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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
a1b0be94-109a-42d6-8369-aa7b1d612b98 | exploring-automatically-perturbed-natural | 2305.15520 | null | https://arxiv.org/abs/2305.15520v1 | https://arxiv.org/pdf/2305.15520v1.pdf | Exploring Automatically Perturbed Natural Language Explanations in Relation Extraction | Previous research has demonstrated that natural language explanations provide valuable inductive biases that guide models, thereby improving the generalization ability and data efficiency. In this paper, we undertake a systematic examination of the effectiveness of these explanations. Remarkably, we find that corrupted explanations with diminished inductive biases can achieve competitive or superior performance compared to the original explanations. Our findings furnish novel insights into the characteristics of natural language explanations in the following ways: (1) the impact of explanations varies across different training styles and datasets, with previously believed improvements primarily observed in frozen language models. (2) While previous research has attributed the effect of explanations solely to their inductive biases, our study shows that the effect persists even when the explanations are completely corrupted. We propose that the main effect is due to the provision of additional context space. (3) Utilizing the proposed automatic perturbed context, we were able to attain comparable results to annotated explanations, but with a significant increase in computational efficiency, 20-30 times faster. | ['Xingran Chen', 'Wanyun Cui'] | 2023-05-24 | null | null | null | null | ['relation-extraction'] | ['natural-language-processing'] | [ 4.92474645e-01 6.32936120e-01 -5.45110762e-01 -4.90362614e-01
-5.99179149e-01 -4.47049618e-01 8.18752050e-01 2.66344577e-01
-2.68971145e-01 1.00979722e+00 7.72564292e-01 -6.08597100e-01
-2.02372864e-01 -4.68747526e-01 -6.81759417e-01 -5.19762576e-01
8.17634091e-02 2.67967939e-01 -9.81215239e-02 -2.68968880e-01
9.09709275e-01 2.99906790e-01 -1.78016675e+00 2.63045251e-01
1.16811895e+00 3.11366379e-01 -2.72161122e-02 4.67405438e-01
-4.87868130e-01 9.21724916e-01 -4.52205747e-01 -5.79379261e-01
9.18717906e-02 -4.44620311e-01 -9.23957109e-01 -1.75740626e-02
3.75855833e-01 -1.55649558e-01 -2.81565767e-02 7.35286534e-01
2.47629642e-01 1.73949644e-01 6.58802629e-01 -1.30689764e+00
-1.12928689e+00 1.11686718e+00 -2.64702439e-01 4.31577675e-02
3.83276224e-01 1.18329540e-01 1.08784270e+00 -1.00622594e+00
5.62412858e-01 1.22798419e+00 7.01447964e-01 1.07290447e+00
-1.26788199e+00 -7.09406316e-01 7.17302144e-01 -2.28085723e-02
-1.00513339e+00 -4.44527656e-01 6.02313638e-01 -2.58807659e-01
1.01653790e+00 3.68505597e-01 3.42574209e-01 1.26735926e+00
2.73466349e-01 6.33479059e-01 1.17505074e+00 -7.60976017e-01
1.12882808e-01 7.13117838e-01 5.58849454e-01 4.74948049e-01
6.37947738e-01 2.56352007e-01 -5.96233070e-01 -3.14695418e-01
1.55963361e-01 -3.58002447e-02 -3.58972013e-01 -1.72903240e-01
-9.95629430e-01 9.55113947e-01 4.76946384e-01 4.92506593e-01
-2.88559109e-01 1.29931197e-01 3.17605168e-01 1.67880192e-01
4.85239506e-01 8.43553960e-01 -6.21867120e-01 -7.90162906e-02
-6.68453276e-01 3.90951842e-01 7.18621373e-01 9.90589857e-01
7.02407300e-01 2.05203339e-01 6.30144328e-02 7.34803677e-01
2.48257577e-01 3.98814827e-01 7.43849277e-01 -8.36881518e-01
4.63772506e-01 6.77959919e-01 2.35007107e-01 -1.17680264e+00
-3.13986778e-01 -5.32475233e-01 -5.04596889e-01 -9.56238359e-02
4.72804308e-01 -7.63926357e-02 -6.97522521e-01 2.08158016e+00
-8.43787566e-02 -2.75068849e-01 1.97194964e-01 9.11384284e-01
7.28564918e-01 2.33152658e-01 3.81555796e-01 -2.40811005e-01
1.01000273e+00 -8.57719064e-01 -7.63187587e-01 -6.62071168e-01
7.77106285e-01 -6.66171968e-01 1.59927320e+00 2.32655779e-01
-8.97163033e-01 -4.80472833e-01 -8.96105468e-01 -1.31547675e-01
-4.19890463e-01 -1.25582099e-01 1.06439245e+00 7.70822048e-01
-9.31024969e-01 4.81279582e-01 -3.25184345e-01 -5.57270467e-01
4.05484773e-02 4.37871486e-01 -1.52709514e-01 3.63938627e-03
-1.21398568e+00 8.64412963e-01 2.07271084e-01 -1.49128214e-01
-3.84233773e-01 -7.35445380e-01 -8.54199588e-01 6.86457306e-02
4.55713272e-01 -6.74135208e-01 1.12807572e+00 -1.48227036e+00
-9.46734011e-01 4.99718577e-01 -7.37036288e-01 -6.19901001e-01
4.57224190e-01 -4.61740255e-01 -2.35894993e-01 -2.09764153e-01
2.13978618e-01 8.23187709e-01 3.47311974e-01 -1.70707726e+00
-4.54949468e-01 -1.24739975e-01 3.43425810e-01 -1.39358006e-02
-4.19981658e-01 -3.62239510e-01 -1.30689025e-01 -8.62919867e-01
2.17213720e-01 -1.04409087e+00 -2.12381214e-01 -3.67604345e-01
-4.26082909e-01 -2.63704717e-01 5.23858190e-01 -1.99247777e-01
1.34530532e+00 -2.01904440e+00 -1.85166635e-02 -1.33228684e-02
2.29816198e-01 1.67490214e-01 3.52635384e-02 4.83509868e-01
-2.18074739e-01 7.33653843e-01 -2.94769645e-01 -4.61363345e-01
-1.68109536e-02 3.87947977e-01 -8.47942948e-01 6.50994703e-02
2.85270423e-01 6.87467039e-01 -8.59278917e-01 -3.69649917e-01
-2.41989363e-02 3.27141076e-01 -8.97655308e-01 -5.07494695e-02
-1.72771066e-01 1.66209728e-01 -4.34101015e-01 4.94591087e-01
4.34506774e-01 -2.39182726e-01 2.17158228e-01 2.51618415e-01
-1.30382419e-01 6.45802557e-01 -1.05385768e+00 1.15425467e+00
-4.18119758e-01 7.57775247e-01 -3.77427250e-01 -6.28439188e-01
9.58721638e-01 2.40181923e-01 -6.18012175e-02 -5.89541256e-01
-2.45756153e-02 3.38833928e-01 5.80186732e-02 -4.60948020e-01
1.02526355e+00 -4.37612683e-01 1.75999254e-01 6.45796955e-01
-3.94219548e-01 1.00510329e-01 8.40035155e-02 3.93825054e-01
7.47616827e-01 -3.83530278e-03 3.99422228e-01 -4.09060270e-01
4.45568413e-01 2.25135908e-01 5.48809409e-01 1.20668900e+00
-3.58419456e-02 5.18657982e-01 3.95042539e-01 -4.31402892e-01
-8.46505046e-01 -6.62428975e-01 -1.57450527e-01 1.06007123e+00
2.26341993e-01 -3.83390963e-01 -4.28297311e-01 -7.05582976e-01
1.77420780e-01 1.32407117e+00 -9.76245224e-01 -2.92331487e-01
-4.44700509e-01 -5.68486035e-01 3.96734864e-01 7.37899840e-01
2.94274151e-01 -1.03869414e+00 -5.94181299e-01 5.62214516e-02
-2.19236642e-01 -8.01915884e-01 -3.18948179e-01 3.16062123e-01
-1.36808610e+00 -1.01950765e+00 -1.94964945e-01 -2.22948790e-01
8.36756229e-01 7.02568710e-01 1.20268369e+00 8.74381721e-01
2.58985102e-01 5.02235949e-01 -5.21084428e-01 -6.61917448e-01
-4.32612181e-01 2.13130891e-01 2.44829938e-01 -4.90679532e-01
5.40514112e-01 -3.48924965e-01 -3.71097893e-01 2.28050113e-01
-9.09645379e-01 2.05081198e-02 6.67673886e-01 9.00350630e-01
1.94884673e-01 -2.36972913e-01 9.18809056e-01 -1.56721354e+00
7.88818538e-01 -5.91947198e-01 -3.94301526e-02 9.92534496e-03
-1.43509603e+00 5.36714077e-01 7.59237528e-01 -3.99227440e-01
-1.36880481e+00 -3.85082543e-01 5.30071929e-02 9.36256275e-02
-3.13994616e-01 4.42583174e-01 2.29978979e-01 9.03439820e-02
8.85330439e-01 1.72529630e-02 -3.00559103e-01 -2.43707508e-01
4.49074626e-01 4.83854860e-01 2.04756007e-01 -7.68577933e-01
6.68541014e-01 4.23118979e-01 -2.84825832e-01 -5.45572281e-01
-1.00372803e+00 -2.30252191e-01 -4.65432435e-01 3.20747606e-02
4.42634374e-01 -6.49591923e-01 -2.05835819e-01 -3.60616505e-01
-9.17999566e-01 -1.95911050e-01 -1.35808736e-01 5.56154311e-01
-2.05170155e-01 3.40220004e-01 -4.16611373e-01 -1.02756119e+00
-1.89460754e-01 -1.06916916e+00 6.32723868e-01 3.31607878e-01
-8.85331392e-01 -1.21214998e+00 -1.90349877e-01 5.32095730e-01
3.28437179e-01 1.13501795e-01 1.24468613e+00 -8.27318251e-01
-5.06901741e-01 -7.19139203e-02 -9.87716317e-02 3.08280066e-02
1.38257876e-01 1.64176464e-01 -1.11405826e+00 -8.70479923e-03
-1.24492273e-01 -2.91494817e-01 9.43864405e-01 1.92942336e-01
9.44504559e-01 -4.77404565e-01 -2.99865633e-01 8.94258171e-02
1.34956098e+00 3.76380831e-02 5.00469863e-01 6.40687585e-01
4.54052299e-01 7.42439628e-01 7.99356937e-01 2.98850298e-01
4.91350353e-01 3.07785869e-01 3.28297943e-01 -2.47579709e-01
-1.02413982e-01 -3.81247014e-01 2.64855176e-01 5.08323848e-01
-6.17622957e-02 -2.45549500e-01 -8.36407304e-01 8.46757710e-01
-1.84363258e+00 -9.83818948e-01 -5.02183437e-01 2.24120307e+00
6.48543000e-01 3.06605250e-01 -2.46281594e-01 3.32289249e-01
4.56247032e-01 -2.34050099e-02 -5.39023459e-01 -8.65385532e-01
-1.21416591e-01 -2.28630498e-01 3.71606827e-01 7.35216498e-01
-4.39969003e-01 7.64010847e-01 7.61110973e+00 1.55960351e-01
-1.12071717e+00 -2.54602194e-01 4.29959416e-01 -2.64142215e-01
-9.98464882e-01 1.92224592e-01 -6.68479085e-01 2.83267021e-01
8.35353136e-01 -2.89267808e-01 2.52587467e-01 8.38547766e-01
3.03990930e-01 -1.90387622e-01 -1.18175066e+00 2.48316363e-01
1.64334670e-01 -1.19164848e+00 4.39302832e-01 1.41679749e-01
8.71937931e-01 -3.06057155e-01 1.31112218e-01 4.08428609e-01
2.10974157e-01 -9.70019639e-01 9.19706047e-01 3.35691869e-01
2.04105631e-01 -8.29371929e-01 9.47963297e-01 3.67305726e-01
-6.11947536e-01 -4.37394202e-01 -4.23136622e-01 -6.60569906e-01
-2.86862284e-01 5.40949047e-01 -1.09106064e+00 4.92706835e-01
3.85663629e-01 4.22028601e-01 -6.84887946e-01 5.79157174e-01
-5.34441650e-01 9.22839642e-01 2.23402634e-01 -1.32911757e-01
3.44686270e-01 7.57435039e-02 4.75495487e-01 1.26157725e+00
1.89462394e-01 2.52956659e-01 -3.11461240e-01 9.17951405e-01
-3.73119698e-03 1.18011437e-01 -8.58405948e-01 -3.60769443e-02
6.68228149e-01 7.93255091e-01 -6.75833583e-01 -4.95502323e-01
-4.92980927e-01 5.48014641e-01 4.40876573e-01 3.79198581e-01
-8.60889912e-01 -8.17146823e-02 6.03475153e-01 1.82329580e-01
5.45624383e-02 7.21690431e-02 -9.94213879e-01 -1.08658659e+00
-9.00431871e-02 -1.01856709e+00 3.94902200e-01 -8.55505228e-01
-1.10618758e+00 6.00062191e-01 3.17320861e-02 -7.39062369e-01
-3.11622798e-01 -4.09855366e-01 -4.96125221e-01 8.03124547e-01
-1.48308313e+00 -7.36688554e-01 -2.45751843e-01 8.49161074e-02
6.92221940e-01 -7.94264078e-02 8.60529959e-01 6.40623420e-02
-5.19098938e-01 6.56408072e-01 -8.61771181e-02 -3.08532923e-01
8.40812624e-01 -1.39380181e+00 2.00215429e-01 9.06813920e-01
1.85085669e-01 1.42397606e+00 1.18902981e+00 -6.18713796e-01
-1.13290095e+00 -8.15093040e-01 1.35042369e+00 -6.68289900e-01
3.33955109e-01 -5.08241914e-02 -1.20056272e+00 7.62250543e-01
2.36530051e-01 -6.27081990e-01 9.01533246e-01 6.76049948e-01
-4.19355810e-01 1.85067281e-01 -1.16803467e+00 8.72859299e-01
1.06438220e+00 -4.54809427e-01 -9.56911385e-01 9.13252309e-03
7.70156801e-01 -2.42260650e-01 -3.34907025e-01 2.31505722e-01
6.05441451e-01 -1.25002205e+00 6.59518957e-01 -6.76882148e-01
8.60293031e-01 -2.06278950e-01 -2.26214796e-01 -1.24425757e+00
-3.22966427e-01 -1.99034959e-01 2.88950980e-01 1.50599432e+00
8.33417833e-01 -9.02858615e-01 5.99749863e-01 1.33432591e+00
-5.82403131e-02 -7.17426181e-01 -4.46214378e-01 -3.87936473e-01
2.57217735e-01 -6.28974020e-01 8.06614518e-01 1.09561300e+00
1.38041779e-01 1.78975508e-01 -3.16943407e-01 1.87589303e-01
3.62127364e-01 2.23617315e-01 7.64901280e-01 -1.00248027e+00
-1.20135888e-01 -2.58126646e-01 1.66448578e-01 -8.95729601e-01
3.67905021e-01 -6.76456392e-01 8.15520287e-02 -1.37948096e+00
4.58689600e-01 -5.74492335e-01 -2.04650119e-01 5.58398247e-01
-7.13605702e-01 1.13589898e-01 2.45241448e-01 3.31533879e-01
-2.18915969e-01 3.93695176e-01 9.99843597e-01 1.92148298e-01
-9.12397951e-02 -1.92864537e-01 -1.59582973e+00 7.34552145e-01
8.40773225e-01 -6.67844355e-01 -6.48403287e-01 -7.39187002e-01
2.87196726e-01 -4.10569072e-01 3.43822896e-01 -6.44196451e-01
4.78299372e-02 -3.41086924e-01 2.51407862e-01 -1.19957887e-01
5.25821410e-02 -6.97764397e-01 2.57923622e-02 5.01604319e-01
-8.67983878e-01 1.53611526e-01 5.28251410e-01 6.57639325e-01
-6.46077842e-02 -4.09814686e-01 3.19979459e-01 -1.00260347e-01
-6.06915474e-01 -5.01959443e-01 -5.48552275e-01 8.66449787e-04
6.11883640e-01 -4.24267173e-01 -3.75710040e-01 -4.63567644e-01
-2.69384176e-01 7.99422711e-03 6.86310112e-01 4.98456031e-01
3.14974457e-01 -1.22611105e+00 -3.99032176e-01 2.91030686e-02
1.84760228e-01 -3.47495764e-01 -7.12917000e-02 6.04703903e-01
6.47572950e-02 7.96015203e-01 1.40979573e-01 -3.76172274e-01
-1.19957614e+00 4.62623000e-01 1.92775235e-01 -6.82877824e-02
-5.36426246e-01 7.16965735e-01 3.32674593e-01 -5.41102171e-01
3.51408631e-01 -5.65167546e-01 -1.31526487e-02 -8.18557665e-02
5.03372908e-01 3.35911483e-01 -9.07949805e-02 -3.57402235e-01
-4.04551059e-01 4.13420469e-01 -3.05396646e-01 1.21791176e-01
1.19323599e+00 -2.52924412e-01 1.87155977e-01 6.35935664e-01
6.16589367e-01 3.81716996e-01 -1.02097523e+00 -4.59093973e-02
8.03691745e-02 -6.03232205e-01 -9.80131179e-02 -1.02140069e+00
-7.78041899e-01 7.14302421e-01 2.25515515e-01 3.80970955e-01
9.25141096e-01 -2.59178039e-02 3.03992808e-01 4.00652111e-01
4.46370631e-01 -7.52882123e-01 -1.24188803e-01 3.86338860e-01
9.74887729e-01 -1.36261737e+00 2.16717079e-01 -4.02124375e-01
-8.78146887e-01 9.82048392e-01 6.52589083e-01 1.44908726e-01
-1.41140753e-02 -7.27436095e-02 3.03850174e-01 -1.80536509e-01
-1.12675393e+00 1.22454286e-01 2.01739565e-01 5.43277919e-01
1.06442070e+00 -5.27759120e-02 -6.05280101e-01 7.92011738e-01
-3.27578664e-01 -1.34264380e-02 7.50467181e-01 6.56404793e-01
-3.36212486e-01 -1.21350133e+00 -5.08028030e-01 3.32300425e-01
-3.72467399e-01 -3.27818066e-01 -9.47078288e-01 1.25372553e+00
2.78644469e-02 1.31823063e+00 -1.83355629e-01 -2.69117981e-01
4.34057295e-01 3.23108494e-01 1.26130227e-02 -6.48801267e-01
-8.95931780e-01 -5.80541015e-01 2.12809965e-01 -5.08927226e-01
-4.64254618e-01 -6.69990897e-01 -1.55828643e+00 -5.90947688e-01
-5.14908075e-01 5.10300100e-01 4.13205624e-01 1.27088618e+00
5.43165922e-01 6.25812471e-01 3.30339938e-01 -4.16308820e-01
-5.67745566e-01 -8.78713429e-01 -2.79427320e-01 5.34684241e-01
4.53304112e-01 -8.28742564e-01 -9.20433760e-01 2.86171567e-02] | [9.184073448181152, 6.253123760223389] |
9dbc0d95-f7a1-4acb-ac9f-a2f55d8143fb | incentive-theoretic-bayesian-inference-for | 2307.03748 | null | https://arxiv.org/abs/2307.03748v1 | https://arxiv.org/pdf/2307.03748v1.pdf | Incentive-Theoretic Bayesian Inference for Collaborative Science | Contemporary scientific research is a distributed, collaborative endeavor, carried out by teams of researchers, regulatory institutions, funding agencies, commercial partners, and scientific bodies, all interacting with each other and facing different incentives. To maintain scientific rigor, statistical methods should acknowledge this state of affairs. To this end, we study hypothesis testing when there is an agent (e.g., a researcher or a pharmaceutical company) with a private prior about an unknown parameter and a principal (e.g., a policymaker or regulator) who wishes to make decisions based on the parameter value. The agent chooses whether to run a statistical trial based on their private prior and then the result of the trial is used by the principal to reach a decision. We show how the principal can conduct statistical inference that leverages the information that is revealed by an agent's strategic behavior -- their choice to run a trial or not. In particular, we show how the principal can design a policy to elucidate partial information about the agent's private prior beliefs and use this to control the posterior probability of the null. One implication is a simple guideline for the choice of significance threshold in clinical trials: the type-I error level should be set to be strictly less than the cost of the trial divided by the firm's profit if the trial is successful. | ['Jake A. Soloff', 'Michael Sklar', 'Michael I. Jordan', 'Stephen Bates'] | 2023-07-07 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [ 3.96634847e-01 3.56685728e-01 -7.70134330e-01 -1.45219296e-01
-4.65406567e-01 -7.33628452e-01 1.39386758e-01 2.14390695e-01
-9.51499343e-01 9.03478086e-01 1.67942792e-01 -9.46584105e-01
-4.66989130e-02 -7.25937724e-01 -6.42246842e-01 -1.01011884e+00
2.26793483e-01 4.77546543e-01 -3.36000741e-01 6.18802667e-01
5.74131310e-01 2.59286106e-01 -7.54742324e-01 -2.92535633e-01
6.63894415e-01 6.22944832e-01 8.22296441e-02 3.17661345e-01
4.81076032e-01 7.56248355e-01 -5.90485334e-01 -2.15266004e-01
2.58750528e-01 -5.06848991e-01 -3.76059592e-01 -4.42011505e-02
-5.86679399e-01 -5.01986265e-01 2.67717332e-01 1.42530596e+00
5.02670050e-01 -6.56624287e-02 6.39983892e-01 -9.33424830e-01
-3.47838193e-01 6.42127991e-01 -5.80602586e-01 7.44802803e-02
1.25425607e-01 7.32760429e-01 9.53192472e-01 -1.01644313e-02
6.82740033e-01 1.21144879e+00 4.52000760e-02 2.39251837e-01
-1.53864205e+00 -6.88219011e-01 3.51001441e-01 -2.02887416e-01
-1.21544397e+00 -4.87382054e-01 6.01180553e-01 -6.88965321e-01
7.88239986e-02 2.33461052e-01 5.20439446e-01 1.24164319e+00
7.02236950e-01 1.00352671e-02 1.53012311e+00 -3.43484730e-01
1.01093400e+00 1.61205575e-01 -5.82008623e-02 1.25294328e-01
9.07264650e-01 3.12319279e-01 -2.69700259e-01 -6.59949601e-01
7.55336702e-01 -4.36458364e-02 -3.41521293e-01 -2.94284195e-01
-1.10020530e+00 1.00957859e+00 -2.12508991e-01 1.17907025e-01
-1.03691733e+00 1.15037866e-01 3.48873116e-04 2.05989107e-01
-1.05867103e-01 7.26763666e-01 -4.38211232e-01 -3.77223939e-02
-4.34193164e-01 2.94571340e-01 7.39026785e-01 4.93290015e-02
2.16473192e-01 -4.00103182e-01 -3.10725838e-01 -2.51253042e-02
6.54607654e-01 4.43740457e-01 2.66830772e-01 -1.30240083e+00
2.00116575e-01 1.92680806e-01 8.33094835e-01 -7.13895082e-01
-6.50492311e-02 -2.47103870e-01 -4.77082312e-01 4.14263934e-01
5.38949966e-01 -7.17722893e-01 -5.30125737e-01 1.95428419e+00
4.89398092e-01 -6.18279949e-02 5.14500178e-02 1.10231793e+00
-4.00063470e-02 4.53539938e-01 2.78264821e-01 -6.53934956e-01
1.61228049e+00 -4.25621644e-02 -5.82039058e-01 -4.23287787e-02
4.16127563e-01 -6.10828280e-01 4.72773254e-01 5.46743274e-01
-1.14472711e+00 6.23605773e-02 -7.88564324e-01 3.95934552e-01
2.28359461e-01 -1.41539127e-01 4.15019155e-01 6.96281791e-01
-4.59727168e-01 4.96503800e-01 -8.26789737e-01 4.15721349e-02
5.66613734e-01 3.87587845e-01 -5.03836460e-02 1.51998892e-01
-8.43984604e-01 5.51502883e-01 -3.73053662e-02 2.80749321e-01
-1.06821525e+00 -5.71778715e-01 -2.34835029e-01 3.32688928e-01
9.23946619e-01 -1.00867963e+00 1.06142867e+00 -7.41020799e-01
-1.38703501e+00 6.75188184e-01 -2.76434943e-02 -2.31622651e-01
6.33588970e-01 3.77944112e-01 3.50470133e-02 -1.33680582e-01
6.78248644e-01 1.73785184e-02 4.65684861e-01 -6.72526598e-01
-6.58380806e-01 -8.52819085e-01 6.69053718e-02 1.31983489e-01
2.32777357e-01 1.98528126e-01 1.16365172e-01 -4.43115801e-01
4.89090644e-02 -1.23784280e+00 -5.83876908e-01 -4.25997049e-01
-6.98316097e-01 -1.97118700e-01 2.59351194e-01 -1.14584856e-01
6.20825410e-01 -2.18725228e+00 -1.04687154e-01 3.87548655e-01
2.40903825e-01 -3.74242008e-01 1.13431193e-01 2.46018276e-01
1.24132723e-01 5.23100495e-01 6.23534918e-02 1.92232758e-01
-5.79940993e-03 -2.84641147e-01 -1.71735257e-01 9.03564930e-01
-4.18091327e-01 5.74430525e-01 -5.95137000e-01 -2.00144183e-02
-2.35381991e-01 1.35400100e-02 -5.09486556e-01 1.96937844e-01
-1.80942371e-01 6.67818189e-01 -1.22488701e+00 2.76087731e-01
5.52252054e-01 -6.03492200e-01 8.55870485e-01 1.53077275e-01
-3.12996060e-01 4.98095393e-01 -1.25685573e+00 7.92682707e-01
4.15872820e-02 3.47176254e-01 5.96591115e-01 -1.16689444e+00
5.42338669e-01 4.37393397e-01 4.32132632e-01 -1.52497619e-01
4.82737929e-01 3.11393291e-01 5.31110406e-01 -5.33348858e-01
-1.58715263e-01 -6.15405381e-01 2.57475406e-01 8.87858570e-01
-6.11020982e-01 2.82840431e-01 -3.13067019e-01 -1.00224532e-01
1.35647988e+00 -2.79686421e-01 3.70076597e-01 -3.82350594e-01
8.47441554e-02 6.26112223e-02 1.13198090e+00 1.15401351e+00
-2.30123118e-01 -9.24511254e-03 1.20557261e+00 -9.01916698e-02
-9.00602639e-01 -5.87352097e-01 -2.56176531e-01 6.20011747e-01
-2.17480928e-01 2.42692173e-01 -4.17004734e-01 -4.31207329e-01
3.74166578e-01 8.56900275e-01 -8.34040523e-01 -1.18037499e-02
-1.07862093e-01 -6.85532629e-01 -3.60901542e-02 1.54504374e-01
3.65309030e-01 -6.89252019e-01 -9.93667245e-01 2.66144931e-01
1.32641315e-01 -7.53056407e-01 -4.80338246e-01 1.47323027e-01
-7.67918348e-01 -1.14424813e+00 -5.69446623e-01 -1.16845004e-01
8.43052447e-01 1.43197596e-01 3.61648500e-01 -3.94671142e-01
8.75461474e-02 2.47149676e-01 8.51684809e-02 -6.00492537e-01
-4.72103626e-01 -6.07352853e-01 -1.61265731e-02 8.56286958e-02
3.07466835e-01 -2.21179947e-01 -8.17333221e-01 2.75030762e-01
-7.26766586e-01 -3.01312596e-01 4.84288275e-01 5.53323507e-01
4.14087921e-01 4.00342762e-01 4.63097930e-01 -1.01032662e+00
7.68780708e-01 -7.27023423e-01 -1.24022293e+00 2.24968180e-01
-7.35956311e-01 1.26131251e-01 1.98635474e-01 -6.85666442e-01
-9.71333504e-01 -2.76405513e-01 6.81738853e-01 -1.76286608e-01
-2.72048771e-01 9.79645252e-01 -6.20549738e-01 2.65114158e-01
4.12640065e-01 -1.36282071e-01 4.36507940e-01 -5.04100978e-01
-1.36815816e-01 5.79447746e-01 1.29098082e-02 -7.72423625e-01
2.16100574e-01 3.01013142e-01 1.78125650e-01 -5.58253229e-01
-5.26820183e-01 1.76720947e-01 2.22013444e-01 1.23673521e-01
9.77563083e-01 -1.03132093e+00 -1.31704760e+00 2.50265896e-01
-1.08112395e+00 -3.30669403e-01 -8.35776404e-02 1.21774912e+00
-3.87554049e-01 -5.22514582e-02 -4.61113639e-02 -8.59032154e-01
2.36505151e-01 -1.61674953e+00 3.29031706e-01 3.02932024e-01
-2.49994740e-01 -7.24246144e-01 4.70062569e-02 7.72385120e-01
3.79645497e-01 2.64970571e-01 8.92490268e-01 -8.76375675e-01
-9.34338331e-01 -1.87418461e-01 2.45431185e-01 -1.06778070e-01
3.44901979e-01 6.66866526e-02 -4.80939090e-01 -2.78159231e-01
6.06392801e-01 -1.01777770e-01 2.01004311e-01 1.21809995e+00
1.00149989e+00 -8.73535156e-01 -4.05210078e-01 1.33377954e-01
8.56475234e-01 1.01645994e+00 2.51131922e-01 3.91130984e-01
5.08030541e-02 8.15838456e-01 2.15859547e-01 7.39682972e-01
3.18743855e-01 7.31445253e-01 8.30399692e-02 2.97253788e-01
6.44271433e-01 -1.72334507e-01 2.14469954e-01 -2.33459204e-01
-4.47063670e-02 -1.52196124e-01 -4.66004938e-01 1.78123295e-01
-1.75519669e+00 -8.01154137e-01 2.40486681e-01 2.72969532e+00
1.02624083e+00 2.84684777e-01 2.17329025e-01 -3.73613119e-01
9.03498769e-01 -2.09578186e-01 -9.43859458e-01 -3.46219689e-01
1.06972389e-01 -3.36475253e-01 8.41247499e-01 5.48966527e-01
-2.63637990e-01 4.63983387e-01 5.64545059e+00 4.54293072e-01
-1.16748130e+00 -1.42938584e-01 1.26409435e+00 -2.62723148e-01
-5.83051682e-01 5.22397518e-01 -6.88203156e-01 6.20591164e-01
1.07636130e+00 -7.70033479e-01 3.19388419e-01 3.50846767e-01
9.18558657e-01 -7.04562783e-01 -1.21043491e+00 3.37681502e-01
-5.47506213e-01 -1.36350691e+00 -6.00905895e-01 9.31789935e-01
4.52010304e-01 -3.02114010e-01 1.32103324e-01 -4.01268870e-01
1.03254461e+00 -1.07409477e+00 7.02527881e-01 4.82001066e-01
2.95791328e-01 -6.44991398e-01 8.04576635e-01 6.44115984e-01
-1.29171968e-01 -9.02788118e-02 -8.10738280e-02 -3.71895015e-01
-1.14850784e-02 8.51382971e-01 -9.26813602e-01 -1.04961991e-01
4.31327283e-01 -3.16793770e-02 2.45413765e-01 7.67850757e-01
-4.17681873e-01 1.03007209e+00 -2.09171936e-01 -1.98085785e-01
1.96102053e-01 -6.80567205e-01 5.35878301e-01 3.23175400e-01
1.90174833e-01 6.30699933e-01 1.00396745e-01 1.27718973e+00
-9.48218852e-02 1.42980918e-01 -5.38654983e-01 -7.16889441e-01
7.20779300e-01 8.22325289e-01 -9.19464648e-01 -2.01859176e-01
-1.87008783e-01 2.03038827e-01 -2.65202671e-01 6.25558615e-01
-2.51471817e-01 -7.33436598e-03 6.37468815e-01 6.73361644e-02
2.71328300e-01 2.69733787e-01 -5.47313869e-01 -7.07022607e-01
5.62478974e-02 -1.15522265e+00 3.21266234e-01 -6.40676677e-01
-8.32344413e-01 -3.21640581e-01 4.48954478e-02 -4.92273808e-01
-2.26555839e-01 -3.41157824e-01 -5.98279238e-01 1.28868771e+00
-9.70156550e-01 -1.94855735e-01 6.12571597e-01 1.42528817e-01
-1.79969817e-01 -2.07764715e-01 4.04874533e-01 -6.24744475e-01
-9.04090345e-01 1.28809199e-01 1.85675204e-01 1.80033579e-01
4.01978791e-01 -8.06234837e-01 -3.91718715e-01 5.04569650e-01
-4.52368319e-01 9.96504247e-01 9.37467396e-01 -9.97262955e-01
-1.66045618e+00 -3.62732798e-01 7.02352583e-01 -2.02152524e-02
9.71634150e-01 -1.18448824e-01 -7.04285860e-01 7.52449512e-01
-1.36825755e-01 -2.55642503e-01 8.28773677e-01 2.83318698e-01
-4.10481393e-02 -1.79627508e-01 -1.28241980e+00 1.00072992e+00
1.55761570e-01 -9.36679095e-02 -3.72982532e-01 3.92257184e-01
5.33979595e-01 -1.03840409e-02 -9.37087655e-01 1.74656853e-01
6.32757187e-01 -5.90235114e-01 5.15439510e-01 -9.15275633e-01
2.53031284e-01 -3.42423111e-01 -1.37331128e-01 -1.23315930e+00
-3.17391664e-01 -9.21558619e-01 7.38648713e-01 8.73743415e-01
5.13457477e-01 -1.05510473e+00 9.46331739e-01 1.32816005e+00
5.90424180e-01 -5.93102098e-01 -1.22998774e+00 -3.71104032e-01
3.38501036e-01 4.57423478e-02 5.15034676e-01 8.53602052e-01
7.50631467e-02 3.86023074e-01 -1.60487384e-01 4.17149395e-01
8.52912545e-01 3.59687984e-01 8.02473247e-01 -1.09701538e+00
-4.35204864e-01 -4.15272385e-01 2.47906037e-02 -6.67683303e-01
-1.25372997e-02 -1.82947442e-01 -5.06728999e-02 -9.80913758e-01
5.63959777e-01 -5.76327920e-01 -1.80017233e-01 4.69767064e-01
-2.92279035e-01 -9.10676539e-01 1.00879498e-01 2.91678071e-01
-6.72669187e-02 1.81395426e-01 1.38250542e+00 7.25249201e-02
-4.93703187e-01 3.54338557e-01 -1.40984094e+00 6.06704116e-01
5.96414745e-01 -7.55270302e-01 -3.63195747e-01 7.93280378e-02
3.87602806e-01 8.63866448e-01 4.41396117e-01 -1.23155555e-02
5.27620196e-01 -1.00167596e+00 1.45375878e-01 -1.76023707e-01
-2.79938221e-01 -7.56098092e-01 6.66814387e-01 8.55389595e-01
-6.70810401e-01 -1.57539427e-01 -1.79309800e-01 5.35801649e-01
2.55404919e-01 -5.43773651e-01 6.02093339e-01 -1.29481986e-01
5.17436266e-01 -1.07412338e-01 -6.11889184e-01 6.58002645e-02
1.01849484e+00 2.39103228e-01 -5.82642257e-01 -4.72769320e-01
-7.05311894e-01 3.04892302e-01 6.58546269e-01 -1.28977418e-01
2.29803458e-01 -7.34200954e-01 -9.61284101e-01 -2.10187569e-01
-3.24811846e-01 -2.63824135e-01 7.69995898e-02 1.09345162e+00
9.66791734e-02 4.79999661e-01 2.85198390e-01 -1.21550873e-01
-1.00046062e+00 5.37219584e-01 3.80854189e-01 -1.15328789e-01
-4.31218863e-01 3.51742417e-01 5.13352334e-01 1.86064258e-01
5.67713045e-02 -1.60178557e-01 5.03998669e-03 1.05380349e-01
5.72053194e-01 3.89377266e-01 -1.56760171e-01 -1.80336729e-01
-4.14439023e-01 -2.14035790e-02 -2.42320791e-01 -5.90734482e-01
1.37073600e+00 3.38385142e-02 -3.27633649e-01 3.80871981e-01
9.26228344e-01 3.98435533e-01 -1.30447292e+00 -1.55166686e-01
-2.49126598e-01 -5.91795743e-01 4.44477350e-01 -9.27959621e-01
-1.09711587e+00 9.72269550e-02 1.47943661e-01 1.95314616e-01
5.15399277e-01 -1.34109676e-01 -3.07590097e-01 1.36101454e-01
4.27935362e-01 -1.04649198e+00 -5.87084554e-02 -3.22714627e-01
5.93878448e-01 -1.16441834e+00 3.55823189e-01 4.49690670e-02
-7.11447299e-01 5.10238469e-01 -6.66275844e-02 1.91466004e-01
6.70095921e-01 8.35802704e-02 -6.59530163e-02 -2.57196218e-01
-9.85883772e-01 3.59695524e-01 3.00858300e-02 2.67937362e-01
1.44456789e-01 5.03992200e-01 -8.71839046e-01 6.72117233e-01
5.49358241e-02 4.40693200e-01 8.45162392e-01 8.58684301e-01
-3.92777234e-01 -1.20198929e+00 -9.53101158e-01 5.71632743e-01
-8.82456541e-01 3.08107119e-02 -2.55925983e-01 5.36683977e-01
-3.83568346e-03 1.17932248e+00 4.06226888e-02 3.11132103e-01
9.21002626e-02 -2.21656367e-01 -2.03490797e-02 -6.03116632e-01
-1.46679401e-01 7.58340776e-01 1.42681105e-02 -3.98023635e-01
-3.57299179e-01 -1.19391990e+00 -9.69254673e-01 -3.05295795e-01
-1.76223248e-01 6.33383989e-01 7.12367952e-01 9.39734757e-01
4.25905496e-01 6.07078187e-02 1.06011188e+00 -7.91517794e-02
-9.78391290e-01 -5.82115293e-01 -9.28242743e-01 -5.28265312e-02
2.35364348e-01 -5.04653573e-01 -8.32535326e-01 -3.60425591e-01] | [7.965751647949219, 5.249485969543457] |
f2d71fee-1010-416f-bc16-72ebb6750845 | words-as-gatekeepers-measuring-discipline | 2212.09676 | null | https://arxiv.org/abs/2212.09676v2 | https://arxiv.org/pdf/2212.09676v2.pdf | Words as Gatekeepers: Measuring Discipline-specific Terms and Meanings in Scholarly Publications | Scholarly text is often laden with jargon, or specialized language that can facilitate efficient in-group communication within fields but hinder understanding for out-groups. In this work, we develop and validate an interpretable approach for measuring scholarly jargon from text. Expanding the scope of prior work which focuses on word types, we use word sense induction to also identify words that are widespread but overloaded with different meanings across fields. We then estimate the prevalence of these discipline-specific words and senses across hundreds of subfields, and show that word senses provide a complementary, yet unique view of jargon alongside word types. We demonstrate the utility of our metrics for science of science and computational sociolinguistics by highlighting two key social implications. First, though most fields reduce their use of jargon when writing for general-purpose venues, and some fields (e.g., biological sciences) do so less than others. Second, the direction of correlation between jargon and citation rates varies among fields, but jargon is nearly always negatively correlated with interdisciplinary impact. Broadly, our findings suggest that though multidisciplinary venues intend to cater to more general audiences, some fields' writing norms may act as barriers rather than bridges, and thus impede the dispersion of scholarly ideas. | ['Katherine A. Keith', 'David Bamman', 'Jesse Dodge', 'Li Lucy'] | 2022-12-19 | null | null | null | null | ['word-sense-induction'] | ['natural-language-processing'] | [-1.55640200e-01 5.42731099e-02 -5.51344633e-01 2.78807670e-01
-4.51465547e-01 -9.06195700e-01 7.06201375e-01 7.69088507e-01
-4.53455538e-01 5.62208652e-01 1.12437832e+00 -8.18232477e-01
-6.58832431e-01 -8.81338656e-01 -4.43694651e-01 -2.13236108e-01
6.53138041e-01 -1.02770083e-01 -2.98048198e-01 -2.94817537e-01
1.11418128e+00 2.58911878e-01 -1.17181945e+00 -1.60599813e-01
1.09710360e+00 8.53209123e-02 2.67322719e-01 2.37791270e-01
-7.69605339e-01 4.74055141e-01 -1.00552380e+00 -3.03634107e-01
-2.09410846e-01 -3.74662459e-01 -6.85414612e-01 -1.54851541e-01
6.49584770e-01 1.82857081e-01 -2.37046555e-01 8.22531879e-01
4.00134653e-01 7.81495720e-02 8.30837727e-01 -7.03587294e-01
-1.11450338e+00 1.03747034e+00 -6.04714215e-01 5.77094078e-01
4.90500420e-01 6.96478263e-02 1.28073680e+00 -4.03729647e-01
1.04804146e+00 1.44868100e+00 6.53497756e-01 -1.35409504e-01
-1.42417848e+00 -9.73421097e-01 1.63827121e-01 -3.53039652e-01
-1.17878115e+00 -4.57047433e-01 4.70045567e-01 -1.08233047e+00
4.77331638e-01 3.38140339e-01 9.13722456e-01 1.09780598e+00
3.01987559e-01 -2.22550675e-01 1.26306748e+00 -4.36326563e-01
1.00444853e-01 4.76209596e-02 6.21387780e-01 1.57887965e-01
1.08316231e+00 -6.87387884e-01 -9.17734206e-01 -5.63415825e-01
5.93105614e-01 2.02671856e-01 -3.20585102e-01 6.90533280e-01
-1.77393949e+00 6.64232612e-01 -8.78371950e-03 8.49873543e-01
-3.08462620e-01 -3.99125041e-03 2.43598193e-01 2.66791344e-01
6.13357008e-01 1.10365379e+00 1.02421045e-01 -7.62355924e-01
-1.05663729e+00 1.92811266e-01 1.08894730e+00 5.59314013e-01
5.57180882e-01 -4.74077165e-01 -3.20857376e-01 1.10764694e+00
2.32304156e-01 3.31811726e-01 3.44290197e-01 -9.15449142e-01
1.65269539e-01 8.88953030e-01 -3.97734530e-02 -1.47500253e+00
-1.82839155e-01 -6.20005667e-01 -4.53921825e-01 -2.06128985e-01
5.62599123e-01 -1.32408142e-01 -2.92304575e-01 1.58103883e+00
-2.78152943e-01 -1.60329819e-01 -5.20082235e-01 6.83127165e-01
8.36415350e-01 4.51920927e-01 4.89860296e-01 -1.55903220e-01
1.56346238e+00 -1.30367011e-01 -7.98269212e-01 -5.26911199e-01
5.42146683e-01 -1.31913507e+00 1.31198430e+00 2.13575080e-01
-9.55148757e-01 -2.17200086e-01 -9.06453371e-01 -2.81781584e-01
-6.08161211e-01 -4.07137722e-01 4.61393416e-01 6.30112410e-01
-1.17004871e+00 6.43746197e-01 -3.57263029e-01 -7.33997524e-01
6.28023267e-01 -4.59297538e-01 -1.33375168e-01 2.67063081e-02
-7.80904830e-01 8.97457600e-01 -9.42820832e-02 -6.78968430e-01
1.62015423e-01 -1.18728578e+00 -3.58581960e-01 -7.74550438e-02
1.15586378e-01 -6.25093699e-01 6.12399876e-01 -4.66259569e-01
-7.40740657e-01 1.07387412e+00 -2.71430552e-01 3.88600051e-01
1.31911516e-01 -1.94488198e-01 -6.02944553e-01 -5.45929708e-02
8.82702410e-01 1.71850935e-01 1.57498986e-01 -1.16166437e+00
-3.04406255e-01 -3.78730476e-01 1.24595314e-01 -1.43512279e-01
-1.04766595e+00 3.20205241e-02 -1.48604885e-01 -9.03714418e-01
1.17109105e-01 -4.61344332e-01 2.58060582e-02 -6.05272688e-02
-4.31167364e-01 -5.63457787e-01 5.53411007e-01 -4.73272771e-01
1.69607472e+00 -2.10204124e+00 -9.89194214e-02 3.61384839e-01
8.51338446e-01 -5.95690727e-01 1.05417341e-01 9.39096212e-01
2.28948012e-01 9.23154116e-01 2.60923803e-01 9.36725885e-02
5.38965724e-02 1.02116920e-01 -2.91519850e-01 6.32440686e-01
-2.16559187e-01 6.72813714e-01 -1.27844429e+00 -5.39394915e-01
-2.03828171e-01 1.89773232e-01 -1.97602972e-01 -4.77573514e-01
4.05465811e-02 1.10342294e-01 -4.27159935e-01 8.33851635e-01
5.23999333e-01 -7.27945149e-01 2.48348922e-01 2.25175276e-01
-9.19169962e-01 7.41198957e-01 -8.40638936e-01 1.34242332e+00
-5.68958580e-01 1.27787960e+00 1.79322138e-02 -7.70070851e-01
8.35784614e-01 -1.75719723e-01 3.37013483e-01 -3.82927150e-01
2.93186028e-02 2.78276712e-01 2.93571591e-01 -4.08778369e-01
9.29390013e-01 -1.45636588e-01 -1.92935318e-02 9.38325524e-01
-2.75250733e-01 -2.59544522e-01 3.32225233e-01 4.61192101e-01
1.13595021e+00 -3.64030033e-01 2.66246915e-01 -1.02376556e+00
-2.44411767e-01 6.94140270e-02 3.02693516e-01 1.03178608e+00
5.87407500e-02 2.94804662e-01 7.94922292e-01 2.58358628e-01
-1.01000082e+00 -7.39597976e-01 -5.28408766e-01 1.25362384e+00
4.12335992e-01 -7.62572885e-01 -1.76324233e-01 1.67940900e-01
4.06396747e-01 8.62521112e-01 -4.03789073e-01 1.57482550e-01
1.19472921e-01 -4.84071314e-01 5.24213195e-01 1.37545029e-02
1.31334677e-01 -5.15222967e-01 -4.15257454e-01 1.54951021e-01
-2.47188762e-01 -9.12942111e-01 -4.82757837e-01 -1.94343001e-01
-6.17974579e-01 -9.92505670e-01 -8.44109833e-01 -4.62537736e-01
5.03836453e-01 8.23804617e-01 1.38122129e+00 4.41219777e-01
-2.68943906e-01 5.36292136e-01 -3.74334574e-01 -8.77900541e-01
-3.88910592e-01 3.25292051e-01 -1.53272137e-01 -6.46830976e-01
4.87039238e-01 -6.09980762e-01 -4.32607412e-01 7.21528977e-02
-1.02560067e+00 -1.85586795e-01 3.04576755e-01 5.30237138e-01
6.85386658e-02 -5.64269662e-01 7.23246217e-01 -9.80499923e-01
1.12290609e+00 -1.10018444e+00 8.70158598e-02 5.29800095e-02
-8.62754822e-01 -3.55463296e-01 3.07360530e-01 -4.74584013e-01
-6.82715356e-01 -1.38714123e+00 2.68967956e-01 3.69221359e-01
1.27683297e-01 8.49045753e-01 4.59665388e-01 -1.87408715e-01
1.02736747e+00 -3.35707903e-01 1.61321163e-01 -3.95414352e-01
-2.04092283e-02 1.04207480e+00 -9.00189206e-03 -9.70424175e-01
7.79960036e-01 5.24386048e-01 -2.08504617e-01 -1.29367566e+00
-9.15300727e-01 -6.90792024e-01 8.21867436e-02 -4.18107748e-01
6.99055731e-01 -9.30199921e-01 -1.08218288e+00 -6.42975494e-02
-9.03568268e-01 -9.00757089e-02 -5.00710718e-02 7.75171876e-01
2.25604624e-01 4.85250831e-01 -4.92617279e-01 -6.85143709e-01
-5.63559774e-03 -8.27620506e-01 1.01773083e+00 4.38464135e-01
-1.08153701e+00 -1.21679342e+00 7.12640435e-02 6.50780737e-01
3.95600736e-01 3.86278719e-01 1.11449182e+00 -4.47324574e-01
-1.84217468e-01 2.00518042e-01 -3.36423725e-01 -1.97320923e-01
4.15909290e-01 1.63619787e-01 -5.37520647e-01 1.73251778e-01
-4.34679955e-01 -9.65894386e-02 6.67387903e-01 2.72036225e-01
1.24792981e+00 -2.58600712e-01 -6.85865819e-01 1.75283626e-01
1.33110917e+00 -1.32545549e-02 3.03396881e-01 8.88046622e-01
5.87457240e-01 6.04687810e-01 2.04572454e-01 6.63153350e-01
3.89469355e-01 6.41815141e-02 -3.15593660e-01 8.43789577e-02
2.77053356e-01 -9.64823812e-02 1.54767722e-01 1.07170463e+00
-3.87731731e-01 -3.50171179e-01 -1.37618244e+00 8.77977252e-01
