-3.28427762e-01 -1.11000001e+00 3.32133442e-01 3.74001443e-01
-1.26446828e-01 2.18835905e-01 3.87381524e-01 7.02470839e-01
-1.92258388e-01 9.56150889e-01 1.71549082e+00 -7.41847232e-02
-2.94571012e-01 3.28303099e-01 -5.02281308e-01 8.63470674e-01
6.39139354e-01 -6.90898478e-01 -2.31892258e-01 -5.14120281e-01
1.94447279e-01 4.06992286e-01 8.25915754e-01 -1.29608965e+00
3.09320748e-01 -1.95745919e-02 8.42698932e-01 -1.30590928e+00
7.13574469e-01 -1.28471315e+00 4.48213220e-01 2.77117699e-01
1.13623708e-01 2.12309524e-01 2.84190625e-01 8.89351964e-01
-8.47152248e-02 -7.38106817e-02 5.78594208e-01 -1.91256359e-01
-1.37936488e-01 1.11396253e+00 5.31113088e-01 -2.38098547e-01
1.09938383e+00 -4.25563246e-01 5.05627580e-02 -3.03840876e-01
-7.47844815e-01 3.82242650e-01 9.49162960e-01 3.65554988e-01
1.11997652e+00 -1.66644275e+00 -7.32640386e-01 4.46010977e-01
-1.23835925e-03 1.21583712e+00 5.81848919e-01 2.59171039e-01
-1.03593040e+00 -2.48196535e-02 2.42798589e-03 -8.43862295e-01
-7.79041529e-01 6.45504117e-01 3.29184175e-01 -1.90669194e-01
-1.08166051e+00 4.49687332e-01 2.92630345e-01 -1.04884481e+00
-6.61737695e-02 -3.50022554e-01 3.46636146e-01 -6.70671523e-01
1.44550860e-01 4.22369808e-01 6.58824283e-04 -6.17364943e-01
2.57993937e-02 8.44401658e-01 1.42309099e-01 1.19213693e-01
1.39922500e+00 5.31381220e-02 -1.11727759e-01 7.33783990e-02
1.24159992e+00 1.49486527e-01 -1.81971872e+00 -5.18352389e-02
-6.34701788e-01 -7.23887503e-01 -1.12323739e-01 -4.73237783e-01
-1.10951018e+00 7.19919264e-01 3.39680254e-01 -2.99462348e-01
7.16668248e-01 -1.62354171e-01 1.14635468e+00 1.69242755e-01
5.50473452e-01 -5.94815016e-01 -1.18869022e-01 4.52844769e-01
1.33748090e+00 -1.08211136e+00 1.75151989e-01 -8.50366771e-01
-2.91587442e-01 1.09976959e+00 5.24335265e-01 -7.26590931e-01
8.61680686e-01 4.02181186e-02 -2.35275179e-01 -3.43374401e-01
-3.44677061e-01 3.91058743e-01 3.12774152e-01 9.92614508e-01
-5.70308208e-01 9.59427953e-02 3.93738389e-01 4.53155130e-01
-6.15942299e-01 1.47774629e-02 3.62795830e-01 7.56955147e-01
-1.72543198e-01 -9.40805495e-01 -7.54980266e-01 4.71925765e-01
2.78625693e-02 -9.07922685e-02 -2.13032126e-01 7.24484265e-01
3.13417673e-01 4.79761064e-01 3.12126458e-01 -4.34727281e-01
5.38825452e-01 -4.59570706e-01 2.07912549e-01 -4.68212992e-01
-4.53751951e-01 -1.12851381e-01 -7.14197397e-01 -5.93259096e-01
6.04982255e-03 -3.98847997e-01 -1.26146758e+00 -4.98231173e-01
2.12517027e-02 9.30779502e-02 8.25934291e-01 5.79722881e-01
8.07713985e-01 3.63487005e-01 9.95815039e-01 -1.52476037e+00
-5.42455852e-01 -6.19511783e-01 -4.99352127e-01 4.32250738e-01
6.34258807e-01 -4.74292815e-01 -4.65690583e-01 -2.70720243e-01] | [8.446930885314941, -3.583808660507202] |
6a7b83e1-106f-462f-8e34-d8c0d38b0160 | stop-overkilling-simple-tasks-with-black-box | 2302.02804 | null | https://arxiv.org/abs/2302.02804v2 | https://arxiv.org/pdf/2302.02804v2.pdf | Stop overkilling simple tasks with black-box models and use transparent models instead | In recent years, the employment of deep learning methods has led to several significant breakthroughs in artificial intelligence. Different from traditional machine learning models, deep learning-based approaches are able to extract features autonomously from raw data. This allows for bypassing the feature engineering process, which is generally considered to be both error-prone and tedious. Moreover, deep learning strategies often outperform traditional models in terms of accuracy. | ['Andrea Albarelli', 'Andrea Gasparetto', 'Alessandro Zangari', 'Matteo Marcuzzo', 'Matteo Rizzo'] | 2023-02-06 | null | null | null | null | ['feature-engineering'] | ['methodology'] | [-3.08389038e-01 -6.58272654e-02 3.20677124e-02 -1.98065311e-01
-1.88741818e-01 -4.78468299e-01 9.18471932e-01 5.39149046e-01
-5.59565008e-01 7.05629647e-01 -4.18217719e-01 2.80864984e-02
-3.00611317e-01 -1.19874358e+00 -3.12461972e-01 -6.81623220e-01
-1.19901054e-01 3.69209707e-01 1.78186253e-01 -3.96513134e-01
3.57441366e-01 8.32148790e-01 -1.86219990e+00 -2.48939902e-01
7.12004364e-01 1.26722002e+00 4.20632660e-02 3.62386465e-01
-3.69259387e-01 5.96721649e-01 -4.31573957e-01 -1.99928582e-01
1.97631761e-01 -1.16988614e-01 -7.47313023e-01 1.01537034e-02
-1.59503147e-01 -5.38841188e-02 -4.49025691e-01 9.54963326e-01
8.36723223e-02 5.72753176e-02 5.15295386e-01 -1.09505641e+00
-3.74561220e-01 -1.37012646e-01 -1.47627056e-01 8.79848283e-03
3.55605543e-01 6.38294071e-02 8.95119548e-01 -9.72844601e-01
4.63887811e-01 9.65678811e-01 7.33574033e-01 3.24445456e-01
-1.11471236e+00 -2.68791348e-01 -1.64539501e-01 3.91035497e-01
-1.14099479e+00 -3.70489091e-01 9.04784679e-01 -4.65269715e-01
7.59816110e-01 -1.69379637e-02 9.83157277e-01 9.59012091e-01
4.87766504e-01 9.06960249e-01 1.15051949e+00 -6.76521301e-01
4.83363062e-01 2.45770574e-01 6.23860769e-02 6.57195389e-01
6.61096275e-01 3.94358784e-01 -3.52266312e-01 2.32572421e-01
7.27353811e-01 3.13159704e-01 6.80556074e-02 -6.83462858e-01
-1.07088697e+00 1.03798234e+00 3.39575231e-01 7.02044785e-01
-5.80647707e-01 3.47256660e-02 4.52525735e-01 4.93023187e-01
5.16226515e-02 7.80546725e-01 -5.19284606e-01 -4.65921402e-01
-6.26896203e-01 4.00609821e-01 8.61292422e-01 5.25002360e-01
1.03235900e+00 3.29619259e-01 3.68891537e-01 4.47615385e-01
1.41948298e-01 1.27485886e-01 7.39163637e-01 -7.50769794e-01
-1.53461844e-01 1.22870159e+00 1.49448097e-01 -1.21089876e+00
-5.55846453e-01 -7.30251610e-01 -1.03964996e+00 5.64607143e-01
3.54651034e-01 -3.36703323e-02 -7.64418364e-01 1.21225643e+00
3.27390879e-01 -2.04990000e-01 1.99893162e-01 5.83168566e-01
4.91739690e-01 4.24176753e-01 -2.82986108e-02 1.99556693e-01
9.36248779e-01 -6.82186842e-01 -7.44046926e-01 -1.62988827e-01
6.57504559e-01 -4.73213434e-01 8.85727644e-01 7.26786911e-01
-6.88524544e-01 -6.00554883e-01 -1.15138447e+00 2.79030576e-02
-7.68649042e-01 -2.07717463e-01 1.12191808e+00 6.09243333e-01
-8.12387943e-01 8.18392813e-01 -9.17238891e-01 -4.25163448e-01
5.79153955e-01 7.39309609e-01 -7.32362628e-01 1.18897848e-01
-1.08452082e+00 1.03464603e+00 7.70458758e-01 3.27467889e-01
-8.90135348e-01 -3.84435028e-01 -7.82577515e-01 1.96724623e-01
5.14056802e-01 -5.10761738e-01 1.21983957e+00 -1.10406852e+00
-1.69711304e+00 5.04915535e-01 -6.65283203e-02 -6.76897526e-01
4.77468759e-01 -5.03308594e-01 -3.77378881e-01 -2.48579159e-02
-4.15411860e-01 2.44346961e-01 9.09631371e-01 -1.09329748e+00
-7.24253893e-01 -2.06159040e-01 9.57131758e-03 -2.09302679e-01
-6.30952656e-01 -2.41674259e-01 6.61457423e-03 -3.17042500e-01
1.57228068e-01 -7.10952044e-01 -5.04370868e-01 -1.97126586e-02
-9.73311290e-02 -3.91626865e-01 1.09320343e+00 5.68098389e-03
7.39738762e-01 -1.98698866e+00 1.51341274e-01 4.76772338e-02
3.72045845e-01 7.18861759e-01 1.88260779e-01 6.04258418e-01
2.31950447e-01 -7.32223615e-02 5.26814796e-02 -1.82981417e-02
-9.85002667e-02 3.69543195e-01 -1.47073761e-01 4.69743460e-01
3.75371873e-01 1.09561026e+00 -1.01354194e+00 -2.10211232e-01
8.25813413e-01 4.62262064e-01 -3.71886969e-01 2.78713912e-01
-3.01535338e-01 6.45598352e-01 -6.38581276e-01 5.54494202e-01
1.98236153e-01 -1.93854332e-01 1.41661063e-01 3.02798033e-01
-2.96164066e-01 4.73819077e-02 -1.02231622e+00 1.59383881e+00
-5.57113171e-01 7.88838863e-01 -1.63672239e-01 -1.57835007e+00
1.17455506e+00 3.81297827e-01 6.20370686e-01 -8.42638731e-01
4.93004858e-01 3.09847981e-01 -2.20799726e-02 -5.38323343e-01
3.26194853e-01 -3.29271138e-01 -1.49334595e-01 7.22251013e-02
1.42958373e-01 -2.38027260e-01 -3.53071801e-02 -4.92244273e-01
9.73388314e-01 1.20027103e-01 6.97084904e-01 -2.91960537e-01
8.78048480e-01 1.33256227e-01 7.19488859e-01 4.31783676e-01
-1.16422221e-01 1.90057352e-01 3.46506774e-01 -1.10676825e+00
-7.75539100e-01 -7.45229900e-01 2.44363435e-02 5.93502581e-01
-4.31223698e-02 -3.13083053e-01 -6.50367260e-01 -5.70820212e-01
1.14646502e-01 2.75178552e-01 -7.43346512e-01 -3.81085157e-01
-4.38907683e-01 -3.27629715e-01 2.11976886e-01 5.29757738e-01
7.44147301e-01 -1.08673024e+00 -1.07328033e+00 4.65655684e-01
3.15957516e-01 -1.15601921e+00 5.31243980e-01 4.37347919e-01
-1.20567179e+00 -9.19348955e-01 -4.76395756e-01 -4.91549402e-01
4.02318001e-01 1.45087585e-01 9.98216808e-01 4.37684923e-01
-6.96000695e-01 2.03086212e-01 -6.40034795e-01 -8.96138668e-01
-3.67262691e-01 5.34104228e-01 1.44163251e-01 1.26617789e-01
8.41488421e-01 -5.74504554e-01 -5.67514420e-01 -6.76791519e-02
-8.48066449e-01 -2.27481589e-01 9.78334963e-01 8.36526692e-01
4.00698513e-01 2.34645754e-01 8.99249315e-01 -4.52951103e-01
5.00171840e-01 -3.69254112e-01 -9.10396576e-01 -4.11275551e-02
-8.68296742e-01 1.48155794e-01 1.00969803e+00 -8.96132737e-02
-6.96914852e-01 1.93601474e-01 -1.79632306e-01 -1.65618569e-01
-6.65966272e-01 6.30837142e-01 -9.25132334e-02 -3.32002103e-01
3.46360594e-01 5.39256930e-01 2.66222805e-01 -5.79993665e-01
-2.72767455e-03 6.35673225e-01 2.45142907e-01 -2.22335309e-01
8.14354599e-01 4.80370671e-01 2.34871209e-01 -1.07094431e+00
-7.52640784e-01 -2.21759200e-01 -1.12411261e+00 -1.87325791e-01
7.90740430e-01 -3.99891168e-01 -7.59251833e-01 6.62477255e-01
-9.52844143e-01 3.21121290e-02 -5.20511746e-01 4.19620275e-01
-4.24712330e-01 1.92157537e-01 -1.84947282e-01 -7.14176059e-01
-9.50028375e-02 -1.01473594e+00 6.32068932e-01 6.26422942e-01
-2.93823987e-01 -1.32818365e+00 1.29583046e-01 -1.79402485e-01
6.54838681e-01 4.81750727e-01 8.98470044e-01 -9.60959315e-01
-4.41044301e-01 -9.79483724e-01 1.38211027e-02 4.41009790e-01
2.86563456e-01 -8.66575837e-02 -1.09361744e+00 -2.32610881e-01
5.42511716e-02 -1.74433455e-01 4.75646883e-01 2.31939889e-02
1.11844349e+00 7.23053589e-02 -3.47298592e-01 3.11309874e-01
1.50649750e+00 4.55716550e-01 4.64466691e-01 7.74797320e-01
2.96768367e-01 5.16400039e-01 5.63081264e-01 5.02474427e-01
2.01834157e-01 5.30870080e-01 7.45349050e-01 -3.05696070e-01
2.58031458e-01 -1.54141724e-01 -2.00022340e-01 5.59122264e-01
-1.59333676e-01 2.49872282e-01 -1.06078660e+00 6.73107266e-01
-1.83442938e+00 -9.30517673e-01 3.59549411e-02 2.08723402e+00
3.54888409e-01 2.55917013e-01 5.45343682e-02 6.40315473e-01
2.42586449e-01 -1.48781434e-01 -5.06132782e-01 -5.38352668e-01
9.83039215e-02 3.46032739e-01 1.51246980e-01 4.28085700e-02
-1.28686464e+00 1.05336809e+00 7.53398371e+00 2.25830644e-01
-1.39704931e+00 -5.14008589e-02 5.74758165e-02 2.77994007e-01
-1.17310481e-02 -1.50739267e-01 -6.10943019e-01 3.97081017e-01
1.03423941e+00 -2.96624452e-01 2.71470428e-01 1.08615971e+00
8.60818997e-02 -3.38701665e-01 -1.32699716e+00 8.68709683e-01
-3.60311717e-01 -1.39710796e+00 6.72631487e-02 3.17299038e-01
4.08531606e-01 -2.12741464e-01 2.67979074e-02 4.88536477e-01
2.54903644e-01 -1.31778181e+00 4.55100656e-01 5.81293106e-01
1.59472719e-01 -9.92095947e-01 1.14877415e+00 5.75015306e-01
-9.71597672e-01 -5.37216008e-01 -4.23009634e-01 -5.84657967e-01
-1.17242187e-01 8.02351713e-01 -6.06589437e-01 6.93439007e-01
7.42256999e-01 6.47692800e-01 -5.24268746e-01 1.16779041e+00
-1.70001730e-01 3.38459104e-01 -1.11776695e-01 -2.50890493e-01
5.05445600e-01 -3.22926700e-01 3.97699863e-01 8.48643541e-01
1.21326469e-01 -2.67678410e-01 2.16239482e-01 6.21569932e-01
1.91340610e-01 -9.11503285e-02 -1.03770447e+00 -2.56316900e-01
2.72317827e-01 1.38100731e+00 -8.69572580e-01 -1.73081443e-01
-5.20292521e-01 8.25196862e-01 3.36232394e-01 -1.55923283e-02
-4.42362189e-01 -7.17718720e-01 7.25558758e-01 6.72290325e-02
4.11299974e-01 -6.93437815e-01 -2.49342889e-01 -8.70768666e-01
-1.48243308e-01 -7.90558934e-01 -3.03944778e-02 2.00542659e-02
-8.81813228e-01 7.74820149e-01 -2.06400231e-01 -1.23295534e+00
-5.76313317e-01 -7.81482160e-01 -5.20970166e-01 5.12475610e-01
-1.72140181e+00 -8.11910152e-01 -4.72327739e-01 4.49513674e-01
5.79814255e-01 -4.42230999e-01 1.19163883e+00 2.29295701e-01
-5.34819186e-01 2.53800452e-01 2.46434957e-01 3.14953387e-01
7.52516091e-02 -1.15614247e+00 3.85832965e-01 6.54291272e-01
1.47663519e-01 6.10423923e-01 9.32839155e-01 -9.35438741e-03
-1.81129742e+00 -7.01907814e-01 6.89265132e-01 -1.38467818e-01
6.82293832e-01 -3.26756328e-01 -9.01416898e-01 4.60259110e-01
1.35729596e-01 3.15007836e-01 5.76257050e-01 1.55469596e-01
-6.28751293e-02 -4.35552508e-01 -1.07004023e+00 3.84907901e-01
5.23103833e-01 -2.79042482e-01 -7.68797100e-01 -5.26384674e-02
1.75924808e-01 1.13209352e-01 -8.77756834e-01 4.51580912e-01
6.28306270e-01 -1.13999152e+00 7.52922475e-01 -7.49325633e-01
4.38181460e-02 -3.45816510e-03 -2.23639347e-02 -1.36228347e+00
-3.01673651e-01 -5.29973328e-01 -3.59311253e-01 9.60608363e-01
-4.29496765e-02 -9.51799870e-01 9.31824803e-01 6.92995369e-01
1.50696309e-02 -8.33103180e-01 -1.05857670e+00 -1.21091557e+00
2.33188972e-01 -2.56063253e-01 8.22031856e-01 7.35141873e-01
-1.02182239e-01 1.44206226e-01 -1.81788858e-02 -2.37793669e-01
4.94181693e-01 2.51127094e-01 7.97000885e-01 -2.13030148e+00
-5.54283224e-02 -7.64177203e-01 -1.01669836e+00 -5.71141005e-01
3.84383261e-01 -5.03470719e-01 1.07287273e-01 -1.72266066e+00
-2.76085496e-01 -5.15627265e-01 -4.26095605e-01 5.61979592e-01
1.22039944e-01 1.49704292e-01 -1.11717880e-01 1.29645184e-01
-3.15082282e-01 6.46831810e-01 9.92201746e-01 2.85880789e-02
-2.39959672e-01 1.17953412e-01 -5.61290324e-01 1.12772775e+00
1.08167291e+00 -4.11191434e-01 -2.84298390e-01 -3.40913832e-01
3.42032373e-01 -3.44512284e-01 2.33111382e-01 -1.28472507e+00
1.40434295e-01 -2.56641865e-01 4.03217942e-01 -9.10901427e-02
2.68233985e-01 -1.27341270e+00 -2.11123511e-01 7.87432194e-01
7.74887204e-02 -3.31921652e-02 1.78750977e-01 5.25055289e-01
-6.08024597e-01 -3.45018625e-01 8.01998019e-01 -3.25055629e-01
-1.05451167e+00 2.38650665e-01 -7.83993840e-01 -4.80777472e-01
1.57137513e+00 -5.27290642e-01 1.66753396e-01 -1.51713938e-01
-5.79623878e-01 -9.98461470e-02 4.37136710e-01 4.01930660e-01
5.89116454e-01 -1.07662141e+00 -2.64859915e-01 3.27903658e-01
3.59653309e-02 2.11760998e-01 -4.53851297e-02 7.27582693e-01
-5.79111099e-01 6.85914516e-01 -5.15587389e-01 -5.56583405e-01
-8.71395290e-01 8.43386769e-01 3.91489118e-01 -1.93453148e-01
-1.00316226e+00 5.12971997e-01 -1.77640438e-01 -8.83846954e-02
1.95319429e-01 -3.02597377e-02 -3.35663259e-01 6.36041313e-02
5.70498884e-01 4.92585689e-01 2.34098315e-01 -2.72915483e-01
-3.39400321e-01 5.00998020e-01 -2.52191693e-01 6.70394301e-02
1.55775940e+00 9.52240750e-02 -6.45014867e-02 6.24002576e-01
9.37319458e-01 -3.46715122e-01 -1.04972148e+00 -3.67847234e-02
7.53314614e-01 -3.78229856e-01 3.31113130e-01 -5.28729081e-01
-1.03944683e+00 1.09428692e+00 5.99072874e-01 5.92996597e-01
1.02767003e+00 -3.86611938e-01 8.07218075e-01 7.32758343e-01
7.41065443e-01 -1.25674498e+00 2.16930993e-02 5.60308576e-01
7.18941391e-01 -1.52455235e+00 -1.01214424e-01 3.19659039e-02
-8.18846971e-02 1.46540701e+00 5.04498541e-01 -4.05142456e-01
7.55201995e-01 3.89095217e-01 2.79511753e-02 -2.32255161e-01
-5.83237708e-01 -2.17539623e-01 1.13711789e-01 6.72557175e-01
4.88996685e-01 -2.56225109e-01 -2.97623098e-01 4.36079383e-01
-5.87560609e-02 3.25635016e-01 4.08575237e-01 1.20495331e+00
-7.66967118e-01 -1.31708217e+00 -2.15904921e-01 2.72936314e-01
-4.02846783e-01 4.48438019e-01 -4.52532917e-01 1.11478841e+00
2.22960804e-02 9.27260876e-01 -8.07229988e-03 -3.85830790e-01
3.39561880e-01 1.28085017e-01 3.91440332e-01 -5.67465365e-01
-5.83243787e-01 -4.07591581e-01 -3.26729327e-01 -5.53267181e-01
-3.08462262e-01 -5.19000053e-01 -1.31240368e+00 -2.53046095e-01
-3.27851802e-01 3.43036324e-01 8.13237250e-01 1.31917155e+00
5.32922387e-01 5.33882201e-01 7.95481384e-01 -1.00113356e+00
-6.22301340e-01 -8.33637953e-01 -4.80902672e-01 3.38242128e-02
3.75447184e-01 -1.04488242e+00 -1.12839185e-01 -1.85834318e-01] | [8.504777908325195, 2.715895414352417] |
e6bcd16a-1449-4e0d-a919-64bf58c1891d | self-supervised-learning-of-occlusion-aware | 2108.03893 | null | https://arxiv.org/abs/2108.03893v3 | https://arxiv.org/pdf/2108.03893v3.pdf | Self-supervised Learning of Occlusion Aware Flow Guided 3D Geometry Perception with Adaptive Cross Weighted Loss from Monocular Videos | Self-supervised deep learning-based 3D scene understanding methods can overcome the difficulty of acquiring the densely labeled ground-truth and have made a lot of advances. However, occlusions and moving objects are still some of the major limitations. In this paper, we explore the learnable occlusion aware optical flow guided self-supervised depth and camera pose estimation by an adaptive cross weighted loss to address the above limitations. Firstly, we explore to train the learnable occlusion mask fused optical flow network by an occlusion-aware photometric loss with the temporally supplemental information and backward-forward consistency of adjacent views. And then, we design an adaptive cross-weighted loss between the depth-pose and optical flow loss of the geometric and photometric error to distinguish the moving objects which violate the static scene assumption. Our method shows promising results on KITTI, Make3D, and Cityscapes datasets under multiple tasks. We also show good generalization ability under a variety of challenging scenarios. | ['Guizhong Liu', 'Jiaojiao Fang'] | 2021-08-09 | null | null | null | null | ['3d-geometry-perception'] | ['computer-vision'] | [-0.18027233 -0.13399471 -0.14541802 -0.7032686 -0.7337041 -0.45668903
0.29760435 -0.510343 -0.32250574 0.7340088 0.29001442 0.10816291
-0.14109047 -0.5962165 -0.6722056 -0.67765427 0.18418522 0.28343746
0.19556312 0.21972351 0.15531862 0.55796605 -1.5534258 -0.01798168
1.0570085 0.987519 0.1783243 0.37647423 -0.1677506 1.0653807
-0.08448693 -0.05627716 0.6181986 -0.11264757 -0.81647414 0.50544983
1.1233104 -0.8240388 -0.61144334 0.88287115 0.54834646 0.43625215
0.19098248 -1.4219958 -0.40713224 -0.18765318 -0.77015084 0.35973427
0.44522023 0.40115568 0.72267836 -0.9002245 0.84629935 1.3761191
0.43709004 0.56599814 -0.96432257 -0.5733257 0.656105 0.32026875
-1.1587917 -0.5206979 1.005929 -0.6226726 0.8362468 -0.20117193
0.56884694 1.0623535 -0.12647532 0.8358236 1.0755367 -0.17758594
0.13813335 0.12938862 0.06838111 1.1202401 0.24007906 0.25866598
-0.6254105 0.17317045 0.93098736 0.09034711 -0.4056834 -0.8544772
-1.078804 0.65420693 0.41924155 -0.12735921 0.10518733 0.16725762
0.15539981 0.01182891 0.88566226 0.0467778 -0.7129681 0.11165955
-0.5927164 0.02191819 0.48883045 0.99197096 1.112913 0.27420124
0.02325238 0.52075416 0.48804983 0.7249124 0.09617666 -1.4401217
0.90110093 0.5659189 0.30546984 -1.1868666 -0.47751585 -0.53606325
-0.6832118 0.5291924 0.62534744 -0.17485182 -0.8099296 1.9434639
0.80945534 0.45672265 -0.06114772 1.1617571 0.89205647 0.4134318
-0.18552993 -0.21626051 0.6908944 -1.2411642 -0.76461124 -0.44967267
0.62617856 -0.6332757 0.9924011 0.05621332 -1.1413454 -0.6870639
-0.9103102 -0.49768218 0.04530677 -0.00672406 0.8068065 0.48792806
-0.73624355 0.3751006 -1.0188147 -0.24920122 0.5865371 0.01360081
-0.49048153 -0.3959241 -0.9115731 0.75119 0.04645349 0.2587079
-0.9315896 -0.92176896 -1.1644484 -0.20992295 0.3632157 -0.9684927
0.69287604 -0.78855467 -1.5194181 1.1623873 -0.23298366 -0.11025154
1.0307567 -0.53077346 -0.0390082 0.17291813 0.2981942 0.82813007
0.60616857 -1.1689212 -0.74696904 -0.6433734 0.2406428 0.538009
-0.15078199 -0.6349986 -0.3796843 -0.41210636 0.5368179 -0.8345237
-0.19777168 0.5931277 -0.24478841 0.01167114 0.9000101 -0.48537466
0.70055485 -2.1588345 0.03489261 -0.26223117 0.20013152 0.01940149
-0.31019592 -0.18326406 0.13910972 -0.26043144 -0.22525299 -0.76615757
-0.26510736 0.27704751 -0.3640686 0.7987175 0.07224091 0.79772717
-1.2151035 -0.48867655 0.7492954 0.5525616 -0.5913703 0.36746374
-0.1082638 0.9462828 -0.6177224 0.47571984 1.05167 -0.22435652
-0.5417032 -0.2358083 -0.1431138 0.3294389 -1.5020655 2.2588918
-0.41701224 0.7816832 0.10298137 -0.79751354 0.81813115 -0.0884398
0.6738496 -0.7855623 -0.0179559 0.07815983 -0.5341775 -0.8226393
0.1089795 0.09275463 0.37580216 0.3977541 -0.05256715 -0.34189153
-0.15973474 0.07516422 0.8474695 0.56083554 -0.04659839 -0.18022466
0.7673883 -0.10879266 1.0619525 0.47525206 -0.6824772 0.9434336
0.0627714 -0.85040903 -0.79816145 -1.2503711 -0.12497196 0.46814248
0.613671 0.16066772 -0.24012034 -0.8600998 -0.01171395 0.3711598
-0.4618138 0.03626174 -0.6470391 -0.5916011 0.09575217 0.4387966
1.1032225 -0.5768672 -0.4340531 0.02299485 -0.6262916 -1.6439753
-0.6196806 -0.22132874 -0.9503274 -1.1705433 -0.50051576 -0.7229584
0.7145799 0.82086724 0.9346533 -0.19149716 -0.32226646 0.5580408
0.01709731 -0.05617663 0.16331974 -0.01985833 0.01758782 0.21434116
0.03916185 -0.66643125 -0.9368501 0.501518 -0.55697733 0.18667567
-0.01686733 0.56056035 0.72830695 -0.07128376 0.09839803 -0.58141124
-0.21259241 -0.09486894 -0.83679533 0.06159304 -0.36996004 0.07427182
0.15177928 -0.28320968 -1.5535307 0.38637006 0.09089792 -0.6702121
-0.22295074 -0.15720777 -0.5108067 -0.24961664 0.4491483 -0.1346906
-0.14317702 -0.3048345 0.4085506 0.17166798 0.3883854 -0.4710494
0.9219124 1.3119164 0.26705888 -0.54832757 -1.5700823 -0.62700623
-0.801002 -0.30526483 1.1676143 -1.2753054 -0.5606518 0.6310747
-1.2689098 -0.44450298 -0.37950227 0.8049345 -0.6613558 0.64308506
-0.4510846 -0.8072982 -0.25069633 -1.0697086 1.1189823 0.11407539
0.19846697 -1.1203135 0.15008564 0.6691062 0.10293179 0.61430335
0.38215718 0.22900935 -1.1704783 0.22127306 -0.51030993 0.38356557
0.30357414 -0.15397178 -1.2014737 -0.3389423 0.2915518 -0.29193214
0.8859146 0.6183128 1.0273628 0.08401138 -0.18847752 1.2832903
1.38142 -0.01807609 0.47620106 0.34336686 1.1518337 0.8622506
0.57211834 0.31553927 0.59262794 0.5322238 0.6092487 -0.23834096
-0.27760705 -0.3958812 0.17244005 0.4809618 0.06855839 -0.27968875
-0.64701027 0.44683504 -1.9359674 -0.9659321 -0.2557308 2.2350245
0.36640668 0.1315262 -0.33650485 -0.09437766 0.55007905 0.52680665
-0.921432 0.2959122 -0.42094043 -0.28402472 0.43973875 0.9213476
-1.2543297 1.0048488 5.219976 0.38124952 -1.0292407 0.20359868
0.75747937 -0.25621077 -0.36844024 0.02315825 -0.8908971 0.20733094
0.07211806 0.24599107 0.18736131 0.6994012 0.46171346 -0.32029733
-1.111983 1.3571631 0.23676448 -1.2017395 -0.10100907 0.07619861
1.1142864 0.16966814 0.05086168 -0.22499043 0.2565377 -0.5905258
0.6663949 0.5306897 0.5800558 -0.41507044 0.4237382 0.21676995
-1.1066338 -0.07719497 -0.29962024 -0.26971635 0.6609371 0.841132
-0.03268939 0.49773192 0.879576 1.205576 -0.21567082 1.0981947
-0.4000689 0.19193383 -0.39582387 0.5536702 0.30337915 -0.42682815
0.74032974 0.66097313 0.02768008 0.20007396 0.3556238 0.89825946
0.13853107 -0.13979174 -0.5555927 0.7483513 0.16486673 1.0561483
-0.5051033 -0.12857996 -0.48835602 1.1184143 0.51699907 0.6345125
-0.73775405 0.15114342 1.025167 0.18434681 0.05661743 -0.4522573
-0.34233102 -1.7203096 0.2660148 -0.14972807 0.42934823 -0.96507746
-1.2246696 0.35127828 -0.03416362 -1.4678426 0.15962987 -0.4771297
-0.5253401 0.54820925 -1.9965143 -0.89658237 -0.94225556 0.71205586
0.6832909 0.10135162 0.28655484 0.542746 -0.6483246 0.30816498
-0.07144511 0.17214441 0.8012617 -0.9845753 0.18814419 0.8570555
0.06711496 0.13096505 0.3821355 -0.33303738 -1.0089251 -1.2648453
0.80755764 -0.5537141 0.2926942 -0.43340302 -0.8550542 0.61058325
-0.12646814 0.49441805 0.1481782 -0.3656071 -0.43248725 -0.51403314
-1.1941409 0.3365375 1.7352589 -0.4098564 -0.3078572 0.536151
0.9635199 -0.74504185 -0.47036934 0.44720855 0.41918495 -1.0894175
1.1835821 -0.5062152 0.3102536 -0.46782032 -0.20859985 -0.9506872
-0.0695482 -0.7198555 -0.05634637 1.190998 0.04652767 -0.76247007
1.1549599 0.86932737 -0.17715292 -0.447036 -1.0586902 -0.65378827
-0.03807747 -0.4563426 0.32787105 1.0352453 -0.5827412 0.30188575
-0.5994456 0.26811728 1.2016398 0.18000543 1.0666984 -1.1460735
-0.09797901 -0.14200632 -0.4823804 -1.5394969 0.3920999 -0.55317146
-0.12039679 -1.510889 0.05791716 -0.5587211 -0.05474168 0.2013577
-0.15033625 -0.02441862 0.01534864 0.14326297 -0.6715587 0.8690489
1.6531391 -0.24460489 -0.37705103 -0.06628577 -0.19307412 0.8075486
0.5972801 -0.35725734 -0.78365916 -1.0588522 0.10681522 0.07752986
0.53717846 -0.8740456 0.18100198 -0.3789071 0.43319973 -0.8488474
0.4439195 -0.8413243 -0.12032115 0.34742227 -0.24675693 -0.05969281
0.07754768 0.79396445 -0.07906467 0.2586725 0.8982291 -0.19787693
-0.90739155 0.9147789 0.23038054 0.43621483 0.94522554 -0.36748698
-0.21315455 -0.40885136 -0.55183524 0.5134737 0.3938242 0.51252806
0.69340616 -1.1299838 -0.701175 0.4622917 0.16887027 0.40960327
0.64466035 0.62666583 -0.6683175 0.39886022 -0.13721524 -0.9322255
-0.9602305 0.44504014 0.7611143 -0.08718172 -0.74388343 0.8632377
0.6666999 -0.5908878 0.49467525 -0.3632504 0.06245727 -0.29641184
0.44946146 0.61859375 -0.11322827 -0.6958994 -0.38860252 1.085258
0.12614858 -0.02809164 1.2967249 -0.8042608 0.18851589 0.27548572
1.6278375 -0.21429928 -2.0460145 -0.4354979 -0.4675858 -1.0000163
0.2790047 -0.36342755 -1.5467459 1.0744532 0.7984789 -0.42321077
0.8782132 -0.12480668 0.8575236 0.30897298 0.40121555 -0.88323206
0.27433658 0.5268192 0.5831408 -1.7228898 0.07096885 -0.65444565
-0.31201217 0.84732914 1.1105644 -0.15658875 0.7813557 -0.06617343
0.33849144 -0.06897124 -0.48926368 -0.41722813 0.21908094 0.7844382
0.01427157 -0.40123782 -0.06494574 -0.29408705 0.13656235 -0.18270405
0.3749412 0.67852926 -0.23421721 -0.7407542 -0.13215066 -0.0756349
-0.09603573 0.21717322 0.02816969 0.77630925 0.22733386 0.9553276
0.19202238 0.01733283 0.41782483 -0.28110468 0.70658463 -0.41454327
0.13144235 -0.08260357 -0.16310519 -0.8703789 -0.87670755 -0.66159314
-1.1172757 -0.21236522 -0.31180987 -0.33961916 0.38455608 1.0262927
0.43395546 0.09694443 0.9709115 -0.8899358 -0.08279307 -0.5195508
-0.4211632 0.81669885 0.73757577 -0.9067626 -0.8677935 0.0641152 ] | [8.672439575195312, -2.167828321456909] |
6b62b94d-73cb-4d3b-aa9b-63a750308102 | leveraging-large-language-models-in | 2305.07961 | null | https://arxiv.org/abs/2305.07961v2 | https://arxiv.org/pdf/2305.07961v2.pdf | Leveraging Large Language Models in Conversational Recommender Systems | A Conversational Recommender System (CRS) offers increased transparency and control to users by enabling them to engage with the system through a real-time multi-turn dialogue. Recently, Large Language Models (LLMs) have exhibited an unprecedented ability to converse naturally and incorporate world knowledge and common-sense reasoning into language understanding, unlocking the potential of this paradigm. However, effectively leveraging LLMs within a CRS introduces new technical challenges, including properly understanding and controlling a complex conversation and retrieving from external sources of information. These issues are exacerbated by a large, evolving item corpus and a lack of conversational data for training. In this paper, we provide a roadmap for building an end-to-end large-scale CRS using LLMs. In particular, we propose new implementations for user preference understanding, flexible dialogue management and explainable recommendations as part of an integrated architecture powered by LLMs. For improved personalization, we describe how an LLM can consume interpretable natural language user profiles and use them to modulate session-level context. To overcome conversational data limitations in the absence of an existing production CRS, we propose techniques for building a controllable LLM-based user simulator to generate synthetic conversations. As a proof of concept we introduce RecLLM, a large-scale CRS for YouTube videos built on LaMDA, and demonstrate its fluency and diverse functionality through some illustrative example conversations. | ['Zhenning Tan', 'Manoj Tiwari', 'Zexi Chen', 'Brian Chu', 'Harsh Lara', 'Ajay Patel', 'Gabriel Schubiner', 'Jun Xie', 'Changbo Long', 'Hakim Sidahmed', 'David Allen', 'Sameer Ahuja', 'Luke Friedman'] | 2023-05-13 | null | null | null | null | ['dialogue-management', 'common-sense-reasoning'] | ['natural-language-processing', 'reasoning'] | [-7.58990496e-02 4.66810405e-01 -5.98072931e-02 -5.84929824e-01
-5.09616375e-01 -8.76004815e-01 7.34267175e-01 -2.87948728e-01
-5.27797379e-02 6.89529061e-01 8.99047554e-01 -4.42414701e-01
-7.69321248e-02 -5.27879477e-01 -3.03558469e-01 2.25227252e-01
-2.90341936e-02 6.40948296e-01 -6.64500578e-04 -8.06111753e-01
1.84580758e-01 1.98762730e-01 -1.74344993e+00 1.01461470e+00
8.68480742e-01 3.88366640e-01 6.26153350e-01 1.12186599e+00
-4.37354475e-01 8.60233486e-01 -6.65354550e-01 -8.32982734e-02
-9.00371596e-02 -3.58107448e-01 -1.01561463e+00 1.05073899e-02
1.83925316e-01 -4.81272042e-01 8.34154114e-02 -2.13336665e-02
5.31076014e-01 5.80944419e-01 8.56670514e-02 -1.31310463e+00
-5.97890079e-01 8.05532455e-01 3.57601494e-01 -1.36171222e-01
1.20969379e+00 2.16761708e-01 1.05601573e+00 -6.11058772e-01
6.69637799e-01 1.52879107e+00 4.87408519e-01 1.08870471e+00
-1.24703920e+00 -3.10113817e-01 3.86897594e-01 -2.79560030e-01
-8.62115383e-01 -7.38971531e-01 1.32541060e-01 -2.24224165e-01
1.21265483e+00 9.75946307e-01 7.87307024e-01 1.39601457e+00
-2.87745088e-01 8.33308220e-01 8.40663493e-01 -4.42732990e-01
8.36475119e-02 8.59807551e-01 1.98921636e-01 5.76968491e-01
-4.87737477e-01 -3.06275606e-01 -8.87352109e-01 -4.29893464e-01
5.70824683e-01 -7.37300515e-02 -5.12575984e-01 -1.49296850e-01
-1.21776044e+00 6.40745878e-01 -4.85590994e-02 1.49133593e-01
3.47161666e-02 -2.45422587e-01 1.87156707e-01 6.30455673e-01
2.80813217e-01 9.59143400e-01 -7.04096854e-01 -8.46539438e-01
-3.78452957e-01 3.22958291e-01 1.67667902e+00 1.11318123e+00
3.27439725e-01 -5.06974638e-01 -4.03597951e-01 1.20604575e+00
5.00581920e-01 4.51667368e-01 6.31080508e-01 -1.30263054e+00
2.47520745e-01 6.43618941e-01 4.83079314e-01 -7.92531490e-01
-6.43296599e-01 -1.76517069e-01 -1.19230926e-01 -1.91905558e-01
3.01082611e-01 -3.05492371e-01 -1.01240352e-01 1.79613733e+00
3.08469862e-01 9.99970958e-02 2.98751861e-01 7.32738018e-01
9.22994614e-01 7.60252357e-01 -4.46345052e-03 -9.42665562e-02
1.32394528e+00 -1.14853466e+00 -5.30693114e-01 -3.86001691e-02
7.26008594e-01 -7.18140602e-01 1.83988464e+00 4.67123777e-01
-1.00982845e+00 -4.01541293e-01 -6.60509765e-01 -1.85582250e-01
-2.91622490e-01 -1.11654751e-01 8.22557151e-01 7.63487577e-01
-1.18404925e+00 3.46101642e-01 -4.55462962e-01 -9.04815912e-01
-2.43805319e-01 4.70526546e-01 -2.21756399e-01 1.43845573e-01
-1.29348195e+00 8.46293569e-01 -3.17194015e-01 -1.32140711e-01
-2.53029853e-01 -8.60243499e-01 -6.72806919e-01 1.22380324e-01
2.69386232e-01 -8.28202665e-01 1.86659145e+00 -7.40316808e-01
-2.39960742e+00 3.65913630e-01 -1.56434581e-01 -1.20217010e-01
3.42434883e-01 -4.17529285e-01 -5.60607135e-01 -2.24864930e-01
-2.06560478e-01 5.93509614e-01 2.71302342e-01 -1.09377384e+00
-6.16378486e-01 2.65417248e-01 8.57196450e-01 6.01759434e-01
-5.33027470e-01 -3.29156034e-02 -5.35500944e-01 -6.52455762e-02
-5.95479071e-01 -1.18620789e+00 -2.69078523e-01 -2.04779893e-01
-9.35435817e-02 -1.38192773e-01 7.05775440e-01 -3.04226518e-01
1.43687296e+00 -1.89402163e+00 6.38017207e-02 2.44955018e-01
5.64172566e-02 4.66623664e-01 -3.55324596e-01 9.17349398e-01
5.00114441e-01 3.47075015e-01 5.38674474e-01 -7.21190870e-01
2.15386569e-01 2.37992853e-01 -2.32903853e-01 -3.35977763e-01
-4.40621912e-01 6.95254385e-01 -9.48527575e-01 -2.39457106e-04
3.54748219e-01 5.15799820e-01 -1.16136134e+00 6.46209359e-01
-5.31479955e-01 6.65700495e-01 -3.24852765e-01 4.33366969e-02
1.04207523e-01 -4.69488502e-01 5.42727709e-01 3.70094888e-02
-2.83308864e-01 7.36506402e-01 -1.14656317e+00 1.86386204e+00
-1.49978840e+00 5.81651986e-01 3.63682866e-01 -3.17999870e-02
7.16866314e-01 5.25689363e-01 2.68199176e-01 -5.57702482e-01
-1.82011828e-01 -1.37119383e-01 -2.50075310e-01 -6.91505849e-01
8.55624795e-01 3.20831746e-01 -2.25580812e-01 1.09924698e+00
-2.10835245e-02 -8.65495428e-02 -1.89317185e-02 6.60703540e-01
9.61319983e-01 5.85569330e-02 2.51203775e-01 -1.76236048e-01
5.68960667e-01 -9.61368233e-02 -6.77802116e-02 9.77478802e-01
2.21487805e-01 3.36593807e-01 4.15189601e-02 -2.57725030e-01
-4.78205711e-01 -8.11688602e-01 1.66135028e-01 1.75798523e+00
-4.92056906e-02 -1.02328181e+00 -6.66770279e-01 -4.42864835e-01
-2.41483804e-02 1.08726859e+00 -5.91009744e-02 8.87871534e-02
-3.11827183e-01 -3.87957215e-01 3.22342187e-01 -4.93334047e-03
3.57854992e-01 -1.11236370e+00 -3.81699592e-01 3.25236231e-01
-5.71834385e-01 -1.21865773e+00 -8.39078844e-01 -5.49257696e-01
-4.31071699e-01 -7.24624097e-01 -1.35351911e-01 -3.84503007e-01
2.99250156e-01 7.14559734e-01 1.46519780e+00 2.43669748e-01
-3.83268408e-02 1.02016926e+00 -3.80594850e-01 -1.10962853e-01
-8.64124417e-01 3.07020664e-01 2.00667948e-01 -1.35827035e-01
1.50784001e-01 -7.97482848e-01 -7.70890534e-01 7.11498737e-01
-4.52937067e-01 5.66951990e-01 1.20371267e-01 5.47818422e-01
-2.81298757e-01 -7.70785153e-01 8.65992725e-01 -1.28976905e+00
1.37265527e+00 -6.55183315e-01 -1.01419561e-01 2.08218858e-01
-5.37237585e-01 -5.85699640e-02 6.14151597e-01 -4.42676693e-01
-1.44034040e+00 -2.25879624e-01 -2.26573363e-01 4.63541925e-01
-2.06785381e-01 4.56128418e-01 -8.55340660e-02 1.59143791e-01
7.87925184e-01 -1.13678657e-01 3.46336067e-01 -4.50022042e-01
9.09382403e-01 1.36599398e+00 1.75904140e-01 -8.37010086e-01
2.27843851e-01 1.08552970e-01 -8.75261843e-01 -9.10054147e-01
-7.22980976e-01 -7.22058296e-01 -2.55070925e-01 -2.87735850e-01
4.99515474e-01 -1.08196008e+00 -1.45082390e+00 -1.55940294e-01
-9.92584348e-01 -8.13355565e-01 -1.19889960e-01 2.99681187e-01
-6.39823616e-01 1.00447938e-01 -5.62785625e-01 -9.38699841e-01
-5.74496150e-01 -1.04579294e+00 7.63243198e-01 4.30395424e-01
-1.17983782e+00 -1.07054329e+00 1.57297909e-01 8.78991365e-01
1.08627295e+00 -5.14050603e-01 5.21524429e-01 -8.55807066e-01
-5.84813893e-01 -9.71536413e-02 1.24520153e-01 -1.78113170e-02
2.65820861e-01 -2.81891041e-02 -1.06023705e+00 -3.05883437e-01
-4.35567617e-01 -3.79221946e-01 -9.40982029e-02 -2.12832704e-01
8.48439157e-01 -5.19506931e-01 -3.99083734e-01 2.01808855e-01
7.00643480e-01 4.32540849e-02 3.95924807e-01 6.37805611e-02
4.03051525e-01 5.64864397e-01 4.15428996e-01 5.01584888e-01
9.81862128e-01 1.01297987e+00 -1.30942225e-01 7.88835958e-02
-3.91401127e-02 -4.30249512e-01 6.55628383e-01 1.02449203e+00
-9.37872902e-02 -6.82937324e-01 -4.11769271e-01 1.58999190e-01
-1.96726561e+00 -1.11367941e+00 1.70258567e-01 2.30325103e+00
9.05262172e-01 -4.23796065e-02 2.06643850e-01 -6.71461582e-01
2.32857734e-01 3.30160335e-02 -2.80282319e-01 -9.75176692e-01
4.26893771e-01 -8.69405363e-03 -7.46994317e-02 1.02567518e+00
-4.60066348e-01 9.00575876e-01 6.52320194e+00 3.40154946e-01
-9.29562628e-01 -2.62951944e-05 1.74096674e-01 -3.22001785e-01
-6.76745236e-01 -1.06822707e-01 -8.24588895e-01 2.92707890e-01
1.40526021e+00 -2.74874181e-01 9.87906814e-01 7.96136022e-01
8.21275830e-01 3.65944579e-02 -1.26684177e+00 7.73880899e-01
1.22646382e-02 -1.59296072e+00 1.17815770e-01 -4.24277820e-02
3.68092537e-01 1.15261115e-01 -2.21185461e-01 7.74073184e-01
7.31536210e-01 -9.22264457e-01 1.96832001e-01 6.54087484e-01
6.73026323e-01 -4.36171502e-01 2.13534683e-01 6.90996468e-01
-9.79516864e-01 -1.23556577e-01 9.72560346e-02 -4.32971358e-01
3.71861368e-01 -6.41822517e-02 -1.27623117e+00 9.69440565e-02
4.70439047e-01 4.47646350e-01 -1.53864503e-01 5.62514544e-01
2.41653338e-01 5.19056737e-01 -5.25968790e-01 -3.39198768e-01
-2.01871797e-01 -1.42325208e-01 4.76130813e-01 1.44456506e+00
2.50175804e-01 2.96678692e-01 4.11566347e-01 6.20660126e-01
-1.96078539e-01 2.19545275e-01 -5.44310331e-01 -2.41508763e-02
8.22054148e-01 1.48303306e+00 -6.66254386e-02 -3.07976663e-01
-6.36569202e-01 1.04829121e+00 3.04549128e-01 2.72377580e-01
-4.27535236e-01 8.78442898e-02 1.21960080e+00 1.46705821e-01
-3.71978045e-01 -3.16196948e-01 1.05354190e-01 -1.48322165e+00
-3.28820169e-01 -1.41293526e+00 1.47682667e-01 -7.76161730e-01
-1.07722318e+00 7.06719637e-01 -9.84014571e-02 -1.03346133e+00
-6.61446154e-01 -2.72574484e-01 -7.27709174e-01 8.77492368e-01
-9.35447395e-01 -1.04528916e+00 -5.22883058e-01 5.49964666e-01
1.06592083e+00 -2.78148830e-01 1.57198167e+00 2.26504266e-01
-2.80532986e-01 6.40223801e-01 -4.07001786e-02 -5.46831965e-01
9.80079949e-01 -1.31358421e+00 5.82615435e-01 -5.94453479e-04
1.51381195e-01 1.18040907e+00 8.93238485e-01 -2.96997190e-01
-1.54151165e+00 -6.57453239e-01 8.25552642e-01 -9.47367728e-01
5.32882512e-01 -8.98590505e-01 -7.01022327e-01 7.69634545e-01
3.59680891e-01 -5.42397320e-01 1.28154635e+00 7.73002446e-01
-5.11818342e-02 3.14910784e-02 -1.13304424e+00 1.10110414e+00
1.36201310e+00 -7.44561315e-01 -4.00702477e-01 5.44378221e-01
9.21434700e-01 -5.91708243e-01 -9.13509727e-01 -2.26573527e-01
1.09027171e+00 -9.55498755e-01 8.49996805e-01 -7.60559618e-01
2.47311573e-02 -6.52475581e-02 -1.59037299e-02 -1.59622169e+00
-4.99425903e-02 -1.47715926e+00 -2.61029620e-02 1.40446782e+00
7.41727889e-01 -6.26057923e-01 5.65616310e-01 1.45139754e+00
-1.12120628e-01 -4.80812430e-01 -1.85405955e-01 -3.54275927e-02
-3.64450544e-01 -5.78365624e-01 7.00360477e-01 7.09300935e-01
9.38520193e-01 7.24453747e-01 -5.76385498e-01 1.05780251e-01
-1.35036796e-01 8.53378847e-02 1.33621109e+00 -1.08464956e+00
-7.62205482e-01 -3.85461450e-02 3.34598571e-01 -1.60788298e+00
-5.50819114e-02 -7.71152377e-01 -8.90381411e-02 -1.62506580e+00
-7.17771500e-02 -7.14804411e-01 1.88957646e-01 2.04301789e-01
2.06340000e-01 -1.64129838e-01 4.83582795e-01 1.78735048e-01
-9.45560634e-01 2.99139589e-01 1.22902632e+00 2.85091043e-01
-8.65345180e-01 5.28793514e-01 -1.09113586e+00 6.75457716e-01
8.23913813e-01 6.75972402e-02 -1.04161561e+00 -2.71032423e-01
2.67908514e-01 2.83826500e-01 -7.48017132e-02 -8.25871646e-01
3.29634249e-01 -1.99326143e-01 -1.47333384e-01 1.28681688e-02
6.20784998e-01 -6.76346540e-01 1.73666716e-01 -2.71717936e-01
-8.99163246e-01 2.16214582e-02 2.85086036e-01 5.78147352e-01
2.51231492e-01 1.65738419e-01 8.68311450e-02 -2.74560690e-01
-6.45486891e-01 -1.68629140e-01 -9.65318799e-01 -1.25685990e-01
6.75120950e-01 1.02069065e-01 -5.78169346e-01 -1.18349993e+00
-9.90359604e-01 5.52034259e-01 3.38936210e-01 1.04461300e+00
2.77858257e-01 -8.73488903e-01 -2.85692304e-01 3.29598874e-01
1.75951511e-01 -4.19419467e-01 3.63753557e-01 3.35900545e-01
-2.01156542e-01 4.60602254e-01 3.79176140e-02 -3.97207141e-01
-1.49416339e+00 1.41159222e-01 3.68991852e-01 -2.21101567e-01
-5.21692991e-01 5.96784592e-01 -6.20691031e-02 -1.20357132e+00
2.20974147e-01 -3.25479478e-01 -4.94704813e-01 -1.72251701e-01
9.04104590e-01 -2.21397430e-02 -4.87974361e-02 -1.61000565e-01
-7.08481446e-02 -2.65181232e-02 -6.42099530e-02 -3.31916481e-01
1.12077117e+00 -7.88838804e-01 9.99583751e-02 6.64150059e-01
8.82693648e-01 6.71410799e-01 -1.04318178e+00 -2.32983872e-01
-2.21735761e-01 -3.74019235e-01 -3.74484777e-01 -1.28475142e+00
-4.32593346e-01 3.84831190e-01 2.80335456e-01 4.52244192e-01
6.04963601e-01 -1.48606613e-01 8.79662693e-01 8.65312874e-01
6.22181177e-01 -9.91697788e-01 -2.59059854e-02 5.45939624e-01
1.06184924e+00 -1.19963801e+00 -2.76110679e-01 -3.64328027e-01
-1.12452030e+00 1.25489080e+00 6.01047814e-01 4.06897396e-01
7.09864557e-01 2.68555284e-01 5.03570199e-01 6.54045818e-03
-1.50667298e+00 5.43601550e-02 1.78364739e-01 5.43819070e-01
8.91854942e-01 2.07104772e-01 -1.40743732e-01 7.87581801e-01
-4.98173952e-01 2.86312461e-01 7.86603272e-01 5.99522054e-01
-3.11441123e-01 -1.44871724e+00 8.30654055e-02 3.50119263e-01
-8.71634856e-02 -1.06089287e-01 -2.84140438e-01 6.59867227e-01
-3.41058493e-01 1.53270364e+00 -9.10266712e-02 -5.65219402e-01
5.34540236e-01 1.20899312e-01 6.37743296e-03 -1.06395805e+00
-8.92885983e-01 -4.59192246e-01 1.02302265e+00 -8.85250390e-01
-3.13349187e-01 -4.77420092e-01 -1.26598644e+00 -5.31095326e-01
-1.40030831e-01 5.79658508e-01 6.15370810e-01 9.67372477e-01
1.01418161e+00 2.83665061e-01 6.48999393e-01 -9.60968256e-01
-4.60447967e-01 -1.00190187e+00 -7.48841465e-02 3.97817194e-01
1.41380832e-01 -3.10889304e-01 -2.43888259e-01 -5.75718060e-02] | [12.524327278137207, 7.700018882751465] |
37ffe439-289c-4579-ae56-16fc6a466484 | superpoint-self-supervised-interest-point | 1712.07629 | null | http://arxiv.org/abs/1712.07629v4 | http://arxiv.org/pdf/1712.07629v4.pdf | SuperPoint: Self-Supervised Interest Point Detection and Description | This paper presents a self-supervised framework for training interest point
detectors and descriptors suitable for a large number of multiple-view geometry
problems in computer vision. As opposed to patch-based neural networks, our
fully-convolutional model operates on full-sized images and jointly computes
pixel-level interest point locations and associated descriptors in one forward
pass. We introduce Homographic Adaptation, a multi-scale, multi-homography
approach for boosting interest point detection repeatability and performing
cross-domain adaptation (e.g., synthetic-to-real). Our model, when trained on
the MS-COCO generic image dataset using Homographic Adaptation, is able to
repeatedly detect a much richer set of interest points than the initial
pre-adapted deep model and any other traditional corner detector. The final
system gives rise to state-of-the-art homography estimation results on HPatches
when compared to LIFT, SIFT and ORB. | ['Tomasz Malisiewicz', 'Daniel DeTone', 'Andrew Rabinovich'] | 2017-12-20 | null | null | null | null | ['interest-point-detection', 'homography-estimation'] | ['computer-vision', 'computer-vision'] | [-5.36346324e-02 -2.03582063e-01 -8.40472504e-02 -1.10369809e-01
-1.06474054e+00 -6.16829515e-01 8.77483666e-01 -2.28807837e-01
-1.80361018e-01 1.64732739e-01 -3.06855589e-02 3.39757323e-01
1.38533115e-01 -8.39951694e-01 -1.17058027e+00 -2.76097625e-01
-5.03329933e-02 8.19787145e-01 5.07062912e-01 -4.45118308e-01
6.78659737e-01 9.24454868e-01 -1.74543989e+00 1.13845289e-01
4.34283346e-01 9.25563216e-01 -9.10695270e-02 6.94725633e-01
5.93768895e-01 1.10409021e-01 -7.31013939e-02 -4.58848476e-01
6.00218177e-01 4.34064157e-02 -4.04771656e-01 2.49885559e-01
1.75119233e+00 -4.74788219e-01 -4.76039529e-01 7.52979577e-01
5.08139491e-01 -1.01337165e-01 6.04097843e-01 -9.76043761e-01
-5.53620100e-01 5.97049147e-02 -8.53264093e-01 -2.13993952e-01
6.34925067e-01 7.16357380e-02 9.90054965e-01 -1.61712289e+00
1.24411798e+00 1.31643713e+00 1.23211992e+00 6.07240908e-02
-1.34641933e+00 -4.31312203e-01 -2.81135619e-01 2.34801583e-02
-1.27876711e+00 -3.16180348e-01 8.62958372e-01 -2.32389763e-01
1.32294428e+00 -1.42540872e-01 8.30212831e-01 1.23222172e+00
5.65638542e-01 7.80117631e-01 6.66588426e-01 -4.56763715e-01
-2.61354595e-02 -2.85992086e-01 -4.72274601e-01 8.18438947e-01
2.90547371e-01 5.20537198e-01 -6.67956114e-01 -1.44873098e-01
1.47304106e+00 2.19928529e-02 -2.28493705e-01 -1.53238571e+00
-1.82499659e+00 7.85941362e-01 6.74752712e-01 1.99264090e-04
-3.73582900e-01 3.87656629e-01 4.44470644e-01 1.80173174e-01
1.99186713e-01 6.88589215e-01 -4.60721999e-01 1.22360826e-01
-1.33823764e+00 4.09252465e-01 6.24327600e-01 1.40767217e+00
1.13318288e+00 1.29230535e-02 3.31794918e-01 5.33371329e-01
1.26431823e-01 7.10188508e-01 2.71600842e-01 -1.06991327e+00
3.77131879e-01 4.94651794e-01 2.08245460e-02 -1.22328389e+00
-4.10886049e-01 -5.79741478e-01 -6.39403164e-01 3.91625851e-01
3.15735549e-01 2.42537484e-01 -1.02023005e+00 1.15205383e+00
3.99552166e-01 8.77480283e-02 -1.55131400e-01 8.46410453e-01
5.75439572e-01 1.73227817e-01 -5.57645500e-01 5.55936575e-01
1.21782410e+00 -1.05097759e+00 8.75247717e-02 -3.58152241e-01
5.61755896e-01 -1.13644099e+00 7.63615668e-01 4.81566817e-01
-1.04537547e+00 -6.78977549e-01 -1.26706970e+00 -5.78446507e-01
-5.04988968e-01 3.62712979e-01 7.63666809e-01 3.28258812e-01
-1.12091827e+00 6.79986715e-01 -5.76644719e-01 -6.64164782e-01
3.17992151e-01 4.92895156e-01 -8.12776804e-01 -1.93425864e-01
-7.95969307e-01 7.15751231e-01 1.90842509e-01 -2.35672086e-01
-8.35042298e-01 -8.43036115e-01 -1.21892619e+00 6.01892322e-02
2.19850957e-01 -1.18653011e+00 9.99914944e-01 -8.09501171e-01
-1.28344345e+00 1.33580637e+00 9.40929949e-02 -6.19470477e-01
8.18165302e-01 -4.99829292e-01 -1.20657377e-01 3.78622562e-01
3.32434207e-01 1.11369550e+00 1.29829216e+00 -1.29657412e+00
-7.95488417e-01 -4.89155710e-01 -2.45022297e-01 2.34273523e-01
4.32603329e-01 -2.47234881e-01 -4.63289171e-01 -6.40522718e-01
5.31854391e-01 -9.89079714e-01 -1.41808718e-01 2.31677130e-01
-2.92399257e-01 8.81342962e-02 1.07542622e+00 -3.39977622e-01
3.31499010e-01 -1.89946949e+00 1.43978864e-01 2.83171237e-01
-1.15348781e-02 4.93076853e-02 -1.17507115e-01 7.09629655e-01
-1.68433174e-01 -4.00028527e-01 2.46516839e-02 -2.33576909e-01
-9.23049450e-02 -3.03929329e-01 -3.44832152e-01 1.01736462e+00
2.07361102e-01 9.07011628e-01 -8.95870686e-01 -4.19425219e-01
9.15672898e-01 7.17578173e-01 -6.96280718e-01 -4.82808091e-02
7.82708544e-03 1.13429390e-01 -6.78167790e-02 9.06909883e-01
1.01114583e+00 -7.04222843e-02 -2.72239447e-01 -5.24637759e-01
-2.76778609e-01 -1.16164528e-01 -1.35298729e+00 2.21607304e+00
-7.20328212e-01 8.42322886e-01 -1.05531387e-01 -5.43366432e-01
1.11848891e+00 -1.00631215e-01 5.16886234e-01 -3.31649184e-01
8.55874494e-02 5.89474976e-01 -4.51190412e-01 1.70259923e-01
8.72494280e-01 4.03600544e-01 -8.49149451e-02 -1.05660304e-01
5.50008714e-01 -5.77929199e-01 -4.40954901e-02 1.78962555e-02
8.42161775e-01 6.18536115e-01 5.66033781e-01 -3.40429217e-01
7.96150267e-01 1.87218130e-01 7.06742108e-02 5.64152598e-01
8.13349783e-02 1.29000020e+00 1.68818265e-01 -8.11516881e-01
-1.58164763e+00 -1.08268940e+00 -3.31765741e-01 6.94846869e-01
2.72793174e-01 -3.16112965e-01 -4.61688370e-01 -5.15452862e-01
4.71926987e-01 1.70103788e-01 -5.83835423e-01 -4.61120903e-02
-9.09357727e-01 1.73641846e-01 2.50105709e-01 5.24564862e-01
7.06198633e-01 -8.33864272e-01 -8.01420450e-01 1.04293875e-01
3.50587457e-01 -1.11954081e+00 -7.29092658e-01 7.68833756e-02
-8.90677333e-01 -1.32564116e+00 -9.50926721e-01 -1.03470683e+00
5.79265356e-01 4.73382562e-01 1.39964223e+00 -2.49700010e-01
-3.75590533e-01 7.61297762e-01 -6.82015792e-02 1.16689168e-01
-1.15658343e-01 1.67300239e-01 -1.02943860e-01 -2.82479469e-02
3.60107660e-01 -4.98010665e-01 -9.57311809e-01 4.90699857e-01
-5.23879290e-01 -2.93236256e-01 9.59062159e-01 1.05706108e+00
7.92638123e-01 -8.98038328e-01 -1.76754534e-01 -5.10275722e-01
5.95979318e-02 1.62440583e-01 -9.88645673e-01 2.22445577e-01
-4.34311569e-01 5.76549135e-02 5.79761624e-01 -2.42620528e-01
-5.63512802e-01 6.55636549e-01 -4.47807610e-02 -8.47867012e-01
-2.85631835e-01 -1.41148001e-01 5.37997531e-03 -9.74785447e-01
8.48052144e-01 3.24933320e-01 -3.10933534e-02 -2.87452012e-01
7.65581548e-01 1.33892611e-01 9.07392442e-01 -2.83019155e-01
1.10972440e+00 8.03673506e-01 3.61364633e-01 -1.00997281e+00
-2.28181422e-01 -9.23259616e-01 -1.30656421e+00 -1.53010916e-02
6.58106089e-01 -1.23532796e+00 -5.29920995e-01 5.60489178e-01
-1.24051416e+00 1.58494905e-01 -3.13301414e-01 5.92207372e-01
-1.15611827e+00 6.92838013e-01 -4.03526723e-01 -1.87430844e-01
-4.38656956e-01 -1.10255396e+00 2.00423622e+00 -4.02885489e-02
4.56035361e-02 -9.48782086e-01 4.30350304e-01 -1.01397626e-01
6.74728379e-02 4.07356620e-01 3.05091560e-01 -4.48079795e-01
-1.04540014e+00 -6.00814462e-01 -2.53564209e-01 1.14613481e-01
-4.39314961e-01 3.09413880e-01 -7.72637010e-01 -6.08247221e-01
-2.55195111e-01 -5.32238185e-01 1.03675663e+00 5.46919703e-01
6.18224978e-01 1.68710992e-01 -4.97713506e-01 1.31835675e+00
1.70117259e+00 -3.14861029e-01 7.44577408e-01 7.94415534e-01
8.02841902e-01 2.22362399e-01 7.22377181e-01 2.99214095e-01
3.49992812e-01 9.36276734e-01 6.97801471e-01 -4.02777612e-01
-1.09092079e-01 -6.05465055e-01 1.04660489e-01 3.53922620e-02
-1.67303264e-01 2.31124297e-01 -8.81532788e-01 6.51649833e-01
-1.61871815e+00 -8.56574297e-01 -1.77687272e-01 2.40308547e+00
1.44029394e-01 -1.37136877e-01 1.10443793e-01 -1.26699790e-01
6.67458534e-01 1.97862610e-01 -4.34890181e-01 -1.08575471e-01
-3.14395487e-01 1.71250984e-01 1.05917287e+00 3.06256235e-01
-1.55070233e+00 1.29801321e+00 6.53000546e+00 7.58045733e-01
-1.25307882e+00 -3.46224308e-01 4.06200327e-02 4.21980649e-01
1.22925788e-01 1.98491126e-01 -9.07676280e-01 -1.24866053e-01
3.93756241e-01 1.20303951e-01 2.22295344e-01 1.46443713e+00
-4.96460140e-01 -3.04197297e-02 -1.28802478e+00 1.38938427e+00
4.85275924e-01 -1.62433684e+00 8.48469287e-02 2.10196957e-01
1.06559229e+00 4.90211993e-01 1.37181967e-01 1.68403834e-02
1.40433654e-01 -5.08381903e-01 6.58393025e-01 2.20299214e-01
8.52609932e-01 -8.00924003e-01 6.21488988e-01 -7.02343434e-02
-1.34869707e+00 1.74180672e-01 -6.16626978e-01 6.09719574e-01
1.60973415e-01 2.38318950e-01 -7.63809025e-01 8.21390808e-01
6.83687091e-01 1.14846873e+00 -9.85066593e-01 1.34438658e+00
-5.72126620e-02 -3.19505990e-01 -5.33296824e-01 4.90947187e-01
4.25352365e-01 -1.80169456e-02 6.48412287e-01 1.17353344e+00
5.66978753e-01 -7.10283697e-01 6.95375502e-02 6.92552626e-01
-3.70561033e-02 1.59691587e-01 -1.14278281e+00 5.41684568e-01
2.76195288e-01 1.45362186e+00 -7.15255558e-01 -3.75394344e-01
-4.49719042e-01 1.12998497e+00 2.77065158e-01 6.70179650e-02
-6.04995787e-01 -6.67020977e-01 3.50763381e-01 3.64500791e-01
8.16004932e-01 -5.34177899e-01 -5.98726943e-02 -1.54131126e+00
-1.79778948e-01 -8.84359896e-01 1.70856357e-01 -8.28244090e-01
-1.04259241e+00 4.85056221e-01 -9.21481289e-03 -1.70570588e+00
-6.48797691e-01 -9.95001733e-01 -6.70014560e-01 6.46616757e-01
-1.41814578e+00 -1.72700059e+00 -4.72926557e-01 7.22098470e-01
6.78575516e-01 -3.76847953e-01 5.28184116e-01 -5.36706261e-02
2.88755715e-01 5.39954245e-01 2.53170907e-01 1.68784216e-01
1.07652724e+00 -1.25988269e+00 1.19843829e+00 8.97310376e-01
4.78182137e-01 7.20776200e-01 5.60155332e-01 -5.25164425e-01
-1.64561605e+00 -8.33128333e-01 6.17243707e-01 -7.90310204e-01
4.66988683e-01 -4.99054790e-01 -7.35625386e-01 8.60281348e-01
2.35379294e-01 3.50948036e-01 -1.02249086e-01 -9.33439210e-02
-6.01145923e-01 -2.03818940e-02 -1.16317976e+00 4.66698527e-01
1.20130122e+00 -6.14929676e-01 -5.90268672e-01 3.84423703e-01
2.60331213e-01 -8.40316713e-01 -9.51573670e-01 5.44062853e-01
7.82108903e-01 -1.52689421e+00 1.52981997e+00 -1.81218266e-01
3.83451194e-01 -1.91247880e-01 -9.79152229e-03 -1.19742656e+00
-3.72010618e-01 -7.39945292e-01 -1.23304743e-02 5.62108576e-01
1.52779333e-02 -6.10401094e-01 1.11592889e+00 -1.57048732e-01
-2.57011533e-01 -5.44390202e-01 -9.95158851e-01 -8.13822269e-01
-1.74074452e-02 5.60746640e-02 4.01789874e-01 8.39566946e-01
-4.12937373e-01 2.06703395e-01 -4.18900847e-01 1.64549381e-01
1.07704198e+00 4.23335075e-01 1.40731323e+00 -1.24629343e+00
-7.09497929e-02 -4.02513683e-01 -1.06440437e+00 -1.28568804e+00
-4.06655110e-03 -6.74805939e-01 -2.00648382e-01 -1.11689556e+00
-3.48867923e-01 3.91759500e-02 4.10081059e-01 3.70028988e-02
3.45190614e-01 4.61797237e-01 -2.55177151e-02 3.03875059e-01
-4.38218772e-01 4.29750472e-01 1.21736073e+00 5.79660125e-02
-9.61345360e-02 -8.44652429e-02 1.56613007e-01 7.04493999e-01
3.54898512e-01 -1.02623835e-01 -3.54388393e-02 -1.94350660e-01
9.34922099e-02 -5.08590825e-02 7.60682821e-01 -1.40833437e+00
3.87968659e-01 2.75228679e-01 8.86919081e-01 -1.23462319e+00
5.04537284e-01 -8.54447424e-01 1.34155095e-01 5.08473217e-01
2.68902220e-02 4.19602871e-01 1.00783326e-01 5.90412736e-01
-2.90332317e-01 8.00207481e-02 8.54958475e-01 -5.67129374e-01
-1.03241980e+00 4.59488004e-01 2.02337965e-01 -2.53342162e-03
9.38642025e-01 -6.53798938e-01 -3.41208845e-01 -3.45948219e-01
-4.66191232e-01 2.87553184e-02 1.27893221e+00 5.82515180e-01
9.33867514e-01 -1.36833405e+00 -4.87496197e-01 7.73180902e-01
4.57587242e-01 2.85132695e-02 2.31866628e-01 7.27298796e-01
-1.17313015e+00 7.91210353e-01 -6.49286807e-01 -1.08689606e+00
-1.07599127e+00 7.48512030e-01 5.09288847e-01 -1.73596904e-01
-9.43817794e-01 5.62084913e-01 2.14132905e-01 -5.60612679e-01
7.71513954e-02 -3.99011672e-01 1.32689163e-01 -1.16143771e-01
3.82431075e-02 4.16421473e-01 2.82993972e-01 -1.08664608e+00
-4.15996313e-01 1.45774174e+00 -1.12137295e-01 2.25952622e-02
1.17725503e+00 1.25670135e-01 1.48559317e-01 -1.00721531e-01
1.47710323e+00 2.20918432e-02 -1.37977159e+00 -6.61208257e-02
-3.81596461e-02 -7.77150989e-01 -3.81931174e-03 -3.88096958e-01
-8.10418427e-01 9.37481403e-01 6.90781355e-01 -2.94384152e-01
5.57081997e-01 -1.17404222e-01 5.79897046e-01 6.45089865e-01
8.00121188e-01 -9.85809147e-01 1.27924174e-01 4.53393668e-01
1.25803578e+00 -1.47599328e+00 3.45597148e-01 -4.46208268e-01
-4.57112014e-01 1.64183581e+00 6.28999472e-01 -1.01158679e+00
3.89867514e-01 -7.76394978e-02 -6.77745342e-02 -4.04910833e-01
-3.91682118e-01 -1.58466503e-01 6.09336138e-01 9.94027257e-01
1.68057039e-01 -3.71022254e-01 1.12277739e-01 -6.06858015e-01
-2.12789789e-01 -2.25990877e-01 4.51456100e-01 8.00668359e-01
-3.53344500e-01 -1.05392551e+00 -6.31963253e-01 -2.79196482e-02
-7.33076856e-02 -2.20532976e-02 -4.96907294e-01 1.48734474e+00
-1.33995041e-01 1.47821859e-01 1.35886922e-01 -4.75744337e-01
6.74831450e-01 -1.79261327e-01 7.30381966e-01 -3.67995471e-01
-6.92651570e-01 2.23461807e-01 -1.77596018e-01 -1.03504086e+00
-3.69723588e-01 -8.20686340e-01 -6.45820260e-01 -1.48380741e-01
-2.09727585e-01 -3.89507294e-01 6.21591985e-01 3.42236876e-01
5.67698717e-01 -1.19294092e-01 6.12410605e-01 -1.78511977e+00
-7.26150692e-01 -7.67054081e-01 -4.47346687e-01 3.69267702e-01
5.33882380e-01 -8.50083232e-01 -4.14754152e-01 -2.49820992e-01] | [8.061495780944824, -2.1492040157318115] |
b578fdd3-b36f-4a92-8188-31c63209ae6d | learning-attentive-and-hierarchical | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2306_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123600103.pdf | Learning Attentive and Hierarchical Representations for 3D Shape Recognition | This paper proposes a novel method for 3D shape representation learning, namely Hyperbolic Embedded Attentive Representation (HEAR). Different from existing multi-view based methods, HEAR develops a unified framework to address both multi-view redundancy and single-view incompleteness. Specifically, HEAR firstly employs a hybrid attention (HA) module, which consists of a view-agnostic attention (VAA) block and a view-specific attention (VSA) block. These two blocks jointly explore distinct but complementary spatial saliency of local features for each single-view image. Subsequently, a multi-granular view pooling (MVP) module is introduced to aggregate the multi-view features with different granularities in a coarse-to-fine manner. The resulting feature set implicitly has hierarchical relations, which are therefore projected into a Hyperbolic space by adopting the Hyperbolic embedding. A hierarchical representation is learned by Hyperbolic multi-class logistic regression based on the Hyperbolic geometry. Experimental results clearly show that HEAR outperforms the state-of-the-art approaches on three 3D shape recognition tasks including generic 3D shape retrieval, 3D shape classification and sketch-based 3D shape retrieval. | ['Fan Zhu', 'Jie Qin', 'Yuming Shen', 'Li Liu', 'Ling Shao', 'Jiaxin Chen'] | null | null | null | null | eccv-2020-8 | ['3d-shape-retrieval', '3d-shape-recognition', '3d-shape-representation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-1.70430899e-01 -1.41135558e-01 -1.09815642e-01 -4.78437096e-01
-1.09993267e+00 -5.48028290e-01 8.82969141e-01 2.04589237e-02
1.97009325e-01 -4.86504473e-02 4.92962897e-01 1.49230689e-01
-3.57065767e-01 -9.78101969e-01 -4.72458810e-01 -8.66820514e-01
3.18472832e-01 4.26375747e-01 2.64544278e-01 -2.09495679e-01
4.90219712e-01 9.24150407e-01 -1.53739953e+00 6.77230597e-01
1.62952185e-01 1.10688615e+00 1.76695272e-01 4.36313540e-01
1.93236920e-03 2.89504528e-01 -1.32741913e-01 -8.67553279e-02
3.10046971e-01 7.82983229e-02 -5.22354722e-01 2.23866716e-01
4.33254391e-01 -5.06266534e-01 -4.82835770e-01 6.52276218e-01
8.33305597e-01 2.25250539e-03 8.09448063e-01 -1.18548846e+00
-1.36918211e+00 -6.25561178e-02 -8.78171861e-01 2.75311291e-01
3.37940633e-01 -6.64005242e-03 1.35299814e+00 -1.76567590e+00
5.88558555e-01 1.61684048e+00 2.45945767e-01 3.35969865e-01
-1.28378260e+00 -3.89808476e-01 4.22750413e-01 1.63926333e-01
-1.48427689e+00 -2.88189411e-01 1.24691474e+00 -2.88454205e-01
1.24920571e+00 1.50570616e-01 8.06134760e-01 7.99707830e-01
1.14886113e-01 1.03491414e+00 1.06012869e+00 -1.16392344e-01
1.60280392e-02 -4.71318550e-02 -9.11882222e-02 8.43718708e-01
3.07498891e-02 -4.63205576e-02 -5.96547067e-01 -2.84389019e-01
1.18892074e+00 4.50205088e-01 -2.10345183e-02 -1.11317825e+00
-1.17484426e+00 8.90643477e-01 7.30699718e-01 2.68134445e-01
-4.81676042e-01 -2.49226883e-01 2.53587037e-01 1.95743650e-01
7.85525262e-01 7.43372217e-02 -2.55418748e-01 4.43518698e-01
-5.96122563e-01 2.44161293e-01 4.36582685e-01 1.01819670e+00
7.34227836e-01 -8.29473957e-02 -2.92084724e-01 1.03030777e+00
5.89966238e-01 5.89135647e-01 2.40509436e-01 -5.11765242e-01
5.91059685e-01 1.17778039e+00 -2.04617679e-01 -1.16426361e+00
-2.06019685e-01 -2.51062185e-01 -8.06950808e-01 2.90741175e-01
-9.83224064e-02 7.05164731e-01 -7.68849313e-01 1.54681206e+00
4.18466687e-01 -2.40582094e-01 -6.35640100e-02 1.08289647e+00
1.26296127e+00 5.69831729e-01 -1.65519133e-01 2.75228947e-01
1.38434660e+00 -1.03524649e+00 -2.41574049e-01 1.62962690e-01
7.19414279e-02 -6.40255630e-01 9.90845203e-01 1.31244222e-02
-1.50110972e+00 -5.44649482e-01 -1.22731555e+00 -5.84793270e-01
-7.06007123e-01 2.17175096e-01 2.75385469e-01 2.68609971e-01
-1.00742340e+00 9.56222266e-02 -5.85585415e-01 -3.15167040e-01
7.01308370e-01 2.39426553e-01 -6.00116014e-01 -2.61794150e-01
-7.72060990e-01 7.43373334e-01 -1.83028296e-01 7.31965154e-02
-8.21247160e-01 -6.35451674e-01 -1.03435862e+00 2.60349244e-01
2.58662701e-01 -9.84589994e-01 7.62795627e-01 -1.86759755e-01
-1.37524128e+00 1.25243270e+00 -3.00583273e-01 2.73121864e-01
-5.39640300e-02 -5.92098944e-02 -7.14311004e-02 4.61919367e-01
7.41094500e-02 6.43249094e-01 1.27997220e+00 -1.64586997e+00
-1.09202430e-01 -1.01510286e+00 1.41288787e-01 7.32441902e-01
-4.60770354e-02 -9.45041180e-02 -5.10567486e-01 -7.20710635e-01
7.28603601e-01 -5.99495769e-01 2.26125062e-01 2.22405896e-01
-1.94704235e-01 -5.20620883e-01 9.75642741e-01 -3.15454632e-01
1.03486311e+00 -2.08351469e+00 8.24147701e-01 1.75458938e-01
3.89466912e-01 -9.40398499e-02 -2.89042413e-01 7.09121764e-01
-1.43559784e-01 -5.63971028e-02 -1.64622843e-01 -4.62251335e-01
1.61971431e-02 1.11512572e-01 -4.80977982e-01 4.40201342e-01
4.99174207e-01 1.30591297e+00 -8.22782040e-01 -1.02849066e-01
2.49088004e-01 6.03895307e-01 -5.99882424e-01 4.17555183e-01
9.12534446e-02 2.54380018e-01 -6.82920694e-01 1.07560194e+00
7.90843010e-01 -5.05717933e-01 -2.36880526e-01 -2.98965245e-01
-2.62638479e-01 1.78758074e-02 -9.17192459e-01 1.84887958e+00
-6.36644304e-01 3.43622603e-02 3.43884435e-03 -9.01042044e-01
1.29886138e+00 2.35698119e-01 3.54838043e-01 -3.96357417e-01
-2.85195082e-01 2.49168217e-01 -4.75920469e-01 -3.54728192e-01
3.37929517e-01 -1.76659718e-01 -7.84384459e-02 7.13111818e-01
3.87339264e-01 -3.86712343e-01 -6.38612330e-01 9.97162312e-02
8.17615449e-01 1.91320837e-01 5.11354148e-01 -1.97166502e-01
9.08345938e-01 -6.38413668e-01 6.29827008e-02 4.81129825e-01
-1.11468047e-01 1.09167731e+00 5.15554070e-01 -7.97826052e-01
-1.22082603e+00 -1.42347431e+00 2.75004841e-02 1.04650617e+00
3.44616205e-01 -2.49049023e-01 -1.23437643e-01 -8.67895126e-01
3.19779843e-01 1.29939124e-01 -8.51896226e-01 -1.71330109e-01
-4.66669381e-01 -1.64245352e-01 1.15167312e-01 6.92096770e-01
5.39162815e-01 -1.18132305e+00 -6.23207271e-01 -3.79649065e-02
3.73420298e-01 -8.35491061e-01 -7.30247140e-01 -5.91958873e-02
-9.80149329e-01 -9.58250701e-01 -1.04394257e+00 -8.19691002e-01
7.28955567e-01 9.82332170e-01 7.99178898e-01 -9.36021209e-02
-2.88672656e-01 1.19168746e+00 -3.38868409e-01 -2.79925346e-01
2.25973219e-01 1.09342788e-03 -3.28021087e-02 3.38896841e-01
4.73214298e-01 -8.95211339e-01 -8.03476810e-01 2.55799681e-01
-8.73073220e-01 1.69716887e-02 8.83279920e-01 1.21831179e+00
8.44863176e-01 -6.88344598e-01 6.23249531e-01 -4.06234324e-01
4.06826526e-01 -4.98965174e-01 -4.25645709e-01 4.73307490e-01
-2.31140703e-01 -7.39256963e-02 3.30669641e-01 -8.97337496e-02
-9.56419528e-01 -9.69688892e-02 2.05563512e-02 -9.52143192e-01
-2.96792418e-01 2.31761515e-01 -4.89377648e-01 -5.59558123e-02
8.10168535e-02 7.37567723e-01 1.42018631e-01 -6.34041071e-01
5.98096669e-01 5.63202024e-01 -2.12297915e-03 -3.83277565e-01
7.36516237e-01 7.93417275e-01 -1.63861047e-02 -9.93097484e-01
-8.59680593e-01 -6.42719686e-01 -1.10667825e+00 -1.21133961e-01
1.02325249e+00 -1.11365116e+00 -5.49483836e-01 4.97221053e-01
-1.22819412e+00 2.62027681e-01 -3.51737142e-01 1.52896509e-01
-8.49215627e-01 3.29119533e-01 -5.03929555e-01 -7.85489023e-01
-4.02074158e-01 -1.07789457e+00 1.91429472e+00 7.13667413e-03
2.74789244e-01 -7.48141646e-01 -4.46803346e-02 3.87602895e-01
2.81690270e-01 -9.77921579e-03 1.34329104e+00 -7.20325172e-01
-1.01232898e+00 -1.16588153e-01 -6.68899298e-01 1.23702914e-01
7.66898617e-02 -4.66879576e-01 -1.17217815e+00 -5.60794711e-01
1.56174272e-01 -4.95297968e-01 1.01379836e+00 2.86525637e-01
1.25424266e+00 -1.48495302e-01 -1.01774916e-01 5.59030235e-01
1.42365694e+00 1.46387313e-02 4.38717037e-01 6.79942966e-02
6.84898198e-01 5.43035984e-01 2.65916795e-01 4.71974880e-01
5.92195392e-01 7.51146078e-01 5.15379488e-01 -4.00244482e-02
-1.81866530e-03 -4.60153550e-01 1.56197082e-02 1.01131749e+00
-2.43228927e-01 8.05598870e-02 -5.88719964e-01 4.71328735e-01
-1.70386648e+00 -1.10337663e+00 4.47509199e-01 2.09345126e+00
1.20960683e-01 -8.81031305e-02 1.06951877e-01 7.52160475e-02
4.74278957e-01 7.59425759e-01 -6.55585885e-01 -3.45937729e-01
-2.44360998e-01 -2.25689057e-02 -2.48493537e-01 3.00583482e-01
-1.14485312e+00 7.40770996e-01 5.03831625e+00 6.65665686e-01
-9.82195973e-01 -1.39717581e-02 4.65564996e-01 1.64847285e-01
-7.26622224e-01 -2.36590877e-01 -6.06556952e-01 -2.04685166e-01
1.68865477e-03 1.04340814e-01 2.24804565e-01 7.45380044e-01
-3.07938546e-01 3.03025335e-01 -1.13217783e+00 1.18080676e+00
6.12103462e-01 -1.32913721e+00 7.12982297e-01 2.22915456e-01
6.15183532e-01 -2.12356802e-02 3.17500442e-01 2.35583127e-01
-4.05069768e-01 -8.32442880e-01 5.79296410e-01 8.91648352e-01
8.24298918e-01 -6.55291021e-01 4.30472195e-01 3.05613041e-01
-1.66734219e+00 -2.36553341e-01 -4.83563542e-01 2.67751336e-01
5.89311076e-03 1.36000246e-01 -1.43037841e-01 8.45594466e-01
6.98992491e-01 1.08303535e+00 -5.16188443e-01 8.44646692e-01
-1.85329560e-02 -2.74403870e-01 -1.04256645e-02 -7.26902857e-02
2.91080952e-01 -2.38556832e-01 9.88970995e-01 7.44468391e-01
3.04059327e-01 6.25187993e-01 -1.41683882e-02 1.12864244e+00
3.20949480e-02 1.10355474e-01 -1.27651381e+00 2.19833970e-01
3.47822696e-01 1.38552344e+00 -5.95119119e-01 -3.07794422e-01
-8.43317986e-01 1.04449451e+00 5.35555959e-01 4.40123796e-01
-2.33090669e-01 -1.55810699e-01 4.68316019e-01 8.58243033e-02
9.93719459e-01 -3.82235080e-01 -3.17714512e-01 -1.26736653e+00
5.67566529e-02 -5.13336599e-01 4.17534232e-01 -9.88746285e-01
-1.70060802e+00 5.00279784e-01 1.28759101e-01 -1.52367687e+00
2.28524119e-01 -5.83105564e-01 -6.08731031e-01 1.07967663e+00
-1.74460912e+00 -1.75577688e+00 -1.63389787e-01 6.01097047e-01
8.37650597e-01 -4.82767016e-01 1.08523488e+00 -5.18557094e-02
-1.57486022e-01 3.27953845e-01 -3.05837512e-01 -1.76466554e-01
4.72029001e-01 -1.22731709e+00 3.18841070e-01 8.08740109e-02
2.52632558e-01 5.97414613e-01 -3.18094820e-01 -4.22302723e-01
-1.73136985e+00 -7.96135962e-01 1.01822293e+00 -6.67183340e-01
3.91164184e-01 -4.54853863e-01 -1.03779197e+00 4.45332319e-01
1.48302643e-02 3.45411062e-01 8.34502816e-01 -6.73970655e-02
-8.11354876e-01 -4.45380174e-02 -9.44996774e-01 4.15325910e-01
1.22765994e+00 -9.90706623e-01 -9.38349187e-01 -1.76588148e-01
6.28163278e-01 -2.98846364e-01 -1.08286202e+00 4.75098372e-01
8.07157993e-01 -1.22714961e+00 1.37391067e+00 -8.17291498e-01
5.70300639e-01 -1.95289880e-01 -6.29344106e-01 -1.02952623e+00
-6.99004471e-01 -8.54683593e-02 -3.04210901e-01 8.92388821e-01
4.38176282e-02 -4.60132867e-01 5.22417247e-01 1.69145316e-01
-2.68750846e-01 -1.43124771e+00 -1.13992083e+00 -6.08793199e-01
1.24756500e-01 -2.52176486e-02 7.16927946e-01 6.95886493e-01
-1.21825114e-01 5.73874414e-01 -7.06344694e-02 2.75039822e-01
6.61366522e-01 9.88469303e-01 6.54497266e-01 -1.28872097e+00
-1.07442290e-01 -7.01916397e-01 -6.36437297e-01 -1.25777602e+00
1.78373419e-02 -1.16347039e+00 -3.95038575e-01 -1.41723478e+00
5.74739933e-01 -9.13221240e-02 -4.66426969e-01 2.70095259e-01
-7.42493570e-02 2.77193010e-01 3.75313818e-01 3.07553709e-01
-7.12552667e-01 1.13865614e+00 1.58021104e+00 -1.58020154e-01
-2.66007274e-01 1.24979675e-01 -6.31295502e-01 6.60642266e-01
3.87465030e-01 -7.69301057e-02 -2.50428379e-01 -4.16986763e-01
-6.48531914e-02 2.70465106e-01 7.61740148e-01 -4.88261789e-01
2.71413833e-01 1.41346186e-01 6.38860703e-01 -1.03690171e+00
7.74766922e-01 -6.94860399e-01 -5.15321493e-01 1.73248470e-01
-2.31492907e-01 2.14844346e-01 4.94192652e-02 8.46889496e-01
-3.41868699e-01 2.16809556e-01 5.79468191e-01 -3.87344629e-01
-4.94765043e-01 6.53034031e-01 1.12441145e-01 -1.98187053e-01
8.48455667e-01 -3.27708483e-01 -1.43447012e-01 -4.13229674e-01
-8.86365235e-01 1.64886579e-01 2.23760232e-01 7.04893708e-01
1.28902960e+00 -1.85824704e+00 -6.56885862e-01 8.87372315e-01
5.43965578e-01 -2.20362134e-02 5.97792625e-01 6.89101696e-01
2.10878089e-01 5.87010503e-01 -3.17796648e-01 -8.68753910e-01
-1.18046606e+00 6.59474134e-01 8.80027190e-02 -3.45045388e-01
-8.45710337e-01 9.50951695e-01 6.33206427e-01 -6.39137685e-01
1.58503979e-01 -1.28694519e-01 -3.80387843e-01 2.34754071e-01
4.10938382e-01 2.94273585e-01 -1.00844659e-01 -8.77689004e-01
-3.02356631e-01 1.26502466e+00 1.06670663e-01 -1.12631366e-01
1.76601791e+00 -1.58736780e-01 -1.97284058e-01 8.18957150e-01
1.43587756e+00 -1.67695716e-01 -1.33968782e+00 -4.59476858e-01
-2.45790228e-01 -6.32644296e-01 -1.23825140e-01 -4.27196294e-01
-8.34138572e-01 1.27468514e+00 4.01308209e-01 2.83597976e-01
1.08767748e+00 4.68095273e-01 5.91061473e-01 2.99495161e-01
3.23038727e-01 -6.64443374e-01 5.38749218e-01 6.91105962e-01
1.81189716e+00 -1.28321946e+00 1.19123131e-01 -3.49865526e-01
-6.00969434e-01 1.16232634e+00 5.62204599e-01 -4.33229953e-01
1.05379844e+00 -3.74874175e-01 -3.30136716e-01 -5.92563331e-01
-9.21326995e-01 -2.40086973e-01 8.80024552e-01 4.83175099e-01
7.99933001e-02 -2.21943349e-01 1.43708140e-01 8.71091127e-01
4.00740117e-01 -5.94045579e-01 -4.83309068e-02 8.05767238e-01
-3.47763598e-01 -9.32649553e-01 -1.66757867e-01 2.77282834e-01
1.36738181e-01 7.33923838e-02 -5.23502767e-01 7.87809432e-01
-1.39277965e-01 3.11644524e-01 4.15531509e-02 -3.32757831e-01
6.88899338e-01 3.26986611e-01 7.35222340e-01 -7.32080638e-01
-4.69166279e-01 3.27059716e-01 -5.36108434e-01 -5.97550690e-01
-3.98947775e-01 -7.04513192e-01 -7.82564104e-01 2.17722014e-01
-8.83152261e-02 -4.82514739e-01 3.02754521e-01 6.50956810e-01
5.91608524e-01 2.24952936e-01 9.90735412e-01 -1.45004070e+00
-7.21579432e-01 -5.75268805e-01 -7.01116383e-01 3.54221851e-01
4.58100349e-01 -9.57354546e-01 -1.99815392e-01 -4.11592096e-01] | [8.14376163482666, -3.8878486156463623] |
3fce106d-8e0d-49ad-8df8-b80631bf2431 | what-s-there-in-the-dark | null | null | https://ieeexplore.ieee.org/document/8803299 | https://ieeexplore.ieee.org/document/8803299 | What's There in the Dark | Scene Parsing is an important cog for modern autonomousdriving systems. Most of the works in semantic segmenta-tion pertains to day-time scenes with favourable weather andillumination conditions. In this paper, we propose a noveldeep architecture, NiSeNet, that performs semantic segmen-tation of night scenes using a domain mapping approach ofsynthetic to real data. It is a dual-channel network, wherewe designed a Real channel using DeepLabV3+ coupled withan MSE loss to preserve the spatial information. In addition,we used an Adaptive channel reducing the domain gap be-tween synthetic and real night images, which also comple-ments the failures of Real channel output. Apart from thedual channel, we introduced a novel fusion scheme to fuse theoutputs of two channels. In addition to that, we compiled anew dataset Urban Night Driving Dataset (UNDD); it consistsof7125unlabelled day and night images; additionally, it has75night images with pixel-level annotations having classesequivalent to Cityscapes dataset. We evaluated our approachon the Berkley Deep Drive dataset, the challenging Mapil-lary dataset and UNDD dataset to exhibit that the proposedmethod outperforms the state-of-the-art techniques in termsof accuracy and visual quality | ['Sukhendu Das', 'Saptakatha Adak', 'Sauradip Nag'] | 2019-09-24 | null | null | null | null | ['scene-parsing'] | ['computer-vision'] | [-1.45250261e-01 2.27308631e-01 1.79868892e-01 -7.48024464e-01
-6.71724498e-01 -7.04591095e-01 7.47377813e-01 -1.41421080e-01
-8.18372130e-01 9.06841338e-01 -1.11497171e-01 -2.70482093e-01
1.28280640e-01 -1.01161015e+00 -1.03541160e+00 -5.52516699e-01
2.93335199e-01 3.60336185e-01 6.38295233e-01 -6.84424698e-01
-1.68646023e-01 3.56541425e-01 -1.89618063e+00 2.88346767e-01
9.15657640e-01 1.08405912e+00 4.49484795e-01 6.53184652e-01
-2.13026851e-01 6.80568099e-01 -4.89497334e-01 -3.01723689e-01
4.57676709e-01 -3.82325888e-01 -6.59440398e-01 1.63896546e-01
2.98575550e-01 -9.07173008e-02 -3.65515381e-01 1.15872645e+00
4.41831082e-01 7.90489614e-02 1.78902581e-01 -1.27920020e+00
-9.78323519e-02 3.12837481e-01 -3.83795857e-01 3.99598390e-01
-7.63034299e-02 -1.32524848e-01 4.13707495e-01 -7.10763395e-01
6.60965800e-01 1.14794588e+00 6.03973150e-01 2.28522018e-01
-1.02015483e+00 -7.57494271e-01 8.56245980e-02 6.65605485e-01
-1.28274882e+00 -4.80292797e-01 6.84399962e-01 -3.54799062e-01
7.95324564e-01 1.31609160e-02 4.79460061e-01 9.92527544e-01
2.05700189e-01 6.07769132e-01 1.35271668e+00 -2.68218428e-01
3.36810499e-01 3.15098166e-01 2.46897444e-01 4.57529008e-01
1.33882299e-01 4.86017346e-01 -2.52621949e-01 6.20008945e-01
3.38083237e-01 -2.16857538e-01 -3.21717858e-02 -2.48802781e-01
-9.20523763e-01 6.77913010e-01 5.25039852e-01 -4.02914546e-02
-1.85655504e-01 -7.55961761e-02 4.55663681e-01 5.42743325e-01
3.02676201e-01 -1.36166573e-01 -4.13920432e-01 -3.87281738e-02
-6.08944297e-01 2.50340790e-01 6.35312438e-01 1.10014927e+00
1.09623063e+00 2.27069169e-01 2.36836120e-01 7.46481061e-01
2.35938579e-01 8.68413031e-01 4.48712170e-01 -8.75624597e-01
6.28756404e-01 2.95920402e-01 4.34306078e-02 -7.71404028e-01
-7.58998871e-01 -2.67256767e-01 -7.26895988e-01 4.85165596e-01
2.36349806e-01 -3.77572000e-01 -1.23676288e+00 1.36845565e+00
2.03256324e-01 2.43326858e-01 6.00600243e-01 1.28834653e+00
1.09831452e+00 5.75472474e-01 2.26459503e-02 2.52180219e-01
1.26409328e+00 -9.65147316e-01 -7.23625362e-01 -3.97288591e-01
6.64305627e-01 -6.65159404e-01 8.82063031e-01 3.89123917e-01
-5.84203362e-01 -1.00995409e+00 -1.31551659e+00 -9.83788744e-02
-8.45857084e-01 5.46068698e-02 5.03216803e-01 4.18066204e-01
-1.06760395e+00 2.73381889e-01 -5.54743767e-01 -6.66216910e-01
1.82853103e-01 -1.69647172e-01 -6.34160459e-01 2.74168681e-02
-1.41399074e+00 9.30335701e-01 1.01695263e+00 2.39272252e-01
-1.05278659e+00 -3.75899374e-01 -9.90402043e-01 -2.60518551e-01
3.60153854e-01 -1.63406312e-01 1.14866889e+00 -1.14688051e+00
-1.31855464e+00 9.44808602e-01 2.24427700e-01 -9.89026487e-01
5.78740537e-01 -1.24427356e-01 -8.63527298e-01 3.34692486e-02
2.15285629e-01 9.43141878e-01 5.21866143e-01 -1.23023760e+00
-8.65132987e-01 -4.55076784e-01 2.90482074e-01 5.00041306e-01
1.61439195e-01 -4.79094595e-01 -6.21802032e-01 -6.40888333e-01
2.70563066e-01 -9.03988302e-01 -1.42025232e-01 -2.33393624e-01
-4.51905340e-01 2.74942040e-01 1.05744410e+00 -6.81911051e-01
7.78582633e-01 -2.60041356e+00 -1.62265271e-01 2.25931909e-02
-1.37944534e-01 2.27900133e-01 -1.24126591e-01 1.84777707e-01
-1.98034152e-01 -3.48709553e-01 -4.86708730e-01 -3.17666143e-01
-1.15398094e-02 6.54077470e-01 -2.98051625e-01 4.80751306e-01
-5.96072935e-02 6.41300619e-01 -6.97877586e-01 -5.93654275e-01
7.29476392e-01 2.73264408e-01 -2.33527109e-01 3.07167441e-01
-1.97611421e-01 2.86023378e-01 -6.28705993e-02 4.87051636e-01
1.03774786e+00 4.07443643e-01 -6.97103143e-02 8.67956057e-02
-3.36902678e-01 -2.11250231e-01 -1.27994359e+00 2.00096512e+00
-5.43664932e-01 8.77012670e-01 2.58720338e-01 -1.10184276e+00
1.12085986e+00 1.23376527e-03 9.13066491e-02 -1.39067781e+00
2.27204248e-01 3.11035603e-01 -2.84730732e-01 -6.23023868e-01
6.69141829e-01 2.24610511e-02 -3.69465888e-01 -2.65483260e-01
7.11030364e-02 -4.77733940e-01 1.52445138e-01 -5.55050150e-02
6.59274876e-01 2.80659437e-01 -1.17669865e-01 -4.78015512e-01
6.49352252e-01 3.94923210e-01 5.66924691e-01 7.17962623e-01
-3.26502323e-01 7.93957770e-01 3.02865148e-01 -5.24788082e-01
-1.22391129e+00 -1.10406446e+00 -2.52514005e-01 7.19351828e-01
6.12208307e-01 2.07816493e-02 -9.95789826e-01 -6.41673684e-01
-8.13556612e-02 7.72611499e-01 -5.11913359e-01 -2.22612526e-02
-2.97431171e-01 -9.10329342e-01 7.79734373e-01 4.97613013e-01
1.51765728e+00 -8.43288302e-01 -8.87656689e-01 1.49966151e-01
-1.99490696e-01 -1.49239135e+00 4.37923335e-02 5.96154273e-01
-3.86076391e-01 -1.23401177e+00 -3.60539824e-01 -8.02847981e-01
9.17747468e-02 3.37283283e-01 9.62886512e-01 -6.19094014e-01
-2.69586980e-01 -1.55097336e-01 -5.47190309e-01 -6.53717637e-01
-3.52355301e-01 7.30998442e-02 -3.95683140e-01 -8.25294033e-02
6.47553504e-01 -3.06779444e-01 -3.96692365e-01 3.25237006e-01
-8.20154130e-01 3.36080998e-01 4.87880468e-01 6.38598502e-01
8.53756189e-01 3.15397501e-01 3.33079964e-01 -7.70565510e-01
1.05442107e-01 -5.15143931e-01 -9.58985090e-01 -9.26067233e-02
-3.59443754e-01 3.67093869e-02 8.27699304e-01 3.45835894e-01
-1.35784245e+00 3.78038496e-01 -5.41566730e-01 -2.31601164e-01
-5.88684678e-01 -8.29535797e-02 -4.14911091e-01 1.01840533e-01
8.50485146e-01 1.05579771e-01 -1.56865269e-01 -4.66472566e-01
5.50052047e-01 9.44852412e-01 9.67755854e-01 -1.19375676e-01
5.65744817e-01 7.73614645e-01 -7.90335014e-02 -9.35743213e-01
-6.43966496e-01 -5.33299208e-01 -5.63441277e-01 -1.45526275e-01
1.33167100e+00 -1.28696907e+00 -2.78484583e-01 9.23600793e-01
-1.04296672e+00 -4.22608435e-01 -2.90833503e-01 4.87314165e-01
-6.46008849e-01 1.46270320e-01 -2.26442233e-01 -4.40272987e-01
3.14629381e-03 -9.57816660e-01 1.02128255e+00 3.99702013e-01
5.11516333e-01 -8.45043838e-01 -5.81511259e-02 2.67681386e-02
1.68138713e-01 4.51114863e-01 4.71422046e-01 -3.92980039e-01
-3.03144425e-01 1.58048347e-02 -6.22179270e-01 6.75392747e-01
-2.58837402e-01 -4.56493467e-01 -1.37035215e+00 1.58569068e-01
2.64268573e-02 -1.53699204e-01 1.09954977e+00 2.83988655e-01
1.17168069e+00 3.31924736e-01 -5.38052619e-02 8.51789951e-01
1.73417473e+00 4.57402915e-01 9.37392533e-01 8.35728109e-01
6.63232803e-01 6.05015576e-01 9.43556309e-01 3.82393152e-01
7.49090970e-01 6.62160695e-01 7.18868017e-01 -6.08044088e-01
-3.44498158e-01 -1.30183160e-01 1.35503262e-01 1.97698146e-01
2.08850533e-01 -4.01207358e-01 -8.71822059e-01 6.52863622e-01
-2.04970193e+00 -6.92170858e-01 -5.26306152e-01 2.06727982e+00
2.55565912e-01 4.29269344e-01 -1.32398963e-01 1.44662768e-01
5.89376450e-01 3.23849320e-01 -5.11101186e-01 -6.53448403e-01
-7.05982625e-01 1.69283040e-02 1.24272454e+00 4.85595316e-01
-1.50495017e+00 1.16556311e+00 5.54769230e+00 6.08324230e-01
-1.08814275e+00 3.86313856e-01 4.22355294e-01 3.17167461e-01
-2.92648003e-02 -1.74504891e-01 -8.62917483e-01 6.41404152e-01
1.20837796e+00 2.20726743e-01 2.98393607e-01 1.01568973e+00
3.36048245e-01 -3.34412366e-01 -3.88975382e-01 1.06086898e+00
2.05958620e-01 -9.71020401e-01 -4.67408329e-01 -2.57865906e-01
4.97857720e-01 5.41770339e-01 -2.50627726e-01 2.70970553e-01
4.98601675e-01 -6.88112557e-01 9.61509883e-01 4.78988707e-01
6.03830397e-01 -9.72636819e-01 1.20089686e+00 3.21873754e-01
-1.18851125e+00 -2.37156436e-01 -3.74443769e-01 8.11394751e-02
1.78965032e-01 5.70241332e-01 -8.43403876e-01 8.52293670e-01
1.22571611e+00 8.28479528e-01 -6.29669666e-01 9.12925005e-01
-3.59902024e-01 2.69018173e-01 -3.03081572e-01 4.87094849e-01
4.14885342e-01 -2.18260288e-01 2.94176489e-01 1.26837909e+00
2.73215622e-01 6.61806837e-02 2.79405385e-01 4.04517174e-01
2.09935874e-01 -2.22685754e-01 -8.39856923e-01 3.61555189e-01
3.19408119e-01 1.10095537e+00 -6.87706769e-01 -5.03567636e-01
-3.36503565e-01 1.17058527e+00 -1.91354275e-01 4.23365802e-01
-1.21085453e+00 -8.01697254e-01 7.58269131e-01 4.80431542e-02
3.16794872e-01 -2.73583174e-01 -4.92645085e-01 -9.92857039e-01
-4.14556786e-02 -3.89833957e-01 3.51988256e-01 -9.77714837e-01
-6.92600131e-01 8.59820604e-01 2.25258589e-01 -1.15295160e+00
-6.68007359e-02 -4.52365249e-01 -1.78191945e-01 8.31847787e-01
-2.22100067e+00 -1.04682589e+00 -8.17633033e-01 6.39480650e-01
8.59361410e-01 -1.16330907e-01 5.93722045e-01 4.59723473e-01
-4.88470525e-01 2.80849218e-01 3.27455282e-01 -8.46446976e-02
6.97882831e-01 -1.35347593e+00 6.43479943e-01 1.18821061e+00
-2.94091851e-01 -4.18826312e-01 8.62725198e-01 -3.91429335e-01
-9.62395132e-01 -1.56040072e+00 5.95004976e-01 -4.44465950e-02
3.63994151e-01 -5.66853523e-01 -7.73225963e-01 5.55598438e-01
2.60351092e-01 6.59427270e-02 -6.19531274e-02 -4.98653412e-01
-2.69703753e-02 -3.15319389e-01 -1.13571131e+00 1.29121363e-01
9.69151735e-01 -3.31475884e-01 -5.30951202e-01 3.51747185e-01
8.29043508e-01 -6.70342386e-01 -3.82943243e-01 3.47129107e-01
2.19903618e-01 -1.34406543e+00 7.39204884e-01 9.63878557e-02
1.10628881e-01 -5.37282109e-01 -5.06679535e-01 -1.31806839e+00
2.38310988e-03 -2.13563796e-02 3.72662663e-01 9.30192649e-01
2.86711216e-01 -7.22730041e-01 8.31897318e-01 4.76823049e-03
-7.92115629e-01 -4.70493771e-02 -9.00768399e-01 -9.75106359e-01
4.38129976e-02 -5.82867324e-01 8.59086215e-01 6.35582209e-01
-7.56648362e-01 1.85217947e-01 -2.49563158e-01 3.80385250e-01
4.02636170e-01 -1.31771103e-01 8.68808210e-01 -1.13693523e+00
4.23681624e-02 8.97112638e-02 -7.92647719e-01 -7.60195017e-01
2.83867151e-01 -5.64225435e-01 3.80558819e-01 -1.58137548e+00
-4.06627774e-01 -3.48013133e-01 -1.59940913e-01 4.62312073e-01
3.62885952e-01 4.48799700e-01 1.17169032e-02 -1.98812738e-01
-5.74397147e-01 6.02589548e-01 9.61142540e-01 -9.46436822e-02
-1.68687865e-01 -3.59068811e-01 -3.18188548e-01 6.60793126e-01
8.05315197e-01 -3.21688116e-01 -7.71330178e-01 -6.60993934e-01
-8.50459337e-02 -1.63016066e-01 6.76739097e-01 -1.43680775e+00
3.07826325e-02 9.57833901e-02 3.49526405e-01 -9.50170398e-01
5.16251683e-01 -9.73343432e-01 3.16566110e-01 3.17784339e-01
2.56938577e-01 8.78781155e-02 3.61649424e-01 4.79524255e-01
-5.89090288e-01 7.42069334e-02 1.00122797e+00 -2.46192053e-01
-1.85901296e+00 -1.21765420e-01 -2.44264811e-01 -1.06641417e-02
1.39869416e+00 -3.93262476e-01 -4.73038107e-01 -8.57018009e-02
-7.08368897e-01 4.79990989e-01 5.99056780e-01 6.86447144e-01
4.77682352e-01 -1.02121842e+00 -5.94511271e-01 5.27881503e-01
3.18902105e-01 2.30895072e-01 5.65568566e-01 5.98078966e-01
-9.68403876e-01 5.37619531e-01 -3.86535257e-01 -6.82094693e-01
-9.99081969e-01 4.34546590e-01 6.39317214e-01 3.92546564e-01
-8.02066982e-01 7.62552798e-01 3.15728575e-01 -7.52190173e-01
5.74254710e-03 -4.52795595e-01 -5.07164478e-01 1.80961549e-01
3.96670341e-01 2.00814009e-01 5.75181723e-01 -9.25087512e-01
-3.90838146e-01 3.83688837e-01 2.41939560e-01 -1.68732315e-01
1.10167205e+00 -4.21089530e-01 2.35726148e-01 3.68030876e-01
1.20075154e+00 -6.72479868e-01 -1.39219165e+00 7.23801777e-02
-2.65721917e-01 -5.04566073e-01 2.28582904e-01 -7.69759595e-01
-1.06089664e+00 9.44600880e-01 1.37855673e+00 6.34774193e-02
1.33176696e+00 -1.95714802e-01 8.38318110e-01 2.65092999e-01
4.76331264e-01 -1.60482824e+00 -3.95616502e-01 6.98406816e-01
4.92731214e-01 -1.77600193e+00 -5.24918437e-01 -2.80992657e-01
-7.87169933e-01 1.00670362e+00 7.52915382e-01 -1.56838819e-01
7.88071394e-01 1.35852903e-01 2.92259127e-01 -1.90248773e-01
-2.97835261e-01 -6.27387166e-01 -2.34515235e-01 8.60879183e-01
-1.37150228e-01 2.63334453e-01 -3.38526845e-01 5.21963537e-01
-5.65033793e-01 6.67543262e-02 6.64896905e-01 9.24728274e-01
-6.81230843e-01 -6.98585451e-01 -4.34846014e-01 5.04232980e-02
-4.93389219e-02 1.49626855e-03 -2.05564480e-02 9.85862911e-01
7.87680387e-01 1.15721488e+00 3.84675890e-01 -5.14374495e-01
7.82264769e-01 -3.19545478e-01 -7.08522871e-02 -3.98076475e-01
-1.64267406e-01 -1.37755066e-01 3.59519690e-01 -7.17712045e-01
-3.99454802e-01 -4.45319712e-01 -1.48483682e+00 -2.49103934e-01
-4.52512540e-02 1.69664562e-01 1.23734546e+00 9.46740568e-01
3.29129785e-01 6.03235543e-01 5.66063046e-01 -6.64896786e-01
1.40731037e-01 -6.93695486e-01 -7.60223508e-01 2.69741029e-01
3.71795416e-01 -6.56477690e-01 -1.15697749e-01 6.31683245e-02] | [8.80954360961914, -1.4530682563781738] |
8c513dbc-7310-413d-942c-f757a60ccb9e | multi-level-domain-adaptation-for-lane | 2206.10692 | null | https://arxiv.org/abs/2206.10692v2 | https://arxiv.org/pdf/2206.10692v2.pdf | Multi-level Domain Adaptation for Lane Detection | We focus on bridging domain discrepancy in lane detection among different scenarios to greatly reduce extra annotation and re-training costs for autonomous driving. Critical factors hinder the performance improvement of cross-domain lane detection that conventional methods only focus on pixel-wise loss while ignoring shape and position priors of lanes. To address the issue, we propose the Multi-level Domain Adaptation (MLDA) framework, a new perspective to handle cross-domain lane detection at three complementary semantic levels of pixel, instance and category. Specifically, at pixel level, we propose to apply cross-class confidence constraints in self-training to tackle the imbalanced confidence distribution of lane and background. At instance level, we go beyond pixels to treat segmented lanes as instances and facilitate discriminative features in target domain with triplet learning, which effectively rebuilds the semantic context of lanes and contributes to alleviating the feature confusion. At category level, we propose an adaptive inter-domain embedding module to utilize the position prior of lanes during adaptation. In two challenging datasets, ie TuSimple and CULane, our approach improves lane detection performance by a large margin with gains of 8.8% on accuracy and 7.4% on F1-score respectively, compared with state-of-the-art domain adaptation algorithms. | ['Guangliang Cheng', 'Jia Shi', 'Boheng Zhang', 'Chenguang Li'] | 2022-06-21 | null | null | null | null | ['lane-detection'] | ['computer-vision'] | [ 1.46667689e-01 -1.17062978e-01 -3.00905555e-01 -5.38499594e-01
-8.91004562e-01 -5.79988837e-01 4.90907818e-01 2.53263153e-02
-4.66132730e-01 8.47649932e-01 7.74646550e-02 -3.56370509e-01
2.07886621e-01 -6.24092460e-01 -7.67023385e-01 -7.58022547e-01
3.77354831e-01 -1.25621125e-01 9.57969606e-01 -2.62755901e-01
3.86677712e-01 4.17482316e-01 -1.73519921e+00 1.49203390e-01
1.37722480e+00 8.54259849e-01 1.13609210e-01 3.29625696e-01
-3.79165173e-01 4.63051260e-01 -1.65554672e-01 -2.42219254e-01
4.38776761e-01 -9.63011663e-03 3.34273390e-02 4.47631121e-01
8.35162938e-01 -3.09731781e-01 -2.84800887e-01 1.10196793e+00
5.11334658e-01 8.39024857e-02 6.92192018e-01 -1.44468236e+00
-1.67446285e-01 -1.70389846e-01 -1.16930377e+00 2.02111170e-01
-1.78197667e-01 5.18954992e-01 6.08837008e-01 -8.40764940e-01
4.53312606e-01 1.20940518e+00 8.34987581e-01 2.99720407e-01
-1.12361753e+00 -8.88406396e-01 4.98708785e-01 5.71492612e-01
-1.16457999e+00 -4.74647254e-01 1.23060584e+00 -5.98862946e-01
4.22828794e-01 -2.20780641e-01 1.39810026e-01 9.57588851e-01
1.07126117e-01 9.33255672e-01 1.17444241e+00 -3.07488620e-01
-3.17515135e-02 3.37575585e-01 2.15155214e-01 6.34287059e-01
3.76287073e-01 1.43712163e-01 -4.03468937e-01 3.14640224e-01
5.24357378e-01 -3.45615089e-01 2.19454005e-01 -1.05046177e+00
-1.06702077e+00 8.26840043e-01 1.79515243e-01 -3.05408955e-01
1.14193019e-02 -2.63656080e-01 8.57650399e-01 -1.17095456e-01
2.32536197e-01 -9.79745537e-02 -3.42867643e-01 7.09529147e-02
-5.76817930e-01 2.27438077e-01 1.95574209e-01 1.11341512e+00
9.41721737e-01 1.49934739e-01 -1.49752244e-01 1.10351074e+00
2.96119153e-01 7.02103138e-01 1.97297707e-01 -8.44272256e-01
9.60287511e-01 6.63050115e-01 1.35381117e-01 -9.42548513e-01
-3.99546653e-01 -6.04698360e-01 -5.54566383e-01 5.09125233e-01
5.49292088e-01 -1.73343018e-01 -1.05742908e+00 1.71004772e+00
5.25632977e-01 4.95319545e-01 3.65442961e-01 9.66533780e-01
4.91707057e-01 5.67589700e-01 4.03004736e-01 1.60571411e-01
1.37339759e+00 -1.24638140e+00 -4.94108766e-01 -8.01436424e-01
7.84250617e-01 -8.17092776e-01 1.07896411e+00 1.34853005e-01
-5.52579820e-01 -1.04069793e+00 -1.43268621e+00 -6.07214831e-02
-5.08615911e-01 5.77409983e-01 3.04689050e-01 5.87871134e-01
-5.38204908e-01 -1.48345888e-01 -2.44732022e-01 -2.23277718e-01
4.56811816e-01 6.74744397e-02 -5.13718128e-01 -3.45492750e-01
-1.31817782e+00 9.05390501e-01 5.81378818e-01 -9.46579501e-02
-6.38977528e-01 -8.51468444e-01 -1.06099451e+00 -3.76521498e-01
5.46957254e-01 -2.58374184e-01 6.70772135e-01 -6.90969884e-01
-1.14368820e+00 8.57371151e-01 -1.97881564e-01 -5.53446829e-01
5.69300473e-01 -1.63164362e-01 -8.62425625e-01 -1.60889357e-01
3.94188046e-01 9.65104818e-01 8.80058229e-01 -1.53667569e+00
-1.38621151e+00 -2.91883945e-01 -1.38815492e-01 4.04512227e-01
-5.82661945e-03 -5.88665605e-01 -4.70685482e-01 -4.48000699e-01
8.18337500e-02 -8.51819932e-01 -1.16444618e-01 1.10028066e-01
-1.25251696e-01 -1.85549289e-01 1.20779538e+00 -7.07188129e-01
1.06872082e+00 -2.56228781e+00 -4.62923646e-01 -8.15562755e-02
-3.39738838e-02 4.93594617e-01 -1.15624338e-01 -1.35880664e-01
1.71198681e-01 -4.23952371e-01 -4.16610479e-01 -9.94095281e-02
8.12317729e-02 3.53988886e-01 -4.02117461e-01 5.32982767e-01
6.52603269e-01 5.22126555e-01 -7.53749669e-01 -6.98859036e-01
5.79597533e-01 1.85540885e-01 -4.82122213e-01 8.91813114e-02
9.37764253e-03 4.89546180e-01 -3.44868571e-01 4.02548552e-01
1.41397953e+00 4.08094436e-01 -1.79082647e-01 -3.81571442e-01
-3.59066367e-01 7.34063685e-02 -1.45048833e+00 1.53816760e+00
-4.63094264e-01 6.01629078e-01 9.85997845e-04 -1.10701454e+00
1.25901890e+00 -2.93658465e-01 7.93403834e-02 -1.06768620e+00
-1.01369560e-01 2.75156975e-01 5.18593974e-02 -3.86817396e-01
5.24894357e-01 5.74499927e-02 -4.68565762e-01 -2.93913305e-01
-2.40864038e-01 1.06793083e-01 3.52625288e-02 -9.56519395e-02
6.64621413e-01 4.24228281e-01 1.27962649e-01 -2.51027316e-01
1.07471752e+00 1.87052310e-01 9.15011406e-01 4.37528104e-01
-9.08793628e-01 3.15317601e-01 4.97231364e-01 -2.46685505e-01
-1.24548864e+00 -1.24266636e+00 -1.77567884e-01 1.11509883e+00
6.58982038e-01 4.00406718e-01 -5.71347773e-01 -9.56060231e-01
4.96082902e-01 8.16108584e-01 -5.75023770e-01 -2.61500210e-01
-6.67485356e-01 -5.46530128e-01 5.04617870e-01 7.64597476e-01
9.70155954e-01 -4.54497755e-01 -5.08343697e-01 1.71449751e-01
9.86012146e-02 -1.56040311e+00 -5.03705561e-01 3.28168631e-01
