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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
a746c19a-f824-4818-81a7-9ce0a4395e93 | how-to-transfer-algorithmic-reasoning | 2110.14056 | null | https://arxiv.org/abs/2110.14056v1 | https://arxiv.org/pdf/2110.14056v1.pdf | How to transfer algorithmic reasoning knowledge to learn new algorithms? | Learning to execute algorithms is a fundamental problem that has been widely studied. Prior work~\cite{veli19neural} has shown that to enable systematic generalisation on graph algorithms it is critical to have access to the intermediate steps of the program/algorithm. In many reasoning tasks, where algorithmic-style reasoning is important, we only have access to the input and output examples. Thus, inspired by the success of pre-training on similar tasks or data in Natural Language Processing (NLP) and Computer Vision, we set out to study how we can transfer algorithmic reasoning knowledge. Specifically, we investigate how we can use algorithms for which we have access to the execution trace to learn to solve similar tasks for which we do not. We investigate two major classes of graph algorithms, parallel algorithms such as breadth-first search and Bellman-Ford and sequential greedy algorithms such as Prim and Dijkstra. Due to the fundamental differences between algorithmic reasoning knowledge and feature extractors such as used in Computer Vision or NLP, we hypothesise that standard transfer techniques will not be sufficient to achieve systematic generalisation. To investigate this empirically we create a dataset including 9 algorithms and 3 different graph types. We validate this empirically and show how instead multi-task learning can be used to achieve the transfer of algorithmic reasoning knowledge. | ['Jian Tang', 'Petar Velickovic', 'Andreea Deac', 'Louis-Pascal A. C. Xhonneux'] | 2021-10-26 | null | http://proceedings.neurips.cc/paper/2021/hash/a2802cade04644083dcde1c8c483ed9a-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/a2802cade04644083dcde1c8c483ed9a-Paper.pdf | neurips-2021-12 | ['learning-to-execute'] | ['computer-code'] | [ 3.62345427e-01 2.92900234e-01 3.10824439e-02 -3.44972491e-01
-1.86781362e-01 -8.17283273e-01 6.14242613e-01 6.30323052e-01
-6.08828247e-01 5.53888798e-01 -1.15392067e-01 -1.07848513e+00
-4.00244772e-01 -1.00724363e+00 -7.71373570e-01 -1.95667431e-01
-2.25762978e-01 7.32490003e-01 4.50534582e-01 -3.55875462e-01
5.66005349e-01 6.47789061e-01 -1.29324079e+00 5.25899112e-01
7.60085285e-01 5.87734699e-01 1.13510147e-01 1.13213325e+00
-6.49918556e-01 9.49484110e-01 -3.73521298e-01 -4.78462607e-01
3.27322930e-01 -6.45151436e-01 -1.41571677e+00 -1.38775319e-01
2.96477079e-01 -5.00442162e-02 -2.20189750e-01 1.15313506e+00
8.44938457e-02 1.49597585e-01 8.52745354e-01 -1.41253018e+00
-1.83124989e-01 8.50506365e-01 -5.04946887e-01 4.19800848e-01
5.49674273e-01 1.24156810e-01 1.02996361e+00 -2.28686005e-01
8.62970293e-01 1.08668458e+00 6.11272633e-01 5.17663062e-01
-1.39128089e+00 -2.66725212e-01 1.27031237e-01 3.21830213e-01
-1.02007365e+00 -1.54764771e-01 6.20910525e-01 -4.31991845e-01
1.08785903e+00 8.81429762e-02 7.40502119e-01 5.29597163e-01
2.57812798e-01 9.87628400e-01 1.09090757e+00 -8.13501894e-01
1.27738774e-01 3.72508019e-02 2.06145898e-01 1.15667284e+00
2.29257584e-01 -2.73715481e-02 -2.52572894e-01 -1.86998991e-03
7.43934810e-01 -3.13904464e-01 -1.62671298e-01 -7.12359071e-01
-1.23225427e+00 8.96408796e-01 6.77815616e-01 4.49598014e-01
-8.77366960e-02 3.39360565e-01 5.37863195e-01 9.42495525e-01
-9.47566107e-02 8.04858565e-01 -5.58678925e-01 -7.54194036e-02
-8.01284611e-01 2.68105119e-01 1.45419848e+00 8.62608969e-01
1.06173921e+00 -2.33631983e-01 2.71249920e-01 3.47458750e-01
-2.63578426e-02 4.86109778e-02 3.38181019e-01 -9.42651212e-01
4.67009306e-01 7.98235536e-01 -2.54866689e-01 -9.38858330e-01
-5.89902282e-01 2.98746629e-03 -5.17726123e-01 3.54499102e-01
7.16863811e-01 -2.47788399e-01 -8.10604930e-01 1.49144387e+00
1.36817813e-01 -1.04018964e-01 2.53181309e-01 6.38082862e-01
3.72812539e-01 3.42509985e-01 -3.63822021e-02 -3.40349436e-01
1.10777330e+00 -8.23315442e-01 -1.32101834e-01 -2.55127490e-01
1.30166733e+00 -6.19495034e-01 1.10759079e+00 5.84547937e-01
-9.36428308e-01 -4.63228405e-01 -9.82609510e-01 4.79845039e-04
-6.62189186e-01 -4.26778853e-01 1.07766020e+00 6.64261878e-01
-1.11394489e+00 9.87982213e-01 -6.66335344e-01 -5.70348263e-01
4.33544278e-01 4.37509567e-01 -1.98763803e-01 -2.71962613e-01
-8.18292320e-01 1.03124440e+00 7.42690384e-01 -1.50378272e-01
-5.48915386e-01 -3.78705531e-01 -5.71383655e-01 -1.95998792e-02
7.31513798e-01 -1.10874939e+00 1.29771924e+00 -1.19508696e+00
-1.32604873e+00 9.40666258e-01 1.73832476e-01 -5.73341846e-01
4.94798094e-01 4.11994793e-02 5.09903766e-02 2.58324295e-01
-2.46385008e-01 4.20922756e-01 8.41526985e-01 -9.35663939e-01
-6.56614959e-01 -3.44838738e-01 6.37556791e-01 1.86610609e-01
-1.62089676e-01 -5.58557846e-02 -1.82624847e-01 -2.29463160e-01
3.10816523e-02 -1.10637999e+00 -2.29141206e-01 -1.05234489e-01
-2.19041497e-01 -2.95370281e-01 5.01066804e-01 -3.20552289e-01
8.73480380e-01 -1.86725247e+00 4.59854335e-01 5.21288753e-01
5.33387423e-01 2.23644108e-01 3.20077315e-02 5.94462931e-01
5.38400598e-02 1.04167394e-01 -1.48833215e-01 2.92433590e-01
9.06706974e-02 5.03829420e-01 -9.91121531e-02 3.19185078e-01
-1.01779386e-01 1.16203475e+00 -1.08165419e+00 -5.72715163e-01
7.72590265e-02 -1.76023364e-01 -7.01499045e-01 6.22310117e-02
-6.40793025e-01 1.47289947e-01 -6.53249979e-01 1.44083217e-01
1.30968750e-01 -4.29853261e-01 3.13299268e-01 2.17012927e-01
1.89417049e-01 3.52321267e-01 -1.19368935e+00 1.79389191e+00
-6.71024144e-01 6.66904688e-01 1.31490663e-01 -1.46937752e+00
5.39556205e-01 8.81976634e-02 1.92333356e-01 -6.49695694e-01
1.14857718e-01 1.64867595e-01 5.35945892e-01 -4.28768426e-01
1.53323650e-01 -3.26285183e-01 2.12242939e-02 7.26850629e-01
-6.91216264e-04 -7.41971970e-01 5.37427604e-01 3.22916985e-01
1.59590757e+00 1.81198403e-01 4.60043997e-01 -3.62783760e-01
5.79203784e-01 6.73844159e-01 3.65304835e-02 8.97992551e-01
1.07843310e-01 4.99632359e-02 7.92550206e-01 -6.27801478e-01
-7.38305032e-01 -8.98120224e-01 4.68368948e-01 1.49482965e+00
-1.97185174e-01 -6.01709604e-01 -7.25486279e-01 -7.85226345e-01
-1.14922374e-01 8.60276461e-01 -3.48934054e-01 -1.49978563e-01
-6.20435596e-01 -4.08242792e-01 4.20161128e-01 5.04329681e-01
3.94509166e-01 -1.28498340e+00 -9.87416804e-01 1.47387177e-01
4.15807992e-01 -9.59439337e-01 -8.42624158e-03 4.77017432e-01
-1.13735580e+00 -1.33524609e+00 -1.94555044e-01 -6.90409958e-01
7.64955461e-01 1.05774730e-01 1.38806105e+00 4.21260357e-01
-4.30363983e-01 8.76936555e-01 -4.46318924e-01 -4.06774014e-01
-6.94886804e-01 3.36705476e-01 -2.72077560e-01 -5.31857312e-01
2.37537131e-01 -7.12667108e-01 -2.72147536e-01 -1.08422257e-01
-9.35738385e-01 2.59016067e-01 9.49683070e-01 5.22454321e-01
1.37599126e-01 5.66266298e-01 4.18324247e-02 -1.43674064e+00
9.35932279e-01 -9.62253734e-02 -6.70422614e-01 4.08008516e-01
-7.28136361e-01 5.92677057e-01 8.69100749e-01 -2.58910567e-01
-6.53616190e-01 5.84025085e-02 9.66918394e-02 -2.70157993e-01
-3.33742201e-01 8.67646635e-01 1.84486240e-01 -3.08881998e-01
8.62449944e-01 2.38435641e-01 5.58291785e-02 2.94662938e-02
5.82590401e-01 1.50521576e-01 2.62943983e-01 -1.11195612e+00
9.18147027e-01 1.68687597e-01 3.14446688e-01 -9.39616144e-01
-6.25967503e-01 -3.90747994e-01 -6.87511563e-01 -5.79770990e-02
8.54325116e-01 -2.40147844e-01 -9.04920876e-01 5.17236330e-02
-1.11152029e+00 -8.55243623e-01 -3.38744402e-01 2.18672782e-01
-1.03363693e+00 3.23262513e-01 -5.07897973e-01 -5.74865758e-01
-9.41857919e-02 -1.12263989e+00 6.43847644e-01 2.66722105e-02
-4.22519296e-01 -1.40128088e+00 -9.66924056e-03 6.91642240e-02
4.72620070e-01 7.29905069e-02 1.47894394e+00 -1.01184571e+00
-6.47239208e-01 8.30828324e-02 -2.96261281e-01 9.20639038e-02
-9.98333097e-03 -1.79425970e-01 -5.79887092e-01 -2.57466137e-01
3.39539424e-02 -4.49481755e-01 7.13557482e-01 -2.45224740e-02
1.20431101e+00 -8.14751089e-02 -3.86471480e-01 4.43434030e-01
1.48631084e+00 2.47915406e-02 5.37660539e-01 4.56154317e-01
7.54072666e-01 7.75312662e-01 2.39418596e-01 -2.63066947e-01
3.19009274e-01 3.58827382e-01 -2.13042405e-02 3.27688962e-01
-3.90132889e-02 -2.22832218e-01 1.14554197e-01 6.52360916e-01
-2.69930929e-01 -3.07244668e-03 -1.43644249e+00 4.68943149e-01
-1.89076376e+00 -8.27608645e-01 -1.13743871e-01 2.04236221e+00
7.78140664e-01 6.10524893e-01 2.76088268e-01 3.24943751e-01
2.83757269e-01 -8.56299773e-02 -3.00037950e-01 -8.39751959e-01
6.82934225e-01 5.49484074e-01 5.09519219e-01 5.71000874e-01
-7.66305566e-01 1.03076255e+00 5.81450415e+00 4.76244777e-01
-9.33837414e-01 -2.50397384e-01 2.24793091e-01 3.35063130e-01
-2.46537477e-01 2.68228322e-01 -4.28187639e-01 -1.13890298e-01
1.08588600e+00 -3.68959963e-01 9.34631348e-01 7.34926760e-01
-3.68430674e-01 -2.63349384e-01 -1.67927837e+00 7.95975327e-01
-3.29790227e-02 -1.15955865e+00 7.22278878e-02 -4.22092825e-02
4.83469337e-01 -5.37748821e-02 -2.64880836e-01 6.32273853e-01
7.10064650e-01 -1.16969132e+00 1.69196114e-01 3.21510971e-01
2.53396094e-01 -5.84757626e-01 5.21967649e-01 6.88330889e-01
-1.14706981e+00 -5.62644191e-02 -2.08955541e-01 -4.70413953e-01
-2.09333092e-01 2.40200788e-01 -1.07931566e+00 7.46703029e-01
2.38236949e-01 4.70389754e-01 -6.55987859e-01 8.31103861e-01
-5.09243071e-01 2.43469447e-01 -4.60727543e-01 -3.80188525e-01
3.83904219e-01 -2.36029223e-01 2.31554136e-01 1.15755355e+00
-5.63076735e-02 2.09962994e-01 4.09828335e-01 6.49732709e-01
7.43411994e-03 7.75655434e-02 -9.76118326e-01 -4.51064587e-01
-2.17784166e-01 1.10537541e+00 -1.33242047e+00 -3.39875042e-01
-7.16271818e-01 9.86677885e-01 5.64457119e-01 2.48726964e-01
-5.00434220e-01 -4.21419054e-01 1.47889495e-01 1.61629617e-01
4.88900065e-01 -5.79853475e-01 -2.29645267e-01 -9.74535584e-01
-1.97967261e-01 -1.04586554e+00 5.91065884e-01 -6.53390884e-01
-1.12050521e+00 3.79111201e-01 2.88736194e-01 -5.17911375e-01
-5.14951527e-01 -9.55143332e-01 -6.47516608e-01 7.17072487e-01
-1.46881688e+00 -9.75952148e-01 -1.97893932e-01 7.77382851e-01
4.79131758e-01 -1.67504191e-01 7.67807961e-01 -1.86497331e-01
-6.48298264e-02 2.69036055e-01 -4.38952416e-01 2.22746328e-01
3.26628327e-01 -1.62148964e+00 3.55138719e-01 6.51838660e-01
6.43347085e-01 7.90435791e-01 8.27755868e-01 -4.59988892e-01
-2.09031367e+00 -6.97166443e-01 4.85293955e-01 -5.42825341e-01
1.10238600e+00 -1.22047208e-01 -8.73165131e-01 9.22088981e-01
8.05629790e-02 7.11452812e-02 3.71172935e-01 4.50617075e-01
-2.76278585e-01 -1.88035686e-02 -6.63250566e-01 6.33306921e-01
1.27074802e+00 -7.02001274e-01 -1.11745644e+00 4.85392183e-01
4.47434574e-01 -4.56505328e-01 -5.54865956e-01 1.68101519e-01
2.54373521e-01 -1.06051362e+00 7.37800062e-01 -9.49400604e-01
4.24864203e-01 -9.65114236e-02 1.43046035e-02 -1.39368331e+00
5.25313709e-03 -6.67248487e-01 -6.62040934e-02 8.09131444e-01
4.28029001e-01 -7.20250368e-01 9.18779790e-01 5.45013964e-01
2.61460450e-02 -5.85901380e-01 -4.57094729e-01 -8.10834706e-01
2.50006706e-01 -6.00760937e-01 1.67804852e-01 8.77086341e-01
2.36046106e-01 6.93919063e-01 4.38586980e-01 -1.44337211e-02
4.03023899e-01 5.95911860e-01 1.07595396e+00 -1.35725200e+00
-6.65172458e-01 -8.05535436e-01 -3.83021355e-01 -9.82586622e-01
5.41272581e-01 -1.36856878e+00 -3.54617313e-02 -1.79290569e+00
7.29684532e-03 -5.24078786e-01 -2.94238091e-01 5.17078221e-01
1.10088766e-01 -2.64326334e-01 8.35174173e-02 8.27781856e-02
-6.87334657e-01 -9.00466070e-02 1.26507890e+00 -7.90856332e-02
-2.95350552e-01 7.42754340e-02 -4.72975045e-01 8.53617966e-01
6.21573091e-01 -5.01585841e-01 -7.90265143e-01 -4.55275953e-01
8.38292062e-01 2.25779772e-01 2.63448119e-01 -1.09533465e+00
6.71283543e-01 -3.23135763e-01 4.82798256e-02 -7.67120272e-02
-1.57793909e-01 -1.06308508e+00 -1.72259882e-01 8.44119251e-01
-3.34164709e-01 1.20857581e-01 2.94784099e-01 4.97361749e-01
-9.67146009e-02 -6.04712605e-01 4.40576226e-01 -6.30064845e-01
-1.11974800e+00 1.03381410e-01 -4.03064728e-01 2.29328960e-01
1.05014586e+00 -1.53260723e-01 -2.33837008e-01 -2.86710739e-01
-7.49248147e-01 3.89856279e-01 6.55530572e-01 4.48834375e-02
5.29392362e-01 -6.78685069e-01 -3.69528532e-01 2.72182599e-02
9.76172462e-02 -8.85142237e-02 -3.33920628e-01 9.74265814e-01
-9.79797959e-01 3.61461550e-01 -3.66908729e-01 -3.85383159e-01
-1.12624347e+00 1.07421947e+00 2.72639513e-01 -2.25576252e-01
-6.40660167e-01 6.96651220e-01 -5.66271655e-02 -4.51415777e-01
-2.79480293e-02 -5.39007366e-01 1.37841657e-01 -8.45577121e-02
2.39753902e-01 1.56726062e-01 4.26116139e-02 1.67343900e-01
-3.29370052e-01 4.86273378e-01 -3.23320359e-01 -1.22383796e-02
1.41351104e+00 2.83127517e-01 -3.54810119e-01 6.03458762e-01
9.52469170e-01 -1.45072564e-01 -6.16756201e-01 -1.21878386e-01
4.51758802e-01 -2.79211015e-01 -4.52899635e-02 -5.21967351e-01
-6.58322334e-01 9.71759737e-01 5.49629740e-02 7.43238211e-01
1.12309468e+00 -9.62496456e-03 4.21978444e-01 1.01622033e+00
6.61677241e-01 -1.02325022e+00 1.04818515e-01 7.22717524e-01
5.52742243e-01 -1.17873907e+00 3.89103472e-01 -4.84037519e-01
-2.43294567e-01 1.55773413e+00 4.16839838e-01 -2.24797726e-01
4.92695391e-01 3.12695652e-01 -2.38721281e-01 -5.96016586e-01
-7.28993177e-01 -3.99706006e-01 8.91582072e-02 5.14905632e-01
2.98569173e-01 -3.01041573e-01 -1.41955674e-01 -8.45688060e-02
-3.74847770e-01 2.39646628e-01 4.49407458e-01 1.27577400e+00
-4.46711212e-01 -1.23803866e+00 -8.18631873e-02 6.62960887e-01
-1.76237702e-01 -2.53653973e-02 -6.34443164e-01 1.17627323e+00
-2.27673322e-01 4.90769893e-01 -2.33323514e-01 -1.61312371e-01
2.01113328e-01 4.82214987e-01 1.09780848e+00 -9.80865121e-01
-5.34174263e-01 -4.67584193e-01 3.27056646e-01 -5.52691817e-01
-4.86768723e-01 -3.43109995e-01 -1.37004900e+00 -4.69909549e-01
-6.77130073e-02 2.56061405e-01 5.67667603e-01 1.28109086e+00
-5.22770472e-02 7.73803413e-01 -6.36467664e-03 -6.46184325e-01
-5.65037310e-01 -5.03968358e-01 -4.60010260e-01 2.56919891e-01
1.32522553e-01 -3.63219947e-01 -3.20868492e-01 2.01176196e-01] | [8.844597816467285, 7.191004276275635] |
edfd33f4-2dc6-4a8a-8ab5-70b938e71054 | natural-language-assisted-sign-language | 2303.12080 | null | https://arxiv.org/abs/2303.12080v1 | https://arxiv.org/pdf/2303.12080v1.pdf | Natural Language-Assisted Sign Language Recognition | Sign languages are visual languages which convey information by signers' handshape, facial expression, body movement, and so forth. Due to the inherent restriction of combinations of these visual ingredients, there exist a significant number of visually indistinguishable signs (VISigns) in sign languages, which limits the recognition capacity of vision neural networks. To mitigate the problem, we propose the Natural Language-Assisted Sign Language Recognition (NLA-SLR) framework, which exploits semantic information contained in glosses (sign labels). First, for VISigns with similar semantic meanings, we propose language-aware label smoothing by generating soft labels for each training sign whose smoothing weights are computed from the normalized semantic similarities among the glosses to ease training. Second, for VISigns with distinct semantic meanings, we present an inter-modality mixup technique which blends vision and gloss features to further maximize the separability of different signs under the supervision of blended labels. Besides, we also introduce a novel backbone, video-keypoint network, which not only models both RGB videos and human body keypoints but also derives knowledge from sign videos of different temporal receptive fields. Empirically, our method achieves state-of-the-art performance on three widely-adopted benchmarks: MSASL, WLASL, and NMFs-CSL. Codes are available at https://github.com/FangyunWei/SLRT. | ['Brian Mak', 'Fangyun Wei', 'Ronglai Zuo'] | 2023-03-21 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zuo_Natural_Language-Assisted_Sign_Language_Recognition_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zuo_Natural_Language-Assisted_Sign_Language_Recognition_CVPR_2023_paper.pdf | cvpr-2023-1 | ['sign-language-recognition'] | ['computer-vision'] | [ 2.36495018e-01 -4.68021899e-01 -4.74763483e-01 -4.45184529e-01
-4.36463237e-01 -5.47847092e-01 6.81954801e-01 -9.07098711e-01
-3.80851179e-01 3.92188221e-01 5.28741121e-01 7.01316074e-02
-5.65836690e-02 -2.40774289e-01 -4.94011492e-01 -9.49485719e-01
2.97437578e-01 -2.32632041e-01 2.09143400e-01 -1.82443559e-01
-8.59153271e-03 3.77545327e-01 -1.69744849e+00 3.92147481e-01
8.37673843e-01 1.10068285e+00 9.75529552e-02 3.60327542e-01
-2.82701910e-01 1.12825537e+00 -2.43443400e-01 -3.55407715e-01
3.46658736e-01 -5.91102242e-01 -3.41979265e-01 2.37328976e-01
9.71026897e-01 -4.89083886e-01 -7.03417718e-01 1.14338958e+00
3.85637969e-01 3.73048298e-02 6.72005892e-01 -1.59954131e+00
-9.74268198e-01 2.15751797e-01 -5.09011745e-01 -4.05327082e-01
1.70677930e-01 5.58130682e-01 1.03865027e+00 -8.39364052e-01
7.52637625e-01 1.28723288e+00 4.61071879e-01 8.98407817e-01
-7.79559731e-01 -8.88729453e-01 3.88863295e-01 2.69226462e-01
-1.17281914e+00 -3.41689050e-01 8.59478176e-01 -4.97196496e-01
4.54147190e-01 3.21125776e-01 8.77974093e-01 1.18235505e+00
-3.42255592e-01 1.30680490e+00 1.38820899e+00 -2.80994236e-01
-6.57651275e-02 -3.18756849e-01 1.88801475e-02 1.02879047e+00
-5.54735772e-02 2.37896979e-01 -7.27478504e-01 1.13500454e-01
8.04641485e-01 2.38207698e-01 -5.60246170e-01 -4.09661382e-01
-1.41330576e+00 4.66409385e-01 5.22059441e-01 1.09317519e-01
-2.47805953e-01 4.00226742e-01 2.33921140e-01 2.30144441e-01
-2.47263193e-01 -2.03235596e-01 -3.34918648e-01 8.13791249e-03
-6.23206913e-01 -8.06653649e-02 3.95303518e-01 1.00336587e+00
4.98361349e-01 1.93763986e-01 -3.17069471e-01 8.85720670e-01
8.21161568e-01 1.11496174e+00 6.59499526e-01 -9.55198526e-01
4.06622618e-01 7.14120626e-01 -8.28647912e-02 -6.53634369e-01
-1.76385954e-01 1.03934899e-01 -9.24230635e-01 4.83893543e-01
5.86190641e-01 8.01875666e-02 -1.75472021e+00 1.86044788e+00
8.41047987e-02 1.44040942e-01 5.95009997e-02 1.32859957e+00
1.16140056e+00 3.73228639e-01 3.85411382e-01 1.95978552e-01
1.40196407e+00 -1.05658555e+00 -6.29712105e-01 -5.58707304e-02
3.89460444e-01 -6.59289956e-01 1.28138375e+00 3.17098945e-01
-6.91936016e-01 -3.09532344e-01 -6.78661644e-01 -2.27343440e-01
-4.05525804e-01 6.03065968e-01 7.32242405e-01 3.57509792e-01
-8.07079494e-01 1.28746867e-01 -7.51786709e-01 -3.31709534e-01
5.03595471e-01 1.38149664e-01 -4.39818233e-01 -2.74985105e-01
-9.14061666e-01 6.60286665e-01 1.20278329e-01 3.94983619e-01
-6.14916921e-01 -3.25139880e-01 -9.78830934e-01 -4.26253051e-01
3.33230704e-01 -5.79387069e-01 1.03058016e+00 -1.27785325e+00
-1.65711093e+00 9.65265810e-01 -2.94840604e-01 1.49827048e-01
7.02969849e-01 -8.80624652e-02 -6.14470124e-01 8.70722830e-02
-2.34584764e-01 8.95871580e-01 1.22680902e+00 -1.29182029e+00
-6.36263132e-01 -6.06266335e-02 -6.55969009e-02 1.64565757e-01
-6.18557781e-02 3.22374135e-01 -6.68931425e-01 -8.81015837e-01
3.94114316e-01 -9.60189998e-01 1.70592651e-01 6.66114330e-01
-4.43823993e-01 -3.37610781e-01 8.41119647e-01 -8.74698162e-01
8.42233837e-01 -2.38715792e+00 2.47546077e-01 4.59862739e-01
2.61889279e-01 5.03417730e-01 -6.65131629e-01 8.84925425e-02
2.68149406e-01 -2.33076468e-01 -4.25170779e-01 -1.61013693e-01
1.96124390e-01 5.87615848e-01 -2.93006122e-01 4.27511573e-01
1.71801120e-01 1.16999888e+00 -9.72601354e-01 -5.85354924e-01
4.08334255e-01 6.94316804e-01 -1.33502886e-01 -3.95072475e-02
-1.80668876e-01 4.18734998e-01 -4.31340098e-01 1.26312709e+00
5.74998379e-01 -2.42501125e-01 5.62510900e-02 -5.60852289e-01
9.55775604e-02 -1.05985269e-01 -1.14579296e+00 1.60396445e+00
-2.64715791e-01 5.99230170e-01 -4.48077805e-02 -6.05907142e-01
7.45819926e-01 1.58489272e-01 3.94898742e-01 -7.36480236e-01
2.56455600e-01 4.50016439e-01 -4.20392007e-02 -6.05664253e-01
7.85832759e-03 -5.88155426e-02 1.23906605e-01 2.36365139e-01
1.44026056e-01 7.80052394e-02 2.77731061e-01 -7.25531280e-02
6.60610616e-01 4.09424543e-01 1.05546109e-01 3.50137085e-01
6.14451051e-01 -3.87264699e-01 6.85089231e-01 4.97071534e-01
-4.63166863e-01 6.42904758e-01 9.70225185e-02 -2.40020171e-01
-5.58151364e-01 -1.51796818e+00 3.02220453e-02 1.11842644e+00
3.42269391e-01 -2.04502225e-01 -2.15039968e-01 -8.31266284e-01
2.15702921e-01 2.65078098e-01 -4.47805375e-01 1.05534516e-01
-6.02044702e-01 -2.76098669e-01 6.58927560e-01 7.12911129e-01
6.75657928e-01 -1.40334582e+00 -5.80496609e-01 -3.52175206e-01
-1.37046844e-01 -1.13652682e+00 -8.45564067e-01 -3.02569807e-01
-4.85466123e-01 -1.19893205e+00 -1.26937819e+00 -9.52156067e-01
9.22451258e-01 2.36772418e-01 5.29483914e-01 6.49124607e-02
-2.86357641e-01 6.07829154e-01 -4.54174370e-01 -1.63581539e-02
-1.32535145e-01 -6.34329975e-01 4.92058471e-02 3.53654683e-01
4.64410752e-01 -4.01969105e-01 -5.25093257e-01 3.54622513e-01
-1.02467132e+00 3.25738490e-01 8.35539639e-01 9.87742484e-01
6.15612388e-01 -6.43438339e-01 9.38449726e-02 -3.88640761e-02
3.77675861e-01 4.17516045e-02 -4.81385171e-01 7.49221206e-01
-1.57129675e-01 2.27117419e-01 4.75877076e-01 -8.27487051e-01
-9.07545269e-01 1.69802919e-01 1.18891388e-01 -5.44826567e-01
-3.26222837e-01 1.74281850e-01 -2.98238307e-01 -4.74794894e-01
1.14895515e-01 6.50672853e-01 2.57985651e-01 -4.04546648e-01
8.73984575e-01 8.02737713e-01 7.24597752e-01 -3.66285324e-01
8.39235961e-01 6.54033244e-01 -1.80809982e-02 -8.88325095e-01
-6.51698112e-01 -4.33607697e-01 -4.17773694e-01 -4.68427390e-01
8.37049305e-01 -8.20192456e-01 -6.93420172e-01 1.03373992e+00
-9.86784101e-01 -4.04476702e-01 -3.19614947e-01 7.67528832e-01
-5.86606264e-01 7.18658209e-01 -6.06297374e-01 -6.53408468e-01
-2.97021776e-01 -1.00882149e+00 1.12899268e+00 4.91893560e-01
-5.83832823e-02 -5.33676386e-01 -3.16901594e-01 2.70323187e-01
2.79554605e-01 2.16734707e-01 6.78131342e-01 -1.84606090e-01
-8.11649680e-01 8.12025624e-04 -7.13709950e-01 8.28297913e-01
4.35213745e-01 -4.55877222e-02 -8.44685853e-01 -7.59731531e-02
-6.80868804e-01 -5.77323377e-01 1.18663394e+00 3.96910310e-01
7.97385573e-01 -3.70221615e-01 -9.98487249e-02 9.28646863e-01
1.12468135e+00 1.63410544e-01 5.55055678e-01 3.93496007e-02
9.92349923e-01 4.80073929e-01 2.21866518e-01 1.57590449e-01
4.56585109e-01 5.42639673e-01 8.28688070e-02 -2.05242351e-01
-7.32391059e-01 -4.42277282e-01 6.46729469e-01 9.55118060e-01
-4.99502212e-01 -7.21429195e-03 -6.42327547e-01 4.77837592e-01
-1.84809172e+00 -8.42581153e-01 1.13593094e-01 1.95075154e+00
7.97880709e-01 -3.89589101e-01 -1.62167568e-02 -7.87958428e-02
5.29666781e-01 3.28378797e-01 -7.72786498e-01 1.18789591e-01
-6.57405555e-01 6.15626797e-02 6.37180507e-01 2.87400663e-01
-1.05288494e+00 1.09305871e+00 5.08684158e+00 8.12696517e-01
-1.36949968e+00 -5.20449504e-02 -1.35438830e-01 -1.38085499e-01
-2.09509492e-01 -9.96572375e-02 -4.80517060e-01 6.27765775e-01
-5.58593646e-02 2.29817316e-01 6.45282388e-01 6.60321236e-01
1.18840858e-01 1.94078326e-01 -7.76693463e-01 1.22447062e+00
3.33909124e-01 -9.88679767e-01 2.73617059e-01 -1.19469322e-01
7.16657937e-01 3.59073669e-01 3.40763219e-02 1.83272734e-02
4.11880761e-01 -7.42305219e-01 7.66020775e-01 9.03584599e-01
1.04739809e+00 -1.44572213e-01 4.02928501e-01 -2.08382867e-02
-1.31501067e+00 -5.31417765e-02 8.29579830e-02 2.72551030e-01
4.47631508e-01 6.66884556e-02 -7.16021731e-02 2.93612093e-01
4.59020793e-01 1.07239950e+00 -4.44920003e-01 1.18298674e+00
-8.45692873e-01 4.91213590e-01 -6.19115591e-01 -1.16182044e-01
2.83155441e-01 -2.51407593e-01 5.02602994e-01 1.18094933e+00
2.54606426e-01 2.09030211e-02 1.47061676e-01 7.97725618e-01
7.91575387e-02 3.65458541e-02 -3.61081898e-01 -2.25277960e-01
1.87388018e-01 8.63271296e-01 -3.44405621e-01 -3.62433344e-01
-7.06193626e-01 1.25311697e+00 -1.74564406e-01 8.28745365e-01
-5.99046469e-01 -3.02468598e-01 9.23582852e-01 -3.04873288e-01
2.39687860e-01 -3.73268962e-01 4.79226969e-02 -1.48209524e+00
3.98306668e-01 -8.84950459e-01 3.09067190e-01 -9.37694073e-01
-1.53083265e+00 3.04805517e-01 -1.89265564e-01 -1.53783393e+00
-4.28812243e-02 -1.06944752e+00 -3.95068944e-01 8.37846398e-01
-1.89669514e+00 -1.74091828e+00 -6.05386853e-01 9.63337302e-01
2.74879843e-01 -2.67999709e-01 5.17018139e-01 2.97587633e-01
-9.96392965e-02 7.26608694e-01 1.41755715e-01 6.14359081e-01
6.71004534e-01 -8.64734352e-01 1.70132443e-01 7.55704224e-01
3.46154034e-01 5.10901213e-01 3.57468873e-02 -6.03801012e-01
-1.35805047e+00 -9.59040344e-01 8.48543108e-01 -9.88912433e-02
7.21511900e-01 -5.67557663e-02 -6.66277289e-01 4.87978071e-01
-2.45055467e-01 2.44528636e-01 4.53472257e-01 -3.58328283e-01
-8.64662647e-01 -3.40981372e-02 -7.97730207e-01 8.07064772e-01
1.43389916e+00 -8.02056968e-01 -7.05582857e-01 3.10552329e-01
3.34203064e-01 -3.27663571e-01 -5.30241787e-01 5.33569574e-01
1.16555917e+00 -6.57046676e-01 1.04183269e+00 -6.70627177e-01
2.56761491e-01 -6.76720202e-01 -3.63408983e-01 -1.07689917e+00
6.90287724e-02 -3.42042714e-01 -2.18384475e-01 1.05180681e+00
-1.19606815e-02 -8.22756469e-01 5.43973744e-01 4.29225206e-01
1.21724255e-01 -4.85451818e-01 -9.37923431e-01 -1.08150876e+00
-3.11802536e-01 -4.34153050e-01 5.35513401e-01 9.26172793e-01
-2.64541060e-01 -1.74528867e-01 -7.03854382e-01 -9.28900689e-02
8.78052473e-01 2.28865281e-01 6.41472399e-01 -9.80556846e-01
-2.09260061e-01 -7.20545769e-01 -6.50671840e-01 -1.24594629e+00
1.88732043e-01 -9.74044859e-01 1.98100865e-01 -1.60563529e+00
-3.75481807e-02 -3.52950633e-01 -5.98761082e-01 9.93603766e-01
7.58162886e-02 4.01222110e-01 7.33025134e-01 4.24692899e-01
-5.39040327e-01 6.54315114e-01 1.75280344e+00 -2.62856215e-01
-1.75185785e-01 -2.44949892e-01 -3.18826735e-01 1.00147223e+00
5.25450766e-01 -6.57398850e-02 -1.59556285e-01 -5.27053893e-01
-1.36602804e-01 -2.74123788e-01 8.13783765e-01 -8.51056576e-01
1.94358215e-01 -4.77347285e-01 1.16775237e-01 -3.96940172e-01
4.53696489e-01 -7.62076676e-01 -1.35581449e-01 4.11595017e-01
-3.20444554e-01 -4.34954852e-01 -1.59712404e-01 4.90684867e-01
-3.62878680e-01 1.54343888e-01 6.01279378e-01 -1.37757482e-02
-1.30344510e+00 3.77144724e-01 -1.83149114e-01 5.32149710e-02
6.55201972e-01 -4.08068568e-01 -4.97803360e-01 -3.89265925e-01
-5.21614075e-01 2.79312164e-01 3.33324790e-01 7.35519469e-01
9.11174119e-01 -1.52450359e+00 -6.22761607e-01 5.14540493e-01
5.23376346e-01 -2.77349859e-01 3.99166197e-01 9.15907681e-01
-3.36915553e-01 2.12655440e-01 -3.50909024e-01 -5.55037379e-01
-1.46666229e+00 5.53788580e-02 3.52376491e-01 4.39831436e-01
-8.63607466e-01 8.82996142e-01 2.61621058e-01 -4.35814947e-01
4.37415570e-01 -7.86648929e-01 -3.04480689e-03 -1.21726036e-01
4.73734409e-01 1.31022677e-01 -6.07354581e-01 -1.08400059e+00
-4.26998943e-01 1.10114610e+00 2.20507979e-01 3.41368206e-02
8.74360442e-01 6.06512763e-02 -2.55790167e-02 2.28301525e-01
1.07389152e+00 1.26887169e-02 -1.43732345e+00 -5.82355142e-01
-2.01853484e-01 -6.08419716e-01 -7.36832172e-02 -1.09389639e+00
-1.21136332e+00 7.49267220e-01 7.65461206e-01 -5.84589660e-01
1.43960071e+00 1.73822343e-01 9.59997714e-01 3.96090329e-01
1.80172846e-01 -1.04076791e+00 -1.64555144e-02 4.78775918e-01
9.26401973e-01 -1.39440310e+00 -2.58442104e-01 -2.14103982e-01
-8.90596688e-01 9.59945619e-01 3.10771227e-01 1.70633253e-02
5.48531950e-01 4.15572599e-02 6.34596705e-01 2.15843730e-02
-1.62600055e-01 -7.28621483e-01 8.08160305e-01 7.85458446e-01
6.29305318e-02 3.78774047e-01 -4.30656314e-01 4.52054769e-01
7.88965821e-02 3.98246139e-01 3.12247351e-02 9.11443055e-01
-3.09165776e-01 -1.04470229e+00 -1.76518664e-01 3.75806659e-01
1.94043860e-01 -2.41172716e-01 -5.64773023e-01 7.98912287e-01
4.25508589e-01 6.12885833e-01 -2.55192965e-01 -4.40010607e-01
2.47626618e-01 1.60262376e-01 7.02591836e-01 -3.20120342e-02
-4.57879603e-02 -2.62119770e-02 -1.21503681e-01 -6.06218815e-01
-8.28861117e-01 -6.21721208e-01 -1.39231980e+00 1.83180198e-01
-7.08977357e-02 -5.64133286e-01 6.27656341e-01 1.12347984e+00
4.94630337e-02 2.25874111e-01 1.83717579e-01 -6.12313509e-01
-5.84697783e-01 -8.07877481e-01 -8.19541454e-01 8.15269053e-01
5.39244115e-01 -9.33756351e-01 -4.01106566e-01 3.36183786e-01] | [9.203834533691406, -6.50338077545166] |
6b0f6b51-f84f-4349-a878-c53278af26d5 | for-sale-state-action-representation-learning | 2306.02451 | null | https://arxiv.org/abs/2306.02451v1 | https://arxiv.org/pdf/2306.02451v1.pdf | For SALE: State-Action Representation Learning for Deep Reinforcement Learning | In the field of reinforcement learning (RL), representation learning is a proven tool for complex image-based tasks, but is often overlooked for environments with low-level states, such as physical control problems. This paper introduces SALE, a novel approach for learning embeddings that model the nuanced interaction between state and action, enabling effective representation learning from low-level states. We extensively study the design space of these embeddings and highlight important design considerations. We integrate SALE and an adaptation of checkpoints for RL into TD3 to form the TD7 algorithm, which significantly outperforms existing continuous control algorithms. On OpenAI gym benchmark tasks, TD7 has an average performance gain of 276.7% and 50.7% over TD3 at 300k and 5M time steps, respectively, and works in both the online and offline settings. | ['David Meger', 'Doina Precup', 'Shixiang Shane Gu', 'Edward J. Smith', 'Wei-Di Chang', 'Scott Fujimoto'] | 2023-06-04 | null | null | null | null | ['openai-gym', 'continuous-control'] | ['playing-games', 'playing-games'] | [-2.77313776e-03 -1.09155178e-01 -5.92413068e-01 1.62608445e-01
-5.31884432e-01 -4.87072617e-01 6.95091724e-01 3.61297935e-01
-7.15702653e-01 5.53792596e-01 2.24776804e-01 -3.56850266e-01
-2.66106009e-01 -6.05091870e-01 -7.67812967e-01 -8.17962468e-01
-6.62046313e-01 5.30469902e-02 1.03709966e-01 -2.94607520e-01
1.32920027e-01 4.67597902e-01 -1.53417051e+00 -2.33937338e-01
4.91794974e-01 8.45052183e-01 1.56921864e-01 8.89075100e-01
2.13110551e-01 1.00918329e+00 -6.45564735e-01 6.60411656e-01
1.97279379e-01 -2.57540554e-01 -6.41565681e-01 -3.45526598e-02
2.62010694e-01 -4.02540892e-01 -1.03928745e+00 4.75117058e-01
8.14413190e-01 5.35054207e-01 4.43875611e-01 -1.35513985e+00
-5.29119253e-01 2.91035146e-01 -4.71636832e-01 4.11366910e-01
1.78100720e-01 7.61140406e-01 7.04634786e-01 -3.54550391e-01
3.78443509e-01 1.23949158e+00 3.40675056e-01 7.35746264e-01
-1.60407770e+00 -5.69560051e-01 4.29937035e-01 3.76703918e-01
-1.03800237e+00 -3.97353292e-01 3.89381409e-01 -1.96805894e-01
1.17908704e+00 -2.03439057e-01 9.77944374e-01 1.03938866e+00
5.84715068e-01 9.52619195e-01 1.37755179e+00 -2.51299679e-01
5.72233319e-01 -5.41302145e-01 -2.56699715e-02 6.80187583e-01
4.16274369e-02 5.80695689e-01 -6.31866097e-01 8.05678517e-02
1.11528432e+00 -1.92567799e-02 -1.46582663e-01 -6.62451744e-01
-1.24081481e+00 7.39280581e-01 6.15343571e-01 -1.53400108e-01
-4.22372341e-01 1.11523342e+00 5.79256237e-01 4.79933858e-01
-8.65716115e-03 5.03343523e-01 -3.92460704e-01 -7.71812320e-01
-2.92775422e-01 4.06810910e-01 4.60635066e-01 7.35477686e-01
4.97684866e-01 5.59498608e-01 -2.73673564e-01 5.76635957e-01
1.28140703e-01 5.69832444e-01 2.82297760e-01 -1.26760054e+00
1.34949103e-01 4.41410810e-01 2.25842044e-01 -5.11843562e-01
-5.68522811e-01 -2.93154150e-01 -6.33553326e-01 6.45449877e-01
1.31096274e-01 -3.09424907e-01 -9.61665988e-01 1.82416761e+00
5.58670938e-01 5.03288627e-01 1.15527079e-01 8.08549166e-01
2.79040068e-01 8.64134669e-01 3.95822600e-02 -1.14777535e-02
9.67524350e-01 -9.70248938e-01 -7.98492432e-01 -4.27334636e-01
5.64423263e-01 -1.33267507e-01 1.34619212e+00 3.38468462e-01
-9.20885086e-01 -6.51064813e-01 -1.25073445e+00 1.16819166e-01
-2.61498332e-01 -4.61010635e-02 8.16987276e-01 2.52674907e-01
-1.05753696e+00 8.77540708e-01 -1.22250032e+00 -1.91819727e-01
2.97408909e-01 6.15814030e-01 -1.31080255e-01 -2.34012231e-01
-1.05268991e+00 1.08401251e+00 1.19862810e-01 -1.00783892e-01
-1.43346989e+00 -7.82524705e-01 -1.09868944e+00 2.00176872e-02
6.98113561e-01 -4.40003723e-01 1.38124478e+00 -5.83114065e-02
-1.88309026e+00 1.59724742e-01 3.10082614e-01 -6.15387678e-01
1.30944312e-01 -5.82243025e-01 -2.86973745e-01 1.92228392e-01
-1.45107999e-01 7.33813047e-01 8.52163911e-01 -9.07025337e-01
-3.59213084e-01 -3.09559166e-01 4.70900297e-01 4.84563589e-01
-1.69796333e-01 -5.93462706e-01 -3.29775780e-01 -2.39386007e-01
-2.53355891e-01 -1.21733773e+00 -5.84214747e-01 3.06008130e-01
2.42525816e-01 -3.70134771e-01 8.41989636e-01 -1.26172394e-01
1.06702650e+00 -2.29164600e+00 3.94768447e-01 -1.86542928e-01
1.77659720e-01 3.18637282e-01 -2.56679654e-01 7.24931121e-01
9.22342092e-02 -1.81727469e-01 4.47689667e-02 -3.20262946e-02
2.27608234e-01 6.80081010e-01 -1.12330154e-01 7.05734730e-01
3.19460303e-01 9.64092135e-01 -1.19943857e+00 -2.60237962e-01
7.10174263e-01 4.12835628e-01 -5.86810350e-01 1.86545312e-01
-3.52731675e-01 4.71249163e-01 -5.10505915e-01 2.56292433e-01
2.06623763e-01 -5.76000549e-02 2.41192684e-01 1.28876165e-01
-2.29615256e-01 3.80600393e-01 -1.22927725e+00 2.06487870e+00
-6.62761748e-01 6.60255134e-01 3.84004377e-02 -9.00830030e-01
5.67379057e-01 2.35640153e-01 8.79356325e-01 -1.10774863e+00
3.54448482e-02 -3.14771235e-01 -1.58861633e-02 -4.58314776e-01
5.05152524e-01 8.42619538e-02 -5.63917220e-01 5.44345319e-01
1.17851533e-01 -3.46531421e-01 8.23626295e-02 2.29783013e-01
1.50269318e+00 5.10260105e-01 2.50190556e-01 -2.37343207e-01
-8.74671265e-02 -5.72623238e-02 4.45957512e-01 7.59916306e-01
-5.84195375e-01 -2.57138237e-02 7.06861138e-01 -4.58523810e-01
-7.64978588e-01 -1.11723924e+00 2.14476675e-01 1.10550678e+00
3.13650966e-01 -6.37322068e-01 -3.39852661e-01 -4.10986215e-01
2.93535918e-01 5.35461783e-01 -6.82771921e-01 -8.00765693e-01
-6.13220692e-01 -3.88219237e-01 5.29678226e-01 8.06558728e-01
3.86539578e-01 -1.13112676e+00 -9.54824507e-01 4.34290677e-01
3.49798411e-01 -9.44961846e-01 -5.26435077e-01 6.01208210e-01
-8.00159216e-01 -1.14982176e+00 -1.97670400e-01 -5.46638370e-01
3.04628521e-01 2.96753347e-01 9.48569477e-01 -3.04748286e-02
-4.91847873e-01 8.23482394e-01 -3.77941370e-01 -1.74948916e-01
-2.42777377e-01 -8.83813128e-02 3.96576554e-01 -5.66109478e-01
-1.32565543e-01 -5.77121139e-01 -6.48217678e-01 6.82297498e-02
-9.61297691e-01 -2.18515724e-01 6.08942628e-01 1.21647155e+00
5.64517915e-01 1.42286497e-03 5.94246387e-01 -2.76952654e-01
6.86777532e-01 -2.17009202e-01 -7.14436233e-01 -6.98853657e-02
-8.46981406e-01 3.73253256e-01 5.75241208e-01 -7.85373867e-01
-5.81426859e-01 8.20581913e-02 3.95345269e-03 -5.94178617e-01
-3.59600522e-02 3.76180053e-01 1.57226548e-01 -9.58998650e-02
6.90162420e-01 3.71249199e-01 4.18503940e-01 -1.41195610e-01
6.71182573e-01 3.91148269e-01 4.11211759e-01 -9.62142825e-01
6.81210577e-01 1.93959177e-01 1.22309141e-01 -9.15899575e-01
-6.06510937e-01 -3.50286573e-01 -3.50032926e-01 -3.15154940e-01
7.07462847e-01 -8.87169182e-01 -1.18228483e+00 3.81668776e-01
-4.87771869e-01 -1.18647754e+00 -6.97378457e-01 4.69339609e-01
-1.06861663e+00 1.34314507e-01 -8.49452794e-01 -7.43773758e-01
-6.41420558e-02 -1.11145353e+00 9.99065816e-01 4.20486391e-01
1.04770429e-01 -9.06638324e-01 2.87264556e-01 1.13275386e-02
4.28191692e-01 2.69041210e-01 9.64756787e-01 2.78756656e-02
-4.92389888e-01 5.72053492e-02 1.64741397e-01 3.48366529e-01
1.09378517e-01 -3.04803640e-01 -7.68537521e-01 -7.51719952e-01
-4.35480326e-01 -8.31710935e-01 6.52379334e-01 3.55248541e-01
1.20276678e+00 -1.78936459e-02 -2.85467096e-02 3.07898670e-01
1.46991849e+00 2.90691584e-01 7.25587964e-01 3.43730509e-01
5.43540716e-01 -3.02624330e-02 8.53400707e-01 6.90555990e-01
4.46223199e-01 7.10077643e-01 6.89250171e-01 -7.83505738e-02
-9.71574411e-02 -4.44521159e-01 7.89083540e-01 8.56155276e-01
2.33767122e-01 7.74629861e-02 -6.15378559e-01 4.41696346e-01
-1.92299557e+00 -8.03792834e-01 5.02324402e-01 2.16450834e+00
9.50213790e-01 2.91218191e-01 1.55203193e-01 1.55167088e-01
2.14552835e-01 5.15974879e-01 -1.10524857e+00 -5.96956015e-01
4.00642633e-01 5.52000105e-01 7.00771391e-01 2.85368234e-01
-1.10482728e+00 1.14783394e+00 7.03851843e+00 7.14145124e-01
-1.09455895e+00 -8.82447064e-02 7.09725395e-02 -3.67699385e-01
1.82949558e-01 -5.90904839e-02 -5.20048320e-01 2.53096849e-01
1.22857726e+00 -3.71087007e-02 8.32940221e-01 8.06510329e-01
5.25192380e-01 -2.71896183e-01 -1.19080508e+00 8.73669088e-01
-1.76077351e-01 -1.32036984e+00 -4.33367074e-01 1.69962093e-01
7.38912404e-01 1.81035876e-01 2.02383652e-01 1.05697155e+00
6.54938459e-01 -1.24071026e+00 4.73210007e-01 1.50470898e-01
8.51406097e-01 -8.79821241e-01 2.70841300e-01 3.58046114e-01
-1.31070340e+00 -4.07616198e-01 -2.93447345e-01 -3.36142421e-01
-1.18367095e-02 3.70352678e-02 -4.96464252e-01 2.73738205e-01
7.14999557e-01 8.91583264e-01 -3.37339103e-01 8.41546535e-01
-2.90921956e-01 6.80122137e-01 -3.75553012e-01 -1.58143282e-01
5.03829896e-01 -6.44300804e-02 2.31119573e-01 6.19785786e-01
-1.45005643e-01 -4.46589366e-02 6.38781130e-01 6.18362427e-01
2.04874262e-01 -5.55392087e-01 -8.08628857e-01 -3.25232029e-01
5.46362519e-01 1.11911643e+00 -3.32031220e-01 -2.02469513e-01
-1.38738647e-01 8.72176588e-01 3.86212051e-01 2.83856392e-01
-1.14628232e+00 -2.44667649e-01 1.25496674e+00 -1.67694911e-01
4.47206497e-01 -9.05312181e-01 1.92860648e-01 -7.85721183e-01
-7.66118094e-02 -9.85397637e-01 6.94654211e-02 -5.11342585e-01
-7.28379309e-01 -6.17526285e-02 1.79808721e-01 -1.20604312e+00
-2.68037170e-01 -6.00265086e-01 -5.50986111e-01 2.82270283e-01
-1.60563433e+00 -7.41948843e-01 -1.26067087e-01 5.88908732e-01
6.26016915e-01 3.09122354e-01 1.03481674e+00 2.86038578e-01
-7.91362464e-01 4.46253300e-01 4.21020508e-01 -7.10695088e-02
4.93610591e-01 -1.54440749e+00 4.36196327e-01 4.26640958e-01
9.39301103e-02 4.69786376e-01 6.66343272e-01 -2.10564718e-01
-2.15851116e+00 -1.11573458e+00 -1.17284264e-02 -1.28251672e-01
7.15538323e-01 -4.31835651e-01 -5.84714532e-01 5.24234593e-01
3.05387437e-01 2.19267011e-01 4.17870253e-01 7.69624189e-02
-1.56892881e-01 -1.05238453e-01 -8.41845214e-01 9.86820042e-01
1.07868779e+00 -5.90247273e-01 -2.95186847e-01 2.53136694e-01
8.78707588e-01 -6.76535308e-01 -1.26253784e+00 8.02265331e-02
4.91808414e-01 -4.95678693e-01 1.05028725e+00 -7.70138860e-01
1.36744484e-01 -3.93259346e-01 -3.63114327e-02 -1.79894102e+00
-4.48987305e-01 -8.02513719e-01 -6.01305664e-01 5.80004036e-01
-4.30033579e-02 -5.14744282e-01 6.64295733e-01 1.33920506e-01
-3.23200315e-01 -1.19115257e+00 -1.03682578e+00 -1.09799004e+00
2.60102749e-01 -3.80790442e-01 3.22860092e-01 5.68490684e-01
1.98870510e-01 3.00305247e-01 -4.89463657e-01 1.71634451e-01
4.60270256e-01 1.30585656e-01 8.92547429e-01 -6.84328318e-01
-3.61871094e-01 -2.07475588e-01 -6.19029880e-01 -1.40070438e+00
2.13447019e-01 -5.63013017e-01 1.99803427e-01 -1.76769400e+00
-1.85808986e-01 -4.80193555e-01 -6.46509230e-01 6.19266927e-01
9.01110750e-03 -2.04673678e-01 2.78901607e-01 -3.23888242e-01
-7.97034740e-01 1.17410183e+00 1.42123270e+00 -2.86712706e-01
-1.65793315e-01 -2.96766758e-01 -4.26263630e-01 3.07553172e-01
1.03429234e+00 -3.86805266e-01 -7.62828887e-01 -3.90059292e-01
-1.71141282e-01 1.92094818e-01 4.09833074e-01 -1.31626010e+00
6.57318980e-02 -5.69442391e-01 1.63413987e-01 -3.65993440e-01
7.19383597e-01 -6.72890723e-01 -3.65030497e-01 8.89541745e-01
-6.34029567e-01 3.33416551e-01 4.59896833e-01 1.06270325e+00
1.98009700e-01 3.35851371e-01 7.50757098e-01 1.28266096e-01
-1.11835730e+00 3.87715280e-01 -7.26068854e-01 2.77528286e-01
1.20460665e+00 3.34397890e-02 -4.37935710e-01 -4.77048367e-01
-7.13253796e-01 7.75614858e-01 3.87065232e-01 5.72472632e-01
7.84941554e-01 -1.52169740e+00 -1.53957114e-01 6.37009889e-02
1.36645108e-01 7.38493502e-02 3.54075462e-01 8.06597590e-01
-3.66121352e-01 1.62422016e-01 -4.40082431e-01 -6.45144701e-01
-8.54225814e-01 5.22076190e-01 2.34882236e-01 -2.77497679e-01
-1.06577468e+00 3.03820401e-01 -2.19379768e-01 -4.19643104e-01
3.98938894e-01 -5.66503644e-01 -1.29390538e-01 -3.54284108e-01
3.88507396e-01 3.83911073e-01 -2.77570337e-01 -9.86849368e-02
-1.96607858e-01 2.51350701e-01 6.51949719e-02 -1.97716311e-01
1.33511245e+00 1.49827436e-01 5.26643336e-01 7.38968372e-01
9.90288079e-01 -5.86433113e-01 -1.98505461e+00 -1.89210430e-01
-6.35539070e-02 -2.78551400e-01 2.83197701e-01 -6.41594887e-01
-9.84834015e-01 9.16735888e-01 1.08179879e+00 -1.29446089e-01
8.18674445e-01 -1.79655939e-01 7.82887876e-01 6.05439723e-01
6.44700587e-01 -1.47528803e+00 6.79542959e-01 6.71301425e-01
6.24250233e-01 -1.13471949e+00 1.17664084e-01 2.41533175e-01
-6.48010135e-01 7.49826431e-01 8.44725847e-01 -5.08848846e-01
4.78906333e-01 4.47199166e-01 -1.78319216e-01 -2.68845379e-01
-1.27446651e+00 -3.41564327e-01 -2.77074218e-01 8.55240583e-01
6.38613701e-02 2.42294729e-01 6.15213141e-02 7.67235607e-02
3.04406762e-01 1.36890039e-02 6.82034254e-01 1.57844591e+00
-4.65084404e-01 -1.13617945e+00 7.85898324e-03 3.74880731e-01
2.75233928e-02 2.99895436e-01 1.78969428e-01 9.98696625e-01
-1.78408429e-01 8.33709002e-01 1.33489490e-01 -6.52229369e-01
5.06549716e-01 -6.85732067e-02 7.42607117e-01 -6.87297285e-01
-5.02998650e-01 -1.31532118e-01 7.38024637e-02 -1.13042974e+00
-2.54442692e-01 -5.83385646e-01 -1.63291228e+00 -3.63365024e-01
-1.14087559e-01 -7.80265704e-02 6.31482124e-01 6.80382013e-01
6.03427529e-01 1.08987296e+00 7.39092708e-01 -1.00021303e+00
-1.12326682e+00 -7.35194743e-01 -6.32128835e-01 1.26080006e-01
4.59934980e-01 -1.17976344e+00 -1.34347573e-01 -3.68621737e-01] | [4.277558326721191, 1.614282488822937] |
da33d513-8bdb-492f-a97d-a3f24dc93c15 | progressive-identification-of-true-labels-for | 2002.08053 | null | https://arxiv.org/abs/2002.08053v3 | https://arxiv.org/pdf/2002.08053v3.pdf | Progressive Identification of True Labels for Partial-Label Learning | Partial-label learning (PLL) is a typical weakly supervised learning problem, where each training instance is equipped with a set of candidate labels among which only one is the true label. Most existing methods elaborately designed learning objectives as constrained optimizations that must be solved in specific manners, making their computational complexity a bottleneck for scaling up to big data. The goal of this paper is to propose a novel framework of PLL with flexibility on the model and optimization algorithm. More specifically, we propose a novel estimator of the classification risk, theoretically analyze the classifier-consistency, and establish an estimation error bound. Then we propose a progressive identification algorithm for approximately minimizing the proposed risk estimator, where the update of the model and identification of true labels are conducted in a seamless manner. The resulting algorithm is model-independent and loss-independent, and compatible with stochastic optimization. Thorough experiments demonstrate it sets the new state of the art. | ['Xin Geng', 'Lei Feng', 'Jiaqi Lv', 'Masashi Sugiyama', 'Miao Xu', 'Gang Niu'] | 2020-02-19 | null | https://proceedings.icml.cc/static/paper_files/icml/2020/6080-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/6080-Paper.pdf | icml-2020-1 | ['partial-label-learning'] | ['methodology'] | [ 3.42510581e-01 2.58945853e-01 -6.14981055e-01 -6.04937255e-01
-1.09606147e+00 -4.79185432e-01 2.95217752e-01 2.24087507e-01
-5.16348183e-01 7.98770428e-01 -4.06025112e-01 -6.99578449e-02
-4.27228093e-01 -4.36011463e-01 -6.07775271e-01 -1.01880348e+00
2.95386136e-01 8.40663373e-01 -1.05121635e-01 4.11592484e-01
1.08274125e-01 2.85372198e-01 -1.68725121e+00 -1.93415239e-01
9.05572295e-01 1.29096174e+00 -5.34413978e-02 4.15091634e-01
-8.78553614e-02 7.69303560e-01 -2.41500527e-01 -6.01417899e-01
3.08514237e-01 -3.19391847e-01 -9.87235904e-01 4.09656256e-01
3.91916543e-01 7.01863244e-02 3.69679272e-01 1.18759751e+00
5.05805969e-01 5.89401573e-02 7.95148432e-01 -1.38778114e+00
-2.66549855e-01 6.38790607e-01 -4.58648264e-01 -2.64201939e-01
-8.52249488e-02 -2.57044822e-01 1.25082362e+00 -9.88388956e-01
3.33131969e-01 9.56205249e-01 8.71279538e-01 5.86671889e-01
-1.31532431e+00 -4.58683461e-01 3.74675065e-01 8.39193612e-02
-1.46686876e+00 -3.57607156e-01 7.31569290e-01 -4.66903508e-01
2.59372234e-01 3.74323785e-01 2.20010519e-01 7.60135531e-01
-2.81407595e-01 8.86011243e-01 1.17734611e+00 -7.14078069e-01
4.78545398e-01 7.70125926e-01 6.22817934e-01 9.68423963e-01
2.54130036e-01 -3.03872693e-02 -4.41996515e-01 -3.18866670e-01
1.34242654e-01 -1.18610874e-01 -4.04846780e-02 -8.35513234e-01
-7.19554543e-01 8.36156189e-01 -1.56197801e-01 9.22782570e-02
2.20031217e-02 2.12255330e-03 3.24175924e-01 3.22839379e-01
9.31664467e-01 5.39521910e-02 -6.30369663e-01 2.14410231e-01
-8.93066883e-01 1.14787824e-01 1.04229248e+00 8.07925522e-01
8.54110301e-01 -3.39804381e-01 -1.17388554e-01 1.10721171e+00
4.50055510e-01 3.54883403e-01 2.80133724e-01 -1.04637969e+00
2.52951086e-01 5.41327834e-01 1.61527932e-01 -6.44414186e-01
-5.41734874e-01 -8.79598677e-01 -6.27153218e-01 6.37798384e-02
6.20198905e-01 -1.77376252e-02 -3.41154784e-01 2.01171041e+00
6.76842093e-01 2.16022611e-01 -2.09209085e-01 5.63014627e-01
2.03418463e-01 3.75115156e-01 1.17736459e-01 -7.89935589e-01
1.08343935e+00 -1.10372174e+00 -7.35178649e-01 -4.59238105e-02
9.42218900e-01 -5.81122994e-01 1.16605687e+00 4.35626179e-01
-1.02116740e+00 -3.91832530e-01 -7.56492078e-01 8.31145346e-02
-1.34118795e-01 4.55306917e-01 5.95817268e-01 1.05105269e+00
-7.39230633e-01 6.07765198e-01 -5.50134957e-01 -1.54203966e-01
3.16055477e-01 4.59491819e-01 -1.65145397e-01 1.80894867e-01
-8.29622507e-01 8.29919100e-01 5.30113935e-01 8.13019946e-02
-7.92482615e-01 -7.91935146e-01 -7.59236634e-01 4.19973731e-02
6.80459201e-01 -5.57430148e-01 1.29839122e+00 -1.03362823e+00
-1.56342518e+00 1.23896468e+00 -2.68042713e-01 -3.45974803e-01
8.01334083e-01 -1.40791535e-01 -1.08766630e-01 -1.64840683e-01
1.45994142e-01 1.27584577e-01 9.86123145e-01 -1.49767697e+00
-9.26198483e-01 -3.35534424e-01 -1.13220572e-01 1.34378985e-01
-6.48698568e-01 8.04316849e-02 -3.17510039e-01 -5.46767175e-01
1.02937251e-01 -8.39027107e-01 -2.58484632e-01 -2.20012409e-03
-4.16858196e-01 -4.65002924e-01 4.95382100e-01 -3.43742281e-01
1.41422093e+00 -2.05833793e+00 8.69546435e-04 3.25399339e-01
2.05854163e-01 1.47672277e-02 1.71920612e-01 1.87915608e-01
-2.50952989e-02 1.21735744e-01 -6.24916553e-01 -9.78569508e-01
2.67936319e-01 1.13642499e-01 -2.72474438e-01 8.13281953e-01
1.26246378e-01 5.86006284e-01 -9.00209129e-01 -7.88799167e-01
1.11281388e-01 1.04809955e-01 -5.07552743e-01 4.78632867e-01
-2.47352600e-01 3.75381649e-01 -5.34725249e-01 7.46494472e-01
7.02187777e-01 -4.16224778e-01 2.97528267e-01 -1.38528123e-01
7.17383763e-03 -4.17228490e-02 -1.50527024e+00 1.27263117e+00
-6.31204128e-01 -1.29815161e-01 1.67137846e-01 -1.46720982e+00
7.69320130e-01 1.17112286e-01 6.96236014e-01 2.99069192e-02
2.13803440e-01 4.53549951e-01 -6.18350863e-01 -4.76813525e-01
4.09107655e-03 -5.46360731e-01 -9.53033417e-02 6.69007421e-01
2.55895764e-01 1.72391355e-01 1.42002404e-01 -1.84127092e-01
6.48460090e-01 1.59285769e-01 5.06726980e-01 -4.56539631e-01
8.04863989e-01 -3.29683572e-01 7.98231721e-01 8.84125352e-01
-2.27132842e-01 3.96956325e-01 4.22503352e-01 -3.10545802e-01
-8.15811992e-01 -1.02684808e+00 -6.15163088e-01 1.23802149e+00
7.35041648e-02 -1.86597228e-01 -9.18101847e-01 -1.17780209e+00
8.33094120e-02 6.61895394e-01 -5.66584706e-01 -1.74007386e-01
-3.40315521e-01 -9.87803280e-01 2.00876474e-01 1.21715732e-01
2.42144182e-01 -6.16793692e-01 -1.90569729e-01 1.11706957e-01
-6.58035800e-02 -8.46122146e-01 -5.50902843e-01 5.33318162e-01
-8.67867708e-01 -1.18406367e+00 -4.39940661e-01 -9.98156965e-01
8.28526378e-01 -1.89986974e-01 1.09792519e+00 3.54079939e-02
-6.51712641e-02 3.24044108e-01 -4.79306541e-02 -2.99283385e-01
-4.52303082e-01 2.31450602e-01 3.49843502e-01 5.96003592e-01
2.01932088e-01 -4.28741455e-01 -1.40129924e-01 4.36023980e-01
-7.46745050e-01 -3.07139903e-01 3.97065848e-01 1.15893114e+00
1.01823187e+00 2.28930026e-01 7.25250840e-01 -1.38843465e+00
4.22805816e-01 -4.73843515e-01 -8.92889738e-01 6.83020532e-01
-1.16715360e+00 3.34028393e-01 8.03029776e-01 -3.70197892e-01
-1.14932370e+00 4.32777852e-01 -6.35379180e-02 -1.58455774e-01
-3.16108391e-02 3.04483682e-01 -5.33292353e-01 -2.94187486e-01
3.95553917e-01 6.12791367e-02 -1.11968808e-01 -8.75031352e-01
3.43703896e-01 6.41238153e-01 2.02709883e-01 -8.54385555e-01
7.51249194e-01 4.77833360e-01 2.29768470e-01 -4.07508880e-01
-1.52089739e+00 -6.05323493e-01 -7.62180269e-01 -3.08425307e-01
3.91746402e-01 -6.10111773e-01 -8.49260151e-01 4.78228331e-01
-8.03136349e-01 -2.70598918e-01 -5.54383159e-01 4.26092982e-01
-6.83513939e-01 4.99760687e-01 -5.07781625e-01 -1.09062791e+00
-1.45147026e-01 -1.12889910e+00 9.97363806e-01 -1.15013227e-01
5.36547787e-02 -1.26860058e+00 2.32528478e-01 5.04909992e-01
-1.66174788e-02 2.46058349e-02 8.46440732e-01 -8.33375871e-01
-2.72597611e-01 -3.77942473e-01 -1.91615701e-01 7.59820640e-01
2.67815194e-03 -2.16155753e-01 -1.12395966e+00 -5.00155330e-01
5.06313086e-01 -5.37076831e-01 8.80201519e-01 4.66002822e-01
1.48392975e+00 -3.82678419e-01 -3.13765824e-01 7.88775027e-01
1.53427029e+00 -1.83549628e-01 -1.44091040e-01 2.35156402e-01
6.60511792e-01 8.21752667e-01 7.78852463e-01 5.75183749e-01
3.63168776e-01 7.22806692e-01 2.76220739e-01 5.23238368e-02
2.65058845e-01 -2.35172525e-01 2.46528938e-01 1.07631719e+00
2.31271997e-01 -8.57119337e-02 -6.56952262e-01 2.49996856e-01
-2.05896831e+00 -7.02342629e-01 -2.53798794e-02 2.44958019e+00
9.83589530e-01 -3.97755485e-03 6.96202442e-02 3.76829803e-01
8.31116974e-01 1.68956537e-02 -7.12050200e-01 -1.38849765e-01
-6.68910518e-02 1.09840654e-01 6.29986942e-01 7.98577785e-01
-1.38220096e+00 6.76247418e-01 6.83443308e+00 1.12042940e+00
-7.69381821e-01 4.29929316e-01 8.92150640e-01 -9.73525196e-02
-2.44858950e-01 5.41535579e-02 -1.09688318e+00 5.02705216e-01
8.41213882e-01 -1.41657725e-01 3.14198643e-01 1.08062160e+00
6.81114718e-02 -4.40792441e-02 -1.34346402e+00 9.39095318e-01
4.02346998e-02 -8.36896300e-01 -1.89906403e-01 3.44905555e-02
8.13997209e-01 -4.88700151e-01 2.36149937e-01 2.34794974e-01
2.61482626e-01 -8.01719785e-01 9.23396051e-01 5.85682869e-01
7.57337034e-01 -9.04114187e-01 7.32869148e-01 7.21734047e-01
-9.23531592e-01 -3.59256685e-01 -3.44053090e-01 -1.86263360e-02
1.92265600e-01 1.11064959e+00 -4.08364773e-01 4.39211905e-01
3.83673579e-01 6.71745598e-01 -4.06779766e-01 8.12712967e-01
-2.50347167e-01 8.99368882e-01 -4.48838860e-01 2.58076563e-02
-6.18677922e-02 -2.66379565e-01 3.35694045e-01 1.17131674e+00
1.47678897e-01 -1.87621951e-01 5.00389040e-01 6.14869595e-01
-3.81144375e-01 4.86258209e-01 -2.90509224e-01 4.11892295e-01
5.00799835e-01 1.28434181e+00 -6.57346010e-01 -1.78665996e-01
-3.52387190e-01 7.26348698e-01 7.91873693e-01 1.15838476e-01
-8.30930114e-01 5.45441583e-02 3.66225779e-01 -9.67331976e-02
-5.72584607e-02 3.48974347e-01 -5.12525082e-01 -1.29106069e+00
3.59958470e-01 -6.93143606e-01 6.82814956e-01 -5.21430857e-02
-1.56519055e+00 2.04809725e-01 -5.03783524e-02 -1.18307292e+00
-5.75474948e-02 -5.56159794e-01 -2.74197757e-01 6.48289382e-01
-1.71195674e+00 -9.48911726e-01 2.82756668e-02 4.95804995e-01
3.22164208e-01 -8.75930712e-02 7.86453843e-01 4.82903481e-01
-8.69007528e-01 1.02060080e+00 5.53925216e-01 -3.74655455e-01
7.00630069e-01 -1.43818545e+00 -4.26101863e-01 7.34320462e-01
1.33390322e-01 3.18810731e-01 6.35476172e-01 -2.10710883e-01
-9.23694909e-01 -1.23870683e+00 1.14008892e+00 -4.88191605e-01
6.79122150e-01 -4.71149504e-01 -7.80347407e-01 4.42468077e-01
-4.62969184e-01 1.71632424e-01 9.18005347e-01 2.64396161e-01
-3.24300498e-01 -4.93446380e-01 -1.34848011e+00 1.77217990e-01
1.01610398e+00 -5.27802706e-01 -2.07520694e-01 7.12590277e-01
6.60779953e-01 -1.41135991e-01 -8.44823480e-01 4.77999419e-01
4.96533036e-01 -9.12231445e-01 8.84501398e-01 -5.64033151e-01
1.06885344e-01 -2.03067586e-01 -4.67625335e-02 -9.62726295e-01
-1.71295941e-01 -5.24124563e-01 -3.46902013e-01 1.45682430e+00
4.45010990e-01 -7.29117334e-01 9.98512983e-01 7.14516640e-01
4.41025421e-02 -9.97232258e-01 -1.10183716e+00 -9.97681916e-01
-5.89283593e-02 -6.58270657e-01 3.74299020e-01 1.08630633e+00
6.27033459e-03 2.81195372e-01 -6.32808924e-01 2.32525662e-01
1.22947097e+00 6.75863549e-02 2.58318096e-01 -1.43121266e+00
-5.27773559e-01 -4.28971380e-01 -3.52627449e-02 -9.89433110e-01
5.56242585e-01 -1.16982496e+00 2.31222451e-01 -9.56138730e-01
4.29507822e-01 -8.62735033e-01 -6.32701576e-01 3.80499512e-01
-3.11378181e-01 -2.51588039e-02 -6.87254667e-02 3.14110756e-01
-8.03231418e-01 4.17940438e-01 7.40023613e-01 -3.87879834e-02
-4.90492433e-02 6.31522596e-01 -5.94967842e-01 8.52616131e-01
7.76924312e-01 -7.92242825e-01 -5.08219957e-01 -1.71286792e-01
2.51378506e-01 -1.81142509e-01 1.53905436e-01 -7.15657532e-01
2.05000371e-01 -9.37700123e-02 -1.14746451e-01 -3.06142539e-01
2.20665652e-02 -9.94041085e-01 4.42660302e-02 4.06869620e-01
-9.60943341e-01 -4.53168780e-01 -4.02088612e-01 7.65364170e-01
-1.58084810e-01 -7.90044069e-01 1.11338449e+00 1.26840457e-01
-3.72655481e-01 5.26023269e-01 5.46096005e-02 2.62415707e-01
1.16385424e+00 1.85778126e-01 2.46598631e-01 -8.66512880e-02
-8.22259724e-01 2.96000630e-01 3.18719596e-01 -9.19079259e-02
1.86519116e-01 -1.28826416e+00 -6.93241060e-01 -5.60846478e-02
1.67837128e-01 -7.77929090e-03 2.95673963e-02 7.94774771e-01
-4.13783863e-02 3.65209609e-01 4.86214101e-01 -5.08360147e-01
-1.19959676e+00 8.12640071e-01 4.22723085e-01 -7.53562152e-01
-2.21292242e-01 1.06049371e+00 2.31949389e-01 -7.88027227e-01
6.54943764e-01 -4.37595919e-02 -1.76964179e-01 1.94078371e-01
3.52796793e-01 4.87940878e-01 7.47806951e-02 -4.94353175e-01
-2.12413460e-01 7.35344887e-01 6.29382655e-02 1.27472386e-01
1.14462590e+00 -4.04641986e-01 -2.80127883e-01 6.93400085e-01
1.39833295e+00 -1.11174487e-01 -1.32467747e+00 -6.44550622e-01
5.11999846e-01 -3.90200049e-01 6.10508323e-02 -6.90361142e-01
-1.04740894e+00 6.50628686e-01 6.09604239e-01 2.76140898e-01
1.17640233e+00 9.96285975e-02 5.10661542e-01 5.02688944e-01
5.74517131e-01 -1.41218448e+00 -1.53266296e-01 3.29658806e-01
3.22061449e-01 -1.44029462e+00 -2.76208054e-02 -6.24383867e-01
-4.29617614e-01 9.98808563e-01 3.87616277e-01 9.43002552e-02
9.29813385e-01 1.21348664e-01 -1.86372608e-01 8.19183812e-02
-7.63404250e-01 -1.77728996e-01 2.66798288e-01 4.47844625e-01
3.26979756e-01 3.56763192e-02 -6.47349894e-01 8.93682420e-01
1.59126028e-01 -1.21726677e-01 4.69170660e-02 6.24131978e-01
-4.97873098e-01 -1.29328978e+00 -2.40039289e-01 4.38385755e-01
-5.72419822e-01 7.79298767e-02 -1.03274509e-01 3.75173509e-01
3.79503846e-01 9.15832937e-01 -3.82301778e-01 -1.35819778e-01
2.40280569e-01 3.58246267e-01 3.06213617e-01 -7.01274872e-01
-3.88061345e-01 -1.05056338e-01 5.81489615e-02 -4.01385128e-01
-5.29118538e-01 -7.53203571e-01 -9.24983859e-01 -7.92314932e-02
-7.20382452e-01 3.36623698e-01 6.35062993e-01 1.12548089e+00
-5.83657213e-02 1.68504000e-01 1.19497681e+00 -4.84583378e-01
-1.19262159e+00 -5.99880993e-01 -9.69977736e-01 3.57199579e-01
2.23350301e-01 -7.45071471e-01 -6.94687188e-01 1.55692637e-01] | [9.20728588104248, 4.119364261627197] |
556b8d09-954c-4866-99ca-0c1317c9776a | red-attack-resource-efficient-decision-based | 1901.10258 | null | http://arxiv.org/abs/1901.10258v2 | http://arxiv.org/pdf/1901.10258v2.pdf | RED-Attack: Resource Efficient Decision based Attack for Machine Learning | Due to data dependency and model leakage properties, Deep Neural Networks
(DNNs) exhibit several security vulnerabilities. Several security attacks
exploited them but most of them require the output probability vector. These
attacks can be mitigated by concealing the output probability vector. To
address this limitation, decision-based attacks have been proposed which can
estimate the model but they require several thousand queries to generate a
single untargeted attack image. However, in real-time attacks, resources and
attack time are very crucial parameters. Therefore, in resource-constrained
systems, e.g., autonomous vehicles where an untargeted attack can have a
catastrophic effect, these attacks may not work efficiently. To address this
limitation, we propose a resource efficient decision-based methodology which
generates the imperceptible attack, i.e., the RED-Attack, for a given black-box
model. The proposed methodology follows two main steps to generate the
imperceptible attack, i.e., classification boundary estimation and adversarial
noise optimization. Firstly, we propose a half-interval search-based algorithm
for estimating a sample on the classification boundary using a target image and
a randomly selected image from another class. Secondly, we propose an
optimization algorithm which first, introduces a small perturbation in some
randomly selected pixels of the estimated sample. Then to ensure
imperceptibility, it optimizes the distance between the perturbed and target
samples. For illustration, we evaluate it for CFAR-10 and German Traffic Sign
Recognition (GTSR) using state-of-the-art networks. | ['Muhammad Shafique', 'Faiq Khalid', 'Muhammad Abdullah Hanif', 'Hassan Ali', 'Semeen Rehman', 'Rehan Ahmed'] | 2019-01-29 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 5.06622732e-01 -2.27488026e-01 2.39326254e-01 -5.85092269e-02
-6.23419821e-01 -8.01818490e-01 4.73768026e-01 -2.11932957e-01
-6.94178522e-01 6.93232715e-01 -6.31290019e-01 -6.83943391e-01
8.94950926e-02 -9.54462826e-01 -9.64139163e-01 -9.51650679e-01
3.51124443e-02 -1.75214082e-01 5.08795261e-01 9.84724909e-02
3.40186149e-01 7.58349657e-01 -1.27300930e+00 1.57452703e-01
6.87078655e-01 1.15908253e+00 -7.12279454e-02 7.58017361e-01
8.71996433e-02 5.32296777e-01 -1.12230527e+00 -3.61144364e-01
4.89040047e-01 -4.20361161e-01 -2.56897688e-01 -4.21934754e-01
3.77760112e-01 -4.94527191e-01 -4.62827116e-01 1.68719149e+00
5.12752116e-01 -1.68944702e-01 6.25290155e-01 -1.54812825e+00
-1.59999460e-01 4.84353662e-01 -3.49185586e-01 2.08811224e-01
-1.89362839e-01 3.90021712e-01 3.56053561e-01 -3.57367575e-01
3.49074244e-01 1.14101100e+00 3.81564021e-01 9.01401579e-01
-1.14382160e+00 -1.35960329e+00 -1.06314726e-01 3.53154659e-01
-1.42697918e+00 -4.51685697e-01 9.52089369e-01 -1.64118782e-01
3.63616854e-01 4.28218991e-01 3.14544708e-01 1.31480634e+00
3.43909085e-01 3.98745775e-01 1.36364663e+00 -2.96559125e-01
4.66259003e-01 3.46437156e-01 -1.19339801e-01 4.77758855e-01
4.55639333e-01 5.81264734e-01 1.12911975e-02 -3.75371814e-01
5.15552521e-01 -2.06319079e-01 -4.27888155e-01 -5.27516156e-02
-6.81226134e-01 6.77503288e-01 4.06590372e-01 2.41564065e-01
-1.47115424e-01 4.07585919e-01 3.26171607e-01 4.45859253e-01
-1.55908778e-01 -1.90565493e-02 -3.29313278e-01 1.01643823e-01
-7.44873881e-01 5.56046441e-02 7.94743717e-01 4.89773542e-01
6.17403626e-01 5.40177047e-01 1.40722528e-01 3.43648225e-01
4.25596267e-01 9.19362485e-01 2.48630852e-01 -2.15969026e-01
6.07104242e-01 1.60978839e-01 -3.04205134e-03 -1.29826057e+00
-2.16212615e-01 -4.05751586e-01 -1.05431139e+00 8.89695466e-01
5.14844894e-01 -5.57881773e-01 -1.22287631e+00 1.87375045e+00
3.38435888e-01 5.18766046e-01 5.34752131e-01 7.15235054e-01
4.12209898e-01 7.14709222e-01 4.82812151e-02 -1.30541414e-01
1.27687991e+00 -4.84367102e-01 -5.23794949e-01 -1.64620265e-01
3.95478427e-01 -6.29055262e-01 7.06840038e-01 4.45972323e-01
-5.55084348e-01 -4.53934222e-01 -1.53556097e+00 7.18144476e-01
-5.78034580e-01 1.47707060e-01 -2.09469095e-01 1.40327144e+00
-6.01498425e-01 3.52489144e-01 -8.41534257e-01 1.46887273e-01
5.04055798e-01 4.28733051e-01 -3.68446290e-01 1.93054572e-01
-1.42404950e+00 7.68436193e-01 5.56175113e-01 2.07095012e-01
-1.14260924e+00 -4.98704523e-01 -5.06279349e-01 -3.04850545e-02
2.05930144e-01 2.94174235e-02 5.43039739e-01 -1.08889282e+00
-1.45742214e+00 4.31200147e-01 3.37862611e-01 -9.21770930e-01
7.47070253e-01 9.88878682e-02 -7.14792788e-01 1.31761044e-01
-5.63590050e-01 3.97643626e-01 1.32401013e+00 -1.36069763e+00
-4.18362528e-01 -2.51379222e-01 -6.27832636e-02 -2.45336637e-01
-4.47421283e-01 1.70108259e-01 -7.11323752e-04 -8.20391595e-01
-3.28006856e-02 -9.76466715e-01 -1.88317776e-01 3.60845737e-02
-7.24879265e-01 4.54692006e-01 1.36617482e+00 -6.13625169e-01
1.11807859e+00 -2.45894051e+00 -3.86449128e-01 6.59243464e-01
-1.00847900e-01 8.76800060e-01 -8.54726657e-02 8.77394229e-02
-6.92574829e-02 3.82627219e-01 -4.04717147e-01 6.99056759e-02
-4.02119681e-02 4.08845469e-02 -6.08518124e-01 7.30366051e-01
1.39241159e-01 4.25950974e-01 -3.58923197e-01 -1.73356488e-01
1.77036136e-01 5.74914157e-01 -4.08417940e-01 9.52394456e-02
-8.22942331e-02 2.78354198e-01 -5.01467824e-01 3.94296110e-01
1.18207836e+00 4.41441506e-01 -8.60651955e-03 -4.28974032e-01
6.09963499e-02 -3.82813424e-01 -1.44184005e+00 6.75063252e-01
-3.86064261e-01 7.87569106e-01 1.94641147e-02 -1.06985450e+00
1.24783182e+00 2.88731545e-01 -8.34935457e-02 -4.20538545e-01
5.50358534e-01 4.37803209e-01 3.09589237e-01 -1.49752796e-01
-6.19037263e-03 1.08075425e-01 4.78872620e-02 3.12066048e-01
-3.85120779e-01 2.05914617e-01 -4.97247994e-01 -1.90311939e-01
1.23574555e+00 -2.44235978e-01 7.71043524e-02 1.87226143e-02
8.31827223e-01 -3.03150952e-01 5.50254941e-01 9.10316646e-01
-3.97009283e-01 3.41205955e-01 6.42984152e-01 -4.14034069e-01
-9.43631947e-01 -9.61627662e-01 -8.49738717e-02 2.17754126e-01
2.52118468e-01 1.51625022e-01 -1.03903913e+00 -9.37801421e-01
-1.28363371e-01 6.61579013e-01 -4.54410583e-01 -5.59578240e-01
-6.69417202e-01 -6.59948826e-01 1.36870885e+00 2.80490547e-01
1.23623037e+00 -9.80255544e-01 -9.24902141e-01 4.57022451e-02
2.82611728e-01 -1.25219548e+00 -8.70659202e-02 1.48632601e-01
-5.42456567e-01 -1.01964128e+00 -5.06552637e-01 -6.49679184e-01
8.12796116e-01 -2.00911075e-01 3.32828641e-01 2.08458588e-01
-2.16718465e-01 -1.81968719e-01 -2.05200821e-01 -4.59612101e-01
-8.47772777e-01 -3.64452630e-01 1.33221686e-01 3.87341499e-01
2.46976122e-01 -4.80780929e-01 -4.97005731e-01 4.90292341e-01
-1.14567256e+00 -4.66841519e-01 6.69896662e-01 8.41533482e-01
3.18629652e-01 4.95950937e-01 5.26259065e-01 -5.96075177e-01
4.45497185e-01 -2.20104828e-01 -1.07870793e+00 3.72324854e-01
-3.12891185e-01 1.55412421e-01 1.14939356e+00 -9.67179060e-01
-6.54694080e-01 1.64124459e-01 -3.33141595e-01 -6.42958283e-01
-3.11036021e-01 9.37378779e-02 -5.94728708e-01 -6.82663202e-01
8.01705062e-01 4.34022576e-01 3.22600752e-02 -9.63928550e-02
1.20202444e-01 8.77056599e-01 5.11824727e-01 -2.45142922e-01
1.38546348e+00 3.83250117e-01 2.70006657e-01 -8.98852587e-01
1.37345986e-02 2.56726950e-01 5.97305931e-02 -2.72202045e-01
7.12914884e-01 -4.93913561e-01 -9.74194050e-01 1.03216183e+00
-1.23175442e+00 -2.19693005e-01 2.40522236e-01 3.68312895e-01
-2.07468972e-01 4.39407289e-01 -1.70199305e-01 -9.66717482e-01
-3.58929545e-01 -1.39309192e+00 4.39432830e-01 4.47227508e-01
3.33811134e-01 -5.52128077e-01 -4.28059399e-01 -8.26292560e-02
5.40453017e-01 6.69797063e-01 8.46595347e-01 -1.00209022e+00
-7.46369362e-01 -5.40664256e-01 -2.91371375e-01 7.66569197e-01
1.15581647e-01 9.62277651e-02 -9.64650154e-01 -4.10236508e-01
2.64090568e-01 -2.33468041e-01 6.20605052e-01 1.51103869e-01
1.13997054e+00 -7.07673132e-01 -2.72303700e-01 7.99834013e-01
1.52027678e+00 8.36368918e-01 8.90578806e-01 5.17692506e-01
5.16733527e-01 1.64917588e-01 5.11709571e-01 2.98413724e-01
-3.03326249e-01 5.64231873e-01 6.57967269e-01 3.32819559e-02
1.41186625e-01 -4.29380648e-02 4.19952750e-01 7.74939582e-02
3.89352828e-01 -4.92688805e-01 -7.44689822e-01 5.42858765e-02
-1.29232204e+00 -9.01309133e-01 1.70684993e-01 2.39301753e+00
7.81746387e-01 6.43517137e-01 -2.77042538e-01 5.34002781e-01
8.89315903e-01 3.03803474e-01 -7.87750065e-01 -3.58659357e-01
-1.61929041e-01 1.80260673e-01 8.88625741e-01 4.98184055e-01
-1.10757279e+00 9.30095613e-01 4.79595470e+00 1.22256505e+00
-1.50522554e+00 -7.05539361e-02 7.34256744e-01 2.95270622e-01
-1.35621667e-01 -5.84761351e-02 -8.55084360e-01 7.10810363e-01
9.50538516e-01 1.41793251e-01 2.76400656e-01 7.02757061e-01
8.90740529e-02 4.01861742e-02 -6.15358293e-01 9.18555200e-01
-3.69158015e-02 -1.12952507e+00 3.06018032e-02 1.06933348e-01
3.53928685e-01 -3.81711334e-01 3.11285287e-01 7.85342827e-02
1.14492401e-01 -9.53218043e-01 6.37417853e-01 2.99508065e-01
8.54231834e-01 -1.02911055e+00 7.47413218e-01 3.61038715e-01
-1.03663850e+00 -1.98888302e-01 -4.63286161e-01 5.20883918e-01
-1.61647484e-01 2.56120712e-01 -4.36454505e-01 2.20566884e-01
5.31879127e-01 -1.79708064e-01 -3.47052395e-01 9.15100694e-01
-3.60244751e-01 8.81711900e-01 -5.69279790e-01 -2.46811166e-01
3.20300758e-01 -9.84562635e-02 8.04173470e-01 9.00301516e-01
3.22443366e-01 -9.55231935e-02 -3.11314672e-01 8.88605595e-01
5.66141456e-02 -1.39488071e-01 -7.03073621e-01 1.17950216e-01
8.70711863e-01 9.62868869e-01 -7.68409073e-01 -1.30330428e-01
3.66804004e-02 8.94465446e-01 -7.21741840e-02 5.60731828e-01
-1.23757231e+00 -9.16375339e-01 6.04359984e-01 -3.65905821e-01
4.19635653e-01 -2.12852508e-01 -2.48582706e-01 -7.31046200e-01
1.11517698e-01 -1.40026772e+00 1.12091988e-01 -2.93274075e-01
-7.17238843e-01 8.46303523e-01 -2.11944699e-01 -1.33001173e+00
2.66606989e-03 -5.37597716e-01 -6.43747807e-01 8.41461897e-01
-1.26715636e+00 -8.65798295e-01 -1.94573686e-01 7.11154759e-01
2.26704866e-01 -5.08778691e-01 6.09046459e-01 2.70224512e-01
-7.16867685e-01 1.24028170e+00 1.91933170e-01 4.50804085e-01
4.15655345e-01 -6.93126500e-01 4.98993903e-01 1.25375021e+00
1.77762210e-01 3.59573752e-01 8.30942273e-01 -6.33519173e-01
-1.18296003e+00 -1.05718708e+00 2.17154592e-01 1.95178851e-01
5.95722556e-01 -4.77210969e-01 -8.20537865e-01 3.85053724e-01
-1.47850975e-01 2.59201407e-01 2.02117398e-01 -9.05567348e-01
-4.90289867e-01 -4.21866715e-01 -1.68522549e+00 9.08553123e-01
4.59166884e-01 -4.67738330e-01 -1.73812822e-01 -5.18924706e-02
5.39255857e-01 -4.81976122e-01 -4.16864395e-01 4.95461375e-01
6.39448285e-01 -8.99037600e-01 1.01340008e+00 -2.80204952e-01
-8.46318752e-02 -4.71233577e-01 -3.05713981e-01 -1.10136151e+00
3.22771281e-01 -8.53275061e-01 -4.85118106e-02 1.19115901e+00
3.69584918e-01 -1.07783318e+00 9.76282001e-01 4.08329159e-01
3.04160357e-01 -5.13613224e-01 -1.28900182e+00 -1.00883079e+00
-3.70465145e-02 -4.55477417e-01 6.94890261e-01 5.67499757e-01
-6.03167772e-01 -4.22737390e-01 -4.20552284e-01 9.04410064e-01
8.71022165e-01 -5.86347401e-01 7.47252166e-01 -7.33105958e-01
-3.74849066e-02 -2.23963737e-01 -9.22190249e-01 -7.39958644e-01
7.26729482e-02 -2.03268915e-01 1.00081585e-01 -6.32642508e-01
-4.82330024e-01 -6.05554581e-01 -3.10936838e-01 3.38142455e-01
-2.59987861e-02 4.05074626e-01 1.80779174e-01 7.18191918e-03
1.87960520e-01 2.71431446e-01 7.22956538e-01 -2.51504362e-01
-1.33647844e-01 3.23838115e-01 -1.30086809e-01 6.61612868e-01
1.10409522e+00 -8.33584785e-01 -4.66778517e-01 6.87225955e-03
-1.93261743e-01 -2.43482031e-02 7.05024898e-01 -1.50985682e+00
2.91804731e-01 -8.74296129e-02 2.10794121e-01 -5.23069024e-01
2.72305936e-01 -1.27106142e+00 2.30628029e-01 1.06105506e+00
-1.55740872e-01 -5.16774245e-02 3.95167619e-01 7.29638040e-01
-2.00355411e-01 -5.47711372e-01 1.22485781e+00 3.32962245e-01
-5.07971883e-01 1.27869591e-01 -5.05515695e-01 -1.07344881e-01
1.28772402e+00 -4.16216105e-01 -6.45861864e-01 -2.09196389e-01
-3.39286119e-01 -2.39808708e-01 1.83243319e-01 1.56043410e-01
8.50488842e-01 -1.24636590e+00 -5.44221461e-01 5.19975662e-01
-2.96697855e-01 -2.85668194e-01 2.57057607e-01 4.27629083e-01
-7.99638271e-01 7.82985166e-02 -2.84015954e-01 -4.21218783e-01
-1.61783385e+00 5.54276884e-01 6.16259575e-01 -4.42480147e-02
-3.56699288e-01 6.84370041e-01 -8.02440941e-02 -7.03059649e-03
4.64267582e-01 -2.17471182e-01 -2.75206655e-01 -2.91432351e-01
5.71770787e-01 1.68898270e-01 -1.32342100e-01 -4.52339113e-01
-4.40337390e-01 6.43835783e-01 -4.81230160e-03 -2.65759081e-01
7.46439934e-01 1.46459267e-01 1.32133931e-01 -1.91357002e-01
1.39992738e+00 8.76677558e-02 -1.30292642e+00 -1.01108856e-01
-7.81293213e-02 -4.19512808e-01 4.25319932e-02 -6.40570581e-01
-1.27229619e+00 7.88196325e-01 1.07000434e+00 2.98507750e-01
1.36535418e+00 -6.94371283e-01 8.12023401e-01 5.54794669e-01
2.15485290e-01 -7.87899077e-01 -1.02578953e-01 2.48268887e-01
5.63578486e-01 -1.00246906e+00 -3.08941871e-01 -1.98911279e-01
-3.77111882e-01 1.23349142e+00 7.12068796e-01 -1.67102009e-01
9.22611594e-01 5.89588344e-01 3.13387960e-01 2.11444914e-01
-3.98499817e-01 2.23160654e-01 7.56221786e-02 7.10901797e-01
-5.75133979e-01 -9.24244076e-02 -2.07854390e-01 3.42482626e-01
-1.61439180e-01 -2.25564852e-01 6.05881870e-01 8.14070046e-01
-4.13570285e-01 -8.98488283e-01 -7.33448446e-01 8.56646001e-02
-6.10410631e-01 -9.05699506e-02 -1.35655269e-01 7.25932777e-01
9.02868584e-02 1.06614566e+00 -1.72227576e-01 -8.97956908e-01
2.19988331e-01 -1.48659691e-01 4.52780351e-02 1.24728635e-01
-4.93390471e-01 -3.29701394e-01 -5.74683771e-02 -3.36604297e-01
1.52933327e-02 -2.54160434e-01 -1.08537889e+00 -3.19811583e-01
-4.23330694e-01 1.58644766e-02 9.81155753e-01 9.37121451e-01
7.15930089e-02 3.53161275e-01 1.13201141e+00 -4.44894820e-01
-1.13335252e+00 -7.05773711e-01 -3.25406551e-01 9.79480669e-02
4.71094728e-01 -4.75879759e-01 -8.91755223e-01 -2.61990875e-01] | [5.408843994140625, 7.871450424194336] |
47fdbae6-81f4-4233-b6c8-ed26da6a5c20 | sim-to-real-reinforcement-learning-for | 1806.07851 | null | http://arxiv.org/abs/1806.07851v2 | http://arxiv.org/pdf/1806.07851v2.pdf | Sim-to-Real Reinforcement Learning for Deformable Object Manipulation | We have seen much recent progress in rigid object manipulation, but
interaction with deformable objects has notably lagged behind. Due to the large
configuration space of deformable objects, solutions using traditional
modelling approaches require significant engineering work. Perhaps then,
bypassing the need for explicit modelling and instead learning the control in
an end-to-end manner serves as a better approach? Despite the growing interest
in the use of end-to-end robot learning approaches, only a small amount of work
has focused on their applicability to deformable object manipulation. Moreover,
due to the large amount of data needed to learn these end-to-end solutions, an
emerging trend is to learn control policies in simulation and then transfer
them over to the real world. To-date, no work has explored whether it is
possible to learn and transfer deformable object policies. We believe that if
sim-to-real methods are to be employed further, then it should be possible to
learn to interact with a wide variety of objects, and not only rigid objects.
In this work, we use a combination of state-of-the-art deep reinforcement
learning algorithms to solve the problem of manipulating deformable objects
(specifically cloth). We evaluate our approach on three tasks --- folding a
towel up to a mark, folding a face towel diagonally, and draping a piece of
cloth over a hanger. Our agents are fully trained in simulation with domain
randomisation, and then successfully deployed in the real world without having
seen any real deformable objects. | ['Stephen James', 'Jan Matas', 'Andrew J. Davison'] | 2018-06-20 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [ 9.41209272e-02 2.37818345e-01 1.94569588e-01 -7.99014047e-02
-4.13512528e-01 -9.53771293e-01 3.23980570e-01 -2.83178955e-01
-3.64055455e-01 8.36511731e-01 -3.08938831e-01 -1.55144036e-01
-2.95980573e-01 -7.06992865e-01 -1.10523820e+00 -7.82543361e-01
-3.92729998e-01 1.06992292e+00 4.27155733e-01 -6.47532761e-01
3.42580490e-03 8.51982355e-01 -1.48913944e+00 1.66322384e-02
5.58336496e-01 6.49041355e-01 3.80326062e-01 9.24248934e-01
4.27897245e-01 5.02298772e-01 -1.92502335e-01 -1.21067099e-01
5.88413298e-01 -1.48265705e-01 -1.01733744e+00 3.12964678e-01
3.59590411e-01 -5.04567742e-01 -1.70957550e-01 6.84119582e-01
5.40853739e-01 4.06833887e-01 4.72507864e-01 -9.67076957e-01
-6.17729425e-01 2.55986065e-01 -3.53984743e-01 -3.98936838e-01
4.39617753e-01 6.32963300e-01 5.66432118e-01 -4.18159992e-01
9.55750108e-01 1.39434874e+00 5.75985312e-01 1.01377988e+00
-1.20124388e+00 -2.35142156e-01 2.86414921e-01 -5.05753398e-01
-6.39880002e-01 -5.80912232e-02 5.89348376e-01 -4.84390318e-01
9.21442449e-01 2.01434091e-01 9.93702888e-01 1.14611673e+00
3.34267974e-01 6.84820712e-01 1.17448223e+00 -3.50349039e-01
2.90851921e-01 -2.13843778e-01 -4.97037023e-01 9.37112331e-01
9.22899693e-02 4.13248777e-01 1.16004996e-01 -1.84026569e-01
1.14855385e+00 -9.77191376e-04 -1.82463169e-01 -8.63888443e-01
-1.35554850e+00 5.91849327e-01 4.98292357e-01 -3.55785936e-02
-3.33643943e-01 6.27535105e-01 3.61618072e-01 3.43414962e-01
2.75288314e-01 6.44352019e-01 -7.71490991e-01 -2.41956368e-01
-4.88519460e-01 1.12754273e+00 1.24463558e+00 8.11544359e-01
3.83093029e-01 1.48196191e-01 2.90207267e-01 4.90898252e-01
3.35397214e-01 4.42541808e-01 -4.42358293e-02 -1.30356848e+00
3.77494037e-01 2.65772223e-01 5.75722516e-01 -6.92895591e-01
-3.92147899e-01 2.60403097e-01 -4.44059700e-01 1.23030639e+00
6.52886569e-01 -6.25523269e-01 -1.09968054e+00 1.59583879e+00
6.43735051e-01 -1.25726759e-01 -5.80506064e-02 1.43979990e+00
8.04558918e-02 5.45599639e-01 7.46139511e-02 2.09841996e-01
1.02490997e+00 -7.40512788e-01 -1.88475624e-01 -5.42635396e-02
2.04281032e-01 -7.25137532e-01 9.57153738e-01 5.76972067e-01
-1.49574304e+00 -5.32339156e-01 -8.27758670e-01 1.30617559e-01
-2.97748625e-01 -3.67491990e-01 7.30100989e-01 9.33214426e-02
-8.60171199e-01 1.22094464e+00 -1.43030906e+00 -5.58032095e-01
2.22766325e-01 7.57092118e-01 -4.38945234e-01 1.38137415e-01
-8.97272408e-01 1.26366973e+00 2.73027301e-01 3.31434727e-01
-1.26193285e+00 -5.06906450e-01 -5.76166093e-01 -3.78853202e-01
5.94607890e-01 -9.49462593e-01 1.31986165e+00 -1.08632195e+00
-2.12806487e+00 7.70149231e-01 4.60318267e-01 -1.93570748e-01
1.11296308e+00 -4.57930297e-01 2.41895735e-01 -8.65206048e-02
-4.83795464e-01 7.32998252e-01 9.14122522e-01 -1.79626858e+00
-3.82194400e-01 -1.80859402e-01 6.41332805e-01 3.69077057e-01
1.96637690e-01 -1.46056369e-01 4.15354632e-02 -7.02631891e-01
-2.01818034e-01 -1.61965108e+00 -4.26265985e-01 5.03157854e-01
-1.19703941e-01 -2.09423304e-01 1.22520006e+00 -4.76188779e-01
3.23387265e-01 -1.75100577e+00 8.71457100e-01 9.51409563e-02
-1.27796859e-01 4.29312557e-01 -1.86012879e-01 6.72725797e-01
7.22238645e-02 -1.40240550e-01 -4.13606256e-01 4.56793718e-02
2.77327150e-01 5.33856511e-01 -3.17330450e-01 4.37061936e-01
3.43886584e-01 1.00815010e+00 -1.11513090e+00 -1.39981657e-01
1.82650015e-01 4.84633476e-01 -7.87283778e-01 4.14103746e-01
-8.31286788e-01 7.00466514e-01 -7.12269068e-01 6.04241848e-01
4.58701611e-01 1.93414330e-01 2.45275751e-01 7.46359974e-02
-1.73823193e-01 -2.19637930e-01 -1.36183810e+00 1.66586506e+00
-4.10773337e-01 2.27387860e-01 6.63961232e-01 -1.11592948e+00
7.30129182e-01 3.50838959e-01 5.88469803e-01 -5.84484823e-02
1.04872644e-01 2.74442792e-01 2.61256933e-01 -1.02312160e+00
2.75852621e-01 -5.08937180e-01 -5.70036247e-02 1.99824095e-01
-2.02669576e-01 -8.52340043e-01 -7.52968788e-02 -4.01622534e-01
1.24110508e+00 1.06505740e+00 -1.79473415e-01 -2.95679182e-01
1.17482044e-01 4.63518173e-01 8.08883160e-02 4.22478288e-01
7.71759152e-02 5.63999295e-01 1.38376221e-01 -7.59788632e-01
-1.30445623e+00 -1.23598719e+00 9.16921645e-02 1.15021956e+00
1.70508504e-01 3.89324784e-01 -8.64438534e-01 -5.81106305e-01
5.81396759e-01 4.50230330e-01 -6.73105299e-01 6.80444613e-02
-1.17119575e+00 -3.54233623e-01 1.96443722e-01 6.44370973e-01
2.03530297e-01 -1.58090198e+00 -1.12466729e+00 5.10563314e-01
4.75721449e-01 -6.75673664e-01 -1.55125782e-01 2.64719605e-01
-9.70984280e-01 -1.03582799e+00 -7.61374831e-01 -8.92263174e-01
6.69227958e-01 4.41068076e-02 9.60288048e-01 1.47090137e-01
-5.58498144e-01 7.23102629e-01 -2.55917817e-01 -4.20393318e-01
-6.27160370e-01 -7.98792541e-02 8.83480459e-02 -5.49755871e-01
-5.02654850e-01 -5.17783582e-01 -6.49376035e-01 4.13425714e-01
-8.78175616e-01 -1.22928053e-01 4.86703634e-01 6.56981170e-01
4.11754370e-01 -1.41621262e-01 3.62557679e-01 -7.06740618e-01
6.09036982e-01 -2.23910779e-01 -7.28889346e-01 1.64286733e-01
2.17295624e-02 -3.15792747e-02 6.62187576e-01 -9.85663056e-01
-8.97193670e-01 3.00853014e-01 -1.43059105e-01 -4.98665333e-01
-3.26918215e-01 2.77771384e-01 1.04108028e-01 -2.73533881e-01
4.75406140e-01 -3.73038322e-01 3.04950923e-01 -2.90998578e-01
4.35107589e-01 1.87277094e-01 2.51342118e-01 -1.16952527e+00
1.02478576e+00 4.95821208e-01 3.00841928e-01 -6.72584295e-01
-3.95346344e-01 2.25712895e-01 -7.95663416e-01 -2.15024590e-01
1.07017946e+00 -3.94952059e-01 -1.24597466e+00 4.48439509e-01
-1.01455164e+00 -1.08185530e+00 -5.37287235e-01 2.67690867e-01
-1.08353496e+00 2.26089805e-01 -5.11174142e-01 -6.65159285e-01
-1.46621093e-01 -1.25667548e+00 1.15464246e+00 6.46000952e-02
-4.02909458e-01 -1.03493917e+00 2.53055722e-01 1.64383188e-01
6.39128327e-01 8.31607044e-01 8.88512075e-01 -1.93024591e-01
-6.89500451e-01 -3.91825885e-01 2.90086478e-01 2.26205379e-01
1.17390506e-01 2.34719485e-01 -4.56980586e-01 -7.26775646e-01
-1.89848617e-01 -7.80218542e-01 2.86929309e-01 2.73078352e-01
1.25568044e+00 -2.70495832e-01 -4.92056578e-01 2.53931016e-01
1.39546299e+00 1.73409864e-01 5.25317729e-01 3.09249938e-01
7.45892882e-01 6.90357566e-01 7.67129421e-01 2.30897576e-01
2.08480045e-01 8.38634968e-01 8.34492624e-01 -3.99538167e-02
-1.40205398e-02 7.99049884e-02 2.85657853e-01 2.51036495e-01
-7.67512381e-01 -3.23960572e-01 -1.00405622e+00 3.40541154e-01
-1.79728568e+00 -1.04634619e+00 7.94116482e-02 2.03360200e+00
8.30790281e-01 1.30604535e-01 1.71112284e-01 -2.87869304e-01
3.21155250e-01 -2.10948914e-01 -7.60879099e-01 -5.01802921e-01
5.93579233e-01 3.24241757e-01 2.49717563e-01 5.24600327e-01
-1.01801431e+00 9.17055964e-01 5.69982767e+00 2.99671322e-01
-1.44597399e+00 -3.12356323e-01 7.92771652e-02 -3.69640626e-02
5.16810603e-02 1.64325923e-01 -4.93224293e-01 3.15554827e-01
4.34458375e-01 3.70416045e-01 8.74731481e-01 8.88370872e-01
2.96682328e-01 -1.33184358e-01 -1.27368915e+00 3.04276437e-01
-2.86271185e-01 -9.51784670e-01 -1.60564050e-01 -1.14923259e-02
7.21734166e-01 -1.37991697e-01 -7.23665133e-02 4.29451168e-01
7.76590407e-01 -1.05448627e+00 9.07850921e-01 6.94679856e-01
3.35124969e-01 -4.77122843e-01 2.84934759e-01 6.45866811e-01
-9.58753169e-01 -6.82368036e-03 -3.55130106e-01 -1.42884716e-01
2.69397020e-01 -1.11159645e-01 -7.03432858e-01 3.50889057e-01
6.25198483e-01 3.02539289e-01 2.42759973e-01 8.76543462e-01
4.97488976e-02 3.07898551e-01 -3.65654081e-01 -2.91472107e-01
3.88689786e-01 -1.95290297e-01 6.28177106e-01 1.05717051e+00
8.61387029e-02 2.85823524e-01 7.04864204e-01 5.81525505e-01
-6.56960160e-02 -4.68076646e-01 -7.22248614e-01 1.17827713e-01
-1.55838653e-01 1.25817096e+00 -7.47213542e-01 -1.19853035e-01
-2.05179676e-01 9.85698998e-01 5.31230927e-01 3.10766637e-01
-1.06220615e+00 -2.79025108e-01 8.01992416e-01 4.42893326e-01
4.63749081e-01 -7.70764947e-01 1.46746829e-01 -9.17477727e-01
6.55865371e-02 -1.00288486e+00 -1.65432513e-01 -7.67020226e-01
-1.45190883e+00 2.32069299e-01 2.14532256e-01 -1.20477581e+00
-1.63817674e-01 -7.17573762e-01 -4.86333251e-01 7.06642091e-01
-1.24298334e+00 -1.25063109e+00 -3.03907931e-01 2.55163699e-01
6.79890335e-01 2.28032276e-01 8.75344157e-01 -7.80685246e-02
1.36147186e-01 -5.37081845e-02 1.49810463e-01 8.51110145e-02
5.89899540e-01 -1.24742699e+00 2.05776423e-01 1.25363052e-01
-3.96132499e-01 5.84346712e-01 8.69519770e-01 -7.58024812e-01
-2.05265427e+00 -8.78738642e-01 -1.13222905e-01 -7.98297942e-01
7.59795368e-01 -4.95954394e-01 -1.19435930e+00 8.33135486e-01
3.00448835e-01 4.14217740e-01 -1.25555366e-01 -2.54928499e-01
1.58649057e-01 1.06832020e-01 -1.31730974e+00 7.05578923e-01
1.19509602e+00 -9.82829742e-03 -7.63677776e-01 4.06492263e-01
5.19302070e-01 -1.10981703e+00 -1.05733025e+00 6.07598543e-01
9.42500532e-01 -3.50181788e-01 1.00439119e+00 -1.07686710e+00
6.27061486e-01 -5.07503927e-01 -2.32675355e-02 -1.55590796e+00
-2.58987725e-01 -7.66002595e-01 -7.65665546e-02 9.21691895e-01
1.66306496e-01 -4.61052358e-01 8.95595133e-01 7.54519403e-01
-2.91267812e-01 -1.11275601e+00 -8.08423579e-01 -8.17281067e-01
5.50519407e-01 1.34140283e-01 2.01985613e-01 9.88525212e-01
-2.09118783e-01 -1.77666500e-01 -3.35900724e-01 3.52245003e-01
4.02054518e-01 3.99674535e-01 9.60637450e-01 -1.12431443e+00
-6.70850635e-01 -1.83369949e-01 -2.06430823e-01 -9.26560104e-01
2.82333851e-01 -8.06793690e-01 4.75181460e-01 -1.69505751e+00
-4.49467264e-02 -7.41966248e-01 3.64630699e-01 6.60119116e-01
-4.04923316e-03 -2.02172101e-02 4.69408751e-01 6.30603060e-02
-3.28726411e-01 4.35190499e-01 1.99148870e+00 -1.13456575e-02
-2.17526361e-01 1.81513548e-01 -7.90356770e-02 9.30516303e-01
6.96856022e-01 -4.26623434e-01 -1.85247064e-01 -6.73171401e-01
8.32296982e-02 3.52333039e-01 6.66700542e-01 -9.68177557e-01
-5.65505028e-02 -6.87286198e-01 4.38636601e-01 4.46436368e-02
4.48099583e-01 -1.22147298e+00 2.67668098e-01 7.66518891e-01
-2.67503351e-01 2.29879901e-01 4.10697132e-01 6.02533817e-01
2.45307893e-01 -2.33795837e-01 1.02165818e+00 -5.33342421e-01
-4.03926522e-01 2.23657578e-01 -5.13932347e-01 1.60847548e-02
1.45124447e+00 7.23574450e-03 -7.32995793e-02 7.44061247e-02
-1.09920084e+00 2.40974545e-01 8.28406751e-01 5.66494882e-01
4.71219748e-01 -9.25143540e-01 -6.27432823e-01 -8.31290558e-02
-4.89972651e-01 3.70716423e-01 -6.42580986e-02 2.94877917e-01
-8.56635809e-01 -1.74609140e-01 -4.29413050e-01 -4.77531850e-01
-1.18432379e+00 6.72149897e-01 5.60779274e-01 -2.33789042e-01
-7.44605482e-01 6.22872174e-01 -1.45872965e-01 -8.18929315e-01
9.50144455e-02 -5.05150139e-01 3.59213322e-01 -4.30521041e-01
-1.94802314e-01 3.66469532e-01 -6.70457184e-02 -8.27440023e-02
-1.67495117e-01 9.92643476e-01 1.35680214e-02 -9.34741572e-02
1.71144629e+00 4.78875279e-01 -9.06361453e-03 1.56025514e-01
7.85119474e-01 -1.37220904e-01 -1.88560605e+00 3.62376273e-01
-2.07254872e-01 -3.86193424e-01 -5.34071565e-01 -8.32817793e-01
-9.00621951e-01 7.95858920e-01 5.16311288e-01 3.53703260e-01
6.11446857e-01 -1.18248546e-02 6.84395432e-01 8.03504109e-01
8.16351533e-01 -1.00244713e+00 5.05685151e-01 6.53724551e-01
1.35936892e+00 -1.22162068e+00 2.29164496e-01 -2.33426541e-01
-6.07500732e-01 1.25953472e+00 6.81588769e-01 -9.12795365e-01
6.33020520e-01 7.96871662e-01 -6.52716085e-02 -1.07453562e-01
-4.97979730e-01 7.31155500e-02 -1.92446318e-02 7.45930970e-01
3.34096014e-01 3.28820907e-02 1.24024823e-01 -2.27484375e-01
5.22123277e-02 1.99754357e-01 3.01930964e-01 1.56065786e+00
-5.85496545e-01 -1.28370452e+00 -3.36478233e-01 1.78143620e-01
-3.60839158e-01 6.24059081e-01 -1.97716370e-01 1.28681540e+00
1.45416752e-01 4.06687886e-01 -2.61458009e-01 2.27522820e-01
8.15138340e-01 -1.14212576e-02 1.10724282e+00 -6.55975044e-01
-8.45450759e-01 -2.57622629e-01 -8.60240683e-02 -4.87048149e-01
-4.30238932e-01 -7.79962301e-01 -1.62666857e+00 -3.05623114e-01
-4.86898310e-02 -2.81362921e-01 7.13096678e-01 8.38592827e-01
1.34993732e-01 5.02222717e-01 2.64622808e-01 -1.83045959e+00
-1.03510451e+00 -5.61048687e-01 -2.57913917e-01 4.98785615e-01
4.75699812e-01 -8.96699607e-01 -2.76689529e-02 2.12455854e-01] | [4.831493854522705, 0.48212817311286926] |
589e182f-d753-445f-bf0a-29ba629888bc | interactive-log-parsing-via-light-weight-user | 2301.12225 | null | https://arxiv.org/abs/2301.12225v2 | https://arxiv.org/pdf/2301.12225v2.pdf | Interactive Log Parsing via Light-weight User Feedback | Template mining is one of the foundational tasks to support log analysis, which supports the diagnosis and troubleshooting of large scale Web applications. This paper develops a human-in-the-loop template mining framework to support interactive log analysis, which is highly desirable in real-world diagnosis or troubleshooting of Web applications but yet previous template mining algorithms fails to support it. We formulate three types of light-weight user feedbacks and based on them we design three atomic human-in-the-loop template mining algorithms. We derive mild conditions under which the outputs of our proposed algorithms are provably correct. We also derive upper bounds on the computational complexity and query complexity of each algorithm. We demonstrate the versatility of our proposed algorithms by combining them to improve the template mining accuracy of five representative algorithms over sixteen widely used benchmark datasets. | ['John C. S. Lui', 'Jian Tan', 'Ye Li', 'Hong Xie', 'Liming Wang'] | 2023-01-28 | null | null | null | null | ['log-parsing'] | ['computer-code'] | [ 4.53846902e-01 -1.88856706e-01 -4.62189525e-01 -2.22792700e-01
-4.92347986e-01 -6.48409426e-01 9.22990590e-02 3.64528716e-01
1.07819833e-01 4.35042351e-01 -5.15447319e-01 -1.02255118e+00
-5.96070230e-01 -9.96386111e-01 -5.93249977e-01 -2.83633649e-01
-3.17888439e-01 4.68847364e-01 7.38208115e-01 1.41429275e-01
4.51333344e-01 3.74647230e-01 -1.73494565e+00 6.33204699e-01
8.80668163e-01 1.26840055e+00 -3.19183730e-02 8.79989743e-01
1.65180027e-01 7.08347201e-01 -2.63520777e-01 -2.80591607e-01
4.33578849e-01 -3.64360362e-01 -6.94332659e-01 6.29999042e-01
-1.03800237e-01 -2.09698290e-01 -1.55889586e-01 1.04166675e+00
1.60995692e-01 -2.09971279e-01 3.06835234e-01 -1.70466888e+00
-2.84546673e-01 4.00129825e-01 -6.84980690e-01 2.99867094e-01
8.48987818e-01 -2.70048957e-02 9.60826099e-01 -7.89304018e-01
6.06058955e-01 7.34239757e-01 2.46530339e-01 1.84093621e-02
-9.47522700e-01 -4.74695861e-01 6.19045757e-02 7.00091124e-01
-1.19165039e+00 -4.06532049e-01 6.69872582e-01 -1.65080100e-01
1.32620382e+00 6.58116639e-01 4.87084299e-01 3.77757549e-01
2.64907569e-01 8.11017632e-01 9.77364779e-01 -5.94668627e-01
2.98286349e-01 3.91234100e-01 4.21023577e-01 9.81000185e-01
9.05565083e-01 4.54208329e-02 -9.05891359e-01 -7.91239023e-01
3.98342609e-01 3.47613692e-01 -3.49112004e-02 -2.92958200e-01
-9.31785822e-01 5.49327075e-01 -4.04123783e-01 -2.19311535e-01
-3.17483693e-01 -2.12722525e-01 3.16313654e-01 8.27745795e-01
6.15593381e-02 -6.96139038e-02 -6.86236024e-01 -2.26223558e-01
-5.66044748e-01 1.73167050e-01 1.20017779e+00 1.39497054e+00
8.15255046e-01 -3.29761684e-01 7.51984566e-02 2.52891034e-01
1.52109221e-01 4.97539461e-01 3.73370588e-01 -6.44742608e-01
2.37415150e-01 1.13743651e+00 1.07110403e-01 -9.58924949e-01
-1.51843444e-01 -3.29740494e-01 -4.79265571e-01 -7.14003071e-02
9.57198739e-02 2.30371058e-01 -3.34839910e-01 1.13028729e+00
4.69987303e-01 1.70047805e-01 -4.45005208e-01 3.31260592e-01
-6.42636567e-02 3.30712318e-01 -3.83608043e-01 -1.03079641e+00
1.47901261e+00 -7.49614298e-01 -6.88655138e-01 -4.54223864e-02
7.68439293e-01 -7.05538571e-01 1.27270937e+00 1.02688253e+00
-8.80371988e-01 -1.50144547e-01 -1.09645379e+00 5.42698085e-01
-6.35470077e-02 2.33791508e-02 8.34885657e-01 8.90961587e-01
-6.52350485e-01 4.06941712e-01 -8.30170751e-01 -3.61264169e-01
2.48437207e-02 5.37706256e-01 -1.59722045e-01 9.07642767e-02
-5.10967016e-01 4.17200267e-01 3.89323980e-01 -3.04058313e-01
-6.56995535e-01 -3.52137476e-01 -4.15284067e-01 3.86297703e-02
1.16680896e+00 -5.64984858e-01 1.57144701e+00 -3.68604660e-01
-9.57330287e-01 8.42376590e-01 -5.11315465e-01 -3.35504264e-01
6.79171801e-01 -1.15041204e-01 -5.11781871e-01 -4.46557738e-02
-8.04018602e-02 -5.09128630e-01 6.62030220e-01 -7.97718108e-01
-1.13329697e+00 -5.47817051e-01 -5.32349683e-02 -2.94618666e-01
-6.71571374e-01 1.49189219e-01 -4.87981975e-01 -2.57048607e-01
5.90480305e-02 -8.51194382e-01 -2.64256239e-01 -2.54471116e-02
-4.11979347e-01 -4.81630594e-01 1.15170789e+00 -1.06366634e-01
1.87879181e+00 -1.92528284e+00 -5.49933612e-01 5.22027135e-01
3.42945874e-01 2.22299285e-02 1.08290091e-01 6.08786166e-01
5.54832295e-02 1.53536096e-01 1.24699473e-01 2.94118047e-01
-8.99698643e-05 2.09718466e-01 -3.76466393e-01 3.93995523e-01
1.78087533e-01 7.40929425e-01 -7.70463824e-01 -6.02188587e-01
-8.20759088e-02 -4.70585674e-01 -5.90499878e-01 5.47110379e-01
-3.95416081e-01 -3.80707592e-01 -6.06687427e-01 8.13499510e-01
3.33433002e-01 -7.64118612e-01 4.77666914e-01 1.92980021e-01
1.45486280e-01 4.35793012e-01 -1.37204480e+00 8.64940107e-01
-1.60365000e-01 1.15508854e-01 -1.30620241e-01 -9.45932448e-01
7.82280803e-01 4.29330736e-01 6.69341326e-01 -5.87634206e-01
-1.07398376e-01 2.68176973e-01 -2.04605516e-02 -8.21473956e-01
3.04813776e-02 3.99221987e-01 -9.44320112e-02 1.39260435e+00
-5.12794435e-01 6.11574769e-01 6.88075066e-01 2.17465460e-01
1.72087073e+00 -4.39709783e-01 9.66903389e-01 -3.83021683e-02
6.16128922e-01 9.68601033e-02 7.44683325e-01 7.66229033e-01
-1.42902866e-01 -6.37470484e-02 8.17344010e-01 -5.85253894e-01
-7.15927839e-01 -8.85707498e-01 2.68573523e-01 1.30463529e+00
-6.24291897e-02 -8.44394922e-01 -4.44371551e-01 -1.02912354e+00
6.64587840e-02 3.62474054e-01 -4.53045398e-01 8.18876922e-03
-4.19339657e-01 -7.18532741e-01 1.56097770e-01 3.99028718e-01
2.13311285e-01 -1.11519134e+00 -1.06196070e+00 4.16223347e-01
-8.16523843e-03 -1.18329692e+00 -4.55918044e-01 4.29667920e-01
-1.02875781e+00 -1.91043508e+00 1.81239545e-01 -6.42260313e-01
7.80534744e-01 7.36812055e-01 1.07461357e+00 5.28249145e-01
-3.28450263e-01 2.33285561e-01 -6.03650868e-01 -7.07722485e-01
-4.36968386e-01 -5.96050620e-02 2.65780002e-01 -1.79062542e-02
6.64132535e-01 -8.51224363e-01 -3.20430756e-01 7.16357827e-01
-9.35846746e-01 -9.90071967e-02 7.42427766e-01 6.24841392e-01
7.64140666e-01 5.91088653e-01 7.59685278e-01 -1.55691588e+00
1.04615664e+00 -4.99480128e-01 -8.25725734e-01 6.33445442e-01
-1.48239195e+00 1.80648595e-01 6.85501933e-01 -6.05450869e-01
-5.74456573e-01 4.10179496e-01 2.95174032e-01 -1.90144017e-01
-3.53632448e-03 5.74462891e-01 -3.38257730e-01 -1.35324076e-01
7.16478407e-01 4.42325562e-01 -2.50408590e-01 -3.36458892e-01
-9.79105011e-02 9.29949522e-01 5.89984715e-01 -5.58639228e-01
1.03452063e+00 1.90711662e-01 3.73884179e-02 -3.85038197e-01
-5.18067062e-01 -7.60286391e-01 -2.39580497e-01 -9.13373530e-02
-9.21369940e-02 -1.60966843e-01 -1.42744434e+00 3.82636487e-02
-9.70130980e-01 7.72645175e-02 8.35894048e-02 -2.12288350e-01
-7.16669858e-01 6.11854970e-01 -3.86932820e-01 -1.10329843e+00
-7.16429293e-01 -9.15315628e-01 5.78489423e-01 -7.30913430e-02
-1.97722852e-01 -4.84038085e-01 2.33512209e-03 3.87714505e-02
7.99279287e-02 1.26078174e-01 1.18037438e+00 -1.09274459e+00
-7.57892489e-01 -4.69024748e-01 -2.42435575e-01 -1.76563621e-01
3.92518848e-01 -1.92799404e-01 -6.76884472e-01 -2.24391237e-01
1.44199222e-01 1.12513900e-01 3.44931751e-01 -4.03783590e-01
1.50178254e+00 -8.40427995e-01 -4.11522448e-01 2.72013366e-01
1.24967098e+00 5.20790339e-01 4.40907419e-01 3.24620634e-01
7.68070444e-02 2.45300636e-01 1.02021933e+00 9.97483313e-01
-3.33643937e-03 4.57116067e-01 2.18322799e-01 4.49475944e-01
6.33737147e-01 -2.84956098e-01 5.33341825e-01 8.69370878e-01
1.90522730e-01 -2.18857944e-01 -6.84638143e-01 6.00772619e-01
-2.17663670e+00 -7.18405426e-01 -2.81015903e-01 2.31534767e+00
1.02668881e+00 6.45587265e-01 4.27808553e-01 9.25433457e-01
4.00951803e-01 -4.57925141e-01 -6.58255339e-01 -4.62275863e-01
4.65717107e-01 3.08085531e-01 3.13261896e-01 3.55534144e-02
-8.45417976e-01 4.33373541e-01 6.29911137e+00 6.56422019e-01
-7.50153780e-01 -6.93499744e-02 1.93885893e-01 3.06101590e-01
-2.03578874e-01 2.05655545e-01 -6.68570459e-01 2.61908829e-01
1.04916883e+00 -9.68160927e-01 4.59391713e-01 1.45959091e+00
4.00086880e-01 -2.20743507e-01 -1.59626615e+00 9.08301115e-01
-2.98132390e-01 -1.23639774e+00 1.52211990e-02 2.33273298e-01
4.64055628e-01 -7.02517807e-01 -1.35926947e-01 -9.49504077e-02
2.74841804e-02 -6.12275839e-01 2.25279897e-01 -5.32830693e-02
9.19759691e-01 -6.61116540e-01 4.67705131e-01 9.23656821e-01
-1.29741681e+00 -3.23623657e-01 -1.84438586e-01 -3.03081512e-01
1.41109014e-02 8.90033603e-01 -1.38357580e+00 4.65595126e-01
5.08766949e-01 2.67450154e-01 -7.37240195e-01 1.08395135e+00
-7.58946314e-02 6.40876830e-01 -5.15346467e-01 -3.34456384e-01
-5.24884939e-01 1.76340312e-01 2.72528261e-01 1.09973538e+00
3.04784238e-01 1.20016560e-01 3.83416384e-01 3.78939986e-01
-9.54795480e-02 1.25160784e-01 -7.64171362e-01 -3.21151733e-01
7.91426420e-01 1.07046747e+00 -8.96458447e-01 -5.05397797e-01
-4.86294985e-01 6.94756567e-01 -2.52594743e-02 -3.27172391e-02
-5.60439944e-01 -7.19814956e-01 3.87704015e-01 7.52280772e-01
2.20234573e-01 -2.62379199e-01 -5.63208282e-01 -9.76853132e-01
6.36602342e-01 -1.30256748e+00 8.08761477e-01 -1.18735172e-01
-1.25481296e+00 3.78527045e-01 -1.07792735e-01 -1.33506978e+00
-5.98269045e-01 -7.18254924e-01 -8.62686992e-01 1.52852476e-01
-1.09055960e+00 -8.39406013e-01 -4.96073693e-01 1.08105028e+00
5.28341651e-01 -1.12040997e-01 9.69816864e-01 2.94309497e-01
-4.12070721e-01 8.87274563e-01 -3.65241766e-01 -2.72457182e-01
6.23701274e-01 -1.07818663e+00 3.72528434e-01 1.09375572e+00
2.76969343e-01 8.73258173e-01 8.35758567e-01 -9.00644958e-01
-2.03192186e+00 -9.45763588e-01 8.62459958e-01 -1.78141549e-01
7.05309868e-01 -2.57442445e-01 -8.21620226e-01 6.39998853e-01
-1.97911590e-01 -7.85307884e-02 8.17728341e-01 1.11031346e-01
-4.42478329e-01 -3.73569787e-01 -1.03618693e+00 1.81486100e-01
1.17350233e+00 -6.18762612e-01 -2.80787617e-01 5.67424357e-01
6.74586356e-01 -6.52193725e-02 -5.59624910e-01 4.94900495e-01
6.65285349e-01 -1.11250961e+00 6.00563884e-01 -8.28475833e-01
-1.03783920e-01 -4.04396474e-01 2.97016129e-02 -3.76535773e-01
-1.22065932e-01 -1.22776186e+00 -9.05295014e-01 5.70993900e-01
3.13111544e-01 -6.44807518e-01 1.15940452e+00 4.40154403e-01
1.94500849e-01 -1.07080042e+00 -5.72889268e-01 -9.21601772e-01
-9.53496337e-01 -8.48050892e-01 4.98177379e-01 6.07747257e-01
6.06550336e-01 3.14770460e-01 -3.74668390e-01 -9.95150115e-03
7.78617024e-01 3.83962363e-01 8.90314698e-01 -1.31142092e+00
-7.58197963e-01 -1.61681548e-01 -2.45569080e-01 -1.01097167e+00
-3.02056551e-01 -6.71697378e-01 -2.82067984e-01 -1.07460153e+00
5.69419801e-01 -3.25005025e-01 -3.95223260e-01 7.04774678e-01
-2.81206556e-02 -1.08947031e-01 -1.32184565e-01 1.82505414e-01
-9.89067912e-01 -1.29878089e-01 7.88287461e-01 1.42722771e-01
-3.93345565e-01 6.51017904e-01 -8.22196364e-01 6.04106724e-01
5.73427558e-01 -7.28555620e-01 -6.81085050e-01 2.03205541e-01
7.63468444e-01 3.78942549e-01 -2.19769571e-02 -6.24028027e-01
6.62899733e-01 -7.18319714e-01 3.64261307e-02 -6.91071033e-01
-4.40611631e-01 -9.36410487e-01 2.91985601e-01 8.39741528e-01
-5.05967848e-02 5.49582124e-01 -2.38666847e-01 8.55782747e-01
7.51683861e-02 -2.44226959e-02 4.63434428e-01 6.88022152e-02
-6.30002320e-01 4.76754725e-01 -4.17587012e-01 -8.97351950e-02
1.30795312e+00 -3.34970057e-01 -2.01273367e-01 -3.60003501e-01
-3.97513360e-01 2.32082635e-01 3.49808812e-01 2.05993488e-01
8.44223678e-01 -1.02555823e+00 -2.37827271e-01 5.51067173e-01
4.21830475e-01 -2.74523616e-01 -5.34928620e-01 9.17644799e-01
-4.48790610e-01 5.06528914e-01 -1.31250098e-01 -3.57597202e-01
-1.71325779e+00 1.04087436e+00 -1.28918663e-01 -6.43689513e-01
-4.15246129e-01 1.91012546e-01 -1.38547450e-01 2.15167999e-01
5.53005755e-01 -7.15772286e-02 3.41170043e-01 -4.76166457e-01
8.07559431e-01 6.26277924e-01 4.83858138e-01 4.68811929e-01
-5.47574222e-01 5.82182407e-02 -3.71156633e-01 3.87494788e-02
1.35165370e+00 -3.52476798e-02 -3.04596663e-01 2.52793372e-01
6.68455124e-01 1.32539496e-01 -5.24271905e-01 -4.61284667e-01
7.73668647e-01 -6.36102736e-01 -5.02258778e-01 -6.41515076e-01
-5.52669346e-01 6.25545323e-01 3.91705751e-01 6.37270033e-01
1.75836420e+00 -1.74907804e-01 9.84776258e-01 1.02498567e+00
8.32247198e-01 -1.06759286e+00 1.53724894e-01 1.87653482e-01
4.09536600e-01 -9.71406579e-01 1.83935434e-01 -8.44628155e-01
-2.95922667e-01 1.24175382e+00 6.41293705e-01 -5.60144521e-02
8.33455563e-01 6.99773073e-01 -4.22482431e-01 -5.08936048e-02
-1.68067002e+00 -8.04968819e-04 5.05163781e-02 5.39865911e-01
2.99765468e-01 7.93651119e-02 -7.65402079e-01 1.05308700e+00
1.18042298e-01 3.92009258e-01 6.08428776e-01 1.47719479e+00
-7.63298512e-01 -1.60310245e+00 -2.00694963e-01 7.28195548e-01
-5.33342481e-01 2.73824334e-01 -6.81612194e-01 5.24718583e-01
-1.49217084e-01 1.12030208e+00 -4.96733814e-01 -8.73396456e-01
4.37409699e-01 1.19666919e-01 3.93191487e-01 -5.83713830e-01
-4.12796021e-01 1.76607102e-01 1.75910950e-01 -7.02153504e-01
2.20032483e-01 -3.65385145e-01 -1.20110440e+00 -5.67328870e-01
-5.10079384e-01 1.66745856e-01 2.14206219e-01 7.92748511e-01
5.01717985e-01 1.57719120e-01 1.04022121e+00 1.26948059e-01
-9.04416978e-01 -7.86354005e-01 -5.16834915e-01 3.20728481e-01
5.01278155e-02 -4.44702238e-01 -1.31079361e-01 3.83634567e-01] | [8.127935409545898, 6.622040748596191] |
3a9172ae-5915-4c6f-8065-66ea0f52b3f3 | identifying-the-context-shift-between-test | 2207.01059 | null | https://arxiv.org/abs/2207.01059v2 | https://arxiv.org/pdf/2207.01059v2.pdf | Identifying the Context Shift between Test Benchmarks and Production Data | Machine learning models are often brittle on production data despite achieving high accuracy on benchmark datasets. Benchmark datasets have traditionally served dual purposes: first, benchmarks offer a standard on which machine learning researchers can compare different methods, and second, benchmarks provide a model, albeit imperfect, of the real world. The incompleteness of test benchmarks (and the data upon which models are trained) hinder robustness in machine learning, enable shortcut learning, and leave models systematically prone to err on out-of-distribution and adversarially perturbed data. The mismatch between a single static benchmark dataset and a production dataset has traditionally been described as a dataset shift. In an effort to clarify how to address the mismatch between test benchmarks and production data, we introduce context shift to describe semantically meaningful changes in the underlying data generation process. Moreover, we identify three methods for addressing context shift that would otherwise lead to model prediction errors: first, we describe how human intuition and expert knowledge can identify semantically meaningful features upon which models systematically fail, second, we detail how dynamic benchmarking - with its focus on capturing the data generation process - can promote generalizability through corroboration, and third, we highlight that clarifying a model's limitations can reduce unexpected errors. Robust machine learning is focused on model performance beyond benchmarks, and as such, we consider three model organism domains - facial expression recognition, deepfake detection, and medical diagnosis - to highlight how implicit assumptions in benchmark tasks lead to errors in practice. By paying close attention to the role of context, researchers can design more comprehensive benchmarks, reduce context shift errors, and increase generalizability. | ['Matthew Groh'] | 2022-07-03 | null | null | null | null | ['facial-expression-recognition'] | ['computer-vision'] | [ 5.74126422e-01 1.75697938e-01 -2.98764855e-01 -5.26284277e-01
-9.32347953e-01 -8.14707518e-01 4.30454373e-01 3.74827199e-02
-1.66932479e-01 6.02755904e-01 2.58835822e-01 -2.68614233e-01
-4.18613218e-02 -7.03373611e-01 -9.77270484e-01 -2.76648819e-01
2.22182751e-01 3.17075878e-01 -2.67283112e-01 -1.40516758e-01
8.13441873e-02 3.70154381e-01 -1.66670680e+00 6.72723293e-01
5.97897470e-01 9.31672394e-01 -3.22309405e-01 2.92130291e-01
9.45769995e-02 8.02840829e-01 -9.46109653e-01 -7.54260123e-01
3.69495749e-01 -4.61688876e-01 -9.30304050e-01 6.50675744e-02
5.98741770e-01 -2.82778263e-01 4.77170274e-02 8.99561048e-01
5.54843664e-01 -1.30784914e-01 7.30869830e-01 -1.74221027e+00
-8.77721608e-01 4.87782180e-01 -1.98643692e-02 -1.08458057e-01
4.07427579e-01 7.06968904e-01 1.02697539e+00 -6.25194013e-01
9.71167088e-01 1.14225388e+00 9.96267080e-01 9.45415914e-01
-1.47205877e+00 -7.82433271e-01 1.14303745e-01 -2.07587644e-01
-1.16880190e+00 -6.17598712e-01 6.81286693e-01 -6.56217515e-01
7.20094860e-01 5.44597030e-01 5.13089120e-01 1.67799485e+00
8.05006456e-03 6.34781778e-01 1.08692777e+00 -1.47215948e-01
1.19864456e-01 4.83123101e-02 -5.09571880e-02 6.12317264e-01
3.81248385e-01 5.70899963e-01 -3.55406463e-01 -2.12629557e-01
3.84506285e-01 -1.95224851e-01 -3.49079251e-01 -3.08655918e-01
-1.08854103e+00 6.14960790e-01 2.99699813e-01 3.20594907e-02
-2.28740945e-02 1.49121836e-01 6.65437996e-01 3.23285252e-01
1.98294178e-01 1.08308601e+00 -7.66540408e-01 -2.77562141e-01
-8.09025764e-01 3.20236742e-01 7.05725074e-01 8.70232701e-01
6.21266901e-01 1.71836942e-01 -1.88789040e-01 8.09233546e-01
-9.91805494e-02 3.81258219e-01 6.05881810e-01 -1.01869380e+00
2.65163779e-01 6.60568953e-01 -2.40244344e-02 -9.81269717e-01
-3.78584236e-01 -4.38378602e-01 -4.76958573e-01 2.09454030e-01
7.29373932e-01 -1.71939760e-01 -8.64656627e-01 2.13421869e+00
4.30303253e-02 9.04107466e-02 1.60054132e-01 7.53779531e-01
8.65774870e-01 7.79689923e-02 9.62956697e-02 8.08512121e-02
1.07612813e+00 -5.89866936e-01 -2.08701879e-01 -2.77038127e-01
1.11504662e+00 -6.18081272e-01 1.74854195e+00 3.50705534e-01
-9.27802145e-01 -4.77550179e-01 -9.11644280e-01 -5.60078844e-02
-4.84141082e-01 -3.03650200e-01 5.85503876e-01 6.47048712e-01
-7.54913151e-01 7.84218013e-01 -6.23712361e-01 -3.74351919e-01
5.61035037e-01 1.99241713e-01 -5.65491319e-01 -2.36809645e-02
-1.07571387e+00 1.03879142e+00 2.30538815e-01 -1.12875752e-01
-9.77775812e-01 -1.16221344e+00 -8.40551853e-01 -7.32007399e-02
2.16441497e-01 -8.20314884e-01 1.33006001e+00 -1.52268434e+00
-9.67485845e-01 1.14332640e+00 2.41954215e-02 -4.08725142e-01
5.90252697e-01 3.37471906e-03 -6.68147445e-01 -2.38985345e-01
6.54978231e-02 8.78386974e-01 5.87678611e-01 -1.45575118e+00
-3.60836536e-01 -2.37425879e-01 4.03281599e-02 -9.26029533e-02
-1.08528472e-01 5.51283546e-03 -6.95918500e-02 -7.16488123e-01
-2.39146203e-01 -9.52478588e-01 3.13187167e-02 2.18746718e-02
-4.57097054e-01 1.67586803e-01 6.54609263e-01 -3.50416601e-01
1.17454588e+00 -2.28000808e+00 -2.88633645e-01 7.89354891e-02
1.12226538e-01 1.40954480e-01 -3.92748624e-01 1.47616044e-01
-4.77498472e-01 6.88084483e-01 -3.37424487e-01 -8.82079545e-03
5.79833016e-02 2.42325395e-01 -5.42653799e-01 1.72465265e-01
6.20905340e-01 1.07897735e+00 -8.83728802e-01 -1.82679132e-01
-1.23193868e-01 2.67648071e-01 -8.66039217e-01 -2.17519980e-03
-3.19856614e-01 3.83269072e-01 -2.64774747e-02 9.46785152e-01
3.68846536e-01 -2.49693274e-01 -1.35630652e-01 -3.43368322e-01
3.89684618e-01 1.56217873e-01 -8.28222752e-01 1.25010967e+00
-3.98484141e-01 5.31573415e-01 -1.50656864e-01 -6.58691883e-01
8.52735639e-01 3.83928195e-02 3.97933275e-01 -7.64958382e-01
-1.12927780e-01 2.54042685e-01 3.00280124e-01 -5.83228588e-01
3.28540683e-01 -4.15116310e-01 -1.46829888e-01 5.01374602e-01
-2.01609388e-01 -3.91367465e-01 -8.66350252e-03 -1.57960251e-01
1.21679461e+00 1.97247997e-01 7.41892084e-02 -1.31947592e-01
-6.20459802e-02 2.84496665e-01 8.52283120e-01 6.75302386e-01
-4.17175591e-01 8.76959383e-01 7.16812372e-01 -5.93414307e-01
-1.10614240e+00 -1.08773351e+00 -2.98550040e-01 1.11272359e+00
-1.44904941e-01 -4.89360005e-01 -7.95686543e-01 -9.35207963e-01
3.74020636e-01 1.03291464e+00 -9.52797890e-01 -6.79689050e-01
-1.49253190e-01 -8.69524539e-01 1.09540153e+00 7.09322810e-01
1.64203271e-01 -1.07607865e+00 -6.49760306e-01 -7.22694024e-02
-1.49637744e-01 -9.10623252e-01 -3.10182810e-01 1.65752321e-01
-7.65092492e-01 -1.45155442e+00 -2.50601888e-01 -4.19447094e-01
6.99142814e-01 -2.17682555e-01 1.63203442e+00 4.11127120e-01
-3.44272316e-01 4.29100573e-01 -1.92357048e-01 -6.15607202e-01
-8.06537569e-01 -8.45638663e-02 -4.69669513e-02 -3.39681953e-01
5.19832432e-01 -4.05287325e-01 -3.98306459e-01 6.19507670e-01
-1.14554262e+00 1.08288631e-01 3.95345002e-01 1.07459867e+00
6.59429789e-01 -4.10868138e-01 6.91625953e-01 -1.34475207e+00
7.65811801e-01 -5.76643825e-01 -3.97379905e-01 5.10845542e-01
-8.73821795e-01 6.73919544e-02 6.05967462e-01 -5.11038303e-01
-7.01019228e-01 -1.13931745e-01 -1.40263671e-02 -4.87062454e-01
-1.98853329e-01 4.18628722e-01 -2.53056884e-01 2.27701709e-01
1.18039250e+00 3.54948221e-03 2.63427973e-01 -3.45287919e-01
1.46182835e-01 4.12570029e-01 4.61668789e-01 -1.07079053e+00
5.36777854e-01 3.45140934e-01 -9.24180672e-02 -2.22502902e-01
-8.52528989e-01 1.73290923e-01 -5.14070988e-01 6.24045245e-02
6.14451587e-01 -7.70017862e-01 -4.92432207e-01 3.27806264e-01
-8.77862692e-01 -7.64684677e-01 -6.65305793e-01 1.15182862e-01
-6.87744915e-01 -5.34127019e-02 -3.95410269e-01 -2.54186332e-01
-1.44863114e-01 -1.37081313e+00 1.13416481e+00 -1.56005651e-01
-7.68626928e-01 -1.09229755e+00 -2.05675755e-02 3.90426010e-01
3.48553330e-01 7.30285048e-01 1.19738722e+00 -9.69827235e-01
-1.87027678e-01 -1.29847795e-01 -1.59589246e-01 5.76498330e-01
3.35023075e-01 4.55871731e-01 -1.23389113e+00 -2.25989506e-01
-1.70535371e-01 -5.73451996e-01 4.65480864e-01 1.89006012e-02
1.58292150e+00 -3.84824157e-01 -1.57331422e-01 8.05400133e-01
1.05654800e+00 6.49594814e-02 5.75031996e-01 2.88940966e-01
5.49464703e-01 7.80213177e-01 5.13471603e-01 1.30231082e-01
1.40090436e-01 4.41276103e-01 4.28213328e-01 -2.81751484e-01
-2.68160522e-01 -4.33061153e-01 2.67430425e-01 3.35956395e-01
3.28638673e-01 3.93321225e-03 -1.13467526e+00 3.80536795e-01
-1.42605889e+00 -9.90136266e-01 1.32188216e-01 2.34125924e+00
9.95483100e-01 2.98586190e-01 1.36717647e-01 -2.04079468e-02
5.34923851e-01 -1.78123593e-01 -7.19780445e-01 -6.47946656e-01
-3.36306840e-01 6.81722611e-02 2.64872134e-01 3.09866935e-01
-7.15417504e-01 8.56636703e-01 7.19661093e+00 5.74268281e-01
-1.46400261e+00 -3.83345410e-02 1.01410866e+00 -3.70892674e-01
-6.20210826e-01 -1.19664378e-01 -4.68677282e-01 6.28957689e-01
8.24889660e-01 -1.78015783e-01 3.74203593e-01 1.02391398e+00
9.21451524e-02 3.24143946e-01 -1.88778389e+00 8.89613330e-01
-3.56036015e-02 -1.44114125e+00 3.05043429e-01 1.17673002e-01
7.32007682e-01 -8.46657455e-02 3.04244548e-01 4.76294488e-01
3.25877428e-01 -1.61002457e+00 8.20222497e-01 5.31138778e-01
1.07318926e+00 -5.35321653e-01 6.62157178e-01 9.52835828e-02
-6.39210522e-01 -9.23120826e-02 -1.78513035e-01 -9.82626900e-02
-3.36930096e-01 4.39679027e-01 -9.82820213e-01 2.16573432e-01
5.61250508e-01 4.61142182e-01 -8.33248436e-01 7.26476371e-01
-3.12280208e-02 6.48653686e-01 -1.00343220e-01 2.65371680e-01
2.06971280e-02 3.27040702e-01 2.37992451e-01 1.26516533e+00
1.22265697e-01 -2.23867163e-01 4.50410172e-02 9.91950393e-01
-2.29072183e-01 -6.67670742e-02 -9.08206940e-01 -1.72066346e-01
5.45301557e-01 9.18896616e-01 -3.40319872e-01 -7.96279907e-02
-4.25685704e-01 6.18430555e-01 2.39816874e-01 3.68371040e-01
-9.31674302e-01 -2.96498779e-02 1.02282310e+00 3.07820350e-01
-2.82348394e-01 4.48937953e-01 -7.02816665e-01 -1.06784689e+00
1.64523609e-02 -1.56049573e+00 4.91276354e-01 -9.03719425e-01
-1.44823456e+00 4.56462443e-01 -6.84858859e-02 -1.16885829e+00
-3.69213223e-01 -7.25498974e-01 -5.44476569e-01 6.94813192e-01
-1.10200369e+00 -1.17586637e+00 -3.61100107e-01 3.86795580e-01
3.80747497e-01 -1.50911903e-04 1.01213920e+00 1.94178000e-01
-6.35815561e-01 1.12345171e+00 -1.05969772e-01 3.64268124e-01
9.85601962e-01 -1.08853614e+00 4.70259845e-01 5.90314925e-01
1.33734256e-01 7.22934723e-01 6.01974785e-01 -5.87817252e-01
-1.14758325e+00 -1.27760231e+00 4.29417491e-01 -9.39606071e-01
6.34045422e-01 -2.49212608e-01 -9.90232408e-01 8.90978873e-01
-3.59986782e-01 1.61803767e-01 9.81998444e-01 3.38393420e-01
-8.21023047e-01 -2.75899500e-01 -1.43413866e+00 6.70653224e-01
1.28757155e+00 -6.08775079e-01 -4.71342385e-01 2.57238597e-01
6.63883567e-01 -5.41561782e-01 -9.61011827e-01 8.16116273e-01
7.42663324e-01 -9.40243363e-01 7.43032694e-01 -1.25611544e+00
7.97748804e-01 -8.74606967e-02 -2.98736304e-01 -1.41925585e+00
-9.17375535e-02 -5.69888353e-01 1.46042421e-01 1.30288815e+00
7.28984237e-01 -5.93291938e-01 6.86122715e-01 1.12568295e+00
-1.94109887e-01 -1.14470315e+00 -5.63258469e-01 -8.47662926e-01
4.18615699e-01 -7.56807625e-01 1.16982019e+00 1.33442903e+00
-1.36399582e-01 3.16181406e-02 5.35474755e-02 -1.98982388e-01
1.68420613e-01 8.18021689e-03 1.09432435e+00 -1.08140516e+00
-3.06965828e-01 -7.43318915e-01 -5.22012711e-01 -3.94414008e-01
2.93760091e-01 -1.02611387e+00 -1.38324916e-01 -9.59593952e-01
2.49207824e-01 -6.59559608e-01 -2.73090929e-01 7.28840947e-01
-2.06844822e-01 2.86906481e-01 2.72199094e-01 1.37791380e-01
-3.13604444e-01 1.39662847e-01 1.18202627e+00 -2.35581677e-02
-6.42201677e-03 -1.42111406e-01 -1.20680797e+00 7.93443680e-01
6.24437332e-01 -4.82651234e-01 -5.84834754e-01 -6.18779838e-01
2.48015195e-01 -1.97770104e-01 6.72139883e-01 -8.65711331e-01
-2.19769731e-01 -4.17069286e-01 5.47982931e-01 1.27039954e-01
1.09540164e-01 -7.79062092e-01 4.26208407e-01 4.84028846e-01
-5.96514702e-01 2.50034928e-01 5.54362297e-01 2.40738556e-01
-1.38541415e-01 -2.23033279e-01 7.98479199e-01 -1.44357324e-01
-6.24896765e-01 1.55113131e-01 1.25374690e-01 5.65657914e-01
1.02117503e+00 -4.07334089e-01 -6.58584833e-01 -1.94754839e-01
-8.02985370e-01 1.10629700e-01 1.04492342e+00 6.02786064e-01
3.17557096e-01 -1.27264118e+00 -6.61240458e-01 2.64945239e-01
5.15791714e-01 4.92215157e-03 -1.26346638e-02 5.83896518e-01
-3.93106490e-01 1.02727778e-01 -2.22864076e-01 -5.32226443e-01
-1.22852504e+00 6.23730838e-01 6.62764430e-01 -9.11557674e-02
-1.96686342e-01 8.43975902e-01 3.62532437e-01 -6.55998528e-01
2.46887967e-01 -5.98808050e-01 4.70723957e-01 -5.99266961e-02
2.42809072e-01 1.59062803e-01 2.47075260e-01 -4.34784800e-01
-3.44618648e-01 1.67131364e-01 -1.50059612e-04 1.34364739e-01
1.14599717e+00 2.85915285e-01 1.41616315e-01 4.86775905e-01
1.15659702e+00 8.10661465e-02 -1.45980978e+00 1.94397360e-01
6.56092837e-02 -4.47818637e-01 -3.62256050e-01 -1.37214482e+00
-8.89833927e-01 8.04955304e-01 4.93775278e-01 1.42354846e-01
1.27253294e+00 -1.35644302e-01 5.65525830e-01 3.18556756e-01
1.92983374e-01 -9.69036937e-01 -4.25929530e-03 2.10057005e-01
9.28988516e-01 -1.39478815e+00 -1.25345528e-01 -2.25347430e-01
-7.64865458e-01 9.49398994e-01 9.93129432e-01 2.45442018e-01
2.65681386e-01 4.32599515e-01 9.83298942e-02 -1.31279632e-01
-9.00525451e-01 1.40868649e-01 2.53781587e-01 7.95094550e-01
5.05881727e-01 3.44425142e-02 9.19095129e-02 9.23189998e-01
-6.92968130e-01 8.96965787e-02 3.22330236e-01 7.81813025e-01
3.80350761e-02 -1.15366828e+00 -3.73447120e-01 7.49961615e-01
-5.64135969e-01 -2.23865613e-01 -7.64078319e-01 1.18632269e+00
3.99700642e-01 7.68049121e-01 1.80649549e-01 -5.79489648e-01
5.13953686e-01 4.16374326e-01 4.92951304e-01 -7.89940596e-01
-7.46922851e-01 -5.49308300e-01 2.96968192e-01 -6.83111966e-01
5.76751381e-02 -5.87117910e-01 -1.07077718e+00 -3.45516771e-01
-1.35158077e-01 2.61849556e-02 4.62943584e-01 9.34537888e-01
8.15550208e-01 4.13162619e-01 4.10343975e-01 -2.79686004e-01
-7.76098430e-01 -7.87025690e-01 -1.29698277e-01 1.11969852e+00
1.88796550e-01 -6.12133861e-01 -4.23831731e-01 1.69172242e-01] | [9.08038330078125, 4.790229320526123] |
3fc5f687-8b6d-40dd-aeae-7b85b7f45ebb | longtonotes-ontonotes-with-longer-coreference | 2210.03650 | null | https://arxiv.org/abs/2210.03650v1 | https://arxiv.org/pdf/2210.03650v1.pdf | Longtonotes: OntoNotes with Longer Coreference Chains | Ontonotes has served as the most important benchmark for coreference resolution. However, for ease of annotation, several long documents in Ontonotes were split into smaller parts. In this work, we build a corpus of coreference-annotated documents of significantly longer length than what is currently available. We do so by providing an accurate, manually-curated, merging of annotations from documents that were split into multiple parts in the original Ontonotes annotation process. The resulting corpus, which we call LongtoNotes contains documents in multiple genres of the English language with varying lengths, the longest of which are up to 8x the length of documents in Ontonotes, and 2x those in Litbank. We evaluate state-of-the-art neural coreference systems on this new corpus, analyze the relationships between model architectures/hyperparameters and document length on performance and efficiency of the models, and demonstrate areas of improvement in long-document coreference modeling revealed by our new corpus. Our data and code is available at: https://github.com/kumar-shridhar/LongtoNotes. | ['Mrinmaya Sachan', 'Andrew McCallum', 'Manzil Zaheer', 'Alessandro Stolfo', 'Raghuveer Thirukovalluru', 'Nicholas Monath', 'Kumar Shridhar'] | 2022-10-07 | null | null | null | null | ['coreference-resolution'] | ['natural-language-processing'] | [-4.97310385e-02 2.46900499e-01 -4.04415578e-01 -4.36077118e-01
-1.44063461e+00 -1.11341250e+00 5.88627338e-01 1.71392977e-01
-6.24665916e-01 8.36947203e-01 1.05567467e+00 -8.30468256e-03
-2.04406291e-01 -5.71675226e-02 -4.57421362e-01 -3.00330997e-01
1.85774252e-01 1.24836361e+00 6.96226284e-02 -3.02243173e-01
2.53235459e-01 3.09767611e-02 -1.15606105e+00 5.61802328e-01
5.94524920e-01 3.69032443e-01 2.93460906e-01 5.39175332e-01
-1.03073284e-01 3.09331119e-01 -6.85125232e-01 -5.47385454e-01
-8.90809170e-04 -1.55499980e-01 -1.44054842e+00 -3.87460887e-01
7.71303833e-01 4.19014208e-02 -5.08726776e-01 8.57604444e-01
6.19275451e-01 1.35318249e-01 1.67438269e-01 -9.62268114e-01
-4.34598386e-01 1.59520960e+00 -4.72458124e-01 3.89755845e-01
3.01477462e-01 -5.84046185e-01 1.49561048e+00 -6.38294101e-01
1.01898503e+00 1.28087842e+00 7.70270348e-01 6.94868147e-01
-9.83848631e-01 -9.60104346e-01 -1.38038725e-01 2.37236395e-01
-1.34747446e+00 -9.51925159e-01 4.29599494e-01 -2.52081662e-01
1.20007193e+00 3.51450026e-01 3.03303242e-01 1.20767057e+00
-1.76068485e-01 8.02952707e-01 4.02438134e-01 -8.54550660e-01
-2.16921180e-01 -9.00530592e-02 6.97320461e-01 1.22024588e-01
1.68599620e-01 -3.08772773e-02 -6.17802262e-01 -2.46647984e-01
2.68433750e-01 -4.76519406e-01 -7.41676211e-01 -3.61791961e-02
-9.78634357e-01 6.85951233e-01 1.46558625e-03 5.83798707e-01
1.00195162e-01 -6.94619864e-02 6.69972599e-01 2.08787963e-01
3.80462766e-01 7.09950089e-01 -7.44385958e-01 -5.27627647e-01
-9.31659102e-01 3.17874014e-01 1.09774077e+00 1.26232612e+00
1.18107587e-01 -5.97229540e-01 6.82477951e-02 1.14837527e+00
1.02685027e-01 1.74758941e-01 6.20280981e-01 -1.44031298e+00
8.47237289e-01 3.49088877e-01 2.13831514e-01 -3.84449869e-01
-6.43719614e-01 -4.46199983e-01 -2.99982935e-01 -4.17970031e-01
6.61270618e-01 -1.61208391e-01 -4.25140142e-01 1.94067752e+00
1.10296449e-02 -8.17424431e-02 2.95534045e-01 6.91289783e-01
9.15611982e-01 5.17009079e-01 1.33393602e-02 -4.01157856e-01
1.77017581e+00 -1.01412392e+00 -7.62253940e-01 -2.90284127e-01
8.85481536e-01 -1.15610278e+00 9.26693678e-01 2.78464884e-01
-1.27667046e+00 -2.96129048e-01 -9.48875010e-01 -2.89634407e-01
-3.85908447e-02 -9.03051943e-02 5.54700911e-01 3.79205406e-01
-8.16395879e-01 4.98907685e-01 -6.97818935e-01 -5.78790784e-01
4.10202108e-02 2.91858971e-01 -3.53753865e-01 1.77433118e-01
-1.39654386e+00 8.79582226e-01 6.27391636e-01 -3.29329342e-01
-2.24768862e-01 -9.01608109e-01 -5.64274311e-01 3.62171441e-01
4.63005394e-01 -3.61793637e-01 2.01013541e+00 -5.07628143e-01
-9.82208669e-01 1.08174467e+00 -1.52211040e-01 -3.64039928e-01
1.38010800e-01 -7.10871875e-01 -5.23998857e-01 2.48944461e-02
2.03558266e-01 5.66910028e-01 -1.01544037e-01 -1.14323878e+00
-8.15292120e-01 -3.84606540e-01 1.30534366e-01 4.36841726e-01
-3.69412541e-01 4.58386064e-01 -1.03195477e+00 -5.39912462e-01
-8.51390809e-02 -1.30788839e+00 2.71859825e-01 -6.73013508e-01
-5.53940237e-01 -3.54798019e-01 4.33984578e-01 -6.37021661e-01
1.46538305e+00 -2.09435463e+00 1.21558458e-01 -2.38003582e-01
-3.39046121e-02 1.58329546e-01 -2.93952316e-01 7.68135846e-01
-1.31789699e-01 8.46014172e-02 -9.65379104e-02 -6.02724552e-01
7.88757950e-02 1.58917472e-01 -4.53672707e-01 2.34820992e-01
-5.62852204e-01 5.38702130e-01 -9.00227308e-01 -6.79807186e-01
-3.78865838e-01 3.74076575e-01 -5.98216295e-01 5.35085611e-02
-2.37418488e-01 3.09837073e-01 -3.70451748e-01 3.16957921e-01
3.32601815e-01 -7.58134797e-02 5.29495835e-01 -3.50967437e-01
-3.10605913e-01 1.02840328e+00 -9.25696135e-01 2.18352246e+00
-3.52429152e-01 8.80856991e-01 1.49000719e-01 -3.63655448e-01
6.70580029e-01 8.21984529e-01 6.26166523e-01 -4.05334502e-01
2.35428706e-01 2.06760824e-01 1.12706684e-01 -3.15727890e-01
1.08485973e+00 -3.37348804e-02 -3.37718368e-01 7.63355672e-01
1.57353014e-01 -1.03398249e-01 6.54844463e-01 3.71136189e-01
1.03841579e+00 -8.34018067e-02 1.50777087e-01 -3.79554093e-01
3.08242410e-01 2.95531303e-01 6.83507025e-01 5.57582319e-01
-5.24194799e-02 7.33560801e-01 3.16286117e-01 -6.27105385e-02
-1.13168526e+00 -6.76461160e-01 -5.67910314e-01 1.28175330e+00
-1.53971314e-01 -9.38512981e-01 -8.22495699e-01 -5.69667399e-01
-1.64354309e-01 8.43918920e-01 -3.17319930e-01 1.29172415e-01
-9.99321103e-01 -6.79814398e-01 1.14944410e+00 6.99753165e-01
-7.90965371e-03 -1.21156108e+00 -3.96599889e-01 1.78526297e-01
-7.71276653e-01 -1.15926123e+00 -1.03732812e+00 2.90586203e-01
-5.82023382e-01 -1.43990564e+00 -4.58249807e-01 -7.70518541e-01
4.25795466e-02 1.02722719e-01 1.60873115e+00 -6.24827668e-02
-8.77272934e-02 1.89368919e-01 -5.51121652e-01 -3.89906228e-01
-4.93938953e-01 5.06413996e-01 3.70665416e-02 -7.57766843e-01
7.65988469e-01 -4.40811545e-01 -3.68804596e-02 7.35615492e-02
-6.75637543e-01 -2.90390551e-01 3.42494667e-01 8.72213304e-01
3.43604028e-01 -3.42842698e-01 3.02293122e-01 -1.21980798e+00
8.11772525e-01 -3.48529279e-01 -4.98013645e-01 2.69811779e-01
-5.36211133e-01 1.54367149e-01 5.10407314e-02 -2.84131289e-01
-1.26607955e+00 -2.99521059e-01 -4.24537957e-02 -2.62597531e-01
-4.57059965e-02 6.25212014e-01 -1.39657930e-01 5.77514708e-01
5.26516855e-01 -3.73980016e-01 -4.25471246e-01 -1.07982814e+00
5.06175578e-01 1.06597948e+00 1.09046423e+00 -7.51329541e-01
2.90243149e-01 9.83075276e-02 -5.76743841e-01 -5.12780368e-01
-1.25269353e+00 -8.51960361e-01 -9.45775747e-01 3.18856388e-01
5.62513411e-01 -8.85657370e-01 -7.10278988e-01 1.45650089e-01
-1.42072105e+00 -2.56632477e-01 -2.34357238e-01 6.05008304e-01
-3.88105571e-01 3.41261446e-01 -9.92876828e-01 -4.28103417e-01
-8.25868011e-01 -7.82279730e-01 8.58083844e-01 2.26074114e-01
-1.03292203e+00 -8.99865389e-01 7.16412783e-01 6.34858251e-01
-1.41781136e-01 -2.42317855e-01 9.49910462e-01 -1.08525276e+00
5.64827695e-02 -2.80164257e-02 2.36230362e-02 -1.30606607e-01
-1.04795791e-01 -1.58464164e-02 -8.19384634e-01 -5.71695149e-01
-3.13554972e-01 -4.39208120e-01 7.61019766e-01 3.55672449e-01
5.23994327e-01 -1.34307027e-01 -6.34164095e-01 6.25127971e-01
1.12618995e+00 3.51049185e-01 3.49720478e-01 8.93540204e-01
3.51225495e-01 6.84913516e-01 7.26046383e-01 4.49920654e-01
4.13909912e-01 9.91838813e-01 2.07547620e-02 4.39307630e-01
-1.93283066e-01 -6.35563359e-02 1.31057560e-01 1.27033079e+00
-1.09260924e-01 -5.10283351e-01 -1.05229950e+00 8.20502520e-01
-1.84242964e+00 -1.10645878e+00 -2.31002212e-01 2.01253366e+00
1.33601189e+00 1.72727451e-01 -4.43564579e-02 -1.04836620e-01
8.58634949e-01 9.94361266e-02 -2.17739388e-01 -2.76699007e-01
-2.48429403e-01 1.71518192e-01 3.76776934e-01 8.71935725e-01
-1.15101743e+00 1.06418228e+00 6.45731211e+00 2.49347463e-01
-6.14861250e-01 7.07757547e-02 -9.60894600e-02 -6.97714806e-01
-1.67666137e-01 2.35425785e-01 -1.37002587e+00 3.31838638e-01
1.26895380e+00 -4.59007204e-01 2.05637470e-01 5.74153423e-01
-1.31983301e-02 2.78574735e-01 -1.37045312e+00 6.84456825e-01
1.53681725e-01 -1.31640470e+00 -2.02551022e-01 -1.47375045e-02
4.88453925e-01 5.76760769e-01 -4.22303021e-01 4.17834938e-01
6.78130686e-01 -6.72569275e-01 8.21902514e-01 1.25754133e-01
7.65767813e-01 -7.97935367e-01 9.56790149e-01 7.87653029e-02
-1.08234346e+00 -1.05122201e-01 -3.62849295e-01 3.07600528e-01
4.75415945e-01 1.34412378e-01 -6.08464718e-01 4.64709103e-01
8.90327513e-01 6.43361390e-01 -2.97678024e-01 1.02042675e+00
-9.09141079e-02 4.58659500e-01 -2.80021936e-01 3.64312828e-01
1.37557700e-01 1.11202694e-01 7.68685162e-01 1.52794528e+00
3.03634167e-01 3.83909523e-01 3.19383256e-02 5.68700612e-01
-5.31581998e-01 -5.89676648e-02 -9.94572043e-02 -1.84178352e-02
1.39248300e+00 1.20236111e+00 -7.15025589e-02 -2.72863299e-01
-4.66196865e-01 5.85895121e-01 4.30848449e-01 1.45132169e-01
-5.66895962e-01 -5.26025295e-01 7.28678644e-01 -3.29049565e-02
1.36469901e-01 8.06324631e-02 9.17702541e-02 -1.08859158e+00
-2.09147438e-01 -1.05445635e+00 1.02631605e+00 -7.92777359e-01
-1.12608635e+00 9.17911112e-01 2.68766969e-01 -1.04376543e+00
-4.77471590e-01 -2.21693948e-01 -5.03930092e-01 8.28121364e-01
-1.08452582e+00 -1.15861607e+00 -2.07406193e-01 3.43586057e-01
7.83497214e-01 -2.39967182e-01 1.15010142e+00 5.78002274e-01
-5.82981467e-01 7.59673297e-01 3.79127234e-01 4.74227369e-01
1.32343602e+00 -1.23776889e+00 5.27837038e-01 4.59108740e-01
5.08608341e-01 7.83409357e-01 9.69321012e-01 -5.57133675e-01
-9.03833330e-01 -5.96643567e-01 1.42613137e+00 -6.02395892e-01
8.40858459e-01 -8.13324004e-02 -9.56555068e-01 1.33257067e+00
6.88942373e-01 -6.60906494e-01 8.84149432e-01 6.93763494e-01
-5.52631319e-01 2.13056594e-01 -6.24476790e-01 4.25562888e-01
1.13913071e+00 -5.09574771e-01 -1.28772855e+00 9.45783257e-02
7.22655177e-01 -7.38095939e-01 -1.13162541e+00 2.61563212e-01
6.37596667e-01 -5.66077590e-01 8.68064106e-01 -7.35343516e-01
3.73464942e-01 1.54722854e-01 -2.66138941e-01 -1.36116648e+00
-5.16077042e-01 -7.11968601e-01 -1.86474085e-01 1.67116535e+00
5.65578103e-01 -9.27280635e-02 7.81056345e-01 5.77797651e-01
-6.78212762e-01 -3.30899566e-01 -9.19599056e-01 -5.98301589e-01
3.93343687e-01 -1.72596753e-01 5.18470287e-01 1.20767617e+00
8.00274253e-01 7.80770242e-01 -1.02153711e-01 -7.65862018e-02
5.88021398e-01 3.10814500e-01 4.73140687e-01 -1.65617740e+00
-3.54803056e-01 -4.78942484e-01 4.30472136e-01 -1.02825010e+00
4.13211286e-01 -1.01847982e+00 1.63244270e-02 -1.17717052e+00
5.46688974e-01 -6.02251351e-01 -2.92460889e-01 7.12648153e-01
2.32372917e-02 8.74749497e-02 3.16589981e-01 7.07377672e-01
-4.69089508e-01 1.72516510e-01 8.45858514e-01 -1.87404633e-01
-3.47180307e-01 -2.37113684e-01 -9.62408900e-01 7.49184191e-01
6.80020869e-01 -7.24175930e-01 -1.22635230e-01 -7.90924251e-01
-2.15279050e-02 2.37263858e-01 -4.14167583e-01 -6.33332372e-01
4.49945360e-01 6.12888485e-02 -5.04082292e-02 -7.85287976e-01
5.60268760e-01 -4.29504454e-01 7.78448656e-02 1.34017646e-01
-6.98912144e-01 1.69918641e-01 5.23376167e-01 1.31847173e-01
-3.52654159e-01 -5.91766596e-01 5.13706028e-01 -2.28651941e-01
-6.16972804e-01 -1.06836744e-01 -7.89551288e-02 6.19952083e-01
3.91235113e-01 3.19646955e-01 -8.29937935e-01 -3.51148397e-01
-6.44803941e-01 2.89123029e-01 4.21144992e-01 8.87103438e-01
4.17383462e-02 -1.09800899e+00 -9.68223155e-01 -3.43538612e-01
1.45241722e-01 -1.47136539e-01 -3.11268624e-02 7.67247856e-01
-1.52756140e-01 1.10498333e+00 -1.94472000e-01 -2.08965436e-01
-1.88728142e+00 2.84419686e-01 4.96038944e-02 -5.07257700e-01
-6.76049888e-01 8.55808735e-01 1.24151772e-02 -6.04203761e-01
5.34260571e-01 2.97830030e-02 -4.79383856e-01 4.11247730e-01
6.21309757e-01 1.98174834e-01 -7.96967186e-03 -8.19113255e-01
-4.24317211e-01 4.00543004e-01 -4.94183391e-01 -3.50077927e-01
1.56974578e+00 -4.00265872e-01 -1.09101549e-01 4.00487602e-01
9.81819272e-01 3.65672290e-01 -8.55069280e-01 -3.12282890e-01
4.82665092e-01 -1.85399160e-01 1.00198798e-01 -8.91507745e-01
-8.57291102e-01 5.85813165e-01 1.20496891e-01 -1.26540691e-01
7.28640258e-01 4.73770797e-01 1.01904035e+00 5.38817883e-01
1.89023912e-01 -1.07861567e+00 -3.20142835e-01 8.91049385e-01
1.01532865e+00 -8.53053868e-01 1.18593663e-01 -2.49980986e-01
-5.20696282e-01 9.93797064e-01 5.24370253e-01 3.43344241e-01
3.64936709e-01 6.36364460e-01 2.80118853e-01 -2.33554035e-01
-9.50613976e-01 6.09674565e-02 1.76598668e-01 2.40445390e-01
1.00322270e+00 -1.17792539e-01 -4.36946273e-01 1.09135866e+00
-6.85071290e-01 -1.65151611e-01 6.94785416e-01 8.47580492e-01
-4.12090302e-01 -1.49051154e+00 -4.32904661e-01 -3.13656665e-02
-8.32054615e-01 -5.24369478e-01 -6.64386153e-01 1.04541433e+00
-1.41702056e-01 8.58287692e-01 4.37139034e-01 -1.71128124e-01
3.66066605e-01 2.45745614e-01 4.88551348e-01 -6.33169889e-01
-6.54562712e-01 2.21184611e-01 9.28421021e-01 -2.66887844e-01
-4.06402647e-01 -9.49970841e-01 -1.44359732e+00 -5.05095780e-01
-5.27987480e-01 7.37019658e-01 3.56194079e-01 9.89120662e-01
2.15519994e-01 3.88027996e-01 4.08499278e-02 -7.38813341e-01
-4.23163623e-01 -1.48258770e+00 -4.67063695e-01 6.13415122e-01
-2.41879504e-02 -4.64026392e-01 -2.66624600e-01 2.72541881e-01] | [9.263245582580566, 9.575939178466797] |
6103ef0d-cd32-47a1-97f4-840598db21bf | multi-margin-based-decorrelation-learning-for | 2005.11945 | null | https://arxiv.org/abs/2005.11945v1 | https://arxiv.org/pdf/2005.11945v1.pdf | Multi-Margin based Decorrelation Learning for Heterogeneous Face Recognition | Heterogeneous face recognition (HFR) refers to matching face images acquired from different domains with wide applications in security scenarios. This paper presents a deep neural network approach namely Multi-Margin based Decorrelation Learning (MMDL) to extract decorrelation representations in a hyperspherical space for cross-domain face images. The proposed framework can be divided into two components: heterogeneous representation network and decorrelation representation learning. First, we employ a large scale of accessible visual face images to train heterogeneous representation network. The decorrelation layer projects the output of the first component into decorrelation latent subspace and obtains decorrelation representation. In addition, we design a multi-margin loss (MML), which consists of quadruplet margin loss (QML) and heterogeneous angular margin loss (HAML), to constrain the proposed framework. Experimental results on two challenging heterogeneous face databases show that our approach achieves superior performance on both verification and recognition tasks, comparing with state-of-the-art methods. | ['Nannan Wang', 'Zhifeng Li', 'Jie Li', 'Bing Cao', 'Xinbo Gao'] | 2020-05-25 | null | null | null | null | ['heterogeneous-face-recognition'] | ['computer-vision'] | [ 1.61703095e-01 -8.83879140e-02 1.49679109e-01 -5.13975561e-01
-6.54380739e-01 -2.69507527e-01 6.93549931e-01 -9.56410289e-01
2.33059675e-02 5.20581484e-01 1.62657782e-01 -1.71301141e-02
-2.93038040e-01 -4.50906515e-01 -4.02521193e-01 -9.26356256e-01
1.77128926e-01 2.31505334e-01 -4.39555049e-01 -7.18487054e-02
-1.80666462e-01 8.93955588e-01 -1.44232678e+00 2.14306742e-01
6.39216959e-01 1.33768427e+00 -2.65613019e-01 1.54829666e-01
3.64789397e-01 7.74294794e-01 -3.43005717e-01 -5.90366960e-01
4.59062070e-01 -4.80986238e-01 -4.39192712e-01 6.28441945e-02
3.69017094e-01 -2.70417742e-02 -5.39251626e-01 1.09374344e+00
9.08317685e-01 1.09179713e-01 8.86910081e-01 -1.43452537e+00
-1.10869527e+00 1.84155136e-01 -8.95092487e-01 1.91498429e-01
-1.20158279e-02 -3.15650523e-01 5.09368241e-01 -1.39453959e+00
6.20021582e-01 1.62594771e+00 7.05510199e-01 5.44343531e-01
-1.00581813e+00 -9.86431777e-01 -3.95353399e-02 3.45767885e-01
-1.65640247e+00 -6.91187143e-01 1.10667431e+00 -5.79419315e-01
5.83029449e-01 2.01055452e-01 -6.87211473e-03 1.04286492e+00
9.70014483e-02 3.32662761e-01 1.36288190e+00 -1.64858088e-01
-2.47597694e-02 1.25930309e-01 -6.11148477e-02 8.64286482e-01
2.69169718e-01 4.15525258e-01 -5.82282305e-01 -6.70117810e-02
5.95393240e-01 -2.51196697e-02 -4.16393757e-01 -6.72280729e-01
-6.52283311e-01 6.79295719e-01 4.40110028e-01 -8.56069475e-03
-2.94648111e-01 -3.03030610e-01 2.99182862e-01 4.32929873e-01
6.72623813e-01 -2.10046217e-01 -2.16495886e-01 6.05848610e-01
-6.42276287e-01 -2.22231448e-01 5.11787415e-01 7.64462531e-01
6.43158257e-01 3.84508401e-01 -4.07149643e-01 1.13633060e+00
6.90275729e-01 7.49347687e-01 5.22440553e-01 -4.89820480e-01
6.28118098e-01 3.97868574e-01 -2.18956426e-01 -1.06958699e+00
-3.15929204e-01 -4.31336522e-01 -1.15892875e+00 4.82651234e-01
-4.91845198e-02 -2.24404201e-01 -7.30965793e-01 1.93894410e+00
5.02087414e-01 4.78320599e-01 5.11830688e-01 9.76763487e-01
1.09556878e+00 3.34105134e-01 -3.85409109e-02 -1.40099049e-01
1.40727425e+00 -9.35656607e-01 -7.45650411e-01 1.34440707e-02
-5.80181461e-03 -8.98668766e-01 6.30927980e-01 2.96524558e-02
-8.19864035e-01 -6.33163273e-01 -1.39536655e+00 4.40149941e-02
-3.86191458e-01 4.58644956e-01 3.29393893e-01 7.16466010e-01
-1.00038135e+00 5.56974947e-01 -3.93225580e-01 -2.84827147e-02
6.20004773e-01 3.89413744e-01 -8.74962926e-01 -2.00358868e-01
-1.18075180e+00 8.33663583e-01 2.33535655e-02 6.05340660e-01
-1.10918427e+00 -7.10241973e-01 -1.15957797e+00 3.37201767e-02
1.31585792e-01 -4.73357022e-01 4.52212721e-01 -9.90207970e-01
-1.64319575e+00 9.85988736e-01 -8.49398784e-03 4.53104414e-02
6.59991562e-01 3.49479285e-03 -8.07919502e-01 -2.77338214e-02
-1.69220611e-01 7.14504123e-02 1.38468182e+00 -1.34093738e+00
6.90629706e-02 -8.52556229e-01 -4.73308742e-01 1.80806518e-01
-3.49973887e-01 2.96705514e-01 -5.73099375e-01 -7.92416811e-01
1.65774047e-01 -1.04569757e+00 3.39268416e-01 4.20390293e-02
-2.99678653e-01 -1.36590257e-01 1.26825082e+00 -1.01710296e+00
7.73946166e-01 -2.14977288e+00 4.11617815e-01 1.59935415e-01
5.29496260e-02 3.40637684e-01 -5.06427050e-01 -5.69484346e-02
-6.61644161e-01 -3.35362941e-01 -2.11676508e-01 -6.42902851e-01
1.05404191e-01 -2.70109177e-01 -3.70248169e-01 9.98573244e-01
5.30430198e-01 8.71355653e-01 -3.76236737e-01 -2.39716604e-01
1.57673046e-01 1.27712071e+00 -2.15212554e-01 4.04532522e-01
4.80011225e-01 6.63828492e-01 -1.24903783e-01 9.50453162e-01
1.29775739e+00 -1.07812293e-01 2.59304076e-01 -4.83126879e-01
9.95665416e-02 -4.29315388e-01 -1.13231289e+00 1.42072058e+00
-6.16329014e-01 5.64995766e-01 1.70758441e-01 -1.01512253e+00
1.11538851e+00 3.33534062e-01 2.91745752e-01 -6.61373377e-01
3.88562500e-01 4.10580039e-02 -2.10452706e-01 -3.98376614e-01
-1.21632433e-02 -1.73437044e-01 4.38502245e-02 3.89141798e-01
3.35968137e-01 4.61748451e-01 -4.41678971e-01 -3.73291582e-01
7.17704713e-01 2.62766212e-01 1.64276198e-01 -4.07767594e-01
1.19511354e+00 -9.14328635e-01 8.37375224e-01 7.04063401e-02
-5.06739914e-01 5.71687818e-01 5.11564910e-01 -5.44641197e-01
-8.67629945e-01 -1.04351127e+00 -3.91466171e-01 8.06683898e-01
2.99033254e-01 2.27637812e-01 -6.01097524e-01 -8.95196915e-01
1.98427483e-01 4.25440520e-02 -1.09582531e+00 -3.63923043e-01
-7.04443634e-01 -9.07521188e-01 4.12527859e-01 6.33539557e-01
9.18659806e-01 -1.15112543e+00 9.03103277e-02 -5.82192898e-01
2.06058294e-01 -1.01106048e+00 -5.48997521e-01 -1.59664720e-01
-3.29103231e-01 -1.17864203e+00 -8.52415442e-01 -8.84374797e-01
8.36311758e-01 2.87709504e-01 8.01826417e-01 -3.08641791e-01
-6.01868272e-01 3.10310900e-01 -2.18328144e-02 -7.75805563e-02
-1.79521274e-02 -3.73404413e-01 2.99458504e-01 5.86939156e-01
4.21696335e-01 -5.02698123e-01 -5.40373921e-01 3.91782761e-01
-7.17320323e-01 -5.03996670e-01 7.45034158e-01 1.10300040e+00
5.39517999e-01 -1.22928500e-01 6.52380168e-01 -7.40908146e-01
3.93378943e-01 -6.29559696e-01 -8.01140785e-01 7.71169305e-01
-5.30129671e-01 2.94916451e-01 5.53725719e-01 -3.89349252e-01
-1.41638434e+00 -1.01628937e-01 2.07616463e-01 -9.25870955e-01
9.62286144e-02 1.24035850e-01 -6.66778743e-01 -4.68820512e-01
4.32267308e-01 1.14176810e-01 -1.90605763e-02 -3.85671228e-01
2.89919347e-01 6.59755111e-01 6.36847377e-01 -4.24490809e-01
1.10355639e+00 5.31511664e-01 2.66839266e-01 -7.05652833e-01
-5.99930346e-01 -1.77839130e-01 -4.77636933e-01 -2.61262804e-01
7.12909043e-01 -1.16862583e+00 -1.05725741e+00 5.23930907e-01
-8.22444260e-01 1.53122574e-01 1.09455399e-01 4.88738090e-01
-3.85044903e-01 3.78990650e-01 -2.91877985e-01 -8.08837593e-01
-5.74180186e-01 -1.10960865e+00 1.12768710e+00 4.46108431e-01
6.10330522e-01 -7.91210532e-01 2.06164330e-01 5.32599688e-01
4.52856630e-01 4.56731498e-01 6.11944497e-01 -4.41509277e-01
-3.62301022e-01 -2.33086988e-01 -5.58359146e-01 5.87380052e-01
1.75191388e-01 -3.45365793e-01 -1.35830677e+00 -6.91893160e-01
7.74138421e-02 -4.56962734e-01 9.93343294e-01 5.83310239e-02
1.23391747e+00 -1.84926301e-01 -1.30974546e-01 1.08616221e+00
1.49289060e+00 1.81412756e-01 5.67719340e-01 5.73105589e-02
7.04770923e-01 8.11358154e-01 3.59961659e-01 4.57725793e-01
1.77949890e-01 8.55796874e-01 2.48083130e-01 -7.45480135e-02
-3.87047470e-01 6.94661355e-03 3.93691927e-01 5.63121796e-01
-4.69811447e-02 7.34185725e-02 -7.41919518e-01 7.08075762e-02
-1.79261494e+00 -9.41027164e-01 7.50981688e-01 2.14435315e+00
1.86214268e-01 -4.42918122e-01 -2.85341024e-01 -1.76183805e-01
1.13742197e+00 2.69418716e-01 -9.43121493e-01 -8.45672265e-02
-4.72425729e-01 1.58097640e-01 1.12266108e-01 5.04129529e-01
-1.32981062e+00 8.43232989e-01 4.95271921e+00 6.99866116e-01
-1.23936677e+00 3.41337591e-01 7.13246584e-01 1.46683559e-01
-2.44282350e-01 -3.45960170e-01 -7.09725916e-01 3.79478723e-01
6.56509638e-01 2.50165798e-02 5.93850851e-01 1.01084936e+00
-3.97297829e-01 7.04678893e-01 -8.30291569e-01 1.48353565e+00
7.44543731e-01 -9.43474948e-01 3.20481230e-03 -7.39675686e-02
7.81424284e-01 -2.05399215e-01 6.91444814e-01 4.23065960e-01
1.60703152e-01 -1.15395379e+00 4.87817973e-01 7.58556724e-01
1.02987421e+00 -9.88394320e-01 7.34650671e-01 -3.00812513e-01
-1.50287950e+00 -1.62332386e-01 -6.27897978e-01 5.85454166e-01
-3.43002856e-01 1.50862530e-01 -4.32946980e-01 8.60298574e-01
6.63917124e-01 7.20306456e-01 -5.32201290e-01 5.80255508e-01
-1.21212222e-01 -1.00006685e-01 1.22120447e-01 4.86851275e-01
-3.19810897e-01 -6.06025636e-01 3.78211826e-01 7.70951033e-01
2.56800741e-01 -2.28017047e-02 -1.91733584e-01 7.49853909e-01
-5.17935336e-01 3.19068916e-02 -6.33334696e-01 1.74170211e-01
5.36293864e-01 1.67399383e+00 -2.47329623e-01 1.90010760e-02
-4.60670829e-01 1.23728180e+00 6.63107872e-01 5.06693780e-01
-9.48442280e-01 -4.81504917e-01 9.09320891e-01 -5.20457089e-01
4.07055229e-01 1.71723172e-01 1.48608359e-02 -1.46855223e+00
2.34536693e-01 -6.83700502e-01 3.60632449e-01 -5.51683128e-01
-1.43196476e+00 1.33092332e+00 -2.87526816e-01 -1.38154411e+00
-1.74721368e-02 -7.85513401e-01 -4.31989729e-01 1.16760635e+00
-1.99124980e+00 -1.76789880e+00 -6.20729744e-01 7.68420756e-01
1.54508546e-01 -8.41285527e-01 6.79379046e-01 5.86890340e-01
-1.10494697e+00 1.07864547e+00 2.57396549e-01 3.49752903e-01
7.62797892e-01 -8.11383843e-01 8.68948177e-02 9.44422543e-01
2.47461703e-02 6.12481117e-01 1.79715902e-01 -3.71881545e-01
-1.45859182e+00 -1.47183740e+00 5.67709923e-01 -2.66661704e-01
3.31743747e-01 -3.86558354e-01 -6.36575162e-01 5.48741281e-01
1.10952727e-01 6.95858002e-01 9.71151650e-01 -1.17160283e-01
-1.04057384e+00 -5.19138217e-01 -1.36635816e+00 3.42246324e-01
9.33452606e-01 -9.63527441e-01 -3.47867787e-01 2.64137238e-01
4.27162528e-01 -1.49784848e-01 -9.85008299e-01 8.59127700e-01
8.98447394e-01 -7.88650155e-01 1.17707324e+00 -7.95230448e-01
1.43255219e-01 -4.54783380e-01 -4.39973235e-01 -1.00738418e+00
-3.71184886e-01 -4.37482893e-01 -3.38662148e-01 1.43087924e+00
1.08192995e-01 -7.95727491e-01 7.60965168e-01 3.98486495e-01
-9.38020721e-02 -7.21307814e-01 -1.13070071e+00 -8.45741212e-01
4.21993770e-02 1.17769621e-01 9.27946448e-01 1.03675532e+00
-4.45579529e-01 3.07643205e-01 -6.00366294e-01 6.03360176e-01
1.12363636e+00 4.36228514e-01 4.16207165e-01 -1.26794314e+00
-7.25590810e-02 -1.77949399e-01 -5.73142827e-01 -3.58725041e-01
9.54248548e-01 -9.56522822e-01 -1.65468171e-01 -7.88737833e-01
4.43055987e-01 -3.45992535e-01 -5.95826209e-01 4.27977592e-01
-1.26824483e-01 6.33648694e-01 1.18654467e-01 1.77731737e-01
-3.46947223e-01 1.07751811e+00 8.82718921e-01 -3.21395308e-01
2.87180897e-02 -3.11794251e-01 -5.99497974e-01 6.24787331e-01
6.55743897e-01 -1.75469652e-01 -4.41240817e-01 -2.94496804e-01
-3.74901384e-01 1.37535855e-01 2.82618642e-01 -1.10825300e+00
5.47938384e-02 8.02990422e-02 9.36082304e-01 -4.19477642e-01
5.78824639e-01 -8.76974821e-01 4.18164164e-01 3.10466319e-01
-2.47328267e-01 -7.04980716e-02 1.37886807e-01 5.39078474e-01
-3.26623142e-01 3.30980718e-01 1.32580554e+00 3.18222433e-01
-5.64474523e-01 6.59537971e-01 4.23624098e-01 -3.93913016e-02
1.24430215e+00 1.40213534e-01 -6.44914627e-01 1.48367226e-01
-6.94705546e-01 -1.63179964e-01 1.67898342e-01 8.15866947e-01
8.40356648e-01 -1.82332122e+00 -8.45473886e-01 8.24961245e-01
2.53199190e-01 -5.90823054e-01 7.08037555e-01 6.07006490e-01
-4.28826630e-01 6.33313894e-01 -4.33336437e-01 -3.57606769e-01
-1.42631042e+00 7.95434177e-01 5.16454875e-01 -1.69867367e-01
-4.35713708e-01 1.09109545e+00 4.10676301e-01 -6.45170450e-01
1.14567891e-01 4.58898813e-01 -4.43018168e-01 1.48739487e-01
7.15844572e-01 3.96586567e-01 2.32490093e-01 -1.35980415e+00
-7.11340129e-01 9.48710859e-01 5.02495049e-03 4.25325930e-02
1.24892664e+00 -4.84089553e-02 -2.66109556e-01 -1.35297194e-01
1.76743674e+00 -3.06129426e-01 -1.39906681e+00 -4.58898425e-01
-2.05933273e-01 -5.14567792e-01 2.77848430e-02 -5.66862226e-01
-1.37841702e+00 8.30095708e-01 1.12125981e+00 -4.05345887e-01
1.36717558e+00 -4.23986092e-02 2.99327016e-01 2.64727682e-01
1.57034218e-01 -8.65642548e-01 3.72079283e-01 2.96056151e-01
1.39228809e+00 -1.49466765e+00 -6.28498569e-02 -3.29072297e-01
-7.55400836e-01 9.10470545e-01 8.83086979e-01 -1.79993406e-01
1.18189394e+00 3.49094416e-03 1.53266743e-01 -8.15505236e-02
-5.34958482e-01 -1.07608818e-01 5.88885546e-01 5.39778888e-01
1.49608672e-01 -1.24664210e-01 -5.96393533e-02 7.86272526e-01
-2.99035907e-02 -3.83150876e-01 -1.21497683e-01 5.81203818e-01
6.70988709e-02 -9.66203272e-01 -5.08506179e-01 3.89568061e-02
-3.21263313e-01 1.67227417e-01 -3.59026074e-01 5.55153072e-01
1.53031647e-01 8.89882445e-01 -2.05103680e-02 -6.85170591e-01
2.77470589e-01 9.19847116e-02 5.54436803e-01 -1.82114869e-01
-1.60585672e-01 -1.90767273e-01 -4.16419566e-01 -5.95494390e-01
-3.63001168e-01 -4.66360033e-01 -6.81530535e-01 -1.69483706e-01
-3.35770667e-01 1.89673141e-01 6.18449152e-01 6.95353687e-01
4.98249263e-01 5.08656323e-01 1.24334097e+00 -8.37765634e-01
-7.11916685e-01 -1.09968424e+00 -8.79873037e-01 4.68154877e-01
5.44843733e-01 -1.02819014e+00 -4.28141087e-01 -1.00686327e-01] | [13.158742904663086, 0.48823562264442444] |
62fd8b05-d6a9-4903-a9ac-85c033ddb8e3 | detecting-twenty-thousand-classes-using-image | 2201.02605 | null | https://arxiv.org/abs/2201.02605v3 | https://arxiv.org/pdf/2201.02605v3.pdf | Detecting Twenty-thousand Classes using Image-level Supervision | Current object detectors are limited in vocabulary size due to the small scale of detection datasets. Image classifiers, on the other hand, reason about much larger vocabularies, as their datasets are larger and easier to collect. We propose Detic, which simply trains the classifiers of a detector on image classification data and thus expands the vocabulary of detectors to tens of thousands of concepts. Unlike prior work, Detic does not need complex assignment schemes to assign image labels to boxes based on model predictions, making it much easier to implement and compatible with a range of detection architectures and backbones. Our results show that Detic yields excellent detectors even for classes without box annotations. It outperforms prior work on both open-vocabulary and long-tail detection benchmarks. Detic provides a gain of 2.4 mAP for all classes and 8.3 mAP for novel classes on the open-vocabulary LVIS benchmark. On the standard LVIS benchmark, Detic obtains 41.7 mAP when evaluated on all classes, or only rare classes, hence closing the gap in performance for object categories with few samples. For the first time, we train a detector with all the twenty-one-thousand classes of the ImageNet dataset and show that it generalizes to new datasets without finetuning. Code is available at \url{https://github.com/facebookresearch/Detic}. | ['Philipp Krähenbühl', 'Rohit Girdhar', 'Ishan Misra', 'Armand Joulin', 'Xingyi Zhou'] | 2022-01-07 | null | null | null | null | ['open-vocabulary-object-detection'] | ['computer-vision'] | [-9.30524766e-02 -5.46067320e-02 -4.15565163e-01 -3.81222934e-01
-6.96485162e-01 -8.95762503e-01 6.44062519e-01 1.40101030e-01
-6.49698079e-01 3.95884603e-01 -7.75256008e-02 -1.60930738e-01
4.16340292e-01 -6.80663884e-01 -6.31044030e-01 -3.37824643e-01
5.30846976e-02 4.89189476e-01 8.83368492e-01 -3.72410305e-02
-2.03830585e-01 3.00015718e-01 -1.70104647e+00 4.00242656e-01
2.35538512e-01 1.02869141e+00 9.44910496e-02 5.90578377e-01
1.77938547e-02 3.41395617e-01 -4.10295606e-01 -4.31408912e-01
5.46816587e-01 1.88011497e-01 -5.26420653e-01 -1.51184037e-01
1.24573779e+00 -5.01146257e-01 -5.76424122e-01 1.15391874e+00
4.93586481e-01 -3.39894146e-01 8.20096433e-01 -1.22566581e+00
-7.61416137e-01 6.05520129e-01 -6.76779866e-01 6.83779359e-01
-1.66974217e-01 2.20445350e-01 1.54503500e+00 -1.22027230e+00
7.47700870e-01 1.11782360e+00 6.79600120e-01 8.00832212e-01
-1.42812598e+00 -1.19054830e+00 1.65920675e-01 2.31317133e-02
-1.67787647e+00 -4.00251627e-01 6.52584666e-03 -7.78188050e-01
1.04258144e+00 1.17464602e-01 3.73666078e-01 1.00209236e+00
-6.19191229e-02 9.63494062e-01 8.24781120e-01 -1.44699350e-01
2.23889221e-02 3.79214108e-01 6.32325768e-01 6.29184425e-01
7.56400108e-01 1.39112815e-01 -1.77194878e-01 -1.21185958e-01
5.23807108e-01 1.91892684e-01 1.12441823e-01 -5.05463958e-01
-1.17206120e+00 1.04349172e+00 7.26220608e-01 1.53397232e-01
1.52749896e-01 3.58714968e-01 7.03931808e-01 1.74110234e-01
5.28207481e-01 5.01989365e-01 -6.43601060e-01 4.15806055e-01
-8.27363610e-01 3.32661480e-01 6.77524984e-01 1.35223281e+00
7.36166000e-01 -1.67511523e-01 -1.33300468e-01 6.98760331e-01
1.08172670e-01 7.27501869e-01 4.56967860e-01 -5.75289309e-01
3.62572163e-01 6.05145037e-01 -1.46843061e-01 -6.20590508e-01
-6.06754422e-01 -6.51209652e-01 -5.64332128e-01 4.98748943e-02
4.97680128e-01 -2.56221322e-03 -1.18572795e+00 1.57375705e+00
2.26340026e-01 5.96474968e-02 -3.13477337e-01 6.62525952e-01
1.25161624e+00 5.10203660e-01 2.17825711e-01 2.40347877e-01
1.78912628e+00 -8.65092337e-01 -1.88452467e-01 -5.21125853e-01
9.32704926e-01 -5.21202683e-01 1.05880940e+00 2.66298383e-01
-4.74358112e-01 -5.24361670e-01 -1.12351441e+00 -1.59826338e-01
-6.62705123e-01 2.20312536e-01 6.84013426e-01 7.06236720e-01
-1.08254218e+00 -8.18099156e-02 -6.60646379e-01 -6.57862604e-01
8.49267781e-01 3.23825598e-01 -3.03645551e-01 -1.53131653e-02
-9.13072646e-01 8.32947969e-01 5.99840760e-01 -5.81617475e-01
-9.08632696e-01 -9.17508006e-01 -7.21221685e-01 9.61646363e-02
4.31424648e-01 -2.71739364e-01 1.42389929e+00 -8.05191576e-01
-5.65553546e-01 1.16565847e+00 1.66840658e-01 -7.52695084e-01
2.70513356e-01 -2.93185823e-02 -3.60409349e-01 -1.23601258e-02
3.28536361e-01 1.49787116e+00 7.06578493e-01 -7.34478056e-01
-9.40261304e-01 -2.75746733e-01 1.53102770e-01 -2.33404756e-01
-6.52812183e-01 1.22497134e-01 -5.24706185e-01 -4.75425124e-01
-1.70260310e-01 -1.04476821e+00 1.28905680e-02 2.36101463e-01
-1.44659638e-01 -5.62158406e-01 9.94230628e-01 -1.83600038e-01
1.12945342e+00 -2.21534014e+00 -3.20719898e-01 -4.68185879e-02
7.27231324e-01 4.76322591e-01 -2.94868737e-01 -1.09183704e-02
1.58266053e-02 2.36670345e-01 1.16491809e-01 -1.22632965e-01
-9.69760213e-03 2.37513542e-01 -4.72793579e-01 5.18823624e-01
2.09216416e-01 9.37454581e-01 -7.06035674e-01 -5.15283406e-01
-1.54789733e-02 1.79412156e-01 -8.45222652e-01 -3.78684133e-01
-4.07761931e-01 -7.30683133e-02 -4.26105291e-01 5.96379220e-01
6.37163937e-01 -6.48579657e-01 -4.42133136e-02 -7.05540553e-02
-3.55295613e-02 2.00855836e-01 -1.02117062e+00 1.39170873e+00
-2.18360811e-01 9.26837504e-01 -1.96524262e-01 -8.29037428e-01
8.57505620e-01 1.07819736e-01 2.24402651e-01 -5.12396038e-01
2.62412637e-01 2.08711699e-01 2.31219500e-01 -2.68118884e-02
2.84290671e-01 1.04499608e-01 -2.69932777e-01 2.33410105e-01
4.06695157e-01 -1.43539831e-01 4.73808229e-01 5.32155514e-01
1.30306971e+00 -5.16118884e-01 4.67374504e-01 -5.08982182e-01
2.55307972e-01 4.64185588e-02 5.17085373e-01 1.01779461e+00
-4.78003621e-01 4.90055323e-01 3.57438833e-01 -7.70253778e-01
-1.23266029e+00 -1.21175206e+00 -6.86433792e-01 1.52659965e+00
7.14248940e-02 -6.85828567e-01 -6.16903484e-01 -8.33322763e-01
4.43618596e-01 4.79265392e-01 -5.86210430e-01 1.19367503e-01
-2.20580995e-01 -7.49048650e-01 8.42179596e-01 8.54932249e-01
3.49790782e-01 -8.86870265e-01 -4.83727932e-01 -9.45538729e-02
1.20022818e-01 -1.36870873e+00 -5.51050425e-01 2.64290750e-01
-6.91501856e-01 -1.11221802e+00 -4.73810762e-01 -9.26797569e-01
4.94330436e-01 4.39189613e-01 1.25579441e+00 2.60568678e-01
-9.01859939e-01 2.67895758e-01 -1.40233129e-01 -8.71096253e-01
-2.63633966e-01 3.74940753e-01 2.39748672e-01 -4.30155158e-01
9.28491473e-01 -1.27734259e-01 -5.87693930e-01 5.17570972e-01
-6.35588288e-01 -1.81900069e-01 5.64591467e-01 8.54215026e-01
4.23571348e-01 -3.65864664e-01 6.71638846e-01 -8.27130735e-01
9.45017114e-02 -4.65524912e-01 -1.18918753e+00 6.23572916e-02
-6.04665697e-01 -5.16053550e-02 1.92825273e-01 -5.43323755e-01
-5.92607796e-01 3.08675975e-01 5.31028025e-02 -2.69879282e-01
-1.46681637e-01 -1.27049848e-01 1.89934388e-01 -7.11629987e-02
1.31192994e+00 -2.32597813e-02 -2.80812323e-01 -4.47894245e-01
5.85444689e-01 7.15549052e-01 3.41455817e-01 -2.45792583e-01
8.51389825e-01 6.96912467e-01 -3.58545244e-01 -7.95938015e-01
-1.08156443e+00 -1.01543140e+00 -8.00025225e-01 9.43351611e-02
9.33457315e-01 -1.27907908e+00 -5.88113308e-01 1.11872233e-01
-9.75132704e-01 -1.99860007e-01 -1.40290931e-01 3.64079386e-01
-1.36806890e-01 1.08995065e-01 -7.06010640e-01 -4.25392032e-01
-4.38307136e-01 -1.09085870e+00 1.10586762e+00 2.73959301e-02
-1.99652851e-01 -5.97379506e-01 -1.47928312e-01 2.39260972e-01
2.11209953e-01 -2.58349895e-01 4.23981875e-01 -9.68823791e-01
-7.83817828e-01 -5.51630080e-01 -7.75575161e-01 5.46188891e-01
-3.02190065e-01 -1.57254919e-01 -1.17222118e+00 -5.43699622e-01
-6.36752367e-01 -6.92903340e-01 1.47525084e+00 1.66963756e-01
1.17965937e+00 -3.23934965e-02 -7.86127150e-01 4.52412486e-01
1.33316827e+00 -1.70104921e-01 3.21042508e-01 2.74327278e-01
5.61766207e-01 1.14464045e-01 4.28456962e-01 3.38463038e-01
1.71713203e-01 8.21920574e-01 3.59522015e-01 -2.13388465e-02
-3.53587359e-01 -1.78618401e-01 2.38250867e-01 3.00023675e-01
3.16204667e-01 -2.03959346e-01 -1.28808260e+00 6.87345445e-01
-1.54918206e+00 -9.44798231e-01 -1.19820736e-01 1.90970838e+00
8.44988823e-01 4.46364462e-01 2.73476869e-01 -3.28136116e-01
6.55202091e-01 -3.68177220e-02 -6.46383524e-01 -4.51072901e-02
-6.01766557e-02 1.96655855e-01 1.13649178e+00 3.00305814e-01
-1.71666050e+00 1.21883357e+00 6.57636118e+00 8.99357259e-01
-9.30908561e-01 5.09217560e-01 5.53102791e-01 -5.21956563e-01
2.34810919e-01 -3.74080874e-02 -1.64201331e+00 1.95923582e-01
7.39435554e-01 -2.24737637e-03 1.80356614e-02 1.27287567e+00
-3.44962418e-01 1.51900286e-02 -1.27949357e+00 1.05193758e+00
6.57739565e-02 -1.57791579e+00 -5.31863719e-02 1.48559883e-01
8.75516951e-01 9.68695641e-01 3.00318636e-02 6.54423177e-01
4.89286840e-01 -9.22233880e-01 8.08340192e-01 -1.74400553e-01
1.00533843e+00 -3.55300128e-01 6.30530834e-01 5.28082013e-01
-1.24999905e+00 -3.09889495e-01 -8.31770837e-01 -1.46299601e-01
-3.90297264e-01 3.13087583e-01 -1.10718369e+00 -4.38437462e-01
1.08677459e+00 6.83085024e-01 -1.02787590e+00 1.09136987e+00
-1.47360295e-01 8.65565777e-01 -6.52858078e-01 -2.86670655e-01
3.02045912e-01 3.96695733e-01 3.11590284e-01 1.58620906e+00
-8.46973583e-02 2.09979266e-02 4.94424611e-01 7.22998917e-01
-4.83266920e-01 1.30500659e-01 -8.12039256e-01 2.35018611e-01
6.77159548e-01 1.54029047e+00 -8.19054663e-01 -6.55297577e-01
-7.29128599e-01 4.31867182e-01 5.39612889e-01 1.33742899e-01
-8.90326977e-01 -2.56818026e-01 7.26598978e-01 2.97984570e-01
6.63328111e-01 -1.15078494e-01 -3.46634418e-01 -1.23614383e+00
-1.02375358e-01 -7.63812602e-01 6.09932959e-01 -4.55400825e-01
-1.33789337e+00 3.81756067e-01 1.08403243e-01 -1.01328862e+00
1.63908258e-01 -1.06716609e+00 -2.99606055e-01 3.84078443e-01
-1.32796252e+00 -1.18777263e+00 -2.14820668e-01 3.55635762e-01
5.71050644e-01 -2.97337323e-01 6.88651860e-01 5.61192036e-01
-4.36626703e-01 8.45792294e-01 2.28054658e-01 6.51297748e-01
9.16777790e-01 -1.12079012e+00 7.49377131e-01 6.53911531e-01
5.04668593e-01 5.36155522e-01 5.19275665e-01 -5.27768910e-01
-9.16620195e-01 -1.36694086e+00 6.98171735e-01 -8.84158432e-01
1.01043046e+00 -9.92322981e-01 -7.85409749e-01 1.08837438e+00
-7.21527189e-02 5.70322156e-01 4.47150916e-01 3.45164478e-01
-1.12679100e+00 -1.62612960e-01 -9.28632557e-01 3.31494153e-01
1.27700794e+00 -5.04486203e-01 -4.31392431e-01 6.82169199e-01
7.68322706e-01 -2.86369652e-01 -7.24912822e-01 3.84473294e-01
5.62845170e-01 -5.86337984e-01 1.13722730e+00 -6.37951970e-01
1.34364190e-02 -3.15811098e-01 -2.76870698e-01 -7.84254909e-01
-6.12933993e-01 6.87648132e-02 1.00283973e-01 9.04007077e-01
8.26267660e-01 -8.87341619e-01 7.48248994e-01 3.01668167e-01
5.65329660e-03 -4.35792595e-01 -8.13434720e-01 -1.14805758e+00
3.27116340e-01 -5.44588864e-01 2.88158059e-01 6.91693962e-01
-1.85211957e-01 6.12029612e-01 -1.49896577e-01 1.40263915e-01
6.59887314e-01 -7.29615986e-02 9.48840261e-01 -1.51741195e+00
-2.96889573e-01 -5.16899288e-01 -8.63481343e-01 -1.20082343e+00
4.80454341e-02 -1.34465456e+00 3.27303866e-03 -1.11652493e+00
6.92964196e-01 -6.88477576e-01 -1.90622672e-01 8.58548582e-01
-3.63068916e-02 8.83809149e-01 3.64651054e-01 4.37205702e-01
-9.17675734e-01 1.80960074e-01 8.04869294e-01 -4.54798102e-01
1.42596483e-01 -1.85622394e-01 -7.43183851e-01 9.40541446e-01
9.18050110e-01 -5.54974258e-01 -3.85996222e-01 -3.98549199e-01
1.92228064e-01 -7.76498497e-01 6.25634670e-01 -1.19741917e+00
1.71471834e-01 3.12507510e-01 4.80949283e-01 -8.55173528e-01
1.62372649e-01 -4.71827954e-01 -3.63875866e-01 5.14953911e-01
-3.78100127e-01 -1.74372807e-01 5.11675179e-01 7.95975506e-01
1.36077791e-01 -8.89142379e-02 1.08086991e+00 9.04820766e-03
-1.14678621e+00 4.93048638e-01 -3.28880787e-01 4.16315854e-01
1.18585205e+00 -3.19535006e-03 -6.59097075e-01 3.25362608e-02
-6.63516939e-01 3.46296787e-01 2.96774566e-01 7.30526507e-01
3.10752720e-01 -1.18060029e+00 -7.51760662e-01 1.74759567e-01
8.17216218e-01 -5.90652227e-02 -1.29649177e-01 6.05515242e-01
-4.88601744e-01 6.63201451e-01 6.81990981e-02 -1.06381845e+00
-1.48305070e+00 8.72310162e-01 4.24208552e-01 -8.40823203e-02
-7.80194938e-01 1.04294705e+00 9.23054457e-01 -4.74292517e-01
3.22098345e-01 -5.84665954e-01 -1.60928741e-02 1.40200943e-01
8.18742931e-01 5.55998161e-02 7.47909769e-02 -5.22574782e-01
-4.79893088e-01 3.86743158e-01 -4.96204287e-01 3.24598610e-01
1.06479919e+00 1.91700399e-01 1.79771245e-01 1.27917200e-01
1.19563496e+00 -2.98378110e-01 -1.13815939e+00 -6.57064557e-01
1.96838826e-01 -2.10869521e-01 9.33904648e-02 -6.46823943e-01
-8.81439149e-01 8.87406170e-01 8.15411031e-01 -3.67038660e-02
6.84045076e-01 6.35613143e-01 5.98953605e-01 6.83477819e-01
3.68859619e-01 -1.06118929e+00 1.32205695e-01 6.98106527e-01
7.42169738e-01 -1.48227656e+00 1.62086353e-01 -3.81173342e-01
-4.10880119e-01 8.33406389e-01 8.60257089e-01 -1.89096630e-01
9.07701194e-01 4.42025363e-01 2.75876541e-02 -3.81828308e-01
-8.87062073e-01 -3.71740967e-01 4.61308658e-01 4.57906961e-01
3.98152769e-01 3.87189180e-01 -8.67073238e-02 3.28529984e-01
-6.49323612e-02 -1.23948053e-01 3.00423235e-01 5.60587585e-01
-8.95800889e-01 -7.77711272e-01 -2.51543999e-01 7.83059359e-01
-2.37407580e-01 -3.95653307e-01 -2.80001879e-01 9.81659174e-01
3.38636875e-01 7.05981374e-01 3.63406032e-01 -1.83319747e-01
2.41886497e-01 8.95242766e-02 2.01804295e-01 -1.12565792e+00
-2.55094469e-01 -1.73195302e-01 -2.36724466e-02 -4.84650642e-01
-7.09915981e-02 -7.36007333e-01 -1.17149997e+00 -2.72393137e-01
-6.52330458e-01 -3.22908968e-01 5.18937409e-01 6.09268606e-01
4.33333337e-01 1.83991849e-01 8.17681942e-03 -6.49816990e-01
-8.19183707e-01 -1.02494359e+00 -4.98802453e-01 3.57840240e-01
1.85000837e-01 -8.31534505e-01 -7.30432495e-02 -4.67821360e-02] | [9.460306167602539, 1.4285650253295898] |
5641945c-eff9-41e0-95d0-2728c1ce4679 | compressed-sensing-mri-reconstruction | 2210.14586 | null | https://arxiv.org/abs/2210.14586v2 | https://arxiv.org/pdf/2210.14586v2.pdf | Compressed Sensing MRI Reconstruction Regularized by VAEs with Structured Image Covariance | Objective: This paper investigates how generative models, trained on ground-truth images, can be used \changes{as} priors for inverse problems, penalizing reconstructions far from images the generator can produce. The aim is that learned regularization will provide complex data-driven priors to inverse problems while still retaining the control and insight of a variational regularization method. Moreover, unsupervised learning, without paired training data, allows the learned regularizer to remain flexible to changes in the forward problem such as noise level, sampling pattern or coil sensitivities in MRI. Approach: We utilize variational autoencoders (VAEs) that generate not only an image but also a covariance uncertainty matrix for each image. The covariance can model changing uncertainty dependencies caused by structure in the image, such as edges or objects, and provides a new distance metric from the manifold of learned images. Main results: We evaluate these novel generative regularizers on retrospectively sub-sampled real-valued MRI measurements from the fastMRI dataset. We compare our proposed learned regularization against other unlearned regularization approaches and unsupervised and supervised deep learning methods. Significance: Our results show that the proposed method is competitive with other state-of-the-art methods and behaves consistently with changing sampling patterns and noise levels. | ['Neill D. F. Campbell', 'Matthias J. Ehrhardt', 'Ivor J. A. Simpson', 'Margaret Duff'] | 2022-10-26 | null | null | null | null | ['mri-reconstruction'] | ['computer-vision'] | [ 3.18360537e-01 6.04198456e-01 1.26066923e-01 -5.09013891e-01