-1.34676552e+00 -1.00590980e+00 -4.97776598e-01 1.83869469e+00
1.04902422e+00 3.36374521e-01 2.91354209e-03 -1.77126944e-01
5.34559071e-01 3.77636641e-01 -1.32646769e-01 -4.26234871e-01
-3.10001284e-01 3.48800272e-01 7.80384004e-01 4.18431550e-01
-8.01798552e-02 7.24202096e-01 6.87761736e+00 8.18641543e-01
-9.98342335e-01 -1.12807617e-01 7.34035969e-01 -3.69761586e-02
-1.21041632e+00 2.17601046e-01 -5.16092420e-01 6.43438578e-01
6.02035880e-01 -7.29894102e-01 1.36547759e-01 4.07547295e-01
5.66063464e-01 -2.30615333e-01 -9.16499972e-01 7.99031019e-01
-4.10671756e-02 -1.69600654e+00 -4.20638397e-02 3.42251867e-01
7.52496362e-01 -2.18589827e-01 -3.15405317e-02 -1.43273205e-01
4.05047417e-01 -1.18981647e+00 7.10501969e-01 4.86732543e-01
8.47808719e-01 -3.43698710e-01 3.34429234e-01 3.67181376e-02
-4.25913066e-01 2.63405114e-01 -4.01009977e-01 -8.35945845e-01
-8.99033155e-03 1.20140469e+00 -3.59722108e-01 2.33512521e-01
7.14619875e-01 8.48671257e-01 -3.76092851e-01 7.43203580e-01
1.18552878e-01 1.04979217e+00 -1.13011532e-01 -4.35095906e-01
2.09795475e-01 -5.62368214e-01 8.28563750e-01 1.42084503e+00
6.16562903e-01 1.49779320e-01 -2.05426335e-01 1.02492285e+00
-2.62302667e-01 1.09288424e-01 -7.72544622e-01 -1.06312215e+00
1.16120946e+00 1.22336817e+00 -1.05628526e+00 -2.62027055e-01
-5.00067174e-01 3.29574853e-01 1.55181423e-01 3.92812163e-01
-1.97208434e-01 -5.71930408e-01 1.06375325e+00 6.50262773e-01
-5.15802920e-01 -5.79960823e-01 -1.25747395e+00 -8.89303088e-01
-2.63335913e-01 -7.78021276e-01 1.26362234e-01 -4.31797236e-01
-1.61068201e+00 -2.41499081e-01 -2.33677834e-01 -8.30957592e-01
4.05854762e-01 -4.89029348e-01 -5.79197645e-01 1.02009010e+00
-1.06292319e+00 -6.91365898e-01 -4.34057176e-01 1.33562461e-03
2.09969610e-01 2.95934118e-02 5.91285169e-01 1.61758531e-02
-1.76482692e-01 5.25338709e-01 3.80009025e-01 -1.70458943e-01
1.13868582e+00 -9.92168903e-01 2.10467070e-01 3.05984408e-01
-3.75342704e-02 1.36619079e+00 8.36496294e-01 -9.30097640e-01
-1.49259114e+00 -4.83609825e-01 1.14940727e+00 -6.32393956e-01
1.26773238e+00 -1.84626803e-01 -8.01357090e-01 5.74278116e-01
4.62653220e-01 -9.78516698e-01 1.34662271e+00 9.18979526e-01
-3.62895399e-01 3.31292570e-01 -7.42039263e-01 1.20311713e+00
1.56639588e+00 -6.14427269e-01 -7.62332320e-01 5.67652643e-01
7.39654303e-01 2.59699598e-02 -1.39359069e+00 -4.71976578e-01
6.17434382e-01 -6.28383100e-01 9.17166471e-01 -2.45323688e-01
1.01978028e+00 1.53552890e-01 2.82218933e-01 -1.30005550e+00
-8.24056208e-01 -8.42296600e-01 5.86915553e-01 1.58187461e+00
3.53556246e-01 -8.73504281e-01 3.29281121e-01 6.37485027e-01
-3.65502447e-01 -2.70328552e-01 -6.59837246e-01 -4.62625504e-01
6.12376869e-01 -4.10719097e-01 3.74154150e-01 1.75788164e+00
4.71287102e-01 3.81454825e-01 2.84786791e-01 -4.34439510e-01
5.11201441e-01 6.54995218e-02 5.02629757e-01 -1.68199050e+00
4.93535772e-02 -1.02942884e+00 -3.59062940e-01 -6.67213142e-01
1.65955514e-01 -9.34848547e-01 -4.75461364e-01 -1.75154614e+00
2.91839302e-01 -5.57996452e-01 1.23978190e-01 2.55003750e-01
-2.59070456e-01 7.23246932e-02 2.13225201e-01 3.69477272e-01
-1.13944754e-01 -2.63749778e-01 1.52752185e+00 -1.32943854e-01
-4.20423120e-01 -6.62399352e-01 -1.85441351e+00 5.17738998e-01
5.82059324e-01 -1.74322769e-01 -1.15888700e-01 -4.22941059e-01
1.00400007e+00 -4.27895814e-01 5.31449914e-01 -5.79102635e-01
1.55713096e-01 -8.93079877e-01 4.36076581e-01 1.55227939e-02
-2.77508825e-01 -2.81100720e-01 1.19438156e-01 1.52782977e-01
-4.38175142e-01 -1.36995651e-02 2.29719952e-01 3.01418245e-01
6.38304129e-02 -1.46057636e-01 2.75698572e-01 -5.66339754e-02
-7.94210508e-02 -3.07960510e-01 -8.14139485e-01 5.45731187e-01
5.29334486e-01 -6.86227500e-01 -9.04184759e-01 -4.52887207e-01
-9.50253829e-02 6.34315014e-02 9.07966435e-01 5.65514624e-01
1.01248227e-01 -9.09947336e-01 -8.57912183e-01 -5.81096888e-01
2.82821596e-01 -3.05105507e-01 5.30998409e-02 7.84675777e-01
-5.35959184e-01 1.23076186e-01 -1.21120930e-01 -1.55420020e-01
-8.12964261e-01 -1.47914350e-01 -3.58827770e-01 3.69647950e-01
-6.57182932e-01 7.83136904e-01 2.07323238e-01 2.06246808e-01
-1.18727602e-01 -2.21103817e-01 -3.20318818e-01 6.83560133e-01
4.65848416e-01 8.38337243e-01 -4.57674414e-01 -6.13697231e-01
-3.54927719e-01 5.31134963e-01 2.96034012e-02 -1.77647188e-01
1.35960245e+00 -8.12086742e-03 -6.06132805e-01 1.05237949e+00
1.02350521e+00 6.62842572e-01 -2.32508078e-01 8.61424804e-02
6.44008741e-02 -8.66312683e-01 1.09200470e-01 -6.25063539e-01
-4.13203418e-01 2.95086563e-01 -1.21884666e-01 6.86499119e-01
4.89879012e-01 3.28956068e-01 5.76376915e-01 9.45698172e-02
-1.08053430e-03 -1.45495594e+00 -3.11339777e-02 4.27795410e-01
7.75281549e-01 -7.91806817e-01 2.30539754e-01 -4.61691916e-01
-2.40629122e-01 1.19151962e+00 3.80597085e-01 3.28469574e-01
7.08259881e-01 2.68446863e-01 -2.47755721e-02 -5.25068402e-01
-4.09967452e-01 1.16508618e-01 1.37863159e-01 2.92992175e-01
1.21603918e+00 2.26254657e-01 -1.33400619e+00 5.44804394e-01
-7.23017395e-01 -8.96618702e-03 1.05460119e+00 6.83695138e-01
-6.02308631e-01 -8.88908684e-01 -7.86609471e-01 9.92074430e-01
-9.08783555e-01 -2.81245381e-01 -8.45960259e-01 6.58482969e-01
2.07761899e-01 1.20900428e+00 4.80182022e-01 -2.66531229e-01
-1.17018223e-01 -7.81763196e-02 2.42502894e-02 -7.30526209e-01
-7.72043526e-01 -3.93291451e-02 2.73704380e-01 5.91487251e-02
-4.49224293e-01 -6.81791544e-01 -1.00645697e+00 -1.00878048e+00
-1.49245977e-01 2.04166800e-01 5.07526100e-01 9.61646557e-01
5.26823997e-01 5.97718477e-01 8.15797225e-02 -2.55167216e-01
5.76394750e-03 -1.14261389e+00 -7.64945745e-01 3.72975975e-01
1.05350569e-01 -7.12320268e-01 -4.76564407e-01 2.56021563e-02] | [9.537113189697266, 8.298088073730469] |
b8ec5e75-cd4e-4452-b846-649b8e23e674 | infillmore-neural-frame-lexicalization-for | 2103.04941 | null | https://arxiv.org/abs/2103.04941v3 | https://arxiv.org/pdf/2103.04941v3.pdf | InFillmore: Frame-Guided Language Generation with Bidirectional Context | We propose a structured extension to bidirectional-context conditional language generation, or "infilling," inspired by Frame Semantic theory (Fillmore, 1976). Guidance is provided through two approaches: (1) model fine-tuning, conditioning directly on observed symbolic frames, and (2) a novel extension to disjunctive lexically constrained decoding that leverages frame semantic lexical units. Automatic and human evaluations confirm that frame-guided generation allows for explicit manipulation of intended infill semantics, with minimal loss in distinguishability from human-generated text. Our methods flexibly apply to a variety of use scenarios, and we provide a codebase and interactive demo available from https://nlp.jhu.edu/demos/infillmore. | ['Benjamin Van Durme', 'Felix Yu', 'Anton Belyy', 'Nathaniel Weir', 'Jiefu Ou'] | 2021-03-08 | null | https://aclanthology.org/2021.starsem-1.12 | https://aclanthology.org/2021.starsem-1.12.pdf | joint-conference-on-lexical-and-computational-1 | ['text-infilling'] | ['natural-language-processing'] | [ 5.85448205e-01 5.27094185e-01 -5.33006251e-01 -5.84356785e-01
-9.35038090e-01 -8.11819255e-01 1.24526942e+00 7.58772492e-02
-6.65689781e-02 1.19078290e+00 1.08849001e+00 -6.16267085e-01
1.72814488e-01 -9.57809031e-01 -5.51476121e-01 -9.21617076e-02
7.98246451e-03 2.30474830e-01 1.69316292e-01 -3.93332720e-01
2.63289243e-01 1.09340467e-01 -1.51663566e+00 6.14865839e-01
5.92381835e-01 5.43552876e-01 5.08664072e-01 6.18967354e-01
-2.41041735e-01 1.26129222e+00 -2.59684324e-01 -1.69115335e-01
-2.56090939e-01 -6.09175146e-01 -1.01106131e+00 -3.18424068e-02
-1.92459580e-02 -3.36234748e-01 -1.36469990e-01 7.05997229e-01
2.48093054e-01 2.86737740e-01 4.32989329e-01 -1.05163336e+00
-7.85976887e-01 1.14517629e+00 9.57498252e-02 2.57017642e-01
1.12003231e+00 1.95675030e-01 1.15070605e+00 -1.14425254e+00
9.36648726e-01 1.74614918e+00 5.54061294e-01 7.46457398e-01
-1.54348147e+00 -4.93107855e-01 4.35259461e-01 -1.37097955e-01
-1.34171832e+00 -8.90277684e-01 6.10808313e-01 -5.85197270e-01
1.30264866e+00 3.93818796e-01 7.55078971e-01 1.47029853e+00
-4.87536984e-03 8.19572628e-01 9.82718647e-01 -7.75009930e-01
3.97229791e-01 -4.46213365e-01 5.36564663e-02 5.42885423e-01
1.59210429e-01 3.57616216e-01 -7.25300610e-01 -5.04354060e-01
6.69607878e-01 -4.94742632e-01 -1.62239984e-01 1.78894415e-01
-1.40393209e+00 8.02251637e-01 -9.34066772e-02 2.93560594e-01
-1.99908651e-02 7.52595425e-01 1.48088902e-01 -3.20240892e-02
5.41356921e-01 1.96161643e-01 -2.51141906e-01 -4.06136721e-01
-1.05038488e+00 6.60262585e-01 7.24910915e-01 1.33503747e+00
7.21691370e-01 2.20147789e-01 -2.90577173e-01 4.08465236e-01
7.29835510e-01 7.60661185e-01 4.26187426e-01 -1.25870955e+00
4.19351876e-01 2.96219021e-01 4.72908229e-01 -7.71477580e-01
-2.65046120e-01 1.65444270e-01 -1.34412646e-01 -2.28313670e-01
1.87245250e-01 -3.29170138e-01 -8.93405914e-01 2.26517844e+00
1.17941603e-01 9.77991372e-02 -4.77692019e-03 5.75734973e-01
6.19973421e-01 5.79487860e-01 5.78393340e-01 -2.74258733e-01
1.20376468e+00 -4.52790618e-01 -7.64058352e-01 -6.47423744e-01
7.70443082e-01 -5.53906500e-01 1.35324156e+00 -2.86538035e-01
-1.21129107e+00 -1.21901132e-01 -8.99977028e-01 -3.15583616e-01
-2.57817596e-01 -1.79570556e-01 9.56869006e-01 5.71225703e-01
-1.14705455e+00 2.04825103e-01 -1.18978119e+00 -5.18398285e-01
3.00131142e-01 -1.05826324e-02 3.15360390e-02 1.44977212e-01
-1.48689878e+00 7.31283069e-01 5.89931428e-01 -8.89926404e-02
-1.11594105e+00 -2.96959728e-01 -1.23932791e+00 -1.48626953e-01
5.24871171e-01 -8.56691241e-01 1.71520352e+00 -1.22270918e+00
-1.36609936e+00 8.78651977e-01 -6.95240438e-01 -5.90424597e-01
4.00368959e-01 -4.11352843e-01 -3.85448515e-01 6.41279444e-02
5.49010634e-01 8.75041068e-01 6.55642569e-01 -1.23183942e+00
-6.24729335e-01 1.19370647e-01 5.22680521e-01 3.41724902e-01
1.24094784e-01 3.42601269e-01 -1.30123258e-01 -1.13497198e+00
-5.56518137e-02 -7.27512062e-01 -1.76214069e-01 -1.51120216e-01
-3.22829306e-01 -1.16676331e-01 5.13560772e-01 -4.87607211e-01
1.73837698e+00 -1.86658442e+00 1.39191404e-01 2.37579253e-02
1.60700604e-02 -3.95075351e-01 -8.43109563e-02 7.43470013e-01
-6.32125363e-02 5.63527286e-01 -4.58095193e-01 -2.36834601e-01
2.24412918e-01 3.54821920e-01 -8.86056006e-01 1.90009847e-01
2.00405002e-01 9.80194271e-01 -1.36355650e+00 -7.22069085e-01
2.12622344e-01 2.02045724e-01 -6.72155499e-01 -7.02268183e-02
-8.56486857e-01 3.67728353e-01 -4.44274098e-01 8.68700504e-01
5.60550299e-03 -2.37488151e-01 6.37832820e-01 2.05673993e-01
-2.49419108e-01 6.47990465e-01 -1.04788005e+00 1.87017143e+00
-2.83595473e-01 5.48284888e-01 -2.08198830e-01 -2.70285875e-01
5.21041751e-01 4.45867389e-01 -8.15725923e-02 -3.70492786e-01
-4.71772850e-02 3.53086106e-02 -4.61131662e-01 -4.12261367e-01
7.85897076e-01 -3.76381963e-01 -6.02455080e-01 5.72915614e-01
-4.25377022e-03 -2.25540861e-01 4.68964785e-01 5.13154089e-01
8.93285036e-01 8.28782022e-01 6.77695930e-01 -5.64485252e-01
2.73156673e-01 1.81267142e-01 7.43420243e-01 9.23704445e-01
2.48465072e-02 4.44463074e-01 5.47464669e-01 -2.77449518e-01
-8.94979358e-01 -1.16904342e+00 -1.68505803e-01 1.36433125e+00
2.11874187e-01 -9.75757599e-01 -7.43958235e-01 -3.86267453e-01
-2.51695842e-01 1.46446812e+00 -5.97208977e-01 1.05812147e-01
-8.74617636e-01 -5.13076186e-01 9.11316097e-01 8.01660120e-01
2.53953218e-01 -1.09271908e+00 -8.89648318e-01 2.51567483e-01
-6.02077186e-01 -9.24851358e-01 -3.12521994e-01 -3.71497422e-02
-7.76878953e-01 -8.84745955e-01 -1.57395035e-01 -4.55314487e-01
4.54065531e-01 1.53661054e-02 1.24603021e+00 3.13007116e-01
2.12676242e-01 6.46900654e-01 -4.89490658e-01 -2.08976671e-01
-5.78091323e-01 6.00688765e-03 1.06736738e-02 -4.24816638e-01
2.59710401e-01 -4.48555857e-01 -1.23252414e-01 1.91365823e-01
-9.09340024e-01 6.27675354e-01 -1.63015455e-01 7.45702505e-01
4.68203694e-01 -4.85438794e-01 3.25352877e-01 -9.34096813e-01
8.07624757e-01 -6.81650817e-01 -5.23906052e-01 3.32388580e-01
-3.60975653e-01 7.89290071e-02 2.31958091e-01 -1.48800626e-01
-1.54805017e+00 -1.17553532e-01 1.81140378e-02 6.06522262e-02
-3.27133983e-01 9.34393644e-01 -1.45528898e-01 6.34761751e-01
8.78578782e-01 7.65273198e-02 -4.76981461e-01 -7.59667754e-02
7.76800931e-01 3.34712416e-01 5.15792191e-01 -1.33480573e+00
6.67593002e-01 5.45474529e-01 -3.65090102e-01 -8.30987215e-01
-7.08953083e-01 2.98449397e-01 -4.48449999e-01 -3.42804551e-01
9.43854511e-01 -1.08756030e+00 -8.48541781e-02 -4.54912782e-02
-1.11134398e+00 -9.43935215e-01 -4.37945306e-01 3.28710794e-01
-8.35758686e-01 2.85863221e-01 -5.33387244e-01 -9.65921223e-01
5.85434176e-02 -8.86886001e-01 1.26344860e+00 -1.07914560e-01
-7.96553075e-01 -1.10181522e+00 -3.74080939e-03 2.43330002e-02
1.74244925e-01 3.75707865e-01 9.90945339e-01 -4.29248869e-01
-4.74070102e-01 2.87727088e-01 1.23468470e-02 -3.17292511e-01
-2.96090525e-02 1.71400040e-01 -8.72853339e-01 -7.70706981e-02
-2.34445617e-01 -4.05160218e-01 5.10994077e-01 2.46351838e-01
6.23303592e-01 -4.14587080e-01 -4.92844522e-01 3.61000061e-01
1.15588319e+00 2.37429470e-01 6.47118330e-01 3.67025614e-01
3.53886545e-01 2.34891191e-01 5.38612187e-01 6.10044181e-01
6.63599491e-01 5.46322644e-01 5.45499399e-02 3.11152518e-01
-3.70427728e-01 -8.88994932e-01 5.43038964e-01 4.26889181e-01
-2.17957616e-01 -5.59316456e-01 -1.34329379e+00 6.30756199e-01
-2.16976333e+00 -1.40268600e+00 7.70532861e-02 1.84101224e+00
9.59674418e-01 2.29132652e-01 -1.11058138e-01 -1.95931181e-01
9.03802574e-01 5.44232011e-01 -1.79547831e-01 -7.90244564e-02
-3.64157915e-01 1.51110455e-01 2.77319342e-01 1.18393612e+00
-9.67749953e-01 1.48558486e+00 7.37493086e+00 8.74018133e-01
-6.23674810e-01 3.38895649e-01 4.46415454e-01 -1.38766870e-01
-9.55916584e-01 7.55328476e-01 -8.11734796e-01 4.87107992e-01
9.39106584e-01 -6.36581242e-01 5.82426369e-01 5.77189982e-01
5.29093504e-01 -2.70345032e-01 -1.21664715e+00 6.20217264e-01
-1.11380890e-02 -1.69073832e+00 2.40315065e-01 -1.66596368e-01
5.25250554e-01 -4.79952209e-02 -3.26425850e-01 2.83417791e-01
8.63099754e-01 -8.61304045e-01 1.59869003e+00 5.79176426e-01
1.05835485e+00 -3.33423644e-01 6.50403500e-02 2.52794981e-01
-1.44642413e+00 -1.04381397e-01 -2.70968769e-02 -6.19420826e-01
8.06072712e-01 3.69737059e-01 -5.44804037e-01 1.97323546e-01
2.38492206e-01 7.28576183e-01 -4.72992539e-01 1.81641698e-01
-8.63729596e-01 9.76507902e-01 -3.89784276e-01 5.11641204e-02
1.71967685e-01 1.86961949e-01 9.32696283e-01 1.43711412e+00
2.81117916e-01 3.61510575e-01 3.73425335e-01 1.07770824e+00
3.56617197e-02 -1.83212027e-01 -7.75156975e-01 -1.62697017e-01
9.34741557e-01 6.16515636e-01 -8.73854876e-01 -5.07194877e-01
-2.99503386e-01 5.14347672e-01 4.71019596e-02 6.35593832e-01
-1.04015255e+00 5.75035438e-02 5.15783250e-01 2.66912013e-01
8.26647654e-02 -5.22986591e-01 -3.19456100e-01 -1.53746760e+00
-2.51802981e-01 -6.68567598e-01 3.23438078e-01 -1.12780964e+00
-7.94936955e-01 3.88072968e-01 7.30004251e-01 -8.30161035e-01
-8.16886127e-01 -2.16512203e-01 -4.23040628e-01 5.89564800e-01
-9.48935866e-01 -1.55830944e+00 -3.15860361e-02 4.84189272e-01
7.71936893e-01 7.36426637e-02 1.03143668e+00 -6.36051297e-02
-3.80419433e-01 2.76649922e-01 -4.03119534e-01 2.50187796e-02
3.32791746e-01 -1.15222025e+00 7.01727331e-01 1.13833475e+00
-7.40553066e-02 8.80759478e-01 8.30539703e-01 -1.11002004e+00
-1.20512152e+00 -1.04248655e+00 1.25642872e+00 -6.95818484e-01
9.09567773e-01 -6.42878175e-01 -4.22861308e-01 1.44467843e+00
3.38990569e-01 -3.66717398e-01 7.52032936e-01 1.21097215e-01
-2.58781165e-01 6.02184594e-01 -1.07498467e+00 1.06963980e+00
1.55449712e+00 -6.99929178e-01 -1.03052640e+00 4.41056252e-01
9.44394886e-01 -5.43160498e-01 -5.36309302e-01 1.32063270e-01
5.77280223e-01 -6.85666621e-01 7.95880675e-01 -5.61639488e-01
4.23283428e-01 -5.57823002e-01 -8.59938085e-01 -6.93169951e-01
-3.59626263e-01 -8.20520222e-01 -2.18680665e-01 1.20171475e+00
5.54486513e-01 -5.68806827e-01 3.04781675e-01 1.05773246e+00
-2.49323666e-01 -3.37420613e-01 -7.97723770e-01 -5.28083324e-01
-9.28503461e-03 -1.13325572e+00 7.80341744e-01 9.51647222e-01
2.50835747e-01 4.43174273e-01 -2.54997432e-01 6.90934062e-02
4.26977098e-01 9.57186967e-02 4.79776084e-01 -7.85008192e-01
-4.36350733e-01 -2.51458913e-01 -1.73952609e-01 -1.14835513e+00
4.36669290e-01 -1.11275339e+00 1.81630999e-02 -1.46790779e+00
-7.33216703e-02 -3.81457180e-01 2.82980353e-01 9.25670505e-01
1.01876080e-01 1.06824450e-01 4.45820957e-01 2.24039704e-01
-6.44846320e-01 4.82946903e-01 6.02263510e-01 9.16051269e-02
-2.93155372e-01 -5.92382729e-01 -9.57699835e-01 1.14034021e+00
6.86376870e-01 -4.86460626e-01 -6.76981688e-01 -8.20182621e-01
6.66596174e-01 2.88033485e-01 5.36277473e-01 -7.76391566e-01
-3.93841900e-02 -7.96949863e-01 2.04174101e-01 -2.50009567e-01
1.43964618e-01 -2.49289662e-01 5.16703427e-01 1.82286650e-01
-6.41647518e-01 1.15466468e-01 2.88007885e-01 5.09218335e-01
7.04624057e-02 -3.18953276e-01 2.48280302e-01 -3.65609109e-01
-9.49373543e-01 -2.78551042e-01 -1.00415969e+00 3.07332605e-01
8.65599155e-01 -2.77451307e-01 -5.02150476e-01 -4.96247649e-01
-1.08298731e+00 2.34369874e-01 7.03567088e-01 2.86208898e-01
5.06662667e-01 -1.48127878e+00 -4.51105028e-01 -1.07641080e-02
1.28757417e-01 -1.70577481e-01 -2.77116209e-01 4.10284251e-01
-4.94643152e-01 2.95883894e-01 3.19906443e-01 -4.13134187e-01
-8.35303366e-01 2.77262777e-01 7.54460543e-02 7.44955018e-02
-5.48511446e-01 6.53729618e-01 1.69668987e-01 -1.76112160e-01
-1.01372734e-01 -6.22819364e-01 2.79790282e-01 -3.93220149e-02
6.46714389e-01 2.95106649e-01 -4.31074739e-01 -7.71384776e-01
-4.13396895e-01 1.53069317e-01 3.07124525e-01 -8.16223264e-01
1.10644209e+00 -5.37019432e-01 -1.61331102e-01 6.56098306e-01
4.01835680e-01 1.46201283e-01 -1.13218534e+00 -6.43368885e-02
1.84429288e-01 -4.65081275e-01 -4.38897312e-02 -8.55609417e-01
-1.79509908e-01 3.41793776e-01 9.67062935e-02 2.44151995e-01
9.00385380e-01 1.22778855e-01 4.74874854e-01 5.68621397e-01
7.17266500e-01 -9.95638013e-01 -3.91192645e-01 7.06841290e-01
9.93586600e-01 -6.66696668e-01 1.69680994e-02 -3.21082890e-01
-5.42718887e-01 9.05528247e-01 4.53419447e-01 9.29698795e-02
6.52426779e-01 5.12677073e-01 -1.73713610e-01 -1.21147610e-01
-1.13443935e+00 -2.80255586e-01 -3.65854979e-01 6.13257349e-01
7.52313495e-01 2.90336698e-01 -5.65583706e-01 7.09453404e-01
-3.04857254e-01 3.13176453e-01 5.94421029e-01 1.47561181e+00
-6.22629941e-01 -1.19204164e+00 -5.45204699e-01 2.98291236e-01
-1.74347594e-01 -4.75655645e-01 -2.76220083e-01 8.95649374e-01
2.06352219e-01 1.10332894e+00 1.79134488e-01 -1.26546055e-01
-5.70788011e-02 4.15548831e-01 5.25508106e-01 -9.14987683e-01
-3.32352817e-01 3.37168366e-01 6.60251260e-01 -6.61096811e-01
-7.76808202e-01 -8.98576617e-01 -1.55067837e+00 -3.42835158e-01
-1.48205951e-01 3.45434844e-02 -3.56050348e-03 1.16261351e+00
3.58521074e-01 1.79189637e-01 -5.22321276e-02 -1.12953889e+00
1.27236219e-02 -7.97603071e-01 -1.35614172e-01 2.70262748e-01
1.54159144e-01 -8.36375892e-01 -2.05498725e-01 5.87494969e-01] | [10.922453880310059, 9.033588409423828] |
f2bb6e35-a509-4c36-a7a4-21b07c316db6 | flare-aware-cross-modal-enhancement-network | 2305.13659 | null | https://arxiv.org/abs/2305.13659v1 | https://arxiv.org/pdf/2305.13659v1.pdf | Flare-Aware Cross-modal Enhancement Network for Multi-spectral Vehicle Re-identification | Multi-spectral vehicle re-identification aims to address the challenge of identifying vehicles in complex lighting conditions by incorporating complementary visible and infrared information. However, in harsh environments, the discriminative cues in RGB and NIR modalities are often lost due to strong flares from vehicle lamps or sunlight, and existing multi-modal fusion methods are limited in their ability to recover these important cues. To address this problem, we propose a Flare-Aware Cross-modal Enhancement Network that adaptively restores flare-corrupted RGB and NIR features with guidance from the flare-immunized thermal infrared spectrum. First, to reduce the influence of locally degraded appearance due to intense flare, we propose a Mutual Flare Mask Prediction module to jointly obtain flare-corrupted masks in RGB and NIR modalities in a self-supervised manner. Second, to use the flare-immunized TI information to enhance the masked RGB and NIR, we propose a Flare-Aware Cross-modal Enhancement module that adaptively guides feature extraction of masked RGB and NIR spectra with prior flare-immunized knowledge from the TI spectrum. Third, to extract common informative semantic information from RGB and NIR, we propose an Inter-modality Consistency loss that enforces semantic consistency between the two modalities. Finally, to evaluate the proposed FACENet in handling intense flare, we introduce a new multi-spectral vehicle re-ID dataset, called WMVEID863, with additional challenges such as motion blur, significant background changes, and particularly intense flare degradation. Comprehensive experiments on both the newly collected dataset and public benchmark multi-spectral vehicle re-ID datasets demonstrate the superior performance of the proposed FACENet compared to state-of-the-art methods, especially in handling strong flares. The code and dataset will be released soon. | ['Chenglong Li', 'Zi Wang', 'Zhiqi Ma', 'Aihua Zheng'] | 2023-05-23 | null | null | null | null | ['vehicle-re-identification'] | ['computer-vision'] | [ 4.51803625e-01 -9.28209245e-01 8.71562399e-03 -4.96765524e-01
-7.21901119e-01 -7.62129128e-01 4.05731231e-01 -4.50087070e-01
-1.53076768e-01 5.39203465e-01 -2.61963513e-02 1.63724616e-01
-7.27541968e-02 -7.66258657e-01 -6.59945190e-01 -1.09230804e+00
3.47335398e-01 -2.06162170e-01 -1.85991097e-02 -5.08889854e-01
-1.09465219e-01 6.90956652e-01 -2.27299833e+00 1.71838447e-01
1.12056077e+00 1.23146009e+00 3.11750859e-01 3.00174594e-01
-6.72364384e-02 5.65600693e-01 -2.95750231e-01 -8.30307901e-02
4.92073804e-01 -2.59383768e-01 -1.89281538e-01 3.19146544e-01
8.08632612e-01 -3.25931728e-01 -3.96740556e-01 1.19329631e+00
4.19325292e-01 3.50625426e-01 4.50775951e-01 -1.62716317e+00
-5.28111756e-01 -2.54160076e-01 -8.26080084e-01 3.43426943e-01
1.85388565e-01 6.04451418e-01 3.60241532e-01 -8.71630847e-01
2.49221370e-01 1.21508324e+00 8.22039545e-01 4.10050780e-01
-9.48704660e-01 -1.13250625e+00 4.47362214e-02 5.36031663e-01
-1.56173956e+00 -7.87964284e-01 1.17882109e+00 -3.08770239e-01
7.37778544e-01 3.47620726e-01 2.78121859e-01 7.26474106e-01
-1.63622081e-01 3.56933415e-01 1.36141956e+00 9.60791931e-02
-1.63959831e-01 2.72382628e-02 3.48187573e-02 6.08745277e-01
1.69605508e-01 5.27120590e-01 -6.33520901e-01 -6.31685406e-02
3.38528231e-02 2.82514662e-01 -5.04575491e-01 1.53108761e-01
-7.29667723e-01 5.96851349e-01 5.37362874e-01 6.26466051e-02
-2.67657399e-01 -4.20107655e-02 6.98718429e-02 2.84812659e-01
6.41011298e-01 -4.11561519e-01 -4.69137788e-01 5.39650619e-01
-6.45190358e-01 -5.81052750e-02 1.09539986e-01 4.59899992e-01
1.46557033e+00 2.63301611e-01 3.13353091e-02 9.98514414e-01
6.88905537e-01 1.42510760e+00 -8.49027187e-03 -7.33965039e-01
3.38575959e-01 4.42413688e-01 9.18634012e-02 -1.00305521e+00
-5.50493896e-01 -3.08640271e-01 -9.55643237e-01 3.26467156e-01
-1.34380847e-01 -4.23031189e-02 -1.07254362e+00 1.75569034e+00
6.19530499e-01 5.99419653e-01 3.62404376e-01 1.23179531e+00
1.20428884e+00 6.26362562e-01 2.12957427e-01 -5.88066995e-01
1.29143965e+00 -6.74808145e-01 -8.88605833e-01 -6.05197608e-01
9.17149708e-02 -8.35760951e-01 5.70681214e-01 -1.19045086e-01
-3.45391124e-01 -9.03647482e-01 -1.02325487e+00 2.71886915e-01
-4.51801598e-01 6.49285242e-02 5.70480168e-01 7.20237136e-01
-9.47536588e-01 1.15548015e-01 -3.51325303e-01 -2.23696679e-01
2.91825861e-01 7.27696493e-02 -4.52483863e-01 -4.74453688e-01
-1.39185691e+00 7.80781686e-01 2.14545667e-01 6.01460099e-01
-8.23814809e-01 -7.85204589e-01 -1.04013968e+00 -5.14422774e-01
3.80186528e-01 -4.38400477e-01 5.70937872e-01 -1.24066520e+00
-1.09003580e+00 7.08179295e-01 -6.75534248e-01 5.17924726e-02
6.05088994e-02 1.47856683e-01 -1.02552819e+00 3.00151736e-01
1.08425416e-01 3.91627252e-01 1.19366848e+00 -1.85547316e+00
-6.67274654e-01 -6.17966712e-01 -1.42501965e-01 3.47827792e-01
-1.71792194e-01 7.90839717e-02 -4.60069478e-01 -3.80499780e-01
3.46433744e-02 -8.26460481e-01 2.07316011e-01 -1.55830055e-01
3.87659743e-02 1.82940660e-03 1.33392143e+00 -8.03683579e-01
7.98525870e-01 -2.47471642e+00 -3.60532820e-01 1.41349301e-01
-2.89785955e-02 2.84095019e-01 -4.68254238e-01 -1.78315803e-01
-1.38075382e-01 -1.48315802e-01 -5.46160102e-01 -1.74205959e-01
-1.04793407e-01 2.89090484e-01 -1.07189059e-01 8.24182451e-01
3.35569352e-01 6.57135725e-01 -7.30634212e-01 -3.37413549e-01
5.60260713e-01 9.62285101e-01 1.29969656e-01 1.30435348e-01
-8.65788013e-02 6.00222111e-01 -3.02597731e-01 1.03797674e+00
1.55444753e+00 2.80267149e-01 -2.45844647e-01 -7.51374424e-01
-9.81943086e-02 -4.19785202e-01 -1.08841968e+00 1.35141253e+00
-4.14701134e-01 3.87340873e-01 4.32342172e-01 -4.89401191e-01
9.46867406e-01 9.32704881e-02 5.60719073e-01 -1.27630115e+00
3.12026709e-01 1.32638723e-01 -6.36349440e-01 -6.71364903e-01
3.48337889e-01 -2.80136615e-01 2.32297748e-01 7.91918412e-02
-2.97546804e-01 2.57838350e-02 -8.21846202e-02 -1.83411241e-01
8.16428363e-01 1.07495077e-01 -4.78463531e-01 1.31152689e-01
1.00767064e+00 -1.83388248e-01 8.38421226e-01 2.57729918e-01
-3.75805736e-01 4.53129053e-01 -4.73565340e-01 -2.54759222e-01
-4.44572419e-01 -1.01646376e+00 -2.79710382e-01 1.12846875e+00
6.30832732e-01 3.31200361e-02 -3.68160963e-01 -5.54280698e-01
3.53071719e-01 5.74066699e-01 -5.02769947e-01 -2.46529266e-01
-3.16627532e-01 -1.15890718e+00 4.23224598e-01 3.86305004e-01
1.08688331e+00 -7.07736373e-01 -2.40140721e-01 -1.89390138e-01
-5.95810115e-01 -1.32206166e+00 -4.46224898e-01 -1.01103254e-01
-3.41279924e-01 -1.37682164e+00 -9.35711265e-02 -5.63435972e-01
4.80418921e-01 1.08206773e+00 6.78396642e-01 2.65958548e-01
-4.14678276e-01 5.20734131e-01 -4.02758837e-01 -1.55702323e-01
-2.07981721e-01 -5.95229864e-01 2.12251823e-02 5.99381149e-01
4.90081429e-01 -2.75935650e-01 -6.41806185e-01 5.82650483e-01
-1.12272537e+00 -1.64942652e-01 3.52348328e-01 4.86657143e-01
7.12665319e-01 5.24334013e-01 4.99051303e-01 -3.12678665e-01
-1.07965022e-02 -4.73246217e-01 -5.38541913e-01 1.57365903e-01
-3.69282991e-01 -3.24942321e-01 3.93132687e-01 -2.28353322e-01
-1.74063683e+00 3.60127389e-01 -4.78368029e-02 -5.56574881e-01
-3.49727243e-01 2.37855375e-01 -5.88716090e-01 -5.93709767e-01
5.52474737e-01 2.24816039e-01 3.22000906e-02 -3.21993321e-01
3.77376586e-01 9.03705418e-01 8.16113114e-01 -1.39790908e-01
1.37063301e+00 9.20810640e-01 9.47085172e-02 -8.30903590e-01
-8.17042470e-01 -7.44772375e-01 -2.89867353e-02 -6.10134304e-01
7.96119690e-01 -1.43004060e+00 -7.59875000e-01 9.40996289e-01
-8.90152037e-01 -1.12834655e-01 2.57786483e-01 2.86943495e-01
-1.51716530e-01 5.05002201e-01 -2.31649220e-01 -9.18381572e-01
-3.87322605e-01 -1.05576360e+00 1.17360628e+00 6.81893766e-01
6.92221045e-01 -4.91947949e-01 -9.59306732e-02 9.28899705e-01
5.52115083e-01 4.10801619e-01 3.74477118e-01 2.01818541e-01
-5.59129834e-01 8.21885243e-02 -5.77351153e-01 4.37814772e-01
6.23864591e-01 -1.54266253e-01 -1.43297148e+00 -3.26866806e-01
-5.76946624e-02 -9.91208181e-02 1.35671723e+00 -2.99481209e-02
8.49382281e-01 1.19420879e-01 -3.30702543e-01 1.02043891e+00
1.63111281e+00 1.49381459e-01 4.57204163e-01 3.41233343e-01
1.04941368e+00 7.10906565e-01 8.35593581e-01 4.13223147e-01
6.53264761e-01 5.34351468e-01 7.97666132e-01 -3.20777953e-01
-4.24831152e-01 2.08321020e-01 4.43277150e-01 4.09977995e-02
1.30397424e-01 -1.27040520e-01 -4.90965694e-01 4.78660703e-01
-1.70821106e+00 -1.12950134e+00 -4.59562391e-01 2.05953360e+00
5.74956775e-01 -4.84085262e-01 -1.81852683e-01 -1.03514828e-01
8.99172604e-01 4.25018132e-01 -7.81230927e-01 3.30219358e-01
-1.00913596e+00 -6.41626492e-02 8.20396245e-01 4.80655193e-01
-1.31975996e+00 6.54567719e-01 4.89941597e+00 7.74135888e-01
-1.29997313e+00 4.03805256e-01 3.97163719e-01 -2.08959100e-03
-6.35353565e-01 -2.00385168e-01 -6.60365164e-01 5.15185773e-01
7.64030695e-01 1.81658462e-01 7.97975361e-01 2.69928545e-01
4.39343005e-01 -3.60091299e-01 -5.22618592e-01 1.28772306e+00
6.15920484e-01 -8.67339909e-01 -4.07944143e-01 -1.60195366e-01
7.54518807e-01 2.84142703e-01 1.29380211e-01 -9.16533768e-02
1.49219126e-01 -9.40268099e-01 6.20441854e-01 7.39836156e-01
9.06789541e-01 -9.40400124e-01 6.56541467e-01 -1.84126005e-01
-1.76658106e+00 -2.81082720e-01 -1.76410645e-01 4.30036575e-01
4.64165173e-02 6.90448761e-01 -2.05976546e-01 9.71546113e-01
1.02267933e+00 7.98720300e-01 -6.22570097e-01 7.63690770e-01
-6.68666437e-02 2.95837194e-01 -5.14750481e-01 7.45167077e-01
-1.52776137e-01 -2.63133526e-01 4.22559053e-01 1.08014381e+00
2.13083163e-01 1.07390709e-01 2.57813871e-01 7.14212894e-01
1.41875550e-01 -2.20848024e-01 -4.36116397e-01 2.10372686e-01
4.08306211e-01 1.58990836e+00 -3.02089781e-01 7.00660869e-02
-5.30098259e-01 9.79475915e-01 -3.69727522e-01 7.73679137e-01
-1.18844068e+00 -4.62757200e-02 1.24864602e+00 -9.40379798e-02
2.04933420e-01 -3.52169052e-02 3.57065052e-02 -8.54760468e-01
1.55352369e-01 -6.36650562e-01 4.36965823e-01 -1.08705544e+00
-1.38867712e+00 6.20303631e-01 -1.71328098e-01 -1.20560443e+00
2.41652861e-01 -4.29208457e-01 -3.66243184e-01 9.92566288e-01
-2.51579452e+00 -1.65697801e+00 -9.56102312e-01 1.06142998e+00
2.24339008e-01 7.73340231e-03 3.54722738e-01 6.57178283e-01
-6.59086525e-01 3.84582251e-01 1.21605933e-01 -1.81828439e-01
9.81048346e-01 -5.97287536e-01 -2.10919797e-01 9.62114692e-01
-4.97769475e-01 -1.36115646e-03 6.79459929e-01 -6.72628999e-01
-1.90090728e+00 -1.84511530e+00 3.62953156e-01 -2.54865140e-02
3.96148682e-01 1.14352457e-01 -1.00576055e+00 3.35675448e-01
-7.45807886e-02 4.90522236e-01 4.81405437e-01 -3.95748734e-01
-6.59295499e-01 -8.00938249e-01 -1.36382008e+00 -2.99635343e-02
1.03865194e+00 -8.88395250e-01 -1.32309839e-01 4.02411371e-01
5.29825628e-01 -1.97249681e-01 -5.86405516e-01 9.99309897e-01
2.17488974e-01 -9.07768309e-01 1.16047192e+00 2.12805718e-01
-2.30672807e-01 -1.07104516e+00 -4.81742591e-01 -1.29005277e+00
-2.42166549e-01 -4.29327816e-01 2.02477321e-01 1.59823394e+00
-6.66587055e-02 -8.01611602e-01 2.65488505e-01 1.44468039e-01
-2.39953548e-01 1.90566868e-01 -8.72866273e-01 -5.70327699e-01
-4.69589591e-01 -6.09374762e-01 8.30935180e-01 8.98537040e-01
-8.76727700e-01 -1.17936879e-01 -4.07881320e-01 8.44184816e-01
9.53729331e-01 3.02882463e-01 6.97346866e-01 -1.25873077e+00
1.27948165e-01 -1.66804433e-01 -1.46816462e-01 -3.27117801e-01
6.57659054e-01 -8.52125406e-01 5.24589241e-01 -1.40681696e+00
3.54902089e-01 -3.43927979e-01 -4.76124257e-01 5.96802950e-01
-4.28508997e-01 8.93421173e-01 1.89714313e-01 9.62037668e-02
-5.88493764e-01 7.40110815e-01 9.13614094e-01 -5.63258529e-01
-5.52821085e-02 -5.06222904e-01 -7.23033965e-01 6.51697874e-01
6.40418530e-01 -2.65427202e-01 -3.82472336e-01 -5.19800842e-01
1.43570423e-01 -2.40807496e-02 8.53130937e-01 -1.09083307e+00
2.07354069e-01 -4.93464351e-01 5.39477348e-01 -7.45207608e-01
4.31711018e-01 -9.51950014e-01 6.26581013e-01 -4.72683869e-02
4.75829303e-01 -2.87369251e-01 5.82915127e-01 6.79853678e-01
-2.55243719e-01 2.12166309e-01 8.48392963e-01 1.44991502e-01
-1.11779559e+00 5.11851072e-01 -2.63117790e-01 -3.87660205e-01
1.00471961e+00 -3.80367994e-01 -8.09791386e-01 -1.13716535e-01
-1.35696799e-01 4.34232265e-01 6.91125810e-01 7.28512108e-01
5.97775757e-01 -1.31685567e+00 -8.74910831e-01 4.50282574e-01
4.93552655e-01 -3.06233853e-01 1.20605397e+00 8.62562776e-01
-7.66176805e-02 2.26590894e-02 -1.11036390e-01 -6.04708552e-01
-1.53079641e+00 5.11649311e-01 6.89302623e-01 4.78429705e-01
-2.05958277e-01 6.25951648e-01 3.84457171e-01 -3.29705805e-01
-8.66238475e-02 3.89932066e-01 -3.04867327e-01 1.98717207e-01
7.99144149e-01 3.37230861e-01 3.85613382e-01 -1.48473930e+00
-7.67794073e-01 9.94588912e-01 4.92173672e-01 4.50605959e-01
1.16866922e+00 -6.54294074e-01 -3.62683892e-01 7.91039690e-02
1.18768346e+00 -1.76744480e-02 -1.40150523e+00 -3.61310542e-01
-7.54867613e-01 -5.88292599e-01 4.45015490e-01 -9.26365614e-01
-1.61809826e+00 4.00340199e-01 1.22421956e+00 -2.43422925e-01
1.90715539e+00 -1.73252091e-01 8.87562335e-01 2.13654377e-02
1.57944635e-01 -9.75185931e-01 -1.23697869e-01 4.32550222e-01
6.86208546e-01 -1.48276913e+00 -2.36163229e-01 -6.18733346e-01
-5.40480494e-01 8.62667441e-01 7.01901972e-01 4.21285361e-01
5.19888580e-01 3.25106651e-01 3.87386441e-01 -1.08077027e-01
-3.38214487e-01 -8.11840296e-01 3.20265830e-01 8.57485116e-01
-8.55910927e-02 -3.40554956e-03 2.06943601e-01 3.35186332e-01
2.25463137e-01 -8.70228931e-02 1.86029561e-02 6.16284728e-01
-5.38583755e-01 -6.41328454e-01 -9.38265860e-01 1.91483453e-01
1.86002646e-02 1.02097914e-01 -2.61204243e-01 2.83465803e-01
6.97141528e-01 1.67836428e+00 6.97558746e-02 -6.91587389e-01
2.32534915e-01 2.46127229e-02 3.89409065e-01 1.29893407e-01
-4.56279278e-01 1.63039729e-01 -4.95823333e-05 -7.58394122e-01
-1.00011194e+00 -6.61668003e-01 -1.34406364e+00 -5.87004066e-01
-3.06124777e-01 -2.18368486e-01 8.85310829e-01 9.99110460e-01
3.29113632e-01 5.07044971e-01 1.14478731e+00 -9.65954304e-01
8.17031786e-02 -7.43226230e-01 -6.21054292e-01 6.72795355e-01
7.74464965e-01 -9.86002326e-01 -8.36228251e-01 2.70026401e-02] | [10.241764068603516, -1.759351134300232] |
573b8b6d-2250-4ec3-a30e-be7589e4f196 | secrets-of-3d-implicit-object-shape | 2101.06860 | null | https://arxiv.org/abs/2101.06860v2 | https://arxiv.org/pdf/2101.06860v2.pdf | Mending Neural Implicit Modeling for 3D Vehicle Reconstruction in the Wild | Reconstructing high-quality 3D objects from sparse, partial observations from a single view is of crucial importance for various applications in computer vision, robotics, and graphics. While recent neural implicit modeling methods show promising results on synthetic or dense data, they perform poorly on sparse and noisy real-world data. We discover that the limitations of a popular neural implicit model are due to lack of robust shape priors and lack of proper regularization. In this work, we demonstrate highquality in-the-wild shape reconstruction using: (i) a deep encoder as a robust-initializer of the shape latent-code; (ii) regularized test-time optimization of the latent-code; (iii) a deep discriminator as a learned high-dimensional shape prior; (iv) a novel curriculum learning strategy that allows the model to learn shape priors on synthetic data and smoothly transfer them to sparse real world data. Our approach better captures the global structure, performs well on occluded and sparse observations, and registers well with the ground-truth shape. We demonstrate superior performance over state-of-the-art 3D object reconstruction methods on two real-world datasets. | ['Raquel Urtasun', 'Shenlong Wang', 'Justin Liang', 'Sivabalan Manivasagam', 'Wei-Chiu Ma', 'ZiHao Wang', 'Shivam Duggal'] | 2021-01-18 | null | null | null | null | ['3d-object-reconstruction'] | ['computer-vision'] | [ 1.82374522e-01 1.53102949e-01 -2.02580437e-01 -4.31460023e-01
-8.25055420e-01 -3.76769036e-01 6.81600392e-01 -1.86462671e-01
1.18568383e-01 4.47431445e-01 2.81523883e-01 -2.80771833e-02
3.82771641e-02 -7.17050076e-01 -1.06511939e+00 -6.59627914e-01
1.91312477e-01 1.17139578e+00 2.63628155e-01 1.02552488e-01
8.72195810e-02 8.18596482e-01 -1.51762199e+00 2.39348412e-01
6.63333237e-01 9.94433045e-01 4.67190802e-01 2.35562190e-01
-8.68191570e-02 4.54012424e-01 -4.93118316e-02 -3.27819251e-02
3.61004770e-01 -1.15734294e-01 -4.34669554e-01 6.17215812e-01
7.37388015e-01 -5.08811176e-01 -4.15319741e-01 1.00090671e+00
3.56732488e-01 -1.72218725e-01 8.98747444e-01 -8.49101305e-01