-6.75092220e-01 -9.06429589e-01 -4.01360631e-01 -8.46010923e-01
5.25259078e-01 5.56768000e-01 7.46741474e-01 -5.33154786e-01
-2.10262269e-01 -6.41877996e-03 -1.70632780e-01 -2.78312564e-01
-2.45979786e-01 4.69689034e-02 -7.38837495e-02 2.87143290e-01
5.07828057e-01 -3.71342748e-01 -7.45805979e-01 5.99782646e-01
-4.22044277e-01 1.18242584e-01 9.93869424e-01 9.44227874e-01
5.22749901e-01 -1.30122870e-01 8.87247682e-01 -8.59335065e-01
1.48777604e-01 -6.14437819e-01 -7.18262434e-01 2.16646399e-03
-6.68112814e-01 1.13802664e-01 6.00853026e-01 -2.35109195e-01
-1.44893122e+00 3.73370618e-01 1.14705451e-02 -3.99028212e-01
-5.72351635e-01 -1.36799037e-01 -5.22783518e-01 -1.65088736e-02
5.03242075e-01 4.07205522e-01 5.18192202e-02 -2.80559063e-01
5.21273375e-01 6.74196839e-01 8.56282651e-01 -4.37384665e-01
9.37803984e-01 6.19786441e-01 1.21215098e-01 -5.64880252e-01
-6.92396998e-01 -9.21243131e-01 -8.61090004e-01 -2.54157245e-01
1.04855740e+00 -1.37430239e+00 -1.50218055e-01 3.49211693e-01
-8.26413393e-01 -1.57627285e-01 1.80211756e-02 2.40761399e-01
-4.64286327e-01 4.95279819e-01 1.55315727e-01 -7.97107935e-01
1.93728834e-01 -1.07401228e+00 1.16321671e+00 2.61924148e-01
4.48377967e-01 -9.33522880e-01 -2.51969069e-01 4.87929642e-01
2.26957783e-01 4.23707604e-01 8.77188325e-01 -4.37272161e-01
-5.84062219e-01 -1.71385109e-01 -7.16990411e-01 5.10650694e-01
2.34443936e-02 -4.10709709e-01 -1.13541222e+00 -1.00456111e-01
-5.85008025e-01 -4.09091972e-02 1.08126569e+00 5.88457137e-02
9.20264900e-01 3.26196343e-01 -5.37100196e-01 4.27903742e-01
1.43226635e+00 2.48088539e-01 6.12087607e-01 6.26941741e-01
8.76342237e-01 8.05397868e-01 1.16637945e+00 3.52451384e-01
6.74229443e-01 6.86866343e-01 3.69881034e-01 -3.02618206e-01
-5.24474859e-01 -4.44575816e-01 3.66746843e-01 3.46611738e-01
5.95035434e-01 -1.28554497e-02 -7.86564231e-01 9.38799500e-01
-1.95674670e+00 -7.43209004e-01 -3.55749041e-01 2.14172220e+00
4.75688756e-01 7.46115506e-01 3.36432666e-01 9.70822722e-02
9.25954759e-01 2.36262426e-01 -8.51417720e-01 -3.35383236e-01
-2.65339226e-01 -5.54911315e-01 1.20321035e+00 5.06587386e-01
-1.57457244e+00 1.17071486e+00 4.85345602e+00 1.09339178e+00
-1.09614170e+00 4.58030403e-02 5.39817870e-01 4.17442679e-01
-4.07601744e-02 -1.02387220e-01 -1.29932559e+00 6.32316887e-01
7.12826848e-01 1.79386854e-01 -1.24853425e-01 1.06903028e+00
3.90347570e-01 -3.07893693e-01 -6.87388539e-01 8.32955599e-01
1.32648692e-01 -9.67308700e-01 -1.82981506e-01 -7.93627836e-03
7.82981277e-01 -1.36163712e-01 2.12029889e-01 6.35170758e-01
-2.95143429e-04 -4.41230804e-01 7.08625734e-01 3.37201387e-01
8.00788105e-01 -9.86124694e-01 6.97768927e-01 3.60963225e-01
-1.40527380e+00 -3.34005982e-01 -4.26825464e-01 2.20725819e-01
1.22026026e-01 3.93021643e-01 -9.61335719e-01 6.16528451e-01
4.75875914e-01 8.29459727e-01 -7.62004197e-01 9.62281287e-01
-3.47822011e-02 5.40299654e-01 -9.91199687e-02 4.93624628e-01
5.60012102e-01 -8.57605040e-02 5.51208496e-01 1.43995225e+00
1.09080561e-01 -2.77279109e-01 4.55075383e-01 4.82232720e-01
3.65258217e-01 -1.58721864e-01 -7.27445245e-01 8.03746283e-01
6.01818800e-01 1.13540840e+00 -3.42537105e-01 -2.60382593e-01
-6.47220373e-01 8.05292010e-01 2.06015676e-01 2.17535600e-01
-1.25597680e+00 -6.19433165e-01 9.56910729e-01 3.86543691e-01
6.89384639e-01 -4.66376692e-01 -5.54365695e-01 -6.24022007e-01
2.09186405e-01 -4.77115124e-01 2.19285056e-01 -4.41390663e-01
-1.19140565e+00 1.14793867e-01 2.57156286e-02 -1.71505094e+00
7.59329274e-02 -7.63335228e-01 -5.17050922e-01 8.75439167e-01
-2.38749003e+00 -1.46488321e+00 -5.44841886e-01 5.09431601e-01
8.48824680e-01 -2.26824120e-01 7.04689100e-02 6.73609555e-01
-5.58326542e-01 9.53698218e-01 3.32290053e-01 3.32993343e-02
1.24998939e+00 -1.05911720e+00 2.89362937e-01 1.03738952e+00
-6.43004298e-01 1.00979269e-01 6.35324180e-01 -4.92985576e-01
-1.15964675e+00 -1.57309306e+00 5.13525426e-01 -3.64949942e-01
5.84429383e-01 -4.74064767e-01 -9.35217917e-01 1.86111152e-01
-1.19769283e-01 2.41766676e-01 2.98768580e-01 -2.95996845e-01
-4.66233701e-01 -6.22351289e-01 -1.24941170e+00 5.38206100e-01
1.18095839e+00 -3.22828025e-01 -4.69512641e-01 -8.52721334e-02
6.00069284e-01 -3.11546206e-01 -5.42790711e-01 6.66316748e-01
4.18406457e-01 -1.04566407e+00 1.07937765e+00 -2.28883877e-01
1.85943052e-01 -9.34408545e-01 -2.93856114e-01 -1.10585332e+00
-3.67004037e-01 2.72472445e-02 2.10152447e-01 1.40160096e+00
3.74306411e-01 -7.31400669e-01 8.51077497e-01 2.07401782e-01
-7.13922560e-01 -4.79947597e-01 -1.02017236e+00 -8.22019219e-01
1.75401762e-01 -4.83393997e-01 3.78468513e-01 6.93193436e-01
-4.59974468e-01 2.07126141e-01 -3.70860487e-01 5.88219464e-01
8.65108609e-01 -2.39855126e-02 1.10170639e+00 -1.18980694e+00
1.49489000e-01 -2.77641982e-01 -6.70358658e-01 -1.32757974e+00
3.89143258e-01 -6.45942807e-01 3.79962891e-01 -1.25029075e+00
-7.21318200e-02 -6.60391986e-01 -4.32774931e-01 2.54211187e-01
-3.06815863e-01 2.68827379e-01 4.27351594e-02 4.01836149e-02
-8.26620340e-01 4.92616057e-01 1.19381547e+00 -2.16321692e-01
-4.26335000e-02 -2.39997223e-01 -6.01507545e-01 5.74783087e-01
7.75706530e-01 -3.72385085e-02 -5.54412842e-01 -4.56720322e-01
-4.63568389e-01 -3.09392840e-01 4.28467035e-01 -1.44202578e+00
2.41502643e-01 -2.52828568e-01 4.16345000e-01 -9.87612665e-01
2.32153878e-01 -9.09571707e-01 -4.16544318e-01 2.19176561e-01
-1.19682476e-01 -2.25308046e-01 7.50699401e-01 1.06885076e+00
-2.68440962e-01 1.26078054e-01 1.15537453e+00 2.99057007e-01
-1.75450683e+00 -7.36846998e-02 -2.91876704e-01 6.42666668e-02
1.43136585e+00 -7.46014833e-01 -4.88871902e-01 1.45772114e-01
-3.85016203e-01 5.92643023e-01 4.45453197e-01 6.46560371e-01
3.70881647e-01 -1.44737577e+00 -6.05909467e-01 5.46674132e-01
6.50086761e-01 -7.68238306e-02 5.69290698e-01 8.88808608e-01
-2.17191398e-01 5.72723985e-01 -5.00573397e-01 -8.29584897e-01
-9.82271969e-01 4.82589751e-01 2.39053115e-01 -1.04393266e-01
-4.06227350e-01 6.54492378e-01 5.12056470e-01 -4.06351238e-01
7.77019709e-02 -2.14806363e-01 -3.27510566e-01 1.50698852e-02
2.00286359e-01 4.21059996e-01 4.50213291e-02 -9.06879008e-01
-5.17648280e-01 9.60825264e-01 -9.93459225e-02 2.74327606e-01
8.49006295e-01 -5.33502400e-01 5.42400301e-01 3.20034653e-01
1.13383412e+00 1.35116667e-01 -2.10779285e+00 -1.28068790e-01
6.62545636e-02 -5.52622736e-01 1.22576512e-01 -9.39595759e-01
-6.30468965e-01 9.96803820e-01 1.19981611e+00 -3.85369509e-01
9.75698411e-01 -2.74002880e-01 9.09675062e-01 2.79091410e-02
1.23074949e-01 -1.41168547e+00 -1.14090294e-01 6.19638741e-01
4.17911947e-01 -1.83645988e+00 -2.25126103e-01 -6.59478962e-01
-1.02084351e+00 9.70347464e-01 9.90669608e-01 -2.40839884e-01
3.81164908e-01 1.47005543e-01 1.05469093e-01 3.33623260e-01
-3.58712018e-01 -3.81715328e-01 4.17874306e-01 8.89187276e-01
9.03016850e-02 1.11263573e-01 -2.31309727e-01 3.35029483e-01
3.84508282e-01 -2.04908624e-01 2.05827609e-01 7.34112620e-01
-9.79759932e-01 -1.02248847e+00 -3.70779574e-01 2.96109974e-01
2.60954916e-01 2.26624012e-01 1.54771730e-01 8.98283541e-01
7.76838362e-01 6.91963792e-01 9.19178277e-02 -5.34903109e-01
6.72796190e-01 2.76969254e-01 4.79552336e-02 -2.30760440e-01
2.83101469e-01 -1.04350865e-01 3.32333475e-01 -3.32015216e-01
-1.30422994e-01 -8.67498457e-01 -1.20599842e+00 -4.52165939e-02
-1.34667307e-01 -1.33537620e-01 5.92416883e-01 7.44135678e-01
7.66409636e-01 5.56992769e-01 7.28975773e-01 -7.27246284e-01
-5.77383339e-01 -7.01098800e-01 -4.41017330e-01 4.16887373e-01
5.04209995e-01 -1.20706236e+00 -3.67236733e-01 4.24834713e-02] | [8.091313362121582, -1.4787602424621582] |
0b190aa1-05c8-4668-a9ee-527f1c2a44d5 | model-free-robust-average-reward | 2305.10504 | null | https://arxiv.org/abs/2305.10504v1 | https://arxiv.org/pdf/2305.10504v1.pdf | Model-Free Robust Average-Reward Reinforcement Learning | Robust Markov decision processes (MDPs) address the challenge of model uncertainty by optimizing the worst-case performance over an uncertainty set of MDPs. In this paper, we focus on the robust average-reward MDPs under the model-free setting. We first theoretically characterize the structure of solutions to the robust average-reward Bellman equation, which is essential for our later convergence analysis. We then design two model-free algorithms, robust relative value iteration (RVI) TD and robust RVI Q-learning, and theoretically prove their convergence to the optimal solution. We provide several widely used uncertainty sets as examples, including those defined by the contamination model, total variation, Chi-squared divergence, Kullback-Leibler (KL) divergence and Wasserstein distance. | ['Shaofeng Zou', 'Ashley Prater-Bennette', 'George Atia', 'Alvaro Velasquez', 'Yue Wang'] | 2023-05-17 | null | null | null | null | ['q-learning'] | ['methodology'] | [-9.46553051e-02 1.47927210e-01 -1.53442860e-01 -2.84148812e-01
-1.56614661e+00 -6.67192578e-01 4.12135839e-01 3.84165466e-01
-5.77938855e-01 1.20682693e+00 -3.60872559e-02 -5.20292819e-01
-8.72720420e-01 -3.80550265e-01 -5.56758761e-01 -8.12185109e-01
-6.48148298e-01 6.53675377e-01 -6.36978671e-02 3.86751026e-01
4.45558846e-01 4.20153111e-01 -9.91038859e-01 -5.00973463e-01
1.06403434e+00 1.43551731e+00 7.32845590e-02 7.52213717e-01
3.26322407e-01 8.54832828e-01 -6.12257719e-01 -3.23107660e-01
3.58672410e-01 -2.37834409e-01 -5.07662117e-01 -2.55843759e-01
-4.77086902e-01 -7.17699289e-01 -8.14485997e-02 1.21154189e+00
5.79425335e-01 5.36417246e-01 1.10534811e+00 -1.71788895e+00
-3.28642458e-01 7.20025778e-01 -5.50718963e-01 1.67012945e-01
2.08733886e-01 2.35186815e-01 1.08174896e+00 -4.93105888e-01
3.80784750e-01 1.58439493e+00 2.53395319e-01 5.00206411e-01
-1.17843211e+00 -2.79457092e-01 2.35028222e-01 -9.59629640e-02
-1.33025670e+00 -8.76465291e-02 5.66192083e-02 -4.93430167e-01
5.69755197e-01 1.60836577e-02 9.95254144e-02 9.75430787e-01
8.06036055e-01 7.82756984e-01 1.24607515e+00 -2.11139858e-01
9.23680484e-01 -7.10688680e-02 7.71927228e-03 2.33419448e-01
4.92358983e-01 6.40755236e-01 9.75112338e-03 -3.84463549e-01
7.48594344e-01 1.77774370e-01 -2.33901478e-02 -6.01553619e-01
-1.08246672e+00 9.30164754e-01 -2.16947809e-01 -3.60931695e-01
-5.84594667e-01 5.44520974e-01 -4.34596501e-02 4.07477140e-01
1.20461099e-01 2.50616610e-01 -3.28971654e-01 -4.01912063e-01
-5.59498250e-01 7.42134392e-01 1.01780879e+00 1.18656588e+00
2.22905308e-01 3.29343788e-02 -9.66295421e-01 1.96274668e-01
9.51484561e-01 1.04482794e+00 -1.99150369e-01 -1.53597999e+00
7.52760470e-01 -4.47401464e-01 1.27759111e+00 -3.49064559e-01
-2.03005299e-01 -3.73101056e-01 -4.98579949e-01 2.55108774e-01
5.41652858e-01 -8.42963815e-01 -6.92299545e-01 1.80799460e+00
2.47756764e-01 -7.47306794e-02 3.92877877e-01 7.21107721e-01
-2.56242305e-01 6.09194577e-01 -2.94976085e-02 -8.62728298e-01
6.15790069e-01 -3.97354692e-01 -1.01121438e+00 1.79440200e-01
2.83099771e-01 -4.02272463e-01 6.82924151e-01 4.50976372e-01
-1.13067317e+00 1.59198195e-02 -9.28946018e-01 7.84489095e-01
3.52100879e-01 -7.56081641e-01 -1.31330892e-01 6.34588301e-01
-7.07774282e-01 9.96484458e-01 -8.31438601e-01 2.28489026e-01
4.21053141e-01 2.10121408e-01 3.45964879e-01 -1.73947051e-01
-1.15812445e+00 9.72573459e-01 2.41148531e-01 7.03749806e-02
-1.79879284e+00 -5.34011543e-01 -5.52857280e-01 -2.37392168e-02
9.58397865e-01 -4.43622112e-01 1.76988447e+00 -1.35929912e-01
-1.61038232e+00 8.40260312e-02 1.96218729e-01 -7.01456189e-01
1.01695466e+00 -2.93302268e-01 -1.59358546e-01 -2.29698811e-02
3.48575711e-02 1.48987010e-01 1.01406944e+00 -1.29484642e+00
-6.42774403e-01 -2.60654271e-01 5.41196764e-02 6.13865703e-02
4.97102499e-01 1.32453412e-01 3.19481522e-01 -4.21700716e-01
-1.32563099e-01 -8.81137788e-01 -7.21363842e-01 -3.51392835e-01
-5.14504373e-01 -1.59116909e-01 4.68201004e-02 -3.63697082e-01
1.31389570e+00 -1.83109760e+00 3.28403674e-02 4.52874064e-01
-1.27246499e-01 -3.38572204e-01 -9.24055427e-02 5.29727638e-01
4.48774189e-01 3.12864870e-01 -4.25126702e-01 -1.14009805e-01
6.09922707e-01 2.78026938e-01 -3.70978534e-01 5.77546358e-01
2.82949746e-01 6.38088763e-01 -1.12675202e+00 -4.75938618e-01
-2.27522701e-02 -2.08667874e-01 -1.35978535e-01 3.01977903e-01
-4.72898871e-01 4.27343369e-01 -6.89121485e-01 5.83625555e-01
6.23705685e-01 1.32927820e-01 -9.73286945e-03 4.06210244e-01
-1.60283223e-01 -3.22723091e-01 -1.59877741e+00 1.14974189e+00
-2.89853483e-01 -1.18724920e-01 1.95288971e-01 -6.92963064e-01
6.36228621e-01 2.23626807e-01 6.98108137e-01 -3.44059110e-01
3.63797516e-01 2.15063572e-01 -6.91438243e-02 -2.09209517e-01
3.73159200e-01 -7.03830779e-01 -4.40137208e-01 8.06909263e-01
-1.14937909e-01 -1.92728281e-01 8.11594352e-02 1.78781256e-01
1.04131329e+00 4.86567505e-02 3.17584217e-01 -3.57153803e-01
-4.50932086e-02 -5.90075374e-01 6.74929142e-01 1.24967086e+00
-7.47047186e-01 2.35162750e-01 1.03552616e+00 2.68825740e-01
-8.85435045e-01 -1.65256703e+00 -2.72175252e-01 5.88303983e-01
1.41486034e-01 8.71086121e-02 -5.16475737e-01 -6.78949177e-01
4.39712524e-01 1.09542274e+00 -5.59071720e-01 -1.58510491e-01
1.03902876e-01 -6.43831193e-01 1.52812213e-01 4.22055691e-01
7.02282367e-03 -5.69552839e-01 -6.27138078e-01 4.87276375e-01
1.46033630e-01 -7.94779480e-01 -3.96141678e-01 3.01234782e-01
-8.46819401e-01 -8.51419091e-01 -8.58042359e-01 2.55164653e-01
3.03512633e-01 -2.53200203e-01 7.69111156e-01 -1.01862502e+00
2.03810796e-01 7.62003362e-01 -8.52833539e-02 -8.70271146e-01
-3.45951200e-01 -6.77403808e-01 4.98582423e-01 -9.24472287e-02
-1.04468748e-01 -4.25068103e-02 -3.90765637e-01 3.94192129e-01
-7.72110283e-01 -8.10632527e-01 4.67720270e-01 5.72384059e-01
1.11243284e+00 2.20031887e-01 8.26182306e-01 -3.30893785e-01
1.07822478e+00 -5.40171742e-01 -1.11580420e+00 4.92066056e-01
-7.30156362e-01 5.15600026e-01 -1.04664741e-02 -2.43730292e-01
-1.15167236e+00 -2.65746474e-01 4.05838072e-01 -5.00525713e-01
2.27463230e-01 4.72397923e-01 -2.03652516e-01 3.53504241e-01
2.83186138e-01 -1.81546450e-01 1.00229355e-02 -2.78083503e-01
4.40388739e-01 7.32676804e-01 2.60713935e-01 -1.02704227e+00
5.07023931e-01 2.43805528e-01 3.85767579e-01 -2.93491870e-01
-1.16815817e+00 -1.32731363e-01 -4.05542403e-01 -2.61702716e-01
6.97729588e-01 -7.39022315e-01 -1.02596712e+00 1.69501394e-01
-8.72621298e-01 -5.29882193e-01 -7.67771006e-01 7.85308957e-01
-1.23803246e+00 2.61697888e-01 -4.35321748e-01 -1.75533390e+00
-6.56767786e-02 -1.02666569e+00 7.31784701e-01 8.23727250e-02
-2.32758299e-02 -9.28841054e-01 3.88306648e-01 -1.49700880e-01
1.66949883e-01 6.19775593e-01 7.05527425e-01 -5.95796764e-01
-5.39793074e-01 -1.86181247e-01 -1.26299083e-01 6.34544969e-01
-2.09800646e-01 -6.13894686e-02 -5.28415322e-01 -4.09146905e-01
2.11200893e-01 -2.05352142e-01 6.51502311e-01 9.06946778e-01
8.78987968e-01 -4.96750712e-01 -1.33633256e-01 5.59832118e-02
1.51592958e+00 5.81813395e-01 3.25969100e-01 2.04730406e-01
1.92403480e-01 5.57727754e-01 1.23651576e+00 1.16958070e+00
2.24968746e-01 1.26955912e-01 7.14442968e-01 1.04325604e+00
8.19592893e-01 -1.85114309e-01 7.06536472e-01 1.94522321e-01
1.18813299e-01 -2.52745211e-01 -8.94290864e-01 3.50996166e-01
-2.16417551e+00 -1.03709459e+00 1.39475480e-01 2.83231330e+00
5.01490533e-01 3.08725357e-01 3.27149093e-01 -1.88742757e-01
7.10446537e-01 -2.81817675e-01 -8.86749446e-01 -5.76133847e-01
1.34228766e-01 -5.10671223e-03 8.90308917e-01 6.91264272e-01
-1.08044720e+00 3.72940212e-01 7.26399565e+00 7.71946728e-01
-2.26051778e-01 1.55827656e-01 6.06739819e-01 -3.47071439e-01
-3.35096538e-01 -2.06506744e-01 -7.73639798e-01 5.82385361e-01
1.46552777e+00 -5.45773208e-01 3.35215181e-01 8.63310516e-01
6.48189127e-01 -3.35176319e-01 -1.34445667e+00 7.64998674e-01
-7.39376366e-01 -6.54764175e-01 -5.76306641e-01 3.50337088e-01
1.09625971e+00 -3.28385942e-02 2.57552654e-01 3.91493917e-01
1.18221557e+00 -1.00929964e+00 8.87975276e-01 9.84872520e-01
7.89917707e-01 -1.22886634e+00 1.00767648e+00 2.50299633e-01
-5.85768104e-01 -6.29095674e-01 -4.03277159e-01 9.23828930e-02
5.13317227e-01 8.09421182e-01 -3.78338367e-01 4.81861621e-01
3.57263565e-01 3.97450238e-01 2.57260561e-01 1.21718454e+00
-1.79382652e-01 2.48029307e-01 -4.39573169e-01 -1.93952531e-01
4.99828905e-01 -4.12316352e-01 6.67674601e-01 7.29237616e-01
5.89867949e-01 -5.19811697e-02 1.72906905e-01 9.32755709e-01
1.94455400e-01 -3.01152587e-01 -3.09217393e-01 -2.08190918e-01
6.77681208e-01 7.76434004e-01 -5.44362485e-01 -1.52670935e-01
1.03643097e-01 5.80189645e-01 -1.03884481e-01 5.04290402e-01
-8.95039797e-01 -4.89491761e-01 8.90603721e-01 -4.79483515e-01
3.27949494e-01 -2.13377923e-01 -8.82069245e-02 -9.04582739e-01
9.84782074e-03 -5.66131949e-01 4.68168497e-01 -3.71817440e-01
-1.69013512e+00 -1.05263226e-01 3.62042040e-01 -1.10294735e+00
-6.43662214e-01 -4.61609960e-01 -4.73304391e-01 9.64911461e-01
-1.34530509e+00 -2.33241275e-01 4.86899108e-01 5.40303349e-01
6.59961551e-02 1.04145505e-01 4.72569615e-01 -3.18718553e-01
-6.28629446e-01 2.13101566e-01 1.07129979e+00 -3.44940126e-01
5.50754070e-01 -1.58653247e+00 -5.28807715e-02 6.36233509e-01
-5.17849922e-01 2.59572208e-01 1.06859469e+00 -7.15304255e-01
-1.52521288e+00 -1.14398241e+00 1.02828942e-01 -6.17165029e-01
1.05402184e+00 7.28510320e-02 -4.59399015e-01 6.50573671e-01
-3.10283631e-01 -8.05792511e-02 2.68123001e-01 -1.55101985e-01
3.19006704e-02 -1.11015551e-01 -1.56852865e+00 4.88325149e-01
7.07295895e-01 -1.58025771e-01 -4.82594103e-01 3.34715277e-01
9.89363909e-01 -4.21115197e-02 -1.24055684e+00 2.99705118e-01
5.93117893e-01 -6.43699050e-01 7.69868433e-01 -1.00173926e+00
6.23369664e-02 -1.18976738e-02 -6.02437198e-01 -1.52193320e+00
-2.00222582e-01 -1.27045226e+00 -5.05714536e-01 1.12136626e+00
3.14518988e-01 -6.42115533e-01 2.74617195e-01 8.35869253e-01
1.24086358e-01 -7.42408395e-01 -1.28385103e+00 -1.42689657e+00
4.32760119e-01 -8.32081139e-01 5.81355453e-01 1.96149409e-01
1.52545780e-01 -1.36877254e-01 -4.23319370e-01 8.11358988e-02
1.43059552e+00 -9.02742222e-02 2.03015029e-01 -1.07602656e+00
-4.63375062e-01 -4.06695396e-01 3.14287990e-02 -7.85201430e-01
-2.44591031e-02 -3.05441350e-01 5.56486070e-01 -1.58861661e+00
3.43756050e-01 -4.69714195e-01 -7.30338216e-01 -8.08454454e-02
-1.82989821e-01 -8.67462337e-01 5.01239598e-01 1.36186615e-01
-1.24660146e+00 9.01429713e-01 1.33223271e+00 6.23849072e-02
-1.96778178e-01 5.80776691e-01 -7.05585480e-01 5.84025145e-01
8.34800005e-01 -7.85225570e-01 -5.35355866e-01 9.74832382e-03
2.96887606e-01 7.06127942e-01 8.60401243e-03 -7.36496329e-01
-2.54377007e-01 -9.09858823e-01 -3.75861861e-02 -6.92494154e-01
1.97224189e-02 -7.02952802e-01 -2.75426675e-02 7.07650959e-01
-7.06396878e-01 -8.93504024e-02 -2.46210575e-01 1.21952653e+00
1.48955494e-01 -6.50940657e-01 8.76370430e-01 -4.58760001e-02
-1.86297879e-01 6.49218738e-01 -6.79690719e-01 4.33644831e-01
1.37077355e+00 4.51183110e-01 -9.85265076e-02 -7.92496741e-01
-9.40301597e-01 7.54544556e-01 3.12146824e-02 -6.15228117e-02
7.43467152e-01 -1.22785163e+00 -6.98858321e-01 -4.66536790e-01
-7.11864382e-02 -1.09979577e-01 2.37595543e-01 8.70267630e-01
4.07212116e-02 5.20973623e-01 3.42072211e-02 -4.21018451e-01
-4.99293506e-01 7.72584379e-01 3.72446209e-01 -5.90043783e-01
1.69217005e-01 5.89920938e-01 -2.19805270e-01 -1.81824118e-01
4.41413283e-01 -4.93963718e-01 3.03060621e-01 6.52370527e-02
6.64028466e-01 1.01582348e+00 -4.16493624e-01 -9.02905241e-02
-1.69995070e-01 2.06089318e-01 1.53497532e-01 -9.25362289e-01
9.88214135e-01 -5.22804141e-01 1.98251560e-01 7.42075861e-01
8.62260759e-01 -6.04210556e-01 -1.92105913e+00 -3.05250406e-01
5.01660943e-01 -4.66306627e-01 -5.69141135e-02 -8.22905600e-01
-3.26265156e-01 1.03643644e+00 7.80721068e-01 3.14591736e-01
5.03938615e-01 -5.03480323e-02 4.33325291e-01 5.65872908e-01
6.13338828e-01 -1.34284389e+00 4.79677059e-02 6.09704912e-01
7.54840195e-01 -1.16641843e+00 -5.02660163e-02 3.07323098e-01
-1.04157388e+00 6.60322130e-01 1.56597763e-01 -7.04670846e-02
9.22319889e-01 4.04382467e-01 -3.37003499e-01 3.49174291e-01
-9.73976076e-01 -3.88936013e-01 -1.16002806e-01 9.14800823e-01
-1.60088539e-01 5.03874063e-01 -2.34165609e-01 8.12043786e-01
1.52388036e-01 -4.39263359e-02 5.90620100e-01 1.14459920e+00
-7.99064696e-01 -7.99286127e-01 -5.12776554e-01 5.55701435e-01
-6.75758839e-01 2.05893993e-01 2.20685929e-01 4.76706415e-01
-4.88702387e-01 1.29115367e+00 1.90745592e-02 6.37099296e-02
3.62021267e-01 1.94315575e-02 7.53788829e-01 -3.10679048e-01
2.25213945e-01 2.66530275e-01 -4.31034081e-02 -5.73229671e-01
-1.42461702e-01 -9.89292979e-01 -1.19418192e+00 -3.94686669e-01
-9.84851494e-02 1.46295756e-01 7.14634478e-01 1.08391488e+00
9.35458988e-02 2.96531051e-01 8.66771281e-01 -5.37859440e-01
-1.92310512e+00 -9.72067237e-01 -1.16640925e+00 -9.66660678e-02
4.32256967e-01 -7.73811221e-01 -5.01155317e-01 -5.13258815e-01] | [4.395322799682617, 2.65397310256958] |
32b0d651-1509-4851-9414-529b27bdd5ed | projected-canonical-decomposition-for | null | null | https://openreview.net/forum?id=ByeAK1BKPB | https://openreview.net/pdf?id=ByeAK1BKPB | Projected Canonical Decomposition for Knowledge Base Completion | The leading approaches to tensor completion and link prediction are based on the canonical polyadic (CP) decomposition of tensors. While these approaches were originally motivated by low rank approximations, the best performances are usually obtained for ranks as high as permitted by computation constraints. For large scale factorization problems where the factor dimensions have to be kept small, the performances of these approaches tend to drop drastically. The other main tensor factorization model, Tucker decomposition, is more flexible than CP for fixed factor dimensions, so we expect Tucker-based approaches to yield better performance under strong constraints on the number of parameters. However, as we show in this paper through experiments on standard benchmarks of link prediction in knowledge bases, ComplEx, a variant of CP, achieves similar performances to recent approaches based on Tucker decomposition on all operating points in terms of number of parameters. In a control experiment, we show that one problem in the practical application of Tucker decomposition to large-scale tensor completion comes from the adaptive optimization algorithms based on diagonal rescaling, such as Adagrad. We present a new algorithm for a constrained version of Tucker which implicitly applies Adagrad to a CP-based model with an additional projection of the embeddings onto a fixed lower dimensional subspace. The resulting Tucker-style extension of ComplEx obtains similar best performances as ComplEx, with substantial gains on some datasets under constraints on the number of parameters. | ['Nicolas Usunier', 'Joan Bruna', 'Guillaume Obozinski', 'Timothée Lacroix'] | 2019-09-25 | null | null | null | null | ['knowledge-base-completion', 'knowledge-base-completion'] | ['graphs', 'knowledge-base'] | [-3.65908742e-01 -3.07347868e-02 -3.53266358e-01 1.09138653e-01
-5.53443313e-01 -5.92348337e-01 5.15296280e-01 5.06814271e-02
-2.92824268e-01 6.13613129e-01 4.88881320e-01 -4.79756713e-01
-5.03961384e-01 -6.48119211e-01 -7.87375033e-01 -6.54499114e-01
-3.83078575e-01 7.71732926e-01 2.29836404e-01 -4.98954445e-01
1.53906429e-02 4.88005191e-01 -1.22480512e+00 2.51169056e-01
6.11035466e-01 6.99830353e-01 -4.19901907e-01 8.33637536e-01
-3.57519239e-02 3.20409507e-01 4.28484334e-03 -7.10289121e-01
6.81665957e-01 6.67105466e-02 -1.05571163e+00 2.23925486e-01
7.58896768e-01 -8.77198204e-02 -4.82348382e-01 8.53373766e-01
2.22705036e-01 2.69044340e-01 4.74095285e-01 -1.22866285e+00
-5.16715288e-01 6.22023106e-01 -4.61717904e-01 -3.68123618e-03
4.31542516e-01 -4.73674804e-01 1.48206913e+00 -1.17101836e+00
7.39775181e-01 1.41147220e+00 9.09952581e-01 4.56676409e-02
-1.99085474e+00 -3.48132700e-02 6.33007735e-02 2.61626363e-01
-1.26452231e+00 -2.96414167e-01 5.07736742e-01 -6.14465833e-01
7.05073833e-01 4.95052159e-01 5.80440760e-01 6.51558995e-01
-2.31547773e-01 7.60395229e-01 9.69590724e-01 -5.08142054e-01
-1.19942043e-03 -4.58649099e-02 4.86321092e-01 6.82489693e-01
3.93046230e-01 -8.12526196e-02 -4.60648775e-01 -7.42364645e-01
5.91243029e-01 -6.33846670e-02 -3.13435346e-01 -8.69242191e-01
-1.72853339e+00 9.16628242e-01 3.73694032e-01 3.01516742e-01
-2.31996745e-01 2.17174828e-01 6.12322867e-01 5.46770811e-01
6.13447368e-01 2.87436306e-01 -5.93470991e-01 -7.35106096e-02
-8.11565638e-01 5.75482845e-01 1.19032490e+00 7.93525100e-01
7.09569454e-01 -1.99830115e-01 -1.69685423e-01 8.75067770e-01
1.88565582e-01 2.64789224e-01 2.18697906e-01 -1.06967294e+00
6.51325405e-01 2.71103621e-01 2.31851399e-01 -1.21595931e+00
-3.64445001e-01 -5.08535266e-01 -6.18356049e-01 -6.28436804e-02
8.05761755e-01 5.70467673e-02 -5.45421243e-01 1.52202547e+00
5.77755332e-01 4.52809855e-02 7.88727950e-04 8.90445530e-01
3.91293839e-02 6.45941317e-01 -3.29444408e-01 -1.61473021e-01
1.32359111e+00 -1.00392604e+00 -5.15395224e-01 3.77474844e-01
1.16094363e+00 -1.02299941e+00 8.64575446e-01 6.03668034e-01
-8.33154440e-01 -2.43386611e-01 -9.42320824e-01 -1.71342924e-01
-2.69031674e-01 1.66079551e-01 1.08609390e+00 7.86876500e-01
-1.18593729e+00 1.07090855e+00 -8.78287554e-01 -5.78472376e-01
-1.74802125e-01 4.91640121e-01 -7.34842062e-01 -2.98392057e-01
-1.00915480e+00 7.22757995e-01 3.89122635e-01 1.94161355e-01
-3.65790755e-01 -7.11566091e-01 -4.77519214e-01 3.90587673e-02
3.52048784e-01 -7.74339199e-01 8.45287025e-01 -6.01436853e-01
-1.29951048e+00 4.82417971e-01 -2.91209146e-02 -3.52838516e-01
4.92356211e-01 -2.67681718e-01 -2.23607361e-01 1.18583500e-01
-1.19419083e-01 2.87764102e-01 8.58646750e-01 -9.62196946e-01
-2.25650951e-01 -3.78790915e-01 2.89205283e-01 -6.46505430e-02
-6.01892292e-01 -4.34416570e-02 -4.04034764e-01 -7.90411770e-01
4.30599630e-01 -1.50929725e+00 -2.05632433e-01 1.29326191e-02
-1.93591475e-01 -3.09647739e-01 6.30127370e-01 -7.45545864e-01
1.14048851e+00 -1.85048616e+00 8.45027804e-01 4.06069994e-01
3.25911522e-01 1.52867332e-01 -3.02942216e-01 6.72251701e-01
-5.15840650e-01 4.42596991e-03 9.46472436e-02 -3.90682399e-01
1.50523216e-01 5.74658036e-01 -3.55625987e-01 6.28914893e-01
-2.85401911e-01 3.52553010e-01 -9.18030560e-01 -2.20114663e-01
-1.88699737e-01 3.91326189e-01 -1.09168839e+00 -9.35229436e-02
-9.37052742e-02 -4.53558154e-02 -3.16181213e-01 4.03655648e-01
6.75860226e-01 -1.14683405e-01 3.74933004e-01 -6.23656273e-01
-6.30075336e-02 1.79668799e-01 -1.55266809e+00 1.70703149e+00
-3.54192317e-01 5.32752812e-01 2.64821529e-01 -1.19807148e+00
6.05838001e-01 4.65453506e-01 7.95956194e-01 8.43818858e-02
-3.65468115e-01 4.15746421e-01 -6.82894588e-02 -1.09773800e-01
7.61827052e-01 -6.09676950e-02 2.31338724e-01 3.88129324e-01
-7.21638091e-03 1.82091758e-01 6.86907530e-01 6.26809835e-01
1.29011691e+00 3.45592678e-01 -2.46419922e-01 -4.77438748e-01
4.53333855e-01 8.79916623e-02 5.67559242e-01 2.70333320e-01
2.29327530e-01 3.81157607e-01 9.13127065e-01 -6.01523221e-01
-1.37119985e+00 -7.72788405e-01 -3.86029154e-01 1.24522805e+00
-4.28030670e-01 -1.12159634e+00 -5.69070399e-01 -6.70479536e-01
2.44671866e-01 2.31807843e-01 -5.36643922e-01 2.91147113e-01
-5.47804236e-01 -1.06274819e+00 3.86868685e-01 3.18675101e-01
-3.11893225e-02 -9.29695666e-02 3.27104121e-01 3.10261548e-01
-9.02896672e-02 -1.18215036e+00 -3.96037102e-01 -8.80927965e-03
-1.40540314e+00 -9.86540020e-01 -9.44442868e-01 -2.69802600e-01
5.41699052e-01 3.50229770e-01 8.61438334e-01 2.18264796e-02
9.67204645e-02 5.01863718e-01 -4.21692759e-01 3.65588397e-01
-3.66671592e-01 1.51087388e-01 5.93266487e-01 1.55310556e-01
-1.09867966e-02 -8.23781967e-01 -4.44990426e-01 4.71690387e-01
-1.01722634e+00 -1.00615785e-01 3.85617316e-01 1.14448237e+00
3.15654188e-01 -1.04782015e-01 -5.81089780e-02 -8.06924403e-01
6.42114758e-01 -3.64445776e-01 -4.81218934e-01 1.53093845e-01
-7.12680459e-01 6.30435050e-01 6.89739823e-01 -3.76483619e-01
-4.07712281e-01 -8.95731524e-02 1.78406596e-01 -7.09076405e-01
5.14088929e-01 6.52954400e-01 4.69944067e-02 -3.69399756e-01
7.45281816e-01 -1.49752364e-01 -3.77326943e-02 -9.73057210e-01
6.67835772e-01 3.91803324e-01 1.45900100e-01 -1.18926823e+00
1.01349437e+00 4.26508844e-01 3.99487048e-01 -9.11175668e-01
-7.74399817e-01 -4.85270947e-01 -7.57506251e-01 8.91581178e-02
3.79775763e-01 -9.01494741e-01 -6.99907184e-01 -1.01634562e-01
-1.01566732e+00 1.33365663e-02 -9.41161588e-02 7.27669358e-01
-6.26367331e-01 8.92187715e-01 -7.49425173e-01 -3.16810697e-01
-1.29252765e-02 -9.21627641e-01 7.92974353e-01 -5.93681574e-01
-7.34582245e-02 -1.07828844e+00 4.25241828e-01 6.02083504e-01
2.26106495e-01 7.11105540e-02 1.12064421e+00 -5.14554739e-01
-5.13675511e-01 -3.72664630e-01 -1.44454747e-01 5.30625761e-01
-2.58160084e-01 8.90145153e-02 -5.11204183e-01 -6.01228833e-01
-4.33792770e-01 -1.82258226e-02 8.41742218e-01 -1.25675768e-01
6.71721399e-01 -4.36707795e-01 -2.86533862e-01 6.84718490e-01
1.39303350e+00 -6.51216209e-01 4.86713052e-01 2.94616878e-01
1.03641033e+00 4.66133863e-01 5.43401778e-01 3.49867612e-01
2.60013819e-01 1.02794611e+00 2.52354890e-01 1.60813674e-01
-1.13386000e-02 -2.84566224e-01 5.50560892e-01 1.38613033e+00
-7.41671920e-01 3.08120757e-01 -8.76500428e-01 5.57741880e-01
-2.12669516e+00 -8.12609315e-01 -5.48209012e-01 2.46911263e+00
7.32264817e-01 1.02583133e-01 4.63885486e-01 2.76234925e-01
4.79051948e-01 9.87423733e-02 -3.41003016e-02 -5.41385174e-01
1.28370652e-03 4.16057408e-02 8.71497929e-01 5.99571645e-01
-9.38217700e-01 7.27045894e-01 6.68102169e+00 8.68757129e-01
-6.55718863e-01 3.74008447e-01 3.38324085e-02 -1.35394961e-01
-3.10582519e-01 4.39533472e-01 -7.13940263e-01 1.75089687e-01
9.88788486e-01 3.68333422e-02 6.83198810e-01 9.06134129e-01
7.02984184e-02 1.35273874e-01 -1.33295012e+00 9.33402419e-01
-1.76568434e-01 -1.19770110e+00 -7.00799562e-03 3.95965189e-01
9.21335340e-01 1.67761311e-01 -1.78678304e-01 2.62103945e-01
2.78516263e-01 -6.29812121e-01 4.10971731e-01 3.80252182e-01
4.17605937e-01 -6.35957658e-01 4.17491525e-01 4.38826084e-02
-1.06650531e+00 5.38375378e-02 -7.62093961e-01 -1.04215555e-01
-9.90020186e-02 9.74998951e-01 -5.18650293e-01 7.98448980e-01
3.33968580e-01 6.51267648e-01 -3.68280947e-01 9.88236904e-01
4.06097341e-03 6.60674155e-01 -6.90652847e-01 3.26139152e-01
2.74638325e-01 -5.91037095e-01 8.64704549e-01 1.17108393e+00
2.86995918e-01 -5.12814075e-02 3.93990695e-01 2.99696624e-01
5.43263629e-02 3.83504987e-01 -4.25698459e-01 -1.50786936e-01
1.00143412e-02 1.34031379e+00 -3.92399132e-01 -2.90902436e-01
-4.96436805e-01 7.98728824e-01 5.52278042e-01 4.57461089e-01
-4.97465640e-01 -7.48825446e-02 8.43216062e-01 4.48010057e-01
5.44281125e-01 -7.81727195e-01 5.70750348e-02 -1.59352231e+00
2.80139893e-01 -8.77568781e-01 5.85584521e-01 -4.08564836e-01
-1.14090264e+00 3.84094536e-01 7.05259219e-02 -1.25746095e+00
-1.15078703e-01 -9.71883893e-01 -9.05419365e-02 7.46543884e-01
-1.14291298e+00 -1.24602103e+00 3.78295898e-01 6.57600462e-01
-7.10949302e-03 -4.18733507e-02 9.60706651e-01 5.21651328e-01
-6.09354258e-01 4.38873947e-01 7.85757840e-01 -4.02816311e-02
7.26638198e-01 -1.37493980e+00 -6.45284541e-03 8.44600320e-01
3.27291131e-01 7.49836147e-01 1.11916375e+00 -3.95151228e-01
-1.90363026e+00 -7.35731483e-01 8.63863349e-01 -3.68423283e-01
1.16323423e+00 -3.07983845e-01 -8.30458939e-01 9.91988182e-01
-1.43233046e-01 3.16765547e-01 6.67073369e-01 9.58415151e-01
-6.67266190e-01 -3.35425079e-01 -7.37034559e-01 5.99669874e-01
9.91235375e-01 -5.31321228e-01 -4.42299157e-01 9.22262430e-01
6.23528957e-01 -2.03853339e-01 -1.65261471e+00 1.51969910e-01
6.35611653e-01 -7.89341688e-01 1.05635130e+00 -1.01791322e+00
2.74051815e-01 -5.69562197e-01 -5.11555493e-01 -1.28594720e+00
-6.27851009e-01 -9.13378477e-01 -4.70112413e-01 9.00613725e-01
2.29546815e-01 -5.50164342e-01 8.49942029e-01 6.64483011e-01
1.15931183e-01 -8.28039289e-01 -1.02180576e+00 -1.04422474e+00
1.66384593e-01 -3.91696304e-01 3.55777055e-01 1.07970202e+00
2.36798555e-01 3.91513914e-01 -5.30084372e-01 1.44514799e-01
8.17648351e-01 7.38881901e-02 1.03249061e+00 -1.25783408e+00
-7.38094926e-01 -1.25948668e-01 -6.49483740e-01 -1.10892498e+00
6.21280335e-02 -1.16059220e+00 -5.94662189e-01 -1.17837393e+00
1.15642697e-01 -7.04641223e-01 -1.35817736e-01 3.81562889e-01
1.00147724e-01 2.66519755e-01 4.56226766e-01 5.06763339e-01
-2.97893167e-01 3.99036258e-01 1.15170956e+00 -1.35028780e-01
-9.50917080e-02 -5.14651202e-02 -3.68413746e-01 6.45827830e-01
3.39658588e-01 -2.42669120e-01 -1.58215329e-01 -4.99141604e-01
6.97203696e-01 2.54803300e-01 1.26465350e-01 -7.23997593e-01
1.58906549e-01 1.76934917e-02 1.90688372e-02 -4.71518219e-01
3.62998903e-01 -7.05420434e-01 4.51195449e-01 2.52339184e-01
-1.62500888e-01 1.36385038e-01 -9.18773413e-02 6.46554530e-01
-1.24140114e-01 -2.29752421e-01 4.14661318e-01 1.32627502e-01
-3.10528517e-01 3.08967978e-01 -9.76628885e-02 -3.92170697e-01
5.81386149e-01 -1.25252411e-01 -1.89359292e-01 -1.91988453e-01
-1.22852814e+00 6.92533189e-03 4.09415454e-01 1.98806390e-01
1.72656059e-01 -1.71216488e+00 -8.08187604e-01 -1.16890535e-01
-1.19508002e-02 -3.42799246e-01 -6.71307668e-02 1.37541354e+00
-5.86150527e-01 6.54839993e-01 6.75422847e-02 -5.80550849e-01
-1.34163642e+00 8.94792676e-01 9.16148350e-02 -7.37322927e-01
-5.29595971e-01 4.68195915e-01 -6.19514212e-02 -4.56329286e-01
6.52787387e-02 -3.25801760e-01 1.93712458e-01 3.30004275e-01
3.27405125e-01 6.94668591e-01 3.95660877e-01 -6.69593215e-01
-1.78085089e-01 6.12090409e-01 -2.29337007e-01 -1.27852838e-02
1.50884092e+00 -1.46323554e-02 -5.15752256e-01 2.59508878e-01
1.35899138e+00 2.41537869e-01 -8.40056956e-01 -5.06435394e-01
1.24241687e-01 -6.67414904e-01 1.94526032e-01 -2.67126769e-01
-1.06795418e+00 6.07030332e-01 3.54177713e-01 2.71195918e-01
7.34712005e-01 -1.83768332e-01 6.73424065e-01 7.41509914e-01
5.35191834e-01 -9.10352468e-01 -2.42554843e-01 5.42992711e-01
8.72034490e-01 -7.90326118e-01 5.12008011e-01 -5.88725984e-01
-1.06671959e-01 1.36957896e+00 1.72225367e-02 -2.96160728e-01
6.63878441e-01 -1.21648759e-01 -3.92023623e-01 -2.53626797e-02
-8.75516772e-01 -1.00130767e-01 4.41190124e-01 2.92616397e-01
4.90548432e-01 2.98861146e-01 -7.01536119e-01 3.52496170e-02
-1.97004527e-01 -2.06029475e-01 6.25468493e-01 5.60554445e-01
-2.53144383e-01 -1.64715338e+00 -8.15183282e-01 3.75345826e-01
-5.00624955e-01 -9.21363607e-02 -3.18708532e-02 6.36281252e-01
-7.31983706e-02 5.74397862e-01 -2.21480995e-01 -4.02775913e-01
2.61985660e-01 1.31980479e-01 7.03719616e-01 -3.83934468e-01
-4.80892956e-01 1.74515575e-01 5.81960022e-01 -7.37099528e-01
-5.72547495e-01 -8.65393102e-01 -7.23525882e-01 -8.18470597e-01
-3.01587582e-01 4.24067676e-01 7.47406602e-01 7.98795164e-01
3.29996139e-01 -2.57514901e-02 5.95211625e-01 -9.08362210e-01
-9.02404487e-01 -8.77514780e-01 -8.75836551e-01 5.23914218e-01
5.19887358e-02 -7.51871109e-01 -3.13646913e-01 -5.08575998e-02] | [7.0649824142456055, 4.764883041381836] |
f2000929-13b1-4a65-81d7-2e7c43ecdf68 | ubernet-training-a-universal-convolutional | 1609.02132 | null | http://arxiv.org/abs/1609.02132v1 | http://arxiv.org/pdf/1609.02132v1.pdf | UberNet: Training a `Universal' Convolutional Neural Network for Low-, Mid-, and High-Level Vision using Diverse Datasets and Limited Memory | In this work we introduce a convolutional neural network (CNN) that jointly
handles low-, mid-, and high-level vision tasks in a unified architecture that
is trained end-to-end. Such a universal network can act like a `swiss knife'
for vision tasks; we call this architecture an UberNet to indicate its
overarching nature.
We address two main technical challenges that emerge when broadening up the
range of tasks handled by a single CNN: (i) training a deep architecture while
relying on diverse training sets and (ii) training many (potentially unlimited)
tasks with a limited memory budget. Properly addressing these two problems
allows us to train accurate predictors for a host of tasks, without
compromising accuracy.
Through these advances we train in an end-to-end manner a CNN that
simultaneously addresses (a) boundary detection (b) normal estimation (c)
saliency estimation (d) semantic segmentation (e) human part segmentation (f)
semantic boundary detection, (g) region proposal generation and object
detection. We obtain competitive performance while jointly addressing all of
these tasks in 0.7 seconds per frame on a single GPU. A demonstration of this
system can be found at http://cvn.ecp.fr/ubernet/. | ['Iasonas Kokkinos'] | 2016-09-07 | null | null | null | null | ['human-part-segmentation'] | ['computer-vision'] | [ 3.24633360e-01 2.31001392e-01 6.37794435e-02 -2.20564395e-01
-6.93476379e-01 -3.79878819e-01 5.87389946e-01 1.59774870e-01
-6.77553535e-01 4.67278183e-01 -3.26896727e-01 -3.66931111e-01
5.49694002e-01 -5.26653707e-01 -1.11082077e+00 -4.64588314e-01
6.81180432e-02 4.65700597e-01 8.75417829e-01 -2.03066573e-01
1.26656175e-01 4.90116477e-01 -1.87006545e+00 2.77235180e-01
8.08584988e-01 1.34709644e+00 6.74695671e-01 9.73453820e-01
1.67853609e-01 5.24055898e-01 -4.49794263e-01 -3.00795913e-01
4.05689150e-01 -1.02439150e-01 -8.39830220e-01 1.88197538e-01
7.03099966e-01 -4.00660634e-01 1.17057905e-01 9.46772158e-01
4.17955458e-01 4.12980467e-02 4.19670612e-01 -1.14326143e+00
-2.36622512e-01 2.39670992e-01 -6.88334882e-01 3.35452735e-01
2.19859686e-02 3.03492606e-01 7.57523060e-01 -8.77797306e-01
6.76604629e-01 9.67304409e-01 6.90543532e-01 5.79072356e-01
-1.14154649e+00 -2.56698072e-01 2.10567221e-01 -6.62257969e-02
-1.10566676e+00 -4.57318842e-01 4.81076300e-01 -3.89585376e-01
8.79410625e-01 8.65125880e-02 7.16603935e-01 1.01588213e+00
6.99408576e-02 1.08628941e+00 7.94605494e-01 -3.02999467e-01
3.33796203e-01 -9.16207656e-02 2.66710520e-01 7.57719398e-01
2.29264140e-01 -6.91588828e-03 -3.68584901e-01 1.40610158e-01
9.97277856e-01 -8.71789381e-02 -8.73791799e-02 -3.37685734e-01
-1.37213004e+00 7.80427217e-01 7.90715158e-01 2.87770659e-01
-3.01494658e-01 4.27156955e-01 5.83378851e-01 -1.07471729e-02
3.10324460e-01 3.29485297e-01 -6.86925769e-01 2.20759273e-01
-1.00303400e+00 3.45018089e-01 7.30276883e-01 9.67214763e-01
8.65757585e-01 5.77758662e-02 -7.06539080e-02 7.00874388e-01
1.21800952e-01 2.65782088e-01 3.37185591e-01 -9.15795326e-01
1.66432217e-01 2.53984362e-01 2.68521011e-01 -5.80609322e-01
-7.91181862e-01 -6.90620422e-01 -8.02252293e-01 5.16867578e-01
6.81053996e-01 -3.59281540e-01 -1.25436962e+00 1.62755370e+00
4.63315070e-01 1.23688690e-01 -2.22078845e-01 1.14606655e+00
9.33487952e-01 4.73474532e-01 2.39843547e-01 3.44730228e-01
1.75206125e+00 -1.42557323e+00 -1.53692082e-01 -6.54009163e-01
3.45938832e-01 -8.19802344e-01 9.29836392e-01 3.19514096e-01
-1.33757174e+00 -7.87948310e-01 -1.03040290e+00 -5.04966736e-01
-4.36701208e-01 4.30454284e-01 6.99708283e-01 3.16224396e-01
-1.59824598e+00 5.54327130e-01 -9.89152074e-01 -4.01962787e-01
6.39293373e-01 5.39019048e-01 -1.86855182e-01 1.76453739e-01
-7.26649523e-01 8.34380746e-01 5.02355695e-01 1.81292012e-01
-9.08747733e-01 -5.89656472e-01 -8.36001635e-01 9.26089957e-02
5.13999045e-01 -1.16787052e+00 1.47762001e+00 -1.25319731e+00
-1.24462485e+00 1.19496250e+00 -7.71699622e-02 -6.61229849e-01
7.13443995e-01 -3.25669765e-01 6.43776283e-02 2.02700838e-01
3.05202723e-01 1.24806154e+00 9.31889534e-01 -1.23032153e+00
-1.02985239e+00 -3.38382870e-01 -3.60565260e-02 1.52316377e-01
2.84642011e-01 2.42674351e-01 -6.70964479e-01 -7.28265047e-01
1.51443630e-01 -8.39752913e-01 -3.21397334e-01 3.86621267e-01
-4.93883193e-01 -2.06101686e-01 9.28077757e-01 -5.83610177e-01
5.20103693e-01 -1.95044160e+00 1.88288882e-01 -2.49854267e-01
4.35174316e-01 6.27525032e-01 -4.16581519e-02 -2.83147451e-02
5.10544293e-02 -1.64641827e-01 -5.12751758e-01 -7.64949977e-01
-2.03524217e-01 -7.17224255e-02 3.44670713e-02 4.05613989e-01
3.09904367e-01 1.11852241e+00 -8.06370139e-01 -3.34046215e-01
3.31925601e-01 4.44730431e-01 -4.02782053e-01 2.29012430e-01
-4.78543907e-01 5.77577293e-01 -2.90424168e-01 7.55179405e-01
6.24988258e-01 -5.02040029e-01 -1.76712289e-01 -1.65562496e-01
-3.03045303e-01 9.36530009e-02 -9.91475880e-01 1.92706275e+00
-3.87709647e-01 7.35714316e-01 6.36171699e-01 -1.00986719e+00
6.96342230e-01 4.81984057e-02 2.01425388e-01 -5.16545892e-01
4.28210258e-01 4.14927989e-01 -1.34466454e-01 -2.46349230e-01
5.54543793e-01 1.28538966e-01 3.10514141e-02 2.77936637e-01
4.13031250e-01 1.25479072e-01 2.06855714e-01 -2.39477940e-02
9.41012979e-01 2.08484456e-01 2.69227386e-01 -5.66657960e-01
4.81441110e-01 1.29863426e-01 4.51984257e-01 8.19961607e-01
-5.98519087e-01 8.27706873e-01 4.35196519e-01 -7.63589859e-01
-1.22334516e+00 -9.88760054e-01 -1.27071932e-01 1.42157125e+00
3.17163616e-01 2.39035767e-02 -8.56687665e-01 -4.91103083e-01
-7.62580633e-02 2.63127387e-01 -7.06037283e-01 1.21897921e-01
-7.35955834e-01 -5.60491800e-01 3.98427159e-01 7.60242343e-01
7.34244227e-01 -1.32304370e+00 -1.22934949e+00 1.43805474e-01
3.54287252e-02 -1.24134684e+00 -3.75925571e-01 4.54141796e-01
-7.65678763e-01 -9.73831594e-01 -9.75651264e-01 -1.05121112e+00
4.65234995e-01 3.96074384e-01 1.39368320e+00 9.32819992e-02
-4.15641397e-01 1.30713031e-01 -2.61257589e-01 -4.50851530e-01
-2.80159801e-01 2.67829388e-01 -2.36828551e-01 -6.05933964e-02
-2.14659646e-02 -3.74099076e-01 -9.58240092e-01 2.60518700e-01
-8.76400471e-01 3.83156389e-01 8.33031774e-01 6.83190703e-01
6.11009896e-01 -6.09128416e-01 4.73222136e-01 -6.71343744e-01
8.21641833e-02 -3.82596493e-01 -7.70439982e-01 1.18383199e-01
-1.00706644e-01 -1.98335290e-01 4.25232470e-01 -1.16668984e-01
-8.33213806e-01 3.17490101e-01 -5.19816577e-01 -4.01632309e-01
-5.32054842e-01 6.50111213e-02 4.71546464e-02 -2.13865057e-01
5.79581678e-01 1.80432931e-01 1.20090522e-01 -4.78557914e-01
4.63511288e-01 4.48196411e-01 1.01633799e+00 -3.56851429e-01
4.77305919e-01 5.87978244e-01 1.42098609e-02 -8.17118943e-01
-1.02964437e+00 -6.84156895e-01 -8.63454461e-01 -1.93823203e-01
1.34721136e+00 -1.13171065e+00 -7.15751588e-01 7.24532783e-01
-1.26255715e+00 -6.90943956e-01 -1.88271552e-01 1.16003908e-01
-6.97977245e-01 1.53035149e-01 -5.62477291e-01 -5.15266061e-01
-6.15532637e-01 -1.19976211e+00 1.54350686e+00 5.54275036e-01
-3.26521951e-03 -9.30485129e-01 -2.75476724e-01 4.92907882e-01
5.98026633e-01 5.87338209e-01 3.03824097e-01 -6.36797607e-01
-7.67608047e-01 4.03155833e-02 -6.70049906e-01 2.50850677e-01
-3.69970590e-01 -1.32422388e-01 -1.14409685e+00 -3.24498087e-01
-2.28331164e-02 -5.43809116e-01 1.18090010e+00 5.76603651e-01
1.15264153e+00 6.85425848e-02 -2.94962525e-01 9.34345782e-01
1.46523750e+00 -2.07147092e-01 4.50095505e-01 5.03490269e-01
7.94694245e-01 4.85674024e-01 4.70568776e-01 1.93203583e-01
3.70614529e-01 7.29389250e-01 7.09211111e-01 -4.96100098e-01
-3.33737254e-01 1.72428340e-01 1.48832649e-01 9.03846174e-02
-9.80438963e-02 -6.55048192e-02 -1.05538547e+00 7.91911185e-01
-2.02550077e+00 -6.73190594e-01 -4.35479313e-01 1.89123321e+00
6.99229538e-01 2.76839495e-01 4.45866704e-01 -2.10217506e-01
9.01625454e-01 1.22559927e-01 -5.93265653e-01 -3.77841502e-01
-2.84845810e-02 2.39220068e-01 5.06519079e-01 2.76433617e-01
-1.47842872e+00 1.20173347e+00 5.65152216e+00 7.77466834e-01
-1.27373695e+00 2.20910192e-01 8.91187608e-01 8.41462314e-02
1.79122016e-01 -2.74068236e-01 -8.18926871e-01 3.97183836e-01
7.57490695e-01 3.51145893e-01 1.61541119e-01 1.01757073e+00
-1.22682154e-01 -3.72013777e-01 -9.33899999e-01 9.85578656e-01
3.97103140e-03 -1.49278784e+00 -1.55914098e-01 -1.63009793e-01
6.88569069e-01 6.32774413e-01 -6.58601448e-02 1.39759645e-01
1.51572078e-01 -9.93311405e-01 1.14193428e+00 3.08253139e-01
7.98731506e-01 -5.20759106e-01 5.95570505e-01 3.75634730e-01
-1.10462642e+00 -1.86609756e-02 -3.93207282e-01 1.35946900e-01
1.09591305e-01 6.60746276e-01 -5.47468603e-01 4.64015633e-01
8.08275044e-01 4.08725023e-01 -6.32605851e-01 1.25762212e+00
-1.99593470e-01 2.74997205e-01 -4.77192551e-01 2.13138834e-01
5.54421067e-01 2.93077417e-02 5.46728075e-01 1.36972976e+00
7.91623890e-02 -1.84149265e-01 2.36610800e-01 1.09020281e+00
-1.47459030e-01 -3.02663475e-01 -1.98837757e-01 4.19775546e-01
1.28485397e-01 1.45412493e+00 -1.36084759e+00 -3.64985019e-01
-3.60665649e-01 1.09513390e+00 3.96244287e-01 1.69802055e-01
-9.08296466e-01 -3.21255982e-01 6.40772045e-01 -2.47887219e-03
6.25752687e-01 -2.93720901e-01 -5.75957775e-01 -1.10621238e+00
3.45718814e-03 -4.74773020e-01 2.06123963e-01 -8.98416579e-01
-9.53861415e-01 6.92985058e-01 -3.51454765e-01 -8.20646524e-01
-1.44504324e-01 -7.53171504e-01 -7.50104964e-01 7.85869539e-01
-1.69498301e+00 -1.20913637e+00 -6.00530505e-01 4.83168960e-01
7.67505705e-01 1.81947928e-02 5.26241004e-01 1.56374887e-01
-6.75680697e-01 3.64701241e-01 -1.88597620e-01 2.49532908e-01
4.27947015e-01 -1.45296192e+00 7.94892371e-01 9.38291788e-01
-9.73017737e-02 2.83307612e-01 7.27395117e-01 -4.06274110e-01
-1.01324415e+00 -1.32933140e+00 7.79484570e-01 -3.35980654e-01
4.86668646e-01 -4.51566577e-01 -8.94675195e-01 7.21062183e-01
2.77981818e-01 4.33341801e-01 2.27072909e-01 -6.87125996e-02
-1.18909761e-01 1.84014335e-01 -9.99461472e-01 5.86454272e-01
9.13590670e-01 -1.79995060e-01 -4.88209575e-01 3.91485244e-01
8.15979481e-01 -7.96273112e-01 -4.59588766e-01 2.80209064e-01
3.70039046e-01 -1.42081666e+00 1.07984805e+00 -3.61570358e-01
6.59409106e-01 -3.53224695e-01 1.84078351e-01 -8.61527503e-01
-1.26470461e-01 -7.38550782e-01 -2.30120674e-01 7.56460607e-01
2.78503567e-01 -4.98253375e-01 6.85222447e-01 3.77774596e-01
-5.65804422e-01 -1.06283915e+00 -9.31486011e-01 -6.36914074e-01
-3.94103788e-02 -3.37939739e-01 3.38885546e-01 4.55410272e-01
-5.98522842e-01 3.46456736e-01 -1.03522569e-01 9.51159745e-02
5.57279408e-01 2.90132940e-01 6.66930795e-01 -1.23416603e+00
-2.36538276e-01 -6.90404713e-01 -4.24064249e-01 -1.30020416e+00
-2.68235710e-02 -7.81075180e-01 2.08340064e-01 -1.54937696e+00
1.78829446e-01 -3.61947566e-01 -3.51666845e-02 4.54964548e-01
-1.67778939e-01 5.59381902e-01 2.85261363e-01 2.71255016e-01
-8.30990911e-01 2.66484231e-01 1.13732386e+00 2.04871625e-01
4.65759784e-02 2.26782471e-01 -6.52954221e-01 8.79632115e-01
8.17470551e-01 -1.57412156e-01 -1.12748407e-02 -6.72127426e-01
-2.09330916e-02 1.96532663e-02 9.04148698e-01 -1.25383210e+00
3.42666745e-01 2.13888109e-01 4.68089372e-01 -5.74835062e-01
3.32719505e-01 -5.08941650e-01 -3.77286315e-01 4.95947689e-01
-1.64468393e-01 -1.26691505e-01 5.19610226e-01 4.92619306e-01
-1.41799241e-01 -2.02783182e-01 1.22438419e+00 -2.41592109e-01
-1.06863451e+00 3.10760677e-01 -2.21787959e-01 5.65327611e-03
1.11553717e+00 -2.89528340e-01 -4.27433163e-01 3.41044515e-02
-9.02843773e-01 2.38248259e-01 6.89189374e-01 1.84273422e-01
3.26359540e-01 -8.62082660e-01 -6.44282341e-01 2.95558810e-01
-8.44476558e-03 4.75427091e-01 2.29288325e-01 8.72172892e-01
-8.97711754e-01 4.64474946e-01 -5.16207159e-01 -9.03637707e-01
-1.12981534e+00 3.87144089e-01 4.63141829e-01 -1.84277609e-01
-8.19600165e-01 1.23349798e+00 2.38489985e-01 -7.88560510e-02
2.52872229e-01 -3.11950833e-01 -1.12115636e-01 1.01933547e-03
4.43898290e-01 1.85329661e-01 1.84016749e-01 -6.19209290e-01
-3.57446879e-01 4.12258267e-01 -1.37987500e-02 1.12956278e-01
1.34307790e+00 -6.30563572e-02 -1.28773049e-01 1.43001392e-01
1.05612898e+00 -5.67954898e-01 -1.77456987e+00 6.51097856e-03
-9.31022912e-02 -1.89313695e-01 2.65188068e-01 -7.82973886e-01
-1.17786527e+00 8.96090746e-01 4.48895991e-01 2.98746139e-01
1.10992086e+00 1.11353837e-01 1.08326197e+00 1.50538042e-01
3.23745370e-01 -9.97614026e-01 -2.49320157e-02 6.14310920e-01
7.99453378e-01 -1.51732397e+00 -2.10838348e-01 -4.23267245e-01
-6.57716870e-01 1.17612183e+00 5.99154353e-01 -4.22933668e-01
5.26964247e-01 3.49832267e-01 1.60207432e-02 -2.76517808e-01
-6.86474800e-01 -7.47769833e-01 3.48867744e-01 5.96591949e-01
2.74008304e-01 -2.52043996e-02 -1.09632649e-01 3.77700210e-01
-9.10867564e-03 -7.53299519e-02 6.08277202e-01 9.31702495e-01
-7.54658163e-01 -8.01994443e-01 -2.91692466e-01 3.57788444e-01
-5.17105579e-01 3.30691077e-02 -3.23551446e-01 8.14176083e-01
4.14022386e-01 7.32675314e-01 3.06238294e-01 -1.81693211e-02
1.09862149e-01 -6.34920746e-02 2.81613410e-01 -8.09539080e-01
-8.02425027e-01 1.13903917e-01 -9.60868374e-02 -7.86612570e-01
-3.90789598e-01 -5.90430260e-01 -1.13274205e+00 -9.37861204e-02
3.60331163e-02 -3.03724945e-01 8.18855226e-01 9.90163863e-01
4.31479782e-01 7.26839185e-01 6.46345168e-02 -1.52059948e+00
-4.60703433e-01 -8.13857019e-01 -4.44573373e-01 1.89352661e-01
5.38665056e-01 -6.05015516e-01 -5.50671965e-02 1.51415035e-01] | [9.385237693786621, 0.12311527878046036] |
e9c32e11-2547-4997-a4fb-dc226e8e5454 | can-chatgpt-enable-its-the-case-of-mixed | 2306.08094 | null | https://arxiv.org/abs/2306.08094v1 | https://arxiv.org/pdf/2306.08094v1.pdf | Can ChatGPT Enable ITS? The Case of Mixed Traffic Control via Reinforcement Learning | The surge in Reinforcement Learning (RL) applications in Intelligent Transportation Systems (ITS) has contributed to its growth as well as highlighted key challenges. However, defining objectives of RL agents in traffic control and management tasks, as well as aligning policies with these goals through an effective formulation of Markov Decision Process (MDP), can be challenging and often require domain experts in both RL and ITS. Recent advancements in Large Language Models (LLMs) such as GPT-4 highlight their broad general knowledge, reasoning capabilities, and commonsense priors across various domains. In this work, we conduct a large-scale user study involving 70 participants to investigate whether novices can leverage ChatGPT to solve complex mixed traffic control problems. Three environments are tested, including ring road, bottleneck, and intersection. We find ChatGPT has mixed results. For intersection and bottleneck, ChatGPT increases number of successful policies by 150% and 136% compared to solely beginner capabilities, with some of them even outperforming experts. However, ChatGPT does not provide consistent improvements across all scenarios. | ['Weizi Li', 'Bibek Poudel', 'Michael Villarreal'] | 2023-06-13 | null | null | null | null | ['general-knowledge', 'management'] | ['miscellaneous', 'miscellaneous'] | [-2.2058128e-01 8.0360621e-02 -3.5129663e-01 -2.8279167e-01
-8.6241162e-01 -6.5459764e-01 5.9519422e-01 -1.0101039e-01
-6.6825491e-01 1.1944137e+00 3.6956513e-01 -6.6077387e-01
-1.8493491e-01 -6.4309531e-01 -3.8538200e-01 -3.5954425e-01
6.1558355e-02 9.6351367e-01 1.7499581e-01 -6.3815391e-01
2.5595176e-01 4.3233058e-01 -1.4044195e+00 9.6755564e-02
1.1637554e+00 2.5690299e-01 1.8503578e-01 7.6729423e-01
-1.5924340e-01 1.3328212e+00 -4.3674237e-01 -3.2109487e-01
1.6440110e-01 4.0699426e-02 -9.2314142e-01 -2.2049083e-01
3.2688659e-01 -4.3369988e-01 -6.6868526e-01 5.6035703e-01
7.5891685e-01 8.1113786e-01 6.0373950e-01 -1.7779955e+00
-4.8448092e-01 4.8542613e-01 -3.7222019e-01 1.2897809e-01
5.8115458e-01 1.1540529e+00 1.0117135e+00 1.7369475e-02
4.8641217e-01 1.8315736e+00 3.6526427e-01 7.7751309e-01
-1.2827182e+00 -8.5964912e-01 4.3097815e-01 3.6209676e-01
-9.3276632e-01 -5.4734236e-01 2.9465497e-01 -2.7693999e-01
1.3353310e+00 -1.8021379e-01 5.2548051e-01 1.5103402e+00
1.0119785e-01 1.0708331e+00 1.4824333e+00 -6.9703430e-02
4.2734638e-01 8.5159540e-02 1.5203071e-01 6.2505311e-01
-6.5732993e-02 4.6722007e-01 -4.0947452e-01 3.9808344e-02
5.6142849e-01 -6.4410287e-01 5.2371520e-01 1.6435349e-01
-1.0432692e+00 8.8357514e-01 -2.8416185e-02 -2.5024638e-01
-6.4701474e-01 4.2024377e-01 2.4229833e-01 2.6883456e-01
-4.1616764e-02 6.3546109e-01 -1.7836215e-01 -8.8094771e-01
-3.7860358e-01 8.8002425e-01 9.2850304e-01 1.1334617e+00
5.5181915e-01 1.9690067e-01 -7.9834473e-01 8.7830490e-01
3.3464816e-01 9.5965612e-01 -1.3945417e-01 -1.6540875e+00
8.0298322e-01 2.4336247e-02 5.8376163e-01 -9.1124511e-01
-7.3216492e-01 9.6672745e-03 -2.0377001e-01 2.7452973e-01
6.1605591e-01 -8.2670444e-01 -8.5835636e-01 2.0322125e+00
1.3656323e-01 2.6734889e-01 9.2842638e-02 5.9211028e-01
3.9932272e-01 6.9129825e-01 9.6041286e-01 2.5558138e-01
1.1922572e+00 -7.9127985e-01 -6.7373621e-01 -8.8372010e-01
5.9218341e-01 -4.3244204e-01 1.3096522e+00 4.3034685e-01
-1.1946661e+00 -4.7954535e-01 -6.2366223e-01 -4.1295309e-02
-2.9576096e-01 -2.8906330e-01 6.7079002e-01 7.6625562e-01
-1.0835483e+00 1.6834900e-01 -6.4074719e-01 -5.3294992e-01
6.0245949e-01 2.1125719e-01 3.5991624e-02 -4.4020876e-01
-1.5803906e+00 1.6386504e+00 1.6769120e-01 -2.7835512e-01
-1.3256161e+00 -7.6166004e-01 -9.1234481e-01 -7.3294453e-03
6.1793876e-01 -7.1623713e-01 1.8188016e+00 -3.2985440e-01
-1.9009762e+00 3.9029554e-01 -2.9942885e-01 -5.9315383e-01
5.8988297e-01 -2.7189767e-01 -4.9910399e-01 -3.1655099e-02
4.3627357e-01 1.1212916e+00 4.4018987e-01 -1.0787275e+00
-1.1186472e+00 6.0754370e-02 4.5450050e-01 5.1286340e-01
3.7328854e-01 7.4350730e-02 -2.0896027e-02 3.0167649e-02
-6.6650951e-01 -1.0021461e+00 -4.1710460e-01 -4.5278248e-01
-7.5333841e-02 -6.7320037e-01 5.7243395e-01 -6.3558137e-01
9.9981159e-01 -1.8473321e+00 -2.8237757e-01 1.6862382e-01
1.0497403e-01 2.6772502e-01 -5.6209505e-01 5.3713286e-01
3.2542893e-01 3.3273607e-01 1.4709094e-01 -1.6381721e-01
6.1519200e-01 7.1539885e-01 -2.2444244e-01 4.1619278e-02
3.9591858e-01 1.0334830e+00 -1.2740679e+00 -6.2772489e-01
6.0734367e-01 -4.9248144e-02 -6.7547870e-01 4.1125603e-02
-7.0018095e-01 6.2260860e-01 -4.8580024e-01 2.7980429e-01
4.7280836e-01 1.2246817e-01 -1.7649574e-03 4.8884410e-01
-3.0599976e-01 5.5349493e-01 -1.0759841e+00 1.2922981e+00
-6.6271710e-01 8.9098591e-01 -1.9824289e-02 -6.2396705e-01
4.0737846e-01 2.3099908e-01 3.2589081e-01 -1.3827171e+00
-1.2395824e-01 -3.5915667e-01 2.4110463e-01 -9.4053018e-01
6.0649550e-01 -4.1054231e-01 -1.4027423e-01 6.0184020e-01
-1.4754894e-01 -2.6849657e-01 5.9002578e-01 1.9411801e-01
1.1246306e+00 1.7981654e-01 -2.2729969e-02 6.0228914e-02
1.5519021e-01 4.4669276e-01 5.7383436e-01 1.3508892e+00
-6.2899107e-01 -1.6960555e-01 8.2247072e-01 -4.2920664e-02
-7.6325196e-01 -9.9264777e-01 4.9854380e-01 1.5538299e+00
-5.1386464e-02 -1.1996995e-01 -5.5978888e-01 -6.1349481e-01
1.2043235e-01 1.7044709e+00 -2.8647268e-01 -1.7294721e-01
-6.0051483e-01 -5.1987052e-01 1.0615594e+00 4.2738780e-01
9.3805093e-01 -1.2372313e+00 -6.1575234e-01 2.6535159e-01
-5.7334006e-01 -1.6329435e+00 -2.0041017e-01 -3.7008530e-01
-3.8201082e-01 -8.3994770e-01 -2.7762049e-01 -4.1242984e-01
9.7351700e-02 3.0119303e-01 1.2340797e+00 -1.6540998e-01
1.3165361e-01 7.5842023e-01 -1.2654787e-01 -6.2172359e-01
-6.6434658e-01 1.0490346e-02 6.5780833e-02 -4.4638997e-01
4.5614392e-01 -1.9405931e-01 -2.2407748e-01 5.1653814e-01
-3.6533520e-01 1.7138466e-01 6.8873298e-01 3.3465448e-01
-8.6525030e-02 3.1151584e-01 7.0547128e-01 -5.2974296e-01
1.2560942e+00 -6.8694073e-01 -5.0105256e-01 2.4463831e-01
-6.7943084e-01 1.6696189e-01 1.9629794e-01 -2.4118303e-01
-1.5027268e+00 -7.0773512e-01 -2.2330187e-01 2.1109809e-01
-7.9001456e-01 4.7009632e-01 -3.4137614e-02 1.5148818e-01
6.6440505e-01 -2.7997356e-02 4.5302201e-02 1.4509974e-01
5.0820649e-01 4.5658475e-01 2.8001960e-02 -1.1708076e+00
7.0716864e-01 9.3136966e-02 -9.1611594e-02 -1.0597128e+00
-8.2967591e-01 -2.0645392e-01 -1.4333659e-01 -5.9224015e-01
9.9760234e-01 -7.5959754e-01 -1.3623184e+00 2.5581181e-01
-7.8776824e-01 -1.2282336e+00 -2.0188519e-01 6.4102989e-01
-7.6228672e-01 1.7269349e-02 -4.7687450e-01 -1.0514591e+00
3.1511694e-01 -1.3425909e+00 5.9783018e-01 3.8998502e-01
-3.9456987e-01 -1.0659461e+00 8.8584468e-02 6.9812620e-01
7.0134401e-01 -5.3003184e-03 1.2482977e+00 -5.0599951e-01
-4.2862880e-01 1.7259182e-01 -2.2579689e-01 7.7942036e-02
-4.8716132e-02 -5.7236072e-02 -8.7298405e-01 9.0781607e-02
-5.5729818e-01 -5.6080621e-01 3.4232292e-01 5.3919393e-01
6.5119886e-01 -1.7785738e-01 -1.7519416e-01 -1.4409804e-01
9.4357765e-01 4.5624164e-01 7.1840757e-01 4.3761462e-01
2.7347824e-01 7.2604132e-01 8.8903075e-01 3.3432621e-01
9.7587395e-01 4.5931616e-01 1.2518060e-01 3.5051280e-01
-4.9619257e-02 -4.8786119e-01 8.2911605e-01 -1.5050329e-01
-7.4466981e-02 -5.4788029e-01 -1.2545201e+00 3.1927302e-01
-1.9295862e+00 -1.3949860e+00 1.1383261e-01 1.8894745e+00
6.3012516e-01 3.9024243e-01 4.3392920e-01 -4.5623466e-01
5.6313634e-01 1.2388898e-01 -8.2646096e-01 -5.5590624e-01
9.0320498e-02 2.0573196e-01 4.7450241e-01 7.8692001e-01
-6.9849944e-01 1.4565188e+00 7.2348456e+00 8.6987460e-01
-6.4881039e-01 -2.5773823e-02 5.1886266e-01 -1.8783806e-01
-1.2608480e-01 6.3442802e-03 -8.7110198e-01 1.7981856e-01
1.3988546e+00 -1.0082177e-01 8.2703757e-01 5.1523769e-01
9.1149235e-01 -4.3859386e-01 -8.6989629e-01 8.3626622e-01
-3.8076699e-01 -9.8197776e-01 -2.5282817e-02 -6.5551372e-03
5.6954044e-01 3.0465963e-01 1.9026984e-01 1.3665247e+00
1.3604059e+00 -1.3188273e+00 6.8176591e-01 5.0150996e-01
1.6602491e-01 -1.0080448e+00 4.0741056e-01 6.2307948e-01
-7.3799169e-01 -3.9384273e-01 -2.2397327e-01 -3.0798212e-01
3.9584476e-01 -2.5545809e-01 -1.1660798e+00 2.5992271e-01
3.4953469e-01 6.9475532e-01 -2.2567447e-01 9.5513701e-01
-5.3014481e-01 1.0597631e+00 -2.4175011e-01 -1.8713564e-01
8.0099225e-01 -2.0627649e-01 6.2353754e-01 1.2075466e+00
-1.9545114e-01 2.3641710e-01 5.8520442e-01 8.6488414e-01
2.0644268e-01 -2.8762758e-01 -5.6191063e-01 -2.0395154e-01
6.2323302e-01 9.7501642e-01 -3.5005289e-01 -3.6284530e-01
-2.7361879e-01 3.3537298e-01 2.7401379e-01 8.3221781e-01
-1.2684901e+00 1.2336881e-02 1.3111602e+00 -8.9958068e-03
-4.8878111e-02 -6.1552632e-01 -2.5694975e-01 -4.6898243e-01
-5.2812654e-01 -1.2212291e+00 3.1102017e-01 -1.0537719e+00
-1.0505151e+00 -4.2536487e-03 4.7453555e-01 -8.7108397e-01
-2.0368421e-01 -4.6310556e-01 -4.9514437e-01 8.5832554e-01
-1.8211416e+00 -9.6208197e-01 -2.1081272e-01 6.2774533e-01
7.7005255e-01 -1.7639427e-01 4.6158582e-01 2.5449756e-01
-8.0818254e-01 3.4453890e-01 -3.4957927e-01 7.2784737e-02
7.7370262e-01 -1.1572051e+00 3.5349664e-01 5.6643665e-01
-4.8703155e-01 5.5343801e-01 9.0445507e-01 -8.2095295e-01
-1.3869834e+00 -9.8813832e-01 5.0402766e-01 -7.0177442e-01
5.2986157e-01 -3.5905596e-02 -3.9222360e-01 9.0427965e-01
5.5496371e-01 -7.4056238e-01 5.4216492e-01 2.8382650e-01
-8.4786683e-02 1.1142769e-01 -1.3180174e+00 1.2501446e+00
9.6527356e-01 -4.0354133e-01 -5.5718923e-01 3.7466043e-01
6.7073351e-01 -5.4464102e-01 -6.7053866e-01 -1.8526514e-01
4.6609876e-01 -5.4477102e-01 8.8560241e-01 -1.3283912e+00
2.3928350e-01 -1.9874109e-01 2.0469490e-02 -1.5769016e+00
-5.3781205e-01 -9.6534634e-01 1.4237416e-01 1.0804302e+00
5.2937925e-01 -8.0147702e-01 5.3276467e-01 1.3672831e+00
-1.8366033e-01 -2.1046947e-01 -7.0715660e-01 -6.3015187e-01
2.9446292e-01 -1.0570010e+00 4.9813500e-01 5.0787628e-01
-1.3324334e-01 3.9020193e-01 -3.4269130e-01 3.2716095e-01
6.0539311e-01 -6.1062276e-01 8.9374226e-01 -9.9511141e-01
3.0693404e-02 -7.1965700e-01 9.5633522e-02 -1.2754360e+00
6.0842496e-01 -7.0867604e-01 3.1369799e-01 -1.9548919e+00
-2.9549494e-01 -8.1749129e-01 -1.2379879e-02 6.5697378e-01
1.8868152e-02 -6.7718166e-01 2.4076937e-01 -3.8758895e-01
-8.7378234e-01 5.7648891e-01 1.3760233e+00 -1.5871038e-01
-2.1499275e-01 2.4682488e-01 -9.8681521e-01 5.8181220e-01
1.2443800e+00 -4.6323517e-01 -7.0319599e-01 -5.5982322e-01
2.9534987e-01 3.1760347e-01 4.3250600e-01 -1.1317725e+00
3.7874642e-01 -9.2240006e-01 6.3041575e-02 -2.9280153e-01
3.7517112e-01 -5.5195028e-01 -3.3240402e-01 3.1910542e-01
-7.9263633e-01 3.6379105e-01 7.9112750e-01 5.7940871e-01
3.1559795e-01 -7.2229397e-03 6.8324536e-01 -2.6151887e-01
-1.1320453e+00 1.7632805e-01 -1.2627685e+00 7.0011491e-01
1.1023191e+00 -1.3921230e-01 -6.0469377e-01 -8.6319339e-01
-6.8107802e-01 1.1097294e+00 -9.2976280e-02 5.8862889e-01
6.5382951e-01 -9.7890204e-01 -8.3334154e-01 -1.0844054e-01
-1.3991910e-01 -2.7101162e-01 3.7797898e-01 1.0096796e+00
-2.6733041e-01 7.1855026e-01 -4.7304335e-01 -2.5135323e-01
-9.0340459e-01 2.1664935e-01 4.7590873e-01 -3.2671684e-01
-3.2679147e-01 3.9506790e-01 -2.3595756e-01 -7.3264736e-01
3.5502949e-01 -3.8519803e-01 -3.6070147e-01 3.5578001e-02
4.7786489e-01 8.7638897e-01 -3.6106217e-01 -3.3394423e-01
-1.1090835e-01 1.7768830e-01 -3.6208719e-02 -6.9813198e-01
9.9935675e-01 -3.5726371e-01 3.4093234e-01 2.5038955e-01
3.3443713e-01 -3.1838125e-01 -1.4473612e+00 -8.8630244e-02
1.1712542e-01 -2.4347441e-01 7.1185976e-02 -1.2271152e+00
-6.9069290e-01 7.5019127e-01 4.2709956e-01 -1.3650373e-01
4.4541749e-01 -3.7495673e-01 6.9285268e-01 9.1315407e-01
5.7924587e-01 -1.4375223e+00 -1.3113004e-02 1.0267060e+00
6.4610195e-01 -1.3398119e+00 -4.1834360e-01 2.3594286e-01
-1.2928535e+00 7.4103761e-01 1.0227056e+00 1.4377181e-01
4.1287398e-01 -5.6208927e-02 -1.1749214e-01 -1.1987180e-01
-1.1824871e+00 -6.6124594e-01 -2.0480551e-01 1.0501988e+00
7.9290330e-02 2.7950218e-01 3.0255955e-01 5.2315973e-02
-1.6829091e-01 2.2777371e-01 6.0375935e-01 1.0521545e+00
-6.1112428e-01 -9.5659673e-01 -3.2457525e-01 3.6196655e-01
1.1940648e-02 -4.8105642e-02 -7.4180208e-02 1.0475336e+00
-1.7239439e-01 1.6108366e+00 -7.0091501e-02 -2.3648763e-01
5.5047798e-01 1.7148310e-01 3.5090062e-01 -5.4854506e-01
-4.3717816e-01 -4.6289831e-01 7.6633584e-01 -7.2865129e-01
-2.8712362e-01 -9.1166085e-01 -1.4940001e+00 -7.5180000e-01
4.0673068e-01 2.2684424e-01 3.0510345e-01 1.2126825e+00
3.6712319e-01 5.7264698e-01 1.6816109e-01 -4.8337308e-01
-5.8055454e-01 -1.0307274e+00 -1.9638205e-01 4.1446861e-02
3.0430630e-01 -9.8085511e-01 -3.3038508e-02 -4.3553752e-01] | [4.838532447814941, 1.4411981105804443] |
6bc00b3e-d515-40b0-a4a8-b21015dc7e2f | transferable-multi-level-attention-neural | null | null | https://doi.org/10.26434/chemrxiv.12588170.v1 | https://chemrxiv.org/ndownloader/articles/12588170/versions/1/export_pdf | Transferable Multi-level Attention Neural Network for Accurate Prediction of Quantum Chemistry Properties via Multi-task Learning | The development of efficient models for predicting specific properties through machine learning is of great importance for the innovation of chemistry and material science. However, predicting electronic structure properties like frontier molecular orbital HOMO and LUMO energy levels and their HOMO-LUMO gaps from the small-sized molecule data to larger molecules remains a challenge. Here we develop a multi-level attention strategy that enables chemical interpretable insights to be fused into multi-task learning of up to 110,000 records of data in QM9 for random split evaluation. The good transferability for predicting larger molecules outside the training set is demonstrated in both QM9 and Alchemy datasets. The efficient and accurate prediction of 12 properties including dipole moment, HOMO, and Gibbs free energy within chemical accuracy is achieved by using our specifically designed interpretable multi-level attention neural network, named as DeepMoleNet. Remarkably, the present multi-task deep learning model adopts the atom-centered symmetry functions (ACSFs) descriptor as one of the prediction targets, rather than using ACSFs as input in the conventional way. The proposed multi-level attention neural network is applicable to high-throughput screening of numerous chemical species to accelerate rational designs of drug, material, and chemical reactions. | ['Jing Ma', 'Yanwen Guo', 'Yanyan Jiang', 'Zheng Cheng', 'Qingqing Jia', 'Liqiang Lin'] | 2020-06-30 | null | null | null | null | ['formation-energy'] | ['miscellaneous'] | [ 3.87224793e-01 -2.11701348e-01 -3.94520551e-01 -2.45322436e-01
-9.12798822e-01 -4.86113369e-01 3.25574219e-01 6.44628406e-01
-2.92920589e-01 1.28471959e+00 7.60771632e-02 -6.56574845e-01
-1.08282290e-01 -7.76971042e-01 -9.50579286e-01 -1.20375395e+00
-4.99166781e-04 3.43162298e-01 -2.77149498e-01 -1.21051811e-01
5.57209671e-01 8.40197325e-01 -1.09320927e+00 4.68879730e-01
1.44343555e+00 1.19692171e+00 1.26409948e-01 2.66274601e-01
3.74646261e-02 7.63564646e-01 -3.46144795e-01 -3.77623647e-01
-6.82266876e-02 -2.96526432e-01 -9.70960259e-01 -4.76191580e-01
3.89792740e-01 6.04708260e-03 9.96017680e-02 8.45061243e-01
8.45606923e-01 4.28804547e-01 1.00636649e+00 -6.70925081e-01
-1.18673599e+00 2.81714052e-01 -3.48195940e-01 3.17507163e-02
2.97062308e-01 3.85700703e-01 1.23781705e+00 -1.10207081e+00
3.46154034e-01 1.02426028e+00 3.10437739e-01 2.77449399e-01
-1.29034293e+00 -8.73348057e-01 -1.12352021e-01 5.41970909e-01
-1.28702676e+00 -2.37348571e-01 6.99393094e-01 -5.07616460e-01
1.41950047e+00 7.03590736e-02 2.74427503e-01 1.07509518e+00
7.23114073e-01 2.50691980e-01 7.26902306e-01 -1.33762076e-01
2.90337145e-01 -3.49038877e-02 1.16251767e-01 6.08008087e-01
2.95279980e-01 4.41059284e-02 -4.87753421e-01 -1.65757447e-01
3.22840899e-01 2.52749503e-01 -2.49467090e-01 -3.13979164e-02
-1.02851510e+00 9.97292161e-01 9.36483324e-01 1.15073420e-01
-6.99596405e-01 1.98140323e-01 4.08596486e-01 -3.08995564e-02
4.12154019e-01 1.12950993e+00 -8.01683187e-01 2.48368382e-01
-3.05495352e-01 2.10602786e-02 3.61921340e-01 4.31689441e-01
7.18151331e-01 1.84749201e-01 -2.00385809e-01 3.45183730e-01
1.93218976e-01 2.77316302e-01 4.03757215e-01 -5.49510837e-01
3.20296079e-01 6.76537275e-01 2.05577821e-01 -6.53325856e-01
-5.99050224e-01 -4.69324201e-01 -9.43001211e-01 1.16570383e-01
-7.22993612e-02 -1.57486588e-01 -6.90692306e-01 1.46842492e+00
3.46694112e-01 -4.56353456e-01 1.85872525e-01 5.26538789e-01
1.11010849e+00 1.03957140e+00 5.16830444e-01 -4.40160066e-01
1.12399733e+00 -1.04455054e+00 -5.71268380e-01 1.04616351e-01
8.17012906e-01 -4.11462903e-01 8.21406901e-01 3.91413063e-01
-8.65949750e-01 -6.76663876e-01 -1.11743963e+00 -4.86278623e-01
-7.55199969e-01 2.09909573e-01 1.22182667e+00 3.82714421e-01
-3.57569396e-01 1.09622955e+00 -5.65775871e-01 2.98142731e-01
7.94938684e-01 1.00386155e+00 -6.57963872e-01 1.23082034e-01
-1.23396766e+00 9.62696433e-01 8.24184597e-01 1.04026169e-01
-9.50833738e-01 -8.00491571e-01 -6.91062570e-01 4.27012444e-01
2.39253357e-01 -6.43394172e-01 8.79636168e-01 -9.58580673e-01
-1.63931501e+00 3.08526188e-01 -3.20259422e-01 -2.28950247e-01
-2.28871211e-01 -1.22827783e-01 -3.59962463e-01 1.41308317e-02
5.41624576e-02 6.22760355e-01 4.63847369e-01 -7.97773063e-01
-1.13898739e-01 -5.10975480e-01 -1.23063274e-01 1.90641955e-01
-3.37576151e-01 -1.38420358e-01 5.50526202e-01 -2.94084013e-01
-2.73948789e-01 -5.17196417e-01 -4.23397213e-01 -2.84998804e-01
-5.89369714e-01 -5.05009770e-01 2.20660627e-01 -3.95130694e-01
1.01475549e+00 -1.53453219e+00 3.74907941e-01 7.11065680e-02
3.19969982e-01 2.80269504e-01 -2.49504462e-01 7.35603034e-01
-6.34282291e-01 3.12447608e-01 4.73287553e-02 2.91792989e-01
-3.62867832e-01 -5.81775248e-01 -2.42770985e-01 4.49483782e-01
5.40987372e-01 1.04156744e+00 -8.52878928e-01 -4.86944616e-02
3.23131830e-01 4.92667943e-01 -7.86349475e-01 1.17582902e-01
-5.17737925e-01 7.14483261e-01 -8.23420703e-01 8.69894862e-01
5.41740298e-01 -7.03211606e-01 3.13970707e-02 -3.53185713e-01
-3.09372663e-01 4.83593255e-01 -3.70036334e-01 1.44491041e+00
-4.32856917e-01 2.82817453e-01 -7.54482567e-01 -1.01770079e+00
8.82925332e-01 3.37704599e-01 6.95409060e-01 -9.73408043e-01
1.12471364e-01 2.76781529e-01 6.58507407e-01 -3.50081772e-01
1.74160331e-01 -5.41112125e-01 5.59253283e-02 5.87118976e-02
2.11175948e-01 1.84192017e-01 7.39242211e-02 -3.08155805e-01
5.49270093e-01 -1.31836291e-02 7.75592387e-01 -2.46476531e-01
9.04330194e-01 -8.36191103e-02 3.83060575e-01 4.07699525e-01
8.52494091e-02 8.62184539e-02 4.34606552e-01 -7.86899567e-01
-1.05208826e+00 -6.12950325e-01 -6.18779063e-01 1.26879883e+00
-6.34203553e-02 -4.80132431e-01 -3.95818353e-01 -3.66216391e-01
3.61599326e-02 5.54960310e-01 -6.21914983e-01 -6.29156232e-01
-1.57692358e-01 -1.13256824e+00 2.39788778e-02 4.77461010e-01
2.41504028e-01 -1.38084447e+00 1.73939601e-01 5.07713199e-01
3.37164313e-01 -7.70209372e-01 -2.45834753e-01 6.63007975e-01
-6.23639047e-01 -9.96540785e-01 -4.90415424e-01 -5.49695134e-01
3.20049047e-01 -6.30960539e-02 7.01772869e-01 -2.71538705e-01
-1.85370415e-01 -5.53862393e-01 -6.84673041e-02 -7.78360307e-01
-3.23934317e-01 3.31465006e-01 4.96246479e-02 5.79192936e-02
2.98649192e-01 -5.02220690e-01 -8.92809272e-01 -7.44410306e-02
-4.78897154e-01 5.27504645e-02 7.55669296e-01 9.83587027e-01
8.84717047e-01 -2.97408491e-01 1.00321496e+00 -8.16294491e-01
5.95539868e-01 -5.96424341e-01 -5.03865719e-01 2.35116914e-01
-6.37159467e-01 4.52268749e-01 1.49413896e+00 -1.89038560e-01
-7.46377528e-01 5.76565005e-02 -5.27387977e-01 2.32963599e-02
-6.06789161e-03 7.10879207e-01 -3.80889863e-01 -4.86763775e-01
6.51619077e-01 3.07153046e-01 -4.50115710e-01 -4.98826057e-01
1.87893331e-01 6.39619291e-01 1.78542629e-01 -6.33669376e-01
2.92065442e-01 -4.85935062e-02 5.87046266e-01 -9.08100724e-01
-9.78547692e-01 -2.29879007e-01 -6.52214944e-01 4.25709337e-01
1.03935528e+00 -9.54042017e-01 -1.55514932e+00 1.18190214e-01
-1.34944665e+00 -2.77217794e-02 1.22406699e-01 5.18242300e-01
-3.20505679e-01 3.20401490e-01 -3.78528863e-01 -6.14495099e-01
-8.88275027e-01 -1.48132682e+00 1.16612422e+00 1.76600680e-01
1.11883692e-02 -1.03679800e+00 -7.18148574e-02 6.06298923e-01
1.18575364e-01 5.99418044e-01 1.51444268e+00 -8.36105168e-01
-8.09508502e-01 -2.71093249e-01 -2.81161398e-01 1.30720288e-01
3.38758349e-01 4.71362658e-02 -1.00104558e+00 -2.20516711e-01
-4.77025986e-01 -7.00104773e-01 1.05106056e+00 6.52678192e-01
1.89081836e+00 -3.27073216e-01 -2.71830142e-01 8.50014150e-01
1.28418207e+00 7.12305427e-01 5.21576166e-01 1.68504357e-01
1.10299897e+00 1.03376612e-01 3.55924815e-01 4.93378222e-01
6.25444800e-02 4.95159358e-01 5.60536683e-01 -2.72855103e-01
5.52558124e-01 -3.18352640e-01 3.19375783e-01 4.31207955e-01
-5.40690660e-01 -5.54870427e-01 -7.62889445e-01 -6.48552850e-02
-1.34649730e+00 -1.09565306e+00 -1.25749573e-01 2.19570613e+00
9.10663426e-01 -5.29639572e-02 -5.27683273e-02 -1.80936605e-01
4.48527128e-01 8.43144134e-02 -1.18714237e+00 -7.78520286e-01
-9.11594033e-02 5.81951261e-01 3.76320422e-01 4.66903061e-01
-1.07790136e+00 9.38150346e-01 5.95536470e+00 1.09841704e+00
-1.35089552e+00 -2.98056692e-01 1.16675448e+00 7.81147256e-02
-4.51404065e-01 -1.20394873e-02 -8.33961070e-01 4.99138147e-01
1.14856935e+00 -2.61547476e-01 3.52673620e-01 7.35038340e-01
3.83212000e-01 2.42898077e-01 -1.18157136e+00 8.61709952e-01
-4.05929446e-01 -1.97962368e+00 6.15138531e-01 1.78863242e-01
9.41149652e-01 -1.00886673e-01 2.76807189e-01 1.61768943e-01
-4.10003960e-01 -1.64378130e+00 3.09131861e-01 3.13502550e-01
1.20121896e+00 -9.65187132e-01 4.90094483e-01 1.05741598e-01
-1.01090002e+00 -1.48595601e-01 -4.71674502e-01 -2.35908814e-02
-3.56823117e-01 1.06357165e-01 -7.21977532e-01 6.44748271e-01
3.82388607e-02 9.03575540e-01 -3.78271848e-01 7.97716558e-01
1.22802563e-01 1.60878465e-01 1.17892392e-01 -4.03475821e-01
3.46358567e-01 -5.67383945e-01 -1.94371417e-02 6.65822208e-01
1.07070625e-01 8.51857364e-02 6.35612980e-02 1.06813776e+00
-5.97021341e-01 4.48212445e-01 -4.07781869e-01 -6.66886568e-01
3.04015666e-01 1.06548131e+00 -3.80179733e-01 -1.85153037e-01
-3.27867985e-01 8.35459173e-01 3.44167024e-01 4.67012525e-01
-8.94208074e-01 -5.10644197e-01 7.21065104e-01 -1.67998016e-01
1.77990288e-01 1.31969586e-01 -2.08267063e-01 -9.94732738e-01
-3.46580535e-01 -8.19086373e-01 1.22441091e-01 -6.30373597e-01
-1.06081200e+00 5.54913998e-01 -4.43191707e-01 -6.45662785e-01
2.03948900e-01 -1.14625609e+00 -9.28164840e-01 1.13544345e+00
-1.66710091e+00 -8.75867426e-01 3.41958478e-02 2.54449308e-01
4.34405416e-01 -2.81317770e-01 1.06138968e+00 2.53452778e-01
-1.02710080e+00 5.00279725e-01 8.58951449e-01 -3.65069270e-01
6.92883611e-01 -1.37507641e+00 1.72740012e-01 1.71021551e-01
-1.48751020e-01 7.34459519e-01 5.05808055e-01 -4.79041338e-01
-1.50176668e+00 -1.18940008e+00 6.81434989e-01 -3.78865510e-01
5.25047183e-01 -3.00648719e-01 -8.17422032e-01 3.58074337e-01
-1.49135515e-02 5.43457568e-02 1.17497802e+00 9.09894705e-02
-2.41064224e-02 -9.47541371e-02 -8.18685710e-01 4.48716730e-01
7.07168758e-01 -6.30985856e-01 1.01690330e-01 9.49632943e-01
1.00128603e+00 -2.27607340e-01 -1.33311749e+00 4.42632139e-01
5.00202060e-01 -6.66325748e-01 1.21275389e+00 -1.29370344e+00
8.37363780e-01 -5.63829951e-02 6.03867769e-02 -1.11844552e+00
-5.22070587e-01 -6.21981144e-01 -1.77033469e-01 4.90690649e-01
7.74632275e-01 -6.53250337e-01 6.08966827e-01 4.18439209e-01
-3.73721570e-01 -1.29354632e+00 -8.97158384e-01 -3.14740241e-01
4.50091064e-01 7.42081925e-02 6.15511417e-01 6.87132537e-01
-5.27686998e-02 9.31765914e-01 -3.81136835e-01 -3.73601913e-03
1.40895769e-01 6.30201221e-01 3.47279221e-01 -1.26801610e+00
-2.68583387e-01 -3.38908434e-01 -1.94637716e-01 -9.62384880e-01
3.85556459e-01 -1.11187100e+00 -6.78760469e-01 -1.37975013e+00
3.41493756e-01 5.09158187e-02 -6.18556201e-01 3.72993052e-01
-1.46309897e-01 -2.20597491e-01 -1.64175257e-01 -6.83204159e-02
-6.70875907e-01 1.28853655e+00 1.74419570e+00 -5.16801000e-01
-2.24123374e-01 1.12943009e-01 -1.02601755e+00 3.53771031e-01
5.63383520e-01 -3.74552637e-01 -1.12488486e-01 8.04111212e-02
4.04410660e-01 1.60337478e-01 1.32566988e-01 -8.68693471e-01
-1.85469523e-01 -4.96503741e-01 8.56122017e-01 -3.80749464e-01
3.46965820e-01 -5.28804004e-01 -2.27899440e-02 5.99723637e-01
-4.49338943e-01 -2.35636204e-01 2.99986184e-01 6.14692390e-01
-1.84458956e-01 2.54899468e-02 7.52609193e-01 -6.54480532e-02
-5.02562523e-01 1.01834285e+00 -2.02786833e-01 -3.38803649e-01
9.81098413e-01 -2.84059763e-01 -2.90177256e-01 3.37860063e-02
-5.59408605e-01 1.72902972e-01 2.06452787e-01 1.55997172e-01
5.75887442e-01 -1.21005630e+00 -3.29937398e-01 8.53243917e-02
3.27041417e-01 1.26284778e-01 4.19209242e-01 6.40354812e-01
-6.20553792e-01 1.01526916e+00 -3.09861541e-01 -1.52406052e-01
-9.22487676e-01 7.32303917e-01 7.01918125e-01 -1.51090905e-01
9.28745195e-02 6.31988525e-01 6.12200081e-01 -1.49875343e-01
-1.05329588e-01 -3.87093663e-01 -2.56464452e-01 5.42430282e-02
4.00145799e-01 1.69580072e-01 1.50576860e-01 -5.03608882e-01
-4.54522580e-01 4.76290017e-01 -3.37418526e-01 1.01041126e+00
1.60476255e+00 4.18477803e-01 -9.02102366e-02 1.62120685e-01
1.46874237e+00 -2.69388765e-01 -1.22749770e+00 2.15381771e-01
-2.03384370e-01 2.91603096e-02 2.16057286e-01 -9.17985797e-01
-4.61073250e-01 1.18228614e+00 4.05889153e-01 -2.65925318e-01
7.87337601e-01 -2.03279555e-01 7.93833196e-01 1.01892281e+00
2.75268033e-02 -7.79881120e-01 2.53797591e-01 4.21678990e-01
9.86997128e-01 -1.70614707e+00 2.67140865e-01 -4.12199572e-02
-5.48414826e-01 1.49124670e+00 6.78025126e-01 2.42150292e-01
1.56294942e-01 -5.84140599e-01 -7.15976238e-01 -4.53066021e-01
-7.64385998e-01 1.55604845e-02 5.48455596e-01 4.03424352e-01
9.27848101e-01 -7.18271360e-02 -1.00144118e-01 6.70750797e-01
1.84541658e-01 -2.63195366e-01 2.56158620e-01 3.38575721e-01
-5.60471714e-01 -1.22009301e+00 1.07034490e-01 4.07507271e-01
-6.24024332e-01 -5.97367883e-01 -6.95406735e-01 4.85804856e-01
2.45362297e-01 5.46120763e-01 -2.61878818e-01 -1.06571749e-01
-2.58323736e-02 -1.76665951e-02 5.90441704e-01 -6.40130818e-01
-5.66632450e-01 -2.88155228e-01 -2.30105557e-02 -2.84660995e-01
-3.51755977e-01 -1.53510422e-01 -1.35881495e+00 -3.25507522e-01
-6.76187813e-01 3.92445952e-01 4.48417366e-01 1.05886304e+00
7.41738558e-01 5.53702354e-01 1.04833734e+00 -9.18013692e-01
-4.68309283e-01 -8.19823623e-01 -3.53172749e-01 3.33615869e-01
6.62824214e-01 -5.57536960e-01 6.09526671e-02 -2.94992924e-01] | [5.097573757171631, 5.805250644683838] |
551bdfe1-500b-4ba7-9539-f68216240279 | modality-alignment-between-deep | null | null | https://aclanthology.org/2022.lrec-1.295 | https://aclanthology.org/2022.lrec-1.295.pdf | Modality Alignment between Deep Representations for Effective Video-and-Language Learning | Video-and-Language learning, such as video question answering or video captioning, is the next challenge in the deep learning society, as it pursues the way how human intelligence perceives everyday life. These tasks require the ability of multi-modal reasoning which is to handle both visual information and text information simultaneously across time. In this point of view, a cross-modality attention module that fuses video representation and text representation takes a critical role in most recent approaches. However, existing Video-and-Language models merely compute the attention weights without considering the different characteristics of video modality and text modality. Such na ̈ıve attention module hinders the current models to fully enjoy the strength of cross-modality. In this paper, we propose a novel Modality Alignment method that benefits the cross-modality attention module by guiding it to easily amalgamate multiple modalities. Specifically, we exploit Centered Kernel Alignment (CKA) which was originally proposed to measure the similarity between two deep representations. Our method directly optimizes CKA to make an alignment between video and text embedding representations, hence it aids the cross-modality attention module to combine information over different modalities. Experiments on real-world Video QA tasks demonstrate that our method outperforms conventional multi-modal methods significantly with +3.57% accuracy increment compared to the baseline in a popular benchmark dataset. Additionally, in a synthetic data environment, we show that learning the alignment with our method boosts the performance of the cross-modality attention. | ['Kyomin Jung', 'Yongil Kim', 'Hyeongu Yun'] | null | null | null | null | lrec-2022-6 | ['video-question-answering'] | ['computer-vision'] | [ 1.22395270e-01 -3.64571631e-01 -1.82276204e-01 -1.82274640e-01
-8.35572600e-01 -5.39543867e-01 9.58103597e-01 -1.01782165e-01
-5.81543982e-01 2.46150002e-01 5.38298368e-01 -2.99483895e-01
1.75763518e-01 -4.67505962e-01 -9.63484347e-01 -5.80598950e-01
4.35140371e-01 3.08161112e-03 9.72206220e-02 -1.71841830e-01
5.34275472e-02 -1.36941344e-01 -1.41318333e+00 6.39791489e-01
7.21253455e-01 1.08288085e+00 3.56400371e-01 6.37630463e-01
-3.78745556e-01 1.01081431e+00 -3.22636127e-01 -6.40883029e-01
4.93445471e-02 -3.10223103e-01 -6.95647597e-01 1.47199899e-01
7.95043111e-01 -4.55276012e-01 -7.55023360e-01 1.01989114e+00
3.97370785e-01 2.53994673e-01 5.16973794e-01 -1.36451197e+00
-1.08795297e+00 3.89178187e-01 -7.78549016e-01 2.90263176e-01
4.78938907e-01 3.19115043e-01 1.10515356e+00 -8.68027806e-01
4.23676431e-01 1.37347639e+00 3.00857067e-01 4.82899725e-01
-9.35108125e-01 -4.11520898e-01 4.57348645e-01 7.07448840e-01
-1.07969117e+00 -3.37692618e-01 8.83450568e-01 -5.12673736e-01
8.39667022e-01 1.85649917e-01 4.44207966e-01 1.24871576e+00
1.20728694e-01 1.10624123e+00 8.08690488e-01 -2.11091831e-01
-1.77776530e-01 2.60329619e-02 -5.93958481e-04 6.83553398e-01
-1.08838156e-01 -2.80065060e-01 -6.86776698e-01 2.38505557e-01
6.94475472e-01 4.26324874e-01 -4.46421772e-01 -4.85344112e-01
-1.73878801e+00 7.08598137e-01 4.73430395e-01 3.96875441e-01
-4.10825998e-01 1.88789546e-01 5.77329338e-01 2.10808530e-01
1.19638205e-01 1.85525283e-01 -2.00519502e-01 -1.55232385e-01
-5.90168118e-01 -5.94907627e-02 2.35028103e-01 7.19210565e-01
5.38905144e-01 5.43994792e-02 -4.40335155e-01 7.44194865e-01
4.99181986e-01 6.32545948e-01 6.19200826e-01 -8.21545720e-01
8.59365165e-01 7.87787735e-01 -7.87023827e-02 -1.07526493e+00
-1.53696209e-01 -6.18257299e-02 -9.57477510e-01 -8.76005664e-02
4.39247161e-01 4.05708142e-02 -9.56163824e-01 1.99066985e+00
1.87467232e-01 3.75556827e-01 2.94622004e-01 1.04431903e+00
9.53581214e-01 7.74007142e-01 1.84049636e-01 -5.12969829e-02
1.61978734e+00 -1.35096359e+00 -8.98400426e-01 -3.18650484e-01
2.88453817e-01 -7.04523027e-01 1.32792711e+00 2.55831536e-02
-1.12721646e+00 -6.72158480e-01 -1.05137074e+00 -4.26701695e-01
-3.63865018e-01 9.32840258e-02 5.43878853e-01 1.77487403e-01
-8.93470764e-01 -3.31061892e-02 -6.84051096e-01 -3.87724191e-01
3.17052037e-01 1.01877563e-01 -6.41186416e-01 -3.03083330e-01
-1.36191726e+00 7.50171006e-01 3.40943694e-01 1.32377207e-01
-6.88489795e-01 -6.03947997e-01 -1.03272283e+00 3.55861336e-02
5.67710042e-01 -9.78908360e-01 1.12387276e+00 -1.33468401e+00
-1.34476399e+00 6.98798895e-01 -3.20136130e-01 -2.28426099e-01
4.27622616e-01 -4.13038611e-01 -4.65012997e-01 5.36371171e-01
-6.91355988e-02 7.13506758e-01 1.24608648e+00 -1.06647980e+00
-6.28353775e-01 -3.14368814e-01 5.46060145e-01 5.04655480e-01
-6.12709939e-01 -1.89788759e-01 -9.17987466e-01 -7.03431964e-01
-7.08194599e-02 -6.61295056e-01 2.53101379e-01 1.52608261e-01
-9.78585184e-02 -2.58116901e-01 9.54874873e-01 -7.34501481e-01
1.04967213e+00 -2.31190038e+00 6.59971595e-01 -3.41860712e-01
3.44520479e-01 2.26648450e-01 -2.98372328e-01 3.80515158e-01
-8.47260654e-02 -1.11043528e-01 -2.86749769e-02 -4.82682198e-01
1.88106149e-01 1.60388991e-01 -2.59581685e-01 4.22026634e-01
3.44868809e-01 1.15775204e+00 -8.65309596e-01 -3.94934118e-01
2.16981545e-01 7.63766229e-01 -6.14324808e-01 3.06062490e-01
-2.22639546e-01 3.23784113e-01 -2.30842873e-01 7.37581789e-01
4.97277886e-01 -5.40373325e-01 1.66808978e-01 -7.32075155e-01
2.58463502e-01 -2.37159878e-01 -9.06421781e-01 2.22196007e+00
-4.32611495e-01 6.44975245e-01 5.31117059e-02 -1.23176920e+00
4.56422687e-01 5.04304111e-01 3.09195936e-01 -9.12575066e-01
1.20218955e-01 -8.82315487e-02 -4.23918515e-02 -8.16574156e-01
4.75625813e-01 -1.85285956e-01 -1.51486322e-01 3.82729799e-01
2.52897590e-01 3.56018335e-01 1.16691999e-01 4.80400890e-01
8.14697385e-01 2.11971685e-01 1.40288100e-01 1.92681104e-01
7.97101796e-01 -3.31753880e-01 2.64193594e-01 4.84753758e-01
-4.06590223e-01 6.02662802e-01 3.83834422e-01 -3.52298468e-01
-9.39512849e-01 -1.03477132e+00 2.80612886e-01 1.42802691e+00
3.67730796e-01 -3.43697816e-01 -4.55382377e-01 -8.15109074e-01
-4.37523313e-02 3.14958930e-01 -6.81682527e-01 -3.07708204e-01
-4.60828781e-01 -4.35192943e-01 3.18324268e-01 6.25475407e-01
5.70113361e-01 -8.92155409e-01 -4.66010928e-01 -1.33451581e-01
-6.31153226e-01 -1.48910189e+00 -8.08206201e-01 -3.46596032e-01
-4.78001028e-01 -1.08075976e+00 -8.97284806e-01 -8.62251163e-01
4.84269172e-01 6.35797262e-01 9.24924314e-01 9.40124691e-02
8.85833427e-02 9.46114779e-01 -4.94425207e-01 7.48575404e-02
-5.86489961e-03 -1.09405003e-01 2.14979388e-02 5.41441739e-01
4.84385490e-01 -5.38518846e-01 -6.01221800e-01 5.33593707e-02
-1.26393032e+00 4.48059946e-01 7.45841503e-01 1.00679958e+00
2.67907768e-01 -4.22264516e-01 4.96779621e-01 -3.91791999e-01
4.13342208e-01 -6.96251869e-01 -1.62321597e-01 6.17430091e-01
1.20371161e-02 2.01441705e-01 5.28319001e-01 -6.93062425e-01
-1.00417197e+00 -2.00715467e-01 -1.86477229e-02 -8.43257368e-01
4.47640456e-02 7.05240667e-01 -4.97557163e-01 1.76016510e-01
3.18308212e-02 3.73798490e-01 2.13150214e-02 -1.82823196e-01
6.48412645e-01 5.79015434e-01 7.57483482e-01 -5.52819371e-01
8.03494036e-01 5.49425006e-01 -2.61292905e-01 -6.38375580e-01
-6.94934189e-01 -4.39103514e-01 -5.38349390e-01 -2.94620335e-01
1.24193728e+00 -1.13272882e+00 -9.41296279e-01 4.93281573e-01
-1.23960984e+00 4.13658144e-03 2.05895931e-01 5.80273092e-01
-4.76048827e-01 6.10644698e-01 -4.49620813e-01 -5.23010790e-01
-1.95263609e-01 -1.28371036e+00 1.17674625e+00 2.76431471e-01
3.45241457e-01 -1.10295439e+00 -1.85452178e-01 7.64354050e-01
3.62792641e-01 9.34113115e-02 8.61936390e-01 -4.92140174e-01
-6.79413855e-01 8.00843164e-02 -5.80071330e-01 1.92125693e-01
1.09850183e-01 -2.42424548e-01 -9.81735051e-01 -4.01631922e-01
-2.91452892e-02 -5.24964452e-01 1.06933963e+00 5.52872233e-02
1.14564073e+00 -1.66283861e-01 7.88702667e-02 6.05596244e-01
1.39923441e+00 1.39883071e-01 5.77992857e-01 5.08099616e-01
1.19820166e+00 4.55697626e-01 3.64963084e-01 1.82882577e-01
8.66965353e-01 7.38536000e-01 5.61307549e-01 -1.20938411e-02
-1.12734489e-01 -2.30288833e-01 6.16083205e-01 9.79244947e-01
2.36523468e-02 -1.79876536e-01 -8.58026683e-01 4.46462214e-01
-2.14195800e+00 -1.17176020e+00 3.13744396e-01 1.97820842e+00
6.68269277e-01 -9.03741550e-03 5.84627278e-02 -6.17333055e-02
6.82483912e-01 3.61641586e-01 -6.85578346e-01 -5.39578013e-02
-1.19478732e-01 -3.86989474e-01 5.89894503e-02 3.37386549e-01
-1.28105569e+00 6.32286012e-01 5.36186314e+00 7.42829144e-01
-1.28061068e+00 1.73546076e-01 3.80678236e-01 -1.44466490e-01
-4.77425277e-01 -3.83285522e-01 -4.31096971e-01 5.14130116e-01
8.22795630e-01 -2.06020266e-01 4.74890828e-01 4.43330348e-01
-1.86395515e-02 1.63638126e-02 -1.16145217e+00 1.29752111e+00
5.93164504e-01 -1.30075181e+00 4.28741664e-01 -2.12666854e-01
5.01981616e-01 -1.71351016e-01 2.67910331e-01 5.95632911e-01
-2.43056431e-01 -9.83311296e-01 7.74284899e-01 7.02058494e-01
6.76302910e-01 -5.59564173e-01 6.16835058e-01 1.69425264e-01
-1.40957093e+00 -2.04622611e-01 -4.09789868e-02 3.50905061e-02
4.07786727e-01 7.93114901e-02 -2.85053819e-01 8.67444277e-01
6.42582953e-01 9.56413090e-01 -7.13960111e-01 7.27932990e-01
1.41070023e-01 2.02224448e-01 6.25561923e-02 2.64088064e-01
3.97687256e-01 -1.57339126e-01 4.96901244e-01 1.03525960e+00
2.39715323e-01 -1.19757630e-01 1.89530268e-01 6.15999460e-01
-2.20532253e-01 -2.54633147e-02 -5.09444356e-01 -3.90648991e-01
1.77143961e-01 1.15304053e+00 -2.78103441e-01 -4.64034617e-01
-1.00302386e+00 1.18947864e+00 3.53172392e-01 6.18657708e-01
-1.19402480e+00 -1.21528879e-01 7.30224550e-01 -2.74876654e-01
4.60596561e-01 -3.09133828e-01 5.90654798e-02 -1.70731521e+00
1.68393999e-01 -1.05139267e+00 4.71210510e-01 -9.69324589e-01
-1.43523979e+00 6.11014128e-01 -7.34550506e-02 -1.31768608e+00
-1.75210819e-01 -7.67121553e-01 -4.17315900e-01 7.60278165e-01
-1.66888833e+00 -1.66985774e+00 -3.72004777e-01 9.68975544e-01
7.52602100e-01 -1.42834619e-01 4.39769596e-01 5.77224493e-01
-6.66278481e-01 7.32009053e-01 8.45744312e-02 2.51751125e-01
8.88140261e-01 -1.16312623e+00 1.40007555e-01 9.99634206e-01
2.67614901e-01 6.34352863e-01 4.09757763e-01 -2.29950875e-01
-1.95678747e+00 -9.04681385e-01 4.93456602e-01 -4.29252148e-01
9.75966334e-01 -1.96942419e-01 -1.11210227e+00 7.29902983e-01
5.95570803e-01 3.17490511e-02 7.10268557e-01 -1.46769747e-01
-7.47879803e-01 -1.37895852e-01 -6.66879594e-01 6.15446627e-01
7.42227495e-01 -1.02016425e+00 -8.39930296e-01 9.27965119e-02
1.02704227e+00 -3.62989187e-01 -7.84474552e-01 3.44115376e-01
6.24258518e-01 -8.36359024e-01 1.15415657e+00 -7.31615186e-01
7.37622440e-01 -4.03035045e-01 -5.02325416e-01 -1.02021253e+00
-1.42628208e-01 -3.19249362e-01 -4.86663610e-01 1.18974841e+00
1.31818607e-01 -3.79687637e-01 2.91478723e-01 3.61955941e-01
-6.70694709e-02 -6.32593453e-01 -8.95062566e-01 -4.60775882e-01
9.37336087e-02 -3.57993633e-01 3.96453232e-01 1.13883603e+00
8.51577669e-02 5.83372593e-01 -6.43300891e-01 1.95337802e-01
4.39891309e-01 2.36098930e-01 6.11926138e-01 -8.25186074e-01
-4.44630325e-01 -5.42437077e-01 -4.23837394e-01 -1.29687691e+00
2.84619719e-01 -8.24779689e-01 -1.66144043e-01 -1.59343886e+00
4.79427040e-01 2.64775813e-01 -5.31467199e-01 4.23682451e-01
-6.16810799e-01 3.75656486e-01 5.95988393e-01 1.37913376e-01
-9.54018772e-01 7.17520118e-01 1.33086050e+00 -4.23322529e-01
1.73716113e-01 -5.46241820e-01 -7.79917479e-01 6.06129885e-01
4.86185819e-01 1.97847098e-01 -5.47239959e-01 -9.70836997e-01