-8.10299993e-01 -2.98821837e-01 4.93567258e-01 -3.30181956e-01
-5.15093744e-01 1.00265372e+00 4.51596528e-01 3.45131606e-01
-5.76952755e-01 -4.63788867e-01 -1.02671587e+00 -1.18204105e+00
-3.31716090e-02 8.30584347e-01 1.24262668e-01 3.10529377e-02
-9.22350436e-02 4.66402739e-01 -1.21754730e+00 2.20903799e-01
7.48705983e-01 6.31195188e-01 5.77642381e-01 2.66932040e-01
4.08393562e-01 7.42047429e-01 -3.34963411e-01 9.90014449e-02
1.67744592e-01 -3.65913033e-01 -5.99058688e-01 3.69900644e-01
5.60419448e-02 -1.56676665e-01 -4.55019295e-01 1.18682086e+00
5.78334212e-01 3.66240174e-01 1.23342729e+00 -5.57061970e-01
-9.20285165e-01 8.42536330e-01 -4.12580550e-01 2.79004127e-01
-1.85565114e-01 2.09268898e-01 4.76620555e-01 -9.04739738e-01
1.04918635e+00 1.01885462e+00 6.14985347e-01 7.46665478e-01
-1.91758525e+00 -9.80857089e-02 -5.26514202e-02 6.86642304e-02
-1.17286384e+00 -5.37057817e-01 9.28339899e-01 -7.77575433e-01
4.91635531e-01 1.21101342e-01 3.39149296e-01 1.42078710e+00
6.00694835e-01 3.52799058e-01 1.37635124e+00 -3.19856405e-01
5.27864754e-01 4.89982367e-01 -1.04717083e-01 5.22919238e-01
1.85942173e-01 2.06127197e-01 -3.64064686e-02 -4.88898046e-02
9.82148707e-01 -2.14130767e-02 -7.83754289e-01 -6.68143392e-01
-1.13871610e+00 1.08251202e+00 4.94355738e-01 5.38127244e-01
-7.49004006e-01 2.53659457e-01 4.47385386e-02 -1.40360922e-01
5.43875635e-01 4.20590699e-01 -1.70687526e-01 3.63833010e-01
-1.01103795e+00 -2.34448254e-01 5.15150368e-01 7.38461614e-01
7.55600214e-01 4.44453150e-01 -4.06591773e-01 8.37899745e-01
7.62307823e-01 5.44600546e-01 6.82642400e-01 -1.14906549e+00
-1.59278497e-01 1.31415883e-02 -2.10962996e-01 -8.53141904e-01
-5.33288658e-01 -6.89713359e-01 -1.04994571e+00 2.99814045e-01
1.63794026e-01 -2.86737025e-01 -1.23181403e+00 1.97374713e+00
2.06561923e-01 2.27259904e-01 -2.37326443e-01 1.10827661e+00
8.89383316e-01 3.90089303e-01 -1.47969678e-01 -7.22356617e-01
1.15114939e+00 -3.98315847e-01 -1.07647562e+00 -1.98136732e-01
2.52032101e-01 -6.51869237e-01 7.42541015e-01 3.56747478e-01
-1.22844672e+00 -4.62830991e-01 -1.02742016e+00 3.28888804e-01
6.63574114e-02 -3.19371298e-02 2.66901314e-01 6.87276602e-01
-1.16683185e+00 8.09960544e-01 -1.28351891e+00 7.89601430e-02
3.50749731e-01 3.81064206e-01 -2.05649748e-01 7.69838020e-02
-1.19515300e+00 1.00417829e+00 2.91301578e-01 2.55953699e-01
-1.30639482e+00 -9.60225284e-01 -8.61470997e-01 -2.15955704e-01
1.51455045e-01 -7.87098229e-01 6.36147559e-01 -8.97495925e-01
-1.64894283e+00 7.83671498e-01 1.56780556e-01 -4.64101672e-01
5.67072093e-01 1.24829533e-02 -2.57539690e-01 3.92355740e-01
1.41177759e-01 5.89626670e-01 1.32035089e+00 -1.64811385e+00
4.70802069e-01 -4.70555663e-01 -4.78685498e-01 -6.66181743e-02
1.47457302e-01 -5.41293025e-01 -9.56153646e-02 -6.85480893e-01
3.37120354e-01 -1.10277915e+00 -5.81810415e-01 -2.28250086e-01
-3.80458266e-01 4.81840014e-01 4.45689261e-01 -6.67049766e-01
6.32892311e-01 -2.11193895e+00 5.64174891e-01 4.80964482e-01
2.71772087e-01 -2.48295322e-01 8.67446437e-02 -8.13837051e-02
-4.03954804e-01 -1.13213666e-01 -7.56301820e-01 -2.49181777e-01
-1.64613605e-01 6.55630708e-01 5.47995279e-03 9.92760539e-01
3.39288823e-02 8.48471045e-01 -7.91732609e-01 -5.32983840e-01
4.33800370e-01 8.94705951e-01 -7.68809438e-01 2.47084439e-01
-2.30042383e-01 1.37092340e+00 -3.89092833e-01 -1.56710505e-01
7.54680395e-01 -3.17922473e-01 2.69559145e-01 -5.17245054e-01
2.19858363e-01 -2.90908158e-01 -1.33530307e+00 1.99853921e+00
-4.71931338e-01 3.55127692e-01 2.83030778e-01 -1.47516429e+00
7.33822048e-01 4.25018102e-01 6.43123984e-01 -3.32538158e-01
3.84568930e-01 1.73827007e-01 1.06606342e-01 -7.06456006e-01
-2.55197227e-01 -5.80676019e-01 3.72956663e-01 1.92247167e-01
6.19327664e-01 -4.07584906e-01 -2.74489541e-02 2.31544152e-01
8.51786435e-01 2.34472483e-01 -1.23957768e-01 -7.50259757e-01
5.34046769e-01 -4.51163262e-01 6.30869448e-01 9.55399096e-01
9.23074260e-02 9.78793383e-01 3.39135051e-01 -1.26412570e-01
-9.29099858e-01 -1.46678019e+00 -8.82811189e-01 4.58192766e-01
-2.91248500e-01 2.30543345e-01 -8.54202330e-01 -4.21947539e-01
-2.48114854e-01 9.31870341e-01 -9.59120572e-01 -2.61450738e-01
-6.20641887e-01 -1.26338935e+00 2.55686641e-02 2.85398781e-01
7.89274126e-02 -9.89822865e-01 -4.22621340e-01 3.96912217e-01
-2.60675073e-01 -1.10255039e+00 -2.86264151e-01 5.44942319e-01
-1.15850627e+00 -7.97585368e-01 -7.98551619e-01 -5.15459657e-01
9.40329909e-01 -5.64893901e-01 1.26728499e+00 -4.37206864e-01
-5.75796068e-01 8.29608202e-01 -1.23061396e-01 -2.39362624e-02
-6.34616315e-01 -4.27679807e-01 2.46970430e-01 3.83825302e-01
-4.31533605e-01 -1.11815083e+00 -8.36894512e-01 1.28630310e-01
-1.26663876e+00 -4.13546294e-01 6.16009772e-01 1.14993656e+00
1.04910338e+00 -8.79837722e-02 3.93009126e-01 -1.25561821e+00
4.71678942e-01 -6.18039846e-01 -3.98977697e-01 8.42411220e-02
-8.73938441e-01 7.55147874e-01 2.83168554e-01 -5.12566149e-01
-1.43643582e+00 1.34025455e-01 -1.71806008e-01 -6.63196146e-01
-1.65757984e-01 3.95608991e-01 -1.00773714e-01 -2.57007461e-02
1.05819607e+00 3.26019645e-01 3.33121449e-01 -4.05892640e-01
6.27663791e-01 1.09640703e-01 4.94171262e-01 -7.27760017e-01
5.49622118e-01 9.40475702e-01 2.28659377e-01 -8.75029743e-01
-6.18357718e-01 4.20155302e-02 -8.48304570e-01 -3.12655181e-01
1.25641644e+00 -6.00024343e-01 -3.33091974e-01 4.01710048e-02
-9.50887144e-01 -2.20793575e-01 -7.33170331e-01 9.17777121e-01
-9.66453016e-01 3.50945234e-01 -7.69367099e-01 -5.51836073e-01
-2.21770197e-01 -1.61004937e+00 8.67039084e-01 -1.79487430e-02
8.72096866e-02 -1.41404235e+00 2.52489090e-01 1.50397271e-01
5.05107403e-01 5.27215421e-01 8.30123365e-01 -4.93693203e-01
-5.81174850e-01 1.98851585e-01 2.94144899e-01 6.65518522e-01
1.41213477e-01 -4.53520983e-01 -9.71528590e-01 -2.81165957e-01
8.64875317e-01 -2.00828150e-01 9.94218767e-01 1.12421668e+00
1.26892865e+00 -2.86229402e-01 -1.80980459e-01 9.10982609e-01
1.39766204e+00 -1.01081833e-01 7.71741450e-01 -7.78146535e-02
5.05191743e-01 4.73936766e-01 -1.49347827e-01 3.60461950e-01
-1.93994239e-01 4.59676415e-01 3.96545708e-01 -3.66437398e-02
-2.27432743e-01 9.39242095e-02 4.49817963e-02 8.80686402e-01
-3.15939844e-01 1.47006214e-01 -6.33560181e-01 3.72254193e-01
-1.78646314e+00 -7.52480268e-01 -1.51709378e-01 2.01545596e+00
8.19275618e-01 3.38538066e-02 -3.75960439e-01 -9.54930410e-02
6.23162031e-01 1.15818590e-01 -5.24029911e-01 -8.11588392e-03
2.34999228e-02 4.35685724e-01 4.96500522e-01 7.21514940e-01
-7.84706950e-01 3.64522219e-01 6.22301435e+00 7.43487418e-01
-1.10871077e+00 5.51666021e-01 6.59367502e-01 -3.92809324e-02
-5.87763429e-01 -1.55456021e-01 -1.48512065e-01 3.70630026e-01
9.92962599e-01 3.02200884e-01 4.68005210e-01 6.06576979e-01
3.25909466e-01 -1.33708874e-02 -1.29620206e+00 8.17118585e-01
3.29145610e-01 -1.36487925e+00 -6.13364503e-02 -4.40905569e-03
8.77263129e-01 1.30280823e-01 3.22211236e-01 7.34131858e-02
2.94074953e-01 -1.10703599e+00 4.58476096e-01 1.13456094e+00
6.46226406e-01 -4.09065872e-01 7.32819140e-01 3.87472838e-01
-3.55529249e-01 3.59752268e-01 -3.08276534e-01 5.63961327e-01
4.38690126e-01 1.07388783e+00 -6.98422253e-01 3.72547090e-01
5.95145822e-01 5.83792627e-01 -1.78769186e-01 5.28695047e-01
-4.08163756e-01 6.83862031e-01 -2.48959482e-01 4.77209061e-01
2.12422255e-02 -5.10676146e-01 9.32268083e-01 1.05169773e+00
6.60715848e-02 3.33912641e-01 -4.28391434e-02 1.43222034e+00
2.01658353e-01 -1.27228871e-01 -5.05729496e-01 2.64385909e-01
-1.50844783e-01 1.23492944e+00 -8.58498394e-01 -6.51838779e-02
-6.16190434e-02 8.42443287e-01 1.13647170e-01 7.05153048e-01
-7.21846819e-01 3.20320070e-01 1.76009983e-01 5.00843406e-01
2.81911284e-01 -9.81249139e-02 -1.40251368e-01 -1.29806519e+00
-1.69548616e-01 -6.00612998e-01 3.48650068e-01 -8.61546159e-01
-1.51580000e+00 7.35235870e-01 3.13543379e-01 -9.80168104e-01
-3.77498686e-01 -6.01480067e-01 -4.10085052e-01 6.85473561e-01
-1.17873776e+00 -8.37537050e-01 -6.94832066e-03 7.27638602e-01
3.07740241e-01 -1.51573926e-01 7.93888092e-01 2.03217626e-01
-2.19849765e-01 2.56073564e-01 4.92293745e-01 6.99884146e-02
5.48792005e-01 -1.33171916e+00 -5.93612194e-01 6.89596117e-01
1.97478697e-01 7.75721014e-01 9.79975581e-01 -6.44583821e-01
-1.29181373e+00 -7.93611705e-01 -8.75005275e-02 -5.10925710e-01
7.34257519e-01 -3.74047577e-01 -9.00781453e-01 9.36101377e-01
1.82687610e-01 5.52592218e-01 5.81928611e-01 -1.07000686e-01
1.93800986e-01 -9.88717079e-02 -1.45598292e+00 2.01202482e-01
7.55451620e-01 -3.50033849e-01 -8.08634102e-01 4.72692698e-01
4.74663824e-01 -6.18664622e-01 -1.14223409e+00 5.08553565e-01
7.53851309e-02 -9.22137976e-01 1.23298526e+00 -4.26754713e-01
3.94579053e-01 4.40157810e-03 -7.92450309e-02 -1.72040927e+00
-5.28565764e-01 -5.65072119e-01 -8.82102475e-02 9.15892184e-01
3.94270152e-01 -6.20063722e-01 6.43877804e-01 7.14795649e-01
-3.37806731e-01 -5.16886055e-01 -1.16366875e+00 -6.63542807e-01
3.49515349e-01 -5.67125499e-01 -4.44721952e-02 8.06419492e-01
-3.16556007e-01 2.84699291e-01 -4.62381721e-01 2.90843576e-01
1.22522783e+00 -2.97666103e-01 4.84998524e-02 -1.06334484e+00
-6.82669103e-01 -7.70491958e-02 -4.50438172e-01 -6.00122094e-01
2.95775861e-01 -1.17864478e+00 1.86728567e-01 -1.33361399e+00
1.23491086e-01 -2.86427349e-01 -3.59852940e-01 -5.19729704e-02
4.19947393e-02 6.61636293e-02 -9.02790800e-02 2.86845565e-01
-2.46840999e-01 8.98744881e-01 1.55180347e+00 -2.15163440e-01
-1.41652927e-01 -3.68575677e-02 -4.03140068e-01 9.75892365e-01
2.98770279e-01 -8.05188239e-01 -6.19626939e-01 -3.37996751e-01
1.01546839e-01 3.34249318e-01 3.65095466e-01 -8.85423005e-01
3.52428705e-02 9.66662839e-02 6.20373249e-01 -9.44754407e-02
2.04923943e-01 -1.00369716e+00 4.62779194e-01 4.48117286e-01
-2.69637674e-01 -4.97285068e-01 3.97678800e-02 7.63130426e-01
7.88828656e-02 -5.71235418e-01 1.09432638e+00 -4.73007560e-01
-1.12589985e-01 1.85108066e-01 -4.74963844e-01 3.91183436e-01
6.18422985e-01 2.98741329e-02 2.52867520e-01 -3.80455256e-01
-1.65146148e+00 -1.32651791e-01 3.06065008e-02 9.84840095e-02
6.56768560e-01 -1.19957614e+00 -9.30160761e-01 2.27904752e-01
-2.38734320e-01 8.47561583e-02 5.55747449e-01 1.23613977e+00
-3.36666673e-01 1.84207827e-01 -2.10924581e-01 -1.01402497e+00
-5.31860888e-01 7.35446990e-01 6.83306813e-01 -2.44910970e-01
-8.96038055e-01 7.77692914e-01 5.55022359e-01 -6.20598555e-01
6.36551753e-02 -2.85722464e-01 -2.00966343e-01 -1.89608689e-02
1.51713490e-01 1.07442111e-01 1.80946037e-01 -5.75410247e-01
-3.49509269e-01 4.47066158e-01 3.96827348e-02 -3.34438086e-01
1.74155927e+00 -1.19586147e-01 -1.19821198e-01 5.93939126e-01
1.18602216e+00 -1.68715358e-01 -1.28019810e+00 -2.51661003e-01
-2.69675225e-01 -1.15267836e-01 5.13609111e-01 -6.16406798e-01
-1.44809449e+00 6.99359059e-01 1.14774024e+00 -1.42391562e-01
9.03414071e-01 2.34697431e-01 1.41336769e-01 5.53406067e-02
3.65750313e-01 -1.24646533e+00 2.83373237e-01 2.75616534e-02
1.28079438e+00 -1.20260715e+00 2.04824150e-01 -2.49718949e-01
-7.06317067e-01 7.97713101e-01 1.15714535e-01 -5.37488997e-01
1.16940475e+00 4.53485429e-01 6.18427002e-04 -4.78195518e-01
-2.43150920e-01 4.99956943e-02 5.95731616e-01 7.62747109e-01
4.15634334e-01 5.50472550e-02 -4.78821099e-01 6.73392713e-01
1.01888746e-01 -6.91858754e-02 5.34150183e-01 5.84337234e-01
-1.70378402e-01 -9.76038992e-01 -5.83805442e-01 3.80153954e-01
-3.94161105e-01 -1.76246166e-02 2.48128995e-01 7.23221362e-01
2.39220008e-01 5.15034080e-01 -6.79626465e-02 2.64613092e-01
1.42989278e-01 5.15824370e-02 7.77383029e-01 -6.70927525e-01
-2.04137310e-01 4.91382360e-01 -3.14939320e-01 -5.81233382e-01
-5.84861577e-01 -8.62119138e-01 -1.38619816e+00 4.61736619e-01
-3.13237399e-01 2.42211834e-01 6.45897031e-01 1.04859674e+00
1.81121469e-01 9.42793429e-01 5.10540426e-01 -9.87589538e-01
-6.03359461e-01 -1.13442886e+00 -8.70411754e-01 6.91275418e-01
2.38461688e-01 -9.75860000e-01 -7.69256890e-01 3.30764472e-01] | [13.50645923614502, -2.35205340385437] |
d5c4eb39-72ab-4cd0-8294-8e2cc65c96b8 | first-and-second-order-bounds-for-adversarial | 2305.00832 | null | https://arxiv.org/abs/2305.00832v3 | https://arxiv.org/pdf/2305.00832v3.pdf | First- and Second-Order Bounds for Adversarial Linear Contextual Bandits | We consider the adversarial linear contextual bandit setting, which allows for the loss functions associated with each of $K$ arms to change over time without restriction. Assuming the $d$-dimensional contexts are drawn from a fixed known distribution, the worst-case expected regret over the course of $T$ rounds is known to scale as $\tilde O(\sqrt{Kd T})$. Under the additional assumption that the density of the contexts is log-concave, we obtain a second-order bound of order $\tilde O(K\sqrt{d V_T})$ in terms of the cumulative second moment of the learner's losses $V_T$, and a closely related first-order bound of order $\tilde O(K\sqrt{d L_T^*})$ in terms of the cumulative loss of the best policy $L_T^*$. Since $V_T$ or $L_T^*$ may be significantly smaller than $T$, these improve over the worst-case regret whenever the environment is relatively benign. Our results are obtained using a truncated version of the continuous exponential weights algorithm over the probability simplex, which we analyse by exploiting a novel connection to the linear bandit setting without contexts. | ['Chen-Yu Wei', 'Gergely Neu', 'Tim van Erven', 'Jack Mayo', 'Julia Olkhovskaya'] | 2023-05-01 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [-2.97804568e-02 5.11534572e-01 -4.76011306e-01 -1.72266930e-01
-1.36399841e+00 -1.01473558e+00 5.89631498e-02 2.49397740e-01
-1.08866167e+00 1.09114540e+00 -3.07885587e-01 -6.40551031e-01
-7.94425726e-01 -1.01722860e+00 -1.31093562e+00 -1.00183105e+00
-4.93021488e-01 4.36762154e-01 -1.66558683e-01 1.68986067e-01
9.14411396e-02 2.69320190e-01 -9.88385558e-01 -4.29640710e-01
6.60296440e-01 1.62902462e+00 -3.22888255e-01 7.67754197e-01
-2.31387705e-01 5.66286862e-01 -5.56742072e-01 -1.03933084e+00
6.81665242e-01 -4.12218362e-01 -6.18302166e-01 -3.53845149e-01
1.73481852e-01 -3.52301747e-01 -5.64699531e-01 1.33707356e+00
3.47457141e-01 4.54248667e-01 3.94960046e-01 -8.67658496e-01
-1.26134902e-01 8.03539813e-01 -5.92680871e-01 2.00623274e-01
5.18115237e-02 -8.85805711e-02 1.13874793e+00 2.34984100e-01
4.62529957e-01 1.00709999e+00 4.80305910e-01 5.34368694e-01
-1.38149810e+00 -1.00303555e+00 6.60536885e-01 -2.39828676e-01
-1.10053742e+00 -2.52756953e-01 6.07024789e-01 -2.27606535e-01
7.27591693e-01 4.23413068e-01 5.54646432e-01 4.28440630e-01
6.67146072e-02 8.77706170e-01 1.12718427e+00 -5.63962758e-01
5.03055513e-01 2.36812726e-01 -3.91106009e-02 6.58350229e-01
1.08995222e-01 5.10558069e-01 -2.76119858e-01 -2.33292505e-01
5.21941483e-01 1.19635612e-01 -1.72551677e-01 -3.01702917e-01
-5.25436699e-01 1.04731071e+00 4.14359838e-01 -3.35936770e-02
-1.60444498e-01 8.00224602e-01 3.89716178e-01 5.35473168e-01
5.68770468e-01 6.39557168e-02 -5.91515422e-01 -3.17388654e-01
-7.86377966e-01 4.15480375e-01 7.45812833e-01 1.24444151e+00
5.86067855e-01 -1.18827775e-01 -2.20909342e-01 4.91630256e-01
6.49120435e-02 7.94656873e-01 -1.94804698e-01 -1.20293784e+00
1.05655253e+00 -6.88204134e-04 7.14810073e-01 -3.09870809e-01
-6.74994662e-02 -5.55661738e-01 -4.27421898e-01 1.22756220e-01
8.63710940e-01 -5.34248710e-01 -7.04497755e-01 2.30498385e+00
1.25095516e-01 -2.16040134e-01 -2.02087179e-01 2.98556596e-01
-1.99599698e-01 4.82060492e-01 4.66937758e-02 -7.92725980e-01
8.22656214e-01 -3.09585124e-01 -4.64759856e-01 -3.73700768e-01
4.74185407e-01 -4.30230886e-01 9.23511982e-01 4.24056321e-01
-1.60918117e+00 3.30348015e-01 -7.80805647e-01 5.28080165e-01
-8.99456814e-02 -7.58975804e-01 7.80519187e-01 1.23356485e+00
-6.65667713e-01 5.16376197e-01 -6.29370928e-01 3.19837779e-01
7.37933457e-01 5.44439375e-01 -9.86987874e-02 -2.80371368e-01
-1.00255609e+00 5.31052351e-01 8.46416652e-02 -1.04895659e-01
-1.11705899e+00 -7.95957625e-01 -5.25702477e-01 3.41778286e-02
8.28343868e-01 -2.55579829e-01 1.37298429e+00 -7.40539312e-01
-1.30849922e+00 6.78584874e-01 -1.55159310e-01 -6.40607297e-01
9.93631601e-01 1.00952744e-01 1.52323186e-01 -1.79995131e-02
-5.80665916e-02 -1.69540316e-01 6.95051968e-01 -8.80892158e-01
-1.08147871e+00 -6.99609339e-01 5.35788774e-01 2.34429687e-01
-6.94465935e-02 -2.03040287e-01 -2.16992632e-01 -5.14702141e-01
-1.62218690e-01 -1.09939206e+00 -4.04164106e-01 -1.81245893e-01
-9.43098068e-02 9.71163958e-02 9.48134735e-02 -3.91719133e-01
1.31678820e+00 -1.98637915e+00 -9.08918157e-02 4.84492451e-01
-1.31213143e-01 -8.88480395e-02 2.25579783e-01 1.29108533e-01
1.92528173e-01 3.88583124e-01 -2.37442479e-01 -2.25417718e-01
4.25606877e-01 1.61398500e-01 -2.72425026e-01 5.07315993e-01
-7.14609206e-01 5.12029409e-01 -6.25331700e-01 -1.03193112e-02
-6.00057431e-02 -1.58902928e-01 -6.06201410e-01 1.79423392e-02
-4.65698570e-01 -1.68354109e-01 -6.26961350e-01 1.54066086e-01
6.72128379e-01 1.07868180e-01 2.10823819e-01 5.63282549e-01
1.14164621e-01 1.40071824e-01 -1.24271488e+00 1.47285461e+00
-7.29242444e-01 1.92375213e-01 3.94802511e-01 -1.36714947e+00
2.56010056e-01 2.38506749e-01 5.49553633e-01 -7.88788199e-01
2.19597206e-01 -1.24968523e-02 -2.94789493e-01 -2.76451185e-02
3.69371921e-02 -1.06312239e+00 -5.21435916e-01 7.52728164e-01
-1.63549617e-01 -9.20722485e-02 -1.24976940e-01 2.11063877e-01
1.38564813e+00 -3.65794003e-01 -1.07119299e-01 -8.10658932e-02
-7.72669464e-02 -2.32714817e-01 6.95242345e-01 1.26644433e+00
-4.38414454e-01 -3.16145480e-01 1.01493204e+00 -3.12794238e-01
-9.37003970e-01 -1.36499918e+00 -2.26515476e-02 1.31872499e+00
2.89047778e-01 2.99979866e-01 -8.23253810e-01 -9.26988542e-01
3.32149804e-01 1.02902067e+00 -1.03560221e+00 1.23258248e-01
-3.35350662e-01 -9.93736267e-01 4.72120553e-01 3.81903529e-01
4.74764675e-01 -6.90394402e-01 -3.54619652e-01 3.32656860e-01
1.99516416e-01 -6.54221237e-01 -6.98222458e-01 7.43497431e-01
-7.51408517e-01 -8.09853315e-01 -5.99975586e-01 -2.68633157e-01
6.01325989e-01 -1.91019714e-01 9.21456873e-01 -5.73195398e-01
-1.59617960e-01 6.21221960e-01 -8.13362375e-02 -9.40495491e-01
1.50914371e-01 -2.53175408e-01 -1.33989647e-01 -4.79272187e-01
2.12597057e-01 -4.48734343e-01 -1.05153549e+00 6.24154806e-02
-9.34151292e-01 -6.69517994e-01 2.67573923e-01 8.07374060e-01
9.09868240e-01 4.86196935e-01 4.92056161e-01 -1.06285667e+00
4.60897148e-01 -3.01359683e-01 -1.15251160e+00 4.23023492e-01
-7.57207334e-01 2.17344746e-01 6.64185405e-01 -6.35471642e-01
-1.05888557e+00 -2.81661540e-01 1.00226952e-02 -4.34306800e-01
4.10836279e-01 2.49464974e-01 -1.28587142e-01 -7.99046904e-02
4.71699417e-01 2.50884324e-01 -2.08919957e-01 -2.87482977e-01
6.04009688e-01 2.47113392e-01 2.34671384e-01 -9.72778797e-01
7.02277899e-01 4.33273256e-01 5.34846634e-02 -1.29439846e-01
-8.93580735e-01 2.36799806e-01 2.80760229e-01 1.39903933e-01
4.16641176e-01 -6.61137879e-01 -1.65200806e+00 -4.36746031e-02
-3.40171129e-01 -7.89487600e-01 -8.18783998e-01 4.55200523e-01
-9.87132728e-01 6.29463792e-02 -3.28344613e-01 -1.55459106e+00
-2.93208569e-01 -8.85035455e-01 2.63275892e-01 1.24438956e-01
3.83355618e-01 -9.94620025e-01 -2.56619304e-01 4.85965788e-01
3.63586754e-01 2.25527376e-01 1.19391251e+00 -3.69524568e-01
-7.10757196e-01 -6.54715240e-01 -9.01015848e-02 5.15178502e-01
-2.41472960e-01 -7.01961875e-01 -5.41241884e-01 -4.96371835e-01
2.02647224e-01 -2.28762925e-01 7.82818019e-01 7.47774899e-01
1.34340072e+00 -9.79257166e-01 -3.44432950e-01 5.64211130e-01
1.69567418e+00 8.11095059e-01 4.17452753e-01 3.62537384e-01
-9.44246259e-03 1.73093647e-01 6.44020438e-01 7.05902219e-01
1.26512691e-01 3.95044982e-01 7.29875803e-01 4.06978160e-01
7.00062335e-01 -2.41017178e-01 2.04693362e-01 -3.26567478e-02
1.20521575e-01 -2.73640841e-01 -6.42933309e-01 6.64111137e-01
-1.55028689e+00 -1.10863137e+00 6.90772235e-01 2.87468839e+00
9.79257584e-01 7.53164232e-01 3.04748148e-01 -1.35973236e-02
5.76813996e-01 1.29400268e-01 -9.54410672e-01 -9.52015281e-01
2.13040099e-01 8.78351212e-01 1.44980288e+00 7.72354126e-01
-7.59859562e-01 5.96956730e-01 6.06642866e+00 1.41547012e+00
-9.17571247e-01 2.50037402e-01 8.18015575e-01 -1.04906261e+00
-5.26266098e-01 6.92910049e-03 -5.97616255e-01 9.73301649e-01
9.85239506e-01 -5.83890438e-01 9.86269116e-01 1.04540610e+00
-2.02680856e-01 -4.76990074e-01 -1.22639358e+00 8.66484463e-01
-5.16501784e-01 -1.22050691e+00 -6.67122126e-01 4.84854966e-01
6.60427928e-01 4.85896273e-03 5.59222519e-01 3.37885976e-01
1.13758814e+00 -9.54945266e-01 8.04009259e-01 3.38797271e-02
1.26450157e+00 -1.43683732e+00 4.01646316e-01 6.97383821e-01
-8.95048201e-01 -4.84337091e-01 -1.90156028e-01 6.20976649e-02
4.08677198e-02 6.11994863e-01 -5.96179426e-01 3.85215163e-01
8.64743471e-01 -3.69450450e-01 6.37648582e-01 8.02352667e-01
-2.38065179e-02 5.12944043e-01 -7.74272621e-01 -1.43720537e-01
5.15812099e-01 -2.05912232e-01 3.46401900e-01 9.41083550e-01
3.65466028e-01 4.75811958e-01 -3.51281166e-02 2.90325552e-01
-4.84076321e-01 -2.31323019e-02 -3.13727051e-01 1.80027280e-02
6.22865021e-01 4.64183420e-01 -2.85589635e-01 1.03692338e-03
-1.02795534e-01 6.69816315e-01 3.15748215e-01 3.40569407e-01
-9.31070566e-01 -6.69176817e-01 8.78405094e-01 3.52178700e-02
6.39577150e-01 1.16099171e-01 -2.98641980e-01 -6.79538906e-01
2.60404557e-01 -3.31862390e-01 7.54514456e-01 3.94671187e-02
-1.32059741e+00 1.53409407e-01 -5.90425879e-02 -7.84668028e-01
3.64906676e-02 -4.16614234e-01 -2.63930649e-01 1.03139138e+00
-1.39535892e+00 -5.55812120e-01 6.28766418e-01 8.57731700e-01
-1.24103473e-02 -4.71692942e-02 8.86168838e-01 1.08605204e-02
-5.18006265e-01 1.29075348e+00 9.48316395e-01 -1.22126658e-02
2.43806586e-01 -1.37660849e+00 -2.72108316e-01 5.77828586e-01
-3.71303469e-01 4.09039140e-01 7.79791832e-01 -2.67875433e-01
-1.37638950e+00 -9.56709206e-01 2.57862210e-01 -1.87739208e-01
5.98364472e-01 -1.98018938e-01 -2.98127085e-01 1.03174722e+00
-3.01032841e-01 3.00812632e-01 9.71229434e-01 2.98711807e-01
-5.00797927e-01 -7.31860161e-01 -1.94067359e+00 4.21307534e-01
1.15004349e+00 -5.88155270e-01 -9.24884453e-02 2.63477743e-01
6.54705107e-01 -5.57864189e-01 -1.09819174e+00 1.09306157e-01
7.30912805e-01 -7.81675220e-01 6.83723807e-01 -6.46612346e-01
-7.84780234e-02 4.68841851e-01 -5.10824025e-01 -1.03833520e+00
3.52148376e-02 -1.14820063e+00 -1.57217011e-02 8.42339277e-01
6.38069093e-01 -8.64776492e-01 1.22894967e+00 1.23606706e+00
4.29458916e-01 -7.11475313e-01 -1.79752982e+00 -8.46877396e-01
8.62489402e-01 -7.86953092e-01 4.22795027e-01 6.16110384e-01
1.61311224e-01 -3.19787979e-01 -2.56661832e-01 -7.89423510e-02
9.78636801e-01 -1.10877603e-02 3.72655720e-01 -6.04433894e-01
-7.48938799e-01 -4.57653672e-01 1.83453299e-02 -1.10412633e+00
7.04100430e-02 -6.20485544e-01 1.78806707e-02 -9.64527786e-01
3.87246490e-01 -1.19757104e+00 -8.96283746e-01 3.09851855e-01
3.93449664e-02 -3.08382750e-01 2.86769360e-01 -3.63668293e-01
-7.00747490e-01 2.65643150e-01 1.07712162e+00 -2.18820482e-01
-9.84488949e-02 4.47772592e-01 -1.08655059e+00 5.79593122e-01
5.02454758e-01 -7.47767091e-01 -5.10477006e-01 -5.32238483e-01
5.34958363e-01 9.13649738e-01 -1.20219328e-01 -3.33589971e-01
-4.14859578e-02 -6.41261220e-01 6.63936511e-02 -3.28025579e-01
4.84450161e-01 -8.57381642e-01 7.48397633e-02 4.75392044e-01
-6.71799302e-01 -5.10316432e-01 1.51257545e-01 9.92464542e-01
3.28322738e-01 -2.50311434e-01 9.50859725e-01 -4.12739038e-01
2.69738913e-01 5.79867482e-01 -8.25551450e-02 2.59746879e-01
1.41908193e+00 -1.17776476e-01 -2.24091008e-01 -6.74072027e-01
-7.49683201e-01 3.68015975e-01 3.40191066e-01 -3.55060577e-01
1.06059469e-01 -1.06542265e+00 -1.46906421e-01 -2.73416758e-01
-2.70415515e-01 4.16869283e-01 6.81607068e-01 4.61732805e-01
-1.00532517e-01 3.37050825e-01 2.92497516e-01 1.41593500e-03
-8.24355483e-01 8.79243076e-01 5.37658632e-01 -7.36634314e-01
-4.03540544e-02 1.54809713e+00 1.50878325e-01 1.40596196e-01
6.40513837e-01 -4.19425070e-02 7.43237376e-01 -1.14675850e-01
4.70091224e-01 5.06063640e-01 -8.91639367e-02 5.37047051e-02
-1.99458495e-01 1.76390290e-01 -4.23879147e-01 -6.63049042e-01
1.43044925e+00 -2.33835593e-01 1.13930941e-01 5.39810210e-03
1.25946879e+00 7.95916542e-02 -1.61025918e+00 -6.01886213e-01
-2.26136386e-01 -7.79176056e-01 -5.53862378e-02 -1.12881541e+00
-1.12377870e+00 6.66820467e-01 8.30025375e-01 5.56205750e-01
1.24352777e+00 4.64532115e-02 5.62716603e-01 1.49358779e-01
8.38959396e-01 -1.35360146e+00 -2.68270701e-01 2.54095674e-01
4.51606393e-01 -9.31023836e-01 -9.36555937e-02 1.82189345e-02
-3.50094646e-01 4.48862761e-01 6.57812357e-02 -1.98304415e-01
7.39441872e-01 3.86817902e-01 -2.92581171e-01 3.72305751e-01
-5.97991288e-01 4.85268794e-02 -3.35715443e-01 3.44158649e-01
1.49680182e-01 5.26096404e-01 -4.00155604e-01 8.49465489e-01
-2.61955619e-01 -8.73872358e-03 1.17381126e-01 9.97909546e-01
-5.67647576e-01 -1.35666549e+00 -2.71616131e-01 7.58873701e-01
-1.18894279e+00 6.07764758e-02 2.16122955e-01 5.97777545e-01
1.58719152e-01 9.14963365e-01 1.80084482e-01 6.40451387e-02
3.66688728e-01 1.82399929e-01 8.16224635e-01 -1.75863966e-01
-4.07421261e-01 1.40493110e-01 -8.78930017e-02 -5.58481395e-01
-6.98465407e-02 -6.49455070e-01 -1.14267683e+00 -9.48123217e-01
-2.18989015e-01 4.05839711e-01 6.88002169e-01 8.96870792e-01
-1.80567250e-01 1.02218531e-01 9.80704248e-01 -3.00994843e-01
-8.64121139e-01 -5.78254402e-01 -9.74171221e-01 3.23956251e-01
4.05497879e-01 -4.36306477e-01 -5.43507218e-01 -4.49985564e-01] | [4.64522123336792, 3.408402442932129] |
f31bcfba-fee3-402a-9780-5a4fb5d30c89 | on-selecting-distance-metrics-in-n | 2306.09243 | null | https://arxiv.org/abs/2306.09243v1 | https://arxiv.org/pdf/2306.09243v1.pdf | On Selecting Distance Metrics in $n$-Dimensional Normed Vector Spaces of Cells: A Novel Criterion and Similarity Measure Towards Efficient and Accurate Omics Analysis | Single-cell omics enable the profiles of cells, which contain large numbers of biological features, to be quantified. Cluster analysis, a dimensionality reduction process, is used to reduce the dimensions of the data to make it computationally tractable. In these analyses, cells are represented as vectors in $n$-Dimensional space, where each dimension corresponds to a certain cell feature. The distance between cells is used as a surrogate measure of similarity, providing insight into the cell's state, function, and genetic mechanisms. However, as cell profiles are clustered in 3D or higher-dimensional space, it remains unknown which distance metric provides the most accurate spatiotemporal representation of similarity, limiting the interpretability of the data. I propose and prove a generalized proposition and set of corollaries that serve as a criterion to determine which of the standard distance measures is most accurate for conveying cell profile heterogeneity. Each distance method is evaluated via statistical, geometric, and topological proofs, which are formalized into a set of criteria. In this paper, I present the putative, first-ever method to elect the most accurate and precise distance metrics with any profiling modality, which are determined to be the Wasserstein distance and cosine similarity metrics, respectively, in general cases. I also identify special cases in which the criterion may select non-standard metrics. Combining the metric properties selected by the criterion, I develop a novel, custom, optimal distance metric that demonstrates superior computational efficiency, peak annotation, motif identification, and footprinting for transcription factor binding sites when compared with leading methods. | ['Elizabeth Engle', 'Arthur Lee', 'Okezue Bell'] | 2023-06-13 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [ 9.03360695e-02 -5.21567583e-01 -2.08227977e-01 3.73721425e-03
-3.76048088e-01 -1.03652000e+00 6.47252023e-01 5.53088367e-01
-4.02251095e-01 6.90481901e-01 -4.86313511e-04 -3.26124907e-01
-7.15225279e-01 -6.83698177e-01 -1.98147707e-02 -1.04003823e+00
-1.06652632e-01 3.33458483e-01 2.15384141e-02 -2.71572564e-02
5.71234167e-01 8.60483527e-01 -1.44932294e+00 -1.75361648e-01
5.97342253e-01 9.94532704e-01 1.32765368e-01 6.01262093e-01
-1.45543680e-01 -2.94096619e-02 -3.69294435e-01 3.01145576e-02
-4.71504629e-02 -5.46582699e-01 -5.50458312e-01 -7.95478821e-02
-2.14683339e-01 5.15995681e-01 -5.72786517e-02 7.80891299e-01
4.43381459e-01 -6.08741269e-02 1.22837698e+00 -1.15274608e+00
-5.27536392e-01 2.51458921e-02 -2.71500885e-01 2.92018682e-01
4.89773154e-01 -2.10045651e-01 9.92628396e-01 -8.70713055e-01
9.54697251e-01 7.95642853e-01 4.50890988e-01 3.79628330e-01
-1.38957369e+00 1.90274958e-02 -2.22579256e-01 -1.32423133e-01
-1.67335415e+00 -3.02902460e-01 6.42520666e-01 -6.76167786e-01
4.82484192e-01 6.88929796e-01 6.76363528e-01 5.49835026e-01
3.35672557e-01 3.00365239e-01 9.31786001e-01 -3.77850294e-01
6.09719813e-01 4.31984439e-02 5.07765859e-02 6.97461724e-01
5.22400916e-01 -1.68152228e-01 -3.56455743e-01 -4.45145637e-01
4.24015492e-01 1.53755367e-01 -4.57990080e-01 -7.51878321e-01
-1.32729161e+00 5.35471857e-01 -4.92473282e-02 9.53550458e-01
1.72282979e-02 -8.83343294e-02 2.54511237e-01 3.35634388e-02
2.24157453e-01 7.37170160e-01 -2.75924146e-01 -2.46995360e-01
-6.57527328e-01 2.83709943e-01 5.78357816e-01 6.59144938e-01
6.29272044e-01 -3.63322318e-01 4.56832387e-02 6.10321224e-01
3.26350719e-01 2.04770133e-01 5.62101662e-01 -9.36199069e-01
-9.28245857e-02 9.62803304e-01 1.70569971e-01 -1.38774467e+00
-6.18054450e-01 -3.03444088e-01 -8.41746271e-01 -2.14032650e-01
8.03416669e-01 1.66957512e-01 -2.37043485e-01 1.81448960e+00
4.41132426e-01 -1.25622690e-01 -4.79554683e-02 7.77196467e-01
2.96641439e-01 4.28095877e-01 -1.21995158e-01 -5.99878490e-01
1.32998741e+00 7.16758445e-02 -6.09902203e-01 5.38106024e-01
1.01225924e+00 -5.21071017e-01 1.04068291e+00 1.15928464e-01
-7.81286836e-01 -1.62794083e-01 -1.08764970e+00 2.64268994e-01
-8.96051168e-01 -2.10482016e-01 5.58025420e-01 5.89779973e-01
-8.95452917e-01 7.79513061e-01 -6.22535586e-01 -7.64673412e-01
3.25969942e-02 2.06487805e-01 -3.55205834e-01 2.43662357e-01
-9.30613935e-01 6.30234122e-01 1.08976364e-02 3.56682688e-02
-3.29808235e-01 -4.34197277e-01 -6.02154613e-01 -1.35635927e-01
-2.37419352e-01 -6.25085533e-01 4.58247691e-01 -5.02171814e-02
-1.23483884e+00 1.07422054e+00 -3.22089344e-01 1.37079209e-01
1.50573596e-01 5.64370394e-01 -4.50870365e-01 1.92027897e-01
1.65268093e-01 1.49236321e-01 1.27598599e-01 -1.16516149e+00
-4.46271688e-01 -6.64323211e-01 -2.63598800e-01 1.37651414e-01
-3.08117598e-01 -4.47745398e-02 -2.19908923e-01 -2.55928397e-01
4.73743886e-01 -7.43410766e-01 -2.17940807e-01 1.24891158e-02
-3.97410125e-01 -2.70419776e-01 6.18700087e-01 -1.09171480e-01
1.43752372e+00 -2.35305834e+00 3.81407589e-01 5.12501597e-01
3.48418713e-01 -8.97392407e-02 1.00244157e-01 7.54232705e-01
2.55083274e-02 4.71852481e-01 -3.97442579e-01 1.08765975e-01
7.91205093e-02 1.02945253e-01 -6.31208569e-02 9.97555912e-01
4.77217846e-02 6.68140531e-01 -9.53746676e-01 -5.91756046e-01
8.29031095e-02 3.50951672e-01 -1.45687804e-01 -1.20245442e-01
2.19428986e-01 3.71172160e-01 -5.63331306e-01 7.94182479e-01
3.28120202e-01 -2.73246109e-01 3.11675817e-01 -2.08677292e-01
-3.07730973e-01 -2.49324307e-01 -1.06276333e+00 1.42781568e+00
9.73988995e-02 5.79225838e-01 -2.43532673e-01 -1.20282567e+00
1.16621983e+00 1.33247644e-01 8.70476365e-01 -3.50870788e-01
1.86449975e-01 3.21781754e-01 -7.92690590e-02 -4.98653919e-01
2.91485161e-01 -1.02826886e-01 -4.40279067e-01 3.55490535e-01
-2.35033825e-01 8.08052495e-02 4.34359848e-01 3.51231806e-02
1.26769638e+00 -2.32134432e-01 3.63085896e-01 -9.23543930e-01
5.48031867e-01 -6.04928881e-02 5.79566777e-01 4.23722565e-01
-4.15910453e-01 5.68920791e-01 8.12169075e-01 -2.00557336e-01
-9.71970975e-01 -1.23378503e+00 -5.21848083e-01 6.22780919e-01
5.33954740e-01 -2.67379016e-01 -6.46295309e-01 -1.92166612e-01
3.06452274e-01 1.26345888e-01 -6.72160208e-01 -8.01251382e-02
-2.70552725e-01 -1.11165416e+00 6.48342609e-01 1.05026521e-01
-1.03293143e-01 -1.88675568e-01 -6.66931689e-01 1.62625895e-03
3.75462659e-02 -6.76748216e-01 -3.69414926e-01 3.02902907e-01
-9.30873990e-01 -1.38581622e+00 -5.92756689e-01 -5.77291906e-01
7.75034845e-01 2.42882475e-01 5.22989631e-01 -4.33544517e-02
-4.96900290e-01 3.19078565e-01 -2.78255045e-01 1.15830593e-01
-4.36884584e-03 -1.30936086e-01 4.87796336e-01 6.02162592e-02
4.23945636e-01 -5.43258250e-01 -6.53842807e-01 4.47572082e-01
-9.33606923e-01 -5.19014955e-01 2.07238480e-01 7.47092485e-01
9.10735309e-01 -3.80934104e-02 4.66897041e-01 -5.60143232e-01
7.93646276e-01 -5.36517262e-01 -3.75249684e-01 2.79907763e-01
-7.10279047e-01 2.01713830e-01 7.65665352e-01 -1.50750697e-01
-3.66666943e-01 -1.24995910e-01 3.57762545e-01 2.76753251e-02
-1.74401835e-01 6.38973117e-01 -4.42826599e-01 -2.49534860e-01
6.86902940e-01 4.87776756e-01 1.51297286e-01 -3.04483443e-01
2.27203965e-01 6.45660579e-01 3.99039954e-01 -5.92673182e-01
4.56182510e-01 7.98969388e-01 5.25394022e-01 -1.03720427e+00
-2.40899712e-01 -8.41619134e-01 -7.64219701e-01 -6.20036796e-02
6.81721985e-01 -2.37660125e-01 -1.05812824e+00 9.00489166e-02
-9.19816256e-01 4.36526500e-02 -9.87072438e-02 4.76663619e-01
-8.18171859e-01 5.82666993e-01 -5.87838888e-01 -8.20044756e-01
1.85251728e-01 -8.85179818e-01 8.82741153e-01 6.26765192e-02
-5.07617831e-01 -1.13021004e+00 4.64756399e-01 -2.30509818e-01
1.35862187e-01 5.76676428e-01 1.33529162e+00 -5.92021704e-01
-2.03050524e-01 -2.33119965e-01 -9.68680754e-02 -3.51924151e-01
2.42733598e-01 4.79110211e-01 -6.79683328e-01 -1.51147366e-01
-1.51491687e-01 4.91764873e-01 5.09109259e-01 4.39849347e-01
1.15350187e+00 -4.65583168e-02 -1.02203107e+00 8.12107921e-01
1.57195234e+00 5.61991692e-01 5.13301969e-01 1.19237863e-01
2.39618033e-01 6.71493351e-01 4.47501540e-01 4.78942811e-01
9.29776505e-02 6.66186094e-01 1.79368332e-01 1.38414711e-01
3.84715676e-01 -4.69468161e-02 -6.19131327e-02 9.75996137e-01
-1.88064739e-01 -3.94964606e-01 -9.30216789e-01 4.50634539e-01
-1.72051883e+00 -8.61742616e-01 -1.00097936e-02 2.37017775e+00
4.74279970e-01 -2.93572813e-01 3.83159846e-01 4.82952356e-01
7.86728263e-01 -1.33532882e-01 -3.58116686e-01 -3.16463560e-01
-4.98139322e-01 -2.49906912e-01 4.54359770e-01 4.51971769e-01
-8.10462236e-01 3.14641625e-01 7.20264626e+00 6.32324219e-01
-1.01405406e+00 -2.96323150e-01 6.02537870e-01 8.55783895e-02
-5.23927748e-01 -9.52970162e-02 -7.29186118e-01 6.81082010e-01
7.66372800e-01 -4.47894841e-01 7.53839016e-02 4.21516925e-01
4.34566140e-01 -1.76563233e-01 -1.29793477e+00 1.07157755e+00
-1.90397859e-01 -1.51260066e+00 3.31438333e-02 5.44094026e-01
4.65463459e-01 -5.32344818e-01 7.52248615e-02 -3.85653436e-01
-2.47288775e-02 -9.65522647e-01 2.96734512e-01 8.41465354e-01
8.57293487e-01 -7.36068428e-01 5.34287155e-01 2.37594277e-01
-1.25191987e+00 -3.58598158e-02 -3.90166134e-01 -8.74799788e-02
-4.61612595e-03 8.12161386e-01 -6.76198959e-01 4.26202804e-01
2.42534980e-01 4.94581044e-01 -2.08955139e-01 1.03755319e+00
5.76644957e-01 1.74726412e-01 -3.72113913e-01 -5.23250461e-01
1.24410607e-01 -5.64045012e-01 5.58543801e-01 1.28568041e+00
7.12891221e-01 3.23826402e-01 -2.06972271e-01 9.30695474e-01
2.94448078e-01 3.52106750e-01 -1.06943882e+00 -3.14249307e-01
8.65918458e-01 1.23992455e+00 -1.18864763e+00 3.98396961e-02
-1.58018544e-01 7.01411068e-01 2.06024736e-01 3.75667512e-01
-4.90041256e-01 -7.42976367e-01 1.04009235e+00 3.36068034e-01
-1.89079508e-01 -5.47457337e-01 -4.71777588e-01 -9.62971985e-01
-9.00187343e-02 -3.18039954e-01 4.09463376e-01 -2.48227477e-01
-1.20491993e+00 2.32064456e-01 -1.44065663e-01 -1.36047804e+00
-3.02530468e-01 -9.61960018e-01 -3.49335611e-01 6.87668920e-01
-9.94868696e-01 -4.74312216e-01 -1.10687554e-01 3.18381280e-01
-2.03745738e-01 1.14830928e-02 9.34423268e-01 1.66792214e-01
-6.04325712e-01 4.08705980e-01 6.71402872e-01 1.78751703e-02
3.20258647e-01 -1.23265648e+00 -2.05783129e-01 4.31406289e-01
-2.07084492e-02 7.59439766e-01 8.32172513e-01 -4.04430240e-01
-1.70311153e+00 -6.90767765e-01 1.11893487e+00 -5.23123384e-01
7.68236458e-01 -4.67130482e-01 -6.23077154e-01 6.70375675e-02
-5.08661568e-01 -6.80737421e-02 1.20014870e+00 9.95658562e-02
-8.40936676e-02 -1.87718168e-01 -1.11148977e+00 7.83313334e-01
1.12675619e+00 -6.77227378e-01 -1.66691199e-01 1.85896724e-01
2.37750754e-01 3.30234468e-02 -1.27151883e+00 3.40084612e-01
7.32076168e-01 -8.95240486e-01 8.96233320e-01 -5.77683270e-01
6.59246594e-02 -7.89451301e-01 -5.84065914e-01 -1.12124181e+00
-5.34360409e-01 -5.98138511e-01 1.61273390e-01 9.82971489e-01
5.69259763e-01 -6.44658446e-01 7.12694764e-01 3.93026650e-01
6.17898591e-02 -1.10028088e+00 -1.14139915e+00 -9.85538244e-01
9.60488468e-02 -1.54574633e-01 7.16382980e-01 1.19115889e+00
8.74992490e-01 1.24924451e-01 2.27057472e-01 -2.55536407e-01
5.98062217e-01 1.62773311e-01 4.29224193e-01 -1.38693547e+00
1.19110495e-01 -8.04936469e-01 -9.48938131e-01 -9.57800746e-01
-6.13390766e-02 -9.33844805e-01 -1.13742918e-01 -1.17516994e+00
1.26849607e-01 -5.76861680e-01 -4.18986917e-01 -1.87218696e-01
1.85830325e-01 8.86736438e-02 -6.21186420e-02 2.61823714e-01
-3.73184144e-01 2.54753202e-01 1.15806198e+00 3.48458290e-02
-2.57784367e-01 -2.30672464e-01 -7.93634176e-01 4.73991007e-01
6.74312294e-01 -5.02850581e-03 -3.01933974e-01 6.80473745e-02
2.84094125e-01 1.84605703e-01 2.43006498e-01 -9.31391954e-01
3.20652157e-01 -4.13268119e-01 4.97027904e-01 -3.80603492e-01
3.17927957e-01 -6.83871508e-01 3.76718849e-01 4.79735345e-01
-3.48567516e-01 8.40485916e-02 -1.64736465e-01 6.05467558e-01
-2.33417556e-01 -1.47471532e-01 7.01049984e-01 2.68909838e-02
-4.89749759e-01 3.00271511e-01 -4.87260789e-01 -1.92700490e-01
1.16953087e+00 -7.74953187e-01 -3.25145364e-01 -1.07596084e-01
-7.18928397e-01 -1.38604566e-01 7.43570328e-01 -5.01190536e-02
4.96028662e-01 -1.49704754e+00 -3.62174422e-01 1.84809733e-02
3.19898486e-01 -4.20358866e-01 4.61560860e-02 1.01145363e+00
-5.62450826e-01 7.37736285e-01 -1.10791221e-01 -8.30766559e-01
-9.94178951e-01 8.01133096e-01 3.65249187e-01 8.73466674e-03
-9.39116776e-02 4.72991258e-01 2.28901699e-01 -1.77493945e-01
-1.67881832e-01 -3.66481811e-01 -1.55517623e-01 2.68447220e-01
3.42518955e-01 5.29266834e-01 -6.28963485e-02 -7.32917190e-01
-6.53194249e-01 9.90046799e-01 4.99279857e-01 -1.02680614e-02
9.31937397e-01 -4.86142755e-01 -2.84107089e-01 8.20290804e-01
1.34012234e+00 -1.48858905e-01 -8.63948882e-01 8.26017931e-02
3.46048355e-01 -4.37856644e-01 -4.24106449e-01 -2.45876014e-01
-5.32597601e-01 6.98189914e-01 4.28749889e-01 6.82612360e-01
9.60754216e-01 2.01690391e-01 3.15813214e-01 3.11681539e-01
4.03729975e-01 -1.12360418e+00 -2.31116191e-01 3.99614394e-01
5.51218331e-01 -6.74733758e-01 -1.63143784e-01 -3.45045358e-01
-4.69669849e-02 1.17114532e+00 2.02907190e-01 1.03937350e-01
8.49521697e-01 3.90946984e-01 -1.24107458e-01 -3.19770962e-01
-5.99052846e-01 -1.19212039e-01 3.99461612e-02 6.35461688e-01
5.59597969e-01 1.86917841e-01 -7.75428772e-01 4.92153615e-01
-1.79012775e-01 -1.63904965e-01 2.46430010e-01 7.08288133e-01
-8.85453045e-01 -7.71508753e-01 -1.87693313e-01 3.51963699e-01
-2.38350764e-01 3.11592847e-01 -4.37746018e-01 7.68108726e-01
1.22221127e-01 8.62960994e-01 1.76196128e-01 -5.45064569e-01
1.05417043e-01 3.71434242e-02 3.23344439e-01 -8.32847878e-02
1.27295852e-01 -8.68883058e-02 -2.14077890e-01 -4.97344375e-01
-2.74878442e-01 -6.43735588e-01 -1.29611456e+00 -5.42315006e-01
-2.21476376e-01 5.18481910e-01 5.91588497e-01 9.99181986e-01
5.77800453e-01 -3.19744870e-02 8.38891745e-01 -4.76885170e-01
-2.65060306e-01 -4.09529239e-01 -1.13752007e+00 4.08425152e-01
6.59606159e-02 -7.93516695e-01 -4.33620006e-01 3.03980559e-02] | [7.407649993896484, 4.325754642486572] |
0abdb60a-4077-479d-aab6-8a4df2490931 | from-axioms-over-graphs-to-vectors-and-back | 2303.16519 | null | https://arxiv.org/abs/2303.16519v2 | https://arxiv.org/pdf/2303.16519v2.pdf | From axioms over graphs to vectors, and back again: evaluating the properties of graph-based ontology embeddings | Several approaches have been developed that generate embeddings for Description Logic ontologies and use these embeddings in machine learning. One approach of generating ontologies embeddings is by first embedding the ontologies into a graph structure, i.e., introducing a set of nodes and edges for named entities and logical axioms, and then applying a graph embedding to embed the graph in $\mathbb{R}^n$. Methods that embed ontologies in graphs (graph projections) have different formal properties related to the type of axioms they can utilize, whether the projections are invertible or not, and whether they can be applied to asserted axioms or their deductive closure. We analyze, qualitatively and quantitatively, several graph projection methods that have been used to embed ontologies, and we demonstrate the effect of the properties of graph projections on the performance of predicting axioms from ontology embeddings. We find that there are substantial differences between different projection methods, and both the projection of axioms into nodes and edges as well ontological choices in representing knowledge will impact the success of using ontology embeddings to predict axioms. | ['Robert Hoehndorf', 'Fernando Zhapa-Camacho'] | 2023-03-29 | null | null | null | null | ['ontology-embedding'] | ['knowledge-base'] | [-1.31107479e-01 7.10901141e-01 -1.28962830e-01 -4.70086396e-01
3.49175006e-01 -6.55872226e-01 8.00325692e-01 3.56585354e-01
-1.41166270e-01 2.47111037e-01 7.79056549e-01 -6.44761980e-01
-4.69905645e-01 -1.40200615e+00 -5.97880960e-01 -1.96098853e-02
-2.19277009e-01 6.75169468e-01 2.92669386e-01 -4.83791620e-01
-2.22160414e-01 2.39427224e-01 -1.46405983e+00 2.17047185e-01
4.75224763e-01 6.33635461e-01 -3.53850067e-01 3.53441238e-01
-5.21640539e-01 1.25227022e+00 -2.90390968e-01 -7.72999227e-01
1.43873155e-01 -4.15681720e-01 -1.05879486e+00 -2.48402804e-01
2.00584665e-01 1.55789042e-02 -7.68815696e-01 1.03128064e+00
-3.68192270e-02 2.10055392e-02 8.37658167e-01 -1.64852011e+00
-1.33505940e+00 1.04231024e+00 3.97078753e-01 1.22548575e-02
6.93349421e-01 -9.56478938e-02 1.65281713e+00 -5.12637854e-01
9.87610281e-01 1.36978006e+00 7.08676636e-01 5.34090698e-01
-1.49357140e+00 -4.19065475e-01 -2.19540060e-01 3.62699151e-01
-1.38069499e+00 -3.28876883e-01 7.21132696e-01 -8.26610923e-01
9.64135647e-01 1.68145359e-01 7.91430652e-01 6.84602857e-01
3.11007768e-01 1.26234367e-01 7.77750194e-01 -6.91639304e-01
4.45217848e-01 5.85623026e-01 4.70067501e-01 9.63734031e-01
5.25286674e-01 -1.79164320e-01 -1.54264599e-01 -4.41136181e-01
4.42110568e-01 -3.74446690e-01 -2.07969114e-01 -8.71383011e-01
-1.05223417e+00 1.22831583e+00 6.52553141e-01 3.09253812e-01
-1.43654436e-01 4.51811880e-01 5.08170724e-01 3.58578920e-01
-8.48304555e-02 8.11483443e-01 -2.51682997e-01 4.59120035e-01
-2.17770711e-01 2.27259323e-01 9.37537134e-01 1.01736295e+00
9.36704993e-01 -4.52376269e-02 2.91872531e-01 4.59805757e-01
4.60278451e-01 6.53221384e-02 4.09471929e-01 -8.78219366e-01
1.01562008e-01 1.06302643e+00 -1.69124931e-01 -1.12972820e+00
-4.03954089e-01 3.46326619e-01 1.71405375e-01 4.90861721e-02
1.00921899e-01 -7.58177694e-03 -6.27116382e-01 1.67616677e+00
2.52756402e-02 -2.45102912e-01 3.33809823e-01 5.61305881e-01
1.07860816e+00 4.18864012e-01 2.28332296e-01 5.09150088e-01
1.60218740e+00 -4.84843373e-01 -6.14373446e-01 -2.04825550e-01
1.12937081e+00 -4.44320068e-02 1.14250481e+00 -3.69100183e-01
-7.89379954e-01 -1.42954022e-01 -1.08731914e+00 -3.21625680e-01
-9.04761732e-01 -2.96791792e-01 8.34857583e-01 6.72765434e-01
-1.05550230e+00 5.31838357e-01 -4.63279456e-01 -4.81234252e-01
2.03531370e-01 4.43849176e-01 -7.53913701e-01 -5.47489002e-02
-1.57897162e+00 1.21775258e+00 7.72086859e-01 -5.64210653e-01
-4.07489985e-01 -5.39191425e-01 -1.45612502e+00 4.44468290e-01
9.57709700e-02 -6.97967410e-01 6.14863932e-01 -9.03966606e-01
-9.28748727e-01 8.52686524e-01 2.60138839e-01 -5.59583962e-01
-1.94149673e-01 4.84920949e-01 -7.09318876e-01 -5.38900606e-02
1.71701059e-01 6.44491315e-01 1.77115485e-01 -1.04100382e+00
-4.20986414e-01 -3.57606500e-01 9.97985721e-01 -4.30310890e-02
-7.49813378e-01 -3.38095240e-04 -1.19480096e-01 -5.05694300e-02
2.02587709e-01 -9.33774829e-01 7.20669031e-02 4.34557348e-02
-2.35809028e-01 -3.36501449e-01 5.65435112e-01 -4.49875891e-01
1.13942134e+00 -2.20737457e+00 2.81399101e-01 4.55293894e-01
4.24981117e-01 -3.35846752e-01 -8.59563351e-02 6.98514640e-01
-2.83070832e-01 6.65361226e-01 3.81344594e-02 2.74044722e-01
7.65560687e-01 5.26073396e-01 -3.45528901e-01 3.32885802e-01
2.06831828e-01 9.20630217e-01 -8.78142416e-01 -4.10031855e-01
2.65076280e-01 5.52627623e-01 -1.03359628e+00 -1.84365913e-01
-4.25789863e-01 -3.30546588e-01 -3.76891881e-01 9.58204046e-02
3.56996693e-02 -2.65593797e-01 9.59225178e-01 -2.58743465e-01
9.97649357e-02 8.85373414e-01 -1.19608104e+00 9.68108475e-01
-3.67487162e-01 9.23890471e-01 -5.81990957e-01 -8.13394845e-01
7.83394933e-01 5.15889943e-01 2.78772354e-01 -3.90171587e-01
4.08554152e-02 4.38998034e-03 2.49509752e-01 -4.00891244e-01
4.90427345e-01 -5.23557663e-01 -1.07880525e-01 4.60578144e-01
3.21928471e-01 -3.02694768e-01 4.94458884e-01 4.08473551e-01
1.29676461e+00 1.13204375e-01 3.84113073e-01 -5.26222944e-01
3.25605333e-01 2.88849056e-01 5.82161963e-01 2.09278375e-01
7.17595965e-02 2.36045588e-02 1.06275856e+00 -6.95202470e-01
-1.01753783e+00 -1.14641309e+00 -2.96553224e-01 9.70815897e-01
-3.11659649e-02 -8.89983177e-01 -3.08029294e-01 -7.60759830e-01
3.94386977e-01 1.27005053e+00 -7.69753397e-01 -5.26843786e-01
-1.35266945e-01 -1.78922594e-01 7.20245242e-01 6.24473333e-01
-2.61520028e-01 -1.03939211e+00 -4.41248983e-01 -1.87320225e-02
9.11923200e-02 -1.33403909e+00 -9.47295353e-02 2.29965374e-01
-6.72833145e-01 -1.23339331e+00 3.86989325e-01 -8.70094717e-01
9.69700694e-01 -2.61966318e-01 1.25625265e+00 2.45799944e-01
1.04971334e-01 6.52979553e-01 -4.03486401e-01 -3.32552344e-01
-6.10059142e-01 -3.42860967e-01 2.11923331e-01 -2.38052666e-01
6.11015975e-01 -6.32181406e-01 1.47403523e-01 -2.00126953e-02
-1.11612236e+00 -1.13474779e-01 -2.65130308e-02 5.78236878e-01
1.04402445e-01 4.89449471e-01 9.94192585e-02 -1.15030289e+00
8.65280509e-01 -4.30152118e-01 -5.80822229e-01 3.83049548e-01
-6.20474994e-01 6.16137803e-01 6.86168015e-01 -7.88736194e-02
-4.22295272e-01 -2.91104138e-01 7.62286335e-02 -3.30563426e-01
2.44647175e-01 9.47721064e-01 -1.11360930e-01 8.48152861e-02
8.40824664e-01 -1.48944438e-01 -6.19022467e-04 5.20591363e-02
8.43001783e-01 3.10073525e-01 -5.01044244e-02 -6.63713098e-01
8.43904614e-01 3.79336864e-01 8.54298323e-02 -7.30464101e-01
-3.98311496e-01 8.89926553e-02 -5.83239853e-01 1.71134710e-01
1.11886799e+00 -6.11700535e-01 -4.56524760e-01 -6.49125755e-01
-8.55433404e-01 -1.60203636e-01 -8.13424885e-01 5.32135606e-01
-4.29998070e-01 2.61869431e-01 -3.88101846e-01 -4.08869773e-01
1.79590479e-01 -1.26172936e+00 3.62431854e-01 -2.31324151e-01
-5.91679931e-01 -1.48434818e+00 2.38327354e-01 1.12997338e-01
2.72151053e-01 1.44596487e-01 1.95363605e+00 -9.30942535e-01
-3.18701625e-01 -3.15749794e-01 -2.69532949e-01 3.41743737e-01
2.36419976e-01 1.66915730e-01 -6.37362182e-01 7.69750699e-02
-4.93807703e-01 -1.74235225e-01 2.67285317e-01 5.08059189e-02
3.80208701e-01 -5.02298892e-01 -4.58633870e-01 3.78360718e-01
1.63176656e+00 -2.35518515e-02 8.29863906e-01 3.75862241e-01
5.47814310e-01 5.53680062e-01 -1.64194360e-01 3.16926017e-02
6.99624777e-01 5.13183534e-01 1.19208887e-01 3.87320638e-01
7.58598149e-02 -4.92766231e-01 5.29838681e-01 7.94082344e-01
3.42813767e-02 -6.07131720e-02 -1.30538571e+00 5.43626547e-01
-1.48030555e+00 -9.84325349e-01 6.25398010e-02 1.89971864e+00
7.82679737e-01 1.52129337e-01 -5.89198321e-02 4.85347845e-02
5.73233783e-01 6.23509809e-02 1.86584771e-01 -6.22925341e-01
-2.16379128e-02 5.18554747e-01 5.08666754e-01 7.22757757e-01
-6.11576736e-01 8.81760418e-01 6.96519899e+00 -1.92221537e-01
-7.99374878e-01 5.45313880e-02 -3.42627406e-01 2.65472412e-01
-9.11020935e-01 6.69617295e-01 -2.69569784e-01 3.22708249e-01
9.77892995e-01 -5.65922260e-01 6.09754503e-01 7.64627337e-01
-2.79158026e-01 4.39014792e-01 -1.54573274e+00 3.08767289e-01
-4.59012315e-02 -1.51563954e+00 4.56918091e-01 1.34351939e-01
5.49667120e-01 -2.06360742e-01 -3.75923216e-01 5.76290369e-01
8.57771158e-01 -1.09100807e+00 7.28450596e-01 4.44536239e-01
4.56729501e-01 -3.83874089e-01 7.38598406e-01 -1.36434749e-01
-1.18569016e+00 -3.65725189e-01 -5.56138813e-01 -2.59806752e-01
-1.12222657e-01 2.46440917e-01 -9.51129556e-01 5.24000883e-01
4.52119857e-01 3.81423503e-01 -5.41864216e-01 3.75866294e-01
-5.63967645e-01 3.60784620e-01 -2.05915764e-01 -1.98605299e-01
6.31434396e-02 -5.10426223e-01 3.35364372e-01 9.65155542e-01
1.30171075e-01 7.55671263e-02 6.52521104e-02 1.24619639e+00
-3.35711896e-01 7.91560486e-02 -9.94847953e-01 -8.29892218e-01
5.69857895e-01 8.97689462e-01 -5.42604923e-01 -3.64412308e-01
-7.65071034e-01 3.46891820e-01 3.27849180e-01 3.72892231e-01
-8.60622883e-01 -4.69097525e-01 9.85542536e-01 4.99258608e-01
2.90517330e-01 -2.98341751e-01 6.64858371e-02 -1.08758307e+00
-7.20871612e-03 -7.11120903e-01 5.00003755e-01 -1.15487039e+00
-1.03618550e+00 3.55639219e-01 1.54895321e-01 -6.67532027e-01
8.16228762e-02 -9.97439027e-01 -6.11732364e-01 6.42715514e-01
-1.10144413e+00 -1.08448923e+00 2.00873781e-02 3.28353256e-01
-3.55915636e-01 -1.74066842e-01 1.15670562e+00 6.35677204e-02
-1.52744845e-01 3.13980132e-01 -1.40696645e-01 5.76234639e-01
1.48223326e-01 -1.41182196e+00 2.46985152e-01 5.27438700e-01
5.61471343e-01 9.67199743e-01 7.38047838e-01 -4.49473590e-01
-1.56147146e+00 -9.23979342e-01 1.06919050e+00 -8.10000777e-01
1.09384763e+00 -3.50960881e-01 -6.22695684e-01 1.63336098e+00
1.66207194e-01 1.86267540e-01 8.23288023e-01 5.53605497e-01
-8.21971238e-01 1.24909051e-01 -1.10491407e+00 9.58213925e-01
1.05501652e+00 -1.01486933e+00 -1.11746562e+00 2.87392527e-01
1.01456285e+00 -2.09180196e-03 -1.24823344e+00 2.72234201e-01
5.86871624e-01 -7.71089554e-01 9.64571893e-01 -1.28043830e+00
4.97765064e-01 -3.21257830e-01 -5.09337187e-01 -1.26443219e+00
-7.38291323e-01 1.01386495e-01 9.39472839e-02 9.76125896e-01
7.23973393e-01 -1.13255298e+00 5.19944906e-01 8.24602425e-01
1.37249961e-01 -5.85089087e-01 -6.51217401e-01 -6.39258444e-01
2.00886741e-01 -2.23533094e-01 9.23575222e-01 1.31020665e+00
6.79208279e-01 5.64214230e-01 3.73401195e-01 2.73191690e-01
1.64543316e-01 -5.31951785e-02 6.64100468e-01 -1.49410057e+00
-1.08130351e-02 -3.60065520e-01 -1.36909068e+00 -2.58099198e-01
4.93305117e-01 -1.51690090e+00 -3.99801970e-01 -1.95607924e+00
-9.07488987e-02 -4.64844227e-01 -2.64964312e-01 6.66132569e-01
2.77405292e-01 -1.93501413e-01 1.75640166e-01 -6.60664737e-02
-1.44033462e-01 4.28359032e-01 7.12850988e-01 -3.34807426e-01
5.62047325e-02 -8.94490302e-01 -8.23484123e-01 7.83428788e-01
5.59109032e-01 -5.27535796e-01 -6.29622221e-01 -5.06006718e-01
7.09649384e-01 -2.47744873e-01 3.60669047e-01 -1.04121447e+00
1.16257735e-01 -2.18938947e-01 -6.68582916e-02 3.80472064e-01
2.41163120e-01 -8.40940356e-01 5.17673671e-01 2.86420524e-01
-4.93093729e-01 3.68214399e-01 2.14759424e-01 3.75108093e-01
-7.62590915e-02 -5.96521378e-01 5.85472047e-01 -1.01846576e-01
-8.51015210e-01 -7.71779492e-02 -3.63102406e-01 2.67961353e-01
8.76796424e-01 -2.52583295e-01 -2.10066184e-01 -3.37054580e-01
-8.75589609e-01 3.72890085e-02 8.56223702e-01 4.98702824e-01
4.57763553e-01 -1.55282032e+00 -3.52322459e-01 3.90366912e-02
3.64701420e-01 -5.34629464e-01 -4.82872128e-01 6.26572609e-01
-6.83152258e-01 4.11930203e-01 -3.58922929e-01 -9.65172648e-02
-9.54838574e-01 4.86467570e-01 5.24783075e-01 -8.94327462e-02
-7.18746960e-01 4.97420728e-01 1.00679033e-01 -9.79388952e-01
-6.45027161e-02 -5.19790411e-01 -1.76869556e-01 -4.39981781e-02
-1.09670408e-01 3.24402787e-02 -2.24892005e-01 -5.95341325e-01
-5.24498343e-01 2.01528341e-01 2.69042403e-01 -8.14548656e-02
1.33638728e+00 5.06686628e-01 -5.10627270e-01 5.30345082e-01
9.99236465e-01 1.92339733e-01 -3.57955635e-01 -1.74954742e-01
1.29863545e-01 -2.93534696e-01 -6.84663951e-02 -1.93199500e-01
-7.72686243e-01 5.61989188e-01 1.21460900e-01 7.05571532e-01
3.26433599e-01 1.40568331e-01 2.97310650e-01 4.39839512e-01
4.81804878e-01 -9.64397728e-01 -1.36452720e-01 5.68949938e-01
6.69565856e-01 -6.42543435e-01 3.01092923e-01 -4.16680306e-01
-5.21761596e-01 1.20109451e+00 2.49756813e-01 -2.36078441e-01
6.77025378e-01 1.39072075e-01 -7.57786855e-02 -7.93359697e-01
-5.75914741e-01 -2.58370042e-01 1.50631279e-01 5.85101724e-01
5.41897833e-01 3.29954088e-01 -2.43046045e-01 5.20226717e-01
-5.53986192e-01 -7.54224211e-02 7.21006572e-01 8.86218429e-01
-3.12462538e-01 -1.04608738e+00 1.15800700e-04 5.72513223e-01
1.13123478e-02 -1.69821456e-01 -6.81594610e-01 1.14993179e+00
2.90648431e-01 6.33468091e-01 1.92114487e-01 -6.38868034e-01
4.17433679e-01 6.50588036e-01 5.32410562e-01 -1.06211412e+00
-1.77399740e-01 -1.06738567e+00 4.77088958e-01 -1.50598958e-01
-8.51037949e-02 -5.33903837e-01 -1.78998363e+00 -4.00822133e-01
-5.84583342e-01 3.95631343e-01 4.57954615e-01 1.05865979e+00
4.11976367e-01 5.81058800e-01 4.01392207e-02 -9.14066732e-02
-4.67674285e-01 -5.82681239e-01 -6.07220232e-01 7.35893488e-01
-2.12814987e-01 -9.23101664e-01 -4.79000479e-01 -3.11279017e-03] | [8.902565002441406, 7.750180721282959] |
5c53564d-db49-4a92-bc6f-260cc16c6408 | semi-supervised-image-captioning-by | 2301.11174 | null | https://arxiv.org/abs/2301.11174v1 | https://arxiv.org/pdf/2301.11174v1.pdf | Semi-Supervised Image Captioning by Adversarially Propagating Labeled Data | We present a novel data-efficient semi-supervised framework to improve the generalization of image captioning models. Constructing a large-scale labeled image captioning dataset is an expensive task in terms of labor, time, and cost. In contrast to manually annotating all the training samples, separately collecting uni-modal datasets is immensely easier, e.g., a large-scale image dataset and a sentence dataset. We leverage such massive unpaired image and caption data upon standard paired data by learning to associate them. To this end, our proposed semi-supervised learning method assigns pseudo-labels to unpaired samples in an adversarial learning fashion, where the joint distribution of image and caption is learned. Our method trains a captioner to learn from a paired data and to progressively associate unpaired data. This approach shows noticeable performance improvement even in challenging scenarios including out-of-task data (i.e., relational captioning, where the target task is different from the unpaired data) and web-crawled data. We also show that our proposed method is theoretically well-motivated and has a favorable global optimal property. Our extensive and comprehensive empirical results both on (1) image-based and (2) dense region-based captioning datasets followed by comprehensive analysis on the scarcely-paired COCO dataset demonstrate the consistent effectiveness of our semisupervised learning method with unpaired data compared to competing methods. | ['In So Kweon', 'Jinsoo Choi', 'Tae-Hyun Oh', 'Dong-Jin Kim'] | 2023-01-26 | null | null | null | null | ['relational-captioning', 'relational-captioning'] | ['computer-vision', 'natural-language-processing'] | [ 6.28549576e-01 2.76493073e-01 -3.18829209e-01 -5.54853737e-01
-1.62165391e+00 -9.23642993e-01 5.77724159e-01 -1.93472981e-01
-4.14721280e-01 8.52197170e-01 -3.07685807e-02 -1.12945333e-01
3.33244264e-01 -4.30595726e-01 -1.39952111e+00 -7.70895541e-01
2.99558848e-01 8.96549881e-01 -8.95115957e-02 8.34635571e-02
-2.09220111e-01 -2.18870137e-02 -1.22533631e+00 1.89601094e-01
8.86529207e-01 1.08821857e+00 4.80291694e-01 4.47123468e-01
4.39189859e-02 7.94354141e-01 -1.77017137e-01 -6.53960228e-01
4.53072697e-01 -5.46305776e-01 -9.41268623e-01 5.33014596e-01
8.66008878e-01 -2.74948597e-01 -3.40037286e-01 1.14181733e+00
3.31193030e-01 -2.02290088e-01 5.86843967e-01 -1.63826120e+00
-1.07772338e+00 4.55069005e-01 -7.26828158e-01 -1.08396567e-01
1.98155582e-01 2.68031001e-01 9.26890612e-01 -8.09169888e-01
7.11338639e-01 7.73012936e-01 3.92022938e-01 6.26474619e-01
-1.30417609e+00 -6.72495604e-01 -3.28706913e-02 -1.31261691e-01
-1.29988945e+00 -3.01972657e-01 9.84941602e-01 -4.30178642e-01
-8.82081091e-02 6.61579072e-02 6.58722892e-02 1.43310738e+00
-4.37176496e-01 1.03489566e+00 1.33984232e+00 -3.85212451e-01
2.46166602e-01 4.39023376e-01 -1.47542983e-01 4.82606888e-01
2.84742080e-02 -1.35583624e-01 -3.54897767e-01 -8.47034752e-02
5.61374128e-01 4.43227915e-03 -2.66300738e-01 -5.81016719e-01
-1.53414845e+00 7.12580323e-01 6.30196154e-01 -7.66952708e-02
-3.19250941e-01 1.90796062e-01 2.95479149e-01 -5.17876446e-02
4.87990171e-01 1.56617373e-01 -5.49692869e-01 1.61312580e-01
-1.02077067e+00 1.29556693e-02 4.89142746e-01 1.25593066e+00
8.25312912e-01 -3.32430094e-01 -1.76010206e-01 8.18267286e-01
1.32236302e-01 1.00123787e+00 3.48936737e-01 -8.32204044e-01
8.55720341e-01 1.71275437e-01 3.57723504e-01 -8.44767272e-01
6.26352876e-02 -3.25000823e-01 -1.03827369e+00 -2.09422022e-01
7.12058008e-01 2.33664382e-02 -1.00350463e+00 2.15609264e+00
3.09371591e-01 3.64152223e-01 2.80452073e-01 1.05407989e+00
7.25639641e-01 6.56820416e-01 2.87080646e-01 -1.16786294e-01
1.39537048e+00 -1.36250770e+00 -6.45341635e-01 -5.43663800e-01
5.00829816e-01 -6.60520554e-01 1.27241373e+00 -1.59344673e-02
-7.44971573e-01 -5.20952225e-01 -7.18456745e-01 -9.93289798e-02
-2.81020105e-01 2.37429455e-01 4.15452480e-01 2.51373023e-01
-8.46670210e-01 -5.31212147e-03 -5.37487149e-01 -2.17300802e-01
7.84634829e-01 1.69464827e-01 -7.99220085e-01 -5.17268240e-01
-1.01377809e+00 5.28483689e-01 5.11593699e-01 -4.06277850e-02
-1.01579833e+00 -6.79444611e-01 -1.04170215e+00 -1.49489880e-01
4.63271558e-01 -5.42871892e-01 1.13210404e+00 -1.28253531e+00
-9.22633469e-01 1.29678106e+00 -2.82503106e-02 -4.95500833e-01
7.14529991e-01 -9.61212739e-02 -1.05269723e-01 5.07014394e-01
5.07188499e-01 1.19230950e+00 8.15743804e-01 -1.82771933e+00
-3.37665230e-01 -2.78159678e-01 -8.42092857e-02 2.62637556e-01
-3.89849275e-01 -2.11460173e-01 -8.20900142e-01 -6.72988772e-01
-2.35253781e-01 -1.20058882e+00 -2.75007933e-01 3.22661042e-01
-5.73835969e-01 1.69630587e-01 8.83188426e-01 -7.24956453e-01
4.39050406e-01 -2.14397669e+00 -7.91812316e-02 -1.18004233e-01
9.29051787e-02 2.13170543e-01 -4.95146871e-01 1.41327813e-01
-2.13737562e-01 8.92358124e-02 -6.85549200e-01 -6.34023011e-01
-8.88246149e-02 3.18222404e-01 -4.97997701e-01 4.80854392e-01
3.24923784e-01 1.33848274e+00 -1.23791575e+00 -1.04403138e+00
1.77849829e-01 4.19238180e-01 -2.37664521e-01 6.20384455e-01
-3.49205166e-01 7.33540297e-01 -4.39539164e-01 7.48209238e-01
7.69304872e-01 -7.97802806e-01 -4.22915723e-03 -4.09775138e-01
5.11948168e-01 -4.78323996e-01 -5.92076480e-01 1.92223930e+00
-6.00446939e-01 5.35701394e-01 1.00905888e-01 -1.15437269e+00
7.85092592e-01 3.51482391e-01 5.07557988e-01 -7.49412537e-01
6.47087917e-02 1.87981963e-01 -6.57234848e-01 -6.19841993e-01
2.39557683e-01 -2.95043081e-01 -5.02865314e-01 6.18498445e-01
2.47555286e-01 -1.11857265e-01 3.16220112e-02 4.66821998e-01
9.17329371e-01 3.67481232e-01 1.00478135e-01 -1.17018037e-02
3.37927938e-01 9.78628322e-02 3.86633337e-01 7.54147828e-01
-4.75443423e-01 1.12597704e+00 3.24110717e-01 -1.15538254e-01
-1.34982657e+00 -1.20653093e+00 -1.99536145e-01 8.30273926e-01
4.77525830e-01 9.15166810e-02 -7.75579214e-01 -1.15356779e+00
-2.07525998e-01 4.16193306e-01 -8.58284771e-01 8.54117051e-02
-3.40751499e-01 -6.13425493e-01 3.94634455e-01 4.44658905e-01
6.31127000e-01 -1.14193726e+00 1.02837935e-01 -1.71123639e-01
-6.30418181e-01 -1.78703058e+00 -8.07298660e-01 1.61507741e-01
-4.19087201e-01 -1.02692509e+00 -8.66895676e-01 -1.35832405e+00
1.15902853e+00 3.50643128e-01 1.14940512e+00 -2.16138110e-01
1.53780896e-02 4.84156400e-01 -4.58281398e-01 -1.33030191e-01
-4.53012794e-01 -9.21203662e-03 -1.32184684e-01 3.87353897e-01
1.37306109e-01 -4.27783936e-01 -6.48301125e-01 4.18249547e-01
-1.26937044e+00 3.37398648e-01 8.61500680e-01 1.14603949e+00
9.89634216e-01 -3.15118670e-01 8.33131492e-01 -1.04841888e+00
1.34640336e-01 -8.66264999e-01 -5.41879535e-01 5.08401215e-01
-2.06926748e-01 -5.07524610e-03 8.51122200e-01 -5.85401535e-01
-9.68962014e-01 6.99323177e-01 2.37610310e-01 -6.95556283e-01
-3.33074391e-01 2.91002214e-01 -3.96449149e-01 5.03902277e-03
4.79397029e-01 4.75929469e-01 3.18505645e-01 -2.27468669e-01
6.26808286e-01 8.99699509e-01 1.21225977e+00 -7.97868013e-01
1.32862914e+00 7.19793737e-01 -2.83314705e-01 -3.92410606e-01
-1.24949181e+00 -5.16098142e-01 -6.20107830e-01 -2.04791322e-01
1.11124992e+00 -1.30977547e+00 -3.18961829e-01 3.89284730e-01
-1.17701900e+00 -3.04880112e-01 -2.55861193e-01 3.64941835e-01
-8.15306544e-01 4.45240229e-01 -4.53956127e-01 -4.87435758e-01
-2.87286907e-01 -1.20052314e+00 1.47525036e+00 1.46397769e-01
2.71041662e-01 -1.04437876e+00 3.10799181e-02 1.06885231e+00
1.03235111e-01 5.79798460e-01 4.72831190e-01 -8.68063390e-01
-6.51126325e-01 -3.74494135e-01 -6.16723776e-01 6.47353292e-01
1.00855380e-01 -4.85980570e-01 -9.50125575e-01 -2.06624925e-01
-1.84273779e-01 -1.08040893e+00 4.93293434e-01 -2.85579953e-02
1.40400553e+00 -2.77983248e-01 -2.36730471e-01 3.86315852e-01
1.56659770e+00 -2.67732501e-01 5.27658224e-01 1.66888714e-01
8.71293604e-01 7.20087290e-01 9.63727713e-01 1.81488574e-01
4.03480113e-01 6.30476713e-01 6.78581476e-01 -4.46881711e-01
-1.46659195e-01 -5.61539948e-01 2.38996565e-01 6.60903692e-01
6.09009445e-01 -5.38341761e-01 -8.31807733e-01 8.20982516e-01
-2.01882577e+00 -7.99375415e-01 9.22044367e-02 2.05774689e+00
1.17830122e+00 -1.66436091e-01 -2.70577762e-02 -3.62393171e-01
1.19779503e+00 5.10052703e-02 -6.93081737e-01 4.15128112e-01
-2.63946623e-01 -9.85508487e-02 6.26589477e-01 2.36629859e-01
-1.39731538e+00 9.23572302e-01 5.57624245e+00 8.73131633e-01
-9.26586866e-01 2.83084214e-01 1.13744926e+00 -2.63257995e-02
-2.97758222e-01 -1.27356216e-01 -3.73654902e-01 8.97949100e-01
8.75446677e-01 5.62337637e-02 3.22681397e-01 9.82908249e-01
-6.19491413e-02 1.19323649e-01 -1.27580273e+00 1.26843858e+00
4.71640706e-01 -1.36049330e+00 1.19622841e-01 -2.72610952e-04
9.65892255e-01 2.84308106e-01 1.64220512e-01 1.49305031e-01
3.33152294e-01 -9.92866397e-01 8.79461586e-01 1.99642792e-01
1.14091229e+00 -3.88721108e-01 9.36388671e-01 3.33024025e-01
-8.65588725e-01 2.07686841e-01 4.43294868e-02 3.73968720e-01
4.69277799e-01 5.72015643e-01 -5.08539498e-01 5.67520380e-01
6.52083933e-01 7.90035009e-01 -7.43226171e-01 8.19203556e-01
-2.22954482e-01 7.45403826e-01 -3.39401305e-01 3.91986161e-01
3.27884465e-01 -3.34396333e-01 1.03642955e-01 9.06207204e-01
1.66741535e-02 -7.39372447e-02 3.37641597e-01 8.71545076e-01
-4.97120023e-01 2.03262996e-02 -7.34941363e-01 -6.24087900e-02
4.56956506e-01 1.46133947e+00 -7.49263465e-01 -4.12639201e-01
-5.28102934e-01 1.10562778e+00 4.97801989e-01 4.63419974e-01
-1.25452888e+00 -8.56526941e-02 1.47410050e-01 2.02452280e-02
3.90350193e-01 1.24673277e-01 -1.03461415e-01 -1.37083542e+00
4.28166598e-01 -7.49543428e-01 3.23634267e-01 -1.17711782e+00
-1.73964512e+00 8.29875410e-01 -4.06706743e-02 -1.52832806e+00
-1.03106417e-01 -4.47754651e-01 -4.11138594e-01 5.70383012e-01
-1.73263299e+00 -1.63803887e+00 -5.87606966e-01 6.60845339e-01
3.22171628e-01 5.55070909e-03 6.87659919e-01 4.55937594e-01
-5.61952829e-01 7.28681803e-01 2.92806953e-01 5.16235352e-01
1.00735593e+00 -1.15728843e+00 -1.48386993e-02 8.59095871e-01
3.38062435e-01 1.60852373e-01 6.09838188e-01 -4.03526336e-01
-1.17044628e+00 -1.51252091e+00 5.74906826e-01 -6.80679739e-01
8.03135514e-01 -8.09186220e-01 -7.72865593e-01 8.37954938e-01
2.98490763e-01 8.22472036e-01 7.85114408e-01 -4.54885066e-01
-5.62258005e-01 -8.17879066e-02 -1.22045088e+00 3.50509107e-01
8.44007730e-01 -7.45728672e-01 -5.90807915e-01 8.55233014e-01
1.00406694e+00 -2.80690730e-01 -7.26258397e-01 2.59907842e-01
3.00838560e-01 -4.80335444e-01 1.00198841e+00 -4.48660642e-01
7.96347320e-01 -5.36472499e-01 -2.84181476e-01 -1.04073083e+00
2.24354342e-01 -6.15790606e-01 1.18949093e-01 1.49748993e+00
4.53623354e-01 -3.89241546e-01 9.73960459e-01 7.46218264e-01
-5.81088010e-03 -6.13590717e-01 -7.94257164e-01 -9.11390543e-01
4.47354652e-02 -2.87259579e-01 3.36257786e-01 1.06937122e+00
-3.75241071e-01 4.08030719e-01 -6.44685864e-01 2.97016948e-01
1.04236746e+00 4.49340284e-01 8.83319914e-01 -7.20109165e-01
-2.74769157e-01 2.82570839e-01 -2.82651395e-01 -1.08642364e+00
5.28880179e-01 -1.08324087e+00 3.96587670e-01 -1.31534660e+00
6.66067123e-01 -7.52495110e-01 -2.57905990e-01 7.04921722e-01
-3.59866142e-01 9.61334467e-01 9.85804647e-02 4.22833592e-01
-1.13137293e+00 7.44307280e-01 1.32189238e+00 -4.16081905e-01
3.76332879e-01 -3.40624452e-01 -7.63880014e-01 3.70035887e-01
6.42731547e-01 -8.05099607e-01 -5.65991879e-01 -3.40185761e-01
-3.63473035e-02 1.77549601e-01 7.05350339e-01 -6.81139648e-01
1.10170126e-01 -1.41874969e-01 1.64747879e-01 -4.16012108e-01
1.74199790e-01 -1.05267859e+00 -8.78890604e-02 6.48335442e-02
-5.60797393e-01 -3.10550481e-01 -5.58834225e-02 9.79092896e-01
-4.60121304e-01 -1.45425737e-01 7.89645612e-01 5.68919443e-02
-6.07470214e-01 6.47247136e-01 3.56802404e-01 5.10549784e-01
1.42767322e+00 1.78178884e-02 -3.94344181e-01 -5.33697426e-01
-5.99040329e-01 4.80421275e-01 6.71762764e-01 5.45596838e-01
3.96269947e-01 -1.52632558e+00 -8.27292204e-01 -9.69498456e-02
7.23239183e-01 4.60615009e-01 2.89556712e-01 6.21279120e-01
-3.80227089e-01 1.17941409e-01 -2.02730939e-01 -8.23802650e-01
-9.96881723e-01 9.64012265e-01 -9.81876701e-02 -2.81408727e-01
-3.77883613e-01 6.36012316e-01 5.10632813e-01 -6.59886718e-01
1.18910231e-01 2.10732490e-01 1.76585823e-01 -1.02233969e-01
4.30558681e-01 -3.58850688e-01 -2.70874321e-01 -8.56273413e-01
-2.95367867e-01 4.25730258e-01 -1.77761003e-01 -7.39086866e-02
1.29393518e+00 -3.28028202e-01 -7.04926401e-02 1.60618886e-01
1.71046972e+00 -3.14743638e-01 -1.61969709e+00 -4.44722116e-01
-3.33789498e-01 -4.16491896e-01 -1.91984817e-01 -8.15782309e-01
-1.24947000e+00 6.11939192e-01 3.87066305e-01 -1.87593684e-01
1.07273912e+00 5.06159782e-01 1.02162504e+00 1.92843124e-01
4.54490781e-01 -8.71077597e-01 4.15595382e-01 -1.35429472e-01
7.78605998e-01 -2.01877928e+00 -2.33576700e-01 -4.93101120e-01
-8.98237646e-01 5.86888611e-01 8.14292669e-01 -5.36269434e-02
3.88880223e-01 7.79037178e-02 2.10064664e-01 -4.25166525e-02
-5.30129015e-01 -7.56099075e-02 3.14503796e-02 7.81644821e-01
-1.32215977e-01 -4.84247990e-02 1.83336109e-01 6.07946992e-01
-1.56883430e-02 -1.38702132e-02 5.09252846e-01 8.57439697e-01
1.89475834e-01 -1.05380964e+00 -4.20320570e-01 3.44571143e-01
-2.35510141e-01 -2.05556229e-01 -2.43230551e-01 7.06236422e-01
-1.83010101e-02 8.78289878e-01 8.05613250e-02 -2.30358973e-01
-2.03607023e-01 -1.73611358e-01 1.98261797e-01 -5.32942355e-01
-1.28619120e-01 -3.12923551e-01 -1.98472321e-01 -4.49357629e-01
-7.71535158e-01 -5.61897218e-01 -1.08711791e+00 2.28122413e-01
-2.46961802e-01 2.98810273e-01 5.99285781e-01 9.89293754e-01
4.71174777e-01 9.45857242e-02 9.23185587e-01 -8.19526374e-01
-5.31690836e-01 -7.34769881e-01 -3.64267707e-01 1.11469543e+00
3.44734669e-01 -4.41037506e-01 -5.20645678e-01 6.60564184e-01] | [10.767768859863281, 1.253160834312439] |
0ed4041a-1d16-4341-9765-36153c3839b6 | speech-enhancement-modeling-towards-robust | 1305.1426 | null | http://arxiv.org/abs/1305.1426v1 | http://arxiv.org/pdf/1305.1426v1.pdf | Speech Enhancement Modeling Towards Robust Speech Recognition System | Form about four decades human beings have been dreaming of an intelligent
machine which can master the natural speech. In its simplest form, this machine
should consist of two subsystems, namely automatic speech recognition (ASR) and
speech understanding (SU). The goal of ASR is to transcribe natural speech
while SU is to understand the meaning of the transcription. Recognizing and
understanding a spoken sentence is obviously a knowledge-intensive process,
which must take into account all variable information about the speech
communication process, from acoustics to semantics and pragmatics. While
developing an Automatic Speech Recognition System, it is observed that some
adverse conditions degrade the performance of the Speech Recognition System. In
this contribution, speech enhancement system is introduced for enhancing speech
signals corrupted by additive noise and improving the performance of Automatic
Speech Recognizers in noisy conditions. Automatic speech recognition
experiments show that replacing noisy speech signals by the corresponding
enhanced speech signals leads to an improvement in the recognition accuracies.
The amount of improvement varies with the type of the corrupting noise. | ['V. M. Thakare', 'Urmila Shrawankar'] | 2013-05-07 | null | null | null | null | ['robust-speech-recognition'] | ['speech'] | [ 4.96738315e-01 3.67432415e-01 2.99139649e-01 -6.69665456e-01
-4.34892476e-01 -4.34395850e-01 6.28806531e-01 -1.75960630e-01
-3.09188604e-01 4.75281000e-01 5.75337470e-01 -4.93247837e-01
1.72488391e-01 -5.22337854e-01 -2.50458539e-01 -5.95446527e-01
4.62159932e-01 1.27635047e-01 -5.72132207e-02 -4.85524476e-01
2.54496664e-01 3.68000209e-01 -1.66601694e+00 2.69066840e-01
7.57723808e-01 6.89675927e-01 8.23913336e-01 1.08471513e+00
-5.29322505e-01 7.46466875e-01 -9.22668755e-01 2.36686364e-01
-2.09352538e-01 -5.30132115e-01 -1.18153727e+00 5.90748012e-01
-5.30855775e-01 1.40817359e-01 -1.67238355e-01 1.21568191e+00
2.19123989e-01 3.22292000e-01 4.27708715e-01 -6.53626978e-01
-5.42587519e-01 6.25015199e-01 2.63824940e-01 3.25315803e-01
5.53364694e-01 -3.84403989e-02 6.11452162e-01 -7.11140156e-01
1.63839072e-01 1.40788424e+00 -8.66332427e-02 6.92672372e-01
-8.45368922e-01 -2.74078697e-01 9.65431482e-02 1.14912435e-01
-1.16822374e+00 -9.51614678e-01 6.47568226e-01 -2.86571264e-01
1.14260995e+00 4.73751396e-01 3.26569051e-01 7.80751824e-01
3.45455781e-02 6.28450990e-01 9.43027318e-01 -8.71747077e-01
2.53518909e-01 5.76041520e-01 5.88119030e-01 1.36316776e-01
-3.36948127e-01 -6.48012664e-03 -4.14630234e-01 2.80448884e-01
2.03282073e-01 -4.36178774e-01 -5.83422720e-01 5.27893841e-01
-9.94364798e-01 4.72264379e-01 -3.16073261e-02 1.01010394e+00
-6.57514453e-01 -3.64496082e-01 3.50298643e-01 5.57370484e-01
2.92124033e-01 2.68682092e-01 -4.36475188e-01 -5.23131847e-01
-6.23338163e-01 -4.00056928e-01 9.06257212e-01 5.69384038e-01
4.80337292e-01 4.48721290e-01 2.06425294e-01 1.09191859e+00
5.64402461e-01 8.10318053e-01 8.09155107e-01 -5.92087865e-01
2.65618742e-01 6.85924709e-01 1.96178872e-02 -7.48120248e-01
-1.19007565e-01 -3.81753623e-01 -6.05896294e-01 -7.36667365e-02
2.33990654e-01 -1.69530258e-01 -1.16078579e+00 1.35105169e+00
1.90410376e-01 -2.36715153e-01 7.64097452e-01 8.09463143e-01
9.44608808e-01 1.29790747e+00 2.13944659e-01 -6.09935284e-01
1.58000565e+00 -6.71604872e-01 -1.33048677e+00 -5.13844550e-01
4.73390013e-01 -1.06895137e+00 1.11329949e+00 4.54126924e-01
-8.58611286e-01 -7.40822732e-01 -1.01626503e+00 3.51209044e-02
-3.58592212e-01 2.26656362e-01 5.90036511e-02 9.47841823e-01
-9.55088258e-01 6.22628890e-02 -6.15885079e-01 -4.33272272e-01
-3.08568716e-01 3.87219518e-01 -4.90399241e-01 1.24917947e-01
-1.33053231e+00 1.01308954e+00 5.32910883e-01 4.09400105e-01
-6.14888847e-01 -3.10135167e-02 -8.48449945e-01 2.52960026e-01
3.01245689e-01 -2.22624347e-01 1.58164251e+00 -1.45287549e+00
-1.79939640e+00 6.50284648e-01 -6.96462691e-01 -3.37581098e-01
-6.35762066e-02 -5.95001429e-02 -9.80705738e-01 4.18047234e-02
-2.79827505e-01 -4.66263033e-02 8.86183798e-01 -1.10683286e+00
-6.20483756e-01 -5.76335430e-01 -4.64626104e-01 3.77260715e-01
1.03657355e-03 3.24297220e-01 -1.00186191e-01 -4.21006829e-01
4.02931184e-01 -6.29392743e-01 1.16985641e-01 -7.65004575e-01
-9.16273072e-02 -3.02907825e-01 1.07058418e+00 -9.83863294e-01
1.11472833e+00 -2.52543283e+00 -1.70519799e-02 1.43486932e-01
-3.94877642e-01 6.98860645e-01 4.96045388e-02 3.89245331e-01
-2.08078563e-01 -6.51967153e-02 -1.38076201e-01 -2.52839595e-01
-9.75528806e-02 6.39580846e-01 -3.91238540e-01 7.39524364e-02
5.98518848e-02 4.47288036e-01 -5.39165914e-01 -9.86188464e-03
4.29346472e-01 6.66208148e-01 -5.94201125e-03 4.64024127e-01
-8.74728337e-02 6.12068892e-01 -4.62967753e-01 2.55216151e-01
2.86912560e-01 1.87367275e-01 9.98312086e-02 1.54830396e-01
-2.34426513e-01 8.02329540e-01 -1.21948719e+00 1.19502437e+00
-5.34793317e-01 7.50994742e-01 5.89189529e-01 -1.23371732e+00
1.24754775e+00 1.01404655e+00 -1.78695202e-01 -8.03045571e-01
3.20573628e-01 3.81535769e-01 3.08962643e-01 -8.49826157e-01
4.72181052e-01 -4.46013689e-01 2.39001229e-01 2.22129554e-01
-8.98474529e-02 -2.48667598e-01 -1.70757160e-01 5.87477256e-03
6.66053712e-01 -6.54429078e-01 5.22817671e-01 -2.44565845e-01
1.13742507e+00 -1.89216778e-01 2.93590009e-01 3.56345415e-01
-2.13406667e-01 1.25001222e-01 -2.93962285e-02 -1.49092257e-01
-8.55262339e-01 -7.47668505e-01 1.36174662e-02 9.52533007e-01
-1.03044249e-01 1.96187776e-02 -9.75680947e-01 3.65555473e-03
-6.12517238e-01 1.14778471e+00 3.51809518e-04 -1.64277136e-01
-4.30750042e-01 -4.63100463e-01 3.35739523e-01 2.62041926e-01
5.29425681e-01 -1.40683663e+00 -1.86631680e-01 3.45277160e-01
-2.09619224e-01 -1.16804636e+00 -2.56222725e-01 1.56352907e-01
-6.27433956e-01 -6.55286431e-01 -4.91147578e-01 -9.44304228e-01
5.04920661e-01 3.34605217e-01 6.23817086e-01 1.75790116e-01
2.52567232e-02 3.85904431e-01 -7.33647227e-01 -5.74533641e-01
-1.36652267e+00 -7.34589174e-02 2.40224093e-01 2.05656454e-01
4.49871331e-01 -3.88728619e-01 -1.04494490e-01 3.07799906e-01
-9.75058436e-01 -6.54021427e-02 6.51897252e-01 6.00078404e-01
7.94765279e-02 5.41437149e-01 7.83146918e-01 -5.43726861e-01
8.49087298e-01 -1.04700305e-01 -2.02782840e-01 2.79027373e-01
-3.11694205e-01 1.49033412e-01 6.10358417e-01 -1.82501763e-01
-1.53354347e+00 -2.67977603e-02 -6.30288303e-01 3.71397883e-01
-8.49063456e-01 4.78561997e-01 -6.07715011e-01 3.33824009e-01
5.35137177e-01 7.93614864e-01 2.65534222e-01 -6.24217629e-01
1.08504817e-01 1.68054771e+00 5.55127978e-01 -2.29064673e-01
3.46719086e-01 -5.94901331e-02 -5.70929110e-01 -1.78066564e+00
-6.51019514e-01 -8.63389194e-01 -4.99476254e-01 -2.98832148e-01
9.10075128e-01 -6.02416277e-01 -5.00193536e-01 6.51901186e-01
-1.35022426e+00 -5.29712662e-02 1.54197533e-02 6.90766752e-01
-2.96957344e-01 3.87118489e-01 -2.97019839e-01 -1.38507652e+00
-3.89996648e-01 -1.39227331e+00 8.64112914e-01 3.31692100e-01
-2.89199740e-01 -8.24534893e-01 -2.16391578e-01 8.57962787e-01
4.06938195e-01 -7.64898717e-01 8.55984926e-01 -9.28393602e-01
-2.27079347e-01 -1.62693113e-01 1.62196964e-01 1.07043517e+00
5.85187733e-01 -6.55659139e-02 -1.36450303e+00 7.44966716e-02
4.92882937e-01 7.97511414e-02 3.62813950e-01 3.10086548e-01
5.54890275e-01 -4.07599419e-01 8.96717161e-02 -1.10619083e-01
9.69595194e-01 1.05973637e+00 8.68862569e-01 -1.28181472e-01
1.61615431e-01 8.69004965e-01 4.09161925e-01 -5.04580252e-02
2.39576232e-02 4.99779850e-01 -1.12215362e-01 -2.94851456e-02
-1.04423128e-01 2.24721851e-03 6.73887730e-01 1.40735650e+00
2.12271631e-01 -3.71290147e-01 -1.00310373e+00 5.58219433e-01
-1.48433292e+00 -8.74632716e-01 -2.12540835e-01 1.99395657e+00
7.01231480e-01 1.33418441e-01 -2.93594807e-01 7.20759332e-01
7.54768729e-01 1.40630215e-01 -1.51538461e-01 -8.73747289e-01
-5.20294905e-02 -8.57560486e-02 -6.62731379e-02 1.17360687e+00
-6.88016832e-01 9.53555226e-01 6.30114889e+00 6.44724965e-01
-1.29283690e+00 -1.41993016e-01 3.51165354e-01 6.07779205e-01
-1.97294340e-01 -1.81342229e-01 -4.99964595e-01 2.88141042e-01
1.25862706e+00 -1.80534318e-01 5.86881280e-01 6.24547660e-01
8.54280949e-01 -3.84831995e-01 -7.53662586e-01 8.27131391e-01
1.10171631e-01 -8.12424839e-01 1.72529325e-01 -2.85189092e-01
2.06281468e-01 -1.78620666e-01 -2.06151158e-01 1.82408378e-01
-1.17430590e-01 -1.17989695e+00 5.97517133e-01 4.48573142e-01
1.47336513e-01 -7.12706029e-01 9.19519067e-01 8.48561049e-01
-9.82518852e-01 6.96353838e-02 4.28837948e-02 -4.10386473e-01
2.85142362e-01 5.01323402e-01 -1.20538986e+00 3.49075884e-01
4.02180523e-01 1.62112013e-01 -1.39397591e-01 5.51939487e-01
-3.88781279e-01 1.00093126e+00 -3.23586643e-01 -2.73012698e-01
2.08523780e-01 -1.63467780e-01 8.05884004e-01 1.15715468e+00
1.60748735e-01 7.07593620e-01 -1.28309786e-01 5.79530180e-01
2.96313614e-01 4.82762367e-01 -5.35683990e-01 -5.31525671e-01
4.18146133e-01 7.43222535e-01 -5.15666068e-01 -4.68399614e-01
-2.60459274e-01 7.31520057e-01 -4.47618574e-01 4.69818562e-01
-1.44944951e-01 -3.14519674e-01 6.24795258e-01 7.43390471e-02
-1.63374953e-02 -3.38162661e-01 -2.91051418e-01 -8.41203749e-01
1.73196763e-01 -1.07064629e+00 -8.63453001e-02 -9.60666895e-01
-7.59132028e-01 8.51272941e-01 -3.29376966e-01 -6.48862183e-01
-4.54783648e-01 -5.47819376e-01 -4.05261248e-01 1.20929003e+00
-1.18014860e+00 -6.14677310e-01 -1.91848390e-02 2.69034177e-01
1.16750872e+00 -2.59622246e-01 8.83352101e-01 1.89818710e-01
-3.94384831e-01 5.96901216e-03 -1.09021470e-01 1.09103821e-01
1.88650906e-01 -9.51383948e-01 1.86097041e-01 9.24917758e-01
2.97360599e-01 6.28113925e-01 1.11927891e+00 -5.05689442e-01
-1.19003069e+00 -4.86959517e-01 1.33700573e+00 -3.25450748e-02
4.89103109e-01 -1.74776822e-01 -1.19955432e+00 4.63559449e-01
2.70493746e-01 -7.30236709e-01 6.19013369e-01 -4.91205677e-02
1.16088413e-01 -1.71222925e-01 -9.09191787e-01 3.73644441e-01
3.85090679e-01 -8.76844585e-01 -1.20681512e+00 1.15117520e-01
8.37058783e-01 -4.10664901e-02 -5.71775913e-01 8.30919221e-02
1.73757806e-01 -6.42469347e-01 5.30200601e-01 -4.65477824e-01
-1.70452461e-01 -5.34154058e-01 -3.13583046e-01 -1.36677110e+00
2.71214992e-01 -6.65129125e-01 5.36521256e-01 1.18842363e+00
6.93259537e-01 -7.50650942e-01 3.99368584e-01 7.98262000e-01
-2.20478565e-01 1.91948339e-02 -8.76107037e-01 -5.39600551e-01
-3.13572884e-01 -6.94076002e-01 3.12562585e-01 6.99128091e-01
3.57252330e-01 7.67560542e-01 -1.59273058e-01 5.38119256e-01
2.34402433e-01 -3.98749024e-01 5.32701612e-01 -1.00794399e+00
-1.51737615e-01 -1.64885163e-01 -4.68309224e-01 -1.25126886e+00
9.15936157e-02 -3.83837759e-01 4.93232638e-01 -1.59156835e+00
-2.93032229e-01 1.07339002e-01 1.85123011e-01 2.03451097e-01
-8.73262584e-02 -5.97843707e-01 6.63108900e-02 9.13858637e-02
1.84657949e-03 7.28940606e-01 1.20007479e+00 -1.05973154e-01
-4.32769001e-01 4.25703824e-01 -4.02194023e-01 7.73140728e-01
7.97973812e-01 -2.91144699e-01 -6.56137049e-01 -3.37563157e-01
-2.13348880e-01 3.58187377e-01 5.83928153e-02 -8.20989013e-01
3.59546721e-01 -2.45182216e-02 -2.61373401e-01 -4.64331269e-01
4.63550389e-01 -9.17526484e-01 2.68699855e-01 3.54104728e-01
-4.77620542e-01 -1.29176617e-01 2.37421960e-01 3.15260261e-01
-7.24308848e-01 -5.74342549e-01 8.95179808e-01 -6.15709722e-02
-9.65361655e-01 -5.29894590e-01 -9.75910485e-01 -3.15537840e-01
7.77856767e-01 -4.21979487e-01 -1.28779843e-01 -7.41773784e-01
-1.07166350e+00 -5.35262711e-02 -9.75977480e-02 5.51757038e-01
6.05914950e-01 -5.99755406e-01 -4.54870284e-01 5.30722380e-01
-1.80918768e-01 -7.99927935e-02 2.83347815e-01 5.39323449e-01
-4.01239127e-01 7.73614645e-01 2.09761426e-01 -3.30157965e-01
-1.60375357e+00 3.05120021e-01 4.54647332e-01 2.49206707e-01
-3.26351225e-01 6.22630239e-01 3.83723974e-02 -3.26100230e-01
5.22395849e-01 -4.96859163e-01 -5.50851285e-01 -2.46097147e-01
8.80334854e-01 1.45199671e-01 2.77493715e-01 -1.08533394e+00
-2.66204953e-01 2.38745898e-01 -7.54291425e-03 -4.21222776e-01
1.18060982e+00 -5.26786745e-01 -1.73856393e-01 6.97657228e-01
9.59314585e-01 -1.46780729e-01 -4.23922300e-01 -1.36013836e-01
1.91261664e-01 -2.11378202e-01 4.16996807e-01 -9.78652894e-01
-6.51505649e-01 9.04722750e-01 4.67573345e-01 6.49883926e-01
1.18258190e+00 3.85971600e-03 7.16449380e-01 6.21422887e-01
2.52396673e-01 -1.31153321e+00 -2.95915335e-01 6.99174881e-01
8.97036672e-01 -1.29583681e+00 -4.95007694e-01 -5.64508200e-01
-8.70003402e-01 9.82796073e-01 1.54936135e-01 3.66655499e-01
8.11448097e-01 2.11274073e-01 4.92561519e-01 -1.85721368e-01
-6.74450815e-01 -3.19446892e-01 3.28838468e-01 4.47000951e-01
4.66573477e-01 1.58018634e-01 -3.53459865e-01 5.62646508e-01
-4.33361858e-01 -7.88141936e-02 4.67036992e-01 6.86697423e-01
-1.09289384e+00 -1.07647467e+00 -7.22809136e-01 -6.21611960e-02
-4.29604560e-01 8.84293765e-03 -6.48847759e-01 4.41015661e-01
-1.12219803e-01 1.63220489e+00 -3.10897678e-02 -2.63192058e-01
5.77415347e-01 4.32909727e-01 -1.48986623e-01 -8.90834033e-01
-4.79350686e-01 1.21516846e-01 4.37422872e-01 -1.22135445e-01
-6.20077968e-01 -4.99458790e-01 -1.54352510e+00 -2.01236717e-02
-3.77930999e-01 6.57823980e-01 1.22960353e+00 1.53000581e+00
-6.00671992e-02 5.28488517e-01 6.14334762e-01 -1.83044776e-01
-5.62521994e-01 -1.12950945e+00 -6.40841305e-01 2.57012963e-01
6.01269245e-01 -1.41664132e-01 -6.02338612e-01 3.69142205e-01] | [14.38803768157959, 6.36789083480835] |
b4d36a1d-26d9-445f-945e-0bc93154ecc8 | polynet-polynomial-neural-network-for-3d | 2110.07882 | null | https://arxiv.org/abs/2110.07882v1 | https://arxiv.org/pdf/2110.07882v1.pdf | PolyNet: Polynomial Neural Network for 3D Shape Recognition with PolyShape Representation | 3D shape representation and its processing have substantial effects on 3D shape recognition. The polygon mesh as a 3D shape representation has many advantages in computer graphics and geometry processing. However, there are still some challenges for the existing deep neural network (DNN)-based methods on polygon mesh representation, such as handling the variations in the degree and permutations of the vertices and their pairwise distances. To overcome these challenges, we propose a DNN-based method (PolyNet) and a specific polygon mesh representation (PolyShape) with a multi-resolution structure. PolyNet contains two operations; (1) a polynomial convolution (PolyConv) operation with learnable coefficients, which learns continuous distributions as the convolutional filters to share the weights across different vertices, and (2) a polygonal pooling (PolyPool) procedure by utilizing the multi-resolution structure of PolyShape to aggregate the features in a much lower dimension. Our experiments demonstrate the strength and the advantages of PolyNet on both 3D shape classification and retrieval tasks compared to existing polygon mesh-based methods and its superiority in classifying graph representations of images. The code is publicly available from https://myavartanoo.github.io/polynet/. | ['Kyoung Mu Lee', 'Yue Zhang', 'Reyhaneh Neshatavar', 'Shih-Hsuan Hung', 'Mohsen Yavartanoo'] | 2021-10-15 | null | null | null | null | ['3d-shape-retrieval', '3d-object-classification', '3d-shape-recognition', '3d-shape-representation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-2.45078444e-01 -3.17005455e-01 5.71313910e-02 -2.80846924e-01
-3.55973899e-01 -4.04839098e-01 5.41777551e-01 2.56085426e-01
-1.27568752e-01 1.36802316e-01 1.09098658e-01 -2.51462251e-01
-9.07124504e-02 -1.46742857e+00 -8.86505008e-01 -6.32215321e-01
-2.82770187e-01 6.47666335e-01 2.87454158e-01 -2.17583433e-01
3.46000701e-01 1.35004795e+00 -1.50647330e+00 4.50779527e-01
4.88915890e-01 1.36068952e+00 1.45588377e-02 4.07797694e-01
-7.02102542e-01 -4.28719223e-02 -2.92193174e-01 -2.55103886e-01
4.78719026e-01 4.81481344e-01 -6.48001552e-01 -1.47141591e-01
6.25407040e-01 -1.30999357e-01 -6.99827909e-01 9.56556082e-01
7.33545959e-01 1.99930042e-01 8.72335672e-01 -1.11039853e+00
-1.19030106e+00 3.09757978e-01 -7.78533340e-01 -4.38413536e-03
1.44483209e-01 2.46730614e-02 7.77317584e-01 -1.38407242e+00
5.38713634e-01 1.74773753e+00 9.12708580e-01 4.35952023e-02
-7.77156830e-01 -7.03207374e-01 -5.01015820e-02 -1.60654396e-01
-1.74842668e+00 2.26768795e-02 8.75348747e-01 -3.93280804e-01
1.14588273e+00 2.60971427e-01 7.99261808e-01 6.47485018e-01
2.80552745e-01 6.63802028e-01 6.33833766e-01 -7.18622357e-02
3.92056182e-02 -4.19548392e-01 9.84169766e-02 9.62754726e-01
3.89416158e-01 -1.65724441e-01 -1.18991017e-01 -3.90948534e-01
1.56707788e+00 2.88389355e-01 -2.32374445e-02 -4.60763365e-01
-9.68819559e-01 7.24240124e-01 9.71882641e-01 2.32952997e-01
-3.55394930e-01 3.31484348e-01 4.97827172e-01 1.87085405e-01
7.22251534e-01 2.40140885e-01 -3.23904097e-01 2.89817870e-01
-4.95529920e-01 5.13340175e-01 7.24186957e-01 1.18227935e+00
7.87085831e-01 2.60304332e-01 -2.55308896e-01 1.05109513e+00
2.71238238e-01 6.21936381e-01 2.25296229e-01 -4.21651632e-01
5.50576150e-01 1.10525608e+00 -3.22784126e-01 -1.54101098e+00
-5.92005253e-01 -1.70426682e-01 -1.09486318e+00 3.47469866e-01
7.52863428e-03 3.10447544e-01 -1.21084547e+00 1.25779605e+00
4.16817129e-01 1.68186650e-01 -2.39464119e-01 8.34635973e-01
1.65492809e+00 7.43914247e-01 -7.65185878e-02 5.70680797e-01
1.35830545e+00 -7.57204890e-01 -2.36472264e-01 2.45626435e-01
3.62807870e-01 -7.42439866e-01 8.93706739e-01 -9.91225988e-02
-1.14141238e+00 -6.23404741e-01 -1.03612399e+00 -4.49808478e-01
-9.08438206e-01 1.88902572e-01 9.01979268e-01 4.33379024e-01
-1.19580853e+00 7.05457330e-01 -6.75039947e-01 -1.31623372e-01
9.75467622e-01 6.29225254e-01 -4.85582829e-01 -1.32144183e-01
-1.09374905e+00 5.30401170e-01 3.37617368e-01 2.97600091e-01
-3.15418422e-01 -5.77005684e-01 -1.24178040e+00 2.49104068e-01
-1.28622837e-02 -6.54920280e-01 7.21296787e-01 -5.11445284e-01
-1.31542003e+00 1.01518261e+00 4.05458398e-02 1.73164215e-02
2.27016330e-01 1.09185874e-01 -1.03349499e-02 2.85034105e-02
-1.48818851e-01 6.26684308e-01 7.48895466e-01 -1.03265989e+00
-1.92102134e-01 -5.18507123e-01 7.71794245e-02 4.86392856e-01
1.55852363e-02 -1.53465092e-01 -8.83971572e-01 -8.45290542e-01
5.04460037e-01 -6.76876247e-01 -1.95831075e-01 4.80529368e-01
-2.07102403e-01 -7.37753034e-01 8.05734813e-01 -3.99211615e-01
9.58042979e-01 -2.19742918e+00 2.10957453e-02 4.92436796e-01
5.40659249e-01 4.64301825e-01 -4.17610466e-01 5.40553212e-01
-1.21976204e-01 3.35379183e-01 -2.22868368e-01 -2.60163218e-01
1.44115593e-02 2.92738736e-01 -1.17202424e-01 4.18623388e-01
5.15755594e-01 1.12914264e+00 -7.46424973e-01 -2.91231424e-01
3.10301006e-01 7.99014151e-01 -4.15988982e-01 -1.33498043e-01
-7.25127682e-02 -1.41300894e-02 -6.33444011e-01 8.68994713e-01
1.35014713e+00 -2.58505076e-01 -2.12971941e-01 -4.40263063e-01
-9.32430550e-02 1.07860379e-01 -1.37547159e+00 1.33505881e+00
-1.52378961e-01 1.73736945e-01 -8.48921109e-03 -9.62833047e-01
1.40645611e+00 5.14271110e-02 4.16620344e-01 -4.68835920e-01
2.85804749e-01 3.74270856e-01 -3.22366983e-01 -2.02271268e-01
5.60585737e-01 3.10882926e-01 1.56923272e-02 1.93298608e-01
6.44489145e-03 -3.36752504e-01 -1.96628973e-01 -2.28125736e-01
8.54167581e-01 -7.36066177e-02 8.77069756e-02 -3.97924334e-01
4.45674568e-01 -3.10337871e-01 4.02240098e-01 6.65463269e-01
7.35946223e-02 8.54537249e-01 4.80651498e-01 -1.02468991e+00
-1.08249319e+00 -1.09243500e+00 -2.60317534e-01 6.05787635e-01
1.34419665e-01 -3.41312230e-01 -2.82760531e-01 -3.71159256e-01
6.25805914e-01 -4.35688235e-02 -6.34871840e-01 1.04355782e-01
-7.95014143e-01 -4.47532564e-01 6.01938903e-01 7.23706186e-01
7.52699852e-01 -1.15287995e+00 -1.22960219e-02 -1.02039203e-01
5.21016777e-01 -9.28329825e-01 -5.79429865e-01 -4.63388711e-02
-9.16749477e-01 -9.97457087e-01 -6.83604419e-01 -1.16670227e+00
8.86010528e-01 4.55637544e-01 1.05330789e+00 4.23273325e-01
-2.39126667e-01 3.22989523e-01 -1.60024494e-01 -5.33559501e-01
2.26803094e-01 3.55984196e-02 -1.71710122e-02 -6.01099245e-02
3.49461496e-01 -7.72460341e-01 -5.20718575e-01 7.38897771e-02
-9.43488181e-01 1.32348463e-01 5.49402654e-01 6.20921552e-01
1.17070460e+00 1.09761447e-01 3.32141876e-01 -9.21539783e-01
7.50788510e-01 -5.41040480e-01 -6.60059631e-01 1.26133831e-02
-9.89712924e-02 -2.32145071e-01 4.51237768e-01 -2.95351654e-01
-5.03780186e-01 4.88742664e-02 -3.59113485e-01 -9.15625334e-01
3.09659378e-03 6.35477126e-01 -6.31349608e-02 -5.04382968e-01
3.06536943e-01 2.38408908e-01 8.25623944e-02 -6.36833191e-01
3.03535879e-01 5.27287304e-01 1.05541624e-01 -8.18369925e-01
8.52635562e-01 4.92057234e-01 3.37809414e-01 -1.10429299e+00
-3.48674476e-01 -2.39143878e-01 -6.97858632e-01 1.00625262e-01
8.59483123e-01 -9.02727008e-01 -7.16460228e-01 7.68769801e-01
-1.32545185e+00 -1.66514263e-01 -1.58484891e-01 1.62222907e-01
-4.24893826e-01 3.94447833e-01 -7.34449744e-01 -5.38794756e-01
-6.86683655e-01 -1.11404788e+00 1.22227228e+00 4.13926959e-01
2.33289123e-01 -8.23981583e-01 -2.75266081e-01 -1.19282752e-01
5.55190682e-01 5.68865955e-01 1.33116734e+00 -5.51416695e-01
-7.18446016e-01 -2.77622700e-01 -7.88385808e-01 1.66577980e-01
5.42873852e-02 8.89132097e-02 -6.37378275e-01 -2.51351148e-01
-3.19015443e-01 -9.22150314e-02 7.66367614e-01 4.31955993e-01
1.79810202e+00 -1.67392164e-01 -2.09282741e-01 9.65493679e-01
1.44627631e+00 6.07280545e-02 5.93590617e-01 3.03739253e-02
9.85402107e-01 1.49726659e-01 -2.23323740e-02 2.26929680e-01
4.37945843e-01 5.19762218e-01 3.96391243e-01 -2.95837045e-01
-3.14563811e-01 -2.83649445e-01 -2.28216052e-01 1.02040219e+00
-4.26071525e-01 -1.29144341e-01 -1.03834558e+00 3.56477231e-01
-1.77661729e+00 -5.58093727e-01 -2.02411488e-01 2.02445889e+00
4.80779201e-01 1.01740859e-01 -1.99014515e-01 -1.94550261e-01
8.69601369e-01 4.57716137e-01 -5.47922909e-01 -7.15352416e-01
-3.38106573e-01 6.19123399e-01 4.57789660e-01 1.78065434e-01
-1.31353426e+00 9.87170637e-01 5.76482487e+00 1.09595466e+00
-1.19237995e+00 -3.28917652e-01 4.77610260e-01 4.05283660e-01
-3.36686015e-01 -5.06163239e-01 -6.80742443e-01 2.30708838e-01
1.56210989e-01 -1.34751629e-02 3.59893739e-01 8.90821576e-01
-1.74097136e-01 3.90048981e-01 -8.72424662e-01 1.42875552e+00
1.50783867e-01 -1.66456163e+00 7.77870715e-01 -1.29045518e-02
6.55122459e-01 3.89835417e-01 9.16869566e-02 2.92275518e-01
1.56252515e-02 -1.29256260e+00 5.01708210e-01 5.59997499e-01
1.07823455e+00 -8.63529325e-01 6.44961298e-01 -5.71068041e-02
-1.73816133e+00 3.26528102e-01 -9.72984016e-01 -8.68875682e-02
-2.57033139e-01 5.15469372e-01 -4.25554246e-01 7.30270982e-01
6.76344872e-01 8.47140789e-01 -4.10484225e-01 1.24442589e+00
-6.79521039e-02 1.38891712e-01 -5.44265270e-01 -1.22822553e-01
3.93770397e-01 -4.84519601e-01 6.31947517e-01 1.31749225e+00
4.10218179e-01 3.13563406e-01 3.49798352e-01 1.08201182e+00
-5.05415559e-01 2.89809406e-01 -9.80736136e-01 -1.02571979e-01
5.97770154e-01 1.15541220e+00 -9.14381921e-01 -2.59189963e-01
-5.06485581e-01 6.97050333e-01 5.43883026e-01 3.48028451e-01
-4.40668285e-01 -6.22859120e-01 7.74251282e-01 1.46893188e-01
6.95284188e-01 -7.14476705e-01 -4.64164615e-01 -7.98141539e-01
5.26676849e-02 -5.38386106e-01 3.48888785e-01 -7.20849037e-01
-1.65693760e+00 7.24665165e-01 -9.57534760e-02 -1.17894685e+00
7.39661336e-01 -1.08006406e+00 -8.10261190e-01 1.14677799e+00
-1.54839432e+00 -1.34152532e+00 -3.63217294e-01 5.46655536e-01
3.22437465e-01 -1.89416945e-01 9.66786802e-01 4.66999412e-01
-3.58316302e-01 5.63876808e-01 -9.54786222e-03 5.14569104e-01
1.81437746e-01 -1.09192538e+00 9.06056345e-01 1.36659190e-01
2.77791005e-02 5.70139229e-01 -1.30243540e-01 -7.72575736e-01
-1.82882810e+00 -1.29048467e+00 7.04767644e-01 -2.89481938e-01
3.11600268e-01 -4.92888778e-01 -1.16140580e+00 5.40672421e-01
-2.92697579e-01 3.33111376e-01 4.88164514e-01 -5.88577613e-02
-5.03720939e-01 1.31268531e-01 -1.18839395e+00 6.44637465e-01
1.28266835e+00 -4.57571417e-01 -4.91214335e-01 3.55461240e-01
7.27358222e-01 -8.60411882e-01 -1.07950950e+00 5.33557296e-01
4.71479088e-01 -6.17692709e-01 1.18925571e+00 -5.71944594e-01
4.86181974e-01 -3.37467670e-01 -2.54294932e-01 -1.03161335e+00
-5.95372617e-01 -2.01578990e-01 5.12055270e-02 7.81941950e-01
1.27298236e-01 -9.00416315e-01 7.50350833e-01 5.15595913e-01
-3.27141434e-01 -1.40373671e+00 -1.01060522e+00 -6.16816580e-01
3.13530356e-01 -3.44241560e-01 1.02920628e+00 9.96301830e-01
-5.48121274e-01 -1.03502162e-02 1.02616593e-01 3.06489885e-01
3.51037651e-01 4.07321811e-01 6.51728451e-01 -1.45546806e+00
2.87293583e-01 -7.03962684e-01 -8.36130202e-01 -1.21279836e+00
8.93193204e-03 -1.46946669e+00 -4.70224589e-01 -1.74670041e+00
-5.76300025e-02 -8.38694394e-01 -2.89581954e-01 5.47229588e-01
1.57246366e-01 2.82740414e-01 5.76467738e-02 1.68052867e-01
-3.68768632e-01 7.38421440e-01 1.64947510e+00 -2.50878423e-01
-1.76112995e-01 -8.98826271e-02 -4.67084527e-01 7.62174547e-01
7.63731360e-01 -2.12792918e-01 -1.04352325e-01 -9.51838911e-01
1.92343682e-01 -1.29128292e-01 2.91481763e-01 -8.74985933e-01
2.08829656e-01 8.96475390e-02 7.31163263e-01 -1.04900861e+00
4.48332280e-01 -5.95158041e-01 1.21605471e-01 2.49501884e-01
2.62288570e-01 3.99964422e-01 6.31879210e-01 5.52183747e-01
-1.28379375e-01 -1.22168511e-02 5.30095518e-01 -3.58006448e-01
-4.94053602e-01 1.05943096e+00 -1.50331348e-01 -2.73098409e-01
6.56376302e-01 -4.54971194e-01 -3.46084446e-01 -1.32341273e-02
-4.85365152e-01 2.75813431e-01 4.16136324e-01 5.36110044e-01
1.09073389e+00 -1.93465018e+00 -6.32966399e-01 5.73110819e-01
-4.90862653e-02 4.93089467e-01 3.44438761e-01 3.55836898e-01
-1.08865917e+00 2.93976247e-01 -4.91350502e-01 -5.11624336e-01
-9.84418929e-01 2.84028500e-01 3.53282392e-01 -1.55651465e-01
-9.08681571e-01 9.90977228e-01 2.28800327e-01 -9.82314825e-01
3.02726418e-01 -5.48397124e-01 -2.79379398e-01 -1.68086320e-01
3.06204915e-01 2.08890662e-01 2.32275248e-01 -6.80589557e-01
-2.30125844e-01 9.57803786e-01 1.51767051e-02 6.88355029e-01
1.60576427e+00 3.45987827e-01 -5.23009121e-01 3.30555201e-01
1.25955999e+00 -1.62575707e-01 -8.05277526e-01 -4.83993828e-01
-3.37129474e-01 -6.00122333e-01 -1.57419555e-02 -3.42828721e-01
-1.34164715e+00 8.59956264e-01 3.24240297e-01 1.88041240e-01
7.38428354e-01 -4.38869037e-02 9.14938390e-01 4.87640053e-01
4.25081760e-01 -6.54503405e-01 -2.06368461e-01 1.04826522e+00
1.52306139e+00 -1.02464151e+00 1.27809361e-01 -7.74683774e-01
-1.29074886e-01 1.54055953e+00 5.12629509e-01 -7.20329642e-01
1.29071295e+00 1.35249957e-01 -1.66574121e-01 -7.81381905e-01
-3.02506089e-01 -1.16583563e-01 5.55329263e-01 5.38202703e-01
3.33664000e-01 3.50847453e-01 -2.65025109e-01 6.43852532e-01
-1.07278176e-01 -3.32983881e-01 6.62498996e-02 8.96177351e-01
-2.18481004e-01 -9.85711396e-01 -2.09056199e-01 8.34972799e-01
-8.30357298e-02 -1.43720552e-01 -4.10856396e-01 7.61378169e-01
1.48478404e-01 9.16713849e-02 3.28122199e-01 -4.16987360e-01
6.07762754e-01 -7.75478035e-02 4.12719488e-01 -8.52027297e-01
-5.94346523e-01 -2.04360545e-01 -2.50044167e-01 -5.74133039e-01
-8.83454382e-02 -3.62075984e-01 -1.37307274e+00 -4.32182372e-01
-4.19735573e-02 -2.46999860e-01 6.12350404e-01 4.95326936e-01
7.09834099e-01 4.00270760e-01 4.85041410e-01 -1.48102295e+00
-2.74311692e-01 -6.84234619e-01 -6.49745047e-01 4.40205544e-01
-2.42807996e-02 -9.90051568e-01 -9.58211571e-02 -6.88651800e-01] | [8.061229705810547, -3.6976425647735596] |
4d14d25a-37be-4af4-b8d4-f962a119e80f | hft-cnn-learning-hierarchical-category | null | null | https://aclanthology.org/D18-1093 | https://aclanthology.org/D18-1093.pdf | HFT-CNN: Learning Hierarchical Category Structure for Multi-label Short Text Categorization | We focus on the multi-label categorization task for short texts and explore the use of a hierarchical structure (HS) of categories. In contrast to the existing work using non-hierarchical flat model, the method leverages the hierarchical relations between the pre-defined categories to tackle the data sparsity problem. The lower the HS level, the less the categorization performance. Because the number of training data per category in a lower level is much smaller than that in an upper level. We propose an approach which can effectively utilize the data in the upper levels to contribute the categorization in the lower levels by applying the Convolutional Neural Network (CNN) with a fine-tuning technique. The results using two benchmark datasets show that proposed method, Hierarchical Fine-Tuning based CNN (HFT-CNN) is competitive with the state-of-the-art CNN based methods. | ['Jiyi Li', 'Kazuya Shimura', 'Fumiyo Fukumoto'] | 2018-10-01 | null | null | null | emnlp-2018-10 | ['extreme-multi-label-classification'] | ['methodology'] | [ 1.98861491e-02 1.92962497e-01 -2.72596627e-01 -5.15987813e-01
-2.44815499e-01 -3.89362633e-01 3.02427024e-01 2.56405026e-01
-3.92082691e-01 4.58286315e-01 5.96749723e-01 -1.89856365e-01
-2.10550323e-01 -8.47537577e-01 -1.26074925e-01 -6.70840859e-01
3.55784059e-01 2.49744222e-01 2.71414965e-01 -2.48446286e-01
5.11179507e-01 -1.14244319e-01 -1.84012711e+00 8.96659195e-01
5.85368931e-01 1.36230803e+00 9.81729999e-02 2.01353684e-01
-4.14661705e-01 9.22069907e-01 -3.04452300e-01 -1.72903612e-01
2.18577251e-01 2.26839017e-02 -1.06452560e+00 1.97533101e-01
5.58136940e-01 -1.01375923e-01 9.88274999e-03 1.00099874e+00
1.75946712e-01 4.13191766e-01 7.31689513e-01 -9.72896457e-01
-7.59339511e-01 9.23145056e-01 -7.69182622e-01 1.45851195e-01
-1.69633806e-01 -6.58515930e-01 1.30961764e+00 -8.76145184e-01
3.72634113e-01 1.37882733e+00 1.18298662e+00 5.37400782e-01
-8.66177559e-01 -8.21055889e-01 5.14219701e-01 9.60135013e-02
-1.45909917e+00 -8.03107694e-02 5.11496186e-01 -7.25546718e-01
8.55187654e-01 -1.15586430e-01 6.07462414e-02 6.45325422e-01
5.97921461e-02 3.46685141e-01 1.32318103e+00 -6.66215003e-01
5.36409281e-02 1.63060635e-01 1.11108065e+00 5.98701835e-01
1.26378238e-01 -2.64777154e-01 -3.37648988e-01 -6.08948283e-02
3.17124605e-01 2.39545941e-01 2.57689148e-01 9.82918937e-05
-1.00129175e+00 1.20410347e+00 7.45066166e-01 7.81341195e-01
-2.96198785e-01 3.82895768e-03 8.24515522e-01 3.25159132e-01
6.74413860e-01 4.01318252e-01 -9.14114535e-01 4.19044614e-01
-1.13980258e+00 -9.15770531e-02 7.46776223e-01 1.01216018e+00
9.49123085e-01 -9.47089046e-02 -5.30162334e-01 1.16351306e+00
2.47026771e-01 -7.15159297e-01 7.97988296e-01 -7.52444446e-01
5.74448943e-01 1.13919580e+00 -3.09163451e-01 -9.35537219e-01
-6.05140448e-01 -7.39912450e-01 -1.31904757e+00 -1.68404415e-01
2.17156589e-01 -1.36794358e-01 -1.05545080e+00 1.67391098e+00
3.94802749e-01 -2.83410260e-03 4.64463420e-02 4.88347381e-01
1.28312063e+00 5.40053964e-01 4.47758645e-01 -4.28773314e-02
1.58576202e+00 -1.34114635e+00 -6.74288809e-01 3.62265930e-02
7.50412524e-01 -7.07594395e-01 1.08087885e+00 3.80861163e-01
-4.69207644e-01 -1.03992891e+00 -9.81478512e-01 -4.25360799e-01
-6.93053246e-01 2.92467028e-01 5.91480911e-01 5.97426713e-01
-1.14487636e+00 5.09837031e-01 -2.75950611e-01 -4.44467276e-01
4.80706006e-01 4.41969901e-01 -3.48381072e-01 5.90114854e-03
-1.44104052e+00 6.70519292e-01 8.59627187e-01 -1.89802378e-01
-6.68611586e-01 -7.26426840e-01 -6.49741709e-01 5.04263282e-01
1.07745111e-01 -5.31559527e-01 1.11794221e+00 -1.14181113e+00
-1.22879255e+00 7.97882318e-01 5.98425828e-02 -2.95932978e-01
1.27377197e-01 5.48699126e-02 -1.64534189e-02 -3.01651098e-02
2.65220374e-01 9.34511602e-01 5.69171548e-01 -1.39965332e+00
-1.17589474e+00 -2.17051774e-01 4.19844329e-01 3.65696520e-01
-1.10097969e+00 8.39387625e-02 -1.86146200e-01 -7.01906741e-01
3.86682212e-01 -7.61870086e-01 -2.19048262e-01 -4.60702002e-01
-4.80951905e-01 -9.48191881e-01 1.01852226e+00 -3.56096953e-01
1.63581645e+00 -1.93497038e+00 9.94548053e-02 5.55925742e-02
5.99718392e-01 -7.22172484e-02 2.04038113e-01 3.51505518e-01
4.83040847e-02 3.32381099e-01 2.41594151e-01 -5.13737679e-01
9.19225067e-02 2.65611261e-01 -1.32078789e-02 2.42470011e-01
-4.02206331e-01 4.96201187e-01 -6.73635662e-01 -6.71692729e-01
-1.91732422e-02 2.84096986e-01 -7.13378668e-01 7.09584579e-02
1.97016761e-01 2.46657327e-01 -3.92737508e-01 4.83832747e-01
6.49814367e-01 -3.97864580e-01 2.16371059e-01 -5.29046476e-01
-4.17002887e-01 4.76756841e-01 -1.12265050e+00 1.38865685e+00
-5.17989337e-01 2.54686654e-01 -3.26549932e-02 -1.31545913e+00
7.32064068e-01 5.21702707e-01 3.18092793e-01 -2.69800991e-01
3.70307267e-01 -5.03486320e-02 4.28193994e-02 -2.44059354e-01
5.87001979e-01 -2.99039215e-01 -3.54268581e-01 4.04653490e-01
4.06001300e-01 4.68094498e-01 2.67252237e-01 2.01128647e-01
8.25924098e-01 -2.37829596e-01 4.96037751e-01 -8.98667812e-01
6.01034582e-01 -8.40481296e-02 5.96765935e-01 9.31319237e-01
-6.85999542e-02 3.85990173e-01 4.91436243e-01 -6.97713494e-01
-1.21789539e+00 -2.66744137e-01 -3.64620656e-01 1.96766424e+00
-6.44122139e-02 -3.39379638e-01 -9.02608931e-01 -9.25755739e-01
1.14660889e-01 2.61048466e-01 -1.26115906e+00 1.40788943e-01
-2.94855475e-01 -7.50954270e-01 6.31579161e-01 6.66324556e-01
7.92261660e-01 -1.20107126e+00 -2.84219831e-01 2.37545207e-01
-1.94636449e-01 -1.06204295e+00 -4.75345820e-01 6.69024944e-01
-8.01293969e-01 -8.15602422e-01 -2.83209085e-01 -1.19732523e+00
6.47375822e-01 4.53149170e-01 1.15545762e+00 3.96550179e-01
2.30246782e-01 -1.77471057e-01 -9.14776802e-01 -2.22116008e-01
-2.63096571e-01 7.59976208e-01 -4.84411120e-02 2.30079606e-01
5.80538750e-01 -3.83910507e-01 -4.98971939e-01 3.01789850e-01
-8.11836302e-01 3.04078311e-01 3.97204965e-01 1.08749020e+00
4.89755273e-01 5.39760947e-01 8.82253706e-01 -1.35121572e+00
5.45893788e-01 -8.33661079e-01 -1.63895249e-01 4.34554778e-02
-7.86970913e-01 -1.24745443e-01 9.40234363e-01 -5.33330262e-01
-1.02677584e+00 5.43009229e-02 -3.36113498e-02 -1.74268842e-01
-3.26760799e-01 8.02147925e-01 1.93886474e-01 3.38409580e-02
5.21603703e-01 -1.99251726e-01 -4.81036365e-01 -7.16138780e-01
3.24038118e-01 1.20017016e+00 7.46882111e-02 -5.28506875e-01
4.30338442e-01 3.16336930e-01 -1.82555020e-01 -5.04485190e-01
-1.65835786e+00 -1.06152570e+00 -1.31804097e+00 -1.63740456e-01
9.52753603e-01 -9.49109435e-01 -6.32077157e-01 1.61334485e-01
-8.78406703e-01 -2.50713915e-01 1.07144408e-01 1.38060283e-02
-1.08573601e-01 2.53418297e-01 -7.56626606e-01 -4.22933847e-01
-6.24800563e-01 -9.21578407e-01 9.98282790e-01 1.58363163e-01
-1.76239640e-01 -1.05139017e+00 -3.22112113e-01 5.27466059e-01
5.29963791e-01 6.77091107e-02 1.21917510e+00 -1.01255083e+00
4.80817370e-02 -1.79870844e-01 -5.12981057e-01 4.65194076e-01
2.44063333e-01 -5.18860877e-01 -1.10057628e+00 -4.37626362e-01
3.42414454e-02 -6.83347404e-01 1.13680816e+00 3.18014413e-01
1.42875576e+00 -5.81368148e-01 -3.28863025e-01 4.30892587e-01
1.53173208e+00 -1.04926504e-01 2.92637616e-01 6.16181850e-01
1.14538860e+00 7.48928249e-01 4.53275532e-01 3.35380763e-01
7.07030416e-01 5.86884141e-01 1.37930453e-01 -3.62359196e-01
-2.57611722e-01 5.42156249e-02 -9.81629342e-02 1.36091530e+00
1.05079226e-01 -1.61324859e-01 -8.34295452e-01 5.28654575e-01
-1.94679248e+00 -7.41657078e-01 -3.53605539e-01 1.67342842e+00
1.06342053e+00 1.09283447e-01 7.77894631e-03 3.64689231e-01
1.00929165e+00 -1.79966301e-01 -2.26684257e-01 -5.84216833e-01
1.45991907e-01 1.06076203e-01 4.55579877e-01 3.94585133e-01
-1.54224813e+00 1.08096457e+00 6.40063334e+00 1.02196658e+00
-7.89018214e-01 5.88829875e-01 7.18516469e-01 4.23779845e-01
2.10984141e-01 -1.17649488e-01 -1.32866454e+00 1.85153410e-01
9.42554832e-01 7.40079954e-02 8.55849609e-02 8.83263826e-01
-2.56973561e-02 1.75031990e-01 -9.71195936e-01 5.00154793e-01
6.67199045e-02 -1.17951787e+00 1.90953955e-01 7.61712641e-02
1.08808088e+00 4.20499779e-02 -4.84821200e-02 6.59369409e-01
4.78916645e-01 -8.83788526e-01 7.90327549e-01 8.10958818e-02
8.30057442e-01 -6.97707176e-01 1.10208702e+00 5.61501265e-01
-1.57712519e+00 -5.23887575e-01 -6.97599828e-01 -1.82026133e-01
-5.10068715e-01 4.99700725e-01 -6.10876322e-01 5.35107970e-01
9.81644988e-01 7.18344092e-01 -7.13154078e-01 6.73402905e-01
4.77624089e-02 9.13620591e-01 -3.74842063e-02 1.18846670e-01
7.60928154e-01 1.71593174e-01 -9.50559974e-02 1.79635429e+00
5.28083220e-02 1.92057580e-01 5.83115637e-01 3.01245928e-01
-4.03214961e-01 4.19139475e-01 -2.05400452e-01 3.99687707e-01
4.94833708e-01 1.63229835e+00 -9.87745583e-01 -6.84006095e-01
-8.15453172e-01 3.65295440e-01 7.49086976e-01 -3.22522074e-02
-3.35090727e-01 -4.99744028e-01 6.37644827e-02 -3.44013199e-02
4.16180342e-01 -3.45924590e-03 -7.84075737e-01 -1.03274429e+00
-4.94044781e-01 -8.14350486e-01 9.69027162e-01 -2.85052866e-01
-1.68757069e+00 8.25307906e-01 7.90248588e-02 -1.27167082e+00
2.55012095e-01 -5.68285108e-01 -1.49806425e-01 7.62037694e-01
-1.55153954e+00 -1.51043808e+00 -5.07529795e-01 5.83947897e-01
1.03761744e+00 -2.90776074e-01 1.03760779e+00 5.40240109e-01
-5.35987794e-01 8.33663940e-01 3.23145360e-01 2.72814721e-01
5.55318534e-01 -1.34362006e+00 -2.13698357e-01 5.60036182e-01
-2.36563385e-01 7.07784295e-01 2.47117206e-01 -5.00382900e-01
-3.99822712e-01 -1.33690965e+00 1.08686173e+00 -2.77087629e-01
6.00262046e-01 -5.22085249e-01 -8.83826256e-01 6.28585935e-01
5.37411630e-01 -1.91970870e-01 1.05097520e+00 4.36860740e-01
-8.24869096e-01 -1.98401824e-01 -1.20444524e+00 -7.77877576e-04
6.79255009e-01 -2.99277216e-01 -8.97345841e-01 2.25227088e-01
8.08792889e-01 -1.40897185e-01 -1.03137410e+00 4.26038235e-01
7.04963326e-01 -6.54128253e-01 7.56102085e-01 -6.01002991e-01
6.78137600e-01 -1.30902693e-01 -3.87136042e-01 -1.13077974e+00
-9.16154087e-01 -2.09580548e-02 5.20112086e-03 1.41991568e+00
2.81158417e-01 -3.31862509e-01 6.30457342e-01 5.24822697e-02
-4.21807319e-01 -7.63105333e-01 -7.92612016e-01 -4.75115597e-01
5.90700805e-01 -3.92032154e-02 4.66016710e-01 1.32389212e+00
-5.14110588e-02 6.94023788e-01 -3.66188943e-01 -6.71928823e-02
6.26974523e-01 2.34671548e-01 1.72090739e-01 -1.72116244e+00
1.06487140e-01 -4.87984419e-01 -1.62354916e-01 -8.40872765e-01
3.73734623e-01 -1.21754718e+00 2.37467960e-01 -1.60772204e+00
7.06570029e-01 -7.56051242e-01 -8.95934880e-01 1.02509844e+00
-3.25695068e-01 6.84628308e-01 1.62159428e-01 5.84632039e-01
-8.22117150e-01 1.14946067e-01 1.03699374e+00 -5.57307042e-02
-1.79760866e-02 -1.77071422e-01 -1.23321378e+00 8.77149999e-01
7.34053433e-01 -8.39230120e-01 -2.01417014e-01 -3.04461747e-01
-2.65565366e-02 -4.55275476e-01 -3.50500107e-01 -1.06726110e+00
5.32194614e-01 2.42105983e-02 3.54607284e-01 -7.68624663e-01
-1.09340735e-01 -7.02892661e-01 -2.23219201e-01 3.10115844e-01
-8.65410686e-01 -1.28546372e-01 8.08754005e-03 3.61251712e-01
-3.51892740e-01 -4.62749273e-01 1.14964783e+00 -3.22833061e-01
-6.78887546e-01 3.23832244e-01 -3.56335074e-01 -2.14793384e-01
6.25571907e-01 -2.39648610e-01 -2.41029903e-01 8.89669284e-02
-8.78757775e-01 2.48788446e-01 -2.18752809e-02 5.22463500e-01
1.58333883e-01 -1.38825560e+00 -5.99163830e-01 9.16029979e-03
2.51266599e-01 -1.03452712e-01 1.67815089e-01 5.72049439e-01
-2.10007921e-01 7.06740975e-01 -3.31491709e-01 -3.88665169e-01
-1.19522953e+00 7.63581634e-01 2.41527408e-01 -7.56485224e-01
-4.90481198e-01 8.70298088e-01 7.23209202e-01 -7.30458677e-01
5.40520728e-01 -1.16585135e-01 -1.19700229e+00 5.10076046e-01
2.45612651e-01 3.09001863e-01 3.14792633e-01 -1.01733267e+00
-1.79402664e-01 6.69143140e-01 -4.34081405e-01 3.60264033e-01
1.37507248e+00 -3.98465693e-01 -4.88195509e-01 6.12676740e-01
1.34146404e+00 -5.01724482e-01 -9.27419126e-01 -4.72628862e-01
1.66487873e-01 -9.36443806e-02 4.66328710e-01 -7.99938619e-01
-1.02229142e+00 8.38616610e-01 5.25265157e-01 7.20314801e-01
1.06926668e+00 -1.22255325e-01 5.55720389e-01 3.60951662e-01
1.82091534e-01 -1.38908696e+00 2.28816450e-01 1.01339173e+00
7.17764139e-01 -9.84033763e-01 5.02776392e-02 -5.59676528e-01
-6.26428962e-01 1.27503312e+00 9.18005347e-01 -1.78786054e-01
1.02650964e+00 1.72622740e-01 1.38552755e-01 -2.32221141e-01
-7.80698538e-01 -3.57319444e-01 4.77737248e-01 2.68095672e-01
7.90182710e-01 8.53551924e-02 -4.76602286e-01 9.54720855e-01
-1.56347036e-01 -4.33888808e-02 3.61317664e-01 6.56288266e-01
-7.45207787e-01 -1.04910278e+00 -2.89020151e-01 8.92561197e-01
-8.63292873e-01 -3.61504436e-01 -1.65091231e-01 5.54252028e-01
8.42208922e-01 1.14357674e+00 5.62680997e-02 -4.69506323e-01
3.09172064e-01 2.06529468e-01 2.18832474e-02 -1.23017228e+00
-1.11821580e+00 1.79754049e-01 4.68188785e-02 -1.48335427e-01
-7.63396084e-01 -5.46399176e-01 -1.00238812e+00 -2.80065358e-01
-6.99082553e-01 1.97291464e-01 3.31474334e-01 1.17022920e+00
1.88722149e-01 7.28736877e-01 8.28544915e-01 -9.05681252e-01
-5.37168562e-01 -1.60976839e+00 -6.89457953e-01 5.95357955e-01
3.04434776e-01 -1.01803219e+00 -5.89073777e-01 1.45756498e-01] | [9.678513526916504, 4.389142036437988] |
7c07e654-a1cb-4051-a413-bd47707798f4 | cross-lingual-morphological-tagging-for-low | 1606.04279 | null | http://arxiv.org/abs/1606.04279v1 | http://arxiv.org/pdf/1606.04279v1.pdf | Cross-Lingual Morphological Tagging for Low-Resource Languages | Morphologically rich languages often lack the annotated linguistic resources
required to develop accurate natural language processing tools. We propose
models suitable for training morphological taggers with rich tagsets for
low-resource languages without using direct supervision. Our approach extends
existing approaches of projecting part-of-speech tags across languages, using
bitext to infer constraints on the possible tags for a given word type or
token. We propose a tagging model using Wsabie, a discriminative
embedding-based model with rank-based learning. In our evaluation on 11
languages, on average this model performs on par with a baseline
weakly-supervised HMM, while being more scalable. Multilingual experiments show
that the method performs best when projecting between related language pairs.
Despite the inherently lossy projection, we show that the morphological tags
predicted by our models improve the downstream performance of a parser by +0.6
LAS on average. | ['Jan Buys', 'Jan A. Botha'] | 2016-06-14 | cross-lingual-morphological-tagging-for-low-1 | https://aclanthology.org/P16-1184 | https://aclanthology.org/P16-1184.pdf | acl-2016-8 | ['morphological-tagging'] | ['natural-language-processing'] | [-9.97869000e-02 2.14730442e-01 -4.63446617e-01 -6.63757384e-01
-1.41252410e+00 -1.01310098e+00 6.06258571e-01 3.16318214e-01
-8.78053069e-01 6.74687266e-01 6.11033380e-01 -7.10178077e-01
3.57920736e-01 -5.84207833e-01 -6.38063073e-01 -3.95491421e-01
-1.18242688e-01 7.74144530e-01 3.41006458e-01 -5.26713021e-02
-7.30208680e-02 2.39789844e-01 -9.11625266e-01 4.76222783e-01
7.89876997e-01 1.12706922e-01 6.70802414e-01 3.03305328e-01
-5.09338617e-01 3.35924596e-01 -1.95605069e-01 -8.43273044e-01
3.50490332e-01 -1.35684133e-01 -9.12701249e-01 -2.88605720e-01
6.76097095e-01 7.24259093e-02 -7.02000186e-02 1.18905902e+00
2.62050837e-01 -6.65219650e-02 4.34544593e-01 -3.61553311e-01
-4.11900908e-01 1.40181661e+00 -8.73748213e-02 2.87127346e-01
3.54446530e-01 -1.21324748e-01 1.70669544e+00 -1.07414865e+00
9.83843625e-01 1.29533982e+00 7.58024037e-01 7.26236761e-01
-1.56904483e+00 -3.12961519e-01 2.24177554e-01 -2.37566739e-01
-1.29587615e+00 -7.59294450e-01 4.05940890e-01 -4.03376192e-01
1.40705204e+00 8.45339075e-02 2.23428950e-01 7.76361108e-01
1.28376693e-01 7.99040377e-01 1.23349273e+00 -8.08050334e-01
-6.18352517e-02 2.28319049e-01 1.31219402e-01 9.15950060e-01
3.27813327e-01 -3.78520824e-02 -6.49691343e-01 -2.11108819e-01
2.18386725e-01 -4.60255027e-01 3.67479771e-01 -1.97620332e-01
-1.21949351e+00 7.41568863e-01 -2.12464377e-01 5.76021910e-01
-2.04012562e-02 -1.99500732e-02 4.95821357e-01 1.53366074e-01
7.46138990e-01 4.83719140e-01 -9.76296842e-01 -1.06252201e-01
-1.10627985e+00 -2.80837446e-01 9.15271878e-01 1.08192587e+00
8.91622126e-01 -7.58541971e-02 1.45777255e-01 1.14940953e+00
4.15225476e-01 5.43317854e-01 3.95659804e-01 -7.05643773e-01
6.75369501e-01 1.95109516e-01 -1.44782767e-01 -9.43874493e-02
-3.62800002e-01 -6.00605831e-02 -3.79902944e-02 -2.45836139e-01
6.18964851e-01 -1.55617595e-01 -7.16218948e-01 1.76421547e+00
1.47946447e-01 -1.66762322e-01 2.65547484e-01 3.11571777e-01
2.84378171e-01 7.81011641e-01 6.37293816e-01 -2.18064189e-01
1.54925585e+00 -7.78855324e-01 -5.02856910e-01 -6.74747050e-01
1.19694996e+00 -1.16568458e+00 1.07714319e+00 1.87614769e-01
-1.17514217e+00 -3.60598654e-01 -6.89109862e-01 -3.22455347e-01
-3.61609668e-01 3.43481869e-01 1.01727247e+00 7.66398907e-01
-1.16946387e+00 7.01755345e-01 -1.17882204e+00 -4.77851182e-01
-6.95995986e-02 4.02010888e-01 -6.56883359e-01 -1.68283165e-01
-9.49002683e-01 1.05638099e+00 5.70647955e-01 -2.20980972e-01
-4.99592632e-01 -4.93700743e-01 -1.07685292e+00 -1.38258100e-01
-7.12440386e-02 -1.37963563e-01 1.38033164e+00 -5.59577346e-01
-1.40778220e+00 1.31516075e+00 -4.09389049e-01 -3.70555580e-01
2.86935698e-02 -4.30036962e-01 -5.06742597e-01 -1.88552558e-01
4.37322855e-01 6.19908929e-01 -2.95786802e-02 -9.54646230e-01
-9.63302672e-01 -1.99524015e-01 -2.71138817e-01 6.76732436e-02
-3.30843121e-01 7.92815447e-01 -3.74490619e-01 -5.51085055e-01
8.64097849e-02 -8.76125753e-01 -4.15705144e-01 -5.38227260e-01
-3.89968753e-01 -5.47993362e-01 9.39523205e-02 -8.99044037e-01
1.18232965e+00 -1.98210132e+00 -2.14956716e-01 7.66296461e-02
-4.40521598e-01 3.20865512e-01 -3.27160805e-01 6.84303045e-01
9.59575996e-02 2.88437843e-01 -2.14342058e-01 -6.30795300e-01
1.74287856e-01 8.67403030e-01 -3.52521628e-01 4.38279450e-01
4.90406573e-01 6.01222336e-01 -1.10661352e+00 -6.51568174e-01
1.66543387e-02 1.98177621e-01 -5.40837765e-01 2.08096310e-01
-1.30681232e-01 1.20786957e-01 -1.48671046e-01 6.69200778e-01
4.83359545e-01 4.59870785e-01 1.02863431e+00 2.96905488e-01
-6.05225146e-01 1.38213015e+00 -1.00785291e+00 1.96919882e+00
-9.17341113e-01 2.73073673e-01 8.96316022e-02 -8.50970805e-01
9.58378911e-01 3.30580890e-01 7.73963034e-02 -1.60500184e-01
-3.13866675e-01 7.51866341e-01 3.66214335e-01 -2.68194973e-01
4.76649642e-01 -4.71652776e-01 -5.72933316e-01 3.22896421e-01
5.65397084e-01 5.21745645e-02 3.72008443e-01 5.05480133e-02
1.20264387e+00 4.21699822e-01 5.01975000e-01 -6.36949480e-01
3.32641572e-01 2.37763003e-01 1.09041154e+00 4.26321119e-01
3.61636095e-02 1.11245289e-01 2.43053898e-01 -3.94158661e-01
-1.11108482e+00 -1.28846371e+00 -4.76038516e-01 1.82538116e+00
-4.15952265e-01 -6.24106050e-01 -5.66952765e-01 -9.41661835e-01
-1.61813572e-01 9.31044996e-01 -3.49733494e-02 4.37943995e-01
-1.05344582e+00 -8.85934293e-01 9.24536228e-01 6.18241072e-01
-3.11180323e-01 -1.14003038e+00 1.42294198e-01 6.07610583e-01
1.45269623e-02 -1.25630593e+00 -4.39443260e-01 7.86635816e-01
-9.87107992e-01 -6.20409906e-01 -1.02646314e-01 -1.42688000e+00
6.33161366e-01 -3.06186557e-01 1.28401804e+00 -2.37178087e-01
2.04608113e-01 -2.39628419e-01 -3.61710042e-01 -1.48360685e-01
-7.52283931e-01 3.86435121e-01 3.97510111e-01 -5.14199674e-01
8.23024571e-01 -4.30651933e-01 3.95882726e-02 -7.65275434e-02
-4.66889918e-01 -2.73270249e-01 7.92837977e-01 1.04152036e+00
5.75483203e-01 -5.45660555e-01 4.71143564e-03 -1.32135534e+00
5.91956563e-02 -4.33621049e-01 -6.39487922e-01 2.04184651e-01
-4.21171784e-01 5.48763275e-01 8.65888834e-01 -2.50448465e-01
-1.23215067e+00 5.68248391e-01 -5.43620050e-01 4.78304178e-01
-5.10457039e-01 4.03153241e-01 -6.21645749e-01 3.39075804e-01
4.02349621e-01 1.44772157e-01 -4.86022413e-01 -1.09889042e+00
6.18816018e-01 7.22969294e-01 6.65520608e-01 -9.56967711e-01
8.84169996e-01 2.39294648e-01 -2.60760546e-01 -7.01176524e-01
-9.14773464e-01 -9.51590061e-01 -1.18945301e+00 5.10652900e-01
8.69342029e-01 -1.09266806e+00 -2.65791267e-02 -1.20508827e-01
-1.28915000e+00 -3.84918123e-01 -7.51640126e-02 8.52474034e-01
-3.53095919e-01 4.98744160e-01 -1.11203408e+00 -6.56479061e-01
-1.40544578e-01 -8.04603517e-01 1.13733435e+00 -3.93027693e-01
-2.70091921e-01 -1.33011305e+00 5.28783977e-01 1.94130570e-01
-1.42387822e-01 -4.71209288e-01 1.05227542e+00 -1.29768014e+00
-3.37330312e-01 5.30510815e-03 1.97186217e-01 2.24324152e-01
-5.54934097e-03 -1.01061270e-01 -8.48357975e-01 -1.54145524e-01
-6.20678604e-01 -2.62064278e-01 9.64159608e-01 -1.13704419e-02
3.98474514e-01 -4.60991204e-01 -3.97979617e-01 6.43631518e-01
1.47198272e+00 -4.03693430e-02 2.16897219e-01 3.71356368e-01
6.26003027e-01 7.33235240e-01 6.26265764e-01 6.53486028e-02
3.71569037e-01 5.69744706e-01 -1.86213464e-01 5.91797754e-02
-1.15106530e-01 -7.33958125e-01 1.01680040e+00 1.59606302e+00
1.40716150e-01 -5.91621622e-02 -1.15852344e+00 9.32766259e-01
-1.54748845e+00 -8.07301521e-01 -2.11854637e-01 2.11975622e+00
1.33203363e+00 1.67530522e-01 -4.59481701e-02 -3.23652416e-01
6.82226479e-01 2.40586475e-02 1.72267243e-01 -9.13119256e-01
-1.01059482e-01 7.48226404e-01 8.65626216e-01 9.99747574e-01
-1.13371074e+00 1.82319546e+00 6.81813812e+00 8.56427014e-01
-7.41137862e-01 4.41318274e-01 6.46816939e-02 3.09890181e-01
-6.49603009e-01 5.26846945e-01 -1.47751081e+00 2.84980476e-01
1.28383374e+00 1.89013809e-01 2.26263762e-01 8.29424798e-01
1.64084151e-01 9.51793641e-02 -1.15428853e+00 3.01175594e-01
-1.30485475e-01 -9.67123687e-01 -1.96058258e-01 2.34903350e-01
5.05879283e-01 6.69650376e-01 -2.31711119e-01 2.51142681e-01
1.25762784e+00 -6.61587238e-01 8.33641171e-01 9.50010866e-02
9.93329823e-01 -6.19620383e-01 7.73156404e-01 4.12488133e-01
-1.24971831e+00 3.11435103e-01 -8.16589177e-01 -2.18007207e-01
3.58459502e-01 5.74587584e-01 -1.19567418e+00 2.63107508e-01
2.73074090e-01 5.35019755e-01 -3.41964036e-01 7.89137721e-01
-8.98385525e-01 1.35574520e+00 -4.93683219e-01 -1.33397371e-01
4.58828270e-01 -1.82788670e-01 4.45018351e-01 2.07167864e+00
4.16536242e-01 -1.20910205e-01 6.11048460e-01 2.99216121e-01
-2.09443178e-02 5.70900679e-01 -6.85897231e-01 -1.92280263e-02
7.33764052e-01 1.29290628e+00 -6.65566504e-01 -4.08350468e-01
-7.19927013e-01 9.62034106e-01 8.02227676e-01 -1.76746905e-01
-1.27387062e-01 -2.87125468e-01 8.00833941e-01 -9.32660997e-02
4.93195176e-01 -5.97764254e-01 -1.83003843e-01 -1.25103176e+00
-4.64671254e-02 -5.38506746e-01 5.25337934e-01 -1.33846506e-01
-1.51717865e+00 3.74402016e-01 -1.70650259e-01 -8.54311526e-01
-4.92745370e-01 -1.14427984e+00 -4.52210337e-01 1.08043444e+00
-1.56913185e+00 -1.46242881e+00 9.16339099e-01 5.20761535e-02
4.42525804e-01 -2.78096497e-01 1.18520916e+00 1.90893203e-01
-4.62425977e-01 5.69386363e-01 3.48380119e-01 6.00339770e-01
8.62335205e-01 -1.74677110e+00 8.47690582e-01 1.22084188e+00
9.85192716e-01 1.03181922e+00 5.51398754e-01 -7.33030677e-01
-1.14113438e+00 -1.12888205e+00 2.03258443e+00 -6.67319655e-01
1.14854133e+00 -8.62279296e-01 -7.97849715e-01 9.94332612e-01
3.08162898e-01 4.49592248e-02 1.27128613e+00 9.20394003e-01
-7.17970848e-01 2.05421727e-02 -7.39429772e-01 2.97926754e-01
1.35738564e+00 -7.99642503e-01 -1.24323499e+00 4.63553280e-01
5.38724959e-01 9.14504752e-02 -9.47772682e-01 -1.37704790e-01
4.05720770e-01 -5.55802763e-01 5.55705965e-01 -7.61758924e-01
1.05328590e-01 -1.46854341e-01 -3.98016632e-01 -1.48191106e+00
-6.19861901e-01 -6.14352703e-01 6.23648345e-01 1.57625818e+00
8.55304837e-01 -3.42835635e-01 6.41869485e-01 2.08494008e-01
-5.84316552e-01 -1.64907023e-01 -1.09508979e+00 -1.15467536e+00
5.51590025e-01 -7.08005846e-01 3.14211041e-01 1.05733812e+00
5.70932329e-01 4.98976171e-01 -1.19511373e-01 5.08781374e-01
4.74954188e-01 -1.01833478e-01 2.89241731e-01 -1.23675466e+00
-6.07549727e-01 -1.40575260e-01 -5.63812196e-01 -1.04205775e+00
8.08331907e-01 -1.49065924e+00 4.74311441e-01 -1.41219044e+00
3.00926626e-01 -8.51084709e-01 -3.66270840e-01 1.01261723e+00
-9.28675160e-02 2.45108992e-01 1.09957948e-01 1.08415909e-01
-5.34126103e-01 -4.39562388e-02 3.29369187e-01 9.98302549e-02
4.98013757e-02 -2.27287680e-01 -5.21689832e-01 1.07225132e+00
7.01249003e-01 -9.18075621e-01 2.93313324e-01 -8.32475066e-01
3.73485982e-01 -2.20731556e-01 -1.78943694e-01 -6.79037809e-01
3.48685905e-02 -1.78739816e-01 3.29054929e-02 -2.92771161e-01
9.24261287e-02 -4.73308057e-01 -2.69756079e-01 5.13423562e-01
-2.92431027e-01 1.48492381e-01 2.22351491e-01 1.87062293e-01
-1.73216701e-01 -6.02118969e-01 5.13036847e-01 -2.28560835e-01
-9.64520633e-01 1.42278939e-01 -6.28248751e-01 3.14430803e-01
4.09565270e-01 1.65007234e-01 6.93149492e-02 2.10557297e-01
-7.81797945e-01 -2.07330480e-01 5.05712271e-01 1.41084895e-01
9.10972357e-02 -1.08664954e+00 -7.82256126e-01 1.59504145e-01
3.04453462e-01 -5.94794333e-01 -6.09195709e-01 3.07956487e-01
-3.41627240e-01 5.31090975e-01 3.86311463e-03 -1.10949904e-01
-1.29925096e+00 3.58541608e-01 -1.33802876e-01 -6.15422785e-01
-2.86517888e-01 1.00838065e+00 -1.33856505e-01 -1.04307091e+00
-2.20197707e-01 -3.46742421e-01 -8.02730117e-03 -3.98516022e-02
2.04615429e-01 -7.95314927e-03 2.17736512e-01 -9.31118846e-01
-6.25006378e-01 3.39874148e-01 -1.06396787e-01 -5.28834403e-01
1.50515866e+00 1.14408053e-01 -2.72354901e-01 5.83212256e-01
1.01387942e+00 9.98417258e-01 -8.92546475e-01 -3.90501231e-01
7.79709578e-01 -1.78214848e-01 -1.08688794e-01 -6.71650648e-01
-4.45134550e-01 1.05095375e+00 -1.24901719e-02 -3.00325871e-01
4.29620087e-01 3.66935939e-01 9.42285657e-01 5.56416035e-01
6.96541250e-01 -1.23508561e+00 -6.63499355e-01 9.20714796e-01
7.80197559e-04 -9.94806409e-01 -2.07350940e-01 -6.41157329e-01
-4.34424907e-01 9.49536741e-01 1.79983065e-01 -7.62698054e-02
4.48348165e-01 7.48806596e-01 3.77786607e-01 2.06203833e-01
-8.71412694e-01 -6.56238556e-01 2.24462256e-01 7.26617157e-01
1.07189405e+00 5.33061445e-01 -7.79116750e-01 4.30441946e-01
-5.63063383e-01 -6.75141931e-01 1.66492894e-01 6.97277069e-01
-7.44632542e-01 -2.06368017e+00 -1.02602102e-01 2.59945273e-01
-9.88952398e-01 -8.53976548e-01 -2.04519123e-01 6.51648223e-01
5.55629253e-01 8.62128377e-01 3.37373883e-01 -1.17901877e-01
8.89569074e-02 6.13091886e-01 4.69674289e-01 -1.39442921e+00
-6.11102760e-01 2.56704092e-01 8.15215528e-01 -3.43291819e-01
-2.74722606e-01 -1.29167151e+00 -1.26859415e+00 4.12365124e-02
-2.95645446e-01 4.43109512e-01 4.89394605e-01 9.70894754e-01
-1.07566956e-02 -2.52972007e-01 4.74307209e-01 -4.07962829e-01
-7.91659653e-01 -1.02535403e+00 -7.45778441e-01 3.48569334e-01
-2.92583704e-01 -2.81141102e-01 -1.50711104e-01 2.78648347e-01] | [10.446590423583984, 9.929909706115723] |
e4154fee-7011-4316-9e7b-5492a1a5052d | breast-cancer-detection-using | 2202.06109 | null | https://arxiv.org/abs/2202.06109v1 | https://arxiv.org/pdf/2202.06109v1.pdf | Breast Cancer Detection using Histopathological Images | Cancer is one of the most common and fatal diseases in the world. Breast cancer affects one in every eight women and one in every eight hundred men. Hence, our prime target should be early detection of cancer because the early detection of cancer can be helpful to cure cancer effectively. Therefore, we propose a saliency detection system with the help of advanced deep learning techniques, such that the machine will be taught to emulate actions of pathologists for localization of diagnostically pertinent regions. We study identification of five diagnostic categories of breast cancer by training a CNN (VGG16, ResNet architecture). We have used BreakHis dataset to train our model. We focus on both detection and classification of cancerous regions in histopathology images. The diagnostically relevant regions are salient. The detection system will be available as an open source web application which can be used by pathologists and medical institutions. | ['Harsh Maan', 'Jitendra Maan'] | 2022-02-12 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 3.06518674e-01 6.59501255e-01 -3.68081659e-01 -7.33695775e-02
-6.18215919e-01 -1.09244540e-01 4.13686872e-01 6.18288994e-01
-6.74707830e-01 6.21535599e-01 7.82639012e-02 -5.82191348e-01
1.56464428e-01 -6.90349102e-01 -2.91825354e-01 -8.51784706e-01
1.02112414e-02 3.41183752e-01 5.38912594e-01 -1.11073375e-01
2.02490091e-01 8.50692987e-01 -9.20505106e-01 2.84584671e-01
8.19243193e-01 9.20956671e-01 7.19683826e-01 7.22726405e-01
-4.29077707e-02 7.46304333e-01 -3.43189985e-01 -2.43958965e-01
-1.30591571e-01 -5.18404722e-01 -8.57542515e-01 -2.53327280e-01
-6.25713915e-02 -1.31301090e-01 -2.39374280e-01 1.43397427e+00
4.24730211e-01 -4.06263441e-01 5.86640894e-01 -7.95035481e-01
-2.29583725e-01 3.71926248e-01 -9.69737291e-01 7.32257545e-01
-1.25525191e-01 -9.32851210e-02 6.70256793e-01 -8.13321233e-01
8.29274952e-01 7.97523618e-01 3.34621787e-01 8.75130951e-01
-7.20685720e-01 -4.68009442e-01 -3.04779887e-01 3.84067714e-01
-1.20502079e+00 -4.47686434e-01 8.26431513e-01 -9.94992182e-02
3.70283872e-01 3.47290665e-01 6.23505294e-01 8.95264506e-01
7.42518902e-01 1.03282225e+00 7.61473060e-01 -4.91643608e-01
1.66551173e-01 2.37034202e-01 1.40547737e-01 1.04821932e+00
3.52110237e-01 6.30820990e-02 -1.45893367e-02 -3.25844362e-02
8.37518275e-01 5.92042267e-01 -1.71798363e-01 -2.26623967e-01
-1.07466781e+00 9.97767210e-01 1.06798279e+00 4.38532710e-01
-2.68681884e-01 2.21143961e-01 3.11292499e-01 -1.01129137e-01
2.92839140e-01 3.24887186e-01 -5.54967597e-02 3.58703017e-01
-7.73070693e-01 -4.66853306e-02 2.61320144e-01 2.08465591e-01
4.28649813e-01 -2.40094781e-01 -1.95610598e-01 6.30302072e-01
3.52090061e-01 3.13484669e-01 9.25581098e-01 -5.29159188e-01
-3.74692917e-01 9.86702919e-01 -1.81859732e-01 -1.03378248e+00
-6.82437718e-01 -8.27486515e-01 -1.21475184e+00 1.54672518e-01
1.32015705e-01 -8.44695792e-02 -1.09004140e+00 1.16228807e+00
4.45527047e-01 1.17946275e-01 -4.25244719e-02 1.01636207e+00
1.04876351e+00 2.64284104e-01 3.23555887e-01 4.71914858e-02
1.79252243e+00 -7.71973193e-01 -6.67760313e-01 -2.44060308e-01
9.05385256e-01 -5.38356602e-01 5.57162225e-01 -1.36491090e-01
-9.07125115e-01 -2.39671648e-01 -9.60192144e-01 -2.09396899e-01
-4.46694493e-01 4.10834104e-01 8.33636403e-01 -1.14476401e-02
-1.06570125e+00 3.68588567e-01 -1.18269169e+00 -7.86974967e-01
9.00657117e-01 1.96703345e-01 -3.36137742e-01 1.24106921e-01
-9.61964011e-01 1.04263914e+00 4.11967605e-01 -4.53373082e-02
-1.00501525e+00 -5.80199063e-01 -8.11230719e-01 -3.28864641e-02
1.42368406e-01 -6.75186694e-01 1.36638188e+00 -1.16488111e+00
-5.40112555e-01 1.38874006e+00 -3.15546215e-01 -6.47694051e-01
3.81358922e-01 1.72968715e-01 -2.41404355e-01 6.47373617e-01
2.51222789e-01 9.97519195e-01 4.89439636e-01 -7.82177448e-01
-1.05346441e+00 -3.95425290e-01 -1.03115074e-01 -1.45824524e-02
-2.57978022e-01 1.45848304e-01 -3.21866870e-01 -7.55777538e-01
7.86742270e-02 -7.90432215e-01 -5.50748110e-01 4.35066551e-01
-6.26884818e-01 -1.98775217e-01 1.09678352e+00 -7.77296543e-01
8.49654198e-01 -2.17315483e+00 -4.65451442e-02 -6.92877546e-02
4.97464001e-01 2.67573744e-01 1.20863505e-01 -1.39761403e-01
-1.72405317e-01 1.39595419e-01 7.67129809e-02 1.25343010e-01
-6.08530164e-01 -1.80102572e-01 6.26187772e-02 6.97213829e-01
5.13134122e-01 1.11751378e+00 -1.07188094e+00 -6.98661685e-01
1.00005604e-01 3.16714287e-01 -1.13604985e-01 1.15836821e-01
-3.97649333e-02 3.51977378e-01 -6.39927268e-01 8.40099931e-01
4.41888899e-01 -5.62385798e-01 -6.91873059e-02 -5.38521074e-02
2.43794024e-01 6.23620413e-02 -2.59282380e-01 1.36625707e+00
-1.00123487e-01 8.14383566e-01 1.16163224e-01 -8.86471689e-01
6.09429061e-01 3.79628360e-01 5.37657380e-01 -4.80745405e-01
2.33228073e-01 8.68148580e-02 2.65457392e-01 -5.34854114e-01
3.18405628e-02 -1.47588342e-01 9.93043408e-02 3.17736268e-01
-3.39717001e-01 2.41785869e-01 -4.12862375e-03 5.53431332e-01
1.29750097e+00 -4.97371644e-01 5.35247147e-01 -4.41198438e-01
4.02274489e-01 3.97465616e-01 6.66419446e-01 4.32328433e-01
-5.19444227e-01 3.79070014e-01 6.52006686e-01 -7.05524147e-01
-8.22634220e-01 -1.02207613e+00 -1.32880375e-01 7.49106526e-01
1.19308092e-01 3.02979410e-01 -5.45199871e-01 -8.15268755e-01
-8.91399086e-02 5.88845134e-01 -1.03346598e+00 -5.80762029e-01
-4.40515280e-01 -6.80755436e-01 2.70390004e-01 5.27605653e-01
7.73514390e-01 -1.25784791e+00 -9.11414266e-01 -3.57108004e-03
5.29138856e-02 -7.77212977e-01 -2.07595766e-01 2.42370501e-01
-7.12081611e-01 -1.43262994e+00 -1.05011702e+00 -1.28417408e+00
1.11928260e+00 4.83402789e-01 7.60166645e-01 5.43859720e-01
-1.06897616e+00 -5.77897429e-02 -2.51581132e-01 -9.66505349e-01
-5.85708380e-01 1.20274909e-01 -3.60507011e-01 -1.96810842e-01
5.98998368e-01 6.63331077e-02 -1.06350124e+00 -7.01781288e-02
-7.82091677e-01 5.91851890e-01 9.14275467e-01 6.86625063e-01
6.29517138e-01 2.70161312e-02 4.90613282e-01 -1.02541590e+00
4.32148099e-01 -4.53067899e-01 -2.51612306e-01 3.60786207e-02
-3.46498340e-02 -1.96693718e-01 2.93410420e-01 -7.07690939e-02
-8.22685540e-01 3.07991862e-01 -2.80195981e-01 -1.65122926e-01
-3.84347916e-01 5.56359828e-01 1.83802128e-01 -1.70751110e-01
6.08098507e-01 1.03828035e-01 -9.89172980e-02 -1.50874019e-01
-2.59414405e-01 6.25371456e-01 6.82653189e-01 2.35301226e-01
3.66701126e-01 5.93010485e-01 3.17070007e-01 -7.38916218e-01
-9.33806181e-01 -5.90263724e-01 -2.62435198e-01 -3.55384946e-01
8.58793139e-01 -7.23886311e-01 -5.08807957e-01 2.29625463e-01
-9.85586405e-01 -1.34485513e-01 -4.32575755e-02 1.78414986e-01
-1.01783372e-01 8.59312490e-02 -7.48786867e-01 -2.12891698e-01
-5.73834360e-01 -9.24310625e-01 1.11271942e+00 7.40866303e-01
-2.61929035e-01 -1.17764890e+00 6.88137440e-03 -6.24564243e-03
3.21613222e-01 4.58108485e-01 1.12188876e+00 -6.17618561e-01
-4.34722126e-01 -3.50809455e-01 -4.06590134e-01 -1.75117210e-01
2.94756919e-01 1.40021294e-01 -8.63495409e-01 -3.99740040e-01
-2.02185929e-01 -7.39880949e-02 1.26304507e+00 8.03827524e-01
1.37391222e+00 5.09894826e-02 -1.13549674e+00 5.45771718e-01
1.31404686e+00 2.12573171e-01 4.03727233e-01 3.88586640e-01
5.81018984e-01 5.60808659e-01 6.34701073e-01 -6.43834621e-02
1.91221103e-01 2.96151966e-01 6.95513189e-01 -7.14200974e-01
-3.10898244e-01 -2.69639418e-02 -2.84459472e-01 2.08371356e-01
-3.95579338e-02 -8.11474919e-02 -1.27996457e+00 8.69240046e-01
-1.62966633e+00 -8.02497387e-01 -1.34920508e-01 1.82600236e+00
7.36586392e-01 8.77893344e-02 -1.54584974e-01 4.07629348e-02
8.47611010e-01 -2.45674565e-01 -6.42677546e-01 -1.76924482e-01
2.74103135e-01 1.22307921e-02 4.24947560e-01 1.27863362e-01
-1.40176356e+00 8.68366599e-01 5.77253580e+00 5.74949026e-01
-1.50004470e+00 2.71549709e-02 1.33501947e+00 8.95732492e-02
5.54025173e-02 -4.37713206e-01 -6.82917714e-01 3.96394432e-01
6.32581294e-01 -4.20093387e-01 -3.79213512e-01 8.39470148e-01
5.15440583e-01 -4.50939327e-01 -1.03459167e+00 6.81085110e-01
-2.24503711e-01 -1.44829929e+00 -2.50623887e-03 9.56767122e-04
6.15533650e-01 -1.96502939e-01 1.71431467e-01 7.76812434e-02
3.22826833e-01 -1.06520391e+00 3.95434462e-02 4.93254751e-01
7.84156322e-01 -9.26646113e-01 1.08095717e+00 4.49307710e-01
-7.67720044e-01 1.61972977e-02 -4.05923605e-01 3.40713084e-01
-2.08586708e-01 6.15525365e-01 -1.37071109e+00 -2.23872699e-02
6.45755649e-01 7.04288840e-01 -8.36993575e-01 1.32015979e+00
-2.98634917e-01 3.99245709e-01 8.01027045e-02 -3.30998778e-01
2.56904840e-01 5.31981170e-01 3.70498151e-01 1.18735111e+00
3.85317624e-01 8.25092494e-02 1.30070493e-01 5.89563966e-01
-1.66833922e-01 -1.16492435e-02 -4.70762700e-01 -3.71714085e-02
7.44429454e-02 1.56689823e+00 -1.27561152e+00 -2.98856258e-01
-1.07194655e-01 9.74769354e-01 2.96283871e-01 1.66395023e-01
-6.60535097e-01 -5.10486007e-01 2.87531167e-01 3.13254148e-01
1.72976255e-02 6.11608684e-01 -3.49148214e-01 -7.38117218e-01
-5.37446618e-01 -5.23336887e-01 4.41976845e-01 -7.20409453e-01
-8.28370988e-01 4.09166634e-01 -2.99112678e-01 -8.32181871e-01
-1.07242957e-01 -6.30490005e-01 -1.01018679e+00 5.68134904e-01
-1.58247662e+00 -9.89605248e-01 -5.05781770e-01 3.30399424e-01
5.98831236e-01 -1.09770939e-01 8.37879360e-01 -6.67657033e-02
-6.56424761e-01 3.90519887e-01 -1.56335875e-01 4.12329167e-01
5.37673771e-01 -1.40647960e+00 2.07748458e-01 6.84374988e-01
-4.19471055e-01 3.69397730e-01 7.29845524e-01 -6.77242219e-01
-7.46641219e-01 -1.40103877e+00 1.06975687e+00 1.07636765e-01
6.23936772e-01 -1.19579531e-01 -8.74466598e-01 4.85356092e-01
1.82131737e-01 1.54159486e-01 5.44505954e-01 -5.28678656e-01
4.59950536e-01 -1.04179755e-02 -1.29376864e+00 7.57981777e-01
3.96721154e-01 -1.90794498e-01 -3.36324096e-01 7.16294587e-01
4.82595295e-01 -4.25497174e-01 -4.38197643e-01 4.98293489e-01
1.97801903e-01 -7.69566298e-01 8.42507958e-01 -4.17892456e-01
7.55322695e-01 -1.64976045e-01 5.59903860e-01 -1.29301703e+00
-5.34051478e-01 8.31759349e-03 7.62848035e-02 6.71376407e-01
4.61981237e-01 -5.37954569e-01 1.07126319e+00 3.43888640e-01
-1.35249406e-01 -1.03629375e+00 -6.88715577e-01 6.81240065e-03
3.74579951e-02 3.47052753e-01 9.38740298e-02 7.38119543e-01
4.64855880e-02 9.27126780e-02 2.46796981e-01 8.65170136e-02
5.01646817e-01 5.70897348e-02 2.45038405e-01 -1.05780661e+00
-6.74247965e-02 -5.77096283e-01 -7.55470812e-01 -3.97861600e-01
-1.78866655e-01 -8.84447157e-01 -1.59695327e-01 -1.66094244e+00
7.94233084e-01 -1.60557136e-01 -5.98788559e-01 7.75864601e-01
-3.56291026e-01 4.93784904e-01 -2.71690816e-01 3.55884880e-02
-5.58080971e-01 6.42980114e-02 1.44311595e+00 -4.53669637e-01
1.67440936e-01 1.83172002e-01 -1.00835931e+00 9.99537289e-01
1.08483410e+00 -4.84385580e-01 -6.44825175e-02 2.26420872e-02
-3.98528464e-02 -9.50484304e-04 7.63777316e-01 -1.02005064e+00
3.36968660e-01 -2.01298937e-01 8.18058550e-01 -6.63650513e-01
-5.29081263e-02 -7.47129023e-01 -1.35976374e-01 1.13519919e+00
-5.52100062e-01 -2.92608470e-01 2.43216664e-01 3.05210948e-01
-1.66777626e-01 -2.93891788e-01 1.00509179e+00 -3.31255734e-01
-8.53986025e-01 2.65341997e-01 -5.97411335e-01 -4.87434596e-01
1.66909981e+00 -4.03975882e-02 -5.24486721e-01 -2.11018994e-01
-6.59763694e-01 2.80366302e-01 4.77935910e-01 1.68316841e-01
8.70912015e-01 -1.08082020e+00 -7.76124299e-01 9.74107981e-02
2.10311428e-01 -2.58724894e-02 2.15110272e-01 8.36727381e-01
-1.02515435e+00 4.69568849e-01 -2.22345874e-01 -4.53143924e-01
-1.52551854e+00 6.60052657e-01 4.71138626e-01 -3.45965147e-01
-5.65583289e-01 1.12621069e+00 5.23846149e-01 1.80302247e-01
3.13896567e-01 -1.80628553e-01 -5.98336697e-01 -4.21978444e-01
8.28387916e-01 3.30802985e-03 -1.66195519e-02 -4.36422199e-01
-5.37600815e-01 1.27793998e-01 -8.49456429e-01 4.58895028e-01
1.19516122e+00 2.76287943e-01 -2.27259159e-01 5.43407239e-02
1.09726453e+00 -3.00190747e-01 -8.86488736e-01 -4.07917798e-02
3.63756903e-02 -8.04213807e-02 4.02073771e-01 -9.12243724e-01
-1.25583744e+00 9.26194072e-01 9.13365722e-01 5.60485534e-02
1.17480075e+00 2.64303654e-01 5.57658076e-01 1.49154976e-01
2.27457583e-01 -6.95816398e-01 1.02984626e-02 -6.14564633e-03
6.18563116e-01 -1.43307018e+00 1.02365710e-01 -4.02316630e-01
-4.92696583e-01 1.08535600e+00 5.47844291e-01 -4.18630064e-01
6.88795090e-01 2.41854280e-01 3.61944318e-01 -4.51221406e-01
-8.69176865e-01 -5.34895733e-02 2.50460029e-01 4.91654336e-01
6.57724738e-01 1.96497142e-01 -2.71658927e-01 3.46812129e-01
1.33283436e-01 -1.32708937e-01 3.65991354e-01 9.95719671e-01
-9.22509372e-01 -6.36236966e-01 -2.21987844e-01 8.37474287e-01
-9.04781938e-01 -3.27452719e-02 -6.77147865e-01 6.43680990e-01
-2.23954488e-02 6.28804982e-01 3.11487019e-01 4.71426323e-02
-1.32907122e-01 -3.99233937e-01 9.72031034e-04 -6.92425013e-01
-2.95469880e-01 5.36875799e-02 -2.95435250e-01 -2.30948821e-01
-2.47059703e-01 -5.03854573e-01 -1.55817842e+00 -1.04500838e-01
-5.55002131e-02 7.55256116e-02 6.63577437e-01 6.50991201e-01
1.84588745e-01 8.78589809e-01 4.38425124e-01 -5.97873867e-01
-4.85213920e-02 -8.09045017e-01 -7.17855036e-01 1.96978718e-01
4.09783810e-01 -4.70466226e-01 -4.31546494e-02 6.15109056e-02] | [15.157504081726074, -2.8251090049743652] |
5017ad37-2eab-4929-a390-1c57ad9fd08f | deepseqslam-a-trainable-cnn-rnn-for-joint | 2011.08518 | null | https://arxiv.org/abs/2011.08518v1 | https://arxiv.org/pdf/2011.08518v1.pdf | DeepSeqSLAM: A Trainable CNN+RNN for Joint Global Description and Sequence-based Place Recognition | Sequence-based place recognition methods for all-weather navigation are well-known for producing state-of-the-art results under challenging day-night or summer-winter transitions. These systems, however, rely on complex handcrafted heuristics for sequential matching - which are applied on top of a pre-computed pairwise similarity matrix between reference and query image sequences of a single route - to further reduce false-positive rates compared to single-frame retrieval methods. As a result, performing multi-frame place recognition can be extremely slow for deployment on autonomous vehicles or evaluation on large datasets, and fail when using relatively short parameter values such as a sequence length of 2 frames. In this paper, we propose DeepSeqSLAM: a trainable CNN+RNN architecture for jointly learning visual and positional representations from a single monocular image sequence of a route. We demonstrate our approach on two large benchmark datasets, Nordland and Oxford RobotCar - recorded over 728 km and 10 km routes, respectively, each during 1 year with multiple seasons, weather, and lighting conditions. On Nordland, we compare our method to two state-of-the-art sequence-based methods across the entire route under summer-winter changes using a sequence length of 2 and show that our approach can get over 72% AUC compared to 27% AUC for Delta Descriptors and 2% AUC for SeqSLAM; while drastically reducing the deployment time from around 1 hour to 1 minute against both. The framework code and video are available at https://mchancan.github.io/deepseqslam | ['Michael Milford', 'Marvin Chancán'] | 2020-11-17 | null | null | null | null | ['sequential-image-classification', 'sequential-place-learning', 'sequential-place-recognition'] | ['computer-vision', 'robots', 'robots'] | [ 1.15291461e-01 -4.33698744e-01 -1.05748899e-01 -5.81258059e-01
-1.11260056e+00 -9.11957860e-01 7.48422623e-01 -4.14178818e-02
-1.10087025e+00 5.78023255e-01 -2.16964155e-01 -3.67698491e-01
2.46252209e-01 -5.41428387e-01 -1.09448087e+00 -6.33267820e-01
-2.96453744e-01 2.76499897e-01 4.66246605e-01 -3.35691005e-01
2.26178110e-01 4.29163754e-01 -1.82977068e+00 -4.36182171e-02
5.14077544e-01 1.00561249e+00 3.62754017e-01 9.98736322e-01
1.61168545e-01 5.08412182e-01 -4.84638214e-01 -2.40823671e-01
6.12718761e-01 -1.27735719e-01 -5.30737460e-01 -2.93370336e-01
8.27700019e-01 -3.87113094e-01 -5.44758439e-01 8.74190748e-01
6.76886559e-01 4.96085703e-01 3.62187594e-01 -1.33628035e+00
-2.06329063e-01 -1.24274053e-01 -2.06153303e-01 3.39037985e-01
4.65521991e-01 6.22793734e-01 9.10010219e-01 -7.96407282e-01
7.57720172e-01 9.15234327e-01 8.69696498e-01 3.56072247e-01
-9.32889104e-01 -6.19333923e-01 1.49936691e-01 5.62646389e-01
-1.50696516e+00 -6.67400777e-01 8.50081518e-02 -3.39382887e-01
1.37801456e+00 5.32824323e-02 5.36165118e-01 1.18450511e+00
3.16394895e-01 6.76521182e-01 7.14716733e-01 2.48434618e-01
2.60087520e-01 -3.73270720e-01 -3.59494418e-01 7.27958381e-01
4.95518409e-02 3.38440806e-01 -3.89919519e-01 1.00158207e-01
2.96770960e-01 1.53364614e-01 -2.31417403e-01 -4.96971428e-01
-1.30840945e+00 7.23520041e-01 6.89035416e-01 -1.23151347e-01
-3.69874477e-01 4.07592207e-01 6.29715025e-01 4.89548445e-01
2.44709626e-01 3.24088871e-01 -4.93248731e-01 -5.75972021e-01
-1.07297826e+00 5.75953245e-01 5.70386112e-01 1.17031777e+00
1.07459033e+00 -1.31175265e-01 2.19683975e-01 7.39460230e-01
8.51151124e-02 9.28069413e-01 3.65477860e-01 -8.96161437e-01
4.92725015e-01 1.97599128e-01 3.49786520e-01 -9.56060410e-01
-6.96951032e-01 -2.16639444e-01 -5.17217636e-01 1.22298459e-02
3.27125013e-01 -2.19100192e-01 -1.18155169e+00 1.66758919e+00
3.11679244e-01 3.05206239e-01 1.34153262e-01 1.17926002e+00
7.26870894e-01 8.24676394e-01 -2.42956340e-01 3.12889129e-01
1.13759243e+00 -1.32832098e+00 -2.09270179e-01 -7.03349531e-01
9.19808567e-01 -6.80656016e-01 9.02143061e-01 8.19010735e-02
-6.87490642e-01 -4.71093386e-01 -1.13277078e+00 -1.21850856e-01
-6.13343239e-01 5.01677059e-02 2.22060159e-01 3.32049698e-01
-1.25619423e+00 2.48949945e-01 -8.48964691e-01 -7.57035136e-01
1.59228027e-01 3.28049332e-01 -7.38992274e-01 -4.50505763e-01
-1.08755994e+00 1.00377905e+00 9.34237391e-02 4.15748179e-01
-1.19597602e+00 -6.93890035e-01 -1.14076972e+00 -2.19550416e-01
3.56749833e-01 -4.30035114e-01 1.31559610e+00 -7.65100062e-01
-1.28534448e+00 7.88192868e-01 -3.64044845e-01 -8.63975286e-01
6.27908409e-01 -5.45503616e-01 -3.63662094e-01 9.49546229e-03
3.55539024e-01 1.12495971e+00 3.90276343e-01 -1.04066277e+00
-8.51540089e-01 -9.87875015e-02 1.50385201e-01 3.34906757e-01
4.30237472e-01 -1.98222324e-01 -7.62943089e-01 -1.57581165e-01
6.78016385e-03 -1.41240263e+00 -5.58507085e-01 1.81834787e-01
3.39651247e-03 3.15411896e-01 6.39782786e-01 -5.27754247e-01
5.29625416e-01 -2.18735456e+00 -1.06175572e-01 1.48555428e-01
-3.46132904e-01 3.30383599e-01 -5.92277884e-01 5.86926699e-01
2.87326783e-01 -2.15045184e-01 -4.52027798e-01 -3.77206981e-01
-1.08284302e-01 4.67597455e-01 -1.71438053e-01 9.00099218e-01
5.44297956e-02 8.60684097e-01 -1.08345616e+00 2.95008309e-02
3.41849774e-01 4.64602470e-01 -4.60090131e-01 2.74006248e-01
-2.09662274e-01 3.83264959e-01 8.08294788e-02 6.19537234e-01
7.31979668e-01 3.74813713e-02 6.49944395e-02 1.85818017e-01
-3.15830469e-01 3.96298677e-01 -9.89946425e-01 2.07103539e+00
-6.20373428e-01 1.00386143e+00 -2.00756460e-01 -6.80830419e-01
9.87165689e-01 -6.99839145e-02 2.92076051e-01 -1.39143932e+00
-1.25846416e-01 4.22674745e-01 -1.97113156e-01 -3.38957876e-01
8.58051836e-01 3.35895389e-01 -2.91980743e-01 3.02689038e-02
-6.30460531e-02 -8.08394775e-02 3.76515508e-01 -6.16954193e-02
1.38842726e+00 2.61995584e-01 1.52652070e-01 5.76017192e-03
4.83435690e-01 2.18133509e-01 6.41469061e-01 8.83206785e-01
-3.67417485e-01 7.87387908e-01 2.01493829e-01 -8.01521301e-01
-1.21079147e+00 -8.88043344e-01 1.41540483e-01 1.00824988e+00
5.14295876e-01 -3.22896034e-01 -4.12042022e-01 -5.16124070e-01
6.34481758e-02 4.18040395e-01 -5.66470504e-01 8.12155679e-02
-6.32641137e-01 -5.67875445e-01 9.29667354e-01 3.96359324e-01
6.44836843e-01 -8.66471648e-01 -1.06620884e+00 2.16919750e-01
-2.57446378e-01 -1.49694252e+00 -3.78016263e-01 8.43190253e-02
-3.21751952e-01 -1.08134353e+00 -8.14668953e-01 -6.02927804e-01
4.10882145e-01 6.35055780e-01 1.03698337e+00 6.66348711e-02
-2.08429873e-01 2.43741930e-01 -4.68612015e-01 -1.02625184e-01
1.49107739e-01 3.51775229e-01 5.17251007e-02 -1.50389105e-01
2.10013658e-01 -4.03414190e-01 -8.50810885e-01 5.79496026e-01
-5.82480192e-01 -1.17352135e-01 5.62937796e-01 7.92103052e-01
6.25806689e-01 -6.78157747e-01 7.33113661e-02 -2.51863033e-01
-5.55975959e-02 -7.28059411e-01 -1.11312509e+00 1.10293152e-02
-3.92571509e-01 -3.39989476e-02 5.42848289e-01 -8.59845504e-02
-5.72233975e-01 2.01333851e-01 -2.02008203e-01 -5.33323228e-01
-2.08089888e-01 4.22291130e-01 2.44758680e-01 -2.46124223e-01
7.05073178e-01 5.02748787e-01 1.16263758e-02 -6.72337115e-02
3.56740654e-01 4.36495632e-01 7.58793652e-01 -3.04867744e-01
6.93214118e-01 7.22673297e-01 -7.16227107e-03 -7.88679183e-01
-5.16163409e-01 -9.85474110e-01 -4.81290549e-01 -1.15911968e-01
8.42201293e-01 -1.34469008e+00 -6.65169001e-01 4.99710470e-01
-1.00064003e+00 -8.08938682e-01 2.91637242e-01 5.30038118e-01
-7.55365431e-01 1.26700893e-01 -2.64868230e-01 -7.29043305e-01
-2.65060365e-01 -1.21090674e+00 1.26818180e+00 1.82630643e-01
5.16300127e-02 -6.32482231e-01 4.16913807e-01 1.90021992e-01
5.12399375e-01 2.06550792e-01 7.39140138e-02 -4.85144645e-01
-8.90823007e-01 -1.85236678e-01 -2.43560478e-01 -2.98126917e-02
-1.87473565e-01 -1.85479492e-01 -8.45262587e-01 -4.94416505e-01
-7.06904829e-01 -3.44607204e-01 1.02391338e+00 9.47634876e-02
6.88893259e-01 -2.14887142e-01 -3.65808398e-01 9.09832835e-01
1.58190191e+00 2.88596570e-01 7.21799195e-01 9.10062969e-01
5.68883836e-01 3.60512197e-01 1.12708569e+00 3.97283554e-01
7.68837452e-01 8.87688398e-01 7.28848100e-01 9.90371499e-03
1.98708609e-01 -1.26410261e-01 4.69662249e-01 1.62254870e-01
1.08582467e-01 -4.69919086e-01 -1.23460281e+00 1.04281950e+00
-1.99811280e+00 -1.12753320e+00 7.56004080e-02 2.27559447e+00
2.30348840e-01 -2.42707543e-02 1.16909482e-01 -3.67201954e-01
3.82556111e-01 5.87626874e-01 -5.66092789e-01 -2.95796365e-01
-1.79652512e-01 -3.00556421e-01 1.24353004e+00 5.69120347e-01
-1.25441170e+00 1.10928607e+00 5.33246565e+00 3.93672526e-01
-1.52643585e+00 1.18219963e-04 4.35156435e-01 -4.65926588e-01
-8.79465267e-02 -3.12527688e-03 -8.21218252e-01 3.98592889e-01
1.40041304e+00 2.47488856e-01 4.95388031e-01 9.02403235e-01
2.39584908e-01 -4.51522410e-01 -9.66663778e-01 1.24176991e+00
8.74938145e-02 -1.45298576e+00 -3.01717252e-01 1.21965759e-01
7.36731410e-01 1.20827055e+00 -6.62453175e-02 2.85813838e-01
2.60692447e-01 -1.09730697e+00 9.54235315e-01 2.83588618e-01
6.89922094e-01 -8.31846714e-01 9.69395936e-01 2.37709433e-01
-1.34623122e+00 -8.26531798e-02 -5.05549431e-01 2.26549618e-02
3.69948357e-01 2.27519363e-01 -9.18220103e-01 6.78474784e-01
1.04763520e+00 8.71537149e-01 -4.68533039e-01 1.37325215e+00
-8.23136047e-02 2.65817434e-01 -6.80107415e-01 -7.06793144e-02
9.11154568e-01 8.55533332e-02 4.59923297e-01 1.38812590e+00
5.61962545e-01 1.31910092e-05 2.94971287e-01 2.63133645e-01
6.65595010e-02 -1.29580036e-01 -9.16957080e-01 4.05708969e-01
5.46017647e-01 1.17871547e+00 -4.93512481e-01 -2.70956576e-01
-5.56189060e-01 9.46559548e-01 2.82355666e-01 3.71837467e-01
-1.15923214e+00 -4.04968828e-01 1.27772212e+00 -1.91142574e-01
7.10158885e-01 -7.08335221e-01 1.85165145e-02 -1.05287349e+00
2.28930920e-01 -6.70318484e-01 6.96067959e-02 -7.30087280e-01
-6.18683159e-01 7.44182289e-01 -1.69412300e-01 -1.49486959e+00
-4.17178869e-01 -4.58661795e-01 -3.49542826e-01 7.31568873e-01
-1.91967773e+00 -1.00850999e+00 -5.12364268e-01 4.43677038e-01
6.06589556e-01 2.62905378e-02 7.07348883e-01 5.82992911e-01
-2.67811298e-01 6.36842251e-01 3.13987851e-01 2.11127445e-01
7.56833792e-01 -8.93000603e-01 1.15589023e+00 1.01618612e+00
1.74404293e-01 3.17835361e-01 7.90684283e-01 -3.47234935e-01
-1.58446980e+00 -1.33085537e+00 9.59866703e-01 -2.95674324e-01
5.14817774e-01 -5.50221682e-01 -7.06566930e-01 6.55546308e-01
1.33449465e-01 4.47279543e-01 3.93331021e-01 -5.42526916e-02
-5.56081593e-01 -1.31484956e-01 -9.64160204e-01 7.66950011e-01
1.27408540e+00 -5.13863564e-01 -8.87142643e-02 2.84802765e-01
5.29584944e-01 -8.94541621e-01 -6.89774096e-01 3.82604092e-01
7.65571475e-01 -1.07785749e+00 8.22446048e-01 -1.02136865e-01
4.74027134e-02 -7.07688570e-01 -5.21969080e-01 -1.28036129e+00
-4.78726551e-02 -6.24202669e-01 5.12737095e-01 5.64875305e-01
7.04287887e-01 -6.89081669e-01 6.10741377e-01 1.71569183e-01
-3.14439833e-01 -6.84698820e-01 -1.18954396e+00 -9.86567318e-01
-4.06626374e-01 -6.34777486e-01 6.16975129e-01 5.01407504e-01
-5.42401910e-01 -3.24249417e-02 -4.27707672e-01 6.28313959e-01
3.05773348e-01 3.72639410e-02 1.26445389e+00 -7.32403934e-01
1.37674913e-01 -1.76118851e-01 -8.57739389e-01 -1.24590147e+00
2.14035034e-01 -5.41959465e-01 6.75926924e-01 -1.48019338e+00
-2.48734280e-01 -4.04560596e-01 -1.25920609e-01 5.86501718e-01
2.28168026e-01 3.62716347e-01 2.26279363e-01 1.61683977e-01
-9.27359581e-01 7.16186345e-01 6.91720724e-01 -1.36710450e-01
-4.02653636e-03 -3.81028473e-01 -3.24257128e-02 2.60681570e-01
6.81376696e-01 -4.79332089e-01 -4.40712810e-01 -6.77910745e-01
2.92056620e-01 2.44494379e-02 6.41841352e-01 -1.34148860e+00
3.74966532e-01 -1.31791651e-01 1.12446405e-01 -6.98879898e-01
4.88298357e-01 -6.52588725e-01 1.90211013e-01 4.34447259e-01
-1.17458425e-01 6.58948839e-01 4.99893129e-01 6.43184364e-01
-3.02779168e-01 3.08180958e-01 5.83643556e-01 -9.75980386e-02
-1.49775839e+00 3.75060499e-01 -3.82509232e-01 4.64844257e-02
1.03978288e+00 -3.04284722e-01 -4.28703010e-01 -4.60131556e-01
-2.30735809e-01 5.78893363e-01 7.03414798e-01 8.76363516e-01
5.71572125e-01 -1.11261380e+00 -6.02684617e-01 2.96692252e-01
4.59965110e-01 8.66115689e-02 3.91130537e-01 9.12329197e-01
-9.88067508e-01 6.82844758e-01 -3.55555832e-01 -9.93804455e-01
-1.20186877e+00 3.24217975e-01 5.39727747e-01 9.92599055e-02
-6.87545955e-01 8.00875962e-01 1.89861953e-01 -8.75198901e-01
2.20277440e-03 -5.17906785e-01 1.74507022e-01 -1.23371948e-02
4.71826196e-01 1.37766644e-01 3.02367598e-01 -9.60980475e-01
-7.83816397e-01 7.22631335e-01 -7.05052074e-03 -3.26633126e-01
1.18453276e+00 -3.42285126e-01 3.57213765e-01 2.77254045e-01
1.46170712e+00 -4.39886183e-01 -1.66677499e+00 -1.49808805e-02
-1.59106888e-02 -4.46597844e-01 -1.92287117e-01 -5.74634016e-01
-9.23502803e-01 6.24578536e-01 7.85330892e-01 -4.42512780e-01
7.48364687e-01 -2.56658047e-01 9.94608641e-01 8.24755490e-01
6.03952169e-01 -8.37257862e-01 -3.90652180e-01 9.87163723e-01
6.35926843e-01 -1.42639172e+00 -2.15592235e-01 1.07251927e-01
-6.50333703e-01 8.37939560e-01 5.43476760e-01 -2.20630154e-01
3.28359574e-01 -2.90516764e-03 4.29887444e-01 2.19714865e-02
-9.09584284e-01 -4.57062244e-01 2.17599750e-01 4.64729726e-01
1.10792235e-01 -3.63994054e-02 1.09468006e-01 -1.98863432e-01
-3.86049658e-01 -7.81062692e-02 4.39364612e-01 1.14572477e+00
-4.01436359e-01 -6.61864817e-01 -1.03677206e-01 3.28337476e-02
-1.84374154e-01 -1.94452584e-01 1.41784931e-02 8.16053748e-01
-3.56469750e-02 8.69170845e-01 3.57726485e-01 -5.22546053e-01
4.37545061e-01 -8.99930447e-02 1.64296687e-01 -2.54107833e-01
-4.98909444e-01 -2.67251283e-01 4.15304989e-01 -1.28789914e+00
-4.36599314e-01 -9.45691466e-01 -1.22903609e+00 -4.79954332e-01
1.49469957e-01 -2.85001118e-02 9.20095205e-01 1.05049670e+00
6.72472596e-01 -2.57465802e-02 4.68141407e-01 -1.29300296e+00
-1.00791797e-01 -6.06291890e-01 -5.10093160e-02 2.38382984e-02
8.63906622e-01 -5.81724823e-01 -2.18988240e-01 -2.71976531e-01] | [7.614264965057373, -1.9346951246261597] |
8f7aa095-fd87-4f6c-b7f3-0c7cbbc5f66d | semantic-guided-image-augmentation-with-pre | 2302.02070 | null | https://arxiv.org/abs/2302.02070v1 | https://arxiv.org/pdf/2302.02070v1.pdf | Semantic-Guided Image Augmentation with Pre-trained Models | Image augmentation is a common mechanism to alleviate data scarcity in computer vision. Existing image augmentation methods often apply pre-defined transformations or mixup to augment the original image, but only locally vary the image. This makes them struggle to find a balance between maintaining semantic information and improving the diversity of augmented images. In this paper, we propose a Semantic-guided Image augmentation method with Pre-trained models (SIP). Specifically, SIP constructs prompts with image labels and captions to better guide the image-to-image generation process of the pre-trained Stable Diffusion model. The semantic information contained in the original images can be well preserved, and the augmented images still maintain diversity. Experimental results show that SIP can improve two commonly used backbones, i.e., ResNet-50 and ViT, by 12.60% and 2.07% on average over seven datasets, respectively. Moreover, SIP not only outperforms the best image augmentation baseline RandAugment by 4.46% and 1.23% on two backbones, but also further improves the performance by integrating naturally with the baseline. A detailed analysis of SIP is presented, including the diversity of augmented images, an ablation study on textual prompts, and a case study on the generated images. | ['Wanxiang Che', 'Feng Wang', 'Yunlong Feng', 'Yutai Hou', 'Xiao Xu', 'Xinghao Wang', 'Bohan Li'] | 2023-02-04 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 7.03625858e-01 3.52165788e-01 -2.90043473e-01 -2.39688754e-01
-5.36144614e-01 -4.41250861e-01 7.35522568e-01 -2.95402765e-01
-5.48523307e-01 6.82605624e-01 3.79242748e-01 -7.28371665e-02
4.90394562e-01 -6.20901883e-01 -7.76401281e-01 -8.45174611e-01
4.86058831e-01 2.42561430e-01 1.29988521e-01 -4.03784066e-01
7.91582540e-02 1.12158865e-01 -1.38721895e+00 1.65949389e-01
1.13086915e+00 7.93398976e-01 5.11300504e-01 4.45862055e-01
-3.72390777e-01 9.06529725e-01 -7.65629709e-01 -4.47063953e-01
2.37497464e-01 -6.87500596e-01 -8.30944717e-01 6.54070437e-01
6.64399624e-01 -5.11652052e-01 -6.53279841e-01 1.20791745e+00
4.60912555e-01 1.31582186e-01 4.20370728e-01 -1.60751867e+00
-1.30734682e+00 6.99396193e-01 -8.19918811e-01 8.19751024e-02
-1.44236475e-01 3.45385164e-01 5.59810460e-01 -8.21779966e-01
8.04727316e-01 1.23174155e+00 4.12744403e-01 9.50018406e-01
-1.09568012e+00 -7.69385576e-01 3.98537666e-01 1.24256887e-01
-1.07758725e+00 -3.24371934e-01 8.05446327e-01 -1.44262478e-01
4.21703726e-01 2.71984667e-01 4.58999753e-01 1.23066545e+00
-4.34625298e-01 9.48232591e-01 1.25395727e+00 -4.56046671e-01
-1.19264722e-01 2.31950954e-01 3.62649970e-02 5.79530716e-01
2.35252991e-01 -3.43221314e-02 -3.90202940e-01 2.87608802e-01
1.02687907e+00 4.68757376e-02 -4.25991446e-01 -5.51171089e-03
-1.41156340e+00 8.04458261e-01 8.58516812e-01 2.29799449e-01