-8.41821492e-01 -1.16839416e-01 -6.56837523e-01 -3.88950378e-01
1.28721863e-01 2.49824896e-01 7.44748056e-01 -1.26863241e+00
1.02168429e+00 1.28416193e+00 7.50060201e-01 4.82245654e-01
-1.74250627e+00 -3.47679466e-01 5.09598441e-02 -3.18979084e-01
-1.34141994e+00 -6.66181982e-01 9.94271457e-01 -5.44370413e-01
8.50797534e-01 -4.83453237e-02 7.81840444e-01 1.15507221e+00
-1.08040802e-01 7.79796243e-01 8.29046607e-01 -2.55538553e-01
3.33015174e-01 3.49780209e-02 -4.59551960e-01 9.74070907e-01
2.57704735e-01 3.72672111e-01 -3.92985255e-01 -3.28062564e-01
1.53097761e+00 1.80553645e-01 -3.50616395e-01 -1.23148358e+00
-1.17069495e+00 7.82806516e-01 4.88109022e-01 5.56499064e-02
-4.76149440e-01 2.94359595e-01 -1.64690271e-01 1.16706200e-01
5.57941079e-01 2.74868786e-01 -4.63962942e-01 1.14322647e-01
-9.54724371e-01 2.61122704e-01 8.30991805e-01 1.15734160e+00
8.12374055e-01 5.63787460e-01 2.46174037e-01 8.12302232e-01
6.22941315e-01 8.42905223e-01 6.52278885e-02 -1.22220409e+00
2.71305501e-01 3.43391836e-01 1.60415784e-01 -8.41596901e-01
-2.88129020e-02 -7.62646258e-01 -7.67908692e-01 5.43876052e-01
3.24282736e-01 4.31473702e-01 -1.24926388e+00 1.70571005e+00
3.41453254e-01 4.16301578e-01 -2.31703326e-01 9.67387557e-01
8.58128786e-01 5.51981270e-01 -4.46617454e-01 9.22977328e-02
8.72191966e-01 -6.59743786e-01 -2.96915799e-01 -3.17510188e-01
1.57620639e-01 -7.65690029e-01 8.66409957e-01 2.89381206e-01
-1.30861759e+00 -4.96351779e-01 -9.04746771e-01 -2.14352861e-01
1.65107340e-01 2.27191076e-01 4.97672647e-01 1.90180019e-01
-9.39263940e-01 4.38570440e-01 -1.08203936e+00 -1.05560318e-01
6.28125310e-01 2.95405746e-01 -5.78749657e-01 -4.98948425e-01
-1.59541398e-01 5.91887653e-01 -3.31441872e-02 -2.83317447e-01
-1.36612689e+00 -9.33274806e-01 -1.19204354e+00 -4.88665439e-02
3.76535684e-01 -9.83453751e-01 9.72060859e-01 -8.98533106e-01
-1.57748675e+00 1.22636461e+00 -9.36944261e-02 -9.32645872e-02
3.31078798e-01 -1.61283627e-01 2.98728049e-01 1.47573158e-01
8.20697173e-02 7.08276868e-01 1.15547240e+00 -1.90514040e+00
8.66214857e-02 -4.64934856e-01 -3.88331741e-01 6.99928999e-02
2.55187362e-01 -6.76191747e-01 -4.84404087e-01 -8.33952248e-01
7.20677435e-01 -8.23883176e-01 -2.98470616e-01 6.40771627e-01
-1.59138158e-01 2.28484422e-01 1.02480328e+00 -5.10342777e-01
1.87590525e-01 -2.09101129e+00 4.83980149e-01 1.59978226e-01
1.89089432e-01 2.00380832e-01 -4.35633719e-01 1.37686595e-01
4.14424203e-02 -4.17515039e-01 -4.77394849e-01 -8.81971776e-01
-5.30799702e-02 5.98666668e-01 -4.85727906e-01 7.15626180e-01
2.37534180e-01 1.09999549e+00 -7.99711287e-01 -1.62076011e-01
2.65800565e-01 7.35285819e-01 -8.02455366e-01 5.30336082e-01
-5.87361217e-01 7.81854868e-01 -4.02189046e-01 8.31596434e-01
6.00382209e-01 -6.15809321e-01 -7.42497593e-02 -3.28359932e-01
3.26833576e-02 2.64889777e-01 -1.25515890e+00 2.31485295e+00
-2.26828098e-01 3.31668258e-01 5.34226000e-01 -1.10314941e+00
1.21023810e+00 3.61155510e-01 6.30977392e-01 -5.09542108e-01
4.79580387e-02 4.51057285e-01 -2.92352945e-01 -4.08609062e-01
1.15501404e-01 -4.06463325e-01 3.95609528e-01 5.24851859e-01
3.72581631e-01 -9.39431787e-01 -4.44171607e-01 5.56791872e-02
9.19227958e-01 5.74203610e-01 -6.01041242e-02 -2.22480178e-01
1.24458492e-01 -2.68861890e-01 6.19508386e-01 5.16869843e-01
3.53837609e-01 1.17793083e+00 7.17516467e-02 -6.47935569e-01
-1.54872048e+00 -1.33582151e+00 -1.80004567e-01 4.54172492e-01
1.67236142e-02 2.41910331e-02 -2.71356642e-01 -3.39786321e-01
3.21483493e-01 5.85294545e-01 -5.27205288e-01 3.60378176e-02
-5.65614879e-01 -2.60452688e-01 5.57999574e-02 5.35376012e-01
1.39928699e-01 -9.02948976e-01 -4.14890885e-01 2.46936873e-01
-4.00279872e-02 -1.31337523e+00 -4.13762510e-01 2.57614821e-01
-1.32823193e+00 -9.15388763e-01 -7.76248157e-01 -9.90061164e-01
1.02882195e+00 3.95816863e-01 1.37373662e+00 1.47796631e-01
-2.54598469e-01 6.44656003e-01 5.24298511e-02 -2.10645333e-01
-4.57659751e-01 -3.93790603e-01 -6.38049394e-02 8.96357670e-02
-1.90944463e-01 -1.05985844e+00 -2.89117187e-01 3.69065046e-01
-8.35628808e-01 2.62502700e-01 7.01961160e-01 1.05929613e+00
1.03031242e+00 -5.08527875e-01 2.93951064e-01 -6.51841164e-01
-1.80631336e-02 -3.28719109e-01 -8.58548522e-01 4.67771590e-02
-3.56086493e-01 3.35215688e-01 2.07000732e-01 -5.83257973e-01
-1.01627481e+00 4.06177402e-01 -2.76406735e-01 -9.91162241e-01
-2.44297892e-01 2.93232977e-01 -1.52571097e-01 -1.76712722e-01
5.82884133e-01 4.43000942e-01 3.89023721e-01 -8.50408912e-01
3.56320828e-01 6.85821921e-02 6.68008626e-01 -1.00971031e+00
1.04171479e+00 8.81372809e-01 3.04563791e-01 -9.63665068e-01
-8.05882931e-01 -2.62209952e-01 -8.10465515e-01 2.56712496e-01
6.27325773e-01 -1.21007907e+00 -3.28027546e-01 4.03627872e-01
-1.22204733e+00 -7.14110732e-01 -6.60540462e-01 4.88396764e-01
-1.04443836e+00 3.85649681e-01 -6.47296906e-01 -5.60670912e-01
-7.40899444e-02 -1.25527501e+00 1.54732585e+00 -1.81675807e-01
5.35102226e-02 -9.52063143e-01 6.85552657e-02 2.50875503e-01
2.99383789e-01 3.84687543e-01 9.86107111e-01 6.99127167e-02
-1.14227009e+00 -7.81028271e-02 -3.05396795e-01 3.64474237e-01
1.15603894e-01 -2.46057823e-01 -8.05811048e-01 -4.96559948e-01
2.12738171e-01 -7.33454168e-01 8.15081596e-01 6.71378076e-01
1.20910168e+00 -2.06008881e-01 -2.41114840e-01 1.20186734e+00
1.50873578e+00 -2.64976948e-01 5.23980677e-01 -1.97720557e-01
7.97743917e-01 4.60720688e-01 3.84639241e-02 3.50903392e-01
3.76472801e-01 7.37100959e-01 8.09309959e-01 -1.80738807e-01
-4.89056706e-01 -5.40158749e-01 1.95612296e-01 1.03373516e+00
-1.33473918e-01 1.59887105e-01 -7.36179113e-01 6.60797715e-01
-1.80349112e+00 -7.33682275e-01 2.16444004e-02 2.17580295e+00
9.04444098e-01 6.89181238e-02 -2.16663539e-01 -1.47089347e-01
1.00952484e-01 1.28922522e-01 -6.62698448e-01 2.32898548e-01
-2.22147375e-01 3.23882610e-01 9.44806486e-02 6.67731404e-01
-7.97909498e-01 7.31741607e-01 6.04688454e+00 5.61676919e-01
-1.21985960e+00 1.75513774e-01 4.35374260e-01 4.33471464e-02
-8.37459922e-01 -3.37866768e-02 -5.28544307e-01 1.35549724e-01
1.33464620e-01 3.71288896e-01 3.80501509e-01 8.31464529e-01
-1.59536704e-01 1.95897773e-01 -1.32943165e+00 1.17326260e+00
3.26913953e-01 -1.56717372e+00 1.52924672e-01 2.87753224e-01
1.11808813e+00 3.56551141e-01 1.20431513e-01 -7.39530399e-02
3.45531285e-01 -1.18780601e+00 1.09750021e+00 6.45568848e-01
1.12450194e+00 -2.67682552e-01 2.85208970e-01 7.06183851e-01
-8.77419233e-01 2.82903939e-01 -6.01251483e-01 1.91611305e-01
2.43325919e-01 6.64739966e-01 -6.35169268e-01 2.74664879e-01
3.79721165e-01 9.93942559e-01 -9.78969708e-02 1.05750537e+00
-3.57072443e-01 4.99094427e-01 -5.44304311e-01 4.06003326e-01
1.15811810e-01 -2.67186403e-01 8.76919866e-01 8.08079481e-01
4.45428282e-01 2.83486456e-01 4.09123898e-01 1.40206420e+00
-4.13379110e-02 -3.54207516e-01 -8.54562402e-01 5.84972464e-02
1.37601376e-01 8.76454473e-01 -6.37217522e-01 -1.90911591e-01
-4.16606009e-01 6.85099781e-01 3.88567418e-01 5.82244277e-01
-2.80582041e-01 4.01953429e-01 5.85652649e-01 5.76008916e-01
7.38531470e-01 -8.12672615e-01 -5.98202944e-01 -1.28663528e+00
-3.16719972e-02 -7.07514882e-01 -1.62989169e-01 -1.00893319e+00
-1.31791651e+00 4.32809591e-01 -1.56451553e-01 -1.29802632e+00
-2.06283793e-01 -6.22277081e-01 -3.84047210e-01 8.01749945e-01
-1.62978935e+00 -1.36437511e+00 -3.38257432e-01 6.41434669e-01
5.28961062e-01 -4.75130200e-01 9.58787262e-01 1.55132696e-01
1.58307388e-01 1.23461671e-01 4.27991152e-02 3.90435010e-02
3.68585140e-01 -1.03064322e+00 4.30508733e-01 5.04667580e-01
5.12751222e-01 3.74189317e-01 6.24902010e-01 -6.51771188e-01
-1.82718313e+00 -8.66481364e-01 2.89840698e-01 -4.32333380e-01
2.02396035e-01 -3.91085446e-01 -1.14555991e+00 8.13219070e-01
-1.55752584e-01 5.79024434e-01 2.46442646e-01 -4.38175872e-02
-4.94319081e-01 1.38044104e-01 -1.16423857e+00 2.21734628e-01
1.10036922e+00 -5.62824011e-01 -6.77436650e-01 1.39037326e-01
6.71042919e-01 -7.59622991e-01 -6.34508669e-01 4.43141222e-01
7.08120763e-01 -9.34009612e-01 1.29933858e+00 -5.31400263e-01
6.41086876e-01 -2.58364916e-01 -4.33383912e-01 -1.25441432e+00
-4.62697119e-01 -4.14777726e-01 -5.41197836e-01 6.10196292e-01
7.76113570e-02 -3.78122836e-01 1.10279679e+00 4.01217043e-01
-4.68438923e-01 -1.03295279e+00 -7.78295696e-01 -7.03055203e-01
1.75597845e-03 -3.58130127e-01 4.70293105e-01 9.67425287e-01
-8.76752555e-01 2.78048724e-01 -3.42310011e-01 2.53414005e-01
1.18213534e+00 5.15518069e-01 1.02318156e+00 -1.47761095e+00
-5.94603598e-01 -2.44893551e-01 -4.05140281e-01 -1.77546418e+00
2.07823917e-01 -9.31095064e-01 1.50615826e-01 -1.51591384e+00
1.81266680e-01 -7.59482563e-01 2.24559307e-01 4.24451828e-01
3.70622009e-01 3.22840333e-01 1.01666316e-01 2.76352525e-01
-3.10231030e-01 1.14878619e+00 1.68285990e+00 -2.83360839e-01
3.52801718e-02 -1.87058002e-02 -3.46908599e-01 8.04170549e-01
2.15455487e-01 -6.41582489e-01 -4.43077296e-01 -9.35158849e-01
2.08418697e-01 2.60641545e-01 6.42979741e-01 -6.55438721e-01
1.30915999e-01 -2.30755523e-01 5.37481010e-01 -7.53937662e-01
8.20762455e-01 -1.04466891e+00 2.99588859e-01 3.29098612e-01
2.16030274e-02 -3.62346351e-01 1.06730267e-01 7.00174987e-01
-1.05523139e-01 -2.30216026e-01 9.52129841e-01 -3.38615477e-01
-3.42167675e-01 7.27402627e-01 2.28370026e-01 1.62291631e-01
5.33007264e-01 -4.94318426e-01 6.57990053e-02 -4.53283578e-01
-8.10140848e-01 -6.77049533e-02 9.34624434e-01 3.40234399e-01
9.77525830e-01 -1.67544723e+00 -9.47645307e-01 8.29891801e-01
1.15639186e-02 6.58450603e-01 -1.30004525e-01 5.28153360e-01
-5.04036307e-01 1.57629743e-01 -2.62312919e-01 -1.18971395e+00
-9.50600445e-01 2.34571248e-01 3.92884910e-01 5.60699515e-02
-1.03086102e+00 9.47809875e-01 5.80087185e-01 -8.16195786e-01
4.00122613e-01 -4.15337890e-01 5.06529629e-01 -4.93455172e-01
1.95167616e-01 -1.11630663e-01 -4.26794440e-02 -7.65689969e-01
3.89657170e-02 8.99649143e-01 3.77399564e-01 -1.62078559e-01
1.93188298e+00 1.78929403e-01 -9.07112285e-02 4.89583462e-01
1.20323837e+00 -2.49243025e-02 -1.88382876e+00 -7.81923890e-01
-3.44205230e-01 -8.45647454e-01 1.68139458e-01 -4.59111214e-01
-1.17462349e+00 9.68472600e-01 3.17359865e-01 -1.53323919e-01
6.66900098e-01 1.69633850e-01 7.54119813e-01 3.55329245e-01
6.64797783e-01 -6.52926385e-01 5.32710671e-01 7.19701529e-01
1.29744112e+00 -1.32736266e+00 3.07378501e-01 -4.84298825e-01
-2.95759082e-01 1.08100426e+00 2.79827237e-01 -3.53300303e-01
1.06014359e+00 4.47621137e-01 -1.39689073e-01 -5.17103612e-01
-7.54853368e-01 -4.97272834e-02 5.72672486e-01 8.23805511e-01
2.64350493e-02 -3.66339646e-02 4.87381041e-01 3.70356560e-01
1.78158674e-02 9.60629713e-03 2.02278107e-01 7.04953253e-01
-4.22229826e-01 -9.88131106e-01 -3.77750516e-01 3.15790325e-01
3.80152054e-02 2.24254563e-01 -1.75333973e-02 4.22374219e-01
4.36048135e-02 1.43897370e-01 1.08262025e-01 1.97856590e-01
3.55398536e-01 -1.34863690e-01 1.03672004e+00 -9.49896395e-01
3.87768410e-02 3.42254609e-01 -1.64455965e-01 -6.92441642e-01
-3.11852962e-01 -9.05580580e-01 -1.08023906e+00 -3.32358107e-02
-1.01112381e-01 -2.37186238e-01 7.76803613e-01 9.49371994e-01
4.46748495e-01 1.59407690e-01 4.49137479e-01 -1.49162197e+00
-6.65158451e-01 -7.28826046e-01 -6.23572826e-01 5.33845842e-01
6.00434244e-01 -9.23731863e-01 -1.87625483e-01 2.59167194e-01] | [8.70626163482666, -3.3521416187286377] |
87eea680-c95f-4ec4-82b8-e863f33b337a | chromatic-learning-for-sparse-datasets | 2006.03779 | null | https://arxiv.org/abs/2006.03779v1 | https://arxiv.org/pdf/2006.03779v1.pdf | Chromatic Learning for Sparse Datasets | Learning over sparse, high-dimensional data frequently necessitates the use of specialized methods such as the hashing trick. In this work, we design a highly scalable alternative approach that leverages the low degree of feature co-occurrences present in many practical settings. This approach, which we call Chromatic Learning (CL), obtains a low-dimensional dense feature representation by performing graph coloring over the co-occurrence graph of features---an approach previously used as a runtime performance optimization for GBDT training. This color-based dense representation can be combined with additional dense categorical encoding approaches, e.g., submodular feature compression, to further reduce dimensionality. CL exhibits linear parallelizability and consumes memory linear in the size of the co-occurrence graph. By leveraging the structural properties of the co-occurrence graph, CL can compress sparse datasets, such as KDD Cup 2012, that contain over 50M features down to 1024, using an order of magnitude fewer features than frequency-based truncation and the hashing trick while maintaining the same test error for linear models. This compression further enables the use of deep networks in this wide, sparse setting, where CL similarly has favorable performance compared to existing baselines for budgeted input dimension. | ['Vladimir Feinberg', 'Peter Bailis'] | 2020-06-06 | null | null | null | null | ['feature-compression'] | ['computer-vision'] | [-2.30142400e-01 -5.25356568e-02 -4.34649944e-01 -1.72995955e-01
-1.09180629e+00 -7.02436805e-01 2.32421532e-01 5.81926942e-01
-2.52997637e-01 4.82498854e-01 2.57022709e-01 -2.22380146e-01
-1.97865739e-01 -9.49468374e-01 -8.15798640e-01 -7.63448596e-01
-7.64395118e-01 6.58686101e-01 -1.00157641e-01 1.75971672e-01
3.15545291e-01 4.19545472e-01 -1.86300957e+00 3.42943072e-01
2.88553566e-01 1.08634996e+00 -1.66541532e-01 5.63616335e-01
-6.84107020e-02 4.72390741e-01 -2.98717052e-01 -2.19228998e-01
6.16977155e-01 -9.02688950e-02 -7.96774566e-01 -6.49125278e-02
9.05578613e-01 -6.64524198e-01 -8.58109295e-01 7.87663758e-01
3.02043170e-01 1.72980279e-01 4.51938659e-01 -1.18288457e+00
-7.39382625e-01 7.08772838e-01 -7.15951264e-01 1.99704811e-01
5.01371682e-01 2.93597230e-03 1.62881184e+00 -8.10985446e-01
5.60302734e-01 1.39439619e+00 7.27444410e-01 1.08255446e-01
-1.48407280e+00 -5.58340609e-01 -7.94013590e-02 4.62163091e-01
-1.62782669e+00 -1.65235713e-01 5.32775640e-01 2.69231331e-02
1.20368063e+00 3.49712282e-01 8.20536196e-01 5.71825206e-01
-1.57132536e-01 7.33460963e-01 8.47178102e-01 -4.56942976e-01
3.42763871e-01 -2.59479731e-01 2.71769226e-01 9.11842704e-01
8.17580760e-01 -2.83346251e-02 -6.92522466e-01 -6.23738408e-01
3.06043535e-01 3.77343267e-01 -1.04127072e-01 -6.59059405e-01
-9.84133601e-01 1.27370536e+00 5.99122345e-01 9.93553475e-02
1.03909232e-01 4.77589011e-01 5.94343424e-01 5.27726531e-01
2.61712849e-01 3.99280936e-01 -3.29229683e-01 -1.96256071e-01
-1.21418476e+00 3.80942225e-01 1.08880997e+00 9.28111792e-01
1.17321038e+00 -2.74510741e-01 -1.26098186e-01 4.71344739e-01
-2.83434708e-02 5.02950966e-01 3.89005542e-01 -9.08739448e-01
4.20548558e-01 6.46970868e-01 -3.19103658e-01 -1.26076710e+00
-4.53092515e-01 -3.30348670e-01 -8.62040281e-01 -2.44166315e-01
1.86429217e-01 2.07938462e-01 -9.56718266e-01 1.80979419e+00
6.60635293e-01 4.38095219e-02 -6.47299811e-02 6.69068992e-01
4.22216117e-01 5.71957350e-01 -3.43757898e-01 9.45579410e-02
1.28422940e+00 -7.80049562e-01 -2.06687316e-01 2.14400098e-01
1.17467642e+00 -3.30995500e-01 1.25343525e+00 3.98228914e-01
-1.00710309e+00 -2.12764591e-02 -1.24504745e+00 -5.49386382e-01
-3.77922356e-01 -3.13615113e-01 1.14806366e+00 6.47317767e-01
-1.15490842e+00 5.18022954e-01 -9.92914259e-01 -1.44740403e-01
6.30260289e-01 4.92983550e-01 -5.06972730e-01 -9.30999756e-01
-7.46745467e-01 4.50487733e-01 3.61402929e-01 -5.70600152e-01
-8.17273617e-01 -9.90170896e-01 -1.16520035e+00 3.17156225e-01
2.51120120e-01 -5.20309269e-01 8.56598377e-01 -8.26401934e-02
-6.10525429e-01 5.94357252e-01 -3.19915831e-01 -6.06161416e-01
-7.38203898e-02 -1.86609507e-01 -6.65961113e-03 5.66968918e-01
-7.12697282e-02 6.66088581e-01 6.00393414e-01 -8.11682463e-01
-5.00828624e-01 -4.59805638e-01 1.62369296e-01 1.15126520e-01
-8.62595677e-01 -6.11863136e-01 -4.84502137e-01 -4.69038010e-01
3.06312352e-01 -1.08574271e+00 -2.25157887e-01 1.05394430e-01
-4.01920855e-01 -3.96222740e-01 9.83962357e-01 -1.34461388e-01
1.27506375e+00 -2.52187872e+00 9.72063690e-02 7.72229373e-01
8.33181798e-01 -1.46480978e-01 -3.23361427e-01 7.57128358e-01
3.24217677e-01 -1.34408744e-02 -2.20253050e-01 -4.95162100e-01
2.65424818e-01 4.32297051e-01 -2.65039235e-01 8.23488593e-01
1.56880505e-02 8.09108257e-01 -6.70649111e-01 -3.66090059e-01
-2.47786015e-01 4.23626184e-01 -1.09593391e+00 -1.55440010e-02
-2.47595817e-01 -3.21933836e-01 -2.46639147e-01 7.41445184e-01
8.58592153e-01 -8.03852260e-01 3.06149662e-01 -1.96210161e-01
3.61140996e-01 5.12600005e-01 -1.25392938e+00 1.84500086e+00
-2.99211949e-01 5.09253681e-01 -2.40211338e-01 -1.08621848e+00
6.77246451e-01 -2.01480821e-01 6.21117413e-01 -7.01171517e-01
-2.54138231e-01 3.71788859e-01 -3.21651757e-01 -6.83664531e-02
5.38167357e-01 2.25355521e-01 -2.68480361e-01 7.90152371e-01
5.79648912e-02 2.18287855e-01 5.27967334e-01 7.98949242e-01
1.59978974e+00 -5.05309165e-01 2.17000674e-02 -1.69865578e-01
-6.25975803e-02 1.93966821e-01 2.60428995e-01 8.45253944e-01
3.53812605e-01 4.69660461e-01 7.07362592e-01 -7.52710760e-01
-1.26386464e+00 -7.92176425e-01 -2.80405819e-01 1.14648712e+00
1.48775786e-01 -1.05373025e+00 -4.90747720e-01 -5.36500037e-01
8.07406902e-01 1.01341650e-01 -7.09722459e-01 -5.21122217e-01
-5.14704049e-01 -6.61658823e-01 5.62588334e-01 5.07625639e-01
2.06287041e-01 -3.71155679e-01 -4.83557403e-01 -5.46487682e-02
1.60624743e-01 -1.01718605e+00 -4.17973816e-01 7.64564931e-01
-9.45339859e-01 -1.13766921e+00 -3.59614849e-01 -7.29266286e-01
5.42348683e-01 5.53842664e-01 1.31605494e+00 4.15964365e-01
-6.30817115e-01 4.13691282e-01 -5.74846208e-01 1.77734837e-01
1.92576155e-01 3.56847376e-01 -4.03058454e-02 -3.93608749e-01
9.26354885e-01 -8.64812315e-01 -8.26651037e-01 -2.62086540e-01
-1.20493281e+00 -2.81494498e-01 7.43080795e-01 9.87141490e-01
8.89428437e-01 -3.11445706e-02 2.09602922e-01 -9.63733912e-01
1.68739140e-01 -8.40034664e-01 -5.80946326e-01 1.01272278e-02
-7.28313267e-01 4.48881060e-01 7.46782362e-01 -4.57724899e-01
5.87220006e-02 1.63131319e-02 1.39297888e-01 -5.79044580e-01
1.20962292e-01 7.45412171e-01 1.10057041e-01 -4.55444962e-01
6.29609764e-01 3.51700395e-01 -3.09564453e-02 -5.72805047e-01
4.89755541e-01 4.73202527e-01 3.77520710e-01 -6.80913806e-01
1.10604799e+00 5.65199733e-01 5.47635972e-01 -5.07297635e-01
-1.01450884e+00 -6.01246178e-01 -2.76681602e-01 6.66495144e-01
1.79072946e-01 -1.21959627e+00 -7.20680058e-01 -5.58420569e-02
-5.21409154e-01 -2.10277721e-01 -5.98456144e-01 4.52083886e-01
-3.80933702e-01 5.73743761e-01 -8.30632091e-01 -3.48598450e-01
-2.59298354e-01 -7.00199008e-01 1.25867307e+00 -3.55948895e-01
-2.02226210e-02 -6.22485459e-01 2.07628757e-01 -1.70572530e-02
3.17410678e-01 2.67014235e-01 1.48090649e+00 -8.00607920e-01
-8.56365204e-01 -3.97187829e-01 -3.33233953e-01 3.27649079e-02
-2.56075799e-01 -5.59624672e-01 -6.45055771e-01 -9.19179976e-01
-4.49940741e-01 -9.67927098e-01 1.13749623e+00 1.39519170e-01
1.53945839e+00 -6.65533006e-01 -1.69290513e-01 1.20299399e+00
1.57742524e+00 -6.20420814e-01 4.72479850e-01 2.83682525e-01
8.70483339e-01 8.95529389e-02 4.85498279e-01 1.06218469e+00
6.53032422e-01 5.11406064e-01 4.04869467e-01 -6.19723909e-02
-2.64491558e-01 -3.00333560e-01 8.29342231e-02 8.40920568e-01
5.25934637e-01 3.09871007e-02 -6.81244373e-01 5.92041194e-01
-1.31863642e+00 -8.07349801e-01 1.50309384e-01 2.29135728e+00
1.01503730e+00 -2.29688473e-02 3.12648654e-01 4.66182560e-01
3.07575047e-01 2.55134046e-01 -6.12092793e-01 -4.49049056e-01
-2.81675071e-01 4.35090005e-01 7.43740618e-01 1.20417602e-01
-9.57663715e-01 6.26055121e-01 6.22343588e+00 1.03099167e+00
-9.50273335e-01 -4.20077816e-02 6.44952476e-01 -6.51355922e-01
-6.54836416e-01 -4.95741740e-02 -1.01120484e+00 4.13577229e-01
1.21367538e+00 -1.67161569e-01 7.60596991e-01 1.07006001e+00
-5.12021720e-01 -1.33344948e-01 -1.45624471e+00 1.29689622e+00
2.31944278e-01 -1.60170591e+00 1.91933483e-01 4.45912749e-01
8.24493885e-01 1.94435149e-01 4.55018520e-01 3.52539122e-01
3.66028070e-01 -1.11554682e+00 4.17597920e-01 -3.16897660e-01
1.44337475e+00 -9.41420734e-01 4.55890298e-01 4.83940504e-02
-1.41777551e+00 -4.25797492e-01 -8.54219615e-01 1.91856161e-01
-2.39862949e-01 9.24933612e-01 -4.89738375e-01 2.20746219e-01
9.38766479e-01 7.08710134e-01 -7.00087368e-01 1.07233453e+00
4.16752845e-01 6.59213305e-01 -8.52515936e-01 6.50429279e-02
5.35382986e-01 2.35312805e-01 1.75620615e-01 1.30088747e+00
5.17629802e-01 -8.48169401e-02 3.01317215e-01 4.47973818e-01
-4.74819422e-01 1.14123233e-01 -7.54323721e-01 -2.18810663e-01
8.27920496e-01 1.07368577e+00 -3.47139537e-01 -3.01246047e-01
-6.41030967e-01 6.91313982e-01 7.63232112e-01 1.94214899e-02
-4.84841228e-01 -4.69863266e-01 5.80919266e-01 2.86725104e-01
9.21432257e-01 -4.98691738e-01 -1.41778335e-01 -1.04085732e+00
2.09942684e-01 -1.13640761e+00 8.38265002e-01 -1.83430649e-02
-1.17612326e+00 3.48308712e-01 -3.96363661e-02 -1.18941343e+00
-2.24985287e-01 -3.85990828e-01 -7.89148435e-02 4.46865141e-01
-1.69515634e+00 -9.69233811e-01 -3.18429500e-01 9.49590147e-01
-4.24801521e-02 -2.36728732e-02 8.73839617e-01 3.51104558e-01
-2.45820105e-01 1.20646620e+00 4.70811367e-01 -2.94233859e-01
5.39654195e-01 -1.37225842e+00 2.99872220e-01 4.84403938e-01
4.45553362e-01 5.78328311e-01 2.98891157e-01 -4.10879850e-01
-2.12861919e+00 -1.04763436e+00 6.83173954e-01 6.95059970e-02
5.74865639e-01 -5.73121667e-01 -9.24139857e-01 5.06181955e-01
-3.68543625e-01 5.83270252e-01 1.10542727e+00 4.69292581e-01
-1.17163038e+00 -2.82043904e-01 -1.26114810e+00 2.88881153e-01
1.20115745e+00 -7.16634750e-01 -2.60421097e-01 6.27907276e-01
9.76508975e-01 -3.34493339e-01 -9.25525367e-01 1.41096652e-01
3.85982633e-01 -7.66814888e-01 9.27952826e-01 -7.80839264e-01
3.43491226e-01 -5.27869258e-03 -5.77497184e-01 -8.63518357e-01
-6.08015895e-01 -6.97047591e-01 -9.21961486e-01 6.20240748e-01
2.49735743e-01 -3.29648256e-01 9.82141018e-01 4.10269320e-01
5.54944873e-02 -1.07364106e+00 -1.05466902e+00 -8.51951957e-01
-4.76831384e-02 -7.58059993e-02 8.10882390e-01 9.11633313e-01
1.15900174e-01 -7.25604743e-02 -3.31632286e-01 1.43350288e-01
8.82138908e-01 4.60379392e-01 8.20881724e-01 -1.21485162e+00
-5.12257874e-01 -1.48137152e-01 -8.22651088e-01 -1.22202158e+00
7.07611442e-02 -1.29368138e+00 -3.81092340e-01 -1.00668859e+00
6.24491453e-01 -7.84519374e-01 -3.99042040e-01 6.70085609e-01
6.56752959e-02 6.23024583e-01 4.06231992e-02 3.99019331e-01
-1.11202574e+00 6.11358762e-01 8.58007371e-01 5.27552217e-02
-8.51115212e-02 -7.64995396e-01 -8.47641647e-01 1.07263096e-01
5.06357551e-01 -7.61137068e-01 -1.78034544e-01 -5.18673778e-01
5.03632963e-01 -2.37239718e-01 2.53365308e-01 -1.10472572e+00
4.47555780e-01 8.27238485e-02 4.67824847e-01 -7.11579859e-01
3.44619811e-01 -8.23000968e-01 -1.56428188e-01 5.00261307e-01
-4.66616184e-01 3.82353067e-01 1.73413396e-01 8.24075818e-01
-2.90936857e-01 -1.12857586e-02 5.97631633e-01 6.00559078e-02
-3.33430439e-01 6.16265178e-01 1.92237601e-01 3.34066093e-01
9.54919398e-01 -3.55836570e-01 -5.95246255e-01 -2.92459220e-01
-3.01532567e-01 1.10564031e-01 5.96676946e-01 -1.35617509e-01
6.87070251e-01 -1.51305187e+00 -5.16254127e-01 4.28017795e-01
2.17520356e-01 2.32415885e-01 3.00901920e-01 9.43152428e-01
-4.30447638e-01 4.80878383e-01 -3.11306324e-02 -5.94282031e-01
-1.05068338e+00 7.45539129e-01 -2.31616065e-01 -4.02686119e-01
-1.11735320e+00 9.00931239e-01 8.07262142e-04 1.77067015e-02
4.42292452e-01 5.89793585e-02 4.17851061e-01 -9.62679386e-02
6.97429538e-01 4.11135703e-01 2.51864225e-01 -2.21026152e-01
-3.12081903e-01 3.85922074e-01 -3.93340975e-01 1.72393546e-01
1.69219995e+00 -8.52360670e-03 -3.25792640e-01 1.61113933e-01
1.93386793e+00 1.63849164e-03 -1.20590758e+00 -4.85285938e-01
2.58542188e-02 -8.96646857e-01 2.13936329e-01 -2.60433435e-01
-1.23023903e+00 6.06586277e-01 4.42613333e-01 3.69615555e-01
1.10478234e+00 2.81979144e-01 1.26398945e+00 9.64992404e-01
4.49917018e-01 -8.91431034e-01 2.19065487e-01 4.21888620e-01
4.13295805e-01 -9.48266625e-01 1.52253702e-01 -2.28785902e-01
-2.02756748e-01 1.03227854e+00 2.35995352e-01 -4.70257252e-01
6.54239535e-01 4.47390586e-01 -6.73944890e-01 -3.76599073e-01
-1.06246734e+00 -1.31018385e-01 2.80434996e-01 3.93335313e-01
-1.42710637e-02 2.25497298e-02 -8.71784985e-02 2.36643672e-01
-4.11044031e-01 -3.80666614e-01 2.17164844e-01 1.09554076e+00
-6.06097043e-01 -1.03098345e+00 -2.20196158e-01 1.11780798e+00
-4.33434665e-01 -3.54169071e-01 2.57746782e-02 6.80361807e-01
-2.25180075e-01 6.98496103e-01 3.95589978e-01 -5.69575250e-01
-1.45743921e-01 -1.01312034e-01 4.99620259e-01 -6.60128355e-01
-2.48939604e-01 -2.02067465e-01 -1.52684450e-01 -9.78159964e-01
1.71600580e-02 -5.99193513e-01 -1.14026225e+00 -8.89305055e-01
-2.31596804e-03 1.70232803e-01 4.75768983e-01 6.18270814e-01
7.16131389e-01 -2.92267203e-01 8.77503753e-01 -6.90015852e-01
-1.04006910e+00 -5.77199042e-01 -8.68927956e-01 6.09805346e-01
5.66215992e-01 -6.71549082e-01 -7.69962490e-01 -3.95820320e-01] | [7.08128023147583, 4.959749221801758] |
b9019ec0-a34e-48ce-9fb9-0eaf01de7728 | contrast-with-major-classifier-vectors-for | 2301.05376 | null | https://arxiv.org/abs/2301.05376v1 | https://arxiv.org/pdf/2301.05376v1.pdf | Contrast with Major Classifier Vectors for Federated Medical Relation Extraction with Heterogeneous Label Distribution | Federated medical relation extraction enables multiple clients to train a deep network collaboratively without sharing their raw medical data. In order to handle the heterogeneous label distribution across clients, most of the existing works only involve enforcing regularization between local and global models during optimization. In this paper, we fully utilize the models of all clients and propose a novel concept of \textit{major classifier vectors}, where a group of class vectors is obtained in an ensemble rather than the weighted average method on the server. The major classifier vectors are then distributed to all clients and the local training of each client is Contrasted with Major Classifier vectors (FedCMC), so the local model is not prone to overfitting to the local label distribution. FedCMC requires only a small amount of additional transfer of classifier parameters without any leakage of raw data, extracted representations, and label distributions. Our extensive experiments show that FedCMC outperforms the other state-of-the-art FL algorithms on three medical relation extraction datasets. | ['Yaohui Jin', 'Hao He', 'Chunhui Du'] | 2023-01-13 | null | null | null | null | ['medical-relation-extraction'] | ['medical'] | [ 7.54410774e-02 2.97198355e-01 -4.23213214e-01 -8.53500724e-01
-9.51408803e-01 -2.43773296e-01 1.40467793e-01 2.69271374e-01
-3.61008227e-01 7.43160427e-01 8.78138915e-02 -1.77038506e-01
-1.61915496e-01 -6.89837098e-01 -4.20501262e-01 -1.14680696e+00
-7.42398668e-03 8.83375406e-01 1.45560101e-01 2.45742947e-01
-4.94819134e-01 4.56247360e-01 -1.07094157e+00 9.59232032e-01
5.63625276e-01 1.50713778e+00 -2.36561358e-01 5.26015520e-01
-3.37902963e-01 1.47099054e+00 -7.79766500e-01 -6.10035539e-01
1.60474747e-01 -3.25106651e-01 -1.13682318e+00 7.68382614e-03
2.33354941e-01 -4.32435423e-01 -7.02388063e-02 9.52804327e-01
7.68708706e-01 -6.43681288e-02 4.54782099e-01 -1.27817106e+00
-4.61302370e-01 8.61184478e-01 -3.26899141e-01 -1.76585868e-01
-1.13724664e-01 2.73847412e-02 7.63863921e-01 -4.97704715e-01
8.67474020e-01 8.53706777e-01 7.28862703e-01 5.91342986e-01
-1.13193953e+00 -8.20663929e-01 2.56147027e-01 1.55873090e-01
-1.45626092e+00 -4.77171570e-01 5.36119223e-01 -3.11365426e-01
8.50052893e-01 4.74378049e-01 2.38872394e-01 1.02860701e+00
1.66522130e-01 9.95103061e-01 8.98543954e-01 -1.19124703e-01
3.41330886e-01 4.21106905e-01 5.32430947e-01 8.60388637e-01
1.80471614e-02 -4.04528618e-01 -6.96831405e-01 -7.48694420e-01
2.12716550e-01 3.64944011e-01 -4.55587775e-01 -4.18183893e-01
-1.01237249e+00 8.62307549e-01 4.65565920e-01 3.38791877e-01
-4.47790623e-01 2.75156256e-02 5.70821941e-01 3.21555823e-01
7.58109808e-01 -6.26601055e-02 -9.13319468e-01 2.42515087e-01
-8.03145826e-01 -7.25236833e-02 1.10067916e+00 9.80398536e-01
7.94977546e-01 -6.48103416e-01 -4.67695087e-01 8.00406337e-01
2.43406072e-01 2.47119844e-01 5.42470098e-01 -9.40855026e-01
5.81131935e-01 9.27835822e-01 -4.22785372e-01 -9.43537295e-01
-6.81484699e-01 -8.41441274e-01 -1.09085977e+00 -2.46722773e-01
1.92429811e-01 -5.38691401e-01 -5.89125693e-01 1.67901647e+00
6.48848057e-01 -2.10893732e-02 1.56609714e-01 6.61208272e-01
9.84542489e-01 2.05209047e-01 8.41460750e-02 -3.20153952e-01
1.13910985e+00 -1.32362270e+00 -9.21001971e-01 1.91451386e-01
1.02583909e+00 -5.05505085e-01 4.79612708e-01 2.06541032e-01
-8.26897919e-01 6.77561909e-02 -6.95622444e-01 -1.72215495e-02
-3.67410928e-01 5.36497608e-02 9.39814031e-01 4.28074539e-01
-1.03163636e+00 5.49406528e-01 -8.96341264e-01 -8.97583067e-02
9.91717517e-01 6.73260868e-01 -5.02823532e-01 -1.97099179e-01
-9.59515393e-01 7.95639813e-01 1.15516186e-01 3.17190588e-02
-6.64590418e-01 -6.78640485e-01 -5.58091640e-01 1.76159665e-01
3.90065312e-01 -6.61756098e-01 1.31851697e+00 -9.24215317e-01
-1.38491499e+00 8.04784656e-01 3.39465849e-02 -5.59051186e-02
6.53850973e-01 8.99791792e-02 -1.91330791e-01 2.01246962e-02
-9.01728794e-02 2.24151388e-01 3.39309573e-01 -1.18129563e+00
-8.24204028e-01 -4.80836779e-01 -3.25939238e-01 7.47605488e-02
-5.51834047e-01 -3.98415662e-02 -7.10277259e-01 -2.77734160e-01
1.69727057e-01 -6.43109441e-01 -1.95846364e-01 2.56795526e-01
-6.35496318e-01 -6.30376875e-01 6.43392146e-01 -3.25273037e-01
1.15864933e+00 -2.00640082e+00 -6.83967173e-02 3.94534349e-01
6.23129606e-01 1.10474110e-01 -1.81061849e-01 1.90483049e-01
8.28900710e-02 -6.82654902e-02 -1.12979606e-01 -6.42559886e-01
-1.67744145e-01 4.78308827e-01 1.43954009e-01 7.39914477e-01
-9.98074934e-02 7.44791985e-01 -7.78627753e-01 -1.02765548e+00
-3.68522227e-01 6.01766229e-01 -6.38012171e-01 2.60658532e-01
-1.97551712e-01 1.98796868e-01 -7.37680018e-01 7.15831161e-01
8.35505247e-01 -8.51967335e-01 7.85309374e-01 -3.84272069e-01
6.94934070e-01 3.20032150e-01 -1.03591776e+00 1.69029951e+00
-3.28242213e-01 -2.32137702e-02 2.77505100e-01 -1.16062009e+00
5.99044085e-01 5.39724469e-01 9.33609307e-01 -3.83161217e-01
3.88590574e-01 2.46009216e-01 -1.13781542e-01 -6.86184108e-01
-2.78779835e-01 7.14824274e-02 9.92349237e-02 9.48582113e-01
4.12563115e-01 8.86678994e-01 -2.37390473e-01 2.21180245e-01
1.29856420e+00 -1.08954608e-01 -4.52418625e-02 -9.69925150e-02
3.75051737e-01 -1.15176320e-01 9.07628477e-01 7.46632874e-01
-1.28230020e-01 3.88463944e-01 6.18579328e-01 -7.54375935e-01
-2.26896256e-01 -7.09461749e-01 -2.88062692e-01 1.52674913e+00
-1.51995823e-01 -4.69855845e-01 -7.52682805e-01 -1.54918122e+00
8.21980312e-02 3.57122004e-01 -7.25392759e-01 -8.22478309e-02
-5.42434454e-01 -1.07500184e+00 5.92686772e-01 4.49954122e-01
2.38140509e-01 -1.05980492e+00 -4.52802747e-01 3.40730339e-01
-3.94206882e-01 -8.72279763e-01 -6.10656679e-01 5.31354845e-01
-6.71034038e-01 -1.21889710e+00 -4.41870242e-01 -7.03337133e-01
8.35830331e-01 -2.14422032e-01 1.25925207e+00 3.15748841e-01
-4.10822541e-01 -1.04471028e-01 -2.59217560e-01 -2.83101916e-01
-2.68095911e-01 4.63011533e-01 -3.07588100e-01 4.99876976e-01
5.74577272e-01 -3.55107069e-01 -7.07624674e-01 1.56716064e-01
-7.34275699e-01 -2.43281219e-02 5.75977564e-01 9.96239364e-01
6.22767985e-01 -1.89625323e-01 4.60914195e-01 -1.55007303e+00
4.30958450e-01 -7.85283923e-01 -5.19225411e-02 6.72356665e-01
-6.61621690e-01 7.74250627e-02 4.73064601e-01 -6.55533075e-01
-1.01222241e+00 2.10547239e-01 8.26896429e-02 -2.44533747e-01
3.16576334e-03 4.89516795e-01 -1.75571844e-01 2.23693311e-01
5.81834197e-01 -1.20954856e-01 2.49018878e-01 -7.99441040e-01
2.89867997e-01 1.03170013e+00 2.62170464e-01 -5.07245719e-01
1.03192002e-01 4.11428660e-01 -1.97246492e-01 -2.84615513e-02
-1.20602238e+00 -3.42995524e-01 -4.81442213e-01 -5.37810177e-02
9.21088099e-01 -6.69820011e-01 -1.12587893e+00 5.49420178e-01
-1.09015858e+00 -4.45580244e-01 -4.16337311e-01 3.63341421e-01
-1.64896607e-01 -2.17577577e-01 -9.95163560e-01 -5.03378928e-01
-7.08108306e-01 -1.30381572e+00 9.64659989e-01 8.32159966e-02
-1.05828829e-01 -8.99374008e-01 -2.91098263e-02 5.77767968e-01
6.35740280e-01 2.05736116e-01 9.23044086e-01 -1.15266764e+00
-4.11278635e-01 -5.50975502e-01 -2.65729457e-01 3.12229544e-01
3.51583838e-01 -3.14350188e-01 -1.18126798e+00 -2.63863295e-01
1.40574709e-01 -7.13436782e-01 7.69887805e-01 5.88330580e-03
1.51728666e+00 -5.93692243e-01 -7.21346736e-01 8.80487084e-01
1.30620837e+00 4.75402363e-02 1.11830942e-01 -1.25018448e-01
7.06535518e-01 6.74723685e-01 7.07452297e-02 3.56069237e-01
5.42745769e-01 5.00115871e-01 3.72103870e-01 -4.61738348e-01
1.67613346e-02 1.77221149e-01 1.21613117e-02 9.09964025e-01
2.24778816e-01 -2.51184881e-01 -7.90656090e-01 2.58805126e-01
-2.16193199e+00 -5.51580846e-01 2.51600146e-01 1.84728944e+00
1.30820239e+00 -4.41674292e-01 -2.84846514e-01 -3.01021278e-01
5.14969170e-01 -6.76325783e-02 -8.06299448e-01 -3.15282159e-02
5.66180497e-02 2.75180876e-01 4.92462993e-01 3.03858936e-01
-1.21582806e+00 5.47109783e-01 5.92254114e+00 9.18468237e-01
-1.05816603e+00 6.16637826e-01 1.08049929e+00 -5.24389982e-01
-7.88719133e-02 -3.68777782e-01 -7.16600835e-01 4.09793198e-01
8.72322798e-01 3.74025941e-01 2.39687264e-01 9.45479870e-01
-5.57889283e-01 1.72122076e-01 -1.33246970e+00 8.41597319e-01
1.08009055e-01 -1.54714155e+00 -1.67442262e-01 2.75625437e-01
7.16080546e-01 4.41906333e-01 -1.04484640e-01 2.24204049e-01
7.59573877e-01 -1.08186293e+00 2.14915186e-01 7.73151159e-01
7.89071858e-01 -7.01026797e-01 1.34533346e+00 5.37724793e-01
-8.12535584e-01 -1.77801147e-01 -2.68183380e-01 5.85297465e-01
-1.70001924e-01 6.35884583e-01 -6.34366274e-01 6.90382302e-01
8.41497481e-01 3.76841009e-01 -5.65560162e-01 3.87378901e-01
3.14999849e-01 4.57949281e-01 -6.02723062e-01 -1.52628841e-02
6.48623332e-03 2.08955333e-01 -2.87654877e-01 1.18656862e+00
-1.45676956e-01 4.48838547e-02 6.94560885e-01 3.64102960e-01
-4.03090894e-01 5.10798514e-01 -9.37040197e-04 2.84296721e-01
4.32459712e-01 1.57481396e+00 -5.06218255e-01 -5.99119186e-01
-6.02549791e-01 7.61567235e-01 7.81388164e-01 2.74326980e-01
-6.32759273e-01 -1.02369353e-01 6.05586290e-01 -2.14112937e-01
1.56326100e-01 6.72429204e-01 -2.60877490e-01 -1.20046294e+00
-4.80303820e-03 -1.03032160e+00 8.53590190e-01 -1.33869410e-01
-1.89431310e+00 9.34969485e-01 -4.09545451e-01 -9.36612248e-01
-1.78688526e-01 -2.32298821e-01 -3.41769159e-01 9.09144759e-01
-1.34440422e+00 -1.35546875e+00 -2.98309714e-01 1.07112026e+00
-5.64155020e-02 -4.08679724e-01 1.24689639e+00 5.34639359e-01
-7.86877394e-01 1.17105234e+00 6.98269755e-02 4.81397986e-01
9.51672733e-01 -9.81056988e-01 -4.26014543e-01 1.06396966e-01
-2.33471215e-01 6.66486621e-01 8.68844017e-02 -4.72467780e-01
-1.20120168e+00 -1.40910220e+00 1.17039704e+00 -3.13415498e-01
2.84358025e-01 -3.89281392e-01 -9.29031014e-01 7.53235221e-01
2.56982267e-01 7.99835861e-01 1.36403179e+00 3.69703710e-01
-3.87148380e-01 -6.18654251e-01 -1.51452076e+00 1.10133514e-01
7.94270277e-01 -3.70551318e-01 1.78637341e-01 8.82300377e-01
6.54169798e-01 -3.37453902e-01 -1.27086365e+00 3.64575744e-01
3.99798900e-01 -7.14744449e-01 6.47483945e-01 -9.53248262e-01
1.48248449e-01 -8.37331340e-02 -3.41661036e-01 -8.50339234e-01
-1.87144250e-01 -4.63759929e-01 -4.38415438e-01 1.27638030e+00
6.90312803e-01 -8.81087601e-01 9.76331413e-01 7.65623510e-01
1.40530407e-01 -1.31950533e+00 -8.96227479e-01 -2.40368098e-01
3.91981192e-02 -2.35926628e-01 9.47270036e-01 1.28015232e+00
-9.69583727e-03 1.97410122e-01 -1.67878330e-01 4.25280370e-02