1.43776655e-01 1.39248595e-01 5.12167692e-01 -8.46525550e-01
3.60947579e-01 -2.43061021e-01 2.90506303e-01 -4.32454437e-01
6.10547423e-01 -9.12735760e-01 -1.60747379e-01 1.80333152e-01
-3.02610964e-01 3.75887394e-01 2.09222883e-01 7.93462634e-01
-4.98350084e-01 2.30905041e-01 4.90697950e-01 -5.04972320e-03
-1.17216289e+00 3.89080107e-01 -1.33790791e-01 1.29105389e-01
1.07247543e+00 -1.68239735e-02 -6.29727006e-01 -5.17716408e-01
-5.80750883e-01 4.53414500e-01 3.14206839e-01 9.18466508e-01
7.11212456e-01 -1.70511019e+00 -5.24333894e-01 2.12188005e-01
3.72531265e-01 -2.99334556e-01 6.34079337e-01 1.13120973e+00
-2.66476780e-01 4.66477364e-01 -2.92160004e-01 -7.47090578e-01
-1.18216944e+00 8.26323211e-01 3.00736189e-01 -5.81538528e-02
-4.68608111e-01 6.62214339e-01 5.11644900e-01 -1.72233969e-01
4.06503826e-01 -2.17934549e-01 -3.91828328e-01 2.67376721e-01
8.65690112e-01 -3.39113269e-03 -2.52514422e-01 -9.85957503e-01
-3.22295696e-01 6.75626814e-01 -1.34752572e-01 -1.42025903e-01
1.07973266e+00 -4.69734728e-01 -2.56204454e-04 6.95615113e-01
1.36458993e+00 -3.27244997e-01 -1.32411098e+00 -4.20721203e-01
-2.75919020e-01 -5.25419295e-01 4.99845929e-02 -6.24943316e-01
-1.12814927e+00 1.17926180e+00 4.98911530e-01 8.79016966e-02
1.24022365e+00 3.78658548e-02 9.34064448e-01 3.36492777e-01
-3.07945684e-02 -7.36779332e-01 5.57002544e-01 4.34591055e-01
8.21851790e-01 -1.63358033e+00 -2.66746759e-01 8.03082250e-03
-9.61536527e-01 1.10038388e+00 8.61564398e-01 2.50023603e-01
3.86767983e-01 -2.99521267e-01 1.75515682e-01 -1.11227259e-01
-8.84431601e-01 -2.72267610e-01 7.34807372e-01 4.15146798e-01
3.36525381e-01 -1.55907750e-01 -8.26788992e-02 5.44835567e-01
3.90033156e-01 -1.29581481e-01 3.08446079e-01 8.72412562e-01
-3.18892956e-01 -9.70242202e-01 -3.78970981e-01 1.17297731e-01
-4.39712793e-01 -1.24802612e-01 -1.25150941e-02 6.28098011e-01
3.77272889e-02 8.90759885e-01 1.60487607e-01 -4.63706881e-01
2.21945763e-01 1.29294813e-01 4.48972285e-01 -1.91042230e-01
-3.13043058e-01 7.17937946e-02 -2.95454919e-01 -6.08474553e-01
-9.15412664e-01 -4.57582325e-01 -1.04933941e+00 -2.81046331e-01
5.10546230e-02 -7.24724159e-02 5.30320406e-01 1.21688771e+00
3.55012238e-01 6.86184645e-01 3.90536219e-01 -8.60510528e-01
-3.07467997e-01 -7.89019048e-01 -2.28417397e-01 7.23140299e-01
6.05083346e-01 -8.15682590e-01 -3.81676286e-01 2.25854889e-01] | [10.44334888458252, 1.1424341201782227] |
9a8dd3e5-5359-4769-b922-b42e8da81f50 | piclick-picking-the-desired-mask-in-click | 2304.11609 | null | https://arxiv.org/abs/2304.11609v1 | https://arxiv.org/pdf/2304.11609v1.pdf | PiClick: Picking the desired mask in click-based interactive segmentation | Click-based interactive segmentation enables productive pixel-level annotation and image editing with simple user clicks, whereas target ambiguity remains a problem hindering precise segmentation. That is, in scenes with rich context, one click may refer to multiple potential targets residing in corresponding masks, while most interactive segmentors can only generate one single mask and fail to capture the rich context. To resolve target ambiguity, we here propose PiClick to produce semantically diversified masks. PiClick leverages a transformer network design wherein mutually interactive mask queries are integrated to infuse target priors. Moreover, a Target Reasoning Module is designed in PiClick to automatically imply the best-matched mask from all proposals, significantly relieving target ambiguity as well as extra human intervention. Extensive experiments conducted on all 9 interactive segmentation datasets not only demonstrate the state-of-the-art segmentation performance of PiClick, but also reduces human interventions with multiple proposal generation and target reasoning. To promote direct usage and future endeavors, we release the source code of PiClick together with a plug-and-play annotation tool at https://github.com/cilinyan/PiClick. | ['Efstratios Gavves', 'Guoliang Kang', 'Xu Tang', 'Yao Hu', 'XiaoLong Jiang', 'Jie Liu', 'Haochen Wang', 'Cilin Yan'] | 2023-04-23 | null | null | null | null | ['interactive-segmentation'] | ['computer-vision'] | [ 5.68330944e-01 2.97700882e-01 -1.99789181e-01 -3.81689727e-01
-9.34393585e-01 -9.74934995e-01 1.99971229e-01 -1.01820856e-01
-3.45599592e-01 4.85947877e-01 -3.34101282e-02 -2.89697051e-01
2.55573809e-01 -4.78466094e-01 -5.64773679e-01 -2.94121057e-01
3.94554108e-01 2.73341417e-01 8.77757549e-01 1.39510617e-01
1.87989667e-01 2.84255534e-01 -1.46001387e+00 3.98711652e-01
1.16868663e+00 8.09184790e-01 7.82247543e-01 5.48085690e-01
-2.08002090e-01 1.86472163e-01 -6.51709199e-01 -4.88098115e-01
3.84562463e-01 -2.05943376e-01 -5.66857219e-01 4.81741503e-03
7.41752386e-01 -3.74722302e-01 1.17077613e-02 8.82102311e-01
5.59881985e-01 7.10793212e-02 3.42066199e-01 -1.17658567e+00
-3.54625046e-01 6.00744665e-01 -6.87945426e-01 2.38952618e-02
4.30766404e-01 6.25080347e-01 1.19053686e+00 -9.91321921e-01
8.08200181e-01 8.28394949e-01 5.63179791e-01 4.43732888e-01
-1.43417740e+00 -8.48888934e-01 4.29937303e-01 -1.08565032e-01
-1.36832011e+00 -4.30711299e-01 5.46997905e-01 -3.51204276e-01
6.99581802e-01 7.88211167e-01 7.83355832e-01 9.25903201e-01
-2.06850618e-01 1.11453140e+00 7.66820133e-01 -1.85107693e-01
4.64456901e-03 -6.20223768e-02 -8.27215463e-02 7.66890585e-01
7.95386732e-02 -3.27836007e-01 -6.37251735e-01 2.12572202e-01
9.44830596e-01 -1.23996027e-02 -3.24067861e-01 -3.05145323e-01
-1.29843879e+00 3.80605251e-01 3.64460886e-01 -3.14180106e-02
-3.68435502e-01 1.95559964e-01 1.07936613e-01 -2.51964539e-01
2.38129392e-01 5.76560378e-01 -3.30198169e-01 -1.51678443e-01
-1.19915617e+00 2.67143190e-01 6.58268392e-01 1.28954005e+00
8.54398489e-01 -3.78463447e-01 -6.70944571e-01 7.87717640e-01
8.66096318e-02 3.95826340e-01 -6.83129430e-02 -1.31469381e+00
4.52602386e-01 6.69345915e-01 2.41320744e-01 -7.12720096e-01
-2.52842814e-01 -5.30862987e-01 -2.32567027e-01 2.40260392e-01
4.10265177e-01 -9.12261680e-02 -1.07997882e+00 1.47227657e+00
5.74787796e-01 1.31417409e-01 -6.97538853e-01 9.04820919e-01
8.10180187e-01 3.57753783e-01 6.67477399e-02 6.51834533e-02
1.47809315e+00 -1.20915389e+00 -5.76215267e-01 -4.67145264e-01
3.32352102e-01 -1.04677916e+00 1.60215282e+00 4.31754738e-01
-1.07327831e+00 -2.28538513e-01 -8.20636988e-01 -3.35809708e-01
-2.03256726e-01 2.24553332e-01 6.21255934e-01 3.62837553e-01
-7.03049779e-01 4.22882408e-01 -8.93752933e-01 -1.25263318e-01
8.35076034e-01 1.33929670e-01 -2.42044441e-02 -1.02967046e-01
-7.08774567e-01 4.17325646e-01 3.38082135e-01 9.43368897e-02
-5.16807079e-01 -1.01640010e+00 -5.46190679e-01 -4.72518653e-02
1.09275103e+00 -6.00172520e-01 1.51521122e+00 -7.89861798e-01
-1.25479519e+00 7.24699259e-01 -2.99521416e-01 -1.03414923e-01
8.48987460e-01 -5.88297069e-01 -7.06916675e-02 3.11675161e-01
6.00003839e-01 1.24271047e+00 5.90867996e-01 -1.36556733e+00
-8.43805671e-01 -5.84637281e-04 1.09353356e-01 2.64213234e-01
-2.64971647e-02 -4.11999300e-02 -1.45308602e+00 -6.62940860e-01
2.47223899e-01 -8.03171933e-01 -3.21163565e-01 4.66777980e-01
-8.94463897e-01 -4.66700159e-02 8.27512741e-01 -6.81041956e-01
1.54940808e+00 -2.12629247e+00 -2.84521133e-01 1.68435395e-01
2.82247633e-01 2.50215471e-01 -4.92782183e-02 3.39490831e-01
2.87227511e-01 3.43621075e-01 -3.77310991e-01 -4.93117422e-01
1.10393949e-01 1.83391452e-01 -2.21795365e-01 3.86664607e-02
-3.91699038e-02 1.06093168e+00 -9.66602743e-01 -8.48812878e-01
4.24471438e-01 2.85439312e-01 -5.86354077e-01 -4.43804637e-02
-7.32142389e-01 4.11954999e-01 -4.08382028e-01 1.00150537e+00
7.63208032e-01 -5.01929462e-01 8.97923857e-02 -4.85594362e-01
-2.79272169e-01 2.58838922e-01 -1.18739498e+00 1.97619498e+00
-2.35408351e-01 6.55396640e-01 1.80763111e-01 -1.56289995e-01
3.78647625e-01 -5.35494648e-02 3.38476270e-01 -7.37863183e-01
-1.66858047e-01 4.31602672e-02 -2.33562082e-01 -2.93971032e-01
6.79668903e-01 5.72139978e-01 3.71440873e-02 5.93291700e-01
-3.79086316e-01 -2.49416843e-01 3.52641284e-01 4.02138680e-01
9.41927791e-01 5.63831210e-01 1.98379457e-01 7.89047331e-02
-1.20201804e-01 2.22535089e-01 7.33452201e-01 1.18637192e+00
-2.31957465e-01 9.11735535e-01 5.50879180e-01 -1.29024014e-01
-8.31013739e-01 -1.30004382e+00 -1.70920417e-02 1.24730074e+00
4.67020899e-01 -6.04994237e-01 -8.20211053e-01 -7.56087661e-01
-2.53025115e-01 9.53322232e-01 -3.46335441e-01 5.58221340e-01
-5.71037769e-01 -2.65369028e-01 4.48496252e-01 4.12451833e-01
6.48480892e-01 -1.11025393e+00 -9.88122344e-01 5.24778478e-02
-4.17631477e-01 -1.00689042e+00 -1.13338053e+00 -1.63769703e-02
-5.92022121e-01 -1.17008448e+00 -5.53631008e-01 -6.23472929e-01
7.78451562e-01 4.08117563e-01 1.15919864e+00 1.72904074e-01
-6.22824311e-01 2.33848184e-01 -2.81822741e-01 -2.66425073e-01
-6.60233647e-02 1.87609673e-01 -6.39861524e-01 -3.69610667e-01
-5.74567318e-02 -5.55557251e-01 -1.12905395e+00 6.80976391e-01
-9.48509336e-01 8.03887069e-01 5.86618245e-01 4.72158551e-01
1.00236237e+00 -3.38036776e-01 1.94597274e-01 -1.02430654e+00
3.49178284e-01 -5.46459816e-02 -7.61864603e-01 2.07319573e-01
-4.74803865e-01 -3.17132682e-01 1.72498763e-01 -3.63923103e-01
-1.05041409e+00 3.31318021e-01 -2.38588713e-02 -4.43641573e-01
-2.19227105e-01 2.32731789e-01 -2.30000064e-01 2.43654862e-01
5.79047918e-01 7.27875307e-02 -2.01391205e-01 -4.65552568e-01
7.53243148e-01 4.11766320e-01 6.91492796e-01 -4.52159196e-01
5.15035212e-01 4.22975719e-01 -6.11601353e-01 -3.92768115e-01
-9.15046155e-01 -4.72457647e-01 -5.22402048e-01 -3.66548985e-01
9.07861829e-01 -7.35300899e-01 -5.66315591e-01 1.91791341e-01
-1.15333581e+00 -7.84668386e-01 -1.23804703e-01 -5.26175164e-02
-2.80117601e-01 1.54532000e-01 -3.05251479e-01 -5.21671236e-01
-3.55689168e-01 -1.30252004e+00 1.19984543e+00 5.00585556e-01
-5.24739981e-01 -4.15262282e-01 -5.66333830e-01 4.38469857e-01
1.32332027e-01 1.94509894e-01 5.58610141e-01 -4.97253984e-01
-1.15479398e+00 -2.49862030e-01 -4.09311414e-01 -6.14626855e-02
1.32683307e-01 1.01968922e-01 -1.03487766e+00 8.55854452e-02
-7.32503712e-01 5.14598973e-02 6.66586816e-01 3.77117991e-01
1.51432753e+00 -1.43549949e-01 -6.70414150e-01 5.47286153e-01
1.11851263e+00 2.21028835e-01 5.29858768e-01 2.00286001e-01
9.68516946e-01 4.06158239e-01 9.09546435e-01 3.59714240e-01
1.69677943e-01 8.80561650e-01 3.67693812e-01 -4.03943568e-01
-3.12712252e-01 -4.72047925e-01 -2.35992655e-01 1.85317531e-01
2.86430746e-01 -4.23075914e-01 -9.36794460e-01 4.33463067e-01
-1.89450312e+00 -7.48437822e-01 -9.90738720e-02 2.12596321e+00
1.00228548e+00 3.15046012e-01 5.29349074e-02 -4.57901597e-01
9.38338816e-01 2.08656669e-01 -8.70226383e-01 1.46585852e-01
2.05214862e-02 1.87893078e-01 6.77586138e-01 7.14111090e-01
-1.04734731e+00 1.37287414e+00 5.96405220e+00 1.10914588e+00
-1.01967931e+00 1.33671641e-01 8.08069050e-01 -4.61947054e-01
-5.83753526e-01 2.12521106e-01 -6.41620040e-01 5.53585470e-01
-1.59485586e-04 1.13589123e-01 3.37643564e-01 8.33209991e-01
6.38143897e-01 -6.08832181e-01 -8.41855824e-01 9.79142606e-01
-2.40891755e-01 -1.50674748e+00 -8.32784474e-02 -5.73201105e-02
6.01081252e-01 -4.67748195e-02 -9.17185657e-03 -1.11573949e-01
2.92341530e-01 -8.26543331e-01 9.03380990e-01 5.43315530e-01
9.85316753e-01 -4.55737025e-01 3.24795879e-02 1.36663795e-01
-1.12919271e+00 2.42680699e-01 1.34402126e-01 1.70463681e-01
5.25553465e-01 6.24298930e-01 -1.09331095e+00 1.88978717e-01
6.84068620e-01 4.24537092e-01 -6.19034469e-01 1.33089399e+00
-5.67959666e-01 6.25419438e-01 -4.59887266e-01 1.71332303e-02
4.58869943e-03 -1.40820608e-01 6.56731427e-01 1.38145399e+00
2.78446209e-02 4.18356024e-02 3.69159639e-01 1.27897692e+00
-4.81593460e-02 3.49379820e-03 -1.50228351e-01 -3.44293495e-03
1.03168392e+00 1.43454742e+00 -1.32798743e+00 -4.11830813e-01
-1.66608348e-01 1.34467661e+00 -3.05895992e-02 5.40647149e-01
-1.13983297e+00 -5.17775238e-01 6.30661130e-01 2.49033079e-01
5.30564547e-01 -2.45826393e-01 -9.06560481e-01 -8.47327113e-01
2.99205393e-01 -6.74051464e-01 2.09488422e-01 -1.04861116e+00
-8.33922207e-01 3.48303556e-01 -1.04144271e-02 -1.08992684e+00
2.32358009e-01 -1.87659159e-01 -7.83753335e-01 8.58205318e-01
-9.93016660e-01 -1.20538759e+00 -5.64382613e-01 2.50937015e-01
8.92209470e-01 5.65673649e-01 4.33294713e-01 3.68971914e-01
-6.93097234e-01 6.27007425e-01 -3.83937895e-01 1.43061625e-03
9.44843531e-01 -1.17556083e+00 5.47474205e-01 1.06715345e+00
1.74537987e-01 7.66996801e-01 7.78204799e-01 -9.65313494e-01
-1.03848815e+00 -9.31826353e-01 4.66898501e-01 -3.98583442e-01
3.62388492e-01 -5.22298276e-01 -7.08910525e-01 6.54534459e-01
1.22453928e-01 -2.68339496e-02 6.27984583e-01 -4.91750203e-02
-1.60191759e-01 4.42382470e-02 -9.08723712e-01 1.20107055e+00
1.44091260e+00 -3.19484979e-01 -1.05183259e-01 3.29221636e-01
9.83433306e-01 -7.85209656e-01 -3.91657203e-01 2.79928952e-01
8.17659557e-01 -1.23165286e+00 9.32561934e-01 1.09730668e-01
3.06179434e-01 -7.42671192e-01 1.25635520e-01 -7.53190100e-01
-6.94851279e-02 -9.12371993e-01 1.55469790e-01 1.31837833e+00
8.11269283e-01 -4.19286072e-01 9.80265677e-01 1.01173401e+00
-5.13428450e-01 -9.27719533e-01 -7.80263543e-01 -3.71181279e-01
-6.35468900e-01 -9.47188556e-01 5.04664540e-01 5.31304538e-01
-3.00187141e-01 -4.22854461e-02 -9.21524614e-02 3.02557081e-01
4.29083735e-01 2.51138330e-01 8.96484375e-01 -7.55250454e-01
-3.91693115e-01 -5.49648225e-01 1.30469605e-01 -1.48930097e+00
-3.06905478e-01 -6.69633865e-01 3.79681975e-01 -1.84650123e+00
4.79504541e-02 -7.84996569e-01 1.23081505e-01 8.17982614e-01
-4.18934613e-01 6.77753448e-01 3.87461752e-01 3.39441746e-01
-9.79532301e-01 3.79125103e-02 1.56252050e+00 5.78253530e-02
-5.61600327e-01 7.90926963e-02 -1.01675606e+00 7.53685892e-01
6.34714127e-01 -3.74822438e-01 -5.53751111e-01 -4.96971339e-01
1.88547969e-01 -8.15857574e-02 4.96133536e-01 -8.57335925e-01
3.16134185e-01 -3.10997397e-01 5.08034766e-01 -8.50621164e-01
4.11272883e-01 -4.87988710e-01 2.89069355e-01 1.61532715e-01
-2.39164516e-01 -1.44864678e-01 3.18650782e-01 4.83264208e-01
1.59730971e-01 -2.20039830e-01 4.01860982e-01 -2.80974269e-01
-8.93965483e-01 3.13074708e-01 -2.74899721e-01 4.23239730e-02
9.75523472e-01 -6.54440582e-01 -2.23007649e-01 -1.73334584e-01
-8.34652066e-01 4.89937901e-01 7.81233490e-01 2.40973175e-01
5.01314402e-01 -8.96370590e-01 -1.88128993e-01 3.71119417e-02
1.63391426e-01 5.69118738e-01 4.31247056e-01 8.65302384e-01
-6.79716468e-01 1.62329286e-01 2.37251848e-01 -6.91765010e-01
-1.40785491e+00 1.86638206e-01 2.20203042e-01 1.26885980e-01
-9.19331372e-01 1.11119974e+00 5.19670308e-01 -2.03194082e-01
2.39177123e-01 -4.22989279e-01 2.64666468e-01 -5.44088446e-02
4.55932796e-01 3.18119437e-01 -1.90065056e-01 -1.05622290e-02
-3.66534889e-01 2.33736739e-01 -2.82682627e-01 -4.43862855e-01
8.01640391e-01 -3.39566767e-01 1.10348374e-01 1.49347574e-01
5.26257277e-01 3.57774347e-01 -1.80542672e+00 1.36667658e-02
-1.70609742e-01 -6.90785885e-01 -1.29016057e-01 -1.33007336e+00
-8.85325491e-01 4.61263925e-01 4.05936062e-01 1.90762617e-02
1.09185541e+00 2.10180879e-01 7.61275053e-01 6.18008450e-02
2.46414751e-01 -1.16465890e+00 1.63940251e-01 1.81623071e-01
8.12799513e-01 -1.07492971e+00 -3.22104134e-02 -9.94086564e-01
-8.29544187e-01 8.11481476e-01 9.97772396e-01 2.44003460e-01
2.32045472e-01 4.09569681e-01 1.90055549e-01 -2.37620294e-01
-4.90351707e-01 -2.07292140e-01 1.90443888e-01 4.80165154e-01
4.60208446e-01 9.42295864e-02 -3.27070683e-01 5.20224094e-01
-1.22992732e-01 -2.42963545e-02 4.10025954e-01 8.49744380e-01
-4.47786242e-01 -1.11585653e+00 -3.12786430e-01 6.22874677e-01
-2.62091279e-01 -4.59571451e-01 -2.59451866e-01 5.52031934e-01
2.30654225e-01 9.11878586e-01 1.01994760e-01 -3.98254432e-02
1.63329259e-01 -3.20511758e-02 1.94957405e-01 -8.56295645e-01
-3.95734161e-01 4.87360656e-01 2.06894249e-01 -1.00451815e+00
-1.01189807e-01 -5.64675033e-01 -1.39154899e+00 1.02383442e-01
-3.28790069e-01 -2.32119560e-01 5.62930286e-01 7.10613430e-01
7.33786881e-01 6.36804163e-01 -6.92171380e-02 -9.51385260e-01
1.55918106e-01 -8.15880179e-01 -1.86552942e-01 7.16466829e-02
-1.15755938e-01 -5.10480762e-01 2.56303344e-02 2.73466408e-01] | [9.481090545654297, -0.09500307589769363] |
299f5ef4-326d-419e-b50c-6f2ceb16b231 | deeply-self-supervising-edge-to-contour | 1808.00739 | null | https://arxiv.org/abs/1808.00739v5 | https://arxiv.org/pdf/1808.00739v5.pdf | Deeply Self-Supervised Contour Embedded Neural Network Applied to Liver Segmentation | Objective: Herein, a neural network-based liver segmentation algorithm is proposed, and its performance was evaluated using abdominal computed tomography (CT) images. Methods: A fully convolutional network was developed to overcome the volumetric image segmentation problem. To guide a neural network to accurately delineate a target liver object, the network was deeply supervised by applying the adaptive self-supervision scheme to derive the essential contour, which acted as a complement with the global shape. The discriminative contour, shape, and deep features were internally merged for the segmentation results. Results and Conclusion: 160 abdominal CT images were used for training and validation. The quantitative evaluation of the proposed network was performed through an eight-fold cross-validation. The result showed that the method, which uses the contour feature, segmented the liver more accurately than the state-of-the-art with a 2.13% improvement in the dice score. Significance: In this study, a new framework was introduced to guide a neural network and learn complementary contour features. The proposed neural network demonstrates that the guided contour features can significantly improve the performance of the segmentation task. | ['Yeong-Gil Shin', 'Jeongjin Lee', 'Minkyung Lee', 'Jingyu Lee', 'Minyoung Chung'] | 2018-08-02 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [-1.23592228e-01 1.15484171e-01 -1.61723897e-01 -6.81011915e-01
-3.20158452e-01 -2.47412205e-01 2.04287320e-01 2.65824765e-01
-5.97178757e-01 4.70848888e-01 1.16160512e-01 -1.76475629e-01
5.92836738e-02 -6.75944448e-01 -2.49018103e-01 -1.00057590e+00
-5.11893868e-01 4.95703578e-01 1.12889282e-01 1.37357429e-01
2.02491805e-01 6.86140180e-01 -6.05789363e-01 -5.13253249e-02
1.13338292e+00 1.25591278e+00 1.21804632e-01 3.39626461e-01
-1.92812115e-01 6.38613582e-01 -3.50291252e-01 6.01002946e-02
4.74486083e-01 -7.97517955e-01 -8.29489946e-01 3.22197467e-01
-1.31369144e-01 -5.03915906e-01 -7.83551261e-02 1.10055661e+00
6.97917044e-01 -9.31464285e-02 5.93316138e-01 -5.42220592e-01
-5.06595254e-01 7.74633944e-01 -4.95238185e-01 3.93017173e-01
-1.75733030e-01 -1.42781576e-02 3.14174533e-01 -8.12712610e-01
3.69284779e-01 4.54219073e-01 8.33409905e-01 2.43148774e-01
-8.02471161e-01 -6.34551466e-01 -5.27732015e-01 -2.50607342e-01
-1.35276306e+00 3.05322766e-01 7.90837467e-01 -6.16181374e-01
3.28120500e-01 5.79340160e-02 1.35922527e+00 1.68710217e-01
6.22513115e-01 7.64161885e-01 1.19601929e+00 -2.87284911e-01
-1.02622423e-03 6.79048225e-02 1.43448338e-01 1.43158865e+00
4.37207311e-01 4.19135392e-01 3.77247036e-01 1.66165099e-01
1.18063211e+00 4.47180048e-02 -3.98072451e-01 -6.77005947e-01
-1.12177896e+00 1.16224051e+00 1.23353040e+00 6.84167862e-01
-5.64140916e-01 -8.85133594e-02 5.80168903e-01 -2.71058083e-01
4.78270382e-01 5.18443406e-01 -2.62773812e-01 4.29273158e-01
-1.19884920e+00 -4.89798695e-01 9.29116905e-01 5.56624413e-01
4.11993772e-01 1.86358392e-01 -4.06897783e-01 4.05258447e-01
6.16479814e-01 3.16834807e-01 9.06181931e-01 -4.90810812e-01
-2.28548288e-01 7.83656478e-01 -4.85305846e-01 -9.27667677e-01
-9.49814975e-01 -8.53243530e-01 -1.25711942e+00 1.87981784e-01
4.07452852e-01 -4.45381612e-01 -1.40301180e+00 1.11597645e+00
4.08641636e-01 -1.74495444e-01 -1.85843080e-01 1.47062695e+00
1.29400718e+00 2.07546219e-01 3.10101449e-01 -1.83776781e-01
1.37356162e+00 -1.21397102e+00 -6.40013814e-01 3.38172197e-01
6.66323125e-01 -6.35556161e-01 3.49225909e-01 1.67758584e-01
-9.44589972e-01 -4.35971230e-01 -1.09977961e+00 4.76180136e-01
-4.94791120e-02 3.36625874e-01 8.54278982e-01 7.54644215e-01
-1.06181896e+00 6.15159690e-01 -1.00945508e+00 -2.18361318e-01
7.60830164e-01 5.46741545e-01 -2.97324866e-01 1.93845928e-01
-1.08199918e+00 9.14137483e-01 8.13754439e-01 2.15353206e-01
-9.26476181e-01 -6.22089624e-01 -9.55278933e-01 9.10933968e-03
-9.50763002e-02 -7.89091766e-01 9.85592663e-01 -1.21941781e+00
-1.70235085e+00 9.78201270e-01 5.03130615e-01 -5.67498982e-01
5.68053842e-01 2.55394667e-01 8.54271129e-02 8.04877579e-01
-1.11146085e-02 7.36091435e-01 4.45477873e-01 -1.20446694e+00
-3.17080528e-01 -1.94219843e-01 -5.28877139e-01 2.18814090e-01
8.57203268e-03 3.34708542e-02 -5.26061058e-01 -8.98726881e-01
3.86229306e-01 -7.96621859e-01 -4.17740524e-01 1.49356201e-01
-3.20761055e-01 -6.72316924e-02 4.96078700e-01 -9.78850961e-01
9.01087582e-01 -1.82472742e+00 -5.25574498e-02 5.14909685e-01
3.58026713e-01 3.00307542e-01 2.74125993e-01 -3.08122605e-01
-8.62460583e-02 1.65302143e-01 -7.53219306e-01 4.46475446e-02
-5.28541684e-01 6.84323385e-02 6.61081314e-01 9.62658703e-01
-2.24377532e-02 1.15282297e+00 -8.14785182e-01 -1.07843113e+00
3.39644819e-01 5.25400996e-01 -3.93275857e-01 4.70579118e-01
2.47408450e-01 9.37893569e-01 -3.40345621e-01 5.80053329e-01
7.65809953e-01 -4.60995525e-01 9.77607518e-02 -3.58400464e-01
-1.32234216e-01 -5.46126485e-01 -4.96900648e-01 1.94765639e+00
-1.70910001e-01 3.40089798e-01 2.08154336e-01 -9.23209429e-01
1.21805239e+00 5.52725613e-01 9.27399039e-01 -8.39606524e-01
7.78509498e-01 4.15758222e-01 2.75772184e-01 -7.99572587e-01
-2.13233918e-01 -4.72652316e-01 1.74140647e-01 3.98526162e-01
1.39692008e-01 -4.31187034e-01 9.17681530e-02 -1.74068123e-01
4.04252797e-01 1.45192966e-01 4.05917853e-01 -8.96896839e-01
8.74180496e-01 4.61070508e-01 5.14065027e-01 3.45628411e-01
-6.54473186e-01 6.30337417e-01 1.57100379e-01 -8.79689097e-01
-7.34158933e-01 -6.74347281e-01 -5.76091409e-01 5.66871107e-01
4.62931156e-01 3.79968256e-01 -1.01384616e+00 -1.03810620e+00
-1.18944541e-01 2.07637042e-01 -8.21552813e-01 5.16946465e-02
-8.79921615e-01 -1.07950544e+00 4.80690211e-01 5.22750258e-01
1.00749230e+00 -1.14352059e+00 -9.46722269e-01 1.92434460e-01
-5.97170033e-02 -6.79181218e-01 -6.45086229e-01 3.04850578e-01
-1.24717426e+00 -1.22527099e+00 -1.32192123e+00 -1.20260608e+00
1.07398665e+00 -4.28703241e-02 9.39262986e-01 5.26089787e-01
-5.31721652e-01 4.19223756e-02 -1.78410709e-01 -8.32264572e-02
-4.34687287e-01 -1.59952249e-02 -4.30306762e-01 -2.68015593e-01
-8.59258994e-02 -1.48730725e-01 -1.23963094e+00 1.46607026e-01
-8.76097262e-01 1.72192320e-01 1.04548979e+00 1.08195269e+00
5.47203898e-01 -1.31932944e-01 3.59708339e-01 -7.15015531e-01
4.03494120e-01 -4.78010148e-01 -4.56925482e-01 -1.94912348e-02
-9.07381773e-01 -4.40085530e-02 6.16067171e-01 -9.50516462e-02
-9.79889214e-01 2.11622149e-01 -2.34132022e-01 -1.85660958e-01
-1.11110069e-01 7.86549449e-01 5.81496179e-01 -5.51191092e-01
4.46973711e-01 3.21442038e-01 5.77429950e-01 -2.70197392e-01
7.12349713e-02 3.38092655e-01 3.89156610e-01 -2.46379375e-01
2.84458727e-01 3.78451794e-01 2.67135799e-01 -3.32300931e-01
-6.02164686e-01 -6.26959205e-01 -1.01707494e+00 -2.77171165e-01
1.16327024e+00 -7.49012053e-01 -3.53418350e-01 4.42595750e-01
-8.33385944e-01 -5.34862339e-01 -1.72083691e-01 8.98104310e-01
-3.28819305e-01 6.57734573e-01 -1.02554643e+00 -2.42038846e-01
-1.05069137e+00 -1.54173660e+00 4.61166024e-01 6.59879982e-01
2.41177827e-01 -1.41821027e+00 7.07369819e-02 3.97093855e-02
1.00270069e+00 6.37346804e-01 6.96987867e-01 -1.04443157e+00
-2.69692093e-01 -1.69193700e-01 -7.31097996e-01 2.80430794e-01
2.35383213e-01 -1.93663895e-01 -4.08192575e-01 -5.33308566e-01
1.97948545e-01 -1.83803707e-01 7.38538504e-01 7.44036913e-01
1.17997062e+00 -1.56335205e-01 -9.48930010e-02 1.24425304e+00
1.68193781e+00 3.42701644e-01 8.72146264e-02 2.97481567e-01
3.82954359e-01 1.48066044e-01 2.30327204e-01 4.12013888e-01
1.67649493e-01 1.25956804e-01 6.37432933e-01 -1.01709104e+00
-3.90842378e-01 2.03985155e-01 -3.87301981e-01 9.79732692e-01
-1.87649429e-01 3.02664489e-01 -1.09320259e+00 4.56614524e-01
-1.25065768e+00 -2.74252832e-01 -2.09304377e-01 1.58443284e+00
9.30164576e-01 -1.66266561e-01 -1.89789012e-03 -2.30498374e-01
6.39118552e-01 -1.94446385e-01 -2.89597034e-01 -1.22623622e-01
1.86381653e-01 4.10978198e-01 5.41971326e-01 3.51421505e-01
-1.61911440e+00 4.53327298e-01 6.31645632e+00 4.63101685e-01
-1.68698466e+00 9.83723551e-02 8.88079405e-01 6.93526089e-01
1.69062942e-01 -4.00549442e-01 -1.84780546e-02 2.92165786e-01
5.06698430e-01 -8.24951902e-02 1.69466063e-02 7.45870888e-01
9.51915309e-02 -2.89709538e-01 -6.66086853e-01 6.89810932e-01
3.22009444e-01 -1.19995356e+00 -2.22659633e-02 -1.37610108e-01
8.75927687e-01 2.03908473e-01 -6.09672256e-02 1.78331509e-01
4.55085449e-02 -1.22081625e+00 3.29762667e-01 6.52833283e-01
9.53460157e-01 -5.31059086e-01 1.56403804e+00 2.91045457e-01
-1.12521291e+00 1.50974929e-01 -4.75740731e-02 3.12916994e-01
-4.62508835e-02 5.79353869e-01 -1.26312721e+00 8.52919221e-01
3.32668066e-01 5.11204422e-01 -6.49921536e-01 1.59735966e+00
-1.72942013e-01 5.76990843e-01 -2.57012427e-01 -5.67333773e-02
6.12330496e-01 -4.39850211e-01 2.57068723e-01 1.66659975e+00
2.67496347e-01 2.84425557e-01 5.31909764e-01 9.63288724e-01
-4.01075408e-02 6.78972244e-01 -1.87170997e-01 3.51390153e-01
-1.35990992e-01 1.60082865e+00 -1.49434924e+00 -5.49771905e-01
-2.14791760e-01 8.50015521e-01 -1.40077546e-01 1.20692641e-01
-9.17508900e-01 -2.60589451e-01 -4.59574223e-01 -1.55575708e-01
8.68448988e-02 -1.26093009e-03 -6.11286819e-01 -9.74348247e-01
-5.32937706e-01 -2.98350930e-01 5.25378108e-01 -2.63681829e-01
-1.13095951e+00 8.24027658e-01 -1.62160292e-01 -1.02023506e+00
1.86513722e-01 -4.38410223e-01 -8.07245314e-01 9.46080148e-01
-1.76809227e+00 -1.15165126e+00 -8.47806275e-01 3.38882178e-01
3.66937011e-01 -1.43948555e-01 9.10280406e-01 1.26490027e-01
-4.31211144e-01 2.77043611e-01 -6.89364746e-02 8.52557421e-01
2.97548324e-01 -1.49010289e+00 -5.30783892e-01 7.22745299e-01
-3.99852931e-01 5.52170634e-01 1.67206272e-01 -7.11508095e-01
-1.02636790e+00 -1.16010487e+00 3.42823714e-01 3.51951897e-01
2.31920645e-01 5.06644845e-01 -6.33819997e-01 4.43282872e-01
5.31763256e-01 7.89036393e-01 9.14833307e-01 -8.61686587e-01
3.94163370e-01 1.31026253e-01 -1.68496978e+00 -7.48559088e-02
2.18150958e-01 1.32422388e-01 -7.34126031e-01 7.63092935e-02
3.81206751e-01 -7.46478140e-01 -1.33799195e+00 7.97135532e-01
5.09642124e-01 -1.15215397e+00 9.42233026e-01 -1.39671072e-01
5.54989934e-01 -2.09104270e-01 3.28116775e-01 -1.31793129e+00
-4.99508619e-01 -7.48413131e-02 5.21256961e-02 4.42584157e-01
2.26586834e-01 -4.83270556e-01 7.75695682e-01 2.54710644e-01
-5.02608716e-01 -1.24797368e+00 -8.15450788e-01 -2.03862622e-01
4.85082835e-01 3.79307628e-01 4.35805321e-01 9.97146308e-01
-5.09779155e-02 -1.20049499e-01 3.19638133e-01 -6.60321265e-02
7.79150665e-01 2.97328442e-01 9.69335437e-02 -1.18840802e+00
1.92288578e-01 -7.30869353e-01 -3.84150952e-01 -1.06915975e+00
-2.46744022e-01 -1.44459283e+00 6.09085336e-02 -1.74523485e+00
5.00237763e-01 -6.74785495e-01 -5.69175720e-01 3.15877855e-01
-2.30128154e-01 4.73495394e-01 2.08720863e-01 3.22606415e-01
-3.01612884e-01 4.61163193e-01 2.17332172e+00 -2.67522126e-01
-9.37901512e-02 2.71342248e-01 -4.97722030e-01 6.48271918e-01
9.70819116e-01 -3.00670087e-01 -3.76336873e-02 -1.94080174e-01
-5.05549788e-01 4.83590454e-01 2.99795300e-01 -1.05369294e+00
5.22712588e-01 2.11507857e-01 1.07087171e+00 -7.19274342e-01
-3.46059263e-01 -1.08866727e+00 -1.02601513e-01 1.29536045e+00
-1.54325351e-01 -2.43063390e-01 2.53472656e-01 1.23636648e-01
-3.57809871e-01 -3.03738147e-01 1.15186167e+00 -4.71583933e-01
-4.38638002e-01 5.32634139e-01 -1.09952815e-01 -1.36312112e-01
1.39894652e+00 -8.10810402e-02 2.51264513e-01 -6.55081794e-02
-8.47146869e-01 1.61068976e-01 1.07284993e-01 -4.19537872e-01
8.59790564e-01 -1.25351632e+00 -7.99722850e-01 6.03971541e-01
-3.36232662e-01 3.93469632e-01 -5.69534563e-02 1.59394622e+00
-1.31953883e+00 5.18549025e-01 -4.48541105e-01 -1.02169561e+00
-8.61339152e-01 4.46835786e-01 1.03273499e+00 -5.56796372e-01
-8.91387641e-01 8.76472294e-01 5.92071302e-02 -4.79509383e-01
1.47469744e-01 -6.69507384e-01 -4.86160249e-01 -1.60293281e-01
1.11997098e-01 -4.26320471e-02 9.42192972e-03 -6.25897467e-01
-2.31810242e-01 7.05682695e-01 3.22990119e-01 3.37267458e-01
1.31127548e+00 -7.73966685e-02 -4.89101827e-01 -1.45452783e-01
1.27146280e+00 -2.67824411e-01 -1.18355310e+00 -1.92643076e-01
-4.26439829e-02 -2.75533736e-01 4.24032182e-01 -1.20168638e+00
-1.72574496e+00 8.54669213e-01 1.02169180e+00 1.03361860e-01
1.16018641e+00 -3.44501138e-01 8.76024902e-01 -2.93250084e-01
2.82294184e-01 -5.41905701e-01 4.57649752e-02 3.93531054e-01
7.91247666e-01 -1.53924048e+00 2.28660539e-01 -3.20227474e-01
-6.41231060e-01 1.77105463e+00 5.36498070e-01 -3.49030703e-01
9.27243173e-01 2.72503883e-01 4.31444436e-01 -4.55560267e-01
3.32961082e-01 -8.88863727e-02 4.27503973e-01 1.96244791e-01
7.32618093e-01 2.89679408e-01 -5.07745862e-01 6.54802442e-01
9.57083032e-02 2.78998405e-01 1.55499890e-01 7.45501399e-01
-5.98952949e-01 -5.63782752e-01 -3.22362363e-01 4.39498216e-01
-7.20104933e-01 -5.06223775e-02 1.49979487e-01 1.08871615e+00
2.77715117e-01 4.61790413e-01 -5.59406541e-02 7.94280544e-02
-8.42214301e-02 -9.80937332e-02 3.51719588e-01 -4.60149556e-01
-1.12027478e+00 2.43977472e-01 -4.05101478e-01 -3.33278477e-01
-5.79540551e-01 -1.22748718e-01 -1.76069260e+00 4.61561121e-02
-4.16270822e-01 3.87860388e-01 7.62137890e-01 7.25567937e-01
-1.72165647e-01 7.84065902e-01 9.35396433e-01 -9.02025104e-01
-5.98365307e-01 -1.15491068e+00 -4.01210606e-01 5.49058437e-01
3.48209053e-01 -4.93175805e-01 -3.39899063e-01 3.14451791e-02] | [14.46501636505127, -2.7685413360595703] |
cc3a1b60-524d-4500-b135-d256a9d741c7 | rethinking-zero-shot-action-recognition | null | null | https://link.springer.com/chapter/10.1007/978-3-031-19772-7_7 | https://www.ecva.net/papers/eccv_2022/papers_ECCV/papers/136640102.pdf | Rethinking Zero-shot Action Recognition: Learning from Latent Atomic Actions | To avoid time-consuming annotating and retraining cycle in
applying supervised action recognition models, Zero-Shot Action Recognition (ZSAR) has become a thriving direction. ZSAR requires models to recognize actions that never appear in training set through bridging visual features and semantic representations. However, due to the complexity of actions, it remains challenging to transfer knowledge learned from source to target action domains. Previous ZSAR methods mainly focus on mitigating representation variance between source and target actions through integrating or applying new action-level features. However, the action-level features are coarse-grained and make the learned one-to-one bridge fragile to similar target actions. Meanwhile, integration or application of features usually requires extra computation or annotation. These methods didn’t notice that two actions with different names may still share the same atomic action components. It enables humans to quickly understand an unseen action given bunch of atomic actions learned from seen actions. Inspired by this, we propose Jigsaw Network (JigsawNet) which recognizes complex actions through unsupervisedly decomposing them into combinations of atomic actions and
bridging group to group relationships between visual features and semantic representations. To enhance the robustness of learned group-to-group bridge, we propose Group Excitation (GE) module to model intra-sample
knowledge and Consistency Loss to enforce the model learn from inter-sample knowledge. Our JigsawNet achieves state-of-the-art performance on three benchmarks and surpasses previous works with noticeable margins. | ['and Alexander G. Hauptmann', 'Wenhe Liu', 'Lijun Yu', 'Yijun Qian'] | 2022-03-28 | null | null | null | eccv-2022-10 | ['zero-shot-action-recognition'] | ['computer-vision'] | [ 4.95860010e-01 1.25873485e-03 -3.81794870e-01 -3.58874798e-01
-5.89493275e-01 -2.80282825e-01 6.32092476e-01 -2.49115393e-01
-5.08715771e-02 5.01090884e-01 4.78340149e-01 3.56602192e-01
-1.87456414e-01 -6.58166468e-01 -7.54158199e-01 -6.45755291e-01
1.13427900e-01 2.28051692e-01 6.82458639e-01 -2.82491475e-01
1.04168147e-01 3.51779461e-01 -1.77192760e+00 9.33229744e-01
6.12478316e-01 1.06617749e+00 1.03901640e-01 2.54860222e-01
-3.57857496e-01 1.28559077e+00 -7.42912292e-01 -1.43028468e-01
3.44516873e-01 -8.52418423e-01 -8.44558895e-01 5.17805517e-01
5.96055388e-01 -3.13861161e-01 -4.90979224e-01 8.97936702e-01
1.28983423e-01 4.54865068e-01 4.88886744e-01 -1.55637586e+00
-1.04204881e+00 5.16956270e-01 -6.99543357e-01 1.78745702e-01
4.52018827e-01 5.20707488e-01 1.04415488e+00 -7.19750822e-01
6.86579943e-01 1.46966183e+00 3.72742385e-01 7.42825150e-01
-1.01120090e+00 -6.98244691e-01 6.63383722e-01 6.16046488e-01
-1.14031494e+00 -4.03930575e-01 6.88879430e-01 -3.39100510e-01
1.29013216e+00 1.48971930e-01 5.92431664e-01 1.26586401e+00
-1.45114744e-02 1.15263855e+00 8.52351010e-01 -7.18828291e-02
1.14853688e-01 -3.43517184e-01 1.74807057e-01 6.43963635e-01
-5.98941883e-03 -3.44978310e-02 -8.37015927e-01 2.77837813e-01
8.33767533e-01 5.35951436e-01 -1.38921648e-01 -5.01731396e-01
-1.25585043e+00 4.79217529e-01 6.05473399e-01 3.78706872e-01
-3.23855191e-01 3.76575202e-01 5.56563675e-01 3.22222590e-01
-1.30446330e-01 3.05837840e-01 -5.32872260e-01 -2.68644422e-01
-4.28789854e-01 -1.32776573e-01 1.61231488e-01 9.03894663e-01
9.02288795e-01 2.06114069e-01 -4.35379744e-01 7.20411241e-01
-1.13572098e-01 4.12068814e-01 7.81577587e-01 -8.59786093e-01
3.83599073e-01 1.18570125e+00 -1.33244365e-01 -9.32865977e-01
-1.18676342e-01 -8.67026299e-02 -8.22849393e-01 2.68468529e-01
3.77065867e-01 4.66872871e-01 -1.18650687e+00 1.61215174e+00
1.81264043e-01 4.95164782e-01 1.99321538e-01 8.96910429e-01
8.99224102e-01 4.74806219e-01 3.71413082e-01 -1.44290999e-01
1.17978537e+00 -1.41928256e+00 -6.63403869e-01 -4.17143136e-01
7.98877001e-01 -3.60513002e-01 1.37364423e+00 2.63546675e-01
-7.23555505e-01 -9.95785773e-01 -8.42266858e-01 -9.00681913e-02
-5.23055196e-01 1.75689474e-01 6.33321464e-01 1.19716758e-02
-4.28937793e-01 6.93428516e-01 -9.71816301e-01 -3.33632916e-01
7.37526417e-01 3.16952646e-01 -8.01546574e-01 -1.27046362e-01
-1.04784000e+00 6.97703481e-01 4.66341138e-01 -1.21773906e-01
-1.17080426e+00 -7.02263236e-01 -9.63023186e-01 6.76315799e-02
1.00421286e+00 -2.04614386e-01 1.13558435e+00 -1.45171189e+00
-1.50234962e+00 6.55085087e-01 -2.55057458e-02 -3.80251110e-01
1.77487642e-01 -5.54937780e-01 -6.39794528e-01 1.13727875e-01
2.69819856e-01 5.37152827e-01 9.38139737e-01 -9.84260798e-01
-6.40635788e-01 -4.24433768e-01 2.44729653e-01 1.09812655e-01
-1.90348297e-01 -4.13767248e-02 -3.09990853e-01 -7.37209141e-01
1.99851021e-01 -5.78319728e-01 3.83896418e-02 1.95886403e-01
-2.83647865e-01 -4.12952542e-01 1.11387861e+00 -3.38782907e-01
1.00717616e+00 -2.50665689e+00 1.55885175e-01 -2.48802841e-01
2.60379136e-04 4.11930650e-01 -5.05376995e-01 4.03209895e-01
-3.03873420e-01 -2.51599431e-01 -8.45798329e-02 -6.30882233e-02
1.60614885e-02 5.31229198e-01 -5.25280654e-01 2.31112391e-01
4.36385185e-01 1.06367803e+00 -1.11591029e+00 -3.18811208e-01
4.52094793e-01 2.18880251e-01 -3.27355683e-01 3.14960003e-01
-2.54219502e-01 5.06639719e-01 -4.71967429e-01 8.71108413e-01
2.98844129e-01 -3.53479058e-01 -2.05005966e-02 -4.42354023e-01
3.47930700e-01 6.13577776e-02 -1.27066875e+00 1.82996166e+00
-3.10803235e-01 4.00665291e-02 -5.03282845e-01 -1.47102010e+00
1.01621747e+00 1.27313673e-01 5.30408025e-01 -8.24473202e-01
1.02797158e-01 -7.34565556e-02 2.24508401e-02 -5.49104452e-01
-5.89720160e-02 -1.01994559e-01 -2.43334681e-01 3.89475226e-01
4.19610083e-01 2.12807655e-01 1.30665854e-01 1.23289168e-01
1.32982838e+00 5.39912045e-01 6.08721435e-01 2.52277106e-01
6.95196807e-01 6.22429624e-02 8.93712044e-01 5.80755889e-01
-3.74205768e-01 5.26713073e-01 3.63975585e-01 -7.94433177e-01
-5.33516288e-01 -1.17238855e+00 4.57944363e-01 1.52127087e+00
4.52387631e-01 -4.39159274e-01 -3.48555297e-01 -1.15353191e+00
2.83942353e-02 6.01780355e-01 -8.97765934e-01 -6.47216320e-01
-4.98493969e-01 -1.38527855e-01 4.36320633e-01 1.06431484e+00
8.95641923e-01 -1.47173905e+00 -7.76869416e-01 1.49178028e-01
3.60298976e-02 -1.08914196e+00 -4.91064519e-01 1.22995868e-01
-6.51499033e-01 -1.49636149e+00 -2.74652332e-01 -5.40820181e-01
6.56981051e-01 5.02821028e-01 1.14588010e+00 4.96173464e-02
-4.10539299e-01 4.49460655e-01 -7.55085468e-01 -2.27124140e-01
-9.69760865e-02 -4.55710143e-01 1.41694605e-01 3.50765973e-01
7.53898978e-01 -5.35000145e-01 -5.33412397e-01 6.87734544e-01
-8.38963211e-01 7.53726736e-02 7.34981656e-01 7.27633059e-01
7.88112998e-01 7.11086392e-03 5.94962180e-01 -5.88651240e-01
9.57718119e-02 -2.84519523e-01 -2.20083028e-01 6.39582038e-01
-1.49502993e-01 5.85634597e-02 8.48323584e-01 -7.53617764e-01
-1.01006615e+00 3.48187417e-01 4.66517150e-01 -9.21912611e-01
-3.43772739e-01 1.14829786e-01 -5.41967034e-01 2.67127395e-01
6.02624118e-01 4.62285727e-01 4.72417381e-03 -4.62577581e-01
6.47686124e-01 3.60135794e-01 7.28745997e-01 -4.16458875e-01
5.54897189e-01 6.07348979e-01 -1.94073796e-01 -4.35724527e-01
-1.18099523e+00 -4.84170824e-01 -8.64778042e-01 -1.81926236e-01
1.15890825e+00 -8.94620657e-01 -5.07391691e-01 5.44555187e-01
-8.99780691e-01 -3.30826491e-01 -7.24695265e-01 3.39986354e-01
-7.03322768e-01 3.30477595e-01 -3.18516403e-01 -4.74499643e-01
-2.10223477e-02 -1.00258446e+00 1.15518534e+00 1.74466193e-01
-2.12836862e-01 -5.22536397e-01 -7.80598819e-02 4.23245996e-01
1.38714120e-01 2.10088775e-01 5.98216772e-01 -7.91751087e-01
-5.42919397e-01 -2.10414872e-01 -1.81012794e-01 6.16404772e-01
6.58713758e-01 -7.06095248e-02 -6.72044992e-01 -8.97655189e-02
-1.27538159e-01 -6.94630206e-01 9.58824396e-01 9.21456367e-02
1.40716779e+00 -2.53539622e-01 -2.35162064e-01 3.74065220e-01
1.12075365e+00 2.86547661e-01 8.37852955e-01 1.60054103e-01
9.54432905e-01 5.07698059e-01 9.44817364e-01 3.49847466e-01
-2.28271186e-02 7.32812285e-01 4.11752224e-01 1.40577182e-01
-4.12493825e-01 -4.44506109e-01 7.07672536e-01 3.46034914e-01
-2.20352545e-01 1.28156379e-01 -6.05719209e-01 4.96282220e-01
-2.13446212e+00 -1.35414135e+00 2.00307503e-01 2.00181770e+00
6.34237885e-01 1.66533217e-01 1.72065631e-01 3.25006731e-02
6.38985217e-01 2.96830237e-01 -8.87817681e-01 -2.00656001e-02
-2.17081700e-03 3.13931674e-01 2.21944094e-01 -9.29186586e-03
-1.10027516e+00 1.30033684e+00 5.34312105e+00 1.09063482e+00
-9.61602390e-01 6.75255209e-02 3.68625939e-01 -8.97550285e-02
2.58365452e-01 -1.76140592e-02 -7.75022745e-01 3.94970715e-01
3.91648769e-01 -3.70543227e-02 2.49365523e-01 1.22423911e+00
1.03987977e-01 -5.30509762e-02 -1.32719386e+00 1.07621109e+00
1.85474515e-01 -1.22874665e+00 3.86869758e-01 -3.99109125e-01
7.52707362e-01 -3.38936329e-01 -2.26991966e-01 7.53258467e-01
4.71023262e-01 -1.14653075e+00 4.26287919e-01 5.68730712e-01
6.38814330e-01 -5.90877056e-01 6.05303347e-01 1.84008747e-01
-1.61195147e+00 -3.91460568e-01 -5.25978565e-01 -2.65909523e-01
1.20621495e-01 7.16357976e-02 -2.41760328e-01 5.96462309e-01
7.51715302e-01 1.32868540e+00 -5.32389939e-01 6.04366660e-01
-4.64573592e-01 2.34077021e-01 1.39044270e-01 3.44376951e-01
3.62644970e-01 -4.71974760e-02 1.83419898e-01 6.54073119e-01
1.80864215e-01 2.93023378e-01 6.07854068e-01 7.00122833e-01
1.31261975e-01 -2.44508609e-01 -7.21265614e-01 -3.51404697e-01
1.91767156e-01 1.00104320e+00 -5.97866416e-01 -6.35386348e-01
-7.53442645e-01 1.34079242e+00 2.84675300e-01 3.21078449e-01
-9.92228448e-01 -1.63775086e-01 8.84196520e-01 6.30255342e-02
4.54969287e-01 -5.32598281e-03 -1.40615171e-02 -1.19680870e+00
6.71140775e-02 -1.05053675e+00 6.79581523e-01 -7.65515745e-01
-1.35229266e+00 4.11831141e-01 9.34776142e-02 -1.69159889e+00
-7.17691481e-02 -4.99705791e-01 -6.83069706e-01 3.09753835e-01
-1.02317107e+00 -1.41051209e+00 -6.21192873e-01 1.03867674e+00
1.09802938e+00 -3.88570338e-01 7.32120931e-01 -1.93772893e-02
-4.78014499e-01 6.49646997e-01 -3.70580107e-01 2.85351306e-01
6.85005426e-01 -9.26248908e-01 2.36016557e-01 7.63824522e-01
3.72888237e-01 3.80506456e-01 3.10505986e-01 -6.66034400e-01
-1.23153722e+00 -1.19826424e+00 3.79024148e-01 -5.64567387e-01
7.93233156e-01 -1.38439864e-01 -1.24985731e+00 8.76745999e-01
-1.01196311e-01 5.47894716e-01 5.86264729e-01 -9.58339199e-02
-7.44466603e-01 -2.04407200e-01 -8.75269830e-01 4.56569254e-01
1.71586680e+00 -6.38739049e-01 -9.97167110e-01 2.63499200e-01
7.02216446e-01 -1.45377545e-02 -5.75330317e-01 5.71085393e-01
3.74026090e-01 -1.05239344e+00 9.45793331e-01 -1.19904363e+00
5.66976190e-01 -6.58356845e-01 -2.64205217e-01 -1.11566925e+00
-4.59845901e-01 -2.61787653e-01 -3.34562659e-01 1.01383960e+00
-4.14389670e-02 -3.24048281e-01 7.24288762e-01 4.83116090e-01
-3.13190192e-01 -6.72572255e-01 -7.65914083e-01 -1.28364027e+00
-3.56670499e-01 -2.77621031e-01 6.76524818e-01 1.01298046e+00
1.60945971e-02 3.67477745e-01 -5.66837132e-01 2.13382822e-02
3.46786648e-01 1.90815255e-01 1.07333815e+00 -1.00725901e+00
-5.96971750e-01 -4.24422204e-01 -8.78238916e-01 -1.03328800e+00
2.36743808e-01 -7.54322410e-01 1.12245709e-01 -1.47901750e+00
2.83716381e-01 -9.43260565e-02 -6.95592821e-01 1.13183153e+00
-2.03358173e-01 3.28158438e-01 2.83374578e-01 2.14948118e-01
-1.03167975e+00 8.51206422e-01 1.42741489e+00 -4.09900844e-01
-2.31091678e-01 -3.44356358e-01 -5.43958485e-01 7.67913997e-01
5.63163757e-01 -3.34186256e-01 -8.27823579e-01 -3.23370218e-01
-2.17982382e-01 -2.20758244e-01 4.96708661e-01 -1.36211729e+00
8.50139856e-02 -6.60322726e-01 2.81230003e-01 -3.87853503e-01
3.19520116e-01 -9.53587592e-01 2.32774943e-01 4.23331708e-01
-3.64997655e-01 -2.26675928e-01 3.67046196e-05 8.37440968e-01
-4.44690466e-01 7.96314552e-02 7.67399251e-01 -2.91071326e-01
-1.49278224e+00 3.50676209e-01 -1.32159367e-02 3.46255638e-02
1.48606861e+00 -5.69266975e-01 -4.19367552e-01 -1.42575696e-01
-9.68171179e-01 1.87017098e-01 2.61304677e-01 7.86536813e-01
6.99506342e-01 -1.56588447e+00 -3.49396169e-01 2.97339916e-01
5.35727561e-01 -1.84471952e-03 5.67778826e-01 8.63568962e-01
-6.23731315e-02 -3.54329962e-03 -6.01502955e-01 -4.65846390e-01
-1.38206446e+00 8.58255327e-01 3.23280215e-01 -2.13125676e-01
-9.03007865e-01 1.02195728e+00 6.24550521e-01 -1.45030215e-01
3.60823005e-01 -2.64334083e-01 -1.40308157e-01 -1.31883055e-01
8.81190240e-01 3.31370443e-01 -3.65268350e-01 -7.81360388e-01
-5.80824554e-01 7.69335210e-01 -2.69147962e-01 4.48796809e-01
1.29688740e+00 2.38451719e-01 1.28483981e-01 6.22822821e-01
1.00792527e+00 -4.71883595e-01 -1.65494227e+00 -3.44385624e-01
-8.07299465e-02 -7.74443507e-01 -3.92828912e-01 -8.27028871e-01
-1.18709302e+00 8.53107452e-01 5.38289726e-01 -2.40676373e-01
1.18320763e+00 3.02318007e-01 7.61203766e-01 5.13850212e-01
4.73688781e-01 -1.33801889e+00 9.72325683e-01 4.38252479e-01
1.09516656e+00 -1.36669540e+00 -6.93812147e-02 -4.74631339e-01
-1.09132552e+00 1.02377927e+00 1.16977561e+00 -4.09923464e-01
3.92247170e-01 -1.80672929e-01 -5.29394811e-03 -4.24475044e-01
-6.00084543e-01 -4.69834179e-01 3.71327698e-01 7.45607972e-01
-9.52867046e-02 -5.98622262e-02 -9.92186666e-02 6.90561891e-01
3.47827017e-01 1.66242406e-01 1.69804931e-01 9.99404073e-01
-5.40174782e-01 -1.04925776e+00 1.58367790e-02 4.07932758e-01
-4.86253388e-02 1.87010437e-01 -5.66375434e-01 9.26757991e-01
3.61226708e-01 7.73077011e-01 1.82890922e-01 -7.18719363e-01
4.98823166e-01 1.86039418e-01 4.07255113e-01 -8.43140364e-01
-2.19611391e-01 -2.55390495e-01 -1.65036634e-01 -1.17485344e+00
-6.97113514e-01 -4.03389066e-01 -1.60484183e+00 7.75254741e-02
-2.00708881e-02 -2.52161443e-01 -2.03379065e-01 1.21456397e+00
5.78695536e-01 7.30081201e-01 4.34354872e-01 -5.80304682e-01
-6.29992068e-01 -9.87470031e-01 -6.42199337e-01 1.04865289e+00
1.12034734e-02 -1.18743491e+00 -2.40569994e-01 2.34636098e-01] | [8.439752578735352, 0.803927481174469] |
c8534324-bdd9-4bc6-bb56-7c61d6013a54 | efficient-solution-of-portfolio-optimization | 2306.12639 | null | https://arxiv.org/abs/2306.12639v1 | https://arxiv.org/pdf/2306.12639v1.pdf | Efficient Solution of Portfolio Optimization Problems via Dimension Reduction and Sparsification | The Markowitz mean-variance portfolio optimization model aims to balance expected return and risk when investing. However, there is a significant limitation when solving large portfolio optimization problems efficiently: the large and dense covariance matrix. Since portfolio performance can be potentially improved by considering a wider range of investments, it is imperative to be able to solve large portfolio optimization problems efficiently, typically in microseconds. We propose dimension reduction and increased sparsity as remedies for the covariance matrix. The size reduction is based on predictions from machine learning techniques and the solution to a linear programming problem. We find that using the efficient frontier from the linear formulation is much better at predicting the assets on the Markowitz efficient frontier, compared to the predictions from neural networks. Reducing the covariance matrix based on these predictions decreases both runtime and total iterations. We also present a technique to sparsify the covariance matrix such that it preserves positive semi-definiteness, which improves runtime per iteration. The methods we discuss all achieved similar portfolio expected risk and return as we would obtain from a full dense covariance matrix but with improved optimizer performance. | ['Hande Y. Benson', 'Cassidy K. Buhler'] | 2023-06-22 | null | null | null | null | ['dimensionality-reduction', 'portfolio-optimization'] | ['methodology', 'time-series'] | [-1.13647938e-01 1.78029984e-01 -2.07771569e-01 -2.44036868e-01
-8.22150171e-01 -8.64688873e-01 2.42997352e-02 -8.47902671e-02
-1.91978097e-01 6.77390099e-01 3.32011342e-01 -9.06718016e-01
-7.24291384e-01 -8.06444943e-01 -5.35772622e-01 -5.06548882e-01
-1.59976661e-01 5.77934384e-01 -3.74062330e-01 9.92082506e-02
6.58894062e-01 5.84155858e-01 -1.02697504e+00 3.36617947e-01
5.58916271e-01 1.46726513e+00 -2.91558027e-01 5.45047104e-01
-9.22824070e-02 9.87499774e-01 -4.33592021e-01 -7.24444389e-01
1.24291456e+00 -2.23168120e-01 -3.24008137e-01 -2.01351821e-01
2.81213552e-01 -4.59878653e-01 -1.08456552e-01 8.92645955e-01
3.01134586e-01 2.09791198e-01 6.30852580e-01 -9.99807596e-01
-3.87631327e-01 8.51962864e-01 -7.72427917e-01 3.14200908e-01
-9.02366117e-02 -5.85263297e-02 1.52010190e+00 -6.18352413e-01
1.54898539e-01 9.49638009e-01 9.46445107e-01 4.89923134e-02
-1.18257499e+00 -5.62104821e-01 1.54083043e-01 -1.76605359e-01
-1.02209949e+00 -2.33666599e-01 4.55208480e-01 -4.91206825e-01
1.27126205e+00 6.74657822e-01 7.32140303e-01 2.37227321e-01
5.07522702e-01 4.92974848e-01 9.00553405e-01 -2.38204673e-01
3.09167147e-01 4.59402204e-01 1.44646456e-02 5.08331776e-01
5.65922320e-01 6.18180513e-01 -3.32334697e-01 -5.07530928e-01
6.50204837e-01 3.48064363e-01 -4.44612987e-02 -4.72333938e-01
-9.55690861e-01 1.07267904e+00 -7.26652965e-02 -1.84043571e-01
-4.17890757e-01 2.91571587e-01 8.47918317e-02 6.73485518e-01
4.72119510e-01 1.07283556e+00 -7.46639788e-01 -3.34401459e-01
-1.18583655e+00 3.52147430e-01 1.53243923e+00 7.09401369e-01
6.11191213e-01 1.83412239e-01 -2.03197867e-01 4.51285809e-01
2.84569025e-01 7.56809592e-01 1.12253815e-01 -1.37042773e+00
8.86493146e-01 4.21015799e-01 2.19806582e-01 -1.16714418e+00
-4.06944811e-01 -9.17193711e-01 -4.45495754e-01 5.41416705e-01
6.61542177e-01 -6.52285874e-01 -3.80315959e-01 1.25316417e+00
-8.67706016e-02 -5.20324968e-02 7.64416754e-02 7.20586121e-01
-3.35636288e-01 5.81095457e-01 -6.82907581e-01 -2.92661577e-01
7.77010143e-01 -1.06352460e+00 -3.44201773e-01 -4.62161779e-01
8.09580445e-01 -8.08531225e-01 7.03687906e-01 3.85034412e-01
-1.11887789e+00 3.22853595e-01 -1.12047040e+00 5.50114989e-01
5.94956353e-02 -1.30552664e-01 9.70570326e-01 1.16082692e+00
-8.09213817e-01 1.10911417e+00 -6.76114678e-01 4.30597603e-01
4.15680557e-01 6.87820196e-01 5.58030270e-02 5.05012751e-01
-7.14105844e-01 9.08112586e-01 2.25779936e-01 1.04180358e-01
-2.59239972e-01 -1.33101559e+00 -4.51536834e-01 6.44128740e-01
5.42241096e-01 -5.53857684e-01 1.09559727e+00 -9.58625317e-01
-1.51229560e+00 1.79343686e-01 2.69137204e-01 -9.47267890e-01
6.36385739e-01 -3.77433449e-01 6.05506264e-02 -2.38755673e-01
-3.11912209e-01 -1.74784809e-01 7.31327593e-01 -4.46836442e-01
-7.88013220e-01 -2.32397139e-01 -1.18737794e-01 1.84944317e-01
-6.76135480e-01 -8.27747211e-02 1.20942764e-01 -7.70529091e-01
-9.06157047e-02 -1.13122642e+00 -6.13830090e-01 -3.30004930e-01
-7.99101517e-02 6.05405867e-01 1.79309651e-01 -7.40691900e-01
1.42956913e+00 -1.71282232e+00 -2.60012120e-01 9.43560481e-01
1.46893859e-01 -8.79837126e-02 2.15454325e-02 3.97122294e-01
-3.55610371e-01 1.03703626e-01 2.03360636e-02 7.50855133e-02
1.95538044e-01 -2.04372227e-01 -8.20569873e-01 4.46683764e-01
-6.31588325e-02 9.12887692e-01 -3.82705539e-01 1.10907100e-01
3.40514481e-02 -1.80142105e-01 -9.10388291e-01 6.99851066e-02
-7.49032870e-02 -1.16000012e-01 -4.44548577e-01 4.89345014e-01
5.23273110e-01 -4.19849306e-01 3.25191379e-01 1.24325134e-01
-5.58159575e-02 2.02935711e-01 -1.63224041e+00 8.79991710e-01
-4.82518971e-01 4.21839118e-01 -1.14666373e-01 -1.07508194e+00
9.75566268e-01 -2.08447818e-02 7.27468610e-01 -4.42217767e-01
6.36137426e-02 2.76115537e-01 7.42253438e-02 1.95435241e-01
3.97454172e-01 -5.33441007e-01 2.64615007e-02 1.16647518e+00
-4.80190575e-01 8.33560899e-02 2.15332448e-01 -5.56629635e-02
1.29498613e+00 -3.40875685e-01 1.12575546e-01 -6.15387976e-01
7.33940303e-02 2.02514529e-02 7.58179903e-01 8.10051203e-01
2.77980149e-01 3.89393687e-01 1.03300786e+00 -6.47342145e-01
-1.04373431e+00 -1.02812052e+00 1.05417117e-01 9.57021117e-01
-5.69695652e-01 -1.81380793e-01 -3.93645525e-01 -5.71694553e-01
7.27727652e-01 7.25177109e-01 -5.91329098e-01 6.47694394e-02
-4.97171223e-01 -1.10857189e+00 1.05276980e-01 6.70144916e-01
9.59110707e-02 -3.47379565e-01 -4.81134385e-01 3.68593782e-01
6.60157323e-01 -4.71848786e-01 -8.01615477e-01 3.56922477e-01
-9.93638694e-01 -9.83203650e-01 -8.45729649e-01 -6.60041198e-02
6.44485950e-01 4.17283848e-02 1.10697496e+00 -1.66320965e-01
-2.22477943e-01 4.09105867e-01 1.26123250e-01 -7.31144547e-01
-2.94091124e-02 -1.30925655e-01 1.05967157e-01 -1.96482204e-02
1.22278363e-01 -5.85529804e-01 -5.76679587e-01 2.17736825e-01
-3.02925617e-01 -2.52997011e-01 7.34100640e-01 9.69930708e-01
5.59873104e-01 2.60811210e-01 4.39004213e-01 -1.10194099e+00
8.63955617e-01 -3.99350435e-01 -1.28910410e+00 5.98100960e-01
-1.44397819e+00 7.00185955e-01 4.62584138e-01 -4.04728204e-01
-1.09369314e+00 -4.07202132e-02 3.51280391e-01 -4.42680359e-01
9.59490359e-01 6.40000403e-01 2.96008140e-01 -2.36894697e-01
4.37329412e-01 -2.99285680e-01 1.70141891e-01 -4.62037742e-01
2.55714655e-01 2.54679203e-01 -1.34808734e-01 -5.40549576e-01
8.39809299e-01 2.67944217e-01 4.60933059e-01 -1.00623749e-01
-1.12376344e+00 -2.46364266e-01 -2.14075044e-01 1.19280450e-01
2.49565259e-01 -6.60406172e-01 -9.20097828e-01 -9.11552310e-02
-4.96573031e-01 -3.01027089e-01 -6.93335176e-01 8.23723853e-01
-5.85056126e-01 3.12647782e-02 -5.19096732e-01 -9.70481753e-01
-6.91567421e-01 -7.61951268e-01 4.35722351e-01 4.12155464e-02
-3.42902780e-01 -1.32929933e+00 3.04839522e-01 2.72972047e-01
6.19959116e-01 1.28893763e-01 8.40934813e-01 -1.00098228e+00
-7.31662154e-01 -4.50127155e-01 -2.82987416e-01 4.49027091e-01
2.29048971e-02 -3.34113300e-01 -3.65107149e-01 -3.82459521e-01
4.63749886e-01 1.04079023e-01 8.20194483e-01 4.88301277e-01
8.87296855e-01 -7.24242747e-01 3.27643082e-02 8.84876847e-01
1.40190101e+00 2.76073396e-01 2.41387427e-01 5.79065681e-01
3.64890039e-01 6.94447458e-01 7.77491152e-01 9.80606437e-01
-1.55488640e-01 1.10898219e-01 -1.11740619e-01 1.03499942e-01
5.82366109e-01 -8.97255763e-02 4.15027946e-01 5.74488819e-01
4.47882824e-02 1.63370669e-01 -1.10081553e+00 1.69327036e-01
-2.06134844e+00 -1.20177543e+00 2.52247363e-01 2.45503402e+00
6.26346588e-01 3.84262860e-01 1.40423685e-01 -2.38919616e-01
2.32878596e-01 1.10962257e-01 -8.26262891e-01 -7.26326644e-01
6.53799176e-02 3.48176867e-01 1.40672088e+00 7.20831454e-01
-8.48264754e-01 2.78475493e-01 7.81380844e+00 6.89739048e-01
-9.31501210e-01 -2.46064082e-01 9.18718934e-01 -1.00802922e+00
-6.05078936e-01 2.01265812e-01 -1.13424885e+00 4.05434102e-01
1.22280538e+00 -8.07525694e-01 7.27428794e-01 1.28570187e+00
-9.74469446e-03 5.50142266e-02 -1.25848889e+00 8.73521805e-01
-2.81708948e-02 -1.81341565e+00 -3.31964880e-01 4.80711550e-01
1.01676369e+00 -2.45337978e-01 5.38233578e-01 3.35265607e-01
5.22434831e-01 -1.11324596e+00 5.98124206e-01 8.55736494e-01
1.83787823e-01 -1.21455204e+00 8.87054026e-01 2.70200998e-01
-9.66559649e-01 -5.48585176e-01 -6.74073458e-01 -3.15066725e-01
6.15645722e-02 1.00879788e+00 -8.24873686e-01 6.96934089e-02
4.62582946e-01 5.06969631e-01 -3.45678151e-01 9.21304286e-01
1.98454306e-01 4.48975593e-01 -7.31310368e-01 -1.79852784e-01
1.78694814e-01 -7.77234912e-01 3.77323836e-01 9.09799874e-01
5.89084208e-01 1.14047248e-03 -1.05133057e-01 8.89270425e-01
2.00078428e-01 1.58315882e-01 -5.06658792e-01 -4.34555382e-01
1.70727521e-01 9.88909483e-01 -4.58820671e-01 -1.22306533e-01
-6.54843807e-01 4.01464731e-01 2.50307739e-01 1.67581782e-01
-6.60732508e-01 -4.75509793e-01 7.60333419e-01 2.12864444e-01
4.38002825e-01 -2.07209393e-01 -7.82124102e-01 -1.01177907e+00
2.62869567e-01 -9.22856092e-01 4.76596326e-01 -1.01737335e-01
-1.22748947e+00 1.85775951e-01 -1.28270581e-01 -1.08736718e+00
-5.31586528e-01 -7.37080038e-01 -5.96767187e-01 9.88357425e-01
-1.11538208e+00 -3.65593284e-01 5.25553048e-01 9.24535021e-02
-4.76525836e-02 -7.24485755e-01 4.99167740e-01 7.59056807e-02
-5.94784439e-01 8.75800788e-01 8.07003021e-01 -1.45165980e-01
4.90082204e-01 -1.44055414e+00 3.18718612e-01 6.78366721e-01
2.25400478e-01 9.11080003e-01 6.91368282e-01 -8.53811920e-01
-1.66627169e+00 -8.60761642e-01 5.22836745e-01 -5.04356146e-01
1.16491175e+00 4.64475304e-02 -5.37613153e-01 8.93534839e-01
3.10629662e-02 -4.23372746e-01 1.06196499e+00 6.24883294e-01
-1.44197583e-01 -5.02042174e-01 -9.57617998e-01 5.94124317e-01
5.65506041e-01 -3.98018330e-01 -4.46189523e-01 5.05943239e-01
4.09954399e-01 -2.98338801e-01 -1.21510518e+00 1.62275240e-01
8.05837095e-01 -9.74478960e-01 1.08162785e+00 -6.93945348e-01
2.45806463e-02 2.22134247e-01 -1.11871853e-01 -1.05036092e+00
-3.62463176e-01 -1.09351969e+00 -2.98560292e-01 9.16639328e-01
1.03997779e+00 -9.63995993e-01 1.40651298e+00 1.54554653e+00
4.02259439e-01 -1.15041649e+00 -5.47997594e-01 -1.11588168e+00
2.53952354e-01 -3.91908437e-01 7.81549394e-01 5.46459079e-01
9.74035859e-02 5.19358478e-02 -6.01039648e-01 -4.94686700e-02
7.97337294e-01 6.34280324e-01 6.04620278e-01 -9.15991008e-01
-9.44914281e-01 -5.98914444e-01 -1.05127497e-02 -9.37355578e-01
-1.97858363e-02 -9.49509442e-01 -3.73706311e-01 -1.02416575e+00
2.54270881e-01 -6.43470943e-01 -6.15543783e-01 4.18294221e-01
-5.47726192e-02 -2.67783999e-01 4.28915739e-01 1.65755168e-01
-4.05238181e-01 3.00908417e-01 7.76532233e-01 3.35662700e-02
-4.93526578e-01 4.16967988e-01 -1.16651869e+00 8.28884840e-01
9.59463418e-01 -7.25911081e-01 -4.35750306e-01 -2.57914782e-01
7.29778111e-01 3.86521608e-01 -2.68859118e-01 -7.04070151e-01
3.14224452e-01 -4.96318638e-01 4.67341244e-01 -6.06760800e-01
2.09291905e-01 -5.78271627e-01 4.76422280e-01 5.81414878e-01
-3.99213135e-01 2.59941965e-01 8.77783448e-02 5.58859646e-01
9.18987468e-02 -5.80217838e-01 6.67048693e-01 -1.21656992e-01
-5.97337112e-02 2.33532712e-01 -1.88249350e-01 1.41426191e-01
1.09050715e+00 5.10787638e-03 -1.12545326e-01 -5.49213648e-01
-7.66410291e-01 3.77030581e-01 3.45973700e-01 3.13045792e-02
4.95381355e-01 -1.29852915e+00 -5.47715008e-01 1.94219634e-01
-5.92360675e-01 -6.50666416e-01 -5.32823168e-02 7.98609316e-01
-7.12750912e-01 8.09508860e-01 -3.92402783e-02 1.77100170e-02
-1.10986423e+00 5.55707633e-01 4.11970794e-01 -9.81017709e-01
-4.18901622e-01 1.08400965e+00 1.49091244e-01 -1.33705124e-01
2.70518601e-01 -1.74714535e-01 1.54526338e-01 1.20375030e-01
8.62973869e-01 7.29826033e-01 1.67192239e-02 2.52152264e-01
-1.57718718e-01 5.59779525e-01 -3.07667881e-01 -4.11115646e-01
1.62952256e+00 5.45870215e-02 -2.14701593e-01 8.11599195e-02
1.19123340e+00 3.45397800e-01 -1.48760343e+00 -2.90471930e-02
4.81273890e-01 -8.81219089e-01 2.18979508e-01 -6.23213947e-01
-1.37314022e+00 6.43686593e-01 4.30458009e-01 1.71983883e-01
9.09952879e-01 -5.00929296e-01 7.87001371e-01 1.04968274e+00
3.62102181e-01 -1.37477779e+00 -7.38247409e-02 4.57329035e-01
8.69697809e-01 -7.91186094e-01 6.78501129e-01 -1.29199512e-02
-9.17055190e-01 1.11575329e+00 1.09119348e-01 -5.96290290e-01
1.20360172e+00 6.44898713e-01 -1.39443293e-01 2.14852821e-02
-9.56898928e-01 2.86351383e-01 2.99111456e-01 6.17007911e-02
2.48106234e-02 5.17635085e-02 -1.18864790e-01 1.00579417e+00
-6.57788396e-01 -4.38833058e-01 4.22954410e-01 8.92002702e-01
-5.40853500e-01 -1.39462674e+00 -4.97100323e-01 1.24155593e+00
-9.00966585e-01 -3.59011590e-01 -2.52026021e-01 4.53528494e-01
-5.47371686e-01 8.69118333e-01 2.00114936e-01 -2.94144541e-01
3.14696640e-01 -2.37816349e-01 4.44900155e-01 -6.34177148e-01
-7.62327671e-01 4.89251725e-02 2.68001378e-01 -8.28243077e-01
3.35818231e-01 -9.46018934e-01 -9.12197411e-01 -7.00036943e-01
-3.98881048e-01 3.44595611e-01 4.60839361e-01 7.73548484e-01
4.00289267e-01 2.78951049e-01 9.42241549e-01 -3.81452352e-01
-1.49126410e+00 -3.07318211e-01 -1.07958603e+00 -1.31854579e-01
5.74236587e-02 -3.94225150e-01 -7.02815294e-01 -3.53422105e-01] | [4.990437030792236, 3.9385507106781006] |
73686482-cf8d-4d79-b27c-28b1c5f0b183 | unsupervised-part-based-disentangling-of | 1903.06946 | null | https://arxiv.org/abs/1903.06946v3 | https://arxiv.org/pdf/1903.06946v3.pdf | Unsupervised Part-Based Disentangling of Object Shape and Appearance | Large intra-class variation is the result of changes in multiple object characteristics. Images, however, only show the superposition of different variable factors such as appearance or shape. Therefore, learning to disentangle and represent these different characteristics poses a great challenge, especially in the unsupervised case. Moreover, large object articulation calls for a flexible part-based model. We present an unsupervised approach for disentangling appearance and shape by learning parts consistently over all instances of a category. Our model for learning an object representation is trained by simultaneously exploiting invariance and equivariance constraints between synthetically transformed images. Since no part annotation or prior information on an object class is required, the approach is applicable to arbitrary classes. We evaluate our approach on a wide range of object categories and diverse tasks including pose prediction, disentangled image synthesis, and video-to-video translation. The approach outperforms the state-of-the-art on unsupervised keypoint prediction and compares favorably even against supervised approaches on the task of shape and appearance transfer. | ['Björn Ommer', 'Leonard Bereska', 'Timo Milbich', 'Dominik Lorenz'] | 2019-03-16 | unsupervised-part-based-disentangling-of-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Lorenz_Unsupervised_Part-Based_Disentangling_of_Object_Shape_and_Appearance_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Lorenz_Unsupervised_Part-Based_Disentangling_of_Object_Shape_and_Appearance_CVPR_2019_paper.pdf | cvpr-2019-6 | ['unsupervised-facial-landmark-detection'] | ['computer-vision'] | [ 6.35945201e-01 1.46108091e-01 -2.17710763e-01 -4.36035484e-01
-8.14329743e-01 -9.69198167e-01 9.44572866e-01 -6.00303411e-02
-4.23380621e-02 4.83700246e-01 1.35354385e-01 3.75410229e-01
-1.32881179e-02 -3.21056873e-01 -9.29405034e-01 -8.61056507e-01
3.62499684e-01 9.50971365e-01 1.65951550e-01 -5.22434302e-02
1.06562739e-02 7.86654711e-01 -1.65317285e+00 4.08500731e-01
4.47737992e-01 8.58632684e-01 1.07685424e-01 5.85660398e-01
9.86956656e-02 3.76645267e-01 -2.78216362e-01 -4.84943986e-01
5.49443483e-01 -4.54664052e-01 -5.24742186e-01 7.30384707e-01
1.22775626e+00 -2.81409025e-02 -1.95550188e-01 1.04943466e+00
3.19911301e-01 2.68219616e-02 1.08682191e+00 -1.27774000e+00
-6.15419328e-01 1.01366162e-01 -4.10911083e-01 -3.71015102e-01
3.89666557e-01 5.78118674e-02 1.17411888e+00 -9.99145091e-01
9.23031986e-01 1.28085840e+00 3.57173234e-01 4.49071348e-01
-2.01789927e+00 -3.50684553e-01 1.35144264e-01 3.20673771e-02
-1.24225855e+00 -7.40524411e-01 1.10063708e+00 -7.39191830e-01
4.04126108e-01 4.73376364e-01 6.36720717e-01 1.14682639e+00
-9.61406380e-02 9.37469065e-01 1.30694902e+00 -4.35677052e-01
-2.90477145e-02 3.76958460e-01 -5.34406662e-01 6.50518298e-01
2.59435922e-01 7.16534853e-02 -4.84264135e-01 -4.89469618e-02
8.25544119e-01 -4.12509926e-02 -1.58685729e-01 -1.28919494e+00
-1.51518011e+00 5.29174984e-01 2.80111194e-01 -6.05540015e-02
-1.30800650e-01 1.33319706e-01 7.78346658e-02 2.97943264e-01
3.50421697e-01 6.84980750e-01 -7.11343229e-01 1.49619147e-01
-9.31458056e-01 3.76867026e-01 7.49443412e-01 1.12730384e+00
7.59060681e-01 1.21637277e-01 -4.67654178e-03 7.60948241e-01
1.80470929e-01 7.53396213e-01 1.87993780e-01 -1.14625657e+00
1.61525205e-01 4.33455348e-01 4.55583781e-02 -1.04809308e+00
-1.15116239e-01 -3.26155066e-01 -7.31291413e-01 5.21230459e-01
6.88272357e-01 3.45869124e-01 -1.02641690e+00 1.89610958e+00
2.67842054e-01 -4.12715636e-02 -1.91674694e-01 7.64389396e-01
6.92289770e-01 1.80049524e-01 -1.56344175e-01 -2.03109831e-01
1.23085904e+00 -7.74392247e-01 -4.57785636e-01 -2.15596795e-01
8.48804265e-02 -1.10320807e+00 7.66229093e-01 3.89229268e-01
-1.13536429e+00 -6.21664941e-01 -8.13994408e-01 -8.32911506e-02
-2.28696927e-01 2.78772801e-01 5.10664165e-01 3.87637585e-01
-6.77938581e-01 6.23436749e-01 -7.57981896e-01 -3.11603993e-01
3.54013681e-01 5.39184332e-01 -7.40385354e-01 -9.36849266e-02
-4.39965665e-01 8.47166300e-01 2.40661517e-01 -2.64520347e-01
-7.00100243e-01 -6.45758450e-01 -9.92926121e-01 -1.71974674e-01
2.74115533e-01 -8.82666111e-01 9.28627074e-01 -1.39355350e+00
-1.52201557e+00 1.26312840e+00 1.91576108e-02 8.12074989e-02
7.02565432e-01 7.04894736e-02 -9.78273377e-02 1.31928623e-02
-1.61256760e-01 1.00002003e+00 1.41530824e+00 -1.63369858e+00
-9.99966860e-02 -3.61432731e-01 6.85060211e-03 2.94539601e-01
-3.40015516e-02 -1.07733473e-01 -4.12851602e-01 -8.16936851e-01
3.81701767e-01 -1.41506708e+00 -7.79691115e-02 5.65735519e-01
-3.23501706e-01 2.35132411e-01 7.79992819e-01 -5.08364081e-01
4.51408565e-01 -2.13905692e+00 7.79552758e-01 -3.94669129e-04
1.12330876e-01 -5.42007759e-02 -3.28377843e-01 2.36897752e-01
-3.03176999e-01 -1.90326288e-01 -2.51296312e-01 -4.81720269e-01
8.65277201e-02 3.15121204e-01 -2.57055223e-01 6.72345161e-01
3.12960684e-01 8.90945494e-01 -8.40765893e-01 -5.13259470e-01
3.05364072e-01 3.73384744e-01 -6.82781219e-01 2.74307489e-01
-3.18040788e-01 9.17303324e-01 -2.50545323e-01 6.19926214e-01
5.39061129e-01 -3.07032943e-01 1.01225533e-01 -5.52167654e-01
3.13848823e-01 -8.57387185e-02 -1.18759191e+00 1.98080695e+00
-3.79908353e-01 7.29791224e-01 -6.25643358e-02 -9.60901737e-01
7.66340792e-01 3.38315398e-01 5.47302961e-01 -2.58754611e-01
5.61774559e-02 3.63327384e-01 3.09827209e-01 -4.50322926e-01
1.84713736e-01 -3.92751366e-01 8.35334882e-02 3.16789776e-01
2.99004048e-01 -6.61788285e-01 -1.08709417e-01 2.58514620e-02
6.60723865e-01 5.65101743e-01 3.33777130e-01 -1.88727081e-01
1.99054733e-01 -2.95155793e-01 3.63327950e-01 2.13733509e-01
2.84307133e-02 1.09830749e+00 2.84274727e-01 -3.40251207e-01
-1.26194775e+00 -1.26688492e+00 -3.07592332e-01 9.28696394e-01
1.65409580e-01 -6.89919144e-02 -4.14184600e-01 -7.43248880e-01
2.45181426e-01 5.13424814e-01 -7.93269753e-01 -3.63967344e-02
-5.02491951e-01 -4.14722621e-01 1.36626825e-01 3.76934618e-01
-1.36126205e-01 -1.01830578e+00 -2.95487106e-01 -9.63616297e-02
-1.44261122e-01 -1.37849641e+00 -7.02996612e-01 -1.15259938e-01
-8.34117770e-01 -1.06427479e+00 -7.03006923e-01 -6.38471961e-01
1.02612174e+00 1.85477719e-01 1.15176094e+00 -2.92654037e-01
-5.94747663e-01 6.65475130e-01 -8.54660198e-02 -3.13062131e-01
-5.72316110e-01 -1.62676036e-01 1.54364675e-01 5.66865027e-01
-1.98984623e-01 -8.63750815e-01 -5.40875435e-01 4.13980752e-01
-8.16293001e-01 3.47967863e-01 8.08097601e-01 7.45531738e-01
7.83959448e-01 -4.95098919e-01 1.32860675e-01 -9.30434644e-01
-1.60019249e-02 -4.78130206e-02 -5.21764755e-01 4.28471595e-01
-3.05409193e-01 3.83975476e-01 3.99424076e-01 -8.91715646e-01
-9.09470737e-01 6.06764555e-01 4.23374712e-01 -8.08280826e-01
-3.55784684e-01 -3.92822474e-02 -5.70211768e-01 -2.07157165e-01
6.01522744e-01 2.99852163e-01 2.79281676e-01 -5.94742715e-01
6.58103764e-01 1.31510705e-01 4.92550552e-01 -6.82301700e-01
1.10509217e+00 6.37821198e-01 4.30010885e-01 -8.92379820e-01
-6.44795597e-01 -3.68360132e-01 -1.18894434e+00 -1.00951403e-01
9.69431937e-01 -8.93790305e-01 -3.18272859e-01 3.62050235e-01
-1.26538277e+00 -9.37265977e-02 -5.19290447e-01 4.31253105e-01
-9.74538624e-01 2.96347499e-01 -1.80035174e-01 -3.08929890e-01
1.42381936e-01 -1.31772101e+00 1.46984196e+00 -1.81547895e-01
-2.93237358e-01 -1.01172042e+00 2.21040882e-02 3.51686567e-01
1.85820684e-01 5.04647970e-01 9.12055135e-01 -6.03024185e-01
-9.29837704e-01 -3.61271948e-01 -1.00635134e-01 2.55982667e-01
3.81083876e-01 1.60867780e-01 -9.34412181e-01 -4.89448071e-01
-2.38866508e-01 -5.52473783e-01 7.25788236e-01 1.12864949e-01
1.09765553e+00 -3.94881785e-01 -1.21766113e-01 6.13706589e-01
1.24509001e+00 -3.18008691e-01 4.07338560e-01 -1.29678756e-01
1.10437322e+00 9.24480617e-01 3.35374534e-01 1.38828009e-01
1.16760649e-01 1.32675898e+00 4.41139191e-01 -9.23477113e-02
-4.64905232e-01 -2.86442608e-01 2.90726304e-01 8.50358188e-01
-3.53895873e-01 -1.29305245e-02 -5.42629421e-01 2.70422071e-01
-1.61133397e+00 -9.48520780e-01 -5.95676079e-02 2.32542515e+00
7.97287345e-01 -1.29313618e-01 1.30804360e-01 -1.46503851e-01
4.95668471e-01 1.05006486e-01 -5.73853433e-01 1.51255829e-02
-1.72549009e-01 7.55087882e-02 4.10875738e-01 4.17846084e-01
-1.05491126e+00 8.35622907e-01 6.00090361e+00 7.53441036e-01
-1.12618446e+00 -4.70784046e-02 3.48971963e-01 9.06496122e-02
-5.16447961e-01 1.15571536e-01 -4.24606919e-01 3.63473073e-02
2.53318369e-01 -2.23864596e-02 4.70380723e-01 6.75248206e-01
-1.96831480e-01 2.60097504e-01 -1.60338879e+00 1.09254408e+00
4.23317552e-01 -1.04035342e+00 2.63257593e-01 1.28670037e-01
1.00374997e+00 -2.19724238e-01 3.26312453e-01 -1.27934501e-01
5.01585640e-02 -8.82783651e-01 1.11374974e+00 6.98656261e-01
1.06055617e+00 -3.35050195e-01 1.21750802e-01 2.34734669e-01
-9.43944812e-01 3.30653638e-01 -2.32723624e-01 3.47499132e-01
-1.43735232e-02 3.79329622e-02 -6.26930475e-01 3.92095655e-01
1.82574540e-01 7.60035336e-01 -6.56825840e-01 8.84348929e-01
-2.46158257e-01 1.78329125e-01 -2.11207017e-01 4.27271992e-01
-2.95503795e-01 -2.67373949e-01 8.58726740e-01 1.04713869e+00
1.05197325e-01 -2.06912234e-02 3.18395138e-01 8.97433639e-01
-2.86092088e-02 6.66108504e-02 -6.47559881e-01 1.06942318e-02
1.14296086e-01 1.29061139e+00 -7.35219181e-01 -2.79957473e-01
-4.23963696e-01 1.13353610e+00 2.06184536e-01 3.30651909e-01
-5.83414495e-01 1.19676538e-01 7.23994136e-01 2.98876047e-01
4.97481942e-01 -3.35639447e-01 -3.06629062e-01 -1.47537434e+00
-1.57773029e-02 -1.01988947e+00 -5.30756898e-02 -8.69559705e-01
-1.25985301e+00 5.02076864e-01 2.79392898e-01 -1.65825999e+00
-3.41362953e-01 -7.93998897e-01 -2.55102575e-01 6.28525972e-01
-1.10291636e+00 -1.67420793e+00 -2.17827022e-01 5.09739280e-01
7.36394823e-01 -1.93065017e-01 1.05506504e+00 1.11544937e-01
-3.14508751e-02 5.90923429e-01 1.44668341e-01 -6.52951226e-02
1.02491891e+00 -1.27477789e+00 1.86565712e-01 5.58102727e-01
7.40599990e-01 4.52001631e-01 9.74286079e-01 -3.25966775e-01
-1.67942977e+00 -9.69208479e-01 6.17840171e-01 -9.84637260e-01
5.50553024e-01 -6.60729766e-01 -7.56671488e-01 6.62681639e-01
-1.31178331e-02 5.35380960e-01 6.01111889e-01 3.33079626e-03
-8.61859977e-01 1.47245759e-02 -9.47877169e-01 6.08377039e-01
1.11835527e+00 -7.10714459e-01 -4.35934544e-01 4.33928400e-01
5.90528667e-01 -4.77647722e-01 -7.43314981e-01 4.50494826e-01
9.30602491e-01 -8.82490695e-01 1.18104136e+00 -7.79234290e-01
5.79012692e-01 -2.59312809e-01 -5.36884785e-01 -1.28213727e+00
-3.77241433e-01 -5.00729978e-01 -2.02664420e-01 1.03104198e+00
4.78458434e-01 -2.99933612e-01 7.22208798e-01 7.35406995e-01
7.36332834e-02 -5.68956614e-01 -6.84988141e-01 -8.54935527e-01
-6.60264418e-02 -1.23911060e-01 3.28805476e-01 1.12789214e+00
-4.91251320e-01 4.75507975e-01 -5.36188185e-01 6.80792779e-02
8.46182585e-01 5.08545876e-01 1.07233477e+00 -1.31021106e+00
-7.84077585e-01 -4.80856717e-01 -7.90913165e-01 -8.96771252e-01
2.91388333e-01 -1.10259748e+00 -2.66493652e-02 -1.02896190e+00
4.66055989e-01 -3.98528248e-01 -6.25803694e-02 4.66073364e-01
8.09803158e-02 5.44418454e-01 5.11800289e-01 5.08606195e-01
-4.70652461e-01 6.10185981e-01 1.66090810e+00 -2.96280086e-01
2.95407437e-02 2.05803365e-01 -4.51395482e-01 8.58383536e-01
4.48885113e-01 -4.45051014e-01 -4.53319013e-01 -4.46754634e-01
1.08351015e-01 1.60088807e-01 5.84313571e-01 -6.53587580e-01
-9.84389633e-02 -3.93080324e-01 4.22599971e-01 -2.32413322e-01
6.63906395e-01 -1.08651054e+00 4.73941863e-01 1.44469112e-01
-4.48698193e-01 -5.58317900e-02 1.22696713e-01 8.28990459e-01
-1.10493146e-01 6.90795928e-02 9.18453217e-01 -1.16541564e-01
-4.18872625e-01 5.16687274e-01 2.00343281e-01 1.10895701e-01
8.02733183e-01 -4.74150687e-01 1.81165710e-01 -3.86836976e-01
-1.12022877e+00 -4.17290717e-01 8.63298655e-01 7.24261105e-01
5.18250644e-01 -1.54019773e+00 -8.38693023e-01 4.67216313e-01
4.71197307e-01 -4.21496592e-02 -1.02041978e-02 8.76138270e-01
-9.61948410e-02 1.13705195e-01 -3.48841876e-01 -8.36812198e-01
-1.59088469e+00 6.08881772e-01 2.30006188e-01 2.58879140e-02
-4.40492630e-01 7.76707470e-01 7.44817853e-01 -5.75247288e-01
-2.89626751e-05 -1.36092782e-01 9.07480344e-02 1.01367176e-01
7.00412020e-02 1.28598645e-01 -7.45791197e-02 -1.21481740e+00
-1.85631499e-01 9.65960324e-01 6.61603548e-03 -1.95863828e-01
1.27384710e+00 5.86343668e-02 -1.31170124e-01 6.74420059e-01
1.45370412e+00 2.07110897e-01 -1.47485435e+00 -3.90145034e-01
-3.32541794e-01 -7.56117940e-01 -3.84816825e-01 -7.43449807e-01
-9.43987131e-01 9.09369111e-01 5.31058609e-01 -1.96489125e-01
8.80896449e-01 2.71830380e-01 1.52797073e-01 2.95288116e-01
5.23054004e-01 -6.44062281e-01 4.26133305e-01 2.98911035e-01
1.38880193e+00 -1.46426821e+00 2.64662266e-01 -6.33209586e-01
-6.35154665e-01 1.03611124e+00 4.22663838e-01 -2.21776366e-01
6.52978122e-01 -7.61891082e-02 -1.08045153e-01 -4.92527261e-02
-6.11614048e-01 -3.79861370e-02 1.05672634e+00 6.07714832e-01
4.46706653e-01 2.96701103e-01 1.35684922e-01 -1.34623386e-02
-2.36705467e-01 -5.41873217e-01 2.33500391e-01 6.38912320e-01
-1.14348441e-01 -1.45888352e+00 -2.84081876e-01 3.31753343e-01
-3.69330883e-01 1.07021950e-01 -6.17130816e-01 8.10472429e-01
2.10219741e-01 2.57750779e-01 1.40778245e-02 -2.51230299e-01
2.30958432e-01 4.64081243e-02 1.22125363e+00 -8.33887517e-01
6.45677932e-03 3.10780883e-01 -9.23652127e-02 -4.44505930e-01
-7.18665659e-01 -1.12899971e+00 -7.64365613e-01 9.07349810e-02
-3.69192153e-01 -4.06681746e-01 6.71876490e-01 8.66851091e-01
1.45710841e-01 2.72870570e-01 5.54202318e-01 -1.32051098e+00
-5.41662097e-01 -7.07937419e-01 -5.67209125e-01 9.50441301e-01
4.40165728e-01 -9.29419935e-01 -3.28617334e-01 5.56131899e-01] | [8.55988883972168, -2.853766679763794] |
c2a7ddb1-9103-491e-9a24-37410e6e8d76 | l-hydra-multi-head-physics-informed-neural | 2301.02152 | null | https://arxiv.org/abs/2301.02152v1 | https://arxiv.org/pdf/2301.02152v1.pdf | L-HYDRA: Multi-Head Physics-Informed Neural Networks | We introduce multi-head neural networks (MH-NNs) to physics-informed machine learning, which is a type of neural networks (NNs) with all nonlinear hidden layers as the body and multiple linear output layers as multi-head. Hence, we construct multi-head physics-informed neural networks (MH-PINNs) as a potent tool for multi-task learning (MTL), generative modeling, and few-shot learning for diverse problems in scientific machine learning (SciML). MH-PINNs connect multiple functions/tasks via a shared body as the basis functions as well as a shared distribution for the head. The former is accomplished by solving multiple tasks with MH-PINNs with each head independently corresponding to each task, while the latter by employing normalizing flows (NFs) for density estimate and generative modeling. To this end, our method is a two-stage method, and both stages can be tackled with standard deep learning tools of NNs, enabling easy implementation in practice. MH-PINNs can be used for various purposes, such as approximating stochastic processes, solving multiple tasks synergistically, providing informative prior knowledge for downstream few-shot learning tasks such as meta-learning and transfer learning, learning representative basis functions, and uncertainty quantification. We demonstrate the effectiveness of MH-PINNs in five benchmarks, investigating also the possibility of synergistic learning in regression analysis. We name the open-source code "Lernaean Hydra" (L-HYDRA), since this mythical creature possessed many heads for performing important multiple tasks, as in the proposed method. | ['George Em Karniadakis', 'Zongren Zou'] | 2023-01-05 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [-2.10015029e-01 4.70794030e-02 9.55613032e-02 -2.30673775e-01
-1.01876497e+00 -2.57368624e-01 8.98567736e-01 -2.06161693e-01
-2.93300122e-01 9.54291046e-01 -2.45538782e-02 -3.63447964e-01
-4.57685918e-01 -9.60814238e-01 -1.15447581e+00 -1.10266757e+00
4.44169194e-02 9.66121912e-01 1.18805155e-01 3.04543711e-02
2.79368479e-02 5.21061242e-01 -1.34762776e+00 1.46551589e-02
8.51988256e-01 9.40257668e-01 3.10183913e-01 7.51800179e-01
-5.20856082e-01 1.01047266e+00 -4.44299847e-01 -6.44951642e-01
-1.55675307e-01 -2.08098382e-01 -5.44791758e-01 -6.90480649e-01
7.10110134e-03 4.68605384e-02 -1.47052139e-01 9.63651001e-01
6.93791926e-01 7.58925915e-01 1.23354363e+00 -1.28500593e+00
-6.59443855e-01 1.01304126e+00 -6.80164039e-01 5.21605164e-02
-4.54201370e-01 3.99215549e-01 7.04666138e-01 -9.28668380e-01
3.87335151e-01 1.60042071e+00 6.95795357e-01 5.88894606e-01
-1.25056279e+00 -7.39709914e-01 -2.12998331e-01 7.38519207e-02
-1.16761208e+00 -2.21919507e-01 5.46865702e-01 -7.51891792e-01
6.32918775e-01 -5.86237982e-02 3.30457449e-01 1.72564149e+00
5.16131759e-01 7.34249294e-01 9.36795533e-01 1.93227325e-02
7.66931832e-01 -9.32054147e-02 4.22653437e-01 5.80782473e-01
5.51355667e-02 1.52398244e-01 -6.59935594e-01 -3.34039241e-01
9.00637031e-01 8.03790316e-02 -6.90365657e-02 -1.47237275e-02
-1.01005483e+00 9.54278648e-01 3.96616876e-01 8.61119926e-02
-3.04245651e-01 6.17349803e-01 4.28517461e-01 -1.93627298e-01
7.79376984e-01 3.04455847e-01 -4.05626684e-01 -1.07612289e-01
-1.07925832e+00 5.39803386e-01 9.45401728e-01 8.51620615e-01
9.21129704e-01 3.18988323e-01 -7.74806559e-01 8.34863067e-01
2.69380271e-01 4.35728937e-01 5.36413670e-01 -1.11296439e+00
1.51264608e-01 2.43297383e-01 -8.63350481e-02 -1.91760495e-01
-5.20341158e-01 -6.52931809e-01 -1.45845139e+00 2.47321919e-01
4.45121497e-01 -4.55884010e-01 -1.07903111e+00 1.90206575e+00
3.74470204e-01 6.05290830e-01 -3.68364379e-02 5.14873683e-01
8.91910970e-01 1.04677057e+00 3.62486631e-01 5.46661839e-02
1.36883938e+00 -9.62120950e-01 -2.84382612e-01 -3.96661125e-02
4.28352714e-01 -3.04848433e-01 1.08288646e+00 3.38505924e-01
-1.04406214e+00 -7.05083549e-01 -6.17038071e-01 -1.84509277e-01
-5.22552967e-01 -2.16925666e-01 6.42552972e-01 2.66818374e-01
-9.04015899e-01 1.07351410e+00 -9.01252270e-01 4.72125709e-02
6.29878581e-01 1.39670461e-01 4.45506163e-02 5.22254519e-02
-1.23317182e+00 7.48838305e-01 5.03751576e-01 6.66640177e-02
-1.47584701e+00 -1.36283183e+00 -8.52390885e-01 5.11596382e-01
9.74586233e-02 -1.19148624e+00 1.14847004e+00 -1.34961531e-01
-1.80908394e+00 3.82651091e-01 -3.55246104e-02 -5.36811888e-01
6.14017069e-01 -2.71860152e-01 -1.88563801e-02 -4.32224125e-01
-6.21121451e-02 5.83396673e-01 9.43329334e-01 -1.11818230e+00
-2.13216498e-01 -3.91150177e-01 -3.25817674e-01 -1.90377265e-01
-1.06515244e-01 -1.36305377e-01 -3.68071556e-01 -6.45248771e-01
-4.92440194e-01 -6.90524876e-01 -3.19670647e-01 -1.23737641e-01
-6.41675115e-01 -4.95659381e-01 7.29947150e-01 -6.03114128e-01
8.07139814e-01 -1.97367299e+00 6.08794749e-01 1.31567484e-02
3.65584224e-01 2.33997762e-01 -1.08880565e-01 1.59878567e-01
8.60762894e-02 2.84364298e-02 -5.98235548e-01 -8.63477707e-01
3.54236156e-01 3.59162003e-01 -1.81892768e-01 2.59558886e-01
2.70385090e-02 1.35943687e+00 -8.66014659e-01 -3.55803549e-01
3.58799726e-01 6.89655781e-01 -2.00435203e-02 3.65487903e-01
-5.86414874e-01 6.69133902e-01 -3.06369960e-01 6.86219692e-01
5.64711511e-01 -3.33175838e-01 -4.85346854e-01 -8.22330788e-02
2.44969539e-02 -2.58688390e-01 -1.01805592e+00 1.76649702e+00
-8.50603044e-01 2.13192046e-01 8.77124369e-02 -8.56336772e-01
9.60655868e-01 1.13739744e-01 2.92797327e-01 -3.12743813e-01
2.43646756e-01 2.45598614e-01 -7.35058263e-02 -2.27218628e-01
1.39582336e-01 -3.16396952e-01 -6.21973537e-03 3.14974755e-01
6.91866279e-01 -1.77566081e-01 2.61724621e-01 1.40498847e-01
9.13345456e-01 4.64988112e-01 2.39818767e-01 -4.75054026e-01
4.74983484e-01 -5.48960090e-01 4.75756168e-01 1.01216745e+00
5.22252358e-02 4.28421676e-01 5.64160407e-01 -2.71533340e-01
-1.05407953e+00 -1.29740703e+00 -2.03109086e-01 1.32299197e+00
-3.68188679e-01 -9.97883976e-02 -6.61607146e-01 -3.35690677e-01
2.77621955e-01 1.06004679e+00 -7.42831111e-01 -3.43881696e-01
-2.39183426e-01 -1.15627277e+00 7.33344138e-01 7.23409951e-01
2.12744430e-01 -1.20604265e+00 -6.54773265e-02 3.45788777e-01
2.97694355e-01 -8.25722873e-01 -1.90156534e-01 6.48714721e-01
-7.83109188e-01 -7.82471478e-01 -1.32676065e+00 -1.46710679e-01
1.50014022e-02 -3.85121614e-01 1.18687463e+00 -6.15157664e-01
-3.08014244e-01 -5.35994768e-02 1.20080166e-01 -5.87462544e-01
-5.02701461e-01 2.12110877e-01 8.75018090e-02 -5.58176003e-02
-1.19587235e-01 -1.06561351e+00 -3.17404777e-01 8.53663310e-02
-8.57192576e-01 -4.67371792e-02 7.13449121e-01 9.91774261e-01
5.33560216e-01 -2.87475765e-01 6.58442378e-01 -9.52729225e-01
6.85042381e-01 -9.24397171e-01 -7.45473504e-01 3.50091368e-01
-1.76258892e-01 3.44420552e-01 9.09371316e-01 -4.79771018e-01
-1.41057968e+00 -2.19150394e-01 -4.41643625e-01 -8.85711193e-01
-1.03419669e-01 4.20850456e-01 -1.68677434e-01 -6.43789545e-02
8.09787035e-01 1.94459990e-01 -1.91252515e-01 -6.43530667e-01
6.67382181e-01 3.10497671e-01 6.42792583e-01 -1.02131712e+00
6.43705487e-01 7.29312748e-02 5.73248625e-01 -7.38192797e-01
-1.00160658e+00 -3.23074520e-01 -4.99625802e-01 -1.22716442e-01
9.54196393e-01 -8.12110662e-01 -9.94642437e-01 8.54370415e-01
-1.25709343e+00 -5.57861447e-01 -4.32493627e-01 5.14798641e-01
-6.25416100e-01 -4.34692055e-02 -8.73589218e-01 -9.89302874e-01
-7.20889688e-01 -1.21475661e+00 1.12650287e+00 5.52164495e-01
9.09519847e-03 -1.45911753e+00 2.66051114e-01 8.97942409e-02
5.66473007e-01 2.16533273e-01 1.25687122e+00 -8.11575830e-01
-3.52419108e-01 2.03576148e-01 -2.22995847e-01 2.79434085e-01
-3.12607616e-01 1.79781452e-01 -1.43440950e+00 -3.45279388e-02
-5.66971637e-02 -5.04644692e-01 1.20635998e+00 1.01225376e+00
1.59702313e+00 3.35414223e-02 -2.37133786e-01 1.03891623e+00
1.28734791e+00 -7.64464140e-02 5.61241508e-01 -2.37910133e-02
9.45336342e-01 4.37252760e-01 -1.94218934e-01 4.94720608e-01
1.45974070e-01 2.45331049e-01 4.15901154e-01 7.05286488e-02
2.04448123e-03 -1.16297476e-01 3.93312246e-01 6.85519934e-01
-1.86726004e-01 -3.61878484e-01 -8.78232956e-01 2.28098378e-01
-1.96843565e+00 -8.22029531e-01 -3.48346442e-01 2.02597475e+00
7.23976374e-01 7.38357368e-04 -5.51193692e-02 -4.48053271e-01
7.60678828e-01 1.76574916e-01 -1.03192425e+00 -2.19739705e-01
4.01520655e-02 4.64901119e-01 2.17876300e-01 3.17560703e-01
-8.81753862e-01 6.35399640e-01 5.29819536e+00 1.51035213e+00
-8.70440841e-01 3.55924040e-01 9.82633889e-01 -1.35489866e-01
-2.91223526e-01 -1.77393630e-01 -1.21657109e+00 6.01288855e-01
1.27137697e+00 -2.19672427e-01 3.36839139e-01 1.11083519e+00
9.01303440e-02 -2.62169596e-02 -1.15984905e+00 8.27074051e-01
-4.71180230e-01 -1.59911335e+00 1.89061984e-01 -9.49112177e-02
7.60412693e-01 2.48378456e-01 3.08618862e-02 7.97888279e-01
7.35862553e-01 -1.08579564e+00 4.99501497e-01 1.05495560e+00
6.65652335e-01 -9.34282720e-01 7.08001494e-01 6.51228368e-01
-9.09849763e-01 1.39586374e-01 -5.95779836e-01 2.98474163e-01
3.75311226e-01 1.17514253e+00 -2.06956089e-01 8.21461260e-01
6.77633762e-01 4.31666076e-01 -1.51120916e-01 1.10547602e+00
-5.66191331e-04 6.94972098e-01 -4.60360587e-01 -7.38987476e-02
2.54382789e-01 -3.70380849e-01 6.57107890e-01 1.23555648e+00
5.93200564e-01 -4.32592779e-01 -1.20613500e-01 1.72276258e+00
-2.80557871e-01 -2.30314404e-01 -4.40417767e-01 3.73300724e-02
2.78074026e-01 1.62602985e+00 -6.61655545e-01 -3.66041392e-01
3.66398506e-02 6.49355114e-01 4.99518305e-01 4.39722240e-01
-1.00341964e+00 -3.68597209e-01 6.43512130e-01 -2.42077097e-01
2.15317849e-02 -6.64411485e-02 -2.63009220e-01 -1.07921636e+00
-5.49681604e-01 -3.37655306e-01 2.94991523e-01 -8.44402313e-01
-1.84149945e+00 5.17135859e-01 1.57766491e-01 -5.92337012e-01
-3.64666522e-01 -7.93961763e-01 -1.12171149e+00 1.37352657e+00
-1.37888861e+00 -1.26915836e+00 -3.02193820e-01 6.06238544e-01
5.14183283e-01 -2.08669573e-01 6.31529987e-01 1.18774600e-01
-9.51497912e-01 3.48437369e-01 5.14078259e-01 -2.72298604e-01
6.41321003e-01 -1.45127761e+00 5.52587509e-01 5.86201608e-01
2.34131366e-02 6.95238948e-01 5.05507708e-01 -5.75772166e-01
-1.26115119e+00 -1.34347141e+00 2.26617292e-01 -2.50222206e-01
9.51499104e-01 -6.51548684e-01 -1.20353591e+00 5.31836927e-01
-1.32171690e-01 1.05687797e-01 5.53565323e-01 2.16588885e-01
-5.21524176e-02 -1.14642913e-02 -8.74519587e-01 2.77752310e-01
7.59986639e-01 -3.09876651e-01 -2.48190731e-01 4.88745481e-01
8.70729208e-01 -5.07314444e-01 -1.00204289e+00 3.22083473e-01
3.79241377e-01 -8.07901323e-01 1.00123644e+00 -5.27702093e-01
8.82315934e-01 -1.46397017e-02 6.30477294e-02 -1.56343579e+00
-3.72597218e-01 -6.60803378e-01 -7.86735535e-01 1.43095469e+00
2.60812223e-01 -4.37257648e-01 6.19382322e-01 5.48195839e-01
-4.88737881e-01 -8.94012570e-01 -9.24544215e-01 -8.89008582e-01
4.83579159e-01 -5.83471060e-01 5.49834430e-01 6.36490703e-01
-6.56491697e-01 3.91582698e-01 -5.11938691e-01 -1.32753670e-01
9.46281910e-01 -1.99333772e-01 6.72889590e-01 -1.52995682e+00
-5.86506426e-01 -7.46236563e-01 1.20769516e-01 -6.77312016e-01
4.96587574e-01 -1.00644517e+00 2.38101203e-02 -1.34682763e+00
2.63636857e-01 -2.54710972e-01 -1.61549985e-01 3.34333062e-01
-2.91271865e-01 -1.29912972e-01 3.28710638e-02 2.15064228e-01
-2.66332477e-01 9.63968217e-01 1.19059122e+00 -6.86932430e-02
-8.51028189e-02 3.62996459e-01 -5.26985824e-01 7.92338312e-01
5.06397307e-01 -4.76794958e-01 -3.13652158e-01 -3.04172169e-02
8.71859565e-02 8.13532174e-02 7.41454601e-01 -1.17647004e+00
3.94510925e-01 -1.57106161e-01 6.58396661e-01 -5.54867625e-01
3.59054178e-01 -2.17042699e-01 2.21543342e-01 2.54215568e-01
-1.32743508e-01 -3.95562708e-01 2.61202127e-01 6.39901698e-01
-1.28953625e-02 -4.99560535e-01 9.73922610e-01 -4.23001796e-01
-6.19627714e-01 5.88962615e-01 -2.18825728e-01 1.05498396e-01
9.19429004e-01 3.75785738e-01 -3.84512246e-01 -7.72359073e-02
-8.45693946e-01 3.01658005e-01 -1.15115829e-01 1.37949530e-02
1.88473001e-01 -9.45786893e-01 -6.57002151e-01 -6.25185370e-02
-2.00009555e-01 5.47691107e-01 7.46941328e-01 7.27705896e-01
-1.91004321e-01 1.74435675e-01 -2.15340972e-01 -6.30711496e-01
-3.70032549e-01 4.13871586e-01 5.62077343e-01 -5.93075395e-01
-5.71726263e-01 1.20268965e+00 5.15026033e-01 -7.06838310e-01
2.69249439e-01 -3.03717405e-01 -5.22714369e-02 1.74491078e-01
4.46425319e-01 8.28827024e-01 3.56967673e-02 -4.19836352e-03
1.31763846e-01 3.52965593e-01 6.11389503e-02 2.43211235e-03
1.50110030e+00 2.16871053e-01 -2.38543794e-01 9.24058437e-01
1.08550203e+00 -5.57776749e-01 -1.40864360e+00 -2.46138170e-01
-2.24116072e-01 3.83624226e-01 2.16972396e-01 -8.37721705e-01
-1.02141500e+00 1.27395356e+00 2.00357780e-01 -2.04391196e-01
6.43134415e-01 2.86420640e-02 6.38615370e-01 4.35904354e-01
1.96923718e-01 -7.97892332e-01 -4.10911720e-03 7.33160734e-01
7.59237826e-01 -1.10191524e+00 -1.95336759e-01 6.27283752e-02
-5.17650604e-01 1.24903905e+00 7.76060402e-01 -9.21013504e-02
8.54587495e-01 4.98214900e-01 -5.08486986e-01 -1.79857835e-01
-7.46921837e-01 -5.19619472e-02 3.98540705e-01 4.40674394e-01
2.38183945e-01 -7.38515332e-02 1.62126943e-01 1.13229477e+00
-4.69907299e-02 2.05162123e-01 1.46104530e-01 4.48269755e-01
-2.33483419e-01 -8.17432582e-01 -3.36068362e-01 7.65311539e-01
-1.77983120e-01 -3.03822130e-01 1.90056175e-01 7.28044868e-01
1.37901753e-01 2.12578684e-01 6.89797895e-03 -1.15224376e-01
-2.91304067e-02 5.50036430e-01 2.44779900e-01 -4.52818125e-01
-6.57708824e-01 -4.17485051e-02 -4.37726945e-01 -3.87357652e-01
-3.67811471e-02 -4.16446626e-01 -1.02901995e+00 -5.62327266e-01
2.62102298e-02 7.87157118e-02 7.53653526e-01 1.19642234e+00
2.93973207e-01 1.04067826e+00 4.84320261e-02 -1.29939604e+00
-8.45471740e-01 -1.25289142e+00 -9.52882349e-01 9.02823696e-04
3.33309881e-02 -9.15787816e-01 -4.45278138e-01 -1.88670769e-01] | [7.0876078605651855, 3.749976634979248] |
4d6d59e7-d5d2-42b5-b552-a0d1e66dc178 | tree-based-representation-and-generation-of | 2302.07974 | null | https://arxiv.org/abs/2302.07974v1 | https://arxiv.org/pdf/2302.07974v1.pdf | Tree-Based Representation and Generation of Natural and Mathematical Language | Mathematical language in scientific communications and educational scenarios is important yet relatively understudied compared to natural languages. Recent works on mathematical language focus either on representing stand-alone mathematical expressions, especially in their natural tree format, or mathematical reasoning in pre-trained natural language models. Existing works on jointly modeling and generating natural and mathematical languages simply treat mathematical expressions as text, without accounting for the rigid structural properties of mathematical expressions. In this paper, we propose a series of modifications to existing language models to jointly represent and generate text and math: representing mathematical expressions as sequences of node tokens in their operator tree format, using math symbol and tree position embeddings to preserve the semantic and structural properties of mathematical expressions, and using a constrained decoding method to generate mathematically valid expressions. We ground our modifications in GPT-2, resulting in a model MathGPT, and demonstrate that it outperforms baselines on mathematical expression generation tasks. | ['Andrew Lan', 'Alexander Scarlatos'] | 2023-02-15 | null | null | null | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [ 5.15811503e-01 3.51984471e-01 -3.54818180e-02 -4.76457775e-01