-5.22671461e-01 6.22661747e-02 3.03914994e-01 -5.16760461e-02
5.43806434e-01 5.77634811e-01 -3.63738537e-01 3.56810987e-01
-6.02394342e-01 1.40084758e-01 2.52798378e-01 1.03966761e+00
9.00861025e-01 4.27071154e-01 -5.72792351e-01 1.04466724e+00
8.43966305e-02 7.19854772e-01 7.53041446e-01 -1.08567655e+00
4.64450300e-01 5.04430115e-01 -2.36373805e-02 -8.91100109e-01
-1.73966020e-01 -6.18912339e-01 -1.20983672e+00 2.64107492e-02
1.08771980e-01 -1.28317744e-01 -1.59378624e+00 2.05225134e+00
1.89305976e-01 4.17109638e-01 1.23653494e-01 7.77642846e-01
1.18904591e+00 8.29940259e-01 3.00954789e-01 -3.15761678e-02
1.25552368e+00 -1.74838829e+00 -1.10951865e+00 -5.64925492e-01
5.24649441e-01 -7.49418557e-01 1.38849592e+00 -8.27629790e-02
-1.07524526e+00 -7.61758208e-01 -1.05545473e+00 7.35334679e-02
-2.08307147e-01 1.61370710e-02 4.41448420e-01 6.06673837e-01
-1.35126197e+00 9.57349241e-02 -3.89173687e-01 -2.06504822e-01
5.72987020e-01 9.75384861e-02 -4.19482946e-01 -4.36039388e-01
-1.13528252e+00 7.40176260e-01 4.70853716e-01 -3.43110204e-01
-8.28110993e-01 -8.46677423e-01 -8.94209027e-01 -1.46437302e-01
3.57244790e-01 -7.32322156e-01 1.26879263e+00 -1.16101360e+00
-1.26321292e+00 8.07387054e-01 -1.34483501e-01 -3.88038814e-01
4.93093401e-01 -6.69467300e-02 -3.61917198e-01 1.63994670e-01
4.76714671e-01 1.60154784e+00 8.43125224e-01 -1.77637255e+00
-6.01371288e-01 -1.15767695e-01 1.07359327e-02 4.63969201e-01
-7.81295478e-01 -2.67513007e-01 -1.06741846e+00 -1.09072518e+00
1.40371248e-02 -1.17997217e+00 -6.69437170e-01 -2.19577789e-01
-6.96975529e-01 1.47027656e-01 1.07162392e+00 -8.15652192e-01
9.30721879e-01 -2.03831077e+00 -9.41754058e-02 -4.58616577e-02
2.03444049e-01 4.78652358e-01 -8.68305981e-01 -1.26462383e-02
-1.67533129e-01 3.41555327e-01 -6.19746685e-01 -6.90672040e-01
-4.51305240e-01 5.91263056e-01 -3.21930349e-01 1.59967039e-02
2.78199643e-01 1.35576797e+00 -8.77905190e-01 -4.83330876e-01
3.22684348e-01 6.17909253e-01 -4.89518344e-01 2.12327346e-01
-2.85796016e-01 9.15890813e-01 -1.04062766e-01 3.50489974e-01
8.84366155e-01 -5.80326140e-01 -1.34753495e-01 -2.56794959e-01
2.50119179e-01 -1.39964953e-01 -8.17345858e-01 1.90538669e+00
-3.94692153e-01 6.10326052e-01 -3.91398519e-01 -7.18485296e-01
8.96379828e-01 1.19728222e-01 4.74571705e-01 -1.20770407e+00
-8.70573223e-02 -7.11530959e-03 -2.03731805e-01 -3.31094503e-01
8.27716112e-01 2.35811993e-01 2.20100373e-01 3.72960478e-01
2.17210129e-03 -5.07135168e-02 3.67542148e-01 5.23894429e-01
7.65533984e-01 -6.11174591e-02 -2.07338214e-01 9.34267044e-02
4.56774384e-01 3.03830169e-02 5.42375088e-01 8.01536918e-01
-1.11681357e-01 1.01140296e+00 1.41058341e-01 -1.90010682e-01
-1.39436090e+00 -7.84631550e-01 1.51692912e-01 9.28954482e-01
5.60643852e-01 -1.47332564e-01 -9.36384439e-01 -1.00674832e+00
-3.84572536e-01 8.50581944e-01 -7.50493050e-01 -3.09221894e-01
-4.60919112e-01 -9.23854649e-01 5.57509065e-01 5.23543835e-01
1.26624179e+00 -1.13674879e+00 9.08288881e-02 2.09683776e-02
-7.82007396e-01 -1.61521614e+00 -7.52436936e-01 -3.49565029e-01
-9.35458541e-01 -7.47732580e-01 -1.09278178e+00 -1.09704828e+00
1.12651527e+00 8.34042192e-01 1.01255763e+00 4.42283750e-01
1.61293998e-01 1.85114041e-01 -6.57889545e-01 -2.76887983e-01
-6.26444519e-01 2.06690356e-01 -1.92994505e-01 1.87751651e-01
-2.04152808e-01 -4.00244236e-01 -6.81080997e-01 5.20638525e-01
-1.42580259e+00 4.93093818e-01 8.11194420e-01 1.06550181e+00
6.05364561e-01 -8.28476250e-02 7.37686038e-01 -9.07319903e-01
4.29732949e-01 -2.39957571e-01 -8.52799267e-02 2.70272553e-01
-8.61581743e-01 2.70333514e-02 2.14935839e-01 -6.60050511e-01
-1.32005250e+00 2.22893074e-01 -2.01930493e-01 -5.71227551e-01
-8.99549946e-02 2.66194135e-01 -2.66696066e-01 -2.24712461e-01
7.13766754e-01 6.01805329e-01 1.84261292e-01 -3.17699045e-01
9.28924263e-01 4.50794697e-01 1.03894734e+00 -2.31529713e-01
8.09535563e-01 5.91455162e-01 -4.44400728e-01 -7.03732193e-01
-8.84703279e-01 -2.35454187e-01 -3.44425648e-01 -2.01466173e-01
7.54478097e-01 -1.08948386e+00 -4.41919714e-02 6.46660507e-01
-1.25671458e+00 -4.61629927e-01 -5.93637884e-01 2.02438921e-01
-3.82145792e-01 4.17809635e-01 -5.51106036e-01 -3.00242424e-01
-3.55939656e-01 -1.32833648e+00 9.08662319e-01 4.14421082e-01
4.27484289e-02 -8.63151491e-01 -1.98927835e-01 6.17384791e-01
5.68939805e-01 1.84969172e-01 6.90721989e-01 -4.47650373e-01
-5.48047423e-01 -6.83946982e-02 -5.58596969e-01 6.73333406e-01
3.72935802e-01 -3.46776456e-01 -1.01298249e+00 -1.53460711e-01
-1.39809996e-01 -2.08164304e-01 1.07260156e+00 3.60261738e-01
1.23214102e+00 -5.51478863e-01 -3.34652275e-01 6.46370888e-01
1.15571558e+00 3.79521817e-01 1.14845002e+00 6.14453554e-01
9.09759164e-01 5.22445261e-01 5.69281518e-01 1.34908617e-01
4.47262257e-01 5.44441402e-01 7.32703745e-01 -6.70843542e-01
-9.18315113e-01 -4.43096250e-01 1.83730811e-01 8.98084283e-01
6.99202642e-02 -3.40012312e-01 -6.32077575e-01 7.14233994e-01
-1.82585275e+00 -6.30147815e-01 -2.42653728e-01 1.81786394e+00
1.04312587e+00 -1.79365411e-01 -7.62512460e-02 -2.55207401e-02
1.03314924e+00 2.55900264e-01 -7.12266922e-01 2.16413900e-01
-5.72207570e-01 -1.12057440e-01 6.33253634e-01 5.10182559e-01
-1.02050900e+00 1.25535274e+00 6.23689604e+00 9.21486378e-01
-1.04445279e+00 3.56520414e-01 1.11755490e+00 8.83324817e-02
-4.60385412e-01 -1.87775284e-01 -5.83741963e-01 5.72148025e-01
4.70810533e-01 1.51075050e-02 2.53733218e-01 7.77000904e-01
1.03151702e-04 1.62464872e-01 -5.35487890e-01 1.15888512e+00
3.17683011e-01 -1.52848589e+00 6.60157800e-01 1.71905886e-02
1.48201168e+00 1.53700421e-02 5.63692689e-01 8.88022333e-02
4.10929471e-01 -8.29131007e-01 5.76699913e-01 2.00990781e-01
9.46362376e-01 -5.76715469e-01 7.14014769e-01 8.00816119e-02
-8.88364971e-01 1.23766609e-01 -8.41324851e-02 2.97422886e-01
2.94972688e-01 4.97689009e-01 -7.15607882e-01 4.05578583e-01
6.05666518e-01 7.12068141e-01 -7.97245324e-01 9.54216540e-01
-6.13075256e-01 5.06919324e-01 -6.49123043e-02 3.56220365e-01
4.87362385e-01 -1.42479688e-01 3.93166572e-01 1.02047026e+00
1.40753642e-01 7.27795884e-02 1.62312910e-01 7.03772545e-01
-5.43537676e-01 4.69198311e-03 -5.90198636e-01 -1.72514804e-02
4.98553038e-01 1.32069671e+00 -7.44703650e-01 -7.18930662e-01
-3.94816309e-01 1.38447595e+00 -8.97406936e-02 7.10489810e-01
-1.02570570e+00 -1.96810350e-01 5.90637922e-01 -1.31973132e-01
4.26664501e-02 -2.47296505e-02 -5.18503189e-01 -1.05919814e+00
-6.23539239e-02 -1.21126771e+00 2.20777079e-01 -1.08743393e+00
-1.14390814e+00 1.03258884e+00 -2.16816425e-01 -9.94786203e-01
-1.58544313e-02 -9.14197639e-02 -4.33700055e-01 8.62490773e-01
-1.74757051e+00 -1.35471916e+00 -8.35056007e-01 7.32170999e-01
9.19999242e-01 -2.30679065e-01 5.63523114e-01 4.83767122e-01
-6.80049837e-01 7.82831252e-01 -4.11660112e-02 1.26304120e-01
7.51825750e-01 -9.63004768e-01 7.38539815e-01 1.07643509e+00
2.32911244e-01 1.71882659e-01 5.25274694e-01 -8.56534779e-01
-8.36233199e-01 -1.44294608e+00 5.18673837e-01 -1.83958486e-01
2.54871339e-01 6.06013089e-02 -1.16569781e+00 6.91479564e-01
7.29465783e-01 -1.45401269e-01 2.21590787e-01 -4.05296385e-01
-5.01100242e-01 1.14839420e-01 -1.10196936e+00 9.81729150e-01
1.19711804e+00 -2.65033662e-01 -1.29253283e-01 4.21418369e-01
1.31927621e+00 -4.62546855e-01 -4.84248400e-01 4.85483557e-01
-3.20497304e-02 -6.05390429e-01 1.21123385e+00 -5.45008183e-01
6.45604372e-01 -2.89337873e-01 -8.89231861e-02 -1.49708855e+00
-4.37235624e-01 -5.54992557e-01 8.39663818e-02 1.46588063e+00
4.28385168e-01 -5.80319405e-01 9.62822855e-01 6.35490239e-01
-1.53537363e-01 -4.54105675e-01 -4.35063988e-01 -7.99272358e-01
-1.77091464e-01 -2.54800290e-01 6.93927467e-01 1.33170056e+00
-6.92896068e-01 5.37025988e-01 -5.71795583e-01 1.51944652e-01
6.46055281e-01 -3.96509439e-01 8.38868022e-01 -6.00549579e-01
9.55391899e-02 -5.54170847e-01 -1.61210045e-01 -1.15249693e+00
1.73076261e-02 -9.15340185e-01 -1.93009786e-02 -1.67990243e+00
3.87191892e-01 -6.39149964e-01 -3.29586834e-01 8.64372313e-01
-6.33612692e-01 7.86230206e-01 3.95189106e-01 4.16896135e-01
-3.93154889e-01 8.87964129e-01 1.80238593e+00 -4.64848250e-01
-2.63813764e-01 -4.16816831e-01 -1.16653037e+00 6.59553349e-01
1.14449978e+00 -4.40070987e-01 -7.76445389e-01 -7.29936302e-01
-2.04244122e-01 -3.90104473e-01 3.65148515e-01 -8.19177747e-01
1.19849220e-01 1.08005116e-02 3.51143062e-01 -4.76366609e-01
2.69640356e-01 -5.24034798e-01 -1.79007128e-02 4.69660163e-01
-4.85377699e-01 1.81488246e-01 3.93408179e-01 5.68851888e-01
-3.35473359e-01 -2.68483728e-01 8.92553627e-01 -8.74481276e-02
-1.07846892e+00 5.69275379e-01 -2.48025451e-02 2.00614762e-02
1.00630569e+00 -1.22758932e-01 -6.06736004e-01 -7.30432212e-01
-6.18283510e-01 3.50771010e-01 3.93359333e-01 7.91776538e-01
8.14053535e-01 -1.65888608e+00 -9.13394392e-01 1.07855305e-01
1.59128606e-01 6.03407174e-02 5.48093081e-01 6.13211572e-01
-3.68831336e-01 1.07081123e-01 -3.25421184e-01 -6.87811255e-01
-1.30996811e+00 6.68816566e-01 1.64692014e-01 -2.27308139e-01
-6.84513986e-01 9.76571262e-01 6.09397709e-01 -3.59710366e-01
2.42210954e-01 -4.21886612e-03 -2.15792999e-01 -6.07919097e-02
8.45321894e-01 1.22173823e-01 -1.53974146e-01 -7.01503098e-01
4.69717085e-02 5.26080787e-01 -4.34825152e-01 -3.32347840e-01
1.11816263e+00 -4.58342999e-01 -1.41134024e-01 -2.51661986e-01
1.13147938e+00 -4.34400171e-01 -1.49295700e+00 -7.58804560e-01
-4.67528999e-01 -6.56370997e-01 1.09805338e-01 -1.09833026e+00
-1.82150865e+00 6.25032902e-01 6.33888602e-01 3.12240943e-02
1.39787591e+00 6.31156843e-04 1.01976669e+00 1.16003089e-01
-7.73350596e-02 -6.90445125e-01 7.32492745e-01 3.11634511e-01
1.21709478e+00 -1.39979756e+00 -3.04644078e-01 -6.81548715e-01
-1.16029596e+00 5.12676954e-01 1.00555563e+00 2.36876547e-01
1.42076865e-01 -4.34786715e-02 4.65460449e-01 1.40203446e-01
-3.82732064e-01 -4.17171806e-01 2.13637650e-01 9.85432744e-01
1.76754408e-02 -2.43248850e-01 -4.66624321e-03 1.97706312e-01
-8.25401470e-02 -1.73119754e-01 5.92609882e-01 6.30546808e-01
-3.08897316e-01 -1.16751242e+00 -4.18160170e-01 2.23115116e-01
-2.35543549e-01 -3.99565697e-01 -3.80766690e-01 7.90343165e-01
1.52683347e-01 1.07513547e+00 9.49778333e-02 -5.03511727e-01
3.41912150e-01 -1.72045931e-01 2.99264938e-01 -3.35305005e-01
-2.70261079e-01 2.30343230e-02 -4.40595075e-02 -3.65713626e-01
-6.26475990e-01 -2.43545324e-01 -1.16950381e+00 -3.21767300e-01
-3.64538163e-01 -5.94426692e-03 8.68774891e-01 6.55877590e-01
5.48106313e-01 9.26661074e-01 7.12031126e-01 -8.28667402e-01
-1.79798529e-01 -9.74128008e-01 -2.31275603e-01 8.09152424e-01
2.01719508e-01 -4.33930397e-01 -9.22491774e-02 4.34138387e-01] | [11.164286613464355, -0.5688377022743225] |
baab3bc8-7560-4217-bc28-36eb1c88a24b | segmenting-unknown-3d-objects-from-real-depth | 1809.05825 | null | http://arxiv.org/abs/1809.05825v2 | http://arxiv.org/pdf/1809.05825v2.pdf | Segmenting Unknown 3D Objects from Real Depth Images using Mask R-CNN Trained on Synthetic Data | The ability to segment unknown objects in depth images has potential to
enhance robot skills in grasping and object tracking. Recent computer vision
research has demonstrated that Mask R-CNN can be trained to segment specific
categories of objects in RGB images when massive hand-labeled datasets are
available. As generating these datasets is time consuming, we instead train
with synthetic depth images. Many robots now use depth sensors, and recent
results suggest training on synthetic depth data can transfer successfully to
the real world. We present a method for automated dataset generation and
rapidly generate a synthetic training dataset of 50,000 depth images and
320,000 object masks using simulated heaps of 3D CAD models. We train a variant
of Mask R-CNN with domain randomization on the generated dataset to perform
category-agnostic instance segmentation without any hand-labeled data and we
evaluate the trained network, which we refer to as Synthetic Depth (SD) Mask
R-CNN, on a set of real, high-resolution depth images of challenging,
densely-cluttered bins containing objects with highly-varied geometry. SD Mask
R-CNN outperforms point cloud clustering baselines by an absolute 15% in
Average Precision and 20% in Average Recall on COCO benchmarks, and achieves
performance levels similar to a Mask R-CNN trained on a massive, hand-labeled
RGB dataset and fine-tuned on real images from the experimental setup. We
deploy the model in an instance-specific grasping pipeline to demonstrate its
usefulness in a robotics application. Code, the synthetic training dataset, and
supplementary material are available at https://bit.ly/2letCuE. | ['Ken Goldberg', 'Saurabh Gupta', 'Andrew Li', 'Matthew Matl', 'Andrew Lee', 'Michael Danielczuk', 'Jeffrey Mahler'] | 2018-09-16 | null | null | null | null | ['unseen-object-instance-segmentation'] | ['computer-vision'] | [ 4.17035580e-01 3.17190289e-01 2.04532146e-01 -4.54161644e-01
-9.31603551e-01 -8.96722198e-01 3.08323205e-01 -2.04235345e-01
-4.58899021e-01 2.35162705e-01 -4.78093445e-01 -2.66752005e-01
1.73703939e-01 -8.27250957e-01 -1.49498653e+00 -4.73257720e-01
-1.95962518e-01 1.26268375e+00 4.61165965e-01 3.72279286e-02
2.92028248e-01 8.87075722e-01 -1.91439664e+00 5.16487658e-01
4.96985137e-01 1.12982714e+00 6.00885689e-01 7.55808949e-01
-7.13714212e-02 2.64855415e-01 -8.06569695e-01 -5.47020212e-02
1.10863841e+00 1.85156226e-01 -8.23448002e-01 3.77428740e-01
6.29360080e-01 -7.41157889e-01 -2.46617004e-01 8.81991208e-01
4.54522669e-01 -1.99420497e-01 5.28835416e-01 -1.51003861e+00
-4.63243574e-01 5.79661429e-01 -5.09241402e-01 -2.12124735e-01
1.01413466e-01 8.13249588e-01 3.23672712e-01 -9.22872722e-01
7.70382941e-01 1.47242701e+00 7.42949605e-01 7.02700734e-01
-1.38614917e+00 -7.22569108e-01 -5.38563356e-02 -2.99899906e-01
-1.10452592e+00 5.80779426e-02 5.34828186e-01 -4.95403588e-01
1.11909187e+00 -3.35969508e-01 7.86489606e-01 1.29065967e+00
-1.20311692e-01 8.44259858e-01 9.74587917e-01 -7.89260119e-02
5.03181100e-01 -3.18702281e-01 -1.04077891e-01 6.44252419e-01
5.27691007e-01 3.07304054e-01 -1.36423990e-01 1.19542003e-01
1.16582203e+00 -4.06414270e-04 5.99436983e-02 -1.03553057e+00
-1.64243960e+00 6.95804715e-01 7.85380542e-01 -3.14229786e-01
-4.90250587e-01 5.69058418e-01 2.86224067e-01 8.53948109e-03
8.90623108e-02 6.71731949e-01 -8.06449234e-01 4.56175208e-03
-4.69332039e-01 6.77805424e-01 8.11026216e-01 1.56264055e+00
9.07183766e-01 -1.58475161e-01 1.92540407e-01 5.47530234e-01
-8.62410814e-02 7.84313023e-01 2.97206938e-01 -1.68002570e+00
4.99417782e-01 7.61940598e-01 2.64020294e-01 -4.77155834e-01
-6.60158336e-01 -6.35871440e-02 -3.68745357e-01 7.67843366e-01
8.56994331e-01 -1.75163373e-01 -1.73702812e+00 1.25386417e+00
4.53602701e-01 -2.55449355e-01 3.38566080e-02 1.05288780e+00
8.92496467e-01 3.78407091e-01 -8.97666663e-02 7.48343647e-01
1.07225108e+00 -8.15952480e-01 3.30539718e-02 -6.09068036e-01
3.57145756e-01 -5.01547933e-01 1.16824627e+00 6.37085676e-01
-1.09972823e+00 -5.51358879e-01 -9.48214173e-01 -1.02999702e-01
-5.72334945e-01 1.81929842e-01 8.91877472e-01 3.01289737e-01
-9.57238317e-01 7.58810699e-01 -1.12806177e+00 -4.26822901e-01
1.10720062e+00 5.86509585e-01 -4.26198184e-01 -6.13761544e-01
-3.69394720e-01 6.78533614e-01 7.61033893e-01 2.38260888e-02
-1.51074696e+00 -8.14260781e-01 -8.70727539e-01 -4.56687540e-01
2.98670560e-01 -4.38100159e-01 1.51673818e+00 -6.68264568e-01
-1.15845799e+00 1.16188586e+00 4.71063048e-01 -3.68317574e-01
7.18269229e-01 -2.79037356e-01 5.27667165e-01 5.39700747e-01
3.25363934e-01 1.57611525e+00 8.20235908e-01 -1.76028001e+00
-7.61806786e-01 -6.16269588e-01 2.21388891e-01 -4.13152613e-02
4.65146363e-01 -4.69217777e-01 -5.27796030e-01 -3.41473401e-01
6.53274596e-01 -1.04383409e+00 -3.85653496e-01 3.58716100e-01
-4.35572475e-01 6.02080822e-02 8.97011399e-01 -2.97770321e-01
-3.75772595e-01 -1.87226677e+00 6.95308894e-02 1.01389840e-01
-1.22496344e-01 5.16257528e-03 -3.47315907e-01 3.80023047e-02
9.50552076e-02 -1.02151312e-01 -6.40213251e-01 -2.78548419e-01
7.78486282e-02 5.80916047e-01 -2.23155349e-01 5.65784335e-01
5.77322125e-01 9.62412655e-01 -1.07385993e+00 -2.94323534e-01
4.78389561e-01 3.13116133e-01 -7.66079068e-01 3.36463749e-01
-9.42945659e-01 6.27959609e-01 -1.35631755e-01 1.14588642e+00
9.53065693e-01 -2.12765172e-01 -9.77426618e-02 -3.51777136e-01
1.84779003e-01 -1.10506274e-01 -1.03949618e+00 2.22635961e+00
-1.80495083e-01 3.75680774e-01 2.56051779e-01 -9.01693165e-01
9.36769366e-01 -2.90929973e-01 6.18075490e-01 -4.17988002e-01
3.00443649e-01 2.15267658e-01 -7.48432204e-02 -6.16781473e-01
4.34705824e-01 1.27794802e-01 -1.03926316e-01 4.03823704e-01
2.33191222e-01 -1.05709231e+00 1.59170955e-01 1.32793903e-01
1.37343252e+00 6.86653435e-01 -4.78727430e-01 -2.87571438e-02
-4.81193632e-01 8.54804993e-01 9.28209275e-02 8.38611782e-01
-2.33424395e-01 1.11189926e+00 2.97809124e-01 -5.16265213e-01
-1.44473290e+00 -1.23973250e+00 -2.28584811e-01 7.44943380e-01
5.95545709e-01 3.47825378e-01 -7.80173481e-01 -6.30953789e-01
6.15532756e-01 4.60112751e-01 -8.15931976e-01 3.42389569e-02
-7.58190989e-01 -5.39098680e-01 5.67440391e-01 8.45731854e-01
6.39200211e-01 -1.61461234e+00 -1.36882937e+00 1.66201591e-01
4.58367951e-02 -1.29901052e+00 3.27350557e-01 7.84005344e-01
-9.56978798e-01 -1.49176526e+00 -6.89616919e-01 -8.96406949e-01
7.77338564e-01 4.52824920e-01 1.36753833e+00 -7.36064166e-02
-7.98233032e-01 6.04199648e-01 -6.22820914e-01 -7.35144615e-01
-3.04735482e-01 2.60673851e-01 -1.44397527e-01 -7.71748662e-01
4.65437651e-01 -3.92815411e-01 -8.19162190e-01 3.33074510e-01
-1.06077838e+00 -4.12732363e-02 7.62773991e-01 6.91434383e-01
7.09478498e-01 -2.67941236e-01 4.28345293e-01 -5.68131328e-01
1.27350511e-02 -4.27868426e-01 -7.73636222e-01 -2.46427208e-01
-1.38845161e-01 -3.37142497e-02 1.79828331e-01 -6.38967633e-01
-5.55335522e-01 6.77141249e-01 -5.14030531e-02 -8.56657028e-01
-5.49494505e-01 -2.10844710e-01 -2.03116342e-01 -1.29713699e-01
1.02092493e+00 -1.65803894e-01 2.59473622e-01 -2.98902512e-01
6.94977224e-01 4.26483303e-01 8.85893345e-01 -1.01602209e+00
8.13275158e-01 8.20367932e-01 -1.78057805e-01 -2.28507280e-01
-6.40577674e-01 -3.92062277e-01 -8.34129512e-01 1.21769942e-02
9.79293048e-01 -1.03829908e+00 -7.94976592e-01 6.28947854e-01
-1.23008025e+00 -1.28213847e+00 -5.19696593e-01 2.92247713e-01
-1.08129132e+00 -1.90370560e-01 -4.49933141e-01 -4.34818923e-01
6.45667091e-02 -1.17483079e+00 1.64434969e+00 1.00965435e-02
1.85643643e-01 -3.10648084e-01 -5.48405051e-01 2.19079643e-01
4.97916713e-02 7.32913792e-01 6.16875052e-01 -4.54414368e-01
-1.04328132e+00 -2.46212915e-01 -5.40660560e-01 4.58175749e-01
1.62624151e-01 -1.24702975e-01 -8.86061311e-01 -2.35411808e-01
-3.16401541e-01 -8.28466058e-01 7.68953025e-01 3.32162887e-01
1.57679689e+00 5.21780923e-02 -4.85181391e-01 6.63497806e-01
1.45403719e+00 2.26613246e-02 5.46802878e-01 5.34181833e-01
9.03554320e-01 6.71980143e-01 9.37536597e-01 3.81276280e-01
2.90524751e-01 3.13563645e-01 1.22353482e+00 -1.46866918e-01
-2.05563158e-01 -1.15119196e-01 -1.26014039e-01 -2.14109734e-01
1.98068723e-01 3.68637629e-02 -1.30984855e+00 8.92015636e-01
-1.55665326e+00 -7.27366090e-01 1.07827820e-01 1.76057208e+00
7.66622961e-01 3.40845108e-01 3.47830713e-01 1.10120013e-01
6.32213712e-01 -5.10104060e-01 -1.11087596e+00 1.46076933e-01
-6.07556030e-02 4.73903745e-01 1.04979491e+00 1.93416744e-01
-1.16275215e+00 1.21223259e+00 5.64573479e+00 2.54036099e-01
-9.64460731e-01 -1.74398705e-01 4.28941011e-01 -1.90599874e-01
7.79663771e-02 -3.35940778e-01 -5.51283538e-01 2.78295785e-01
3.81752789e-01 3.00750732e-01 6.09748960e-01 1.39794326e+00
-1.77484751e-01 -1.97675183e-01 -1.50422108e+00 1.09544849e+00
-8.53653904e-03 -1.15022659e+00 6.76669925e-03 -1.29928263e-02
6.85532868e-01 5.90363801e-01 -4.65455838e-02 2.99338043e-01
9.32706058e-01 -1.16101253e+00 9.52274144e-01 -1.57455467e-02
8.17207456e-01 -6.04678810e-01 6.56008244e-01 4.08361256e-01
-7.70433545e-01 -3.20013523e-01 -5.47139943e-01 2.07469195e-01
-6.66371211e-02 3.82473320e-01 -1.32570052e+00 1.94807917e-01
1.35637355e+00 6.31953418e-01 -6.74001992e-01 9.62736487e-01
-1.20866083e-01 5.66855185e-02 -5.70448160e-01 8.64403695e-02
2.42749810e-01 1.09595515e-01 1.62399277e-01 1.03984249e+00
1.59368441e-01 2.05491260e-01 8.13545734e-02 1.14397168e+00
-3.53821777e-02 -7.39943206e-01 -8.13414097e-01 1.56344801e-01
6.98296428e-01 1.18263721e+00 -1.14212632e+00 -3.75946522e-01
1.17802665e-01 7.87999511e-01 2.19899416e-01 1.28021523e-01
-6.53151691e-01 -3.34476888e-01 6.78348958e-01 8.38575438e-02
7.62554944e-01 -4.25081164e-01 -7.49104500e-01 -7.25481689e-01
-5.26582710e-02 -1.00018406e+00 -1.02934219e-01 -1.09900069e+00
-1.35062456e+00 5.07777393e-01 3.04407388e-01 -1.28465176e+00
-2.21583009e-01 -1.16188347e+00 5.64581081e-02 6.11386180e-01
-1.29275298e+00 -1.18664432e+00 -1.13088739e+00 4.52904046e-01
7.29697526e-01 1.98887303e-01 6.24905467e-01 -9.24854167e-03
-6.11390471e-02 1.81955278e-01 -2.16261744e-01 2.89483190e-01
4.39791977e-01 -1.44968879e+00 7.68922031e-01 3.92807364e-01
-3.25169981e-01 2.89986014e-01 6.68285012e-01 -6.88283324e-01
-1.56108832e+00 -1.61431313e+00 -4.63466346e-01 -9.48007464e-01
2.15851456e-01 -5.61177611e-01 -7.19092190e-01 8.99593174e-01
-3.16126555e-01 3.15102339e-01 -3.10551729e-02 -5.79481483e-01
-3.80628049e-01 1.81262538e-01 -1.74040794e+00 3.09352696e-01
1.65112185e+00 -1.23437261e-02 -5.79050660e-01 7.66665339e-01
1.03105569e+00 -9.56072330e-01 -9.13063228e-01 9.02704358e-01
6.57390773e-01 -9.28385854e-01 1.35035038e+00 -3.79945338e-01
7.85327435e-01 -4.09953266e-01 -4.93031204e-01 -1.19949031e+00
2.10469291e-01 1.75830719e-04 6.67881519e-02 6.78444564e-01
1.92344368e-01 -3.88329536e-01 1.32789159e+00 6.24938369e-01
-4.67810810e-01 -6.03182614e-01 -5.82674384e-01 -9.64751244e-01
4.09784615e-01 -5.06409943e-01 8.29905868e-01 6.60162628e-01
-5.62115073e-01 -2.92557627e-01 6.24902666e-01 3.41125429e-01
9.51662362e-01 2.89901972e-01 1.38514614e+00 -1.23551154e+00
-1.27821997e-01 -3.05578858e-01 -4.30961132e-01 -1.12151074e+00
1.89686954e-01 -7.62522459e-01 6.84326470e-01 -1.85814643e+00
-2.29379296e-01 -1.01369917e+00 5.03255546e-01 8.04801583e-01
1.38026893e-01 5.06336868e-01 2.45817788e-02 2.62061179e-01
-2.88014323e-01 3.61171067e-01 1.35722828e+00 -4.31211084e-01
-1.18649222e-01 -1.87021703e-01 -2.46367812e-01 6.48402810e-01
8.95161033e-01 -4.35358137e-01 -2.28355438e-01 -7.20100880e-01
-2.92066962e-01 -3.00519198e-01 7.99064517e-01 -1.27637601e+00
-1.75713107e-01 -1.84818789e-01 7.56157696e-01 -9.04824913e-01
6.72275662e-01 -9.83557701e-01 -4.42338660e-02 5.90519965e-01
-8.86762589e-02 9.89215076e-03 4.68247026e-01 3.45554173e-01
3.50101411e-01 -2.13231608e-01 7.08841145e-01 -7.63981521e-01
-8.21936250e-01 3.41409355e-01 3.62757370e-02 -4.44316119e-02
1.20082641e+00 -5.76715291e-01 -5.59940875e-01 4.51676063e-02
-6.07559144e-01 1.63466424e-01 9.86038685e-01 4.68909174e-01
7.76498199e-01 -1.00867438e+00 -4.71923739e-01 2.61862755e-01
2.66881108e-01 1.28804517e+00 -1.46386251e-01 1.33738397e-02
-1.22954237e+00 9.01751965e-02 -5.45073509e-01 -1.23290551e+00
-6.12306714e-01 7.93204665e-01 3.09783578e-01 5.06208777e-01
-8.38575184e-01 1.12354684e+00 1.05907537e-01 -9.21933472e-01
5.87479413e-01 -8.44556034e-01 4.71492529e-01 -5.53323567e-01
1.25649244e-01 1.45715207e-01 1.61446378e-01 -1.84602693e-01
-4.13007408e-01 4.84818280e-01 3.01346540e-01 -1.55665129e-01
1.50334656e+00 4.02610153e-01 1.13611601e-01 2.08819598e-01
1.01521921e+00 -6.28854454e-01 -2.00591326e+00 6.17688894e-02
-3.32040004e-02 -4.81786579e-01 -5.18440068e-01 -9.17720616e-01
-1.11126590e+00 8.23282242e-01 7.45150447e-01 1.02258429e-01
7.26229012e-01 4.82437611e-01 5.68594873e-01 6.99540317e-01
9.42467809e-01 -8.89035046e-01 5.48151135e-01 4.12592351e-01
1.13270557e+00 -1.57722545e+00 -1.67886555e-01 -4.89508241e-01
-3.95392388e-01 1.07584536e+00 1.17782784e+00 -6.74020171e-01
3.93121392e-01 5.58165848e-01 3.48683655e-01 -4.98601615e-01
-2.98317015e-01 -2.30988309e-01 -4.41377342e-01 1.22653615e+00
-3.08156133e-01 -7.66938478e-02 6.48579478e-01 1.57404050e-01
-6.42119408e-01 -4.37695440e-03 5.64135432e-01 1.31816459e+00
-4.94082838e-01 -7.09654689e-01 -5.70083439e-01 4.65648472e-01
4.09993418e-02 3.31099808e-01 -4.39610720e-01 1.00961637e+00
4.16669786e-01 8.14994514e-01 4.99570280e-01 -4.26848233e-01
4.80432183e-01 -2.31839746e-01 9.80211735e-01 -9.81106281e-01
-4.31664765e-01 -4.60529029e-01 -2.19069332e-01 -8.89765561e-01
-5.82101107e-01 -7.29781568e-01 -1.63053417e+00 1.89085621e-02
-3.97650711e-02 -3.74653816e-01 1.28989577e+00 6.60686970e-01
3.49960089e-01 5.68602383e-01 3.01139235e-01 -1.96163154e+00
-5.24613202e-01 -1.15727997e+00 -1.39571756e-01 4.85914409e-01
2.68828690e-01 -8.27679276e-01 -4.47539896e-01 1.99828669e-01] | [6.023920059204102, -0.9887037873268127] |
64170f2a-a359-4fe5-9f0b-3e332c0fd78b | document-summarization-with-text-segmentation | 2301.08817 | null | https://arxiv.org/abs/2301.08817v1 | https://arxiv.org/pdf/2301.08817v1.pdf | Document Summarization with Text Segmentation | In this paper, we exploit the innate document segment structure for improving the extractive summarization task. We build two text segmentation models and find the most optimal strategy to introduce their output predictions in an extractive summarization model. Experimental results on a corpus of scientific articles show that extractive summarization benefits from using a highly accurate segmentation method. In particular, most of the improvement is in documents where the most relevant information is not at the beginning thus, we conclude that segmentation helps in reducing the lead bias problem. | ['Benjamin Han', 'Lesly Miculicich'] | 2023-01-20 | null | null | null | null | ['extractive-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 5.91703176e-01 8.12677920e-01 -6.06508017e-01 -2.04497963e-01
-9.45587277e-01 -8.25717807e-01 5.94417155e-01 6.26689672e-01
-4.96876270e-01 8.81231606e-01 9.26696062e-01 -4.27971184e-01
-1.30725399e-01 -4.41441745e-01 -6.44131780e-01 -3.09811801e-01
4.89523858e-01 7.88122416e-01 2.60071337e-01 -2.29608908e-01
1.17373490e+00 2.53596365e-01 -9.52071488e-01 4.98912334e-01
1.58459663e+00 3.36697906e-01 1.69145927e-01 9.61702108e-01
-5.30220330e-01 5.84198713e-01 -1.12164664e+00 -4.11621213e-01
7.66903460e-02 -9.11959946e-01 -1.33039773e+00 4.69970005e-03
5.12310326e-01 -9.36755016e-02 3.41115817e-02 1.00630271e+00
3.68144095e-01 8.55101496e-02 1.03862977e+00 -3.83208632e-01
-1.74758807e-01 1.45657289e+00 -6.13592565e-01 5.46870768e-01
3.48965466e-01 -2.35461205e-01 1.29128337e+00 -4.62244034e-01
1.03021348e+00 1.04895961e+00 2.74401128e-01 5.55193603e-01
-1.02546954e+00 1.06457002e-01 2.14245737e-01 -6.45904616e-02
-7.82482982e-01 -4.97724175e-01 9.05609012e-01 -3.47515076e-01
1.05226910e+00 6.88973308e-01 7.25483000e-01 6.90756083e-01
4.92571831e-01 1.34100902e+00 6.94550753e-01 -7.66057730e-01
2.00818613e-01 9.97356027e-02 9.45996225e-01 3.80417228e-01
6.38252854e-01 -6.75955832e-01 -3.98365021e-01 5.63620264e-03
1.37058049e-01 -4.62958783e-01 -3.56912076e-01 1.72227934e-01
-9.56908584e-01 9.20987487e-01 8.66753533e-02 6.23559535e-01
-4.49848503e-01 -1.94164172e-01 5.71313918e-01 7.16427267e-02
9.25984740e-01 1.21070790e+00 -4.85817015e-01 -2.15064704e-01
-1.76493537e+00 3.64782006e-01 1.07299805e+00 8.16122770e-01
4.67285961e-01 -1.11351252e-01 -5.51350355e-01 7.07483590e-01
-9.36987028e-02 2.62440830e-01 6.63858414e-01 -1.14369059e+00
7.30359435e-01 6.95557535e-01 -7.26697892e-02 -7.05869257e-01
-4.69156921e-01 -6.25214219e-01 -5.07255793e-01 -5.11574984e-01
1.23492822e-01 -3.74719828e-01 -9.04367745e-01 1.08404934e+00
-1.10068068e-01 -6.32581949e-01 1.51254714e-01 2.30946913e-01
8.43482375e-01 9.04367447e-01 1.80801284e-02 -7.01933503e-01
1.19062209e+00 -1.11332405e+00 -1.02225220e+00 -4.22708809e-01
7.73263216e-01 -7.60000646e-01 5.34666479e-01 3.58041257e-01
-1.26635695e+00 -4.43963856e-01 -1.19168341e+00 -2.87552655e-01
-1.71173796e-01 2.12275103e-01 3.81562233e-01 3.33257735e-01
-1.00920653e+00 9.75394011e-01 -7.21869707e-01 -4.12799060e-01
3.17156315e-01 3.40069711e-01 1.29187584e-01 2.79935777e-01
-9.19258416e-01 9.26578104e-01 9.57471907e-01 -2.64843673e-01
-1.57349020e-01 -5.42901874e-01 -6.48240983e-01 2.61984646e-01
3.83003831e-01 -9.21513438e-01 1.50264978e+00 -1.01324832e+00
-1.35895300e+00 7.36818433e-01 -5.48526585e-01 -7.91309953e-01
4.56076413e-01 -4.83472943e-01 3.24360371e-01 5.97393990e-01
3.11209351e-01 7.93791354e-01 6.26296759e-01 -1.39031792e+00
-7.97059417e-01 -5.34239471e-01 -3.90823960e-01 4.15153295e-01
-3.33264798e-01 -4.37783785e-02 -5.34312129e-01 -7.30879068e-01
1.00356065e-01 -7.28458107e-01 -3.06155980e-01 -1.11804426e+00
-1.07063282e+00 -6.00708961e-01 4.75109786e-01 -1.09774566e+00
1.64012647e+00 -1.58107197e+00 4.25607234e-01 2.26059839e-01
2.93216825e-01 2.07611933e-01 1.43525541e-01 6.08081639e-01
1.77352890e-01 7.88656950e-01 -3.06721479e-01 -2.74793774e-01
-1.47231638e-01 2.13262111e-01 -5.71276426e-01 -1.16261363e-01
-1.66641995e-02 1.13037229e+00 -7.63444602e-01 -9.96171832e-01
-1.90568298e-01 -1.77797079e-01 -5.43518484e-01 7.74949491e-02
-4.56265748e-01 2.23874792e-01 -8.23355913e-01 2.44540587e-01
3.35011542e-01 6.81024343e-02 5.88020235e-02 1.54124722e-01
-4.05688941e-01 5.85748732e-01 -5.31454682e-01 1.66379952e+00
1.40985280e-01 1.05656767e+00 -1.98517546e-01 -1.09367502e+00
7.34693825e-01 1.90963168e-02 3.67693394e-01 -4.60298866e-01
2.32598230e-01 1.91616923e-01 1.04236476e-01 -5.13713002e-01
1.25611353e+00 5.05987629e-02 -1.92618355e-01 4.54457790e-01
1.28425583e-01 -6.00925326e-01 9.48986471e-01 8.02748680e-01
7.30594575e-01 -7.21042976e-02 5.44227719e-01 -6.97934985e-01
2.88567811e-01 5.47426164e-01 3.78555030e-01 1.15197754e+00
7.22561702e-02 7.24579155e-01 1.01396298e+00 2.64234722e-01
-1.09784138e+00 -5.72563112e-01 -5.12705818e-02 1.05139589e+00
-3.01248074e-01 -6.73117697e-01 -1.46767986e+00 -9.16935921e-01
-3.02963436e-01 1.41893637e+00 -6.53555810e-01 -5.20238504e-02
-9.35247064e-01 -6.05693579e-01 4.86792356e-01 4.51430470e-01
1.31289899e-01 -9.83575821e-01 -6.57512546e-01 2.84201205e-01
-4.85830486e-01 -7.41305649e-01 -5.15602171e-01 4.85907406e-01
-1.43605721e+00 -7.27323592e-01 -8.18024576e-01 -7.80885577e-01
5.77658117e-01 1.16236858e-01 1.12008810e+00 6.70855716e-02
1.77568659e-01 2.56156415e-01 -5.97404242e-01 -8.94071043e-01
-8.64346623e-01 1.02853000e+00 -4.41098213e-01 -7.02549279e-01
2.61700034e-01 9.36777238e-03 -3.85056198e-01 -4.05682057e-01
-9.28921103e-01 1.67133719e-01 7.83104599e-01 5.94708502e-01
2.40662098e-01 -1.56444684e-01 7.03620195e-01 -1.42662668e+00
1.19708550e+00 -1.37241468e-01 2.93274838e-02 3.48835677e-01
-9.22521174e-01 4.99723613e-01 5.91854632e-01 4.59241420e-02
-1.21823108e+00 -2.98153996e-01 -3.42325002e-01 5.62586904e-01
-1.72195107e-01 7.79780746e-01 6.58393577e-02 5.51581979e-01
8.52689147e-01 3.30221981e-01 -1.24099046e-01 -4.15227234e-01
4.24080253e-01 6.22206151e-01 2.74749309e-01 -5.38932264e-01
1.83059081e-01 3.27892043e-02 -2.36412898e-01 -1.12390316e+00
-1.09254670e+00 -7.06583738e-01 -8.83017302e-01 -1.75642118e-01
7.08332181e-01 -3.14343870e-01 -5.52194864e-02 1.28141433e-01
-1.39008737e+00 -1.30335435e-01 -8.20701003e-01 1.64365262e-01
-5.70295990e-01 8.63128781e-01 -6.68744743e-01 -5.49121439e-01
-9.05600905e-01 -9.15777266e-01 1.06157315e+00 6.24250710e-01
-7.40141928e-01 -9.57067788e-01 2.92475671e-01 3.79359394e-01
2.09538755e-03 5.31267188e-03 1.12914538e+00 -1.27773654e+00
-1.50741130e-01 -1.70732126e-01 1.86902154e-02 2.48480558e-01
-3.85016873e-02 4.80652303e-01 -5.63771307e-01 1.78356662e-01
1.53863221e-01 8.65471363e-02 1.43349457e+00 9.23835874e-01
8.61859262e-01 -6.73417628e-01 -5.44229925e-01 1.54945865e-01
1.08026373e+00 1.95258826e-01 6.77767336e-01 3.16030562e-01
6.99527323e-01 9.58273828e-01 5.39727092e-01 7.10996538e-02
1.35949075e-01 2.65002787e-01 -1.43756211e-01 7.15402439e-02
-1.38550565e-01 -2.59218723e-01 2.40237489e-01 1.36460698e+00
-1.03131860e-01 -6.76800072e-01 -8.82061422e-01 6.39244318e-01
-2.00196409e+00 -1.00705218e+00 -4.84069943e-01 1.55935431e+00
1.20386314e+00 4.23114002e-01 1.72879115e-01 1.95469633e-01
4.38445836e-01 2.25086033e-01 -2.70056129e-01 -8.39127898e-01
-2.76537031e-01 2.02059925e-01 4.73503232e-01 6.96711183e-01
-8.61958444e-01 1.35426795e+00 7.46514606e+00 9.60226953e-01
-7.17013001e-01 -3.48611683e-01 7.23802686e-01 8.93633440e-03
-5.49415886e-01 1.04885593e-01 -9.61250246e-01 4.07122761e-01
1.02392411e+00 -4.99596000e-01 -2.55448937e-01 6.54153109e-01
4.53309000e-01 -5.80152273e-01 -9.93267775e-01 2.25283548e-01
3.86163265e-01 -1.48248267e+00 7.16465056e-01 -1.49488345e-01
8.68349016e-01 -3.39777738e-01 -3.69952410e-01 1.31561011e-01
1.02986202e-01 -6.74766302e-01 7.77106583e-01 5.66237926e-01
1.43304348e-01 -8.94485891e-01 8.33055377e-01 6.23723388e-01
-3.59237969e-01 1.84915781e-01 -4.67712462e-01 8.91258866e-02
1.54529065e-01 6.43710375e-01 -1.11745214e+00 8.35247993e-01
-1.79839712e-02 6.75774992e-01 -9.50467348e-01 1.03808892e+00
-6.36408269e-01 1.05835807e+00 1.51135325e-02 -3.10720742e-01
4.21441138e-01 -3.78390133e-01 9.22700167e-01 1.81944609e+00
5.53610884e-02 2.79375553e-01 -1.12854028e-02 5.22256136e-01
-2.78935939e-01 5.50085962e-01 -5.73743105e-01 -2.04199806e-01
-5.21358661e-03 1.03224730e+00 -1.26711977e+00 -6.32811010e-01
3.20117801e-01 1.02142537e+00 2.10552365e-01 3.60573471e-01
-2.17820466e-01 -5.61153412e-01 -3.22366744e-01 -4.22506072e-02
3.33434373e-01 -1.57470599e-01 -9.73745763e-01 -1.00866652e+00
-2.36071438e-01 -8.86215210e-01 4.13305581e-01 -5.51863492e-01
-6.08110785e-01 4.20692891e-01 2.47218937e-01 -4.42339689e-01
-4.93998319e-01 -2.63519168e-01 -8.52602720e-01 5.12594044e-01
-1.21732569e+00 -7.66167283e-01 3.61462772e-01 -1.53345719e-01
1.17887807e+00 1.54383093e-01 4.41138864e-01 -3.57474178e-01
-6.13489807e-01 3.18412095e-01 3.00225079e-01 4.36018743e-02
5.82079768e-01 -1.70479500e+00 3.24847639e-01 1.16208076e+00
2.53368109e-01 7.71160662e-01 1.25582004e+00 -8.89880717e-01
-8.87049139e-01 -6.53375089e-01 1.20106745e+00 -5.75891435e-01
3.34314346e-01 3.37332860e-02 -8.55236351e-01 5.94410777e-01
9.41599727e-01 -1.37028885e+00 6.92225039e-01 2.25151286e-01
2.37135783e-01 3.00869495e-01 -7.26255715e-01 6.14953458e-01
6.40130281e-01 -2.37163398e-02 -1.38660932e+00 3.28805625e-01
8.06175172e-01 -3.14725369e-01 -5.04648924e-01 1.26256943e-01
2.01504305e-01 -5.30700982e-01 4.62587416e-01 -7.55574286e-01
9.79680061e-01 1.86194822e-01 4.54109192e-01 -1.55029416e+00
-2.07552224e-01 -7.61856019e-01 -2.52472281e-01 1.35245156e+00
8.52852702e-01 -2.73360193e-01 8.61513734e-01 4.80593562e-01
-4.73872215e-01 -5.59249580e-01 -5.97738385e-01 -3.80018055e-01
5.70522666e-01 -3.00176144e-02 -1.42862305e-01 5.82140326e-01
3.77200454e-01 1.01706243e+00 4.07317430e-02 -6.92626655e-01
5.37423790e-01 2.85613686e-01 5.70052683e-01 -1.44132555e+00
1.83244839e-01 -9.15226400e-01 2.11829290e-01 -1.50134277e+00
3.67975831e-01 -9.23881292e-01 3.11622947e-01 -2.19581461e+00
5.80272138e-01 2.51879871e-01 1.44590825e-01 2.51598001e-01
-6.66176736e-01 -2.95481384e-01 2.63980836e-01 3.55615437e-01
-7.98227906e-01 3.33613664e-01 1.21678948e+00 -4.64207828e-01
-5.71231127e-01 2.10718453e-01 -1.26939189e+00 7.96602011e-01
1.00965333e+00 -6.87585115e-01 -2.46539041e-01 -2.53801823e-01
2.70451814e-01 8.87947828e-02 -4.47590202e-01 -5.94394088e-01
4.16589588e-01 -1.15307771e-01 2.89387792e-01 -1.01291776e+00
-2.87944615e-01 -1.47863820e-01 -5.57440460e-01 5.40050983e-01
-8.47060442e-01 -1.41370326e-01 3.18432450e-01 3.21462929e-01
-2.05449909e-01 -1.01046836e+00 5.92805505e-01 -2.40696937e-01
-3.16626847e-01 -3.72363836e-01 -8.19693804e-01 5.14438450e-01
5.99102795e-01 -2.36913413e-01 -3.64183962e-01 -2.73140430e-01
-4.99722779e-01 2.30921894e-01 3.87019932e-01 2.07814470e-01
3.13942105e-01 -4.52016085e-01 -9.59723890e-01 -3.57745141e-01
-3.92150730e-01 2.40024291e-02 -4.63448539e-02 6.72491729e-01
-7.12825119e-01 8.96139741e-01 -3.15924361e-02 -4.09425527e-01
-1.32433116e+00 3.05933714e-01 -2.15468377e-01 -7.33374953e-01
-4.51873839e-01 6.20373130e-01 -2.15948179e-01 1.16899483e-01
9.17777941e-02 -4.49943751e-01 -7.19918311e-01 6.11695588e-01
3.41151237e-01 6.58269882e-01 1.50972605e-01 -4.16547984e-01
4.85461392e-02 3.48553926e-01 -6.52195036e-01 -3.15370440e-01
1.29707837e+00 -1.43883869e-01 -4.18559730e-01 4.94650215e-01
8.00195992e-01 4.44916576e-01 -7.60378003e-01 1.26829132e-01
5.41752279e-01 1.20088056e-01 1.18607506e-01 -8.26839030e-01
-5.24127126e-01 7.95679808e-01 -3.10120761e-01 7.51363873e-01
7.89318442e-01 1.14711031e-01 6.55832946e-01 6.06371462e-01
-2.28590026e-01 -1.62791562e+00 -1.63180679e-01 8.22327077e-01
8.51716697e-01 -7.72064924e-01 6.91216469e-01 -4.16854799e-01
-8.65580559e-01 1.32326508e+00 3.55063319e-01 -9.26195085e-02
9.86331627e-02 -3.37414891e-02 -2.06118807e-01 -3.95925194e-01
-4.49681342e-01 -1.09958760e-01 4.76586282e-01 1.97593927e-01
5.79263270e-01 -4.58559543e-02 -1.11371970e+00 6.94098234e-01
-5.32404900e-01 -5.01357973e-01 8.48638296e-01 8.51502478e-01
-1.00328922e+00 -1.01826537e+00 -1.98293611e-01 8.22283506e-01
-7.55656183e-01 -1.93121657e-01 -1.24892616e+00 7.45905221e-01
-3.79382491e-01 1.09200597e+00 -3.12687978e-02 1.84308827e-01
2.59104937e-01 3.42096120e-01 4.75108206e-01 -9.33144033e-01
-1.04790318e+00 3.90368134e-01 2.56478429e-01 -4.03606072e-02
-4.64886159e-01 -8.26498926e-01 -1.47906113e+00 -5.04589900e-02
-4.27715749e-01 8.17369878e-01 6.42206788e-01 1.00357234e+00
4.51054513e-01 8.99716437e-01 3.06187451e-01 -6.97994351e-01
-6.90388143e-01 -1.30879915e+00 -2.91368276e-01 2.16212854e-01
1.60952166e-01 1.17438428e-01 -3.09788704e-01 2.17152566e-01] | [12.544794082641602, 9.535594940185547] |
a6fbc3f6-7f57-4eb7-b2b3-045f7acc565b | g2pm-a-neural-grapheme-to-phoneme-conversion | 2004.03136 | null | https://arxiv.org/abs/2004.03136v5 | https://arxiv.org/pdf/2004.03136v5.pdf | g2pM: A Neural Grapheme-to-Phoneme Conversion Package for Mandarin Chinese Based on a New Open Benchmark Dataset | Conversion of Chinese graphemes to phonemes (G2P) is an essential component in Mandarin Chinese Text-To-Speech (TTS) systems. One of the biggest challenges in Chinese G2P conversion is how to disambiguate the pronunciation of polyphones - characters having multiple pronunciations. Although many academic efforts have been made to address it, there has been no open dataset that can serve as a standard benchmark for fair comparison to date. In addition, most of the reported systems are hard to employ for researchers or practitioners who want to convert Chinese text into pinyin at their convenience. Motivated by these, in this work, we introduce a new benchmark dataset that consists of 99,000+ sentences for Chinese polyphone disambiguation. We train a simple neural network model on it, and find that it outperforms other preexisting G2P systems. Finally, we package our project and share it on PyPi. | ['Kyubyong Park', 'Seanie Lee'] | 2020-04-07 | null | null | null | null | ['polyphone-disambiguation'] | ['natural-language-processing'] | [ 1.60052344e-01 -2.05805704e-01 -3.60674225e-02 -7.84819052e-02
-1.08073461e+00 -6.72228277e-01 3.52763325e-01 -3.35059315e-01
-4.04680520e-01 8.73628438e-01 2.73230076e-01 -7.86390483e-01
7.02178717e-01 -4.96828526e-01 -5.00618517e-01 -5.08166015e-01
4.63839054e-01 3.61629128e-01 2.27318361e-01 -3.29363078e-01
1.30406380e-01 1.65714785e-01 -1.05969787e+00 1.66009441e-01
1.01819932e+00 6.53314054e-01 5.90740919e-01 8.72197390e-01
-1.75555870e-01 4.25526023e-01 -9.53930140e-01 -5.38188517e-01
2.28146195e-01 -6.49435699e-01 -8.87512982e-01 -5.13741910e-01
1.87204242e-01 -5.65942489e-02 -2.16953710e-01 1.45795476e+00
9.07727718e-01 -2.48152371e-02 7.16678441e-01 -7.38411725e-01
-1.13136518e+00 1.17823124e+00 -1.92347188e-02 3.34540933e-01
2.67713159e-01 -1.80889651e-01 7.84056008e-01 -8.10146272e-01
2.53788382e-01 1.11825943e+00 4.93735164e-01 8.13051164e-01
-5.80583274e-01 -8.68467987e-01 6.81066280e-03 1.42520770e-01
-1.51371288e+00 -6.91885471e-01 5.08478105e-01 -4.65882802e-03
1.23927939e+00 5.09456575e-01 3.68712276e-01 1.13286245e+00
7.40638077e-02 8.54974747e-01 8.91390085e-01 -8.90391052e-01
-6.54499307e-02 1.89416215e-01 -1.47152906e-02 1.64084673e-01
-1.75624210e-02 -1.59037456e-01 -2.79410809e-01 5.79574049e-01
5.69349945e-01 -5.54352582e-01 -3.61966044e-01 6.67755604e-01
-1.11358106e+00 6.41557395e-01 -1.14417456e-01 6.26572490e-01
1.01591937e-01 1.44699141e-01 2.01113865e-01 3.90704006e-01
2.85002619e-01 5.29217422e-01 -6.02828920e-01 -7.87299812e-01
-6.64180815e-01 -9.85004604e-02 9.19438124e-01 1.57101262e+00
3.15130949e-01 4.18627769e-01 -1.38159752e-01 9.95297372e-01
1.40447579e-02 7.56546438e-01 9.24226105e-01 -6.01783931e-01
6.77573681e-01 -6.26390204e-02 -1.73934743e-01 -6.55861080e-01
1.50798097e-01 -6.09324463e-02 -7.80902207e-01 -3.90444279e-01
4.12228197e-01 -6.65723860e-01 -6.65715635e-01 1.45508289e+00
-2.05283955e-01 1.15964450e-02 2.61275440e-01 5.61627924e-01
9.14252460e-01 1.11886597e+00 -1.54529765e-01 -1.79110363e-01
1.34754658e+00 -1.21414733e+00 -1.03481948e+00 -3.22470516e-01
4.01344210e-01 -1.29903519e+00 1.19288325e+00 3.45410019e-01
-1.16690171e+00 -6.05073214e-01 -1.00896621e+00 -1.14109851e-01
-7.78768599e-01 4.10181999e-01 3.77683669e-01 1.15408671e+00
-1.12086618e+00 6.25302613e-01 -7.01457560e-01 -3.69810551e-01
-2.01690912e-01 2.16895357e-01 -4.49371897e-02 3.47386599e-01
-1.62341630e+00 8.65863860e-01 7.11488247e-01 1.24789976e-01
-4.34292793e-01 -4.20804024e-01 -6.42861426e-01 1.28310993e-01
1.98688582e-01 -1.13357149e-01 1.59424269e+00 -6.41232312e-01
-2.00682855e+00 5.97063184e-01 -1.67159513e-01 -2.48040333e-01
3.09600264e-01 -4.08782631e-01 -1.10339272e+00 -1.16683565e-01
8.16651136e-02 7.67698765e-01 8.50360096e-01 -5.91266096e-01
-1.01296723e+00 1.51322767e-01 -4.16857809e-01 5.04310131e-01
-4.77412134e-01 6.35170281e-01 -1.05522752e+00 -9.66917336e-01
-2.31982872e-01 -9.57725585e-01 6.31953776e-02 -8.98706675e-01
-8.30274582e-01 -3.80114436e-01 8.54338169e-01 -1.00347769e+00
1.59041178e+00 -2.13885164e+00 -3.27690452e-01 -2.33224794e-01
-2.88540453e-01 7.73668587e-01 5.96480705e-02 4.44098562e-01
9.99209359e-02 4.99696881e-01 -5.15313894e-02 -3.48301649e-01
1.51633084e-01 2.83737723e-02 -4.24639851e-01 -5.61960153e-02
3.48492235e-01 9.70527589e-01 -9.23666656e-01 -4.64079529e-01
2.60979295e-01 3.68998170e-01 -3.04446489e-01 9.16714445e-02
-7.48889521e-02 4.39856291e-01 -2.33074859e-01 6.08850658e-01
6.66886032e-01 1.13215119e-01 2.04828635e-01 2.47054949e-01
-3.84453863e-01 8.94958317e-01 -8.82689238e-01 1.30541527e+00
-4.03521687e-01 9.28685427e-01 -1.96823165e-01 -7.67413557e-01
8.87473404e-01 5.79371750e-01 -1.40025020e-01 -6.36422455e-01
2.99282998e-01 4.87668782e-01 2.03564316e-02 -1.74047887e-01
9.81505454e-01 8.97956118e-02 -3.68227005e-01 1.62647828e-01
7.16977641e-02 -5.31277239e-01 3.69558334e-01 -1.40283713e-02
6.92381382e-01 -2.22191408e-01 3.44146103e-01 -2.78791547e-01
4.77449507e-01 -2.01335311e-01 4.44243520e-01 8.28207195e-01
-1.88875839e-01 1.09824896e+00 3.00994873e-01 4.47409041e-02
-8.92628372e-01 -1.06583452e+00 -1.08725450e-03 9.56587315e-01
-1.07241891e-01 -4.76610154e-01 -1.20985556e+00 -4.53060180e-01
-6.54859245e-01 8.83342922e-01 6.31449744e-02 4.21249717e-02
-1.05715764e+00 -6.22553885e-01 1.09236598e+00 5.31439602e-01
5.43111324e-01 -1.38985348e+00 8.60877112e-02 4.13241059e-01
-4.99625832e-01 -1.47837543e+00 -1.16726112e+00 2.90667176e-01
-4.36248392e-01 -6.96753263e-01 -9.25834715e-01 -1.50768507e+00
3.30267757e-01 2.73077697e-01 9.50270474e-01 -8.28928128e-02
1.17079981e-01 -7.11574480e-02 -7.76195526e-01 -6.78241730e-01
-9.80260670e-01 5.71287274e-01 9.72311124e-02 -3.69795829e-01
8.28772247e-01 -9.91399512e-02 -9.74725857e-02 2.44737029e-01
-7.00975180e-01 -2.57428978e-02 4.95660812e-01 4.72140819e-01
5.14356554e-01 2.51609951e-01 5.79476178e-01 -9.32953358e-01
8.92098129e-01 -2.78247625e-01 -7.96013892e-01 2.92832881e-01
-3.83000612e-01 -1.78015605e-01 1.09151495e+00 -4.44649845e-01
-1.05211115e+00 -2.48585809e-02 -6.95691824e-01 4.93160188e-02
-2.67792344e-01 5.58555305e-01 -4.97504801e-01 6.61418438e-02
2.71084219e-01 4.84158874e-01 -4.20553774e-01 -7.15511382e-01
5.19364297e-01 1.35660875e+00 8.64622295e-01 -4.66724157e-01
5.41458428e-01 -2.34249756e-01 -7.89196014e-01 -1.36548221e+00
-5.17364681e-01 -3.56806636e-01 -5.04617870e-01 1.65009230e-01
1.05627394e+00 -8.80373776e-01 -4.31593776e-01 8.91695201e-01
-1.43105984e+00 -3.08293760e-01 8.88044611e-02 5.95713675e-01
-3.24055105e-01 6.42755389e-01 -8.44365895e-01 -4.19356197e-01
-3.51956457e-01 -1.19551599e+00 6.66150331e-01 5.56110680e-01
-2.01672196e-01 -8.75244856e-01 -1.18942431e-03 2.48448804e-01
7.46448338e-01 -4.84680176e-01 7.19610512e-01 -6.74996138e-01
-4.41612393e-01 -1.60423517e-01 -1.12881035e-01 7.01803446e-01
4.95074779e-01 2.23861709e-01 -1.07718205e+00 -1.06014125e-01
2.98170596e-02 2.84809563e-02 6.11256182e-01 4.18961555e-01
1.23049688e+00 -3.85290861e-01 -1.02621876e-01 7.23448932e-01
9.51500237e-01 8.37948143e-01 8.72825325e-01 4.62293960e-02
7.01404572e-01 1.28068671e-01 3.00915599e-01 1.30977407e-01
4.61089849e-01 4.90144372e-01 -8.64045992e-02 8.34990442e-02
-4.07665402e-01 -4.41391975e-01 6.81289017e-01 1.93867290e+00
-2.92480886e-02 -5.96651375e-01 -8.21284115e-01 5.55964530e-01
-1.20085299e+00 -5.63219011e-01 -1.16751097e-01 2.07340050e+00
1.19569588e+00 1.64418727e-01 -2.55577803e-01 8.82817879e-02
1.21251845e+00 6.50358945e-02 -8.02289695e-03 -6.81016147e-01
-3.85185331e-01 4.10566837e-01 4.49979216e-01 5.87311089e-01
-1.26596200e+00 1.71119487e+00 6.84814835e+00 1.27885044e+00
-1.42609560e+00 -1.07582949e-01 6.25666440e-01 3.78595918e-01
-1.36610642e-01 -7.15416903e-03 -1.31305420e+00 7.02710629e-01
1.17504704e+00 -1.83655858e-01 6.00703359e-01 7.37494171e-01
-1.49234682e-02 2.44037256e-01 -1.02644825e+00 1.08130395e+00
1.19226508e-01 -1.16625702e+00 1.10179624e-02 -4.43886846e-01
8.57017040e-01 1.75095603e-01 1.53119832e-01 5.34765303e-01
4.12936896e-01 -1.03858912e+00 6.98062241e-01 -2.43955944e-02
1.05329669e+00 -5.88154137e-01 5.55953205e-01 2.03082845e-01
-1.27334929e+00 3.68295342e-01 -5.78902543e-01 -5.49687557e-02
3.08915585e-01 2.69376397e-01 -1.02092946e+00 3.82139325e-01
6.16931677e-01 4.26450640e-01 -3.82542402e-01 9.41951573e-01
-5.89826167e-01 9.75980222e-01 -1.66770175e-01 -5.97604096e-01
2.72321314e-01 -3.08560163e-01 3.66502970e-01 1.67266035e+00
7.00668335e-01 3.06149740e-02 -3.79858822e-01 6.96076393e-01
-2.40226582e-01 2.16806874e-01 -4.16135073e-01 -6.70643866e-01
9.37129200e-01 9.60175216e-01 -9.21160400e-01 -4.21208233e-01
-6.44939840e-01 1.26128972e+00 2.44021252e-01 3.35177988e-01
-7.32719123e-01 -1.04997122e+00 5.62496006e-01 -4.77380365e-01
5.12487113e-01 -2.93587983e-01 -8.93407241e-02 -1.25656116e+00
-6.93385676e-02 -1.18847382e+00 7.09090903e-02 -6.17497146e-01
-1.31363010e+00 8.29053462e-01 -3.15525562e-01 -1.16907942e+00
-4.31804746e-01 -7.53636360e-01 -7.85352170e-01 1.04313815e+00
-1.63675058e+00 -7.02151000e-01 1.63090914e-01 4.12751824e-01
6.71702445e-01 -3.08178931e-01 9.44474995e-01 5.84414303e-01
-7.39939153e-01 9.65379179e-01 3.34976584e-01 6.52333856e-01
7.75703371e-01 -1.29568470e+00 1.19366455e+00 1.34282053e+00
4.12872046e-01 6.59435093e-01 3.89606357e-01 -6.35386407e-01
-1.27775061e+00 -1.16555762e+00 1.42476749e+00 -5.41807294e-01
6.13705456e-01 -6.03087127e-01 -6.88954830e-01 7.10951149e-01
5.41387558e-01 -3.51102024e-01 4.40154642e-01 -2.21320271e-01
1.83369696e-01 -5.96020883e-03 -4.67573047e-01 9.13655937e-01
7.45364189e-01 -5.91253161e-01 -6.49218857e-01 1.18725359e-01
1.18360853e+00 -7.08325386e-01 -5.54694414e-01 -1.15628399e-01
2.24840954e-01 -6.42251790e-01 6.51180327e-01 -4.24709678e-01
1.92227617e-01 -2.64308184e-01 -3.12492698e-01 -1.59239912e+00
-2.21917272e-01 -1.15365708e+00 3.31252933e-01 1.52578008e+00
7.01226473e-01 -6.00029945e-01 6.06078923e-01 3.03218484e-01
-6.19909942e-01 -1.45904064e-01 -7.78923690e-01 -1.04761255e+00
5.66371500e-01 -6.77337706e-01 9.04149532e-01 9.85120237e-01
1.82519794e-01 4.13446665e-01 -6.40365303e-01 6.81902245e-02
4.84683812e-02 -1.60022229e-01 4.30324078e-01 -6.41300023e-01
-3.12921911e-01 -7.24740207e-01 1.13942720e-01 -1.43914378e+00
7.30089983e-03 -8.47647965e-01 3.58492702e-01 -1.11713779e+00
-3.05302382e-01 -4.19100761e-01 -8.33398476e-02 3.73427629e-01
-4.66664791e-01 2.69713908e-01 3.98086697e-01 -1.52140826e-01
-3.30784202e-01 5.15493095e-01 1.16296935e+00 -4.00988460e-02
-3.23802292e-01 2.19581604e-01 -9.67822373e-01 5.26678801e-01
1.07126439e+00 -4.95658070e-01 1.42312543e-02 -8.38126421e-01
-1.72781534e-02 8.57819691e-02 -4.59903002e-01 -9.85201061e-01
3.41688365e-01 -1.01924501e-01 3.00949216e-02 -8.70526254e-01
1.73628256e-01 -5.05834043e-01 -3.64479981e-02 2.54091948e-01
-1.09603286e-01 3.27644974e-01 9.62661877e-02 1.33057535e-02
-4.96396095e-01 -4.79675919e-01 6.39284670e-01 -1.82135776e-01
-8.56583834e-01 2.62312442e-01 -8.34942162e-01 5.44844627e-01
6.53147697e-01 -9.86168385e-02 -3.94970030e-01 -4.95335013e-01
-1.68801859e-01 -2.72313505e-02 1.68540433e-01 7.12636650e-01
3.66582304e-01 -1.33107364e+00 -7.08124220e-01 3.88155550e-01
-1.09924532e-01 -1.73217043e-01 -1.15289956e-01 4.03432220e-01
-9.79282796e-01 9.26989317e-01 -6.80858716e-02 -1.33262411e-01
-1.03973424e+00 3.94074649e-01 1.05166532e-01 1.14732563e-01
-3.36662054e-01 9.71859276e-01 -1.55477807e-01 -6.77015483e-01
3.84761125e-01 -6.01910293e-01 -2.99845695e-01 -2.67307371e-01
7.31276453e-01 3.23005140e-01 3.05341989e-01 -6.31884336e-01
-2.47808889e-01 3.65261942e-01 -7.62730166e-02 -9.46922600e-02
8.97688448e-01 -3.03517580e-01 -6.13544583e-02 3.74651164e-01
1.28886116e+00 5.00029981e-01 -8.51630926e-01 -1.04114801e-01
1.27621805e-02 -1.88507721e-01 -3.41900855e-01 -5.41352093e-01
-1.02403200e+00 1.11744893e+00 1.80302620e-01 3.69021773e-01
8.91344011e-01 -2.30722323e-01 1.20062923e+00 6.85810566e-01
7.92602971e-02 -1.68328464e+00 -2.25362584e-01 1.24227989e+00
5.23802936e-01 -1.29331267e+00 -7.23439515e-01 -5.04783094e-01
-7.77561009e-01 1.23400319e+00 4.89777565e-01 5.50865829e-02
6.95837021e-01 4.31776136e-01 6.10007942e-01 4.46390539e-01
-4.09257889e-01 -5.06400056e-02 6.07832111e-02 8.98901284e-01
7.42170811e-01 3.69574696e-01 -3.84554714e-01 8.56177866e-01
-9.49913561e-01 -3.14023167e-01 8.35479677e-01 5.49963534e-01
-3.51633042e-01 -1.40737498e+00 -2.92249560e-01 2.47235849e-01
-8.76521885e-01 -7.91221440e-01 -3.86835814e-01 7.21623719e-01
-1.27453685e-01 1.15978467e+00 1.52567066e-02 -4.31572765e-01
-5.54689281e-02 4.87694032e-02 2.12351084e-01 -6.96977139e-01
-4.99143034e-01 3.36714178e-01 2.22819671e-01 -9.56708416e-02
4.98383045e-02 -5.69744229e-01 -1.13323200e+00 -3.65390390e-01
-4.90497857e-01 3.84767443e-01 7.56847858e-01 8.71987283e-01
4.54246886e-02 4.19402003e-01 5.24379492e-01 -5.10254443e-01
-4.90191936e-01 -1.00009668e+00 -7.21548319e-01 1.31608859e-01
5.72757721e-02 1.50018319e-01 -2.02690527e-01 1.41901225e-01] | [14.392314910888672, 6.981148719787598] |
6d33d753-1349-462b-acf6-7fb15223143e | a-self-paced-multiple-instance-learning | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Zhang_A_Self-Paced_Multiple-Instance_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Zhang_A_Self-Paced_Multiple-Instance_ICCV_2015_paper.pdf | A Self-Paced Multiple-Instance Learning Framework for Co-Saliency Detection | As an interesting and emerging topic, co-saliency detection aims at simultaneously extracting common salient objects in a group of images. Traditional co-saliency detection approaches rely heavily on human knowledge for designing hand-crafted metrics to explore the intrinsic patterns underlying co-salient objects. Such strategies, however, always suffer from poor generalization capability to flexibly adapt various scenarios in real applications, especially due to their lack of insightful understanding of the biological mechanisms of human visual co-attention. To alleviate this problem, we propose a novel framework for this task, by naturally reformulating it as a multiple-instance learning (MIL) problem and further integrating it into a self-paced learning (SPL) regime. The proposed framework on one hand is capable of fitting insightful metric measurements and discovering common patterns under co-salient regions in a self-learning way by MIL, and on the other hand tends to promise the learning reliability and stability by simulating the human learning process through SPL. Experiments on benchmark datasets have demonstrated the effectiveness of the proposed framework as compared with the state-of-the-arts. | ['Junwei Han', 'Chao Li', 'Qian Zhao', 'Dingwen Zhang', 'Lu Jiang', 'Deyu Meng'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['co-saliency-detection'] | ['computer-vision'] | [ 3.48822296e-01 -4.62759882e-02 -2.11657271e-01 -9.60281193e-02
-5.34224391e-01 -1.73217237e-01 6.33256972e-01 3.38325649e-01
-2.68468916e-01 6.80374920e-01 -6.43131360e-02 3.16419043e-02
-4.43141580e-01 -3.40155333e-01 -7.21245944e-01 -8.64066362e-01
1.35606313e-02 2.91185025e-02 4.14959937e-01 -1.55247405e-01
4.39698160e-01 1.05947003e-01 -1.94466686e+00 -1.30852625e-01
1.41301715e+00 9.61755574e-01 6.70522928e-01 1.64025724e-01
-3.67358290e-02 6.60182178e-01 -7.41900653e-02 -1.18025713e-01
2.21589342e-01 -5.68079054e-01 -5.50067484e-01 2.19536737e-01
3.10224980e-01 3.90452296e-01 8.58096704e-02 1.19364071e+00
3.92598271e-01 4.10279259e-02 4.51716363e-01 -1.28108358e+00
-5.55189431e-01 2.46893987e-01 -8.45777452e-01 5.51738858e-01
1.23665236e-01 2.57782161e-01 1.16290438e+00 -1.03218436e+00
3.73153299e-01 1.06155920e+00 4.02792215e-01 4.68267471e-01
-1.24316752e+00 -4.97783393e-01 3.41862798e-01 3.80506068e-01
-1.22879052e+00 -4.02279735e-01 1.20501518e+00 -3.56328994e-01
3.18684816e-01 2.81260729e-01 6.80497050e-01 7.25436330e-01
1.84742108e-01 1.16498911e+00 1.37694013e+00 -3.65674466e-01
4.61390465e-01 2.98218042e-01 2.66231000e-02 6.55271411e-01
4.13239837e-01 -7.92087018e-02 -6.87509239e-01 3.39717306e-02
6.49658263e-01 1.82397753e-01 -3.69662344e-01 -9.67682004e-01
-1.49505949e+00 4.50811148e-01 6.61153495e-01 4.32240456e-01
-4.68644083e-01 -2.81013221e-01 3.09123576e-01 6.66391179e-02
4.75554854e-01 6.03531420e-01 -1.85081393e-01 2.74620414e-01
-1.11095977e+00 3.72382551e-01 1.91832781e-01 8.00113499e-01
9.49578524e-01 1.60246976e-02 -2.78999686e-01 7.30594099e-01
8.41949042e-03 2.65742958e-01 8.54837179e-01 -3.77553254e-01
2.81511933e-01 7.54294395e-01 2.40435481e-01 -1.23940098e+00
-3.70961696e-01 -1.06868827e+00 -8.52884829e-01 3.18998635e-01
3.35596889e-01 9.15603563e-02 -7.74514437e-01 1.84176910e+00
5.40070057e-01 3.90288830e-01 -1.25720110e-02 1.14360523e+00
4.10809904e-01 3.76364768e-01 1.88635617e-01 -4.31814015e-01
1.23242044e+00 -1.13344836e+00 -5.87911427e-01 -3.68167043e-01
1.51019022e-01 -4.62248266e-01 1.15047777e+00 2.53638744e-01
-1.01333547e+00 -8.69820654e-01 -1.22189641e+00 2.46024594e-01
-2.40990534e-01 6.36730120e-02 7.14145303e-01 1.98411912e-01
-9.01002586e-01 5.61694920e-01 -7.19053090e-01 -4.32106614e-01
5.96013725e-01 2.04032823e-01 -3.84134650e-02 1.93385169e-01
-8.95227611e-01 8.08936536e-01 5.43124378e-01 6.81115463e-02
-9.36092913e-01 -6.84695482e-01 -6.15782738e-01 1.78576037e-01
6.93785369e-01 -6.51369154e-01 9.28466916e-01 -1.45559001e+00
-1.34524715e+00 7.91506708e-01 -3.09821308e-01 -4.55157042e-01
4.99161661e-01 -2.95554549e-01 -4.08635229e-01 2.31631771e-01
1.31517127e-01 4.71698552e-01 1.30621362e+00 -1.34459257e+00
-8.12467754e-01 -4.97916996e-01 -9.06312168e-02 3.94624978e-01
-5.10176420e-01 -1.83279902e-01 -1.66089669e-01 -6.66189373e-01
2.72871614e-01 -8.16104949e-01 -2.30935395e-01 5.04016429e-02
-2.14855924e-01 -2.15498015e-01 9.27390993e-01 -2.36948952e-01
1.08196652e+00 -2.06913209e+00 3.02994639e-01 -1.24894626e-01
4.89860684e-01 6.00298822e-01 4.86325435e-02 2.24359229e-01
-1.02867134e-01 -3.65462571e-01 -2.09868833e-01 -3.01012427e-01
-4.22715515e-01 -9.45480615e-02 -3.12182277e-01 5.70839942e-01
3.65609258e-01 9.08544481e-01 -1.31727707e+00 -7.48807371e-01
1.91169739e-01 2.50405550e-01 -3.58494878e-01 3.59239608e-01
-1.80137411e-01 6.64755404e-01 -7.29043305e-01 7.01264679e-01
4.59694535e-01 -5.53489387e-01 5.15821278e-02 -3.50149050e-02
-2.27126896e-01 -9.90337953e-02 -1.13327003e+00 1.70801723e+00
-3.29640418e-01 3.84578347e-01 7.73629313e-03 -1.29767048e+00
1.12706923e+00 1.05896637e-01 4.70955789e-01 -6.76920116e-01
-3.56623344e-02 3.33502293e-01 3.32614481e-02 -4.62929815e-01
2.63593614e-01 -1.87305167e-01 2.99866945e-01 3.09374452e-01
1.41547009e-01 4.01813507e-01 -1.75336283e-02 -9.21025798e-02
6.22014701e-01 1.58105552e-01 5.95530450e-01 -6.63247705e-01
8.63658369e-01 -1.35280281e-01 7.42228210e-01 6.54364824e-01
-4.19987828e-01 4.16082472e-01 1.48451999e-01 -4.25339758e-01
-9.22794044e-01 -1.04393888e+00 -3.18275369e-03 9.49521720e-01
6.41242564e-01 -4.67151068e-02 -7.20785499e-01 -5.66986859e-01
-1.46049917e-01 4.60082024e-01 -6.51874006e-01 -3.28805536e-01
-2.98498750e-01 -6.67004824e-01 -7.08201155e-02 3.08449060e-01
5.97068250e-01 -1.28493154e+00 -1.27695775e+00 1.57452032e-01
-8.34496245e-02 -9.79478776e-01 -3.81548524e-01 2.53973395e-01
-9.12918210e-01 -9.34624612e-01 -9.57856178e-01 -9.67070580e-01
6.15200639e-01 8.53983462e-01 8.54232490e-01 1.33480772e-01
-2.78966248e-01 6.64162114e-02 -2.46898040e-01 -5.01275778e-01
7.12555880e-03 1.43260017e-01 2.34039485e-01 7.01193869e-01
2.31607005e-01 -8.64025652e-01 -8.18302631e-01 4.02349651e-01
-8.70027781e-01 3.34848791e-01 8.82985055e-01 9.35172915e-01
6.68709099e-01 4.80817631e-02 1.04449189e+00 -7.75221229e-01
4.09077764e-01 -5.08004010e-01 -6.03449464e-01 4.32245225e-01
-9.15135324e-01 1.58817604e-01 7.13050008e-01 -3.85171384e-01
-1.07858396e+00 1.74969003e-01 5.05701244e-01 -5.35952568e-01
-1.32507548e-01 3.58245164e-01 -2.77771086e-01 -1.58823118e-01
4.85771120e-01 8.03845465e-01 1.20157704e-01 -3.70497108e-01
2.65370846e-01 3.52304429e-01 8.31729949e-01 -5.63807726e-01
9.85472381e-01 5.60773492e-01 8.94358382e-02 -7.85269022e-01
-1.14384651e+00 -7.77450264e-01 -7.49031961e-01 -3.06968927e-01
6.73427641e-01 -8.93737435e-01 -4.91089135e-01 4.91741121e-01
-6.82200551e-01 1.11357845e-01 -2.33938560e-01 2.43607268e-01
-7.68728912e-01 4.11407471e-01 7.04124430e-03 -9.43280280e-01
-3.70632440e-01 -1.07047462e+00 9.70030785e-01 6.64557993e-01
1.02743290e-01 -9.53264177e-01 1.81462303e-01 2.33018696e-01
3.85291070e-01 3.88879001e-01 6.97507143e-01 -5.60700178e-01
-6.22123182e-01 -1.90253481e-02 -3.00805449e-01 9.36794430e-02
3.63005340e-01 -3.96837741e-01 -1.11192012e+00 -4.03237224e-01
1.57956898e-01 -3.53456795e-01 8.08076382e-01 1.69721052e-01
1.16458714e+00 -1.96126938e-01 -3.70308846e-01 4.19950247e-01
1.53403497e+00 -6.16785400e-02 3.88898402e-01 3.95982772e-01
5.41489840e-01 6.17899954e-01 8.66847515e-01 4.06223863e-01
3.17083985e-01 8.16352844e-01 6.45193338e-01 -3.30132186e-01
3.98658998e-02 -4.75358754e-01 2.73256134e-02 6.08009219e-01
-7.13583976e-02 1.57836497e-01 -7.33540952e-01 9.57861066e-01
-2.13383627e+00 -1.01690769e+00 1.78245828e-01 2.26575232e+00
9.00351763e-01 2.96512187e-01 3.96390110e-01 3.21452469e-01
7.78289676e-01 3.34521025e-01 -8.00367892e-01 1.55335575e-01
-1.86657473e-01 -1.37296468e-01 3.00519496e-01 4.82693724e-02
-1.19034457e+00 9.01408374e-01 5.29779720e+00 8.85001838e-01
-1.32262743e+00 9.68203396e-02 5.85116744e-01 2.35784307e-01
-1.82775840e-01 4.23082188e-02 -7.09846973e-01 4.38117355e-01
4.76515651e-01 -4.91455346e-01 3.30231965e-01 9.37136710e-01
2.13385835e-01 -5.67181297e-02 -8.92978489e-01 1.09028065e+00
2.24283189e-01 -1.14587367e+00 -3.70360315e-02 -1.62444934e-01
9.07486796e-01 -3.36165071e-01 3.53763849e-01 1.43930525e-01
-1.31825566e-01 -8.35153043e-01 8.30712914e-01 6.94000065e-01
3.79987270e-01 -6.09930575e-01 5.79263210e-01 5.54127753e-01
-1.10765254e+00 -3.36548686e-01 -4.25260603e-01 -2.82758892e-01
-1.73645034e-01 7.27624357e-01 -6.95270181e-01 6.35648847e-01
6.07941926e-01 9.63931262e-01 -9.66899812e-01 1.45398223e+00
4.17259075e-02 4.42028672e-01 7.22212419e-02 -1.38546601e-01
3.20338219e-01 -3.26091275e-02 7.38008142e-01 9.53635454e-01
1.62727619e-03 -1.79482117e-01 2.48299062e-01 9.57005978e-01
2.88377255e-01 3.03340107e-01 -4.65850055e-01 9.22143310e-02
2.49458551e-01 1.39623535e+00 -1.01823688e+00 -2.21794441e-01
-2.74244308e-01 8.16643298e-01 5.50097704e-01 2.64737427e-01
-7.10489988e-01 -2.24654272e-01 5.06286085e-01 1.01384528e-01
5.28900743e-01 -8.26020166e-02 -4.01229680e-01 -1.24218011e+00
2.13443443e-01 -8.73857021e-01 3.03264201e-01 -5.49596488e-01
-1.21448255e+00 5.49478233e-01 -9.45242047e-02 -1.59169447e+00
6.11338504e-02 -3.07135254e-01 -6.60845816e-01 6.68657660e-01
-1.91258335e+00 -1.27223051e+00 -4.83215749e-01 7.51074433e-01
8.54369521e-01 -1.19979806e-01 3.93389672e-01 3.51245180e-02
-5.11006176e-01 5.62033594e-01 7.38967583e-02 -4.84731764e-01
5.77457309e-01 -1.22881961e+00 4.44838554e-02 1.01045191e+00
3.14814419e-01 4.92854089e-01 9.66100931e-01 -3.68215591e-01
-1.29926455e+00 -9.93564188e-01 6.17831707e-01 -2.28658453e-01
5.71750581e-01 -3.55230242e-01 -1.20673931e+00 1.79468051e-01
7.97394961e-02 2.31169134e-01 2.84628958e-01 1.03045881e-01
-1.66039765e-01 -3.57317358e-01 -9.02888358e-01 5.91073036e-01
1.08820200e+00 -4.07986075e-01 -8.19980264e-01 2.41336852e-01
6.50516212e-01 -1.04476064e-01 -3.68264347e-01 3.96226376e-01
3.98297131e-01 -1.12757254e+00 9.22238171e-01 -5.42722106e-01
4.54959095e-01 -5.46432376e-01 1.42511442e-01 -1.28031015e+00
-4.70189422e-01 -7.06659615e-01 -1.84984878e-01 1.21795630e+00
6.95224404e-02 -5.03454030e-01 6.22928321e-01 3.54680158e-02
-1.50195444e-02 -1.02337170e+00 -7.42957234e-01 -7.66920984e-01
-2.68281132e-01 3.08540165e-02 4.54469591e-01 8.79876971e-01
1.58009082e-01 2.58228987e-01 -5.69130242e-01 2.28393584e-01
1.00132251e+00 5.83079815e-01 6.85729146e-01 -1.37931299e+00
-3.74430835e-01 -5.75623930e-01 -5.17500222e-01 -9.01715219e-01
8.20607468e-02 -8.18987787e-01 2.32992932e-01 -9.98939037e-01
4.17457312e-01 -4.43454564e-01 -7.99502552e-01 2.57477939e-01
-6.84121549e-01 7.08777532e-02 1.88322693e-01 4.53476012e-01
-9.86481547e-01 8.04292917e-01 1.25200999e+00 -9.47610885e-02
-2.79644817e-01 8.13096687e-02 -9.97957706e-01 7.57175624e-01
7.43677199e-01 -2.74660289e-01 -6.62843823e-01 -1.47826284e-01
-1.08988144e-01 -1.17405988e-01 6.81156218e-01 -1.41432881e+00
4.13590938e-01 -3.56574833e-01 2.03788996e-01 -4.37196702e-01
-7.27808326e-02 -7.18214989e-01 -1.60391927e-01 3.97754878e-01
-5.75950980e-01 -1.33533835e-01 7.30478168e-02 9.54451621e-01
-2.88892180e-01 1.42269023e-02 9.65770543e-01 -2.16899022e-01
-9.38062608e-01 2.28527933e-01 1.38628066e-01 2.40957305e-01
1.12775064e+00 -3.00741106e-01 -8.51568580e-02 -1.74635977e-01
-4.33006346e-01 3.57267350e-01 4.51864630e-01 5.82737863e-01
6.95441842e-01 -1.08161962e+00 -4.40685332e-01 3.35076213e-01
3.86107922e-01 -1.66142106e-01 3.99615556e-01 9.94219720e-01
2.60146447e-02 4.71748650e-01 -5.55365086e-01 -8.85802627e-01
-1.02216578e+00 9.73432422e-01 2.12702602e-01 -3.65173310e-01
-6.44817770e-01 5.69283724e-01 5.51759958e-01 6.81760982e-02
2.12082833e-01 -1.05175421e-01 -4.03076172e-01 -1.13937415e-01
6.32593632e-01 2.12824628e-01 -5.51045462e-02 -6.26828671e-01
-3.30005050e-01 6.66646898e-01 -1.06307663e-01 3.16206008e-01
1.35518932e+00 -4.57698017e-01 2.06085622e-01 6.98579192e-01
9.02030230e-01 -1.98453516e-01 -1.53596103e+00 -5.73766351e-01
2.98263758e-01 -5.10211349e-01 4.54380326e-02 -5.76936841e-01
-1.03396094e+00 8.05559039e-01 7.38715351e-01 6.72143027e-02
1.25591874e+00 -4.19525662e-04 7.28208482e-01 2.90358424e-01
5.18151164e-01 -1.03861392e+00 3.06426227e-01 -1.06440768e-01
7.48947561e-01 -1.40431368e+00 1.99013501e-02 -2.53519773e-01
-8.21318448e-01 8.95463347e-01 6.15673482e-01 -3.91570479e-01
6.62047982e-01 -2.82970369e-01 -1.20806433e-01 -1.94367498e-01
-7.32395768e-01 -4.24300164e-01 4.92374063e-01 5.49406111e-01
1.23846196e-01 -4.43694592e-02 -3.51563483e-01 4.74453539e-01
1.50102660e-01 1.01811342e-01 2.82096058e-01 8.00496876e-01
-6.78433716e-01 -6.91768289e-01 -1.86980888e-01 3.12115937e-01
-2.98193395e-01 1.96967319e-01 -1.22167617e-02 6.01758182e-01
1.01389728e-01 7.52751708e-01 -2.72216111e-01 -2.83111483e-01
1.37852475e-01 -6.69119135e-02 2.11466327e-01 -5.64215362e-01
-4.32390451e-01 8.21366832e-02 -5.23790479e-01 -5.71132541e-01
-6.93239868e-01 -9.14813936e-01 -9.05504286e-01 4.37712431e-01
-2.81492680e-01 1.38887018e-01 2.26618767e-01 1.06194925e+00
4.48980093e-01 4.17623162e-01 8.40904236e-01 -8.69682014e-01
-7.59783983e-01 -6.83990896e-01 -6.95957005e-01 6.67857945e-01
5.99139690e-01 -1.06071794e+00 -3.38261038e-01 -1.13354037e-02] | [9.821367263793945, -0.19872663915157318] |
270baf57-d29f-4449-99c2-9891386c9c45 | human-labeling-errors-and-their-impact-on | 2305.12106 | null | https://arxiv.org/abs/2305.12106v1 | https://arxiv.org/pdf/2305.12106v1.pdf | Human labeling errors and their impact on ConvNets for satellite image scene classification | Convolutional neural networks (ConvNets) have been successfully applied to satellite image scene classification. Human-labeled training datasets are essential for ConvNets to perform accurate classification. Errors in human-labeled training datasets are unavoidable due to the complexity of satellite images. However, the distribution of human labeling errors on satellite images and their impact on ConvNets have not been investigated. To fill this research gap, this study, for the first time, collected real-world labels from 32 participants and explored how their errors affect three ConvNets (VGG16, GoogleNet and ResNet-50) for high-resolution satellite image scene classification. We found that: (1) human labeling errors have significant class and instance dependence, which is fundamentally different from the simulation noise in previous studies; (2) regarding the overall accuracy of all classes, when human labeling errors in training data increase by one unit, the overall accuracy of ConvNets classification decreases by approximately half a unit; (3) regarding the accuracy of each class, the impact of human labeling errors on ConvNets shows large heterogeneity across classes. To uncover the mechanism underlying the impact of human labeling errors on ConvNets, we further compared it with two types of simulated labeling noise: uniform noise (errors independent of both classes and instances) and class-dependent noise (errors independent of instances but not classes). Our results show that the impact of human labeling errors on ConvNets is similar to that of the simulated class-dependent noise but not to that of the simulated uniform noise, suggesting that the impact of human labeling errors on ConvNets is mainly due to class-dependent errors rather than instance-dependent errors. | ['Xiaolin Zhu', 'Luoma Wan', 'Rui Sun', 'Xiaobei Chen', 'Xuehong Chen', 'Tao Wei', 'Longkang Peng'] | 2023-05-20 | null | null | null | null | ['scene-classification'] | ['computer-vision'] | [ 1.23720497e-01 -8.85553733e-02 3.62456203e-01 -5.62972665e-01
-2.33226046e-01 -5.24266064e-01 4.01615560e-01 5.38731478e-02
-9.81190383e-01 6.91276908e-01 -1.12504154e-01 -3.67270380e-01
-3.17509584e-02 -1.06922340e+00 -9.60286736e-01 -6.75837159e-01
-2.00727433e-01 2.31404752e-01 2.69719929e-01 -1.07598084e-03
1.81787293e-02 5.12699068e-01 -1.86827338e+00 2.10429952e-01
9.80046690e-01 8.43916297e-01 1.67254776e-01 3.67442489e-01
2.44606864e-02 1.16289699e+00 -1.00923681e+00 -7.80556574e-02
5.16535640e-01 -1.87957600e-01 -7.82770872e-01 4.12877381e-01
5.77500701e-01 -3.21983993e-01 -4.16588187e-01 1.40109098e+00
5.44229865e-01 6.79164603e-02 5.81362426e-01 -1.25100017e+00
-4.54827070e-01 4.81440157e-01 -4.16191816e-01 4.41393614e-01
-2.97051191e-01 3.69309723e-01 4.92188156e-01 -4.46657449e-01
5.17465293e-01 1.03982699e+00 1.00026953e+00 3.22110444e-01
-1.08742213e+00 -9.63609219e-01 1.88853890e-01 -8.45667422e-02
-1.74890649e+00 5.27554415e-02 1.70373902e-01 -8.83542895e-01
7.42001653e-01 1.06221341e-01 6.50414348e-01 8.16960752e-01
2.53780074e-02 2.10222915e-01 1.48897564e+00 -2.93601036e-01
3.93844038e-01 4.04774457e-01 6.05723202e-01 1.84687823e-01
5.36388099e-01 2.51310408e-01 2.34310217e-02 1.73014596e-01
6.90984249e-01 1.48612363e-02 -2.15787023e-01 3.68030339e-01
-1.02663875e+00 7.31459796e-01 9.44142342e-01 4.90840524e-01
-3.51472139e-01 1.41192809e-01 2.27239043e-01 2.50927895e-01
5.26359797e-01 3.22837293e-01 -5.98371267e-01 2.75831103e-01
-8.25976074e-01 1.39159262e-01 4.93597984e-01 7.06919074e-01
1.05784690e+00 2.92867392e-01 -1.39317751e-01 9.24732745e-01
-4.11716215e-02 7.14921176e-01 6.14244699e-01 -5.21114588e-01
3.68421167e-01 8.90017688e-01 1.06263138e-01 -1.14686203e+00
-9.25268054e-01 -7.86744475e-01 -9.58626986e-01 3.16413492e-01
8.25712383e-01 -3.82963479e-01 -1.20455110e+00 1.73661113e+00
-6.23887405e-02 7.58198425e-02 -2.29491651e-01 1.14748514e+00
1.14133525e+00 3.05001467e-01 5.87440133e-01 3.27855974e-01
1.41020381e+00 -4.54786599e-01 -3.52449626e-01 -5.33333600e-01
8.37896705e-01 -3.45497668e-01 1.16387463e+00 -4.14699912e-02
-4.52870488e-01 -7.06104279e-01 -9.16050136e-01 3.11806858e-01
-5.64728379e-01 3.51225346e-01 4.56569970e-01 9.90758240e-01
-9.96555030e-01 5.47355890e-01 -6.37426853e-01 -5.25106490e-01
7.33991206e-01 2.01127738e-01 -2.56875247e-01 -1.38619646e-01
-1.35840476e+00 1.03248787e+00 3.65935892e-01 1.96628347e-01
-1.04432440e+00 -8.70961905e-01 -7.05508232e-01 4.21218947e-02
7.36084878e-02 -1.57125786e-01 1.08663511e+00 -1.39303613e+00
-6.03817523e-01 1.05228460e+00 1.06273063e-01 -3.61853063e-01
5.50826967e-01 1.73746124e-01 -5.10206223e-01 -2.91540563e-01
3.35438222e-01 7.41162479e-01 1.97184190e-01 -1.37790477e+00
-6.51030600e-01 -4.89127159e-01 8.36969167e-02 1.22391276e-01
-1.33142387e-02 -1.43160224e-01 1.70346856e-01 -4.63394225e-01
1.55543894e-01 -1.00299883e+00 -3.57241482e-01 -1.99533612e-01
-2.65940487e-01 -1.99272826e-01 5.36346436e-01 -5.20004869e-01
7.92686880e-01 -2.27505779e+00 -7.20435560e-01 1.66967034e-01
3.20679694e-01 4.10784960e-01 -2.22068861e-01 -8.57741386e-02
-2.49086827e-01 5.61129451e-01 -2.71781206e-01 6.49657147e-03
-1.41425908e-01 2.37313315e-01 -8.97411630e-02 5.69984317e-01
2.50068545e-01 6.55687034e-01 -7.90483296e-01 -4.32861410e-02
3.26942146e-01 5.18082500e-01 -1.78419173e-01 -1.24373667e-01
1.22635059e-01 5.34813762e-01 -1.69834793e-01 5.93687117e-01
8.81334364e-01 -3.03716093e-01 3.46839242e-02 -1.78476781e-01
-3.49809796e-01 1.83086485e-01 -1.23990703e+00 7.04902470e-01
-3.05232584e-01 8.94140124e-01 -2.80270308e-01 -9.92092729e-01
1.00533640e+00 1.13744743e-01 1.28960952e-01 -1.13151312e+00
2.28922904e-01 2.56386340e-01 2.07335874e-01 -3.64940017e-01
5.24797976e-01 -2.76880503e-01 1.18016981e-01 4.56648320e-01
-5.53924255e-02 1.74259067e-01 1.00680992e-01 -7.82164261e-02
9.40722346e-01 -4.61501658e-01 2.04095244e-02 -6.26516461e-01
-1.48028899e-02 1.80716842e-01 7.31421530e-01 1.10166705e+00
-3.41657937e-01 6.29645228e-01 3.73977363e-01 -8.46467912e-01
-9.85185266e-01 -5.87438762e-01 -5.14151394e-01 8.99406612e-01
1.81375712e-01 1.63136739e-02 -8.80503654e-01 -5.26864409e-01
-5.61811328e-02 7.20058918e-01 -8.03008378e-01 1.52322529e-02
-9.98304188e-02 -1.24979556e+00 9.46942031e-01 6.16621315e-01
1.15953720e+00 -1.28329778e+00 -7.72369683e-01 -8.33712146e-02
-1.43905610e-01 -1.22172058e+00 4.80837151e-02 2.29043052e-01
-8.54796886e-01 -1.35091448e+00 -5.01493454e-01 -4.45914000e-01
1.04072607e+00 4.22910392e-01 1.22338355e+00 7.14500785e-01
-3.77640218e-01 1.75338537e-01 -2.89879888e-01 -5.78203738e-01
-1.89959511e-01 1.31987289e-01 -5.02311960e-02 -2.92449482e-02
4.97025609e-01 -3.64319950e-01 -5.69334447e-01 6.40148222e-01
-1.15835071e+00 -3.02001953e-01 3.55440736e-01 4.76352394e-01
2.70590276e-01 7.26305127e-01 3.45412761e-01 -1.04768467e+00
1.86843619e-01 -6.05389059e-01 -6.31431043e-01 9.88514125e-02
-4.51434642e-01 -4.34487700e-01 2.61602640e-01 -4.39580828e-01
-8.73853981e-01 -4.99168113e-02 4.49543186e-02 1.38264922e-02
-6.42388225e-01 6.22589350e-01 1.16783410e-01 -1.23179235e-01
1.02543199e+00 -1.63602568e-02 -3.74583513e-01 -3.30990970e-01
-3.08211327e-01 8.07185531e-01 1.66036800e-01 -2.68490106e-01
5.50044239e-01 5.27326763e-01 -3.06297362e-01 -1.04892218e+00
-8.27317238e-01 -1.69863790e-01 -5.37242234e-01 -4.12222236e-01
1.01301813e+00 -1.40670753e+00 -5.78590214e-01 8.97995532e-01
-8.09706569e-01 -8.47545147e-01 -2.38640293e-01 6.49472237e-01
1.82516828e-01 -9.43351444e-03 -4.25912619e-01 -7.21089184e-01
2.97688995e-03 -1.34161472e+00 7.57151723e-01 5.00471652e-01
-1.31942600e-01 -9.54897523e-01 -3.58099610e-01 2.72093773e-01
6.57133698e-01 1.90420508e-01 5.97029388e-01 -8.84421110e-01
-4.37388241e-01 -3.04042131e-01 -6.49568677e-01 4.45726782e-01
8.32812116e-02 -5.81803098e-02 -1.06746244e+00 -2.00236127e-01
-1.35470062e-01 -4.36889201e-01 7.83945322e-01 5.20614922e-01
9.33930576e-01 -3.30969878e-02 -1.20516911e-01 2.55288541e-01
1.45033205e+00 1.76471397e-01 1.04851091e+00 7.60138750e-01
5.71168900e-01 6.38949692e-01 2.67480820e-01 2.18869761e-01
4.30221915e-01 4.71550256e-01 3.70631367e-01 -4.80497479e-01
-1.73220977e-01 1.24855218e-02 -6.75403178e-02 -3.15866359e-02
-3.09826404e-01 -1.49166211e-01 -1.34058356e+00 6.49409890e-01
-1.50526190e+00 -7.88745522e-01 -6.65074050e-01 2.39352012e+00
4.32528824e-01 -1.42376900e-01 -1.06884502e-01 1.98221162e-01
1.07924747e+00 -8.67401287e-02 -3.08959305e-01 1.93584666e-01
-2.37898335e-01 -7.64169469e-02 1.01018667e+00 2.48493180e-01
-1.09952068e+00 8.30288410e-01 6.72633743e+00 5.81792891e-01
-1.30174172e+00 1.01029709e-01 8.44926476e-01 1.64014682e-01
9.47471112e-02 -1.74501821e-01 -8.45506728e-01 5.95781803e-01
8.12552571e-01 2.37437293e-01 1.45457000e-01 8.09678018e-01
2.25078061e-01 -4.65982854e-01 -4.35632437e-01 7.60415137e-01
-1.68556944e-01 -8.19913447e-01 -2.04104468e-01 1.39381886e-01
9.36927378e-01 3.52267563e-01 -2.11562321e-01 5.29045582e-01
7.32357442e-01 -1.19263124e+00 8.13202918e-01 2.52390385e-01
6.57320857e-01 -6.40210092e-01 1.25697756e+00 4.68696237e-01
-8.27369511e-01 -1.06934652e-01 -5.63932776e-01 -6.25870764e-01
-2.37906557e-02 8.70316267e-01 -5.01473486e-01 6.90585151e-02
1.18513441e+00 4.92954761e-01 -9.51120794e-01 9.93760288e-01
-3.91484559e-01 9.50225711e-01 -3.10341775e-01 2.76721835e-01
2.96417177e-01 1.74950644e-01 9.80816260e-02 1.16769469e+00
2.08468065e-01 4.70449567e-01 1.84729055e-01 8.66234064e-01
-4.04494889e-02 -2.48968348e-01 -4.34694588e-01 -2.54079755e-02
7.15921462e-01 1.19042552e+00 -1.16441989e+00 -3.87572378e-01
-4.09432679e-01 4.61134076e-01 2.59548962e-01 5.88807583e-01
-8.28366637e-01 -1.84710816e-01 5.58654547e-01 3.38901609e-01
1.30622029e-01 -6.00147154e-03 -6.91588461e-01 -9.86333489e-01
-1.85959503e-01 -7.39237547e-01 2.29499146e-01 -8.24973464e-01
-1.18807685e+00 7.95989096e-01 -1.03947736e-01 -1.11001408e+00
3.86731476e-01 -4.57827330e-01 -4.61018085e-01 8.99060845e-01
-1.35331357e+00 -9.06582654e-01 -9.22994673e-01 5.45452416e-01
1.14056356e-01 -3.34901884e-02 6.13319278e-01 5.99808514e-01
-6.35704637e-01 5.10873854e-01 -2.63858825e-01 7.82916665e-01
3.96141201e-01 -7.54510045e-01 3.77398521e-01 7.67134368e-01
-5.52152656e-03 5.28665960e-01 4.80961442e-01 -7.16447711e-01
-4.56879109e-01 -1.47951913e+00 9.83657897e-01 -4.19528663e-01
2.75138825e-01 -4.69808392e-02 -1.02260721e+00 9.22243237e-01
-2.94303060e-01 2.69618571e-01 6.94500208e-01 -3.96435149e-02
-3.87745887e-01 3.61633636e-02 -1.42198622e+00 2.93824822e-01
9.47486341e-01 -3.98111761e-01 -2.23807409e-01 3.46578479e-01
5.62803686e-01 -4.18914050e-01 -6.33520901e-01 6.48513019e-01
2.50153244e-01 -1.14697361e+00 7.07374752e-01 -4.90721196e-01
4.08853203e-01 -5.53702235e-01 -2.81838745e-01 -1.38111508e+00
-4.80899125e-01 6.01238668e-01 1.01785600e+00 1.06806374e+00
5.53091347e-01 -8.71016502e-01 5.05767167e-01 1.05672801e+00
-1.60070181e-01 -4.52138134e-04 -7.40564287e-01 -7.77471900e-01
7.62578472e-02 -5.92310429e-01 6.63995862e-01 1.41655374e+00
-7.67193377e-01 -1.36547446e-01 -1.08174488e-01 5.86191714e-01
4.72733498e-01 -4.74037617e-01 6.07706606e-01 -1.21132028e+00
2.29367800e-02 -1.91339448e-01 -7.03423858e-01 -3.96610856e-01
1.87278036e-02 -7.55074143e-01 1.95344716e-01 -1.58506274e+00
2.91309416e-01 -8.60378861e-01 -9.03940424e-02 7.38797486e-01
-2.01900452e-01 8.17283332e-01 3.64526123e-01 3.76738548e-01
-4.04373229e-01 8.46805871e-02 9.05572355e-01 -9.14094150e-02
-3.87079045e-02 1.48496740e-02 -5.87121665e-01 7.80928016e-01
7.40743577e-01 -8.07436585e-01 -3.33004482e-02 -7.19706059e-01
3.16612452e-01 -4.47343171e-01 8.56737196e-01 -1.39022064e+00
2.83662051e-01 4.39637750e-02 5.24613023e-01 -3.32465351e-01
-3.31060261e-01 -9.55288529e-01 5.51302135e-01 5.98157227e-01
-2.40484625e-01 -2.77307332e-01 5.42538941e-01 2.80450791e-01
-1.93227947e-01 -2.37870902e-01 8.28360856e-01 -3.27640772e-01
-9.33466673e-01 -4.82759736e-02 -9.15287256e-01 -1.85716003e-02
8.71048391e-01 -2.90840000e-01 -5.70537269e-01 -2.67925769e-01
-7.10955024e-01 8.99529085e-02 3.67348433e-01 2.58676022e-01
6.23716554e-03 -1.13167238e+00 -7.49479830e-01 1.69351831e-01
-1.04353828e-02 6.91653341e-02 5.19303322e-01 7.77144969e-01
-6.39206350e-01 3.18579048e-01 -2.92723268e-01 -7.52710998e-01
-1.15729225e+00 -5.21975867e-02 8.36425066e-01 -6.82966113e-02
-2.66046703e-01 8.23565602e-01 3.18976015e-01 -9.87539589e-01
2.48035446e-01 -2.14035884e-01 -2.03096420e-01 1.18173294e-01
7.13790178e-01 6.03187144e-01 2.67503500e-01 -7.66878426e-01
-3.55417788e-01 5.91699362e-01 1.51537001e-01 2.71899313e-01
1.23214221e+00 3.27642709e-02 -6.76776171e-02 5.07289767e-01
7.65437305e-01 -5.73797584e-01 -1.04430842e+00 -2.81360626e-01
-2.00194716e-01 -4.98860896e-01 2.02978820e-01 -1.01252341e+00
-1.48037195e+00 8.31037521e-01 8.75882626e-01 2.09172502e-01
9.91790056e-01 -1.84130043e-01 2.70067394e-01 2.18049601e-01
6.04027092e-01 -9.41982090e-01 -1.63894996e-01 5.30170262e-01
6.10013247e-01 -1.45235920e+00 -7.90137872e-02 -3.71866256e-01
-8.61293077e-01 6.78002477e-01 8.44759762e-01 -1.07019514e-01
6.85118675e-01 8.75975490e-02 4.41808969e-01 -4.45844203e-01
-1.51699096e-01 -4.56421226e-01 1.31813109e-01 6.10590935e-01
5.62430918e-01 3.40303570e-01 -2.89584160e-01 7.36027181e-01
-2.62650698e-01 2.88038582e-01 5.06066382e-01 7.58619666e-01
-4.10976082e-01 -4.71366167e-01 -5.36067009e-01 7.03264117e-01
-3.59166443e-01 -8.58630836e-02 -3.31106007e-01 1.00731707e+00
5.43258429e-01 1.11006260e+00 5.61583638e-01 -5.37808597e-01
6.77454114e-01 -3.28637362e-02 -6.93874061e-02 -6.54739678e-01
-9.43185866e-01 -6.77289546e-01 1.01932585e-01 -4.13391262e-01
-5.41605592e-01 -3.87336373e-01 -1.24882269e+00 -5.68948805e-01
-4.41354811e-01 -7.78978989e-02 7.14347541e-01 1.03462875e+00
2.09963664e-01 6.01882517e-01 3.50870013e-01 -8.65905046e-01
-3.24701607e-01 -1.33020151e+00 -9.59266126e-01 6.59897625e-01
-9.15068910e-02 -6.95717692e-01 -9.13232148e-01 -1.28082469e-01] | [9.4462251663208, -1.1872762441635132] |
9053903d-ecb2-4357-910b-a228548dd33b | learning-social-navigation-from | 2210.03582 | null | https://arxiv.org/abs/2210.03582v2 | https://arxiv.org/pdf/2210.03582v2.pdf | Learning Social Navigation from Demonstrations with Conditional Neural Processes | Sociability is essential for modern robots to increase their acceptability in human environments. Traditional techniques use manually engineered utility functions inspired by observing pedestrian behaviors to achieve social navigation. However, social aspects of navigation are diverse, changing across different types of environments, societies, and population densities, making it unrealistic to use hand-crafted techniques in each domain. This paper presents a data-driven navigation architecture that uses state-of-the-art neural architectures, namely Conditional Neural Processes, to learn global and local controllers of the mobile robot from observations. Additionally, we leverage a state-of-the-art, deep prediction mechanism to detect situations not similar to the trained ones, where reactive controllers step in to ensure safe navigation. Our results demonstrate that the proposed framework can successfully carry out navigation tasks regarding social norms in the data. Further, we showed that our system produces fewer personal-zone violations, causing less discomfort. | ['Emre Ugur', 'Yigit Yildirim'] | 2022-10-07 | null | null | null | null | ['social-navigation'] | ['robots'] | [-2.05681637e-01 3.88394028e-01 9.92635638e-02 -4.94919479e-01
-5.98331913e-02 -2.21906334e-01 3.27522784e-01 -1.91471472e-01
-7.19136715e-01 1.03427207e+00 1.09886028e-01 -2.79204935e-01
-1.29318461e-01 -9.80245650e-01 -7.98232555e-01 -3.86478007e-01
-3.49607795e-01 2.79471606e-01 2.45496362e-01 -9.07233179e-01
3.22248548e-01 7.68890679e-02 -1.91563070e+00 -2.52031654e-01
1.21743011e+00 5.38178742e-01 8.17179456e-02 6.25493228e-01
5.86164057e-01 7.41237819e-01 -4.29709941e-01 -3.40885192e-01
8.55068937e-02 -3.87967020e-01 -4.39963043e-01 -4.21943039e-01
-6.04644604e-02 -4.54600215e-01 -2.92384893e-01 1.06084406e+00
3.45898956e-01 5.32249212e-01 7.83600211e-01 -1.47525239e+00
-8.45935464e-01 4.73598063e-01 -3.90805490e-02 -4.34833437e-01
6.82255030e-01 4.95843321e-01 8.28811169e-01 -7.37485215e-02
4.53363746e-01 1.41629839e+00 8.78150761e-01 8.06878984e-01
-1.06384647e+00 -7.59176493e-01 3.73421431e-01 1.70960471e-01
-9.32344198e-01 -3.05551827e-01 5.77412248e-01 -4.25979793e-01
1.04323697e+00 -6.60935864e-02 8.77174854e-01 1.91028988e+00
5.83368599e-01 5.86169422e-01 7.77326107e-01 2.66528856e-02
7.32880652e-01 1.96222942e-02 -1.04653917e-01 7.33587861e-01
6.70881033e-01 2.05089837e-01 -4.30435747e-01 -7.89669678e-02
5.92239857e-01 -6.94210082e-02 3.29698510e-02 -6.78530037e-01
-8.71191919e-01 8.06765735e-01 7.84826338e-01 -7.29619116e-02
-5.36823094e-01 3.23240787e-01 3.73420715e-01 1.18080087e-01
-1.06306121e-01 5.17389297e-01 -4.42796052e-01 -6.24493301e-01
-9.33942795e-02 3.83185565e-01 1.03995645e+00 1.12136030e+00
6.19248688e-01 -1.92425251e-01 2.66710371e-01 5.69907010e-01
5.97330391e-01 7.06576705e-01 3.85087222e-01 -1.45153010e+00
2.14507699e-01 5.99898517e-01 5.45548260e-01 -1.23484933e+00
-8.44429672e-01 -1.39782280e-01 -4.64723885e-01 5.59751451e-01
4.45078224e-01 -3.95156860e-01 -5.95068097e-01 2.17365503e+00
2.15279251e-01 -2.58499861e-01 -3.17395665e-02 7.88262725e-01
1.57242835e-01 2.34894320e-01 2.82741249e-01 5.16719759e-01
9.30306733e-01 -9.95329320e-01 -4.76628006e-01 -7.18623102e-01
6.85366273e-01 1.76509172e-01 1.47665918e+00 5.38648248e-01
-6.41853094e-01 -3.33186150e-01 -1.21439183e+00 3.04567758e-02
-6.49752498e-01 -1.44334927e-01 6.86215520e-01 9.75008607e-01
-1.22888005e+00 4.97402757e-01 -1.02623343e+00 -1.03415930e+00
3.28354388e-01 4.45950538e-01 -3.88247401e-01 3.75034094e-01
-1.21380389e+00 1.30997539e+00 2.81570368e-02 1.53442591e-01
-1.03666282e+00 2.29268931e-02 -1.09796035e+00 5.56202568e-02
1.95078686e-01 -7.59211898e-01 1.29201257e+00 -7.49914229e-01
-1.93027878e+00 6.72002852e-01 5.99767454e-02 -5.19528091e-01
5.92844188e-01 -5.48102200e-01 -2.63258815e-01 -4.41213042e-01
4.32294399e-01 7.63807535e-01 4.23819005e-01 -1.29276431e+00
-6.96889281e-01 -2.26615056e-01 4.04183090e-01 2.72547901e-01
-3.46882582e-01 -4.77979988e-01 1.40694119e-02 -1.21203903e-02
-3.97883534e-01 -1.23215830e+00 -6.76732123e-01 9.28585455e-02
-3.55720699e-01 2.85104960e-02 6.07159138e-01 -3.89945060e-01
8.25673044e-01 -2.04067016e+00 6.03856929e-02 3.74374479e-01
5.08441627e-02 -1.00547649e-01 -1.14678569e-01 3.62153500e-01
6.19989574e-01 -8.67652148e-02 -1.86779484e-01 -4.49227124e-01
4.71626788e-01 3.19332778e-01 8.70210007e-02 4.09176737e-01
-1.96688533e-01 5.75383544e-01 -1.31395435e+00 -4.19270359e-02
4.15403277e-01 3.75272572e-01 -1.06187224e+00 1.46372318e-01
-1.30389091e-02 7.53832638e-01 -4.02748972e-01 4.00410026e-01
2.50104606e-01 2.01598004e-01 3.73582751e-01 6.16330504e-01
-1.39724404e-01 3.57915133e-01 -7.61817157e-01 1.65021384e+00
-6.96337342e-01 5.46927452e-01 2.02959597e-01 -5.41431248e-01
8.25119376e-01 -1.51587859e-01 2.32413292e-01 -9.39792395e-01
6.29988730e-01 3.07024568e-01 7.35113472e-02 -6.39568090e-01
5.44876158e-01 3.78162801e-01 -6.60669982e-01 2.27198243e-01
-1.17006786e-01 -1.34969011e-01 1.14994608e-01 -3.34263533e-01
1.28396308e+00 4.77409601e-01 4.59242374e-01 -3.50575954e-01
3.89186114e-01 -1.21002778e-01 5.80569863e-01 1.11108494e+00
-9.17296231e-01 1.38078123e-01 5.05466104e-01 -3.33869845e-01
-7.89483368e-01 -1.28480816e+00 3.16045821e-01 1.53532493e+00
6.02659822e-01 -1.31820351e-01 -1.00027919e+00 -5.52401423e-01
-7.63888359e-02 1.06948948e+00 -8.79988968e-01 -6.73291326e-01
-5.59552550e-01 -4.62866277e-01 5.81951618e-01 5.43098211e-01
6.26419008e-01 -1.28803253e+00 -1.22789419e+00 2.15154365e-01
-1.01082094e-01 -8.27045679e-01 -9.70658511e-02 1.98729366e-01
-3.75678986e-01 -1.05267906e+00 -2.44605854e-01 -8.70300472e-01
6.74298346e-01 9.25476477e-02 6.70544147e-01 -1.23364821e-01
2.48600423e-01 3.57508421e-01 -2.79477894e-01 -3.45126599e-01
-3.20799917e-01 2.83509791e-01 5.23973942e-01 -2.48622194e-01
4.77845579e-01 -9.04021740e-01 -7.72105932e-01 6.32312655e-01
-1.38935521e-01 -2.92132527e-01 4.69682038e-01 7.42058456e-01
-2.54631221e-01 -6.08217008e-02 5.89156270e-01 -5.26797593e-01
1.13629162e+00 -6.81966960e-01 -4.04387057e-01 -9.28126201e-02
-4.11792576e-01 -3.06996834e-02 8.18803191e-01 -3.78366441e-01
-1.03680408e+00 -2.45678686e-02 -8.66749585e-02 4.92565364e-01
-6.05420649e-01 1.32000670e-01 -2.58502096e-01 5.30132689e-02
1.06012547e+00 -1.12481490e-01 1.42016396e-01 9.34978127e-02
6.19698644e-01 8.66192341e-01 4.10685837e-01 -6.55556977e-01
3.94625038e-01 5.22932649e-01 -2.81881332e-01 -8.18170309e-01
-3.45176488e-01 -1.00814469e-01 -6.05449319e-01 -5.10464787e-01
1.07293117e+00 -7.63695359e-01 -1.37604570e+00 5.67905307e-01
-8.26191187e-01 -9.87119317e-01 3.53761092e-02 3.32859218e-01
-9.67911422e-01 -9.79239196e-02 -4.91469204e-01 -1.22082055e+00
4.53652292e-02 -1.20416808e+00 1.03981745e+00 5.54975569e-01
-7.60263085e-01 -9.20331776e-01 2.41853237e-01 2.04364762e-01
7.03289509e-01 1.68875799e-01 5.30958414e-01 -3.90416175e-01
-2.69426733e-01 -2.32946962e-01 9.69389752e-02 4.41615470e-03
2.28372321e-01 -2.89169431e-01 -9.31351185e-01 -1.69802129e-01
-3.09755355e-01 -4.83154804e-01 2.77993381e-01 3.60084027e-01
6.98431373e-01 -2.65839875e-01 -4.72057372e-01 2.31924683e-01
6.94829881e-01 6.07526600e-01 7.21076131e-01 8.41511130e-01
3.92096967e-01 8.62210870e-01 6.19597912e-01 3.70737493e-01
9.87974763e-01 4.56017792e-01 6.45303905e-01 1.76415354e-01
4.35459971e-01 -4.78932828e-01 8.03039551e-01 2.18157887e-01
-1.52423427e-01 -2.63999701e-01 -9.49514151e-01 5.73364675e-01
-2.38663149e+00 -8.69514942e-01 2.56739184e-02 1.97757816e+00
3.32024425e-01 4.96299654e-01 2.42936835e-01 -2.83883423e-01
5.66654325e-01 -1.58767983e-01 -8.26987565e-01 -4.91072118e-01
2.77209431e-01 -3.58686328e-01 5.80526114e-01 5.62180698e-01
-1.13391411e+00 1.03426707e+00 6.83747816e+00 1.20544098e-01
-8.90568018e-01 -6.30522054e-03 2.92242497e-01 6.06393814e-03
2.07473040e-02 -3.03687602e-01 -3.90096039e-01 4.74547416e-01
1.09890735e+00 1.52494222e-01 5.74953437e-01 1.37770283e+00
6.37270451e-01 -4.08588618e-01 -1.10683692e+00 6.80212498e-01
2.75950115e-02 -4.54434514e-01 -3.96317095e-01 8.70148316e-02
5.84800959e-01 1.90691035e-02 2.54802316e-01 8.17656338e-01
9.60697234e-01 -9.07188594e-01 1.05892003e+00 5.40775001e-01
1.04755156e-01 -7.50723302e-01 8.26329410e-01 4.49028105e-01
-9.93419588e-01 -5.34641445e-01 -1.93523929e-01 -7.08226740e-01
3.25214297e-01 -6.43752366e-02 -7.16501474e-01 3.58479321e-02
1.02247667e+00 6.86727703e-01 -5.33933640e-01 9.68498468e-01
-5.36496162e-01 1.31568015e-01 -3.83271992e-01 -7.81386316e-01
3.03070158e-01 -4.50355351e-01 4.18626219e-01 6.99260354e-01
3.66797775e-01 -4.94545162e-01 -8.23867619e-02 8.57055843e-01
2.19602391e-01 -2.72926092e-01 -1.27953160e+00 4.77169693e-01
4.50032145e-01 7.97335744e-01 -6.40162289e-01 1.75848961e-01
-1.75169244e-01 1.13060069e+00 4.68944043e-01 5.18247128e-01
-1.12477839e+00 -5.58916152e-01 1.18830407e+00 -4.95585874e-02
1.59093589e-02 -5.41841924e-01 -6.75748587e-01 -9.47295308e-01
-1.86161790e-02 -5.97419620e-01 1.23310247e-02 -7.58217931e-01
-1.25524211e+00 3.57599974e-01 -1.88703373e-01 -1.26467335e+00
-3.00012916e-01 -8.23433399e-01 -6.62011325e-01 2.65046775e-01
-9.84593928e-01 -9.42057908e-01 -4.65173095e-01 2.47182176e-01
2.60189801e-01 7.60704419e-03 9.21731472e-01 2.62726903e-01
-6.02954388e-01 5.11155665e-01 -8.20904970e-05 2.12278701e-02
7.26741612e-01 -1.14559734e+00 5.89135587e-01 7.50881672e-01
-7.30620503e-01 1.06265259e+00 8.90074611e-01 -6.78208292e-01
-9.92682397e-01 -8.89894545e-01 4.66908157e-01 -6.38002813e-01
7.04850554e-01 -7.52958357e-01 -2.49072865e-01 9.78890598e-01
7.88200870e-02 -6.63551807e-01 7.22065449e-01 5.17546952e-01
-8.40363204e-02 -3.09420880e-02 -1.39216542e+00 1.49745440e+00
1.63078046e+00 -4.84670877e-01 -5.27372718e-01 -1.60411820e-01
5.71441054e-01 -1.48874864e-01 -2.41171464e-01 2.84855887e-02
9.44343925e-01 -1.09737706e+00 8.29189777e-01 -5.16532838e-01
3.25012982e-01 -2.32883915e-01 -2.44348854e-01 -1.58765483e+00
-3.43455285e-01 -5.00442266e-01 1.44814402e-01 7.98912227e-01
7.59689152e-01 -1.02720106e+00 8.62282872e-01 1.06826591e+00
-1.97566152e-01 -3.55711162e-01 -1.00137532e+00 -7.37113774e-01
-4.89551388e-02 -4.41932261e-01 5.66099942e-01 6.19283676e-01
5.33998191e-01 4.01302427e-01 -5.73390841e-01 2.30101913e-01
4.71500516e-01 -5.25152028e-01 1.18516743e+00 -1.23386204e+00
1.17738307e-01 -6.60166025e-01 -6.00220978e-01 -1.11424816e+00
3.88428241e-01 -4.02079701e-01 6.95164323e-01 -1.69294572e+00
-3.59938174e-01 -3.90395343e-01 -4.89647798e-02 4.23199326e-01
5.62315807e-02 -1.47703305e-01 -1.16268873e-01 -4.04326886e-01
-9.04642999e-01 1.06226254e+00 8.90017331e-01 -1.79509699e-01
-4.89035577e-01 -4.27654572e-02 -1.00340068e+00 9.72589433e-01
1.05629694e+00 -2.41019532e-01 -6.14905477e-01 -3.42025578e-01
3.91443133e-01 -4.24064010e-01 2.38229275e-01 -1.55489433e+00
4.36082751e-01 -3.10925275e-01 1.05695136e-01 -1.65720086e-03
5.50222993e-01 -1.03127742e+00 -1.35494992e-01 8.40359509e-01
-2.57658571e-01 -1.36470735e-01 5.00286557e-02 7.89949834e-01
3.23974282e-01 2.22488120e-01 6.70434475e-01 -3.89533192e-02
-7.42105544e-01 -2.16357127e-01 -1.14749694e+00 -2.22258359e-01
1.38291812e+00 -2.60923147e-01 -3.39181513e-01 -8.09456527e-01
-5.20246625e-01 6.95940614e-01 7.35894084e-01 6.60555780e-01
5.25163352e-01 -1.10821819e+00 -8.89015198e-02 2.44709224e-01
2.76250601e-01 -1.58006340e-01 2.15519488e-01 7.06974030e-01
-8.69374990e-01 2.22003520e-01 -7.15899289e-01 -4.29276377e-01
-5.12882769e-01 3.37830752e-01 6.13583386e-01 9.01623890e-02
-4.04371887e-01 7.19176769e-01 3.79770473e-02 -1.35418022e+00
3.47699016e-01 -5.73505461e-01 -2.52870262e-01 -2.94086933e-01
2.19246536e-01 4.57138300e-01 -2.40813375e-01 -3.64248037e-01
-5.17737269e-01 3.14364851e-01 3.49456370e-01 -2.62382567e-01
1.20454395e+00 -4.36275780e-01 3.08574587e-01 5.41118264e-01
6.86252832e-01 -2.88163722e-01 -1.57207239e+00 5.77914774e-01
-8.42321012e-03 -2.18645841e-01 -4.78723109e-01 -7.25901067e-01
-4.05406296e-01 5.04626632e-01 5.30797362e-01 1.61050811e-01
6.32809281e-01 -4.17952001e-01 9.14657950e-01 1.02247322e+00
1.15528262e+00 -1.47599912e+00 -2.65636332e-02 8.81397069e-01
6.65396810e-01 -1.55179501e+00 -6.19437635e-01 -7.34352246e-02
-7.00348556e-01 6.64113700e-01 1.16293013e+00 -2.89951414e-01
7.99288273e-01 -4.71331365e-02 1.70322329e-01 1.05622429e-02
-4.77323115e-01 -1.85761452e-01 -3.62549067e-01 1.21516252e+00
-1.80873554e-02 3.32293957e-01 -7.08161816e-02 7.88938284e-01
-7.78108954e-01 -1.09721176e-01 6.71494544e-01 1.22174931e+00
-6.08041108e-01 -7.00949848e-01 -3.05214167e-01 3.72178406e-01
2.27275223e-01 3.97731155e-01 -5.77998281e-01 8.12518001e-01
1.20927937e-01 1.38656306e+00 -1.06865041e-01 -8.34480345e-01
7.26222813e-01 -1.80276170e-01 1.23655111e-01 -4.10716146e-01
-4.17048752e-01 -7.41040051e-01 4.56878483e-01 -1.01533520e+00
-1.43384293e-01 -5.79536200e-01 -1.36901128e+00 -5.29531479e-01
1.60191044e-01 -1.91760048e-01 4.87065643e-01 7.92306066e-01
4.53544915e-01 4.70039159e-01 2.89419115e-01 -1.34594274e+00
-1.95184454e-01 -8.73524845e-01 -4.63541687e-01 4.38980103e-01
3.12805146e-01 -1.13945770e+00 -3.78794968e-01 -2.66087025e-01] | [4.776705741882324, 0.937761127948761] |
fb8bd677-f91e-4016-9180-a4ba1867dc82 | on-the-effectiveness-of-gan-generated-cardiac | 2005.09026 | null | https://arxiv.org/abs/2005.09026v2 | https://arxiv.org/pdf/2005.09026v2.pdf | On the effectiveness of GAN generated cardiac MRIs for segmentation | In this work, we propose a Variational Autoencoder (VAE) - Generative Adversarial Networks (GAN) model that can produce highly realistic MRI together with its pixel accurate groundtruth for the application of cine-MR image cardiac segmentation. On one side of our model is a Variational Autoencoder (VAE) trained to learn the latent representations of cardiac shapes. On the other side is a GAN that uses "SPatially-Adaptive (DE)Normalization" (SPADE) modules to generate realistic MR images tailored to a given anatomical map. At test time, the sampling of the VAE latent space allows to generate an arbitrary large number of cardiac shapes, which are fed to the GAN that subsequently generates MR images whose cardiac structure fits that of the cardiac shapes. In other words, our system can generate a large volume of realistic yet labeled cardiac MR images. We show that segmentation with CNNs trained with our synthetic annotated images gets competitive results compared to traditional techniques. We also show that combining data augmentation with our GAN-generated images lead to an improvement in the Dice score of up to 12 percent while allowing for better generalization capabilities on other datasets. | ['Pierre-Marc Jodoin', 'Youssef Skandarani', 'Nathan Painchaud', 'Alain Lalande'] | 2020-05-18 | null | https://openreview.net/forum?id=f9Pl3Qj3_Q | https://openreview.net/pdf?id=f9Pl3Qj3_Q | midl-2019-7 | ['cardiac-segmentation'] | ['medical'] | [ 2.48469383e-01 7.47597337e-01 4.96339440e-01 -3.15206379e-01
-1.02140367e+00 -4.82731372e-01 3.63000363e-01 -4.20784533e-01
-1.56785980e-01 8.89185429e-01 1.91262141e-01 -1.21841155e-01
5.46750247e-01 -1.09896863e+00 -9.53556597e-01 -9.05039072e-01
9.94144678e-02 1.04548502e+00 3.35454680e-02 -1.12302817e-01
-3.73922706e-01 5.12060463e-01 -8.19757342e-01 3.81528080e-01
7.67738461e-01 8.20395231e-01 7.53578395e-02 9.39556122e-01
1.49503604e-01 7.95162201e-01 -7.60154903e-01 -2.70352513e-01
4.04911816e-01 -7.98946500e-01 -7.81357765e-01 5.98583780e-02
3.22345912e-01 -5.71693361e-01 -3.32486778e-01 7.71843672e-01
5.99280179e-01 -5.29594235e-02 1.03617167e+00 -1.09823596e+00
-6.81920648e-01 6.90421045e-01 -2.86893338e-01 6.30370080e-02
-1.80877417e-01 4.66030657e-01 4.28744853e-01 -4.97656494e-01
9.91728008e-01 8.46838176e-01 6.58380270e-01 1.09945226e+00
-1.52758622e+00 -4.49769914e-01 -6.86070383e-01 -4.79430944e-01
-1.13589907e+00 1.04185846e-03 8.59816968e-01 -5.97673476e-01
3.18213999e-01 2.66782343e-01 8.47566247e-01 1.28485084e+00
4.90060806e-01 5.58482945e-01 1.13553524e+00 -1.65106580e-01
4.64503050e-01 9.56803374e-03 -5.34306288e-01 5.69130242e-01
-9.14852843e-02 1.26289770e-01 2.31240034e-01 -7.69760013e-02
1.38034773e+00 -3.14305760e-02 -2.58452594e-01 -3.78317058e-01
-1.30894434e+00 1.22118723e+00 8.14429045e-01 3.77465844e-01
-9.06918526e-01 4.54466671e-01 3.90818834e-01 -1.04558393e-01
3.91749829e-01 7.16497481e-01 7.96445608e-02 3.07972670e-01
-1.21960092e+00 2.95038700e-01 5.41987777e-01 6.72191322e-01
3.71560872e-01 6.78007483e-01 -5.70063770e-01 6.63649619e-01
7.39396140e-02 4.89801496e-01 6.55998945e-01 -1.15350211e+00
1.00568652e-01 3.92981380e-01 -1.46158502e-01 -6.09839380e-01
-1.23790838e-01 -6.81825697e-01 -1.18849897e+00 4.67173100e-01
3.64971250e-01 -4.24017102e-01 -1.36006248e+00 1.80913913e+00
3.84443730e-01 1.81064054e-01 2.18394175e-01 9.02172983e-01
7.99565792e-01 6.88446164e-01 1.89082429e-01 3.63850892e-02
1.16461706e+00 -8.35874319e-01 -5.82743227e-01 -1.08825207e-01
3.28133076e-01 -4.68622029e-01 9.12638903e-01 6.40436932e-02
-1.37355781e+00 -3.72897178e-01 -1.09344804e+00 5.53245917e-02
5.47536388e-02 -1.31736159e-01 5.03740549e-01 6.75768256e-01
-1.40345967e+00 5.61948538e-01 -1.16343224e+00 1.79797232e-01
7.81378984e-01 2.27536082e-01 -2.48875380e-01 5.58100678e-02
-1.01434982e+00 6.64875209e-01 3.24027061e-01 -1.38633460e-01
-1.54088521e+00 -9.20410395e-01 -9.26981688e-01 -1.13715798e-01
-1.03896104e-01 -1.06676710e+00 9.40418839e-01 -1.28867233e+00
-1.41491973e+00 8.99137378e-01 3.78095597e-01 -7.33866096e-01
1.00414824e+00 4.06948447e-01 -8.98261592e-02 4.77640182e-01
1.55197531e-01 1.35585892e+00 1.00122619e+00 -1.52980065e+00
2.82651722e-01 -3.69026535e-03 -2.92256773e-01 -5.95878139e-02
2.55713433e-01 -3.12874913e-01 -1.72236890e-01 -1.07675385e+00
-1.57121047e-02 -1.10884273e+00 -5.97325921e-01 -2.21498579e-01
-4.80319440e-01 4.94768471e-01 6.52504444e-01 -1.04049420e+00
6.16783977e-01 -1.84547937e+00 3.48545045e-01 2.98012704e-01
4.68042910e-01 1.09717764e-01 -1.15999199e-01 -1.94958374e-02
-3.86944622e-01 2.72835314e-01 -8.37570488e-01 -3.97359371e-01
-5.01386106e-01 3.97330016e-01 -1.52026623e-01 2.88260162e-01
2.28151515e-01 1.42151618e+00 -7.84190297e-01 -5.75768352e-01
2.73097605e-01 7.55793929e-01 -8.05127382e-01 5.14992774e-01
-2.71651804e-01 1.29666197e+00 -3.64215821e-01 2.30153501e-01
5.54033637e-01 -3.32716644e-01 1.32598564e-01 -1.53258756e-01
5.44597447e-01 -3.24935317e-01 -6.65299714e-01 1.74521434e+00
-4.67008799e-01 4.65527326e-01 -6.35291934e-02 -9.53898787e-01
7.41653740e-01 5.21349251e-01 6.89031661e-01 -3.91241908e-01
3.45917761e-01 1.66870609e-01 1.04952917e-01 -1.13264501e-01
1.21638089e-01 -5.32153189e-01 -6.09831251e-02 5.85731328e-01
2.74011821e-01 -5.25037527e-01 -1.51715189e-01 2.25951105e-01
9.89404500e-01 7.94308558e-02 -7.53890425e-02 -2.34885156e-01
3.58362466e-01 1.15466788e-01 3.90150607e-01 5.73896110e-01
-2.41823420e-02 1.23265684e+00 5.23639679e-01 -6.20977581e-01
-1.79898047e+00 -1.26562929e+00 -3.22759897e-02 1.52138278e-01
-5.11025310e-01 2.74086267e-01 -1.48796928e+00 -8.12452793e-01
-5.15597939e-01 8.74256909e-01 -1.07021654e+00 -6.91785514e-02
-7.77524650e-01 -7.28915751e-01 6.97590590e-01 8.18236113e-01
3.90761435e-01 -1.56754124e+00 -7.14482129e-01 3.98684055e-01
-3.12543482e-01 -1.04249012e+00 -4.35932070e-01 -5.83243854e-02
-1.08905399e+00 -7.77943730e-01 -1.35123503e+00 -7.51971006e-01
9.93485034e-01 -8.01211357e-01 1.35052454e+00 -1.32471994e-02
-3.99293333e-01 1.84581071e-01 -2.53775030e-01 -2.63314873e-01
-1.22526824e+00 -6.97433054e-02 -2.53809720e-01 5.24687991e-02
-4.40980136e-01 -8.95487905e-01 -9.54313159e-01 4.20537069e-02
-1.26109552e+00 3.34954441e-01 4.19521362e-01 1.05920160e+00
1.03049076e+00 -4.22816932e-01 4.87045974e-01 -1.11566353e+00
4.39467520e-01 -5.47888756e-01 -4.46853667e-01 5.24525903e-02
-4.17605758e-01 1.50688201e-01 6.88147783e-01 -5.72401047e-01
-8.25249672e-01 1.08471118e-01 -4.40485954e-01 -7.98440456e-01
-1.15615569e-01 1.08280927e-01 2.25097001e-01 3.42060328e-02
9.58033860e-01 5.13472319e-01 2.19848707e-01 1.65864974e-02
5.79053342e-01 2.38162711e-01 8.06929290e-01 -5.00350356e-01
7.16776967e-01 6.46659493e-01 5.26432879e-02 -3.21489215e-01
-3.38399857e-01 4.17008191e-01 -6.46539569e-01 -2.81467885e-01
1.54742610e+00 -8.00468087e-01 -1.77910626e-01 2.85792619e-01
-9.66217935e-01 -8.85002613e-01 -8.60051394e-01 3.03346515e-01
-9.63783503e-01 4.92484905e-02 -9.87523139e-01 -4.03956681e-01
-7.56997049e-01 -1.50406206e+00 1.06200564e+00 1.21918388e-01
-1.16702802e-01 -1.07312536e+00 1.46716684e-01 4.78144526e-01
5.89324713e-01 1.15119433e+00 8.26313078e-01 -5.24945498e-01
-5.97594082e-01 -9.19925272e-02 1.40781686e-01 7.03852355e-01
-1.75815374e-01 -3.03030342e-01 -8.46371770e-01 -2.99434155e-01
1.15043394e-01 -3.15541506e-01 5.38931966e-01 9.63992953e-01
1.28055918e+00 -3.56960833e-01 -6.67930394e-02 8.74139428e-01
1.43895853e+00 4.26226795e-01 1.03901887e+00 -2.30879262e-01
8.03797185e-01 2.65873402e-01 -1.30517920e-02 4.17759046e-02
3.26715671e-02 4.99229729e-01 3.81471008e-01 -6.94493532e-01
-4.53951776e-01 -3.13542485e-01 -7.83571377e-02 7.27250636e-01
-1.48243725e-01 -2.37644643e-01 -1.00782669e+00 6.14373863e-01
-1.36287844e+00 -7.69959271e-01 1.08367261e-02 1.86181378e+00
8.23088288e-01 -7.07862899e-02 4.15366180e-02 -1.58824489e-01
7.07787693e-01 -5.47952168e-02 -6.79038882e-01 -4.20879394e-01
1.61758989e-01 7.69953370e-01 4.79903966e-01 4.18380171e-01
-8.68904531e-01 8.03844810e-01 6.29652071e+00 6.22119248e-01
-1.37999701e+00 6.97860241e-01 1.24612129e+00 1.38873622e-01
-4.26427364e-01 -3.26533198e-01 5.59461825e-02 5.05083799e-01
8.62772584e-01 1.37229607e-01 5.15690804e-01 8.45665097e-01
1.64974965e-02 2.54571229e-01 -8.83230865e-01 7.65144169e-01
1.63463593e-01 -1.64453804e+00 2.83094436e-01 2.26595655e-01
1.16725278e+00 1.91714836e-03 1.17337547e-01 1.60206482e-01
5.14006019e-01 -1.50588536e+00 6.98834598e-01 6.19184256e-01
1.42180240e+00 -8.08623850e-01 8.36971343e-01 2.84255683e-01
-5.93348920e-01 4.13919330e-01 -1.37457147e-01 7.64276087e-01
3.40268701e-01 3.63421530e-01 -1.13045156e+00 3.69338632e-01
3.61241937e-01 5.54246642e-03 -2.96996087e-01 4.92653638e-01
-2.55706877e-01 6.62164450e-01 -1.25351712e-01 4.56156045e-01
3.43883663e-01 -2.93918699e-01 4.75627333e-01 8.07300568e-01
2.61980355e-01 1.71843544e-01 5.01937419e-03 1.55009365e+00
-3.06077480e-01 1.01265855e-01 -5.54318607e-01 2.11586393e-02
5.78945652e-02 1.30668616e+00 -9.94259655e-01 -6.22047007e-01
2.81918377e-01 1.21119070e+00 7.43476450e-02 3.26040566e-01
-9.44911599e-01 1.64267361e-01 4.64345925e-02 3.45412672e-01
2.60143608e-01 1.33888423e-01 -3.54317844e-01 -1.05107594e+00
-2.54670471e-01 -9.03233647e-01 1.91480756e-01 -1.03032863e+00
-1.22621846e+00 1.13598883e+00 -2.57950068e-01 -1.07280517e+00
-6.84362769e-01 -1.49593428e-01 -7.66647160e-01 1.14354348e+00
-8.48286092e-01 -1.38742352e+00 -4.38323319e-01 5.19923151e-01
4.73859817e-01 -2.67241508e-01 1.04036295e+00 1.97070181e-01
-9.77474079e-02 4.74437326e-01 -1.04569025e-01 5.99053979e-01
2.30013609e-01 -1.45788825e+00 5.65004826e-01 7.55739093e-01
1.38256595e-01 2.86697865e-01 7.08042085e-01 -7.54954219e-01
-5.47754705e-01 -1.44690835e+00 1.64076552e-01 -5.14596343e-01
3.74679640e-02 -1.74340889e-01 -8.45654011e-01 8.95225286e-01
3.43221724e-01 3.83470863e-01 5.78470826e-01 -7.80620396e-01
8.74253139e-02 2.57046670e-01 -1.69711208e+00 4.84843016e-01
5.23865938e-01 -3.26248705e-01 -4.35144961e-01 3.79042119e-01
6.11342311e-01 -8.52620721e-01 -1.29460335e+00 3.37982178e-01
3.09687197e-01 -7.94824719e-01 1.13407671e+00 -5.66037953e-01
1.11072803e+00 -1.98497206e-01 1.17873445e-01 -1.60502303e+00
2.41776332e-02 -3.08425963e-01 2.52290845e-01 7.68519521e-01
3.40496212e-01 -3.56770068e-01 9.05843377e-01 4.98424679e-01
-1.53498605e-01 -7.95758665e-01 -8.99854064e-01 -3.90282571e-01
5.70530474e-01 -2.75921941e-01 6.14750683e-01 1.12638915e+00
-8.12598884e-01 1.24397837e-01 -3.33723217e-01 -5.86456098e-02
7.98356593e-01 -1.13145605e-01 5.59543133e-01 -7.67599761e-01
-5.19509077e-01 -1.42583162e-01 -3.78951162e-01 -5.08067131e-01
6.32849336e-02 -1.16515660e+00 -1.69784069e-01 -1.45457685e+00
2.16982394e-01 -4.89290625e-01 -1.04871251e-01 3.26723605e-01
2.21917611e-02 7.94617176e-01 1.71537608e-01 1.33412734e-01
7.05981404e-02 3.99791807e-01 1.85767889e+00 -2.09599227e-01
-1.42081723e-01 -2.69682258e-01 -4.58237946e-01 4.85143065e-01
6.84809744e-01 -7.01293766e-01 -3.54982883e-01 -2.91384906e-01
-2.12393895e-01 7.08730936e-01 6.69275582e-01 -1.07815981e+00
-2.46407151e-01 3.11839014e-01 8.57328117e-01 -2.42361307e-01
2.04512000e-01 -5.28160691e-01 7.04028070e-01 6.13341510e-01
-3.77717167e-01 -1.23025607e-02 1.03388160e-01 3.24968189e-01
-1.16949365e-01 -3.94195914e-02 1.13731754e+00 -4.92264837e-01
4.01437357e-02 4.67018038e-01 -2.66130567e-01 2.75747538e-01
1.12604940e+00 -2.40730476e-02 2.67629325e-01 -6.62078619e-01
-1.23413384e+00 -1.63044930e-01 6.25607431e-01 5.80839030e-02
5.58590412e-01 -1.39602494e+00 -9.47353780e-01 2.50933141e-01
-2.37996340e-01 3.48306596e-01 5.34138262e-01 6.20792508e-01
-1.08515775e+00 3.96458842e-02 -6.87323451e-01 -8.47431064e-01
-7.17221618e-01 5.43765426e-01 8.80988896e-01 -6.88099205e-01
-8.13142419e-01 6.51789069e-01 4.35189426e-01 -4.92897362e-01
-3.42511356e-01 -2.01629192e-01 -8.79816487e-02 -3.36387783e-01
2.22646132e-01 -2.06963345e-01 -9.37941819e-02 -6.44045472e-01
-3.16034853e-02 1.88996390e-01 2.84803271e-01 -4.25586551e-01
1.54495394e+00 4.70098615e-01 1.42312735e-01 2.53134053e-02
1.02043223e+00 6.90660179e-02 -1.51941788e+00 1.94960833e-01
-6.52658761e-01 -1.32878264e-02 1.19591340e-01 -9.45277512e-01
-1.80323708e+00 8.37353468e-01 9.70768988e-01 -1.88270472e-02
1.02988434e+00 3.73426569e-03 9.72316623e-01 -3.81716102e-01
2.73818552e-01 -5.88514507e-01 9.27480459e-02 -1.58287078e-01
1.08795106e+00 -1.01015222e+00 -4.33529586e-01 -2.99353868e-01
-1.09383047e+00 8.98250580e-01 3.13366205e-01 -5.75914621e-01
3.10783625e-01 4.43552643e-01 3.14807296e-01 -3.65273923e-01
-2.61078596e-01 2.32530996e-01 2.55214751e-01 8.34505379e-01
2.09925577e-01 4.25128132e-01 -1.10166021e-01 3.52100432e-01
-3.31039906e-01 -3.65656167e-02 6.35603368e-01 3.36684078e-01
1.97194904e-01 -1.10592639e+00 -3.07475090e-01 4.63667840e-01
-6.72203183e-01 -2.11618185e-01 2.98891924e-02 7.26653337e-01
2.29134217e-01 3.24357808e-01 2.64713347e-01 -7.55442679e-02
-2.07081083e-02 1.80679545e-01 4.76054907e-01 -6.28284395e-01
-8.32303405e-01 6.66714981e-02 -3.52676868e-01 -5.14782012e-01
-8.92126188e-02 -5.28123736e-01 -1.32262993e+00 -8.74657333e-02
2.58024216e-01 6.59244880e-02 7.06705689e-01 7.06171751e-01
9.38308463e-02 1.08800948e+00 4.68871862e-01 -7.85136819e-01
-3.32739770e-01 -9.93611336e-01 -5.13450503e-01 8.23768854e-01
9.07857269e-02 -3.56481731e-01 -8.30349177e-02 3.29333484e-01] | [14.149157524108887, -1.9984126091003418] |
ce5fbff6-7c7d-46f8-b3ec-e60ee6ff9d7c | hard-sample-aware-network-for-contrastive | 2212.08665 | null | https://arxiv.org/abs/2212.08665v3 | https://arxiv.org/pdf/2212.08665v3.pdf | Hard Sample Aware Network for Contrastive Deep Graph Clustering | Contrastive deep graph clustering, which aims to divide nodes into disjoint groups via contrastive mechanisms, is a challenging research spot. Among the recent works, hard sample mining-based algorithms have achieved great attention for their promising performance. However, we find that the existing hard sample mining methods have two problems as follows. 1) In the hardness measurement, the important structural information is overlooked for similarity calculation, degrading the representativeness of the selected hard negative samples. 2) Previous works merely focus on the hard negative sample pairs while neglecting the hard positive sample pairs. Nevertheless, samples within the same cluster but with low similarity should also be carefully learned. To solve the problems, we propose a novel contrastive deep graph clustering method dubbed Hard Sample Aware Network (HSAN) by introducing a comprehensive similarity measure criterion and a general dynamic sample weighing strategy. Concretely, in our algorithm, the similarities between samples are calculated by considering both the attribute embeddings and the structure embeddings, better revealing sample relationships and assisting hardness measurement. Moreover, under the guidance of the carefully collected high-confidence clustering information, our proposed weight modulating function will first recognize the positive and negative samples and then dynamically up-weight the hard sample pairs while down-weighting the easy ones. In this way, our method can mine not only the hard negative samples but also the hard positive sample, thus improving the discriminative capability of the samples further. Extensive experiments and analyses demonstrate the superiority and effectiveness of our proposed method. | ['Cancan Chen', 'Jingcan Duan', 'Liang Li', 'Wenxuan Tu', 'Ke Liang', 'Zhen Wang', 'Xinwang Liu', 'Sihang Zhou', 'Xihong Yang', 'Yue Liu'] | 2022-12-16 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [-7.40746632e-02 -4.22654161e-03 -2.39169165e-01 -3.60618442e-01
-6.15851618e-02 -1.91141680e-01 3.88317287e-01 3.70959193e-01
-2.65436083e-01 3.01761508e-01 -1.03519946e-01 1.99475706e-01
-7.46409178e-01 -1.25534701e+00 -4.43803519e-02 -1.16133809e+00
5.62832244e-02 6.88574135e-01 2.90247977e-01 -1.42956406e-01
1.04030678e-02 1.26519412e-01 -1.35065174e+00 -2.64478296e-01
1.25147808e+00 1.04845154e+00 2.19286859e-01 -1.39060140e-01
-1.95082352e-01 1.87889636e-01 -3.77546877e-01 -4.10009742e-01
2.79165059e-01 -3.47666323e-01 -4.33713585e-01 2.19447851e-01
-6.10413961e-02 -9.90718380e-02 -2.84361482e-01 1.34325135e+00
5.75078309e-01 1.07418142e-01 5.02870917e-01 -1.36115181e+00
-4.48328406e-01 8.36478591e-01 -9.55780864e-01 2.43253224e-02
-3.65287811e-02 1.93270057e-01 1.20681477e+00 -8.24167550e-01
3.35022718e-01 1.14122093e+00 3.29327732e-01 3.71335685e-01
-9.22716856e-01 -8.81936312e-01 3.40985239e-01 6.19253755e-01
-1.81561923e+00 -2.17790172e-01 1.23276651e+00 -1.51500523e-01
1.35790288e-01 3.92059058e-01 8.51947308e-01 6.87456667e-01
-5.14660060e-01 7.79970467e-01 7.93273509e-01 -3.68740922e-03
2.28509054e-01 2.42141590e-01 2.89496362e-01 6.68794394e-01
5.27727902e-01 -2.54410714e-01 -5.04333861e-02 3.08561008e-02
1.59129992e-01 3.40648264e-01 -5.47940493e-01 -5.69676936e-01
-9.77323771e-01 7.31392086e-01 5.01529276e-01 4.73571867e-01
-1.97864383e-01 -4.49848622e-01 5.15874147e-01 7.71981180e-02
3.82417589e-01 5.53205726e-04 -2.74799585e-01 1.81591585e-01
-8.35496545e-01 -6.76772520e-02 4.94979769e-01 8.81309688e-01
1.08168745e+00 -2.81370699e-01 -8.69017467e-02 8.55255604e-01
4.07698154e-01 1.40253708e-01 5.11267126e-01 -3.61506313e-01
6.49238229e-01 1.34484720e+00 -3.49460572e-01 -1.86048961e+00
-3.32312226e-01 -8.32665324e-01 -1.39256728e+00 -3.40844512e-01
2.59990990e-01 8.18380713e-03 -5.28229713e-01 1.64658237e+00
7.29171515e-01 1.03244193e-01 -1.93593323e-01 8.51072490e-01
7.55705357e-01 3.93669009e-01 -5.99801838e-02 -6.41183138e-01
1.28921866e+00 -6.51370227e-01 -6.65487409e-01 -1.67248435e-02
5.94883084e-01 -4.79830682e-01 1.03785884e+00 4.02647376e-01
-7.05530763e-01 -5.43562710e-01 -1.15891469e+00 2.09020063e-01
-2.18213558e-01 1.44403711e-01 6.54445052e-01 5.50507486e-01
-5.68158329e-01 6.23061597e-01 -5.27793169e-01 -1.32956669e-01
6.08492374e-01 3.56716454e-01 -1.30402967e-01 -3.47534508e-01
-1.18310118e+00 4.17002589e-02 5.38954556e-01 3.58335167e-01
-4.52965707e-01 -3.08522344e-01 -7.08374560e-01 3.04605961e-01
8.71808290e-01 -3.77560109e-01 4.16109204e-01 -6.97824121e-01
-9.24657702e-01 5.20777822e-01 -8.52710530e-02 -9.19105932e-02
6.44154727e-01 2.60761023e-01 -6.70064688e-01 2.86538392e-01
5.76131120e-02 4.11278307e-01 7.16250658e-01 -1.23181331e+00
-6.65227592e-01 -6.37014508e-01 -1.38989642e-01 2.59658247e-01
-1.03757083e+00 -6.51992202e-01 -7.18525052e-01 -5.48381448e-01
3.49687159e-01 -5.92571199e-01 -2.12401018e-01 -4.11305651e-02
-5.82695365e-01 -6.79617822e-01 1.00149286e+00 -9.70391259e-02
1.61690974e+00 -2.34128618e+00 2.58486092e-01 7.20376074e-01
9.97383773e-01 2.56160647e-01 -1.92051545e-01 1.52629048e-01
9.12175179e-02 8.53364095e-02 -4.25384432e-01 -1.65289178e-01
3.32942940e-02 1.41945943e-01 2.08176956e-01 4.01287258e-01
1.97436526e-01 5.13844430e-01 -1.19194293e+00 -8.50232065e-01
2.60069609e-01 1.93380788e-01 -2.56174684e-01 7.62480497e-02
1.16242364e-01 -1.14858821e-01 -5.49408078e-01 5.64548194e-01
1.10388434e+00 -1.85075104e-01 5.08758962e-01 -6.75686598e-01
3.44628990e-01 -1.05144374e-01 -1.44403672e+00 1.05423677e+00
-1.46343363e-02 2.97830310e-02 3.69061172e-01 -1.27717388e+00
1.17898333e+00 -2.55365551e-01 5.64437270e-01 -6.79300010e-01
1.69730142e-01 2.86714256e-01 3.18735749e-01 -5.97946763e-01
3.63645434e-01 -4.88578565e-02 2.96364665e-01 3.08427155e-01
-3.23526591e-01 3.12155157e-01 3.29476327e-01 4.78709579e-01
8.26358795e-01 -3.25288475e-01 1.38855204e-01 -3.92015159e-01
7.89994657e-01 -3.93729091e-01 8.91631126e-01 2.07881108e-01
-3.51387322e-01 3.52197438e-01 5.48466921e-01 -1.88770488e-01
-5.46185315e-01 -7.54722774e-01 -1.55222863e-01 6.64549291e-01
7.32920706e-01 -6.84638500e-01 -7.10463643e-01 -9.89720762e-01
-1.86331216e-02 2.06382483e-01 -6.66513026e-01 -6.29207432e-01
-4.81653333e-01 -1.03628421e+00 1.40894428e-01 3.49388063e-01
6.15035474e-01 -7.86406696e-01 7.77465180e-02 1.29555002e-01
-1.87632948e-01 -6.39188647e-01 -5.06130934e-01 7.46011436e-02
-6.91068769e-01 -1.47243500e+00 -4.95810688e-01 -1.01035607e+00
9.11679208e-01 7.22240090e-01 8.12819302e-01 7.02091753e-01
-9.91703048e-02 4.84149717e-03 -5.55864096e-01 7.83616677e-02
4.16479185e-02 2.03033373e-01 1.15324490e-01 3.80400777e-01
7.37062752e-01 -6.81090236e-01 -7.81706512e-01 4.90030676e-01
-9.29584205e-01 -9.23129544e-02 7.53045917e-01 7.92000234e-01
7.24533141e-01 7.85778046e-01 7.78129339e-01 -7.40304828e-01
7.26847351e-01 -7.35594273e-01 -1.93423614e-01 2.54794717e-01
-8.49140346e-01 -1.70081601e-01 9.74187851e-01 -5.15142739e-01
-6.53235614e-01 -2.62814075e-01 -5.88269010e-02 -3.47843975e-01
6.13862835e-02 4.57987458e-01 -8.32879186e-01 1.50980540e-02
1.67070091e-01 4.50929761e-01 1.68334574e-01 -2.92840570e-01
-4.56376672e-02 6.81150377e-01 1.07362628e-01 -4.50230360e-01
1.01913428e+00 5.07313430e-01 2.55214781e-01 -6.16704762e-01
-4.61490512e-01 -3.84868383e-01 -4.59749222e-01 -1.87785044e-01
4.06933993e-01 -5.55247068e-01 -9.74691808e-01 4.21087295e-01
-4.81643587e-01 1.09467566e-01 -1.11183442e-01 3.19261551e-01
2.03240529e-01 1.06325078e+00 -5.81586182e-01 -8.22751045e-01
-4.16180730e-01 -1.15768635e+00 8.61480534e-01 2.52281547e-01
-3.36490348e-02 -8.99866760e-01 -4.24793571e-01 1.79775193e-01
5.77013455e-02 1.44341663e-01 1.07632756e+00 -7.90324628e-01
-4.97805834e-01 -2.82898813e-01 -4.63749230e-01 3.15700144e-01
4.54500049e-01 1.31700724e-01 -5.62882245e-01 -3.00827324e-01
1.11927450e-01 -1.07615657e-01 9.32016373e-01 1.15749594e-02
1.51100671e+00 -2.95345187e-01 -5.45520902e-01 4.41346467e-01
1.30268979e+00 -7.05236662e-03 4.15370554e-01 1.91615716e-01
9.82626259e-01 8.04746330e-01 7.52419949e-01 5.84310651e-01
5.42245865e-01 3.36924255e-01 6.45693839e-01 -2.78752577e-02
3.20581049e-01 -1.15158856e-01 2.39194632e-02 1.15745187e+00
2.97269244e-02 -1.57603055e-01 -4.87514883e-01 5.08841872e-01
-1.80642414e+00 -8.67249787e-01 -4.54390228e-01 2.19261575e+00
9.38980699e-01 4.16731417e-01 1.72459453e-01 6.68549478e-01
1.01031852e+00 1.73319265e-01 -6.22457445e-01 3.96705151e-01
-1.12431377e-01 -4.40298319e-02 3.22845280e-02 6.55667111e-02
-9.05635715e-01 3.87212008e-01 4.14957094e+00 1.46076679e+00
-8.22152257e-01 -2.53116369e-01 7.81722009e-01 -6.41239807e-02
-6.28538251e-01 -1.09357424e-02 -7.40188241e-01 8.94915760e-01
1.21642493e-01 -2.89710343e-01 3.36330026e-01 8.97074997e-01
1.98057324e-01 3.20021287e-02 -9.19482887e-01 9.75655377e-01
6.09477051e-03 -9.50984180e-01 1.30526513e-01 2.03665093e-01
2.83038437e-01 -4.77930665e-01 -1.53850960e-02 4.11727995e-01
-1.23273805e-02 -5.82760274e-01 3.77110600e-01 4.19278175e-01
5.11882782e-01 -1.17066514e+00 8.04802537e-01 4.61546242e-01
-1.54868782e+00 -1.57483697e-01 -5.78679085e-01 1.16554573e-01
-2.35249326e-01 1.32147765e+00 -4.50977147e-01 8.93988788e-01
6.32211924e-01 1.13319910e+00 -8.02670240e-01 9.38637853e-01
2.10902225e-02 5.00384510e-01 -3.53837103e-01 -2.57877380e-01
1.81322962e-01 -7.27510631e-01 3.80198419e-01 9.03080761e-01
2.93712109e-01 6.79248869e-02 4.16208088e-01 7.49477744e-01
-1.83332890e-01 3.47189397e-01 -2.53867775e-01 6.86818063e-02
7.91298985e-01 1.76259649e+00 -9.21778798e-01 -4.55262274e-01
-1.05087459e-01 6.15720272e-01 3.96130621e-01 1.02126971e-01
-6.78166986e-01 -8.22094440e-01 5.33927619e-01 1.65762261e-01
2.05015898e-01 -1.61713600e-01 -2.36090943e-01 -1.16862953e+00
3.29972774e-01 -7.55416036e-01 5.86997926e-01 -1.56135812e-01
-1.52407849e+00 4.41677600e-01 -2.78276891e-01 -1.53428054e+00
3.79897773e-01 -2.50875652e-01 -7.64157176e-01 5.69319844e-01
-1.33650470e+00 -7.04522789e-01 -6.99905097e-01 5.71400583e-01
2.26656541e-01 4.55759611e-04 3.36946428e-01 5.47238767e-01
-9.76524174e-01 8.66087794e-01 1.39560044e-01 2.00325578e-01
5.73232412e-01 -1.13601744e+00 -1.47265092e-01 6.29241824e-01
-3.86214107e-01 8.84821057e-01 3.06136191e-01 -6.80550218e-01
-1.28653097e+00 -1.18702495e+00 3.01530540e-01 1.41828731e-01
5.07234275e-01 -4.44016844e-01 -1.24698246e+00 -8.57015699e-02
-3.90554033e-02 -6.31846339e-02 8.51610243e-01 5.90877719e-02
-2.34548658e-01 -6.35747671e-01 -1.29266644e+00 6.71651304e-01
1.16384864e+00 -1.70957521e-01 -1.83873445e-01 3.83909374e-01
7.48971701e-01 1.82982296e-01 -8.89210820e-01 5.87993205e-01
2.71286964e-01 -1.02063262e+00 8.02423477e-01 2.37921178e-02
3.18144292e-01 -6.20198488e-01 2.21495599e-01 -1.17083621e+00
-5.88906288e-01 -1.19650707e-01 -2.84056664e-01 1.56647480e+00
-4.68798801e-02 -6.99883997e-01 8.86890829e-01 2.20075548e-01
-5.28459325e-02 -9.96788204e-01 -6.92767978e-01 -6.83053851e-01
-3.23041618e-01 -1.78580910e-01 9.54475880e-01 1.25158751e+00
-2.80933417e-02 5.67310929e-01 -2.75104344e-01 3.96301895e-02
8.45139086e-01 3.04830849e-01 6.02227330e-01 -1.68641853e+00
-2.53926545e-01 -7.24550664e-01 -4.05588120e-01 -9.00531352e-01
-1.40734375e-01 -9.32636261e-01 -1.81012735e-01 -1.49215996e+00
3.78503263e-01 -8.58326733e-01 -3.30174863e-01 3.57077926e-01
-5.86866379e-01 2.72741239e-03 -1.40252247e-01 3.47009450e-01
-8.14546108e-01 7.65521586e-01 1.45158374e+00 -2.99760669e-01
-1.40160650e-01 -9.67948660e-02 -9.70860898e-01 5.25750399e-01
7.04129040e-01 -2.67713845e-01 -6.16135478e-01 -1.45130754e-01
3.30691248e-01 -4.26363140e-01 1.07936583e-01 -1.09588742e+00
1.39543831e-01 -2.95829713e-01 5.49804568e-01 -6.50306642e-01
7.11288443e-03 -1.01814628e+00 -1.05672307e-01 6.30065739e-01
-2.20559053e-02 -3.38921696e-01 -2.77451605e-01 1.01713657e+00
-1.18006051e-01 -2.16423482e-01 6.64676905e-01 3.68440077e-02
-6.80561900e-01 7.18259454e-01 7.65448213e-02 6.64441213e-02
1.10998881e+00 -3.34056765e-01 -2.26573586e-01 -4.66601513e-02
-4.84970808e-01 7.83265114e-01 3.92805099e-01 2.73083150e-01
6.99792981e-01 -1.48016417e+00 -5.74515343e-01 2.67789841e-01
2.48412699e-01 5.59366226e-01 4.98043239e-01 9.87756312e-01
-5.46579473e-02 -2.41658241e-01 1.60610095e-01 -5.59835970e-01
-1.23426747e+00 7.32591510e-01 4.96086963e-02 -1.89588949e-01
-4.28133339e-01 6.41206622e-01 2.57105827e-01 -3.00397009e-01
2.68125325e-01 -5.14426008e-02 -5.89604080e-01 5.56343138e-01
5.10740995e-01 5.24568617e-01 -1.03800498e-01 -3.96186173e-01
-5.14275134e-01 4.63338673e-01 -3.18763018e-01 6.25049293e-01
1.38784814e+00 -2.07636669e-01 -4.36047107e-01 1.47785321e-01
1.33703709e+00 1.76726848e-01 -8.08332682e-01 -3.33446145e-01
-7.86569715e-02 -5.71467698e-01 -1.79068968e-01 -1.36732265e-01
-1.55062354e+00 9.96105075e-01 4.45501626e-01 6.69985712e-01
1.43185318e+00 -3.85240838e-02 1.00204754e+00 2.79055506e-01
2.14457348e-01 -1.24651039e+00 1.69348523e-01 3.86510827e-02
2.82288641e-01 -1.22460270e+00 1.22400425e-01 -7.13710368e-01
-2.84276009e-01 1.17526603e+00 9.19478416e-01 -3.82739007e-02
6.67961061e-01 -5.14458641e-02 -5.17607808e-01 -2.80292839e-01
-3.35812151e-01 -1.99893191e-01 2.05111414e-01 5.76047182e-01
8.30222666e-02 2.36505464e-01 -6.88809037e-01 9.20552552e-01
1.72300451e-02 -2.83155322e-01 3.12348694e-01 4.43168283e-01
-5.74651718e-01 -1.06948614e+00 -8.69375765e-02 8.83612275e-01
9.23564956e-02 -1.24578523e-02 -5.00672817e-01 6.20080471e-01
3.10812384e-01 9.66682673e-01 7.70261362e-02 -9.41095471e-01
1.53093085e-01 -3.77554268e-01 2.65430193e-02 -4.44287658e-01
-2.89735109e-01 2.81729475e-02 -1.85296565e-01 -1.37564391e-01
-3.78654599e-01 -4.11428303e-01 -1.36575603e+00 -4.07552809e-01
-6.07675374e-01 4.02700037e-01 6.81692213e-02 9.39089596e-01
3.91943216e-01 4.21790779e-01 1.06848466e+00 -5.74694812e-01
-5.89354277e-01 -8.21924508e-01 -7.95393527e-01 5.53014100e-01
2.28649378e-02 -6.63483381e-01 -6.65336967e-01 -6.26920760e-01] | [7.464078903198242, 5.9523210525512695] |
b50d7fe3-76af-4415-a1f2-d19470d34aa9 | ji-yu-self-attentionde-ju-fa-gan-zhi-yi-yu | null | null | https://aclanthology.org/2020.ccl-1.57 | https://aclanthology.org/2020.ccl-1.57.pdf | 基于Self-Attention的句法感知汉语框架语义角色标注(Syntax-Aware Chinese Frame Semantic Role Labeling Based on Self-Attention) | 框架语义角色标注(Frame Semantic Role Labeling, FSRL)是基于FrameNet标注体系的语义分析任务。语义角色标注通常对句法有很强的依赖性,目前的语义角色标注模型大多基于双向长短时记忆网络Bi-LSTM,虽然可以获取句子中的长距离依赖信息,但无法很好获取句子中的句法信息。因此,引入self-attention机制来捕获句子中每个词的句法信息。实验结果表明,该模型在CFN(Chinese FrameNet,汉语框架网)数据集上的F1达到83.77%,提升了近11%。 | ['Xiaoqi Han', 'Qinghua Chai', 'Zhiqiang Wang', 'Ru Li', 'Xiaohui Wang'] | null | null | null | null | ccl-2020-10 | ['semantic-role-labeling'] | ['natural-language-processing'] | [-2.38124922e-01 -3.40215713e-01 1.08119205e-01 6.03859685e-02
-2.45346259e-02 -7.59229064e-01 3.56923521e-01 7.16018021e-01
-4.40208763e-01 1.12891376e+00 1.20446718e+00 9.86844748e-02
4.18850314e-03 -7.94665635e-01 -6.09756172e-01 -1.02100623e+00
-3.29883546e-01 9.70905066e-01 6.72467649e-01 -1.09164071e+00
3.69346380e-01 3.81807685e-01 -1.21508801e+00 7.65924037e-01
4.26653057e-01 9.33097839e-01 1.02135515e+00 5.53846002e-01
-5.44839911e-02 1.28417945e+00 -1.32688093e+00 5.11614799e-01
-8.98659050e-01 -3.99774432e-01 -1.09930193e+00 4.79902476e-01
-1.70992419e-01 -6.57459319e-01 -1.34525096e+00 1.29825199e+00
3.13688278e-01 5.40865242e-01 5.31179130e-01 -4.90441352e-01
-1.31925738e+00 9.91473496e-01 4.35279876e-01 1.33169746e+00
2.55830079e-01 -1.54154450e-01 6.01120651e-01 -4.30440098e-01
5.16877532e-01 2.23817062e+00 1.35963047e+00 7.72088826e-01
-1.31340301e+00 -7.64893651e-01 1.01754403e+00 2.05010116e-01
-1.16167140e+00 -1.69105828e-02 4.61804479e-01 -2.38151550e-02
1.39354384e+00 1.13442928e-01 1.24170446e+00 1.93573618e+00
4.63988394e-01 1.38251054e+00 1.25972676e+00 -7.07883760e-02
4.37790193e-02 4.55555737e-01 -1.98843300e-01 -2.34282762e-01
5.55348277e-01 2.50326812e-01 -3.06098878e-01 7.99017251e-01
8.69130015e-01 -1.14738487e-01 -3.60016562e-02 1.16538763e+00
-1.71271479e+00 1.32056034e+00 4.90637034e-01 1.59834814e+00
-4.26644459e-02 -2.59703666e-01 -1.03489853e-01 3.28382969e-01
6.10120833e-01 -3.94291520e-01 2.72575170e-01 2.59685099e-01
3.70735824e-01 1.48480430e-01 4.69293267e-01 1.15152562e+00
8.50900054e-01 9.46954787e-01 -8.15200388e-01 5.54957688e-01
1.59911430e+00 5.90610564e-01 1.40944588e+00 -1.01524520e+00
-9.43927348e-01 2.30877370e-01 -4.44179177e-01 -1.47574127e+00
-9.01778519e-01 -1.00767660e+00 -4.84162956e-01 -1.23803139e+00
-9.78191048e-02 -5.51175714e-01 -1.16170943e+00 1.57466650e+00
1.07688397e-01 -2.10803360e-01 5.61349452e-01 8.75483274e-01
1.02212262e+00 1.22837520e+00 6.22784257e-01 -3.85890156e-01
1.33825290e+00 -6.80243492e-01 -1.07666552e+00 -5.10132551e-01
-3.97092611e-01 -1.21235669e+00 1.50977957e+00 -2.26794019e-01
-9.11743820e-01 -7.12597072e-01 -1.43474162e+00 -2.04534411e-01
-6.99548498e-02 -6.46665633e-01 9.02956545e-01 8.88305843e-01
-1.39185107e+00 3.66575778e-01 -1.11680365e+00 -1.18892980e+00
2.18205675e-01 5.89599133e-01 2.94070423e-01 -1.73550716e-03
-2.09349465e+00 1.06253982e+00 4.71594155e-01 7.94183433e-01
-1.75532806e+00 -5.66539109e-01 -1.23893678e+00 2.49238238e-01
8.49351525e-01 -1.54080600e-01 2.11091614e+00 -1.63490033e+00
-8.51285398e-01 7.10833132e-01 -8.46843243e-01 -4.76296127e-01
-8.92099440e-02 3.98216816e-03 -6.59359872e-01 2.97120184e-01
3.33191007e-01 5.28124690e-01 8.10529649e-01 -1.80997348e+00
-9.22687292e-01 -7.32272208e-01 6.40266836e-01 6.03070818e-02
5.47630429e-01 1.04050088e+00 3.45092684e-01 -1.07943308e+00
-3.33760530e-01 -1.60988212e+00 8.04614276e-02 -1.27052498e+00
-5.56111746e-02 -2.17924267e-01 1.01490808e+00 -1.11964792e-01
1.31329918e+00 -2.23149729e+00 -4.67265934e-01 3.34134310e-01
1.90983832e-01 9.23574865e-01 1.76372007e-01 1.25762010e+00
3.02741021e-01 1.04349339e+00 -9.09529030e-01 6.33582532e-01
-3.98110360e-01 5.38827598e-01 1.02775073e+00 -6.60238042e-02
-1.17868528e-01 1.04381397e-01 -9.84916747e-01 3.90840508e-02
5.79217911e-01 4.90938544e-01 4.37506407e-01 1.50342166e-01
5.26005402e-02 3.39388758e-01 5.53809479e-02 5.03422260e-01
7.63948500e-01 -6.36103392e-01 7.64121413e-01 7.33071119e-02
-3.63963217e-01 1.26437083e-01 -2.09960556e+00 2.04684162e+00
1.16589837e-01 9.63177562e-01 1.50535256e-01 -7.94202626e-01
6.99486911e-01 1.43411291e+00 8.02181005e-01 -3.56716067e-01
6.22562587e-01 4.52568293e-01 8.43487680e-01 -1.75649524e-01
8.74725342e-01 -1.12768888e+00 7.22601533e-01 2.35361114e-01
7.56948888e-02 7.82387733e-01 6.15648985e-01 -1.09080112e+00
1.06416154e+00 -4.54059035e-01 -9.92120579e-02 -1.18159187e+00
8.83384883e-01 -1.77948222e-01 6.46861255e-01 4.67749164e-02
-7.68515408e-01 3.84897768e-01 -4.49533254e-01 -2.01851293e-01
-8.15092772e-02 -9.17681575e-01 -4.09217328e-02 1.72099757e+00
1.84419349e-01 -8.29379082e-01 -8.99402022e-01 -7.65884995e-01
-6.69719040e-01 8.96822929e-01 -1.89810827e-01 -4.05787528e-01
-1.26370835e+00 -5.97259283e-01 -1.85110316e-01 3.80489796e-01
1.53921926e+00 -1.86781836e+00 -6.15375340e-01 1.74833074e-01
-3.81401539e-01 -6.82556987e-01 -6.28735662e-01 -3.82063210e-01
-1.27993321e+00 -7.96244860e-01 1.34235397e-01 -1.68337882e+00
9.54563797e-01 3.29388469e-01 8.87685478e-01 4.18289185e-01
6.29123390e-01 2.27617502e-01 -6.30370677e-01 -1.05739939e+00
-1.09123461e-01 3.36461782e-01 6.27952367e-02 1.01761830e+00
7.29016423e-01 -1.82856902e-01 -1.03900373e+00 2.14087725e-01
-1.29243302e+00 1.50512069e-01 -2.93435752e-01 3.04137677e-01
9.57640707e-01 8.92424062e-02 6.20459206e-02 -1.08270395e+00
1.06068158e+00 1.09639309e-01 -6.12790763e-01 3.32371563e-01
-7.44064510e-01 -7.92933226e-01 9.14726079e-01 -5.68251193e-01
-1.63604510e+00 -1.03441703e+00 -5.34093916e-01 -3.67905140e-01
-2.09029093e-01 2.79468656e-01 -1.78234681e-01 6.72091305e-01
3.11168060e-02 -2.28263624e-02 2.16913462e-01 -5.71320415e-01
-1.63844332e-01 2.94124901e-01 2.96046525e-01 -9.22658384e-01
5.89469597e-02 2.49682248e-01 -3.81613076e-01 -5.49012661e-01
-5.43489158e-01 2.55195200e-02 -6.83973730e-01 4.66193408e-01
1.36795664e+00 -1.59059262e+00 -2.25802846e-02 5.90353608e-01
-9.15214479e-01 3.94051582e-01 -6.61741197e-01 1.54442644e+00
-3.55532438e-01 -5.68326533e-01 -8.93046141e-01 -5.82970858e-01
4.19232100e-02 -2.01468229e+00 1.70667440e-01 1.86558652e+00
1.07301466e-01 -1.18780875e+00 -9.30241048e-02 2.76690125e-01
7.23226190e-01 -9.48429853e-02 8.31702709e-01 -6.67204678e-01
-3.14929038e-01 3.93955439e-01 2.48845160e-01 1.24145520e+00
9.30771708e-01 -3.81125323e-02 -2.48822927e-01 -1.93682283e-01
1.27582297e-01 4.32356179e-01 5.57532191e-01 1.13701630e+00
3.94372910e-01 -3.45672637e-01 3.13392550e-01 5.99144876e-01
1.71013272e+00 1.64041531e+00 1.08767200e+00 1.77752388e+00
1.41419068e-01 -2.97667906e-02 3.44364703e-01 6.04269095e-02
3.39221358e-01 8.06288719e-01 -1.60491139e-01 8.67246330e-01
-1.20668805e+00 -1.10881560e-01 9.10279870e-01 1.06093788e+00
-3.20114911e-01 -1.23034783e-01 -1.03715891e-02 5.30094326e-01
-2.92257214e+00 -1.88150370e+00 1.47209436e-01 2.10545325e+00
2.12883562e-01 1.40234366e-01 -2.96822637e-01 1.16062731e-01
1.81546187e+00 3.68264377e-01 2.78816652e-03 -4.70337719e-01
1.83443412e-01 1.12860072e+00 3.42775434e-01 5.67961812e-01
-1.08848500e+00 1.16735065e+00 2.09383106e+00 8.08930397e-01
-2.04346347e+00 3.09290618e-01 5.50418794e-01 -7.71020830e-01
-6.41209781e-01 6.14669263e-01 -6.78478658e-01 1.80933952e+00
1.37458754e+00 -1.02239954e+00 2.84485519e-01 5.33163510e-02
8.82105887e-01 3.21494639e-01 -5.40387630e-01 1.46611845e+00
1.83384225e-01 -2.02788210e+00 -1.08705178e-01 -2.43549407e-01
5.00578046e-01 -4.05454934e-01 -6.94676578e-01 3.42704654e-01
3.91625494e-01 -9.57306087e-01 1.48883235e+00 1.40939420e-02
5.08002043e-01 -1.17064023e+00 1.53252363e+00 -4.58819211e-01
-1.20114887e+00 -6.43379509e-01 -1.47732913e-01 -7.07185328e-01
9.90517318e-01 -1.64032757e-01 2.62282670e-01 4.28676128e-01
8.04979429e-02 6.17970347e-01 -5.68861783e-01 1.64312422e+00
-1.23836374e+00 2.75643259e-01 2.44243130e-01 -1.08874214e+00
6.58908129e-01 1.62903354e-01 6.24655366e-01 1.38988841e+00
-5.35804108e-02 8.16236317e-01 3.36615562e-01 6.26791120e-01
-4.52557981e-01 -1.86291873e-01 3.64755183e-01 5.07914603e-01
5.83580971e-01 7.49790370e-01 -1.79938745e+00 -7.31770515e-01
-1.30071387e-01 -5.73350728e-01 -4.01111722e-01 -2.15128288e-01
-1.84474841e-01 -1.92132071e-01 2.45173827e-01 -1.27686290e-02
-3.14812094e-01 -6.66259229e-01 6.95431173e-01 -5.72076380e-01
-5.56451082e-01 -1.02844274e+00 3.33260030e-01 -7.85059810e-01
-8.57964218e-01 5.84167182e-01 5.72820961e-01 -6.53468192e-01
4.02349919e-01 -1.32921708e+00 3.04840863e-01 3.73458564e-01
-3.44181150e-01 -1.20624018e+00 1.34057775e-01 1.96651414e-01
1.58084857e+00 3.69329154e-01 8.57815504e-01 8.20396185e-01
-8.45341682e-01 1.49615288e-01 6.38319373e-01 5.94196618e-01
6.89705238e-02 -1.99888968e+00 -5.81196427e-01 7.08267808e-01
8.47370148e-01 1.04658753e-01 -1.19651832e-01 -6.86439991e-01
-8.27398896e-01 -6.47301614e-01 8.41602623e-01 4.77734417e-01
3.12113762e-01 -5.92010736e-01 -2.31920466e-01 1.56777465e+00
7.52196670e-01 4.26306456e-01 4.78870243e-01 -5.06505191e-01
9.42766309e-01 -8.53615046e-01 -1.59727013e+00 1.05674481e+00
1.30185330e+00 1.12018220e-01 -9.40045476e-01 1.20156117e-01
1.16857743e+00 -1.32140875e+00 -1.35083246e+00 3.09679598e-01
1.21628046e+00 -6.11023128e-01 5.37809320e-02 -7.14752913e-01
7.72728026e-02 -3.94759983e-01 -2.85393775e-01 -1.24542701e+00
-5.40558279e-01 -9.61547554e-01 7.30378985e-01 1.13749921e+00
1.50716141e-01 -6.80765867e-01 7.18070507e-01 1.16659510e+00
-1.13043082e+00 -4.80671637e-02 -1.89792812e+00 -5.21228492e-01
-7.97401816e-02 -4.95122612e-01 3.37103039e-01 -1.33579910e-01
-7.26432800e-01 7.93218553e-01 1.60596609e-01 -4.07869190e-01
-5.17345488e-01 -1.08082628e+00 -5.96643090e-01 -1.09553826e+00
9.90130424e-01 -3.78261924e-01 -3.53746533e-01 2.65774727e-02
-2.19272450e-01 -1.06578343e-01 -1.45603883e+00 -1.94672585e+00
-1.56562567e-01 -8.48934203e-02 -7.28379428e-01 -4.77795452e-02
8.45724404e-01 1.14892043e-01 1.05327082e+00 6.97494805e-01
-3.14987928e-01 -1.69073015e-01 1.65609419e+00 -1.70066565e-01
-4.17467773e-01 3.64882387e-02 -8.37134182e-01 1.15992403e+00
3.88270110e-01 -2.10610032e-03 -5.94893396e-01 -8.73733580e-01
-4.20352891e-02 -1.14287376e+00 -1.09086382e+00 -9.52945650e-01
-3.68677638e-02 -1.08846724e-01 6.34949028e-01 8.81268144e-01
-1.79949656e-01 -4.77430314e-01 3.92747521e-01 4.97899652e-01
-1.38401970e-01 8.66191924e-01 4.45282966e-01 -4.04345036e-01
-1.78448975e-01 -1.15673065e+00 7.89164841e-01 -9.93141353e-01
-1.12981415e+00 -5.81965029e-01 -9.46322083e-01 -2.30668515e-01
1.05985081e+00 8.26826170e-02 -5.16751051e-01 1.97160661e-01
-1.24969828e+00 -4.02296633e-01 -5.65511361e-02 1.10153413e+00
1.07742095e+00 -1.63368189e+00 -1.51797485e+00 1.44752488e-01
-7.40053952e-01 6.14600480e-01 4.74547029e-01 5.78881383e-01
-7.80378938e-01 6.08744383e-01 -7.91266561e-01 -4.10759985e-01
-1.66623962e+00 7.20561564e-01 -3.04916620e-01 -5.77666499e-02
-4.90055859e-01 8.03602099e-01 -2.85993934e-01 7.37229466e-01
1.70512825e-01 -7.04281449e-01 -1.02497292e+00 4.10231739e-01
7.53736138e-01 1.86698198e-01 6.43960685e-02 -1.67831028e+00
-6.67304695e-01 1.13190830e+00 -5.50586104e-01 -2.58125901e-01
5.75885594e-01 -2.92914301e-01 -5.36655247e-01 8.44926611e-02
7.52212703e-01 9.27637145e-02 -7.43621111e-01 6.75185978e-01
-4.24081922e-01 -2.03259796e-01 -5.38896263e-01 -1.62945819e+00
-8.65279496e-01 5.32044172e-01 6.70237005e-01 -4.98324364e-01
8.64035070e-01 -4.21832830e-01 1.13913405e+00 -1.44740865e-01
1.57895136e+00 -1.84542847e+00 -7.85787404e-01 1.29625189e+00
9.98464406e-01 -1.59849906e+00 -8.66403580e-02 -2.67206401e-01
2.26826772e-01 1.53228390e+00 1.07551014e+00 -5.09613514e-01
8.90874147e-01 -3.87072951e-01 -5.09426534e-01 -7.83228502e-02
-2.64335752e-01 -9.62350428e-01 3.98538969e-02 9.27274108e-01
9.67094481e-01 -1.51472107e-01 -1.33953726e+00 5.33962190e-01
-8.19521010e-01 3.83678913e-01 8.71676922e-01 1.36210859e+00
-6.89574242e-01 -6.38314486e-01 -1.31939280e+00 1.86109602e-01
-1.34405196e+00 6.47669137e-01 5.72145760e-01 9.55420136e-01
1.03273058e+00 2.42764258e+00 -2.66500246e-02 -9.31789994e-01
1.56334651e+00 -1.75796553e-01 8.72670293e-01 -8.22538555e-01
-1.39908695e+00 -3.96966487e-02 4.12504077e-01 1.83119953e-01
-1.11731744e+00 -4.29265946e-02 -1.32318175e+00 -8.28977227e-01
-1.78447384e-02 1.56023419e+00 4.70662624e-01 1.31093335e+00
-2.14616567e-01 1.33485091e+00 1.81282774e-01 -9.54784095e-01
-2.11714860e-02 -1.32347679e+00 -4.71128464e-01 -5.86974509e-02
5.27367353e-01 -1.21308434e+00 -9.61311936e-01 -2.51653612e-01] | [-3.315978765487671, 6.9076313972473145] |
6c342397-acbf-44fd-92b0-054041cbf745 | survival-text-regression-for-time-to-event | null | null | https://aclanthology.org/2021.findings-acl.104 | https://aclanthology.org/2021.findings-acl.104.pdf | Survival text regression for time-to-event prediction in conversations | null | ['Andreas Vlachos', 'Christine de Kock'] | null | null | null | null | findings-acl-2021-8 | ['time-to-event-prediction'] | ['time-series'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.2067484855651855, 3.815051555633545] |
de6c7c34-97a0-4b27-af69-95be76794487 | interpretable-embeddings-from-molecular | 1912.12175 | null | https://arxiv.org/abs/1912.12175v1 | https://arxiv.org/pdf/1912.12175v1.pdf | Interpretable Embeddings From Molecular Simulations Using Gaussian Mixture Variational Autoencoders | Extracting insight from the enormous quantity of data generated from molecular simulations requires the identification of a small number of collective variables whose corresponding low-dimensional free-energy landscape retains the essential features of the underlying system. Data-driven techniques provide a systematic route to constructing this landscape, without the need for extensive a priori intuition into the relevant driving forces. In particular, autoencoders are powerful tools for dimensionality reduction, as they naturally force an information bottleneck and, thereby, a low-dimensional embedding of the essential features. While variational autoencoders ensure continuity of the embedding by assuming a unimodal Gaussian prior, this is at odds with the multi-basin free-energy landscapes that typically arise from the identification of meaningful collective variables. In this work, we incorporate this physical intuition into the prior by employing a Gaussian mixture variational autoencoder (GMVAE), which encourages the separation of metastable states within the embedding. The GMVAE performs dimensionality reduction and clustering within a single unified framework, and is capable of identifying the inherent dimensionality of the input data, in terms of the number of Gaussians required to categorize the data. We illustrate our approach on two toy models, alanine dipeptide, and a challenging disordered peptide ensemble, demonstrating the enhanced clustering effect of the GMVAE prior compared to standard VAEs. The resulting embeddings appear to be promising representations for constructing Markov state models, highlighting the transferability of the dimensionality reduction from static equilibrium properties to dynamics. | ['Yasemin Bozkurt Varolgunes', 'Tristan Bereau', 'Joseph F. Rudzinski'] | 2019-12-22 | null | null | null | null | ['physical-intuition'] | ['reasoning'] | [ 1.16790406e-01 5.44921122e-02 1.28642526e-02 -5.87388799e-02
-3.79461974e-01 -7.45039046e-01 9.32100475e-01 3.34988266e-01
-3.76260430e-01 6.94058597e-01 2.97160298e-01 -3.80556047e-01
-4.10582483e-01 -7.58542418e-01 -5.85638821e-01 -1.43417943e+00
-1.62795395e-01 7.20535219e-01 -2.10771173e-01 -2.87548393e-01
7.76450038e-02 7.29373991e-01 -1.43813956e+00 -9.73039567e-02
7.90048301e-01 5.70088863e-01 2.33507887e-01 6.54376328e-01
-9.47303548e-02 2.62855202e-01 -2.61193424e-01 3.47764939e-02
-1.66022539e-01 -5.91119051e-01 -5.92014134e-01 -7.56927580e-02
-1.09233044e-01 3.32128704e-02 -3.90804470e-01 9.20747101e-01
3.55260044e-01 4.71215725e-01 1.28197443e+00 -8.96376491e-01
-5.92880487e-01 1.42307594e-01 6.75375983e-02 1.49542153e-01
-1.51328325e-01 3.49853337e-01 1.14075768e+00 -5.92340648e-01
1.00467706e+00 1.12117493e+00 4.93778408e-01 4.95562047e-01
-1.91681707e+00 1.12618744e-01 -1.34297684e-01 -1.18870258e-01
-1.24168909e+00 -3.58316511e-01 1.00365365e+00 -9.45112348e-01
1.21868372e+00 1.05654053e-01 7.07910538e-01 1.21388006e+00
5.42652607e-01 1.87115550e-01 1.00264204e+00 -4.56230283e-01
7.80214012e-01 1.38253346e-01 3.35838616e-01 3.78494620e-01
1.44117206e-01 3.82452488e-01 -2.07339302e-01 -5.89646816e-01
6.89605832e-01 1.66205019e-01 -1.75343946e-01 -1.04650927e+00
-1.03145945e+00 1.09657323e+00 1.67482466e-01 3.38252455e-01
-6.09385073e-01 4.21479046e-02 4.05771464e-01 8.68636146e-02
2.32375622e-01 5.54209530e-01 -4.70466971e-01 -2.09724009e-01
-7.22091258e-01 3.04118156e-01 6.82931185e-01 2.63439655e-01
9.45227563e-01 1.24857360e-02 4.41555053e-01 2.55312264e-01
5.45490265e-01 3.58590931e-01 2.80580491e-01 -1.08641791e+00
-6.76916167e-02 7.13391304e-01 2.37882152e-01 -8.70212078e-01
-2.46703833e-01 -7.35781863e-02 -9.75089014e-01 3.69450748e-01
2.88617849e-01 -1.20067924e-01 -8.34427357e-01 2.01537752e+00
4.88693774e-01 -6.14002466e-01 3.31767857e-01 7.23957121e-01
1.71150476e-01 8.82835329e-01 1.69468448e-01 -3.13715905e-01
1.20695484e+00 -1.67999506e-01 -6.66956782e-01 1.63359761e-01
4.72362965e-01 -1.48063600e-01 8.41937363e-01 1.35600939e-01
-8.31220984e-01 -4.66149896e-01 -1.03944802e+00 -4.00880864e-03
-5.81265628e-01 -3.93336177e-01 8.36754739e-01 4.91792887e-01
-1.05402601e+00 9.92402613e-01 -1.27497351e+00 -3.32643420e-01
9.31378603e-02 4.79936600e-01 -4.63496029e-01 4.06886071e-01
-1.19245827e+00 8.47700775e-01 5.73193967e-01 -3.86090577e-02
-8.30729425e-01 -4.41561729e-01 -8.78366709e-01 1.11841902e-01
-1.14381775e-01 -7.16424465e-01 7.03512967e-01 -6.79939032e-01
-1.40832806e+00 3.35402519e-01 -2.81590521e-01 -2.23427981e-01
2.94813067e-01 3.99423242e-01 -1.56538025e-01 3.73118341e-01
-3.11596543e-01 7.61168301e-01 9.94821250e-01 -1.15889919e+00
2.77699381e-01 -3.38453978e-01 -1.65874094e-01 3.65615599e-02
-1.05438806e-01 -4.55585450e-01 1.44675195e-01 -4.23240930e-01
7.98552930e-02 -9.54530716e-01 -4.93994147e-01 -2.72619009e-01
-1.26348585e-01 -2.06385657e-01 5.86364090e-01 -4.91237462e-01
1.24785340e+00 -2.09982944e+00 9.72907960e-01 3.20897430e-01
7.13319182e-01 -6.17884547e-02 2.03782365e-01 1.04669702e+00
-3.61726016e-01 1.58307984e-01 -4.45600748e-01 -1.26056597e-01
2.60775954e-01 4.26826358e-01 -3.32663983e-01 6.14096701e-01
5.07375896e-01 8.65249455e-01 -6.73268497e-01 -2.21209824e-01
5.10482490e-01 9.37728763e-01 -7.75929868e-01 5.43566458e-02
-4.49359506e-01 5.37146032e-01 -4.32715833e-01 4.74868789e-02
4.57528114e-01 -3.83391887e-01 5.27830541e-01 -1.53886631e-01
-1.46156758e-01 2.22874761e-01 -9.71949100e-01 1.38416529e+00
1.34753808e-01 6.10154569e-01 7.16102421e-02 -1.11607873e+00
7.07502365e-01 1.23886265e-01 7.34103858e-01 -2.76780099e-01
2.92246908e-01 -1.19719476e-01 3.37950766e-01 -3.07607651e-01
3.98697734e-01 -6.32946968e-01 -1.29371181e-01 6.78466558e-01
3.61930907e-01 1.06014676e-01 1.05653368e-01 4.05750573e-01
8.00530255e-01 8.94790217e-02 2.62177885e-01 -5.93677044e-01
2.39927888e-01 1.42117172e-01 3.29483390e-01 5.68600178e-01
-2.54346728e-01 3.41787368e-01 7.35315859e-01 -5.70623577e-01
-1.61661506e+00 -1.14683557e+00 -4.44025815e-01 5.14328122e-01
-7.96539858e-02 -5.40743172e-01 -9.35205281e-01 -1.88633248e-01
9.18760896e-02 4.50117588e-01 -9.85011220e-01 -4.14137691e-01
-3.60582471e-01 -1.14615369e+00 1.25443742e-01 1.83454528e-01
-4.10878241e-01 -1.07713580e+00 -6.72130585e-01 4.30859268e-01
1.40517075e-02 -7.04462588e-01 -4.30315249e-02 7.75623024e-01
-9.63917255e-01 -1.01010442e+00 -3.49688709e-01 -2.82673866e-01
6.44381702e-01 -2.54556417e-01 9.04076636e-01 -4.27937210e-01
-5.04691839e-01 3.32538038e-01 6.66984618e-02 6.84672892e-02
-9.19294775e-01 -1.28227487e-01 5.34451902e-01 -2.04665482e-01
6.68611825e-01 -9.28394735e-01 -7.04744458e-01 -7.98461884e-02
-1.03948474e+00 -1.99881896e-01 4.27771360e-01 9.42459404e-01
6.12436175e-01 2.38035038e-01 3.04522783e-01 -2.03159198e-01
7.88708568e-01 -6.21724308e-01 -4.56071794e-01 -1.20963797e-01
-5.67208946e-01 7.23772883e-01 7.50484288e-01 -4.64458019e-01
-6.56035960e-01 -5.33072390e-02 -8.93299952e-02 -4.17604387e-01
-5.00259042e-01 4.29192841e-01 -2.18506098e-01 2.38661200e-01
6.44089162e-01 6.17285490e-01 5.71608424e-01 -5.96312821e-01
6.61074102e-01 5.45621753e-01 1.75703913e-01 -7.50930965e-01
3.85447323e-01 6.38105750e-01 1.02416322e-01 -1.13802207e+00
2.95299627e-02 -2.06432179e-01 -1.03974187e+00 1.16196625e-01
1.12940753e+00 -5.37868321e-01 -1.05722129e+00 1.68143272e-01
-1.00382829e+00 -1.95665374e-01 -6.02225363e-01 4.33057606e-01
-7.80166805e-01 6.11510098e-01 -6.51912212e-01 -8.46452296e-01
-2.56100800e-02 -1.42590153e+00 8.89218450e-01 -1.06051452e-01
-5.67182422e-01 -1.30702245e+00 7.02499807e-01 -3.23012583e-02
1.81971297e-01 6.02078557e-01 1.53053904e+00 -5.48849881e-01
-4.20780748e-01 -1.36309758e-01 2.54136384e-01 2.42774233e-01
1.48329377e-01 3.33251864e-01 -8.83149147e-01 -4.15027618e-01
1.39834881e-01 2.47098114e-02 9.25673306e-01 6.94156110e-01
4.33747768e-01 -1.89206883e-01 -3.85524303e-01 3.37729990e-01
1.42247748e+00 2.57989973e-01 3.19320261e-01 1.44589264e-02
6.21638954e-01 7.18227863e-01 -2.43499070e-01 4.20463860e-01
1.87363565e-01 4.03860480e-01 2.42766097e-01 4.56082188e-02
3.65931183e-01 -2.74716198e-01 4.71787274e-01 1.04301047e+00
-1.86822470e-02 -8.00664276e-02 -9.27858889e-01 3.49092752e-01
-1.67273891e+00 -1.09447336e+00 1.41237795e-01 2.05361342e+00
7.77431309e-01 1.18126546e-03 4.07993197e-01 1.42494887e-01
5.23341715e-01 3.24310541e-01 -8.55321229e-01 -6.31655872e-01
-1.34102240e-01 3.95660698e-02 1.31297439e-01 7.99230993e-01
-1.01325345e+00 6.26976371e-01 7.02497244e+00 5.33080816e-01
-1.11073565e+00 -1.47520214e-01 3.16663802e-01 9.01682600e-02
-5.23144007e-01 1.36531010e-01 -6.22642338e-01 7.06009984e-01
1.36707067e+00 1.45053407e-02 5.53990304e-01 6.94459796e-01
3.78519326e-01 1.55277858e-02 -1.09027147e+00 6.52411222e-01
-4.33219016e-01 -1.58985603e+00 2.11399496e-01 5.98897219e-01
3.89292240e-01 1.15994655e-01 -5.10473095e-04 1.08983126e-02
4.49446529e-01 -1.10751522e+00 4.92022961e-01 6.37795389e-01
6.75540686e-01 -7.14519322e-01 2.65441895e-01 4.10319567e-01
-7.71389961e-01 9.76234972e-02 -5.01684248e-01 -2.25577518e-01
1.21476009e-01 4.82907325e-01 -4.78653997e-01 1.25337973e-01
3.17927659e-01 5.27359784e-01 -1.94641471e-01 2.11638361e-01
3.86675090e-01 4.25508112e-01 -5.27474999e-01 -1.42239213e-01
3.20261031e-01 -8.67243350e-01 7.01350272e-01 9.52948213e-01
-1.44120887e-01 5.20162247e-02 -1.45478681e-01 1.36187410e+00
3.32062274e-01 -2.45425045e-01 -6.35020375e-01 -8.00710201e-01
2.56672591e-01 9.22139466e-01 -8.50957394e-01 -1.75111175e-01
-1.04112677e-01 6.85758471e-01 2.48822227e-01 4.98051196e-01
-6.18969142e-01 -2.77409434e-01 1.19923639e+00 -5.26253842e-02
6.68043137e-01 -7.12117612e-01 1.97778374e-01 -1.20552850e+00
-4.23500575e-02 -9.67328191e-01 -4.71747555e-02 -2.94773340e-01
-1.14416194e+00 4.22836870e-01 -3.14569399e-02 -7.33085513e-01
-4.75586474e-01 -8.14076364e-01 -4.22542781e-01 1.00121546e+00
-1.00946343e+00 -6.72880054e-01 2.34212548e-01 3.31990212e-01
8.56574997e-02 3.65591198e-02 1.17740464e+00 -1.06554255e-01
-6.60664558e-01 6.57160114e-03 9.93306100e-01 -1.37463123e-01
1.29063398e-01 -1.44992757e+00 4.44055855e-01 4.36078906e-01
2.86974609e-02 1.15379918e+00 1.15300059e+00 -6.05834663e-01
-1.64333224e+00 -6.03090823e-01 3.65405917e-01 -8.88822436e-01
8.23563099e-01 -7.21114576e-01 -1.08084381e+00 3.63446295e-01
5.57110235e-02 -2.87273377e-01 9.87330556e-01 1.03625603e-01
-1.49063438e-01 5.29407024e-01 -8.96779358e-01 4.61181104e-01
6.25448108e-01 -8.94239128e-01 -7.21208870e-01 1.59043327e-01
6.58092856e-01 2.92346209e-01 -1.16634035e+00 4.60279733e-03
5.88218212e-01 -1.01022649e+00 1.04951990e+00 -1.09952092e+00
3.95968825e-01 -3.04095685e-01 -3.26175570e-01 -1.13365483e+00
-5.06323755e-01 -7.24706650e-01 -4.15456742e-01 8.84278417e-01
2.17200622e-01 -5.50435662e-01 8.14797997e-01 7.01912701e-01
2.82304496e-01 -8.21064949e-01 -8.81056070e-01 -5.41881680e-01
6.05021596e-01 -3.25092614e-01 4.34094697e-01 9.38434124e-01
1.08445868e-01 3.30163151e-01 2.84995772e-02 5.79292774e-02
8.02053809e-01 9.69003513e-02 4.47223961e-01 -1.35024333e+00
-4.06457692e-01 -5.40994167e-01 -6.09228134e-01 -8.27310681e-01
2.16201723e-01 -8.09893250e-01 -3.12938124e-01 -1.05882573e+00
2.60881513e-01 -1.57909900e-01 -2.84441561e-01 -5.27024679e-02
9.45432857e-02 -4.30172712e-01 -6.10463955e-02 4.48911428e-01
-3.88583094e-01 1.04952848e+00 7.39132464e-01 1.50879170e-03
-4.82736140e-01 -4.89359260e-01 -6.34135842e-01 6.12519264e-01
6.02437735e-01 -4.08668339e-01 -3.34550679e-01 3.66739593e-02
3.02454233e-01 -1.11979768e-02 4.26531076e-01 -6.04695261e-01
-1.57462340e-02 -2.29851380e-01 4.25594032e-01 -5.05837917e-01
4.64761883e-01 -6.52901471e-01 4.07106131e-01 3.96959096e-01
-2.41172895e-01 -3.68925892e-02 1.88311711e-01 8.69650960e-01
-1.36776701e-01 1.03640936e-01 7.65640020e-01 -4.35627922e-02
-4.59674597e-01 1.05620839e-01 -8.31716478e-01 -1.05130777e-01
8.10266793e-01 -4.29090351e-01 3.80089171e-02 -1.58560395e-01
-1.13910937e+00 -1.28190249e-01 9.31820989e-01 1.65109128e-01
4.17165071e-01 -1.13978922e+00 -2.16126963e-01 5.28393149e-01
-1.79365039e-01 -1.85645401e-01 4.46071059e-01 6.55920684e-01
-4.39620763e-01 5.89993179e-01 -3.63634616e-01 -8.64282489e-01
-8.55145931e-01 8.40661705e-01 5.16033173e-01 -1.12811096e-01
-6.60278857e-01 2.83361882e-01 3.05179626e-01 -5.10880113e-01
-3.04257751e-01 -3.95325482e-01 7.26969764e-02 3.84094179e-01
1.94778323e-01 2.85734296e-01 -2.09959283e-01 -7.51892567e-01
-3.32459509e-01 5.17132998e-01 -3.56724709e-02 6.17323257e-03
1.57656527e+00 -2.16401473e-01 -2.33046457e-01 5.03504455e-01
1.35104275e+00 -1.89437628e-01 -1.67952609e+00 9.47406814e-02
-2.51634140e-02 1.97135210e-01 1.75674468e-01 -1.64213359e-01
-3.23516071e-01 1.21563900e+00 5.11421144e-01 5.60068548e-01
4.77528334e-01 1.46829203e-01 4.58044529e-01 4.66339737e-01
8.88294578e-02 -9.39741373e-01 -2.19543472e-01 3.59914690e-01
5.34019709e-01 -1.13952088e+00 -9.89142880e-02 2.17602119e-01
-4.70268935e-01 1.20174325e+00 -1.35754496e-01 -2.50883549e-01
6.72693849e-01 2.08807915e-01 -1.61940575e-01 -5.48562407e-01
-7.82130957e-01 -5.05676530e-02 1.24281980e-01 6.63096666e-01
1.83523029e-01 1.63615108e-01 8.41737315e-02 4.84010160e-01
-1.06331646e-01 -3.93802553e-01 3.84066969e-01 8.50119889e-01
-5.07728338e-01 -1.08462691e+00 -1.49095207e-01 7.51969665e-02
-1.33638367e-01 -5.08693680e-02 -5.72050512e-01 8.60726118e-01
-8.19046721e-02 5.10698974e-01 1.28626227e-01 -2.06545010e-01
-1.27395287e-01 5.52152038e-01 3.07760835e-01 -3.17418575e-01
-1.04413867e-01 2.88571537e-01 -2.84770072e-01 -3.89564723e-01
-2.01644197e-01 -8.29072177e-01 -1.12465763e+00 -4.86102343e-01
-2.92178929e-01 4.99392807e-01 6.74548864e-01 9.74388301e-01
7.63695002e-01 4.19517100e-01 4.56714183e-01 -1.13760436e+00
-7.49030352e-01 -7.28467941e-01 -5.82069635e-01 5.55089116e-01
7.92007804e-01 -7.70132542e-01 -6.86679184e-01 -1.43643911e-03] | [5.146078586578369, 5.196671485900879] |
c5f47ac8-e824-4ce3-a477-e61e85355166 | appearance-invariance-in-convolutional | 1707.00755 | null | http://arxiv.org/abs/1707.00755v1 | http://arxiv.org/pdf/1707.00755v1.pdf | Appearance invariance in convolutional networks with neighborhood similarity | We present a neighborhood similarity layer (NSL) which induces appearance
invariance in a network when used in conjunction with convolutional layers. We
are motivated by the observation that, even though convolutional networks have
low generalization error, their generalization capability does not extend to
samples which are not represented by the training data. For instance, while
novel appearances of learned concepts pose no problem for the human visual
system, feedforward convolutional networks are generally not successful in such
situations. Motivated by the Gestalt principle of grouping with respect to
similarity, the proposed NSL transforms its input feature map using the feature
vectors at each pixel as a frame of reference, i.e. center of attention, for
its surrounding neighborhood. This transformation is spatially varying, hence
not a convolution. It is differentiable; therefore, networks including the
proposed layer can be trained in an end-to-end manner. We analyze the
invariance of NSL to significant changes in appearance that are not represented
in the training data. We also demonstrate its advantages for digit recognition,
semantic labeling and cell detection problems. | ['Mehran Javanmardi', 'Mehdi Sajjadi', 'Tolga Tasdizen', 'Nisha Ramesh'] | 2017-07-03 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 4.69834030e-01 4.07047011e-02 4.13010642e-02 -6.19747043e-01
2.33151093e-01 -5.57040751e-01 6.62123024e-01 3.06902856e-01
-7.28307962e-01 6.52586997e-01 -2.67896295e-01 -2.24548727e-02
1.81176528e-01 -8.27038884e-01 -9.46646273e-01 -8.13197374e-01
1.98967084e-01 -6.09974600e-02 4.29951489e-01 -8.44874159e-02
1.47379577e-01 1.16251123e+00 -1.59024537e+00 4.52654809e-01
5.26557505e-01 9.82989490e-01 2.33786881e-01 4.23854947e-01
-2.51933048e-03 6.62599981e-01 -5.43565869e-01 1.72436610e-02
3.11778367e-01 -4.89948452e-01 -6.51862442e-01 3.27271551e-01
9.04657006e-01 -4.63059284e-02 -5.20146430e-01 1.26528060e+00
2.95333892e-01 3.63292694e-01 8.55969965e-01 -1.11002910e+00
-1.01020110e+00 6.52643666e-02 -2.09804341e-01 3.64690632e-01
3.26930732e-02 -8.09796825e-02 6.90420270e-01 -9.54492748e-01
7.87494361e-01 1.21889317e+00 5.54746628e-01 7.27677703e-01
-1.47171748e+00 -2.75919855e-01 4.47053224e-01 1.21002123e-01
-1.35593224e+00 -4.26194429e-01 8.33497703e-01 -1.55441359e-01
7.74844587e-01 2.50870377e-01 6.14795089e-01 9.80089068e-01
3.72181654e-01 7.45628417e-01 9.87584114e-01 -4.49142963e-01
5.21961749e-01 4.11801375e-02 6.12358078e-02 5.78686297e-01
2.60775417e-01 -3.36362608e-02 -3.21584851e-01 3.44033599e-01
1.10561585e+00 3.29267383e-01 -2.33526260e-01 -8.17138374e-01
-1.31423688e+00 4.64396745e-01 1.12265933e+00 5.01144171e-01
-3.79761606e-01 2.74255216e-01 2.74367779e-01 4.20747668e-01
2.42263570e-01 3.39023978e-01 -2.62545526e-01 5.28955042e-01
-6.99130297e-01 -6.76623285e-02 2.85478324e-01 7.43364692e-01
8.35973978e-01 9.90516692e-02 -1.58200189e-01 7.17394710e-01
8.64265263e-02 3.02401036e-01 6.53629184e-01 -6.81470752e-01
-1.64909944e-01 7.94532835e-01 -2.16990262e-01 -8.92221808e-01
-6.93438888e-01 -7.22129047e-01 -1.05792725e+00 7.36395836e-01
8.13248575e-01 8.29948038e-02 -1.11800766e+00 2.06388235e+00
2.99762338e-01 4.01440300e-02 4.10750002e-01 8.37944090e-01
6.99279130e-01 3.05703878e-01 4.23887111e-02 4.89449129e-02
1.15853906e+00 -7.66476452e-01 -4.53253895e-01 -2.93788075e-01
6.51472032e-01 -5.08330643e-01 9.47671533e-01 1.26843125e-01
-9.92926359e-01 -8.80951226e-01 -1.10834980e+00 -1.60538897e-01
-8.30854416e-01 1.68043569e-01 6.57719135e-01 3.59551013e-01
-1.22383404e+00 6.91124916e-01 -7.70433426e-01 -9.76819813e-01
6.71785772e-01 3.47418278e-01 -5.97820997e-01 -1.80102408e-01
-8.04241776e-01 9.56843555e-01 5.44830859e-01 2.40036562e-01
-6.64603233e-01 -3.67686570e-01 -8.70417833e-01 2.22769275e-01
-1.81188554e-01 -5.14591336e-01 8.80911112e-01 -1.65481842e+00
-1.20527780e+00 9.47135389e-01 -2.19027087e-01 -5.40890276e-01
5.16744971e-01 2.29843706e-01 -3.77039075e-01 1.44900382e-01
-7.86927938e-02 9.37368751e-01 8.25928330e-01 -1.13680947e+00
-6.40294850e-01 -4.17405307e-01 2.38072529e-01 2.14552820e-01
-4.78070617e-01 -1.78454056e-01 -1.11649022e-01 -7.34936059e-01
6.06601119e-01 -7.75535643e-01 -3.37057412e-01 6.70544863e-01
-1.85965165e-01 -1.32584795e-01 1.02999198e+00 -1.88085720e-01
4.97688085e-01 -2.35735369e+00 -1.68311417e-01 4.55546558e-01
-8.44607055e-02 2.25390136e-01 -3.12934160e-01 1.57726705e-01
-5.16269445e-01 -2.60489821e-01 -2.47220576e-01 3.34058516e-02
-2.39482135e-01 2.84042299e-01 1.19881667e-02 7.06374526e-01
3.91826361e-01 1.00957680e+00 -1.01589942e+00 -2.43560597e-01