7.40309060e-01 3.27417791e-01 5.59675336e-01 -1.19787002e+00
-6.03028655e-01 -1.97351828e-01 -5.42789474e-02 -5.98399222e-01
8.60738829e-02 -1.43958914e+00 2.66055074e-02 -1.43053651e+00
6.60096824e-01 -8.78740132e-01 -7.39619792e-01 1.25954247e+00
-1.89395905e-01 2.38297388e-01 4.98414300e-02 3.88892263e-01
-9.10808504e-01 2.40379825e-01 1.12188935e+00 -1.53111994e-01
-1.48909852e-01 1.46970063e-01 -8.23317170e-01 6.51842535e-01
5.82544684e-01 -8.67362738e-01 -4.09170091e-01 -4.50113624e-01
3.90938520e-02 1.73895955e-01 -4.75330502e-02 -5.91107011e-01
5.63514531e-01 -5.50631732e-02 6.75160885e-01 -4.69404340e-01
3.11605483e-02 -9.68779922e-01 2.10842416e-01 3.15767199e-01
-6.77333534e-01 -2.92068094e-01 -3.60827029e-01 3.61902416e-01
3.42685245e-02 8.82055890e-03 8.16775858e-01 -8.53901580e-02
2.56994605e-01 5.81384659e-01 -1.64449811e-01 -2.45356355e-02
9.56329048e-01 4.50284064e-01 -6.73454881e-01 -2.15980485e-01
-9.91877556e-01 5.39804399e-01 7.95045719e-02 4.17579301e-02
2.15601221e-01 -1.29309511e+00 -7.38054633e-01 3.33514929e-01
-4.53973748e-02 4.33806449e-01 4.60155666e-01 1.05403399e+00
-3.96969378e-01 8.28263983e-02 3.25339973e-01 -6.93370819e-01
-1.26324809e+00 7.10162997e-01 5.58513641e-01 -8.17558050e-01
-7.86944568e-01 1.04971278e+00 2.33178437e-01 -8.25650930e-01
6.69523120e-01 5.77047169e-02 -5.66395894e-02 3.15904409e-01
6.00513339e-01 6.08431734e-02 3.35266113e-01 -3.28452885e-01
-5.92300653e-01 1.85860932e-01 -5.32806575e-01 3.49483937e-01
1.63035274e+00 1.83740005e-01 -4.67343301e-01 2.53512740e-01
1.71377242e+00 -3.44110221e-01 -9.97494221e-01 -6.39110506e-01
-2.48203184e-02 8.15368071e-02 1.16976552e-01 -1.10554814e+00
-2.01326871e+00 5.73776901e-01 8.01522851e-01 -1.60902530e-01
1.26549339e+00 3.64167482e-01 8.65480840e-01 4.49817777e-01
3.99569660e-01 -1.03762388e+00 -2.64867753e-01 1.58283293e-01
5.52741110e-01 -1.07933617e+00 7.93239027e-02 -3.30725133e-01
-7.76587963e-01 8.90946269e-01 5.55560291e-01 1.97348654e-01
1.11932921e+00 7.16174424e-01 5.87528229e-01 -3.88853014e-01
-1.24979532e+00 3.78074162e-02 2.95221340e-02 5.16049683e-01
4.26913053e-01 1.39274269e-01 -5.12234829e-02 9.66480613e-01
1.14500262e-01 1.91024035e-01 -1.02259167e-01 1.14796090e+00
8.53553861e-02 -1.37355030e+00 1.99497547e-02 7.92405188e-01
-8.21087420e-01 -2.02044332e-03 -1.45418540e-01 3.46740603e-01
5.97696483e-01 8.48781049e-01 1.45844281e-01 -3.87145340e-01
1.20073602e-01 3.66328388e-01 4.25723061e-04 -6.01914525e-01
-1.12781906e+00 4.52219993e-01 9.84359011e-02 -9.01794434e-01
-3.55191648e-01 -4.74853188e-01 -1.39092481e+00 -2.70159721e-01
-5.89570165e-01 9.48413163e-02 5.58678448e-01 9.09354687e-01
5.53816080e-01 6.80627465e-01 6.25215650e-01 -3.77824903e-01
-6.66853666e-01 -1.01883304e+00 -5.64938664e-01 4.81541574e-01
3.79622817e-01 -3.30304265e-01 -2.15255886e-01 -3.03154793e-02] | [6.047317028045654, 6.448250770568848] |
114b8fee-309d-4b4e-b050-ac3af8f23a54 | timestamp-supervised-action-segmentation-in | 2212.11694 | null | https://arxiv.org/abs/2212.11694v2 | https://arxiv.org/pdf/2212.11694v2.pdf | Timestamp-Supervised Action Segmentation from the Perspective of Clustering | Video action segmentation under timestamp supervision has recently received much attention due to lower annotation costs. Most existing methods generate pseudo-labels for all frames in each video to train the segmentation model. However, these methods suffer from incorrect pseudo-labels, especially for the semantically unclear frames in the transition region between two consecutive actions, which we call ambiguous intervals. To address this issue, we propose a novel framework from the perspective of clustering, which includes the following two parts. First, pseudo-label ensembling generates incomplete but high-quality pseudo-label sequences, where the frames in ambiguous intervals have no pseudo-labels. Second, iterative clustering iteratively propagates the pseudo-labels to the ambiguous intervals by clustering, and thus updates the pseudo-label sequences to train the model. We further introduce a clustering loss, which encourages the features of frames within the same action segment more compact. Extensive experiments show the effectiveness of our method. | ['Fuchun Sun', 'Fanjiang Xu', 'Lingyu Si', 'Enhan Li', 'Dazhao Du'] | 2022-12-22 | null | null | null | null | ['action-segmentation'] | ['computer-vision'] | [ 5.20522058e-01 1.25104412e-01 -4.15032655e-01 -6.54455364e-01
-6.91473186e-01 -6.44670486e-01 2.32878283e-01 -3.49318646e-02
-3.04685324e-01 7.01013923e-01 2.90782183e-01 1.04833193e-01
1.35717645e-01 -3.16215664e-01 -6.55450284e-01 -7.30204105e-01
2.45543122e-01 3.21376204e-01 7.40207851e-01 3.74550283e-01
1.28500730e-01 -7.84156471e-02 -1.34386802e+00 4.14271832e-01
1.08820641e+00 8.93955529e-01 2.12185338e-01 1.24791943e-01
-2.43746713e-01 1.01284790e+00 -5.64118981e-01 -2.11729601e-01
2.40025178e-01 -8.48802567e-01 -1.04043925e+00 7.53696620e-01
8.20227042e-02 -6.11204743e-01 -2.84159124e-01 1.16969669e+00
-1.68629184e-01 2.86805540e-01 3.06779861e-01 -1.47581792e+00
-1.53639823e-01 7.97533035e-01 -7.64461398e-01 -6.26684949e-02
3.80250037e-01 1.12704933e-02 9.96259749e-01 -5.32438338e-01
6.58270895e-01 1.30202365e+00 5.61866879e-01 6.84900165e-01
-9.60400164e-01 -7.96750665e-01 6.06888831e-01 3.26631844e-01
-1.26897681e+00 -2.60068506e-01 6.17982447e-01 -4.33510214e-01
2.63221890e-01 2.23610308e-02 7.41163552e-01 1.00079370e+00
-4.22088712e-01 1.12488019e+00 8.99090886e-01 -1.85838521e-01
2.71950275e-01 -2.99134582e-01 1.51473641e-01 6.46036267e-01
-2.11041737e-02 -2.23542005e-01 -2.92777717e-01 1.01417946e-02
8.44728172e-01 2.53500760e-01 -3.41539592e-01 -2.06876531e-01
-1.39117634e+00 5.70879221e-01 6.35728389e-02 2.33032212e-01
-2.15822309e-01 1.67239413e-01 4.86752123e-01 -2.38855034e-01
3.67479742e-01 4.92141061e-02 -4.63113159e-01 -4.86477137e-01
-8.88126373e-01 -7.57061094e-02 4.69721019e-01 1.26200950e+00
9.08433735e-01 -3.61730427e-01 -2.81344324e-01 7.60152340e-01
3.64005417e-01 -3.06797549e-02 1.87743634e-01 -1.66445935e+00
6.03311598e-01 6.31580651e-01 3.74439836e-01 -8.62776279e-01
-1.66518599e-01 6.06286265e-02 -5.76265097e-01 -3.41554046e-01
5.95483661e-01 -2.20931068e-01 -9.08992052e-01 1.77547133e+00
6.38225973e-01 7.35243261e-01 -1.49242416e-01 1.07511365e+00
5.05366802e-01 7.00867534e-01 2.13155895e-01 -6.14953637e-01
1.06542969e+00 -1.48241091e+00 -1.09276295e+00 -1.28184140e-01
6.04276478e-01 -6.04983807e-01 1.02582967e+00 1.37964532e-01
-8.63864422e-01 -6.78964078e-01 -8.26736689e-01 9.98179764e-02
1.45987689e-01 1.25360340e-01 3.46227288e-01 1.95534050e-01
-6.74041927e-01 6.10977411e-01 -1.20851421e+00 -1.35571629e-01
5.71993411e-01 8.69802684e-02 -5.31995557e-02 -1.46096230e-01
-1.02736402e+00 1.99890345e-01 6.77904785e-01 9.23098996e-02
-8.54856253e-01 -3.66549730e-01 -9.84759212e-01 -1.40468851e-01
9.78891671e-01 -1.12308249e-01 1.48457706e+00 -1.49449313e+00
-1.39654279e+00 6.20954812e-01 -4.79454279e-01 -2.99049467e-01
6.57630682e-01 -3.99992228e-01 -2.96294451e-01 3.97214115e-01
4.17130738e-01 8.57452452e-01 9.06211674e-01 -1.45483601e+00
-1.18126070e+00 -9.20146480e-02 2.03196585e-01 3.04805458e-01
-1.10831872e-01 -2.32905839e-02 -1.13730729e+00 -8.85209620e-01
4.43180650e-01 -1.10353827e+00 -3.41948211e-01 -2.94738203e-01
-6.15960896e-01 -3.80912691e-01 8.87524486e-01 -5.72420657e-01
1.51574278e+00 -2.39899492e+00 7.13236481e-02 -1.60977200e-01
2.31000230e-01 2.05991283e-01 1.56259257e-02 -2.65112445e-02
4.40873094e-02 1.42908499e-01 -5.21823406e-01 -4.27369326e-01
-1.26712471e-01 4.58607405e-01 -2.88481504e-01 2.72941738e-01
1.34889603e-01 6.11246407e-01 -1.25740337e+00 -1.01289988e+00
1.46671221e-01 1.41363159e-01 -3.39975357e-01 3.25223982e-01
-4.98588860e-01 9.53339458e-01 -6.81777835e-01 6.55211091e-01
6.40731990e-01 -4.04571205e-01 1.08916275e-01 -1.70229211e-01
-1.34578288e-01 1.33359194e-01 -1.15441048e+00 1.87241006e+00
1.53204247e-01 3.76776785e-01 -5.72815053e-02 -7.91321278e-01
3.15953285e-01 3.65971327e-01 9.59185839e-01 -2.71234453e-01
5.08016720e-02 5.88171072e-02 -3.42206478e-01 -6.08763456e-01
4.54348117e-01 -3.18456404e-02 4.82682809e-02 6.16967022e-01
-2.88446814e-01 2.30393171e-01 5.43544054e-01 1.74918070e-01
8.46056521e-01 7.27807879e-01 -6.35979995e-02 2.91417956e-01
5.63058257e-01 1.50792266e-03 1.30909908e+00 4.31328118e-01
-6.19991302e-01 8.78981233e-01 5.69243312e-01 -2.25653142e-01
-8.14642727e-01 -7.40825057e-01 7.57235587e-02 9.53000605e-01
6.65875196e-01 -6.23411953e-01 -1.22511578e+00 -1.15287769e+00
-4.06442285e-01 4.52703595e-01 -3.49453211e-01 -7.27926847e-03
-8.27382743e-01 -4.46104705e-01 4.14885223e-01 6.08909667e-01
6.85135067e-01 -1.10365140e+00 -6.99680090e-01 3.68670493e-01
-7.67418683e-01 -1.49813163e+00 -8.31594348e-01 -2.27004021e-01
-9.80481863e-01 -1.31722939e+00 -6.38796687e-01 -7.79070556e-01
8.64480078e-01 3.93990189e-01 8.01187456e-01 2.04427332e-01
2.13580206e-01 1.39412910e-01 -7.34135032e-01 1.68879151e-01
-3.38118285e-01 -6.53736591e-02 -1.44851059e-01 3.63030523e-01
4.06889409e-01 -2.19832063e-01 -5.98477364e-01 8.19880068e-01
-9.79375303e-01 3.30168486e-01 3.96205634e-01 5.16120434e-01
1.04085219e+00 2.70880103e-01 5.03208280e-01 -1.03757548e+00
4.71928902e-02 -3.37463558e-01 -4.12550032e-01 3.12158287e-01
-1.84401840e-01 -1.15940936e-01 6.15342438e-01 -4.64331031e-01
-1.16361487e+00 4.41447139e-01 1.01791568e-01 -6.37576401e-01
-4.36680764e-01 2.81673044e-01 -5.47017395e-01 4.50940728e-01
1.64241344e-02 4.09417115e-02 -2.41030648e-01 -5.06362855e-01
2.98293144e-01 4.68903422e-01 6.75151289e-01 -5.47619462e-01
5.91619551e-01 6.55038536e-01 -3.53767931e-01 -3.91044497e-01
-1.30693138e+00 -6.65691376e-01 -8.08363259e-01 -3.79838258e-01
1.17565358e+00 -8.43502164e-01 -3.10052097e-01 6.31935179e-01
-1.05225813e+00 -4.94271278e-01 -2.97944844e-01 5.60739577e-01
-6.36836469e-01 8.92085493e-01 -7.98188925e-01 -6.10436022e-01
2.37985969e-01 -1.35350239e+00 1.11844647e+00 5.35766900e-01
-1.89766467e-01 -7.16892660e-01 -2.89407253e-01 4.52172786e-01
-3.79242152e-01 3.33999336e-01 6.76958859e-01 -4.93745059e-01
-6.68906569e-01 1.80073641e-02 -3.06408256e-01 4.14436638e-01
4.39918309e-01 1.23655297e-01 -6.84192538e-01 -2.53554024e-02
-2.19930690e-02 -3.06541532e-01 7.82886386e-01 3.63408327e-01
1.33970451e+00 -1.45956412e-01 -5.36600888e-01 4.67524081e-01
9.88994479e-01 5.66362262e-01 4.60295558e-01 1.99954212e-01
1.02262259e+00 6.46310687e-01 1.36362648e+00 4.97204542e-01
3.78075570e-01 6.64162576e-01 2.97958881e-01 1.63175222e-02
3.54999751e-02 -4.09199864e-01 3.97248656e-01 8.05337846e-01
5.77086136e-02 -1.91768453e-01 -6.67258918e-01 4.37174410e-01
-2.26050711e+00 -9.23814118e-01 -1.91435397e-01 2.25939178e+00
1.06241131e+00 3.53858143e-01 3.02679420e-01 1.36531323e-01
1.17274296e+00 2.57537723e-01 -7.70051062e-01 3.59708190e-01
1.74100429e-01 -3.07954580e-01 2.71001250e-01 3.53487283e-01
-1.39078414e+00 1.12362874e+00 5.32262897e+00 8.03025961e-01
-6.22983575e-01 1.82749033e-01 7.77402520e-01 1.08242661e-01
-1.49647202e-02 4.63540614e-01 -7.50377536e-01 8.11165869e-01
5.20292461e-01 2.33243763e-01 2.49695957e-01 7.58662999e-01
5.03980279e-01 -3.45687777e-01 -1.24370325e+00 8.80367100e-01
1.34886969e-02 -8.56963933e-01 -8.16447809e-02 -1.03702545e-01
9.85540509e-01 -3.82389843e-01 -3.99184585e-01 1.34635389e-01
2.56892622e-01 -5.30177474e-01 9.76316690e-01 3.66730064e-01
7.48998821e-01 -7.62769163e-01 5.18208086e-01 3.58199030e-01
-1.41669631e+00 -1.57814082e-02 -2.10595444e-01 6.18579648e-02
5.40923834e-01 5.09774983e-01 -4.56993759e-01 4.69635040e-01
5.62843382e-01 1.14070892e+00 -5.47704875e-01 1.05587208e+00
-4.67836857e-01 7.43069828e-01 -2.23424509e-01 4.27440882e-01
4.04494494e-01 -6.38222277e-01 2.22455174e-01 9.66508031e-01
8.58409777e-02 3.74285102e-01 8.01237285e-01 7.12949812e-01
-2.60277302e-03 -2.43334994e-01 -1.48115873e-01 -4.24915887e-02
8.03525150e-01 1.03824389e+00 -1.17270350e+00 -6.94923043e-01
-4.88377541e-01 1.17334509e+00 -2.97397245e-02 5.68329871e-01
-1.24497783e+00 -1.69109419e-01 3.89881909e-01 -2.51225401e-02
1.31764069e-01 -2.28745013e-01 -1.58581078e-01 -1.22819328e+00
2.04022959e-01 -7.83152580e-01 5.83078861e-01 -5.97764552e-01
-8.44510436e-01 3.52242231e-01 1.66660041e-01 -1.59415996e+00
-1.87552050e-01 9.45911929e-02 -2.84270793e-01 2.36120895e-01
-1.30361021e+00 -8.34011555e-01 -5.33479631e-01 3.99077326e-01
9.23976064e-01 4.25698668e-01 3.28043908e-01 5.27367651e-01
-8.71783137e-01 3.18815708e-01 -2.28363574e-02 3.46239507e-01
7.70427585e-01 -1.23683858e+00 1.85257182e-01 9.85799015e-01
6.35802969e-02 3.44095469e-01 3.74853998e-01 -8.75045359e-01
-5.61144829e-01 -1.36841238e+00 6.78206444e-01 -2.34548897e-01
3.89093280e-01 -1.84468552e-01 -8.82944882e-01 8.79458249e-01
-6.79551512e-02 -9.25105959e-02 6.04382515e-01 -2.99768060e-01
-8.55205506e-02 6.59490675e-02 -7.53626108e-01 5.20534635e-01
1.25273931e+00 -3.50742519e-01 -5.60392320e-01 4.84317809e-01
9.85923588e-01 -4.35725987e-01 -6.11444831e-01 5.07721066e-01
2.96028048e-01 -9.84753966e-01 6.78450108e-01 -1.85586572e-01
4.84206140e-01 -7.30962038e-01 1.64778426e-01 -9.52730954e-01
1.73115321e-02 -7.47773051e-01 -5.76024763e-02 1.62254488e+00
2.20824897e-01 -1.05888285e-01 9.31367815e-01 6.55095935e-01
-2.76210755e-01 -5.80455005e-01 -7.73067772e-01 -6.88310981e-01
-5.56945503e-01 -5.19137383e-01 5.14821589e-01 1.05733490e+00
-5.44319265e-02 2.75253683e-01 -4.02815759e-01 8.21674168e-02
5.17448843e-01 3.70309472e-01 5.78523278e-01 -1.04460001e+00
-7.82313049e-02 -1.73045292e-01 -4.85688299e-02 -1.46859908e+00
3.39030832e-01 -4.47993547e-01 6.85313284e-01 -1.44233513e+00
3.37307721e-01 -6.56980813e-01 -1.21749170e-01 6.46455467e-01
-5.53204477e-01 1.62691817e-01 1.84294462e-01 6.04341626e-01
-1.36523354e+00 5.28753996e-01 1.45550799e+00 6.74024299e-02
-3.09581459e-01 1.61831781e-01 -2.00078353e-01 1.11517906e+00
6.34540975e-01 -6.50832951e-01 -5.89682877e-01 -3.66585582e-01
-2.91565329e-01 1.85145855e-01 5.52377850e-02 -1.14935672e+00
-2.85980813e-02 -3.89899790e-01 7.38390908e-02 -6.93763137e-01
1.33584887e-01 -1.01671278e+00 2.02099681e-01 2.63929009e-01
-4.39987391e-01 -1.95963293e-01 -3.39956939e-01 8.90259266e-01
-4.48173076e-01 -4.20874923e-01 7.93980837e-01 -2.61382282e-01
-8.62615883e-01 4.88143086e-01 -2.49783069e-01 2.98058361e-01
1.40676367e+00 -3.86940390e-01 9.66201350e-02 -3.52092773e-01
-8.62257004e-01 5.77747345e-01 7.50826418e-01 4.76478696e-01
4.40629750e-01 -1.35822070e+00 -4.42837268e-01 -5.58670871e-02
-2.48401985e-02 6.72547460e-01 3.17294419e-01 8.93287063e-01
-4.47412372e-01 1.39902145e-01 1.75938800e-01 -7.82412648e-01
-1.17379105e+00 6.75792217e-01 1.18920028e-01 -6.80626854e-02
-7.90831327e-01 7.25341797e-01 5.06628573e-01 6.57594278e-02
5.15360653e-01 -5.37574589e-01 -2.99384356e-01 2.52295822e-01
5.21863461e-01 4.21746254e-01 -4.60322052e-01 -9.10772383e-01
-2.38617837e-01 5.95581234e-01 -8.16144496e-02 -1.81534275e-01
8.44658256e-01 -5.09823143e-01 -1.66438550e-01 6.58912182e-01
1.10691988e+00 -2.22326502e-01 -2.01196957e+00 -2.09930256e-01
3.52892071e-01 -6.59085274e-01 -4.77636039e-01 -4.37196404e-01
-1.27394962e+00 5.88013709e-01 2.34023005e-01 6.18790276e-02
1.21140122e+00 -7.70461746e-03 1.20741296e+00 8.69426206e-02
4.53746527e-01 -1.53334141e+00 2.59379297e-01 5.34134269e-01
2.21139252e-01 -1.04611635e+00 -2.10086718e-01 -8.27975929e-01
-9.39806104e-01 7.88950086e-01 7.99996614e-01 1.96849182e-01
2.59714127e-01 -9.99426693e-02 1.73697844e-01 7.43819997e-02
-3.15660357e-01 -2.32496247e-01 -1.16534032e-01 4.38379854e-01
1.09514996e-01 -1.28945097e-01 -5.58893442e-01 6.42352879e-01
3.93349558e-01 2.03533322e-01 5.65910399e-01 1.18144345e+00
-4.24920171e-01 -1.10819614e+00 -3.42248142e-01 2.44890854e-01
-4.94805604e-01 2.01128259e-01 -3.65182430e-01 4.68997031e-01
3.99982214e-01 1.17606330e+00 2.15950096e-03 -3.52891445e-01
1.98505461e-01 2.41627872e-01 5.91025837e-02 -7.09703028e-01
-1.98649168e-01 6.07203603e-01 -2.69900002e-02 -6.91965878e-01
-9.58927035e-01 -9.16981876e-01 -1.88119721e+00 9.20459032e-02
-5.26451409e-01 4.42450404e-01 2.78399177e-02 1.11930656e+00
2.51073897e-01 5.31552076e-01 7.08689034e-01 -7.39602685e-01
-2.02037036e-01 -9.22636449e-01 -5.64480662e-01 8.33011568e-01
1.93076625e-01 -6.69406235e-01 -2.92937487e-01 7.59686589e-01] | [8.523780822753906, 0.55406254529953] |
7c075d85-23ed-435a-9d28-13d0d3d1639d | attend-and-guide-ag-net-a-keypoints-driven | 2110.12183 | null | https://arxiv.org/abs/2110.12183v1 | https://arxiv.org/pdf/2110.12183v1.pdf | Attend and Guide (AG-Net): A Keypoints-driven Attention-based Deep Network for Image Recognition | This paper presents a novel keypoints-based attention mechanism for visual recognition in still images. Deep Convolutional Neural Networks (CNNs) for recognizing images with distinctive classes have shown great success, but their performance in discriminating fine-grained changes is not at the same level. We address this by proposing an end-to-end CNN model, which learns meaningful features linking fine-grained changes using our novel attention mechanism. It captures the spatial structures in images by identifying semantic regions (SRs) and their spatial distributions, and is proved to be the key to modelling subtle changes in images. We automatically identify these SRs by grouping the detected keypoints in a given image. The ``usefulness'' of these SRs for image recognition is measured using our innovative attentional mechanism focusing on parts of the image that are most relevant to a given task. This framework applies to traditional and fine-grained image recognition tasks and does not require manually annotated regions (e.g. bounding-box of body parts, objects, etc.) for learning and prediction. Moreover, the proposed keypoints-driven attention mechanism can be easily integrated into the existing CNN models. The framework is evaluated on six diverse benchmark datasets. The model outperforms the state-of-the-art approaches by a considerable margin using Distracted Driver V1 (Acc: 3.39%), Distracted Driver V2 (Acc: 6.58%), Stanford-40 Actions (mAP: 2.15%), People Playing Musical Instruments (mAP: 16.05%), Food-101 (Acc: 6.30%) and Caltech-256 (Acc: 2.59%) datasets. | ['Ardhendu Behera', 'Nik Bessis', 'Yonghuai Liu', 'Zachary Wharton', 'Asish Bera'] | 2021-10-23 | null | null | null | null | ['fine-grained-image-recognition'] | ['computer-vision'] | [-5.16677313e-02 -1.75873533e-01 -1.82559475e-01 -3.69176865e-01
-4.26513940e-01 -3.46807003e-01 7.71912515e-01 5.75850345e-02
-6.23626471e-01 4.73115504e-01 8.91546234e-02 9.04107764e-02
-1.53247267e-01 -6.14227414e-01 -8.58681023e-01 -6.95797622e-01
-5.44853993e-02 -2.53789667e-02 6.53674722e-01 -4.84436184e-01
5.21434367e-01 9.78206456e-01 -2.00782847e+00 2.89926440e-01
5.32290637e-01 1.47353244e+00 2.03279167e-01 6.46144986e-01
-5.05170226e-03 8.74125302e-01 -7.96764195e-01 -1.09777093e-01
3.44674170e-01 -1.01519421e-01 -6.07049823e-01 -8.30234438e-02
8.65037024e-01 -8.42199177e-02 -3.98648500e-01 1.15086603e+00
4.19546455e-01 1.22378081e-01 6.96540892e-01 -1.32509637e+00
-7.36817896e-01 -7.89147243e-02 -7.65928030e-01 8.66087854e-01
-1.89040080e-01 2.97168076e-01 7.02068746e-01 -9.31685328e-01
4.69375163e-01 1.31109118e+00 5.87509632e-01 3.35446835e-01
-6.16320729e-01 -9.58736062e-01 2.87289739e-01 5.60328066e-01
-1.35920548e+00 -3.30002010e-01 5.55028319e-01 -7.14016497e-01
1.14751744e+00 4.14903820e-01 3.36571187e-01 8.76504600e-01
6.41880572e-01 5.54422617e-01 9.87031579e-01 -1.66672751e-01
-8.03111121e-02 -7.10310340e-02 3.53422076e-01 6.84288144e-01
1.07010610e-01 3.72889817e-01 -5.52857161e-01 2.86595017e-01
6.44378483e-01 3.11961919e-01 -1.02072172e-01 -2.44117156e-01
-1.33977389e+00 6.16691649e-01 9.59950089e-01 4.17997628e-01
-4.05263245e-01 4.12094295e-01 4.90033358e-01 1.61874563e-01
3.59436989e-01 2.79906720e-01 -5.57769120e-01 -1.21412061e-01
-6.16102040e-01 2.08064586e-01 -4.34762752e-03 7.84705997e-01
7.95763373e-01 2.92951707e-02 -6.56601369e-01 7.19233096e-01
1.81525394e-01 3.74223441e-01 8.94096494e-01 -4.60234433e-01
2.28745043e-01 7.73012578e-01 1.56266764e-01 -1.31684554e+00
-4.97579724e-01 -7.90717125e-01 -1.04143858e+00 5.32058060e-01
3.02655518e-01 2.30450779e-01 -1.19506335e+00 1.41529214e+00
1.90469816e-01 1.32846445e-01 -1.68858513e-01 9.69210863e-01
1.01693439e+00 5.67534268e-01 1.52272314e-01 1.75420597e-01
1.57141054e+00 -1.13644612e+00 -3.64072263e-01 -2.39964411e-01
2.59002835e-01 -5.27689219e-01 1.13138127e+00 2.86024004e-01
-7.66060114e-01 -1.26608562e+00 -9.49823856e-01 -1.03434421e-01
-7.25741506e-01 2.45493531e-01 2.85552740e-01 2.71639138e-01
-9.55745637e-01 4.28182662e-01 -4.06804919e-01 -4.74458873e-01
5.99726856e-01 3.48079622e-01 -5.11094928e-01 8.16181973e-02
-8.95374894e-01 8.31165016e-01 4.52833652e-01 2.61599809e-01
-1.23721814e+00 -7.81172097e-01 -7.51720965e-01 8.17301795e-02
2.03173593e-01 -5.19784808e-01 9.10653353e-01 -1.16559458e+00
-9.97382700e-01 1.01210797e+00 -2.84046531e-02 -4.78045493e-01
4.50637817e-01 -3.16852778e-01 -5.23478031e-01 2.17768084e-02
1.74939528e-01 9.20400202e-01 1.14941835e+00 -1.10731077e+00
-1.10148489e+00 -4.47015435e-01 1.60125613e-01 2.30175778e-02
-1.76545441e-01 2.42853671e-01 -5.49397945e-01 -8.65237832e-01
-2.69493133e-01 -8.76344323e-01 -5.10297343e-02 1.15322746e-01
-4.26168323e-01 -4.95858669e-01 1.07600868e+00 -5.46643436e-01
9.14536715e-01 -2.23042703e+00 -5.62799200e-02 -3.93695831e-02
3.33822757e-01 5.16200304e-01 -1.95140839e-01 -4.93066665e-03
-2.97508508e-01 -1.82161126e-02 1.14278816e-01 1.02440929e-02
-1.92517042e-02 -6.01884164e-02 -1.62093118e-01 4.68795687e-01
4.23028886e-01 1.15329051e+00 -6.03361845e-01 -2.34115168e-01
4.47129726e-01 3.20565313e-01 -2.33626530e-01 1.47633612e-01
1.06214628e-01 2.84584194e-01 -3.31157029e-01 8.63971949e-01
6.00733519e-01 1.70914739e-01 -5.59329450e-01 -5.56131303e-01
-2.16204971e-01 -5.30480266e-01 -1.07726097e+00 1.29538214e+00
-2.63513118e-01 8.14472377e-01 -2.60531217e-01 -9.62791204e-01
1.26205719e+00 -1.97838411e-01 1.30060703e-01 -1.17662895e+00
1.41732499e-01 -1.63522027e-02 1.18297517e-01 -6.99547887e-01
5.60083091e-01 9.00527183e-03 -3.35960984e-01 -1.67968348e-01
5.35600632e-03 4.04955357e-01 -1.48365991e-02 -1.69524610e-01
1.05325735e+00 -8.85246247e-02 4.83109117e-01 -4.01134700e-01
6.52353525e-01 1.41817862e-02 6.41896963e-01 7.69303918e-01
-6.80951715e-01 4.58115071e-01 2.00248450e-01 -8.91120851e-01
-8.09754312e-01 -7.67265022e-01 1.42554082e-02 1.26298463e+00
4.34999913e-01 -1.27891630e-01 -4.46997702e-01 -7.99288392e-01
3.00509125e-01 3.16253155e-01 -1.10210705e+00 -4.36307281e-01
-4.15187329e-01 -3.43291789e-01 6.07223392e-01 8.60368371e-01
8.01384747e-01 -1.50850785e+00 -1.01083207e+00 -1.33709222e-01
6.05567619e-02 -9.52402174e-01 -4.16599929e-01 2.25486174e-01
-4.66607004e-01 -1.17475188e+00 -6.09318435e-01 -7.73178101e-01
5.69886446e-01 5.37255466e-01 1.16623878e+00 8.25677663e-02
-7.42741227e-01 1.73945397e-01 -2.64667779e-01 -5.09339631e-01
6.02288730e-02 -2.01243296e-01 -1.83339491e-02 3.40085328e-01
4.85121697e-01 4.09256592e-02 -7.57945120e-01 6.49688363e-01
-8.18480372e-01 -2.08943978e-01 8.95858228e-01 9.31645572e-01
7.34812617e-01 1.68531388e-01 5.48862517e-01 -6.09041452e-01
4.08005118e-01 -3.59555244e-01 -6.08274221e-01 8.20221081e-02
-2.81945497e-01 -8.17022696e-02 5.96109152e-01 -2.69603431e-01
-9.06273365e-01 5.00265770e-02 -1.98574085e-02 -7.15303302e-01
-8.01662207e-01 1.35227501e-01 -1.73754513e-01 -2.47680485e-01
6.89970374e-01 2.45604560e-01 -2.00406402e-01 -5.53344548e-01
2.98634410e-01 6.98325694e-01 9.25320804e-01 -1.01077378e-01
8.87647033e-01 3.86940777e-01 -7.51397982e-02 -7.19770193e-01
-7.26506114e-01 -9.01271403e-01 -8.30748439e-01 -1.99373409e-01
1.01475477e+00 -1.01093900e+00 -8.28854740e-01 7.05961585e-01
-7.59609222e-01 -1.56249285e-01 -1.96180046e-01 3.14407855e-01
-4.82146531e-01 2.04188287e-01 -2.20688045e-01 -4.26994383e-01
-3.76127273e-01 -1.15189195e+00 1.43222868e+00 4.97784913e-01
-5.66389151e-02 -5.33688068e-01 -1.86675817e-01 3.86396587e-01
5.76167881e-01 4.27808583e-01 8.08027983e-01 -3.92856926e-01
-5.19422114e-01 -2.15160206e-01 -6.12265050e-01 3.03550363e-01
1.19623147e-01 5.18361032e-02 -1.12964737e+00 -3.55939239e-01
-4.24151808e-01 -2.86788613e-01 1.12684655e+00 3.31255168e-01
1.59670734e+00 -5.93561828e-02 -6.57686055e-01 6.42085314e-01
1.45187020e+00 4.10331011e-01 8.84814858e-01 6.28317952e-01
8.27767372e-01 4.86321837e-01 9.40133333e-01 3.63524258e-01
1.13515660e-01 1.02536142e+00 9.12480116e-01 -2.83783585e-01
-4.31705356e-01 6.33241236e-02 3.90083462e-01 2.62656838e-01
-8.75939578e-02 -8.27870816e-02 -7.99424648e-01 8.12016070e-01
-1.79569745e+00 -1.07094145e+00 -1.79288983e-01 1.88100803e+00
4.34749186e-01 3.43043178e-01 1.55400977e-01 3.56084527e-03
6.84421718e-01 1.15959056e-01 -6.90308452e-01 -4.51892346e-01
7.81643987e-02 2.18006253e-01 5.88808656e-01 -5.36919311e-02
-1.47083771e+00 9.99656379e-01 4.94335938e+00 9.45143938e-01
-1.27960312e+00 1.74530111e-02 6.96559906e-01 8.36234614e-02
3.71508360e-01 -5.52886307e-01 -1.02120531e+00 6.08180225e-01
8.29601228e-01 1.67211816e-01 -4.58472259e-02 8.78954291e-01
2.88134485e-01 4.35197465e-02 -7.84132540e-01 1.18461883e+00
2.59582371e-01 -1.30163753e+00 -2.44619418e-02 -1.59700960e-01
6.38822496e-01 1.49157286e-01 2.18889996e-01 5.14433801e-01
-8.15235078e-02 -1.33701468e+00 9.01820362e-01 8.13156426e-01
8.91996324e-01 -8.63134444e-01 1.04510987e+00 3.60391140e-01
-1.29199803e+00 -5.16802728e-01 -6.51215672e-01 3.34181786e-02
-4.43249673e-01 3.33319575e-01 -5.93890607e-01 4.42814112e-01
1.39548218e+00 8.49784255e-01 -1.09174716e+00 1.15831327e+00
5.79624809e-02 4.36908871e-01 8.00606161e-02 7.23920092e-02
5.12601495e-01 2.84072518e-01 3.37000310e-01 1.45438612e+00
1.98732972e-01 -1.78689152e-01 -1.94735136e-02 6.54655159e-01
-1.03534289e-01 -9.06202868e-02 -5.51664293e-01 3.17814112e-01
8.34788755e-02 1.49439037e+00 -7.36196041e-01 -2.87114561e-01
-3.41461658e-01 9.86913204e-01 2.81080484e-01 1.38455242e-01
-9.41167772e-01 -7.86236107e-01 1.08570361e+00 6.04840852e-02
8.19054961e-01 1.84538518e-03 2.06273869e-01 -8.12910259e-01
8.29849858e-03 -9.79558647e-01 3.86508465e-01 -9.01857018e-01
-1.20357537e+00 8.06105733e-01 -1.60886765e-01 -1.34146142e+00
-5.44213280e-02 -7.72503555e-01 -6.41981781e-01 9.17415082e-01
-1.85924530e+00 -1.13657534e+00 -1.15812492e+00 9.20880079e-01
7.58415699e-01 -3.73591036e-01 5.90668082e-01 3.33357245e-01
-5.88873148e-01 8.53795528e-01 1.18706383e-01 2.58774251e-01
6.97627723e-01 -1.19530153e+00 5.63611984e-01 8.78705442e-01
1.70728366e-03 3.21886599e-01 4.57334340e-01 -4.17839199e-01
-1.14915979e+00 -1.72084367e+00 6.31165981e-01 -5.01594245e-01
4.50545102e-01 -1.72759607e-01 -8.39410007e-01 4.40876812e-01
9.06491354e-02 6.35895550e-01 2.23767027e-01 -1.56764403e-01
-5.82244396e-01 -5.27434587e-01 -1.10794091e+00 3.05058718e-01
1.09859836e+00 -3.82163942e-01 -6.17252946e-01 1.77009284e-01
5.42489886e-01 -3.69782090e-01 -6.92067087e-01 4.30931330e-01
4.84304428e-01 -1.08442986e+00 1.13314068e+00 -7.98337340e-01
2.19818428e-01 -6.80807829e-01 -3.95772010e-02 -1.36213553e+00
-7.94250190e-01 -1.77831814e-01 2.51211729e-02 9.60837483e-01
-1.51873410e-01 -5.06470203e-01 6.15956724e-01 -5.65637946e-02
-4.75129217e-01 -5.62761128e-01 -9.31966901e-01 -8.68451893e-01
-8.55745524e-02 -3.19360435e-01 5.14816225e-01 6.41578555e-01
-5.65200269e-01 1.49064898e-01 -3.23429227e-01 2.88841814e-01
3.97275299e-01 1.20722733e-01 9.16212618e-01 -1.33671999e+00
-6.56281039e-02 -6.38594747e-01 -1.17127633e+00 -7.24531472e-01
8.78493711e-02 -6.27083778e-01 1.35053650e-01 -1.31996191e+00
2.86838710e-01 -3.64993215e-01 -7.84894764e-01 7.30427325e-01
-2.43527427e-01 7.26698220e-01 3.34725857e-01 2.30125874e-01
-7.49684334e-01 4.39293623e-01 1.07696116e+00 -5.74087977e-01
1.99004069e-01 -6.84135482e-02 -7.36751437e-01 5.80823481e-01
8.77606988e-01 -1.93707705e-01 -2.27883250e-01 -1.30787164e-01
-1.75336808e-01 -4.36716199e-01 8.98828804e-01 -1.36384749e+00
2.02821404e-01 -9.01161358e-02 7.25593209e-01 -7.11608231e-01
1.01727143e-01 -9.32430625e-01 2.00097919e-01 7.24850476e-01
-2.96263844e-01 1.82314649e-01 6.03086233e-01 6.80974305e-01
-4.41753596e-01 -2.62997579e-02 9.60893691e-01 2.15056986e-02
-1.60422027e+00 3.77744496e-01 -1.67369440e-01 6.54423889e-03
1.53673339e+00 -3.99735838e-01 -4.44640249e-01 -7.47891366e-02
-4.99807686e-01 -8.23881030e-02 1.29479736e-01 1.00782692e+00
7.04104066e-01 -1.40242839e+00 -6.04542613e-01 3.65090281e-01
6.85858727e-01 -2.47618437e-01 4.78756011e-01 6.86596274e-01
-4.83658642e-01 6.87302649e-01 -7.40498304e-01 -7.25958586e-01
-1.60969806e+00 7.49302685e-01 6.58005595e-01 -4.25327336e-03
-6.57613337e-01 9.57665801e-01 6.22142434e-01 -5.43205328e-02
2.98557490e-01 -7.85847306e-01 -4.74250704e-01 2.70417798e-02
6.62215889e-01 3.00514072e-01 3.58892292e-01 -1.14597690e+00
-6.22360885e-01 1.02710795e+00 -1.67738914e-01 7.88762212e-01
1.17443156e+00 -3.13461348e-02 1.45915538e-01 2.42619172e-01
1.27977884e+00 -1.55486479e-01 -1.36219740e+00 -2.41032884e-01
-1.90808013e-01 -6.14609718e-01 7.40535706e-02 -1.05326080e+00
-1.25206113e+00 1.02103293e+00 1.19916987e+00 1.24963727e-02
1.16453576e+00 8.57710838e-02 6.42277718e-01 7.54448175e-02
2.86275804e-01 -8.35226357e-01 2.29493409e-01 5.43483138e-01
1.10879910e+00 -1.43185699e+00 -2.87673861e-01 -1.01445653e-01
-6.24191284e-01 1.15375519e+00 8.74795616e-01 -1.16365187e-01
6.24222815e-01 -2.89227180e-02 1.46585420e-01 -4.55577344e-01
-6.71690524e-01 -4.54475313e-01 7.99014091e-01 7.37932265e-01
1.49444491e-01 1.69243351e-01 1.67006120e-01 4.37524736e-01
1.89207383e-02 -1.51691258e-01 1.78826049e-01 6.45190835e-01
-6.84419453e-01 -4.32738900e-01 -4.69680160e-01 4.85511184e-01
-3.99996281e-01 1.06206484e-01 -3.50172371e-01 9.16148424e-01
7.92753875e-01 7.65443504e-01 4.29497361e-01 -5.24484873e-01
6.24569237e-01 -2.13562578e-01 8.83120894e-02 -4.74336267e-01
-7.98133969e-01 -3.13706577e-01 -2.55629361e-01 -9.61775482e-01
-3.80304188e-01 -4.24787074e-01 -9.81227338e-01 -1.09868653e-01
1.82136502e-02 -2.11063266e-01 3.69869381e-01 7.30837405e-01
6.51219845e-01 9.05716479e-01 4.67225015e-01 -8.71016562e-01
-3.94542485e-01 -1.02946711e+00 -5.23736417e-01 6.38813019e-01
5.41874111e-01 -1.03786027e+00 -1.45844504e-01 8.35419968e-02] | [9.599624633789062, 2.009425163269043] |
ef565d20-05dc-4171-95ee-4b6c118d8d01 | unified-open-domain-question-answering-with | 2012.14610 | null | https://arxiv.org/abs/2012.14610v3 | https://arxiv.org/pdf/2012.14610v3.pdf | UniK-QA: Unified Representations of Structured and Unstructured Knowledge for Open-Domain Question Answering | We study open-domain question answering with structured, unstructured and semi-structured knowledge sources, including text, tables, lists and knowledge bases. Departing from prior work, we propose a unifying approach that homogenizes all sources by reducing them to text and applies the retriever-reader model which has so far been limited to text sources only. Our approach greatly improves the results on knowledge-base QA tasks by 11 points, compared to latest graph-based methods. More importantly, we demonstrate that our unified knowledge (UniK-QA) model is a simple and yet effective way to combine heterogeneous sources of knowledge, advancing the state-of-the-art results on two popular question answering benchmarks, NaturalQuestions and WebQuestions, by 3.5 and 2.6 points, respectively. The code of UniK-QA is available at: https://github.com/facebookresearch/UniK-QA. | ['Scott Yih', 'Yashar Mehdad', 'Sonal Gupta', 'Michael Schlichtkrull', 'Dmytro Okhonko', 'Stan Peshterliev', 'Vladimir Karpukhin', 'Xilun Chen', 'Barlas Oguz'] | 2020-12-29 | null | https://aclanthology.org/2022.findings-naacl.115 | https://aclanthology.org/2022.findings-naacl.115.pdf | findings-naacl-2022-7 | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-2.29473352e-01 4.65742201e-01 -2.87489146e-01 1.50397616e-02
-1.53614902e+00 -1.10376632e+00 7.61829317e-01 6.22193694e-01
-2.33820140e-01 9.81248677e-01 5.12082636e-01 -4.91697580e-01
-5.19969404e-01 -1.05523741e+00 -8.80179107e-01 4.78953533e-02
4.32504326e-01 1.00075138e+00 8.83687556e-01 -7.88689077e-01
2.17959717e-01 -1.22700907e-01 -1.36299360e+00 6.74418509e-01
1.32999337e+00 8.93242419e-01 -2.03444123e-01 6.56037986e-01
-7.38418102e-01 1.24647868e+00 -4.81876135e-01 -1.07504404e+00
-8.95272270e-02 -2.30762839e-01 -1.74272501e+00 -5.41155040e-01
9.30645645e-01 -7.64878765e-02 -5.34858644e-01 8.93526137e-01
3.73336792e-01 2.51842234e-02 4.28820580e-01 -9.98549163e-01
-1.17248011e+00 9.46581364e-01 -2.28259757e-01 1.31235912e-01
9.65891302e-01 -3.04437488e-01 1.44103718e+00 -7.03934848e-01
8.08074474e-01 1.21862376e+00 4.97746557e-01 3.33937794e-01
-1.05906093e+00 -1.69040918e-01 -6.15344048e-02 7.51365006e-01
-1.31835639e+00 -5.11385143e-01 5.13939738e-01 -2.10554436e-01
1.16446638e+00 4.47409391e-01 1.45840809e-01 6.94404066e-01
-3.06069463e-01 7.95959413e-01 1.09455633e+00 -6.06267810e-01
-9.92575660e-02 1.09559223e-02 7.84722030e-01 7.58324206e-01
4.91830915e-01 -6.43782139e-01 -7.08486736e-01 -4.65113997e-01
2.30022788e-01 -3.58085304e-01 -5.46621263e-01 -3.07582021e-01
-1.18842161e+00 5.72394431e-01 4.07645077e-01 2.58716464e-01
-4.12110955e-01 -5.92459626e-02 2.40301609e-01 6.11069262e-01
3.19374919e-01 7.49401093e-01 -7.75740147e-01 -4.80125509e-02
-6.29405677e-01 4.55841988e-01 1.49707377e+00 1.17289710e+00
9.36962306e-01 -8.58970702e-01 -5.06422043e-01 1.02609384e+00
2.23974168e-01 8.30692291e-01 3.02273691e-01 -1.21442187e+00
7.82449007e-01 1.06573069e+00 3.05747747e-01 -9.27443206e-01
-2.22931579e-01 -4.28813010e-01 -2.98359632e-01 -6.65081084e-01
6.66377246e-01 4.81824689e-02 -7.67587960e-01 1.33484828e+00
5.86951911e-01 -1.87613487e-01 4.49334383e-01 5.53005695e-01
1.50508273e+00 5.82214653e-01 -3.51149067e-02 5.65897562e-02
1.78860080e+00 -1.32220256e+00 -8.92659605e-01 -6.86968043e-02
6.53890550e-01 -7.29795396e-01 1.22539186e+00 1.20082654e-01
-1.09489977e+00 -4.45841737e-02 -5.19637764e-01 -7.02179432e-01
-9.06923950e-01 -2.13001415e-01 3.41488361e-01 5.01079977e-01
-1.29618835e+00 2.44698569e-01 -3.83803844e-01 -4.79079515e-01
3.63045841e-01 -6.87946007e-02 -2.23251238e-01 -6.11497521e-01
-1.60396039e+00 9.53018904e-01 4.60969537e-01 -4.99062687e-01
-6.56264305e-01 -1.10205150e+00 -5.75933397e-01 1.61444247e-01
1.10292685e+00 -1.32487929e+00 1.61223876e+00 -2.25807250e-01
-1.35806715e+00 7.95330822e-01 -2.67938524e-01 -4.01301712e-01
1.16039872e-01 -4.32425857e-01 -4.57279086e-01 4.63432759e-01
2.73242921e-01 3.20529491e-01 3.70012850e-01 -1.33843565e+00
-4.20420319e-01 -4.58663732e-01 7.33758867e-01 2.83619344e-01
-3.11462849e-01 -2.93196496e-02 -8.50811660e-01 -2.49216303e-01
-2.20350146e-01 -5.31430066e-01 2.08597109e-01 -2.34564304e-01
-4.83467340e-01 -7.39241600e-01 4.10450339e-01 -1.01649618e+00
1.53276026e+00 -1.42892897e+00 2.02966526e-01 8.65974426e-02
6.20383680e-01 2.69719690e-01 -1.48715928e-01 1.01999199e+00
4.25441653e-01 6.88558891e-02 -2.99735278e-01 5.64630069e-02
3.96718413e-01 3.02508026e-01 -5.13628662e-01 -3.91182750e-01
-9.27125216e-02 1.50092185e+00 -1.14632249e+00 -7.44558811e-01
-2.81657934e-01 1.95354700e-01 -3.93288404e-01 -2.88977828e-02
-9.05015647e-01 -1.54298827e-01 -6.18030190e-01 7.58132458e-01
3.71585935e-01 -8.57284784e-01 1.03553496e-01 -1.10133350e-01
4.09470767e-01 8.18547904e-01 -9.00574446e-01 2.01560998e+00
-4.54169959e-01 2.48577192e-01 -8.63712206e-02 -4.22666073e-01
6.07954681e-01 2.94264764e-01 1.56152785e-01 -8.25570583e-01
-2.22722724e-01 3.47010612e-01 -5.18664062e-01 -5.92732728e-01