-3.40070754e-01 -9.46268618e-01 8.75302553e-01 4.71607238e-01
-2.06313357e-01 7.25836873e-01 3.86175215e-01 -9.67968464e-01
6.13947213e-02 -1.13396072e+00 -9.72597718e-01 -7.86904022e-02
5.75142168e-02 3.57450306e-01 -4.80917454e-01 -2.64000118e-01
2.64845133e-01 6.01491213e-01 -1.32170331e+00 8.40298086e-02
1.20479906e+00 4.16856021e-01 -2.16849983e-01 9.46061015e-01
-9.53494370e-01 1.50181961e+00 -1.01175022e+00 -1.10102630e+00
-2.73164492e-02 -4.17030573e-01 -9.49792206e-01 -9.79967043e-02
7.39396393e-01 -2.88172871e-01 -2.74430305e-01 8.74971390e-01
1.80383667e-01 3.36636215e-01 9.11130965e-01 -1.19777918e+00
-1.30182958e+00 9.82014060e-01 -3.47104296e-02 -2.30832502e-01
7.97988236e-01 -1.14747979e-01 9.74608898e-01 -5.10075927e-01
5.81359506e-01 1.53809905e+00 6.17404878e-01 7.24155784e-01
-1.50522792e+00 -5.14825523e-01 3.41483839e-02 -8.28477666e-02
-1.41696668e+00 -4.13809985e-01 6.39131010e-01 -4.63313669e-01
1.14712858e+00 2.24562839e-01 8.43144178e-01 8.40717196e-01
1.53122604e-01 8.55879724e-01 7.88235843e-01 -6.37246072e-01
-4.64655235e-02 5.32945916e-02 2.91711777e-01 1.12400401e+00
-5.70196286e-03 -3.58374953e-01 -3.14303964e-01 -4.95320894e-02
1.15982270e+00 -5.42736650e-01 -7.34039471e-02 2.84162965e-02
-1.10541916e+00 9.22753990e-01 2.46682274e-03 2.30924413e-01
-6.33493364e-02 6.22536242e-01 2.38782629e-01 3.96225154e-01
2.71534294e-01 7.58934617e-01 -2.40303546e-01 -4.75084633e-01
-1.01870656e+00 8.67971241e-01 8.35079670e-01 1.38883960e+00
4.59207356e-01 3.29278767e-01 -4.63224471e-01 6.20227039e-01
1.58885613e-01 2.48695433e-01 1.97924078e-01 -1.14062500e+00
2.26454258e-01 6.24911606e-01 -2.04211265e-01 -9.95692015e-01
-1.81635767e-02 -2.52517164e-01 -4.88508761e-01 -1.24641933e-01
5.54757714e-01 -8.22938010e-02 -9.12091434e-01 1.91721654e+00
-5.98488785e-02 2.45632619e-01 2.53941715e-01 3.46623778e-01
1.29423654e+00 9.83298421e-01 5.73536992e-01 2.43496776e-01
1.29524004e+00 -7.12081969e-01 -8.66709769e-01 1.20297544e-01
1.19067395e+00 -7.05066204e-01 1.22405577e+00 1.92260161e-01
-1.61334527e+00 -3.76785904e-01 -8.59296441e-01 -9.84610438e-01
-7.54740715e-01 9.90267545e-02 1.04583907e+00 8.75944555e-01
-1.29760075e+00 6.29884839e-01 -5.78503788e-01 2.48705208e-01
5.19596219e-01 5.43925054e-02 -2.38179043e-01 -1.12784795e-01
-1.26859152e+00 1.02231085e+00 3.04307759e-01 -3.27979624e-02
-3.30795944e-01 -1.46525908e+00 -1.73110664e+00 3.05343032e-01
-1.11494020e-01 -1.10932028e+00 1.50440061e+00 -7.28591144e-01
-1.66707993e+00 1.13273561e+00 -5.57690024e-01 -6.80659294e-01
4.43905890e-01 -9.67068523e-02 -3.76011357e-02 -1.69060647e-01
-1.18279949e-01 7.85771072e-01 2.46735692e-01 -9.60454941e-01
4.66314377e-03 2.26678327e-02 4.52065200e-01 1.25635326e-01
-9.63216126e-02 2.54200399e-01 -2.88368970e-01 -9.22861099e-01
-4.79862355e-02 -4.07704026e-01 -3.05236667e-01 2.48847544e-01
-3.23927373e-01 -7.46334553e-01 1.31886140e-01 -6.39671087e-01
1.38587916e+00 -1.72729313e+00 3.43636394e-01 1.44133061e-01
4.15342301e-01 8.66568238e-02 -1.51570797e-01 1.53138816e-01
-1.94909573e-01 6.40126765e-01 -3.28057140e-01 -5.03424942e-01
6.65248334e-01 4.80465889e-01 -5.49321830e-01 9.00148600e-03
6.46421134e-01 1.45897329e+00 -1.07580101e+00 -7.39870071e-01
4.71938163e-01 5.69788814e-01 -6.09877706e-01 1.78895086e-01
-6.77966118e-01 2.16704179e-02 -4.53996092e-01 4.63677347e-01
6.79197609e-01 1.06820539e-02 6.87134787e-02 1.08641110e-01
-1.19566381e-01 8.77389193e-01 -9.95976031e-01 1.68354583e+00
-9.71196413e-01 8.71218741e-01 -2.59892523e-01 -9.25857365e-01
7.73585916e-01 2.55271375e-01 1.81703672e-01 -4.04479980e-01
1.67928815e-01 -9.28505138e-03 -2.07621276e-01 -3.74593198e-01
6.44662082e-01 -4.72733080e-01 -3.54353398e-01 5.10697544e-01
2.97961026e-01 -1.21014321e+00 5.70166528e-01 6.63348198e-01
7.99422741e-01 5.90678275e-01 2.34325290e-01 -6.14761785e-02
5.34920633e-01 -2.47595057e-01 -9.22045410e-02 9.19632435e-01
2.38952726e-01 3.38593632e-01 7.83074319e-01 -4.07418221e-01
-9.20085669e-01 -1.47415793e+00 -3.19239378e-01 1.01896000e+00
-5.37422478e-01 -1.03065395e+00 -7.03538835e-01 -2.10650563e-01
-6.54045632e-03 1.47935688e+00 -4.10746723e-01 -3.54533717e-02
-7.25044906e-01 -3.17644328e-01 1.05315840e+00 4.62605834e-01
2.55385548e-01 -9.11937058e-01 -2.42646873e-01 2.61853486e-01
1.03947863e-01 -1.36148572e+00 -1.25978082e-01 -4.19004038e-02
-6.00789011e-01 -5.62604606e-01 -7.07062960e-01 -9.66069818e-01
7.57770240e-01 -3.73181283e-01 1.63028765e+00 3.31467271e-01
-2.49235421e-01 3.71293038e-01 -9.61688086e-02 -8.08514118e-01
-7.15418637e-01 -2.90333509e-01 -4.28111523e-01 -4.65808868e-01
2.91109592e-01 -5.53089142e-01 3.49635243e-01 -8.72463882e-01
-1.04920959e+00 3.94544214e-01 1.89968482e-01 5.21155059e-01
2.78966367e-01 -5.12994826e-02 -3.74998301e-02 -1.06761014e+00
9.37979579e-01 -2.01639950e-01 -5.27432024e-01 3.58014524e-01
-3.39264907e-02 5.60992897e-01 7.66920209e-01 -3.60438764e-01
-9.71672297e-01 -1.68080911e-01 -4.99367923e-01 -2.99139708e-01
-6.08003549e-02 6.48128450e-01 -9.88372788e-02 -1.53932378e-01
3.37796628e-01 5.77538311e-01 -1.86750397e-01 -7.74417073e-02
9.10854280e-01 1.90019593e-01 6.58043742e-01 -1.45766068e+00
9.41112638e-01 -2.21629933e-01 4.24907058e-01 -9.63495791e-01
-8.85093272e-01 2.22553059e-01 -5.26214600e-01 2.26870701e-01
7.17900515e-01 -8.77765298e-01 -7.46878207e-01 4.28850561e-01
-1.71403003e+00 -3.22944522e-01 -4.00840133e-01 1.50009140e-01
-5.79041183e-01 6.13711655e-01 -8.22466016e-01 -9.59188581e-01
-6.25151023e-02 -9.93284225e-01 1.25880969e+00 1.65351450e-01
-6.19538784e-01 -1.48633146e+00 -4.74063575e-01 2.59577513e-01
5.25940776e-01 4.59964514e-01 1.54932177e+00 -2.22087875e-01
-6.99459672e-01 -4.03371565e-02 -3.62773567e-01 1.86018541e-01
-7.07199797e-02 3.64397079e-01 -6.95461094e-01 6.00950360e-01
-2.78731465e-01 -3.13994974e-01 6.35264754e-01 1.78176686e-01
1.59482133e+00 -5.69624841e-01 1.64391145e-01 9.48333740e-01
1.05927658e+00 -2.47675508e-01 7.17840910e-01 -5.81938997e-02
6.27624154e-01 5.27454138e-01 -3.42663914e-01 2.45256320e-01
8.62549841e-01 3.49100798e-01 -1.38891160e-01 -1.08735830e-01
-2.08195690e-02 -6.98776305e-01 2.83087075e-01 7.24144340e-01
1.26807049e-01 -1.91326231e-01 -8.92024100e-01 6.40829504e-01
-1.21973169e+00 -9.36323047e-01 -2.79833138e-01 1.87983155e+00
1.60402572e+00 -1.53563241e-03 -3.37749273e-01 6.59907050e-03
3.59014347e-02 1.33567080e-01 3.46557498e-02 -1.03554428e+00
-3.59004766e-01 1.33425069e+00 2.73847312e-01 8.38145912e-01
-8.65346551e-01 1.24831569e+00 6.89384699e+00 8.51226807e-01
-8.14378858e-01 -3.36996108e-01 4.11438406e-01 1.02869652e-01
-9.12072301e-01 -4.30598110e-02 -5.72339892e-01 1.82749107e-01
1.02227008e+00 -7.80312657e-01 4.67361867e-01 3.71747404e-01
1.18434384e-01 3.89057606e-01 -1.32032585e+00 1.05988336e+00
2.09297668e-02 -1.59521580e+00 8.85613382e-01 -5.12974739e-01
4.49858606e-01 -7.92376578e-01 -6.06571836e-03 6.06841147e-01
7.94378400e-01 -1.80155122e+00 7.84896851e-01 5.32685578e-01
6.54076219e-01 -4.92505878e-01 3.34600627e-01 7.29857013e-02
-1.15638125e+00 3.92522126e-01 -1.05820715e-01 -5.74694932e-01
1.91401660e-01 3.70715410e-01 -6.43294096e-01 5.27829707e-01
2.97088493e-02 8.35273147e-01 -7.15742826e-01 5.33376098e-01
-7.96169579e-01 6.45293415e-01 -3.96060139e-01 -4.69351709e-01
4.25785989e-01 -3.87422323e-01 3.44230622e-01 1.58458638e+00
2.78714836e-01 3.22356373e-01 -1.60905421e-02 1.79872715e+00
-2.05268949e-01 2.49214113e-01 -7.03899443e-01 -7.09001422e-01
3.82244527e-01 9.41149533e-01 -3.71111259e-02 -7.67895222e-01
-4.89269137e-01 8.09394479e-01 3.49277586e-01 3.69313657e-01
-9.75302219e-01 -6.06791258e-01 8.13171446e-01 5.82097061e-02
-2.39493608e-01 -7.88222909e-01 -6.41010404e-01 -1.38282704e+00
4.14766110e-02 -7.52320707e-01 1.06800005e-01 -1.16110456e+00
-1.06069589e+00 1.52695388e-01 1.64512828e-01 -5.99750400e-01
-3.81625682e-01 -1.02512574e+00 -6.88688338e-01 1.31021416e+00
-1.50878370e+00 -1.34017706e+00 1.30328014e-01 2.52915978e-01
3.49252164e-01 2.19590768e-01 1.14082456e+00 9.26767290e-02
-3.18732172e-01 1.01078725e+00 -3.19737107e-01 5.16675413e-01
3.62314805e-02 -1.55657291e+00 6.60439551e-01 6.14717603e-01
3.82817030e-01 1.41471612e+00 8.06882679e-01 -2.70776719e-01
-1.73061025e+00 -1.00253296e+00 1.58422518e+00 -7.41904616e-01
8.32771897e-01 -7.82503068e-01 -1.02726269e+00 9.09602225e-01
1.74451709e-01 -2.52030730e-01 7.78002739e-01 -1.07042141e-01
-4.66918349e-01 6.24257684e-01 -1.02532864e+00 1.19753659e+00
1.00473070e+00 -5.98581016e-01 -7.90013671e-01 5.97291708e-01
1.19539666e+00 -8.67779434e-01 -9.99124467e-01 5.36943935e-02
5.49376607e-01 -1.90342158e-01 1.17743194e+00 -1.21395826e+00
1.03186989e+00 -2.19482079e-01 3.44631001e-02 -1.11480832e+00
1.02036316e-02 -1.02069783e+00 -1.02813572e-01 1.12301481e+00
4.41084385e-01 -2.75783002e-01 9.03227985e-01 9.31753457e-01
-1.65643632e-01 -6.86968565e-01 -6.57336712e-01 -5.07821500e-01
9.11331356e-01 -7.36409724e-01 8.28167737e-01 1.14105690e+00
1.89696271e-02 4.62648064e-01 8.96452516e-02 -2.58804888e-01
3.84610951e-01 1.30720362e-01 8.28710198e-01 -9.60339546e-01
-2.39235029e-01 -8.95941436e-01 -6.95364296e-01 -1.44815528e+00
9.29064155e-01 -1.31273639e+00 -2.90289849e-01 -1.71205008e+00
-2.21308947e-01 -2.06846222e-01 3.37035358e-01 5.24240494e-01
-1.72431007e-01 -1.76241502e-01 1.95453331e-01 -6.81454480e-01
-1.73114896e-01 7.08968103e-01 1.32813215e+00 -3.90689552e-01
1.57058969e-01 -3.97400856e-01 -7.39517212e-01 6.14036918e-01
4.53149557e-01 3.01345368e-03 -3.82375866e-01 -8.98706138e-01
3.66218030e-01 6.42443150e-02 5.91446638e-01 -6.93316340e-01
7.82002956e-02 -5.30103922e-01 4.13082421e-01 -1.15981042e-01
2.83098787e-01 -3.57678920e-01 -2.28227004e-01 1.04665942e-02
-7.48414695e-01 1.86858505e-01 6.97289586e-01 -1.49089709e-01
-1.39403135e-01 -5.41251957e-01 5.20817876e-01 -3.44458491e-01
-5.13821602e-01 -7.56824613e-02 -6.28236532e-01 2.65580595e-01
6.25734031e-01 -3.20628732e-01 -2.22556770e-01 -4.99404877e-01
-5.97704530e-01 2.92042375e-01 2.91705459e-01 3.54456604e-01
6.61420107e-01 -1.28117442e+00 -7.99817085e-01 1.84260026e-01
-1.32673725e-01 1.95355535e-01 -5.15434332e-02 2.01739445e-01
-9.85032797e-01 4.54108030e-01 6.57185018e-02 -4.51363660e-02
-1.14859390e+00 2.54848152e-01 5.27342141e-01 -2.91741014e-01
-1.67255566e-01 1.01551354e+00 -1.48202796e-02 -9.97852683e-01
2.35281557e-01 -9.99709606e-01 -8.17053467e-02 -4.24707800e-01
3.79596084e-01 -1.10127337e-01 -1.77589834e-01 -5.33253491e-01
4.11556698e-02 6.09641016e-01 2.57691056e-01 -7.82963634e-02
1.15107036e+00 4.17395502e-01 -5.57742596e-01 3.11970651e-01
1.16344404e+00 3.08410674e-01 -3.37923527e-01 -2.44976610e-01
-3.62686552e-02 -1.93617702e-01 -8.04389194e-02 -6.69615328e-01
-4.70186919e-01 1.25444984e+00 -2.95017511e-01 -1.70704439e-01
8.15295279e-01 -2.02382386e-01 7.57407486e-01 6.03764772e-01
-2.63030268e-02 -4.54635322e-01 -2.57980973e-01 9.30225909e-01
7.37087548e-01 -7.81282008e-01 1.54251188e-01 -7.04219460e-01
-1.58296421e-01 1.45788479e+00 4.00107235e-01 -3.84301841e-02
4.01684135e-01 6.22207820e-01 -1.05797298e-01 -1.08051382e-01
-8.48887324e-01 2.97275167e-02 3.13579649e-01 6.26671374e-01
1.31348860e+00 1.69243366e-01 -2.69812644e-01 8.15752149e-01
-1.07475007e+00 3.22533339e-01 7.99424708e-01 1.07331574e+00
-1.95673238e-02 -1.20288980e+00 -1.27670348e-01 4.55329180e-01
-6.34577990e-01 -7.14899421e-01 -6.69625223e-01 6.99777603e-01
7.27089643e-02 6.67753339e-01 2.63248473e-01 1.05193034e-01
3.67583275e-01 4.08222646e-01 1.11301935e+00 -8.38973999e-01
-6.10936105e-01 -6.77980065e-01 1.69391423e-01 -7.60739073e-02
-1.35804832e-01 -4.21424270e-01 -1.55830681e+00 -6.81548715e-01
2.52516747e-01 2.07360908e-01 4.63431388e-01 9.62326527e-01
2.10962564e-01 7.16287136e-01 -8.92595798e-02 -5.07860065e-01
-6.28073752e-01 -7.20397472e-01 -1.53455958e-01 5.27020156e-01
2.09472910e-01 -3.52467090e-01 -1.12084858e-01 5.61381280e-01] | [9.647650718688965, 7.3288092613220215] |
3fa09bf1-bfb7-4306-8616-341903ae93ea | generating-formulaic-text-by-splicing | 2101.08248 | null | https://arxiv.org/abs/2101.08248v4 | https://arxiv.org/pdf/2101.08248v4.pdf | Data-to-text Generation by Splicing Together Nearest Neighbors | We propose to tackle data-to-text generation tasks by directly splicing together retrieved segments of text from "neighbor" source-target pairs. Unlike recent work that conditions on retrieved neighbors but generates text token-by-token, left-to-right, we learn a policy that directly manipulates segments of neighbor text, by inserting or replacing them in partially constructed generations. Standard techniques for training such a policy require an oracle derivation for each generation, and we prove that finding the shortest such derivation can be reduced to parsing under a particular weighted context-free grammar. We find that policies learned in this way perform on par with strong baselines in terms of automatic and human evaluation, but allow for more interpretable and controllable generation. | ['Karl Stratos', 'Arturs Backurs', 'Sam Wiseman'] | 2021-01-20 | null | https://aclanthology.org/2021.emnlp-main.352 | https://aclanthology.org/2021.emnlp-main.352.pdf | emnlp-2021-11 | ['conditional-text-generation'] | ['natural-language-processing'] | [ 7.32710779e-01 9.08502460e-01 -3.25592071e-01 -4.92124856e-01
-1.82358849e+00 -1.19314694e+00 8.40561628e-01 1.66813523e-01
-3.57475013e-01 1.21488965e+00 4.89442021e-01 -5.74473977e-01
2.91794151e-01 -1.13145077e+00 -1.19828391e+00 -4.72896248e-01
1.62825048e-01 8.71794522e-01 2.73163915e-02 -2.91369528e-01
8.59025568e-02 1.97578356e-01 -1.14711523e+00 6.48648798e-01
8.62873077e-01 2.32451349e-01 1.61205396e-01 1.08565438e+00
-3.53815079e-01 2.50382096e-01 -8.63335311e-01 -1.10353494e+00
4.22402650e-01 -7.80941308e-01 -9.16695118e-01 -1.16283469e-01
5.92553616e-01 -3.04688603e-01 1.08455539e-01 6.16853654e-01
4.96473461e-01 -5.49879931e-02 6.36292219e-01 -9.25939262e-01
-9.10934567e-01 1.51197851e+00 -1.98477685e-01 -2.21353382e-01
5.87100744e-01 2.44394884e-01 1.22857273e+00 -5.17124534e-01
1.09238768e+00 1.45423770e+00 2.95736432e-01 9.73929346e-01
-1.50193238e+00 -3.32732052e-01 2.60657847e-01 -8.56287241e-01
-7.87661433e-01 -7.04057336e-01 1.18263662e-01 -1.82912558e-01
1.28221858e+00 3.80784899e-01 4.44239080e-01 1.49527478e+00
-9.74364812e-04 8.49558413e-01 5.49273670e-01 -9.17120099e-01
7.28368163e-02 -1.92447200e-01 -1.35010645e-01 6.93217874e-01
3.97212803e-01 1.14087865e-01 -5.68227947e-01 -3.63195866e-01
4.17709470e-01 -7.23876536e-01 -7.25814188e-03 -1.92511361e-02
-1.46762431e+00 9.92995203e-01 -2.30895191e-01 1.25907898e-01
7.26027563e-02 5.33055842e-01 3.98586154e-01 2.58020043e-01
4.33647156e-01 7.35404789e-01 -7.22503364e-01 -2.14934036e-01
-8.26782644e-01 7.82616258e-01 1.07010567e+00 1.56716549e+00
6.42315626e-01 1.30433917e-01 -8.80708158e-01 5.42538047e-01
-9.45942104e-03 9.49054301e-01 5.47209084e-01 -9.04419124e-01
9.19820130e-01 1.25728752e-02 2.59622782e-01 -1.17092781e-01
1.55646756e-01 -1.56967312e-01 -3.54752570e-01 -7.28066862e-02
4.90726292e-01 -8.00054252e-01 -1.02058339e+00 2.13890290e+00
3.34529221e-01 -2.53077865e-01 3.36648554e-01 3.01156759e-01
2.43358582e-01 8.24982047e-01 2.79739857e-01 -2.13606283e-01
1.19378865e+00 -8.50880325e-01 -4.08637285e-01 -4.74218369e-01
9.50523555e-01 -6.77590847e-01 1.04257929e+00 1.19938873e-01
-1.56461203e+00 -2.60788798e-01 -7.07832932e-01 -3.25315684e-01
-1.68983221e-01 1.51219562e-01 5.42326629e-01 6.64001584e-01
-1.29165959e+00 7.15364158e-01 -6.64073467e-01 -8.58597457e-02
1.19970828e-01 1.93323776e-01 -4.10402596e-01 1.33377612e-01
-1.13807070e+00 6.05425239e-01 6.58228397e-01 -1.69500053e-01
-9.47875440e-01 -6.97559774e-01 -9.54805315e-01 -1.17180906e-01
3.43190938e-01 -1.29279757e+00 1.89698696e+00 -7.92140543e-01
-1.71342373e+00 7.63357341e-01 -5.37175715e-01 -6.99073792e-01
7.97342241e-01 -3.54787558e-01 -7.11044893e-02 -6.93792626e-02
4.49969441e-01 1.01963913e+00 9.50170398e-01 -1.13466537e+00
-8.40223014e-01 -1.99596047e-01 1.02949813e-01 2.17288375e-01
1.73881784e-01 3.31447236e-02 -5.12087941e-01 -7.69265175e-01
-2.02178001e-01 -9.35807228e-01 -3.70683759e-01 -4.72191334e-01
-1.00503850e+00 -3.06688040e-01 3.06882679e-01 -7.05839813e-01
9.64689314e-01 -1.59616446e+00 1.22962847e-01 6.39733747e-02
-2.24771380e-01 1.77545846e-01 -4.72243965e-01 7.44318604e-01
1.33280352e-01 7.47536778e-01 -3.58035743e-01 -7.68693388e-01
4.16631311e-01 1.17750905e-01 -8.10144365e-01 -1.58019006e-01
4.01757240e-01 1.22849965e+00 -1.11836278e+00 -5.83376110e-01
-3.14788759e-01 1.28190190e-01 -7.22277522e-01 4.18712586e-01
-9.38555479e-01 6.96161538e-02 -5.31174302e-01 2.73493230e-01
2.61568010e-01 1.95306093e-01 4.15131897e-01 3.14921826e-01
1.50993034e-01 7.09657311e-01 -8.19882333e-01 1.94757915e+00
-6.68489099e-01 3.94654542e-01 -1.89939648e-01 -5.06238341e-01
7.26882935e-01 3.81648988e-01 -2.87495613e-01 -3.80917847e-01
-2.55962878e-01 3.11894625e-01 -2.97980338e-01 -3.13326299e-01
9.29360330e-01 -1.86774611e-01 -5.91372788e-01 7.50328422e-01
3.47610377e-02 -4.02371466e-01 7.42956996e-01 5.37919521e-01
1.14466846e+00 6.10023975e-01 -4.65807356e-02 1.39197379e-01
5.35012707e-02 5.35775907e-02 3.97899806e-01 1.24573362e+00
6.04342937e-01 7.32273757e-01 8.24069500e-01 1.16509777e-02
-1.46937656e+00 -1.15286589e+00 3.29929322e-01 1.34228480e+00
-4.46156502e-01 -4.62736368e-01 -1.13558674e+00 -1.05243146e+00
4.26436514e-02 1.37991452e+00 -4.34045672e-01 -1.73645020e-02
-1.06813335e+00 -4.21563208e-01 1.03501511e+00 5.45268476e-01
-5.21150120e-02 -1.33451414e+00 -2.98612267e-01 5.02303779e-01
-2.01132596e-01 -1.05206919e+00 -7.13714838e-01 3.40581685e-01
-9.29647028e-01 -7.04171538e-01 -5.64738035e-01 -8.59078109e-01
7.93742061e-01 -1.81972846e-01 1.51125729e+00 -5.96289746e-02
1.98712796e-02 -3.88493910e-02 -2.29428977e-01 -3.08819711e-01
-1.15608811e+00 7.21145153e-01 -5.05801737e-01 -4.89592373e-01
-1.33220926e-01 -3.04845005e-01 -2.31213838e-01 -1.82802454e-01
-1.10275412e+00 3.48097175e-01 9.00621891e-01 7.84924150e-01
6.43063784e-01 -4.95111674e-01 6.21581614e-01 -1.51606917e+00
9.54373300e-01 -3.30952346e-01 -6.68877661e-01 5.31780303e-01
-3.89646530e-01 6.65997505e-01 8.00165296e-01 -1.35538995e-01
-1.26017189e+00 4.95107546e-02 -9.76265594e-02 9.27779824e-02
-3.44778001e-01 1.72045395e-01 -3.22225630e-01 7.90936530e-01
7.39025772e-01 1.71939000e-01 -2.98463613e-01 -2.68913507e-01
1.01441658e+00 3.33792150e-01 6.26374304e-01 -1.21749330e+00
1.05014324e+00 7.98842013e-02 -1.78733900e-01 -1.99096933e-01
-7.12982476e-01 2.00990424e-01 -3.77996057e-01 5.71212888e-01
7.98718929e-01 -8.43237758e-01 -3.55647177e-01 1.49535507e-01
-1.58046591e+00 -8.17338049e-01 -7.20969975e-01 -9.01189297e-02
-6.70672834e-01 4.55802232e-02 -6.50497556e-01 -7.83849478e-01
-6.64259493e-01 -8.52334261e-01 1.74261665e+00 -6.94435909e-02
-5.12939811e-01 -9.22661066e-01 1.95291653e-01 1.58064499e-01
2.39616737e-01 8.95066336e-02 1.25534499e+00 -8.59340250e-01
-7.38816857e-01 -2.51327813e-01 1.08392239e-01 2.32919857e-01
-1.63478211e-01 2.75430590e-01 -6.38798177e-01 -2.41911933e-02
-6.09444678e-01 -3.88216197e-01 8.18699896e-01 1.36597410e-01
7.35623240e-01 -7.81019688e-01 -6.19112313e-01 4.31333870e-01
1.14107716e+00 4.13002186e-02 5.13951957e-01 5.86295547e-03
6.53017521e-01 5.78054547e-01 2.81527996e-01 1.53927609e-01
2.66340643e-01 3.82339120e-01 1.08048052e-03 1.91445738e-01
-1.53652430e-01 -9.54990625e-01 6.42505646e-01 3.06258738e-01
4.65179294e-01 -9.08975840e-01 -5.57653189e-01 7.85637498e-01
-1.70007801e+00 -8.74774516e-01 9.75260064e-02 2.16844630e+00
1.35979474e+00 2.86919594e-01 -9.86210927e-02 -2.12741792e-01
8.60638440e-01 1.01568900e-01 -2.91697145e-01 -7.03497112e-01
-2.31870949e-01 7.71426380e-01 7.12802052e-01 8.62253487e-01
-7.95085073e-01 1.33243573e+00 7.32082224e+00 9.88578141e-01
-7.38432407e-01 -2.11892091e-02 6.68713331e-01 -3.15188468e-01
-1.26742280e+00 4.62895244e-01 -1.20647788e+00 3.61727774e-01
1.17891777e+00 -1.83364719e-01 4.35958028e-01 6.97264671e-01
1.73616499e-01 8.85162354e-02 -1.38571858e+00 3.67373973e-02
-8.87943804e-02 -1.44312406e+00 5.86945534e-01 -7.29458705e-02
9.49070752e-01 -3.51624817e-01 -6.58719614e-02 3.45308840e-01
1.20483208e+00 -1.05173337e+00 1.08221352e+00 3.36500376e-01
8.16626549e-01 -6.54711246e-01 3.08333278e-01 4.06083405e-01
-7.23314524e-01 2.69874096e-01 -2.41945073e-01 3.59431982e-01
5.15542150e-01 7.67162442e-01 -1.23045635e+00 6.04082823e-01
-8.87551159e-02 1.07186139e-01 -3.27546090e-01 4.12724614e-01
-8.22576582e-01 6.71890438e-01 -1.27542943e-01 -1.89865991e-01
1.12564296e-01 -1.96215063e-01 6.29362643e-01 1.51025617e+00
8.19120109e-01 -2.24876717e-01 1.25096515e-01 1.03663969e+00
-5.85371077e-01 1.82076152e-02 -8.02349865e-01 -3.48875970e-01
6.95367098e-01 1.15140831e+00 -4.66509670e-01 -8.03181767e-01
5.98834939e-02 1.19747746e+00 4.93045926e-01 3.68773311e-01
-7.02306330e-01 -5.92480421e-01 3.85833800e-01 4.98157963e-02
6.63420200e-01 -9.72689092e-02 -2.23506540e-01 -1.01189983e+00
2.28446707e-01 -1.01444829e+00 3.06003153e-01 -8.88704896e-01
-1.15842354e+00 5.88556826e-01 1.12826601e-01 -5.76814413e-01
-1.04977930e+00 -2.05993861e-01 -6.69313312e-01 1.10861886e+00
-1.19874597e+00 -1.27323568e+00 3.97232682e-01 2.35436335e-01
6.20307326e-01 8.44432879e-03 9.70183074e-01 -3.08270991e-01
-3.41823459e-01 9.76094544e-01 -7.22433478e-02 2.15424135e-01
6.61596000e-01 -1.56990147e+00 1.26597416e+00 1.15879786e+00
3.97622645e-01 8.44109654e-01 6.99063659e-01 -8.93603504e-01
-1.43558955e+00 -1.30263162e+00 1.39185250e+00 -6.32349312e-01
4.53542650e-01 -7.01891720e-01 -4.75945920e-01 1.08436346e+00
5.66836298e-01 -3.37673247e-01 3.00358653e-01 4.45402740e-03
-6.17288828e-01 2.30423495e-01 -1.00495803e+00 9.27849710e-01
1.41681504e+00 -3.89609367e-01 -5.92271984e-01 5.40483057e-01
1.23086154e+00 -7.97478676e-01 -4.42093939e-01 1.46243917e-02
4.30203021e-01 -6.68574631e-01 5.19367516e-01 -1.03623080e+00
7.71765888e-01 -1.12033017e-01 -8.82212967e-02 -1.60514128e+00
8.50798190e-02 -1.38772857e+00 1.35184646e-01 1.43635809e+00
1.15006769e+00 -5.03620744e-01 9.45868969e-01 5.04318357e-01
-5.39340675e-01 -5.40821493e-01 -7.09485710e-01 -7.50492513e-01
6.50097013e-01 -4.46389079e-01 9.43139136e-01 2.19294742e-01
-6.62366748e-02 5.24086356e-01 -8.17668214e-02 -1.34734195e-02
3.59936565e-01 1.75829872e-01 8.79758358e-01 -6.90514088e-01
-7.07431197e-01 -4.39835936e-01 3.27815622e-01 -1.26278853e+00
4.34006780e-01 -1.08416772e+00 5.34379840e-01 -1.65482748e+00
3.14176753e-02 -6.62305534e-01 4.45037007e-01 5.69855452e-01
-4.89962935e-01 2.57961024e-02 1.03206351e-01 -1.89410701e-01
-3.68535370e-01 2.21845925e-01 1.19112182e+00 -1.48252472e-01
-7.99587145e-02 -7.71250799e-02 -1.16545653e+00 2.31604382e-01
8.08809817e-01 -6.82486594e-01 -4.75885093e-01 -8.10906470e-01
5.34449637e-01 4.78359103e-01 -3.04708909e-03 -4.55176592e-01
-2.12210752e-02 -2.67935902e-01 1.60214245e-01 -2.38309011e-01
7.45763609e-05 -5.66021279e-02 1.20483726e-01 6.38406873e-02
-1.05861795e+00 2.01020479e-01 5.76265305e-02 5.05853891e-01
1.60728425e-01 -6.49370074e-01 1.78611934e-01 -4.10711676e-01
-3.01548541e-02 1.98601589e-01 -3.79057527e-01 5.70248961e-01
7.34342515e-01 1.10890888e-01 -5.99341631e-01 -4.12747711e-01
-5.38050294e-01 -6.03023507e-02 4.68856514e-01 1.67284802e-01
1.41450390e-01 -1.07247722e+00 -9.80380595e-01 1.88408911e-01
-6.02523647e-02 3.15829128e-01 -2.10120127e-01 6.85690567e-02
-4.36259985e-01 5.38535893e-01 3.73137802e-01 -1.49332196e-01
-9.24956024e-01 2.97861993e-01 1.04907118e-01 -7.06163168e-01
-2.74407893e-01 9.19596434e-01 5.56271076e-02 -6.39899373e-01
-1.28779411e-02 -4.84262258e-01 5.97399771e-01 -1.79475859e-01
2.70466566e-01 3.59240584e-02 7.37846270e-02 -1.51206940e-01
1.72815770e-01 2.19008878e-01 -2.26588547e-01 -9.01519597e-01
9.26768601e-01 6.94807768e-02 -1.43547384e-02 6.67604879e-02
1.08060551e+00 5.53952038e-01 -1.16917706e+00 3.76001932e-02
9.39336792e-03 -1.89050689e-01 -5.73988974e-01 -9.96454060e-01
-7.81201601e-01 6.00145102e-01 -2.33024955e-01 2.81342238e-01
4.99433458e-01 5.76716885e-02 1.14875221e+00 6.73244536e-01
2.71179676e-01 -9.36790287e-01 -1.00002989e-01 5.31248450e-01
6.58564150e-01 -8.54033172e-01 -3.70682269e-01 -4.31810766e-01
-6.70160890e-01 1.02855837e+00 3.86028171e-01 4.96949395e-03
-1.11231199e-02 5.41586280e-01 -6.15614168e-02 4.25709665e-01
-1.12049603e+00 -9.85915139e-02 6.31281808e-02 8.96934748e-01
6.34385943e-01 1.59242228e-01 -3.17512631e-01 3.42115790e-01
-8.14082563e-01 -7.28276595e-02 4.50712562e-01 1.08300769e+00
-3.83945793e-01 -1.92739344e+00 -8.42842460e-02 4.47257906e-01
-6.46557629e-01 -6.03039861e-01 -3.15945476e-01 8.36268306e-01
2.65097879e-02 1.00602520e+00 1.63803875e-01 -3.13496590e-03
2.21497238e-01 4.58703965e-01 6.30492806e-01 -1.05720699e+00
-8.39409232e-01 2.00289004e-02 7.14110434e-01 -2.58249462e-01
1.17331408e-01 -6.94938421e-01 -1.46041620e+00 -3.09925675e-01
-1.71365783e-01 4.65028405e-01 6.45395756e-01 9.21164691e-01
6.85679972e-01 3.37720484e-01 5.21782160e-01 -5.18369794e-01
-8.34382117e-01 -9.00531352e-01 -1.79551631e-01 4.87368017e-01
1.59479514e-01 2.29513109e-01 -1.93934903e-01 4.33652490e-01] | [11.480525970458984, 8.974837303161621] |
ca0655a7-ae99-4c01-a247-f21cdcb7c113 | multiscale-combinatorial-grouping-for-image | 1503.00848 | null | http://arxiv.org/abs/1503.00848v4 | http://arxiv.org/pdf/1503.00848v4.pdf | Multiscale Combinatorial Grouping for Image Segmentation and Object Proposal Generation | We propose a unified approach for bottom-up hierarchical image segmentation
and object proposal generation for recognition, called Multiscale Combinatorial
Grouping (MCG). For this purpose, we first develop a fast normalized cuts
algorithm. We then propose a high-performance hierarchical segmenter that makes
effective use of multiscale information. Finally, we propose a grouping
strategy that combines our multiscale regions into highly-accurate object
proposals by exploring efficiently their combinatorial space. We also present
Single-scale Combinatorial Grouping (SCG), a faster version of MCG that
produces competitive proposals in under five second per image. We conduct an
extensive and comprehensive empirical validation on the BSDS500, SegVOC12, SBD,
and COCO datasets, showing that MCG produces state-of-the-art contours,
hierarchical regions, and object proposals. | ['Jordi Pont-Tuset', 'Jonathan T. Barron', 'Pablo Arbelaez', 'Jitendra Malik', 'Ferran Marques'] | 2015-03-03 | null | null | null | null | ['object-proposal-generation'] | ['computer-vision'] | [ 6.85313568e-02 -4.43128198e-02 -2.54439890e-01 -4.04290468e-01
-1.34035838e+00 -5.31249404e-01 4.12532747e-01 1.01716653e-01
-3.59993488e-01 1.93382874e-01 -1.49297252e-01 -2.81067818e-01
2.46910349e-01 -6.11383855e-01 -4.13270563e-01 -4.66297746e-01
4.64847684e-02 6.81564391e-01 1.09331751e+00 5.49988598e-02
5.40165901e-01 9.05932486e-01 -1.31395280e+00 2.09199846e-01
8.88242185e-01 9.85560834e-01 1.54231570e-03 7.25265384e-01
-2.74636507e-01 -1.05456986e-01 -4.83624369e-01 -4.59365755e-01
3.78946841e-01 -3.23689610e-01 -1.01692915e+00 8.80331278e-01
6.97792470e-01 -1.65946811e-01 3.08473200e-01 9.01679933e-01
2.97032177e-01 5.30583933e-02 9.31871414e-01 -9.22239125e-01
-1.37726709e-01 6.17757499e-01 -1.12535048e+00 6.91253841e-02
-1.90036800e-02 -1.03463028e-02 1.03532064e+00 -1.30533898e+00
9.04044032e-01 1.48342597e+00 6.82392657e-01 3.74450296e-01
-1.52336848e+00 -4.27985996e-01 4.13212091e-01 -2.66967535e-01
-1.42655647e+00 -5.87010123e-02 7.46088088e-01 -3.99239480e-01
6.73422039e-01 3.07250053e-01 8.01990569e-01 3.53554815e-01
-2.37212870e-02 1.26863658e+00 1.13163435e+00 -5.64871490e-01
4.43561614e-01 -3.46049547e-01 5.15990376e-01 9.68155921e-01
2.56908476e-01 -1.62297189e-01 -2.84833163e-01 -1.40817791e-01
1.01819754e+00 -4.45730329e-01 8.83919820e-02 -6.77808106e-01
-1.05812097e+00 9.08065915e-01 5.03985465e-01 1.21942073e-01
-3.63285124e-01 -3.43197472e-02 1.96444333e-01 -3.19940150e-01
4.30836141e-01 3.54281634e-01 -2.06853777e-01 4.70144153e-01
-1.40140033e+00 3.30212325e-01 6.20878220e-01 1.14729834e+00
7.62299538e-01 -2.36139800e-02 -2.94864744e-01 1.02660513e+00
3.85563165e-01 1.96678653e-01 4.55554686e-02 -1.36188614e+00
1.70737267e-01 6.10841632e-01 6.11806251e-02 -8.63142908e-01
-5.67696154e-01 -6.94017410e-01 -6.33384407e-01 2.06619427e-01
3.69692653e-01 2.24837974e-01 -1.52406967e+00 1.00722241e+00
6.50775135e-01 -5.95841631e-02 -2.69988239e-01 8.87148678e-01
7.56134987e-01 5.71617126e-01 1.40277267e-01 -1.05869129e-01
1.45818186e+00 -1.35350692e+00 -3.23269337e-01 -6.93316087e-02
2.43978515e-01 -7.29322433e-01 7.09942341e-01 4.04645324e-01
-1.31613302e+00 -7.48649120e-01 -9.88933206e-01 -3.02814338e-02
-9.54787731e-02 1.31216317e-01 8.63249898e-01 6.31674230e-01
-1.05262947e+00 4.48752493e-01 -8.51509452e-01 -1.30215034e-01
7.63762951e-01 3.23402107e-01 -3.78474556e-02 3.15489434e-02
-3.34698528e-01 3.67144674e-01 4.79159325e-01 -3.42077278e-02
-8.77187192e-01 -5.20319998e-01 -7.92885959e-01 -1.21614039e-01
3.06944698e-01 -5.05992234e-01 1.32157648e+00 -5.07431686e-01
-1.38060725e+00 1.03373802e+00 -1.59922943e-01 -6.20965660e-01
4.97696489e-01 6.86894506e-02 1.95949167e-01 5.08942544e-01
3.34936112e-01 1.60950124e+00 7.47095406e-01 -1.78758931e+00
-8.55971992e-01 -2.82760799e-01 -3.66680473e-01 9.41603258e-02
3.13143015e-01 2.20646456e-01 -1.19200027e+00 -9.71014559e-01
8.16255450e-01 -7.26737082e-01 -8.41120362e-01 -1.45829812e-01
-9.24277246e-01 -5.10955155e-01 5.35733998e-01 -4.94696945e-01
9.99869227e-01 -2.01604128e+00 1.71063751e-01 4.89147961e-01
1.00640945e-01 -7.78180510e-02 -1.41110614e-01 -1.46549031e-01
3.30308974e-01 5.48213243e-01 -7.33648539e-01 -7.91035712e-01
-8.96829590e-02 7.11086467e-02 -4.48561795e-02 3.01061779e-01
2.87171215e-01 1.03080773e+00 -4.48458046e-01 -1.02593994e+00
1.28752336e-01 8.21557362e-03 -6.96273744e-01 -3.30606364e-02
-1.76345274e-01 2.25417644e-01 -4.41094637e-01 1.18395388e+00
9.44391668e-01 -2.04251006e-01 2.17649341e-02 -2.15781271e-01
-4.47142094e-01 -2.22303301e-01 -1.25459838e+00 1.59876478e+00
3.16825539e-01 4.23312545e-01 1.96093768e-01 -7.86719859e-01
1.17962480e+00 -1.32522091e-01 5.26092410e-01 -5.39648592e-01
1.24125488e-01 3.11328441e-01 -4.50743645e-01 2.55759329e-01
6.94879115e-01 -1.80050731e-01 -2.22218320e-01 2.16558844e-01
2.62091428e-01 -6.12631500e-01 4.05074507e-01 3.46920937e-01
6.20551467e-01 1.04439203e-02 3.48117292e-01 -4.84084576e-01
2.41603494e-01 3.55116874e-01 7.98844814e-01 1.06313896e+00
-4.18951184e-01 1.15163100e+00 5.12734175e-01 -2.11373895e-01
-1.03062546e+00 -1.15097046e+00 -3.36423069e-01 1.00463569e+00
3.85982543e-01 -8.32664147e-02 -1.09669375e+00 -7.73221910e-01
-1.19453125e-01 4.11896318e-01 -4.54469949e-01 5.82085788e-01
-7.43750513e-01 -9.07249331e-01 2.90632308e-01 6.05620325e-01
6.64394200e-01 -1.21582413e+00 -5.21392822e-01 3.10623914e-01
1.26166493e-01 -1.23031962e+00 -7.15675473e-01 1.56675339e-01
-1.24437261e+00 -9.53046262e-01 -9.10753131e-01 -1.08818591e+00
8.27629507e-01 3.13855529e-01 1.15946293e+00 1.46193936e-01
-6.18412018e-01 1.53352827e-01 -4.47156549e-01 6.74823821e-02
-2.08450362e-01 2.61491001e-01 -4.43835229e-01 -6.81232885e-02
-2.80338258e-01 1.22680573e-03 -4.64414120e-01 6.42822325e-01
-7.19128966e-01 3.84608924e-01 7.06217289e-01 6.00384355e-01
1.11343670e+00 -3.41480345e-01 4.25272673e-01 -8.96796346e-01
1.61228493e-01 7.08294213e-02 -9.10085857e-01 2.64978290e-01
-3.58587027e-01 -1.42882347e-01 -3.29608135e-02 -9.41484496e-02
-1.08245730e+00 5.23512661e-01 -3.43781739e-01 -2.85765141e-01
-4.78244513e-01 -1.70162451e-02 -6.73037469e-02 -4.63703036e-01
3.51678401e-01 7.97534660e-02 -4.51005071e-01 -8.06440055e-01
8.46852183e-01 3.39431256e-01 9.07613337e-01 -7.48755693e-01
7.61229336e-01 6.67241633e-01 -1.36156335e-01 -8.82361293e-01
-7.39955068e-01 -9.09715950e-01 -1.11640477e+00 -1.66104779e-01
1.10003901e+00 -5.38619459e-01 -3.56942654e-01 5.74319899e-01
-1.24531436e+00 -5.17389417e-01 -2.14579120e-01 9.04396921e-02
-5.63210309e-01 4.85194296e-01 -8.94080877e-01 -8.67287397e-01
-3.31821471e-01 -1.27388823e+00 1.41012931e+00 4.29314941e-01
9.31942090e-03 -6.11316681e-01 -1.99151084e-01 5.88511944e-01
-7.45664462e-02 3.74959588e-01 7.00470269e-01 -4.96759415e-01
-9.89477694e-01 3.49263661e-02 -6.55287921e-01 1.13253504e-01
-3.00312400e-01 3.59854311e-01 -5.27098536e-01 -1.76766202e-01
-4.54847336e-01 -2.99258322e-01 1.34612787e+00 7.20455587e-01
1.10226190e+00 7.25982413e-02 -4.79784191e-01 7.22848833e-01
1.38032579e+00 3.87620389e-01 5.53667545e-01 1.42184958e-01
6.93810523e-01 6.46124721e-01 8.36283028e-01 2.34200284e-01
2.51177162e-01 7.85858154e-01 1.63459569e-01 -7.09352314e-01
-3.46471786e-01 2.24512015e-02 -1.80067971e-01 7.55141735e-01
-5.00592403e-02 -2.22044252e-02 -9.09670532e-01 7.69963205e-01
-1.56219280e+00 -5.01034737e-01 -4.46133733e-01 1.75770068e+00
7.53595889e-01 4.57919210e-01 4.63431358e-01 3.71002890e-02
8.59111547e-01 -2.74483841e-02 -4.36714888e-01 -3.40741187e-01
-1.89546227e-01 4.04474437e-01 5.22427201e-01 7.01945901e-01
-1.53582227e+00 1.42789006e+00 7.25995874e+00 1.24333823e+00
-6.23851538e-01 -2.97767092e-02 1.18235338e+00 5.24097979e-01
-1.18485123e-01 -1.08412556e-01 -1.22396708e+00 -3.46961766e-02
3.13544899e-01 3.67700905e-01 -2.13066176e-01 9.75152075e-01
-6.87980428e-02 -3.00556242e-01 -5.75887561e-01 6.23086989e-01
-1.61396578e-01 -1.70770550e+00 2.71131307e-01 2.50263158e-02
1.17545140e+00 -5.47918603e-02 -1.02176927e-01 2.09935065e-02
4.29585338e-01 -6.11890018e-01 1.02687562e+00 1.60591334e-01
6.59085453e-01 -8.20588887e-01 5.59869826e-01 1.20793536e-01
-1.58309793e+00 2.63841957e-01 -3.36549044e-01 6.10854685e-01
3.64850014e-01 6.10589325e-01 -6.53223038e-01 5.69823802e-01
5.07116973e-01 3.20788890e-01 -9.15870428e-01 1.53163636e+00
-9.59320590e-02 7.18312323e-01 -3.74803036e-01 2.20851332e-01
5.59988916e-01 -1.80088803e-01 4.00814176e-01 1.61005402e+00
-1.34934321e-01 4.32547241e-01 6.91203356e-01 9.33541417e-01
-1.46043777e-01 2.55357146e-01 2.69071627e-02 4.32819426e-01
4.84880626e-01 1.32935536e+00 -1.83573246e+00 -6.07182980e-01
-1.20066687e-01 9.43079889e-01 1.99846566e-01 3.08780670e-01
-6.88363850e-01 -2.17485562e-01 9.36923772e-02 -2.98553526e-01
6.18919253e-01 -5.28441727e-01 -7.69811392e-01 -7.99804389e-01
-2.98972070e-01 -6.48500741e-01 4.70513284e-01 -3.32927287e-01
-1.14501035e+00 6.36534274e-01 1.17448062e-01 -8.73177111e-01
1.79454520e-01 -4.21909928e-01 -5.90001822e-01 3.21850926e-01
-1.17452466e+00 -1.09491742e+00 -1.29556745e-01 2.24524260e-01
1.17274106e+00 2.70285904e-01 3.31305861e-01 7.64002129e-02
-6.69201374e-01 4.28846240e-01 -2.57181585e-01 3.41583282e-01
1.71832636e-01 -1.35559559e+00 9.62500155e-01 9.60602939e-01
4.21680897e-01 3.21485043e-01 3.77170593e-01 -7.86216497e-01
-6.75498128e-01 -1.17993701e+00 5.28237939e-01 -2.68536713e-02
3.10904086e-01 -3.43192548e-01 -1.01509893e+00 2.81619072e-01
5.47173470e-02 8.85452479e-02 2.72138029e-01 -1.82349339e-01
-1.51275113e-01 2.12128252e-01 -1.22174942e+00 5.39701343e-01
1.04691041e+00 -7.04410374e-02 -5.67758620e-01 1.39403611e-01
9.49100852e-01 -3.46257687e-01 -6.77151442e-01 6.22599840e-01
4.95458961e-01 -1.13814342e+00 1.12702382e+00 -2.67819792e-01
-2.19868142e-02 -3.21409047e-01 -4.68677804e-02 -7.87817419e-01
-4.27329689e-01 -5.99348068e-01 3.14365685e-01 1.29907262e+00
6.19209826e-01 -2.58230835e-01 1.20243955e+00 1.47407815e-01
-4.49844837e-01 -1.14113569e+00 -9.02587771e-01 -8.61146986e-01
8.45913813e-02 -5.40835023e-01 2.37883389e-01 5.21798491e-01
-5.00921190e-01 -4.14484292e-02 9.05313939e-02 2.01930944e-02
1.24612498e+00 4.83365655e-01 5.56814730e-01 -1.13914406e+00
-3.69227305e-02 -6.76973999e-01 -2.70200938e-01 -1.42237782e+00
-1.86265841e-01 -6.83677256e-01 5.25409698e-01 -1.67876112e+00
3.68499905e-01 -6.23562574e-01 -8.75250921e-02 4.25202459e-01
-2.26035148e-01 8.63790393e-01 3.79044384e-01 3.29316169e-01
-9.64728415e-01 3.01327944e-01 1.42103088e+00 -1.07434005e-01
-4.19978231e-01 6.56655505e-02 -4.74584937e-01 9.53710794e-01
7.07286417e-01 -4.87929881e-01 2.74270535e-01 -4.67174463e-02
-5.73634744e-01 -1.83988959e-02 2.43763939e-01 -1.06173718e+00
2.35740751e-01 -2.03609630e-01 2.89696217e-01 -1.33756411e+00
2.79393911e-01 -1.82771936e-01 -4.07170594e-01 5.32639980e-01
-1.90927193e-01 -2.03570321e-01 3.01338732e-01 4.07748491e-01
-2.08092496e-01 -1.65013418e-01 1.33991325e+00 -1.43645465e-01
-9.89039898e-01 3.16245079e-01 -3.84635121e-01 1.28521651e-01
9.61400390e-01 -3.61714751e-01 -4.95369732e-02 1.39651880e-01
-9.72830892e-01 4.35230076e-01 4.36350673e-01 1.44414157e-01
8.83312702e-01 -1.14225006e+00 -7.16051579e-01 1.03370972e-01
-1.01952828e-01 2.07522660e-01 5.44062443e-02 7.57720649e-01
-7.49357641e-01 5.17812133e-01 6.01549335e-02 -9.52508748e-01
-1.41346157e+00 2.73590773e-01 1.65412873e-01 -2.53357977e-01
-6.72757328e-01 1.27626598e+00 4.86949891e-01 -3.28500122e-01
1.38316020e-01 -6.09476447e-01 -5.42259328e-02 2.17115626e-01
8.54406431e-02 3.87911886e-01 -1.22641310e-01 -7.19253838e-01
-4.98635322e-01 1.06008434e+00 -6.40702173e-02 -4.43409652e-01
9.72030163e-01 -4.01646458e-02 -5.08749001e-02 1.10042453e-01
9.62417126e-01 -1.84519887e-01 -1.53790152e+00 -4.45498433e-03
4.69461799e-01 -3.35407436e-01 4.85335886e-02 -5.15515506e-01
-1.07714117e+00 7.72243679e-01 5.05785763e-01 2.74654895e-01
1.03381741e+00 4.21918988e-01 7.22073555e-01 6.45257235e-02
5.98422527e-01 -1.00999773e+00 3.34025055e-01 3.83703142e-01
8.30197990e-01 -1.00943637e+00 1.25116602e-01 -1.05063474e+00
-6.08126879e-01 1.07005882e+00 5.06911218e-01 -2.65731335e-01
4.49301690e-01 2.05429107e-01 5.19617461e-02 -2.98837334e-01
-3.40110481e-01 -5.70566237e-01 7.02365279e-01 3.76762986e-01
-5.03812358e-02 2.56555915e-01 -3.10919791e-01 4.17951316e-01
7.77598768e-02 -4.80386108e-01 2.78652370e-01 7.72226214e-01
-1.09919202e+00 -1.00144887e+00 -6.27376139e-01 2.77080387e-01
-3.30618054e-01 8.12380761e-02 -6.43420935e-01 9.32431638e-01
2.07867786e-01 8.03567469e-01 5.88493571e-02 -1.05452567e-01
2.34268561e-01 -5.75900637e-02 4.38666105e-01 -8.72220695e-01
-4.43537802e-01 7.71150947e-01 2.97396407e-02 -7.33354628e-01
-2.71446675e-01 -7.38009334e-01 -1.64074969e+00 2.08352134e-01
-5.85860014e-01 1.65090770e-01 6.61541581e-01 8.81157517e-01
6.25079870e-02 3.87293220e-01 5.40597558e-01 -1.35963917e+00
-8.61766040e-02 -6.94719195e-01 -5.63263237e-01 1.63343400e-01
1.59832276e-02 -5.99486947e-01 -1.82187185e-01 1.25104874e-01] | [9.435242652893066, 0.2911987602710724] |
7c490637-de0e-43ae-911a-aeacc431788d | anomaly-detection-in-surveillance-videos | 2206.01524 | null | https://arxiv.org/abs/2206.01524v2 | https://arxiv.org/pdf/2206.01524v2.pdf | Anomaly detection in surveillance videos using transformer based attention model | Surveillance footage can catch a wide range of realistic anomalies. This research suggests using a weakly supervised strategy to avoid annotating anomalous segments in training videos, which is time consuming. In this approach only video level labels are used to obtain frame level anomaly scores. Weakly supervised video anomaly detection (WSVAD) suffers from the wrong identification of abnormal and normal instances during the training process. Therefore it is important to extract better quality features from the available videos. WIth this motivation, the present paper uses better quality transformer-based features named Videoswin Features followed by the attention layer based on dilated convolution and self attention to capture long and short range dependencies in temporal domain. This gives us a better understanding of available videos. The proposed framework is validated on real-world dataset i.e. ShanghaiTech Campus dataset which results in competitive performance than current state-of-the-art methods. The model and the code are available at https://github.com/kapildeshpande/Anomaly-Detection-in-Surveillance-Videos | ['Sonali Agarwal', 'Sanjay Kumar Sonbhadra', 'Narinder Singh Punn', 'Kapil Deshpande'] | 2022-06-03 | null | null | null | null | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [ 6.75986186e-02 -3.14397842e-01 2.77651791e-02 -4.64706272e-01
-3.09620678e-01 -4.01381791e-01 4.43233341e-01 1.26133695e-01
-3.16115111e-01 4.23627734e-01 5.82917258e-02 -1.44977078e-01
1.68138579e-01 -4.81063485e-01 -7.92665601e-01 -7.81034350e-01
-3.25413138e-01 -1.81157008e-01 4.94511306e-01 -1.41510993e-01
3.59098494e-01 3.56349766e-01 -1.79346144e+00 5.31001627e-01
8.34901512e-01 1.17357981e+00 -7.83927888e-02 7.94772863e-01
-5.31867109e-02 1.11551857e+00 -4.14045393e-01 -2.70956635e-01
4.72132444e-01 -5.56149721e-01 -5.16245306e-01 3.90853077e-01
7.13426769e-01 -7.35056281e-01 -4.90040720e-01 1.17891252e+00
1.93412334e-01 2.10068479e-01 3.90022486e-01 -1.60734224e+00
-4.85829920e-01 9.08673629e-02 -6.66160822e-01 1.19600070e+00
4.55050230e-01 3.34130883e-01 6.90693200e-01 -8.53402138e-01
2.72962809e-01 8.64221632e-01 5.78120589e-01 4.74195272e-01
-5.54851055e-01 -6.05696440e-01 4.17583078e-01 8.03342938e-01
-1.11683440e+00 -3.65578294e-01 8.21084619e-01 -4.55010772e-01
9.14297044e-01 2.91026682e-01 7.41098046e-01 1.45082915e+00
2.87369907e-01 8.75664115e-01 7.20820606e-01 -1.94673285e-01
-1.54435888e-01 2.93845367e-02 2.11874664e-01 8.31198275e-01
3.19592565e-01 -7.75016621e-02 -3.28120321e-01 -4.05718498e-02
4.82434750e-01 4.55215245e-01 -3.25378209e-01 -7.71311074e-02
-1.08794188e+00 6.00783706e-01 1.57624140e-01 5.07216811e-01
-6.65600717e-01 -1.43482506e-01 8.47745240e-01 6.87008977e-01
5.55072188e-01 6.20221347e-02 -3.56490880e-01 -3.81904542e-01
-8.57090831e-01 4.82099801e-02 2.83686548e-01 7.12958038e-01
2.44539976e-01 5.32736719e-01 -1.57837898e-01 5.47664881e-01
3.05166870e-01 2.07731023e-01 5.53830385e-01 -8.19254577e-01
3.69663239e-01 5.78204513e-01 1.12199731e-01 -1.21313369e+00
-2.41848648e-01 -4.09549892e-01 -7.22509027e-01 2.32147112e-01
5.83334923e-01 -1.96364924e-01 -1.12975478e+00 1.48163235e+00
3.04224193e-01 8.42421174e-01 5.33190966e-02 1.07653451e+00
5.97451091e-01 9.51907396e-01 1.68142557e-01 -2.85128832e-01
1.16073763e+00 -1.02974927e+00 -9.64066148e-01 -3.91959734e-02
6.40152216e-01 -7.57492065e-01 1.05745625e+00 6.23783350e-01
-8.47636759e-01 -7.10256279e-01 -9.23294961e-01 3.30718488e-01
-4.61081177e-01 4.93725203e-02 3.62842947e-01 4.09576416e-01
-8.92944872e-01 3.46190393e-01 -1.06391323e+00 -7.18176842e-01
6.30752265e-01 1.73606440e-01 -5.45714915e-01 -1.66174755e-01
-1.07289147e+00 6.63172722e-01 4.42786276e-01 2.91909158e-01
-1.06126487e+00 -4.30122972e-01 -8.05774093e-01 -2.41227865e-01
5.47322154e-01 -1.33177608e-01 1.14694941e+00 -1.60143495e+00
-1.10824609e+00 7.15825677e-01 -8.28138515e-02 -7.02302694e-01
5.29515266e-01 -5.88847160e-01 -7.38594294e-01 2.72901922e-01
-1.79512084e-01 2.55010873e-01 1.03839600e+00 -7.19373405e-01
-8.47172618e-01 -2.81728148e-01 1.35581167e-02 2.17924178e-01
-4.51657027e-01 1.13717616e-01 -3.22758019e-01 -8.92326295e-01
-1.45037174e-01 -8.16898525e-01 2.62425803e-02 -2.31391743e-01
-2.01720059e-01 -1.73568219e-01 1.36336672e+00 -1.00109529e+00
1.37014949e+00 -2.16762066e+00 -2.96293646e-01 9.96090695e-02
-3.13866399e-02 6.88257575e-01 1.91409700e-02 1.89069554e-01
-3.76103044e-01 -1.75050095e-01 -4.48072255e-02 7.52447322e-02
-4.70094562e-01 2.30234414e-01 -1.48041872e-02 7.19966173e-01
5.35872817e-01 4.26155061e-01 -8.48221064e-01 -5.31815529e-01
4.82855856e-01 2.91456193e-01 -5.63343525e-01 5.02938628e-01
-7.41765946e-02 7.86929965e-01 -6.70877516e-01 1.05685151e+00
6.43853962e-01 1.07316241e-01 -5.79786181e-01 -2.16153875e-01
-1.49293497e-01 -3.34301889e-01 -9.35629129e-01 1.66117001e+00
1.10260971e-01 8.17187488e-01 -2.58098304e-01 -1.50335491e+00
6.83746874e-01 6.10876203e-01 7.15013206e-01 -5.50360084e-01
3.07157069e-01 5.82644157e-02 1.50959611e-01 -1.14232540e+00
1.84199065e-01 4.86195743e-01 2.89376497e-01 -3.54746543e-02
2.95968652e-01 7.14356124e-01 3.77075136e-01 1.15346327e-01
1.30296791e+00 3.81613165e-01 2.59583026e-01 -7.50113055e-02
1.06158745e+00 8.42743590e-02 7.09787548e-01 7.19091773e-01
-7.71285772e-01 3.84797812e-01 5.16909957e-01 -7.88112938e-01
-1.15516138e+00 -7.98216343e-01 -7.06168637e-03 1.06758571e+00
4.63799983e-02 -2.55659163e-01 -7.47229755e-01 -9.73186910e-01
-3.19882900e-01 4.88084048e-01 -7.27578640e-01 -2.73472756e-01
-6.04613841e-01 -6.23823762e-01 4.24475521e-01 4.25592631e-01
7.23766208e-01 -1.21663618e+00 -7.55045950e-01 -8.39612167e-03
-2.35587090e-01 -1.24536884e+00 -3.33670527e-01 -2.48779118e-01
-8.82207394e-01 -1.17497098e+00 -6.38957918e-01 -6.15414381e-01
6.45747900e-01 1.45425141e-01 9.90068138e-01 1.75095588e-01
-3.28653932e-01 3.76487911e-01 -7.32884586e-01 -4.56023812e-01
-3.23658377e-01 -2.09284514e-01 1.44212782e-01 3.66241097e-01
8.75803649e-01 -3.79063427e-01 -5.96650779e-01 2.42607221e-01
-9.22376096e-01 -2.55997479e-01 6.65290236e-01 6.68299377e-01
4.38719243e-01 -6.99091852e-02 5.67719281e-01 -6.77300155e-01
2.70820200e-01 -8.36937308e-01 -7.60782599e-01 8.74221325e-02
-2.81059057e-01 -2.72211164e-01 8.53427649e-01 -5.07504821e-01
-1.10288739e+00 1.14972340e-02 -2.04910353e-01 -8.51090372e-01
-7.29451060e-01 1.29424855e-01 -8.11986160e-03 1.22062802e-01
4.91491854e-01 3.18057060e-01 -1.15417950e-01 -3.14372808e-01
-2.67911196e-01 5.67576408e-01 5.63299775e-01 -1.03524886e-01
7.15396941e-01 3.81966293e-01 -1.38633803e-01 -9.82883692e-01
-7.90227711e-01 -6.31190717e-01 -5.22545278e-01 -6.26134396e-01
1.08149004e+00 -9.32934105e-01 -2.75597662e-01 7.27422237e-01
-8.45310807e-01 2.55707204e-02 -9.77231339e-02 6.92734778e-01
-4.29787368e-01 4.50502455e-01 -5.66590846e-01 -9.55971479e-01
-2.71129340e-01 -9.97277737e-01 7.21574605e-01 2.96824038e-01
5.59826531e-02 -6.92301154e-01 -1.43874325e-02 1.89966813e-01
4.03587162e-01 6.12620056e-01 2.13105023e-01 -1.06917095e+00
-5.37438512e-01 -3.72213840e-01 -4.45489883e-02 6.67813718e-01
1.17858490e-02 3.66681844e-01 -9.08769608e-01 -2.02043474e-01
6.96878582e-02 -4.65048216e-02 7.18110502e-01 5.14302433e-01
1.33700359e+00 -4.72383082e-01 2.99493112e-02 6.65629625e-01
1.32986045e+00 4.27763939e-01 6.98930919e-01 4.49246734e-01
8.53959799e-01 2.97698110e-01 1.13051701e+00 5.62071800e-01
1.72914099e-02 4.19810116e-01 6.57778978e-01 -8.17915518e-03
3.17667395e-01 1.17928728e-01 7.83065379e-01 5.78677654e-01
-3.84784311e-01 -4.91252869e-01 -8.48599076e-01 6.28706634e-01
-1.94289029e+00 -1.41005349e+00 -2.32823253e-01 1.98415816e+00
2.01397195e-01 2.49726862e-01 3.34972560e-01 3.01616877e-01
8.20895433e-01 2.52594054e-01 -4.50160801e-01 -6.01439536e-01
6.18423410e-02 -2.57936031e-01 4.02954251e-01 1.15168020e-01
-1.57800376e+00 7.69755781e-01 4.71389151e+00 5.75298846e-01
-1.09306026e+00 1.54941306e-01 7.38601685e-01 -2.24279061e-01
5.28443933e-01 -3.72523576e-01 -3.79900455e-01 9.59508538e-01
1.25255764e+00 1.91930607e-01 1.44639522e-01 9.42089200e-01
4.75290000e-01 -1.12722963e-02 -8.59115958e-01 1.02916205e+00
2.58290946e-01 -9.20366943e-01 1.16376579e-01 -2.28945628e-01
5.82718432e-01 1.91265211e-01 -9.08530429e-02 2.84928262e-01
-3.19313198e-01 -7.28723824e-01 4.76330131e-01 6.97457910e-01
2.77329177e-01 -8.05128276e-01 1.26981139e+00 7.76216388e-02
-1.08830917e+00 -3.76521289e-01 -2.88301080e-01 1.00553185e-01
1.02246478e-02 2.04612002e-01 -4.57026392e-01 4.10133570e-01
1.22718322e+00 1.11769545e+00 -6.29033566e-01 1.24180603e+00
2.95190394e-01 9.32825744e-01 -2.38989890e-01 4.12021011e-01
5.49736917e-01 -1.64516807e-01 8.39176118e-01 1.39565122e+00
6.20525122e-01 1.16732642e-01 2.07813784e-01 2.51918942e-01
2.33358443e-01 6.71217814e-02 -9.82052028e-01 -2.38197390e-02
2.68212263e-03 1.14294279e+00 -5.68899035e-01 -3.58155280e-01
-8.93026292e-01 1.13735414e+00 -2.56543368e-01 3.31077069e-01
-1.22359121e+00 -1.73415735e-01 6.17946506e-01 1.01029694e-01
3.28665018e-01 -1.49506284e-02 3.78492773e-01 -1.38074982e+00
1.70505375e-01 -1.00737536e+00 8.61266494e-01 -6.51545107e-01
-1.20862949e+00 6.53638840e-01 1.80428445e-01 -1.64765716e+00
-2.73791820e-01 -4.45043147e-01 -8.66138935e-01 3.43428671e-01
-1.29856193e+00 -9.95009482e-01 -8.84553611e-01 9.72519219e-01
1.17827582e+00 -3.65421116e-01 3.90337348e-01 6.51944995e-01
-8.66252184e-01 5.09417951e-01 -1.60450742e-01 3.81584495e-01
7.12926805e-01 -1.09955692e+00 1.69403300e-01 1.52460802e+00
-1.24381930e-02 -1.30319959e-02 1.02866018e+00 -4.97081727e-01
-9.92806017e-01 -1.11462343e+00 2.51579911e-01 -3.55295748e-01
6.31134033e-01 1.09348364e-01 -1.30991876e+00 8.84862542e-01
4.40484792e-01 5.65298080e-01 5.60244739e-01 -5.84017694e-01
7.20394775e-02 -1.57203600e-01 -1.21651292e+00 3.09280336e-01
1.03295255e+00 -1.20157808e-01 -6.17212832e-01 2.53642261e-01
3.03122938e-01 -3.20020378e-01 -5.96804857e-01 5.04187644e-01
4.37952310e-01 -1.22726297e+00 6.46693408e-01 -7.79094875e-01
2.58867979e-01 -5.37018061e-01 -4.03527580e-02 -1.03536034e+00
-1.70018151e-01 -3.94971162e-01 -6.09938800e-01 1.08675683e+00
1.52007043e-01 -5.12000561e-01 6.35244727e-01 3.31842393e-01
-4.44387078e-01 -6.89730346e-01 -7.12693572e-01 -6.33850336e-01
-7.10245788e-01 -3.99926394e-01 1.49873376e-01 1.12807977e+00
-2.34375477e-01 -1.33205771e-01 -6.32901847e-01 4.94999468e-01
5.79210043e-01 -4.77684051e-01 6.06303632e-01 -1.04799426e+00
1.39033226e-02 -5.87848797e-02 -1.05880570e+00 -4.59008932e-01
-1.12798787e-03 -3.58336627e-01 -1.91418931e-01 -9.73956525e-01
-3.56685221e-02 1.07049994e-01 -7.35541224e-01 4.13331598e-01
-7.88995177e-02 4.70466465e-01 -6.63794577e-02 1.01238228e-02
-9.19946432e-01 4.07954395e-01 8.29941809e-01 1.42156303e-01
-1.22968622e-01 8.86204019e-02 -2.07148895e-01 9.55966055e-01
1.19912791e+00 -3.25684190e-01 -4.56158787e-01 -3.38915080e-01
-2.66688973e-01 -1.20814838e-01 4.21125263e-01 -1.42126262e+00
-2.77139209e-02 -1.39377192e-01 6.29767716e-01 -6.58692122e-01
9.46824178e-02 -1.25672495e+00 -9.21463370e-02 4.98525918e-01
-2.93883055e-01 6.82641685e-01 3.40879232e-01 7.31318593e-01
-3.67025316e-01 -1.93307951e-01 7.44312108e-01 -1.88775018e-01
-1.18058491e+00 4.82431293e-01 -5.34793079e-01 -1.21340685e-01
1.47989535e+00 -3.95879656e-01 -2.08241329e-01 -4.50163037e-01
-7.93906212e-01 2.31226638e-01 3.46061736e-01 7.10101843e-01
6.31742179e-01 -1.28849113e+00 -8.75813365e-01 4.04103518e-01
2.85313606e-01 -3.73040944e-01 6.27728999e-01 1.18695831e+00
-8.16307366e-01 3.60221088e-01 -6.53862000e-01 -7.26663649e-01
-1.49580252e+00 6.39171779e-01 3.45201254e-01 3.41361910e-02
-7.18263268e-01 5.90179563e-01 -1.65509760e-01 3.53397161e-01
3.19895208e-01 -4.10943419e-01 -4.67607647e-01 -1.19770125e-01
7.23476708e-01 3.56537670e-01 -4.86612692e-02 -9.39084947e-01
-4.97086525e-01 3.20823789e-01 -2.59718508e-01 4.23176646e-01
1.21633708e+00 -2.01124877e-01 2.49435514e-01 3.98480505e-01
1.07418764e+00 -1.60910681e-01 -1.41521835e+00 -6.41733110e-02
1.06001720e-01 -8.84011686e-01 7.96981454e-02 -4.09331948e-01
-1.28273606e+00 6.66139722e-01 1.24009252e+00 3.96988869e-01
1.45713961e+00 -2.16753572e-01 7.83330441e-01 4.00063515e-01
-1.80530369e-01 -1.02343416e+00 1.67666495e-01 2.97638506e-01
7.09164262e-01 -1.69074094e+00 -2.43076384e-01 2.11055595e-02
-9.01412606e-01 9.95834708e-01 1.09617758e+00 -4.72319245e-01
4.90742952e-01 -1.86596233e-02 1.19274981e-01 -1.50567248e-01
-6.64243281e-01 -1.83348730e-01 3.51535141e-01 6.17256105e-01
4.19601083e-01 -4.64067250e-01 -1.55539200e-01 1.58825487e-01
4.10953879e-01 1.29123684e-02 6.06369138e-01 9.95863438e-01
-4.11133319e-01 -6.87152565e-01 -5.19175291e-01 7.54461288e-01
-1.17923450e+00 1.06519587e-01 4.46285792e-02 7.93191195e-01
3.15088511e-01 9.69413519e-01 2.46565893e-01 -3.97426873e-01
2.57353365e-01 2.92155445e-01 1.77369654e-01 -1.57496020e-01
-3.14134002e-01 -4.64560911e-02 1.67091098e-02 -9.95924592e-01
-7.60137200e-01 -7.54522860e-01 -9.02362645e-01 -2.04703391e-01
-1.58876240e-01 1.35773078e-01 2.81809181e-01 7.24174798e-01
2.73868173e-01 5.80481350e-01 5.96070945e-01 -7.13767648e-01
-8.89339373e-02 -1.06334007e+00 -2.45380044e-01 8.24703515e-01
8.32103431e-01 -7.43157983e-01 -4.49315250e-01 3.60030413e-01] | [7.859673976898193, 1.534198522567749] |
fc5f4f39-e0e2-4620-89c1-0add7963a59b | can-we-evaluate-domain-adaptation-models | 2305.18712 | null | https://arxiv.org/abs/2305.18712v1 | https://arxiv.org/pdf/2305.18712v1.pdf | Can We Evaluate Domain Adaptation Models Without Target-Domain Labels? A Metric for Unsupervised Evaluation of Domain Adaptation | Unsupervised domain adaptation (UDA) involves adapting a model trained on a label-rich source domain to an unlabeled target domain. However, in real-world scenarios, the absence of target-domain labels makes it challenging to evaluate the performance of deep models after UDA. Additionally, prevailing UDA methods typically rely on adversarial training and self-training, which could lead to model degeneration and negative transfer, further exacerbating the evaluation problem. In this paper, we propose a novel metric called the \textit{Transfer Score} to address these issues. The transfer score enables the unsupervised evaluation of domain adaptation models by assessing the spatial uniformity of the classifier via model parameters, as well as the transferability and discriminability of the feature space. Based on unsupervised evaluation using our metric, we achieve three goals: (1) selecting the most suitable UDA method from a range of available options, (2) optimizing hyperparameters of UDA models to prevent model degeneration, and (3) identifying the epoch at which the adapted model performs optimally. Our work bridges the gap between UDA research and practical UDA evaluation, enabling a realistic assessment of UDA model performance. We validate the effectiveness of our metric through extensive empirical studies conducted on various public datasets. The results demonstrate the utility of the transfer score in evaluating UDA models and its potential to enhance the overall efficacy of UDA techniques. | ['Lihua Xie', 'Yuecong Xu', 'Hanjie Qian', 'Jianfei Yang'] | 2023-05-30 | null | null | null | null | ['unsupervised-domain-adaptation'] | ['methodology'] | [ 2.41861045e-01 -1.86986610e-01 -2.08543897e-01 -2.89807409e-01
-8.94612491e-01 -1.02752531e+00 7.48769581e-01 1.67739801e-02
-3.52048934e-01 6.26400054e-01 6.05428480e-02 -2.61629313e-01
-5.70512190e-02 -7.14001298e-01 -6.90709651e-01 -7.65376806e-01
2.78529495e-01 3.36383790e-01 7.69759249e-03 1.32106528e-01
2.60524545e-02 5.45091510e-01 -1.02678764e+00 -7.11897686e-02
1.09810889e+00 1.13332903e+00 -2.43689548e-02 1.91658512e-01
1.01605229e-01 2.29673028e-01 -7.94672549e-01 -4.99718338e-01
4.25961018e-01 -6.35019720e-01 -7.66694129e-01 -1.87710151e-02
2.61381388e-01 -2.36974522e-01 -3.68604772e-02 1.06260598e+00
6.10956430e-01 2.26675779e-01 1.04150593e+00 -1.23952913e+00
-6.36895537e-01 2.62306839e-01 -2.41815627e-01 3.05318266e-01
-5.44621348e-02 3.69520247e-01 8.31777513e-01 -8.62693191e-01
5.46110332e-01 9.60922837e-01 8.84522676e-01 7.06873775e-01
-1.39126122e+00 -9.04816747e-01 1.34933397e-01 -3.05497367e-02
-1.50622416e+00 -5.83889246e-01 8.20346594e-01 -6.39818132e-01
6.20569289e-01 -4.13961932e-02 1.41553894e-01 1.20214915e+00
1.19963080e-01 5.87048292e-01 1.12690890e+00 -4.45598066e-01
7.04234064e-01 3.54056567e-01 -1.81894228e-02 2.07134858e-01
8.12199190e-02 2.79306412e-01 -3.68609637e-01 -2.65119374e-01
5.88088691e-01 -2.92837769e-01 8.56404305e-02 -6.89822495e-01
-8.71721745e-01 8.45861614e-01 4.15049106e-01 5.46107860e-03
-4.50426102e-01 -3.79158705e-01 4.05111492e-01 2.78469294e-01
5.76095223e-01 7.65956879e-01 -4.07503068e-01 9.44161937e-02
-5.93364179e-01 2.21833333e-01 3.10077071e-01 7.63135552e-01
5.79466522e-01 -3.38110700e-02 -3.71632904e-01 1.19305074e+00
1.51483580e-01 4.48441952e-01 6.80079818e-01 -7.20740259e-01
4.33724195e-01 6.08111322e-01 1.55451432e-01 -6.24288678e-01
-3.21216583e-01 -6.07169330e-01 -6.52767241e-01 3.51817578e-01
6.51186526e-01 -2.98661858e-01 -8.77008498e-01 2.10031986e+00
4.92396355e-01 1.41817704e-01 3.07611376e-01 7.78401673e-01
3.92420083e-01 3.93264115e-01 5.13545513e-01 8.46553966e-03
9.68477547e-01 -6.85417593e-01 -3.48954678e-01 -3.47723424e-01
7.87927866e-01 -6.55880153e-01 1.37181711e+00 -1.46668567e-03
-8.13520491e-01 -5.61413705e-01 -1.24981892e+00 4.17197019e-01
-3.91111881e-01 -3.57867568e-03 2.04935148e-01 6.89551473e-01
-7.32768416e-01 4.16023612e-01 -6.28989220e-01 -5.22601247e-01
6.45253003e-01 4.00526524e-01 -3.14674884e-01 -2.53102958e-01
-1.34822810e+00 9.49639440e-01 3.56338054e-01 -2.18371183e-01
-1.01511586e+00 -7.01021373e-01 -6.56325459e-01 7.26082027e-02
7.63437673e-02 -6.48648322e-01 1.26996291e+00 -1.12777638e+00
-1.36324930e+00 8.22476149e-01 1.49406865e-01 -6.49209619e-01
6.52154505e-01 -6.02583811e-02 -4.07393247e-01 -2.55602539e-01
1.27947658e-01 5.80895543e-01 7.74953306e-01 -1.16427338e+00
-5.54491937e-01 -4.64301646e-01 -3.37521732e-02 4.16304410e-01
-7.53948689e-01 -2.07540646e-01 -3.64362925e-01 -8.81177843e-01
-1.78003624e-01 -1.00234139e+00 -1.14657059e-01 -1.65914044e-01
-2.25634173e-01 -6.55008405e-02 7.30566025e-01 -5.51122606e-01
1.23037601e+00 -2.50049019e+00 -1.38373062e-01 3.44701141e-01
1.03172371e-02 4.56875861e-01 -3.43071669e-01 1.41877785e-01
-4.19592373e-02 5.01621626e-02 -4.24606383e-01 -5.29397503e-02
-7.42246583e-02 2.43569147e-02 -3.53869349e-01 3.56578261e-01
3.18211496e-01 6.79338157e-01 -8.19727659e-01 -4.32993501e-01
1.00065403e-01 2.96807826e-01 -5.22899210e-01 3.02205294e-01
-4.62798290e-02 6.14288270e-01 -5.45781553e-01 4.80830163e-01
5.88616908e-01 -2.56200016e-01 1.31566554e-01 6.50580898e-02
3.61575425e-01 2.57370293e-01 -1.03137195e+00 1.28782630e+00
-6.18977547e-01 5.46565115e-01 -3.35739285e-01 -9.42462027e-01
1.03225148e+00 1.31800026e-01 4.29987252e-01 -9.96031165e-01
1.34316415e-01 1.00821167e-01 1.90893523e-02 -7.45203197e-02
1.93577304e-01 -1.13131188e-01 -2.31174186e-01 4.53713417e-01
1.21882193e-01 1.87502742e-01 -2.04626590e-01 -1.03884600e-01
1.14773726e+00 7.39705190e-02 4.40869331e-01 -3.29072326e-01
2.85864770e-01 -1.44763878e-02 5.80838203e-01 7.59449542e-01
-6.25737548e-01 5.44983268e-01 2.52164960e-01 -2.55872279e-01
-1.12369072e+00 -1.38159823e+00 -3.94739956e-01 1.15918314e+00
1.83711559e-01 2.08756715e-01 -1.04196703e+00 -1.09497118e+00
1.11224525e-01 9.75292027e-01 -9.24307466e-01 -7.42259741e-01
-2.84593821e-01 -9.01842117e-01 6.87031209e-01 6.84233189e-01
5.03971577e-01 -9.23870325e-01 -2.73843735e-01 7.71920979e-02
-2.17677727e-01 -1.06513238e+00 -6.63394153e-01 2.35142499e-01
-9.27394331e-01 -9.10492897e-01 -7.74713814e-01 -7.19202995e-01
7.84710944e-01 1.09856389e-01 1.10686839e+00 -3.40386987e-01
1.58483908e-01 2.23905891e-01 -2.56116480e-01 -4.42353129e-01
-7.49054074e-01 2.92117536e-01 2.65229732e-01 7.05763996e-02
5.20430028e-01 -4.55115557e-01 -6.50561631e-01 6.85280204e-01
-9.23356235e-01 -2.37043545e-01 4.96580780e-01 8.00129354e-01
6.46535039e-01 5.39416708e-02 1.15889430e+00 -1.10772705e+00
8.80828738e-01 -6.90882444e-01 -5.87952971e-01 3.36630464e-01
-1.14317703e+00 -8.44254065e-03 8.06027532e-01 -7.61471331e-01
-1.06323874e+00 -1.95900604e-04 9.03508887e-02 -5.30107856e-01
-1.75157383e-01 4.15713608e-01 -4.88512695e-01 3.57324332e-02
1.28610611e+00 2.42882073e-02 -1.03082933e-01 -3.77931893e-01
3.09448689e-01 6.31526947e-01 6.69852972e-01 -5.55488825e-01
9.25948501e-01 1.98207051e-01 -3.17967296e-01 -4.07997847e-01
-7.42859542e-01 -4.56980556e-01 -7.76269734e-01 -2.24794358e-01
5.46166062e-01 -8.81142974e-01 5.36448918e-02 5.02679110e-01
-7.12514281e-01 -5.29245973e-01 -3.59300315e-01 3.64561617e-01
-4.44671094e-01 1.36630267e-01 -1.20395891e-01 -5.43798506e-01
-3.51558447e-01 -1.09114206e+00 6.68261349e-01 2.49071077e-01
-5.77115536e-01 -1.25230682e+00 1.63915083e-01 3.40238124e-01
3.98965955e-01 3.04327130e-01 1.17802072e+00 -1.06160283e+00
-2.44849287e-02 -3.85176063e-01 -1.09831713e-01 6.44743860e-01
3.07422727e-01 -3.96573812e-01 -1.14174008e+00 -4.50452417e-01
-2.63067633e-01 -2.77101725e-01 4.32106048e-01 4.24432516e-01
9.77673829e-01 -1.49130300e-01 -3.09847862e-01 5.24258733e-01
1.21402419e+00 4.59338546e-01 5.08957446e-01 5.85041463e-01
4.88616288e-01 3.44692528e-01 7.43953347e-01 3.19134206e-01
7.77293295e-02 7.65380442e-01 3.19686234e-01 -1.23656489e-01
-2.81729817e-01 -4.45996016e-01 3.81516010e-01 4.57387835e-01
4.30691421e-01 -2.58434325e-01 -1.21501863e+00 8.60950649e-01
-1.49588418e+00 -5.21043003e-01 5.31192362e-01 2.51326942e+00
8.60291302e-01 2.55952179e-01 2.86382258e-01 -4.20832336e-02
8.22583616e-01 -2.33376786e-01 -1.10536790e+00 -3.19696486e-01
-1.01959340e-01 9.15255696e-02 4.87731606e-01 2.32077211e-01
-1.25865805e+00 8.78368080e-01 6.49946499e+00 8.49248767e-01
-1.18049967e+00 1.03140190e-01 7.51491547e-01 -2.67729014e-02
-1.28027990e-01 -2.04987481e-01 -6.39296472e-01 6.22056782e-01
9.48437572e-01 -4.27257627e-01 4.06443000e-01 1.03046560e+00
4.46862280e-02 2.68326074e-01 -1.24123073e+00 6.44994259e-01
-1.15830056e-01 -9.87701774e-01 6.74522966e-02 1.89789504e-01
9.95327413e-01 6.03853986e-02 3.05531889e-01 6.02193475e-01
5.51041424e-01 -9.46753085e-01 5.49374521e-01 5.67358322e-02
9.78426278e-01 -8.12014043e-01 6.82373047e-01 1.94682822e-01
-7.11818457e-01 -1.57630220e-01 -1.00545757e-01 2.97860861e-01
-3.17115813e-01 2.85153180e-01 -1.28799057e+00 2.39500180e-01
5.50528228e-01 3.73688728e-01 -6.87412083e-01 1.06284070e+00
-4.62616496e-02 8.99028957e-01 -2.40201041e-01 2.35577047e-01
2.78606951e-01 -1.76501051e-02 4.64174658e-01 1.12042189e+00
2.09763750e-01 -1.94949865e-01 -3.67948636e-02 8.17238867e-01
-4.54560876e-01 3.17776710e-01 -7.12974310e-01 1.66856553e-02
9.19187069e-01 6.71273053e-01 -5.81231236e-01 -5.41784428e-02
-2.96886593e-01 8.41687083e-01 3.13191414e-01 5.75612724e-01
-8.00610065e-01 -2.38410696e-01 8.26180160e-01 1.88810840e-01
1.08995706e-01 1.49062037e-01 -7.28140056e-01 -8.07720304e-01
1.14651993e-01 -9.48267341e-01 6.86277449e-01 -3.81294161e-01
-1.52696621e+00 6.01954222e-01 7.55800083e-02 -1.54061484e+00
-3.62525523e-01 -2.64455706e-01 -6.22179389e-01 1.02725399e+00
-1.30405784e+00 -1.02665567e+00 -2.89733887e-01 5.23986936e-01
4.43185061e-01 -4.78854984e-01 8.92469645e-01 3.55188847e-01
-7.33691394e-01 1.24001479e+00 6.07389033e-01 1.10607773e-01
1.01483536e+00 -1.03564918e+00 7.00249553e-01 8.65006864e-01
-1.39163792e-01 4.39013034e-01 6.45549595e-01 -3.77319366e-01
-7.65251160e-01 -1.39082980e+00 4.75087494e-01 -6.62058771e-01
4.57062274e-01 -2.85478830e-01 -1.18732429e+00 5.53290188e-01
-3.23977053e-01 -4.38034274e-02 8.50059211e-01 2.55821139e-01
-5.26531279e-01 -1.96559191e-01 -1.38730991e+00 5.87053299e-01
7.63405740e-01 -5.11589825e-01 -3.92567784e-01 1.39842093e-01
4.31955427e-01 -2.08598092e-01 -1.03256953e+00 5.98151922e-01
5.63620448e-01 -5.63681066e-01 1.01721346e+00 -8.20615351e-01
3.26007009e-01 -5.41271232e-02 -1.52935833e-01 -1.64042985e+00
-4.95722681e-01 -1.11308411e-01 2.70820092e-02 1.48912084e+00
6.25511169e-01 -7.18439460e-01 6.80822432e-01 8.20733964e-01
1.17138945e-01 -6.75145507e-01 -9.42145467e-01 -1.00888968e+00
5.98512769e-01 -2.46703342e-01 8.03057909e-01 1.32783210e+00
-3.35394263e-01 3.63726914e-01 1.13139357e-02 3.09420139e-01
5.56154430e-01 -1.52444214e-01 7.45988548e-01 -1.24224412e+00
-8.38872865e-02 -6.14896238e-01 -2.96925724e-01 -6.47122741e-01
3.17259490e-01 -9.10542071e-01 1.02406614e-01 -1.17637849e+00
1.09804243e-01 -7.88366258e-01 -7.31531203e-01 4.73160446e-01
-2.97884762e-01 1.30505249e-01 5.79591021e-02 5.65912604e-01
-4.24397647e-01 4.91988361e-01 9.19943213e-01 -1.43901542e-01
-3.94257575e-01 7.58842751e-02 -8.23334396e-01 6.57691836e-01
8.85806620e-01 -6.21527493e-01 -6.42672062e-01 -4.31094468e-01
-3.09822232e-01 -2.57348776e-01 2.94834793e-01 -9.78034139e-01
-1.18331775e-01 -3.61168832e-01 6.18127882e-01 7.54051283e-02
-1.40814306e-02 -8.06181312e-01 -5.74059449e-02 2.05159754e-01
-4.77126688e-01 -1.87546045e-01 4.18230891e-01 5.59497237e-01
-1.22538142e-01 -2.15608165e-01 1.27067733e+00 3.57586890e-01
-8.93739462e-01 1.83493972e-01 -4.95446175e-02 3.14219385e-01
1.11832333e+00 -1.86755762e-01 -2.10517302e-01 -1.96830228e-01
-5.43674946e-01 2.08570033e-01 7.44101226e-01 6.25422359e-01
1.70785412e-01 -1.54576147e+00 -6.32402956e-01 2.96558440e-01
6.34147406e-01 -5.53160906e-02 2.54408687e-01 3.19744438e-01
-1.97278515e-01 1.66968301e-01 -3.49989980e-01 -5.49724162e-01
-1.00300622e+00 5.65916300e-01 5.42816281e-01 -3.84821624e-01
-3.89478177e-01 5.75560987e-01 6.88442409e-01 -5.14701247e-01
3.52528036e-01 2.32547939e-01 -1.17151223e-01 -1.95510045e-01
3.24669391e-01 2.29742453e-01 3.36700618e-01 -6.21658027e-01
-3.44611377e-01 1.55985430e-01 -1.81242421e-01 4.49865200e-02
9.25927162e-01 -1.47536471e-01 6.29163980e-01 1.23370074e-01
1.14734542e+00 -1.96864501e-01 -1.59562683e+00 -5.55582285e-01
-7.20375329e-02 -1.88253164e-01 2.38291677e-02 -1.16736305e+00
-8.46138775e-01 7.11862981e-01 9.39273059e-01 4.79851626e-02
1.34244370e+00 -1.44486547e-01 6.69939935e-01 1.01700678e-01
7.87861049e-02 -1.20106268e+00 6.17477521e-02 3.62959236e-01
7.07734525e-01 -1.29591703e+00 -2.10666716e-01 3.50822583e-02
-8.74955595e-01 5.02896309e-01 8.02429497e-01 1.11779734e-01
5.18903196e-01 1.22960499e-02 4.78303105e-01 1.84662193e-01
-4.58688408e-01 1.24701373e-01 3.73275250e-01 7.73089528e-01
7.41024837e-02 1.05356425e-01 -2.87627690e-02 9.05668259e-01
-2.06418142e-01 -6.76701069e-02 6.99159130e-02 7.62811303e-01
-1.85783617e-02 -1.06747019e+00 -3.93082261e-01 3.46570194e-01
-4.06199276e-01 1.48447633e-01 -5.50953507e-01 7.71031320e-01
-6.11883663e-02 8.33825111e-01 -6.78065419e-02 -5.03191233e-01
7.20908999e-01 4.19979423e-01 1.70517385e-01 -4.78600025e-01
-4.21326190e-01 -2.33252272e-01 -1.98663220e-01 -1.44899219e-01
4.87450622e-02 -7.81980217e-01 -9.49259222e-01 -1.76055059e-01
-2.79537857e-01 4.06659506e-02 6.71662807e-01 1.01846063e+00
5.76758742e-01 4.02587682e-01 9.05519366e-01 -3.69082481e-01
-9.12305415e-01 -1.13011277e+00 -4.12660629e-01 8.50755453e-01
1.75230131e-01 -9.32573855e-01 -1.65683493e-01 1.80548117e-01] | [10.213519096374512, 3.0953054428100586] |
d9ecd320-035d-435c-b8ff-059ba61cef2b | generalized-face-anti-spoofing-via-multi-task | 2211.15955 | null | https://arxiv.org/abs/2211.15955v1 | https://arxiv.org/pdf/2211.15955v1.pdf | Generalized Face Anti-Spoofing via Multi-Task Learning and One-Side Meta Triplet Loss | With the increasing variations of face presentation attacks, model generalization becomes an essential challenge for a practical face anti-spoofing system. This paper presents a generalized face anti-spoofing framework that consists of three tasks: depth estimation, face parsing, and live/spoof classification. With the pixel-wise supervision from the face parsing and depth estimation tasks, the regularized features can better distinguish spoof faces. While simulating domain shift with meta-learning techniques, the proposed one-side triplet loss can further improve the generalization capability by a large margin. Extensive experiments on four public datasets demonstrate that the proposed framework and training strategies are more effective than previous works for model generalization to unseen domains. | ['Shang-Hong Lai', 'Chien-Yi Wang', 'Chu-Chun Chuang'] | 2022-11-29 | null | null | null | null | ['face-parsing', 'face-anti-spoofing'] | ['computer-vision', 'computer-vision'] | [ 4.67583716e-01 -1.24827541e-01 -5.34739316e-01 -2.77509809e-01
-4.14793104e-01 -3.78772378e-01 5.46956897e-01 -5.83924830e-01
4.62795943e-02 4.97341603e-01 -2.29509428e-01 -4.00391161e-01
2.04341486e-01 -7.29686439e-01 -5.01564622e-01 -9.48852956e-01
-2.08414420e-01 1.56142071e-01 2.32547998e-01 -1.43380269e-01
2.10315511e-01 8.47031057e-01 -1.45789135e+00 5.00054777e-01
6.93739951e-01 1.27805686e+00 -2.49151081e-01 4.65553224e-01
-2.93959421e-03 3.58250320e-01 -7.56735742e-01 -6.48740828e-01
4.09980029e-01 -9.58214998e-02 -6.34192646e-01 1.75124452e-01
4.51764673e-01 -6.51505172e-01 -5.49990416e-01 1.09384263e+00
6.73113942e-01 -3.43992412e-01 3.72679532e-01 -1.65384197e+00
-5.20471692e-01 6.57816380e-02 -8.65456879e-01 2.74141073e-01
4.96266782e-01 7.39415213e-02 1.57858357e-01 -6.74242735e-01
2.93014675e-01 2.01015639e+00 6.16840661e-01 1.12212098e+00
-6.00101113e-01 -1.45699501e+00 1.19735841e-02 4.50325578e-01
-1.35411632e+00 -9.00670290e-01 1.04716241e+00 -1.04273409e-01
6.63339376e-01 -2.34496444e-01 8.07343647e-02 1.38195264e+00
2.34540552e-03 6.18780136e-01 1.29083610e+00 -3.18176985e-01
-8.60137567e-02 3.10699016e-01 -2.66864806e-01 9.04937088e-01
4.53274816e-01 5.67586780e-01 -6.10994577e-01 -4.45080012e-01
6.76811576e-01 -2.66949981e-02 -1.99168578e-01 -3.57741863e-01
-4.69507843e-01 8.48982275e-01 2.98726887e-01 -1.56925201e-01
-4.17407043e-02 4.51607164e-03 6.98386252e-01 2.48642668e-01
7.06485748e-01 -2.56836742e-01 -5.93352616e-01 4.07334000e-01
-7.73717880e-01 5.56618311e-02 5.86881936e-01 9.28482831e-01
6.97824359e-01 9.33564901e-02 3.45029563e-01 6.43676639e-01
4.03965831e-01 9.95129287e-01 2.31780738e-01 -6.63493335e-01
6.37544870e-01 1.54698282e-01 -2.09529281e-01 -1.34271789e+00
-1.72390208e-01 -2.30829328e-01 -7.16300726e-01 2.67122723e-02
2.06706554e-01 -2.63932943e-01 -1.08314478e+00 1.75937319e+00
8.73783648e-01 8.24810445e-01 9.21811387e-02 2.41847664e-01
6.14439309e-01 5.17823219e-01 1.80025354e-01 -5.16463459e-01
1.47524285e+00 -8.72076035e-01 -7.92796493e-01 -3.83666843e-01
5.43579876e-01 -8.08039904e-01 4.45455521e-01 2.81025946e-01
-5.70204139e-01 -4.69054520e-01 -1.02230883e+00 2.27062002e-01
-2.42732003e-01 -1.82676494e-01 8.26790512e-01 1.58951557e+00
-7.98961103e-01 2.58457422e-01 -6.41260982e-01 -2.56525993e-01
1.22305429e+00 7.12235570e-01 -5.36796510e-01 -3.88298422e-01
-1.07125592e+00 3.39704752e-01 4.42939878e-01 -5.93514815e-02
-1.30469978e+00 -4.79813069e-01 -8.60054195e-01 -1.34090200e-01
2.64528215e-01 -4.43487555e-01 8.35989356e-01 -7.94075310e-01
-1.53513801e+00 1.18381643e+00 -5.28894186e-01 -4.63944227e-01
4.11177933e-01 7.58765861e-02 -9.57552314e-01 5.38515210e-01
-1.41729832e-01 6.46623194e-01 1.70383120e+00 -1.37129784e+00
-5.47269344e-01 -9.39527154e-01 -1.07121363e-01 -3.44879538e-01
-6.11096621e-01 2.92889506e-01 -2.39332840e-01 -6.13558412e-01
1.32160902e-01 -9.28757012e-01 2.40644246e-01 3.25271674e-02
-9.88307744e-02 -2.32209116e-01 1.65408146e+00 -8.55620325e-01
1.02121508e+00 -2.19841671e+00 -3.83752316e-01 5.75908422e-02
9.51970965e-02 7.65664995e-01 -3.52338195e-01 -7.08455145e-02
-4.18467745e-02 2.59207159e-01 -1.83503896e-01 -4.12011534e-01
-5.21280408e-01 -4.34263609e-03 -3.89723152e-01 8.67178380e-01
2.51472622e-01 4.51712132e-01 -7.42415428e-01 -9.41240430e-01
1.38293281e-01 7.31619954e-01 -7.28759825e-01 4.28329930e-02
-8.04321170e-02 7.68906951e-01 -4.54273731e-01 1.00649285e+00
1.73282814e+00 6.43740073e-02 5.87690286e-02 1.10104963e-01
7.14967132e-01 1.48157313e-01 -8.48459542e-01 1.39832520e+00
-5.71882725e-01 3.85774761e-01 5.25056005e-01 -9.40294206e-01
9.30163741e-01 3.31019521e-01 3.77620339e-01 -5.89875221e-01
2.53267378e-01 1.22415312e-01 -3.43475670e-01 -7.72036910e-01
-1.38147265e-01 -1.40621141e-01 2.24595591e-01 2.57898688e-01
2.07411990e-01 2.16009036e-01 -3.89916331e-01 3.65173928e-02
7.35691786e-01 7.58637441e-03 -1.37231141e-01 -1.96395949e-01
1.06186438e+00 -4.77322400e-01 9.00717139e-01 3.32664192e-01
-7.63192475e-01 -5.77952489e-02 4.12324578e-01 -4.38271999e-01
-6.84393287e-01 -7.38855064e-01 -4.32590485e-01 1.22120261e+00
3.09391260e-01 -2.16179818e-01 -1.08034742e+00 -1.28146541e+00
1.88913196e-01 9.31412876e-02 -4.23210144e-01 -2.41252676e-01
-7.27640271e-01 -7.15798080e-01 8.66371214e-01 1.60842031e-01
9.27882433e-01 -7.10582316e-01 -3.39317322e-03 -2.69561231e-01
-5.92742488e-02 -1.46367073e+00 -1.47242576e-01 -4.42090660e-01
-9.26871538e-01 -1.19014800e+00 -5.61412632e-01 -1.14184487e+00
6.56446397e-01 6.28978968e-01 4.87571925e-01 5.01667023e-01
-2.33983949e-01 9.05106738e-02 -2.76789337e-01 -5.08940458e-01
-5.67281485e-01 -2.06921712e-01 3.95019114e-01 2.53270745e-01
5.01382470e-01 -4.80932951e-01 -6.26497507e-01 4.31068391e-01
-7.89244652e-01 -3.88242513e-01 2.35157743e-01 8.40114594e-01
-3.05265188e-02 3.17286104e-01 7.41081238e-01 -9.58871186e-01
2.41606966e-01 -6.06230259e-01 -4.84231651e-01 3.44638586e-01
-7.70064890e-01 -2.00582266e-01 4.82928216e-01 -5.47475815e-01
-1.32624614e+00 -2.03464124e-02 -8.18795338e-02 -5.81290424e-01
-2.25597158e-01 -5.84314764e-01 -8.16038191e-01 -7.05448031e-01
5.66489458e-01 2.95029759e-01 2.95566827e-01 -5.94982982e-01
1.38607726e-01 1.10080838e+00 3.35912704e-01 -4.40805674e-01
1.07058418e+00 9.42892969e-01 3.09320331e-01 -1.06628370e+00
-3.18387747e-01 -2.87208676e-01 -3.86502475e-01 1.61037177e-01
4.05633032e-01 -9.72294569e-01 -1.07077861e+00 1.04483640e+00
-1.42438567e+00 2.35120654e-01 5.18800318e-01 1.76030040e-01
-3.01117212e-01 1.04331791e+00 -7.13292480e-01 -9.26824689e-01
-3.04361433e-01 -1.22878873e+00 1.12008560e+00 2.17978761e-01
5.77740371e-01 -1.09801447e+00 -2.33338863e-01 6.28435314e-01
2.52335966e-01 8.08697864e-02 7.92869985e-01 -7.42768824e-01
-3.54137599e-01 -2.96511501e-01 -3.03876013e-01 5.57594121e-01
3.58121932e-01 -8.72913823e-02 -1.55363023e+00 -6.99635804e-01
4.48136717e-01 -4.02159810e-01 1.11223066e+00 1.36105314e-01
1.45583606e+00 -5.60609162e-01 -7.87533581e-01 1.16895998e+00
1.09191597e+00 2.61560202e-01 5.42684555e-01 1.31313443e-01
7.27471650e-01 9.39843118e-01 5.60035765e-01 6.38537884e-01
1.82486102e-01 3.66978198e-01 7.04000235e-01 1.11792341e-01
-1.00597776e-01 -4.92683619e-01 5.24007320e-01 -7.72345737e-02
4.09190953e-01 -3.65790457e-01 -7.33308554e-01 1.96638197e-01
-1.04289579e+00 -1.08914363e+00 4.36568767e-01 2.00822306e+00
3.21214139e-01 -8.91644359e-02 1.04163289e-01 3.44690591e-01
1.27153420e+00 5.41703463e-01 -6.41718030e-01 -3.03961307e-01
-1.29959166e-01 1.59663498e-01 5.60995519e-01 5.14722049e-01
-1.32493412e+00 1.34893322e+00 6.60673618e+00 1.14893353e+00
-1.23369181e+00 3.02152097e-01 8.33836675e-01 2.29154766e-01
-6.90824464e-02 -5.81775010e-02 -1.19499028e+00 6.17424965e-01
7.87881315e-01 1.93855733e-01 6.81097269e-01 9.57906067e-01
-1.47647053e-01 3.54106009e-01 -8.11442196e-01 1.37163699e+00
3.69478226e-01 -1.13680458e+00 2.30339214e-01 2.47626603e-01
5.97823501e-01 -3.49984497e-01 4.46577311e-01 1.19650669e-01
-3.53257656e-02 -1.07626712e+00 1.82261422e-01 -3.98989767e-01
1.16094089e+00 -7.85012424e-01 4.82004911e-01 2.29851902e-01
-1.21489906e+00 -5.72609544e-01 -6.09619141e-01 3.08053106e-01
7.09664226e-02 4.44235504e-01 -9.38151240e-01 3.46907318e-01
6.01880312e-01 6.40795469e-01 -3.87131631e-01 4.99340832e-01
-2.56326973e-01 4.87435311e-01 -2.97060102e-01 3.72428417e-01
-2.41190538e-01 3.92324775e-01 5.24483323e-01 1.14724112e+00
2.12997034e-01 1.08253350e-02 -3.26395445e-02 2.11046070e-01
-3.37623864e-01 -7.71756172e-02 -7.64254868e-01 -2.98595354e-02
8.21378946e-01 1.06730378e+00 -3.99668753e-01 -1.18157543e-01
-1.74270973e-01 1.14238727e+00 -8.65442827e-02 3.23232114e-01
-8.50548506e-01 -3.06468695e-01 9.30960774e-01 2.20498636e-01
3.11243594e-01 8.06619450e-02 -4.82213125e-02 -9.94592786e-01
-5.60552366e-02 -1.06998444e+00 5.37698925e-01 -6.20805174e-02
-1.11135280e+00 6.29910409e-01 -2.06569031e-01 -8.66279781e-01
-1.88212693e-01 -9.55413401e-01 -4.81448352e-01 6.20022476e-01
-1.96531928e+00 -1.48237133e+00 -2.47121334e-01 1.02395666e+00
3.25745970e-01 -5.94357193e-01 6.82096362e-01 3.51820081e-01
-7.97229171e-01 1.08474040e+00 -1.22963451e-01 3.23158294e-01
7.19032288e-01 -2.89953887e-01 4.20272887e-01 6.82794094e-01
-2.36524537e-01 5.35508513e-01 4.00068015e-01 -7.30105162e-01
-1.45401287e+00 -9.57382321e-01 5.21630645e-01 -9.48264748e-02
3.49806339e-01 -4.31044221e-01 -6.27116621e-01 5.79974532e-01
-3.80642675e-02 4.31710660e-01 6.89448714e-01 -2.54861832e-01
-1.02570677e+00 -4.74895060e-01 -2.08627248e+00 -4.44643460e-02
1.34935164e+00 -9.18541014e-01 -4.47750241e-02 5.43993235e-01
7.57962763e-01 -2.10261032e-01 -5.02402127e-01 7.02577531e-01
6.48639441e-01 -1.01714969e+00 1.20338905e+00 -8.22863340e-01
-1.41308457e-02 1.95065573e-01 -1.07909329e-01 -7.43908226e-01
-1.02378123e-01 -1.01247931e+00 -4.89110321e-01 1.42908490e+00
-9.84665751e-02 -9.53172088e-01 1.17371368e+00 -4.40500043e-02
4.47435409e-01 -5.62863946e-01 -1.29797363e+00 -9.89824891e-01
2.86569029e-01 -2.58449972e-01 1.05398381e+00 8.34449470e-01
-1.95198908e-01 -1.12054490e-01 -5.76865256e-01 7.20912158e-01
1.13481474e+00 -4.31528181e-01 7.20954001e-01 -1.27095103e+00
-9.18388739e-02 -2.07224954e-02 -5.50870478e-01 -1.37179923e+00
7.88920760e-01 -6.50130570e-01 -4.32191342e-01 -6.83506966e-01
1.14211543e-02 -5.85900903e-01 -2.87073284e-01 1.74637616e-01
3.69517356e-02 1.72758982e-01 -2.97924615e-02 1.92202121e-01
-1.63658276e-01 3.66459042e-01 1.22115803e+00 -3.22875500e-01
8.57699364e-02 1.95990220e-01 -4.92714465e-01 7.06130862e-01
7.97906101e-01 -7.39769399e-01 -4.94136572e-01 -5.34242213e-01
-5.10402441e-01 2.56541044e-01 4.15176988e-01 -1.06315601e+00
1.34979188e-01 -2.18627051e-01 5.81091285e-01 -2.99024284e-01
5.09349644e-01 -9.41601932e-01 -3.78790617e-01 9.29526687e-01
1.30171567e-01 -2.20083475e-01 4.47843850e-01 7.47673571e-01
2.23019216e-02 7.75812864e-02 1.26819313e+00 1.17958255e-01
-5.72434664e-01 8.46143007e-01 2.58056492e-01 -8.78496468e-02
1.29166627e+00 -4.30966109e-01 -6.33291125e-01 -1.88356519e-01
-4.46337104e-01 -9.64461640e-02 5.56801796e-01 5.11050165e-01
8.33418071e-01 -1.03750479e+00 -7.07546592e-01 8.66344631e-01
3.40498909e-02 -2.93140769e-01 5.19929409e-01 3.50432009e-01
-5.58466375e-01 5.12339592e-01 -2.60255545e-01 -6.79695964e-01
-1.58145666e+00 1.04894042e+00 3.03735435e-01 -1.49184480e-01
-8.40694904e-02 1.26729476e+00 3.80234092e-01 -2.56356895e-01
3.32131177e-01 5.06087899e-01 5.00733149e-04 -2.12413251e-01
9.83100951e-01 5.98836482e-01 -1.22514844e-01 -9.48622465e-01
-6.74587846e-01 6.89681709e-01 -2.63636768e-01 1.24987960e-01
1.05828226e+00 -2.73353368e-01 -1.35339886e-01 -6.45156741e-01
1.53591621e+00 1.38889715e-01 -1.38727582e+00 -3.25383514e-01
-2.01665178e-01 -9.62498665e-01 -3.58899459e-02 -4.58802342e-01
-1.33510804e+00 1.22649217e+00 1.04427516e+00 -3.11544295e-02
1.48066878e+00 5.30936606e-02 1.04004550e+00 2.22891539e-01
6.33795619e-01 -7.92217910e-01 2.84710944e-01 2.09817976e-01
4.58386034e-01 -1.45843637e+00 -1.77429438e-01 -1.01448619e+00
-3.60870808e-01 9.42738831e-01 5.92915952e-01 3.36542688e-02
9.47377443e-01 2.93891132e-01 -1.34997815e-01 1.34960502e-01
-3.29007298e-01 2.75388092e-01 -2.06667691e-01 1.34388828e+00
-1.34354249e-01 -9.22410190e-02 1.81783080e-01 2.22670063e-01
1.18550256e-01 -6.89279428e-03 1.71971265e-02 7.55391359e-01
-5.93104899e-01 -1.41568780e+00 -7.03411460e-01 4.71763052e-02
-8.65603507e-01 9.86949950e-02 -2.51907408e-01 4.13915247e-01
3.38037431e-01 1.32344961e+00 -9.65962633e-02 -7.44700313e-01
-2.59744346e-01 -3.50785293e-02 7.50088394e-01 -4.53244448e-01
-9.28591192e-02 -2.84867257e-01 -2.73314476e-01 -5.01966655e-01
-3.99581641e-01 -3.63728195e-01 -1.13708901e+00 -9.33065236e-01
-5.62092423e-01 -4.82146144e-02 8.19947779e-01 8.29633594e-01
5.89076519e-01 -2.31221437e-01 1.50030458e+00 -6.25796437e-01
-9.14844513e-01 -7.60393441e-01 -5.46995819e-01 4.61530268e-01
8.42689574e-01 -8.27175081e-01 -6.63910151e-01 -3.60199064e-02] | [13.003393173217773, 1.237642765045166] |
00504c38-7c04-4e43-9282-60868b8d6ef5 | parametric-implicit-face-representation-for-1 | 2306.07579 | null | https://arxiv.org/abs/2306.07579v1 | https://arxiv.org/pdf/2306.07579v1.pdf | Parametric Implicit Face Representation for Audio-Driven Facial Reenactment | Audio-driven facial reenactment is a crucial technique that has a range of applications in film-making, virtual avatars and video conferences. Existing works either employ explicit intermediate face representations (e.g., 2D facial landmarks or 3D face models) or implicit ones (e.g., Neural Radiance Fields), thus suffering from the trade-offs between interpretability and expressive power, hence between controllability and quality of the results. In this work, we break these trade-offs with our novel parametric implicit face representation and propose a novel audio-driven facial reenactment framework that is both controllable and can generate high-quality talking heads. Specifically, our parametric implicit representation parameterizes the implicit representation with interpretable parameters of 3D face models, thereby taking the best of both explicit and implicit methods. In addition, we propose several new techniques to improve the three components of our framework, including i) incorporating contextual information into the audio-to-expression parameters encoding; ii) using conditional image synthesis to parameterize the implicit representation and implementing it with an innovative tri-plane structure for efficient learning; iii) formulating facial reenactment as a conditional image inpainting problem and proposing a novel data augmentation technique to improve model generalizability. Extensive experiments demonstrate that our method can generate more realistic results than previous methods with greater fidelity to the identities and talking styles of speakers. | ['Guanbin Li', 'Yipeng Qin', 'Peiwen Lai', 'Ricong Huang'] | 2023-06-13 | parametric-implicit-face-representation-for | http://openaccess.thecvf.com//content/CVPR2023/html/Huang_Parametric_Implicit_Face_Representation_for_Audio-Driven_Facial_Reenactment_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Huang_Parametric_Implicit_Face_Representation_for_Audio-Driven_Facial_Reenactment_CVPR_2023_paper.pdf | cvpr-2023-1 | ['image-inpainting'] | ['computer-vision'] | [ 3.71551871e-01 3.98323536e-01 -1.06929906e-01 -4.96736318e-01
-5.16707659e-01 -3.93217444e-01 5.64643145e-01 -7.07668066e-01
1.23424686e-01 6.02997541e-01 4.10897911e-01 1.75406694e-01
-1.81216896e-01 -6.54224813e-01 -8.01084995e-01 -7.83194780e-01
1.26262024e-01 7.70464838e-02 -2.11753815e-01 -4.52463627e-01
-1.46409739e-02 7.75406420e-01 -1.79796457e+00 2.20697135e-01
8.36661577e-01 1.23802888e+00 -2.94124722e-01 4.84272599e-01
-1.87936246e-01 6.94924474e-01 -5.32468557e-01 -4.61820185e-01
2.77388453e-01 -3.76086831e-01 -4.14956748e-01 3.52476865e-01
3.58000070e-01 -3.47217143e-01 -1.22568622e-01 8.10295284e-01
4.25133437e-01 7.30167255e-02 6.12900376e-01 -1.43688059e+00
-6.94957316e-01 2.66377777e-01 -8.06700528e-01 -5.53088009e-01
4.80813980e-01 -8.50527436e-02 7.32396841e-01 -9.11019623e-01
4.99513954e-01 1.57485545e+00 7.36195803e-01 8.34710777e-01
-1.40740669e+00 -9.84918058e-01 1.83062136e-01 9.83776972e-02
-1.74059629e+00 -8.50545108e-01 1.30775619e+00 -3.38816077e-01
1.46007746e-01 5.06201506e-01 7.01108217e-01 1.18775916e+00
-2.15323716e-01 4.96015400e-01 1.12199295e+00 -4.97341901e-01
1.95049778e-01 2.79373020e-01 -4.71595019e-01 7.47431040e-01
-3.34835321e-01 6.21074513e-02 -6.63545489e-01 -3.89231235e-01
1.32050276e+00 -1.78864345e-01 -4.75941539e-01 -4.10089195e-01
-6.97791100e-01 8.49401236e-01 3.01783979e-01 7.98836052e-02
-3.01074088e-01 2.52764106e-01 2.06060752e-01 1.83660328e-01
4.44246173e-01 2.18071043e-01 2.96829967e-03 8.64694491e-02
-9.00509953e-01 3.26234549e-01 5.46148896e-01 9.09302711e-01
8.89258265e-01 3.14729273e-01 3.11018191e-02 1.09704614e+00
4.26677585e-01 3.44399691e-01 4.25830901e-01 -1.28665674e+00
1.48930371e-01 3.04731607e-01 1.07678942e-01 -1.23098230e+00
-5.03517538e-02 -1.64147854e-01 -7.61176169e-01 4.03779328e-01
8.58040247e-03 -6.60655499e-02 -7.77821541e-01 2.18403840e+00
5.29995680e-01 5.52104652e-01 -1.62999272e-01 9.64714050e-01
7.03150153e-01 9.31342006e-01 -8.23840946e-02 -3.50911438e-01
1.25607634e+00 -7.45245814e-01 -1.02946377e+00 2.27635965e-01
1.24524646e-01 -6.64330184e-01 1.13907707e+00 4.12913591e-01
-1.21420693e+00 -6.43237889e-01 -1.08816171e+00 -1.59517974e-01
4.13699411e-02 5.85276932e-02 5.63605189e-01 8.04070890e-01
-1.12913406e+00 4.04251456e-01 -4.98021722e-01 5.00853136e-02
1.30355909e-01 5.66906691e-01 -4.80625480e-01 3.46913934e-01
-1.10306597e+00 4.82636333e-01 -8.87584314e-02 2.61630207e-01
-6.00906432e-01 -7.80249059e-01 -8.92910659e-01 2.77133379e-02
4.87995893e-01 -7.54455566e-01 1.11504269e+00 -1.87132370e+00
-2.38499165e+00 7.61690438e-01 -1.29126713e-01 -4.02413011e-02
5.42308569e-01 -2.19049394e-01 -3.19405437e-01 2.47017145e-01
-2.72578508e-01 8.47236097e-01 1.32768524e+00 -1.56254971e+00
8.95094406e-03 -2.22660244e-01 1.15063757e-01 2.23348171e-01
-4.98774797e-01 -3.75236422e-02 -4.44421411e-01 -9.52743769e-01
1.20759517e-01 -9.69717860e-01 -9.66338813e-02 7.03203738e-01
-2.26069137e-01 7.75954351e-02 1.07644987e+00 -6.36934578e-01
8.54113162e-01 -2.10016418e+00 4.63707596e-01 3.38383794e-01
1.29246697e-01 2.57325530e-01 -1.90076590e-01 1.41748324e-01
-2.91423470e-01 3.36331911e-02 -2.13091940e-01 -5.93654394e-01
-6.67625144e-02 2.80018836e-01 -4.58351493e-01 3.46338630e-01
4.00751650e-01 6.51114047e-01 -4.52152878e-01 -5.43978453e-01
1.70390695e-01 1.11788452e+00 -8.79269600e-01 3.26299369e-01
-1.53252259e-01 7.08473802e-01 -4.89226550e-01 5.13822317e-01
8.75060737e-01 4.11360145e-01 9.43184346e-02 -4.16193128e-01
-1.43093675e-01 2.12811790e-02 -1.33809400e+00 1.56125319e+00
-8.28600645e-01 3.39690089e-01 7.13637769e-01 -8.46043706e-01
1.24102318e+00 5.78788340e-01 5.12771726e-01 -5.09214282e-01
2.31189042e-01 8.53622258e-02 -5.24724483e-01 -3.89843583e-01
4.34306771e-01 -5.13297319e-01 1.55956686e-01 5.63171841e-02
-2.06871182e-02 -3.92299920e-01 -5.84461451e-01 -1.87207058e-01
3.14292908e-01 4.13189590e-01 -1.65410563e-02 -2.13014990e-01
7.58937597e-01 -6.38222814e-01 7.05854177e-01 -4.83769923e-02
1.93651780e-01 8.80740404e-01 6.52846992e-01 -3.42858106e-01
-9.05362129e-01 -9.90771294e-01 -5.17697148e-02 8.40386152e-01
-5.90317771e-02 -3.98543477e-01 -1.05832231e+00 -1.72762215e-01
-1.90663844e-01 5.72342157e-01 -8.40491652e-01 -2.69621104e-01
-5.94351649e-01 -1.76624298e-01 6.18727922e-01 3.97415370e-01
4.40273613e-01 -7.13829637e-01 -2.81966418e-01 -8.77488330e-02
-2.28957072e-01 -1.06724775e+00 -6.44896746e-01 -1.81335643e-01
-8.00922513e-01 -6.02283776e-01 -7.51589239e-01 -6.40840769e-01
7.17494130e-01 2.37242458e-03 6.68888092e-01 -1.00899093e-01
-1.62206888e-01 3.52584630e-01 -2.46024668e-01 -2.84885764e-01
-3.79652768e-01 -3.12243819e-01 2.54672170e-01 7.04444170e-01
-4.01506573e-01 -1.11093712e+00 -5.04325569e-01 3.64260614e-01