4.08068925e-01 4.38700736e-01 -3.38918418e-01 -1.32745281e-02
-8.55462998e-02 4.63064581e-01 -1.12974100e-01 5.36663353e-01
6.73251629e-01 -3.61431330e-01 1.30346427e-02 -4.83422995e-01
-1.17066845e-01 -1.34084523e-01 -1.13096094e+00 1.87188697e+00
-2.76530951e-01 9.59104240e-01 -2.86863357e-01 -1.23113203e+00
1.00041330e+00 5.54984389e-03 2.06828102e-01 -8.64802599e-01
7.28886575e-02 1.53575391e-01 4.63509560e-01 -8.74538571e-02
1.73147559e-01 -1.07324064e-01 2.88430631e-01 3.07731599e-01
2.38260835e-01 4.13600951e-01 2.88857847e-01 -7.39156618e-04
1.02184772e+00 2.63159480e-02 3.25320154e-01 -4.43806171e-01
6.52206838e-01 -2.42522195e-01 3.89951140e-01 7.78365672e-01
-3.15976471e-01 7.22028613e-01 3.10474217e-01 -6.81915283e-01
-1.05878723e+00 -1.33918166e+00 -5.20132005e-01 1.08697104e+00
3.01022828e-01 1.79403126e-01 -7.57397592e-01 -7.93697953e-01
6.30155578e-02 4.01918530e-01 -9.49827433e-01 -4.33163077e-01
-4.20325220e-01 -4.49824959e-01 3.98613364e-01 7.67341256e-01
7.31167734e-01 -1.27068520e+00 -9.34467852e-01 2.87651867e-01
3.99479687e-01 -1.08931422e+00 -3.54933083e-01 5.13260126e-01
-9.08484459e-01 -1.04265285e+00 -8.87072265e-01 -1.22617483e+00
1.30699468e+00 2.66660035e-01 6.50759161e-01 4.55393605e-02
-4.40818280e-01 4.57528591e-01 -2.81943172e-01 -2.09348485e-01
-3.01486909e-01 -1.55294284e-01 2.62058750e-02 2.86549240e-01
4.50387925e-01 -5.25771320e-01 -7.39380479e-01 3.15240473e-01
-1.12596512e+00 -1.61091406e-02 6.23866498e-01 1.04773951e+00
6.40773535e-01 -8.07469860e-02 6.58212543e-01 -7.49747157e-01
1.95149705e-01 6.09666146e-02 -2.49887034e-01 4.07674551e-01
-1.42304972e-01 1.02303781e-01 7.04855323e-01 -5.22249341e-01
-8.73835146e-01 4.09533471e-01 1.03496596e-01 -2.59572923e-01
-6.06261492e-01 3.40782881e-01 -1.92422628e-01 -4.16974723e-01
9.53309000e-01 3.06208223e-01 1.61925793e-01 -2.61559606e-01
3.30192059e-01 4.39826488e-01 7.56765604e-01 -3.09703171e-01
7.33204067e-01 7.47109354e-01 2.14767843e-01 -1.05949426e+00
-6.41999543e-01 -3.47886354e-01 -1.10745263e+00 -2.22234115e-01
8.61746728e-01 -6.13376856e-01 -5.37740469e-01 5.64467669e-01
-1.18178153e+00 -1.57078266e-01 -7.06316411e-01 4.13390994e-01
-5.78314364e-01 3.13181967e-01 -2.72629231e-01 -5.60920119e-01
-2.71989796e-02 -7.53104627e-01 6.21386528e-01 4.42374736e-01
-2.05964953e-01 -1.11082304e+00 -3.98088932e-01 -3.76758695e-01
3.32066864e-01 4.05314773e-01 1.10782731e+00 -8.29399705e-01
-3.28993082e-01 -3.41648489e-01 -4.77631480e-01 5.53985536e-01
5.61903059e-01 -9.05356482e-02 -1.15953970e+00 -5.92010856e-01
-1.65618151e-01 -2.57325023e-01 9.38967407e-01 4.71024424e-01
1.23236597e+00 -1.50578236e-02 -2.96816319e-01 6.55394614e-01
1.41591942e+00 3.56429040e-01 7.60571718e-01 4.60711092e-01
4.43415880e-01 6.23364270e-01 2.38174260e-01 1.62421703e-01
-1.09498285e-01 5.24805069e-01 6.10553145e-01 -7.02721715e-01
-3.37336361e-01 -1.01631898e-02 1.43045843e-01 1.91946015e-01
-5.27155250e-02 -6.50722533e-02 -6.47638142e-01 6.00181222e-01
-1.84015799e+00 -8.05103064e-01 8.26852992e-02 2.35670519e+00
5.16948938e-01 5.63520901e-02 -4.99293543e-02 9.67565775e-02
8.61658573e-01 -6.36255518e-02 -8.88322115e-01 -4.59033847e-01
-6.20590806e-01 2.45781124e-01 4.49240088e-01 1.42418995e-01
-1.21278107e+00 8.79593492e-01 5.97007322e+00 5.48888564e-01
-1.29775095e+00 -2.03331590e-01 6.89969897e-01 2.94945478e-01
4.03771177e-02 -1.47617862e-01 -4.19678599e-01 1.90894142e-01
4.15017784e-01 1.23301558e-02 2.40445167e-01 7.82961071e-01
-4.81367670e-02 5.39723188e-02 -1.56899476e+00 8.64781201e-01
1.01990633e-01 -1.27737284e+00 3.11932892e-01 -1.51237711e-01
7.22217262e-01 -1.82171345e-01 3.46685708e-01 3.65466848e-02
3.57819051e-02 -1.18138969e+00 6.81082308e-01 4.48910236e-01
6.76249743e-01 -7.50553071e-01 6.15485072e-01 7.90989026e-02
-9.94042158e-01 -1.24959253e-01 -8.48540843e-01 -8.48206580e-02
-2.49009445e-01 1.73635319e-01 -7.64687657e-01 2.30028644e-01
5.19768536e-01 7.98655331e-01 -8.33400965e-01 1.21739662e+00
-9.21718702e-02 -2.52754483e-02 -2.02562362e-01 -4.99959998e-02
4.56563532e-01 2.54811123e-02 2.33377039e-01 1.27252865e+00
2.83041805e-01 -9.92562994e-02 -4.59356047e-02 1.01986754e+00
-1.26043916e-01 1.16056323e-01 -6.90003335e-01 4.19510566e-02
1.69471934e-01 1.15746474e+00 -1.13281834e+00 -3.48330677e-01
-5.08443654e-01 1.35628855e+00 3.69340479e-01 7.34297514e-01
-4.23615485e-01 -5.61890721e-01 7.64174759e-01 8.39079693e-02
4.88639623e-01 -2.11119112e-02 -2.46757194e-01 -8.93266678e-01
-6.48168176e-02 -6.15687370e-01 2.36649141e-01 -8.10600877e-01
-1.31708372e+00 6.20859504e-01 -3.71040285e-01 -1.25718844e+00
-3.41781639e-02 -1.10821509e+00 -6.16985500e-01 8.84723365e-01
-1.49345601e+00 -1.15844846e+00 -3.35921943e-01 9.10178125e-01
3.96504253e-01 -1.46084592e-01 9.91514504e-01 8.31254050e-02
-2.65290916e-01 7.55986691e-01 3.86413574e-01 3.16494048e-01
6.12483680e-01 -1.29309893e+00 3.74567598e-01 9.90189254e-01
3.83500308e-01 6.60721421e-01 4.84881461e-01 -2.56958961e-01
-9.71047223e-01 -1.29727185e+00 7.57365644e-01 -1.16599109e-02
4.02021855e-01 -4.77900624e-01 -1.00075221e+00 6.17075026e-01
4.17621993e-02 5.14558196e-01 4.58693266e-01 -1.48009703e-01
-4.88849670e-01 -2.59491563e-01 -1.30113792e+00 7.39099622e-01
1.05929887e+00 -6.13961339e-01 -4.37452257e-01 2.60926843e-01
4.10376608e-01 -2.65840590e-01 -6.60363019e-01 2.69743115e-01
5.95668972e-01 -8.99600089e-01 1.04771662e+00 -8.79937768e-01
1.53990895e-01 -5.22749901e-01 -3.93045366e-01 -1.29078031e+00
-6.17724180e-01 4.57617119e-02 2.94204444e-01 9.24303055e-01
3.86181384e-01 -7.52518356e-01 9.40050244e-01 5.43243110e-01
-2.33177334e-01 -4.90184456e-01 -1.05364108e+00 -1.12812972e+00
1.70359582e-01 -9.82309729e-02 3.80466133e-01 9.41166103e-01
-9.63908061e-02 1.96829259e-01 2.44013011e-03 2.00636730e-01
5.81581473e-01 2.36503452e-01 4.91851598e-01 -1.32905269e+00
-6.37189969e-02 -4.23418313e-01 -1.08989203e+00 -1.10488045e+00
1.03986450e-01 -1.05506301e+00 -3.11920252e-02 -1.44585538e+00
4.60730344e-02 -3.44891369e-01 -8.23514223e-01 6.95711315e-01
1.85592473e-01 5.63327909e-01 1.95273533e-01 1.05118519e-03
-4.09158736e-01 3.62382799e-01 1.43601012e+00 -3.26291829e-01
1.24897342e-02 -1.08544305e-01 -4.14807260e-01 7.63825476e-01
8.39659929e-01 -1.33188009e-01 -3.63778800e-01 -2.16324762e-01
-2.58369595e-01 -5.67219019e-01 6.03379190e-01 -1.34130645e+00
3.04044664e-01 -2.30107494e-02 1.06971037e+00 -3.50701094e-01
1.75001219e-01 -1.18982553e+00 7.66722262e-02 6.70233727e-01
-5.36694527e-01 -5.62601984e-02 3.01221997e-01 5.05367815e-01
-1.80806696e-01 -4.08644587e-01 1.20372736e+00 -1.62972406e-01
-9.75785553e-01 2.41768107e-01 -4.44652170e-01 -3.37917119e-01
9.95560646e-01 -8.28300118e-01 -3.71768147e-01 -2.02173874e-01
-8.21199119e-01 -2.44894579e-01 7.50391364e-01 2.60343611e-01
7.63401151e-01 -1.55046463e+00 -4.73467439e-01 4.79472995e-01
3.92202795e-01 -3.12757865e-02 2.18131721e-01 6.79933608e-01
-5.83190918e-01 3.02324891e-01 -7.17774212e-01 -7.45794177e-01
-1.13697135e+00 8.88176620e-01 5.25717437e-01 3.00248384e-01
-6.57170057e-01 7.98571944e-01 7.04493880e-01 -2.52178967e-01
3.14223140e-01 -5.93065441e-01 -2.67993748e-01 -1.74714446e-01
5.62088132e-01 -1.42773271e-01 1.42677441e-01 -6.81427360e-01
-4.35750872e-01 5.36149800e-01 -1.72909975e-01 3.07479531e-01
1.22011399e+00 -3.78428772e-02 -5.89954946e-03 6.21244311e-01
1.25074708e+00 -4.63573486e-01 -1.46341991e+00 -5.20459592e-01
-4.38598208e-02 -4.22245413e-01 -1.14860155e-01 -7.53768384e-01
-1.16713285e+00 8.52791071e-01 1.05983925e+00 1.80925548e-01
1.16231751e+00 -1.04316294e-01 2.25588366e-01 7.50239730e-01
7.50708655e-02 -1.05916190e+00 1.30801752e-01 5.57162762e-01
8.34529996e-01 -1.16109729e+00 -3.42577279e-01 -2.22319797e-01
-3.18334669e-01 1.41829824e+00 7.53969073e-01 -3.37374866e-01
4.53451306e-01 -8.99645500e-03 1.89321250e-01 -1.02411136e-02
-4.18694139e-01 -3.83503646e-01 2.60059625e-01 9.93065000e-01
5.10203540e-01 -2.65770435e-01 -1.11278422e-01 -2.66637988e-02
2.24769086e-01 -3.62827033e-01 4.60408449e-01 9.43549693e-01
-6.20738268e-01 -8.99417579e-01 -1.76438928e-01 3.06552202e-01
-2.89301239e-02 8.56471658e-02 -3.14227909e-01 9.00623083e-01
2.66312957e-01 4.80093092e-01 5.52452803e-01 -1.44460291e-01
4.99962658e-01 1.18735977e-01 6.65862262e-01 -6.28819346e-01
-3.40826571e-01 -1.04558252e-01 -3.67535114e-01 -3.35259497e-01
-5.86432397e-01 -6.80453956e-01 -1.45748580e+00 -9.78315100e-02
-1.16015404e-01 -1.70939177e-01 5.54159164e-01 7.87158489e-01
9.64043364e-02 7.11288512e-01 6.08657718e-01 -8.05332601e-01
-2.35175908e-01 -8.61356080e-01 -8.18717599e-01 7.00631380e-01
3.63514751e-01 -6.06953621e-01 -1.66159749e-01 2.86872447e-01] | [9.475353240966797, 2.246067762374878] |
a4de35d0-7fc3-4f44-a714-ff6c13388d18 | hopeful-men-lt-edi-eacl2021-hope-speech-1 | null | null | https://aclanthology.org/2021.ltedi-1.23 | https://aclanthology.org/2021.ltedi-1.23.pdf | Hopeful Men@LT-EDI-EACL2021: Hope Speech Detection Using Indic Transliteration and Transformers | This paper aims to describe the approach we used to detect hope speech in the HopeEDI dataset. We experimented with two approaches. In the first approach, we used contextual embeddings to train classifiers using logistic regression, random forest, SVM, and LSTM based models. The second approach involved using a majority voting ensemble of 11 models which were obtained by fine-tuning pre-trained transformer models (BERT, ALBERT, RoBERTa, IndicBERT) after adding an output layer. We found that the second approach was superior for English, Tamil and Malayalam. Our solution got a weighted F1 score of 0.93, 0.75 and 0.49 for English, Malayalam and Tamil respectively. Our solution ranked 1st in English, 8th in Malayalam and 11th in Tamil. | ['Radhika Mamidi', 'Anshul Wadhawan', 'Nikhil E', 'Ishan Sanjeev Upadhyay'] | null | null | null | null | eacl-ltedi-2021-4 | ['hope-speech-detection', 'transliteration'] | ['natural-language-processing', 'natural-language-processing'] | [-1.34078011e-01 3.73504698e-01 1.24022402e-02 -1.43738940e-01
-7.88941622e-01 -4.89997029e-01 1.08301020e+00 3.50774415e-02
-5.04745603e-01 9.16738868e-01 6.30946159e-01 -6.57809913e-01
-2.82648772e-01 -7.01426983e-01 -2.85202086e-01 -4.46356952e-01
-7.47592747e-02 5.61038673e-01 1.18162192e-01 -4.46375340e-01
7.14731753e-01 4.14376587e-01 -1.02501583e+00 5.06000221e-01
6.61598504e-01 8.33414674e-01 -2.79522352e-02 1.11818409e+00
-9.41117778e-02 1.20828032e+00 -5.89019895e-01 -4.35202539e-01
-1.15464978e-01 -4.55165505e-01 -1.07591987e+00 -3.08157831e-01
7.22370809e-03 -3.54800075e-02 -4.03224021e-01 6.02294385e-01
4.85343933e-01 1.82813898e-01 8.18435311e-01 -9.96015787e-01
-8.43397021e-01 8.99769247e-01 -2.39668369e-01 1.81579083e-01
5.01442969e-01 -2.34532163e-01 7.78529167e-01 -1.10435748e+00
4.79186505e-01 1.30746567e+00 8.23863566e-01 4.88594621e-01
-8.12845409e-01 -5.85650921e-01 -4.82367635e-01 3.25686038e-01
-1.29644358e+00 -6.04145169e-01 6.14645362e-01 -2.10193008e-01
1.31892014e+00 2.62547970e-01 7.42912591e-01 1.21260881e+00
2.18136266e-01 6.56795502e-01 1.64925027e+00 -8.81751895e-01
1.14557154e-01 6.68571770e-01 -1.30198166e-01 9.94662762e-01
-4.88647461e-01 1.51647240e-01 -5.68995237e-01 -3.64164203e-01
3.19131017e-01 -4.04697478e-01 9.59154889e-02 3.42462987e-01
-1.08992100e+00 1.30416834e+00 3.37585598e-01 6.98960006e-01
-4.07217175e-01 -2.65033305e-01 3.06355715e-01 3.79107684e-01
5.06859839e-01 3.96675766e-01 -6.07744694e-01 -3.04427922e-01
-1.28764427e+00 -5.91067113e-02 6.98098421e-01 7.01702595e-01
1.96370825e-01 2.11943522e-01 7.80737177e-02 1.21469188e+00
4.54002023e-01 1.26554519e-01 1.01718390e+00 -6.17941678e-01
4.04633135e-01 1.99897632e-01 1.60030276e-01 -7.51009107e-01
-3.53266299e-01 -2.88707525e-01 -4.70909894e-01 -6.57545403e-02
2.44613484e-01 -2.72962332e-01 -1.27484035e+00 1.21207714e+00
-2.03816742e-01 -7.26579651e-02 5.56588709e-01 3.63300949e-01
9.82595026e-01 8.03061604e-01 -1.38103487e-02 -1.63076699e-01
1.15858674e+00 -1.33426595e+00 -5.92033088e-01 -2.00966131e-02
2.54480332e-01 -1.32976639e+00 8.27552617e-01 5.96124828e-01
-1.03863811e+00 -3.50917548e-01 -9.76125300e-01 2.42881015e-01
-7.23621070e-01 2.20375583e-01 2.73338467e-01 8.63418579e-01
-1.24827635e+00 6.33759022e-01 -5.31592131e-01 -6.65442109e-01
1.21372476e-01 3.17024797e-01 -4.31353271e-01 1.93525568e-01
-1.21405029e+00 1.45516026e+00 3.72921675e-01 -9.84313712e-02
-9.06866968e-01 1.11737318e-01 -6.18023813e-01 -2.72378027e-01
-3.35592479e-01 -1.07205333e-02 1.20184970e+00 -6.46738589e-01
-1.75414658e+00 8.59142780e-01 -3.40424865e-01 -7.40282714e-01
4.12600517e-01 1.02948546e-02 -8.01916122e-01 3.26882266e-02
-1.75051108e-01 6.08408630e-01 6.59744978e-01 -9.13600147e-01
-7.49859929e-01 -2.35536501e-01 -1.02028891e-01 -9.08604488e-02
-2.21540570e-01 3.07895511e-01 3.79125118e-01 -4.38024312e-01
1.99628994e-01 -6.72802746e-01 2.21777335e-02 -8.59763622e-01
-3.60277295e-01 -3.86560559e-01 9.04934049e-01 -1.15797162e+00
1.13698184e+00 -1.70262814e+00 -6.01939624e-03 2.20918745e-01
-6.82799593e-02 3.55912417e-01 8.08876455e-02 7.04611719e-01
7.45889079e-03 2.73052722e-01 6.32113516e-02 -1.30860507e-01
-3.73859592e-02 3.21763009e-01 -1.65106416e-01 1.93825454e-01
2.88401812e-01 6.15328252e-01 -7.94966340e-01 -6.03491783e-01
3.26048642e-01 8.14362884e-01 -2.50993609e-01 2.57003069e-01
2.71604359e-01 8.63005519e-02 -6.47471175e-02 8.13031852e-01
3.31842959e-01 3.10887247e-01 1.37217548e-02 9.86776724e-02
-3.35740805e-01 7.11397290e-01 -8.49444270e-01 1.11417568e+00
-7.94245362e-01 9.41889584e-01 -2.80289292e-01 -1.08665574e+00
1.38513827e+00 7.37476468e-01 -3.08904308e-03 -5.68105400e-01
2.05699891e-01 5.39029896e-01 -1.75861560e-03 -6.60821974e-01
6.36889696e-01 -3.23881418e-01 2.32881084e-02 2.89276503e-02
2.59009540e-01 -1.41870186e-01 -2.00628787e-01 2.18759626e-02
1.04141092e+00 -4.00476754e-02 2.76493847e-01 -1.13828108e-01
5.65594673e-01 -1.36279032e-01 1.93412945e-01 6.51823044e-01
-4.59520996e-01 6.42156780e-01 5.94066262e-01 -3.92645955e-01
-9.76571620e-01 -1.03727865e+00 -1.46361455e-01 9.92438436e-01
-5.99272788e-01 -5.42797804e-01 -5.82573712e-01 -8.80323470e-01
-6.20539188e-01 1.24313056e+00 -5.62705278e-01 -1.53387547e-01
-6.46595895e-01 -5.72624981e-01 7.73647785e-01 1.50841445e-01
7.06066966e-01 -1.13648176e+00 -5.72810829e-01 2.12077916e-01
-4.19239789e-01 -7.31601417e-01 5.28888293e-02 5.92286289e-01
-8.83958817e-01 -4.75752234e-01 -6.94134653e-01 -8.30097139e-01
8.74858946e-02 -4.00800139e-01 1.00197458e+00 -2.38020197e-01
3.66282165e-02 2.24522054e-02 -6.05905473e-01 -3.22411329e-01
-4.55667049e-01 1.19263776e-01 -1.65066957e-01 -4.37955886e-01
3.85464758e-01 -4.13088322e-01 -1.46056190e-01 -4.71626259e-02
-3.12914342e-01 -1.64965898e-01 5.84694266e-01 1.15530729e+00
9.12664272e-03 -8.03246535e-03 4.66370463e-01 -7.57734299e-01
8.98352742e-01 -5.88525116e-01 -8.11034292e-02 8.21129531e-02
-7.36853242e-01 1.31922022e-01 5.45451045e-01 -4.20039982e-01
-7.71332324e-01 1.87549800e-01 -6.22579277e-01 -2.73208506e-02
-1.45354643e-01 4.43844795e-01 2.89671391e-01 3.97679769e-02
6.32856250e-01 3.38431984e-01 -3.72271061e-01 -5.00694752e-01
2.85067737e-01 1.27541816e+00 2.14536414e-01 -1.64906144e-01
5.23892164e-01 -9.89260823e-02 -3.78569812e-01 -1.19029009e+00
-4.86363411e-01 -1.57753423e-01 -6.39043987e-01 -2.79155582e-01
8.84193242e-01 -6.25277162e-01 -1.80562243e-01 3.96860868e-01
-1.24828637e+00 -1.15403384e-02 1.70460548e-02 6.01108670e-01
-3.01877588e-01 3.18973185e-03 -7.19970644e-01 -1.28687906e+00
-5.43984711e-01 -9.26750481e-01 6.78835928e-01 6.65296316e-02
-6.61964417e-01 -1.10537219e+00 1.47520900e-01 4.53372121e-01
7.00954735e-01 1.48914754e-01 1.03964746e+00 -9.37462509e-01
3.32696348e-01 -1.94136605e-01 -6.11912385e-02 3.74107510e-01
1.62190616e-01 1.87754795e-01 -1.18943119e+00 2.94291861e-02
-1.50250122e-02 -5.64937830e-01 8.48753989e-01 1.39701977e-01
5.33857942e-01 -4.94860947e-01 -9.24475864e-02 -2.14932002e-02
1.25537574e+00 6.18886828e-01 8.66678834e-01 5.60314417e-01
1.80881530e-01 3.66227031e-01 5.02990544e-01 1.94062412e-01
4.60656971e-01 4.45087999e-01 1.03426576e-01 1.37809128e-01
-4.64413874e-02 -5.41729212e-01 6.86581135e-01 9.52968717e-01
-1.40715197e-01 -2.11386994e-01 -1.04221737e+00 8.35094154e-01
-1.38852835e+00 -9.95919764e-01 -9.49982479e-02 2.01198840e+00
7.24174380e-01 4.12566692e-01 4.51167107e-01 4.84558344e-01
4.06915724e-01 5.04056923e-02 2.96249032e-01 -1.19475257e+00
9.46123376e-02 6.85727000e-01 3.97160351e-01 7.58270502e-01
-1.20334065e+00 1.16453910e+00 6.93448591e+00 6.08378351e-01
-1.17840815e+00 2.84257710e-01 4.31156814e-01 8.73050094e-02
-6.90012798e-02 1.93527177e-01 -6.80519462e-01 4.92029935e-01
1.80022275e+00 2.03514978e-01 4.36611325e-01 5.68567216e-01
1.39506027e-01 -1.74940929e-01 -5.81389487e-01 5.62422216e-01
9.73000154e-02 -1.10521281e+00 -2.37234503e-01 -7.51127750e-02
5.98725557e-01 3.46792758e-01 6.39902204e-02 5.93587399e-01
5.25264859e-01 -1.38938606e+00 6.28379107e-01 4.56683606e-01
2.09050372e-01 -8.85038376e-01 8.65387797e-01 4.43074226e-01
-8.06960762e-01 -3.62826437e-02 -3.51252466e-01 -1.53543353e-01
-3.37779373e-01 2.09626570e-01 -1.27159321e+00 2.24771068e-01
6.20763183e-01 2.98940301e-01 -5.44204593e-01 7.57728696e-01
-4.10253733e-01 9.29423332e-01 -2.97404498e-01 -5.53881705e-01
6.78012133e-01 1.86554685e-01 2.90945292e-01 1.56085610e+00
5.53790450e-01 -3.09294432e-01 -9.16865766e-02 3.43276501e-01
2.30961129e-01 1.50307164e-01 -6.68890774e-01 -8.57969001e-02
4.63705122e-01 1.16634691e+00 -5.54092109e-01 -4.39458281e-01
-1.71416685e-01 8.67860198e-01 2.46055514e-01 4.86972518e-02
-8.50384176e-01 -7.46704578e-01 -1.39601305e-01 -6.11673035e-02
4.00281101e-01 -1.79424092e-01 -1.40832439e-01 -8.17881405e-01
-3.02870244e-01 -9.60251391e-01 4.19370294e-01 -8.59389722e-01
-9.42310035e-01 1.11442232e+00 -7.01082572e-02 -6.54498875e-01
-6.50457323e-01 -7.15528190e-01 -8.32207739e-01 9.73967016e-01
-1.08178687e+00 -1.11535645e+00 1.46874502e-01 3.06077540e-01
7.43092477e-01 -7.84168541e-01 1.19313228e+00 2.05199316e-01
-3.18555772e-01 4.75924104e-01 1.84364721e-01 8.63384157e-02
4.18955654e-01 -1.29401708e+00 6.31299168e-02 4.28058773e-01
4.36609983e-01 4.90679353e-01 8.48477840e-01 -2.19827175e-01
-8.39381456e-01 -7.47384846e-01 1.81273365e+00 -3.52533340e-01
8.02925646e-01 -2.02549756e-01 -5.47725976e-01 6.94219410e-01
7.83541679e-01 -4.74995434e-01 5.70630252e-01 1.82852238e-01
-2.56478101e-01 1.54948220e-01 -1.57061064e+00 2.42116302e-01
3.40877503e-01 -6.57337129e-01 -9.71551001e-01 5.34582138e-01
2.49646857e-01 -1.29160911e-01 -1.05718517e+00 3.97001281e-02
7.61334002e-01 -1.10697532e+00 6.87415123e-01 -7.24407315e-01
7.67969608e-01 9.67216864e-02 -4.82663155e-01 -1.60639000e+00
-1.95379645e-01 -4.01336074e-01 -1.21301459e-02 1.28881800e+00
9.51570928e-01 -7.10794389e-01 6.09707177e-01 1.03410907e-01
1.15251355e-01 -8.78539681e-01 -1.07833827e+00 -5.15103519e-01
5.06374419e-01 -3.42267126e-01 2.13299781e-01 8.52741480e-01
1.69074982e-01 6.43458366e-01 -6.06018305e-01 -1.91623092e-01
1.44931436e-01 -2.94595659e-01 2.68037230e-01 -9.09843147e-01
-1.68998852e-01 -2.81430244e-01 -5.44388533e-01 -5.13206959e-01
7.04841986e-02 -8.15824389e-01 -1.99216887e-01 -1.80282295e+00
1.74420625e-02 -3.55506331e-01 -5.79437554e-01 7.17348099e-01
3.08280170e-01 3.75536293e-01 -1.40934475e-02 -5.87282628e-02
-2.80366465e-02 2.59850711e-01 5.53117871e-01 -1.90501362e-01
-4.99324054e-02 2.36310109e-01 -6.24656498e-01 8.85998666e-01
1.39930260e+00 -6.17356658e-01 -2.77642399e-01 -1.31346032e-01
-2.22124264e-01 2.67202873e-03 2.50582695e-01 -9.52991366e-01
1.05023764e-01 3.00021432e-02 6.96694970e-01 -7.57406592e-01
5.74077308e-01 -4.98094499e-01 -6.37065768e-02 6.88020766e-01
-4.57341492e-01 2.78845698e-01 8.29024166e-02 9.78701189e-03
-2.64446616e-01 -6.33896947e-01 9.39301252e-01 -3.38266611e-01
-6.08170509e-01 -5.03647387e-01 -8.46831918e-01 -2.10242197e-01
7.85473585e-01 -6.68231919e-02 -2.38417149e-01 -5.20455301e-01
-8.73221159e-01 -8.47759768e-02 -2.59096660e-02 6.09517038e-01
9.50953841e-01 -1.30412328e+00 -8.17879260e-01 2.88551778e-01
-2.10975051e-01 -9.49421883e-01 -3.70259881e-01 8.38646591e-01
-5.06068945e-01 4.88463521e-01 -2.40863279e-01 -9.10069197e-02
-1.50875914e+00 9.16721523e-02 3.61557484e-01 -3.26680481e-01
-1.94391370e-01 9.99303043e-01 -8.66740227e-01 -8.36298943e-01
2.97573447e-01 -9.85657349e-02 -5.83353937e-01 5.66608757e-02
1.52605191e-01 5.13778806e-01 1.13112964e-01 -8.18990052e-01
-5.95835328e-01 4.39574629e-01 -6.99580088e-02 -5.23662448e-01
1.31950450e+00 2.73734242e-01 -1.47873163e-01 6.09281659e-01
1.26384950e+00 2.36902758e-01 -2.56485909e-01 2.78275967e-01
3.40944439e-01 -2.21580535e-01 3.93109202e-01 -1.22702563e+00
-5.99010229e-01 9.27337945e-01 9.53558505e-01 5.22526503e-01
1.10829341e+00 -1.31704956e-01 6.58648729e-01 4.14512366e-01
1.42179072e-01 -1.26074481e+00 -2.12249383e-01 8.88033688e-01
8.38901579e-01 -1.12529182e+00 -2.00013727e-01 2.82925665e-01
-6.55981660e-01 1.42582250e+00 1.42155573e-01 -1.26127720e-01
6.47319376e-01 1.80972576e-01 8.60873163e-02 7.91495442e-02
-1.05417585e+00 9.15374905e-02 9.51570868e-02 6.13397419e-01
1.02193296e+00 2.63728738e-01 -5.74578702e-01 3.55395257e-01
-7.23407865e-01 -5.08204065e-02 5.62370777e-01 8.35332453e-01
-7.28073835e-01 -1.07278728e+00 -3.65379184e-01 5.41786075e-01
-6.91734791e-01 -3.01473677e-01 -8.38623047e-01 9.75160182e-01
2.73937315e-01 1.35883558e+00 -1.79892361e-01 -9.80359375e-01
3.22851896e-01 4.28681165e-01 2.83682048e-01 -1.37307167e-01
-7.51276612e-01 2.20337883e-01 6.42805696e-01 1.56396896e-01
-2.08224326e-01 -5.10895908e-01 -9.74245012e-01 -3.18363458e-01
-3.36415261e-01 4.96875018e-01 8.36373508e-01 9.48588192e-01
-8.18485320e-02 4.07470345e-01 6.69659257e-01 -4.12001014e-01
-5.22603393e-01 -1.48909020e+00 -3.66192013e-01 -2.93898284e-01
3.88177991e-01 -5.20798981e-01 -4.72897351e-01 -1.11686520e-01] | [9.565314292907715, 10.649045944213867] |
ca932dfb-994a-48bb-97f4-688ba71deb0b | coarse-to-fine-video-denoising-with-dual | 2205.00214 | null | https://arxiv.org/abs/2205.00214v2 | https://arxiv.org/pdf/2205.00214v2.pdf | Coarse-to-Fine Video Denoising with Dual-Stage Spatial-Channel Transformer | Video denoising aims to recover high-quality frames from the noisy video. While most existing approaches adopt convolutional neural networks~(CNNs) to separate the noise from the original visual content, however, CNNs focus on local information and ignore the interactions between long-range regions in the frame. Furthermore, most related works directly take the output after basic spatio-temporal denoising as the final result, leading to neglect the fine-grained denoising process. In this paper, we propose a Dual-stage Spatial-Channel Transformer for coarse-to-fine video denoising, which inherits the advantages of both Transformer and CNNs. Specifically, DSCT is proposed based on a progressive dual-stage architecture, namely a coarse-level and a fine-level stage to extract dynamic features and static features, respectively. At both stages, a Spatial-Channel Encoding Module is designed to model the long-range contextual dependencies at both spatial and channel levels. Meanwhile, we design a Multi-Scale Residual Structure to preserve multiple aspects of information at different stages, which contains a Temporal Features Aggregation Module to summarize the dynamic representation. Extensive experiments on four publicly available datasets demonstrate our proposed method achieves significant improvements compared to the state-of-the-art methods. | ['Huadong Ma', 'Huiyuan Fu', 'Chuanming Wang', 'Mengshi Qi', 'Wulian Yun'] | 2022-04-30 | null | null | null | null | ['video-denoising'] | ['computer-vision'] | [ 5.95339909e-02 -6.88051879e-01 8.92261714e-02 -4.87112850e-01
-6.69696450e-01 -1.84051111e-01 1.42655179e-01 -1.39009386e-01
-4.33415592e-01 3.80406916e-01 4.99474138e-01 1.29795223e-01
7.44744614e-02 -8.37677658e-01 -6.88238680e-01 -9.35672998e-01
2.44275019e-01 -7.39234090e-01 6.01166368e-01 -3.06846619e-01
8.03055533e-04 2.16191873e-01 -1.30025494e+00 5.82884729e-01
8.33040714e-01 1.38578176e+00 3.16906512e-01 5.37247956e-01
-9.50711593e-02 1.02779901e+00 -4.64924157e-01 -1.80795208e-01
9.77366939e-02 -5.55148840e-01 -1.36621758e-01 1.39858112e-01
3.54616851e-01 -7.54120350e-01 -9.90414560e-01 1.45963430e+00
4.58177000e-01 1.40213087e-01 -5.91609208e-03 -7.76289284e-01
-7.72424281e-01 2.49930799e-01 -7.22520471e-01 5.81336737e-01
2.34411657e-01 3.86548489e-01 6.64065003e-01 -7.60375857e-01
3.68053496e-01 1.50239050e+00 8.45632672e-01 3.17497998e-01
-9.93003070e-01 -7.47811079e-01 5.31184435e-01 4.77521092e-01
-1.29970884e+00 -5.35310805e-01 9.73344564e-01 -1.57776132e-01
7.11360455e-01 -5.63834561e-03 7.88722157e-01 8.81322563e-01
4.53060687e-01 7.15336978e-01 1.01831102e+00 9.88426059e-02
-1.36326300e-02 -8.08170795e-01 7.80857913e-03 6.45453691e-01
8.58329982e-02 1.86139077e-01 -5.56113958e-01 2.18474731e-01
1.05351484e+00 4.10539865e-01 -5.41447163e-01 9.53758731e-02
-1.18128276e+00 4.61697817e-01 5.54538131e-01 4.72065836e-01
-6.22789383e-01 4.01133716e-01 6.72800958e-01 3.75392467e-01
6.78927898e-01 -4.40698504e-01 -4.09231812e-01 -8.70593563e-02
-1.10409474e+00 1.83617651e-01 3.21547300e-01 9.10239697e-01
7.96003878e-01 1.81580886e-01 -5.47685444e-01 7.32464671e-01
3.34868401e-01 2.82865226e-01 3.36780459e-01 -1.10777116e+00
6.97055697e-01 3.83507997e-01 -9.41490605e-02 -1.29234147e+00
-5.94495535e-02 -6.19682968e-01 -1.52888870e+00 1.77253205e-02
-3.93564068e-02 -4.10093740e-02 -1.03141177e+00 1.61987662e+00
2.01452255e-01 7.30111957e-01 -1.23419397e-01 1.18397105e+00
1.02741480e+00 9.68878984e-01 2.21132681e-01 -5.41455686e-01
1.26519799e+00 -1.11507189e+00 -1.13413823e+00 6.99505359e-02
5.80719532e-03 -8.16299438e-01 6.41028225e-01 4.47499007e-01
-1.40375423e+00 -9.67879653e-01 -1.07712340e+00 -5.24932563e-01
8.12932476e-02 7.85661582e-03 2.71433681e-01 1.30141005e-01
-1.13273895e+00 6.77237153e-01 -1.05070102e+00 1.50461107e-01
4.82188284e-01 8.66051465e-02 -3.17819536e-01 -4.63856012e-01
-1.33598495e+00 3.29206020e-01 1.24165535e-01 8.15780699e-01
-1.06697202e+00 -5.03311753e-01 -9.56546068e-01 2.24555716e-01
3.97994757e-01 -6.67773545e-01 9.78130639e-01 -1.00496912e+00
-1.34117007e+00 2.18690380e-01 -6.31039679e-01 -1.15588725e-01
5.31851113e-01 -2.34749153e-01 -5.43109417e-01 2.93139011e-01
1.43208534e-01 3.23345840e-01 1.10234761e+00 -1.12748134e+00
-8.55585515e-01 -2.25686193e-01 1.36562645e-01 1.53268114e-01
-3.57099503e-01 5.19470014e-02 -1.21190906e+00 -1.30953693e+00
4.17746425e-01 -2.82689691e-01 -3.01431954e-01 9.33711752e-02
-2.34088123e-01 1.05919622e-01 1.07886624e+00 -1.07769895e+00
1.66887963e+00 -2.31580877e+00 3.61833513e-01 -6.32705390e-02
5.83601475e-01 3.91046882e-01 -3.23002428e-01 1.25880897e-01
-1.45633951e-01 -1.47488741e-02 -2.10783049e-01 -4.70360965e-01
-2.35864818e-01 2.31822684e-01 -1.51076302e-01 4.44111913e-01
2.00265974e-01 8.68615210e-01 -9.72441018e-01 -4.13766563e-01
4.40117657e-01 9.03848946e-01 -4.75614488e-01 3.17992538e-01
1.29121989e-01 6.32443845e-01 -6.64922893e-01 6.51135087e-01
1.07180893e+00 -1.27026495e-02 -1.80346109e-02 -7.44739354e-01
-2.75695086e-01 7.84531981e-02 -1.17640352e+00 1.92925096e+00
-3.20687205e-01 3.67946565e-01 5.22922635e-01 -8.61395240e-01
6.80488586e-01 3.00902337e-01 4.57276165e-01 -1.03257084e+00
2.03943595e-01 2.00573936e-01 -3.34580481e-01 -5.80032766e-01
4.68227684e-01 -1.27404809e-01 8.88530836e-02 -2.01659516e-01
5.44523224e-02 4.66747046e-01 1.45753354e-01 5.47785908e-02
1.23923135e+00 3.43878388e-01 -3.39279473e-02 -7.67717659e-02
9.22523379e-01 -5.21455944e-01 1.24401426e+00 5.59034050e-01
-4.35286909e-01 1.02640641e+00 5.54635644e-01 -6.18421793e-01
-8.54461849e-01 -8.75836134e-01 3.31334889e-01 9.22631741e-01
6.46239221e-01 -5.55629969e-01 -7.97292531e-01 -4.69621420e-01
-2.77349710e-01 -1.66788865e-02 -6.33332610e-01 -3.02614599e-01
-9.55223680e-01 -6.25703573e-01 3.11144561e-01 6.00766003e-01
1.07324755e+00 -9.29790616e-01 -2.94394135e-01 4.75232661e-01
-5.90195179e-01 -1.10725427e+00 -8.67345393e-01 -2.28013527e-02
-9.42841291e-01 -9.11712289e-01 -8.73709440e-01 -9.19163287e-01
3.16431433e-01 5.73791981e-01 9.81489241e-01 4.90971923e-01
7.17422739e-02 -2.55375445e-01 -5.42143404e-01 2.44891509e-01
1.48307994e-01 -1.50761023e-01 -4.17509079e-01 2.75399983e-01
1.92277417e-01 -8.66867959e-01 -1.07598352e+00 1.60206452e-01
-1.25018919e+00 1.09361231e-01 7.22678781e-01 9.52667236e-01
8.18884671e-01 5.04555106e-01 3.07809263e-01 -5.72601795e-01
5.06388187e-01 -4.23132718e-01 -3.65365803e-01 1.29045084e-01
-1.88929409e-01 -1.41467556e-01 8.80483091e-01 -2.70774215e-01
-1.21024275e+00 -1.06166638e-01 -4.28358138e-01 -8.24145257e-01
-2.86490079e-02 4.63990480e-01 -6.22081995e-01 -6.12746459e-03
-9.29322988e-02 5.72711706e-01 -2.45351478e-01 -8.71118248e-01
1.17066324e-01 3.96830201e-01 8.06838989e-01 -4.78460193e-01
8.53437006e-01 6.69891477e-01 -1.40280604e-01 -4.01782304e-01
-9.57470357e-01 -3.91438454e-01 -5.73508620e-01 -1.43169686e-01
1.11051965e+00 -1.40674818e+00 -7.07552433e-01 9.74654436e-01
-1.20994198e+00 -2.12173209e-01 -1.78344578e-01 2.91716367e-01
-1.75619334e-01 6.07940316e-01 -1.21732235e+00 -2.41523609e-01
-2.34199867e-01 -1.47565627e+00 1.32291734e+00 3.35832030e-01
5.32589793e-01 -6.60930514e-01 -3.14005882e-01 -1.25002069e-02
5.69162071e-01 1.98302224e-01 6.69655979e-01 2.51322478e-01
-9.03293967e-01 2.22817570e-01 -5.33472955e-01 7.97824979e-01
5.78714460e-02 -1.75074890e-01 -7.64009118e-01 -4.48988527e-01
3.35086554e-01 4.05145660e-02 1.35790086e+00 6.10511363e-01
1.52793789e+00 -1.98029399e-01 8.25908501e-03 1.19224477e+00
1.55312943e+00 1.50837064e-01 1.17356277e+00 3.09139073e-01
1.10909152e+00 1.44984722e-01 5.82914770e-01 4.17912543e-01
4.13433969e-01 4.22009021e-01 4.23185587e-01 -2.50559300e-01
-3.54800195e-01 -1.51894525e-01 4.22939837e-01 1.16917861e+00
-1.68708056e-01 -1.28065020e-01 -3.51010948e-01 4.74988371e-01
-1.91415310e+00 -1.01377082e+00 -1.91927254e-01 1.74831617e+00
6.83698297e-01 1.24695994e-01 -3.33982587e-01 1.15841389e-01
8.27084005e-01 7.51344979e-01 -5.42043209e-01 2.00011596e-01
-3.78702074e-01 1.30863577e-01 4.19098645e-01 4.26779091e-01
-1.28775430e+00 7.06088245e-01 5.42297077e+00 1.22503197e+00
-1.09415746e+00 2.29465082e-01 8.69596541e-01 -5.58285415e-02
-2.29343131e-01 -1.06924526e-01 -4.23926026e-01 7.70691574e-01
5.39084375e-01 2.59324461e-01 4.20625806e-01 3.47161144e-01
6.19533181e-01 3.42724174e-02 -6.21716619e-01 1.08053219e+00
-9.65384841e-02 -1.33924520e+00 1.41444966e-01 -2.49080241e-01
6.82861209e-01 -2.31357113e-01 -9.87738743e-02 1.83006674e-01
-1.65645510e-01 -8.39373231e-01 1.08716047e+00 9.17483926e-01
9.09515381e-01 -1.07268834e+00 9.08019900e-01 5.96700460e-02
-1.89132547e+00 -1.48180619e-01 -3.50071788e-01 2.06373613e-02
2.80574173e-01 8.38689327e-01 7.30086625e-01 8.83068860e-01
1.27411532e+00 1.46212792e+00 -5.15399277e-01 9.06504035e-01
-2.67426252e-01 5.05309582e-01 4.48664092e-02 5.99179685e-01
3.46937537e-01 -3.60995382e-01 3.88541996e-01 1.37413049e+00
3.46760839e-01 5.09010255e-01 2.37143949e-01 4.57685530e-01
-1.71712264e-01 -3.53590876e-01 6.16000332e-02 3.49495500e-01
2.60000646e-01 1.26414490e+00 -4.83654916e-01 -4.77906078e-01
-7.28103697e-01 1.26555336e+00 1.05513968e-02 6.25352204e-01
-8.73253226e-01 -5.91157317e-01 9.17136014e-01 1.35854557e-02
7.16752827e-01 -2.55545586e-01 -2.62897909e-01 -1.57094502e+00
4.35481846e-01 -1.03836846e+00 2.32016876e-01 -7.82447934e-01
-1.28508294e+00 6.55506432e-01 -3.62081528e-01 -1.46127915e+00
2.66502976e-01 -2.34139413e-01 -5.90529799e-01 1.08672202e+00
-1.94155729e+00 -1.18401408e+00 -6.99122667e-01 9.96690452e-01
6.85737789e-01 2.96414346e-01 1.68823183e-01 7.90242255e-01
-7.57810533e-01 3.31807733e-01 2.86968499e-01 3.88504595e-01
7.15781331e-01 -8.87173116e-01 3.46782744e-01 1.35800028e+00
-5.78582585e-01 6.52998686e-01 3.75493795e-01 -7.91658998e-01
-1.44139600e+00 -1.33485365e+00 5.87026179e-01 2.95775533e-01
4.48551416e-01 -2.51030922e-01 -1.20955515e+00 5.39355099e-01
1.87942043e-01 6.04541540e-01 -1.48058146e-01 -3.90712202e-01
-4.25719231e-01 -5.36534488e-01 -9.64432359e-01 3.72807056e-01
1.17568231e+00 -6.44153357e-01 -3.54857475e-01 -2.57165968e-01
9.76433814e-01 -5.29369116e-01 -8.51873457e-01 5.12679160e-01
4.75383162e-01 -1.41559649e+00 1.01415062e+00 -6.44427389e-02
7.24951923e-01 -8.17843020e-01 -2.31169477e-01 -1.06715822e+00
-6.23052537e-01 -6.70902610e-01 -3.17226440e-01 1.29823101e+00
-3.82699043e-01 -2.35888571e-01 4.83194768e-01 4.44533080e-02
-4.34500247e-01 -8.05786610e-01 -8.80064368e-01 -3.86173934e-01
-3.03650975e-01 -5.03842413e-01 5.73931813e-01 7.11785078e-01
-7.56422162e-01 2.45285615e-01 -7.11931050e-01 3.18382621e-01
6.91033244e-01 2.79934215e-03 3.45959514e-01 -8.71179521e-01
-1.35350883e-01 -2.45186776e-01 -3.68443549e-01 -1.64695859e+00
-3.98016945e-02 -2.27272615e-01 3.18411916e-01 -1.51427615e+00
3.95354599e-01 -1.75552800e-01 -5.83757818e-01 2.14335293e-01
-4.22802389e-01 4.67368305e-01 1.20817922e-01 2.64834106e-01
-8.61977279e-01 8.04737866e-01 1.56864047e+00 -9.29532498e-02
2.36139707e-02 -5.00297010e-01 -7.22246349e-01 8.62500966e-01
2.33177930e-01 -3.45152140e-01 -2.73412079e-01 -1.00960183e+00
-1.92254156e-01 2.08068103e-01 6.33397639e-01 -1.11483467e+00
3.46755028e-01 -2.16203839e-01 7.67509758e-01 -7.38120437e-01
3.16791028e-01 -8.78956974e-01 1.10628031e-01 3.09589744e-01
-1.49272457e-01 1.87780663e-01 -1.78572908e-02 9.73628461e-01
-8.56551588e-01 3.18916023e-01 9.86359656e-01 -1.99128106e-01
-9.70362484e-01 8.32185328e-01 -2.39273667e-01 -1.00026459e-01
6.82144761e-01 -6.82229549e-02 -2.33418718e-01 -3.56621206e-01
-7.33964384e-01 1.40872702e-01 5.12931287e-01 3.65881026e-01
8.99203062e-01 -1.40970051e+00 -7.31659114e-01 2.93835312e-01
-2.30325148e-01 2.76563764e-01 9.20255542e-01 9.67088699e-01
-6.87503278e-01 3.99136283e-02 -1.50731578e-01 -5.26292861e-01
-1.02805042e+00 5.46002507e-01 4.09877807e-01 -3.46620500e-01
-9.10709918e-01 8.51753354e-01 4.09491152e-01 2.34196261e-01
2.81090081e-01 -4.93099719e-01 -2.12021351e-01 -1.29047418e-02
8.89565110e-01 2.50920057e-01 -2.85563786e-02 -9.89184678e-01
-2.10949868e-01 7.77328372e-01 2.23994888e-02 1.60836592e-01
1.57634592e+00 -6.36500716e-01 -4.80945021e-01 1.96990043e-01
1.47688520e+00 -1.66760325e-01 -1.72274220e+00 -4.21394229e-01
-5.01841962e-01 -7.88051665e-01 4.37775701e-01 -4.18802321e-01
-1.77291751e+00 1.03121018e+00 5.95947862e-01 -5.92304058e-02
1.92166853e+00 -4.94900703e-01 1.33572853e+00 -6.26735464e-02
1.07998036e-01 -9.45031703e-01 1.10742949e-01 6.33996844e-01
7.72968411e-01 -1.02534020e+00 7.31560029e-03 -6.19857669e-01
-2.83647239e-01 1.19170582e+00 6.50876284e-01 -3.81730437e-01
7.68905699e-01 3.37267935e-01 2.90342403e-04 8.21028836e-03
-7.11976767e-01 -7.35299289e-02 2.19907671e-01 4.22567368e-01
3.72098565e-01 -3.75491589e-01 -2.88817555e-01 8.21678758e-01
5.12199461e-01 3.24830294e-01 3.05368811e-01 9.96283650e-01
-3.68590444e-01 -9.77407455e-01 -2.90207058e-01 3.17128479e-01
-6.12407267e-01 -4.03231919e-01 3.06642950e-01 4.01014209e-01
6.50488675e-01 1.08996022e+00 2.13232916e-02 -6.89258695e-01
5.54430425e-01 -5.50190508e-01 4.87807244e-02 -2.26240188e-01
-7.61906683e-01 6.26140237e-01 -2.54879266e-01 -1.11642194e+00
-7.48474061e-01 -5.94313562e-01 -1.04545891e+00 -5.08162439e-01
4.77112196e-02 -2.98963599e-02 8.32170397e-02 8.62357736e-01
3.41222703e-01 1.00262225e+00 6.43636763e-01 -1.08641958e+00
-6.51553124e-02 -8.53937566e-01 -6.90296888e-01 5.02054334e-01
8.79048049e-01 -3.60985368e-01 -2.89515644e-01 3.42352092e-01] | [11.215350151062012, -2.010462760925293] |
630fa2c0-e900-4e3e-8bd8-cf608ce91ccc | treec-a-method-to-generate-interpretable | 2304.08310 | null | https://arxiv.org/abs/2304.08310v1 | https://arxiv.org/pdf/2304.08310v1.pdf | TreeC: a method to generate interpretable energy management systems using a metaheuristic algorithm | Energy management systems (EMS) have classically been implemented based on rule-based control (RBC) and model predictive control (MPC) methods. Recent research are investigating reinforcement learning (RL) as a new promising approach. This paper introduces TreeC, a machine learning method that uses the metaheuristic algorithm covariance matrix adaptation evolution strategy (CMA-ES) to generate an interpretable EMS modeled as a decision tree. This method learns the decision strategy of the EMS based on historical data contrary to RBC and MPC approaches that are typically considered as non adaptive solutions. The decision strategy of the EMS is modeled as a decision tree and is thus interpretable contrary to RL which mainly uses black-box models (e.g. neural networks). The TreeC method is compared to RBC, MPC and RL strategies in two study cases taken from literature: (1) an electric grid case and (2) a household heating case. The results show that TreeC obtains close performances than MPC with perfect forecast in both cases and obtains similar performances to RL in the electrical grid case and outperforms RL in the household heating case. TreeC demonstrates a performant application of machine learning for energy management systems that is also fully interpretable. | ['Thierry Coosemans', 'Maarten Messagie', 'Muhammad Andy Putratama', 'Luis Ramirez Camargo', 'Julian Ruddick'] | 2023-04-17 | null | null | null | null | ['energy-management'] | ['time-series'] | [ 1.60875365e-01 3.10912609e-01 -6.10216334e-02 3.71515453e-02
-1.46659970e-01 -4.90539819e-01 9.42381144e-01 3.74969333e-01
-2.30463251e-01 1.22076321e+00 -2.22273812e-01 -6.04565263e-01
-4.95346546e-01 -8.93457055e-01 -4.34609056e-01 -1.11380696e+00
3.23745981e-02 6.83196664e-01 -2.33625293e-01 -4.45634544e-01
2.41069704e-01 5.83221257e-01 -1.66463709e+00 4.27647792e-02
1.08497071e+00 8.11172009e-01 3.53811920e-01 7.45453775e-01
8.47517699e-02 7.97290266e-01 -5.22607744e-01 3.94776851e-01
2.56820917e-01 -6.96803212e-01 -7.04851687e-01 -4.48116427e-03
-7.00988710e-01 1.73795491e-01 5.61071992e-01 7.83004582e-01
3.40174645e-01 5.50375760e-01 1.07316136e+00 -1.38659048e+00
1.10895429e-02 5.38594365e-01 -2.89101928e-01 -1.60010308e-02
1.96072116e-01 2.96503812e-01 4.66035724e-01 -4.12995331e-02
3.56110543e-01 1.07115257e+00 5.15894711e-01 2.95105666e-01
-1.33334684e+00 -1.27974272e-01 -2.09426507e-02 8.01734567e-01
-1.04863644e+00 1.64645225e-01 7.47425258e-01 -5.15478611e-01
1.50759304e+00 5.04760146e-01 1.10968971e+00 5.79234600e-01
3.85877401e-01 5.78032076e-01 1.56987858e+00 -9.32694614e-01
1.17888403e+00 3.77479434e-01 -1.63505197e-01 1.75737947e-01
2.15174213e-01 7.85669744e-01 2.37176210e-01 -4.96345274e-02
1.44723758e-01 -4.67913330e-01 -1.63704306e-01 -2.82556355e-01
-5.73733330e-01 1.24675012e+00 1.92680672e-01 6.16260231e-01
-1.00079858e+00 -1.74501926e-01 3.92440617e-01 2.84138203e-01
2.26787567e-01 7.33933449e-01 -7.23439395e-01 -5.76293468e-02
-9.07967746e-01 2.51542002e-01 1.24200690e+00 4.96421635e-01
4.94286776e-01 9.66515839e-01 4.36855704e-02 4.75597858e-01
2.76843995e-01 5.13910413e-01 9.68647957e-01 -8.20447385e-01
-2.14128613e-01 6.85806096e-01 3.11262935e-01 -7.31312215e-01
-6.64983511e-01 -2.43211895e-01 -8.40337336e-01 8.77230048e-01
-5.65121546e-02 -7.33690202e-01 -5.69486320e-01 1.28648949e+00
3.39902729e-01 -8.89248922e-02 4.44905102e-01 5.43280423e-01
2.37429470e-01 1.20469725e+00 7.42797926e-02 -8.39968562e-01
9.78502870e-01 -9.59566355e-01 -9.06576395e-01 2.70852923e-01
4.64201003e-01 -4.42996979e-01 1.93397328e-01 8.64343345e-01
-7.98409760e-01 -5.43770075e-01 -1.19246089e+00 9.48225200e-01
-9.81248677e-01 1.35532871e-01 2.11598083e-01 6.63479269e-01
-1.09080148e+00 1.05141866e+00 -7.84543037e-01 -6.79360569e-01
-1.84023589e-01 4.81450140e-01 1.82318687e-01 6.86302185e-01
-1.10791695e+00 1.53815544e+00 1.29306567e+00 1.17617315e-02
-4.38325405e-01 -3.69197875e-01 -8.24842751e-01 2.82663275e-02
5.07357657e-01 -4.31317061e-01 1.29603255e+00 -1.34432626e+00
-2.25966167e+00 1.38557464e-01 9.52760503e-02 -1.04478514e+00
5.84352136e-01 1.43384337e-01 -4.03555721e-01 4.36096750e-02
-5.62621713e-01 3.36242974e-01 8.99351239e-01 -1.52823806e+00
-9.54653382e-01 1.68160386e-02 -2.76405215e-01 8.12832490e-02
2.08276510e-01 -5.44045627e-01 8.03058326e-01 -4.64448035e-01
-3.90949488e-01 -1.00458014e+00 -3.24442834e-01 -9.08841848e-01
-2.49200672e-01 -5.83476245e-01 9.68407691e-01 -8.30527246e-01
1.41627431e+00 -1.11803997e+00 5.22771597e-01 5.65453470e-01
-7.11007535e-01 6.57302856e-01 3.90040964e-01 9.54369366e-01
-4.64736938e-01 -2.17002966e-02 -3.94107670e-01 2.23604783e-01
2.33881369e-01 7.62380242e-01 1.14504687e-01 1.96416631e-01
1.43130198e-01 5.83112359e-01 -7.77293563e-01 -1.67183369e-01
9.02805090e-01 1.33128658e-01 -1.30263309e-03 3.51680875e-01
-6.12125933e-01 3.74392956e-01 -4.88362610e-01 2.20411554e-01
4.67656493e-01 1.38513044e-01 5.78592777e-01 -2.99516231e-01
-4.71761227e-01 -6.33766353e-01 -1.41593289e+00 8.82758081e-01
-7.81349599e-01 4.68161017e-01 -5.21437004e-02 -1.79814315e+00
1.05521452e+00 6.09440863e-01 4.84054744e-01 -5.23934245e-01
2.82179356e-01 1.12433247e-01 -1.14520509e-02 -7.36730814e-01
6.27104864e-02 -2.43650541e-01 5.21835029e-01 3.53518695e-01
-2.77603567e-02 -4.72206742e-01 3.54822516e-01 -5.41202188e-01
7.62189388e-01 5.78840554e-01 1.09043634e+00 -6.29170597e-01
1.08292603e+00 2.11950704e-01 6.50094092e-01 4.12637323e-01
1.53079689e-01 -1.74220607e-01 2.34144732e-01 -5.65488338e-01
-1.08336997e+00 -4.03279215e-01 8.50923210e-02 4.63265121e-01
-2.56889880e-01 -1.94157679e-02 -7.52474725e-01 -4.62424099e-01
7.58409202e-02 1.81526363e+00 -5.24656475e-01 -2.79902399e-01
-4.63484317e-01 -9.87211168e-01 -3.01667005e-01 1.19678974e-01
5.98325253e-01 -1.15999210e+00 -1.20222116e+00 7.04342604e-01
2.29365617e-01 -7.20537901e-01 5.88016093e-01 5.81083357e-01
-8.93322110e-01 -1.21804488e+00 -2.91081846e-01 -2.82128632e-01
4.56327140e-01 -5.25452256e-01 1.18082404e+00 -8.88260752e-02
-2.83465236e-01 6.14098489e-01 -6.84060812e-01 -9.37137902e-01
-1.21288061e+00 4.14201394e-02 2.58116294e-02 -3.75471972e-02
2.09534124e-01 -6.22921467e-01 -3.57184052e-01 9.89305004e-02
-8.28293443e-01 2.05409437e-01 5.48304379e-01 1.02309489e+00
5.22117972e-01 9.13688481e-01 8.21755648e-01 -7.24839985e-01
7.41656244e-01 -5.90555787e-01 -1.20135438e+00 4.92140800e-01
-1.34357941e+00 2.81385690e-01 1.02188063e+00 -3.20134282e-01
-1.25690401e+00 2.76913106e-01 9.12635401e-02 -3.09743732e-01
-6.85933650e-01 1.76692873e-01 1.45511359e-01 -2.68156696e-02
3.02144021e-01 4.13534284e-01 1.15177557e-01 -4.92626846e-01
-1.49389151e-02 4.62165385e-01 2.05514655e-01 -4.77317989e-01
9.65530515e-01 -1.45748019e-01 4.50643092e-01 -8.50955188e-01
-1.13803968e-02 -1.26397163e-01 -5.56726694e-01 -4.83063549e-01
1.11039960e+00 -3.18643123e-01 -9.74809229e-01 2.90703923e-01
-8.98425877e-01 -4.58134800e-01 -5.00550270e-01 4.63339031e-01
-9.36950624e-01 -1.28008977e-01 -1.56290114e-01 -1.60406339e+00
-5.76671839e-01 -9.82148886e-01 5.19824207e-01 6.34198904e-01
-2.51591891e-01 -1.46672606e+00 1.23160794e-01 -1.36924133e-01
4.92670268e-01 9.70405400e-01 1.33522618e+00 -7.83179998e-01
-3.60296555e-02 -6.78142607e-02 6.16342008e-01 4.94052827e-01
5.11624925e-02 3.64528060e-01 -7.56829083e-01 -5.60354531e-01
1.37262121e-01 1.08228348e-01 1.85457319e-01 4.62004870e-01
6.50346100e-01 -5.36273539e-01 -2.98085779e-01 1.01577900e-01
2.07632470e+00 1.29837084e+00 4.61972296e-01 8.51523101e-01
-1.89015806e-01 6.99327052e-01 5.85514188e-01 6.29988909e-01
1.96847036e-01 7.92128265e-01 6.00772262e-01 -9.22717303e-02
6.22031629e-01 6.00297265e-02 2.89255559e-01 4.95230079e-01
-4.57390010e-01 -2.41794854e-01 -7.53538787e-01 9.66887102e-02
-2.20345926e+00 -1.11911762e+00 -6.00691065e-02 1.95173764e+00
4.64406759e-01 1.18394159e-01 2.74902552e-01 4.85273451e-01
6.28889561e-01 -3.93153638e-01 -5.13344705e-01 -1.38326418e+00
-1.15922205e-01 4.15867716e-01 5.53363919e-01 3.66590708e-01
-1.00674641e+00 2.54354417e-01 5.23151255e+00 9.01789367e-01
-9.81446207e-01 -1.93032846e-01 6.14458263e-01 2.68435031e-01
1.98956043e-01 -5.73838316e-02 -5.20460427e-01 6.74540401e-01
1.36001074e+00 -6.53523445e-01 7.99439430e-01 7.59744883e-01
8.40684772e-01 -7.10033596e-01 -9.75931287e-01 7.15183139e-01
-3.55890423e-01 -1.14696968e+00 -1.20679438e-01 -5.39472625e-02
1.17211139e+00 -5.01469076e-01 -5.04996717e-01 5.21310389e-01
4.49111968e-01 -8.70219231e-01 5.61871529e-01 8.73536170e-01
-1.43414363e-01 -1.09970450e+00 1.24442339e+00 6.73830211e-01
-1.03354275e+00 -7.20402837e-01 -2.20566228e-01 -2.70193140e-03
1.26862526e-01 1.59651428e-01 -6.80241823e-01 1.32580566e+00
7.01396704e-01 4.61815655e-01 -4.91694838e-01 8.95405233e-01
-7.94120058e-02 8.54970574e-01 -4.21183169e-01 -4.16859150e-01
5.34229100e-01 -7.59499907e-01 6.84707105e-01 1.07262850e+00
4.60501075e-01 2.13416368e-01 -2.24838536e-02 9.52241480e-01
8.83214355e-01 1.79680213e-01 -5.65090001e-01 1.04022466e-01
2.51770645e-01 1.27284765e+00 -7.15062439e-01 -4.50842321e-01
-9.65063497e-02 6.00562096e-01 -2.45565340e-01 2.99692869e-01
-6.00629866e-01 -1.86577052e-01 4.85527515e-02 -2.08518088e-01
6.28156304e-01 2.60018438e-01 5.42816445e-02 -4.84519124e-01
-4.85786259e-01 -7.99339175e-01 4.32949066e-01 -7.91195989e-01
-1.17330253e+00 5.37213862e-01 4.77019608e-01 -1.18479002e+00
-1.20542860e+00 -6.55243933e-01 -1.13616252e+00 8.20738971e-01
-1.42937672e+00 -5.64149618e-01 -9.34804305e-02 1.28379777e-01
7.26545453e-01 -5.89745164e-01 1.11989844e+00 -5.60673118e-01
-8.98822606e-01 -1.39429420e-01 9.68167067e-01 -6.61682665e-01
-1.96357578e-01 -1.76875150e+00 -5.02903759e-01 4.13273394e-01
-5.77764809e-01 -1.92285731e-01 1.10700262e+00 -3.99207443e-01
-1.46514440e+00 -8.91372144e-01 3.13654423e-01 3.07470411e-01
4.41325575e-01 2.89851695e-01 -8.82247627e-01 1.22974627e-01
1.02922845e+00 -7.05782712e-01 4.12376046e-01 -5.90532243e-01
6.96980774e-01 -1.95810437e-01 -1.57264733e+00 2.69505680e-01
-7.54339248e-02 2.43213251e-01 -7.90744305e-01 2.25573018e-01
2.98502874e-02 3.34039098e-04 -1.22674286e+00 5.43975055e-01
3.39013577e-01 -9.72179770e-01 5.60443819e-01 -5.10222793e-01
-2.48227213e-02 -3.15038532e-01 1.81657135e-01 -1.96039367e+00
-1.63482592e-01 -8.43895912e-01 -5.11265635e-01 1.18732047e+00
2.45040119e-01 -8.82545233e-01 4.38996434e-01 5.56822121e-01
1.35690972e-01 -1.04540062e+00 -9.61724579e-01 -1.02697217e+00
2.94514209e-01 8.31207931e-02 8.24693203e-01 8.40735137e-01
-1.72699407e-01 1.11864887e-01 -6.69073453e-03 5.91797978e-02
5.45018435e-01 3.17767799e-01 1.99950144e-01 -1.11833942e+00
-5.52814484e-01 -6.47983015e-01 -2.25097224e-01 2.48516023e-01
3.88582826e-01 -6.70091093e-01 -1.44953832e-01 -1.65569544e+00
-4.43733692e-01 -1.63576212e-02 -2.68554747e-01 3.42876941e-01
1.43182129e-01 -6.24515355e-01 3.42503458e-01 -3.55030000e-01
3.41047309e-02 6.32697940e-01 6.50371552e-01 -1.73718005e-01
-3.60254437e-01 4.60253239e-01 4.84956726e-02 8.60215724e-01
1.36098850e+00 -5.57010174e-02 -5.71906149e-01 4.73089844e-01
-3.19513492e-02 3.66884589e-01 1.54714957e-01 -1.05320191e+00
3.38145383e-02 -4.27666515e-01 5.45713603e-01 -6.97833180e-01
-2.00062022e-01 -1.34223878e+00 8.85448873e-01 1.16510201e+00
-6.41629621e-02 4.17217106e-01 4.58645612e-01 5.80157757e-01
-5.88496067e-02 -7.63529778e-01 9.47791815e-01 -3.11397433e-01
-9.27156925e-01 -4.66772586e-01 -1.04118562e+00 -4.96347010e-01
1.52338147e+00 -4.67092723e-01 9.57479402e-02 -4.88713145e-01
-9.16918457e-01 5.80940247e-01 1.09529108e-01 2.10298017e-01
2.07430854e-01 -1.16785824e+00 -7.02105999e-01 6.83575403e-04
-4.79115605e-01 -3.14310819e-01 -1.76496673e-02 5.83816350e-01
-5.73092461e-01 7.23063886e-01 -2.61884570e-01 -4.68568414e-01
-1.13076711e+00 6.88946605e-01 9.10451651e-01 -6.01327598e-01
-3.15600783e-01 -1.81398079e-01 -6.09345853e-01 -3.11039776e-01
-7.49263465e-02 -8.79029110e-02 -7.56380320e-01 3.08584303e-01
1.03840185e-03 9.19812799e-01 2.16549546e-01 -4.79482710e-01
7.60441571e-02 5.87068737e-01 6.59366310e-01 8.10643658e-02
1.66431499e+00 1.67870764e-02 3.22794612e-03 4.79312837e-01
9.07479942e-01 -8.86988163e-01 -1.02433228e+00 2.30597794e-01
5.03853202e-01 1.85264707e-01 3.17693144e-01 -1.28671145e+00
-9.44378793e-01 4.33959395e-01 1.12474585e+00 6.59673214e-01
1.60677695e+00 -7.92383134e-01 1.72162041e-01 5.92057467e-01
4.25441325e-01 -1.75475407e+00 -4.86361772e-01 1.70287460e-01
1.15737236e+00 -1.05959356e+00 2.45593175e-01 1.66003913e-01
-5.62534392e-01 1.71074343e+00 4.87674415e-01 -3.65841575e-02
7.89931238e-01 3.04227799e-01 -3.88093114e-01 3.18025798e-01
-9.88105536e-01 -3.77496332e-01 1.63520858e-01 6.67035818e-01
-4.69885692e-02 3.07838053e-01 -6.56369686e-01 3.82918268e-01
-1.13936044e-01 1.46601140e-01 4.35824841e-01 1.09100711e+00
-2.81924814e-01 -1.11607552e+00 -7.31820107e-01 3.96106720e-01
-1.54625690e-02 4.12887365e-01 -7.58302631e-03 1.27161932e+00
3.70345145e-01 1.21174860e+00 -4.30210494e-03 -2.23969728e-01
5.14342606e-01 2.58522242e-01 4.10937816e-01 -1.21232517e-01
-9.50565040e-01 2.31977254e-02 8.14395249e-02 -2.54233629e-01
-5.26705623e-01 -7.87064731e-01 -1.30871785e+00 -2.59228051e-01
-4.63935465e-01 6.69704556e-01 8.63144636e-01 1.04795754e+00
-6.17748611e-02 7.83721328e-01 9.84917402e-01 -1.10029328e+00
-8.07038724e-01 -8.48482370e-01 -6.48888350e-01 1.77138433e-01
-4.37009856e-02 -7.42696822e-01 -3.63556892e-01 4.89035025e-02] | [5.391025543212891, 2.3565211296081543] |
2c80c8a2-13d5-4502-8a97-2bd4a34dc2e3 | gradicon-approximate-diffeomorphisms-via-1 | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Tian_GradICON_Approximate_Diffeomorphisms_via_Gradient_Inverse_Consistency_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Tian_GradICON_Approximate_Diffeomorphisms_via_Gradient_Inverse_Consistency_CVPR_2023_paper.pdf | GradICON: Approximate Diffeomorphisms via Gradient Inverse Consistency | We present an approach to learning regular spatial transformations between image pairs in the context of medical image registration. Contrary to optimization-based registration techniques and many modern learning-based methods, we do not directly penalize transformation irregularities but instead promote transformation regularity via an inverse consistency penalty. We use a neural network to predict a map between a source and a target image as well as the map when swapping the source and target images. Different from existing approaches, we compose these two resulting maps and regularize deviations of the Jacobian of this composition from the identity matrix. This regularizer -- GradICON -- results in much better convergence when training registration models compared to promoting inverse consistency of the composition of maps directly while retaining the desirable implicit regularization effects of the latter. We achieve state-of-the-art registration performance on a variety of real-world medical image datasets using a single set of hyperparameters and a single non-dataset-specific training protocol. The code is available at https://github.com/uncbiag/ICON. | ['Marc Niethammer', 'Sylvain Bouix', 'Nikolaos Makris', 'Richard Jarrett Rushmore', 'Raúl San José Estépar', 'Roland Kwitt', 'François-Xavier Vialard', 'Hastings Greer', 'Lin Tian'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['image-registration', 'medical-image-registration'] | ['computer-vision', 'medical'] | [ 4.72379446e-01 2.96202838e-01 -1.04085423e-01 -5.68430185e-01
-1.07923329e+00 -2.80925661e-01 6.90243900e-01 2.63159364e-01
-7.17117786e-01 4.79679346e-01 2.92521387e-01 -1.78583339e-01
-1.30460352e-01 -5.45487046e-01 -7.43372500e-01 -7.94441283e-01
-1.53946817e-01 6.43349349e-01 5.93061633e-02 -1.51799396e-01
1.00357138e-01 3.09705883e-01 -8.03586960e-01 -9.56127420e-03
8.65519464e-01 5.91411173e-01 -4.61767241e-02 4.50568318e-01
4.23157066e-01 4.25908655e-01 -9.61266086e-02 -2.05308720e-01
6.48177803e-01 -4.82782036e-01 -1.02214932e+00 2.42825132e-02
7.65800238e-01 -1.21840753e-01 -3.17999452e-01 1.12592053e+00
6.84711754e-01 1.26331210e-01 5.53917050e-01 -7.66450047e-01
-6.38154566e-01 3.22396040e-01 -6.56050920e-01 2.71125853e-01
3.18244606e-01 -3.87058198e-03 7.21416175e-01 -6.39435530e-01
9.15899813e-01 6.95656896e-01 1.03560305e+00 5.03732026e-01
-1.62181759e+00 -3.09214294e-01 -2.91840762e-01 -1.60177648e-01
-1.42796111e+00 -6.50985718e-01 6.32084370e-01 -5.01693130e-01
6.60853028e-01 3.59584123e-01 5.04312336e-01 6.20866060e-01
6.79784179e-01 1.87068611e-01 1.15090024e+00 -4.97513324e-01
-1.09080307e-01 -7.44800568e-02 -1.62333846e-02 8.37075830e-01
2.50178706e-02 2.27968872e-01 -2.21521571e-01 -3.07106078e-01
9.41357911e-01 -1.15371577e-01 -5.29129088e-01 -6.30095780e-01
-1.59869683e+00 7.32893467e-01 6.68162704e-01 6.04453385e-01
-5.52842975e-01 1.14128925e-01 2.14944288e-01 2.41727769e-01
4.85239893e-01 5.39744735e-01 -1.94511488e-01 2.03448549e-01
-9.26925778e-01 5.24064414e-02 5.50774515e-01 4.58464712e-01
7.99792707e-01 -1.24191955e-01 -1.16506450e-01 8.18368793e-01
1.52242780e-01 2.43100852e-01 7.16410697e-01 -1.09977877e+00
2.44308442e-01 2.60142297e-01 -2.44741291e-01 -1.01308119e+00
-6.24919653e-01 -3.78562003e-01 -9.93919790e-01 2.84276843e-01
5.93130827e-01 -6.16889149e-02 -1.02598858e+00 1.90389299e+00
5.44488370e-01 3.96299839e-01 -2.89009988e-01 9.00093555e-01
7.18285799e-01 2.06696093e-01 -1.10986561e-01 -1.40969843e-01
1.33676505e+00 -8.20388436e-01 -6.60298884e-01 -1.91904739e-01
7.79773772e-01 -1.03663206e+00 9.36693430e-01 -3.82743701e-02
-1.31325912e+00 -2.47875288e-01 -9.44711566e-01 -1.76003173e-01
-8.18211734e-02 -9.11439583e-02 4.94642109e-01 3.44089717e-01
-1.43023419e+00 7.85044432e-01 -1.21493101e+00 -3.20988059e-01
3.31145316e-01 6.59307063e-01 -7.54470348e-01 1.05589263e-01
-8.25148523e-01 1.22485304e+00 1.56025574e-01 8.85122418e-02
-3.09105664e-01 -1.15485013e+00 -9.32586253e-01 -4.42070514e-01
8.01715776e-02 -7.08388567e-01 9.82377231e-01 -1.03413916e+00
-1.36887121e+00 1.33027041e+00 -1.59302786e-01 -1.92784652e-01
6.97959423e-01 -1.83320194e-02 -2.14580104e-01 -5.90771660e-02
3.95922333e-01 8.21773946e-01 5.55945516e-01 -1.13108110e+00
-9.78945494e-02 -2.82385945e-01 -3.89237791e-01 4.54545051e-01
-9.44098681e-02 2.31889896e-02 -6.77351058e-01 -7.71093369e-01
3.82142037e-01 -1.18613076e+00 -5.26639163e-01 2.24156037e-01
-4.61786568e-01 3.20027441e-01 4.14100349e-01 -8.69914770e-01
8.19058120e-01 -2.14658880e+00 1.55916229e-01 5.64632952e-01
2.57808685e-01 -3.83296311e-02 -3.53525192e-01 1.05755241e-03
-4.95841950e-01 -3.72625291e-01 -7.34973788e-01 -3.75975251e-01
-4.40927088e-01 1.66880131e-01 1.28290877e-01 1.11530733e+00
-1.72119081e-01 9.33992028e-01 -8.73351753e-01 -4.07822162e-01
2.37684622e-01 9.21677232e-01 -5.79912186e-01 6.51385263e-02
2.80509382e-01 1.05277586e+00 -2.68539131e-01 2.26627812e-01
6.78701401e-01 -3.20435166e-01 2.72547901e-01 -3.99380982e-01
1.41686738e-01 3.11536402e-01 -1.13412869e+00 1.96073854e+00
-3.18836004e-01 4.58401054e-01 1.11570038e-01 -1.05180991e+00
6.02773368e-01 5.83388507e-01 1.04326630e+00 -7.15428293e-01
2.83423632e-01 2.93002009e-01 2.65460312e-01 -1.88699529e-01
2.57838160e-01 -3.90441388e-01 2.48706967e-01 6.30099416e-01
1.09122358e-01 -2.31773451e-01 8.02015290e-02 -1.22365989e-01
8.72274816e-01 1.36525929e-01 4.33123708e-01 -6.38862967e-01
5.59753060e-01 -1.85167581e-01 5.44210196e-01 5.14023066e-01
-9.96384472e-02 1.05884647e+00 2.30816454e-01 -6.01953387e-01
-1.13827074e+00 -1.08118474e+00 -5.17631471e-01 7.12735057e-01
8.92414078e-02 -8.34323466e-02 -8.82891655e-01 -4.97717828e-01
-9.69973020e-03 3.47204834e-01 -8.27902377e-01 -1.10097580e-01
-9.99172747e-01 -1.17452216e+00 6.17489815e-01 2.34974429e-01
2.72389174e-01 -8.79875422e-01 -2.48813242e-01 9.82178748e-02
-3.17876399e-01 -9.50990319e-01 -1.06066835e+00 1.35893196e-01
-1.01217782e+00 -1.11985278e+00 -6.74432337e-01 -9.73358750e-01
1.20692110e+00 -5.13102449e-02 1.17194915e+00 2.97912568e-01
-3.26902479e-01 2.69943118e-01 1.85560927e-01 -5.13054915e-02
-6.15546823e-01 9.22483355e-02 4.49490696e-02 8.30041617e-02
-7.30485618e-02 -7.31338382e-01 -6.05444849e-01 3.46088529e-01
-9.87433553e-01 -1.48833543e-02 3.45133334e-01 9.75349426e-01
7.90712357e-01 -4.89207327e-01 -3.39357294e-02 -9.50253308e-01
6.42456412e-01 -2.54723489e-01 -5.21447301e-01 2.72931904e-01
-7.80375540e-01 2.11211175e-01 2.72776127e-01 -5.35963953e-01
-7.25584507e-01 1.75976738e-01 -2.24356189e-01 -1.64157614e-01
-1.67612091e-01 4.02652770e-01 3.24199349e-01 -4.79003131e-01
8.72427523e-01 1.03564933e-01 3.76166403e-01 -2.66842753e-01
3.39942366e-01 1.61037475e-01 8.13174069e-01 -5.94118714e-01
8.20489049e-01 7.14458287e-01 4.23746780e-02 -5.17097533e-01
-4.61863250e-01 -4.57700580e-01 -9.21916306e-01 2.60194298e-02
9.69674230e-01 -6.34857833e-01 -2.26788044e-01 4.54573840e-01
-1.07839429e+00 -4.82718766e-01 -5.40032566e-01 7.61102676e-01
-6.64588332e-01 3.78694862e-01 -5.86341739e-01 1.59992874e-01
-4.47524667e-01 -1.36617756e+00 9.23125505e-01 -5.01248091e-02
-3.05919766e-01 -1.35527050e+00 6.26710296e-01 2.09627777e-01
5.02019763e-01 3.08133215e-01 6.56242013e-01 -5.71621180e-01
-3.21371228e-01 -2.76278406e-01 3.24553140e-02 7.65942186e-02
4.71954286e-01 -3.48445863e-01 -6.10076308e-01 -4.66199040e-01
3.18285793e-01 3.78699452e-02 5.19443810e-01 7.36621022e-01
7.07372665e-01 -3.59287739e-01 -2.47997358e-01 1.11276293e+00
1.35196078e+00 1.48390576e-01 6.80525959e-01 4.61789519e-01
7.99914122e-01 4.32962358e-01 9.29066241e-02 -3.03718708e-02
3.20405453e-01 9.60207343e-01 6.88888133e-03 -5.95931649e-01
-2.43528828e-01 5.85268326e-02 -1.52120501e-01 8.55195522e-01
-3.38412136e-01 2.91597933e-01 -1.17992997e+00 5.73121488e-01
-1.70219934e+00 -6.66859448e-01 2.25594942e-03 2.37285638e+00
1.21829760e+00 -2.72086769e-01 -7.80928954e-02 -4.27470118e-01
6.58497453e-01 2.96357602e-01 -2.20432356e-01 -1.93989594e-02
6.40442595e-02 2.19799325e-01 7.84712911e-01 1.07838929e+00
-1.18166888e+00 7.34765410e-01 6.81823158e+00 4.86599118e-01
-1.47839642e+00 3.56507987e-01 7.00397968e-01 -6.62554651e-02
-2.22765073e-01 -6.75465092e-02 -2.38258466e-01 1.75816298e-01
6.87724292e-01 5.15551027e-03 6.27387762e-01 4.66515869e-01
1.13913074e-01 5.94417080e-02 -1.14472842e+00 7.51123309e-01
1.77540317e-01 -1.45908725e+00 -1.96467549e-01 2.05168217e-01
7.46395767e-01 3.63942146e-01 5.17029986e-02 -4.38834041e-01
3.14208329e-01 -1.10019124e+00 4.68127996e-01 4.75977987e-01
9.30662513e-01 -2.18529761e-01 6.96143448e-01 -8.15889090e-02
-8.95366073e-01 5.02036631e-01 -7.94077516e-02 3.51642847e-01
3.13724339e-01 5.05811393e-01 -7.82648802e-01 5.80267370e-01
5.41010976e-01 6.62480593e-01 -5.26539087e-01 1.00681710e+00
-1.03145808e-01 1.96416795e-01 -4.40415949e-01 7.91394949e-01
-1.40647041e-02 -6.17923915e-01 6.62818909e-01 1.02507317e+00
1.26067147e-01 -2.37068953e-03 7.48979151e-02 7.01336384e-01
-1.46978078e-02 2.89511442e-01 -6.15987480e-01 5.57283401e-01
1.04571387e-01 1.23797035e+00 -7.65738964e-01 -1.69671535e-01
-3.72592032e-01 9.76052642e-01 3.11205149e-01 2.80893177e-01
-6.83098197e-01 -4.75850552e-02 4.62992519e-01 3.26757908e-01
-1.36954607e-02 -1.21982761e-01 -5.72556853e-01 -1.22526610e+00
2.68386900e-01 -9.91557062e-01 4.08456743e-01 -4.14962620e-01
-1.13882339e+00 8.20891500e-01 1.95303932e-01 -1.23804426e+00
-2.39521384e-01 -2.74553806e-01 -7.49546528e-01 1.05372858e+00
-1.23844182e+00 -1.17700553e+00 -1.92020625e-01 5.78561604e-01
-5.80379553e-02 -1.63949654e-02 9.58662629e-01 5.36993206e-01
-3.09581220e-01 5.74328959e-01 1.82874069e-01 3.22421998e-01
1.06923270e+00 -1.08272088e+00 4.01973695e-01 8.31459284e-01
1.74794301e-01 8.65930140e-01 6.81535780e-01 -6.41398489e-01
-1.00274348e+00 -8.49772274e-01 9.76603687e-01 -4.82707083e-01
6.03228152e-01 -3.10635921e-02 -9.76492763e-01 1.07923079e+00
3.47017795e-01 4.89571691e-01 5.64423382e-01 -1.82823576e-02
-2.59618104e-01 -1.95529554e-02 -1.09319854e+00 6.92537487e-01
7.95979738e-01 -3.69897157e-01 -4.55125868e-01 4.96751159e-01
1.98099479e-01 -9.34057653e-01 -1.21528995e+00 4.76573467e-01
5.79763770e-01 -8.28584552e-01 1.24870133e+00 -4.67986196e-01
2.14839190e-01 -3.39190036e-01 1.52084097e-01 -1.36685741e+00
-4.38778400e-01 -6.08903706e-01 4.30818409e-01 7.17603624e-01
5.43014884e-01 -9.78010356e-01 7.52310514e-01 7.65261829e-01
-1.25805959e-01 -8.02291572e-01 -1.21991301e+00 -6.71308458e-01
3.47761571e-01 7.97586888e-02 3.77979875e-01 1.37332463e+00
5.19231660e-04 1.46231502e-02 -3.23197305e-01 1.25270680e-01
8.06581140e-01 -3.08165997e-01 6.49854422e-01 -8.03345501e-01
-2.78672129e-01 -5.68238795e-01 -4.73039061e-01 -5.90794981e-01
8.12109485e-02 -1.13782072e+00 2.83998519e-01 -1.32771027e+00
1.22877352e-01 -6.86218262e-01 -2.10073739e-01 7.97446489e-01
-2.34038338e-01 6.73588395e-01 -1.70240998e-02 5.31402946e-01
-1.26223817e-01 2.80408710e-01 1.32077658e+00 7.71165490e-02
-4.99088287e-01 -1.74925223e-01 -6.75204694e-01 6.37296021e-01
8.49911273e-01 -7.54042804e-01 -3.09180647e-01 -5.33131182e-01
-2.32494801e-01 6.03928305e-02 2.89274603e-01 -8.20321679e-01
2.88207084e-01 -1.04661301e-01 1.93025038e-01 6.03130646e-02
1.09511212e-01 -6.93336427e-01 5.92143416e-01 5.96018076e-01
-3.72590959e-01 2.30912894e-01 1.51612535e-01 8.27171504e-02
-1.46675333e-01 -9.41092074e-02 1.12194598e+00 -1.78737547e-02
-1.37096882e-01 4.49584395e-01 -1.13050290e-03 6.59997985e-02
7.09453523e-01 -1.91342697e-01 -2.60272741e-01 -2.74664313e-01
-7.04238892e-01 -1.00630380e-01 7.08808243e-01 2.72263736e-01
3.51786673e-01 -1.45439339e+00 -9.27573204e-01 3.59609723e-01
-1.42038375e-01 1.73044633e-02 8.89025703e-02 1.55218697e+00
-7.72755206e-01 1.14127472e-01 -3.52557182e-01 -7.48021066e-01
-1.35809445e+00 2.10405901e-01 9.04432416e-01 -5.70175767e-01
-9.54548717e-01 5.80152810e-01 4.21602339e-01 -1.03762805e+00
-1.28047109e-01 4.15946208e-02 5.40202409e-02 -3.49890620e-01
2.73911029e-01 9.09418017e-02 3.07210624e-01 -1.04096317e+00
-5.06409764e-01 7.85678685e-01 -2.06307083e-01 -3.64864171e-01
1.37588501e+00 1.44694205e-02 -4.26965863e-01 -2.30857613e-03
1.42108154e+00 5.52842878e-02 -1.17490041e+00 -4.12336737e-01
-6.26505911e-02 -4.29955125e-01 1.28244147e-01 -5.95404983e-01
-1.33698452e+00 3.18710536e-01 8.04498851e-01 -1.34187803e-01
9.64968681e-01 -1.95138678e-02 5.57036042e-01 4.77829725e-02
-5.55640971e-03 -7.94685066e-01 -2.12586552e-01 3.53868902e-01
9.40929651e-01 -1.26062298e+00 4.42384958e-01 -4.83494639e-01
-5.45821905e-01 7.82824337e-01 2.50653356e-01 -5.04204988e-01
8.25653076e-01 4.30533469e-01 6.08035207e-01 -2.54897922e-01
-1.27053291e-01 2.16287345e-01 7.95234203e-01 6.78129137e-01
8.27628136e-01 -1.26808867e-01 -4.12107885e-01 -2.63535857e-01
-2.50614196e-01 -4.10088599e-02 2.51749814e-01 1.02360463e+00
1.59324348e-01 -1.43607700e+00 -5.34750700e-01 4.14382279e-01
-5.62588751e-01 -3.06851178e-01 1.82765666e-02 8.05982649e-01
-5.82795218e-02 3.31410617e-01 1.96671501e-01 -7.50033781e-02
2.63491273e-01 -1.45412982e-01 6.63403928e-01 -5.41321993e-01
-8.08526576e-01 3.65589648e-01 -7.52718523e-02 -6.95856094e-01
-6.72591567e-01 -8.99064898e-01 -1.18030322e+00 -3.02259654e-01
-2.14912891e-01 4.60270457e-02 4.81632054e-01 9.21658754e-01
3.18503827e-01 3.11834008e-01 3.59980285e-01 -1.08174062e+00
-3.88552785e-01 -7.41977572e-01 -3.64230394e-01 7.74872065e-01
4.47255641e-01 -4.21431392e-01 -1.32264093e-01 3.02863568e-01] | [14.005461692810059, -2.578321695327759] |
e1a706f5-9b59-4424-a6b1-c2256f147439 | can-natural-language-processing-become | null | null | https://aclanthology.org/P15-1120 | https://aclanthology.org/P15-1120.pdf | Can Natural Language Processing Become Natural Language Coaching? | null | ['Marti A. Hearst'] | 2015-07-01 | can-natural-language-processing-become-1 | https://aclanthology.org/P15-1120 | https://aclanthology.org/P15-1120.pdf | ijcnlp-2015-7 | ['grammatical-error-detection'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.3713765144348145, 3.7093167304992676] |
e5d3422e-73f7-4d1d-ab0f-a6c44e4cf535 | scenerf-self-supervised-monocular-3d-scene | 2212.02501 | null | https://arxiv.org/abs/2212.02501v3 | https://arxiv.org/pdf/2212.02501v3.pdf | SceneRF: Self-Supervised Monocular 3D Scene Reconstruction with Radiance Fields | 3D reconstruction from 2D image was extensively studied, training with depth supervision. To relax the dependence to costly-acquired datasets, we propose SceneRF, a self-supervised monocular scene reconstruction method using only posed image sequences for training. Fueled by the recent progress in neural radiance fields (NeRF) we optimize a radiance field though with explicit depth optimization and a novel probabilistic sampling strategy to efficiently handle large scenes. At inference, a single input image suffices to hallucinate novel depth views which are fused together to obtain 3D scene reconstruction. Thorough experiments demonstrate that we outperform all recent baselines for novel depth views synthesis and scene reconstruction, on indoor BundleFusion and outdoor SemanticKITTI. Our code is available at https://astra-vision.github.io/SceneRF. | ['Raoul de Charette', 'Anh-Quan Cao'] | 2022-12-05 | null | null | null | null | ['3d-scene-reconstruction', 'image-to-video'] | ['computer-vision', 'computer-vision'] | [ 4.21435505e-01 1.57320537e-02 2.12723091e-02 -7.16272235e-01
-8.52117717e-01 -6.81071401e-01 6.93808198e-01 -5.07268131e-01
-1.72913760e-01 7.96513021e-01 5.68003595e-01 -1.08659498e-01
3.53902280e-01 -8.76825869e-01 -1.14691222e+00 -5.81809163e-01
5.45100152e-01 2.80169725e-01 -1.25657603e-01 1.23862758e-01
6.43721372e-02 6.21851683e-01 -1.53070271e+00 3.48478436e-01
4.67752635e-01 7.81166375e-01 8.49272966e-01 1.05022538e+00
-6.58134818e-02 1.06314433e+00 -3.51940934e-03 -1.53576210e-01
5.88988721e-01 -3.52418721e-01 -7.80835211e-01 4.21256453e-01
8.98266315e-01 -9.90200043e-01 -7.63914049e-01 8.16252112e-01
6.33370996e-01 2.11778462e-01 3.22994441e-01 -8.20991755e-01
-4.07947093e-01 1.60737813e-01 -2.04683751e-01 -9.52706560e-02
8.03349376e-01 3.64202380e-01 8.69158387e-01 -1.26802981e+00
8.92220676e-01 1.17300928e+00 6.89833760e-01 5.90553343e-01
-1.39312065e+00 -3.08934242e-01 4.09211917e-03 -7.46786371e-02
-1.14967656e+00 -6.86673462e-01 8.26353252e-01 -2.69911289e-01
1.22144234e+00 1.75609291e-01 6.05998099e-01 1.40187633e+00
-8.53503719e-02 8.20193648e-01 1.27503908e+00 -2.14105785e-01
1.04384668e-01 -5.67160826e-03 -3.24396998e-01 7.30760634e-01
-1.50108665e-01 4.95758891e-01 -7.38334358e-01 1.82964563e-01
1.03379571e+00 2.31554478e-01 -5.77575147e-01 -6.70072675e-01
-1.32705879e+00 7.15876281e-01 6.43048346e-01 -2.03412995e-01
-3.27140480e-01 5.72138190e-01 -3.08991410e-02 1.35365725e-01
5.81867635e-01 3.84178400e-01 -5.89016378e-01 -2.47090049e-02
-7.27931261e-01 4.59643364e-01 4.82852966e-01 1.01819491e+00
1.17828166e+00 1.80885732e-01 2.57382840e-01 6.78104222e-01
4.03339118e-01 9.04241562e-01 3.16806277e-03 -1.70012093e+00
3.52283388e-01 1.53578565e-01 9.24307704e-02 -4.56562102e-01
-4.00486648e-01 -3.95537972e-01 -6.47341371e-01 1.48589447e-01
2.04727292e-01 4.00466472e-02 -9.96475518e-01 1.39836979e+00
5.18102646e-01 3.74846429e-01 1.70161903e-01 9.59661782e-01
1.18033648e+00 6.24145627e-01 -4.64055300e-01 2.47698441e-01
7.99904168e-01 -1.22427964e+00 -3.80579352e-01 -3.59596193e-01
4.91106331e-01 -8.44613016e-01 1.04063272e+00 4.98822391e-01
-1.11351430e+00 -5.80871105e-01 -8.37566376e-01 -7.24985659e-01
-7.13980347e-02 -2.04197422e-01 9.17548060e-01 5.12693584e-01
-1.24148583e+00 5.11272430e-01 -6.66682363e-01 -2.66620696e-01
6.77551150e-01 -9.78149399e-02 -6.19370997e-01 -6.57506227e-01
-6.98671281e-01 6.34831071e-01 3.50961715e-01 -3.04400269e-02
-1.48545074e+00 -9.10843253e-01 -1.23279154e+00 -4.93534803e-01
1.87127367e-01 -1.40613329e+00 1.46624303e+00 -5.94396532e-01
-1.59943438e+00 1.11964464e+00 -1.91569746e-01 -5.07595003e-01
5.22440374e-01 -4.67168093e-01 1.93618238e-01 4.38123256e-01
2.62145996e-01 1.19016707e+00 6.50176346e-01 -1.64886367e+00
-2.90279299e-01 -3.62019241e-01 3.67261529e-01 7.33734429e-01
5.31680286e-01 -5.01594841e-01 -4.15772676e-01 -3.18652034e-01
4.02795702e-01 -6.85729802e-01 -4.73733693e-01 2.34924808e-01
-5.04743755e-01 4.52861071e-01 5.40102482e-01 -5.68042338e-01
2.03725368e-01 -1.94561768e+00 3.63180667e-01 -2.62738675e-01
9.94831920e-02 -4.35218990e-01 -7.66303986e-02 3.39161187e-01
-1.05123804e-03 -3.43079597e-01 -5.29015720e-01 -1.12342525e+00
-5.93285412e-02 5.07390022e-01 -5.58453083e-01 6.47778809e-01
-1.13970995e-01 1.03886807e+00 -1.04567218e+00 -1.58208832e-01
9.67863142e-01 6.13687873e-01 -9.74714220e-01 5.82095146e-01
-5.27734399e-01 1.13019717e+00 -1.80533439e-01 7.68090844e-01
1.06088030e+00 -4.21918750e-01 -1.46678627e-01 -2.12406918e-01
-1.71252042e-01 5.02000451e-01 -8.60332906e-01 2.90412283e+00
-6.93333745e-01 5.79369068e-01 1.08676970e-01 -7.42651641e-01
6.37275577e-01 4.80944067e-02 5.30123591e-01 -6.62835658e-01
-5.74665703e-02 5.58783561e-02 -9.81245935e-01 -3.59914005e-01
5.47598839e-01 -4.18585122e-01 1.25120178e-01 1.05529234e-01
4.15926278e-01 -1.14844954e+00 -4.18630421e-01 3.15409750e-01
1.05256450e+00 9.72742796e-01 1.69545095e-02 1.15510747e-01
1.55460507e-01 1.92482606e-01 1.81106061e-01 8.02550614e-01
1.25792816e-01 1.24921906e+00 -2.90444970e-01 -3.75056446e-01
-1.16182768e+00 -1.67632651e+00 -3.82408977e-01 5.08349538e-01
3.15848827e-01 -2.94424981e-01 -4.16983694e-01 -4.52376038e-01
-6.47319481e-02 8.65302980e-01 -3.30160558e-01 3.01465571e-01
-3.76999587e-01 -4.59903061e-01 3.52669954e-01 1.81520775e-01
7.80329645e-01 -8.87061000e-01 -7.56769001e-01 3.07656601e-02
-4.14318860e-01 -1.51461244e+00 -1.32442757e-01 4.00738329e-01
-9.61977124e-01 -8.35576475e-01 -7.73900270e-01 -4.21892166e-01
2.99608648e-01 8.80888760e-01 1.46073091e+00 -2.95052320e-01
-3.50904942e-01 8.74741077e-01 -2.09556878e-01 -3.90664935e-02
-1.57404274e-01 -2.66686767e-01 -1.94768429e-01 -2.97769219e-01
-1.92530602e-01 -9.24514592e-01 -8.86002958e-01 -2.27902420e-02
-8.29153359e-01 7.38942504e-01 2.46960402e-01 7.18276858e-01
8.18637013e-01 -2.98927814e-01 -6.30775690e-02 -9.00699615e-01
-4.17474359e-01 -6.24799430e-01 -6.04337096e-01 -2.70937592e-01
-1.10613056e-01 -1.09331615e-01 5.07089436e-01 2.71750480e-01
-1.43246722e+00 3.70446235e-01 -4.47320879e-01 -7.26395488e-01
-6.44029140e-01 -3.22268158e-02 -2.32132450e-01 -4.98776771e-02
8.03228199e-01 5.25258958e-01 -1.84107214e-01 -5.94323695e-01
7.57200658e-01 1.98310718e-01 7.63180494e-01 -4.72880095e-01
8.43393385e-01 1.07181072e+00 1.70372985e-02 -8.13815057e-01
-1.37021720e+00 -6.43586218e-01 -7.61348724e-01 -8.25334936e-02
1.09856868e+00 -1.74457085e+00 -3.93721789e-01 4.93116617e-01
-1.23929060e+00 -8.77797723e-01 -4.63896275e-01 5.57257056e-01
-1.10328496e+00 4.28327143e-01 -4.19093311e-01 -6.21674538e-01
-4.90461290e-02 -9.93406653e-01 1.76018548e+00 -4.43676896e-02
1.45149544e-01 -1.05437493e+00 2.22463593e-01 8.44262898e-01
4.73220088e-02 4.27250385e-01 1.35137141e-01 4.83315200e-01
-1.37325168e+00 4.20264781e-01 -3.38122427e-01 4.81233895e-01
1.83502883e-01 -5.56798279e-01 -1.57593274e+00 -9.71951261e-02
1.96958646e-01 -5.69409490e-01 1.02652621e+00 4.74925786e-01
1.18989146e+00 -2.58528162e-02 -4.95650657e-02 1.48877823e+00
1.73473859e+00 -2.59157389e-01 7.06202447e-01 2.38307282e-01
1.20225644e+00 4.90942240e-01 4.25760716e-01 5.16909301e-01
7.40660250e-01 6.71508253e-01 7.19921529e-01 -1.91476896e-01
-7.08765090e-01 -7.18251586e-01 3.23783517e-01 4.94681567e-01
1.90107822e-01 -2.26540774e-01 -6.91304326e-01 4.40053105e-01
-1.51466393e+00 -9.69838262e-01 -1.64627701e-01 1.87464654e+00
7.11430669e-01 -1.91475198e-01 -3.13340992e-01 -1.41789377e-01
-5.83965145e-02 5.02409637e-01 -7.23130584e-01 7.70130157e-02
-6.45226002e-01 3.52967411e-01 7.52264321e-01 9.34465647e-01
-8.51798296e-01 1.15714371e+00 6.26156759e+00 4.40786093e-01
-8.92710388e-01 2.08816707e-01 5.30409098e-01 -3.46812308e-01
-9.42068934e-01 2.15290710e-01 -6.18964851e-01 -3.26612517e-02
6.12274468e-01 3.74909967e-01 7.23351836e-01 7.56311655e-01
1.81988209e-01 -3.90938789e-01 -1.03930151e+00 1.51133478e+00
9.78112668e-02 -1.49037337e+00 -4.95009404e-03 -7.12779686e-02
1.05940771e+00 7.43881762e-01 1.36065530e-02 -6.91900551e-02
4.85999376e-01 -1.01285398e+00 7.87436903e-01 6.37426853e-01
8.44598770e-01 -3.47120911e-01 1.08264193e-01 3.15429419e-01
-8.83235693e-01 1.47970527e-01 -4.67613459e-01 -2.20024437e-01
5.41631579e-01 8.09942544e-01 -9.45921004e-01 9.36332643e-01
9.19292450e-01 1.37723088e+00 -3.64647180e-01 7.00803995e-01
-4.10080880e-01 2.28129327e-01 -4.22273219e-01 6.43476844e-01
1.86362118e-01 -2.01742187e-01 5.48129976e-01 8.72989535e-01
3.22744668e-01 6.45760596e-02 1.13374852e-01 1.10659564e+00
-7.43490085e-02 -2.92069495e-01 -1.16108477e+00 4.57242876e-01
5.16250022e-02 1.05749249e+00 -2.36018091e-01 -3.00858408e-01
-3.94014060e-01 1.55445707e+00 2.31404617e-01 5.54532766e-01
-7.32104123e-01 3.40164632e-01 6.99309826e-01 1.10647656e-01
3.30197453e-01 -5.62843859e-01 -5.27356565e-01 -1.63868117e+00
-1.46204516e-01 -4.45808470e-01 -1.73425097e-02 -1.48231161e+00
-1.12231421e+00 2.87764728e-01 1.40121073e-01 -1.11120439e+00
-2.11066097e-01 -5.81690371e-01 -1.23534322e-01 7.78185964e-01
-1.98708344e+00 -1.20192039e+00 -6.84636176e-01 9.11518633e-01
8.06504667e-01 3.00401002e-01 8.41764152e-01 2.00817823e-01
2.58401874e-02 -5.49779758e-02 -1.26240581e-01 -4.16762978e-01
5.55073202e-01 -1.44025362e+00 7.65659332e-01 8.10491860e-01
3.76876056e-01 2.19598755e-01 7.27508008e-01 -4.40708280e-01
-1.66187263e+00 -1.15864229e+00 4.37326699e-01 -8.70114803e-01
1.96689218e-01 -6.10080540e-01 -4.65413034e-01 7.88422763e-01
2.83737242e-01 2.59628445e-01 5.46657026e-01 -3.31647187e-01
-4.44072455e-01 1.48345485e-01 -1.16375828e+00 5.34906089e-01
1.73196602e+00 -8.58948112e-01 -2.62940317e-01 6.39917612e-01
1.04506147e+00 -7.60756195e-01 -8.45416188e-01 5.48488796e-01
3.37364525e-01 -1.60691357e+00 1.47452259e+00 5.45921139e-02
5.64171374e-01 -4.94426578e-01 -1.00572860e+00 -1.11039221e+00
2.16088649e-02 -5.33500552e-01 -1.52510822e-01 5.25274873e-01
2.45543364e-02 -5.13477504e-01 8.65970135e-01 2.00054601e-01
-5.22863328e-01 -3.59088480e-01 -7.26743698e-01 -4.23419923e-01
-1.24265939e-01 -6.65698528e-01 4.38722372e-01 8.03665996e-01
-8.45416546e-01 4.38869864e-01 -5.52093029e-01 2.80444920e-01
1.11306560e+00 3.77250433e-01 1.28333724e+00 -6.91238582e-01
-8.48847926e-01 2.07182337e-02 -1.61543787e-01 -1.71790624e+00
2.37408593e-01 -9.70745265e-01 1.43502459e-01 -1.79576564e+00
1.13950714e-01 -3.14083755e-01 2.27765650e-01 2.10206032e-01
1.10856093e-01 6.59147620e-01 1.00312836e-01 1.49661094e-01
-5.53752542e-01 9.05548632e-01 1.49384356e+00 1.66881129e-01
-2.46546548e-02 -1.72754988e-01 -4.16370243e-01 6.98851526e-01
7.44484484e-01 -2.24252552e-01 -6.56438231e-01 -8.26893210e-01
2.44325027e-01 3.35078120e-01 8.45417619e-01 -1.00120127e+00
-1.52664870e-01 -9.50412601e-02 6.92538798e-01 -9.32105780e-01
1.01393282e+00 -7.66228735e-01 4.54401642e-01 1.05696715e-01
-1.07030105e-02 -3.06444257e-01 1.23252869e-01 5.81675529e-01
-4.67850007e-02 6.38736086e-03 6.50526464e-01 -6.45259261e-01
-9.54908967e-01 5.74219465e-01 4.30425256e-02 -1.97406188e-01
5.47981799e-01 -3.37589562e-01 -1.60522535e-01 -5.22927284e-01
-7.63127744e-01 -5.43817058e-02 1.00687361e+00 3.21099252e-01
9.35624003e-01 -1.28103304e+00 -6.43113494e-01 2.78968662e-01
1.85785770e-01 7.14562535e-01 6.73586369e-01 4.51066256e-01
-8.87951672e-01 3.32113743e-01 -2.45683361e-02 -9.82141495e-01
-9.05398369e-01 3.80386055e-01 5.71097374e-01 1.53642386e-01
-1.00597155e+00 1.04918027e+00 7.79806316e-01 -1.14057529e+00
-5.38753644e-02 -3.90488446e-01 4.72501725e-01 -6.14467978e-01
2.35249490e-01 7.90814832e-02 -1.30988717e-01 -5.42439520e-01
-1.35288671e-01 7.50793993e-01 4.29159552e-01 -3.60457361e-01
1.41604412e+00 -6.55579984e-01 1.86679289e-02 6.57978237e-01
1.45989454e+00 -2.13201833e-03 -1.88407815e+00 -3.54817718e-01
-5.95384240e-01 -8.61591518e-01 3.09970796e-01 -6.87505066e-01
-7.64618933e-01 9.52816427e-01 4.61396009e-01 -4.91835684e-01
1.15193200e+00 9.03192237e-02 8.37334216e-01 4.43143547e-01
7.50688076e-01 -4.54646349e-01 1.79965451e-01 6.00062728e-01
8.48169506e-01 -1.54130816e+00 1.36946425e-01 -5.00474572e-01
-4.15060431e-01 8.99932206e-01 3.83830547e-01 -3.26848149e-01
6.61739051e-01 1.77326113e-01 -9.96823423e-03 -2.58531004e-01
-6.40195727e-01 -2.88990885e-01 1.85139328e-01 5.96811354e-01
2.29714915e-01 1.21830322e-01 7.55946398e-01 -1.52904674e-01
-5.04657745e-01 -2.61478070e-02 5.00757694e-01 6.82799101e-01
-3.11831862e-01 -9.24155891e-01 -2.04140931e-01 7.83113670e-03
-4.23615351e-02 -4.10722643e-01 -6.35187104e-02 3.76583368e-01
1.04598537e-01 8.22801530e-01 5.82010224e-02 -1.56819746e-01
1.95613265e-01 -3.27730805e-01 1.13263917e+00 -9.10675228e-01
-1.87362790e-01 6.91224337e-02 1.18249774e-01 -1.02118790e+00
-7.70564914e-01 -6.97043240e-01 -1.17916846e+00 -1.87459812e-01
4.41823825e-02 -4.92847234e-01 8.12025249e-01 6.82582557e-01
2.68693596e-01 3.44711185e-01 7.64774740e-01 -1.34852362e+00
2.00673137e-02 -6.75132334e-01 -3.55936050e-01 2.51675636e-01
6.04119301e-01 -5.85746288e-01 -4.51199234e-01 1.17607929e-01] | [8.98232364654541, -2.97721266746521] |
75626f5e-0fb2-4048-a553-6fd504e954ff | making-sense-of-violence-risk-predictions | 2204.13976 | null | https://arxiv.org/abs/2204.13976v1 | https://arxiv.org/pdf/2204.13976v1.pdf | Making sense of violence risk predictions using clinical notes | Violence risk assessment in psychiatric institutions enables interventions to avoid violence incidents. Clinical notes written by practitioners and available in electronic health records (EHR) are valuable resources that are seldom used to their full potential. Previous studies have attempted to assess violence risk in psychiatric patients using such notes, with acceptable performance. However, they do not explain why classification works and how it can be improved. We explore two methods to better understand the quality of a classifier in the context of clinical note analysis: random forests using topic models, and choice of evaluation metric. These methods allow us to understand both our data and our methodology more profoundly, setting up the groundwork to work on improved models that build upon this understanding. This is particularly important when it comes to the generalizability of evaluated classifiers to new data, a trustworthiness problem that is of great interest due to the increased availability of new data in electronic format. | ['Marco Spruit', 'Floortje Scheepers', 'Uzay Kaymak', 'Kalliopi Zervanou', 'Emil Rijcken', 'Pablo Mosteiro'] | 2022-04-29 | null | null | null | null | ['topic-models'] | ['natural-language-processing'] | [ 2.10368618e-01 1.90678686e-01 -2.50066996e-01 -4.79979247e-01
-8.77504408e-01 -6.24232471e-01 4.03281271e-01 6.92964852e-01
-8.63567770e-01 8.38678956e-01 8.06558132e-01 -6.24313772e-01
-5.16160905e-01 -7.73311317e-01 1.06805630e-01 -4.10695046e-01
1.10687204e-01 6.08173490e-01 -1.39574707e-01 5.01862429e-02
6.32915676e-01 7.09146559e-01 -7.32490182e-01 4.23977554e-01
6.06136322e-01 2.81531215e-01 -2.63465911e-01 2.63527453e-01
8.26777518e-02 9.67312157e-01 -5.14668643e-01 -8.69775891e-01
4.36222553e-02 -3.87371719e-01 -1.07344627e+00 -1.75071478e-01
7.31794164e-02 -7.88410187e-01 -6.12809032e-04 6.36621833e-01
5.41708171e-01 -4.97201532e-02 8.52110267e-01 -1.13231981e+00
-3.37993264e-01 7.40286231e-01 -8.64377022e-02 5.84547520e-01
3.38850260e-01 3.07610244e-01 8.56186509e-01 -2.94935852e-01
9.62528586e-01 9.16134119e-01 8.18827212e-01 5.16100407e-01
-1.23848486e+00 -9.38278615e-01 -3.44455451e-01 2.71288723e-01
-1.04475307e+00 -6.60023034e-01 2.70017087e-01 -9.09967005e-01
8.89052570e-01 2.21294254e-01 8.91251981e-01 1.36126339e+00
4.03761417e-01 2.03687493e-02 1.26902890e+00 -2.30412111e-01
1.75918952e-01 4.08639997e-01 6.18321419e-01 2.61330217e-01
9.82368171e-01 -1.30403727e-01 -4.46573436e-01 -7.62847364e-01
6.52897477e-01 2.01347739e-01 -2.48537675e-01 1.15500636e-01
-8.60172212e-01 1.11197460e+00 -2.18664259e-02 5.27012765e-01
-4.12078083e-01 -2.01515794e-01 7.22231030e-01 -3.03842244e-03
3.70610625e-01 5.96755624e-01 -1.28462523e-01 -8.10972393e-01
-1.21312296e+00 2.53510475e-01 9.09273803e-01 2.36206561e-01
-1.66877210e-01 -3.78773361e-01 -1.25341341e-01 7.12366104e-01
3.04845959e-01 2.54426032e-01 1.31249756e-01 -9.57567692e-01
5.73978603e-01 4.54773128e-01 -4.86453064e-02 -9.38024104e-01
-5.36553264e-01 -1.30966023e-01 -4.77027476e-01 2.04360068e-01
6.79225326e-01 -2.05086902e-01 -4.04342383e-01 1.28168702e+00
1.25945762e-01 -2.30110571e-01 3.34718376e-02 5.41949928e-01
4.95059788e-01 -5.40128164e-02 4.26630408e-01 -4.87921774e-01
1.41271722e+00 -2.45445639e-01 -8.96381319e-01 -2.71730535e-02
1.11997914e+00 -6.22008860e-01 5.22174835e-01 5.56156337e-01
-9.51547146e-01 3.88492018e-01 -5.41940212e-01 2.94906199e-02
4.40115109e-03 -5.05013227e-01 5.60816228e-01 1.11555171e+00
-8.13757956e-01 7.50143290e-01 -1.14490247e+00 -7.43110359e-01
9.22720313e-01 2.32017130e-01 -5.59795678e-01 -1.32822394e-01
-9.02546048e-01 1.38302386e+00 -4.94341515e-02 -6.35834560e-02
-1.42379045e-01 -5.78270733e-01 -6.20010555e-01 2.92551741e-02
2.88860798e-01 -6.73141479e-01 1.24027574e+00 -4.17873859e-01
-6.10589266e-01 8.14883590e-01 -7.39641907e-03 -3.54099214e-01
5.77260733e-01 1.60592943e-01 -7.29042068e-02 3.08047831e-01
1.46106273e-01 -1.16852470e-01 2.43973672e-01 -8.84199619e-01
-3.96053970e-01 -1.00815248e+00 -3.76910418e-01 -1.15626670e-01
-3.12228769e-01 6.73865795e-01 4.28641289e-01 -4.69131261e-01
-1.07837670e-01 -7.27220058e-01 -3.52287352e-01 -7.82869160e-02
-1.34487182e-01 -2.94936001e-02 3.73366922e-01 -9.48569775e-01
1.39707124e+00 -1.73154902e+00 -5.56966662e-01 2.94690222e-01
4.59605545e-01 2.66258866e-01 6.59026265e-01 8.22064698e-01
1.16969191e-01 7.55667925e-01 -2.38928854e-01 -6.41827956e-02
-4.14155751e-01 1.06705483e-02 -7.18846917e-02 5.98834157e-01
7.86910430e-02 5.82632124e-01 -7.36963630e-01 -8.01381230e-01
3.29038575e-02 5.68836033e-01 -6.10313058e-01 -1.40992686e-01
8.14240873e-01 2.44496822e-01 -6.02686942e-01 4.61216003e-01
3.80437195e-01 -2.32259072e-02 5.17870426e-01 4.35975224e-01
-1.16515219e-01 6.75265789e-01 -7.64815867e-01 7.72455990e-01
2.71717217e-02 6.55885041e-01 1.09331496e-01 -7.77617395e-01
8.59112620e-01 4.83907163e-01 6.54063284e-01 -2.15982601e-01
3.08763951e-01 9.84904990e-02 2.01626733e-01 -7.53516257e-01
3.58775377e-01 -8.61413896e-01 3.96321982e-01 7.78512478e-01
-3.05778921e-01 -1.42460885e-02 -2.03310072e-01 1.39044091e-01
1.31956291e+00 -4.73188937e-01 5.80411971e-01 -2.02235565e-01
-1.97266713e-01 2.68532962e-01 5.77124894e-01 6.95984304e-01
-4.92933989e-01 6.06940269e-01 7.08005905e-01 -4.04924870e-01
-9.54751313e-01 -4.29759860e-01 -1.02144086e+00 3.97557348e-01
-6.21208489e-01 -5.50911546e-01 -7.37718523e-01 -5.53100824e-01
-1.05496980e-01 1.07160032e+00 -6.80993915e-01 -1.67439416e-01
-2.11322501e-01 -9.84452069e-01 7.66197860e-01 6.29372120e-01
-9.40413550e-02 -1.08196497e+00 -1.34488392e+00 4.30180550e-01
-1.84283063e-01 -9.26927388e-01 5.22708148e-02 2.22421527e-01
-1.28792298e+00 -1.40967965e+00 -4.75696027e-01 -8.34513679e-02
1.94815308e-01 -6.11436404e-02 7.31599748e-01 3.36760461e-01
-2.96877772e-01 5.12404382e-01 -5.12237132e-01 -7.24591672e-01
-5.02535284e-01 -5.44146523e-02 2.36586877e-03 -5.24639368e-01
8.91710818e-01 -4.05925900e-01 -4.09722537e-01 -1.72927454e-01
-9.72156227e-01 -4.30409700e-01 2.58898228e-01 7.62942493e-01
-6.24438114e-02 -3.47425610e-01 3.93610388e-01 -1.35960281e+00
9.20135558e-01 -8.01028609e-01 6.16968907e-02 9.09283161e-02
-1.13180411e+00 -2.64840394e-01 1.79323778e-02 4.54625227e-02
-7.27251351e-01 -6.08082771e-01 -1.98689610e-01 5.61303720e-02
-3.03202868e-01 6.83807075e-01 3.33637595e-01 1.54002547e-01
8.72894943e-01 -5.23808360e-01 3.29380065e-01 -4.47761148e-01
-3.27494115e-01 1.03844905e+00 -1.18005186e-01 -2.66818434e-01
2.82906741e-01 5.33815980e-01 -2.76282784e-02 -8.13187599e-01
-2.57736027e-01 -5.90697467e-01 -7.15004981e-01 4.99149784e-03
8.50908518e-01 -4.47230339e-01 -7.23469496e-01 4.29068059e-02
-8.00892174e-01 -3.56006145e-01 -9.20248851e-02 8.63593519e-01
-3.82828534e-01 2.10925177e-01 -5.78364253e-01 -1.26899338e+00
-5.22881031e-01 -1.22272134e+00 8.48448098e-01 -2.63045579e-01
-8.27329695e-01 -1.09878314e+00 4.40093219e-01 7.21062183e-01
2.81164318e-01 3.35994959e-01 1.05560470e+00 -1.18506145e+00
-1.88713241e-02 -3.11554968e-01 -2.64959596e-02 4.59059887e-02
1.74827904e-01 1.90113381e-01 -8.79588366e-01 -1.03430763e-01
3.54659885e-01 -2.47941643e-01 6.66698396e-01 4.96375740e-01
7.04787195e-01 -5.35169184e-01 -5.44037700e-01 2.17107117e-01
1.25459659e+00 5.01844525e-01 6.22776270e-01 8.38174760e-01
4.35428262e-01 9.26526368e-01 4.09350514e-01 7.43330121e-01
4.63524342e-01 6.79672837e-01 -1.20167442e-01 1.99846327e-01
3.82712603e-01 7.75515586e-02 -7.00328201e-02 4.44459021e-01
-3.68878543e-01 1.43168971e-01 -1.73908067e+00 5.05497396e-01
-1.54890072e+00 -1.30673647e+00 -3.39530498e-01 2.26772237e+00
6.14315450e-01 1.51736304e-01 4.82506514e-01 4.25269455e-01
4.83346701e-01 -2.51711875e-01 -1.72333986e-01 -7.42352188e-01
4.40073967e-01 1.73124433e-01 3.31814170e-01 1.74897566e-01
-5.26634336e-01 3.50365222e-01 6.91609383e+00 1.07332237e-01
-8.78539383e-01 1.89135864e-01 8.91634107e-01 -2.72152960e-01
-3.35013717e-01 6.12305738e-02 -5.86903155e-01 3.85839701e-01
1.30813468e+00 -1.42567039e-01 2.18872666e-01 7.32200682e-01
6.04316175e-01 -2.38135949e-01 -9.62813079e-01 8.60812247e-01
1.05147891e-01 -1.13199484e+00 -3.16228211e-01 5.97039521e-01
5.70401996e-02 -1.39445350e-01 -1.26570851e-01 7.14542419e-02
2.46678680e-01 -1.22482133e+00 4.93977815e-01 5.97104430e-01
5.80501556e-01 -6.90994203e-01 1.15287936e+00 1.86748296e-01
-3.34453046e-01 -1.26298577e-01 -1.93174228e-01 -2.18035907e-01
2.24819437e-01 4.02991176e-01 -1.48949325e+00 6.40695021e-02
8.29468608e-01 3.68048519e-01 -4.13594425e-01 1.18016136e+00
1.70984089e-01 1.02901542e+00 -4.06164341e-02 1.10792220e-02
9.37209502e-02 -1.61250606e-01 4.43477571e-01 1.12979829e+00
1.94527656e-01 6.25039637e-01 -3.10980082e-01 5.41349351e-01
1.96129188e-01 4.49267775e-01 -1.00754702e+00 -3.10685664e-01
4.75525677e-01 1.07140553e+00 -1.04129016e+00 3.39835919e-02
-6.07422054e-01 3.22779298e-01 3.99631560e-01 -1.83747169e-02
-4.80683558e-02 1.05824709e-01 6.49959803e-01 7.54093111e-01
-3.17404211e-01 -5.92027865e-02 -6.73900604e-01 -9.46910739e-01
-2.97595829e-01 -1.20835233e+00 7.09444046e-01 -4.98235554e-01
-1.29755068e+00 4.94242609e-01 2.00164303e-01 -7.64859855e-01
-4.35874552e-01 -4.22693163e-01 -3.25243086e-01 6.41180634e-01
-8.71997654e-01 -8.45213473e-01 -2.82370932e-02 3.57502311e-01
1.16980515e-01 1.56315640e-01 9.65003848e-01 -2.73125708e-01
-5.80158532e-01 5.29501855e-01 3.21225315e-01 3.15072119e-01
7.94058681e-01 -9.35553193e-01 2.24015750e-02 4.11382765e-01
1.20323654e-02 9.09019113e-01 4.69612598e-01 -9.45134997e-01
-8.54003549e-01 -5.85041940e-01 9.69931006e-01 -1.02761638e+00
5.15455544e-01 -4.53918912e-02 -1.07926714e+00 8.57787549e-01
-6.72261938e-02 -6.43480837e-01 1.47508848e+00 6.52204096e-01
-2.19616547e-01 3.88333589e-01 -1.48418438e+00 4.46548760e-01
8.37623477e-01 -5.67472756e-01 -9.80330229e-01 2.40402430e-01
6.13562986e-02 8.72966796e-02 -1.23033774e+00 8.85458067e-02
8.37222159e-01 -1.05394053e+00 5.25613129e-01 -8.72735381e-01
6.50414765e-01 7.57264376e-01 7.23853111e-02 -1.09075141e+00
-3.34052444e-01 -2.70978481e-01 6.15598321e-01 1.19343019e+00
4.49452966e-01 -8.18182826e-01 6.88627720e-01 1.64522207e+00
3.81757081e-01 -8.85582209e-01 -8.88323367e-01 -3.95212859e-01
4.76349980e-01 -6.29278243e-01 5.53766787e-01 1.46877468e+00
6.99839950e-01 5.93274720e-02 6.36928156e-02 -1.84318900e-01
3.63146007e-01 -2.52128094e-01 3.57410342e-01 -1.51699519e+00
-2.35338649e-03 -3.90499651e-01 -8.47023070e-01 4.45454657e-01
-1.65994138e-01 -6.82298899e-01 -4.61516559e-01 -1.80273366e+00
8.91291618e-01 -6.20549321e-01 -9.12909582e-02 6.62235796e-01
-3.80420119e-01 -4.74217767e-03 4.79653627e-01 5.28742194e-01
1.17860651e-02 4.76000197e-02 8.42552662e-01 3.45961004e-01
-3.88898700e-01 -2.19232887e-02 -1.25918746e+00 6.72374964e-01
1.03162110e+00 -1.08566391e+00 -2.32949093e-01 -1.33335203e-01
1.00242808e-01 2.91286111e-01 3.33653390e-01 -7.26872802e-01
2.01453909e-01 -3.92856240e-01 2.97834307e-01 -6.20227382e-02
2.21441537e-01 -8.53015602e-01 1.74557999e-01 4.26436514e-01
-4.83061820e-01 2.11660594e-01 1.73879877e-01 2.88577408e-01
8.42245817e-02 -8.28764081e-01 4.28022504e-01 -3.14972997e-01
2.16911629e-01 -3.56288515e-02 -7.86876082e-01 2.58403003e-01
8.50286067e-01 -4.23319012e-01 -3.56078923e-01 -6.23440564e-01
-8.25441599e-01 -1.60155892e-01 6.71832442e-01 -5.60938604e-02
4.63448048e-01 -8.99932027e-01 -7.57724524e-01 -1.36809319e-01
3.50255556e-02 -4.64310199e-01 7.06457421e-02 1.12391043e+00
-5.64752221e-01 6.50722206e-01 -2.86142826e-01 -3.35114240e-03
-1.45699406e+00 6.40596569e-01 8.57464895e-02 -2.73690134e-01
-8.87466371e-01 3.09486061e-01 -1.64361730e-01 4.71575446e-02
8.46501067e-02 -1.19299874e-01 -5.24512231e-01 3.08817506e-01
8.34062517e-01 5.89276850e-01 2.16653660e-01 -7.47247040e-01
-4.99554515e-01 -4.27723303e-02 -3.17203641e-01 -5.76872170e-01
1.80880666e+00 1.10116169e-01 -1.11909375e-01 4.81612146e-01
7.85787582e-01 1.65444463e-01 -5.53184569e-01 3.71027350e-01
2.38909379e-01 -7.25995302e-01 1.91396490e-01 -6.65407717e-01
-6.10757172e-01 9.22713637e-01 2.85402387e-01 2.63792962e-01
7.39938915e-01 -1.66959256e-01 2.84001112e-01 2.47452885e-01
3.70570421e-01 -8.27445924e-01 -2.38806918e-01 -3.00363488e-02
6.39386475e-01 -1.04475224e+00 9.33861956e-02 -2.63103694e-01
-9.29100871e-01 1.11735225e+00 9.29118544e-02 3.43892485e-01
7.04670906e-01 2.95275807e-01 1.81658208e-01 -4.72498715e-01
-7.16823459e-01 9.17806476e-02 -3.30300301e-01 8.72180164e-01
6.11779630e-01 1.17597610e-01 -9.13193882e-01 9.33867574e-01
-4.00947690e-01 2.27902412e-01 9.80039299e-01 1.11413705e+00
-2.41645083e-01 -1.26084960e+00 -6.75758302e-01 1.16155529e+00
-1.09847820e+00 -2.98646390e-01 -7.58842409e-01 8.82541597e-01
-1.94770768e-01 1.21267056e+00 -1.09133095e-01 -2.77532756e-01
2.32495904e-01 4.12185401e-01 2.11493403e-01 -8.07267785e-01
-9.44969654e-01 -1.15110777e-01 4.61388320e-01 -3.26092392e-01
-1.68977097e-01 -1.35140598e+00 -8.08454812e-01 -4.85872746e-01
-4.59912300e-01 3.51103812e-01 6.80933595e-01 1.02049232e+00
1.68636411e-01 2.03407839e-01 5.89457415e-02 -9.93350074e-02
-7.39399672e-01 -8.63492131e-01 -6.89660549e-01 2.61662453e-01
7.24001229e-02 -5.33527374e-01 -2.96419501e-01 -2.06497535e-01] | [8.22862720489502, 5.858125686645508] |
65970858-dbb4-45c1-90af-130e89df0ef6 | an-advanced-combination-of-semi-supervised | 2207.10777 | null | https://arxiv.org/abs/2207.10777v1 | https://arxiv.org/pdf/2207.10777v1.pdf | An advanced combination of semi-supervised Normalizing Flow & Yolo (YoloNF) to detect and recognize vehicle license plates | Fully Automatic License Plate Recognition (ALPR) has been a frequent research topic due to several practical applications. However, many of the current solutions are still not robust enough in real situations, commonly depending on many constraints. This paper presents a robust and efficient ALPR system based on the state-of-the-art YOLO object detector and Normalizing flows. The model uses two new strategies. Firstly, a two-stage network using YOLO and a normalization flow-based model for normalization to detect Licenses Plates (LP) and recognize the LP with numbers and Arabic characters. Secondly, Multi-scale image transformations are implemented to provide a solution to the problem of the YOLO cropped LP detection including significant background noise. Furthermore, extensive experiments are led on a new dataset with realistic scenarios, we introduce a larger public annotated dataset collected from Moroccan plates. We demonstrate that our proposed model can learn on a small number of samples free of single or multiple characters. The dataset will also be made publicly available to encourage further studies and research on plate detection and recognition. | ['Xinyi Dai', 'Khalid Oublal'] | 2022-07-21 | null | null | null | null | ['license-plate-recognition'] | ['computer-vision'] | [ 1.26214489e-01 -7.01959014e-01 9.75338593e-02 -2.94239014e-01
-8.79848301e-01 -5.85711300e-01 4.49275881e-01 -4.92576361e-01
-4.31753963e-01 5.52962542e-01 -2.41196826e-01 5.16653620e-02
4.71655071e-01 -5.92590690e-01 -6.72457695e-01 -7.34266698e-01
3.91767979e-01 3.63031089e-01 7.81104207e-01 -3.23519289e-01
6.62632287e-01 8.35916877e-01 -1.39618802e+00 1.50610879e-01
7.81770051e-01 6.62778497e-01 1.74877271e-02 7.46338725e-01
3.58590670e-02 7.73963273e-01 -7.27680266e-01 -5.72836518e-01
8.62822413e-01 -3.46245617e-01 -2.00774789e-01 6.49980426e-01
5.14949441e-01 -3.24396253e-01 -4.26770449e-01 1.25818014e+00
3.85896474e-01 3.86519670e-01 8.63485098e-01 -1.17420757e+00
-7.67370164e-01 8.90529901e-02 -7.59283900e-01 2.35394448e-01
-9.38738417e-03 2.24966079e-01 5.06494224e-01 -1.17945910e+00
5.48445642e-01 1.01524031e+00 8.66136551e-01 4.71097022e-01
-5.93310833e-01 -6.46552980e-01 -2.21477285e-01 3.50680023e-01
-1.59309626e+00 -5.42036235e-01 7.72906482e-01 -4.65049416e-01
6.92423403e-01 3.35405618e-01 2.72237837e-01 5.82393467e-01
-5.69384359e-03 8.14104319e-01 1.23045564e+00 -6.64866328e-01
5.54515310e-02 4.60815758e-01 2.74781078e-01 7.11884201e-01
2.94210315e-01 -4.90402520e-01 -2.58799374e-01 4.23278272e-01
9.23117280e-01 -5.87890111e-02 -4.89035621e-02 -5.28463759e-02
-7.34248936e-01 7.67117679e-01 -1.07047275e-01 3.41829509e-01
3.15542147e-02 -2.57595390e-01 1.93293512e-01 -1.17068745e-01
3.20119888e-01 2.76809990e-01 -1.90832987e-02 3.80536318e-02
-1.41187191e+00 6.06071129e-02 7.65775442e-01 1.06723702e+00
7.87627339e-01 3.15153569e-01 9.38566253e-02 1.18197882e+00
4.59246933e-01 6.93637431e-01 6.29682600e-01 -5.29225826e-01
7.88555503e-01 6.19945467e-01 8.99627060e-03 -1.28475940e+00
-1.35427460e-01 -2.58568488e-03 -6.73497975e-01 4.67815638e-01
6.82460248e-01 -4.46506962e-02 -1.00355697e+00 7.85071671e-01
-1.48272872e-01 2.89148629e-01 1.08506702e-01 1.01060307e+00
6.80749953e-01 1.19377589e+00 -3.68837506e-01 -5.13219857e-04
1.45256960e+00 -1.30532908e+00 -7.58819997e-01 -4.08790827e-01
4.43200827e-01 -1.27915406e+00 8.30103159e-01 5.15377820e-01
-8.75635505e-01 -4.79756951e-01 -1.31660807e+00 -1.07143104e-01
-7.57867575e-01 9.88289356e-01 7.14189261e-02 1.15048265e+00
-8.01073015e-01 2.05381624e-02 -5.52059293e-01 -6.79146051e-01
3.19252759e-01 5.52234709e-01 -5.67253411e-01 -3.04980755e-01
-6.84596062e-01 1.09650528e+00 3.22236210e-01 6.03812575e-01
-5.38100183e-01 2.31180508e-02 -7.84357190e-01 -1.37051523e-01
4.86637831e-01 2.92828739e-01 7.31370807e-01 -1.04302847e+00
-1.77680469e+00 9.11471784e-01 4.57456447e-02 -2.12537691e-01
7.27306962e-01 -2.04032362e-01 -7.91800678e-01 5.24317771e-02
-5.56737110e-02 2.19614699e-01 7.77733088e-01 -9.96166050e-01
-6.94515288e-01 -2.83431113e-01 -4.35307205e-01 4.04900908e-01
-3.37319791e-01 6.40440822e-01 -9.67024267e-01 -7.01955259e-01
7.59891421e-02 -9.83504534e-01 9.77645144e-02 -1.88053325e-01
-5.98978698e-01 9.48298201e-02 9.32676017e-01 -1.12415171e+00
9.51726854e-01 -2.20385766e+00 -4.95971888e-01 3.18847239e-01
-2.95078516e-01 7.50838876e-01 -6.26388118e-02 2.22377926e-01
7.59984031e-02 4.87775169e-02 -4.35580522e-01 -2.53851801e-01
-1.11980990e-01 -1.38776481e-01 -1.12221293e-01 9.33397174e-01
4.26087052e-01 4.09998208e-01 -1.04532726e-01 -4.09596890e-01
4.52976525e-01 3.16175550e-01 -1.50977373e-01 6.59141839e-02
3.10171217e-01 -1.88735604e-01 -7.20629245e-02 1.08101428e+00
1.07689464e+00 3.17554355e-01 -3.21258605e-01 -1.12249382e-01
-4.37866956e-01 -4.72034991e-01 -1.68364060e+00 7.71602273e-01
3.03999763e-02 1.12404454e+00 1.81715712e-01 -7.91081369e-01
1.45718288e+00 6.14647102e-03 4.56367955e-02 -4.96156275e-01
2.94920176e-01 4.55603808e-01 -1.34953126e-01 -6.13767982e-01
8.74773204e-01 4.95907366e-02 2.49524843e-02 -8.98422971e-02
-2.99243666e-02 -6.15849644e-02 7.11399853e-01 -1.64426848e-01
6.46693647e-01 5.74268475e-02 2.03416690e-01 -7.87206963e-02
1.21611834e+00 1.86116830e-01 6.77511930e-01 6.77676022e-01
-4.34993237e-01 7.98395813e-01 2.14088723e-01 -3.12079042e-01
-1.23090875e+00 -6.30265176e-01 -1.17262332e-02 9.63243484e-01
3.19460660e-01 2.58216023e-01 -8.66668761e-01 -4.19913918e-01
-2.29700208e-01 4.60463554e-01 -3.03793043e-01 2.77956635e-01
-1.01357877e+00 -1.27717113e+00 9.22410965e-01 3.75863433e-01
1.01336515e+00 -9.39752102e-01 -8.08325037e-02 8.67497642e-03
1.40090749e-01 -1.42579067e+00 -5.92601776e-01 -1.91617489e-01
-5.89516938e-01 -1.03365684e+00 -1.05540752e+00 -1.18907404e+00
8.17545474e-01 4.47145492e-01 4.16995853e-01 -1.56920493e-01
-2.30152979e-01 7.82669038e-02 -1.76278576e-01 -2.72519052e-01
-6.93169713e-01 -1.30263194e-01 -5.70730530e-02 4.91418391e-01
7.00845003e-01 1.86605722e-01 -6.50755642e-03 8.49052608e-01
-9.81818914e-01 -1.59589186e-01 9.50423419e-01 4.45137739e-01
2.49039978e-01 2.07754016e-01 3.11732084e-01 -8.46127033e-01
6.11995399e-01 -1.05358407e-01 -1.18499780e+00 4.30163950e-01
-2.59743154e-01 -3.49287570e-01 7.30520844e-01 -2.95474857e-01
-1.31555355e+00 4.81112659e-01 -2.19265372e-01 -1.92588985e-01
-5.60145795e-01 3.51496637e-02 -3.56216788e-01 -5.33597887e-01
3.32814127e-01 7.08369315e-01 -1.15628071e-01 -3.73463333e-01
2.51019388e-01 1.07889700e+00 7.68144488e-01 -1.79875702e-01
1.19189179e+00 3.66957337e-01 -2.02748895e-01 -1.31338108e+00
-1.65715203e-01 -9.84916091e-01 -9.03313756e-01 -5.28001189e-01
9.59734619e-01 -7.54714966e-01 -4.98951107e-01 1.22245812e+00
-1.00007749e+00 6.22052923e-02 2.52818257e-01 5.82072139e-01
-2.94305354e-01 5.95180511e-01 -8.26706290e-01 -9.40871537e-01
-2.03961525e-02 -1.36356246e+00 6.75241590e-01 6.23375535e-01
6.17544174e-01 -7.04405665e-01 2.49911204e-01 6.51931107e-01
3.82605255e-01 -1.11190349e-01 3.38291973e-01 -1.05290270e+00
-8.21112394e-01 -6.92828357e-01 -4.66880560e-01 7.03253150e-01
-1.29666878e-03 4.93891716e-01 -9.16121423e-01 -4.57012504e-02
-1.54642969e-01 -1.88292742e-01 7.98918128e-01 1.65243179e-01
4.44666654e-01 -1.95387334e-01 4.93554771e-02 5.35329580e-01
1.71695590e+00 4.33005840e-01 1.03556263e+00 7.13952124e-01
7.54865468e-01 3.20890844e-01 4.83287901e-01 1.57176614e-01
8.48663077e-02 6.44930303e-01 -3.89204696e-02 -1.02693655e-01
-1.63289979e-01 1.10862948e-01 6.81922853e-01 1.03088164e+00
-3.80054504e-01 -3.57494235e-01 -1.17281258e+00 2.37111032e-01
-1.50374639e+00 -9.98589575e-01 -3.53147894e-01 2.00503230e+00
4.03707683e-01 6.26235008e-02 2.22969368e-01 1.57000527e-01
1.25147808e+00 -1.01220496e-01 -1.50745556e-01 -3.21329236e-01
-5.90022981e-01 -3.32941443e-01 1.05129445e+00 3.45381141e-01
-1.52085996e+00 1.21972299e+00 6.05089378e+00 8.60675752e-01
-1.36513329e+00 -1.56701088e-01 7.54183233e-01 5.52088618e-01
5.81634939e-01 -4.20854717e-01 -1.30454099e+00 3.83903861e-01
5.87228179e-01 9.50233638e-02 3.36413443e-01 8.60451221e-01
1.02680989e-01 -1.05388798e-01 -4.60660070e-01 1.25592065e+00
8.48944545e-01 -1.21378160e+00 -1.52382746e-01 1.21126212e-02
6.58285737e-01 -9.22019407e-02 5.52446721e-03 3.24908018e-01
5.74332140e-02 -8.15416336e-01 7.08630800e-01 5.20891368e-01
5.65438211e-01 -7.54376769e-01 1.13615382e+00 2.38892138e-01
-1.06433809e+00 -1.06877692e-01 -7.21409202e-01 2.49708280e-01
-1.03228867e-01 6.78450763e-02 -9.06603336e-01 2.78159231e-01
2.63451934e-01 7.23284602e-01 -1.05589342e+00 1.44753814e+00
-3.91148031e-02 6.73704922e-01 -3.19555670e-01 -2.37918362e-01
3.97682190e-01 -4.65231687e-01 4.32178378e-01 1.68994284e+00
4.73911017e-01 -1.17030300e-01 -6.61124364e-02 6.00704432e-01
-1.97048917e-01 6.12737894e-01 -4.57482666e-01 6.95001185e-02
9.54723358e-02 1.48601425e+00 -1.25006568e+00 -1.45603582e-01
-7.80692995e-01 9.86809313e-01 -1.82743087e-01 2.59602427e-01
-8.96472573e-01 -5.88046610e-01 9.04357210e-02 1.01025142e-01
3.60708684e-01 -3.21158320e-01 -8.03609416e-02 -1.45784044e+00
-1.89029165e-02 -8.96072805e-01 2.82506824e-01 -6.90696299e-01
-1.22397006e+00 4.01434809e-01 -2.10280448e-01 -1.50660086e+00
1.22352920e-01 -1.27894402e+00 -6.83850229e-01 7.63969839e-01
-1.65150416e+00 -1.10445321e+00 -2.17574507e-01 5.93274057e-01
9.74993587e-01 -7.92326152e-01 4.89387959e-01 4.67001557e-01
-1.19642282e+00 5.65267026e-01 6.66617274e-01 8.64951313e-01
9.45962429e-01 -8.91388178e-01 1.44117370e-01 1.63858330e+00
4.24071103e-02 2.72589386e-01 4.86067832e-01 -5.16951025e-01
-1.24801612e+00 -1.04619193e+00 5.67250788e-01 -1.21176943e-01
5.28678775e-01 -6.25637174e-01 -1.00390017e+00 7.11880088e-01
1.53249234e-01 7.97411650e-02 6.74914539e-01 -6.60034895e-01
-1.56102598e-01 -3.62730443e-01 -1.25848186e+00 4.39758241e-01
7.44002983e-02 3.30514833e-02 -6.20832980e-01 2.93321162e-01
-1.83407754e-01 -2.53715456e-01 -3.67064714e-01 -7.02702478e-02
4.13218051e-01 -8.18368196e-01 8.19020689e-01 -1.15648694e-02
2.30733439e-01 -7.38188803e-01 -2.40591034e-01 -8.43343735e-01
-1.77623972e-01 -3.58727038e-01 3.84871721e-01 1.58087897e+00
2.87320435e-01 -4.76748347e-01 8.25835586e-01 7.43290126e-01
-1.19250432e-01 -8.92314762e-02 -8.71374130e-01 -9.31955695e-01
-9.61364508e-02 -2.50611395e-01 1.52529269e-01 8.09295952e-01
-4.43441570e-01 -6.75016344e-02 -7.77183771e-01 3.63245755e-01
5.32050788e-01 -2.76909679e-01 8.68165612e-01 -9.21485782e-01
-1.09952437e-02 -5.43669224e-01 -9.34253931e-01 -7.15227962e-01
-5.90040088e-02 -5.60582876e-01 4.11201060e-01 -1.27634943e+00
7.91236311e-02 -1.44198105e-01 -1.61424596e-02 1.96553424e-01
5.40491492e-02 8.75060618e-01 6.46370232e-01 5.27476132e-01
-6.92207158e-01 2.08764076e-01 7.98080683e-01 -3.01780075e-01
-1.97551847e-01 1.68136626e-01 -2.96036184e-01 1.00043261e+00
8.86433423e-01 -4.88265365e-01 6.76411018e-02 -9.67683867e-02
-1.81792319e-01 -2.71899760e-01 4.25790669e-03 -1.31286252e+00
5.13530552e-01 -1.39838263e-01 7.21905291e-01 -7.88637638e-01
3.62890482e-01 -7.18114376e-01 -2.75981039e-01 1.86527833e-01
1.20462505e-02 1.12892844e-01 3.05448651e-01 3.52001995e-01
-3.63673747e-01 -6.26137972e-01 1.15689933e+00 -1.84138656e-01
-1.17861807e+00 -3.58415321e-02 -8.53666306e-01 -7.36721307e-02
1.40928268e+00 -5.50033092e-01 -3.70434701e-01 -1.23623535e-01
-2.84762114e-01 -1.79591730e-01 4.97508585e-01 3.16282630e-01
6.18457556e-01 -1.06025517e+00 -8.99586499e-01 3.86444032e-01
-3.20114428e-03 -3.91315669e-01 3.78244184e-02 4.68106568e-01
-1.51644123e+00 5.44357717e-01 -6.62669182e-01 -4.16469425e-01
-1.48261344e+00 2.76662648e-01 3.58601213e-01 -2.24949181e-01
-3.93786401e-01 6.33350015e-01 -1.96040064e-01 -4.57191020e-01
2.69027799e-03 1.19762085e-01 -4.96566534e-01 2.33466756e-02
6.23955667e-01 6.94804788e-01 1.90672085e-01 -1.31966400e+00
-4.05946016e-01 8.82371724e-01 -6.89545795e-02 -2.03616723e-01
1.35790217e+00 4.00352925e-02 -2.03275353e-01 3.07790071e-01
9.27332520e-01 3.37786466e-01 -1.19682515e+00 5.61452657e-02
1.43246744e-02 -7.41478741e-01 -2.40298674e-01 -5.78915000e-01
-1.02597213e+00 8.43485296e-01 6.25108600e-01 -8.15502182e-03
1.08254659e+00 -5.57644963e-01 4.80645597e-01 7.30332911e-01
4.89918478e-02 -1.49230456e+00 -5.12414239e-02 5.99831998e-01
6.78635001e-01 -1.23323619e+00 7.45889638e-03 -5.56257308e-01
-7.11340904e-01 1.68961930e+00 6.80870175e-01 -4.44111407e-01
3.08445454e-01 3.71026903e-01 3.87784302e-01 5.19664943e-01
4.13552113e-02 -3.31102638e-03 2.54552752e-01 5.96172094e-01
2.48219028e-01 -2.40879789e-01 -1.34053290e-01 4.85218704e-01
1.63137704e-01 -1.22905530e-01 1.17257333e+00 7.79790282e-01
-5.51149666e-01 -9.29065883e-01 -1.11771142e+00 3.32895756e-01
-6.50406361e-01 2.83449199e-02 -3.46819013e-01 1.23364878e+00
7.10916370e-02 6.80151582e-01 -1.74696445e-02 -1.02391496e-01
2.66510755e-01 1.80796459e-01 1.37999237e-01 -3.55396509e-01
-3.12865257e-01 3.97477865e-01 -2.28074357e-01 1.51925534e-02
-4.92688984e-01 -8.42876077e-01 -9.19020534e-01 -1.76059499e-01
-5.37375629e-01 -2.83234358e-01 9.03405786e-01 8.11336517e-01
-1.30038619e-01 1.22887209e-01 5.02151310e-01 -8.31902206e-01
-4.31205809e-01 -9.30544853e-01 -9.03228641e-01 3.25986564e-01
-4.01778035e-02 -3.42293173e-01 -3.31549138e-01 5.17678916e-01] | [9.836054801940918, -4.950724124908447] |
22ead3c9-324b-4fa9-9b17-b2a8367fa8bd | towards-scale-consistent-monocular-visual | 2203.05712 | null | https://arxiv.org/abs/2203.05712v1 | https://arxiv.org/pdf/2203.05712v1.pdf | Towards Scale Consistent Monocular Visual Odometry by Learning from the Virtual World | Monocular visual odometry (VO) has attracted extensive research attention by providing real-time vehicle motion from cost-effective camera images. However, state-of-the-art optimization-based monocular VO methods suffer from the scale inconsistency problem for long-term predictions. Deep learning has recently been introduced to address this issue by leveraging stereo sequences or ground-truth motions in the training dataset. However, it comes at an additional cost for data collection, and such training data may not be available in all datasets. In this work, we propose VRVO, a novel framework for retrieving the absolute scale from virtual data that can be easily obtained from modern simulation environments, whereas in the real domain no stereo or ground-truth data are required in either the training or inference phases. Specifically, we first train a scale-aware disparity network using both monocular real images and stereo virtual data. The virtual-to-real domain gap is bridged by using an adversarial training strategy to map images from both domains into a shared feature space. The resulting scale-consistent disparities are then integrated with a direct VO system by constructing a virtual stereo objective that ensures the scale consistency over long trajectories. Additionally, to address the suboptimality issue caused by the separate optimization backend and the learning process, we further propose a mutual reinforcement pipeline that allows bidirectional information flow between learning and optimization, which boosts the robustness and accuracy of each other. We demonstrate the effectiveness of our framework on the KITTI and vKITTI2 datasets. | ['DaCheng Tao', 'Jing Zhang', 'Sen Zhang'] | 2022-03-11 | null | null | null | null | ['monocular-visual-odometry'] | ['robots'] | [-1.65331796e-01 -1.34338573e-01 -1.76094174e-01 -3.31314117e-01
-6.09628737e-01 -5.97128272e-01 5.76050103e-01 -5.38668990e-01
-4.64919090e-01 8.27585578e-01 -2.58596212e-01 -2.24068984e-01
2.83890784e-01 -8.02601099e-01 -1.15550852e+00 -7.27219105e-01
3.37125748e-01 3.68877590e-01 3.93130749e-01 -2.38374561e-01
2.66916215e-01 4.86923665e-01 -1.71117246e+00 -3.10097039e-01
1.11007237e+00 1.11427522e+00 5.12553513e-01 5.38905859e-01
6.21749312e-02 6.96916521e-01 -2.97627181e-01 -1.16372392e-01
7.53500819e-01 -2.02649102e-01 -3.74601632e-01 8.19758698e-02
8.16939890e-01 -7.55788863e-01 -8.23604167e-01 1.22024608e+00
5.20895123e-01 4.23749685e-02 2.03152761e-01 -1.49886537e+00
-3.16599667e-01 -3.78448844e-01 -4.69275892e-01 -1.64561376e-01
3.41872603e-01 6.99907124e-01 7.27662861e-01 -7.54034996e-01
1.01556873e+00 1.04171205e+00 4.92944181e-01 5.53087592e-01
-1.01736236e+00 -8.14910769e-01 1.63732037e-01 3.91271532e-01
-1.34489059e+00 -3.85740548e-01 9.49439585e-01 -5.63267469e-01
8.19436550e-01 -1.43027201e-01 1.01859212e+00 9.35353577e-01
2.91583717e-01 6.64974451e-01 9.24914777e-01 7.59839639e-02
1.64785177e-01 3.22670974e-02 -5.37722290e-01 6.93729460e-01
2.46100157e-01 7.36645281e-01 -5.49056470e-01 3.24217141e-01
1.09202921e+00 1.34681417e-02 -4.49136198e-01 -1.01596355e+00
-1.03643250e+00 7.41685152e-01 8.36161852e-01 -1.83072031e-01
-2.63331592e-01 2.34800443e-01 1.07088462e-01 4.04860109e-01
2.50433266e-01 3.43555093e-01 -3.12511355e-01 -2.72070849e-03
-7.72809148e-01 4.64520603e-01 4.30487603e-01 1.00055575e+00
1.24238765e+00 3.77227336e-01 2.61194050e-01 4.23764378e-01
3.43220145e-01 8.46493006e-01 4.47149366e-01 -1.27153802e+00
7.67059624e-01 5.03627360e-01 3.16457868e-01 -1.15647316e+00
-2.74181992e-01 -2.66752154e-01 -7.24802732e-01 6.58029258e-01
4.48644519e-01 -9.90647897e-02 -8.91804695e-01 2.04785466e+00
6.58056915e-01 4.32095200e-01 1.08274028e-01 1.37474155e+00
5.69351256e-01 5.36940098e-01 -4.07682538e-01 6.22322895e-02
6.68417156e-01 -8.41003895e-01 -5.97661018e-01 -5.70790350e-01
4.82205898e-01 -5.83609283e-01 7.61908948e-01 1.74327253e-03
-8.65571856e-01 -6.57810271e-01 -1.45822811e+00 -2.64030755e-01
-2.48343289e-01 -2.65342295e-01 3.34331334e-01 3.28400791e-01
-1.05847371e+00 4.92737025e-01 -8.22775126e-01 -1.16267521e-02
1.59175813e-01 3.70761573e-01 -4.66320395e-01 -2.28423879e-01
-1.33216536e+00 9.73610699e-01 3.40606906e-02 2.01991081e-01
-9.40182567e-01 -7.97007382e-01 -1.29699361e+00 -3.83262515e-01
4.03127313e-01 -9.56928074e-01 9.22909200e-01 -1.09789503e+00
-1.77336884e+00 8.21395993e-01 -2.15002969e-01 -3.83478791e-01
9.26396012e-01 -2.83342391e-01 -1.10589527e-01 7.91423470e-02
1.16576426e-01 9.98428702e-01 8.33858728e-01 -1.50896847e+00
-6.83442891e-01 -4.89829153e-01 2.87216544e-01 5.35900295e-01
2.87313879e-01 -7.49127746e-01 -6.38254941e-01 -3.75199586e-01
3.77651125e-01 -1.11613321e+00 -2.32612610e-01 4.52279806e-01
-2.20812440e-01 4.75722820e-01 7.29979932e-01 -7.09043443e-01
6.89155757e-01 -1.97134233e+00 3.74280095e-01 -8.53771344e-03
1.28765613e-01 2.03346789e-01 -1.31110370e-01 -1.17803290e-01
2.00089902e-01 -4.24851120e-01 -2.50927620e-02 -4.26691681e-01
-8.79039690e-02 4.29346949e-01 -4.26106989e-01 7.46743858e-01
1.58468381e-01 1.06561053e+00 -9.67356622e-01 -3.78324658e-01
5.99442482e-01 5.50301313e-01 -7.59903610e-01 4.29354310e-01
-1.56482652e-01 9.75300670e-01 -2.16961890e-01 4.28123146e-01
9.43751991e-01 -3.65872569e-02 -2.23798845e-02 -1.33520380e-01
-3.32306445e-01 3.83599877e-01 -1.27180159e+00 1.96846700e+00
-5.32213926e-01 7.61442780e-01 8.53907689e-02 -8.49559128e-01
8.27752411e-01 -1.02445576e-03 4.83024508e-01 -1.19401610e+00
7.14700148e-02 4.11850929e-01 -2.50126362e-01 -2.76761770e-01
6.95867896e-01 -1.98297873e-01 1.14013284e-01 -6.83728606e-02
-2.53885269e-01 -7.63756514e-01 -1.51273698e-01 -1.67446896e-01
5.97624958e-01 4.80964750e-01 6.38615713e-02 1.71755880e-01
6.55875087e-01 1.04544006e-01 1.01221240e+00 3.32486302e-01
-3.69506538e-01 8.55972767e-01 9.33071226e-02 -4.60814238e-01
-1.34745812e+00 -1.09615862e+00 -2.45324913e-02 1.87628403e-01
8.90295982e-01 1.77791089e-01 -4.37839597e-01 -5.20193875e-01
2.70316303e-01 3.71286601e-01 -5.13271570e-01 -3.09093326e-01
-7.52073109e-01 -1.70206845e-01 3.09150517e-01 3.87814373e-01
8.17839801e-01 -6.53978288e-01 -7.73388684e-01 2.34311461e-01
-2.53738046e-01 -1.45606804e+00 -6.28496230e-01 -2.34965786e-01
-8.51890683e-01 -1.05778313e+00 -6.50248230e-01 -7.08333373e-01
4.90668982e-01 6.03310108e-01 8.96206319e-01 -3.38011459e-02
6.74179941e-02 -7.19953105e-02 1.04690067e-01 -3.09947710e-02
-4.49273795e-01 -2.38268986e-01 2.18525201e-01 1.71189420e-02
-1.54297277e-01 -5.85891843e-01 -8.51194501e-01 6.06271982e-01
-7.04659522e-01 4.72830892e-01 2.69459873e-01 9.56202269e-01
6.10017061e-01 -3.81199926e-01 3.36983681e-01 -2.95676172e-01
-2.85313338e-01 -2.76809454e-01 -1.36394501e+00 -2.43682206e-01
-5.14869094e-01 1.16775602e-01 6.58327222e-01 -3.16613674e-01
-9.69929755e-01 1.89917818e-01 -2.04562321e-01 -9.32270408e-01
2.49824300e-01 2.25308985e-01 -3.62477332e-01 -4.01734918e-01
4.08794075e-01 3.06110919e-01 2.41779909e-01 -6.02494739e-02
1.66102752e-01 3.77687603e-01 7.64528453e-01 -1.54716149e-01
1.27258062e+00 9.35035408e-01 1.26996011e-01 -6.09362721e-01
-6.41503513e-01 -3.94225717e-01 -5.94697773e-01 -2.44715601e-01
8.85160446e-01 -1.35515273e+00 -5.92793822e-01 7.73450971e-01
-1.17216992e+00 -4.62897241e-01 -2.03294769e-01 6.12988889e-01
-8.26679707e-01 5.12886822e-01 -3.89778763e-01 -4.38565075e-01
-6.72875792e-02 -1.40902781e+00 1.08144569e+00 2.07157820e-01
2.68267661e-01 -8.70766938e-01 2.17572525e-01 3.58863652e-01
2.11708039e-01 2.17018858e-01 4.64580148e-01 2.52807409e-01
-1.21214473e+00 -1.06514655e-01 -3.04745555e-01 2.80660748e-01
3.84137523e-03 -3.01642060e-01 -9.30624068e-01 -5.17425358e-01
-4.53710034e-02 -4.69832510e-01 7.22541451e-01 3.54254752e-01
6.57054186e-01 8.06110576e-02 -1.86343223e-01 1.26522136e+00
1.57607520e+00 1.89691395e-01 6.39826596e-01 5.85004389e-01
1.08641553e+00 5.65294385e-01 7.82864928e-01 1.66755155e-01
7.90736139e-01 1.08931887e+00 7.56737351e-01 -1.28044888e-01
-2.10978195e-01 -5.67453563e-01 3.58621776e-01 7.25315571e-01
1.97302982e-01 1.04702383e-01 -6.79581165e-01 5.99886119e-01
-1.83241248e+00 -8.90706658e-01 2.53118604e-01 2.50774884e+00
6.94753408e-01 1.74086422e-01 -3.10402304e-01 -5.04930578e-02
4.87902045e-01 3.40267539e-01 -9.33716416e-01 4.76507051e-03
-2.35236093e-01 -3.79039973e-01 7.63476729e-01 7.98507571e-01
-9.36861217e-01 1.06224990e+00 5.15513086e+00 4.09296781e-01
-1.76885211e+00 -9.05925501e-03 3.07952136e-01 -9.29961130e-02
-4.64917392e-01 1.71454102e-01 -7.32618928e-01 4.00328696e-01
5.39743602e-01 -1.59629464e-01 5.77011287e-01 7.63762474e-01
2.99557686e-01 -1.24359533e-01 -1.03413391e+00 1.22695255e+00
-4.50961925e-02 -1.49879777e+00 -1.33421570e-01 1.94685563e-01
1.12771153e+00 4.13314581e-01 1.08338028e-01 1.97043031e-01
3.33515525e-01 -7.83020496e-01 9.77266729e-01 2.91978747e-01
9.79237556e-01 -6.43293083e-01 6.62405610e-01 5.66839755e-01
-1.35207510e+00 -5.77069968e-02 -3.25648636e-01 -9.46470574e-02
3.34000647e-01 3.56656641e-01 -5.99649370e-01 9.05967832e-01
4.78776872e-01 9.34994280e-01 -2.72478729e-01 1.03752041e+00
-2.92255998e-01 -1.63893431e-01 -3.67055625e-01 3.83483797e-01
2.70567805e-01 -4.06156242e-01 6.87139690e-01 4.54961360e-01
3.98256928e-01 -2.10466474e-01 2.29013681e-01 9.49106693e-01
7.74006313e-03 -3.26500893e-01 -8.85886252e-01 5.16297400e-01
4.91189867e-01 7.91363537e-01 -8.01310167e-02 -2.96554267e-01
-6.27560079e-01 8.83904219e-01 4.49561954e-01 5.01413822e-01
-9.03036356e-01 -4.53233235e-02 1.17737794e+00 -9.61904321e-03
3.98647219e-01 -4.11616296e-01 -3.96380663e-01 -1.52585244e+00
2.52338290e-01 -6.64562464e-01 -1.69482127e-01 -7.57802427e-01
-8.88966143e-01 4.77201879e-01 -2.53215492e-01 -1.70994949e+00
-6.43180788e-01 -3.68172407e-01 -2.57010341e-01 9.69698429e-01
-2.09954691e+00 -9.73757267e-01 -6.69005156e-01 5.07960379e-01
3.82895797e-01 -8.22917733e-04 2.07027704e-01 4.34635341e-01
-3.18578184e-01 4.33877468e-01 1.81568041e-01 -8.28989744e-02
8.25415611e-01 -1.05061579e+00 6.41507089e-01 8.30469251e-01
-1.50497109e-01 2.02070653e-01 7.60180712e-01 -4.83163446e-01
-1.64723146e+00 -1.18438351e+00 6.48637116e-01 -4.33034867e-01
3.84671777e-01 -1.49623916e-01 -9.77761805e-01 6.07299566e-01
-2.43715331e-01 3.64483863e-01 -3.09431493e-01 -5.82419574e-01
-3.21101993e-01 -3.29950362e-01 -1.03657341e+00 7.63656974e-01
1.18078077e+00 -7.09116459e-01 -3.44311386e-01 -1.11324973e-01
8.50832403e-01 -9.73160446e-01 -4.50255632e-01 4.99779284e-01
6.44721448e-01 -1.11420393e+00 1.15400815e+00 -1.10827312e-01
4.36620831e-01 -5.72907984e-01 -2.74936378e-01 -1.46732652e+00
1.44463912e-01 -5.09827077e-01 -2.59924438e-02 7.69461632e-01
8.16566721e-02 -8.69716942e-01 9.79753256e-01 3.57510239e-01
-2.34850571e-01 -5.31993032e-01 -1.28450203e+00 -9.15364146e-01
2.29738101e-01 -5.68577707e-01 7.56339967e-01 8.51400793e-01
-4.06039387e-01 1.33637756e-01 -5.42888880e-01 4.39342469e-01
7.64244497e-01 3.85318249e-01 1.36554110e+00 -8.09289277e-01
-3.92904878e-01 -2.47474894e-01 -7.44724035e-01 -1.70047712e+00
3.22777212e-01 -6.34120226e-01 3.83598953e-01 -1.20512390e+00
-1.96776450e-01 -5.12551546e-01 1.46633759e-01 -8.52690563e-02
-2.25802958e-01 2.52494425e-01 3.72787327e-01 3.37555796e-01
-2.17292741e-01 9.59050238e-01 1.74456644e+00 -7.88437277e-02
-3.31176341e-01 -1.58699989e-01 -2.63491459e-02 6.64462149e-01
5.34093738e-01 -1.49569258e-01 -5.76339602e-01 -7.78994977e-01
1.13722853e-01 4.87538725e-01 5.87072670e-01 -1.08915865e+00
4.39868778e-01 -3.30827326e-01 2.51598507e-01 -6.77675009e-01
6.39550805e-01 -8.50017786e-01 2.00064719e-01 4.37058389e-01
8.35192204e-02 -1.32235400e-02 1.09416500e-01 6.16083205e-01
-3.97952378e-01 2.52876610e-01 1.03498709e+00 1.70374662e-01
-9.96613145e-01 7.02730119e-01 1.58431649e-01 2.01718509e-01
9.83191907e-01 -4.13617134e-01 -1.65224314e-01 -5.16934395e-01
-1.63925201e-01 5.22543073e-01 1.09905100e+00 6.04133368e-01
7.32122719e-01 -1.42568851e+00 -3.57616842e-01 6.36436284e-01
2.96915382e-01 5.41288376e-01 3.47004414e-01 7.44235158e-01
-7.60736048e-01 4.24272776e-01 -4.08391744e-01 -8.90591979e-01
-8.56641233e-01 5.39020479e-01 6.77133381e-01 -1.98682278e-01
-6.57505751e-01 4.40305501e-01 5.65051496e-01 -8.21578801e-01
7.61403218e-02 -2.78563380e-01 1.24727242e-01 -3.85926753e-01
3.02662730e-01 2.04467401e-01 -3.08037587e-02 -8.72612000e-01
-2.04161167e-01 1.02153575e+00 2.62298197e-01 -3.51009995e-01
9.76779819e-01 -5.68385303e-01 3.86036664e-01 3.21322232e-01
1.45228457e+00 -1.76595047e-01 -2.09550452e+00 -3.10269564e-01
-3.33632410e-01 -8.33375931e-01 1.55250430e-01 -1.70279652e-01
-1.29901564e+00 8.80869210e-01 6.78648174e-01 -4.86460388e-01
7.89109051e-01 -4.15333986e-01 1.11369205e+00 2.34733552e-01
5.82862556e-01 -1.04962873e+00 4.77489457e-02 6.51730895e-01
7.83577561e-01 -1.77488756e+00 -9.74290818e-02 -4.53724951e-01
-4.76799995e-01 8.87052715e-01 8.30139041e-01 -2.45135203e-01
4.08171833e-01 -3.91234830e-02 3.50108951e-01 9.11535323e-02
-6.46025777e-01 -1.57987282e-01 2.60979086e-01 6.26197338e-01
-2.47104689e-01 -1.26962811e-01 4.70934361e-02 -3.22797499e-03
-2.03413069e-01 2.84472536e-02 3.56932253e-01 9.57803369e-01
-3.61805767e-01 -9.40152705e-01 -3.49196136e-01 -2.09866434e-01
1.21583305e-01 1.39496192e-01 7.09527433e-02 9.20182049e-01
8.42783824e-02 6.96427524e-01 2.26489827e-01 -4.84162092e-01
3.84756505e-01 -3.44057471e-01 5.34777701e-01 -1.91006303e-01
-5.90091906e-02 -2.45389193e-01 -8.95421505e-02 -9.78589356e-01
-1.35985911e-01 -5.87838829e-01 -1.34869444e+00 -5.49723327e-01
-7.25989342e-02 -2.38545850e-01 7.43504167e-01 1.01603019e+00
3.42534006e-01 2.50867963e-01 9.21475589e-01 -1.34322178e+00
-5.87843895e-01 -3.96123081e-01 -2.60133147e-01 5.20632386e-01
8.97653222e-01 -8.46371531e-01 -4.79842991e-01 -1.89867169e-01] | [8.255399703979492, -2.206068754196167] |
9b6df952-6892-4767-a3ac-35874e8f930f | indicnlg-suite-multilingual-datasets-for | 2203.05437 | null | https://arxiv.org/abs/2203.05437v2 | https://arxiv.org/pdf/2203.05437v2.pdf | IndicNLG Benchmark: Multilingual Datasets for Diverse NLG Tasks in Indic Languages | Natural Language Generation (NLG) for non-English languages is hampered by the scarcity of datasets in these languages. In this paper, we present the IndicNLG Benchmark, a collection of datasets for benchmarking NLG for 11 Indic languages. We focus on five diverse tasks, namely, biography generation using Wikipedia infoboxes, news headline generation, sentence summarization, paraphrase generation and, question generation. We describe the created datasets and use them to benchmark the performance of several monolingual and multilingual baselines that leverage pre-trained sequence-to-sequence models. Our results exhibit the strong performance of multilingual language-specific pre-trained models, and the utility of models trained on our dataset for other related NLG tasks. Our dataset creation methods can be easily applied to modest-resource languages as they involve simple steps such as scraping news articles and Wikipedia infoboxes, light cleaning, and pivoting through machine translation data. To the best of our knowledge, the IndicNLG Benchmark is the first NLG benchmark for Indic languages and the most diverse multilingual NLG dataset, with approximately 8M examples across 5 tasks and 11 languages. The datasets and models are publicly available at https://ai4bharat.iitm.ac.in/indicnlg-suite. | ['Pratyush Kumar', 'Mitesh M. Khapra', 'Amogh Mishra', 'Anoop Kunchukuttan', 'Ratish Puduppully', 'Raj Dabre', 'Prachi Sahu', 'Himani Shrotriya', 'Aman Kumar'] | 2022-03-10 | null | null | null | null | ['paraphrase-generation', 'headline-generation', 'paraphrase-generation', 'abstractive-sentence-summarization'] | ['computer-code', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 1.99487522e-01 1.52426079e-01 -2.75747031e-01 5.51189817e-02
-1.42591453e+00 -9.38426197e-01 1.14132082e+00 6.60444871e-02
-3.58004570e-01 1.41837955e+00 9.53596652e-01 -5.43785751e-01
2.84175634e-01 -6.80286705e-01 -8.69749308e-01 -1.39787659e-01
2.11637512e-01 6.48663521e-01 -2.89865345e-01 -7.43220091e-01
5.32075465e-01 -7.29963928e-02 -1.09031796e+00 6.09618545e-01
1.52244961e+00 4.21570778e-01 2.59399861e-01 1.01172030e+00
-3.63817006e-01 9.85384822e-01 -1.01303530e+00 -6.54393017e-01
1.37947574e-01 -7.20109284e-01 -1.11899149e+00 -3.60874414e-01
7.63479292e-01 -1.89166769e-01 -2.75700957e-01 6.26477003e-01
9.72465396e-01 -1.60375416e-01 6.93698943e-01 -1.17136240e+00
-1.13571775e+00 1.28013504e+00 -1.90641180e-01 2.72750072e-02
9.25012648e-01 2.65873820e-01 1.02337325e+00 -8.94711554e-01
1.31899416e+00 1.23871148e+00 5.78990638e-01 6.71283901e-01
-8.75059545e-01 -3.22430909e-01 -3.88748318e-01 9.25718993e-02
-8.97908568e-01 -7.46641219e-01 4.15276676e-01 -4.17237699e-01
1.23666584e+00 2.57428646e-01 1.36853695e-01 1.48164690e+00
3.42527241e-01 9.39077556e-01 1.19833493e+00 -6.37988746e-01
-1.86866894e-01 -9.83647909e-03 4.50075902e-02 5.47466218e-01
1.54112875e-01 -2.65878022e-01 -7.39312947e-01 -1.28754720e-01
5.78345060e-02 -7.67026544e-01 -3.29844147e-01 5.39618969e-01
-1.74194074e+00 8.35096121e-01 8.62856954e-03 1.59703240e-01
-3.26028079e-01 -5.28264418e-03 6.90911531e-01 5.08463085e-01
6.20429873e-01 9.45986509e-01 -4.42364961e-01 -3.10925663e-01
-9.86559629e-01 6.03540421e-01 1.28833234e+00 1.43606877e+00
4.36136901e-01 1.46893010e-01 -6.91348493e-01 1.01545215e+00
-1.66804075e-01 7.16529071e-01 7.85413384e-01 -7.93960810e-01
1.22977507e+00 4.49985117e-01 -2.76813712e-02 -7.17124939e-01
-1.54277340e-01 3.72376107e-02 -8.45448077e-01 -4.48898077e-01
4.88340408e-01 -6.65537357e-01 -7.59614944e-01 1.50304770e+00
-1.27813388e-02 -3.58661950e-01 6.16044700e-01 3.09531718e-01
1.45697796e+00 1.17908788e+00 -1.69885725e-01 -1.31256521e-01
1.27317238e+00 -1.19288278e+00 -6.29189134e-01 -3.15409034e-01
8.02093744e-01 -1.45747650e+00 1.25330675e+00 1.29820973e-01
-1.19376314e+00 -4.29674894e-01 -8.43862593e-01 -6.13156796e-01
-6.56131685e-01 5.22699118e-01 5.52232325e-01 2.63130486e-01
-1.13994217e+00 4.11351472e-01 -2.73205429e-01 -6.67444885e-01
9.68598202e-02 -3.88009101e-01 -3.48605603e-01 -2.19442114e-01
-1.43299115e+00 1.06645656e+00 6.96647406e-01 -2.87819915e-02
-8.58741641e-01 -7.29302466e-01 -8.20409179e-01 -5.73908806e-01
2.94852287e-01 -7.02218592e-01 1.38224542e+00 -4.02766496e-01
-1.54392278e+00 9.38267291e-01 -7.02190921e-02 -6.67093575e-01
6.31416440e-01 -4.06544685e-01 -4.57505345e-01 -1.76129699e-01
5.40499926e-01 7.62840033e-01 3.59990388e-01 -7.45646119e-01
-3.24058115e-01 1.05964676e-01 -4.37914282e-02 1.80723324e-01
9.65171903e-02 2.48567551e-01 -3.34638894e-01 -9.32788014e-01
-5.69088519e-01 -8.77123117e-01 -2.17914492e-01 -9.32965755e-01
-1.05014503e+00 -1.81508228e-01 2.59400338e-01 -1.42154336e+00
1.48890841e+00 -1.38032007e+00 5.00024185e-02 -4.27382588e-01
-3.24452639e-01 2.53535479e-01 -5.81526101e-01 1.25802493e+00
3.54987741e-01 3.51523548e-01 -2.69124836e-01 -1.13006964e-01
1.20249964e-01 -1.25104249e-01 -6.76682413e-01 -7.84157887e-02
2.53556281e-01 1.31558299e+00 -1.00230825e+00 -4.87359673e-01
-2.73792744e-01 1.34883404e-01 -4.26778942e-01 2.02026248e-01
-5.76818943e-01 4.11117852e-01 -1.68654397e-01 6.26834750e-01
3.10301661e-01 7.90786967e-02 -1.84362456e-02 -1.14928996e-02
-3.60871494e-01 9.24255669e-01 -6.76858068e-01 1.97550142e+00
-6.77825153e-01 8.11204255e-01 -4.16934013e-01 -4.86056149e-01
1.07660413e+00 4.19753402e-01 -2.19336301e-01 -6.93813384e-01
-1.69519469e-01 6.46228611e-01 -1.42416641e-01 -5.62602520e-01
1.11368239e+00 4.61232364e-01 -7.50852883e-01 7.98544466e-01
4.06925708e-01 -5.51865399e-01 1.12311387e+00 6.73669100e-01
8.24561834e-01 2.62871027e-01 6.26907408e-01 -2.61307776e-01
5.01410902e-01 3.90623927e-01 3.28950346e-01 1.05883467e+00
3.60848904e-01 6.24108613e-01 5.21721125e-01 -1.89211637e-01
-1.37418830e+00 -8.57537389e-01 1.79070324e-01 1.11412871e+00
-4.46601033e-01 -6.63315177e-01 -8.22850347e-01 -5.82550824e-01
-1.38572529e-01 1.14974153e+00 -3.19174230e-01 2.11387798e-01
-8.75791967e-01 -9.96160030e-01 1.17151511e+00 1.29016250e-01
4.75551844e-01 -1.56941772e+00 2.20558159e-02 4.19736892e-01
-6.35135651e-01 -1.21325982e+00 -7.15161383e-01 -4.19645160e-01
-5.79349041e-01 -9.68232810e-01 -8.14794898e-01 -8.66710484e-01
3.10387015e-01 -2.18237624e-01 1.61868966e+00 -3.07732582e-01
-3.02039087e-01 1.26691282e-01 -5.15874505e-01 -5.70630133e-01
-1.11052155e+00 7.50017464e-01 -1.92400545e-01 -5.27878463e-01
-2.22580582e-02 -2.23097727e-01 -2.23541409e-01 -2.69892693e-01
-7.37446487e-01 5.00057518e-01 6.87628746e-01 7.37703860e-01
3.59797984e-01 -9.31303322e-01 1.10174382e+00 -1.25421214e+00
1.18144298e+00 -6.15166366e-01 -5.00620246e-01 6.94662631e-01
-3.73859048e-01 1.30978987e-01 8.71043324e-01 -1.09117210e-01
-1.24589765e+00 -4.31311369e-01 -1.52889192e-01 7.11513877e-01
2.13678479e-02 8.81956279e-01 -1.01417027e-01 5.22079885e-01
9.04550374e-01 5.50881088e-01 -2.46199310e-01 -5.74701309e-01
1.01202440e+00 9.45788145e-01 7.50295043e-01 -6.56833887e-01
7.05868602e-01 -1.10979699e-01 -3.62729877e-01 -7.66863227e-01
-8.68836820e-01 -3.50479819e-02 -5.09689271e-01 -6.12639226e-02
6.40673876e-01 -1.06196606e+00 -2.22018622e-02 5.97867489e-01
-1.51496780e+00 -5.00646830e-01 -1.81722879e-01 1.83080092e-01
-5.04179418e-01 2.91496515e-01 -9.08793211e-01 -3.23990732e-01
-1.23159409e+00 -8.79669547e-01 9.72081065e-01 1.21676430e-01
-4.79352355e-01 -1.05770993e+00 4.21635896e-01 6.79789066e-01
4.72460747e-01 4.91901726e-01 7.58230627e-01 -8.27124953e-01
-3.94393265e-01 -8.61528888e-02 5.27421944e-04 3.00326139e-01
6.69539645e-02 1.24126151e-01 -5.00076711e-01 -9.68359485e-02
-5.77190697e-01 -8.43278408e-01 8.55941176e-01 1.00041308e-01
6.62283838e-01 -8.91663194e-01 7.61047974e-02 5.13726413e-01
1.08831024e+00 -8.35493207e-02 6.70993507e-01 5.98348200e-01
7.63962686e-01 5.74500322e-01 5.99115372e-01 2.66418695e-01
7.71818876e-01 3.72176528e-01 -2.88943619e-01 1.08900204e-01
-5.44434845e-01 -6.50369585e-01 7.12247372e-01 1.55627310e+00
8.58839825e-02 -7.26412535e-01 -1.08826780e+00 8.07886958e-01
-1.74476027e+00 -9.56262827e-01 -4.17429835e-01 1.81542242e+00
1.54944527e+00 -2.62217641e-01 -1.23325430e-01 -4.64939773e-01
5.53871989e-01 3.96772653e-01 -2.56966650e-01 -7.55498767e-01
-6.68314695e-01 2.04166979e-01 3.67479861e-01 7.43731678e-01
-9.22100663e-01 1.39243054e+00 5.93068123e+00 1.07705295e+00
-8.49904180e-01 1.49850950e-01 6.32471204e-01 -1.62369281e-01
-4.28769112e-01 1.12242647e-01 -1.25343776e+00 5.74543059e-01
1.10718453e+00 -8.12629461e-01 5.52914023e-01 7.00752854e-01
3.34738910e-01 1.42184332e-01 -9.82967079e-01 5.72202206e-01
4.57446903e-01 -1.75688183e+00 4.19891417e-01 -3.36209327e-01
1.23762298e+00 3.64185452e-01 -2.44829297e-01 5.59761405e-01
7.13260472e-01 -1.12374735e+00 7.14249253e-01 5.52943349e-01
9.08765674e-01 -5.23212075e-01 6.03987575e-01 3.73343408e-01
-6.78010821e-01 3.25565904e-01 -4.24559563e-01 5.59494682e-02
6.35899723e-01 6.81699455e-01 -1.18997765e+00 6.89914525e-01
2.35329330e-01 7.51482069e-01 -8.09937775e-01 7.62129605e-01
-6.40346825e-01 7.40299165e-01 -3.70273963e-02 -2.88763940e-01
3.13312560e-01 -1.99700624e-01 6.58747137e-01 1.66470110e+00
6.61247611e-01 -4.53217268e-01 1.31731167e-01 6.53274179e-01
-5.06305873e-01 5.59839547e-01 -7.98154652e-01 -4.29649383e-01
5.54332078e-01 1.37918293e+00 -1.18908376e-01 -6.01371467e-01
-2.35960379e-01 9.48267817e-01 3.71551812e-01 3.50007266e-01
-6.26199841e-01 -8.54849279e-01 4.94811982e-02 -4.34221365e-02
-1.91796780e-01 -1.69893771e-01 -2.08831966e-01 -1.49107277e+00
-7.94739723e-02 -1.48085749e+00 4.84299660e-01 -8.19348693e-01
-1.35035825e+00 8.65890384e-01 1.83281060e-02 -1.03332651e+00
-9.23765779e-01 -2.91994184e-01 -7.68782437e-01 9.38554704e-01
-1.42974579e+00 -1.43027854e+00 -1.91247955e-01 2.22472057e-01
8.22325289e-01 -6.04087353e-01 6.62846267e-01 2.24701375e-01
-5.37976325e-01 6.44332647e-01 3.52905899e-01 3.01145971e-01
1.13867617e+00 -1.32007778e+00 1.15864623e+00 1.09025729e+00
1.44197941e-01 6.54277027e-01 5.87384343e-01 -9.24297690e-01
-1.33308685e+00 -1.41785920e+00 1.76721251e+00 -6.27128363e-01
9.24324512e-01 -6.88357890e-01 -5.19143641e-01 8.00156891e-01
9.01324987e-01 -7.63283670e-01 5.76538265e-01 -2.07201719e-01
-1.50408641e-01 1.59037307e-01 -7.08039343e-01 9.87986445e-01
1.00857770e+00 -3.72896165e-01 -6.56211972e-01 8.75347853e-01
8.86016488e-01 -6.14782095e-01 -9.81100142e-01 1.80319175e-01
3.27622950e-01 -4.54208761e-01 7.50044405e-01 -7.14581072e-01
1.13279903e+00 -1.68661118e-01 -6.96605956e-03 -1.65845358e+00
1.47812113e-01 -1.03997731e+00 2.64641922e-02 1.81007898e+00
1.00700390e+00 -7.14537859e-01 2.33273819e-01 -2.87906658e-02
-3.49323571e-01 -4.45389867e-01 -6.49207294e-01 -6.58258498e-01
3.69963169e-01 -1.67323276e-01 6.71572566e-01 9.05633152e-01
4.53670286e-02 8.70359242e-01 -6.77731395e-01 -4.76892471e-01
4.15466428e-01 3.34196568e-01 1.21499610e+00 -6.06042862e-01
-3.13888043e-01 -3.41942936e-01 3.74767691e-01 -8.03578496e-01
2.13228092e-01 -1.37272191e+00 1.00243531e-01 -1.91802275e+00
2.04281762e-01 -1.66696385e-01 6.21736526e-01 5.36590457e-01
-5.27890921e-01 2.25795344e-01 3.02690655e-01 2.08891615e-01
-5.84137082e-01 5.11079729e-01 1.09148502e+00 -6.88357577e-02
-2.77779907e-01 -3.01940739e-01 -9.34382200e-01 3.76495659e-01
1.15089917e+00 -3.73559684e-01 -2.52187818e-01 -7.66921997e-01
4.04316187e-01 -1.48449838e-02 -4.32251394e-02 -6.54934466e-01
1.15718946e-01 -1.39725104e-01 7.48403445e-02 -6.53872013e-01
-1.57267034e-01 3.99186373e-01 5.07385656e-02 3.17832261e-01
-5.70181072e-01 4.01631325e-01 2.64753789e-01 -1.30765602e-01
-2.53267795e-01 -3.60894948e-01 2.14095592e-01 -5.13840437e-01
-6.76474035e-01 1.08195260e-01 -4.41084653e-01 9.14762497e-01
5.71389794e-01 1.64998263e-01 -1.04252744e+00 -5.46814859e-01
-4.97321300e-02 3.06545645e-01 3.12348783e-01 7.29549408e-01
2.85566598e-01 -1.32702136e+00 -1.58091605e+00 -2.98793912e-01
1.83428124e-01 -2.22413927e-01 -3.33187473e-03 6.51868343e-01
-9.70465839e-01 8.83178830e-01 -2.41556644e-01 -4.68544699e-02
-9.37820733e-01 1.52762830e-01 -1.31958127e-01 -7.59597659e-01
-1.93580747e-01 4.80068386e-01 -2.43258864e-01 -1.03543699e+00
-2.37522066e-01 -8.33237618e-02 -2.88409054e-01 1.61292002e-01
6.89951897e-01 6.56909883e-01 5.99713763e-03 -6.25042856e-01
-8.70256275e-02 1.76904440e-01 -2.46693134e-01 -3.79601717e-01
1.26806140e+00 6.19263528e-03 -4.78245586e-01 2.99840748e-01
1.07532001e+00 3.46115977e-01 -4.86613274e-01 -6.03241771e-02
2.57819325e-01 4.16244902e-02 -4.96653587e-01 -1.16199481e+00
-4.61827040e-01 6.09314263e-01 -2.51901746e-01 -1.03786081e-01
8.72997463e-01 -4.85990644e-02 1.07025254e+00 5.04871130e-01
1.80070758e-01 -1.20027566e+00 -2.29609641e-03 1.08985257e+00
1.47763681e+00 -1.11392462e+00 3.41772251e-02 -9.78117585e-02
-9.73108113e-01 8.93006742e-01 5.59766769e-01 6.08412400e-02
-1.18581327e-02 1.11072540e-01 3.02761227e-01 1.84716463e-01
-9.66597378e-01 2.69322153e-02 5.29991746e-01 4.93954331e-01
9.02118027e-01 1.06506109e-01 -8.46767664e-01 5.00656784e-01
-9.98248875e-01 7.02995956e-02 9.16902483e-01 7.61803329e-01
-7.74480775e-02 -1.32673943e+00 -2.51694292e-01 5.54959238e-01
-7.47564852e-01 -8.49601924e-01 -7.52488911e-01 8.43196929e-01
-1.60802156e-01 1.01887381e+00 -2.72229165e-01 8.27391632e-04
1.89882204e-01 1.57339692e-01 2.74795473e-01 -7.99987316e-01
-7.65139222e-01 -2.68366665e-01 7.91810930e-01 -2.53140450e-01
-7.44790211e-02 -6.63226843e-01 -8.30569565e-01 -5.00198364e-01
8.97630230e-02 1.93286091e-01 5.93268573e-01 5.86075962e-01
7.23271072e-01 1.94266781e-01 2.76728064e-01 -6.64168179e-01
-4.86096382e-01 -1.52280962e+00 6.43148199e-02 3.63309085e-01
-1.04775846e-01 3.05522054e-01 -2.05168203e-01 2.84858465e-01] | [11.802274703979492, 9.476014137268066] |
728d8f18-8aa8-4d32-b81f-e6e11919e50d | holistic-sentence-embeddings-for-better-out | 2210.07485 | null | https://arxiv.org/abs/2210.07485v1 | https://arxiv.org/pdf/2210.07485v1.pdf | Holistic Sentence Embeddings for Better Out-of-Distribution Detection | Detecting out-of-distribution (OOD) instances is significant for the safe deployment of NLP models. Among recent textual OOD detection works based on pretrained language models (PLMs), distance-based methods have shown superior performance. However, they estimate sample distance scores in the last-layer CLS embedding space and thus do not make full use of linguistic information underlying in PLMs. To address the issue, we propose to boost OOD detection by deriving more holistic sentence embeddings. On the basis of the observations that token averaging and layer combination contribute to improving OOD detection, we propose a simple embedding approach named Avg-Avg, which averages all token representations from each intermediate layer as the sentence embedding and significantly surpasses the state-of-the-art on a comprehensive suite of benchmarks by a 9.33% FAR95 margin. Furthermore, our analysis demonstrates that it indeed helps preserve general linguistic knowledge in fine-tuned PLMs and substantially benefits detecting background shifts. The simple yet effective embedding method can be applied to fine-tuned PLMs with negligible extra costs, providing a free gain in OOD detection. Our code is available at https://github.com/lancopku/Avg-Avg. | ['Xu sun', 'Rundong Gao', 'Xiaohan Bi', 'Sishuo Chen'] | 2022-10-14 | null | null | null | null | ['sentence-embeddings', 'sentence-embeddings'] | ['methodology', 'natural-language-processing'] | [-2.32799366e-01 6.48214743e-02 -3.29897463e-01 -8.66942704e-02
-1.22588265e+00 -7.37429798e-01 7.05340207e-01 8.69672298e-01
-4.12436306e-01 2.70730823e-01 4.52629924e-01 -4.32382882e-01
2.85011604e-02 -7.17557728e-01 -4.91540372e-01 -4.28459704e-01
-6.86731413e-02 1.36172399e-01 3.02890331e-01 -1.31687939e-01
3.09437931e-01 5.32331049e-01 -1.20789421e+00 3.19462270e-01
9.91121650e-01 7.62807846e-01 -7.73761496e-02 7.78835416e-01
-5.70511758e-01 4.79830116e-01 -8.99112046e-01 -3.90146881e-01
8.35281909e-02 -6.82155788e-02 -5.46098948e-01 -2.82759160e-01
8.28977883e-01 -3.50872874e-01 -5.95913708e-01 8.65132451e-01
8.77441645e-01 -3.03339548e-02 8.15474689e-01 -9.29710567e-01
-1.06762016e+00 5.17950654e-01 -5.40061235e-01 6.07750058e-01
3.26878697e-01 1.02710985e-01 1.49593174e+00 -1.20455801e+00
6.10143185e-01 1.24755788e+00 6.96128786e-01 2.41369456e-01
-1.47959888e+00 -2.92862743e-01 1.14128232e-01 1.17850723e-02
-1.38329315e+00 -4.21310782e-01 4.42313313e-01 -3.54825020e-01
1.55414522e+00 2.81369358e-01 1.86105773e-01 1.03136718e+00
2.09769338e-01 1.02112472e+00 7.19518781e-01 -8.28631639e-01
1.54215142e-01 3.99028391e-01 4.83987093e-01 7.21569061e-01
5.30952334e-01 -2.62378663e-01 -6.92583203e-01 -3.47329974e-01
4.56309468e-02 -1.32737741e-01 -1.50991648e-01 -1.74534038e-01
-1.04074681e+00 9.87170219e-01 2.86346197e-01 5.77721715e-01
-1.74507231e-01 1.46387771e-01 7.22651362e-01 1.80339500e-01
9.02398586e-01 6.35251582e-01 -4.26052362e-01 -2.19546333e-01
-1.02762723e+00 2.90068865e-01 8.20160627e-01 8.34518135e-01
6.34699047e-01 -1.85511127e-01 -5.69310963e-01 9.58504200e-01
1.56597182e-01 3.34332585e-01 4.75881726e-01 -5.61568379e-01
6.62546813e-01 7.92346835e-01 -8.77134800e-02 -1.05002522e+00
-2.62209356e-01 -4.98290896e-01 -2.66516626e-01 -4.45569940e-02
3.89276147e-01 -2.53692456e-02 -7.52683580e-01 1.34424460e+00
4.32130545e-01 -1.95028216e-01 -8.28722045e-02 3.33002210e-01
6.01896167e-01 7.84735620e-01 -5.88024035e-02 1.78033620e-01
1.56491268e+00 -8.92117083e-01 -6.45797729e-01 -4.78562832e-01
1.16182029e+00 -8.69924605e-01 1.36757863e+00 2.07745895e-01
-6.88173056e-01 -2.66493410e-01 -1.16821802e+00 -4.40208018e-01
-7.42239892e-01 2.22327664e-01 5.22804618e-01 7.80683160e-01
-1.00729978e+00 4.95375127e-01 -7.17664719e-01 -4.90333706e-01
4.83089328e-01 7.82312676e-02 -2.41572663e-01 -2.16060698e-01
-1.18568790e+00 9.00731504e-01 3.49223226e-01 -2.21933618e-01
-5.73430419e-01 -1.09624732e+00 -9.47112918e-01 2.30739489e-01
3.37853432e-01 -1.63066357e-01 9.95847523e-01 -1.53438032e-01
-1.02617967e+00 8.78347218e-01 -4.39199328e-01 -4.99141097e-01
4.62628782e-01 -4.75897253e-01 -3.19062203e-01 2.67486185e-01
2.82186687e-01 5.34232080e-01 6.52949989e-01 -8.32451224e-01
-4.74094540e-01 -1.30032003e-01 -3.06760054e-02 -8.80349427e-02
-9.79817688e-01 7.15971440e-02 -5.26784301e-01 -6.68495595e-01
-2.01803610e-01 -6.73766434e-01 1.66381791e-01 -4.03980836e-02
-6.58758581e-01 -6.08957469e-01 5.92290521e-01 -7.62652516e-01
1.76349747e+00 -2.36722732e+00 -3.58842283e-01 1.77869610e-02
2.91239679e-01 3.69653672e-01 -2.65548289e-01 9.09478068e-01
1.87440217e-01 3.23574632e-01 -7.35169873e-02 -4.81104702e-01
6.36378765e-01 -1.56494290e-01 -3.61951709e-01 5.01041114e-01
5.62139153e-01 8.73856604e-01 -9.15135741e-01 -6.15882456e-01
2.86391169e-01 5.37046909e-01 -7.12371230e-01 8.86261314e-02
-1.45681769e-01 -5.03765225e-01 -2.34713003e-01 6.71133459e-01
5.83700001e-01 -1.10574067e-01 1.77549422e-01 -1.97412986e-02
-2.02314556e-01 8.15436125e-01 -1.05394804e+00 1.48362291e+00
-4.59687561e-01 1.04747808e+00 -3.38048428e-01 -7.74916172e-01
8.84276628e-01 6.92965984e-02 2.15579763e-01 -5.68841934e-01
-1.16949931e-01 3.48487407e-01 -3.36615965e-02 -4.80415016e-01
7.33274758e-01 2.34024405e-01 -2.60383546e-01 3.63187075e-01
2.46927202e-01 8.44491199e-02 4.62810934e-01 4.62629229e-01
1.50662756e+00 -2.59025484e-01 4.01355088e-01 -3.88880700e-01
4.41346288e-01 -6.35802448e-02 2.56078094e-01 9.79408145e-01
-2.82022297e-01 3.87527406e-01 8.75371575e-01 -9.15588215e-02
-9.60845709e-01 -1.31804156e+00 -4.30949688e-01 1.23260999e+00
-2.81718284e-01 -8.10595632e-01 -6.56123161e-01 -9.10995662e-01
4.48592335e-01 1.03217793e+00 -5.72806001e-01 -1.13020860e-01
-5.28493822e-01 -8.31378639e-01 8.67091417e-01 4.93082345e-01
-4.47426140e-02 -8.11632276e-01 -1.59721538e-01 3.35125566e-01
-2.01153662e-03 -1.09139657e+00 -6.88802898e-01 4.01054412e-01
-5.70235550e-01 -6.86534584e-01 -6.21222615e-01 -8.86177719e-01
5.02803206e-01 1.56346783e-01 9.68274951e-01 -6.00706078e-02
-6.50523245e-01 3.09901625e-01 -3.99677098e-01 -4.53147382e-01
-5.11118352e-01 5.65379739e-01 5.10033369e-02 -2.34386921e-01
1.01100314e+00 -2.23986194e-01 -3.14169228e-01 -2.66006589e-01
-8.30931842e-01 -6.89196229e-01 5.33327758e-01 7.82844067e-01
3.56640965e-01 2.94684246e-02 5.89029968e-01 -8.31781268e-01
9.20474112e-01 -4.27707136e-01 -4.03130263e-01 1.31766275e-01
-7.37431645e-01 3.47219586e-01 5.91610610e-01 -3.72554153e-01
-8.36354554e-01 -4.60631996e-01 -2.89774507e-01 -2.15673104e-01
-3.00294191e-01 3.54425102e-01 8.84671062e-02 1.61391512e-01
6.67129159e-01 2.04990864e-01 -3.10263574e-01 -6.71177864e-01
3.34640622e-01 9.60201740e-01 2.04659894e-01 -4.46280479e-01
8.22020531e-01 3.82897556e-01 -4.87648755e-01 -1.17615008e+00
-9.50723350e-01 -7.92776406e-01 -4.69560504e-01 1.76814631e-01
6.83729291e-01 -8.53600502e-01 -3.67924422e-01 9.24352333e-02
-1.15783107e+00 -2.26851970e-01 -7.70136416e-02 3.65910858e-01
3.98806036e-02 6.45400822e-01 -8.35850239e-01 -8.50311756e-01
-2.74634480e-01 -7.34439135e-01 1.21711612e+00 -1.24802493e-01
-5.84582508e-01 -1.15179515e+00 1.90061048e-01 1.22321568e-01
3.50749254e-01 1.48443252e-01 1.21246052e+00 -1.02382445e+00
-3.10583442e-01 -4.90570784e-01 -2.29971096e-01 4.52063203e-01
2.03635171e-01 9.92280021e-02 -1.09328628e+00 -2.94788301e-01
-3.43435407e-01 -1.73332423e-01 1.00672579e+00 2.22443894e-01
9.21332955e-01 -2.16134027e-01 -3.38043392e-01 1.47878662e-01
1.49133909e+00 -4.06358153e-01 4.23822165e-01 5.80044448e-01
5.18951297e-01 5.07351995e-01 5.67732155e-01 5.11504173e-01
3.50309700e-01 5.36125422e-01 1.66232586e-01 3.80012542e-02
-2.29992598e-01 -3.00952792e-01 7.36301661e-01 9.62449372e-01
6.97938740e-01 -4.35401708e-01 -9.55221951e-01 9.76463199e-01
-1.46232986e+00 -8.27326715e-01 -1.10192649e-01 2.05383635e+00
9.31173623e-01 6.17299020e-01 -6.05168706e-03 2.28697032e-01
5.52058518e-01 4.71558273e-01 -3.94433856e-01 -9.12560821e-01
-3.59915607e-02 2.86046296e-01 6.28539562e-01 5.96489429e-01
-1.10439348e+00 9.30656791e-01 6.08332205e+00 1.12991154e+00
-8.15594614e-01 1.97381184e-01 2.02591464e-01 -2.80093998e-01
-4.08767968e-01 -3.20852339e-01 -1.33927023e+00 6.50993645e-01
1.02420855e+00 -1.52348310e-01 1.88891843e-01 9.19619262e-01
1.60034239e-01 -6.20002076e-02 -1.16744256e+00 6.14187896e-01
2.68124133e-01 -1.34989059e+00 7.91331679e-02 2.19802558e-01
6.69465721e-01 3.20901543e-01 3.16600390e-02 5.43573737e-01
2.16400009e-02 -7.82998502e-01 5.92136919e-01 6.06633835e-02
5.81929147e-01 -7.70726919e-01 8.62718284e-01 3.01324338e-01
-1.11114252e+00 7.27572665e-02 -5.93825698e-01 1.43156752e-01
-5.73276803e-02 9.81773078e-01 -1.04697716e+00 3.67627382e-01
6.86320722e-01 4.47963566e-01 -7.90259600e-01 7.57148087e-01
-1.77509800e-01 8.12765896e-01 -3.85008365e-01 -3.28323722e-01
4.93832052e-01 1.80915505e-01 5.28520465e-01 1.78384757e+00
2.06840754e-01 -5.66786051e-01 -9.02190357e-02 8.74893606e-01
-2.30949834e-01 3.75573635e-01 -5.86273015e-01 -5.31204164e-01
8.26529384e-01 1.14885318e+00 -5.17465413e-01 -3.53899896e-01
-5.64587057e-01 9.56023693e-01 6.39317095e-01 1.01045556e-01
-8.21649432e-01 -1.06344020e+00 9.91277099e-01 -8.55983645e-02
6.30801082e-01 -1.60381779e-01 -1.82900801e-01 -1.14541030e+00
2.04687595e-01 -7.96386123e-01 4.37298715e-01 -1.35996267e-01
-1.48396647e+00 3.63102943e-01 -2.09213778e-01 -9.94798541e-01
-8.39960063e-04 -8.67891133e-01 -5.67857623e-01 6.87318623e-01
-1.73121655e+00 -6.47591114e-01 -1.08100679e-02 3.98480846e-03
6.67345762e-01 5.62165566e-02 8.49206567e-01 5.41566372e-01
-6.55177116e-01 9.78754640e-01 3.36640298e-01 3.32331479e-01
9.56722438e-01 -1.56597900e+00 5.70002973e-01 9.71172333e-01
3.51780444e-01 8.52438390e-01 7.93818533e-01 -3.53376210e-01
-1.37288618e+00 -1.07218170e+00 1.54194963e+00 -7.25174487e-01
9.12486792e-01 -7.67826378e-01 -9.96794581e-01 5.38490534e-01
2.12965637e-01 -1.07352033e-01 7.21837342e-01 3.43209356e-01
-5.68267226e-01 -7.94625208e-02 -1.00474358e+00 5.86227834e-01
8.41347158e-01 -8.23795378e-01 -9.32100713e-01 4.83017534e-01
8.67950380e-01 -2.13153914e-01 -1.03826272e+00 -1.47961825e-02
1.96029544e-01 -8.60068619e-01 1.10317862e+00 -3.17111939e-01
4.06883091e-01 -1.61292896e-01 -2.26089418e-01 -1.14904559e+00
-3.78357589e-01 -3.53679448e-01 -3.96625161e-01 1.66138542e+00
5.40345728e-01 -8.58495593e-01 6.33690119e-01 2.19501674e-01
-1.23094425e-01 -8.43046248e-01 -8.11362684e-01 -1.17695761e+00
3.04161906e-01 -5.88132679e-01 4.29439634e-01 8.48722756e-01
2.15837911e-01 2.18113422e-01 -7.71691799e-02 4.45849895e-01
5.86710393e-01 -1.68800771e-01 6.34720385e-01 -9.10692453e-01
-2.94555843e-01 -4.99542326e-01 -3.45022321e-01 -1.05296278e+00
2.34779567e-01 -1.20959592e+00 -7.39786997e-02 -1.66581213e+00
5.98273948e-02 -1.90471202e-01 -4.65877324e-01 3.67232352e-01
-3.35833251e-01 2.98734605e-01 1.82378218e-01 -9.76739526e-02
-5.52463353e-01 4.11362350e-01 5.10493517e-01 -2.34057650e-01
-8.51700306e-02 -6.10371172e-01 -8.22626591e-01 5.53493917e-01
9.57864225e-01 -8.13732207e-01 -1.26482220e-02 -5.14295995e-01
1.27447039e-01 -7.47786045e-01 1.22253127e-01 -8.96573961e-01
-3.72286178e-02 2.67679811e-01 2.90596277e-01 -7.26468384e-01
1.30516201e-01 -3.25761914e-01 -7.73430884e-01 6.82035029e-01
-5.09817660e-01 -7.91109204e-02 4.12335753e-01 5.24990380e-01
-2.48634845e-01 -4.22836185e-01 4.78153855e-01 1.40073434e-01
-6.47467017e-01 6.62205294e-02 -5.21650970e-01 4.70270455e-01
7.56947100e-01 -3.83472443e-02 -4.74558592e-01 4.24406193e-02
-1.36579961e-01 2.22767085e-01 2.67116398e-01 5.31502068e-01
2.69774228e-01 -1.22187912e+00 -6.72683716e-01 8.18816125e-02
4.52685475e-01 -4.66458499e-01 3.84737626e-02 8.61688435e-01
-5.95411241e-01 8.07576656e-01 4.39045191e-01 -5.13634861e-01
-1.11835766e+00 4.84540522e-01 6.58152476e-02 -4.35591489e-01
-5.31119466e-01 9.23632741e-01 1.25609413e-01 -6.78726673e-01
3.00575942e-01 -5.44103324e-01 1.70902416e-01 3.89171481e-01
6.70089960e-01 5.45390248e-01 1.96573123e-01 -1.99307665e-01
-7.62777567e-01 2.74476230e-01 -2.66770512e-01 1.66230008e-01
1.40232563e+00 -1.22912750e-01 8.37889221e-03 5.60350657e-01
1.69005597e+00 6.81601882e-01 -9.41413462e-01 -3.30208153e-01
3.67226899e-01 -4.55841184e-01 9.27210823e-02 -5.98847330e-01
-4.22887623e-01 1.05877173e+00 3.45570475e-01 4.91709262e-01
6.27352059e-01 2.41483480e-01 9.54349279e-01 3.87787610e-01
1.72316339e-02 -1.27413321e+00 2.22595945e-01 5.21723032e-01
5.50010383e-01 -1.20490468e+00 1.44281425e-03 -2.65529692e-01
-4.66312915e-01 1.18331373e+00 3.30260515e-01 -2.84656107e-01
4.20038342e-01 4.01423424e-01 4.69728000e-02 -1.26244664e-01
-8.31764996e-01 -1.53796837e-01 3.11266690e-01 4.71028358e-01
6.29255176e-01 -7.97110796e-02 -2.69047886e-01 2.77980447e-01
-1.16757810e-01 -4.31680351e-01 3.37774217e-01 7.71245420e-01
-7.08767116e-01 -1.10839379e+00 -2.08501711e-01 6.88922524e-01
-6.43308580e-01 -4.90939528e-01 -3.90144020e-01 9.14755285e-01
-3.83404642e-02 7.85596371e-01 1.23132944e-01 -6.21170923e-02
3.52309048e-01 3.90055835e-01 2.50867218e-01 -7.69827545e-01
-5.95803618e-01 -1.09149128e-01 3.00649077e-01 -5.28851628e-01
1.48001611e-01 -7.61565328e-01 -1.13984978e+00 -4.38212723e-01
-4.61924046e-01 6.23790966e-03 3.92692447e-01 6.91407323e-01
6.16857052e-01 5.42270839e-01 4.60263193e-01 -5.74147701e-01
-9.41750407e-01 -1.03980243e+00 -5.66239119e-01 2.65299141e-01
5.30371308e-01 -3.83427650e-01 -7.97943711e-01 -4.24069881e-01] | [10.758840560913086, 8.726883888244629] |
d8b3bbba-901a-4c71-b499-80ee1a66470f | 3d-human-motion-generation-from-the-text-via | 2211.10003 | null | https://arxiv.org/abs/2211.10003v1 | https://arxiv.org/pdf/2211.10003v1.pdf | 3d human motion generation from the text via gesture action classification and the autoregressive model | In this paper, a deep learning-based model for 3D human motion generation from the text is proposed via gesture action classification and an autoregressive model. The model focuses on generating special gestures that express human thinking, such as waving and nodding. To achieve the goal, the proposed method predicts expression from the sentences using a text classification model based on a pretrained language model and generates gestures using the gate recurrent unit-based autoregressive model. Especially, we proposed the loss for the embedding space for restoring raw motions and generating intermediate motions well. Moreover, the novel data augmentation method and stop token are proposed to generate variable length motions. To evaluate the text classification model and 3D human motion generation model, a gesture action classification dataset and action-based gesture dataset are collected. With several experiments, the proposed method successfully generates perceptually natural and realistic 3D human motion from the text. Moreover, we verified the effectiveness of the proposed method using a public-available action recognition dataset to evaluate cross-dataset generalization performance. | ['Hanseok Ko', 'Jeongmin Bae', 'David K. Han', 'Junyeop Lee', 'Youngsuk Ryu', 'Gwantae Kim'] | 2022-11-18 | null | null | null | null | ['action-classification'] | ['computer-vision'] | [ 3.72116268e-01 -6.44878224e-02 -1.13577582e-01 -5.24564147e-01
-3.97445887e-01 -8.67394954e-02 9.66505051e-01 -9.34315383e-01
-4.14486080e-01 3.77197862e-01 8.35917652e-01 1.56035749e-02
2.63955444e-01 -6.62872910e-01 -5.21273077e-01 -9.04833496e-01
2.68073529e-01 2.70600468e-01 -1.77005097e-01 -5.99855222e-02
2.79270858e-01 5.40471017e-01 -1.35107744e+00 4.42425638e-01
4.09824640e-01 8.58312607e-01 2.27254659e-01 1.10316038e+00
-1.86367929e-01 1.36511576e+00 -8.45777929e-01 -5.68123199e-02
2.28969023e-01 -8.18558037e-01 -7.10365176e-01 4.83176261e-01
3.70362997e-02 -9.78891551e-01 -6.96169734e-01 4.60538924e-01
9.36961353e-01 7.55048156e-01 1.02943826e+00 -1.27373910e+00
-8.49287629e-01 5.80285430e-01 -1.47369280e-01 -4.83411163e-01
5.54621398e-01 4.39798743e-01 7.02987432e-01 -6.97150826e-01
8.24211836e-01 1.69403493e+00 1.82557464e-01 1.22954965e+00
-6.33151174e-01 -5.37413597e-01 8.58572125e-02 2.42603824e-01
-1.04491723e+00 -1.94983199e-01 9.84198630e-01 -4.21963811e-01
1.30272079e+00 2.80534983e-01 6.63116336e-01 1.89551139e+00
2.16204479e-01 1.48781013e+00 6.04614913e-01 -4.72090751e-01
2.23494023e-01 -4.74197239e-01 -1.57077089e-01 5.52196860e-01
-5.22717476e-01 -2.15664674e-02 -4.95257705e-01 2.79102683e-01
1.08716702e+00 1.96994111e-01 -1.05306640e-01 -3.15281749e-02
-1.40884602e+00 5.46508372e-01 2.96271056e-01 4.17555064e-01
-5.21992207e-01 3.46251398e-01 5.83581209e-01 -1.36813730e-01
2.95256853e-01 1.57041685e-03 -9.93625894e-02 -6.61150515e-01
-7.37727582e-01 3.82972300e-01 4.99060422e-01 9.74227846e-01
-1.63216248e-01 5.57406723e-01 -5.41147411e-01 6.66109025e-01
4.21653539e-01 8.45220327e-01 1.17513728e+00 -8.22238803e-01
6.51673615e-01 6.65302932e-01 1.00176327e-01 -9.62524712e-01
-4.63795036e-01 3.64668995e-01 -1.13631523e+00 1.52037621e-01
2.89430737e-01 -1.62788138e-01 -1.18109024e+00 1.54938483e+00
9.25832763e-02 1.73095927e-01 3.86899263e-01 1.35980427e+00
1.12548554e+00 9.77555335e-01 2.87279606e-01 -1.24410205e-01
8.88080180e-01 -1.10862875e+00 -1.03091407e+00 3.07468683e-01
8.41780007e-01 -5.16306281e-01 1.52251184e+00 4.06678975e-01
-9.16176975e-01 -1.01026726e+00 -7.77180791e-01 -2.92741656e-01
-1.02775795e-02 8.85478139e-01 4.71611470e-01 4.65961426e-01
-5.96905112e-01 5.20218730e-01 -1.25718582e+00 -4.27532196e-01
5.83276227e-02 7.25716539e-03 -2.39013925e-01 2.87379563e-01
-1.02283752e+00 5.01322985e-01 3.84511381e-01 3.79442632e-01
-8.57212722e-01 4.93674241e-02 -9.94421661e-01 -1.15303293e-01
-1.54093400e-01 -7.06029117e-01 1.07910812e+00 -1.02572155e+00
-2.31495142e+00 6.22457922e-01 -1.64045036e-01 -4.28700596e-01
9.22111750e-01 -5.43117583e-01 -2.79033601e-01 1.76805958e-01
-2.64533073e-01 9.04602945e-01 9.68713582e-01 -7.83253253e-01
-1.69227079e-01 -4.41542536e-01 -1.72259301e-01 4.23905969e-01
-4.04272825e-01 -2.17898399e-01 -2.48495564e-01 -1.24072862e+00
1.21413566e-01 -1.04075575e+00 -3.77423614e-02 -2.08436385e-01
-5.70177257e-01 -3.61298978e-01 9.68579650e-01 -8.95069301e-01
1.00440264e+00 -1.98108768e+00 5.25281489e-01 -1.31160319e-01
-2.90903062e-01 1.85560212e-01 -4.03908730e-01 3.95710945e-01
1.60860538e-01 -4.23704796e-02 -1.47558227e-01 -3.72847736e-01
2.93269604e-01 3.63157839e-01 -7.54831970e-01 1.44752726e-01
2.77575582e-01 1.14512539e+00 -6.83919787e-01 -2.85778910e-01
5.69382727e-01 6.37177825e-01 -5.36116123e-01 6.52034104e-01
-3.57377261e-01 8.39919031e-01 -7.01799273e-01 4.69071716e-01
1.99086756e-01 3.20073552e-02 3.08720488e-02 -4.08756956e-02
2.28506669e-01 4.32896540e-02 -1.01905882e+00 2.01844120e+00
-5.44062078e-01 5.70720136e-01 -6.03902638e-01 -9.22952890e-01
1.33327532e+00 3.97879362e-01 5.03903687e-01 -6.27591014e-01
4.87327576e-01 -5.64837977e-02 -2.13851452e-01 -1.23488605e+00
7.07676291e-01 2.65657231e-02 -2.31859852e-02 6.56524658e-01
-1.40821919e-01 -3.57743949e-01 -1.41743451e-01 -1.62969843e-01
9.44060087e-01 9.28875208e-01 -1.53788969e-01 5.10108590e-01
6.50747657e-01 -3.59043479e-02 2.33110443e-01 4.28171486e-01
-6.35380223e-02 5.85748792e-01 1.87366933e-01 -6.60688400e-01
-9.99462545e-01 -8.43890131e-01 4.79484111e-01 1.08228731e+00
8.90601054e-02 -3.24873701e-02 -7.43407309e-01 -7.33497381e-01
-4.03967172e-01 8.67587626e-01 -5.09625196e-01 -1.81621224e-01
-8.53949070e-01 -5.15302718e-01 8.65380347e-01 1.00581706e+00
1.00917900e+00 -1.57607341e+00 -8.63579631e-01 8.36960673e-02
-4.54717547e-01 -1.11610067e+00 -3.99969071e-01 -5.76520503e-01
-8.06778848e-01 -8.48569274e-01 -1.28231251e+00 -9.27092195e-01
6.80301905e-01 -1.79825947e-01 3.65510225e-01 -1.96174040e-01
-2.16568515e-01 4.07435209e-01 -8.73509943e-01 -2.05077678e-01
-5.85683167e-01 -2.63555378e-01 1.39227748e-01 1.68197662e-01
4.58328754e-01 -3.27492386e-01 -4.08223122e-01 3.10386777e-01
-1.06752074e+00 2.84760922e-01 5.75401545e-01 1.07442224e+00
1.90247700e-01 -3.52232695e-01 3.27178180e-01 1.29904747e-02
8.63653362e-01 1.52669400e-01 -9.57782045e-02 8.61549452e-02
-1.55815128e-02 2.64655799e-01 5.90531647e-01 -9.18773115e-01