6.57315195e-01 3.75056028e-01 1.15855865e-01 4.24368858e-01
2.93215632e-01 -2.74892569e-01 6.92293584e-01 1.00397980e+00
1.49409103e+00 -2.63786763e-02 9.40971971e-02 -2.80162245e-01
5.30578434e-01 4.68050808e-01 -2.13482454e-01 9.75326240e-01
3.06750834e-01 2.15394780e-01 6.19641066e-01 1.94408461e-01
-5.42888165e-01 -1.19290447e+00 1.05506867e-01 1.33399951e+00
2.27112606e-01 -9.44344044e-01 -6.74166918e-01 -1.04618990e+00
2.59082794e-01 9.36713040e-01 -4.74736333e-01 8.70530382e-02
-3.47547024e-01 -2.20322803e-01 6.95842981e-01 3.71130556e-01
4.05245900e-01 -8.07484150e-01 1.19599678e-01 1.33342460e-01
-8.68509114e-01 -1.25856686e+00 -1.89190328e-01 -2.61406004e-01
-6.79925501e-01 -1.30985010e+00 -7.16136932e-01 -4.72173363e-01
2.51621127e-01 3.57776284e-01 2.03890872e+00 2.64042079e-01
1.23388670e-01 1.07207596e+00 -7.90909469e-01 -2.89230734e-01
-2.22398564e-01 4.86017972e-01 -5.85916936e-01 -3.26221198e-01
3.32589716e-01 -4.16136473e-01 -6.72148466e-01 1.19545184e-01
-1.08652163e+00 -1.58261657e-01 4.11835283e-01 4.66594219e-01
4.90731031e-01 -2.55958527e-01 1.00788021e+00 -1.07288587e+00
8.46012294e-01 -8.63780499e-01 -5.76690853e-01 9.31055844e-01
-9.25453246e-01 3.33347946e-01 4.50989246e-01 8.79392028e-02
-1.20636988e+00 -4.92801338e-01 -1.65653050e-01 6.98249638e-02
-5.02370708e-02 8.18708241e-01 -3.99216227e-02 -1.15511745e-01
9.06487465e-01 9.23925862e-02 -4.03355211e-01 -7.96297729e-01
1.09707296e+00 6.55707777e-01 4.57321405e-01 -8.65680873e-01
9.02612209e-01 3.95440727e-01 -2.86475569e-01 -5.26040554e-01
-1.26235628e+00 -9.02032375e-01 -4.18285280e-01 -7.76105002e-02
6.12590253e-01 -7.69996166e-01 -6.51919365e-01 1.12380922e-01
-1.21203399e+00 -3.36835295e-01 -4.06149298e-01 -1.66662753e-01
-2.13541552e-01 7.72871673e-01 -5.89928091e-01 -2.74867982e-01
-6.47281587e-01 -4.39489603e-01 1.03090096e+00 1.45198748e-01
-3.80054116e-03 -1.26168668e+00 3.73606920e-01 1.23183835e+00
5.18616915e-01 -1.46730050e-01 1.16671145e+00 -1.02192080e+00
-7.72443712e-01 -8.09161663e-02 -3.89016181e-01 1.25324026e-01
-1.16699837e-01 -4.02572721e-01 -7.19845831e-01 1.12510897e-01
-5.20159364e-01 -8.17138672e-01 1.00429142e+00 -2.14100197e-01
7.73650408e-01 -4.87976551e-01 -3.43411416e-01 -1.08366676e-01
1.52835214e+00 -2.72506058e-01 7.04998195e-01 4.54023689e-01
7.02561557e-01 6.98933125e-01 2.66523361e-01 2.43444797e-02
1.24672019e+00 5.31326532e-01 4.13757652e-01 3.96530002e-01
-5.92263043e-01 -5.00449300e-01 4.88930941e-02 1.10860097e+00
3.28134783e-02 -4.95689243e-01 -1.23250031e+00 1.01584446e+00
-1.94056571e+00 -7.46314526e-01 -4.46044356e-01 1.93491852e+00
1.29810464e+00 -1.42563179e-01 2.96680331e-02 -1.57478526e-01
3.63391370e-01 9.36651900e-02 -3.54436845e-01 1.47676259e-01
-2.73696542e-01 4.80711669e-01 3.60715419e-01 7.33208001e-01
-6.29798055e-01 9.53238308e-01 5.89999533e+00 1.16748047e+00
-2.05768883e-01 3.23348850e-01 4.27062996e-02 2.65693307e-01
-8.46877873e-01 2.08847508e-01 -8.70782435e-01 1.37282565e-01
1.03117383e+00 -5.04939556e-01 5.24177969e-01 5.31509936e-01
-5.80045998e-01 -9.50635746e-02 -9.11595047e-01 5.68232059e-01
2.38615200e-01 -1.64002216e+00 4.05466318e-01 -4.03454542e-01
8.30071151e-01 2.63806731e-01 -2.48338297e-01 6.03786528e-01
6.81288183e-01 -7.75844514e-01 2.65938908e-01 7.58644283e-01
4.36246157e-01 -2.65003294e-01 7.53623784e-01 2.97952056e-01
-1.01496410e+00 1.62285686e-01 -1.00977145e-01 1.01649858e-01
1.17947318e-01 5.82116246e-01 -6.03637695e-01 1.44572735e+00
8.33601058e-01 4.58121270e-01 -8.85650396e-01 9.32182252e-01
-3.40020299e-01 9.01309550e-01 -3.88155371e-01 -1.93734095e-01
9.74672437e-02 1.47205457e-01 3.52826267e-01 1.09355021e+00
-9.86399278e-02 3.09011132e-01 4.00836766e-02 5.38282514e-01
-6.53824568e-01 4.50948477e-01 -2.83148289e-01 -7.06344172e-02
6.42108083e-01 1.15457654e+00 -2.80745178e-01 -7.15043247e-01
-6.83305919e-01 7.82019854e-01 7.28576005e-01 4.65267628e-01
-5.34892380e-01 -8.04776609e-01 1.26076834e-02 6.02834411e-02
3.96568686e-01 -3.89379226e-02 -2.13722382e-02 -1.42031431e+00
4.93942410e-01 -1.06055510e+00 1.05390358e+00 -1.10609961e+00
-1.59384584e+00 6.07257068e-01 1.69761479e-01 -5.26303470e-01
-2.44014606e-01 -3.99334908e-01 9.03008357e-02 8.07649612e-01
-2.14482737e+00 -1.12557828e+00 -4.34395343e-01 8.79894793e-01
2.01169252e-01 1.85481742e-01 9.18991625e-01 5.12178004e-01
-8.79735351e-02 4.26174879e-01 3.61814380e-01 2.43260860e-01
8.12990367e-01 -1.56848776e+00 3.29686284e-01 5.41986465e-01
5.43669522e-01 6.29116714e-01 3.68064761e-01 -5.48943162e-01
-1.75014985e+00 -8.59121621e-01 1.31839144e+00 -1.33792686e+00
1.21305215e+00 -6.71950579e-02 -1.23396599e+00 6.61934614e-01
7.01665699e-01 -2.28436410e-01 6.62395656e-01 3.82032901e-01
-9.36430156e-01 -1.69801950e-01 -7.95955956e-01 4.30193990e-01
1.03004837e+00 -6.09107792e-01 -1.26748312e+00 8.54018986e-01
1.28753066e+00 -4.74864811e-01 -1.39234972e+00 2.98279017e-01
1.88930348e-01 -6.03495777e-01 1.11058152e+00 -7.39626646e-01
5.01214445e-01 -3.36826771e-01 -2.01472536e-01 -1.17604172e+00
-1.09897763e-01 -6.25459194e-01 -6.75230443e-01 1.34018481e+00
8.39562237e-01 -8.55439842e-01 6.12593234e-01 4.60516214e-01
-1.42594755e-01 -7.17225134e-01 -9.29239571e-01 -8.34158599e-01
2.37482443e-01 -1.91466451e-01 6.03230178e-01 9.54227507e-01
2.28614450e-01 7.91884363e-01 1.40656933e-01 2.72974700e-01
5.19367516e-01 3.41810465e-01 5.84206760e-01 -1.16818130e+00
-2.17365906e-01 -4.17884558e-01 1.07076518e-01 -1.43397367e+00
1.82021737e-01 -1.13505936e+00 -3.98037106e-01 -2.52864766e+00
2.04675674e-01 -3.47082049e-01 -2.52936512e-01 6.83098972e-01
-4.20626193e-01 1.73429683e-01 1.09645672e-01 2.39029601e-01
-1.46858835e+00 5.37906170e-01 1.30145800e+00 -1.67245910e-01
2.10666239e-01 -2.93045163e-01 -1.08307719e+00 2.52596408e-01
7.36050189e-01 -4.03007269e-01 -5.97278893e-01 -9.16905642e-01
7.66082048e-01 2.35803679e-01 4.82936263e-01 -6.82771206e-01
7.03250885e-01 6.51792288e-02 -3.93496513e-01 -4.29222912e-01
3.52911085e-01 -7.27640629e-01 -3.37450236e-01 6.96313456e-02
-4.91986245e-01 4.80069518e-02 2.80313522e-01 6.89722955e-01
-4.31652874e-01 -2.44604036e-01 1.61736198e-02 -2.66620249e-01
-6.55730724e-01 1.58151552e-01 4.85333614e-02 1.03610003e+00
4.07175809e-01 4.69288558e-01 -1.27761638e+00 -7.34467983e-01
-4.30801153e-01 6.58395529e-01 4.61935960e-02 4.49103266e-01
4.43115383e-01 -1.17955816e+00 -9.44683552e-01 -6.70220852e-01
5.09201825e-01 -8.94867107e-02 2.85874218e-01 8.56831133e-01
-5.06461084e-01 1.00790751e+00 3.46826226e-01 -2.19637200e-01
-1.01117921e+00 6.57445073e-01 1.15628965e-01 -6.79895818e-01
-3.16628426e-01 6.86584651e-01 -3.52620274e-01 -7.37744987e-01
6.28029704e-02 -3.23246270e-01 -4.54540849e-01 -2.39191186e-02
4.72107381e-01 6.93116903e-01 5.37307739e-01 -2.18545899e-01
-3.83407831e-01 5.77166617e-01 -5.97764440e-02 4.28301692e-02
1.01792037e+00 -3.13995123e-01 -6.63380325e-01 2.49523938e-01
8.49978507e-01 2.69399405e-01 -3.22129399e-01 -8.97982180e-01
3.21303815e-01 -1.06870912e-01 -1.84026912e-01 -1.50892651e+00
-5.34125090e-01 4.97572780e-01 -1.25063822e-01 6.45996451e-01
1.09447420e+00 6.03341758e-01 9.58027065e-01 1.00744259e+00
4.23863888e-01 -6.06465101e-01 1.85924061e-02 7.78261840e-01
9.82330620e-01 -1.16401792e+00 -1.60143971e-01 -6.38424814e-01
-5.42632282e-01 7.75622189e-01 3.88004154e-01 3.60793740e-01
5.28229713e-01 -1.47351369e-01 1.82598665e-01 -6.46778166e-01
-1.03569186e+00 -5.54923296e-01 7.48239338e-01 5.54794014e-01
2.41404220e-01 -1.20542556e-01 -2.05681205e-01 6.76715076e-01
-3.22462469e-01 1.18841112e-01 2.90938765e-01 9.38596785e-01
-5.14990985e-01 -1.16272795e+00 -2.53369242e-01 5.51468313e-01
-5.75865805e-01 -5.59760630e-01 -6.55818522e-01 7.38158405e-01
-4.31213289e-01 1.30641210e+00 -4.22634691e-01 -6.73565641e-02
5.21979094e-01 4.53264207e-01 4.87864763e-01 -6.46854281e-01
-6.04145467e-01 -6.72606647e-01 6.39671981e-01 -5.18783391e-01
-5.25577009e-01 -1.87449798e-01 -1.00740540e+00 -3.43455076e-01
-3.93233299e-01 6.95179522e-01 2.57981986e-01 7.75879860e-01
9.40948188e-01 3.34565639e-01 -1.39052823e-01 3.37103933e-01
-5.32154381e-01 -8.32832396e-01 -1.16510421e-01 3.38394731e-01
2.14751244e-01 -4.74473745e-01 -1.37519941e-01 1.40657127e-01] | [10.699372291564941, 7.954188346862793] |
8e4c6ca7-af17-4a54-8c76-26da59a6f4b7 | self-supervised-learning-of-multi-object | 2205.08316 | null | https://arxiv.org/abs/2205.08316v2 | https://arxiv.org/pdf/2205.08316v2.pdf | Self-Supervised Learning of Multi-Object Keypoints for Robotic Manipulation | In recent years, policy learning methods using either reinforcement or imitation have made significant progress. However, both techniques still suffer from being computationally expensive and requiring large amounts of training data. This problem is especially prevalent in real-world robotic manipulation tasks, where access to ground truth scene features is not available and policies are instead learned from raw camera observations. In this paper, we demonstrate the efficacy of learning image keypoints via the Dense Correspondence pretext task for downstream policy learning. Extending prior work to challenging multi-object scenes, we show that our model can be trained to deal with important problems in representation learning, primarily scale-invariance and occlusion. We evaluate our approach on diverse robot manipulation tasks, compare it to other visual representation learning approaches, and demonstrate its flexibility and effectiveness for sample-efficient policy learning. | ['Abhinav Valada', 'Tim Welschehold', 'Eugenio Chisari', 'Jan Ole von Hartz'] | 2022-05-17 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 3.45676333e-01 -6.79404885e-02 -7.01275408e-01 -1.69010554e-02
-7.35052228e-01 -8.22163165e-01 8.39286447e-01 1.89019546e-01
-5.16378582e-01 8.70056272e-01 3.21720578e-02 -1.50606185e-01
-2.51830548e-01 -3.98646683e-01 -8.96888793e-01 -6.32096231e-01
-6.45252168e-02 7.93535411e-01 2.13813499e-01 -1.49923578e-01
5.03548861e-01 9.60513830e-01 -1.40422392e+00 -1.55941978e-01
5.09891450e-01 8.65327835e-01 4.69926149e-01 4.84655052e-01
2.04822093e-01 1.03852069e+00 -3.48522127e-01 3.51036131e-01
5.60222745e-01 -8.33894685e-02 -7.40806162e-01 3.79242420e-01
6.51129067e-01 -5.81038833e-01 -5.36104739e-01 9.63228405e-01
2.38779128e-01 5.26681185e-01 7.91466773e-01 -1.32138515e+00
-3.83639842e-01 7.25032762e-02 -7.18034923e-01 -4.20895442e-02
2.60729909e-01 3.26361060e-01 8.48119199e-01 -7.10162580e-01
9.27228272e-01 1.56321609e+00 3.63950133e-01 4.80158120e-01
-1.50549090e+00 -5.03432751e-01 3.60292196e-01 3.05451304e-01
-7.21026421e-01 -4.01541293e-01 6.43188000e-01 -5.53000808e-01
1.00393832e+00 -4.31914777e-01 5.48400819e-01 1.31853282e+00
2.06975311e-01 8.87452126e-01 1.28707147e+00 -4.06506896e-01
2.85001665e-01 -1.83358192e-01 -3.03565115e-01 8.29457641e-01
1.45778567e-01 5.84629059e-01 -2.74410397e-01 -1.79524750e-01
1.37653351e+00 1.56305686e-01 -2.24752486e-01 -1.28270698e+00
-1.44636452e+00 1.01345718e+00 5.98030925e-01 -8.11306760e-02
-4.75878596e-01 6.67803586e-01 3.13006759e-01 3.95164579e-01
5.91340624e-02 7.54466116e-01 -4.94755715e-01 -2.62109190e-01
-4.25448149e-01 5.68156004e-01 7.66062975e-01 1.20258701e+00
8.33345473e-01 9.06910300e-02 -4.91831936e-02 7.37835169e-01
1.23611495e-01 6.40176773e-01 2.91319907e-01 -1.57720292e+00
2.62094408e-01 2.03085437e-01 3.92591715e-01 -8.05947840e-01
-2.76268423e-01 6.37853816e-02 -5.75042844e-01 8.41557920e-01
5.61659276e-01 -2.87420787e-02 -1.13296092e+00 1.71875679e+00
2.55649298e-01 1.26237288e-01 -2.48606727e-02 7.90512085e-01
4.24496969e-03 4.84861851e-01 1.74898893e-01 -7.54557699e-02
8.12794745e-01 -1.13151801e+00 -4.14980590e-01 -4.55923706e-01
2.58657336e-01 -8.00004065e-01 9.24639881e-01 4.16012108e-01
-8.98931861e-01 -3.95166487e-01 -8.34749401e-01 -2.10897356e-01
-1.46443680e-01 2.03971624e-01 7.69028366e-01 -1.27203450e-01
-9.23138559e-01 7.30297685e-01 -1.07755005e+00 -7.28481531e-01
6.23241425e-01 5.89307010e-01 -5.71548939e-01 -3.27415794e-01
-4.49640363e-01 1.34809744e+00 4.54729348e-01 -2.83824712e-01
-1.41399276e+00 -4.66275483e-01 -9.38536823e-01 -9.90593359e-02
7.66359866e-01 -7.23603547e-01 1.60796356e+00 -8.60477448e-01
-1.90518582e+00 4.90761429e-01 1.16621293e-01 -2.90871680e-01
4.73120660e-01 -4.81330693e-01 4.82704222e-01 2.88493246e-01
2.92157684e-03 8.03960979e-01 1.23470438e+00 -1.41340327e+00
-6.92154050e-01 -3.04790229e-01 3.97048086e-01 3.03130060e-01
5.11754788e-02 -1.11776493e-01 -1.05406605e-01 -5.54054260e-01
-6.09658798e-03 -1.19534612e+00 -6.77017510e-01 5.92224538e-01
2.03271471e-02 -2.14631602e-01 1.08286357e+00 -2.60410935e-01
2.23464489e-01 -1.88331521e+00 6.00626945e-01 2.16112360e-01
-4.32412326e-02 2.40924954e-01 -3.68818134e-01 5.37932515e-01
2.88235009e-01 -4.62430686e-01 -1.53846264e-01 -8.00956413e-02
1.66102946e-02 7.65384078e-01 -7.07694948e-01 6.99788809e-01
3.17851186e-01 9.26318169e-01 -1.23618388e+00 -3.26181263e-01
5.50869465e-01 2.59799749e-01 -6.58635378e-01 3.40585232e-01
-5.60391963e-01 9.76119697e-01 -6.55858696e-01 6.41538441e-01
1.99266262e-02 -2.87744850e-01 2.24287242e-01 1.99156880e-01
-1.12093017e-01 8.20377916e-02 -9.97240126e-01 2.11094117e+00
-6.47669792e-01 5.49894750e-01 3.93691152e-01 -1.45364809e+00
7.99211502e-01 6.10498860e-02 6.24851704e-01 -4.60413456e-01
9.00684521e-02 1.41675845e-01 -1.46609858e-01 -5.29356718e-01
4.28821355e-01 -1.38977453e-01 2.21911119e-03 1.50281757e-01
2.22106338e-01 -7.68142641e-01 1.27624393e-01 -3.06338146e-02
1.18534613e+00 7.34311759e-01 6.50325060e-01 -9.76146311e-02
1.78010166e-01 4.13135022e-01 4.15810347e-01 8.55934083e-01
-1.53875247e-01 2.96349078e-01 4.41276044e-01 -2.33315393e-01
-1.34103465e+00 -1.16815031e+00 1.55046076e-01 9.43835080e-01
2.90344745e-01 -1.60809696e-01 -1.41949192e-01 -6.73462689e-01
5.04505277e-01 3.84774446e-01 -4.41595703e-01 -1.43892229e-01
-9.63484406e-01 2.63702795e-02 3.02556325e-02 7.15474665e-01
1.69681907e-01 -1.32130098e+00 -1.04231167e+00 4.14976418e-01
2.78310716e-01 -1.32145715e+00 -1.01100415e-01 2.32147902e-01
-1.05242288e+00 -1.22566557e+00 -6.59984112e-01 -7.57053614e-01
7.28120267e-01 5.14635146e-01 8.89209270e-01 -9.91546065e-02
-5.48792839e-01 1.00356293e+00 -3.09063256e-01 -4.57251936e-01
-3.63134831e-01 -3.12346108e-02 1.10744841e-01 -5.39452612e-01
-1.30790427e-01 -6.75066829e-01 -3.60822558e-01 1.51769087e-01
-6.44776821e-01 -2.38739207e-01 8.26590598e-01 1.29945993e+00
6.33067727e-01 -4.82916892e-01 4.11614269e-01 -6.38042331e-01
4.53657031e-01 -2.80384004e-01 -1.07400477e+00 1.14733828e-02
-4.90089506e-01 2.79064149e-01 5.34098268e-01 -6.95562840e-01
-8.08864236e-01 3.73702854e-01 3.06838900e-01 -8.16790104e-01
-3.42791229e-01 2.01595187e-01 3.00558776e-01 -3.74520898e-01
5.11842608e-01 -6.02322333e-02 4.73652720e-01 -3.65405768e-01
6.64753199e-01 2.27235824e-01 5.92473030e-01 -1.05892432e+00
8.91179979e-01 7.18101382e-01 2.76096463e-01 -7.50531375e-01
-6.59512818e-01 -5.77718675e-01 -7.48090863e-01 3.75536196e-02
6.85368776e-01 -9.16396081e-01 -7.01761127e-01 5.03387339e-02
-1.12210345e+00 -7.04208195e-01 -6.25371993e-01 8.07136893e-01
-1.38570547e+00 3.85487854e-01 -4.91150081e-01 -5.89938283e-01
1.13866299e-01 -1.30264354e+00 1.31840110e+00 -6.09370647e-03
8.24308116e-03 -9.19463277e-01 1.65627301e-01 -1.81472912e-01
3.75789940e-01 3.71713936e-01 1.03188682e+00 -2.53112465e-01
-9.83052969e-01 -2.85019167e-02 -2.14080766e-01 1.24526113e-01
1.74294278e-01 -2.78365403e-01 -6.34776354e-01 -6.70631349e-01
-2.15296775e-01 -9.47933316e-01 8.53980422e-01 3.98122191e-01
1.28134871e+00 -5.34208529e-02 -5.70953727e-01 4.49738503e-01
1.42733145e+00 6.27293438e-02 2.53008574e-01 4.75935280e-01
5.67391753e-01 6.46150887e-01 1.04570079e+00 3.85187507e-01
-9.43574402e-03 8.12226295e-01 7.14246929e-01 6.01154715e-02
-6.00458719e-02 -4.40128207e-01 3.40723842e-01 2.07487926e-01
-1.89131677e-01 4.60376665e-02 -7.51119316e-01 4.98080343e-01
-2.27679944e+00 -8.89558613e-01 5.68449497e-01 2.06558442e+00
5.15730798e-01 -1.79048583e-01 6.32029921e-02 -2.79601991e-01
3.78968537e-01 4.67845313e-02 -9.24914062e-01 -5.78036234e-02
1.84246585e-01 4.03245449e-01 6.86902523e-01 4.03486639e-01
-1.22763085e+00 1.24629593e+00 6.69301510e+00 4.80144233e-01
-1.09472978e+00 3.63160893e-02 -4.49458547e-02 -5.64821735e-02
1.05276003e-01 3.56182843e-01 -4.97223288e-01 -1.35202631e-01
2.68476248e-01 6.92117810e-02 7.74487615e-01 1.03266442e+00
9.30460077e-03 -2.93261260e-01 -1.27786899e+00 1.16672266e+00
-8.66136551e-02 -1.16609335e+00 -1.59020513e-01 2.04791531e-01
7.34947324e-01 2.46023998e-01 -1.81094296e-02 4.92862642e-01
1.07256854e+00 -9.71522748e-01 4.97775018e-01 1.92304894e-01
6.73808575e-01 -4.62692350e-01 8.71299803e-02 4.86294657e-01
-9.10988569e-01 -6.03706360e-01 -6.55886173e-01 -1.86852172e-01
1.42402455e-01 6.13274332e-03 -8.53299141e-01 2.95460373e-01
4.95415241e-01 1.13410687e+00 2.16299295e-02 1.13093603e+00
-6.03299558e-01 1.68634489e-01 -4.45892632e-01 1.83521714e-02
5.26295662e-01 -2.89569259e-01 5.54856837e-01 7.77967751e-01
2.40145341e-01 5.01374714e-03 8.18566918e-01 5.82074881e-01
-1.73633378e-02 -6.24743886e-02 -1.17567265e+00 -4.21037525e-02
2.74298459e-01 1.17700577e+00 -6.36019051e-01 -2.79950827e-01
-5.36715686e-01 8.75069737e-01 8.78556132e-01 5.69480777e-01
-5.47760785e-01 6.67579472e-02 9.39437211e-01 -1.87702432e-01
6.29514396e-01 -8.70937645e-01 2.99912930e-01 -1.15117717e+00
-1.87754184e-01 -9.97587383e-01 -4.14965022e-03 -5.29120505e-01
-1.20487535e+00 -3.85844670e-02 3.31265122e-01 -1.27827847e+00
-6.38604760e-01 -1.15218496e+00 -1.41081646e-01 4.20247465e-01
-1.84527624e+00 -1.25846326e+00 -1.06545143e-01 7.16723979e-01
7.46794999e-01 -2.04622135e-01 8.58349264e-01 -9.76079479e-02
-3.58366743e-02 1.45404916e-02 3.90742898e-01 -7.08578154e-02
8.51090312e-01 -1.35297751e+00 7.09532127e-02 3.05126816e-01
2.36916825e-01 6.12289786e-01 5.66014826e-01 -3.41478676e-01
-1.84905040e+00 -8.38749290e-01 -8.65570754e-02 -4.05086398e-01
8.53681266e-01 -2.97155291e-01 -6.00049019e-01 1.08346212e+00
1.32836059e-01 3.19836169e-01 3.74749564e-02 2.61162315e-02
-4.11882907e-01 2.82405138e-01 -9.52482045e-01 4.91930366e-01
1.15747273e+00 -4.22444493e-01 -7.16834307e-01 4.94888127e-01
3.92249972e-01 -6.33004248e-01 -8.12648773e-01 5.05547047e-01
6.33411765e-01 -4.24982578e-01 1.12044299e+00 -9.46259856e-01
2.80843794e-01 -2.07529783e-01 -1.48065418e-01 -1.52577698e+00
-3.25046033e-01 -6.44691706e-01 -2.64297903e-01 6.63969874e-01
-4.30048220e-02 -5.65427244e-01 6.12871289e-01 1.70099854e-01
-4.65222485e-02 -5.64451814e-01 -8.73347044e-01 -9.82294977e-01
2.71713406e-01 3.41319218e-02 1.67123482e-01 8.63391638e-01
-9.40815881e-02 2.97134101e-01 -4.68539923e-01 2.20440514e-02
6.64317429e-01 4.92162943e-01 1.16567326e+00 -1.19760215e+00
-4.96528298e-01 -3.53510559e-01 -4.70799595e-01 -1.41570747e+00
6.71632946e-01 -7.87589371e-01 4.08700317e-01 -1.58878386e+00
-5.09196939e-03 -5.80550790e-01 -3.77432890e-02 6.97967768e-01
1.23641863e-01 -1.71571031e-01 4.92159933e-01 4.79981422e-01
-4.90944237e-01 8.32529187e-01 1.54870999e+00 -2.55212456e-01
2.51437929e-02 -1.47589445e-01 -7.07595870e-02 8.22941959e-01
8.10534358e-01 -3.56958091e-01 -5.87801039e-01 -5.04831791e-01
-2.45331138e-01 1.37933269e-01 5.96704483e-01 -9.04093862e-01
1.10114977e-01 -5.45605421e-01 4.06544894e-01 -1.29210994e-01
6.14923060e-01 -1.09818959e+00 -4.42369640e-01 6.76860332e-01
-3.89816850e-01 1.45554051e-01 2.45567799e-01 1.02099288e+00
-4.83557507e-02 -1.69290215e-01 8.24873924e-01 -4.25132811e-01
-1.13019204e+00 5.00330389e-01 -3.04690480e-01 -9.24434066e-02
1.25160563e+00 1.15544848e-01 -2.62619078e-01 -3.25767308e-01
-6.45096183e-01 2.73950398e-01 7.80060351e-01 5.52006900e-01
6.84755921e-01 -1.13992441e+00 -4.24156338e-01 1.03357732e-01
2.33120114e-01 1.15748882e-01 -4.09517527e-01 6.82119310e-01
-4.13596958e-01 2.69261420e-01 -5.67293644e-01 -8.10197413e-01
-9.90926504e-01 8.46070468e-01 8.22592229e-02 -2.84887373e-01
-9.94850099e-01 4.51115876e-01 2.52608508e-01 -5.03604650e-01
3.77347708e-01 -4.33455676e-01 3.51075944e-03 -3.95172060e-01
1.60821855e-01 2.10369661e-01 -5.32296181e-01 -3.80924165e-01
1.13978080e-01 8.05605590e-01 -1.28442496e-01 -2.45890126e-01
1.37139547e+00 1.43663585e-01 -6.64118379e-02 2.93835372e-01
9.94965196e-01 -3.67152631e-01 -1.88440430e+00 -3.87983501e-01
2.30232090e-01 -5.97432494e-01 -1.53753489e-01 -4.10267293e-01
-8.37722123e-01 9.45415914e-01 5.64600110e-01 -3.35786939e-01
7.38109827e-01 8.72877166e-02 4.76593107e-01 1.04969299e+00
9.44824994e-01 -9.90426958e-01 4.41964626e-01 7.18691885e-01
1.04628158e+00 -1.56537771e+00 2.48002991e-01 -2.53342301e-01
-3.76710266e-01 1.18181276e+00 6.51315808e-01 -7.69260645e-01
5.25692463e-01 2.04149783e-01 4.15905155e-02 -1.33974746e-01
-6.35420382e-01 -3.73332232e-01 -6.10384978e-02 9.17365670e-01
-5.83166517e-02 -1.82113096e-01 1.98201300e-03 -5.93734562e-01
1.60214067e-01 1.35523528e-02 4.05800909e-01 1.54309571e+00
-6.77704692e-01 -1.43581426e+00 -2.37077564e-01 3.36228341e-01
-2.35510379e-01 3.37417781e-01 -1.87786058e-01 1.12556148e+00
-3.38264614e-01 6.39711440e-01 -5.49139827e-02 2.89916694e-01
3.91930729e-01 -2.77440757e-01 1.19986570e+00 -8.30737472e-01
-2.02222794e-01 5.25357015e-02 -2.17669234e-01 -9.81622458e-01
-6.59533620e-01 -7.58860230e-01 -1.23814213e+00 1.79336295e-01
-1.15402699e-01 -1.34280488e-01 7.57645130e-01 8.63483191e-01
2.75203019e-01 3.38102013e-01 3.59137714e-01 -1.39783585e+00
-1.26385367e+00 -8.10456276e-01 -4.03842747e-01 4.69544798e-01
6.88158870e-01 -1.34156013e+00 2.81688236e-02 6.35953108e-03] | [4.6399736404418945, 0.7080104947090149] |
5c640414-964a-4186-9f46-bab263624f80 | 1st-place-solution-to-icdar-2021-rrc-ictext | 2104.03544 | null | https://arxiv.org/abs/2104.03544v1 | https://arxiv.org/pdf/2104.03544v1.pdf | 1st Place Solution to ICDAR 2021 RRC-ICTEXT End-to-end Text Spotting and Aesthetic Assessment on Integrated Circuit | This paper presents our proposed methods to ICDAR 2021 Robust Reading Challenge - Integrated Circuit Text Spotting and Aesthetic Assessment (ICDAR RRC-ICTEXT 2021). For the text spotting task, we detect the characters on integrated circuit and classify them based on yolov5 detection model. We balance the lowercase and non-lowercase by using SynthText, generated data and data sampler. We adopt semi-supervised algorithm and distillation to furtherly improve the model's accuracy. For the aesthetic assessment task, we add a classification branch of 3 classes to differentiate the aesthetic classes of each character. Finally, we make model deployment to accelerate inference speed and reduce memory consumption based on NVIDIA Tensorrt. Our methods achieve 59.1 mAP on task 3.1 with 31 FPS and 306M memory (rank 1), 78.7\% F2 score on task 3.2 with 30 FPS and 306M memory (rank 1). | ['Yi Niu', 'Li Zhu', 'Pengfei Li', 'Qiyao Wang'] | 2021-04-08 | null | null | null | null | ['text-spotting'] | ['computer-vision'] | [ 4.32647854e-01 -1.39405876e-01 9.39088836e-02 -3.80510837e-01
-8.19799781e-01 -6.21177256e-01 2.44904846e-01 7.99986571e-02
-1.87493891e-01 3.95103216e-01 -2.01273561e-01 -6.78418517e-01
1.38886347e-01 -6.23560667e-01 -6.75651073e-01 -2.65516043e-01
5.04282475e-01 4.19529945e-01 6.26593679e-02 3.33337300e-02
6.70478225e-01 4.79985267e-01 -1.54567635e+00 8.04861546e-01
9.04564857e-01 1.56112432e+00 2.36288592e-01 1.19803381e+00
-1.12695292e-01 8.83536696e-01 -9.55720723e-01 -4.03331250e-01
7.60447830e-02 -1.52534470e-01 -7.92698920e-01 -2.43408382e-01
6.02842629e-01 -4.39724177e-01 -2.79632092e-01 1.00555539e+00
8.49195778e-01 -5.30502558e-01 9.23553467e-01 -1.26241601e+00
-4.14882809e-01 3.87395203e-01 -9.86911535e-01 2.79303286e-02
4.59653437e-01 3.32888454e-01 8.79909575e-01 -8.87937903e-01
3.91891778e-01 1.27870965e+00 3.69993001e-01 4.06802833e-01
-9.41762865e-01 -7.98039675e-01 -1.38041496e-01 2.43647590e-01
-1.40714514e+00 -6.11521423e-01 3.80643874e-01 -4.11724091e-01
1.21373963e+00 4.81559277e-01 3.75502050e-01 8.69260967e-01
4.70135659e-01 9.35541868e-01 1.10538745e+00 -2.89893985e-01
4.47859317e-01 -1.27275348e-01 2.64699489e-01 6.82573617e-01
3.10892034e-02 -4.12081420e-01 -6.48149431e-01 1.83121055e-01
6.33632421e-01 -3.21105033e-01 2.48984009e-01 2.01084882e-01
-9.50088024e-01 3.29109102e-01 8.03913772e-02 -2.94729650e-01
1.20066099e-01 3.26491922e-01 4.37958032e-01 4.21949893e-01
1.42235905e-01 4.68168229e-01 -2.63649970e-01 -4.40998048e-01
-1.16273785e+00 -4.37854826e-02 4.61586624e-01 1.45007086e+00
3.26872677e-01 1.29005834e-01 -7.05449104e-01 1.14381182e+00
4.15969551e-01 1.10766315e+00 3.18500280e-01 -6.64310217e-01
6.09663129e-01 6.86342180e-01 -1.40065044e-01 -7.65953660e-01
-3.80014062e-01 -3.97560745e-01 -7.93450475e-01 5.64593017e-01
1.89311001e-02 -1.48480888e-02 -1.10212862e+00 8.76140654e-01
-3.91822219e-01 -3.59563321e-01 -1.12785265e-01 6.74518824e-01
8.41782391e-01 6.45569265e-01 -3.22885603e-01 2.64613450e-01
1.51462293e+00 -1.19188046e+00 -6.06419921e-01 1.99722908e-02
7.94151783e-01 -1.09086823e+00 1.45743155e+00 1.07283556e+00
-1.07517838e+00 -7.60906637e-01 -1.80560207e+00 -4.56555516e-01
-3.56289953e-01 1.04177666e+00 1.27582923e-01 9.83734846e-01
-9.68695164e-01 6.01723671e-01 -5.95829487e-01 -1.86804533e-02
7.50086188e-01 4.65588272e-01 2.86596775e-01 5.86609542e-02
-5.58272362e-01 4.27633762e-01 3.80238257e-02 -2.57944167e-01
-8.03254664e-01 -4.81503010e-01 -4.00559843e-01 -3.82760540e-02
3.32001224e-02 -1.46497265e-01 1.08475995e+00 -6.55507565e-01
-1.77614951e+00 1.05904472e+00 3.30862105e-02 -4.77783829e-01
6.54160559e-01 -5.33059120e-01 -8.38599622e-01 1.98837593e-01
-8.38104859e-02 9.56583917e-01 8.64266872e-01 -6.59863234e-01
-7.65428185e-01 -2.75336564e-01 -2.61912584e-01 1.50998518e-01
-4.15214151e-01 -3.82578745e-02 -7.61262834e-01 -6.63662910e-01
-1.75567865e-01 -6.62577033e-01 3.18684638e-01 2.81187654e-01
-8.98564100e-01 -6.21686876e-02 9.88785863e-01 -6.56304479e-01
1.43060017e+00 -2.27112007e+00 -4.54319298e-01 3.12972635e-01
5.10778248e-01 2.15461150e-01 -1.49300575e-01 -9.46559682e-02
2.06130892e-02 1.29052848e-01 1.09492831e-01 -4.59531426e-01
2.78854042e-01 -4.39386636e-01 -2.49127448e-01 2.55366087e-01
2.25350454e-01 8.68153691e-01 -3.28436881e-01 -3.54143411e-01
3.03495973e-01 1.92179546e-01 -5.27943075e-01 -7.87059516e-02
-2.33779520e-01 -2.58518934e-01 -2.08793595e-01 8.85753334e-01
9.59273517e-01 -2.69400328e-01 -1.08984716e-01 -2.56509483e-01
-8.58896598e-02 2.77774781e-01 -9.98334706e-01 1.65743399e+00
-5.18984258e-01 1.04625642e+00 -2.13848889e-01 -3.33264649e-01
1.26803470e+00 -3.70066881e-01 5.74730374e-02 -1.23739791e+00
1.98574752e-01 7.09834173e-02 -1.07484542e-01 -2.05808133e-01
8.67915094e-01 6.58800483e-01 -2.36244857e-01 4.23365772e-01
-2.47652650e-01 -1.84464276e-01 1.87353194e-02 3.19024146e-01
1.38244486e+00 1.48583338e-01 -2.62858063e-01 -5.43199420e-01
2.14874417e-01 -6.52186945e-02 8.63023698e-02 6.59016252e-01
-5.25389872e-02 8.38073194e-01 8.39442134e-01 -8.65614265e-02
-1.06692624e+00 -1.29864430e+00 -3.06993723e-01 1.13145566e+00
2.66144276e-01 -6.25916064e-01 -9.22836304e-01 -4.72771227e-01
-2.97381073e-01 8.36429060e-01 -2.89684743e-01 -2.04212904e-01
-4.28419486e-02 -7.54098237e-01 9.01396275e-01 6.94226325e-01
7.44444728e-01 -8.18363726e-01 -9.84070599e-01 -2.96209365e-01
1.93894178e-01 -1.18763769e+00 -4.11134124e-01 2.56510884e-01
-7.28922486e-01 -8.35641980e-01 -7.01066256e-01 -8.34748805e-01
7.60544598e-01 -1.97703853e-01 1.13424981e+00 -1.55247837e-01
-7.89712071e-01 -5.76494560e-02 -2.31255576e-01 -2.53524899e-01
-1.03827961e-01 1.40926495e-01 -1.16073646e-01 -3.47659826e-01
6.86249807e-02 4.44042049e-02 -7.38491118e-01 5.50755501e-01
-5.05326331e-01 6.10502422e-01 7.47321308e-01 4.55806851e-01
7.08822131e-01 -4.28186841e-02 2.75479704e-01 -7.35141814e-01
4.55921561e-01 1.21509530e-01 -6.25970244e-01 3.59211087e-01
-6.03544831e-01 1.42619997e-01 5.61273396e-01 -4.22044009e-01
-7.51832843e-01 2.35384136e-01 -1.29907086e-01 -2.07410961e-01
9.75475535e-02 -1.95118412e-01 -5.79372048e-01 -8.81904270e-03
7.04039693e-01 -8.47284868e-02 -3.80985111e-01 -3.77641529e-01
2.46609122e-01 1.19272566e+00 5.93401551e-01 -4.63272899e-01
5.23723662e-01 1.71811253e-01 -1.87800780e-01 -8.68160307e-01
-1.97544545e-01 2.50624157e-02 -4.49436843e-01 -2.33123854e-01
1.05007315e+00 -9.40396845e-01 -8.88352633e-01 7.28686452e-01
-1.00639248e+00 -5.01546025e-01 -2.43964861e-03 -6.07237518e-02
-3.24657083e-01 1.86597303e-01 -6.20336473e-01 -7.52157807e-01
-8.37791502e-01 -1.40160632e+00 1.62363040e+00 1.61756024e-01
-4.54131842e-01 -2.70341992e-01 -4.57117289e-01 3.53671610e-01
4.02723432e-01 -2.51699425e-02 1.17864347e+00 -3.64806831e-01
-3.54746640e-01 -2.85335004e-01 -8.17803741e-01 2.21463025e-01
-2.54894763e-01 2.01375917e-01 -1.23794854e+00 -1.58585399e-01
-4.01857287e-01 -4.43614960e-01 7.47688830e-01 1.74658194e-01
1.74521923e+00 2.09404290e-01 -3.20381254e-01 5.83230853e-01
1.20211661e+00 4.45169926e-01 1.08994734e+00 -2.88917534e-02
8.14491749e-01 2.13442862e-01 5.48554838e-01 6.35243177e-01
6.83460981e-02 6.82088256e-01 2.61357188e-01 -2.26169124e-01
-2.98863202e-01 -2.27902204e-01 5.18857121e-01 7.20049620e-01
5.43008983e-01 -4.90509003e-01 -1.02670753e+00 -2.07532141e-02
-1.38994896e+00 -2.58790553e-01 -4.62824881e-01 2.22524953e+00
4.49742913e-01 8.28683674e-01 6.44029900e-02 4.07451719e-01
7.61970997e-01 -1.50992155e-01 -9.16257501e-01 -6.33718550e-01
-4.29529279e-01 3.01977515e-01 6.56384826e-01 3.97622257e-01
-8.74280035e-01 8.84428978e-01 6.28934526e+00 1.51423991e+00
-1.13873076e+00 -3.61641675e-01 1.46303511e+00 -3.60352337e-01
-2.50264376e-01 -4.01396006e-01 -9.12504256e-01 5.89403093e-01
9.39793468e-01 2.31805220e-01 3.09128463e-01 6.40195012e-01
5.67121431e-03 -4.42656845e-01 -1.32968545e+00 1.64158988e+00
2.29738817e-01 -1.13971949e+00 4.26538102e-02 -4.33021262e-02
6.49639845e-01 -2.94971526e-01 5.07390797e-01 3.86451423e-01
2.76155416e-02 -1.50537074e+00 8.09114516e-01 4.92824346e-01
1.64156711e+00 -7.76584566e-01 4.59650159e-01 -1.61086917e-01
-1.17460752e+00 1.31984547e-01 -4.18798923e-01 7.74911279e-03
-4.03040588e-01 6.13999605e-01 -5.91719508e-01 7.16332346e-02
6.63660228e-01 6.75669312e-01 -1.13314748e+00 8.09458375e-01
4.20879200e-02 4.10027653e-01 -3.45764816e-01 -4.04569238e-01
-3.52056086e-01 1.96285635e-01 1.43909663e-01 1.10980463e+00
3.34236622e-01 -2.50645220e-01 -2.37637699e-01 7.54348040e-01
-3.78792346e-01 2.87123900e-02 -1.59332696e-02 -6.33503497e-02
4.76694852e-01 1.21606076e+00 -1.18116641e+00 -4.66136396e-01
1.28296196e-01 1.65809464e+00 -2.02839747e-01 1.05191782e-01
-1.27195144e+00 -9.95317578e-01 1.67303294e-01 5.18187061e-02
1.08072251e-01 1.12625724e-02 -1.19994414e+00 -7.73896515e-01
2.26196915e-01 -8.99593294e-01 1.79735257e-03 -1.21056581e+00
-7.86769569e-01 6.05700850e-01 -5.61272740e-01 -1.31097162e+00
3.39974761e-01 -1.18332732e+00 -6.33987546e-01 5.13047040e-01
-1.03634143e+00 -8.33550751e-01 -5.12699187e-01 3.92443873e-02
6.81652904e-01 -2.55118966e-01 6.49715722e-01 4.57929879e-01
-7.78670430e-01 1.27994800e+00 3.60575438e-01 2.74193678e-02
8.97615612e-01 -1.36920166e+00 9.09478664e-01 3.93970251e-01
-2.86734790e-01 7.84312561e-02 3.20672989e-01 -7.52842963e-01
-1.69703066e+00 -9.94197845e-01 4.94489521e-01 -4.12367463e-01
5.48522651e-01 -8.96876276e-01 -3.74054998e-01 -2.14548241e-02
1.92774236e-02 -4.16638583e-01 5.16921759e-01 -1.14872679e-01
-4.86279428e-01 -2.20083296e-01 -1.21302319e+00 7.83974767e-01
1.08246958e+00 -7.26785064e-01 1.24591649e-01 3.51395756e-01
4.47750896e-01 -4.63448107e-01 -7.82539368e-01 2.84563839e-01
7.35836565e-01 -7.06906974e-01 6.59147143e-01 1.68379799e-01
9.92777288e-01 -3.62403244e-01 -4.04484957e-01 -5.38871408e-01
1.60949647e-01 -6.34716988e-01 7.27986619e-02 1.11338949e+00
8.39124858e-01 -4.62764949e-02 9.66030598e-01 3.17060441e-01
-1.11037552e-01 -8.26629877e-01 -7.23309815e-01 -7.16144502e-01
8.66494477e-02 -7.52015233e-01 5.32083929e-01 2.88709939e-01
1.37264118e-01 6.19029403e-01 -2.78845221e-01 -1.58265859e-01
3.59444439e-01 9.61947143e-02 7.27127254e-01 -9.63854790e-01
-5.74670315e-01 -8.73062730e-01 -5.30960619e-01 -1.52932239e+00
-4.15027857e-01 -5.82641006e-01 -6.10974766e-02 -1.21064901e+00
1.84619606e-01 -2.57966340e-01 -7.21753836e-02 4.56944287e-01
8.86591971e-02 5.05897224e-01 2.22494170e-01 3.14524919e-02
-1.15382731e+00 4.89222437e-01 8.42524290e-01 -4.55126226e-01
-2.41615865e-02 -2.07100034e-01 -6.01597905e-01 3.68771255e-01
9.26371038e-01 7.12150987e-03 -3.71513456e-01 -6.14357829e-01
3.34370583e-01 -5.04165664e-02 1.81102246e-01 -1.53161430e+00
1.87197894e-01 4.32444245e-01 8.62654686e-01 -1.03631115e+00
4.10038918e-01 -3.81991565e-01 -2.95308173e-01 5.14339447e-01
-4.99425083e-01 2.65586555e-01 5.48191071e-01 2.49423072e-01
4.18026179e-01 -1.57771751e-01 8.16422224e-01 6.34132624e-01
-6.12382829e-01 -1.99149027e-01 -6.57125831e-01 5.10161072e-02
9.10876393e-01 -4.34357285e-01 -6.29334688e-01 -2.90416300e-01
-5.71559608e-01 3.96277517e-01 6.86652184e-01 3.79074901e-01
8.83171558e-01 -1.17278075e+00 -5.81481099e-01 3.28970492e-01
3.41180921e-01 -2.76303083e-01 3.59991670e-01 4.45707470e-01
-1.09117162e+00 1.44677997e-01 -4.20315564e-01 -6.04273379e-01
-1.41163385e+00 1.83704838e-01 2.60516107e-01 -1.14031799e-01
-4.18816596e-01 8.60960841e-01 1.26999602e-01 2.31044199e-02
5.00171185e-01 -3.47778469e-01 -1.42238513e-01 -1.05598599e-01
6.71243608e-01 7.82734632e-01 3.47389847e-01 -1.73896804e-01
-4.53243554e-01 7.26450861e-01 -3.17075551e-01 -3.10518146e-01
7.21675038e-01 1.36084601e-01 1.65212706e-01 3.57082069e-01
1.37138271e+00 -6.36692420e-02 -1.09717715e+00 3.09601188e-01
9.97821242e-02 8.01808015e-02 2.40778476e-01 -1.34136283e+00
-8.36051345e-01 1.28468740e+00 1.36239564e+00 -3.03732038e-01
1.38326383e+00 -5.05183518e-01 8.53933513e-01 3.91478390e-01
1.61333773e-02 -1.65673268e+00 3.42693239e-01 5.00393212e-01
6.12455249e-01 -1.02532768e+00 1.89836502e-01 -4.50657517e-01
-7.08338499e-01 1.21252584e+00 8.40279698e-01 -7.20657930e-02
2.18866423e-01 8.20078254e-01 -2.43976247e-02 -1.83203176e-01
-6.75383687e-01 3.03755641e-01 3.72837543e-01 4.84386832e-01
5.13897896e-01 2.98032254e-01 1.38111293e-01 6.66219354e-01
-4.08697754e-01 -3.58168453e-01 3.51495862e-01 8.61229181e-01
-4.09731448e-01 -7.11256862e-01 -2.41545543e-01 8.25290740e-01
-3.07723820e-01 -3.40336084e-01 -6.55181885e-01 1.76940605e-01
-8.68580267e-02 1.17894220e+00 3.88340771e-01 -9.73183692e-01
1.94876537e-01 -2.73891300e-01 3.33673596e-01 -2.78757453e-01
-4.29725558e-01 2.69802004e-01 9.55787599e-02 -4.14785266e-01
6.02258205e-01 -1.72629341e-01 -1.40307963e+00 -4.15340275e-01
-1.67553067e-01 -3.07173312e-01 9.54347432e-01 6.17922246e-01
7.36196220e-01 1.05192554e+00 5.94745219e-01 -4.63685960e-01
-2.95638740e-01 -9.15827870e-01 -3.52992237e-01 -1.07733160e-01
8.47412739e-03 -2.83851743e-01 6.25545904e-02 4.88197319e-02] | [11.864747047424316, 2.2575185298919678] |
00e6be64-9ee8-4e11-8ed8-4acbb00e4cc7 | analyzing-how-bert-performs-entity-matching | null | null | https://dl.acm.org/doi/10.14778/3529337.3529356 | https://dl.acm.org/doi/pdf/10.14778/3529337.3529356 | Analyzing how BERT performs entity matching | State-of-the-art Entity Matching (EM) approaches rely on transformer architectures, such as BERT, for generating highly contex-tualized embeddings of terms. The embeddings are then used to predict whether pairs of entity descriptions refer to the same real-world entity. BERT-based EM models demonstrated to be effective, but act as black-boxes for the users, who have limited insight into the motivations behind their decisions.