-1.12867177e+00 3.94716382e-01 3.71567607e-01 9.00939524e-01
6.46056414e-01 -3.82545114e-01 7.12266922e-01 -7.61934102e-01
4.74819243e-01 -2.27145880e-01 -5.24809599e-01 -3.86033244e-02
-1.70691669e-01 6.74006045e-02 6.90392435e-01 -8.03962409e-01
-1.38127851e+00 2.06066743e-01 -4.54492360e-01 -9.05315161e-01
2.31206000e-01 1.30224109e-01 -6.91104054e-01 -3.87848437e-01
5.01656413e-01 7.86380917e-02 4.05417204e-01 -3.80390286e-01
6.46750808e-01 7.17596292e-01 7.39968419e-01 -1.01311171e+00
9.54328120e-01 6.12480819e-01 5.58464937e-02 -1.06531191e+00
-4.17574346e-01 7.35732317e-02 -5.40602386e-01 -3.22503388e-01
5.97177744e-01 -7.79805064e-01 -8.00124466e-01 4.20852184e-01
-1.12704384e+00 -1.23732954e-01 -4.73107666e-01 3.22340816e-01
-8.80215943e-01 3.44060689e-01 -4.15345877e-01 -8.72353852e-01
-2.22490102e-01 -1.25590003e+00 1.30154276e+00 1.47511110e-01
-2.46064559e-01 -7.14307129e-01 -8.00765026e-03 3.73130888e-01
4.43962991e-01 6.89375401e-01 8.61481071e-01 2.22127408e-01
-4.30000514e-01 9.27065313e-02 1.65719986e-02 3.95435959e-01
2.39882231e-01 2.79507041e-01 -1.27341545e+00 -1.27894774e-01
1.35860011e-01 -4.41287339e-01 3.83763850e-01 3.47785205e-01
1.31663215e+00 -6.32727683e-01 -1.31786883e-01 1.01240838e+00
1.02776062e+00 8.20976868e-02 9.22991216e-01 -1.12602562e-01
5.30384600e-01 1.04059327e+00 3.93125415e-01 6.17519021e-01
1.32182479e-01 1.16341543e+00 5.02142668e-01 -2.21349195e-01
-1.35424912e-01 -5.24602115e-01 4.15720731e-01 7.47142255e-01
-3.32923859e-01 2.02642858e-01 -1.96927071e-01 2.42859766e-01
-1.66202688e+00 -8.31495285e-01 3.75995815e-01 2.12449908e+00
9.21221852e-01 -2.49214470e-01 1.56417236e-01 3.15028399e-01
7.44498134e-01 6.13773055e-02 -4.92674828e-01 -7.56107748e-01
2.70531233e-03 5.01066089e-01 -7.09875822e-02 5.81310987e-01
-7.51217604e-01 7.94848859e-01 5.70697260e+00 8.86428595e-01
-1.42825258e+00 6.13387302e-02 6.96695745e-01 -9.52047408e-02
-6.39009893e-01 -9.78819653e-02 -4.94155407e-01 1.68798760e-01
6.14869356e-01 -1.09447531e-01 4.85339522e-01 7.76220262e-01
3.62590730e-01 4.65446740e-01 -1.01459181e+00 1.22866571e+00
1.65758640e-01 -1.29261374e+00 4.04700279e-01 3.32968049e-02
5.62207639e-01 -9.39551890e-01 3.53268981e-01 1.24012575e-01
-3.53416800e-01 -1.24605930e+00 1.18147469e+00 6.11840487e-01
1.15739882e+00 -9.74345326e-01 1.15810379e-01 7.05748349e-02
-1.15479839e+00 9.09838304e-02 -6.73752502e-02 1.36362970e-01
1.41216189e-01 3.07803780e-01 -4.83707696e-01 4.14217204e-01
5.25317848e-01 3.21815073e-01 -2.44823955e-02 5.30889750e-01
-1.22529492e-01 3.48580241e-01 -3.37489665e-01 1.95579842e-01
4.38920641e-03 -4.33845818e-01 6.78470969e-01 9.57667887e-01
3.75526786e-01 3.19378346e-01 -1.74064308e-01 1.09387350e+00
-1.92770869e-01 2.84104973e-01 -6.43142819e-01 1.63499057e-01
2.68681198e-01 1.20064652e+00 -2.46707976e-01 6.63286448e-02
-1.60054743e-01 8.95584762e-01 6.69059083e-02 3.91988218e-01
-1.06076944e+00 -4.49918270e-01 9.64750528e-01 5.22211730e-01
1.19059533e-01 -2.46996433e-01 -4.18174788e-02 -1.04116845e+00
2.10117698e-02 -8.37024271e-01 -1.49997801e-01 -9.32018280e-01
-9.06773686e-01 7.55026042e-01 1.37917280e-01 -1.16470957e+00
-3.30724716e-01 -4.65215683e-01 -5.04753292e-01 7.61933446e-01
-1.46149671e+00 -1.48389137e+00 -3.22306871e-01 8.57690275e-01
3.63723159e-01 1.47220075e-01 7.98440635e-01 3.59447390e-01
-2.17921257e-01 9.16022003e-01 -2.94117242e-01 -1.22672215e-01
6.26632810e-01 -5.66362739e-01 3.26155946e-02 2.17212945e-01
-3.87329161e-02 5.51973760e-01 6.31459951e-01 -3.64737436e-02
-1.61176574e+00 -9.91567612e-01 2.11698279e-01 6.55249730e-02
2.61119068e-01 -5.88049531e-01 -8.61955702e-01 6.90867186e-01
5.73053211e-02 -6.73678070e-02 6.57167196e-01 -1.84534956e-02
-4.61026400e-01 -4.99964178e-01 -1.41995907e+00 8.40810120e-01
9.81580198e-01 -4.83627647e-01 -2.80748695e-01 -7.16654360e-02
7.84919143e-01 -3.21966380e-01 -7.27070928e-01 4.82498795e-01
7.99450755e-01 -9.08161700e-01 1.02733064e+00 -1.81437150e-01
2.77771533e-01 -3.26244712e-01 -1.34383827e-01 -1.12722480e+00
-1.09332383e-01 -1.12533700e+00 -6.15843013e-02 1.55221224e+00
7.81138241e-02 -6.10008121e-01 7.85548449e-01 4.44064260e-01
-1.20650634e-01 -9.34587717e-01 -1.04712915e+00 -4.57333416e-01
-6.23755483e-03 -3.38271618e-01 7.80282319e-01 1.06235373e+00
-1.99882090e-01 1.45671844e-01 -6.84456587e-01 1.83024168e-01
3.86259824e-01 -2.01311111e-02 8.63843381e-01 -1.09787774e+00
-3.23064178e-01 -2.98051000e-01 -4.52379227e-01 -1.15337288e+00
4.41231787e-01 -5.57904899e-01 -6.91001937e-02 -8.87879014e-01
-4.86127399e-02 -6.59892619e-01 1.28249362e-01 4.69257087e-01
3.73213619e-01 3.92836630e-01 1.86055720e-01 1.61598623e-02
3.52104269e-02 1.05913270e+00 1.34883547e+00 -6.55253837e-03
-4.57346708e-01 -1.15493745e-01 -7.71259427e-01 8.49058092e-01
4.44124818e-01 -1.62583560e-01 -7.09541976e-01 -3.81821066e-01
-1.74251534e-02 4.18323398e-01 4.98512208e-01 -9.75821793e-01
-3.16928178e-02 -2.37904757e-01 2.53062993e-01 -1.05102710e-01
1.07884276e+00 -7.85481155e-01 5.21136820e-01 9.39095914e-02
-3.39439213e-01 -2.12067127e-01 2.84084201e-01 3.81628931e-01
-3.82249355e-01 -6.57458976e-02 1.11234474e+00 1.59996957e-01
-1.46905631e-01 3.21824372e-01 -3.17777961e-01 -2.46419489e-01
9.99943018e-01 -4.69182342e-01 2.70632386e-01 -9.19550896e-01
-7.35689878e-01 -1.03879482e-01 4.22080189e-01 5.13061821e-01
7.72646487e-01 -1.55791056e+00 -6.69079304e-01 5.74554026e-01
-1.37661904e-01 -4.56559695e-02 1.73573434e-01 6.04563594e-01
-5.04152834e-01 1.01698227e-02 -3.90426844e-01 -6.81470573e-01
-1.29506743e+00 2.80159742e-01 4.76061612e-01 1.54831886e-01
-4.68967319e-01 7.49993742e-01 6.46286428e-01 -4.82501656e-01
2.15768844e-01 -1.32641762e-01 -1.84883446e-01 -8.65957364e-02
4.67059612e-01 1.21288881e-01 -1.15501963e-01 -9.91608202e-01
-1.04280278e-01 1.06597209e+00 1.66329086e-01 -3.32773089e-01
1.24655902e+00 1.82929263e-02 -5.09450771e-02 2.16537744e-01
1.25189948e+00 4.15224135e-01 -1.44017136e+00 -4.39961813e-02
-5.88676751e-01 -5.36107302e-01 1.30319642e-02 -3.50125909e-01
-1.29076099e+00 8.79206896e-01 4.86710548e-01 -2.65103549e-01
1.47249484e+00 -2.17544839e-01 7.15838075e-01 2.79366113e-02
3.18305016e-01 -8.31862330e-01 3.79190296e-01 5.62760681e-02
1.24543238e+00 -5.30683935e-01 -1.33630127e-01 -8.75972986e-01
-5.79800904e-01 1.22200251e+00 4.78760928e-01 -1.28420010e-01
6.48627937e-01 4.23560023e-01 1.00703947e-01 8.42943043e-02
-6.18897974e-01 4.31994289e-01 3.09554100e-01 6.20593905e-01
3.71116757e-01 4.11819667e-03 -1.80752993e-01 8.13190043e-01
-3.02967161e-01 9.63614583e-02 3.32590014e-01 7.00806856e-01
-1.07634112e-01 -1.20085740e+00 -6.09542847e-01 -2.11096585e-01
-3.19189310e-01 2.46740282e-01 -4.25800741e-01 7.69133270e-01
3.39878678e-01 7.18010247e-01 -5.12653627e-02 -3.90854836e-01
4.52868402e-01 -8.86107702e-03 6.79099321e-01 -3.49341810e-01
-3.01301837e-01 2.40649164e-01 -5.36570065e-02 -4.93427694e-01
-4.56333995e-01 -4.07606870e-01 -1.18532121e+00 -3.53818893e-01
-3.98000360e-01 2.46526778e-01 7.56600976e-01 6.01999402e-01
5.68112612e-01 3.36238742e-01 9.90143478e-01 -1.29476249e+00
-3.12391430e-01 -5.90419531e-01 -5.65827072e-01 2.86276072e-01
3.61658305e-01 -9.44333911e-01 -2.14967832e-01 1.55635774e-01] | [12.944238662719727, -0.35720232129096985] |
ec548eae-519f-4387-a795-1a57efbfd141 | an-efficient-distributed-learning-algorithm | 1310.8418 | null | http://arxiv.org/abs/1310.8418v4 | http://arxiv.org/pdf/1310.8418v4.pdf | An efficient distributed learning algorithm based on effective local functional approximations | Scalable machine learning over big data is an important problem that is
receiving a lot of attention in recent years. On popular distributed
environments such as Hadoop running on a cluster of commodity machines,
communication costs are substantial and algorithms need to be designed suitably
considering those costs. In this paper we give a novel approach to the
distributed training of linear classifiers (involving smooth losses and L2
regularization) that is designed to reduce the total communication costs. At
each iteration, the nodes minimize locally formed approximate objective
functions; then the resulting minimizers are combined to form a descent
direction to move. Our approach gives a lot of freedom in the formation of the
approximate objective function as well as in the choice of methods to solve
them. The method is shown to have $O(log(1/\epsilon))$ time convergence. The
method can be viewed as an iterative parameter mixing method. A special
instantiation yields a parallel stochastic gradient descent method with strong
convergence. When communication times between nodes are large, our method is
much faster than the Terascale method (Agarwal et al., 2011), which is a state
of the art distributed solver based on the statistical query model (Chuet al.,
2006) that computes function and gradient values in a distributed fashion. We
also evaluate against other recent distributed methods and demonstrate superior
performance of our method. | ['Leon Bottou', 'Dhruv Mahajan', 'S. Sundararajan', 'S. Sathiya Keerthi', 'Nikunj Agrawal'] | 2013-10-31 | null | null | null | null | ['l2-regularization'] | ['methodology'] | [-6.93200648e-01 -2.46406436e-01 5.13404906e-02 -6.50729954e-01
-1.17845333e+00 -3.05947304e-01 1.88855946e-01 2.87431806e-01
-7.79263198e-01 8.37348104e-01 -3.23149562e-01 -1.56064451e-01
-3.05296898e-01 -1.02131641e+00 -9.07969296e-01 -1.05713725e+00
-9.32356045e-02 9.25867021e-01 1.15174957e-01 -5.99090047e-02
6.28489777e-02 5.40192068e-01 -1.34947014e+00 1.29941165e-01
8.25356901e-01 1.23452759e+00 2.30627730e-01 8.57277632e-01
-3.80710781e-01 1.05713439e+00 -6.11231625e-01 -4.08802539e-01
4.42675829e-01 -2.72382349e-01 -1.12806904e+00 -4.34156984e-01
3.26849163e-01 -2.17304364e-01 1.30760074e-01 1.02423716e+00
7.56270289e-01 3.03853989e-01 1.38337538e-01 -1.28895056e+00
1.77076384e-02 8.69396806e-01 -7.42564976e-01 5.06915003e-02
-1.77779078e-01 -2.86242545e-01 8.95501375e-01 -9.30930674e-01
4.18351322e-01 1.11911678e+00 9.00387406e-01 2.55061924e-01
-1.26346338e+00 -5.20578861e-01 -5.97991087e-02 2.75187761e-01
-1.79861569e+00 -2.38769621e-01 4.72035646e-01 -7.95197785e-02
9.76341546e-01 5.02864718e-01 3.60869288e-01 8.41597244e-02
1.46522045e-01 7.76188195e-01 9.12132800e-01 -5.89400828e-01
7.61200905e-01 4.60062027e-01 -2.53371000e-02 7.50385821e-01
-2.53875856e-03 -2.09395632e-01 -7.08392203e-01 -8.97832513e-01
9.72775817e-02 2.10198209e-01 2.02502504e-01 -2.86743164e-01
-7.28211462e-01 1.20403087e+00 7.63544440e-01 2.00192988e-01
-4.87091333e-01 5.46336949e-01 6.29696131e-01 5.66947222e-01
1.02921510e+00 7.48174312e-03 -8.05179358e-01 -1.66660786e-01
-1.12993193e+00 5.08143604e-01 1.41685283e+00 7.06569791e-01
1.16146767e+00 -3.72758687e-01 6.55585602e-02 9.91601229e-01
1.25981152e-01 4.43270922e-01 3.30356866e-01 -1.16675162e+00
4.54718381e-01 2.03445956e-01 6.52800724e-02 -1.02952349e+00
-4.09920216e-01 -1.13390483e-01 -1.06155157e+00 2.22304061e-01
4.12571073e-01 -4.42727596e-01 -1.21832728e-01 1.56010079e+00
9.10676360e-01 1.57982767e-01 -2.30174497e-01 8.47357154e-01
1.01000533e-01 8.30881476e-01 -1.76724583e-01 -1.77316442e-01
8.78054082e-01 -1.22222543e+00 -3.03078353e-01 1.81940004e-01
1.16413414e+00 -7.18206286e-01 7.46835411e-01 6.01656854e-01
-1.31321788e+00 -1.77607119e-01 -6.14720285e-01 -5.97453229e-02
-2.43334308e-01 -2.20844775e-01 1.05646229e+00 6.85106635e-01
-1.52105868e+00 9.56764817e-01 -1.28345239e+00 -1.26096264e-01
3.19965452e-01 5.40269732e-01 9.61716622e-02 -1.52551606e-01
-7.16398001e-01 7.56781340e-01 3.28358710e-02 3.27267945e-01
-5.87868214e-01 -7.65095055e-01 -2.41304353e-01 -1.12616159e-02
3.67804050e-01 -8.05972934e-01 1.30868721e+00 -9.56139266e-01
-1.65794015e+00 5.77981472e-01 -2.02556580e-01 -4.28829223e-01
6.17819011e-01 -9.83779654e-02 3.76254976e-01 -1.70797408e-01
-1.37221918e-01 2.46026129e-01 4.68466401e-01 -6.45138919e-01
-8.19778502e-01 -6.33353055e-01 -1.49724990e-01 2.27642268e-01
-6.73513293e-01 2.78080106e-01 -2.84270436e-01 -2.43996054e-01
-8.39300230e-02 -1.02971673e+00 -6.35832369e-01 1.48968384e-01
-7.15641826e-02 -4.86116767e-01 4.58097488e-01 -3.50369930e-01
9.96866047e-01 -1.90150547e+00 2.52424210e-01 5.64042568e-01
4.41261441e-01 -1.99402377e-01 1.21884473e-01 6.06989682e-01
6.22370243e-01 4.09538895e-02 -2.97331870e-01 -5.35905361e-01
1.46598235e-01 1.73629194e-01 -2.69102640e-02 8.14977109e-01
-5.22143722e-01 5.39087296e-01 -6.11594379e-01 -5.96619308e-01
-1.96373686e-01 4.43200886e-01 -7.72034585e-01 1.42572433e-01
-2.89658874e-01 -4.21569608e-02 -6.54955089e-01 3.23601037e-01
8.79731596e-01 -4.55665320e-01 2.27808252e-01 1.89847991e-01
-9.01322588e-02 3.95433418e-02 -1.58593547e+00 1.58642757e+00
-9.21601415e-01 3.69127601e-01 9.23726797e-01 -1.39203811e+00
7.63294995e-01 1.95671722e-01 8.81677866e-01 -1.41197711e-01
-5.56862615e-02 6.37355626e-01 -5.00357211e-01 -2.43502289e-01
3.75955522e-01 -7.79898539e-02 9.14578966e-04 8.21942031e-01
-3.80950212e-01 -3.37227434e-01 1.15501381e-01 3.53697598e-01
1.25778127e+00 -3.17446828e-01 -1.93879232e-01 -5.24076223e-01
3.48463058e-01 8.31215158e-02 4.99147207e-01 8.56734574e-01
1.31252378e-01 2.77131736e-01 4.76611078e-01 -8.36022973e-01
-9.60512638e-01 -4.88414019e-01 -1.63816869e-01 1.62077999e+00
-1.26840144e-01 -5.20570219e-01 -8.44812274e-01 -3.32779765e-01
3.13850999e-01 1.64250568e-01 -3.63442063e-01 2.39421442e-01
-6.34127438e-01 -9.63543773e-01 3.48421514e-01 3.18163723e-01
4.87010479e-01 -9.38369513e-01 -5.94821811e-01 2.15651929e-01
-2.13113427e-02 -5.46029985e-01 -4.73920524e-01 4.40658331e-01
-1.07168186e+00 -7.30010927e-01 -5.97250938e-01 -7.51892447e-01
6.96232378e-01 1.78932503e-01 1.24228048e+00 3.28839570e-01
-5.10780275e-01 2.18785003e-01 -4.97176163e-02 -3.89074922e-01
-1.11607648e-01 2.71998852e-01 -5.82082905e-02 5.30080460e-02
4.66578752e-01 -5.47414124e-01 -5.33995926e-01 2.29440019e-01
-8.68208408e-01 1.64492186e-02 2.43792146e-01 9.12759125e-01
7.99627781e-01 1.81783214e-01 1.03478380e-01 -1.26098108e+00
7.66607165e-01 -7.91127086e-01 -1.00610721e+00 3.15682203e-01
-8.63122046e-01 2.72838116e-01 1.02302468e+00 -2.27022737e-01
-8.23430538e-01 1.96758181e-01 7.45094344e-02 -3.51621330e-01
3.75762105e-01 6.99825585e-01 2.98990995e-01 -4.89429891e-01
7.28777647e-01 -8.12964514e-02 1.43589318e-01 -6.78221643e-01
4.32920933e-01 9.11382914e-01 -2.04470918e-01 -1.02014995e+00
2.91060209e-01 5.36388576e-01 6.38601035e-02 -6.57795131e-01
-5.99682152e-01 -5.63650310e-01 -2.48962373e-01 1.73662677e-01
2.08969936e-01 -7.49913156e-01 -1.02995920e+00 5.02931118e-01
-1.19496799e+00 -6.78094387e-01 -3.72733146e-01 4.05595213e-01
-4.84648883e-01 -1.14272580e-01 -9.58091140e-01 -7.42839754e-01
-6.77341938e-01 -9.28531706e-01 9.16974783e-01 -2.99096238e-02
1.63510203e-01 -1.03009391e+00 4.47789043e-01 1.28146604e-01
1.04198718e+00 1.73460945e-01 4.67555851e-01 -5.66149712e-01
-5.61698079e-01 -3.93320471e-01 -1.86079890e-01 2.57822454e-01
-4.98386562e-01 -5.59689216e-02 -6.69577956e-01 -5.71127713e-01
3.36889118e-01 -6.41965747e-01 6.71235681e-01 5.32206655e-01
1.55266595e+00 -6.25710309e-01 -4.31442082e-01 9.59667921e-01
1.62667727e+00 -6.03719473e-01 3.73478085e-02 1.20504990e-01
7.10090995e-01 5.30411243e-01 3.18168998e-01 9.38342094e-01
3.83295834e-01 5.97726285e-01 1.09888345e-01 -3.89582738e-02
3.76937389e-01 2.42340982e-01 2.03518048e-01 1.11780441e+00
1.58500497e-03 1.14491776e-01 -9.92188156e-01 3.05089027e-01
-2.20844293e+00 -6.97758794e-01 -2.36807734e-01 2.32414746e+00
1.21220529e+00 -3.91273081e-01 2.40377277e-01 -2.57825553e-01
6.75357044e-01 -1.37053490e-01 -6.65843427e-01 -7.76998699e-01
1.45272315e-01 4.01679784e-01 7.77184069e-01 6.68567240e-01
-6.91530645e-01 7.64095008e-01 5.83938265e+00 1.05336475e+00
-1.10810626e+00 5.80375671e-01 8.90229940e-01 -6.40175521e-01
-1.78082094e-01 -1.49350148e-02 -7.30474830e-01 6.00557745e-01
1.37812483e+00 -3.25210154e-01 1.07576966e+00 1.26230693e+00
-3.25504169e-02 -2.32858464e-01 -9.58535254e-01 1.09768200e+00
-1.51304752e-01 -1.36691105e+00 -7.10483611e-01 1.35843843e-01
8.66309166e-01 6.81886673e-01 -3.61281604e-01 2.99057096e-01
8.11474264e-01 -1.04074144e+00 5.16330838e-01 2.80481160e-01
4.35546339e-01 -1.00754905e+00 7.91629910e-01 7.09596932e-01
-1.16552305e+00 -1.10694841e-01 -9.12932396e-01 -2.65392154e-01
4.00272161e-02 1.20611215e+00 -2.60018170e-01 1.75916106e-01
1.20738006e+00 1.46781027e-01 -2.79792964e-01 1.00081110e+00
1.72329932e-01 6.31964922e-01 -1.16076601e+00 -5.18953443e-01
1.51719108e-01 -4.89907801e-01 6.18833639e-02 1.06101394e+00
3.06074142e-01 -2.15417705e-02 4.36423481e-01 6.45428479e-01
-3.09714228e-01 5.96961081e-01 -2.97595829e-01 4.62353706e-01
6.30820155e-01 1.42298710e+00 -6.53338611e-01 -3.40263963e-01
-4.30981994e-01 8.65749359e-01 9.53384221e-01 2.37507135e-01
-6.87418580e-01 -5.32684982e-01 4.78388250e-01 -9.98647586e-02
3.45602781e-01 -1.18154198e-01 -4.53313321e-01 -1.01133156e+00
4.16387916e-01 -7.58700907e-01 5.81115186e-01 -1.15499660e-01
-1.53690422e+00 7.39668727e-01 -1.30210653e-01 -5.05506337e-01
-2.61059701e-01 -2.48396173e-01 -5.31919122e-01 9.92816985e-01
-1.25559628e+00 -8.41800928e-01 -3.23733181e-01 8.78099263e-01
1.02299072e-01 1.32725030e-01 9.73554969e-01 4.82680827e-01
-5.18734694e-01 5.42529464e-01 9.36146021e-01 -3.03087682e-01
5.84694386e-01 -1.24397731e+00 -2.70196982e-02 4.00164515e-01
9.23802480e-02 3.61595929e-01 6.50515676e-01 -2.95238137e-01
-1.72752178e+00 -9.26915467e-01 1.13117397e+00 -1.61616638e-01
6.33731186e-01 -5.57213008e-01 -7.44481623e-01 4.93525654e-01
1.81967635e-02 7.35284269e-01 6.42643452e-01 4.68059748e-01
-8.04522112e-02 -8.36284876e-01 -1.36643445e+00 4.72628325e-03
5.04556835e-01 -3.33596647e-01 3.93294781e-01 1.02638924e+00
3.52612555e-01 -4.86439168e-01 -6.90922141e-01 -1.44193396e-01
2.44364038e-01 -1.01459682e+00 6.65159881e-01 -4.67951804e-01
7.72369057e-02 7.60689080e-02 -1.00181676e-01 -1.25826240e+00
6.16489584e-03 -1.06945491e+00 -2.20121384e-01 9.30908859e-01
2.87387848e-01 -9.24644291e-01 1.02692461e+00 1.18644917e+00
3.29746664e-01 -1.05478430e+00 -1.13769412e+00 -7.76728213e-01
4.76525396e-01 -3.96630257e-01 6.20345950e-01 6.74374461e-01
1.56839147e-01 -4.04567346e-02 -2.29760900e-01 -1.53548911e-01
8.30602467e-01 2.72653908e-01 9.03012335e-01 -1.04103661e+00
-6.14718378e-01 -4.32182014e-01 -1.12542786e-01 -9.76891339e-01
1.67136863e-01 -1.02488744e+00 1.14035234e-01 -1.07907987e+00
2.53002465e-01 -1.12154579e+00 -3.01045150e-01 5.05379677e-01
9.08050090e-02 2.19591912e-02 1.46799490e-01 4.07804608e-01
-7.61429667e-01 4.33398783e-01 5.48393488e-01 4.76284511e-02
-3.89255315e-01 2.49515995e-01 -4.48468238e-01 4.51613843e-01
7.25450754e-01 -8.83625686e-01 -3.08559462e-02 -8.03329349e-01
6.73903883e-01 2.60003358e-01 1.01523764e-01 -7.57345736e-01
8.90921533e-01 -2.93565720e-01 -9.92972627e-02 -1.68435976e-01
1.13322087e-01 -8.09369385e-01 1.65466234e-01 4.99056160e-01
-5.62313080e-01 3.24606866e-01 -1.69173971e-01 3.86784971e-01
-3.43402386e-01 -3.38471234e-01 7.18585312e-01 -2.06699982e-01
-2.61103157e-02 5.25649965e-01 -1.23326294e-01 2.10249379e-01
1.17436135e+00 4.24978554e-01 4.97995764e-02 -4.29209143e-01
-3.32338572e-01 5.44188440e-01 2.64400452e-01 -3.54690790e-01
2.31813028e-01 -1.21036327e+00 -9.59598541e-01 -1.33469671e-01
-6.72393501e-01 4.59704638e-01 7.52661303e-02 1.22104073e+00
-9.46870923e-01 1.33990839e-01 3.85336280e-01 -4.85437930e-01
-1.02514851e+00 4.62120801e-01 3.17058146e-01 -4.66437489e-01
-5.27253687e-01 1.24225640e+00 -3.74115437e-01 -6.27129912e-01
4.86462593e-01 -2.45512426e-02 6.65814519e-01 -1.67617634e-01
6.12728119e-01 9.35720742e-01 3.99588853e-01 9.92898867e-02
-5.00126481e-01 2.79069215e-01 -6.63746521e-02 -1.62637755e-01
1.84861124e+00 -1.26065373e-01 -8.96839499e-01 3.38569790e-01
1.69680178e+00 6.88798800e-02 -9.76630747e-01 -4.60332215e-01
3.84266824e-02 -5.34040093e-01 3.24097693e-01 -4.70186621e-01
-1.41305363e+00 7.59952307e-01 3.72881889e-01 5.07804453e-01
9.59932983e-01 9.43748653e-02 1.05358970e+00 7.16341496e-01
6.72135949e-01 -1.51857066e+00 -3.63325506e-01 4.55669403e-01
5.08153796e-01 -1.31892037e+00 2.48953223e-01 -8.00423324e-02
-3.12029272e-01 1.22281504e+00 4.08647448e-01 -4.43652779e-01
1.06518555e+00 7.27315187e-01 -2.51488294e-02 -1.51367888e-01
-9.46408570e-01 4.66481328e-01 -3.03945363e-01 2.14749556e-02
3.72655511e-01 1.55290693e-01 -5.37355959e-01 3.91990215e-01
-2.50929654e-01 1.35671958e-01 7.97115117e-02 1.10086954e+00
-3.47276419e-01 -1.20757341e+00 -3.40381086e-01 6.01234913e-01
-5.49127996e-01 -9.63237956e-02 -8.80930107e-03 1.92389518e-01
-1.27595872e-01 8.49011540e-01 9.67426747e-02 -4.46630158e-02
-1.78319201e-01 6.59755478e-03 2.36160904e-01 -4.99109656e-01
-9.37685788e-01 -1.52454779e-01 -2.05447808e-01 -8.79197896e-01
-1.89714044e-01 -3.49590510e-01 -1.30388319e+00 -9.42319453e-01
-5.06082118e-01 6.78916216e-01 1.23767292e+00 7.68545866e-01
6.15316391e-01 -3.46609727e-02 1.00984192e+00 -8.91677797e-01
-1.36848831e+00 -7.09348917e-01 -8.16741109e-01 2.18641251e-01
-6.01992011e-02 2.59304475e-02 -6.61776245e-01 -2.28239328e-01] | [6.293936252593994, 4.988681316375732] |
8eb3103b-4579-4821-83a5-62107c64cf4d | skeletal-video-anomaly-detection-using-deep | 2301.00114 | null | https://arxiv.org/abs/2301.00114v2 | https://arxiv.org/pdf/2301.00114v2.pdf | Skeletal Video Anomaly Detection using Deep Learning: Survey, Challenges and Future Directions | The existing methods for video anomaly detection mostly utilize videos containing identifiable facial and appearance-based features. The use of videos with identifiable faces raises privacy concerns, especially when used in a hospital or community-based setting. Appearance-based features can also be sensitive to pixel-based noise, straining the anomaly detection methods to model the changes in the background and making it difficult to focus on the actions of humans in the foreground. Structural information in the form of skeletons describing the human motion in the videos is privacy-protecting and can overcome some of the problems posed by appearance-based features. In this paper, we present a survey of privacy-protecting deep learning anomaly detection methods using skeletons extracted from videos. We present a novel taxonomy of algorithms based on the various learning approaches. We conclude that skeleton-based approaches for anomaly detection can be a plausible privacy-protecting alternative for video anomaly detection. Lastly, we identify major open research questions and provide guidelines to address them. | ['Shehroz S. Khan', 'Alex Mihailidis', 'Pratik K. Mishra'] | 2022-12-31 | null | null | null | null | ['video-anomaly-detection'] | ['computer-vision'] | [ 3.25016469e-01 -7.94360321e-03 -1.08933970e-01 -4.50175583e-01
-1.70674354e-01 -4.23645735e-01 2.46293366e-01 3.67185891e-01
-5.23390532e-01 4.01484400e-01 3.44151072e-02 -9.21713039e-02
1.61837161e-01 -5.11091888e-01 -7.00682342e-01 -9.13805842e-01
-4.69802767e-01 -3.28481823e-01 1.16035223e-01 2.28551731e-01
2.45011747e-01 9.65506971e-01 -1.80863059e+00 5.86508214e-01
3.85624111e-01 1.19770253e+00 -8.99789035e-01 4.61777627e-01
-2.34073661e-02 5.82527876e-01 -5.70158839e-01 -7.34028757e-01
7.45246053e-01 -2.99550205e-01 -5.49606025e-01 2.34106854e-01
1.06537366e+00 -1.08982038e+00 -7.07615852e-01 1.21001732e+00
1.57365710e-01 7.21314549e-02 5.02215505e-01 -1.81662786e+00
-1.55272499e-01 -1.91313133e-01 -9.72324967e-01 6.15948737e-01
4.64732200e-01 1.18825048e-01 5.57577074e-01 -5.95344961e-01
4.41867858e-01 1.25364947e+00 5.93102634e-01 8.84564996e-01
-1.01337850e+00 -8.16990256e-01 4.72671896e-01 5.25009692e-01
-1.30138195e+00 -7.96661675e-01 8.48515153e-01 -4.54485506e-01
5.18261969e-01 6.12722218e-01 7.72258401e-01 1.44396973e+00
1.74290970e-01 1.03073168e+00 6.29564643e-01 -2.05083326e-01
5.15223481e-02 -3.79097641e-01 -9.69287753e-02 1.10291696e+00
7.24973261e-01 9.33928192e-02 -7.06399143e-01 -1.01816440e+00
8.07310581e-01 3.98993075e-01 -1.40625700e-01 -7.46081054e-01
-7.33075082e-01 8.01917851e-01 -9.46197733e-02 -1.66593306e-02
-2.14852303e-01 3.03397119e-01 8.74937713e-01 4.06123936e-01
3.93359601e-01 1.17198475e-01 -2.86188811e-01 1.37376860e-01
-8.61233592e-01 2.51905620e-01 5.10755539e-01 6.84609771e-01
6.39038265e-01 2.96327442e-01 -1.60620719e-01 2.76435345e-01
5.34257770e-01 -1.20944204e-02 3.74415934e-01 -1.17799652e+00
-9.21533853e-02 4.64873493e-01 -7.02956766e-02 -1.36829865e+00
-2.16518342e-01 4.71976191e-01 -7.06450462e-01 5.21064222e-01
8.77260745e-01 -1.63516611e-01 -8.81449938e-01 1.58193827e+00
5.22505879e-01 4.18326348e-01 -3.71563673e-01 6.02100253e-01
5.17461717e-01 7.05615282e-02 2.13686079e-01 -4.02521461e-01
1.38921630e+00 -6.74822927e-01 -9.33508277e-01 2.68962104e-02
7.38573968e-01 -4.71742958e-01 3.92904222e-01 4.12715018e-01
-8.03525329e-01 1.01580925e-01 -8.40297282e-01 9.08119529e-02
-3.56528997e-01 -1.63381219e-01 6.09046519e-01 1.06507945e+00
-1.01336980e+00 5.08304775e-01 -1.17767394e+00 -5.96379638e-01
9.24361527e-01 6.15321159e-01 -7.19535649e-01 8.29721019e-02
-6.88731194e-01 5.08726716e-01 -1.47934109e-02 -3.75694036e-02
-7.33242393e-01 -4.37189668e-01 -1.11451113e+00 -1.72313988e-01
5.18944442e-01 -4.08115745e-01 9.74628627e-01 -1.41425157e+00
-1.00480115e+00 1.25567293e+00 -3.02888691e-01 -5.10191381e-01
8.11389029e-01 -3.00806195e-01 -4.68490690e-01 5.20994186e-01
-2.60600209e-01 4.54037398e-01 1.48938429e+00 -7.67584145e-01
-7.15871155e-01 -6.99973226e-01 -6.97973818e-02 -1.38635054e-01
-4.02150929e-01 3.86596769e-01 -1.82891414e-01 -9.75831032e-01
6.60075024e-02 -8.12823713e-01 -2.77766317e-01 1.04924548e+00
-1.51196480e-01 6.31892607e-02 1.41282141e+00 -7.72530496e-01
1.13014889e+00 -2.46863794e+00 -3.48127693e-01 3.24785203e-01
3.91152173e-01 3.62323314e-01 -1.18817940e-01 1.35286406e-01
-1.65194988e-01 3.18010271e-01 -1.51082143e-01 -2.30612248e-01
-4.35262442e-01 2.31641367e-01 9.11609177e-03 1.10486507e+00
1.69702694e-01 5.32407343e-01 -6.91607058e-01 -7.28778362e-01
2.05839828e-01 1.99582011e-01 -6.20403945e-01 2.89128032e-02
7.82962814e-02 5.84404945e-01 -6.09433949e-01 1.11324632e+00
8.23743522e-01 4.71354336e-01 4.04764786e-02 5.34560271e-02
3.10949087e-01 -3.19951773e-01 -9.69015718e-01 1.52938175e+00
2.90253431e-01 8.04994226e-01 5.70269883e-01 -9.66136038e-01
4.63458210e-01 4.82297301e-01 9.40764725e-01 -1.18078537e-01
1.20419882e-01 2.63211466e-02 7.22517669e-02 -7.97451258e-01
7.05767050e-02 2.34358415e-01 3.09145987e-01 5.08574009e-01
-1.28798455e-01 6.46227062e-01 -9.11694914e-02 5.47174253e-02
1.36931312e+00 -2.62282491e-02 4.42037284e-01 3.42835821e-02
7.35047638e-01 -5.22095621e-01 7.34447777e-01 8.23914409e-01
-9.69910324e-01 5.35921097e-01 4.53775495e-01 -1.08079219e+00
-8.01196933e-01 -9.62836623e-01 -1.50493264e-01 9.36686993e-01
-2.41639629e-01 -5.35573840e-01 -8.42595518e-01 -1.37091100e+00
2.19105467e-01 1.38003856e-01 -7.11174846e-01 -3.42755139e-01
-5.88438511e-01 -5.94601810e-01 8.93886089e-01 4.54737514e-01
3.18830937e-01 -7.57738650e-01 -7.24471509e-01 -7.22538009e-02
-2.15324208e-01 -1.07325900e+00 -3.81918639e-01 -3.51228058e-01
-1.03896630e+00 -1.37412441e+00 -3.51200283e-01 -3.46230060e-01
1.09763384e+00 2.49938533e-01 5.39216995e-01 6.35115385e-01
-7.01170146e-01 1.11145747e+00 -3.51505607e-01 -7.41040826e-01
-3.41958463e-01 -5.83143115e-01 4.03139681e-01 6.86780810e-01
8.98611903e-01 -4.84832555e-01 -5.69071174e-01 2.01859444e-01
-1.05325449e+00 -7.83625305e-01 1.70730084e-01 5.69252074e-01
3.45692009e-01 -6.88574389e-02 2.59808172e-02 -6.63808167e-01
1.42496347e-01 -4.73534226e-01 -5.15422046e-01 7.68089890e-02
-2.64182657e-01 -2.30733007e-01 2.81970948e-01 -5.56600809e-01
-7.08015501e-01 5.70205271e-01 4.50901128e-02 -7.99901247e-01
-6.14073992e-01 -2.77231932e-01 -2.48655185e-01 -6.15643919e-01
4.99140412e-01 1.45551249e-01 4.69889045e-01 -4.58837926e-01
-1.67254478e-01 3.96601439e-01 2.60975599e-01 -4.56456155e-01
7.57331729e-01 1.07695699e+00 2.01909304e-01 -1.45493686e+00
-4.63782489e-01 -5.96139014e-01 -7.74759173e-01 -3.63353938e-01
8.49688649e-01 -7.28381038e-01 -5.54003060e-01 7.17671871e-01
-1.17144263e+00 3.99823785e-01 -3.29123467e-01 3.01812947e-01
-5.98496675e-01 1.12977910e+00 -5.06614149e-01 -1.01765919e+00
-1.09049834e-01 -1.03344023e+00 9.57649291e-01 -5.98253217e-03
-4.57007945e-01 -7.45859146e-01 -1.59245849e-01 2.95040697e-01
1.20215587e-01 7.30743170e-01 7.37446487e-01 -9.81072783e-01
-5.97405434e-01 -5.86389184e-01 1.39018953e-01 3.32334548e-01
3.28311563e-01 2.75330544e-01 -1.30212188e+00 -4.66641814e-01
2.38698170e-01 4.16939072e-02 7.81452835e-01 5.15518606e-01
1.83053792e+00 -7.52873182e-01 -4.64209855e-01 7.97650278e-01
8.50677311e-01 2.35305622e-01 6.38911307e-01 4.18971419e-01
7.42467165e-01 7.60377944e-01 4.77716208e-01 5.83111644e-01
-2.22724795e-01 3.71043205e-01 7.67666876e-01 8.55762511e-04
4.18945223e-01 8.66401121e-02 5.92552543e-01 -1.76908210e-01
-2.36248076e-01 -8.95328298e-02 -5.31978250e-01 3.94156426e-01
-1.93514156e+00 -1.29460669e+00 -1.51745602e-01 2.26935816e+00
3.30511868e-01 -5.24038136e-01 5.20698071e-01 1.45456702e-01
8.50455701e-01 2.67742246e-01 -5.00888526e-01 -5.13747692e-01
3.86688672e-02 -1.55624509e-01 5.01260161e-01 -1.90022197e-02
-1.67677081e+00 5.94519138e-01 7.29856873e+00 3.42631400e-01
-8.80583525e-01 -6.77247867e-02 6.26476824e-01 -4.35880989e-01
1.99615836e-01 -2.12925792e-01 -4.16883022e-01 4.99944121e-01
7.66921937e-01 1.26450816e-02 6.33890182e-02 9.55968380e-01
1.78899407e-01 -1.01580858e-01 -1.43538916e+00 1.18782902e+00
2.58510023e-01 -9.29989338e-01 2.69342661e-01 4.21276927e-01
1.74074709e-01 -2.50866473e-01 2.12417111e-01 -4.91017967e-01
-2.01190323e-01 -9.87415731e-01 4.03927922e-01 2.20396698e-01
5.37182748e-01 -6.48287237e-01 5.82237780e-01 -8.68013650e-02
-9.43796694e-01 -2.94037282e-01 -4.07300621e-01 2.75188144e-02
-1.12215266e-01 1.57228515e-01 -5.02430618e-01 2.43417442e-01
1.03640449e+00 6.66541338e-01 -5.39188623e-01 1.24357247e+00
1.83573067e-01 5.03107488e-01 -5.70689619e-01 4.51420665e-01
7.69548491e-02 -2.32686833e-01 1.00448143e+00 1.19742429e+00
3.31854731e-01 1.96122639e-02 1.59613937e-01 3.75071049e-01
2.72344351e-02 2.06407383e-01 -1.21666133e+00 -1.59428477e-01
5.98727986e-02 1.03653049e+00 -6.23921275e-01 1.31859407e-01
-1.01288092e+00 1.26832700e+00 -1.03372820e-01 3.46231133e-01
-6.08308911e-01 5.42279668e-02 1.34757161e+00 5.18265665e-01
1.22502685e-01 -1.20429702e-01 2.86590129e-01 -1.32574821e+00
2.34883860e-01 -1.22874963e+00 1.01937580e+00 -7.28035644e-02
-1.22443414e+00 1.39395982e-01 9.91727412e-02 -1.33423948e+00
-6.30605547e-03 -7.15061665e-01 -6.59542322e-01 9.47315916e-02
-1.09186709e+00 -1.08742142e+00 -1.24437436e-01 8.75368714e-01
4.10052180e-01 -4.45024580e-01 8.76762450e-01 2.68189251e-01
-7.00958550e-01 9.09906805e-01 8.94065481e-03 6.34743094e-01
8.01915586e-01 -8.45199823e-01 1.70251951e-01 1.18662500e+00
3.06259185e-01 6.08371854e-01 4.91778225e-01 -6.17848396e-01
-1.28363502e+00 -1.00526869e+00 2.77585804e-01 -6.45979285e-01
4.27400321e-01 -3.28575850e-01 -1.13415158e+00 1.04823875e+00
-8.10504556e-02 5.03464282e-01 1.11790276e+00 -1.61314189e-01
-4.71726477e-01 -7.64025748e-02 -1.77980471e+00 5.90090632e-01
9.89550650e-01 -4.60109293e-01 -2.48323023e-01 1.75027743e-01
1.62999377e-01 -2.83586711e-01 -3.90313953e-01 2.38506690e-01
8.01361322e-01 -1.05168962e+00 9.97938573e-01 -1.07620430e+00
-1.74755618e-01 -2.63963044e-01 3.73770520e-02 -6.03400230e-01
-2.83486187e-01 -8.05954456e-01 -6.14118934e-01 9.04599607e-01
-4.44684289e-02 -6.58265889e-01 1.16733575e+00 1.10700858e+00
4.19472098e-01 -3.06406975e-01 -1.26601315e+00 -7.89665341e-01
-3.18032146e-01 -1.82351336e-01 3.57824951e-01 1.06783175e+00
-1.56396374e-01 -6.20933712e-01 -5.32925427e-01 3.76361310e-01
7.10927486e-01 -5.79706609e-01 7.42297649e-01 -1.29520094e+00
3.18010509e-01 -1.70389503e-01 -1.23883867e+00 -2.66205311e-01
4.45479572e-01 -5.55100143e-01 -4.02246952e-01 -6.16063416e-01
4.13463116e-02 2.15682492e-01 -3.43462735e-01 7.01313376e-01
1.28131613e-01 2.84175903e-01 -1.97780743e-01 -3.88649181e-02
-3.99105489e-01 2.09924042e-01 7.10111618e-01 -1.07792959e-01
1.39274299e-01 2.60779262e-01 -4.59466547e-01 1.30052507e+00
7.09064841e-01 -6.95199907e-01 -1.60641223e-01 -2.72430182e-01
-2.43810877e-01 -5.67621529e-01 5.16338110e-01 -7.91119874e-01
1.52417466e-01 -2.73572385e-01 6.80192828e-01 -3.31853688e-01
3.88572812e-01 -1.44321680e+00 -1.87963814e-01 6.39110327e-01
-1.06603624e-02 2.46803969e-01 3.85227144e-01 1.05452597e+00
-6.27870485e-02 -2.76860714e-01 9.84167933e-01 -3.14457744e-01
-7.00044155e-01 6.12739921e-01 -1.01074171e+00 -2.28707105e-01
1.47897851e+00 -6.20034218e-01 9.55783129e-02 -5.75695872e-01
-7.10638881e-01 4.86356728e-02 7.20014513e-01 4.30734962e-01
7.28453934e-01 -1.22562337e+00 -5.80968797e-01 5.70387125e-01
2.91041195e-01 -3.97337079e-01 1.75842330e-01 8.13506007e-01
-7.23819673e-01 -8.37708488e-02 -4.48627651e-01 -5.21259785e-01
-1.91113245e+00 7.82652020e-01 4.33052778e-01 4.42550987e-01
-6.93348587e-01 7.59765148e-01 3.46743554e-01 2.70178467e-01
5.87710559e-01 -3.43115963e-02 5.72303031e-03 6.07945547e-02
8.50333035e-01 6.86100423e-01 -4.87001278e-02 -8.44487131e-01
-6.78260505e-01 2.43297711e-01 -3.98405492e-01 1.99815363e-01
9.62040663e-01 -9.95346233e-02 -2.59480745e-01 -5.24038374e-02
1.16418731e+00 1.69799671e-01 -1.13166142e+00 -1.22440368e-01
2.82541007e-01 -1.00030565e+00 -4.57168967e-02 -1.56595305e-01
-1.47547293e+00 7.17855930e-01 7.89634168e-01 1.40351737e-02
1.23903584e+00 -3.03814590e-01 6.16375268e-01 5.05348265e-01
2.50365198e-01 -1.10212421e+00 8.21770877e-02 -4.14653867e-02
6.78453743e-01 -1.35415661e+00 1.89532712e-01 -3.98743808e-01
-2.58765459e-01 1.31340110e+00 7.44677186e-01 2.05222573e-02
7.49577940e-01 2.04922557e-01 2.03538001e-01 -1.90802470e-01
-1.85063481e-01 8.40421245e-02 3.81562769e-01 1.02712381e+00
2.66759008e-01 -4.46305156e-01 -1.37655452e-01 1.64490804e-01
3.93816978e-01 -4.45881367e-01 5.93462586e-01 1.36447275e+00
-1.76329747e-01 -1.30359149e+00 -6.29325211e-01 7.75245309e-01
-1.10841346e+00 2.82422870e-01 -8.71266067e-01 6.45151198e-01
5.72651215e-02 7.93029785e-01 3.58788937e-01 1.46225421e-03
6.54798523e-02 4.07748371e-01 5.27221382e-01 -4.22384262e-01
-3.42141718e-01 1.56766884e-02 -1.30169084e-02 -1.16524601e+00
-5.20103097e-01 -1.17482173e+00 -8.24894190e-01 -4.13170636e-01
-1.42800584e-01 -2.67419100e-01 2.10382938e-01 7.70456254e-01
4.50231224e-01 -2.96490997e-01 3.54093999e-01 -5.35281122e-01
-2.63003528e-01 -3.56030732e-01 -6.70803905e-01 5.09098649e-01
9.36119974e-01 -6.26354218e-01 -5.45326948e-01 1.32315144e-01] | [7.854331016540527, 1.3986703157424927] |
c776f3f0-e701-4c6b-aed1-1c05011dee6d | extracting-clinician-s-goals-by-what-if | 2110.15165 | null | https://arxiv.org/abs/2110.15165v3 | https://arxiv.org/pdf/2110.15165v3.pdf | Extracting Expert's Goals by What-if Interpretable Modeling | Although reinforcement learning (RL) has tremendous success in many fields, applying RL to real-world settings such as healthcare is challenging when the reward is hard to specify and no exploration is allowed. In this work, we focus on recovering clinicians' rewards in treating patients. We incorporate the what-if reasoning to explain the clinician's treatments based on their potential future outcomes. We use generalized additive models (GAMs) - a class of accurate, interpretable models - to recover the reward. In both simulation and a real-world hospital dataset, we show our model outperforms baselines. Finally, our model's explanations match several clinical guidelines when treating patients while we found the commonly-used linear model often contradicts them. | ['Anna Goldenberg', 'Rich Caruana', 'George Alexandru Adam', 'Chun-Hao Chang'] | 2021-10-28 | null | null | null | null | ['additive-models'] | ['methodology'] | [-2.29421128e-02 8.01974595e-01 -8.23320687e-01 -3.65839750e-01
-7.93883502e-01 -2.15192258e-01 5.42523712e-02 3.95104229e-01
-5.58748782e-01 1.25428140e+00 5.69350362e-01 -8.23320985e-01
-4.21788484e-01 -3.62573087e-01 -6.28247917e-01 -4.08581227e-01
-2.28575245e-01 7.80302107e-01 -5.69916964e-01 2.49056257e-02
2.57683277e-01 9.08589289e-02 -7.77875721e-01 2.93238193e-01
1.04552245e+00 4.13496763e-01 1.71335444e-01 6.75041616e-01
3.25344503e-01 1.37272775e+00 -3.54311526e-01 -3.51299524e-01
3.00765663e-01 -5.04402041e-01 -5.98392785e-01 -3.78855877e-02
-5.71061432e-01 -1.00294268e+00 -3.72769147e-01 5.52669883e-01
4.79505837e-01 1.29789263e-02 7.84722865e-01 -1.40875137e+00
-7.02188373e-01 8.48743558e-01 -3.55195820e-01 1.94172394e-02
4.46469247e-01 8.00776660e-01 9.96727645e-01 1.02998689e-01
5.16283035e-01 1.38471913e+00 4.37502891e-01 1.01963735e+00
-1.27468920e+00 -5.11157334e-01 2.98056453e-01 1.20463513e-01
-6.04224324e-01 -1.16678305e-01 2.90119171e-01 -3.72194141e-01
9.27405357e-01 -1.09710723e-01 6.94063187e-01 1.29196167e+00
6.45633042e-01 1.01088536e+00 1.34267330e+00 -2.77951896e-01
6.89718366e-01 1.88095812e-02 -8.93160701e-02 2.51727372e-01
8.26669857e-02 7.33389199e-01 -2.31922209e-01 -6.14966989e-01
1.13326454e+00 5.49882233e-01 -1.32594377e-01 -2.61970293e-02
-1.09782434e+00 1.11077869e+00 4.17026162e-01 -4.15693164e-01
-1.06769848e+00 5.89921176e-01 3.38262878e-04 2.21475884e-02
1.25486761e-01 7.53125846e-01 -5.49962580e-01 -5.62089503e-01
-6.17032588e-01 5.30888081e-01 8.68277669e-01 7.93326080e-01
1.64966304e-02 -6.46166801e-02 -6.22580051e-01 5.46786904e-01
3.43046784e-01 1.73417389e-01 3.42381984e-01 -1.56099510e+00
3.70292515e-01 2.27333307e-01 9.48240101e-01 -2.85951018e-01
-6.11165822e-01 -4.36366081e-01 -4.50550497e-01 1.87489614e-01
4.98744011e-01 -5.88511109e-01 -8.19942415e-01 1.85809350e+00
-1.17546879e-01 4.36030000e-01 3.25613290e-01 8.99143815e-01
2.06070855e-01 2.41239682e-01 5.21016240e-01 -7.27582216e-01
9.86725509e-01 -7.66499639e-01 -8.24792504e-01 -6.03837609e-01
5.88293612e-01 -2.43177697e-01 1.04260123e+00 5.30466259e-01
-1.55408776e+00 1.91945970e-01 -3.89686525e-01 5.13444602e-01
5.30284286e-01 -2.23391339e-01 8.53676140e-01 2.88931519e-01
-9.67150152e-01 8.13795984e-01 -1.01069176e+00 -1.14228539e-01
7.23558664e-01 4.12521303e-01 2.61894375e-01 -1.46124795e-01
-1.04716706e+00 1.05621195e+00 -1.03705846e-01 -1.02698416e-01
-1.24930584e+00 -7.68723965e-01 -4.23475355e-01 3.43302816e-01
7.56180525e-01 -1.13624072e+00 2.05500793e+00 -7.38644660e-01
-1.43885303e+00 4.37792778e-01 -5.02507500e-02 -9.55411017e-01
8.31132114e-01 -1.87100038e-01 3.08549628e-02 1.70385502e-02
-1.10031992e-01 8.02598178e-01 3.68848920e-01 -1.13149834e+00
-3.94632459e-01 -1.08256273e-01 1.12359531e-01 3.55425805e-01
2.69642830e-01 -2.24489480e-01 3.01271737e-01 -4.83461261e-01
-2.31890216e-01 -1.03148019e+00 -1.12975514e+00 -2.83924252e-01
-4.15863663e-01 -1.38824120e-01 -1.45916983e-01 -4.69823956e-01
1.24535942e+00 -1.64641643e+00 -2.67419308e-01 7.05966577e-02
2.74096578e-01 -2.47820079e-01 1.01957873e-01 7.01040447e-01
-3.49826366e-02 2.53654838e-01 -1.92923665e-01 -2.36295775e-01
1.09878674e-01 5.82557023e-01 -3.83874685e-01 1.85133606e-01
4.81045321e-02 1.07460213e+00 -1.10236490e+00 -4.69064474e-01
1.10397391e-01 4.21308875e-02 -1.05800271e+00 5.44003844e-01
-4.38844025e-01 5.98420441e-01 -7.13809371e-01 4.21480417e-01
2.40460202e-01 -5.68900824e-01 4.67820197e-01 7.33777046e-01
4.61259514e-01 3.28510702e-01 -8.05495262e-01 1.10169768e+00
-4.18328851e-01 -9.66370553e-02 -2.37848982e-01 -8.05740476e-01
4.34252799e-01 2.85877377e-01 7.28324473e-01 -4.32310075e-01
1.52711704e-01 3.10781617e-02 5.20503521e-01 -7.20347166e-01
3.28102559e-02 -6.31674886e-01 2.44491976e-02 8.66072834e-01
-5.54087937e-01 -2.10131660e-01 -2.05774248e-01 3.49993974e-01
1.44328654e+00 -8.55707284e-03 7.71710753e-01 -5.68080787e-03
-5.22019148e-01 2.32152522e-01 8.11458588e-01 1.11192560e+00
-4.24297512e-01 3.94566268e-01 9.10467088e-01 -3.28323036e-01
-7.70279944e-01 -1.04230797e+00 5.49491532e-02 5.50473750e-01
-1.77284554e-01 -4.43229266e-02 -5.89184284e-01 -5.57759404e-01
3.74011368e-01 1.42408216e+00 -7.64591634e-01 -3.27541441e-01
-2.19193503e-01 -8.14822793e-01 1.91666692e-01 8.61686647e-01
-2.57303804e-01 -1.36292374e+00 -1.14538407e+00 5.98964632e-01
-5.20330407e-02 -7.71025777e-01 -6.13430679e-01 1.97802946e-01
-1.04477477e+00 -1.22636473e+00 -7.03382730e-01 4.21503000e-02
6.29313231e-01 -1.42099306e-01 1.30052662e+00 -3.25075798e-02
-3.18904549e-01 8.43159497e-01 -4.78495583e-02 -7.10837066e-01
-5.46862543e-01 -6.56158149e-01 2.26354524e-01 -6.31181002e-01
2.98434168e-01 -3.20585638e-01 -1.06827247e+00 4.49638441e-02
-6.02783859e-01 8.09645355e-02 5.80339015e-01 1.11180913e+00
5.29234469e-01 -4.87924248e-01 9.71208572e-01 -1.28327000e+00
1.30318069e+00 -8.46111059e-01 -3.92784685e-01 2.59159416e-01
-1.21174729e+00 2.03759044e-01 4.88626778e-01 -7.66287804e-01
-1.04047430e+00 -4.20303904e-02 2.32268691e-01 -5.01457095e-01
-6.52854741e-02 4.71278161e-01 4.72666591e-01 6.78835809e-01
6.82991743e-01 -1.71449482e-01 2.90530473e-02 -1.52012080e-01
2.37997577e-01 4.81498599e-01 2.05372140e-01 -7.20764399e-01
-1.88299071e-03 -1.50076980e-02 -7.75598884e-02 1.73229009e-01
-8.34382236e-01 1.03973925e-01 2.27587819e-01 -6.87019378e-02
6.29481256e-01 -8.66686702e-01 -1.63724172e+00 -4.48799759e-01
-9.50178385e-01 -1.03803682e+00 -6.62677586e-01 9.21801865e-01
-1.31965947e+00 -1.09142892e-01 -7.89993405e-01 -1.43765759e+00
-7.62827098e-02 -1.43750799e+00 5.89546382e-01 4.17335570e-01
-5.27151942e-01 -1.09957755e+00 2.07498685e-01 2.74207324e-01
2.81386197e-01 5.16631156e-02 9.83952641e-01 -6.03284478e-01
-7.20737696e-01 2.12556973e-01 5.55707999e-02 -8.60607624e-02
8.15316439e-02 -2.27775469e-01 -5.98101377e-01 -1.82655647e-01
7.69519284e-02 -5.73851466e-01 5.71608007e-01 1.18983078e+00
1.53157473e+00 -7.52930760e-01 -4.20271158e-01 2.76870310e-01
1.06881130e+00 8.28645706e-01 4.00475532e-01 2.13283375e-01
-7.66971111e-02 5.96486747e-01 8.88344765e-01 1.02052641e+00
7.07381427e-01 3.28175634e-01 6.89227462e-01 1.10272346e-02
6.25005960e-01 -6.21119261e-01 2.42420360e-01 5.57109900e-02
-8.99951681e-02 -3.37255359e-01 -1.05453312e+00 2.72030413e-01
-2.43082500e+00 -9.39109266e-01 2.14760408e-01 2.24129748e+00
9.94232476e-01 2.66097337e-01 2.89786518e-01 -5.22773087e-01
2.52312392e-01 -5.93522727e-01 -1.24042749e+00 -8.48615944e-01
2.27586508e-01 -9.22058243e-03 7.85758555e-01 5.06324887e-01
-3.71297687e-01 6.43979430e-01 7.86768150e+00 3.59768748e-01
-6.67812526e-01 -5.19092493e-02 1.35306275e+00 -4.50796515e-01
-4.09714252e-01 2.76681501e-02 -3.48115474e-01 3.42224270e-01
1.18217611e+00 -3.63377810e-01 7.85047531e-01 8.87487769e-01
8.82804751e-01 -2.98664659e-01 -1.61686146e+00 9.06297445e-01
-4.35044825e-01 -1.21921182e+00 -4.41327125e-01 1.98685870e-01
7.97607422e-01 -3.06079984e-01 3.40961158e-01 5.20456314e-01
1.49253988e+00 -1.54141283e+00 4.14369762e-01 7.90066957e-01
5.78729272e-01 -7.51818597e-01 6.25574529e-01 7.57283151e-01
-1.17480785e-01 -6.25382006e-01 -2.43069842e-01 -2.88024336e-01
2.33771309e-01 2.69620627e-01 -1.68509805e+00 -2.29908922e-03
3.22028279e-01 6.37361288e-01 1.45960629e-01 9.94179010e-01
-3.41249287e-01 9.20839727e-01 -2.36135405e-02 5.44048287e-02
5.39865971e-01 -9.52792093e-02 1.54059650e-02 9.02521014e-01
2.98607975e-01 6.98011935e-01 3.12041610e-01 1.10805058e+00
-8.15707725e-03 -4.56973687e-02 -4.93136585e-01 -1.60467729e-01
3.07810038e-01 8.68966579e-01 -4.32091743e-01 -3.62219453e-01
1.77592095e-02 4.98086393e-01 2.05366060e-01 5.61650217e-01
-8.73674273e-01 5.20343184e-01 7.18102157e-01 -3.78599879e-03
-4.37485069e-01 2.64669389e-01 -4.85590726e-01 -8.45201731e-01
-5.17324269e-01 -1.01446950e+00 6.00808263e-01 -1.14353776e+00
-1.35669863e+00 1.96928412e-01 1.28965005e-01 -1.12689555e+00
-1.04319835e+00 -3.85076404e-01 -5.15531182e-01 9.23635006e-01
-1.44165707e+00 -6.37393057e-01 2.80574232e-01 4.39851999e-01
5.55339754e-01 1.79164588e-01 6.43459082e-01 -3.22514892e-01
-5.14805794e-01 4.44769293e-01 3.98455895e-02 -2.63985127e-01
8.12155128e-01 -1.33849847e+00 -5.85161373e-02 9.00370441e-03
-4.06236321e-01 5.55724859e-01 9.66551602e-01 -8.31881166e-01
-1.16722834e+00 -5.78933716e-01 3.74164909e-01 -3.60522032e-01
5.27400196e-01 2.12996468e-01 -7.30856717e-01 1.06990731e+00
2.36137554e-01 -2.05612451e-01 5.84655225e-01 -7.21311057e-03
1.66409820e-01 3.12379092e-01 -1.46263814e+00 8.05366755e-01
9.43953812e-01 -6.68926910e-02 -5.39705038e-01 6.71904445e-01
8.36988568e-01 -4.62940156e-01 -7.13212430e-01 5.90214804e-02
4.33097959e-01 -7.36564279e-01 9.37583387e-01 -1.51119578e+00
9.37366545e-01 3.54588121e-01 9.69924033e-02 -1.83253455e+00
-2.08604157e-01 -8.18993568e-01 -3.32289711e-02 4.82308596e-01
5.03607988e-01 -7.50252306e-01 7.77503788e-01 1.42228138e+00
3.42119485e-02 -1.19699502e+00 -6.95980787e-01 -4.90813673e-01
8.07012394e-02 -3.66916209e-01 6.45345390e-01 9.59581137e-01
4.64233041e-01 -1.85973883e-01 -6.02896750e-01 -5.65523393e-02
6.44346952e-01 -5.03405370e-02 3.05506945e-01 -8.31856847e-01
-9.32983935e-01 -3.85699660e-01 3.30656976e-01 -8.26907814e-01
2.17823610e-01 -6.15616500e-01 1.40180752e-01 -1.83029819e+00
7.21786499e-01 -6.99077964e-01 -5.21901906e-01 8.62568378e-01
-3.47166419e-01 -7.50066996e-01 3.22816938e-01 6.15489967e-02
-5.23321390e-01 5.18776059e-01 1.51018536e+00 9.70184281e-02
-5.10640919e-01 3.74757171e-01 -1.04455733e+00 6.80124283e-01
8.94820929e-01 -7.61343241e-01 -6.49634242e-01 -7.19363689e-02
2.60141760e-01 1.24892330e+00 3.09495836e-01 -4.64620925e-02
1.98379844e-01 -1.12264967e+00 1.88056648e-01 -2.46151760e-01
3.26562136e-01 -6.51372433e-01 1.71387300e-01 9.42335486e-01
-9.49111223e-01 2.09585562e-01 2.22913295e-01 6.85838878e-01
3.03564668e-01 -2.05951840e-01 6.26975954e-01 -4.47009295e-01
6.27898499e-02 2.78814644e-01 -3.22103173e-01 3.48223478e-01
1.14082360e+00 1.70813695e-01 -3.44621003e-01 -1.15562665e+00
-1.03559399e+00 6.16294622e-01 3.88296098e-01 -5.95162325e-02
9.44675624e-01 -9.67195272e-01 -7.59961307e-01 -4.64749366e-01
-2.17919052e-01 -1.87072814e-01 4.37767804e-01 7.54479408e-01
-9.46034864e-02 4.02966022e-01 -1.71871021e-01 -1.95532665e-01
-7.43574500e-01 8.64261448e-01 4.77303952e-01 -7.51796782e-01
-5.49946487e-01 3.24515581e-01 5.62343419e-01 -8.18043128e-02
1.83477819e-01 -5.07680297e-01 3.16797271e-02 -4.90558416e-01
3.82947683e-01 4.28003401e-01 -6.44868374e-01 1.67988241e-01
-2.45285794e-01 -1.81881323e-01 -7.44874403e-02 -3.55804265e-01
1.60117579e+00 -3.08169536e-02 4.94779497e-01 4.37229097e-01
2.98305482e-01 -5.07277310e-01 -1.72718370e+00 3.70620377e-02
1.27276495e-01 -4.39910471e-01 5.07229045e-02 -1.29876316e+00
-7.26887107e-01 9.40432847e-01 4.45865780e-01 -4.98701595e-02
1.07603955e+00 -9.89967734e-02 4.41276878e-01 3.65035057e-01
4.41154361e-01 -1.00306916e+00 1.01300672e-01 -1.96983084e-01
7.62851954e-01 -1.48172605e+00 3.21995914e-02 1.35878265e-01
-1.50321317e+00 6.93768322e-01 6.27654374e-01 -1.26116633e-01
4.50095087e-01 1.80845425e-01 1.04954960e-02 8.63639787e-02
-1.61155009e+00 5.02552725e-02 -2.38297030e-01 4.42488313e-01
3.01021099e-01 6.02830350e-01 -2.78095812e-01 1.12539852e+00
-5.98568581e-02 6.12129450e-01 1.19289982e+00 8.31952572e-01
-5.79244718e-02 -1.12395334e+00 -3.84649634e-01 8.95987809e-01
-5.58865905e-01 -1.43240169e-01 -5.55461086e-02 6.50677741e-01
-2.93615639e-01 1.01589835e+00 -1.29703462e-01 3.71779017e-02
3.48480493e-01 -1.30842075e-01 4.48404402e-01 -7.30418921e-01
-6.12058342e-01 2.71309316e-01 -1.13899522e-01 -8.21151137e-01
-1.85609907e-01 -7.24208236e-01 -1.61603332e+00 -3.49059492e-01
3.52895588e-01 1.48254456e-02 4.06159163e-01 7.81026244e-01
4.07625377e-01 7.74920106e-01 6.14002705e-01 -3.53276163e-01
-1.22978878e+00 -5.86534917e-01 -6.59391642e-01 4.02505755e-01
4.59169775e-01 -7.18933225e-01 -3.22231859e-01 -2.69675404e-01] | [4.038642883300781, 2.6845877170562744] |
76466212-75cd-442d-a840-22689aea7a14 | a-two-stage-data-association-approach-for-3d | 2101.08684 | null | https://arxiv.org/abs/2101.08684v1 | https://arxiv.org/pdf/2101.08684v1.pdf | A two-stage data association approach for 3D Multi-object Tracking | Multi-object tracking (MOT) is an integral part of any autonomous driving pipelines because itproduces trajectories which has been taken by other moving objects in the scene and helps predicttheir future motion. Thanks to the recent advances in 3D object detection enabled by deep learning,track-by-detection has become the dominant paradigm in 3D MOT. In this paradigm, a MOT systemis essentially made of an object detector and a data association algorithm which establishes track-to-detection correspondence. While 3D object detection has been actively researched, associationalgorithms for 3D MOT seem to settle at a bipartie matching formulated as a linear assignmentproblem (LAP) and solved by the Hungarian algorithm. In this paper, we adapt a two-stage dataassociation method which was successful in image-based tracking to the 3D setting, thus providingan alternative for data association for 3D MOT. Our method outperforms the baseline using one-stagebipartie matching for data association by achieving 0.587 AMOTA in NuScenes validation set. | ['Vincent Frémont', 'Minh-Quan Dao'] | 2021-01-21 | null | null | null | null | ['3d-multi-object-tracking'] | ['computer-vision'] | [-3.19922954e-01 -3.42829138e-01 -3.26029330e-01 2.13094261e-02
-6.68533564e-01 -6.50178492e-01 1.06128383e+00 2.00605497e-01
-7.93426216e-01 4.83247280e-01 -3.89147311e-01 -2.92315066e-01
-2.05284759e-01 -5.87154627e-01 -9.56326127e-01 -9.17697787e-01
4.39667627e-02 1.10958755e+00 1.06716466e+00 -2.05771267e-01
1.45309707e-02 9.21473920e-01 -2.01506281e+00 -3.42355669e-02
4.64508921e-01 8.57287109e-01 4.11524177e-01 8.92483890e-01
-2.46450529e-01 6.51289761e-01 -4.07366484e-01 -1.30140278e-02
4.35816199e-01 -2.57193744e-01 -3.54428470e-01 6.73283711e-02
9.35674191e-01 -8.67848396e-02 -2.83082843e-01 8.77675891e-01
3.26723009e-01 2.00601727e-01 6.79831088e-01 -1.63585746e+00
9.24434215e-02 1.29761636e-01 -9.08219874e-01 5.35220325e-01
1.36079401e-01 -1.07591443e-01 8.12020302e-01 -8.66136074e-01
7.67898202e-01 1.26168215e+00 7.61110783e-01 4.44810808e-01
-1.24421966e+00 -7.39945114e-01 -7.23640323e-02 4.93208766e-01
-1.18424237e+00 -3.97146791e-01 5.57743728e-01 -7.47708201e-01
8.18605661e-01 3.41511488e-01 8.87210369e-01 7.30471492e-01
4.52448726e-01 1.03097534e+00 9.35209572e-01 -3.06991190e-01
-6.39074366e-04 1.47211581e-01 2.14070976e-01 5.09768784e-01
5.68996072e-01 4.41261530e-01 -4.76102382e-01 2.74587274e-01
5.89969695e-01 -8.89471695e-02 4.08655822e-01 -6.84296310e-01
-1.15279746e+00 6.24386966e-01 5.58310926e-01 2.24165529e-01
-4.57178801e-01 3.72203797e-01 2.53089935e-01 2.52490938e-01
2.62751997e-01 -3.25782364e-03 -1.37729608e-02 4.96890992e-02
-9.32416260e-01 7.43620992e-01 4.88325655e-01 1.03885007e+00
8.10383201e-01 3.04432046e-02 -7.90217295e-02 2.00331002e-01
5.17764270e-01 9.85317945e-01 1.63313225e-01 -7.68922448e-01
1.45285830e-01 7.40531087e-01 3.30144256e-01 -8.44015002e-01
-6.92135036e-01 -3.64206284e-01 -5.67160964e-01 8.73548150e-01
6.76880360e-01 4.46313620e-02 -8.67057025e-01 1.24068666e+00
9.96988416e-01 1.18959695e-01 -2.89852649e-01 1.08007193e+00
7.15094149e-01 4.17121202e-01 -1.54427677e-01 2.16250792e-01
1.43006694e+00 -7.24711061e-01 -5.70496500e-01 -1.70748368e-01
6.19543314e-01 -6.90657794e-01 1.97500363e-01 1.67220622e-01
-8.55826080e-01 -9.11940694e-01 -9.86932039e-01 2.88989805e-02
-4.81836349e-01 -2.61556834e-01 4.76898193e-01 3.83915097e-01
-1.03851402e+00 1.82000294e-01 -8.94275188e-01 -6.77176833e-01
6.04208708e-01 5.30508578e-01 -5.02278507e-01 3.04764837e-01
-6.76272631e-01 1.31829441e+00 7.28797019e-01 -9.19629261e-02
-1.02707875e+00 -4.86360997e-01 -4.84910399e-01 -5.15600860e-01
4.73335296e-01 -7.85620987e-01 1.06141996e+00 -5.40288270e-01
-1.00560379e+00 1.30915785e+00 -9.82075036e-02 -7.55489469e-01
7.29057074e-01 -3.93649369e-01 -3.39623392e-01 -1.18064210e-01
2.59874642e-01 8.85010362e-01 9.04782772e-01 -1.18961310e+00
-1.33860815e+00 -3.75663579e-01 3.79226170e-02 1.30104274e-01
3.27413619e-01 1.96151838e-01 -5.17266572e-01 -2.14022294e-01
3.12235981e-01 -1.20194173e+00 -1.32916540e-01 3.81454617e-01
-2.62761474e-01 -6.41009510e-01 1.14657688e+00 2.60142535e-02
8.45572531e-01 -1.93542635e+00 3.04504156e-01 -1.01220861e-01
3.43171120e-01 3.21325302e-01 2.38807142e-01 2.58403569e-01
8.48617107e-02 -5.62020004e-01 3.23131502e-01 -6.43339813e-01
9.86881182e-02 1.51616067e-01 -7.28633720e-03 8.83813620e-01
2.21127421e-01 9.29400563e-01 -1.09026527e+00 -9.08441782e-01
6.63688481e-01 1.86408043e-01 -2.78969526e-01 1.28521442e-01
-3.83385867e-01 3.14511180e-01 -4.10914570e-01 8.80295873e-01
6.72542036e-01 -1.11568712e-01 -4.35373247e-01 1.75591752e-01
-8.11236739e-01 -2.41942212e-01 -1.25105679e+00 1.75744998e+00
9.19914395e-02 1.00583041e+00 2.32977763e-01 -7.04248667e-01
9.20066774e-01 2.25422129e-01 8.22260857e-01 -5.26867270e-01
3.08554173e-01 4.13135946e-01 1.14219703e-01 -4.17477638e-01
7.85112560e-01 -1.42849023e-02 -4.40691225e-02 1.64337724e-01
-1.08915143e-01 -2.11997554e-01 3.58414620e-01 5.64346500e-02
1.09265685e+00 5.68742692e-01 4.02979366e-02 -3.42171758e-01
4.31256801e-01 8.70389462e-01 3.77801090e-01 1.07362378e+00
-4.81934100e-01 3.29419941e-01 -3.37463081e-01 -6.71622038e-01
-1.24475634e+00 -9.17635620e-01 -3.81378680e-01 7.72500277e-01
5.10362446e-01 3.60724926e-02 -2.18060553e-01 -5.63597381e-01
4.18183506e-01 2.79723078e-01 -6.37036860e-01 1.38144359e-01
-7.49335408e-01 -6.11307800e-01 3.75537038e-01 3.96618515e-01
3.71582180e-01 -8.25353265e-01 -1.17261124e+00 3.99163008e-01
1.00022942e-01 -1.04793799e+00 -2.19559774e-01 4.94282812e-01
-6.12698853e-01 -1.12234175e+00 -5.33694088e-01 -5.56220710e-01
3.86596471e-01 6.06270730e-01 8.27273786e-01 -9.36888605e-02
-4.01950181e-01 1.89802155e-01 -2.84841508e-01 -1.02901721e+00
-4.15402919e-01 -9.64026377e-02 3.68189901e-01 -7.23747313e-02
7.69567132e-01 -2.15071291e-01 -3.64297271e-01 4.52632129e-01
-4.77116287e-01 1.36907130e-01 7.36100137e-01 3.32794219e-01
7.05736816e-01 -1.47341132e-01 8.33781362e-02 -2.96622992e-01
-2.64493257e-01 -4.58170563e-01 -1.16316974e+00 -1.27324864e-01
-1.79093152e-01 -4.39231023e-02 1.76132590e-01 -4.78885025e-01
-4.60594982e-01 7.17142344e-01 -1.92492977e-02 -8.47928047e-01
-4.02645081e-01 -1.95919588e-01 -8.38659033e-02 -3.10808331e-01
6.98154867e-01 1.45663530e-01 5.70050143e-02 -4.29996043e-01
4.24083918e-01 5.36764979e-01 6.12638474e-01 -1.66500852e-01
1.22424996e+00 8.62611294e-01 4.74466026e-01 -8.25452387e-01
-8.01468015e-01 -1.03382778e+00 -1.10029769e+00 -8.53843391e-01
1.24553418e+00 -9.77134526e-01 -9.60088193e-01 4.32649434e-01
-1.33580029e+00 -6.97479700e-04 -3.30692053e-01 5.60385227e-01
-4.78367507e-01 -1.41140565e-01 8.22686106e-02 -9.99416590e-01
-1.96164176e-02 -8.75226140e-01 1.26667619e+00 3.04032505e-01
-1.49316192e-01 -8.22328448e-01 3.98502886e-01 1.40105635e-01
9.45669189e-02 5.02685845e-01 2.09515810e-01 -6.81956708e-01
-9.43086445e-01 -5.76905489e-01 -2.56969959e-01 -3.87523770e-01
-2.11686313e-01 -6.99954256e-02 -9.93427575e-01 -1.90972403e-01
-2.30759636e-01 4.52370308e-02 8.19806993e-01 4.91942436e-01
5.10283589e-01 2.95153111e-01 -9.03233171e-01 4.11883891e-01
1.38597500e+00 2.28379041e-01 2.92291567e-02 8.17532420e-01
6.84717357e-01 5.45794904e-01 8.25816751e-01 2.28264481e-01
5.14231205e-01 1.22006214e+00 8.73266101e-01 -1.58123404e-01
-6.34411395e-01 8.90451670e-02 2.19896182e-01 2.32598096e-01
-1.31051525e-01 -2.49881193e-01 -1.05448699e+00 6.03947818e-01
-2.23339605e+00 -1.12691712e+00 -9.78625178e-01 2.07098627e+00
2.61173844e-01 6.34800136e-01 6.73533857e-01 4.68444526e-02
6.68511510e-01 -8.72885212e-02 -6.05962515e-01 1.20510899e-01
-2.44220674e-01 -3.71940732e-01 8.33285034e-01 3.23819607e-01
-1.53124690e+00 8.06441188e-01 5.70756483e+00 6.37304187e-01
-8.05005729e-01 6.22701496e-02 -2.60951847e-01 -5.31784706e-02
3.51497918e-01 -1.31415144e-01 -1.73308957e+00 3.16038489e-01
7.19199121e-01 -2.43465342e-02 -8.47625360e-02 7.32785046e-01
2.66099930e-01 -3.33579391e-01 -1.09809530e+00 9.49761689e-01
-9.14793685e-02 -1.37521589e+00 -1.06845893e-01 2.45429605e-01
5.06145477e-01 4.89756256e-01 -1.60743684e-01 1.42359346e-01
3.87136489e-01 -3.44671279e-01 1.12244964e+00 4.35647994e-01
2.68115669e-01 -3.76025707e-01 5.67742407e-01 6.26720369e-01
-1.49845231e+00 -1.36360586e-01 -2.51038104e-01 1.45217419e-01
3.12760562e-01 3.22266966e-01 -9.11908507e-01 7.33006954e-01
9.73736942e-01 8.06482017e-01 -4.28116083e-01 1.73240399e+00
1.40101567e-01 8.44705626e-02 -5.94354808e-01 -2.09558681e-01
4.82232600e-01 -1.35746211e-01 1.16353786e+00 1.14748430e+00
1.97940707e-01 -2.42232144e-01 5.23747444e-01 6.57810986e-01
3.02415133e-01 -3.74296397e-01 -7.27903128e-01 2.48747841e-01
3.64202857e-01 1.35774469e+00 -9.80991125e-01 -2.92648911e-01
-4.33418870e-01 4.63848650e-01 1.21961407e-01 -1.80687159e-01
-1.00461340e+00 -2.44417921e-01 7.84077764e-01 4.28226203e-01
5.90152800e-01 -4.65120047e-01 -5.66670969e-02 -6.22797191e-01
-2.22950980e-01 -3.42366219e-01 3.60161752e-01 -6.57010794e-01
-8.12435150e-01 5.18740118e-01 8.09958205e-02 -1.57957220e+00
7.17505515e-02 -6.30454481e-01 -2.89599031e-01 5.74604571e-01
-1.73137665e+00 -1.28747392e+00 -4.04683024e-01 4.17298704e-01
7.17380047e-01 -4.19723317e-02 4.24472034e-01 5.74857712e-01
-5.15785813e-01 7.17819715e-03 2.71683455e-01 -9.33882445e-02
5.99903762e-01 -1.40143192e+00 2.78062075e-01 1.09069622e+00
3.38743269e-01 1.58148371e-02 1.09606946e+00 -9.21980441e-01
-1.78353059e+00 -1.12160003e+00 7.86805928e-01 -9.91367340e-01
9.43965912e-01 -6.17057681e-01 -8.06320488e-01 5.26271105e-01
1.45689994e-01 1.49761751e-01 9.01920795e-02 -1.60502762e-01
1.39174104e-01 -2.72058189e-01 -6.68474376e-01 3.15697521e-01
1.18250406e+00 -2.37563569e-02 -6.43188417e-01 4.78105187e-01
3.90078992e-01 -7.84031808e-01 -4.94254827e-01 2.48747736e-01
6.49120808e-01 -7.98622012e-01 1.07325733e+00 -4.07595545e-01
-2.03426555e-01 -8.38016093e-01 3.09577715e-02 -7.92657197e-01
-4.18270946e-01 -5.70789158e-01 -4.63917732e-01 7.33082056e-01
1.32037565e-01 -1.67236581e-01 1.19635594e+00 1.14347503e-01
-6.54404044e-01 -1.62670538e-01 -1.21055794e+00 -1.21003211e+00
-2.98637331e-01 -3.10983390e-01 3.79737109e-01 5.01039386e-01
-5.93250513e-01 1.78631425e-01 -3.20128858e-01 3.64055485e-01
1.07923675e+00 3.48145127e-01 1.25574076e+00 -1.68042600e+00
-5.88438772e-02 -6.81445003e-01 -7.14582980e-01 -1.31009638e+00
-6.24026470e-02 -1.02847600e+00 4.52920109e-01 -1.42643356e+00
1.97595544e-03 -7.06026554e-01 -2.18683720e-01 3.77911806e-01
-1.80544239e-02 3.51674497e-01 1.75571844e-01 3.83895665e-01
-1.11887598e+00 2.76912987e-01 1.05313778e+00 -1.27818570e-01
-7.77596235e-02 2.54573673e-01 1.82201639e-01 3.58498305e-01
4.40738201e-01 -8.55705380e-01 1.14360549e-01 -4.78782326e-01
1.29159793e-01 -1.26771122e-01 6.22807562e-01 -1.46378469e+00
8.29685926e-01 5.46714589e-02 3.66895378e-01 -1.54812443e+00
7.19190776e-01 -1.32131970e+00 5.41609347e-01 8.21030557e-01
2.94430315e-01 2.32747629e-01 5.50605893e-01 6.48407400e-01
5.74959069e-02 -7.54325539e-02 7.58378863e-01 -8.35331157e-02
-1.14776969e+00 4.19665903e-01 -4.98703241e-01 -2.71261394e-01
1.52366626e+00 -7.67097235e-01 -2.32195690e-01 2.60165900e-01
-5.73513269e-01 5.64728200e-01 3.93317789e-01 6.72883809e-01
3.74277890e-01 -1.39643526e+00 -9.17545199e-01 1.01487637e-01
2.49481663e-01 9.63618979e-02 -8.60308558e-02 1.19136381e+00
-3.86338830e-01 7.03565717e-01 -2.39062071e-01 -1.29582942e+00
-1.65615392e+00 5.45664430e-01 3.66166860e-01 2.05137670e-01
-9.23605084e-01 9.84420121e-01 -5.08848056e-02 -1.28747761e-01
2.49499381e-01 -1.22198395e-01 -3.54362488e-01 3.17020655e-01
5.23835242e-01 5.16977012e-01 1.11418098e-01 -9.28076327e-01
-6.39445901e-01 7.38161147e-01 -3.22530046e-02 1.47746682e-01
1.28518140e+00 -1.22833110e-01 2.02700987e-01 7.58082211e-01
5.53311110e-01 -4.26633179e-01 -1.47400582e+00 -1.28821269e-01
2.92458564e-01 -5.02745390e-01 1.61321983e-01 -2.50065923e-01
-5.11221230e-01 6.31128132e-01 8.34692240e-01 4.76625711e-01
4.33959007e-01 2.82550007e-01 5.06003737e-01 2.38092676e-01
5.47503114e-01 -7.69247711e-01 2.99904719e-02 5.49776256e-01
5.28505564e-01 -1.72264969e+00 -1.95492264e-02 4.22549322e-02
-6.18528649e-02 1.08717108e+00 7.05855250e-01 -7.34116808e-02
6.38610721e-01 2.88418889e-01 9.76864323e-02 -4.64858860e-01
-5.31120718e-01 -6.58710659e-01 4.58264053e-01 5.34021080e-01
-3.02150417e-02 -2.11797118e-01 1.01579696e-01 -2.97761917e-01
1.04095034e-01 -2.78298825e-01 1.00226559e-01 1.04056931e+00
-7.99369872e-01 -1.07361531e+00 -7.40136981e-01 2.74908185e-01
3.34701547e-03 4.69953388e-01 -1.89260960e-01 1.25834107e+00
4.47895348e-01 8.79257143e-01 2.66271979e-01 -3.19643468e-01
5.82011402e-01 -2.63227731e-01 5.94688237e-01 -5.15323162e-01
-6.27790093e-01 4.89607304e-02 -9.78378281e-02 -4.97817457e-01
-7.47872353e-01 -1.23237264e+00 -1.35732865e+00 -1.74431980e-01
-8.46992910e-01 7.60567710e-02 9.74262655e-01 1.00289083e+00
1.90558210e-01 6.40647233e-01 2.48263732e-01 -1.17788291e+00
1.00076562e-02 -5.30579388e-01 -3.03312033e-01 1.48717642e-01
6.61362469e-01 -1.07029915e+00 -2.79358793e-02 -2.52743900e-01] | [6.606133937835693, -2.2527785301208496] |
6399c356-3d55-4da2-a1e3-877627045829 | fdvts-s-solution-for-2nd-cov19d-competition | 2207.01758 | null | https://arxiv.org/abs/2207.01758v1 | https://arxiv.org/pdf/2207.01758v1.pdf | FDVTS's Solution for 2nd COV19D Competition on COVID-19 Detection and Severity Analysis | This paper presents our solution for the 2nd COVID-19 Competition, occurring in the framework of the AIMIA Workshop in the European Conference on Computer Vision (ECCV 2022). In our approach, we employ an effective 3D Contrastive Mixup Classification network for COVID-19 diagnosis on chest CT images, which is composed of contrastive representation learning and mixup classification. For the COVID-19 detection challenge, our approach reaches 0.9245 macro F1 score on 484 validation CT scans, which significantly outperforms the baseline method by 16.5%. In the COVID-19 severity detection challenge, our approach achieves 0.7186 macro F1 score on 61 validation samples, which also surpasses the baseline by 8.86%. | ['Yuejie Zhang', 'Rui Feng', 'Jilan Xu', 'Junlin Hou'] | 2022-07-05 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [ 3.24591100e-01 1.18910149e-02 -1.68820411e-01 1.83212459e-02
-1.36594343e+00 -3.70311737e-01 5.93759775e-01 3.73611093e-01
-6.38448536e-01 4.75043952e-01 1.11559853e-01 -2.23326281e-01
8.86700824e-02 -3.49594265e-01 -3.03245008e-01 -5.06377399e-01
6.76617324e-02 6.67171419e-01 4.14465278e-01 1.50683805e-01
-3.48101020e-01 4.41759855e-01 -1.16987634e+00 5.73083043e-01
2.98839748e-01 1.24113810e+00 1.67492434e-01 1.23169756e+00
4.35700387e-01 8.96580577e-01 -5.57835817e-01 -1.67267382e-01
4.89844419e-02 -2.43242189e-01 -7.93307900e-01 -3.79653543e-01
8.40085387e-01 -2.71557361e-01 -4.86292779e-01 7.85653889e-01
6.56703413e-01 -1.74565852e-01 1.23811269e+00 -1.04757226e+00
-4.42377524e-03 5.02143681e-01 -9.84619617e-01 9.66493726e-01
-7.39288470e-03 5.39816134e-02 8.96337390e-01 -1.01013470e+00
6.98913395e-01 8.08885157e-01 9.47547913e-01 4.91451114e-01
-9.65813160e-01 -6.37870610e-01 -2.52008349e-01 1.78548992e-01
-1.36029899e+00 2.00509697e-01 3.09216082e-01 -8.07793558e-01
9.01134610e-01 4.85491365e-01 6.13039076e-01 1.26810360e+00
3.67582589e-01 1.01447248e+00 1.19183230e+00 -2.07151607e-01
-2.04750057e-02 -5.53990267e-02 5.68522036e-01 8.06156397e-01
4.07341331e-01 2.65502989e-01 1.47448167e-01 -3.15407157e-01
7.53525436e-01 1.31209260e-02 -1.03646532e-01 -2.13549271e-01
-1.11181021e+00 9.00436699e-01 6.45961702e-01 3.80989134e-01
-5.22671402e-01 1.82335034e-01 7.11520791e-01 -1.24544732e-01
2.31099188e-01 3.74782830e-01 -1.26486436e-01 1.35224566e-01
-8.88244629e-01 3.56919676e-01 3.58429790e-01 2.96566010e-01
-2.36606508e-01 2.25020722e-02 -6.82732284e-01 1.06526673e+00
2.43149355e-01 7.27509916e-01 6.65728331e-01 -7.06915855e-01
3.97585034e-01 2.07651794e-01 -1.11758508e-01 -6.25651836e-01
-8.90896082e-01 -7.83287823e-01 -1.03330767e+00 2.65967786e-01
1.91332117e-01 -3.16917330e-01 -1.25408530e+00 1.46886826e+00
2.16347873e-02 3.46384227e-01 -2.16839351e-02 8.66284430e-01
1.61412323e+00 4.66689676e-01 2.30255038e-01 -3.21073085e-01
1.53454185e+00 -1.07204115e+00 -3.39397043e-01 -2.30057328e-03
6.12966537e-01 -6.05645120e-01 6.16362870e-01 6.46446228e-01
-1.20488048e+00 -5.85340083e-01 -1.25462067e+00 4.14736927e-01
-1.27819523e-01 1.35681182e-01 2.95663893e-01 7.04654157e-01
-8.81905854e-01 4.40058634e-02 -6.91537082e-01 1.93609353e-02
5.32908320e-01 1.92988023e-01 -8.79191160e-02 -3.95099729e-01
-9.98763382e-01 1.07048595e+00 1.56510860e-01 -2.91863799e-01
-1.37705100e+00 -1.19028342e+00 -5.54435074e-01 -2.98839390e-01
2.75767326e-01 -9.21673536e-01 1.46966326e+00 -2.91441292e-01
-6.89744413e-01 1.21739781e+00 3.34142685e-01 -7.73740649e-01
9.95091677e-01 -8.65205675e-02 -6.07331574e-01 4.99445379e-01
7.66801685e-02 6.69666648e-01 8.00227940e-01 -1.20895314e+00
-8.91578853e-01 -2.26482615e-01 -1.28587574e-01 1.58419415e-01
4.08051401e-01 4.98680659e-02 -5.43397129e-01 -7.84341812e-01
-2.19610438e-01 -1.09276092e+00 -5.93051016e-01 -1.84920803e-01
-1.76058441e-01 -1.24593966e-01 9.44861770e-01 -6.27934039e-01
1.01727319e+00 -2.01731610e+00 -7.51920566e-02 2.29339838e-01
9.96885955e-01 6.65909111e-01 2.41431314e-02 -4.04728383e-01
-5.01501739e-01 -1.32755756e-01 -4.79703695e-01 -1.72391936e-01
-3.87152672e-01 -1.71512626e-02 1.05321497e-01 4.47895765e-01
1.07167162e-01 9.00921524e-01 -8.76174152e-01 -6.14663541e-01
4.29344475e-01 4.44445163e-01 -3.19671094e-01 1.05916038e-01
1.91777319e-01 1.43249035e-01 -2.94200748e-01 5.06612837e-01
9.23497915e-01 -5.75911880e-01 1.01028606e-01 -1.63833082e-01
3.33610207e-01 -2.27305785e-01 -9.70075667e-01 1.28174758e+00
-6.23024285e-01 6.51923358e-01 -5.04004322e-02 -9.88342106e-01
4.62747544e-01 4.41400081e-01 9.65053678e-01 -5.23659706e-01
2.85158217e-01 2.39222318e-01 3.95084694e-02 -3.98505449e-01
2.37385765e-01 -1.79653004e-01 -1.26850605e-01 2.93058217e-01
-8.44806284e-02 -4.04694945e-01 1.22155301e-01 2.54376799e-01
1.42973757e+00 -5.94925284e-01 4.18711990e-01 -2.02386037e-01
8.03246200e-01 -6.06442355e-02 2.29585379e-01 1.00665152e+00
-5.51412344e-01 1.01647234e+00 3.05362552e-01 -3.92137587e-01
-7.22569525e-01 -1.44645727e+00 -5.59388518e-01 5.33215225e-01
-8.38998035e-02 -3.16905528e-01 -6.19200408e-01 -1.41472375e+00
-1.13419265e-01 3.27747524e-01 -8.83804917e-01 -1.59692615e-02
-8.04921031e-01 -9.78413343e-01 8.21862042e-01 9.41401303e-01
1.88935101e-01 -9.77719486e-01 -6.79005027e-01 1.69391334e-01
-2.43075550e-01 -1.35055435e+00 -2.93891847e-01 1.59554422e-01
-6.04921222e-01 -1.30763233e+00 -1.18614590e+00 -8.09085310e-01
2.64430374e-01 2.39949331e-01 1.25183284e+00 1.86265126e-01
-9.06850874e-01 4.16521430e-01 -2.58264482e-01 -5.12983501e-01
-4.92602050e-01 -7.81992450e-02 -2.50138044e-01 -5.20219266e-01
-8.20641033e-03 4.34319898e-02 -6.19918644e-01 9.09365118e-02
-6.05487704e-01 -1.15242876e-01 5.49792290e-01 1.15872633e+00
9.82121706e-01 -3.43657583e-01 5.72690725e-01 -1.02803051e+00
3.73040020e-01 -5.06081998e-01 -5.75694323e-01 4.48888950e-02
-7.57890880e-01 -5.46573043e-01 2.51132786e-01 -1.54525980e-01
-8.07673573e-01 6.43594656e-03 -4.08000410e-01 -7.60282338e-01
-5.98543650e-03 2.72410214e-01 6.46251023e-01 7.64392763e-02
8.39118838e-01 -4.14762720e-02 -1.56481728e-01 -2.67111748e-01
1.75552443e-01 6.13002360e-01 8.97812843e-01 -1.33086532e-01
5.03189862e-01 4.25932765e-01 3.24088186e-01 -8.49741280e-01
-9.37634826e-01 -7.66637325e-01 -3.37800533e-01 -6.59305677e-02
1.31483579e+00 -1.11749709e+00 -2.61160016e-01 4.28798437e-01
-8.18115115e-01 1.31797371e-02 -5.12608886e-01 7.03138053e-01
-5.30489504e-01 4.23458278e-01 -7.39644587e-01 -3.69814813e-01
-7.54445493e-01 -1.54924822e+00 9.70421970e-01 2.52653025e-02
-2.49825493e-01 -9.37686980e-01 3.26216251e-01 5.84808946e-01
3.55414510e-01 6.03048325e-01 7.21620500e-01 -7.73794889e-01
5.74558526e-02 -3.39112163e-01 -5.82874179e-01 5.61100066e-01
-2.24891201e-01 -1.76701546e-01 -1.16591287e+00 -2.83775359e-01
-3.82844031e-01 -4.20867085e-01 1.33541679e+00 9.33159351e-01
1.41054296e+00 5.18938482e-01 -4.71055955e-01 5.90629518e-01
1.44406378e+00 3.42740327e-01 3.35380077e-01 -4.50704768e-02
7.74816513e-01 8.93439054e-02 5.38511693e-01 5.19595921e-01
1.38289869e-01 6.25585020e-01 7.36821592e-01 -4.02808845e-01
-7.29998708e-01 1.76791593e-01 -2.37516820e-01 5.25288165e-01
-2.28737503e-01 -1.10954396e-01 -1.28838682e+00 5.53853989e-01
-1.44788277e+00 -7.63698339e-01 -6.37876630e-01 1.73195577e+00
3.28961521e-01 2.66029656e-01 2.35845834e-01 5.48727103e-02
6.44725680e-01 2.61930883e-01 -1.91482991e-01 -4.90061611e-01
1.07172824e-01 7.47844100e-01 3.99743676e-01 2.57039845e-01
-1.64130235e+00 3.82242262e-01 7.03746319e+00 8.49825144e-01
-1.08511424e+00 4.83069122e-01 6.73867345e-01 -6.25025108e-02
1.05305612e-01 -7.52973676e-01 -7.54046202e-01 3.83906990e-01
9.34118748e-01 4.45888750e-02 -2.61187047e-01 7.89522588e-01
-1.24687776e-01 6.63580000e-02 -7.84270525e-01 1.28913903e+00
3.36719573e-01 -1.57154667e+00 -2.75091738e-01 3.01100183e-02
9.95124638e-01 6.35777295e-01 3.14423770e-01 7.01972187e-01
1.66959375e-01 -1.26114047e+00 2.96350360e-01 2.45244056e-01
1.29402196e+00 -8.00937593e-01 1.10311079e+00 8.18938091e-02
-1.24838102e+00 1.02930285e-01 -7.52715692e-02 5.93904436e-01
7.95655996e-02 4.55489814e-01 -1.31931281e+00 5.96030712e-01
7.36698627e-01 5.88027000e-01 -6.21770203e-01 1.30801284e+00
-1.20687544e-01 7.29463160e-01 -2.39525229e-01 8.91971365e-02
2.89356887e-01 5.35279989e-01 5.97177267e-01 1.78407538e+00
-3.27800065e-02 1.25299647e-01 2.63447881e-01 2.75075465e-01
-3.73953223e-01 -4.69646864e-02 -4.80053186e-01 4.01424080e-01
1.19536445e-01 1.37749481e+00 -4.96618003e-01 -8.52035165e-01
-1.35104775e-01 5.76462448e-01 -2.71545589e-01 -5.26114106e-02
-1.30400240e+00 -6.95255101e-02 1.93231463e-01 -1.09604448e-01
6.42167211e-01 4.78354454e-01 -3.57164502e-01 -8.56439352e-01
-2.27186933e-01 -9.00762320e-01 9.07522261e-01 -7.24742472e-01
-1.42464626e+00 8.15112054e-01 2.41642341e-01 -1.29023957e+00
-3.58675122e-01 -7.99450278e-01 -5.93485057e-01 7.00033724e-01
-1.52554607e+00 -1.23986363e+00 -5.84494948e-01 4.21582490e-01
6.34825528e-01 -2.11765334e-01 8.30718815e-01 2.78176785e-01
-2.70529062e-01 8.78804624e-01 -3.01592704e-03 4.81707230e-02
6.30764365e-01 -1.28547382e+00 2.31708884e-01 5.77290297e-01
-9.81733054e-02 -1.20929368e-01 3.11061144e-01 -3.42662871e-01
-7.70363450e-01 -1.23015261e+00 7.11386383e-01 -5.36370635e-01
5.27632058e-01 2.33443633e-01 -7.56348431e-01 4.29090679e-01
1.83071047e-01 5.51569998e-01 5.52711904e-01 -3.08215231e-01
-4.97691184e-01 1.69138506e-01 -1.47667050e+00 1.84242785e-01
7.32566655e-01 -1.57272115e-01 -8.83895099e-01 4.54804778e-01
5.73485196e-01 -7.08900452e-01 -1.13876700e+00 9.10013318e-01
4.96301293e-01 -6.17915988e-01 1.28907812e+00 -5.44234097e-01
5.19974828e-01 -9.83729437e-02 -3.19860965e-01 -1.05394733e+00
-4.38958079e-01 2.71666385e-02 -9.42372829e-02 3.97068739e-01
5.09006858e-01 -2.35581070e-01 7.52325475e-01 -3.01697314e-01
-1.76009342e-01 -1.00898898e+00 -9.90755856e-01 -6.34033144e-01
4.48946208e-01 -7.10776925e-01 1.01472177e-01 9.07219231e-01
-5.25793016e-01 3.66374642e-01 -1.26356781e-01 -3.53275724e-02
7.95013666e-01 -8.65778774e-02 4.70302731e-01 -1.20660043e+00
-7.10949779e-01 -7.93503404e-01 -5.62766552e-01 -6.46176279e-01
-2.13259205e-01 -1.02654696e+00 -2.75472671e-01 -1.67344487e+00
8.27187657e-01 -2.85836816e-01 -7.31529593e-01 1.48346037e-01
-2.97731102e-01 7.96185672e-01 4.19869155e-01 3.87749225e-02
-4.12734360e-01 -1.96684867e-01 1.37107360e+00 -5.46545982e-01
2.12785304e-01 1.97245568e-01 -5.56882083e-01 8.91871214e-01
6.47457659e-01 -4.86673087e-01 -3.53381425e-01 3.77389938e-02
-2.82689482e-01 4.58620012e-01 4.01794344e-01 -1.00388038e+00
-3.56012613e-01 5.87975718e-02 6.26008689e-01 -1.15373516e+00
5.58939695e-01 -6.09272361e-01 -1.49193421e-01 1.11818242e+00
-3.27989012e-01 1.62221491e-01 3.08058977e-01 4.32725549e-01
-1.67792484e-01 -3.37126344e-01 1.27649117e+00 -1.13766685e-01
-5.79453051e-01 5.09023547e-01 -5.70894778e-01 6.36355996e-01
1.31667459e+00 7.28680715e-02 -4.69728678e-01 5.98817840e-02
-1.00059140e+00 1.24614254e-01 -1.72392637e-01 2.76865959e-01
8.47402811e-01 -1.04744077e+00 -1.21975601e+00 2.08082944e-01
3.64009589e-01 -3.50796022e-02 5.54552853e-01 1.19435954e+00
-7.61810660e-01 3.19265127e-01 -5.86444996e-02 -1.05640316e+00
-1.62748086e+00 4.81333137e-01 6.88221991e-01 -9.78437185e-01
-8.35854709e-01 1.10237217e+00 5.49202740e-01 -1.64971396e-03
4.28355858e-02 -3.09733272e-01 -5.14786839e-01 4.61607501e-02
7.28729665e-01 4.12299812e-01 3.81139427e-01 -6.03801370e-01
-6.92727506e-01 5.98039269e-01 -5.73076487e-01 -5.62482104e-02
1.17846310e+00 5.23597360e-01 4.62326825e-01 5.84544279e-02
1.52353942e+00 -1.92132473e-01 -6.54181123e-01 -1.42617106e-01
-2.53590018e-01 -1.97967291e-01 4.19422209e-01 -1.12368238e+00
-1.09293854e+00 8.96373212e-01 1.26008630e+00 -1.66026980e-01
9.70468640e-01 3.31869543e-01 8.02622676e-01 -4.34130542e-02
-1.27525419e-01 -7.90885329e-01 2.26978689e-01 4.92179930e-01
8.12538266e-01 -1.35140860e+00 1.53048858e-01 -3.21709871e-01
-9.49808836e-01 9.13088858e-01 4.47657317e-01 -4.94337142e-01
8.37396741e-01 3.82194161e-01 2.75805965e-02 -4.86057669e-01
-9.00272787e-01 -3.71102169e-02 5.72240293e-01 8.60694766e-01
3.29315126e-01 6.07413292e-01 -4.53446358e-01 5.68346441e-01
6.37678280e-02 -2.01687276e-01 5.76378167e-01 7.59273112e-01
-4.32895213e-01 -6.67676389e-01 -3.45074922e-01 6.87589943e-01
-8.23346376e-01 5.19863926e-02 -2.27299094e-01 1.04407859e+00
2.55153090e-01 5.98687649e-01 -4.16710638e-02 -4.03532982e-01
4.50948149e-01 -3.50597918e-01 6.20739639e-01 -5.82772195e-01
-8.13054740e-01 2.45150939e-01 2.79925376e-01 -3.84604514e-01
-2.92826593e-01 -7.65892625e-01 -1.33546555e+00 1.50957838e-01
-1.11219361e-01 -8.93625394e-02 5.47136128e-01 6.03610992e-01
-2.83112854e-01 1.08645058e+00 6.27089500e-01 -2.64380246e-01
-7.39410222e-01 -7.99302936e-01 -4.91891921e-01 2.35073134e-01
5.62161803e-01 -8.56610060e-01 -1.99597895e-01 -9.35240239e-02] | [15.358924865722656, -1.9139043092727661] |
e2f498c7-420e-4cc1-b686-6627a7429770 | causal-modeling-of-soil-processes-for | 2211.05675 | null | https://arxiv.org/abs/2211.05675v1 | https://arxiv.org/pdf/2211.05675v1.pdf | Causal Modeling of Soil Processes for Improved Generalization | Measuring and monitoring soil organic carbon is critical for agricultural productivity and for addressing critical environmental problems. Soil organic carbon not only enriches nutrition in soil, but also has a gamut of co-benefits such as improving water storage and limiting physical erosion. Despite a litany of work in soil organic carbon estimation, current approaches do not generalize well across soil conditions and management practices. We empirically show that explicit modeling of cause-and-effect relationships among the soil processes improves the out-of-distribution generalizability of prediction models. We provide a comparative analysis of soil organic carbon estimation models where the skeleton is estimated using causal discovery methods. Our framework provide an average improvement of 81% in test mean squared error and 52% in test mean absolute error. | ['Ranveer Chandra', 'Emre Kiciman', 'John Crawford', 'Eduardo Rodrigues', 'Sara Malvar', 'Andy Neal', 'Swati Sharma', 'Somya Sharma'] | 2022-11-10 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [ 5.41942596e-01 6.95533976e-02 -9.24180388e-01 4.72773537e-02
-1.79550216e-01 -6.20900393e-01 4.33076769e-01 6.68009937e-01
-8.08997750e-02 1.30940735e+00 2.85599649e-01 -1.02905357e+00
-5.41835308e-01 -1.34333217e+00 -9.91638601e-01 -7.01716781e-01
-3.04369539e-01 9.56892781e-03 2.93714106e-01 1.22091189e-01
1.53891265e-01 2.69598991e-01 -1.55925250e+00 5.39335683e-02
1.05048478e+00 4.46977645e-01 5.79028964e-01 6.20132744e-01
1.38705317e-02 2.10175991e-01 -5.41923270e-02 1.86107814e-01
1.17576011e-02 -3.08610737e-01 -4.37559366e-01 -3.10118556e-01
4.68104742e-02 -4.51097310e-01 1.50679141e-01 1.02050316e+00
3.67941380e-01 -2.89064646e-01 9.38169837e-01 -8.61729562e-01
-5.03263950e-01 8.26249719e-01 -9.83313203e-01 -2.16328353e-01
-6.81024138e-03 1.10076927e-01 9.94863093e-01 -4.46969807e-01
3.83599639e-01 1.40227747e+00 8.66601765e-01 -5.48968948e-02
-1.48701596e+00 -7.36112595e-01 2.75218606e-01 -9.43832174e-02
-8.88379514e-01 -4.65800673e-01 -3.59198093e-01 -5.97674429e-01
8.03129852e-01 3.27368915e-01 9.65083301e-01 4.22514319e-01
8.64198089e-01 5.93000710e-01 9.81954336e-01 -4.56424713e-01
3.25268209e-01 -1.03447653e-01 4.55690175e-02 5.03998548e-02
1.16768539e+00 3.85355949e-01 -2.40196079e-01 -3.41763467e-01
8.39059651e-01 1.53119311e-01 6.01153560e-02 -3.67097765e-01
-8.84186208e-01 7.67864764e-01 3.63727719e-01 -1.99638918e-01
-4.94222969e-01 7.57713079e-01 1.07357979e-01 2.14938801e-02
4.76470381e-01 4.66005862e-01 -9.31646347e-01 2.56707311e-01
-8.11606407e-01 5.15171051e-01 6.50006771e-01 7.40826905e-01
7.56239951e-01 1.35343045e-01 8.45336765e-02 7.02251136e-01
7.23311126e-01 1.36208189e+00 1.92593466e-02 -1.19725204e+00
9.53473598e-02 3.59728158e-01 5.16158581e-01 -9.24327493e-01
-2.97489047e-01 -3.95238608e-01 -5.31914353e-01 1.15427524e-01
3.52455437e-01 -2.45966494e-01 -8.36436689e-01 1.79901016e+00
1.95149720e-01 -3.16675991e-01 -2.19088972e-01 2.10035101e-01
6.34770840e-02 5.67888379e-01 8.16838980e-01 -4.70826626e-01
1.28816879e+00 -9.16600898e-02 -7.17268348e-01 -5.23061633e-01
5.39110363e-01 -5.64621150e-01 7.02566504e-01 1.40592651e-02
-1.02901006e+00 1.28961205e-01 -7.59687781e-01 5.99492073e-01
-4.48121488e-01 -1.58503696e-01 1.22523677e+00 7.83985138e-01
-6.17780030e-01 6.16234124e-01 -1.32012773e+00 -5.52216470e-01
4.56073195e-01 7.80969188e-02 -2.25999296e-01 -7.70499706e-02
-9.88326192e-01 1.02928138e+00 3.25785160e-01 7.70297125e-02
-8.10208857e-01 -1.21719790e+00 -4.16153908e-01 4.85356539e-01
2.07033917e-01 -6.02683127e-01 1.03286159e+00 -4.90053564e-01
-1.05166483e+00 3.49598795e-01 -2.95276791e-01 -3.28863204e-01
2.93319792e-01 -6.00173950e-01 -4.60807122e-02 -2.85190284e-01
6.31057620e-01 6.43208444e-01 1.06913783e-01 -1.32024562e+00
-7.36534894e-01 -5.97563148e-01 -2.68270224e-01 2.28288472e-02
-2.64984220e-01 -1.83152616e-01 5.90491474e-01 -4.04478669e-01
5.65726042e-01 -8.78502429e-01 -4.63819981e-01 1.70065001e-01
-4.62445281e-02 6.92623913e-01 3.86084467e-01 -6.65014684e-01
1.15197229e+00 -1.53292513e+00 -3.92894596e-01 1.91535652e-01
-2.63723701e-01 -9.85104814e-02 1.51866516e-02 5.45155346e-01
-5.38659431e-02 5.24449050e-01 -5.60643196e-01 7.27097571e-01
-1.38183668e-01 2.18999252e-01 -1.96144238e-01 3.76185805e-01
4.55757648e-01 6.33205771e-01 -1.12458682e+00 -1.79034546e-02
2.22889274e-01 1.93116978e-01 -8.44304681e-01 -3.93084347e-01
-2.71445662e-01 -8.64190310e-02 -2.45267570e-01 7.85020351e-01
1.06492245e+00 -1.44519448e-01 9.06160772e-01 5.25518775e-01
-5.52598655e-01 4.60977018e-01 -1.19918764e+00 9.89441216e-01
-3.95438880e-01 3.12058300e-01 5.91354705e-02 -9.14124250e-01
5.73517203e-01 9.68465954e-02 4.45969433e-01 -4.78456318e-01
-5.76361418e-01 4.19257581e-01 3.15633327e-01 -3.08631271e-01
3.85099500e-01 -4.30271715e-01 4.83330786e-01 2.40204260e-01
-4.19475824e-01 -7.31539130e-02 -1.98155399e-02 -1.01257004e-01
1.09138107e+00 3.90778840e-01 7.30834424e-01 -1.18750548e+00
-7.95533359e-02 1.46261573e-01 5.80850244e-01 5.36576092e-01
1.65436521e-01 -1.65222451e-01 8.03379714e-01 2.08388194e-01
-1.06713259e+00 -1.32979774e+00 -8.88679385e-01 1.11478496e+00
-2.66615480e-01 -2.10629031e-02 -9.49512273e-02 2.89499182e-02
1.03785706e+00 1.00320661e+00 -8.22276175e-01 -1.79607496e-01
-2.89807320e-02 -1.45507383e+00 5.17567754e-01 7.96339333e-01
3.91250372e-01 -6.67901993e-01 -6.58979177e-01 4.33304548e-01
-5.66528440e-02 -4.36039120e-01 4.29782093e-01 7.87657142e-01
-1.35463846e+00 -1.18871164e+00 -5.30436635e-01 -8.12060907e-02
3.67131114e-01 7.65230417e-01 1.09861171e+00 -6.24256060e-02
-6.21362254e-02 -7.64254704e-02 -7.50602856e-02 -1.02965999e+00
-4.51653451e-01 1.25917196e-01 1.01430729e-01 -8.29043388e-01
5.15918493e-01 -7.28292525e-01 -7.08178997e-01 2.52517968e-01
-6.55535340e-01 -2.07028031e-01 6.78788006e-01 8.80995870e-01
5.39230168e-01 -3.93324979e-02 9.61182952e-01 -8.76783967e-01
-4.11693119e-02 -9.72401321e-01 -6.71335876e-01 2.67859280e-01
-1.10269821e+00 -5.66723682e-02 -2.92407990e-01 -2.94793963e-01
-1.11881256e+00 -2.09879339e-01 4.55711275e-01 7.62764573e-01
2.60633659e-02 1.14364243e+00 -2.47262418e-01 4.85975087e-01
6.20907724e-01 -3.27343255e-01 8.42931494e-02 -2.31429189e-01
5.90087511e-02 2.58293480e-01 3.84620816e-01 -7.38214016e-01
6.81445181e-01 2.56645590e-01 3.85411024e-01 -1.15997660e+00
-6.79646373e-01 -2.50463158e-01 -5.61317742e-01 -5.16277440e-02
5.75458765e-01 -1.19363785e+00 -7.07060635e-01 1.52727708e-01
-7.04507470e-01 -6.73454821e-01 -3.92656438e-02 9.53544319e-01
-4.06063855e-01 -5.85214421e-02 -9.79045331e-02 -1.02049494e+00
-4.56179902e-02 -7.04210520e-01 6.35714471e-01 -4.22517210e-02
-1.07047014e-01 -1.11648238e+00 1.24443665e-01 -2.03347549e-01
4.54410493e-01 3.22288185e-01 1.16291463e+00 2.91074812e-01
-5.01738310e-01 5.48958778e-05 -6.56543911e-01 5.53779267e-02
6.22480571e-01 2.73006290e-01 -1.07885909e+00 -1.10504404e-01
-4.83445495e-01 8.23783204e-02 1.24248683e+00 1.40294862e+00
8.04398596e-01 9.22941789e-03 -5.58541477e-01 3.81186306e-01
1.61723328e+00 1.46531001e-01 1.04954588e+00 6.25833452e-01
2.98796624e-01 8.80679309e-01 7.16461360e-01 6.83228970e-01
4.94689196e-02 1.94862261e-01 6.70472443e-01 6.02149256e-02
4.81606349e-02 -1.92300767e-01 5.09186506e-01 1.50460556e-01
-2.44124115e-01 -1.07419126e-01 -1.20517254e+00 7.79327273e-01
-1.83243656e+00 -1.44539630e+00 -7.19589770e-01 2.21309853e+00
1.12232149e+00 -3.11655924e-03 -6.67507872e-02 1.43981919e-01
6.40091121e-01 -7.05806762e-02 -5.81701756e-01 -1.23518638e-01
-2.29745105e-01 1.43289611e-01 1.47334826e+00 6.32812321e-01
-7.05141723e-01 7.83668518e-01 7.86522150e+00 1.82166725e-01
-6.82986259e-01 -2.31511742e-01 2.66690910e-01 2.03196004e-01
-7.28666067e-01 4.79105026e-01 -8.96755397e-01 2.98092961e-02
1.11076868e+00 -3.99975330e-01 1.28222451e-01 6.29260540e-01
8.27374160e-01 -8.19045603e-01 -8.97623420e-01 -2.62297004e-01
-7.51113474e-01 -1.14829540e+00 -7.56364539e-02 4.01424378e-01
1.04170299e+00 2.53308658e-02 -1.36024296e-01 -7.72367269e-02
9.39088345e-01 -9.14684057e-01 7.64168978e-01 4.65621680e-01
1.00022876e+00 -5.68360388e-01 7.57447124e-01 -5.45821264e-02
-1.16719437e+00 -9.52371731e-02 -8.51172149e-01 -5.81791282e-01
-2.58608311e-01 1.15391004e+00 -4.72759604e-01 2.69910038e-01
7.75236785e-01 7.57952154e-01 -4.53537792e-01 1.06764078e+00
-3.82747918e-01 1.11496913e+00 -4.88900602e-01 1.36923090e-01
5.44569604e-02 -1.96735591e-01 8.55132192e-02 1.15957379e+00
4.55800414e-01 1.86628431e-01 -2.75617927e-01 9.16906714e-01
4.03221637e-01 1.25428334e-01 -7.37469673e-01 -2.59469211e-01
1.00525153e+00 5.76822996e-01 -7.14794159e-01 -3.53360996e-02
-2.30539650e-01 -1.39880423e-02 -4.87597138e-01 4.12305266e-01
-5.93182683e-01 -3.41841191e-01 9.40870464e-01 4.54539895e-01
3.13801408e-01 -3.07705224e-01 -7.21534789e-01 -8.18986177e-01
-1.57012910e-01 -6.16989017e-01 -1.25957161e-01 -7.00936258e-01
-8.14240575e-01 -1.06612980e+00 3.95859718e-01 -6.02282703e-01
1.32278875e-01 -7.05902994e-01 -5.02816856e-01 1.16554034e+00
-1.68905139e+00 -9.00071025e-01 -4.48240310e-01 -3.92707676e-01
6.40006214e-02 2.13554636e-01 1.05995810e+00 3.35631892e-02
-3.13776702e-01 1.29295185e-01 9.73299742e-01 -7.13482618e-01
8.86918128e-01 -1.28936160e+00 4.14940864e-01 6.17291749e-01
-1.00602984e+00 8.00431013e-01 8.01184535e-01 -1.40539289e+00
-1.36087382e+00 -1.26949382e+00 7.01057553e-01 -9.26956907e-02
1.04143488e+00 5.17544709e-02 -8.29245031e-01 7.28573859e-01
-4.13883924e-01 -5.63622892e-01 6.04664207e-01 7.70998836e-01
-3.66918266e-01 -1.70488328e-01 -9.69190061e-01 4.06859875e-01
7.50777006e-01 -2.54118413e-01 1.55672282e-01 2.21029490e-01
5.54824412e-01 2.21088409e-01 -1.22960472e+00 6.14643097e-01
1.47544038e+00 -3.94800454e-01 1.02966642e+00 -7.86189258e-01
1.03939235e+00 -2.03440841e-02 -7.77287960e-01 -1.03378904e+00
-8.25481951e-01 1.62629098e-01 1.37122840e-01 1.19044876e+00
6.40189588e-01 -6.25839591e-01 7.13866770e-01 4.06759471e-01
-5.70961498e-02 -2.55253613e-01 -1.21046916e-01 -1.04712784e+00
8.05916965e-01 -2.09481835e-01 5.69232106e-01 8.89745057e-01
-2.11318240e-01 -1.05159834e-01 1.35944396e-01 5.35764337e-01
9.17731702e-01 -3.77232373e-01 7.11595356e-01 -1.67334640e+00
-1.94586843e-01 -4.91357476e-01 -4.20325771e-02 -4.64889586e-01
-1.71095300e-02 -7.86036909e-01 6.27980977e-02 -1.78658772e+00
7.47348726e-01 -7.32376873e-01 -1.79048538e-01 7.07775593e-01
-5.87687910e-01 -3.06042314e-01 -8.67762119e-02 7.30190286e-03
3.92750084e-01 3.21959376e-01 7.77036667e-01 -2.90424019e-01
-5.26633322e-01 -7.00012892e-02 -8.99721682e-01 6.46736026e-01
9.33206379e-01 -7.45234549e-01 -4.00515050e-01 -5.75338542e-01
6.62873149e-01 2.06030771e-01 6.22541189e-01 -6.04709685e-01
-8.17051113e-01 -1.12283981e+00 6.81951165e-01 -7.14424610e-01
-4.10839796e-01 -5.82499921e-01 6.01098657e-01 1.22295964e+00
-3.90146226e-01 -1.34715930e-01 6.91714466e-01 7.22272635e-01
5.21476090e-01 -3.56857717e-01 7.41508603e-01 -7.03201443e-02
-2.03059942e-01 -1.17165245e-01 -7.69897103e-01 -3.59050244e-01
5.01436889e-01 -1.39371306e-01 -8.31768095e-01 3.29615287e-02
-5.15871286e-01 3.92932445e-01 6.20745838e-01 6.52441829e-02
-7.28192031e-02 -1.25467288e+00 -8.51344645e-01 -2.55245596e-01
3.52872126e-02 -4.29201156e-01 -1.77099198e-01 6.50147915e-01
-7.15407789e-01 6.68537855e-01 -5.21331489e-01 -5.16359746e-01
-7.97177076e-01 6.08629286e-02 1.97816029e-01 -6.89692572e-02
-1.93641022e-01 6.09595418e-01 4.90719795e-01 -3.83871377e-01
-2.58987725e-01 -2.80283034e-01 3.08717877e-01 2.84212232e-01
4.02219772e-01 8.42472136e-01 -2.43111745e-01 3.10022831e-01
-5.71768403e-01 1.44783959e-01 2.71690726e-01 -9.95411500e-02
1.75267494e+00 -2.26093501e-01 -5.26387453e-01 8.34665120e-01
4.04582262e-01 -7.29085729e-02 -1.23151958e+00 1.63523689e-01
2.50641316e-01 -4.27880764e-01 2.18093872e-01 -9.76344585e-01
-5.85476995e-01 7.88640857e-01 7.49917805e-01 1.95692524e-01
6.86460137e-01 -5.26033461e-01 -1.36234298e-01 8.27127159e-01
2.08670929e-01 -9.61779654e-01 -2.20437676e-01 2.30982468e-01
7.83160269e-01 -1.31660485e+00 7.74910331e-01 -3.88700366e-01
1.06527125e-02 7.87275195e-01 4.41534519e-01 4.45192829e-02