-1.48997307e+00 3.06806177e-01 -3.09644043e-01 -6.03841186e-01
-4.09617424e-01 3.88314426e-02 -3.73095542e-01 5.55401921e-01
5.12011051e-01 5.69606125e-01 -1.36190457e-02 -4.47067052e-01
6.11917257e-01 1.04115963e+00 5.31492293e-01 -3.71686190e-01
3.84330004e-01 5.39316177e-01 -2.31107231e-02 -1.01688719e+00
3.90298963e-02 -3.65663841e-02 -8.91831994e-01 -3.76879126e-01
1.37403870e+00 -6.87124610e-01 -8.93265784e-01 1.07350528e+00
-1.52941155e+00 -8.45959365e-01 -8.79021063e-02 1.07004857e+00
-1.12093675e+00 3.99105549e-01 -9.47541833e-01 -1.17961383e+00
-4.44918513e-01 -1.28451800e+00 1.52947199e+00 1.95228476e-02
-4.06740427e-01 -6.90978110e-01 -6.33202959e-03 3.40255946e-01
9.57978442e-02 4.60130304e-01 6.92716360e-01 -4.99603063e-01
-3.86354268e-01 -1.79133847e-01 1.52535856e-01 3.51895392e-01
1.88315526e-01 5.40522225e-02 -6.41244709e-01 9.85275954e-02
-1.24717854e-01 -6.55097306e-01 7.67603397e-01 4.26404744e-01
1.13137734e+00 -4.68584806e-01 6.16202503e-02 5.55087030e-01
6.00410998e-01 5.72494566e-01 8.99181247e-01 8.85349587e-02
8.46343160e-01 5.93595505e-01 6.88839138e-01 7.34248459e-01
6.45764470e-02 7.67466247e-01 1.44553766e-01 1.93050206e-01
-1.18084796e-01 -6.40006065e-01 6.61006451e-01 8.57516170e-01
-4.96232569e-01 -3.85911405e-01 -7.07445025e-01 1.52721867e-01
-1.83037233e+00 -1.21936703e+00 -1.57936309e-02 1.64371586e+00
6.05458617e-01 -1.53688759e-01 8.74546021e-02 1.87291160e-01
4.65781063e-01 2.65783310e-01 -5.71111262e-01 -2.84370422e-01
-1.05110221e-01 1.44128188e-01 -7.93400258e-02 4.11123991e-01
-1.09491456e+00 1.35454273e+00 5.29667997e+00 7.38121033e-01
-1.51219952e+00 -3.85330051e-01 3.77443790e-01 -2.27615729e-01
-3.41026969e-02 -6.30362511e-01 -7.35737205e-01 3.85875970e-01
4.65490013e-01 7.60118589e-02 1.34216994e-01 9.28699017e-01
9.03981328e-01 3.48261207e-01 -8.86935055e-01 1.26211143e+00
3.53861511e-01 -1.07071280e+00 6.98364496e-01 -2.57552028e-01
6.21857584e-01 -5.33763468e-01 1.80029520e-03 4.90820915e-01
2.57496744e-01 -1.11020124e+00 7.30309486e-01 8.38751316e-01
8.27889025e-01 -4.94957387e-01 6.03826880e-01 5.40457726e-01
-9.21232522e-01 -1.90808713e-01 -1.09092280e-01 -3.86940747e-01
4.85350043e-01 -2.25963742e-01 -8.53500783e-01 3.28142047e-01
3.48333389e-01 9.01828825e-01 -3.11740249e-01 2.96022892e-01
-5.74003994e-01 5.88618577e-01 -5.11591733e-02 -6.05309665e-01
4.48195070e-01 -1.12521291e-01 5.76870918e-01 1.21867514e+00
5.39097488e-01 4.16028321e-01 -1.73228867e-02 7.76656508e-01
1.31273851e-01 2.60254145e-01 -6.96737766e-01 -4.24337745e-01
3.09969625e-03 8.60184312e-01 -4.69915122e-01 -5.06304741e-01
1.42008424e-01 1.44678771e+00 -3.73920172e-01 7.22178698e-01
-9.05198932e-01 -5.22713125e-01 3.14541936e-01 -3.79906356e-01
-5.95779791e-02 -6.31830215e-01 -1.80724129e-01 -1.28405106e+00
1.97723970e-01 -7.22228289e-01 -5.67989908e-02 -1.26379311e+00
-8.61872494e-01 6.29992723e-01 1.16491271e-02 -1.57009351e+00
-1.04516184e+00 -7.93962657e-01 -6.06263340e-01 5.02519608e-01
-5.91355860e-01 -1.42469549e+00 -6.91080928e-01 6.66774571e-01
1.02594566e+00 -4.30070698e-01 8.03961277e-01 -2.00243928e-02
-3.34475696e-01 5.89353025e-01 -3.12237799e-01 4.89917696e-01
5.46254456e-01 -7.20622480e-01 5.68842113e-01 5.89796424e-01
2.30447337e-01 4.35567737e-01 4.42748874e-01 -7.08759069e-01
-1.32059741e+00 -1.16455460e+00 6.33349895e-01 -2.96818823e-01
4.16860640e-01 -3.27787280e-01 -7.23115206e-01 7.41166651e-01
-8.72710422e-02 -3.95687610e-01 3.26526672e-01 -8.24739337e-01
8.28833356e-02 4.24187958e-01 -8.71771812e-01 9.62035716e-01
1.40897202e+00 -2.52393097e-01 -8.22063386e-01 2.30686411e-01
7.20896184e-01 -5.71082056e-01 -4.66483325e-01 5.67138195e-01
9.24471796e-01 -5.20621538e-01 8.20016563e-01 -1.14478385e+00
9.27856326e-01 -1.39768980e-02 -2.57956356e-01 -1.20053673e+00
1.20804876e-01 -5.52762210e-01 -7.20238760e-02 1.00271511e+00
-1.63685530e-02 -1.25383273e-01 9.73972976e-01 5.93518734e-01
-9.00983140e-02 -5.09321094e-01 -6.75671935e-01 -8.38367820e-01
3.07979193e-02 -5.33239901e-01 6.48747206e-01 7.65464246e-01
-3.06642409e-02 1.99718580e-01 -8.27358961e-01 -2.89803803e-01
3.15777399e-02 -9.07747969e-02 1.37187982e+00 -5.38328946e-01
-3.10991079e-01 -4.09984171e-01 -7.01700389e-01 -1.75089335e+00
5.82388580e-01 -8.70175540e-01 2.34639123e-01 -1.52767646e+00
-1.14720047e-01 1.67409465e-01 2.78764665e-01 3.70137274e-01
-4.62633893e-02 -1.26039743e-01 4.56252545e-01 3.20292801e-01
-2.07243890e-01 1.13805103e+00 1.59785295e+00 -4.23743486e-01
-6.60287201e-01 9.83174369e-02 2.91686743e-01 7.32377589e-01
6.03805780e-01 1.29120618e-01 -5.18243074e-01 -5.81403494e-01
-4.00650322e-01 2.88104087e-01 5.80434859e-01 -8.35355163e-01
-1.18536942e-01 -2.87058920e-01 6.25107348e-01 -7.71827042e-01
6.60772383e-01 -5.46915948e-01 -2.04831600e-01 6.49139106e-01
-9.12670493e-01 -2.12938994e-01 1.39081543e-02 5.03206432e-01
-1.88773140e-01 1.67224586e-01 2.88106173e-01 2.75173504e-02
-8.94122064e-01 8.85335952e-02 -4.87759233e-01 -3.37839752e-01
9.20106113e-01 -1.70479670e-01 1.01976648e-01 -7.88031399e-01
-1.04200876e+00 1.97897274e-02 4.96373512e-02 9.45004642e-01
1.04313493e+00 -1.80126476e+00 -7.78020322e-01 3.37108642e-01
-4.00259905e-02 -1.74998641e-01 3.10071230e-01 5.29018760e-01
-5.91193259e-01 5.48667848e-01 -4.63034838e-01 -5.67968845e-01
-1.24772334e+00 2.39581734e-01 2.07057953e-01 -9.91641134e-02
-7.88816869e-01 5.83860278e-01 -5.07917395e-03 -6.20730400e-01
4.41946894e-01 -6.92949593e-01 -2.82811224e-01 -3.87048602e-01
7.09348679e-01 5.49470961e-01 -5.08661151e-01 -8.11801791e-01
-9.83270407e-02 5.60412645e-01 5.78383148e-01 -6.43549204e-01
1.07660961e+00 1.16098851e-01 3.54584038e-01 5.06960392e-01
1.13517690e+00 -3.88076127e-01 -1.45989275e+00 2.07440883e-01
-1.92016155e-01 -4.33628798e-01 -3.92470270e-01 -6.07449949e-01
-7.68838108e-01 1.30396914e+00 5.29411435e-01 -3.36629957e-01
1.05753684e+00 -3.61936629e-01 1.08629096e+00 8.07277322e-01
2.51603574e-01 -1.17188287e+00 6.63287938e-01 5.91801107e-01
1.28592181e+00 -1.01706827e+00 -4.66866761e-01 1.49170846e-01
-1.09079337e+00 1.38290107e+00 6.81935310e-01 -1.79031268e-01
1.68367401e-01 -2.13913359e-02 2.17645749e-01 3.20979267e-01
-4.95231330e-01 -3.91385751e-03 2.32523412e-01 6.13929868e-01
5.15042782e-01 1.62747353e-01 -5.87661088e-01 7.82441735e-01
-4.79705960e-01 4.55450386e-01 4.45996612e-01 7.09678769e-01
-1.82835147e-01 -7.51584113e-01 -3.13957721e-01 -3.16502340e-02
-7.40126520e-02 2.55103201e-01 -7.21365154e-01 7.05595732e-01
-2.70400345e-01 9.09169197e-01 2.48199314e-01 -8.15797210e-01
3.39499772e-01 4.20438111e-01 5.61669111e-01 -2.65931845e-01
-1.01007737e-01 1.37526065e-01 -9.55601782e-02 -6.82051718e-01
-5.77531338e-01 -3.71926904e-01 -1.80493665e+00 1.46701604e-01
1.02998786e-01 -3.36360663e-01 6.90821230e-01 9.96896923e-01
2.66272992e-01 4.72655207e-01 6.42596662e-01 -1.25202823e+00
-8.19473684e-01 -1.43623376e+00 -4.67286706e-01 9.19565141e-01
-9.29796994e-02 -5.96230268e-01 -2.41624147e-01 6.10779583e-01] | [5.649774074554443, -0.11890541762113571] |
b791fa06-08f8-4252-8810-e2fbdb7727f5 | character-independent-font-identification | 2001.08893 | null | https://arxiv.org/abs/2001.08893v1 | https://arxiv.org/pdf/2001.08893v1.pdf | Character-independent font identification | There are a countless number of fonts with various shapes and styles. In addition, there are many fonts that only have subtle differences in features. Due to this, font identification is a difficult task. In this paper, we propose a method of determining if any two characters are from the same font or not. This is difficult due to the difference between fonts typically being smaller than the difference between alphabet classes. Additionally, the proposed method can be used with fonts regardless of whether they exist in the training or not. In order to accomplish this, we use a Convolutional Neural Network (CNN) trained with various font image pairs. In the experiment, the network is trained on image pairs of various fonts. We then evaluate the model on a different set of fonts that are unseen by the network. The evaluation is performed with an accuracy of 92.27%. Moreover, we analyzed the relationship between character classes and font identification accuracy. | ['Shota Harada', 'Seiichi Uchida', 'Daichi Haraguchi', 'Brian Kenji Iwana', 'Yuto Shinahara'] | 2020-01-24 | null | null | null | null | ['font-recognition'] | ['computer-vision'] | [ 8.99000987e-02 -7.06814945e-01 3.52482349e-01 -3.95619988e-01
2.56235123e-01 -8.81834507e-01 4.51812536e-01 -7.87400827e-02
-3.36335182e-01 7.68218756e-01 -3.48012030e-01 -3.55751157e-01
4.32858944e-01 -7.48651981e-01 -5.25336325e-01 -6.79394007e-01
5.43230295e-01 1.87111974e-01 3.91451269e-01 -1.69337332e-01
3.65792453e-01 6.16537988e-01 -1.64786291e+00 4.96043086e-01
9.86690700e-01 9.59986210e-01 2.53231198e-01 5.38941026e-01
-4.90696162e-01 4.26631063e-01 -1.12085557e+00 -5.02966285e-01
5.31975389e-01 -3.22936684e-01 -3.18397731e-01 4.30116475e-01
7.44934380e-01 -4.11218166e-01 -2.52361387e-01 1.29500413e+00
4.06410456e-01 -2.50344872e-01 7.55289435e-01 -8.21154952e-01
-8.64533365e-01 3.69432956e-01 -5.88374197e-01 2.45396812e-02
5.13202175e-02 1.12204559e-01 4.44199055e-01 -7.25487173e-01
2.23406494e-01 1.17868900e+00 3.63483042e-01 5.61946034e-01
-8.85360301e-01 -8.74767244e-01 1.59793422e-01 -5.85172176e-02
-1.41876030e+00 -9.96140987e-02 7.89447546e-01 -5.34758151e-01
2.49489382e-01 2.16073349e-01 5.60865641e-01 1.18357551e+00
-2.97137909e-02 4.95559722e-01 1.41051483e+00 -6.31263196e-01
-6.52675182e-02 7.18856752e-01 3.57053757e-01 3.47274631e-01
3.71238440e-01 -1.89992994e-01 -1.48929641e-01 2.48857364e-01
6.91600561e-01 -1.74002141e-01 -3.28571975e-01 -1.09492101e-01
-1.08026505e+00 5.36955178e-01 1.30207986e-02 4.48513776e-01
1.78328604e-01 -4.88621324e-01 4.46993500e-01 3.83169860e-01
-3.43788899e-02 4.07883793e-01 -2.87262827e-01 -9.89239737e-02
-5.38118064e-01 -1.39057159e-01 9.39559519e-01 1.03341329e+00
3.12732160e-01 1.08892351e-01 1.58579320e-01 1.09097624e+00
8.41070786e-02 6.12757802e-01 7.05416203e-01 -7.63625279e-02
6.10137284e-01 6.30219400e-01 2.09032252e-01 -1.07061684e+00
-8.12737346e-02 -1.42835408e-01 -9.37839806e-01 4.59758371e-01
8.82766008e-01 -2.43190840e-01 -8.80292416e-01 1.42011607e+00
-2.12960765e-02 -3.33847672e-01 1.83538944e-01 9.31350172e-01
7.88118124e-01 6.24233723e-01 -3.07957381e-01 -5.43596633e-02
1.44474232e+00 -9.45899606e-01 -7.35241413e-01 -3.65657173e-02
2.97370732e-01 -1.24719107e+00 1.35993004e+00 5.73101997e-01
-5.36756754e-01 -9.93687212e-01 -1.21483183e+00 1.74550787e-01
-8.61528456e-01 8.52980435e-01 4.00392830e-01 1.21852887e+00
-6.32088959e-01 3.24786454e-01 -1.40657768e-01 -6.24045432e-01
-4.83521819e-02 2.48442456e-01 -1.87303141e-01 1.41188905e-01
-1.14287150e+00 7.46316373e-01 5.73355377e-01 1.82333991e-01
-2.48156369e-01 6.50169477e-02 -2.76684821e-01 4.19800868e-03
-4.89727184e-02 2.08252698e-01 9.10332084e-01 -1.47894037e+00
-1.33998454e+00 6.31810486e-01 8.62957723e-03 -6.68570250e-02
9.62738693e-01 9.99598131e-02 -1.15047300e+00 -2.96713352e-01
-1.45047992e-01 4.62256998e-01 9.70389426e-01 -1.14774680e+00
-6.87692344e-01 -1.26142099e-01 9.27903578e-02 1.29276231e-01
-9.80234563e-01 1.94809273e-01 -4.40043867e-01 -4.20806319e-01
5.96958511e-02 -8.10288668e-01 4.36633319e-01 1.02909170e-01
-6.49508834e-01 -1.39492676e-01 1.17421579e+00 -5.83439887e-01
1.12228405e+00 -2.29054785e+00 -4.29377109e-01 1.69003814e-01
-1.13387547e-01 6.07570291e-01 -1.23829760e-01 3.19926113e-01
-1.31851241e-01 1.84874684e-01 9.81470644e-02 -2.27026884e-02
-2.63857543e-01 1.43174484e-01 -4.06294644e-01 4.26497683e-02
3.54013592e-01 2.62966692e-01 -4.11196113e-01 -5.29487610e-01
1.39677912e-01 2.22706676e-01 5.31605124e-01 1.51345521e-01
-1.76892499e-03 4.05522704e-01 -2.74933338e-01 8.14892709e-01
1.18286407e+00 -7.49056488e-02 2.38186538e-01 -3.17872673e-01
-2.83151090e-01 -3.40773940e-01 -1.32803833e+00 8.01945329e-01
-2.12582648e-01 1.03092158e+00 -3.60887527e-01 -5.40468574e-01
1.34731507e+00 7.03345016e-02 -2.22778484e-01 -4.70613956e-01
4.19189692e-01 4.95779335e-01 4.03232008e-01 -5.14353752e-01
8.06083620e-01 3.65878493e-01 3.89964908e-01 3.16503882e-01
-4.60462570e-01 1.54341027e-01 7.59242713e-01 -3.07773501e-01
6.14268541e-01 -2.80320406e-01 -1.63814668e-02 -1.93827778e-01
8.04155171e-01 -2.24660963e-01 3.09332073e-01 6.66353464e-01
-2.31357634e-01 8.56883943e-01 5.99908948e-01 -6.98172450e-01
-1.34929883e+00 -7.91580856e-01 -6.01000130e-01 3.60655546e-01
2.90187478e-01 -1.06547758e-01 -6.16448760e-01 -6.24059975e-01
-8.14376995e-02 3.83326173e-01 -5.98019004e-01 1.33161873e-01
-5.81614256e-01 -6.74674153e-01 5.38513482e-01 6.07214928e-01
8.53564262e-01 -9.97398794e-01 -3.62334818e-01 -1.66528746e-01
1.17653467e-01 -1.36072946e+00 -5.56074798e-01 8.23001042e-02
-6.29574895e-01 -1.23274446e+00 -9.93397832e-01 -1.09186578e+00
1.13143480e+00 3.79439384e-01 9.46312785e-01 6.52023777e-02
1.16072763e-02 -1.45160213e-01 -4.67237413e-01 -6.24256253e-01
-7.53022611e-01 2.73714602e-01 9.27172676e-02 3.01240385e-01
6.52350545e-01 -1.31221175e-01 -3.88467133e-01 7.32466161e-01
-9.38153863e-01 3.21457237e-02 5.33220410e-01 8.36527944e-01
1.90393120e-01 3.82709414e-01 1.05873384e-01 -6.29862428e-01
9.53205585e-01 -3.45932841e-02 -9.76831615e-01 5.92265785e-01
-3.98587465e-01 -1.38842836e-01 1.24892938e+00 -1.04196811e+00
-7.88029432e-01 3.10995489e-01 2.02360123e-01 -2.79361039e-01
-4.39220518e-01 -1.48508837e-02 -3.16472948e-01 -2.42119789e-01
5.29378235e-01 4.10787582e-01 -5.85642736e-03 -5.18358469e-01
-9.61146578e-02 1.25398004e+00 2.05050349e-01 -3.83594900e-01
9.43521142e-01 1.69990227e-01 -2.07184985e-01 -1.03622067e+00
-3.35126579e-01 1.29691616e-01 -7.28044331e-01 -3.68517399e-01
7.56789744e-01 -6.01307869e-01 -6.75003231e-01 1.01178920e+00
-1.24994588e+00 2.14447334e-01 1.32390290e-01 4.90546703e-01
2.03036726e-01 8.10335636e-01 -4.37992305e-01 -9.81877387e-01
-1.15454361e-01 -1.31962883e+00 6.77827835e-01 6.30255401e-01
3.58654223e-02 -7.79201388e-01 -3.80509526e-01 -9.11899135e-02
2.04682305e-01 8.59715343e-02 1.00912368e+00 -6.36504769e-01
-4.27542806e-01 -3.66299242e-01 -5.05343139e-01 6.29539788e-01
6.74616694e-01 5.27655005e-01 -9.64419842e-01 -2.72064894e-01
-2.19524443e-01 -2.31877699e-01 5.47602713e-01 -6.52557164e-02
1.17510605e+00 1.20980769e-01 -1.26864789e-02 5.07891774e-01
1.34342790e+00 7.58611321e-01 6.65864766e-01 4.67633963e-01
6.86928451e-01 6.15047455e-01 5.30729473e-01 3.04440677e-01
-2.82264382e-01 5.64353645e-01 3.14979285e-01 -1.16045266e-01
7.25121498e-02 -1.34956896e-01 1.54110938e-01 7.49053776e-01
3.72707658e-02 -5.58375955e-01 -7.43143201e-01 2.18723044e-01
-1.40756321e+00 -6.80815876e-01 -4.53279644e-01 2.42489958e+00
7.60256171e-01 3.65106374e-01 6.25933185e-02 2.09605947e-01
1.26680326e+00 6.23176396e-02 -5.96580148e-01 -7.79136717e-01
-6.54252529e-01 -2.08454594e-01 5.73755562e-01 -2.17759069e-02
-1.21773648e+00 8.74502480e-01 6.28586912e+00 7.73307145e-01
-1.55357420e+00 -6.11348391e-01 8.78595829e-01 3.00012410e-01
2.15395287e-01 -1.94933951e-01 -1.12216687e+00 9.67322469e-01
4.40681577e-01 1.99826017e-01 4.41796571e-01 5.95031977e-01
-1.83047354e-02 -1.85966790e-01 -8.79565775e-01 1.08082831e+00
2.92468339e-01 -7.70338833e-01 1.09770149e-01 -3.99941355e-02
6.58699453e-01 -5.73939681e-01 1.59691557e-01 -1.18818097e-01
-1.06289223e-01 -9.57261980e-01 6.47795141e-01 3.16458732e-01
8.99141550e-01 -5.85743189e-01 7.56187737e-01 3.91638637e-01
-1.08578253e+00 -1.73346013e-01 -8.98844123e-01 1.01324387e-01
-2.19075099e-01 4.31179971e-01 -6.23945296e-01 3.31782162e-01
6.60754919e-01 4.68285441e-01 -8.07538629e-01 1.10470855e+00
-1.86922774e-01 2.73924887e-01 -3.67689699e-01 -5.70403397e-01
-3.85685377e-02 -5.28180122e-01 1.91432998e-01 9.52342510e-01
5.41115344e-01 -4.24746186e-01 -8.26428682e-02 9.12601292e-01
-1.40478194e-01 3.79969686e-01 -5.84414959e-01 -3.75861764e-01
4.87762779e-01 1.31246269e+00 -8.20576072e-01 -2.64210790e-01
-7.14604497e-01 1.14403939e+00 4.40463014e-02 3.36950898e-01
-9.23189878e-01 -6.87723041e-01 4.07303244e-01 -2.00436473e-01
2.57983238e-01 -1.05656818e-01 -2.33103395e-01 -1.13426316e+00
5.05054474e-01 -1.01550722e+00 -1.49791703e-01 -9.11367476e-01
-1.57272577e+00 9.49836135e-01 -3.85198414e-01 -1.74502528e+00
2.30549544e-01 -1.26422834e+00 -6.71606839e-01 1.25566602e+00
-1.31823087e+00 -9.49629366e-01 -7.33770430e-01 2.72327363e-01
5.08642614e-01 -6.37170434e-01 6.19896352e-01 2.96659499e-01
-7.69206882e-01 9.97989416e-01 5.63745618e-01 6.95587218e-01
9.25812721e-01 -1.15830326e+00 3.99259746e-01 8.73399019e-01
-2.34311923e-01 5.65620124e-01 4.85172719e-01 -5.48983157e-01
-9.29632306e-01 -9.44690287e-01 6.72215223e-01 -1.54669732e-01
5.16340375e-01 -6.07847512e-01 -8.72337222e-01 1.76901802e-01
3.70137185e-01 -2.13392943e-01 4.27866340e-01 -1.67891920e-01
-5.31937301e-01 -2.04725862e-01 -1.02251446e+00 7.85720289e-01
6.23680592e-01 -3.01773340e-01 -2.57161021e-01 3.90530199e-01
4.39225078e-01 -4.54110801e-01 -7.44703591e-01 4.77707833e-02
7.36547112e-01 -1.06159413e+00 5.58410227e-01 -4.87921298e-01
5.04105568e-01 -4.71596032e-01 -7.02065676e-02 -1.03394926e+00
-1.03083327e-01 -1.38478920e-01 2.75864959e-01 1.52521443e+00
5.27475655e-01 -9.73720670e-01 8.79679441e-01 5.89436591e-01
2.67203063e-01 -4.81306016e-01 -4.22643065e-01 -1.17275023e+00
2.60528713e-01 3.60200346e-01 1.04732704e+00 9.21920240e-01
-4.21738148e-01 1.84364393e-01 -5.14077246e-01 4.17654365e-02
1.82877928e-01 2.10157171e-01 7.48118103e-01 -1.41583800e+00
-2.49308378e-01 -4.18385565e-01 -4.38525647e-01 -9.67351794e-01
-3.74946706e-02 -3.38677675e-01 -2.34497532e-01 -9.29271817e-01
1.92345455e-01 -3.11573803e-01 -5.38818873e-02 1.97028145e-01
-1.54873431e-01 5.32250628e-02 2.80784607e-01 3.81420910e-01
6.25097975e-02 2.81037003e-01 1.43418574e+00 -3.60758215e-01
-2.29129300e-01 1.46353096e-01 -3.17546874e-01 4.92408037e-01
1.32847071e+00 -3.74921411e-02 -3.56922656e-01 -4.88143533e-01
1.93582922e-02 -4.11871493e-01 2.77037453e-03 -1.19642985e+00
-1.44300805e-02 2.24890560e-01 9.45952415e-01 -8.06930602e-01
3.14020991e-01 -1.01769280e+00 -2.21194878e-01 2.85389513e-01
-2.24112004e-01 4.84184474e-01 2.50896305e-01 2.72261888e-01
-2.27037892e-01 -7.23896086e-01 8.86677980e-01 -3.65480245e-03
-6.51283920e-01 5.53774647e-02 -3.63732338e-01 -1.48600668e-01
9.69924986e-01 -8.11882555e-01 -5.83755195e-01 -1.66753635e-01
-3.02758127e-01 -1.69598639e-01 7.86327124e-01 5.72199345e-01
4.11807746e-01 -1.29141533e+00 -4.90800381e-01 4.12793964e-01
3.58823538e-01 -4.94538933e-01 -1.47884181e-02 5.64623237e-01
-7.57628202e-01 2.93175310e-01 -5.32076180e-01 -4.17352706e-01
-1.79233110e+00 6.29574299e-01 4.93357241e-01 1.01093948e-01
-2.31059834e-01 3.45411301e-01 3.65542099e-02 -5.07627249e-01
3.21453303e-01 -1.75705314e-01 -5.90945363e-01 -7.78168216e-02
5.12537837e-01 1.63841501e-01 -1.17867090e-01 -7.26761401e-01
6.08593896e-02 8.52968335e-01 -4.17064220e-01 2.03009546e-01
5.57681084e-01 1.46031752e-01 -1.27829596e-01 4.66577083e-01
1.07652342e+00 2.96327591e-01 -1.03529787e+00 9.81437191e-02
-2.41368622e-01 -8.28487098e-01 -6.86938643e-01 -8.60726297e-01
-1.09168506e+00 1.12592816e+00 1.24619246e+00 5.24161994e-01
1.09582353e+00 -6.68318927e-01 6.19653106e-01 4.89845723e-01
2.01233342e-01 -1.40007091e+00 -1.76337838e-01 5.49467921e-01
5.78094244e-01 -1.51005125e+00 -2.07828179e-01 -6.15686238e-01
-5.09176075e-01 1.63181078e+00 1.03038967e+00 1.09548837e-01
3.19137067e-01 1.36032566e-01 4.56834137e-01 4.40299898e-01
-2.21905097e-01 -2.09237132e-02 9.57595482e-02 6.86672330e-01
7.02882469e-01 1.10875042e-02 -6.60591841e-01 1.87944993e-01
-1.97964847e-01 -2.65992403e-01 8.35663021e-01 6.69479132e-01
-4.13169026e-01 -1.48304331e+00 -6.86216533e-01 4.67752486e-01
-8.21541026e-02 -1.99006990e-01 -8.98553312e-01 8.27450275e-01
3.52186203e-01 1.11699033e+00 1.04498550e-01 -6.54692173e-01
1.93536296e-01 1.87023804e-02 2.37992793e-01 -1.02239855e-01
-5.77403188e-01 -1.47756740e-01 -5.97207397e-02 1.69294730e-01
3.15434113e-02 -6.19753182e-01 -7.54204154e-01 -4.73757029e-01
-4.88169760e-01 7.41072968e-02 7.34808564e-01 4.22502428e-01
2.84598954e-02 3.96150649e-01 9.47557390e-01 -5.15204668e-02
-8.41409385e-01 -9.42150056e-01 -8.51620197e-01 5.09848475e-01
-2.37353086e-01 -3.91231805e-01 -3.06935519e-01 -1.41866952e-01] | [11.952266693115234, 2.0719046592712402] |
05d547f7-d36d-4e36-b304-36d5165e7f7d | semantic-image-matting | 2104.08201 | null | https://arxiv.org/abs/2104.08201v1 | https://arxiv.org/pdf/2104.08201v1.pdf | Semantic Image Matting | Natural image matting separates the foreground from background in fractional occupancy which can be caused by highly transparent objects, complex foreground (e.g., net or tree), and/or objects containing very fine details (e.g., hairs). Although conventional matting formulation can be applied to all of the above cases, no previous work has attempted to reason the underlying causes of matting due to various foreground semantics. We show how to obtain better alpha mattes by incorporating into our framework semantic classification of matting regions. Specifically, we consider and learn 20 classes of matting patterns, and propose to extend the conventional trimap to semantic trimap. The proposed semantic trimap can be obtained automatically through patch structure analysis within trimap regions. Meanwhile, we learn a multi-class discriminator to regularize the alpha prediction at semantic level, and content-sensitive weights to balance different regularization losses. Experiments on multiple benchmarks show that our method outperforms other methods and has achieved the most competitive state-of-the-art performance. Finally, we contribute a large-scale Semantic Image Matting Dataset with careful consideration of data balancing across different semantic classes. | ['Yu-Wing Tai', 'Chi-Keung Tang', 'Yanan sun'] | 2021-04-16 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Sun_Semantic_Image_Matting_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Sun_Semantic_Image_Matting_CVPR_2021_paper.pdf | cvpr-2021-1 | ['transparent-objects', 'semantic-image-matting'] | ['computer-vision', 'computer-vision'] | [ 0.6968312 0.01375507 -0.18538408 -0.37943304 -0.62793714 -0.29222378
0.3324304 -0.24268566 0.23569044 0.507193 0.03004605 -0.04555387
0.09241241 -0.950971 -1.188503 -0.9308327 0.4815702 0.63180226
0.76460123 0.20242216 0.28799412 0.24815746 -1.4619644 0.753747
1.3502524 1.1359446 0.48429415 0.18971047 -0.6107969 0.81798375
-0.4731369 -0.56187737 0.3753022 -0.5185495 -0.5285278 0.66770726
0.80587375 -0.11374372 -0.0670246 1.2132307 -0.114751 -0.20725489
0.80023116 -1.3995192 -0.7090172 0.7140417 -1.1307169 0.04577469
-0.1244491 0.1296919 0.76947165 -0.93937886 0.34342992 1.5507354
0.5225362 0.22482581 -1.4065468 -0.6311528 0.4756227 0.27818736
-1.1351646 -0.15319312 0.93351537 -0.50737923 0.24384755 0.3947833
0.65782636 1.0261393 0.3701747 1.1120191 1.3379747 -0.2341591
0.29269397 0.13668527 0.08913612 0.6838768 0.40391383 -0.5274792
-0.43648922 -0.08207636 0.75897515 0.17094444 -0.23492944 -0.62399006
-1.3170816 0.5729084 0.5296501 -0.06594374 -0.23393758 0.23604238
0.18194257 -0.08844323 0.7265056 0.01723069 -0.3193284 0.34156415
-1.2407321 0.13147236 0.31014398 1.1316453 0.9561797 0.21764113
-0.23248804 1.0375924 0.22969158 0.72756726 0.09058613 -1.153042
0.6131194 0.8225326 0.02081656 -0.9496035 0.07096058 -0.40031934
-0.95776904 0.14692225 0.53417826 0.37736812 -1.4657118 1.4946039
0.44712484 0.5011576 -0.18906015 0.96916366 0.7124503 0.8041854
-0.09433945 -0.13214384 1.5246702 -1.3995235 -0.81198835 -0.51825845
0.15674858 -0.91463697 1.0518626 0.5431572 -1.2237239 -0.40710518
-0.87862074 -0.00671392 -0.26423478 0.18585213 0.7194653 0.6078423
-0.7207278 0.3571196 -0.8281802 -0.27394068 0.64929456 -0.03954743
0.11594165 -0.18978679 -0.8864402 0.6019189 0.41618913 0.19737749
-1.1092767 -0.8830808 -0.7761953 -0.21592663 0.6234583 -0.9453692
0.72313035 -1.6771251 -1.0243267 1.0441396 -0.37886152 -0.370373
0.7078585 -0.25608885 -0.26373392 0.25874817 0.30853957 0.6636912
1.2309419 -1.7932346 -0.5189946 -0.31653446 -0.01692859 0.17362605
-0.15625018 -0.05231814 -0.48759082 -1.05885 0.5141328 -0.81571203
-0.01037943 0.3127954 -0.65336245 0.32896686 0.9449719 -0.8555179
1.0248358 -2.1320784 0.41863778 0.15417498 0.2698629 -0.02351615
0.08064151 -0.0421791 0.10935318 -0.08608457 -0.5917756 -0.24573138
-0.09435052 0.50281155 -0.4521963 0.3497842 0.21941447 0.8426246
-0.68654615 -0.64152634 0.2810333 0.4274225 -0.42673784 0.13695082
-0.5412387 0.26671743 -0.52198607 0.9820971 1.2158343 -0.42595017
-0.04676812 -0.51499796 0.09760107 0.08189222 -1.1538903 1.5183024
-0.22979492 0.27212307 0.41074508 -1.0273727 0.7403931 -0.14868784
0.37978432 -0.405624 0.05374178 0.30016783 -0.1936205 -0.22120109
0.5142781 -0.1342069 0.25508147 0.14084528 -0.33055636 0.06170516
-0.21254647 0.30726272 0.8509824 0.03107496 -0.16859257 -0.5617947
0.3194703 0.04220603 0.81260735 0.59400105 -0.070315 1.0543071
0.34220272 -0.4005837 -1.0275408 -1.4376272 -0.22996674 1.0732627
0.8274497 -0.13071483 -1.0192467 -0.5371989 0.1532676 0.5732417
-0.7948082 -0.18620564 -0.6420083 -1.0223136 0.28749716 0.57694435
0.88084435 -1.0160047 -0.27661932 0.07152899 -0.5339659 -1.2667414
-0.6056118 0.15446176 -1.0528067 -0.89837694 -0.84011364 -0.81943065
0.77075076 0.65696585 1.2921888 0.14706686 -0.30219483 0.12262042
-0.36885697 -0.129119 -0.3402378 -0.07479562 -0.36423957 0.5672803
-0.00971243 -0.5088799 -0.7172257 0.6670326 -1.0361716 0.6326889
0.7482368 0.6762984 0.7534693 0.1770327 0.32117242 -1.1665167
-0.1911108 -0.593138 -0.34600776 0.3805482 -0.46079424 -0.12650432
0.30471358 -0.45201585 -1.2950631 -0.12353744 0.2276043 -0.7147788
-0.14721377 -0.02079939 -0.6177466 -0.10377548 0.17763437 0.4045576
-0.18074435 -0.4340677 0.3495254 0.26695737 0.41091335 -0.8850057
0.9359847 0.895544 -0.13866194 -0.6323045 -1.1387446 -0.11802648
-0.45653895 -0.24446611 1.1306056 -1.043675 0.06742212 0.7029511
-0.9702514 -0.5324353 -0.08205716 -0.12231613 -0.5612262 0.48997524
-0.9141 -0.71507037 -0.22800739 -1.2976837 1.3299439 0.12448999
0.19606853 -0.8168412 -0.5162128 0.9432818 0.4248914 0.38660195
0.9432744 0.02692166 -1.1736826 0.58160174 -0.5978757 0.3483907
0.21273884 0.18502781 -1.0869769 -0.13308795 -0.11299446 -0.09636932
1.408137 0.6678005 1.5603255 -0.3331139 -0.45997146 0.85052305
1.3071883 0.13388017 1.0591553 0.39880994 1.2028279 0.6295862
0.82711476 0.16275038 0.2795178 0.68820626 0.58478826 -0.34820518
-0.37925 -0.10980367 0.3873979 0.5961124 0.06090931 -0.2296463
-0.51981026 0.4999606 -1.9991889 -0.7306098 -0.422099 1.9256529
0.76940733 0.3383212 0.05862227 0.08674155 1.0325747 0.24310638
-0.6310224 -0.07411785 -0.52355117 -0.0719436 0.6588584 0.42528853
-1.1711162 1.1317238 5.816157 1.3004714 -0.8656324 0.22106867
1.0093923 0.03727866 -0.7458322 0.09147123 -0.628399 0.8610269
0.2451762 0.25759202 0.4279157 0.79915196 0.0607357 -0.1786218
-0.7577606 0.92384624 0.22459508 -1.222098 0.4790719 -0.03638799
0.962988 -0.4048262 0.06812452 -0.00999586 0.03146676 -0.9146242
1.1879237 0.39162365 0.5086245 -0.51439154 0.4330213 0.09071177
-1.368547 0.11091001 -0.54294056 0.1150843 0.08272459 1.1185404
-0.28231207 0.53068036 0.7567849 0.7785799 -0.7240762 0.94598716
0.12144852 0.79509324 -0.15008666 0.40217695 0.12334202 -0.45598295
0.543608 1.1216009 0.19794574 -0.18557301 0.36782926 1.306091
0.08632728 0.0058924 -0.31308103 0.24253927 0.31148222 1.3063453
-1.2833645 -0.58785933 -0.29617754 1.2659428 0.12727992 0.42932904
-1.1287714 0.22388782 0.7134169 0.64481723 0.43459207 0.04312964
-0.74760365 -1.4013664 0.2730401 -0.76315445 0.3883067 -0.9416596
-1.4460561 0.27452406 0.10337146 -1.0259063 0.65679055 -0.56474876
-0.74140644 0.5281776 -1.3663584 -1.4900732 -0.5994265 0.49171883
0.84606576 0.05601912 0.13360965 0.5673437 -0.6269876 0.47480422
-0.03691174 -0.13093878 0.71789813 -1.3190353 0.14921421 0.95311236
-0.0813156 0.29837212 0.7015639 -0.7885727 -1.2961555 -1.4773208
0.23314773 -0.46969593 0.5582889 -0.61629295 -1.2109897 0.6991939
0.19159609 -0.08547918 0.169938 -0.26556128 -0.53778726 -0.26303598
-1.1742753 0.5123303 1.1431333 -0.01945745 -0.3683295 0.2977369
0.8675944 -0.5356711 -0.50882524 0.5732959 0.3296821 -1.0941786
1.1486354 -0.20190613 0.7010338 -0.739799 -0.29124737 -1.003036
-0.4463404 -0.20423451 -0.08664041 1.4874 0.16457334 -0.52840453
0.90634084 0.3004979 -0.5041921 -0.720621 -0.8643298 -0.674938
0.01525678 -0.25447482 0.46658385 1.0496087 -0.74926364 0.02154997
-0.68189716 0.27424583 1.1250848 0.61923516 0.69704455 -0.9527643
-0.3660516 -0.53323853 -0.34133813 -1.0045915 0.09948307 -0.86791146
0.07609392 -1.5419635 0.73485214 -0.66162586 -0.14939016 0.42727786
-0.54160476 0.5731189 0.04206504 0.16309093 -0.75992644 0.7553579
1.5872439 -0.50670284 0.27541414 -0.22885194 -0.731618 0.79542595
0.6726755 -0.35546237 -0.37576538 -0.5108798 0.1097687 -0.28142938
0.7258794 -0.91822094 -0.27224442 -0.602052 0.47368047 -0.6632937
0.40110907 -0.86121595 0.41087803 0.12862682 -0.05916939 -0.34128582
0.11882524 0.7757186 -0.0840918 0.0352571 1.1398383 -0.16398129
-0.549671 0.15808435 -0.17120637 0.29339764 0.9499032 -0.48165688
-0.46252486 -0.04376771 -0.50798565 0.17211166 0.82251257 0.40689772
0.5734027 -1.401946 -0.5568226 0.14123537 0.05417365 0.32873482
0.30808315 0.90542185 -0.66327786 -0.23860449 -0.14937146 -0.9060694
-1.1611594 0.45177415 0.30996257 0.02442979 -0.9344112 0.7531913
1.1126169 -0.09932277 0.21821998 -0.525355 0.37536004 -0.13899241
0.44379118 0.30429626 -0.05116731 -0.6198655 -0.31324285 0.7110482
-0.14402077 0.37238553 0.97527105 -0.20385294 -0.34514156 0.5250152
0.7030092 -0.17750429 -1.4194319 -0.18706353 -0.24332817 -0.88611484
-0.13449971 -0.6364831 -1.539861 0.80371577 0.503047 0.08434393
1.188317 0.22402348 1.1090641 -0.23513328 0.5112168 -1.0209188
0.38988128 0.14098312 0.8117436 -1.2303714 0.06462733 -1.0200359
-0.7120425 0.7719489 0.870621 -0.3068908 0.42410377 0.34187645
0.13615048 -0.18473004 -0.5494268 -0.00719434 0.37418795 0.40570837
0.1640788 0.3247399 -0.15041599 0.43627936 0.02460501 -0.47761005
0.42433655 0.7997262 -0.7025877 -0.84963924 -0.692615 0.8174437
-0.43384278 -0.0897618 -0.32137832 0.4097574 0.40085834 0.6988341
0.05851204 -0.21746048 0.0882343 -0.0509721 0.6237425 -0.44122568
-0.43101147 0.46195024 -0.21539348 -0.56229943 -0.514839 -0.6291997
-1.0925219 -0.32675198 -0.3582177 -0.23313124 0.21755283 0.92187333
0.0924444 0.71836287 0.31726232 -0.85156107 0.08461349 -0.79805005
-0.81948227 0.7027993 0.1458322 -1.0647529 -0.5423965 0.3494201 ] | [10.623125076293945, -0.9035617709159851] |
8597edbe-e5da-4765-8ecf-cf96fe0a6f12 | community-detection-with-a-subsampled | 2102.01419 | null | https://arxiv.org/abs/2102.01419v3 | https://arxiv.org/pdf/2102.01419v3.pdf | Community Detection with a Subsampled Semidefinite Program | Semidefinite programming is an important tool to tackle several problems in data science and signal processing, including clustering and community detection. However, semidefinite programs are often slow in practice, so speed up techniques such as sketching are often considered. In the context of community detection in the stochastic block model, Mixon and Xie \cite{mixon2020sketching} have recently proposed a sketching framework in which a semidefinite program is solved only on a subsampled subgraph of the network, giving rise to significant computational savings. In this short paper, we provide a positive answer to a conjecture of Mixon and Xie about the statistical limits of this technique for the stochastic block model with two balanced communities. | ['Afonso S. Bandeira', 'Pedro Abdalla'] | 2021-02-02 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 1.65028542e-01 2.19561830e-01 -4.29734379e-01 2.61536296e-02
-4.22227710e-01 -7.13281751e-01 7.53380656e-02 2.45755792e-01
-2.23503605e-01 6.64922833e-01 2.30144188e-02 -5.40993512e-01
-5.69172263e-01 -7.84549177e-01 -5.12667060e-01 -6.57729268e-01
-4.80098754e-01 6.29917145e-01 -1.86560620e-02 -3.26495245e-02
-4.46674898e-02 1.04916751e-01 -6.79576039e-01 1.63079286e-03
6.44806921e-01 4.34686422e-01 -9.93091688e-02 6.89112365e-01
1.03729749e-02 3.12743962e-01 -1.21893555e-01 -5.18774390e-01
4.71023470e-01 -1.99190974e-01 -3.97575140e-01 4.85605448e-01
3.03349614e-01 2.05723792e-02 -7.66704798e-01 1.31131101e+00
3.99405509e-01 -3.60757172e-01 3.79671097e-01 -1.76741993e+00
-3.02804559e-01 1.18215132e+00 -1.23600090e+00 3.04114874e-02
2.89610505e-01 -1.35359213e-01 1.41411746e+00 -9.37708020e-01
7.49274850e-01 1.18754065e+00 7.26037443e-01 1.98567972e-01
-1.85389507e+00 -1.00389755e+00 1.97424889e-01 -6.17715865e-02
-1.60580003e+00 -1.12675101e-01 8.29737663e-01 -5.75600743e-01
8.82475302e-02 4.91777062e-01 8.26266944e-01 6.71328068e-01
-4.75540251e-01 1.16097689e+00 1.13562512e+00 -1.86246112e-01
3.42225254e-01 7.63739645e-02 4.54335630e-01 8.43215704e-01
1.01130664e+00 -1.01542011e-01 -5.42238235e-01 -8.15994263e-01
7.64836967e-01 3.44642252e-02 -2.18103856e-01 -9.30398226e-01
-1.17070389e+00 1.21793723e+00 2.33444080e-01 -3.55599113e-02
-1.30792215e-01 3.03894192e-01 5.85417747e-01 3.19189638e-01
5.80258727e-01 4.95601334e-02 2.57917762e-01 2.22053394e-01
-1.23828149e+00 3.38593036e-01 1.20501959e+00 1.04434228e+00
4.49099183e-01 -1.56457350e-01 8.24926496e-02 4.67103094e-01
2.14419097e-01 6.47790074e-01 -8.70751739e-01 -7.75430560e-01
9.20786977e-01 5.21227896e-01 1.10603668e-01 -1.46949482e+00
-4.14271444e-01 -5.03720105e-01 -1.47610712e+00 -1.47725552e-01
6.02849126e-01 -3.14159423e-01 -2.56788194e-01 1.55937648e+00
1.94251522e-01 7.08212107e-02 -4.73872244e-01 8.77663136e-01
2.58263886e-01 6.04954660e-01 -3.88080537e-01 -5.01170456e-01
8.52644682e-01 -7.46777475e-01 -5.87691069e-01 -1.50680542e-01
4.67442632e-01 -3.48012120e-01 4.25615042e-01 5.73721290e-01
-8.02168012e-01 3.05904388e-01 -9.62841213e-01 3.32145035e-01
2.28558496e-01 -3.12921628e-02 1.05741882e+00 9.14686561e-01
-1.08466387e+00 3.69398057e-01 -4.62934285e-01 -4.31728244e-01
5.56811392e-01 4.87736821e-01 -3.46349597e-01 -3.75274956e-01
-8.47229302e-01 1.11361496e-01 2.37982094e-01 3.32250386e-01
-6.00174963e-01 -5.05971432e-01 -6.19189918e-01 1.02916777e-01
7.69361198e-01 -5.16676426e-01 3.61011207e-01 -7.52584815e-01
-7.87246108e-01 9.35089707e-01 -4.44511697e-02 -7.11621940e-01
8.44538808e-01 3.02129596e-01 5.02790399e-02 1.39083937e-01
1.38953641e-01 5.00577539e-02 6.51596606e-01 -9.31328297e-01
-1.54157594e-01 -1.88126102e-01 1.88174441e-01 -1.49360403e-01
-2.81098932e-01 1.64312243e-01 -4.38219130e-01 -6.12104237e-01
1.99985623e-01 -1.21014774e+00 -7.11797893e-01 5.15314043e-01
-7.40242541e-01 -1.13950767e-01 4.70901191e-01 -4.82693791e-01
1.48415780e+00 -2.16775107e+00 3.99544805e-01 8.82015109e-01
8.05222154e-01 3.00781410e-02 -1.32864147e-01 6.87626302e-01
-1.26328766e-01 3.58476847e-01 -4.91113991e-01 -2.70746917e-01
1.25533924e-01 1.39958367e-01 -1.00085638e-01 1.01506293e+00
-1.84167311e-01 4.71253067e-01 -1.03637671e+00 -3.83607864e-01
-2.26098225e-01 -1.02655821e-01 -7.07779527e-01 -3.84732246e-01
4.31836059e-04 -7.65648559e-02 -4.33723450e-01 4.66765851e-01
9.43970263e-01 -6.39143467e-01 6.94269598e-01 1.22842297e-01
-1.53925583e-01 -2.04327881e-01 -1.87677205e+00 1.17696142e+00
-2.61098314e-02 9.35739338e-01 7.76363790e-01 -1.18765450e+00
6.86917067e-01 3.41650456e-01 6.15142643e-01 -4.28182585e-03
-1.11953929e-01 1.55383036e-01 1.33954976e-02 2.56488211e-02
8.67789239e-02 -1.22529209e-01 -9.54903290e-02 9.03806806e-01
-4.84270632e-01 1.66188776e-02 6.75190091e-01 8.30735505e-01
1.41245270e+00 -8.34931970e-01 1.43300548e-01 -7.76030481e-01
2.93974310e-01 -2.72718742e-02 3.92153740e-01 1.09365356e+00
-1.46418706e-01 5.82329571e-01 1.16173434e+00 -2.27789789e-01
-1.37310278e+00 -1.01786447e+00 1.85351931e-02 4.45561320e-01
8.64447802e-02 -5.91472507e-01 -5.96553206e-01 -3.11896265e-01
2.77708501e-01 -1.48929089e-01 -5.77796042e-01 5.49894571e-02
-2.91668415e-01 -8.84030282e-01 5.28328955e-01 2.41113931e-01
2.01162875e-01 -1.54466122e-01 7.09274337e-02 3.28716516e-01
-2.67127931e-01 -1.03042877e+00 -9.80858624e-01 3.32772844e-02
-1.01083040e+00 -1.35991025e+00 -1.02098382e+00 -5.92945576e-01
9.12979245e-01 6.31596923e-01 1.02430987e+00 1.83857486e-01
-4.46472883e-01 2.64360964e-01 7.23519102e-02 -1.36368303e-02
-1.70300439e-01 -3.98006570e-03 3.01691800e-01 3.32822919e-01
-3.48276868e-02 -4.77481812e-01 -4.75430042e-01 3.01384538e-01
-7.19580531e-01 1.34288535e-01 3.19149315e-01 9.59453762e-01
4.01320517e-01 1.00760005e-01 2.12307125e-01 -1.12059939e+00
7.80677617e-01 -6.83084309e-01 -9.62685645e-01 2.92833179e-01
-4.97698039e-01 -5.93765937e-02 4.50951338e-01 -3.44268322e-01
-3.45861465e-01 2.91314870e-01 4.95182693e-01 -4.23584074e-01
9.20510948e-01 7.92533278e-01 3.45423352e-03 -4.10693467e-01
4.07598764e-01 2.87958235e-01 1.15315884e-01 -3.75023782e-01
2.51325011e-01 5.49942374e-01 -1.98375121e-01 -5.34884691e-01
1.07989812e+00 9.88613188e-01 4.34805989e-01 -1.10918570e+00
-6.96593404e-01 -8.03660154e-01 -4.48552758e-01 -2.15103507e-01
2.98584074e-01 -1.00320601e+00 -7.80059397e-01 2.38163203e-01
-1.06295240e+00 -1.00102611e-01 2.54065171e-02 2.58789480e-01
-4.90919143e-01 6.38989508e-01 -5.51124752e-01 -1.13803625e+00
-1.11513890e-01 -4.85858411e-01 6.72334373e-01 -1.84383065e-01
-6.10775799e-02 -1.06160498e+00 2.04331622e-01 3.09700638e-01
-1.20354062e-02 4.58406001e-01 8.27752888e-01 -2.97497302e-01
-6.79845929e-01 -3.30531180e-01 -7.00773180e-01 7.29751820e-03
-2.18730405e-01 3.22458953e-01 -3.39622617e-01 -7.16151357e-01
-4.53567296e-01 -3.36305797e-02 7.50700533e-01 4.35536295e-01
9.82463300e-01 -5.49374342e-01 -5.03700078e-01 4.72156644e-01
1.47319829e+00 -4.04481739e-01 3.12443674e-01 -4.84767593e-02
6.34222746e-01 5.14498055e-01 3.24586213e-01 7.19216704e-01
1.73903674e-01 4.65916693e-01 1.66666105e-01 -1.97621688e-01
3.44577879e-01 -3.65554899e-01 2.38169298e-01 7.70238876e-01
1.58470824e-01 -1.92067459e-01 -8.70195270e-01 6.88763082e-01
-1.99603391e+00 -1.12014890e+00 -7.29162931e-01 2.29403591e+00
8.20263982e-01 1.75758436e-01 7.53962278e-01 3.13339025e-01
1.21755636e+00 1.69496387e-01 -3.15016866e-01 -2.74653882e-01
-5.63467085e-01 -5.54210506e-02 9.76475716e-01 5.03248513e-01
-1.02279103e+00 3.49222600e-01 6.64360237e+00 9.25258517e-01
-7.39160419e-01 -8.19592923e-03 5.84678829e-01 -1.63540646e-01
-4.07981455e-01 2.89596468e-01 -4.77860272e-01 4.53641206e-01
4.48886991e-01 -7.92504609e-01 4.84271139e-01 6.38595402e-01
4.57404047e-01 -3.59050930e-01 -1.04952693e+00 1.06694603e+00
-2.02017337e-01 -1.55705440e+00 -3.80955279e-01 5.62143862e-01
9.91720617e-01 -1.40535653e-01 1.68845206e-02 -1.88026279e-01
4.42635357e-01 -8.89587402e-01 4.83103484e-01 6.65099397e-02
7.48312354e-01 -7.58338869e-01 2.82989413e-01 3.67523342e-01
-1.20661938e+00 4.97725308e-02 -3.51521194e-01 -1.57569319e-01
3.09489369e-01 1.24325454e+00 -6.35418713e-01 2.54700929e-01
3.05638880e-01 7.34804094e-01 -2.52067775e-01 1.54782808e+00
9.69430879e-02 9.79785502e-01 -9.47547019e-01 -2.28845865e-01
2.81234324e-01 -7.66635478e-01 1.12802958e+00 1.09201455e+00
-5.60333170e-02 3.63950022e-02 2.04187676e-01 1.01892424e+00
-1.31096646e-01 2.88820658e-02 -7.84076512e-01 -5.31949103e-01
2.76692063e-01 9.39005136e-01 -1.13598287e+00 -1.43138049e-02
-3.56164157e-01 6.32968187e-01 2.26304814e-01 2.98315614e-01
-5.31463742e-01 -2.82743335e-01 4.29406196e-01 4.89952356e-01
3.10762048e-01 -6.24385417e-01 -6.22710288e-01 -1.17669785e+00
1.81112736e-01 -8.04145277e-01 3.87939543e-01 -7.37282783e-02
-1.18276119e+00 -1.02007650e-01 -6.59720004e-02 -9.28066134e-01
3.66956949e-01 -3.68110329e-01 -7.24612176e-01 5.78568339e-01
-7.37450659e-01 -7.10199058e-01 2.41447404e-01 4.64500725e-01
-6.35537319e-03 3.22594345e-01 2.65645444e-01 6.32020473e-01
-7.97813892e-01 3.89952153e-01 7.25315452e-01 1.83684185e-01
2.04077974e-01 -1.32770431e+00 2.75226206e-01 9.40163851e-01
3.47888142e-01 6.33116126e-01 1.05736685e+00 -6.02937043e-01
-1.70012105e+00 -8.67385626e-01 8.45810115e-01 1.02505825e-01
1.19537711e+00 -8.72729242e-01 -5.39584816e-01 2.31329471e-01
-6.20599985e-02 1.25469416e-01 5.74269831e-01 2.55108178e-01
-3.57536584e-01 -7.37775415e-02 -1.00722599e+00 7.38323331e-01
1.20585203e+00 -5.75584650e-01 2.10817695e-01 7.10257173e-01
1.54758215e-01 1.48508281e-01 -5.21964669e-01 -1.36789456e-01
4.67150927e-01 -7.11318672e-01 8.50595236e-01 -3.82811159e-01
1.25097290e-01 -3.24307442e-01 1.12714790e-01 -1.06705070e+00
-2.04536334e-01 -1.21376562e+00 -1.37124598e-01 1.06624079e+00
4.02502716e-01 -4.27651316e-01 1.21091354e+00 3.45801830e-01
8.29054296e-01 -5.94465554e-01 -1.15188229e+00 -1.07209599e+00
-6.10045530e-02 -1.66219160e-01 -5.29888384e-02 8.89004350e-01
2.87509680e-01 2.20169827e-01 -7.60217607e-01 1.25660105e-02
1.42402852e+00 1.46760285e-01 7.03324914e-01 -1.36769021e+00
-3.89652371e-01 -2.83809543e-01 -2.75958419e-01 -1.02472365e+00
7.37396106e-02 -9.72466707e-01 -2.52226174e-01 -1.32898843e+00
7.77429760e-01 -7.50003219e-01 2.52366960e-01 -3.21151316e-02
1.91819549e-01 3.37147266e-01 3.07945728e-01 1.15328118e-01
-7.41641521e-01 2.08984777e-01 1.06777132e+00 -5.04241824e-01
5.81501946e-02 5.26989400e-01 -7.35147774e-01 4.33161110e-01
4.68866527e-01 -7.27248013e-01 -1.95906162e-01 -1.89671725e-01
9.73764420e-01 4.01795685e-01 4.26785320e-01 -5.71586013e-01
6.12700820e-01 -1.67652890e-01 -4.70774114e-01 -6.62369490e-01
5.30328639e-02 -8.02101612e-01 4.69683707e-01 7.29995251e-01
-2.50061750e-01 -7.78810382e-02 -6.48793057e-02 1.26061082e+00
-1.16622455e-01 -2.84982890e-01 6.73761070e-01 1.55058548e-01
-1.18712068e-01 5.19638062e-01 -5.98968923e-01 3.48572791e-01
1.11232126e+00 -2.22462744e-01 -2.37796027e-02 -7.87138224e-01
-8.76014471e-01 7.45236754e-01 2.79561818e-01 -2.80910712e-02
5.01273215e-01 -1.16926599e+00 -1.00155473e+00 -2.06088632e-01
2.98296325e-02 -1.99802116e-01 4.65694666e-02 1.29029191e+00
-6.48602307e-01 5.46577632e-01 2.72606999e-01 -6.09459281e-01
-1.52270854e+00 5.22251964e-01 1.55421540e-01 -4.96882737e-01
-3.54072422e-01 9.48160589e-01 2.25854233e-01 -6.07790388e-02
2.76983857e-01 7.40985721e-02 4.28590953e-01 3.04766715e-01
2.33008459e-01 7.89099216e-01 -4.05932158e-01 -2.72925168e-01
-4.61311966e-01 2.27518141e-01 2.15422735e-02 -1.25658795e-01
1.38954127e+00 -3.34590524e-01 -5.15185297e-01 3.65767866e-01
1.15231526e+00 1.97848111e-01 -8.81938100e-01 -4.22671437e-01
3.09989929e-01 -4.49098140e-01 -7.82840848e-02 -6.60262629e-02
-1.14798737e+00 6.65609539e-01 2.08074018e-01 5.91934741e-01
7.04881430e-01 1.35854244e-01 4.73736346e-01 3.61277640e-01
6.24415159e-01 -1.00937903e+00 -2.56114364e-01 4.22234803e-01
8.03632975e-01 -1.08896661e+00 3.92785788e-01 -9.51960742e-01
-2.74244905e-01 1.08882403e+00 -2.33883760e-03 -5.13575017e-01
1.04348397e+00 2.38355160e-01 -8.03291917e-01 -2.52240330e-01
-7.18751490e-01 -1.11772314e-01 1.46264397e-02 4.99328315e-01
2.03976296e-02 3.92857641e-01 -5.97621560e-01 4.61471200e-01
2.36141697e-01 3.41434218e-03 8.50486100e-01 6.05920374e-01
-1.71929285e-01 -1.03157711e+00 -4.16633070e-01 7.23761439e-01
-3.13550293e-01 -3.15576762e-01 -7.42867827e-01 4.65867430e-01
-3.18052590e-01 9.26492214e-01 -1.52018443e-01 -2.14066021e-02
1.44325510e-01 -6.22541785e-01 3.87226462e-01 -7.59638250e-01
-3.49531263e-01 -5.23027778e-02 2.35389888e-01 -1.85596168e-01
-3.21687818e-01 -1.00026429e+00 -4.13476557e-01 -7.14814126e-01
-6.21741652e-01 3.43571872e-01 3.05208236e-01 7.51952827e-01
1.17063962e-01 -6.96542710e-02 9.31749821e-01 -3.02009076e-01
-6.97043777e-01 -3.93271327e-01 -1.20277524e+00 -1.31019682e-01
2.32456073e-01 -4.25521135e-01 -7.48939276e-01 -4.53132123e-01] | [6.913620948791504, 5.111055850982666] |
fd63bbe8-a570-4040-996c-9014164d4b12 | jointly-optimizing-image-compression-with-low | 2305.15030 | null | https://arxiv.org/abs/2305.15030v1 | https://arxiv.org/pdf/2305.15030v1.pdf | Jointly Optimizing Image Compression with Low-light Image Enhancement | Learning-based image compression methods have made great progress. Most of them are designed for generic natural images. In fact, low-light images frequently occur due to unavoidable environmental influences or technical limitations, such as insufficient lighting or limited exposure time. %When general-purpose image compression algorithms compress low-light images, useful detail information is lost, resulting in a dramatic decrease in image enhancement. Once low-light images are compressed by existing general image compression approaches, useful information(e.g., texture details) would be lost resulting in a dramatic performance decrease in low-light image enhancement. To simultaneously achieve a higher compression rate and better enhancement performance for low-light images, we propose a novel image compression framework with joint optimization of low-light image enhancement. We design an end-to-end trainable two-branch architecture with lower computational cost, which includes the main enhancement branch and the signal-to-noise ratio~(SNR) aware branch. Experimental results show that our proposed joint optimization framework achieves a significant improvement over existing ``Compress before Enhance" or ``Enhance before Compress" sequential solutions for low-light images. Source codes are included in the supplementary material. | ['Sheng Zhong', 'Luxin Yan', 'Liqun Chen', 'Xu Zou', 'Shilv Cai'] | 2023-05-24 | null | null | null | null | ['image-enhancement', 'low-light-image-enhancement'] | ['computer-vision', 'computer-vision'] | [ 8.70332599e-01 -5.76162696e-01 -1.68477912e-02 -4.52010602e-01
-8.15923989e-01 -6.41950145e-02 8.63188133e-03 -3.39566544e-02
-5.17242014e-01 5.94161332e-01 -3.78595777e-02 -1.44212946e-01
9.42410603e-02 -9.21473086e-01 -8.91518891e-01 -1.00616562e+00
4.37384695e-01 -4.44752693e-01 1.42771527e-01 -2.13293046e-01
1.83962762e-01 1.08879343e-01 -1.75043488e+00 3.40934545e-01
9.24020588e-01 1.04172766e+00 7.61027575e-01 7.89644361e-01
-6.33226335e-02 8.69160950e-01 -3.32030684e-01 -3.83963615e-01
4.73177612e-01 -3.99044603e-01 -2.30911165e-01 2.29651764e-01
6.19717836e-01 -8.63500953e-01 -7.40178704e-01 1.49388254e+00
8.01174164e-01 1.69811279e-01 1.95411041e-01 -8.14354956e-01
-5.63061118e-01 2.28291571e-01 -9.97826338e-01 2.33487770e-01
1.11698911e-01 7.09359586e-01 7.70537317e-01 -7.22716391e-01
4.28687155e-01 9.49607074e-01 5.26283383e-01 3.27413231e-01
-1.32168067e+00 -7.26840317e-01 -1.35248944e-01 4.81188774e-01
-1.10324585e+00 -6.78827703e-01 9.72255468e-01 1.78083614e-01
7.85345078e-01 3.73100072e-01 5.64723194e-01 4.48611438e-01
3.77779275e-01 7.14935422e-01 1.41789865e+00 -3.54007870e-01
1.67951956e-02 3.45403329e-02 -4.74620372e-01 6.46861732e-01
2.28848070e-01 3.36931020e-01 -6.00589216e-01 4.76185530e-01
6.75367713e-01 1.87213600e-01 -5.68130851e-01 2.26515993e-01
-9.60892558e-01 4.21417177e-01 7.01074600e-01 3.18069868e-02
-3.15250933e-01 3.56028914e-01 2.82395810e-01 2.43449822e-01
4.24083382e-01 4.41880375e-02 -1.29987419e-01 -6.43286854e-02
-9.76279914e-01 7.39933029e-02 2.68471897e-01 8.18859279e-01
8.38207006e-01 1.10808678e-01 -1.38426617e-01 1.15030932e+00
1.52306721e-01 7.32123315e-01 3.59822124e-01 -1.01214135e+00
6.47419512e-01 1.98414862e-01 1.03662804e-01 -9.88100767e-01
-2.32146293e-01 -7.05776572e-01 -1.35346293e+00 4.69416618e-01
5.07156029e-02 -5.35825565e-02 -6.32118344e-01 1.75508618e+00
1.31018668e-01 1.59951434e-01 8.05449039e-02 1.14466989e+00
9.47896302e-01 9.65585947e-01 -1.02412082e-01 -7.41955221e-01
1.26408708e+00 -8.61174524e-01 -1.08066952e+00 -2.83732921e-01
4.89885323e-02 -1.16857111e+00 1.25493097e+00 3.96978974e-01
-1.64050281e+00 -8.80631089e-01 -1.25845742e+00 -4.45865363e-01
1.03160590e-01 1.79940701e-01 6.55538917e-01 5.39061785e-01
-8.96775484e-01 5.61006546e-01 -5.45263946e-01 2.70096153e-01
4.83502954e-01 2.36405894e-01 -4.29372638e-02 -5.61683893e-01
-9.45157349e-01 3.71921003e-01 1.14564210e-01 1.02712393e-01
-8.92275512e-01 -6.35066152e-01 -7.65354931e-01 6.29164129e-02
4.17079031e-01 -6.13781095e-01 1.06889546e+00 -5.62222183e-01
-1.41702294e+00 6.42401993e-01 -3.37571323e-01 -2.46150985e-01
4.60490614e-01 -1.89701065e-01 -2.43432060e-01 3.38202536e-01
-2.08881661e-01 6.55070245e-01 9.58195984e-01 -1.41421783e+00
-6.83097839e-01 -1.15273252e-01 -8.07406679e-02 4.22139794e-01
-4.00897652e-01 -8.22304115e-02 -6.90502942e-01 -8.05063486e-01
1.65529922e-01 -7.16106534e-01 -2.63199210e-01 3.49164188e-01
-9.31167156e-02 2.79297888e-01 9.63699281e-01 -7.25725949e-01
1.27443314e+00 -2.24231625e+00 -2.44025990e-01 -2.58730263e-01
2.59523004e-01 4.09684449e-01 -2.68712938e-01 4.00394201e-02
-1.26290265e-02 -1.85663059e-01 -4.49694574e-01 -3.14058930e-01
-2.78189778e-01 9.35708731e-02 -8.85046721e-02 4.28394943e-01
3.20107415e-02 6.26358449e-01 -9.05629396e-01 -6.27043486e-01
5.22868395e-01 7.35200047e-01 -6.49657249e-01 3.90812188e-01
-1.01911947e-01 3.40775043e-01 -2.51893345e-02 7.26911902e-01
1.02759862e+00 -1.39272243e-01 -3.21640708e-02 -6.91304028e-01
-1.95302695e-01 -3.98790650e-02 -1.02080500e+00 1.66648173e+00
-8.61024082e-01 8.88961792e-01 3.34630281e-01 -7.25601017e-01
7.59710431e-01 1.57650918e-01 4.83797908e-01 -1.02182245e+00
2.89164007e-01 1.70260772e-01 -2.34673128e-01 -6.70179725e-01
4.37052101e-01 -3.60644907e-01 5.67466855e-01 2.58787662e-01
-3.93184841e-01 -5.07021308e-01 2.79460818e-01 1.43801728e-02
9.13034201e-01 7.13016614e-02 1.08195886e-01 2.23842770e-01
5.33219039e-01 -5.07383227e-01 5.33397377e-01 4.34615195e-01
-2.88888991e-01 6.39315665e-01 -1.32483542e-01 -2.36755788e-01
-1.14243257e+00 -9.31556165e-01 -3.99154842e-01 9.03115213e-01
5.62890053e-01 -3.71553957e-01 -7.12403476e-01 -1.08633123e-01
-4.67100292e-01 5.57413459e-01 1.94428966e-03 -3.09358925e-01
-6.30193174e-01 -1.01261389e+00 -4.05717604e-02 1.75676003e-01
1.25932872e+00 -8.62327039e-01 -7.16887176e-01 1.65739879e-01
-7.56235003e-01 -1.28300071e+00 -5.23619950e-01 2.75770158e-01
-9.14553940e-01 -7.76303291e-01 -7.22590268e-01 -8.28274429e-01
6.30971253e-01 8.72926414e-01 9.14301097e-01 1.89840257e-01
-5.39505363e-01 -6.65318444e-02 -9.57069546e-02 -4.04868990e-01
-3.54389966e-01 -4.77616936e-01 -2.62030184e-01 5.97829819e-02
7.74962902e-02 -7.66099572e-01 -1.22349143e+00 2.72734314e-01
-1.21050680e+00 3.63785475e-01 1.08784997e+00 1.01246321e+00
1.05346918e+00 8.27058494e-01 6.06273413e-02 -5.08936048e-01
5.14179349e-01 1.22934297e-01 -5.45763314e-01 1.00278899e-01
-8.23992014e-01 -5.70483766e-02 9.70708311e-01 -2.56078422e-01
-1.32911003e+00 -1.00877441e-01 -4.00613844e-01 -4.70981777e-01
-1.39842510e-01 3.82254988e-01 -2.27188170e-01 -4.49064940e-01
5.23903191e-01 7.39474237e-01 -1.29816324e-01 -3.51829082e-01
1.32358715e-01 6.83035791e-01 7.12799370e-01 -3.38200897e-01
8.70319247e-01 4.17753100e-01 4.50003445e-01 -9.16436076e-01
-7.08318055e-01 -4.21800315e-01 -5.78299984e-02 -4.45065379e-01
5.74060440e-01 -1.17109144e+00 -7.52647638e-01 4.51753110e-01
-8.55769932e-01 -3.16311330e-01 -3.19796681e-01 5.95238090e-01
-5.84079862e-01 6.05715156e-01 -8.45759213e-01 -5.95542073e-01
-5.13315022e-01 -1.53971410e+00 9.03659940e-01 3.62049699e-01
4.48743194e-01 -5.48501611e-01 -3.53509277e-01 5.34178793e-01
6.37131691e-01 1.57547612e-02 9.78673756e-01 6.32927001e-01
-8.49154830e-01 -1.35626331e-01 -7.91533709e-01 6.29069328e-01
2.42727190e-01 -3.01847517e-01 -9.94390666e-01 -5.93503058e-01
2.70948589e-01 -3.99565876e-01 9.99629736e-01 6.16851270e-01
1.53525329e+00 -3.23942065e-01 -2.11277381e-02 1.13201416e+00
1.85937321e+00 2.72906780e-01 8.22944105e-01 6.22702613e-02
5.64095140e-01 3.67065340e-01 5.96673846e-01 5.91479123e-01
1.09967932e-01 7.52633691e-01 5.99009633e-01 -4.10908580e-01
-7.62559891e-01 -5.43627329e-02 2.59042770e-01 9.56287622e-01
-2.21125156e-01 -1.94764465e-01 -3.09927642e-01 3.64425510e-01
-1.32330120e+00 -1.07869589e+00 -5.87842502e-02 2.28910756e+00
1.27842176e+00 1.12461902e-01 -3.27848434e-01 3.58751714e-01
5.16408920e-01 3.26392859e-01 -7.37440646e-01 -1.35232031e-01
-2.47381374e-01 1.77734405e-01 6.72202826e-01 4.93824691e-01
-8.91310871e-01 4.63729709e-01 5.58207607e+00 1.04629481e+00
-1.12527978e+00 1.07515357e-01 8.46435428e-01 -4.33199316e-01
-2.81327307e-01 -6.51100948e-02 -5.37136495e-01 6.03638530e-01
5.83591461e-01 -6.12584278e-02 7.14699090e-01 5.69570601e-01
5.22035897e-01 -3.34655017e-01 -6.85043335e-01 1.54021347e+00
2.47193929e-02 -1.04107285e+00 1.24908589e-01 -5.64032607e-02
7.59799302e-01 -2.24561140e-01 3.07605207e-01 -1.35844052e-01
-2.24256575e-01 -6.44401908e-01 3.37793738e-01 2.68677205e-01
1.35597181e+00 -8.56594980e-01 5.06700099e-01 2.28417918e-01
-1.22881401e+00 -9.41826031e-02 -7.36973166e-01 2.76483506e-01
2.79659450e-01 1.01042473e+00 -1.96480855e-01 3.83186072e-01
8.19390416e-01 7.58310378e-01 -3.17798585e-01 1.10144556e+00
-2.78292567e-01 5.90188920e-01 -1.84476063e-01 2.24878594e-01
7.46567473e-02 -3.27748477e-01 6.80839717e-01 1.28180683e+00
2.67928898e-01 5.60329676e-01 8.10806751e-02 5.60550094e-01
-3.55308712e-01 1.57615125e-01 -4.09346461e-01 2.25812852e-01
-4.92841750e-02 1.26868927e+00 -6.84960783e-01 -2.58371294e-01
-6.38208210e-01 1.06797159e+00 -2.96448231e-01 3.38786960e-01
-8.41470063e-01 -7.72554159e-01 4.03431952e-01 2.87177682e-01
-3.35526876e-02 -1.07119195e-01 -3.69101852e-01 -1.16116583e+00
1.37410864e-01 -9.92806017e-01 1.66300461e-02 -1.05338573e+00
-8.07517231e-01 5.01771808e-01 -2.68884063e-01 -1.47826838e+00
3.95788372e-01 -4.55143750e-01 -3.93370837e-01 7.81105876e-01
-2.03460670e+00 -9.20618057e-01 -7.77709901e-01 8.35622132e-01
9.12344754e-01 -2.38621198e-02 2.50415921e-01 6.70453131e-01
-4.67863947e-01 6.49915934e-01 2.86623061e-01 -2.50501037e-01
8.62642407e-01 -9.38004315e-01 -1.19044811e-01 1.16100967e+00
-1.44901052e-01 3.56186837e-01 8.34164679e-01 -5.05385220e-01
-1.60800743e+00 -1.10590947e+00 5.44307530e-01 3.13133359e-01
3.62624489e-02 -4.23566252e-02 -8.28895986e-01 1.93552241e-01
3.48451912e-01 1.39075473e-01 6.07179761e-01 -3.87176514e-01
-1.30362645e-01 -6.73497379e-01 -1.23881352e+00 7.60758460e-01
9.87242997e-01 -5.28145194e-01 6.17103800e-02 4.97548193e-01
7.41371691e-01 -4.90937114e-01 -6.63548410e-01 4.68520612e-01
4.45590973e-01 -1.22937393e+00 1.17547333e+00 1.09573923e-01
1.09208632e+00 -4.39305007e-01 -4.02372748e-01 -1.17207599e+00
-3.54634881e-01 -8.15386236e-01 -7.98586160e-02 9.03684974e-01
7.27557987e-02 -3.64518046e-01 6.60342515e-01 3.21009308e-01
-1.60629526e-01 -8.67393911e-01 -6.37381256e-01 -5.09153485e-01
-5.53397059e-01 -4.42691535e-01 3.66990715e-01 6.65349782e-01
-1.96893275e-01 2.54592746e-01 -7.24206090e-01 1.26508281e-01
1.07812297e+00 3.89399350e-01 6.36718631e-01 -5.54678619e-01
-6.31478369e-01 -3.01136285e-01 -9.06876177e-02 -1.42610145e+00
-3.32409263e-01 -6.84612155e-01 3.45009446e-01 -1.47192657e+00
6.93201005e-01 -3.86191636e-01 -4.14573878e-01 2.80571967e-01
-3.71784419e-01 6.05872095e-01 2.09684119e-01 2.31487662e-01
-5.13663232e-01 6.05079770e-01 1.65855229e+00 -3.05731803e-01
-2.13361591e-01 -6.20537363e-02 -6.57635093e-01 5.78318536e-01
5.04165053e-01 -2.08812177e-01 -6.03254199e-01 -6.77726090e-01
2.87265897e-01 2.89927751e-01 2.87409097e-01 -1.27225423e+00
2.59012401e-01 -2.65439302e-01 5.87516010e-01 -4.66563016e-01
4.47737008e-01 -1.05058277e+00 2.38113217e-02 4.87312317e-01
-3.53676617e-01 -2.30723724e-01 1.97322339e-01 5.86520374e-01
-5.21144271e-01 -1.62470087e-01 1.31316876e+00 -1.26454979e-01
-4.76307720e-01 3.25132847e-01 2.92646922e-02 -4.51291919e-01
8.60791028e-01 -2.83686548e-01 -4.06486362e-01 -6.23673618e-01
-3.12413096e-01 -1.74523607e-01 3.28944325e-01 1.58952605e-02
1.02507961e+00 -1.07691228e+00 -7.92424083e-01 3.30829948e-01
-4.82479259e-02 1.03894919e-01 7.05258012e-01 8.51752996e-01
-6.25517607e-01 8.79620314e-02 -3.05146575e-01 -4.27782744e-01
-1.55371201e+00 5.62152565e-01 1.15824841e-01 -1.51797578e-01
-6.62856758e-01 8.54925275e-01 3.37936014e-01 3.76546800e-01
1.65456384e-01 -1.98268145e-01 1.68131039e-01 -3.97415191e-01
8.29409242e-01 4.70606625e-01 5.71787208e-02 -4.32678372e-01
1.36152551e-01 7.82199800e-01 -1.75260544e-01 2.30533123e-01
1.41903412e+00 -4.78850186e-01 -3.17276921e-03 1.59597502e-03
1.28562927e+00 1.05908111e-01 -1.45235586e+00 -3.83987427e-01
-1.04882061e+00 -1.08047760e+00 6.41422808e-01 -7.55365252e-01
-1.46802402e+00 1.12513936e+00 1.11952007e+00 -3.40774879e-02
2.02571869e+00 -4.73299533e-01 1.29545522e+00 3.21518719e-01
3.17464322e-01 -1.10883117e+00 3.13252538e-01 -4.90285940e-02
8.21481824e-01 -1.50632095e+00 3.98170263e-01 -5.33264577e-01
-3.03088576e-01 1.04153001e+00 5.16923845e-01 3.38673323e-01
4.43495810e-01 5.00882268e-01 -6.48126155e-02 3.19943279e-02
-6.54532015e-01 -1.91752419e-01 6.91123586e-03 5.80518186e-01
4.83583093e-01 -2.42622033e-01 -2.76325226e-01 1.97107226e-01
-1.14924684e-01 1.60803229e-01 4.27650034e-01 7.91455865e-01
-6.04602218e-01 -8.49840045e-01 -4.01693970e-01 4.55617577e-01
-7.03334212e-01 -3.30236107e-01 3.59211087e-01 3.22347283e-01
3.77995074e-01 1.14886665e+00 -1.51118189e-01 -3.00438911e-01
3.30637246e-01 -5.77572405e-01 4.86363351e-01 -1.34085491e-01
-1.74718544e-01 4.48154896e-01 -3.22939515e-01 -5.98431408e-01
-4.93638068e-01 -2.84713149e-01 -9.30276155e-01 -6.59334421e-01
-2.63579756e-01 -3.95077884e-01 7.47326493e-01 4.73588914e-01
4.65688296e-02 7.13105261e-01 8.85751128e-01 -9.17402029e-01
-2.05981225e-01 -6.42940462e-01 -5.46179175e-01 7.16939688e-01
4.34660584e-01 -1.95004284e-01 -3.26398998e-01 4.10872757e-01] | [10.861457824707031, -2.371697187423706] |
cd458ae1-34e9-4e79-b40a-7ee919536e4c | are-you-for-real-detecting-identity-fraud-via | 1908.06820 | null | https://arxiv.org/abs/1908.06820v1 | https://arxiv.org/pdf/1908.06820v1.pdf | Are You for Real? Detecting Identity Fraud via Dialogue Interactions | Identity fraud detection is of great importance in many real-world scenarios such as the financial industry. However, few studies addressed this problem before. In this paper, we focus on identity fraud detection in loan applications and propose to solve this problem with a novel interactive dialogue system which consists of two modules. One is the knowledge graph (KG) constructor organizing the personal information for each loan applicant. The other is structured dialogue management that can dynamically generate a series of questions based on the personal KG to ask the applicants and determine their identity states. We also present a heuristic user simulator based on problem analysis to evaluate our method. Experiments have shown that the trainable dialogue system can effectively detect fraudsters, and achieve higher recognition accuracy compared with rule-based systems. Furthermore, our learned dialogue strategies are interpretable and flexible, which can help promote real-world applications. | ['Cheng-qing Zong', 'Jiajun Zhang', 'Weikang Wang', 'Qian Li', 'Zhifei Li'] | 2019-08-19 | are-you-for-real-detecting-identity-fraud-via-1 | https://aclanthology.org/D19-1185 | https://aclanthology.org/D19-1185.pdf | ijcnlp-2019-11 | ['dialogue-management'] | ['natural-language-processing'] | [-1.36595652e-01 3.68044406e-01 -1.54638007e-01 -5.15408874e-01
-4.39582169e-01 -3.81492794e-01 1.38224572e-01 1.56351894e-01
-7.11704269e-02 8.39868188e-01 -1.87428117e-01 -6.28464103e-01
4.62505743e-02 -1.14195669e+00 1.00520207e-02 -2.21414551e-01
1.86943099e-01 8.49402428e-01 3.42180938e-01 -6.68305337e-01
6.18035614e-01 3.22273970e-01 -1.26428294e+00 4.61349308e-01
1.45434451e+00 6.15207434e-01 7.74193108e-02 5.42970240e-01
-4.04030114e-01 9.47889030e-01 -9.46757376e-01 -9.37170386e-01
-1.35750612e-04 -5.94024599e-01 -1.18979430e+00 2.31086135e-01
-1.70218512e-01 -7.11431444e-01 7.17311203e-02 1.11241603e+00
4.93591577e-01 -3.55050936e-02 4.49494541e-01 -1.22017384e+00
-3.58586282e-01 6.94023550e-01 -1.94653630e-01 -4.10810374e-02
8.97320271e-01 2.55093519e-02 9.04045701e-01 -4.82013673e-01
5.30888200e-01 1.18212211e+00 6.32399082e-01 7.64900804e-01
-1.02522957e+00 -5.90632737e-01 1.99405849e-01 5.23974299e-01
-1.01711547e+00 -1.07568502e-01 7.08265483e-01 -4.87503231e-01
8.07838500e-01 2.67493546e-01 7.71156132e-01 6.85277283e-01
-1.17210671e-01 1.01445854e+00 8.23806584e-01 -7.53844678e-01
2.74198949e-01 6.44884944e-01 6.55138433e-01 1.10976362e+00
2.85723627e-01 -4.99581873e-01 -5.82293093e-01 -6.85317636e-01
1.02057350e+00 -1.84640944e-01 -2.73756653e-01 -3.17302257e-01
-6.32135212e-01 1.16769838e+00 -3.00681144e-01 1.33021921e-01
-4.36089426e-01 -5.59499383e-01 4.16191280e-01 5.11114717e-01
1.86661467e-01 4.43828613e-01 -7.08792657e-02 -2.93053210e-01
-3.06598037e-01 4.01916414e-01 1.78085506e+00 8.55288267e-01
7.39678323e-01 -2.12658674e-01 -1.41469270e-01 1.14033020e+00
3.86205077e-01 -5.03108138e-03 6.12635076e-01 -8.63615215e-01
5.83660722e-01 1.17247868e+00 4.46189672e-01 -1.05233181e+00
-2.08460733e-01 1.85998559e-01 -5.37175119e-01 1.05291784e-01
5.66969931e-01 -4.08611000e-01 -2.98713058e-01 1.23940337e+00
4.31246877e-01 -9.82217118e-02 2.29632765e-01 9.54708695e-01
8.17134082e-01 2.67561555e-01 1.53033510e-01 -3.37249368e-01
1.63055277e+00 -6.57113016e-01 -9.77242112e-01 -1.25311360e-01
8.20790410e-01 -5.74105382e-01 9.17966485e-01 5.49839437e-01
-9.24282193e-01 -2.47057348e-01 -9.13299561e-01 1.38535500e-01
-1.43414512e-01 2.24205852e-01 9.86166418e-01 1.12531126e+00
-7.49598801e-01 5.52419424e-01 -5.65738797e-01 -4.57811654e-01
1.29141454e-02 4.36862230e-01 4.80919331e-02 1.95543647e-01
-1.63132310e+00 9.61162865e-01 3.17973226e-01 6.55702800e-02
-2.43970811e-01 -1.19678155e-02 -8.58522952e-01 1.19853027e-01
4.89352435e-01 -6.79552436e-01 1.71797585e+00 -8.38473618e-01
-1.90714550e+00 1.02621913e+00 -5.62807173e-02 -4.28368658e-01
7.71923304e-01 -1.25510216e-01 -3.42789918e-01 1.58559322e-01
7.20128864e-02 -8.52265656e-02 3.78671497e-01 -1.08916986e+00
-9.66458380e-01 -4.30699766e-01 3.63901317e-01 4.03306365e-01
-5.42855680e-01 2.03843236e-01 -6.08844578e-01 -4.02226418e-01
7.60885328e-02 -6.19423628e-01 -2.10519329e-01 -4.22074586e-01
-2.78165966e-01 -4.20170456e-01 5.27472019e-01 -9.40883160e-01
1.63952351e+00 -1.62749052e+00 -3.18468004e-01 4.63492483e-01
-5.47902845e-02 5.51751435e-01 2.81188846e-01 6.35464489e-01
1.81087226e-01 -1.68041289e-02 -2.80468911e-01 3.55521962e-02
4.96733822e-02 2.18400896e-01 -3.09130847e-01 -2.49626726e-01
6.24352172e-02 4.79025126e-01 -9.98212099e-01 -6.36869967e-01
-7.31282607e-02 -2.04922944e-01 -3.92917335e-01 6.33742392e-01
-1.08197354e-01 -3.43869627e-02 -8.97070706e-01 6.45666361e-01
5.72359920e-01 -4.32971977e-02 9.01685715e-01 3.05279851e-01
3.63583744e-01 2.51825929e-01 -1.46288919e+00 1.13230884e+00
-3.44165891e-01 6.39461651e-02 2.83477783e-01 -1.07078397e+00
1.28391099e+00 3.15653622e-01 2.01357394e-01 -3.94976467e-01
-1.75011888e-01 8.02320987e-02 -1.55221015e-01 -9.74391520e-01
6.45731688e-01 -4.96306792e-02 -2.04918712e-01 8.67596567e-01
-1.91996560e-01 4.98128645e-02 2.30798900e-01 3.96034718e-01
1.07034123e+00 -8.13080836e-03 4.81728435e-01 9.50814858e-02
8.07319403e-01 1.52170181e-01 7.23089755e-01 9.10694718e-01
-2.80551523e-01 8.19342509e-02 9.57182288e-01 -5.28834283e-01
-3.78340542e-01 -7.05627501e-01 1.41135558e-01 9.08846080e-01
4.16547388e-01 -3.69856566e-01 -1.18838561e+00 -1.09491193e+00
2.95903891e-01 7.72058606e-01 -9.85014439e-02 -1.03892542e-01
-5.45711935e-01 -8.41770768e-01 4.76995617e-01 2.31760785e-01
8.61951470e-01 -1.38802433e+00 -5.14016628e-01 5.88210285e-01
-5.47493517e-01 -8.50123584e-01 -3.28249007e-01 -2.43780866e-01
-8.95892203e-01 -1.46448100e+00 -1.84917316e-01 -1.01796818e+00
7.27845550e-01 5.49078882e-02 1.10836565e+00 4.74730730e-01
-2.11304024e-01 3.56184989e-01 -4.15483773e-01 -4.93089229e-01
-9.81900036e-01 1.76368758e-01 -1.15874454e-01 2.48432964e-01
8.08409095e-01 -2.29597986e-01 -3.82663518e-01 7.31638491e-01
-6.88668609e-01 1.34137213e-01 3.14518094e-01 1.05302572e+00
-5.18653542e-02 3.13811488e-02 1.14546430e+00 -1.60979140e+00
1.57676697e+00 -2.65678883e-01 -5.95916569e-01 8.50106061e-01
-8.69679749e-01 1.19081452e-01 4.42406744e-01 -2.55856782e-01
-1.57198465e+00 2.10808322e-01 -1.04842439e-01 2.87266880e-01
-1.63093776e-01 5.33988595e-01 -2.94291496e-01 -8.38868618e-02
6.76972032e-01 1.36206076e-01 2.64216572e-01 -4.19757545e-01
-3.76585424e-02 1.26195848e+00 4.76512253e-01 -8.19806874e-01
3.68122965e-01 -1.23804137e-01 -5.46983480e-01 -6.34178996e-01
-4.63682055e-01 -6.52592897e-01 -4.90948468e-01 -3.39953601e-01
3.93838465e-01 -5.53657293e-01 -1.34889865e+00 7.87764013e-01
-1.32193947e+00 -1.74096271e-01 7.03002959e-02 2.11428002e-01
-5.88291526e-01 9.30490851e-01 -1.10344839e+00 -1.27803719e+00
-6.44047260e-01 -7.67670214e-01 5.04151344e-01 3.32839489e-01
-4.00365710e-01 -1.03150082e+00 -3.75925936e-02 7.94484615e-01
-4.29233201e-02 1.75236426e-02 1.14706290e+00 -8.97411644e-01
-4.68834847e-01 -3.48582089e-01 -3.75611410e-02 2.41819367e-01
2.54250407e-01 -3.73555094e-01 -8.12222183e-01 2.70799994e-02
2.92691618e-01 -2.49336064e-01 4.63678569e-01 -2.97521859e-01
9.61180091e-01 -5.35647452e-01 -2.78446555e-01 4.85034324e-02
7.90678620e-01 5.66477656e-01 7.31966376e-01 4.24883187e-01
3.04009587e-01 8.51671338e-01 1.00561070e+00 5.86436212e-01
6.91914976e-01 7.56666362e-01 9.36017185e-03 6.18535234e-03
6.08431339e-01 -1.89121202e-01 4.50910181e-01 5.07544756e-01
-2.45296121e-01 -3.85785513e-02 -1.03239858e+00 2.84250319e-01
-2.52537847e+00 -1.06023228e+00 -1.20456085e-01 2.14264345e+00
1.17990661e+00 1.34265140e-01 2.47308210e-01 1.07447878e-01
1.02273655e+00 -4.65519905e-01 -5.32555401e-01 -6.10699534e-01
2.64950752e-01 -1.99520588e-03 1.09624982e-01 6.43100500e-01
-8.48816991e-01 1.12392426e+00 6.61290026e+00 5.62978506e-01
-6.07552290e-01 -2.43596613e-01 5.76735675e-01 5.72290897e-01
-3.66000347e-02 9.98409316e-02 -8.59101474e-01 3.25515777e-01
4.98580724e-01 -4.78033036e-01 2.70455480e-01 9.18835521e-01
1.28251761e-01 -4.03291911e-01 -1.08995342e+00 6.91938818e-01
2.68632956e-02 -1.03016675e+00 2.12463066e-01 -4.11689766e-02
2.12664679e-01 -1.02838433e+00 -6.07755005e-01 5.47364533e-01
6.56158030e-01 -5.89599609e-01 1.83885887e-01 6.60259306e-01
2.19414815e-01 -8.14136207e-01 9.77606058e-01 8.68005157e-01
-7.05443561e-01 -1.84659362e-01 -3.12098891e-01 -1.67739108e-01
9.73937064e-02 3.99109840e-01 -1.51926470e+00 8.13893437e-01
3.00833941e-01 2.80319810e-01 -2.90700465e-01 1.03571451e+00
-4.07073706e-01 5.83404124e-01 -5.54177724e-02 -3.46362978e-01
-3.08727056e-01 -6.09340250e-01 2.45333791e-01 1.21806645e+00
1.21691823e-01 5.69692194e-01 2.82542139e-01 6.66064441e-01
4.43478450e-02 4.13937002e-01 -5.07597804e-01 2.45435774e-01
3.90312284e-01 1.16606402e+00 -5.43169677e-01 -5.60012221e-01
-3.18259895e-01 1.14757180e+00 4.40443724e-01 2.39857823e-01
-4.49874103e-01 -6.82349920e-01 5.35455346e-01 1.82778344e-01
-6.49917349e-02 1.07175432e-01 -2.66404599e-01 -1.33650315e+00
2.17483759e-01 -1.06181264e+00 8.86446655e-01 -4.03857708e-01
-1.37729585e+00 3.34373862e-01 -2.71826327e-01 -1.05828571e+00
-6.60211205e-01 -4.98036295e-01 -9.32267904e-01 9.50867236e-01
-1.20693028e+00 -7.68400431e-01 -3.56112331e-01 5.59508383e-01
4.20305371e-01 -3.95991266e-01 1.23565829e+00 3.33710723e-02
-8.94050598e-01 8.34871233e-01 -2.63615042e-01 4.31147218e-01
7.06307173e-01 -1.32572556e+00 4.07189280e-01 4.16992486e-01
-1.07011661e-01 8.27933371e-01 3.93560171e-01 -9.49448228e-01
-1.09985638e+00 -7.84551680e-01 1.04169190e+00 -3.82690430e-01
4.01921511e-01 -3.87509584e-01 -1.33921003e+00 3.56189698e-01
1.25019355e-02 -7.50143647e-01 8.00372243e-01 2.53295898e-01
-2.51244288e-02 -1.77282598e-02 -1.39093542e+00 4.81879413e-01
9.07539308e-01 -5.29275477e-01 -6.97542787e-01 4.92246181e-01
1.38561755e-01 -6.42771125e-01 -8.40057373e-01 7.64639005e-02
3.35757047e-01 -1.13748252e+00 6.91812038e-01 -7.78655291e-01
8.10935423e-02 8.11735243e-02 6.07309818e-01 -1.20042002e+00
-1.79369345e-01 -7.51510441e-01 -1.95412621e-01 1.43503845e+00
3.72731626e-01 -9.24203694e-01 1.11006773e+00 1.31029260e+00
3.40301603e-01 -7.37758398e-01 -5.01943111e-01 -5.32536745e-01
-4.97704953e-01 -1.31371483e-01 7.49826014e-01 1.24095166e+00
9.13996041e-01 4.87884879e-01 -3.47512215e-01 2.09118769e-01
3.59351575e-01 3.86090159e-01 9.02688920e-01 -1.31551707e+00
-3.54212284e-01 -2.81159848e-01 -3.20604861e-01 -9.87045288e-01
1.61647469e-01 -5.54092824e-01 -1.38735980e-01 -1.61647546e+00
1.46111578e-01 -4.51817393e-01 -1.06450524e-02 6.61853731e-01
-4.70363408e-01 -4.65821177e-01 -1.39980376e-01 2.80220240e-01
-5.10984004e-01 2.72573769e-01 8.76189291e-01 -1.48405628e-02
-7.17311442e-01 5.82603037e-01 -8.59925866e-01 7.59492218e-01
8.70616674e-01 -1.45201489e-01 -2.72859484e-01 8.65752473e-02
-8.88840556e-02 6.50717258e-01 1.17111094e-01 -4.77163345e-01
4.50097620e-01 -3.69623393e-01 3.83423343e-02 -1.24508627e-01
1.42026870e-02 -4.48339611e-01 -1.99701637e-01 5.75054586e-01
-2.54999131e-01 -3.86424884e-02 -2.77332440e-02 4.64731604e-01
-3.07511300e-01 -6.94969594e-01 4.30595130e-01 -1.98200136e-01
-8.18292916e-01 -9.35680419e-02 -6.57614231e-01 -9.43248868e-02
1.18401682e+00 -3.81543577e-01 -2.18299970e-01 -9.22625840e-01
-7.08509624e-01 7.13615775e-01 2.05048710e-01 3.62515628e-01
6.75288141e-01 -1.08261931e+00 -6.11802518e-01 2.75743484e-01
1.98917821e-01 -1.17900632e-01 -4.24581021e-02 1.19971536e-01
-7.52913833e-01 1.99156836e-01 -2.15258762e-01 -1.84631169e-01
-1.55271828e+00 7.20576718e-02 5.19201875e-01 -6.71619356e-01
-3.14528376e-01 3.83704841e-01 -1.27124459e-01 -5.85472167e-01
4.78757054e-01 1.18413836e-01 -7.34510243e-01 1.57244578e-01
3.73414129e-01 2.64149755e-01 5.46181351e-02 -8.59846324e-02
-1.77712068e-01 1.95819542e-01 -3.87248069e-01 -6.48292527e-02
1.10387707e+00 1.21705517e-01 -1.21208146e-01 4.74725962e-02
3.56867343e-01 -2.27937549e-01 -7.74668872e-01 -3.74750763e-01
5.21433115e-01 -6.11577928e-01 -4.89707083e-01 -9.16779876e-01
-6.69682324e-01 5.38572311e-01 4.03632820e-01 6.81164801e-01
9.33968723e-01 -4.93952304e-01 7.72132695e-01 8.98056388e-01
6.61752045e-01 -1.61996675e+00 -2.19476363e-03 4.46995437e-01
7.61588991e-01 -1.39851332e+00 -1.75603360e-01 -8.38666439e-01
-1.15848196e+00 1.29411864e+00 9.60763276e-01 3.38175356e-01
3.71994287e-01 -1.23836406e-01 4.60476011e-01 -2.26745516e-01
-5.95362067e-01 6.26934785e-03 -1.55849606e-01 5.58429956e-01
3.06900948e-01 2.23113000e-02 -6.86900079e-01 1.15852773e+00
-3.66914421e-02 1.75052583e-01 7.61098564e-01 1.08362496e+00
-5.70626557e-01 -1.71063614e+00 -4.72719640e-01 5.16649783e-01
-3.18323642e-01 2.05263689e-01 -8.75556886e-01 5.69016993e-01
-1.82063699e-01 1.07469964e+00 -3.28696162e-01 -4.21652794e-01
8.33096564e-01 4.87149328e-01 1.64258227e-01 -8.96457911e-01
-1.13450539e+00 -4.09170777e-01 6.78200781e-01 -1.89444751e-01
1.49222435e-02 -6.65058196e-01 -1.34286320e+00 -2.90540248e-01
-5.32345593e-01 6.67790711e-01 3.19347084e-01 6.33631647e-01
3.28468055e-01 -1.19690202e-01 8.25883031e-01 -1.74045190e-02
-9.29226041e-01 -8.05263758e-01 -8.15439165e-01 7.55348623e-01
-2.34277993e-01 -3.58649731e-01 1.25409402e-02 7.70972148e-02] | [12.88281536102295, 7.935924530029297] |
092bdaec-e5dc-4189-bddc-cd7976572c10 | unsupervised-deep-digital-staining-for | 2303.02057 | null | https://arxiv.org/abs/2303.02057v1 | https://arxiv.org/pdf/2303.02057v1.pdf | Unsupervised Deep Digital Staining For Microscopic Cell Images Via Knowledge Distillation | Staining is critical to cell imaging and medical diagnosis, which is expensive, time-consuming, labor-intensive, and causes irreversible changes to cell tissues. Recent advances in deep learning enabled digital staining via supervised model training. However, it is difficult to obtain large-scale stained/unstained cell image pairs in practice, which need to be perfectly aligned with the supervision. In this work, we propose a novel unsupervised deep learning framework for the digital staining of cell images using knowledge distillation and generative adversarial networks (GANs). A teacher model is first trained mainly for the colorization of bright-field images. After that,a student GAN for staining is obtained by knowledge distillation with hybrid non-reference losses. We show that the proposed unsupervised deep staining method can generate stained images with more accurate positions and shapes of the cell targets. Compared with other unsupervised deep generative models for staining, our method achieves much more promising results both qualitatively and quantitatively. | ['Bihan Wen', 'Alex C. Kot', 'Shuyan Zhang', 'Lanqing Guo', 'Ziwang Xu'] | 2023-03-03 | null | null | null | null | ['colorization', 'medical-diagnosis'] | ['computer-vision', 'medical'] | [ 3.87154877e-01 4.09838781e-02 4.67038065e-01 -1.13261834e-01
-8.27116668e-01 -7.30084777e-01 1.91327974e-01 -1.04093097e-01
-4.26637590e-01 1.25971580e+00 -3.07891250e-01 -2.57591903e-02
4.99587625e-01 -9.24907267e-01 -7.53273249e-01 -1.57947624e+00
6.15669429e-01 6.15207493e-01 4.72192131e-02 2.56929338e-01
-1.06830448e-01 8.08279932e-01 -8.76840949e-01 -6.59122467e-02
8.98599684e-01 6.31588519e-01 4.86656949e-02 8.86012435e-01
-9.98300016e-02 1.19237447e+00 -8.18161786e-01 -5.22224545e-01
1.79454572e-02 -7.93638170e-01 -7.50630200e-01 2.14947000e-01
5.16164899e-02 -3.48234713e-01 -1.11224644e-01 1.24711001e+00
7.67339170e-01 -1.23435445e-01 1.05342877e+00 -1.14525223e+00
-8.17543924e-01 1.01197615e-01 -7.33451843e-01 -2.88688280e-02
-3.56071323e-01 2.88884401e-01 3.87308210e-01 -5.50937533e-01
8.23635936e-01 7.67451644e-01 4.33597058e-01 9.49399114e-01
-1.54440022e+00 -7.81028032e-01 -4.53398436e-01 -1.81333736e-01
-1.24950171e+00 -1.12650692e-01 9.00935233e-01 -5.73880553e-01
2.06331238e-01 1.39847457e-01 8.40723276e-01 8.26090991e-01
1.87866688e-01 6.88057959e-01 1.40256274e+00 -3.15816730e-01
3.17210883e-01 1.13995761e-01 -3.63364458e-01 6.84811950e-01
2.13600338e-01 -1.14761591e-01 -7.11320490e-02 3.48970992e-03
1.35889995e+00 4.37013358e-01 -3.88788283e-01 -2.02428892e-01
-1.11013436e+00 5.81017494e-01 4.99812990e-01 3.96743238e-01
-2.89243728e-01 3.84013176e-01 2.24306673e-01 -3.79596427e-02
3.65123630e-01 4.05690819e-01 -2.24436343e-01 -7.57064996e-03
-9.07498240e-01 -1.59716830e-01 5.06378353e-01 5.02529681e-01
8.90811980e-01 2.12072939e-01 -2.26333305e-01 6.07063949e-01
1.60309777e-01 6.75398111e-01 2.91328520e-01 -1.08984113e+00
-3.19889724e-01 5.49537003e-01 2.44677395e-01 -8.40126097e-01
-1.22182049e-01 -4.00013208e-01 -1.26505446e+00 6.14242077e-01
7.47440755e-01 -3.10525119e-01 -1.11647725e+00 1.65797710e+00
4.59687769e-01 2.81011134e-01 2.85937607e-01 9.38284934e-01
6.07011914e-01 6.44269705e-01 1.89864203e-01 -3.67193371e-01
9.10107017e-01 -9.31608677e-01 -9.51717079e-01 9.19976681e-02
5.45640349e-01 -6.39612794e-01 1.05259156e+00 3.33250761e-01
-1.12504482e+00 -2.87956536e-01 -9.65837955e-01 -2.73332179e-01
-4.20134723e-01 6.23516403e-02 7.57269025e-01 4.26649988e-01
-9.64741170e-01 2.33562365e-01 -1.03752899e+00 -1.79490969e-01
9.63700116e-01 4.16758627e-01 -3.47827196e-01 -6.26477739e-03
-6.66274726e-01 4.64940429e-01 -9.46972221e-02 1.65585503e-01
-1.18787515e+00 -6.95730805e-01 -5.21763682e-01 -3.71095948e-02
-1.41098216e-01 -7.16098428e-01 1.07410550e+00 -1.25530434e+00
-1.83254528e+00 1.21634507e+00 -8.30247998e-02 -3.11965376e-01
5.46968281e-01 1.99580044e-01 1.05310060e-01 3.22587818e-01
-1.73507750e-01 7.60836244e-01 5.68729103e-01 -1.69905055e+00
-1.46919727e-01 -1.62519678e-01 -2.01214552e-01 -1.07681036e-01
-1.50213167e-01 -2.21926033e-01 -3.76672685e-01 -6.70776665e-01
-2.63863206e-01 -5.88777781e-01 -2.65393525e-01 3.52097005e-01
-4.78913128e-01 4.13814574e-01 1.02675116e+00 -8.54055583e-01
5.12251973e-01 -2.31449270e+00 1.63982674e-01 1.21367835e-01
4.36853796e-01 3.08664739e-01 -2.86750738e-02 1.64049417e-01
6.87946305e-02 1.33585688e-02 -3.31150323e-01 -4.01785314e-01
-1.60792366e-01 4.24434155e-01 -2.37423062e-01 6.31386817e-01
-2.01119166e-02 1.24416363e+00 -1.03677332e+00 -8.40162277e-01
8.34735110e-02 9.98167694e-01 -3.50328773e-01 4.35054332e-01
-1.58518344e-01 1.04365110e+00 -3.03617746e-01 8.35583389e-01
7.65542388e-01 -5.23144484e-01 1.60314396e-01 -1.83647543e-01
3.14066410e-01 -4.26587969e-01 -3.38869393e-01 1.23861754e+00
-2.19224066e-01 6.43433392e-01 1.85898021e-01 -1.05074668e+00
8.33336353e-01 2.63303488e-01 3.26648176e-01 -3.49867344e-01
3.69541675e-01 2.49420181e-01 -2.94872195e-01 -1.95822477e-01
-2.18047768e-01 -7.36973882e-01 3.51750135e-01 3.02182853e-01
9.46739241e-02 -5.62275767e-01 -1.42213982e-02 1.12772748e-01
9.69721556e-01 -7.23405629e-02 -8.59902501e-02 -4.17464934e-02
4.43781316e-01 -2.39822641e-01 6.05215728e-01 1.77270755e-01
-9.07848626e-02 8.90812874e-01 7.52285838e-01 -3.43925804e-01
-1.14697742e+00 -1.06500769e+00 8.06855112e-02 3.97342891e-01
1.65296882e-01 4.22289580e-01 -1.06857419e+00 -7.67028153e-01
-2.41601825e-01 3.83874804e-01 -8.15517068e-01 -6.67044446e-02
-2.84013212e-01 -7.94842422e-01 7.08205521e-01 6.57554686e-01
4.94076222e-01 -1.22078633e+00 -2.09870756e-01 1.24874957e-01
2.64332779e-02 -7.56111681e-01 -3.02722901e-01 5.35125434e-01
-7.92723298e-01 -1.15050328e+00 -1.36725557e+00 -1.27417421e+00
1.49866247e+00 -1.40094370e-01 1.00658083e+00 3.67350012e-01
-3.33524525e-01 4.99369688e-02 -8.08033645e-02 -4.93732721e-01
-3.91517252e-01 -3.37460220e-01 -3.96603227e-01 1.27978280e-01
9.02038813e-02 -6.64415240e-01 -9.59895968e-01 2.67574877e-01
-1.25997424e+00 1.00147441e-01 6.70556426e-01 1.38858163e+00
1.31593812e+00 3.29296201e-01 4.37096655e-01 -1.38787973e+00
9.56822261e-02 3.80853154e-02 -7.78289318e-01 1.84203386e-01
-2.79451042e-01 -1.63101062e-01 9.32187259e-01 -4.42626804e-01
-1.22161222e+00 2.12495178e-01 -2.69577891e-01 -4.86822933e-01
-2.14713305e-01 1.89289257e-01 -3.13151985e-01 -3.62752378e-01
4.53281075e-01 4.16288793e-01 2.91268408e-01 -1.14967279e-01
-1.17122401e-02 2.78298080e-01 8.11851323e-01 -3.90193433e-01
1.06187737e+00 9.52636242e-01 1.64529756e-01 -4.54766959e-01
-8.16608548e-01 -7.88831860e-02 -4.77736562e-01 -1.11382212e-02
1.24761009e+00 -7.12316632e-01 -5.35389125e-01 1.00148296e+00
-9.49573457e-01 -1.08102870e+00 -4.03496742e-01 1.98724613e-01
-5.99108458e-01 2.83603400e-01 -1.10087979e+00 -5.58732569e-01
-3.66091907e-01 -9.71254587e-01 9.90758061e-01 5.29307544e-01
2.43024170e-01 -1.37289488e+00 3.07185173e-01 3.44561338e-01
3.94318610e-01 7.62838781e-01 1.02951229e+00 -1.55421495e-01
-6.40576780e-01 -3.39529246e-01 -2.87131667e-01 5.39503932e-01
4.02120680e-01 2.85116911e-01 -9.32634175e-01 -3.26883912e-01
-5.31718656e-02 -6.61326587e-01 7.36050606e-01 5.12362301e-01
1.44956481e+00 -2.30345652e-01 -4.46603358e-01 9.66388285e-01
1.50575709e+00 4.68556434e-01 1.14210653e+00 9.24270302e-02
8.40593874e-01 2.44587615e-01 3.47029448e-01 -5.19428682e-03
2.61182860e-02 1.77143708e-01 4.23989534e-01 -1.04453099e+00
-1.53143242e-01 -1.94974884e-01 -5.61336912e-02 6.41094863e-01
-1.42859757e-01 -3.38518828e-01 -6.85058653e-01 6.49750054e-01
-1.49255228e+00 -8.42331707e-01 4.83885109e-02 1.89460301e+00
1.29186654e+00 -7.11871311e-02 -1.44529417e-01 1.83212787e-01
6.91145897e-01 -4.18645561e-01 -7.83621788e-01 5.40556572e-02
-2.63602108e-01 5.27679563e-01 5.30381799e-01 3.68716359e-01
-8.86328220e-01 8.86670291e-01 6.22218180e+00 7.67561316e-01
-1.44398832e+00 2.09226802e-01 1.27537870e+00 8.55761841e-02
-3.88343036e-01 -3.41805488e-01 -2.99874783e-01 5.16660631e-01
3.08516890e-01 1.39824644e-01 2.47170284e-01 5.05125105e-01
3.55136506e-02 -8.70483667e-02 -8.86605382e-01 1.02640033e+00
-1.66520208e-01 -1.28942740e+00 1.18274353e-02 1.94493800e-01
1.16525090e+00 -4.54145044e-01 2.86961704e-01 -5.50598465e-02
6.15872025e-01 -1.17478395e+00 2.69899100e-01 7.06243515e-01
1.09112334e+00 -1.00068903e+00 1.11452818e+00 1.74942613e-01
-5.97816467e-01 4.33163941e-01 -5.09917557e-01 3.43507916e-01
1.23862088e-01 7.80270815e-01 -7.61514723e-01 1.31799623e-01
5.18742085e-01 5.61193049e-01 -3.13223034e-01 7.96370745e-01
-4.47652340e-01 7.67614841e-01 -8.62397626e-02 2.96970103e-02
3.25255878e-02 -2.40489796e-01 -5.34792915e-02 1.09296751e+00
2.53127605e-01 5.69912717e-02 -1.62294164e-01 9.08350825e-01
-3.95030707e-01 -1.14062928e-01 -4.54039156e-01 -3.68289858e-01
1.70774952e-01 1.56648099e+00 -1.26213646e+00 -4.46761489e-01
-1.76470548e-01 1.25941312e+00 3.44205886e-01 7.23428488e-01
-1.03035975e+00 -4.61521178e-01 2.83439606e-01 1.33716583e-01
2.66144693e-01 1.24393612e-01 -3.08209240e-01 -1.05261433e+00
-5.81377685e-01 -6.33478403e-01 1.47672385e-01 -9.94014442e-01
-1.51769412e+00 3.42056751e-01 -6.53961360e-01 -1.09698677e+00
1.09014258e-01 -6.20528162e-01 -7.04230547e-01 7.90571392e-01
-1.76920724e+00 -1.18064439e+00 -4.81363088e-01 5.89455962e-01
-5.17069884e-02 4.73189317e-02 9.22279298e-01 2.81255960e-01
-4.97919381e-01 5.46791792e-01 3.77627671e-01 5.06588280e-01
8.29640567e-01 -1.63759434e+00 -3.30052018e-01 7.86354780e-01
-1.03725962e-01 4.15773183e-01 5.80645680e-01 -4.58283812e-01
-1.03146207e+00 -1.24696493e+00 4.50915694e-01 -1.66460961e-01
4.14891809e-01 -1.35086223e-01 -8.55615675e-01 5.90135694e-01
4.67878520e-01 3.95328760e-01 1.01208067e+00 -6.02454722e-01
-2.68813707e-02 -1.96112797e-01 -1.49627137e+00 6.52012169e-01
3.60872358e-01 -5.58780789e-01 1.69812888e-02 4.84601915e-01
3.78783464e-01 -4.28560078e-01 -8.02116454e-01 7.07559213e-02
3.88423175e-01 -7.61613846e-01 7.74039030e-01 -1.74874067e-01
5.62475026e-01 -7.40848184e-01 2.68759966e-01 -1.32699013e+00
-1.91488162e-01 -5.68010926e-01 2.38638535e-01 1.31910515e+00
3.05702001e-01 -7.35709608e-01 1.09128392e+00 3.30894947e-01
-7.75478929e-02 -7.87189424e-01 -6.06246412e-01 -4.37952340e-01
3.18041682e-01 4.64399636e-01 3.85989904e-01 1.07279813e+00
-2.24529222e-01 2.36909121e-01 -1.19315363e-01 6.00100346e-02
8.43314826e-01 1.27364546e-01 8.79352987e-01 -9.23161745e-01
-3.80814314e-01 -2.83168554e-01 -4.76346999e-01 -8.79311621e-01
1.78469390e-01 -6.20562434e-01 1.31662592e-01 -1.64051390e+00
4.68996257e-01 -6.36603087e-02 -4.51849431e-01 6.59235299e-01
-2.13016212e-01 9.15849566e-01 -2.93688715e-01 2.22667769e-01
-6.10074461e-01 4.33865130e-01 1.92024148e+00 -5.31544387e-01
4.63942699e-02 -4.38489228e-01 -6.09813631e-01 6.06760085e-01
9.62873340e-01 -4.04508501e-01 -3.30060005e-01 -3.93686593e-01
2.08957449e-01 -2.90416437e-03 4.90861863e-01 -8.69128883e-01
1.43313244e-01 -1.20461591e-01 8.63660812e-01 -4.93328929e-01
1.02197193e-01 -8.78368497e-01 4.44724917e-01 5.38979530e-01
-1.42043456e-01 -4.52223122e-01 -5.94423618e-04 3.90390843e-01
-4.92984504e-01 3.42098437e-02 1.27100003e+00 -2.48957962e-01
6.69156238e-02 2.90328234e-01 -4.44706559e-01 -2.07854167e-01
1.10640061e+00 -2.63504863e-01 -3.82950574e-01 -3.82204890e-01
-7.63337255e-01 5.65856509e-02 9.53707218e-01 -6.54575884e-01
5.57795167e-01 -1.39736331e+00 -2.52629757e-01 1.15402870e-01
-1.72188833e-01 6.09649837e-01 2.90400714e-01 9.98097658e-01
-1.08475137e+00 -1.61272958e-01 -3.35090786e-01 -4.50924248e-01
-1.07272673e+00 6.51975751e-01 4.28646117e-01 -6.53063118e-01
-2.77703971e-01 1.02162504e+00 7.13969111e-01 -2.93598324e-01
5.89792617e-02 -1.86495125e-01 -8.47224053e-03 -2.83489943e-01
4.21697080e-01 1.45719647e-01 -3.30093592e-01 -3.50936413e-01
-5.87625615e-02 6.29490018e-01 -5.05360514e-02 1.37842402e-01
1.25963521e+00 2.68355161e-02 -4.57431585e-01 2.98813730e-01
1.15892744e+00 1.40672579e-01 -1.70428336e+00 2.47085452e-01
-6.28587604e-01 -2.32024640e-01 2.90359501e-02 -8.90386581e-01
-1.76946712e+00 1.00466299e+00 4.41641659e-01 1.11752152e-01
1.38801730e+00 -8.64156187e-02 9.90827203e-01 1.90645665e-01
2.87292480e-01 -7.83413231e-01 3.35014105e-01 1.18251489e-02
4.32836622e-01 -9.45078909e-01 -3.37228566e-01 -4.71393555e-01
-4.18656528e-01 9.81476307e-01 6.29464626e-01 -1.84571564e-01
4.20377642e-01 7.50840306e-01 5.91245651e-01 -2.36003131e-01
-4.28622454e-01 -1.19232936e-02 -2.61747390e-01 9.38390374e-01
4.50154543e-01 -1.43429011e-01 -1.29675016e-01 4.71122086e-01
2.44310707e-01 3.28080177e-01 2.78380871e-01 8.18529606e-01
-6.04335852e-02 -1.02925146e+00 -2.58503199e-01 2.56846905e-01
-8.43480766e-01 7.24030361e-02 -5.38998544e-01 5.95944405e-01
8.75598639e-02 3.31949085e-01 1.02677688e-01 5.00853220e-03
-1.01275407e-01 -8.57080072e-02 6.49882495e-01 -4.97744918e-01
-2.41815865e-01 4.92315978e-01 -6.12850845e-01 -9.42549184e-02
-6.59104943e-01 -3.42347652e-01 -1.67361403e+00 -2.22308546e-01
-3.27277690e-01 2.49600202e-01 8.23472887e-02 6.08651876e-01
1.20896608e-01 7.73227513e-01 6.23549104e-01 -7.16703892e-01
1.61599770e-01 -7.44133532e-01 -1.01322651e+00 5.45306861e-01
3.11687678e-01 -5.16903758e-01 -5.84924519e-01 9.20149565e-01] | [14.644976615905762, -2.958805799484253] |
9cec4a8b-1665-437e-9520-1863c0d40676 | second-order-unsupervised-neural-dependency | 2010.14720 | null | https://arxiv.org/abs/2010.14720v1 | https://arxiv.org/pdf/2010.14720v1.pdf | Second-Order Unsupervised Neural Dependency Parsing | Most of the unsupervised dependency parsers are based on first-order probabilistic generative models that only consider local parent-child information. Inspired by second-order supervised dependency parsing, we proposed a second-order extension of unsupervised neural dependency models that incorporate grandparent-child or sibling information. We also propose a novel design of the neural parameterization and optimization methods of the dependency models. In second-order models, the number of grammar rules grows cubically with the increase of vocabulary size, making it difficult to train lexicalized models that may contain thousands of words. To circumvent this problem while still benefiting from both second-order parsing and lexicalization, we use the agreement-based learning framework to jointly train a second-order unlexicalized model and a first-order lexicalized model. Experiments on multiple datasets show the effectiveness of our second-order models compared with recent state-of-the-art methods. Our joint model achieves a 10% improvement over the previous state-of-the-art parser on the full WSJ test set | ['Kewei Tu', 'Wenjuan Han', 'Yong Jiang', 'Songlin Yang'] | 2020-10-28 | null | https://aclanthology.org/2020.coling-main.347 | https://aclanthology.org/2020.coling-main.347.pdf | coling-2020-8 | ['dependency-grammar-induction'] | ['natural-language-processing'] | [-1.22494586e-01 6.72317147e-01 -1.46558613e-01 -9.53838468e-01
-9.88953590e-01 -5.45731306e-01 2.40025267e-01 3.50297570e-01
-5.63328505e-01 7.32275367e-01 3.38628501e-01 -5.79979062e-01
1.93219781e-01 -8.91415775e-01 -6.81377232e-01 -5.82678676e-01
-5.45306131e-02 8.15849185e-01 4.70002085e-01 -9.69952568e-02
-1.94644868e-01 1.61441881e-02 -1.05427539e+00 -1.52944056e-02
1.18812084e+00 3.00590575e-01 4.62760448e-01 4.50903386e-01
-4.61300164e-01 6.14974737e-01 -5.29744387e-01 -7.31906414e-01
-9.91888344e-02 -3.44217807e-01 -6.62212372e-01 -3.93723100e-01
2.70579994e-01 -3.17987621e-01 -1.26921624e-01 1.03929698e+00
3.27788860e-01 -1.37351863e-02 4.60254788e-01 -8.14658403e-01
-8.17460179e-01 1.47863162e+00 -3.40374708e-01 2.95737326e-01
6.16934076e-02 -4.56625193e-01 1.44333243e+00 -6.93738580e-01
5.12567878e-01 1.41758573e+00 3.95526141e-01 5.93138576e-01
-1.34639394e+00 -8.15796494e-01 7.37663805e-01 -1.09701529e-01
-1.11111927e+00 -2.62227654e-01 9.08851206e-01 -1.32682666e-01
1.48480654e+00 -2.50405908e-01 3.43776673e-01 1.01165438e+00
7.33789131e-02 8.18515718e-01 1.05808115e+00 -7.23904312e-01
1.39608890e-01 -3.43224376e-01 7.78680384e-01 8.51396620e-01
5.03474474e-01 2.44316924e-02 -5.32851040e-01 -1.10748418e-01
6.24310374e-01 -6.55080259e-01 3.80910605e-01 -1.41105160e-01
-8.48023891e-01 1.33266413e+00 7.61728510e-02 3.73713672e-01
-1.32073283e-01 -1.49011593e-02 2.10710347e-01 -2.76966747e-02
7.32524872e-01 2.53106475e-01 -8.71231139e-01 -1.16793014e-01
-1.00613964e+00 -7.52267689e-02 1.13341415e+00 1.19697523e+00
6.98363483e-01 -2.91630672e-03 3.57047021e-02 9.47591484e-01
5.71120977e-01 4.26521689e-01 2.98883587e-01 -6.39656246e-01
8.30012321e-01 4.63148147e-01 -4.05758679e-01 -3.48952591e-01
-6.91183388e-01 -4.30977434e-01 -6.35396779e-01 -4.36871290e-01
3.86993438e-01 -2.00437278e-01 -1.22177207e+00 2.18196368e+00
4.03289139e-01 -1.40530676e-01 1.60510927e-01 4.44201171e-01
9.66631770e-01 6.16147757e-01 5.66592336e-01 -4.26511586e-01
1.43358195e+00 -9.25501227e-01 -7.25104153e-01 -4.18623209e-01
8.01813483e-01 -6.04604781e-01 7.98840642e-01 1.80460930e-01
-1.22581267e+00 -5.10401070e-01 -1.05137336e+00 -2.25164518e-01
-1.51830778e-01 5.13619259e-02 9.72262561e-01 8.00160766e-01
-9.93842542e-01 4.60253745e-01 -1.17417657e+00 -3.40034664e-01
7.16125071e-02 4.72679287e-01 -4.78980452e-01 7.43416771e-02
-1.35872614e+00 8.23654115e-01 7.96949685e-01 1.21966479e-02
-5.03183007e-01 -4.83307034e-01 -1.14634812e+00 2.64998734e-01
3.29892457e-01 -3.51869673e-01 1.28211355e+00 -3.60907525e-01
-1.75071943e+00 8.16079140e-01 -4.68915850e-01 -9.44524035e-02
-1.04683809e-01 -4.23970222e-01 -1.44246981e-01 -1.49755210e-01
3.19437802e-01 5.97432852e-01 2.99742758e-01 -9.60378349e-01
-7.41237164e-01 -3.32103997e-01 2.45385155e-01 1.27998024e-01
-2.61663646e-01 2.40253538e-01 -5.40206671e-01 -7.43893027e-01
5.46693563e-01 -8.24860752e-01 -3.40294003e-01 -9.27256942e-01
-5.83292663e-01 -8.27437997e-01 2.46950507e-01 -7.91621447e-01
1.39961946e+00 -1.84726226e+00 2.78191000e-01 -1.22879319e-01
3.61061394e-02 1.46999747e-01 -4.20876533e-01 3.74168336e-01
-2.02798974e-02 3.96727979e-01 -3.19969803e-01 -7.67479002e-01
1.39862970e-01 8.09488475e-01 -3.15418057e-02 2.42608320e-02
4.92555797e-01 8.28714967e-01 -9.27511334e-01 -7.35300779e-01
-1.80514842e-01 2.72309661e-01 -6.07118130e-01 5.32706320e-01
-1.82025149e-01 4.13347214e-01 -4.18755203e-01 5.01759827e-01
7.41087854e-01 -6.46516308e-02 7.35059381e-01 7.09184706e-02
-7.86968768e-02 1.10587132e+00 -9.00522768e-01 1.77819264e+00
-5.23292422e-01 -3.47035490e-02 -5.53287454e-02 -1.03729308e+00
8.05579424e-01 3.81499946e-01 -3.08409915e-03 -4.95362043e-01
6.91304356e-02 2.60582119e-01 3.11005503e-01 -2.29491502e-01
8.61383602e-02 -2.28324279e-01 -6.56491518e-01 5.16462445e-01
6.57115042e-01 -2.17207298e-02 7.07646966e-01 2.80091882e-01
1.17483032e+00 3.54447991e-01 4.32302892e-01 -4.09247607e-01
3.21575403e-01 -4.14438218e-01 1.09645951e+00 7.61762083e-01
1.66563436e-01 3.07656586e-01 7.51811504e-01 -3.87746304e-01
-7.83248603e-01 -1.26617420e+00 -9.00997296e-02 1.68926799e+00
-1.43869415e-01 -6.20459080e-01 -9.68818188e-01 -1.05220628e+00
-3.64842802e-01 1.03547859e+00 -4.83559847e-01 2.16321468e-01
-1.28371775e+00 -1.08756137e+00 7.08774209e-01 9.28763449e-01
2.66115099e-01 -1.22910833e+00 -2.14242980e-01 6.80330098e-01
-2.83297718e-01 -1.52998006e+00 -2.16762424e-01 8.07172894e-01
-9.39821422e-01 -7.71726072e-01 -2.35178798e-01 -1.09679043e+00
6.69346511e-01 -4.60096955e-01 1.53356588e+00 9.64248776e-02
2.50186563e-01 -1.81673497e-01 -6.57024264e-01 -3.91811907e-01
-8.24710011e-01 6.65198386e-01 -9.21918154e-02 -6.41911685e-01
7.18513191e-01 -9.12516177e-01 -3.87969171e-03 -2.89207935e-01
-8.01662624e-01 4.04848121e-02 8.66429508e-01 9.57458019e-01
4.29071188e-01 -2.67303437e-01 5.15386522e-01 -1.23174691e+00
5.13141096e-01 -3.73969704e-01 -7.98494697e-01 3.84062916e-01
-4.98143196e-01 6.13929510e-01 7.19694734e-01 -4.28661197e-01
-1.35658324e+00 2.65101820e-01 -4.55266535e-01 3.41070324e-01
-2.87259281e-01 7.33047009e-01 -5.91902018e-01 5.95768571e-01
4.82142828e-02 8.90358314e-02 -6.22181118e-01 -7.60959744e-01
6.46208823e-01 4.02624905e-01 4.73273754e-01 -8.83504152e-01
7.19650984e-01 -7.09671080e-02 -1.51153594e-01 -6.42995834e-01
-1.19645584e+00 -2.63779789e-01 -1.28998959e+00 4.78709131e-01
1.00774658e+00 -7.47220218e-01 -3.60148132e-01 4.01042134e-01
-1.75209022e+00 -2.51472265e-01 9.00557116e-02 5.96432209e-01
-2.73501426e-01 5.82002819e-01 -1.00098538e+00 -6.75557137e-01
-3.45502585e-01 -9.87774193e-01 1.00951159e+00 1.32531092e-01
-1.24998175e-01 -9.48707700e-01 3.29177290e-01 1.53296053e-01
5.07997088e-02 -2.51504779e-01 1.52026641e+00 -1.09427691e+00
-2.75591761e-01 1.13672828e-02 -2.86332797e-02 4.14576620e-01
3.19306590e-02 -2.37636685e-01 -6.04459524e-01 1.16614625e-03
-5.29698916e-02 -3.63979250e-01 8.79008770e-01 3.21306527e-01
5.89222074e-01 -1.26938343e-01 -3.67756844e-01 5.67685544e-01
1.10018468e+00 7.05213696e-02 2.82270104e-01 7.07640126e-02
8.20994675e-01 6.08893096e-01 4.56703365e-01 6.04223460e-02
9.46613789e-01 4.03507262e-01 2.61112154e-01 -3.19786519e-02
-9.58284480e-04 -5.22273898e-01 3.14842194e-01 1.64925587e+00
-1.32127643e-01 -3.08168530e-01 -8.55488539e-01 7.19459593e-01
-1.79130518e+00 -4.15602684e-01 -1.61258280e-01 1.94453692e+00
1.25302947e+00 3.75205100e-01 -9.52567682e-02 -2.60254055e-01
6.74279511e-01 1.40053228e-01 -1.84421301e-01 -6.65858984e-01
-1.01629518e-01 7.00460255e-01 3.64379376e-01 6.41463876e-01
-1.32327330e+00 1.73610568e+00 6.58882666e+00 5.87928176e-01
-7.21994340e-01 4.72135425e-01 4.35650080e-01 1.41429111e-01
-3.70059520e-01 4.12826896e-01 -1.43927765e+00 2.84399003e-01
1.18163908e+00 2.63599664e-01 1.78306043e-01 7.84362078e-01
-2.41079211e-01 -1.90052480e-01 -1.28856838e+00 3.18975091e-01
5.90245659e-03 -6.26483738e-01 -1.81499973e-01 5.91782220e-02
5.23620963e-01 3.12568784e-01 -5.75857818e-01 7.04578996e-01
8.72042716e-01 -8.30889761e-01 6.18712425e-01 -1.11790352e-01
5.81243098e-01 -5.95066130e-01 7.29891658e-01 6.67820513e-01
-1.20365250e+00 1.95023671e-01 -6.32283390e-01 -1.21993378e-01
5.85163295e-01 7.36047804e-01 -3.92510086e-01 4.83122915e-01
6.22846544e-01 1.33899659e-01 -4.23993140e-01 4.48441267e-01
-1.24657738e+00 1.15978646e+00 -6.64124429e-01 -2.96963364e-01
2.35304624e-01 -4.81211133e-02 4.75090325e-01 1.39534032e+00
1.28411531e-01 3.84219170e-01 2.88930088e-01 6.50190353e-01
-7.88592398e-02 1.94477707e-01 -1.65754527e-01 -7.47628585e-02
7.07184434e-01 1.15366423e+00 -9.10115302e-01 -3.20559382e-01
-6.34460390e-01 8.23071063e-01 1.10635853e+00 1.86793789e-01
-7.34136939e-01 -2.22190097e-01 2.89586961e-01 -1.83443516e-01
7.16185570e-01 -7.10347831e-01 -2.33966142e-01 -1.18558276e+00
2.33142361e-01 -5.18604875e-01 4.46809173e-01 -4.77277972e-02
-1.52032793e+00 9.17986631e-01 3.33554536e-01 -3.45907241e-01
-6.18327618e-01 -7.23591328e-01 -4.45312530e-01 8.56558919e-01
-1.57088673e+00 -1.54283977e+00 3.95156980e-01 9.74881873e-02
6.44086301e-01 7.56843062e-03 1.41096199e+00 -7.22027943e-03
-5.77641666e-01 8.42281103e-01 -1.56426087e-01 5.25723994e-01
5.34839928e-01 -1.43399572e+00 8.61541390e-01 1.04421914e+00
4.49849635e-01 7.50533938e-01 4.50021237e-01 -7.95680344e-01
-8.78051341e-01 -8.10075819e-01 1.56014884e+00 -4.10681635e-01
7.37414002e-01 -9.27150190e-01 -9.76805747e-01 9.82683957e-01
2.74644315e-01 9.38281068e-04 9.48899448e-01 8.84740770e-01
-5.99773049e-01 9.60803255e-02 -6.72113776e-01 3.75279129e-01
1.27759898e+00 -1.84761584e-01 -1.10122621e+00 1.63912177e-01
9.04621601e-01 -3.11303020e-01 -8.36119473e-01 5.70939839e-01
4.89497542e-01 -7.75117874e-01 4.03793275e-01 -5.71726501e-01
2.75822431e-01 2.26794794e-01 -1.04100168e-01 -1.36884439e+00
-6.88410342e-01 -5.03512204e-01 -1.86675623e-01 1.54580331e+00
7.83924639e-01 -7.24298120e-01 6.90035582e-01 6.73631728e-01
-2.31306285e-01 -6.79328978e-01 -1.23702705e+00 -8.65552902e-01
6.44100547e-01 -4.53005850e-01 4.82957482e-01 6.91809773e-01
1.95588753e-01 8.22385550e-01 -5.28874956e-02 4.65081513e-01
6.92581117e-01 7.14117289e-02 4.60011065e-01 -1.51008725e+00
-5.58790505e-01 -1.44618481e-01 -2.41132438e-01 -1.29581702e+00
6.75152957e-01 -9.00646210e-01 5.20189762e-01 -1.58048999e+00
3.37816745e-01 -5.59934795e-01 -4.56444114e-01 7.29308784e-01
-5.85716188e-01 -1.72467530e-01 4.74580526e-02 -7.16282651e-02
-4.39633518e-01 3.91875148e-01 6.75584674e-01 2.37414896e-01
-2.53516018e-01 -2.08584920e-01 -7.70063996e-01 9.34601247e-01
7.54611433e-01 -1.08682132e+00 -2.90178299e-01 -8.92438471e-01
2.16686249e-01 1.41413799e-02 -2.59889543e-01 -6.68250859e-01
2.76596695e-01 -4.40704487e-02 6.61105476e-03 -7.86233068e-01
1.30686924e-01 -3.02825540e-01 -4.64248359e-01 4.62928787e-02
-2.19331861e-01 1.42788917e-01 8.27629492e-02 2.89118528e-01
-1.82008132e-01 -4.99721080e-01 4.87689912e-01 -1.28472045e-01
-4.16663051e-01 8.95485207e-02 -4.13244247e-01 1.92079008e-01
6.12405479e-01 2.91682214e-01 -2.76232392e-01 -6.37106970e-02
-8.67444396e-01 1.73326284e-01 -9.47161578e-03 5.31871676e-01
2.47387603e-01 -9.16740537e-01 -9.26620960e-01 3.73396248e-01
-9.44837108e-02 2.94704199e-01 -6.35553598e-02 4.37945962e-01
-2.63678342e-01 5.43327451e-01 2.13618621e-01 -3.63655955e-01
-1.10812962e+00 3.67921978e-01 -2.34593824e-01 -1.01366186e+00
-3.83756727e-01 1.21511364e+00 3.94968688e-01 -8.61774981e-01
1.19647011e-01 -5.12695193e-01 -2.23862484e-01 -2.26004884e-01
9.33238566e-02 -1.18148260e-01 -2.57693212e-02 -7.57548392e-01
-5.17268240e-01 4.49863732e-01 -4.54206824e-01 -2.94823021e-01
1.49487507e+00 1.13634445e-01 -4.20766860e-01 4.08237100e-01
8.62045050e-01 7.17282295e-02 -8.82810712e-01 -2.71904528e-01
3.69064003e-01 3.05235445e-01 -8.54751095e-02 -8.28553975e-01
-6.99508250e-01 9.50442433e-01 -6.26920238e-02 7.75032640e-02
1.00330245e+00 4.87249970e-01 8.35349441e-01 4.88250226e-01
4.36259925e-01 -1.24790430e+00 -3.06243002e-01 1.04686415e+00
4.43811923e-01 -1.21986949e+00 -1.71598122e-01 -1.05001926e+00
-3.22101235e-01 1.02860308e+00 6.60286069e-01 -2.29914665e-01
9.29090619e-01 6.75284624e-01 1.79925457e-01 3.11317295e-02
-8.46077144e-01 -2.60196805e-01 2.08048701e-01 5.03001332e-01
9.38802063e-01 3.20888638e-01 -9.49115098e-01 1.35546207e+00
-4.81087059e-01 -4.46501195e-01 8.88562500e-02 1.03491139e+00
-4.30389911e-01 -1.84073734e+00 3.38741168e-02 2.04055130e-01
-7.23711610e-01 -7.25727320e-01 -1.56409904e-01 9.41068947e-01
8.46934021e-02 1.18438065e+00 6.85956031e-02 -2.95242667e-02
1.61732957e-01 4.23815697e-01 7.43807077e-01 -1.16005003e+00
-3.43793213e-01 3.67078215e-01 2.90286094e-01 -3.75499487e-01
-4.60048229e-01 -5.32990813e-01 -1.49825394e+00 2.60282516e-01
-5.58354676e-01 3.07300210e-01 8.05428684e-01 1.31450915e+00
7.43220299e-02 3.56272012e-01 3.04773659e-01 -6.75365806e-01
-6.94802821e-01 -1.41832435e+00 -5.71572840e-01 3.71352769e-02
-2.09798381e-01 -6.25938535e-01 -2.41707608e-01 6.79118931e-02] | [10.320820808410645, 9.666071891784668] |
3bd6affc-5f02-4d00-bede-a6a643a6220f | balancing-accuracy-and-integrity-for | 2207.08057 | null | https://arxiv.org/abs/2207.08057v1 | https://arxiv.org/pdf/2207.08057v1.pdf | Balancing Accuracy and Integrity for Reconfigurable Intelligent Surface-aided Over-the-Air Federated Learning | Over-the-air federated learning (AirFL) allows devices to train a learning model in parallel and synchronize their local models using over-the-air computation. The integrity of AirFL is vulnerable due to the obscurity of the local models aggregated over-the-air. This paper presents a novel framework to balance the accuracy and integrity of AirFL, where multi-antenna devices and base station (BS) are jointly optimized with a reconfigurable intelligent surface (RIS). The key contributions include a new and non-trivial problem jointly considering the model accuracy and integrity of AirFL, and a new framework that transforms the problem into tractable subproblems. Under perfect channel state information (CSI), the new framework minimizes the aggregated model's distortion and retains the local models' recoverability by optimizing the transmit beamformers of the devices, the receive beamformers of the BS, and the RIS configuration in an alternating manner. Under imperfect CSI, the new framework delivers a robust design of the beamformers and RIS configuration to combat non-negligible channel estimation errors. As corroborated experimentally, the novel framework can achieve comparable accuracy to the ideal FL while preserving local model recoverability under perfect CSI, and improve the accuracy when the number of receive antennas is small or moderate under imperfect CSI. | ['Ping Zhang', 'Wei Ni', 'Wanli Ni', 'Hui Tian', 'Jingheng Zheng'] | 2022-07-17 | null | null | null | null | ['robust-design'] | ['miscellaneous'] | [ 1.90077230e-01 3.06963742e-01 1.40096262e-01 1.90270200e-01
-1.10788751e+00 -9.78691876e-01 1.29830256e-01 -3.45661879e-01
1.59015268e-01 5.26263654e-01 -1.72273070e-01 -3.22099626e-01
-7.57285893e-01 -9.62350070e-01 -1.23817039e+00 -1.20099545e+00
-3.83282661e-01 1.79474548e-01 -3.37637246e-01 7.72077963e-02
-2.13922903e-01 6.35793865e-01 -1.51831949e+00 1.26154587e-01
4.99634206e-01 1.59293008e+00 4.44057994e-02 9.97696221e-01
4.00135964e-01 7.98846602e-01 -6.31179094e-01 -2.76120147e-03
8.21415424e-01 6.69457540e-02 -2.70130724e-01 -2.17102826e-01
4.31428641e-01 -3.18830103e-01 -4.49710131e-01 8.52591455e-01
8.72986674e-01 -1.91584662e-01 3.51827703e-02 -9.99254346e-01
7.83447027e-02 3.24992985e-01 1.28614187e-01 -5.51470876e-01
1.88144609e-01 7.04059452e-02 8.15171540e-01 -4.45838362e-01
7.81109706e-02 7.87178278e-01 1.11497855e+00 -1.61113031e-02
-1.04780018e+00 -1.00053525e+00 1.08801350e-01 -1.84143782e-01
-1.87602890e+00 -8.00878167e-01 4.64902818e-01 -1.44541249e-01
4.82497334e-01 5.97256184e-01 4.22314435e-01 6.74629509e-01
5.90513647e-01 3.35524738e-01 7.44245887e-01 -6.09018922e-01
6.13092303e-01 8.19619000e-02 -1.67039007e-01 7.81276643e-01
5.77256799e-01 4.87956196e-01 -7.25358844e-01 -5.50716221e-01
7.79092252e-01 -1.32492676e-01 -2.23643765e-01 -5.24814546e-01
-1.01516151e+00 2.23636284e-01 3.10588807e-01 2.33157784e-01
-5.30137539e-01 3.77664506e-01 -2.35728845e-01 5.34752846e-01
2.59584524e-02 4.11115438e-01 -7.78698444e-01 3.05075645e-01
-8.27393413e-01 1.39275759e-01 1.05565357e+00 1.25972748e+00
7.83945441e-01 2.71154374e-01 -6.04845062e-02 5.97308017e-02
5.61514318e-01 1.21156931e+00 -1.67341828e-01 -8.88372302e-01
6.15266860e-01 2.77678937e-01 4.50904667e-01 -1.26957834e+00
-5.36214113e-01 -1.58704889e+00 -9.42203820e-01 -2.10841686e-01
1.53174236e-01 -6.99906886e-01 -1.46935731e-01 1.75491774e+00
5.70781708e-01 5.17395854e-01 2.88739920e-01 5.56520104e-01
1.61714286e-01 5.77760160e-01 -6.96164548e-01 -3.74495119e-01
7.32970476e-01 -7.34569311e-01 -8.07470441e-01 -3.42944205e-01
8.59107375e-01 -6.32038116e-01 4.02217954e-01 7.21163929e-01
-1.19864297e+00 -5.12762070e-01 -1.44919801e+00 6.91722333e-01
1.04916580e-01 4.72033024e-01 4.99976844e-01 9.58261728e-01
-1.16689694e+00 1.63466349e-01 -7.52703130e-01 2.25655064e-01
1.93311721e-01 1.03445005e+00 -1.01787493e-01 -1.63606465e-01
-8.86889338e-01 5.90745568e-01 -1.05008274e-01 4.13813800e-01
-1.07507443e+00 -9.13502276e-01 -5.86437583e-01 1.60346255e-02
2.05004230e-01 -8.19824457e-01 1.15360999e+00 -9.66810703e-01
-1.77286911e+00 -7.76065793e-03 -6.95462618e-03 -4.88090724e-01
1.72047779e-01 -1.13453932e-01 -6.70582950e-01 -1.74390420e-01
-3.25984776e-01 -5.04610240e-01 6.98324263e-01 -1.44781566e+00
-6.21238470e-01 -6.21710062e-01 7.32015818e-02 -1.08717658e-01
-3.46731603e-01 -7.88559258e-01 -3.16420913e-01 -5.37334681e-01
4.15267229e-01 -1.23882520e+00 -2.60253787e-01 1.39052272e-01
-2.87689507e-01 6.32264137e-01 6.91651821e-01 -5.67697942e-01
1.27490187e+00 -2.09008503e+00 -6.58028275e-02 7.69909143e-01
-7.33560026e-02 5.08874618e-02 -2.05274701e-01 5.12218118e-01
2.97933877e-01 -4.23321903e-01 3.05255353e-01 -4.21426654e-01
-1.49951624e-02 3.18170875e-01 -4.71151710e-01 9.09252524e-01
-6.74635589e-01 6.86651826e-01 -6.16540074e-01 5.58161996e-02
1.33365482e-01 6.24411285e-01 -6.12362862e-01 3.84296149e-01
1.66131288e-01 6.58537924e-01 -7.78606772e-01 6.22370243e-01
6.43532276e-01 -3.42117697e-01 5.94149172e-01 -7.26189971e-01
-1.58600524e-01 2.10889559e-02 -1.70784807e+00 1.60404384e+00
-9.61709201e-01 2.18964964e-01 8.40841770e-01 -6.72378421e-01
8.92228305e-01 4.60880697e-01 7.12662041e-01 -1.01235926e+00
3.64961416e-01 3.34279895e-01 -4.34914798e-01 -6.79715723e-02
-1.81953348e-02 6.88859448e-02 -3.40061069e-01 6.03713751e-01
9.49558169e-02 1.12652726e-01 -9.73454654e-01 -1.45436833e-02
1.33545876e+00 -1.83277190e-01 1.67625278e-01 -4.53487933e-01
5.88991940e-01 -6.09001935e-01 4.61368084e-01 1.06165469e+00
4.95217919e-01 1.48701102e-01 -4.37864512e-01 -3.51029962e-01
-5.16334891e-01 -1.07168412e+00 4.15537953e-02 7.92017937e-01
4.92163658e-01 -3.57007325e-01 -5.36120653e-01 -4.30144101e-01
1.25798136e-01 6.46006286e-01 -4.75958169e-01 -4.21937436e-01
-2.93023199e-01 -4.72754449e-01 7.29927957e-01 2.00567111e-01
6.12332404e-01 3.16187114e-01 -3.93550217e-01 2.23728150e-01
5.63290678e-02 -9.32960510e-01 -1.81044191e-01 1.83408484e-01
-5.51747501e-01 -8.81731272e-01 -8.92563090e-02 -4.27861214e-01
7.53115833e-01 3.88213068e-01 5.91452539e-01 8.76313522e-02
1.80906102e-01 8.49846303e-01 -1.27971753e-01 -4.07365233e-01
-1.00927025e-01 -3.08985002e-02 3.90639484e-01 5.85278988e-01
-8.15662980e-01 -4.97527301e-01 -2.38621131e-01 6.97650075e-01
-5.55722773e-01 -1.76009566e-01 4.44944113e-01 6.26701832e-01
7.99391031e-01 5.19070089e-01 2.38902539e-01 -4.65367705e-01
-2.28415951e-01 -4.88087565e-01 -9.28149104e-01 5.85929155e-01
-5.37927091e-01 2.85278022e-01 8.23280096e-01 -1.19231090e-01
-1.04742646e+00 3.28077078e-01 1.74951956e-01 -1.65101528e-01
3.83557856e-01 6.14016578e-02 -8.41222167e-01 -1.19662166e+00
4.76335853e-01 4.70231511e-02 -2.06263736e-01 -5.14050484e-01
1.20511428e-01 8.48193109e-01 5.37329495e-01 -6.43062890e-01
1.04220712e+00 6.91958368e-01 5.07057488e-01 -6.78362727e-01
-1.07198668e+00 -1.68129921e-01 -4.43900734e-01 -4.54452634e-01
4.49141935e-02 -1.24285948e+00 -9.52519000e-01 5.17761707e-01
-9.32232380e-01 -3.74519318e-01 -2.61029571e-01 4.41764683e-01
-4.72873211e-01 -1.10931527e-02 2.92006172e-02 -1.08257258e+00
-4.41597670e-01 -9.04130995e-01 1.32781720e+00 -8.68694708e-02
2.65485018e-01 -1.08608663e+00 -4.26728874e-02 2.67168462e-01
8.04992855e-01 4.48247522e-01 5.15785336e-01 -2.10532576e-01
-9.26178813e-01 -5.44269919e-01 3.89568627e-01 1.26377121e-01
4.60112654e-02 -6.73778832e-01 -1.17627800e+00 -9.05380726e-01
2.50592977e-01 2.06667081e-01 7.30471686e-02 2.06796989e-01
1.07087338e+00 -9.47963476e-01 -4.40693468e-01 1.13348758e+00
1.55890870e+00 -7.20029790e-03 4.55579311e-01 5.85349649e-02
4.22998399e-01 1.29257321e-01 5.26692688e-01 7.69127667e-01
4.51121241e-01 9.41088796e-01 7.10073769e-01 1.98590904e-02
1.10990240e-03 -1.30185932e-01 4.98408526e-01 1.00519407e+00
3.71345431e-01 -4.19428855e-01 -4.28174704e-01 1.41748473e-01
-1.73875570e+00 -4.35298443e-01 5.84858209e-02 2.78363919e+00
2.96932518e-01 -3.27784151e-01 -4.72635269e-01 2.03000292e-01
4.25727725e-01 -1.31916732e-01 -6.78133547e-01 1.38819264e-02
-3.44775707e-01 1.76376656e-01 1.26391065e+00 6.86612427e-01
-7.96441615e-01 2.62063384e-01 6.27772903e+00 4.97342169e-01
-1.14421761e+00 4.19679314e-01 4.66185622e-02 -1.94644421e-01
-4.38791275e-01 5.35659492e-02 -7.47408330e-01 2.90011048e-01
1.20932603e+00 1.64485916e-01 7.34025836e-01 6.62775338e-01
-3.76516655e-02 2.41327018e-01 -1.01067173e+00 1.16770220e+00
1.84922546e-01 -1.45323145e+00 -2.43728936e-01 2.89280206e-01
8.73182297e-01 4.78535518e-02 1.34641334e-01 -6.53613657e-02
3.91903996e-01 -5.43819487e-01 1.10307193e+00 1.09870279e+00
8.76894534e-01 -8.50979388e-01 7.60683596e-01 4.67820764e-01
-1.36838019e+00 -6.92827404e-01 1.23089842e-01 -2.29244456e-01
-8.42669234e-02 7.51174986e-01 -5.55043995e-01 1.10438704e+00
6.39094651e-01 2.24958822e-01 -4.13388729e-01 9.91134226e-01
1.47110015e-01 6.43368900e-01 -6.33233428e-01 2.90756613e-01
-1.61806136e-01 2.21042246e-01 6.84182286e-01 5.53497314e-01
6.34751737e-01 1.80565238e-01 2.26856947e-01 2.98893601e-01
3.17519665e-01 -3.04012984e-01 -3.54907155e-01 3.97510052e-01
9.10492480e-01 8.51277888e-01 1.19418919e-01 1.27241567e-01
-4.66924459e-02 4.18121755e-01 -2.13997468e-01 3.82240504e-01
-6.05077326e-01 -9.17780772e-02 6.85266078e-01 1.53050184e-01
1.27101660e-01 -4.93317366e-01 -4.58384186e-01 -7.98798680e-01
3.29871178e-02 -9.28908587e-01 2.22511873e-01 -4.69357491e-01
-8.92209589e-01 3.63460511e-01 -5.16934812e-01 -1.39553690e+00
-2.21454337e-01 -3.63544971e-01 -6.05079234e-02 5.15899837e-01
-1.23376727e+00 -1.44475675e+00 -3.18605959e-01 7.68372595e-01
-3.21823597e-01 -3.67084086e-01 1.13096154e+00 4.86986071e-01
-5.10928214e-01 1.06065488e+00 6.60692513e-01 -5.81329823e-01
3.38779628e-01 -4.71116990e-01 -1.87581852e-01 8.17902625e-01
-1.51814044e-01 5.20102441e-01 7.82966852e-01 -4.32505518e-01
-2.55897498e+00 -1.56600165e+00 4.58145261e-01 -2.50720769e-01
3.63128781e-01 -6.10469043e-01 -2.19277933e-01 6.49052262e-01
-2.89214909e-01 1.84201732e-01 8.32107604e-01 -1.75678059e-01
-1.26778051e-01 -7.44284809e-01 -1.22985423e+00 2.21001282e-01
8.91247511e-01 -5.65960884e-01 2.28874087e-01 6.47666514e-01
5.59802353e-01 -6.28754675e-01 -9.41534340e-01 5.08209944e-01
8.05103302e-01 -5.07014573e-01 9.70958591e-01 -6.25635777e-03
-9.46688592e-01 -6.61391318e-01 -9.43603456e-01 -8.85217130e-01
-2.26932973e-01 -1.25371516e+00 -6.31643713e-01 1.03110635e+00
1.62838623e-01 -8.55771065e-01 6.79057598e-01 3.08165520e-01
-2.74676025e-01 -4.60933924e-01 -1.28119159e+00 -8.08066308e-01
-5.97312152e-01 -8.07051063e-01 1.12867987e+00 5.39959311e-01
-6.17992699e-01 -8.69838148e-02 -7.03987896e-01 1.22985530e+00
9.37210977e-01 -9.74396542e-02 1.04925013e+00 -1.30995917e+00
-9.35524166e-01 3.34718585e-01 -2.03243941e-01 -1.21493936e+00
-5.52099794e-02 -7.08352447e-01 7.25835338e-02 -1.07000542e+00
-6.31624818e-01 -1.10133839e+00 -2.68516719e-01 5.82910061e-01
6.78816080e-01 -4.13410440e-02 1.42735913e-01 1.22538045e-01
-9.15000439e-01 4.99903768e-01 7.74847150e-01 -2.37186268e-01
-2.07287848e-01 5.73000491e-01 -7.13632643e-01 4.30258900e-01
5.74446738e-01 -2.82176882e-01 -3.21937233e-01 -9.26950037e-01
5.11461735e-01 2.19228342e-01 5.02843857e-01 -1.62637711e+00
8.44131887e-01 2.51463801e-01 2.23807096e-01 -2.26628989e-01
3.79003227e-01 -1.54169977e+00 1.02527320e+00 5.06945848e-01
-4.03615572e-02 -1.96848497e-01 2.67142355e-01 6.16568744e-01
2.99175054e-01 3.25003028e-01 4.97076511e-01 5.13577819e-01
1.30766742e-02 1.63049281e-01 -1.25240862e-01 -6.03962660e-01
1.02104545e+00 -1.16181791e-01 -2.54826933e-01 -5.43340921e-01
-3.65721583e-01 2.77775437e-01 3.64683330e-01 1.79970756e-01
6.05177283e-02 -1.14153254e+00 -1.29509434e-01 8.44752252e-01
-3.02327812e-01 7.40146711e-02 5.24880946e-01 7.71144331e-01
-1.91522874e-02 6.14782393e-01 3.19432110e-01 -6.54935956e-01
-9.13491547e-01 8.93926397e-02 8.61497104e-01 -1.48486838e-01
2.71003563e-02 7.90425658e-01 1.31023020e-01 -8.08832407e-01
6.12819374e-01 4.49226461e-02 5.63713312e-01 -4.52469617e-01
7.40130782e-01 5.98093331e-01 7.28430033e-01 -6.51249409e-01
-3.33605647e-01 8.70511532e-01 5.19268155e-01 2.18666196e-01
1.33587372e+00 -5.86347938e-01 8.10006931e-02 1.45129278e-01
1.02061307e+00 3.88207972e-01 -1.29243326e+00 -4.45457429e-01
-5.41355908e-01 -3.12182486e-01 7.08788574e-01 -9.58763123e-01
-1.36174369e+00 4.87981349e-01 9.75242078e-01 -1.37565866e-01
1.31256366e+00 -2.69911438e-01 8.21072817e-01 5.17724872e-01
1.39619172e+00 -8.17906201e-01 -2.66748101e-01 4.46390599e-01
5.54150045e-01 -2.06585005e-01 -5.36082201e-02 -1.97777256e-01
1.86333373e-01 8.80013645e-01 1.43606693e-01 7.76088312e-02
9.47630763e-01 9.52009380e-01 -1.79573819e-01 -3.14571597e-02
-5.54755330e-01 2.87956923e-01 2.30341181e-01 5.56011677e-01
-4.03463483e-01 1.89912245e-01 4.25803691e-01 1.37421358e+00
-2.29945540e-01 -1.08162634e-01 7.43199065e-02 1.09365666e+00
-3.26753795e-01 -9.33853924e-01 -7.68448770e-01 3.79182816e-01
-1.14747539e-01 3.35181922e-01 -4.49567921e-02 2.64921814e-01
5.43869615e-01 1.20301938e+00 -1.10711269e-01 -9.33360457e-01
6.43730760e-01 -5.32145858e-01 5.19603610e-01 -1.53752968e-01
-7.05732703e-01 5.82161844e-02 4.84819338e-02 -1.07550645e+00
-5.67449369e-02 -5.72698236e-01 -1.10268545e+00 -1.44575745e-01
-7.82107294e-01 2.97535568e-01 8.72804582e-01 1.17755926e+00
1.04310608e+00 7.05117166e-01 1.17803419e+00 -9.28580463e-01
-8.25165033e-01 -1.27366677e-01 -7.77231574e-01 -7.00034618e-01
7.31497347e-01 -6.69703841e-01 -5.26135266e-01 -4.50642586e-01] | [6.314572811126709, 1.0774126052856445] |
54a15aa9-02eb-4289-898c-ceb1cb124341 | learning-true-rate-distortion-optimization | 2201.01586 | null | https://arxiv.org/abs/2201.01586v1 | https://arxiv.org/pdf/2201.01586v1.pdf | Learning True Rate-Distortion-Optimization for End-To-End Image Compression | Even though rate-distortion optimization is a crucial part of traditional image and video compression, not many approaches exist which transfer this concept to end-to-end-trained image compression. Most frameworks contain static compression and decompression models which are fixed after training, so efficient rate-distortion optimization is not possible. In a previous work, we proposed RDONet, which enables an RDO approach comparable to adaptive block partitioning in HEVC. In this paper, we enhance the training by introducing low-complexity estimations of the RDO result into the training. Additionally, we propose fast and very fast RDO inference modes. With our novel training method, we achieve average rate savings of 19.6% in MS-SSIM over the previous RDONet model, which equals rate savings of 27.3% over a comparable conventional deep image coder. | ['André Kaup', 'Alexander Kopte', 'Kristian Fischer', 'Fabian Brand'] | 2022-01-05 | null | null | null | null | ['ms-ssim'] | ['computer-vision'] | [ 3.62491518e-01 -7.00240210e-02 -3.92025650e-01 -3.62829387e-01
-4.49232012e-01 -1.48472860e-01 4.63705093e-01 1.30233034e-01
-5.15842617e-01 3.37711900e-01 3.13223988e-01 -4.21713591e-01
8.32030401e-02 -8.13416600e-01 -7.55929530e-01 -3.97902906e-01
-1.11855447e-01 9.91648436e-02 2.99926847e-01 3.94915557e-03
4.59438652e-01 4.42040414e-01 -1.47803640e+00 3.64751339e-01
8.22453976e-01 1.25614309e+00 4.95275259e-01 1.09467673e+00
4.30898592e-02 1.13407838e+00 -5.80493748e-01 -4.82587934e-01
4.44228947e-01 -5.49875438e-01 -5.99212050e-01 1.37396827e-01
2.94723958e-01 -1.05921698e+00 -7.19853520e-01 8.36545527e-01
7.09257424e-01 1.38571551e-02 5.38506925e-01 -6.52891278e-01
-3.09988320e-01 5.80280483e-01 -5.67062318e-01 2.84727722e-01
-1.46355741e-02 8.61833319e-02 8.29187036e-01 -6.55654788e-01
5.12882888e-01 8.98025572e-01 5.26937187e-01 4.72485274e-01
-1.16248894e+00 -4.86989111e-01 -2.75409937e-01 6.39230132e-01
-1.55164468e+00 -8.33877146e-01 5.19796968e-01 -1.01755910e-01
1.23359847e+00 2.96998501e-01 5.77307165e-01 3.20546478e-01
3.15808058e-02 6.45038068e-01 6.57348037e-01 -4.72851753e-01
1.94210768e-01 -2.02802896e-01 -5.68911672e-01 2.89830804e-01
8.99260864e-02 1.61800981e-01 -2.73425788e-01 4.73018169e-01
9.58754420e-01 -8.33489820e-02 -3.34517241e-01 -1.72953397e-01
-1.07566249e+00 5.11844814e-01 4.43333536e-01 2.86817193e-01
-3.04628521e-01 4.80768114e-01 5.43921471e-01 3.20268184e-01
3.89478862e-01 1.50458932e-01 -1.70414746e-01 -4.43059951e-01
-1.73399341e+00 1.37939915e-01 6.39758885e-01 9.94957983e-01
5.39671898e-01 2.10216120e-01 -3.41141760e-01 9.31357384e-01
2.35421047e-01 2.57039577e-01 5.32216012e-01 -1.48938918e+00
6.50680304e-01 1.87684983e-01 -1.72944874e-01 -1.08720767e+00
4.10693586e-02 -6.10571146e-01 -1.08754051e+00 2.38752708e-01
-6.44142646e-03 2.97842652e-01 -8.35745275e-01 1.27352071e+00
-4.28676186e-03 2.03380525e-01 -3.56013812e-02 8.97149861e-01
4.17704523e-01 8.75345170e-01 -3.50747377e-01 -5.20919621e-01
9.34450626e-01 -1.18902004e+00 -9.33931410e-01 9.02970061e-02
5.55183113e-01 -8.84091318e-01 8.04568648e-01 6.75271809e-01
-1.63088369e+00 -7.66314864e-01 -1.44292498e+00 -5.27796507e-01
2.24760577e-01 1.35866120e-01 4.18998718e-01 7.11221039e-01
-1.21193743e+00 9.16338086e-01 -7.82198191e-01 8.09274912e-02
6.44892752e-01 5.05884826e-01 1.23309597e-01 -1.79943874e-01
-7.65613735e-01 5.91854990e-01 5.68806231e-01 -2.38303050e-01
-1.17220724e+00 -6.61440074e-01 -5.74375689e-01 4.37634706e-01
4.71950322e-01 -5.18630743e-01 1.28949308e+00 -7.17899084e-01
-1.84270835e+00 6.03531182e-01 -3.95983964e-01 -9.93536711e-01
7.06438124e-01 -4.24206793e-01 -2.58772373e-01 5.23324907e-01
-3.64023209e-01 8.42727065e-01 1.03451097e+00 -1.03594065e+00
-4.71942693e-01 1.45766765e-01 2.01751083e-01 1.68322086e-01
-4.70512509e-01 -3.71459424e-02 -9.30554211e-01 -9.23487365e-01
2.82868231e-03 -4.60888863e-01 -2.86061078e-01 3.48731607e-01
1.57867894e-02 3.66728932e-01 7.01111972e-01 -1.05252409e+00
1.87343097e+00 -2.19152713e+00 2.64582131e-02 4.67010066e-02
4.21360075e-01 6.58337355e-01 -3.20382379e-02 9.76243615e-02
-7.63100898e-03 2.39897162e-01 -3.77479017e-01 -3.77320707e-01
2.84359213e-02 1.50323793e-01 -1.33404493e-01 3.61338049e-01
2.74805557e-02 7.28337705e-01 -5.54667354e-01 -5.87935090e-01
5.00552535e-01 6.64096773e-01 -1.24049520e+00 3.94658953e-01
-1.07238837e-01 2.35628411e-01 1.10247023e-01 3.80036950e-01
9.03073668e-01 -2.76465207e-01 4.06315237e-01 -3.61017019e-01
-1.11543514e-01 5.29141903e-01 -9.76984322e-01 1.80419123e+00
-6.30808353e-01 8.44553530e-01 -1.84614599e-01 -1.16495025e+00
7.17196524e-01 3.59468907e-01 5.91735065e-01 -9.46268737e-01
3.65768433e-01 2.27726400e-01 1.05731085e-01 -2.50469327e-01
8.08524847e-01 1.69103250e-01 6.52969778e-01 3.58450949e-01
4.84463312e-02 -1.15233503e-01 4.14490223e-01 9.89907607e-02
1.06350493e+00 1.50189146e-01 3.90014917e-01 3.44765708e-02
7.16435373e-01 -5.09973288e-01 3.44773680e-01 5.42546570e-01
-7.84149617e-02 9.89719868e-01 4.17348504e-01 -2.05980271e-01
-1.66690934e+00 -9.52671349e-01 -3.20443213e-01 7.03827500e-01
1.65112149e-02 -8.67285311e-01 -1.04215920e+00 -2.20300779e-01
-5.56865633e-01 5.83654642e-01 -4.09030430e-02 -5.43004349e-02
-8.58818889e-01 -5.51340759e-01 5.11476874e-01 2.68541098e-01
1.05715430e+00 -5.65245986e-01 -7.14924872e-01 3.49632472e-01
-3.34121525e-01 -1.39775419e+00 -5.50703168e-01 -6.64678635e-03
-1.30237472e+00 -5.89133918e-01 -9.17964995e-01 -5.33873558e-01
3.86971772e-01 2.60123312e-01 1.09894776e+00 2.73239225e-01
-9.89474133e-02 -2.68821150e-01 -4.04870152e-01 -3.41801085e-02
-6.20985985e-01 2.25111708e-01 -4.11035895e-01 -1.39839098e-01
4.20684693e-04 -7.99767733e-01 -1.06733239e+00 3.75253469e-01
-1.06127274e+00 3.38984996e-01 6.51165187e-01 5.13736308e-01
8.00692737e-01 2.37089857e-01 7.03350753e-02 -4.92337227e-01
1.36163518e-01 -1.98168620e-01 -6.79951727e-01 4.92483117e-02
-1.03463590e+00 1.23057269e-01 7.14850605e-01 -3.07793915e-01
-8.15136373e-01 -1.86797798e-01 -5.60598910e-01 -6.51054263e-01
1.72898412e-01 2.91976959e-01 -1.98710620e-01 -1.00444324e-01
4.54491496e-01 3.73171657e-01 3.31621841e-02 -4.31178838e-01
2.01297954e-01 6.59660339e-01 5.29021263e-01 -1.60846338e-01
7.19414532e-01 3.42019528e-01 -9.18715075e-02 -6.37448192e-01
-3.84599119e-01 -2.46400282e-01 -5.23579895e-01 -1.91447169e-01
7.65951157e-01 -1.07042515e+00 -6.14035904e-01 2.74689555e-01
-1.03350925e+00 -5.18212736e-01 -4.77617681e-01 5.65671504e-01
-7.65272975e-01 7.04948604e-01 -6.63487554e-01 -3.85820001e-01
-3.78506929e-01 -1.24597931e+00 7.53280044e-01 -1.60714552e-01
2.71798912e-02 -7.48062909e-01 -2.52317697e-01 3.43020797e-01
8.36505830e-01 -1.55400991e-01 6.41326308e-01 -1.21068927e-02
-9.49007273e-01 3.90023626e-02 -5.54819584e-01 8.62837493e-01
-6.57743514e-02 -2.00849280e-01 -9.01965022e-01 -4.17017430e-01
2.66421497e-01 -9.49642360e-02 9.45913553e-01 6.76940143e-01
1.77204120e+00 -4.96194929e-01 -2.94614732e-02 1.24308240e+00
1.68751931e+00 3.42090100e-01 1.39474618e+00 3.57456863e-01
5.40132403e-01 1.44242853e-01 5.06205320e-01 7.36969233e-01
2.92308122e-01 9.44029808e-01 4.36507642e-01 9.31087270e-05
-6.32830620e-01 -1.08702399e-01 4.09309775e-01 1.20783794e+00
-5.25281131e-01 -4.63680387e-01 -5.49430072e-01 3.00128728e-01
-1.41122174e+00 -1.16454661e+00 1.31805733e-01 2.27976775e+00
1.03973067e+00 2.66680032e-01 -1.67209044e-01 6.66681767e-01
5.28414071e-01 5.00588953e-01 -4.82571423e-01 -5.26878595e-01
-2.63325702e-02 5.05665898e-01 8.58042240e-01 5.47094166e-01
-9.47875738e-01 5.85762203e-01 6.76148033e+00 1.10626543e+00
-1.17209208e+00 2.37908036e-01 7.09174454e-01 -3.64158452e-01
-2.45410904e-01 1.08825438e-01 -5.69893420e-01 3.80076975e-01
1.38960111e+00 -1.40853509e-01 7.22724080e-01 7.49498963e-01
4.35616463e-01 -6.02970831e-02 -1.06980419e+00 1.34352374e+00
1.54881343e-01 -1.43971658e+00 9.80499014e-02 4.71753716e-01
6.87483966e-01 -2.40865335e-01 5.26954420e-02 8.10003579e-02
-2.31842026e-01 -1.03795302e+00 7.92983413e-01 1.09216057e-01
1.21407211e+00 -7.74502337e-01 5.72740316e-01 1.87516704e-01
-1.12227559e+00 -3.75498571e-02 -6.14389896e-01 5.57908416e-02
3.58405411e-01 8.06522667e-01 -8.21200192e-01 5.24531305e-01
6.00687802e-01 9.21616912e-01 -5.13562024e-01 1.14528358e+00
-1.37203559e-02 7.27906048e-01 -1.25868261e-01 4.78649974e-01
-1.26067609e-01 3.28794681e-02 2.96617508e-01 1.26236916e+00
6.30900264e-01 -2.37336885e-02 -4.34862942e-01 5.13250053e-01
-3.46215218e-01 2.32267287e-02 -1.98238924e-01 1.50344307e-02
1.55408949e-01 8.47199023e-01 -5.17711639e-01 -5.60830534e-01
-3.41094971e-01 1.38082421e+00 -1.01986498e-01 1.68522656e-01
-1.20277381e+00 -3.94296438e-01 4.92337137e-01 3.83052617e-01
8.11762333e-01 -4.18105632e-01 -2.99632639e-01 -1.22526026e+00
1.07852779e-01 -9.58073556e-01 -6.86302185e-02 -5.45824409e-01
-4.30088162e-01 4.02347237e-01 -1.15756476e-02 -1.51633954e+00
-3.91836107e-01 -4.28127617e-01 -1.56268150e-01 5.46614230e-01
-1.89640653e+00 -6.70887530e-01 -3.61985415e-01 3.98380041e-01
7.48927176e-01 -6.19683303e-02 4.86001849e-01 9.92581666e-01
-3.34698498e-01 8.72133136e-01 3.06986094e-01 -1.35553507e-02
5.47054112e-01 -7.95094609e-01 4.28784996e-01 1.13842738e+00
-1.08392618e-03 4.75403309e-01 6.53711677e-01 -2.99691141e-01
-1.29685104e+00 -1.06826770e+00 8.75071406e-01 3.95207286e-01
2.29955167e-01 -1.31409541e-01 -1.00051510e+00 3.22590172e-01
4.57454920e-01 9.51105878e-02 4.78599578e-01 -4.69425708e-01
-3.21671367e-01 -3.80962580e-01 -1.09382367e+00 5.36865175e-01
1.14692986e+00 -4.94344801e-01 -1.23254210e-01 4.11411785e-02
8.82571936e-01 -5.73895991e-01 -9.65657651e-01 5.07183313e-01
6.44338071e-01 -1.39582491e+00 1.17083824e+00 1.88136995e-01
9.85602856e-01 -3.89815629e-01 -5.13697147e-01 -7.73239911e-01
-2.88578361e-01 -5.29495239e-01 -7.38121092e-01 1.10782814e+00
1.64409578e-02 -9.26058888e-02 7.85600722e-01 3.72151956e-02
-3.04816902e-01 -6.23982251e-01 -8.98180068e-01 -7.89895415e-01
-2.56515861e-01 -7.03594387e-01 6.25107586e-01 5.99455535e-01
-4.82462287e-01 -4.45760377e-02 -5.39217234e-01 -4.08624038e-02
6.09610677e-01 -2.15003163e-01 7.57328749e-01 -6.73914611e-01
-8.07507455e-01 -5.87215304e-01 -5.36221445e-01 -1.76074266e+00
-3.04920405e-01 -8.57127607e-01 -5.78266047e-02 -1.48611903e+00
1.93678409e-01 -2.67891735e-01 -2.34476477e-01 -3.66442814e-03
1.34400755e-01 4.51767325e-01 2.78226227e-01 3.33896428e-01
-5.21232486e-01 6.10734761e-01 1.34288418e+00 -1.32438883e-01
-1.46590158e-01 -2.83362806e-01 -3.73393118e-01 4.09288108e-01
8.36841702e-01 -3.04838926e-01 -6.16618276e-01 -8.84416342e-01
3.74041386e-02 1.08596779e-01 2.23199189e-01 -1.48133087e+00
7.62128532e-02 1.04264244e-01 3.25293809e-01 -7.05527484e-01
3.33574116e-01 -7.79452503e-01 2.38878444e-01 5.51282823e-01
-3.44302177e-01 -4.97940416e-03 7.62602389e-02 3.17235023e-01
-5.14574647e-01 -4.08456057e-01 1.02061653e+00 1.05568655e-02
-6.88735127e-01 3.31514865e-01 -2.20774174e-01 -1.19565934e-01
9.18031037e-01 -4.11893576e-01 -1.22717157e-01 -4.98686373e-01
-4.33965951e-01 -2.86696553e-01 5.89774489e-01 -1.49926879e-02
1.01593518e+00 -1.19193947e+00 -7.12157965e-01 1.57061145e-01
-3.04295212e-01 1.34413555e-01 3.71309102e-01 7.56337106e-01
-1.15805733e+00 5.31059325e-01 -2.31756106e-01 -5.87426007e-01
-1.21179616e+00 6.46312833e-01 2.24947497e-01 -4.55494851e-01
-6.64288640e-01 4.71960723e-01 -1.94956958e-02 2.36734807e-01
3.12607288e-01 -3.39025527e-01 -9.76042375e-02 -3.48183900e-01
8.20749402e-01 4.62518662e-01 1.86333299e-01 -5.17851830e-01
1.54479071e-01 5.45968950e-01 -2.58399919e-02 -2.32150882e-01
1.25805008e+00 -4.35233533e-01 -7.78099746e-02 1.00661874e-01
1.45176470e+00 -5.24392873e-02 -1.26786602e+00 -1.38068590e-02
-2.80041516e-01 -9.06351745e-01 4.61200655e-01 -3.32176685e-01
-1.34737122e+00 1.22021961e+00 9.52592969e-01 -1.27752557e-01
1.80367720e+00 -3.24165016e-01 1.06431448e+00 1.43291533e-01
2.28542864e-01 -9.70700204e-01 1.06307693e-01 3.19715172e-01
7.64294386e-01 -1.04751635e+00 4.03451174e-01 -3.82896334e-01
-2.19990313e-01 1.16063881e+00 2.59147376e-01 -3.36983502e-02
6.05459273e-01 3.96090508e-01 -4.09209937e-01 2.74809033e-01
-8.32260907e-01 -4.47009169e-02 2.17798620e-01 3.84651840e-01
6.38399065e-01 -3.34754527e-01 -6.94477618e-01 -6.76200241e-02
-2.54271746e-01 3.97582620e-01 5.19866705e-01 7.45308876e-01
-4.00609463e-01 -1.36710823e+00 -7.94442520e-02 3.65504026e-01
-8.08100462e-01 -3.80232692e-01 6.36950791e-01 3.62562478e-01
1.13248460e-01 7.96942472e-01 3.32940012e-01 -6.58134341e-01
8.36675689e-02 -4.34908599e-01 5.32076359e-01 -2.10606113e-01
-4.23769325e-01 3.03450108e-01 -1.18169449e-01 -7.12312639e-01
-6.25914037e-01 -3.45896572e-01 -1.11660469e+00 -7.20677972e-01
-1.43716216e-01 -2.64189452e-01 7.64431953e-01 6.43005967e-01
4.15251702e-01 7.87705541e-01 7.92625129e-01 -9.34267879e-01
-3.03865105e-01 -7.02437103e-01 -3.11306685e-01 2.52749711e-01
4.31257069e-01 -1.29332423e-01 -2.38711670e-01 4.50500131e-01] | [11.362932205200195, -1.5939404964447021] |
1f378824-c547-470d-a170-40d3a3489c05 | towards-safe-propofol-dosing-during-general | 2303.10180 | null | https://arxiv.org/abs/2303.10180v1 | https://arxiv.org/pdf/2303.10180v1.pdf | Towards Safe Propofol Dosing during General Anesthesia Using Deep Offline Reinforcement Learning | Automated anesthesia promises to enable more precise and personalized anesthetic administration and free anesthesiologists from repetitive tasks, allowing them to focus on the most critical aspects of a patient's surgical care. Current research has typically focused on creating simulated environments from which agents can learn. These approaches have demonstrated good experimental results, but are still far from clinical application. In this paper, Policy Constraint Q-Learning (PCQL), a data-driven reinforcement learning algorithm for solving the problem of learning anesthesia strategies on real clinical datasets, is proposed. Conservative Q-Learning was first introduced to alleviate the problem of Q function overestimation in an offline context. A policy constraint term is added to agent training to keep the policy distribution of the agent and the anesthesiologist consistent to ensure safer decisions made by the agent in anesthesia scenarios. The effectiveness of PCQL was validated by extensive experiments on a real clinical anesthesia dataset. Experimental results show that PCQL is predicted to achieve higher gains than the baseline approach while maintaining good agreement with the reference dose given by the anesthesiologist, using less total dose, and being more responsive to the patient's vital signs. In addition, the confidence intervals of the agent were investigated, which were able to cover most of the clinical decisions of the anesthesiologist. Finally, an interpretable method, SHAP, was used to analyze the contributing components of the model predictions to increase the transparency of the model. | ['Yu Yao', 'Beiming Wang', 'Yaoyao Zhu', 'Jiao Chen', 'Xiuding Cai'] | 2023-03-17 | null | null | null | null | ['q-learning'] | ['methodology'] | [-1.27883747e-01 4.70395058e-01 -3.56683075e-01 -2.18662664e-01
-4.65811819e-01 -4.25090164e-01 6.16842136e-02 6.05912685e-01
-6.82568789e-01 1.01460290e+00 2.13531740e-02 -7.15068400e-01
-5.72639525e-01 -5.38509130e-01 -4.98736024e-01 -7.99245715e-01
-8.07925984e-02 4.81180340e-01 -2.63151944e-01 3.28235142e-02
2.46616423e-01 4.18691903e-01 -1.17914605e+00 -2.10847929e-02
1.03189254e+00 8.01223218e-01 2.18149945e-01 4.65389460e-01
1.62751630e-01 8.21576238e-01 -6.25391960e-01 -1.72062129e-01
4.62476790e-01 -6.80451691e-01 -2.73545593e-01 -1.26221672e-01
3.58696245e-02 -5.53583503e-01 3.87433395e-02 8.30340624e-01
6.83836937e-01 2.07126856e-01 3.77254486e-01 -9.13298428e-01
4.63315308e-01 3.86653513e-01 -1.62505675e-02 -6.70380443e-02
2.28363842e-01 4.88347739e-01 5.67807496e-01 -2.57366765e-02
4.53091592e-01 1.01894033e+00 1.10828057e-01 6.87021852e-01
-1.00742650e+00 -6.21160328e-01 2.06872180e-01 2.37736162e-02
-9.53056335e-01 9.79115441e-02 5.87563932e-01 -4.18226033e-01
4.96436298e-01 1.62268832e-01 1.10159564e+00 8.13572288e-01
8.86263371e-01 2.94563115e-01 1.32642317e+00 -3.38784575e-01
7.00009525e-01 4.86225247e-01 -1.27548486e-01 5.78186989e-01
4.41159070e-01 6.65350378e-01 -2.46546775e-01 -2.91957051e-01
7.59906232e-01 1.83009863e-01 -5.03525257e-01 -9.32393551e-01
-9.05563951e-01 1.09890497e+00 2.94555455e-01 2.58341357e-02
-1.11324418e+00 -4.36894447e-02 3.98923904e-01 7.50574619e-02
-1.10091150e-01 8.66778255e-01 -6.05621636e-01 -3.78710926e-01
-5.78869104e-01 3.19524080e-01 1.00157964e+00 5.50442517e-01
2.27468506e-01 -6.12171926e-02 -4.17604595e-01 2.85654843e-01
1.83259919e-01 4.11497474e-01 4.98308986e-01 -1.22235143e+00
3.41094375e-01 5.44725358e-01 6.51664913e-01 -6.17092133e-01
-6.35676384e-01 -6.43569291e-01 -4.31158483e-01 5.26768684e-01
3.67939264e-01 -4.05351520e-01 -7.62822390e-01 1.39096892e+00
4.47024047e-01 -1.29456669e-01 2.47420609e-01 9.06751394e-01
-2.70766877e-02 3.60188633e-01 3.37897807e-01 -8.11084569e-01
1.19560337e+00 -3.96868408e-01 -9.58553195e-01 -1.03060445e-02
5.38929164e-01 -3.74879032e-01 9.95496452e-01 7.33658135e-01
-1.12398136e+00 -4.09130245e-01 -8.47791493e-01 7.81903386e-01
7.25146979e-02 -2.61018366e-01 5.70619285e-01 6.73619866e-01
-5.64007521e-01 7.56089866e-01 -1.06605971e+00 4.25510406e-02
4.19678479e-01 4.94507879e-01 6.62230775e-02 -1.38446301e-01
-1.26213205e+00 1.18631899e+00 4.82191473e-01 1.17713436e-01
-1.11396730e+00 -8.69099736e-01 -8.29856217e-01 2.43528008e-01
6.37184680e-01 -8.29603374e-01 1.41869521e+00 -8.33650887e-01
-1.86322546e+00 1.94516718e-01 1.93665758e-01 -6.80111229e-01
7.49128819e-01 -1.31953478e-01 -5.91707416e-03 3.27601701e-01
-3.19939107e-01 4.39857692e-01 5.94529986e-01 -1.24964666e+00
-5.99643111e-01 -2.93627471e-01 1.94502279e-01 5.03950596e-01
-2.76817698e-02 -4.80037421e-01 -5.69458976e-02 -1.36244908e-01
-3.26572537e-01 -1.09520900e+00 -7.76633859e-01 -2.81782895e-01
1.46593809e-01 5.98871373e-02 2.21855253e-01 -3.44132334e-01
1.13004744e+00 -2.26349282e+00 3.42943296e-02 4.35952991e-01
3.19006220e-02 2.65462935e-01 6.91565350e-02 6.34999037e-01
-1.44096576e-02 -2.33452007e-01 -2.01533347e-01 7.24770501e-02
-2.27955177e-01 4.73185569e-01 -2.35304147e-01 3.50586981e-01
-2.25991488e-01 5.72472274e-01 -9.91478264e-01 -3.85996014e-01
4.60833788e-01 1.92156687e-01 -7.22536027e-01 7.08681643e-01
-3.65763009e-01 8.02529454e-01 -5.26992917e-01 2.88963497e-01
3.04390311e-01 8.24035108e-02 5.80215216e-01 -1.52942864e-02
-2.48248666e-03 -9.31677967e-03 -8.86980236e-01 1.33276689e+00
-7.36050189e-01 -1.26106232e-01 2.17368677e-01 -7.06541657e-01
8.66261005e-01 3.14133793e-01 9.90107298e-01 -9.71110404e-01
3.64594191e-01 1.72718056e-02 6.42487168e-01 -7.83774257e-01
-5.52759506e-02 -5.18869340e-01 1.78157285e-01 2.76674479e-01
-4.59407687e-01 -6.39962256e-01 7.76071697e-02 -1.80237293e-02
6.73752069e-01 3.04128882e-02 6.95818663e-01 -4.09729362e-01
1.84105575e-01 1.20601326e-01 6.95701182e-01 7.42527425e-01
-4.97555107e-01 -2.52945144e-02 4.98912036e-01 -4.62917775e-01
-4.39834863e-01 -8.72007310e-01 -1.35108709e-01 3.58846098e-01
3.50318849e-01 -2.04156283e-02 -7.12793827e-01 -8.08095872e-01
1.36374474e-01 1.16537285e+00 -5.67437589e-01 -4.28736299e-01
-3.88803631e-01 -7.14880288e-01 -2.78233945e-01 2.33717948e-01
-1.41691165e-02 -1.12047756e+00 -1.52588618e+00 6.55867875e-01
-1.18479781e-01 -8.88504565e-01 -3.29141676e-01 4.65724885e-01
-1.14636648e+00 -1.48247838e+00 -3.24994981e-01 -1.11434847e-01
6.51114464e-01 -2.00031221e-01 7.88846374e-01 -1.08011536e-01
-3.19637179e-01 5.38007557e-01 -1.59739032e-01 -9.96453285e-01
-6.06115341e-01 -2.44549558e-01 2.10087940e-01 -3.24726701e-01
1.81581557e-01 -1.50366381e-01 -9.97322977e-01 2.64128208e-01
-9.24264312e-01 -2.06928596e-01 7.62732923e-01 9.95144308e-01
5.79225838e-01 6.84111565e-02 6.82585597e-01 -1.04192615e+00
8.58098447e-01 -2.55918205e-01 -9.54112232e-01 8.93092528e-02
-1.24133265e+00 3.01955581e-01 8.17674994e-01 -4.93995547e-01
-8.94367635e-01 8.32687616e-02 1.73374385e-01 -4.86270040e-01
3.02852802e-02 4.23659801e-01 2.16366332e-02 2.50855625e-01
4.85598415e-01 3.52121554e-02 5.51283300e-01 -3.93568687e-02
-9.94244665e-02 3.20661545e-01 8.42216238e-02 -6.02563143e-01
2.85793394e-01 9.75121260e-02 1.65960833e-01 -3.11173826e-01
-7.59000421e-01 -1.98551118e-01 -2.52938122e-01 -2.17048198e-01
6.39123857e-01 -8.17333937e-01 -1.11706424e+00 -4.53686975e-02
-5.78319311e-01 -6.55254066e-01 -4.57522243e-01 1.03935671e+00
-8.08389485e-01 1.53434440e-01 -1.81852773e-01 -9.93583262e-01
-2.55680650e-01 -1.28654134e+00 4.08518225e-01 4.29310799e-01
-1.66745260e-01 -1.10816360e+00 2.28641450e-01 2.83506155e-01
4.02346343e-01 3.55948746e-01 1.16823411e+00 -5.70853353e-01
-4.23622072e-01 -1.26332730e-01 4.80723888e-01 4.29343343e-01
3.81748557e-01 -5.41999079e-02 -5.57443678e-01 -5.40218413e-01
3.09329271e-01 -5.77938631e-02 1.11122929e-01 6.98819160e-01
1.26079929e+00 -4.96283382e-01 -2.67974883e-01 4.32496965e-01
1.46859157e+00 9.21164155e-01 4.92511094e-01 3.45483780e-01
6.38034344e-02 7.00689495e-01 1.24552417e+00 8.08328450e-01
2.90439278e-01 5.37136197e-01 9.12140608e-01 -1.90489352e-01
3.55887115e-01 -1.81687742e-01 2.00730607e-01 2.98584789e-01
1.87530629e-02 -6.21175431e-02 -7.09056616e-01 1.45359531e-01
-1.79746401e+00 -6.16355658e-01 4.13840711e-01 2.59521222e+00
7.78716803e-01 2.40813479e-01 1.11907847e-01 -2.26573437e-01
2.88245171e-01 -3.75539988e-01 -6.90581381e-01 -7.43924737e-01
6.62613094e-01 3.87570947e-01 6.91689193e-01 5.36345243e-01
-6.15858555e-01 5.73294878e-01 6.32015467e+00 1.89534992e-01
-1.33249247e+00 -3.92993331e-01 6.12300038e-01 -1.80025131e-01
-5.34768663e-02 -1.04365073e-01 -5.01689374e-01 6.77290678e-01
1.12117136e+00 -2.60355264e-01 4.60614949e-01 9.43961859e-01
9.86406624e-01 -5.46707928e-01 -1.09257638e+00 7.27182984e-01
-2.97281981e-01 -9.62161899e-01 -2.81758815e-01 7.65920058e-02
4.16815698e-01 -3.36908728e-01 -1.97901890e-01 3.65024954e-01
2.51765192e-01 -8.80962789e-01 4.03906018e-01 7.33171523e-01
4.77493405e-01 -9.74491417e-01 1.17277706e+00 5.14075994e-01
-3.72975171e-01 -5.82170367e-01 -8.95954296e-02 -1.08015724e-01
-1.21293431e-02 3.38899821e-01 -1.35981464e+00 4.99308944e-01
3.79019439e-01 2.27193311e-01 -1.18742101e-01 1.09131420e+00
-1.69895679e-01 5.57060957e-01 -1.18102752e-01 -3.39922100e-01
2.79071748e-01 -3.51614326e-01 1.91142574e-01 7.21945465e-01
5.90310432e-02 3.16857845e-01 3.60467046e-01 4.95575011e-01
3.34546894e-01 3.53127629e-01 -5.72396278e-01 1.13066792e-01
3.87265801e-01 1.00741446e+00 -4.11647916e-01 -1.25509977e-01
6.51408732e-02 3.08483928e-01 2.41192598e-02 3.51174206e-01
-8.39067698e-01 -1.94773510e-01 6.28468633e-01 2.23740011e-01
1.85984913e-02 -4.97546457e-02 -2.90442824e-01 -5.65261602e-01
-1.43983841e-01 -1.05931282e+00 5.16377568e-01 -3.85425180e-01
-7.17464507e-01 3.91536564e-01 1.62844434e-01 -1.49989223e+00
-5.70097327e-01 -5.93509734e-01 -3.82298559e-01 9.10056591e-01
-1.71019804e+00 -2.31542185e-01 -2.80778497e-01 4.65221792e-01
3.23747098e-01 1.06149592e-01 9.13823485e-01 -2.49527525e-02
-4.52352047e-01 2.55456984e-01 -4.94461097e-02 -4.32156652e-01
6.73974633e-01 -1.17126822e+00 -7.98482597e-01 4.10220712e-01
-5.56520879e-01 6.90404534e-01 8.29606175e-01 -5.09866416e-01
-1.47245502e+00 -8.05843771e-01 6.88983798e-02 -1.37304887e-01
2.87144303e-01 8.34216624e-02 -7.26201773e-01 2.02527940e-01
5.06627932e-02 -2.41480887e-01 6.51359618e-01 -2.79589534e-01
3.62511218e-01 -4.75080639e-01 -1.20984674e+00 5.07719994e-01
3.01058531e-01 1.06743850e-01 -4.66659814e-01 2.29459435e-01
4.95846450e-01 -6.85919166e-01 -9.70313609e-01 5.78887463e-01
4.85209316e-01 -7.68486619e-01 5.90396047e-01 -7.02860594e-01
2.48785868e-01 -8.13466460e-02 3.82502556e-01 -1.67467153e+00
6.28871843e-02 -6.47660375e-01 -2.17741489e-01 3.26224416e-01
3.08846116e-01 -7.33171046e-01 6.89556122e-01 9.18103039e-01
-1.09171562e-01 -1.09875178e+00 -8.41150343e-01 -5.98265827e-01
1.68483499e-02 1.00729600e-01 3.25779051e-01 6.06785059e-01
1.88796252e-01 2.28169654e-02 -2.99630016e-01 2.66966403e-01
4.68372464e-01 2.24356577e-01 6.80933952e-01 -8.93079281e-01
-4.75705832e-01 -3.12524855e-01 -1.10125162e-01 -4.85111773e-01
-9.11808684e-02 -4.05329138e-01 2.86056161e-01 -1.64634788e+00
1.02201082e-01 -5.78229487e-01 -5.77086627e-01 4.74873066e-01
-4.49255586e-01 -6.35228455e-01 1.53406590e-01 -1.14173971e-01
-4.77908105e-02 5.84110916e-01 1.69000351e+00 -1.96950510e-02
-6.78911567e-01 5.31005085e-01 -6.16594613e-01 3.94995689e-01
9.32146013e-01 -6.51042044e-01 -9.06045973e-01 9.53371599e-02
-6.08242862e-02 7.63773739e-01 -4.66765985e-02 -7.04960763e-01
5.27559146e-02 -6.73101008e-01 1.52008474e-01 -2.11321756e-01
1.41024500e-01 -1.37602293e+00 1.68867201e-01 1.29161489e+00
-2.95151144e-01 1.33560419e-01 4.98179525e-01 4.50380325e-01
-3.01721454e-01 -1.53759956e-01 1.01759231e+00 -1.13235503e-01
-1.56411991e-01 3.43231052e-01 -3.18193287e-01 -3.38857323e-02
1.39485264e+00 2.92465203e-02 1.39001936e-01 -5.56493819e-01
-6.89017951e-01 5.15688002e-01 1.96777150e-01 2.19720230e-01
5.82458913e-01 -4.82255667e-01 -6.13269508e-01 3.23349386e-02
1.04077041e-01 7.77995586e-02 3.92952651e-01 8.27321708e-01
-7.10135102e-01 4.98760372e-01 -2.64162093e-01 -4.48074013e-01
-9.99627590e-01 8.69398236e-01 8.26524973e-01 -5.48689187e-01
-5.34092903e-01 1.77215472e-01 1.89211443e-01 -1.43558040e-01
5.11848032e-01 -5.60557008e-01 -8.98034126e-02 -1.19935960e-01
3.28745008e-01 -6.54571690e-03 1.40896514e-01 2.71149993e-01
-3.10305625e-01 2.84372419e-01 -1.94102019e-01 1.81072187e-02
1.15257049e+00 1.12390652e-01 3.15448225e-01 1.63521141e-01
5.37931502e-01 -2.87556229e-03 -1.55823624e+00 2.66959339e-01
-1.09521151e-01 -4.54170287e-01 3.99349146e-02 -1.41428971e+00
-8.91699314e-01 7.34433353e-01 9.44893539e-01 -7.99910352e-02
1.21859288e+00 -6.30163074e-01 2.53609866e-01 3.01166266e-01
5.96971571e-01 -1.11773801e+00 5.04992157e-02 -1.77826993e-02
6.18753731e-01 -1.20992076e+00 1.18540086e-01 -1.66336671e-01
-1.10936904e+00 1.06277061e+00 7.39792764e-01 1.73450932e-01
4.81536984e-01 9.56739709e-02 5.99558353e-01 -1.71877984e-02
-6.96730852e-01 3.46926838e-01 -1.26736447e-01 2.23259076e-01
7.86885247e-02 3.43740016e-01 -5.46618223e-01 3.11726481e-01
1.70244661e-03 2.93661445e-01 7.80795693e-01 1.02991915e+00
-2.78755844e-01 -1.13415980e+00 -1.80294707e-01 4.03012842e-01
-4.78309572e-01 7.52202123e-02 2.24109799e-01 1.04089904e+00
-4.16733511e-02 8.24881911e-01 -1.68543443e-01 2.15275735e-01
6.14090860e-01 2.70578861e-02 5.16040981e-01 -7.22803295e-01
-6.59972191e-01 1.14564501e-01 -2.37259507e-01 -8.41299117e-01
-1.61192626e-01 -4.73658085e-01 -1.40074348e+00 2.92984158e-01
-5.94394542e-02 6.94732428e-01 9.96741593e-01 6.60846055e-01
3.65495354e-01 8.43253791e-01 1.00939596e+00 -2.93265939e-01
-1.10033584e+00 -7.64767289e-01 -5.92673540e-01 2.41989732e-01
3.67466390e-01 -6.06926739e-01 -4.08935726e-01 -3.47502202e-01] | [4.054839134216309, 2.6534931659698486] |
36e6f914-a839-47d1-8733-4cc57264b7aa | revisiting-the-plastic-surgery-hypothesis-via | 2303.10494 | null | https://arxiv.org/abs/2303.10494v1 | https://arxiv.org/pdf/2303.10494v1.pdf | Revisiting the Plastic Surgery Hypothesis via Large Language Models | Automated Program Repair (APR) aspires to automatically generate patches for an input buggy program. Traditional APR tools typically focus on specific bug types and fixes through the use of templates, heuristics, and formal specifications. However, these techniques are limited in terms of the bug types and patch variety they can produce. As such, researchers have designed various learning-based APR tools with recent work focused on directly using Large Language Models (LLMs) for APR. While LLM-based APR tools are able to achieve state-of-the-art performance on many repair datasets, the LLMs used for direct repair are not fully aware of the project-specific information such as unique variable or method names. The plastic surgery hypothesis is a well-known insight for APR, which states that the code ingredients to fix the bug usually already exist within the same project. Traditional APR tools have largely leveraged the plastic surgery hypothesis by designing manual or heuristic-based approaches to exploit such existing code ingredients. However, as recent APR research starts focusing on LLM-based approaches, the plastic surgery hypothesis has been largely ignored. In this paper, we ask the following question: How useful is the plastic surgery hypothesis in the era of LLMs? Interestingly, LLM-based APR presents a unique opportunity to fully automate the plastic surgery hypothesis via fine-tuning and prompting. To this end, we propose FitRepair, which combines the direct usage of LLMs with two domain-specific fine-tuning strategies and one prompting strategy for more powerful APR. Our experiments on the widely studied Defects4j 1.2 and 2.0 datasets show that FitRepair fixes 89 and 44 bugs (substantially outperforming the best-performing baseline by 15 and 8), respectively, demonstrating a promising future of the plastic surgery hypothesis in the era of LLMs. | ['Lingming Zhang', 'Yifeng Ding', 'Chunqiu Steven Xia'] | 2023-03-18 | null | null | null | null | ['program-repair', 'program-repair'] | ['computer-code', 'reasoning'] | [-3.06371659e-01 2.11616233e-01 -3.86321038e-01 6.25811890e-02
-9.21226084e-01 -4.80154663e-01 2.24337459e-01 3.67826551e-01
2.61397690e-01 3.40550840e-01 -5.65602407e-02 -7.08995223e-01
-1.23709701e-01 -7.74146438e-01 -1.04856515e+00 -7.27468431e-02
-1.70831531e-01 3.15486975e-02 4.76440966e-01 -4.23713177e-01
6.02395296e-01 5.42503642e-03 -1.57997906e+00 4.02289361e-01
1.09016514e+00 2.83575624e-01 2.52049178e-01 6.03262186e-01
-3.23597461e-01 6.74161375e-01 -7.54444063e-01 -3.93728286e-01
1.92372113e-01 -4.09659773e-01 -8.76348078e-01 -2.76817441e-01
3.90311033e-01 -1.18484735e-01 1.51188746e-01 9.08121109e-01
1.37335390e-01 -5.01661539e-01 3.44008347e-03 -1.30921805e+00
-4.44250703e-01 9.11881387e-01 -6.70044124e-01 1.13423523e-02
6.74581528e-01 3.92663598e-01 1.08564222e+00 -6.75844014e-01
6.77349925e-01 1.04928124e+00 1.00045455e+00 5.64442754e-01
-1.28531754e+00 -2.33491912e-01 -3.52361724e-02 -8.48423541e-02
-1.31337678e+00 -5.61850704e-02 5.88840187e-01 -8.45228136e-01
1.61847317e+00 3.47111642e-01 3.78727466e-01 9.71917093e-01
5.79093754e-01 4.63586122e-01 9.46291447e-01 -9.20480728e-01
2.29829192e-01 1.19994320e-01 5.51638961e-01 1.03608608e+00
4.19242233e-01 2.00000763e-01 -3.66219908e-01 -6.05884314e-01
4.78568852e-01 3.65456827e-02 -3.90926808e-01 -1.04579188e-01
-9.73028600e-01 8.12567770e-01 1.94830596e-01 4.31649327e-01
-6.68002740e-02 2.52567381e-01 4.93271381e-01 4.68074292e-01
1.02295130e-01 1.12826002e+00 -7.80424535e-01 -3.79860491e-01
-9.74645138e-01 4.52112108e-01 9.04407203e-01 1.08744991e+00
1.11872363e+00 -1.06245495e-01 -2.58068204e-01 5.90930104e-01
2.42317244e-01 4.26211394e-02 3.87246519e-01 -6.56486034e-01
5.20836890e-01 1.24911261e+00 -1.21684177e-02 -9.70837116e-01
-2.83885121e-01 -3.69326472e-01 -4.19310108e-02 2.93433964e-01
2.03275099e-01 -6.09226199e-03 -7.73690104e-01 1.61441243e+00
1.94098890e-01 1.35130346e-01 -1.77066356e-01 5.18328846e-01
3.66021931e-01 3.38503599e-01 -4.07150775e-01 4.61939946e-02
1.03189838e+00 -1.05012631e+00 -2.37820417e-01 -4.86030668e-01
1.25898409e+00 -6.48847342e-01 1.42195857e+00 6.11960292e-01
-7.56231308e-01 -1.89489111e-01 -1.09207034e+00 2.44898885e-01
-1.30704835e-01 1.39081731e-01 8.33876252e-01 7.28696585e-01
-1.07260895e+00 7.79661775e-01 -9.42258954e-01 -4.60014760e-01
9.15533379e-02 1.78487182e-01 -3.93215686e-01 -3.75997454e-01
-5.53321481e-01 7.45595813e-01 9.01625827e-02 -1.31671846e-01
-8.95403862e-01 -1.10877335e+00 -8.65561366e-01 9.39680338e-02
8.38922322e-01 -6.06796026e-01 1.32117879e+00 -5.98838568e-01
-1.01115978e+00 4.70473826e-01 -1.93301365e-02 -3.09791237e-01
3.69737782e-02 -3.60703737e-01 -1.10791281e-01 -2.31711298e-01
2.02221245e-01 1.31404012e-01 7.41753876e-01 -1.14948559e+00
-3.52018207e-01 -1.08045759e-02 6.28434598e-01 -8.01197350e-01
-3.30143988e-01 3.21749151e-01 -3.40589672e-01 -6.13244772e-01
-1.70734510e-01 -9.29938018e-01 -2.59796917e-01 -3.95902336e-01
-2.99713582e-01 -1.94704235e-01 5.54241776e-01 -8.07581902e-01
1.99406469e+00 -2.21295881e+00 3.60066980e-01 5.62909953e-02
3.06712478e-01 3.00707310e-01 -4.07736570e-01 8.54007185e-01
-1.59926102e-01 5.44385016e-01 -3.08023572e-01 -1.64845124e-01
1.68024763e-01 1.90020204e-01 -4.02462304e-01 7.24497586e-02
3.97564232e-01 7.39455700e-01 -8.41753721e-01 -2.19027981e-01
-2.35621378e-01 -7.27484049e-03 -1.20905876e+00 3.23114276e-01
-7.54575312e-01 -1.92744285e-01 -2.23923668e-01 9.95693088e-01
4.56658572e-01 9.55373496e-02 6.79788813e-02 2.33996540e-01
-2.67036349e-01 4.01340187e-01 -1.03834498e+00 1.71125972e+00
-6.09628499e-01 1.93283588e-01 3.77222821e-02 -5.76903641e-01
9.32446361e-01 2.04769135e-01 -8.35526362e-03 -3.72596502e-01
-3.65931123e-01 6.09452546e-01 -2.89746206e-02 -8.50651145e-01
3.79446715e-01 2.61217147e-01 -4.20332313e-01 5.93230784e-01
1.07404925e-01 -2.18273252e-02 3.07552874e-01 1.16193064e-01
2.10735321e+00 3.23404074e-01 3.93225282e-01 -8.20423365e-02
2.32691854e-01 2.19522893e-01 9.33327615e-01 7.74306118e-01
2.99762338e-01 6.91542089e-01 1.04333603e+00 -4.05596912e-01
-7.79892802e-01 -6.38679981e-01 1.73825055e-01 7.66455710e-01
-2.76168257e-01 -1.32145989e+00 -9.71725464e-01 -1.20356154e+00
-4.17810120e-02 7.21382558e-01 -5.84076166e-01 -3.65428180e-01
-7.20955372e-01 -3.90008867e-01 7.16920912e-01 4.56982851e-01
-1.27961010e-01 -1.10417104e+00 -7.67286062e-01 2.94709712e-01
-4.07937430e-02 -4.49263960e-01 -4.04815465e-01 1.86549067e-01
-7.28503466e-01 -1.31301427e+00 -9.85746905e-02 -4.40538704e-01
6.52379215e-01 1.75645486e-01 1.23650301e+00 9.42775130e-01
-5.79589903e-01 3.16065907e-01 -7.77978539e-01 -1.05265647e-01
-7.31612802e-01 1.89410880e-01 -3.27736139e-01 -5.42653441e-01
1.82399303e-01 -6.87360764e-01 -1.62527512e-03 2.69627243e-01
-9.22462404e-01 -3.22110653e-01 7.83863783e-01 9.48536396e-01
1.41593322e-01 7.20366314e-02 4.56341624e-01 -9.42004383e-01
6.14770889e-01 -6.73073888e-01 -7.87763059e-01 4.56873745e-01
-6.91782236e-01 2.11430505e-01 5.37470460e-01 -3.83808047e-01
-5.77713013e-01 -1.63686007e-01 -2.53779292e-01 -3.95551592e-01
-9.83366743e-02 1.07714331e+00 -1.90101027e-01 -1.99614167e-01
9.73042965e-01 -1.77190229e-01 -2.86092460e-01 -7.31157422e-01
-4.00019214e-02 5.37783921e-01 2.61736661e-01 -1.07158983e+00
7.24520028e-01 -3.40653867e-01 -3.81712526e-01 -3.35233271e-01
-4.72509563e-01 -3.37743044e-01 -2.28855744e-01 1.25923663e-01
4.40587372e-01 -4.77244318e-01 -2.83877552e-01 3.04926693e-01
-1.29776728e+00 -4.30019796e-01 -1.32501245e-01 -1.45971715e-01
-3.76620293e-01 5.56808889e-01 -4.84969854e-01 -7.18523264e-01
-2.03998134e-01 -1.61200941e+00 1.07162642e+00 -1.25686437e-01
-6.64556444e-01 -6.25439405e-01 3.12793165e-01 1.98975369e-01
4.98372644e-01 2.83985436e-01 1.57644260e+00 -3.70781928e-01
-8.92735004e-01 -4.17665213e-01 2.29307443e-01 2.53084123e-01
6.19378760e-02 2.47581333e-01 -6.14775777e-01 -2.26385340e-01
-7.53963441e-02 -4.00087126e-02 5.83378851e-01 -1.29594684e-01
8.51721346e-01 -4.32549745e-01 -4.30268615e-01 3.76810223e-01
1.68271792e+00 -1.07347488e-01 7.10280538e-01 7.16664374e-01
5.56913614e-01 5.29063344e-01 7.80672312e-01 3.92139971e-01
3.08631063e-01 8.48021388e-01 8.13252926e-01 3.25802803e-01
-9.21949148e-02 -3.56522352e-01 7.20286548e-01 5.23657084e-01
2.13311434e-01 5.13300486e-02 -1.34140265e+00 8.47678721e-01
-1.95735180e+00 -6.68531001e-01 -4.35949236e-01 2.38494802e+00
9.96414244e-01 2.03611597e-01 2.81325765e-02 3.51216882e-01
3.51228952e-01 -2.81512052e-01 -2.78080180e-02 -6.69946015e-01
4.33855861e-01 4.08875406e-01 6.44677058e-02 3.38976830e-01
-7.36402273e-01 8.46182048e-01 5.64749718e+00 7.29849875e-01
-1.09436655e+00 1.02298245e-01 -3.26181650e-01 1.97095484e-01
-4.90508586e-01 6.88585520e-01 -1.00434613e+00 3.45168412e-01
9.18799639e-01 -6.87437952e-02 5.52020133e-01 1.24147737e+00
-6.79662228e-02 -3.02247733e-01 -1.41648424e+00 4.08374250e-01
-1.09997004e-01 -1.55944133e+00 -2.81435251e-01 1.14640534e-01
6.33325815e-01 -1.93653211e-01 -2.94618696e-01 7.01275587e-01
-1.67598370e-02 -9.58768368e-01 7.97628760e-01 6.24055862e-01
2.93338776e-01 -5.73589802e-01 9.48751986e-01 6.29319727e-01
-8.38607430e-01 -3.47947240e-01 -1.72550797e-01 -2.79777944e-01
-4.72081341e-02 8.87295246e-01 -9.89137888e-01 7.61327922e-01
9.46409583e-01 5.07647514e-01 -1.09011769e+00 1.44140983e+00
-4.65674192e-01 8.70181024e-01 -1.41345203e-01 2.37689540e-01
-9.56049487e-02 2.81598121e-01 6.23030782e-01 1.35418344e+00
5.63325942e-01 -5.56361735e-01 3.26495171e-01 1.21448529e+00
2.45916754e-01 -1.49156243e-01 -5.57815194e-01 -1.92064241e-01
5.17438710e-01 1.13998449e+00 -4.16732788e-01 1.34074062e-01
-4.49420154e-01 4.06382769e-01 5.78129411e-01 1.99961048e-02
-7.75338709e-01 -4.81627673e-01 5.95988393e-01 5.20227909e-01
4.20360476e-01 -2.57724732e-01 -4.49707657e-01 -1.02535391e+00
5.20774782e-01 -1.49780738e+00 1.75346985e-01 -5.55966020e-01
-9.72746074e-01 5.67685843e-01 8.20579603e-02 -1.03005326e+00
-2.20563620e-01 -2.50961572e-01 -9.71276045e-01 6.63025737e-01
-1.16150284e+00 -1.09691739e+00 -1.76124927e-02 3.06436070e-03
4.79256988e-01 -3.64611112e-02 9.23712313e-01 2.30714679e-01
-8.00353825e-01 8.68976533e-01 -4.20052141e-01 -2.97432452e-01
8.05952191e-01 -1.28631139e+00 4.90978599e-01 1.33440936e+00
-5.54265268e-02 1.36134124e+00 7.65708566e-01 -8.96005630e-01
-1.78907108e+00 -1.10181999e+00 1.02526236e+00 -6.78756714e-01
7.78329194e-01 -5.67002714e-01 -1.27435911e+00 7.17062175e-01
1.04365095e-01 -1.36163041e-01 3.20893735e-01 3.63876671e-01
-6.81063592e-01 3.69957276e-02 -1.03696084e+00 4.70331579e-01
8.13057542e-01 -4.27578092e-01 -5.75007617e-01 3.00604135e-01
1.03964353e+00 -3.98205251e-01 -8.98281157e-01 4.38246340e-01
9.62280333e-02 -1.11690366e+00 5.71958661e-01 -4.00643349e-01
8.37100625e-01 -5.19667685e-01 -2.14442074e-01 -1.19135439e+00
-1.51335597e-01 -8.30355763e-01 -2.78391361e-01 1.62784088e+00
5.78785241e-01 -6.11907065e-01 3.24036241e-01 5.38632452e-01
-6.81238711e-01 -1.12114155e+00 -6.44007146e-01 -9.21265483e-01
-8.03574547e-02 -6.81885719e-01 7.50623286e-01 9.10014808e-01
3.31276029e-01 -5.87180555e-02 -2.60642111e-01 3.94246101e-01
2.75610462e-02 2.12796092e-01 1.11763883e+00 -1.05869186e+00
-1.02490139e+00 -3.35843176e-01 -2.99330294e-01 -4.90247965e-01
2.04140365e-01 -8.91638398e-01 2.85976976e-01 -1.29812026e+00
1.92077041e-01 -6.87072873e-01 5.91132268e-02 1.11722159e+00
-3.78361762e-01 -2.32376710e-01 3.73097286e-02 -4.38659899e-02
-3.16494823e-01 4.83561680e-02 5.20166397e-01 -7.86065981e-02
-3.46344620e-01 -1.85873397e-02 -8.44318867e-01 7.27840483e-01
5.74112356e-01 -8.51992548e-01 -1.25573859e-01 -5.03370166e-01
7.78763413e-01 1.85252503e-01 5.32466888e-01 -9.62008655e-01
8.55740085e-02 -2.23081261e-01 -6.98290288e-01 -1.55462101e-01
-3.19299191e-01 -4.82073188e-01 2.85020798e-01 4.71571386e-01
3.65231116e-03 3.23205918e-01 4.55741376e-01 4.15638417e-01
-2.04271495e-01 -8.07446063e-01 3.00832778e-01 -1.89753458e-01
-6.59972131e-01 -1.05035909e-01 -3.94689411e-01 -7.68181831e-02
9.90223944e-01 -5.76005839e-02 -5.84076643e-01 3.30358475e-01
-2.35805422e-01 3.03965751e-02 9.23547745e-01 6.24100149e-01
4.77426231e-01 -7.80294597e-01 -3.79246712e-01 3.40671837e-01
4.91376400e-01 -1.61772832e-01 -1.05571702e-01 1.05394733e+00
-3.80444258e-01 1.22237906e-01 9.25419107e-02 -5.98752439e-01
-1.23706520e+00 8.18778813e-01 9.85230654e-02 -5.09674251e-01
-6.71379566e-01 7.23635375e-01 -1.10522591e-01 -6.98209047e-01
-2.69154776e-02 -6.18832231e-01 2.71607012e-01 -4.00510520e-01
4.02346879e-01 2.55715847e-01 6.05830133e-01 1.72804847e-01
-4.60119069e-01 4.45411116e-01 -1.58218801e-01 2.66860843e-01
1.59218967e+00 4.84700918e-01 -6.28748775e-01 2.03663200e-01
7.95748293e-01 4.63577271e-01 -7.40512729e-01 1.60353541e-01
5.66899776e-01 -5.84845483e-01 -1.74527735e-01 -1.09624982e+00
-8.32640350e-01 7.02984691e-01 -1.23962341e-03 1.51175007e-01
1.06643462e+00 1.32430062e-01 6.48384929e-01 2.45366380e-01
9.61381018e-01 -5.10521531e-01 1.77065164e-01 5.49172044e-01