In this paper, we perform a multi-facet analysis of the components of pre-trained and fine-tuned BERT architectures applied to an EM task. The main findings resulting from our extensive experimental evaluation are (1) the fine-tuning process applied to the EM task mainly modifies the last layers of the BERT components, but in a different way on tokens belonging to descriptions of matching / non-matching entities; (2) the special structure of the EM datasets, where records are pairs of entity descriptions is recognized by BERT; (3) the pair-wise semantic similarity of tokens is not a key knowledge exploited by BERT-based EM models. | ['Francesco Guerra', 'Andrea Baraldi', 'Francesco Del Buono', 'Matteo Paganelli'] | 2022-04-01 | null | null | null | proceedings-of-the-vldb-endowment-2022-4 | ['entity-resolution'] | ['natural-language-processing'] | [-4.17230874e-01 2.97103077e-01 -1.57147199e-02 -4.65191722e-01
-5.07996023e-01 -5.09161830e-01 1.06493390e+00 5.42878449e-01
-8.43549013e-01 2.90303916e-01 3.36279839e-01 -2.29156718e-01
-2.15077430e-01 -1.09403920e+00 -7.32001960e-01 -2.37888873e-01
-9.83996466e-02 9.50086594e-01 2.98870146e-01 -4.70081687e-01
1.42746627e-01 3.88150930e-01 -1.69070733e+00 5.17040431e-01
6.87157810e-01 7.57846832e-01 2.08460793e-01 3.51208657e-01
-5.70412338e-01 5.80335021e-01 -4.18613166e-01 -9.92128849e-01
1.03348009e-01 3.47397663e-02 -9.95295763e-01 -4.38486159e-01
3.89887393e-01 -1.30289853e-01 -5.59868395e-01 9.59464192e-01
3.22683454e-01 3.33597302e-01 9.05765116e-01 -1.31739533e+00
-1.06480205e+00 8.71093214e-01 1.53061554e-01 1.21792838e-01
3.76555115e-01 -1.72527313e-01 1.46849823e+00 -1.17261600e+00
1.00799239e+00 1.21225882e+00 6.79327905e-01 5.76226711e-01
-1.21874857e+00 -4.93888438e-01 1.02413163e-01 6.73290133e-01
-1.77949297e+00 -4.70793396e-01 4.31480557e-01 -4.05774862e-01
1.34058833e+00 2.41535559e-01 3.21405917e-01 9.73985255e-01
-1.36711419e-01 7.27107644e-01 5.72710812e-01 -5.50193429e-01
1.08244710e-01 9.18883562e-01 1.06780119e-01 2.72826791e-01
2.84139484e-01 7.67106144e-03 -2.78423429e-01 -2.77778864e-01
2.16012508e-01 -1.03521012e-01 -4.40331884e-02 -5.19382834e-01
-9.35487688e-01 1.03874898e+00 4.89332408e-01 8.63645494e-01
-5.83867550e-01 -1.01030827e-01 3.19056928e-01 4.22381371e-01
1.77389935e-01 8.31024528e-01 -4.61997837e-01 -2.58941144e-01
-1.00991023e+00 3.91014993e-01 9.94846046e-01 1.12522531e+00
1.03723431e+00 -5.10214984e-01 -2.16076151e-01 7.52448320e-01
5.14347911e-01 1.19798869e-01 6.49339736e-01 -3.32624018e-01
4.68728036e-01 8.97724271e-01 2.12995350e-01 -1.05661452e+00
-1.65171757e-01 -2.51732200e-01 -3.28094631e-01 -2.47096717e-01
8.87400061e-02 3.60479169e-02 -5.66853225e-01 1.94975746e+00
2.19178617e-01 2.06374135e-02 -1.19339228e-02 5.00216544e-01
1.17531002e+00 4.10978377e-01 4.29628760e-01 5.25015116e-01
1.53940976e+00 -7.59404778e-01 -5.19058108e-01 -3.31165463e-01
1.00595379e+00 -7.99278677e-01 5.84951699e-01 -5.45707941e-01
-1.21945667e+00 -6.69241726e-01 -8.61217380e-01 -2.96319187e-01
-1.29829347e+00 5.59903346e-02 3.54902387e-01 5.32022953e-01
-1.12369263e+00 7.14654446e-01 -4.57380086e-01 -5.45988619e-01
1.52713954e-01 3.94979864e-01 -5.96826315e-01 -1.19621046e-01
-1.72372520e+00 1.58407569e+00 5.67169487e-01 -4.60502878e-02
-5.62114060e-01 -7.74564087e-01 -1.23795712e+00 6.59604371e-01
1.88784972e-02 -6.82246864e-01 1.12163293e+00 -7.30761766e-01
-7.62780905e-01 1.45313060e+00 -4.73916501e-01 -4.18567806e-01
3.83076906e-01 7.17977807e-02 -8.27688575e-01 -2.39470769e-02
2.67614067e-01 7.99614668e-01 4.02442932e-01 -1.24634099e+00
-8.11834395e-01 -1.78688258e-01 2.36833036e-01 7.55918473e-02
-4.50299948e-01 1.68126881e-01 -2.00773880e-01 -4.25190389e-01
-3.00516397e-01 -7.04415739e-01 4.96933721e-02 -2.66114920e-01
-1.43725887e-01 -5.20680726e-01 4.41246659e-01 -4.13735569e-01
1.59567916e+00 -2.08726358e+00 -5.11111245e-02 2.12460414e-01
1.45520151e-01 5.17308712e-01 -2.64634103e-01 1.09388459e+00
-1.70376495e-01 2.92118192e-01 2.21134320e-01 -6.21854007e-01
6.45137906e-01 -1.53178927e-02 -2.64632255e-01 1.08571857e-01
2.86082357e-01 1.16102111e+00 -9.18806851e-01 -4.77835059e-01
2.16682613e-01 3.60619217e-01 -3.55824351e-01 4.60914820e-01
2.38484591e-02 -1.49833322e-01 -3.31509233e-01 3.43182683e-01
7.37564206e-01 -2.89288253e-01 1.65585116e-01 -3.29998374e-01
-5.29373400e-02 7.87557244e-01 -1.24018276e+00 1.44904864e+00
-6.31878734e-01 7.59310424e-01 5.76882018e-03 -9.01758134e-01
7.04452276e-01 4.68285292e-01 2.23951444e-01 -7.80570149e-01
8.13742206e-02 2.68966228e-01 -1.30855694e-01 -5.79684079e-01
9.87764120e-01 -4.03176360e-02 -1.17758326e-01 5.11455476e-01
4.03332680e-01 3.33052218e-01 5.03599823e-01 2.42788211e-01
1.12809432e+00 -7.71564096e-02 3.07862133e-01 -2.71728933e-01
6.20174289e-01 1.57191277e-01 3.55066895e-01 9.58677351e-01
5.54051325e-02 3.74913603e-01 2.44502891e-02 -2.54270434e-01
-1.02434897e+00 -1.09448469e+00 -4.18188602e-01 1.07574487e+00
3.57082307e-01 -8.20266962e-01 -4.20079678e-01 -6.25760674e-01
2.89357215e-01 1.13826430e+00 -8.36486876e-01 -5.39021432e-01
-5.14804482e-01 -4.86842722e-01 4.82015669e-01 6.01522326e-01
2.26816252e-01 -1.09134007e+00 -5.63900650e-01 4.81284678e-01
-2.77077317e-01 -1.21227193e+00 -1.78906471e-01 1.99495524e-01
-4.67862010e-01 -8.96164596e-01 -4.58594054e-01 -8.97270918e-01
6.97214663e-01 9.23626944e-02 1.64768219e+00 2.68504649e-01
-9.57742184e-02 5.03407896e-01 -3.98610324e-01 -4.87892717e-01
-5.74792206e-01 3.48492414e-01 -1.05182879e-01 -3.48242037e-02
1.36393809e+00 -6.65381551e-01 -5.75173497e-01 6.05423689e-01
-9.24691617e-01 -2.76737541e-01 5.37107170e-01 7.59964764e-01
2.40346994e-02 -9.18667614e-02 4.27206159e-01 -9.09699023e-01
5.65934479e-01 -8.89965892e-01 -2.88608044e-01 5.72116375e-01
-7.48937130e-01 3.95300537e-01 2.61947900e-01 -5.49964666e-01
-9.04354155e-01 -4.95838314e-01 -4.12140846e-01 -1.73528135e-01
-2.51790315e-01 4.34296191e-01 -1.14541516e-01 1.12536877e-01
5.81290662e-01 9.81102288e-02 -4.75834727e-01 -6.93101406e-01
5.60933650e-01 8.23951304e-01 2.57458448e-01 -6.76958323e-01
1.00250101e+00 3.74055922e-01 -4.70214069e-01 -4.25491065e-01
-5.29725492e-01 -8.72201264e-01 -7.37382174e-01 1.96813196e-01
7.63497949e-01 -9.81809676e-01 -5.29842317e-01 -5.51396757e-02
-1.18272793e+00 4.42957394e-02 -3.04939330e-01 4.27586317e-01
-2.13293374e-01 9.67456773e-02 -6.17329061e-01 -5.55717528e-01
-2.62446284e-01 -8.63711059e-01 1.00269639e+00 3.18349391e-01
-6.04415953e-01 -1.25791276e+00 2.11236596e-01 1.82922438e-01
7.73930490e-01 -3.51521492e-01 1.32354307e+00 -1.26597929e+00
-4.18804109e-01 -4.85399395e-01 -3.17508548e-01 -4.24928740e-02
-8.19007233e-02 -6.26737671e-03 -1.24263394e+00 -1.49423987e-01
-3.39117467e-01 6.23585433e-02 6.62488699e-01 -3.69458944e-02
5.37247419e-01 -7.54402727e-02 -6.88611925e-01 3.86920363e-01
1.29885972e+00 5.53334989e-02 7.61387408e-01 8.17610860e-01
3.62311602e-01 6.97537959e-01 5.55968225e-01 3.26074094e-01
7.37476468e-01 9.33629096e-01 3.29638809e-01 -7.52215013e-02
8.02305043e-02 -6.30164802e-01 1.20647237e-01 5.86418211e-01
2.83607334e-01 -1.57731354e-01 -6.80459499e-01 1.07150435e+00
-1.95971704e+00 -1.11691511e+00 5.86719513e-02 2.20198536e+00
7.69475937e-01 2.41510142e-02 -9.41118747e-02 -3.95548582e-01
8.68638933e-01 -2.39749905e-02 -1.04932219e-01 -4.32992578e-01
-1.18235126e-01 3.78472090e-01 3.48724991e-01 1.60140038e-01
-1.08425510e+00 9.75570142e-01 6.15760183e+00 8.22887421e-01
-8.17021430e-01 1.41099095e-01 -7.34612569e-02 1.53461173e-01
-4.05819327e-01 3.40025932e-01 -9.95955288e-01 5.66677213e-01
1.11423409e+00 -5.42134404e-01 1.96257100e-01 8.94855499e-01
-2.65454918e-01 1.33941188e-01 -1.66453838e+00 8.37718070e-01
-6.42894357e-02 -1.36094677e+00 2.20601737e-01 2.54527867e-01
4.63012010e-01 1.47916853e-01 -8.20035562e-02 8.76059592e-01
5.63126981e-01 -1.10989463e+00 8.12919915e-01 2.49905944e-01
2.99264193e-01 -4.58290607e-01 1.03079271e+00 3.49780262e-01
-1.29534745e+00 1.08702201e-03 -6.92714393e-01 1.99172199e-01
3.38560104e-01 4.20361698e-01 -9.74706769e-01 6.43567145e-01
5.63514113e-01 4.01554167e-01 -7.20810950e-01 1.07730782e+00
-3.85293812e-01 1.91418648e-01 -2.38255724e-01 -8.92998874e-02
4.85398680e-01 -1.09206483e-01 5.04612803e-01 1.64684081e+00
3.30501825e-01 -4.04554695e-01 -3.59765053e-01 1.54663956e+00
-3.88101578e-01 5.31251095e-02 -6.04580343e-01 -3.70644420e-01
8.51486564e-01 1.43189836e+00 -1.40628830e-01 -4.69313741e-01
-6.43725574e-01 7.76742935e-01 5.83191216e-01 2.90566087e-01
-6.78403556e-01 -5.14105856e-01 8.43526483e-01 2.86955714e-01
6.98870480e-01 9.57205594e-02 5.57339527e-02 -1.15762579e+00
1.67387083e-01 -5.85490108e-01 6.38886929e-01 -7.65460432e-01
-1.40715730e+00 6.31477237e-01 9.65792239e-02 -9.81896281e-01
-6.56370163e-01 -5.47386169e-01 -7.44366765e-01 9.77312207e-01
-1.76143110e+00 -1.19769967e+00 -1.85582727e-01 3.87499869e-01
8.76390934e-02 -1.29306138e-01 1.30073833e+00 7.90803671e-01
-2.56526321e-01 7.53426015e-01 2.75467038e-01 5.58195233e-01
7.30196476e-01 -1.27705586e+00 8.08941245e-01 7.05509126e-01
5.18022835e-01 1.19796312e+00 7.46503055e-01 -3.64953190e-01
-1.02026463e+00 -8.57591748e-01 1.84630668e+00 -7.23810434e-01
6.77363217e-01 -4.68632519e-01 -1.04087520e+00 7.88598955e-01
3.01721036e-01 -1.12654537e-01 7.63864815e-01 5.00131667e-01
-5.03943563e-01 2.78053939e-01 -1.21103394e+00 5.75797796e-01
9.45117116e-01 -9.15879071e-01 -1.20291352e+00 3.89156118e-02
3.98196578e-01 -8.71648639e-02 -1.00981081e+00 1.33933604e-01
5.27695119e-01 -8.18677247e-01 1.05085778e+00 -1.12596714e+00
4.06092077e-01 -3.54016945e-02 -2.03272209e-01 -1.32218313e+00
-5.08108199e-01 -1.80386052e-01 -1.43064290e-01 1.62093496e+00
6.25030100e-01 -5.50871909e-01 5.20218790e-01 1.15585852e+00
1.90394819e-01 -4.20101225e-01 -8.14416647e-01 -7.72590280e-01
9.02363658e-02 -1.33897901e-01 9.19277191e-01 1.19346797e+00
1.48888588e-01 2.40662888e-01 2.86064059e-01 1.23039752e-01
1.83499083e-01 1.76148728e-01 6.85001373e-01 -1.34483802e+00
-1.64875999e-01 -5.14475882e-01 -7.03419268e-01 -1.07199752e+00
3.92853916e-01 -1.06523132e+00 -3.06172937e-01 -1.56521714e+00
3.00036669e-01 -6.33141994e-01 -5.24895191e-01 1.67292491e-01
-1.96134970e-01 -2.49637738e-01 1.09828323e-01 1.76048636e-01
-5.09484351e-01 5.60679257e-01 4.34698343e-01 -2.01922521e-01
8.94510150e-02 -1.03596607e-02 -6.80421889e-01 5.09797573e-01
3.08675021e-01 -8.06217313e-01 -1.57204658e-01 -6.45876706e-01
5.77539980e-01 -5.21124184e-01 3.00687611e-01 -6.45399988e-01
4.04632628e-01 1.44067481e-01 6.02630414e-02 -4.21851844e-01
4.27997351e-01 -1.14884973e+00 2.91589350e-01 4.61858232e-03
-4.61139321e-01 3.76080364e-01 1.66367069e-01 2.92544782e-01
-5.88956237e-01 -8.23001027e-01 5.00888884e-01 -3.41314197e-01
-9.79651392e-01 9.15519074e-02 -2.77857989e-01 3.89306396e-01
8.18867087e-01 -2.25846857e-01 -1.70028239e-01 -2.95929223e-01
-6.35811687e-01 3.15873325e-01 5.26179552e-01 6.23735368e-01
1.11675248e-01 -1.49348736e+00 -6.56292260e-01 1.21628819e-02
5.39160311e-01 -2.34163210e-01 1.24742895e-01 6.57856286e-01
-9.29539874e-02 5.06081641e-01 -1.97086468e-01 -3.23002785e-01
-1.23545969e+00 5.60133338e-01 4.96775806e-01 -5.68158329e-01
-4.26127195e-01 8.36273193e-01 1.77938670e-01 -7.43836701e-01
9.13077444e-02 -7.37660229e-02 -2.50174880e-01 3.83216739e-01
4.78219777e-01 2.42812440e-01 4.49499398e-01 -7.94208229e-01
-5.41929185e-01 2.07301527e-01 -4.10263330e-01 -1.81161791e-01
1.51756287e+00 -1.57171354e-01 -2.15232242e-02 2.36126512e-01
1.39757597e+00 -2.15352729e-01 -4.04387534e-01 -6.36221826e-01
3.67573947e-01 -4.93435413e-01 -8.56492594e-02 -5.50591290e-01
-6.28914714e-01 8.62360477e-01 3.68893087e-01 2.83811599e-01
5.13921261e-01 4.93666232e-02 8.51351500e-01 4.98545498e-01
4.08116072e-01 -1.37199152e+00 -6.00956678e-01 3.99020553e-01
3.96747947e-01 -1.17389131e+00 -3.12535524e-01 -1.27678812e-01
-4.88216728e-01 1.02135432e+00 4.98864710e-01 1.51699662e-01
5.46958208e-01 4.09234017e-02 -1.06408753e-01 -4.54353184e-01
-8.80992591e-01 -4.69208330e-01 4.68066990e-01 5.57011545e-01
4.82833803e-01 -3.37364897e-02 -2.39094809e-01 6.19515717e-01
-1.07178830e-01 -1.89754546e-01 1.23619929e-01 9.99281585e-01
-2.95640320e-01 -1.20871758e+00 4.06000279e-02 2.54619956e-01
-3.81398171e-01 -4.09810185e-01 -5.66928983e-01 1.03053081e+00
7.26481303e-05 9.75581229e-01 2.20152438e-01 -3.74258041e-01
4.63369310e-01 4.66982394e-01 2.37048119e-01 -6.79540873e-01
-9.56034362e-01 -9.05222595e-01 3.32976699e-01 -3.91312957e-01
-1.69427514e-01 -5.43649554e-01 -1.08896613e+00 -4.55751389e-01
-4.69712436e-01 4.84498411e-01 7.17034459e-01 1.06013727e+00
8.31312239e-01 1.58409104e-01 4.70597982e-01 -4.91380423e-01
-7.16874242e-01 -1.12916553e+00 -5.16983926e-01 1.16975188e+00
-6.88068420e-02 -7.93586016e-01 -4.84641314e-01 -2.60990113e-01] | [9.471957206726074, 8.47412395477295] |
29b32180-d0ca-490d-804f-89b1a829cf50 | conversational-implicatures-in-english | 1911.10704 | null | https://arxiv.org/abs/1911.10704v1 | https://arxiv.org/pdf/1911.10704v1.pdf | Conversational implicatures in English dialogue: Annotated dataset | Human dialogue often contains utterances having meanings entirely different from the sentences used and are clearly understood by the interlocutors. But in human-computer interactions, the machine fails to understand the implicated meaning unless it is trained with a dataset containing the implicated meaning of an utterance along with the utterance and the context in which it is uttered. In linguistic terms, conversational implicatures are the meanings of the speaker's utterance that are not part of what is explicitly said. In this paper, we introduce a dataset of dialogue snippets with three constituents, which are the context, the utterance, and the implicated meanings. These implicated meanings are the conversational implicatures. The utterances are collected by transcribing from listening comprehension sections of English tests like TOEFL (Test of English as a Foreign Language) as well as scraping dialogues from movie scripts available on IMSDb (Internet Movie Script Database). The utterances are manually annotated with implicatures. | ['Radhika Mamidi', 'Elizabeth Jasmi George'] | 2019-11-25 | null | null | null | null | ['implicatures'] | ['natural-language-processing'] | [ 3.61970484e-01 7.46786356e-01 1.88192561e-01 -8.84908140e-01
-3.91254187e-01 -1.07605195e+00 7.68756807e-01 1.21056683e-01
-1.95054591e-01 7.22469687e-01 9.54735339e-01 -2.45018989e-01
1.73181146e-01 -3.60327244e-01 -4.88174886e-01 -5.39718926e-01
6.73631504e-02 5.81876874e-01 2.98663806e-02 -6.63432002e-01
2.31770486e-01 -2.26920575e-01 -1.18596518e+00 9.99658823e-01
3.84210736e-01 7.84488797e-01 3.81037682e-01 9.59332287e-01
-6.93939030e-01 1.56722105e+00 -1.17772663e+00 -6.48896813e-01
-3.51725161e-01 -7.66447067e-01 -1.56700861e+00 6.78386867e-01
1.46756887e-01 -5.22846103e-01 2.51325779e-02 8.81870925e-01
-8.46818760e-02 -7.82777667e-02 3.98292512e-01 -1.08986318e+00
-3.81911278e-01 1.10155773e+00 1.88385382e-01 2.08976582e-01
9.57015514e-01 -6.86089247e-02 1.22991741e+00 -6.51127279e-01
6.92027390e-01 1.73937345e+00 3.75989854e-01 7.42642105e-01
-1.09760690e+00 -1.43420650e-02 1.91376492e-01 -2.35522073e-02
-7.42880106e-01 -9.36296523e-01 6.68454170e-01 -4.55520719e-01
1.35908699e+00 5.41936815e-01 4.16380852e-01 1.35620332e+00
9.55742896e-02 8.19689631e-01 7.01200604e-01 -6.82563841e-01
2.05835178e-01 5.61480939e-01 6.28848135e-01 1.11766420e-01
-7.78965414e-01 -5.41702926e-01 -5.29011726e-01 -4.68538433e-01
4.17755455e-01 -6.64651275e-01 -4.72516745e-01 1.67396352e-01
-1.27792013e+00 8.55782092e-01 -2.61551559e-01 5.68883181e-01
-5.06620526e-01 -1.54378638e-01 6.78074777e-01 8.50700080e-01
3.99163872e-01 4.10166800e-01 -7.59929359e-01 -6.93405926e-01
2.68688232e-01 3.27660620e-01 1.56426954e+00 9.21405613e-01
2.75260687e-01 -4.03929323e-01 2.30763823e-01 1.42488635e+00
2.39344373e-01 9.83995646e-02 6.78889811e-01 -1.34674549e+00
4.65071142e-01 6.33996964e-01 4.98656034e-01 -8.81456017e-01
-3.15814316e-01 4.14654881e-01 -4.91243489e-02 -5.40752351e-01
8.02602470e-01 -3.29583734e-01 7.35647529e-02 1.99639511e+00
5.93584776e-01 -3.05770934e-01 8.46091509e-01 6.44673705e-01
1.14745736e+00 7.90430903e-01 2.13370770e-02 -5.54709077e-01
1.63894904e+00 -1.04004312e+00 -1.35796273e+00 -4.19628471e-01
9.66543436e-01 -9.42826748e-01 1.31627512e+00 4.61285740e-01
-1.06041384e+00 -3.95090252e-01 -9.29317355e-01 -1.82408899e-01
-2.45616660e-02 -3.19901466e-01 3.59427392e-01 4.30378228e-01
-8.41902494e-01 7.74304494e-02 -1.74373239e-01 -4.82895374e-01
-6.00425005e-01 1.72843188e-01 -5.05133092e-01 1.56575739e-01
-1.34485400e+00 9.01245058e-01 3.03832799e-01 7.79948905e-02
-7.82877803e-01 -1.94192920e-02 -1.04561353e+00 -1.06507666e-01
5.15111864e-01 -4.34323549e-02 2.13379717e+00 -1.63622022e+00
-1.70516396e+00 1.09608281e+00 -2.75445223e-01 -2.87713230e-01
1.35082021e-01 -2.07727432e-01 -4.96136606e-01 2.68544257e-01
2.01870948e-01 3.34449530e-01 7.05715120e-01 -1.03545952e+00
-7.32190311e-01 -4.26973730e-01 7.69673169e-01 4.38784808e-01
-3.55518535e-02 3.28614861e-01 -2.83052862e-01 -4.88465667e-01
3.58903825e-01 -8.07109058e-01 3.23339939e-01 -5.75980604e-01
-7.06701696e-01 -6.25145972e-01 8.54933918e-01 -7.43439972e-01
1.08463323e+00 -2.40137577e+00 2.78230697e-01 -4.29460734e-01
2.26363048e-01 -3.33835572e-01 -2.02561393e-01 9.15361881e-01
-3.17957014e-01 4.38502505e-02 -1.46599978e-01 -3.53556216e-01
1.73358470e-01 8.14329147e-01 -4.47521925e-01 3.27894658e-01
1.49412364e-01 5.30968964e-01 -9.74997461e-01 -3.06918472e-01
5.97075634e-02 -2.71422695e-02 -3.37271154e-01 8.48198175e-01
-6.11053646e-01 6.47499621e-01 -5.15498042e-01 1.03653759e-01
2.37442568e-01 8.90724808e-02 5.96642613e-01 -1.31674893e-02
-1.21613266e-03 1.11383402e+00 -7.58235037e-01 1.34455967e+00
-4.91063684e-01 6.39979899e-01 5.71699798e-01 -7.11204112e-01
5.55321515e-01 9.51312065e-01 -1.74219072e-01 -8.91823322e-02
8.38797092e-02 5.50188050e-02 3.43236774e-01 -1.20820093e+00
5.60957432e-01 -6.15322113e-01 -4.56530631e-01 7.08765566e-01
3.45470235e-02 -4.28453833e-01 2.02949643e-01 4.54611152e-01
1.04474282e+00 -4.59614396e-01 6.66049600e-01 -2.78749347e-01
8.81684363e-01 8.52327868e-02 3.90632480e-01 5.14842689e-01
-8.20436850e-02 1.78389087e-01 1.14354467e+00 -4.24035609e-01
-7.31787384e-01 -9.30031776e-01 -2.81435668e-01 1.38608778e+00
-1.13182351e-01 -6.14071846e-01 -1.07661045e+00 -7.16454446e-01
-5.24489939e-01 1.30773771e+00 -4.44478154e-01 3.05915624e-01
-6.65554345e-01 -4.40650135e-02 5.96593261e-01 8.47880542e-02
3.18053305e-01 -1.32300436e+00 -3.43359649e-01 5.55404544e-01
-7.74337947e-01 -1.49085736e+00 -4.89973158e-01 5.48604205e-02
-4.39671636e-01 -1.21533906e+00 1.87067643e-01 -7.88525403e-01
4.31852907e-01 8.86391092e-04 1.35119665e+00 3.44481677e-01
4.73517388e-01 5.42609930e-01 -7.02978790e-01 -1.90033630e-01
-1.32762671e+00 -6.17282569e-01 3.55854537e-03 5.56011535e-02
4.42575186e-01 -5.17987549e-01 1.99667796e-01 5.13226509e-01
-8.74267399e-01 2.39898860e-01 -1.51153728e-01 9.51197386e-01
-2.98972372e-02 -8.10177531e-03 6.34189904e-01 -1.30810785e+00
9.77600276e-01 -7.00112045e-01 7.27762356e-02 2.37166062e-01
4.61568415e-01 -1.69302344e-01 7.94540703e-01 -3.30777287e-01
-1.42674828e+00 -3.13534766e-01 -4.21309412e-01 3.77495527e-01
-7.24266291e-01 3.52024734e-01 -5.37325978e-01 7.72417188e-01
5.02442420e-01 2.28033382e-02 -1.21918768e-02 -6.17304087e-01
3.14865440e-01 1.26895750e+00 6.14049852e-01 -1.04584789e+00
-1.77662089e-01 -1.17029704e-01 -9.01100934e-01 -1.36420417e+00
-1.05212641e+00 -4.90098566e-01 -5.83294034e-01 -3.76242965e-01
9.40928042e-01 -5.81254542e-01 -5.64264476e-01 3.48845661e-01
-1.72562194e+00 -2.68442780e-01 -1.36979267e-01 3.65456581e-01
-7.99579322e-01 5.01562119e-01 -9.07572925e-01 -9.61720347e-01
9.70591679e-02 -1.32129264e+00 7.17898488e-01 -1.36625230e-01
-1.04179621e+00 -1.12169075e+00 -2.02727184e-01 6.55567110e-01
3.69327106e-02 -1.04336761e-01 1.48250020e+00 -1.29348958e+00
2.59333134e-01 -8.42023343e-02 2.21192077e-01 6.23242199e-01
5.00855684e-01 -6.40992895e-02 -1.10435498e+00 2.02385813e-01
8.82315159e-01 -8.04143786e-01 -2.07392395e-01 1.24148831e-01
5.23452997e-01 -9.85648692e-01 1.38082027e-01 -4.37252939e-01
5.88314950e-01 5.21410286e-01 2.87372798e-01 -2.36838207e-01
1.48681169e-02 1.59114182e+00 5.19607067e-01 2.63646096e-01
6.06472492e-01 5.31142175e-01 1.88321188e-01 5.69460273e-01
2.90348142e-01 -2.25951433e-01 7.29278088e-01 1.30436516e+00
6.11568809e-01 -5.36722898e-01 -8.12937737e-01 5.69076836e-01
-1.83962321e+00 -8.55976522e-01 -5.04588544e-01 1.75157607e+00
1.27087760e+00 2.16202453e-01 -1.17283829e-01 -2.17268541e-02
8.81169558e-01 3.30441743e-01 -3.98455739e-01 -9.19833779e-01
-1.32868320e-01 -5.02067566e-01 -6.81061327e-01 1.03627789e+00
-8.15414250e-01 8.97484124e-01 6.14934587e+00 4.08216655e-01
-5.91488063e-01 2.40172654e-01 5.93509018e-01 3.56750965e-01
-3.39197874e-01 3.36827636e-02 -2.96226442e-01 1.34545401e-01
1.22648263e+00 -1.94229390e-02 5.25618851e-01 6.51082873e-01
3.13418478e-01 -5.89579880e-01 -1.87651324e+00 6.45076692e-01
1.04204096e-01 -1.05687928e+00 -2.25092351e-01 -3.86872083e-01
2.78991014e-02 -2.80575454e-01 -4.08333689e-01 3.22805375e-01
5.10457829e-02 -8.64463091e-01 9.24952626e-01 7.81009793e-02
3.09739590e-01 -2.86998779e-01 8.80405486e-01 9.54049468e-01
-6.14963353e-01 7.91862458e-02 -2.17087701e-01 -6.56123996e-01
2.08763674e-01 -5.27542457e-02 -1.01813567e+00 -2.32531205e-02
4.56598341e-01 5.34416914e-01 -1.63329974e-01 -8.16223919e-02
-5.41853070e-01 8.30017805e-01 -1.25402048e-01 -2.95699716e-01
2.09645852e-01 -2.45961532e-01 9.69755650e-01 1.15587068e+00
-3.46170872e-01 7.25531876e-01 9.29715857e-02 6.61515534e-01
6.14913460e-03 2.37836897e-01 -7.50229895e-01 -4.52848285e-01
5.35731852e-01 1.03439724e+00 -4.58935440e-01 -6.06067836e-01
-5.74158907e-01 8.97569776e-01 8.29262584e-02 3.08075398e-01
-4.51171547e-01 -3.51233631e-02 7.10056484e-01 -1.94317922e-01
-4.02315050e-01 8.02918300e-02 2.59934932e-01 -1.01134276e+00
6.92913681e-02 -1.23806453e+00 2.07864657e-01 -1.03212178e+00
-1.64583254e+00 8.12890053e-01 2.25517988e-01 -7.57999361e-01
-7.69217670e-01 -5.73744655e-01 -6.91937566e-01 7.82652795e-01
-5.78514934e-01 -6.74253881e-01 2.15328455e-01 3.51790130e-01
1.29099488e+00 -3.37721892e-02 1.31819415e+00 -2.13001862e-01
-3.83115381e-01 -8.60219747e-02 -5.31155348e-01 2.95180351e-01
5.86298525e-01 -1.22328389e+00 1.12994157e-01 1.88886762e-01
4.08591479e-02 8.55695724e-01 1.34646988e+00 -2.97673017e-01
-1.18635297e+00 -3.14727396e-01 1.31266928e+00 -5.83447993e-01
1.11102653e+00 -4.26752269e-01 -1.15008461e+00 1.05711305e+00
9.83124971e-01 -6.11442327e-01 1.15673053e+00 1.75544024e-01
-4.78992760e-02 3.97588521e-01 -1.24537814e+00 4.81599391e-01
7.92920291e-01 -7.80607224e-01 -1.62510347e+00 6.68203831e-01
1.22735739e+00 -5.34640133e-01 -6.26763463e-01 -5.47173843e-02
3.39224190e-01 -1.10158479e+00 5.10051191e-01 -1.01928973e+00
5.70923150e-01 1.24308489e-01 -7.84443796e-01 -1.11184931e+00
5.54159045e-01 -5.42068183e-01 2.20645487e-01 1.47403800e+00
5.01659513e-01 -5.09087861e-01 2.05737129e-01 1.23163021e+00
-3.07172239e-01 -4.61484015e-01 -9.94941473e-01 -2.22702965e-01
1.40570588e-02 -9.00877595e-01 5.11082828e-01 1.13493848e+00
8.31497252e-01 1.16199565e+00 -1.21867217e-01 -6.53715655e-02
1.42106280e-01 2.33919062e-02 7.74602175e-01 -8.38447332e-01
-5.08401692e-01 4.83008027e-02 9.54945460e-02 -1.51793754e+00
5.97025156e-01 -4.89179224e-01 3.21645916e-01 -1.00426316e+00
-3.20156366e-02 3.91638419e-03 6.42712057e-01 4.08055574e-01
1.59183994e-01 -5.32197952e-01 2.50257012e-02 1.68617934e-01
-4.88216519e-01 5.60849667e-01 1.12377584e+00 -1.18981913e-01
-1.80441171e-01 7.06888810e-02 -4.12522227e-01 1.30702722e+00
6.28122985e-01 -4.29512531e-01 -5.39250493e-01 -3.38050187e-01
7.09652081e-02 9.39509571e-01 1.48274645e-01 -2.05878392e-01
5.65459132e-02 -2.59737134e-01 -2.92150140e-01 -3.69843662e-01
5.55864096e-01 -6.91830873e-01 7.12340251e-02 3.97240035e-02
-9.62461710e-01 2.91471452e-01 7.13466257e-02 3.27734232e-01
-6.18952572e-01 -7.56368160e-01 4.34175074e-01 -2.79315978e-01
-6.72230542e-01 -7.16005623e-01 -1.04187799e+00 2.39394903e-01
7.77333677e-01 -4.04932685e-02 -3.71293128e-01 -1.03576159e+00
-1.18369055e+00 2.09046394e-01 3.71543407e-01 5.70744097e-01
6.50382817e-01 -8.53609324e-01 -5.91918170e-01 7.83207566e-02
1.57862142e-01 -1.67572454e-01 2.38358200e-01 7.52344608e-01
-3.05714667e-01 4.00198400e-01 1.11514099e-01 -4.85814840e-01
-1.33462465e+00 2.64444411e-01 3.53051573e-01 2.27733478e-01
-5.06012976e-01 1.02629983e+00 7.46160507e-01 -8.08879733e-01
2.92611390e-01 -5.05049407e-01 -5.16445339e-01 -2.69131977e-02
8.04163337e-01 -2.03680471e-01 -1.92074627e-01 -1.03141332e+00
-1.43466696e-01 -2.10959017e-01 -5.02053089e-02 -3.41524661e-01
1.17456985e+00 -7.16042638e-01 -8.07077289e-01 1.32793903e+00
1.26819634e+00 3.56914788e-01 -8.21710348e-01 -2.70835042e-01
3.62918109e-01 -3.28011453e-01 -5.58822930e-01 -8.60986114e-01
-3.38001460e-01 5.07043004e-01 -3.56176794e-01 1.00490344e+00
6.34205043e-01 6.18634045e-01 6.91492796e-01 7.11857438e-01
3.65421504e-01 -1.02054906e+00 2.60333300e-01 8.68133605e-01
1.62488961e+00 -1.01322305e+00 -6.28121078e-01 -7.27695584e-01
-1.02220941e+00 1.28787923e+00 6.43619239e-01 3.36514115e-01
4.91795123e-01 1.71761125e-01 4.81564462e-01 -4.21976238e-01
-1.33908904e+00 2.79002666e-01 -1.69410631e-01 2.80783147e-01
8.22628558e-01 1.73361436e-01 -4.81508315e-01 1.17526138e+00
-7.06383944e-01 -7.11982012e-01 9.51990843e-01 8.14621985e-01
-4.33116049e-01 -1.04685950e+00 -4.15128261e-01 2.26624697e-01
-4.46650624e-01 -1.31493220e-02 -1.04592347e+00 6.67153478e-01
-8.00473094e-02 1.73167610e+00 1.52511731e-01 -3.66303265e-01
5.48458695e-01 3.97253811e-01 1.35537058e-01 -1.08981478e+00
-7.86849022e-01 -1.58870630e-02 1.14571965e+00 -5.16344488e-01
-6.35130703e-01 -5.77076137e-01 -1.50544095e+00 -2.88927168e-01
-2.83511102e-01 7.62641191e-01 4.18736488e-01 1.47796178e+00
-3.97811502e-01 5.84232733e-02 6.29194856e-01 -4.22521293e-01
-5.40652812e-01 -1.37133062e+00 -6.27399027e-01 5.97650528e-01
4.93977249e-01 -1.47085607e-01 -6.16679490e-01 2.33789638e-01] | [12.567962646484375, 7.895253658294678] |
c462ab2f-b142-4fe4-aa8a-de553bd1b9ae | k-radar-4d-radar-object-detection-dataset-and | 2206.08171 | null | https://arxiv.org/abs/2206.08171v3 | https://arxiv.org/pdf/2206.08171v3.pdf | K-Radar: 4D Radar Object Detection for Autonomous Driving in Various Weather Conditions | Unlike RGB cameras that use visible light bands (384$\sim$769 THz) and Lidars that use infrared bands (361$\sim$331 THz), Radars use relatively longer wavelength radio bands (77$\sim$81 GHz), resulting in robust measurements in adverse weathers. Unfortunately, existing Radar datasets only contain a relatively small number of samples compared to the existing camera and Lidar datasets. This may hinder the development of sophisticated data-driven deep learning techniques for Radar-based perception. Moreover, most of the existing Radar datasets only provide 3D Radar tensor (3DRT) data that contain power measurements along the Doppler, range, and azimuth dimensions. As there is no elevation information, it is challenging to estimate the 3D bounding box of an object from 3DRT. In this work, we introduce KAIST-Radar (K-Radar), a novel large-scale object detection dataset and benchmark that contains 35K frames of 4D Radar tensor (4DRT) data with power measurements along the Doppler, range, azimuth, and elevation dimensions, together with carefully annotated 3D bounding box labels of objects on the roads. K-Radar includes challenging driving conditions such as adverse weathers (fog, rain, and snow) on various road structures (urban, suburban roads, alleyways, and highways). In addition to the 4DRT, we provide auxiliary measurements from carefully calibrated high-resolution Lidars, surround stereo cameras, and RTK-GPS. We also provide 4DRT-based object detection baseline neural networks (baseline NNs) and show that the height information is crucial for 3D object detection. And by comparing the baseline NN with a similarly-structured Lidar-based neural network, we demonstrate that 4D Radar is a more robust sensor for adverse weather conditions. All codes are available at https://github.com/kaist-avelab/k-radar. | ['Kevin Tirta Wijaya', 'Seung-Hyun Kong', 'Dong-Hee Paek'] | 2022-06-16 | null | null | null | null | ['radar-object-detection'] | ['robots'] | [ 1.15338609e-01 -2.87490517e-01 1.12184085e-01 -5.09213924e-01
-7.46351480e-01 -6.83775246e-01 4.00370985e-01 -4.75417316e-01
-5.18302917e-01 4.47020531e-01 -2.78533936e-01 -6.01280928e-01
-2.77834922e-01 -1.30057693e+00 -5.86078465e-01 -8.05270314e-01
-3.31358045e-01 4.19983745e-01 7.25790262e-02 -3.08761716e-01
-7.15258569e-02 1.01635218e+00 -1.47761381e+00 -3.36322755e-01
6.33444071e-01 1.30463445e+00 -9.20616742e-03 6.76544785e-01
7.05421418e-02 1.14867836e-01 -5.36955357e-01 -1.09485790e-01
8.42993319e-01 2.70409226e-01 2.28734761e-01 -2.94789106e-01
1.15005231e+00 -7.68540859e-01 -7.58841574e-01 8.95730674e-01
6.66694462e-01 -9.51623544e-02 5.91171741e-01 -1.23424995e+00
-5.44369221e-01 -9.60880294e-02 -8.38646233e-01 2.61806309e-01
-1.08577110e-01 2.74202049e-01 6.75640285e-01 -7.35651195e-01
1.78054586e-01 1.07563329e+00 5.53237975e-01 1.92263722e-01
-7.82777846e-01 -1.23355067e+00 -5.76880164e-02 2.05902144e-01
-1.43795502e+00 -1.51522696e-01 6.95758522e-01 -3.83873761e-01
8.39100957e-01 1.60778791e-01 6.79951608e-01 1.10829091e+00
2.15710893e-01 1.28515705e-01 1.28936243e+00 4.01855744e-02
1.85005546e-01 -2.56021649e-01 3.26394796e-01 6.02546930e-01
7.30396748e-01 7.81261981e-01 -5.50958633e-01 5.70602976e-02
6.98652804e-01 1.01754807e-01 -3.44305992e-01 -7.75243267e-02
-1.09489560e+00 7.18783557e-01 8.05313468e-01 -2.30829105e-01
-2.30534881e-01 5.07809222e-01 -2.28285909e-01 2.60090739e-01
4.54642713e-01 1.66071683e-01 -3.71392667e-01 1.52019635e-01
-6.23753428e-01 2.92988956e-01 2.59053707e-01 1.13823545e+00
1.20278883e+00 4.76039857e-01 1.58620730e-01 3.86747450e-01
2.89362043e-01 1.63699102e+00 -4.25983340e-01 -1.06332564e+00
5.54047823e-01 3.98514181e-01 3.00864518e-01 -7.43088663e-01
-7.69014359e-01 -7.34509587e-01 -9.15008247e-01 6.06200039e-01
2.86926717e-01 -3.18668216e-01 -1.29491484e+00 1.37168181e+00
4.28509206e-01 2.08085001e-01 3.26757342e-01 1.33535826e+00
1.05627501e+00 4.62831497e-01 -5.16562939e-01 1.02073193e-01
1.35940099e+00 -1.82548225e-01 -2.34773695e-01 -5.98877490e-01
4.60112751e-01 -4.68411952e-01 7.87743330e-01 2.75066346e-01
-3.08067739e-01 -4.70881701e-01 -1.08378649e+00 2.20936295e-02
-5.37359059e-01 1.23059064e-01 8.08233619e-01 8.40223849e-01
-6.97443843e-01 -4.03899588e-02 -6.74170375e-01 -3.14529389e-01
4.70890045e-01 -3.25964676e-04 -6.22798912e-02 -5.06968856e-01
-1.16722810e+00 7.91357040e-01 -2.63001412e-01 5.40626168e-01
-8.47330928e-01 -1.03122258e+00 -7.07626343e-01 -3.69107753e-01
2.76813120e-01 -3.74841154e-01 7.96450794e-01 9.01037976e-02
-9.82806325e-01 5.82546890e-01 4.04329114e-02 -5.30634165e-01
1.94202349e-01 -3.23926538e-01 -6.43582284e-01 2.65333503e-01
-9.29944508e-04 7.99528837e-01 6.99124098e-01 -1.23540926e+00
-6.18471682e-01 -6.50362313e-01 2.29890555e-01 -5.01975999e-04
9.91778970e-02 -2.92284995e-01 -7.83449113e-02 -6.89512938e-02
2.99910098e-01 -9.07532334e-01 -1.05327360e-01 1.93617567e-01
-4.03514475e-01 1.82685569e-01 1.11202717e+00 -1.23935901e-01
3.07871819e-01 -2.11500001e+00 -5.63630462e-01 7.84976929e-02
3.38549346e-01 1.11300377e-02 -1.66989461e-01 6.14202879e-02
2.44583488e-01 -6.40216470e-03 -1.99867249e-01 5.14970794e-02
-2.36106236e-02 3.72886211e-01 -6.85495913e-01 6.24267161e-01
9.41587240e-02 6.42585516e-01 -5.12832522e-01 -2.21170764e-02
3.63959581e-01 6.32096648e-01 2.84140781e-02 -1.39916465e-01
-1.73054814e-01 3.53525251e-01 -5.15965223e-01 1.28732085e+00
1.48366916e+00 3.57690364e-01 -5.67789316e-01 -6.37847126e-01
-4.82177198e-01 8.76736566e-02 -8.74702156e-01 1.04722285e+00
-3.38092566e-01 1.07964385e+00 2.59643018e-01 -6.89530790e-01
1.23418188e+00 -1.56277403e-01 4.08651948e-01 -1.21354473e+00
3.11954636e-02 8.08769763e-02 -8.96647871e-02 -4.42500830e-01
5.52961290e-01 -3.46014947e-01 -2.08276063e-01 3.18329245e-01
-3.35371941e-01 -6.24204338e-01 -3.64661701e-02 1.11748733e-01
1.34751093e+00 -2.36904219e-01 -4.54184234e-01 2.53554225e-01
-1.63043991e-01 3.32351416e-01 6.35628700e-01 9.92990673e-01
-1.76090732e-01 5.96244991e-01 6.27210364e-02 -6.64734423e-01
-8.41055155e-01 -1.44345188e+00 -5.95492125e-01 4.27585304e-01
3.68523002e-01 6.47241101e-02 -3.39242965e-02 -3.71717304e-01
4.90310580e-01 5.97774088e-01 -2.42192835e-01 1.32418443e-02
-3.82747412e-01 -1.19127810e+00 8.47948909e-01 4.57745522e-01
8.18811178e-01 -3.68635327e-01 -8.27442646e-01 -1.77069008e-01
-4.55062510e-03 -1.72217226e+00 2.53427804e-01 2.32674509e-01
-8.05722833e-01 -9.86944139e-01 -1.16756499e-01 1.26427323e-01
3.49249035e-01 9.74528372e-01 9.44325864e-01 -3.11931342e-01
-5.72506547e-01 6.27376080e-01 -3.68379980e-01 -9.37929869e-01
4.60594565e-01 -3.98757637e-01 3.83292496e-01 -2.52918422e-01
5.03562748e-01 -7.21552968e-01 -5.67485571e-01 4.01829422e-01
-3.92576218e-01 -5.13311401e-02 9.58449423e-01 8.76643583e-02
4.38248813e-01 1.29506141e-01 1.18593931e-01 -2.43292615e-01
-1.49386942e-01 -2.69621402e-01 -1.22089529e+00 -3.66341084e-01
-3.05429280e-01 -3.66038144e-01 2.02794820e-01 -1.44297741e-02
-6.87697291e-01 -8.76197070e-02 8.35497454e-02 -5.64683020e-01
-3.62785429e-01 2.71361232e-01 -3.03487349e-02 -5.48817158e-01
7.97920704e-01 1.04674995e-01 -3.84800076e-01 -3.83226544e-01
1.93753317e-01 7.59163201e-01 6.54253900e-01 -4.25616622e-01
1.60207403e+00 1.02996528e+00 4.55027014e-01 -1.38746583e+00
-9.94117260e-01 -3.35941911e-01 -5.06189704e-01 -5.31399906e-01
6.85594201e-01 -1.41413116e+00 -7.08199918e-01 3.77457589e-01
-8.71553123e-01 -4.41034198e-01 -7.30815306e-02 1.00921047e+00
-1.10720612e-01 6.70237094e-02 -1.14240132e-01 -1.21067953e+00
-3.36005867e-01 -6.47854924e-01 1.22926974e+00 3.26875389e-01
4.07533616e-01 -2.95659781e-01 -1.79985598e-01 7.20519722e-01
4.99220014e-01 5.17807305e-01 7.12969065e-01 2.20305786e-01
-1.35805905e+00 -3.23339999e-01 -6.88002646e-01 1.77739814e-01
-1.40167966e-01 -2.13941522e-02 -1.12834573e+00 -1.77260518e-01
-2.03678817e-01 -2.22180635e-01 1.16642892e+00 5.03857136e-01
9.11093235e-01 3.03891063e-01 -3.08574259e-01 1.18800724e+00
1.42754090e+00 1.02571763e-01 6.07171774e-01 1.17508881e-01
7.91625798e-01 4.57872987e-01 8.14260542e-01 2.85661817e-01
4.87932891e-01 4.46389705e-01 9.83788013e-01 -1.30410418e-01
7.34329829e-03 3.03159684e-01 5.28326035e-01 6.14269525e-02
-3.79694372e-01 -2.13278636e-01 -9.86942589e-01 2.96455294e-01
-1.21527028e+00 -8.25898886e-01 -6.04119241e-01 2.19702506e+00
9.51737314e-02 3.06625724e-01 -3.28558654e-01 -1.55288026e-01
5.33558011e-01 4.03360486e-01 -7.69077599e-01 9.83487256e-03
-3.13675314e-01 3.35564435e-01 1.29059744e+00 4.36354488e-01
-1.06363082e+00 6.22092307e-01 5.53394413e+00 3.78820866e-01
-1.21977949e+00 -1.12613318e-02 -5.94428591e-02 -4.94330227e-01
-4.81635481e-01 4.34343070e-02 -1.18665707e+00 1.87262684e-01
9.84266222e-01 2.59167641e-01 3.03210586e-01 6.34268105e-01
3.40421081e-01 -4.33986366e-01 -5.72582006e-01 1.04860127e+00
-2.39245772e-01 -1.27668965e+00 -2.55972564e-01 4.66627032e-01
2.26296023e-01 8.96039367e-01 8.06162953e-02 3.06393653e-01
4.32851195e-01 -7.85963118e-01 6.09921277e-01 5.22727668e-01
1.10650253e+00 -7.18493521e-01 5.54517388e-01 2.85216928e-01
-1.18404818e+00 -1.32627580e-02 -7.87298560e-01 -1.62983969e-01
5.02266698e-02 1.27158248e+00 -6.02267444e-01 6.75115645e-01
1.18208408e+00 6.43964529e-01 -3.22292328e-01 7.60143876e-01
-3.36985081e-01 6.11934662e-01 -7.83639669e-01 2.18646243e-01
3.54020625e-01 -4.46471393e-01 7.86665857e-01 9.41869676e-01
8.34442735e-01 3.15816134e-01 1.10931188e-01 9.06494141e-01
-1.76632270e-01 -8.17153394e-01 -9.82126474e-01 3.60751078e-02
5.82230330e-01 1.73013186e+00 -3.69658738e-01 9.06576514e-02
-4.76719350e-01 1.55929238e-01 -2.88341880e-01 6.35695875e-01
-1.00608516e+00 -3.73811990e-01 1.34190381e+00 2.32837528e-01
3.24121445e-01 -9.63508427e-01 -2.46160761e-01 -8.40713978e-01
8.45776498e-02 -1.42949671e-01 -3.16705033e-02 -1.19366968e+00
-1.27066457e+00 1.55265346e-01 7.00456463e-03 -1.45476925e+00
4.25450832e-01 -9.17199910e-01 -5.82868934e-01 8.75900328e-01
-2.16897154e+00 -1.26843059e+00 -1.07049441e+00 3.31813425e-01
-2.20071167e-01 4.05626968e-02 4.40424889e-01 3.13003361e-01
-5.58758974e-01 8.64884928e-02 1.07234441e-01 3.18631887e-01
7.68086553e-01 -1.04786718e+00 5.84215939e-01 7.24266052e-01
-9.76984948e-02 5.01065291e-02 5.79687297e-01 -7.49763012e-01
-2.03235388e+00 -1.47266483e+00 2.75106937e-01 -4.57005352e-01
6.41641796e-01 -7.01672614e-01 -5.05875766e-01 6.17727697e-01
-3.23963553e-01 5.09788871e-01 4.95434642e-01 3.09859831e-02
-6.21647060e-01 -8.92804384e-01 -1.15658712e+00 2.27013707e-01
1.26119041e+00 -5.75164318e-01 -1.34737328e-01 5.25691807e-01
7.95539677e-01 -6.49381995e-01 -6.17881596e-01 8.48405480e-01
7.41555393e-01 -1.04841399e+00 1.18448830e+00 2.68475637e-02
4.50498657e-03 -6.27684832e-01 -6.19261980e-01 -8.83334100e-01
-2.36385390e-01 -1.20935723e-01 -8.72298256e-02 8.57300103e-01
1.41442373e-01 -8.21276069e-01 1.00025797e+00 4.03061435e-02
-4.16464806e-01 -3.44507873e-01 -1.35195124e+00 -1.09527910e+00
-2.07034364e-01 -9.75359261e-01 5.43605983e-01 6.98920012e-01
-9.34070051e-01 2.36535743e-01 -1.23020887e-01 1.12457263e+00
1.29407132e+00 5.34654975e-01 9.56386387e-01 -1.47968662e+00
2.71061152e-01 6.63270503e-02 -5.59280157e-01 -9.07648444e-01
-6.34559020e-02 -6.18331432e-01 -5.26370816e-02 -1.46440065e+00
-2.95122027e-01 -7.70199358e-01 3.71457450e-02 4.14962262e-01
4.93134677e-01 5.17919898e-01 -6.48466125e-02 1.31031364e-01
-7.67865628e-02 6.07723236e-01 9.87859488e-01 -3.58271718e-01
1.22929461e-01 1.91529877e-02 -4.89304304e-01 6.75910950e-01
7.28722751e-01 -4.26320732e-01 -6.40117675e-02 -8.28575969e-01
4.30275321e-01 -3.73023599e-02 9.77082372e-01 -1.51437533e+00
1.43918931e-01 -3.74282926e-01 7.42375553e-01 -1.29636717e+00
7.95664132e-01 -8.07249904e-01 -1.36040235e-02 3.08206797e-01