8.09357643e-01 3.51420432e-01 -4.26808983e-01 5.58170006e-02
-9.07522082e-01 -1.09515928e-01 -5.42540908e-01 8.55474949e-01
1.00609469e+00 7.04922676e-01 -7.76614249e-01 1.65323392e-01
-9.77638513e-02 -5.51816374e-02 5.94818115e-01 9.31880891e-01
-1.28064966e+00 -9.23267186e-01 -5.40036321e-01 1.00358558e+00
-5.05215049e-01 -4.05022264e-01 -1.05097197e-01 9.28716838e-01
-1.82378083e-01 9.59890664e-01 -5.19532673e-02 1.15974963e-01
2.84364894e-02 1.99238151e-01 4.08373773e-01 -7.56136596e-01
4.15768288e-03 2.78343648e-01 2.53691763e-01 -1.82322055e-01
-6.90001547e-01 -1.45730436e+00 -1.12217510e+00 -9.22608852e-01
-9.94508684e-01 -1.92215443e-01 1.01227760e+00 5.87568521e-01
1.02510490e-01 7.12511718e-01 5.11862934e-01 -5.90453088e-01
-5.81991434e-01 -1.00668359e+00 -9.60519016e-01 -4.65567023e-01
1.29325643e-01 -1.00972211e+00 -1.71495482e-01 1.77781329e-01] | [9.316872596740723, -1.3645097017288208] |
9280fd5b-191e-46bc-bf1c-1ea728466e29 | masking-by-moving-learning-distraction-free | 1909.03752 | null | https://arxiv.org/abs/1909.03752v4 | https://arxiv.org/pdf/1909.03752v4.pdf | Masking by Moving: Learning Distraction-Free Radar Odometry from Pose Information | This paper presents an end-to-end radar odometry system which delivers robust, real-time pose estimates based on a learned embedding space free of sensing artefacts and distractor objects. The system deploys a fully differentiable, correlation-based radar matching approach. This provides the same level of interpretability as established scan-matching methods and allows for a principled derivation of uncertainty estimates. The system is trained in a (self-)supervised way using only previously obtained pose information as a training signal. Using 280km of urban driving data, we demonstrate that our approach outperforms the previous state-of-the-art in radar odometry by reducing errors by up 68% whilst running an order of magnitude faster. | ['Ingmar Posner', 'Dan Barnes', 'Rob Weston'] | 2019-09-09 | null | null | null | null | ['radar-odometry'] | ['robots'] | [ 1.82014808e-01 3.71318400e-01 1.80682689e-01 -8.39578211e-01
-1.18284059e+00 -5.18352509e-01 8.93670201e-01 -1.00995727e-01
-6.09649241e-01 8.70537758e-01 1.06037296e-02 -2.62870669e-01
-5.75279832e-01 -8.31032515e-01 -6.09824061e-01 -4.94419426e-01
-5.78303337e-01 9.73224401e-01 3.26046973e-01 -4.73558962e-01
3.37148130e-01 7.59297252e-01 -1.61916244e+00 -8.11542213e-01
6.87649548e-01 1.02424598e+00 -1.95233405e-01 8.41388404e-01
5.55356264e-01 3.46685797e-01 -3.86838794e-01 -2.74306417e-01
7.60360420e-01 9.91577581e-02 2.31369603e-02 -2.87671894e-01
1.25083292e+00 -3.25472564e-01 -7.62076855e-01 9.27222013e-01
7.04352140e-01 -2.20268108e-02 6.44513547e-01 -8.53889883e-01
-1.15521550e-01 -3.29756886e-02 7.64266495e-03 4.79898825e-02
3.93823266e-01 2.05206260e-01 7.81125546e-01 -9.14705694e-01
5.18817663e-01 9.34752703e-01 9.37261879e-01 1.58155784e-01
-1.64569998e+00 -7.39843488e-01 -5.42269528e-01 -1.44133285e-01
-1.62215555e+00 -6.99004889e-01 4.46989864e-01 -5.92186630e-01
8.96473348e-01 1.65870450e-02 6.51476026e-01 8.48115027e-01
7.73915052e-01 1.71741918e-02 1.18616354e+00 -2.25038230e-01
2.71109700e-01 6.83310702e-02 1.50207564e-01 6.59443617e-01
7.38397717e-01 1.30253839e+00 -6.64986193e-01 -8.65572169e-02
4.58014697e-01 -2.99773157e-01 1.14069566e-01 -1.06079924e+00
-1.32366180e+00 9.74211574e-01 5.03965139e-01 -2.98940808e-01
-3.48110944e-01 5.39253473e-01 1.14676720e-02 5.48219919e-01
4.52828944e-01 5.34843683e-01 -2.33269036e-01 -2.41486818e-01
-1.21992350e+00 5.24685979e-01 8.48052621e-01 1.10792923e+00
1.10238981e+00 5.49720526e-01 3.82266849e-01 1.06217518e-01
6.79696083e-01 1.51874685e+00 3.72436363e-03 -7.92055488e-01
3.86279702e-01 1.56381726e-01 3.37335408e-01 -1.03615189e+00
-7.57891655e-01 -8.92757833e-01 -2.98228651e-01 9.20457959e-01
2.09975079e-01 -2.89811730e-01 -9.82050657e-01 1.55903411e+00
2.63479769e-01 1.57414302e-01 4.96405482e-01 9.81909990e-01
3.51690531e-01 -2.68651620e-02 -5.85157156e-01 2.61338025e-01
1.13207197e+00 -2.21095771e-01 -5.25760233e-01 -8.41551602e-01
3.33732158e-01 -6.27029002e-01 2.51619279e-01 4.89025652e-01
-3.87358665e-01 -5.77447176e-01 -1.88310957e+00 3.32239300e-01
-4.03482765e-01 -2.51321316e-01 6.49962366e-01 1.01425338e+00
-8.88447881e-01 5.30749261e-01 -5.12397766e-01 -2.52812505e-01
1.50178308e-02 2.74117142e-01 -4.67441678e-01 -1.71681285e-01
-1.40740037e+00 1.62580907e+00 3.63577724e-01 2.62582719e-01
-1.05227256e+00 -9.09319341e-01 -1.33666742e+00 -7.55345464e-01
3.08249891e-02 -5.48900366e-01 1.00414526e+00 4.50138375e-02
-1.62883759e+00 7.22912073e-01 7.44557083e-02 -9.27587509e-01
6.74732804e-01 -6.81809723e-01 -8.91345561e-01 -6.24875240e-02
1.32822692e-01 5.36220729e-01 1.00986707e+00 -1.01967371e+00
-4.19784695e-01 -6.19538546e-01 -4.32926536e-01 6.67668581e-02
8.90700638e-01 -7.34421492e-01 2.96651065e-01 -1.53665215e-01
7.57841110e-01 -1.16907573e+00 -3.07999432e-01 -3.37352008e-02
2.02992886e-01 6.48736954e-01 4.73845810e-01 -3.81522030e-01
6.99975431e-01 -1.83640015e+00 -1.48114622e-01 4.55308080e-01
2.61253506e-01 -1.47421107e-01 2.11947665e-01 6.11239970e-01
2.93870807e-01 -7.73208499e-01 -4.02778596e-01 -1.05754219e-01
2.85537452e-01 2.83325195e-01 -5.18786371e-01 1.05153930e+00
2.53328890e-01 9.39263582e-01 -1.01883209e+00 -7.67451674e-02
6.55525088e-01 4.85120773e-01 -1.16426192e-01 7.13883787e-02
1.92550927e-01 4.02648121e-01 -6.91305622e-02 5.79200983e-01
9.40316439e-01 7.30814576e-01 -1.89900547e-01 -1.75881982e-01
-2.50869274e-01 4.41699475e-01 -1.37824166e+00 2.02690673e+00
-5.63641429e-01 1.23539412e+00 -1.61421910e-01 -5.72507977e-01
1.75197089e+00 -2.94896334e-01 2.08315030e-01 -8.38775873e-01
6.95312321e-02 5.64907193e-01 2.63324846e-02 -1.67290971e-01
7.33980715e-01 -5.82531095e-01 -5.85118294e-01 5.05184382e-02
1.66857794e-01 -8.12853813e-01 -4.05821443e-01 1.87454969e-02
1.29700518e+00 3.64166200e-01 4.27488238e-01 -6.12620473e-01
4.38066065e-01 2.01421365e-01 2.15129822e-01 6.13853157e-01
-2.56339490e-01 3.25931966e-01 -5.77575445e-01 -4.51038212e-01
-8.81886482e-01 -1.84463882e+00 -7.36917078e-01 3.45739722e-01
4.24552917e-01 -2.26383116e-02 5.92073463e-02 -1.24911457e-01
8.95299911e-01 7.23615825e-01 -6.78401709e-01 -3.15443039e-01
-3.35864753e-01 -3.79913777e-01 9.82101023e-01 4.44081843e-01
3.34064215e-01 -2.75907874e-01 -8.97946239e-01 4.38362807e-01
4.18216079e-01 -1.13373160e+00 7.22144023e-02 3.75639409e-01
-1.08895302e+00 -8.22510064e-01 8.13257992e-02 -1.50220796e-01
3.35983545e-01 2.68610269e-01 1.09943509e+00 -7.27619052e-01
-4.39504743e-01 4.24765646e-01 4.55973260e-02 -6.50409341e-01
3.33635062e-02 -3.54456812e-01 5.50982177e-01 -1.57599658e-01
6.91099167e-01 -7.99165487e-01 -2.76564866e-01 1.77787840e-01
-2.46218413e-01 -5.34663975e-01 1.06682539e+00 6.70936286e-01
2.10356772e-01 -7.18385994e-01 3.73347908e-01 -5.37125468e-01
4.12942529e-01 -2.36486122e-01 -1.18284905e+00 -5.92721403e-01
-1.15441811e+00 7.58460879e-01 -2.53524557e-02 1.90258473e-02
-7.22210169e-01 4.08661395e-01 1.36861637e-01 -1.37398243e-01
-2.36584187e-01 4.87096906e-01 9.00834799e-02 -7.89509475e-01
1.20670187e+00 2.60030895e-01 4.70589221e-01 -1.81913510e-01
7.20637023e-01 6.24722838e-01 1.09280360e+00 -2.53306121e-01
1.67463815e+00 7.25614190e-01 5.89096308e-01 -9.22413528e-01
-5.09923518e-01 -7.36849904e-01 -1.05459845e+00 -3.25905204e-01
3.50509048e-01 -1.33661330e+00 -5.73952377e-01 -7.74624124e-02
-5.61829150e-01 -6.20472757e-03 -3.98420781e-01 9.89692807e-01
-7.73013234e-01 8.26843157e-02 2.71970958e-01 -1.12737679e+00
-1.49420679e-01 -4.69762534e-01 1.37429690e+00 -5.87075576e-03
-1.80314437e-01 -9.09536898e-01 9.17817354e-01 1.82698160e-01
5.10089934e-01 7.17552185e-01 -3.00622195e-01 -3.08482051e-01
-9.61617589e-01 -7.32511342e-01 -3.65400985e-02 -7.88391009e-02
-3.79366279e-01 -5.50461471e-01 -1.07057714e+00 -4.65512812e-01
-2.70373285e-01 -1.38038874e-01 1.03540909e+00 1.28490925e-01
-6.05167925e-01 2.01431215e-01 -4.10866946e-01 7.64875948e-01
1.73304498e+00 -3.07304502e-01 8.42485905e-01 5.89938998e-01
2.84471810e-01 5.73205948e-01 1.11192727e+00 1.07990749e-01
4.20709759e-01 7.05970526e-01 7.03425944e-01 3.06542665e-01
2.45479286e-01 -3.98718089e-01 4.59270120e-01 1.87345445e-01
2.36688405e-02 5.80528855e-01 -7.98241079e-01 6.72732651e-01
-1.72599757e+00 -1.03033972e+00 -1.51289806e-01 2.55204773e+00
2.25992501e-01 5.44967115e-01 -2.28191853e-01 1.41813785e-01
1.64576709e-01 3.42313468e-01 -2.66509891e-01 -3.16818833e-01
2.80552864e-01 3.67737383e-01 1.56134248e+00 1.06239033e+00
-8.81886899e-01 7.88099647e-01 6.96970224e+00 6.51765391e-02
-9.35672998e-01 -9.83688235e-02 -8.17184269e-01 -2.66972352e-02
-2.03535527e-01 3.42931390e-01 -9.74510610e-01 -3.18878368e-02
1.40976715e+00 -2.05356209e-03 1.85928643e-01 8.82263899e-01
-5.03208563e-02 -4.00214791e-01 -9.41393137e-01 8.77595961e-01
3.26094717e-01 -1.25884902e+00 -7.14675486e-01 4.16063666e-01
2.14564398e-01 7.36180604e-01 -2.46242940e-01 4.96654719e-01
4.56112742e-01 -8.95583570e-01 9.87729371e-01 7.88951337e-01
8.52019548e-01 -8.31628025e-01 8.85941446e-01 4.28343624e-01
-1.27231026e+00 2.92302459e-01 -4.44229871e-01 -5.42846739e-01
3.82673770e-01 9.86422002e-01 -1.07526481e+00 9.83961403e-01
2.35268429e-01 6.03447199e-01 -5.42321682e-01 1.02887082e+00
-4.85673636e-01 9.01833922e-02 -5.68833470e-01 3.82747836e-02
2.41467655e-01 -2.51479268e-01 8.94163251e-01 1.20703220e+00
3.57582450e-01 -1.86645627e-01 1.32766783e-01 6.67665780e-01
5.80066562e-01 -5.77670217e-01 -1.25586879e+00 4.21535552e-01
5.96921027e-01 1.25059688e+00 5.70289828e-02 1.79819372e-02
3.09724417e-02 3.89437318e-01 1.87790040e-02 -1.11241542e-01
-7.40908086e-01 -7.90381670e-01 1.16748834e+00 9.46765468e-02
4.13511753e-01 -9.57202375e-01 -1.07971363e-01 -8.16693068e-01
8.47324431e-02 -3.03291261e-01 -7.38342851e-02 -6.19861603e-01
-1.07364285e+00 5.60962677e-01 2.08194852e-01 -1.72416985e+00
-7.85466671e-01 -7.38531411e-01 -1.59766972e-01 9.63313818e-01
-1.79761481e+00 -1.04119897e+00 -6.31718099e-01 9.86614171e-03
-1.68448448e-01 -3.91199082e-01 7.90530920e-01 8.29021931e-02
4.45641726e-01 1.99565351e-01 1.06067568e-01 7.93365669e-03
8.69424403e-01 -1.36722887e+00 5.95589280e-01 1.04945350e+00
4.43219602e-01 4.20922607e-01 1.46673071e+00 -7.40819931e-01
-1.70824301e+00 -1.00632989e+00 1.01776123e+00 -8.72225046e-01
8.46910715e-01 -5.82933903e-01 -2.82564878e-01 7.34271049e-01
-1.50602296e-01 2.30540335e-01 2.06689224e-01 2.19122604e-01
-5.25386512e-01 -5.06252527e-01 -1.25150430e+00 1.95019752e-01
1.14245450e+00 -6.54999733e-01 -1.24174452e+00 -7.81328306e-02
5.01468122e-01 -8.68951380e-01 -8.84409606e-01 6.20399892e-01
9.67830658e-01 -9.72098112e-01 1.09988737e+00 -6.91367984e-02
-4.46572304e-01 -6.59515679e-01 -5.40238976e-01 -1.36045313e+00
-3.67757261e-01 -8.54436755e-01 -1.60387203e-01 5.81604600e-01
2.84390718e-01 -1.16728914e+00 8.15577269e-01 -5.70297055e-02
-2.76578248e-01 -2.36684978e-01 -1.26002061e+00 -1.37456322e+00
-2.65763968e-01 -9.60140467e-01 2.57335693e-01 4.04071540e-01
-8.03371817e-02 5.36647558e-01 -4.51396257e-01 7.86553085e-01
1.41424310e+00 1.38252795e-01 1.23727775e+00 -1.84844339e+00
-1.26935914e-01 3.91832851e-02 -1.39500701e+00 -1.14643812e+00
1.88178699e-02 -7.39318132e-01 5.70613861e-01 -1.06875491e+00
-8.25703382e-01 -3.47928733e-01 1.72831744e-01 -2.08130658e-01
4.42432523e-01 5.36879182e-01 -7.42069036e-02 6.35498837e-02
-2.15448305e-01 5.10746419e-01 4.51043129e-01 -9.06363353e-02
2.17189938e-02 1.13502532e-01 -3.54835361e-01 3.04520309e-01
5.05116463e-01 -6.95990443e-01 -3.40358853e-01 -8.02905411e-02
3.20921540e-01 -1.19578533e-01 7.48994946e-01 -1.61437750e+00
3.93881857e-01 -5.92980115e-03 3.82366002e-01 -1.02892447e+00
7.22245216e-01 -9.61352646e-01 2.82775670e-01 7.85656810e-01
1.71787083e-01 -3.59126627e-02 4.25152540e-01 9.68994856e-01
-1.43303975e-01 -7.33086020e-02 6.82541132e-01 2.99735427e-01
-1.07505143e+00 6.56642094e-02 -2.67954499e-01 1.22001991e-01
9.77790117e-01 -6.40513480e-01 -3.84694695e-01 -5.47606766e-01
-3.37855488e-01 1.87806770e-01 3.61819148e-01 3.34479094e-01
7.45315790e-01 -1.43823600e+00 -9.49995339e-01 6.55031860e-01
8.14990163e-01 -2.67026216e-01 -4.35564369e-02 8.00767660e-01
-5.21807313e-01 8.54693174e-01 -3.91719788e-01 -9.16069865e-01
-7.72031724e-01 1.47269651e-01 4.84101623e-01 1.18302673e-01
-8.61706853e-01 4.35988069e-01 -6.73066974e-01 -9.21924472e-01
-2.25087106e-01 -2.66229123e-01 2.98930377e-01 -1.04015775e-01
4.37695563e-01 1.60196617e-01 3.31771165e-01 -7.78304338e-01
-8.21770966e-01 8.68690073e-01 3.59522760e-01 -7.70962656e-01
1.17356253e+00 -2.67611295e-01 6.51025891e-01 5.99169493e-01
9.38626409e-01 3.91039580e-01 -1.51619613e+00 -3.97202432e-01
2.33325988e-01 -8.23224545e-01 5.01213193e-01 -6.17588043e-01
-3.25942695e-01 7.01449931e-01 9.08612967e-01 -1.84257478e-01
4.41491902e-01 -6.45639375e-02 3.43074918e-01 8.56827080e-01
9.17519033e-01 -9.88156855e-01 -2.04547077e-01 8.47871780e-01
7.84760475e-01 -1.46352994e+00 5.20965993e-01 -5.96794933e-02
-4.45039541e-01 1.04176092e+00 -5.80507591e-02 -8.17480087e-01
7.65587866e-01 5.31016350e-01 4.66578692e-01 -5.15604556e-01
-4.01962161e-01 -7.91011572e-01 5.57361424e-01 1.27363348e+00
1.62867848e-02 1.27878323e-01 -2.52535776e-03 -6.77098259e-02
-6.77788615e-01 -4.30806518e-01 4.27276343e-01 1.03807676e+00
-1.00531828e+00 -7.38960922e-01 -5.01082838e-01 2.13042438e-01
4.71773028e-01 -5.29455021e-02 -1.67184055e-01 1.31964743e+00
-2.18568355e-01 6.38488710e-01 2.26282850e-02 -8.34679604e-01
7.30608821e-01 1.80811182e-01 8.35678399e-01 -3.80953163e-01
-4.47942950e-02 -4.41902488e-01 4.48947430e-01 -8.42501581e-01
-2.36168385e-01 -8.50410163e-01 -1.06030285e+00 -2.75201738e-01
-2.64855057e-01 -8.59262869e-02 9.65750933e-01 1.00283980e+00
5.11017680e-01 4.07473415e-01 5.82930148e-01 -1.05452394e+00
-1.05279863e+00 -9.46468651e-01 -6.42350972e-01 -2.58924007e-01
6.44803464e-01 -1.21412230e+00 -4.62532789e-01 -3.57922494e-01] | [7.3736653327941895, -2.116434335708618] |
7f05e6ab-cd0c-42e0-979c-cfb8e6370894 | 3d-u-net-learning-dense-volumetric | 1606.06650 | null | http://arxiv.org/abs/1606.06650v1 | http://arxiv.org/pdf/1606.06650v1.pdf | 3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation | This paper introduces a network for volumetric segmentation that learns from
sparsely annotated volumetric images. We outline two attractive use cases of
this method: (1) In a semi-automated setup, the user annotates some slices in
the volume to be segmented. The network learns from these sparse annotations
and provides a dense 3D segmentation. (2) In a fully-automated setup, we assume
that a representative, sparsely annotated training set exists. Trained on this
data set, the network densely segments new volumetric images. The proposed
network extends the previous u-net architecture from Ronneberger et al. by
replacing all 2D operations with their 3D counterparts. The implementation
performs on-the-fly elastic deformations for efficient data augmentation during
training. It is trained end-to-end from scratch, i.e., no pre-trained network
is required. We test the performance of the proposed method on a complex,
highly variable 3D structure, the Xenopus kidney, and achieve good results for
both use cases. | ['Soeren S. Lienkamp', 'Özgün Çiçek', 'Thomas Brox', 'Olaf Ronneberger', 'Ahmed Abdulkadir'] | 2016-06-21 | null | null | null | null | ['3d-instance-segmentation-1'] | ['computer-vision'] | [ 2.40831092e-01 7.73052335e-01 -2.44437128e-01 -6.73583984e-01
-5.72987795e-01 -4.80476648e-01 2.52767682e-01 3.13026637e-01
-7.63972044e-01 6.99856043e-01 -2.88046539e-01 -5.51010035e-02
4.63512301e-01 -7.43928134e-01 -1.05241179e+00 -4.27052855e-01
-3.07239175e-01 1.05311596e+00 3.23946536e-01 1.79548293e-01
9.13355127e-02 7.24843264e-01 -9.95611548e-01 -1.32039145e-01
6.44603729e-01 1.06209600e+00 2.58334458e-01 6.18400812e-01
-3.36263210e-01 3.89386594e-01 -2.21793458e-01 -8.34474787e-02
5.26679754e-01 -2.21295834e-01 -1.10031998e+00 5.39952695e-01
5.12639821e-01 -4.85587597e-01 -8.28773528e-02 7.53810108e-01
3.26454967e-01 1.44027616e-03 4.72029299e-01 -7.27734149e-01
-2.54484624e-01 6.78516448e-01 -6.87555850e-01 5.86820655e-02
-4.83461190e-03 1.21357888e-01 7.47386038e-01 -1.06880653e+00
9.50371265e-01 8.99151921e-01 6.49183035e-01 5.36877930e-01
-1.49577105e+00 -3.15732121e-01 7.24054947e-02 -5.59952915e-01
-1.28424335e+00 -1.50563270e-01 8.40330243e-01 -6.23686790e-01
6.99725628e-01 -3.38868015e-02 1.08804858e+00 6.62455916e-01
-1.24739803e-01 1.04509366e+00 9.50941086e-01 -2.61125952e-01
4.82158393e-01 8.88341144e-02 1.84067324e-01 8.68464351e-01
9.40373242e-02 -1.78424075e-01 -4.26001847e-02 -1.19996846e-01
1.18989515e+00 5.66392653e-02 -2.50502527e-01 -1.11590767e+00
-1.15352094e+00 6.48403466e-01 6.65804982e-01 3.54535133e-01
-4.30186838e-01 1.73754245e-01 5.29367864e-01 -3.14224279e-03
6.21073008e-01 2.28703544e-01 -4.52963859e-01 1.26577124e-01
-1.37381029e+00 2.45393559e-01 7.45468497e-01 9.41140950e-01
1.07928395e+00 5.57043473e-04 -9.61632654e-03 6.67327881e-01
2.91604549e-01 1.82138771e-01 4.07588124e-01 -8.33971024e-01
3.41430634e-01 5.44751704e-01 -1.37234656e-02 -4.16256726e-01
-6.55066967e-01 -4.35652733e-01 -8.62104774e-01 2.74241179e-01
5.23320496e-01 -2.17473999e-01 -1.58418941e+00 1.51066601e+00
6.18800044e-01 2.02001914e-01 -1.09701261e-01 1.04295421e+00
8.46293211e-01 3.16276848e-01 -2.86047291e-02 -7.29189292e-02
9.59719598e-01 -9.80354488e-01 -3.91050279e-01 -1.73999339e-01
6.44433737e-01 -4.35779244e-01 1.04493797e+00 2.71957606e-01
-1.32462823e+00 -3.66187125e-01 -9.70933437e-01 -2.11624995e-01
-1.60046980e-01 6.46656603e-02 6.50928020e-01 5.10217965e-01
-1.09395385e+00 8.36670339e-01 -1.26831245e+00 -3.67361754e-01
8.05320621e-01 6.12907767e-01 -4.28293526e-01 2.23373808e-02
-6.84059083e-01 5.84200144e-01 6.57738388e-01 1.76698714e-01
-1.20071912e+00 -7.12665021e-01 -1.06118762e+00 1.82852987e-02
4.30646449e-01 -6.54727221e-01 1.23489380e+00 -8.25774729e-01
-1.55325747e+00 1.21098101e+00 4.62865978e-02 -4.99965400e-01
8.01350296e-01 -2.40106866e-01 7.09176064e-01 3.33331853e-01
-2.73910239e-02 9.58671808e-01 4.73468512e-01 -1.55718195e+00
-1.91047594e-01 -4.68404442e-01 1.10222995e-01 -2.17483584e-02
-8.76387283e-02 -4.11980271e-01 -7.36934185e-01 -5.66932440e-01
2.67360896e-01 -1.01085281e+00 -5.56609094e-01 2.74248809e-01
-5.79729795e-01 1.20325528e-01 1.01147342e+00 -5.36666572e-01
3.97151142e-01 -1.80151248e+00 2.10026518e-01 6.07354939e-01
4.59837615e-01 2.19125405e-01 1.33847799e-02 4.82810140e-02
1.41354278e-02 7.39300251e-02 -8.95323694e-01 -7.92808652e-01
-2.93720603e-01 4.15485591e-01 2.51782298e-01 6.19305134e-01
2.43791461e-01 9.95834708e-01 -8.55071723e-01 -7.02396691e-01
2.53709257e-01 4.81103778e-01 -6.92518413e-01 4.73868012e-01
-4.36534554e-01 9.30772781e-01 -4.76086974e-01 6.39175177e-01
7.68424869e-01 -3.85532200e-01 1.30584359e-01 -1.26474395e-01
-1.18731670e-01 -1.83535114e-01 -1.05830848e+00 2.33093953e+00
-5.03144145e-01 3.27502280e-01 2.52632499e-01 -1.18057251e+00
8.51139784e-01 2.60420382e-01 7.35424519e-01 -1.85407951e-01
3.87823105e-01 3.00427347e-01 -2.95246392e-01 -3.29280585e-01
1.79353729e-01 -1.51449531e-01 3.58101130e-02 5.64073682e-01
5.94991386e-01 -4.53700662e-01 3.86984766e-01 1.56140253e-01
9.57572222e-01 5.14856756e-01 4.27816778e-01 -4.16520834e-01
4.54321921e-01 5.29615059e-02 6.44162416e-01 4.20703948e-01
-3.16947363e-02 7.72840083e-01 5.41052043e-01 -6.90772951e-01
-1.20839250e+00 -9.64967370e-01 -2.69759774e-01 6.70688808e-01
1.06693484e-01 -1.72123685e-01 -9.55722809e-01 -7.70070910e-01
-1.14919707e-01 1.75071239e-01 -8.05461824e-01 4.19927984e-01
-6.56545520e-01 -4.61017102e-01 3.10405821e-01 6.96885943e-01
4.37574357e-01 -1.03517520e+00 -9.02210653e-01 1.83926240e-01
1.57672152e-01 -1.14710855e+00 -2.94247806e-01 4.26184118e-01
-1.27437091e+00 -1.03926396e+00 -1.03496909e+00 -1.00474572e+00
1.08649194e+00 -1.20756976e-01 1.20168638e+00 2.75459766e-01
-3.35047126e-01 2.61435658e-01 -5.40946750e-03 -1.74125582e-01
-1.91089764e-01 1.42188892e-01 -2.83380657e-01 -2.43356243e-01
-1.10610053e-01 -6.70782030e-01 -5.05419016e-01 8.79002437e-02
-9.19677854e-01 2.76381895e-02 5.93630135e-01 8.54593694e-01
1.40335989e+00 -3.31368357e-01 2.82048941e-01 -1.44445264e+00
2.31709883e-01 -1.68371320e-01 -8.45714748e-01 -2.08174642e-02
-2.27087215e-01 1.35646492e-01 5.76425254e-01 -3.12363774e-01
-7.99237311e-01 8.10454726e-01 -3.70942205e-01 -6.93643093e-01
-4.54600900e-01 5.68045557e-01 -1.33219874e-02 -4.57238853e-01
4.56708997e-01 -8.35285634e-02 8.23348910e-02 -6.00653470e-01
3.97429079e-01 3.74021500e-01 6.36455119e-01 -6.00203931e-01
1.00818121e+00 6.29551351e-01 2.09786594e-02 -7.73528039e-01
-6.44274771e-01 -5.16208470e-01 -1.25725782e+00 -1.12572730e-01
9.41065192e-01 -8.13146949e-01 -5.33018589e-01 2.64196694e-01
-1.10866868e+00 -8.87412548e-01 -5.29401362e-01 4.62124377e-01
-8.25446010e-01 3.66751313e-01 -7.57159710e-01 -5.39567053e-01
-4.78900194e-01 -1.32020319e+00 1.16899836e+00 2.03266382e-01
-1.47783309e-01 -1.06937540e+00 2.14391679e-01 1.60017490e-01
2.04071283e-01 7.87683189e-01 6.89703643e-01 -7.75234342e-01
-6.61807477e-01 -2.52574056e-01 -1.40239015e-01 3.34845603e-01
-5.70545122e-02 -1.24681167e-01 -1.03192747e+00 -3.34424049e-01
-1.53147459e-01 -7.12751091e-01 8.11440349e-01 4.21419978e-01
1.51930094e+00 -8.36354867e-02 -2.41960555e-01 7.54660547e-01
1.48225808e+00 -1.04854219e-01 2.74608433e-01 -6.56979252e-03
9.06334579e-01 5.66371024e-01 3.23014706e-01 3.07731777e-01
3.11337590e-01 2.83825040e-01 6.23709679e-01 -6.49979293e-01
1.30157068e-01 -2.23106109e-02 -2.48757616e-01 7.16417551e-01
-2.11374789e-01 5.34635372e-02 -1.00777555e+00 6.87609136e-01
-1.69713318e+00 -4.90092397e-01 2.76942134e-01 2.20306158e+00
9.63997424e-01 2.73538142e-01 2.24344388e-01 -7.66236708e-03
4.92068857e-01 5.21487445e-02 -6.98511660e-01 -1.24414302e-01
3.28544945e-01 4.17974323e-01 5.47931492e-01 6.35228097e-01
-1.16113806e+00 1.01875019e+00 5.64147282e+00 3.27574342e-01
-1.30224347e+00 7.37137571e-02 9.31984365e-01 -1.69952080e-01
-2.23276317e-01 -2.14391768e-01 -4.05433446e-01 1.06740251e-01
5.94714701e-01 1.32321954e-01 2.83995271e-02 1.06628799e+00
1.68448880e-01 -2.16731623e-01 -1.36211133e+00 7.58786261e-01
7.68899173e-03 -1.48592460e+00 -2.54400373e-01 2.15755422e-02
6.80936337e-01 1.28576145e-01 -3.34684581e-01 2.38998875e-01
2.35080481e-01 -9.08666193e-01 6.64773166e-01 3.26337188e-01
8.79721582e-01 -8.33179235e-01 8.12138796e-01 5.01604557e-01
-1.15982211e+00 4.21967030e-01 -2.80278325e-01 3.36548477e-01
4.32742834e-01 7.66403139e-01 -9.99632776e-01 2.51445293e-01
5.47561526e-01 4.52529103e-01 -3.05302680e-01 1.20938921e+00
-1.97482675e-01 4.76170421e-01 -5.79746544e-01 3.09142798e-01
3.98401767e-01 -2.49097735e-01 3.31803560e-01 1.27210486e+00
8.21036622e-02 9.85741615e-02 6.38747454e-01 9.69286263e-01
-4.76372331e-01 1.60673648e-01 -5.87205589e-01 3.45890410e-02
2.46707164e-02 1.43802559e+00 -1.26904118e+00 -4.25817728e-01
-2.17921257e-01 8.75799060e-01 5.11203825e-01 2.47830421e-01
-5.92768073e-01 -1.94517002e-01 9.40867439e-02 2.45934382e-01
4.07209486e-01 -2.05530599e-01 -3.22179168e-01 -9.61865544e-01
-5.34799136e-03 -2.78199285e-01 1.75770029e-01 -4.85090345e-01
-1.00089562e+00 5.82884908e-01 4.07030620e-02 -8.55949938e-01
-2.88672388e-01 -3.70535135e-01 -5.74434102e-01 7.72380054e-01
-1.53434432e+00 -1.10671401e+00 -5.33801317e-01 3.62587541e-01
4.26819444e-01 2.34066248e-01 7.53730297e-01 2.50436723e-01
-3.78154993e-01 3.32600594e-01 -3.47783893e-01 4.41007882e-01
2.85651565e-01 -1.52625370e+00 4.37554806e-01 7.74623394e-01
6.73229024e-02 4.91145968e-01 4.82994437e-01 -6.89893782e-01
-1.11800385e+00 -1.31183159e+00 6.03335261e-01 1.78578515e-02
3.04269969e-01 -6.75172031e-01 -1.03875470e+00 8.73529077e-01
8.03470090e-02 5.58357179e-01 6.38854980e-01 -8.05777833e-02
4.31409851e-03 2.15098158e-01 -1.54528129e+00 2.64250785e-01
1.00535011e+00 -2.69518584e-01 -4.79590744e-01 6.33903980e-01
7.18060374e-01 -9.86570776e-01 -1.17117369e+00 4.07952577e-01
2.52549052e-01 -7.50741482e-01 8.32954347e-01 -3.40386987e-01
4.66552734e-01 -2.79608428e-01 2.34516159e-01 -1.20433176e+00
7.96707422e-02 -5.68687141e-01 -1.69334307e-01 7.50499010e-01
3.65583926e-01 -2.07381725e-01 1.24325943e+00 6.15924060e-01
-3.52739573e-01 -1.21428859e+00 -8.71524394e-01 -5.33313751e-01
7.68430978e-02 -2.34764695e-01 4.09854352e-01 7.88970649e-01
-2.14643076e-01 2.81189442e-01 -1.64366454e-01 -5.48195168e-02
8.94877017e-01 3.04025084e-01 8.89352918e-01 -1.33447945e+00
-1.97525606e-01 -2.79051811e-02 -4.74157572e-01 -1.29306507e+00
1.31708637e-01 -9.93187308e-01 3.27771097e-01 -1.63299906e+00
2.16989636e-01 -5.98539650e-01 1.95665732e-02 6.53722405e-01
1.11629821e-01 4.47385937e-01 6.79447204e-02 1.15575053e-01
-4.63929117e-01 4.76791024e-01 1.51065445e+00 -9.18770134e-02
-4.83436853e-01 6.52578324e-02 -2.05886468e-01 8.80225301e-01
8.38118196e-01 -3.15102249e-01 -4.28356856e-01 -3.78163546e-01
-2.86175966e-01 2.87602097e-01 2.77205139e-01 -8.85914862e-01
-1.86948553e-02 3.62444259e-02 5.52722633e-01 -8.23958039e-01
2.28795350e-01 -9.79246259e-01 -2.17489704e-01 4.02765483e-01
-2.65170008e-01 -2.74837166e-01 1.69165596e-01 4.64698017e-01
-6.11053854e-02 -3.40095878e-01 1.08381283e+00 -5.42627335e-01
-3.80875528e-01 7.70935476e-01 -3.73490155e-02 1.26925141e-01
1.17922270e+00 -3.40296268e-01 3.08201522e-01 -5.96954823e-02
-1.17542052e+00 2.42223859e-01 6.40016377e-01 -2.21108958e-01
5.93116999e-01 -1.09219253e+00 -3.33021283e-01 2.01458976e-01
-7.06274435e-02 1.00939643e+00 1.21534042e-01 7.40235209e-01
-8.54216218e-01 1.95652172e-01 -2.81493664e-01 -9.69552398e-01
-1.04892075e+00 4.62836713e-01 3.98855329e-01 -4.43279505e-01
-9.35743749e-01 9.10726786e-01 2.18598083e-01 -6.45064712e-01
3.36678326e-01 -7.18007922e-01 -1.34855062e-01 -9.45437253e-02
3.00736099e-01 -7.27557093e-02 1.80845901e-01 -5.43810725e-01
-1.97913051e-01 5.02780795e-01 -7.20735937e-02 -8.52115080e-02
1.73860443e+00 8.41433629e-02 -9.73873213e-02 5.13249457e-01
1.25185931e+00 -1.70723349e-01 -1.50976861e+00 -3.48181903e-01
-1.35891307e-02 -3.88610899e-01 2.32713930e-02 -3.99563372e-01
-1.56267071e+00 8.92298579e-01 4.63873446e-01 -7.73077533e-02
8.35914731e-01 -5.27482573e-03 9.49754953e-01 4.71744448e-01
4.47821528e-01 -9.68860984e-01 4.30062450e-02 4.38283831e-01
6.89890921e-01 -1.25747848e+00 1.77618891e-01 -4.49668378e-01
-3.92556548e-01 1.12772858e+00 6.77578688e-01 -3.90491545e-01
7.45223939e-01 3.83675396e-01 1.43381968e-01 -3.53121638e-01
-2.96744257e-01 -9.36350971e-02 6.90135732e-02 5.36102414e-01
5.01113653e-01 -2.03849539e-01 -1.75443023e-01 1.89055353e-01
-1.56804934e-01 2.34095037e-01 6.42982602e-01 1.16164064e+00
-3.98597836e-01 -8.88760328e-01 -2.25239053e-01 4.96895939e-01
-3.00645471e-01 1.09533168e-01 -2.93300509e-01 9.05905843e-01
9.48450416e-02 1.94454357e-01 1.86438590e-01 9.74287987e-02
3.26919585e-01 3.80334817e-02 6.01048827e-01 -9.85351801e-01
-6.28150165e-01 3.49926025e-01 -2.34455794e-01 -7.01877654e-01
-5.75396240e-01 -5.20057321e-01 -1.73742735e+00 7.40800574e-02
-2.14584529e-01 3.47280763e-02 7.00580657e-01 9.86864984e-01
-9.81115364e-03 5.85852325e-01 3.85681659e-01 -1.47995079e+00
-1.81271583e-01 -8.39586198e-01 -4.93917763e-01 4.03627187e-01
3.35277200e-01 -7.34603167e-01 -7.50610558e-03 3.38294506e-01] | [14.627257347106934, -2.2278640270233154] |
8462d3ac-5880-4e77-9282-0e55d6b0df6d | a-biologically-inspired-evaluation-of | 2208.09658 | null | https://arxiv.org/abs/2208.09658v2 | https://arxiv.org/pdf/2208.09658v2.pdf | A biologically-inspired multi-modal evaluation of molecular generative machine learning | While generative models have recently become ubiquitous in many scientific areas, less attention has been paid to their evaluation. For molecular generative models, the state-of-the-art examines their output in isolation or in relation to its input. However, their biological and functional properties, such as ligand-target interaction is not being addressed. In this study, a novel biologically-inspired benchmark for the evaluation of molecular generative models is proposed. Specifically, three diverse reference datasets are designed and a set of metrics are introduced which are directly relevant to the drug discovery process. In particular we propose a recreation metric, apply drug-target affinity prediction and molecular docking as complementary techniques for the evaluation of generative outputs. While all three metrics show consistent results across the tested generative models, a more detailed comparison of drug-target affinity binding and molecular docking scores revealed that unimodal predictiors can lead to erroneous conclusions about target binding on a molecular level and a multi-modal approach is thus preferrable. The key advantage of this framework is that it incorporates prior physico-chemical domain knowledge into the benchmarking process by focusing explicitly on ligand-target interactions and thus creating a highly efficient tool not only for evaluating molecular generative outputs in particular, but also for enriching the drug discovery process in general. | ['Siamac Fazli', 'Vsevolod Peshkov', 'Ferdinand Molnár', 'Rustam Zhumagambetov', 'Karina Pats', 'Abylay Salimzhanov', 'Anuar Suleimenov', 'Albina Li', 'Maxim Mametkulov', 'Mukhamejan Karatayev', 'Alisher Amanatay', 'Abay Artykbayev', 'Elizaveta Vinogradova'] | 2022-08-20 | null | null | null | null | ['molecular-docking'] | ['medical'] | [ 2.63992310e-01 -3.39692123e-02 -1.92500725e-01 -4.80816737e-02
-7.43978679e-01 -7.64047921e-01 8.82953703e-01 5.07464111e-01
-3.27227116e-01 1.31698334e+00 -4.32881750e-02 -3.28789741e-01
-3.96562606e-01 -8.60849738e-01 -6.96271241e-01 -1.16897774e+00
7.02598989e-02 7.64804900e-01 9.92644802e-02 -2.30324239e-01
1.59659684e-01 8.57120633e-01 -1.61104131e+00 1.35136530e-01
1.11354601e+00 4.98233765e-01 1.35357067e-01 2.41925761e-01
9.37175378e-02 2.84635365e-01 -7.44265795e-01 -5.69068611e-01
-1.70655042e-01 -7.99019158e-01 -5.45033753e-01 -4.02904630e-01
1.10725567e-01 4.44062650e-01 9.90443379e-02 8.09618175e-01
7.92851806e-01 1.04986317e-01 1.02166677e+00 -6.95076525e-01
-4.48810667e-01 3.51944506e-01 1.69537261e-01 -1.41617030e-01
5.99294186e-01 3.39930356e-01 9.41191912e-01 -8.73886526e-01
7.71406710e-01 1.01604295e+00 3.90127420e-01 3.95259649e-01
-1.72028995e+00 -4.32932585e-01 -3.18759121e-02 1.95007727e-01
-1.39850163e+00 -1.41775861e-01 6.03954017e-01 -7.51238704e-01
8.29091072e-01 5.27843416e-01 6.02113247e-01 1.33755791e+00
4.85354692e-01 3.91879022e-01 1.23314297e+00 -2.50498354e-01
6.40277088e-01 2.17444673e-01 -5.72991893e-02 2.70836860e-01
4.09327477e-01 2.39826888e-01 -4.87384945e-01 -3.43074292e-01
3.98947448e-01 -2.57471688e-02 -2.58591175e-01 -4.81171638e-01
-9.40189481e-01 8.48263621e-01 3.82123142e-01 6.22456610e-01
-6.03038073e-01 -3.90455872e-02 2.39316329e-01 -1.30601510e-01
3.48592132e-01 6.82581902e-01 -3.47605079e-01 -4.25809659e-02
-7.35240281e-01 6.47544622e-01 8.90501559e-01 2.59168088e-01
7.91542828e-01 -1.14284366e-01 -2.42072865e-01 4.77589875e-01
2.42512569e-01 3.11738223e-01 3.27918619e-01 -2.44004026e-01
-1.39631145e-02 6.39455080e-01 1.83152705e-02 -8.85942519e-01
-4.85331178e-01 -7.09800780e-01 -7.35018015e-01 2.78342962e-01
3.17429245e-01 1.36884749e-01 -8.37541819e-01 1.82901549e+00
4.96698409e-01 1.53122813e-01 2.23540038e-01 8.44501317e-01
1.06642056e+00 4.53755111e-01 6.93195164e-01 -3.88196945e-01
1.22935283e+00 -3.50827605e-01 -4.39324647e-01 2.20987022e-01
6.28684163e-01 -8.18322182e-01 7.48351812e-01 4.94428515e-01
-9.23139393e-01 -3.52173954e-01 -8.91931534e-01 2.27375686e-01
-8.17785978e-01 -1.35498255e-01 8.56107950e-01 7.31439888e-01
-6.56376362e-01 8.37248564e-01 -6.84310377e-01 -2.33254433e-01
1.51197881e-01 6.88976824e-01 -4.90207762e-01 1.65859517e-02
-1.39111567e+00 1.30942178e+00 4.89876300e-01 -1.27104998e-01
-8.06164324e-01 -8.07043076e-01 -5.31706333e-01 8.34898129e-02
1.15691617e-01 -9.56536293e-01 7.15872943e-01 -5.77508390e-01
-1.69444299e+00 6.50605083e-01 -2.41082590e-02 -3.74236792e-01
5.10435462e-01 1.37392834e-01 -3.64280194e-01 -1.47242561e-01
-1.99603662e-01 5.02216399e-01 2.63735771e-01 -1.24203444e+00
-1.82688460e-02 -3.62260312e-01 4.67550531e-02 9.52067971e-02
5.38286157e-02 -2.80484617e-01 -1.94970608e-01 -6.36924803e-01
-2.99984366e-01 -8.49318564e-01 -3.47225457e-01 -4.92146790e-01
-4.21960205e-01 -1.29357100e-01 4.32461858e-01 -2.45821908e-01
1.38943493e+00 -1.73093414e+00 7.44706213e-01 4.33130890e-01
2.09391505e-01 5.13278425e-01 1.36117507e-02 1.07918012e+00
-4.22994822e-01 1.24743424e-01 -4.45047319e-01 -2.79614031e-02
-8.38159174e-02 1.50169924e-01 -1.77226216e-01 4.31822240e-01
3.48592252e-01 9.52539504e-01 -9.01004910e-01 -1.73041552e-01
3.92216355e-01 1.00021017e+00 -6.81601822e-01 -1.55272484e-02
-5.12582600e-01 8.62042487e-01 -6.76571012e-01 5.80177367e-01
5.11979222e-01 -2.37699956e-01 4.08360988e-01 -1.10488817e-01
-9.19187814e-02 2.48270020e-01 -7.87324667e-01 1.56333554e+00
-3.59632552e-01 2.59371512e-02 -6.83123291e-01 -7.60709763e-01
9.35476482e-01 3.81705165e-01 6.69583321e-01 -6.20712936e-01
1.58063188e-01 3.26194972e-01 3.82750630e-01 -1.49660081e-01
4.29763794e-02 -4.78099048e-01 2.11089313e-01 1.82939991e-01
1.35078132e-01 7.23969415e-02 4.48677868e-01 -1.48086846e-01
1.07169974e+00 4.98477310e-01 5.41985750e-01 -2.77686357e-01
8.65978420e-01 4.57638241e-02 2.73035049e-01 2.21342176e-01
4.75134671e-01 4.86289829e-01 5.10737777e-01 -3.23871821e-01
-7.48817980e-01 -1.06042802e+00 -5.37468076e-01 7.59862542e-01
9.09118205e-02 -6.07751966e-01 -8.30082059e-01 -5.01354337e-01
6.35678917e-02 6.31938398e-01 -9.88436103e-01 -4.13155705e-01
-1.18533239e-01 -1.42019570e+00 3.72404963e-01 -5.66487573e-02
-9.59622860e-02 -1.05581605e+00 -4.38569367e-01 4.34034467e-01
2.21265391e-01 -6.28020048e-01 1.72713980e-01 5.27911842e-01
-8.77431095e-01 -1.32335854e+00 -8.96940708e-01 -8.65358487e-02
4.29081917e-01 -3.43752563e-01 1.10361302e+00 -2.32953086e-01
-4.62967753e-01 9.27698016e-02 -3.26605946e-01 -5.01742482e-01
-7.74377465e-01 9.16461349e-02 -7.71295950e-02 2.69293319e-02
3.89234483e-01 -8.60053182e-01 -8.44506919e-01 4.29984003e-01
-1.06522512e+00 -1.98707998e-01 9.78322029e-01 7.79612660e-01
8.44330013e-01 -4.29249078e-01 6.66469097e-01 -9.71993804e-01
7.26366282e-01 -5.52645028e-01 -4.14662033e-01 2.00911447e-01
-7.87474573e-01 3.30529481e-01 6.39247060e-01 -4.50737447e-01
-7.44733691e-01 1.09815724e-01 -4.99482512e-01 -5.34812771e-02
-3.33198756e-01 8.52238774e-01 -5.15880883e-01 -1.40109241e-01
9.61859345e-01 2.36540601e-01 -8.55240747e-02 -6.50771856e-01
1.74974892e-02 5.60015701e-02 1.33332282e-01 -7.79003143e-01
5.18540025e-01 1.51636600e-01 5.64626873e-01 -8.90671432e-01
-3.57112437e-01 -3.05734366e-01 -5.18763244e-01 -3.17815691e-02
7.95079112e-01 -4.90925789e-01 -9.79225457e-01 -7.28481682e-04
-1.17564404e+00 -4.00477350e-02 -2.12973580e-01 5.95664442e-01
-6.37298822e-01 4.96391952e-01 -5.92134707e-02 -5.46820879e-01
-2.45653823e-01 -1.68029952e+00 9.61465180e-01 -2.53582560e-02
-3.37774903e-01 -1.11303592e+00 5.39425373e-01 9.12513398e-03
2.10496724e-01 7.21668541e-01 1.24521041e+00 -9.79369819e-01
-5.31843960e-01 -3.93326402e-01 9.96344909e-02 5.50981872e-02
1.52634203e-01 -5.84979840e-02 -1.16859365e+00 1.33855399e-02
-4.31338847e-01 -2.21321266e-03 7.60255992e-01 4.17394608e-01
7.81387866e-01 -7.71091506e-03 -4.95666444e-01 2.61931837e-01
1.38230419e+00 4.55012709e-01 9.36346352e-01 2.52074122e-01
3.65914822e-01 5.10245979e-01 5.54047227e-01 4.38460052e-01
-1.72481537e-01 1.09921920e+00 6.61424458e-01 -2.61302620e-01
-2.45660041e-02 -1.81393638e-01 2.83221304e-01 1.09878302e-01
-5.31536162e-01 -5.86630285e-01 -8.62606645e-01 -1.40129983e-01
-1.83304501e+00 -9.93824780e-01 -3.18266332e-01 2.56505775e+00
9.82844353e-01 -3.26077710e-03 2.80980080e-01 8.69100764e-02
4.21387136e-01 -3.42782050e-01 -5.25946021e-01 -2.40307540e-01
-3.55996430e-01 6.78140938e-01 1.34948581e-01 5.21831095e-01
-7.07997441e-01 6.44368351e-01 6.58150387e+00 1.01928258e+00
-1.37967062e+00 -1.30538970e-01 6.56575680e-01 1.12683401e-01
-4.58608896e-01 1.07381798e-01 -8.02293777e-01 4.82903868e-01
1.02470040e+00 -1.28164604e-01 8.27964023e-02 6.56770527e-01
2.46481627e-01 -2.13503301e-01 -1.29135573e+00 8.88359427e-01
-2.12543994e-01 -1.55076766e+00 4.30986762e-01 5.84804475e-01
4.30054128e-01 -2.05883548e-01 1.35512844e-01 1.43776136e-02
-1.18873090e-01 -1.36255980e+00 4.73006129e-01 8.07810664e-01
6.46728218e-01 -8.97564352e-01 7.37699151e-01 1.70165747e-01
-6.65363252e-01 3.63864601e-01 -8.88663307e-02 1.99308157e-01
-3.37879881e-02 6.83197200e-01 -8.09641659e-01 8.89049470e-01
1.50941581e-01 4.68122542e-01 -6.44671738e-01 1.15747118e+00
-2.56377123e-02 3.87782156e-01 -2.20596790e-01 -1.42347544e-01
2.64424652e-01 -3.90626490e-01 4.67181474e-01 1.47929561e+00
2.55180120e-01 -1.65419370e-01 2.97604837e-02 1.12893450e+00
1.86746910e-01 5.39234877e-01 -5.36769092e-01 -2.93693930e-01
-5.74127631e-03 1.15792716e+00 -6.84993207e-01 -1.24114700e-01
-2.55706180e-02 7.04487145e-01 -1.18388429e-01 4.52772200e-01
-1.02995026e+00 -1.10512361e-01 8.55072618e-01 2.48076186e-01
4.27133851e-02 7.54333809e-02 -5.19710854e-02 -7.98457980e-01
-3.63777161e-01 -7.75885761e-01 8.11307132e-02 -2.94175744e-01
-1.00965357e+00 6.09335721e-01 2.11905062e-01 -1.07796466e+00
-3.14076692e-01 -8.49302411e-01 -4.75598097e-01 1.21932304e+00
-1.21126544e+00 -1.05742681e+00 -1.83283165e-01 3.38962913e-01
1.09471634e-01 -6.11856878e-02 1.15258408e+00 5.44972658e-01
-5.00045478e-01 3.86703312e-01 3.02204400e-01 -7.19288766e-01
7.37465322e-01 -1.09656417e+00 1.57939689e-03 3.58624339e-01
1.32622525e-01 9.77616847e-01 1.22644758e+00 -5.73380411e-01
-1.25942051e+00 -8.10335398e-01 6.01000607e-01 -5.93609035e-01
4.53166097e-01 -2.53857672e-01 -9.28544581e-01 -8.34364220e-02
-1.16262101e-01 -2.85960972e-01 1.14946365e+00 1.36823386e-01
-9.19239372e-02 1.77143514e-01 -9.44888771e-01 5.50617993e-01
8.57889831e-01 -2.38905936e-01 -1.21523045e-01 4.47486222e-01
1.77469030e-01 -3.55937779e-01 -1.13293958e+00 5.54100871e-01
4.77766216e-01 -1.20660615e+00 1.18534756e+00 -8.00871074e-01
2.85431385e-01 -6.20087624e-01 -5.46725206e-02 -1.11956143e+00
-2.54928946e-01 -4.63230550e-01 -5.49625009e-02 9.43448424e-01
6.97287261e-01 -6.91450059e-01 7.87438035e-01 2.56418854e-01
-1.63659006e-01 -1.16962767e+00 -9.17287707e-01 -7.19671965e-01
1.28625214e-01 -2.97152281e-01 4.36151803e-01 7.21574008e-01
-1.19189888e-01 6.80364013e-01 -1.46946386e-01 -8.43488723e-02
2.67213136e-01 1.17926501e-01 7.99767554e-01 -1.48540092e+00
-5.94572246e-01 -7.79380322e-01 -6.54062390e-01 -5.45007706e-01
7.90314302e-02 -1.05566883e+00 -3.97710055e-01 -1.38802111e+00
2.34637126e-01 -3.04352582e-01 -3.53121251e-01 1.88945204e-01
-8.01192969e-02 3.38121712e-01 -7.37681985e-02 1.36677936e-01
-2.29810774e-01 6.41227365e-01 1.25762832e+00 -1.04258157e-01
-2.17659265e-01 2.07258731e-01 -5.64582229e-01 3.32753628e-01
5.36964715e-01 -4.26504761e-01 -4.24774319e-01 4.23368782e-01
3.83349895e-01 -8.16533193e-02 4.88129020e-01 -9.70224321e-01
-1.93091840e-01 -2.51487345e-01 4.25973654e-01 -3.52866679e-01
3.69309723e-01 -6.46946549e-01 8.43706012e-01 6.32750094e-01
4.41413298e-02 -3.15335751e-01 2.77455389e-01 6.68658614e-01
-3.77263635e-01 -2.34571099e-01 8.31626236e-01 2.25826427e-02
-3.04476857e-01 4.33128685e-01 -3.24760288e-01 -3.29751879e-01
1.12674844e+00 -5.77887774e-01 9.58549380e-02 -5.43453507e-02
-1.05206585e+00 -3.86644006e-01 6.18844807e-01 1.44394755e-01
3.52315634e-01 -1.04190075e+00 -5.14482677e-01 -2.26974815e-01
3.46609801e-01 -3.48359555e-01 2.33299449e-01 9.71237957e-01
-5.26423156e-01 8.22270334e-01 -1.49378657e-01 -6.53808475e-01
-1.27116168e+00 7.42679358e-01 4.47325289e-01 -4.56164658e-01
-3.01934741e-02 4.34883744e-01 6.79971099e-01 -1.41571209e-01
3.91919762e-02 -1.40597612e-01 -2.06355140e-01 2.79951781e-01
1.77592933e-01 2.33957663e-01 3.89996707e-01 -6.09689891e-01
-4.50069398e-01 5.06292939e-01 8.05718154e-02 2.04444081e-01
1.44257224e+00 3.58810455e-01 -1.73406955e-02 3.09310555e-01
8.95126104e-01 -4.47666533e-02 -9.07151997e-01 1.79150015e-01
1.53883457e-01 -1.34295717e-01 -1.89240098e-01 -1.20787168e+00
-4.48107779e-01 8.22583675e-01 5.78771353e-01 5.17614484e-02
9.34999406e-01 3.10592763e-02 1.37001008e-01 3.39232534e-01
3.72861803e-01 -5.74776828e-01 -3.76708768e-02 2.60883212e-01
9.44797337e-01 -9.69862521e-01 2.18676716e-01 -5.21994293e-01
-3.28737795e-01 1.10722172e+00 1.78167060e-01 1.75268322e-01
2.62894988e-01 -6.40604496e-02 -3.41633052e-01 -3.29263300e-01
-4.63558555e-01 -3.41443419e-01 6.07550859e-01 5.41512847e-01
1.00488174e+00 -1.91672534e-01 -8.16838980e-01 3.75770986e-01
1.05393328e-01 -1.36562102e-02 9.73679405e-03 7.90167451e-01
-2.05292165e-01 -1.80270779e+00 -3.04182947e-01 -4.29831212e-03
-5.14340878e-01 -1.66635633e-01 -7.88497329e-01 9.07877564e-01
2.26418361e-01 5.26889682e-01 -4.92200345e-01 -2.26482227e-01
3.50498348e-01 1.90831289e-01 7.75419414e-01 -7.02818513e-01
-8.61467600e-01 1.74146280e-01 -1.10762015e-01 -3.39733422e-01
-5.75789034e-01 -4.10262525e-01 -1.11905026e+00 -2.47381076e-01
-4.58228320e-01 5.91872036e-01 8.43750536e-01 8.17219198e-01
5.99806011e-01 6.37456298e-01 2.57661730e-01 -1.01894462e+00
-3.07902601e-02 -7.70431697e-01 -4.08149183e-01 2.55862683e-01
-2.14883573e-02 -1.02123857e+00 -4.15635891e-02 -5.50461654e-03] | [5.02547550201416, 5.579461574554443] |
fd2257f8-d51a-4c23-8174-1a5db63d30d8 | attentive-pooling-for-group-activity | 2208.14847 | null | https://arxiv.org/abs/2208.14847v1 | https://arxiv.org/pdf/2208.14847v1.pdf | Attentive pooling for Group Activity Recognition | In group activity recognition, hierarchical framework is widely adopted to represent the relationships between individuals and their corresponding group, and has achieved promising performance. However, the existing methods simply employed max/average pooling in this framework, which ignored the distinct contributions of different individuals to the group activity recognition. In this paper, we propose a new contextual pooling scheme, named attentive pooling, which enables the weighted information transition from individual actions to group activity. By utilizing the attention mechanism, the attentive pooling is intrinsically interpretable and able to embed member context into the existing hierarchical model. In order to verify the effectiveness of the proposed scheme, two specific attentive pooling methods, i.e., global attentive pooling (GAP) and hierarchical attentive pooling (HAP) are designed. GAP rewards the individuals that are significant to group activity, while HAP further considers the hierarchical division by introducing subgroup structure. The experimental results on the benchmark dataset demonstrate that our proposal is significantly superior beyond the baseline and is comparable to the state-of-the-art methods. | ['Zhizhong Zhang', 'Yongqiang Tang', 'Wensheng Zhang', 'Yuan Xie', 'Ding Li'] | 2022-08-31 | null | null | null | null | ['group-activity-recognition'] | ['computer-vision'] | [ 3.03326815e-01 1.12944625e-01 -3.21731627e-01 -2.90326089e-01
-2.02442229e-01 1.71315849e-01 8.04484367e-01 -1.74366646e-02
-4.31499749e-01 7.13734329e-01 9.42777693e-01 6.26543701e-01
-2.90470839e-01 -8.54194164e-01 -2.47370407e-01 -1.02049780e+00
-2.17912927e-01 -1.11021481e-01 4.93318230e-01 1.43879026e-01
3.94332528e-01 -5.41161895e-02 -1.52553070e+00 3.74504954e-01
1.01746345e+00 1.03264165e+00 1.45330712e-01 1.97084859e-01
5.65301217e-02 1.19594419e+00 -7.43630290e-01 1.35365240e-02
-9.45932232e-03 -5.71517706e-01 -7.62462080e-01 4.12579089e-01
1.82868332e-01 -4.44746107e-01 -3.61052752e-01 6.98431015e-01
5.86138427e-01 5.64173818e-01 3.48776489e-01 -1.10102081e+00
-9.86233950e-01 7.35167265e-01 -3.88968617e-01 4.95605111e-01
5.09772241e-01 1.77188113e-01 1.28898442e+00 -8.30070138e-01
1.89265162e-01 1.39181387e+00 2.95598596e-01 4.50392157e-01
-7.41561651e-01 -6.03690386e-01 1.01079297e+00 3.85754377e-01
-1.40444946e+00 -1.95623815e-01 7.64842153e-01 -3.98103267e-01
9.40296948e-01 3.30152333e-01 8.10313106e-01 8.47105920e-01
-2.82531828e-01 1.05932057e+00 1.06362653e+00 -1.00753792e-01
1.08280435e-01 -9.43095684e-02 5.91359615e-01 6.67704046e-01
2.74711818e-01 -3.34010452e-01 -7.11374700e-01 -1.76353425e-01
6.85666323e-01 4.43802744e-01 -6.19550467e-01 -7.24575892e-02
-1.40241504e+00 4.40565705e-01 8.19888055e-01 4.97626543e-01
-6.21293426e-01 1.84722185e-01 1.32862031e-01 -3.52829427e-01
3.74720156e-01 3.21675360e-01 -3.16614546e-02 1.96702451e-01
-6.21496320e-01 2.43153766e-01 3.82371992e-01 6.49242938e-01
8.10177445e-01 -2.55385727e-01 -1.12233591e+00 7.38288820e-01
4.60196465e-01 -8.12014341e-02 8.19600940e-01 -8.34506512e-01
5.47740161e-01 1.30139685e+00 8.63047913e-02 -9.49328482e-01
-3.35862696e-01 -5.15941978e-01 -7.88468957e-01 -3.60117197e-01
4.43103723e-02 -4.64733988e-02 -7.29972482e-01 1.74894381e+00
1.80090070e-01 5.78328133e-01 7.00042546e-02 7.80163109e-01
1.06214440e+00 7.55534649e-01 4.47696298e-01 -4.68031079e-01
1.37366974e+00 -1.45697105e+00 -1.03301978e+00 -3.14035416e-01
4.12055314e-01 -2.18957916e-01 9.94005978e-01 1.40097335e-01
-1.07472384e+00 -8.29889596e-01 -1.24365246e+00 -1.85917672e-02
-4.44852263e-01 5.30138500e-02 8.99294615e-01 6.55153096e-01
-8.40254784e-01 4.80656743e-01 -7.25667477e-01 -4.58889037e-01
7.86820412e-01 5.73901892e-01 -3.70226055e-01 7.00297281e-02
-1.36546862e+00 3.77519339e-01 6.67048097e-01 4.11617845e-01
-8.43843699e-01 -3.84566605e-01 -6.79728687e-01 4.32111531e-01
7.46587813e-01 -6.86008155e-01 9.93216991e-01 -1.05361640e+00
-1.40978420e+00 3.92900497e-01 -3.02238464e-01 -1.90804884e-01
2.55647749e-01 -3.00292253e-01 -1.51703626e-01 1.43255830e-01
-1.77700624e-01 6.67736292e-01 4.55765784e-01 -9.64459598e-01
-8.06203067e-01 -3.77699137e-01 3.25389922e-01 7.52708197e-01
-6.93913221e-01 2.63477355e-01 -6.30832672e-01 -5.73742211e-01
4.85275909e-02 -5.20448029e-01 -1.87646672e-01 -3.39831561e-01
-2.45426759e-01 -7.52829969e-01 9.05933261e-01 -6.19148493e-01
1.79810798e+00 -2.01846147e+00 2.27924734e-01 -6.65217452e-03
2.47842759e-01 4.23265815e-01 1.84050515e-01 2.93882161e-01
2.45775089e-01 3.18231404e-01 -3.08412462e-01 -4.18302149e-01
6.35096431e-02 3.56755972e-01 1.23853073e-01 2.06485033e-01
2.11838916e-01 9.67372477e-01 -9.00098026e-01 -7.67706752e-01
-9.80553851e-02 3.35764349e-01 -7.23487973e-01 5.33668756e-01
1.89042002e-01 3.69852632e-01 -8.03816020e-01 8.65412354e-01
5.76360762e-01 -2.62231767e-01 2.46196583e-01 1.71909388e-02
-1.16550319e-01 1.26005262e-01 -1.19597363e+00 1.26869130e+00
1.06770106e-01 3.61512415e-02 -1.48466349e-01 -7.45715737e-01
8.58795106e-01 2.18267411e-01 4.17646021e-01 -4.73132700e-01
7.38128051e-02 -1.17194220e-01 2.42906928e-01 -4.43963081e-01
7.03483641e-01 3.14366817e-01 -7.39055276e-02 3.56368959e-01
1.60806790e-01 5.90592086e-01 2.31426239e-01 5.82413822e-02
1.02641368e+00 2.48723716e-01 6.53299570e-01 -2.83677578e-01
1.04761004e+00 -8.05541575e-01 1.00036311e+00 7.76791632e-01
-7.66204596e-01 4.34823066e-01 5.98976612e-01 -3.66090983e-01
-2.63882458e-01 -7.95110404e-01 4.17744190e-01 1.36021340e+00
4.67132360e-01 -4.46781397e-01 -9.43564832e-01 -9.12523329e-01
-3.47959787e-01 8.36814567e-02 -9.66542661e-01 -2.57597059e-01
-7.88757205e-01 -1.22813714e+00 3.35069239e-01 8.73431027e-01
1.41304302e+00 -1.55883813e+00 -2.76694506e-01 1.41707629e-01
-3.08517843e-01 -7.64133573e-01 -7.96529531e-01 -3.02391887e-01
-7.93237686e-01 -9.53616798e-01 -7.04900861e-01 -7.85015464e-01
7.65156746e-01 3.22036803e-01 7.27426946e-01 4.63815361e-01
1.56973630e-01 1.79712176e-01 -4.01685148e-01 -2.66477674e-01
5.06143570e-01 3.58667940e-01 -8.26739520e-02 7.53554702e-01
5.34472942e-01 -4.64235991e-01 -1.04021871e+00 6.97170556e-01
-7.21443355e-01 -1.06356114e-01 8.67552757e-01 6.79174244e-01
4.57511872e-01 -3.44319083e-03 7.40652263e-01 -8.52955401e-01
8.95517170e-01 -3.54939848e-01 9.47919413e-02 4.28190231e-01
-4.47759390e-01 -2.01888964e-01 3.49421024e-01 -4.57646847e-01
-1.46704221e+00 -1.73647240e-01 1.11632697e-01 2.02406853e-01
-7.59306550e-02 2.89467782e-01 -7.76038766e-01 -2.75908615e-02
2.03192309e-01 4.09514070e-01 -4.05773252e-01 -4.28463519e-01
-5.82069680e-02 6.83855295e-01 3.62486213e-01 -5.72904646e-01
1.98915988e-01 3.97718191e-01 -2.18124524e-01 -6.30212903e-01
-8.37214112e-01 -4.37652379e-01 -5.66612780e-01 -3.44495684e-01
1.11698496e+00 -9.02138889e-01 -9.06553149e-01 7.44491518e-01
-9.12288487e-01 -2.72339493e-01 -1.40247911e-01 4.33038592e-01
-3.51622999e-01 4.83103871e-01 -5.71801364e-01 -9.11763549e-01
-3.15146327e-01 -9.88034725e-01 7.75409877e-01 7.69553423e-01
-8.61028489e-03 -8.93176138e-01 -1.48353711e-01 6.66292191e-01
3.66855890e-01 2.60099620e-01 4.83076334e-01 -7.59968579e-01
-9.75293636e-01 -1.34235635e-01 -3.64713669e-01 3.89569372e-01
2.18261361e-01 -2.19807863e-01 -1.05186534e+00 -1.52010515e-01
-2.37604752e-01 -1.25287250e-01 1.05451739e+00 2.20680386e-01
1.35894489e+00 -4.34157640e-01 -4.36274439e-01 2.45221660e-01
8.82774174e-01 4.48859990e-01 1.04127657e+00 -2.02734228e-02
7.99095929e-01 6.98641241e-01 6.69799924e-01 3.31075847e-01
4.43996310e-01 7.48688042e-01 3.12012315e-01 -7.61940377e-03
4.07008789e-02 -3.62980664e-01 5.35259664e-01 7.92382360e-01
-9.02128220e-01 -3.03685158e-01 -3.83869797e-01 4.10031915e-01
-2.26333761e+00 -1.21691370e+00 1.24721773e-01 2.10216236e+00
4.58920121e-01 2.15256080e-01 1.61739662e-01 8.21864828e-02
8.14390779e-01 6.46974683e-01 -2.48849049e-01 -1.24582045e-01
-1.44719049e-01 -8.52254108e-02 2.30044663e-01 3.32584321e-01
-1.32234800e+00 7.37933159e-01 6.60409451e+00 9.59076345e-01
-3.66603166e-01 2.69683361e-01 5.00445306e-01 2.65615750e-02
1.52006537e-01 -1.82352647e-01 -1.06409204e+00 6.92475379e-01
4.66194183e-01 -2.19412878e-01 1.53182805e-01 5.99396110e-01
3.35912704e-01 -1.34244338e-01 -9.40340459e-01 6.47165954e-01
3.28052163e-01 -8.31321597e-01 1.68222055e-01 2.74989754e-01
8.01878870e-01 -6.87908947e-01 -1.17751561e-01 6.38240337e-01
1.82914883e-01 -8.75042796e-01 2.23040968e-01 9.19342160e-01
-1.18298471e-01 -5.43025792e-01 1.07086432e+00 3.07486385e-01
-1.57949400e+00 -4.30397511e-01 -2.42603287e-01 -5.00658810e-01
7.27672279e-02 6.15207739e-02 -4.14163768e-01 7.40189016e-01
6.55310094e-01 9.34802711e-01 -9.65741158e-01 1.03244364e+00
-4.73882794e-01 5.93202531e-01 4.81297970e-02 -3.09694439e-01
4.33437765e-01 -8.05230513e-02 3.19802582e-01 1.08087957e+00
1.89398065e-01 2.90315479e-01 5.94434083e-01 5.26221871e-01
-2.44986758e-01 3.47545117e-01 -1.57131910e-01 -5.65316044e-02
2.79768258e-01 1.13093996e+00 -5.13587177e-01 -6.49160445e-01
-4.69814241e-01 7.69647837e-01 4.63973850e-01 2.57274717e-01
-7.43076622e-01 -4.89116788e-01 4.89721924e-01 8.47281143e-02
3.27746153e-01 9.10586491e-02 -1.35855153e-02 -9.72828984e-01
9.42137614e-02 -4.87226248e-01 8.56039166e-01 -7.05823183e-01
-1.03946590e+00 3.42946976e-01 1.41216889e-01 -1.07201028e+00
2.21091717e-01 -3.34207773e-01 -9.92695689e-01 8.62650335e-01
-1.28432560e+00 -1.30587387e+00 -7.54369795e-01 5.06395638e-01
5.54166019e-01 -4.49334159e-02 4.66513813e-01 3.19603473e-01
-9.55824196e-01 6.26017153e-01 -6.77539289e-01 1.40663818e-01
4.59413677e-01 -1.13088822e+00 -2.76681691e-01 1.01166403e+00
-3.81083578e-01 9.45360243e-01 1.95012227e-01 -5.85080862e-01
-5.77372253e-01 -1.12850964e+00 9.28455710e-01 -3.07199508e-01
2.68451750e-01 -2.54876465e-01 -1.08973968e+00 8.05644631e-01
4.37340051e-01 -3.20726000e-02 1.05376530e+00 4.39864993e-02
-7.02129006e-02 -2.65932471e-01 -1.11217117e+00 7.53982604e-01
1.46288562e+00 -1.35550365e-01 -9.28775311e-01 -1.97580317e-03
1.03660679e+00 4.17419411e-02 -8.81066859e-01 4.40473318e-01
6.24039054e-01 -1.09238267e+00 1.00660205e+00 -5.10329962e-01
2.70789474e-01 -7.02717781e-01 -4.89430204e-02 -1.03815651e+00
-9.79263425e-01 -4.84215170e-01 -5.02617598e-01 1.44311380e+00
5.52962720e-02 -9.09145355e-01 8.55186820e-01 5.30981481e-01
-3.54903579e-01 -8.82999837e-01 -6.31357133e-01 -6.40402675e-01
-4.31481093e-01 2.91622907e-01 9.27354276e-01 8.62749696e-01
2.10295096e-01 1.31402522e-01 -6.28214836e-01 2.22015396e-01
2.33501732e-01 -1.48675203e-01 6.10420108e-01 -1.08262491e+00
-5.49265444e-01 -6.56278253e-01 -6.01893008e-01 -1.17176843e+00
-9.58274230e-02 -5.42138278e-01 -6.16656579e-02 -1.80281663e+00
5.92551708e-01 8.73904899e-02 -9.90856051e-01 6.69270992e-01
-8.65142286e-01 3.99593979e-01 8.68113115e-02 2.04185724e-01
-1.09311306e+00 6.15679443e-01 1.45147908e+00 -1.85325071e-01
-5.48604786e-01 6.49217004e-03 -1.15228558e+00 7.54146814e-01
7.87800491e-01 3.97347361e-02 -3.06335092e-01 -8.29362795e-02
-3.73604447e-01 -5.55613518e-01 2.90946245e-01 -1.16148722e+00
3.62171531e-01 -2.74573743e-01 3.46475363e-01 -4.62091267e-01
3.85150313e-01 -4.90031898e-01 -1.04798585e-01 5.56863606e-01
-4.61719573e-01 -4.44121987e-01 -4.16390747e-01 6.30920827e-01
-4.34814751e-01 7.70349009e-03 4.97216344e-01 -3.44532430e-01
-1.02750552e+00 6.53121412e-01 -1.60236463e-01 -1.54685423e-01
1.14996254e+00 -3.00433010e-01 -5.62903941e-01 -1.70562729e-01
-7.38065660e-01 6.54401302e-01 -4.38463204e-02 4.07899857e-01
4.68495250e-01 -1.64481974e+00 -5.17214477e-01 8.67920071e-02
3.39195549e-01 -9.76158231e-02 7.41465867e-01 1.09065855e+00
-1.57160819e-01 3.69557202e-01 -1.30758017e-01 -2.49723658e-01
-1.42327559e+00 5.57753623e-01 3.66760910e-01 -7.52709806e-01
-1.54863998e-01 7.41517544e-01 5.91377616e-01 -1.33524686e-01
3.71054679e-01 -4.36101049e-01 -8.15272033e-01 1.89858630e-01
8.85595024e-01 6.60122395e-01 -4.55047250e-01 -7.73534596e-01
-3.84736598e-01 5.90100884e-01 -1.23013891e-01 2.63077855e-01
8.86917114e-01 3.57108423e-03 -1.66703582e-01 2.81910866e-01
7.27549672e-01 -1.49841398e-01 -1.47158468e+00 -3.14229548e-01
-9.49853137e-02 -6.61976814e-01 -8.87590870e-02 -5.62213838e-01
-9.49684978e-01 8.92911494e-01 4.30340350e-01 4.56974447e-01
1.28168941e+00 -6.82805628e-02 5.45579553e-01 2.21939951e-01
2.81270593e-01 -1.15753400e+00 3.40430379e-01 1.18442535e-01
8.13763022e-01 -1.01961124e+00 3.62961739e-01 -7.63386071e-01
-6.06474698e-01 5.15560627e-01 1.09253228e+00 -1.78393438e-01
4.10460353e-01 -5.36482275e-01 -5.01586914e-01 -4.37555574e-02
-7.66882718e-01 -4.18660790e-01 3.85432929e-01 5.37588537e-01
3.06222200e-01 1.69142604e-01 -7.97028601e-01 1.10871148e+00
4.27048117e-01 2.31603265e-01 2.82642096e-01 1.06901205e+00
-7.94921219e-01 -9.37480867e-01 -3.53806049e-01 4.61264372e-01
-5.09610832e-01 2.83272624e-01 -5.85691094e-01 6.26923621e-01
7.56251872e-01 9.15727377e-01 2.28343144e-01 -2.83895254e-01
5.12691200e-01 1.86044723e-01 2.79055208e-01 -6.53665483e-01
-8.79506409e-01 2.08664350e-02 -6.98717078e-03 -5.42595327e-01
-8.84507775e-01 -6.58415735e-01 -1.03249502e+00 6.08828180e-02
-5.22577278e-02 7.07897097e-02 -2.49123886e-01 9.15580392e-01
2.40692943e-01 8.74956489e-01 5.87516904e-01 -9.53613162e-01
1.46359392e-02 -1.37067258e+00 -6.76588893e-01 4.01209682e-01
-1.39909862e-02 -8.70553315e-01 -4.79612470e-01 -9.18937996e-02] | [8.172174453735352, 0.6370940208435059] |
247a6fca-a0fe-407a-ac09-1c1fd3d3f5bd | facial-emotion-recognition-state-of-the-art | 2105.03588 | null | https://arxiv.org/abs/2105.03588v1 | https://arxiv.org/pdf/2105.03588v1.pdf | Facial Emotion Recognition: State of the Art Performance on FER2013 | Facial emotion recognition (FER) is significant for human-computer interaction such as clinical practice and behavioral description. Accurate and robust FER by computer models remains challenging due to the heterogeneity of human faces and variations in images such as different facial pose and lighting. Among all techniques for FER, deep learning models, especially Convolutional Neural Networks (CNNs) have shown great potential due to their powerful automatic feature extraction and computational efficiency. In this work, we achieve the highest single-network classification accuracy on the FER2013 dataset. We adopt the VGGNet architecture, rigorously fine-tune its hyperparameters, and experiment with various optimization methods. To our best knowledge, our model achieves state-of-the-art single-network accuracy of 73.28 % on FER2013 without using extra training data. | ['Zhuofa Chen', 'Yousif Khaireddin'] | 2021-05-08 | null | null | null | null | ['facial-emotion-recognition'] | ['computer-vision'] | [-7.34533593e-02 -1.51514009e-01 -1.36302695e-01 -6.54584587e-01
-5.59363604e-01 -1.25217913e-02 3.71744074e-02 -3.15062582e-01
-4.70953763e-01 5.95690966e-01 -2.36000776e-01 1.09728865e-01
9.96524245e-02 -4.04081702e-01 -3.29512686e-01 -4.71981674e-01
-2.11982012e-01 2.12142020e-01 -3.07379991e-01 -2.30495259e-01
-4.42870229e-01 6.98278308e-01 -1.56018651e+00 3.55468392e-01
5.56863427e-01 1.69889522e+00 -6.15432143e-01 3.73027623e-01
1.83820650e-01 6.98977947e-01 -4.86330837e-01 -8.78553510e-01
-1.44445971e-01 -1.32937670e-01 -5.60366333e-01 -1.44549131e-01
4.00139004e-01 -1.98269174e-01 -5.08692086e-01 9.66313362e-01
8.94790053e-01 -1.19614050e-01 4.89954352e-01 -1.46417117e+00
-3.68301392e-01 1.04141049e-01 -5.39177179e-01 -7.87658766e-02
2.76373953e-01 4.15859744e-03 5.25129497e-01 -1.01471686e+00
5.47208071e-01 1.13130295e+00 9.57120895e-01 1.14007628e+00
-1.06570375e+00 -1.25333595e+00 -1.11504443e-01 3.13671619e-01
-1.78383362e+00 -7.58723378e-01 8.18819582e-01 -2.69204944e-01
9.75925803e-01 9.21476483e-02 7.93418765e-01 1.36761868e+00
2.05006123e-01 7.06488371e-01 1.06947541e+00 -1.69164941e-01
1.75399370e-02 9.60797220e-02 -8.18189457e-02 1.04712629e+00
-2.54297117e-03 -3.01948208e-02 -4.95813608e-01 -1.18394673e-01
5.80293417e-01 -7.83649459e-02 -2.99001455e-01 2.34902710e-01
-5.29643059e-01 7.20748305e-01 5.42381227e-01 2.54223317e-01
-5.01776159e-01 3.31951857e-01 6.89606845e-01 7.92859569e-02
6.06013417e-01 2.26581171e-01 -6.29783869e-01 -1.66954681e-01
-8.79428625e-01 -9.02674254e-03 7.52500713e-01 6.76677346e-01
3.85272413e-01 3.13402295e-01 -6.10639453e-02 1.16772997e+00
1.83583125e-01 2.91171402e-01 4.45275903e-01 -7.63452828e-01
-6.32781861e-03 5.51321924e-01 -3.82529169e-01 -1.33009481e+00
-6.74875379e-01 -2.81504810e-01 -1.31663334e+00 -1.20737348e-02
-2.81612612e-02 -3.48542005e-01 -1.01242280e+00 1.79470897e+00
1.71066776e-01 2.02367693e-01 -1.46766394e-01 7.57930815e-01
1.50804925e+00 1.06927022e-01 5.01869321e-01 -1.60384700e-01
1.40759468e+00 -6.48250699e-01 -8.16153467e-01 -1.07732095e-01
4.08854902e-01 -5.19919634e-01 6.39739931e-01 5.08189142e-01
-7.88159668e-01 -3.21361184e-01 -7.81079292e-01 9.89786983e-02
-3.19995254e-01 4.66979116e-01 1.08476090e+00 8.55471432e-01
-1.09567356e+00 4.40141231e-01 -8.87745380e-01 -1.75879836e-01
1.16083169e+00 8.05811703e-01 -7.82099962e-01 -8.10199305e-02
-1.14323652e+00 6.93208337e-01 1.42519816e-03 5.88048875e-01
-7.11110234e-01 -5.92754602e-01 -9.37629282e-01 7.80826434e-02
1.02213986e-01 -6.07479393e-01 1.18423951e+00 -1.39288175e+00
-1.65599561e+00 1.27329731e+00 -1.39322191e-01 -2.94738144e-01
4.01931673e-01 -3.73272132e-03 -7.17827797e-01 2.00787306e-01
-5.94755113e-01 7.35885501e-01 7.72208214e-01 -8.88080597e-01
6.63771406e-02 -5.39373338e-01 -3.39933783e-01 -3.20228875e-01
-4.85912770e-01 4.23334807e-01 -5.68921089e-01 -3.43800813e-01
-1.79017991e-01 -1.04497218e+00 -2.74446517e-01 3.20719182e-01
-3.69217426e-01 -1.49640992e-01 6.72785342e-01 -5.25186121e-01
1.22888887e+00 -2.07534456e+00 -2.77349114e-01 2.49413773e-01
5.55767059e-01 6.67002559e-01 -9.43705216e-02 -3.18170160e-01
-2.96286404e-01 3.40544134e-01 3.07867587e-01 -4.40951228e-01
-1.16210535e-01 -1.17108397e-01 3.82634759e-01 6.04830086e-01
3.33349347e-01 1.13581824e+00 -5.31260431e-01 -6.80962265e-01
1.10679030e-01 1.13373375e+00 -7.14264214e-01 1.82058588e-01
1.41538337e-01 2.32860804e-01 -5.57855725e-01 1.06092632e+00
7.76491582e-01 -3.74216318e-01 2.78120160e-01 -7.99784303e-01
5.76542914e-01 -4.63681251e-01 -7.20190823e-01 1.31200445e+00
-3.78616840e-01 7.43502796e-01 1.87576383e-01 -8.12739372e-01
1.03135240e+00 5.34652948e-01 7.60909259e-01 -6.37524128e-01
8.34582806e-01 1.05960399e-01 -3.99156027e-02 -5.53696156e-01
1.24476165e-01 -3.45319867e-01 1.21809125e-01 -1.54161558e-01
4.58062887e-01 2.92758077e-01 -1.84070677e-01 -3.06815118e-01
8.90395820e-01 -1.89296380e-01 2.86935180e-01 3.93963456e-02
3.94744784e-01 -5.34965754e-01 6.30679905e-01 1.86177433e-01
-6.17257535e-01 5.84306359e-01 5.17263353e-01 -7.27059901e-01
-4.72331464e-01 -5.43728352e-01 -4.15478051e-01 7.36795962e-01
-2.72067875e-01 -4.40797567e-01 -8.90487194e-01 -7.01704621e-01
-1.64488748e-01 4.93439436e-02 -1.01178563e+00 -4.09831136e-01
-2.52247959e-01 -9.77163017e-01 9.45826352e-01 4.90049481e-01
6.76534891e-01 -1.22223246e+00 -3.87034684e-01 5.92156686e-02
-1.51407599e-01 -1.47124875e+00 -1.52875885e-01 -1.32176578e-02
-4.92639482e-01 -1.29554677e+00 -6.35631204e-01 -7.05563486e-01
7.13077784e-01 -4.01396751e-01 1.21829927e+00 2.82363892e-01
-8.07646751e-01 3.20880115e-01 -1.65830821e-01 -5.96486092e-01
7.13496804e-02 -1.01952240e-01 -5.15045505e-03 1.67698547e-01
5.71830809e-01 -3.58610392e-01 -8.06373656e-01 2.20317647e-01
-6.19137883e-01 -1.70369431e-01 3.66131216e-01 1.00987065e+00
7.80219495e-01 -8.29123780e-02 3.88990283e-01 -8.41632366e-01
5.20547926e-01 -4.16659772e-01 -2.51326442e-01 3.40080291e-01
-4.48701411e-01 -4.99917984e-01 4.05830801e-01 -3.86491239e-01
-9.73178089e-01 1.84315781e-03 -3.96722943e-01 -8.34276199e-01
-2.98954636e-01 4.76110339e-01 -1.75160810e-01 -6.60083294e-01
6.19363129e-01 -3.04837048e-01 4.39091586e-02 -1.07638054e-01
-4.89411280e-02 7.63497233e-01 3.68006438e-01 -5.04103184e-01
9.82206166e-02 2.43126392e-01 1.88406006e-01 -7.54015028e-01
-8.65664840e-01 -6.29428253e-02 -3.53425533e-01 -5.93730092e-01
7.95962870e-01 -1.10710168e+00 -1.34434271e+00 7.64367700e-01
-1.04244542e+00 -1.25855982e-01 2.48436734e-01 4.19329166e-01
-3.59074652e-01 -2.14230880e-01 -7.79287755e-01 -7.78046489e-01
-9.07717943e-01 -1.24097192e+00 1.14118934e+00 3.92428488e-01
-2.37980843e-01 -9.17894900e-01 -1.88965708e-01 2.82501459e-01
6.04659259e-01 8.13629806e-01 3.71162236e-01 -4.74496126e-01
6.35761023e-02 -4.83104527e-01 -4.30318326e-01 5.69696426e-01
1.44961879e-01 2.34571084e-01 -1.37361765e+00 -1.65316179e-01
-2.94942737e-01 -8.27827573e-01 7.60462999e-01 4.34496820e-01
2.03630495e+00 -1.55918330e-01 -3.76589894e-01 1.11617863e+00
1.40604353e+00 1.18168846e-01 8.70597899e-01 1.51543394e-01
6.38444245e-01 4.02889699e-01 3.13375711e-01 6.23983026e-01
2.41911322e-01 5.15757799e-01 4.75601494e-01 -4.12629843e-01
2.38722861e-02 9.84472707e-02 -1.10410722e-02 3.82190734e-01
-3.50478142e-01 -3.02023977e-01 -9.42452550e-01 1.04071684e-01
-1.41370821e+00 -9.26124692e-01 2.77135283e-01 1.68354678e+00
6.84106112e-01 -2.81576425e-01 -1.07370362e-01 -8.70698839e-02
5.38611650e-01 -1.32797947e-02 -5.69639385e-01 -4.71713990e-01
-1.65670037e-01 7.61425257e-01 1.46945268e-01 -8.40351358e-02
-1.22995651e+00 1.03752625e+00 6.41415358e+00 8.87713671e-01
-1.65872443e+00 3.72336968e-03 1.26656497e+00 -2.52228886e-01
2.46173024e-01 -9.87204015e-01 -7.19823599e-01 2.56642014e-01
1.03431129e+00 9.67811123e-02 4.32661265e-01 9.61951375e-01
-5.10267541e-02 3.44776481e-01 -8.82596195e-01 1.51147711e+00
3.53297263e-01 -1.12988961e+00 -5.72560243e-02 -6.49334043e-02
4.98677224e-01 -6.24595210e-02 2.31505215e-01 3.95887196e-01
-9.92578939e-02 -1.67998266e+00 2.38250926e-01 5.46653867e-01
1.52807558e+00 -9.97868896e-01 1.02604854e+00 -4.12998468e-01
-1.11063421e+00 1.70124128e-01 -1.22152910e-01 4.74908471e-01
-2.19882995e-01 4.34898883e-01 -5.59086800e-01 2.82542080e-01
1.00483906e+00 7.78658271e-01 -3.80352527e-01 8.91164660e-01
-1.15908450e-03 5.15600622e-01 -2.25092158e-01 -8.41689408e-02
-1.82732284e-01 2.18502283e-01 -3.76510099e-02 1.33027446e+00
2.55832136e-01 2.67553568e-01 -1.72164873e-03 5.03719330e-01
-6.91083133e-01 1.87347755e-01 -2.09047019e-01 -2.69527018e-01
1.14308022e-01 1.68200421e+00 -4.00220215e-01 -4.86087091e-02
-3.20407540e-01 7.92710483e-01 5.13320148e-01 3.87083501e-01
-1.05443621e+00 -3.10373127e-01 1.08933234e+00 -3.40311578e-03
3.69160771e-02 1.70279741e-01 -5.26836328e-03 -1.10126173e+00
-9.39898863e-02 -9.48441207e-01 3.90637308e-01 -7.37769067e-01
-1.42908716e+00 1.26294494e+00 -3.27386647e-01 -1.06303227e+00
-3.67928334e-02 -1.00272834e+00 -3.97643626e-01 5.22016227e-01
-1.63844931e+00 -1.35125339e+00 -7.52668738e-01 9.14581299e-01
-2.52177790e-02 -2.89094120e-01 1.23013270e+00 6.08040214e-01
-9.79755461e-01 1.32159328e+00 -2.85315245e-01 4.15185481e-01
7.95042634e-01 -7.13934362e-01 -1.87786937e-01 1.33087412e-01
-3.33052546e-01 4.90206122e-01 3.81791413e-01 -1.50180310e-01
-1.44027138e+00 -1.29955518e+00 7.28128731e-01 1.98179986e-02
4.64813173e-01 -3.48830193e-01 -7.03685999e-01 6.36081696e-01
-5.21105528e-02 7.26943791e-01 1.07947683e+00 2.07124412e-01
-4.65165138e-01 -3.32049817e-01 -1.57433224e+00 5.56337893e-01
9.80707407e-01 -5.29171109e-01 2.06337735e-01 3.23532104e-01
2.49637350e-01 -6.41466320e-01 -1.24063468e+00 7.86040187e-01
9.31069911e-01 -8.60756278e-01 9.21228230e-01 -7.73369849e-01
5.47598243e-01 2.92178094e-01 -9.19132680e-02 -1.30613518e+00
-2.00165749e-01 -5.26881397e-01 -9.24983099e-02 1.16201460e+00
2.57034838e-01 -5.82485437e-01 9.36233222e-01 1.01386774e+00
1.50971428e-01 -1.41613448e+00 -9.15220857e-01 -3.93415362e-01
-2.19604239e-01 -6.08738959e-01 7.37187088e-01 9.90993738e-01
-1.10375233e-01 7.84425288e-02 -4.72587168e-01 -2.17785373e-01
4.60873336e-01 -2.30756968e-01 5.87024152e-01 -1.26629472e+00
1.42693698e-01 -5.62537074e-01 -8.05721104e-01 -1.60525322e-01
6.87651038e-01 -8.07893157e-01 -1.08879514e-01 -1.02838421e+00
4.60151911e-01 -4.42983031e-01 -5.24847388e-01 8.90338778e-01
-4.10142206e-02 7.07057595e-01 7.30009526e-02 -3.06463808e-01
-8.67336392e-01 8.75170231e-01 1.22675490e+00 -1.25820488e-01
1.17127914e-02 -3.34931910e-01 -7.69303560e-01 7.80943990e-01
8.33032846e-01 -3.63672018e-01 1.37453359e-02 -2.66294122e-01
-5.46085238e-02 5.36490977e-02 4.16757852e-01 -8.24122369e-01
6.68341741e-02 1.18530251e-01 9.80953991e-01 1.30155057e-01
5.89979887e-01 -8.52506101e-01 2.18589589e-01 1.81006819e-01
-5.41086979e-02 -1.57099500e-01 6.44405603e-01 1.55517802e-01
-5.04359066e-01 3.24632645e-01 9.15038288e-01 1.99502688e-02
-5.76559007e-01 1.08988094e+00 -1.38698176e-01 4.16713618e-02
9.14430499e-01 -1.33055106e-01 -2.28501454e-01 -2.46996343e-01
-9.01395321e-01 1.25849962e-01 1.70362219e-01 3.40078831e-01
7.63849080e-01 -1.45439589e+00 -5.89319170e-01 2.09718928e-01
2.43599951e-01 -3.43581915e-01 5.48616827e-01 9.08798516e-01
-2.62022018e-01 2.19519049e-01 -4.43770498e-01 -4.93265510e-01
-1.73419476e+00 -1.54244376e-03 8.49048674e-01 -2.18193248e-01
-9.09612328e-02 9.35355783e-01 3.28365229e-02 -3.40955496e-01
3.19894671e-01 2.11060345e-01 -3.91020864e-01 8.61818157e-03
6.82673216e-01 1.62536711e-01 2.95602947e-01 -8.91069114e-01
-7.21748948e-01 6.66776121e-01 -3.56836803e-02 5.53490937e-01
1.42710769e+00 4.05729204e-01 -1.11444809e-01 -1.40191928e-01
1.77101660e+00 -5.74213147e-01 -7.56003976e-01 3.10001173e-03
-5.11366665e-01 -3.84786457e-01 4.76758480e-01 -9.02468860e-01
-1.80355227e+00 8.61146510e-01 9.76205587e-01 -5.99569499e-01
1.58784318e+00 -1.26630068e-01 7.51520455e-01 3.52816314e-01
4.44820940e-01 -9.09431994e-01 1.03543900e-01 3.36692989e-01
8.55766356e-01 -1.48355067e+00 -1.01266645e-01 -4.02249724e-01
-6.51575863e-01 1.27231705e+00 9.40822244e-01 9.23319981e-02
1.16609120e+00 3.55924904e-01 1.79405645e-01 -5.06031454e-01
-7.55659997e-01 -8.33784416e-02 5.48264384e-01 6.29940033e-01
8.32049072e-01 2.45428354e-01 -7.89892003e-02 1.21872556e+00
-1.96674958e-01 6.52855754e-01 -1.17223985e-01 5.98650038e-01
1.34337544e-01 -7.56080747e-01 1.56168237e-01 6.95452213e-01
-1.04168582e+00 7.78736593e-03 -3.31190944e-01 8.14510584e-01
1.51283875e-01 8.45730305e-01 -9.21920836e-02 -6.84582353e-01
6.16762757e-01 -2.35699806e-02 6.92394733e-01 -1.17823169e-01
-8.59343708e-01 -7.31641650e-02 2.48378068e-01 -1.03157961e+00
-4.50951338e-01 -4.40558761e-01 -1.21881163e+00 -4.99511421e-01
-2.90384531e-01 -2.72847295e-01 8.32377732e-01 7.89846182e-01
5.91096222e-01 5.60222387e-01 6.24665439e-01 -8.21549058e-01
-6.12639859e-02 -8.69068384e-01 -6.04581535e-01 4.58500206e-01
1.63099945e-01 -7.83490121e-01 -2.51616329e-01 -1.55147985e-01] | [13.528135299682617, 1.6870688199996948] |
f86f380d-66d6-41ae-b6d5-da26b6dcb387 | scoot-a-perceptual-metric-for-facial-sketches | 1908.08433 | null | https://arxiv.org/abs/1908.08433v2 | https://arxiv.org/pdf/1908.08433v2.pdf | Scoot: A Perceptual Metric for Facial Sketches | Human visual system has the strong ability to quick assess the perceptual similarity between two facial sketches. However, existing two widely-used facial sketch metrics, e.g., FSIM and SSIM fail to address this perceptual similarity in this field. Recent study in facial modeling area has verified that the inclusion of both structure and texture has a significant positive benefit for face sketch synthesis (FSS). But which statistics are more important, and are helpful for their success? In this paper, we design a perceptual metric,called Structure Co-Occurrence Texture (Scoot), which simultaneously considers the block-level spatial structure and co-occurrence texture statistics. To test the quality of metrics, we propose three novel meta-measures based on various reliable properties. Extensive experiments demonstrate that our Scoot metric exceeds the performance of prior work. Besides, we built the first large scale (152k judgments) human-perception-based sketch database that can evaluate how well a metric is consistent with human perception. Our results suggest that "spatial structure" and "co-occurrence texture" are two generally applicable perceptual features in face sketch synthesis. | ['Ming-Ming Cheng', 'Yu-Huan Wu', 'Shengchuan Zhang', 'Deng-Ping Fan', 'Paul L. Rosin', 'Yun Liu', 'Bo Ren', 'Rongrong Ji'] | 2019-08-21 | scoot-a-perceptual-metric-for-facial-sketches-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Fan_Scoot_A_Perceptual_Metric_for_Facial_Sketches_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Fan_Scoot_A_Perceptual_Metric_for_Facial_Sketches_ICCV_2019_paper.pdf | iccv-2019-10 | ['face-sketch-synthesis'] | ['computer-vision'] | [-1.08909167e-01 -5.46819925e-01 -2.16291279e-01 -5.49093008e-01
-2.69555539e-01 -3.08641136e-01 9.16836500e-01 -1.35723159e-01
7.21810609e-02 3.13674927e-01 2.03863055e-01 7.10978592e-03
-2.33881325e-01 -6.35608077e-01 -1.84582815e-01 -4.27739620e-01
1.42790675e-01 -2.28573412e-01 4.07450348e-01 -3.03371549e-01
6.54589951e-01 7.30630457e-01 -1.91274476e+00 3.53075325e-01
6.97427034e-01 1.27564657e+00 7.49856085e-02 2.63043642e-01
-3.01325917e-01 2.84737468e-01 -4.51078236e-01 -6.45411372e-01
2.12916747e-01 -4.11035687e-01 -3.90701294e-01 -1.22737952e-01
7.19937563e-01 -3.30353856e-01 1.62847508e-02 1.03302431e+00
5.14526665e-01 -2.00720847e-01 8.87117922e-01 -1.54266870e+00
-7.97065973e-01 -2.93554179e-02 -6.63090229e-01 -8.36442187e-02
6.69998527e-01 3.61333452e-02 9.67356920e-01 -1.25912488e+00
4.73253489e-01 1.55491388e+00 6.93067014e-01 2.40782201e-01
-9.09986258e-01 -9.50187325e-01 1.51390368e-02 9.01032239e-02
-1.69286227e+00 -5.92056572e-01 1.16143060e+00 -3.41954380e-01
5.21680295e-01 4.47871327e-01 5.06343484e-01 7.00310409e-01
2.21938550e-01 3.88809472e-01 1.58220100e+00 -4.72210169e-01
1.05279259e-01 2.40298644e-01 -3.32232714e-01 8.11595201e-01
3.67631689e-02 1.53614953e-01 -7.51918018e-01 -2.24745229e-01
1.19346058e+00 -1.59800649e-01 1.69939846e-01 -1.56142831e-01
-1.04945886e+00 4.54428047e-01 1.64774582e-01 5.57262063e-01
-1.60683244e-01 4.74863499e-03 2.11207151e-01 3.40448350e-01
3.64232600e-01 1.11587152e-01 3.27990465e-02 -2.20563039e-01
-1.03168869e+00 2.86972791e-01 4.25791353e-01 9.07891273e-01
6.27710700e-01 1.22090816e-01 -3.05599660e-01 1.24639678e+00
4.48621601e-01 7.68748879e-01 2.22684786e-01 -1.04379416e+00
-5.08332029e-02 7.23597109e-01 -8.89365748e-03 -1.67501903e+00
-1.14318207e-01 1.94737300e-01 -8.81469607e-01 5.08755028e-01
1.98099777e-01 4.91519958e-01 -5.94533026e-01 1.79071259e+00
2.76895225e-01 1.35652423e-01 -4.90644217e-01 9.53328669e-01
1.17411840e+00 2.81472325e-01 2.58281410e-01 -7.47533962e-02
1.42647350e+00 -4.64549392e-01 -8.57625067e-01 2.50843763e-01
-1.73365653e-01 -1.35668457e+00 1.48983741e+00 3.66084903e-01
-1.21598423e+00 -1.02661574e+00 -1.18376446e+00 1.63848832e-01
-4.16430712e-01 4.26086426e-01 6.97492898e-01 1.01123381e+00
-1.13097274e+00 6.28230751e-01 -1.35305181e-01 -4.85862046e-01
1.38318971e-01 8.22471008e-02 -6.56414688e-01 1.45809650e-01
-9.74362671e-01 8.65873516e-01 -1.55833811e-01 -1.20732613e-01
-4.78969127e-01 -4.70504761e-01 -4.35770541e-01 -1.08240135e-01
1.05152346e-01 -3.58217001e-01 8.11679661e-01 -9.56930637e-01
-1.67847490e+00 7.97093272e-01 -3.23990554e-01 5.32060921e-01
3.10525477e-01 2.99188077e-01 -7.35466182e-01 3.46653283e-01
-3.23362947e-02 9.04097378e-01 1.05705714e+00 -1.53010798e+00
-4.49031353e-01 -3.51525515e-01 4.45911884e-02 1.21701479e-01
-5.40451884e-01 2.96342432e-01 -4.77545291e-01 -1.00534225e+00
3.06097001e-01 -7.52475798e-01 3.55142593e-01 6.78271234e-01
1.57785900e-02 -5.08352757e-01 7.35149741e-01 -5.54149151e-01
1.39482605e+00 -2.26221275e+00 -3.02964032e-01 5.17459571e-01
7.89899454e-02 2.70861417e-01 -3.40164781e-01 6.09247744e-01
1.87504441e-01 3.08784038e-01 -8.08537155e-02 -1.24751560e-01
2.01563030e-01 9.34287906e-02 -2.54922599e-01 2.19241574e-01
2.55698383e-01 7.91412473e-01 -6.09795272e-01 -8.80562425e-01
1.47051677e-01 4.63676095e-01 -3.85029852e-01 1.85369819e-01
2.49956980e-01 6.68133423e-02 -3.75797927e-01 8.69240463e-01
1.02398932e+00 3.63230184e-02 1.47271425e-01 -4.57282573e-01
2.72966898e-03 -2.27407485e-01 -1.21637607e+00 1.42527270e+00
-2.23659679e-01 5.25768518e-01 -1.39711589e-01 -4.36930150e-01
1.42723167e+00 2.51885682e-01 2.45192572e-01 -1.18606448e+00
-1.02381952e-01 2.49779746e-01 -1.35581493e-01 -4.33245689e-01
5.76517642e-01 -2.74827689e-01 3.02391738e-01 3.25499892e-01
-1.00012146e-01 -1.75847843e-01 -7.66387880e-02 2.98711117e-02
6.21257782e-01 1.43224537e-01 4.21233058e-01 -6.55162930e-01
7.12375164e-01 -7.01481760e-01 3.42866600e-01 4.33315456e-01
-4.04760540e-01 7.89254606e-01 7.47343361e-01 -1.44558534e-01
-1.11906219e+00 -1.26777196e+00 -2.00006142e-01 1.09001529e+00
3.97085935e-01 -5.68006337e-01 -7.21023500e-01 -4.35507566e-01
2.61351824e-01 1.12202270e-02 -7.79489219e-01 5.29077537e-02
-1.35920167e-01 -1.66085735e-01 8.77671778e-01 6.10548437e-01
6.33343577e-01 -1.13181114e+00 -3.12694728e-01 -2.89264828e-01
8.82668495e-02 -9.35509026e-01 -5.97608149e-01 -8.72013807e-01
-6.98763847e-01 -9.76247787e-01 -8.44463348e-01 -7.06285596e-01
6.09977603e-01 6.00210190e-01 1.01326275e+00 5.85156679e-01
-2.94265300e-01 4.40283567e-01 -4.66881692e-01 -1.71243787e-01
-1.36620015e-01 -4.52493012e-01 1.51243806e-01 2.57523656e-01
1.40080675e-01 -7.03453720e-01 -8.20494950e-01 7.72382557e-01
-9.70181882e-01 4.83867899e-02 6.46939874e-01 5.79259813e-01
3.64825130e-01 -9.64128152e-02 7.13404715e-01 -2.75745779e-01
1.06121504e+00 -7.44045079e-02 -8.60819966e-02 5.26009619e-01
-6.94066703e-01 -4.33241315e-02 4.21327114e-01 -4.45520580e-01
-9.81512249e-01 -5.62591076e-01 -1.79133654e-01 -3.60400677e-01
-1.64385095e-01 3.37873101e-01 -2.58330554e-01 -4.93968844e-01
3.21305990e-01 2.20069900e-01 2.04000384e-01 -3.13006282e-01
1.41109511e-01 5.33508360e-01 4.04488474e-01 -9.25168216e-01
6.90839708e-01 4.61786628e-01 2.40325287e-01 -1.15169358e+00
-1.10739820e-01 -2.60465711e-01 -5.97892582e-01 -5.11742234e-01
4.56531644e-01 -5.60239375e-01 -1.03567648e+00 4.60450947e-01
-1.04189444e+00 1.91173911e-01 2.34951422e-01 2.12999269e-01
-4.53561395e-01 7.51388907e-01 -2.11948410e-01 -1.06701398e+00
-1.31850600e-01 -1.10569060e+00 1.24361765e+00 1.75759196e-01
-2.13338643e-01 -6.90703809e-01 -2.00886987e-02 3.47300358e-02
6.86365902e-01 1.86271623e-01 9.12129402e-01 2.00957075e-01
-2.70104498e-01 5.03211394e-02 -7.89438367e-01 2.48336032e-01
3.29052567e-01 4.84739006e-01 -1.02441311e+00 -9.29898098e-02
-4.20697868e-01 -1.46002382e-01 5.38468838e-01 1.69317201e-01
1.18366313e+00 -1.96379915e-01 -5.32214865e-02 3.63249779e-01
1.21882975e+00 2.87836969e-01 9.43683386e-01 -2.51882821e-01
3.68622154e-01 7.77303815e-01 7.35681772e-01 5.14254868e-01
3.66150320e-01 1.05786681e+00 -5.98149039e-02 -9.07837451e-02
-3.38797957e-01 -5.09615242e-01 2.54818767e-01 9.17129219e-01
-5.04768550e-01 4.73124310e-02 -6.85596645e-01 1.81147337e-01
-1.52792430e+00 -9.16371882e-01 2.39782795e-01 2.12465739e+00
6.87814057e-01 6.87863380e-02 2.78709292e-01 4.09630924e-01
5.30361533e-01 2.60553360e-01 -4.78274375e-03 -5.52735448e-01
-2.42131174e-01 3.81106585e-01 -3.16658355e-02 3.68250102e-01
-7.22778738e-01 9.65185106e-01 6.80223846e+00 1.37270510e+00
-1.22793829e+00 -1.44715428e-01 6.65466368e-01 3.57654512e-01
-5.17269313e-01 3.40953073e-03 -4.67034370e-01 5.86077869e-01
4.60466683e-01 1.22951414e-03 3.34531397e-01 7.58729696e-01
7.33243451e-02 -1.94217101e-01 -9.09657896e-01 1.29902470e+00
3.08514863e-01 -1.11884487e+00 3.89919460e-01 3.89842577e-02
4.43415195e-01 -8.99154603e-01 4.72512811e-01 2.20390316e-02
-3.04657668e-01 -1.35974872e+00 8.18898141e-01 8.39588046e-01
1.34387302e+00 -6.30257905e-01 4.46435899e-01 -1.72209814e-01
-1.69875598e+00 2.79193848e-01 -4.17943150e-01 4.80754272e-04
-1.60260424e-01 1.85082763e-01 -4.73492056e-01 4.35226738e-01
5.77700436e-01 5.75182855e-01 -8.90090942e-01 8.69351745e-01
1.39581310e-02 2.52625257e-01 -1.48834810e-01 -2.94955879e-01
2.98355911e-02 -1.95847824e-01 2.37856910e-01 1.21025944e+00
4.39113855e-01 2.31252030e-01 6.66583842e-03 8.92879546e-01
3.07878882e-01 5.60397804e-01 -6.27143145e-01 -1.26985908e-01
7.77396798e-01 1.25998473e+00 -9.44540322e-01 -2.44524240e-01
-4.17420268e-01 8.30659568e-01 -2.61115041e-02 2.20319778e-01
-6.82326972e-01 -2.67094344e-01 7.80269563e-01 3.47274184e-01
-1.82279516e-02 -3.55656743e-01 -4.95477796e-01 -7.16156006e-01
2.23606989e-01 -8.91869366e-01 -9.00997519e-02 -9.56089377e-01
-1.43925750e+00 5.79074919e-01 6.60192268e-03 -1.34824872e+00
1.80077240e-01 -6.79656804e-01 -7.50200331e-01 8.54586482e-01
-1.34526193e+00 -1.56201386e+00 -6.11078858e-01 6.37078226e-01
3.18895310e-01 -4.14725304e-01 7.27096081e-01 4.17001575e-01
-2.37958223e-01 1.01025033e+00 -4.55662876e-01 -4.60439622e-02
8.85992229e-01 -8.61176908e-01 3.54396671e-01 3.51347029e-01
1.11274436e-01 8.63430679e-01 5.08062124e-01 -7.23517656e-01
-1.37357330e+00 -4.50992197e-01 7.37384796e-01 -3.45469803e-01
2.64967471e-01 -2.49915674e-01 -8.61843169e-01 -1.10694245e-01
3.31647731e-02 -2.55364813e-02 4.42799836e-01 1.64654683e-02
-7.46779323e-01 -3.48456234e-01 -1.22390628e+00 6.93631291e-01
1.17258620e+00 -7.38667488e-01 -3.78602177e-01 -2.99331039e-01
1.55788764e-01 1.61136940e-01 -1.09517491e+00 7.79139459e-01
1.31989145e+00 -1.38146913e+00 1.19397461e+00 -3.29867810e-01
4.43916380e-01 -3.13396513e-01 -4.43517447e-01 -8.29667389e-01
-4.31667835e-01 -4.12823796e-01 2.48948559e-01 1.33312237e+00
1.45579010e-01 -5.09765863e-01 6.27503395e-01 2.61591911e-01
3.15988749e-01 -8.71847928e-01 -8.73181701e-01 -9.04891968e-01
-6.42196909e-02 -2.45917201e-01 9.77269232e-01 9.87051725e-01
2.08114274e-02 2.59157643e-03 -4.30746585e-01 -2.22842231e-01
4.67034250e-01 8.44167471e-02 8.75606060e-01 -1.40930486e+00
2.78665185e-01 -9.83911872e-01 -6.00938976e-01 -6.58917487e-01
1.39014507e-02 -6.43479824e-01 -2.64662355e-01 -1.26133990e+00
3.10889572e-01 -7.67477751e-01 -3.25529695e-01 3.19830686e-01
-1.54214099e-01 4.99569654e-01 3.10542315e-01 1.72103450e-01
-3.57147694e-01 6.11846507e-01 1.43658793e+00 1.01637140e-01
1.57279268e-01 -4.76241261e-01 -5.29772222e-01 6.91236913e-01
6.04122877e-01 1.13139994e-01 -4.43727762e-01 -1.94737054e-02
1.41015679e-01 1.30598277e-01 4.29667354e-01 -1.10005069e+00
1.74404815e-01 -6.18328869e-01 5.53634763e-01 -4.34587061e-01
5.68041027e-01 -6.57698870e-01 1.34242326e-01 1.68345779e-01
-1.93360612e-01 3.36100847e-01 1.04696251e-01 2.57871360e-01
-3.77054363e-01 4.91080210e-02 8.61364424e-01 4.52467538e-02
-9.17349935e-01 3.57775301e-01 -5.79848466e-03 -3.17845732e-01
7.05402553e-01 -6.57635808e-01 -1.87284157e-01 -5.42762697e-01
-1.92933291e-01 -3.95727158e-01 5.65476120e-01 7.91823387e-01
1.01735985e+00 -1.82654107e+00 -6.75722897e-01 6.09089196e-01
3.83800626e-01 -1.08530724e+00 2.66193569e-01 7.15791285e-01
-4.84294266e-01 3.91240716e-01 -7.13019311e-01 -5.21858931e-01
-1.44982529e+00 3.07316035e-01 -9.39747617e-02 3.39411050e-01
-2.18244180e-01 6.48511887e-01 3.22526962e-01 -2.16219962e-01
1.46096215e-01 -2.87851065e-01 -2.58766234e-01 6.75738007e-02
5.34666717e-01 4.02919739e-01 -5.59926480e-02 -8.99769306e-01
-4.70514774e-01 1.05925226e+00 3.44446331e-01 -2.99701571e-01
8.10898304e-01 1.00345882e-02 -2.10482463e-01 2.96519518e-01
1.01716554e+00 2.21159056e-01 -9.48610544e-01 -5.52481078e-02
-3.28561962e-02 -9.96511996e-01 -2.07272604e-01 -7.36178875e-01
-9.57408309e-01 9.14611638e-01 7.84496903e-01 2.27931321e-01
1.22025001e+00 -3.18761468e-01 5.96852839e-01 -3.21633518e-02
5.12967348e-01 -1.16283727e+00 5.29024243e-01 2.84191191e-01
1.31574881e+00 -1.09860897e+00 4.91471142e-02 -6.63565218e-01
-7.84384429e-01 1.05773151e+00 7.57556796e-01 -1.71471968e-01
9.31756854e-01 2.18838245e-01 -1.01195700e-01 -1.19106665e-01
-5.79298437e-01 -1.12011090e-01 9.04303610e-01 4.39123482e-01
6.79710865e-01 2.22354844e-01 -5.18281579e-01 4.57890719e-01
-3.06545377e-01 4.87541258e-02 -2.86014304e-02 7.85949826e-01
-6.14604652e-01 -1.19336796e+00 -3.25556010e-01 3.87214094e-01
-2.06011385e-01 1.00383960e-01 -5.61168849e-01 6.94933891e-01
2.25376248e-01 8.22295606e-01 -5.37013672e-02 -7.25196123e-01
2.87200034e-01 -1.55853033e-01 8.68392766e-01 -1.76804155e-01
-2.42610902e-01 -6.80114925e-02 -1.52600095e-01 -6.54886246e-01
-6.34429276e-01 -2.53857046e-01 -6.96484804e-01 -7.23935306e-01
-1.49930909e-01 -1.02576271e-01 6.81316555e-01 7.03084290e-01
2.73297638e-01 -1.20874941e-02 6.81113601e-01 -6.68756306e-01
-1.54967427e-01 -1.01051617e+00 -7.47301698e-01 5.79851270e-01
2.00166069e-02 -1.18981683e+00 -4.79960404e-02 -5.92758991e-02] | [12.783318519592285, 0.2269609570503235] |
914a8051-2205-4a2c-ac41-aaf32ea14b88 | sibert-enhanced-chinese-pre-trained-language | null | null | https://aclanthology.org/2020.lrec-1.293 | https://aclanthology.org/2020.lrec-1.293.pdf | SiBert: Enhanced Chinese Pre-trained Language Model with Sentence Insertion | Pre-trained models have achieved great success in learning unsupervised language representations by self-supervised tasks on large-scale corpora. Recent studies mainly focus on how to fine-tune different downstream tasks from a general pre-trained model. However, some studies show that customized self-supervised tasks for a particular type of downstream task can effectively help the pre-trained model to capture more corresponding knowledge and semantic information. Hence a new pre-training task called Sentence Insertion (SI) is proposed in this paper for Chinese query-passage pairs NLP tasks including answer span prediction, retrieval question answering and sentence level cloze test. The related experiment results indicate that the proposed SI can improve the performance of the Chinese Pre-trained models significantly. Moreover, a word segmentation method called SentencePiece is utilized to further enhance Chinese Bert performance for tasks with long texts. The complete source code is available at https://github.com/ewrfcas/SiBert{\_}tensorflow. | ['Chenjie Cao', 'Xiuyan Jiang', 'Jiahao Chen'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['cloze-test'] | ['natural-language-processing'] | [ 6.67233467e-02 -2.59346604e-01 -2.17048451e-01 -6.19112194e-01
-1.03456187e+00 -4.26958442e-01 1.83901265e-01 3.37521434e-01
-6.42451763e-01 4.88561004e-01 5.24159312e-01 -4.57083762e-01
5.01822308e-02 -7.18440175e-01 -2.99688876e-01 -3.04124087e-01
3.38115185e-01 3.11605394e-01 2.70431548e-01 -4.94020879e-01
5.12567282e-01 -6.54853955e-02 -8.90145481e-01 7.27581084e-01
1.40014207e+00 6.08920932e-01 7.68179953e-01 6.76670730e-01
-4.88750905e-01 8.15857172e-01 -6.23150945e-01 -3.85870367e-01
-1.80048421e-01 -5.92061281e-01 -1.22843969e+00 -5.16928844e-02
7.61961639e-02 -2.55891740e-01 -3.10907692e-01 9.09942150e-01
7.12455153e-01 4.20529783e-01 4.52837467e-01 -5.68780243e-01
-1.13992167e+00 1.07405078e+00 -3.13280344e-01 5.89142084e-01
2.84260303e-01 8.49600658e-02 1.21556008e+00 -9.11390662e-01
2.71748066e-01 1.03686702e+00 3.86487037e-01 5.67969263e-01
-8.18050385e-01 -7.38212228e-01 2.33340323e-01 5.20497203e-01
-1.27530670e+00 -4.19707090e-01 8.66346180e-01 -1.15548179e-01
1.13938129e+00 2.27287963e-01 -2.84533426e-02 1.00845194e+00
-1.81628719e-01 1.29783905e+00 9.30916309e-01 -5.89574337e-01
-2.47598767e-01 1.57791480e-01 6.54427409e-01 4.39708352e-01
-9.45732072e-02 -4.18579787e-01 -1.88350543e-01 3.04675221e-01
4.89683688e-01 -8.37736502e-02 -4.74253356e-01 3.47020179e-01
-1.19632125e+00 9.57717896e-01 5.74100077e-01 7.84258902e-01
-1.28821447e-01 -2.60174245e-01 7.04287410e-01 4.69329268e-01
7.30356038e-01 7.37118840e-01 -8.79187584e-01 -1.96884379e-01
-7.56388605e-01 -4.23471518e-02 6.07460737e-01 1.22220266e+00
7.20095754e-01 4.98861354e-03 -7.02284455e-01 1.17639530e+00
1.18137449e-01 1.58451170e-01 1.16294372e+00 -5.50809205e-01
1.00376821e+00 7.50360250e-01 -2.28205144e-01 -9.03772771e-01
-4.29058373e-01 -5.73849559e-01 -7.12925732e-01 -9.18015242e-01
2.96752632e-01 -4.99672234e-01 -5.99685967e-01 1.38323057e+00
3.38289738e-02 -1.37605578e-01 2.14956194e-01 7.64011919e-01
1.24897933e+00 1.06722331e+00 1.95213675e-01 -1.09753080e-01
1.21218491e+00 -1.59875095e+00 -8.81238639e-01 -2.82280952e-01
1.28949142e+00 -1.02605331e+00 1.62569177e+00 3.25969867e-02
-9.74000990e-01 -9.27347720e-01 -7.80723989e-01 -5.70343792e-01
-3.62456352e-01 5.11591792e-01 4.55291510e-01 5.40286958e-01
-7.56458044e-01 2.68590420e-01 -5.41682959e-01 -3.25625122e-01
2.91771382e-01 1.38801828e-01 6.69451952e-02 -4.13469732e-01
-1.55723011e+00 6.74993455e-01 6.53611660e-01 3.14716399e-01
-5.61586678e-01 -4.94072080e-01 -7.72777021e-01 2.16390520e-01
2.73437411e-01 -3.21060926e-01 1.34032023e+00 -9.31959927e-01
-1.64878345e+00 8.89593303e-01 -3.37628245e-01 -2.25884289e-01
8.70910287e-03 -6.11986101e-01 -4.47404981e-01 3.08154583e-01
2.84834534e-01 6.35411799e-01 5.35260141e-01 -6.72300279e-01
-2.38856182e-01 -2.72750050e-01 9.81266052e-02 4.56085503e-01
-8.04146230e-01 3.39304090e-01 -7.90753722e-01 -8.47723901e-01
-8.58269632e-02 -6.01608694e-01 -3.51304322e-01 -7.94931531e-01