1.01684320e+00 -7.93233097e-01 -2.25380421e-01 -6.14905477e-01
-8.06663856e-02 1.08712387e+00 9.35673118e-01 -1.75602302e-01
1.73809990e-01 6.46135569e-01 -2.88673908e-01 -2.21967652e-01
-1.00056970e+00 1.61863118e-01 -1.81335490e-02 5.65175593e-01
4.56659615e-01 -1.81333452e-01 -5.12078822e-01 1.12409139e+00
-2.03084290e-01 7.02353602e-04 9.09037590e-01 1.49434590e+00
-4.25918937e-01 -1.68870294e+00 -6.23919725e-01 3.33782107e-01
-2.77005941e-01 -1.54806972e-01 -4.68284309e-01 8.31180036e-01
3.70580763e-01 9.14283812e-01 -7.27083206e-01 -6.76664889e-01
4.74830717e-01 1.25403270e-01 7.02275455e-01 -1.29905808e+00
-1.12992787e+00 -1.64827660e-01 2.63204962e-01 -8.60620260e-01
2.37237871e-01 -6.04145110e-01 -1.07859635e+00 -2.24729143e-02
-6.37569964e-01 9.91152301e-02 4.39211190e-01 1.04814672e+00
6.59049511e-01 6.37465537e-01 2.82204837e-01 -5.68583071e-01
-6.19308233e-01 -7.68740416e-01 -7.48604462e-02 -1.07374758e-01
3.16627949e-01 -7.61036634e-01 -2.74307072e-01 -1.43316448e-01] | [7.628232479095459, 7.721411228179932] |
60142c83-0e8e-45ef-8478-a3eab235e813 | cascade-rcnn-for-midog-challenge | 2109.01085 | null | https://arxiv.org/abs/2109.01085v2 | https://arxiv.org/pdf/2109.01085v2.pdf | Cascade RCNN for MIDOG Challenge | Mitotic counts are one of the key indicators of breast cancer prognosis. However, accurate mitotic cell counting is still a difficult problem and is labourious. Automated methods have been proposed for this task, but are usually dependent on the training images and show poor performance on unseen domains. In this work, we present a multi-stage mitosis detection method based on a Cascade RCNN developed to be sequentially more selective against false positives. On the preliminary test set, the algorithm scores an F1-score of 0.7492. | ['April Khademi', 'Susan Done', 'Dimitri Androutsos', 'Fariba Dambandkhameneh', 'Salar Razavi'] | 2021-09-02 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [ 2.44639859e-01 -2.20717698e-01 -3.35043192e-01 -1.62721843e-01
-7.98290551e-01 -2.92855889e-01 5.95438242e-01 5.69970310e-01
-7.97455907e-01 1.16744769e+00 -2.92638510e-01 -1.48495466e-01
2.03279600e-01 -8.13991308e-01 -1.96272537e-01 -1.13883781e+00
2.38812938e-01 6.01833940e-01 5.90689659e-01 8.47525224e-02
4.59349066e-01 6.49410009e-01 -1.40339828e+00 3.37451398e-01
6.26804233e-01 7.05798030e-01 1.35928780e-01 9.99460101e-01
1.84041977e-01 6.30738795e-01 -7.45714426e-01 -2.51509130e-01
-4.85199988e-01 -5.62528670e-01 -6.36242986e-01 -1.99963480e-01
-3.24531756e-02 1.42419368e-01 -2.47458238e-02 9.93135214e-01
5.41526496e-01 -4.25980836e-01 9.83314097e-01 -7.33079731e-01
8.44451562e-02 3.71793509e-01 -4.28558528e-01 6.51249170e-01
-1.06902070e-01 -7.62124583e-02 6.47016704e-01 -8.49718988e-01
5.61466992e-01 5.12684882e-01 4.73859638e-01 7.17002153e-01
-1.29194379e+00 -6.74739599e-01 -6.46138072e-01 1.16523437e-01
-1.54222310e+00 -3.59090120e-01 4.44829762e-01 -2.65382499e-01
7.24937081e-01 3.14327598e-01 6.31388962e-01 9.45879936e-01
4.18635458e-01 7.76035964e-01 1.33388174e+00 -3.12524587e-01
3.28877985e-01 3.28399658e-01 4.44779433e-02 4.94108081e-01
6.31737888e-01 -3.25135738e-02 -3.42260599e-01 3.25400740e-01
9.08757508e-01 -1.53860450e-01 -3.35643142e-01 -4.64105532e-02
-9.81696010e-01 9.11774099e-01 4.66218263e-01 1.11341357e+00
5.64474910e-02 -2.01568864e-02 3.80750477e-01 -7.28511140e-02
3.99828643e-01 6.25780165e-01 -3.07604015e-01 1.26407608e-01
-1.18507814e+00 -1.17132060e-01 6.03774190e-01 2.37676531e-01
3.09896111e-01 -3.72781634e-01 -2.16554642e-01 5.07378995e-01
-2.31549200e-02 2.11990431e-01 7.79923856e-01 -2.19381869e-01
-2.68368185e-01 9.64510262e-01 -1.55757576e-01 -9.12005305e-01
-8.27550650e-01 -7.34935462e-01 -1.44173932e+00 -3.03880009e-03
7.65585601e-01 2.62022555e-01 -1.04375339e+00 1.09121132e+00
8.42137411e-02 5.84418736e-02 2.15907022e-01 6.78321242e-01
8.80469561e-01 5.13764083e-01 3.10469478e-01 -5.15869796e-01
1.08106458e+00 -7.17161238e-01 -6.15185380e-01 -1.06992401e-01
8.77054274e-01 -6.51534915e-01 4.57294494e-01 4.43819582e-01
-7.69076347e-01 -3.14266920e-01 -1.03211081e+00 7.29067251e-02
-6.64007127e-01 6.66644275e-01 5.64693928e-01 7.34448612e-01
-1.06860578e+00 4.39907223e-01 -8.96199703e-01 -5.19623041e-01
4.97706473e-01 7.23607063e-01 -3.00649613e-01 1.22270159e-01
-7.42312610e-01 8.79398227e-01 6.74542725e-01 1.06665149e-01
-7.82308698e-01 -6.84867054e-02 -3.64986241e-01 -3.26999500e-02
1.58227593e-01 -2.89142311e-01 1.06811547e+00 -4.86313015e-01
-1.38549471e+00 1.22852767e+00 -2.36850128e-01 -4.50299680e-01
5.41855574e-01 2.24406257e-01 -8.35635513e-02 2.10036784e-01
3.97380628e-02 7.48460591e-01 4.05062467e-01 -1.05841839e+00
-9.46963310e-01 -1.80779472e-01 -4.21145886e-01 -6.31287470e-02
-2.11415052e-01 -1.77860335e-01 -3.52720082e-01 -5.82651317e-01
-9.48439091e-02 -7.91852891e-01 -3.86507064e-01 -1.16455913e-01
-2.51338892e-02 -3.06960762e-01 6.10729694e-01 -5.23998857e-01
1.13253152e+00 -2.15830803e+00 2.84925085e-02 3.04444939e-01
4.43104766e-02 6.06062531e-01 2.97990888e-01 -2.03064337e-01
1.60707369e-01 2.79398918e-01 -1.73882380e-01 -3.23582627e-02
-7.55270481e-01 -6.82222322e-02 3.54375958e-01 5.80105484e-01
4.42045391e-01 8.50900590e-01 -9.26759899e-01 -7.53596187e-01
1.17574096e-01 4.55885470e-01 -1.66624650e-01 1.28314361e-01
3.80289704e-02 5.02366662e-01 -2.23809645e-01 1.04571748e+00
2.71916628e-01 -6.37753189e-01 2.49804825e-01 -1.40628368e-01
-2.15599220e-02 -3.95194501e-01 -7.20907390e-01 1.00924480e+00
-3.94685566e-02 6.35098994e-01 -3.17473888e-01 -8.94459903e-01
9.95172739e-01 3.56374979e-01 2.40395427e-01 -4.75484669e-01
7.22345173e-01 6.15897119e-01 9.78540778e-02 -2.14068875e-01
3.69804889e-01 -2.92596847e-01 8.91500786e-02 -4.05254448e-03
1.29531965e-01 -8.72847214e-02 5.57274580e-01 -2.18794048e-01
1.14336264e+00 -2.57609576e-01 7.64856040e-01 -3.38952631e-01
8.55273247e-01 9.84213054e-02 5.49334645e-01 3.44231874e-01
-3.54947984e-01 6.23089135e-01 5.95063090e-01 -4.90862072e-01
-8.06863606e-01 -5.14127553e-01 -2.02063739e-01 4.09520537e-01
2.79065571e-03 4.30255011e-02 -5.83535492e-01 -6.22600079e-01
-3.55701655e-01 3.13102782e-01 -7.20861733e-01 -3.83130424e-02
-4.77388054e-01 -1.37091064e+00 5.28171420e-01 5.99161386e-01
5.99355578e-01 -9.16328311e-01 -5.83599031e-01 3.47637177e-01
-9.06937718e-02 -1.04871511e+00 8.73772129e-02 5.73196352e-01
-9.59749520e-01 -1.35027719e+00 -1.10045671e+00 -1.08672464e+00
9.42169130e-01 1.06450267e-01 8.78511190e-01 3.98172706e-01
-6.07354581e-01 -4.92027640e-01 -3.41392368e-01 -3.84860933e-01
-5.50788760e-01 4.86518055e-01 -3.15513551e-01 -1.61123890e-02
5.56619406e-01 -4.18613665e-02 -5.91965675e-01 4.90719736e-01
-7.98084915e-01 -4.66395468e-02 1.03553343e+00 1.19466305e+00
9.37440276e-01 1.57812312e-01 6.48222268e-01 -1.06813753e+00
2.39750072e-01 -1.73382387e-01 -6.59401715e-01 1.60372481e-01
-5.41392684e-01 -1.67383462e-01 5.39802432e-01 -5.18524587e-01
-7.70977318e-01 5.17663062e-01 -1.24958649e-01 5.89982830e-02
-2.53359824e-01 4.05046850e-01 3.25192101e-02 -3.70002449e-01
6.72520459e-01 1.67747229e-01 -1.66345671e-01 -1.83912832e-03
-5.09231687e-01 5.63889682e-01 6.52821064e-01 1.63580716e-01
3.27073514e-01 4.47035819e-01 5.51242530e-01 -1.02857184e+00
-7.57750452e-01 -7.15298176e-01 -4.69369084e-01 -4.58953857e-01
7.74839461e-01 -8.43952537e-01 -4.69889909e-01 8.38734388e-01
-8.69180501e-01 -2.64464915e-01 1.42360315e-01 5.08558869e-01
-2.68087953e-01 2.58326143e-01 -6.35100007e-01 -6.05985105e-01
-4.26440269e-01 -8.68259192e-01 9.10395086e-01 7.31444478e-01
-2.91851550e-01 -8.31510723e-01 1.41117379e-01 3.11724633e-01
3.90838057e-01 4.72884923e-01 7.71500349e-01 -7.66059995e-01
-3.64132136e-01 -7.65873909e-01 -2.92601645e-01 -1.55405030e-02
1.81665272e-01 3.49818617e-01 -1.25989056e+00 -2.89137453e-01
-2.69391030e-01 -2.19953790e-01 1.13271773e+00 5.76798141e-01
1.23728824e+00 2.89833128e-01 -9.68727887e-01 2.34753162e-01
1.50490928e+00 2.76591510e-01 7.27352798e-01 3.85917664e-01
2.83041988e-02 3.05083811e-01 6.38682067e-01 -3.15945372e-02
-1.45774767e-01 1.08110733e-01 3.39472204e-01 -3.84220481e-01
-1.88671406e-02 4.72566299e-02 -1.74544752e-01 5.56507707e-01
-2.82124102e-01 -5.52647650e-01 -1.04387498e+00 6.22673154e-01
-1.59421277e+00 -7.07954288e-01 -3.00281256e-01 1.97419822e+00
8.45982671e-01 5.19108474e-01 1.39380440e-01 8.11864972e-01
8.87462676e-01 -2.77256846e-01 -3.03252876e-01 -7.92341605e-02
-3.66727859e-01 3.33152950e-01 4.16925281e-01 1.86293140e-01
-1.27102578e+00 9.35260296e-01 6.59315062e+00 8.12857270e-01
-1.18544304e+00 -1.45434961e-01 1.25407434e+00 1.27493486e-01
4.21620965e-01 -3.21888030e-01 -9.58291233e-01 3.96007538e-01
8.59304726e-01 7.56208822e-02 -3.15031886e-01 5.22975326e-01
-1.24724507e-01 -7.56378055e-01 -9.22890842e-01 8.75989854e-01
9.42261443e-02 -1.20739579e+00 -6.26358539e-02 1.00698054e-01
6.05352879e-01 -4.19310361e-01 -1.83850780e-01 3.01570833e-01
-1.43278018e-01 -1.24840391e+00 -1.78778961e-01 4.61649179e-01
9.79778051e-01 -8.80350232e-01 1.64768314e+00 4.77091491e-01
-8.44655931e-01 1.00309499e-01 -2.42025346e-01 2.09986165e-01
-1.97025761e-01 6.17711604e-01 -1.19913411e+00 -1.02242995e-02
5.19970059e-01 5.12090445e-01 -8.95989537e-01 1.55750072e+00
-1.36034280e-01 5.74903011e-01 -3.92394692e-01 -8.18332732e-01
-6.68643191e-02 4.08207804e-01 -3.81472260e-02 1.42362702e+00
4.68468577e-01 5.76147716e-03 -1.84542641e-01 3.28199565e-01
4.89353687e-02 2.06227362e-01 -3.54446262e-01 -2.07640052e-01
-2.50924081e-02 1.54707980e+00 -1.68402183e+00 -1.72601447e-01
-1.03702813e-01 5.93831182e-01 2.47134581e-01 -1.86410233e-01
-5.02327979e-01 -2.20337927e-01 -1.07272565e-01 3.23961265e-02
2.41662472e-01 1.47415444e-01 -2.34070346e-01 -9.08166111e-01
-4.81258065e-01 -5.99803388e-01 6.03292584e-01 -3.34329069e-01
-1.01168871e+00 4.33934540e-01 -4.08864111e-01 -1.02476823e+00
3.18736173e-02 -8.48446608e-01 -4.84429210e-01 4.41327304e-01
-1.51796377e+00 -9.43446040e-01 -3.57360095e-01 1.97118163e-01
5.95260322e-01 -1.18956424e-01 1.09694862e+00 1.75069124e-01
-7.21405327e-01 6.71354532e-01 -3.83340195e-02 1.03315368e-01
5.57839096e-01 -1.19017994e+00 -2.41525576e-01 6.73216701e-01
-2.06617326e-01 -6.11251257e-02 6.10260248e-01 -6.30515635e-01
-6.89715266e-01 -1.01721764e+00 1.14612257e+00 -1.08813919e-01
3.90891910e-01 9.94573161e-02 -8.84946406e-01 7.29704276e-02
-1.79217324e-01 -1.17681818e-02 8.90587568e-01 -2.78046578e-01
3.41611564e-01 -3.26777287e-02 -1.21483493e+00 4.34478045e-01
2.95632124e-01 -1.42953366e-01 1.34848086e-02 1.87907934e-01
-1.70067757e-01 -4.29893315e-01 -7.66784251e-01 7.53462017e-01
4.29324299e-01 -8.55402529e-01 5.48194826e-01 1.37289539e-01
1.88067928e-01 -5.33913314e-01 2.24334002e-01 -9.46346641e-01
-4.37176555e-01 -4.86930199e-02 1.37575716e-01 1.00093758e+00
6.14537776e-01 -2.15585023e-01 1.27050579e+00 -2.55296249e-02
3.26132476e-01 -9.53511357e-01 -1.12265980e+00 -4.02042270e-01
-1.17024258e-01 3.87936145e-01 -4.16182578e-02 5.51334620e-01
8.55195746e-02 6.96592152e-01 -3.68067399e-02 9.91503522e-02
3.58039647e-01 -1.85959503e-01 4.77550089e-01 -1.53622162e+00
6.82411566e-02 -7.65522480e-01 -8.88651371e-01 -3.05011243e-01
-9.98139977e-02 -6.42468810e-01 1.09992720e-01 -1.09291363e+00
4.36375380e-01 -6.60126507e-02 -3.99524927e-01 3.99072081e-01
-3.18067759e-01 7.02715456e-01 -5.28223813e-01 2.59844691e-01
-5.03275573e-01 -8.93664658e-02 1.09203148e+00 -2.84673631e-01
-1.23084858e-01 1.39641017e-01 -4.52198565e-01 7.39010990e-01
1.47031522e+00 -5.09280741e-01 -3.58401984e-02 2.62597352e-01
3.23759198e-01 4.71645258e-02 1.62068352e-01 -1.34718645e+00
4.23798412e-01 -1.08526051e-01 9.22122180e-01 -8.98496091e-01
3.04649353e-01 -6.03333294e-01 2.24220157e-01 9.08447742e-01
-1.82893723e-01 -1.93573311e-01 1.30037114e-01 6.30809546e-01
-3.15198511e-01 -3.00313830e-01 9.94011998e-01 -2.12559298e-01
-4.88398850e-01 3.82502563e-02 -6.42512619e-01 -3.96593094e-01
1.27978575e+00 -4.34267193e-01 -2.75286406e-01 6.82877973e-02
-5.27044833e-01 1.03418134e-01 4.56715316e-01 -1.43214822e-01
6.03832483e-01 -8.43922436e-01 -6.58117771e-01 3.68958414e-02
2.11513162e-01 2.09052503e-01 -2.32250273e-01 8.17581594e-01
-8.67971122e-01 5.16199052e-01 -2.98435420e-01 -7.01178670e-01
-1.70446658e+00 4.54848737e-01 4.75112736e-01 -8.32271338e-01
-1.55914769e-01 1.02313638e+00 -1.19719952e-01 2.74353981e-01
1.67647764e-01 -3.15318853e-01 -1.02316928e+00 2.37713918e-01
4.82779443e-01 4.17000473e-01 1.56651139e-01 -4.43663538e-01
-3.43723446e-01 5.88668764e-01 -3.11620086e-01 3.15762967e-01
1.31681013e+00 3.19961011e-01 -2.44940355e-01 4.78032529e-01
7.73243904e-01 -4.25062120e-01 -6.49432003e-01 1.57524779e-01
2.17168778e-01 -1.52844757e-01 9.12369862e-02 -8.12964320e-01
-1.00419843e+00 7.83531189e-01 7.39076316e-01 3.40538591e-01
1.19112754e+00 -7.68619180e-02 4.09959912e-01 4.17701930e-01
3.24213862e-01 -1.11940026e+00 3.37265171e-02 2.48095319e-01
4.28685457e-01 -1.42142916e+00 2.95183599e-01 -6.31205440e-01
-1.27195284e-01 1.31183457e+00 6.14219368e-01 -5.96144423e-02
5.50392807e-01 3.53602082e-01 1.39627516e-01 -1.94808707e-01
-6.90151155e-01 -1.58512399e-01 -1.54297486e-01 3.63422543e-01
7.19610035e-01 1.00710921e-01 -7.88212895e-01 4.34528053e-01
9.54357535e-02 4.50849503e-01 5.46513617e-01 8.96675050e-01
-5.45871854e-01 -8.53979349e-01 -2.66431779e-01 6.45509958e-01
-8.47195327e-01 3.61071914e-01 -6.35827541e-01 8.63618672e-01
-1.45974252e-02 7.95121372e-01 -1.20832980e-01 -2.29625732e-01
-5.25233038e-02 -6.44481555e-02 3.95244092e-01 -2.89562345e-01
-5.26456892e-01 2.20683888e-01 -1.23551689e-01 -1.09774582e-02
-8.44579279e-01 -5.18446445e-01 -1.10804272e+00 -9.08401608e-02
-7.50542462e-01 1.01265892e-01 4.98856276e-01 6.65196538e-01
-1.77959651e-01 5.10979772e-01 4.67699885e-01 -7.17413902e-01
-1.24130480e-01 -1.01285005e+00 -5.87028444e-01 -6.64719120e-02
3.37018609e-01 -5.91327906e-01 -6.05805039e-01 1.95728108e-01] | [15.049129486083984, -3.1035561561584473] |
189808ca-08b8-4663-bde1-62cf7fc764bd | coastalcph-at-semeval-2016-task-11-the | null | null | https://aclanthology.org/S16-1160 | https://aclanthology.org/S16-1160.pdf | CoastalCPH at SemEval-2016 Task 11: The importance of designing your Neural Networks right | null | ["H{\\'e}ctor Mart{\\'\\i}nez Alonso", 'Natalie Schluter', 'Joachim Bingel'] | 2016-06-01 | null | null | null | semeval-2016-6 | ['complex-word-identification'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.36746072769165, 3.638685464859009] |
550140c3-f969-4d02-9359-370bb62af4a3 | approximate-information-state-based | 2306.05991 | null | https://arxiv.org/abs/2306.05991v1 | https://arxiv.org/pdf/2306.05991v1.pdf | Approximate information state based convergence analysis of recurrent Q-learning | In spite of the large literature on reinforcement learning (RL) algorithms for partially observable Markov decision processes (POMDPs), a complete theoretical understanding is still lacking. In a partially observable setting, the history of data available to the agent increases over time so most practical algorithms either truncate the history to a finite window or compress it using a recurrent neural network leading to an agent state that is non-Markovian. In this paper, it is shown that in spite of the lack of the Markov property, recurrent Q-learning (RQL) converges in the tabular setting. Moreover, it is shown that the quality of the converged limit depends on the quality of the representation which is quantified in terms of what is known as an approximate information state (AIS). Based on this characterization of the approximation error, a variant of RQL with AIS losses is presented. This variant performs better than a strong baseline for RQL that does not use AIS losses. It is demonstrated that there is a strong correlation between the performance of RQL over time and the loss associated with the AIS representation. | ['Aditya Mahajan', 'Amit Sinha', 'Nima Akbarzadeh', 'Erfan Seyedsalehi'] | 2023-06-09 | null | null | null | null | ['q-learning'] | ['methodology'] | [ 7.84563944e-02 4.62048560e-01 -5.68767607e-01 5.46139702e-02
-1.02401114e+00 -5.75900972e-01 5.44881940e-01 7.00399041e-01
-7.26601601e-01 1.07422829e+00 2.47841328e-01 -3.74223411e-01
-3.50349277e-01 -6.99766636e-01 -7.67683625e-01 -8.62980306e-01
-2.86418319e-01 6.69437051e-01 1.58120334e-01 -1.56917404e-02
5.70092015e-02 3.76801580e-01 -1.36364388e+00 -1.26226768e-01
4.42505658e-01 1.10178912e+00 9.82125849e-03 7.48754382e-01
1.19249441e-01 1.19587708e+00 -5.38124502e-01 -1.51359320e-01
4.31191385e-01 -7.12275028e-01 -1.03038597e+00 1.40071753e-02
-2.12894663e-01 -5.74641645e-01 -4.93924707e-01 1.04406345e+00
3.68223816e-01 3.08295548e-01 4.28649426e-01 -1.28974462e+00
1.25578076e-01 7.28552520e-01 -1.09832130e-01 3.25612836e-02
4.08317626e-01 1.78754300e-01 1.24409497e+00 1.10335499e-02
8.08188260e-01 1.28116286e+00 6.11974955e-01 4.94252503e-01
-1.19971192e+00 -2.42070690e-01 2.81703919e-01 1.55431882e-01
-8.44283521e-01 -1.21493764e-01 3.31136405e-01 -3.14701200e-02
8.21249604e-01 -1.05528481e-01 8.45415235e-01 8.48673999e-01
5.59683740e-01 1.16280794e+00 1.45726073e+00 -5.06092906e-01
8.47503543e-01 -1.59527749e-01 6.21450394e-02 6.47343457e-01
2.67926663e-01 6.42127693e-01 -5.38321435e-01 -2.02709064e-01
7.07143903e-01 5.95446676e-03 -5.12406183e-03 -8.44323575e-01
-8.93991649e-01 9.53140259e-01 6.57383054e-02 -3.54266278e-02
-5.78217864e-01 5.11034966e-01 5.82341254e-01 8.07964146e-01
1.14292152e-01 3.49157095e-01 -2.15881765e-01 -6.86665893e-01
-7.27786958e-01 4.56058711e-01 1.17197239e+00 6.79227471e-01
6.72156155e-01 6.34767488e-02 -2.84581780e-01 5.68688028e-02
1.56696454e-01 4.71379310e-01 2.62228698e-01 -1.37659943e+00
4.16753501e-01 2.49628097e-01 6.98269606e-01 -2.41040364e-01
-4.97765630e-01 -2.69116789e-01 -4.40384179e-01 6.22533917e-01
6.60422742e-01 -3.10422033e-01 -5.57181418e-01 2.10944843e+00
2.82542318e-01 -6.58833161e-02 3.79159480e-01 6.20763958e-01
-2.09330201e-01 6.16999805e-01 -1.13544397e-01 -6.09182239e-01
7.61045992e-01 -6.66891456e-01 -9.22976434e-01 4.07576077e-02
4.98719335e-01 -6.25530779e-02 8.34810734e-01 6.49585724e-01
-1.32100058e+00 -1.75003886e-01 -1.20628524e+00 2.01006338e-01
1.17973499e-01 -5.53550422e-01 6.69035196e-01 7.54581928e-01
-1.08405793e+00 1.02821910e+00 -1.29380393e+00 -2.94221073e-01
2.50542283e-01 2.80283302e-01 1.04173571e-02 -1.44182378e-02
-1.11737299e+00 1.13830245e+00 5.44871688e-01 -1.76195532e-01
-1.45313299e+00 -2.71469712e-01 -6.57369852e-01 1.67348757e-02
8.35595131e-01 -5.03592432e-01 1.84034479e+00 -9.63953197e-01
-1.90419626e+00 1.68159485e-01 3.88273932e-02 -1.07508922e+00
1.09412718e+00 -2.07868367e-01 6.95119575e-02 5.89870393e-01
-1.69442832e-01 2.24110052e-01 9.14131582e-01 -1.06419897e+00
-9.87256289e-01 -3.10474873e-01 5.63315690e-01 4.22381163e-01
2.24016577e-01 -3.71973217e-01 1.11654125e-01 -1.55227467e-01
-1.25509784e-01 -9.53876138e-01 -6.19542956e-01 -1.82503894e-01
6.13508299e-02 -3.27029794e-01 4.48687613e-01 -4.07437682e-01
1.04868436e+00 -1.80790901e+00 1.47101387e-01 1.24053851e-01
7.68321827e-02 -3.21804471e-02 -1.14714928e-01 9.91652668e-01
1.54542118e-01 -2.07598135e-01 -2.60714173e-01 -4.45947230e-01
3.22814971e-01 5.80044568e-01 -5.26886523e-01 7.95632303e-01
-1.09147720e-01 7.36646891e-01 -1.25290668e+00 -3.05942953e-01
2.60924220e-01 -6.46590516e-02 -4.67464507e-01 1.61664248e-01
-5.40149093e-01 3.62617016e-01 -5.05138934e-01 3.00856203e-01
2.25317314e-01 -1.30214334e-01 3.89232159e-01 6.20747805e-01
-1.60664737e-01 4.94650304e-01 -1.35343444e+00 1.56818986e+00
-3.29187840e-01 2.88717091e-01 1.09934717e-01 -1.00886643e+00
4.80847716e-01 6.44028842e-01 8.58799279e-01 -7.03771412e-01
7.64311338e-03 8.14019144e-02 -2.56252855e-01 2.58126156e-03
6.42712414e-01 -4.93611157e-01 -1.62939787e-01 8.98511291e-01
3.52385491e-02 -9.95048434e-02 3.87562573e-01 6.05972186e-02
1.35995221e+00 4.54205841e-01 4.13153708e-01 -2.20647119e-02
1.82772994e-01 9.40859914e-02 4.53521103e-01 1.31056464e+00
-4.29049343e-01 6.92479014e-02 9.77043867e-01 -3.30612034e-01
-1.23572934e+00 -1.11389482e+00 3.74527723e-02 8.35651875e-01
2.42271349e-01 -3.20065379e-01 -7.65484869e-01 -6.04044437e-01
-6.31463602e-02 8.49673033e-01 -5.74824810e-01 -2.75994807e-01
-4.32416826e-01 -4.55907673e-01 4.38753337e-01 3.05465668e-01
4.34442431e-01 -1.24231064e+00 -1.25659907e+00 6.21158540e-01
7.52286240e-02 -9.26565766e-01 -1.03943765e-01 4.39679384e-01
-1.22333181e+00 -9.70194995e-01 -5.22516131e-01 -6.17958680e-02
2.03006938e-01 -5.34420013e-02 1.09270954e+00 -5.15406966e-01
4.11199987e-01 8.52089942e-01 -4.02663916e-01 -4.31743473e-01
-8.41500998e-01 4.31766175e-02 3.05431206e-02 -2.95027822e-01
-1.71936508e-02 -3.90880853e-01 -4.58405584e-01 -2.10659981e-01
-1.15010917e+00 -2.21734956e-01 6.47677183e-01 8.26225221e-01
5.83864689e-01 2.06015036e-01 3.80801022e-01 -5.20801663e-01
6.93571210e-01 -2.98553675e-01 -9.99954641e-01 1.73918426e-01
-7.72738636e-01 6.27275407e-01 6.34920001e-01 -2.58339137e-01
-8.80640388e-01 1.23422503e-01 1.72464296e-01 -2.38001108e-01
3.13929431e-02 5.12018681e-01 1.62568808e-01 2.88063407e-01
2.48128980e-01 2.52448797e-01 3.00007075e-01 -2.39326805e-01
3.18368196e-01 2.68878698e-01 1.41241774e-01 -6.51543200e-01
4.33849752e-01 6.02106690e-01 4.09304768e-01 -6.14856362e-01
-9.73665535e-01 -3.05259645e-01 -3.38949025e-01 -2.08717152e-01
5.20603657e-01 -6.45233214e-01 -1.29999030e+00 1.31581187e-01
-6.42213166e-01 -7.33665466e-01 -1.08309019e+00 7.64871418e-01
-1.45750964e+00 4.38882828e-01 -7.81522691e-01 -1.51763034e+00
-4.76403832e-02 -9.58416402e-01 6.38916969e-01 -1.33506253e-01
1.98047776e-02 -8.35125804e-01 3.92879069e-01 -1.31855994e-01
-5.32076545e-02 1.95368782e-01 8.89340281e-01 -4.72367048e-01
-7.16564715e-01 -2.52087712e-01 2.65638500e-01 2.58391231e-01
-3.27865124e-01 -4.52815503e-01 -6.38736427e-01 -7.42779970e-01
2.14000732e-01 -5.32287598e-01 8.17157745e-01 5.52343130e-01
5.82455575e-01 -6.00014865e-01 9.93937775e-02 -1.69937778e-02
1.71484232e+00 3.27561200e-01 4.98339593e-01 6.40279293e-01
-7.16080219e-02 4.77677435e-01 7.82615364e-01 8.52553070e-01
3.16080898e-01 3.52841079e-01 7.38290906e-01 3.77323091e-01
3.95682454e-01 -5.11752486e-01 8.71043801e-01 5.12473583e-01
-1.27812564e-01 -8.96361619e-02 -5.94455898e-01 4.02061552e-01
-2.03916049e+00 -1.14891863e+00 3.58312160e-01 2.66772652e+00
7.98884511e-01 4.13018137e-01 5.10903358e-01 3.20311397e-01
3.42250139e-01 1.74273640e-01 -8.98271680e-01 -7.30790079e-01
1.41462944e-02 2.36612707e-01 9.44186270e-01 7.11781025e-01
-9.49932158e-01 6.53490961e-01 7.10888577e+00 6.51128113e-01
-6.64744437e-01 1.93610266e-01 2.25575879e-01 -2.57482547e-02
3.06923385e-03 1.64169699e-01 -7.49626458e-01 3.21684599e-01
1.34744954e+00 -3.27414036e-01 6.68481886e-01 8.38858426e-01
2.74109244e-01 -4.29861993e-01 -1.17960513e+00 5.49111247e-01
-5.92638195e-01 -1.02338326e+00 -1.76769242e-01 3.29980195e-01
7.64141440e-01 -2.58746594e-02 1.12023484e-02 5.00984073e-01
1.09352183e+00 -9.18614328e-01 1.05337965e+00 5.01146078e-01
4.72660691e-01 -1.18622601e+00 8.02914858e-01 6.18092477e-01
-1.00746334e+00 -5.07120848e-01 -3.17430079e-01 -3.65775764e-01
1.63739428e-01 2.02579468e-01 -7.66360819e-01 7.16998160e-01
1.55236572e-01 4.97916549e-01 -2.80633979e-02 9.48637187e-01
-3.07187527e-01 7.43793428e-01 -4.78784055e-01 -1.01291619e-01
7.28895724e-01 -3.55778515e-01 6.91952467e-01 7.60927081e-01
8.85349587e-02 -9.35088098e-02 4.41371560e-01 4.56597894e-01
2.70913154e-01 -2.52527803e-01 -4.37172860e-01 -2.97705173e-01
2.89517313e-01 5.42589664e-01 -5.08018315e-01 -2.26786152e-01
-2.51149893e-01 7.70718396e-01 2.51569182e-01 3.41905326e-01
-5.33455431e-01 1.26493216e-01 4.42112029e-01 -1.90918490e-01
4.37501252e-01 -3.05203170e-01 1.53950617e-01 -8.36649239e-01
6.78707808e-02 -8.59463751e-01 6.31825387e-01 -2.44766548e-01
-1.01561701e+00 2.46669009e-01 1.83297209e-02 -1.16901684e+00
-8.79239857e-01 -9.99794900e-02 7.51142725e-02 4.44367766e-01
-1.47136009e+00 -5.84387958e-01 4.43672955e-01 4.78130609e-01
3.56045336e-01 1.53715655e-01 8.73726010e-01 -3.44904602e-01
-2.90498912e-01 1.75432459e-01 7.03996599e-01 -2.85109907e-01
6.06438480e-02 -1.66889751e+00 6.65887492e-03 6.19828820e-01
3.44386846e-02 1.24718882e-01 1.15179646e+00 -5.28811693e-01
-1.62026060e+00 -7.47672141e-01 3.63292456e-01 -3.27826262e-01
8.49719882e-01 1.31524652e-01 -6.09843612e-01 8.51130843e-01
5.20026237e-02 -2.74741471e-01 9.37756971e-02 -6.11147322e-02
-3.02975494e-02 -9.59726050e-02 -1.10379767e+00 4.87185448e-01
6.69942677e-01 -4.73679662e-01 -6.33439362e-01 1.02886319e-01
6.66629910e-01 -2.36796603e-01 -7.76400268e-01 6.47471249e-02
5.58096111e-01 -9.63595331e-01 5.68074524e-01 -8.43643188e-01
3.14512759e-01 -1.03100307e-01 -1.20162435e-01 -1.13028300e+00
4.58220355e-02 -1.11596787e+00 -5.76974750e-01 5.49202919e-01
-2.53381114e-02 -5.46208322e-01 9.36314106e-01 4.14990366e-01
1.46016002e-01 -7.10136831e-01 -1.44342017e+00 -1.19913590e+00
3.37017745e-01 -4.48670298e-01 2.48507306e-01 2.24604368e-01
1.91193014e-01 2.30177566e-01 -3.86555225e-01 4.16161641e-02
8.83516610e-01 3.12099069e-01 4.59617704e-01 -1.08253944e+00
-3.48687828e-01 -2.73006111e-01 -2.43840873e-01 -1.21976376e+00
4.12033424e-02 -5.02354801e-01 2.29211345e-01 -1.46276653e+00
9.40466374e-02 -4.74326253e-01 -6.37721002e-01 1.27192497e-01
2.84486413e-01 -4.59452480e-01 5.19991755e-01 2.14914203e-01
-1.23307729e+00 8.24233174e-01 1.20126915e+00 2.49881223e-01
-4.30210322e-01 4.60186213e-01 -1.87337145e-01 6.36907518e-01
8.91848564e-01 -7.22794592e-01 -3.49496752e-01 1.23553500e-01
4.59267288e-01 9.61602509e-01 2.21460879e-01 -1.08348095e+00
7.19931200e-02 -3.03927332e-01 -1.86200544e-01 -8.36912513e-01
4.48626012e-01 -8.26360881e-01 2.70516593e-02 1.04241765e+00
-8.15953612e-01 9.92148221e-02 -1.92239374e-01 1.07276857e+00
-4.59779100e-03 -6.58948004e-01 6.57863796e-01 -2.72392362e-01
-3.94067258e-01 3.24830770e-01 -8.11253905e-01 1.95995152e-01
9.41480875e-01 -7.27784559e-02 2.34498799e-01 -9.11566138e-01
-7.34814465e-01 2.37540588e-01 4.92523044e-01 -1.01624139e-01
2.41506621e-01 -9.53989208e-01 -4.28993583e-01 -3.72906506e-01
-2.40296975e-01 -1.65390715e-01 5.84023520e-02 7.63610423e-01
-1.76235586e-01 6.45191669e-01 -2.03674391e-01 -3.02435935e-01
-7.10755646e-01 7.46762574e-01 3.12344819e-01 -9.56114113e-01
-8.57898831e-01 4.90143187e-02 -2.95311004e-01 -5.18730283e-02
5.91793418e-01 -3.92562062e-01 8.56644735e-02 7.65752196e-02
5.29449701e-01 6.73823416e-01 -1.66497618e-01 -1.42573506e-01
1.90643698e-01 -5.91683760e-02 1.70308501e-01 -8.33810508e-01
1.33317745e+00 -4.89363313e-01 2.79309720e-01 8.55456173e-01
7.23635674e-01 -3.15486848e-01 -2.18462253e+00 -4.23295766e-01
3.23464900e-01 -1.21278264e-01 -1.09495908e-01 -5.24461389e-01
-6.18247271e-01 7.77220845e-01 6.72592282e-01 4.59735930e-01
7.66975522e-01 -2.97655165e-01 7.40966439e-01 8.15896332e-01
8.83638978e-01 -1.41659856e+00 8.72823745e-02 7.39837348e-01
4.94579583e-01 -8.79656613e-01 1.08644083e-01 3.73998940e-01
-7.52395630e-01 1.05083120e+00 -2.19553515e-01 -2.58743703e-01
2.12411851e-01 3.11888337e-01 -2.35265270e-01 1.80815980e-01
-1.02372968e+00 -5.15842676e-01 -5.27241886e-01 7.10069716e-01
-3.93826723e-01 1.46480039e-01 -1.44433782e-01 3.36682200e-01
-2.89900094e-01 2.42096066e-01 7.64230788e-01 1.24424815e+00
-6.93755269e-01 -1.01051354e+00 -2.14035451e-01 1.40403733e-01
-7.17404187e-01 2.41206259e-01 4.18497361e-02 8.16626132e-01
-5.28476000e-01 1.03032243e+00 -5.31911142e-02 2.63597339e-01
5.37567399e-02 2.98486166e-02 7.83607602e-01 -2.52756894e-01
-4.16796833e-01 -1.40604407e-01 -9.70945135e-03 -9.17653024e-01
-4.99664396e-01 -7.98317254e-01 -1.32942951e+00 -4.63765919e-01
-3.02902348e-02 3.49698722e-01 5.01691341e-01 1.05284286e+00
-1.07025027e-01 3.58152777e-01 5.89906514e-01 -3.95232797e-01
-1.52461147e+00 -5.79938352e-01 -8.27632129e-01 1.26476482e-01
7.13646412e-01 -7.08926260e-01 -2.70812988e-01 -3.01780492e-01] | [4.237282752990723, 2.2137224674224854] |
dc1685db-8875-473a-88d3-3d81ef282b79 | metaann-um-gerador-de-ferramentas-para | null | null | https://aclanthology.org/W13-4802 | https://aclanthology.org/W13-4802.pdf | MetaAnn: Um Gerador de Ferramentas para Anota\cc\~ao de Textos (MetaAnn: a Generator of Text Annotation Tools) [in Portuguese] | null | ['Norton Trevisan Roman', 'Tiago Emanuel Infante Miss{\\~a}o'] | 2013-01-01 | null | null | null | ws-2013-1 | ['text-annotation'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.352250099182129, 3.6770882606506348] |
9613fba1-7998-468f-bef7-510c46137420 | dual-view-molecule-pre-training | 2106.10234 | null | https://arxiv.org/abs/2106.10234v2 | https://arxiv.org/pdf/2106.10234v2.pdf | Dual-view Molecule Pre-training | Inspired by its success in natural language processing and computer vision, pre-training has attracted substantial attention in cheminformatics and bioinformatics, especially for molecule based tasks. A molecule can be represented by either a graph (where atoms are connected by bonds) or a SMILES sequence (where depth-first-search is applied to the molecular graph with specific rules). Existing works on molecule pre-training use either graph representations only or SMILES representations only. In this work, we propose to leverage both the representations and design a new pre-training algorithm, dual-view molecule pre-training (briefly, DMP), that can effectively combine the strengths of both types of molecule representations. The model of DMP consists of two branches: a Transformer branch that takes the SMILES sequence of a molecule as input, and a GNN branch that takes a molecular graph as input. The training of DMP contains three tasks: (1) predicting masked tokens in a SMILES sequence by the Transformer branch, (2) predicting masked atoms in a molecular graph by the GNN branch, and (3) maximizing the consistency between the two high-level representations output by the Transformer and GNN branches separately. After pre-training, we can use either the Transformer branch (this one is recommended according to empirical results), the GNN branch, or both for downstream tasks. DMP is tested on nine molecular property prediction tasks and achieves state-of-the-art performances on seven of them. Furthermore, we test DMP on three retrosynthesis tasks and achieve state-of-the-art results on them. | ['Tie-Yan Liu', 'Houqiang Li', 'Wengang Zhou', 'Tao Qin', 'Yingce Xia', 'Jinhua Zhu'] | 2021-06-17 | null | null | null | null | ['retrosynthesis'] | ['medical'] | [ 8.22925925e-01 2.21759126e-01 -5.88861823e-01 -1.27262264e-01
-3.39102179e-01 -6.92484558e-01 7.48253286e-01 5.96323311e-01
-1.56290364e-02 7.80423939e-01 -3.90864816e-03 -8.63531590e-01
1.47591323e-01 -1.10704303e+00 -1.01827288e+00 -8.27290416e-01
-1.57946900e-01 3.56906354e-01 1.75427705e-01 -2.20405012e-01
4.09677178e-01 5.43501079e-01 -1.25129926e+00 6.58445954e-01
9.74904895e-01 8.39244187e-01 3.31284404e-01 4.88069028e-01
-4.22117740e-01 9.10339892e-01 -1.35762826e-01 -5.14576316e-01
1.57381967e-01 -6.85897112e-01 -8.43061507e-01 -2.80138642e-01
9.69574973e-02 3.11773032e-01 -1.41649246e-01 8.50677371e-01
4.74717259e-01 1.71818540e-01 8.57516825e-01 -7.03185380e-01
-5.48667908e-01 7.52558112e-01 -5.05014479e-01 -1.00396961e-01
5.31035364e-01 2.73675740e-01 1.22792423e+00 -1.06118691e+00
9.12303627e-01 1.16689587e+00 2.58848786e-01 5.18323183e-01
-1.37697232e+00 -5.83506227e-01 3.68893653e-01 1.79994136e-01
-1.00437748e+00 -3.24191749e-01 5.90040088e-01 -6.57912612e-01
1.56936181e+00 1.52482809e-02 4.52127218e-01 9.92744863e-01
3.25245470e-01 6.32763326e-01 8.35839868e-01 -3.09772044e-01
3.88753057e-01 -2.29081050e-01 1.17669642e-01 8.90595317e-01
6.29040301e-02 1.42518073e-01 -5.59707046e-01 -2.99610972e-01
3.14268887e-01 1.03463590e-01 -2.72498280e-01 -3.33634645e-01
-9.65735435e-01 9.50194299e-01 9.85203743e-01 8.55374709e-02
-5.61225533e-01 -2.97889952e-02 3.78816575e-01 2.40190521e-01
2.90880144e-01 7.51761198e-01 -4.77179140e-01 4.46217239e-01
-6.81838453e-01 -6.64456263e-02 8.02941620e-01 7.27452934e-01
7.80048430e-01 -2.53181398e-01 -1.52263820e-01 5.69246292e-01
3.07905436e-01 -1.17324643e-01 3.90990138e-01 1.23571731e-01
5.90723515e-01 7.95787454e-01 -1.75167978e-01 -5.97752810e-01
-5.81654489e-01 -4.23106492e-01 -7.69853592e-01 7.82180876e-02
3.25285584e-01 1.03345901e-01 -1.34076917e+00 1.70011568e+00
4.06643242e-01 1.74456894e-01 3.05261314e-01 6.99470520e-01
1.18533790e+00 1.09909785e+00 4.26371694e-01 -2.53080338e-01
1.34292448e+00 -1.23634565e+00 -2.56645173e-01 -3.95733006e-02
9.48835731e-01 -5.66047609e-01 5.26299536e-01 5.50767362e-01
-9.38531101e-01 -6.48571849e-01 -1.30267715e+00 -2.54552662e-01
-7.62663126e-01 2.48826649e-02 7.76160777e-01 4.00462270e-01
-5.36065459e-01 1.20820749e+00 -5.16478479e-01 -1.28982708e-01
3.06963772e-01 5.92471063e-01 -5.12082160e-01 -3.03499430e-01
-1.30668116e+00 8.87004197e-01 7.47868776e-01 -7.07755685e-02
-1.00882995e+00 -7.71177471e-01 -1.03095937e+00 1.32271871e-01
6.03281856e-01 -9.21147525e-01 7.44599521e-01 -9.86382782e-01
-1.60834777e+00 7.79294491e-01 -2.54833877e-01 -4.81015801e-01
3.57513756e-01 1.34943455e-01 -2.50669122e-01 4.90195118e-02
-6.22806326e-02 6.90434933e-01 8.08430851e-01 -1.02453816e+00
-2.58948356e-01 -4.48702067e-01 1.88668028e-01 1.07232705e-01
8.50274712e-02 -2.38593698e-01 -3.07643712e-01 -5.70402145e-01
7.58148283e-02 -8.60911965e-01 -4.64113206e-01 -3.67267668e-01
-9.19776678e-01 -3.81849796e-01 3.31673473e-01 -6.63890004e-01
1.11484683e+00 -1.80299771e+00 6.69097185e-01 5.25292337e-01
2.12537557e-01 3.24163675e-01 -4.34196919e-01 8.13907206e-01
-7.49424040e-01 3.70786309e-01 -1.77484542e-01 -3.55465189e-02
-3.78242522e-01 -7.09043443e-02 -2.86759555e-01 2.79722750e-01
4.35034662e-01 8.17090988e-01 -1.27083421e+00 6.02841303e-02
8.26564357e-02 3.27273220e-01 -5.96488774e-01 2.87060350e-01
-7.63994336e-01 3.50840032e-01 -5.05264878e-01 7.68769503e-01
6.92699134e-01 -5.52496135e-01 7.43633449e-01 -3.28622729e-01
-2.57681698e-01 8.26955259e-01 -7.56728530e-01 1.61587894e+00
-2.81604469e-01 -3.95618454e-02 -5.08564174e-01 -9.55675721e-01
1.07856631e+00 2.30188832e-01 2.68070906e-01 -5.51988125e-01
-1.51453465e-01 2.74237990e-01 3.59711975e-01 -2.58578449e-01
2.77130991e-01 -3.13678026e-01 3.51367384e-01 1.90714926e-01
2.54243553e-01 9.70065147e-02 5.61860919e-01 2.90610164e-01
1.00855446e+00 4.83110219e-01 7.53329754e-01 -1.36245405e-02
9.05800998e-01 -1.68770134e-01 3.10901403e-01 4.92755026e-01
5.73965490e-01 2.13495463e-01 9.81844962e-01 -6.21377349e-01
-7.52561569e-01 -8.40777278e-01 2.73109972e-01 1.35699761e+00
-9.90476385e-02 -8.72961402e-01 -4.86312628e-01 -9.58589911e-01
1.71917991e-03 5.65200925e-01 -6.09815001e-01 -3.93194824e-01
-3.39302778e-01 -7.12797165e-01 1.14517234e-01 3.90671670e-01
1.35637522e-01 -1.17071056e+00 -1.09328352e-01 3.93576384e-01
3.27191025e-01 -7.66359568e-01 -3.61097425e-01 7.44135201e-01
-8.24595690e-01 -1.31680775e+00 -4.56931949e-01 -5.71109474e-01
6.61077142e-01 3.41218412e-01 9.29544985e-01 2.05407381e-01
-4.77856882e-02 -5.07392645e-01 -3.77387524e-01 -3.71300071e-01
-4.82215047e-01 3.28336865e-01 -2.15177611e-01 1.00335129e-01
-9.31065157e-03 -5.97281992e-01 -6.03872240e-01 8.01086575e-02
-7.55120039e-01 3.32821697e-01 7.97927916e-01 7.89954901e-01
8.44307601e-01 -4.57020253e-01 5.06387830e-01 -1.23283243e+00
2.68618166e-01 -6.79025590e-01 -5.62334478e-01 3.21421653e-01
-5.81090450e-01 3.92134309e-01 1.03826940e+00 -3.59684438e-01
-3.32180619e-01 4.67966765e-01 -2.66110629e-01 -3.56340200e-01
1.88919693e-01 1.12594581e+00 -4.47044194e-01 4.24964651e-02
5.67318022e-01 2.54191428e-01 -4.22503799e-02 -4.54743505e-01
4.68072653e-01 2.70710647e-01 2.76123613e-01 -5.53053141e-01
5.35357654e-01 1.37410369e-02 3.96464378e-01 -8.54748905e-01
-6.95330381e-01 -4.08004612e-01 -6.62662625e-01 2.11133704e-01
8.19779694e-01 -8.07089090e-01 -8.07262361e-01 1.50074298e-02
-1.16201532e+00 -2.04710498e-01 6.07789084e-02 2.47116745e-01
-3.04336041e-01 4.97771561e-01 -3.95864904e-01 -6.27490699e-01
-4.40292865e-01 -1.41789758e+00 8.92478943e-01 9.30377692e-02
-1.29744500e-01 -7.85752356e-01 -5.46566024e-02 2.60519713e-01
-1.44272819e-01 4.10323858e-01 1.59854960e+00 -1.14759016e+00
-6.77104414e-01 -3.10827252e-02 -7.34329224e-02 2.61406060e-02
3.12396228e-01 9.69131961e-02 -9.83767867e-01 -3.39672327e-01
-6.09384060e-01 -3.40453357e-01 1.46057332e+00 2.49908090e-01
1.19819903e+00 -2.36504957e-01 -6.44852936e-01 6.28097057e-01
1.25362980e+00 3.54222476e-01 7.94323146e-01 2.09740087e-01
7.59614050e-01 6.40425920e-01 2.68669248e-01 2.20530093e-01
-7.99864065e-03 6.68936968e-01 7.53116965e-01 -1.40425771e-01
1.46357521e-01 -7.68673360e-01 6.63906336e-01 3.75270367e-01
-3.55174363e-01 -5.93411326e-01 -5.96763611e-01 -2.47559305e-02
-1.76089990e+00 -9.83947456e-01 -2.03418300e-01 2.41490293e+00
8.71217489e-01 2.09670991e-01 3.28816056e-01 1.72111630e-01
5.39752662e-01 4.03525472e-01 -8.38627458e-01 -6.45698845e-01
1.04628064e-01 5.02116501e-01 3.78723383e-01 4.81626540e-01
-1.05576432e+00 1.22657788e+00 5.38610888e+00 9.20944154e-01
-1.30765867e+00 -3.96866024e-01 6.41200542e-01 3.34140956e-01
-4.47053403e-01 3.17881584e-01 -8.59146237e-01 3.61910999e-01
8.56908798e-01 1.62821099e-01 3.41899097e-01 8.44566643e-01
2.19615903e-02 1.79797515e-01 -1.49574733e+00 7.10477650e-01
-2.00147495e-01 -1.98761666e+00 7.37103224e-01 1.56812131e-01
3.83146793e-01 -1.72070503e-01 -7.90536180e-02 3.73945981e-01
9.64597836e-02 -1.45162976e+00 4.87553328e-01 2.43810847e-01
7.41856277e-01 -7.07448721e-01 4.58123982e-01 3.10064554e-01
-1.44803798e+00 1.63584873e-01 -3.56494993e-01 1.35431131e-02
-1.64647058e-01 5.47436953e-01 -1.13307071e+00 1.26697981e+00
-1.40864164e-01 1.12260628e+00 -3.18694770e-01 7.22730219e-01
-5.47459126e-01 3.99389684e-01 -2.33009905e-02 -2.00490698e-01
3.73897284e-01 -6.96402311e-01 3.06450576e-01 1.28671670e+00
-1.07801501e-02 1.10755451e-01 5.68270802e-01 8.97420943e-01
-4.52372938e-01 2.33126760e-01 -6.48756683e-01 -4.35355961e-01
2.45119810e-01 1.28787243e+00 -7.20951319e-01 -4.06608224e-01
-3.49545956e-01 8.61498952e-01 3.61762822e-01 3.69106680e-01
-6.00419462e-01 -3.36171746e-01 5.66064179e-01 1.00752890e-01
5.52658796e-01 4.27761823e-02 2.70876456e-02 -8.11813414e-01
-4.22862083e-01 -9.55834746e-01 6.27486646e-01 -5.75092137e-01
-1.12164378e+00 6.17504716e-01 -2.94500172e-01 -9.71351147e-01
-5.95631339e-02 -1.03826869e+00 -7.39777803e-01 1.05604422e+00
-1.80079401e+00 -1.01875877e+00 -2.37740669e-02 3.09672087e-01
4.27230090e-01 -7.98670948e-02 8.68583024e-01 -3.50601273e-03
-8.06825876e-01 4.66642320e-01 -1.46738142e-01 -2.70781517e-01
4.78252739e-01 -1.21376836e+00 4.80631709e-01 4.66442227e-01
5.63117974e-02 9.02678251e-01 4.74760413e-01 -8.37288737e-01
-1.61547399e+00 -1.20451641e+00 7.12507546e-01 -1.09251216e-01
4.37441915e-01 -4.74496275e-01 -9.12934542e-01 4.77864414e-01
-1.67943835e-01 -1.85170844e-01 9.40734506e-01 1.21293157e-01
-8.03483605e-01 3.63186002e-01 -9.31616426e-01 4.19962615e-01
1.06796050e+00 -3.75233680e-01 -2.81785876e-01 7.62192786e-01
7.90911555e-01 -5.03778040e-01 -7.74425745e-01 4.76631343e-01
4.82543796e-01 -7.66396582e-01 1.14789987e+00 -1.26343012e+00
8.32706928e-01 -3.55249882e-01 -7.20276237e-02 -1.29789841e+00
-3.85088682e-01 -7.13144779e-01 -2.34832719e-01 6.93745017e-01
9.58766222e-01 -5.78065753e-01 6.77898884e-01 -1.34590551e-01
-2.97972798e-01 -1.25350642e+00 -5.09075761e-01 -7.28049576e-01
-2.77931392e-02 6.82048053e-02 4.94601578e-01 8.18005919e-01
1.56653255e-01 1.07328868e+00 -2.95860887e-01 7.39296898e-02
3.13285291e-02 5.53689659e-01 7.71402180e-01 -1.17819440e+00
-4.67293501e-01 -4.58018810e-01 -1.48604870e-01 -1.37471986e+00
1.23534285e-01 -1.59841835e+00 -2.57321030e-01 -1.65260005e+00
3.69257778e-01 -8.08626413e-02 -2.99948305e-01 6.39788032e-01
-2.57458299e-01 -3.74386400e-01 1.72501713e-01 7.26193935e-02
-1.94068775e-01 4.69415098e-01 1.14741004e+00 -4.26703364e-01
-4.21393096e-01 -1.00917378e-02 -6.56072617e-01 4.93085742e-01
5.62697589e-01 -3.27751279e-01 -3.55576217e-01 1.71334803e-01
4.55067933e-01 1.22084066e-01 -6.53764084e-02 -5.54279387e-01
4.54225913e-02 -2.53790140e-01 4.63734120e-01 -7.24997282e-01
3.87678951e-01 -3.15852255e-01 1.28781125e-01 9.51410353e-01
-2.73170322e-01 -1.90461785e-01 4.59133014e-02 6.97660387e-01
3.66008462e-04 -3.84243608e-01 8.33655536e-01 -3.72436047e-01
-4.38411355e-01 5.87900460e-01 -1.48145869e-01 -5.53694963e-01
8.55181217e-01 -2.74671376e-01 -4.12559152e-01 -5.24814613e-02
-6.30665541e-01 2.61752993e-01 4.26284373e-01 2.21271262e-01
7.19278574e-01 -8.55245054e-01 -3.57531071e-01 1.72165364e-01
1.71136081e-01 -5.71749770e-05 -2.76005536e-01 8.08098733e-01
-4.24886942e-01 4.51024741e-01 2.96601281e-02 -3.19637865e-01
-1.37277913e+00 1.06914055e+00 3.46572876e-01 -7.35853434e-01
-1.75881490e-01 9.01118755e-01 4.94390666e-01 -4.37047392e-01
4.50997725e-02 -4.09685761e-01 -3.76326919e-01 2.86331683e-01
5.15068233e-01 3.94117907e-02 2.60773838e-01 -4.17841792e-01
-5.14659166e-01 5.22978008e-01 -4.19184297e-01 4.09688264e-01
1.35058260e+00 7.73396313e-01 -1.55181065e-01 1.45289347e-01
1.13543057e+00 -2.46018469e-02 -8.11055839e-01 -1.18083164e-01
1.98088542e-01 -4.18969058e-02 -2.58630246e-01 -7.67500103e-01
-7.39502609e-01 1.00456429e+00 8.60559344e-02 -1.17116002e-02
7.75332868e-01 -7.84359425e-02 6.40624583e-01 5.61511040e-01
1.31538674e-01 -5.51823020e-01 3.25891078e-01 6.15266860e-01
9.35141683e-01 -9.30057108e-01 3.62472147e-01 -7.09685147e-01
-6.62912607e-01 1.48056567e+00 3.13029379e-01 1.63118429e-02
2.74357826e-01 -2.97600627e-01 -6.33169115e-01 -4.26203430e-01
-8.27843308e-01 -3.31654370e-01 4.81738597e-01 3.34661782e-01
7.53525019e-01 -2.18519606e-02 -3.26338768e-01 4.72001761e-01
3.81825194e-02 -1.70148879e-01 1.33486122e-01 9.98439550e-01
-5.44327497e-01 -1.47369063e+00 1.86484270e-02 4.17885333e-01
-3.06031317e-01 -4.58893061e-01 -1.04715383e+00 6.55532300e-01
1.56968668e-01 9.25156534e-01 -4.78872150e-01 -6.72749698e-01
3.73347402e-01 2.93274313e-01 6.51863277e-01 -9.30148900e-01
-7.95937240e-01 -1.60861500e-02 2.22570986e-01 -4.56130087e-01
-3.02315593e-01 -2.29909346e-01 -1.47716951e+00 -2.13150501e-01
-4.63519186e-01 3.84773016e-01 6.11451507e-01 8.77072096e-01
4.62974370e-01 5.78341067e-01 7.94535577e-01 -9.97006834e-01
-4.07564700e-01 -9.68550920e-01 -1.81794479e-01 1.85259536e-01
2.74656504e-01 -5.50719082e-01 3.64597589e-02 -1.52235106e-01] | [4.9324870109558105, 5.953561782836914] |
fdc4e964-f226-4bbd-a51e-de990e725f32 | you-don-t-know-when-i-will-arrive | 2211.12803 | null | https://arxiv.org/abs/2211.12803v2 | https://arxiv.org/pdf/2211.12803v2.pdf | You Don't Know When I Will Arrive: Unpredictable Controller Synthesis for Temporal Logic Tasks | In this paper, we investigate the problem of synthesizing controllers for temporal logic specifications under security constraint. We assume that there exists a passive intruder (eavesdropper) that can partially observe the behavior of the system. For the purpose of security, we require that the system's behaviors are unpredictable in the sense that the intruder cannot determine for sure that the system will exactly accomplish the task in $K$ steps ahead. This problem is particularly challenging since future information is involved in the synthesis process. We propose a novel information structure that predicts the effect of control in the future. A sound and complete algorithm is developed to synthesize a controller which ensures both task completion and security guarantee. The proposed approach is illustrated by a case study of robot task planning. | ['Xiang Yin', 'Rahul Mangharam', 'Shuo Yang', 'Yu Chen'] | 2022-11-23 | null | null | null | null | ['robot-task-planning'] | ['robots'] | [ 6.96506083e-01 7.27552235e-01 -6.15504384e-02 -1.10611439e-01
-2.13299975e-01 -7.32774198e-01 4.33209598e-01 2.28965972e-02
-1.11184366e-01 6.89589739e-01 -4.48149830e-01 -6.69146776e-01
-7.29011074e-02 -9.18926954e-01 -7.26901412e-01 -2.99246252e-01
-2.09941640e-01 1.13669701e-01 4.34538692e-01 -1.91974461e-01
7.49742389e-02 6.28632605e-01 -1.11222291e+00 -1.33205056e-02
1.82077065e-01 7.73708165e-01 1.24600440e-01 8.30459416e-01
5.12362897e-01 9.05533731e-01 -6.89055562e-01 1.88922837e-01
3.28914106e-01 -3.75614852e-01 -1.02231050e+00 2.54344612e-01
-4.79253024e-01 -4.79668021e-01 -1.18136935e-01 1.34002578e+00
-4.15616393e-01 -1.62489414e-01 3.31244946e-01 -1.73048949e+00
2.85626680e-01 7.85257280e-01 -5.79898879e-02 -4.67736900e-01
4.67886329e-01 4.81172860e-01 6.54725909e-01 1.00821733e-01
7.60228872e-01 9.73983943e-01 -9.16955769e-02 8.73508215e-01
-1.24716747e+00 -6.92741871e-01 4.30875033e-01 -1.42940775e-01
-1.64678562e+00 -4.39747334e-01 8.62624407e-01 -4.90905225e-01
9.56649303e-01 3.30004245e-01 1.13406450e-01 9.33918476e-01
9.85874772e-01 3.72405320e-01 9.33752120e-01 -4.05671775e-01
5.07682562e-01 3.84160995e-01 1.46843627e-01 4.85398531e-01
4.12155360e-01 7.18155980e-01 -5.75470515e-02 -4.80048746e-01
3.68507594e-01 -1.12230361e-01 -2.42953107e-01 -2.87556082e-01
-1.09053755e+00 3.01412791e-01 -2.96410501e-01 3.02555412e-01
-3.18384022e-01 4.44743425e-01 2.22405553e-01 7.46716559e-01
-2.54232168e-01 5.60420394e-01 -4.85970587e-01 -7.46951848e-02
-2.98207313e-01 3.07671189e-01 1.11653519e+00 1.31267428e+00
2.73975343e-01 3.93875480e-01 3.62274557e-01 -6.51277006e-01
2.84580082e-01 5.73834598e-01 -2.45846361e-01 -9.68902767e-01
3.38488013e-01 3.86474490e-01 7.82462418e-01 -9.38888073e-01
-1.28576651e-01 9.27522853e-02 -2.07025230e-01 8.60297501e-01
1.94791839e-01 -4.49276209e-01 -6.02302670e-01 1.78066456e+00
3.14741373e-01 1.72860548e-01 4.53841776e-01 5.94888628e-01
-2.83178300e-01 8.77661407e-01 -1.32318005e-01 -7.69131005e-01
1.04798007e+00 -1.40877426e-01 -1.15858674e+00 -2.40213975e-01
4.12377119e-01 -4.85314101e-01 3.53586644e-01 6.21083200e-01
-1.10158777e+00 -1.94867492e-01 -1.63123500e+00 8.97786260e-01
-7.04807714e-02 -9.32724029e-02 3.94508094e-01 6.93634391e-01
-7.71091342e-01 2.82797277e-01 -1.35257661e+00 5.23778498e-02
-4.27756011e-01 7.52482772e-01 -3.18232030e-01 1.53421342e-01
-1.00237918e+00 8.27728987e-01 6.33433104e-01 1.94394425e-01
-1.49945390e+00 -3.77650782e-02 -8.90507221e-01 1.83411717e-01
9.73148704e-01 -1.78006113e-01 1.58710885e+00 -7.04158902e-01
-1.66346920e+00 1.87157124e-01 2.04499643e-02 -6.31968200e-01
5.50916612e-01 2.72359908e-01 -5.51297486e-01 1.54243065e-02
-3.73766869e-01 -2.27155387e-02 1.09545743e+00 -1.22104609e+00
-8.22519243e-01 -4.64312434e-01 7.20990360e-01 -2.96598554e-01
9.98778716e-02 -7.69432113e-02 1.71851158e-01 1.15257576e-01
-9.47538242e-02 -1.21842694e+00 -4.82811719e-01 -2.41584927e-01
-5.04516006e-01 2.25981817e-01 1.02034569e+00 -6.61517829e-02
1.23160803e+00 -2.18027115e+00 -1.30993739e-01 4.98678327e-01
-3.16077694e-02 1.49438843e-01 1.93778962e-01 8.13989758e-01
-8.82960260e-02 1.19072288e-01 -1.96004003e-01 4.11140360e-02
7.11544752e-02 2.46992901e-01 -7.87147403e-01 7.49612153e-01
1.95614383e-01 3.24704230e-01 -5.58505416e-01 -2.38089561e-01
1.73348919e-01 -2.33344976e-02 -4.05119210e-01 3.35764468e-01
-7.26672530e-01 5.68007648e-01 -9.42605793e-01 2.37980157e-01
4.54373717e-01 1.47716254e-01 7.89439261e-01 5.12151957e-01
-3.74136895e-01 9.11251009e-02 -1.42864144e+00 9.29931641e-01
-5.41203618e-01 1.51030749e-01 4.74280506e-01 -5.13727307e-01
6.89165831e-01 9.83769715e-01 2.40786783e-02 -5.41161094e-03
6.13734007e-01 1.25537455e-01 9.77864638e-02 -6.00559898e-02
3.03036302e-01 -3.44197154e-01 -5.67709267e-01 6.87683642e-01
-5.23040056e-01 -3.31840724e-01 -2.41178393e-01 -1.08896531e-01
1.38415337e+00 -2.02545777e-01 6.50027215e-01 -1.11827500e-01
8.92592967e-01 2.88342446e-01 7.59152591e-01 7.20199049e-01
-1.61456868e-01 -3.42837542e-01 7.53929913e-01 -1.84066206e-01
-6.77875698e-01 -5.80275714e-01 4.72098857e-01 1.25634938e-01
2.74985731e-01 -2.95038313e-01 -6.19125307e-01 -6.80891395e-01
-2.67481416e-01 1.09719813e+00 -4.66488153e-01 -5.45585513e-01
-5.26180267e-01 2.46347293e-01 2.31397584e-01 7.01376498e-02
2.04248637e-01 -1.04866493e+00 -1.34854150e+00 3.73697907e-01
3.66619118e-02 -1.28058422e+00 -4.36107665e-01 2.82876998e-01
-6.56717837e-01 -1.29191661e+00 5.06559014e-01 -4.86189455e-01
9.80502248e-01 -5.39148003e-02 2.65832931e-01 2.32912198e-01
-8.69336911e-03 3.84519398e-01 -4.39484604e-02 -4.12452340e-01
-1.11415684e+00 -3.23375046e-01 1.49137542e-01 -1.18585885e-01
-1.56387448e-01 -1.70514449e-01 3.13552953e-02 2.00560480e-01
-9.92581844e-01 -1.72148228e-01 2.40460914e-02 5.10023355e-01
3.00699025e-01 1.24921894e+00 2.49928996e-01 -9.17642653e-01
7.24601269e-01 -1.31674483e-01 -1.49212885e+00 2.64357060e-01
-6.50134742e-01 1.75740555e-01 1.07930100e+00 -3.69870871e-01
-1.08460748e+00 6.82069361e-01 3.38360995e-01 -8.35541859e-02
-4.40451115e-01 3.36879760e-01 -5.71391165e-01 -1.17152095e-01
1.76332816e-01 2.06271172e-01 6.17994480e-02 5.01745716e-02
-1.78848475e-01 3.99368167e-01 2.79333353e-01 -6.03432834e-01
9.92033303e-01 4.43207592e-01 4.96732175e-01 -5.51281095e-01
2.19062008e-02 -1.31621035e-02 -2.32195243e-01 -2.81183571e-01
5.38800478e-01 -3.82979929e-01 -1.70132899e+00 2.38090619e-01
-1.32341015e+00 -2.16882914e-01 -1.71424318e-02 7.35994577e-02
-8.98376644e-01 4.03344072e-02 -1.45893320e-01 -1.63607395e+00
9.72719416e-02 -1.40596938e+00 6.06050551e-01 -2.27459684e-01
-4.46158409e-01 -5.22456765e-01 -2.58502401e-02 -2.14268535e-01
1.32552654e-01 5.40396988e-01 9.39097881e-01 -6.70818388e-01
-1.02416539e+00 -6.63059771e-01 5.46607494e-01 4.90142405e-02
2.92661369e-01 8.33597630e-02 -6.48330390e-01 -4.40551370e-01
5.47554493e-01 1.37794182e-01 -2.96940476e-01 7.01918826e-02
5.35930932e-01 -7.28930831e-01 -4.39856082e-01 -9.96565595e-02
1.47614729e+00 1.14055264e+00 5.04872561e-01 -1.21944368e-01
2.38584378e-03 7.84235120e-01 1.11064661e+00 5.10903478e-01
4.58742455e-02 5.91659069e-01 6.59242749e-01 6.89771056e-01
6.19943738e-01 -2.52840787e-01 5.67688465e-01 4.75361906e-02
4.80490983e-01 -3.63490820e-01 -8.79653037e-01 3.48865539e-01
-1.78164291e+00 -7.12809205e-01 2.12123126e-01 2.33251286e+00
5.82103431e-01 5.27981102e-01 -2.97426134e-01 5.76498270e-01
5.33946514e-01 -8.82727653e-02 -4.68294293e-01 -6.50885701e-01
7.34539330e-01 -8.24693516e-02 6.38214111e-01 1.20583868e+00
-8.23075712e-01 9.14441049e-01 6.12877321e+00 2.57447153e-01
-1.00403929e+00 -2.62635946e-01 1.42852157e-01 4.33778048e-01
-1.09144725e-01 3.70960712e-01 -7.81729043e-01 1.98972803e-02
1.32159305e+00 -6.37826860e-01 4.46107298e-01 9.06789124e-01
2.59237409e-01 -3.60430568e-01 -1.30855203e+00 1.40009806e-01
-3.76655221e-01 -9.35099304e-01 -4.65589792e-01 -1.38623327e-01
2.07124248e-01 -8.15087736e-01 1.77535906e-01 -6.60838559e-02
6.44238353e-01 -6.80115938e-01 9.68053281e-01 3.02192003e-01
6.07495368e-01 -1.25040948e+00 4.64128286e-01 1.03055930e+00
-1.17837977e+00 -4.18367177e-01 1.24515161e-01 -3.54957193e-01
4.14029568e-01 -5.72462566e-02 -1.34109533e+00 5.58803618e-01
-4.30045836e-02 -9.60883871e-02 1.60680607e-01 4.56129372e-01
-4.76368964e-01 1.66715577e-01 -2.89732128e-01 -1.52996719e-01
1.09934501e-01 -1.08218119e-02 7.45170891e-01 5.18845677e-01
3.16431284e-01 5.73614180e-01 2.75289387e-01 8.92462254e-01
6.23293042e-01 -6.01595938e-01 -1.08920360e+00 -4.09143746e-01
4.92942870e-01 5.28569996e-01 -7.57768571e-01 -4.30261344e-02
-5.96655048e-02 7.84078181e-01 -2.46946692e-01 3.41653019e-01
-8.01957250e-01 -3.32634926e-01 6.31419957e-01 -4.37845849e-02
1.39041409e-01 -7.34337449e-01 -2.33562425e-01 -4.52794492e-01
5.89431301e-02 -8.47606599e-01 1.87213629e-01 -4.76020813e-01
-4.58913952e-01 7.91007638e-01 9.95955244e-02 -1.16387463e+00
-6.34030819e-01 -4.27367210e-01 -3.01483691e-01 4.97419417e-01
-1.15720463e+00 -8.09237719e-01 2.86012053e-01 7.09759235e-01
2.26691931e-01 2.97369640e-02 1.11068797e+00 -3.91580075e-01
-5.37199914e-01 8.22795704e-02 -7.19048083e-01 -4.19762760e-01
1.09836131e-01 -7.63695538e-01 2.38935381e-01 1.27792895e+00
-4.39192623e-01 7.93050349e-01 1.34730959e+00 -1.07563472e+00
-2.04539537e+00 -9.81270969e-01 9.82411504e-01 -2.44553909e-01
8.69315147e-01 -3.41056317e-01 -2.92855561e-01 1.15818620e+00
-9.18141007e-02 -1.85927927e-01 1.64823964e-01 -6.16956949e-01
-7.13957697e-02 -8.77032951e-02 -1.43017519e+00 8.17555308e-01
5.00295997e-01 -5.65175116e-01 -4.80844498e-01 1.26770496e-01
1.07104266e+00 -4.21125233e-01 -5.14873505e-01 4.13888663e-01
2.91416079e-01 -5.39843261e-01 4.68537271e-01 -5.99584639e-01
-1.97566867e-01 -6.69217169e-01 -1.05344661e-01 -7.58303702e-01
1.02746978e-01 -1.18944430e+00 -8.15560967e-02 7.75376022e-01
5.26550531e-01 -7.41131186e-01 8.26176643e-01 1.28233027e+00
1.79452389e-01 5.32014333e-02 -1.22819543e+00 -9.90552902e-01
-3.31483245e-01 -7.15479553e-01 7.26541340e-01 4.91784781e-01
7.56671906e-01 -2.46503782e-02 -7.72300780e-01 9.82859671e-01
3.41264248e-01 1.38103515e-01 5.58495522e-01 -7.24983692e-01
-2.75210172e-01 3.03427666e-01 -1.48385212e-01 -5.77950060e-01
6.01616442e-01 -2.54449666e-01 7.32274950e-01 -7.23052442e-01
-2.19495818e-01 -2.05202371e-01 3.03338934e-02 4.32363570e-01
4.72271204e-01 -7.22434700e-01 3.36636662e-01 -2.34024540e-01
-4.47995752e-01 3.01580518e-01 1.04215825e+00 -1.84327662e-01
-2.71202385e-01 7.29877889e-01 -2.71980703e-01 5.01139104e-01
9.43965614e-01 -5.68882942e-01 -8.42599690e-01 -1.41813494e-02
1.27740338e-01 1.34056938e+00 4.44189794e-02 -1.00566220e+00
4.60684597e-01 -8.04323673e-01 -6.09006286e-01 -4.47494805e-01
4.26249981e-01 -1.98024297e+00 8.94448102e-01 1.16188180e+00
-4.71211493e-01 1.55851245e-01 1.96107715e-01 8.05782735e-01
-4.98145968e-02 -3.05189341e-01 5.48627615e-01 1.63741112e-01
-4.49443907e-01 2.66513050e-01 -1.10738683e+00 -8.80496979e-01
1.60150683e+00 1.58978879e-01 -2.28093266e-01 -5.27308583e-01
-9.67868865e-01 4.06672150e-01 5.14034331e-01 2.64659971e-01
8.61184299e-01 -7.68435597e-01 5.74896932e-02 3.41550559e-01
1.61379874e-02 -4.43275899e-01 4.72358130e-02 3.45075548e-01
-1.57975256e-01 7.02629805e-01 -1.81316078e-01 -2.19556559e-02
-1.56652272e+00 1.08657992e+00 1.17550969e-01 -2.47939572e-01
-3.48903716e-01 1.15381628e-01 9.12621096e-02 -1.18533336e-02
2.70694315e-01 -4.08289105e-01 -1.00025989e-01 -7.33766258e-01
7.23812401e-01 -1.15826139e-02 -2.74565250e-01 -2.41609737e-01
-5.57663739e-01 -4.52464931e-02 4.08368409e-02 -5.62129200e-01
9.33116615e-01 -3.19605798e-01 -2.77848691e-01 5.24779797e-01
6.50247395e-01 -2.14221422e-02 -1.09743786e+00 -8.92908648e-02
1.75653383e-01 -5.17109275e-01 -7.01047704e-02 -5.15917957e-01
-7.29571283e-01 4.26655054e-01 2.48876706e-01 4.24320102e-01
1.04510117e+00 -4.55883056e-01 5.58112741e-01 6.45018280e-01
1.24230158e+00 -8.60279977e-01 -2.31909081e-01 7.09016085e-01
6.59903347e-01 -6.35414720e-01 -1.23771504e-01 -8.41327965e-01
-5.68875372e-01 1.06485403e+00 6.42930031e-01 -8.69503245e-02
6.92810833e-01 1.00542951e+00 -2.46873498e-01 -8.37418810e-02
-1.10292721e+00 3.05778533e-01 -3.09036314e-01 6.28925622e-01
-3.34122181e-01 2.67928839e-01 -2.73900867e-01 6.72912717e-01
-1.18941516e-02 1.63855329e-01 1.04231596e+00 1.54403746e+00
-6.00061476e-01 -1.21505392e+00 -5.17295063e-01 -3.49320918e-01
-5.01539588e-01 4.01644886e-01 -4.98986423e-01 8.59040618e-01
-2.04145387e-01 1.38098598e+00 -3.70237172e-01 -5.88964522e-01
4.11092728e-01 5.91741577e-02 2.94884920e-01 -8.61792564e-01
-2.23675951e-01 7.76253790e-02 4.54564452e-01 -8.87928188e-01
-1.30726412e-01 -4.84868467e-01 -1.37217760e+00 -1.61023825e-01
-2.36733630e-01 3.19452345e-01 5.26384890e-01 8.65830839e-01
8.48877057e-03 6.41356349e-01 1.02214932e+00 -1.20344549e-01
-6.07938886e-01 -2.96482533e-01 -6.60936654e-01 -3.82197082e-01
6.90880001e-01 -4.64670509e-01 -4.33492780e-01 1.69350639e-01] | [4.886463165283203, 2.2774369716644287] |
53b0ef83-9167-45b7-a833-e85773242c07 | spherical-space-feature-decomposition-for | 2303.08942 | null | https://arxiv.org/abs/2303.08942v1 | https://arxiv.org/pdf/2303.08942v1.pdf | Spherical Space Feature Decomposition for Guided Depth Map Super-Resolution | Guided depth map super-resolution (GDSR), as a hot topic in multi-modal image processing, aims to upsample low-resolution (LR) depth maps with additional information involved in high-resolution (HR) RGB images from the same scene. The critical step of this task is to effectively extract domain-shared and domain-private RGB/depth features. In addition, three detailed issues, namely blurry edges, noisy surfaces, and over-transferred RGB texture, need to be addressed. In this paper, we propose the Spherical Space feature Decomposition Network (SSDNet) to solve the above issues. To better model cross-modality features, Restormer block-based RGB/depth encoders are employed for extracting local-global features. Then, the extracted features are mapped to the spherical space to complete the separation of private features and the alignment of shared features. Shared features of RGB are fused with the depth features to complete the GDSR task. Subsequently, a spherical contrast refinement (SCR) module is proposed to further address the detail issues. Patches that are classified according to imperfect categories are input to the SCR module, where the patch features are pulled closer to the ground truth and pushed away from the corresponding imperfect samples in the spherical feature space via contrastive learning. Extensive experiments demonstrate that our method can achieve state-of-the-art results on four test datasets and can successfully generalize to real-world scenes. Code will be released. | ['Luc van Gool', 'Radu Timofte', 'Yulun Zhang', 'Shuang Xu', 'Chengli Tan', 'Xiang Gu', 'Jiangshe Zhang', 'Zixiang Zhao'] | 2023-03-15 | null | null | null | null | ['depth-map-super-resolution'] | ['computer-vision'] | [ 6.52518988e-01 -1.75791726e-01 2.03454912e-01 -4.50794011e-01
-1.35956478e+00 3.75140039e-03 3.11295539e-01 -5.58180690e-01
-2.05929443e-01 7.76225388e-01 2.62770236e-01 5.80354154e-01
-1.67257085e-01 -8.97328854e-01 -5.80078959e-01 -1.16499996e+00
2.65313447e-01 -1.63257286e-01 4.07855451e-01 -2.80998617e-01
4.09478635e-01 5.46659589e-01 -2.05281568e+00 6.46253705e-01
9.10493851e-01 1.27782905e+00 5.58102310e-01 4.89185065e-01
3.36894654e-02 7.21785545e-01 -1.74802944e-01 1.70769393e-01
2.43158996e-01 -2.89086014e-01 -7.32871294e-01 1.97037399e-01
5.51910520e-01 -6.06257856e-01 -2.98892319e-01 1.17119324e+00
4.03284758e-01 1.77239269e-01 3.96311045e-01 -8.16585541e-01
-6.08005166e-01 -7.17420280e-02 -7.38061547e-01 1.56884417e-01
3.83264422e-01 -4.95816097e-02 7.04995215e-01 -1.36996531e+00
6.10185385e-01 1.13051236e+00 2.57940173e-01 4.42605942e-01
-1.06389582e+00 -6.80345714e-01 3.30260433e-02 2.62627423e-01
-1.41990614e+00 -5.29280961e-01 1.08341539e+00 -1.65402636e-01
6.59838557e-01 2.94085622e-01 4.79197353e-01 7.77807593e-01
4.86006625e-02 5.33172131e-01 1.41982341e+00 -2.66523272e-01
1.30590796e-01 5.02536334e-02 -2.06599146e-01 5.12468219e-01
-1.00435138e-01 2.86697209e-01 -9.23253298e-01 1.34840906e-01
1.12069404e+00 1.39696181e-01 -6.55976474e-01 -2.41744548e-01
-1.11472309e+00 6.84714019e-01 7.59249568e-01 2.64854908e-01
-5.11457920e-01 -1.64583653e-01 -1.65984869e-01 2.18787715e-02
7.36385524e-01 2.07453400e-01 -5.16589880e-01 1.62805289e-01
-7.43172169e-01 1.23801492e-01 1.58082306e-01 8.44176173e-01
1.38522863e+00 -2.71654904e-01 -6.10349402e-02 9.96634722e-01
2.00828522e-01 4.67042327e-01 3.13429683e-01 -1.13608837e+00
3.76403809e-01 5.75417101e-01 1.32309616e-01 -1.07926714e+00
-3.11707199e-01 -3.31686467e-01 -1.07442522e+00 1.78878844e-01
6.97692856e-02 4.32578713e-01 -9.41121340e-01 1.30275762e+00
5.85691869e-01 2.61090547e-01 1.59846276e-01 1.30156469e+00
1.08543730e+00 6.45749569e-01 -4.34026748e-01 -2.13141069e-01
1.29667687e+00 -7.49146461e-01 -4.42950994e-01 -3.28924924e-01
-5.13814390e-02 -6.63361251e-01 9.80071366e-01 3.74697953e-01
-1.00873554e+00 -6.62775338e-01 -9.54691231e-01 -6.08224034e-01
-1.92482129e-01 6.84897825e-02 4.35403645e-01 2.02330083e-01
-1.01906288e+00 5.73543787e-01 -7.23901212e-01 8.32779333e-02
6.15458786e-01 1.60530299e-01 -4.35865998e-01 -7.11978137e-01
-1.10165262e+00 4.64532733e-01 1.14708766e-01 3.65681082e-01
-7.58663177e-01 -7.16654122e-01 -1.09270620e+00 -2.91584015e-01
1.50579900e-01 -5.73649049e-01 6.28458917e-01 -7.90579975e-01
-1.44976079e+00 8.45538020e-01 -4.92910892e-01 3.61997634e-01
1.42797783e-01 -1.92104559e-02 -3.66696954e-01 5.06703436e-01
2.68983871e-01 4.89436209e-01 1.12426984e+00 -1.60967827e+00
-9.78554308e-01 -7.06553221e-01 -6.47677854e-02 5.39409101e-01
-1.34711396e-02 -6.98009282e-02 -5.31349778e-01 -4.89030540e-01
7.30464578e-01 -4.95975554e-01 -1.29668145e-02 5.01665026e-02
-2.84857959e-01 1.02715313e-01 6.06434107e-01 -8.56057167e-01
7.53305018e-01 -2.39640355e+00 5.48283517e-01 1.13078691e-01
4.31296438e-01 -3.73374730e-01 -2.19586954e-01 -2.45778337e-01
-6.29062727e-02 -3.16188961e-01 -3.80597532e-01 -4.07800019e-01
-4.91333753e-01 7.28882700e-02 -8.57853666e-02 6.44882441e-01
4.71359938e-01 7.60998547e-01 -8.36010575e-01 -4.74678576e-01
4.60453659e-01 7.98116744e-01 -3.70207965e-01 2.52811909e-01
1.34269997e-01 8.87772858e-01 -5.06078720e-01 9.05008376e-01
1.23197091e+00 -2.56442606e-01 -5.08251369e-01 -5.98295927e-01
-3.20302427e-01 6.18792251e-02 -1.25744963e+00 1.78183091e+00
-4.44196820e-01 2.62706846e-01 1.30560920e-01 -6.04254186e-01
9.69240367e-01 -2.56524175e-01 5.82001388e-01 -1.15133369e+00
-1.16483951e-02 2.38043278e-01 -4.85203981e-01 -3.66689742e-01
7.13147819e-01 -2.45766521e-01 1.22895747e-01 1.56377345e-01
-5.67946099e-02 -3.49750191e-01 -4.71465707e-01 -2.00244218e-01
6.92918777e-01 4.27043140e-02 8.12908541e-03 1.14308238e-01
9.11910176e-01 -1.64842874e-01 6.87101483e-01 2.71515846e-01
2.33241338e-02 1.17483640e+00 1.41005600e-02 -3.92762512e-01
-1.00097048e+00 -1.14751899e+00 -3.28809559e-01 8.33742023e-01
4.99164522e-01 -8.01991597e-02 -4.47865486e-01 -2.69707352e-01
-2.31147826e-01 1.40070483e-01 -7.23136902e-01 -1.54120848e-01
-5.63898802e-01 -7.60049939e-01 -1.05041541e-01 4.16208923e-01
8.94735038e-01 -1.02105367e+00 -6.34384632e-01 9.81381629e-03
-5.63318729e-01 -1.31708682e+00 -1.26043916e-01 -2.65099276e-02
-7.07715154e-01 -1.10308695e+00 -7.74999022e-01 -6.75344944e-01
7.52397299e-01 6.68835580e-01 7.41159558e-01 -1.06065243e-01
-4.27256286e-01 1.54082358e-01 -4.59105074e-01 1.88214466e-01
8.14302936e-02 -1.97138488e-01 -2.44804621e-01 3.43113303e-01
1.00599349e-01 -4.74640429e-01 -9.66753423e-01 4.22948331e-01
-1.07729578e+00 3.46140563e-01 7.18937933e-01 1.02078319e+00
1.08712363e+00 3.34059477e-01 2.55726337e-01 -4.89158750e-01
7.10221659e-03 -2.33805463e-01 -4.80590701e-01 -2.69357543e-02
-2.51966298e-01 3.83717567e-02 4.62955356e-01 -1.61642626e-01
-1.30247366e+00 1.03332557e-01 -4.48616296e-02 -3.80946249e-01
-1.78344756e-01 2.36418113e-01 -4.76622432e-01 -4.29244578e-01
3.49105597e-01 7.14349449e-01 -3.05781625e-02 -4.00903434e-01
2.61720479e-01 8.23155761e-01 6.03091002e-01 -3.19269717e-01
8.14870238e-01 1.04051316e+00 3.53022479e-02 -9.02539372e-01
-1.11595619e+00 -5.73383808e-01 -5.75064063e-01 -1.15380116e-01
9.23717856e-01 -1.22573805e+00 -3.17170084e-01 9.08258498e-01
-7.76247501e-01 -2.57206291e-01 -2.66682774e-01 3.49091738e-01
-7.45529711e-01 3.89238328e-01 -6.01448059e-01 -6.12707257e-01
-2.11230919e-01 -1.12109327e+00 1.57025421e+00 6.33698344e-01
4.98053670e-01 -4.46880400e-01 -8.84583071e-02 5.69717646e-01
5.09854555e-01 1.98808149e-01 6.76194966e-01 3.01906943e-01
-1.07334816e+00 2.51593441e-01 -7.16610730e-01 6.81238592e-01
4.33986664e-01 -3.22471142e-01 -1.22628236e+00 -3.41804355e-01
3.27958941e-01 -3.07392091e-01 8.99982274e-01 4.84579951e-01
1.28971696e+00 6.62736595e-02 -9.48988497e-02 1.20973098e+00
1.76839018e+00 -1.99620411e-01 9.39504325e-01 5.81179917e-01
9.63204384e-01 7.49220550e-01 1.09823012e+00 4.20312911e-01
6.46251321e-01 5.97611725e-01 5.50232351e-01 -1.89366564e-01
-3.73603493e-01 -6.43600821e-02 1.92071944e-01 5.01017869e-01
-1.04441002e-01 3.62366349e-01 -5.00510573e-01 4.08005238e-01
-1.41770434e+00 -7.66807973e-01 -1.26812965e-01 2.09147048e+00
8.15400660e-01 -2.80017674e-01 -2.93331295e-01 2.24949062e-01
6.53009415e-01 3.98017734e-01 -7.96171844e-01 2.63714463e-01
-5.72124004e-01 4.00454164e-01 3.81618589e-01 5.36642492e-01
-1.03636301e+00 9.19602275e-01 5.15827465e+00 8.55164349e-01
-1.21503913e+00 8.60142112e-02 7.42774367e-01 -8.30605924e-02
-4.08379436e-01 -1.11250222e-01 -6.64514959e-01 4.40033168e-01
3.33766162e-01 2.29177892e-01 8.26876700e-01 4.69449669e-01
1.24156047e-02 -4.93165433e-01 -7.85388291e-01 1.50108802e+00
3.10097575e-01 -1.08549881e+00 -2.19435394e-01 -1.08069265e-02
9.10387635e-01 2.09972590e-01 2.70495355e-01 -3.68098877e-02
-1.64249286e-01 -1.08920598e+00 4.95123148e-01 9.10942376e-01
1.23485076e+00 -8.97673190e-01 6.81561828e-01 1.24907665e-01
-1.27504790e+00 -1.96289927e-01 -6.79284394e-01 2.06754342e-01
1.17897972e-01 7.40494609e-01 -5.99987470e-02 9.67402101e-01
1.18962097e+00 1.03158152e+00 -6.37164176e-01 6.98344111e-01
-1.45965904e-01 -3.09384614e-01 -1.39882550e-01 5.62878251e-01
-1.80930167e-01 -2.09300622e-01 2.88607150e-01 6.06324613e-01
3.19453001e-01 3.78583252e-01 -2.82671213e-01 1.03373933e+00
2.94569045e-01 -2.33445629e-01 -2.24691838e-01 4.86689538e-01
2.73600727e-01 1.50079417e+00 -3.87116104e-01 -1.15692854e-01
-4.86123443e-01 1.30596650e+00 3.46153855e-01 5.36595821e-01
-5.92924774e-01 -2.99655050e-01 7.50071287e-01 5.01126982e-02
4.11396623e-01 -2.17292868e-02 -3.49150211e-01 -1.36320925e+00
3.21028709e-01 -8.75715494e-01 1.77241147e-01 -1.15412760e+00
-1.35014451e+00 7.23656774e-01 -2.35274479e-01 -1.44141066e+00
7.32277259e-02 -4.70274419e-01 -1.39760062e-01 1.28635132e+00
-2.22171950e+00 -1.20593953e+00 -1.06474900e+00 1.01334012e+00
4.21802193e-01 1.87438548e-01 5.14958143e-01 2.40975529e-01
-3.13585699e-01 2.12774411e-01 7.80968964e-02 -1.77162159e-02
6.79149032e-01 -8.83511484e-01 -1.52975498e-02 7.94135809e-01
-4.67329144e-01 2.37186775e-01 3.50658298e-01 -5.32412171e-01
-1.35214865e+00 -1.21333921e+00 3.61634612e-01 -3.71448040e-01
2.76038378e-01 -2.64113426e-01 -1.05259073e+00 2.51537889e-01
-4.58732933e-01 5.75926900e-01 3.03089231e-01 -4.00924057e-01
-3.38915795e-01 -3.60562533e-01 -1.42776823e+00 1.64959088e-01
1.10147679e+00 -9.04539943e-01 -3.42555583e-01 -2.85974424e-03
7.58181930e-01 -7.55584300e-01 -1.06085050e+00 5.87219656e-01
5.11890233e-01 -1.35669994e+00 1.17006230e+00 4.16537561e-02
7.96241641e-01 -5.75125575e-01 -6.04715645e-01 -1.31845808e+00
-3.39308769e-01 -4.88269329e-02 -4.92657684e-02 1.14908862e+00
-8.31781179e-02 -7.91826129e-01 7.48420835e-01 5.20745575e-01
-1.76403463e-01 -8.12106609e-01 -1.10221839e+00 -3.82593900e-01
-1.66207120e-01 -2.52516121e-01 7.36756444e-01 8.65326226e-01
-4.56371039e-01 1.45028140e-02 -1.91781402e-01 5.92992604e-01
1.09884584e+00 3.99998039e-01 5.95743179e-01 -1.03373384e+00
-5.56548275e-02 -1.65976048e-01 -3.63814831e-01 -9.77375627e-01
-6.40526339e-02 -5.81367731e-01 2.03671679e-01 -1.60944605e+00
4.38975960e-01 -6.53586686e-01 -2.77144760e-01 1.97761562e-02
-3.36592168e-01 4.96180594e-01 -1.45888865e-01 2.27250353e-01
-4.67998654e-01 9.32849228e-01 1.54556060e+00 -2.89176330e-02
-1.51720077e-01 -2.70725369e-01 -8.28040540e-01 4.30223584e-01
5.71009278e-01 -2.75673300e-01 -1.61793262e-01 -4.35806692e-01
1.96575955e-01 1.41214594e-01 6.57226145e-01 -9.30686414e-01
2.32415110e-01 -2.55823791e-01 7.75557756e-01 -8.20285439e-01
7.42311776e-01 -7.99180448e-01 -8.09158105e-03 -7.31040463e-02
7.07141832e-02 -3.61694485e-01 -1.24584720e-01 5.29011548e-01
-4.69506919e-01 3.47017020e-01 9.17767465e-01 -8.52236822e-02
-7.31196702e-01 5.74421585e-01 3.28582019e-01 -1.71594560e-01
8.85315776e-01 -5.05760014e-01 -5.42227328e-01 -1.91104487e-01
-6.03870332e-01 1.84632748e-01 8.10014367e-01 3.18428576e-01
1.30246544e+00 -1.31024218e+00 -8.04336965e-01 7.30572760e-01
4.12440211e-01 8.08304012e-01 9.77413774e-01 8.79861712e-01
-4.27892923e-01 1.41194299e-01 -3.29509169e-01 -9.23694432e-01
-1.05610037e+00 4.09404069e-01 4.88806427e-01 5.11611253e-02
-7.88255930e-01 8.94146204e-01 4.61012542e-01 -4.22621936e-01
-9.90948230e-02 -4.07447278e-01 -1.99446797e-01 -1.09087460e-01
8.50174606e-01 3.92598122e-01 1.02399670e-01 -8.79306972e-01
-3.71944487e-01 1.15249419e+00 -6.70881122e-02 -8.18979442e-02
1.76979041e+00 -6.11545384e-01 -5.04800200e-01 3.36056113e-01
1.43789816e+00 -5.78219481e-02 -1.69256067e+00 -5.73367000e-01
-4.57150400e-01 -1.01131988e+00 4.48175132e-01 -5.21259367e-01
-1.31673157e+00 6.61999524e-01 8.58551443e-01 -1.91285551e-01
1.70287132e+00 3.38813454e-01 6.95453763e-01 -2.57448941e-01
6.14975929e-01 -9.49505329e-01 2.03467950e-01 3.19202960e-01
9.25387979e-01 -1.41746426e+00 1.15002252e-01 -6.91442072e-01
-5.07772148e-01 1.08218110e+00 6.94067001e-01 -2.49178901e-01
5.69408178e-01 3.35941650e-02 -1.38554007e-01 -3.38233948e-01
-4.03553128e-01 -3.21020424e-01 2.70846158e-01 8.32323790e-01
-8.95634368e-02 -1.63547441e-01 2.31900930e-01 6.34212673e-01
-1.14097908e-01 2.69288570e-02 4.55795079e-01 6.84077799e-01
-3.82324904e-01 -6.37619853e-01 -5.04170954e-01 2.90872812e-01
-1.23918466e-01 -1.60369068e-01 1.70126259e-02 5.58888376e-01
2.33986422e-01 9.13802683e-01 2.66845226e-01 -4.78727728e-01
3.27041745e-01 -4.15875018e-01 7.29898632e-01 -5.08496344e-01
6.93036020e-02 8.21972340e-02 -3.15973312e-01 -1.07559288e+00
-6.52671754e-01 -6.12759233e-01 -1.21966779e+00 -4.56414558e-02
-1.89529940e-01 -3.37888271e-01 6.50291145e-01 7.17094719e-01
3.03896219e-01 5.71127832e-01 1.08542085e+00 -1.27862048e+00
-8.38018060e-02 -8.19200456e-01 -7.90994346e-01 2.67496079e-01
8.24981213e-01 -7.49384344e-01 -7.28054702e-01 -2.90556192e-01] | [9.798482894897461, -2.3864047527313232] |
53884ee9-9743-4022-b113-16e1041f16ce | feature-selection-as-a-one-player-game | null | null | https://hal.inria.fr/inria-00484049/ | https://hal.archives-ouvertes.fr/inria-00484049/document | Feature Selection as a One-Player Game | This paper formalizes Feature Selection as a Reinforcement Learning problem, leading to a provably optimal though intractable selection policy. As a second contribution, this paper presents an approximation thereof, based on a one-player game approach and relying on the Monte-Carlo tree search UCT (Upper Confidence Tree) proposed by Kocsis and Szepesvari (2006). The Feature Uct SElection (FUSE) algorithm extends UCT to deal with i) a finite unknown horizon (the target number of relevant features); ii) the huge branching factor of the search tree, reflecting the size of the feature set. Finally, a frugal reward function is proposed as a rough but unbiased estimate of the relevance of a feature subset. A proof of concept of FUSE is shown on benchmark data sets. | ['Michèle Sebag', 'Romaric Gaudel'] | 2010-05-17 | null | null | null | international-conference-on-machine-learning | ['automated-feature-engineering'] | ['methodology'] | [ 2.48818114e-01 3.87301087e-01 -2.52443314e-01 -9.60318279e-03
-9.26404893e-01 -5.83261549e-01 5.10907292e-01 1.56346142e-01
-5.08937836e-01 1.16778874e+00 -3.69044960e-01 -2.01527208e-01
-6.82607114e-01 -9.09792602e-01 -4.45980191e-01 -6.06901467e-01
-3.76284957e-01 5.29554665e-01 1.96456715e-01 -1.89397156e-01
3.95696282e-01 3.36241543e-01 -1.79231513e+00 -2.44447395e-01
9.05751765e-01 1.35514438e+00 3.17054182e-01 6.59410179e-01
3.60536456e-01 3.08221132e-01 -6.12246096e-01 -4.56128389e-01
6.22980595e-01 -3.76417369e-01 -8.73619974e-01 3.04292262e-01
-8.78933892e-02 -3.10278207e-01 2.57324398e-01 1.07542503e+00
3.66453916e-01 3.62383068e-01 5.79650164e-01 -1.28465486e+00
4.38432805e-02 8.63360643e-01 -2.83833772e-01 -5.20511121e-02
4.05122936e-01 2.29106024e-01 1.35628974e+00 -6.15070522e-01
6.03091478e-01 1.03745246e+00 3.18352789e-01 2.79054075e-01
-1.04709625e+00 -3.28539819e-01 2.68078029e-01 2.66711831e-01
-1.33533263e+00 4.34421562e-02 2.83913642e-01 -2.29408696e-01
1.01367509e+00 4.12104309e-01 1.09894848e+00 6.10008121e-01
2.88474381e-01 9.10644531e-01 1.23922646e+00 -7.17807412e-01
1.00789773e+00 -1.42454520e-01 -2.31418461e-01 4.34163362e-01
3.94900441e-01 7.75314689e-01 -3.42859954e-01 -6.52381480e-01
6.30800366e-01 -2.76536256e-01 6.46543801e-02 -7.56957829e-01
-7.36921132e-01 1.20283425e+00 1.93956554e-01 6.17400697e-03
-7.78290272e-01 1.12348117e-01 4.15544689e-01 6.17898762e-01
7.89311081e-02 7.06117928e-01 -5.83228827e-01 -1.68492600e-01
-7.70494521e-01 4.96094644e-01 1.02433062e+00 7.84565091e-01
6.49613261e-01 1.60125569e-01 -3.67163897e-01 2.36213997e-01
8.96997601e-02 3.00868958e-01 5.26895165e-01 -9.05439556e-01
1.71276212e-01 3.43687207e-01 5.55086672e-01 -2.56919622e-01
-2.07884371e-01 -6.49244428e-01 -1.96095705e-01 5.96283257e-01
2.87338585e-01 -5.40371835e-01 -6.04340374e-01 1.70018470e+00
4.34544861e-01 -1.75035372e-02 2.17490718e-01 7.14557469e-01
-1.48977399e-01 3.05812329e-01 3.76001075e-02 -7.14890659e-01
1.10835922e+00 -7.06425190e-01 -4.78280663e-01 3.52389961e-02
3.20226640e-01 -3.06049317e-01 5.99302888e-01 8.47278476e-01
-9.21861053e-01 -2.16919318e-01 -1.06040931e+00 8.48503172e-01
-2.36163318e-01 -1.39202014e-01 1.06878316e+00 1.09752810e+00
-9.69026148e-01 1.01566064e+00 -4.96341497e-01 -2.42601499e-01
3.59305203e-01 5.97022355e-01 -9.91410017e-02 4.11448851e-02
-1.24883938e+00 7.09760725e-01 7.72702694e-01 -1.03374258e-01
-1.03695869e+00 -2.94454217e-01 -6.06919825e-01 2.56115973e-01
1.04431248e+00 -6.31210983e-01 1.48618770e+00 -1.15736270e+00
-1.83531880e+00 2.31502607e-01 3.58806580e-01 -9.04939532e-01
9.65701103e-01 -2.24605680e-01 6.45520687e-02 1.66579351e-01
1.50258899e-01 3.23866725e-01 1.09619176e+00 -7.71767914e-01
-1.07144260e+00 -3.34187955e-01 5.82711585e-02 4.49864060e-01
1.70772076e-01 -2.65607033e-02 2.35841930e-01 -4.99638349e-01
-2.98414618e-01 -8.42212558e-01 -9.37721133e-01 -5.25646627e-01
-2.83961415e-01 -3.29069734e-01 7.89113939e-02 -1.56322762e-01
1.20826626e+00 -1.70873630e+00 7.85939768e-02 4.76626098e-01
-2.10520610e-01 -3.69098014e-03 -9.43716168e-02 6.62874520e-01
3.07494365e-02 -2.90155411e-03 -2.75927126e-01 2.67401308e-01
2.43695185e-01 1.81056783e-01 -2.13975519e-01 3.42216641e-01
1.98027626e-01 6.47698045e-01 -1.08945143e+00 -1.65066406e-01
7.39114657e-02 -2.44486630e-01 -4.81069326e-01 -6.34823218e-02
-3.06975842e-01 -1.14295818e-01 -9.36134458e-01 7.13183582e-01
5.27344823e-01 6.33568922e-03 2.85240978e-01 4.70288008e-01
-3.25555503e-01 -1.26309291e-01 -1.57901847e+00 1.11221290e+00
-1.95851341e-01 -1.91657484e-01 -2.23287210e-01 -8.06073129e-01
9.29673254e-01 2.46144310e-01 5.28218269e-01 -2.75195330e-01
4.59641218e-01 4.22677875e-01 -1.42156899e-01 -4.27587293e-02
4.31166172e-01 -2.69419283e-01 -4.18321937e-01 5.82638621e-01
2.02671126e-01 -4.08024281e-01 4.14975554e-01 -2.54389316e-01
1.27157342e+00 3.63657802e-01 9.94479358e-01 -4.88228947e-01
4.49696004e-01 1.66811924e-02 5.21347225e-01 1.25248230e+00
-2.61566728e-01 3.36110860e-01 6.68822765e-01 -3.80909085e-01
-6.66753173e-01 -7.82377899e-01 -1.63346946e-01 9.80339468e-01
-9.03267562e-02 -1.96481392e-01 -6.76124871e-01 -8.18130970e-01
2.85291374e-01 9.27731156e-01 -8.57722759e-01 -2.91923672e-01
-3.25672477e-02 -6.77315950e-01 -5.17283147e-03 1.78795502e-01
1.60460711e-01 -1.27363491e+00 -1.39934981e+00 3.83201569e-01
3.36477518e-01 -5.83712637e-01 -1.33296356e-01 8.81769717e-01
-9.18579102e-01 -1.18179226e+00 -4.70597357e-01 -3.05492103e-01
2.32554317e-01 -2.67832935e-01 1.08330417e+00 -3.18465084e-01
-3.27727407e-01 3.11405212e-01 -5.92187822e-01 -5.48255861e-01
-3.17977607e-01 -6.28882200e-02 1.23015270e-01 -7.47518381e-03
2.21234858e-01 -3.02258760e-01 -4.29444790e-01 1.05111398e-01
-5.35189211e-01 -3.95025849e-01 7.28536010e-01 1.26185298e+00
7.84496307e-01 3.64149898e-01 7.77155221e-01 -6.33575261e-01
7.97652364e-01 -3.88126224e-01 -1.23469901e+00 4.44698125e-01
-9.41631734e-01 3.08857709e-01 6.43460274e-01 -4.00877625e-01
-9.49603140e-01 4.26852763e-01 2.77000248e-01 -3.19248140e-01
3.74952750e-03 3.73606294e-01 2.90321261e-02 -1.69984385e-01
5.62498212e-01 2.28709102e-01 -1.44561213e-02 -1.60197690e-01
2.49917671e-01 2.86644280e-01 8.37917998e-02 -7.35639393e-01
6.32375300e-01 8.60619172e-02 2.77403563e-01 -2.02432215e-01
-5.34611166e-01 -8.29538107e-02 -4.91710454e-01 -9.91325006e-02
2.69096136e-01 -5.12455940e-01 -1.11677635e+00 1.55552000e-01
-4.29144859e-01 -3.25101197e-01 -9.55223680e-01 5.77446222e-01
-1.22347212e+00 2.05824926e-01 -2.47811720e-01 -1.34876263e+00
-3.40005875e-01 -1.06096160e+00 7.56140947e-01 2.90523082e-01
-1.07084572e-01 -4.19591874e-01 9.42689925e-02 -2.49027565e-01
9.38147753e-02 3.84871721e-01 6.94884598e-01 -8.79680395e-01
-4.55625832e-01 -6.67442739e-01 2.18365073e-01 6.40535280e-02
-2.24293306e-01 -6.57466501e-02 -6.64465308e-01 -3.76356483e-01
1.41713265e-02 -5.44402122e-01 6.83651567e-01 5.92159033e-01
7.36194968e-01 -2.82110363e-01 -6.35030121e-02 2.43127435e-01
1.63332069e+00 4.93251354e-01 1.77582115e-01 6.67799890e-01
-2.45592251e-01 6.48386359e-01 1.46682644e+00 1.29856467e+00
-1.45273417e-01 4.84934032e-01 7.17858553e-01 6.04293525e-01
3.86571616e-01 -3.15910161e-01 6.01018108e-02 -2.38993794e-01
-1.05203673e-01 -8.56139511e-02 -5.39553225e-01 4.53567594e-01
-2.02470374e+00 -8.90512884e-01 4.13993090e-01 2.69120932e+00
5.63815355e-01 1.71182722e-01 4.58836138e-01 2.34644443e-01
7.59959221e-01 -3.85599166e-01 -7.83276737e-01 -6.51171148e-01
-1.56070724e-01 2.78895259e-01 6.77923262e-01 4.44382012e-01
-1.00745130e+00 9.17475045e-01 7.01572037e+00 1.11284494e+00
-6.56415462e-01 -3.53439659e-01 6.58275723e-01 -3.29234675e-02
-1.18003398e-01 1.64328054e-01 -6.52177751e-01 3.23744386e-01
8.18239570e-01 -6.22324884e-01 6.17218792e-01 1.05957007e+00
-1.25583773e-02 -6.33498728e-01 -7.75049508e-01 5.76394498e-01
-4.05679971e-01 -8.75513017e-01 -1.15139604e-01 2.54191667e-01
5.57007611e-01 -1.23788513e-01 1.64342180e-01 5.75846672e-01
8.20148647e-01 -8.20744276e-01 9.93783891e-01 1.73830137e-01
6.27736092e-01 -1.44912422e+00 9.21451569e-01 3.34727377e-01
-9.52507854e-01 -7.29288816e-01 -4.13936406e-01 -2.89489388e-01
-3.37978080e-02 4.50545937e-01 -9.02066231e-01 8.47260296e-01
5.75705469e-01 7.40018114e-02 -5.32852829e-01 1.36178935e+00
-3.16724300e-01 3.61838073e-01 -5.63819051e-01 -5.19077361e-01
4.25955921e-01 -2.28331283e-01 6.66545868e-01 6.84312999e-01
5.66746891e-01 3.60705480e-02 3.31494242e-01 5.42312205e-01
3.65162313e-01 1.95432723e-01 -3.24538022e-01 8.01497027e-02
6.31304920e-01 1.10643494e+00 -9.36199903e-01 -6.93349726e-03
2.67732311e-02 8.09181929e-01 1.82371169e-01 2.01405182e-01
-4.18111205e-01 -4.44237739e-01 2.92716831e-01 -4.78371412e-01
7.55035996e-01 5.14276445e-01 -2.34999284e-01 -8.73062491e-01
-8.00522193e-02 -6.62509680e-01 6.66179478e-01 -2.06679493e-01
-1.19979477e+00 6.45521760e-01 1.61327675e-01 -1.44161427e+00
-7.54475772e-01 -4.53989834e-01 -2.83273458e-01 8.29121470e-01
-1.24257672e+00 -6.20451152e-01 3.63123775e-01 4.27483737e-01
3.64644349e-01 -2.97931850e-01 8.47204089e-01 -5.52091956e-01
-3.41891706e-01 5.41079760e-01 3.06945145e-01 -5.51332831e-01
9.05185491e-02 -1.48124242e+00 2.27877453e-01 6.07672632e-01
-3.33344877e-01 1.79859534e-01 1.04991400e+00 -6.97312653e-01
-1.43745387e+00 -6.90521836e-01 5.93939662e-01 6.84797415e-04
8.07089925e-01 -2.31761318e-02 -3.31611902e-01 3.08477610e-01
-6.05880506e-02 4.44867425e-02 5.50081789e-01 5.68430759e-02
1.29077926e-01 1.51384957e-02 -1.52146161e+00 5.98263919e-01
7.05890179e-01 1.41630873e-01 -3.96225810e-01 6.98648989e-02
4.48801726e-01 -2.39422128e-01 -7.63929367e-01 4.51941639e-01
6.46850467e-01 -1.04900610e+00 5.75594783e-01 -4.83670533e-01
-1.15468659e-01 5.97248562e-02 -1.70590371e-01 -1.36623526e+00
-3.16530854e-01 -1.09745312e+00 -1.63805559e-01 6.46255612e-01
2.36041903e-01 -5.59969902e-01 8.29916894e-01 3.81669194e-01
2.67401755e-01 -1.05871046e+00 -1.36932564e+00 -1.05656171e+00
3.05577293e-02 -3.05153996e-01 7.64013350e-01 4.34753716e-01
2.84290612e-01 1.41663492e-01 -2.51631677e-01 -2.42778808e-01
8.31976593e-01 4.26490366e-01 2.30833068e-01 -1.45985770e+00
-8.01274478e-01 -5.96023202e-01 -2.62997597e-01 -4.19046521e-01
-4.40911353e-02 -3.93218935e-01 1.62302554e-01 -9.20568645e-01
9.16678756e-02 -3.33912343e-01 -3.82757157e-01 3.33917201e-01
-1.12437658e-01 -3.59949917e-01 2.46375561e-01 -2.07782581e-01
-7.84068763e-01 6.66443229e-01 1.10214293e+00 2.28738517e-01
-3.16787124e-01 7.37008631e-01 -6.14765048e-01 4.15764749e-01
7.08238423e-01 -3.78543019e-01 -3.93191397e-01 5.30909956e-01
3.16217124e-01 7.91173339e-01 -1.46860513e-03 -7.47350633e-01
-6.05988838e-02 -4.68927026e-01 2.77417779e-01 -5.98092198e-01
1.25503555e-01 -7.76253164e-01 1.30482465e-02 9.32576060e-01
-4.48385954e-01 -1.14845678e-01 -1.66923236e-02 8.59017909e-01
-1.20901689e-02 -9.10240531e-01 5.63698888e-01 -3.15200567e-01
-6.63656592e-01 1.82025224e-01 -6.18121922e-01 -2.66596615e-01
1.25826800e+00 -4.04526502e-01 2.64602363e-01 -4.25560415e-01
-8.69915485e-01 4.26730007e-01 2.15869352e-01 -4.60587535e-03
5.57464957e-01 -1.09106934e+00 -5.64013839e-01 3.33385244e-02
1.32200956e-01 -3.23317349e-01 1.65727854e-01 3.41553599e-01
-6.92636967e-02 3.79799068e-01 -3.93974751e-01 -4.68728989e-02
-9.56033349e-01 7.75064528e-01 2.89301723e-01 -7.61425614e-01
-3.31515729e-01 7.33164310e-01 -1.50841236e-01 -1.72980905e-01
1.99859098e-01 -2.37084292e-02 -1.97317272e-01 2.53711015e-01
5.55520594e-01 4.77965444e-01 8.08193162e-02 1.17136883e-02
-1.57635510e-01 1.28436789e-01 2.51944996e-02 -4.40381765e-01
1.23209512e+00 -1.57453686e-01 2.95894027e-01 1.52157947e-01
4.93153632e-01 -4.16852683e-01 -1.58381224e+00 -2.31114402e-01
2.66683698e-01 -5.59360623e-01 7.78544247e-02 -1.01528490e+00
-4.84990180e-01 3.38901967e-01 6.77060246e-01 3.94488275e-01
1.24828160e+00 -4.01806563e-01 1.42585620e-01 6.45779252e-01
1.04115784e+00 -1.23887908e+00 -3.26904684e-01 4.25421387e-01
7.30397522e-01 -1.03422797e+00 1.01468131e-01 -1.91237047e-01
-9.29529727e-01 1.30108416e+00 4.75502491e-01 -2.61554301e-01
4.85621095e-01 1.94762692e-01 -5.12891293e-01 5.89980297e-02
-1.07464302e+00 -7.05981255e-01 -1.20163128e-01 7.28326619e-01
-2.72415280e-01 3.19729984e-01 -6.35496736e-01 8.46533000e-01
-1.94476172e-01 1.06317297e-01 4.49789733e-01 1.16927969e+00
-6.76070809e-01 -1.04563129e+00 -4.30414945e-01 6.11567616e-01
-4.04900700e-01 -5.02965450e-02 -2.93591619e-01 8.37806344e-01
-6.15558401e-02 8.53675961e-01 -2.31665030e-01 -2.97990590e-01
3.21740299e-01 -1.80636823e-01 6.32237852e-01 -5.42360425e-01
-7.91077018e-01 3.52703840e-01 -1.51355788e-02 -5.05558848e-01
5.95328175e-02 -9.12077546e-01 -9.52884436e-01 1.13650255e-01
-6.18787408e-01 6.35760725e-01 3.53270769e-01 8.92940283e-01
-5.50737139e-03 3.16324830e-01 9.49698865e-01 -8.47526431e-01
-1.33767903e+00 -8.00283968e-01 -1.08219087e+00 -2.04662770e-01
4.78923619e-02 -8.23252678e-01 -4.06840026e-01 -6.43510282e-01] | [4.267866611480713, 2.3955116271972656] |
93285ae6-3c9f-4ce2-a5b0-3570cab39fd2 | approximating-pareto-frontier-through | null | null | https://openreview.net/forum?id=S9MPX7ejmv | https://openreview.net/pdf?id=S9MPX7ejmv | Approximating Pareto Frontier through Bayesian-optimization-directed Robust Multi-objective Reinforcement Learning | Many real-word decision or control problems involve multiple conflicting objectives and uncertainties, which requires learned policies are not only Pareto optimal but also robust. In this paper, we proposed a novel algorithm to approximate a representation for robust Pareto frontier through Bayesian-optimization-directed robust multi-objective reinforcement learning (BRMORL). Firstly, environmental uncertainty is modeled as an adversarial agent over the entire space of preferences by incorporating zero-sum game into multi-objective reinforcement learning (MORL). Secondly, a comprehensive metric based on hypervolume and information entropy is presented to evaluate convergence, diversity and evenness of the distribution for Pareto solutions. Thirdly, the agent’s learning process is regarded as a black-box, and the comprehensive metric we proposed is computed after each episode of training, then a Bayesian optimization (BO) algorithm is adopted to guide the agent to evolve towards improving the quality of the approximated Pareto frontier. Finally, we demonstrate the effectiveness of proposed approach on challenging high-dimensional multi-objective tasks across several environments, and show our scheme can produce robust policies under environmental uncertainty. | ['Wulong Liu', 'Bin Wang', 'Dong Li', 'Jianye Hao', 'Xiangkun He'] | 2021-01-01 | null | null | null | null | ['multi-objective-reinforcement-learning'] | ['methodology'] | [ 4.42392714e-02 -4.74968813e-02 1.06220260e-01 -9.66177583e-02
-1.07586956e+00 -5.37023306e-01 3.33756328e-01 1.59936771e-01
-6.13659203e-01 1.43003416e+00 1.97601885e-01 5.81791103e-02
-8.94446850e-01 -7.49568701e-01 -7.38625288e-01 -1.19541478e+00
-1.93304002e-01 5.44126987e-01 1.89254787e-02 1.38152193e-03
6.42558217e-01 1.94584981e-01 -1.26476896e+00 -3.40888619e-01
1.30544138e+00 1.11694741e+00 1.36187017e-01 6.74508035e-01
1.95732653e-01 5.12290657e-01 -8.45825911e-01 -3.94221306e-01
1.50987700e-01 1.41318263e-02 -5.82602382e-01 -1.19785242e-01
-4.08209562e-01 -3.67809147e-01 6.91752657e-02 1.39874411e+00
8.22225869e-01 6.76506162e-01 1.06047869e+00 -1.46927440e+00
-4.29635376e-01 6.94075584e-01 -4.81004685e-01 -2.25386038e-01
4.32740115e-02 5.23746014e-01 8.62944424e-01 -3.33821088e-01
1.57127976e-01 1.64868414e+00 8.57552364e-02 4.28612232e-01
-9.18765187e-01 -3.99483651e-01 3.56233150e-01 1.38060167e-01
-1.11757183e+00 3.84111069e-02 6.98101759e-01 -1.57683805e-01
4.14336234e-01 1.38733998e-01 6.51708961e-01 9.05499876e-01
7.84047902e-01 7.25542009e-01 1.19124401e+00 -1.13607779e-01
8.37175786e-01 -4.14665192e-02 -5.73483407e-01 4.54017073e-01
4.72022146e-01 6.43900990e-01 -1.72179833e-01 -3.42080295e-01
1.98623091e-01 -3.00706774e-01 -4.18005437e-02 -3.73697340e-01
-8.35187137e-01 9.02241826e-01 2.36098975e-01 -2.35496238e-01
-6.68180466e-01 2.94653684e-01 2.41412133e-01 1.12117678e-01
8.21785256e-02 6.50705993e-01 -2.50578225e-01 -1.00769728e-01
-4.95122641e-01 5.98595083e-01 5.85145175e-01 6.21372581e-01
3.70734483e-01 5.63047945e-01 -5.22232354e-01 5.49040079e-01
8.41659844e-01 7.19804108e-01 3.23401630e-01 -1.32230592e+00
3.95184308e-01 2.65014827e-01 8.13807487e-01 -1.09016848e+00
-2.74532288e-01 -6.87647641e-01 -5.83196402e-01 8.12470198e-01
1.30973741e-01 -7.12188601e-01 -6.31046116e-01 1.74370539e+00
6.01287842e-01 -7.67133161e-02 4.25639927e-01 8.94103706e-01
2.12191492e-01 8.69719505e-01 1.15164064e-01 -7.05578327e-01
8.58466446e-01 -5.56834102e-01 -8.08766842e-01 -1.15805715e-01
-2.18442649e-01 -3.80565971e-01 6.44625247e-01 6.68235779e-01
-1.13631070e+00 -2.32917681e-01 -1.27401149e+00 9.39016163e-01
-1.54895097e-01 -4.33157682e-01 1.03667825e-01 6.46835387e-01
-5.67270160e-01 6.45337284e-01 -4.16787922e-01 2.53077686e-01
5.60945868e-01 3.02675188e-01 2.50542045e-01 -6.76071346e-02
-1.34429240e+00 9.79012489e-01 1.11410272e+00 1.82885051e-01
-1.59179521e+00 -3.68180305e-01 -5.85817814e-01 9.87859666e-02
9.34907854e-01 -5.57435751e-01 9.73419845e-01 -7.73600698e-01
-1.92215157e+00 -1.85841486e-01 5.08792639e-01 -1.57809541e-01
6.40263557e-01 -1.66642621e-01 -3.65832061e-01 3.46428826e-02
-2.79454023e-01 3.38913590e-01 1.09151089e+00 -1.65084255e+00
-8.86374772e-01 -3.42689365e-01 2.46399254e-01 7.05483377e-01
-1.01282328e-01 -8.31242800e-02 1.71847984e-01 -4.98691589e-01
-4.96148527e-01 -6.64190710e-01 -4.77635711e-01 -6.20938480e-01
-2.66207576e-01 -2.46586457e-01 2.37352535e-01 -6.43059969e-01
1.41478562e+00 -1.83112550e+00 5.82432389e-01 6.08121812e-01
-3.72799397e-01 1.41637802e-01 -2.24668890e-01 4.03824687e-01
5.76333404e-01 1.30008772e-01 -5.82027256e-01 1.05048344e-01
5.10597646e-01 2.28258505e-01 -3.13005507e-01 4.28966701e-01
1.34186611e-01 5.72505116e-01 -1.07141721e+00 -3.92056912e-01
1.32373627e-02 1.02784924e-01 -4.40898150e-01 3.17144364e-01
-6.66910827e-01 4.61644441e-01 -9.16718483e-01 7.60636330e-01
6.60355747e-01 4.05399919e-01 4.77312459e-03 6.13545328e-02
-1.21576995e-01 -6.87689841e-01 -1.73997176e+00 1.38930905e+00
-3.88164639e-01 -2.78656542e-01 1.88585997e-01 -8.57729554e-01
1.05338991e+00 1.17049620e-01 3.40539992e-01 -4.67020184e-01
4.50623840e-01 2.20550746e-01 -1.12939514e-02 -3.81159782e-01
5.13929725e-01 -5.37302718e-02 -2.22405106e-01 3.66251051e-01
-8.89731273e-02 -5.16273439e-01 1.54056728e-01 -2.82549351e-01
7.46717036e-01 4.65105385e-01 3.55736196e-01 -3.98084015e-01
8.10208082e-01 -1.27448171e-01 8.73732805e-01 8.05652142e-01
-3.57811302e-01 -1.77852958e-02 6.30106807e-01 -1.19281404e-01
-7.56832778e-01 -1.19083595e+00 1.25775456e-01 9.10372674e-01
3.15663606e-01 4.04161006e-01 -3.87965888e-01 -6.32135212e-01
2.77560711e-01 1.24845588e+00 -4.76118952e-01 -4.01217639e-01
-1.53601557e-01 -9.83375967e-01 3.65440428e-01 5.21493107e-02
4.30838615e-01 -1.14851475e+00 -7.52594292e-01 5.38116574e-01
1.89776599e-01 -3.96277249e-01 -4.28601950e-01 4.95747589e-02
-6.48617446e-01 -9.64220643e-01 -7.70052314e-01 -3.14597458e-01
4.12526339e-01 -5.96368194e-01 6.74790561e-01 -4.28699702e-01
-1.79921106e-01 4.02634948e-01 -1.47789037e-02 -6.85646534e-01
-3.93658221e-01 -5.02763331e-01 3.32395017e-01 1.19630694e-01
-2.71999538e-01 -4.88653421e-01 -7.11924970e-01 3.38328153e-01
-9.64913905e-01 -7.83767700e-01 7.33090341e-01 8.50584269e-01
7.40761578e-01 6.75262749e-01 1.13079453e+00 -7.73297399e-02
1.16297376e+00 -5.40486276e-01 -1.10639131e+00 5.31053185e-01
-5.53548217e-01 4.89788979e-01 7.90728807e-01 -5.00809848e-01
-1.34867287e+00 -3.83931369e-01 1.10674620e-01 -3.67652535e-01
1.22615777e-01 5.24931490e-01 -3.96907955e-01 1.36943251e-01
3.75005364e-01 3.05446953e-01 4.97204587e-02 5.32584339e-02
3.54042590e-01 6.36229336e-01 4.04635042e-01 -1.20362782e+00
8.32234740e-01 9.26433653e-02 4.28679377e-01 -1.37887090e-01
-6.91357255e-01 1.93563864e-01 -9.72987860e-02 -5.87968886e-01
8.03828537e-01 -5.45383453e-01 -1.33653688e+00 4.58228141e-01
-7.74575353e-01 -8.04973021e-02 -2.24854082e-01 6.60602808e-01
-8.85645151e-01 1.14221200e-01 1.57839864e-01 -1.57359290e+00
-3.05958450e-01 -1.20436728e+00 3.77296984e-01 5.64872324e-01
3.59400690e-01 -9.36739385e-01 3.87883455e-01 2.68861681e-01
2.72387773e-01 8.39891613e-01 1.04563487e+00 -4.86787856e-01
-3.88523251e-01 8.48180577e-02 2.66259909e-01 4.38091427e-01
-1.25213146e-01 1.72740474e-01 -3.98367792e-01 -6.26669347e-01
4.85903062e-02 -7.18335748e-01 5.02322674e-01 5.07568121e-01
1.07752085e+00 -7.16293812e-01 5.89937083e-02 3.78677249e-01
1.69888842e+00 7.81695843e-01 4.05302972e-01 5.55001676e-01
4.44357097e-03 7.45105863e-01 1.23293972e+00 9.41199183e-01
3.01203102e-01 8.42429921e-02 1.07506454e+00 6.85192704e-01
6.74062490e-01 -9.86636803e-02 5.40991783e-01 3.82866561e-01
-9.55409855e-02 -4.02362019e-01 -6.45091951e-01 3.69722933e-01
-2.08725977e+00 -1.11688662e+00 6.79976344e-01 2.35192585e+00
7.67625988e-01 2.74678886e-01 2.71997929e-01 -2.62312263e-01
9.35358167e-01 1.20519534e-01 -1.16864753e+00 -5.87654889e-01
-1.20378546e-01 -2.63445616e-01 4.34569985e-01 6.11742914e-01
-8.64743650e-01 5.31589568e-01 5.59230757e+00 1.19269979e+00
-7.94254541e-01 -2.20354557e-01 6.94989979e-01 -2.35466093e-01
-5.26237488e-01 -2.71290392e-01 -4.52051997e-01 6.68943763e-01
6.05296969e-01 -5.28858542e-01 9.19077039e-01 6.59002185e-01
3.57330114e-01 -2.71424532e-01 -5.14380455e-01 6.03909254e-01
-2.19229653e-01 -8.84220183e-01 2.65011983e-03 -2.20268704e-02
1.07217395e+00 -4.54675436e-01 3.86294574e-01 4.10664469e-01
1.03676307e+00 -1.26375616e+00 6.56391263e-01 9.18025672e-01
3.85591269e-01 -1.73704398e+00 8.59839499e-01 5.05016983e-01
-8.22498083e-01 -7.97254086e-01 -5.43312132e-01 3.80683631e-01
9.10263881e-02 2.63465762e-01 -4.00974452e-01 9.90877569e-01
5.01746297e-01 3.35980654e-02 -1.15900502e-01 1.20657599e+00
-2.74056673e-01 1.41951382e-01 -1.61846057e-01 -4.74125296e-01
6.92493677e-01 -5.76803207e-01 1.07683611e+00 5.97841680e-01
5.89455426e-01 8.22651312e-02 2.93724060e-01 7.47695923e-01
2.06797287e-01 4.61223051e-02 -2.81626284e-01 -1.34529054e-01
7.38906562e-01 1.14347386e+00 -3.10019404e-01 9.60581005e-02
2.68335521e-01 3.59942794e-01 2.10607767e-01 3.86135072e-01
-1.05311358e+00 -5.86015940e-01 4.63080734e-01 -7.61423707e-01
2.88797796e-01 -8.67017452e-03 -4.19243798e-02 -6.23394072e-01
-2.40202636e-01 -9.32633162e-01 6.36398733e-01 -5.73798418e-01
-1.38219380e+00 3.38521838e-01 7.39838779e-02 -1.11135197e+00
-1.92626581e-01 -3.76423061e-01 -8.41840088e-01 7.32315421e-01
-1.60344863e+00 -5.85785568e-01 1.13945618e-01 4.91891414e-01
3.29269469e-01 -7.08775699e-01 3.99012148e-01 -3.07991445e-01
-7.01281607e-01 3.66258174e-01 8.05506766e-01 -5.76376498e-01
6.05262637e-01 -1.32492518e+00 -7.54628241e-01 6.48114800e-01
-6.68875992e-01 1.29281417e-01 9.65779603e-01 -7.41372228e-01
-1.53541303e+00 -1.20915484e+00 -3.14744890e-01 3.41927335e-02
7.58566320e-01 2.99533844e-01 -3.96419138e-01 2.31984444e-03
3.10829341e-01 -4.21585709e-01 3.58144403e-01 -3.77735198e-01
1.23157308e-01 -3.59153807e-01 -1.64664507e+00 8.18926752e-01
5.67284167e-01 3.66015315e-01 -6.15745783e-01 4.79437932e-02
8.61479044e-01 -2.64886945e-01 -1.10244513e+00 6.11692667e-01
5.33632934e-01 -4.92343038e-01 8.53194714e-01 -8.35107028e-01
4.44869369e-01 -5.10596335e-01 -5.36646187e-01 -1.99400973e+00
-4.36464369e-01 -8.28244984e-01 -2.94628471e-01 1.14090800e+00
1.70027807e-01 -6.00258708e-01 2.83330590e-01 1.73164055e-01
-1.64363876e-01 -1.07149708e+00 -1.19531941e+00 -7.98307955e-01
3.52250695e-01 2.50456319e-03 8.67645025e-01 3.69936526e-01
-3.47254485e-01 -2.25799531e-01 -5.45319736e-01 5.21127224e-01
1.22763205e+00 -1.45435715e-02 2.50543654e-01 -8.62137496e-01
-3.87101352e-01 -5.54766119e-01 1.13749810e-01 -3.22711557e-01
2.20995069e-01 -5.26048839e-01 3.66544783e-01 -1.48144221e+00
-1.39739662e-02 -1.58385128e-01 -7.33935177e-01 -9.22433585e-02
-3.65352511e-01 -6.01011693e-01 2.81238496e-01 -1.61695287e-01
-8.58080983e-01 1.30497813e+00 1.51125669e+00 -4.24065560e-01
-1.73793629e-01 2.58472532e-01 -7.93836176e-01 7.42553413e-01
8.75878155e-01 -5.82436442e-01 -5.78985929e-01 -1.07574277e-01
3.26833367e-01 5.45573235e-01 -1.05008138e-02 -9.68891084e-01
1.78798690e-01 -1.05672169e+00 3.82797211e-01 -5.47028482e-01
2.96106070e-01 -9.12009835e-01 5.32132536e-02 7.28825569e-01
-2.84817219e-01 -7.56870508e-02 1.98262334e-01 1.01072025e+00
-6.94025606e-02 -6.05977595e-01 1.02413046e+00 -1.85159415e-01
-4.78139162e-01 3.87145966e-01 -2.80182213e-01 3.22721779e-01
1.33823085e+00 1.69191837e-01 -2.68858969e-01 -3.09240073e-01
-5.93178034e-01 1.02074885e+00 6.20689318e-02 1.01660125e-01
8.17254186e-01 -1.32955098e+00 -1.05552804e+00 -4.63945448e-01
-2.42443681e-01 2.01120898e-02 3.88137013e-01 2.54530311e-01
-3.05700749e-01 3.68132740e-02 -4.81984198e-01 -2.29062319e-01
-8.01381648e-01 5.02555132e-01 3.86741847e-01 -5.29257894e-01
2.57335007e-01 5.02096772e-01 -3.92247885e-01 -6.88940346e-01
5.01549423e-01 2.78740704e-01 -4.14160341e-01 3.61883849e-01
3.48534971e-01 9.32816923e-01 -4.67662156e-01 -1.65686175e-01
-3.23017776e-01 5.39198995e-01 3.92042786e-01 -5.79054356e-01
1.48726583e+00 -1.99338302e-01 -2.39224285e-02 2.61169672e-01
5.96234322e-01 -1.67611495e-01 -1.78597772e+00 -2.03248430e-02
-1.87968522e-01 -5.06590545e-01 9.92428958e-02 -1.27559125e+00
-6.39023304e-01 4.44119364e-01 5.98436415e-01 -1.31620094e-01
1.05066431e+00 -6.70767367e-01 3.35063964e-01 6.24568999e-01
4.36231047e-01 -1.63346088e+00 5.12193322e-01 5.71083188e-01
1.27933669e+00 -1.09361327e+00 1.03436060e-01 5.02679884e-01
-1.08716094e+00 1.19402742e+00 6.52804852e-01 -2.85136729e-01
4.88012373e-01 8.45745429e-02 -2.70927548e-01 2.46988520e-01
-8.19339871e-01 -6.40548766e-02 2.44000778e-01 7.20048666e-01
-5.29502690e-01 2.87733842e-02 -2.91365594e-01 6.31350875e-01
1.05668880e-01 -2.79778600e-01 4.07604963e-01 8.65294158e-01
-9.56692457e-01 -6.77567780e-01 -8.55987430e-01 2.81360745e-01
-3.28154236e-01 2.89352775e-01 1.86811015e-01 4.38434213e-01
1.67962939e-01 1.20501733e+00 -2.74044067e-01 -1.00207865e-01
1.80640936e-01 -1.70641109e-01 4.32286590e-01 -1.21749192e-01
-3.63991469e-01 -3.53461839e-02 -6.98461607e-02 -4.09105331e-01
-7.15874508e-02 -5.61381400e-01 -1.14036095e+00 -4.82743196e-02
-1.07584924e-01 5.24718761e-01 7.01908350e-01 8.53144765e-01
1.34589076e-01 8.70364249e-01 1.05446661e+00 -5.27009666e-01
-1.46312273e+00 -7.55280256e-01 -6.35389388e-01 -1.02927692e-01
2.61560768e-01 -8.85635376e-01 -5.16744792e-01 -6.93174422e-01] | [4.303105354309082, 2.4221835136413574] |
342fd090-e4ad-4d9f-8805-48be75e7aded | association-graph-learning-for-multi-task | 2210.04637 | null | https://arxiv.org/abs/2210.04637v1 | https://arxiv.org/pdf/2210.04637v1.pdf | Association Graph Learning for Multi-Task Classification with Category Shifts | In this paper, we focus on multi-task classification, where related classification tasks share the same label space and are learned simultaneously. In particular, we tackle a new setting, which is more realistic than currently addressed in the literature, where categories shift from training to test data. Hence, individual tasks do not contain complete training data for the categories in the test set. To generalize to such test data, it is crucial for individual tasks to leverage knowledge from related tasks. To this end, we propose learning an association graph to transfer knowledge among tasks for missing classes. We construct the association graph with nodes representing tasks, classes and instances, and encode the relationships among the nodes in the edges to guide their mutual knowledge transfer. By message passing on the association graph, our model enhances the categorical information of each instance, making it more discriminative. To avoid spurious correlations between task and class nodes in the graph, we introduce an assignment entropy maximization that encourages each class node to balance its edge weights. This enables all tasks to fully utilize the categorical information from related tasks. An extensive evaluation on three general benchmarks and a medical dataset for skin lesion classification reveals that our method consistently performs better than representative baselines. | ['Marcel Worring', 'Cees G. M. Snoek', 'XianTong Zhen', 'Zehao Xiao', 'Jiayi Shen'] | 2022-10-10 | null | null | null | null | ['skin-lesion-classification', 'classification'] | ['medical', 'methodology'] | [ 7.02176630e-01 2.42736399e-01 -4.72217023e-01 -4.45244551e-01
-3.99505287e-01 -6.19636893e-01 4.65977341e-01 4.16583508e-01
-2.83156246e-01 8.32282543e-01 1.96566656e-01 -2.71435622e-02
-4.42620784e-01 -5.86587965e-01 -5.61106145e-01 -8.26541066e-01
-1.99322607e-02 3.22349846e-01 -1.03918016e-02 8.34438056e-02
-4.14947048e-03 -2.32637644e-01 -1.13517570e+00 5.79341590e-01
1.01482391e+00 9.84910369e-01 1.82354480e-01 5.43399714e-02
9.73817706e-02 7.68350005e-01 -3.68986189e-01 -7.03009605e-01
1.46558406e-02 -2.75120318e-01 -1.08695078e+00 3.40025395e-01
3.44509631e-01 1.22897223e-01 -2.58381575e-01 9.45092678e-01
2.12510720e-01 1.89132154e-01 1.20875299e+00 -1.72448862e+00
-6.19962931e-01 5.69735944e-01 -8.02674174e-01 -1.00824825e-01
-4.49828841e-02 -1.90606385e-01 1.24521232e+00 -5.67422152e-01
6.24525309e-01 9.31553006e-01 5.65796494e-01 6.19929373e-01
-1.37542593e+00 -6.87919736e-01 4.66783136e-01 3.24917912e-01
-1.22627902e+00 -5.46926036e-02 8.63332212e-01 -6.60389900e-01
4.27327573e-01 1.99710801e-01 4.14672405e-01 1.39773774e+00
2.39053041e-01 8.30156028e-01 1.07656717e+00 -2.17353314e-01
1.23232760e-01 3.55464280e-01 5.31450152e-01 5.53618968e-01
3.47379416e-01 -3.20070475e-01 -5.09010375e-01 -1.24489702e-01
3.05844516e-01 2.15728015e-01 -4.26238179e-01 -8.11127722e-01
-1.29446173e+00 8.27963710e-01 5.69214821e-01 2.76335269e-01
-3.13458353e-01 -2.88742576e-02 5.96151173e-01 4.48964000e-01
8.20975721e-01 1.95836559e-01 -5.28015018e-01 3.94343823e-01
-4.29140151e-01 -5.07332794e-02 8.03254068e-01 8.96225154e-01
8.79204452e-01 -5.84369063e-01 -6.46493554e-01 1.06425595e+00
1.64809138e-01 -4.86473553e-02 4.32363987e-01 -6.33797228e-01
6.66417539e-01 7.63458133e-01 -4.31106448e-01 -9.06877995e-01
-4.88036811e-01 -9.23428357e-01 -1.08836138e+00 -2.67116576e-01
3.35339665e-01 -2.58675009e-01 -8.78298283e-01 2.07886124e+00
2.87899315e-01 3.11430484e-01 -9.97547656e-02 6.11485124e-01
6.21257961e-01 8.39155763e-02 2.75660515e-01 5.52579935e-04
1.28660321e+00 -1.04824746e+00 -6.32625699e-01 -6.15163088e-01
9.23329890e-01 -4.32227910e-01 8.49624455e-01 1.89604938e-01
-4.36754674e-01 -3.76091659e-01 -8.09345007e-01 -8.05527531e-03
-3.05700004e-01 1.16065852e-01 6.98062658e-01 2.13191047e-01
-7.96949089e-01 4.29116845e-01 -4.17798847e-01 -2.74990439e-01
6.60655916e-01 2.99193501e-01 -5.55235028e-01 -3.96993667e-01
-1.13245642e+00 7.09050894e-01 4.85635400e-01 -1.84752271e-01
-7.88200319e-01 -7.29938269e-01 -1.01473582e+00 8.08514133e-02
6.12481117e-01 -8.58504653e-01 9.55295980e-01 -8.02477181e-01
-7.42037475e-01 9.93235886e-01 -1.29475489e-01 -6.17982559e-02
2.75842249e-01 1.41585916e-01 -4.27800755e-04 -2.53084987e-01
2.71911144e-01 6.57016218e-01 8.69271338e-01 -1.46302319e+00
-7.21431375e-01 -5.57803094e-01 1.04936704e-01 3.69204700e-01
-9.10092473e-01 -5.50315976e-01 -4.83740300e-01 -6.53487086e-01
-1.58141751e-03 -1.21211851e+00 -5.05068563e-02 -4.25014086e-02
-7.92183518e-01 -5.77668667e-01 7.12176979e-01 -4.66723859e-01
1.15165699e+00 -2.28446841e+00 6.76102102e-01 2.67485082e-01
6.98753595e-01 -1.42057374e-01 -2.92135388e-01 2.20699772e-01
-9.20911357e-02 6.21932857e-02 -4.08904254e-01 -6.65455341e-01
-1.66999195e-02 5.17213464e-01 3.41404043e-02 2.79853016e-01
1.52351856e-01 9.35666025e-01 -9.85479891e-01 -4.45066452e-01
-1.11264683e-01 3.95269543e-01 -5.23528099e-01 6.57341704e-02
1.19546643e-02 6.66958988e-01 -6.71350658e-01 5.49080849e-01
5.00099480e-01 -6.84301674e-01 3.26624513e-01 -2.97070056e-01
6.25955164e-01 2.24294022e-01 -8.72952044e-01 1.82582307e+00
-7.19158113e-01 4.61047024e-01 8.28965008e-02 -1.48257816e+00
6.81675732e-01 3.11044127e-01 5.93432069e-01 -4.04057324e-01
-1.33170247e-01 4.22370546e-02 2.04942584e-01 -4.56932604e-01
-3.11677437e-03 -2.74379458e-02 -2.02717870e-01 3.72471511e-01
2.91210860e-01 3.67273837e-02 1.77997440e-01 2.97375828e-01
1.03044045e+00 -1.77594334e-01 2.87042856e-01 -1.75641626e-01
3.92897844e-01 -2.87504375e-01 7.14766860e-01 5.41475058e-01
-1.16172433e-01 2.50014424e-01 7.08639383e-01 -1.00393258e-01
-6.54123068e-01 -9.43387091e-01 -2.92837143e-01 1.29403853e+00
5.50991036e-02 -2.29711071e-01 -5.07275522e-01 -1.33646882e+00
2.94831246e-01 3.52075279e-01 -1.11890090e+00 -4.96158361e-01
-2.20333561e-01 -8.80625606e-01 7.04995766e-02 3.69926989e-01
3.18451196e-01 -9.00082111e-01 5.23946024e-02 9.34385061e-02
-5.16820967e-01 -1.13446403e+00 -5.81180215e-01 5.51578522e-01
-8.23088050e-01 -1.29998243e+00 -7.06355512e-01 -1.00996614e+00
9.02431726e-01 2.61411726e-01 1.15862465e+00 -9.68450308e-02
-2.80187368e-01 3.46816510e-01 -3.90829116e-01 -2.64497578e-01
-8.14982206e-02 5.28717935e-01 -2.74419069e-01 4.73456681e-01
3.33098412e-01 -5.90365589e-01 -4.98848826e-01 4.39898193e-01
-8.92325401e-01 2.49721333e-01 5.83828986e-01 1.03252769e+00
5.64456224e-01 5.92721207e-03 8.11841130e-01 -1.26073456e+00
4.38608795e-01 -1.05157113e+00 9.49020535e-02 5.26600718e-01
-5.82172215e-01 6.02695830e-02 5.24404526e-01 -4.57010925e-01
-8.96420777e-01 3.07472739e-02 3.97984803e-01 -3.04762602e-01
-1.32543929e-02 6.69694901e-01 -1.82410896e-01 1.38112539e-02
5.18136740e-01 1.29406258e-01 -8.97176191e-03 -4.35264111e-01
1.76114097e-01 7.63761520e-01 1.49592161e-01 -4.68125314e-01
6.31745279e-01 4.29690361e-01 1.79913908e-01 -4.37176764e-01
-1.33020175e+00 -5.72674513e-01 -7.86831021e-01 -9.91155729e-02
8.07778358e-01 -7.82359540e-01 -5.53871334e-01 3.81231397e-01
-8.68222058e-01 -4.90533829e-01 -1.75995633e-01 4.92949635e-01
-2.92748213e-01 2.52173573e-01 -4.85576421e-01 -3.76769572e-01
1.42674064e-02 -1.18058276e+00 1.11469698e+00 -1.65240262e-02
-7.83227086e-02 -1.70237076e+00 1.52228074e-02 4.32234466e-01
1.98084563e-01 1.62048057e-01 1.16649342e+00 -9.17490959e-01
-1.39523193e-01 -1.29353208e-02 -3.40952963e-01 3.69739681e-01
6.61398709e-01 -5.34661353e-01 -8.63088071e-01 -5.49431086e-01
-1.81942716e-01 -7.20511794e-01 1.51531219e+00 2.21872032e-01
1.58549964e+00 -1.87494904e-01 -8.50723088e-01 6.00342155e-01
1.02102280e+00 -6.96507469e-02 1.33443996e-01 6.93838969e-02
9.85497177e-01 1.07748628e+00 3.90979528e-01 2.96961546e-01
6.85675085e-01 8.34343374e-01 4.14685756e-01 -2.28030652e-01
-2.67144293e-01 -1.46307230e-01 1.28421918e-01 7.42430449e-01
1.27917334e-01 -3.34999293e-01 -8.48480999e-01 5.70543289e-01
-1.90587318e+00 -5.31271040e-01 -5.10493480e-02 2.19048405e+00
1.09660959e+00 -1.23522386e-01 -1.02880940e-01 1.12041011e-01
8.85370731e-01 2.10107654e-01 -7.89671957e-01 1.91802531e-01
1.09117486e-01 -8.50598980e-03 2.26132691e-01 3.76707762e-01
-1.41199052e+00 6.07750177e-01 5.46744204e+00 8.96348059e-01
-8.13337445e-01 2.35015973e-01 7.52953410e-01 -6.11025356e-02
-3.78969699e-01 -1.74341023e-01 -6.98496699e-01 5.11351883e-01
3.56432021e-01 -1.98101148e-01 1.83271140e-01 5.61230898e-01
-3.04174125e-01 1.59007937e-01 -1.46399510e+00 7.19147503e-01
1.54326081e-01 -7.23591208e-01 1.00701869e-01 3.16843748e-01
9.09954250e-01 -1.44412607e-01 3.15639853e-01 4.91157889e-01
4.89226878e-01 -8.61639023e-01 1.93888396e-01 3.90914977e-01
8.22119415e-01 -4.87433732e-01 6.12621605e-01 3.13080609e-01
-1.20551157e+00 -1.90177023e-01 -3.18776906e-01 1.38920516e-01
-2.03282356e-01 8.54962409e-01 -8.65786374e-01 8.04259121e-01
3.87224823e-01 1.19898593e+00 -8.11466277e-01 9.88923550e-01
-2.41405934e-01 5.42800069e-01 1.55677572e-01 2.50056952e-01
1.47329986e-01 -2.42550805e-01 2.12397337e-01 1.15168381e+00
1.16883136e-01 -2.73427606e-01 6.09569192e-01 5.32039285e-01
-4.38835502e-01 9.72093865e-02 -8.26842606e-01 1.46073952e-01
3.65376711e-01 1.47583568e+00 -6.00473940e-01 -1.97142825e-01
-6.73026323e-01 1.17621028e+00 7.53819942e-01 4.46601599e-01
-6.39508605e-01 -2.52864093e-01 6.89200044e-01 -1.84389293e-01
-6.98183328e-02 8.27950686e-02 -2.58793890e-01 -1.22255862e+00
1.37339190e-01 -4.95290399e-01 8.62263799e-01 -2.11491153e-01
-1.86019504e+00 3.96988869e-01 -2.14996096e-03 -1.17443967e+00
-2.44958997e-01 -5.25376856e-01 -3.64454389e-01 7.94454157e-01
-1.84236908e+00 -1.30126643e+00 -4.85440850e-01 8.30658972e-01
4.41419154e-01 -1.88523412e-01 6.62653208e-01 3.73068541e-01
-6.85372591e-01 8.43180239e-01 5.05542196e-02 2.51886219e-01
1.11437905e+00 -1.39385772e+00 1.12107610e-02 3.57308984e-01
1.84933111e-01 4.18562859e-01 1.33926660e-01 -7.58652091e-01
-9.70206559e-01 -1.43212235e+00 8.06592226e-01 -4.48038161e-01
6.78365529e-01 -5.34506142e-01 -1.15430915e+00 8.65606666e-01
-7.16053415e-03 2.02436969e-01 1.11789024e+00 4.67865109e-01
-6.35397196e-01 -1.20249294e-01 -8.05258036e-01 2.04081491e-01
1.33775997e+00 -6.33733392e-01 -3.88388485e-01 7.19998538e-01
6.15396261e-01 -1.57379180e-01 -8.74142289e-01 5.10355651e-01
2.11502925e-01 -3.94971400e-01 8.06374252e-01 -7.76823521e-01
6.92892790e-01 -2.45577674e-02 -1.01809703e-01 -1.86091697e+00
-4.02865738e-01 -6.21988252e-02 -4.80375662e-02 1.20243347e+00
7.39535809e-01 -6.12468600e-01 6.63298428e-01 4.23800409e-01
-2.72343904e-01 -8.27708483e-01 -8.64236534e-01 -7.00790107e-01
1.96636483e-01 -1.20436996e-01 1.24583624e-01 1.43729389e+00
1.03807889e-01 6.08652532e-01 -5.22190273e-01 -6.54599741e-02
8.94892097e-01 1.65688992e-01 2.55482048e-01 -1.54722023e+00
-4.04567897e-01 -2.97464967e-01 -2.10461378e-01 -7.05088258e-01
6.38538897e-01 -1.68111324e+00 -1.85145006e-01 -1.71432173e+00
8.50699842e-01 -8.60722065e-01 -5.85598230e-01 1.05165064e+00
-6.49776399e-01 3.58413696e-01 1.55116633e-01 3.24960798e-01
-7.95127213e-01 5.82470775e-01 1.56204951e+00 -3.69886816e-01
-7.25377426e-02 1.24500811e-01 -1.03802681e+00 5.27822673e-01
6.96775198e-01 -6.20920718e-01 -7.02568591e-01 -6.30204201e-01
1.61305800e-01 -7.53455386e-02 4.01016891e-01 -7.04465806e-01
3.51867050e-01 -6.65606037e-02 3.05009902e-01 3.57664563e-02
2.41222739e-01 -8.32760096e-01 -1.02874592e-01 4.30946082e-01
-7.11508095e-01 -6.59919620e-01 -1.16978623e-01 7.72742391e-01
-2.42032319e-01 -2.60683268e-01 6.23267710e-01 1.71284318e-01
-4.99640495e-01 7.46130228e-01 1.66500404e-01 2.49148771e-01
1.20414102e+00 9.76113006e-02 -4.33157682e-01 -1.70974106e-01
-1.09279263e+00 6.24255836e-01 2.79196441e-01 6.66361868e-01
4.22865659e-01 -1.50176632e+00 -8.40027153e-01 2.30121017e-02
5.39029360e-01 5.70845492e-02 5.04067600e-01 9.17847455e-01
5.11829376e-01 3.75245333e-01 -2.28003591e-01 -6.59307778e-01
-1.25992262e+00 6.58434629e-01 1.95894212e-01 -5.54342091e-01
-1.56321794e-01 7.06200659e-01 7.72222936e-01 -5.95442951e-01
3.63928586e-01 -3.24651063e-03 -5.81624329e-01 5.30652106e-01
2.48733044e-01 2.60055333e-01 1.52459592e-01 -2.81377941e-01
-3.33421707e-01 5.06768763e-01 -4.61850971e-01 2.03069195e-01
1.26147592e+00 -6.05488606e-02 -2.10363224e-01 5.36495984e-01
1.43819225e+00 -2.98220098e-01 -1.30004811e+00 -7.45031714e-01
1.28860667e-01 -3.04073423e-01 2.69961916e-02 -9.25450206e-01
-1.26758313e+00 9.34239924e-01 1.37326896e-01 1.51688680e-01
1.14868009e+00 3.55493337e-01 3.06721568e-01 3.36638808e-01
2.94952184e-01 -8.76582325e-01 3.39118391e-01 4.60945219e-01
7.32972980e-01 -1.40378225e+00 -2.55986929e-01 -8.63548458e-01
-7.93991506e-01 8.36526573e-01 7.47134864e-01 3.27156812e-01
6.64704502e-01 9.73115712e-02 -3.52888346e-01 -1.43559098e-01
-8.98741782e-01 -3.11170578e-01 5.90509593e-01 7.59823620e-01
5.79638124e-01 1.88736886e-01 -3.94890457e-01 6.78481042e-01
2.10461095e-01 -1.06672339e-01 2.59063303e-01 7.48263180e-01
-2.04896435e-01 -1.36297214e+00 -7.27343466e-03 9.39469278e-01
-4.06899571e-01 -4.56893556e-02 -5.25600374e-01 6.41966999e-01
2.58096546e-01 9.36790407e-01 2.05156624e-01 -4.20078576e-01
2.04927489e-01 1.78670973e-01 4.57485437e-01 -9.58650112e-01
-4.08193201e-01 -3.27664554e-01 -1.19116111e-02 -2.51365423e-01
-2.72396296e-01 -5.85019529e-01 -8.78139615e-01 8.65282789e-02
-1.54520839e-01 1.80898950e-01 3.75223935e-01 8.81629288e-01
3.64918590e-01 9.35313106e-01 8.07078600e-01 -4.91686523e-01
-5.27230501e-01 -9.49259639e-01 -8.22205544e-01 7.57764339e-01
2.87736148e-01 -9.66636777e-01 -4.06661361e-01 -1.27949910e-02] | [9.711673736572266, 3.7600643634796143] |
8482f46f-48c2-4d9d-9a0c-7879811ef808 | a-morphology-aware-network-for-morphological | 1702.03654 | null | http://arxiv.org/abs/1702.03654v1 | http://arxiv.org/pdf/1702.03654v1.pdf | A Morphology-aware Network for Morphological Disambiguation | Agglutinative languages such as Turkish, Finnish and Hungarian require
morphological disambiguation before further processing due to the complex
morphology of words. A morphological disambiguator is used to select the
correct morphological analysis of a word. Morphological disambiguation is
important because it generally is one of the first steps of natural language
processing and its performance affects subsequent analyses. In this paper, we
propose a system that uses deep learning techniques for morphological
disambiguation. Many of the state-of-the-art results in computer vision, speech
recognition and natural language processing have been obtained through deep
learning models. However, applying deep learning techniques to morphologically
rich languages is not well studied. In this work, while we focus on Turkish
morphological disambiguation we also present results for French and German in
order to show that the proposed architecture achieves high accuracy with no
language-specific feature engineering or additional resource. In the
experiments, we achieve 84.12, 88.35 and 93.78 morphological disambiguation
accuracy among the ambiguous words for Turkish, German and French respectively. | ['Ozan Sonmez', 'Eray Yildiz', 'H. Bahadir Sahin', 'Mustafa Tolga Eren', 'Caglar Tirkaz'] | 2017-02-13 | null | null | null | null | ['morphological-disambiguation'] | ['natural-language-processing'] | [-3.72725427e-01 -6.76533163e-01 2.87984967e-01 -3.65622252e-01
-3.97281319e-01 -8.61559331e-01 5.02979457e-01 7.95909166e-01
-9.78022039e-01 6.93920434e-01 -9.28502828e-02 -5.95456064e-01
1.69022709e-01 -7.82150507e-01 -1.90960079e-01 -5.50753653e-01
-6.19112663e-02 5.61461985e-01 4.80846735e-03 -3.75137329e-01
2.29262859e-01 7.59705603e-01 -1.04042041e+00 2.25691095e-01
1.16863465e+00 6.04911268e-01 3.09205383e-01 3.81838799e-01
-4.60176915e-01 -9.56548378e-02 -7.87700057e-01 -5.67457616e-01
2.78636396e-01 5.34752980e-02 -8.32133532e-01 4.99412380e-02
5.75097442e-01 1.79306835e-01 3.01553279e-01 1.52137375e+00
6.36902392e-01 4.38880563e-01 6.18995368e-01 -3.86871040e-01
-9.19464111e-01 9.53345656e-01 -5.94922602e-01 6.30178988e-01
2.44519874e-01 -1.58010930e-01 1.01662624e+00 -1.09799767e+00
3.95961702e-01 1.51616597e+00 2.51163453e-01 3.46691191e-01
-7.39096045e-01 -5.40867388e-01 4.50159758e-01 3.64841521e-01
-1.47151542e+00 -1.72772810e-01 6.75457001e-01 -2.99752772e-01
1.25736380e+00 -6.23836089e-03 4.03338760e-01 4.38044757e-01
3.81839603e-01 7.08902001e-01 1.15165770e+00 -7.96411633e-01
1.93195924e-01 -1.16403170e-01 6.50426328e-01 7.62188256e-01
4.99512523e-01 -2.75921524e-01 -9.85519886e-02 2.57691413e-01
5.11796892e-01 -4.73718047e-01 -8.64804834e-02 2.60931611e-01
-1.01247871e+00 7.78005242e-01 5.10935962e-01 8.75762522e-01
-3.65296900e-01 -4.62510943e-01 5.86219907e-01 3.73602867e-01
-8.98327027e-03 5.01723886e-01 -6.89173639e-01 2.18239486e-01
-6.80262923e-01 -8.33240747e-02 5.66939235e-01 6.06361866e-01
7.38884091e-01 5.76476812e-01 3.44508469e-01 1.13437271e+00
3.47324312e-01 5.95579088e-01 1.06756556e+00 -4.65300344e-02
3.80882919e-01 7.97992766e-01 -2.10977480e-01 -8.92911196e-01
-5.97170472e-01 -1.79673418e-01 -6.26641393e-01 2.29992494e-01
4.71343488e-01 -3.65283161e-01 -1.34286737e+00 1.37902915e+00
3.10217798e-01 -2.12386832e-01 4.82287437e-01 9.14974153e-01
1.12264371e+00 7.80422270e-01 4.87221241e-01 -2.78608888e-01
1.86628449e+00 -6.18975937e-01 -7.46329963e-01 -4.80004936e-01
3.51016730e-01 -1.30434871e+00 1.21069539e+00 4.48650569e-01
-9.49354589e-01 -6.94458723e-01 -1.21135414e+00 -1.36914834e-01
-8.59546661e-01 4.36885029e-01 5.69983065e-01 7.40994036e-01
-1.00069988e+00 1.87155947e-01 -9.92803931e-01 -8.15357625e-01
-1.41832575e-01 6.83920205e-01 -6.68649912e-01 3.18393111e-01
-1.24467933e+00 1.12537754e+00 8.88049245e-01 1.79351732e-01
-1.42082095e-01 2.71308385e-02 -1.04141831e+00 -3.69339213e-02
6.69570863e-02 -3.32234532e-01 1.25740254e+00 -1.06692266e+00
-1.32295918e+00 1.19133198e+00 -1.32630616e-01 -5.88332474e-01
-1.43001974e-03 -3.76897275e-01 -7.40804613e-01 2.71762870e-02
-3.71751934e-02 5.88348329e-01 4.49363768e-01 -7.54462481e-01
-9.32355881e-01 -6.29171431e-01 -1.35562330e-01 2.69932956e-01
-1.87007874e-01 6.04914546e-01 -1.96192771e-01 -8.74266267e-01
2.89246768e-01 -8.05095732e-01 -8.33340362e-02 -7.94115305e-01
-2.13148370e-01 -6.51845694e-01 6.87113702e-01 -8.70491445e-01
1.07431722e+00 -2.00555015e+00 -1.24578878e-01 -1.42876655e-01
-3.50225955e-01 9.11178350e-01 -3.61949429e-02 1.76476777e-01
-1.78617463e-02 2.44461715e-01 -1.97441205e-01 -1.24091044e-01
2.53151990e-02 2.64313966e-01 4.91958857e-02 4.10425812e-01
5.96485853e-01 7.38707006e-01 -6.62313044e-01 -5.85309148e-01
5.73766053e-01 3.05297405e-01 -1.28462479e-01 -1.27107590e-01
9.96131130e-05 1.01371102e-01 -2.52818763e-01 8.78321707e-01
8.37133944e-01 6.55040622e-01 4.97902066e-01 -3.10285032e-01
-3.92663360e-01 5.50633192e-01 -1.27143943e+00 1.45366669e+00
-5.85072398e-01 5.38574219e-01 2.11733818e-01 -1.14317298e+00
1.16823769e+00 2.44056627e-01 -1.28570750e-01 -7.20452368e-01
5.03560662e-01 5.24320185e-01 5.61512649e-01 -2.98685253e-01
7.14578211e-01 -3.30223411e-01 -3.27090114e-01 1.27765656e-01
1.40746579e-01 7.08701983e-02 6.54193997e-01 -3.92017156e-01
5.49278855e-01 -1.91889063e-01 8.62414718e-01 -5.22124887e-01
9.50582504e-01 2.86047719e-02 7.58133948e-01 2.29384080e-01
-4.81993407e-01 3.36480469e-01 -2.44246185e-01 -6.62750661e-01
-5.45055807e-01 -1.28037584e+00 -3.32806379e-01 1.13616371e+00
-4.56022546e-02 -1.15386814e-01 -8.28077316e-01 -6.48576140e-01
-2.26987213e-01 6.92479432e-01 -8.62796307e-02 1.44894898e-01
-1.22772276e+00 -1.00979793e+00 4.88578379e-01 4.52508390e-01
6.87290728e-01 -1.63362122e+00 -5.58239400e-01 3.61879826e-01
1.60869449e-01 -1.30940986e+00 -2.79567838e-01 2.91680336e-01
-7.81094134e-01 -8.94338191e-01 -2.71397859e-01 -1.53766012e+00
4.72547382e-01 2.34382134e-02 9.99160051e-01 2.88658030e-02
-1.88671559e-01 -1.68100849e-01 -5.26359916e-01 -4.88259673e-01
-3.98183018e-01 2.93592870e-01 3.65590870e-01 -1.33156508e-01
1.02836919e+00 -2.30424106e-01 -2.09307924e-01 -1.11180469e-01
-8.24475229e-01 -7.60780036e-01 8.02803755e-01 6.10918105e-01
5.87618232e-01 1.95218831e-01 5.24933398e-01 -6.46214724e-01
7.06985593e-01 4.29340638e-02 -7.10860491e-01 1.44289821e-01
-1.55841023e-01 2.12893471e-01 8.79631639e-01 -4.91259754e-01
-1.12569022e+00 3.50011498e-01 -3.82006824e-01 2.28565067e-01
-6.61251128e-01 8.64889741e-01 -6.11193955e-01 9.88515913e-02
5.10947108e-01 4.64046858e-02 -5.49604058e-01 -8.20780814e-01
2.22001404e-01 7.57858694e-01 5.20215750e-01 -5.06571829e-01
4.83964622e-01 1.34169891e-01 -2.72161007e-01 -1.23360682e+00
-4.22398597e-01 -5.77711821e-01 -1.23475242e+00 3.38319093e-01
9.59396780e-01 -6.55993819e-01 -4.65320289e-01 5.83659410e-01
-1.30279779e+00 1.22661240e-01 2.19247654e-01 6.72207892e-01
2.43707909e-03 5.54439068e-01 -6.64194107e-01 -6.54202223e-01
-8.04771304e-01 -1.15509748e+00 9.18415308e-01 4.66373503e-01
-2.48497188e-01 -1.23698485e+00 -7.28499815e-02 -6.50714487e-02
6.20159246e-02 -1.28903501e-02 1.05284607e+00 -1.11960411e+00
-5.43308444e-02 -4.11969945e-02 -4.27817665e-02 2.40865603e-01
5.73029578e-01 2.03403220e-01 -7.43137658e-01 -2.25009054e-01
-1.88272875e-02 4.24437858e-02 9.16913092e-01 3.40945721e-01
1.53212935e-01 -7.87941087e-03 6.16270490e-02 3.94693911e-01
1.42302620e+00 6.38435423e-01 3.13426763e-01 5.38494706e-01
6.50618315e-01 5.30423403e-01 9.26451325e-01 2.41880789e-02
2.70772576e-01 4.56333011e-01 5.03052101e-02 6.61244392e-02
-1.93313301e-01 1.99305683e-01 6.46209717e-01 1.09235144e+00
4.59169537e-01 -1.62294567e-01 -1.22741675e+00 8.16162705e-01
-1.49675941e+00 -6.00494921e-01 -1.76888138e-01 2.20014524e+00
9.91308630e-01 9.88582447e-02 -1.64873265e-02 3.86346400e-01
1.07975602e+00 1.96505859e-02 -7.84958676e-02 -1.12391722e+00
-3.67072850e-01 7.99064040e-01 1.39988676e-01 8.13357770e-01
-1.64290869e+00 1.75930786e+00 5.17080641e+00 7.08428383e-01
-1.69319534e+00 -1.36843890e-01 3.65683049e-01 3.72814149e-01
2.24017873e-01 -3.30089569e-01 -9.84746814e-01 -1.02902194e-02
6.85374320e-01 -8.35262090e-02 1.59161791e-01 4.93230164e-01
2.31404856e-01 -2.13805303e-01 -8.45474541e-01 1.07519090e+00
3.34513523e-02 -6.78314626e-01 4.46918428e-01 -3.98319185e-01
4.69890237e-01 -4.44633029e-02 -8.70612785e-02 2.37981379e-01
3.46038967e-01 -1.06098330e+00 5.77479362e-01 -1.02992162e-01
2.22578228e-01 -1.12200332e+00 8.68866086e-01 9.83249620e-02
-1.46121621e+00 7.87283629e-02 -7.73028016e-01 -5.90155795e-02
7.42641790e-03 5.18875957e-01 -1.21618736e+00 4.31958854e-01
4.65721130e-01 2.92464316e-01 -7.04508841e-01 1.06792915e+00
-6.63888276e-01 6.55290067e-01 -5.09029627e-01 -3.33851218e-01
4.18300420e-01 -2.15638667e-01 5.43036044e-01 1.72802627e+00
3.15762669e-01 -5.77670336e-02 5.97820818e-01 4.13655281e-01
7.44117871e-02 9.02222395e-01 -3.22706103e-01 -4.07954492e-02
5.59450686e-01 1.45895362e+00 -1.07587361e+00 -2.88320899e-01
-2.42367879e-01 7.71252394e-01 5.49918473e-01 1.57643497e-01
-3.57847571e-01 -5.85191309e-01 8.35961699e-01 -2.11059347e-01
2.93864340e-01 -8.23745430e-01 -4.42799211e-01 -1.03157783e+00
-1.35473711e-02 -8.39523375e-01 6.60205245e-01 -2.61597365e-01
-1.30078483e+00 9.19461310e-01 -3.88698786e-01 -8.23197722e-01
-2.00079814e-01 -1.30085635e+00 -6.52292371e-01 1.04183137e+00
-1.46729648e+00 -1.14890373e+00 1.26221597e-01 3.57339352e-01
9.55042958e-01 -4.41702336e-01 8.81893575e-01 3.23769897e-01
-6.87761962e-01 6.99615538e-01 -1.81159273e-01 5.25377631e-01
8.33214045e-01 -1.59076285e+00 5.86214364e-01 1.30457866e+00
6.13817334e-01 8.27049375e-01 5.46270967e-01 -6.67121470e-01
-1.36768091e+00 -9.20777082e-01 1.21472597e+00 -1.22875497e-01
7.74817407e-01 -2.15270996e-01 -9.79941010e-01 4.18012798e-01
7.03487754e-01 -1.42533481e-01 6.23372495e-01 3.10114007e-02
-2.94741187e-02 -1.97308704e-01 -1.08766568e+00 7.72903323e-01
5.58446467e-01 -2.37470999e-01 -1.05904913e+00 -1.52827561e-01
4.74638492e-01 -1.98456988e-01 -5.06753504e-01 2.37263903e-01
1.96593583e-01 -6.58081770e-01 6.70734465e-01 -5.46513796e-01
-3.36781949e-01 -6.01060927e-01 -1.97869837e-01 -1.42781556e+00
-3.62807721e-01 -4.17156070e-01 5.29368818e-01 1.39992321e+00
5.47137201e-01 -5.06029963e-01 5.08425951e-01 1.22135840e-01
-3.31279278e-01 -1.78192139e-01 -8.57319891e-01 -7.38978148e-01
5.13578236e-01 -3.21959198e-01 6.02747619e-01 1.11924136e+00
7.56271183e-02 7.02701926e-01 3.81168664e-01 4.04810607e-01
1.40540972e-01 3.13625306e-01 1.94969803e-01 -1.23856103e+00
1.05153814e-01 -7.29502559e-01 -7.55900145e-01 -5.36804557e-01
5.86351752e-01 -7.96884120e-01 4.44469862e-02 -1.58082402e+00
-3.52521360e-01 -1.46981061e-01 -4.64638680e-01 5.46867311e-01
-5.36063731e-01 2.14851275e-01 2.93805301e-01 -3.08454663e-01
-3.21488798e-01 2.72958547e-01 8.99047732e-01 -2.79158533e-01
-3.96318942e-01 -1.78926215e-01 -5.82946420e-01 9.30489421e-01
1.25421083e+00 -7.22842887e-02 7.76095456e-03 -8.56304526e-01
-8.73649642e-02 -5.46310067e-01 -2.86982328e-01 -8.69841635e-01
8.76848847e-02 -1.61262661e-01 4.08182949e-01 -6.68410897e-01
1.11412898e-01 -6.68417931e-01 -5.46737969e-01 4.93755162e-01
2.86894768e-01 8.16156983e-01 5.50319910e-01 -1.48077667e-01
-5.64893723e-01 -5.06708682e-01 1.05949426e+00 -3.09612781e-01
-1.24594438e+00 6.20566681e-02 -8.28386009e-01 -1.54472655e-02
8.94706607e-01 -2.03278795e-01 -9.81515199e-02 1.96156755e-01
-8.24418068e-01 4.35754191e-03 3.10982734e-01 6.68067336e-01
6.03399932e-01 -1.18977880e+00 -8.48337591e-01 2.47104198e-01
-1.77765995e-01 -8.09578747e-02 -3.12735409e-01 3.34500372e-01
-9.47999656e-01 5.55909991e-01 -4.98550773e-01 -2.20514327e-01
-1.61533475e+00 6.96851313e-01 2.27603585e-01 -2.29709476e-01
-1.25178963e-01 7.60961652e-01 1.17346935e-01 -5.53180099e-01
1.24942139e-01 -5.63402414e-01 -7.75607049e-01 2.25684986e-01
4.20300364e-01 1.84804261e-01 6.45000875e-01 -1.09324503e+00
-7.72804379e-01 7.74797559e-01 -2.80764967e-01 -2.18143940e-01
8.18722069e-01 -6.64423108e-02 -3.79817069e-01 2.60112524e-01
7.78916121e-01 6.19241834e-01 -2.30469763e-01 -1.33669734e-01
6.63367927e-01 -7.97110498e-02 -7.41206482e-02 -9.14496183e-01
-9.37297165e-01 1.21793878e+00 6.38975441e-01 1.81548029e-01
1.19797814e+00 -3.26081455e-01 8.67367506e-01 6.86190367e-01
3.55845094e-01 -1.18240285e+00 -4.42726880e-01 1.23458409e+00
6.88465416e-01 -1.26996231e+00 -2.68945813e-01 -4.05755073e-01
-4.86664176e-01 1.21390128e+00 9.38234985e-01 -3.03566098e-01
5.18471479e-01 2.71471322e-01 6.48374140e-01 1.33039817e-01
-4.35209155e-01 -5.36597371e-01 3.26240718e-01 7.67174423e-01
1.10197210e+00 3.93108040e-01 -9.71938968e-01 6.47981226e-01
-6.43157363e-01 -8.24494720e-01 4.51709539e-01 8.04679275e-01
-8.12098563e-01 -1.44288456e+00 -7.34481275e-01 3.23050730e-02
-9.40603316e-01 -3.85249346e-01 -8.82843852e-01 8.03583741e-01
1.66629419e-01 1.14530373e+00 1.66233137e-01 -5.05139083e-02
3.11387211e-01 2.45761797e-01 5.23012042e-01 -9.00015593e-01
-9.71945882e-01 2.96502322e-01 2.63415635e-01 1.87305540e-01
-2.05309570e-01 -7.18893170e-01 -1.89627957e+00 -1.57746211e-01
-2.26695403e-01 3.09709191e-01 6.31447256e-01 9.58139718e-01
3.64013314e-02 1.91586971e-01 1.41233996e-01 -5.16935706e-01
-3.51931065e-01 -1.07182252e+00 -5.30403674e-01 3.81107450e-01
6.78427145e-02 -4.99576926e-01 1.83805287e-01 7.56348819e-02] | [10.377156257629395, 10.146698951721191] |
b1b42ceb-ac8c-456e-818d-b0f82d127f9f | unsupervised-audio-source-separation-using | 2005.13769 | null | https://arxiv.org/abs/2005.13769v1 | https://arxiv.org/pdf/2005.13769v1.pdf | Unsupervised Audio Source Separation using Generative Priors | State-of-the-art under-determined audio source separation systems rely on supervised end-end training of carefully tailored neural network architectures operating either in the time or the spectral domain. However, these methods are severely challenged in terms of requiring access to expensive source level labeled data and being specific to a given set of sources and the mixing process, which demands complete re-training when those assumptions change. This strongly emphasizes the need for unsupervised methods that can leverage the recent advances in data-driven modeling, and compensate for the lack of labeled data through meaningful priors. To this end, we propose a novel approach for audio source separation based on generative priors trained on individual sources. Through the use of projected gradient descent optimization, our approach simultaneously searches in the source-specific latent spaces to effectively recover the constituent sources. Though the generative priors can be defined in the time domain directly, e.g. WaveGAN, we find that using spectral domain loss functions for our optimization leads to good-quality source estimates. Our empirical studies on standard spoken digit and instrument datasets clearly demonstrate the effectiveness of our approach over classical as well as state-of-the-art unsupervised baselines. | ['Jayaraman J. Thiagarajan', 'Vivek Narayanaswamy', 'Andreas Spanias', 'Rushil Anirudh'] | 2020-05-28 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [ 4.74202156e-01 -6.94252923e-02 -1.27217010e-01 -2.86337048e-01
-1.56537497e+00 -7.29119658e-01 5.14295995e-01 -1.49193227e-01
-1.20843522e-01 4.49343234e-01 5.86204410e-01 1.05150819e-01
-3.23925704e-01 -2.46383324e-01 -5.18533826e-01 -8.35230410e-01
6.12021536e-02 3.03380638e-01 -1.77470431e-01 -1.37126520e-01
-1.56990543e-01 1.85481042e-01 -1.39669299e+00 -1.05363034e-01
9.88767505e-01 8.82264614e-01 7.64894113e-02 8.57656419e-01
-1.10428602e-01 7.86505103e-01 -6.00767076e-01 -2.01599494e-01
2.61232227e-01 -8.91176462e-01 -4.18426633e-01 1.83841914e-01
3.97364616e-01 -1.63622230e-01 -2.37376302e-01 1.10069501e+00
7.42880702e-01 2.64216274e-01 7.07206905e-01 -1.09801984e+00
-2.80866981e-01 9.03979480e-01 -4.23371762e-01 4.30493690e-02
4.12896946e-02 -5.82064316e-03 1.23451245e+00 -1.04772317e+00
1.49477869e-01 1.05022502e+00 8.36243927e-01 3.49372357e-01
-1.55063093e+00 -6.37917042e-01 1.06246457e-01 -1.27322629e-01
-1.20948148e+00 -1.22930741e+00 1.36856842e+00 -6.25123799e-01
6.28756881e-01 1.96535103e-02 1.92438647e-01 1.20439661e+00
-3.86452407e-01 6.48598671e-01 8.30018520e-01 -7.75025785e-01
2.66756952e-01 1.00459173e-01 -3.01223062e-02 3.51581305e-01
-1.77119881e-01 1.69024304e-01 -1.07300687e+00 -1.35618865e-01
5.21945834e-01 -4.66110915e-01 -5.45426309e-01 -4.97999936e-01
-1.12576115e+00 7.67520845e-01 1.11785889e-01 3.06583285e-01
-3.58378887e-01 2.19157860e-01 2.15582892e-01 8.02570134e-02
7.58436501e-01 5.39382994e-01 -4.67908114e-01 -2.19648942e-01
-1.57321763e+00 5.54216914e-02 7.80668557e-01 6.66349888e-01
5.85931420e-01 6.08742177e-01 -1.26554370e-01 1.12272096e+00
5.42697668e-01 5.39843202e-01 4.55355644e-01 -1.07620454e+00
4.04164255e-01 2.73954719e-02 4.85673212e-02 -7.01689363e-01
-9.04315040e-02 -8.63880157e-01 -7.02181578e-01 1.54936552e-01
6.17221475e-01 -3.97239208e-01 -9.88640189e-01 2.13468981e+00
1.80120856e-01 4.34524268e-01 2.77036458e-01 8.84255290e-01
4.00554329e-01 5.52975059e-01 -2.94452608e-01 -2.03389436e-01
1.06104398e+00 -9.84659731e-01 -8.09600353e-01 -5.09045839e-01
3.99978645e-02 -9.09352899e-01 1.08313692e+00 5.67513645e-01
-1.28110671e+00 -6.07230663e-01 -1.14122570e+00 4.71910387e-02
-2.46533733e-02 3.24644774e-01 3.02989632e-01 7.98818231e-01
-8.70288014e-01 5.97229421e-01 -9.88961995e-01 -1.74059775e-02
1.84670687e-01 2.01235697e-01 5.13965674e-02 1.88619316e-01
-1.03879678e+00 4.78140593e-01 7.28203803e-02 1.28170535e-01
-1.34057200e+00 -8.51752102e-01 -8.36306691e-01 1.58364087e-01
3.33258450e-01 -4.83833998e-01 1.50979114e+00 -1.10430419e+00
-1.95684576e+00 5.00159264e-01 -3.06154221e-01 -5.08690298e-01
3.61940026e-01 -5.47473133e-01 -4.84636337e-01 3.48219007e-01
2.99843214e-02 2.60239959e-01 1.34748435e+00 -1.24796295e+00
-4.17765021e-01 -2.54095122e-02 -2.46270597e-01 2.52880901e-01
-4.89582032e-01 1.78528309e-01 -3.10071051e-01 -9.94745493e-01
1.48293614e-01 -8.22810292e-01 -9.23186913e-03 -2.48060614e-01
-4.55182463e-01 2.26180270e-01 4.73853946e-01 -9.64144289e-01
9.31345403e-01 -2.38831854e+00 4.74760860e-01 2.05859900e-01
4.21078205e-02 -3.57822627e-02 -3.68985236e-02 3.29531193e-01
-7.16895536e-02 -1.98702306e-01 -5.41192770e-01 -8.85589600e-01
2.29620621e-01 -1.63244858e-01 -7.58443117e-01 4.64452833e-01
2.22932532e-01 2.95826048e-01 -8.73188555e-01 -3.03665221e-01
3.18569392e-02 8.25443506e-01 -5.94880402e-01 4.47678715e-01
-1.81092739e-01 7.37235963e-01 -1.01413177e-02 3.36920828e-01
3.03542107e-01 -1.40658796e-01 1.32092312e-01 -1.73796371e-01
4.11714427e-02 8.43993366e-01 -1.33363450e+00 2.11565495e+00
-5.76920927e-01 8.06360364e-01 5.75998306e-01 -9.02687728e-01
7.14132130e-01 6.82703555e-01 5.84705949e-01 -1.07440546e-01
3.84350084e-02 6.36200249e-01 1.33654540e-02 -2.00257793e-01
2.31573239e-01 -2.94786394e-01 1.17491052e-01 5.07849097e-01
6.33123636e-01 -3.06351274e-01 2.17873845e-02 2.23358460e-02
7.32912600e-01 2.41880193e-01 -4.71838284e-03 -1.71626359e-01
2.80547738e-01 -2.48659015e-01 5.97565234e-01 6.94479764e-01
6.64173141e-02 9.09145117e-01 5.78109846e-02 4.86562967e-01
-7.60085285e-01 -1.13867414e+00 -7.22933337e-02 1.30820048e+00
-2.59365380e-01 -3.83559197e-01 -8.54108036e-01 -3.34442466e-01
-2.33482972e-01 7.35797048e-01 -1.26781374e-01 -2.29277819e-01
-6.30751133e-01 -7.08081961e-01 8.45508218e-01 4.52970684e-01
2.15172619e-01 -6.07067347e-01 -5.03246963e-01 4.86763954e-01
-3.27872783e-01 -1.05393219e+00 -5.86713970e-01 5.76645195e-01
-7.77433336e-01 -6.61333680e-01 -9.76015508e-01 -5.63873291e-01
3.46028060e-01 1.56901523e-01 1.02277958e+00 -6.51737034e-01
5.14622591e-02 4.92661893e-01 -1.80437744e-01 -5.52812696e-01
-4.33841407e-01 1.38849199e-01 2.27317318e-01 4.19829160e-01
7.48544186e-02 -1.12749076e+00 -4.17279840e-01 8.01395997e-02
-6.02856457e-01 -1.23535581e-01 4.18766946e-01 7.63694167e-01
4.59926009e-01 1.84351936e-01 7.02996671e-01 -4.93984669e-01
7.26060212e-01 -3.99406999e-01 -6.04849577e-01 9.82439891e-02
-6.08623564e-01 3.07344675e-01 5.51522791e-01 -5.77929556e-01
-1.26080608e+00 6.02268353e-02 1.89748313e-02 -5.74499488e-01
-9.79917794e-02 5.49349248e-01 -3.54693025e-01 2.06527546e-01
8.61548364e-01 2.50801712e-01 -2.27355406e-01 -7.80153811e-01
5.76254487e-01 6.98777556e-01 8.63926053e-01 -6.40372097e-01
8.52319062e-01 4.14300948e-01 -2.73387343e-01 -7.80769110e-01
-9.72600758e-01 -4.56456751e-01 -5.46828747e-01 2.20213993e-03
7.25971878e-01 -1.17094123e+00 -1.97420251e-02 5.78145206e-01
-1.02149785e+00 -4.31161910e-01 -5.45477629e-01 8.46449256e-01
-6.07247531e-01 2.19542161e-01 -4.97922748e-01 -1.07071042e+00
-3.06358904e-01 -1.07309163e+00 1.09258950e+00 1.36300465e-02
-3.10306937e-01 -9.86232936e-01 3.83290082e-01 2.16624811e-01
5.20518780e-01 -4.79279049e-02 6.34496689e-01 -6.55346096e-01
-4.52684194e-01 -8.70640948e-03 2.72863388e-01 6.35725915e-01
3.69180232e-01 -1.14119612e-01 -1.60222328e+00 -2.07710221e-01
3.58390391e-01 -2.51828462e-01 1.01949799e+00 4.90245610e-01
6.71553195e-01 -2.88465291e-01 -2.64060441e-02 9.05906081e-01
1.05817604e+00 1.19627275e-01 1.38085619e-01 -9.59775373e-02
6.52798533e-01 6.19254231e-01 1.47433132e-01 3.48742038e-01
1.59185514e-01 6.08541548e-01 1.13549739e-01 -1.55991644e-01
-5.09429812e-01 -5.04663944e-01 5.16537428e-01 1.01000214e+00
1.79044038e-01 -2.70153463e-01 -7.88367748e-01 8.24071884e-01
-1.67060602e+00 -8.42944205e-01 2.97871411e-01 2.23554921e+00
1.34736300e+00 1.54725015e-01 2.58875012e-01 5.53750575e-01
5.83655715e-01 3.17541957e-01 -6.46050870e-01 2.13775799e-01
-1.28351256e-01 4.20721710e-01 2.73657799e-01 6.46014690e-01
-1.09242797e+00 5.53356171e-01 6.16696215e+00 9.01925087e-01
-1.33501768e+00 2.92020768e-01 2.28984609e-01 -4.02528405e-01
-3.05379868e-01 3.22791114e-02 -6.15739942e-01 3.56678694e-01
1.15221894e+00 -5.47541007e-02 8.16950977e-01 5.84540009e-01
1.73132390e-01 1.70164883e-01 -1.25022435e+00 1.14414215e+00
1.04566015e-01 -9.06346738e-01 -5.54832935e-01 -8.95314142e-02
5.49394429e-01 5.76078445e-02 1.86506525e-01 -1.05699729e-02
3.92218053e-01 -9.09262955e-01 1.23176634e+00 4.22218442e-01
6.26359284e-01 -5.35179198e-01 1.32870302e-01 3.43291432e-01
-1.02390397e+00 1.48775220e-01 1.70028750e-02 1.56398699e-01
2.17226520e-01 8.55001509e-01 -8.08110833e-01 4.83253837e-01
4.36967641e-01 5.89299202e-01 -6.48146421e-02 9.50072467e-01
-5.73333621e-01 1.34013748e+00 -5.68296671e-01 5.51002979e-01
-5.28122894e-02 -6.63352162e-02 8.15581977e-01 1.33070028e+00
4.47196156e-01 -2.43869856e-01 9.01650414e-02 9.19450343e-01
-1.52619928e-01 -1.20414682e-01 -2.57896632e-01 -2.73565799e-01
5.25461435e-01 1.07786083e+00 -4.95484114e-01 -9.34149921e-02
-1.96483850e-01 8.01068723e-01 9.26409811e-02 6.50576711e-01
-7.24339366e-01 -5.53498745e-01 6.25545919e-01 -6.47584721e-02
4.18161869e-01 -4.60747391e-01 -3.28745812e-01 -1.31734073e+00
6.54890668e-03 -1.04513967e+00 1.45384029e-01 -5.09494960e-01
-1.35010159e+00 6.40408516e-01 -1.25649674e-02 -1.11056113e+00
-7.48924792e-01 -3.62660080e-01 -6.53453171e-01 1.05327070e+00
-1.76691723e+00 -1.01223934e+00 -2.43671183e-02 7.79882073e-01
5.49024105e-01 -3.03480119e-01 8.09106886e-01 3.89334202e-01
-4.79338169e-01 6.16347134e-01 3.04438531e-01 9.58650783e-02
9.87571537e-01 -1.24695849e+00 1.56208813e-01 1.25755465e+00
7.36943960e-01 5.72700739e-01 8.25584650e-01 -2.61494666e-01
-1.12385178e+00 -7.97867775e-01 7.52304137e-01 -2.26210356e-01
6.05928481e-01 -6.44377351e-01 -8.21539104e-01 4.50377196e-01
4.13521826e-01 -6.52857497e-02 9.53137875e-01 3.29314470e-01
-5.49203694e-01 -2.65371323e-01 -7.48052478e-01 4.33634192e-01
8.73116553e-01 -1.00388432e+00 -6.16869211e-01 2.61849701e-01
5.96557438e-01 -2.83755749e-01 -4.85767841e-01 2.00922310e-01
3.89677912e-01 -9.26655769e-01 1.06053281e+00 -4.53048557e-01
2.67889798e-01 -3.85967672e-01 -3.77335519e-01 -1.60207653e+00
-8.20950866e-02 -1.12588513e+00 -3.19737285e-01 1.77507448e+00
4.94779885e-01 -4.42322493e-01 4.23722535e-01 4.54710752e-01
-2.61937588e-01 -2.42268220e-01 -8.10962558e-01 -8.06497693e-01
-6.48436174e-02 -7.56304562e-01 4.65648383e-01 8.20885181e-01
-1.87248103e-02 5.13459861e-01 -6.34598255e-01 3.98912132e-01
8.82297456e-01 1.47013113e-01 6.58730090e-01 -1.18470609e+00
-9.21762645e-01 -4.73524660e-01 2.07132220e-01 -1.02895987e+00
3.33517641e-01 -7.00317621e-01 5.16650498e-01 -1.30836415e+00
-3.34536880e-01 -4.12817657e-01 -4.47007716e-01 3.59613091e-01
-1.65432870e-01 1.18071541e-01 1.94694772e-01 4.05496359e-01
-1.55717716e-01 7.24270463e-01 5.75226843e-01 -2.41916060e-01
-4.14017230e-01 5.60527816e-02 -8.58502090e-01 1.06758666e+00
7.94953406e-01 -6.38663709e-01 -6.68507397e-01 -6.71984494e-01
7.92815387e-02 1.11333758e-01 3.80660951e-01 -1.23066175e+00
3.69942456e-01 1.23613458e-02 6.56212717e-02 -2.29247063e-01
8.28502119e-01 -7.21365392e-01 5.62860444e-02 -9.09553841e-02
-6.73183024e-01 -6.18220627e-01 1.04345970e-01 6.03469372e-01
-5.28316855e-01 -2.59656191e-01 7.55226254e-01 1.12163700e-01
-1.57382190e-01 -5.24736457e-02 -1.41389444e-01 5.08435249e-01
3.17253262e-01 3.66958953e-03 1.02429546e-01 -7.53260851e-01
-6.25195205e-01 -2.05253646e-01 1.21843591e-01 1.31996915e-01
3.38370591e-01 -1.19541192e+00 -8.41345429e-01 2.92769939e-01
-1.48037583e-01 -4.42336388e-02 -3.02330423e-02 8.70989978e-01
1.86201379e-01 1.91031039e-01 1.84565768e-01 -4.85122204e-01
-9.05295134e-01 2.04316452e-01 5.03715515e-01 -2.53001541e-01
-3.89307499e-01 1.14445651e+00 3.01809788e-01 -2.42617056e-01
5.67792296e-01 -3.02474529e-01 7.11035728e-02 2.22056836e-01
3.10865909e-01 4.63575453e-01 1.14921823e-01 -6.56704247e-01
-3.51774067e-01 5.53329289e-01 3.99673581e-01 -7.86857963e-01
1.36318004e+00 -1.95536807e-01 1.89013153e-01 6.86796188e-01
1.11403167e+00 7.34935999e-01 -1.64262569e+00 -4.21626627e-01
-4.24381644e-02 -3.43571544e-01 3.54385018e-01 -8.29191387e-01
-1.03049636e+00 1.14763367e+00 4.44959313e-01 1.92214280e-01
1.41440678e+00 -1.63320392e-01 6.80117726e-01 1.67852283e-01
1.00742191e-01 -1.02127182e+00 1.01535909e-01 3.87279034e-01
8.08572531e-01 -9.66194987e-01 -3.37741107e-01 -2.86058605e-01
-3.76212507e-01 9.90842879e-01 -5.20785451e-02 -1.35453656e-01
7.40824997e-01 4.38603967e-01 3.01167786e-01 8.21032077e-02
-3.75714123e-01 -2.14905828e-01 6.13672912e-01 4.80629623e-01
5.67030728e-01 -1.08176664e-01 3.23888928e-01 7.83324957e-01
-4.65917259e-01 -1.64948478e-01 2.13870764e-01 7.14237571e-01
-2.76507378e-01 -1.24755216e+00 -5.79048038e-01 1.10410623e-01
-7.03778028e-01 -3.56160551e-01 -3.17735434e-01 2.17204675e-01
-5.34685701e-02 1.39572155e+00 -2.41752937e-01 -7.99739286e-02
1.75643176e-01 4.99395609e-01 4.40521359e-01 -5.88379145e-01
-4.36821342e-01 7.65854895e-01 1.56042993e-01 -1.49536252e-01
-7.48500526e-01 -7.90530205e-01 -1.06614423e+00 3.70123565e-01
-3.75082672e-01 2.87756324e-01 6.81114674e-01 8.71730983e-01
4.00947154e-01 7.24461734e-01 5.88338375e-01 -1.11256754e+00
-8.57095420e-01 -1.06165266e+00 -5.64255536e-01 2.13622332e-01
8.31212461e-01 -5.37848413e-01 -5.93778074e-01 4.91761357e-01] | [15.429659843444824, 5.564156532287598] |
ada3cccc-8405-46df-a8f4-2ac4f9cd629b | real-time-object-detection-for-streaming | 2203.12338 | null | https://arxiv.org/abs/2203.12338v2 | https://arxiv.org/pdf/2203.12338v2.pdf | Real-time Object Detection for Streaming Perception | Autonomous driving requires the model to perceive the environment and (re)act within a low latency for safety. While past works ignore the inevitable changes in the environment after processing, streaming perception is proposed to jointly evaluate the latency and accuracy into a single metric for video online perception. In this paper, instead of searching trade-offs between accuracy and speed like previous works, we point out that endowing real-time models with the ability to predict the future is the key to dealing with this problem. We build a simple and effective framework for streaming perception. It equips a novel DualFlow Perception module (DFP), which includes dynamic and static flows to capture the moving trend and basic detection feature for streaming prediction. Further, we introduce a Trend-Aware Loss (TAL) combined with a trend factor to generate adaptive weights for objects with different moving speeds. Our simple method achieves competitive performance on Argoverse-HD dataset and improves the AP by 4.9% compared to the strong baseline, validating its effectiveness. Our code will be made available at https://github.com/yancie-yjr/StreamYOLO. | ['Jian Sun', 'Xiaoping Li', 'Zeming Li', 'Songtao Liu', 'Jinrong Yang'] | 2022-03-23 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Yang_Real-Time_Object_Detection_for_Streaming_Perception_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Yang_Real-Time_Object_Detection_for_Streaming_Perception_CVPR_2022_paper.pdf | cvpr-2022-1 | ['real-time-object-detection'] | ['computer-vision'] | [ 9.62170288e-02 -2.88889408e-01 -2.09564507e-01 -6.43703818e-01
-3.41696143e-01 -4.61031079e-01 5.84868312e-01 1.42981231e-01
-5.26414871e-01 2.03712136e-01 1.36876926e-01 -3.33948314e-01
1.82700396e-01 -8.29987943e-01 -7.38796592e-01 -4.64023381e-01
-3.41009945e-01 -2.14222461e-01 9.38713014e-01 -3.59617293e-01
3.56751353e-01 2.50480622e-01 -2.02244782e+00 3.78842473e-01
8.37346554e-01 1.45109892e+00 5.58055282e-01 1.16443753e+00
2.48452932e-01 1.20370293e+00 -3.00897092e-01 -4.60450083e-01
4.75764751e-01 -9.82567444e-02 -2.11537629e-01 -1.53080165e-01
3.74120414e-01 -6.56714439e-01 -7.46277332e-01 7.67340422e-01
5.72600543e-01 6.32816181e-02 1.98468447e-01 -1.62312305e+00
-9.92448255e-02 5.98868728e-01 -5.04023910e-01 7.66814947e-01
2.07899198e-01 5.70663750e-01 9.76320326e-01 -6.99597359e-01
1.85411021e-01 1.15244210e+00 4.41783875e-01 2.50206620e-01
-6.71386302e-01 -7.84184992e-01 5.41015565e-01 9.62288201e-01
-1.00414944e+00 -9.46577847e-01 5.07799387e-01 -2.64306188e-01
8.77727330e-01 2.83627987e-01 6.18209779e-01 1.06704807e+00
3.69278610e-01 7.73988724e-01 5.37406743e-01 1.10335477e-01
3.24759722e-01 -4.90550324e-02 2.17164736e-02 3.52252096e-01
8.68751109e-02 2.45868891e-01 -1.02545202e+00 2.08920613e-01
2.79126972e-01 -5.23571670e-03 -3.28649104e-01 -1.33315772e-01
-1.32180488e+00 3.87453198e-01 2.94034481e-01 -2.39286721e-01
-4.02894914e-01 4.54347938e-01 6.39907241e-01 3.25544894e-01
3.47123623e-01 -4.31862138e-02 -3.05982143e-01 -7.87735164e-01
-9.18648779e-01 2.60523289e-01 6.29300356e-01 1.15924382e+00
5.55596888e-01 -5.71443960e-02 -3.94698560e-01 5.73667467e-01
2.78347164e-01 6.34469450e-01 2.35157460e-01 -1.56225193e+00
6.27561331e-01 1.44859836e-01 1.42978460e-01 -1.15061343e+00
-3.11240673e-01 -4.40690666e-01 -5.23762465e-01 2.72513688e-01
1.44690439e-01 -4.36562970e-02 -5.32071888e-01 1.60018206e+00
2.50035256e-01 7.31132686e-01 1.02412142e-02 1.09005189e+00
3.65733594e-01 7.92576611e-01 2.52156816e-02 -1.56157449e-01
1.16139388e+00 -1.21445441e+00 -6.30344570e-01 -2.88983792e-01
2.79960275e-01 -6.76164627e-01 9.44609702e-01 5.24444699e-01
-1.25772738e+00 -8.32758844e-01 -1.21725404e+00 -5.75167947e-02
1.47811547e-02 -3.40064883e-01 4.67010260e-01 3.89847189e-01
-1.08060193e+00 5.64668119e-01 -1.04155195e+00 -2.91411877e-01
4.11682516e-01 -2.06014682e-02 1.24133810e-01 -3.32745574e-02
-1.15477037e+00 5.69046140e-01 3.19813937e-01 -5.02282381e-02
-1.08648717e+00 -8.90992343e-01 -6.92728698e-01 5.92443608e-02
4.26436245e-01 -8.21369469e-01 1.59021103e+00 -8.58721673e-01
-1.49990094e+00 3.10266584e-01 -3.78862232e-01 -8.98270845e-01
1.01132977e+00 -3.98549914e-01 -4.35480326e-01 4.02723700e-01
-4.46836762e-02 7.75088251e-01 8.04266930e-01 -1.10039866e+00
-1.34233809e+00 -9.69284996e-02 3.19555521e-01 3.86735559e-01
-3.56489450e-01 -1.00208022e-01 -7.87344515e-01 -3.88373375e-01
4.43220250e-02 -6.12074554e-01 -1.31067619e-01 3.19542289e-01
1.31281778e-01 4.97246273e-02 8.65373075e-01 -4.75978225e-01
1.34626007e+00 -2.25584698e+00 -2.63502002e-01 -1.44500375e-01
3.90413612e-01 2.22350582e-01 -1.71888620e-01 3.97708356e-01
4.51614261e-01 -1.75719008e-01 -2.76875440e-02 -5.74946463e-01
3.20273414e-02 1.86030507e-01 -4.15748358e-01 2.88922489e-01
1.86161980e-01 6.65998101e-01 -1.18656671e+00 -4.28524792e-01
3.35537374e-01 4.01267260e-01 -6.81107461e-01 3.38347375e-01
-9.34722126e-02 1.79965734e-01 -3.08892548e-01 5.46421289e-01
8.50794554e-01 -6.41923398e-02 -1.78683937e-01 -4.76921201e-02
-3.21143657e-01 3.82945210e-01 -1.13973689e+00 1.52422667e+00
-4.63807344e-01 7.27247596e-01 8.02018419e-02 -7.43980587e-01
8.84010792e-01 1.93405285e-01 3.96025807e-01 -1.16572201e+00
-1.81102939e-02 1.52414575e-01 -8.88382569e-02 -7.25520730e-01
7.89305329e-01 2.16721892e-01 2.71775156e-01 6.90130666e-02
-3.17702770e-01 2.99137205e-01 2.77429998e-01 3.89776915e-01
1.41697121e+00 2.74995923e-01 -1.73014194e-01 1.18902922e-01
4.65895236e-01 -2.31547534e-01 9.54993784e-01 7.86552608e-01
-7.41782546e-01 6.42971992e-01 5.02680480e-01 -4.49849099e-01
-9.97875690e-01 -1.19270897e+00 6.86025769e-02 1.24802673e+00
6.80270612e-01 -4.74465042e-01 -3.48733842e-01 -3.83280098e-01
-7.42606353e-03 8.82239044e-01 -3.64807785e-01 -2.38497987e-01
-5.43629229e-01 -4.65272099e-01 3.95450175e-01 5.74213684e-01
6.47061706e-01 -7.28067935e-01 -1.25376654e+00 3.49592805e-01
-5.42275846e-01 -1.50129354e+00 -4.86582279e-01 -2.00237423e-01
-6.00681007e-01 -7.64902592e-01 -3.32014620e-01 -2.15704203e-01
-2.75915246e-02 7.93368995e-01 1.02815747e+00 9.66410190e-02
-8.19395762e-03 2.62570202e-01 -6.58308148e-01 -5.77023506e-01
-3.59089822e-01 -9.22707990e-02 -1.12902615e-02 1.70025751e-01
2.46699512e-01 -7.74239540e-01 -1.11789155e+00 3.92946392e-01
-7.82083929e-01 4.14873391e-01 3.11246753e-01 3.39316517e-01
5.30814052e-01 8.04143900e-04 5.69879711e-01 -3.77514899e-01
2.10708812e-01 -7.96979189e-01 -4.05156732e-01 -1.53967962e-01
-6.88676894e-01 -3.44035745e-01 6.67327642e-01 -3.13247830e-01
-9.80395675e-01 -2.12474391e-01 -2.50059247e-01 -5.71958363e-01
-3.67291234e-02 7.24587142e-02 -1.27126155e-02 3.77005249e-01
4.71919239e-01 3.23233992e-01 -1.19100250e-02 -7.84734562e-02
2.44052023e-01 6.79876149e-01 6.27708852e-01 -1.92127571e-01
5.49168885e-01 8.27384472e-01 -5.75691052e-02 -4.61954176e-01
-7.77566791e-01 -6.08465612e-01 -3.76153439e-01 -6.84644461e-01
4.79811639e-01 -1.25238848e+00 -8.77823651e-01 5.32473326e-01
-1.06840181e+00 -5.22007048e-01 -2.38161281e-01 4.59178537e-01
-8.23427260e-01 4.95217949e-01 -5.55288553e-01 -9.44602072e-01
-1.81969479e-01 -9.78641450e-01 8.90099943e-01 1.72273040e-01
2.76019037e-01 -5.34619808e-01 -1.34985477e-01 2.88888365e-01
7.48717844e-01 9.82190073e-02 -4.19632271e-02 -1.95330933e-01
-8.82795870e-01 7.14414045e-02 -5.60936034e-01 3.41551870e-01
-1.99301749e-01 1.90504804e-01 -1.23364806e+00 -2.50400037e-01
2.40024272e-02 -9.53906402e-03 1.33683050e+00 2.57165760e-01
1.28132463e+00 -1.98146090e-01 -1.32278323e-01 8.39993834e-01
1.37271571e+00 2.53582388e-01 7.19998598e-01 4.74926740e-01
5.29085338e-01 6.29316807e-01 9.61713016e-01 9.65320528e-01
1.03351748e+00 6.55316591e-01 8.25365901e-01 2.59162217e-01
-2.69502223e-01 -2.43800610e-01 7.53771484e-01 8.20820808e-01
-1.70376062e-01 -6.42050505e-01 -7.63089657e-01 4.94243771e-01
-2.23366976e+00 -1.23950005e+00 -1.09808996e-01 2.08185768e+00
3.72276932e-01 6.99793994e-01 1.82513714e-01 3.67028683e-01
4.31441516e-01 3.74333024e-01 -8.02880883e-01 -3.93548667e-01
-4.15645353e-02 -4.63366657e-01 7.27411926e-01 6.03123367e-01
-9.33093607e-01 7.38268673e-01 5.41878939e+00 7.22875416e-01
-1.24893379e+00 -3.73679283e-03 6.37888134e-01 -3.20722401e-01
-2.75807530e-01 -4.50013094e-02 -7.32311010e-01 8.57617795e-01
1.30687714e+00 -3.52743775e-01 3.55920672e-01 5.85243583e-01
8.09256315e-01 -2.03854784e-01 -1.00002885e+00 8.96349609e-01
-6.00781180e-02 -1.17664516e+00 -8.04196820e-02 -1.99327916e-01
1.88750178e-01 3.79941881e-01 -6.28695777e-03 2.17908248e-01
-6.71619922e-02 -6.75212920e-01 1.29033482e+00 8.40734780e-01
5.17869949e-01 -6.43525720e-01 7.09379554e-01 4.48371232e-01
-1.44049442e+00 -4.04504746e-01 -3.78606826e-01 -4.45787102e-01
4.52537209e-01 5.58694720e-01 -6.10772550e-01 5.43610871e-01
1.03195310e+00 1.02933407e+00 -5.11934757e-01 1.26695275e+00
7.48156616e-03 8.21331918e-01 -3.48421365e-01 1.80392101e-01
2.43482981e-02 1.06875673e-01 7.06258237e-01 1.50478578e+00
5.09771287e-01 1.06335972e-02 1.60545707e-01 4.81530458e-01
1.90139726e-01 -7.52736032e-02 -3.05619389e-01 3.84060711e-01
7.13233173e-01 1.12428665e+00 -4.73450035e-01 -5.01756251e-01
-5.94516397e-01 8.17308426e-01 4.56931368e-02 3.09995323e-01
-1.14065969e+00 -3.29204857e-01 9.98174191e-01 3.58098835e-01
5.65534353e-01 -4.28382963e-01 -2.89958358e-01 -1.10903084e+00
3.54725927e-01 -5.92934549e-01 3.22036684e-01 -6.36847258e-01
-9.73943412e-01 6.05355084e-01 -1.40096053e-01 -1.58790505e+00
2.15030704e-02 -2.23347902e-01 -7.17863679e-01 4.15201157e-01
-2.26297617e+00 -8.46485972e-01 -7.58000195e-01 3.92433256e-01
8.56928885e-01 2.05047414e-01 1.89229712e-01 5.81082761e-01
-5.36288261e-01 5.25249243e-01 -2.99316615e-01 -3.28929067e-01
8.32802534e-01 -1.05822837e+00 6.59939229e-01 1.16612983e+00
-3.30378264e-01 -1.49062872e-01 1.12514699e+00 -1.95250794e-01
-1.34426498e+00 -1.18007767e+00 8.09544325e-01 -3.03430974e-01
7.16750145e-01 -2.95457780e-01 -9.22488213e-01 3.38364661e-01
1.81726664e-01 2.14509323e-01 3.24708819e-01 -5.08413255e-01
-4.57712889e-01 -5.81165612e-01 -8.21816981e-01 4.69757497e-01
1.34991634e+00 -1.54484697e-02 -8.64543840e-02 -9.03786421e-02
1.18563867e+00 -3.61031473e-01 -6.67982042e-01 4.57116485e-01
7.62781560e-01 -1.56466353e+00 9.05327022e-01 -5.99153116e-02
4.95375216e-01 -6.00970864e-01 -3.59239101e-01 -9.07119513e-01
-2.43385613e-01 -6.91554368e-01 -5.05313814e-01 1.07754421e+00
2.13594794e-01 -7.01042593e-01 5.75459421e-01 2.67121851e-01
-4.82742012e-01 -7.27365375e-01 -8.59303355e-01 -8.34890902e-01
-3.68698239e-01 -9.73511577e-01 6.87152386e-01 3.58898222e-01
-5.64557500e-02 1.37904122e-01 -3.89251173e-01 4.96752143e-01
6.68531537e-01 8.03789347e-02 8.07456493e-01 -8.12058568e-01
-2.95635343e-01 -4.81059223e-01 -6.98109508e-01 -1.55882967e+00
-2.06300229e-01 -6.12843931e-01 2.63042897e-01 -1.39212751e+00
5.18046767e-02 -4.61231172e-01 -5.84608793e-01 8.22689906e-02
-1.67386934e-01 1.03934243e-01 5.45245290e-01 2.08121464e-01
-1.11852145e+00 7.54874766e-01 1.22131371e+00 1.65058836e-01
-5.18337637e-02 2.04453692e-01 -6.51292145e-01 5.56781173e-01
1.21043980e+00 -2.90642679e-01 -5.67313790e-01 -6.26695216e-01
2.09587872e-01 1.23279855e-01 4.30219114e-01 -1.48183715e+00
5.86823404e-01 -2.93192118e-01 3.94372232e-02 -6.09164059e-01
5.04724622e-01 -6.55100942e-01 -2.41856724e-01 4.31702316e-01
-3.75552028e-01 1.97825119e-01 1.45404696e-01 8.76014829e-01
-1.93716794e-01 1.87729776e-01 6.39767885e-01 1.69895142e-01
-1.18404722e+00 5.67132890e-01 -4.20445651e-01 -1.15362685e-02
1.09663665e+00 -3.41562480e-01 -5.22774756e-01 -7.13308334e-01
-3.10839415e-01 8.35628927e-01 5.41158617e-01 7.01119900e-01
7.19854534e-01 -1.03585577e+00 -9.62356091e-01 9.47586671e-02
3.14960599e-01 -1.66096181e-01 5.00794232e-01 8.50116789e-01
-5.79897881e-01 5.13544679e-02 -1.43383071e-01 -7.48302877e-01
-1.01477432e+00 5.56937397e-01 2.50692189e-01 1.26310810e-01
-6.77804232e-01 8.35248053e-01 1.13276832e-01 9.97130424e-02
5.21779299e-01 -3.75117898e-01 -1.95704877e-01 -9.78043824e-02
9.04084325e-01 7.04945564e-01 -2.69427523e-02 -5.92986822e-01
-3.09358716e-01 2.03807682e-01 1.80190042e-01 -3.67485657e-02
1.16602612e+00 -8.86015296e-01 2.92661607e-01 6.60086095e-01
9.03478801e-01 -1.60294667e-01 -1.94600701e+00 -2.72575647e-01
-1.90170646e-01 -7.58071899e-01 1.67645067e-01 -5.68168938e-01
-9.43691373e-01 9.90634501e-01 7.64277816e-01 3.40083182e-01
1.40546763e+00 -1.47210672e-01 1.15285611e+00 8.21586922e-02
3.91673505e-01 -1.04735160e+00 4.06400152e-02 5.11969507e-01
7.20909655e-01 -1.40993702e+00 -2.50767052e-01 -5.02260566e-01
-7.50362515e-01 1.08225334e+00 7.06261694e-01 9.75568518e-02
6.92205548e-01 4.91437584e-01 2.72398710e-01 2.50207424e-01
-1.56392777e+00 -3.54365021e-01 -5.57851419e-02 5.07165313e-01
7.54194930e-02 1.38372362e-01 -7.59607404e-02 3.25879902e-01
-2.63544530e-01 1.26199499e-01 7.27254152e-01 8.64884377e-01
-7.59897649e-01 -7.72533774e-01 1.06394505e-02 4.10119295e-01
-2.84531504e-01 8.69599357e-02 4.71619636e-01 8.99108127e-02
2.76016027e-01 1.36837542e+00 4.52534467e-01 -7.02710986e-01
5.16236603e-01 -3.90734613e-01 -5.18525429e-02 -2.21803159e-01
-1.95137948e-01 -1.07017979e-01 1.15176745e-01 -1.15061975e+00
-3.68988961e-01 -7.12158740e-01 -1.15613973e+00 -5.72599471e-01
1.11087188e-01 -3.17247987e-01 5.99300742e-01 5.55851519e-01
6.46263421e-01 7.66189992e-01 9.22408879e-01 -8.80715847e-01
-4.21807677e-01 -6.11997485e-01 -2.57171571e-01 2.32172385e-01
7.03381062e-01 -4.95145530e-01 -4.81472611e-01 2.35740989e-01] | [8.445693969726562, -0.6712128520011902] |
9770a64e-ff12-4882-bd97-47bd3f2af072 | analysis-of-multilingual-sequence-to-sequence | 1811.03451 | null | http://arxiv.org/abs/1811.03451v1 | http://arxiv.org/pdf/1811.03451v1.pdf | Analysis of Multilingual Sequence-to-Sequence speech recognition systems | This paper investigates the applications of various multilingual approaches
developed in conventional hidden Markov model (HMM) systems to
sequence-to-sequence (seq2seq) automatic speech recognition (ASR). On a set
composed of Babel data, we first show the effectiveness of multi-lingual
training with stacked bottle-neck (SBN) features. Then we explore various
architectures and training strategies of multi-lingual seq2seq models based on
CTC-attention networks including combinations of output layer, CTC and/or
attention component re-training. We also investigate the effectiveness of
language-transfer learning in a very low resource scenario when the target
language is not included in the original multi-lingual training data.