5.48210979e-01 -3.28830123e-01 2.49330342e-01 6.27794206e-01
1.17523998e-01 3.22704762e-01 8.83142412e-01 -5.48171960e-02
-8.51422250e-01 6.88665569e-01 -4.19690102e-01 -1.42935500e-01
7.41506815e-01 -5.19556701e-01 -1.00047398e+00 -4.71393466e-01
-7.22213760e-02 4.54703510e-01 2.99780548e-01 1.87551990e-01
7.65431523e-01 -1.11489725e+00 -7.28993833e-01 2.78092384e-01
3.94180596e-01 3.75681520e-01 3.80803794e-01 7.95563757e-01
-4.33424681e-01 3.33471060e-01 -1.00453183e-01 -8.62246454e-01
-1.04725766e+00 -1.17399578e-03 5.17204106e-01 5.19555748e-01
-7.34062791e-01 6.97023332e-01 6.52596876e-02 -6.46241903e-01
-2.84957904e-02 -6.25362217e-01 1.24795973e-01 6.50205314e-02
4.85551417e-01 2.47367531e-01 2.39958689e-01 -4.89738017e-01
-4.82688725e-01 1.02123487e+00 3.07690620e-01 -1.16581134e-02
1.37463820e+00 -7.19259232e-02 7.59133399e-02 1.86553881e-01
7.02046752e-01 6.73395172e-02 -1.32738495e+00 -3.11408430e-01
-5.60699701e-01 -5.39034724e-01 6.06951058e-01 -5.82671642e-01
-1.13019395e+00 1.10633647e+00 8.88762414e-01 -3.39838043e-02
1.07319891e+00 4.14923206e-02 8.42260599e-01 1.03569698e+00
5.88199377e-01 -8.80639791e-01 -2.45849460e-01 9.16655481e-01
5.56935370e-01 -1.14453244e+00 2.51832247e-01 -3.29664946e-01
-1.06402732e-01 1.16878736e+00 7.04255760e-01 6.55183569e-02
4.98228908e-01 5.56815028e-01 3.02536249e-01 -5.15253782e-01
-4.70882595e-01 -7.11432636e-01 -1.68498695e-01 7.75736094e-01
-1.22953586e-01 1.57167554e-01 5.28291523e-01 2.87316471e-01
-4.22728121e-01 -3.31241846e-01 7.69073308e-01 8.76773655e-01
-8.92324448e-01 -5.42570710e-01 -8.20070803e-01 6.55298054e-01
3.07057023e-01 2.74211019e-02 -2.45574042e-01 9.68162537e-01
5.00189126e-01 9.19813871e-01 3.99488568e-01 -5.98983526e-01
4.10986394e-01 -2.86973596e-01 6.47392571e-01 -3.77121121e-01
1.67199403e-01 -1.21383630e-01 2.37987846e-01 -5.66480637e-01
-3.01991105e-01 -5.05496502e-01 -1.22029424e+00 -4.86535639e-01
-1.95012018e-01 -1.27288491e-01 1.02975106e+00 8.02403390e-01
3.15419167e-01 1.96669936e-01 6.96036458e-01 -9.83414710e-01
-1.22060843e-01 -8.41648042e-01 -9.29604769e-01 -5.83259821e-01
6.52604759e-01 -9.91134942e-01 -6.23046339e-01 -4.89587396e-01] | [7.848667621612549, -1.5012108087539673] |
4957c1c3-c8ef-40f7-9b46-8356ddd5af58 | scrolls-standardized-comparison-over-long | 2201.03533 | null | https://arxiv.org/abs/2201.03533v2 | https://arxiv.org/pdf/2201.03533v2.pdf | SCROLLS: Standardized CompaRison Over Long Language Sequences | NLP benchmarks have largely focused on short texts, such as sentences and paragraphs, even though long texts comprise a considerable amount of natural language in the wild. We introduce SCROLLS, a suite of tasks that require reasoning over long texts. We examine existing long-text datasets, and handpick ones where the text is naturally long, while prioritizing tasks that involve synthesizing information across the input. SCROLLS contains summarization, question answering, and natural language inference tasks, covering multiple domains, including literature, science, business, and entertainment. Initial baselines, including Longformer Encoder-Decoder, indicate that there is ample room for improvement on SCROLLS. We make all datasets available in a unified text-to-text format and host a live leaderboard to facilitate research on model architecture and pretraining methods. | ['Omer Levy', 'Jonathan Berant', 'Mor Geva', 'Wenhan Xiong', 'Ankit Gupta', 'Adi Haviv', 'Ori Yoran', 'Avia Efrat', 'Maor Ivgi', 'Elad Segal', 'Uri Shaham'] | 2022-01-10 | null | null | null | null | ['long-range-modeling'] | ['natural-language-processing'] | [ 4.78540063e-01 4.07607406e-01 -4.57846463e-01 -4.44115400e-01
-1.24884057e+00 -8.22927952e-01 8.53888810e-01 4.86383259e-01
-3.98270100e-01 9.36538875e-01 9.45306778e-01 -6.71539903e-01
7.13652968e-02 -6.29155815e-01 -8.63815486e-01 1.90501511e-02
3.53398174e-01 5.03627598e-01 -7.08628446e-02 -3.83774936e-01
3.21272165e-01 -2.36533076e-01 -9.59604442e-01 8.11411440e-01
7.93645740e-01 7.21720994e-01 1.52592346e-01 8.89478147e-01
-4.31061745e-01 1.12647259e+00 -7.14016140e-01 -8.12619865e-01
-9.32824761e-02 -3.82688582e-01 -1.19333637e+00 -1.81861192e-01
7.02096820e-01 -3.47076356e-01 -4.14746433e-01 6.62405670e-01
5.24877667e-01 1.49934262e-01 6.66285574e-01 -9.73447919e-01
-9.48473632e-01 1.23992980e+00 -4.06661034e-01 3.59619677e-01
6.86340094e-01 3.14352572e-01 1.64267707e+00 -8.26565862e-01
8.35980177e-01 1.23900938e+00 4.69131112e-01 4.50233191e-01
-1.04790676e+00 -2.92832792e-01 2.90736586e-01 5.65363839e-02
-6.67761445e-01 -7.84066498e-01 4.15026993e-01 -3.43436569e-01
1.63145697e+00 2.43663743e-01 1.40120471e-02 1.53654039e+00
2.63362706e-01 1.19752753e+00 3.81831437e-01 -3.45012039e-01
-9.85574210e-04 -1.52993426e-01 4.32770193e-01 3.59363467e-01
3.23141098e-01 -4.71542865e-01 -6.11285686e-01 1.20160125e-01
1.21062472e-01 -2.39073455e-01 -2.56138623e-01 4.48594958e-01
-1.49816227e+00 6.90777659e-01 5.07708900e-02 8.62959996e-02
-2.35613465e-01 1.71572063e-02 7.85839736e-01 4.46817398e-01
5.42902827e-01 8.92305493e-01 -7.22119212e-01 -4.66088653e-01
-1.03791356e+00 5.25858283e-01 1.41570520e+00 1.29647803e+00
3.97695899e-01 -9.26342160e-02 -4.92907584e-01 9.11391199e-01
-2.06507370e-01 5.37974954e-01 4.27891880e-01 -9.25879061e-01
1.36929476e+00 5.79583466e-01 1.20671138e-01 -9.22052383e-01
-2.83110768e-01 -2.13687077e-01 -8.34973931e-01 -5.79158485e-01
2.50873417e-01 -6.81599200e-01 -6.52368963e-01 1.49823308e+00
-1.87696680e-01 -2.53132582e-01 3.91973138e-01 4.62730229e-01
1.42947412e+00 1.11949027e+00 -9.79509130e-02 -1.08845167e-01
1.22928584e+00 -1.31185019e+00 -7.02811182e-01 -7.79068351e-01
7.49554813e-01 -6.44823790e-01 1.46185470e+00 4.00202245e-01
-1.47093344e+00 -3.21998656e-01 -9.55529809e-01 -9.45812345e-01
-6.38138354e-01 4.18794453e-02 3.10443640e-01 -1.84258670e-01
-9.64771628e-01 6.25278473e-01 -5.27803540e-01 -3.41606110e-01
3.37970823e-01 -1.68149069e-01 -2.32581005e-01 -1.36598513e-01
-1.40486932e+00 1.11702251e+00 4.39151406e-01 -1.09167350e-02
-5.37872553e-01 -7.29208469e-01 -1.23646605e+00 3.54943901e-01
6.40198052e-01 -8.64759922e-01 1.78413868e+00 -5.83078921e-01
-1.38708448e+00 8.20908666e-01 -2.60227621e-01 -6.57534420e-01
5.25963902e-01 -4.67463076e-01 -2.63047665e-01 7.79081136e-02
1.91418618e-01 6.36062503e-01 4.83126104e-01 -6.50564194e-01
-5.74747324e-01 4.19032127e-02 3.70968103e-01 2.59176582e-01
-1.36222661e-01 1.47538096e-01 -4.69087243e-01 -6.85632050e-01
-5.01869559e-01 -4.85370278e-01 -7.19475150e-02 -3.19856793e-01
-9.00568664e-01 -4.62492079e-01 3.27573776e-01 -9.48427856e-01
1.41813338e+00 -1.74952447e+00 2.46401235e-01 -4.16185379e-01
1.53021038e-01 -1.18121423e-01 -4.41590816e-01 9.15440321e-01
4.10105940e-03 3.68852377e-01 -3.90183866e-01 -4.96140808e-01
3.72852236e-01 1.49318337e-01 -7.90875018e-01 -8.90508294e-02
5.04745066e-01 1.30014277e+00 -8.62732828e-01 -6.01609886e-01
-1.38448671e-01 -1.41321495e-01 -4.99676108e-01 1.25142792e-03
-8.69269431e-01 -3.98418158e-02 -5.06509900e-01 3.89126718e-01
1.14438310e-01 -5.03661752e-01 -3.09276044e-01 7.96528980e-02
-5.11709861e-02 1.19452739e+00 -5.56896031e-01 1.86745310e+00
-6.22004211e-01 1.08049631e+00 -5.98317664e-03 -7.63288796e-01
4.13719893e-01 2.76752204e-01 2.13078722e-01 -6.14238620e-01
1.23276405e-01 -5.62215969e-02 -2.19229147e-01 -8.31565320e-01
1.02795434e+00 1.61888793e-01 -2.81040460e-01 6.87723279e-01
4.11970206e-02 -4.15302277e-01 8.48924458e-01 7.16994464e-01
1.30067110e+00 -6.55895174e-02 2.26130128e-01 -7.34134391e-02
1.55331299e-01 2.19970152e-01 6.68049678e-02 9.02589977e-01
2.62097478e-01 4.82455194e-01 9.41085517e-01 -9.03843939e-02
-1.17421401e+00 -8.69041085e-01 -1.00240618e-01 1.31292117e+00
-2.07611695e-01 -9.43439126e-01 -5.29687047e-01 -6.70921504e-01
5.27901426e-02 1.25533056e+00 -4.60825324e-01 1.82934836e-01
-5.15790403e-01 -4.60583359e-01 6.11271381e-01 6.25621498e-01
2.12823734e-01 -1.19459164e+00 -2.83455491e-01 2.05537096e-01
-5.08659244e-01 -1.39554000e+00 -6.54367626e-01 2.55487502e-01
-6.75631225e-01 -9.71750736e-01 -5.68910956e-01 -7.34298885e-01
2.92670667e-01 -1.53258160e-01 1.63334930e+00 -2.33262926e-01
-8.81134868e-02 2.31323794e-01 -4.52796221e-01 -6.50017798e-01
-5.94018102e-01 6.75548196e-01 -4.52615470e-01 -6.22221231e-01
2.68285215e-01 -4.27567810e-01 -2.53615886e-01 -2.33943850e-01
-8.38212073e-01 4.65163678e-01 6.78449154e-01 8.12231600e-01
2.65998125e-01 -3.65813076e-01 8.80317569e-01 -1.25439847e+00
1.02258158e+00 -7.60345936e-01 -2.82017797e-01 3.43004823e-01
-3.00248295e-01 2.05766946e-01 9.29526925e-01 -1.78602755e-01
-8.89126718e-01 -6.69056654e-01 -4.60681915e-01 2.49919325e-01
-1.20793186e-01 9.95928526e-01 -9.20459908e-03 7.77952731e-01
7.60707378e-01 2.11228773e-01 -2.12856069e-01 -3.43668252e-01
5.42720377e-01 8.92256081e-01 5.64803660e-01 -6.78849638e-01
4.79020298e-01 -3.55592966e-02 -4.62979913e-01 -1.00284636e+00
-1.29967630e+00 -3.46382469e-01 -2.57126898e-01 2.43286729e-01
7.17384517e-01 -9.32844281e-01 -5.41914105e-01 6.98510706e-02
-1.40782166e+00 -7.54183471e-01 -4.21249390e-01 1.49697140e-01
-4.33878988e-01 3.60703826e-01 -9.69541430e-01 -2.90242821e-01
-7.04242945e-01 -8.53445530e-01 1.27307856e+00 -1.61646456e-02
-7.03025281e-01 -1.13072324e+00 -1.84986830e-01 4.45165128e-01
3.50313574e-01 1.14025913e-01 1.12738383e+00 -9.49844062e-01
-4.14941251e-01 -1.62361547e-01 -2.20454395e-01 3.39361727e-01
-7.49688819e-02 1.93269700e-01 -6.68848455e-01 -1.34468647e-02
-2.30087802e-01 -8.22417915e-01 1.16119409e+00 1.98071316e-01
1.28233695e+00 -7.38431573e-01 -1.30351350e-01 2.51315355e-01
9.84119594e-01 -2.00195402e-01 4.13959503e-01 1.96614996e-01
5.49652576e-01 5.70540547e-01 2.61981189e-01 4.02966827e-01
8.57793331e-01 7.14493692e-02 1.07755817e-01 5.34349047e-02
1.14235081e-01 -4.66571450e-01 5.00955760e-01 1.23844314e+00
5.74405253e-01 -6.77147210e-01 -9.63181615e-01 6.42711520e-01
-1.78682220e+00 -1.11695611e+00 3.15452516e-02 1.52822280e+00
1.39188027e+00 5.16586006e-01 -2.12806493e-01 -1.53066084e-01
3.08549970e-01 6.27163291e-01 -7.31982648e-01 -6.51122630e-01
-2.40034968e-01 1.63108781e-01 2.43817776e-01 7.40536690e-01
-9.88550365e-01 8.32933307e-01 7.06583881e+00 6.71771586e-01
-1.07243872e+00 -3.70670915e-01 6.29358172e-01 -3.90964061e-01
-6.35398030e-01 -1.51279971e-01 -9.02532697e-01 5.28956652e-01
1.20986617e+00 -6.02764666e-01 3.16366524e-01 6.75135732e-01
2.76239067e-01 -2.11050168e-01 -1.69329321e+00 7.28486359e-01
1.72225356e-01 -1.55060613e+00 3.25468063e-01 -5.06474733e-01
7.18693852e-01 4.09958690e-01 -2.05915853e-01 7.47474849e-01
4.90690291e-01 -1.22064185e+00 5.75723708e-01 3.32889527e-01
8.02619576e-01 -2.19310254e-01 5.73124766e-01 7.15418160e-01
-6.06752694e-01 1.26056314e-01 -2.80824333e-01 -3.01652223e-01
2.35649332e-01 5.01750946e-01 -7.61909783e-01 5.55184484e-01
2.64260143e-01 1.20231581e+00 -7.25340724e-01 4.90052223e-01
-3.97676736e-01 7.12152183e-01 -2.39223525e-01 -4.26106960e-01
3.79554212e-01 1.99075975e-02 4.05592233e-01 1.68354869e+00
2.99146231e-02 1.16794311e-01 1.05258726e-01 8.23527932e-01
-5.90961933e-01 1.18746623e-01 -5.61664999e-01 -5.23478389e-01
4.46810305e-01 9.68778491e-01 -3.15995604e-01 -6.39298260e-01
-8.00135314e-01 7.52648473e-01 2.59449154e-01 5.67643702e-01
-8.15112770e-01 -7.69417286e-01 2.69932032e-01 -8.55534673e-02
8.87309853e-03 -1.90824911e-01 -6.48996830e-01 -1.57225132e+00
2.19507352e-01 -1.10388255e+00 4.13808107e-01 -1.10795188e+00
-1.47490799e+00 4.56505179e-01 1.04326576e-01 -6.56703413e-01
-5.90067744e-01 -5.54040551e-01 -5.79513371e-01 6.85423911e-01
-1.54793227e+00 -8.46376896e-01 -1.32396355e-01 2.10213885e-01
1.18716180e+00 2.49505155e-02 5.43249905e-01 1.41662493e-01
-6.32592916e-01 5.42695642e-01 2.08232924e-01 3.19228172e-01
9.17777658e-01 -1.28589654e+00 9.62760210e-01 6.80341005e-01
2.01811150e-01 6.15816772e-01 7.53144681e-01 -5.16296446e-01
-1.60566461e+00 -1.07029831e+00 1.40091681e+00 -8.53580654e-01
9.32903826e-01 -4.73574400e-01 -8.02608609e-01 1.12208664e+00
8.09212089e-01 -6.20251596e-01 5.21330714e-01 2.62209952e-01
-2.25354791e-01 2.94161476e-02 -5.16425073e-01 9.27780509e-01
8.84661198e-01 -7.22236753e-01 -1.16753793e+00 5.90247273e-01
1.15458298e+00 -7.52059102e-01 -6.15552545e-01 1.48126081e-01
3.40948284e-01 -3.48563880e-01 6.97066963e-01 -9.15073395e-01
1.47308588e+00 2.98739016e-01 2.50880290e-02 -1.38273716e+00
8.03708211e-02 -8.06936979e-01 -3.24775189e-01 1.17668736e+00
9.96703625e-01 -3.59207034e-01 6.22229457e-01 7.49781489e-01
-4.08349484e-01 -9.86804605e-01 -4.36866611e-01 -4.49187458e-01
5.09858251e-01 -4.39867198e-01 4.40821320e-01 8.68638515e-01
5.17935574e-01 1.15901053e+00 -2.86818296e-02 -4.20383304e-01
1.74230352e-01 1.91671461e-01 7.46757627e-01 -8.88454616e-01
-2.70467609e-01 -7.69271314e-01 3.80955607e-01 -1.48270869e+00
3.74783218e-01 -1.12704563e+00 1.72214240e-01 -2.16894317e+00
2.03627080e-01 2.25911230e-01 2.77610511e-01 6.32298946e-01
-3.05620193e-01 -3.11914653e-01 2.47821696e-02 -2.23941892e-01
-9.18680549e-01 5.61372221e-01 1.41722310e+00 -3.21166456e-01
7.41667897e-02 -2.13960603e-01 -1.15867639e+00 6.18755817e-01
6.00040793e-01 -1.35061234e-01 -5.08888721e-01 -1.14709151e+00
6.41756356e-01 4.13623005e-01 8.45993832e-02 -6.82888210e-01
4.03583616e-01 -4.16068345e-01 1.64556846e-01 -7.16146827e-01
2.47432753e-01 -3.66953939e-01 -5.76466262e-01 -1.15398630e-01
-1.26044190e+00 3.20011437e-01 3.15697342e-01 2.30225563e-01
-4.24846292e-01 -3.07878882e-01 4.80529845e-01 -1.53762341e-01
-2.21093357e-01 2.68426716e-01 -2.48785108e-01 9.19789791e-01
5.56896508e-01 9.45357457e-02 -8.02343667e-01 -8.41101885e-01
-3.34008396e-01 7.91433811e-01 1.61115915e-01 5.57935715e-01
4.50406402e-01 -6.59763157e-01 -1.06133115e+00 -1.56868905e-01
-2.72691362e-02 5.02009630e-01 9.10436660e-02 5.01914501e-01
-5.75167954e-01 8.07913184e-01 2.67478049e-01 -6.07658960e-02
-8.20944190e-01 4.12923872e-01 -3.22369263e-02 -4.90491062e-01
-6.79749012e-01 7.65193403e-01 1.15178883e-01 -2.29701564e-01
5.65016866e-01 -9.56352055e-01 -1.17352523e-01 2.46393025e-01
5.87452948e-01 1.81988105e-01 7.12432265e-02 -7.03453869e-02
-4.05849703e-02 6.39761165e-02 -2.73395181e-01 -1.03591681e-01
1.43083632e+00 -2.66843200e-01 -1.62579015e-01 6.96578562e-01
1.29271221e+00 3.20662074e-02 -9.14509714e-01 -3.40519667e-01
3.89591575e-01 1.33540422e-01 -2.59964794e-01 -1.02667475e+00
-3.96162480e-01 1.14865208e+00 -8.45860422e-01 2.41659209e-01
8.34533393e-01 1.05452284e-01 1.10570431e+00 1.07848454e+00
-1.26640663e-01 -1.17722917e+00 1.41542360e-01 1.20881164e+00
1.23379993e+00 -1.30419827e+00 2.41009921e-01 2.10354775e-02
-8.26816738e-01 1.08309734e+00 6.28116488e-01 4.58076894e-02
3.17867190e-01 5.50022781e-01 -3.38932753e-01 -4.56218719e-02
-1.54380071e+00 8.34552199e-02 1.46245837e-01 6.84824213e-02
7.43781328e-01 -2.16653913e-01 -1.77104264e-01 1.01613677e+00
-6.97145283e-01 9.57769901e-03 7.86045969e-01 8.07797372e-01
-3.86033922e-01 -7.26057887e-01 1.41885161e-01 8.65742624e-01
-6.47739291e-01 -5.65661132e-01 -6.67885661e-01 5.27860940e-01
-5.00857472e-01 1.03932643e+00 8.57887119e-02 1.04214046e-02
3.62245321e-01 3.41295779e-01 3.29025894e-01 -8.61810803e-01
-7.01708317e-01 -3.56829464e-01 7.75094151e-01 -3.13241631e-01
1.39913261e-01 -4.77708906e-01 -1.21291375e+00 -5.89572608e-01
1.68798640e-02 1.85917512e-01 5.26461542e-01 1.08647716e+00
4.74802077e-01 7.11417675e-01 1.44584775e-01 -5.11788964e-01
-7.54164994e-01 -1.27428412e+00 -1.26844004e-01 2.74923593e-01
6.57325089e-01 1.20055310e-01 -1.83824748e-01 2.54472107e-01] | [11.651307106018066, 8.900276184082031] |
2a1e156d-5999-4168-a21d-ed40adc4a16a | open-cykg-an-open-cyber-threat-intelligence | null | null | https://www-sciencedirect-com.proxy.library.uu.nl/science/article/pii/S0950705121007863?dgcid=rss_sd_all | https://www-sciencedirect-com.proxy.library.uu.nl/science/article/pii/S0950705121007863/pdfft?md5=526ecd890d0da054e1084d452a43ec78&pid=1-s2.0-S0950705121007863-main.pdf | Open-CyKG: An Open Cyber Threat Intelligence Knowledge Graph | Instant analysis of cybersecurity reports is a fundamental challenge for security experts as an immeasurable amount of cyber information is generated on a daily basis, which necessitates automated information extraction tools to facilitate querying and retrieval of data. Hence, we present Open-CyKG: an Open Cyber Threat Intelligence (CTI) Knowledge Graph (KG) framework that is constructed using an attention-based neural Open Information Extraction (OIE) model to extract valuable cyber threat information from unstructured Advanced Persistent Threat (APT) reports. More specifically, we first identify relevant entities by developing a neural cybersecurity Named Entity Recognizer (NER) that aids in labeling relation triples generated by the OIE model. Afterward, the extracted structured data is canonicalized to build the KG by employing fusion techniques using word embeddings. As a result, security professionals can execute queries to retrieve valuable information from the Open-CyKG framework. Experimental results demonstrate that our proposed components that build up Open-CyKG outperform state-of-the-art models. | ['MarcoSpruit', 'InjySarhan'] | 2021-12-05 | null | null | null | knowledge-based-systems-2021-12 | ['deep-attention', 'deep-attention', 'open-information-extraction'] | ['computer-vision', 'natural-language-processing', 'natural-language-processing'] | [-1.70328002e-02 3.09931666e-01 -1.64799735e-01 2.80281663e-01
-6.76652312e-01 -8.40936244e-01 5.07333159e-01 8.20641398e-01
-3.45273852e-01 4.54883456e-01 4.07598108e-01 -6.68185472e-01
-4.05720592e-01 -1.29580200e+00 -4.25744087e-01 -4.73131193e-03
-3.51569951e-01 1.60421267e-01 2.77630263e-03 -2.88322210e-01
3.60887766e-01 5.51609933e-01 -9.36771095e-01 3.34235519e-01
8.38921905e-01 1.35350323e+00 -4.60649788e-01 2.95792073e-01
-4.88420933e-01 1.06037712e+00 -6.14507437e-01 -1.03243577e+00
2.82784075e-01 4.95871812e-01 -1.07401931e+00 -6.11924767e-01
-1.77113190e-01 2.92355008e-02 -4.12240297e-01 1.35144770e+00
1.71793044e-01 1.42721042e-01 4.93769258e-01 -1.45682430e+00
-1.10386693e+00 4.82628703e-01 -8.12951773e-02 4.89622205e-01
6.45324409e-01 -7.06115887e-02 1.16849875e+00 -9.69680786e-01
9.59221959e-01 9.30274248e-01 5.01220286e-01 1.90665826e-01
-3.79305959e-01 -6.52891397e-01 2.50188231e-01 5.86904168e-01
-1.28056324e+00 2.55578846e-01 8.22242022e-01 -3.79082710e-01
1.33555031e+00 2.80356854e-01 5.20825148e-01 1.08704162e+00
5.37430465e-01 3.94787252e-01 6.21013343e-01 -3.32417130e-01
1.68858364e-01 2.47665584e-01 8.44598293e-01 6.01171970e-01
9.76095438e-01 7.00088069e-02 -5.74094832e-01 -4.56423849e-01
2.04023346e-01 3.71580690e-01 -3.43982637e-01 8.73311162e-02
-8.80544007e-01 8.16325545e-01 8.10267329e-01 5.01955628e-01
-8.45184982e-01 -4.28232729e-01 6.77613020e-01 2.10947156e-01
5.44874489e-01 8.44010055e-01 -3.87650788e-01 7.76401833e-02
-1.23707853e-01 -1.05174994e-02 1.21072578e+00 7.98002064e-01
6.82874739e-01 -1.53394222e-01 1.71497583e-01 1.80316836e-01
2.61025399e-01 1.07036114e-01 2.84452736e-01 -1.29080683e-01
9.67186034e-01 1.30893111e+00 -1.20412178e-01 -1.71698976e+00
-2.38325238e-01 -4.84704435e-01 -5.53286910e-01 -2.15954289e-01
-3.12526375e-01 -2.82441556e-01 -5.87527812e-01 1.13245487e+00
5.66470325e-01 2.29530796e-01 4.84068215e-01 5.84901273e-01
8.54110658e-01 5.99369645e-01 6.17845714e-01 -4.95199766e-03
1.74278224e+00 -3.85179818e-01 -9.57660496e-01 -6.11828826e-02
6.74570441e-01 -1.46656811e-01 3.12426805e-01 3.09022367e-01
-1.66889206e-01 3.47961187e-02 -1.35606062e+00 9.02567059e-02
-1.54096079e+00 -5.19994557e-01 5.86881638e-01 5.30051351e-01
-4.52727228e-01 3.00938070e-01 -4.16188598e-01 -1.73221752e-01
5.40359735e-01 -2.92806271e-02 -6.51798248e-01 -2.49979362e-01
-1.72764969e+00 9.99962509e-01 1.31049287e+00 -2.82378625e-02
-8.71527970e-01 -9.79327977e-01 -1.11347067e+00 1.85450464e-01
1.09797013e+00 -2.87889242e-01 6.16179824e-01 5.04807644e-02
-5.53102851e-01 3.19945663e-01 5.12777269e-01 -6.69754863e-01
-5.91819525e-01 -4.86431539e-01 -1.09765732e+00 4.28578317e-01
5.62018417e-02 -9.05825943e-02 6.27845228e-01 -1.21863770e+00
-5.08318603e-01 -8.44241202e-01 3.74945849e-01 -1.66475158e-02
-8.47915530e-01 5.45740366e-01 -9.05849263e-02 -5.61407685e-01
-3.54480684e-01 -5.46240330e-01 -1.61448196e-01 -7.46349394e-01
-5.08624256e-01 -4.06762093e-01 1.15503573e+00 -1.42835355e+00
1.85416603e+00 -1.83636844e+00 -3.83641645e-02 4.95256364e-01
7.50236094e-01 5.51808059e-01 -2.97591705e-02 8.88202190e-01
-4.26459968e-01 5.84984481e-01 -1.19537465e-01 2.69199491e-01
-2.06052642e-02 -1.43800825e-01 -8.48407984e-01 -1.88830540e-01
4.57260877e-01 1.01667690e+00 -9.25385833e-01 -3.09778512e-01
-1.30839184e-01 2.19470382e-01 -1.82981148e-01 2.71327555e-01
-3.29327196e-01 -1.60007507e-01 -9.04325783e-01 9.70817506e-01
5.39218426e-01 -2.16962546e-01 1.22111365e-01 -3.85177672e-01
-1.24010794e-01 1.30167350e-01 -5.72651565e-01 1.43066394e+00
-4.30957288e-01 1.55179754e-01 -1.70634046e-01 -8.07150304e-01
8.59625101e-01 5.41488051e-01 2.98264295e-01 -4.82418329e-01
3.34011436e-01 -1.04544282e-01 -5.57879031e-01 -5.14719546e-01
6.87238753e-01 1.26883686e-01 -6.41966283e-01 4.88966435e-01
3.89882714e-01 4.42511469e-01 7.32040554e-02 8.56998622e-01
1.46406698e+00 -4.48019266e-01 4.53627735e-01 3.05500597e-01
6.15318835e-01 3.24259192e-01 6.32261693e-01 1.90056384e-01
2.51005944e-02 -2.25769132e-01 5.87212026e-01 -8.72718990e-01
-5.68852663e-01 -8.24802577e-01 2.97529280e-01 4.31140453e-01
-7.70180300e-02 -9.01598871e-01 -8.25959921e-01 -1.32748222e+00
-1.55172080e-01 8.68019342e-01 -6.26102448e-01 -3.93385291e-01
-1.25764668e-01 -5.76643527e-01 5.78800857e-01 4.84674513e-01
3.33085358e-01 -1.15048110e+00 -4.79398727e-01 3.16693187e-01
-3.51367235e-01 -1.42768312e+00 4.47941618e-03 -3.75162438e-02
-1.44233152e-01 -1.74618602e+00 2.56886601e-01 -2.46254131e-01
6.02560103e-01 2.32249454e-01 8.54069650e-01 -9.28392559e-02
-5.04883111e-01 6.14827812e-01 -9.20005322e-01 -7.92299449e-01
-1.66276500e-01 -5.79392910e-02 2.74814218e-01 1.08533539e-01
8.45682859e-01 -2.36857384e-01 -6.33717477e-02 -3.68003368e-01
-1.47177351e+00 -4.20068800e-01 4.05294597e-01 2.34702542e-01
2.34615237e-01 5.97857177e-01 8.38207960e-01 -6.33540332e-01
1.18219781e+00 -1.07693183e+00 -8.47468674e-01 7.40314603e-01
-5.75766087e-01 7.79522508e-02 6.16035223e-01 -1.98773339e-01
-1.06882572e+00 -5.49692392e-01 1.73999056e-01 -5.42065263e-01
-9.36011747e-02 1.40731668e+00 -4.31214750e-01 -1.67474926e-01
5.84584057e-01 -9.09158885e-02 -4.39713299e-01 -3.92622232e-01
5.08976638e-01 9.19955611e-01 5.02174914e-01 -3.83545339e-01
1.23586977e+00 1.29206225e-01 -1.21182024e-01 -5.69786489e-01
-1.16435838e+00 -4.85736787e-01 -4.33215767e-01 -2.29582667e-01
1.16027474e+00 -6.36694968e-01 -7.53919125e-01 1.36090785e-01
-1.39707112e+00 6.96710765e-01 -1.49348840e-01 3.54411691e-01
1.88026205e-01 4.90104258e-01 -5.63924611e-01 -8.72267663e-01
-9.41021144e-01 -6.17954254e-01 5.89183152e-01 1.71783105e-01
6.86796010e-02 -9.45416093e-01 3.26694369e-01 5.29004097e-01
1.81763217e-01 5.68400145e-01 1.18039501e+00 -1.48044193e+00
-7.37172604e-01 -8.08074236e-01 -4.81374681e-01 2.51975834e-01
1.14370726e-01 -6.19049370e-01 -6.55265927e-01 -2.32358892e-02
-2.96741482e-02 -2.96289742e-01 4.19445008e-01 -6.18929207e-01
9.02562737e-01 -7.55640447e-01 -2.82829881e-01 3.52653682e-01
1.52457094e+00 4.85252500e-01 7.65641272e-01 6.02229774e-01
8.60633492e-01 7.40632951e-01 6.19030774e-01 5.08879602e-01
5.78819513e-01 9.20445547e-02 6.05088234e-01 5.89213014e-01
4.28916633e-01 -6.65538788e-01 1.60370409e-01 8.82069290e-01
6.46475470e-03 -1.05017655e-01 -1.16539145e+00 6.37004197e-01
-1.44743943e+00 -9.85169947e-01 2.55209506e-01 1.86648667e+00
5.79017460e-01 1.07724527e-02 -6.22985840e-01 2.55293369e-01
4.73715663e-01 1.77910551e-01 -2.73281634e-01 -1.91077381e-01
2.18820781e-01 2.97019213e-01 3.80921870e-01 1.24272294e-01
-1.13705575e+00 9.82054651e-01 4.45159197e+00 9.26201880e-01
-6.70972824e-01 2.26850376e-01 1.59720138e-01 4.07726496e-01
-3.65860283e-01 2.88995892e-01 -9.97411609e-01 2.03028709e-01
1.44747746e+00 -6.01674974e-01 3.08512390e-01 9.99144495e-01
-5.01955092e-01 4.26143199e-01 -5.26224196e-01 8.52825046e-01
4.13665563e-01 -1.52714586e+00 4.50971007e-01 3.28547984e-01
3.25698793e-01 -1.86732799e-01 -2.43275017e-01 5.34640551e-01
5.95014870e-01 -5.74132144e-01 2.10394487e-01 8.27680111e-01
5.23079455e-01 -9.70009029e-01 8.77000034e-01 2.83353508e-01
-1.33236706e+00 -6.79090142e-01 -3.00346047e-01 5.39855421e-01
1.60589099e-01 6.41649485e-01 -1.05111468e+00 1.50644410e+00
6.80854678e-01 5.70053160e-01 -7.09305465e-01 8.13012183e-01
-4.87350404e-01 3.87449563e-01 -1.02710612e-01 1.16335623e-01
2.69364655e-01 -2.04074576e-01 5.41793823e-01 8.86702120e-01
4.60866064e-01 5.40079534e-01 2.20634848e-01 7.56825686e-01
-3.51790935e-01 8.49390998e-02 -1.25343931e+00 -9.23579276e-01
5.59857786e-01 1.51815295e+00 -3.32777679e-01 -3.33208948e-01
-5.37038743e-01 4.62803841e-01 5.53388834e-01 3.39443177e-01
-6.05219543e-01 -9.18656647e-01 5.85458159e-01 -1.31151408e-01
1.80741221e-01 -2.61954129e-01 1.37577921e-01 -1.51896775e+00
1.17626593e-01 -8.60476971e-01 1.14843023e+00 -8.82498920e-01
-1.35576558e+00 1.19105554e+00 1.90867081e-01 -1.35142660e+00
-2.08293274e-01 -7.22575009e-01 -4.96394217e-01 6.59865022e-01
-1.48423529e+00 -1.41334248e+00 -2.11108923e-01 6.62533402e-01
5.39683588e-02 -2.89403409e-01 1.01680577e+00 1.58592597e-01
-8.03805292e-01 9.77470130e-02 -5.83742023e-01 6.82514012e-01
1.35667786e-01 -7.40667522e-01 6.01138771e-01 1.26096809e+00
2.86223292e-01 9.78628039e-01 -7.07827285e-02 -1.44051325e+00
-1.70826149e+00 -1.41949630e+00 1.06654012e+00 -7.95258284e-01
1.29644299e+00 -2.78446913e-01 -1.07621455e+00 7.55814552e-01
3.95337284e-01 -3.06447186e-02 1.16837907e+00 6.54300526e-02
-9.06228781e-01 1.01234436e-01 -1.01563215e+00 4.39507335e-01
7.59094000e-01 -8.93786311e-01 -1.12260425e+00 1.71526402e-01
1.54961646e+00 2.49268755e-01 -1.24380422e+00 2.85979778e-01
-6.41652271e-02 -1.17544383e-01 1.01465297e+00 -1.29372716e+00
-6.70152754e-02 -2.52825677e-01 -7.90439993e-02 -1.30989051e+00
-3.86924483e-02 -4.22431588e-01 -4.96136874e-01 1.09311855e+00
4.05945122e-01 -8.47248971e-01 2.52776951e-01 9.40430939e-01
-1.83306068e-01 -4.26236689e-01 -8.84390414e-01 -6.21981263e-01
-4.16028738e-01 -8.82999599e-01 8.93491566e-01 1.37535882e+00
7.48915076e-01 1.24784678e-01 -5.79447336e-02 7.46905565e-01
8.26452851e-01 -9.04695615e-02 1.97903842e-01 -1.66120327e+00
4.00756717e-01 3.11744273e-01 -6.85255289e-01 6.77566975e-02
3.15887600e-01 -9.84846830e-01 -8.09653461e-01 -1.48224235e+00
-8.66976902e-02 -1.72482520e-01 -7.06445217e-01 7.96304107e-01
-1.90495655e-01 -2.47046798e-01 4.06934738e-01 -7.75366426e-02
-8.24146807e-01 4.82923120e-01 5.59274673e-01 -3.22883010e-01
-1.91783300e-03 -6.21736526e-01 -1.11268473e+00 6.21728420e-01
7.23761201e-01 -6.06836319e-01 -3.76267284e-01 -9.45008993e-02
7.22354293e-01 3.10512275e-01 5.36905289e-01 -1.21590567e+00
5.52383006e-01 -1.11886360e-01 3.36603001e-02 -6.09185398e-01
2.39407465e-01 -9.89868164e-01 4.17216793e-02 6.83384836e-02
-3.62900086e-02 2.54784107e-01 3.49337816e-01 7.66258895e-01
-5.23072600e-01 -2.59851635e-01 -1.52584285e-01 7.70311803e-02
-8.30476463e-01 6.67506337e-01 -1.85328573e-01 1.33416466e-02
1.04750240e+00 2.51503050e-01 -8.50848258e-01 -1.63501397e-01
-7.33832419e-01 7.15290233e-02 -1.45898446e-01 7.72261024e-01
1.12522674e+00 -1.45719731e+00 -2.48499870e-01 2.57256150e-01
5.86183310e-01 -2.77857602e-01 3.91004443e-01 3.99044752e-01
-1.35675028e-01 7.66128838e-01 -1.79371446e-01 6.42983317e-01
-8.55503321e-01 1.30466270e+00 -4.43315983e-01 -9.64233458e-01
-5.28774500e-01 3.40188473e-01 -2.71635830e-01 -4.82614934e-01
1.44727394e-01 -1.51344255e-01 -8.29915822e-01 1.98623493e-01
1.08459663e+00 1.78294376e-01 3.37601036e-01 -6.28425777e-01
-4.37962800e-01 5.24247587e-02 -3.88626128e-01 -3.26329954e-02
1.46806860e+00 3.48435551e-01 -3.73077035e-01 -3.97566557e-01
1.08142996e+00 -1.33114845e-01 -2.38694385e-01 -2.85331041e-01
5.24086773e-01 -4.27157044e-01 5.75121352e-03 -9.81422484e-01
-7.99615383e-01 7.69432783e-01 9.30963159e-02 4.65192348e-01
1.10846901e+00 -1.27405077e-01 9.93474424e-01 9.04380918e-01
8.05870473e-01 -9.44656610e-01 -8.49060789e-02 6.71676993e-01
1.10944867e+00 -8.13892484e-01 -2.93597639e-01 -5.62743902e-01
-7.58272052e-01 9.64817941e-01 5.37162244e-01 7.88022205e-02
9.48662043e-01 1.96856126e-01 -8.64427388e-02 -6.96052670e-01
-7.46553004e-01 -1.41644031e-01 4.24096525e-01 7.74117649e-01
-3.19530845e-01 -5.50457872e-02 -3.03632189e-02 1.28909361e+00
6.35524169e-02 -1.17445409e-01 3.04295421e-01 1.17968464e+00
-3.54866445e-01 -1.10794747e+00 -5.79661548e-01 3.76083702e-01
-6.31196856e-01 -4.01801139e-01 -7.16700315e-01 5.47277749e-01
-1.31574824e-01 1.16839826e+00 -4.32300448e-01 -9.68678415e-01
3.16070527e-01 3.16355646e-01 -3.76965523e-01 -5.96080601e-01
-7.47599006e-01 -7.25980699e-01 3.54387403e-01 -6.38873279e-01
1.36674538e-01 -8.91037732e-02 -1.15749681e+00 5.36054075e-02
-3.63334030e-01 7.26383209e-01 7.74486184e-01 8.00605118e-01
8.44477713e-01 5.33878922e-01 3.43717188e-01 -2.07177356e-01
-3.64486486e-01 -9.20400262e-01 -2.70044595e-01 3.63515168e-01
-1.43550292e-01 -6.67881668e-01 -2.40046814e-01 -1.28126234e-01] | [6.628945350646973, 7.508791446685791] |
3e7d6330-1797-43f3-8655-b7a7cd1b3e47 | generative-plug-and-play-posterior-sampling | 2306.07233 | null | https://arxiv.org/abs/2306.07233v1 | https://arxiv.org/pdf/2306.07233v1.pdf | Generative Plug and Play: Posterior Sampling for Inverse Problems | Over the past decade, Plug-and-Play (PnP) has become a popular method for reconstructing images using a modular framework consisting of a forward and prior model. The great strength of PnP is that an image denoiser can be used as a prior model while the forward model can be implemented using more traditional physics-based approaches. However, a limitation of PnP is that it reconstructs only a single deterministic image. In this paper, we introduce Generative Plug-and-Play (GPnP), a generalization of PnP to sample from the posterior distribution. As with PnP, GPnP has a modular framework using a physics-based forward model and an image denoising prior model. However, in GPnP these models are extended to become proximal generators, which sample from associated distributions. GPnP applies these proximal generators in alternation to produce samples from the posterior. We present experimental simulations using the well-known BM3D denoiser. Our results demonstrate that the GPnP method is robust, easy to implement, and produces intuitively reasonable samples from the posterior for sparse interpolation and tomographic reconstruction. Code to accompany this paper is available at https://github.com/gbuzzard/generative-pnp-allerton . | ['Gregery T. Buzzard', 'Charles A. Bouman'] | 2023-06-12 | null | null | null | null | ['image-denoising'] | ['computer-vision'] | [ 1.94964215e-01 -1.38141245e-01 2.45454088e-01 -9.24118757e-02
-7.12286711e-01 -2.55441368e-01 7.38166392e-01 -4.33994472e-01
6.30357414e-02 8.80249143e-01 1.07037522e-01 -1.09757625e-01
-1.21003643e-01 -9.83144164e-01 -8.63592505e-01 -1.05689204e+00
1.95731729e-01 5.68225205e-01 4.44096118e-01 -1.69561088e-01
2.73046847e-02 4.84907418e-01 -1.43230820e+00 2.11719871e-01
8.83308411e-01 6.93530321e-01 6.25373483e-01 7.44508445e-01
-4.43165526e-02 6.71803594e-01 -3.07368398e-01 -8.02283138e-02
2.28958815e-01 -7.79872656e-01 -3.59190345e-01 -8.23907778e-02
-5.01951538e-02 -4.41943109e-01 -3.90596986e-01 9.80352283e-01
6.47520185e-01 -2.45340168e-02 7.85637915e-01 -1.08221388e+00
-5.64544082e-01 1.32176906e-01 -5.84880829e-01 1.19215250e-03
4.44426358e-01 2.58034974e-01 2.48851299e-01 -8.07721138e-01
7.14137852e-01 1.36791229e+00 9.71302927e-01 4.58639801e-01
-1.76937318e+00 -4.21171129e-01 -5.01101613e-01 -9.40983649e-03
-1.37854898e+00 -4.82625246e-01 7.79956937e-01 -3.87647152e-01
5.06647348e-01 4.89358157e-01 9.00194943e-01 1.02761292e+00
5.77133060e-01 7.42392659e-01 1.39648509e+00 -6.70771003e-01
3.91755253e-01 -9.99200270e-02 -2.46261269e-01 3.59652758e-01
-8.43475983e-02 4.04213965e-01 -5.44994712e-01 -5.29418409e-01
1.30325782e+00 -1.10791922e-01 -5.44626772e-01 -4.48811591e-01
-8.74140382e-01 6.89965129e-01 3.26442838e-01 -2.46908516e-02
-7.28748441e-01 5.10305166e-01 -3.87394913e-02 5.56001533e-03
5.82596421e-01 -2.18627244e-01 1.73665076e-01 -1.71424001e-01
-1.35135996e+00 5.64552307e-01 9.17994440e-01 7.99124181e-01
7.66815782e-01 1.31772414e-01 -1.29016131e-01 8.90061319e-01
7.10242927e-01 7.89210916e-01 1.90768450e-01 -1.58086371e+00
-3.09894502e-01 -3.85209024e-01 2.61465460e-01 -7.49192953e-01
1.31873131e-01 -3.07929337e-01 -9.66070831e-01 7.44389653e-01
2.29141742e-01 1.97748572e-01 -1.07995832e+00 1.56451237e+00
4.76579070e-01 3.33554864e-01 -2.22992912e-01 7.13464320e-01
7.04850197e-01 1.09606814e+00 -2.30823755e-01 -1.71487987e-01
1.12870884e+00 -6.84954286e-01 -6.41045451e-01 8.15754384e-02
-2.34845236e-01 -9.71495688e-01 7.63258100e-01 6.49906039e-01
-1.54241681e+00 -4.66053903e-01 -8.28159809e-01 -1.36141777e-01
3.55848968e-01 -2.93881416e-01 1.06484361e-01 4.94724095e-01
-1.53314710e+00 9.02461946e-01 -1.17786252e+00 -3.68406653e-01
3.59956622e-01 -3.39424014e-02 -6.76303171e-03 -2.96013415e-01
-8.31542969e-01 8.63462269e-01 -3.12861390e-02 3.44055258e-02
-1.06434166e+00 -8.79560232e-01 -5.99157929e-01 -1.77909106e-01
-2.06890598e-01 -1.21704519e+00 1.32725000e+00 -7.81008363e-01
-1.95560825e+00 6.39707088e-01 -5.46377540e-01 -4.01108712e-01
7.49559224e-01 1.27875298e-01 -1.07795991e-01 3.61911416e-01
1.85813785e-01 6.47221923e-01 1.03422952e+00 -1.72530770e+00
-1.52623743e-01 2.13737175e-01 -3.91048402e-01 -1.51715688e-02
5.03181517e-01 -3.48298848e-02 -5.34215450e-01 -7.16045737e-01
2.28315383e-01 -8.26029122e-01 -2.91463763e-01 4.03356314e-01
-3.21968466e-01 2.89399683e-01 7.39727974e-01 -8.67820382e-01
7.95591712e-01 -2.23679399e+00 1.68717325e-01 2.01721400e-01
2.05348670e-01 -7.99841434e-02 1.31482407e-01 8.93727839e-01
-3.51862051e-02 -1.49632156e-01 -7.24739313e-01 -5.06760478e-01
-5.88445030e-02 6.79146051e-01 -4.32373464e-01 5.00736654e-01
-1.48994923e-01 6.66874111e-01 -8.25244546e-01 -2.63364911e-01
4.19983506e-01 9.79862094e-01 -6.25702143e-01 -2.09034439e-02
-1.91761389e-01 8.97199810e-01 -2.48685002e-01 3.34001809e-01
1.17161870e+00 -1.79833218e-01 4.61689867e-02 -1.71580791e-01
-1.86529577e-01 9.05354768e-02 -1.20439935e+00 1.63238871e+00
-3.40609252e-01 3.39859515e-01 6.58997178e-01 -8.36687565e-01
6.85385704e-01 4.50391948e-01 3.69161010e-01 -3.80602837e-01
-1.83016449e-01 5.02743840e-01 -2.30420902e-01 -2.34593034e-01
4.11144197e-01 -6.24501526e-01 4.56330687e-01 4.41205859e-01
3.42601212e-03 -6.83224380e-01 1.12008557e-01 4.05849546e-01
1.02776790e+00 5.85556924e-01 2.95832068e-01 -5.95174968e-01
2.66192555e-01 9.25944448e-02 5.93859076e-01 8.83998156e-01
8.51185322e-02 1.23659658e+00 2.72553623e-01 -8.52953419e-02
-1.08282042e+00 -1.46739352e+00 -4.63262439e-01 2.65585065e-01
9.39102843e-02 -3.79290789e-01 -6.99897587e-01 -1.47063313e-02
-1.40542522e-01 9.08223987e-01 -3.35783929e-01 3.10458958e-01
-5.17775118e-01 -8.68949771e-01 2.67662823e-01 1.48534238e-01
5.72511196e-01 -9.82745171e-01 -5.31269789e-01 3.65988851e-01
-3.80447596e-01 -6.75627947e-01 -1.12963930e-01 4.52752709e-02
-1.02265358e+00 -7.19974279e-01 -1.11679208e+00 -3.89308363e-01
6.55516624e-01 8.92708153e-02 1.23952842e+00 6.80706650e-02
-8.47794265e-02 6.91823363e-01 -1.10169463e-01 -2.37223789e-01
-1.04191947e+00 -6.07534468e-01 -1.88493788e-01 1.72417201e-02
-3.40144336e-01 -1.24136305e+00 -8.56665015e-01 2.17147529e-01
-1.12320530e+00 4.27613109e-01 2.82683671e-01 7.21965790e-01
9.80166912e-01 1.38013527e-01 7.63142928e-02 -6.78965092e-01
5.29569745e-01 -6.61398888e-01 -5.03974795e-01 -2.11819112e-01
-3.73426229e-01 -1.00474589e-01 4.77307558e-01 -2.18097910e-01
-1.21468234e+00 -4.93855849e-02 -5.68738520e-01 -5.63238859e-01
-1.19513944e-01 4.44209576e-01 -4.45788465e-02 -6.20947964e-02
6.43475175e-01 5.22445142e-01 3.88843417e-01 -7.28244901e-01
3.16945672e-01 3.02357197e-01 6.97204232e-01 -8.38684201e-01
7.70724297e-01 9.11930442e-01 2.56214708e-01 -9.30530906e-01
-2.73358643e-01 -9.77966711e-02 -7.23516047e-02 -3.80582452e-01
6.59011364e-01 -7.81825960e-01 -2.94698536e-01 7.46521831e-01
-1.15397048e+00 -6.40737236e-01 -7.62617707e-01 3.39065462e-01
-8.43631268e-01 5.03804266e-01 -7.33074665e-01 -8.58412206e-01
-1.27829522e-01 -1.20393741e+00 1.01959705e+00 2.66955376e-01
-1.11450605e-01 -1.00767958e+00 1.99281722e-01 -8.53000283e-02
8.04917514e-01 3.47137332e-01 5.83566546e-01 1.72377765e-01
-7.36850858e-01 -1.48974573e-02 -2.27549449e-02 4.64183450e-01
-1.55199859e-02 1.89121664e-01 -9.47472811e-01 -3.21238101e-01
6.95619524e-01 8.10097307e-02 7.96719074e-01 1.07587636e+00
9.57512498e-01 -1.55552238e-01 -3.34220916e-01 7.97207236e-01
1.74534810e+00 3.52919251e-02 1.25671232e+00 9.41379443e-02
3.60788077e-01 2.21221596e-01 -1.07080504e-01 3.67437541e-01
2.80681163e-01 7.50868261e-01 2.95133352e-01 1.30244076e-01
-6.16642296e-01 -2.61752933e-01 3.08165044e-01 8.79086673e-01
-3.53558362e-01 -2.32896864e-01 -8.18152070e-01 3.53998870e-01
-1.76833391e+00 -1.06377125e+00 -5.92296958e-01 2.09049058e+00
9.45999026e-01 -2.32705355e-01 -8.18402618e-02 4.09383737e-02
6.02686226e-01 -3.84843796e-02 -1.78975224e-01 -1.32539541e-01
-1.05942629e-01 4.87708509e-01 3.45793545e-01 8.43765616e-01
-6.54213071e-01 2.67326713e-01 6.54472923e+00 1.26069915e+00
-8.89287531e-01 5.34457684e-01 4.81675476e-01 5.53177372e-02
-6.49906516e-01 3.40056807e-01 -5.22078395e-01 8.45900238e-01
9.30288374e-01 -9.33029428e-02 5.33331215e-01 4.55064952e-01
5.21955967e-01 -5.60497284e-01 -7.06243098e-01 9.12124693e-01
-2.41101831e-01 -1.46563816e+00 -3.88967022e-02 2.80445993e-01
7.09820449e-01 1.79495066e-01 -2.00942531e-01 -2.26459667e-01
5.40700793e-01 -8.52604091e-01 1.04863119e+00 1.04390848e+00
7.77222991e-01 -4.13057685e-01 2.27951422e-01 6.26058817e-01
-9.05112207e-01 4.68138397e-01 -4.38038886e-01 3.76103036e-02
8.22261453e-01 1.21025121e+00 -1.89898223e-01 8.08816671e-01