-5.51760018e-01 -7.25777626e-01 4.23989177e-01 -8.78979921e-01
1.71209490e+00 -2.03457332e+00 -1.70904681e-01 -2.66645312e-01
-2.04936728e-01 5.39895475e-01 -6.05740964e-01 6.54764175e-01
-2.27343440e-02 2.60562092e-01 -5.84676601e-02 -3.12576532e-01
-2.93104172e-01 -7.80537203e-02 -1.28308684e-01 9.81940478e-02
3.11041057e-01 1.23182392e+00 -9.71747756e-01 -6.09381676e-01
-1.21817909e-01 -4.69412580e-02 -5.04535854e-01 5.36603510e-01
-3.58323365e-01 6.73952818e-01 -7.48033404e-01 3.21861982e-01
6.21303678e-01 -4.35352623e-01 9.56658870e-02 2.23889604e-01
-1.59601439e-02 7.84689605e-01 -6.53853774e-01 2.05730438e+00
-6.29150629e-01 5.07303834e-01 -2.53567785e-01 -1.20179868e+00
1.17189276e+00 3.04243207e-01 1.01274669e-01 -9.87577736e-01
2.38668814e-01 6.00758269e-02 1.92180231e-01 -1.09036613e+00
7.23824561e-01 2.44540691e-01 -1.08584724e-01 4.05262560e-01
1.03856124e-01 2.16678455e-02 3.88404846e-01 2.80481875e-01
8.41259778e-01 1.74743608e-01 7.60092819e-03 -3.67220700e-01
1.03827822e+00 2.50552088e-01 4.97862995e-01 5.22259176e-01
-2.27116421e-01 7.37149060e-01 3.05629045e-01 1.01138778e-01
-7.48862088e-01 -6.65218353e-01 -6.61135614e-02 1.59252226e+00
-3.05545852e-02 -5.67012906e-01 -9.71690416e-01 -7.22761631e-01
-2.75084466e-01 8.73747945e-01 -2.72346795e-01 -2.13811666e-01
-8.09953332e-01 -5.32175243e-01 6.47086084e-01 7.07301676e-01
6.98361397e-01 -1.36232710e+00 7.33891279e-02 2.77002126e-01
-5.82088470e-01 -1.16136169e+00 -6.59621358e-01 -8.51962417e-02
-1.16673720e+00 -7.45885670e-01 -1.00219572e+00 -1.31021404e+00
7.80959725e-01 4.96577978e-01 1.01585984e+00 4.56090540e-01
1.54091880e-01 8.77576098e-02 -1.09439886e+00 -2.18564570e-01
-1.80235207e-01 6.75549328e-01 -2.78527021e-01 -1.18635513e-01
9.03313279e-01 -6.88583329e-02 -7.27665424e-01 3.97354364e-01
-8.52922201e-01 1.65890560e-01 5.33119082e-01 7.73744047e-01
4.04257357e-01 -2.12663502e-01 1.06944036e+00 -1.04787123e+00
1.08100426e+00 -6.41721010e-01 -2.58648306e-01 4.12704915e-01
-3.17614943e-01 9.31759104e-02 7.94095397e-01 -4.99109417e-01
-1.37182093e+00 -3.58306646e-01 -4.95818257e-01 5.55825792e-02
-2.97749132e-01 9.02716875e-01 -3.19570154e-01 5.53061128e-01
6.31639183e-01 2.50592679e-01 -1.41939521e-01 -9.50261414e-01
4.91893351e-01 1.04458308e+00 1.66091800e-01 -5.96428871e-01
4.47248191e-01 -1.26562119e-01 -9.16546106e-01 -8.08569849e-01
-1.23862290e+00 -8.18336666e-01 -8.79026353e-01 1.72294691e-01
9.15525198e-01 -9.45361435e-01 -4.30495232e-01 4.55348104e-01
-1.24578750e+00 -5.84370732e-01 1.33043915e-01 5.62305987e-01
-1.40509516e-01 5.83898246e-01 -9.18572843e-01 -4.24301237e-01
-6.19110286e-01 -1.00647557e+00 7.03990698e-01 3.73631895e-01
-1.28600329e-01 -1.09513927e+00 -2.05128845e-02 9.12951112e-01
4.85734254e-01 -6.06542230e-01 9.17322278e-01 -1.04525268e+00
-3.28520507e-01 -2.00902015e-01 -3.20298404e-01 8.52565825e-01
1.64716169e-01 -4.08433914e-01 -7.92359769e-01 -1.56625897e-01
1.60812318e-01 -4.97289151e-01 1.03687441e+00 2.40269378e-01
1.58130836e+00 -2.61592299e-01 -3.00123878e-02 4.13566083e-01
1.18719149e+00 5.67544848e-02 5.40905535e-01 4.18922514e-01
7.77334213e-01 7.03734994e-01 7.32824802e-01 2.15483475e-02
5.36181450e-01 2.94101655e-01 -2.70516604e-01 1.64938852e-01
-1.63926423e-01 -3.04318190e-01 3.70864213e-01 1.65501058e+00
2.43763521e-01 -3.45004082e-01 -9.80530858e-01 6.08594120e-01
-1.77603412e+00 -7.08276927e-01 -4.58224118e-01 1.79309845e+00
1.25405312e+00 1.38250738e-01 -2.92801112e-01 -5.68093397e-02
8.21887732e-01 2.07421541e-01 -3.89350116e-01 -4.42699343e-01
-9.14165378e-02 2.92952031e-01 2.35124290e-01 4.41615105e-01
-1.16988850e+00 1.52972949e+00 4.97344065e+00 1.12617409e+00
-1.00110602e+00 3.40126067e-01 6.77316487e-01 3.29450250e-01
-4.32462215e-01 -7.68605843e-02 -1.05029917e+00 4.07791823e-01
1.07062578e+00 -2.34626144e-01 -9.91611183e-03 7.45933473e-01
4.87903684e-01 -4.37118821e-02 -7.95295238e-01 7.09653795e-01
2.58583754e-01 -1.20077264e+00 7.95959607e-02 -5.31010091e-01
9.10740674e-01 1.73432112e-01 -7.44827604e-03 7.35461056e-01
-7.15220273e-02 -7.81699181e-01 5.15647195e-02 4.72645432e-01
6.79483533e-01 -5.68015933e-01 7.90395677e-01 5.83762944e-01
-1.00432849e+00 -5.26252985e-02 -7.07794964e-01 -2.93606311e-01
1.60877764e-01 3.64219904e-01 -7.50903785e-01 4.93405610e-01
4.60501909e-01 7.89842486e-01 -1.08139896e+00 1.19881296e+00
-7.27678180e-01 1.18525231e+00 1.72032222e-01 -5.17625391e-01
4.81981426e-01 -1.31753132e-01 1.03946760e-01 1.47124672e+00
6.03027716e-02 1.08931132e-01 -4.04476514e-03 5.70943236e-01
-1.95114002e-01 7.81426907e-01 -1.65321872e-01 -2.42866129e-01
3.63389283e-01 1.15974391e+00 -6.08206451e-01 -3.79373282e-01
-5.55925548e-01 9.84474301e-01 5.77803671e-01 6.43895745e-01
-5.60428083e-01 -8.11126709e-01 2.49748856e-01 -1.10449739e-01
3.33229423e-01 -3.67906958e-01 -5.40611148e-01 -1.52610290e+00
-2.97009293e-02 -8.80974233e-01 4.91440982e-01 -7.95929849e-01
-1.46561885e+00 6.58693552e-01 -2.81501144e-01 -1.24201238e+00
-9.10577029e-02 -5.74634254e-01 -8.17654252e-01 9.67764080e-01
-1.69639528e+00 -1.01832390e+00 -1.61752656e-01 6.95847631e-01
1.19474053e+00 -2.66869128e-01 8.94899070e-01 3.92725617e-01
-9.67963278e-01 7.87734807e-01 3.94663066e-01 7.46893644e-01
8.43220294e-01 -9.64325011e-01 2.87410200e-01 9.60290372e-01
8.61417279e-02 8.99569392e-01 6.06810078e-02 -6.60861433e-01
-1.14602780e+00 -1.02233362e+00 1.27070558e+00 -3.33685786e-01
7.46470451e-01 -2.65406996e-01 -1.19088602e+00 6.08505189e-01
5.47004640e-01 -3.29949617e-01 8.32213581e-01 3.67711663e-01
-8.55665952e-02 -1.42321944e-01 -4.78107274e-01 5.23006141e-01
6.82118654e-01 -6.68986738e-01 -8.92845273e-01 6.13518715e-01
1.16382062e+00 -2.72849590e-01 -6.84592605e-01 2.95406103e-01
-1.31921157e-01 -6.09403431e-01 7.39958048e-01 -8.53174746e-01
7.30090857e-01 7.44034885e-04 2.01772675e-01 -1.16635334e+00
-3.74378115e-01 -3.87205660e-01 1.24734238e-01 1.45063174e+00
7.51028001e-01 -2.90111542e-01 6.54908299e-01 3.20775777e-01
-4.51555580e-01 -8.84766877e-01 -3.74734312e-01 -6.41932487e-01
5.42852283e-01 -4.63827819e-01 4.71351892e-01 9.83350456e-01
1.87161386e-01 8.74782264e-01 9.14320052e-02 -4.62285802e-02
3.82565595e-02 1.54239312e-01 5.39508045e-01 -7.48387575e-01
-1.00517318e-01 -4.52281386e-01 2.21672848e-01 -1.69517648e+00
3.16267282e-01 -1.22712612e+00 6.27723709e-02 -1.58995509e+00
1.51469842e-01 -4.68707055e-01 -4.40848649e-01 2.46822551e-01
-8.18962932e-01 -1.66307688e-01 1.18399493e-01 1.29919872e-01
-8.92261386e-01 6.30430222e-01 1.66906917e+00 -4.55144271e-02
-2.84880787e-01 3.61697197e-01 -9.66161549e-01 4.20468897e-01
1.24333632e+00 -4.32802409e-01 -4.46886003e-01 -1.10552597e+00
1.28437161e-01 -6.66566268e-02 -2.68527955e-01 -7.13387370e-01
3.55408967e-01 8.27100724e-02 3.23498607e-01 -7.07747698e-01
-3.93908247e-02 -2.22887933e-01 -1.03084278e+00 2.52356291e-01
-9.06549335e-01 1.99066013e-01 1.59665614e-01 1.96017727e-01
-4.56350803e-01 -8.46221447e-01 3.84415895e-01 -3.94439697e-01
-8.60814989e-01 1.91230774e-01 -3.05181175e-01 4.98551488e-01
5.39156139e-01 1.37140393e-01 -3.31504703e-01 -5.76821268e-01
-4.76196289e-01 6.86590135e-01 -2.46576786e-01 7.61368632e-01
7.38288164e-01 -1.11798728e+00 -1.04802787e+00 -2.61487123e-02
2.30981484e-01 7.99217820e-02 4.99746174e-01 7.95360446e-01
-6.66684687e-01 7.34732449e-01 5.18916398e-02 -2.91817516e-01
-9.50253129e-01 5.82442343e-01 -3.42617668e-02 -4.35903549e-01
-3.45487714e-01 1.35594845e+00 1.72534175e-02 -6.61598086e-01
2.18503639e-01 -3.10047954e-01 -6.82036996e-01 6.75019026e-02
6.48696899e-01 2.84374326e-01 2.07662974e-02 -5.26180983e-01
-5.18040545e-02 3.03233892e-01 -4.59928364e-01 1.32710516e-01
1.17999053e+00 -4.37465250e-01 -2.50964969e-01 3.38076711e-01
1.67304289e+00 -4.03250121e-02 -7.79348552e-01 -5.87836325e-01
3.73773545e-01 -2.19444394e-01 1.12614054e-02 -6.52755380e-01
-9.84957278e-01 1.26243937e+00 9.69542712e-02 -1.15620732e-01
1.12563825e+00 3.83049808e-02 1.13987792e+00 7.84265280e-01
-1.17636964e-01 -1.37719047e+00 3.91997218e-01 1.04378951e+00
9.59773898e-01 -1.39778769e+00 -2.49218792e-01 -2.74456888e-01
-9.39291358e-01 1.18482733e+00 1.01454103e+00 -2.53582507e-01
5.45899808e-01 -3.16736102e-01 3.39480788e-01 1.45610780e-01
-4.56660509e-01 -4.56140250e-01 4.26316470e-01 3.16974282e-01
9.77182031e-01 -8.72484893e-02 -7.74224699e-01 9.70283389e-01
-3.99230778e-01 -1.73261568e-01 3.30230623e-01 8.58330488e-01
-4.57951546e-01 -1.32761216e+00 -7.77357221e-02 5.43039501e-01
-4.68246430e-01 -6.17037475e-01 -1.54806674e-01 3.71840745e-01
-1.31420359e-01 1.34706819e+00 7.15990812e-02 -3.45719308e-01
1.67448461e-01 1.38390973e-01 1.10850140e-01 -1.10885251e+00
-8.45529139e-01 1.42057896e-01 1.46859944e-01 -2.72825330e-01
-3.64797711e-01 -3.48868638e-01 -1.33767438e+00 2.11072415e-01
-5.59015453e-01 4.63999808e-01 3.31829935e-01 1.02532148e+00
2.53443390e-01 4.97417867e-01 7.15723693e-01 -2.92532086e-01
-5.76654136e-01 -1.52474165e+00 -1.98106050e-01 4.39941645e-01
-3.73558812e-02 -7.77191967e-02 -1.22394785e-01 6.48765117e-02] | [10.824393272399902, 9.068069458007812] |
cb37fc90-b08c-4972-9ffc-4d781bc7cbd9 | boosted-prompt-ensembles-for-large-language | 2304.05970 | null | https://arxiv.org/abs/2304.05970v1 | https://arxiv.org/pdf/2304.05970v1.pdf | Boosted Prompt Ensembles for Large Language Models | Methods such as chain-of-thought prompting and self-consistency have pushed the frontier of language model reasoning performance with no additional training. To further improve performance, we propose a prompt ensembling method for large language models, which uses a small dataset to construct a set of few shot prompts that together comprise a ``boosted prompt ensemble''. The few shot examples for each prompt are chosen in a stepwise fashion to be ``hard'' examples on which the previous step's ensemble is uncertain. We show that this outperforms single-prompt output-space ensembles and bagged prompt-space ensembles on the GSM8k and AQuA datasets, among others. We propose both train-time and test-time versions of boosted prompting that use different levels of available annotation and conduct a detailed empirical study of our algorithm. | ['Jimmy Ba', 'Andrew Wang', 'Michael R. Zhang', 'Silviu Pitis'] | 2023-04-12 | null | null | null | null | ['gsm8k'] | ['natural-language-processing'] | [ 1.13930799e-01 2.60506600e-01 6.09264597e-02 -7.38274872e-01
-1.28315699e+00 -7.02064812e-01 9.39201951e-01 4.64517474e-01
-6.47733569e-01 7.62063265e-01 4.01674271e-01 -6.01181924e-01
-1.14039367e-03 -5.25723696e-01 -4.81944829e-01 -5.50068438e-01
1.27865925e-01 7.55841970e-01 3.35010052e-01 -3.56988013e-01
-3.82602327e-02 -2.21834466e-01 -1.25482857e+00 3.79150778e-01
7.29211152e-01 6.73941612e-01 1.16934612e-01 9.51679885e-01
-3.81229162e-01 1.13767028e+00 -4.55302238e-01 -4.27995741e-01
-1.94634367e-02 -6.13008916e-01 -8.94191802e-01 -2.01197311e-01
2.65105873e-01 -4.75252122e-01 -2.72084743e-01 7.12652802e-01
7.78621197e-01 6.19263172e-01 4.41645205e-01 -9.93331850e-01
-6.57033861e-01 1.13136578e+00 1.37171689e-02 4.50315684e-01
2.52422333e-01 3.74774903e-01 1.07172883e+00 -1.19233489e+00
3.68818283e-01 1.04833686e+00 7.28197694e-01 9.03689742e-01
-1.58227003e+00 -6.67457759e-01 4.10268068e-01 1.25808656e-01
-9.79232967e-01 -6.23787940e-01 5.13158739e-01 -4.49569821e-01
1.30151904e+00 2.16785803e-01 5.05581856e-01 1.29191327e+00
-1.67612374e-01 8.10196996e-01 1.33266270e+00 -6.21192157e-01
9.27105129e-01 9.12066028e-02 7.12168455e-01 5.79387128e-01
3.74965160e-03 3.69936883e-01 -9.45790768e-01 -6.49395168e-01
2.82169133e-01 -2.88992524e-01 -6.92970976e-02 2.64327675e-01
-1.18129957e+00 9.77838695e-01 6.64767176e-02 7.24670961e-02
-5.60589552e-01 1.37339354e-01 2.46311054e-01 2.32438296e-01
8.69988620e-01 6.93699896e-01 -6.75438941e-01 -4.25623536e-01
-1.07779181e+00 5.54199755e-01 9.87950385e-01 7.82004118e-01
7.44198799e-01 -1.60721075e-02 -7.39940286e-01 8.95677090e-01
1.78935185e-01 2.58452684e-01 5.08965075e-01 -8.67619038e-01
1.90417171e-01 3.54068398e-01 2.83152074e-01 -3.01198721e-01
-4.34761077e-01 -2.81808496e-01 -4.92568702e-01 -2.18679279e-01
3.67133558e-01 -5.61336100e-01 -1.13331497e+00 1.90290451e+00
3.55500102e-01 5.04974723e-01 9.95523632e-02 7.12304294e-01
7.27224052e-01 7.35163867e-01 6.88730359e-01 -2.93768942e-01
1.29021645e+00 -1.08755457e+00 -5.28794646e-01 -4.71655697e-01
7.66998112e-01 -4.71259058e-01 1.11546385e+00 3.82723719e-01
-9.46530700e-01 -5.40754616e-01 -6.73048496e-01 -4.29435074e-02
-1.25132263e-01 -7.12166950e-02 9.04562414e-01 3.26116920e-01
-1.17882347e+00 5.54219007e-01 -8.38620722e-01 -3.59744728e-01
1.45889387e-01 -8.34194645e-02 8.26235861e-03 1.50425732e-01
-1.39243960e+00 9.83221591e-01 5.35160363e-01 -4.40098733e-01
-1.28032482e+00 -7.89715350e-01 -6.37935996e-01 1.11780912e-01
5.40585637e-01 -6.43706143e-01 2.06317282e+00 -6.48544967e-01
-1.37204444e+00 6.84841752e-01 -2.29744807e-01 -6.62485600e-01
3.03540736e-01 -4.53451157e-01 -2.36788988e-01 -1.71470046e-01
-6.81677610e-02 9.07994330e-01 5.15378177e-01 -9.81530428e-01
-7.46410549e-01 -8.07145759e-02 7.69922286e-02 1.39036834e-01
-2.13519424e-01 3.44650894e-01 -1.36244088e-01 -4.98174518e-01
-7.21085891e-02 -8.56081963e-01 -6.43614829e-01 -4.09995407e-01
-3.52976888e-01 -6.71193063e-01 1.03108890e-01 -5.81051409e-01
1.45640218e+00 -1.73746920e+00 4.38522473e-02 -9.62112620e-02
2.09415242e-01 1.36284783e-01 -3.42512995e-01 5.01037180e-01
1.33775815e-01 1.42430708e-01 3.81070958e-03 -5.28211832e-01
9.50886235e-02 3.79984766e-01 -7.09555745e-01 -1.20040379e-01
2.70204008e-01 8.09268534e-01 -1.34142661e+00 -5.79086900e-01
-6.10769056e-02 1.33966608e-02 -8.17432046e-01 6.76285386e-01
-7.88358748e-01 4.40082312e-01 -4.00810510e-01 3.46434802e-01
1.12582400e-01 -4.72203344e-01 4.48255539e-02 3.21555078e-01
-1.37283625e-02 7.14896858e-01 -8.12551737e-01 1.86310065e+00
-3.34373027e-01 1.78308636e-01 -2.73000717e-01 -5.34264028e-01
1.02826762e+00 6.20318174e-01 1.95181683e-01 -4.36358035e-01
4.31638844e-02 9.81849283e-02 1.49654308e-02 -4.02033269e-01
7.71690130e-01 -5.31255782e-01 -3.45084399e-01 8.98687243e-01
6.04162455e-01 -2.08581015e-01 2.36203820e-01 8.27642679e-01
1.42294371e+00 1.56149834e-01 4.12560344e-01 -1.44383729e-01
7.53895268e-02 1.47200078e-01 5.55782139e-01 1.20933366e+00
-2.45113835e-01 3.91113073e-01 2.97441274e-01 -2.71562487e-01
-1.07057095e+00 -9.33121860e-01 1.56194165e-01 1.83647990e+00
-3.49124610e-01 -7.12411344e-01 -6.70023143e-01 -8.11525643e-01
-2.36059772e-03 1.51115263e+00 -5.96447349e-01 -1.41920164e-01
-2.50387549e-01 -6.60136998e-01 6.28125489e-01 8.85773301e-01
-8.06857571e-02 -1.09502602e+00 -7.27945089e-01 5.37733734e-01
-3.61345977e-01 -7.34154105e-01 -2.65571505e-01 1.09509754e+00
-6.59850359e-01 -6.80243790e-01 -5.57117939e-01 -3.95738035e-01
4.55600232e-01 -1.53235346e-01 1.50260985e+00 5.38122058e-02
2.14206517e-01 4.59668607e-01 -5.96729994e-01 -5.59532702e-01
-7.86513686e-01 1.09468825e-01 3.96404684e-01 -3.33144128e-01
6.56114697e-01 -4.49951887e-01 -3.14261824e-01 -5.22729754e-02
-5.17242014e-01 4.31603014e-01 2.61534840e-01 1.27423322e+00
5.17172754e-01 -5.62051535e-01 8.07027161e-01 -9.23500359e-01
9.43383396e-01 -6.97117269e-01 -3.89120311e-01 3.55131000e-01
-6.86598063e-01 2.72985309e-01 5.96347392e-01 -6.99110448e-01
-1.24938881e+00 -8.35034102e-02 -2.39846587e-01 -1.77180544e-01
-3.60764891e-01 5.87581813e-01 3.66957456e-01 3.94177884e-01
1.10873425e+00 2.73889184e-01 -3.96044821e-01 -7.58149147e-01
6.87009096e-01 7.39861369e-01 7.24748552e-01 -1.00068069e+00
5.39171696e-01 -8.67717713e-02 -6.14013135e-01 -2.38509908e-01
-1.33377016e+00 -5.77286363e-01 -2.70991355e-01 -2.98192084e-01
6.50919974e-01 -1.02089834e+00 -3.06980103e-01 2.65134096e-01
-1.28788555e+00 -7.63461709e-01 -5.50075054e-01 3.33356589e-01
-4.95194376e-01 -6.34023873e-03 -7.38994658e-01 -1.23292553e+00
-5.89085102e-01 -6.65622354e-01 1.01442981e+00 4.17666197e-01
-7.55203664e-01 -8.41711223e-01 6.25258625e-01 -1.36594534e-01
4.36744720e-01 -4.88779873e-01 8.03631544e-01 -1.63334608e+00
-2.12778509e-01 -4.87624332e-02 2.45538592e-01 3.46422523e-01
-2.03463241e-01 -2.94976294e-01 -1.22873223e+00 -2.00803608e-01
-4.44608890e-02 -8.72653723e-01 8.09533119e-01 4.96911444e-02
7.40770817e-01 -3.86082739e-01 -3.31352055e-01 1.94751635e-01
1.07437587e+00 1.24161303e-01 1.44400686e-01 1.81123182e-01
1.83259517e-01 3.81074876e-01 7.51234114e-01 5.91207743e-01
5.44169724e-01 4.74006444e-01 -1.01107746e-01 9.16293859e-02
1.68496475e-01 -6.39809668e-01 4.05833632e-01 1.15679562e+00
1.45350546e-01 -3.46695632e-01 -1.21610940e+00 6.37275457e-01
-1.85709560e+00 -9.66873109e-01 2.18059808e-01 2.01633310e+00
1.36296606e+00 1.26717746e-01 4.58821319e-02 -7.09867626e-02
3.28523964e-01 1.20661706e-01 -4.11371022e-01 -4.37417358e-01
-3.36683840e-02 4.08087671e-01 6.47194311e-02 6.35834694e-01
-9.40567613e-01 1.13953829e+00 6.86724949e+00 7.46563852e-01
-9.24100637e-01 4.34633136e-01 7.43432641e-01 -3.04278314e-01
-6.63244188e-01 4.54421461e-01 -1.05617738e+00 5.55581987e-01
1.50544250e+00 -3.88583541e-01 4.27309215e-01 9.44625616e-01
-3.74610350e-02 -4.03293192e-01 -1.37122369e+00 4.96401370e-01
-1.86387692e-02 -1.27985871e+00 -2.10910782e-01 -4.84652728e-01
7.73666918e-01 3.51753443e-01 -4.59843695e-01 9.88296270e-01
1.16773152e+00 -7.19539881e-01 8.72161508e-01 5.92015505e-01
6.12552345e-01 -4.13240492e-01 4.15110976e-01 9.82715249e-01
-6.50256932e-01 -1.84977889e-01 -3.22041929e-01 -2.02018335e-01
4.76039857e-01 4.90464956e-01 -1.20598292e+00 3.30084264e-01
4.93388653e-01 2.98064891e-02 -6.71247900e-01 8.44169199e-01
-4.07623827e-01 1.25059187e+00 -3.89274716e-01 -3.31502259e-01
2.65234828e-01 2.45067969e-01 5.26621163e-01 1.37075651e+00
2.40652472e-01 6.52246416e-01 5.26238978e-01 8.71328950e-01
1.63428277e-01 -3.97835337e-02 -3.00052881e-01 -1.43372074e-01
8.93676758e-01 1.28323734e+00 -5.09414434e-01 -9.90035594e-01
-2.95864403e-01 6.73260987e-01 5.85386097e-01 3.31706792e-01
-6.35243714e-01 9.95040610e-02 2.46782854e-01 -2.51128048e-01
1.11886384e-02 -3.37724574e-02 -2.61858016e-01 -1.18859398e+00
-5.61691105e-01 -9.05644178e-01 6.29012048e-01 -1.03180003e+00
-1.42506468e+00 6.77544951e-01 1.23314463e-01 -6.59590364e-01
-7.64461756e-01 -3.03480893e-01 -7.88596869e-01 1.01470661e+00
-1.07564974e+00 -7.61617541e-01 -2.36756533e-01 2.21956626e-01
7.39640415e-01 -1.16098501e-01 1.25588739e+00 -9.43616480e-02
-2.08152235e-01 3.94270688e-01 -1.52760357e-01 -6.26211092e-02
7.73653984e-01 -1.48156595e+00 8.12842011e-01 8.07797194e-01
3.66733998e-01 6.87738478e-01 1.04205239e+00 -7.98343837e-01
-9.34087932e-01 -1.00328958e+00 1.26427329e+00 -8.79241765e-01
6.91878557e-01 -4.37172681e-01 -1.02848995e+00 9.81613696e-01
1.65961102e-01 -1.43305168e-01 8.92266333e-01 5.15956759e-01
-3.73539180e-01 1.47489443e-01 -6.94864213e-01 5.23560703e-01
9.16868687e-01 -6.31491780e-01 -1.16580629e+00 3.70827019e-01
9.42145646e-01 -4.67553109e-01 -6.66415095e-01 2.61527926e-01
2.89184749e-01 -7.46859074e-01 6.04535580e-01 -1.08596480e+00
2.31848493e-01 1.40151590e-01 -1.86823070e-01 -1.63644922e+00
-5.32891810e-01 -7.21798360e-01 -3.83540809e-01 1.32609117e+00
4.45307046e-01 -4.16290641e-01 4.40728664e-01 8.15899968e-01
-2.77931780e-01 -9.67224061e-01 -7.63075531e-01 -7.46634245e-01
2.29730621e-01 -6.70134246e-01 6.92571342e-01 6.97733700e-01
4.23417479e-01 6.59769833e-01 -3.49695385e-01 -2.46946454e-01
3.84872556e-01 -1.39169797e-01 6.88280761e-01 -1.24082649e+00
-6.34426236e-01 -6.74516186e-02 3.17556858e-01 -1.13921332e+00
1.36167243e-01 -1.07945132e+00 6.36970520e-01 -1.31092882e+00
5.18666565e-01 -6.84457719e-01 -6.58387840e-01 9.59990382e-01
-7.76762486e-01 -2.65577108e-01 9.25401077e-02 3.28575745e-02
-1.00779212e+00 4.27768052e-01 6.42482936e-01 1.53682038e-01
-1.48028299e-01 -2.83027411e-01 -8.15598071e-01 5.97361207e-01
4.94610846e-01 -5.97572744e-01 -3.23195934e-01 -2.89301842e-01
2.95992076e-01 2.47953981e-01 3.94807011e-01 -7.56828547e-01
4.30822372e-01 -2.31925428e-01 2.60114074e-01 -5.27927756e-01
3.79085869e-01 -2.59102196e-01 -2.61146133e-03 2.35972136e-01
-1.03707576e+00 2.93423589e-02 2.15217650e-01 6.29175901e-01
2.06189826e-01 -1.89981937e-01 6.01660848e-01 -2.43209571e-01
-8.37960780e-01 1.40610501e-01 -4.25557196e-01 3.00540477e-01
8.01485479e-01 3.56652558e-01 -4.30498004e-01 -4.74652559e-01
-7.81127989e-01 3.81426364e-01 3.32963556e-01 3.82724613e-01
3.25735003e-01 -1.22734082e+00 -9.42186475e-01 -8.86040926e-03
2.99601674e-01 3.97591516e-02 2.53503621e-01 8.45919609e-01
2.72588551e-01 5.45428872e-01 2.15106979e-01 -6.02110505e-01
-9.76241291e-01 4.98443335e-01 1.59469008e-01 -4.64028567e-01
-5.86920917e-01 1.42491627e+00 -1.44645959e-01 -5.18130839e-01
2.12261200e-01 -2.17845023e-01 3.20060402e-02 9.55137163e-02
6.67928636e-01 1.39919922e-01 1.02699064e-02 -1.70761302e-01
-3.92050713e-01 -2.98831552e-01 -2.00693905e-01 -6.49787605e-01
1.24571872e+00 9.26880389e-02 1.52999133e-01 9.43059742e-01
4.73185837e-01 -1.91005439e-01 -1.35471630e+00 -5.90596378e-01
3.43311697e-01 -2.30179340e-01 -3.30684893e-03 -1.40522921e+00
-1.60946429e-01 8.00150454e-01 2.71331787e-01 2.15808854e-01
9.06667173e-01 3.18776459e-01 5.52418411e-01 4.81053591e-01
5.65500259e-01 -1.22175300e+00 1.42681435e-01 7.22587585e-01
6.89721644e-01 -1.09327257e+00 -3.53747129e-01 3.00231367e-01
-8.10095251e-01 7.40453959e-01 8.72639537e-01 1.48471147e-01
2.51126081e-01 5.65640450e-01 2.64884025e-01 -1.13670148e-01
-1.75070846e+00 -1.37710944e-01 2.17931852e-01 3.20665091e-01
6.42637312e-01 1.90351740e-01 -2.66187847e-01 1.37448490e+00
-2.47159407e-01 3.57791901e-01 1.90416127e-01 8.02756786e-01
-7.52719104e-01 -8.80141616e-01 -2.89350063e-01 6.18372798e-01
-2.11277097e-01 -5.39393306e-01 -3.35134029e-01 2.08887041e-01
1.10792860e-01 1.08172512e+00 1.49281785e-01 -6.03780329e-01
4.81396988e-02 8.76867890e-01 3.20188880e-01 -1.01258767e+00
-7.52758086e-01 -6.73262402e-02 5.08380890e-01 -3.66496801e-01
1.18749484e-01 -5.46786308e-01 -1.26268339e+00 -1.85886323e-01
-5.39642930e-01 5.05313873e-01 3.74967933e-01 1.17700672e+00
2.97667265e-01 5.51128387e-01 3.07313472e-01 -7.16635287e-01
-1.32479525e+00 -1.42288435e+00 -2.73907810e-01 4.01066303e-01
9.81609970e-02 -5.25504470e-01 -4.22754258e-01 -3.45358363e-04] | [10.860560417175293, 8.227086067199707] |
81cab239-befe-4d2f-88fc-e78eb0ba2d53 | efficient-meshy-neural-fields-for-animatable | 2303.12965 | null | https://arxiv.org/abs/2303.12965v1 | https://arxiv.org/pdf/2303.12965v1.pdf | Efficient Meshy Neural Fields for Animatable Human Avatars | Efficiently digitizing high-fidelity animatable human avatars from videos is a challenging and active research topic. Recent volume rendering-based neural representations open a new way for human digitization with their friendly usability and photo-realistic reconstruction quality. However, they are inefficient for long optimization times and slow inference speed; their implicit nature results in entangled geometry, materials, and dynamics of humans, which are hard to edit afterward. Such drawbacks prevent their direct applicability to downstream applications, especially the prominent rasterization-based graphic ones. We present EMA, a method that Efficiently learns Meshy neural fields to reconstruct animatable human Avatars. It jointly optimizes explicit triangular canonical mesh, spatial-varying material, and motion dynamics, via inverse rendering in an end-to-end fashion. Each above component is derived from separate neural fields, relaxing the requirement of a template, or rigging. The mesh representation is highly compatible with the efficient rasterization-based renderer, thus our method only takes about an hour of training and can render in real-time. Moreover, only minutes of optimization is enough for plausible reconstruction results. The disentanglement of meshes enables direct downstream applications. Extensive experiments illustrate the very competitive performance and significant speed boost against previous methods. We also showcase applications including novel pose synthesis, material editing, and relighting. The project page: https://xk-huang.github.io/ema/. | ['Jiwen Lu', 'Jie zhou', 'Xiu Li', 'Yansong Tang', 'Yiji Cheng', 'Xiaoke Huang'] | 2023-03-23 | null | null | null | null | ['inverse-rendering'] | ['computer-vision'] | [ 5.96995763e-02 1.66516885e-01 2.23971233e-01 -6.66788965e-02
-6.06758118e-01 -5.17493367e-01 6.31969392e-01 -4.56246942e-01
-1.34964362e-01 8.37200463e-01 -1.33899599e-01 -2.46538818e-01
1.57937720e-01 -1.13919568e+00 -1.13147175e+00 -5.68592787e-01
2.09839076e-01 7.26141632e-01 -1.80855347e-03 -5.03180981e-01
-9.14535224e-02 7.00912118e-01 -1.51041627e+00 1.61174294e-02
8.98806870e-01 9.78525102e-01 6.96112216e-02 7.28637874e-01
-1.00731682e-02 5.22119999e-01 -3.72030824e-01 -6.55802429e-01
4.27679390e-01 -4.59868878e-01 -5.83961129e-01 -5.18009216e-02
8.56990218e-01 -5.77033162e-01 -5.15243471e-01 5.94492912e-01
6.53273046e-01 3.56854051e-01 5.32102048e-01 -1.05852568e+00
-7.68090427e-01 2.36147106e-01 -5.29358864e-01 -4.41030264e-01
3.13695252e-01 4.79968965e-01 7.64320552e-01 -8.13126922e-01
9.02091801e-01 1.33645189e+00 1.05513370e+00 6.03516042e-01
-1.33222830e+00 -5.86204529e-01 -6.43914863e-02 -2.90803295e-02
-1.37072122e+00 -3.39767963e-01 8.01116645e-01 -3.80440593e-01
6.41471386e-01 7.03307509e-01 1.27458441e+00 1.17937386e+00
1.46480054e-01 6.60637558e-01 8.54834974e-01 2.23986283e-02
1.54570654e-01 -3.97670299e-01 -5.36027491e-01 9.80128050e-01
-1.39094904e-01 2.44059607e-01 -3.89721960e-01 -6.74115717e-02
1.70023191e+00 -1.47404626e-01 -4.13648754e-01 -4.50123787e-01
-1.40253758e+00 7.61223376e-01 5.53975224e-01 -2.40765169e-01
-4.08291221e-01 7.08147168e-01 3.67913902e-01 1.73262924e-01
5.47890484e-01 4.93403465e-01 -8.11332539e-02 -2.10297227e-01
-9.59300935e-01 8.51322293e-01 6.42313123e-01 1.04406929e+00
4.86216009e-01 3.13449651e-01 -1.06965683e-01 5.76170266e-01
5.33292908e-03 5.87969005e-01 -1.52714834e-01 -1.40643573e+00
3.65313917e-01 1.45820826e-01 9.84853208e-02 -1.36784506e+00
-3.73008221e-01 -2.53127277e-01 -1.28781807e+00 5.51102042e-01
5.28659880e-01 -2.06883950e-03 -6.68041110e-01 1.54391146e+00
7.23922610e-01 3.23111832e-01 -5.29979169e-01 1.10336113e+00
8.79655242e-01 9.11481559e-01 7.92800449e-03 -5.61534502e-02
1.35040855e+00 -1.06335723e+00 -7.02445388e-01 2.18851000e-01
2.39184886e-01 -7.03362465e-01 1.25787485e+00 5.41416943e-01
-1.54836404e+00 -3.99772555e-01 -9.05490220e-01 -7.31739521e-01
1.33909613e-01 -5.18962704e-02 1.02770662e+00 2.28413790e-01
-1.06204438e+00 1.12478793e+00 -8.67025256e-01 2.17153013e-01
6.08657062e-01 3.05391967e-01 -2.76448041e-01 2.60588497e-01
-1.08915257e+00 6.42359912e-01 -1.40568152e-01 3.07098478e-01
-6.02752745e-01 -1.05267930e+00 -7.74385989e-01 -4.89550009e-02
3.51666868e-01 -1.32488537e+00 1.15702772e+00 -8.39218557e-01
-2.14441228e+00 7.53945887e-01 1.85949117e-01 -1.35889068e-01
1.34377480e+00 -1.92412347e-01 1.41477674e-01 8.25124457e-02
-1.53747737e-01 7.01506734e-01 9.09413338e-01 -1.19163752e+00
-1.12027779e-01 -7.34667033e-02 7.08609298e-02 3.32003653e-01
6.11896180e-02 -4.42320198e-01 -6.61429048e-01 -1.07251191e+00
-7.74650043e-03 -8.69180262e-01 -3.40848714e-01 8.30332220e-01
-3.20418000e-01 8.25809408e-03 6.15563452e-01 -1.01100159e+00
1.01178694e+00 -1.82706702e+00 5.27307451e-01 1.09987266e-01
4.59796488e-01 -1.58091038e-02 3.45925521e-03 1.93115517e-01
1.17795773e-01 6.63806275e-02 -4.04702634e-01 -5.01219988e-01
1.64800569e-01 1.13065146e-01 -3.39953125e-01 6.11849785e-01
-1.74200162e-02 1.19563735e+00 -8.24089527e-01 -4.96983379e-01
2.86259800e-01 1.05019534e+00 -8.07565212e-01 2.28632063e-01
-3.95009220e-01 9.02553976e-01 -3.21586788e-01 5.68061233e-01
7.69860089e-01 -1.26645491e-01 1.66250557e-01 -4.88537014e-01
-1.76926732e-01 2.20578223e-01 -1.24074817e+00 2.02293873e+00
-7.79049277e-01 6.35457277e-01 3.09785217e-01 -6.49380147e-01
7.91371763e-01 2.06563368e-01 3.41877431e-01 -8.26250494e-01
2.64294028e-01 4.32059728e-02 -5.76253593e-01 -3.28879297e-01
6.50427401e-01 -1.58720940e-01 -5.50837666e-02 4.27488953e-01
-3.19991201e-01 -5.52837193e-01 -1.35194123e-01 -6.51085190e-03
6.44836009e-01 8.65839183e-01 3.57779488e-02 -2.03052253e-01
-7.65013462e-03 -6.99203983e-02 5.05656183e-01 2.09696919e-01
4.38630342e-01 8.88299108e-01 4.25212294e-01 -8.26904833e-01
-1.51288903e+00 -1.19046354e+00 -1.01779595e-01 8.98703754e-01
4.20584410e-01 -4.78338391e-01 -8.36327374e-01 -1.05602235e-01
-1.20533928e-01 4.81186569e-01 -6.46658659e-01 1.21063605e-01
-1.30046606e+00 -3.52836788e-01 5.40331423e-01 4.73050565e-01
2.95268148e-01 -1.11701429e+00 -6.89578593e-01 2.44954199e-01
-2.85045058e-01 -7.03211010e-01 -4.50230598e-01 -5.33985138e-01
-1.08091736e+00 -7.18357980e-01 -9.57312047e-01 -6.63587034e-01
6.18714273e-01 -2.24378286e-03 1.38487434e+00 3.04453641e-01
-2.91468352e-01 -1.00510381e-01 2.63511688e-01 1.09464832e-01
-4.23187673e-01 -8.17258433e-02 -1.49764314e-01 -3.13367128e-01
-9.24207091e-01 -1.06243706e+00 -8.49401474e-01 4.04837221e-01
-7.55535722e-01 8.32935989e-01 1.51490137e-01 8.18446696e-01
9.98880625e-01 -3.68339717e-01 4.64271270e-02 -8.14397395e-01
4.26964283e-01 -1.20397650e-01 -7.15757191e-01 6.24782220e-02
-1.23554088e-01 8.62804428e-03 7.81408370e-01 -6.92421794e-01
-9.81808662e-01 1.06763206e-01 -4.34813380e-01 -6.53168321e-01
1.13140531e-01 5.51038943e-02 -2.31241137e-01 -1.87808022e-01
6.71224475e-01 -1.10173389e-01 7.40311593e-02 -5.87113321e-01
6.93510056e-01 3.44749875e-02 8.50997746e-01 -9.06660795e-01
8.59274626e-01 5.28153718e-01 1.24214299e-01 -8.29417408e-01
-2.00485766e-01 3.31835598e-01 -4.75788355e-01 -5.17312229e-01
6.85745776e-01 -8.66345942e-01 -1.21339905e+00 5.20267010e-01
-1.34946382e+00 -7.99774706e-01 -4.73284364e-01 1.93505719e-01
-9.35570121e-01 3.62474442e-01 -8.51580083e-01 -4.35283720e-01
-5.62404931e-01 -1.26700580e+00 1.27769828e+00 2.91784434e-03
-4.21900004e-01 -7.19735920e-01 -4.99814488e-02 4.66949731e-01
3.35140049e-01 9.01953697e-01 6.91105962e-01 6.34933949e-01
-1.01858592e+00 5.90663739e-02 -9.83694792e-02 -2.08999068e-01
-2.10740864e-01 2.96109527e-01 -8.14350486e-01 -2.18554646e-01
-3.10025662e-01 -1.49237603e-01 5.20419419e-01 2.65117049e-01
1.59898865e+00 -6.72584474e-01 -1.10983647e-01 1.26470566e+00
1.21701682e+00 -2.51152277e-01 8.51301253e-01 1.12358131e-01
1.36900747e+00 4.11597520e-01 3.54950964e-01 6.36488795e-01
4.28731650e-01 1.04568756e+00 4.17750478e-01 -3.81717831e-01
-3.41106057e-01 -4.48966593e-01 1.19988747e-01 9.93458450e-01
-8.62262249e-01 -1.14286467e-01 -6.34084404e-01 1.40392065e-01
-1.74786377e+00 -9.22009885e-01 -4.73386258e-01 2.17760873e+00
9.91386771e-01 -2.80646235e-01 3.29597145e-01 -7.41149932e-02
5.41691720e-01 1.82157010e-01 -5.51603854e-01 -4.27792549e-01
1.11091204e-01 5.39663196e-01 3.15805286e-01 6.14158809e-01
-7.51833320e-01 1.02977061e+00 5.51502371e+00 1.07949984e+00
-1.14877558e+00 1.44361496e-01 6.20192826e-01 -2.77833611e-01
-7.81750023e-01 -1.28355622e-01 -2.41389289e-01 2.61478812e-01
5.30090928e-01 -5.62429102e-03 8.02223206e-01 7.99637735e-01
2.49687403e-01 2.18382969e-01 -9.46231008e-01 1.26567066e+00
-1.93122596e-01 -1.93148673e+00 2.17325151e-01 1.16121091e-01
6.80025637e-01 -4.00439322e-01 5.57634421e-03 3.80177796e-02
1.07812874e-01 -1.21784079e+00 1.31847751e+00 7.68415511e-01
1.33076727e+00 -9.64351654e-01 1.98009443e-02 2.26503655e-01
-1.17480266e+00 5.22797585e-01 -2.67814964e-01 -1.08363576e-01
5.68513513e-01 3.99797380e-01 -2.11539358e-01 4.23008531e-01
5.59847772e-01 4.98578370e-01 -1.85737014e-01 7.52540052e-01
-2.08304793e-01 2.57166713e-01 -4.30194676e-01 7.60231242e-02
-4.55119461e-02 -5.62306881e-01 5.99387407e-01 8.84693265e-01
3.68680447e-01 2.83054590e-01 -4.72962223e-02 1.22105885e+00
-2.75500059e-01 1.33187652e-01 -4.39458162e-01 1.91405848e-01
2.40182132e-01 1.13737643e+00 -7.89045870e-01 -2.97865421e-01
1.52350396e-01 1.18058228e+00 4.62476432e-01 2.42454201e-01
-1.33865166e+00 -1.81477532e-01 6.59131825e-01 4.76943761e-01
1.03414416e-01 -5.38245320e-01 -6.83762550e-01 -1.11032295e+00
2.36070365e-01 -8.80241215e-01 -1.98795557e-01 -7.15440869e-01
-1.07922828e+00 7.00906157e-01 -1.37833118e-01 -1.26601386e+00
-1.18972495e-01 -2.96785593e-01 -5.09878635e-01 7.53067553e-01
-9.30447757e-01 -1.30189288e+00 -5.20198703e-01 5.21027565e-01
4.39683914e-01 3.96991313e-01 7.71522582e-01 4.43819255e-01
-4.13449466e-01 7.87078321e-01 -9.65209305e-03 -7.49220550e-02
3.72608215e-01 -9.99088168e-01 9.85205054e-01 5.36249280e-01
-3.03989816e-02 3.20068508e-01 7.33660638e-01 -6.66340709e-01
-1.70110726e+00 -9.71330345e-01 5.05635262e-01 -3.33303630e-01
2.38088563e-01 -5.80391526e-01 -1.00813580e+00 4.71430004e-01
-5.88548519e-02 -2.83534732e-02 2.14587823e-01 -1.37475267e-01
-2.02888742e-01 3.78457680e-02 -1.01520181e+00 1.05383563e+00
1.56684184e+00 -2.97646403e-01 -5.22785215e-03 4.70247716e-01
7.98618078e-01 -1.02298629e+00 -1.05868483e+00 2.87205786e-01
8.55129898e-01 -1.01307154e+00 1.28914917e+00 -2.45224148e-01
7.45126128e-01 -3.45462978e-01 2.91164935e-01 -9.70223665e-01
-4.39125806e-01 -1.16580367e+00 -3.57987493e-01 9.44921434e-01
6.38647228e-02 -3.79607677e-01 6.64405107e-01 7.35842288e-01
-1.58353627e-01 -1.19913065e+00 -7.14697719e-01 -5.57314813e-01
1.06638543e-01 -4.19086695e-01 9.12249207e-01 1.07891262e+00
-4.75064784e-01 6.67821094e-02 -7.68262804e-01 -1.29689977e-01
6.85963154e-01 3.80252600e-01 1.14510858e+00 -1.11176693e+00
-5.80869377e-01 -5.73208153e-01 -1.77975953e-01 -1.32516348e+00
-7.15577900e-02 -7.97048748e-01 2.84222979e-02 -1.50336146e+00
-4.03009176e-01 -8.80981028e-01 6.94107592e-01 3.68468702e-01
3.85895297e-02 6.05610669e-01 2.62761146e-01 1.95129797e-01
-1.06115140e-01 8.63192260e-01 2.00227928e+00 -2.17403471e-02
-2.70264894e-01 -1.39788434e-01 -2.90393651e-01 8.42954874e-01
7.51781762e-01 -2.67636746e-01 -3.07307214e-01 -7.49947250e-01
5.53507566e-01 4.94009465e-01 7.60726690e-01 -8.72263789e-01
-1.33787975e-01 -3.27701479e-01 4.25755084e-01 -3.00332516e-01
6.13678992e-01 -3.65626544e-01 9.19519722e-01 2.45870769e-01
-1.71671063e-01 2.50742853e-01 1.45056382e-01 3.18973631e-01
2.71521419e-01 2.89820462e-01 8.77904594e-01 -2.04460010e-01
-3.37209255e-01 7.39957333e-01 -7.74630159e-02 1.58382673e-02
7.32982635e-01 -3.58823985e-01 1.55034408e-01 -4.32436347e-01
-6.82310879e-01 -1.99511394e-01 7.90966511e-01 2.68072456e-01
5.74212611e-01 -1.60910892e+00 -7.13135719e-01 1.51287943e-01
-5.38263917e-01 6.62699163e-01 6.12474322e-01 6.32126331e-01
-1.30791759e+00 -2.60412306e-01 -4.76017863e-01 -5.18342137e-01
-9.98337269e-01 3.99303436e-01 2.74990857e-01 -4.55769040e-02
-1.28757548e+00 7.99744487e-01 4.41833347e-01 -4.02267456e-01
1.47236556e-01 -3.65140826e-01 2.46510461e-01 -2.31530756e-01
4.59456533e-01 6.66448951e-01 -7.20213056e-02 -6.41132653e-01
7.30406567e-02 7.75067568e-01 4.38235193e-01 -1.45470545e-01
1.47091401e+00 9.86561701e-02 -8.50160420e-02 1.83664531e-01
1.00490069e+00 -1.52393496e-02 -1.65342486e+00 1.99958727e-01
-8.30039799e-01 -5.97581148e-01 -4.96120006e-02 -3.78603429e-01
-1.32834518e+00 9.14396226e-01 1.37821436e-01 -1.46813691e-01
8.96440089e-01 -2.28014857e-01 1.31671762e+00 1.24173678e-01
5.55075407e-01 -8.94796431e-01 -1.11734591e-01 3.02293569e-01
1.44738364e+00 -5.93389571e-01 1.89188749e-01 -7.99882770e-01
-4.04376477e-01 1.16906774e+00 5.05759001e-01 -5.23278773e-01
2.60598987e-01 3.75993162e-01 -1.05978161e-01 -1.49592221e-01
-3.99208874e-01 4.23029125e-01 3.47104162e-01 4.94333118e-01
2.69722730e-01 3.22472662e-01 -2.60285974e-01 2.69060194e-01
-7.78326690e-01 -1.04025051e-01 2.82030821e-01 5.49229264e-01
1.56839132e-01 -9.66779530e-01 -3.62207800e-01 1.27718717e-01
-8.21610391e-02 4.53317240e-02 -5.88288046e-02 7.97521710e-01
1.58124611e-01 2.42754847e-01 2.18458161e-01 -2.44496301e-01
3.66250604e-01 -4.55349386e-01 7.94280410e-01 -2.51691248e-02
-6.42062008e-01 7.77177662e-02 1.25977382e-01 -9.73052800e-01
-2.80139334e-02 -3.61278594e-01 -1.23799908e+00 -1.03547347e+00
-5.45612313e-02 -2.17756733e-01 5.98673165e-01 4.85523254e-01
5.23225009e-01 5.06478846e-01 3.44606459e-01 -1.59672081e+00
-1.85928568e-01 -4.23093617e-01 -2.86183327e-01 4.60569650e-01
1.35503814e-01 -8.13656151e-01 1.16797939e-01 1.81098089e-01] | [9.148468017578125, -3.15118145942688] |
1963c01a-4db0-46d7-aa3f-aa895e2188df | skin-cancer-classification-using-inception | 2111.02402 | null | https://arxiv.org/abs/2111.02402v1 | https://arxiv.org/pdf/2111.02402v1.pdf | Skin Cancer Classification using Inception Network and Transfer Learning | Medical data classification is typically a challenging task due to imbalance between classes. In this paper, we propose an approach to classify dermatoscopic images from HAM10000 (Human Against Machine with 10000 training images) dataset, consisting of seven imbalanced types of skin lesions, with good precision and low resources requirements. Classification is done by using a pretrained convolutional neural network. We evaluate the accuracy and performance of the proposal and illustrate possible extensions. | ['Mauro Femminella', 'Gianluca Reali', 'Osvaldo Gervasi', 'Marco Simonetti', 'Damiano Perri', 'Priscilla Benedetti'] | 2021-11-03 | null | null | null | null | ['skin-cancer-classification'] | ['medical'] | [ 4.87475038e-01 -2.59965565e-03 -4.87003148e-01 -4.93955791e-01
-4.63736534e-01 -5.16726732e-01 2.39474282e-01 3.15631092e-01
-7.20391512e-01 9.76888299e-01 -4.82748806e-01 -3.92179847e-01
-2.89090365e-01 -7.40853131e-01 -3.76883954e-01 -5.51079392e-01
-1.66235670e-01 6.60447717e-01 8.42315406e-02 -6.53853863e-02
3.68622184e-01 6.29202366e-01 -1.51620305e+00 7.23319948e-01
9.32668865e-01 1.25879872e+00 -5.92085660e-01 1.10994709e+00
1.99183658e-01 9.40661848e-01 -9.22945023e-01 -7.42688417e-01
2.18282118e-01 7.61618419e-03 -1.31896698e+00 1.87259570e-01
6.20978892e-01 -2.68212527e-01 -7.18423724e-02 9.55423594e-01
7.49263406e-01 -4.99154598e-01 7.31599450e-01 -1.01292074e+00
-2.29678676e-01 -5.14808185e-02 -6.37802839e-01 6.93801165e-01
1.35511011e-01 1.52577907e-01 5.28264463e-01 -3.82448375e-01
7.33137310e-01 8.65252316e-01 7.65185297e-01 5.97880423e-01
-9.50076878e-01 -4.70130175e-01 -4.12805557e-01 3.77741128e-01
-1.25795686e+00 1.30648389e-01 3.32836956e-01 -4.51085269e-01
9.22406197e-01 6.29175901e-01 5.95586896e-01 1.13406241e+00
4.16522294e-01 6.38830662e-01 1.24697340e+00 -3.27459872e-01
1.62465900e-01 1.95736468e-01 4.17701542e-01 5.06439209e-01
4.01972651e-01 -1.37345731e-01 -6.49873242e-02 -1.87349170e-01
5.26097357e-01 -1.43777192e-01 1.44886121e-01 6.13366859e-03
-6.14638686e-01 6.41263723e-01 5.51404953e-01 3.24013174e-01
-4.04634923e-01 -1.59159318e-01 7.70280480e-01 5.66553175e-01
5.10425031e-01 5.36675513e-01 -1.71867177e-01 4.55504991e-02
-7.25136817e-01 2.12204233e-01 8.21541727e-01 2.61142224e-01
-1.96162999e-01 -3.51236850e-01 -3.69395353e-02 8.24333966e-01
-3.67327482e-01 1.47910193e-01 8.41720045e-01 -2.44249865e-01
2.60497838e-01 7.54166782e-01 -2.37190556e-02 -9.97172952e-01
-7.47887015e-01 -4.66979295e-01 -1.25079536e+00 2.92083710e-01
4.75119710e-01 -9.89587307e-02 -1.32819033e+00 7.77200878e-01
2.24420637e-01 -3.58791620e-01 -7.21260235e-02 9.23775256e-01
8.93326223e-01 1.74244642e-02 2.29532287e-01 1.27769798e-01
1.11156762e+00 -8.08014929e-01 -6.62276208e-01 3.17259550e-01
7.59420037e-01 -7.83864796e-01 6.51968241e-01 1.07145154e+00
-1.15042937e+00 -5.86445391e-01 -9.58052576e-01 2.25521494e-02
-7.24553168e-01 5.60320675e-01 6.74599886e-01 9.19978917e-01
-9.21490431e-01 6.32686079e-01 -6.60189271e-01 -3.48168582e-01
8.90636444e-01 8.10630441e-01 -6.08896375e-01 1.66654941e-02
-1.15166938e+00 1.04589963e+00 5.78114569e-01 6.08474873e-02
-3.86295289e-01 -5.48248827e-01 -4.29467410e-01 -3.26706558e-01
-1.38434008e-01 -2.65421569e-01 1.14728880e+00 -1.19360495e+00
-1.12922263e+00 1.41607940e+00 3.20413947e-01 -7.36420989e-01
7.59218574e-01 -7.15049822e-03 -6.74730539e-01 3.09740663e-01
-4.64741617e-01 3.69246304e-01 6.55681610e-01 -6.49814248e-01
-7.29732096e-01 -6.44949675e-01 1.54327914e-01 -1.58565529e-02
-4.27180380e-01 -1.32432997e-01 1.04240666e-03 -3.97254854e-01
-2.80277550e-01 -8.46744895e-01 -3.35657716e-01 1.67941034e-01
-6.61644638e-01 -2.62431920e-01 3.75909030e-01 -7.78322041e-01
1.15646946e+00 -1.88189316e+00 -5.42430840e-02 5.69460690e-01
4.66578096e-01 9.45711374e-01 -1.10382564e-01 2.14262784e-01
-4.60126311e-01 1.51546478e-01 1.29323289e-01 2.23249301e-01
-5.04364848e-01 2.77948077e-03 7.71666318e-02 6.02523804e-01
4.49412853e-01 8.35759580e-01 -5.30066550e-01 -4.68607724e-01
5.36382198e-01 4.23995793e-01 -2.62906909e-01 1.33012652e-01
1.08105272e-01 1.26562968e-01 -1.44148141e-01 1.01280034e+00
1.05694270e+00 -5.30677199e-01 3.10098469e-01 -4.75493431e-01
3.53115439e-01 -9.69612077e-02 -9.86981630e-01 1.02286100e+00
-4.03735936e-01 5.34021616e-01 -2.80584812e-01 -1.13028729e+00
6.00807965e-01 1.99598178e-01 2.99719363e-01 -8.18924189e-01
5.39068639e-01 2.67867953e-01 4.90537137e-01 -1.06371486e+00
1.53168693e-01 -1.17702849e-01 1.71480909e-01 3.73717397e-02
2.50130981e-01 1.17311172e-01 3.72579545e-01 -2.56192625e-01
1.03626812e+00 -5.08171618e-01 6.66046858e-01 -9.88419801e-02
5.52793503e-01 2.25284234e-01 -1.35429397e-01 4.92951304e-01
-4.96584535e-01 4.91936773e-01 7.00806260e-01 -1.08357096e+00
-1.11477959e+00 -9.24383700e-01 -6.52243495e-01 4.71647948e-01
-4.62473296e-02 2.79995918e-01 -9.33439672e-01 -7.85146713e-01
6.15937151e-02 -7.47193992e-02 -1.16305244e+00 -8.29314888e-02
-3.28853309e-01 -1.29812658e+00 8.22313666e-01 5.43424606e-01
3.77693534e-01 -9.42944586e-01 -7.07998157e-01 -1.10756494e-01
2.10328639e-01 -8.22273672e-01 2.91457295e-01 1.53098688e-01
-5.76582611e-01 -1.77495539e+00 -7.24183619e-01 -7.76448190e-01
8.59381258e-01 -1.46876290e-01 1.17897046e+00 2.26719528e-01
-1.27884471e+00 -3.01692665e-01 -2.46245489e-01 -4.72012728e-01
-1.96503967e-01 3.40139419e-01 -6.62062839e-02 -6.39612451e-02
6.70532703e-01 -6.87837601e-02 -7.97286570e-01 9.69120562e-02
-1.23506689e+00 -1.60302758e-01 8.11110914e-01 1.14238763e+00
4.68856901e-01 2.05976009e-01 3.73320937e-01 -1.31819141e+00
8.15810800e-01 -4.31993842e-01 -4.06672001e-01 2.81078219e-01
-2.40143865e-01 -4.73760426e-01 8.53376567e-01 -3.71117324e-01
-5.65803111e-01 1.25652691e-02 -4.34505343e-01 -1.89723954e-01
-4.40432906e-01 1.86677545e-01 2.85044700e-01 -6.48446977e-01
1.19752979e+00 -1.13020875e-01 2.22957283e-01 -2.97250479e-01
-2.00979579e-02 1.15211952e+00 3.86081427e-01 1.20374942e-02
2.71847427e-01 4.53441799e-01 9.66891572e-02 -7.20875621e-01
-9.19161856e-01 -4.46926296e-01 -6.53111875e-01 -1.94072664e-01
5.69242775e-01 -6.17921889e-01 -9.20878828e-01 8.83636713e-01
-8.25861752e-01 1.65767744e-02 -5.70694543e-02 3.48759830e-01
-3.80855292e-01 1.03429265e-01 -8.76509905e-01 -5.34123480e-01
-5.85388899e-01 -1.05710673e+00 7.69837618e-01 2.72582501e-01
-3.35389644e-01 -1.05755854e+00 1.03032231e-01 5.20944178e-01
3.90994579e-01 7.55339444e-01 7.68108308e-01 -1.02465487e+00
1.27276048e-01 -8.14744592e-01 -5.01744509e-01 5.81643164e-01
3.64464492e-01 2.86163688e-01 -1.00163925e+00 -5.92342138e-01
-2.04774812e-01 -8.48179221e-01 8.37795794e-01 3.87016863e-01
2.13280916e+00 -2.13649228e-01 -3.43872845e-01 5.94087005e-01
1.69875097e+00 2.60838687e-01 8.82192194e-01 2.53926784e-01
4.53758597e-01 8.54695916e-01 5.57899773e-01 1.99168056e-01
-6.88382387e-02 1.83616981e-01 5.38814068e-01 -7.08670676e-01
5.89708202e-02 3.55467051e-01 -7.56782711e-01 4.82392102e-01
-2.96211213e-01 -4.31758016e-01 -1.06842935e+00 5.98033726e-01
-1.15703177e+00 -5.78902006e-01 -6.33243844e-02 1.99409688e+00
9.34498191e-01 3.19788158e-01 2.84348458e-01 6.58006489e-01
7.49402821e-01 -1.56201795e-01 -6.03983641e-01 -7.38170445e-01
-6.67935470e-03 9.25194323e-01 5.57809472e-01 4.11097407e-01
-1.32238054e+00 5.80720603e-01 7.60956287e+00 1.08133638e+00
-1.58649862e+00 -2.02756599e-01 1.34046674e+00 -2.54589587e-01
3.92909408e-01 -1.07444620e+00 -4.76686835e-01 6.22938097e-01
8.88086557e-01 1.03673369e-01 -7.38213956e-02 6.33006811e-01
-4.43098515e-01 -1.69914752e-01 -7.38313854e-01 9.40371931e-01
2.26218745e-01 -1.33723223e+00 1.28014266e-01 -1.43741429e-01
6.45903945e-01 -2.65957862e-01 5.47271669e-01 -2.04706136e-02
4.88061905e-02 -1.78421164e+00 -3.51068050e-01 4.68488604e-01
1.28811622e+00 -9.89766896e-01 1.30608261e+00 -1.69954170e-02
-4.15223300e-01 6.24833219e-02 -3.91380310e-01 -1.07747264e-01
-5.54165721e-01 7.62143373e-01 -1.12502980e+00 4.49482054e-01
9.16490972e-01 5.11903882e-01 -6.98026657e-01 1.12368381e+00
2.81820178e-01 5.20933092e-01 -6.48089349e-02 -2.57418543e-01
1.42583773e-01 3.09837699e-01 1.93546414e-02 1.27418947e+00
-2.05823705e-01 -1.38681531e-02 -1.85806930e-01 2.80658603e-01
-1.08535886e-02 3.23693067e-01 -5.43590903e-01 -4.60077263e-02
-3.76695469e-02 1.46714509e+00 -6.25420392e-01 -4.01731998e-01
3.98945063e-02 8.35344732e-01 2.60261416e-01 -4.65109237e-02
-6.77335322e-01 -7.72567391e-01 3.35360765e-01 1.95950851e-01
4.23683487e-02 7.26743340e-01 -3.91184509e-01 -8.56583834e-01
1.19908191e-01 -1.16052771e+00 5.75113535e-01 -4.20886815e-01
-1.51645803e+00 8.68654072e-01 -5.26938796e-01 -1.52345765e+00
-2.22431511e-01 -1.26977158e+00 -3.21322352e-01 7.44117320e-01
-1.60920453e+00 -1.16932213e+00 -5.53642690e-01 3.80376220e-01
2.18960762e-01 -2.47993127e-01 9.71775234e-01 4.02029246e-01
-4.69644040e-01 9.63047385e-01 1.10186070e-01 2.15603828e-01
7.83991456e-01 -1.41327977e+00 2.57118642e-01 9.84505862e-02
-7.12539971e-01 4.05280143e-01 4.68677521e-01 -4.01797384e-01
-6.86795294e-01 -1.16141737e+00 7.89856851e-01 -1.96278349e-01
4.97967154e-01 4.40796986e-02 -7.27477014e-01 2.09144816e-01
1.42105550e-01 3.37507248e-01 1.31738198e+00 -3.60936485e-02
-2.42122665e-01 -6.78914115e-02 -1.77522838e+00 2.72280842e-01
4.42569613e-01 -2.47676551e-01 -2.89520770e-01 7.61849046e-01
1.12607285e-01 -7.91353047e-01 -1.25624573e+00 6.74943030e-01
7.09799230e-01 -1.03148472e+00 9.32184756e-01 -1.28628635e+00
8.89025092e-01 1.12604454e-01 2.00073853e-01 -1.20350349e+00
1.01588108e-01 -2.13067550e-02 1.01470493e-01 4.33848202e-01
2.20276132e-01 -4.38975245e-01 1.14197874e+00 1.30967915e-01
5.53704023e-01 -1.39685142e+00 -9.73310113e-01 -4.76802468e-01
2.21799940e-01 9.77554321e-02 3.59696776e-01 1.06674182e+00
1.91578418e-02 -7.10140467e-02 -5.67014158e-01 -1.37839034e-01
4.73441482e-01 -5.63112125e-02 4.67269212e-01 -1.19490194e+00
-1.95569564e-02 -2.20843792e-01 -1.21241367e+00 -1.89161718e-01
-3.45228553e-01 -5.80808699e-01 -4.96985137e-01 -1.07894027e+00
2.81317085e-01 -2.94605702e-01 -6.59273863e-01 3.55214596e-01
-1.97151005e-01 1.00968063e+00 -2.76537627e-01 -7.82505348e-02
-6.29436255e-01 -3.18304628e-01 1.22598457e+00 -2.57599264e-01
2.49628544e-01 1.50287718e-01 -3.44024539e-01 7.37855494e-01
1.15440559e+00 -1.45004615e-01 -1.58948511e-01 -2.03705117e-01
2.56151110e-01 4.89801764e-02 3.12615424e-01 -1.32249248e+00
2.03962460e-01 -2.00150292e-02 1.11565197e+00 -6.73765898e-01
2.23177865e-01 -5.86957872e-01 -4.67857579e-03 1.26695585e+00
-7.14592993e-01 -6.80708960e-02 2.66913623e-01 1.84969947e-01
-6.20214283e-01 -1.79686472e-01 1.11222219e+00 -2.52734870e-01
-3.46795082e-01 4.01815534e-01 -2.03129292e-01 -1.41364589e-01
1.32023633e+00 -2.63127357e-01 -5.41119576e-01 1.46889791e-01
-9.70912576e-01 8.56701210e-02 2.80137181e-01 2.53382683e-01
5.21811068e-01 -1.16716576e+00 -4.72698569e-01 3.66552532e-01
3.40971559e-01 -3.17962289e-01 6.31449044e-01 7.66176462e-01
-1.23645771e+00 6.64351046e-01 -1.04445243e+00 -6.64532721e-01
-1.70196950e+00 4.17070657e-01 8.77259612e-01 -5.16840518e-01
8.36161673e-02 1.00850856e+00 -2.05204204e-01 -5.96203625e-01
2.49102011e-01 -1.01875193e-01 -6.97247028e-01 1.13455288e-01
7.59960711e-01 4.54615682e-01 6.02462828e-01 -2.71454841e-01
-3.12752694e-01 4.23679233e-01 -5.97595394e-01 5.42402983e-01
1.19105208e+00 5.39052308e-01 -3.29169273e-01 1.60712048e-01
1.19829381e+00 -4.30554122e-01 -4.20483500e-01 1.33590415e-01
-2.36271918e-01 -6.39391840e-01 -2.92946368e-01 -1.25494242e+00
-1.12479424e+00 1.12429273e+00 1.43191600e+00 3.99789780e-01
1.27417684e+00 -6.46590412e-01 6.90005422e-01 3.25002640e-01
1.03750736e-01 -1.26283157e+00 7.85321966e-02 6.12461008e-03
3.26672584e-01 -1.67089999e+00 -8.47060159e-02 -6.56883121e-01
-5.11674047e-01 1.29524720e+00 8.82287443e-01 -5.64754486e-01
6.15860224e-01 2.37300485e-01 4.16238487e-01 -2.94247657e-01
-7.94335783e-01 7.61630991e-03 4.20247257e-01 7.92920053e-01
5.16231000e-01 1.22508287e-01 -8.83828938e-01 2.66667575e-01
-1.48290977e-01 5.06822050e-01 3.77974123e-01 8.00994158e-01
-1.95380032e-01 -1.00893092e+00 -1.04437821e-01 1.09773040e+00
-1.14668453e+00 1.25839682e-02 -5.55228949e-01 9.57008004e-01
4.69787091e-01 6.11600697e-01 1.72894493e-01 -5.61903954e-01
3.36856067e-01 -4.42044169e-01 5.95440924e-01 -2.14800313e-01
-9.13253069e-01 -4.35697913e-01 1.22830033e-01 -6.78929865e-01
-3.23408157e-01 1.43368155e-01 -5.48084021e-01 -3.76024634e-01
-1.42824829e-01 -1.35325775e-01 6.02457047e-01 6.94994867e-01
8.38800147e-02 7.44939446e-01 7.98919141e-01 -1.44775286e-01
-7.73175299e-01 -1.02750576e+00 -6.57377183e-01 8.09753060e-01
5.18861115e-01 -4.32026297e-01 -3.17690670e-01 -2.70672310e-02] | [15.658957481384277, -2.9720678329467773] |
18bbb742-fe8c-41af-bda3-3157bd2cfcf8 | domain-adversarial-graph-convolutional | 2204.05184 | null | https://arxiv.org/abs/2204.05184v3 | https://arxiv.org/pdf/2204.05184v3.pdf | Domain Adversarial Graph Convolutional Network Based on RSSI and Crowdsensing for Indoor Localization | In recent years, the use of WiFi fingerprints for indoor positioning has grown in popularity, largely due to the widespread availability of WiFi and the proliferation of mobile communication devices. However, many existing methods for constructing fingerprint datasets rely on labor-intensive and time-consuming processes of collecting large amounts of data. Additionally, these methods often focus on ideal laboratory environments, rather than considering the practical challenges of large multi-floor buildings. To address these issues, we present a novel WiDAGCN model that can be trained using a small number of labeled site survey data and large amounts of unlabeled crowdsensed WiFi fingerprints. By constructing heterogeneous graphs based on received signal strength indicators (RSSIs) between waypoints and WiFi access points (APs), our model is able to effectively capture the topological structure of the data. We also incorporate graph convolutional networks (GCNs) to extract graph-level embeddings, a feature that has been largely overlooked in previous WiFi indoor localization studies. To deal with the challenges of large amounts of unlabeled data and multiple data domains, we employ a semi-supervised domain adversarial training scheme to effectively utilize unlabeled data and align the data distributions across domains. Our system is evaluated using a public indoor localization dataset that includes multiple buildings, and the results show that it performs competitively in terms of localization accuracy in large buildings. | ['Xuan Song', 'Ryosuke Shibasaki', 'Zipei Fan', 'Mingxin Zhang'] | 2022-04-06 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [ 2.27305129e-01 -1.17563337e-01 -1.43108457e-01 -4.81482536e-01
-7.54357696e-01 -8.32639933e-01 3.70336652e-01 -6.93696514e-02
-1.18862599e-01 8.51685584e-01 2.11317003e-01 -4.19494331e-01
-2.30820596e-01 -1.27749801e+00 -9.57011461e-01 -4.40064520e-01
-1.82859704e-01 2.73974240e-01 1.53527278e-02 -4.15832177e-03
1.98770221e-02 3.89716506e-01 -1.22671986e+00 -2.69445390e-01
1.06979740e+00 1.05388904e+00 -1.05127022e-01 4.66112554e-01
-5.67365065e-02 2.85525471e-01 -5.35615623e-01 -4.04845715e-01
3.03388089e-01 3.49135115e-03 -3.85484636e-01 -3.23730230e-01
7.49708354e-01 -2.77857304e-01 -8.07733238e-01 7.79994845e-01
7.34050274e-01 5.25193289e-02 3.25908631e-01 -1.36618352e+00
-7.95557797e-01 4.53442544e-01 -2.98863381e-01 4.02477570e-02
7.97784090e-01 -3.25504154e-01 6.78818524e-01 -6.99178278e-01
2.63670385e-01 7.77436674e-01 1.40117550e+00 7.46352598e-02
-1.32472038e+00 -9.87456322e-01 8.64542052e-02 -1.91772327e-01
-1.82841289e+00 -3.72301251e-01 8.99650931e-01 -2.03589424e-01
4.49441135e-01 7.05163553e-02 5.56935430e-01 1.61941552e+00
1.38645843e-01 1.99367821e-01 7.69331038e-01 -2.01962590e-01
3.49360108e-01 -2.20208377e-01 -2.69060433e-01 5.82719803e-01
5.39900243e-01 1.51429966e-01 -3.42344105e-01 -4.45182115e-01
8.74043643e-01 5.87233603e-01 -3.29841614e-01 -6.44435287e-01
-1.26135039e+00 4.54091549e-01 8.45602751e-01 3.78194630e-01
-2.42833629e-01 3.52246791e-01 -1.66643560e-01 -5.67590669e-02
2.33784929e-01 3.22172403e-01 -1.28036186e-01 -1.47786140e-01
-1.04264426e+00 1.81563497e-01 7.84294248e-01 1.33280575e+00
9.76886988e-01 -1.48219451e-01 2.51654387e-02 7.93430567e-01
4.49723154e-01 1.12035084e+00 2.20626697e-01 -7.61094987e-01
8.32983553e-01 4.99506563e-01 3.73648167e-01 -1.57402062e+00
-3.60125721e-01 -4.77570713e-01 -8.02921951e-01 -6.65673554e-01
5.66986024e-01 -2.85716742e-01 -8.36516500e-01 1.91363037e+00
1.41698420e-01 7.18721867e-01 -3.28887969e-01 4.75898087e-01
5.41907668e-01 2.99174577e-01 -1.20060422e-01 5.40800273e-01
8.61426711e-01 -6.96551919e-01 -3.66211265e-01 -4.43498552e-01
2.75495112e-01 -4.56167787e-01 1.00934982e+00 -1.06572010e-01
-4.37500000e-01 -6.87696755e-01 -1.22079885e+00 1.16808392e-01
-7.02070236e-01 -1.87778890e-01 7.36273766e-01 1.19086123e+00
-9.91753459e-01 3.03615451e-01 -8.55412543e-01 -5.61887681e-01
6.27669334e-01 6.99084759e-01 -4.29831177e-01 -8.07880163e-01
-1.21082473e+00 8.85911435e-02 -1.19834065e-01 2.10397288e-01
-3.83252919e-01 -5.60563982e-01 -9.42350149e-01 1.11355223e-01
2.61432398e-02 -5.71515858e-01 7.87332118e-01 -2.95118570e-01
-1.09926784e+00 5.34989126e-02 -8.95528197e-02 -7.38855600e-02
1.77843168e-01 -1.20522664e-03 -9.90692019e-01 -2.55980968e-01
5.16134560e-01 1.19504578e-01 4.04650450e-01 -1.22453928e+00
-4.66821432e-01 -5.83691895e-01 1.05012536e-01 -2.77048677e-01
-3.11818928e-01 -8.49503994e-01 -5.71691930e-01 -6.57779932e-01
4.14912671e-01 -1.28269136e+00 -3.09089184e-01 -2.52129287e-01
-3.97215128e-01 4.40574735e-01 6.77649260e-01 -3.34747642e-01
1.19849586e+00 -2.24350572e+00 -6.58411503e-01 8.10406864e-01
2.57328838e-01 -6.74304068e-02 -1.29863024e-01 5.49829423e-01
4.47712868e-01 1.67072803e-01 -1.63218692e-01 -2.72409320e-01
1.59172758e-01 2.63095796e-01 -2.08502620e-01 4.35990304e-01
-2.78996557e-01 9.87459600e-01 -1.28029919e+00 -2.90720105e-01
1.89429298e-01 5.32186925e-01 -6.83005035e-01 4.58105691e-02
3.24720502e-01 6.93665504e-01 -8.09446871e-01 8.29422593e-01
1.04224300e+00 -4.48726088e-01 4.06528234e-01 -1.27439633e-01
2.28597194e-01 3.38203013e-01 -1.33214760e+00 1.96735585e+00
-6.83170438e-01 3.96329820e-01 -2.07763970e-01 -8.63156974e-01
7.72692561e-01 1.14156142e-01 7.24599659e-01 -7.91525543e-01
-1.16740331e-01 3.53931844e-01 -5.56075931e-01 -7.49759898e-02
3.11185598e-01 2.62926549e-01 -4.96345341e-01 4.12399918e-01
-1.25838369e-02 8.17222968e-02 -2.91580498e-01 5.32246009e-02
1.66389835e+00 -1.09924674e-01 -1.07455544e-01 6.90180734e-02
3.31840843e-01 -2.03969717e-01 3.92151207e-01 1.12329221e+00
-1.50104031e-01 7.58992076e-01 -2.44334713e-01 -3.32965374e-01
-6.19947016e-01 -1.56077003e+00 -2.50810057e-01 9.07414198e-01
5.99320710e-01 -3.88821691e-01 -4.48057741e-01 -7.97028780e-01
5.39471924e-01 2.07402259e-01 -6.40661657e-01 3.26355034e-03
-4.05830085e-01 -5.24098873e-01 1.08713257e+00 5.76474607e-01
7.61915922e-01 -4.08276439e-01 3.36192906e-01 4.55704808e-01
-3.55417490e-01 -1.45560110e+00 -5.45662463e-01 3.74378152e-02
-4.98749137e-01 -1.04298007e+00 -5.20465612e-01 -8.65540445e-01
8.88397515e-01 7.22217083e-01 9.29955482e-01 4.31643194e-03
7.92948157e-02 6.79918349e-01 -2.21148476e-01 -1.83708712e-01
2.37063318e-01 5.03870845e-01 2.68938571e-01 1.39125675e-01
4.05697376e-01 -1.09323859e+00 -7.35677063e-01 8.02761734e-01
-5.48630357e-01 -4.33673501e-01 5.18129766e-01 6.51581228e-01
5.83725452e-01 6.52436465e-02 6.29054725e-01 -8.59183908e-01
5.63956082e-01 -8.16223323e-01 -4.20227975e-01 1.66451111e-01
-5.79575598e-01 -2.02636391e-01 8.17624748e-01 -3.38508934e-01
-6.37011588e-01 2.72210874e-02 -2.07440689e-01 7.93973729e-02
-1.35049239e-01 3.40361595e-01 -4.37012225e-01 -7.22890437e-01
6.86096430e-01 1.09959774e-01 -4.63627577e-01 -2.64722824e-01
4.30867434e-01 9.59925652e-01 6.92121327e-01 -7.91188598e-01
1.17802835e+00 5.58905423e-01 -1.29932731e-01 -6.88936293e-01
-7.48577714e-01 -4.81117427e-01 -4.46638048e-01 1.25085132e-03
4.52643931e-01 -9.77340460e-01 -5.70088387e-01 1.94657713e-01
-7.18303144e-01 -1.80517107e-01 3.05177048e-02 5.48876345e-01
-1.11735918e-01 3.85993093e-01 -2.49076009e-01 -4.45852399e-01
8.00785422e-02 -9.48135018e-01 1.13352799e+00 2.85796702e-01
9.09086037e-03 -9.64361787e-01 3.65050107e-01 3.45409691e-01
7.88396060e-01 4.62217450e-01 5.45443654e-01 -4.19586658e-01
-8.33037198e-01 -8.50812912e-01 -2.44480014e-01 -5.94933368e-02
5.47888577e-01 -5.19251823e-01 -1.08942640e+00 -2.81156361e-01
-5.39160728e-01 -7.00188652e-02 1.72547057e-01 2.78767794e-01
1.34926116e+00 -1.91687033e-01 -8.99898171e-01 9.98074830e-01
1.35462272e+00 -2.05744002e-02 6.71456754e-01 1.82203442e-01
9.88873363e-01 1.67463142e-02 2.82016516e-01 2.82268703e-01
7.54576623e-01 7.29109287e-01 3.72563660e-01 -1.47636384e-01
6.48585409e-02 -9.79720831e-01 -1.56087369e-01 5.37698328e-01
-8.99969190e-02 -4.71236289e-01 -9.97967482e-01 4.25437808e-01
-1.70334947e+00 -9.48743284e-01 2.59153266e-02 2.28121996e+00
1.72358021e-01 7.56988898e-02 -1.32683992e-01 1.47000775e-01
7.50488400e-01 2.69869268e-01 -4.19900388e-01 1.93337351e-01
1.82208270e-01 4.05003965e-01 9.60858881e-01 3.32156181e-01
-1.22706819e+00 5.23625374e-01 6.25831652e+00 3.71239066e-01
-1.19579411e+00 -1.98719278e-02 1.22245453e-01 1.48356289e-01
-4.62158889e-01 -3.13328266e-01 -4.65740561e-01 9.05760705e-01
9.11918044e-01 2.10833848e-01 8.35791528e-01 1.05834842e+00
-9.74579901e-02 1.63696870e-01 -9.61834848e-01 1.23166752e+00
-2.25811243e-01 -1.41515660e+00 -4.13659930e-01 4.58672136e-01
9.88098383e-01 4.69415426e-01 1.27650201e-01 5.12617648e-01
2.84656018e-01 -9.98991549e-01 4.59242910e-01 3.20611149e-01
9.64051247e-01 -5.96929073e-01 7.89825916e-01 1.95174247e-01
-1.74254310e+00 -1.18433371e-01 -2.93261081e-01 -1.38535246e-01
5.54517191e-03 7.07209051e-01 -7.89870858e-01 7.31705427e-01
7.37454951e-01 5.75543940e-01 -7.40416229e-01 1.14602685e+00
-1.41853079e-01 5.73763430e-01 -7.70144820e-01 3.74818370e-02
1.53107261e-02 -8.06808472e-02 -1.79709807e-01 1.01624095e+00
7.88940847e-01 -3.05954784e-01 2.87515849e-01 7.53340721e-01
-4.51109439e-01 -2.20948130e-01 -1.14072835e+00 2.00262144e-02
1.12428856e+00 1.03321433e+00 -4.90352869e-01 1.94272876e-01
-6.57126009e-01 8.65625560e-01 2.12653860e-01 6.22632384e-01
-1.08309996e+00 -6.39544964e-01 6.93676829e-01 4.40707356e-01
2.75486350e-01 -5.81033051e-01 -1.84940845e-01 -1.10681701e+00
1.19519420e-01 -4.47186947e-01 -2.02262610e-01 -3.33378822e-01
-1.35083079e+00 3.95804673e-01 -5.29805064e-01 -1.26442313e+00
-1.72873735e-02 -1.32328212e-01 -5.41841924e-01 7.62800395e-01
-1.44258952e+00 -1.25172520e+00 -8.76885414e-01 8.15707326e-01
-3.88538182e-01 2.00638056e-01 1.12929285e+00 1.07054412e+00
-5.22562206e-01 9.15491760e-01 6.70175970e-01 4.90560114e-01
7.53422737e-01 -1.11762607e+00 9.08336639e-01 8.06204796e-01
4.70509410e-01 1.18702710e+00 1.69364035e-01 -4.70307171e-01
-1.56370223e+00 -1.34719503e+00 8.23562264e-01 -6.76639080e-01
4.81757820e-01 -8.95621121e-01 -4.20967281e-01 7.51094937e-01
-5.94327986e-01 7.15678692e-01 1.12825370e+00 2.55776137e-01
-3.29102993e-01 -4.54153240e-01 -1.35631156e+00 3.80373061e-01
1.49954629e+00 -8.11332941e-01 -2.34380111e-01 9.58747417e-02
3.86202216e-01 -3.94686043e-01 -1.01231730e+00 3.61179590e-01
8.99897456e-01 -7.86073983e-01 1.29795706e+00 7.89082870e-02
-1.47301123e-01 -5.67155242e-01 -4.12540197e-01 -1.06391156e+00
-4.21270043e-01 -2.86334664e-01 -7.38296807e-02 1.36070776e+00
2.69031048e-01 -9.12699282e-01 1.12022865e+00 7.09092915e-01
1.11800231e-01 -4.71164465e-01 -9.04105127e-01 -7.06815660e-01
-5.60951829e-01 -6.08510435e-01 1.31225312e+00 1.00812972e+00
-3.08815926e-01 1.19929887e-01 -2.40281448e-01 6.61120355e-01
5.89235127e-01 1.99200828e-02 1.19739437e+00 -1.40694225e+00
-1.79739699e-01 1.06055535e-01 -6.14533842e-01 -1.41077805e+00
-9.34293270e-02 -7.84624994e-01 1.73400328e-01 -1.80804586e+00
-4.78610814e-01 -1.10344863e+00 -5.42934299e-01 3.34241569e-01
7.66349360e-02 6.69984579e-01 -3.69449377e-01 8.39638412e-02
-7.21575975e-01 4.09353822e-01 8.90782297e-01 -4.00982082e-01
-1.95389003e-01 3.69755417e-01 -7.95141637e-01 4.93948847e-01
5.38991392e-01 -4.72327590e-01 -5.64912617e-01 -6.28322959e-01
1.30421415e-01 -1.42041206e-01 4.00066376e-01 -1.68544745e+00
2.31727287e-01 -7.14353099e-03 7.86915660e-01 -2.51851588e-01
2.05416918e-01 -1.18718004e+00 4.06316727e-01 4.03642328e-03
1.25963569e-01 -1.98147111e-02 8.12333971e-02 1.12351143e+00
-5.96291423e-02 4.43150342e-01 2.53967911e-01 2.52138108e-01
-4.96655196e-01 5.84149361e-01 -8.54439363e-02 9.95163769e-02
7.57536232e-01 -5.51143110e-01 -1.58716723e-01 -4.68092740e-01
-1.90225631e-01 1.56139499e-02 9.22265410e-01 4.73613262e-01
3.05062741e-01 -1.75101423e+00 1.59677789e-01 6.10886097e-01
4.57526386e-01 4.95866761e-02 1.64843038e-01 4.86164093e-01
-5.55491447e-01 3.72287929e-01 -9.22450703e-03 -4.56438601e-01
-4.74165112e-01 2.90679961e-01 1.79403991e-01 -8.10689777e-02
-4.51456755e-01 6.68777645e-01 -2.16102481e-01 -1.00743294e+00
3.57577056e-01 -5.22094905e-01 3.01597506e-01 -4.71127719e-01
2.35502854e-01 3.87817889e-01 2.07175344e-01 -5.17433643e-01
-7.51052499e-01 7.51086473e-01 5.10039985e-01 2.20994368e-01
1.17536438e+00 -1.91200867e-01 2.99961656e-01 -6.85325861e-02
1.28877401e+00 7.48567224e-01 -1.10250747e+00 -1.90988511e-01
5.98204136e-02 -7.06042945e-01 -1.67840376e-01 -6.00456417e-01
-9.92034912e-01 3.86686444e-01 7.73362100e-01 3.02015603e-01
6.89460337e-01 -2.39230052e-01 1.15515387e+00 5.35363734e-01
1.21251082e+00 -7.13855267e-01 -2.26734117e-01 3.62480819e-01
7.41945580e-02 -1.18677330e+00 -1.94170013e-01 -3.24039191e-01
1.49289653e-01 8.00857127e-01 2.89714575e-01 -1.66648597e-01
8.95932019e-01 1.67535171e-01 -1.33043295e-02 -4.84575145e-02
4.83239412e-01 1.13156307e-02 1.61421284e-01 1.12550759e+00
2.85141796e-01 1.31273285e-01 1.82704255e-01 9.16455686e-01
-3.56154174e-01 2.23729819e-01 3.59201096e-02 1.11286664e+00
-1.79526404e-01 -1.44846940e+00 -4.16148514e-01 6.22534096e-01
-2.82629073e-01 4.36511338e-02 -1.77522171e-02 6.84355974e-01
5.08673906e-01 1.19595802e+00 -3.23060304e-01 -8.78603220e-01
4.49140370e-01 -2.69777596e-01 4.08896297e-01 -5.71401954e-01
-8.50542784e-02 -4.60645407e-01 -8.16201940e-02 -7.16123462e-01
-2.54880965e-01 -2.18611911e-01 -1.02593040e+00 -5.91077387e-01
-2.33728781e-01 2.42481470e-01 7.58888602e-01 8.41892600e-01
6.81392789e-01 4.28690404e-01 7.83388078e-01 -1.17425168e+00
-1.43790424e-01 -5.35564363e-01 -8.19731712e-01 3.72250974e-01
2.34984338e-01 -1.00592661e+00 -2.05901727e-01 -4.36925471e-01] | [6.421818256378174, 0.8708997964859009] |
725abf83-ae7f-4371-a209-dba3db186584 | a-simple-information-based-approach-to-1 | 2201.12549 | null | https://arxiv.org/abs/2201.12549v1 | https://arxiv.org/pdf/2201.12549v1.pdf | A Simple Information-Based Approach to Unsupervised Domain-Adaptive Aspect-Based Sentiment Analysis | Aspect-based sentiment analysis (ABSA) is a fine-grained sentiment analysis task which aims to extract the aspects from sentences and identify their corresponding sentiments. Aspect term extraction (ATE) is the crucial step for ABSA. Due to the expensive annotation for aspect terms, we often lack labeled target domain data for fine-tuning. To address this problem, many approaches have been proposed recently to transfer common knowledge in an unsupervised way, but such methods have too many modules and require expensive multi-stage preprocessing. In this paper, we propose a simple but effective technique based on mutual information maximization, which can serve as an additional component to enhance any kind of model for cross-domain ABSA and ATE. Furthermore, we provide some analysis of this approach. Experiment results show that our proposed method outperforms the state-of-the-art methods for cross-domain ABSA by 4.32% Micro-F1 on average over 10 different domain pairs. Apart from that, our method can be extended to other sequence labeling tasks, such as named entity recognition (NER). | ['Xiaojun Wan', 'Xiang Chen'] | 2022-01-29 | null | null | null | null | ['term-extraction', 'aspect-based-sentiment-analysis'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.46713027e-01 -1.74121365e-01 -5.16765751e-02 -6.62971377e-01
-1.05396080e+00 -8.56529474e-01 7.04301596e-01 2.28398383e-01
-4.66065705e-01 6.43465996e-01 1.70948971e-02 -2.58512765e-01
8.83149430e-02 -8.55846941e-01 -4.28549439e-01 -6.58991814e-01
4.98929203e-01 5.26278019e-01 4.16823208e-01 -6.08593166e-01
3.49560797e-01 1.04112029e-01 -1.29247797e+00 4.53184962e-01
9.53864396e-01 8.91777158e-01 2.25206595e-02 3.11092705e-01
-8.95003974e-01 4.21213627e-01 -7.69750834e-01 -8.93444180e-01
-1.76286921e-01 -5.51160216e-01 -8.89191806e-01 9.75393876e-02
-2.78009832e-01 3.08732331e-01 4.46736455e-01 1.17483675e+00
5.04514396e-01 -1.52254095e-02 7.82943070e-01 -1.14298546e+00
-3.79421622e-01 6.63830698e-01 -8.39305639e-01 -1.64176539e-01
3.75532389e-01 -2.23011255e-01 1.14188802e+00 -9.06207144e-01
4.61161941e-01 1.00196004e+00 6.81004405e-01 6.07470572e-01
-6.95601761e-01 -5.64047933e-01 3.94421011e-01 -1.33601287e-02
-9.59137917e-01 -2.66552091e-01 9.60969865e-01 -2.69188881e-01
1.05194652e+00 2.32139915e-01 6.14854038e-01 1.01808047e+00
-6.67646378e-02 1.07780206e+00 1.50113738e+00 -5.03902316e-01
3.11283499e-01 4.70149428e-01 4.04297024e-01 3.30940276e-01
3.44009191e-01 -6.05526805e-01 -5.36543310e-01 -3.28335136e-01
1.91642061e-01 -2.89013058e-01 -2.88450792e-02 -2.12263003e-01
-1.09626591e+00 8.37454438e-01 -2.65510976e-01 5.80737948e-01
-2.81626225e-01 -5.52150488e-01 6.48925066e-01 3.23217034e-01
6.99914753e-01 6.25472188e-01 -1.12259126e+00 -2.84467816e-01
-7.74647355e-01 1.12983502e-01 1.20183098e+00 1.08984756e+00
7.59243608e-01 -1.09285340e-01 -1.19946331e-01 1.05124021e+00
3.44012260e-01 5.38437188e-01 7.94909298e-01 -1.47978365e-01
5.54648161e-01 9.84896064e-01 -3.78924003e-03 -8.70929182e-01
-3.20516020e-01 -2.02368364e-01 -6.92492843e-01 -9.28294957e-02
3.52367133e-01 -2.46737108e-01 -9.51803207e-01 1.40625274e+00
5.08532822e-01 -2.71672040e-01 2.45197326e-01 6.59432411e-01
8.69063258e-01 5.54685056e-01 1.10395037e-01 -2.42199630e-01
2.05782723e+00 -1.12480319e+00 -7.50950933e-01 -4.25506234e-01
5.52716732e-01 -1.28163040e+00 1.08498323e+00 3.96553278e-01
-6.93919420e-01 -2.84401834e-01 -1.07272363e+00 2.33604625e-01
-8.71038437e-01 2.47828946e-01 8.79403472e-01 1.00037575e+00
-6.96198285e-01 2.53110141e-01 -6.81700110e-01 -4.50646281e-01
1.64750338e-01 2.41069123e-01 -5.02798080e-01 3.67219374e-02
-1.28633428e+00 6.64407909e-01 3.59651208e-01 -6.24147542e-02
-3.61356586e-01 -4.75370735e-01 -9.80511546e-01 -7.34182671e-02
4.75142956e-01 -7.58676767e-01 1.28565001e+00 -1.05668092e+00
-1.69572616e+00 9.84770060e-01 -4.07591999e-01 -2.14904115e-01
2.71626143e-03 -4.41526383e-01 -6.39318883e-01 -7.31348321e-02
3.27422321e-01 1.51801467e-01 9.46260154e-01 -9.94071960e-01
-6.59931004e-01 -5.83245993e-01 3.38626683e-01 1.96944565e-01
-6.12894893e-01 4.61604714e-01 -6.06926620e-01 -8.95452082e-01
-1.64017692e-01 -9.06699181e-01 -4.29277956e-01 -6.46941423e-01
-2.86322713e-01 -3.30095470e-01 7.03855038e-01 -5.49963593e-01
1.35829067e+00 -1.94004238e+00 -5.48059382e-02 1.11669995e-01
-1.15546793e-01 4.14678246e-01 -1.30002707e-01 4.78935421e-01
-3.86899635e-02 1.02981657e-01 -6.75374091e-01 -3.03464174e-01
9.85217467e-02 -1.30824447e-01 -1.46280602e-01 -1.08048655e-01
4.33573037e-01 6.70839727e-01 -8.12857389e-01 -5.92074990e-01
-1.45134985e-01 4.76048887e-01 -2.89521843e-01 2.06622735e-01
-3.63038510e-01 3.87043387e-01 -6.84493899e-01 6.83840275e-01
8.28093350e-01 -2.06871197e-01 3.08392942e-01 -2.17944637e-01
6.65942812e-03 5.58824718e-01 -1.24011815e+00 1.83875942e+00
-6.94904447e-01 2.60426015e-01 -8.05981383e-02 -1.12935507e+00
1.12223494e+00 3.92913043e-01 4.21118498e-01 -4.74568367e-01
3.50663006e-01 2.76954919e-01 -2.19821632e-01 -4.20048833e-01
7.28863657e-01 -4.58354712e-01 -6.35127068e-01 5.33955693e-01
2.28189930e-01 -2.71362543e-01 4.67551768e-01 -1.17000397e-02
9.94371355e-01 2.26992562e-01 7.40397274e-01 -2.64100611e-01
1.10351074e+00 2.44233862e-01 6.82787359e-01 2.62401611e-01
-1.66532546e-01 5.21989405e-01 4.16491240e-01 -3.03509772e-01
-8.86484087e-01 -4.75194514e-01 -1.08294776e-02 1.01475286e+00
3.10480352e-02 -7.06971049e-01 -9.17593181e-01 -1.20336497e+00
-4.86541063e-01 5.82214415e-01 -3.57928693e-01 1.39882013e-01
-4.14978594e-01 -1.13219285e+00 5.71494579e-01 4.20686454e-01
6.92678213e-01 -1.06944835e+00 3.67336650e-03 2.86321849e-01
-3.86066198e-01 -1.35843694e+00 -4.01207238e-01 2.39454731e-01
-7.54603982e-01 -9.27565634e-01 -8.21798623e-01 -9.83694673e-01
5.20998895e-01 1.87625676e-01 1.30929065e+00 -3.92979622e-01
9.44957808e-02 2.48176247e-01 -7.39189088e-01 -4.98033077e-01
-3.19922388e-01 4.69597489e-01 -1.44293040e-01 2.08567902e-01
1.06649685e+00 -5.52261651e-01 -5.83490968e-01 3.73526573e-01
-1.00790286e+00 -2.06424609e-01 1.04511476e+00 7.44846582e-01
8.14679265e-01 2.22271249e-01 7.06110120e-01 -1.56124294e+00
9.24603105e-01 -3.29728752e-01 -6.02657139e-01 3.63647878e-01
-6.67609870e-01 3.15086365e-01 7.47836411e-01 -2.27335200e-01
-1.56807280e+00 2.19567433e-01 -4.44883019e-01 2.85729259e-01
-4.36080784e-01 6.34974897e-01 -6.51371658e-01 1.67985946e-01
3.11376274e-01 5.05264580e-01 -3.65454286e-01 -5.80788791e-01
4.01239514e-01 9.70244944e-01 -2.77548991e-02 -4.97201502e-01
6.38873816e-01 3.93903822e-01 -1.52905002e-01 -8.90595019e-01
-1.15733647e+00 -9.27601516e-01 -5.71720064e-01 1.94855452e-01
8.55637133e-01 -1.06060982e+00 -3.82427603e-01 6.97236240e-01
-1.21573889e+00 2.63809830e-01 -2.44749524e-02 3.90042514e-01
-3.18158001e-01 5.79074621e-01 -4.39301282e-01 -6.40975237e-01
-8.33900332e-01 -1.07068670e+00 1.22271883e+00 3.69933605e-01
-3.78546625e-01 -1.03070617e+00 4.04056370e-01 6.78906381e-01
3.78397703e-01 -1.27830729e-01 7.80069828e-01 -1.03309143e+00
-1.97883800e-01 -2.07815185e-01 -2.28413437e-02 5.07988155e-01
4.09548968e-01 -2.56112993e-01 -1.11745155e+00 3.99372019e-02
3.25788647e-01 -6.66673630e-02 6.32712245e-01 7.29058012e-02
6.98591053e-01 -1.20494001e-01 -1.70925811e-01 2.79211223e-01
1.36056411e+00 2.43035525e-01 5.85777938e-01 8.01930845e-01
6.29777968e-01 5.96600413e-01 1.27016830e+00 3.43897045e-01
4.41571087e-01 4.10269916e-01 -5.75357787e-02 4.41780947e-02
1.01250477e-01 6.31982787e-03 6.01700246e-01 1.48328114e+00
4.51015448e-03 -1.36170164e-01 -5.88676214e-01 8.71834517e-01
-1.64477205e+00 -7.08135366e-01 -2.55399734e-01 1.94388938e+00
1.14336741e+00 2.71941572e-01 1.67466715e-01 2.03863576e-01
5.92001140e-01 3.72449964e-01 -2.77356267e-01 -7.54166901e-01
-2.56297410e-01 4.36006695e-01 1.51368275e-01 2.08126590e-01
-1.47597361e+00 1.09030974e+00 5.24069452e+00 1.24152184e+00
-7.74897695e-01 1.85111016e-01 3.75007302e-01 4.57853675e-01
-5.69750667e-01 1.16028577e-01 -9.90055263e-01 2.94479042e-01
9.22987700e-01 -8.18760991e-02 -1.31655382e-02 1.23504472e+00
-1.90756038e-01 1.99799356e-03 -6.88100278e-01 7.82831252e-01
2.82052457e-01 -8.44844580e-01 1.70292526e-01 -3.04625124e-01
9.36851859e-01 -2.78451651e-01 -2.67483741e-01 5.71361959e-01
1.64329708e-01 -4.81250525e-01 7.25406557e-02 1.01435803e-01
5.04682958e-01 -1.01466751e+00 1.14542425e+00 2.05667958e-01
-1.45862126e+00 5.63325226e-01 -3.82984281e-01 9.84330773e-02
1.13635607e-01 9.69064713e-01 -6.11257970e-01 9.92866993e-01
6.98234499e-01 6.61449909e-01 -4.29682553e-01 8.51590931e-01
-5.48410356e-01 6.29674256e-01 -6.08416349e-02 -4.90778923e-01
2.12850943e-01 -3.36140335e-01 6.61431968e-01 1.37570894e+00
2.92604685e-01 -1.83146685e-01 -1.29137009e-01 3.38566214e-01
-9.10372064e-02 6.51586056e-01 -5.85424244e-01 -2.95598567e-01
1.23978823e-01 1.70874977e+00 -9.50412273e-01 -4.69052523e-01
-8.14337492e-01 1.24918902e+00 9.79031324e-02 6.82840720e-02
-7.12383330e-01 -9.00485516e-01 8.02748322e-01 -2.25533307e-01
7.46163487e-01 2.81974375e-02 -3.43337238e-01 -1.51606798e+00
3.31561446e-01 -1.18438745e+00 5.61152935e-01 -5.15721321e-01
-1.57407093e+00 1.07563317e+00 -3.64576012e-01 -1.67841947e+00
-2.88439691e-01 -6.54720128e-01 -4.40854847e-01 7.44659662e-01
-1.88540995e+00 -1.29199266e+00 -3.32332253e-02 6.76671505e-01
7.62485981e-01 -2.35491812e-01 9.59997118e-01 4.41154689e-01
-3.22922796e-01 6.15535796e-01 -6.17059059e-02 1.78317457e-01
1.06067359e+00 -1.30369055e+00 4.87308383e-01 9.96939719e-01
1.76074743e-01 7.64144778e-01 5.89571059e-01 -5.72443366e-01
-1.31714571e+00 -1.10365045e+00 1.38573277e+00 -5.56959748e-01
9.13464904e-01 -2.45773017e-01 -7.55773544e-01 5.51959932e-01
4.96498585e-01 -4.73316938e-01 1.15361834e+00 5.07569253e-01
-4.78251219e-01 -1.86880082e-01 -9.79267061e-01 4.35776591e-01
6.72563314e-01 -5.07679045e-01 -9.94891942e-01 -1.18904095e-02
9.13859665e-01 -1.13227211e-01 -9.51412499e-01 4.20032680e-01
3.70215297e-01 -9.24182415e-01 6.64722264e-01 -5.38868904e-01
4.18828696e-01 -6.41554952e-01 -4.20824885e-02 -1.48658943e+00
5.65026402e-02 -4.77810949e-01 2.48026475e-01 1.94335186e+00
6.97156489e-01 -6.77020550e-01 6.49938881e-01 4.74275619e-01
-8.44382122e-03 -5.82666457e-01 -5.50902963e-01 -7.69666970e-01
-1.34902790e-01 -6.28593326e-01 8.72410178e-01 1.04153502e+00
1.48548231e-01 9.40381527e-01 -2.23338768e-01 1.68912709e-01
3.96845549e-01 6.09604001e-01 6.63043916e-01 -9.90635753e-01
-4.90112454e-01 -3.86569947e-01 -3.26908350e-01 -8.77093971e-01
1.14102952e-01 -5.92068195e-01 9.83939469e-02 -1.43193007e+00
3.78175169e-01 -3.32580924e-01 -4.18596953e-01 3.03565413e-01
-5.66464126e-01 1.69960022e-01 -6.84368312e-02 -9.66739133e-02
-8.60424101e-01 6.92883074e-01 1.13753653e+00 -2.31827766e-01
-1.87610447e-01 3.62891912e-01 -1.09430444e+00 1.02187276e+00
9.42843080e-01 -6.46187961e-01 -3.77545565e-01 -2.83221126e-01
4.70062882e-01 -3.88637543e-01 -4.27324325e-01 -7.20662415e-01
7.56250843e-02 -5.87068312e-02 4.02301215e-02 -7.23899782e-01
2.59222329e-01 -9.43899870e-01 -3.60334903e-01 7.52959847e-02
1.17271893e-01 1.35608360e-01 3.22136551e-01 5.09593189e-01
-8.25546384e-01 -4.63052720e-01 3.78287345e-01 -1.90039679e-01
-9.15497661e-01 9.74664241e-02 -5.22362828e-01 3.10364366e-01
8.75519037e-01 2.51133233e-01 -2.15800866e-01 -1.64760694e-01
-4.36768115e-01 -3.86242643e-02 3.02225530e-01 4.34584349e-01
3.77606302e-01 -1.23570645e+00 -4.67254162e-01 9.47018340e-02
5.95297575e-01 -4.99488860e-02 3.38023603e-01 7.17950404e-01
-1.96442962e-01 5.25557637e-01 -3.14977206e-02 -2.94042617e-01
-1.38885355e+00 5.77222586e-01 -2.80838698e-01 -9.22789574e-01
-6.32476732e-02 7.15855718e-01 1.46441773e-01 -8.85295331e-01
-2.85055101e-01 -8.21936801e-02 -6.82015777e-01 3.51667941e-01
6.24109089e-01 -8.89179355e-04 3.96286696e-01 -6.69363797e-01
-6.71925604e-01 8.49035442e-01 -2.67894089e-01 -1.44735724e-01
1.35419881e+00 -2.39083573e-01 -4.12218362e-01 3.53431731e-01
1.10871661e+00 3.39677513e-01 -4.22939450e-01 -2.17493966e-01
1.94624424e-01 -1.71686843e-01 -3.45077485e-01 -7.89072156e-01
-9.76977587e-01 8.71476829e-01 1.70996904e-01 3.17934275e-01
1.42990279e+00 2.43057776e-02 9.73337114e-01 3.89842242e-01
3.36934179e-01 -1.18812454e+00 -2.08555609e-01 6.70826495e-01
4.76689041e-01 -1.45934916e+00 -8.16210657e-02 -6.15516186e-01
-9.37387466e-01 8.67636442e-01 6.22104466e-01 -2.34394029e-04
6.94532156e-01 2.70180166e-01 3.77956510e-01 -1.76138356e-01
-5.40633798e-01 -5.13418198e-01 2.77422935e-01 5.52111745e-01
6.10234797e-01 9.45709795e-02 -8.13059926e-01 1.20862937e+00
-2.23210007e-01 -1.83817804e-01 1.87427685e-01 1.02971148e+00
-2.20409870e-01 -1.73254943e+00 -2.47487798e-01 2.98463166e-01
-8.96064520e-01 -4.32393104e-01 -4.25618947e-01 7.63491452e-01
-2.45341182e-01 1.07094407e+00 -5.05229235e-01 -3.85879248e-01
5.14063299e-01 3.39760870e-01 2.20761657e-01 -6.04275763e-01
-7.68362105e-01 2.65431404e-01 5.08059382e-01 -3.71299267e-01
-8.63618195e-01 -6.32077694e-01 -9.57830846e-01 -7.90491700e-03
-4.12114471e-01 4.98638183e-01 8.69858205e-01 1.12147665e+00
4.53401536e-01 5.33880115e-01 6.83512270e-01 -2.22060248e-01
-1.14069358e-01 -1.10001981e+00 -5.39032757e-01 5.69105983e-01
-4.39204015e-02 -3.54967058e-01 -2.60254174e-01 9.65317488e-02] | [11.372610092163086, 6.726204872131348] |
9f476898-b1c7-48f0-82c0-74c2e89efe1d | on-exploring-node-feature-and-graph-structure-1 | 2306.12726 | null | https://arxiv.org/abs/2306.12726v1 | https://arxiv.org/pdf/2306.12726v1.pdf | On Exploring Node-feature and Graph-structure Diversities for Node Drop Graph Pooling | A pooling operation is essential for effective graph-level representation learning, where the node drop pooling has become one mainstream graph pooling technology. However, current node drop pooling methods usually keep the top-k nodes according to their significance scores, which ignore the graph diversity in terms of the node features and the graph structures, thus resulting in suboptimal graph-level representations. To address the aforementioned issue, we propose a novel plug-and-play score scheme and refer to it as MID, which consists of a \textbf{M}ultidimensional score space with two operations, \textit{i.e.}, fl\textbf{I}pscore and \textbf{D}ropscore. Specifically, the multidimensional score space depicts the significance of nodes through multiple criteria; the flipscore encourages the maintenance of dissimilar node features; and the dropscore forces the model to notice diverse graph structures instead of being stuck in significant local structures. To evaluate the effectiveness of our proposed MID, we perform extensive experiments by applying it to a wide variety of recent node drop pooling methods, including TopKPool, SAGPool, GSAPool, and ASAP. Specifically, the proposed MID can efficiently and consistently achieve about 2.8\% average improvements over the above four methods on seventeen real-world graph classification datasets, including four social datasets (IMDB-BINARY, IMDB-MULTI, REDDIT-BINARY, and COLLAB), and thirteen biochemical datasets (D\&D, PROTEINS, NCI1, MUTAG, PTC-MR, NCI109, ENZYMES, MUTAGENICITY, FRANKENSTEIN, HIV, BBBP, TOXCAST, and TOX21). Code is available at~\url{https://github.com/whuchuang/mid}. | ['Tongliang Liu', 'Wenbin Hu', 'Bo Du', 'Liu Liu', 'Baosheng Yu', 'Yibing Zhan', 'Chuang Liu'] | 2023-06-22 | on-exploring-node-feature-and-graph-structure | https://openreview.net/forum?id=zc0YnpS90ug | https://openreview.net/pdf?id=zc0YnpS90ug | null | ['graph-classification'] | ['graphs'] | [ 1.31447330e-01 1.51189277e-02 -3.82953823e-01 2.79133506e-02
-5.10248125e-01 -5.39234698e-01 2.92887390e-01 7.12944031e-01
-2.30107665e-01 9.48974133e-01 6.93878159e-02 -2.32051432e-01
-2.98075616e-01 -9.21644628e-01 -6.57784939e-01 -8.84530544e-01
-2.97401369e-01 1.19122993e-02 5.40283382e-01 -1.46859840e-01
1.97890282e-01 4.29458380e-01 -9.51268554e-01 2.17589334e-01
8.41177642e-01 8.36305380e-01 1.48294851e-01 4.52889234e-01
-1.17701910e-01 5.44304967e-01 -6.16858840e-01 -5.63715339e-01
-1.84845515e-02 -2.70588309e-01 -5.23285687e-01 -4.69245404e-01
1.44156337e-01 7.62548894e-02 -6.52277529e-01 1.18262613e+00
6.50889277e-01 -8.47323146e-03 7.14052439e-01 -1.38119459e+00
-7.00765610e-01 5.34968913e-01 -9.30832088e-01 3.03361714e-01
3.84252876e-01 1.09701015e-01 1.31229877e+00 -8.08643520e-01
8.51732254e-01 1.19682288e+00 7.13277161e-01 1.71932086e-01
-1.16120410e+00 -7.02547729e-01 3.26261431e-01 5.87179661e-02
-1.51888478e+00 8.40933546e-02 6.62787676e-01 -3.25830281e-01
1.01666248e+00 4.44405556e-01 4.05608714e-01 7.62001216e-01
5.71998835e-01 7.49894261e-01 9.35858607e-01 1.36676773e-01
1.12052776e-01 -2.54575014e-01 6.10373676e-01 1.01056743e+00
5.60587227e-01 -4.15588766e-01 -4.73520279e-01 -4.61429685e-01
5.39619684e-01 3.19282889e-01 -4.76778060e-01 -2.04084978e-01
-1.11357486e+00 6.12274349e-01 9.00105536e-01 3.44877213e-01
-2.49745458e-01 1.80579394e-01 4.23428327e-01 2.32479259e-01
3.72300595e-01 2.61088669e-01 -5.22591650e-01 2.34350160e-01
-3.84377599e-01 2.59859800e-01 5.35348058e-01 8.50183308e-01
8.92739117e-01 -3.14224422e-01 -2.59741336e-01 6.72140956e-01
2.63131142e-01 -1.05845690e-01 6.01784647e-01 -2.30488077e-01
7.05630422e-01 1.15689623e+00 -1.85256422e-01 -1.35258138e+00
-7.83630192e-01 -2.66754985e-01 -1.07624316e+00 -1.84018493e-01
1.62622198e-01 -1.16320655e-01 -1.15086389e+00 1.85061860e+00
1.98367387e-01 2.45141014e-01 -3.41144912e-02 6.36109948e-01
1.40646160e+00 6.66607559e-01 2.47904465e-01 -2.14089513e-01
1.34320450e+00 -7.18940079e-01 -4.81056243e-01 3.48213874e-02
7.11536646e-01 -4.48624223e-01 1.01538181e+00 2.16707930e-01
-9.02586877e-01 -2.40847304e-01 -1.16628015e+00 -6.21217452e-02
-8.57666194e-01 -1.14554673e-01 6.83820248e-01 4.18132961e-01
-1.07266855e+00 7.95685172e-01 -6.26598120e-01 -3.57086062e-01
4.82879519e-01 5.09204090e-01 -5.88597894e-01 -1.78201869e-01
-1.37530816e+00 5.05906522e-01 5.20568669e-01 2.76322081e-03
-4.27889377e-01 -5.40662289e-01 -7.80150056e-01 2.27102593e-01
3.10454458e-01 -4.49528694e-01 3.18854511e-01 -1.31974325e-01
-1.03680408e+00 6.25979483e-01 -9.92187485e-02 -1.12856753e-01
2.01447979e-01 2.35917941e-01 -3.27415705e-01 3.71198691e-02
-7.40902498e-02 4.85703349e-01 4.21346009e-01 -8.90906215e-01
-1.39215112e-01 -5.15232682e-01 5.33005893e-02 2.32429028e-01
-3.89504820e-01 3.59718911e-02 -5.21252155e-01 -7.44791627e-01
2.85834193e-01 -6.41442180e-01 -2.13872313e-01 -2.56109804e-01
-7.19298482e-01 -4.03715700e-01 6.73575044e-01 -6.60626769e-01
1.46387005e+00 -2.21775961e+00 3.99262965e-01 4.12733585e-01
7.08836854e-01 2.68993467e-01 -2.90392339e-01 7.03168631e-01
-1.56206653e-01 5.74429691e-01 -2.26468921e-01 9.48151499e-02
-1.29255250e-01 8.60431790e-02 2.94246167e-01 5.16319811e-01
2.65765220e-01 9.71598089e-01 -8.38139236e-01 -5.09202898e-01
-2.13772841e-02 3.52162063e-01 -4.79460299e-01 -1.80321500e-01
-1.37364075e-01 -3.73105556e-02 -8.46537054e-01 8.59763205e-01
8.27532053e-01 -6.23932064e-01 4.09636527e-01 -5.48327677e-02
8.84805545e-02 4.81444746e-02 -1.14044070e+00 1.37215042e+00
4.04162973e-01 7.35807642e-02 -6.97499812e-02 -7.14002788e-01
9.09833550e-01 3.32963206e-02 4.76405591e-01 -4.26130295e-01
2.11341947e-01 7.06087276e-02 7.07820505e-02 -1.76187269e-02
4.13623691e-01 3.33200723e-01 -2.95562774e-01 1.61636323e-01
8.62737671e-02 7.91726932e-02 3.19406778e-01 5.46551883e-01
1.53999317e+00 -3.11595500e-01 3.00001800e-01 -2.80047596e-01
4.87240225e-01 -2.92477161e-01 7.11044133e-01 5.04750967e-01
-3.94845784e-01 5.95024168e-01 1.31821287e+00 -3.14064533e-01
-4.45053726e-01 -9.01778698e-01 3.51328179e-02 1.03677952e+00
3.97875249e-01 -7.22425938e-01 -5.72774529e-01 -7.98011363e-01
2.45123297e-01 1.86958268e-01 -7.02722609e-01 -4.30640757e-01
-5.49332976e-01 -1.21577334e+00 6.49903297e-01 3.03647548e-01
3.75245392e-01 -1.00866354e+00 6.29938692e-02 1.47283763e-01
3.65004689e-02 -5.82582653e-01 -6.75638974e-01 2.51344889e-01
-6.07709408e-01 -1.26578891e+00 -6.05130792e-01 -8.66016150e-01
7.28038013e-01 1.99184045e-01 8.25602591e-01 3.61096621e-01
-3.72274518e-01 -4.86185282e-01 -2.95103252e-01 -7.14576021e-02
1.65774480e-01 3.20377320e-01 -8.50714818e-02 -2.96204597e-01
2.20317140e-01 -4.93254870e-01 -6.53097212e-01 5.12008011e-01
-8.35188091e-01 -2.23032180e-02 6.35036707e-01 9.49824810e-01
8.97999644e-01 6.86027715e-03 6.61663413e-01 -7.95117974e-01
9.46541846e-01 -7.28756964e-01 -3.04669559e-01 4.13112223e-01
-4.97063458e-01 -1.73265979e-01 6.51539922e-01 -3.74386579e-01
-3.70990932e-01 -2.67031670e-01 -5.87456934e-02 -3.13853204e-01
4.51289684e-01 8.87794971e-01 -4.25594091e-01 -2.24342093e-01
4.34431076e-01 2.99417883e-01 -1.27402022e-01 -3.72585118e-01
1.69692457e-01 3.41162682e-01 1.26642331e-01 -3.88514280e-01
4.63828266e-01 2.61531980e-03 6.59204945e-02 -6.10422492e-01
-1.85474500e-01 -2.48569772e-01 -2.76061445e-01 5.85215352e-02
7.51065433e-01 -7.40081966e-01 -8.63314569e-01 6.93261504e-01
-8.30725491e-01 -1.12858199e-01 9.79302526e-02 1.98254466e-01
-1.81850363e-02 5.70534408e-01 -8.46587002e-01 -3.24544132e-01
-4.47967440e-01 -1.22985375e+00 6.77919209e-01 4.04016197e-01
5.96727654e-02 -6.30006373e-01 -1.31577864e-01 7.78093282e-03
1.63914278e-01 7.48267949e-01 1.51280642e+00 -8.13686848e-01
-5.40777862e-01 -4.10207868e-01 -6.36624694e-01 7.46098086e-02
2.88155735e-01 1.05008051e-01 -5.40362895e-01 -5.33084214e-01
-6.93769753e-01 -3.09829324e-01 1.01974034e+00 2.34113753e-01
1.30389750e+00 -2.90368259e-01 -6.63742721e-01 5.85609674e-01
1.40138757e+00 2.12147430e-01 5.96026659e-01 -7.18761012e-02
8.27624381e-01 1.85018316e-01 2.49666691e-01 2.38921732e-01
3.87159675e-01 4.26848829e-01 5.38091183e-01 -8.30097198e-02
-4.41797487e-02 -2.97551662e-01 1.76067933e-01 8.02847683e-01
-6.61699325e-02 -7.26402760e-01 -7.93347061e-01 3.52070093e-01
-1.79730487e+00 -5.93450069e-01 -8.92191380e-02 1.99060261e+00
7.06999719e-01 2.76259124e-01 3.36500034e-02 2.07288377e-03
7.57743657e-01 4.29757446e-01 -8.07903349e-01 -2.84229726e-01
-2.54536182e-01 1.29741609e-01 5.90605795e-01 1.97536781e-01
-9.20901120e-01 9.46215510e-01 5.27280092e+00 9.64420557e-01
-1.00531721e+00 -7.94099346e-02 7.95865357e-01 3.94610241e-02
-2.69077808e-01 -2.15793356e-01 -6.44684017e-01 7.34816194e-01
4.98576880e-01 -3.31211299e-01 2.71523923e-01 5.93880594e-01
-2.80626826e-02 7.80960247e-02 -7.96913147e-01 6.57068729e-01
-1.07963197e-01 -1.43969882e+00 9.19854566e-02 1.19266219e-01
5.32527745e-01 2.12872088e-01 -2.65248436e-02 3.62423390e-01
4.47537303e-01 -1.09462535e+00 1.44218534e-01 4.28781122e-01
6.88913226e-01 -7.65064597e-01 7.78066218e-01 9.58357304e-02
-1.46901608e+00 1.38399422e-01 -4.35329765e-01 3.07947189e-01
-2.19040230e-01 4.67828244e-01 -5.90688407e-01 1.02211285e+00
7.54444301e-01 6.94763780e-01 -7.46562481e-01 8.96641016e-01
-3.07147093e-02 3.05477142e-01 -2.85611242e-01 -5.00631273e-01
3.24047029e-01 -1.44967372e-02 4.69179392e-01 1.05076993e+00
1.19561531e-01 2.25332260e-01 4.23111498e-01 5.36599219e-01
-4.99197125e-01 3.11635584e-01 -2.49746650e-01 -3.01329941e-01
3.73639494e-01 1.26599669e+00 -8.07660282e-01 -1.24215662e-01
-3.87159169e-01 6.33279204e-01 5.71441293e-01 4.40669507e-01
-7.83167720e-01 -9.02102947e-01 7.79161870e-01 1.89036690e-02
2.73999810e-01 -6.19963519e-02 -1.13998540e-01 -9.56563354e-01
6.69403076e-02 -7.53044963e-01 7.14097142e-01 -5.52761912e-01