Interestingly, we found multilingual features superior to multilingual models,
and this finding suggests that we can efficiently combine the benefits of the
HMM system with the seq2seq system through these multilingual feature
techniques. | ['Jan "Honza\'\' Černocký', 'Martin Karafiát', 'Murali Karthick Baskar', 'Takaaki Hori', 'Shinji Watanabe', 'Matthew Wiesner'] | 2018-11-07 | null | null | null | null | ['sequence-to-sequence-speech-recognition'] | ['speech'] | [-1.94046237e-02 -4.11698110e-02 4.58093807e-02 -3.82081062e-01
-1.56337500e+00 -6.88679814e-01 7.98202455e-01 -4.35895890e-01
-6.88678145e-01 8.66122544e-01 5.02664149e-01 -9.51682508e-01
5.07847309e-01 4.64665592e-02 -8.12013268e-01 -7.00044215e-01
-1.76025331e-02 6.14002347e-01 3.25584668e-03 -4.75028962e-01
-1.56083271e-01 3.44516724e-01 -1.26194143e+00 4.23568904e-01
6.09412313e-01 3.03843737e-01 6.94649041e-01 1.04796648e+00
-7.17899203e-02 8.10535073e-01 -6.63655460e-01 -3.88340712e-01
1.89569388e-02 -4.50423628e-01 -6.54273808e-01 -2.23018393e-01
4.48141515e-01 3.28876823e-02 -1.66635528e-01 7.71130860e-01
7.17892349e-01 1.73087984e-01 2.67374039e-01 -6.73774362e-01
-5.25284648e-01 6.73451066e-01 -2.18900200e-02 4.21493858e-01
5.09047210e-01 1.80407256e-01 9.60073709e-01 -9.30876613e-01
6.22724116e-01 1.56766725e+00 7.71886885e-01 4.55196321e-01
-1.10367346e+00 -5.05042136e-01 -1.18258763e-02 5.06486952e-01
-1.13703990e+00 -9.32909489e-01 1.53057307e-01 -2.65365094e-01
1.81905842e+00 1.08080454e-01 3.15954268e-01 1.49842167e+00
3.15608352e-01 9.22758758e-01 1.27425838e+00 -9.34008598e-01
-2.13060245e-01 2.99957097e-01 -1.34332612e-01 5.08361459e-01
-5.34776568e-01 5.05763233e-01 -5.23731053e-01 5.73622845e-02
4.16962832e-01 -4.31327790e-01 9.97016020e-03 1.27156436e-01
-1.26681912e+00 9.43610370e-01 -5.52791134e-02 6.23720765e-01
-2.54102081e-01 3.65657033e-03 7.99411893e-01 5.69804907e-01
3.14394265e-01 3.37505579e-01 -8.75145316e-01 -4.68118370e-01
-9.46576774e-01 -5.20556569e-01 7.55545318e-01 1.18935061e+00
7.42505848e-01 5.28289437e-01 -5.32709695e-02 1.11342144e+00
1.73293516e-01 7.32102454e-01 7.50411570e-01 -6.86093867e-01
6.35333538e-01 -3.47382903e-01 -3.71683329e-01 9.18342769e-02
-1.56872764e-01 -2.84346759e-01 -3.67305040e-01 -3.95514548e-01
3.69186439e-02 -3.65340859e-01 -9.65356767e-01 1.46987498e+00
1.09493602e-02 1.03092946e-01 5.28426886e-01 4.49567676e-01
3.02233845e-01 9.03451145e-01 2.67767966e-01 -1.59688219e-01
1.20099437e+00 -1.26681399e+00 -8.49638641e-01 -7.30965436e-02
1.27879798e+00 -1.09249270e+00 9.67575490e-01 2.27182508e-01
-7.77358294e-01 -7.20114231e-01 -7.58155942e-01 -1.80075467e-01
-6.32272780e-01 2.17069477e-01 1.63780198e-01 8.79528165e-01
-1.38213038e+00 2.49672145e-01 -8.18386495e-01 -7.37980843e-01
-4.08578068e-01 3.39073271e-01 -4.06859279e-01 -3.46815437e-01
-1.47354209e+00 1.35740173e+00 3.63785446e-01 1.98558524e-01
-1.14733577e+00 -1.46125779e-01 -1.05294192e+00 -8.20021853e-02
9.92236063e-02 3.15104760e-02 1.56106913e+00 -9.36031044e-01
-1.91573226e+00 3.27605784e-01 -4.51859891e-01 -4.69578832e-01
1.01137958e-01 -2.29746521e-01 -7.54384935e-01 -6.42642900e-02
9.05705616e-02 6.06725752e-01 5.13492048e-01 -8.93768966e-01
-6.99056506e-01 -5.39438799e-02 -3.93072993e-01 2.64765978e-01
8.80559732e-04 5.71573555e-01 -3.42054844e-01 -5.24060547e-01
-5.77667594e-01 -1.10297179e+00 -1.96636349e-01 -1.24038339e+00
-2.88864642e-01 -2.22114354e-01 7.29795575e-01 -1.23522472e+00
1.12677741e+00 -2.14870167e+00 8.01060572e-02 -3.90126705e-02
-6.82516158e-01 6.30613923e-01 -5.62687635e-01 8.28865826e-01
-1.81145564e-01 1.72613546e-01 -2.03130636e-02 -4.79383230e-01
-1.56176731e-01 5.43843508e-01 3.12095135e-02 1.49441883e-01
3.39771509e-01 1.05723763e+00 -7.47286916e-01 -3.85077655e-01
3.43708187e-01 7.43536890e-01 -3.29203457e-01 2.32836664e-01
2.43405942e-02 6.70125723e-01 -7.08416775e-02 4.67061460e-01
2.15187609e-01 4.21263903e-01 5.47024846e-01 2.13325247e-01
-5.52663743e-01 9.36637700e-01 -4.49760795e-01 1.66944146e+00
-1.00012994e+00 8.94784987e-01 1.11283295e-01 -6.92586362e-01
8.25755358e-01 8.26682448e-01 -1.96856465e-02 -1.01656246e+00
-1.76854566e-01 5.73056042e-01 2.79584169e-01 -5.03025651e-01
5.56288421e-01 -4.46716189e-01 -1.55569717e-01 2.25538179e-01
4.69197363e-01 9.94380638e-02 -7.20193833e-02 -3.38055342e-02
1.00062847e+00 3.57954115e-01 1.58571213e-01 -2.72825778e-01
5.52691996e-01 -7.44971782e-02 2.71017462e-01 6.78433180e-01
-9.22843963e-02 3.48832428e-01 1.80789798e-01 -7.13953078e-02
-1.41342163e+00 -6.10565364e-01 -8.30103159e-02 1.57957482e+00
-9.62534726e-01 -2.97421753e-01 -7.51727939e-01 -5.41058362e-01
-2.99427509e-01 1.02972329e+00 -3.32804531e-01 2.35176906e-01
-8.56423557e-01 -5.26169896e-01 9.86436486e-01 4.71387625e-01
4.46910458e-03 -1.28961825e+00 -4.87542190e-02 6.36819601e-01
-2.01537728e-01 -1.44582093e+00 -6.65972471e-01 7.99652100e-01
-5.82119346e-01 -3.82278264e-01 -8.10002446e-01 -8.80817652e-01
-1.20873861e-01 1.75306424e-02 8.93674254e-01 -2.94516325e-01
2.19484478e-01 3.42658550e-01 -7.46448576e-01 -9.06574950e-02
-1.09683239e+00 5.83774805e-01 2.39647120e-01 -1.23767778e-01
4.42371756e-01 -2.31687903e-01 2.37156972e-01 1.13138743e-01
-4.55608994e-01 -2.42369860e-01 9.61912632e-01 1.11503720e+00
1.92414984e-01 -5.57468295e-01 8.20023835e-01 -6.92494333e-01
2.92186379e-01 -3.55258346e-01 -4.81357545e-01 4.80355352e-01
-2.84095258e-01 2.08132103e-01 7.42134690e-01 -3.37872773e-01
-1.17513609e+00 9.71532464e-02 -6.85171366e-01 -5.34612000e-01
-3.27054590e-01 7.63630807e-01 -3.15506876e-01 1.67488009e-01
1.95180610e-01 5.14916480e-01 -7.92749897e-02 -7.87497520e-01
4.88165319e-01 1.13686156e+00 3.34118903e-01 -4.87319708e-01
1.49561644e-01 -2.88409263e-01 -4.77393448e-01 -1.32759416e+00
-3.60785991e-01 -6.41000986e-01 -1.09041703e+00 3.46149579e-02
1.06677806e+00 -1.07704449e+00 -4.16120142e-01 3.35771024e-01
-1.24912953e+00 -6.87154710e-01 -3.89251001e-02 8.25378597e-01
-5.99286914e-01 4.50282246e-01 -9.85665739e-01 -9.59769249e-01
-3.24258283e-02 -1.47048044e+00 1.24226201e+00 -4.69432026e-01
5.61747737e-02 -1.25087333e+00 3.18728685e-01 3.99061799e-01
4.85891968e-01 -5.68495572e-01 9.92908716e-01 -1.06851923e+00
-4.04686093e-01 -7.76674971e-02 1.95709318e-01 6.39883697e-01
2.57720258e-02 -5.82278036e-02 -1.24022567e+00 -5.62352002e-01
-2.09141672e-01 -4.51570839e-01 6.38032734e-01 2.66521484e-01
3.88156444e-01 -3.80239695e-01 -7.22378939e-02 3.78850251e-01
1.23988378e+00 4.56880778e-01 6.58982277e-01 2.93061078e-01
8.82295191e-01 7.03788936e-01 3.30982000e-01 -1.47381529e-01
5.17014980e-01 9.18330193e-01 -3.18191439e-01 -1.82475939e-01
-3.06675345e-01 -3.16598326e-01 1.14239788e+00 1.93330419e+00
2.64157474e-01 -2.17984110e-01 -1.17324364e+00 6.52907908e-01
-1.46302092e+00 -9.30540562e-01 -1.94870844e-01 2.19607306e+00
8.13610256e-01 -9.54806581e-02 1.72406033e-01 -3.14052165e-01
8.05417955e-01 -1.61468327e-01 -3.05003710e-02 -1.11553371e+00
-2.84360558e-01 3.41539264e-01 6.35629416e-01 8.23605895e-01
-8.03172827e-01 1.48753846e+00 7.06295919e+00 1.10610056e+00
-1.21266973e+00 5.81944585e-01 3.20513040e-01 1.05478875e-01
-1.45864263e-01 1.40302509e-01 -1.33342636e+00 3.24244142e-01
2.16078353e+00 2.94804186e-01 4.41524416e-01 5.50384045e-01
3.23563248e-01 -4.70548160e-02 -1.12679887e+00 2.91782558e-01
1.51638985e-01 -7.85534263e-01 4.86823320e-02 1.99851811e-01
6.51853681e-01 7.45893896e-01 -7.99524412e-02 7.57399857e-01
5.74925721e-01 -1.01918137e+00 6.45075262e-01 1.39094815e-01
1.10062706e+00 -8.63287866e-01 9.67693925e-01 3.16120982e-01
-1.18455470e+00 6.25852570e-02 -1.63027853e-01 3.40493411e-01
3.53399009e-01 -1.38712391e-01 -1.44152284e+00 8.77491653e-01
4.76499528e-01 5.92581272e-01 -5.22622049e-01 7.55651534e-01
-8.37054253e-02 1.00815690e+00 -2.68379658e-01 -1.23744622e-01
8.84225190e-01 -3.11137922e-03 4.35601562e-01 1.96144342e+00
3.62577707e-01 -5.23478568e-01 3.53696644e-02 3.95361073e-02
3.48028064e-01 3.23362738e-01 -7.65927374e-01 -1.00061283e-01
4.54011500e-01 8.35861444e-01 -1.10411793e-01 -5.24532974e-01
-8.18872392e-01 9.19922233e-01 4.28246498e-01 4.28403109e-01
-4.47054058e-01 -2.44949386e-01 3.92569542e-01 -3.57696891e-01
7.42571831e-01 -5.62687755e-01 2.77345538e-01 -1.05034328e+00
-3.53427112e-01 -1.34211183e+00 1.17689997e-01 -8.73304367e-01
-1.11886144e+00 1.08296752e+00 -1.34676367e-01 -8.42752635e-01
-7.94367969e-01 -6.05959654e-01 -3.08684587e-01 1.29236269e+00
-1.73971176e+00 -1.46162701e+00 7.54688501e-01 5.09120166e-01
9.65348005e-01 -4.89374965e-01 1.04746747e+00 4.68223900e-01
-6.04695737e-01 7.72360146e-01 7.15090215e-01 1.99313059e-01
6.48774624e-01 -1.11764455e+00 7.82113910e-01 8.14258039e-01
2.74302334e-01 5.01267910e-01 2.78117299e-01 -6.85640812e-01
-1.26329958e+00 -1.06408429e+00 1.39563298e+00 -5.79510391e-01
8.85314286e-01 -7.75522590e-01 -1.04193318e+00 1.16109657e+00
8.18028271e-01 -6.52483881e-01 6.45473599e-01 3.45693201e-01
-1.57702804e-01 3.39507282e-01 -3.97973388e-01 4.81998265e-01
5.51487327e-01 -1.21973956e+00 -5.77825904e-01 1.27356440e-01
8.69996846e-01 4.39512618e-02 -9.21124279e-01 1.72479391e-01
5.63213468e-01 -5.75725853e-01 5.17548919e-01 -7.26128817e-01
-9.81891304e-02 1.63762525e-01 -6.27220631e-01 -1.62670970e+00
-2.70704001e-01 -7.70761669e-01 5.19322217e-01 1.16078722e+00
9.31192577e-01 -6.35105550e-01 1.12953141e-01 -1.40403450e-01
-8.17711890e-01 -3.21124315e-01 -1.41837716e+00 -1.13707638e+00
4.85472202e-01 -5.52366197e-01 4.06457275e-01 9.94721532e-01
-1.85160302e-02 6.41004562e-01 -5.48967600e-01 8.31714943e-02
1.23670362e-01 -6.05721951e-01 4.94940847e-01 -5.54112732e-01
-3.58967960e-01 -1.63209930e-01 -1.38822198e-01 -9.43626225e-01
4.45942849e-01 -1.12089121e+00 4.88026559e-01 -1.01197660e+00
-5.00625111e-02 -3.96783441e-01 -2.72812992e-01 6.82159007e-01
-1.19468316e-01 -2.34209776e-01 2.85655081e-01 1.34120017e-01
-3.60889941e-01 6.71223462e-01 9.40852702e-01 1.53342903e-01
-9.40483510e-02 -2.76754051e-01 -1.82035118e-02 6.40124232e-02
5.72641373e-01 -6.09723151e-01 9.08919945e-02 -4.82183099e-01
-3.47288072e-01 5.12324154e-01 -4.63951826e-02 -8.46664429e-01
1.53803989e-01 3.61312002e-01 -6.25860095e-02 -5.64861178e-01
4.37938452e-01 -6.45734310e-01 5.04610986e-02 3.52878124e-01
-3.41216147e-01 4.40644592e-01 5.66791117e-01 3.40526462e-01
-4.13097292e-01 -1.71984732e-01 6.75122559e-01 -1.54416144e-01
-7.60802209e-01 -2.35562876e-01 -1.17016602e+00 -2.80639201e-01
5.19508958e-01 -6.08585961e-02 -1.30470186e-01 -3.56158495e-01
-8.88456285e-01 1.09103449e-01 1.57887295e-01 6.36652946e-01
1.86940148e-01 -1.27141893e+00 -9.35265064e-01 4.86443102e-01
1.66074365e-01 -8.55546474e-01 1.58397973e-01 1.07226193e+00
-2.75609493e-01 1.34854329e+00 -1.87484428e-01 -6.39739871e-01
-1.21459544e+00 5.09608328e-01 4.04125571e-01 -3.76276582e-01
-2.72584915e-01 6.59347415e-01 8.46504048e-03 -1.04314721e+00
1.62338942e-01 -1.30710989e-01 -2.52633300e-02 5.53100705e-02
-1.27699338e-02 2.13501886e-01 4.94908363e-01 -1.13651109e+00
-4.78956968e-01 2.21151575e-01 -2.93632627e-01 -7.47402310e-01
1.06135082e+00 -3.07584614e-01 3.40989679e-01 8.45474660e-01
1.45838487e+00 1.20963650e-02 -9.10747528e-01 -1.28360152e-01
3.75193208e-01 1.51804864e-01 2.58950870e-02 -8.56608272e-01
-5.85530877e-01 1.33682680e+00 5.78745544e-01 -3.56950499e-02
6.11707807e-01 -8.05461034e-02 7.67168880e-01 5.09971917e-01
3.91739219e-01 -1.12747884e+00 -3.01865011e-01 1.16112232e+00
6.68278635e-01 -1.10604978e+00 -6.37837410e-01 8.31107795e-02
-1.00177586e+00 9.61263120e-01 3.03782195e-01 3.05315018e-01
5.70603967e-01 3.80500078e-01 5.35576880e-01 1.80582196e-01
-1.09846997e+00 -4.41838533e-01 -1.34798540e-02 6.59164608e-01
8.05364013e-01 3.55171442e-01 3.19054425e-02 8.70902836e-02
-3.55657637e-01 -3.77960652e-01 5.23575604e-01 8.79228592e-01
-2.58202374e-01 -1.48667288e+00 -3.58744532e-01 7.25605115e-02
-5.56026340e-01 -9.26252604e-01 -1.19620949e-01 9.09478426e-01
-1.59099296e-01 9.96526659e-01 1.94704011e-02 -4.93879974e-01
1.76351890e-01 8.02361250e-01 2.84228861e-01 -8.46792877e-01
-1.04286540e+00 7.34124124e-01 5.51569402e-01 -4.03213531e-01
-3.06335479e-01 -1.17568529e+00 -6.22924626e-01 6.51254924e-03
-5.38095832e-01 3.45072716e-01 1.01226258e+00 1.16956723e+00
3.59753907e-01 5.03690660e-01 6.75229073e-01 -7.90070713e-01
-5.01609385e-01 -1.45504427e+00 -4.93794322e-01 -1.84849486e-01
5.75365961e-01 -2.19497576e-01 -2.31777787e-01 2.15811819e-01] | [14.376910209655762, 7.008882522583008] |
afc5cc65-5318-459e-a7b0-baa33e3b678b | collaborative-blind-image-deblurring | 2305.16034 | null | https://arxiv.org/abs/2305.16034v1 | https://arxiv.org/pdf/2305.16034v1.pdf | Collaborative Blind Image Deblurring | Blurry images usually exhibit similar blur at various locations across the image domain, a property barely captured in nowadays blind deblurring neural networks. We show that when extracting patches of similar underlying blur is possible, jointly processing the stack of patches yields superior accuracy than handling them separately. Our collaborative scheme is implemented in a neural architecture with a pooling layer on the stack dimension. We present three practical patch extraction strategies for image sharpening, camera shake removal and optical aberration correction, and validate the proposed approach on both synthetic and real-world benchmarks. For each blur instance, the proposed collaborative strategy yields significant quantitative and qualitative improvements. | ['Gabriele Facciolo', 'Jean-Michel Morel', 'Thomas Eboli'] | 2023-05-25 | null | null | null | null | ['deblurring', 'blind-image-deblurring'] | ['computer-vision', 'computer-vision'] | [ 1.20068192e-01 -5.13458371e-01 4.27280575e-01 -2.88891256e-01
-6.80688381e-01 -7.28574038e-01 4.90478724e-01 -4.71386015e-01
-2.47261018e-01 7.59720206e-01 6.56598806e-01 -5.07745966e-02
-4.07287538e-01 8.23775865e-03 -7.34164774e-01 -6.93963885e-01
2.04790942e-03 -6.07856631e-01 -2.86010243e-02 1.80609792e-01
5.59121490e-01 5.40374935e-01 -1.20859623e+00 2.61273563e-01
1.16023743e+00 9.31686640e-01 2.85400182e-01 1.07415640e+00
6.61772728e-01 9.93164837e-01 -7.04425693e-01 -3.99253100e-01
4.77607071e-01 -1.30129144e-01 -6.78286910e-01 3.27251554e-01
1.28210211e+00 -7.22619116e-01 -7.18921959e-01 1.51073742e+00
6.50074959e-01 6.98477626e-02 6.14730954e-01 -6.08333647e-01
-1.32736397e+00 3.39580894e-01 -7.68520951e-01 4.77923065e-01
2.07872093e-01 4.02119040e-01 6.42749608e-01 -1.09720731e+00
2.92992800e-01 7.56822109e-01 1.06945848e+00 2.47074366e-01
-1.21732295e+00 -2.35824898e-01 -1.23577207e-01 4.96663570e-01
-1.24732244e+00 -6.89373612e-01 7.50332415e-01 -4.81473744e-01
9.76879478e-01 1.93217814e-01 3.92509758e-01 8.93542051e-01
4.96455014e-01 6.72269642e-01 1.42417860e+00 -5.63301705e-02
6.82332516e-02 -1.83492869e-01 4.46566194e-01 3.38303894e-01
5.95586777e-01 2.69303203e-01 -4.84761238e-01 5.96328601e-02
8.34547043e-01 2.01340273e-01 -1.14033353e+00 -7.05554038e-02
-1.23901379e+00 2.34544098e-01 8.76510680e-01 1.33340061e-01
-5.00514925e-01 4.66944456e-01 -3.54639674e-03 3.67947221e-01
5.99880099e-01 8.39510381e-01 -5.33929467e-01 5.13016172e-02
-1.34907997e+00 2.19785556e-01 6.00500286e-01 1.06394899e+00
6.33247674e-01 -2.45569438e-01 -4.28793788e-01 9.45569515e-01
-6.00716937e-03 4.89656508e-01 4.37548637e-01 -1.10628486e+00
2.30898976e-01 9.26112086e-02 6.84656978e-01 -8.26748431e-01
-2.44031444e-01 -4.36286688e-01 -1.26329017e+00 2.42881596e-01
5.08909702e-01 -5.88970363e-01 -1.10294557e+00 1.33565176e+00
-2.28747249e-01 6.27396107e-01 -8.44486654e-02 1.31239021e+00
6.39248967e-01 2.39075676e-01 -4.73383218e-01 -1.55650020e-01
1.43796563e+00 -1.26984048e+00 -9.68514204e-01 -4.68733549e-01
-1.03777431e-01 -1.17171574e+00 4.58434820e-01 4.71536934e-01
-1.38151240e+00 -5.05500913e-01 -9.98088062e-01 -5.21198452e-01
-1.58444792e-01 4.49954093e-01 5.41247904e-01 5.42953968e-01
-1.61083639e+00 7.51490057e-01 -4.22423720e-01 -1.56277359e-01
8.52726102e-01 1.63522214e-01 -3.18431556e-01 -3.75389457e-01
-9.41253960e-01 1.09203291e+00 -2.79019147e-01 6.36320233e-01
-8.28636944e-01 -8.61401081e-01 -7.84324765e-01 1.57189488e-01
-1.27070189e-01 -1.07987201e+00 1.31002939e+00 -7.73054421e-01
-1.42377758e+00 4.64742839e-01 -5.14185429e-01 -6.87219441e-01
5.72876275e-01 -1.00878716e+00 -3.21735650e-01 -5.56343049e-03
-3.06999087e-01 3.58225763e-01 1.82404220e+00 -1.21885252e+00
-3.14639509e-01 -1.09808430e-01 1.73898656e-02 1.30687222e-01
-8.53285715e-02 2.09450543e-01 -4.00716960e-01 -1.05719769e+00
-2.63957307e-02 -6.59945250e-01 -1.82674631e-01 -1.28222257e-03
-3.57039154e-01 3.29676211e-01 6.91552639e-01 -1.15649796e+00
1.08059359e+00 -2.12875938e+00 4.10690218e-01 -3.96178067e-01
6.71641469e-01 2.46167853e-01 -5.95967174e-02 2.28659678e-02
-2.80080974e-01 -3.61882001e-01 -5.13615727e-01 -6.06856823e-01
-2.06656039e-01 -3.34255517e-01 -4.55359250e-01 9.15487587e-01
3.19656640e-01 9.67903495e-01 -7.94472694e-01 2.72114933e-01
2.61357427e-01 6.11893117e-01 -6.29755378e-01 3.86976123e-01
7.67954737e-02 1.96485505e-01 3.23362112e-01 5.05930305e-01
1.35298479e+00 -4.10038829e-01 -3.20353717e-01 -6.50131106e-01
-1.63807929e-01 -1.30443946e-01 -9.09655094e-01 1.76186585e+00
-4.43021476e-01 1.36379063e+00 2.24169537e-01 -3.09405982e-01
3.89676899e-01 2.82589167e-01 -1.55551717e-01 -3.86767499e-02
1.67146683e-01 2.03645915e-01 -1.48889363e-01 -7.22357273e-01
8.66558790e-01 1.48371935e-01 3.95608783e-01 3.84463966e-01
2.34426335e-01 -3.75169635e-01 -3.45981091e-01 -7.84564614e-02
1.30265236e+00 -2.71916866e-01 6.24807030e-02 -4.60428298e-01
4.49417293e-01 -4.88665223e-01 -1.13079332e-01 9.46231246e-01
-5.41059256e-01 1.17663848e+00 2.27054834e-01 -3.77557218e-01
-1.02350104e+00 -8.44550848e-01 -1.37941509e-01 5.09211540e-01
5.01221061e-01 7.83554465e-02 -1.13070464e+00 -4.27264512e-01
3.02631650e-02 3.23649704e-01 -6.93192959e-01 -8.99003148e-02
-4.91774440e-01 -9.38048959e-01 2.95848429e-01 4.15992677e-01
8.49921465e-01 -6.43015504e-01 -4.60712940e-01 -1.86144486e-01
-1.53348986e-02 -1.11591339e+00 -1.04690766e+00 -7.76632801e-02
-6.18191779e-01 -1.29349685e+00 -1.35306561e+00 -9.52749074e-01
7.74987996e-01 9.40160990e-01 8.06296229e-01 -1.14870563e-01
-2.26097107e-01 2.65703708e-01 1.52975261e-01 4.88034412e-02
-1.05018262e-02 -2.35299364e-01 1.26585737e-02 1.21165305e-01
1.68573678e-01 -5.40607810e-01 -1.17565787e+00 8.87974054e-02
-8.00484061e-01 2.18335949e-02 6.47479594e-01 9.55472589e-01
-1.57773495e-01 2.62804538e-01 1.62892967e-01 -4.45762962e-01
1.10653353e+00 -3.22948337e-01 -6.86873019e-01 1.11586615e-01
-4.85914022e-01 -1.13923438e-02 3.56809765e-01 -4.56788599e-01
-1.40252078e+00 -2.63598025e-01 2.92481780e-01 -7.07639456e-01
-4.09883738e-01 2.41703749e-01 1.60854012e-01 -4.64213997e-01
8.67729902e-01 1.43981099e-01 -3.26269418e-02 -7.69883513e-01
5.73437035e-01 8.27206969e-01 9.07980323e-01 9.08958018e-02
8.94774437e-01 4.53822017e-01 -3.01747352e-01 -7.90801406e-01
-8.40402722e-01 -5.42435169e-01 -5.10387659e-01 -5.41709177e-02
6.82343423e-01 -1.21837139e+00 -5.93044579e-01 1.47803795e+00
-1.70189071e+00 -3.17054689e-01 1.06553443e-01 4.06411648e-01
-2.88214743e-01 6.17059469e-01 -9.22523022e-01 -4.46345627e-01
-3.77157211e-01 -1.30186439e+00 9.83333290e-01 5.51976442e-01
2.53348798e-02 -8.33852649e-01 5.47667295e-02 2.19880879e-01
9.27073181e-01 -3.89594495e-01 3.96147609e-01 9.09256786e-02
-8.37446690e-01 -1.58538103e-01 -9.53065991e-01 7.21695423e-01
7.14292586e-01 -2.31346935e-01 -1.55033255e+00 -4.63277578e-01
4.90785837e-01 2.98648000e-01 1.27606368e+00 8.31840038e-01
1.15647173e+00 -5.32388926e-01 7.23597556e-02 9.53309059e-01
1.59515929e+00 -2.03529313e-01 7.88545251e-01 2.04514518e-01
9.27142441e-01 2.57170200e-01 -2.22709328e-01 1.23178244e-01
8.80357772e-02 5.40764809e-01 3.78941000e-01 -2.21391693e-01
-7.55930662e-01 5.17883062e-01 4.14754510e-01 4.33932424e-01
-1.09854013e-01 -8.47695917e-02 -6.68463469e-01 9.16177690e-01
-1.69669414e+00 -9.05846477e-01 -1.45303011e-01 2.04028320e+00
1.16843581e+00 -2.33389080e-01 -3.89630258e-01 -1.97963387e-01
8.68446469e-01 3.17625850e-01 -5.24612308e-01 -1.56903997e-01
-5.45523405e-01 3.19650531e-01 8.54530036e-01 9.81408954e-01
-1.37110662e+00 7.27446377e-01 7.04857016e+00 3.53112966e-01
-1.09753788e+00 1.74057886e-01 4.38524336e-01 -2.01691106e-01
1.51968017e-01 -2.91709125e-01 -2.82444984e-01 6.15721107e-01
6.12606168e-01 -4.76258919e-02 1.06511736e+00 4.57877189e-01
2.46574819e-01 -2.41911262e-01 -1.12427425e+00 1.22523785e+00
2.04218358e-01 -1.47252011e+00 -3.14545989e-01 -2.81324416e-01
1.26336169e+00 3.49880368e-01 3.68874282e-01 -4.51448083e-01
1.86184734e-01 -1.12556648e+00 6.93122983e-01 1.08603716e+00
6.57478333e-01 -2.75880456e-01 9.98213172e-01 -1.99443430e-01
-5.00849247e-01 -1.86716110e-01 -2.39056081e-01 -3.64652798e-02
4.46556509e-02 9.34368193e-01 -4.48429763e-01 5.07377684e-01
9.62979436e-01 1.19100893e+00 -9.81617689e-01 1.71007323e+00
-4.02101487e-01 4.96617019e-01 1.39985532e-01 4.79538023e-01
-1.00527018e-01 -2.28527069e-01 8.49924207e-01 1.39200389e+00
4.29163516e-01 -1.94661498e-01 -9.80958462e-01 8.75454843e-01
-1.91036075e-01 -7.94111609e-01 -3.72078538e-01 2.75660753e-01
2.22226918e-01 1.18307543e+00 -6.13208003e-02 -3.10511470e-01
-5.57249427e-01 1.59948540e+00 1.04197443e-01 8.79717231e-01
-7.57871449e-01 -6.56649470e-01 1.23751974e+00 -3.45105886e-01
6.14541531e-01 -2.01831117e-01 -6.84043169e-01 -1.42924798e+00
7.47339875e-02 -8.54079008e-01 -2.30064124e-01 -1.29473257e+00
-1.64275217e+00 6.31012082e-01 -3.13523531e-01 -1.15557742e+00
2.15739250e-01 -9.31551218e-01 -6.49575591e-01 1.41333485e+00
-1.86228251e+00 -9.69726682e-01 -8.05034637e-01 5.72156549e-01
5.72298527e-01 8.53476450e-02 3.65662813e-01 2.84033120e-01
-6.65449262e-01 3.95886838e-01 2.46838361e-01 -5.69021851e-02
1.13645291e+00 -1.47849202e+00 6.54617906e-01 1.44874120e+00
-2.35411435e-01 9.67826486e-01 1.12282729e+00 -3.30203116e-01
-1.35070491e+00 -1.18463922e+00 6.22319520e-01 -5.38735867e-01
8.62302065e-01 -2.24270262e-02 -1.18849945e+00 5.64994991e-01
1.02421951e+00 1.36366278e-01 -9.03694928e-02 -1.15937263e-01
-5.68126440e-01 -5.33061549e-02 -1.02456474e+00 5.70761144e-01
9.00720775e-01 -9.24523830e-01 -7.73328125e-01 3.72069865e-01
6.84480011e-01 -7.09529042e-01 -5.76365948e-01 1.35712177e-01
4.35290694e-01 -1.19676709e+00 1.04914165e+00 -2.80332506e-01
7.56638885e-01 -6.19094670e-01 1.43317819e-01 -1.83375442e+00
-6.49851918e-01 -1.14736104e+00 -5.24929643e-01 7.49342024e-01
7.43421614e-02 -7.42688417e-01 3.76261562e-01 4.38491821e-01
-2.48028770e-01 -3.25442076e-01 -6.70851946e-01 -4.29143757e-01
-2.42425017e-02 -1.68614820e-01 6.72109485e-01 7.39367068e-01
-3.60642195e-01 7.94014037e-02 -6.67859912e-01 7.07605779e-01
8.59513760e-01 1.03089131e-01 5.29313326e-01 -7.34963655e-01
-1.90626115e-01 -6.69265389e-01 -2.31447458e-01 -1.42179668e+00
-1.78789079e-01 -2.71780789e-01 3.53960186e-01 -1.47722960e+00
2.49788105e-01 8.23662952e-02 -2.42300510e-01 9.17957723e-02
-5.88809967e-01 4.57873374e-01 -1.63442612e-01 3.80527854e-01
-2.56522417e-01 3.11208606e-01 1.33038020e+00 -3.54188621e-01
-3.39808390e-02 -1.47267625e-01 -8.53176177e-01 7.14844584e-01
5.21984100e-01 1.08022213e-01 -1.19169220e-01 -1.16336894e+00
2.17011839e-01 -2.26738289e-01 9.43452716e-01 -9.91078854e-01
5.92796087e-01 1.13410205e-01 5.65560937e-01 -2.26989925e-01
2.37456888e-01 -8.62979293e-01 1.97020292e-01 1.41159415e-01
-1.60499543e-01 -2.36772999e-01 4.43787634e-01 6.05834484e-01
-2.20076397e-01 -5.30215306e-03 1.01696038e+00 2.56963372e-01
-5.24784446e-01 1.26635477e-01 -2.71049619e-01 -2.90698200e-01
4.54952180e-01 -6.91820607e-02 -9.58607912e-01 -5.19153476e-01
-4.26888436e-01 -2.32801333e-01 6.15071416e-01 4.09226626e-01
6.01680815e-01 -1.02780962e+00 -6.78967595e-01 3.48264635e-01
-1.04147583e-01 -1.39634326e-01 5.99286556e-01 1.23164833e+00
-6.22690558e-01 2.15077922e-01 -1.29124492e-01 -3.99808288e-01
-1.03579664e+00 5.68503261e-01 8.47482443e-01 2.81975687e-01
-5.36676407e-01 1.29671597e+00 3.09563816e-01 3.95803720e-01
1.35478124e-01 -7.52806664e-01 -1.34581253e-01 -5.89247718e-02
8.07067275e-01 5.16122222e-01 3.24050665e-01 -5.41916072e-01
-1.85345948e-01 5.61705112e-01 -2.74284363e-01 2.28753313e-01
1.44871497e+00 -4.70680416e-01 -5.46991348e-01 -4.26587835e-03
1.15467715e+00 2.78044879e-01 -1.69081199e+00 -3.69684577e-01
-1.78295910e-01 -7.56980419e-01 5.24598002e-01 -9.28788781e-01
-1.24282241e+00 5.22482634e-01 8.26668143e-01 2.49432907e-01
1.50173807e+00 -1.68466240e-01 6.26890182e-01 1.23803400e-01
9.10018384e-02 -5.33169627e-01 -1.71563715e-01 4.24799770e-01
1.31653488e+00 -1.27154672e+00 1.05208337e-01 -3.22834522e-01
-2.85999537e-01 8.77943158e-01 1.73027098e-01 -5.14492333e-01
8.02885950e-01 2.48165220e-01 7.23046660e-02 -1.83690295e-01
-5.68403304e-01 1.23794429e-01 6.99814320e-01 4.88859266e-01
2.02559620e-01 -1.42003581e-01 8.60953480e-02 5.06506205e-01
-5.47985286e-02 1.21981882e-01 8.52212369e-01 5.43998361e-01
-3.70779932e-01 -4.02237713e-01 -5.79436660e-01 4.00611520e-01
-4.28653300e-01 -6.39662147e-01 -2.71305829e-01 1.70189321e-01
6.87117316e-03 9.63133216e-01 1.09173641e-01 -1.70232028e-01
1.97699755e-01 -2.94695467e-01 8.41294646e-01 -2.70520806e-01
-6.06883407e-01 7.50901997e-02 -1.78906307e-01 -6.40221834e-01
-4.99792397e-01 -6.00141704e-01 -2.99471885e-01 -2.81002134e-01
-4.56517369e-01 -2.78936118e-01 3.62348258e-01 7.51051903e-01
5.96294522e-01 4.48066652e-01 6.40387118e-01 -1.30083430e+00
-4.38694596e-01 -1.44998550e+00 -5.60591817e-01 2.80704767e-01
1.44613600e+00 -4.41043109e-01 -8.90706182e-01 5.23402333e-01] | [11.615772247314453, -2.748058795928955] |
087722e3-ba3c-46e7-93fa-dfee1a5a26d4 | simple-negation-scope-resolution-through-deep | null | null | https://aclanthology.org/P14-1007 | https://aclanthology.org/P14-1007.pdf | Simple Negation Scope Resolution through Deep Parsing: A Semantic Solution to a Semantic Problem | null | ['Stephan Oepen', 'Woodley Packard', 'Rebecca Dridan', 'Emily M. Bender', 'Jonathon Read'] | 2014-06-01 | null | null | null | acl-2014-6 | ['negation-scope-resolution'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.35286283493042, 3.7072086334228516] |
307e629c-7728-448f-ad55-f1bc9704b22e | segment-anything-in-high-quality | 2306.01567 | null | https://arxiv.org/abs/2306.01567v1 | https://arxiv.org/pdf/2306.01567v1.pdf | Segment Anything in High Quality | The recent Segment Anything Model (SAM) represents a big leap in scaling up segmentation models, allowing for powerful zero-shot capabilities and flexible prompting. Despite being trained with 1.1 billion masks, SAM's mask prediction quality falls short in many cases, particularly when dealing with objects that have intricate structures. We propose HQ-SAM, equipping SAM with the ability to accurately segment any object, while maintaining SAM's original promptable design, efficiency, and zero-shot generalizability. Our careful design reuses and preserves the pre-trained model weights of SAM, while only introducing minimal additional parameters and computation. We design a learnable High-Quality Output Token, which is injected into SAM's mask decoder and is responsible for predicting the high-quality mask. Instead of only applying it on mask-decoder features, we first fuse them with early and final ViT features for improved mask details. To train our introduced learnable parameters, we compose a dataset of 44K fine-grained masks from several sources. HQ-SAM is only trained on the introduced detaset of 44k masks, which takes only 4 hours on 8 GPUs. We show the efficacy of HQ-SAM in a suite of 9 diverse segmentation datasets across different downstream tasks, where 7 out of them are evaluated in a zero-shot transfer protocol. Our code and models will be released at https://github.com/SysCV/SAM-HQ. | ['Fisher Yu', 'Chi-Keung Tang', 'Yu-Wing Tai', 'Yifan Liu', 'Martin Danelljan', 'Mingqiao Ye', 'Lei Ke'] | 2023-06-02 | null | null | null | null | ['2d-semantic-segmentation'] | ['computer-vision'] | [ 4.26589727e-01 2.44607717e-01 -1.39817387e-01 -3.12306970e-01
-1.26235986e+00 -8.42701495e-01 2.45023176e-01 -8.72068778e-02
-4.86873478e-01 4.51057881e-01 3.24445777e-02 -3.50311577e-01
3.10527563e-01 -6.15411520e-01 -9.50530052e-01 -6.94668055e-01
1.47969544e-01 5.40905356e-01 8.36516500e-01 -3.61121795e-03
8.42022523e-02 2.72242546e-01 -1.43943048e+00 4.42436665e-01
8.48236501e-01 9.05074120e-01 5.05931318e-01 1.01938617e+00
1.43282965e-01 4.82139826e-01 -3.68548155e-01 -6.41326189e-01
4.87529874e-01 -2.89713621e-01 -7.49581575e-01 -1.20857961e-01
4.18344945e-01 -4.67941850e-01 -3.14366400e-01 5.81267059e-01
7.16904998e-01 -1.05206922e-01 5.34389853e-01 -9.27287102e-01
-3.46692652e-01 9.07066703e-01 -4.65642184e-01 3.07314545e-01
-6.71999082e-02 7.45595157e-01 1.04427350e+00 -9.02915657e-01
6.87771261e-01 7.13207543e-01 8.32328796e-01 8.15681815e-01
-1.34424627e+00 -5.57797730e-01 -9.44794044e-02 -1.89540982e-01
-1.30418789e+00 -7.68809080e-01 8.29487070e-02 -5.44752419e-01
1.17354560e+00 4.09474075e-01 6.67784572e-01 1.12077475e+00
1.78179234e-01 9.45498586e-01 7.65552342e-01 -1.11384250e-01
6.77266270e-02 -1.25561982e-01 4.05630589e-01 8.69254172e-01
-9.03535113e-02 -8.21064133e-03 -6.63267016e-01 3.54373232e-02
6.92715704e-01 -2.45649710e-01 -1.85211390e-01 6.87071979e-02
-1.46072149e+00 6.54286146e-01 2.97242016e-01 1.48711234e-01
-7.81440139e-02 3.92053664e-01 3.32042515e-01 1.16831064e-02
3.75221431e-01 3.62427741e-01 -5.60036719e-01 -4.42544907e-01
-1.24123025e+00 8.36532339e-02 7.14063227e-01 1.25295126e+00
9.95799243e-01 -3.09589684e-01 -8.56105924e-01 6.82256222e-01
-8.50057378e-02 3.25350612e-01 3.87335360e-01 -1.01056027e+00
3.30632299e-01 3.59995902e-01 -1.10237710e-01 -4.31890309e-01
-4.73430812e-01 -5.13790965e-01 -6.17751896e-01 -7.23038837e-02
3.20374936e-01 -2.33260512e-01 -1.41738927e+00 1.66758776e+00
4.12765801e-01 4.43725497e-01 -2.79983521e-01 7.85330713e-01
7.54522681e-01 8.23142350e-01 -6.82474226e-02 2.34812260e-01
1.42306900e+00 -1.40278184e+00 -1.80957809e-01 -4.93811816e-01
7.20783055e-01 -8.68855059e-01 1.16636026e+00 2.55877733e-01
-1.24433076e+00 -5.94049752e-01 -9.79778647e-01 -5.33981085e-01
-2.48289734e-01 2.06719741e-01 5.55700183e-01 5.35210788e-01
-1.28061640e+00 7.42530465e-01 -1.09212697e+00 -1.25790656e-01
8.94177973e-01 5.60112000e-01 -9.07894745e-02 7.52445981e-02
-7.79583275e-01 6.61853433e-01 2.96846062e-01 -6.68581426e-02
-1.03565729e+00 -1.07217765e+00 -8.73842359e-01 1.65305957e-01
3.87217909e-01 -8.11011195e-01 1.51254165e+00 -6.30105257e-01
-1.60195434e+00 9.87412810e-01 -3.51410687e-01 -4.18819129e-01
5.63516915e-01 -2.69138426e-01 8.62913057e-02 7.02938959e-02
3.65096301e-01 1.13413024e+00 9.57529783e-01 -8.78744543e-01
-6.51737869e-01 1.03730343e-01 2.16513518e-02 -5.34927882e-02
-1.05142333e-01 4.71560173e-02 -1.08332348e+00 -6.92872405e-01
-1.24634609e-01 -9.14357424e-01 -3.71511608e-01 1.20070666e-01
-6.79117680e-01 1.36587232e-01 4.25968796e-01 -5.46736360e-01
1.18831360e+00 -2.36122012e+00 1.38800740e-01 -6.04390763e-02
4.50251430e-01 5.19927442e-01 -3.08085322e-01 3.46532017e-01
2.50268996e-01 3.91507074e-02 -7.70866036e-01 -6.75987303e-01
5.56023791e-02 1.66202262e-01 -1.42612755e-01 2.64351040e-01
5.98926067e-01 1.41832972e+00 -7.22038090e-01 -6.22370064e-01
2.30382234e-01 3.19349080e-01 -7.76320100e-01 4.49653178e-01
-3.44411820e-01 4.11368370e-01 -2.61442333e-01 5.78931272e-01
6.00807190e-01 -5.10608375e-01 -2.33820036e-01 -1.56788066e-01
-1.65628344e-01 2.96332359e-01 -8.80614281e-01 2.08873582e+00
-2.75197327e-01 4.74964797e-01 6.72236830e-03 -6.79855883e-01
5.11989594e-01 1.41910851e-01 2.95743644e-01 -5.73032796e-01
2.61525422e-01 2.23374471e-01 -8.34252685e-03 -3.63525867e-01
4.20880020e-01 6.80886209e-02 -3.15194935e-01 5.18421292e-01
2.69007832e-01 -7.11062551e-02 2.28047177e-01 5.00928581e-01
1.46449697e+00 1.87771142e-01 3.33712548e-02 -2.60513455e-01
1.11271970e-01 1.26843125e-01 6.98877275e-01 8.04972231e-01
-1.50595382e-01 1.11171198e+00 5.59104204e-01 -8.41945931e-02
-1.07757187e+00 -1.04084635e+00 -2.21275717e-01 1.25461304e+00
2.82733172e-01 -5.35111129e-01 -1.01798522e+00 -5.26318371e-01
-3.66028249e-02 5.94099164e-01 -6.37844443e-01 -1.18362449e-01
-8.26508343e-01 -7.61259615e-01 9.22273874e-01 5.01107991e-01
3.16850096e-01 -1.00595725e+00 -8.54019344e-01 4.58540380e-01
-7.97569752e-02 -9.66429532e-01 -9.14699137e-01 3.89027864e-01
-4.88006592e-01 -8.22946310e-01 -6.49322689e-01 -8.13288629e-01
5.52304745e-01 3.23951364e-01 1.05938518e+00 2.72222102e-01
-6.90893173e-01 -1.50697887e-01 -2.85024792e-01 -2.36076668e-01
-3.84119660e-01 6.23018324e-01 -3.49703342e-01 -1.18804686e-02
-5.07483780e-02 -4.77374971e-01 -7.61859298e-01 3.33342254e-01
-8.32166970e-01 6.81936443e-01 6.23066068e-01 7.55913734e-01
7.32130229e-01 -6.06796384e-01 5.81619024e-01 -1.18232656e+00
8.79404545e-02 -4.70490903e-01 -5.09295642e-01 6.87209219e-02
-2.43719548e-01 -3.54894325e-02 4.59320575e-01 -3.94224644e-01
-9.46847618e-01 2.79094130e-01 -5.98395586e-01 -3.32723498e-01
-4.43240069e-02 -2.02997737e-02 -1.55396581e-01 2.86009554e-02
6.16030633e-01 1.16279393e-01 -8.25642422e-02 -6.20138168e-01
7.17449486e-01 5.96017838e-01 6.13784969e-01 -5.64212024e-01
7.31732786e-01 2.51274407e-01 -4.28220212e-01 -4.30374801e-01
-9.89753783e-01 -5.82257688e-01 -6.11420512e-01 6.45888075e-02
1.04035950e+00 -8.61731708e-01 -5.79318941e-01 7.55917728e-01
-1.08693588e+00 -9.92858350e-01 -5.35332620e-01 -7.37488195e-02
-7.07841873e-01 4.93695438e-02 -1.11537719e+00 -1.88045084e-01
-5.43593884e-01 -1.40545607e+00 1.44083130e+00 2.24585533e-01
-1.83679983e-01 -4.47462410e-01 -4.01759706e-03 3.55512112e-01
4.43095356e-01 -5.62948473e-02 7.86524653e-01 -4.35710818e-01
-8.76851201e-01 4.22384180e-02 -4.15676475e-01 2.37487823e-01
-8.97177961e-04 1.34570254e-02 -1.23926270e+00 -2.49396667e-01
-2.05716655e-01 -3.45467627e-01 1.19525886e+00 3.66944879e-01
1.36476171e+00 -1.64699078e-01 -6.30866587e-01 1.04566681e+00
1.37938154e+00 -1.02645442e-01 5.45847356e-01 -5.05466126e-02
8.09803426e-01 2.57637203e-01 6.25215888e-01 1.59861535e-01
3.35752696e-01 6.71470284e-01 2.56119609e-01 -2.92061388e-01
-5.05074322e-01 -2.30575249e-01 2.66406506e-01 9.70997393e-01
1.63901106e-01 -5.01307964e-01 -8.75339270e-01 5.62280238e-01
-1.85790670e+00 -5.85243106e-01 -8.41763616e-02 1.84832335e+00
1.16767478e+00 3.71180564e-01 5.96315674e-02 -1.79001525e-01
6.26156271e-01 2.40447327e-01 -7.62064397e-01 -4.31026101e-01
-1.60618741e-02 5.19198835e-01 6.68196201e-01 6.23042464e-01
-1.11923718e+00 1.51232421e+00 6.12779903e+00 1.23318493e+00
-1.04643548e+00 3.86999875e-01 8.45093489e-01 -3.48375469e-01
-2.66437322e-01 1.08303778e-01 -1.01176012e+00 7.03665316e-01
1.01215816e+00 1.88369490e-02 3.54116946e-01 5.29738843e-01
-1.96441654e-02 -1.03738301e-01 -1.16467679e+00 6.74533427e-01
-1.26591608e-01 -1.77466333e+00 -1.72971278e-01 -3.94970849e-02
5.99152207e-01 4.93280828e-01 -4.13693190e-02 4.15937155e-01
1.92236498e-01 -1.09278297e+00 1.02248442e+00 3.66362900e-01
1.12949240e+00 -3.63055289e-01 4.09171402e-01 4.89500880e-01
-1.06032169e+00 1.57922924e-01 -3.72204512e-01 -1.97073594e-02
4.98274744e-01 6.32436752e-01 -8.57776761e-01 3.64464819e-01
4.78540599e-01 7.36880124e-01 -5.20604610e-01 9.81899381e-01
-1.00018375e-01 9.33972776e-01 -3.69515210e-01 3.53999078e-01
2.82342106e-01 1.07955366e-01 3.73211652e-01 1.57790911e+00
2.42555201e-01 3.23646665e-01 1.29452288e-01 8.54201078e-01
-2.76023209e-01 -1.72446862e-01 -2.82658428e-01 1.90725520e-01
5.13523936e-01 1.43010926e+00 -1.05749547e+00 -6.57499790e-01
-3.72889876e-01 1.29256082e+00 5.21886468e-01 4.36095893e-02
-1.20741332e+00 -1.96122542e-01 8.59012306e-01 1.60599530e-01
6.32488072e-01 -1.03156626e-01 -4.74209934e-01 -1.03774095e+00
-6.23061582e-02 -8.81193340e-01 1.08646840e-01 -4.60853308e-01
-1.11118078e+00 5.52888930e-01 -2.85809875e-01 -9.37946796e-01
1.13468185e-01 -5.17873347e-01 -7.33103096e-01 7.29777515e-01
-1.36752784e+00 -1.38747668e+00 -1.22681111e-01 2.31306002e-01
8.82165968e-01 2.06205651e-01 7.70097673e-01 4.86442387e-01
-8.53787959e-01 9.19853926e-01 -2.47250512e-01 7.18894824e-02
7.05137849e-01 -1.18981504e+00 1.22396445e+00 7.96879709e-01
5.26734963e-02 3.11348200e-01 4.88703400e-01 -5.24798334e-01
-1.39881063e+00 -1.36578774e+00 6.86843455e-01 -6.40029252e-01
5.37067533e-01 -7.16116548e-01 -9.19717193e-01 8.81173790e-01
2.65038367e-02 2.11380616e-01 5.10708511e-01 -3.37644279e-01
-2.24544078e-01 5.72218597e-02 -9.12472427e-01 4.86989737e-01
1.39312863e+00 -1.88942239e-01 -3.45024943e-01 2.94719189e-01
1.28046608e+00 -7.96425343e-01 -7.95903444e-01 3.28059345e-01
4.98670787e-01 -9.89992678e-01 7.96276748e-01 -3.40243876e-01
4.65099305e-01 -2.27592543e-01 3.47363343e-03 -1.00423586e+00
-5.88896215e-01 -7.73261786e-01 2.16442849e-02 1.28771746e+00
7.88648427e-01 -5.93919337e-01 8.88309598e-01 4.81555462e-01
-6.38395965e-01 -1.17258108e+00 -9.79038239e-01 -5.14010668e-01
-4.77006882e-02 -5.89308381e-01 8.23269665e-01 6.05129957e-01
-3.58001411e-01 3.97681952e-01 -4.18799460e-01 1.15603387e-01
3.34384590e-01 2.54365116e-01 8.50380242e-01 -6.85032487e-01
-7.22115755e-01 -3.36123168e-01 -1.82822183e-01 -1.37232113e+00
-9.79466587e-02 -1.23404551e+00 4.22648162e-01 -1.41553378e+00
3.92236710e-01 -7.54506230e-01 -1.56078219e-01 8.65422726e-01
-3.40979159e-01 6.71850741e-01 3.54159176e-01 2.14870259e-01
-6.16185069e-01 2.80634612e-01 1.35872459e+00 -6.51845187e-02
-1.42730057e-01 1.17687061e-01 -7.02204585e-01 5.89266896e-01
9.14840460e-01 -6.74713373e-01 -1.77788198e-01 -8.98684084e-01
-5.73605299e-02 -5.59353046e-02 3.06953669e-01 -1.08957100e+00
1.96642622e-01 9.31671187e-02 1.49181709e-01 -4.14241612e-01
6.44290209e-01 -2.80911177e-01 1.78502604e-01 4.68430042e-01
-1.72746733e-01 -2.55701751e-01 2.64506161e-01 2.93020487e-01
1.73113793e-01 -1.22743636e-01 8.63977909e-01 -7.55158365e-02
-6.70611858e-01 6.40283465e-01 -2.33617768e-01 2.26045147e-01
1.11293125e+00 -2.53875017e-01 -3.50073159e-01 2.16437623e-01
-7.04735518e-01 2.85506546e-01 7.39833355e-01 2.88731039e-01
3.70927602e-01 -8.19632292e-01 -7.09923983e-01 4.48655933e-01
1.10693395e-01 3.01081866e-01 3.79186571e-01 7.80688941e-01
-7.60827601e-01 1.28054485e-01 3.59948836e-02 -7.27582455e-01
-9.38885331e-01 4.42450523e-01 1.71728089e-01 -2.65695453e-01
-1.11202848e+00 1.38266957e+00 4.50888991e-01 -3.83515060e-01
1.49755059e-02 -4.36493337e-01 4.21107024e-01 -7.23718405e-02
4.56541896e-01 3.39909017e-01 9.02621970e-02 -5.30099690e-01
-3.60465527e-01 6.21619463e-01 -9.06807333e-02 -1.38196290e-01
1.22277987e+00 8.96162391e-02 -7.29430765e-02 3.57745051e-01
1.06725502e+00 -8.88775587e-02 -1.57860339e+00 -4.98210974e-02
-1.54892206e-01 -3.06078136e-01 -1.99647948e-01 -7.98018336e-01
-1.05993760e+00 9.59175110e-01 4.27002698e-01 -3.03724647e-01
9.87387002e-01 2.68016249e-01 1.38179290e+00 -1.10842526e-01
5.22536695e-01 -7.99644411e-01 -1.44271404e-01 4.99631524e-01
4.07885194e-01 -1.20919120e+00 -4.10149366e-01 -7.23608553e-01
-6.27763748e-01 5.51127970e-01 6.47186637e-01 -1.03012666e-01
4.38622892e-01 6.04810715e-01 9.98749509e-02 -8.77576247e-02
-9.59453225e-01 -2.94028759e-01 1.12608604e-01 4.20363009e-01
2.42285013e-01 2.85793573e-01 -3.04891735e-01 8.68278325e-01
-3.48076999e-01 -7.32127801e-02 3.34768981e-01 9.16484594e-01
-5.92686951e-01 -1.16617918e+00 1.12811796e-01 8.29854727e-01
-5.20823359e-01 -3.16217393e-01 1.75918806e-02 2.71795958e-01
5.24258077e-01 7.11448669e-01 1.07630864e-01 -4.31619912e-01
1.79924130e-01 -3.91191803e-02 3.06509942e-01 -1.08905709e+00
-8.15779746e-01 -2.21839286e-02 -2.90406682e-03 -9.71384048e-01
8.70623216e-02 -4.77823675e-01 -1.40850747e+00 -2.66667485e-01
-2.16248631e-01 -1.52114570e-01 3.89745712e-01 8.74483049e-01
7.32820213e-01 6.28155589e-01 3.42344046e-01 -1.10636294e+00
-3.04732740e-01 -8.00446510e-01 -3.12811315e-01 1.45701274e-01
3.57258707e-01 -3.07107925e-01 -7.83746913e-02 1.36342004e-01] | [9.573427200317383, 0.0773303434252739] |
88216e49-63aa-4378-9f89-84cff12899ad | robust-multi-domain-mitosis-detection | 2109.15092 | null | https://arxiv.org/abs/2109.15092v1 | https://arxiv.org/pdf/2109.15092v1.pdf | Robust Multi-Domain Mitosis Detection | Domain variability is a common bottle neck in developing generalisable algorithms for various medical applications. Motivated by the observation that the domain variability of the medical images is to some extent compact, we propose to learn a target representative feature space through unpaired image to image translation (CycleGAN). We comprehensively evaluate the performanceand usefulness by utilising the transformation to mitosis detection with candidate proposal and classification. This work presents a simple yet effective multi-step mitotic figure detection algorithm developed as a baseline for the MIDOG challenge. On the preliminary test set, the algorithm scoresan F1 score of 0.52. | ['Sasidhar Kadiyala', 'Ritesh Gangnani', 'Mustaffa Hussain'] | 2021-09-13 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [ 6.91719055e-01 3.75235111e-01 -4.00744259e-01 -1.89479396e-01
-1.42296386e+00 -7.52944171e-01 1.02292955e+00 -1.00315422e-01
-5.89691281e-01 1.13942480e+00 -2.33868808e-01 -3.30522247e-02
-1.87180825e-02 -2.18980178e-01 -5.85274100e-01 -9.58831310e-01
3.25734437e-01 7.01363623e-01 2.55428493e-01 1.27555402e-02
1.49851620e-01 4.22591597e-01 -1.14608657e+00 4.58979994e-01
8.37718546e-01 6.69889748e-01 -1.88100219e-01 1.17914224e+00
4.15306985e-01 3.30478907e-01 -9.27266598e-01 -4.30016160e-01
1.00883678e-01 -1.05301678e+00 -8.87476206e-01 6.40334725e-01
3.64207476e-01 4.52970378e-02 1.16081769e-02 1.00233865e+00
5.33735693e-01 -2.89718747e-01 1.08071220e+00 -1.32699645e+00
-1.91315964e-01 2.81975567e-01 -4.16808814e-01 5.00357807e-01
1.69904560e-01 6.06481209e-02 5.92465401e-01 -8.98541927e-01
1.22474945e+00 7.93851018e-01 4.55232978e-01 6.02941930e-01
-1.46291363e+00 -4.75044638e-01 -4.57026541e-01 7.96411559e-02
-1.46862245e+00 -5.07461846e-01 3.70183408e-01 -4.36895519e-01
5.82226336e-01 3.29496354e-01 3.45304310e-01 1.02202034e+00
6.11102164e-01 6.50532246e-01 1.51192904e+00 -5.74466884e-01
2.77419239e-01 4.52183336e-01 -1.27973929e-01 7.43615150e-01
4.64154780e-01 1.61893919e-01 -6.15804553e-01 -1.23003893e-01
7.50582933e-01 -6.52339220e-01 -2.70786285e-01 -4.41290826e-01
-1.24513745e+00 8.03179383e-01 7.86469877e-02 3.94733846e-01
1.45071223e-01 -4.53690365e-02 4.58818674e-01 5.91340005e-01
3.32701206e-01 5.25618553e-01 -1.37562037e-01 -3.80346142e-02
-1.10671091e+00 4.91533168e-02 5.84549963e-01 8.22911322e-01
4.06898677e-01 -1.91032842e-01 -1.65907472e-01 3.66629660e-01
1.65073257e-02 4.91220027e-01 8.26248229e-01 -1.00352335e+00
-1.87046364e-01 4.74674761e-01 -1.90079361e-01 -5.21775901e-01
-5.72162688e-01 -5.51467896e-01 -9.21125710e-01 2.11321071e-01
8.88801038e-01 -2.26845816e-02 -1.25087798e+00 1.51284838e+00
4.25919890e-01 1.02002896e-01 2.46726871e-01 6.97364569e-01
7.85741448e-01 4.45452519e-02 6.45636348e-03 -2.08206773e-01
1.46998477e+00 -7.73311913e-01 -6.30625129e-01 6.10415302e-02
9.13026452e-01 -9.23068285e-01 6.45652533e-01 4.79857326e-01
-9.69439745e-01 -3.27174515e-01 -1.25517488e+00 -1.20936967e-02
-3.70686501e-01 3.27175707e-01 3.58797103e-01 1.06921625e+00
-1.28732812e+00 4.62300897e-01 -8.54515314e-01 -8.32521558e-01
5.83610237e-01 6.48085177e-01 -6.44081354e-01 -8.53759516e-03
-6.86192930e-01 9.72555578e-01 6.29766047e-01 -5.35217524e-01
-8.22646260e-01 -7.79936135e-01 -5.26990056e-01 -5.38916051e-01
6.36766180e-02 -7.95787871e-01 1.10819924e+00 -9.59912241e-01
-1.49073887e+00 1.56837642e+00 6.52519017e-02 -6.34420633e-01
9.34611261e-01 4.65741843e-01 -2.29672298e-01 5.06094158e-01
7.84809217e-02 1.00116110e+00 7.95057952e-01 -9.37331378e-01
-7.60607362e-01 -2.12932214e-01 -4.72818971e-01 1.59227490e-01
-9.56333205e-02 -1.09641574e-01 -4.83069867e-01 -8.99263024e-01
2.98719164e-02 -1.33327174e+00 -2.07102761e-01 1.44406274e-01
-3.25248897e-01 1.22538127e-01 5.93981385e-01 -7.55605102e-01
7.89979279e-01 -2.26471353e+00 1.97586179e-01 4.34079230e-01
1.96604788e-01 5.19522838e-02 -1.39630958e-01 -9.83850583e-02
-1.66620076e-01 1.03042260e-01 -4.95789587e-01 -9.67398882e-02
-2.34137744e-01 1.19627751e-01 2.37882316e-01 9.87877727e-01
4.44259703e-01 1.03782105e+00 -7.86217093e-01 -8.43152881e-01
4.22606841e-02 4.34137136e-01 -3.37913364e-01 8.18560049e-02
1.19791636e-02 6.24959111e-01 -1.88332096e-01 8.05865943e-01
6.46387637e-01 -4.52077895e-01 3.21888030e-01 1.42818958e-01
3.85533899e-01 -2.85479844e-01 -9.43241060e-01 1.70378935e+00
2.16709793e-01 7.53785074e-01 -1.52641177e-01 -6.49213612e-01
7.92206585e-01 2.82267809e-01 4.33482200e-01 -3.31699073e-01
2.82375872e-01 5.34781218e-01 2.77485341e-01 7.66458455e-03
1.76625580e-01 -1.79451346e-01 -2.98394263e-01 2.25650012e-01
5.32153964e-01 -4.86502349e-01 4.74091500e-01 4.02768329e-02
1.28583705e+00 7.83381984e-02 7.94547677e-01 -7.16083944e-01
9.24117446e-01 3.95264119e-01 3.40563178e-01 4.20702070e-01
-6.52470469e-01 9.65457618e-01 5.47017455e-01 -1.93350583e-01
-1.08225787e+00 -1.03687119e+00 -3.03894669e-01 6.86953545e-01
-1.06370427e-01 -2.11698599e-02 -8.99185300e-01 -9.32061076e-01
-2.80125052e-01 1.52034774e-01 -9.68349159e-01 -1.39531597e-01
-4.24537063e-01 -1.08279848e+00 8.80359113e-01 3.63860309e-01
6.41348779e-01 -6.65724218e-01 -7.46861935e-01 -8.52499604e-02
-7.19548985e-02 -1.23164570e+00 -5.35415053e-01 3.24080974e-01
-8.42761159e-01 -1.22056901e+00 -1.17704773e+00 -1.02064395e+00
7.62754738e-01 -2.02797711e-01 1.16426432e+00 3.19847167e-02
-4.83230531e-01 1.99886575e-01 -1.50548667e-01 -4.15975362e-01
-8.26269627e-01 4.23198253e-01 -2.79561222e-01 -2.45171338e-01
4.49597180e-01 -8.07309449e-02 -5.18996775e-01 2.83432662e-01
-1.05409718e+00 -1.57448694e-01 8.18610787e-01 1.22369659e+00
9.13226247e-01 -1.36815518e-01 3.14316362e-01 -1.15209210e+00
2.33191624e-01 -1.28634053e-03 -5.00958979e-01 2.44299129e-01
-5.17930806e-01 1.66552188e-03 3.61257553e-01 -3.67007971e-01
-8.78273606e-01 4.52208191e-01 1.45672053e-01 1.28331454e-02
-2.74435490e-01 2.05801725e-01 -1.76201269e-01 -4.70768720e-01
1.08082128e+00 2.36216724e-01 1.97754025e-01 1.88317373e-01
2.60069907e-01 5.20817280e-01 8.59563231e-01 -1.24284156e-01
8.85625720e-01 6.46586120e-01 3.74761432e-01 -9.18495476e-01
-4.10869896e-01 -4.88534451e-01 -5.44150531e-01 5.04191443e-02
7.51602590e-01 -1.05688477e+00 -1.89850539e-01 5.09127438e-01
-8.77827644e-01 -3.05936038e-01 -2.69822568e-01 4.46488172e-01
-8.58369231e-01 2.33268231e-01 -4.94046062e-01 -2.04926923e-01
-2.80086815e-01 -8.98882627e-01 1.09553480e+00 2.81441659e-01
-5.92611492e-01 -1.06597948e+00 2.43703932e-01 2.32427090e-01
1.70360491e-01 6.99146748e-01 7.67703950e-01 -8.23250175e-01
-3.87908995e-01 -9.61771756e-02 -1.33424833e-01 -7.48821497e-02
6.86111003e-02 -1.09321345e-02 -1.11001468e+00 -4.62579340e-01
2.33692490e-02 -3.71541917e-01 9.48521435e-01 5.26021719e-01
7.75555253e-01 3.54247510e-01 -5.78942060e-01 6.88138545e-01
1.44000173e+00 1.67613164e-01 7.10086107e-01 5.46375573e-01
5.82243428e-02 5.00555754e-01 6.58131182e-01 1.43095613e-01
8.93566385e-02 4.31773752e-01 -5.73910251e-02 -4.36668456e-01
-5.61286330e-01 1.24088041e-01 8.86787474e-02 3.42870355e-01
9.06420052e-02 -3.11709940e-01 -9.30848897e-01 5.75930297e-01
-1.32582355e+00 -5.28587699e-01 1.52537078e-01 1.90900064e+00
8.73136818e-01 1.78465009e-01 2.51542032e-01 3.27096224e-01
7.36335039e-01 -3.44290525e-01 -4.85256433e-01 -2.63063163e-01
-5.68446815e-01 2.37200752e-01 7.37136483e-01 2.71621168e-01
-1.27262986e+00 7.16993868e-01 7.46032095e+00 8.64782453e-01
-1.04012191e+00 2.26571169e-02 1.14157081e+00 2.46340305e-01
1.32451370e-01 -3.25501025e-01 -7.56981492e-01 1.85164750e-01
1.21676111e+00 -4.44873989e-01 2.56478712e-02 4.57062691e-01
-2.24759489e-01 -2.36607194e-01 -1.15322506e+00 8.64626765e-01
3.48488241e-01 -1.32381177e+00 -8.29049796e-02 2.07911670e-01
9.69648838e-01 -2.71605670e-01 5.78189671e-01 1.08798571e-01
9.88662913e-02 -1.14896178e+00 1.95088968e-01 1.47907048e-01
1.26117444e+00 -7.26589262e-01 1.08775175e+00 3.75327282e-03
-7.12040007e-01 2.57171452e-01 -1.79489017e-01 4.94830042e-01
-3.43865454e-01 2.51760811e-01 -1.50215423e+00 4.90573943e-01
2.74476081e-01 4.73585933e-01 -8.99167180e-01 1.19376159e+00
-5.26719093e-02 3.67137492e-01 -3.26771706e-01 2.27339223e-01
-1.22018084e-02 1.80585116e-01 3.35634828e-01 1.30062163e+00
5.19259691e-01 -1.78202361e-01 2.27462444e-02 3.34261566e-01
-2.44284302e-01 1.24843419e-01 -6.13961756e-01 -2.40442395e-01
1.97918817e-01 1.16731572e+00 -1.39725137e+00 -3.79684806e-01
-1.09348014e-01 1.05786431e+00 -2.08647192e-01 5.04506379e-02
-7.34139383e-01 -4.52411026e-01 2.22529590e-01 -6.70173243e-02
3.99448425e-01 1.77628383e-01 -5.53886294e-01 -9.47302639e-01
-2.74847150e-01 -1.15097439e+00 8.00593913e-01 -2.92233020e-01
-9.35587168e-01 6.66828513e-01 -6.21539801e-02 -1.44931328e+00
-4.90787506e-01 -7.58851707e-01 -2.27031648e-01 5.83164811e-01
-1.74871492e+00 -1.34275508e+00 -2.49904633e-01 4.23356116e-01
6.02940321e-01 -4.75295961e-01 1.10244429e+00 -1.50364950e-01
-1.41640678e-01 1.10433888e+00 2.08241045e-01 -3.16360444e-02
1.08463860e+00 -1.42808187e+00 1.58545360e-01 8.93126249e-01
-2.53427811e-02 3.14458579e-01 9.12070870e-01 -5.27955115e-01
-1.01918423e+00 -1.10086513e+00 8.97266388e-01 -6.48910403e-01
5.07120311e-01 -1.98336795e-01 -6.79545045e-01 5.28264165e-01
3.45526010e-01 1.08905308e-01 9.37588573e-01 -4.85130996e-01
-1.06951423e-01 2.62718201e-01 -1.63062727e+00 4.15174127e-01
6.21514142e-01 -4.35309172e-01 -4.44981873e-01 4.15966690e-01
1.66793287e-01 -6.59962893e-01 -1.11173916e+00 2.52985626e-01
3.99473399e-01 -8.15453470e-01 6.92988813e-01 -2.42187753e-01
4.02203023e-01 -3.36990356e-01 -2.83110384e-02 -1.26240623e+00
-1.60803065e-01 -8.94804060e-01 3.92518267e-02 1.02114081e+00
5.78002691e-01 -3.34628314e-01 1.50747061e+00 4.42431480e-01
1.13641858e-01 -3.20977271e-01 -1.24598765e+00 -8.96833301e-01
5.19171476e-01 2.78841287e-01 2.90906161e-01 6.72859311e-01
1.52136192e-01 2.05721468e-01 -3.71511169e-02 -1.43993929e-01
5.92541337e-01 -1.51416063e-01 9.14436102e-01 -1.03364730e+00
-3.87350053e-01 -3.68484586e-01 -9.45600450e-01 -3.99451047e-01
-8.04435685e-02 -9.98806179e-01 4.35972512e-02 -7.04278529e-01
5.83092093e-01 1.76243857e-01 -4.87492055e-01 1.04818190e-03
-7.68067092e-02 8.58147383e-01 5.45663200e-02 -3.54985520e-03
-6.97227895e-01 -1.39345914e-01 1.33587551e+00 -2.17014268e-01
-1.86142288e-02 -2.03069001e-01 -7.17364550e-01 4.94405776e-01
8.88773263e-01 -5.07740378e-01 -3.47828567e-01 1.09595329e-01
-1.43511653e-01 -1.25547960e-01 1.33198738e-01 -1.11798000e+00
1.61845028e-01 -2.23351270e-02 7.09476948e-01 -5.76489389e-01
3.37843150e-01 -5.90041339e-01 3.06513429e-01 8.85898888e-01
-3.07989955e-01 -4.70191054e-02 2.33851552e-01 4.91254061e-01
-2.36614108e-01 -7.37879947e-02 1.05335450e+00 -1.23996764e-01
-6.83788717e-01 6.95643248e-03 -4.20556933e-01 1.80069029e-01
1.31260037e+00 -5.61173379e-01 -5.69871187e-01 -1.49900049e-01
-6.89060926e-01 -3.34475227e-02 9.75865722e-01 -1.41213760e-01
4.46580857e-01 -1.09083712e+00 -9.56696808e-01 3.71714890e-01
3.52700293e-01 -3.64681274e-01 -9.26545914e-03 9.77691054e-01
-7.28230357e-01 6.24252737e-01 -5.45621812e-01 -9.66675580e-01
-1.69407833e+00 3.74894917e-01 4.21088606e-01 -5.12355447e-01
-2.50496387e-01 1.03957331e+00 1.99395061e-01 -2.07082197e-01
-2.45395020e-01 -3.37651849e-01 -2.29022950e-01 -9.40465182e-02
5.18339455e-01 3.94971937e-01 2.25728422e-01 -7.41960645e-01
-5.14916420e-01 5.19077599e-01 -2.07273439e-01 -1.39310867e-01
9.62947845e-01 1.12927158e-03 2.11911336e-01 1.90331504e-01
1.19694364e+00 -4.08184558e-01 -1.08055186e+00 2.74942461e-02
3.50385070e-01 -4.53936964e-01 -2.00076386e-01 -7.82325387e-01
-6.88102126e-01 5.82587481e-01 1.20828557e+00 -8.89924765e-02
1.21152043e+00 -5.11793792e-02 3.11317414e-01 2.47876495e-01
1.55141428e-01 -1.20578337e+00 3.48491594e-02 2.80784458e-01
6.10455871e-01 -1.50101173e+00 3.07049364e-01 -7.18095005e-01
-6.54795885e-01 1.11095667e+00 2.80454129e-01 -2.06273258e-01
3.99705201e-01 3.45073462e-01 3.50152761e-01 4.89892922e-02
-7.47484028e-01 -1.84527654e-02 2.84720629e-01 1.05613434e+00
5.44015944e-01 -5.14289886e-02 -6.23156548e-01 3.13123673e-01
-1.77387267e-01 2.65402287e-01 8.54703665e-01 9.14308190e-01
-2.18160540e-01 -1.19432628e+00 -5.27935565e-01 3.60702515e-01
-8.79635394e-01 5.14608145e-01 -6.46533072e-01 1.14137387e+00
3.87954228e-02 7.30543911e-01 -1.56105995e-01 -1.10249177e-01
9.57137346e-02 1.25358533e-03 6.06496871e-01 -4.73448277e-01
-5.39616168e-01 3.27913880e-01 1.38718691e-02 -1.73537895e-01
-6.72790229e-01 -8.38428080e-01 -1.03337824e+00 -4.25126515e-02
-3.00981760e-01 -5.83105795e-02 4.03190553e-01 8.18419158e-01
1.29750431e-01 4.99957293e-01 3.57201129e-01 -5.06503880e-01
-4.88057911e-01 -7.45324671e-01 -6.62453234e-01 4.45716739e-01
3.14636022e-01 -4.93075132e-01 -2.20654815e-01 6.59590602e-01] | [15.128247261047363, -3.0437657833099365] |
e0372d54-32e5-4f5d-a5a5-51f2cf023ae7 | evaluating-similitude-and-robustness-of-deep | 2306.16050 | null | https://arxiv.org/abs/2306.16050v2 | https://arxiv.org/pdf/2306.16050v2.pdf | Evaluating Similitude and Robustness of Deep Image Denoising Models via Adversarial Attack | Deep neural networks (DNNs) have shown superior performance comparing to traditional image denoising algorithms. However, DNNs are inevitably vulnerable while facing adversarial attacks. In this paper, we propose an adversarial attack method named denoising-PGD which can successfully attack all the current deep denoising models while keep the noise distribution almost unchanged. We surprisingly find that the current mainstream non-blind denoising models (DnCNN, FFDNet, ECNDNet, BRDNet), blind denoising models (DnCNN-B, Noise2Noise, RDDCNN-B, FAN), plug-and-play (DPIR, CurvPnP) and unfolding denoising models (DeamNet) almost share the same adversarial sample set on both grayscale and color images, respectively. Shared adversarial sample set indicates that all these models are similar in term of local behaviors at the neighborhood of all the test samples. Thus, we further propose an indicator to measure the local similarity of models, called robustness similitude. Non-blind denoising models are found to have high robustness similitude across each other, while hybrid-driven models are also found to have high robustness similitude with pure data-driven non-blind denoising models. According to our robustness assessment, data-driven non-blind denoising models are the most robust. We use adversarial training to complement the vulnerability to adversarial attacks. Moreover, the model-driven image denoising BM3D shows resistance on adversarial attacks. | ['WangMeng Zuo', 'Jiebao Sun', 'Zhichang Guo', 'Yao Li', 'Jie Ning'] | 2023-06-28 | null | null | null | null | ['adversarial-attack', 'image-denoising'] | ['adversarial', 'computer-vision'] | [-2.15548053e-01 -4.29937512e-01 7.33098030e-01 -1.43507332e-01
-6.05836928e-01 -9.39849555e-01 6.95625722e-01 -5.23427427e-01
-2.95258135e-01 5.62833846e-01 2.07085997e-01 -3.42136234e-01
-5.48318774e-02 -7.54597008e-01 -7.68239677e-01 -1.15162218e+00
-6.57453537e-02 -2.06899509e-01 2.39464030e-01 -6.64120495e-01
8.92275572e-02 6.46117985e-01 -1.14875805e+00 1.61446109e-02
7.47740090e-01 8.28936994e-01 -2.40205139e-01 8.84302258e-01
2.32998490e-01 8.21317554e-01 -8.45640779e-01 -9.79162931e-01
7.39911497e-01 -4.68373060e-01 -3.47676426e-01 -3.33589584e-01
5.41121483e-01 -6.58587635e-01 -9.62355494e-01 1.70151734e+00
1.11268401e+00 -2.15127859e-02 5.72389305e-01 -1.26643419e+00
-9.66076314e-01 4.84730899e-01 -4.45022881e-01 1.82357579e-01
5.23537472e-02 4.56099957e-01 3.25435281e-01 -6.01754308e-01
5.34437060e-01 1.64067721e+00 1.01145899e+00 9.40767705e-01
-1.04960001e+00 -8.85771394e-01 1.02352254e-01 -8.22622888e-03
-1.16088188e+00 -4.32519615e-01 6.67529643e-01 -4.20405120e-02
1.95717573e-01 4.09580588e-01 1.49490342e-01 1.72312307e+00
4.16772574e-01 4.35959488e-01 1.45789301e+00 -7.65392482e-02
4.62411314e-01 -3.21956366e-01 -2.16582939e-02 2.91044801e-01
3.27710271e-01 5.90794504e-01 -3.89153868e-01 -6.01504147e-02
7.45299280e-01 7.67905936e-02 -4.91117656e-01 1.24617666e-01
-7.33132839e-01 7.00224221e-01 5.84044933e-01 1.95893705e-01
-3.48576933e-01 1.68348148e-01 5.37061274e-01 8.50772440e-01
2.10742384e-01 -3.16894962e-03 -3.93593371e-01 1.41132966e-01
-5.56537569e-01 2.50773907e-01 7.23481655e-01 8.42372119e-01
5.57872534e-01 6.16080999e-01 -7.51712322e-02 6.41441405e-01
2.86992460e-01 7.23699152e-01 4.60668564e-01 -1.08869326e+00
1.99519143e-01 4.28021476e-02 -1.48258984e-01 -1.21309745e+00
-1.63150460e-01 -3.21531415e-01 -1.67830718e+00 8.27735305e-01
2.76112705e-01 -4.45518970e-01 -1.29205465e+00 1.79324436e+00
2.84705758e-02 3.23847473e-01 5.40327191e-01 9.23118651e-01
9.71741557e-01 4.64657366e-01 1.32122129e-01 3.96566764e-02
1.06276083e+00 -6.21514618e-01 -8.24260771e-01 -1.95275962e-01
-1.18191093e-02 -8.99364233e-01 7.50390053e-01 6.70523763e-01
-8.78823757e-01 -7.18625069e-01 -1.15766621e+00 3.05026382e-01
-4.16440755e-01 -5.70466757e-01 2.10935265e-01 1.19558752e+00
-1.17983997e+00 7.57767618e-01 -6.27449512e-01 -2.58358449e-01
6.04869246e-01 2.27140695e-01 -7.02999473e-01 -4.28736746e-01
-1.28972960e+00 6.79975450e-01 8.11711475e-02 2.98717916e-01
-1.74621618e+00 -5.49557149e-01 -5.61253190e-01 -4.05659497e-01
-2.62021750e-01 -6.79036856e-01 8.00568640e-01 -1.18513691e+00
-1.36100066e+00 7.09582686e-01 3.15240115e-01 -5.64155817e-01
9.42703426e-01 -2.00205222e-01 -9.78067458e-01 2.89397418e-01
-1.77512795e-01 4.61224705e-01 1.41384923e+00 -1.92577636e+00
-1.07376710e-01 -3.20703119e-01 9.29716453e-02 -2.45227680e-01
-5.43035686e-01 3.35430294e-01 -2.90581167e-01 -1.21725452e+00
2.72227257e-01 -5.06231427e-01 -3.11835170e-01 2.69500464e-01
-8.17001879e-01 5.27216733e-01 1.24115169e+00 -9.06033099e-01
8.29347908e-01 -2.52724814e+00 -4.28242907e-02 5.25039136e-01
2.20843002e-01 5.91647565e-01 -5.45443535e-01 4.10276622e-01
-3.91918182e-01 5.28080583e-01 -3.85099322e-01 -3.34016502e-01
4.57527079e-02 6.33464575e-01 -3.47786456e-01 6.83465362e-01
1.55099230e-02 6.22479796e-01 -6.34748816e-01 -2.36928418e-01
-5.05301775e-03 6.07868731e-01 -3.06152225e-01 3.84857267e-01
2.37874359e-01 4.93579984e-01 -2.47603297e-01 8.95739317e-01
1.29762256e+00 6.89592183e-01 -2.17938125e-01 -4.34166044e-01
3.78415257e-01 -5.47202647e-01 -1.36434340e+00 1.13752568e+00
-2.11036891e-01 4.75691527e-01 8.25560629e-01 -9.45526719e-01
1.09662700e+00 5.52050114e-01 -1.87421620e-01 -7.44313717e-01
2.36112237e-01 2.81282663e-01 1.24589674e-01 -5.14935613e-01
3.42471041e-02 -2.69132227e-01 1.73650563e-01 2.20840245e-01
2.04578832e-01 2.28219219e-02 -1.66085467e-01 2.77536601e-01
1.51174450e+00 -3.00472379e-01 -2.44412675e-01 -3.02065551e-01
4.59580332e-01 -4.92291003e-01 6.28866196e-01 1.30954397e+00
-6.30022764e-01 1.10922527e+00 4.91242677e-01 -1.56190217e-01
-9.49544549e-01 -1.34348702e+00 1.58401243e-02 7.49166548e-01
3.31641674e-01 1.28362477e-01 -9.96710539e-01 -7.32737482e-01
1.76121294e-02 4.04073358e-01 -6.08588874e-01 -5.24124980e-01
-3.99912477e-01 -7.14028835e-01 1.25312233e+00 4.22075719e-01
1.05277848e+00 -9.28668261e-01 3.20929140e-01 -7.37285167e-02
1.59202531e-01 -9.64915633e-01 -3.29685032e-01 3.04399222e-01
-8.53631258e-01 -9.35191691e-01 -8.72461021e-01 -9.05868471e-01
7.00564444e-01 1.47060633e-01 1.13268793e+00 1.97983012e-01
-4.72138226e-02 4.30938631e-01 -5.27047276e-01 -3.19361359e-01
-9.91242349e-01 -6.20637357e-01 4.43805397e-01 1.30303606e-01
-1.45969996e-02 -9.03104365e-01 -7.41944373e-01 5.81934214e-01
-1.41731167e+00 -9.00772333e-01 4.45706517e-01 7.08509028e-01
4.83146757e-01 6.01242840e-01 5.14340758e-01 -4.92274165e-01
8.18013549e-01 -4.51071441e-01 -2.56969124e-01 3.95859480e-02
-2.80487835e-01 -2.94360489e-01 9.71301556e-01 -7.88542032e-01
-9.04675841e-01 -3.07063222e-01 -5.76964915e-01 -8.73727977e-01
-4.39674139e-01 3.02192777e-01 -7.81961381e-01 -5.27211189e-01
8.98404241e-01 2.05905870e-01 2.18857422e-01 -7.49136508e-01
2.95555860e-01 4.33137268e-01 9.59693611e-01 -3.50992173e-01
1.53214180e+00 5.92185438e-01 -4.45944741e-02 -6.75944924e-01
-2.07040727e-01 5.96505031e-02 -1.17507502e-01 -1.58229336e-01
8.29898059e-01 -9.74979877e-01 -6.33397341e-01 1.34366596e+00
-1.08336771e+00 -1.40460342e-01 -9.95501652e-02 1.65673658e-01
-7.04227015e-02 6.46148860e-01 -7.49449432e-01 -7.40739882e-01
-4.62039292e-01 -1.17340350e+00 5.10860145e-01 1.60344049e-01
2.67191857e-01 -1.04087019e+00 -2.52407789e-01 9.75054651e-02
5.76727211e-01 5.54433227e-01 8.17626774e-01 -9.98597205e-01
-1.93344787e-01 -4.31104332e-01 -1.12082057e-01 1.36820138e+00
2.00994328e-01 1.63574070e-01 -1.05561543e+00 -4.64651853e-01
5.05217493e-01 -1.81573585e-01 9.18610752e-01 4.70138490e-01
9.50087130e-01 -4.57733870e-01 2.61292577e-01 8.82332981e-01
1.62612879e+00 9.45287123e-02 1.21481645e+00 4.97002542e-01
6.73415959e-01 2.35065877e-01 4.07773294e-02 3.16519961e-02
-3.81115139e-01 3.88250239e-02 1.19451964e+00 -4.25853014e-01
-2.68698514e-01 -9.32056550e-03 7.90013254e-01 6.69086456e-01
-3.86632979e-02 -5.49945235e-01 -7.13036358e-01 3.18098485e-01
-1.41642964e+00 -9.59443390e-01 -3.64010483e-01 1.86087990e+00
6.76113307e-01 4.07257617e-01 -1.37550846e-01 3.96057039e-01
8.41715872e-01 1.53825819e-01 -5.27828634e-01 -4.87628847e-01
-8.12075913e-01 3.65274906e-01 7.54954696e-01 3.70015979e-01
-1.36745071e+00 7.46933877e-01 6.46173000e+00 9.87025023e-01
-7.24080443e-01 1.34198442e-01 6.68219805e-01 3.47066373e-01
-1.47530794e-01 -1.76165342e-01 -2.35514507e-01 5.44596076e-01
6.11979842e-01 2.58428574e-01 6.18835211e-01 6.46875858e-01
2.64641583e-01 3.61120820e-01 -7.51251876e-01 7.68480957e-01
-4.11870815e-02 -1.05854523e+00 4.07278985e-01 -1.98606789e-01
9.57469106e-01 -1.14676483e-01 1.43553272e-01 3.06583196e-02
8.89235675e-01 -1.10343528e+00 6.79955244e-01 7.39155293e-01
4.07765001e-01 -9.37765777e-01 1.27857566e+00 3.36982012e-02
-6.36660576e-01 -1.89208016e-01 -4.57180262e-01 1.92684919e-01
-6.00426272e-02 6.11739159e-01 2.79783666e-01 7.60537624e-01
1.27590287e+00 5.55497706e-01 -6.45777822e-01 7.96087146e-01
-3.19039375e-01 9.26004052e-01 -2.79733129e-02 5.20273685e-01
2.42518276e-01 -2.71241397e-01 8.64375591e-01 9.73781407e-01
2.98598439e-01 -1.63562894e-01 -3.07088941e-01 6.64847732e-01
-2.95012325e-01 -4.41863507e-01 -5.44654965e-01 2.78816640e-01
3.82522911e-01 9.46457267e-01 -4.88919109e-01 1.53563335e-03
-2.63354212e-01 1.12846196e+00 -4.77856070e-01 8.28519464e-01
-6.59413040e-01 -4.40170854e-01 1.24985492e+00 -3.78955394e-01
3.64430517e-01 -9.74015296e-02 -2.91942626e-01 -7.33768106e-01
-6.80744350e-02 -1.69964206e+00 8.70631561e-02 -8.49874973e-01
-2.12034178e+00 6.80243015e-01 -5.51384568e-01 -1.36037183e+00
5.19342542e-01 -7.90712476e-01 -1.03094590e+00 9.27634120e-01
-1.26158524e+00 -1.04801810e+00 -4.00468260e-01 1.23050225e+00
6.58778548e-02 -6.80107415e-01 7.41864204e-01 4.38681394e-01
-7.32930958e-01 8.74007523e-01 5.90726495e-01 7.06701219e-01
8.36200416e-01 -1.08011043e+00 6.12073779e-01 1.49838257e+00
-2.67719507e-01 7.10397065e-01 1.03054738e+00 -5.87218046e-01
-1.58625352e+00 -1.32086146e+00 3.20095532e-02 -2.12881565e-01
7.09095955e-01 -2.13280201e-01 -1.09597266e+00 3.25746268e-01
2.81218290e-01 2.20129624e-01 2.41521135e-01 -5.58597028e-01
-7.52962172e-01 -4.52317059e-01 -1.67851543e+00 8.36992741e-01
9.76159334e-01 -6.89290762e-01 -3.68520737e-01 2.41978779e-01
7.97910810e-01 -1.63766686e-02 -8.39652121e-01 3.35651606e-01
1.98149741e-01 -1.28779221e+00 1.37978923e+00 -5.83502412e-01
5.00889897e-01 -3.09057266e-01 -5.72197735e-01 -1.27101910e+00
-1.14987567e-01 -9.83533382e-01 8.87554511e-02 1.78451693e+00
-1.04443759e-01 -9.23861623e-01 4.59328324e-01 1.61791101e-01
-1.21042937e-01 9.51553211e-02 -1.24357986e+00 -1.22002161e+00
4.59591657e-01 -4.95677531e-01 6.13894165e-01 9.24879014e-01
-1.02989602e+00 -4.42501813e-01 -4.94933575e-01 7.84684896e-01
9.14503753e-01 -8.01467478e-01 7.07588553e-01 -9.31107044e-01
-1.10524356e-01 -3.74920994e-01 -6.97694004e-01 -5.17889857e-01
2.10133363e-02 -5.58161139e-01 2.76123695e-02 -1.41708541e+00
-2.27484003e-01 -1.46162465e-01 -5.37222624e-01 4.90371853e-01
-1.92526221e-01 7.42051542e-01 1.69544518e-02 1.57486796e-01
-2.38599516e-02 2.90332079e-01 1.25577343e+00 -3.46245408e-01
2.00572640e-01 1.37053683e-01 -7.61357903e-01 7.85654962e-01
7.33860672e-01 -6.83864355e-01 -1.13095388e-01 -5.40077567e-01
-1.76356994e-02 -3.35890204e-01 9.10898268e-01 -1.22827423e+00
1.95441902e-01 2.11455896e-01 3.15242201e-01 -1.41329793e-02
5.86595051e-02 -1.23936307e+00 2.50708073e-01 6.56918645e-01
-1.08728884e-02 5.93581684e-02 3.13082814e-01 6.96157992e-01
-4.43115681e-01 -3.62256676e-01 9.18260276e-01 -7.44139925e-02
-6.39409006e-01 2.33219907e-01 -5.87107897e-01 1.28061781e-02
6.14968419e-01 -4.57657278e-01 -5.83227575e-01 -6.66775167e-01
-8.19637120e-01 -1.71588317e-01 3.36141378e-01 3.14283371e-01
6.92388833e-01 -1.20829916e+00 -1.01321316e+00 3.36561263e-01
-3.15403700e-01 -2.34886035e-01 4.52342868e-01 4.19089019e-01
-7.63102889e-01 -6.51356637e-01 -4.09459144e-01 -2.33783379e-01
-1.28469539e+00 6.65951908e-01 6.55694425e-01 1.03426673e-01
-4.14954662e-01 1.12740219e+00 1.70348715e-02 -5.89634895e-01
5.15405655e-01 1.76932216e-01 1.95589513e-01 -1.15712276e-02
3.45327049e-01 6.80350780e-01 2.57762522e-01 -4.99231488e-01
-3.70643735e-01 4.66872603e-01 2.47670664e-03 1.26044527e-01
1.25499451e+00 1.95623431e-02 -4.68561381e-01 -3.65837127e-01
1.21497416e+00 -8.17089081e-02 -1.46517181e+00 4.90769930e-02
-4.13668782e-01 -2.92416155e-01 2.38721922e-01 -9.48368549e-01
-1.61876643e+00 7.63017356e-01 1.28273654e+00 5.01978159e-01
1.62873232e+00 -4.09100920e-01 5.78053534e-01 1.61936820e-01
4.35408764e-02 -8.11455786e-01 1.98886588e-01 4.03410941e-01
1.08556128e+00 -1.06995022e+00 -1.50706396e-01 -7.68454596e-02
-4.61459756e-01 9.64814961e-01 4.69409555e-01 -3.83334309e-01
9.36990678e-01 6.20411456e-01 7.23079622e-01 1.55285835e-01
-1.87446773e-01 7.57313743e-02 -2.29891136e-01 1.26858628e+00
-4.09675270e-01 -4.21755701e-01 1.42661497e-01 6.17245913e-01
-1.39461756e-01 -4.94586945e-01 4.25635546e-01 8.45174730e-01
-1.16122469e-01 -9.97074306e-01 -9.09003139e-01 1.06423050e-01
-5.47574759e-01 -2.69978106e-01 -4.85383928e-01 7.86921501e-01
1.90404370e-01 1.38971412e+00 -2.06178829e-01 -7.36165941e-01
4.66622710e-01 -2.70014852e-01 -9.34184492e-02 2.45878309e-01
-1.16636848e+00 -8.20019841e-02 -1.89275458e-01 -5.18932104e-01
-3.27972561e-01 -2.15898693e-01 -7.60100901e-01 -8.67678463e-01
-5.75438403e-02 -1.82592943e-01 6.45241320e-01 7.40259349e-01
1.59081027e-01 6.36869192e-01 9.33045983e-01 -8.10401976e-01
-8.83263111e-01 -8.10104549e-01 -7.75175989e-01 6.20004535e-01
5.99157333e-01 -1.70162156e-01 -9.94927466e-01 1.38738453e-01] | [5.4359941482543945, 8.03760051727295] |
a258e9e1-8062-4d78-af3e-7734533b56af | hcgmnet-a-hierarchical-change-guiding-map | 2302.10420 | null | https://arxiv.org/abs/2302.10420v2 | https://arxiv.org/pdf/2302.10420v2.pdf | HCGMNET: A Hierarchical Change Guiding Map Network For Change Detection | Very-high-resolution (VHR) remote sensing (RS) image change detection (CD) has been a challenging task for its very rich spatial information and sample imbalance problem. In this paper, we have proposed a hierarchical change guiding map network (HCGMNet) for change detection. The model uses hierarchical convolution operations to extract multiscale features, continuously merges multi-scale features layer by layer to improve the expression of global and local information, and guides the model to gradually refine edge features and comprehensive performance by a change guide module (CGM), which is a self-attention with changing guide map. Extensive experiments on two CD datasets show that the proposed HCGMNet architecture achieves better CD performance than existing state-of-the-art (SOTA) CD methods. | ['Bo Du', 'Chen Wu', 'Chengxi Han'] | 2023-02-21 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 3.40048432e-01 -6.02748990e-01 1.64912835e-01 -5.20162106e-01
-4.68519866e-01 8.13981295e-02 4.82262492e-01 9.93274339e-03
-3.28543931e-01 4.26396996e-01 2.55673856e-01 -2.54783213e-01
-2.02868298e-01 -1.19072676e+00 -4.04617548e-01 -7.22001672e-01
-4.30802941e-01 -7.17753246e-02 5.81122756e-01 -5.42745829e-01
1.44315451e-01 8.46866071e-01 -1.60403872e+00 3.07024032e-01
1.14485502e+00 1.00879991e+00 3.61962587e-01 6.72781765e-01
2.72941105e-02 7.39627838e-01 7.55738765e-02 4.36531693e-01
2.91753829e-01 -2.34802246e-01 -6.76726401e-01 -3.54940817e-02
5.48634171e-01 -4.91788179e-01 -3.45590889e-01 1.29347968e+00
9.55179751e-01 1.26840159e-01 3.34069490e-01 -8.67759228e-01
-9.18161213e-01 3.97342443e-01 -1.05111110e+00 1.11906815e+00
-1.01019286e-01 1.80185929e-01 7.13297248e-01 -1.10201502e+00
6.10136867e-01 1.28872049e+00 9.89888430e-01 -5.05055077e-02
-9.49877858e-01 -8.09469461e-01 6.34530127e-01 6.75347447e-01
-1.61675310e+00 -1.72611937e-01 1.01471579e+00 -4.24398780e-01
1.07272315e+00 3.24145168e-01 8.09919357e-01 1.49687946e-01
1.60659105e-01 7.31387734e-01 1.02910793e+00 -9.20318514e-02
2.02653140e-01 -5.54092407e-01 2.55812913e-01 5.49327314e-01
6.51454851e-02 -6.42434508e-02 -1.55720755e-01 3.46949726e-01
9.30889070e-01 3.11158776e-01 -5.32033801e-01 8.81691054e-02
-9.66567516e-01 6.90343022e-01 1.33364236e+00 5.95898509e-01
-7.71591485e-01 2.36776575e-01 1.57169908e-01 4.15889591e-01
8.27788174e-01 9.83709618e-02 -4.32741612e-01 3.98906440e-01
-9.37749267e-01 -2.64294399e-03 -4.76208627e-02 4.23529029e-01
1.10115457e+00 2.34188768e-03 -3.27276647e-01 8.86507332e-01
1.62620321e-01 5.74395955e-01 4.05085266e-01 -5.15046895e-01
2.68277854e-01 8.48310649e-01 1.11983139e-02 -1.63719356e+00
-8.02107513e-01 -7.65654862e-01 -1.56279373e+00 2.05384672e-01
-4.63992059e-01 6.14269488e-02 -1.08656812e+00 1.26509023e+00
5.24785638e-01 2.56985575e-01 -4.33917671e-01 1.11771560e+00
1.10372901e+00 9.38410044e-01 -3.76461186e-02 -1.25345036e-01
1.01101434e+00 -8.38308215e-01 -6.68367922e-01 -2.57500410e-01
1.12514727e-01 -9.56408307e-02 1.09309316e+00 -3.53494883e-02
-6.82237744e-01 -8.04539084e-01 -1.04219711e+00 -1.45081669e-01
-5.05986094e-01 1.09663598e-01 5.76551735e-01 2.41587892e-01
-1.35355771e+00 5.39174497e-01 -9.11661446e-01 -3.84326488e-01
8.91770959e-01 1.29758954e-01 -1.06948040e-01 -4.72404629e-01
-1.19765520e+00 4.86475736e-01 1.70959502e-01 7.26817846e-01
-6.95532203e-01 -8.27346206e-01 -8.17754805e-01 1.23020694e-01
-1.34302117e-02 -4.58983898e-01 6.38371468e-01 -8.43247592e-01
-1.25115383e+00 7.44489849e-01 -7.50731975e-02 -2.40685523e-01
4.34661627e-01 4.58072536e-02 -6.45237863e-01 2.83723652e-01
-2.26446340e-04 7.85921633e-01 7.66802609e-01 -1.27214956e+00
-9.54515100e-01 -5.66692770e-01 -2.31343612e-01 4.04427558e-01
-1.29603222e-01 1.50734767e-01 -4.59851325e-01 -7.08765030e-01
7.96159029e-01 -3.26047540e-01 -5.53460777e-01 1.85753956e-01
-4.72787134e-02 1.79514859e-03 9.78956997e-01 -9.66321707e-01
1.50198340e+00 -2.10083699e+00 -2.00736299e-01 2.22942114e-01
3.28288645e-01 2.27545485e-01 -3.98847073e-01 6.24020956e-02
-2.33656973e-01 1.42816529e-01 -5.36375821e-01 2.39008561e-01
-2.17810377e-01 3.25235426e-02 4.47301008e-02 8.10513854e-01
3.28425646e-01 1.02516162e+00 -8.25095236e-01 -2.44338512e-01
3.81033152e-01 5.04292309e-01 -4.65281606e-01 -7.47276396e-02
2.89917383e-02 4.18692768e-01 -3.62208694e-01 8.68915200e-01
1.21435869e+00 -3.20279270e-01 -3.27132523e-01 -4.44782585e-01
-6.08807206e-01 -2.16416642e-01 -1.50431097e+00 1.25878322e+00
-1.93101186e-02 6.89675927e-01 7.13917464e-02 -8.42290342e-01
1.04948616e+00 -3.60191107e-01 5.83749771e-01 -1.26057184e+00
-6.23783879e-02 -1.07364796e-01 -3.02582622e-01 -6.98981762e-01
5.71286142e-01 5.84260225e-02 2.29921758e-01 7.14549571e-02
-6.24925792e-01 1.15645237e-01 -2.51323134e-01 -2.16737583e-01
1.05723846e+00 -1.40145108e-01 2.45693162e-01 -4.40276384e-01
5.77467442e-01 1.23881713e-01 9.07668114e-01 7.67036319e-01
-4.74767357e-01 5.95040858e-01 -1.58431083e-01 -8.42051446e-01
-6.99302316e-01 -8.08106601e-01 -2.98071116e-01 9.25747812e-01
5.67543805e-01 2.39256755e-01 -2.36994579e-01 -2.34090298e-01
4.61267643e-02 1.83467194e-01 -7.57514834e-01 1.90702938e-02
-7.02721000e-01 -1.44534230e+00 1.76586151e-01 5.65706253e-01
1.37085664e+00 -9.78643894e-01 -5.13865590e-01 5.40753484e-01
-2.72134244e-01 -6.69081748e-01 -2.57309526e-01 -1.48306236e-01
-7.23478198e-01 -7.25470901e-01 -3.73733282e-01 -9.66879666e-01
7.22999275e-01 8.00138474e-01 1.01275146e+00 -9.97325704e-02
-6.84863865e-01 -1.51844144e-01 -4.14052367e-01 -1.80866897e-01
2.09258482e-01 -1.53388083e-01 -3.48865449e-01 5.29831946e-02
2.29966506e-01 -7.04283535e-01 -1.12541127e+00 2.01608956e-01
-1.08686054e+00 1.53491199e-01 6.75519466e-01 7.33443260e-01
8.17195058e-01 6.56046450e-01 6.17574871e-01 -6.70311153e-01
2.50709683e-01 -3.76706481e-01 -5.85127354e-01 1.38548255e-01
-8.08653474e-01 -4.75357920e-01 1.96231365e-01 3.21979150e-02
-1.28315222e+00 -8.77962857e-02 -4.05360103e-01 -4.90349047e-02
-1.84929177e-01 7.66892254e-01 -2.12095812e-01 -4.73561406e-01
5.56754589e-01 3.41479957e-01 -4.43615407e-01 -6.69755936e-01
3.50418717e-01 9.42822158e-01 8.06627035e-01 6.30151406e-02
8.06209981e-01 8.14842522e-01 -2.67025471e-01 -7.45931923e-01
-6.63433194e-01 -7.68442869e-01 -7.71223605e-01 -3.77173513e-01
7.66531765e-01 -1.48840714e+00 -3.54385674e-01 1.19654357e+00
-6.71059251e-01 -5.76229095e-01 -2.01892510e-01 2.74249762e-01
-3.72444354e-02 1.11384116e-01 -7.77188003e-01 -5.91711938e-01
-7.42451489e-01 -5.55525064e-01 1.06335258e+00 5.30248344e-01
5.52458107e-01 -7.46470749e-01 1.55898765e-01 -6.43180981e-02
1.04507113e+00 6.62660718e-01 7.47904301e-01 3.22430044e-01
-7.93502569e-01 1.08641744e-01 -8.22280765e-01 1.42585650e-01
3.77219528e-01 -5.49399219e-02 -8.10215116e-01 -5.25449514e-01
-2.87863221e-02 1.68217674e-01 1.39511311e+00 8.24010909e-01
1.35007620e+00 -3.63369137e-01 -3.43356162e-01 9.74975884e-01
2.13737035e+00 1.37404069e-01 9.23288465e-01 6.29381955e-01
9.96247113e-01 1.13238640e-01 5.91632962e-01 6.48993850e-01
8.68368387e-01 3.49306762e-01 5.31862259e-01 -9.30185854e-01
-2.86241472e-01 3.06614399e-01 1.57451406e-01 7.25623965e-01
5.95276505e-02 1.10625327e-01 -9.72277224e-01 9.46284592e-01
-1.78082955e+00 -1.19501710e+00 -5.07624149e-01 1.52929962e+00
8.91070604e-01 -1.89029023e-01 -2.13826969e-01 1.16088696e-01
1.07014751e+00 5.31098664e-01 -9.81803000e-01 2.27944970e-01
-4.85801309e-01 1.15819372e-01 6.65964127e-01 4.27787155e-01
-1.44274724e+00 8.10086966e-01 6.02124834e+00 7.68786669e-01
-1.32327902e+00 2.29555398e-01 6.89038038e-01 2.50084370e-01
-2.23113582e-01 -4.20409203e-01 -6.28572643e-01 3.27793956e-01
5.67581020e-02 1.10432491e-01 3.86325896e-01 4.82246250e-01
5.40338993e-01 -1.70385018e-01 -3.83533537e-01 9.85544741e-01
7.78762996e-02 -1.30913532e+00 1.67034611e-01 -4.94647652e-01
1.12223232e+00 7.14148641e-01 -7.79239833e-02 2.23727435e-01
4.12799448e-01 -8.36367428e-01 2.89467782e-01 6.50978506e-01
8.82935643e-01 -7.32038081e-01 7.97478795e-01 -3.70873371e-03
-1.74366510e+00 -3.57757449e-01 -6.28006279e-01 -1.89819247e-01
-5.04957028e-02 1.00434291e+00 -2.72482723e-01 7.86506534e-01
1.13458478e+00 1.24725401e+00 -8.64372849e-01 1.28310728e+00
1.47841543e-01 4.84775096e-01 -3.59406203e-01 3.42063278e-01
3.72305870e-01 -2.10976899e-01 5.57153463e-01 1.40655887e+00
2.31198370e-01 4.05554891e-01 3.06558043e-01 6.95124686e-01
-1.71371049e-03 1.55446986e-02 -1.19760647e-01 5.32329440e-01
4.07982439e-01 1.42540503e+00 -7.82811821e-01 -2.85665780e-01
-1.96966588e-01 1.00439334e+00 1.95238173e-01 3.58836770e-01
-5.63650370e-01 -5.68096340e-01 7.46691048e-01 -7.52300844e-02
6.56524658e-01 -4.89342287e-02 -4.26518172e-01 -1.09704280e+00
3.39163281e-02 -5.60546219e-01 6.54693425e-01 -9.91637290e-01
-1.30879700e+00 5.29540360e-01 -5.87243676e-01 -1.30983341e+00
5.37190378e-01 -4.15774472e-02 -8.26546311e-01 9.08470273e-01
-2.50937080e+00 -1.11311877e+00 -1.13801157e+00 7.43771434e-01
3.88783544e-01 3.74925911e-01 2.88188666e-01 6.37371480e-01
-8.94744217e-01 2.17862695e-01 3.81394774e-01 1.83506057e-01
3.13648731e-01 -1.15931034e+00 6.63430274e-01 1.35612750e+00
-5.14237404e-01 -1.01944953e-02 3.37516218e-01 -7.62144923e-01
-1.08050966e+00 -2.02157497e+00 4.82783735e-01 3.06452096e-01
4.13586378e-01 2.12287515e-01 -1.12043977e+00 4.37356383e-01
-1.89998969e-01 2.99731910e-01 2.60650486e-01 -2.10651398e-01
-1.88295439e-01 -6.17315054e-01 -1.33434653e+00 3.14059794e-01
1.30484891e+00 -4.51392561e-01 -3.48088235e-01 1.26273975e-01
8.18785548e-01 -3.73578846e-01 -9.47926104e-01 7.97431409e-01
8.51330310e-02 -9.60443258e-01 9.73517776e-01 -1.14123911e-01
1.85230106e-01 -9.80104685e-01 -2.21125156e-01 -1.48062575e+00
-1.24624658e+00 -8.11745152e-02 3.07016343e-01 1.21571028e+00
-9.78439152e-02 -6.03009403e-01 2.58784682e-01 8.68666843e-02
-3.97460192e-01 -5.91384530e-01 -8.13387632e-01 -4.55481261e-01
-1.98340818e-01 -2.79103279e-01 9.50135529e-01 1.47269285e+00
-5.24175286e-01 1.66560858e-01 -6.93870857e-02 8.53535831e-01
3.79857600e-01 4.88660485e-01 2.41969079e-01 -1.40368927e+00
2.54652619e-01 -6.49391294e-01 -6.06166959e-01 -9.59425330e-01
-4.35745537e-01 -8.67404580e-01 1.90029845e-01 -2.21842003e+00
3.41492683e-01 -6.27106011e-01 -5.03524184e-01 5.58283448e-01
-6.97928727e-01 4.47278291e-01 -3.05561453e-01 1.82928413e-01
-6.52048409e-01 7.96444595e-01 1.32416201e+00 -3.39007556e-01
-4.92597401e-01 -3.89252752e-01 -6.36680961e-01 4.39467520e-01
6.97505832e-01 -3.19249064e-01 4.98936549e-02 -6.18810952e-01
2.72419274e-01 -3.48557681e-01 3.76475841e-01 -1.30878687e+00
4.73434538e-01 -2.13043168e-01 8.00612867e-01 -1.11417413e+00
-4.21951294e-01 -6.87611639e-01 3.69110882e-01 6.58597767e-01
-3.38569954e-02 -6.67673275e-02 2.20819399e-01 4.65220630e-01
-3.63967329e-01 5.50095201e-01 1.19504917e+00 -6.00801557e-02
-1.32254303e+00 8.99883807e-01 -2.36753926e-01 -2.28006944e-01
7.78613389e-01 -1.26039952e-01 -5.84586322e-01 3.96826770e-03
-4.91920769e-01 6.85139239e-01 2.59421617e-01 2.56247997e-01
7.13074267e-01 -1.49366176e+00 -1.06896317e+00 4.04869705e-01
3.32694769e-01 4.76821214e-01 8.74001324e-01 8.15871835e-01
-7.11456418e-01 -2.01721981e-01 -4.40670073e-01 -7.42760479e-01
-1.00433445e+00 2.65739828e-01 6.21270776e-01 -1.10876724e-01
-1.05456102e+00 8.38023782e-01 3.68220173e-02 -6.32226944e-01
-1.73237860e-01 -5.45928180e-01 -5.20221591e-01 2.57659167e-01
9.10606384e-01 5.79057217e-01 2.23280340e-01 -5.28619170e-01
-4.27869201e-01 9.70535576e-01 1.16372362e-01 3.97304595e-01
1.96884131e+00 -7.95026839e-01 -3.59721571e-01 1.14841692e-01
1.10633707e+00 -4.71370757e-01 -1.42804205e+00 -7.61148751e-01
-3.47099721e-01 -5.97834170e-01 7.42380083e-01 -7.39289284e-01
-1.35009146e+00 6.46057308e-01 1.26367188e+00 1.16118662e-01
1.70575178e+00 -3.08345973e-01 7.83501267e-01 4.82394338e-01
9.33536217e-02 -1.21256125e+00 -1.10245816e-01 6.75225437e-01
1.00443041e+00 -1.41645467e+00 -8.18799436e-03 -4.11018938e-01
-2.21845701e-01 8.88150394e-01 6.80868447e-01 -2.40291357e-01
1.02254879e+00 2.37975299e-01 1.69992864e-01 -4.85746592e-01
-3.91681105e-01 -6.51725948e-01 1.25637785e-01 5.91894150e-01
-2.59032905e-01 1.05204798e-01 -5.85920364e-02 3.71548951e-01
2.13272050e-01 2.09608823e-01 3.34392369e-01 8.20234656e-01
-9.36021328e-01 -2.23766595e-01 -3.13709408e-01 7.17548847e-01
-1.41645402e-01 -2.58597106e-01 3.57842520e-02 4.75457489e-01
2.79621601e-01 8.89231026e-01 5.33084035e-01 -4.67320979e-01
3.63384098e-01 -6.18293285e-01 -1.48225026e-02 -2.22020328e-01
-5.36597669e-01 -7.73917213e-02 -2.68470943e-01 -6.23644590e-01
-7.76202321e-01 -7.05986023e-01 -1.34721935e+00 -3.46362770e-01
-1.35533050e-01 -3.33780795e-01 4.93497163e-01 7.77902722e-01
6.34156823e-01 8.81742358e-01 1.27647185e+00 -8.42782319e-01
4.64154743e-02 -1.17538619e+00 -7.58196652e-01 1.52114093e-01
6.89319909e-01 -2.68921763e-01 -2.49956548e-01 9.73732471e-02] | [9.783567428588867, -1.345316767692566] |
ce2043f5-a7ae-4190-aad5-d3d7133784ba | towards-a-multi-entity-aspect-based-sentiment | null | null | https://aclanthology.org/2022.woah-1.19 | https://aclanthology.org/2022.woah-1.19.pdf | Towards a Multi-Entity Aspect-Based Sentiment Analysis for Characterizing Directed Social Regard in Online Messaging | Online messaging is dynamic, influential, and highly contextual, and a single post may contain contrasting sentiments towards multiple entities, such as dehumanizing one actor while empathizing with another in the same message.These complexities are important to capture for understanding the systematic abuse voiced within an online community, or for determining whether individuals are advocating for abuse, opposing abuse, or simply reporting abuse. In this work, we describe a formulation of directed social regard (DSR) as a problem of multi-entity aspect-based sentiment analysis (ME-ABSA), which models the degree of intensity of multiple sentiments that are associated with entities described by a text document. Our DSR schema is informed by Bandura’s psychosocial theory of moral disengagement and by recent work in ABSA. We present a dataset of over 2,900 posts and sentences, comprising over 24,000 entities annotated for DSR over nine psychosocial dimensions by three annotators. We present a novel transformer-based ME-ABSA model for DSR, achieving favorable preliminary results on this dataset. | ['Christopher Miller', 'Diana Gomez', 'Jeremy Gottlieb', 'Ruta Wheelock', 'Ian Magnusson', 'Sonja Schmer-Galunder', 'Scott Friedman', 'Joan Zheng'] | null | null | null | null | naacl-woah-2022-7 | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 7.30557367e-02 2.87137359e-01 -4.60859984e-01 -7.10992515e-01
-3.11559081e-01 -7.40690589e-01 7.81876504e-01 9.42142725e-01
-3.99877608e-01 4.23882484e-01 1.17778385e+00 -7.52077263e-04
-7.83005655e-02 -5.18918455e-01 -9.61327776e-02 -2.65297890e-01
2.23163180e-02 4.64079976e-01 -2.28108197e-01 -7.69136488e-01
4.74748313e-01 3.80969137e-01 -6.49470508e-01 7.53998876e-01
3.65266830e-01 6.80261314e-01 -8.80158186e-01 6.76080585e-01
-1.71901256e-01 1.49801064e+00 -9.33682144e-01 -1.64611959e+00
-4.28262889e-01 -3.33140045e-01 -1.04203904e+00 -1.01974458e-01
3.63205194e-01 -1.58678278e-01 -1.23218901e-01 1.00238633e+00
4.03619647e-01 -2.60757893e-01 6.76607609e-01 -1.51603794e+00
-5.52123308e-01 9.53338325e-01 -5.54408252e-01 4.10854727e-01
8.74328732e-01 -3.58671218e-01 1.26612496e+00 -5.91657817e-01
1.23903823e+00 1.41765153e+00 1.02786863e+00 6.18791997e-01
-1.09729421e+00 -4.58137721e-01 -3.11822612e-02 -1.37961552e-01
-9.63587642e-01 -5.35276413e-01 8.62855256e-01 -7.07695365e-01
8.07986677e-01 5.62581480e-01 9.22874689e-01 1.57549572e+00
2.08056167e-01 7.05775023e-01 1.06234694e+00 7.08483160e-02
1.00435287e-01 4.87872124e-01 6.10484064e-01 3.02207083e-01
3.71509165e-01 -7.69176543e-01 -1.04308081e+00 -1.16821730e+00
-3.62403274e-01 -9.53405723e-02 1.25381693e-01 8.50671679e-02
-9.18379962e-01 1.15101063e+00 2.88747191e-01 5.57841182e-01
-4.43113416e-01 -9.79065076e-02 7.87169278e-01 1.86209708e-01
8.55154395e-01 7.02767372e-01 -1.90813243e-01 -6.71854019e-01
-5.33035815e-01 6.30924165e-01 1.22620153e+00 5.55578768e-01
2.88745224e-01 -5.86221993e-01 1.73215985e-01 9.89885986e-01
3.36238146e-01 4.03009564e-01 2.79713333e-01 -1.00190675e+00
4.18328345e-01 9.03228879e-01 -5.17820427e-03 -1.84839368e+00
-5.26164174e-01 -8.10529515e-02 -4.48821813e-01 -4.30207938e-01
9.70300213e-02 -2.33566359e-01 1.08801655e-01 1.79359829e+00
5.76646030e-01 -3.18413943e-01 1.98202968e-01 5.36397874e-01
8.83970916e-01 3.17955405e-01 4.46256727e-01 -2.26942047e-01
1.38679039e+00 -4.31244224e-01 -1.13422275e+00 -4.97485965e-01
1.16779315e+00 -6.90841734e-01 2.71612704e-01 3.16432834e-01
-1.09479642e+00 6.26117706e-01 -8.22094381e-01 -3.17720979e-01
-6.08218133e-01 -5.63616037e-01 5.60232937e-01 7.78299868e-01
-6.05922401e-01 6.16566360e-01 -1.48805454e-01 -6.59441411e-01
6.08566761e-01 -1.46123752e-01 -7.69461453e-01 2.28011116e-01
-1.35253596e+00 1.13301396e+00 -1.20322920e-01 -9.76888463e-03
-2.33190358e-02 -5.90050638e-01 -8.17397714e-01 -4.24093157e-01
1.45023629e-01 -3.19450915e-01 1.01828086e+00 -1.06481409e+00
-8.96369278e-01 1.50310075e+00 -6.65228590e-02 -3.50358516e-01
3.95712435e-01 -4.53728795e-01 -9.12667453e-01 9.60706398e-02
4.45154071e-01 5.87982982e-02 6.83289707e-01 -1.13084888e+00
-4.29188199e-02 -9.57932889e-01 4.89148974e-01 3.39296907e-01
-6.48750603e-01 1.16170609e+00 4.94820178e-01 -6.90472305e-01
-2.71989584e-01 -9.03567731e-01 -1.17185198e-01 -2.32636645e-01
-6.15544558e-01 -1.59143880e-01 7.33305633e-01 -7.51953959e-01
1.65178359e+00 -2.15283155e+00 3.98336649e-01 1.83308676e-01
7.09147692e-01 8.63979235e-02 3.41634065e-01 1.03036642e+00
-7.63588294e-04 6.38925672e-01 -3.18930596e-01 -6.15848005e-01
-2.72825230e-02 2.75691569e-01 -5.18823087e-01 7.05633998e-01
3.17147300e-02 5.88137805e-01 -1.17935598e+00 -6.26699209e-01
-5.26262403e-01 3.24995846e-01 -4.72657233e-01 1.01238988e-01
1.80485100e-01 9.36046764e-02 -5.43395281e-01 6.93210006e-01
4.38593060e-01 -1.01587132e-01 5.10781348e-01 -1.54176787e-01
-6.40267655e-02 3.30821455e-01 -6.65191948e-01 1.04929483e+00
-1.84781432e-01 9.78398621e-01 4.47606891e-01 -4.23513681e-01
6.45869613e-01 1.97108626e-01 4.50892329e-01 -1.65024847e-01
5.46161294e-01 1.73544705e-01 -1.22484304e-01 -6.01129413e-01
9.30588186e-01 -4.83616620e-01 -7.81384289e-01 9.31964457e-01
1.80232842e-02 -1.84416413e-01 4.45190296e-02 9.99548495e-01
1.24370694e+00 -5.06739974e-01 6.42933547e-01 -1.80643052e-01
2.08076254e-01 1.18326649e-01 5.64878523e-01 3.53608787e-01
-6.35470331e-01 1.70896903e-01 1.19498897e+00 -4.67867076e-01
-8.63705516e-01 -6.67561293e-01 -3.44825089e-01 1.17278588e+00
2.19683647e-01 -9.20356989e-01 -5.60523808e-01 -7.34467506e-01
5.92592582e-02 8.51253450e-01 -9.12935972e-01 -6.47397572e-03
-4.30857420e-01 -8.76704395e-01 1.09213245e+00 1.65597543e-01
1.40599176e-01 -7.48168230e-01 -4.31506157e-01 1.96382329e-01
-5.46498835e-01 -1.37058210e+00 -1.02307305e-01 -3.65886599e-01
-2.16272995e-01 -1.22233951e+00 -1.62908301e-01 -3.90961878e-02
5.32545686e-01 -3.57530452e-02 1.18183124e+00 3.02778006e-01
-2.56228577e-02 4.41300005e-01 -6.98018253e-01 -5.53262770e-01
-7.72831380e-01 -2.91478224e-02 2.03386799e-01 2.96760827e-01
5.83523393e-01 -4.71423984e-01 -3.22751045e-01 8.40694159e-02
-1.10405898e+00 -2.91631907e-01 -2.18458369e-01 3.73090714e-01
-4.42195445e-01 -6.34598136e-01 5.42582572e-01 -1.56078804e+00
1.17586398e+00 -1.39103079e+00 6.40275121e-01 1.08057730e-01
-2.69250989e-01 -7.19721258e-01 3.41381252e-01 -3.76220435e-01
-8.97141159e-01 -8.02383840e-01 -2.75124997e-01 1.58324435e-01
9.27944705e-02 5.09706795e-01 3.11537117e-01 1.76820591e-01
8.63567948e-01 -5.11854351e-01 -1.14560708e-01 -8.97408500e-02
3.72290879e-01 1.17713881e+00 3.86072546e-02 -4.12030876e-01
6.34310782e-01 7.56952405e-01 -2.37444207e-01 -9.63553250e-01
-1.34483540e+00 -7.71107972e-01 -4.02084827e-01 -7.31120586e-01
9.36294913e-01 -8.07807744e-01 -7.59694517e-01 7.87359953e-01
-1.38135457e+00 6.46530464e-02 3.14143747e-02 4.00443971e-02
-4.82194014e-02 4.29187506e-01 -8.62435818e-01 -8.79478991e-01
-3.82550299e-01 -5.45197308e-01 7.58090913e-01 -1.37278095e-01
-1.25215435e+00 -1.23227692e+00 6.95682764e-01 9.43448007e-01
4.54151183e-01 7.05323100e-01 6.68389320e-01 -1.38632631e+00
7.05265045e-01 -4.97366846e-01 3.55720893e-02 7.81146288e-02
3.50576937e-02 2.92282283e-01 -7.38735318e-01 9.53956321e-03
2.15031207e-01 -8.80243838e-01 1.14786150e-02 -5.53608656e-01
4.08377767e-01 -8.50002706e-01 -2.64799595e-01 -1.53208882e-01
1.06871378e+00 -3.03812385e-01 6.07359707e-01 5.19337773e-01
6.40410423e-01 1.09414017e+00 5.82584441e-01 8.08467388e-01
5.43871224e-01 4.42668974e-01 2.62430429e-01 3.17791164e-01
5.67801416e-01 -2.63607085e-01 6.52776420e-01 9.24935400e-01
3.74730900e-02 -3.02774042e-01 -1.02529657e+00 6.66660786e-01
-1.71017146e+00 -1.25061858e+00 -5.48634648e-01 1.25926733e+00
8.21524560e-01 -9.77357998e-02 1.15923114e-01 -5.69438376e-02
7.15856433e-01 8.01397026e-01 -1.72093436e-01 -1.14602590e+00
-3.14135551e-01 -2.14309633e-01 1.36142001e-01 5.65903902e-01
-7.11808324e-01 7.61745214e-01 6.48230171e+00 4.86248463e-01
-6.39744282e-01 3.01855952e-01 7.33205557e-01 -1.56965017e-01
-6.37827694e-01 1.01564035e-01 -5.38726389e-01 3.70104820e-01
9.62299526e-01 -4.24660951e-01 1.56664565e-01 9.92222488e-01
-1.33415833e-01 -1.43792659e-01 -9.26143765e-01 6.10953271e-01
7.24538565e-01 -1.08932149e+00 -1.85678393e-01 2.82674730e-02
6.50255740e-01 -3.10406759e-02 -2.28515547e-02 3.92510481e-02
2.77327895e-01 -7.61154950e-01 7.91240931e-01 1.85722753e-01
2.68245459e-01 -3.59124690e-01 7.89612889e-01 2.76811212e-01
-4.44202483e-01 -5.24605066e-02 1.03878930e-01 -1.43140435e-01
5.98621368e-01 6.08548939e-01 -5.47317624e-01 2.20758528e-01
6.63587272e-01 1.08406389e+00 -4.68985051e-01 5.74866198e-02
-1.72499046e-01 3.27202171e-01 -1.48883879e-01 -3.83628786e-01
7.50854537e-02 -2.76367068e-01 1.08615208e+00 1.61401379e+00
-2.53996521e-01 4.43059951e-01 -4.86655444e-01 4.49398220e-01
-3.88679534e-01 2.85770357e-01 -9.65879798e-01 -6.16244256e-01
4.21201676e-01 1.70279765e+00 -4.14449245e-01 -3.27599108e-01
-1.77434862e-01 8.86211395e-01 6.53638005e-01 -2.68623326e-02
-6.57475948e-01 -4.48013246e-02 7.13405728e-01 2.93219775e-01
-4.29083735e-01 7.67526478e-02 -1.74816206e-01 -1.41615725e+00
-3.24418843e-01 -9.25101280e-01 4.16924059e-01 -9.50963855e-01
-1.84545326e+00 5.37518203e-01 5.45354784e-02 -7.37596750e-01
-2.22559273e-01 -4.77237165e-01 -6.21796191e-01 2.43739486e-01
-8.76937330e-01 -1.22314143e+00 -3.91875617e-02 3.54874343e-01
8.64322931e-02 2.29781300e-01 7.42397428e-01 3.27969819e-01
-5.45773149e-01 3.08975220e-01 -3.87767464e-01 2.74232984e-01
9.63293493e-01 -1.09230554e+00 1.20494843e-01 4.19186443e-01
-3.15031052e-01 8.95507514e-01 9.57847297e-01 -8.70017469e-01
-1.24790204e+00 -6.93210959e-01 1.49055350e+00 -1.04747856e+00
1.50853837e+00 -4.53970492e-01 -7.94799924e-01 9.97567356e-01
3.28501701e-01 -3.68886858e-01 1.70326388e+00 5.46982527e-01
-7.67870069e-01 3.35239172e-01 -1.57908034e+00 8.07068348e-01
1.26106811e+00 -8.14929843e-01 -9.95535374e-01 5.82218707e-01
3.92133564e-01 -2.30190367e-01 -1.22372925e+00 5.81251085e-02
5.74089766e-01 -9.88981903e-01 3.87380600e-01 -1.27558017e+00
1.02159131e+00 4.18042630e-01 -2.45740816e-01 -1.26362967e+00
-7.82778561e-02 -6.84291780e-01 -1.46778405e-01 1.47429609e+00
4.18279022e-01 -6.12790942e-01 2.03437522e-01 1.16053247e+00
1.77164093e-01 -7.89317906e-01 -8.48662138e-01 -4.83472459e-02
1.27457082e-01 -6.06842756e-01 2.95867294e-01 1.82209146e+00
7.88535476e-01 5.96725166e-01 -4.35923874e-01 -1.43197075e-01
4.78048980e-01 -2.20717594e-01 8.09975445e-01 -1.08928144e+00
-3.13112102e-02 -3.23313534e-01 -5.71819007e-01 -1.69351116e-01
2.83964545e-01 -7.97539949e-01 -5.94302118e-01 -1.35914290e+00
5.71394563e-01 -2.47449711e-01 3.86753947e-01 2.53425568e-01
2.89594196e-02 2.86900491e-01 2.07858428e-01 2.37510547e-01
-8.49637985e-01 4.23536688e-01 6.84522808e-01 1.32144108e-01
1.62857890e-01 -5.39326429e-01 -1.22474742e+00 1.27532828e+00
5.67887008e-01 -7.68401623e-01 1.84192896e-01 -6.23671077e-02
1.45068955e+00 -9.24815312e-02 2.65355289e-01 -1.42899975e-01
1.34268507e-01 -2.72629738e-01 -1.42489210e-01 -1.48756430e-01
5.20800650e-01 -6.41503155e-01 2.12969214e-01 3.06691498e-01
-6.82733357e-01 1.46437168e-01 -1.59209132e-01 6.07661426e-01
-2.60391176e-01 -4.82775271e-01 5.60739398e-01 -1.84195098e-02
-8.69027227e-02 -8.34926814e-02 -5.63067853e-01 6.32748604e-01
1.16142416e+00 1.29325911e-01 -8.39660645e-01 -6.56408310e-01
-7.04528272e-01 6.95645437e-02 5.56782842e-01 3.98966432e-01
5.11276007e-01 -1.18585014e+00 -9.04995501e-01 -5.98640621e-01
4.12246913e-01 -6.52974725e-01 4.77401704e-01 8.08870137e-01
-3.41466904e-01 -3.75120342e-01 1.61451213e-02 1.05735451e-01
-1.49199760e+00 1.97917014e-01 1.76554978e-01 -2.63503909e-01
-1.27002120e-01 6.45815492e-01 -3.38289589e-01 -4.92215574e-01
-4.40795749e-01 5.61226428e-01 -6.00425780e-01 7.69543767e-01
5.95488369e-01 6.13715351e-01 -3.04462314e-01 -1.40930068e+00
-6.04618788e-01 2.66939551e-01 -3.89967799e-01 -2.06623137e-01
1.47170973e+00 -1.09126016e-01 -9.31352735e-01 8.17320406e-01
1.50839448e+00 7.97323167e-01 -1.53162312e-02 -2.79586557e-02
-9.05301794e-03 -6.51456952e-01 -3.02033573e-01 -6.84353471e-01
-4.72711980e-01 2.47407362e-01 -3.75495464e-01 8.38120878e-01
5.17978191e-01 3.86771291e-01 8.20635259e-01 1.39901161e-01
1.31524339e-01 -1.31487000e+00 3.03287983e-01 4.23425049e-01
1.24058080e+00 -9.97577846e-01 3.84561211e-01 -6.33830786e-01
-1.32435679e+00 9.27888870e-01 5.03749847e-01 -2.03564093e-02
7.32878149e-01 1.42497361e-01 1.34627759e-01 -8.52401197e-01
-7.08964467e-01 4.71710414e-01 -2.65868813e-01 5.30291684e-02
4.33600068e-01 1.70555547e-01 -8.13026965e-01 7.10260808e-01
-3.49737585e-01 -4.76792425e-01 1.02585924e+00 1.02679908e+00
1.53878227e-01 -9.00620282e-01 2.48672273e-02 4.16086644e-01
-1.11889327e+00 4.01679687e-02 -1.36548519e+00 5.30692458e-01
-2.60412972e-03 1.02563167e+00 -2.84628896e-03 -4.71445799e-01
2.18815699e-01 -1.03515692e-01 1.64109785e-02 -5.52986562e-01
-1.34058201e+00 -6.44642472e-01 1.10882175e+00 -5.32073140e-01
-9.88745987e-01 -1.00393486e+00 -9.62081254e-01 -8.03601980e-01
-1.44340888e-01 1.78891376e-01 7.33928800e-01 1.16120052e+00
4.60700035e-01 -2.44078636e-01 8.11030209e-01 -1.92205712e-01
-1.90578267e-01 -9.44553733e-01 -4.83763039e-01 9.63839591e-01
1.89955860e-01 -3.72382045e-01 -7.23793089e-01 -3.82099122e-01] | [8.673334121704102, 10.475805282592773] |
f138c49e-4076-4f33-83af-671f65025683 | swipenet-object-detection-in-noisy-underwater | 2010.10006 | null | https://arxiv.org/abs/2010.10006v3 | https://arxiv.org/pdf/2010.10006v3.pdf | SWIPENET: Object detection in noisy underwater images | In recent years, deep learning based object detection methods have achieved promising performance in controlled environments. However, these methods lack sufficient capabilities to handle underwater object detection due to these challenges: (1) images in the underwater datasets and real applications are blurry whilst accompanying severe noise that confuses the detectors and (2) objects in real applications are usually small. In this paper, we propose a novel Sample-WeIghted hyPEr Network (SWIPENET), and a robust training paradigm named Curriculum Multi-Class Adaboost (CMA), to address these two problems at the same time. Firstly, the backbone of SWIPENET produces multiple high resolution and semantic-rich Hyper Feature Maps, which significantly improve small object detection. Secondly, a novel sample-weighted detection loss function is designed for SWIPENET, which focuses on learning high weight samples and ignore learning low weight samples. Moreover, inspired by the human education process that drives the learning from easy to hard concepts, we here propose the CMA training paradigm that first trains a clean detector which is free from the influence of noisy data. Then, based on the clean detector, multiple detectors focusing on learning diverse noisy data are trained and incorporated into a unified deep ensemble of strong noise immunity. Experiments on two underwater robot picking contest datasets (URPC2017 and URPC2018) show that the proposed SWIPENET+CMA framework achieves better accuracy in object detection against several state-of-the-art approaches. | ['Huiyu Zhou', 'Xin Wang', 'Haiping Ma', 'Ning li', 'Junyu Dong', 'Shengke Wang', 'Feixiang Zhou', 'Long Chen'] | 2020-10-19 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [ 3.94102633e-02 -1.47310212e-01 7.39112735e-01 -2.20538273e-01
-5.97364783e-01 -7.77325258e-02 4.03514594e-01 2.45249830e-02
-9.93141711e-01 3.95851582e-01 -8.64791721e-02 2.05814198e-01
-3.68011683e-01 -9.89531934e-01 -8.75277936e-01 -1.12517428e+00
-1.42694935e-02 2.16174439e-01 8.19726646e-01 -4.70728457e-01
1.48485392e-01 1.82031855e-01 -1.79614496e+00 -4.29487228e-03
1.11864114e+00 9.21512425e-01 7.89869487e-01 6.09250605e-01
-1.37291074e-01 4.67752486e-01 -6.60459161e-01 -3.52982044e-01
5.14319241e-01 -2.35815480e-01 -1.42276466e-01 -2.63768047e-01
4.81076598e-01 -4.81594950e-01 -3.93420905e-01 1.40269756e+00
1.00969255e+00 1.99116498e-01 5.53608298e-01 -9.64792669e-01
-3.77992690e-01 6.82303786e-01 -6.72204971e-01 2.89483637e-01
-7.77632892e-02 2.82846898e-01 8.23428392e-01 -1.02318192e+00
1.93377852e-01 1.44246149e+00 7.71627009e-01 7.90805459e-01
-7.95409858e-01 -8.52274597e-01 3.41015011e-01 2.58314580e-01
-1.07327247e+00 -1.11966513e-01 5.29382825e-01 -3.79853934e-01
5.06951928e-01 6.78222850e-02 9.52430010e-01 1.06461895e+00
2.48507373e-02 1.17612159e+00 9.79546726e-01 -2.88798153e-01
3.18618774e-01 9.60203856e-02 2.09271178e-01 5.91526210e-01
6.06097817e-01 1.62538752e-01 -2.83449441e-01 5.39966337e-02
3.81387472e-01 2.15983167e-01 -4.28501695e-01 -3.67650360e-01
-6.17015779e-01 8.02600980e-01 6.80633247e-01 5.00622988e-02
-2.77141601e-01 -1.66840523e-01 5.00741482e-01 1.72505438e-01
6.51663318e-02 2.94633836e-01 -2.28940889e-01 4.14166391e-01
-4.61621493e-01 4.73299056e-01 6.79656148e-01 9.83787835e-01
7.48073936e-01 2.67433017e-01 -1.58531651e-01 1.04176903e+00
5.75256348e-01 8.87019455e-01 5.70321679e-01 -3.86218935e-01
3.82477582e-01 4.87461746e-01 1.22572044e-02 -9.10213411e-01
-6.20609879e-01 -5.98037541e-01 -6.21774256e-01 7.08717942e-01
1.50047258e-01 -2.35763445e-01 -1.16850054e+00 1.43096292e+00
2.46764988e-01 2.92736501e-01 5.54695308e-01 1.21892214e+00
1.34220862e+00 7.99057722e-01 8.78866315e-02 4.67355028e-02
1.22603714e+00 -1.01314926e+00 -5.47530949e-01 -5.10239601e-01
3.05953085e-01 -4.22973037e-01 7.50798762e-01 5.60587645e-01
-8.28679383e-01 -7.00631499e-01 -1.43783629e+00 1.82971492e-01
-3.81304771e-01 7.52913486e-03 2.95053750e-01 6.09207749e-01
-7.11702526e-01 3.02309602e-01 -6.54738784e-01 -2.01477304e-01
6.27719641e-01 1.85005337e-01 -1.32676288e-01 -5.16132891e-01
-1.02279055e+00 1.03527403e+00 6.79074526e-01 4.39259171e-01
-1.38466632e+00 -4.28653151e-01 -1.07260787e+00 1.15742445e-01
4.05066490e-01 -2.82104850e-01 1.21608031e+00 -6.29349709e-01
-1.47920859e+00 3.07505190e-01 5.67034006e-01 -4.13490385e-01
9.09175396e-01 -5.87610304e-01 -1.46989286e-01 2.97884047e-02
7.78745264e-02 5.65555394e-01 9.69670832e-01 -1.56014872e+00
-1.15833104e+00 -3.20036262e-01 -2.25840323e-03 4.81843680e-01
-6.15812659e-01 -1.11275837e-01 -4.77986306e-01 -3.24279130e-01
2.44523719e-01 -5.51355720e-01 -3.13744515e-01 2.26682588e-01
-1.34317398e-01 -2.61326909e-01 1.23526096e+00 -3.00639689e-01
6.83837891e-01 -2.11439657e+00 2.61068314e-01 -7.03374594e-02
1.44703522e-01 8.14990819e-01 -3.56831223e-01 1.57390371e-01
4.35143650e-01 -3.38714182e-01 -4.32595313e-01 -4.71649110e-01
2.58554704e-02 5.96124470e-01 -1.74974009e-01 5.12242615e-01
1.79826215e-01 4.52696294e-01 -1.33306801e+00 -4.89868671e-01
4.21828866e-01 3.65600228e-01 -4.14307296e-01 3.93938661e-01
3.60908136e-02 2.06502870e-01 -4.17127371e-01 7.68763959e-01
1.19110572e+00 4.46952730e-01 -1.86637908e-01 -1.27471447e-01
-4.01521564e-01 -4.37825620e-01 -1.31150603e+00 1.37704933e+00
-3.12045097e-01 4.53288078e-01 5.06338775e-01 -1.03415287e+00
1.05415380e+00 -1.17333950e-02 1.02962581e-02 -7.44209528e-01
2.35086769e-01 4.42356378e-01 9.60770175e-02 -9.65493798e-01
3.82982850e-01 -1.18962273e-01 2.35985160e-01 -3.43862504e-01
2.60529935e-01 -1.57156691e-01 2.45064899e-01 -3.61217298e-02
9.48949158e-01 1.45834908e-01 -9.55119655e-02 -3.86855304e-01
4.07515615e-01 -2.38335475e-01 1.08091307e+00 1.13063323e+00
-3.73263240e-01 7.27681458e-01 -6.86597973e-02 -5.27164221e-01
-8.17538559e-01 -9.48371589e-01 -4.72257704e-01 1.11092269e+00
8.89739990e-01 1.68401554e-01 -5.63593209e-01 -6.18067503e-01
1.01057142e-01 1.65643513e-01 -6.47394359e-01 -1.35025650e-01
-5.97651839e-01 -1.26059842e+00 5.12393415e-01 6.86705232e-01
6.58406496e-01 -1.31385493e+00 -8.74607325e-01 3.84891778e-01
9.22593549e-02 -1.08516097e+00 5.30549176e-02 2.79327840e-01
-6.25993967e-01 -1.03993738e+00 -1.02734005e+00 -1.05543756e+00
6.33837700e-01 6.56829298e-01 6.46808445e-01 2.96629488e-01
-5.27269900e-01 8.63895714e-02 -7.48285234e-01 -1.05240393e+00
-1.41470641e-01 -4.68985915e-01 2.16868311e-01 4.00058515e-02
3.34401757e-01 -3.09759170e-01 -7.27426350e-01 2.37901106e-01
-1.04660535e+00 -2.46303558e-01 1.08132422e+00 1.19630814e+00
8.76204595e-02 -6.40987679e-02 5.36439002e-01 -4.95173454e-01
1.55295879e-01 -4.57287252e-01 -7.99967527e-01 3.15176733e-02
-3.44350874e-01 -3.49323332e-01 4.51714069e-01 -5.40978551e-01
-9.59977627e-01 2.70380210e-02 -3.94908309e-01 -3.77373159e-01
8.36308450e-02 2.97724962e-01 -9.37501565e-02 -5.07168949e-01
6.98220849e-01 4.78766799e-01 -2.03631055e-02 -4.83082175e-01
-2.84845442e-01 7.18797624e-01 6.07529223e-01 -3.93831372e-01
1.12590837e+00 5.58311045e-01 -2.76347816e-01 -1.13356102e+00
-7.94435680e-01 -7.89882183e-01 -2.80590981e-01 -3.52192432e-01
8.95897508e-01 -1.30059481e+00 -6.40580177e-01 9.93039072e-01
-1.00478399e+00 -2.85343766e-01 -2.40911474e-03 6.59451663e-01
-9.57173258e-02 4.91235971e-01 -5.99975049e-01 -1.21425116e+00
-5.93063176e-01 -1.20469773e+00 9.16506529e-01 9.91698325e-01
8.19914043e-01 -5.21023512e-01 -7.14062825e-02 1.25731692e-01
4.42384124e-01 1.90270618e-01 2.44144663e-01 -5.97812831e-01
-5.92905521e-01 -2.17589512e-01 -4.59212035e-01 5.85309744e-01
-2.65472710e-01 -1.00257732e-01 -1.09398293e+00 -6.18351877e-01
-5.97701780e-02 -6.01603329e-01 1.35829592e+00 2.10915327e-01
7.83531785e-01 -1.45934103e-02 -2.96572179e-01 7.11449981e-01
1.63393605e+00 1.59006715e-01 6.55137658e-01 6.69761777e-01
7.55912781e-01 6.35161519e-01 6.30897701e-01 4.34853405e-01
3.85860533e-01 3.24037969e-01 9.17645276e-01 -2.34679535e-01
-3.79227288e-02 -6.40777424e-02 3.24413657e-01 6.76479220e-01
-1.37304127e-01 -1.41794309e-01 -6.69665635e-01 8.58060300e-01
-1.96129167e+00 -7.58716643e-01 -2.93038428e-01 1.86472762e+00
5.32068968e-01 3.07996303e-01 -1.98518321e-01 1.58231094e-01
6.52718484e-01 -5.83491847e-02 -5.77331781e-01 1.36258513e-01
-4.04500425e-01 -1.25111818e-01 7.05554426e-01 1.49272561e-01
-1.37130249e+00 6.98852718e-01 4.66322565e+00 7.66851962e-01
-8.18982124e-01 4.91895229e-02 -6.43957555e-02 3.56643766e-01
6.95420057e-02 -2.44295955e-01 -1.09520400e+00 5.21168828e-01
1.69796109e-01 1.97225869e-01 -1.55742750e-01 1.11740375e+00
-5.08365519e-02 -2.43174434e-02 -6.33577108e-01 8.63856375e-01
2.37849891e-01 -8.62860322e-01 6.61022291e-02 -2.60333240e-01
7.87036896e-01 2.73786366e-01 -2.05372110e-01 8.12051594e-01
4.29602861e-01 -5.41028142e-01 9.43610966e-01 4.28377330e-01
1.04131967e-01 -5.02249181e-01 1.26295197e+00 5.14792204e-01
-1.06146157e+00 -8.09521914e-01 -1.06396246e+00 -6.90227672e-02
-3.70432436e-02 3.32593560e-01 -4.64222103e-01 5.52319944e-01
1.39367592e+00 4.88359898e-01 -3.94496650e-01 1.95214069e+00
-3.13818902e-01 2.79352278e-01 -3.27995121e-01 -3.88572544e-01
5.25856912e-01 -1.84448466e-01 7.85181880e-01 1.46407354e+00
3.61669540e-01 2.50098377e-01 4.72959280e-01 5.05040467e-01
-1.40240667e-02 4.70425375e-02 -3.45288336e-01 7.39764273e-01
4.59967971e-01 1.53021002e+00 -4.72063392e-01 -2.69864261e-01
-5.42286217e-01 5.39810598e-01 3.05165797e-01 1.72430083e-01
-4.88720864e-01 -6.09945655e-01 5.60465395e-01 -3.97455066e-01
5.15474021e-01 6.97937235e-02 -5.67427790e-03 -9.50823486e-01
-5.32859145e-03 -6.99298441e-01 3.01672578e-01 -5.45180023e-01
-1.59899724e+00 6.01200342e-01 -1.67993978e-01 -1.55195618e+00
7.00810373e-01 -9.10072744e-01 -8.97773147e-01 6.57937348e-01
-2.19352937e+00 -1.07513893e+00 -9.38758135e-01 1.40125215e-01
8.54019701e-01 -1.34227216e-01 5.05230308e-01 4.58257705e-01
-7.70304918e-01 5.83893418e-01 2.93220609e-01 5.02690554e-01
5.48801601e-01 -1.36551595e+00 6.85084984e-02 9.95330930e-01
-2.54638940e-01 2.86237448e-01 7.36225605e-01 -5.37312090e-01
-1.44302738e+00 -1.23184741e+00 -9.92633700e-02 -3.95647101e-02
4.40452635e-01 -4.04854268e-01 -1.31249964e+00 1.50926515e-01
-1.94225158e-03 3.51099461e-01 1.95159286e-01 -4.10365403e-01
-2.24372000e-01 -3.67862523e-01 -1.06044233e+00 3.90897453e-01
7.83484817e-01 2.10423633e-01 -6.95247591e-01 3.88139844e-01
5.75022519e-01 -6.24070048e-01 -5.02452075e-01 7.99059272e-01
5.50427914e-01 -9.57862139e-01 9.45317507e-01 -2.03842118e-01
1.80574194e-01 -5.45728445e-01 -7.39463195e-02 -1.30964637e+00
-2.23691121e-01 -3.19355547e-01 2.13759169e-02 8.88003230e-01
2.21049801e-01 -4.89379942e-01 6.44972146e-01 5.65986261e-02
-8.05050552e-01 -3.84287357e-01 -1.01112795e+00 -9.32625115e-01
9.65103332e-04 7.23272562e-02 7.01647550e-02 6.68854892e-01
-3.14490736e-01 5.39925396e-02 -4.96471971e-01 8.10437620e-01
1.16773105e+00 -8.34878832e-02 8.39070439e-01 -1.54896522e+00
-2.14648053e-01 -4.69784558e-01 -6.84790671e-01 -1.04575384e+00
-3.66724283e-01 -3.27589810e-01 9.49058831e-01 -1.65162492e+00
3.91542524e-01 -3.71127635e-01 -3.00306141e-01 3.63745779e-01
-6.34254575e-01 3.90182227e-01 2.74224967e-01 1.01936683e-01
-7.11965919e-01 9.89006221e-01 1.18152440e+00 -3.33323151e-01
1.76960211e-02 1.30864173e-01 -3.24998915e-01 8.31008315e-01
3.45166862e-01 -5.25992155e-01 -1.74939539e-02 -7.42546797e-01
-2.67196059e-01 -2.48811036e-01 3.31841886e-01 -1.50627315e+00
6.16923153e-01 1.19877130e-01 4.22818393e-01 -3.90499890e-01
3.76804382e-01 -7.48703182e-01 -6.65730298e-01 8.87517691e-01
1.49495348e-01 -4.79532719e-01 1.79152325e-01 9.72160637e-01
-1.85844630e-01 -7.35103309e-01 1.21906531e+00 -2.47249454e-01
-1.17667270e+00 2.32169628e-01 -2.26516545e-01 -3.47827785e-02
1.04587746e+00 -2.98472285e-01 -5.38042724e-01 1.28719613e-01
-3.71235698e-01 8.20241630e-01 -7.68327340e-02 6.20167494e-01
9.98790681e-01 -7.83041298e-01 -1.11870909e+00 1.45491436e-01
3.27276051e-01 5.29215813e-01 5.66431165e-01 6.26151681e-01
-7.41514802e-01 -3.57492536e-01 -3.35208654e-01 -7.78020620e-01
-1.12792134e+00 2.81255513e-01 4.49446976e-01 3.20700467e-01
-9.91723835e-01 1.23653471e+00 3.77561003e-01 -6.46601439e-01
6.28652632e-01 -1.04761213e-01 -5.86867392e-01 8.00117776e-02
8.90018940e-01 3.64841670e-01 3.59689863e-03 -4.42611545e-01
-1.55607909e-01 7.75337458e-01 -2.52049446e-01 3.50532532e-01
1.63192165e+00 1.75849050e-02 1.97371617e-01 5.62478900e-02
8.43487620e-01 -3.06518197e-01 -1.77812624e+00 -4.69750047e-01
-1.38175502e-01 -4.04462427e-01 1.53598577e-01 -5.51344812e-01
-9.52054560e-01 9.33388948e-01 1.04454100e+00 1.19666368e-01
1.02458894e+00 -2.31875896e-01 8.05795670e-01 6.72765017e-01
3.06211948e-01 -1.18587601e+00 3.21486264e-01 6.45785749e-01
6.76367700e-01 -1.60595477e+00 -3.18583623e-02 -1.64286181e-01
-4.64239597e-01 1.06468260e+00 1.08070004e+00 -3.33931118e-01
2.69338638e-01 3.75712574e-01 4.63683069e-01 -1.68323338e-01
-1.44647539e-01 -6.12695992e-01 -1.06056519e-01 6.72177255e-01
-4.75452572e-01 -3.03830057e-01 -2.93825030e-01 6.34388268e-01
1.97120100e-01 -4.63326365e-01 7.20238626e-01 1.16706157e+00
-1.17344236e+00 -5.58250129e-01 -4.98810232e-01 2.29432270e-01
-3.12881678e-01 5.17080165e-02 2.90092021e-01 7.17567086e-01
3.66049498e-01 8.94578040e-01 -1.54285550e-01 -4.43006784e-01
7.40205467e-01 -4.74074513e-01 9.59411636e-02 -6.02502108e-01
-4.80154067e-01 -1.84823852e-02 -2.82932222e-01 -8.05861652e-02
-3.95743251e-01 -6.09158933e-01 -1.29031265e+00 4.42042232e-01
-7.11690962e-01 3.76634002e-01 7.60147393e-01 9.83443022e-01
-2.83789724e-01 6.74857914e-01 5.72900414e-01 -1.34056866e+00
-8.39899778e-01 -1.34579527e+00 -5.64071834e-01 3.75632435e-01
5.27390659e-01 -8.42913389e-01 -5.53259671e-01 -4.33612764e-01] | [10.642315864562988, -3.468898057937622] |
8818d99f-cd7c-415c-8235-ab97021cdb34 | ensemble-framework-for-cardiovascular-disease | 2306.09989 | null | https://arxiv.org/abs/2306.09989v1 | https://arxiv.org/pdf/2306.09989v1.pdf | Ensemble Framework for Cardiovascular Disease Prediction | Heart disease is the major cause of non-communicable and silent death worldwide. Heart diseases or cardiovascular diseases are classified into four types: coronary heart disease, heart failure, congenital heart disease, and cardiomyopathy. It is vital to diagnose heart disease early and accurately in order to avoid further injury and save patients' lives. As a result, we need a system that can predict cardiovascular disease before it becomes a critical situation. Machine learning has piqued the interest of researchers in the field of medical sciences. For heart disease prediction, researchers implement a variety of machine learning methods and approaches. In this work, to the best of our knowledge, we have used the dataset from IEEE Data Port which is one of the online available largest datasets for cardiovascular diseases individuals. The dataset isa combination of Hungarian, Cleveland, Long Beach VA, Switzerland & Statlog datasets with important features such as Maximum Heart Rate Achieved, Serum Cholesterol, Chest Pain Type, Fasting blood sugar, and so on. To assess the efficacy and strength of the developed model, several performance measures are used, such as ROC, AUC curve, specificity, F1-score, sensitivity, MCC, and accuracy. In this study, we have proposed a framework with a stacked ensemble classifier using several machine learning algorithms including ExtraTrees Classifier, Random Forest, XGBoost, and so on. Our proposed framework attained an accuracy of 92.34% which is higher than the existing literature. | ['Aman Sharma', 'Aryan Chugh', 'Achyut Tiwari'] | 2023-06-16 | null | null | null | null | ['disease-prediction', 'specificity'] | ['medical', 'natural-language-processing'] | [-4.92144562e-02 -1.61641970e-01 -3.54814351e-01 -2.27791995e-01
3.20575498e-02 -1.70574620e-01 -9.09819752e-02 5.29039741e-01
-1.06491834e-01 9.54634249e-01 9.08765793e-02 -6.03444278e-01
-4.11591589e-01 -8.90526891e-01 1.69880256e-01 -3.98963779e-01
-1.00874841e-01 6.39872372e-01 3.29368599e-02 1.93635508e-01
4.91979122e-01 5.34827411e-01 -1.38921678e+00 4.93857525e-02
9.71866071e-01 9.37567174e-01 -2.11472154e-01 6.40208244e-01
1.29008964e-01 6.76381350e-01 -3.52111429e-01 -1.97960824e-01
9.41603333e-02 -7.50843525e-01 -5.35508394e-01 -1.44196913e-01
-1.70577466e-01 -3.39287788e-01 4.25593793e-01 3.76020461e-01
7.31759727e-01 -3.62062126e-01 6.48329616e-01 -9.14278686e-01
-2.92711794e-01 4.18403476e-01 -3.34836394e-01 3.31561416e-01
2.32286453e-02 -8.73290449e-02 5.26382625e-01 -4.86894280e-01
2.79005200e-01 9.08913851e-01 7.07283437e-01 3.65464658e-01
-8.71381164e-01 -7.00012922e-01 -3.90641779e-01 2.58765280e-01
-9.62706804e-01 -4.36357036e-02 5.37807167e-01 -6.86515689e-01
4.12665784e-01 5.26160002e-01 9.08041954e-01 3.77798975e-01
4.71425533e-01 1.65002614e-01 1.52706993e+00 -7.93420553e-01
1.76086500e-01 5.61497808e-01 5.24801731e-01 7.61114955e-01
5.40730417e-01 1.01039626e-01 5.61285857e-03 -3.71093035e-01
5.79888344e-01 3.80490392e-01 -5.19251414e-02 -6.33247420e-02
-9.43865538e-01 8.15431595e-01 -1.06406538e-03 4.96767998e-01
-3.14302593e-01 -4.46613789e-01 3.41675013e-01 1.38050854e-01
1.68324336e-01 1.97279006e-01 -8.56951177e-01 -7.95931220e-02
-5.86042762e-01 -6.37514740e-02 7.91328013e-01 1.70802951e-01
2.20958397e-01 -1.90350398e-01 1.13753952e-01 6.58711493e-01
4.79306102e-01 5.22049904e-01 6.04642510e-01 -5.49994946e-01
1.30474672e-01 9.54066813e-01 -1.21317118e-01 -1.08448935e+00
-5.40432870e-01 -6.82220459e-01 -1.09660006e+00 9.57504362e-02
1.63598910e-01 -4.04872626e-01 -8.44675481e-01 1.28536177e+00
3.89204770e-01 1.19968206e-01 1.78778827e-01 7.85765886e-01
8.90155375e-01 2.73380965e-01 4.45022166e-01 -6.07589126e-01
1.43545485e+00 -3.00254405e-01 -5.15554786e-01 3.11478347e-01
4.06813860e-01 -7.57794440e-01 4.76677150e-01 5.25340378e-01
-7.09568977e-01 -4.04689997e-01 -7.53635645e-01 5.19507945e-01
-3.02708745e-01 4.66763347e-01 6.48783147e-01 1.11911821e+00
-4.45172429e-01 5.57472348e-01 -7.21774638e-01 -8.09546769e-01
2.08052039e-01 1.89821407e-01 -1.27137154e-01 1.75611734e-01
-1.10299897e+00 9.45555270e-01 3.72405469e-01 -1.17868029e-01
-2.43530333e-01 -5.68496048e-01 -2.11243317e-01 -4.00538072e-02
-3.83368917e-02 -9.52806473e-01 4.70897138e-01 -5.19189537e-01
-1.16326571e+00 7.47795284e-01 4.51713987e-02 -3.09106648e-01
2.19148338e-01 -1.57728732e-01 -5.28123140e-01 2.02792548e-02
7.77862151e-04 -6.44156486e-02 4.94628027e-02 -6.04074657e-01
-6.37762606e-01 -9.13386226e-01 -3.64413470e-01 6.93301996e-03
-1.92875504e-01 6.08646087e-02 3.46535623e-01 -4.55460429e-01
2.72436023e-01 -9.59189534e-01 -3.37062031e-01 -2.99468637e-01
-3.59662801e-01 -9.70875472e-02 7.52819479e-01 -1.04658818e+00
1.44199204e+00 -1.94036782e+00 -1.66741088e-01 4.48938489e-01
7.11845830e-02 4.76353228e-01 7.91627884e-01 3.83928657e-01
-1.23575911e-01 5.64501524e-01 -6.33440167e-02 6.04323804e-01
-5.10241628e-01 2.78275490e-01 2.32424527e-01 1.69346780e-01
1.03690773e-01 2.83338606e-01 -3.95454884e-01 -7.21061945e-01
5.21455765e-01 4.78391379e-01 -2.84873009e-01 -3.92801166e-02
2.84067512e-01 5.02801239e-01 -7.37890899e-01 6.86781883e-01
3.97669703e-01 -3.13902318e-01 4.52091813e-01 -1.04865491e-01
-2.81428576e-01 -1.70614183e-01 -1.23631608e+00 9.06961679e-01
-5.36438227e-02 4.48930450e-02 -4.83054429e-01 -1.25348520e+00
1.31168425e+00 7.07829714e-01 5.99971354e-01 -3.89403999e-01
3.41563582e-01 4.01548326e-01 2.17305318e-01 -7.52002716e-01
-4.70761895e-01 -1.94976747e-01 3.97446424e-01 6.41823113e-02
-4.72963482e-01 5.60342133e-01 2.89491087e-01 -8.61758292e-02
8.68176341e-01 -9.28926393e-02 9.36805189e-01 -2.45991439e-01
8.41652572e-01 1.29341498e-01 7.51627743e-01 3.26853961e-01
-2.78343678e-01 2.92014748e-01 5.29138088e-01 -7.57391512e-01
-8.78602684e-01 -8.97268713e-01 -5.96614778e-01 3.74930322e-01
-2.74225980e-01 -9.12565142e-02 -2.99405605e-01 -3.91790748e-01
8.74178857e-02 3.04459512e-01 -2.06709355e-01 3.23789567e-02
-4.18788642e-01 -1.14857244e+00 3.12504917e-01 1.66417390e-01
9.48898017e-01 -7.96325803e-01 -9.33307111e-01 3.54392469e-01
9.52456892e-03 -3.66396636e-01 3.48910123e-01 -5.79734966e-02
-1.48628306e+00 -1.39212751e+00 -5.59820950e-01 -7.24395096e-01
2.53925264e-01 -2.45838910e-01 1.05628037e+00 4.11546230e-01
-7.28116453e-01 -2.73216724e-01 -4.90128368e-01 -6.40611172e-01
-3.66040528e-01 1.24141246e-01 -2.60862529e-01 -2.22551420e-01
2.58029044e-01 -5.90372622e-01 -1.02761102e+00 1.81808293e-01
-1.81775987e-01 -5.19042090e-02 9.81620371e-01 5.63768804e-01
4.40654129e-01 2.50883371e-01 8.77198875e-01 -1.03243327e+00
3.35221022e-01 -6.35275066e-01 -2.30599880e-01 3.22237164e-01
-1.18512547e+00 -3.01086515e-01 3.34407479e-01 1.18509494e-01
-6.79377258e-01 1.63724050e-02 5.07124588e-02 1.67367384e-01
-3.98671180e-01 5.19352257e-01 2.07787268e-02 3.25507045e-01
5.24312258e-01 8.35076794e-02 2.13375241e-01 -6.83839798e-01
-4.03775811e-01 9.48037446e-01 1.04206331e-01 -2.04624027e-01
2.12320223e-01 3.62255424e-02 4.89885390e-01 -7.73005784e-01
-6.08401716e-01 -5.91852188e-01 -4.12685156e-01 -3.20510060e-01
1.07640696e+00 -6.35158956e-01 -8.01260889e-01 3.15995634e-01
-6.68736696e-01 4.89915848e-01 3.41667742e-01 8.87687325e-01
-1.41748656e-02 2.87397236e-01 -3.17384809e-01 -1.06097114e+00
-8.13514411e-01 -6.29238248e-01 3.29475522e-01 4.40739959e-01
-2.92342156e-01 -9.79550481e-01 1.99819773e-01 4.19840336e-01
4.36263412e-01 8.41943324e-01 1.29889274e+00 -5.76419652e-01
-2.37723187e-01 -1.88267395e-01 -6.89717308e-02 3.85831326e-01
3.33807677e-01 2.73027629e-01 -3.06723386e-01 -1.44383148e-03
-1.65419325e-01 1.02022730e-01 5.82858503e-01 7.04280555e-01
8.03729177e-01 -2.39691630e-01 -5.24441123e-01 1.39059260e-01
1.54470992e+00 8.21052194e-01 6.46720171e-01 3.97985935e-01
2.34817818e-01 2.18228638e-01 4.82282162e-01 5.84708929e-01
3.70741576e-01 2.76274234e-01 2.63498992e-01 -2.03078464e-01
1.10685313e-02 2.38826960e-01 -1.55747741e-01 6.23304427e-01
-5.41431546e-01 -1.28431469e-02 -1.20892799e+00 2.16616973e-01
-1.59328425e+00 -1.25431418e+00 -6.74485028e-01 2.41031361e+00
5.89570045e-01 9.70432088e-02 3.03711712e-01 7.81738698e-01
6.84743464e-01 -5.68344533e-01 -1.63955033e-01 -4.97827202e-01
2.03578621e-01 3.29570740e-01 2.41646826e-01 3.92945595e-02
-1.21090317e+00 7.77945518e-02 5.57876778e+00 -1.03222042e-01
-1.29852295e+00 -1.65276960e-01 1.01351762e+00 3.20596814e-01
2.76473790e-01 1.97797954e-01 -5.24505138e-01 7.20694542e-01
8.48023295e-01 4.15138993e-03 3.50829214e-03 8.51741135e-01
3.44066828e-01 -2.53991544e-01 -4.97841954e-01 5.91230154e-01
-5.64740747e-02 -9.79248345e-01 -3.41939896e-01 5.61635941e-02
3.89063925e-01 -2.83836901e-01 -1.97937205e-01 1.48724347e-01
-7.07137883e-02 -8.21194708e-01 -7.84678012e-02 7.53984749e-01
5.35903037e-01 -5.30090153e-01 1.15660620e+00 3.66415977e-01
-7.24325895e-01 -1.84070170e-01 2.26638973e-01 -3.31317335e-01
7.25233555e-03 1.11231983e+00 -1.10654855e+00 4.98605311e-01
7.92985022e-01 5.64243793e-01 -4.75651324e-01 1.34604001e+00
2.32563570e-01 8.16590071e-01 -3.03403795e-01 -2.53160805e-01
-3.26547146e-01 -1.40979499e-01 2.56553501e-01 5.74151814e-01
6.30909622e-01 3.80721927e-01 2.81525552e-01 2.31483728e-01
4.97659326e-01 6.91582322e-01 -4.60109085e-01 1.65586889e-01
4.10790920e-01 1.09637129e+00 -8.10784996e-01 -5.08606672e-01
-3.02578896e-01 1.34427607e-01 -3.84628624e-01 -1.26433328e-01
-7.81448603e-01 -2.83499479e-01 6.07997328e-02 5.12358010e-01
-1.60219356e-01 3.21377695e-01 -6.87792003e-01 -8.93967927e-01
-2.19571128e-01 -7.50070870e-01 7.79453337e-01 -2.19283417e-01
-1.04160273e+00 2.73712963e-01 -6.19276240e-02 -1.07049024e+00
-1.06786303e-01 -4.54426348e-01 -5.40353596e-01 9.65370595e-01
-1.10039163e+00 -8.35899591e-01 -5.13906181e-01 1.67071164e-01
2.37747222e-01 -4.04336721e-01 1.07242155e+00 5.14412642e-01
-8.64882708e-01 2.23504957e-02 -7.43695796e-02 1.33005306e-01
6.12706006e-01 -1.16336644e+00 -5.34896255e-01 3.01124245e-01
-5.94019592e-01 5.35092354e-01 4.92973655e-01 -9.13444936e-01
-6.84577644e-01 -7.83911049e-01 1.07445323e+00 3.09738256e-02
1.08002340e-02 5.29359937e-01 -4.51843798e-01 2.26202965e-01
-8.92226994e-02 -7.45328665e-02 9.71733510e-01 2.09362328e-01
2.41911575e-01 -4.22430158e-01 -1.38930106e+00 9.39347446e-02
3.36947799e-01 2.74021775e-01 -3.13305587e-01 2.80907899e-01
-2.04651535e-01 -1.08824544e-01 -1.22901404e+00 8.12716186e-01
9.80442286e-01 -1.47154081e+00 9.35222208e-01 -6.09245300e-01
3.57249111e-01 -1.86491802e-01 4.33471426e-02 -8.30584407e-01
-2.66555250e-01 -2.96797715e-02 1.95471287e-01 1.26737714e+00
6.11103892e-01 -7.33074248e-01 7.57962346e-01 6.27649546e-01
1.52326107e-01 -1.03583252e+00 -5.99994659e-01 -2.55528033e-01
-1.14572853e-01 1.62839711e-01 3.67746770e-01 1.00916195e+00
-9.32969525e-02 4.71884251e-01 -3.02796245e-01 -1.03832833e-01
6.58994913e-01 3.03438932e-01 5.84548235e-01 -1.85289717e+00
-2.13542476e-01 -8.32384005e-02 -5.60360074e-01 1.34248570e-01
-7.03255892e-01 -5.50352097e-01 -8.07724953e-01 -1.83636427e+00
5.22202373e-01 -8.39848816e-01 -4.90372121e-01 4.68407989e-01
-2.56300509e-01 -3.03422883e-02 -9.98157263e-02 3.09911400e-01
1.50188535e-01 -2.51439244e-01 1.24491107e+00 2.48186052e-01
-4.19592440e-01 3.30234975e-01 -6.22970700e-01 7.03967810e-01
1.44147742e+00 -5.17748296e-01 -5.05193472e-01 -3.77865285e-02
-4.43372875e-02 5.28133094e-01 4.17192191e-01 -1.09808922e+00
-3.37606579e-01 -4.35356766e-01 8.67095470e-01 -5.48196912e-01
-3.48202735e-02 -8.86376143e-01 7.05476761e-01 1.24140990e+00
-9.23779830e-02 7.59344995e-02 -3.37923497e-01 1.81784764e-01
5.65679222e-02 -3.72178555e-02 6.64961457e-01 -3.71908307e-01
-4.16150153e-01 -3.35886069e-02 -2.45544225e-01 -1.42843604e-01
1.53037214e+00 -1.39538437e-01 -5.10291517e-01 2.77626161e-02
-1.03043890e+00 2.31993124e-02 5.14025204e-02 1.07218049e-01
3.87676269e-01 -9.31993783e-01 -8.49287629e-01 8.66544247e-02
-1.12951726e-01 -5.09914696e-01 2.26419240e-01 1.30820787e+00
-9.35893357e-01 4.84958380e-01 -5.30069649e-01 -4.50792730e-01
-1.64970875e+00 2.67887533e-01 1.22610621e-01 -3.80668133e-01
-5.56092620e-01 8.95465240e-02 -5.20908475e-01 -4.99233082e-02
-5.27704731e-02 -2.18156397e-01 -7.14788675e-01 -6.92560151e-02
2.00700015e-01 9.39836740e-01 -3.60513665e-02 -2.38426745e-01
-5.77436566e-01 8.01926613e-01 1.34279892e-01 3.19813788e-01
1.02721179e+00 2.49742102e-02 -3.45897168e-01 3.67796481e-01
7.71581948e-01 -1.85437892e-02 -6.63906708e-02 3.77584130e-01
1.96638275e-02 -3.21239680e-01 1.48383304e-01 -1.36747575e+00
-9.51875687e-01 6.89963520e-01 1.24436522e+00 4.17428195e-01
1.37480342e+00 -1.77315488e-01 6.10453367e-01 -3.66707295e-02
1.43213332e-01 -7.93357730e-01 -3.42261642e-01 -1.18822128e-01
4.32169795e-01 -1.31442833e+00 2.26743832e-01 -4.91585582e-01
-4.75870758e-01 1.22963226e+00 4.22496349e-01 -5.55675365e-02
1.01006532e+00 8.64240304e-02 3.47865731e-01 8.01132247e-02
-7.84793317e-01 -7.92375952e-02 -4.28167246e-02 4.68597412e-01
8.16450834e-01 3.00788522e-01 -1.10935295e+00 5.57186246e-01
-4.82394621e-02 6.67225182e-01 2.20009223e-01 8.79736900e-01
-8.12957644e-01 -1.16216791e+00 -5.09790540e-01 9.75207448e-01
-9.52171326e-01 1.16788216e-01 -3.29693109e-01 8.73309672e-01
5.13078511e-01 9.16854203e-01 -2.27544606e-01 -2.34231785e-01
2.75781095e-01 3.20998788e-01 1.77043095e-01 -2.80335665e-01
-4.52957034e-01 -1.27700031e-01 3.04622322e-01 1.53417885e-01
-6.73567057e-01 -6.39925838e-01 -1.20472467e+00 -3.12782943e-01
-2.75752664e-01 2.96762288e-01 8.83047819e-01 8.45243633e-01
3.16418111e-01 5.28658450e-01 6.86555445e-01 1.27774268e-01
-3.23317200e-01 -1.05997241e+00 -5.26245952e-01 2.43650541e-01
-6.16382025e-02 -6.58914626e-01 -2.46338576e-01 1.47813529e-01] | [8.462784767150879, 4.89336633682251] |
9ad77b45-eeb3-4b9a-9af7-7a91bfc8aaeb | accounting-for-the-neglected-dimensions-of-ai | 1806.00610 | null | https://arxiv.org/abs/1806.00610v2 | https://arxiv.org/pdf/1806.00610v2.pdf | Between Progress and Potential Impact of AI: the Neglected Dimensions | We reframe the analysis of progress in AI by incorporating into an overall framework both the task performance of a system, and the time and resource costs incurred in the development and deployment of the system. These costs include: data, expert knowledge, human oversight, software resources, computing cycles, hardware and network facilities, and (what kind of) time. These costs are distributed over the life cycle of the system, and may place differing demands on different developers and users. The multidimensional performance and cost space we present can be collapsed to a single utility metric that measures the value of the system for different stakeholders. Even without a single utility function, AI advances can be generically assessed by whether they expand the Pareto surface. We label these types of costs as neglected dimensions of AI progress, and explore them using four case studies: Alpha* (Go, Chess, and other board games), ALE (Atari games), ImageNet (Image classification) and Virtual Personal Assistants (Siri, Alexa, Cortana, and Google Assistant). This broader model of progress in AI will lead to novel ways of estimating the potential societal use and impact of an AI system, and the establishment of milestones for future progress. | ['José Hernández-Orallo', 'Sean Ó hÉigeartaigh', 'Fernando Martínez-Plumed', 'Allan Dafoe', 'Miles Brundage', 'Shahar Avin'] | 2018-06-02 | null | null | null | null | ['board-games'] | ['playing-games'] | [-6.75285906e-02 1.12205923e-01 4.24674004e-02 -6.52941614e-02
-3.63626361e-01 -8.92673016e-01 7.96169877e-01 2.28384510e-01
-6.50846183e-01 3.41039389e-01 2.14655280e-01 -5.38637519e-01
-4.63552237e-01 -6.60792172e-01 -3.13336968e-01 -3.01483497e-02
-8.51080418e-02 4.80496496e-01 3.58588547e-02 -2.71667063e-01
4.03220296e-01 3.50553632e-01 -1.46111917e+00 -3.07363331e-01
6.95775449e-01 9.80036020e-01 1.38644176e-02 6.59680426e-01
2.60650486e-01 7.70701885e-01 -9.09672141e-01 -4.35279936e-01
3.80008698e-01 3.18799727e-02 -4.81601715e-01 -4.94723022e-02
1.93299666e-01 -5.29619753e-01 9.58779186e-04 7.77814448e-01
4.68218625e-01 -3.07018399e-01 5.44166923e-01 -1.69766390e+00
-4.66705263e-01 6.37528896e-01 -4.03895259e-01 3.38021487e-01
3.00056815e-01 4.92690533e-01 1.02899492e+00 -3.74931276e-01
6.36115015e-01 9.06378627e-01 5.20286679e-01 -2.43202224e-01
-1.02018034e+00 -6.33434296e-01 7.25154206e-02 -5.39020225e-02
-1.21437979e+00 -6.67492628e-01 1.76999167e-01 -1.00095475e+00
9.33794558e-01 2.44704217e-01 9.79336977e-01 4.06108171e-01
2.49415234e-01 5.45253456e-01 6.89945161e-01 -2.94464946e-01
4.84060824e-01 1.27800718e-01 1.31615296e-01 3.87666345e-01
5.69478095e-01 -1.95311472e-01 -1.56524092e-01 -2.74372488e-01
5.78433514e-01 -1.10817730e-01 3.81943621e-02 -4.87376183e-01
-1.19315255e+00 2.42394879e-01 6.42065108e-02 5.44785559e-01
-5.69232821e-01 4.22237307e-01 1.75622836e-01 3.70286196e-01
2.14140847e-01 1.11820066e+00 -3.55079949e-01 -6.57508671e-01
-8.88892889e-01 3.09971392e-01 9.30686533e-01 7.99051821e-01
6.69765353e-01 1.71052337e-01 -4.78815637e-04 2.71653205e-01
-9.85543206e-02 2.57160068e-01 1.79159015e-01 -1.41150904e+00
3.43626857e-01 7.57188380e-01 4.96733636e-01 -9.26453948e-01
-5.98086834e-01 -6.36533439e-01 -6.81073964e-02 2.99257278e-01
3.62019837e-01 -5.88417172e-01 -6.56380713e-01 1.38341522e+00
-1.55718029e-01 -2.93193758e-01 -2.93871671e-01 7.03868747e-01
3.13897729e-01 5.24757743e-01 2.34235480e-01 1.10104531e-01
8.57227087e-01 -7.91381598e-01 -3.23232651e-01 -4.25709397e-01
5.95642447e-01 -3.06895226e-01 9.06011999e-01 4.63639170e-01
-1.61093128e+00 -2.90576249e-01 -1.17113996e+00 3.54589462e-01
-3.41899067e-01 4.80330642e-03 7.20236599e-01 7.14287162e-01
-1.22081590e+00 4.66366619e-01 -8.14250946e-01 -3.49600554e-01
1.73422813e-01 4.10989314e-01 1.02038302e-01 6.81208968e-02
-7.32735753e-01 1.29577935e+00 -1.40330344e-01 -1.40845761e-01
-7.97177970e-01 -6.61455333e-01 -3.48224550e-01 3.74257624e-01
4.35804069e-01 -7.40550101e-01 1.23061502e+00 -1.09375405e+00
-9.46210504e-01 3.38688880e-01 7.60053217e-01 -1.20986804e-01
5.08072495e-01 1.09511420e-01 -3.63834023e-01 -3.46953660e-01
2.50569969e-01 3.16546589e-01 3.81584048e-01 -8.07074487e-01
-9.82829690e-01 -5.97009957e-01 4.64517295e-01 4.93005574e-01
-4.82214689e-01 3.60623330e-01 -2.44457051e-01 -2.49259248e-01
-4.70803082e-01 -1.04937506e+00 -2.20069110e-01 -2.26058573e-01
2.77245551e-01 -7.89680928e-02 1.59345910e-01 -6.81275427e-01
1.42021275e+00 -2.04359365e+00 4.20717955e-01 2.09883258e-01
4.21770751e-01 1.60612687e-01 1.94770601e-02 4.10117269e-01
1.95652395e-01 3.57658684e-01 9.07739252e-02 4.13967252e-01
5.31349033e-02 -3.10165524e-01 4.38634515e-01 3.38700563e-01
3.78888287e-02 7.57173002e-01 -6.59365773e-01 -8.61888081e-02
1.44763172e-01 1.24196485e-01 -3.89491260e-01 -4.14986551e-01
2.14022286e-02 -4.10545506e-02 -5.12190640e-01 5.89868248e-01
8.68006144e-03 -3.78683507e-01 2.37693042e-01 1.62975714e-01
-5.39607108e-01 -6.34584352e-02 -1.01223433e+00 1.25282931e+00
-4.29757029e-01 6.67547643e-01 3.22607577e-01 -7.16318786e-01
6.06439233e-01 1.68039635e-01 5.83025634e-01 -1.02410102e+00
2.48176426e-01 1.74389094e-01 6.13013685e-01 -4.27395225e-01
5.64307511e-01 4.48308021e-01 -2.70046532e-01 7.63449132e-01
-1.65422723e-01 -2.92157471e-01 1.81396872e-01 1.00209855e-01
1.55800998e+00 -8.04108903e-02 2.48791113e-01 -4.58468288e-01
-1.27828255e-01 3.80175143e-01 2.40090653e-01 4.14779752e-01
-4.24528033e-01 -1.58185810e-01 5.31729817e-01 -3.68702292e-01
-1.41099298e+00 -8.39095831e-01 6.85561374e-02 1.32781601e+00
1.89532489e-01 -4.54491019e-01 -6.96132660e-01 -1.84672475e-01
2.83820659e-01 7.43040502e-01 -3.66004288e-01 -1.30582765e-01
-2.47393146e-01 -3.91931772e-01 5.45155942e-01 4.58480924e-01
3.89111578e-01 -8.20091069e-01 -1.22542858e+00 1.42343327e-01
1.79155022e-01 -7.66252458e-01 -1.91953272e-01 -3.08615863e-01
-6.01143599e-01 -8.06270242e-01 -7.31804729e-01 -1.28641024e-01
2.64555961e-01 4.63285707e-02 1.18113613e+00 1.09306872e-01
-5.21708071e-01 9.17890906e-01 -8.16945210e-02 -7.23403811e-01
-5.46782576e-02 -1.57907426e-01 1.41662627e-01 -6.06718004e-01
7.93615133e-02 -3.28116506e-01 -5.55062711e-01 3.61046016e-01
-4.09728974e-01 -7.60786096e-03 7.02343225e-01 3.04267317e-01
1.46103963e-01 3.17397892e-01 6.11686468e-01 -4.20063823e-01
1.16291726e+00 -7.75971651e-01 -5.63551128e-01 5.02370596e-01
-9.67771113e-01 -2.49921367e-01 2.31918588e-01 -2.49737188e-01
-8.42378199e-01 -1.42958194e-01 6.05868578e-01 5.45172691e-02
2.31314451e-01 9.22514558e-01 8.87928456e-02 -1.14404619e-01
8.06550384e-01 -3.78562599e-01 7.47165382e-02 -6.82490095e-02
2.91413695e-01 7.60870516e-01 4.60383952e-01 -6.16931200e-01
4.02156502e-01 1.29341289e-01 -2.99878836e-01 -8.08159053e-01
1.59158185e-02 8.99081230e-02 -1.59347162e-01 -5.72273254e-01
6.11060381e-01 -8.44982028e-01 -7.76842475e-01 1.60208181e-01
-6.42292440e-01 -3.83816987e-01 -2.70751655e-01 5.21004438e-01
-4.05547053e-01 -1.34009629e-01 -1.59526676e-01 -8.10638309e-01
-2.32479617e-01 -1.26990974e+00 3.41913402e-01 2.91501999e-01
-5.89840412e-01 -5.33364058e-01 -9.15225074e-02 5.78218460e-01
8.05400133e-01 2.89295137e-01 7.97985852e-01 -2.76249677e-01
-5.95893502e-01 -5.04053295e-01 -4.59781557e-01 1.81998461e-01
-1.51561707e-01 2.84139663e-01 -2.86044955e-01 -2.79276878e-01
-1.98513091e-01 -1.26722693e-01 9.87858772e-02 5.06143689e-01
7.64092386e-01 -3.24616164e-01 -3.98428112e-01 8.24925080e-02
1.22894728e+00 6.85674489e-01 4.02763128e-01 5.83057344e-01
4.22011465e-01 9.04439807e-01 7.80803204e-01 5.73310912e-01
6.40091121e-01 5.50001979e-01 2.64819384e-01 8.26850682e-02
2.79704690e-01 2.02900618e-01 3.12213361e-01 4.44451094e-01
-6.34029269e-01 -2.60048956e-01 -1.58897519e+00 6.45691752e-01
-1.87924445e+00 -6.41876340e-01 2.90971130e-01 2.32669377e+00
1.64116949e-01 3.53729993e-01 5.89746296e-01 -5.18556833e-02
6.63295388e-01 -3.23260754e-01 -9.37399149e-01 -4.97334152e-01
4.75520074e-01 -1.11239009e-01 7.78579354e-01 1.12692110e-01
-3.75914663e-01 6.03174269e-01 7.81591511e+00 5.21968484e-01
-9.44537520e-01 -8.87672678e-02 7.55119324e-01 -3.49600255e-01
-3.41200978e-01 1.61708921e-01 -1.55715451e-01 4.53292191e-01
1.30319011e+00 -9.62099493e-01 7.31029510e-01 9.09619629e-01
3.10587902e-02 -2.83118784e-01 -1.11278951e+00 6.31454587e-01
-1.41428083e-01 -1.24275339e+00 -5.17366707e-01 4.05609876e-01
3.85778427e-01 3.23157042e-01 1.52503029e-01 2.00357065e-01
9.00727749e-01 -1.00521183e+00 9.92471993e-01 4.27717716e-01
9.13745940e-01 -8.21350336e-01 3.78138393e-01 2.83628345e-01
-8.76819849e-01 -5.50404608e-01 2.00575456e-01 -7.06154764e-01
-1.42575279e-01 2.37228419e-03 -4.53227282e-01 8.34716484e-02
8.09145272e-01 1.13821976e-01 -6.26767576e-01 7.00522840e-01
4.49676633e-01 3.21067989e-01 -3.37469310e-01 -2.22648636e-01
2.33271316e-01 -1.07329190e-01 3.77624571e-01 6.91403091e-01
2.57422000e-01 2.79699415e-01 2.07118224e-02 8.14482212e-01
2.09074810e-01 -2.19509259e-01 -4.13821071e-01 -8.04368913e-01
7.82720745e-01 1.30573130e+00 -7.19939470e-01 -1.94227263e-01
-4.65629369e-01 2.83067405e-01 -3.39663476e-02 2.20767558e-01
-6.35257125e-01 -4.65806395e-01 7.98925221e-01 5.12850046e-01
-2.87494838e-01 -3.38326603e-01 -4.51085001e-01 -5.88968992e-01
8.68004467e-03 -8.50165308e-01 1.05555899e-01 -7.49574244e-01
-6.59656346e-01 3.00209343e-01 1.81723878e-01 -7.01042056e-01
-2.97244668e-01 -3.42553526e-01 -4.44979578e-01 6.51985347e-01
-8.17680538e-01 -6.95322752e-01 -4.13388968e-01 1.97742090e-01
-2.19810139e-02 -4.69157934e-01 4.11311656e-01 2.89735526e-01
-6.00873709e-01 4.34766531e-01 1.36728957e-01 -3.73593628e-01
4.11266953e-01 -9.08135831e-01 6.48679793e-01 4.62718338e-01
-3.91339302e-01 4.15252775e-01 4.99044478e-01 -6.68365121e-01
-1.45348573e+00 -4.27487671e-01 3.77621502e-01 -6.64724350e-01
8.15112770e-01 2.10371409e-02 -1.34901255e-01 7.01647341e-01
9.79382023e-02 -6.58632457e-01 5.12807667e-01 4.27545130e-01
7.24861547e-02 -6.47982806e-02 -1.17355764e+00 7.49186337e-01
9.90885556e-01 -1.68225259e-01 -1.87783331e-01 -9.30823311e-02
6.32320881e-01 1.42353058e-01 -1.07775712e+00 1.61641032e-01
9.82056141e-01 -1.02456820e+00 7.38178074e-01 -3.90698969e-01
4.41194534e-01 -3.81327532e-02 -8.30022544e-02 -1.10656655e+00
-6.30033851e-01 -5.33652604e-01 3.98341566e-01 9.77485836e-01
5.36357164e-01 -5.68874717e-01 6.26120389e-01 1.61032403e+00
-1.63345307e-01 -7.59149253e-01 -6.15453660e-01 -6.05126977e-01
1.06023960e-02 -4.19172674e-01 9.76807237e-01 1.03961456e+00
4.33924943e-01 2.70138443e-01 1.75939411e-01 -3.38296950e-01
4.56593871e-01 -2.07997695e-01 6.84715986e-01 -1.46208549e+00
-2.45331779e-01 -8.67157996e-01 -8.85578156e-01 -2.39187822e-01
-4.14418370e-01 -4.68261480e-01 -3.68780911e-01 -1.74253726e+00
6.01530224e-02 -7.35960782e-01 -8.04322585e-02 4.79632169e-01
1.78228721e-01 -2.60418296e-01 5.61095059e-01 6.07147992e-01
-4.69730169e-01 3.83169092e-02 7.83395708e-01 3.59988250e-02
-2.54491389e-01 -1.32112131e-01 -1.14739633e+00 9.10569131e-01
5.23602426e-01 -5.65718277e-04 -2.81274378e-01 -7.51640916e-01
6.23272359e-01 3.80010903e-01 2.20426604e-01 -1.35526550e+00
4.51108932e-01 -4.65034574e-01 1.68439314e-01 1.65247053e-01
3.30492526e-01 -1.09580719e+00 5.68359077e-01 5.54250896e-01
-3.48503977e-01 3.29356760e-01 1.65908076e-02 1.51903465e-01
1.36954218e-01 -6.59182593e-02 3.15587372e-01 -1.77771538e-01
-5.57151675e-01 3.08672916e-02 -4.65333909e-01 -5.00282869e-02
1.43692541e+00 -4.94488984e-01 -5.81996143e-01 -5.07548869e-01
-6.20935082e-01 6.34756684e-01 7.26224720e-01 4.08025503e-01
-4.14364971e-02 -8.56007218e-01 -6.07973397e-01 -2.90457368e-01
7.41155297e-02 -3.78914535e-01 -6.95358356e-03 3.58929425e-01
-6.24926209e-01 2.17758179e-01 -8.14747453e-01 -1.04638703e-01
-9.60332215e-01 2.90722430e-01 1.06253155e-01 7.36561790e-02
-1.92463264e-01 5.42558312e-01 1.99442893e-01 2.46171989e-02
2.65570641e-01 -5.48341200e-02 -1.18414901e-01 3.58547181e-01
2.68630117e-01 9.09596503e-01 -1.96430683e-01 -3.20791662e-01
-3.47841561e-01 2.48006359e-01 9.96496249e-03 -5.55821359e-01
1.34306133e+00 -1.89639956e-01 -6.79311603e-02 4.52379048e-01
3.37851495e-01 -2.33625352e-01 -1.03441763e+00 1.21336997e-01
4.03295346e-02 -2.97143161e-01 2.67507404e-01 -1.06135130e+00
-8.97750735e-01 4.35236961e-01 5.04992187e-01 4.27903295e-01
1.08492708e+00 -1.31039292e-01 2.56548524e-01 2.26542532e-01
5.91314852e-01 -1.56288397e+00 -7.12337941e-02 -9.00864527e-02
7.59784222e-01 -5.90274692e-01 2.58561969e-01 9.21044499e-02
-8.95594001e-01 8.13432038e-01 7.00933337e-01 8.60306434e-03
7.15621412e-01 4.41035062e-01 -3.64852816e-01 -2.52798140e-01
-6.59219325e-01 -7.81414881e-02 -1.17188945e-01 3.89931411e-01
3.74672979e-01 4.83173013e-01 -4.63623971e-01 7.70715415e-01
-2.46502459e-01 2.73119181e-01 8.69235754e-01 1.05680954e+00
-5.35172641e-01 -5.81809938e-01 -7.50409588e-02 8.79586637e-01
-3.30372214e-01 5.95039353e-02 -6.38108850e-01 8.61270010e-01
1.99274316e-01 8.34879100e-01 3.40506941e-01 -7.33107686e-01
5.19646525e-01 -2.21773118e-01 3.27676356e-01 -5.37351012e-01
-7.02087462e-01 -1.49997413e-01 4.42758739e-01 -4.75844145e-01
-5.06986380e-02 -8.05264533e-01 -9.32584226e-01 -5.47042012e-01
-3.59879375e-01 5.97628951e-03 1.19464135e+00 6.78534210e-01
7.72866070e-01 5.90183496e-01 2.59500444e-01 -6.19530737e-01
-4.35831189e-01 -7.40282774e-01 -5.99957526e-01 -7.61035830e-02
-1.29440844e-01 -6.68324351e-01 -2.85152435e-01 -3.02047104e-01] | [8.981947898864746, 6.275793552398682] |
44bae55f-b9e3-4cd4-a994-3472b174c130 | spatio-temporal-forecasting-with-gridded | null | null | https://dl.acm.org/doi/abs/10.1145/3397536.3422247 | https://dl.acm.org/doi/abs/10.1145/3397536.3422247 | Spatio-Temporal Forecasting With Gridded Remote Sensing Data Using Feed-Backward Decoding | We present a novel deep learning approach for spatio-temporal forecasting with remote sensing data, extending a previous model named Spatio-Temporal Convolutional Sequence to Sequence Network (STConvS2S) in several directions. Experiments using datasets from previous studies show that the proposed approaches outperform the original STConvS2S and other baseline models on tasks related to predicting future time-steps. In tests related to predicting a missing time-step, some of the proposed extensions also lead to improvements over the original STConvS2S architecture, although simpler models seem to be beneficial in this scenario. | ['Bruno Martins', 'Jacinto Estima', 'Mário Cardoso'] | 2020-11-01 | null | null | null | null | ['spatio-temporal-forecasting'] | ['time-series'] | [ 2.49729753e-01 -3.36991936e-01 -1.20331727e-01 -7.29441345e-01
-5.05025506e-01 -4.00885731e-01 1.13019693e+00 -2.68572271e-01
-2.69294739e-01 1.01240444e+00 6.49535358e-01 -8.61940861e-01
-2.11532071e-01 -9.18940485e-01 -7.14442849e-01 -6.20313704e-01
-7.17625260e-01 1.85448304e-02 3.53951871e-01 -6.76210463e-01
3.15051079e-01 6.57172441e-01 -1.41656935e+00 8.34346890e-01
5.44194579e-01 1.04292655e+00 3.04210544e-01 8.10930312e-01
-2.00713217e-01 1.18266988e+00 -2.90195674e-01 2.08712846e-01
4.82781202e-01 -1.68545246e-01 -9.71694827e-01 -4.35988575e-01
2.50180453e-01 -5.99280715e-01 -2.85012424e-01 3.68250847e-01
2.29148179e-01 4.64061975e-01 6.46386445e-01 -1.05608940e+00
-7.57467806e-01 6.52351856e-01 -3.50814104e-01 5.58722854e-01
8.82271305e-02 2.39357457e-01 6.93504870e-01 -6.80426836e-01
4.65037912e-01 1.54918873e+00 1.38079333e+00 2.37020448e-01
-8.57242763e-01 -6.08701646e-01 4.95620847e-01 3.01137775e-01
-1.04712665e+00 -2.09709004e-01 2.83674359e-01 -6.34326518e-01
1.67395067e+00 3.46272260e-01 4.30451214e-01 1.39643538e+00
3.49858791e-01 9.29722250e-01 1.10428190e+00 6.55790046e-02
1.79995611e-01 -5.01086771e-01 -4.75536622e-02 7.84751847e-02
-4.56936449e-01 5.27697980e-01 -4.82247360e-02 -1.30702242e-01
6.34172618e-01 4.07254398e-01 7.48184770e-02 5.23871779e-01
-1.45670497e+00 8.45138609e-01 8.30010653e-01 8.45194459e-01
-9.25903440e-01 3.94698709e-01 6.54820621e-01 3.62653852e-01
1.45933056e+00 1.50362864e-01 -8.71267676e-01 6.90323114e-02
-1.50436711e+00 4.27785993e-01 3.94400358e-01 4.49727982e-01
5.88307559e-01 7.02323437e-01 -4.69791621e-01 3.02733123e-01
-9.71201211e-02 5.48604131e-01 4.81015593e-01 -9.12667453e-01
5.49446285e-01 2.77401894e-01 5.65737009e-01 -6.50812030e-01
-8.16614568e-01 -2.35969514e-01 -1.18575168e+00 8.22926238e-02
-7.02646449e-02 -4.36783075e-01 -1.29321027e+00 1.29461515e+00
-1.09224163e-01 8.67179036e-01 7.39994198e-02 8.12107205e-01
7.75855958e-01 1.38731003e+00 4.07937407e-01 1.93056073e-02
6.06468678e-01 -1.01139009e+00 -6.82543933e-01 -2.19745368e-01
7.51466811e-01 -4.01877910e-01 7.58781016e-01 -4.74833744e-03
-5.38841903e-01 -6.94449544e-01 -5.43785691e-01 2.57576525e-01
-9.29584980e-01 2.57379085e-01 7.81917810e-01 3.09045948e-02
-1.56349826e+00 1.09525335e+00 -1.09685397e+00 -6.26423776e-01
1.19438052e-01 -1.11474715e-01 -4.78969961e-02 3.00604522e-01
-1.66816127e+00 9.65165377e-01 6.23396456e-01 5.79755127e-01
-1.05695879e+00 -8.73166323e-01 -8.47275972e-01 1.37293786e-01
-1.21744841e-01 -3.65677983e-01 1.46738088e+00 -8.93546164e-01
-9.68916595e-01 2.80213714e-01 -4.32807207e-01 -9.76163030e-01
6.04213893e-01 -1.41577795e-01 -7.24201620e-01 -2.90217936e-01
3.38891894e-01 6.74867630e-01 4.25970495e-01 -7.12024391e-01
-7.92193770e-01 -2.26393808e-02 -4.65292446e-02 -1.02770485e-01
-8.63016844e-02 1.92821443e-01 5.03590524e-01 -8.51837337e-01
-3.49649549e-01 -7.75908649e-01 -8.89724970e-01 -2.93253958e-01
-4.66061570e-02 -6.20184660e-01 9.02556777e-01 -1.00298786e+00
1.05720425e+00 -1.74136651e+00 -3.27248067e-01 -3.12334627e-01
-1.74502224e-01 8.30401123e-01 -5.60400367e-01 1.07209539e+00
-2.35547915e-01 3.95730674e-01 -3.83430213e-01 -2.74100631e-01
-2.28537723e-01 3.43062162e-01 -8.81998897e-01 3.08103800e-01
2.37280473e-01 1.04210055e+00 -9.17503417e-01 8.53694677e-02
5.59322298e-01 6.25227988e-01 2.30586722e-01 -1.60808399e-01
-6.47915661e-01 6.12554550e-01 -4.08989608e-01 4.33161229e-01
7.54429698e-01 -4.81389791e-01 3.55707877e-03 4.84632581e-01
-4.96167332e-01 5.34837067e-01 -5.89735866e-01 1.16996622e+00
-6.48975313e-01 8.81499290e-01 -4.65075165e-01 -1.08723426e+00
7.85675824e-01 6.32949591e-01 6.27859235e-01 -9.80574489e-01
-5.82392633e-01 2.83788890e-01 -4.08432007e-01 -6.06436431e-01
6.93388462e-01 -1.08852737e-01 2.46341601e-01 3.74755025e-01
-2.64994264e-01 4.31319803e-01 -2.49992553e-02 -4.86738756e-02
8.18028092e-01 6.17619455e-01 3.08943391e-01 -1.59276307e-01
5.29842436e-01 3.35309029e-01 4.12962168e-01 8.42754245e-01
-2.40213871e-01 3.35535437e-01 4.57041198e-03 -1.40054405e+00
-1.22658968e+00 -6.36854649e-01 1.91538900e-01 1.36656928e+00
-5.23721218e-01 -6.13230765e-02 -9.51536894e-02 -7.39564836e-01
-1.32156596e-01 1.01375699e+00 -8.73277545e-01 6.30980194e-01
-6.63696826e-01 -6.98035240e-01 8.39782476e-01 9.99904156e-01
3.81590217e-01 -1.37116396e+00 -6.83693409e-01 6.38078392e-01
-5.23761928e-01 -9.31485057e-01 2.29156278e-02 -4.49899212e-02
-1.08620596e+00 -6.31581843e-01 -1.04034102e+00 -4.07412678e-01
-2.92660445e-01 6.61327064e-01 1.18782258e+00 -1.72552645e-01
1.60004124e-01 -1.67884558e-01 -5.54309428e-01 -5.75230896e-01
-2.13218063e-01 3.20674896e-01 -2.52671599e-01 -1.30033374e-01
5.24955332e-01 -6.75545633e-01 -8.33299160e-01 4.77534384e-02
-9.86932755e-01 8.73563252e-03 1.57744735e-01 5.60167789e-01
1.10679552e-01 -1.59057066e-01 8.29462290e-01 -4.67392743e-01
4.77198362e-01 -9.70752656e-01 -6.16587818e-01 2.53276736e-01
-4.90649700e-01 -9.99209434e-02 8.89519989e-01 -1.97921202e-01
-1.13194883e+00 1.57909766e-01 -5.15719175e-01 -4.42530543e-01
-5.27586460e-01 9.87572193e-01 9.77646589e-01 3.36291909e-01
5.70612431e-01 8.17288280e-01 -3.69164109e-01 -7.41754234e-01
2.32226580e-01 5.14222860e-01 8.39554965e-02 2.68983245e-01
4.36563700e-01 8.00590992e-01 -9.32486802e-02 -8.51135790e-01
-9.66340721e-01 -6.42691851e-01 -7.78422296e-01 -1.50095657e-01
8.53246033e-01 -1.25753832e+00 -6.11400127e-01 5.67001104e-01
-1.37310636e+00 -7.94240236e-01 4.38888557e-02 5.71967244e-01
-2.96102405e-01 2.36378983e-01 -5.73616922e-01 -1.24252629e+00
-4.36838984e-01 -6.23382568e-01 1.55741286e+00 -4.05059338e-01
-2.29455940e-02 -1.56489050e+00 3.85183364e-01 -1.83378950e-01
1.17017734e+00 6.05087161e-01 3.81310076e-01 -6.66824460e-01
-1.91078022e-01 8.24592710e-02 -2.60258764e-01 1.72943294e-01
1.06546953e-01 -1.12449981e-01 -8.67728889e-01 -1.18400805e-01
-2.27371961e-01 -2.21996322e-01 1.58357382e+00 7.48131335e-01
1.20900261e+00 -7.09325910e-01 -3.78427088e-01 7.26733506e-01
1.48447633e+00 3.36860806e-01 7.26705432e-01 6.16295397e-01
5.16509652e-01 6.16914272e-01 6.83839321e-01 5.34412265e-01
6.81478083e-01 4.21636850e-01 7.28907108e-01 -4.89071250e-01
1.89900592e-01 -1.73123747e-01 6.76573217e-01 2.43723422e-01
-4.35213149e-01 -5.23275316e-01 -1.28537667e+00 1.11080956e+00
-2.21868372e+00 -1.53442371e+00 -7.65977323e-01 1.48576140e+00
1.86051756e-01 -5.21348238e-01 1.69449449e-01 -1.91271938e-02
3.93907130e-01 9.29548502e-01 -6.71782374e-01 -4.81889188e-01
-3.49602640e-01 1.56014353e-01 7.85689533e-01 5.79054117e-01
-1.49700415e+00 1.35514164e+00 7.67292881e+00 7.31720269e-01
-1.61483347e+00 2.59821087e-01 6.65495753e-01 4.33594286e-02
-3.44037473e-01 -1.51051879e-01 -6.91556990e-01 3.60257953e-01
1.54770780e+00 2.21335977e-01 2.26213902e-01 6.36537969e-01
8.79311085e-01 1.93923041e-01 -6.36700213e-01 2.63165295e-01
-4.14005697e-01 -1.75374901e+00 1.64481103e-01 -6.58598468e-02
9.42199290e-01 9.34126318e-01 1.88616645e-02 4.91393119e-01
6.96611047e-01 -1.21832871e+00 5.18929183e-01 6.12560153e-01
5.97285450e-01 -3.36017907e-01 1.09519911e+00 5.20595729e-01
-1.53477085e+00 -1.27324119e-01 -4.46578979e-01 -6.01666689e-01
4.81923878e-01 4.68482703e-01 -1.05483305e+00 8.28327119e-01
9.40896451e-01 1.36864531e+00 -4.02195185e-01 7.42281437e-01
-2.65295478e-03 1.17134964e+00 -1.65356174e-01 1.59445673e-01
1.28311169e+00 6.86282963e-02 3.75257820e-01 1.47826731e+00
7.68386841e-01 3.43956649e-02 1.94313720e-01 5.50102293e-01
4.75081086e-01 -1.34590223e-01 -1.06809342e+00 -2.07502618e-02
5.41985370e-02 7.95594156e-01 -4.90533888e-01 -7.04879344e-01
-4.40269023e-01 1.05063307e+00 4.28200103e-02 6.14331186e-01
-8.35331500e-01 -7.01207817e-02 6.60411954e-01 9.95333120e-02
8.01489234e-01 -5.14969707e-01 -2.05104932e-01 -1.13081121e+00
-1.73743725e-01 -3.84814441e-01 5.35782516e-01 -1.15896475e+00
-1.37437189e+00 9.10993397e-01 7.27611631e-02 -1.45821917e+00
-7.10845530e-01 -2.86262900e-01 -8.11304092e-01 1.14402914e+00
-2.25466728e+00 -1.66584921e+00 -4.35499810e-02 7.06630588e-01
8.79102588e-01 -5.34830242e-02 7.61931300e-01 7.84243867e-02
1.73540920e-01 -1.44375414e-01 5.32045066e-01 -1.19194910e-01
3.76940966e-01 -9.40531254e-01 1.31281173e+00 1.12162364e+00
-1.38874158e-01 2.04837501e-01 3.86926442e-01 -8.25167775e-01
-7.16403008e-01 -1.95689380e+00 1.67549229e+00 -2.14187488e-01
7.61000514e-01 -1.18191078e-01 -1.05237055e+00 1.25580716e+00
4.84556258e-01 -1.98673457e-01 3.21948528e-01 2.66129263e-02
-4.85412449e-01 -4.02115509e-02 -9.30581212e-01 3.66127014e-01
1.11425102e+00 -5.28163195e-01 -3.93301487e-01 7.49949276e-01
8.83866191e-01 1.55139744e-01 -6.26039028e-01 5.46634138e-01
4.81273413e-01 -1.02258408e+00 9.50980365e-01 -9.60890293e-01
7.84454167e-01 -2.10897207e-01 -2.50717044e-01 -1.51195955e+00
-7.33473182e-01 -3.12498629e-01 1.09914958e-01 5.15573740e-01
4.61576670e-01 -7.18927026e-01 5.70730984e-01 -1.68933287e-01
-3.52361053e-01 -2.39470348e-01 -1.00865948e+00 -1.21453285e+00
4.09183472e-01 -6.70150518e-01 1.05567598e+00 1.18195891e+00
-6.05974615e-01 -9.29840840e-03 -1.19993293e+00 3.97062629e-01
1.52523696e-01 2.44907215e-01 3.22091192e-01 -1.19206691e+00
2.55882353e-01 -4.11979109e-01 -2.03495264e-01 -1.02421713e+00
2.79905707e-01 -7.33382940e-01 -1.01349160e-01 -2.10699844e+00
-2.86392778e-01 -7.28864670e-02 -4.52659816e-01 9.20328915e-01
-9.88835767e-02 3.06700349e-01 8.85084569e-02 1.49171248e-01
-5.51041901e-01 7.07399428e-01 9.51262355e-01 -2.45184019e-01
-1.86561957e-01 2.95644462e-01 -2.01521218e-01 4.18184876e-01
9.32633877e-01 -2.92294383e-01 -2.49377131e-01 -8.23019981e-01
3.08311462e-01 4.54411447e-01 5.21139920e-01 -9.73810434e-01
5.14175221e-02 -6.18718326e-01 2.83500552e-01 -1.28828108e+00
2.21890926e-01 -7.37228453e-01 4.24890578e-01 6.99354649e-01
-4.74695504e-01 2.39424646e-01 4.68504161e-01 5.41074514e-01
-4.06027019e-01 4.89654779e-01 2.72564590e-01 -4.32189286e-01
-1.25752020e+00 4.21316028e-01 -1.07883406e+00 -7.49575317e-01
8.94650578e-01 -2.48932719e-01 -2.20677987e-01 -8.11270058e-01
-9.11371887e-01 4.37694311e-01 -1.88519925e-01 8.25693607e-01
5.30070722e-01 -1.33827496e+00 -1.22669923e+00 -1.33584365e-02
1.11267582e-01 -5.56650341e-01 4.68236804e-01 7.84972608e-01
-5.71387351e-01 1.22594380e+00 -2.13588580e-01 -7.46019125e-01
-8.48221362e-01 5.41317642e-01 3.67296666e-01 -6.04248106e-01
-6.70988441e-01 6.36894822e-01 1.90349296e-01 -1.01706719e+00
-9.54408273e-02 -7.07955658e-01 -3.85674059e-01 9.82757583e-02
7.17594504e-01 5.31847417e-01 2.69416183e-01 -7.65656054e-01
-3.81768912e-01 2.29420722e-01 3.45967025e-01 -8.46702531e-02
2.21430159e+00 -2.89315701e-01 4.59339358e-02 7.86180794e-01
9.98309553e-01 -8.00170004e-01 -1.15761924e+00 -4.67418790e-01
2.80317932e-01 -2.25677133e-01 -2.86952034e-03 -1.00795567e+00
-1.06939876e+00 1.21790075e+00 7.57970750e-01 3.66038710e-01
1.14413631e+00 -4.02906209e-01 1.05614150e+00 5.23167729e-01
2.63379276e-01 -5.39364934e-01 -5.54900706e-01 1.20907593e+00
1.05325091e+00 -1.41496885e+00 -3.17135543e-01 3.21133852e-01
-8.01767945e-01 1.27808499e+00 2.50987202e-01 -9.75777954e-02
9.86192346e-01 -1.17675208e-01 2.36738876e-01 -1.87064111e-01
-1.21996355e+00 -5.07732511e-01 3.26562077e-01 5.41056812e-01
6.84060276e-01 4.10750031e-01 -2.10126877e-01 -3.04314736e-02
7.42425025e-02 4.07431185e-01 3.23835701e-01 7.89681613e-01
-2.30989873e-01 -6.88028216e-01 -3.05143416e-01 3.37945670e-01
-5.21272421e-01 -4.43427265e-01 -3.31447512e-01 7.81525135e-01
-4.32862826e-02 1.10061288e+00 2.69385189e-01 -4.63959634e-01
-1.30921090e-02 1.23616457e-01 -4.38549280e-01 -2.78355449e-01
-1.15127718e+00 8.95007625e-02 2.04749048e-01 -9.62545991e-01
-1.02571535e+00 -5.96358418e-01 -8.77119541e-01 -6.23811781e-01
2.34632686e-01 -1.14004970e-01 6.36560559e-01 1.08554125e+00
6.22807443e-01 4.96312231e-01 6.58338666e-01 -1.28359687e+00
-3.89131516e-01 -1.38653028e+00 -4.68830585e-01 1.90594956e-01
9.13187444e-01 -4.21835333e-02 -1.39699176e-01 -9.32254866e-02] | [6.629281044006348, 2.6933650970458984] |
846f8a4e-22a8-4722-bbd1-eadd40110b74 | unidirectional-imaging-using-deep-learning | 2212.02025 | null | https://arxiv.org/abs/2212.02025v1 | https://arxiv.org/pdf/2212.02025v1.pdf | Unidirectional Imaging using Deep Learning-Designed Materials | A unidirectional imager would only permit image formation along one direction, from an input field-of-view (FOV) A to an output FOV B, and in the reverse path, the image formation would be blocked. Here, we report the first demonstration of unidirectional imagers, presenting polarization-insensitive and broadband unidirectional imaging based on successive diffractive layers that are linear and isotropic. These diffractive layers are optimized using deep learning and consist of hundreds of thousands of diffractive phase features, which collectively modulate the incoming fields and project an intensity image of the input onto an output FOV, while blocking the image formation in the reverse direction. After their deep learning-based training, the resulting diffractive layers are fabricated to form a unidirectional imager. As a reciprocal device, the diffractive unidirectional imager has asymmetric mode processing capabilities in the forward and backward directions, where the optical modes from B to A are selectively guided/scattered to miss the output FOV, whereas for the forward direction such modal losses are minimized, yielding an ideal imaging system between the input and output FOVs. Although trained using monochromatic illumination, the diffractive unidirectional imager maintains its functionality over a large spectral band and works under broadband illumination. We experimentally validated this unidirectional imager using terahertz radiation, very well matching our numerical results. Using the same deep learning-based design strategy, we also created a wavelength-selective unidirectional imager, where two unidirectional imaging operations, in reverse directions, are multiplexed through different illumination wavelengths. Diffractive unidirectional imaging using structured materials will have numerous applications in e.g., security, defense, telecommunications and privacy protection. | ['Aydogan Ozcan', 'Mona Jarrahi', 'Songyu Sun', 'Che-Yung Shen', 'Bijie Bai', 'Yifan Zhao', 'Tianyi Gan', 'Jingxi Li'] | 2022-12-05 | null | null | null | null | ['blocking'] | ['natural-language-processing'] | [ 8.28419268e-01 2.63251036e-01 1.15501598e-01 -9.29044262e-02
3.74446273e-01 -4.99423563e-01 4.31089848e-01 -1.00760353e+00
-3.94773185e-01 5.92176497e-01 -7.53210485e-02 -3.20244491e-01
-1.55881971e-01 -1.10410416e+00 -6.71399415e-01 -1.65631545e+00
3.11885506e-01 2.51656115e-01 1.47927895e-01 6.57806918e-02
3.44775587e-01 6.27892256e-01 -1.63277674e+00 6.29672766e-01
5.11178553e-01 1.10520780e+00 4.27350640e-01 4.57994074e-01
-8.10748935e-02 7.30066478e-01 -4.23455596e-01 -2.24034954e-02
3.00328940e-01 -3.61929625e-01 -2.34956488e-01 3.19376886e-02
4.83990312e-01 -5.73652029e-01 -8.41619492e-01 7.61942267e-01
1.52467117e-01 -3.95316035e-01 6.27392232e-01 -6.90923572e-01
-1.26241112e+00 1.88086793e-01 -4.50209290e-01 -3.39239061e-01
1.57602385e-01 6.20754182e-01 5.27861416e-01 -7.10241437e-01
7.20943689e-01 6.09412253e-01 3.33204985e-01 9.28359449e-01
-1.09036231e+00 -9.18838799e-01 -2.92855829e-01 -2.31971487e-01
-8.91752422e-01 -6.07783735e-01 8.16762030e-01 -4.24746841e-01
9.83691990e-01 3.57718587e-01 6.26041889e-01 6.64127409e-01
7.99276233e-01 -2.94197053e-02 1.52195179e+00 -3.91860783e-01
-2.47901440e-01 3.58026177e-01 -1.03289731e-01 7.89143920e-01
7.80699402e-02 6.92983925e-01 -5.31663537e-01 3.82884651e-01
7.40068257e-01 3.35009575e-01 -8.65702927e-01 -2.80243307e-01
-1.14013326e+00 8.97187665e-02 5.45068383e-01 2.90627807e-01
-1.66142642e-01 -1.28564715e-01 -6.09652221e-01 5.51808774e-01
1.64123863e-01 9.11298931e-01 1.39964774e-01 5.92045784e-01
-8.35439444e-01 -1.52795359e-01 5.73646724e-01 5.41171193e-01
7.25596368e-01 -1.50019646e-01 -1.96875811e-01 7.71386027e-01
4.66619998e-01 1.13987315e+00 1.09997593e-01 -1.03299534e+00
3.26632738e-01 4.76449639e-01 4.50572670e-01 -5.32441378e-01
-3.64655793e-01 -4.74808514e-01 -7.51926482e-01 6.42221987e-01
3.81816566e-01 -1.18726619e-01 -1.13951135e+00 1.53442216e+00
8.72052610e-02 -5.77923179e-01 4.30329233e-01 1.20340729e+00
8.58908117e-01 1.00885248e+00 -5.01520991e-01 -6.23709977e-01
1.09888196e+00 -6.14886343e-01 -6.51924253e-01 -1.66959316e-01
2.07635477e-01 -9.25784409e-01 8.64418447e-01 5.09872198e-01
-1.43439209e+00 -9.40737426e-02 -1.13242614e+00 1.63716942e-01
8.50748792e-02 2.75461882e-01 4.04453516e-01 3.74283552e-01
-8.08290064e-01 3.81985575e-01 -4.91231352e-01 1.85243294e-01
2.93581992e-01 4.47391212e-01 -3.85467887e-01 -3.79426301e-01
-6.20540082e-01 1.26439303e-01 -8.03225935e-02 -9.14031044e-02
-4.34640944e-01 -1.02667201e+00 -4.09526192e-02 -1.27653465e-01
-2.65812546e-01 -8.16268981e-01 5.87087095e-01 -6.07918978e-01
-2.06301737e+00 1.40548563e+00 -2.65169024e-01 -2.55827010e-01
1.86477482e-01 1.78857088e-01 -5.63880444e-01 4.12978947e-01
1.15614116e-01 3.73350859e-01 1.03462183e+00 -1.51806533e+00
-7.64591277e-01 -5.57414174e-01 -1.77832171e-01 3.43737602e-02
-4.83359218e-01 -2.13372409e-01 -2.33873706e-02 5.00911400e-02
7.26198375e-01 -9.30805027e-01 3.80132049e-01 -3.03343367e-02
-6.28179610e-01 2.81182259e-01 1.65588200e+00 1.62081316e-01
8.35582197e-01 -2.33990741e+00 -4.67588268e-02 -1.53216958e-01
6.02952123e-01 3.72004882e-02 -5.55983894e-02 6.86593115e-01
-3.35517041e-02 -2.67517030e-01 -3.28424692e-01 2.13369355e-03
-6.74703598e-01 -2.24806800e-01 -4.07052040e-01 7.48783946e-01
-2.21796602e-01 7.52687037e-01 -4.62137222e-01 -1.07819075e-03
2.57780384e-02 2.76865274e-01 -7.12623179e-01 4.84209597e-01
-9.57539529e-02 8.66620660e-01 -2.96343774e-01 5.30915856e-01
1.17131245e+00 -2.88189113e-01 1.94665506e-01 -5.84663391e-01
-1.09780705e+00 1.37853846e-01 -2.55252123e-01 7.56548703e-01
-5.30791521e-01 8.47548723e-01 4.91192728e-01 -1.12481320e+00
1.04502499e+00 3.34287077e-01 9.31189060e-01 -1.65104711e+00
-2.97388405e-01 5.06205559e-01 4.04905565e-02 -8.59669507e-01
3.01011384e-01 -8.35900605e-01 1.87748987e-02 8.03405225e-01
-5.25185056e-02 -2.61371553e-01 5.51768169e-02 -2.82020122e-01
9.58926499e-01 -3.15802217e-01 -5.10649741e-01 -1.34069890e-01
3.94495785e-01 4.55558905e-03 2.71114856e-01 9.88283455e-01
3.09382498e-01 7.08061397e-01 1.67787969e-01 -7.78905034e-01
-1.16064751e+00 -1.37478232e+00 -3.07382375e-01 5.60125649e-01
7.18089700e-01 3.33118141e-01 -2.57094085e-01 -4.51773815e-02
-2.05661803e-02 4.43612456e-01 4.01358726e-03 -2.17218518e-01
-8.69700372e-01 -7.77707934e-01 1.98136233e-02 -3.08531970e-01
9.61013734e-01 -1.04653537e+00 -6.13195181e-01 9.84796733e-02
7.36353993e-02 -1.09033704e+00 -5.20893745e-02 -2.47753356e-02
-5.39739192e-01 -1.22055316e+00 -7.51869321e-01 -7.84254193e-01
7.53350317e-01 4.32558447e-01 8.34175766e-01 -1.96837917e-01
-4.10652012e-01 6.57934725e-01 3.59803557e-01 -7.41205588e-02
-3.84511530e-01 -2.81774521e-01 8.61108303e-02 2.60481089e-01
1.21236056e-01 -9.15241718e-01 -1.21465445e+00 4.90889162e-01
-6.46725357e-01 5.20037234e-01 8.20840240e-01 7.42786705e-01
5.00294864e-01 -1.64547980e-01 2.68012643e-01 -8.97342384e-01
7.65774548e-02 -3.37434202e-01 -9.48829472e-01 -7.10811764e-02
-2.97811121e-01 -2.60266989e-01 8.53428960e-01 -4.30389673e-01
-1.50216222e+00 -3.82850260e-01 -1.92392573e-01 4.41807806e-02
-2.41369054e-01 3.72025929e-02 9.61849988e-02 -4.42855984e-01
5.85474193e-01 6.21721625e-01 9.12080482e-02 -3.17685492e-02
-1.47828341e-01 8.07035446e-01 4.41538751e-01 4.10992689e-02
6.52847409e-01 1.22214043e+00 2.64199167e-01 -1.41564500e+00
-7.45138168e-01 -3.12083811e-01 7.35166073e-02 -5.15321851e-01
9.31923091e-01 -7.70120502e-01 -1.25060558e+00 7.85489440e-01
-1.04310644e+00 -3.13689888e-01 -3.70667040e-01 7.02476144e-01
-3.54717344e-01 -3.83862406e-01 -6.99603498e-01 -5.50548196e-01
-2.22785234e-01 -1.12919819e+00 8.90313387e-01 4.88650858e-01
2.20955402e-01 -6.85041785e-01 -8.62543434e-02 7.40093112e-01
7.18586445e-01 -2.59034127e-01 1.48263693e+00 5.26533782e-01
-1.36377823e+00 -7.57125439e-03 -3.15505028e-01 1.52242452e-01
4.25576031e-01 -2.72224635e-01 -9.95710135e-01 -3.21528167e-01
4.86692220e-01 -2.36144602e-01 1.32214999e+00 6.71658158e-01
1.13010311e+00 -4.82960969e-01 -7.72228301e-01 1.19358730e+00
1.49126244e+00 6.22607112e-01 9.46826041e-01 2.24197730e-01
8.37384820e-01 5.07394195e-01 1.23938546e-01 -2.80355290e-02
-1.90855935e-01 4.66528803e-01 5.41847765e-01 -5.36844313e-01
-2.87537813e-01 2.39653409e-01 3.41303796e-01 6.99770212e-01
-3.17304641e-01 -7.82733262e-01 -3.85603994e-01 2.19219148e-01
-1.06800687e+00 -1.07108355e+00 -3.23392630e-01 2.43048501e+00
7.51659214e-01 -2.12215140e-01 -5.88313460e-01 -4.15495127e-01
3.85061502e-01 1.62702098e-01 -5.47713637e-01 -1.72190830e-01
-3.15116346e-01 3.76020908e-01 7.31951773e-01 7.09616721e-01
-8.14277291e-01 7.48784482e-01 6.40974045e+00 2.50551254e-01
-2.04080939e+00 -2.10632235e-01 2.25661173e-01 -4.81050968e-01
-1.09174430e+00 -1.84109166e-01 -7.42785394e-01 5.41438758e-01
1.84502244e-01 5.50957441e-01 2.07672626e-01 4.06851694e-02
3.30304503e-01 -1.96605951e-01 -7.80041277e-01 1.04505360e+00
-5.75909495e-01 -1.74538183e+00 1.48289651e-01 3.66423458e-01
8.63998711e-01 1.11453898e-01 6.37080729e-01 -4.79551882e-01
-1.91254169e-01 -1.04645288e+00 4.07650501e-01 7.95745194e-01
1.20361149e+00 -5.45431733e-01 -8.75108410e-03 3.43283862e-01
-4.29380387e-01 -3.84811044e-01 -2.25409567e-01 -8.78693014e-02
6.96194395e-02 1.11613214e+00 -3.66323203e-01 -8.94006900e-03
1.00140119e+00 4.49534714e-01 5.98155856e-01 2.84071296e-01
3.07551469e-03 5.28041482e-01 4.62569296e-02 1.36244476e-01
1.74404830e-01 -9.01811004e-01 1.11709738e+00 7.00533926e-01
4.88421202e-01 3.45062703e-01 -1.91820771e-01 1.59322524e+00
-3.26844275e-01 -7.08435357e-01 -9.14763033e-01 -8.61210003e-02
9.75906253e-02 1.29419184e+00 -4.06739473e-01 1.77104119e-02
-4.65603977e-01 9.82690156e-01 -1.19899120e-02 9.42022502e-01
-4.07325625e-01 -4.81586665e-01 9.96235073e-01 6.84792399e-01
1.10952534e-01 -4.08948362e-01 -4.99933586e-02 -1.11272180e+00
1.34020463e-01 -1.38687640e-01 -2.92070478e-01 -6.61271214e-01
-1.01606870e+00 8.77986923e-02 -3.68402213e-01 -8.01747561e-01
6.43185735e-01 -8.38511765e-01 -4.72161323e-01 1.02267802e+00
-1.75022352e+00 -9.82880652e-01 -1.65059373e-01 3.74264389e-01
-2.14481037e-02 -4.97912169e-02 5.10413051e-01 2.82156974e-01
-2.64699105e-02 1.80093460e-02 5.44275165e-01 -2.00545624e-01
8.43661666e-01 -6.84315860e-01 -4.18733686e-01 6.15878522e-01
-3.52347463e-01 7.71430969e-01 4.73827153e-01 -4.26075816e-01
-1.83332014e+00 -1.15889645e+00 5.37517726e-01 1.20233320e-01
4.15840223e-02 -4.24136609e-01 -6.18033767e-01 3.50042582e-01
5.88051021e-01 2.00017206e-02 7.30497718e-01 -5.89415789e-01
-9.51122865e-02 -5.00184000e-01 -1.22054231e+00 5.21395385e-01
1.37442279e+00 -4.31225270e-01 -2.01835465e-02 6.31000638e-01
6.19604647e-01 -4.63758022e-01 -5.06487489e-01 6.32175207e-01
9.32182193e-01 -1.48417127e+00 8.97056162e-01 -3.08151186e-01
9.17351961e-01 -2.64341503e-01 7.82032162e-02 -1.02619410e+00
-5.93955994e-01 -5.82838833e-01 3.08843017e-01 8.55258465e-01
4.58721161e-01 -1.39514804e+00 9.30112064e-01 -8.09128881e-02
-5.46132326e-01 -7.00103402e-01 -7.32079625e-01 -5.17190874e-01
-2.17108633e-02 2.70427406e-01 2.99626827e-01 6.39749467e-01
-3.85649472e-01 1.07502826e-01 -1.71337083e-01 5.73175251e-01
9.39175963e-01 8.20325434e-01 1.13616377e-01 -9.03225064e-01
-9.43114311e-02 -3.27381849e-01 1.03133075e-01 -1.49094212e+00
-7.10762367e-02 -8.93990219e-01 4.80229445e-02 -1.41751683e+00
-4.02668491e-02 -1.01227820e+00 6.71112612e-02 1.50949225e-01
6.17057323e-01 4.93205875e-01 2.32113171e-02 7.00802505e-01
3.99778783e-02 5.35127401e-01 1.76932049e+00 -5.03843546e-01
-3.80526632e-01 2.48857290e-01 -4.83435035e-01 3.76286060e-01
6.68558776e-01 -3.89410019e-01 -3.52812856e-01 -5.98224699e-01
4.26485538e-01 -1.43067360e-01 4.38101858e-01 -1.07565045e+00
3.11453342e-01 -3.19049835e-01 4.39117432e-01 -1.58103108e-01
3.92967165e-01 -7.00091481e-01 1.92611471e-01 5.14029026e-01
-4.58686277e-02 -9.63050842e-01 -2.24143431e-01 1.61670655e-01
-1.63578928e-01 6.12575896e-02 1.16268933e+00 -4.91384834e-01
-3.85453731e-01 3.13391984e-01 -7.65662968e-01 -1.83919162e-01
1.10962391e+00 -5.22807539e-01 -9.36232448e-01 -3.98478471e-02
-3.93532336e-01 -2.36013517e-01 8.15274954e-01 -1.44246623e-01
8.27655971e-01 -1.04592371e+00 -1.69205427e-01 5.28020620e-01
-1.50987878e-01 -2.39007905e-01 5.69548190e-01 9.31382120e-01
-8.44672322e-01 6.96497321e-01 -2.21232414e-01 -9.70854402e-01
-1.22138250e+00 1.65347159e-01 8.01708221e-01 -2.10545212e-02
-8.58488679e-01 9.40761030e-01 9.57861781e-01 -5.01440883e-01
-4.46221292e-01 1.05036683e-01 -1.96456224e-01 -4.93064106e-01
4.31304872e-01 -4.13996056e-02 7.50838891e-02 -3.18075001e-01
-1.57820091e-01 1.00449193e+00 4.41149957e-02 1.78445458e-01
1.54595363e+00 -1.82750419e-01 -4.73789513e-01 -1.10719778e-01
1.22514951e+00 5.40598750e-01 -1.40402341e+00 3.22826020e-02
-9.86937463e-01 -3.99734646e-01 1.92349166e-01 -4.72808570e-01
-1.47050440e+00 8.82473409e-01 4.94018823e-01 4.40100968e-01
1.18297327e+00 4.30434227e-01 6.88209713e-01 1.92751005e-01
2.25189999e-01 -1.00880599e+00 2.04159573e-01 2.28883907e-01
6.37436330e-01 -9.22835350e-01 -2.74461240e-01 -6.01710081e-01
1.66117530e-02 1.37161613e+00 5.65450311e-01 5.35899028e-02
4.42145228e-01 5.39675832e-01 3.52460751e-03 -8.76646042e-01
-6.28847063e-01 4.26033407e-01 4.94756848e-02 8.22801113e-01
6.50208354e-01 -5.67313237e-03 -2.73661055e-02 -1.40327364e-01
-1.92787051e-02 -9.05275866e-02 4.69138891e-01 7.51201093e-01
-4.58893418e-01 -8.40604961e-01 -2.59757996e-01 5.20178795e-01
-3.62221092e-01 1.75275862e-01 -1.45805165e-01 5.61300218e-01
3.11450779e-01 6.30710304e-01 5.45665145e-01 1.71450507e-02
2.82502174e-01 -2.93905675e-01 8.35017383e-01 -6.22892737e-01
6.90372139e-02 -1.67573974e-01 -2.13844731e-01 -5.39038479e-01
-5.89149594e-01 -1.62140995e-01 -1.16519129e+00 -3.09685990e-02
9.33227129e-03 -2.32910290e-01 3.87815952e-01 1.03954768e+00
5.63571334e-01 2.98339546e-01 8.17221820e-01 -5.20451665e-01
1.58544153e-01 -2.67268598e-01 -6.95562065e-01 1.92089230e-01
7.45254457e-01 -4.14931744e-01 -5.43962002e-01 -2.02179849e-01] | [10.19332504272461, -2.6068062782287598] |
fb6f39c4-3fc2-415d-bca3-40fbfeef77e2 | ying-yong-ji-yi-zeng-qiang-tiao-jian-sui-ji | null | null | https://aclanthology.org/2019.ijclclp-1.1 | https://aclanthology.org/2019.ijclclp-1.1.pdf | 應用記憶增強條件隨機場域與之深度學習及自動化詞彙特徵於中文命名實體辨識之研究 (Leveraging Memory Enhanced Conditional Random Fields with Gated CNN and Automatic BAPS Features for Chinese Named Entity Recognition) | null | ['Chia-Hui Chang', 'Kuo-Chun Chien'] | null | null | null | null | ijclclp-2019-6 | ['chinese-named-entity-recognition'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.315688133239746, 3.800696849822998] |
c94c95da-0d60-4bb2-aa5d-b0f071cb0a28 | listening-human-behavior-3d-human-pose | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Shibata_Listening_Human_Behavior_3D_Human_Pose_Estimation_With_Acoustic_Signals_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Shibata_Listening_Human_Behavior_3D_Human_Pose_Estimation_With_Acoustic_Signals_CVPR_2023_paper.pdf | Listening Human Behavior: 3D Human Pose Estimation With Acoustic Signals | Given only acoustic signals without any high-level information, such as voices or sounds of scenes/actions, how much can we infer about the behavior of humans? Unlike existing methods, which suffer from privacy issues because they use signals that include human speech or the sounds of specific actions, we explore how low-level acoustic signals can provide enough clues to estimate 3D human poses by active acoustic sensing with a single pair of microphones and loudspeakers (see Fig. 1). This is a challenging task since sound is much more diffractive than other signals and therefore covers up the shape of objects in a scene. Accordingly, we introduce a framework that encodes multichannel audio features into 3D human poses. Aiming to capture subtle sound changes to reveal detailed pose information, we explicitly extract phase features from the acoustic signals together with typical spectrum features and feed them into our human pose estimation network. Also, we show that reflected or diffracted sounds are easily influenced by subjects' physique differences e.g., height and muscularity, which deteriorates prediction accuracy. We reduce these gaps by using a subject discriminator to improve accuracy. Our experiments suggest that with the use of only low-dimensional acoustic information, our method outperforms baseline methods. The datasets and codes used in this project will be publicly available. | ['Yoshimitsu Aoki', 'Akisato Kimura', 'Go Irie', 'Mariko Isogawa', 'Yutaka Kawashima', 'Yuto Shibata'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['3d-human-pose-estimation'] | ['computer-vision'] | [ 3.77758622e-01 -1.86221041e-02 3.14110249e-01 -3.79769146e-01
-7.36113727e-01 -5.33183753e-01 1.66415349e-01 -8.13445002e-02
-3.53272200e-01 5.12778878e-01 5.32659769e-01 4.06616300e-01
1.02540836e-01 -7.03871429e-01 -5.34137011e-01 -6.20543242e-01
-2.75116861e-01 1.78096257e-02 2.59006232e-01 -2.05111399e-01
-3.36004421e-02 1.93827584e-01 -1.79629040e+00 3.82037729e-01
3.87295008e-01 1.23285925e+00 3.25120501e-02 7.58132756e-01
3.19607049e-01 2.01064974e-01 -9.29820895e-01 -2.64878958e-01
3.48846674e-01 -3.10052514e-01 -2.89880455e-01 -8.78954977e-02
4.55374002e-01 -4.38359946e-01 -2.17065319e-01 7.95236051e-01
8.31893802e-01 1.24323301e-01 3.77441138e-01 -9.53297019e-01
-1.99822709e-01 4.32331145e-01 -7.79308602e-02 -1.33167133e-01
1.14185584e+00 2.11184606e-01 9.29841995e-01 -7.23826110e-01
1.39917374e-01 1.32465923e+00 1.03761220e+00 5.71070015e-01
-9.74339843e-01 -7.30743885e-01 -5.85167632e-02 2.73129761e-01
-1.27548194e+00 -8.42078984e-01 1.17704368e+00 -4.47152317e-01
3.88573468e-01 8.03424418e-01 8.20637524e-01 1.74148357e+00
-3.16188931e-02 5.00540257e-01 9.65076089e-01 -2.57268816e-01
2.92838156e-01 2.41105959e-01 -1.88373238e-01 5.48911095e-01
8.26871321e-02 2.43153855e-01 -9.68832731e-01 -5.81660450e-01
5.04469633e-01 6.90141739e-03 -5.90436816e-01 -3.16352397e-01
-1.36946893e+00 3.63207281e-01 2.47481927e-01 1.53135076e-01
-2.29947597e-01 2.02255603e-02 7.54189044e-02 1.20739974e-01
2.10451111e-01 6.34508252e-01 -2.96885550e-01 -3.12851131e-01
-4.82771128e-01 2.44472548e-01 7.81487525e-01 5.06841719e-01
5.54181933e-01 -1.13238320e-01 9.17845517e-02 9.23085630e-01
3.51223797e-01 7.64852524e-01 3.48444790e-01 -1.05908060e+00
3.93274397e-01 1.08268544e-01 2.67766684e-01 -1.62576771e+00
-7.26191759e-01 -4.86767679e-01 -5.49255848e-01 -1.93745628e-01
6.17344022e-01 -4.42606628e-01 -3.50469738e-01 1.92005420e+00
5.01828015e-01 2.90143192e-01 -5.49963415e-01 1.26610541e+00
8.80896747e-01 1.79433599e-01 -4.18052137e-01 -2.99394667e-01
1.36068511e+00 -3.64787757e-01 -7.07086563e-01 -4.52844411e-01
2.43894547e-01 -6.15276575e-01 1.02173364e+00 5.96791387e-01
-9.51778591e-01 -7.18723178e-01 -9.44682360e-01 2.39367098e-01
-9.43408161e-02 -1.41832709e-01 5.57073295e-01 9.35936272e-01
-6.08920693e-01 3.70311230e-01 -8.52770329e-01 1.13691725e-02
1.08667754e-01 2.40789697e-01 -2.88767487e-01 1.48429647e-01
-1.39869022e+00 4.33816612e-01 -3.91457468e-01 3.22819948e-01
-6.95631266e-01 -6.71549082e-01 -9.50842738e-01 -3.04084718e-01
6.36500776e-01 -5.71167469e-01 1.07059062e+00 -4.51819420e-01
-1.58348465e+00 2.97815323e-01 -6.87484443e-02 -2.75182049e-03
4.44339931e-01 -4.62983578e-01 -4.05960768e-01 3.47926050e-01
-1.79354936e-01 1.58288643e-01 9.60681379e-01 -1.20784581e+00
-1.46101743e-01 -5.88932097e-01 5.13152033e-03 1.06692180e-01
-5.35949588e-01 -1.68545358e-02 -2.16015249e-01 -5.52767694e-01
3.52907658e-01 -9.54212606e-01 -2.99254239e-01 2.78741479e-01
-5.96247911e-01 3.32109988e-01 3.59027892e-01 -6.42952383e-01
1.15662229e+00 -2.37699008e+00 -8.28296617e-02 4.19374049e-01
3.03350061e-01 -6.01208657e-02 -1.19358018e-01 2.40691066e-01
3.45936298e-01 -1.96266294e-01 -5.13799600e-02 -4.43058372e-01
4.01004441e-02 -1.18790589e-01 -2.03012347e-01 6.17240667e-01
-3.00579756e-01 5.98430514e-01 -8.24919462e-01 -2.78114557e-01
2.32826956e-02 6.33047462e-01 -8.55643988e-01 1.83084458e-01
1.74414724e-01 8.59790146e-01 -6.39826179e-01 6.74514949e-01
5.43486834e-01 1.56213164e-01 -1.08978413e-01 -2.66872495e-01
2.13194087e-01 5.80801249e-01 -1.36117232e+00 1.72643209e+00
-4.08267140e-01 4.12335336e-01 4.18943286e-01 -1.10622001e+00
9.41015959e-01 5.07419765e-01 5.94419897e-01 -4.63828921e-01
9.48108956e-02 1.26647651e-01 -6.29822463e-02 -8.01214635e-01
4.10320871e-02 -1.99756701e-04 -2.65575320e-01 2.68346876e-01
-3.09941977e-01 -6.35554865e-02 -4.13633615e-01 -2.54121125e-01
1.37438977e+00 -3.01310092e-01 1.70084789e-01 1.49717301e-01
2.48358786e-01 -6.21552885e-01 7.16925442e-01 8.04549217e-01
-2.60614544e-01 8.34299862e-01 8.86692032e-02 -1.16076417e-01
-4.39045489e-01 -1.24700403e+00 -1.60503089e-01 1.09875739e+00
7.58546293e-02 -6.22695446e-01 -6.70797646e-01 -2.62960345e-01
6.18656427e-02 3.10837209e-01 -6.19899988e-01 -2.43301123e-01
-5.83072841e-01 -5.50188899e-01 6.95228696e-01 4.37013835e-01
2.10534677e-01 -7.83915341e-01 -8.65941882e-01 2.08755791e-01
-5.20869970e-01 -1.27316785e+00 -6.14509046e-01 -1.49155501e-02
-4.06067103e-01 -9.94130552e-01 -6.02661967e-01 -3.38591009e-01
4.26223636e-01 1.35898173e-01 8.17068219e-01 -1.31052434e-01
-6.18052125e-01 7.25142896e-01 -4.38629001e-01 -5.53304374e-01
5.83061241e-02 -4.97246563e-01 5.20317137e-01 2.59748280e-01
1.69657752e-01 -1.02105987e+00 -7.88779140e-01 5.13035655e-01
-1.21242076e-01 -2.61550069e-01 3.00780714e-01 5.92665076e-01
2.34864846e-01 1.66388676e-01 3.95442754e-01 -5.87941885e-01
5.27683437e-01 -2.75263220e-01 -1.04893021e-01 -2.31593817e-01
1.26498014e-01 -4.72626179e-01 5.33441365e-01 -7.68671274e-01
-8.34100068e-01 2.42586702e-01 -2.30525836e-01 -1.70102105e-01
-5.50336957e-01 1.36989921e-01 -4.06449705e-01 4.80772369e-02
8.52010429e-01 1.52110144e-01 -2.69133248e-03 -6.37340128e-01
-3.93649889e-03 8.48784983e-01 6.60475612e-01 -4.74312454e-01
7.71770179e-01 5.22345126e-01 9.02032182e-02 -1.13947809e+00
-7.76297271e-01 -4.63360876e-01 -5.13477981e-01 -4.23065752e-01
6.91903472e-01 -9.02337074e-01 -1.08737683e+00 4.71607864e-01
-1.03844726e+00 7.11794570e-03 -1.23105617e-02 9.43301499e-01
-4.96380478e-01 3.91224235e-01 -6.69668436e-01 -1.36261916e+00
1.25475422e-01 -1.00710118e+00 1.24255621e+00 -3.42446148e-01
-6.22098684e-01 -6.17397130e-01 -1.16816023e-02 7.07524419e-01
2.82238036e-01 3.79671097e-01 4.17183042e-01 -5.09114146e-01
-3.79689604e-01 -2.21620500e-01 6.12654805e-01 1.98401332e-01
3.08911443e-01 -5.67575336e-01 -1.24007523e+00 -1.89959794e-01
4.24352109e-01 -4.50840771e-01 4.85288411e-01 4.86819774e-01
1.33872056e+00 -5.80494225e-01 -3.16905916e-01 6.30764365e-01
6.24082088e-01 1.14304461e-01 2.74298787e-01 -2.32733533e-01
7.60620117e-01 1.17920232e+00 4.40881878e-01 7.61427045e-01
2.02337116e-01 8.40768456e-01 3.16376597e-01 2.09596038e-01
9.37427208e-02 -2.54374415e-01 5.93932271e-01 1.03495502e+00
-2.01052442e-01 -7.39715174e-02 -7.08594024e-01 1.50689214e-01
-1.34627008e+00 -9.49214280e-01 -8.73851702e-02 2.24533916e+00
8.82902026e-01 1.16754532e-01 2.01660126e-01 5.40683866e-01
4.61152285e-01 1.63151637e-01 -4.27072853e-01 4.31237742e-02
1.10827729e-01 1.33213148e-01 1.39049858e-01 4.08807486e-01
-8.74416232e-01 9.65222418e-02 6.70551920e+00 4.47820485e-01
-9.99311209e-01 -1.04761720e-01 1.50809586e-01 -5.47468305e-01
-4.25798357e-01 -5.71813822e-01 -4.43168372e-01 5.95907450e-01
7.28831470e-01 5.54931641e-01 5.57655394e-01 5.74135303e-01
3.35588425e-01 7.19907805e-02 -1.45122409e+00 1.13102067e+00
1.97568938e-01 -6.18433177e-01 -4.07658786e-01 9.88427401e-02
6.96757138e-02 -2.68683136e-01 2.95030355e-01 -3.84024419e-02
-1.19470753e-01 -1.01553583e+00 8.17777634e-01 6.51292920e-01
5.50153613e-01 -4.79445606e-01 2.95090258e-01 6.31018698e-01
-1.16082275e+00 -2.26668209e-01 -2.69599766e-01 -4.01599675e-01
1.74269959e-01 9.01212394e-01 -6.96954548e-01 2.55088508e-01
8.39143515e-01 4.91226733e-01 -3.96449566e-01 9.71142352e-01
-3.85395661e-02 7.92324483e-01 -6.94747508e-01 -2.04740539e-01
-5.51383972e-01 1.47596449e-01 8.64385784e-01 8.79720271e-01
6.00636244e-01 4.32166189e-01 2.66296923e-01 8.51586878e-01
3.62670124e-01 6.89641610e-02 -6.03365958e-01 6.16437905e-02
6.47386551e-01 8.74250472e-01 -2.64025390e-01 2.19434783e-01
-4.34615284e-01 6.83380902e-01 -3.13414067e-01 3.82613808e-01
-4.74047869e-01 -2.33320192e-01 8.13017547e-01 3.61630023e-01
8.79205912e-02 -3.20594370e-01 -2.07348034e-01 -1.09636545e+00
3.21713120e-01 -9.64033365e-01 9.83579084e-02 -7.41543531e-01
-1.33646202e+00 4.05115515e-01 -2.24397689e-01 -1.30095947e+00
-4.83393222e-01 -5.42023063e-01 -4.07991707e-01 5.60801804e-01
-7.48201132e-01 -6.57132387e-01 -3.93484354e-01 7.28487551e-01
2.21616223e-01 -2.07903385e-02 9.88588750e-01 3.11225414e-01
-1.83201030e-01 8.96819293e-01 -1.73649654e-01 3.41350228e-01
8.35757136e-01 -8.35281968e-01 1.65588036e-01 3.27448457e-01
3.84583622e-01 7.14025438e-01 9.79808748e-01 -5.17078519e-01
-1.55218339e+00 -5.75491309e-01 6.81126714e-01 -8.34600866e-01
4.37175453e-01 -9.73123789e-01 -6.98395073e-01 3.38374645e-01
-5.12806475e-01 2.26732895e-01 1.09996676e+00 4.27607238e-01
-3.93569022e-01 -1.73498824e-01 -1.05435383e+00 4.57824260e-01
1.44798625e+00 -6.75218821e-01 -7.05789626e-01 8.99326578e-02
6.88975155e-01 -3.56942743e-01 -7.01403499e-01 4.74750847e-01
8.99039626e-01 -1.10287130e+00 1.32710290e+00 -1.14639901e-01
6.92835450e-02 -8.93357024e-03 -4.00854081e-01 -1.25482380e+00
-2.25587547e-01 -6.95226312e-01 -9.44261253e-03 9.95245874e-01
2.06744418e-01 -6.77095890e-01 7.26078391e-01 4.71094757e-01
8.16042572e-02 -6.18057430e-01 -9.86811042e-01 -6.09257758e-01
-2.97232389e-01 -9.78813350e-01 6.68881238e-01 7.87519574e-01
2.73765028e-01 1.27368242e-01 -6.56589687e-01 3.87321979e-01
7.63740897e-01 -5.27309962e-02 7.43689954e-01 -1.46973813e+00
-7.38042951e-01 4.48263995e-02 -5.30830383e-01 -1.22366130e+00
-3.90006113e-03 -4.75833237e-01 3.90654683e-01 -8.00901830e-01
-1.02746159e-01 -4.67321843e-01 -2.37124428e-01 3.45232606e-01
-2.57396623e-02 4.44362134e-01 7.52528831e-02 -1.18529618e-01
-4.59256142e-01 6.39773190e-01 1.30972123e+00 -1.38668507e-01
-1.10958517e-01 4.07769799e-01 -6.75284624e-01 1.06329894e+00
5.24618924e-01 -3.12510163e-01 -4.28667486e-01 -1.80616260e-01
3.09177250e-01 4.38987762e-01 7.18118966e-01 -1.17649376e+00
2.84310341e-01 -1.42854407e-01 5.87798834e-01 -2.71788865e-01
9.67791915e-01 -8.04344416e-01 2.09846869e-01 3.89280498e-01
-6.12639308e-01 -5.32775164e-01 -2.06815600e-01 8.41877401e-01
-1.20519884e-02 5.40666543e-02 2.62346476e-01 -1.94238499e-01
-1.15439340e-01 2.16894329e-01 -4.73617822e-01 -5.39115779e-02
4.76552755e-01 -2.57804960e-01 -1.38771944e-02 -7.40481555e-01
-8.33030224e-01 -1.36828005e-01 1.68929636e-01 4.62869525e-01
4.85340297e-01 -1.38125193e+00 -5.67722261e-01 5.03930151e-01
7.49157518e-02 -1.30177975e-01 2.18853250e-01 7.41806090e-01
1.42296314e-01 3.74644607e-01 -6.21343479e-02 -7.08873808e-01
-1.37032020e+00 2.80781090e-01 2.10612983e-01 5.81659138e-01
-6.29491210e-01 1.18640554e+00 4.40553755e-01 -4.15542394e-01
4.67293322e-01 -3.36979091e-01 -2.06999376e-01 1.04194865e-01
6.69414818e-01 4.17478234e-01 -4.01932672e-02 -6.94798887e-01
-6.89880967e-01 6.21511161e-01 4.55661535e-01 -2.26072133e-01
1.16816652e+00 -4.00240928e-01 1.78741425e-01 7.32755721e-01
1.36473334e+00 7.99945533e-01 -1.19456053e+00 -3.93587828e-01
-4.23890054e-01 -7.13422477e-01 -2.10494831e-01 -6.49435878e-01
-9.51963425e-01 9.45137143e-01 5.34771025e-01 2.80700058e-01
1.17747843e+00 1.37729257e-01 9.17262614e-01 4.34910566e-01
7.72424221e-01 -1.01353610e+00 4.66867298e-01 1.28238693e-01
8.18041205e-01 -1.06340837e+00 -1.50033951e-01 -6.96739137e-01
-2.87697017e-01 7.58094490e-01 4.53755498e-01 2.30745211e-01
7.85559356e-01 4.77400869e-01 3.03778052e-01 8.40767846e-02
-6.52051806e-01 -1.18278615e-01 5.84913015e-01 9.66233134e-01
2.27903947e-01 1.40043795e-01 8.88193697e-02 1.20082068e+00
-1.02326596e+00 -4.24468517e-01 2.41604954e-01 6.31123424e-01
-5.10768175e-01 -8.16720009e-01 -8.11853111e-01 4.07301188e-01
-4.72194016e-01 2.74393052e-01 -5.22849560e-01 1.00619994e-01
4.14668143e-01 1.31845856e+00 -6.25903159e-02 -8.62764359e-01
4.98354495e-01 -1.17861874e-01 5.19325197e-01 -6.36661291e-01
-2.94433445e-01 9.79687944e-02 1.87460840e-01 -9.34601009e-01
-3.61404240e-01 -9.01209354e-01 -9.89119053e-01 -1.04934573e-01
-1.59391731e-01 8.59830305e-02 5.04814029e-01 7.90972531e-01
2.40767345e-01 5.78660250e-01 7.29292154e-01 -9.96385217e-01
-6.83134913e-01 -7.20563173e-01 -7.28141189e-01 4.17381316e-01
7.51772106e-01 -8.38541687e-01 -6.88230872e-01 -1.35781407e-01] | [15.004447937011719, 5.26356840133667] |
ede509cb-5b68-4ca6-b7e8-5d13aa8f3943 | cross-lingual-semantic-representation-for-nlp | null | null | https://aclanthology.org/2020.coling-tutorials.1 | https://aclanthology.org/2020.coling-tutorials.1.pdf | Cross-lingual Semantic Representation for NLP with UCCA | This is an introductory tutorial to UCCA (Universal Conceptual Cognitive Annotation), a cross-linguistically applicable framework for semantic representation, with corpora annotated in English, German and French, and ongoing annotation in Russian and Hebrew. UCCA builds on extensive typological work and supports rapid annotation. The tutorial will provide a detailed introduction to the UCCA annotation guidelines, design philosophy and the available resources; and a comparison to other meaning representations. It will also survey the existing parsing work, including the findings of three recent shared tasks, in SemEval and CoNLL, that addressed UCCA parsing. Finally, the tutorial will present recent applications and extensions to the scheme, demonstrating its value for natural language processing in a range of languages and domains. | ['Nathan Schneider', 'Jakob Prange', 'Daniel Hershcovich', 'Dotan Dvir', 'Omri Abend'] | 2020-12-01 | null | null | null | coling-2020-8 | ['ucca-parsing'] | ['natural-language-processing'] | [ 2.63689160e-01 6.08397424e-01 -4.33672160e-01 -5.10342836e-01
-7.24839985e-01 -1.10791850e+00 5.99033713e-01 7.15128243e-01
-5.44696510e-01 7.33575463e-01 9.76393938e-01 -3.30434978e-01
-2.68648654e-01 -4.94339705e-01 8.52015987e-02 -1.87568337e-01
2.41910443e-01 6.78392112e-01 7.97245726e-02 -5.71269810e-01
4.59731758e-01 4.32714187e-02 -1.49524629e+00 7.18568504e-01
4.92804766e-01 6.66254163e-01 5.20965099e-01 4.59367871e-01
-7.53187060e-01 1.02796459e+00 -8.67582798e-01 -4.97788399e-01
-5.71571410e-01 -3.93921286e-01 -1.73370397e+00 -1.69073567e-01
2.94512361e-01 4.93022889e-01 3.24219078e-01 9.42934930e-01
3.94979239e-01 4.34555143e-01 2.36792892e-01 -6.50616467e-01
-1.15884316e+00 1.08505344e+00 8.10508355e-02 4.21496272e-01
1.02995229e+00 -7.04339027e-01 1.30072486e+00 -7.23389924e-01
1.15512252e+00 1.55368865e+00 8.96591067e-01 1.19984663e+00
-8.56361032e-01 -2.88769513e-01 4.41832185e-01 1.19170453e-02
-9.97334480e-01 -2.64665753e-01 3.02369058e-01 -2.70411819e-01
1.71829283e+00 6.86964020e-02 7.39967823e-01 9.45415556e-01
-2.65039414e-01 8.95223022e-01 9.95082021e-01 -1.11822546e+00
1.10137813e-01 -1.36478379e-01 8.70896697e-01 3.45709622e-01
1.04857825e-01 -1.88114882e-01 -5.38564801e-01 -2.66879350e-01
6.35588586e-01 -6.95194483e-01 -2.61679385e-02 -1.40287608e-01
-1.16281283e+00 8.50727737e-01 4.69099581e-02 9.16845679e-01
3.66511010e-02 2.99169630e-01 1.07894659e+00 2.03783110e-01
3.19033355e-01 6.36474550e-01 -9.06194806e-01 -3.20045829e-01
-3.28466564e-01 4.24551040e-01 6.61139607e-01 1.42275572e+00
1.17235400e-01 9.23290029e-02 -5.65908616e-04 1.37749302e+00
3.67808104e-01 2.76912451e-01 8.18318129e-01 -1.17876732e+00
3.10138404e-01 4.90836859e-01 1.38383545e-02 -4.78574544e-01
-7.48188257e-01 6.96867704e-02 1.03178874e-01 -2.69531399e-01
3.31567705e-01 -2.13500150e-02 -5.42100906e-01 1.73538303e+00
-1.08210901e-02 -6.64982796e-01 7.19117701e-01 6.41116917e-01
1.29677582e+00 6.26953989e-02 1.32852805e+00 -2.05812931e-01
2.13974881e+00 -7.03198969e-01 -1.19175994e+00 -5.45880854e-01
1.10276484e+00 -8.68404567e-01 1.17553294e+00 -1.23192996e-01
-1.14465857e+00 -4.71267194e-01 -8.25716913e-01 -7.31109321e-01
-1.01798511e+00 2.20082164e-01 1.19330728e+00 9.96105492e-01
-1.05269158e+00 2.86726832e-01 -5.92060030e-01 -1.10564899e+00
3.10532421e-01 -4.10982110e-02 -2.90105402e-01 1.97748630e-03
-1.44997978e+00 1.28906977e+00 1.30908728e+00 -2.92612940e-01
-5.22865318e-02 -4.05729592e-01 -1.23618186e+00 -2.09304065e-01
4.98730779e-01 -5.56761503e-01 1.57963288e+00 -1.07939482e+00
-1.12248003e+00 1.62211692e+00 -1.66217446e-01 -3.42075139e-01
-2.92237878e-01 -5.82260609e-01 -7.30861604e-01 9.64248255e-02
5.04113078e-01 8.98548663e-01 -1.45028338e-01 -8.80749643e-01
-8.45825374e-01 -4.96005028e-01 4.32692528e-01 6.90255284e-01
-1.48881096e-02 9.62497532e-01 -2.51492947e-01 -1.04735851e+00
1.39207795e-01 -5.96049845e-01 6.65261298e-02 -5.18921375e-01
2.59993434e-01 -8.84494603e-01 3.45104098e-01 -6.18567109e-01
1.37699652e+00 -1.94107807e+00 5.61803468e-02 -4.12625432e-01
-1.30367532e-01 5.85412271e-02 8.63576233e-02 6.75246716e-01
-2.97930539e-01 4.06245232e-01 -3.88479501e-01 -1.04337648e-01
1.00662500e-01 8.73319328e-01 -2.49423102e-01 -1.08654603e-01
-6.92231283e-02 1.09657300e+00 -1.41377068e+00 -4.42403823e-01
3.79482508e-01 2.11264879e-01 -2.01210871e-01 -2.99922913e-01
-1.75395682e-01 1.29882351e-01 -3.66149187e-01 6.40861392e-01
2.11918741e-01 2.19331548e-01 7.93704927e-01 2.02655002e-01
-3.61360133e-01 8.18190217e-01 -7.32434332e-01 2.31867433e+00
-3.94663185e-01 5.48679054e-01 -3.49267237e-02 -8.35046470e-01
6.94073021e-01 8.77998710e-01 -1.54156908e-01 -3.87372315e-01
1.70791462e-01 3.48984182e-01 -2.18137607e-01 -2.02341616e-01
6.81000590e-01 -4.18534786e-01 -7.11589456e-01 3.19669813e-01
4.23616260e-01 -3.20129693e-01 4.74688232e-01 1.86580822e-01
6.82953477e-01 6.18400514e-01 1.22631335e+00 -9.78074133e-01
8.94877076e-01 4.96643066e-01 4.75507319e-01 5.24255931e-01
-3.56487155e-01 7.10758045e-02 3.59611988e-01 -6.03457391e-01
-8.07596207e-01 -7.65240610e-01 -6.77244723e-01 1.59451592e+00
-6.16235994e-02 -1.05665743e+00 -1.01764619e+00 -7.99373507e-01
-4.58357900e-01 1.24287403e+00 -7.71128953e-01 3.23223889e-01
-6.48855627e-01 -6.96709633e-01 9.69984472e-01 1.10849762e+00
3.00939083e-01 -1.62700915e+00 -8.81158888e-01 2.48056546e-01
-5.32350123e-01 -1.26772439e+00 2.84590751e-01 2.43566900e-01
-7.56522179e-01 -1.47480345e+00 -8.93775225e-02 -1.24437916e+00
2.60778755e-01 6.21012002e-02 1.33906507e+00 2.86604762e-01
-5.24347648e-02 8.23693097e-01 -8.83785605e-01 -5.75889289e-01
-4.21212971e-01 -1.17511310e-01 -3.45760494e-01 -1.38820469e+00
1.14657032e+00 9.43578407e-02 1.76202253e-01 -8.49508420e-02
-6.98205709e-01 -8.54887590e-02 -2.09525436e-01 9.40754950e-01
5.30967593e-01 -2.28707194e-01 7.91089416e-01 -1.40141618e+00
8.59589815e-01 -4.28984195e-01 -5.39344773e-02 3.76016736e-01
-2.96563506e-01 -2.70021856e-01 8.99880007e-02 3.13008249e-01
-1.66328263e+00 -3.44921462e-02 -6.01194799e-01 5.21794200e-01
-5.31991899e-01 7.01670468e-01 -2.91451246e-01 3.04548323e-01
8.13939810e-01 -3.00448447e-01 -3.21597964e-01 -5.64400554e-01
8.60225439e-01 3.98897439e-01 7.65033126e-01 -1.23381114e+00
-2.22485572e-01 1.71238393e-01 -3.87020707e-01 -9.88376617e-01
-1.24833906e+00 -6.28788054e-01 -1.21876049e+00 3.27109098e-02
1.29580188e+00 -8.88645411e-01 -1.43293485e-01 -5.64161502e-02
-1.34035766e+00 -2.26539567e-01 -6.89309776e-01 2.59722292e-01
-8.94035161e-01 4.68176126e-01 -7.07784891e-01 -4.60458815e-01
-4.27307695e-01 -5.39819479e-01 1.04766846e+00 3.41093950e-02
-7.69587934e-01 -1.63884115e+00 7.98930507e-03 4.48133290e-01
7.09638894e-02 2.86107868e-01 1.27912593e+00 -9.36370432e-01
5.53912640e-01 1.49202928e-01 -2.02752024e-01 1.17417693e-01
-8.64403918e-02 -5.51221490e-01 -8.15363169e-01 9.79797319e-02
-1.42350510e-01 -8.35034728e-01 4.98706788e-01 1.18176647e-01
7.83521593e-01 2.98570812e-01 -3.26304138e-01 -1.06401388e-02
1.27648830e+00 4.28316683e-01 4.73368019e-01 5.97617984e-01
3.92008126e-01 1.04961002e+00 9.67750907e-01 -1.01851918e-01
4.93169725e-01 4.38597888e-01 -1.22103229e-01 5.22348642e-01
-6.21651769e-01 -1.12041004e-01 1.67608634e-01 8.42895985e-01
-3.91442358e-01 -2.21551761e-01 -1.08059347e+00 6.95011675e-01
-2.06459951e+00 -7.63348699e-01 -1.45470858e-01 1.73443437e+00
6.60135150e-01 -2.37542361e-01 -1.98666498e-01 -5.57878949e-02
8.59598100e-01 1.24047473e-01 3.71803902e-02 -8.39100599e-01
-5.82037508e-01 6.48454249e-01 1.36365388e-02 4.58875328e-01
-1.27025497e+00 1.95977545e+00 8.08620548e+00 6.46264672e-01
-1.90460950e-01 6.59155428e-01 6.08170293e-02 5.37498832e-01
-1.91389039e-01 1.91235170e-01 -7.03044236e-01 -2.26517305e-01
1.14757061e+00 -2.47105017e-01 1.64795637e-01 8.35118890e-01
-3.29360038e-01 7.28177279e-02 -8.53654087e-01 6.34339213e-01
3.73881966e-01 -1.26683676e+00 1.29610360e-01 -6.15503192e-01
1.70355290e-01 9.29012522e-02 -7.25830317e-01 6.00161314e-01
6.89023077e-01 -7.23396003e-01 9.76838291e-01 -1.20286800e-01
9.96349394e-01 -6.11612201e-01 9.43990767e-01 -2.24420562e-01
-1.44009197e+00 -8.43244642e-02 -7.76201248e-01 -2.85813749e-01
4.88269061e-01 -3.45131487e-01 5.77654615e-02 6.97277367e-01
8.69332850e-01 8.79037440e-01 -5.43037593e-01 3.54812920e-01
-9.16581929e-01 3.25936615e-01 2.44398639e-01 -1.20234592e-02
3.03129852e-01 6.42644316e-02 5.16267717e-01 1.89227736e+00
-6.09059818e-02 7.92551994e-01 4.33410943e-01 3.60310704e-01
2.44818807e-01 6.08463466e-01 -3.50281775e-01 -2.39105448e-01
7.47245193e-01 9.66445446e-01 -1.10600269e+00 -6.12969458e-01
-7.54150212e-01 9.30494487e-01 5.36231399e-01 -8.58315900e-02
-4.50526655e-01 -4.63649660e-01 5.68543196e-01 -3.29579830e-01
-9.95213836e-02 -1.37378395e-01 -3.09989482e-01 -1.05904889e+00
-5.03344476e-01 -4.99103934e-01 1.20567977e+00 -1.04093039e+00
-1.24962413e+00 6.63155854e-01 4.02390271e-01 -6.23971045e-01
-2.49945328e-01 -1.10476184e+00 -9.83310714e-02 7.02052057e-01
-1.02758956e+00 -1.47685885e+00 -3.09434850e-02 4.58838761e-01
9.01036143e-01 -1.14457160e-01 1.63769937e+00 2.63348967e-02
-1.94714308e-01 1.43404365e-01 -4.61733490e-01 1.61385953e-01
6.06202781e-01 -1.71815908e+00 7.23205745e-01 4.22730565e-01
4.21417058e-02 9.77331817e-01 3.00942719e-01 -8.48162293e-01
-6.26209259e-01 -7.06963718e-01 1.42565918e+00 -8.31073701e-01
8.93700480e-01 -2.40721494e-01 -7.72606194e-01 1.33544528e+00
6.99246287e-01 -3.11442643e-01 1.17304099e+00 4.98005897e-01
-3.07068110e-01 8.66280735e-01 -1.26020122e+00 3.12370211e-01
1.58198476e+00 -6.48161352e-01 -1.69295299e+00 3.32169175e-01
8.95659089e-01 -6.06429040e-01 -1.07373655e+00 -1.68541614e-02
6.15208030e-01 -4.49639082e-01 7.75782585e-01 -9.87127423e-01
6.75253421e-02 -2.79420614e-01 -3.51497322e-01 -1.23697531e+00
-5.09501278e-01 -3.20502549e-01 4.91224021e-01 1.32674742e+00
3.13709199e-01 -4.84066278e-01 1.74129203e-01 6.31018519e-01
-8.70311379e-01 9.70639512e-02 -9.49073613e-01 -3.50339383e-01
4.73978221e-01 -9.84130681e-01 5.05302191e-01 1.36386311e+00
9.19366062e-01 7.55983770e-01 2.91602492e-01 -4.11834091e-01
3.52943331e-01 -2.12728992e-01 3.03149372e-02 -1.55216348e+00
1.68814242e-01 -4.10749733e-01 -5.20268500e-01 -8.04479897e-01
6.50959253e-01 -1.30458045e+00 -2.15629369e-01 -1.76211452e+00
-2.81473082e-02 -3.54428589e-01 -2.22857803e-01 9.36352432e-01
4.24125604e-02 1.03272580e-01 2.29263708e-01 1.11493878e-01
-6.01121485e-01 1.08677417e-01 7.83629179e-01 3.77933413e-01
-1.05976418e-01 -6.01689696e-01 -1.18457627e+00 1.16970563e+00
7.92135000e-01 -2.76340961e-01 -5.05870461e-01 -7.31000066e-01
4.32308525e-01 -2.88870573e-01 -1.87932327e-01 -7.31781840e-01
-2.22474337e-02 -1.05588786e-01 2.89305568e-01 -4.34728503e-01
9.38573629e-02 -5.23998976e-01 -2.33162001e-01 9.76859704e-02
-6.11730695e-01 3.95003021e-01 6.76252663e-01 2.66106397e-01
-3.23590070e-01 -8.30020130e-01 6.56739175e-01 -6.79250121e-01
-1.60106814e+00 -6.34698451e-01 -8.88016582e-01 4.47531402e-01
1.03585637e+00 -2.08826780e-01 -5.77839077e-01 1.80072978e-01
-1.29470694e+00 3.75108421e-01 2.63724416e-01 8.61207366e-01
3.51295590e-01 -1.28777051e+00 -2.92487949e-01 -2.79790252e-01
6.29029036e-01 -3.47130239e-01 1.29482657e-01 2.48811513e-01
-4.52513337e-01 1.00184906e+00 -3.87920290e-01 1.19314799e-02
-1.28711104e+00 5.76938272e-01 8.92830938e-02 9.24139768e-02
-7.64564693e-01 6.24140799e-01 7.23948404e-02 -8.20658624e-01
-2.24955138e-02 -2.20526919e-01 -1.11746120e+00 1.38104439e-01
7.55463183e-01 3.72596264e-01 -5.49110305e-03 -1.12084937e+00
-5.20841658e-01 5.19224942e-01 1.26405671e-01 -2.39681065e-01
9.04169679e-01 -6.27955675e-01 -6.91814601e-01 6.86538160e-01
5.47207534e-01 7.92092979e-02 -2.80637950e-01 -1.77723467e-01
8.00610065e-01 5.39110005e-02 -1.02088526e-02 -1.16195953e+00
-3.99461240e-01 5.94127417e-01 3.57636571e-01 1.12813517e-01
8.34192514e-01 5.46109736e-01 5.20805120e-01 5.72049320e-01
4.38747495e-01 -1.58712828e+00 -4.94817019e-01 1.01830626e+00
9.80461657e-01 -5.11158288e-01 -8.60640872e-03 -9.38877463e-01
-1.02605009e+00 1.39999938e+00 5.33403337e-01 7.10982531e-02
6.77829862e-01 1.12456463e-01 2.71510988e-01 -7.84916341e-01
-6.62697494e-01 -5.21482527e-01 -6.13117851e-02 1.02466559e+00
1.34112656e+00 4.47437257e-01 -1.14310467e+00 1.06837761e+00
-6.79510951e-01 -3.08404863e-01 5.28812230e-01 1.29079640e+00
-5.67248821e-01 -1.59297347e+00 -1.26507714e-01 4.12693359e-02
-7.67426431e-01 -3.95479083e-01 -8.36011827e-01 1.22056854e+00
3.77671510e-01 9.92239535e-01 4.95150387e-01 3.39034110e-01
4.77391183e-01 6.82748914e-01 4.84176278e-01 -1.30544448e+00
-9.64646280e-01 1.61679804e-01 9.83073235e-01 -5.28141677e-01
-1.28279793e+00 -1.15208459e+00 -1.93404973e+00 1.72222421e-01
-2.07751930e-01 4.85392690e-01 4.53647524e-01 1.06641972e+00
-2.12866790e-03 3.97424817e-01 -7.43679404e-01 -3.90437812e-01
2.89550185e-01 -1.08560479e+00 -4.95615333e-01 3.32740098e-01
-5.06992698e-01 -8.49707007e-01 1.01252414e-01 2.22729847e-01] | [10.29318904876709, 9.457542419433594] |
e19855bf-9078-4110-8a32-ecce5199480d | pyss3-a-python-package-implementing-a-novel | 1912.09322 | null | https://arxiv.org/abs/1912.09322v2 | https://arxiv.org/pdf/1912.09322v2.pdf | PySS3: A Python package implementing a novel text classifier with visualization tools for Explainable AI | A recently introduced text classifier, called SS3, has obtained state-of-the-art performance on the CLEF's eRisk tasks. SS3 was created to deal with risk detection over text streams and, therefore, not only supports incremental training and classification but also can visually explain its rationale. However, little attention has been paid to the potential use of SS3 as a general classifier. We believe this could be due to the unavailability of an open-source implementation of SS3. In this work, we introduce PySS3, a package that implements SS3 and also comes with visualization tools that allow researchers to deploy robust, explainable, and trusty machine learning models for text classification. | ['Manuel Montes-y-Gómez', 'Sergio G. Burdisso', 'Marcelo Errecalde'] | 2019-12-19 | null | null | null | null | ['text-categorization'] | ['natural-language-processing'] | [-3.77211630e-01 2.22219571e-01 -4.48394986e-03 -4.11958396e-01
-4.26361442e-01 -4.09819394e-01 6.47099972e-01 8.46367538e-01
-1.00474305e-01 5.00537634e-01 3.82939540e-02 -8.47757936e-01
-1.10323668e-01 -5.51885128e-01 -3.44131351e-01 -2.54089147e-01
-1.86357945e-02 3.61598223e-01 2.40852043e-01 1.97410896e-01
2.57088333e-01 5.48579097e-01 -1.57085049e+00 7.73308098e-01
5.15048325e-01 8.04801404e-01 -2.95550048e-01 9.72949922e-01
-2.32283399e-01 6.67732954e-01 -9.73375082e-01 -7.44774401e-01
-3.86959404e-01 -3.22008103e-01 -8.22172403e-01 -3.21620494e-01
3.24037582e-01 -4.34488580e-02 5.97149059e-02 5.54873466e-01
3.39413345e-01 -3.64653200e-01 7.34973013e-01 -1.63375747e+00
-2.68616021e-01 8.21886480e-01 -4.17396009e-01 2.45645240e-01
5.53272843e-01 -1.29851744e-01 9.49874580e-01 -6.67071760e-01
6.06132209e-01 1.14198029e+00 7.97925174e-01 5.09893775e-01
-1.14289463e+00 -6.85961902e-01 1.54968023e-01 4.54186589e-01
-1.15988410e+00 -5.17221950e-02 4.21428531e-01 -4.54175025e-01
1.05380225e+00 7.81478941e-01 7.91306794e-01 1.38220012e+00
3.10055315e-01 1.19313157e+00 1.33775532e+00 -7.34553218e-01
2.89942592e-01 7.28726327e-01 6.60385430e-01 9.10045981e-01
4.57192719e-01 -3.12237013e-02 -8.63052845e-01 -5.89996994e-01
1.60644837e-02 -2.10501090e-01 -2.63981909e-01 -1.82378978e-01
-9.64074314e-01 8.77009273e-01 1.37455463e-01 1.29109785e-01
6.26665652e-02 -1.59693971e-01 6.43594742e-01 1.75674155e-01
8.28996301e-01 1.23398274e-01 -4.24910307e-01 -5.95925033e-01
-9.47324038e-01 2.58188218e-01 1.18544447e+00 7.38537967e-01
2.43991464e-02 -1.70024082e-01 3.30410935e-02 4.29332614e-01
5.69445908e-01 1.92708343e-01 2.27679476e-01 -2.89221972e-01
2.26816967e-01 8.12653661e-01 -1.45414487e-01 -7.94889271e-01
-7.95943379e-01 -3.06883782e-01 -5.83711147e-01 5.54404140e-01
3.86597395e-01 2.63783205e-02 -5.33849537e-01 8.90668690e-01
3.22761685e-01 -1.03297435e-01 6.52422532e-02 5.26018441e-01
1.13114738e+00 2.49335438e-01 1.29191548e-01 1.02038920e-01
1.32647932e+00 -7.75634408e-01 -7.95522451e-01 -8.79977364e-03
9.45686638e-01 -6.97175860e-01 1.00249827e+00 7.98017740e-01
-9.51634049e-01 -3.72193083e-02 -1.18028724e+00 1.24708571e-01
-8.46235514e-01 1.59996629e-01 7.43874550e-01 1.20424366e+00
-5.50867200e-01 6.63407803e-01 -9.42443967e-01 -7.09312081e-01
5.71670234e-01 6.45730644e-02 -3.82735729e-01 2.34549120e-01
-9.40019369e-01 9.93446767e-01 1.98253691e-01 -1.55536175e-01
-2.56706119e-01 -4.27548975e-01 -5.89861035e-01 2.14001775e-01
5.86583056e-02 -5.58288813e-01 1.25420892e+00 -5.60887575e-01
-1.16890788e+00 9.90728199e-01 -1.56483591e-01 -4.30435121e-01
8.66448998e-01 -2.23102182e-01 -5.80818713e-01 1.53544649e-01
-3.46280903e-01 1.85620129e-01 7.92596340e-01 -1.01270390e+00
-6.18077219e-01 -2.93432355e-01 -3.68936092e-01 -1.66845694e-01
-3.29313397e-01 6.08116508e-01 -1.63009927e-01 -7.52932608e-01
-3.41637999e-01 -6.64673150e-01 3.19275439e-01 2.47483402e-01
-4.96517748e-01 -4.95028287e-01 9.78717446e-01 -5.60204685e-01
1.41922998e+00 -2.05657339e+00 -3.16012800e-01 2.77228475e-01
1.72359437e-01 5.14950037e-01 3.81531596e-01 6.22420549e-01
-2.11463764e-01 5.98527133e-01 -4.47430201e-02 -4.83933598e-01
3.28792036e-01 5.51094152e-02 -3.95144522e-01 4.15121943e-01
1.42660409e-01 6.02965176e-01 -5.40416539e-01 -5.77502072e-01
3.82492185e-01 4.05763745e-01 -2.86041707e-01 4.55458574e-02
-1.77858174e-01 -2.16924191e-01 -3.17255259e-01 4.95162815e-01
7.19214261e-01 -3.30784589e-01 9.88876149e-02 3.75049233e-01
-9.24954042e-02 2.95562416e-01 -1.28396046e+00 1.05042315e+00
-9.71194878e-02 9.69168067e-01 2.34243320e-03 -7.12714851e-01
8.01706374e-01 3.31249148e-01 9.01576802e-02 1.58325970e-01
1.55422166e-01 -1.98560506e-02 -5.13278723e-01 -6.40481532e-01
3.19648862e-01 1.46101460e-01 2.68912226e-01 7.88412750e-01
-1.35612026e-01 4.47709449e-02 4.86647077e-02 5.80888331e-01
1.06971729e+00 2.85944372e-01 5.81487477e-01 -2.00997330e-02
2.85134733e-01 1.24553762e-01 -2.00937480e-01 8.30511570e-01
-3.92557122e-02 6.51148975e-01 7.80391037e-01 -5.36766589e-01
-5.65298200e-01 -8.86547685e-01 -5.88141024e-01 1.01035714e+00
-3.06806117e-01 -9.75441396e-01 -6.97479129e-01 -1.45448434e+00
2.60403544e-01 1.23722708e+00 -8.50864649e-01 -7.25270659e-02
1.03499755e-01 -8.95752609e-01 7.35439122e-01 5.24738193e-01
3.00260097e-01 -1.17924893e+00 -7.70231366e-01 5.53584173e-02
2.89721857e-03 -7.92854309e-01 -6.55231774e-02 4.04399842e-01
-6.94216669e-01 -1.45381117e+00 -2.17140898e-01 -1.40243828e-01
6.27320290e-01 1.31710440e-01 9.62469637e-01 4.08320338e-01
-4.88285184e-01 5.28856933e-01 -7.81439483e-01 -1.05081069e+00
-7.21918523e-01 9.30859372e-02 -3.55001360e-01 -2.78652519e-01
7.52271712e-01 2.06317142e-01 -1.61004424e-01 4.71360296e-01
-1.01013136e+00 1.16481870e-01 1.05039626e-01 8.06850672e-01
-1.11407503e-01 1.07032999e-01 5.01362085e-01 -1.15199840e+00
7.96902061e-01 -4.77184087e-01 -3.19963127e-01 3.83872449e-01
-1.18501210e+00 -1.09615192e-01 4.49470401e-01 -8.00033435e-02
-1.04345286e+00 -1.34090289e-01 -3.49825919e-01 2.01370381e-02
-5.70090234e-01 5.36488891e-01 2.51435220e-01 2.08175316e-01
8.45148802e-01 -1.40891373e-01 -5.33303432e-02 -5.58853924e-01
2.02643484e-01 1.20045304e+00 -1.82885081e-02 -2.23395616e-01
5.33687770e-01 3.98158282e-01 -1.38706148e-01 -8.46660554e-01
-8.76462519e-01 -4.71915036e-01 -3.88871461e-01 -2.07763836e-01
5.77136159e-01 -5.15861750e-01 -1.04364729e+00 6.06594622e-01
-1.08154762e+00 -2.90207237e-01 4.61538620e-02 2.68833190e-01
-2.96771467e-01 5.11425674e-01 -4.55305845e-01 -1.03455222e+00
-5.70985138e-01 -4.87045079e-01 7.70637810e-01 4.02937382e-02
-6.85315907e-01 -1.10177720e+00 -2.10844636e-01 3.47929120e-01
3.18714291e-01 1.85272694e-01 7.44554639e-01 -9.30046797e-01
-1.23592034e-01 -6.04203343e-01 -3.21927100e-01 1.54754028e-01
-2.77214527e-01 5.06705284e-01 -1.38294911e+00 -2.77315319e-01
-3.39513421e-01 -3.26440454e-01 7.57678807e-01 1.64019436e-01
1.38007617e+00 -1.83194116e-01 -5.89360595e-01 4.89929825e-01
7.49780715e-01 -7.75319487e-02 3.99294704e-01 7.70268977e-01
1.45505935e-01 7.62155235e-01 7.05928743e-01 7.04226434e-01
3.45571190e-01 6.48381054e-01 3.52845490e-01 -3.04028958e-01
1.53588384e-01 -3.51092488e-01 2.72101372e-01 2.44298697e-01
1.63332775e-01 -4.90976930e-01 -9.23763216e-01 -1.91939667e-01
-2.04110527e+00 -9.32809293e-01 -7.89484203e-01 1.92289877e+00
4.75727856e-01 3.20361286e-01 1.91069424e-01 6.32032335e-01
5.12976289e-01 -1.89767227e-01 -2.67308533e-01 -7.36531079e-01
-4.08292413e-02 9.60542858e-02 1.19214281e-01 2.29973838e-01
-1.12440622e+00 5.08740425e-01 7.23280144e+00 7.56817758e-01
-8.92896354e-01 -1.14854947e-01 5.68799496e-01 -7.17005804e-02
6.70601055e-02 -1.99940741e-01 -8.96784008e-01 3.66896957e-01
1.17669034e+00 -4.47219968e-01 6.07110001e-02 1.06463015e+00
2.44934291e-01 -2.89651245e-01 -1.05235338e+00 6.05251074e-01
1.51916921e-01 -1.34347677e+00 -1.72543734e-01 -1.59089595e-01
-1.58079788e-02 -2.25717649e-01 -6.57414347e-02 2.93575436e-01
1.97968155e-01 -1.00645113e+00 8.82878363e-01 2.78702468e-01
5.76663971e-01 -7.39462972e-01 9.62521255e-01 3.00215095e-01
-8.67293179e-01 1.91939250e-02 -1.09152272e-01 1.62038386e-01
-1.94127977e-01 5.92285395e-01 -1.24441671e+00 5.65983891e-01
1.18576086e+00 7.05787361e-01 -1.06674111e+00 1.17215490e+00
-3.77014428e-01 1.06068814e+00 -2.22577944e-01 -4.61494982e-01
-1.66086569e-01 4.84391451e-01 4.14995909e-01 1.76768279e+00
2.06329972e-01 -1.11012958e-01 -8.63340572e-02 5.77538013e-01
3.39234591e-01 4.57956493e-01 -4.92265373e-01 -5.31867780e-02
1.53269380e-01 1.32434356e+00 -9.28243041e-01 -3.67852867e-01
-5.98093092e-01 8.61976385e-01 3.07332128e-01 1.49637327e-01
-5.60994029e-01 -6.34946823e-01 2.34438121e-01 2.97874242e-01
1.19181857e-01 1.28407562e-02 -6.98946238e-01 -1.19600463e+00
-1.15758285e-01 -9.45811152e-01 9.59389448e-01 -1.13907111e+00
-1.30864871e+00 7.19779670e-01 2.56452322e-01 -1.01212776e+00
-1.53149247e-01 -9.05198812e-01 -7.06986964e-01 7.64346242e-01
-1.42867041e+00 -1.11767387e+00 -5.48861980e-01 4.85734493e-01
2.26080015e-01 -2.25380540e-01 1.19316494e+00 -2.97493905e-01
-7.26695001e-01 6.12052858e-01 1.11114159e-01 -1.46536008e-01
9.59248781e-01 -1.68938649e+00 4.71381307e-01 4.39764678e-01
2.03817204e-01 5.13010561e-01 9.58812952e-01 -6.51337802e-01
-8.28449428e-01 -8.57516050e-01 1.00684929e+00 -8.07883263e-01
7.71843493e-01 -5.50876856e-01 -1.18513429e+00 7.68702626e-01
2.83220738e-01 -2.28080913e-01 1.06489587e+00 4.75161940e-01
-4.85969007e-01 1.53191686e-01 -1.24389791e+00 5.11121154e-01
5.42719781e-01 -2.13314116e-01 -6.55803084e-01 4.86610830e-01
7.00807199e-02 -4.35820997e-01 -7.69754589e-01 -1.44218683e-01
5.25328040e-01 -1.29803169e+00 5.33848166e-01 -6.41659141e-01
3.30122709e-01 7.37353861e-02 4.41239148e-01 -1.37431514e+00
2.68103220e-02 -5.66975534e-01 -2.56146848e-01 1.26690173e+00
6.39782369e-01 -8.65770221e-01 9.21910942e-01 6.53023660e-01
-1.63124334e-02 -5.56502581e-01 -7.12487757e-01 -6.64165854e-01
2.26428717e-01 -8.89271557e-01 4.70735729e-01 1.13242722e+00
8.49510431e-01 1.24492556e-01 -2.59718239e-01 7.46401176e-02
4.57453072e-01 1.66958030e-02 7.13397920e-01 -1.51038623e+00
-1.27405882e-01 -5.50834894e-01 -3.84415418e-01 -4.69513208e-01
1.74660563e-01 -1.08310211e+00 -3.64478022e-01 -1.56004465e+00
3.17574739e-01 -3.63923073e-01 -4.33429889e-02 9.20184016e-01
-3.77689004e-01 5.41604832e-02 2.32975319e-01 1.07741214e-01
-4.90891039e-01 1.33401930e-01 6.74526811e-01 1.58718109e-01
-7.98713863e-02 3.82275373e-01 -8.21780264e-01 6.93585634e-01
1.11305630e+00 -7.97029853e-01 -5.00897504e-03 1.93131641e-01
3.81559342e-01 -3.08323175e-01 4.94807452e-01 -7.48644471e-01
1.43657476e-01 7.05936030e-02 7.49632120e-01 -7.94140339e-01
-5.20184413e-02 -8.49639237e-01 5.99396713e-02 4.25979823e-01
-5.27922094e-01 -5.49046583e-02 5.96536160e-01 4.81503934e-01
2.04124525e-02 -7.26535320e-01 6.81853712e-01 1.47385761e-01
-6.62073269e-02 -2.48432323e-01 -7.62564063e-01 -1.06526166e-01
1.14604795e+00 -1.14317112e-01 -7.50388563e-01 -3.30954343e-01
-7.48891532e-01 3.04923028e-01 4.32714671e-01 5.63214004e-01
5.25500894e-01 -8.59720767e-01 -7.14104712e-01 3.49974841e-01
3.96196872e-01 -5.04003882e-01 1.50931664e-02 4.29383159e-01
-5.77365637e-01 5.80772638e-01 1.16209380e-01 -4.72372711e-01
-1.96531701e+00 6.55147791e-01 2.00712755e-01 -2.02860162e-01
-1.10648072e+00 4.74574059e-01 -3.95883679e-01 -2.11228490e-01
6.46337748e-01 -1.62245244e-01 -3.61547112e-01 4.26630378e-02
8.45395863e-01 5.43589652e-01 3.46015394e-01 -1.25714034e-01
-4.76857334e-01 9.39800218e-02 -2.84364969e-01 -1.65001787e-02
1.10414422e+00 5.57317734e-02 1.77000061e-01 5.20617485e-01
6.76738679e-01 6.51372522e-02 -7.31977403e-01 1.42995536e-01
2.46390775e-01 -4.17209834e-01 1.93613216e-01 -1.39569652e+00
-6.05751276e-01 1.05369103e+00 5.33410490e-01 7.46505558e-01
7.65886366e-01 3.74494754e-02 2.24876225e-01 3.45719188e-01
-4.79132049e-02 -9.43542719e-01 -1.25176951e-01 3.04758400e-01
8.50692332e-01 -1.31880355e+00 2.55111426e-01 -4.27860171e-01
-9.74952996e-01 1.56238544e+00 3.13157737e-01 5.69554865e-01
7.98549831e-01 4.15770918e-01 3.23905736e-01 -3.38548124e-01
-1.21386933e+00 8.98444876e-02 4.32135373e-01 7.78361082e-01
7.81451523e-01 3.65424342e-02 -2.76877999e-01 6.32266998e-01
-2.54316300e-01 2.43534774e-01 7.87511766e-01 1.11318934e+00
-4.42449868e-01 -1.08901143e+00 -6.48486674e-01 6.23375416e-01
-6.25835180e-01 -3.94173600e-02 -6.47467613e-01 9.40770626e-01
-8.90652835e-02 1.16144741e+00 -1.28110141e-01 -2.05016539e-01
4.34658706e-01 3.84192854e-01 1.54536128e-01 -5.61794937e-01
-9.78957295e-01 -1.08188957e-01 4.56547767e-01 -2.42996663e-01
-8.61675590e-02 -8.66606891e-01 -1.31290984e+00 -5.65053642e-01
-3.51008624e-01 4.63792264e-01 1.08974755e+00 7.18486786e-01
2.90360898e-01 3.87510896e-01 3.00978243e-01 -4.21759903e-01
-4.41854805e-01 -9.29113507e-01 -8.38180959e-01 3.14551353e-01
-3.25348377e-02 -4.49806780e-01 -7.05642939e-01 -1.09629072e-01] | [8.758319854736328, 7.395889759063721] |
018bbbbf-e959-4499-94c6-ea05e916d4ed | physnlu-a-language-resource-for-evaluating | 2201.04275 | null | https://arxiv.org/abs/2201.04275v3 | https://arxiv.org/pdf/2201.04275v3.pdf | PhysNLU: A Language Resource for Evaluating Natural Language Understanding and Explanation Coherence in Physics | In order for language models to aid physics research, they must first encode representations of mathematical and natural language discourse which lead to coherent explanations, with correct ordering and relevance of statements. We present a collection of datasets developed to evaluate the performance of language models in this regard, which measure capabilities with respect to sentence ordering, position, section prediction, and discourse coherence. Analysis of the data reveals equations and sub-disciplines which are most common in physics discourse, as well as the sentence-level frequency of equations and expressions. We present baselines that demonstrate how contemporary language models are challenged by coherence related tasks in physics, even when trained on mathematical natural language objectives. | ['Andre Freitas', 'Zili Zhou', 'Jordan Meadows'] | 2022-01-12 | null | https://aclanthology.org/2022.lrec-1.492 | https://aclanthology.org/2022.lrec-1.492.pdf | lrec-2022-6 | ['sentence-ordering'] | ['natural-language-processing'] | [ 1.20547041e-01 5.88557065e-01 -1.93232819e-01 -5.32465875e-01
-5.48621178e-01 -7.88717330e-01 1.19833291e+00 9.62783158e-01
1.95563897e-01 6.68319821e-01 1.16123033e+00 -8.80985260e-01
-4.76428956e-01 -8.24260235e-01 -7.49907494e-01 2.72387359e-02
-6.52225092e-02 5.84447980e-01 -5.24418131e-02 -4.63532925e-01
8.35604906e-01 3.57726008e-01 -1.06473851e+00 6.97001159e-01
1.16816521e+00 2.72383153e-01 1.92770287e-01 1.01443708e+00
-6.57700002e-01 1.89224327e+00 -5.71293294e-01 -3.37998927e-01
-5.19356906e-01 -6.73156142e-01 -1.26821339e+00 -2.00996548e-01
7.94123471e-01 -1.14850827e-01 -6.32686973e-01 8.17283750e-01
-1.25766560e-01 3.19179505e-01 9.84983146e-01 -5.91026366e-01
-1.14254558e+00 1.17493391e+00 4.45250235e-02 7.68384516e-01
1.12231183e+00 1.51939198e-01 1.43466413e+00 -4.47152227e-01
8.40219676e-01 1.67662621e+00 3.87449324e-01 3.52854967e-01
-1.51206529e+00 -1.77529022e-01 2.59715974e-01 5.02974808e-01
-5.63942015e-01 -5.34543335e-01 6.77732527e-01 -7.45911121e-01
1.41277432e+00 5.22863030e-01 7.34750688e-01 1.12419748e+00
7.65748084e-01 3.95306855e-01 1.05213439e+00 -4.55694377e-01
1.40742317e-01 2.96161789e-02 1.07419980e+00 6.41286612e-01
2.28504509e-01 4.52597924e-05 -8.64474058e-01 6.14970364e-02
3.79055351e-01 -6.90559626e-01 -3.46338689e-01 3.20809036e-01
-1.30377686e+00 8.33420932e-01 4.17077959e-01 6.09860599e-01
-4.28803474e-01 5.39923832e-02 8.26931186e-03 2.96367347e-01
4.72258866e-01 1.29388893e+00 -8.53610709e-02 -5.55484369e-02
-8.65975797e-01 1.00329375e+00 1.16234601e+00 7.30290473e-01
2.77080476e-01 -3.44526246e-02 -5.50977528e-01 2.66131461e-01
2.80713737e-01 2.89035857e-01 1.84154406e-01 -1.42873037e+00
6.24445558e-01 6.94647253e-01 -6.10210374e-02 -1.59895945e+00
-4.83335286e-01 -5.01874030e-01 -5.57873011e-01 -4.36483204e-01
2.59693563e-01 4.25533086e-01 -2.78892100e-01 1.76839769e+00
-1.87978476e-01 1.12508111e-01 2.33042449e-01 5.85120320e-01
1.37466490e+00 1.05979478e+00 2.54038334e-01 -4.76675898e-01
1.31395698e+00 -5.05719841e-01 -1.06378174e+00 -3.95477802e-01
6.86694443e-01 -1.02540171e+00 1.10952497e+00 2.37877853e-02
-1.64862156e+00 -6.64724052e-01 -8.77764821e-01 -6.21580899e-01
7.00068399e-02 -3.74771386e-01 5.70031047e-01 7.51527250e-02
-1.06481957e+00 8.08796883e-01 -5.71054041e-01 -3.59985121e-02
2.64649600e-01 -2.19599023e-01 1.74619004e-01 2.34066814e-01
-1.22018015e+00 1.32470500e+00 4.50664043e-01 -3.41847330e-01
-3.04698437e-01 -1.36572337e+00 -9.74698246e-01 4.76573855e-01
1.16692238e-01 -7.83008933e-01 1.44481838e+00 -1.82591692e-01
-1.14275932e+00 9.81085598e-01 -6.65546894e-01 -7.95553148e-01
-1.69001091e-02 -1.67302176e-01 -2.43951216e-01 2.87436694e-01
2.31463984e-01 4.06964749e-01 -1.39112100e-01 -1.06782627e+00
-1.64973766e-01 2.41168607e-02 4.69691396e-01 5.51718511e-02
-1.43371880e-01 -6.32712916e-02 3.71921062e-01 -3.53125721e-01
3.22407305e-01 -4.13460761e-01 -1.77184716e-01 -4.80401993e-01
-6.70770109e-01 -9.36696947e-01 4.21350598e-02 -9.27167833e-01
1.57154167e+00 -1.84484601e+00 5.12792826e-01 -1.74357742e-02
5.57378769e-01 -5.77900290e-01 4.50838134e-02 4.76110041e-01
-2.20109239e-01 5.92980742e-01 -2.23952942e-02 -2.07445741e-01
4.68426555e-01 1.69092759e-01 -9.96444106e-01 2.08861664e-01
3.62641484e-01 1.25654519e+00 -8.59219193e-01 -7.30808973e-01
2.73977071e-01 -9.15154889e-02 -7.93072104e-01 3.71688604e-01
-9.33781385e-01 4.64615047e-01 -4.72102463e-01 1.26922518e-01
1.20820567e-01 -8.99600267e-01 5.95320836e-02 -3.22012529e-02
-3.68833363e-01 1.68363273e+00 -4.79628980e-01 1.41414058e+00
-1.48120567e-01 1.20634413e+00 -5.92316091e-01 -8.81098211e-01
6.13526702e-01 2.71529526e-01 2.31082618e-01 -8.55487049e-01
-1.02955112e-02 -1.98202968e-01 4.76509035e-01 -7.76746690e-01
7.14499116e-01 6.57112058e-03 -8.64994451e-02 7.10262418e-01
-1.97917640e-01 -7.72181869e-01 6.73731327e-01 9.94636655e-01
8.13894212e-01 -2.68451005e-01 3.41208071e-01 -7.79769301e-01
6.32278681e-01 4.00551528e-01 -1.08702898e-01 8.98080528e-01
2.40879804e-01 1.69824094e-01 7.23056018e-01 -5.55392146e-01
-1.12914026e+00 -1.04318357e+00 -2.09670529e-01 7.32945621e-01
5.17458543e-02 -9.46932137e-01 -3.75008076e-01 1.87107757e-01
4.70351614e-03 1.86883366e+00 -5.85782945e-01 -2.23745941e-03
-1.03329384e+00 -1.69492602e-01 9.39475521e-02 1.86782137e-01
6.51895702e-02 -9.40956831e-01 -6.53658569e-01 2.18529254e-01
-4.72241431e-01 -1.46619833e+00 9.61626545e-02 -1.95768073e-01
-8.61730337e-01 -1.11612499e+00 1.11715376e-01 -4.70153242e-01
2.80614257e-01 1.06956556e-01 1.93102229e+00 4.00304168e-01
5.78845404e-02 4.40517783e-01 -1.45580038e-01 -4.45031375e-01
-1.04349685e+00 -1.03707761e-01 -2.51448363e-01 -9.86092627e-01
1.95685998e-01 -5.91560185e-01 -2.15874448e-01 -6.90524280e-01
-7.04462707e-01 4.27780092e-01 4.47389632e-02 4.70737696e-01
-1.01411842e-01 -3.73219162e-01 -9.64190140e-02 -7.26411045e-01
9.73570168e-01 -5.50292015e-01 -3.23937267e-01 5.56494892e-01
-5.07412910e-01 2.86372721e-01 4.41610336e-01 -5.21089956e-02
-8.76774788e-01 -8.40521574e-01 5.34909405e-03 2.21883669e-01
-7.05626011e-02 9.84790683e-01 4.89137471e-01 4.60224658e-01
9.42955852e-01 3.25512707e-01 -2.91105896e-01 -9.25497413e-02
6.57474875e-01 -4.26236540e-02 7.16608346e-01 -1.03916788e+00
6.70788765e-01 -1.82518587e-01 1.84237689e-01 -9.36547577e-01
-1.35570276e+00 7.63894692e-02 -5.06564200e-01 -5.13247140e-02
8.60627770e-01 -7.88142681e-01 -9.28235412e-01 -4.71819103e-01
-1.93369043e+00 -1.34435035e-02 -3.16400647e-01 5.00716388e-01
-3.12452823e-01 5.41874170e-01 -7.50334680e-01 -8.64536166e-01
-8.99542049e-02 -8.38295758e-01 5.65530837e-01 4.19447213e-01
-1.17271566e+00 -1.42723870e+00 5.91584854e-02 3.84484917e-01
3.46795589e-01 4.21714336e-01 1.87023795e+00 -6.87549591e-01
-9.08825040e-01 4.74265188e-01 4.06619981e-02 -3.99616420e-01
-5.06052971e-02 2.02910975e-01 -5.69013894e-01 3.12093109e-01
1.32785663e-01 -4.80573535e-01 8.02648365e-01 6.34182572e-01
1.06292808e+00 -7.79068649e-01 -1.23437203e-01 1.38291731e-01
9.70328331e-01 -1.53933614e-01 3.43286335e-01 2.11867228e-01
3.48297924e-01 8.04626405e-01 -1.46751150e-01 1.48699582e-01
7.77661920e-01 4.00981545e-01 -1.68610923e-02 2.87648201e-01
-2.73370624e-01 -2.94607192e-01 2.26061851e-01 1.18861783e+00
2.21344948e-01 -3.15937340e-01 -1.16254747e+00 5.65624833e-01
-1.68077099e+00 -1.37617481e+00 -7.41060019e-01 1.31605744e+00
1.38608277e+00 3.82228434e-01 -2.76713520e-01 -1.45780444e-01
-2.33270116e-02 4.90729123e-01 -2.27757633e-01 -6.93114042e-01
-3.78759474e-01 5.66894770e-01 -3.00906092e-01 1.20944691e+00
-6.98589683e-01 8.35001647e-01 7.73181581e+00 3.44953358e-01
-6.99173629e-01 -3.06981802e-01 7.04318404e-01 -2.79044304e-02
-1.14189589e+00 7.79489577e-02 -5.31901836e-01 2.15983182e-01
1.26743984e+00 -7.80222535e-01 5.02223253e-01 3.46751034e-01
4.14665103e-01 -6.84201196e-02 -1.69936323e+00 4.73051906e-01
2.62180381e-02 -2.26531601e+00 4.32719648e-01 -3.51225346e-01
7.86746919e-01 -5.12002528e-01 -2.09067553e-01 2.23014250e-01
6.29148424e-01 -1.53397834e+00 8.74732375e-01 6.41709745e-01
-3.05078961e-02 -1.15514450e-01 2.86471605e-01 5.63060522e-01
-8.01127672e-01 1.69422448e-01 -2.92074561e-01 -9.39162254e-01
2.98793674e-01 5.27935207e-01 -7.33543456e-01 6.24178886e-01
2.88598388e-01 9.73696411e-01 -7.31636167e-01 2.41175845e-01
-6.03810191e-01 8.18293035e-01 -7.67861009e-02 -4.62345481e-01
1.11118317e-01 -1.51769593e-01 6.31711245e-01 1.33491123e+00
1.20085694e-01 6.50316238e-01 1.60059199e-01 1.75262201e+00
1.55350074e-01 -8.78644809e-02 -5.87774336e-01 -5.70317328e-01
7.39567101e-01 6.32718563e-01 -3.81210834e-01 -6.10269248e-01
-1.66266307e-01 2.57499754e-01 3.24081808e-01 3.10069740e-01
-7.75797963e-01 4.35487926e-01 6.89864516e-01 9.19939652e-02
-3.55044574e-01 -6.28386974e-01 -1.00805938e+00 -1.12964749e+00
-1.37527317e-01 -8.82378757e-01 1.96348071e-01 -8.93088043e-01
-1.39197910e+00 4.10988212e-01 8.56982172e-02 -3.82974178e-01
-6.38597310e-01 -6.18306339e-01 -8.75446439e-01 1.10485220e+00
-1.44066966e+00 -6.31901622e-01 3.10030635e-02 5.38729988e-02
7.75256634e-01 -4.03858535e-02 9.63894606e-01 -4.27169621e-01
-2.61683255e-01 5.12963124e-02 -4.51528341e-01 7.83008263e-02
2.58062482e-01 -1.41887879e+00 7.39765048e-01 7.74561584e-01
4.87121642e-01 1.21688533e+00 1.57170010e+00 -6.28022909e-01
-1.25235426e+00 -3.50961655e-01 1.72811449e+00 -8.61109674e-01
9.74849820e-01 -1.09979026e-01 -1.07994401e+00 6.23059213e-01
8.35192621e-01 -8.37115943e-01 6.84478283e-01 5.00658393e-01
-3.66415143e-01 4.60628003e-01 -3.96892637e-01 8.00516963e-01
9.89495933e-01 -8.15902472e-01 -1.47013402e+00 9.38199282e-01
1.23742402e+00 -9.23171759e-01 -9.13574338e-01 3.40694189e-01
2.28091151e-01 -9.17952001e-01 1.23306739e+00 -1.33335721e+00
1.75844133e+00 5.98377958e-02 -1.05271555e-01 -1.04283273e+00
-7.67841935e-01 -5.44888139e-01 -4.08653736e-01 9.47973371e-01
5.60473263e-01 -4.53292690e-02 3.62976313e-01 9.87291396e-01
-2.56578684e-01 -6.37066841e-01 -8.08961928e-01 -2.95889258e-01
6.51707828e-01 -2.34720901e-01 2.31103688e-01 1.05785608e+00
3.86198848e-01 8.80780756e-01 2.53205508e-01 -6.35009632e-03
5.10120153e-01 7.42504358e-01 4.91539299e-01 -1.26383209e+00
-3.25968534e-01 -1.05600584e+00 1.36193261e-01 -1.14524841e+00
5.81137657e-01 -1.19802403e+00 -1.83418423e-01 -2.03752708e+00
3.04842740e-01 2.70580977e-01 2.27698565e-01 -1.50283471e-01
-3.39639366e-01 -6.29176438e-01 4.82828647e-01 3.42336267e-01
-5.08204341e-01 4.40291315e-01 1.31025124e+00 -3.76608759e-01
2.34945603e-02 -5.00693142e-01 -1.02038538e+00 7.89313316e-01
5.93031824e-01 -1.19280942e-01 -3.57629389e-01 -7.09609091e-01
5.85324407e-01 2.67037034e-01 6.60322130e-01 -6.88929379e-01
4.56703484e-01 -7.70419598e-01 5.21076798e-01 -6.19314849e-01
7.38277882e-02 -6.81257769e-02 -1.13646142e-01 5.55261731e-01
-1.37681186e+00 5.05431235e-01 3.65965992e-01 1.63072661e-01
-2.31213216e-02 -6.41817898e-02 3.66960704e-01 -3.49838823e-01
-4.16089892e-01 -4.15203542e-01 -4.73298967e-01 3.64373863e-01
4.38222855e-01 1.52595803e-01 -8.84710729e-01 -5.93580723e-01
-5.36034584e-01 3.68738234e-01 1.64781481e-01 3.66209716e-01
5.86421669e-01 -1.16153300e+00 -1.16974688e+00 -4.33046639e-01
-2.46576488e-01 -1.91181526e-01 7.93649927e-02 6.25106096e-01
-5.50386667e-01 1.01740086e+00 8.28075223e-03 -7.14411974e-01
-1.01867986e+00 3.15416217e-01 2.33786210e-01 -7.18396068e-01
-7.37883568e-01 9.85150218e-01 1.87203661e-01 -1.80994362e-01
3.40080261e-02 -9.25712049e-01 -5.65064192e-01 -2.77830929e-01
4.66081798e-01 -4.44388352e-02 -4.18032974e-01 -2.74004698e-01
-3.36822331e-01 4.24269944e-01 2.38245148e-02 -5.85817099e-02
1.31337225e+00 -1.98155716e-01 -6.47161067e-01 8.67865562e-01
6.49827540e-01 4.95375469e-02 -5.55358291e-01 -3.84061456e-01
4.77296352e-01 -1.98829770e-01 -1.70709416e-01 -9.13155615e-01
4.96568996e-03 9.44256663e-01 -2.85135448e-01 5.99586785e-01
3.25699985e-01 3.56194407e-01 6.22952640e-01 4.76457953e-01
-2.99174577e-01 -6.30026400e-01 4.10027206e-01 1.03246605e+00
1.32953477e+00 -9.13795471e-01 5.08471608e-01 -5.67479491e-01
-2.07544550e-01 1.57392955e+00 5.66099882e-01 -2.11402088e-01
3.89850318e-01 1.69313982e-01 -2.61628062e-01 -6.48461401e-01
-1.30346847e+00 1.23799033e-01 6.86816752e-01 -7.88854361e-02
1.17284119e+00 1.64943844e-01 -4.69944566e-01 6.45082057e-01
-1.12083292e+00 -3.81379843e-01 6.92683101e-01 5.10855019e-01
-7.28285670e-01 -6.42309427e-01 -4.20625120e-01 2.10807920e-01
-1.36733741e-01 -4.27828878e-01 -9.27613914e-01 4.61766332e-01
-2.92509824e-01 1.16087520e+00 2.20726997e-01 -9.97542590e-02
1.67870864e-01 2.79198796e-01 8.39249372e-01 -6.44798279e-01
-6.92008853e-01 -3.38017792e-01 4.09327626e-01 -4.52052593e-01
-5.09447813e-01 -5.44827580e-01 -1.54643297e+00 -9.32383955e-01
2.10766733e-01 4.25205797e-01 6.77599758e-02 1.59076440e+00
2.10568845e-01 1.23302937e+00 1.41376317e-01 -3.06069463e-01
-6.70114756e-01 -1.19415593e+00 6.73251748e-02 6.81715548e-01
5.83650589e-01 -2.63031602e-01 -2.66785145e-01 4.37615067e-01] | [11.079421043395996, 9.1515474319458] |
ba987cab-6688-4eeb-860e-fea7461988fe | sequential-dual-deep-learning-with-shape-and | 1708.02716 | null | http://arxiv.org/abs/1708.02716v1 | http://arxiv.org/pdf/1708.02716v1.pdf | Sequential Dual Deep Learning with Shape and Texture Features for Sketch Recognition | Recognizing freehand sketches with high arbitrariness is greatly challenging.