-1.30886674e+00 5.89078009e-01 3.98323312e-03 -7.68887997e-01
4.71411586e-01 -7.51675665e-01 -7.84538448e-01 9.47723448e-01
-1.34494472e+00 -9.42248821e-01 -3.78118277e-01 5.94003618e-01
8.21822435e-02 -4.27829139e-02 6.00892186e-01 3.07568014e-01
-1.07710743e+00 8.25302958e-01 1.99745268e-01 1.09789379e-01
5.02106786e-01 -1.08174038e+00 5.02640843e-01 4.65218246e-01
-3.48932445e-01 7.76386738e-01 4.38758194e-01 -9.22270060e-01
-1.43351901e+00 -1.13922513e+00 6.63661480e-01 7.80800125e-04
6.46366298e-01 -5.74387074e-01 -1.14967883e+00 4.53212082e-01
-1.17213890e-01 1.57120705e-01 6.27394497e-01 -1.50140747e-01
-3.32922488e-01 5.25779463e-02 -1.46169317e+00 6.61279261e-01
1.20996654e+00 -9.66646895e-02 -2.13304058e-01 6.12776399e-01
1.01065862e+00 -3.54364157e-01 -1.21521807e+00 4.76703405e-01
4.46480751e-01 -6.89065278e-01 9.06138897e-01 -7.76861966e-01
1.95238978e-01 -3.90846163e-01 -9.35134888e-02 -1.14793134e+00
-6.01558149e-01 -6.08778179e-01 -1.28996849e-01 1.04327023e+00
6.00230038e-01 -1.13232291e+00 8.39103878e-01 2.47297242e-01
-1.58983618e-01 -1.21868527e+00 -1.15762591e+00 -5.19923568e-01
4.51587066e-02 1.65598303e-01 1.02751684e+00 9.67461109e-01
2.45809823e-01 -4.28420817e-03 -1.31048456e-01 1.59047186e-01
2.95948029e-01 -1.22136228e-01 6.16120517e-01 -1.24686480e+00
-1.60216600e-01 -6.05782986e-01 -7.22399175e-01 -6.79436207e-01
-1.51626587e-01 -1.16562879e+00 -2.97969162e-01 -1.63760805e+00
3.67570311e-01 -3.78887177e-01 -9.04213190e-01 8.56752813e-01
-4.52396691e-01 8.31207857e-02 2.85427668e-03 2.26377845e-01
-4.90286529e-01 3.70209813e-01 1.32347608e+00 -1.79051206e-01
-2.00611949e-01 -3.60449076e-01 -9.20630574e-01 4.63711619e-01
8.97082925e-01 -3.83510053e-01 -2.89766639e-01 -2.29601771e-01
2.27325127e-01 -4.79982048e-02 2.28346661e-01 -6.96257114e-01
1.19376153e-01 -1.91201627e-01 5.22036850e-01 -5.40002227e-01
3.59564215e-01 -2.05474570e-01 1.83555692e-01 7.79387772e-01
-1.00506537e-01 2.87328809e-01 1.89100757e-01 7.00532854e-01
-1.69180389e-02 7.22547174e-02 6.72714412e-01 -2.07813576e-01
-6.03958547e-01 7.73368359e-01 5.61443111e-03 -6.49246722e-02
1.05061448e+00 -2.61479616e-01 -9.26505625e-01 6.09908402e-02
-5.94638884e-01 5.54808319e-01 3.70917648e-01 4.30377632e-01
6.56466663e-01 -1.15878046e+00 -6.74365938e-01 2.94367671e-01
3.59474391e-01 -3.96631099e-02 4.11154956e-01 8.29689205e-01
-5.00715137e-01 9.93822366e-02 -1.36757135e-01 -3.97528887e-01
-1.27589130e+00 4.58107412e-01 2.18133494e-01 -5.96928954e-01
-4.49792564e-01 1.10434496e+00 2.84299552e-01 -4.09848034e-01
1.45061284e-01 -4.54147160e-01 -2.64259875e-01 4.02592793e-02
1.77174613e-01 3.60121518e-01 1.35544866e-01 -4.50427562e-01
-6.39582038e-01 3.84535551e-01 -4.96778578e-01 5.78717530e-01
1.22630322e+00 3.67348224e-01 -3.65307510e-01 -3.24082263e-02
1.18393576e+00 -1.10075951e-01 -9.18233514e-01 -1.14769600e-01
-2.31629461e-02 -3.43665242e-01 -2.50416875e-01 -6.29738450e-01
-1.03793895e+00 3.92466336e-01 3.54913175e-01 3.53830636e-01
1.00945818e+00 1.55149877e-01 6.06896698e-01 2.56077468e-01
3.89931172e-01 -7.90846348e-01 7.04297200e-02 2.76860267e-01
8.87604773e-01 -8.54202628e-01 2.98279852e-01 -6.19963706e-01
-5.25761425e-01 6.31017327e-01 6.18380249e-01 -1.77236602e-01
7.68134415e-01 -6.24150261e-02 -5.36597311e-01 -4.18619394e-01
-6.60297394e-01 7.76726231e-02 1.39003605e-01 2.27835372e-01
3.72839212e-01 2.62716174e-01 -6.97955191e-01 7.63513803e-01
4.13459018e-02 -1.91763312e-01 2.62544215e-01 9.41601813e-01
-4.48739827e-01 -1.02036333e+00 1.38453022e-01 8.53922367e-01
-4.76480722e-01 -1.46002397e-01 -8.22597027e-01 9.56541061e-01
7.54762366e-02 6.52033627e-01 -3.27600509e-01 -6.62725806e-01
3.98461580e-01 1.46281975e-03 6.30559996e-02 -4.22878653e-01
-7.69158721e-01 -1.36547208e-01 7.07793161e-02 -4.39353317e-01
1.39741689e-01 -5.34960449e-01 -1.40912282e+00 -3.92161161e-01
-4.66218948e-01 9.77589116e-02 2.93606043e-01 3.89560252e-01
5.55146515e-01 6.53783321e-01 5.74597895e-01 -4.67768282e-01
-3.08829695e-01 -1.05736244e+00 -6.44407868e-01 2.06688017e-01
1.87263340e-02 -7.37899423e-01 -3.15777659e-01 -7.15196371e-01] | [7.080223560333252, 6.2690558433532715] |
1c37ae94-6250-49a4-81ba-939b5b6d1948 | improving-state-of-the-art-in-one-class-1 | 2203.07206 | null | https://arxiv.org/abs/2203.07206v1 | https://arxiv.org/pdf/2203.07206v1.pdf | Improving State-of-the-Art in One-Class Classification by Leveraging Unlabeled Data | When dealing with binary classification of data with only one labeled class data scientists employ two main approaches, namely One-Class (OC) classification and Positive Unlabeled (PU) learning. The former only learns from labeled positive data, whereas the latter also utilizes unlabeled data to improve the overall performance. Since PU learning utilizes more data, we might be prone to think that when unlabeled data is available, the go-to algorithms should always come from the PU group. However, we find that this is not always the case if unlabeled data is unreliable, i.e. contains limited or biased latent negative data. We perform an extensive experimental study of a wide list of state-of-the-art OC and PU algorithms in various scenarios as far as unlabeled data reliability is concerned. Furthermore, we propose PU modifications of state-of-the-art OC algorithms that are robust to unreliable unlabeled data, as well as a guideline to similarly modify other OC algorithms. Our main practical recommendation is to use state-of-the-art PU algorithms when unlabeled data is reliable and to use the proposed modifications of state-of-the-art OC algorithms otherwise. Additionally, we outline procedures to distinguish the cases of reliable and unreliable unlabeled data using statistical tests. | ['Aleksei Shpilman', 'Dmitry Ivanov', 'Farid Bagirov'] | 2022-03-14 | improving-state-of-the-art-in-one-class | https://openreview.net/forum?id=4KOJ5XJ_z5W | https://openreview.net/pdf?id=4KOJ5XJ_z5W | null | ['one-class-classification'] | ['miscellaneous'] | [ 1.22576416e-01 2.88622588e-01 -6.84754252e-01 -4.47131306e-01
-1.10320508e+00 -4.93980318e-01 3.13162506e-01 1.15862399e-01
-2.98192561e-01 1.27092969e+00 -4.78409290e-01 -4.36117709e-01
6.14180490e-02 -6.02524042e-01 -5.36573887e-01 -1.13187611e+00
1.30477384e-01 8.12078893e-01 3.35298687e-01 2.14695483e-01
3.04828703e-01 3.99543166e-01 -1.91466773e+00 1.09143116e-01
9.15521085e-01 1.25170648e+00 -4.41672295e-01 6.70382202e-01
-2.06297934e-01 8.47162187e-01 -6.30473316e-01 -4.47064579e-01
2.88414896e-01 -2.98804045e-01 -8.44435871e-01 3.84061396e-01
2.47892991e-01 6.11762851e-02 3.27630222e-01 9.79549289e-01
4.26756531e-01 -3.68214585e-02 1.22497582e+00 -1.53444052e+00
-7.48133600e-01 3.33993375e-01 -7.02964008e-01 2.23389044e-02
8.99121985e-02 -8.29303712e-02 7.32268929e-01 -1.05707371e+00
5.55952847e-01 1.11394370e+00 4.45239216e-01 7.04300225e-01
-9.62390006e-01 -5.43693066e-01 2.38380298e-01 2.47011974e-01
-1.40585756e+00 -4.13914353e-01 6.94265962e-01 -3.28885257e-01
4.41444069e-01 1.52063057e-01 3.37037534e-01 1.09898829e+00
4.92112115e-02 1.24414837e+00 1.48800373e+00 -9.36970949e-01
7.20896542e-01 3.22207272e-01 7.83473015e-01 6.29066885e-01
6.41280770e-01 1.89127117e-01 -4.94002283e-01 -2.41556734e-01
7.74880275e-02 -2.53184199e-01 1.12898886e-01 -5.13262451e-01
-7.78269708e-01 6.92855358e-01 -7.32595623e-02 3.52560401e-01
2.88685173e-04 -2.38173589e-01 1.06003009e-01 4.85191077e-01
7.73088336e-01 1.64274037e-01 -5.95199287e-01 1.11270994e-01
-8.75084460e-01 -1.65808827e-01 9.40974355e-01 9.98555005e-01
9.78708029e-01 -2.62057818e-02 -2.20225528e-02 8.82535875e-01
4.22224104e-01 4.97335464e-01 6.88764453e-01 -6.99801803e-01
3.05417359e-01 6.33105814e-01 1.23272590e-01 -1.66538909e-01
-9.21073332e-02 -1.80513278e-01 -3.83382976e-01 2.47493729e-01
6.02557302e-01 -1.93281576e-01 -1.36884356e+00 1.31796074e+00
1.95859715e-01 -1.74928963e-01 3.49387199e-01 6.15994453e-01
6.24792576e-01 3.36446911e-01 1.46614285e-02 -6.96970999e-01
8.14742386e-01 -1.08963954e+00 -7.43701518e-01 -2.05463648e-01
8.54957700e-01 -8.09371769e-01 8.36627543e-01 7.26848543e-01
-6.49171829e-01 -5.32754302e-01 -1.24667394e+00 1.45106083e-02
-7.17998207e-01 4.46706563e-01 5.77234209e-01 1.19055939e+00
-9.33049500e-01 4.55156654e-01 -6.90227687e-01 -5.33792138e-01
1.14808142e-01 6.05253160e-01 -2.70719588e-01 -3.04274708e-01
-8.63756776e-01 9.09816623e-01 6.00951195e-01 6.98926970e-02
-7.20316470e-01 3.68986502e-02 -7.13815331e-01 -1.86088160e-01
6.73683882e-01 -4.88119610e-02 1.32962775e+00 -1.19959164e+00
-1.17899156e+00 7.84489512e-01 -3.50633383e-01 -3.92765068e-02
6.61590576e-01 -1.49290934e-02 -7.07410097e-01 1.24717958e-01
3.15128043e-02 6.97387755e-01 9.63619888e-01 -1.83044326e+00
-8.09377015e-01 -5.41917741e-01 -4.09101665e-01 1.03139624e-01
-3.22994858e-01 -2.62414724e-01 -6.14083171e-01 -5.18271685e-01
5.19327819e-01 -9.99550760e-01 3.42164367e-01 2.09356129e-01
-5.81250906e-01 -6.54909909e-01 1.35943556e+00 -1.72615990e-01
1.07040131e+00 -2.13701439e+00 -3.39875698e-01 4.62583423e-01
-1.48099825e-01 1.92712396e-01 -1.19925942e-02 1.15136094e-01
-1.76288694e-01 3.57277125e-01 -2.66152889e-01 -7.00586855e-01
-2.56986737e-01 4.96528327e-01 -8.05657357e-02 7.32511342e-01
4.32379812e-01 4.56604242e-01 -9.44869578e-01 -9.54709947e-01
2.82010883e-02 1.54531456e-03 2.53053084e-02 1.99745134e-01
-1.93385720e-01 4.38291550e-01 -5.20562172e-01 1.41244149e+00
5.58733821e-01 -2.81745791e-01 9.08364877e-02 1.80735081e-01
-6.16812669e-02 -1.92340419e-01 -1.49539053e+00 7.91180849e-01
-7.87219703e-02 4.41288203e-01 -3.04051757e-01 -1.07713830e+00
1.03030825e+00 5.15532315e-01 2.37761572e-01 -3.28803748e-01
2.01341301e-01 6.25171244e-01 -9.10704508e-02 -3.13495427e-01
5.75742185e-01 -2.52268493e-01 2.55920351e-01 6.70115590e-01
2.70222366e-01 -8.02520961e-02 5.71392894e-01 -3.63667831e-02
7.25547671e-01 3.14168721e-01 3.79392117e-01 -2.10296467e-01
5.57132363e-01 -1.15228169e-01 4.35147166e-01 7.84763336e-01
-5.16611516e-01 7.62311459e-01 6.81156874e-01 -4.23704199e-02
-7.23576188e-01 -9.42663729e-01 -3.41364354e-01 1.04498053e+00
2.51507372e-01 5.66132255e-02 -4.68488276e-01 -1.24226594e+00
-1.24865934e-01 5.32130837e-01 -8.66320074e-01 -4.66671214e-02
-4.88533340e-02 -9.57708538e-01 4.19651754e-02 5.06613255e-01
1.99159786e-01 -9.14289474e-01 7.79845789e-02 -1.32583648e-01
-5.54720648e-02 -7.15856850e-01 1.51311398e-01 6.81954861e-01
-1.08956110e+00 -1.46007776e+00 -6.44793630e-01 -8.27340245e-01
1.12937272e+00 3.81763428e-01 9.27352846e-01 4.26921815e-01
3.86811465e-01 3.95400643e-01 -7.99297512e-01 -5.88361323e-01
-5.94231725e-01 1.28797442e-01 3.20428610e-01 2.85978228e-01
6.60558462e-01 6.77611157e-02 -1.55375719e-01 6.85175896e-01
-1.02985430e+00 -4.74116743e-01 6.53622687e-01 9.65541363e-01
7.37732351e-01 9.21731517e-02 6.60672009e-01 -1.14496863e+00
1.47707656e-01 -5.33448935e-01 -5.07066548e-01 7.51335084e-01
-9.86510456e-01 2.76187271e-01 5.42240739e-01 -6.34517789e-01
-1.12306607e+00 2.28788003e-01 9.33062360e-02 -4.46342349e-01
-2.46888399e-01 3.15028697e-01 -9.48382318e-02 -2.39724740e-01
8.92269790e-01 -1.68860272e-01 -3.41771573e-01 -4.54185724e-01
2.12410003e-01 1.20761001e+00 3.57595742e-01 -7.49352336e-01
6.48338854e-01 5.82387269e-01 -2.09037066e-01 -8.82705629e-01
-9.96116042e-01 -7.90496409e-01 -7.78518558e-01 -4.65674669e-01
4.82456982e-01 -5.13299286e-01 -2.91555613e-01 5.00114322e-01
-7.16661811e-01 -1.92722484e-01 -2.73388207e-01 3.98229450e-01
-4.73322988e-01 7.29671657e-01 -4.96282607e-01 -1.26741064e+00
-7.59453848e-02 -1.30760515e+00 1.02953351e+00 2.53077894e-01
1.47560649e-02 -8.55708957e-01 -2.34693084e-02 3.41492623e-01
-3.01841766e-01 -1.79776609e-01 8.30564201e-01 -1.06521988e+00
-2.22128302e-01 -5.49302936e-01 -8.06432664e-02 6.40334785e-01
8.99027959e-02 3.95862937e-01 -1.51071060e+00 -3.78028244e-01
-1.21641234e-01 -8.09636712e-01 9.33983386e-01 1.62999302e-01
1.10243189e+00 -8.98860674e-03 -3.71620893e-01 -6.80544525e-02
1.25112391e+00 2.59475768e-01 5.92382967e-01 1.09308744e-02
4.12256747e-01 8.12244356e-01 1.22160566e+00 2.03595728e-01
5.94728813e-02 2.37738550e-01 3.32181394e-01 -6.62658438e-02
1.03690639e-01 -1.05465546e-01 3.26909333e-01 7.33813524e-01
-1.68527395e-01 -4.72522408e-01 -7.96317458e-01 4.70946848e-01
-1.76548934e+00 -6.18890107e-01 -2.42273450e-01 2.58443189e+00
8.57046008e-01 4.53111649e-01 4.48523574e-02 8.50892961e-01
9.82755363e-01 -1.41939864e-01 -4.01584983e-01 -1.21594541e-01
-9.71901193e-02 6.75931275e-02 5.55448949e-01 2.84943759e-01
-1.43844867e+00 6.03793621e-01 6.85026026e+00 1.09304821e+00
-1.24398005e+00 1.56908423e-01 1.14565563e+00 2.81157523e-01
1.22311143e-02 9.03355405e-02 -9.77778018e-01 3.16350549e-01
8.93306792e-01 2.99530119e-01 -1.07083783e-01 1.16370881e+00
-2.09344849e-01 -7.94830501e-01 -1.23835742e+00 7.55694926e-01
2.04375073e-01 -6.90871835e-01 -2.00315356e-01 -2.44130567e-02
1.07878280e+00 -1.80139422e-01 -1.14124447e-01 4.39749330e-01
2.77882636e-01 -6.16029799e-01 8.72512698e-01 2.23670840e-01
7.68535912e-01 -5.90445101e-01 1.11660719e+00 4.59758222e-01
-8.77856910e-01 -1.66845739e-01 -4.48962539e-01 2.17443526e-01
-2.67673194e-01 6.95994437e-01 -4.85896498e-01 5.79423964e-01
5.04785240e-01 7.07653761e-01 -1.02832520e+00 1.06738687e+00
-4.25416023e-01 6.53975427e-01 -2.31200650e-01 -7.47144297e-02
1.07180187e-02 -8.42635408e-02 1.30060855e-02 7.60969341e-01
2.75492162e-01 9.79559720e-02 1.79993868e-01 1.84447631e-01
6.00649267e-02 3.21833283e-01 -6.32320762e-01 -7.53299594e-02
2.65651613e-01 1.12328911e+00 -1.07922339e+00 -8.04617047e-01
-5.64260423e-01 5.84825754e-01 2.73418874e-01 3.40927482e-01
-3.27879280e-01 1.02221705e-01 -3.21683735e-01 -3.30560952e-01
9.84736308e-02 3.95169929e-02 -4.65838641e-01 -1.16587651e+00
-4.48477119e-02 -6.98373795e-01 7.97823370e-01 -8.85305882e-01
-1.64967191e+00 4.15280998e-01 -1.45524843e-02 -1.75128675e+00
-3.35914761e-01 -7.87909210e-01 -1.97597250e-01 4.84281182e-01
-1.60188138e+00 -1.00307047e+00 1.51816113e-02 5.82386732e-01
5.17434001e-01 -1.02356538e-01 6.28993690e-01 4.75163050e-02
-4.87445593e-01 5.31407952e-01 3.31370085e-01 3.37616727e-02
1.13946235e+00 -1.15796244e+00 -4.53508407e-01 1.04579139e+00
9.39104930e-02 4.09157813e-01 5.37963271e-01 -7.03903496e-01
-1.14301372e+00 -8.79481435e-01 9.64264631e-01 -4.53439236e-01
5.22762656e-01 -5.30980378e-02 -8.95514846e-01 4.90078270e-01
-3.76010418e-01 4.58788395e-01 8.58966470e-01 1.12505101e-01
-4.01239634e-01 -2.67192926e-02 -1.31158316e+00 4.42768395e-01
4.71259475e-01 -3.41154248e-01 -5.66395760e-01 4.15659249e-01
2.53864616e-01 -9.31325331e-02 -5.44672191e-01 6.23716772e-01
5.43429971e-01 -9.85600054e-01 3.50181967e-01 -4.25901473e-01
2.10542604e-01 -3.91901344e-01 -1.34004354e-01 -1.01770735e+00
2.65279859e-01 -1.61662344e-02 -4.11552340e-01 1.00436091e+00
7.77859390e-01 -6.98426306e-01 1.06007719e+00 7.73157179e-01
-2.38134079e-02 -4.79593605e-01 -8.74728382e-01 -9.89764869e-01
-8.69023576e-02 -6.94547117e-01 -6.13156818e-02 9.02454436e-01
-5.49937375e-02 4.96127531e-02 -3.70347112e-01 4.62478772e-02
4.97136772e-01 4.48906273e-02 4.31580961e-01 -1.48233521e+00
-2.02334821e-01 -6.24162592e-02 -2.37122461e-01 -6.41015887e-01
2.17249602e-01 -4.24981654e-01 2.83414900e-01 -1.25667489e+00
2.71163523e-01 -6.12143099e-01 -3.96729916e-01 7.79881537e-01
-2.45863855e-01 7.67940283e-01 1.45153210e-01 5.58946252e-01
-6.48553133e-01 2.79337764e-01 1.04952157e+00 -2.67732769e-01
-2.87430793e-01 3.81432474e-01 -3.64358604e-01 6.34379864e-01
6.54249907e-01 -6.46095932e-01 -3.77831578e-01 1.43861115e-01
2.65512615e-02 3.31532996e-04 -1.14957340e-01 -9.71418381e-01
4.96961363e-02 -1.35078788e-01 5.03310084e-01 -7.84785032e-01
2.79740721e-01 -1.01307952e+00 -2.88175344e-01 3.69365722e-01
-2.77385145e-01 -4.26741570e-01 -2.32620001e-01 7.45517850e-01
-2.98952758e-01 -7.58850098e-01 1.03399050e+00 -4.80688643e-03
-7.54637301e-01 1.63550600e-01 -4.94974285e-01 -1.76682442e-01
1.20974302e+00 -4.50286150e-01 -5.33917904e-01 -3.78230691e-01
-8.94641697e-01 1.06973834e-01 6.08327866e-01 3.34346712e-01
2.81149983e-01 -1.11210597e+00 -3.02414626e-01 1.81932747e-01
5.03641188e-01 -1.80582225e-01 -8.92972052e-02 8.04838479e-01
-3.53223145e-01 4.13859904e-01 1.04194194e-01 -6.63466632e-01
-1.32906497e+00 8.79469395e-01 7.29246512e-02 -2.15884909e-01
1.41276464e-01 4.30550724e-01 -2.67564058e-01 -4.37186748e-01
4.64512318e-01 -6.12310544e-02 -4.15354162e-01 4.01761919e-01
3.32432210e-01 5.05412519e-01 5.34516633e-01 -5.99084139e-01
-1.21773943e-01 4.13876206e-01 1.37889147e-01 -2.32214183e-01
7.52463937e-01 -9.90981013e-02 -3.10432352e-02 9.13567781e-01
9.54016209e-01 -9.48512405e-02 -1.00929332e+00 -2.22584382e-01
2.08484754e-01 -3.50506395e-01 -2.30372265e-01 -7.28585184e-01
-1.09294355e+00 9.49232578e-01 8.28648686e-01 3.32582265e-01
1.09644532e+00 1.23219699e-01 -7.61076957e-02 6.66809916e-01
7.35969067e-01 -1.51107514e+00 1.71897486e-01 3.28015029e-01
5.06925106e-01 -1.70981300e+00 2.91464388e-01 -7.00058937e-01
-7.58054018e-01 1.27604139e+00 7.93490112e-01 1.94771022e-01
7.94756055e-01 2.53930432e-03 3.15342426e-01 1.52774274e-01
-6.95323050e-01 -4.62134987e-01 3.98965180e-01 6.62426591e-01
4.90556955e-01 -8.34926590e-03 -7.81913042e-01 2.53551006e-01
3.67257982e-01 1.30393565e-01 7.25235224e-01 1.46157897e+00
-3.85809302e-01 -1.43472338e+00 -8.70040178e-01 7.82506168e-01
-3.78696948e-01 1.57720879e-01 -4.74167317e-01 1.00304997e+00
1.21492580e-01 1.39788198e+00 -1.18644625e-01 -1.83309570e-01
-1.61674142e-01 5.71377754e-01 1.91954285e-01 -7.17616558e-01
-1.64164349e-01 2.65482545e-01 1.46187901e-01 -1.13798134e-01
-9.47276115e-01 -5.85065782e-01 -1.17376542e+00 -3.36216018e-02
-8.24970186e-01 1.32332072e-01 5.75473785e-01 1.22654939e+00
-2.02108160e-01 -1.20885037e-01 7.56738305e-01 -7.75093675e-01
-5.51717758e-01 -1.12822044e+00 -9.30507600e-01 1.21832088e-01
1.88607886e-01 -9.75518823e-01 -6.61593914e-01 1.94896162e-01] | [9.30871868133545, 3.9140524864196777] |
85182508-1598-4f6d-9648-ab4632fd6349 | detector-free-structure-from-motion | 2306.15669 | null | https://arxiv.org/abs/2306.15669v1 | https://arxiv.org/pdf/2306.15669v1.pdf | Detector-Free Structure from Motion | We propose a new structure-from-motion framework to recover accurate camera poses and point clouds from unordered images. Traditional SfM systems typically rely on the successful detection of repeatable keypoints across multiple views as the first step, which is difficult for texture-poor scenes, and poor keypoint detection may break down the whole SfM system. We propose a new detector-free SfM framework to draw benefits from the recent success of detector-free matchers to avoid the early determination of keypoints, while solving the multi-view inconsistency issue of detector-free matchers. Specifically, our framework first reconstructs a coarse SfM model from quantized detector-free matches. Then, it refines the model by a novel iterative refinement pipeline, which iterates between an attention-based multi-view matching module to refine feature tracks and a geometry refinement module to improve the reconstruction accuracy. Experiments demonstrate that the proposed framework outperforms existing detector-based SfM systems on common benchmark datasets. We also collect a texture-poor SfM dataset to demonstrate the capability of our framework to reconstruct texture-poor scenes. Based on this framework, we take $\textit{first place}$ in Image Matching Challenge 2023. | ['Xiaowei Zhou', 'Hujun Bao', 'QiXing Huang', 'Sida Peng', 'Yifan Wang', 'Jiaming Sun', 'Xingyi He'] | 2023-06-27 | null | null | null | null | ['keypoint-detection'] | ['computer-vision'] | [ 8.63265023e-02 -5.09926796e-01 8.62946361e-02 -2.86282897e-01
-9.59852159e-01 -5.36122978e-01 4.90893543e-01 -1.13181427e-01
-1.46171033e-01 6.08740896e-02 2.66123731e-02 -5.32847978e-02
-1.50946258e-02 -7.55795121e-01 -8.45444262e-01 -3.91393512e-01
5.26884139e-01 7.06644118e-01 8.32080960e-01 -2.35331357e-01
5.92063844e-01 7.23791480e-01 -1.56250703e+00 7.62751400e-02
7.65180051e-01 9.88430262e-01 5.34632802e-01 3.76782119e-01
2.04121228e-02 4.40693319e-01 2.96869241e-02 -1.92331821e-01
5.85232496e-01 3.66682708e-02 -4.67897266e-01 2.07147464e-01
9.41189945e-01 -7.03843236e-01 -3.79008591e-01 1.15396214e+00
2.52338052e-01 8.64027664e-02 3.56453031e-01 -1.12027645e+00
-1.65142804e-01 -1.27662033e-01 -9.71444309e-01 -8.62710774e-02
7.19371319e-01 2.67086783e-03 9.45650220e-01 -1.48345912e+00
6.59364641e-01 1.26645792e+00 7.53366053e-01 1.28985569e-01
-9.70866621e-01 -8.93320918e-01 2.70907789e-01 1.41326010e-01
-1.71130919e+00 -7.14374006e-01 7.67166674e-01 -3.60830575e-01
9.34224427e-01 2.24104494e-01 7.09461451e-01 4.85730857e-01
9.78511423e-02 4.87840623e-01 7.26794004e-01 -2.31967881e-01
1.00905951e-02 -4.30944294e-01 -1.03591286e-01 7.57116556e-01
2.21475974e-01 2.67463833e-01 -7.35731721e-01 -1.58409312e-01
1.22732723e+00 4.66399878e-01 -1.99514851e-01 -8.74065638e-01
-1.58635700e+00 5.57303429e-01 6.28807664e-01 -2.48608693e-01
-3.51404876e-01 6.63539842e-02 -1.39133990e-01 1.62895545e-01
3.30225646e-01 1.42776340e-01 -2.52676308e-01 8.28321502e-02
-1.12831903e+00 3.54458123e-01 3.49692345e-01 1.21708071e+00
1.11703181e+00 -4.58954483e-01 1.46754861e-01 6.02731764e-01
4.56839651e-01 7.27518022e-01 -4.36296351e-02 -1.08418345e+00
5.89436412e-01 9.01895046e-01 2.96986401e-01 -1.35671258e+00
-3.67877126e-01 -1.42123491e-01 -8.73149276e-01 1.80575609e-01
5.32903858e-02 4.04645860e-01 -8.47573996e-01 1.16759980e+00
8.24081957e-01 3.22036058e-01 -3.63217652e-01 1.10716081e+00
7.04089284e-01 4.56240445e-01 -6.65575683e-01 5.17357886e-02
1.02085161e+00 -1.07539129e+00 -2.69988567e-01 -1.89933777e-01
3.06837708e-01 -1.11981356e+00 7.07633138e-01 3.91018063e-01
-9.33951735e-01 -5.47456980e-01 -1.10820866e+00 -2.40948677e-01
1.89851359e-01 -1.18554063e-01 3.91732395e-01 6.69796169e-02
-9.27259803e-01 3.94339055e-01 -8.80558670e-01 -2.87666321e-01
3.85980695e-01 3.35930496e-01 -5.64253807e-01 -5.61056376e-01
-5.68708062e-01 6.59661472e-01 2.06546679e-01 1.62260368e-01
-7.13097513e-01 -6.62585795e-01 -8.94538462e-01 -1.89659104e-01
7.10458934e-01 -1.06228900e+00 1.15337670e+00 -4.13835555e-01
-1.10583234e+00 7.17952073e-01 -5.63901484e-01 4.69383709e-02
5.60855269e-01 -3.55664819e-01 -8.05642307e-02 2.71590650e-01
4.29549545e-01 4.58193600e-01 8.94421160e-01 -1.37663901e+00
-1.03264463e+00 -4.45006788e-01 1.68913361e-02 4.35333371e-01
2.59314775e-01 -2.19372734e-01 -9.52962041e-01 -3.69696826e-01
1.01397610e+00 -8.82046819e-01 -3.61655593e-01 2.29882002e-01
-4.06653821e-01 7.53172785e-02 7.92892039e-01 -3.35868239e-01
9.93520677e-01 -2.16339207e+00 1.92377225e-01 4.06553358e-01
3.60552639e-01 -1.02071390e-01 -6.11137273e-03 3.11095804e-01
1.30817056e-01 -3.75066549e-01 1.72471106e-02 -6.03838265e-01
-2.10431129e-01 2.37811223e-01 -1.54722482e-01 6.94452167e-01
1.97409540e-02 6.81971192e-01 -8.37984145e-01 -4.76164818e-01
6.74225211e-01 3.45055580e-01 -7.38701284e-01 8.20246711e-02
-8.99778530e-02 4.83585894e-01 -4.96975332e-01 9.82199013e-01
1.13048804e+00 -5.05596459e-01 -2.66109914e-01 -4.97228861e-01
-4.29882884e-01 4.48358655e-02 -1.77389264e+00 2.22454333e+00
-2.05045685e-01 4.01839241e-02 6.39855936e-02 -5.83233654e-01
7.89789319e-01 -5.75150289e-02 6.38379693e-01 -5.88775039e-01
-1.52104124e-01 3.45475167e-01 -4.13442641e-01 5.50615415e-02
7.53609002e-01 1.06716372e-01 5.74351810e-02 1.72225177e-01
-2.13059440e-01 -3.34178060e-01 -2.16635600e-01 3.63530785e-01
9.60363865e-01 3.21663231e-01 4.16379094e-01 -8.17376748e-02
6.04082882e-01 1.22554630e-01 8.20293963e-01 6.02808714e-01
1.51766941e-01 1.06092262e+00 -1.66498676e-01 -5.55237949e-01
-1.08308041e+00 -1.16991830e+00 -8.27280506e-02 5.32836854e-01
8.27421963e-01 -7.19848394e-01 -3.16725254e-01 -5.26453853e-01
8.06284472e-02 2.75329268e-03 -2.70658582e-01 4.59437147e-02
-4.75907117e-01 -3.36973488e-01 -1.32956579e-01 3.97962362e-01
6.27067685e-01 -4.61664826e-01 -6.11247659e-01 7.79651031e-02
-1.98776722e-01 -1.12650418e+00 -8.03618133e-01 -1.72683910e-01
-7.92144001e-01 -1.35727525e+00 -5.25347292e-01 -7.22209096e-01
8.06505322e-01 1.07923293e+00 1.13065910e+00 4.61402684e-01
-5.74629847e-03 2.27089033e-01 -3.59703034e-01 -7.15811253e-02
-7.85017386e-02 -6.19964078e-02 2.70182103e-01 8.30464438e-02
2.88259774e-01 -5.60771763e-01 -9.98269022e-01 8.41574907e-01
-6.08571947e-01 5.16131163e-01 5.70893347e-01 7.07031786e-01
1.06708694e+00 -9.64533389e-02 -5.06878756e-02 -5.36025465e-01
-3.07254553e-01 -1.94072962e-01 -9.69398797e-01 1.45900056e-01
-5.54623485e-01 -1.22621872e-01 2.50965595e-01 -1.56540036e-01
-6.82248712e-01 7.22360551e-01 -1.82857677e-01 -8.30943942e-01
1.08311027e-01 1.36876971e-01 -3.60771179e-01 -4.83656019e-01
2.95643449e-01 3.35903585e-01 -9.69995931e-02 -6.43715858e-01
2.78855860e-01 3.84227514e-01 7.59470642e-01 -4.14203256e-01
1.18077886e+00 7.88950145e-01 1.04254119e-01 -6.32889628e-01
-7.72188127e-01 -9.73530173e-01 -8.53339553e-01 -1.88841835e-01
6.59513652e-01 -1.23551166e+00 -5.92248082e-01 6.53760076e-01
-1.31579149e+00 1.61685720e-01 -2.62616370e-02 4.44172263e-01
-5.74193716e-01 5.97658813e-01 -2.96146303e-01 -3.81210566e-01
-3.34883571e-01 -1.38194489e+00 1.67029011e+00 1.66204304e-01
4.46331948e-02 -4.89427507e-01 1.18411504e-01 3.87096196e-01
3.38176861e-02 1.52413130e-01 3.53591353e-01 -1.95869915e-02
-1.27875578e+00 -1.23931736e-01 -2.80985296e-01 -1.07652903e-01
1.63862705e-01 -5.64649440e-02 -7.85931528e-01 -4.36750740e-01
-2.81987479e-03 8.98552686e-02 7.92672157e-01 2.29295447e-01
7.56109834e-01 6.40951023e-02 -5.20646751e-01 1.20231330e+00
1.54227245e+00 -5.50780892e-02 5.75450182e-01 5.86344182e-01
1.21423781e+00 1.36888608e-01 9.61217761e-01 5.26468396e-01
8.26979280e-01 1.03993583e+00 7.27090538e-01 -6.44982606e-02
-5.43968529e-02 -5.43529510e-01 3.19944769e-02 9.38949645e-01
2.27801647e-05 2.21403494e-01 -9.13641572e-01 3.84818345e-01
-2.01748252e+00 -6.90111697e-01 -2.95605034e-01 2.37359095e+00
3.66288036e-01 -4.34930483e-03 7.66490633e-03 -6.53840303e-02
6.32805526e-01 1.39340863e-01 -5.98838210e-01 4.44249868e-01
-7.64243975e-02 -2.51804702e-02 5.28043032e-01 6.85633302e-01
-1.02073908e+00 9.64665830e-01 5.31533098e+00 6.66169226e-01
-9.57279086e-01 -2.61186082e-02 -2.09796894e-02 -2.07508411e-02
-3.87612492e-01 4.24162358e-01 -1.05297482e+00 3.56039375e-01
1.51572898e-01 -8.45188182e-03 3.38763773e-01 7.95017838e-01
2.24634651e-02 -4.11016166e-01 -1.13842881e+00 1.54883313e+00
5.15747555e-02 -1.58687854e+00 9.39830989e-02 1.93268761e-01
6.99500918e-01 3.12003434e-01 -2.77773380e-01 -1.96659639e-01
1.82141274e-01 -7.09540367e-01 8.46794784e-01 5.91611326e-01
7.96913683e-01 -8.80491257e-01 4.74597901e-01 6.39692664e-01
-1.77483284e+00 1.75933644e-01 -5.62076867e-01 -2.25613359e-02
3.00189137e-01 7.48301148e-01 -5.91569543e-01 1.01440871e+00
9.00812387e-01 1.08256233e+00 -6.52543426e-01 1.18581545e+00
1.01057902e-01 -2.53780395e-01 -5.93788385e-01 5.70919871e-01
-4.25638668e-02 -3.01607043e-01 5.98611116e-01 7.15990782e-01
5.72188675e-01 2.01537430e-01 4.50321257e-01 6.22377515e-01
1.89187989e-01 -1.50366411e-01 -6.17308676e-01 7.61274695e-01
8.16354454e-01 1.36891162e+00 -8.59264612e-01 -1.94178775e-01
-6.58916473e-01 1.23810542e+00 2.67580479e-01 -9.18720756e-03
-6.93533361e-01 4.35007885e-02 7.16204941e-01 5.26400089e-01
4.38375235e-01 -4.51712817e-01 -1.72174841e-01 -1.60060358e+00
3.07449669e-01 -9.20950651e-01 2.35486060e-01 -8.23796928e-01
-1.14069605e+00 4.90635842e-01 -9.93072912e-02 -1.69589818e+00
-2.50010401e-01 -1.00318864e-01 -2.67257005e-01 9.47028100e-01
-1.47158551e+00 -1.31252813e+00 -6.67933762e-01 7.87430406e-01
6.52503908e-01 9.14392397e-02 5.56785107e-01 3.67884248e-01
-3.26120257e-01 3.23764980e-01 6.36850076e-04 -7.42250085e-02
7.06975341e-01 -9.01543677e-01 8.86160493e-01 1.12414825e+00
2.52639383e-01 6.62089705e-01 2.50441164e-01 -8.06607187e-01
-1.73540783e+00 -1.04674053e+00 6.26733065e-01 -7.71830022e-01
2.03457892e-01 -4.71816361e-01 -7.96532631e-01 5.09505808e-01
-4.60015088e-01 2.62233406e-01 6.68762028e-02 -1.31731004e-01
-3.04910332e-01 -1.39324397e-01 -8.96419585e-01 3.67570341e-01
1.35649908e+00 -4.51674342e-01 -4.48337346e-01 1.89028680e-01
6.14108860e-01 -9.51226711e-01 -7.76797295e-01 6.12373948e-01
7.10193455e-01 -1.21322870e+00 1.36668980e+00 8.64608064e-02
9.64982957e-02 -8.54481220e-01 -4.48073953e-01 -1.00808990e+00
-4.90266293e-01 -6.63180888e-01 -1.64167479e-01 1.16280067e+00
-1.94494560e-01 -3.67558002e-01 8.91340494e-01 3.88594002e-01
-2.20506623e-01 -6.32246315e-01 -1.12388623e+00 -6.33702755e-01
-5.15052974e-01 -4.10486341e-01 8.05859804e-01 7.14191258e-01
-6.21813834e-01 3.04599941e-01 -3.63622189e-01 6.45378947e-01
8.98610830e-01 5.11188686e-01 1.33085322e+00 -1.25995028e+00
-3.10492456e-01 -1.66291967e-01 -5.13194501e-01 -1.67878175e+00
-2.54155755e-01 -5.30694306e-01 1.37845904e-01 -1.47793365e+00
3.54975551e-01 -6.30863726e-01 6.65896237e-02 1.22304298e-01
-3.13909769e-01 3.92352909e-01 2.86879241e-01 5.76373637e-01
-8.37924659e-01 3.61979157e-01 1.18881822e+00 2.15292379e-01
-7.79367834e-02 -1.15642682e-01 -5.10239303e-01 1.00371516e+00
1.83087572e-01 -4.60324228e-01 -3.21855187e-01 -6.23126209e-01
3.73996943e-01 2.26285860e-01 6.38041675e-01 -1.10079288e+00
5.55039108e-01 -3.00375044e-01 5.24312854e-01 -1.28870058e+00
4.56549644e-01 -1.04155207e+00 4.82047915e-01 2.68732756e-01
4.89025265e-01 3.64457995e-01 -1.28550470e-01 7.36679018e-01
-1.31490588e-01 1.15002722e-01 6.48566902e-01 -2.46808723e-01
-1.10479605e+00 7.65679896e-01 3.37971598e-01 -1.22405410e-01
9.34750438e-01 -5.28078079e-01 -7.53267482e-02 -2.04297736e-01
-2.47173235e-01 4.58402187e-01 1.24657655e+00 4.82530236e-01
1.00245738e+00 -1.51068234e+00 -6.11925125e-01 6.93724275e-01
4.43482369e-01 8.60669672e-01 3.87385517e-01 8.66494894e-01
-8.75388265e-01 1.54947028e-01 4.33814675e-02 -1.22364330e+00
-1.33036208e+00 5.71286559e-01 2.35019490e-01 4.65003811e-02
-9.68683481e-01 7.36526847e-01 5.05179644e-01 -3.42817873e-01
6.20915927e-02 -3.49420607e-01 1.34760648e-01 -2.90191650e-01
5.96375406e-01 3.14080596e-01 3.18158954e-01 -8.94688964e-01
-6.39303923e-01 1.27011764e+00 -1.39072552e-01 -9.56603233e-03
1.29621255e+00 -5.31039238e-01 -4.05623950e-02 2.13759467e-01
1.03110790e+00 1.08912632e-01 -1.53382468e+00 -3.75175327e-01
-1.71587944e-01 -9.61858630e-01 7.38086253e-02 -2.68407106e-01
-9.07993376e-01 6.51245296e-01 4.18668777e-01 -4.37615752e-01
1.14002287e+00 -2.70568412e-02 1.08978975e+00 1.82158738e-01
8.59527051e-01 -8.30868900e-01 4.84274104e-02 6.11384749e-01
7.77113080e-01 -1.37123978e+00 2.20634148e-01 -6.33781493e-01
-1.19465940e-01 1.10654902e+00 6.57070756e-01 -2.49767914e-01
5.42669892e-01 2.05466628e-01 -1.70637846e-01 -3.91693592e-01
-4.60985154e-01 8.11903030e-02 4.70288157e-01 4.36718285e-01
-1.61293179e-01 -2.36261606e-01 1.83213040e-01 3.17915499e-01
-5.92883602e-02 -2.47992259e-02 2.81371146e-01 9.59296405e-01
-5.15140355e-01 -1.15661073e+00 -6.51691318e-01 3.35969925e-01
6.57916144e-02 -4.11955155e-02 -1.71014503e-01 7.59772420e-01
2.50205062e-02 8.32619309e-01 1.45478114e-01 -7.43821859e-01
6.47808194e-01 -3.80937517e-01 5.60273349e-01 -7.70874262e-01
-2.37612695e-01 2.68857598e-01 -3.06401819e-01 -1.17612171e+00
-5.21426082e-01 -8.39772880e-01 -1.07487798e+00 -5.33086658e-01
-3.75021040e-01 -1.90221593e-01 4.60062355e-01 8.42972994e-01
5.28628469e-01 6.32292181e-02 8.03908646e-01 -1.20088971e+00
-4.14411515e-01 -5.43914378e-01 -4.14822400e-01 4.20552284e-01
5.86190343e-01 -9.06285107e-01 -1.92617446e-01 -1.72431126e-01] | [7.789266109466553, -2.453009605407715] |
e2269792-45f7-4ea6-a5b6-72ef64b89860 | learning-to-generate-move-by-move-commentary | null | null | https://aclanthology.org/P18-1154 | https://aclanthology.org/P18-1154.pdf | Learning to Generate Move-by-Move Commentary for Chess Games from Large-Scale Social Forum Data | This paper examines the problem of generating natural language descriptions of chess games. We introduce a new large-scale chess commentary dataset and propose methods to generate commentary for individual moves in a chess game. The introduced dataset consists of more than 298K chess move-commentary pairs across 11K chess games. We highlight how this task poses unique research challenges in natural language generation: the data contain a large variety of styles of commentary and frequently depend on pragmatic context. We benchmark various baselines and propose an end-to-end trainable neural model which takes into account multiple pragmatic aspects of the game state that may be commented upon to describe a given chess move. Through a human study on predictions for a subset of the data which deals with direct move descriptions, we observe that outputs from our models are rated similar to ground truth commentary texts in terms of correctness and fluency. | ['Taylor Berg-Kirkpatrick', 'Graham Neubig', 'Harsh Jhamtani', 'Varun Gangal', 'Eduard Hovy'] | 2018-07-01 | null | null | null | acl-2018-7 | ['game-of-chess'] | ['playing-games'] | [ 1.12148620e-01 6.22596025e-01 7.12396875e-02 -4.21308458e-01
-1.20894063e+00 -9.39896643e-01 1.02780771e+00 -6.26217900e-03
-2.48637334e-01 1.10639906e+00 1.10600901e+00 -3.09227854e-01
2.85477757e-01 -6.85005963e-01 -6.50233865e-01 -1.15847655e-01
5.04617095e-01 9.70290124e-01 2.17629850e-01 -1.24815738e+00
7.14139283e-01 -3.99635732e-01 -1.23267746e+00 1.20635450e+00
4.51915950e-01 4.34110075e-01 1.54464662e-01 1.13431835e+00
1.94036290e-01 1.43564165e+00 -1.27196884e+00 -1.07363904e+00
3.35934043e-01 -1.07025540e+00 -1.27776492e+00 -9.26391929e-02
2.91987598e-01 -1.74672678e-01 -3.47224146e-01 7.00208843e-01
7.05744922e-01 3.87875855e-01 1.02309895e+00 -9.46677208e-01
-1.02920473e+00 1.36710095e+00 2.30482053e-02 4.08151269e-01
9.67117369e-01 5.43936849e-01 1.36215496e+00 -4.13688242e-01
8.73613000e-01 9.21767592e-01 6.25522316e-01 1.16666365e+00
-9.11237061e-01 -2.99743861e-01 -6.72472175e-03 1.23597480e-01
-8.60660315e-01 -3.14663976e-01 4.88992065e-01 -7.08730400e-01
1.02112615e+00 3.94053370e-01 9.25596535e-01 1.62232161e+00
4.92993206e-01 8.30699980e-01 9.56485033e-01 -2.36273438e-01
2.19923154e-01 -1.07453419e-02 -1.78485662e-01 3.77251178e-01
-2.32083187e-01 5.76229431e-02 -9.47927594e-01 -1.10404305e-01
4.29413646e-01 -9.89367366e-01 -6.70008510e-02 3.11461419e-01
-1.21572530e+00 1.01639175e+00 1.41311258e-01 -1.41491279e-01
-1.71508014e-01 2.24229082e-01 7.15503514e-01 3.46372038e-01
3.72285783e-01 1.45709920e+00 -2.81606108e-01 -8.35562110e-01
-9.29416597e-01 1.29833543e+00 8.97021770e-01 1.12706411e+00
5.28351441e-02 3.15732546e-02 -8.19436908e-01 6.70562863e-01
-1.98337987e-01 3.67174715e-01 8.81541848e-01 -8.51425529e-01
8.99910390e-01 3.51359934e-01 5.32184660e-01 -1.16017318e+00
-3.89582723e-01 -3.25906038e-01 -5.38986981e-01 -2.16730945e-02
2.15079054e-01 -2.95816094e-01 -3.44285160e-01 1.53169322e+00
-2.42568374e-01 -4.49884981e-01 3.08375984e-01 1.04666281e+00
1.47656763e+00 7.49578238e-01 3.14038666e-03 2.52326488e-01
1.16309345e+00 -1.28291488e+00 -3.90964985e-01 -4.09048349e-01
5.69878042e-01 -7.42047966e-01 1.51158881e+00 3.11436802e-01
-1.62391353e+00 -5.84975600e-01 -9.63508368e-01 -4.45694536e-01
-7.04106390e-02 4.07482535e-01 3.69570524e-01 1.15855440e-01
-1.31236744e+00 4.24183697e-01 -1.68251306e-01 3.64547372e-02
-1.64214298e-01 6.09710999e-02 -3.27280387e-02 5.60085833e-01
-1.46816599e+00 9.72235799e-01 5.89689374e-01 -1.56396180e-01
-9.14568782e-01 -5.67914486e-01 -8.70144963e-01 -5.28814122e-02
3.67461056e-01 -9.59444463e-01 2.05891776e+00 -8.21222961e-01
-1.95454085e+00 1.16273463e+00 3.13322097e-01 -5.30557156e-01
8.40458989e-01 4.15486358e-02 -2.13892385e-01 -1.30871609e-01
4.97807443e-01 8.65509868e-01 4.04482991e-01 -7.17655420e-01
-7.27934778e-01 5.35726786e-01 8.87672603e-01 6.28405988e-01
3.10141712e-01 1.55732287e-02 7.95435384e-02 -8.61819983e-01
-5.29766560e-01 -1.03003788e+00 -3.22947294e-01 -9.74397123e-01
-9.07703519e-01 -1.75100476e-01 -3.73111606e-01 -7.09287107e-01
1.30439520e+00 -1.56131566e+00 5.59311450e-01 -3.63026559e-01
2.14021012e-01 6.42796559e-03 -1.94526806e-01 8.43209922e-01
6.92216083e-02 2.37598717e-01 -7.78683722e-02 -1.18690148e-01
2.98173100e-01 -4.92326975e-01 -7.17476904e-01 -2.79415399e-01
9.37147215e-02 1.28180444e+00 -1.07264876e+00 -5.58070540e-02
-1.73432276e-01 -2.51732856e-01 -1.02588105e+00 3.68796706e-01
-7.11836398e-01 6.30448163e-01 -3.50124061e-01 1.47327650e-02
-2.38557965e-01 -3.00111949e-01 -1.53886765e-01 2.72824228e-01
-5.67921158e-03 9.43963110e-01 -4.36477572e-01 1.86050129e+00
-4.62776184e-01 7.51489460e-01 -7.59090960e-01 -3.71621907e-01
7.79922307e-01 3.36085081e-01 -2.29181617e-01 -6.55125558e-01
1.77716583e-01 1.81886956e-01 1.30255282e-01 -5.13969481e-01
1.50361550e+00 -4.67708200e-01 -1.03961051e+00 8.06120813e-01
-1.68458387e-01 -8.23667228e-01 4.41703230e-01 7.26442814e-01
1.12553418e+00 1.17845364e-01 6.46667719e-01 -4.25188005e-01
4.29613173e-01 6.64913297e-01 1.55923098e-01 1.17275596e+00
4.97495346e-02 7.74309993e-01 7.40330815e-01 -7.23414660e-01
-1.24387240e+00 -8.32923830e-01 6.18666112e-01 1.22483134e+00
6.95012212e-02 -9.76767898e-01 -8.17945242e-01 -3.92044067e-01
-5.60263991e-01 1.00139427e+00 -9.49505746e-01 -2.77093023e-01
-4.64664668e-01 -5.44850886e-01 8.54264081e-01 5.70399582e-01
3.07003170e-01 -1.50226152e+00 -8.97083759e-01 4.18123305e-01
-6.79644883e-01 -1.03935432e+00 -6.85272157e-01 -3.14993083e-01
-2.00469583e-01 -1.04124188e+00 -4.84841734e-01 -7.98336565e-01
2.87252486e-01 -3.54766160e-01 1.66985261e+00 1.69451497e-02
-7.52723601e-04 1.71494409e-01 -5.33174813e-01 -3.70357007e-01
-1.08033001e+00 5.08928359e-01 -2.43898198e-01 -7.43053675e-01
2.09566742e-01 -1.74417391e-01 -2.51333207e-01 -8.35210085e-02
-5.46272039e-01 9.64186490e-01 3.49618405e-01 1.01159382e+00
1.74446434e-01 -4.29528832e-01 4.53041166e-01 -1.18554997e+00
1.68372166e+00 -5.25622666e-01 -2.32721567e-01 1.44467637e-01
3.52803282e-02 -2.25470997e-02 9.70045388e-01 -2.60920018e-01
-1.06720161e+00 -2.70159334e-01 -1.67868018e-01 5.87381184e-01
-1.30914435e-01 6.37924969e-01 4.40891385e-01 6.33083284e-01
1.33010817e+00 3.05067271e-01 -4.34191704e-01 1.07694278e-02
5.45547426e-01 7.14253306e-01 8.65536869e-01 -7.84883022e-01
5.64717472e-01 -1.22449383e-01 -3.97907436e-01 -3.83480966e-01
-1.05535960e+00 -1.04945699e-04 -3.61457527e-01 -4.97903079e-01
7.29057729e-01 -9.80046809e-01 -7.69709289e-01 3.86531651e-01
-1.26915121e+00 -8.19094002e-01 -4.16673988e-01 -1.13374047e-01
-1.09829664e+00 -1.31744623e-01 -1.18212509e+00 -5.37678719e-01
-3.85944486e-01 -1.30938029e+00 1.12409067e+00 3.54385257e-01
-1.02589655e+00 -8.40865254e-01 3.27412993e-01 6.00094974e-01
3.72413367e-01 2.67510355e-01 8.89930665e-01 -7.37734139e-01
-1.38518795e-01 -7.01152384e-02 2.46035844e-01 -6.94610476e-02
-1.82630986e-01 -7.53690628e-03 -3.96951735e-01 8.63932222e-02
-2.09497899e-01 -9.56510663e-01 3.85695994e-01 2.83104986e-01
7.40054786e-01 -5.92582464e-01 2.94585496e-01 9.63970348e-02
9.21805203e-01 2.32969880e-01 5.78136504e-01 4.74980295e-01
5.11636376e-01 6.20649457e-01 7.35567570e-01 5.67832351e-01
7.56214023e-01 7.81351864e-01 8.47445452e-04 4.38347846e-01
-6.58047572e-02 -7.86626041e-01 6.34479821e-01 9.75960970e-01
-1.28995106e-01 -5.55463493e-01 -9.91863310e-01 6.58026457e-01
-1.78168559e+00 -1.23868465e+00 -2.63700604e-01 1.54377151e+00
1.15490520e+00 3.71371925e-01 5.31454742e-01 -1.00377910e-01
7.07872033e-01 3.03048819e-01 -2.76329666e-01 -9.86473501e-01
-1.71771571e-01 3.01619262e-01 6.74928948e-02 8.40985417e-01
-8.20227444e-01 1.37256598e+00 7.34939528e+00 8.37301195e-01
-8.49172890e-01 -1.43514454e-01 7.76716828e-01 -5.30020952e-01
-5.45871496e-01 -1.87730938e-01 -6.55704558e-01 5.42932510e-01
1.07365704e+00 -7.09486127e-01 5.46967566e-01 6.15126252e-01
2.93948352e-01 -3.42432931e-02 -1.19775796e+00 6.94571733e-01
3.65369827e-01 -1.79488778e+00 4.43835527e-01 -2.24643961e-01
1.19864953e+00 -2.83937991e-01 2.58109182e-01 6.51288927e-01
1.04406548e+00 -1.36445153e+00 1.47675729e+00 3.51844281e-01
8.93982291e-01 -6.42226398e-01 6.68247938e-01 5.26636422e-01
-6.96816981e-01 -5.80341369e-02 -3.35512161e-01 -9.11991358e-01
4.28115577e-01 -1.41101256e-01 -1.16057420e+00 2.52093166e-01
1.90802246e-01 6.79174423e-01 -7.49390841e-01 7.12259889e-01
-6.86754942e-01 4.56877470e-01 2.91983902e-01 -6.49336338e-01
7.10016131e-01 5.10968156e-02 5.03969848e-01 1.05141306e+00
1.48869306e-01 6.01072550e-01 8.68299752e-02 9.80530977e-01
-8.85439590e-02 3.43735456e-01 -4.16781098e-01 -2.37175003e-01
3.45708102e-01 7.59476900e-01 -6.49538517e-01 -6.17110968e-01
2.24848926e-01 1.07185471e+00 3.93920481e-01 -3.72177847e-02
-1.01718557e+00 -1.22173458e-01 2.82964885e-01 1.77457035e-01
-1.46036735e-03 2.42650092e-01 -3.33617449e-01 -1.09294116e+00
-6.31148592e-02 -1.32185280e+00 1.93004206e-01 -1.30303299e+00
-1.38485157e+00 1.11295295e+00 1.27568087e-02 -1.21586192e+00
-8.54407847e-01 -4.37319636e-01 -7.45742202e-01 8.43108416e-01
-7.04197824e-01 -8.27454448e-01 -6.28740713e-02 7.68310800e-02
1.07867503e+00 -5.15473425e-01 7.44321644e-01 -1.81389108e-01
7.70132942e-03 4.56373841e-01 -3.52394402e-01 2.63095498e-01
6.80021584e-01 -1.47606754e+00 1.27329695e+00 5.69699228e-01
2.70806074e-01 4.63053018e-01 1.31381214e+00 -6.98216140e-01
-3.80094290e-01 -8.64101470e-01 1.07773471e+00 -8.50240111e-01
6.81356728e-01 -1.81472391e-01 -2.29688987e-01 8.14005136e-01
4.94965702e-01 -8.75551403e-01 8.49736333e-01 -2.03941725e-02
-1.32887587e-01 6.36109948e-01 -7.55337298e-01 1.10207951e+00
1.23654425e+00 -7.16549397e-01 -1.15581572e+00 5.83909988e-01
4.98212636e-01 -1.22442496e+00 -2.72657722e-01 -9.15750638e-02
3.22153956e-01 -1.03606200e+00 4.36778128e-01 -1.46919024e+00
1.59064615e+00 -2.38361120e-01 1.29777983e-01 -2.01268911e+00
-3.58140707e-01 -7.41584539e-01 5.04175007e-01 8.16178560e-01
8.24209929e-01 2.05691740e-01 7.55569994e-01 8.72080266e-01
-5.22338510e-01 -6.55723214e-01 -6.43392265e-01 -3.22449207e-01
4.88236070e-01 -1.60009176e-01 6.09200180e-01 7.31328189e-01
8.01441312e-01 7.74362266e-01 -8.24591160e-01 -5.45305729e-01
-7.00752512e-02 5.25820777e-02 1.03884685e+00 -8.61922681e-01
-8.42094600e-01 -6.51926339e-01 -5.37891924e-01 -1.32714331e+00
2.78293550e-01 -1.06498182e+00 4.36740577e-01 -1.65677345e+00
2.54907429e-01 3.47251259e-02 4.88518655e-01 2.24682540e-01
-2.44501904e-01 4.79311138e-01 3.05449992e-01 5.96258603e-02
-9.27390993e-01 4.42335069e-01 1.31672394e+00 -3.09811324e-01
-2.23634735e-01 -1.01038985e-01 -1.28201437e+00 5.96011758e-01
8.49729538e-01 -3.41445893e-01 -6.75777256e-01 -6.68094695e-01
9.03026342e-01 5.69381893e-01 1.01665303e-01 -9.78358090e-01
2.31841862e-01 -4.74428415e-01 -4.22559790e-02 -1.03613779e-01
3.27322453e-01 2.41978377e-01 2.93175817e-01 2.35432401e-01
-1.38811350e+00 6.01280391e-01 1.16004437e-01 4.16410118e-01
-2.04196051e-01 -1.52151063e-01 4.42145586e-01 -6.14545047e-01
-5.41962981e-01 -2.01278359e-01 -9.46825683e-01 9.02094126e-01
1.02578354e+00 -6.37301356e-02 -5.63818812e-01 -1.01903975e+00
-7.93510973e-01 3.02211158e-02 5.53960979e-01 6.72375917e-01
3.33422184e-01 -1.36830556e+00 -1.23701191e+00 -4.41553295e-01
2.93114096e-01 -6.45960793e-02 3.91290128e-01 1.37875557e-01
-1.02519894e+00 5.24597943e-01 -3.76113921e-01 -6.47776425e-02
-9.30922747e-01 7.88054392e-02 3.11432600e-01 -5.56898594e-01
-6.99478209e-01 1.29736257e+00 -2.36328207e-02 -3.12108845e-01
-3.47469568e-01 -4.16218102e-01 -3.49377066e-01 -1.43058628e-01
8.35605085e-01 -1.19853266e-01 -4.89419550e-02 -9.82050359e-01
5.22342362e-02 1.96375679e-02 -2.90862322e-01 -6.58544064e-01
1.12876868e+00 1.31904840e-01 2.96261162e-01 7.28360474e-01
4.85434890e-01 2.50665545e-01 -1.01727843e+00 2.33425185e-01
-1.00799128e-01 -3.08263123e-01 -7.01537967e-01 -1.41976595e+00
-7.06467330e-01 5.86543500e-01 -5.99434495e-01 1.85953811e-01
4.16700393e-01 1.88765880e-02 8.05991352e-01 6.04195476e-01
6.34691417e-01 -1.23450148e+00 2.78248131e-01 7.93455124e-01
1.21751046e+00 -9.71566856e-01 -3.47545505e-01 -2.31649224e-02
-1.62514162e+00 1.00084388e+00 9.59481597e-01 -4.36682373e-01
-8.76057148e-02 -4.47523370e-02 3.71865660e-01 -3.97262305e-01
-1.30574703e+00 3.66336554e-01 2.29570940e-01 4.67119634e-01
7.06009626e-01 4.00528938e-01 -7.43680060e-01 1.23610008e+00
-1.59984243e+00 -2.22808972e-01 1.56660783e+00 5.25397420e-01
-1.26776129e-01 -8.07730675e-01 1.13621820e-02 5.29345751e-01
-3.03277850e-01 -3.72261703e-01 -1.18458939e+00 4.61791098e-01
-2.48466969e-01 1.00835872e+00 -1.04751028e-01 -6.83432817e-01
4.53190327e-01 -4.76047173e-02 4.06534612e-01 -1.16260660e+00
-1.30941200e+00 -7.46987224e-01 6.31592631e-01 -3.35298181e-01
-2.37834111e-01 -5.36667168e-01 -1.13656545e+00 -5.73716044e-01
1.57865047e-01 6.94271028e-01 1.73566908e-01 1.11473227e+00
2.09460124e-01 6.03510320e-01 5.87932616e-02 -7.41150916e-01
-6.64239585e-01 -1.02519429e+00 -5.05790770e-01 8.00646484e-01
-2.40688864e-02 -4.40020710e-01 -1.37187511e-01 2.12993443e-01] | [11.791887283325195, 8.917034149169922] |
9cb2adbf-fe8d-4b63-9060-f838060bba12 | an-image-analogies-approach-for-multi-scale | 2007.11047 | null | https://arxiv.org/abs/2007.11047v1 | https://arxiv.org/pdf/2007.11047v1.pdf | An Image Analogies Approach for Multi-Scale Contour Detection | In this paper we deal with contour detection based on the recent image analogy principle which has been successfully used for super-resolution, texture and curves synthesis and interactive editing. Hand-drawn outlines are initially as benchmarks. Given such a reference image, we present a new method based on this expertise to locate contours of a query image in the same way that it is done for the reference (i.e by analogy). Applying a image analogies for contour detection using hand drawn images as leaning images cannot gives good result for any query image. The contour detection may be improved if we increase the number of learning images such that there will be exist similarity between query image and some reference images. In addition of the hardness of contours drawing task, this will increase considerably the time computation. We investigated in this work, how can we avoid this constraint in order to guaranty that all contour pixels will be located for any query image. Fourteen derived stereo patches, derived from a mathematical study, are the knowledge used in order to locate contours at different scales independently of the light conditions. Comprehensive experiments are conducted on different data sets (BSD 500, Horses of Weizmann). The obtained results show superior performance via precision and recall vs. hand-drawn contours at multiple resolutions to the reported state of the art. | ['Slimane Larabi', 'Neil M. Robertson'] | 2020-07-21 | null | null | null | null | ['contour-detection'] | ['computer-vision'] | [ 5.44300020e-01 4.33667116e-02 3.11436206e-01 -7.45991915e-02
-4.61529404e-01 -6.41267896e-01 5.42646587e-01 2.82624692e-01
-6.46290302e-01 7.65796840e-01 -4.14384484e-01 -2.34835204e-02
-9.54392254e-02 -1.08048856e+00 -4.43576217e-01 -6.26868665e-01
6.24389648e-02 5.27626574e-01 9.01567698e-01 -4.04025674e-01
7.51065731e-01 9.37699020e-01 -1.74364841e+00 1.57444373e-01
6.01090968e-01 6.46615326e-01 3.43282968e-01 7.27584183e-01
-4.53143299e-01 1.99095950e-01 -7.16929376e-01 -5.75806975e-01
2.87962437e-01 -5.52974641e-01 -7.27867544e-01 4.13911551e-01
6.42577589e-01 7.81255066e-02 3.44264597e-01 1.38283777e+00
3.79020989e-01 -7.54164159e-02 1.08921957e+00 -7.53843486e-01
-4.68101561e-01 1.96392938e-01 -8.26414168e-01 2.21542060e-01
2.87840426e-01 -5.08209407e-01 5.43221295e-01 -7.51870930e-01
9.12804902e-01 9.86444831e-01 5.41049242e-01 3.79246712e-01
-1.27860045e+00 -3.15408945e-01 -2.39608124e-01 4.31101061e-02
-1.63642204e+00 9.08005238e-02 1.03358746e+00 -3.64004403e-01
3.33023638e-01 3.51792186e-01 7.94392109e-01 2.98954338e-01
2.69015521e-01 3.29935491e-01 1.31199431e+00 -8.85781229e-01
1.16383508e-01 5.52410781e-01 -1.93685889e-02 6.92940295e-01
3.01489562e-01 1.31905884e-01 -1.55468941e-01 1.60313934e-01
1.26338565e+00 -5.73618054e-01 -3.95896107e-01 -4.41810399e-01
-8.26003015e-01 8.44838679e-01 4.00675118e-01 7.75811553e-01
-3.00110549e-01 -2.99319685e-01 1.50358066e-01 2.38519192e-01
2.57921517e-01 3.65944713e-01 -2.11343780e-01 3.29321980e-01
-9.81677234e-01 1.80887222e-01 8.23221684e-01 8.91803682e-01
7.87026584e-01 6.61812276e-02 2.43060544e-01 6.49537861e-01
-1.74724609e-01 3.57489407e-01 4.31842089e-01 -7.33925223e-01
1.38538284e-02 7.33795524e-01 5.36193997e-02 -1.27741587e+00
-3.77799958e-01 -3.52754563e-01 -6.89633787e-01 1.09443998e+00
1.00686049e+00 -1.33880973e-02 -9.17873383e-01 1.19542694e+00
4.41930860e-01 -6.51301891e-02 2.27788314e-01 8.62586558e-01
6.20240152e-01 6.12736642e-01 -2.21176401e-01 -4.29503351e-01
1.54932666e+00 -4.43079889e-01 -7.25600362e-01 1.45250633e-01
2.74062097e-01 -1.41471684e+00 8.98267448e-01 8.61991823e-01
-9.73357201e-01 -7.28337526e-01 -1.27433300e+00 2.54816890e-01
-7.23059297e-01 2.69381434e-01 2.95741409e-01 8.83714378e-01
-9.66509342e-01 7.41890013e-01 -2.56764382e-01 -5.15873194e-01
3.41024399e-02 1.41765863e-01 -3.22401285e-01 5.40527940e-01
-9.02462661e-01 1.07725477e+00 7.38471925e-01 1.14203505e-01
-9.95544940e-02 -2.37152249e-01 -3.34863603e-01 -2.17148408e-01
3.42598885e-01 -3.07420343e-01 7.95283198e-01 -1.34011006e+00
-1.32827473e+00 1.23161542e+00 1.03795260e-01 -4.71037418e-01
8.35863292e-01 8.02964941e-02 -4.61772889e-01 5.52120805e-01
-1.00816958e-01 5.66739440e-01 9.41232800e-01 -1.45609498e+00
-8.56127977e-01 -3.33003283e-01 -2.14311868e-01 1.88375175e-01
1.20291393e-02 -1.65906981e-01 -5.82091033e-01 -7.00681150e-01
4.09867883e-01 -7.65592754e-01 -1.42477136e-02 1.17608048e-01
-2.59733766e-01 -2.36763924e-01 8.71017337e-01 -6.31964207e-01
9.52707827e-01 -1.74390769e+00 -1.37733608e-01 5.75794518e-01
-2.06227452e-01 3.10262382e-01 1.29219130e-01 2.87783742e-01
-1.95538074e-01 1.09525174e-01 -3.84537846e-01 1.77478388e-01
-5.44150412e-01 -5.07851504e-02 -1.11770153e-01 4.56360459e-01
1.83246270e-01 2.97795057e-01 -3.67280543e-01 -1.24577463e+00
2.34398589e-01 4.66210335e-01 -1.02070056e-01 8.91631469e-02
-7.67390132e-02 2.65823156e-01 -3.19758654e-01 2.40799129e-01
9.85856831e-01 2.50123888e-01 -1.14105284e-01 -4.89861161e-01
-3.42932999e-01 -8.97687554e-01 -1.58207464e+00 1.29846692e+00
-2.08058715e-01 7.05382824e-01 -9.00218356e-03 -8.97982538e-01
1.70437992e+00 4.78744984e-01 1.77309692e-01 -8.07673097e-01
6.62765577e-02 3.12039912e-01 -2.08759487e-01 -4.32760984e-01
4.43955868e-01 -2.77943045e-01 3.75125974e-01 1.82472974e-01
-2.94106990e-01 -6.55226827e-01 4.29983288e-01 -1.67945504e-01
1.16210163e-01 3.07532996e-01 4.93652821e-01 -5.99825025e-01
1.19714260e+00 2.96532094e-01 1.46918058e-01 4.69621390e-01
2.78620981e-02 7.63424098e-01 5.45637667e-01 -4.72523630e-01
-1.24866521e+00 -1.01407421e+00 -5.79377174e-01 7.23691583e-01
3.39426428e-01 1.73015565e-01 -1.02040088e+00 -2.30626434e-01
-3.42154771e-01 3.84050339e-01 -6.03442132e-01 5.66097379e-01
-7.61505306e-01 -6.41795874e-01 3.08998436e-01 1.33487880e-01
6.74880147e-01 -1.35813665e+00 -1.12136853e+00 1.94636345e-01
3.38222444e-01 -6.97615087e-01 -2.31505036e-01 7.27248415e-02
-1.05420387e+00 -1.19194722e+00 -1.16728866e+00 -1.11735129e+00
7.06429064e-01 2.81449724e-02 9.94122684e-01 2.23618656e-01
-6.53067648e-01 1.75491273e-01 -4.01518255e-01 -3.79764497e-01
-6.16804898e-01 -1.83838993e-01 -5.92392564e-01 -5.23477281e-03
5.94411232e-02 -3.85932118e-01 -6.66423559e-01 2.40181521e-01
-1.05685747e+00 2.06573587e-02 7.00822532e-01 6.13626599e-01
7.19514370e-01 3.53885263e-01 5.91060579e-01 -1.13453054e+00
6.32051885e-01 2.45132789e-01 -9.57560420e-01 4.83027905e-01
-6.44790053e-01 2.36626104e-01 5.77500820e-01 -3.06144953e-01
-1.30897868e+00 3.55843365e-01 -1.39944971e-01 -4.88850698e-02
-4.61271822e-01 6.81798067e-03 1.42469943e-01 -4.52340811e-01
9.43093479e-01 3.16323638e-01 1.16869450e-01 -3.99858654e-01
4.09525275e-01 5.57701051e-01 7.67590284e-01 -5.27694166e-01
8.09136152e-01 5.75452387e-01 3.92720401e-01 -1.11499596e+00
-2.79104143e-01 -4.23581600e-01 -9.69023943e-01 -4.73807305e-01
1.07471132e+00 -2.15167269e-01 -5.96206307e-01 1.92069367e-01
-1.32466209e+00 1.79541096e-01 6.99498951e-02 1.97299287e-01
-6.98738635e-01 5.59233904e-01 -1.56042323e-01 -1.26926482e+00
-4.29184943e-01 -9.09254611e-01 8.22926223e-01 6.12585068e-01
1.58670098e-01 -1.09070063e+00 3.47538292e-02 -5.31485379e-02
2.33221531e-01 6.63606703e-01 9.64994252e-01 -4.84514236e-01
-5.42820573e-01 -1.51976019e-01 -3.56446892e-01 2.17231050e-01
6.36547878e-02 2.52193600e-01 -8.97523701e-01 -9.07025114e-02
-3.64137604e-03 -1.07860351e-02 7.56469369e-01 2.30511472e-01
6.47914648e-01 1.37087228e-02 -2.15271309e-01 3.59589219e-01
1.94459462e+00 6.23569727e-01 7.00465918e-01 4.85954106e-01
2.19371378e-01 9.92825687e-01 8.17597687e-01 1.02983773e-01
-4.53545153e-01 6.57859623e-01 1.96027458e-01 -4.25545037e-01
-2.14928210e-01 6.53692186e-02 -2.54106373e-01 2.04773381e-01
-3.74374509e-01 9.63914245e-02 -8.39300931e-01 4.14290816e-01
-1.37440133e+00 -9.40567195e-01 -5.47001779e-01 2.44646573e+00
7.77023733e-01 5.06609499e-01 2.06143886e-01 4.64315504e-01
9.42889392e-01 -1.01698302e-01 -1.75956920e-01 -6.14921272e-01
-2.88604289e-01 4.59854513e-01 4.28240538e-01 5.91848910e-01
-1.11040258e+00 8.41787219e-01 5.45216465e+00 1.16567588e+00
-1.25703418e+00 -2.44369999e-01 6.43550992e-01 7.06657887e-01
-7.82274976e-02 1.67929858e-03 -6.04605436e-01 8.24059173e-02
4.59693462e-01 -3.04112792e-01 1.38354197e-01 5.49365342e-01
-1.19751103e-01 -5.69428980e-01 -7.00173020e-01 1.07184815e+00
1.92009792e-01 -1.13162744e+00 1.66663513e-01 -3.70228291e-01
6.26261413e-01 -7.34177709e-01 -2.94100996e-02 -2.67814040e-01
-3.62794101e-01 -1.00650704e+00 6.89930439e-01 6.84537411e-01
6.48028791e-01 -9.52835202e-01 8.36677611e-01 4.26844060e-01
-1.47236609e+00 4.49618787e-01 -6.79407477e-01 2.03678325e-01
-5.47622051e-03 9.07914713e-02 -9.71814394e-01 7.16615617e-01
3.48508954e-01 9.64592025e-03 -8.53820682e-01 1.21686316e+00
1.88087802e-02 6.08220957e-02 -3.64181399e-01 -2.39842594e-01
1.73925862e-01 -5.65742910e-01 3.07312608e-01 1.37030780e+00
2.26794988e-01 1.74140051e-01 -2.78684020e-01 1.01383114e+00
5.38207710e-01 8.84431064e-01 -7.19427764e-01 4.44674999e-01
3.15232664e-01 1.28305173e+00 -1.44583035e+00 -2.59990126e-01
-3.59829694e-01 1.12368870e+00 -6.04698546e-02 3.36264402e-01
-4.72452223e-01 -6.33979023e-01 -3.69527221e-01 1.90621421e-01
2.37485945e-01 5.77425063e-02 -2.91349828e-01 -4.85994935e-01
-2.51239866e-01 -4.50069100e-01 4.60722685e-01 -7.55319834e-01
-9.90273416e-01 1.01684952e+00 1.57787845e-01 -1.40025663e+00
-8.89370441e-02 -7.00906157e-01 -7.41988003e-01 1.09466696e+00
-1.28973830e+00 -1.04663491e+00 -2.99914062e-01 5.21904409e-01
5.92509747e-01 3.07933893e-02 7.84080803e-01 -1.34496838e-01
-5.98909482e-02 2.71799445e-01 3.13618705e-02 1.79668859e-01
6.25519931e-01 -1.54217970e+00 5.40864654e-02 7.53712952e-01
4.15094376e-01 2.18259275e-01 8.70445132e-01 -6.66652679e-01
-6.72465563e-01 -4.34122562e-01 5.31187713e-01 -1.78326026e-01
3.11107278e-01 2.32587289e-02 -1.16216159e+00 4.20655981e-02
2.93572307e-01 -1.84529170e-01 7.82369375e-02 -4.24777806e-01
4.98334877e-02 -1.06052794e-01 -1.13439214e+00 5.03191888e-01
5.53279817e-01 -3.23368311e-01 -8.71332169e-01 -2.67431084e-02
1.36850730e-01 -1.89668655e-01 -9.34129596e-01 3.16744149e-01
7.12618291e-01 -1.52969849e+00 1.02457130e+00 -2.00407609e-01
-4.78642806e-02 -4.66259956e-01 2.87030816e-01 -8.75405312e-01
1.69876277e-01 -3.57701063e-01 7.47189999e-01 1.22459745e+00
5.56098044e-01 -4.48881924e-01 1.05828011e+00 2.56107420e-01
3.43121529e-01 -6.90387428e-01 -8.67937028e-01 -6.58453107e-01
2.69746155e-01 2.84659732e-02 1.85972825e-01 7.38748372e-01
-3.49332809e-01 1.88695252e-01 6.68731332e-03 2.61545181e-01
8.52824688e-01 5.38567603e-01 7.29985774e-01 -1.66862011e+00
-3.91089693e-02 -6.25999749e-01 -5.32990754e-01 -6.51450634e-01
-2.28302300e-01 -5.83275020e-01 -1.83116823e-01 -1.35059571e+00
-4.54120245e-03 -4.62893158e-01 9.96590853e-02 -2.03877300e-01
-1.25190482e-01 4.72268850e-01 3.94369096e-01 2.83134490e-01
2.78119184e-03 5.00417650e-02 1.66512644e+00 2.49351382e-01
-3.12117070e-01 2.63945609e-01 -1.88691735e-01 1.01716328e+00
9.08328712e-01 -2.51773030e-01 -3.16503584e-01 2.79737890e-01
8.81350338e-02 4.79573727e-01 1.54253960e-01 -1.04633033e+00
4.39809293e-01 -1.99323567e-03 6.71295524e-01 -8.21074545e-01
3.20314825e-01 -1.09472036e+00 2.14645848e-01 5.57141781e-01
-2.41203770e-01 1.37343451e-01 1.76055312e-01 5.54933429e-01
-2.03704745e-01 -8.88194442e-01 1.11110640e+00 -3.57796222e-01
-9.22252417e-01 -2.82958001e-01 -1.17346570e-01 -1.02829799e-01
1.20928073e+00 -8.32518637e-01 -1.44530669e-01 -3.66972953e-01
-9.58990753e-01 -3.49416792e-01 5.27068853e-01 -3.96381132e-02
7.20469594e-01 -9.46206570e-01 -7.31849730e-01 2.04810128e-01
-2.62511782e-02 -2.01311171e-01 1.12104282e-01 6.10144496e-01
-1.13035941e+00 2.68756658e-01 -6.51321590e-01 -3.99436593e-01
-1.52429390e+00 8.56604457e-01 4.71632153e-01 3.22659407e-03
-6.57925546e-01 4.38574761e-01 2.34546676e-01 1.87377408e-01
1.98150277e-01 -3.02592576e-01 -8.24052751e-01 2.18198240e-01
3.51830810e-01 3.92093271e-01 -2.31820606e-02 -6.32029772e-01
-4.43639383e-02 1.45301425e+00 1.04923144e-01 -2.82607347e-01
8.95749986e-01 -3.78862210e-02 5.35899587e-02 4.97336715e-01
1.02273059e+00 3.77855748e-01 -1.16147840e+00 -5.67480922e-02
4.33905333e-01 -5.37998021e-01 -1.07979491e-01 -6.14854276e-01
-8.96756947e-01 8.96129251e-01 1.11654913e+00 5.52778304e-01
1.31552911e+00 -2.05827698e-01 1.70513093e-01 2.57052183e-01
4.78680372e-01 -1.15941525e+00 1.71861425e-01 -1.28846720e-01
9.75626588e-01 -1.25921011e+00 1.14987820e-01 -7.51251459e-01
-8.51743162e-01 1.92166507e+00 4.32382971e-01 -4.38061863e-01
5.09295464e-01 2.78847218e-01 2.13634804e-01 -2.17624784e-01
-1.12706974e-01 -5.24502993e-01 3.93483400e-01 6.75260961e-01
5.21347165e-01 -6.63876683e-02 -8.31202865e-01 1.63981348e-01
-1.32203057e-01 -1.61267042e-01 7.48308241e-01 7.16405511e-01
-9.80477512e-01 -1.11281931e+00 -9.17799532e-01 1.69417635e-01
-4.93341446e-01 1.34951875e-01 -4.31295216e-01 1.33051956e+00
2.38437995e-01 7.32586443e-01 1.13732547e-01 1.58867776e-01
4.70738173e-01 7.98041373e-03 5.43549955e-01 -2.03953698e-01
-3.92118484e-01 3.67243499e-01 2.21024035e-04 -1.08883768e-01
-6.19557977e-01 -3.50928992e-01 -1.47802544e+00 2.17473544e-02
-4.22535419e-01 3.80318522e-01 6.94479406e-01 5.00221848e-01
-4.61616218e-01 1.84641168e-01 3.09267044e-01 -6.69549704e-01
-3.95071715e-01 -8.19799662e-01 -8.77490103e-01 6.53264821e-01
-6.33400492e-03 -5.35973012e-01 -3.59054089e-01 4.98999625e-01] | [10.479490280151367, -1.8881138563156128] |
e89bf493-c18f-4506-8584-9baf90d90d88 | when-hyperspectral-image-classification-meets | 2306.08964 | null | https://arxiv.org/abs/2306.08964v1 | https://arxiv.org/pdf/2306.08964v1.pdf | When Hyperspectral Image Classification Meets Diffusion Models: An Unsupervised Feature Learning Framework | Learning effective spectral-spatial features is important for the hyperspectral image (HSI) classification task, but the majority of existing HSI classification methods still suffer from modeling complex spectral-spatial relations and characterizing low-level details and high-level semantics comprehensively. As a new class of record-breaking generative models, diffusion models are capable of modeling complex relations for understanding inputs well as learning both high-level and low-level visual features. Meanwhile, diffusion models can capture more abundant features by taking advantage of the extra and unique dimension of timestep t. In view of these, we propose an unsupervised spectral-spatial feature learning framework based on the diffusion model for HSI classification for the first time, named Diff-HSI. Specifically, we first pretrain the diffusion model with unlabeled HSI patches for unsupervised feature learning, and then exploit intermediate hierarchical features from different timesteps for classification. For better using the abundant timestep-wise features, we design a timestep-wise feature bank and a dynamic feature fusion module to construct timestep-wise features, adaptively learning informative multi-timestep representations. Finally, an ensemble of linear classifiers is applied to perform HSI classification. Extensive experiments are conducted on three public HSI datasets, and our results demonstrate that Diff-HSI outperforms state-of-the-art supervised and unsupervised methods for HSI classification. | ['Tao Chen', 'Bin Wang', 'Tong He', 'Peng Ye', 'Jiayuan Fan', 'Jiamu Sheng', 'Jingyi Zhou'] | 2023-06-15 | null | null | null | null | ['classification-1'] | ['methodology'] | [ 4.54426825e-01 -6.73892498e-01 -3.89972597e-01 -4.91320968e-01
-6.30642235e-01 -4.59277332e-01 6.39051616e-01 1.75179049e-01
3.06250174e-02 5.37887216e-01 2.46309519e-01 1.58816323e-01
-6.98390245e-01 -1.28448558e+00 -2.91999340e-01 -1.26079178e+00
-3.38566840e-01 4.91382778e-02 3.31300765e-01 4.03246842e-03
9.57729965e-02 5.48482716e-01 -1.93501270e+00 3.88406008e-01
1.18774557e+00 1.54743600e+00 2.10758731e-01 4.61797953e-01
-1.84553266e-01 8.47924709e-01 -1.07788236e-03 5.59557557e-01
1.66548446e-01 -4.30038333e-01 -6.16479278e-01 5.41831076e-01
8.67454559e-02 -1.16939820e-01 -4.76293713e-01 1.01757598e+00
3.31069559e-01 5.17047465e-01 9.98909950e-01 -1.28084481e+00
-8.37879002e-01 6.25649214e-01 -6.83174491e-01 2.30014309e-01
1.74599644e-02 4.06017095e-01 1.02885520e+00 -8.82271588e-01
3.24141771e-01 7.87168026e-01 7.32425213e-01 -2.89704442e-01
-1.01908982e+00 -6.67115867e-01 1.48597613e-01 4.33388323e-01
-1.69756711e+00 -3.02143514e-01 1.05244565e+00 -6.63212121e-01
8.33215773e-01 2.42401406e-01 8.92135978e-01 6.59614146e-01
1.08615078e-01 7.13274717e-01 1.35831249e+00 -1.05762251e-01
9.32682231e-02 -2.12146953e-01 3.72654349e-01 7.18449533e-01
-1.33983582e-01 1.94036484e-01 -7.31373012e-01 -6.69054016e-02
4.53599334e-01 4.64429885e-01 -4.71560180e-01 -3.01258624e-01
-1.07894552e+00 9.13478553e-01 1.03050816e+00 5.78565538e-01
-7.03490973e-01 -3.49724919e-01 2.36170337e-01 1.58261254e-01
6.09966636e-01 -2.94643380e-02 -2.19273537e-01 2.72919178e-01
-1.04923224e+00 -3.82740319e-01 3.12412292e-01 6.72881186e-01
1.55917120e+00 8.96144807e-02 -2.86382049e-01 6.89222336e-01
1.92309916e-01 6.10279977e-01 7.06272423e-01 -3.90660733e-01
1.91274919e-02 9.71260726e-01 -4.12357718e-01 -1.14748323e+00
-5.26209116e-01 -6.37094855e-01 -1.46679246e+00 -1.89696938e-01
-2.93930858e-01 8.21119025e-02 -1.08183622e+00 1.35540307e+00
1.89649165e-01 5.76300144e-01 2.39175186e-01 7.60472596e-01
8.42168748e-01 1.33390594e+00 2.53214270e-01 -5.41797280e-01
1.01768184e+00 -8.55511069e-01 -3.70748222e-01 -2.38386355e-02
4.39182818e-01 -3.73976082e-01 8.85628223e-01 8.96363258e-02
-3.71664047e-01 -8.45646858e-01 -9.89742875e-01 4.28805947e-02
-8.90375257e-01 3.35128129e-01 9.12957311e-01 3.88096541e-01
-6.87919199e-01 6.15337849e-01 -7.79644668e-01 -4.34474617e-01
6.17598295e-01 2.26785034e-01 -1.91022202e-01 -3.66608292e-01
-1.01361430e+00 3.26636881e-01 7.63117492e-01 2.00729325e-01
-8.34031224e-01 -1.03429854e+00 -8.79759848e-01 1.06051743e-01
3.53382170e-01 -4.56574589e-01 4.39146966e-01 -8.13149393e-01
-1.25634491e+00 4.46598619e-01 -2.90510595e-01 -8.01557079e-02
-5.32831885e-02 2.57696837e-01 -6.45363033e-01 2.60514170e-01
1.78585738e-01 3.49895626e-01 1.09247684e+00 -1.29279149e+00
-8.00986826e-01 -4.97534513e-01 -3.18951249e-01 3.29974145e-01
-9.33941245e-01 -6.25988781e-01 4.44170833e-02 -4.04301077e-01
4.85835254e-01 -6.72500134e-01 -1.05384059e-01 -1.45533280e-02
-3.48036855e-01 -1.62416890e-01 1.38241696e+00 -3.41782242e-01
1.19505811e+00 -2.36183786e+00 2.04085603e-01 4.98370051e-01
1.76472545e-01 1.71607405e-01 -1.51543409e-01 4.01689231e-01
2.48411410e-02 -5.24943471e-02 -7.48212516e-01 1.00580990e-01
-1.67715311e-01 1.79879606e-01 -3.73912513e-01 4.95297372e-01
3.50388885e-01 9.61896420e-01 -1.04467535e+00 -4.83511776e-01
5.57597101e-01 4.74679917e-01 -1.27165377e-01 3.20594877e-01
-1.05794057e-01 6.27668321e-01 -7.01861858e-01 8.84103894e-01
7.32001305e-01 -5.21875322e-01 -2.32109264e-01 -6.40959203e-01
-5.46866298e-01 -3.21436077e-01 -1.05835903e+00 1.81909847e+00
-3.36857706e-01 3.54572713e-01 -3.20533901e-01 -1.10838950e+00
8.97975922e-01 3.99203636e-02 7.75255680e-01 -4.28263187e-01
-1.25035942e-01 1.15439087e-01 -4.02992487e-01 -3.94409597e-01
3.13721508e-01 -1.05910294e-01 7.89740309e-02 3.46596152e-01
2.42646933e-01 -3.26176435e-01 -4.82136235e-02 -1.84302311e-02
7.05786347e-01 -1.17062531e-01 4.16323721e-01 -3.38931173e-01
7.58127809e-01 2.48597965e-01 2.89055586e-01 5.49512506e-01
-6.45447848e-03 3.38229984e-01 -7.63480589e-02 -4.56114173e-01
-6.33810997e-01 -7.51597226e-01 -3.37054819e-01 1.18502760e+00
1.57529637e-01 -2.97197729e-01 -8.11393782e-02 -6.03765249e-01
1.26146063e-01 4.44881976e-01 -7.70901263e-01 -3.69693339e-01
1.28596455e-01 -1.13272023e+00 2.68697292e-01 4.22767341e-01
9.33550537e-01 -9.00241613e-01 -4.81255084e-01 1.66878387e-01
-3.87550034e-02 -7.46785820e-01 -1.01978585e-01 5.62546551e-01
-6.49177492e-01 -8.36153507e-01 -3.81515354e-01 -5.51278710e-01
4.57799017e-01 7.26578772e-01 6.62911415e-01 -6.44732490e-02
-6.02673769e-01 2.88389266e-01 -6.06773794e-01 -1.14681147e-01
6.41112924e-02 3.89830738e-01 -1.49588794e-01 4.19139236e-01
2.20194131e-01 -9.90308702e-01 -6.05002761e-01 5.68142533e-02
-1.31007302e+00 1.63161963e-01 7.85750329e-01 1.00476623e+00
8.48386407e-01 7.81112313e-01 3.98846060e-01 -6.97965682e-01
7.65633658e-02 -8.16644192e-01 -4.39448237e-01 3.57188791e-01
-5.51441550e-01 -6.00400418e-02 9.00036991e-01 -3.59532058e-01
-1.24415827e+00 2.44743913e-01 1.47918552e-01 -4.28314149e-01
-3.40169370e-01 1.01557684e+00 -1.36267185e-01 -4.05040354e-01
6.24947608e-01 9.50915694e-01 -3.52662444e-01 -1.61838129e-01
4.76380825e-01 7.51638770e-01 3.81449699e-01 -3.96900088e-01
1.24061275e+00 5.88744521e-01 1.46032467e-01 -1.21096504e+00
-1.20433414e+00 -6.73568428e-01 -1.00091720e+00 -2.20933497e-01
6.97483003e-01 -1.02709591e+00 -2.93765396e-01 7.23874450e-01
-4.30671185e-01 -3.45878661e-01 -4.79052037e-01 3.93019944e-01
-2.75184691e-01 3.27074826e-01 -3.15300137e-01 -6.02833211e-01
-1.42265514e-01 -8.88179004e-01 1.19372094e+00 5.24400532e-01
5.70027947e-01 -1.02705038e+00 -6.22959947e-03 -5.90343699e-02
6.09765291e-01 2.58988261e-01 1.20139933e+00 -3.12899023e-01
-6.54084146e-01 4.12722342e-02 -5.45461297e-01 1.98716432e-01
5.76247990e-01 1.60944223e-01 -1.15084469e+00 -2.52247274e-01
-5.30318879e-02 -5.86130500e-01 1.36510313e+00 4.99315023e-01
1.52620423e+00 -1.39582023e-01 -3.84808362e-01 1.15571952e+00
1.74249256e+00 5.32990918e-02 3.34301472e-01 1.02782749e-01
9.50902522e-01 4.12095219e-01 6.81294024e-01 8.01323295e-01
3.43088955e-01 1.65689308e-02 4.08123106e-01 -1.99830860e-01
5.70445098e-02 -2.00281069e-01 1.33492485e-01 9.07098532e-01
-4.79585789e-02 -5.97223267e-02 -8.20419669e-01 3.02435696e-01
-1.74143028e+00 -1.20160186e+00 -9.02217627e-03 1.84033895e+00
5.41025639e-01 -3.16137910e-01 -7.74983391e-02 2.76853740e-01
5.29043794e-01 7.70859540e-01 -8.45882654e-01 7.10512340e-01
-4.21003073e-01 -1.12776645e-03 3.73741090e-01 1.55100942e-01
-1.44996428e+00 1.12026417e+00 5.47767925e+00 1.00520086e+00
-1.50564432e+00 -7.44120404e-02 5.51426113e-01 3.94979715e-01
-2.45791331e-01 3.75379771e-02 -5.60231149e-01 1.58920139e-01
6.72878385e-01 -2.95890629e-01 5.84619641e-01 6.73990250e-01
1.23738172e-02 -9.49695036e-02 -6.97900414e-01 9.25769866e-01
3.25427130e-02 -1.39084470e+00 3.27366173e-01 -2.18303893e-02
1.00425112e+00 1.92208543e-01 3.38412046e-01 3.03134412e-01
1.05972692e-01 -9.12887335e-01 3.81233037e-01 8.88484895e-01
7.15732813e-01 -7.47100413e-01 5.84328532e-01 4.37895119e-01
-2.01473022e+00 -3.77992332e-01 -6.88150644e-01 2.18905896e-01
-2.74676293e-01 8.68500769e-01 -1.57417178e-01 1.20571065e+00
6.48289263e-01 1.58656991e+00 -8.82644773e-01 9.66983676e-01
1.19301483e-01 6.50608361e-01 -2.59624749e-01 2.86902726e-01
5.61699212e-01 -4.12144125e-01 1.14400096e-01 1.23391223e+00
7.92716444e-01 7.76450455e-01 6.46194339e-01 8.04407299e-01
5.65499105e-02 8.97279009e-02 -8.34455669e-01 -3.75888467e-01
4.51351464e-01 1.52653527e+00 -8.16165924e-01 -3.41805428e-01
-3.04231137e-01 1.02601492e+00 6.08132146e-02 4.42049772e-01
-7.31085300e-01 -3.30262601e-01 3.43893021e-01 -4.05668885e-01
3.12773734e-01 -3.76540214e-01 1.23328619e-01 -1.28466773e+00
-3.36441904e-01 -5.63061178e-01 5.76484680e-01 -9.40530419e-01
-1.54335868e+00 6.61519945e-01 -1.27947837e-01 -1.45766890e+00
-1.88642032e-02 -5.27559996e-01 -5.93407452e-01 7.05926180e-01
-2.10259819e+00 -1.74584579e+00 -1.08540583e+00 1.07427907e+00
4.85966682e-01 -8.95482972e-02 9.15193200e-01 -1.29267707e-01
-6.67061269e-01 -6.09373674e-04 3.37705821e-01 -4.74220626e-02
4.99088019e-01 -1.02774417e+00 -3.03414881e-01 8.12513053e-01
1.56472161e-01 2.13794023e-01 9.24779996e-02 -5.21942437e-01
-1.76097870e+00 -1.59025681e+00 2.90318191e-01 5.66312261e-02
9.13465559e-01 1.55825347e-01 -1.08609390e+00 4.01944339e-01
-3.39197293e-02 4.92877185e-01 9.83129978e-01 -2.04238240e-02
-3.68115813e-01 -5.36182404e-01 -9.32388663e-01 1.38221651e-01
1.15267897e+00 -8.93595159e-01 -2.07983211e-01 5.98354518e-01
6.32585883e-01 7.56578222e-02 -1.13528240e+00 5.67223430e-01
3.65118772e-01 -1.04662013e+00 8.55990291e-01 -8.44835564e-02
2.87089109e-01 -4.94422913e-01 -3.75759393e-01 -1.49045825e+00
-8.47531497e-01 -2.10597232e-01 7.56836832e-02 1.32404900e+00
1.69311404e-01 -5.39065480e-01 5.21995008e-01 -1.02728665e-01
-3.26417059e-01 -4.29347783e-01 -6.54051483e-01 -6.61080182e-01
-1.77914217e-01 -3.87724340e-01 8.01247001e-01 1.30591452e+00
-2.29864120e-01 2.26284906e-01 -2.58284479e-01 7.33283162e-01
7.00538754e-01 6.94332838e-01 3.98475826e-01 -1.51566100e+00
2.53913272e-02 -6.61799669e-01 -3.88043553e-01 -7.28680491e-01
3.87686253e-01 -1.05156147e+00 3.69719900e-02 -1.38745379e+00
5.09058774e-01 -5.20592690e-01 -4.98486400e-01 7.12204337e-01
-2.44609237e-01 2.38328353e-01 -1.32978067e-01 5.21265566e-01
-3.68850499e-01 9.02171850e-01 1.01243019e+00 -5.81786513e-01
-2.79920608e-01 -2.57649988e-01 -4.12195534e-01 4.51503366e-01
6.04645908e-01 -3.80140990e-01 -6.05927408e-01 -2.71272004e-01
-1.80542514e-01 -1.64697215e-01 4.12386388e-01 -1.48221183e+00
5.90883434e-01 -6.39804363e-01 7.75693059e-01 -7.63161957e-01
2.19170511e-01 -9.85126376e-01 1.91347510e-01 1.34161532e-01
-1.85545921e-01 -5.15422940e-01 1.41672030e-01 6.07301712e-01
-4.17255133e-01 2.74378303e-02 8.68361235e-01 -7.23708421e-02
-1.29828274e+00 1.00357282e+00 -3.25081378e-01 -3.79553556e-01
1.01536965e+00 -1.64954990e-01 -4.05743480e-01 -4.14932929e-02
-6.47728741e-01 1.81848839e-01 4.11374032e-01 9.73668322e-02
6.65611088e-01 -1.25969195e+00 -6.67785883e-01 5.91281056e-01
6.40359163e-01 -1.21834511e-02 6.44815683e-01 7.81780481e-01
-9.16367993e-02 1.89352050e-01 -2.66317427e-01 -1.03881502e+00
-8.21376264e-01 4.42462951e-01 1.64567545e-01 -3.02493513e-01
-5.54621518e-01 8.17222178e-01 1.54933497e-01 -2.74051011e-01
-2.72805065e-01 -2.43761823e-01 -3.20188582e-01 4.79769498e-01
3.82882893e-01 1.62090978e-03 -6.18584082e-02 -9.11803722e-01
-1.77244142e-01 9.38194573e-01 3.92518401e-01 2.06372976e-01
1.62292767e+00 -1.49875641e-01 -3.82688254e-01 7.49880910e-01
1.30903554e+00 -3.76737505e-01 -1.29796505e+00 -6.90739572e-01
-1.54656142e-01 -5.39030612e-01 5.00677526e-01 -6.60539985e-01
-1.11455774e+00 1.04738557e+00 8.01002800e-01 3.02580535e-01
1.68965256e+00 -1.45318478e-01 3.64916176e-01 5.27227998e-01
3.51227045e-01 -6.93666637e-01 1.32797256e-01 5.14361322e-01
6.03107750e-01 -1.37828732e+00 7.87919015e-02 -4.26176786e-01
-5.66617012e-01 1.17238581e+00 4.18490112e-01 9.92178693e-02
9.63225842e-01 3.53937261e-02 -2.84288347e-01 -3.18892509e-01
-5.39986193e-01 -5.16346872e-01 5.96768200e-01 7.39657402e-01
2.01563448e-01 1.89310804e-01 4.58901525e-01 4.39953566e-01
2.91342568e-02 -3.95498425e-02 -7.65996575e-02 9.75112677e-01
-8.64652395e-01 -7.05664754e-01 -1.58762306e-01 8.11371326e-01
4.06464249e-01 -1.86270326e-01 -1.21347554e-01 6.24525607e-01
3.59578162e-01 8.15237284e-01 1.05847195e-02 -7.16202676e-01
6.19766209e-03 2.14610338e-01 1.49908572e-01 -6.53132796e-01
-2.80733585e-01 3.73679847e-02 -5.80073059e-01 -4.42364246e-01
-7.44844496e-01 -6.57700956e-01 -1.13740945e+00 -2.61694133e-01
-3.26065004e-01 2.85183221e-01 3.40726376e-01 1.05823147e+00
2.05136672e-01 5.88605046e-01 1.12302899e+00 -1.27437091e+00
-5.13162553e-01 -1.01549649e+00 -1.10776985e+00 1.74180791e-01
5.16884148e-01 -7.06435800e-01 -5.55797756e-01 1.58436105e-01] | [9.934417724609375, -1.5473088026046753] |
bdd3feb7-65f6-4df1-a662-59b174e1696a | reward-uncertainty-for-exploration-in-1 | 2205.12401 | null | https://arxiv.org/abs/2205.12401v1 | https://arxiv.org/pdf/2205.12401v1.pdf | Reward Uncertainty for Exploration in Preference-based Reinforcement Learning | Conveying complex objectives to reinforcement learning (RL) agents often requires meticulous reward engineering. Preference-based RL methods are able to learn a more flexible reward model based on human preferences by actively incorporating human feedback, i.e. teacher's preferences between two clips of behaviors. However, poor feedback-efficiency still remains a problem in current preference-based RL algorithms, as tailored human feedback is very expensive. To handle this issue, previous methods have mainly focused on improving query selection and policy initialization. At the same time, recent exploration methods have proven to be a recipe for improving sample-efficiency in RL. We present an exploration method specifically for preference-based RL algorithms. Our main idea is to design an intrinsic reward by measuring the novelty based on learned reward. Specifically, we utilize disagreement across ensemble of learned reward models. Our intuition is that disagreement in learned reward model reflects uncertainty in tailored human feedback and could be useful for exploration. Our experiments show that exploration bonus from uncertainty in learned reward improves both feedback- and sample-efficiency of preference-based RL algorithms on complex robot manipulation tasks from MetaWorld benchmarks, compared with other existing exploration methods that measure the novelty of state visitation. | ['Pieter Abbeel', 'Kimin Lee', 'Katherine Shu', 'Xinran Liang'] | 2022-05-24 | reward-uncertainty-for-exploration-in | https://openreview.net/forum?id=OWZVD-l-ZrC | https://openreview.net/pdf?id=OWZVD-l-ZrC | iclr-2022-4 | ['robot-manipulation'] | ['robots'] | [-1.56796664e-01 3.30511123e-01 -6.94704056e-01 -2.80616671e-01
-9.81970727e-01 -5.49839973e-01 4.58314478e-01 7.70728812e-02
-7.38549829e-01 1.19004595e+00 3.73933077e-01 -9.58415866e-02
-3.87687027e-01 -6.89683795e-01 -6.36697769e-01 -7.70551383e-01
-3.41652423e-01 6.50532246e-01 1.12095512e-01 -6.27110362e-01
7.09169269e-01 2.92025596e-01 -1.53211105e+00 -2.11578205e-01
1.02540231e+00 1.04698074e+00 3.82148772e-01 3.99270773e-01
-1.07624168e-02 7.36005604e-01 -5.19976616e-01 3.20102006e-01
4.10278440e-01 -4.86209989e-01 -9.81965363e-01 -2.52586722e-01
-3.97601545e-01 -3.94812018e-01 9.21651870e-02 9.71396506e-01
5.59070528e-01 5.15437782e-01 6.04903102e-01 -1.41806781e+00
-4.06648517e-01 1.22761309e+00 -4.33094263e-01 -1.69835001e-01
5.09829342e-01 6.04137719e-01 1.10889709e+00 -4.47644234e-01
6.86191380e-01 1.59656978e+00 -1.47658229e-01 8.94879103e-01
-1.35326922e+00 -5.66516459e-01 3.85984629e-01 2.45504349e-01
-6.12193108e-01 -1.01177275e-01 7.26692915e-01 -1.71878800e-01
7.70077944e-01 1.88328072e-01 7.98783243e-01 1.20159602e+00
2.23124638e-01 1.14552569e+00 1.47044373e+00 -8.42091888e-02
7.99353957e-01 1.76962718e-01 -2.22431064e-01 4.96252090e-01
5.90698160e-02 7.02069998e-01 -7.47209907e-01 -2.02265128e-01
8.37650120e-01 -1.70419365e-01 -2.40331590e-01 -1.06161058e+00
-1.38264835e+00 1.10117137e+00 6.90033317e-01 -5.67850284e-02
-6.92238688e-01 4.13421243e-01 3.37945223e-01 8.08311880e-01
-1.27217516e-01 1.30255163e+00 -6.99805915e-01 -6.81999087e-01
-3.25351477e-01 5.05166888e-01 7.42241502e-01 7.50879765e-01
9.33075130e-01 1.41221732e-02 -6.10446393e-01 5.84814250e-01
3.33518744e-01 4.17043120e-01 6.99221253e-01 -1.50660229e+00
2.42183939e-01 6.34229302e-01 5.76416731e-01 -4.87687469e-01
-4.60785031e-01 -3.80692929e-01 -4.24025089e-01 8.52345526e-01
1.24971218e-01 -3.32724363e-01 -7.02681124e-01 1.86377025e+00
1.95456758e-01 -5.86219072e-01 2.50812352e-01 1.13860834e+00
2.51505315e-01 3.48326445e-01 -1.89247485e-02 -4.08862025e-01
7.85544515e-01 -1.19404793e+00 -6.58642590e-01 -1.72600523e-01
6.72560513e-01 -1.73694983e-01 1.59375012e+00 6.21929586e-01
-1.10936141e+00 -2.04602644e-01 -7.57878840e-01 6.10399485e-01
-6.75490201e-02 -2.74353474e-01 7.68302798e-01 4.36990917e-01
-9.16236818e-01 9.38696027e-01 -6.00400984e-01 -3.85650486e-01
2.20086783e-01 4.72065568e-01 6.24022679e-03 1.74306005e-01
-1.10194194e+00 1.11478817e+00 4.50637132e-01 -3.33298147e-01
-1.20823026e+00 -2.63956755e-01 -6.50545359e-01 1.22076191e-01
9.94976819e-01 -5.08825541e-01 1.63551414e+00 -7.50291288e-01
-2.36946559e+00 1.01881385e-01 3.61988664e-01 -5.92972696e-01
5.77179313e-01 -3.80770445e-01 2.02014208e-01 -1.59865364e-01
8.05983767e-02 1.12678242e+00 8.39709103e-01 -1.32100177e+00
-7.22710311e-01 -8.63375068e-02 2.01749533e-01 5.99740446e-01
-1.13778800e-01 -3.88992488e-01 1.32691056e-01 -3.24422926e-01
-6.70687184e-02 -1.12086630e+00 -7.78389812e-01 -9.24066603e-02
-1.04177988e-03 -5.72689295e-01 4.84052867e-01 1.94722593e-01
9.04164076e-01 -1.73583424e+00 3.13458025e-01 2.30091661e-01
-1.38164967e-01 -1.63007379e-01 -4.10941839e-01 4.63598073e-01
4.90160048e-01 3.48152989e-03 -1.66059107e-01 1.01785831e-01
3.86787355e-01 4.49036479e-01 -4.84641552e-01 2.11983249e-01
1.94100186e-01 9.37657952e-01 -1.51974535e+00 -3.57152224e-01
3.44299302e-02 -1.40645266e-01 -9.44304764e-01 6.11987412e-01
-5.94362378e-01 7.84242868e-01 -7.62391508e-01 7.57105649e-01
1.67980969e-01 -6.66708797e-02 9.57742110e-02 4.19362277e-01
-1.65478483e-01 5.64200640e-01 -1.13524330e+00 1.87518525e+00
-3.56529921e-01 1.08676948e-01 -7.92044774e-03 -7.23627508e-01
1.03847384e+00 9.01176110e-02 4.97036159e-01 -9.32618082e-01
1.30268514e-01 3.85763079e-01 1.15611784e-01 -3.66695195e-01
8.07029426e-01 2.03273565e-01 -2.80145735e-01 7.49769628e-01
-2.30179742e-01 -6.99302793e-01 1.16129391e-01 -1.16040938e-01
1.15457892e+00 6.89252853e-01 5.04166424e-01 -3.13330501e-01
1.35136679e-01 1.64159819e-01 6.84568644e-01 1.08033240e+00
-4.44726586e-01 6.01473674e-02 6.10486507e-01 -2.22042814e-01
-5.63974559e-01 -9.45370257e-01 3.07642728e-01 1.46732950e+00
2.95485944e-01 -7.58593827e-02 -1.97548151e-01 -1.04109144e+00
4.10873853e-02 1.05154562e+00 -6.98890269e-01 -4.88914043e-01
-5.81777453e-01 -4.10654545e-01 1.38335139e-01 4.34616446e-01
2.64098346e-01 -1.73233390e+00 -1.54553807e+00 3.23988169e-01
-1.74143929e-02 -3.00646245e-01 -4.71750498e-01 7.15087712e-01
-1.03043664e+00 -8.38434279e-01 -7.85103440e-01 -2.42545351e-01
5.46721220e-01 9.38167125e-02 1.17717063e+00 -1.59358054e-01
4.59946170e-02 6.43321276e-01 -6.49256647e-01 -3.94382805e-01
-2.70506561e-01 1.06583670e-01 2.97783136e-01 -6.26252353e-01
2.38613095e-02 -4.05211419e-01 -7.40525544e-01 4.41639066e-01
-6.98109925e-01 -2.87768841e-01 9.81849313e-01 1.20342684e+00
4.01229650e-01 -3.03265393e-01 7.51963198e-01 -4.50063556e-01
1.24335623e+00 -5.15316784e-01 -6.60392344e-01 1.84875697e-01
-1.19449699e+00 9.57168877e-01 2.94786900e-01 -7.03724682e-01
-9.63491082e-01 -1.53080905e-02 2.31369063e-01 -2.14447588e-01
-3.60446516e-03 4.46211219e-01 2.06254810e-01 -3.45453550e-03
8.17203224e-01 1.22826532e-01 5.23999035e-02 -1.22237101e-01
6.86546743e-01 3.25770974e-01 6.36512563e-02 -1.17888963e+00
3.68274212e-01 5.94529696e-03 -6.46462664e-02 -1.28297910e-01
-4.38565969e-01 -2.83385128e-01 -8.45977813e-02 -1.71496153e-01
3.95283312e-01 -3.92944723e-01 -1.17443228e+00 -9.39809605e-02
-7.24626780e-01 -7.69082069e-01 -7.50317931e-01 6.65864527e-01
-1.23929763e+00 5.70782982e-02 -2.28581592e-01 -1.09967649e+00
-1.89394265e-01 -1.61250186e+00 8.90338004e-01 3.39178234e-01
-3.18241984e-01 -4.97948885e-01 3.31146359e-01 -3.09762299e-01
6.32785261e-01 -1.15554836e-02 6.94131196e-01 -3.99601996e-01
-7.43066072e-01 4.05774355e-01 5.69491833e-03 -2.37150684e-01
7.80860800e-03 -4.67092603e-01 -5.58759332e-01 -4.77224082e-01
-2.95924954e-02 -1.14707708e+00 8.43757033e-01 2.44518131e-01
9.89920974e-01 -5.49526274e-01 -1.65359393e-01 1.56580210e-01
1.15031719e+00 2.30359390e-01 3.75050455e-01 7.03524768e-01
2.12468356e-02 7.42715478e-01 1.48733318e+00 8.89198244e-01
2.61922359e-01 7.31298387e-01 9.97828305e-01 4.44307923e-01
5.09770989e-01 -3.54337275e-01 7.69344509e-01 2.26748496e-01
-9.99388006e-03 8.83051902e-02 -4.93885219e-01 3.85991842e-01
-2.31675601e+00 -7.04529464e-01 6.25401855e-01 2.32064319e+00
1.28389609e+00 2.63461173e-01 4.42690521e-01 -2.44396880e-01
7.39734620e-02 -8.44989568e-02 -1.17862272e+00 -6.12063825e-01
2.71937907e-01 -7.59077445e-02 6.98259711e-01 4.45327431e-01
-5.09804845e-01 1.07541382e+00 6.06488323e+00 6.74983025e-01
-1.03840923e+00 -2.58096069e-01 3.46489936e-01 -3.43505710e-01
-5.53830683e-01 1.11453153e-01 -6.15745485e-01 1.73510030e-01
6.20948017e-01 -8.70114490e-02 7.94152200e-01 1.16458201e+00
1.41776577e-01 -4.45330679e-01 -1.27367091e+00 8.05443704e-01
-5.10811687e-01 -1.03258145e+00 -2.41188794e-01 8.79882500e-02
6.28353179e-01 5.75646572e-02 2.41614535e-01 9.41551864e-01
1.16444027e+00 -1.15656519e+00 7.48776257e-01 6.00071669e-01
3.90775084e-01 -7.62070477e-01 4.82499182e-01 5.67055583e-01
-6.92698479e-01 -4.13451910e-01 -4.32600409e-01 -1.85396433e-01
1.58087518e-02 1.15753099e-01 -1.09417272e+00 1.71123743e-01
7.91238606e-01 3.91892552e-01 -2.18400389e-01 1.03946567e+00
-4.53017801e-01 2.52026111e-01 -1.93087652e-01 -6.20666802e-01
6.25567555e-01 -2.66379416e-01 7.01800764e-01 6.94188356e-01
3.73620331e-01 -6.35160133e-02 4.07595485e-01 1.13034844e+00
2.52848208e-01 -1.41460858e-02 -5.42955875e-01 -2.19353452e-01
6.11613989e-01 1.05958521e+00 -5.29685557e-01 8.31003208e-03
2.99712360e-01 7.43888378e-01 6.63204968e-01 2.41915524e-01
-7.27468967e-01 -5.89665435e-02 5.99647701e-01 -3.53466362e-01
1.44362524e-01 -1.53536662e-01 -3.15188207e-02 -8.46052647e-01
-2.75069445e-01 -1.06111157e+00 2.61264145e-01 -5.34427762e-01
-1.01143169e+00 3.94863784e-01 4.48362483e-03 -1.28192627e+00
-7.12163806e-01 -2.61915833e-01 -3.06659997e-01 5.03633857e-01
-1.67671347e+00 -5.43281674e-01 -2.17326428e-03 3.56174052e-01
6.95276558e-01 -1.93043232e-01 8.76701474e-01 -5.50996721e-01
-2.55887896e-01 4.74945813e-01 3.20621766e-03 -6.10238075e-01
1.03904986e+00 -1.57600141e+00 -7.16426447e-02 1.10360160e-01
-2.24546075e-01 5.56299210e-01 9.59858060e-01 -5.48538804e-01
-1.68009102e+00 -6.08559549e-01 1.29795432e-01 -2.31408253e-01
5.48389137e-01 1.90063462e-01 -5.80747128e-01 3.50412279e-01
3.25847477e-01 -2.84435689e-01 1.92650318e-01 1.64022610e-01
-3.98588255e-02 1.02076828e-01 -1.23961651e+00 8.80156636e-01
9.26352382e-01 3.65444422e-02 -7.18069553e-01 -1.43871814e-01
9.57176805e-01 -4.57222462e-01 -6.99625075e-01 5.09072542e-01
5.64604580e-01 -1.01959610e+00 7.98892498e-01 -5.72465837e-01
2.21272692e-01 -3.06236476e-01 -8.38481039e-02 -1.83388388e+00
-2.66472310e-01 -9.25264359e-01 -2.19553903e-01 6.14764571e-01
2.83792287e-01 -6.42346203e-01 8.90097320e-01 5.99359751e-01
-6.25044294e-03 -1.01052547e+00 -7.86290407e-01 -9.70480561e-01
7.49712288e-02 -4.00019400e-02 7.40801215e-01 5.96466660e-01
4.67478096e-01 1.68710738e-01 -5.49848080e-01 -3.36866826e-01
5.28110325e-01 2.99817175e-01 8.95855367e-01 -1.18743062e+00
-4.76092130e-01 -9.66336966e-01 4.01424587e-01 -1.20038223e+00
5.34090027e-02 -4.80162382e-01 5.85981607e-01 -1.45484304e+00
-8.70951787e-02 -7.11625516e-01 -5.27642012e-01 5.66954136e-01
-7.61999935e-02 -4.79615182e-01 3.19689810e-01 2.43769407e-01
-9.39851999e-01 1.12854755e+00 1.68003201e+00 -7.06482083e-02
-7.34855592e-01 8.20014477e-02 -7.18310118e-01 4.65213478e-01
1.19070208e+00 -6.12583339e-01 -5.66805363e-01 4.20783907e-02
6.19058609e-01 6.25943959e-01 8.65092129e-02 -7.82564878e-01
2.61118591e-01 -9.53795254e-01 -1.26982316e-01 -4.96084422e-01
2.89277703e-01 -7.85670638e-01 -3.97863805e-01 8.52578759e-01
-9.43210840e-01 6.08763397e-02 -1.37545660e-01 8.10378909e-01
5.74870454e-03 -2.32816473e-01 5.19805849e-01 -4.07265007e-01
-7.58775592e-01 1.32986009e-01 -4.72274959e-01 5.14341220e-02
8.17343533e-01 -1.14562288e-02 -2.10012257e-01 -5.17558575e-01
-3.99881124e-01 7.72111893e-01 4.74191517e-01 5.14114678e-01
7.47459531e-01 -1.24058676e+00 -4.68127996e-01 2.86386274e-02
2.92459220e-01 -5.19509390e-02 -4.83391285e-01 8.36799920e-01
2.01968431e-01 3.99897426e-01 -6.89785659e-01 -4.35243845e-01
-8.35423231e-01 4.61782277e-01 2.01475322e-01 -6.07193887e-01
-2.49366000e-01 7.92401552e-01 -1.90524742e-01 -7.79306352e-01
7.76746452e-01 -6.77862525e-01 -3.31505656e-01 -2.32826304e-02
5.56624029e-03 4.76271480e-01 -4.61531520e-01 2.63895810e-01
-1.70314759e-01 2.73094922e-01 6.40673414e-02 -5.95858634e-01
1.24789000e+00 -1.31348357e-01 2.06345782e-01 5.93469024e-01
5.04971683e-01 -1.26008749e-01 -1.74048018e+00 -1.09588712e-01
3.16080213e-01 -4.39906776e-01 -9.24021006e-02 -1.15806711e+00
-6.84592426e-01 6.48969233e-01 6.54031873e-01 5.55425584e-02
8.56799662e-01 -1.11500479e-01 4.83401626e-01 1.10549426e+00
9.37140763e-01 -1.43969047e+00 6.09384298e-01 6.33749723e-01
1.15422475e+00 -1.68495011e+00 -1.02770656e-01 3.97021234e-01
-1.14011705e+00 1.07913589e+00 9.65604544e-01 -7.52914771e-02
2.64812410e-01 -1.82664037e-01 2.94802617e-02 -1.20018840e-01
-1.04991519e+00 -3.43030632e-01 9.22079608e-02 5.73337138e-01
3.16985518e-01 3.56191158e-01 -4.81556594e-01 3.94397140e-01
-2.37314507e-01 1.60548240e-02 5.14196932e-01 1.14120150e+00
-9.24285293e-01 -1.37234986e+00 -2.93002963e-01 4.23602343e-01
6.98647797e-02 2.66643465e-01 -3.02306682e-01 5.43374360e-01
-4.19436872e-01 9.81439114e-01 -3.05300117e-01 -2.46694937e-01
1.79621860e-01 -3.35139811e-01 5.81329107e-01 -6.26213491e-01
-7.50123858e-01 -2.76166387e-02 -4.73911203e-02 -1.20720971e+00
-3.35692018e-01 -5.54753304e-01 -1.55263329e+00 1.85378999e-01
-2.11578488e-01 5.58178604e-01 5.69891334e-01 6.57820404e-01
2.97659516e-01 4.07530874e-01 7.73929954e-01 -9.32907224e-01
-1.44538796e+00 -8.88571739e-01 -5.12138426e-01 1.58920735e-01
4.59726959e-01 -1.05763829e+00 -3.25265586e-01 -6.86162651e-01] | [3.9662930965423584, 1.766076683998108] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.