Most existing methods either ignore the geometric characteristics or treat
sketches as handwritten characters with fixed structural ordering.
Consequently, they can hardly yield high recognition performance even though
sophisticated learning techniques are employed. In this paper, we propose a
sequential deep learning strategy that combines both shape and texture
features. A coded shape descriptor is exploited to characterize the geometry of
sketch strokes with high flexibility, while the outputs of constitutional
neural networks (CNN) are taken as the abstract texture feature. We develop
dual deep networks with memorable gated recurrent units (GRUs), and
sequentially feed these two types of features into the dual networks,
respectively. These dual networks enable the feature fusion by another gated
recurrent unit (GRU), and thus accurately recognize sketches invariant to
stroke ordering. The experiments on the TU-Berlin data set show that our method
outperforms the average of human and state-of-the-art algorithms even when
significant shape and appearance variations occur. | ['Qi Jia', 'Meiyu Yu', 'Xin Fan', 'Haojie Li'] | 2017-08-09 | null | null | null | null | ['sketch-recognition'] | ['computer-vision'] | [ 1.58280849e-01 -6.82221234e-01 -1.39502540e-01 -2.65311003e-01
-2.98382640e-01 -6.61817253e-01 8.41319203e-01 -4.81966227e-01
-1.83610022e-01 4.02123958e-01 -3.04374397e-01 -7.61318505e-02
-5.51398322e-02 -1.11072624e+00 -5.71858048e-01 -8.39370668e-01
3.33289355e-01 3.50617439e-01 2.27603361e-01 -1.82494134e-01
5.99154472e-01 1.17788875e+00 -1.49399877e+00 4.49345589e-01
5.51579297e-01 1.16910148e+00 1.91804338e-02 6.18676186e-01
-4.44273025e-01 7.18113959e-01 -5.21552145e-01 -3.01733822e-01
3.18480700e-01 -1.56142026e-01 -1.49314076e-01 6.90419972e-02
7.01539159e-01 -5.11250675e-01 -8.14262807e-01 7.42242098e-01
6.19888842e-01 1.95805311e-01 9.27598953e-01 -8.45928371e-01
-1.04796779e+00 2.79110849e-01 -4.62890178e-01 -2.28991017e-01
2.01460123e-01 -7.29696732e-03 8.53148282e-01 -1.10331142e+00
6.69661820e-01 1.19108045e+00 5.66857040e-01 4.25585061e-01
-8.98048282e-01 -8.06818604e-01 3.39468747e-01 2.79271137e-02
-1.47499633e+00 -3.94388646e-01 1.25368643e+00 -1.66045398e-01
7.92487681e-01 3.22280705e-01 6.37720108e-01 1.16803229e+00
7.24169090e-02 1.17492974e+00 9.19077933e-01 -2.26949513e-01
-5.72891720e-02 -3.15934658e-01 1.79027636e-02 1.13044643e+00
8.78918469e-02 4.33836952e-02 -4.11265135e-01 3.73697057e-02
1.54167593e+00 6.52967155e-01 -6.28089979e-02 -4.05958682e-01
-1.04661572e+00 5.15877128e-01 4.19689089e-01 5.23393154e-01
-3.34942371e-01 2.34321713e-01 2.96975940e-01 6.16628170e-01
4.76398505e-02 1.51807815e-01 -2.25570668e-02 2.76303682e-02
-1.00835800e+00 2.64289737e-01 6.70954883e-01 9.85145628e-01
6.40006006e-01 3.53912205e-01 -3.32365900e-01 1.16620183e+00
3.11441217e-02 7.68160343e-01 4.12203401e-01 -4.24512774e-01
3.44456315e-01 7.38914728e-01 6.69903532e-02 -1.55164564e+00
-8.46930891e-02 -1.77685231e-01 -1.23796952e+00 4.98802364e-01
3.69078487e-01 2.09626228e-01 -1.14461589e+00 1.25775146e+00
-3.10359210e-01 1.07356245e-02 -2.14985326e-01 9.06255126e-01
1.10832071e+00 6.11108363e-01 -2.64483809e-01 3.41451049e-01
1.01037908e+00 -1.04968464e+00 -6.66653395e-01 2.11202607e-01
-3.37258130e-02 -1.03317559e+00 9.82719362e-01 6.44391060e-01
-1.06133974e+00 -1.01785028e+00 -1.29942644e+00 -1.21319495e-01
-7.06533074e-01 8.33836198e-01 6.23499274e-01 4.41133857e-01
-7.56855428e-01 9.45579588e-01 -6.22051060e-01 -7.39172846e-02
5.51679850e-01 4.45881665e-01 -2.18839869e-01 6.00020140e-02
-8.94329131e-01 6.82609975e-01 -1.53057009e-01 6.46857977e-01
-6.03967249e-01 -8.08341950e-02 -5.34444928e-01 6.19367957e-02
1.07974946e-01 -2.24699154e-01 8.39936793e-01 -1.01673698e+00
-2.02448297e+00 7.03921914e-01 -7.10292086e-02 8.42243433e-02
9.60555196e-01 -5.24785556e-02 -3.62395048e-01 -1.70154218e-02
-4.46973652e-01 4.54582095e-01 1.35456610e+00 -1.20798814e+00
-2.83744782e-01 -1.94290280e-01 -7.27187544e-02 4.65428941e-02
-3.86512935e-01 -2.79946685e-01 -5.24731398e-01 -9.75163698e-01
3.82627517e-01 -7.95893312e-01 -7.81079829e-02 3.52110237e-01
-3.13564509e-01 -4.50170368e-01 1.19411445e+00 -4.61549163e-01
1.17083573e+00 -2.00181818e+00 2.70012647e-01 4.30666566e-01
1.14573628e-01 5.27638674e-01 -4.31577444e-01 2.82213658e-01
2.65677691e-01 -1.96503073e-01 -4.55851518e-02 -9.78449583e-02
7.58991465e-02 3.10667336e-01 -6.47997200e-01 3.13129395e-01
4.07830268e-01 1.13605452e+00 -6.58567071e-01 -2.07513928e-01
4.57883716e-01 4.13275510e-01 -1.06274836e-01 3.98437649e-01
-1.02936953e-01 2.79669333e-02 -7.76342988e-01 8.84121954e-01
7.56021142e-01 -5.99559732e-02 2.37643614e-01 -3.01912129e-01
-2.77185589e-01 -1.94727063e-01 -1.13029289e+00 1.49337661e+00
-5.36329508e-01 5.54634154e-01 -1.61803633e-01 -8.51902425e-01
1.69209599e+00 6.48337826e-02 -8.04490894e-02 -8.02455127e-01
6.14701547e-02 3.68159980e-01 -1.56296939e-01 -2.94803828e-01
4.25116450e-01 9.79001373e-02 -4.07154672e-02 6.03889644e-01
-1.36321872e-01 -1.13473184e-01 -1.98980510e-01 -3.12165916e-01
8.06649446e-01 3.03946435e-01 -3.79796363e-02 -9.21175703e-02
7.35136986e-01 -7.03516364e-01 2.89726228e-01 9.50821579e-01
2.29890555e-01 8.90273571e-01 6.04005396e-01 -1.11433625e+00
-1.19381368e+00 -1.03192234e+00 1.17266633e-01 1.05354476e+00
1.57624409e-01 1.11160748e-01 -2.53604978e-01 -5.18013418e-01
3.06070417e-01 7.19572231e-02 -8.69821846e-01 7.21529797e-02
-7.63785601e-01 -3.51504683e-01 7.32295871e-01 8.27797711e-01
7.21766233e-01 -1.39755380e+00 -6.19917691e-01 2.56539702e-01
6.13663793e-01 -8.17579448e-01 -3.34174842e-01 -4.89006713e-02
-7.46736646e-01 -8.99177134e-01 -1.23587501e+00 -8.67941260e-01
8.00225079e-01 2.30778411e-01 6.79602683e-01 4.62680936e-01
-3.61080706e-01 1.26227140e-01 -2.48522907e-01 -5.87834120e-02
-3.16892862e-02 2.84730703e-01 -1.89086601e-01 3.72906476e-01
7.45446458e-02 -7.12275326e-01 -5.35797656e-01 3.21807116e-01
-8.84275913e-01 1.05612524e-01 8.99454474e-01 1.21868038e+00
5.25062084e-01 -4.19229805e-01 4.84674603e-01 -7.63959527e-01
7.14010477e-01 2.25456953e-01 -4.92657155e-01 6.99602783e-01
-1.52176410e-01 2.07529739e-01 1.05659175e+00 -6.35434508e-01
-1.01659966e+00 1.71473309e-01 -1.39853051e-02 -6.79834247e-01
-2.26157427e-01 1.49514124e-01 -1.70263842e-01 -6.58035040e-01
1.91960379e-01 5.58862448e-01 -9.63988453e-02 -6.29718721e-01
2.96633035e-01 5.64538240e-01 5.10400534e-01 -8.85751963e-01
7.69820213e-01 4.23238814e-01 9.62275639e-02 -9.70584810e-01
-3.55479211e-01 1.40526593e-01 -9.65290248e-01 -2.20135212e-01
2.73094386e-01 -4.61373389e-01 -8.31563711e-01 9.81750011e-01
-1.31288564e+00 -2.59051234e-01 1.58345774e-01 -5.36182933e-02
-4.00030494e-01 5.55555284e-01 -7.51904368e-01 -9.33716297e-01
-4.62881505e-01 -1.11241627e+00 1.08580601e+00 4.02562171e-01
1.12936452e-01 -7.60581911e-01 -1.91942707e-01 -3.07306558e-01
5.21242440e-01 3.84411544e-01 1.02315652e+00 -3.68905514e-01
-7.61470973e-01 -5.46862304e-01 -6.52546108e-01 7.53667280e-02
3.88881654e-01 4.98273283e-01 -1.02244782e+00 -1.59466207e-01
-5.58291435e-01 -4.24115717e-01 1.09293139e+00 1.52517278e-02
1.62516701e+00 -1.23102307e-01 -1.45161934e-02 7.16562629e-01
1.33352220e+00 3.86913747e-01 7.43241310e-01 -5.54319879e-04
9.23298359e-01 1.26789391e-01 5.06286807e-02 5.76383710e-01
-7.06495866e-02 5.40772498e-01 -7.56185949e-02 -2.94205844e-01
-6.22097403e-02 -4.00159866e-01 1.50485799e-01 1.05346346e+00
-6.66454315e-01 -1.83370605e-01 -6.75481737e-01 1.21681288e-01
-1.90617037e+00 -9.23712015e-01 2.65666872e-01 2.04705739e+00
5.17059624e-01 1.94136932e-01 -1.43362343e-01 6.34278432e-02
7.06179619e-01 6.01313233e-01 -5.78909993e-01 -5.54717541e-01
-5.26993275e-01 5.20473063e-01 3.06047380e-01 2.51955479e-01
-1.07290745e+00 1.16534066e+00 6.16921091e+00 9.73851800e-01
-1.46292126e+00 -5.33456087e-01 6.44966185e-01 1.37242600e-01
-3.72671485e-01 -4.43767756e-01 -4.58569884e-01 2.82205939e-01
2.27572098e-02 2.55439609e-01 7.28214979e-01 7.09146857e-01
-3.09213549e-01 3.28369439e-01 -1.00389111e+00 1.21368277e+00
1.88347593e-01 -1.42180109e+00 7.14736283e-01 -2.67467320e-01
7.51948595e-01 -4.92194712e-01 4.50441182e-01 1.97598204e-01
-7.25925788e-02 -1.32898414e+00 6.07532620e-01 1.22583938e+00
1.19514668e+00 -8.00346494e-01 6.26424193e-01 -2.47084480e-02
-1.46892810e+00 -2.40756255e-02 -7.05242276e-01 -1.36516526e-01
-3.38089615e-01 1.17410250e-01 -7.75672048e-02 6.12822831e-01
1.93705261e-01 8.95745516e-01 -5.13511300e-01 7.24830389e-01
-1.86709911e-01 8.14743116e-02 -2.63014138e-01 -5.34948289e-01
5.83474159e-01 -3.97073478e-01 1.42001256e-01 1.31663096e+00
3.08539540e-01 1.91697150e-01 1.65680915e-01 1.09795833e+00
-1.84620112e-01 -9.79125500e-03 -6.67655587e-01 -3.90472502e-01
3.67997050e-01 1.31357861e+00 -9.06035304e-01 -5.67009926e-01
-3.08816433e-01 1.35054231e+00 4.05054182e-01 4.59327728e-01
-4.18079346e-01 -8.44645321e-01 3.49495620e-01 -3.41312349e-01
5.37322104e-01 -4.18745458e-01 -5.40233374e-01 -1.18757987e+00
1.09703779e-01 -7.29808033e-01 -5.18225692e-02 -5.34350932e-01
-1.56117046e+00 7.03141093e-01 -6.38375700e-01 -1.17387986e+00
-3.68838049e-02 -1.21070933e+00 -1.00875676e+00 9.32413280e-01
-1.41593742e+00 -1.43767464e+00 -4.95679498e-01 4.66529220e-01
5.60024679e-01 -5.74802041e-01 9.25647736e-01 2.39845216e-01
-6.75477147e-01 9.73228872e-01 2.12472573e-01 6.48255765e-01
5.54267168e-01 -9.58525360e-01 5.92881024e-01 2.77662694e-01
2.60141581e-01 7.93565571e-01 -4.83191535e-02 -5.48224926e-01
-1.78772545e+00 -7.61334062e-01 6.73229754e-01 -5.56211174e-02
4.66156453e-01 -5.76044858e-01 -9.91577625e-01 2.86767930e-01
-1.05199769e-01 1.50830835e-01 2.47299492e-01 -7.05778077e-02
-5.95534146e-01 -1.42314851e-01 -7.75992155e-01 6.83954835e-01
1.13926291e+00 -7.60625064e-01 -6.15742028e-01 -1.10879757e-01
1.03482157e-02 -4.11448389e-01 -5.74775338e-01 3.75045776e-01
1.41419506e+00 -9.45771456e-01 9.14831519e-01 -7.16073453e-01
5.43086469e-01 -1.52840108e-01 -6.71384484e-02 -7.87727535e-01
-5.42281032e-01 -4.69722390e-01 -5.59019260e-02 9.29726660e-01
1.70405909e-01 -4.00792837e-01 8.87096584e-01 3.65972251e-01
2.43248433e-01 -9.57007945e-01 -7.03928292e-01 -6.78482592e-01
7.46801123e-02 2.77813785e-02 7.37109840e-01 9.85703468e-01
-4.78694826e-01 -1.46415010e-02 -5.79301417e-01 -2.60577887e-01
4.26926643e-01 7.98104942e-01 6.93907619e-01 -1.37117732e+00
1.14690669e-01 -1.02716780e+00 -5.48457980e-01 -1.19592440e+00
3.22267026e-01 -6.74630404e-01 -1.73922241e-01 -1.10792136e+00
-1.87048428e-02 -5.85662901e-01 -5.00800669e-01 6.08537674e-01
1.66979820e-01 3.19814146e-01 2.69463122e-01 8.23833868e-02
-4.71588850e-01 7.84637749e-01 1.33988500e+00 -3.78815413e-01
-1.13009758e-01 -2.03033052e-02 -1.32668599e-01 7.16474712e-01
5.55395186e-01 1.59233749e-01 -8.20784271e-02 -6.11304045e-01
-1.35810435e-01 2.03686692e-02 5.56881368e-01 -9.94041860e-01
4.17505711e-01 -1.03015073e-01 1.07612979e+00 -8.41078579e-01
2.02832833e-01 -6.69861615e-01 -5.04566431e-02 3.97722781e-01
-4.85837102e-01 1.18445881e-01 1.89777583e-01 4.03257310e-01
-2.01334491e-01 -1.31675795e-01 6.39975607e-01 -2.58154720e-01
-6.64872706e-01 6.04293525e-01 -2.77393073e-01 -5.43122113e-01
3.36454332e-01 -3.50303233e-01 -5.78880310e-04 -3.05772632e-01
-5.32060802e-01 -1.76785961e-01 3.71116638e-01 5.88739693e-01
1.10959208e+00 -1.79112983e+00 -5.41087031e-01 6.84009373e-01
-9.82640833e-02 -2.36455500e-01 4.07518893e-01 1.47903860e-01
-7.24702120e-01 4.42567796e-01 -7.95630515e-01 -2.81641603e-01
-1.02132523e+00 2.77181298e-01 3.82251263e-01 -5.07773831e-02
-7.31560647e-01 7.84673095e-01 -5.94527647e-02 -4.63569373e-01
4.60937321e-01 -2.87679553e-01 -1.64412588e-01 2.06960261e-01
4.91185635e-01 3.10214102e-01 -3.04879304e-02 -4.41099495e-01
-1.76643416e-01 1.08088875e+00 -2.46994838e-01 7.71977529e-02
1.43451619e+00 4.98530954e-01 -1.46492571e-01 4.82001781e-01
1.16823721e+00 -1.61352828e-01 -1.38277674e+00 -2.86656618e-01
-1.05225302e-01 -5.96724391e-01 -2.44217768e-01 -7.11206913e-01
-1.30753529e+00 1.27717924e+00 4.97756720e-01 -5.17285503e-02
8.70323062e-01 -7.14994550e-01 7.95962453e-01 1.03343666e+00
3.03697944e-01 -1.14376426e+00 1.94098443e-01 7.40092337e-01
1.33178830e+00 -8.13459873e-01 -3.82946953e-02 3.64181846e-02
-4.37377185e-01 1.94563222e+00 6.54931426e-01 -8.46796215e-01
5.23891449e-01 2.68326670e-01 -1.64178517e-02 -2.63620447e-02
-5.51311672e-01 1.91861354e-02 6.33490026e-01 1.91900566e-01
4.93111521e-01 1.29655600e-01 -1.43119663e-01 7.18547642e-01
6.60315156e-02 5.61355352e-02 8.45682696e-02 8.38965178e-01
-2.92800784e-01 -1.17289233e+00 -1.64027140e-01 5.49152493e-01
3.79845500e-02 8.03686082e-02 -7.33522892e-01 7.04898298e-01
2.41459534e-02 1.56359017e-01 1.24736533e-01 -5.79617977e-01
4.06285405e-01 8.09978992e-02 7.66072512e-01 -1.59133479e-01
-5.90786397e-01 -9.48678516e-03 -3.53465110e-01 -3.94300103e-01
-1.47469610e-01 -4.45962012e-01 -8.54908705e-01 -3.23596269e-01
-5.29257506e-02 -2.34649241e-01 2.63071179e-01 7.41329193e-01
9.17405561e-02 4.55123365e-01 8.59615684e-01 -1.20929849e+00
-5.73520482e-01 -7.60869682e-01 -6.49720669e-01 2.36765429e-01
3.26642632e-01 -7.43590176e-01 -4.31302302e-02 -2.96890050e-01] | [11.72422981262207, 0.4606890380382538] |
2f545220-4f29-4b06-890b-bd1a35652314 | autoregressive-perturbations-for-data | 2206.03693 | null | https://arxiv.org/abs/2206.03693v3 | https://arxiv.org/pdf/2206.03693v3.pdf | Autoregressive Perturbations for Data Poisoning | The prevalence of data scraping from social media as a means to obtain datasets has led to growing concerns regarding unauthorized use of data. Data poisoning attacks have been proposed as a bulwark against scraping, as they make data "unlearnable" by adding small, imperceptible perturbations. Unfortunately, existing methods require knowledge of both the target architecture and the complete dataset so that a surrogate network can be trained, the parameters of which are used to generate the attack. In this work, we introduce autoregressive (AR) poisoning, a method that can generate poisoned data without access to the broader dataset. The proposed AR perturbations are generic, can be applied across different datasets, and can poison different architectures. Compared to existing unlearnable methods, our AR poisons are more resistant against common defenses such as adversarial training and strong data augmentations. Our analysis further provides insight into what makes an effective data poison. | ['David W. Jacobs', 'Tom Goldstein', 'Micah Goldblum', 'Jonas Geiping', 'Vasu Singla', 'Pedro Sandoval-Segura'] | 2022-06-08 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [ 6.49227723e-02 1.14929564e-01 1.65998600e-02 4.14951928e-02
-7.31814027e-01 -1.22728324e+00 7.56161392e-01 2.11890623e-01
-5.07291853e-01 5.71531832e-01 4.40322965e-01 -4.40234482e-01
3.57202172e-01 -9.00480568e-01 -8.98992419e-01 -7.36060143e-01
2.44073987e-01 1.03602998e-01 -4.65479754e-02 -4.80336249e-01
2.65394688e-01 6.67422593e-01 -1.01781452e+00 1.81035653e-01
7.17239380e-01 3.42560261e-01 -4.67036307e-01 5.75978220e-01
2.18875587e-01 7.89490581e-01 -1.20635462e+00 -7.82884002e-01
5.90850174e-01 -2.00314283e-01 -5.84136903e-01 -3.47638220e-01
2.57827669e-01 -6.81203961e-01 -5.92732072e-01 1.12070882e+00
5.88146627e-01 -2.14462608e-01 5.53683341e-01 -1.56757724e+00
-1.09252143e+00 9.29790437e-01 -5.13767958e-01 1.61703795e-01
3.57955277e-01 5.95603228e-01 4.43576723e-01 -5.65507293e-01
3.47159922e-01 1.29032850e+00 6.36188269e-01 1.33171904e+00
-1.45430183e+00 -1.13049495e+00 -6.51163831e-02 -3.97042215e-01
-1.26551139e+00 -5.24515212e-01 8.34380805e-01 -4.18876171e-01
2.83832639e-01 4.30152088e-01 1.75574243e-01 2.01342487e+00
1.79860983e-02 6.94009900e-01 1.11132801e+00 -1.99496537e-01
5.60601711e-01 3.53068173e-01 5.33920601e-02 1.89491510e-01
6.36027396e-01 1.89615652e-01 -4.68116790e-01 -9.97038841e-01
6.02746546e-01 1.87154226e-02 -3.72095883e-01 -2.07565516e-01
-7.22565114e-01 1.05163336e+00 4.93363112e-01 -9.07950178e-02
-5.32574534e-01 1.38802737e-01 4.91183996e-01 3.64210337e-01
2.73058981e-01 9.61192369e-01 -3.93987447e-01 1.66265726e-01
-3.04167628e-01 4.66905802e-01 8.55842650e-01 6.13535523e-01
4.94369030e-01 5.20381212e-01 2.88131516e-02 3.15960407e-01
2.96814114e-01 7.90823042e-01 6.17105067e-01 -6.42064869e-01
5.39584816e-01 5.76004744e-01 2.95589030e-01 -1.11809599e+00
-6.79435879e-02 -1.05592139e-01 -8.33917677e-01 3.12613845e-01
6.09647751e-01 -5.82996488e-01 -8.18338454e-01 1.99027336e+00
4.90022868e-01 1.31523207e-01 3.39230329e-01 6.68664873e-01
4.86501664e-01 3.43364060e-01 3.08879882e-01 3.44682395e-01
1.06655610e+00 -1.57582358e-01 -4.57660526e-01 -1.45775914e-01
4.88536984e-01 -2.91145682e-01 1.46312571e+00 4.52234715e-01
-7.31997252e-01 -1.01143926e-01 -1.16713083e+00 2.04600513e-01
-7.78702021e-01 -7.81207204e-01 3.00941855e-01 1.35459900e+00
-7.33051717e-01 3.04265559e-01 -7.53725111e-01 -1.55512705e-01
7.34985828e-01 4.51370478e-01 -3.19271445e-01 1.09819099e-01
-1.35021436e+00 7.66825438e-01 1.06081992e-01 -2.35099792e-01
-1.44123316e+00 -9.07463253e-01 -5.88029742e-01 -2.94599354e-01
2.15119094e-01 -5.42602003e-01 9.18303311e-01 -7.24538207e-01
-1.08647954e+00 5.35175741e-01 3.91568780e-01 -8.99411559e-01
6.43771410e-01 -5.57290316e-01 -4.54691052e-01 1.23293489e-01
-1.40426099e-01 4.32955086e-01 1.31301463e+00 -1.61632001e+00
1.49318039e-01 -3.59418422e-01 1.75733358e-01 -5.83365075e-02
-8.26966166e-01 2.97581226e-01 1.52210668e-01 -1.01154613e+00
-6.08284473e-01 -9.23047960e-01 -4.41402674e-01 -3.37380558e-01
-9.98752356e-01 3.29879105e-01 1.19181323e+00 -5.56568205e-01
1.04463029e+00 -2.22453475e+00 1.41625814e-02 2.50280142e-01
5.60421944e-01 6.57363176e-01 -2.46790871e-01 6.08817577e-01
-1.23524852e-01 9.79602516e-01 -1.97713286e-01 -1.61118135e-01
-5.87125272e-02 1.44397272e-02 -1.00208735e+00 5.51773012e-01
2.25938067e-01 7.96124220e-01 -7.52965748e-01 1.52496502e-01
-8.06627646e-02 6.36689186e-01 -6.11683726e-01 3.25248122e-01
-2.50352234e-01 6.81240678e-01 -6.75174713e-01 4.62557644e-01
5.80144167e-01 7.11725205e-02 -2.61377633e-01 -4.37936233e-03
2.85934836e-01 3.04520484e-02 -7.78155684e-01 8.95901322e-01
-7.95059130e-02 2.92450309e-01 1.03502557e-01 -2.66754121e-01
9.50898111e-01 3.39611381e-01 2.03970790e-01 -1.65126055e-01
4.19847697e-01 -6.36922643e-02 -1.02990411e-01 -3.21139485e-01
3.11040074e-01 -1.06851459e-01 -4.49328244e-01 8.04211736e-01
-4.92550701e-01 -8.67387876e-02 -2.64603823e-01 4.93471622e-01
1.46119237e+00 -4.19196665e-01 6.01658933e-02 -9.03170481e-02
3.02939802e-01 -5.93668967e-02 4.55351174e-01 1.09312248e+00
-6.37129843e-02 5.46426177e-01 4.07237113e-01 -5.38282156e-01
-1.27480805e+00 -1.13479948e+00 1.77620620e-01 8.03467214e-01
3.59575427e-03 -3.49137962e-01 -1.21296859e+00 -1.09225047e+00
2.31787249e-01 8.83049190e-01 -7.78906286e-01 -7.49763489e-01
-5.20389795e-01 -9.16899681e-01 1.49980474e+00 3.89903933e-01
5.63085139e-01 -1.01429403e+00 -3.58704954e-01 3.86694707e-02
-1.21785533e-02 -8.47971320e-01 -5.14519691e-01 -1.79374129e-01
-5.25885463e-01 -1.23802471e+00 -2.87855983e-01 -9.35779959e-02
8.05432081e-01 2.22032398e-01 6.74076855e-01 2.63707310e-01
-1.63899928e-01 4.72965568e-01 -3.43959659e-01 -8.37471664e-01
-1.04787683e+00 1.73245475e-01 6.04845762e-01 2.70230889e-01
6.08471513e-01 -6.36538923e-01 -2.84833372e-01 2.81465948e-01
-1.47106707e+00 -5.49723268e-01 2.62183368e-01 5.10776281e-01
-2.52226964e-02 -1.99460052e-02 9.46864784e-01 -1.32574332e+00
1.30667174e+00 -7.72343755e-01 -3.69955748e-01 -2.48901118e-02
-2.96084315e-01 8.80326405e-02 1.14097214e+00 -1.00071681e+00
-6.35709763e-01 -1.62257627e-01 -5.66770472e-02 -7.77572691e-01
-2.68737942e-01 1.51248276e-01 -4.45016116e-01 -2.65247822e-01
1.42824328e+00 1.02582403e-01 1.02108255e-01 -6.25583053e-01
5.30238807e-01 6.59884751e-01 4.52319294e-01 -7.06026614e-01
1.49727356e+00 5.36328077e-01 -3.55608672e-01 -7.64413238e-01
-6.13494635e-01 1.21366397e-01 -4.33453411e-01 2.15402931e-01
7.44886935e-01 -7.67298937e-01 -7.22963274e-01 8.97265792e-01
-1.05915451e+00 -3.96640688e-01 -2.99614400e-01 -9.06116155e-04
-1.31983206e-01 3.87299478e-01 -6.17020488e-01 -7.99379170e-01
-5.46529412e-01 -8.77729535e-01 4.72234845e-01 9.19427946e-02
-5.04918098e-01 -8.90168846e-01 1.08819477e-01 4.64470148e-01
6.49567068e-01 6.96551561e-01 9.30075288e-01 -1.26693106e+00
-3.00835669e-01 -7.27485657e-01 2.00430155e-01 5.69813192e-01
2.57115752e-01 5.28746732e-02 -1.10283661e+00 -5.21879137e-01
3.66558313e-01 -5.07823050e-01 3.20918769e-01 -3.02355081e-01
1.19155681e+00 -1.14263499e+00 -3.24299783e-02 4.75787669e-01
1.14691961e+00 4.81891930e-02 8.51076663e-01 3.30685019e-01
9.64321375e-01 4.08235252e-01 7.95226470e-02 5.43548465e-01
6.72025606e-02 2.21864328e-01 7.44740427e-01 -1.14033960e-01
2.06621557e-01 -7.33729124e-01 5.26666999e-01 2.50264525e-01
3.70474279e-01 -3.56765628e-01 -9.74959254e-01 3.18934411e-01
-1.37061775e+00 -9.90341485e-01 -5.96068278e-02 2.33663321e+00
1.00689924e+00 1.88303933e-01 4.03781325e-01 1.42971918e-01
7.39052773e-01 9.89524946e-02 -9.57830310e-01 -3.18610847e-01
-1.50908425e-01 1.99996546e-01 8.57820690e-01 3.79703373e-01
-1.11234808e+00 1.00904036e+00 7.14751101e+00 4.37740773e-01
-1.18013322e+00 9.54814255e-02 4.19389307e-01 -2.31743097e-01
-3.93411547e-01 -1.00927137e-01 -7.12745070e-01 5.94594300e-01
1.15831006e+00 -2.94063538e-01 6.23362064e-01 7.63326466e-01
2.03181759e-01 4.81110752e-01 -1.00651133e+00 5.84523618e-01
1.51035607e-01 -1.17126632e+00 4.83270526e-01 3.33208263e-01
4.76195276e-01 1.36057481e-01 3.94491524e-01 1.59869164e-01
1.19812393e+00 -1.27724040e+00 3.90148908e-01 2.52427965e-01
4.97631729e-01 -8.85354161e-01 3.60653341e-01 5.28422356e-01
-4.07491207e-01 -3.68831426e-01 -2.61935830e-01 1.32560626e-01
-1.04100935e-01 1.87737927e-01 -9.48608100e-01 7.79618472e-02
4.87726241e-01 9.50266272e-02 -9.34045970e-01 6.32797539e-01
-3.74871761e-01 9.37113643e-01 -3.10043782e-01 1.28891766e-01
-2.11867760e-03 2.67359376e-01 6.71973884e-01 7.44675398e-01
-4.06177193e-02 3.87497656e-02 2.48232454e-01 8.59256804e-01
-3.75488311e-01 -1.65019482e-01 -1.20651340e+00 -4.11346853e-01
9.16853964e-01 9.93620694e-01 -9.10856649e-02 -7.46589303e-02
-1.06263019e-01 6.87671423e-01 2.74699926e-01 4.97100085e-01
-8.14121664e-01 -2.90386498e-01 8.59276116e-01 3.70745122e-01
-3.68114799e-01 -2.31757388e-01 -3.92441213e-01 -1.29205358e+00
-3.06203961e-01 -1.65642047e+00 4.28488553e-01 -6.28409922e-01
-1.66977155e+00 6.24750674e-01 -1.68198682e-02 -1.10316789e+00
-6.73754737e-02 -1.68961942e-01 -6.16178811e-01 8.52132618e-01
-9.27412570e-01 -1.16731095e+00 -2.40001827e-02 8.72791588e-01
9.46237668e-02 -4.60431695e-01 9.77703512e-01 1.57483462e-02
-8.45607996e-01 8.97432208e-01 -1.32743657e-01 6.39098287e-01
6.84026659e-01 -9.63054061e-01 7.45414138e-01 1.11220062e+00
1.07315101e-01 1.09653485e+00 1.05141985e+00 -9.79439914e-01
-1.54115748e+00 -1.21186423e+00 7.28541762e-02 -1.10485697e+00
7.79754341e-01 -7.23118246e-01 -1.32903123e+00 7.42548347e-01
7.29595348e-02 -2.33744383e-01 9.59715486e-01 -2.62452543e-01
-1.07847774e+00 3.38965580e-02 -1.66217661e+00 9.34533954e-01
5.57177007e-01 -7.42812932e-01 -4.24224854e-01 1.31957293e-01
9.62523937e-01 -2.23555211e-02 -7.32075393e-01 1.45716161e-01
1.90924436e-01 -5.82450449e-01 1.07263863e+00 -1.25084352e+00
1.88229993e-01 -1.98596522e-01 -9.36367065e-02 -1.21200442e+00
-4.83941920e-02 -1.08582497e+00 -3.10158700e-01 1.47242641e+00
4.46926147e-01 -8.98021281e-01 8.17901790e-01 1.18522453e+00
3.53841335e-01 -1.62433192e-01 -5.61754048e-01 -7.83131242e-01
4.70175475e-01 -2.27979779e-01 1.05480170e+00 1.44242513e+00
-1.44298866e-01 1.24940947e-01 -8.07202101e-01 6.49754286e-01
8.84098053e-01 -7.62863755e-01 1.34731996e+00 -1.11498702e+00
-1.65223360e-01 -1.16822161e-01 -2.11963862e-01 -3.86738598e-01
1.02705592e-02 -6.47502899e-01 -3.14624548e-01 -6.96538389e-01
-9.18540731e-02 -4.09275532e-01 -1.69331819e-01 7.98259497e-01
-2.09273040e-01 3.01960737e-01 3.15363288e-01 5.52840412e-01
7.21717924e-02 3.95505548e-01 7.10654736e-01 -8.81840438e-02
-3.27292532e-01 -7.54459351e-02 -1.41522050e+00 8.15437138e-01
1.20787907e+00 -9.72571194e-01 -5.93716621e-01 -3.43323290e-01
2.57339060e-01 -4.26651746e-01 5.30322731e-01 -8.77460539e-01
1.52824968e-01 -2.74430960e-01 2.28773072e-01 -2.86157783e-02
1.67669877e-01 -6.64559424e-01 1.51961535e-01 4.95076895e-01
-5.75632155e-01 2.68809557e-01 2.94716656e-01 7.63919413e-01
2.88585991e-01 -2.89646149e-01 9.65275228e-01 -1.47205904e-01
3.44803482e-02 3.53664726e-01 -3.91538024e-01 3.57766956e-01
1.17474484e+00 5.15614003e-02 -6.94975674e-01 -4.48784739e-01
-5.51282704e-01 4.05584686e-02 9.48737800e-01 5.41083455e-01
5.46875536e-01 -1.15582132e+00 -7.99074531e-01 3.08926105e-01
-1.73785731e-01 3.86315072e-03 -7.83647299e-02 7.83378035e-02
-2.76261061e-01 -2.50223935e-01 -2.48391107e-01 3.25595178e-02
-1.23720860e+00 1.16744959e+00 2.65684038e-01 1.01408817e-01
-6.24836385e-01 5.61386526e-01 2.09699050e-01 -4.96899545e-01
2.92115808e-01 3.12345177e-01 -7.49234715e-03 -1.68500692e-01
9.83790696e-01 1.85884044e-01 -1.90332562e-01 -4.90673155e-01
-1.77397728e-01 -1.66598186e-01 -4.49200034e-01 -1.10472538e-01
1.17376697e+00 1.35477588e-01 6.52944520e-02 2.12184176e-01
8.40282857e-01 4.14795935e-01 -1.22236264e+00 -1.33929193e-01
-1.17653154e-01 -5.30318797e-01 -3.43600184e-01 -7.19610751e-01
-9.03145134e-01 6.96739197e-01 1.57743692e-01 8.42249453e-01
7.92167842e-01 -2.71852344e-01 7.11031020e-01 4.23933387e-01
3.74105871e-01 -6.56429291e-01 3.50247651e-01 1.99068472e-01
1.02020180e+00 -9.56646383e-01 -1.32910281e-01 -2.69174546e-01
-8.46661329e-01 6.66381359e-01 7.81074047e-01 -4.21325266e-01
4.35698241e-01 5.16160011e-01 2.27605984e-01 -1.40382662e-01
-6.42351985e-01 4.53231961e-01 -1.32378295e-01 1.08455515e+00
-2.15756208e-01 -2.46712402e-01 1.89757913e-01 6.66157663e-01
-3.35930169e-01 -3.31484407e-01 1.11595082e+00 9.29027736e-01
-1.10713616e-01 -1.20386004e+00 -8.26155782e-01 3.65228772e-01
-8.47526073e-01 3.28927636e-02 -1.02809680e+00 7.47163177e-01
-3.89410704e-01 1.13346612e+00 -4.19018537e-01 -6.49001002e-01
2.74072140e-01 1.19689770e-01 -1.31596938e-01 -6.64838493e-01
-1.04022133e+00 -4.42334533e-01 -9.14770085e-03 -2.69877732e-01
5.23216054e-02 -4.69146937e-01 -1.03339911e+00 -7.65704036e-01
-2.10169971e-01 1.64718106e-01 4.95693952e-01 7.09930658e-01
6.34485662e-01 -2.26310678e-02 8.90527189e-01 -5.18715203e-01
-1.15675402e+00 -7.72793055e-01 -3.22556883e-01 8.83238256e-01
3.33860993e-01 -2.70510256e-01 -6.72590315e-01 3.30201420e-03] | [5.826002597808838, 7.723130702972412] |
8095f427-981f-439e-a141-b079f298740f | multi-output-gaussian-process-based-data | 2202.01980 | null | https://arxiv.org/abs/2202.01980v1 | https://arxiv.org/pdf/2202.01980v1.pdf | Multi-Output Gaussian Process-Based Data Augmentation for Multi-Building and Multi-Floor Indoor Localization | Location fingerprinting based on RSSI becomes a mainstream indoor localization technique due to its advantage of not requiring the installation of new infrastructure and the modification of existing devices, especially given the prevalence of Wi-Fi-enabled devices and the ubiquitous Wi-Fi access in modern buildings. The use of AI/ML technologies like DNNs makes location fingerprinting more accurate and reliable, especially for large-scale multi-building and multi-floor indoor localization. The application of DNNs for indoor localization, however, depends on a large amount of preprocessed and deliberately-labeled data for their training. Considering the difficulty of the data collection in an indoor environment, especially under the current epidemic situation of COVID-19, we investigate three different methods of RSSI data augmentation based on Multi-Output Gaussian Process (MOGP), i.e., by a single floor, by neighboring floors, and by a single building; unlike Single-Output Gaussian Process (SOGP), MOGP can take into account the correlation among RSSI observations from multiple Access Points (APs) deployed closely to each other (e.g., APs on the same floor of a building) by collectively handling them. The feasibility of the MOGP-based RSSI data augmentation is demonstrated through experiments based on the state-of-the-art RNN indoor localization model and the UJIIndoorLoc, i.e., the most popular publicly-available multi-building and multi-floor indoor localization database, where the RNN model trained with the UJIIndoorLoc database augmented by using the whole RSSI data of a building in fitting an MOGP model (i.e., by a single building) outperforms the other two augmentation methods as well as the RNN model trained with the original UJIIndoorLoc database, resulting in the mean three-dimensional positioning error of 8.42 m. | ['Jeremy Smith', 'Kyeong Soo Kim', 'Sihao Li', 'Zhe Tang'] | 2022-02-04 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [ 1.74029525e-02 -2.55645126e-01 8.44955668e-02 -1.18163250e-01
-8.25364530e-01 -2.93567210e-01 2.47550845e-01 -7.96796009e-02
-4.29631233e-01 8.84823322e-01 1.90627337e-01 -4.72607523e-01
-4.90230471e-01 -1.09335232e+00 -1.00123000e+00 -1.02044451e+00
2.49953568e-03 3.68722051e-01 -6.17754757e-02 1.19103253e-01
-3.43620926e-01 3.82582188e-01 -1.36817932e+00 -4.98656958e-01
9.24005508e-01 1.09934628e+00 4.15303499e-01 6.03423774e-01
1.53600037e-01 2.84738153e-01 -8.38792622e-01 -3.08531225e-02
2.19926760e-02 3.05044767e-03 1.02322079e-01 -5.91100276e-01
8.90440643e-02 -1.62291944e-01 -4.92515236e-01 5.17568469e-01
1.06106293e+00 3.61569494e-01 3.08619827e-01 -1.34063017e+00
-5.27479231e-01 5.65865576e-01 -4.94177103e-01 -1.39056116e-01
4.24812108e-01 -5.13626993e-01 4.34839875e-01 -5.99502504e-01
-3.23307328e-02 7.61760712e-01 1.35461617e+00 1.17271505e-01
-1.13138115e+00 -8.44404459e-01 1.26816079e-01 -2.80943185e-01
-2.00891972e+00 -2.40541890e-01 7.27248251e-01 1.13817297e-01
3.46211225e-01 4.62405086e-01 3.51782590e-01 1.76959813e+00
8.17680582e-02 5.20636320e-01 7.40843952e-01 -2.12826446e-01
3.67466211e-01 -8.63427520e-02 -7.56832138e-02 1.74351141e-01
6.36764824e-01 -2.65017394e-02 -4.09882218e-01 -4.95783091e-01
7.48542309e-01 5.64700484e-01 -3.12369496e-01 -1.57571912e-01
-1.38390100e+00 1.63315639e-01 6.77164733e-01 6.51802480e-01
-6.79125607e-01 5.75544775e-01 -3.69438767e-01 -3.97118658e-01
2.20984861e-01 1.96371436e-01 -6.93910897e-01 -3.87077808e-01
-1.23710454e+00 -6.09212928e-03 7.74137318e-01 1.29830682e+00
6.50664389e-01 2.99678911e-02 -2.50116199e-01 8.86832416e-01
8.08619976e-01 1.34329283e+00 3.45115036e-01 -6.06645703e-01
7.30648756e-01 7.63904527e-02 3.63309026e-01 -1.12934899e+00
-7.96885014e-01 -1.05131996e+00 -1.26118743e+00 -6.18904412e-01
3.82642478e-01 -7.03949153e-01 -8.43477726e-01 1.98355901e+00
1.87473938e-01 7.82057226e-01 -6.09179437e-02 1.92630291e-01
5.81323087e-01 6.04202271e-01 -1.21193349e-01 2.02935025e-01
1.04401267e+00 -7.80705750e-01 -5.94939172e-01 -2.54540652e-01
3.97927135e-01 -5.14801681e-01 5.84900081e-01 2.16828346e-01
-3.36971194e-01 -8.04189622e-01 -1.00633156e+00 5.23165584e-01
-5.80964744e-01 2.52542287e-01 5.10292470e-01 1.16774559e+00
-1.13153613e+00 1.39353201e-01 -9.80156720e-01 -3.77788305e-01
2.06149161e-01 5.86013556e-01 -2.09030733e-01 -4.65728402e-01
-1.10832846e+00 3.40296239e-01 -9.09782425e-02 6.27519190e-01
-4.30240750e-01 -6.63337767e-01 -7.49610603e-01 5.96199073e-02
4.04002704e-02 -6.80332661e-01 8.23975146e-01 -5.11061251e-02
-1.23817968e+00 -2.36356273e-01 -4.53254312e-01 -2.45562270e-01
3.32358509e-01 -2.06408277e-01 -9.40154374e-01 -7.08146870e-01
3.66607040e-01 9.55298319e-02 3.66729885e-01 -1.52679694e+00
-6.76421702e-01 -4.51592088e-01 -3.93273264e-01 -2.43380785e-01
-1.70020625e-01 -5.04616082e-01 -5.42797029e-01 -6.17962003e-01
5.92613876e-01 -1.30477905e+00 -4.77020115e-01 -6.23642743e-01
-7.85945833e-01 2.89126277e-01 7.56631136e-01 -7.76627243e-01
1.36008132e+00 -2.16009426e+00 -6.86017096e-01 8.70932400e-01
-1.85705110e-01 -1.04132123e-01 -5.41083328e-03 3.99298579e-01
2.32805729e-01 1.49579778e-01 9.27877426e-02 -8.77223492e-01
1.07673250e-01 4.03623819e-01 -2.66170073e-02 4.46043223e-01
-7.00685024e-01 7.51713753e-01 -1.18408859e+00 -1.18557058e-01
2.78929204e-01 8.76377285e-01 -3.01875323e-01 -3.98443826e-03
4.85292047e-01 9.26442146e-01 -6.67198718e-01 7.09637880e-01
9.30032730e-01 -2.25756034e-01 5.98314591e-02 -9.79644433e-02
-2.91917473e-01 1.27395973e-01 -1.58301485e+00 1.74601614e+00
-8.98217976e-01 2.91422933e-01 -1.74561799e-01 -3.90399605e-01
8.94417346e-01 5.88743985e-01 6.51113808e-01 -6.23310268e-01
-1.57391667e-01 5.03792942e-01 -4.08187598e-01 2.36660596e-02
3.83852869e-01 3.92787099e-01 -3.71225953e-01 4.55566555e-01
-6.57903627e-02 4.01483208e-01 -2.60854036e-01 -3.00721109e-01
1.44914198e+00 9.46509913e-02 -5.97106814e-02 9.99796018e-02
4.70610470e-01 -8.83786857e-01 5.55844367e-01 1.54192865e+00
1.32514253e-01 6.89470053e-01 -4.14336175e-01 1.08610861e-01
-5.03056765e-01 -1.19168985e+00 -3.23964030e-01 9.54258919e-01
1.32749274e-01 -1.47412181e-01 -3.72855812e-01 -5.08124292e-01
2.17463121e-01 7.56728828e-01 -5.16875267e-01 1.93402290e-01
-4.60067272e-01 -1.01492012e+00 1.02315438e+00 4.94008094e-01
8.40391457e-01 -6.04630828e-01 1.77320957e-01 4.24983948e-01
-5.37632763e-01 -1.13101554e+00 -6.14194833e-02 4.14803177e-01
-3.72736484e-01 -5.22604167e-01 -9.15830910e-01 -4.33999300e-01
7.27311611e-01 6.12189233e-01 6.85746968e-01 -9.01667848e-02
4.73076850e-01 6.23040736e-01 -1.30655780e-01 -3.59472752e-01
1.06183834e-01 4.37598348e-01 4.41239446e-01 1.83907911e-01
-7.90151134e-02 -1.00859606e+00 -5.80579221e-01 7.54740655e-01
-3.48240048e-01 -2.66444594e-01 8.26895416e-01 4.67020392e-01
6.26338542e-01 3.78183693e-01 6.00516379e-01 -4.26108122e-01
3.09042186e-01 -9.14600432e-01 -4.25143927e-01 2.96000749e-01
-4.42888498e-01 -2.19315365e-01 5.14996707e-01 -3.21216345e-01
-9.98712599e-01 -8.57610703e-02 -6.95975482e-01 1.21795714e-01
-3.59405726e-01 4.19221789e-01 -6.42075360e-01 -4.43801969e-01
4.83563513e-01 3.45697016e-01 -6.86751544e-01 -5.21255434e-01
3.20557714e-01 9.00923908e-01 5.19902408e-01 -5.48309028e-01
1.12599862e+00 4.67401892e-01 1.54680282e-01 -9.25852776e-01
-7.43296385e-01 -7.48621762e-01 -4.98901695e-01 2.39411253e-03
5.01846850e-01 -9.89721298e-01 -8.63266766e-01 6.31991446e-01
-1.00163794e+00 -2.43316576e-01 1.16755618e-02 7.12903798e-01
-1.86367348e-01 1.61566615e-01 -3.06462377e-01 -1.02861392e+00
-9.97073278e-02 -9.49013054e-01 1.06886733e+00 3.96411180e-01
-1.15503229e-01 -9.59691525e-01 2.85138994e-01 3.97392362e-01
8.21240962e-01 2.47643471e-01 2.92550415e-01 -7.86578357e-01
-5.97938061e-01 -7.56095231e-01 3.09855435e-02 2.73616225e-01
4.58638221e-01 -3.92883420e-01 -1.23960674e+00 -1.98586211e-02
7.43379304e-03 5.32976985e-01 4.36302610e-02 7.00777709e-01
1.12121809e+00 -2.40028501e-01 -8.54544818e-01 6.58133328e-01
1.36273253e+00 1.44972980e-01 7.43559837e-01 4.02398348e-01
9.55735743e-01 -1.86966851e-01 2.99668759e-01 3.27706337e-01
6.60683692e-01 7.18254864e-01 4.88545507e-01 -3.01380515e-01
1.50596961e-01 -6.49925411e-01 -2.20479369e-01 6.66064501e-01
-3.32043916e-01 -6.68503284e-01 -8.93152058e-01 2.64684945e-01
-1.91063261e+00 -8.08509529e-01 -4.41435128e-01 2.65427995e+00
3.08478355e-01 -3.33024152e-02 -3.26477259e-01 2.31405333e-01
6.35765612e-01 1.98703095e-01 -1.31906927e-01 4.86879468e-01
-1.16640724e-01 1.18479319e-01 1.17659688e+00 3.93792629e-01
-1.09484994e+00 2.76442766e-01 5.06003904e+00 7.39075005e-01
-8.26855183e-01 3.17324281e-01 3.49386781e-01 2.98495203e-01
-1.87764332e-01 -5.33072352e-01 -1.07014763e+00 1.06688631e+00
1.22957766e+00 7.05903530e-01 5.24899542e-01 1.00205696e+00
3.93537015e-01 -2.44606733e-01 -8.45289052e-01 1.07981932e+00
-1.88130453e-01 -1.08231008e+00 -6.13617003e-01 4.43182379e-01
8.75650883e-01 4.88561720e-01 2.48624817e-01 6.16005301e-01
1.62926093e-01 -6.98693573e-01 5.59352219e-01 6.78106606e-01
4.28135157e-01 -7.13045776e-01 1.19725883e+00 5.15149057e-01
-1.35955322e+00 -2.37510398e-01 -9.83182341e-03 2.82850415e-01
2.99065441e-01 8.46740305e-01 -6.56552374e-01 1.03470218e+00
8.17498446e-01 2.47443974e-01 -5.25748551e-01 1.46758986e+00
-5.29276252e-01 7.53948033e-01 -8.83192241e-01 3.24810334e-02
2.56752614e-02 -1.04498781e-01 3.18557620e-01 1.04989481e+00
1.03719532e+00 -4.75827098e-01 -1.27892569e-02 5.91366291e-01
-9.46751162e-02 -4.09568310e-01 -4.73364264e-01 6.78852379e-01
9.64958668e-01 1.19015062e+00 -4.02530819e-01 1.85945645e-01
-4.28983003e-01 8.84880066e-01 -3.87985796e-01 1.01720178e+00
-1.09861696e+00 -3.15011829e-01 6.88587189e-01 9.22930837e-02
4.13123101e-01 -5.33602059e-01 -1.93808973e-01 -5.74850023e-01
-5.52120395e-02 -2.39965394e-01 -3.15874219e-01 -7.50122249e-01
-1.08430004e+00 4.35941935e-01 -4.44723815e-01 -1.05861628e+00
-1.30390063e-01 -1.95861384e-01 -6.21610343e-01 1.02561951e+00
-1.43799531e+00 -1.39765167e+00 -6.50744379e-01 6.26501560e-01
-2.54992247e-01 4.15988207e-01 1.16701496e+00 1.08696377e+00
-7.56035984e-01 8.14305305e-01 9.12591457e-01 2.48129085e-01
4.86772448e-01 -1.04326570e+00 6.39830232e-01 9.43588972e-01
3.37036997e-01 9.02302206e-01 5.79475462e-01 -6.43496335e-01
-1.25029254e+00 -1.33325231e+00 1.00376344e+00 -6.23217583e-01
4.43091393e-01 -7.44687319e-01 -4.46333408e-01 8.82198811e-01
-5.14715493e-01 2.06883326e-01 1.02501321e+00 5.06397069e-01
1.98892653e-01 -4.56578344e-01 -1.19459248e+00 6.37672365e-01
1.06782544e+00 -3.78048629e-01 -3.71779539e-02 2.62404948e-01
5.06174743e-01 -3.11370462e-01 -8.53871346e-01 4.27644998e-01
5.88081121e-01 -6.64171875e-01 1.49051630e+00 4.25694436e-01
-5.02045572e-01 -6.08706772e-01 -4.71410990e-01 -1.18263113e+00
-3.56101751e-01 -4.63451058e-01 -4.78671551e-01 1.74039614e+00
4.38126057e-01 -1.07222831e+00 8.11338723e-01 5.85714877e-01
-1.08305432e-01 -5.45260727e-01 -1.27812922e+00 -8.53471577e-01
-7.37492859e-01 -9.72989857e-01 1.02551723e+00 4.80904818e-01
-7.55666554e-01 2.40493380e-03 -5.45397580e-01 9.05033767e-01
6.87642395e-01 -5.24167240e-01 1.14892972e+00 -1.36634171e+00
-2.52970308e-01 1.31785171e-02 -3.09950024e-01 -1.59614944e+00
-1.57021716e-01 -3.96041065e-01 4.47836965e-01 -1.90655291e+00
-5.10467589e-01 -1.31456339e+00 -6.71456337e-01 3.45198870e-01
-2.15011034e-02 2.84692734e-01 -2.07467154e-01 9.08561125e-02
-6.88421249e-01 5.77285469e-01 6.17392957e-01 -3.00742239e-02
-4.19981271e-01 1.09238493e+00 -6.55294001e-01 7.16873825e-01
6.33629501e-01 -6.82606518e-01 -2.27458760e-01 -3.21260482e-01
3.68707925e-01 -1.96421668e-01 4.31733340e-01 -1.78833103e+00
5.63810349e-01 3.36884826e-01 8.35002899e-01 -8.63410652e-01
6.42158628e-01 -1.35528135e+00 7.19620943e-01 1.45862952e-01
4.40585166e-01 -1.86868966e-01 1.17973737e-01 1.00427401e+00
2.14369103e-01 -1.26261180e-02 6.32754946e-03 3.05958062e-01
-3.44148159e-01 3.23326141e-01 -5.03510535e-01 -5.02802968e-01
6.81742370e-01 -3.84054333e-01 -2.20177397e-01 -4.95536417e-01
-4.15318996e-01 6.25666454e-02 1.96803495e-01 3.33877563e-01
2.81608999e-01 -1.52009833e+00 -5.93923740e-02 2.88594216e-01
-2.27859560e-02 1.35996953e-01 4.78531122e-01 9.38433468e-01
3.35560255e-02 5.90196908e-01 3.17052245e-01 -5.44280350e-01
-8.78396094e-01 1.51271716e-01 3.73393506e-01 -4.74995971e-01
-2.75871396e-01 8.06697547e-01 -6.98297098e-02 -1.05865991e+00
5.28081119e-01 -6.33550465e-01 -1.76859964e-02 -2.28207633e-01
2.73416877e-01 7.24038780e-01 1.46491468e-01 -6.27387404e-01
-4.99224812e-01 6.49655402e-01 6.44890249e-01 -1.23373354e-02
1.13299906e+00 -4.14177686e-01 1.34100989e-02 5.07851064e-01
8.68155956e-01 6.77738249e-01 -9.54266310e-01 -2.41443053e-01
2.99539361e-02 -2.13123128e-01 -1.50754318e-01 -8.64749730e-01
-7.05652714e-01 1.89825609e-01 7.79092968e-01 1.52601361e-01
6.21202230e-01 -1.95338130e-01 8.55346680e-01 4.63234186e-01
1.12191677e+00 -7.58202612e-01 -4.58527058e-01 4.62034613e-01
3.59030098e-01 -9.52543557e-01 -1.70975879e-01 -8.88119340e-02
8.90966058e-02 6.31987035e-01 8.76370594e-02 3.33638757e-01
1.14316165e+00 5.84616549e-02 -1.81590766e-02 2.39854887e-01
4.34017807e-01 -3.13784741e-02 8.50033835e-02 8.72788668e-01
1.55497327e-01 2.32727125e-01 4.98576880e-01 1.10036564e+00
-3.00041020e-01 -1.90062039e-02 -1.47916442e-02 1.03699458e+00
-1.13833047e-01 -1.10840881e+00 -7.69974291e-01 5.40530741e-01
-3.43141317e-01 -1.48402154e-01 6.94192231e-01 5.99536240e-01
5.44944465e-01 1.36001348e+00 -2.36061171e-01 -5.31295121e-01
5.47189057e-01 -1.68212608e-01 9.29320138e-03 -3.23405057e-01
-4.84279543e-01 -1.51616454e-01 -1.19759105e-01 -4.26788419e-01
-3.17489445e-01 -5.70714891e-01 -1.10743535e+00 -4.13022757e-01
-3.99559528e-01 2.22163200e-01 1.24844575e+00 9.63615716e-01
5.48845112e-01 9.37751710e-01 7.28944600e-01 -1.18391132e+00
-1.27835318e-01 -1.03443873e+00 -8.51287544e-01 -3.17638725e-01
3.17652553e-01 -7.11840808e-01 -3.77829909e-01 -6.92272663e-01] | [6.410688400268555, 0.922505259513855] |
e614787c-a448-46a8-8538-a7e82406a396 | deep-learning-based-assessment-of-hepatic | 2202.02377 | null | https://arxiv.org/abs/2202.02377v1 | https://arxiv.org/pdf/2202.02377v1.pdf | Deep Learning-based Assessment of Hepatic Steatosis on chest CT | Purpose: Automatic methods are required for the early detection of hepatic steatosis to avoid progression to cirrhosis and cancer. Here, we developed a fully automated deep learning pipeline to quantify hepatic steatosis on non-contrast enhanced chest computed tomography (CT) scans. Materials and Methods: We developed and evaluated our pipeline on chest CT images of 1,431 randomly selected National Lung Screening Trial (NLST) participants. A dataset of 451 CT scans with volumetric liver segmentations of expert readers was used for training a deep learning model. For testing, in an independent dataset of 980 CT scans hepatic attenuation was manually measured by an expert reader on three cross-sectional images at different hepatic levels by selecting three circular regions of interest. Additionally, 100 randomly selected cases of the test set were volumetrically segmented by expert readers. Hepatic steatosis on the test set was defined as mean hepatic attenuation of < 40 Hounsfield unit. Spearman correlation was conducted to analyze liver fat quantification accuracy and the Cohen's Kappa coefficient was calculated for hepatic steatosis prediction reliability. Results: Our pipeline demonstrated strong performance and achieved a mean dice score of 0.970 for the volumetric liver segmentation. The spearman correlation of the liver fat quantification was 0.954 (P <0.0001) between the automated and expert reader measurements. The cohen's kappa coefficient was 0.875 for automatic assessment of hepatic steatosis. Conclusion: We developed a fully automatic deep learning-based pipeline for the assessment of hepatic steatosis in chest CT images. With the fast and cheap screening of hepatic steatosis, our pipeline has the potential to help initiate preventive measures to avoid progression to cirrhosis and cancer. | ['Hugo J. W. L. Aerts', 'Michael T. Lu', 'Roman Zeleznik', 'Jana Taron', 'Jakob Weiss', 'Zhongyi Zhang'] | 2022-02-04 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [-5.52199304e-01 -3.33959401e-01 -2.46263906e-01 -3.35144550e-01
-1.04066324e+00 -7.03751683e-01 2.57305205e-02 4.65138942e-01
-3.24475288e-01 2.78021485e-01 3.90950531e-01 -5.07520556e-01
9.86661762e-02 -9.71773922e-01 -2.22637847e-01 -7.78991997e-01
-9.10570383e-01 9.20976937e-01 7.68770948e-02 6.61039412e-01
-3.72012585e-01 4.48869228e-01 -2.11383421e-02 3.48407805e-01
9.30727899e-01 8.67652953e-01 1.35686606e-01 1.03910851e+00
3.37314963e-01 7.06518829e-01 -7.75457770e-02 -1.44050419e-01
5.93029022e-01 -8.36226761e-01 -6.02344096e-01 2.75780946e-01
2.96786696e-01 -8.42807949e-01 1.18650664e-02 7.92844832e-01
9.44113970e-01 -4.50325072e-01 5.67343533e-01 -5.33234894e-01
-5.21366656e-01 9.41255867e-01 -5.44295251e-01 5.61964750e-01
-1.74688220e-01 8.86464059e-01 4.33500260e-01 -9.38588858e-01
1.24945208e-01 3.67300659e-01 1.25047219e+00 -1.43477262e-03
-1.31308293e+00 -5.60500681e-01 -9.87903774e-01 8.42387974e-02
-1.48776245e+00 4.76122163e-02 -1.55165568e-01 -8.72470021e-01
5.54851592e-01 1.24875106e-01 1.50766873e+00 3.55100967e-02
5.58880866e-01 1.17704608e-01 1.36368811e+00 -5.19319363e-02
9.22503509e-03 -1.82929605e-01 -4.11871523e-01 1.47276354e+00
6.18817806e-01 1.38913676e-01 3.33481580e-01 -3.87456626e-01
1.12506878e+00 -1.99110091e-01 -2.87873626e-01 -4.85504836e-01
-1.59893024e+00 1.08960581e+00 6.16249382e-01 3.42956364e-01
-6.63456857e-01 3.11900247e-02 6.89749360e-01 7.56573752e-02
1.28387466e-01 3.03830594e-01 -2.59118855e-01 3.97810876e-01
-1.18223464e+00 -6.30276203e-01 7.42270648e-01 2.68517524e-01
2.68067271e-02 1.02908418e-01 -1.59628153e-01 4.48487580e-01
5.24549723e-01 6.47218227e-01 7.49545455e-01 -8.55811834e-01
-1.94399074e-01 6.44736826e-01 -1.22553088e-01 -5.07283628e-01
-6.17997706e-01 -5.40898025e-01 -1.28242803e+00 1.77371070e-01
6.68472350e-01 -4.04579848e-01 -8.75965953e-01 8.83918166e-01
2.54126489e-01 -3.15802813e-01 -4.02249783e-01 1.56377733e+00
7.94399738e-01 8.10555220e-02 6.10790372e-01 -2.13318795e-01
1.68342853e+00 -6.84268117e-01 -1.36828035e-01 6.19423568e-01
6.88623488e-01 -9.59901929e-01 9.56743479e-01 1.83824137e-01
-1.15183258e+00 -8.01600218e-02 -7.33593345e-01 3.16884905e-01
2.17378229e-01 5.09690344e-01 3.62382501e-01 9.10197020e-01
-1.20076287e+00 5.06637216e-01 -1.31085432e+00 -4.48663086e-01
6.88052118e-01 5.04405141e-01 -2.41311833e-01 -1.00182205e-01
-9.53385413e-01 1.20375884e+00 1.42811388e-01 -2.11308122e-01
-1.23605180e+00 -1.27270889e+00 -7.57357478e-01 2.13488013e-01
-1.12469390e-01 -1.33631825e+00 1.04326248e+00 -7.32187450e-01
-1.18873477e+00 1.09175336e+00 3.57167751e-01 -6.80199385e-01
7.69670844e-01 3.65020901e-01 1.58307850e-01 5.91860235e-01
2.07790881e-01 1.56471476e-01 2.35445842e-01 -7.99699068e-01
-3.68779540e-01 -2.85894185e-01 -6.80796564e-01 1.62715614e-01
3.83157521e-01 2.92870492e-01 1.64133742e-01 -2.97136754e-01
-3.14602777e-02 -8.41059208e-01 -3.44824791e-01 3.78473043e-01
-1.59555569e-01 9.68590453e-02 3.21842134e-01 -1.35531580e+00
6.92741692e-01 -1.56607032e+00 -4.98837382e-01 4.90894288e-01
7.71966159e-01 2.36582950e-01 1.90686077e-01 -3.38146597e-01
-1.28934160e-01 7.87873805e-01 -1.46200582e-01 5.04977643e-01
-1.06341556e-01 -4.03860480e-01 5.90696037e-01 1.07814837e+00
-4.99242023e-02 1.24660516e+00 -1.22810388e+00 -8.13923240e-01
6.17569923e-01 6.73670471e-01 -3.01470280e-01 4.05736208e-01
3.64367515e-01 5.68052888e-01 -6.71053231e-02 7.10056126e-01
4.54801887e-01 -6.30318820e-01 7.30438054e-01 -4.49660242e-01
-3.46913189e-01 1.07754722e-01 -5.06404519e-01 1.37866998e+00
-2.81231433e-01 5.15525281e-01 2.40477920e-01 -2.80210644e-01
6.43138289e-01 7.12249875e-01 1.03152251e+00 -4.38796818e-01
3.49198550e-01 3.29262137e-01 6.78932369e-01 -5.16817749e-01
-7.70945668e-01 -7.55662143e-01 3.36075485e-01 6.70544088e-01
-1.28519922e-01 -4.08764601e-01 3.39976221e-01 -1.69834390e-01
1.21808279e+00 -2.26839200e-01 8.06235731e-01 -6.12627268e-01
4.34646279e-01 5.51048100e-01 2.36530334e-01 4.32310611e-01
-7.73418546e-01 6.10218108e-01 5.88112831e-01 -1.04405189e+00
-1.52279532e+00 -1.42878854e+00 -3.80778283e-01 4.47479725e-01
-5.86718380e-01 -1.83421224e-01 -4.62523758e-01 -8.83996606e-01
-2.89765187e-02 3.90653685e-02 -4.08345312e-01 4.95630711e-01
-5.26395202e-01 -1.39994800e+00 6.36456013e-01 4.28986430e-01
6.27113879e-01 -4.13894802e-01 -8.80663991e-01 1.42209575e-01
-2.88814783e-01 -4.90959674e-01 -4.76505399e-01 1.87814366e-02
-8.52360606e-01 -1.67151070e+00 -1.23987353e+00 -8.51594865e-01
6.72232449e-01 -1.07354216e-01 1.56016660e+00 4.53948408e-01
-7.18783557e-01 2.75833309e-01 -7.73881402e-05 -5.29511869e-02
-6.67559743e-01 -1.51774988e-01 4.56855027e-03 -7.91234970e-01
1.90992594e-01 -2.97791809e-01 -1.54502761e+00 2.45975748e-01
-3.54484141e-01 -8.70690122e-03 1.02286053e+00 7.06783593e-01
8.61310184e-01 -5.02897650e-02 -7.65055716e-02 -3.25193524e-01
2.42130324e-01 -5.65545440e-01 -7.60519803e-01 1.33099183e-01
-7.14062512e-01 -3.61686170e-01 6.73608780e-01 -1.37311071e-01
-3.28553796e-01 1.55415043e-01 1.55506015e-01 -2.87234128e-01
7.13513121e-02 5.25955796e-01 6.50106370e-01 -2.55065173e-01
5.98154962e-01 1.51604086e-01 3.65011513e-01 -1.74490824e-01
4.59371321e-02 1.16076566e-01 1.77596465e-01 -1.77756563e-01
6.10824883e-01 1.77247122e-01 4.23947155e-01 -5.44347048e-01
-1.07494898e-01 -5.47086239e-01 -8.84833157e-01 -8.50042179e-02
1.14292526e+00 -9.84365761e-01 -6.35490358e-01 1.22924179e-01
-5.62345207e-01 -9.23013210e-01 -2.68765688e-01 9.75191176e-01
-3.78170788e-01 5.60575426e-01 -1.04481125e+00 1.61550590e-03
-1.10771096e+00 -1.58675706e+00 4.67591196e-01 -1.40834302e-01
-2.92949378e-01 -1.22238302e+00 5.23902595e-01 -2.46895067e-02
8.45370412e-01 5.74797511e-01 1.24400759e+00 -8.32197130e-01
-3.62892598e-01 -1.07043535e-01 -7.41240382e-01 2.38743082e-01
1.59220949e-01 8.49969015e-02 4.84354142e-03 -1.15072265e-01
-1.69620946e-01 -6.19662702e-02 4.99473065e-01 8.05288672e-01
9.39181328e-01 -3.38266015e-01 2.34013647e-01 7.58531690e-01
1.84169388e+00 9.81772915e-02 2.91777104e-01 2.45997176e-01
3.25315446e-01 -9.68221873e-02 -1.63093448e-01 5.54958761e-01
3.18282664e-01 7.80705214e-02 3.14196169e-01 -6.42950416e-01
-6.45463765e-01 1.50822103e-01 4.45233136e-02 6.76077068e-01
-3.34847867e-01 3.84195715e-01 -1.58789432e+00 5.77609599e-01
-7.19341695e-01 -7.21776724e-01 -6.89171374e-01 2.06491423e+00
8.55961084e-01 -5.03235042e-01 3.43051851e-01 -4.68065560e-01
7.05162704e-01 -3.60601187e-01 -2.05333576e-01 1.80759863e-03
2.46920913e-01 5.24942040e-01 6.41272783e-01 4.03468311e-01
-1.20202172e+00 1.83365047e-01 6.36631298e+00 -6.88940436e-02
-1.28663898e+00 4.86090630e-01 9.36799407e-01 -3.49396095e-02
1.80444121e-01 -2.74164528e-01 1.66729137e-01 5.03873706e-01
9.06534910e-01 -2.89330989e-01 3.91136825e-01 5.73445261e-01
6.25252306e-01 -4.45269495e-01 -7.13955522e-01 3.54961455e-01
-2.59546191e-01 -1.31985641e+00 -1.94543138e-01 1.09313779e-01
7.42438912e-01 5.82658052e-01 8.51501897e-02 1.17689535e-01
6.17879212e-01 -1.16154087e+00 -7.95090497e-02 3.36247176e-01
1.58553779e+00 -3.65615636e-01 1.35940528e+00 -1.56190544e-01
-1.21262813e+00 3.55971038e-01 -1.36104465e-01 3.75455499e-01
-8.13968405e-02 9.70400929e-01 -1.95053411e+00 1.36978179e-01
5.08841336e-01 5.97398356e-02 -5.88048041e-01 1.50139141e+00
-2.95900762e-01 8.38376880e-01 -4.24880207e-01 2.72938967e-01
2.94777334e-01 -1.85315207e-01 2.26924822e-01 1.48971450e+00
5.22346258e-01 3.55104744e-01 3.64267141e-01 9.11688149e-01
-1.24426663e-01 5.70341408e-01 -1.01293162e-01 2.39224330e-01
1.64626285e-01 1.69350755e+00 -1.16421771e+00 -6.28932655e-01
-5.20396233e-01 3.71295363e-01 -3.16066772e-01 -1.02146968e-01
-9.39249277e-01 1.62472397e-01 -9.16831940e-02 7.44382739e-01
-2.76918620e-01 -8.59336257e-02 -6.26861930e-01 -1.04754329e+00
-8.33265841e-01 -8.56562972e-01 7.36938775e-01 -5.66254377e-01
-1.20396876e+00 2.90774822e-01 -3.32672477e-01 -9.38770354e-01
-3.83163467e-02 -4.62131649e-01 -8.25797081e-01 1.23219299e+00
-1.47166216e+00 -1.05431461e+00 -5.85792601e-01 2.59745806e-01
3.61652851e-01 3.52046192e-02 9.67125237e-01 1.68245912e-01
-2.79107153e-01 2.69049779e-03 -2.54909508e-02 9.07100260e-01
7.54048467e-01 -1.83766532e+00 -3.43101799e-01 6.88741565e-01
-5.73787689e-01 6.87373519e-01 1.69987768e-01 -9.98072386e-01
-1.28549922e+00 -1.45982754e+00 7.84200847e-01 -1.31548241e-01
6.45178437e-01 7.21760988e-01 -4.52858716e-01 6.47801161e-01
3.80349666e-01 5.02936184e-01 1.52661741e+00 -5.69576204e-01
-1.29750103e-01 2.68188000e-01 -1.60287869e+00 8.16924497e-02
-4.69386727e-02 1.05674844e-03 -3.81265223e-01 5.82479477e-01
-2.07235903e-01 -6.47683144e-01 -1.72294605e+00 4.16547477e-01
9.46190476e-01 -9.45633590e-01 1.08227527e+00 -3.30373526e-01
5.16674340e-01 -3.52837443e-01 1.55463785e-01 -1.17458296e+00
-7.34955370e-01 -2.21138708e-02 1.22607306e-01 1.70938194e-01
1.56589702e-01 -8.22355300e-02 7.22246885e-01 4.23423290e-01
-6.47766814e-02 -8.79516244e-01 -7.21869707e-01 -1.58758864e-01
5.34339845e-01 3.21702957e-01 2.57033646e-01 9.53750730e-01
-2.19142679e-02 -2.38220677e-01 4.59148884e-01 2.23016500e-01
9.07078505e-01 4.04461734e-02 2.63415307e-01 -9.25401747e-01
4.14953046e-02 -6.59999728e-01 -2.96955556e-01 -5.70395216e-02
-3.89717102e-01 -1.44647050e+00 -3.26445043e-01 -1.87912834e+00
1.12743902e+00 -3.96455973e-01 -2.33029455e-01 5.57400584e-01
-2.55079232e-02 7.94737577e-01 2.16485649e-01 3.61766040e-01
-2.39132628e-01 -1.48042098e-01 1.33951783e+00 -1.11831315e-01
-9.24124494e-02 -1.01808712e-01 -2.64651954e-01 7.59653807e-01
9.89945769e-01 -3.68573517e-01 1.90963358e-01 -1.10784724e-01
-6.11027218e-02 2.89701998e-01 8.38188946e-01 -9.39276516e-01
-1.75073415e-01 -5.15920809e-03 1.39754152e+00 -4.81667221e-01
-6.59150481e-01 -7.64796197e-01 4.32856053e-01 1.61779153e+00
-1.34810463e-01 1.69648275e-01 -7.06186593e-02 -2.26774037e-01
4.70259368e-01 4.68392298e-02 1.34058487e+00 -9.52223539e-01
-2.87517846e-01 4.76781398e-01 -8.83234918e-01 -9.02362764e-02
8.91158223e-01 4.31461811e-01 -1.05158791e-01 -2.09455818e-01
-1.02775502e+00 1.56582877e-01 4.95561510e-01 -7.01371729e-01
5.72034180e-01 -1.17827487e+00 -1.25339961e+00 -4.92490875e-03
-4.54358220e-01 -1.27457470e-01 -4.53044660e-02 2.02542138e+00
-1.35128856e+00 4.77779597e-01 -3.32232565e-01 -8.56529295e-01
-1.23519421e+00 3.20282340e-01 1.11776531e+00 -7.67465234e-01
-7.10508883e-01 3.73792231e-01 -1.81769431e-01 -1.34501234e-01
-4.37758476e-01 -5.56163669e-01 2.95033395e-01 -2.61221319e-01
4.78986055e-01 6.19606733e-01 1.23400316e-02 -5.44146895e-01
-3.17652255e-01 4.40915197e-01 4.49767292e-01 2.46917978e-01
1.13332784e+00 -2.01256350e-01 -5.23941159e-01 -4.50765900e-02
1.01348555e+00 1.61227450e-01 -7.52348781e-01 -2.80259885e-02
-2.74505824e-01 -2.87305325e-01 4.75682765e-01 -1.48871660e+00
-1.12717605e+00 7.87892282e-01 1.22046447e+00 1.39589235e-01
9.20490324e-01 -2.01147214e-01 6.59472585e-01 -3.29687744e-01
1.55399188e-01 -1.45385459e-01 -1.41491234e-01 1.70112610e-01
5.88803530e-01 -1.38865113e+00 4.08001304e-01 -2.51511753e-01
-6.25340044e-01 1.53181481e+00 1.36226565e-01 -2.87601024e-01
3.51858437e-01 5.22495389e-01 3.01019609e-01 -2.76393145e-01
-2.09548667e-01 -1.50650710e-01 2.78820008e-01 6.41320527e-01
1.03738928e+00 4.99666989e-01 -5.27866781e-01 5.98097861e-01
-1.03019569e-02 2.16611505e-01 3.41485679e-01 3.63022596e-01
-6.94191217e-01 -6.36234820e-01 -5.59165716e-01 8.29066992e-01
-8.63570154e-01 -2.55290508e-01 -2.03561127e-01 1.21804249e+00
2.53134668e-01 3.86112928e-01 -6.80385949e-03 3.21180671e-01
-2.33257264e-01 3.09437156e-01 5.24796784e-01 -5.71249843e-01
-9.51054037e-01 3.95200431e-01 -1.73077092e-01 -3.55328918e-01
-2.48885825e-01 -6.19878948e-01 -1.31316626e+00 -4.48877513e-01
6.31604493e-02 1.54388383e-01 5.33367217e-01 5.22464633e-01
-3.46208811e-02 3.53709102e-01 6.24561727e-01 -4.45303887e-01
-5.87980509e-01 -9.86452401e-01 -5.72735786e-01 8.49041417e-02
3.61465394e-01 -1.20573238e-01 -3.83553892e-01 3.24661344e-01] | [14.454910278320312, -2.68129825592041] |
bc077507-3a40-4cd7-b46e-1e0b5cfc4c44 | graph-convolutional-gaussian-processes | 1905.05739 | null | https://arxiv.org/abs/1905.05739v1 | https://arxiv.org/pdf/1905.05739v1.pdf | Graph Convolutional Gaussian Processes | We propose a novel Bayesian nonparametric method to learn translation-invariant relationships on non-Euclidean domains. The resulting graph convolutional Gaussian processes can be applied to problems in machine learning for which the input observations are functions with domains on general graphs. The structure of these models allows for high dimensional inputs while retaining expressibility, as is the case with convolutional neural networks. We present applications of graph convolutional Gaussian processes to images and triangular meshes, demonstrating their versatility and effectiveness, comparing favorably to existing methods, despite being relatively simple models. | ['Ian Walker', 'Ben Glocker'] | 2019-05-14 | null | null | null | null | ['superpixel-image-classification'] | ['computer-vision'] | [ 1.17193639e-01 4.49496001e-01 -1.01240590e-01 -3.11085105e-01
-2.22212955e-01 -5.63698292e-01 9.44067955e-01 -1.23197131e-01
-1.42028585e-01 6.74342155e-01 1.09604657e-01 -3.35297704e-01
-4.27149355e-01 -9.18224633e-01 -7.55399466e-01 -7.78733373e-01
-4.33729649e-01 1.05494404e+00 2.28350744e-01 3.52643490e-01
3.51459272e-02 1.02250278e+00 -1.10147130e+00 -2.17877388e-01
4.41407084e-01 8.17622006e-01 -3.54075462e-01 9.07329321e-01
-3.33002210e-02 4.38885927e-01 -1.47431746e-01 -3.18132073e-01
-6.44731615e-03 3.21786642e-01 -4.76672083e-01 2.83473670e-01
7.57413805e-01 -3.93774696e-02 -8.69579256e-01 1.04056120e+00
5.32435365e-02 5.69136441e-01 1.61256778e+00 -1.30385709e+00
-1.30684447e+00 1.43328413e-01 -3.73301595e-01 -1.29262328e-01
2.27333605e-02 -2.43304148e-01 9.85374153e-01 -8.50511253e-01
5.13166726e-01 2.05588579e+00 1.13628590e+00 2.23932460e-01
-1.98405683e+00 -2.56091267e-01 2.53465265e-01 -2.69198030e-01
-1.32005954e+00 -1.91414490e-01 6.16018772e-01 -8.13445747e-01
8.88677895e-01 -1.69283092e-01 2.80487955e-01 1.30976474e+00
7.14584768e-01 5.38032234e-01 8.74060869e-01 -9.96641535e-03
4.47938234e-01 -2.93186337e-01 4.64480594e-02 7.20906436e-01
4.02714133e-01 8.54507387e-02 -3.38435799e-01 -6.17885530e-01
1.37423074e+00 1.89313099e-01 2.34818961e-02 -9.73222256e-01
-1.04106653e+00 1.05906951e+00 3.18871081e-01 -9.10977945e-02
-4.67158765e-01 8.82231236e-01 1.74950972e-01 1.52329296e-01
9.26728725e-01 2.59326279e-01 -5.50674438e-01 -5.37713096e-02
-7.54074335e-01 2.70647109e-01 1.23122740e+00 1.46207988e+00
7.37629712e-01 4.22697186e-01 -1.76360056e-01 5.72612107e-01
6.56323075e-01 6.63245857e-01 -5.94454631e-02 -1.08449495e+00
8.07532966e-02 -3.50152589e-02 5.83850928e-02 -9.77889597e-01
-3.53648633e-01 -2.79284686e-01 -9.54361439e-01 3.68192852e-01
4.27389890e-01 -1.99092373e-01 -1.12426531e+00 1.45763659e+00
-6.68823943e-02 2.52267420e-01 -8.10199231e-02 2.96398699e-01
5.66058338e-01 5.03574789e-01 1.34175614e-01 1.41182244e-01
9.57410276e-01 -4.50089663e-01 -6.86215699e-01 -2.66430285e-02
-1.48219205e-02 -6.99916482e-01 6.54041529e-01 3.66141915e-01
-8.45910549e-01 -5.25188923e-01 -7.44503081e-01 -1.63520530e-01
-4.76021200e-01 -8.30314234e-02 1.08977783e+00 6.68934643e-01
-1.45950830e+00 1.02880847e+00 -1.16489398e+00 -4.67021286e-01
6.61718130e-01 5.66334605e-01 -2.70301700e-01 -6.52506668e-03
-7.21864223e-01 7.73714006e-01 8.97404030e-02 1.44969538e-01
-1.14795816e+00 -5.65271139e-01 -1.30877948e+00 6.27153516e-02
1.19654454e-01 -6.63392067e-01 1.28781950e+00 -5.72325468e-01
-1.77766144e+00 4.79136914e-01 -8.90808105e-02 -3.43914241e-01
5.80841064e-01 -2.43904427e-01 -6.72776103e-02 -5.17845526e-02
4.68669012e-02 4.18805510e-01 1.31894612e+00 -1.02526760e+00
-3.12580392e-02 -2.88640231e-01 -9.48024616e-02 8.32340401e-03
1.03286363e-01 -1.29508182e-01 -3.77295047e-01 -7.13060796e-01
3.63726526e-01 -1.16145802e+00 -6.02555156e-01 3.74247789e-01
-2.94461310e-01 -4.69873607e-01 9.95359838e-01 -4.58992541e-01
2.54674733e-01 -1.92235816e+00 3.67143005e-01 3.40064704e-01
4.16008860e-01 -3.82375211e-01 -7.03538880e-02 4.83600765e-01
5.88516407e-02 2.11090490e-01 -3.27614814e-01 -6.25124633e-01
3.16371620e-01 6.13244355e-01 -3.57546687e-01 1.06165671e+00
7.09789634e-01 1.11239016e+00 -9.72792268e-01 -2.95130640e-01
3.36263001e-01 6.56212926e-01 -2.93749809e-01 1.08849704e-01
-4.71104205e-01 3.06239069e-01 -6.00447357e-01 5.27258337e-01
5.22022903e-01 -2.84625947e-01 -1.98565386e-02 1.13847122e-01
2.21434057e-01 1.76096726e-02 -1.26870561e+00 1.41568255e+00
-3.07519883e-01 9.29692566e-01 3.26861143e-01 -1.01603341e+00
1.08715641e+00 3.06381524e-01 3.76627296e-01 3.48085433e-01
9.75877494e-02 -9.78343710e-02 -2.77990997e-01 -8.85083899e-02
2.60971785e-01 -5.94020605e-01 -6.75071999e-02 1.47013351e-01
5.00415802e-01 -7.66444921e-01 -1.24560677e-01 -5.93295880e-02
9.99034405e-01 6.26156211e-01 1.72413021e-01 -7.51950264e-01
1.25509184e-02 -2.11789504e-01 1.87813833e-01 8.70610595e-01
1.13601396e-02 6.69652641e-01 7.23369837e-01 -3.91523212e-01
-9.27410364e-01 -1.66124189e+00 -2.75223106e-01 9.88025010e-01
-1.51962087e-01 -7.24151284e-02 -3.97909619e-02 -4.89742339e-01
5.81813037e-01 5.81654072e-01 -9.91121829e-01 1.73792057e-02
-3.26708019e-01 -5.02940059e-01 5.84349930e-01 1.04337811e+00
-3.56333666e-02 -8.50124955e-01 3.46953779e-01 2.76187390e-01
7.56822109e-01 -1.23802245e+00 -3.25639606e-01 3.57942045e-01
-1.21822631e+00 -9.76049006e-01 -7.58615017e-01 -8.94886613e-01
5.44439018e-01 -1.89473599e-01 1.11199319e+00 -6.23560190e-01
-1.30386457e-01 9.49036896e-01 3.88271540e-01 -6.13590539e-01
-4.21318531e-01 -1.88105375e-01 2.35211819e-01 -2.12089702e-01
3.44906569e-01 -9.94397819e-01 2.43578684e-02 6.92590699e-02
-6.93299770e-01 -4.61732745e-01 3.76752615e-01 7.90066004e-01
5.48322558e-01 1.54417962e-01 2.42842257e-01 -9.89711165e-01
9.64256167e-01 -5.56520760e-01 -8.76266778e-01 -1.30887702e-01
-3.70360434e-01 1.94022715e-01 5.42806327e-01 -7.33568609e-01
-8.29492271e-01 -7.53452927e-02 4.45530802e-01 -8.47760379e-01
-4.23630267e-01 3.66182685e-01 3.35925934e-03 -1.71468824e-01
7.57959962e-01 -2.35474274e-01 3.77459466e-01 -3.31912577e-01
8.10087383e-01 -9.75732226e-03 5.27869046e-01 -9.67422009e-01
8.36188674e-01 7.02339470e-01 7.78551221e-01 -1.23810923e+00
-1.50827080e-01 -4.02572423e-01 -9.09247935e-01 2.17715323e-01
1.03480160e+00 -8.18302214e-01 -6.81928933e-01 3.19635630e-01
-1.32670820e+00 -4.01494056e-01 -3.61256033e-01 7.57764280e-01
-1.14461255e+00 3.16210270e-01 -9.88265634e-01 -8.26648355e-01
2.70956993e-01 -8.82346332e-01 1.28850198e+00 -6.02273233e-02
-2.80062228e-01 -1.90941358e+00 1.24980146e-02 -5.37431836e-01
4.06971186e-01 3.93921077e-01 1.03529930e+00 -8.11312556e-01
-5.05973041e-01 -1.93946823e-01 -2.77890056e-01 5.17105222e-01
3.49403739e-01 3.48268509e-01 -9.40629780e-01 -9.24117342e-02
-1.17370896e-01 5.67164682e-02 7.95649111e-01 8.91850591e-01
1.09195244e+00 -1.95569526e-02 -5.32320321e-01 6.04545355e-01
1.12612069e+00 -1.52967602e-01 2.79297501e-01 -1.56393513e-01
1.01542139e+00 6.41726196e-01 -4.10665013e-02 1.24435715e-01
8.20258260e-02 2.39488453e-01 2.72766143e-01 -2.06149802e-01
6.83673611e-03 -1.28441304e-01 1.55534714e-01 5.69805443e-01
-2.50683188e-01 -7.65448362e-02 -1.03072143e+00 6.25445545e-01
-2.15340924e+00 -7.04222202e-01 -4.72101450e-01 1.94693947e+00
3.60274345e-01 1.14235625e-01 -1.87849224e-01 -5.62459409e-01
8.84946883e-01 1.82422996e-01 -4.24515247e-01 -5.69525123e-01
-4.76696752e-02 7.32776582e-01 9.13028777e-01 5.53991079e-01
-1.50884581e+00 8.30876887e-01 8.35590172e+00 5.37850201e-01
-6.41289115e-01 -3.07072252e-02 1.97739497e-01 2.94095367e-01
-1.80750847e-01 3.11655179e-03 -6.06288075e-01 -1.32284656e-01
8.83347571e-01 -2.38281891e-01 3.59615475e-01 9.93291736e-01
9.96575654e-02 3.52762192e-01 -1.60018837e+00 8.42946053e-01
-1.03219859e-01 -1.34712732e+00 5.10581881e-02 1.83184102e-01
9.12937760e-01 1.10870868e-01 2.48422474e-01 3.96780878e-01
1.08477843e+00 -1.44007897e+00 4.58747566e-01 7.65113115e-01
6.16285920e-01 -7.14637041e-01 5.15165031e-01 2.18365733e-02
-1.00804436e+00 4.89000946e-01 -7.87221253e-01 -1.06359601e-01
5.13255447e-02 5.37705898e-01 -1.20708072e+00 4.49804544e-01
5.85011840e-01 1.03318286e+00 -1.79369554e-01 9.53318477e-01
-4.49841619e-01 5.90917349e-01 -4.93297696e-01 -1.85977384e-01
3.83797944e-01 -5.75928092e-01 8.16333473e-01 1.51783705e+00
2.33790934e-01 -3.83646131e-01 4.18914676e-01 1.26672971e+00
-1.87600583e-01 -1.13633409e-01 -1.31203949e+00 -3.21406692e-01
1.20414265e-01 9.94530320e-01 -8.88961136e-01 -3.43436122e-01
-5.64548850e-01 7.53131032e-01 2.92231023e-01 9.81845856e-01
-4.06026781e-01 -1.56988069e-01 9.63528156e-01 1.06065581e-02
6.65448606e-01 -8.98645639e-01 -3.45450550e-01 -8.31537902e-01
-2.87547588e-01 -4.53646898e-01 1.21834829e-01 -7.17177033e-01
-1.97390342e+00 2.90303439e-01 5.18030941e-01 -7.64177024e-01
-2.77313113e-01 -1.29009926e+00 -5.79030812e-01 1.19748878e+00
-1.01902890e+00 -1.46715808e+00 8.69906992e-02 5.04602134e-01
3.64833266e-01 -1.70270070e-01 8.93435359e-01 -4.95752215e-01
1.27115443e-01 -1.27875701e-01 4.24849093e-01 4.98139001e-02
5.18436491e-01 -1.84641147e+00 1.03421569e+00 6.14227235e-01
3.34212065e-01 8.08446288e-01 7.79880583e-01 -9.84265864e-01
-1.61567438e+00 -1.24718201e+00 1.06805585e-01 -7.04009354e-01
1.26376367e+00 -6.79142237e-01 -9.34063971e-01 1.13822961e+00
-2.86899996e-03 8.23065639e-02 4.32425439e-01 7.10583508e-01
-3.94250482e-01 5.56504786e-01 -1.03274202e+00 6.91965044e-01
8.64843786e-01 -7.50194192e-01 -7.27456748e-01 4.26380783e-01
7.11618364e-01 -3.19683641e-01 -1.04193997e+00 3.83690029e-01
3.87767047e-01 -1.20795414e-01 9.21457469e-01 -1.06387770e+00
1.09434597e-01 -3.32462460e-01 -2.74482876e-01 -1.44256783e+00
-7.54197299e-01 -9.79065359e-01 -4.08270448e-01 7.57493794e-01
2.68788189e-01 -8.72150779e-01 8.46573889e-01 6.86023951e-01
-2.49584904e-04 -4.27792460e-01 -9.62098956e-01 -9.67945874e-01
5.29806256e-01 -6.64706349e-01 3.32925677e-01 7.60939121e-01
-5.55312037e-01 3.73302668e-01 -2.34827161e-01 6.04185581e-01
9.09810901e-01 -1.17004775e-01 7.33558238e-01 -1.79927778e+00
-3.61418605e-01 -6.19836211e-01 -8.95591795e-01 -1.06402528e+00
6.57766998e-01 -8.41128945e-01 2.81290114e-01 -1.70508623e+00
-2.51693398e-01 -2.70091981e-01 6.42531961e-02 2.02800468e-01
4.58691940e-02 4.02015112e-02 -2.85335869e-01 -8.21631588e-03
-2.05950752e-01 8.65603685e-01 1.16526985e+00 -1.08676985e-01
-1.07165821e-01 2.86928654e-01 -2.91711509e-01 1.00015414e+00
6.93409979e-01 -3.34799558e-01 -5.94834507e-01 -1.97100297e-01
-1.15680970e-01 -3.55852455e-01 4.60774630e-01 -8.08859766e-01
2.17410028e-01 -2.61790693e-01 6.31858706e-01 -3.81404996e-01
6.31996810e-01 -6.83573067e-01 3.38211924e-01 -4.99696806e-02
-1.31826490e-01 2.03978196e-01 3.57872009e-01 1.26403880e+00
-3.29383127e-02 -4.72810082e-02 6.84405982e-01 -2.77934503e-02
-5.66021442e-01 7.33300269e-01 -6.27606511e-01 -1.15954159e-02
8.90867651e-01 -1.56962797e-01 9.92698479e-04 -7.01615512e-01
-1.14010561e+00 -1.55533373e-01 5.78941107e-01 4.50127572e-01
5.52601993e-01 -1.36234772e+00 -6.88231707e-01 1.09341368e-01
1.89374536e-02 2.13115975e-01 -2.52212226e-01 6.50045514e-01
-5.21033287e-01 2.15654165e-01 -4.22540978e-02 -8.87051940e-01
-8.81393194e-01 7.12974131e-01 2.24862337e-01 5.86027056e-02
-7.59400427e-01 7.78688848e-01 6.25225842e-01 -5.46787918e-01
1.67011797e-01 -8.58812273e-01 1.33788750e-01 -2.42941543e-01
3.36831510e-02 3.13181162e-01 -1.64211884e-01 -3.97495419e-01
-1.65678829e-01 3.61218065e-01 2.62513191e-01 -1.60170525e-01
1.43120003e+00 1.97384506e-01 -3.58640134e-01 1.01839328e+00
1.18309379e+00 -1.19089782e-01 -1.54994369e+00 -3.10710728e-01
6.56247512e-02 -1.47990748e-01 2.59107053e-02 9.88332257e-02
-5.85645914e-01 9.08851564e-01 1.39612421e-01 2.39074677e-01
2.58906901e-01 3.31984639e-01 -1.53459474e-01 5.50455213e-01
3.14517945e-01 -7.71974385e-01 -3.99493687e-02 8.85967910e-01
1.03228390e+00 -7.56939411e-01 1.90726772e-01 -6.82946265e-01
-3.41161847e-01 1.57182455e+00 1.21106923e-01 -9.34236467e-01
1.21399951e+00 2.82777399e-01 -3.97238016e-01 -4.45042938e-01
-6.75508380e-01 -1.20818697e-01 6.76789045e-01 1.33834982e+00
4.53842521e-01 4.59656775e-01 2.70884693e-01 1.07038125e-01
-6.31521568e-02 -4.05549765e-01 4.44654346e-01 8.25426221e-01
-2.85024315e-01 -9.00378346e-01 -4.35714513e-01 6.53115273e-01
-4.12750751e-01 -1.66866228e-01 -1.20171085e-01 1.14773428e+00
-4.00188625e-01 6.98967576e-01 2.75096685e-01 9.83908698e-02
2.67917484e-01 6.23759292e-02 6.29025280e-01 -9.53817904e-01
-2.35349266e-03 2.01000392e-01 3.51724148e-01 -4.23219055e-01
-3.61037642e-01 -8.06299329e-01 -8.53153169e-01 -1.89561009e-01
-1.76834911e-02 -7.16463998e-02 5.96614301e-01 8.29371929e-01
6.85356855e-02 5.08166552e-01 8.21652412e-02 -1.19507730e+00
-6.95207715e-01 -1.19623601e+00 -9.01758730e-01 2.27730125e-01
4.27457660e-01 -1.24949443e+00 -4.71742868e-01 2.07401097e-01] | [6.998424530029297, 3.7986552715301514] |
4d9df7e8-332a-46d4-9d8b-29618768f03e | inverse-cubature-and-quadrature-kalman | 2303.10322 | null | https://arxiv.org/abs/2303.10322v1 | https://arxiv.org/pdf/2303.10322v1.pdf | Inverse Cubature and Quadrature Kalman filters | Recent developments in counter-adversarial system research have led to the development of inverse stochastic filters that are employed by a defender to infer the information its adversary may have learned. Prior works addressed this inverse cognition problem by proposing inverse Kalman filter (I-KF) and inverse extended KF (I-EKF), respectively, for linear and non-linear Gaussian state-space models. However, in practice, many counter-adversarial settings involve highly non-linear system models, wherein EKF's linearization often fails. In this paper, we consider the efficient numerical integration techniques to address such nonlinearities and, to this end, develop inverse cubature KF (I-CKF) and inverse quadrature KF (I-QKF). We derive the stochastic stability conditions for the proposed filters in the exponential-mean-squared-boundedness sense. Numerical experiments demonstrate the estimation accuracy of our I-CKF and I-QKF with the recursive Cram\'{e}r-Rao lower bound as a benchmark. | ['Arpan Chattopadhyay', 'Kumar Vijay Mishra', 'Himali Singh'] | 2023-03-18 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [-1.26286149e-01 -9.57348868e-02 3.71837795e-01 -1.04919486e-02
-1.04236853e+00 -1.18652475e+00 6.43366337e-01 -3.80921751e-01
-4.83632088e-01 8.67570519e-01 1.34918302e-01 -9.21122551e-01
-4.34471399e-01 -5.38714111e-01 -8.17601621e-01 -7.39621103e-01
-2.15284824e-01 -5.48373237e-02 -2.10575446e-01 -3.65925729e-01
3.92528981e-01 3.68972570e-01 -4.03338611e-01 -2.61408299e-01
8.08477402e-01 1.05824888e+00 -3.14206362e-01 1.11353219e+00
7.05223262e-01 8.56113970e-01 -6.04313493e-01 -7.14454293e-01
7.74861038e-01 -3.56346160e-01 -2.63440371e-01 -6.23781860e-01
-7.52682164e-02 -5.93714833e-01 -9.02235568e-01 1.59407973e+00
4.41255063e-01 1.10228799e-01 7.52477765e-01 -1.16890705e+00
-5.09186685e-01 3.79316509e-01 -2.58788764e-01 2.74039030e-01
2.31870562e-01 6.61820471e-02 3.10441583e-01 -6.71548307e-01
1.69628896e-02 1.34024298e+00 1.07094836e+00 3.83649200e-01
-9.75581110e-01 -1.15807760e+00 2.73288459e-01 2.45187879e-02
-1.73595023e+00 -5.01354039e-01 2.95331001e-01 -2.88630933e-01
5.00959218e-01 3.60248357e-01 2.12337762e-01 9.50803339e-01
7.82841742e-01 5.72214544e-01 1.59001958e+00 -5.58699183e-02
2.79591054e-01 4.22473282e-01 1.33295923e-01 7.21558869e-01
4.27704960e-01 9.41865206e-01 -3.83510619e-01 -6.52408898e-01
7.91332245e-01 -1.44208848e-01 -5.66882312e-01 7.82925934e-02
-8.41641724e-01 8.94118309e-01 2.67952651e-01 -2.55695641e-01
-3.51287186e-01 3.60568702e-01 1.19741641e-01 5.36425233e-01
4.44439471e-01 2.03300178e-01 -3.71461570e-01 -3.14513221e-02
-6.07154608e-01 5.06707728e-01 1.32668018e+00 9.94437814e-01
3.24362576e-01 2.95358956e-01 5.25429733e-02 -1.37792781e-01
8.72139335e-01 1.14098513e+00 3.26770358e-02 -9.82995450e-01
5.58475494e-01 -4.78471339e-01 5.17773986e-01 -1.02104783e+00
-1.52733237e-01 -1.10063004e+00 -1.10686851e+00 2.62587011e-01
6.59791648e-01 -4.85928774e-01 -6.92406297e-01 1.91210246e+00
4.08230722e-01 8.96428764e-01 4.74014103e-01 8.16025317e-01
-1.31033733e-01 3.73979986e-01 -2.73031056e-01 -3.50089043e-01
1.13265514e+00 -3.22590262e-01 -9.44373488e-01 -2.77714700e-01
3.91123235e-01 -8.46655905e-01 4.97053802e-01 6.02976739e-01
-7.63839483e-01 -1.68866292e-01 -1.19363618e+00 5.39723814e-01
-2.24140525e-01 -5.36700748e-02 4.41105992e-01 1.33637524e+00
-6.45593524e-01 5.51600099e-01 -7.43743539e-01 4.99468565e-01
2.03235045e-01 3.52704644e-01 -1.79682598e-01 3.24496031e-01
-1.77910066e+00 9.61881042e-01 1.57464534e-01 5.32120943e-01
-1.35080349e+00 -7.32151031e-01 -4.51724291e-01 -2.03535378e-01
6.29020333e-01 -8.08717847e-01 1.28869498e+00 -4.45825428e-01
-1.85949910e+00 6.12857491e-02 -2.26756139e-03 -8.93622577e-01
7.29013085e-01 -8.73828053e-01 -6.99146450e-01 -9.09324065e-02
-6.96295351e-02 -5.91139495e-01 1.30115247e+00 -1.14568973e+00
-4.58779216e-01 -7.28234529e-01 3.17899793e-01 4.32086736e-02
1.62007838e-01 -2.61313319e-01 2.03913733e-01 -9.30794001e-01
2.11742982e-01 -1.32231498e+00 -7.44461298e-01 -2.32410505e-01
-6.35658979e-01 5.17300963e-01 6.63938999e-01 -8.13948214e-01
1.34797251e+00 -2.14380407e+00 -1.97135568e-01 6.52729690e-01
5.89471012e-02 5.89003563e-01 3.26898336e-01 5.22775948e-01
4.32144888e-02 -1.06402658e-01 1.72282979e-02 -1.96721584e-01
1.00886494e-01 6.46368638e-02 -1.21754241e+00 1.28036439e+00
-4.33386356e-01 5.28964877e-01 -9.51647878e-01 2.98385546e-02
1.29247040e-01 3.94821286e-01 -6.34881914e-01 1.61275879e-01
3.00658762e-01 4.27270591e-01 -8.64484131e-01 1.02886491e-01
8.53657842e-01 1.85030565e-01 -1.93897441e-01 -1.92473710e-01
-2.80930959e-02 -6.44445196e-02 -1.47157657e+00 1.29216146e+00
-6.23366058e-01 2.36751869e-01 7.36651361e-01 -6.56546414e-01
4.66479957e-01 4.02670830e-01 -7.80200660e-02 -1.36141196e-01
5.84653556e-01 2.03688622e-01 -5.52743301e-02 -2.74498127e-02
4.54697162e-01 -6.93175256e-01 -3.41306150e-01 4.69149798e-01
-1.58376560e-01 -3.10851872e-01 -4.95075583e-01 4.21476305e-01
8.83879185e-01 -8.84766132e-02 5.48285782e-01 -3.47537458e-01
9.25516248e-01 -3.80096644e-01 4.64248210e-01 1.37355411e+00
-4.37187523e-01 6.60578310e-02 9.60200131e-02 7.97938332e-02
-2.81380802e-01 -1.34661198e+00 1.87642038e-01 5.79873204e-01
3.37010771e-01 -1.50325596e-01 -7.14369357e-01 -4.92373317e-01
1.69571429e-01 1.03808677e+00 -5.54254413e-01 -4.77513939e-01
-2.43516728e-01 -5.50433636e-01 1.05708468e+00 2.89561659e-01
5.12067556e-01 1.64346263e-01 -1.40554011e-01 3.08805197e-01
1.40187532e-01 -8.58322024e-01 -8.26882541e-01 1.23485453e-01
-5.42155206e-01 -1.00102699e+00 -5.79918683e-01 1.27568692e-01
4.82910514e-01 3.25318396e-01 4.86907601e-01 -4.83240068e-01
2.69911677e-01 6.24451697e-01 -3.99242193e-02 -8.04346085e-01
-4.91619051e-01 -5.60647786e-01 7.15708315e-01 3.30525309e-01
-3.06470960e-01 -4.22297150e-01 -6.82990551e-01 3.74027610e-01
-6.70596600e-01 -3.92013818e-01 6.69155419e-01 1.01760197e+00
2.88163304e-01 1.07107110e-01 5.37350535e-01 -1.02842605e+00
1.21021867e+00 -4.49765772e-01 -1.02730286e+00 2.80157477e-01
-8.34366202e-01 2.70682633e-01 9.68378544e-01 -6.41241908e-01
-1.35211170e+00 -1.03991434e-01 -1.06689356e-01 -5.47999084e-01
2.92752624e-01 5.08830667e-01 -1.72239125e-01 -4.58662301e-01
8.50551844e-01 3.54996502e-01 6.10607825e-02 -5.15488582e-03
9.42336679e-01 8.05509567e-01 1.11886680e+00 -6.01266623e-01
1.46632004e+00 6.69120193e-01 1.49543330e-01 -4.26349908e-01
-1.14436495e+00 -2.99222499e-01 -1.93831444e-01 -6.06693253e-02
4.46526051e-01 -1.12531030e+00 -1.18371391e+00 4.38615203e-01
-9.15937960e-01 -1.78460795e-02 -6.78861812e-02 9.38361347e-01
-6.90706253e-01 6.11421883e-01 -5.14498115e-01 -1.59278750e+00
-2.83523619e-01 -1.08823228e+00 7.41312027e-01 1.88683957e-01
1.70412604e-02 -1.27107751e+00 1.62645042e-01 -3.35475691e-02
3.36032987e-01 1.21349968e-01 5.98867238e-02 -6.02326035e-01
-5.39012551e-01 -6.95297241e-01 6.44939765e-02 5.24171293e-01
-3.82761240e-01 -6.13329530e-01 -8.71678531e-01 -6.07833564e-01
7.62457311e-01 2.55649447e-01 4.49748725e-01 4.15722340e-01
5.38273215e-01 -8.39185894e-01 -2.39241093e-01 8.63564432e-01
1.28839433e+00 3.08572590e-01 4.93168294e-01 2.67434686e-01
9.54455361e-02 -8.96984935e-02 9.21831727e-01 6.91497266e-01
2.22590312e-01 2.74364501e-01 3.96962583e-01 3.97135526e-01
6.35756075e-01 -4.90641445e-01 6.97056472e-01 6.77569330e-01
9.28659141e-02 -3.15239489e-01 -2.62810409e-01 7.14632273e-02
-1.71269608e+00 -9.86242712e-01 -1.76623836e-01 2.44095850e+00
5.66082656e-01 1.82265282e-01 -4.06642884e-01 -2.97525786e-02
4.74759340e-01 -3.44157591e-02 -6.30105317e-01 -2.00406581e-01
1.80651009e-01 1.16835721e-01 1.35616338e+00 7.71887600e-01
-1.28517830e+00 5.69234431e-01 6.13207531e+00 1.02888870e+00
-6.44251347e-01 2.39915654e-01 2.40225986e-01 3.28794152e-01
-4.79954621e-03 5.09918034e-01 -8.17307591e-01 3.97352010e-01
1.26429522e+00 -7.59162426e-01 7.57900417e-01 7.77717948e-01
1.44232661e-01 -1.65950641e-01 -7.17930555e-01 9.48399663e-01
-1.84622630e-01 -1.02117240e+00 -4.53215867e-01 3.62774342e-01
4.96992618e-01 -2.83261508e-01 5.88237643e-01 6.09766006e-01
7.33249664e-01 -7.73629427e-01 8.22208405e-01 9.05509055e-01
6.80197537e-01 -9.51072097e-01 6.55955315e-01 8.40186000e-01
-1.06484556e+00 -3.46463978e-01 -1.78479865e-01 -2.70303875e-01
4.67719764e-01 4.50836062e-01 -6.30826652e-01 8.74113560e-01
2.69343704e-01 1.52863994e-01 1.06038287e-01 6.36174679e-01
-4.03691053e-01 6.82639360e-01 -4.67540860e-01 -9.71173774e-03
2.91423261e-01 -4.11828607e-01 8.69940042e-01 1.00572002e+00
4.66402858e-01 5.74678838e-01 5.10819294e-02 6.08963966e-01
3.94333333e-01 -4.00767863e-01 -3.91474813e-01 2.75686029e-02
5.09435773e-01 8.67210150e-01 -2.56416202e-01 -4.03537810e-01
-1.19098790e-01 7.80650973e-01 -2.93389112e-01 4.08856124e-01
-1.24229395e+00 -5.40121138e-01 1.20514345e+00 -4.01996493e-01
-1.48044238e-02 -5.86619198e-01 -9.33277756e-02 -1.22477412e+00
-4.78130877e-01 -9.53822970e-01 3.07897210e-01 -1.89526498e-01
-1.27106667e+00 3.41881812e-01 -1.13402009e-01 -1.44707847e+00
-5.01184940e-01 -5.69474339e-01 -2.83953756e-01 1.22074628e+00
-1.01321256e+00 -8.38694692e-01 3.40903759e-01 9.27179933e-01
3.50828952e-04 -2.76348144e-01 9.09118116e-01 4.92593199e-02
-2.75328636e-01 7.60827661e-01 7.24369407e-01 1.72083855e-01
5.40432751e-01 -1.13779449e+00 5.13566315e-01 1.34337819e+00
3.10593173e-02 1.17764258e+00 1.19691968e+00 -7.25544274e-01
-1.89360094e+00 -1.04187119e+00 3.35992686e-02 -5.77924728e-01
1.07841873e+00 -1.82618648e-01 -5.02680659e-01 6.90706313e-01
-2.72054166e-01 1.74398646e-01 5.90264738e-01 -2.17533171e-01
-4.04353052e-01 9.27087665e-02 -1.27875769e+00 5.12077451e-01
5.98846138e-01 -1.09218919e+00 -4.13500994e-01 3.77386808e-01
4.06512260e-01 -7.41411209e-01 -1.10206068e+00 1.47883460e-01
7.11272180e-01 -8.53383482e-01 1.47607303e+00 -6.27248526e-01
-4.11498964e-01 -6.10243618e-01 -1.80009007e-01 -1.39663565e+00
-1.27165243e-01 -1.30857420e+00 -4.27579463e-01 7.79852629e-01
1.34339735e-01 -1.16561735e+00 5.79167306e-01 5.79682767e-01
7.00515136e-02 -3.83314580e-01 -1.24691057e+00 -1.07098544e+00
8.08394104e-02 -8.66980076e-01 9.92474407e-02 4.18455780e-01
-1.48055032e-01 -8.70037973e-02 -8.72971952e-01 1.08803260e+00
1.05736077e+00 -1.89094543e-01 1.05829859e+00 -6.15992844e-01
-7.37127841e-01 -3.59242503e-03 -1.81566879e-01 -1.54351795e+00
1.61331490e-01 -6.48453355e-01 1.72469914e-01 -7.24206924e-01
-5.11486709e-01 -3.96357477e-01 -2.82820672e-01 -4.26498532e-01
-3.31233472e-01 -3.28822471e-02 2.34199628e-01 -2.12212279e-02
-4.00622457e-01 6.04371071e-01 1.01097274e+00 6.25244528e-02
1.19459797e-02 7.39412427e-01 -7.88083553e-01 9.22231019e-01
4.71073717e-01 -6.06218636e-01 -5.90178549e-01 3.95331532e-02
1.33444704e-02 6.61166906e-01 4.05722380e-01 -1.19288802e+00
5.46126544e-01 -2.70666368e-02 2.83252269e-01 -4.83866483e-01
6.17896020e-01 -8.45645428e-01 4.73851502e-01 9.64949191e-01
-1.57239869e-01 -2.52133191e-01 -1.70217201e-01 1.42121840e+00
-7.68181086e-02 -2.42574796e-01 6.35037303e-01 -8.55569392e-02
-2.65333503e-01 4.11855519e-01 -7.11150229e-01 1.62714913e-01
9.06405509e-01 2.41340622e-01 -3.73510480e-01 -1.08538568e+00
-7.80481219e-01 9.22372099e-03 -1.20523952e-01 4.97608744e-02
6.95734143e-01 -9.81549919e-01 -5.25438905e-01 3.03240150e-01
-4.59286779e-01 -4.89759237e-01 4.70006466e-01 1.10280073e+00
-2.12512076e-01 5.82227111e-01 5.38451850e-01 -4.73326743e-02
-9.72439587e-01 7.73468792e-01 3.58550966e-01 -4.06825930e-01
-7.21310601e-02 1.02100849e+00 4.29951996e-01 -1.88911587e-01
8.47020820e-02 -9.94066820e-02 3.19281459e-01 -4.57561582e-01
7.49743164e-01 5.14352560e-01 -3.48151535e-01 -7.81088710e-01
-2.83198684e-01 2.59127229e-01 4.21207845e-02 -6.18288934e-01
8.46971750e-01 -5.14650166e-01 1.65820882e-01 3.70622538e-02
1.08285642e+00 3.15742821e-01 -1.29777253e+00 -3.45080912e-01
-6.82802871e-02 -7.53065169e-01 7.80364946e-02 -4.43425864e-01
-6.50729239e-01 7.12469220e-01 8.55836928e-01 -1.59020498e-02
1.05472493e+00 -6.89507902e-01 7.03441441e-01 6.53373003e-01
8.07109535e-01 -9.13401783e-01 -4.77629632e-01 8.57068539e-01
6.33876383e-01 -6.57122612e-01 3.07982892e-01 -3.22013348e-01
-4.42467541e-01 1.07227421e+00 -4.00339961e-02 -5.94695926e-01
1.33304167e+00 6.05271041e-01 2.19426244e-01 3.78864884e-01
-5.58874607e-01 -1.89470388e-02 1.80039480e-01 4.87613797e-01
-5.74844517e-03 1.63232595e-01 -2.07267538e-01 8.84335041e-01
-4.38483000e-01 8.50976631e-03 6.84691012e-01 9.03401256e-01
-2.53427386e-01 -8.62769186e-01 -9.12537158e-01 3.58544976e-01
-7.32519507e-01 7.69090205e-02 8.09335113e-02 6.98205173e-01
-3.15729171e-01 1.21406269e+00 -7.24790454e-01 -5.84538579e-01
2.92210013e-01 -4.37096030e-01 2.76400238e-01 -4.17776871e-03
-5.68220377e-01 2.21180603e-01 -1.06006652e-01 -6.13606989e-01
-3.34628741e-03 -6.55677855e-01 -9.05250192e-01 -3.77499849e-01
-9.51879680e-01 5.26559114e-01 8.41139853e-01 9.92801130e-01
6.34786040e-02 3.62608999e-01 6.75337076e-01 -7.23248661e-01
-1.68470585e+00 -8.54728758e-01 -7.75293231e-01 -2.72319429e-02
3.84427160e-01 -6.50256872e-01 -7.43089139e-01 -3.74178469e-01] | [6.3153181076049805, 3.4102296829223633] |
eeef39db-87cb-44cf-a223-8d2303955749 | using-speech-and-nlp-resources-to-build-an | null | null | https://aclanthology.org/2022.computel-1.14 | https://aclanthology.org/2022.computel-1.14.pdf | Using Speech and NLP Resources to build an iCALL platform for a minority language, the story of An Scéalaí, the Irish experience to date | This paper describes how emerging linguistic resources and technologies can be used to build a language learning platform for Irish, an endangered language. This platform, An Scéalaí, harvests learner corpora - a vital resource both to study the stages of learners’ language acquisition and to guide future platform development. A technical description of the platform is provided, including details of how different speech technologies and linguistic resources are fused to provide a holistic learner experience. The active continuous participation of the community, and platform evaluations by learners and teachers, are discussed. | ['Ailbhe Ni Chasaide', 'Harald Berthelsen', 'Neimhin Robinson Gunning', 'Madeleine Comtois', 'Oisín Nolan', 'Neasa Ní Chiaráin'] | null | null | null | null | computel-acl-2022-5 | ['language-acquisition'] | ['natural-language-processing'] | [-5.90546250e-01 4.11473401e-02 -3.19965750e-01 -5.51182330e-02
-1.01642978e+00 -1.05732834e+00 7.30226576e-01 6.77047610e-01
-6.80633843e-01 6.24041080e-01 7.57345855e-01 -6.29289687e-01
-8.14515725e-02 -5.41730642e-01 -2.03322217e-01 -1.49186984e-01
-4.72716196e-03 2.85779744e-01 4.08507109e-01 -1.06660640e+00
3.34097147e-01 4.69709307e-01 -2.10093904e+00 5.63685179e-01
8.35589528e-01 4.92227376e-02 3.64173830e-01 7.84133613e-01
-8.16713810e-01 1.08570385e+00 -7.94882774e-01 -1.57952622e-01
-3.06964725e-01 -1.12574935e-01 -1.20997965e+00 -5.38831234e-01
3.96524191e-01 -2.00136900e-01 1.47030711e-01 7.47949421e-01
1.06437349e+00 3.20903152e-01 8.99670348e-02 -5.96175432e-01
-3.14750254e-01 9.82996047e-01 5.38332999e-01 3.57709706e-01
1.11526823e+00 -1.81819320e-01 4.66811150e-01 -8.72125924e-01
8.50944519e-01 1.18496954e+00 7.49294996e-01 7.02527046e-01
-4.41046983e-01 -6.51735961e-01 1.55270444e-02 1.35170802e-01
-1.10615194e+00 -8.24453175e-01 4.42420155e-01 -5.72614372e-01
1.05122936e+00 4.75154445e-02 1.01261640e+00 9.04238224e-01
-1.63926244e-01 1.05059123e+00 8.73270571e-01 -1.41066563e+00
-2.04080530e-02 7.32063651e-01 1.55260444e-01 5.09897888e-01
-1.40695274e-01 -6.05168156e-02 -1.45642436e+00 4.75633353e-01
8.04255456e-02 -7.12141275e-01 -1.80252120e-01 7.75206462e-03
-9.30827796e-01 5.82870662e-01 -4.20503706e-01 9.93644774e-01
-1.32205000e-03 -3.90104085e-01 7.76283801e-01 6.37285113e-01
3.60209435e-01 7.16548041e-02 -3.39396387e-01 -1.05319202e+00
-8.08016121e-01 1.92248970e-01 1.05884838e+00 1.02348053e+00
1.07290514e-01 1.75437912e-01 3.08676451e-01 1.11840880e+00
8.41300130e-01 2.57648379e-01 6.20037675e-01 -6.90907419e-01
3.48380893e-01 6.19911909e-01 -3.32257360e-01 -4.96846251e-02
-3.39570791e-01 -1.88116699e-01 4.43535477e-01 3.77243519e-01
4.35668319e-01 -4.41472918e-01 -2.52660692e-01 1.13782191e+00
5.03033638e-01 -2.58740276e-01 4.57203329e-01 1.81773856e-01
1.65087891e+00 4.13811088e-01 5.43670595e-01 6.94833696e-02
1.33647406e+00 -6.51597738e-01 -7.50856698e-01 1.17115021e-01
1.31167591e+00 -1.40014422e+00 1.23789048e+00 4.43825662e-01
-1.54863286e+00 -3.72007191e-01 -9.63166237e-01 -1.61825508e-01
-9.94769275e-01 -3.71610820e-01 3.20009828e-01 1.33037710e+00
-1.46821690e+00 1.82798386e-01 -6.17633522e-01 -7.37310171e-01
3.09283612e-04 1.06395319e-01 -3.41270089e-01 -1.99401289e-01
-1.23792195e+00 9.87231672e-01 7.07887828e-01 -4.44649160e-01
-6.92440927e-01 -1.03687382e+00 -1.01386809e+00 -4.11076188e-01
-4.95764092e-02 1.45673439e-01 1.68722510e+00 -5.69618940e-01
-1.91986573e+00 1.41822457e+00 2.57009804e-01 8.21946003e-03
5.80694139e-01 -4.83941108e-01 -7.08506942e-01 7.24243671e-02
1.74000740e-01 2.38508850e-01 -3.75863224e-01 -9.19306517e-01
-1.20992458e+00 -4.95975539e-02 -1.04876205e-01 6.90568745e-01
-6.39157355e-01 9.21865225e-01 -4.26575601e-01 -4.05175477e-01
-4.51163590e-01 -2.83506155e-01 3.41297746e-01 -5.06055474e-01
4.53576744e-01 -4.28192914e-01 8.41364801e-01 -9.21364963e-01
1.35733020e+00 -2.11558700e+00 -5.09927273e-01 6.43362999e-02
-2.09351316e-01 8.78090501e-01 1.11085422e-01 1.29403770e+00
5.29594958e-01 2.87377000e-01 5.62879920e-01 -7.87418522e-03
1.24233201e-01 1.77210853e-01 1.13465644e-01 1.25860721e-01
-3.01315516e-01 2.78980225e-01 -1.41348767e+00 -4.50003952e-01
5.87595940e-01 1.05800128e+00 -1.69060200e-01 9.61171612e-02
1.73056409e-01 5.07432997e-01 -3.56918901e-01 5.42692423e-01
2.13022754e-01 8.27685356e-01 6.60786182e-02 8.61238360e-01
-1.29218221e+00 7.82523394e-01 -1.20106983e+00 2.04363656e+00
-1.01851380e+00 9.26995456e-01 5.23050189e-01 -4.43804085e-01
9.27429676e-01 9.43137825e-01 1.52938455e-01 -8.64861250e-01
9.72302631e-02 7.80876577e-01 -2.14230821e-01 -8.92303884e-01
5.73218405e-01 -1.65104032e-01 -7.87210651e-03 4.64477122e-01
5.26321352e-01 -3.21251750e-02 3.92854959e-01 1.58231974e-01
6.72671676e-01 5.17834306e-01 2.38999337e-01 -9.06923652e-01
1.00958896e+00 1.54104039e-01 6.37392178e-02 2.72401184e-01
-5.23550928e-01 -1.85830474e-01 -1.05377406e-01 -1.41069487e-01
-6.24420047e-01 -9.37453806e-01 -4.11052644e-01 1.85818291e+00
-4.92154896e-01 -4.97273266e-01 -9.60122943e-01 -3.13961387e-01
-3.51144016e-01 8.41778398e-01 1.22251838e-01 2.23974213e-01
-4.96531665e-01 2.16286644e-01 8.51744592e-01 3.92517984e-01
2.00802580e-01 -1.58287776e+00 -5.97606301e-01 6.11328900e-01
-1.70155093e-01 -8.77690136e-01 -2.35019654e-01 7.64147788e-02
-2.28952393e-01 -6.63021684e-01 -6.07821524e-01 -1.66693091e+00
6.75246119e-02 6.57979250e-02 1.07576823e+00 4.98002321e-01
-2.40832001e-01 1.04480672e+00 -7.47180521e-01 -9.71556365e-01
-1.04533696e+00 3.44566815e-02 1.22206032e-01 -9.87954557e-01
6.85182631e-01 -3.57025921e-01 3.29732685e-03 -3.38784933e-01
-6.70112073e-01 -2.12271810e-01 7.89490435e-03 3.32806468e-01
1.76382989e-01 -4.44448553e-02 9.37245011e-01 -7.40943849e-01
8.16921949e-01 -2.67985255e-01 -2.82242924e-01 4.46180522e-01
-3.12238008e-01 -2.97595352e-01 4.47909534e-01 -5.69045842e-02
-1.44997835e+00 -3.18586409e-01 -8.75465691e-01 5.80764592e-01
-5.46066999e-01 7.86251426e-01 -4.87039834e-01 -4.27512258e-01
6.14430487e-01 4.26563881e-02 -1.41105862e-04 -7.34572649e-01
3.08340788e-01 1.21962202e+00 7.76557386e-01 -9.16136026e-01
2.36368462e-01 -8.82462859e-02 -6.71936870e-01 -1.51459873e+00
-3.47138017e-01 -1.02348983e+00 -1.22951901e+00 -8.78191590e-01
5.81523836e-01 -1.33668935e+00 -5.09125113e-01 4.84323800e-01
-8.17702711e-01 -7.59098828e-01 -5.07974088e-01 8.17869127e-01
-3.21903050e-01 -1.85900599e-01 -5.94570518e-01 -9.62436199e-01
-1.73666567e-01 -9.86632884e-01 6.13120615e-01 6.82872653e-01
-1.90826207e-01 -1.88257492e+00 4.12755579e-01 6.59047008e-01
3.00017208e-01 1.34105265e-01 6.36732996e-01 -9.42835331e-01
2.08448902e-01 -6.89692050e-02 5.63176930e-01 3.85996461e-01
-1.03670761e-01 2.54699647e-01 -1.01284182e+00 -3.08297396e-01
-3.95941854e-01 -6.95928931e-01 -2.01746821e-01 -3.53875935e-01
2.97100563e-02 -3.93826477e-02 1.41715512e-01 7.53336474e-02
1.44270015e+00 3.05019408e-01 5.95970862e-02 9.85046506e-01
3.04318160e-01 9.94850039e-01 4.43319559e-01 1.89194337e-01
6.29006684e-01 3.05791557e-01 -3.38503778e-01 3.79053026e-01
-8.16368818e-01 -5.12209415e-01 8.93789411e-01 2.05131674e+00
1.59458183e-02 3.21405344e-02 -1.74867022e+00 9.63371158e-01
-1.31549013e+00 -9.56386566e-01 -3.08221638e-01 2.10928559e+00
1.01810765e+00 -4.48825844e-02 4.11924988e-01 3.88084501e-01
4.77802962e-01 -1.98221296e-01 3.80286962e-01 -1.07319069e+00
-3.05526480e-02 7.10854530e-01 2.01844007e-01 9.90824759e-01
-4.85353768e-01 1.30992901e+00 7.26450586e+00 7.35952437e-01
-1.08418512e+00 2.73517936e-01 -3.83566052e-01 9.27407891e-02
-4.73944962e-01 -2.47552201e-01 -1.28531194e+00 7.67434016e-02
1.50804532e+00 -7.24998176e-01 2.67627269e-01 5.67376435e-01
4.84598070e-01 1.76530287e-01 -5.15975118e-01 2.21555606e-01
2.22400472e-01 -1.46351802e+00 -1.40800908e-01 -1.30584329e-01
4.78055149e-01 5.08196354e-01 -2.18530655e-01 6.10401511e-01
7.60107517e-01 -8.42986763e-01 1.25119638e+00 2.25353241e-01
7.83687949e-01 -1.22190630e+00 5.11067271e-01 5.65797329e-01
-1.37220752e+00 1.47905588e-01 1.23735361e-01 -2.80120939e-01
6.58068061e-02 -4.05424386e-01 -7.45757520e-01 6.36158049e-01
9.64334965e-01 4.04939264e-01 -5.16839921e-01 1.23029935e+00
-4.07286674e-01 8.08491230e-01 -1.55738205e-01 -2.94976056e-01
3.66553485e-01 5.77125810e-02 5.60134590e-01 1.68717766e+00
3.55216771e-01 -3.06474902e-02 5.34552753e-01 -2.32605740e-01
3.20490867e-01 1.13067293e+00 -8.42045128e-01 -1.78240817e-02
9.57581937e-01 1.00945735e+00 -4.92994219e-01 7.00946599e-02
-7.84221053e-01 3.13777298e-01 2.37781495e-01 -6.91191927e-02
1.57881588e-01 -6.58438385e-01 5.45632005e-01 3.28695059e-01
-2.03751966e-01 -3.55477005e-01 -2.91186813e-02 -5.02223194e-01
-3.19000930e-01 -1.25980175e+00 4.52675581e-01 -2.60566086e-01
-6.32186532e-01 3.10044616e-01 1.39889315e-01 -1.05987561e+00
-4.55007643e-01 -6.14094198e-01 -9.25886273e-01 9.78074789e-01
-1.43239474e+00 -1.41916227e+00 2.86413487e-02 5.22350729e-01
5.85979819e-01 -6.66335344e-01 1.20473444e+00 6.78577840e-01
-4.85700220e-01 7.24342346e-01 3.41709554e-01 4.29368727e-02
4.31794494e-01 -1.24661314e+00 -1.73747480e-01 7.45730042e-01
2.30864272e-01 4.54519182e-01 6.07766032e-01 -4.36722904e-01
-1.29132223e+00 -7.46603787e-01 1.48175490e+00 -3.01942647e-01
9.14804518e-01 -2.21518740e-01 -8.35650861e-01 4.99086082e-01
9.67376828e-01 -5.27465761e-01 1.52609348e+00 1.03558451e-01
1.45440042e-01 2.81484365e-01 -9.53701615e-01 6.14553392e-01
1.07041621e+00 -8.93971682e-01 -8.12901318e-01 3.21152627e-01
6.02841735e-01 -5.18407762e-01 -1.39232445e+00 -1.50397390e-01
7.12192416e-01 -8.66092741e-01 7.43670642e-01 -4.27822292e-01
-2.80468166e-01 3.65751013e-02 1.33986086e-01 -1.20951176e+00
2.76458263e-01 -1.20213675e+00 3.05403262e-01 2.01808405e+00
3.99736464e-01 -4.02773142e-01 4.51544046e-01 2.80979216e-01
-8.43887508e-01 -1.39164522e-01 -1.00468600e+00 -7.10369885e-01
9.75548208e-01 -9.32423949e-01 4.18365330e-01 1.02793694e+00
7.36265421e-01 2.55519360e-01 4.05376524e-01 -8.37991387e-02
3.10130805e-01 -1.05508149e+00 6.04399502e-01 -1.19616985e+00
2.87895262e-01 -9.92688119e-01 -4.51285481e-01 -2.25776821e-01
1.74596399e-01 -1.10888493e+00 8.65318328e-02 -1.73551428e+00
-7.17523277e-01 -3.52697194e-01 -5.81634641e-02 3.76796514e-01
2.05710426e-01 1.88254137e-02 1.96672469e-01 -1.37042388e-01
-4.29966390e-01 -5.97883128e-02 1.01999640e+00 4.71215546e-01
-6.01438582e-01 -1.27654910e-01 -7.86703765e-01 9.23038423e-01
7.83905029e-01 -1.59351140e-01 -5.41144371e-01 -6.75514758e-01
3.03516954e-01 -3.56432587e-01 -3.57290447e-01 -1.37099445e+00
4.64011580e-01 -9.36129168e-02 -1.09141193e-01 -5.49477220e-01
-2.17620894e-01 -4.69348073e-01 -2.90623069e-01 4.75183725e-01
-3.05712879e-01 1.31094277e-01 8.08093131e-01 -3.41875106e-01
-4.18134600e-01 -7.70275176e-01 6.43482387e-01 -1.01932198e-01
-1.08759236e+00 -3.43932897e-01 -9.91590440e-01 5.23925960e-01
1.28799915e+00 -2.68839777e-01 -1.79930758e-02 -2.31885493e-01
-7.50254571e-01 6.51163697e-01 3.02390844e-01 5.11312246e-01
4.03299302e-01 -1.29708445e+00 -9.53419745e-01 3.86536568e-01
2.40927160e-01 -1.42240310e-02 2.18461499e-01 2.61926353e-01
-1.21519899e+00 4.56311256e-01 -4.58569378e-01 -1.77834824e-01
-1.53509438e+00 -1.59591079e-01 2.03172147e-01 -1.48368806e-01
-9.40399706e-01 1.05268049e+00 -7.51477599e-01 -9.88076687e-01
7.24894881e-01 3.29297543e-01 -9.77905273e-01 5.22704065e-01
1.12944877e+00 4.24297243e-01 2.15038598e-01 -1.09858024e+00
-3.37968647e-01 2.73787737e-01 6.10782467e-02 -6.78185761e-01
1.31584227e+00 -3.59822512e-01 2.37074554e-01 1.07117236e+00
9.22694266e-01 7.26632535e-01 -5.90061724e-01 -2.44780570e-01
4.58744526e-01 -1.33233696e-01 3.16426426e-01 -9.88465130e-01
-3.14486563e-01 1.02765679e+00 6.87216997e-01 1.84532627e-01
6.75586402e-01 -1.33456469e-01 5.92130840e-01 6.08535595e-02
1.77527070e-01 -1.86606574e+00 -2.96014369e-01 9.27050591e-01
8.28416288e-01 -7.21546948e-01 -2.45961413e-01 -2.05649048e-01
-3.59406978e-01 1.28843701e+00 4.76257801e-01 4.25655723e-01
1.03602123e+00 5.21461904e-01 9.76368010e-01 7.43801966e-02
-5.85171223e-01 -5.76508939e-01 2.25985646e-01 1.17419410e+00
1.31876719e+00 -3.34999375e-02 -7.00601876e-01 5.98176599e-01
-8.21643472e-01 2.78769638e-02 5.61710894e-01 1.54248905e+00
-6.74659312e-01 -1.80413878e+00 -5.59474111e-01 -3.56983751e-01
-8.35585296e-01 -2.30930820e-01 -4.62478995e-01 1.15348065e+00
5.05933583e-01 9.99363840e-01 -7.87609294e-02 -1.03373259e-01
7.16580331e-01 6.20711923e-01 3.70892197e-01 -1.07399440e+00
-1.28805947e+00 1.00574009e-01 5.43375909e-01 9.39681157e-02
-4.67549026e-01 -1.05563712e+00 -1.48382878e+00 -5.50035715e-01
-8.94466117e-02 5.29147863e-01 1.08523428e+00 8.23181272e-01
-1.92356139e-01 5.77485204e-01 1.93781078e-01 -6.14100575e-01
-1.85070246e-01 -7.22030997e-01 -4.97400641e-01 -2.78509200e-01
-5.07524051e-02 -2.69871473e-01 -1.67549804e-01 4.51605059e-02] | [10.694640159606934, 10.180119514465332] |
e7c0fffd-461b-48f3-8c09-035185b7ac2f | aspanformer-detector-free-image-matching-with | 2208.14201 | null | https://arxiv.org/abs/2208.14201v1 | https://arxiv.org/pdf/2208.14201v1.pdf | ASpanFormer: Detector-Free Image Matching with Adaptive Span Transformer | Generating robust and reliable correspondences across images is a fundamental task for a diversity of applications. To capture context at both global and local granularity, we propose ASpanFormer, a Transformer-based detector-free matcher that is built on hierarchical attention structure, adopting a novel attention operation which is capable of adjusting attention span in a self-adaptive manner. To achieve this goal, first, flow maps are regressed in each cross attention phase to locate the center of search region. Next, a sampling grid is generated around the center, whose size, instead of being empirically configured as fixed, is adaptively computed from a pixel uncertainty estimated along with the flow map. Finally, attention is computed across two images within derived regions, referred to as attention span. By these means, we are able to not only maintain long-range dependencies, but also enable fine-grained attention among pixels of high relevance that compensates essential locality and piece-wise smoothness in matching tasks. State-of-the-art accuracy on a wide range of evaluation benchmarks validates the strong matching capability of our method. | ['Long Quan', 'Yanghai Tsin', 'David McKinnon', 'Tian Fang', 'Mingmin Zhen', 'Yurun Tian', 'Lei Zhou', 'Zixin Luo', 'Hongkai Chen'] | 2022-08-30 | null | null | null | null | ['homography-estimation'] | ['computer-vision'] | [ 1.11052975e-01 -2.70286590e-01 -1.31611601e-01 -3.01903754e-01
-8.11315298e-01 -3.17998260e-01 7.25185633e-01 2.70935714e-01
-5.60224593e-01 4.91877466e-01 2.58434564e-01 1.04120411e-01
-2.38901809e-01 -8.73034179e-01 -7.89961100e-01 -6.41169012e-01
1.01412356e-01 2.68445224e-01 5.52611113e-01 -2.14973971e-01
5.59431314e-01 7.94831455e-01 -1.66102278e+00 7.23506361e-02
1.13547742e+00 1.13887715e+00 4.09172535e-01 2.81913817e-01
-1.13603935e-01 4.52021748e-01 -3.54945749e-01 -4.38483268e-01
4.18096423e-01 -3.15226674e-01 -6.87320411e-01 1.89178929e-01
7.82419622e-01 -2.46501133e-01 -2.89968759e-01 1.18299520e+00
3.71802002e-01 3.86309206e-01 4.44178104e-01 -7.67607808e-01
-7.40607798e-01 4.66219574e-01 -7.31440604e-01 5.76906264e-01
1.32423684e-01 4.92435008e-01 1.36259115e+00 -1.10435784e+00
4.50199664e-01 1.16348028e+00 3.66740227e-01 1.33451223e-01
-1.35195279e+00 -5.18444777e-01 3.58310580e-01 2.73611516e-01
-1.50328135e+00 -4.10529941e-01 9.17358398e-01 -3.80730659e-01
8.64345431e-01 1.84547082e-01 7.12909341e-01 6.38517976e-01
4.47633177e-01 4.49130058e-01 7.32792914e-01 -4.18607175e-01
2.05302477e-01 -1.82994008e-01 -1.26066029e-01 5.96878469e-01
-1.60590652e-02 1.11060560e-01 -6.26141250e-01 1.28088012e-01
1.12387323e+00 -5.29039092e-02 -4.98337835e-01 -5.17676651e-01
-1.50219893e+00 6.63044214e-01 1.07798517e+00 4.14789319e-01
-7.53801942e-01 2.45586917e-01 2.63199300e-01 2.84765232e-02
2.65752941e-01 5.09203136e-01 -7.75491372e-02 1.46409705e-01
-9.45759058e-01 9.24564153e-02 1.80010259e-01 7.78919756e-01
1.09223306e+00 -1.73478007e-01 -7.41211474e-01 6.80709541e-01
-1.51663870e-02 1.39861673e-01 5.14932871e-01 -8.27865362e-01
5.22911370e-01 8.20244491e-01 1.82876274e-01 -1.20910108e+00
-1.49608925e-01 -8.61495912e-01 -1.07271099e+00 3.35844278e-01
3.78192157e-01 3.57421607e-01 -8.55857849e-01 1.73685801e+00
5.25147438e-01 3.55414003e-01 -4.29154813e-01 1.24740672e+00
2.98738658e-01 3.89881790e-01 2.30113804e-01 -3.54144499e-02
1.41250384e+00 -1.14704680e+00 -4.46965426e-01 -1.10823780e-01
5.45885935e-02 -5.36077857e-01 1.13733220e+00 1.18375637e-01
-1.30550253e+00 -9.77846503e-01 -9.95185375e-01 -3.26242208e-01
-3.09905350e-01 -4.95844036e-02 3.29658270e-01 1.18857734e-01
-1.07856083e+00 6.11984968e-01 -6.63518488e-01 -1.12491027e-01
4.21383023e-01 2.02762261e-01 -4.25504416e-01 1.20442644e-01
-1.16412508e+00 7.48724401e-01 2.29842409e-01 2.70069748e-01
-5.69539189e-01 -8.26508284e-01 -9.26555932e-01 4.62846041e-01
1.69003576e-01 -9.75271404e-01 7.94299781e-01 -9.74420130e-01
-1.38134766e+00 8.79090905e-01 -3.38290393e-01 -5.90271711e-01
6.84505284e-01 -1.49991721e-01 1.72572117e-02 1.97036952e-01
4.40186381e-01 9.00364995e-01 1.18665504e+00 -1.04395652e+00
-8.26410115e-01 -3.61187935e-01 8.97781365e-03 3.04706424e-01
-4.05411959e-01 -9.87785235e-02 -9.05839860e-01 -8.48307729e-01
2.40416795e-01 -5.12075961e-01 -4.08560038e-01 2.35686839e-01
-2.60912776e-01 -2.02174932e-01 4.19305116e-01 -4.29524332e-01
1.49796867e+00 -2.04474974e+00 2.99982995e-01 4.15421218e-01
3.06660354e-01 1.37773648e-01 -1.71229586e-01 1.28018409e-01
2.42140307e-03 -1.16026446e-01 -3.72437507e-01 -3.85646790e-01
-9.89471897e-02 -1.23093858e-01 -3.11914772e-01 4.46083844e-01
6.26201391e-01 1.14655817e+00 -9.83374894e-01 -5.08109868e-01
6.69110060e-01 4.10925657e-01 -6.63096011e-01 3.05505693e-01
-1.27560869e-01 6.30678833e-01 -6.01996183e-01 5.01709878e-01
6.50393546e-01 -4.56953704e-01 -2.04824910e-01 -3.53675008e-01
-3.69592160e-01 1.48968017e-02 -1.20639145e+00 1.96627915e+00
-6.90248132e-01 4.95301127e-01 -4.03334424e-02 -9.58677828e-01
1.17462397e+00 -1.59964189e-01 5.80375910e-01 -1.02370751e+00
7.73106962e-02 5.52615523e-02 7.09366146e-03 -2.15247482e-01
6.69309676e-01 4.33446705e-01 1.30670264e-01 2.38068119e-01
-1.92797899e-01 2.31595308e-01 1.78292900e-01 -1.18248522e-01
9.32624698e-01 1.45963103e-01 4.21682805e-01 -4.96817499e-01
1.06233311e+00 -1.41540661e-01 6.05960250e-01 7.86384165e-01
-2.00813994e-01 8.11672509e-01 1.58897817e-01 -6.90896392e-01
-9.98824120e-01 -9.69997048e-01 -3.04747045e-01 1.08822203e+00
5.90140104e-01 -2.82900065e-01 -6.54191136e-01 -5.73818505e-01
2.67372578e-01 2.27257490e-01 -9.13490057e-01 -1.00771181e-01
-8.13713491e-01 -3.06464016e-01 -5.05986884e-02 5.75062454e-01
7.12330937e-01 -1.28302562e+00 -1.05153251e+00 5.76805472e-01
-3.89652960e-02 -1.01487005e+00 -9.06629741e-01 -2.60556154e-02
-6.46628022e-01 -9.75920379e-01 -8.63499820e-01 -6.45300090e-01
8.25304210e-01 3.59636396e-01 1.38095713e+00 2.36409470e-01
-2.95479655e-01 -4.21006307e-02 -1.82779044e-01 2.04400241e-01
-3.01598795e-02 1.66225344e-01 -4.20953572e-01 5.81250370e-01
1.92992762e-01 -5.76600671e-01 -1.04000056e+00 5.56849062e-01
-7.84597933e-01 4.01539216e-03 7.84783304e-01 1.07163608e+00
8.60518992e-01 -3.35672706e-01 4.85073864e-01 -4.65978742e-01
4.98185694e-01 -2.91035235e-01 -9.40717936e-01 3.68088216e-01
-3.75737250e-01 2.69867599e-01 6.77001297e-01 -3.39912236e-01
-8.65490079e-01 1.19938001e-01 -1.05111584e-01 -6.54511034e-01
-7.14186430e-02 1.02939069e-01 -1.06127135e-01 -1.80915773e-01
6.18649244e-01 2.25703806e-01 -1.30125225e-01 -1.74299747e-01
5.13544977e-01 4.14830416e-01 8.78805578e-01 -5.63221097e-01
7.07999408e-01 4.92676258e-01 -4.65616994e-02 -4.43040669e-01
-5.77484548e-01 -5.11457860e-01 -7.90926158e-01 -2.80035466e-01
9.04586136e-01 -6.53338373e-01 -7.25976110e-01 3.65723729e-01
-1.05061555e+00 -3.70302975e-01 -3.31753165e-01 3.35746288e-01
-4.82566774e-01 1.59630999e-01 -4.17506933e-01 -5.45276344e-01
-5.84591329e-01 -1.29767072e+00 1.26114452e+00 3.08959216e-01
-1.03346288e-01 -8.66333127e-01 -7.20398128e-02 1.03909839e-02
6.49682820e-01 2.93313026e-01 6.73150718e-01 -2.55277097e-01
-1.12324190e+00 -2.83632241e-02 -5.49396396e-01 1.64021784e-03
2.23069444e-01 -4.53636348e-02 -8.52070570e-01 -2.74306566e-01
-2.97213107e-01 8.25609863e-02 9.72629607e-01 4.91621733e-01
1.32382095e+00 2.17104927e-02 -3.72078925e-01 7.89914906e-01
1.43917096e+00 -8.70662183e-02 7.26773858e-01 4.67659801e-01
5.74680150e-01 5.18513858e-01 6.84867144e-01 3.83977056e-01
3.64069253e-01 1.06069493e+00 5.74065030e-01 -2.49105856e-01
-3.67975414e-01 -2.25872651e-01 -2.71839947e-01 1.02485314e-01
7.31596053e-02 -2.33937930e-02 -7.06972063e-01 6.05192006e-01
-1.99198258e+00 -9.03734982e-01 1.08923197e-01 2.54949737e+00
7.66934812e-01 2.88252294e-01 -4.86491844e-02 -1.06632642e-01
9.31462586e-01 2.79483497e-01 -5.70894897e-01 -6.08632937e-02
1.46224692e-01 2.68150061e-01 3.25433642e-01 7.42868006e-01
-9.79556739e-01 1.04820335e+00 5.25153017e+00 7.09719241e-01
-1.27547491e+00 -1.72374666e-01 7.02674508e-01 1.10436305e-01
-3.62842172e-01 -1.71920210e-02 -6.97471082e-01 6.70310855e-01
3.17829698e-01 -6.75630793e-02 3.48389357e-01 5.01783371e-01
2.87523031e-01 -2.82263644e-02 -1.00938511e+00 9.48303759e-01
-7.45920390e-02 -1.57657504e+00 6.00947179e-02 -2.69241836e-02
7.74611175e-01 -1.55495256e-01 -1.16476463e-02 4.88526635e-02
-1.48338089e-02 -7.91523933e-01 7.76567578e-01 6.00124478e-01
7.85527647e-01 -7.96034098e-01 4.89474684e-01 2.78525382e-01
-1.58282781e+00 -1.62363783e-01 -3.10572714e-01 1.78335786e-01
2.19197661e-01 6.80844784e-01 -3.24097037e-01 4.90561813e-01
7.21814334e-01 6.99058592e-01 -4.73793656e-01 1.25003219e+00
-2.68288285e-01 -1.39389187e-01 -2.10026518e-01 2.37697944e-01
2.98029393e-01 -2.72230685e-01 4.99030232e-01 1.35390377e+00
2.70217031e-01 -1.12908177e-01 3.68390605e-02 1.14451218e+00
-7.67488927e-02 7.72751644e-02 -3.10040593e-01 8.02773595e-01
7.24427521e-01 1.31987464e+00 -5.85793078e-01 -3.29497039e-01
-3.01517099e-01 9.47145998e-01 7.35801458e-01 2.58321524e-01
-9.69152212e-01 -4.42148030e-01 7.05460906e-01 2.07015499e-01
5.95684171e-01 -1.02496438e-01 -3.66683215e-01 -1.03600907e+00
2.68480927e-01 -5.57081223e-01 4.03448224e-01 -5.63995421e-01
-1.06395769e+00 8.98578584e-01 -3.12323779e-01 -1.42567623e+00
-2.48669580e-01 -5.30518554e-02 -8.86342585e-01 1.20746469e+00
-1.87734175e+00 -1.10327768e+00 -8.78896236e-01 6.47295475e-01
6.90079451e-01 1.37694195e-01 5.21186531e-01 3.59827787e-01
-5.71676314e-01 8.28074872e-01 -3.98013949e-01 8.47595558e-03
7.71627963e-01 -1.11651564e+00 7.66665280e-01 1.09854436e+00
1.22394003e-01 5.74493706e-01 3.23050231e-01 -3.72986376e-01
-1.05538177e+00 -1.28708792e+00 9.51720893e-01 -2.04145133e-01
5.50809205e-01 -2.52264142e-01 -1.33635342e+00 3.44177932e-01
1.49789751e-01 5.58789730e-01 -5.27471006e-02 -2.70195067e-01
-4.42577153e-01 -4.49500293e-01 -1.02770913e+00 5.31571329e-01
1.10025680e+00 -4.00685430e-01 -4.43056852e-01 -1.10659543e-02
5.07413864e-01 -6.74170196e-01 -9.57116783e-01 3.36497515e-01
4.69713181e-01 -1.36016262e+00 1.11038077e+00 -2.48138115e-01
3.90264809e-01 -5.19769609e-01 1.97407052e-01 -1.06288171e+00
-8.18040013e-01 -6.89717650e-01 -5.47806025e-02 1.26303053e+00
1.71804905e-01 -7.00532258e-01 6.96506441e-01 5.45538902e-01
-2.73373008e-01 -9.11051989e-01 -8.80981088e-01 -5.17434418e-01
-2.15353489e-01 -2.96637211e-02 8.07015359e-01 6.87667131e-01
-3.12736869e-01 1.03597276e-01 -4.66217808e-02 3.11789691e-01
8.11256766e-01 4.19115782e-01 7.40311205e-01 -1.00938487e+00
-1.93695530e-01 -9.43865180e-01 -4.90625918e-01 -1.39968646e+00
1.76022321e-01 -5.83156109e-01 7.64279589e-02 -1.38481855e+00
-5.25604896e-02 -7.23804414e-01 -5.11667430e-01 2.63266057e-01
-5.52932441e-01 3.61323565e-01 1.32197306e-01 4.45941389e-01
-4.10081565e-01 6.00539863e-01 1.31817567e+00 -1.51988324e-02
-2.97159880e-01 -1.14502691e-01 -5.82047343e-01 3.02057773e-01
5.62408090e-01 -7.18227029e-02 -2.26661786e-01 -5.48636913e-01
-3.96494597e-01 3.06901429e-02 5.53788066e-01 -1.08764577e+00
5.04119337e-01 -2.19014660e-01 4.08979595e-01 -5.10834515e-01
1.04374394e-01 -8.01811695e-01 7.15149641e-02 3.78714114e-01
-5.33191323e-01 2.46722102e-01 -5.34857158e-03 5.54136813e-01
-3.95719558e-01 1.69164538e-01 1.02356517e+00 1.37324065e-01
-1.08968318e+00 7.27086484e-01 3.24393302e-01 -1.85107086e-02
1.18435264e+00 -3.10975105e-01 -3.13666835e-02 -3.87287326e-02
-3.40865105e-01 4.13529843e-01 5.97495794e-01 5.74941933e-01
5.28697014e-01 -1.49658668e+00 -7.52443910e-01 5.73671818e-01
4.30639416e-01 2.28200972e-01 4.09508079e-01 7.95882940e-01
-3.25863868e-01 3.05036098e-01 -3.05610508e-01 -8.28981638e-01
-7.29792535e-01 7.53382742e-01 5.32956004e-01 -3.45785469e-01
-8.66574466e-01 7.97324777e-01 3.73591304e-01 2.76017070e-01
2.89347559e-01 -3.87106508e-01 -2.75595576e-01 -9.92308483e-02
7.16445267e-01 4.93815131e-02 2.27153361e-01 -6.21985078e-01
-3.39763612e-01 1.01354730e+00 6.65651709e-02 1.43103108e-01
9.01898980e-01 -3.71542484e-01 2.29683891e-02 -2.10159406e-01
1.12075913e+00 -9.84191429e-03 -1.86160088e+00 -4.24116999e-01
-8.24372098e-02 -9.15684581e-01 6.21062852e-02 -3.92626137e-01
-1.27852023e+00 7.51807272e-01 4.22160983e-01 1.07289970e-01
1.31035340e+00 -1.31514013e-01 6.85428381e-01 -3.13552171e-01
2.55221397e-01 -8.35364103e-01 6.61208630e-02 1.76802084e-01
1.07165682e+00 -1.35022950e+00 -2.84790814e-01 -2.04883307e-01
-3.98442447e-01 1.03199542e+00 8.49576175e-01 -2.91480601e-01
2.76243359e-01 2.06616610e-01 -1.58804372e-01 8.59027822e-03
-6.11365378e-01 -4.59556103e-01 5.59707105e-01 4.91895348e-01
3.70394379e-01 -3.15536082e-01 -1.81312785e-02 2.93954667e-02
1.51658639e-01 -2.46370643e-01 -1.34337157e-01 4.89574075e-01
-4.84033555e-01 -8.48191440e-01 -3.18463117e-01 1.19615346e-01
-8.12001526e-02 -9.24630761e-02 9.93874520e-02 6.98459327e-01
5.92279956e-02 5.54727137e-01 5.80878198e-01 -8.68720859e-02
5.69633245e-01 -5.68884134e-01 1.88793778e-01 -2.85979867e-01
-8.02613437e-01 -4.37085554e-02 -4.08321351e-01 -9.19880271e-01
-1.82499737e-01 -3.43012899e-01 -1.16160882e+00 -9.32590216e-02
-1.79906219e-01 3.40782553e-02 2.82075018e-01 7.25574434e-01
5.96278787e-01 5.74221969e-01 7.69888759e-01 -1.07477129e+00
-5.70783734e-01 -5.83628774e-01 -1.53727055e-01 6.97734296e-01
5.03298819e-01 -6.90627754e-01 -1.55938357e-01 -1.78821564e-01] | [8.826288223266602, -1.9917460680007935] |
e6842329-8d4a-4713-8158-9ad5c2a293b0 | on-enhancing-speech-emotion-recognition-using | 1806.06626 | null | http://arxiv.org/abs/1806.06626v1 | http://arxiv.org/pdf/1806.06626v1.pdf | On Enhancing Speech Emotion Recognition using Generative Adversarial Networks | Generative Adversarial Networks (GANs) have gained a lot of attention from
machine learning community due to their ability to learn and mimic an input
data distribution. GANs consist of a discriminator and a generator working in
tandem playing a min-max game to learn a target underlying data distribution;
when fed with data-points sampled from a simpler distribution (like uniform or
Gaussian distribution). Once trained, they allow synthetic generation of
examples sampled from the target distribution. We investigate the application
of GANs to generate synthetic feature vectors used for speech emotion
recognition. Specifically, we investigate two set ups: (i) a vanilla GAN that
learns the distribution of a lower dimensional representation of the actual
higher dimensional feature vector and, (ii) a conditional GAN that learns the
distribution of the higher dimensional feature vectors conditioned on the
labels or the emotional class to which it belongs. As a potential practical
application of these synthetically generated samples, we measure any
improvement in a classifier's performance when the synthetic data is used along
with real data for training. We perform cross-validation analyses followed by a
cross-corpus study. | ['Carol Espy-Wilson', 'Rahul Gupta', 'Saurabh Sahu'] | 2018-06-18 | null | null | null | null | ['cross-corpus'] | ['computer-vision'] | [ 5.77644289e-01 5.34134448e-01 1.18398212e-01 -5.31459391e-01
-1.03295112e+00 -6.33485258e-01 8.32974970e-01 -3.01444024e-01
-2.03461379e-01 9.29129481e-01 2.51129627e-01 -1.07515668e-02
5.70806563e-01 -9.12464738e-01 -7.21464336e-01 -1.06203723e+00
2.17031226e-01 7.22125590e-01 -5.69990635e-01 2.43783649e-02
-3.22364867e-01 3.59866709e-01 -1.46184957e+00 4.51748580e-01
7.63758957e-01 9.55301702e-01 -1.82216257e-01 8.80014002e-01
2.94420458e-02 7.10665405e-01 -1.38747489e+00 -5.20252824e-01
2.92967319e-01 -1.03125083e+00 -3.84082049e-01 1.33936241e-01
-1.23745121e-01 6.46297857e-02 -2.10450411e-01 1.01058030e+00
5.41867077e-01 1.34557769e-01 1.11964738e+00 -1.65121460e+00
-5.13198495e-01 4.05914217e-01 -6.85493201e-02 -3.19905668e-01
3.83184701e-01 3.88389319e-01 7.65569985e-01 -4.25744146e-01
5.27683079e-01 1.09136486e+00 4.59669113e-01 1.07063651e+00
-1.60804880e+00 -6.45296693e-01 -4.54153985e-01 -2.12623522e-01
-1.14359725e+00 -3.60281795e-01 9.64198291e-01 -2.91905582e-01
5.38456798e-01 1.58211276e-01 6.75190091e-01 1.85422099e+00
3.40655446e-02 8.40720296e-01 1.20517361e+00 -5.52093029e-01
7.30777740e-01 5.01866221e-01 -6.16575718e-01 7.16909021e-02
-2.36322135e-01 2.75670260e-01 -2.24343777e-01 -3.94755185e-01
3.65560591e-01 -2.96288341e-01 -3.22568417e-01 -2.09911928e-01
-8.41351628e-01 1.15478981e+00 3.21565062e-01 1.80008784e-01
-8.15020859e-01 1.50157541e-01 3.18017066e-01 6.00120664e-01
5.74490428e-01 5.99088073e-01 -2.91251898e-01 -2.81935245e-01
-8.52138638e-01 4.34980899e-01 1.00887096e+00 5.26914775e-01
5.95026076e-01 5.58857203e-01 -3.57923746e-01 8.86973679e-01
2.24504285e-02 6.55863345e-01 9.75335538e-01 -6.07506216e-01
2.41442069e-01 2.76531249e-01 -4.18950617e-03 -4.80953068e-01
1.07567765e-01 -2.30832443e-01 -8.95519316e-01 4.25892919e-01
5.61393142e-01 -6.30688787e-01 -8.89234960e-01 2.10282946e+00
1.63401470e-01 1.13873079e-01 6.95069671e-01 5.20367324e-01
4.24242228e-01 1.02409482e+00 -1.02133108e-02 -4.47913073e-02
9.38616097e-01 -5.78660429e-01 -4.49277729e-01 -2.40060702e-01
4.13759857e-01 -5.21479547e-01 1.17450106e+00 3.33385319e-01
-9.33842659e-01 -5.08576632e-01 -9.43859518e-01 4.39965934e-01
-4.54148948e-01 1.20941877e-01 3.17913830e-01 8.65147412e-01
-1.01082206e+00 3.54228169e-01 -4.88108099e-01 1.17211185e-01
5.70472062e-01 9.10309777e-02 -3.74503404e-01 -9.55691095e-03
-1.25597548e+00 5.66467106e-01 2.51089424e-01 -1.86684877e-01
-1.04493546e+00 -6.03312850e-01 -9.11554277e-01 1.70777723e-01
-2.29767516e-01 -5.97466648e-01 1.40978527e+00 -1.84293532e+00
-1.70150220e+00 9.53611970e-01 1.92855924e-01 -5.60330272e-01
5.79502106e-01 2.58001983e-01 -5.29053628e-01 -1.21082865e-01
-6.10610060e-02 6.48459792e-01 1.36120856e+00 -1.25561786e+00
-2.01975927e-01 -1.13144763e-01 -2.51166373e-01 -9.28973313e-03
1.07720064e-03 -3.56281340e-01 2.55739361e-01 -9.16096628e-01
-5.02793550e-01 -1.02413237e+00 -5.55321341e-03 -3.60566080e-01
-7.16407120e-01 -1.10371217e-01 6.65012062e-01 -4.50990170e-01
7.22937286e-01 -2.27431631e+00 1.86048642e-01 3.71402770e-01
-1.70679465e-01 4.24367338e-01 -2.57600158e-01 5.31915188e-01
-5.62771499e-01 6.85263947e-02 -3.07202607e-01 -3.79112929e-01
1.36114836e-01 2.87271678e-01 -4.68459278e-01 1.84810206e-01
6.12278342e-01 9.00373280e-01 -8.94475996e-01 1.83512926e-01
-1.85261965e-01 6.73706591e-01 -5.11011481e-01 7.28794396e-01
-4.08691138e-01 5.75784564e-01 -2.60110617e-01 6.38663992e-02
4.22854602e-01 2.67338872e-01 4.91313599e-02 5.51919965e-03
6.34375691e-01 1.83748841e-01 -8.23793411e-01 1.32433438e+00
-6.47393703e-01 7.67338872e-01 -1.95142508e-01 -1.18368340e+00
1.24682581e+00 6.59619212e-01 6.27352670e-02 -4.91033226e-01
2.20224723e-01 -5.55738807e-05 2.52226442e-01 -2.17651948e-01
1.96462467e-01 -6.84730232e-01 -4.48416024e-01 7.33268082e-01
2.45132819e-01 -6.38526142e-01 -2.21594170e-01 7.97887594e-02
1.18716657e+00 -5.78250773e-02 1.09444752e-01 1.79078221e-01
3.52616519e-01 -2.97938675e-01 4.04556751e-01 5.43973327e-01
1.63133830e-01 8.14966619e-01 1.03957772e+00 -2.21024856e-01
-1.35965216e+00 -1.22332573e+00 1.18008636e-01 4.85891074e-01
-7.29640305e-01 -7.10121989e-02 -1.01175654e+00 -1.00428295e+00
-8.74814838e-02 1.31243598e+00 -1.02559137e+00 -7.03662753e-01
-1.06492519e-01 -7.04713345e-01 7.64361501e-01 4.16357458e-01
2.70025790e-01 -1.61652064e+00 -6.05320930e-01 4.02081273e-02
2.20238157e-02 -9.20073569e-01 -2.79427469e-01 4.90518302e-01
-3.38889867e-01 -8.88769269e-01 -8.04312348e-01 -5.50088465e-01
8.39806974e-01 -7.17950702e-01 1.30252302e+00 -5.21403670e-01
-1.16366453e-01 4.22435790e-01 -4.93184358e-01 -7.91357100e-01
-1.20659924e+00 -7.34754950e-02 -3.14617082e-02 5.07681668e-01
3.92409384e-01 -5.97474098e-01 -3.86887342e-01 3.70211527e-02
-1.19468975e+00 1.95152517e-02 4.73482877e-01 1.04576695e+00
2.97293454e-01 1.36094436e-01 8.93912196e-01 -1.16696262e+00
1.06934321e+00 -6.89860046e-01 -1.45070478e-01 -4.35303785e-02
-2.16786996e-01 2.35622913e-01 9.98812675e-01 -7.64771283e-01
-1.00568736e+00 -5.98133318e-02 -3.04409355e-01 -6.38013780e-01
-5.22208571e-01 2.69298255e-01 -4.36438143e-01 6.36938393e-01
8.03501546e-01 4.33572769e-01 2.73622245e-01 -1.21176176e-01
4.36159462e-01 8.33220482e-01 6.47297025e-01 -6.53479159e-01
8.77093017e-01 -2.57110856e-02 -2.77954131e-01 -6.83324933e-01
-5.96671104e-01 3.55480939e-01 -2.89634556e-01 -3.48464809e-02
7.95239031e-01 -6.56347215e-01 -1.99693233e-01 7.00902104e-01
-9.71030712e-01 -8.37635696e-01 -1.07078052e+00 3.75292569e-01
-8.49343717e-01 -4.64588940e-01 -1.69880912e-01 -8.66986573e-01
-2.94520646e-01 -7.96528459e-01 9.95708406e-01 3.74547243e-01
-5.11352897e-01 -1.00267470e+00 3.48802239e-01 -3.21689621e-02
3.92691195e-01 7.77933002e-01 1.02325153e+00 -1.09998095e+00
2.90434361e-01 -6.96599960e-01 3.41195852e-01 1.04938340e+00
2.96830863e-01 -9.04474780e-02 -1.21254230e+00 -2.95418441e-01
1.48310646e-01 -6.34064794e-01 2.75190592e-01 1.78405210e-01
1.13766694e+00 -5.09669125e-01 1.22327767e-01 4.67443675e-01
1.03300023e+00 5.20473003e-01 7.48504460e-01 -1.77961349e-01
2.77543038e-01 6.07084095e-01 2.29711190e-01 3.71427596e-01
-1.25373885e-01 6.51523292e-01 2.25658044e-01 -5.37972189e-02
9.78780817e-03 -6.70280933e-01 4.29307252e-01 4.11324829e-01
3.09062451e-01 -5.97708166e-01 -6.45869911e-01 4.88356352e-01
-1.26347923e+00 -1.02955723e+00 5.65939009e-01 2.21291924e+00
9.84749258e-01 1.74549565e-01 4.34892714e-01 4.26616997e-01
6.91181958e-01 1.36121837e-02 -8.31352413e-01 -7.37755895e-01
-1.53068140e-01 7.45836914e-01 -2.69082129e-01 8.43031257e-02
-7.72992909e-01 5.55198014e-01 5.77225733e+00 8.49817932e-01
-1.43871140e+00 -7.85565376e-02 1.12386250e+00 -2.44114667e-01
-4.60551769e-01 -4.56607550e-01 -1.60251006e-01 8.43438268e-01
1.26974773e+00 -3.36538941e-01 5.81663907e-01 7.87563205e-01
-3.99135090e-02 1.27610847e-01 -1.28100801e+00 1.05963385e+00
8.24223831e-02 -1.01442599e+00 6.90041259e-02 2.09601596e-01
8.76879036e-01 -2.04544544e-01 2.93894768e-01 6.42670810e-01
5.17603517e-01 -1.43867385e+00 7.39619553e-01 5.54589450e-01
1.08887422e+00 -9.61597443e-01 7.97693908e-01 4.66648489e-01
-3.70882839e-01 2.33995333e-01 -1.25360265e-01 6.18410446e-02
6.34095147e-02 4.75146621e-01 -1.11043775e+00 1.44976020e-01
1.02661826e-01 1.22799449e-01 -1.59654900e-01 4.25250739e-01
-4.87325430e-01 1.06039941e+00 -1.30100116e-01 -2.17625406e-02
2.84694016e-01 -2.60482401e-01 3.83209229e-01 1.10833275e+00
3.68795127e-01 -8.25158507e-02 -1.82742909e-01 1.12224281e+00
-3.71051043e-01 -1.19803190e-01 -9.13320661e-01 -5.31117380e-01
4.53538507e-01 1.21544862e+00 -4.33130533e-01 -2.82989651e-01
-9.22035724e-02 1.14232779e+00 2.43327186e-01 5.85245013e-01
-8.12884033e-01 -5.85278809e-01 8.32383513e-01 -1.15611330e-01
4.48156297e-01 2.73437202e-01 -6.77749589e-02 -9.60906863e-01
-1.11310422e-01 -1.14131653e+00 -2.61575612e-03 -9.98435616e-01
-1.54877031e+00 1.07614982e+00 -2.06161261e-01 -1.16273010e+00
-1.12275684e+00 -2.98721820e-01 -1.11773646e+00 1.29753625e+00
-8.11944664e-01 -9.17372406e-01 -1.81045368e-01 4.98359859e-01
3.70381355e-01 -7.13194013e-01 1.28171551e+00 -2.41949245e-01
-3.47253829e-01 8.69354546e-01 1.69465199e-01 4.01626557e-01
3.48596305e-01 -1.44122446e+00 4.74348545e-01 4.26846743e-01
4.13812131e-01 9.01517272e-02 8.58103931e-01 -2.64635444e-01
-9.15128589e-01 -1.20089066e+00 5.00463665e-01 -2.92807311e-01
2.40286544e-01 -6.14019334e-01 -8.08532417e-01 6.40665889e-01
2.33879969e-01 7.01549500e-02 9.20257151e-01 -3.95001262e-01
-2.44712085e-01 -1.84221733e-02 -1.77528548e+00 4.00305808e-01
3.21106702e-01 -5.75567663e-01 -4.40083563e-01 3.06527186e-02
3.03855628e-01 -3.21929723e-01 -7.96292186e-01 8.50111321e-02
4.86056179e-01 -1.02130866e+00 6.34256482e-01 -8.85223269e-01
7.58532405e-01 1.03772748e-02 -3.54512960e-01 -2.27886319e+00
1.23321079e-01 -5.94213367e-01 7.32086897e-02 1.47289944e+00
4.53228265e-01 -6.76815867e-01 9.53955948e-01 5.36567152e-01
1.29494876e-01 -1.04368865e+00 -9.41838384e-01 -5.44573367e-01
2.40106627e-01 -4.61426049e-01 8.01131189e-01 7.53595412e-01
-2.93969959e-01 5.02754271e-01 -2.85402566e-01 -1.78623393e-01
2.96544284e-01 -9.40898135e-02 1.09888542e+00 -7.84776390e-01
-5.91886401e-01 -3.23215336e-01 -6.60248756e-01 -5.69460213e-01
5.18588185e-01 -9.58821118e-01 9.35146362e-02 -9.32987332e-01
-2.33814940e-01 -4.00795132e-01 -5.95060037e-03 3.24224919e-01
-1.23250065e-02 3.75290841e-01 1.42547712e-01 -3.35671633e-01
1.10920392e-01 9.51282620e-01 1.00374043e+00 -1.32643476e-01
-1.02663867e-01 3.94848019e-01 -7.46734083e-01 4.72989053e-01
8.32726538e-01 -5.92234075e-01 -7.27953732e-01 1.25318319e-01
-1.66285206e-02 2.14046106e-01 2.92031944e-01 -8.82337272e-01
-4.63988453e-01 7.95909986e-02 6.56140208e-01 1.89153090e-01
5.45172095e-01 -5.86376607e-01 2.54673928e-01 2.44462594e-01
-5.62044919e-01 -7.29228929e-02 7.93761462e-02 4.89948750e-01
-4.44344580e-01 -3.65925133e-01 8.75525177e-01 3.80947329e-02
-4.15103137e-02 6.46350756e-02 -4.40186977e-01 5.14340103e-01
1.20001733e+00 -2.99621504e-02 1.26263231e-01 -1.00508845e+00
-7.82714903e-01 -4.28696901e-01 4.99873936e-01 4.07043755e-01
4.79464531e-01 -1.53588974e+00 -8.99457991e-01 7.79991150e-01
-4.44245674e-02 -1.05045009e-02 -8.62083212e-02 1.70854673e-01
-1.14667296e-01 -9.76585671e-02 -1.93574518e-01 -3.62061590e-01
-8.20093274e-01 3.53742540e-01 4.63225216e-01 -2.34822646e-01
-2.70615637e-01 7.39003837e-01 2.49699503e-01 -6.28573060e-01
-5.23001552e-02 -5.90498820e-02 1.73545673e-01 -1.88950133e-02
3.04392874e-01 -1.84913147e-02 -8.45227614e-02 -5.04608154e-01
1.06732072e-02 -1.76518649e-01 2.26872340e-01 -3.81502151e-01
1.36335945e+00 4.85372216e-01 2.15244964e-01 7.88607597e-01
1.36792755e+00 -6.72133546e-03 -1.39533818e+00 -1.19124569e-01
-4.12618220e-01 -3.17405641e-01 -2.50551671e-01 -9.39296186e-01
-1.23274839e+00 6.84426427e-01 4.55690861e-01 6.93463624e-01
1.25490224e+00 5.96920922e-02 5.67773879e-01 -2.20165700e-01
1.09387875e-01 -6.23478711e-01 1.53634310e-01 1.67494461e-01
1.01485658e+00 -9.28506017e-01 -7.19608307e-01 1.72893077e-01
-1.06392169e+00 9.39001620e-01 4.17378247e-01 -3.78034860e-01
4.38239962e-01 3.35553229e-01 2.64110476e-01 8.14171210e-02
-9.61708069e-01 1.11393265e-01 2.75935590e-01 1.13184083e+00
4.36152518e-01 2.79143631e-01 1.73942357e-01 7.13124096e-01
-7.34999239e-01 -1.76385060e-01 4.40654188e-01 3.59125704e-01
2.89449066e-01 -1.23062646e+00 -1.64316311e-01 7.61197746e-01
-3.22625309e-01 1.83769554e-01 -4.34981883e-01 6.63315117e-01
1.51648507e-01 7.80755877e-01 4.25112039e-01 -4.93612349e-01
3.16041410e-01 5.45628190e-01 3.20696592e-01 -7.01903462e-01
-5.12763619e-01 -3.63133490e-01 4.77599576e-02 -1.40412480e-01
1.05631668e-02 -9.48350012e-01 -8.59786451e-01 4.01145518e-02
1.39930919e-01 6.88018739e-01 9.07421052e-01 9.03691709e-01
1.82456687e-01 6.35938942e-01 9.57576632e-01 -9.09277856e-01
-7.83162653e-01 -1.30067849e+00 -7.23837018e-01 8.96882117e-01
2.03658342e-01 -2.86035597e-01 -5.80608904e-01 5.13352230e-02] | [11.735801696777344, -0.039917439222335815] |
5e80be49-9977-447e-9e1f-e580edd83777 | bag-of-words-baselines-for-semantic-code | null | null | https://aclanthology.org/2021.nlp4prog-1.10 | https://aclanthology.org/2021.nlp4prog-1.10.pdf | Bag-of-Words Baselines for Semantic Code Search | The task of semantic code search is to retrieve code snippets from a source code corpus based on an information need expressed in natural language. The semantic gap between natural language and programming languages has for long been regarded as one of the most significant obstacles to the effectiveness of keyword-based information retrieval (IR) methods. It is a common assumption that “traditional” bag-of-words IR methods are poorly suited for semantic code search: our work empirically investigates this assumption. Specifically, we examine the effectiveness of two traditional IR methods, namely BM25 and RM3, on the CodeSearchNet Corpus, which consists of natural language queries paired with relevant code snippets. We find that the two keyword-based methods outperform several pre-BERT neural models. We also compare several code-specific data pre-processing strategies and find that specialized tokenization improves effectiveness. | ['Jimmy Lin', 'Andrew Yates', 'Ji Xin', 'Xinyu Zhang'] | null | null | null | null | acl-nlp4prog-2021-8 | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [-4.92565241e-03 -1.88496426e-01 -6.42791569e-01 -9.11021754e-02
-1.26083553e+00 -6.06822371e-01 4.79785174e-01 6.72947228e-01
-5.57780921e-01 9.23304260e-02 4.55668300e-01 -7.89740801e-01
-4.47262198e-01 -5.43204665e-01 -4.38344657e-01 1.42243221e-01
1.64397761e-01 2.98621267e-01 3.72198880e-01 -4.18571025e-01
1.26617551e+00 -1.37469783e-01 -1.79318082e+00 5.32257020e-01
1.11141038e+00 8.33612800e-01 7.20135331e-01 3.40127409e-01
-1.04489696e+00 1.16782093e+00 -6.13863289e-01 -3.22922796e-01
-2.67110735e-01 -1.24665841e-01 -1.49578333e+00 -6.85057521e-01
1.50843978e-01 1.78367466e-01 -1.53735325e-01 1.29747796e+00
2.83436447e-01 6.43021911e-02 4.66709793e-01 -9.06408370e-01
-1.00040150e+00 9.10553873e-01 -3.30377430e-01 5.68191469e-01
8.84430170e-01 -3.96498263e-01 1.34352386e+00 -9.04099584e-01
8.15385997e-01 1.21295285e+00 6.55092180e-01 5.73867619e-01
-9.19470608e-01 -4.51489329e-01 -2.49940023e-01 2.10351348e-01
-1.39788365e+00 -3.81072253e-01 5.86559057e-01 -9.38283384e-01
1.65020192e+00 -2.83077769e-02 6.55705482e-02 6.18525386e-01
2.09130690e-01 8.12997222e-01 3.50930274e-01 -9.81803656e-01
1.55384570e-01 3.24142426e-01 3.75327706e-01 7.13607669e-01
2.75976330e-01 -1.74469471e-01 -3.70659173e-01 -6.98544323e-01
2.17700526e-01 2.88795512e-02 -1.10210925e-01 -2.60062695e-01
-7.90877759e-01 1.23362172e+00 2.94077218e-01 8.91744316e-01
4.77664080e-03 3.64929229e-01 9.42650974e-01 5.25507271e-01
4.27422345e-01 1.20232761e+00 -5.37182868e-01 -4.18794572e-01
-9.24504578e-01 4.15449381e-01 8.35355222e-01 1.22741663e+00
9.73419368e-01 -3.24679404e-01 -8.65786523e-02 1.29233801e+00
6.95900381e-01 3.08941811e-01 1.05593419e+00 -6.62793875e-01
6.30920231e-01 9.52030957e-01 -2.26213057e-02 -1.07939339e+00
-7.83493463e-03 -1.13287501e-01 2.48708919e-01 -3.54348540e-01
7.54399896e-02 3.05102170e-01 -4.78425115e-01 1.26463306e+00
-5.40248632e-01 -5.93915462e-01 2.16635481e-01 4.77450609e-01
9.58844423e-01 4.63966906e-01 2.87567854e-01 1.85798958e-01
1.38226736e+00 -8.82333338e-01 -3.40372175e-01 -6.34912491e-01
1.13008571e+00 -9.43050802e-01 1.18919230e+00 -6.98339120e-02
-8.58535945e-01 -2.50250310e-01 -8.58533323e-01 -1.39435232e-01
-6.15675271e-01 9.03535932e-02 7.81824589e-01 6.18523061e-01
-1.22358406e+00 2.71726280e-01 -5.16058505e-01 -7.23729610e-01
1.05491199e-01 -1.04607210e-01 3.80961248e-03 -3.36842686e-01
-9.83209789e-01 8.75913501e-01 3.82543623e-01 -6.62012815e-01
-7.97055662e-01 -7.04313874e-01 -1.02069604e+00 3.60889405e-01
3.57422233e-01 -1.93575710e-01 1.52501988e+00 -1.26823008e+00
-8.25179100e-01 1.15234637e+00 -3.45250666e-01 -2.24206701e-01
-4.32460994e-01 -4.31668013e-02 -4.07399684e-01 1.00027427e-01
5.20023227e-01 3.53678703e-01 4.39392000e-01 -1.22601688e+00
-5.58920264e-01 -1.77505210e-01 2.24972203e-01 -2.22190954e-02
-5.11644602e-01 8.53477836e-01 -4.07109767e-01 -5.83402991e-01
-1.52599066e-01 -7.73065507e-01 -4.66558635e-02 -4.87918377e-01
3.69354747e-02 -8.65737915e-01 1.83156401e-01 -5.85561216e-01
1.77102363e+00 -2.33892393e+00 -2.35755071e-01 2.18256921e-01
7.25155771e-02 8.81219730e-02 -2.58040845e-01 7.48632789e-01
-1.72003075e-01 6.32110059e-01 -1.14203691e-01 1.51994392e-01
2.89104935e-02 -1.78765908e-01 -5.54284096e-01 -1.05140597e-01
4.74208519e-02 9.09648478e-01 -1.12844777e+00 -6.16814733e-01
-3.98244053e-01 6.74111396e-02 -8.63254249e-01 5.30819893e-02
-4.16269153e-01 -3.54968250e-01 -9.08450246e-01 7.30413854e-01
-2.62950845e-02 -4.13932830e-01 -4.19240026e-03 5.42581677e-01
-2.25911811e-01 8.39494050e-01 -3.62365127e-01 2.05931115e+00
-6.53123200e-01 7.26327479e-01 -4.94268626e-01 -1.04347563e+00
9.96064782e-01 3.97090375e-01 5.27576685e-01 -1.10621762e+00
-1.83713496e-01 5.13229549e-01 -2.99130797e-01 -9.57330823e-01
6.97525263e-01 1.96604040e-02 -4.35711771e-01 4.91127491e-01
-9.34243128e-02 -1.35435462e-01 3.10611188e-01 3.44864428e-01
1.50278604e+00 6.66031465e-02 2.24389836e-01 -8.49480033e-01
4.96342152e-01 7.19849169e-01 1.94593012e-01 1.01261234e+00
2.15716884e-01 3.18706393e-01 4.62946385e-01 -3.55716795e-01
-8.02829802e-01 -5.47818422e-01 -9.16294232e-02 1.49120212e+00
1.72922790e-01 -6.51065707e-01 -6.82666540e-01 -7.79394031e-01
1.48401886e-01 8.41818094e-01 -3.03038090e-01 -2.96740711e-01
-3.95121038e-01 -3.67337793e-01 7.65056729e-01 3.28844756e-01
4.05018814e-02 -1.16222692e+00 -8.33554685e-01 3.80999506e-01
-3.84420633e-01 -6.47562623e-01 -4.74827379e-01 1.82897061e-01
-6.96894348e-01 -1.11372757e+00 -4.83762622e-01 -9.29059863e-01
6.75822198e-01 4.94467199e-01 1.72206891e+00 9.82051075e-01
-3.64972830e-01 6.18009984e-01 -1.09993649e+00 -2.40257725e-01
-8.12420666e-01 3.99970829e-01 -5.18438041e-01 -9.51506317e-01
1.06563246e+00 -1.08544670e-01 -3.78083229e-01 1.13654062e-01
-1.20126450e+00 -5.07906795e-01 5.27484834e-01 6.66297674e-01
-1.22027934e-01 2.35411599e-02 6.47022128e-01 -6.81176245e-01
1.18343174e+00 -7.52598703e-01 -6.42259240e-01 5.03171086e-01
-9.80568349e-01 5.10863721e-01 3.41946095e-01 -1.33359700e-01
-9.96574640e-01 -1.72444195e-01 -4.55816016e-02 2.65540425e-02
5.35782129e-02 1.28159022e+00 4.42193836e-01 -3.30860734e-01
1.18630254e+00 3.31562519e-01 -2.07018867e-01 -5.65090597e-01
4.87048877e-03 9.13969815e-01 8.95228684e-02 -1.24287212e+00
3.69588047e-01 4.95349877e-02 -9.14409399e-01 -6.32323623e-01
-5.06947398e-01 -1.14810932e+00 -1.87657103e-01 -1.43241673e-03
8.44362617e-01 -8.26867223e-01 -2.88756311e-01 -1.58201337e-01
-1.25111866e+00 -6.22552782e-02 7.77615607e-02 3.40520144e-01
-5.29947102e-01 5.27443349e-01 -6.43020868e-01 -4.28674370e-01
-2.45344907e-01 -1.43148720e+00 1.07965302e+00 -5.27928248e-02
-3.63993078e-01 -8.73333037e-01 3.91249955e-01 3.07217658e-01
7.78933823e-01 -5.66482544e-01 1.49990427e+00 -8.75156999e-01
-3.68194759e-01 -4.05433655e-01 -3.96305203e-01 1.75630972e-02
-3.76968132e-03 2.28863042e-02 -8.52936506e-01 -1.69625387e-01
-1.58755600e-01 -5.47133505e-01 8.98139000e-01 -6.48788586e-02
1.12797868e+00 -3.11903983e-01 -5.17762899e-01 5.30185970e-03
2.11222172e+00 7.01153576e-01 6.56183064e-01 5.87520659e-01
3.44934255e-01 5.60972691e-01 4.50378895e-01 4.55630392e-01
4.86211926e-01 5.28525531e-01 2.25791618e-01 4.76647139e-01
3.26340437e-01 -2.56187886e-01 2.46848837e-01 1.08683634e+00
3.27805489e-01 -3.01521868e-02 -1.64842939e+00 9.75157261e-01
-1.63954902e+00 -7.37934351e-01 2.93248504e-01 2.20277119e+00
1.12856829e+00 -1.57720417e-01 -1.51515275e-01 -3.16785276e-01
4.42061752e-01 -2.13533387e-01 -2.33438507e-01 -4.58884448e-01
4.05204326e-01 1.12606585e-01 5.54532349e-01 2.52904445e-01
-7.58847296e-01 9.80452776e-01 6.35784435e+00 8.76854181e-01
-7.77019262e-01 3.91623788e-02 1.54078424e-01 5.61865628e-01
-6.19913220e-01 4.30721402e-01 -6.56639159e-01 4.30169851e-01
9.91389036e-01 -6.60878241e-01 6.99729025e-01 1.31764507e+00
-3.19074661e-01 -2.82158434e-01 -1.33542788e+00 8.62168252e-01
3.63151729e-01 -1.43176687e+00 8.89680872e-04 -2.82825083e-01
7.00187266e-01 3.66387516e-01 -3.41783255e-01 7.12796748e-01
6.41376555e-01 -9.10811067e-01 7.97710359e-01 3.70686501e-01
5.65838456e-01 -4.45512354e-01 7.22357571e-01 1.47676602e-01
-1.29114854e+00 -4.41479266e-01 -4.99118239e-01 6.33213818e-02
-4.83892858e-01 3.11549842e-01 -6.22478902e-01 3.32915515e-01
7.98575044e-01 6.32124603e-01 -7.72807598e-01 1.24418485e+00
1.44074038e-01 3.05642933e-01 1.93500206e-01 -3.99582773e-01
4.42347348e-01 4.11574274e-01 1.99962288e-01 1.44275868e+00
3.51428181e-01 -3.32893729e-01 1.58680856e-01 1.27180135e+00
6.81398390e-03 3.06935847e-01 -8.85474026e-01 -4.94761497e-01
6.29816830e-01 7.06131816e-01 -8.18551660e-01 -4.34094399e-01
-1.01435554e+00 4.05432165e-01 1.78429246e-01 4.73220468e-01
-3.39578032e-01 -8.66740942e-01 4.00876313e-01 -1.39308453e-01
2.65252978e-01 -8.43284130e-02 1.08142219e-01 -1.17706621e+00
1.92863420e-01 -1.16350794e+00 5.55840850e-01 -8.05950820e-01
-1.28184283e+00 4.67269719e-01 9.51228291e-02 -1.05201375e+00
-4.15732473e-01 -4.20150399e-01 -2.88367867e-01 7.15851903e-01
-1.61220539e+00 -4.88187671e-01 -4.00224961e-02 3.19331825e-01
7.66676784e-01 -3.39114368e-01 8.75575960e-01 6.87020779e-01
-7.29033574e-02 4.10525888e-01 3.40805918e-01 4.65900987e-01
4.42142665e-01 -1.03863895e+00 4.05934215e-01 6.15156651e-01
1.52868748e-01 1.26995814e+00 5.88023245e-01 -7.09909916e-01
-1.59826517e+00 -6.59083605e-01 1.15344882e+00 -5.95782578e-01
8.45326900e-01 -1.24615803e-01 -9.80038106e-01 5.09784341e-01
2.45395303e-01 -3.08517247e-01 5.30442059e-01 1.65076941e-01
-8.62084866e-01 1.66962221e-01 -7.97463357e-01 1.44899651e-01
7.32605875e-01 -1.15296423e+00 -9.27542686e-01 3.66881371e-01
9.87124026e-01 6.79707155e-02 -8.92173827e-01 7.02962130e-02
2.18903229e-01 -5.00274301e-01 9.89719808e-01 -6.50178730e-01
8.51660609e-01 -6.26983568e-02 -3.09078097e-01 -1.13699830e+00
-7.31654316e-02 -2.48075172e-01 7.97110558e-01 1.02003670e+00
6.89174592e-01 -5.39303780e-01 3.34448934e-01 7.75773644e-01
-1.56929299e-01 -2.13404164e-01 -7.89861917e-01 -9.15316522e-01
4.12939668e-01 -4.23016131e-01 4.90163028e-01 1.17718351e+00
9.66765821e-01 2.21448481e-01 5.28729022e-01 -3.69057745e-01
1.18721105e-01 2.39373565e-01 3.07243258e-01 -1.36572766e+00
-1.07469752e-01 -8.73315990e-01 -2.08282664e-01 -8.65528047e-01
5.34880877e-01 -1.33291996e+00 5.09799063e-01 -1.50885713e+00
6.04090095e-01 -6.24364018e-01 -3.48158836e-01 6.41645670e-01
3.03126667e-02 -4.57625479e-01 -1.13174699e-01 5.23220003e-01
-7.25727141e-01 1.58863112e-01 4.47669804e-01 -3.22294146e-01
5.91099821e-02 -3.22257429e-01 -8.99677217e-01 4.44780737e-01
5.31723380e-01 -9.06186700e-01 -6.20974660e-01 -7.66055942e-01
1.13534665e+00 4.10619408e-01 7.76014328e-02 -7.48171568e-01
4.17517781e-01 -5.98091260e-02 -6.13671839e-01 4.61840630e-02
-4.90290284e-01 -8.34457397e-01 -3.32164347e-01 3.02220017e-01
-9.80759084e-01 5.05869091e-01 3.46449673e-01 3.51737589e-01
-5.08717954e-01 -1.12743855e+00 4.21869040e-01 -6.86001062e-01
-9.74431336e-01 -1.65774778e-01 -6.90347612e-01 6.24950826e-01
3.33649337e-01 -7.08982125e-02 -4.99722362e-01 3.24156992e-02
1.55164495e-01 1.24038272e-01 6.11938894e-01 8.73378873e-01
6.85639501e-01 -1.04890025e+00 -4.49683517e-01 -1.20365068e-01
9.27871168e-01 -3.94267172e-01 -3.04391682e-01 3.89555663e-01
-6.16757989e-01 9.32695031e-01 8.68182555e-02 -2.97258109e-01
-1.05827296e+00 8.44446480e-01 1.57353669e-01 -8.58772248e-02
-4.65074509e-01 7.80130148e-01 -8.97814557e-02 -4.65186298e-01
1.83947414e-01 -4.48055536e-01 -4.67437834e-01 -1.78126857e-01
5.63373566e-01 1.09079964e-02 2.23259240e-01 -3.13710660e-01
-5.84621191e-01 4.72860128e-01 -6.45536184e-02 9.22919624e-03
1.21073925e+00 5.50354272e-02 -6.39158249e-01 3.74800622e-01
1.39754426e+00 -1.43457800e-02 -1.82790413e-01 -5.98656952e-01
1.10460210e+00 -7.04800427e-01 4.80029061e-02 -7.22221315e-01
-8.97645950e-01 6.37316823e-01 3.42285544e-01 4.57560718e-01
7.70930231e-01 3.85713965e-01 4.31302398e-01 8.73224556e-01
6.90785170e-01 -1.20080197e+00 8.86912644e-02 7.23559916e-01
5.97875237e-01 -1.19464028e+00 -1.98192984e-01 3.13892886e-02
-1.91062003e-01 1.20961225e+00 5.34508288e-01 8.22726041e-02
4.89096493e-01 2.39137754e-01 4.12997529e-02 -6.65000498e-01
-7.59274602e-01 -2.29824454e-01 2.01699898e-01 1.36515349e-01
8.56140196e-01 -3.78249079e-01 -3.81719977e-01 3.18418533e-01
1.23538733e-01 3.30587663e-02 4.38091248e-01 1.32632375e+00
-6.81485891e-01 -1.01800752e+00 -1.63876936e-01 6.38709962e-01
-7.10484445e-01 -7.93009639e-01 -3.57155949e-01 5.42160630e-01
-4.34645832e-01 1.09384906e+00 -5.06127402e-02 -2.59222746e-01
-2.83836387e-02 2.89909720e-01 -3.51273594e-03 -1.08483589e+00
-8.78136218e-01 -3.48256648e-01 8.43456537e-02 -6.90825045e-01
-3.65365237e-01 -5.79126418e-01 -1.32382584e+00 9.50331539e-02
-6.35418415e-01 6.84850037e-01 9.19865549e-01 9.83873546e-01
4.97515887e-01 2.44949952e-01 3.66597295e-01 -7.30219632e-02
-5.75250745e-01 -7.95052707e-01 -4.20713164e-02 4.66659099e-01
1.68969885e-01 -5.87435305e-01 -4.88265902e-01 1.55737728e-01] | [7.521803379058838, 8.081120491027832] |
bd175b09-5481-4fb3-84d8-5ddd3f117e6d | graph-wise-common-latent-factor-extraction | 2112.08830 | null | https://arxiv.org/abs/2112.08830v3 | https://arxiv.org/pdf/2112.08830v3.pdf | Graph-wise Common Latent Factor Extraction for Unsupervised Graph Representation Learning | Unsupervised graph-level representation learning plays a crucial role in a variety of tasks such as molecular property prediction and community analysis, especially when data annotation is expensive. Currently, most of the best-performing graph embedding methods are based on Infomax principle. The performance of these methods highly depends on the selection of negative samples and hurt the performance, if the samples were not carefully selected. Inter-graph similarity-based methods also suffer if the selected set of graphs for similarity matching is low in quality. To address this, we focus only on utilizing the current input graph for embedding learning. We are motivated by an observation from real-world graph generation processes where the graphs are formed based on one or more global factors which are common to all elements of the graph (e.g., topic of a discussion thread, solubility level of a molecule). We hypothesize extracting these common factors could be highly beneficial. Hence, this work proposes a new principle for unsupervised graph representation learning: Graph-wise Common latent Factor EXtraction (GCFX). We further propose a deep model for GCFX, deepGCFX, based on the idea of reversing the above-mentioned graph generation process which could explicitly extract common latent factors from an input graph and achieve improved results on downstream tasks to the current state-of-the-art. Through extensive experiments and analysis, we demonstrate that, while extracting common latent factors is beneficial for graph-level tasks to alleviate distractions caused by local variations of individual nodes or local neighbourhoods, it also benefits node-level tasks by enabling long-range node dependencies, especially for disassortative graphs. | ['Ngai-Man Cheung', 'Thilini Cooray'] | 2021-12-16 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [ 2.24213108e-01 1.60911694e-01 -2.63508290e-01 -1.45666525e-01
-1.93161532e-01 -5.65634370e-01 6.32302582e-01 7.30191171e-01
-1.91570833e-01 6.46703601e-01 3.84786040e-01 -4.32073385e-01
-2.61392146e-01 -9.73311841e-01 -7.86063373e-01 -8.49306643e-01
-2.44079545e-01 2.28844762e-01 1.35194942e-01 -7.24004358e-02
5.77581897e-02 6.03054583e-01 -1.12938261e+00 2.04017609e-01
8.59946251e-01 4.13173139e-01 8.18208158e-02 3.59762937e-01
-2.29789898e-01 3.43508899e-01 -5.72868466e-01 -3.74857992e-01
7.46957883e-02 -4.57806975e-01 -5.63652396e-01 4.83141840e-02
1.35419920e-01 5.10988012e-02 -4.99443620e-01 1.10415161e+00
4.74849850e-01 1.24181785e-01 8.45700145e-01 -1.33237052e+00
-3.46110016e-01 6.84699714e-01 -5.87643325e-01 -5.56314550e-02
7.54531324e-02 1.82716891e-01 1.37636840e+00 -6.42034531e-01
8.14519584e-01 1.17395878e+00 4.68848109e-01 2.32535407e-01
-1.51817894e+00 -5.49059331e-01 4.19954926e-01 -3.23884115e-02
-1.25181210e+00 -2.02205494e-01 1.05140018e+00 -4.99834508e-01
9.32274401e-01 1.58691600e-01 6.05840981e-01 1.12650740e+00
3.34453881e-01 5.35115242e-01 8.54163051e-01 -3.17794472e-01
2.47437134e-01 -1.46968886e-01 1.63766563e-01 8.37435365e-01
4.13489372e-01 -2.70721346e-01 -5.70478082e-01 -4.83763486e-01
5.39326966e-01 2.24580571e-01 -4.45668310e-01 -6.13048673e-01
-1.20495605e+00 9.83427823e-01 7.04046965e-01 5.08125603e-01
-3.91182780e-01 9.83721465e-02 2.92763472e-01 1.74320608e-01
3.69830221e-01 5.41016757e-01 -4.25165832e-01 1.54184416e-01
-7.67645478e-01 1.35086179e-02 7.45628417e-01 7.06924736e-01
9.52002227e-01 -1.57663807e-01 -2.60361791e-01 5.65749586e-01
4.76623774e-01 -1.00087337e-01 3.06560069e-01 -1.09777674e-01
3.24674308e-01 1.12183249e+00 -2.22128421e-01 -1.31656468e+00
-6.50582731e-01 -6.98918581e-01 -9.79882658e-01 -6.02278998e-03
4.25389200e-01 7.97065347e-02 -1.08622503e+00 1.85381973e+00
3.27261597e-01 1.20829485e-01 -2.33185142e-01 7.28855729e-01
8.26951563e-01 5.12370646e-01 2.52123088e-01 -2.49087930e-01
1.47337592e+00 -8.40241492e-01 -7.15939760e-01 -2.80099094e-01
7.54312277e-01 -5.96389532e-01 1.07348323e+00 1.77307710e-01
-4.42990780e-01 -4.27243054e-01 -1.13610685e+00 -4.04452384e-02
-6.60324454e-01 -7.72599652e-02 9.10832524e-01 2.65150011e-01
-8.41402233e-01 8.59849036e-01 -8.17347825e-01 -2.45919257e-01
4.32176381e-01 4.57701385e-01 -6.62803411e-01 -9.33406800e-02
-1.36196744e+00 4.08566147e-01 2.89038807e-01 2.55842600e-03
-6.77528977e-01 -5.88015676e-01 -7.76529908e-01 4.09559339e-01
5.59059680e-01 -7.08422124e-01 2.88764715e-01 -5.87353647e-01
-1.18256295e+00 6.02721035e-01 -2.30859905e-01 -1.16068035e-01
2.78387517e-01 4.44425968e-03 -2.09937856e-01 4.97818477e-02
2.21938677e-02 4.61738259e-01 8.60830009e-01 -9.37511623e-01
9.55736563e-02 -5.85912526e-01 9.84033942e-02 -7.50916228e-02
-6.28873408e-01 -2.82104343e-01 -6.05076909e-01 -6.72106624e-01
6.57880977e-02 -1.05320919e+00 -4.47352767e-01 -6.61291108e-02
-6.45641983e-01 -2.85750955e-01 5.85467875e-01 -5.17842352e-01
1.30781794e+00 -2.02334952e+00 3.71784091e-01 3.98301393e-01
7.39816368e-01 3.00867409e-01 -2.34546915e-01 8.78136635e-01
-3.25534374e-01 4.10034269e-01 -1.71615630e-01 -1.77838400e-01
-5.13694398e-02 -1.47693139e-02 -3.25756446e-02 5.46397150e-01
4.46442217e-01 1.04135561e+00 -1.04101419e+00 -2.33429253e-01
7.65875820e-03 7.89391100e-01 -4.17043477e-01 1.94695622e-01
-3.82194281e-01 4.18969274e-01 -5.15530646e-01 3.15855116e-01
5.25740564e-01 -7.26509094e-01 5.21128058e-01 -5.12006164e-01
2.00898901e-01 4.30656105e-01 -1.08544219e+00 1.75736284e+00
-2.55637288e-01 3.30517679e-01 -1.96442217e-01 -9.80498731e-01
8.71999443e-01 1.67944461e-01 6.05381131e-01 -1.37942985e-01
-3.44477966e-02 -2.53961869e-02 3.83131474e-01 -1.91963002e-01
2.08684236e-01 1.04067922e-01 6.37768805e-02 4.78879392e-01
1.65786937e-01 2.86183894e-01 2.72898108e-01 7.09795415e-01
1.49481952e+00 -4.64986339e-02 4.45998758e-01 -2.65369117e-01
4.17046517e-01 -3.86943966e-01 4.67826754e-01 4.35752779e-01
-1.36794329e-01 3.12365055e-01 1.23649633e+00 -1.75576150e-01
-7.41498113e-01 -8.33278716e-01 7.85092786e-02 9.17725265e-01
-8.33024681e-02 -1.05033278e+00 -4.89962518e-01 -1.03358698e+00
1.22142956e-02 3.46392006e-01 -6.77021146e-01 -3.77859235e-01
-2.10428432e-01 -8.76327395e-01 3.57200950e-01 3.15689564e-01
-6.95942435e-03 -7.83465445e-01 -6.86399937e-02 3.25905085e-01
8.88672099e-02 -1.17536902e+00 -4.14536089e-01 4.78327036e-01
-7.89905012e-01 -1.00977290e+00 -4.75519747e-01 -4.20028120e-01
9.45750535e-01 4.18368369e-01 1.01493061e+00 3.17051709e-01
-2.40957931e-01 -6.33092076e-02 -3.56568307e-01 -2.84324706e-01
-2.85918027e-01 3.12178195e-01 -1.21588528e-01 2.24942967e-01
3.32585573e-01 -8.58659148e-01 -7.43359625e-01 1.51958019e-01
-9.55906808e-01 8.47881436e-02 6.02703214e-01 9.31375861e-01
5.53944886e-01 2.82934308e-01 4.50560033e-01 -1.16286123e+00
7.35400498e-01 -4.41568613e-01 -5.81312418e-01 1.63372934e-01
-7.20165014e-01 3.85668755e-01 8.47927928e-01 -4.64261860e-01
-4.35935915e-01 1.30953807e-02 1.02649229e-02 -4.59633619e-01
6.07791282e-02 8.02586079e-01 -5.07828891e-01 8.68582726e-02
6.38744831e-01 2.33995318e-02 3.45870703e-02 -3.63796264e-01
3.73609662e-01 3.13989609e-01 -1.88055068e-01 -3.21921408e-01
7.22723722e-01 5.13517797e-01 4.18999314e-01 -8.74355018e-01
-5.20048380e-01 -6.51232481e-01 -6.70900464e-01 -2.88553815e-02
7.73725092e-01 -8.12513411e-01 -7.06786811e-01 2.69184053e-01
-1.07863307e+00 -1.66631937e-01 5.37684634e-02 4.46749389e-01
-1.18160889e-01 4.72848356e-01 -5.14939964e-01 -5.11254966e-01
-2.22610027e-01 -1.41779375e+00 1.21428895e+00 6.99004978e-02
-3.09565693e-01 -1.21921754e+00 1.45016938e-01 1.83759615e-01
2.09275246e-01 3.81959230e-01 1.21453536e+00 -8.97051930e-01
-5.54569900e-01 -1.42025188e-01 -2.25906670e-01 -5.13747036e-02
4.18392956e-01 1.04172260e-01 -8.89909685e-01 -4.15490746e-01
-4.79763001e-01 -4.98578176e-02 1.03931177e+00 1.48371875e-01
9.23437655e-01 -1.52166337e-01 -6.49323046e-01 6.06607437e-01
1.35621953e+00 -3.12777132e-01 4.98276323e-01 -4.11360115e-02
1.16846550e+00 6.88919842e-01 1.74488425e-01 1.61413059e-01
1.52579933e-01 7.34560609e-01 3.75142485e-01 -3.76402497e-01
-4.77074198e-02 -4.77175415e-01 4.37120736e-01 8.00324023e-01
1.36311520e-02 -5.65376401e-01 -8.42337608e-01 4.11134869e-01
-1.96794796e+00 -5.81546009e-01 -2.91810960e-01 2.18014216e+00
7.95664430e-01 2.73818374e-01 6.16924502e-02 6.53159171e-02
5.90457678e-01 4.14917499e-01 -4.97810870e-01 -1.05393417e-01
1.10622924e-02 3.42331737e-01 3.73555571e-01 3.78014743e-01
-8.74288917e-01 1.10334516e+00 4.97111654e+00 7.95028567e-01
-1.17885578e+00 -1.46160409e-01 5.52430153e-01 1.84136659e-01
-5.29018104e-01 2.97775269e-01 -7.28419185e-01 3.97287279e-01
7.90363133e-01 3.14347334e-02 3.35429221e-01 6.01032913e-01
3.12321991e-01 2.03814149e-01 -1.16375554e+00 7.91973114e-01
-8.24943483e-02 -1.30440104e+00 1.95072249e-01 4.00173545e-01
6.20034456e-01 2.02135313e-02 -2.87263542e-01 1.30581707e-01
2.27561921e-01 -1.05050921e+00 -4.78115976e-02 3.05487931e-01
5.50319433e-01 -6.72921002e-01 6.15817070e-01 1.91841125e-01
-1.26914966e+00 4.06218350e-01 -3.95672143e-01 1.07623845e-01
2.21167170e-02 1.11500657e+00 -9.31068838e-01 7.59660184e-01
2.28586212e-01 8.30323040e-01 -6.68561935e-01 6.59066856e-01
-6.02998614e-01 6.92162156e-01 -1.95274770e-01 -3.29588167e-02
2.59091794e-01 -4.21521842e-01 5.55230796e-01 1.11643565e+00
1.37721524e-02 -2.10032687e-01 2.58460581e-01 8.59346747e-01
-4.08849329e-01 3.39070886e-01 -8.81221831e-01 -5.77841222e-01
1.87502295e-01 1.58471358e+00 -1.13378692e+00 -2.41290838e-01
-5.05232334e-01 9.67316329e-01 6.59276485e-01 4.85213995e-01
-7.11925089e-01 -3.26667517e-01 7.12768853e-01 2.88771302e-01
4.36351180e-01 -3.63841891e-01 1.10130208e-02 -1.21097696e+00
-8.02157260e-03 -9.26684439e-01 3.90978485e-01 -3.79404515e-01
-1.30038774e+00 4.67814803e-01 -2.51608104e-01 -1.03980041e+00
4.90920767e-02 -4.73619342e-01 -6.26318693e-01 7.81170964e-01
-1.50892818e+00 -1.17812300e+00 -1.67709649e-01 5.10142386e-01
1.70907408e-01 -4.02609073e-02 8.59632909e-01 2.85458952e-01
-5.12928963e-01 6.60480082e-01 -8.99332017e-02 2.05245942e-01
7.37473488e-01 -1.11297894e+00 4.92733687e-01 9.62270617e-01
5.31285226e-01 1.16603029e+00 5.54756045e-01 -9.07693148e-01
-1.56875658e+00 -1.12922978e+00 9.53559160e-01 -2.21003339e-01
8.90224814e-01 -7.52392709e-01 -1.09979486e+00 4.04612899e-01
-1.27981320e-01 1.71637431e-01 9.35435832e-01 4.41170126e-01
-4.90013838e-01 5.18707670e-02 -6.92843676e-01 7.27592230e-01
1.28805876e+00 -6.16976857e-01 -8.95128548e-02 6.11547947e-01
7.16251969e-01 -5.08815013e-02 -9.93447542e-01 3.94211620e-01
3.67752105e-01 -7.76410043e-01 8.09407055e-01 -9.24134493e-01
4.05879110e-01 -4.45011020e-01 2.69629061e-01 -1.54532051e+00
-5.80454469e-01 -7.85820067e-01 -2.46393666e-01 1.19501853e+00
5.45330048e-01 -6.95072949e-01 8.37849796e-01 3.18164855e-01
6.20539747e-02 -7.69618630e-01 -6.91141725e-01 -5.67913949e-01
-2.84300208e-01 -9.19336975e-02 5.83600223e-01 1.29012561e+00
3.26182470e-02 1.03226030e+00 -2.75636137e-01 2.22488806e-01
4.26273316e-01 1.91993415e-01 1.01460326e+00 -1.27954912e+00
-5.77248216e-01 -4.32294130e-01 -6.17191970e-01 -8.21923912e-01
1.41074285e-01 -1.23474765e+00 -2.20274821e-01 -1.63283610e+00
2.30641097e-01 -2.87679225e-01 -4.82392102e-01 5.70937335e-01
-6.44637406e-01 -1.02965944e-01 1.35023788e-01 1.38749838e-01
-3.98905337e-01 6.21528745e-01 1.34795856e+00 -2.58131117e-01
-4.16142642e-01 -2.26477012e-01 -8.06737959e-01 4.32776928e-01
6.09350741e-01 -5.59252739e-01 -6.15830898e-01 -3.90676707e-02
5.75357914e-01 -9.96815860e-02 6.75202385e-02 -5.94959259e-01
1.16769746e-01 -5.57370782e-02 1.62638426e-01 -1.09296568e-01
1.58196941e-01 -7.58391380e-01 2.85311192e-01 4.94672149e-01
-2.84114093e-01 -5.18302508e-02 2.28012465e-02 8.89905870e-01
-1.96115360e-01 3.17598470e-02 5.03076911e-01 -7.15687200e-02
-1.90034568e-01 5.86137831e-01 -1.60381258e-01 -2.13787988e-01
7.66888082e-01 2.51175035e-02 -4.31266367e-01 -4.26196605e-01
-6.46340847e-01 8.92475322e-02 5.15445411e-01 3.60623121e-01
4.06881332e-01 -1.08271432e+00 -5.59280097e-01 2.84675866e-01
3.79191488e-01 -4.23454158e-02 -3.38275060e-02 9.17743325e-01
-1.35281891e-01 2.31544137e-01 2.29281373e-03 -5.15772402e-01
-1.30401421e+00 6.93477035e-01 -2.22387135e-01 -8.54868591e-01
-4.94062304e-01 7.61924088e-01 5.17323375e-01 -2.37263128e-01
-3.59018310e-03 -2.76058733e-01 -3.63862306e-01 1.87728971e-01
2.95827746e-01 7.62008503e-02 2.87268847e-01 -5.71753621e-01
-3.95596176e-01 3.97147477e-01 -3.40669870e-01 3.75462353e-01
1.43702614e+00 4.04720247e-01 -2.38442123e-01 3.56272697e-01
1.27875519e+00 3.98010910e-01 -9.77986515e-01 -1.73040047e-01
-1.01592101e-01 -3.04618329e-01 1.76235691e-01 -4.06292975e-01
-1.13071918e+00 1.01952672e+00 2.73075163e-01 2.03180283e-01
7.99006164e-01 -2.26121023e-02 6.36518896e-01 1.52197242e-01
1.81583986e-01 -7.01887369e-01 1.27259955e-01 2.10530341e-01
5.99740922e-01 -1.21019125e+00 4.33805734e-01 -6.84874058e-01
-3.78758907e-01 1.10641563e+00 3.85165364e-01 6.61757514e-02
6.77312732e-01 5.90782277e-02 -3.20226282e-01 -6.31832838e-01
-7.68910587e-01 -2.80665159e-01 4.98806715e-01 3.23639929e-01
6.21622443e-01 2.07792178e-01 -4.56173867e-01 3.74433875e-01
1.31438181e-01 -3.17211211e-01 2.57178515e-01 7.24517286e-01
-7.17112198e-02 -1.45753562e+00 7.56812021e-02 5.90478361e-01
-3.51766109e-01 -2.77157128e-01 -7.88741946e-01 6.48314118e-01
-7.09084049e-02 6.71107352e-01 -2.41142735e-01 -4.32284117e-01
6.13113008e-02 9.04704705e-02 3.11958343e-01 -7.48937249e-01
-5.53541958e-01 3.24576586e-01 1.44370079e-01 -4.90866899e-01
-3.65039349e-01 -4.02247250e-01 -1.37303579e+00 -1.48976326e-01
-5.27118266e-01 1.57597438e-01 5.35003960e-01 8.10191989e-01
7.63359368e-01 7.03622937e-01 4.37885970e-01 -5.51449299e-01
-1.72699958e-01 -9.83602524e-01 -4.08456534e-01 7.56278694e-01
6.00102209e-02 -7.79233873e-01 -4.54059631e-01 -8.29679370e-02] | [5.465051651000977, 5.922904014587402] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.