9.20746624e-01 6.22890592e-01 2.07453966e-01 1.32501376e+00
-4.57248032e-01 8.39542747e-01 -6.56581640e-01 6.17099643e-01
-2.55316734e-01 -6.03040397e-01 9.74241912e-01 8.71344686e-01
6.82123661e-01 1.81588948e-01 1.16869695e-01 1.27103972e+00
2.47886553e-01 -4.58824158e-01 -2.96473294e-01 4.95536059e-01
2.91882932e-01 1.07055783e+00 -8.12561095e-01 -3.59564364e-01
-3.92222181e-02 1.11422718e+00 -3.18269789e-01 5.52841425e-01
-8.87430668e-01 -7.77544379e-02 2.48045951e-01 6.86459839e-01
5.21423161e-01 -1.80728316e-01 -1.70492336e-01 -1.03796709e+00
-6.13687001e-02 -7.70359874e-01 -1.23816229e-01 -1.24402499e+00
-1.47083831e+00 4.38105792e-01 3.98952186e-01 -1.19610405e+00
-1.46844581e-01 -2.59672463e-01 -7.12511301e-01 1.33106649e+00
-1.47695124e+00 -1.05833685e+00 -3.66354108e-01 5.00651240e-01
3.32131118e-01 5.04108548e-01 8.51765633e-01 2.21859500e-01
-1.50559038e-01 -1.43947080e-01 6.42182887e-01 -4.11770940e-01
3.96310359e-01 -1.14780438e+00 3.51109415e-01 9.56851602e-01
-1.63109168e-01 5.48699677e-01 1.14260077e+00 -8.50048423e-01
-1.27107620e+00 -6.74673080e-01 4.52424943e-01 -9.47849602e-02
3.53083432e-01 -1.16986975e-01 -9.82986510e-01 5.14432609e-01
5.92758000e-01 1.18458360e-01 2.70318031e-01 -5.28518796e-01
1.25661120e-01 2.31006853e-02 -1.46506763e+00 4.44853306e-01
8.17652702e-01 -2.00180963e-01 -3.34184915e-01 3.06521654e-01
1.11835450e-01 -7.40361929e-01 -8.46165061e-01 2.19518006e-01
4.69735533e-01 -1.34539282e+00 1.16119146e+00 6.51472569e-01
5.98592699e-01 -6.77821279e-01 -1.36912465e-01 -1.41050386e+00
-3.40894520e-01 -9.21453118e-01 -1.63511187e-01 1.17336047e+00
-9.44585726e-02 -1.07037210e+00 3.48032355e-01 3.80927444e-01
-2.40003139e-01 -5.85751414e-01 -1.18834043e+00 -9.50676918e-01
2.01769933e-01 -4.77882534e-01 3.34822953e-01 4.86373603e-01
-5.82767427e-01 -1.77991718e-01 -4.11924809e-01 6.04792014e-02
1.11981249e+00 -2.11158823e-02 6.71177924e-01 -1.05528355e+00
-9.66045737e-01 -2.11870700e-01 -5.31627499e-02 -1.40225339e+00
-3.93947601e-01 -8.66662323e-01 3.52427244e-01 -1.96498799e+00
2.78890312e-01 -6.24311626e-01 2.81854451e-01 9.10878778e-02
1.83017343e-01 4.79697526e-01 9.56581682e-02 6.61618888e-01
3.17737073e-01 6.38781667e-01 1.35081041e+00 2.57194996e-01
1.30312862e-02 5.39941192e-02 -3.84959340e-01 7.62909114e-01
6.74450219e-01 -6.79908276e-01 -4.70748693e-01 -3.24299335e-01
1.42594635e-01 2.87843645e-01 9.13483620e-01 -1.03907204e+00
1.88670307e-01 9.31083038e-02 4.60272670e-01 -5.95457554e-01
6.17309570e-01 -5.77079654e-01 1.08721900e+00 4.80674386e-01
3.60102952e-01 -3.19815248e-01 2.75157183e-01 5.96767783e-01
-5.76220527e-02 -4.74262446e-01 9.70211029e-01 -4.12066102e-01
-2.73672968e-01 -3.46087813e-02 -6.24860168e-01 -1.34755954e-01
6.23417974e-01 -3.77227664e-01 -1.94897607e-01 -6.41519785e-01
-9.16159153e-01 -2.27521747e-01 8.78110051e-01 -4.48233426e-01
5.68992615e-01 -1.30250311e+00 -8.99662554e-01 2.38677830e-01
-6.00820899e-01 2.04355299e-01 5.05191326e-01 1.14878833e+00
-1.02266979e+00 -2.22930238e-01 6.28480464e-02 -1.00697064e+00
-8.35655808e-01 2.64523953e-01 4.93017346e-01 -1.46129042e-01
-1.04451776e+00 7.08791018e-01 3.23274404e-01 -4.68150079e-01
-4.19695526e-01 -1.36237189e-01 4.96511936e-01 -4.70859200e-01
5.15074790e-01 3.69824052e-01 -1.43098354e-01 -4.43515629e-01
-1.56657770e-01 4.98421699e-01 5.14923930e-01 -6.11031532e-01
1.54789734e+00 -2.20953420e-01 -5.55688262e-01 2.92160690e-01
6.49703622e-01 2.36116111e-01 -1.53784192e+00 1.15213111e-01
-6.42962217e-01 -5.30425549e-01 2.54291296e-01 -7.81359673e-01
-1.05278432e+00 6.27433717e-01 3.85580897e-01 3.16532433e-01
1.30965602e+00 6.12343336e-03 7.39084482e-01 -6.00998700e-01
4.96096581e-01 -4.91543174e-01 -4.80915308e-01 2.29608119e-01
1.22751033e+00 -5.18792748e-01 2.47248799e-01 -6.09970152e-01
-2.24029854e-01 9.68530536e-01 -1.19161054e-01 -4.64938909e-01
1.12632370e+00 5.64128339e-01 -1.98368669e-01 -1.69547141e-01
-3.20181608e-01 2.35659629e-01 -1.98339764e-03 7.23773301e-01
2.34760657e-01 -8.18930119e-02 -3.61929893e-01 2.38893345e-01
-2.96980906e-02 4.82541531e-01 6.89757168e-01 1.20174229e+00
-2.15626895e-01 -1.57710373e+00 -7.98217177e-01 2.67512381e-01
-2.28371590e-01 -3.18440020e-01 4.35144782e-01 4.75863874e-01
-1.57289430e-02 7.84680188e-01 -1.79625619e-02 1.50295362e-01
-1.04949595e-02 2.09286325e-02 8.09586644e-01 -4.58263487e-01
-2.82441795e-01 5.45576930e-01 -2.58856714e-02 -6.75407946e-01
-7.06757128e-01 -7.38639176e-01 -1.05716228e+00 -6.69341147e-01
-4.88986149e-02 1.64669305e-01 5.87465882e-01 6.74167156e-01
5.76266646e-01 4.97944325e-01 2.21344158e-01 -1.41952658e+00
-3.56848150e-01 -8.37784171e-01 -7.45337725e-01 -4.14622203e-02
1.78060472e-01 -4.97456670e-01 -3.88975650e-01 2.12339610e-01] | [11.760435104370117, -2.3961968421936035] |
3aedbf06-41d3-4e70-984e-a399c9c27a9b | big-gans-are-watching-you-towards | 2006.04988 | null | https://arxiv.org/abs/2006.04988v2 | https://arxiv.org/pdf/2006.04988v2.pdf | Object Segmentation Without Labels with Large-Scale Generative Models | The recent rise of unsupervised and self-supervised learning has dramatically reduced the dependency on labeled data, providing effective image representations for transfer to downstream vision tasks. Furthermore, recent works employed these representations in a fully unsupervised setup for image classification, reducing the need for human labels on the fine-tuning stage as well. This work demonstrates that large-scale unsupervised models can also perform a more challenging object segmentation task, requiring neither pixel-level nor image-level labeling. Namely, we show that recent unsupervised GANs allow to differentiate between foreground/background pixels, providing high-quality saliency masks. By extensive comparison on standard benchmarks, we outperform existing unsupervised alternatives for object segmentation, achieving new state-of-the-art. | ['Stanislav Morozov', 'Artem Babenko', 'Andrey Voynov'] | 2020-06-08 | null | null | null | null | ['unsupervised-object-segmentation'] | ['computer-vision'] | [ 9.31315541e-01 4.19765055e-01 -4.69250143e-01 -4.67868865e-01
-7.55620003e-01 -5.42916834e-01 7.05054283e-01 9.94483382e-02
-5.28184235e-01 7.20210016e-01 -8.56356919e-02 1.33316382e-03
2.66308159e-01 -6.61914706e-01 -8.45109522e-01 -8.69788289e-01
2.80445129e-01 4.54564989e-01 6.64180160e-01 4.74231094e-02
2.29352877e-01 4.59665090e-01 -1.88367820e+00 2.41438359e-01
8.71321857e-01 1.02266943e+00 4.73650515e-01 6.65904343e-01
-1.26067430e-01 6.63420856e-01 -7.11428761e-01 -1.47602975e-01
3.09314519e-01 -4.61239934e-01 -8.59378159e-01 5.38548768e-01
8.54815364e-01 -6.71314746e-02 6.25069439e-02 1.12395036e+00
2.21515894e-01 6.99438527e-02 8.29616308e-01 -1.03537393e+00
-7.04480290e-01 5.49722791e-01 -6.05455935e-01 3.05236518e-01
-2.40660965e-01 4.36791003e-01 1.10943842e+00 -6.14853442e-01
6.36968493e-01 9.24999774e-01 4.00591254e-01 5.95091939e-01
-1.61577415e+00 -2.27396280e-01 2.11476773e-01 1.20133623e-01
-7.73911715e-01 -2.97099471e-01 8.62809420e-01 -3.98859859e-01
8.17601144e-01 2.02038825e-01 6.30010366e-01 1.15786791e+00
-2.91129798e-01 1.02186489e+00 1.45927274e+00 -6.66153848e-01
4.05960232e-01 6.55082464e-02 1.33699715e-01 6.43367469e-01
3.13070625e-01 1.38361216e-01 -4.04237151e-01 4.54266071e-01
1.03308988e+00 -1.59993730e-02 1.84691846e-02 -5.63373089e-01
-1.24983597e+00 8.71295094e-01 7.60644376e-01 4.38960671e-01
-3.34702104e-01 4.00070131e-01 1.99936807e-01 -8.78755450e-02
4.54192162e-01 7.16755152e-01 -5.43017268e-01 1.02738224e-01
-1.37330878e+00 -3.25108111e-01 4.33279604e-01 9.81013536e-01
1.17803276e+00 4.27150846e-01 -4.95213717e-01 6.58549488e-01
1.28108323e-01 3.90225887e-01 4.68524039e-01 -1.13930893e+00
2.36806665e-02 7.39749789e-01 -7.94334114e-02 -6.39374495e-01
-2.89311171e-01 -7.31787562e-01 -8.26370180e-01 6.05121911e-01
5.11135280e-01 7.75540918e-02 -1.65241981e+00 1.62727487e+00
1.10657103e-01 2.23402634e-01 -5.94416773e-03 7.76118815e-01
6.52111948e-01 3.66784364e-01 3.07206422e-01 -8.27993378e-02
1.15061331e+00 -1.54902959e+00 -4.73287374e-01 -6.59643292e-01
3.67454797e-01 -5.90276659e-01 1.19318831e+00 4.42983389e-01
-1.15994370e+00 -7.01398969e-01 -8.77579808e-01 -1.30710810e-01
-5.94686806e-01 2.49356821e-01 1.03392565e+00 8.40705693e-01
-1.44824564e+00 5.19428909e-01 -7.53579795e-01 -2.87058830e-01
1.13739347e+00 5.11001050e-01 -1.21567197e-01 2.20956549e-01
-6.45190954e-01 7.62781978e-01 5.13605356e-01 -4.57379669e-02
-1.30440784e+00 -6.71140194e-01 -8.65380049e-01 3.25104268e-03
3.37478966e-01 -6.37389839e-01 1.12053359e+00 -1.55361640e+00
-1.64293313e+00 1.34176612e+00 -8.84792805e-02 -9.76265073e-01
4.55401093e-01 -1.22866139e-01 1.68214832e-02 4.13529962e-01
2.26388142e-01 1.37777925e+00 1.35843015e+00 -1.47746801e+00
-6.81888521e-01 -7.47040808e-02 1.57037199e-01 -5.86324185e-03
-8.95967633e-02 -1.72621518e-01 -3.27679187e-01 -8.33850145e-01
-9.91412103e-02 -7.02721715e-01 -5.30701697e-01 -1.12373181e-01
-4.28295195e-01 -1.45003498e-01 9.48125839e-01 -3.24193776e-01
5.38984954e-01 -2.07697177e+00 2.14285880e-01 -1.38378903e-01
2.21145302e-01 4.23214465e-01 -2.49806434e-01 -4.03109752e-02
-1.50878191e-01 1.47168934e-01 -7.75860369e-01 -6.01917803e-01
-1.49588481e-01 3.58840406e-01 -2.47673824e-01 3.12667370e-01
7.97195733e-01 1.40095699e+00 -9.15690541e-01 -6.56783760e-01
5.18212080e-01 1.90107644e-01 -4.42722261e-01 1.19582668e-01
-5.48255563e-01 8.37808847e-01 -3.49359252e-02 5.96327662e-01
4.18488324e-01 -4.82294828e-01 -1.33974537e-01 8.42345655e-02
1.19519271e-02 2.46803388e-01 -6.85887992e-01 2.00720263e+00
-3.59775901e-01 8.62891138e-01 -1.74411805e-03 -1.42067158e+00
7.33498037e-01 -1.26259863e-01 3.92195463e-01 -8.33917916e-01
1.68673381e-01 8.97725746e-02 -1.62942614e-02 -4.44833450e-02
2.46218130e-01 9.90143567e-02 2.17677578e-02 2.10418046e-01
6.10424697e-01 -3.52835774e-01 3.50051135e-01 1.43677533e-01
8.98838878e-01 3.25696409e-01 -7.72932023e-02 -5.50011277e-01
3.65417778e-01 3.08605075e-01 3.86537611e-01 9.44312811e-01
-1.40253961e-01 9.73593831e-01 3.36103320e-01 -1.10761933e-02
-9.08086777e-01 -1.18372393e+00 -1.14545517e-01 1.33681512e+00
3.16559106e-01 2.32207738e-02 -1.23689151e+00 -8.30724478e-01
-2.25634396e-01 6.23097897e-01 -7.87504077e-01 -8.76226723e-02
-5.86198509e-01 -6.41498268e-01 3.53025377e-01 6.75390363e-01
5.05755007e-01 -1.37690103e+00 -1.00507307e+00 3.29820476e-02
1.37710914e-01 -1.31868613e+00 -8.15331861e-02 7.29107857e-01
-1.24769843e+00 -8.52791786e-01 -9.32353318e-01 -1.11903930e+00
9.09641981e-01 2.74426818e-01 1.52825308e+00 1.71841696e-01
-6.05538189e-01 4.29413766e-01 -2.03835890e-01 -4.24940854e-01
-2.40588009e-01 2.98532993e-01 -4.05118674e-01 8.46945792e-02
-2.62677744e-02 -6.03816450e-01 -7.62581944e-01 2.81130195e-01
-1.06331658e+00 2.44752854e-01 7.47568369e-01 8.33379626e-01
7.77839124e-01 -2.50965714e-01 7.98336208e-01 -1.10176933e+00
-8.57892707e-02 5.43160252e-02 -6.79516554e-01 4.80667688e-02
-4.80778068e-01 2.06764296e-01 4.72116053e-01 -2.54555643e-01
-1.11477375e+00 3.59590203e-01 -7.42896423e-02 -2.16438860e-01
-6.76442981e-01 -2.06861962e-02 -5.81386350e-02 -1.88953266e-01
6.22016966e-01 2.74834514e-01 -6.19519763e-02 -4.22479689e-01
8.08460653e-01 2.78268129e-01 8.32779229e-01 -3.07924420e-01
9.07028615e-01 7.33974576e-01 -3.98520902e-02 -6.62505567e-01
-1.20161581e+00 -5.77892303e-01 -1.09386110e+00 7.32603669e-02
1.23743010e+00 -8.44326138e-01 -1.08150221e-01 4.69661385e-01
-8.88155222e-01 -8.61160815e-01 -9.53299403e-01 -1.45509308e-02
-8.54389906e-01 3.02860498e-01 -4.13793921e-01 -5.02341270e-01
-1.39765680e-01 -1.16544092e+00 1.26805460e+00 4.80069757e-01
3.91070358e-03 -1.06517649e+00 -3.07536244e-01 4.17637199e-01
4.43888515e-01 2.35200346e-01 7.38518715e-01 -4.80487078e-01
-8.86350214e-01 1.67316839e-01 -4.55846667e-01 5.59320211e-01
2.45602027e-01 -1.40889943e-01 -1.26106369e+00 -7.30263889e-02
-1.42366275e-01 -4.75455701e-01 1.35668266e+00 5.17162263e-01
1.21601474e+00 -3.13214511e-02 -2.27920845e-01 6.69615507e-01
1.30555689e+00 -3.03896785e-01 6.87967241e-01 2.88960814e-01
9.86921728e-01 5.28204322e-01 5.31585932e-01 -2.47342661e-01
-2.78296862e-02 4.78008121e-01 6.14839971e-01 -7.69097626e-01
-6.93839490e-01 -9.73611176e-02 8.52707997e-02 2.95037150e-01
-1.05959244e-01 1.60722267e-02 -7.81683505e-01 8.62224221e-01
-1.85444975e+00 -5.86369634e-01 -1.68699369e-01 1.97937727e+00
1.06129432e+00 4.00526136e-01 2.42658645e-01 9.86971855e-02
7.07164109e-01 2.28790835e-01 -6.66006505e-01 -2.54931331e-01
-2.92159528e-01 7.64049292e-01 7.95654953e-01 1.98836669e-01
-1.47871602e+00 1.52531981e+00 7.33906031e+00 9.35211182e-01
-1.20307279e+00 2.78094649e-01 9.48174477e-01 2.05879524e-01
-1.91094875e-01 2.00950447e-03 -5.81518888e-01 3.56138438e-01
7.67942965e-01 3.83218527e-01 2.59074807e-01 9.10630107e-01
-7.73567855e-02 -4.73782003e-01 -1.06145883e+00 8.75461340e-01
-4.75244373e-02 -1.63249040e+00 3.89887691e-02 -4.96722423e-02
1.32012606e+00 2.33094350e-01 2.44699925e-01 6.98954538e-02
2.85832077e-01 -1.15069914e+00 6.98174596e-01 2.98951030e-01
7.63336957e-01 -3.80773693e-01 4.14519399e-01 2.12682664e-01
-7.90766358e-01 1.24113999e-01 -2.86000878e-01 7.52432505e-04
1.59866959e-01 7.03162432e-01 -6.48876309e-01 3.05558115e-01
6.57605231e-01 9.06516194e-01 -1.10365534e+00 9.78528976e-01
-5.68210840e-01 8.77919197e-01 -2.85619676e-01 2.96006709e-01
6.37015760e-01 -2.31445685e-01 3.23917836e-01 1.44086778e+00
-3.32509637e-01 -5.35630226e-01 1.48346364e-01 1.06068885e+00
-1.98507249e-01 -1.61360368e-01 -3.46203536e-01 -1.19382754e-01
-1.68326586e-01 1.39033616e+00 -1.44734836e+00 -5.32534063e-01
-3.82318310e-02 1.24178791e+00 2.48892948e-01 4.75879073e-01
-8.66671383e-01 -3.02316509e-02 3.44881892e-01 -4.25576270e-02
7.66096473e-01 -3.65077794e-01 -7.11541891e-01 -9.81789172e-01
-2.53022641e-01 -7.41975486e-01 6.71220645e-02 -5.68006694e-01
-1.09443581e+00 3.29231650e-01 -1.91484988e-01 -9.15390432e-01
-1.81801930e-01 -7.33765841e-01 -5.83577037e-01 4.32870060e-01
-1.92873824e+00 -1.31083131e+00 -2.66364604e-01 4.31488842e-01
8.88110876e-01 -1.37095734e-01 7.99279153e-01 1.82226002e-02
-4.51110214e-01 3.19941163e-01 8.27402063e-03 2.56660044e-01
5.46898961e-01 -1.64780867e+00 4.57530349e-01 1.06935239e+00
7.29944944e-01 2.88524568e-01 4.76600111e-01 -3.19578350e-01
-1.08603621e+00 -1.09158456e+00 1.80139959e-01 -5.67358434e-01
4.97267395e-01 -5.90715230e-01 -7.28163004e-01 5.03671408e-01
5.16491473e-01 2.95614749e-01 3.30516994e-01 -1.00363337e-01
-3.21807176e-01 5.33217452e-02 -1.09304667e+00 4.19525862e-01
1.22121000e+00 -4.26072478e-01 -6.41685069e-01 3.86888057e-01
8.89705658e-01 -1.72015846e-01 -3.45393300e-01 4.65457708e-01
-4.58545685e-02 -1.16944420e+00 1.07726288e+00 -4.54429507e-01
6.66242719e-01 -3.63564223e-01 2.73591518e-01 -1.03780997e+00
-2.50575662e-01 -5.80996811e-01 -9.60811004e-02 1.31922102e+00
4.86162186e-01 -5.30313909e-01 9.63331461e-01 1.10315308e-01
-2.75366843e-01 -6.71960950e-01 -6.79751575e-01 -6.38961792e-01
-8.24689046e-02 -2.19809085e-01 8.32429007e-02 6.70773387e-01
-7.59840965e-01 4.23554569e-01 -2.07613651e-02 -8.88275821e-03
7.87148595e-01 3.46399397e-01 4.95560020e-01 -1.29764962e+00
-3.29072297e-01 -8.34385037e-01 -7.47243226e-01 -1.16600311e+00
2.75356591e-01 -9.95405078e-01 3.53657454e-01 -1.72204256e+00
4.00318354e-02 -4.01579946e-01 -5.03033340e-01 5.68810463e-01
-1.80887610e-01 9.42493021e-01 8.95255804e-02 1.82798117e-01
-8.49828184e-01 3.14525306e-01 1.27666390e+00 -4.50274259e-01
-3.78462933e-02 -2.07118988e-01 -7.34233499e-01 9.18833196e-01
8.50612998e-01 -4.45529103e-01 -3.15985531e-01 -5.41885138e-01
-2.92378962e-01 -7.85948634e-01 6.07478380e-01 -1.26672888e+00
6.58664445e-04 -2.22269520e-02 5.00253439e-01 -4.14059669e-01
1.53804764e-01 -6.72529519e-01 -4.87504065e-01 3.08342189e-01
-3.78303170e-01 -5.51933706e-01 3.40381742e-01 7.22158194e-01
-3.27852458e-01 -3.94560367e-01 1.08519089e+00 -2.32442588e-01
-1.18275702e+00 9.80139077e-02 -3.61284524e-01 1.68530941e-01
1.06862938e+00 -5.69380939e-01 -1.32676959e-01 -4.89060357e-02
-7.32070684e-01 -2.16405913e-02 6.99383855e-01 2.57512629e-01
3.38012785e-01 -8.69774103e-01 -4.16648507e-01 2.36297637e-01
1.22249387e-02 3.79011422e-01 -2.45008636e-02 9.11106348e-01
-5.29871821e-01 2.68044531e-01 -4.84307915e-01 -1.13851678e+00
-8.46511722e-01 6.08713031e-01 1.29446566e-01 -2.00848535e-01
-6.41727567e-01 9.82122421e-01 5.18483818e-01 -2.03370526e-01
2.22595111e-01 -3.67478788e-01 -1.24588519e-01 1.60167217e-01
1.65719450e-01 1.16640609e-02 2.33768746e-02 -5.49497962e-01
-2.62799591e-01 5.87173223e-01 7.97129348e-02 -3.25566456e-02
1.28878796e+00 -6.09536432e-02 -1.40528619e-01 5.08874059e-01
9.99823272e-01 -2.03880638e-01 -1.67361534e+00 -3.90244313e-02
4.43345100e-01 -3.39982063e-01 2.16238663e-01 -8.29424024e-01
-1.31658125e+00 1.05090582e+00 6.67656898e-01 2.95116723e-01
1.25389802e+00 1.99153826e-01 6.76151156e-01 3.26173633e-01
5.37441850e-01 -1.13903379e+00 3.49119902e-01 4.82875824e-01
5.44352412e-01 -1.52075851e+00 -1.18816055e-01 -6.28882945e-01
-7.22673118e-01 7.84326434e-01 4.26245630e-01 -5.45336068e-01
2.33498678e-01 2.17469856e-01 2.19044462e-01 4.39956114e-02
-1.56615570e-01 -8.78209829e-01 4.46817130e-01 9.67970014e-01
2.95395732e-01 -3.23947668e-02 4.34636399e-02 2.08730504e-01
1.62150934e-02 -2.32637674e-01 3.38384151e-01 7.90093422e-01
-5.22286534e-01 -1.27650177e+00 1.38771255e-02 4.29845303e-01
-4.14572716e-01 -2.40289792e-01 -5.34486294e-01 6.04166031e-01
1.27935290e-01 7.86336780e-01 2.25066110e-01 2.09294796e-01
-7.17301667e-02 5.91825657e-02 7.43178487e-01 -1.16408527e+00
-6.71027899e-01 2.33698532e-01 -1.81103364e-01 -5.56384146e-01
-9.29247439e-01 -3.43498290e-01 -1.33480644e+00 6.33560419e-01
-4.54375476e-01 -1.70074642e-01 5.64636528e-01 1.09806252e+00
4.13861752e-01 7.64771163e-01 3.72885615e-01 -1.36887252e+00
-1.79933414e-01 -9.00147676e-01 -3.87439936e-01 5.94165921e-01
3.06318253e-01 -6.99359596e-01 -2.59516448e-01 5.78306079e-01] | [9.606121063232422, 0.6644976735115051] |
adbdd2bd-270b-4bb7-ba05-80758bad2f49 | learning-controls-using-cross-modal | 1909.06993 | null | https://arxiv.org/abs/1909.06993v2 | https://arxiv.org/pdf/1909.06993v2.pdf | Learning Visuomotor Policies for Aerial Navigation Using Cross-Modal Representations | Machines are a long way from robustly solving open-world perception-control tasks, such as first-person view (FPV) aerial navigation. While recent advances in end-to-end Machine Learning, especially Imitation and Reinforcement Learning appear promising, they are constrained by the need of large amounts of difficult-to-collect labeled real-world data. Simulated data, on the other hand, is easy to generate, but generally does not render safe behaviors in diverse real-life scenarios. In this work we propose a novel method for learning robust visuomotor policies for real-world deployment which can be trained purely with simulated data. We develop rich state representations that combine supervised and unsupervised environment data. Our approach takes a cross-modal perspective, where separate modalities correspond to the raw camera data and the system states relevant to the task, such as the relative pose of gates to the drone in the case of drone racing. We feed both data modalities into a novel factored architecture, which learns a joint low-dimensional embedding via Variational Auto Encoders. This compact representation is then fed into a control policy, which we trained using imitation learning with expert trajectories in a simulator. We analyze the rich latent spaces learned with our proposed representations, and show that the use of our cross-modal architecture significantly improves control policy performance as compared to end-to-end learning or purely unsupervised feature extractors. We also present real-world results for drone navigation through gates in different track configurations and environmental conditions. Our proposed method, which runs fully onboard, can successfully generalize the learned representations and policies across simulation and reality, significantly outperforming baseline approaches. Supplementary video: https://youtu.be/VKc3A5HlUU8 | ['Vibhav Vineet', 'Sebastian Scherer', 'Ratnesh Madaan', 'Rogerio Bonatti', 'Ashish Kapoor'] | 2019-09-16 | null | null | null | null | ['drone-navigation'] | ['computer-vision'] | [ 2.56225280e-02 1.26906306e-01 -9.99902412e-02 -9.32854936e-02
-6.05947256e-01 -9.04425263e-01 7.22719193e-01 -2.57346213e-01
-5.35691857e-01 8.68540466e-01 1.35043412e-01 -2.76875407e-01
-2.39380538e-01 -5.54662406e-01 -1.16806567e+00 -6.89050436e-01
-2.85076588e-01 3.92269850e-01 5.33041693e-02 -6.17355168e-01
-2.04168737e-01 5.22082686e-01 -1.73137379e+00 -3.23378921e-01
6.54225111e-01 7.24607289e-01 3.24722946e-01 1.00295186e+00
6.93207860e-01 7.31915593e-01 -2.26540968e-01 3.02682281e-01
6.11745417e-01 -1.37571290e-01 -3.80212456e-01 1.09279901e-01
2.37139300e-01 -6.27896667e-01 -8.31148684e-01 6.86282218e-01
4.82805282e-01 7.09743857e-01 6.95454001e-01 -1.29105949e+00
-1.91622645e-01 -4.54765698e-03 -4.29795794e-02 -2.37822279e-01
3.15417290e-01 8.32252026e-01 7.68817842e-01 -2.49319211e-01
6.50627613e-01 1.04130232e+00 3.86643320e-01 8.18933427e-01
-1.20067811e+00 -4.63231027e-01 3.87451649e-01 -1.16613165e-01
-9.41850603e-01 -3.11446637e-01 4.03041691e-01 -7.42987156e-01
1.05659890e+00 -1.64070308e-01 8.02233994e-01 1.72022247e+00
2.99071521e-01 7.10236490e-01 8.55302393e-01 1.59569383e-01
3.99963945e-01 -2.18024150e-01 -5.75619459e-01 9.08115029e-01
6.97395355e-02 7.22723186e-01 -1.71966732e-01 -1.80977285e-01
8.50811899e-01 3.94496888e-01 -5.93064427e-01 -1.05441892e+00
-1.52551377e+00 7.75418460e-01 6.63879931e-01 -2.28688255e-01
-4.93845552e-01 5.51162422e-01 2.27969378e-01 3.45385313e-01
-1.13169357e-01 6.59368038e-01 -6.37475610e-01 -3.86096895e-01
-4.25754547e-01 5.88276982e-01 7.73216903e-01 1.14004946e+00
6.97681725e-01 4.01528865e-01 -5.51224314e-03 5.17609537e-01
1.82725191e-01 9.47122097e-01 3.07257563e-01 -1.22307134e+00
5.88899016e-01 2.89917082e-01 7.35221565e-01 -7.72193134e-01
-4.29711610e-01 -1.74046829e-01 -4.89845246e-01 7.08122790e-01
3.01307052e-01 -7.18791783e-01 -9.90638375e-01 2.13230515e+00
3.46059710e-01 3.51791471e-01 3.89345437e-01 1.13364065e+00
1.36950091e-01 9.67290342e-01 -2.08304018e-01 1.21575892e-01
7.49387205e-01 -9.93549645e-01 -5.32586992e-01 -4.52830106e-01
5.02091467e-01 -1.71726018e-01 9.75669086e-01 2.76411623e-01
-7.32766032e-01 -5.82324207e-01 -1.06340337e+00 2.25513354e-01
-4.87749130e-01 3.05675358e-01 2.79005587e-01 -4.48147282e-02
-9.08041954e-01 7.50271618e-01 -1.27511013e+00 -5.24530888e-01
8.90060365e-02 5.07878482e-01 -6.77860260e-01 1.40469838e-02
-1.07041228e+00 1.01852047e+00 3.19187850e-01 1.04662411e-01
-2.00027084e+00 -4.22443211e-01 -1.25462508e+00 -7.71140829e-02
9.05078471e-01 -8.09312403e-01 1.35967708e+00 -5.36908269e-01
-2.03867865e+00 2.52880365e-01 2.05340117e-01 -4.98320073e-01
4.17664737e-01 -6.26954138e-01 -1.09716117e-01 5.34268878e-02
6.28563687e-02 7.28499472e-01 1.10667884e+00 -1.36819410e+00
-6.42510593e-01 -2.04770103e-01 5.88821530e-01 4.25742358e-01
-1.96285173e-02 -6.92791879e-01 -2.00938061e-01 -4.36600685e-01
-3.26778054e-01 -1.51276445e+00 -5.95146835e-01 2.29589701e-01
-7.24382177e-02 2.05266848e-01 7.31858373e-01 -3.86023164e-01
6.26422584e-01 -1.98343825e+00 1.04799271e+00 -1.01479232e-01
-3.69830541e-02 1.72530487e-01 -1.55532166e-01 7.71107018e-01
4.12055328e-02 -3.05778563e-01 -4.09804076e-01 -2.68510222e-01
2.04633206e-01 5.89976549e-01 -6.30181789e-01 5.78560829e-01
1.58831283e-01 8.00788522e-01 -1.23498356e+00 1.22626677e-01
4.12864894e-01 3.50027472e-01 -6.64307356e-01 6.81364596e-01
-5.21133363e-01 9.12524879e-01 -5.29896677e-01 3.74977708e-01
3.44665721e-02 1.68696851e-01 1.17190637e-01 1.40723884e-01
-1.27686054e-01 -1.35593981e-01 -9.52252030e-01 2.12279439e+00
-7.26372123e-01 5.33497751e-01 2.33481690e-01 -8.94455850e-01
7.12404907e-01 2.01920763e-01 4.05742615e-01 -2.51508474e-01
2.55704463e-01 6.09032623e-03 -1.09551840e-01 -7.96629310e-01
4.48786795e-01 -7.81261623e-02 -5.89214385e-01 1.66300729e-01
5.23797452e-01 -7.17604756e-01 -9.05718505e-02 1.41620219e-01
1.21083808e+00 7.42657542e-01 2.75785089e-01 1.96256861e-02
2.06419542e-01 1.11665495e-01 5.94630957e-01 6.52303219e-01
-2.92015851e-01 3.08084965e-01 4.50712562e-01 -3.43401343e-01
-9.45997536e-01 -1.12248397e+00 4.32042539e-01 8.78129125e-01
3.09200019e-01 -3.81657928e-01 -5.75569153e-01 -7.26337373e-01
1.78721458e-01 7.86107838e-01 -7.64288127e-01 -4.19138372e-01
-3.87562424e-01 -8.87322873e-02 4.10820633e-01 5.14994383e-01
1.43449083e-01 -9.45489705e-01 -1.20770490e+00 1.07765608e-01
4.92105149e-02 -1.31437492e+00 -1.50025964e-01 2.90555179e-01
-7.01643825e-01 -1.05921268e+00 -5.73403478e-01 -4.17359859e-01
5.52868545e-01 1.21566132e-01 6.16594434e-01 -4.24633950e-01
-2.03201815e-01 1.05919039e+00 -3.68377596e-01 -2.21412718e-01
-2.17888549e-01 -2.02840701e-01 7.48651028e-01 -1.89246200e-02
-3.22357386e-01 -5.11163592e-01 -4.25098896e-01 3.48649651e-01
-7.50499427e-01 -1.07415594e-01 5.55406570e-01 1.11475730e+00
5.86805403e-01 -2.35059246e-01 -5.77005092e-03 -3.06691945e-01
4.09596950e-01 -5.74628532e-01 -9.36304212e-01 2.65662614e-02
-4.32254598e-02 2.84646541e-01 1.02988505e+00 -5.80551088e-01
-7.51168072e-01 2.70848751e-01 1.02784514e-01 -1.03681505e+00
-4.65191036e-01 2.69036382e-01 -6.08628839e-02 -1.68066975e-02
6.93987787e-01 2.12004766e-01 3.06672871e-01 -1.11084968e-01
5.06249428e-01 4.12484735e-01 5.05810678e-01 -7.71338344e-01
1.14271319e+00 5.41037261e-01 3.75695666e-03 -9.60867226e-01
-4.88860816e-01 -1.65655851e-01 -5.89669824e-01 -2.59080708e-01
9.61550295e-01 -1.16785717e+00 -8.46715927e-01 2.42633715e-01
-7.36426592e-01 -1.07215858e+00 -3.69782060e-01 7.69468665e-01
-1.16837227e+00 6.21473938e-02 -2.66819239e-01 -7.91921377e-01
3.04696649e-01 -1.36652923e+00 1.19539344e+00 2.55327910e-01
1.62371591e-01 -8.46905529e-01 4.08137947e-01 4.86652330e-02
2.33619452e-01 7.15215325e-01 3.97685587e-01 -1.63105264e-01
-5.31998754e-01 -1.26439363e-01 4.38038766e-01 3.04242015e-01
2.62035914e-02 -1.17234096e-01 -7.14845121e-01 -8.81242156e-01
-2.88819671e-01 -7.88294196e-01 6.59122884e-01 2.03488752e-01
9.28672016e-01 -1.18574448e-01 -4.28859979e-01 7.44517744e-01
1.30067182e+00 1.45943508e-01 2.72976577e-01 1.08097889e-01
8.06855142e-01 6.92305088e-01 8.92680049e-01 5.35421371e-01
4.00833726e-01 7.16700912e-01 9.33408022e-01 1.41535044e-01
4.53436226e-01 -6.94490790e-01 8.70039463e-01 5.03662586e-01
-2.10204259e-01 -2.04075709e-01 -8.65242600e-01 5.87854803e-01
-2.07704949e+00 -8.59316885e-01 5.95342040e-01 2.04639888e+00
5.24059311e-02 6.25616759e-02 -8.97530280e-03 -4.54409689e-01
1.94992289e-01 2.07836598e-01 -9.14444447e-01 -2.01717213e-01
3.06809932e-01 -1.59912258e-01 7.31040895e-01 5.27251184e-01
-1.19466484e+00 1.04690886e+00 5.15644073e+00 3.58647585e-01
-1.20692337e+00 -4.82319035e-02 -1.91825539e-01 -4.30049181e-01
9.01857391e-02 2.11037248e-01 -5.09775758e-01 2.25014344e-01
1.07322693e+00 2.45925054e-01 7.92600095e-01 1.00611734e+00
3.56042355e-01 -2.88651884e-02 -1.23957360e+00 9.95566905e-01
-1.29544083e-02 -1.10024416e+00 -1.93819121e-01 2.11183593e-01
7.28113115e-01 4.23550844e-01 1.55285046e-01 7.92282999e-01
5.95490932e-01 -8.64111543e-01 8.19826663e-01 6.59065723e-01
7.88682699e-01 -6.08721673e-01 4.66608316e-01 7.91203737e-01
-1.01008511e+00 -5.95622718e-01 -3.84127438e-01 -2.62145370e-01
3.09535891e-01 -3.94408524e-01 -6.07461751e-01 5.94565690e-01
6.54596269e-01 1.01096892e+00 -8.89027566e-02 7.02745199e-01
-4.31336880e-01 4.28997666e-01 -4.53746468e-01 -1.63467154e-01
7.13826895e-01 -2.65217602e-01 8.95851910e-01 5.62041104e-01
3.66702259e-01 1.82910159e-01 4.50674981e-01 7.76693642e-01
2.37066507e-01 -5.50180137e-01 -1.39480889e+00 -1.99213132e-01
1.72753949e-02 1.11189198e+00 -2.76249558e-01 -1.29949480e-01
-3.46791565e-01 9.25817490e-01 4.52260911e-01 6.69910133e-01
-1.06921923e+00 -2.79998124e-01 1.15935981e+00 -2.12015316e-01
5.36778927e-01 -6.87333524e-01 6.51038229e-01 -1.52011549e+00
-1.16743036e-01 -8.78469348e-01 -1.12803169e-01 -7.36711562e-01
-8.51357698e-01 6.27189696e-01 2.84613520e-01 -1.76966918e+00
-6.76287353e-01 -8.55961442e-01 -4.25077140e-01 5.81988692e-01
-1.32424295e+00 -1.08700848e+00 -3.14083636e-01 6.72218800e-01
6.46757901e-01 -4.55525875e-01 9.58587110e-01 -8.90172794e-02
-5.87407351e-01 2.14110613e-01 2.63029784e-01 -9.31954533e-02
4.87233669e-01 -1.21238613e+00 3.37445199e-01 8.91043782e-01
3.85106504e-02 5.73849320e-01 8.25096786e-01 -5.16768217e-01
-1.79216838e+00 -1.14836025e+00 -3.39607865e-01 -5.49545348e-01
7.86484003e-01 -6.78593695e-01 -4.58937526e-01 9.26372588e-01
1.83856208e-02 2.68143177e-01 2.67861158e-01 -1.96545348e-01
1.76761970e-02 6.22665323e-02 -8.79107356e-01 8.92784297e-01
1.05168402e+00 -5.69115341e-01 -5.95245898e-01 2.47225150e-01
7.98655391e-01 -7.03315914e-01 -6.92408442e-01 4.04114991e-01
5.94781518e-01 -6.46696568e-01 8.13044310e-01 -1.01546621e+00
2.59331107e-01 -5.73428452e-01 -2.71916360e-01 -1.82211697e+00
-2.55860448e-01 -8.31606925e-01 -1.71584591e-01 4.54266846e-01
3.73797715e-01 -5.66203535e-01 5.16508877e-01 2.84925848e-01
-2.88125604e-01 -8.60044122e-01 -8.05172563e-01 -7.90073991e-01
4.69734147e-02 -3.56530339e-01 3.09115529e-01 4.70834613e-01
-1.33581981e-01 2.21339583e-01 -8.16315472e-01 6.58876777e-01
4.17640060e-01 8.47735852e-02 1.20978403e+00 -8.03605318e-01
-6.23619258e-01 4.31568362e-02 -5.10924995e-01 -1.32883251e+00
6.18044376e-01 -6.10780001e-01 4.11016613e-01 -1.30599499e+00
-4.98206109e-01 -2.02100798e-01 6.06690273e-02 4.66421247e-01
1.34983182e-01 -4.06731725e-01 2.80285895e-01 -6.10406138e-03
-5.13451099e-01 1.21879411e+00 1.23145413e+00 -1.89104706e-01
-2.35128820e-01 5.80603955e-03 -1.29005402e-01 7.00109184e-01
7.31367230e-01 -3.34034890e-01 -8.02937388e-01 -7.19615757e-01
4.00611833e-02 4.99255866e-01 5.97443342e-01 -1.36527479e+00
1.52789295e-01 -3.69984150e-01 1.84544593e-01 -9.68267024e-02
8.58498394e-01 -1.09657133e+00 -7.86646605e-02 5.96502483e-01
-2.47689351e-01 1.72458515e-01 4.76798981e-01 1.25538266e+00
-8.09157938e-02 2.07082033e-01 5.33003569e-01 -1.42655075e-01
-8.95282507e-01 4.72709507e-01 -6.19970977e-01 7.94478282e-02
1.46353984e+00 8.95611793e-02 -5.16781732e-02 -6.25578463e-01
-8.98548603e-01 7.58162141e-01 6.33897483e-01 7.07284391e-01
7.11811185e-01 -1.02597213e+00 -3.94201517e-01 3.88266772e-01
2.58977085e-01 8.09507743e-02 4.80406493e-01 6.75972879e-01
-4.75775182e-01 2.70138234e-01 -5.56906044e-01 -7.52910674e-01
-6.44799352e-01 7.05577493e-01 4.75781411e-01 -1.45275882e-02
-6.56297922e-01 5.59239566e-01 2.35872701e-01 -1.00177002e+00
3.04533280e-02 -6.42079234e-01 -1.07417636e-01 -3.08713198e-01
1.47558644e-01 9.33621451e-02 -4.22026724e-01 -7.24945962e-01
-1.09507121e-01 6.69812739e-01 2.56872296e-01 -4.09219623e-01
1.31402695e+00 5.55986613e-02 4.58686233e-01 5.85628629e-01
9.41118360e-01 -2.34801680e-01 -2.01379108e+00 1.07945949e-01
-5.37841439e-01 -3.28810215e-01 -1.55656375e-02 -5.00674903e-01
-6.68527246e-01 9.24362421e-01 6.22387588e-01 -2.41261840e-01
7.43275583e-01 -2.72911102e-01 6.58103108e-01 9.83212233e-01
8.45608830e-01 -1.07825983e+00 1.93672463e-01 6.85069978e-01
1.01602495e+00 -1.49622810e+00 -3.46909404e-01 3.09712112e-01
-9.68174756e-01 9.38075006e-01 7.71254599e-01 -5.87071300e-01
5.80847204e-01 1.00917764e-01 4.70718741e-02 -4.36730050e-02
-1.01482272e+00 -4.44594264e-01 1.65118631e-02 7.02420652e-01
-4.64580983e-01 2.27082118e-01 5.39985776e-01 3.49088073e-01
-1.02341302e-01 -2.00796887e-01 6.75289631e-01 1.26253378e+00
-2.86042750e-01 -8.07177126e-01 -1.92560002e-01 2.31723428e-01
2.14356855e-01 4.95618612e-01 -1.04600340e-01 9.78148282e-01
1.27440035e-01 6.46045566e-01 -1.27705067e-01 -6.26573741e-01
6.73064411e-01 -1.93889126e-01 5.44270933e-01 -7.50054657e-01
-1.78069860e-01 -2.15109542e-01 -1.25522554e-01 -1.08720279e+00
-2.37046450e-01 -6.93908751e-01 -1.10842228e+00 1.90779530e-02
7.63102844e-02 5.04843518e-02 5.23917854e-01 8.59964192e-01
4.53304052e-01 6.53231502e-01 4.81861591e-01 -1.62438858e+00
-8.66447270e-01 -8.09588611e-01 -3.60149771e-01 2.45586991e-01
9.55766022e-01 -1.31150007e+00 -3.70526373e-01 -3.42679508e-02] | [4.580878734588623, 1.0460129976272583] |
6af09cc7-c227-42c1-b224-b492638194b4 | sound-event-localization-based-on-sound | 2002.05994 | null | http://arxiv.org/abs/2002.05994v1 | http://arxiv.org/pdf/2002.05994v1.pdf | Sound Event Localization based on Sound Intensity Vector Refined By DNN-Based Denoising and Source Separation | We propose a direction-of-arrival (DOA) estimation method for Sound Event
Localization and Detection (SELD). Direct estimation of DOA using a deep neural
network (DNN), i.e. completely-datadriven approach, achieves high accuracy.
However, there is a gap in the accuracy between DOA estimation for single and
overlapping sources because they cannot incorporate physical knowledge.
Meanwhile, although the accuracy of physics-based approaches is inferior to
DNN-based approaches, it is robust for overlapping source. In this study, we
consider a combination of physics-based and DNN-based approaches; the sound
intensity vectors (IVs) for physics-based DOA estimation is refined based on
DNN-based denoising and source separation. This method enables the accurate DOA
estimation for both single and overlapping sources using a spherical microphone
array. Experimental results show that the proposed method achieves
state-of-the-art DOA estimation accuracy on an open dataset of the SELD. | [] | 2020-02-14 | null | null | null | null | ['sound-event-localization-and-detection'] | ['audio'] | [-4.80353266e-01 -8.29895377e-01 6.42851591e-01 -9.61297899e-02
-1.23077166e+00 -4.86988753e-01 3.25757712e-01 1.64748877e-01
-1.41941547e-01 4.75570172e-01 6.22537792e-01 5.42998612e-02
-4.05864269e-01 -9.44876254e-01 -5.08453310e-01 -1.17864215e+00
-1.15508966e-01 -1.02366716e-01 2.49449074e-01 1.49259448e-01
-4.84427512e-02 7.25549877e-01 -1.79987860e+00 -2.54330814e-01
6.67780459e-01 1.30973172e+00 1.18784502e-01 8.05683315e-01
-2.71664292e-01 7.85181940e-01 -8.65837276e-01 2.20929846e-01
2.31155694e-01 -5.08778811e-01 1.64908394e-01 -6.82739258e-01
6.40078783e-01 -4.05560076e-01 -5.12002826e-01 9.35272217e-01
1.14349294e+00 3.18111539e-01 7.22684383e-01 -1.03040469e+00
-4.08005148e-01 1.70562759e-01 -3.52425963e-01 6.34678245e-01
3.51614118e-01 -1.59055606e-01 6.42028451e-01 -1.26681173e+00
-3.56022827e-02 9.96278048e-01 1.13435161e+00 1.38883457e-01
-5.86607277e-01 -7.12656736e-01 -3.55583489e-01 9.61371791e-03
-1.47226274e+00 -5.71917117e-01 1.13133693e+00 -3.13536793e-01
5.68989098e-01 5.45286052e-02 5.95949650e-01 8.69995058e-01
8.49218145e-02 5.06053686e-01 1.03365612e+00 -5.46955347e-01
6.80358708e-01 -3.27518493e-01 6.07839189e-02 5.26289165e-01
2.27749139e-01 2.70420313e-01 -1.09482682e+00 -3.72693330e-01
7.25998998e-01 -3.48767601e-02 -4.29910481e-01 1.52720645e-01
-1.16185927e+00 5.69744349e-01 2.51608014e-01 4.33064371e-01
-8.16222370e-01 3.53206187e-01 -1.07761718e-01 -1.05890661e-01
8.15450430e-01 9.93918031e-02 -3.57412487e-01 -3.16528171e-01
-1.30648458e+00 3.25427443e-01 7.99102485e-01 4.90907758e-01
5.01676381e-01 6.74097598e-01 -8.72466937e-02 9.98744726e-01
7.16079652e-01 1.35398340e+00 1.53503299e-01 -1.13866484e+00
1.07355565e-01 -2.41464436e-01 2.69013226e-01 -1.40666318e+00
-6.08388543e-01 -8.57000709e-01 -9.87140954e-01 2.40402564e-01
5.94008803e-01 -5.56325436e-01 -8.40059519e-01 1.72672880e+00
5.87614417e-01 9.28518534e-01 9.50244889e-02 1.02162373e+00
1.31973362e+00 9.31839705e-01 -2.69779146e-01 -2.80581772e-01
1.09733820e+00 -7.24658072e-01 -9.93681014e-01 7.41862282e-02
-6.59398735e-02 -9.71441984e-01 5.23857296e-01 8.65630031e-01
-1.10807967e+00 -7.59952366e-01 -8.84491980e-01 2.35473216e-01
-2.26394564e-01 -6.78968802e-02 5.66956997e-01 9.34014976e-01
-1.05713427e+00 1.93560302e-01 -8.79440665e-01 -8.62345472e-02
-2.19407938e-02 -2.41152346e-01 8.25587064e-02 6.68100342e-02
-1.02620602e+00 2.86798239e-01 -4.79792565e-01 3.39333266e-01
-1.24080896e+00 -1.26912165e+00 -7.63200700e-01 8.68330225e-02
-1.26492068e-01 -6.81221783e-01 1.11319411e+00 -4.05824661e-01
-1.65002620e+00 1.01089671e-01 -5.70775747e-01 -4.95661765e-01
-8.26460402e-03 -3.82118821e-01 -8.37711036e-01 4.61662382e-01
1.85511172e-01 1.83379903e-01 8.67305458e-01 -1.44625211e+00
-6.94724262e-01 -2.78076269e-02 -2.33654633e-01 -6.41902238e-02
-3.37464213e-01 1.74966186e-01 -1.08739913e-01 -9.14855778e-01
6.07350886e-01 -4.31236655e-01 1.37971882e-02 3.29939663e-01
-2.72109151e-01 -7.33481944e-02 4.39242393e-01 -7.67158687e-01
1.21007395e+00 -2.08033180e+00 -4.36785847e-01 5.82480766e-02
5.10182083e-01 1.03257045e-01 -9.43302885e-02 5.16603172e-01
2.02211991e-01 -3.58756781e-01 -1.36568815e-01 -5.99357128e-01
-1.08628003e-02 -1.78795993e-01 -3.97583932e-01 5.25896013e-01
-1.64543450e-01 1.91919789e-01 -9.96389449e-01 -3.25276107e-01
3.81498247e-01 9.36591387e-01 -6.23065352e-01 3.59695673e-01
2.29353786e-01 6.39788806e-01 -3.62134844e-01 8.12841952e-01
1.26639593e+00 2.84629673e-01 -5.35115838e-01 -3.91116709e-01
-4.26164597e-01 3.65602881e-01 -1.61323261e+00 1.68409789e+00
-1.00994802e+00 8.35075617e-01 4.54923719e-01 -7.55196095e-01
9.86189425e-01 5.71505427e-01 8.09113622e-01 -3.83770525e-01
-1.17628582e-01 4.55396235e-01 -2.24555701e-01 -5.00291109e-01
1.85688555e-01 -3.61403704e-01 3.19770306e-01 3.61000150e-01
1.73329055e-01 -4.39940542e-02 -1.92743421e-01 -1.27548397e-01
1.26461935e+00 -7.94662312e-02 3.10298018e-02 -2.75540888e-01
4.81616199e-01 -3.83545160e-01 7.84474790e-01 1.10252476e+00
-3.95563036e-01 1.09741855e+00 -2.95173496e-01 -2.88651198e-01
-4.49516684e-01 -1.44700801e+00 -2.42989048e-01 6.33055091e-01
1.96659341e-01 -2.57708728e-01 -6.76785827e-01 -2.28220925e-01
-2.26538461e-02 8.49490523e-01 -1.97532222e-01 1.31212488e-01
-6.38067186e-01 -9.21979964e-01 7.76489735e-01 4.47035939e-01
7.60959446e-01 -5.62651575e-01 -3.53272170e-01 6.07604146e-01
-3.83779615e-01 -1.14557433e+00 -2.17745647e-01 2.54664291e-03
-6.74128592e-01 -5.98739326e-01 -8.70685577e-01 -3.45683753e-01
1.40253425e-01 7.61331797e-01 9.86239791e-01 -1.84817672e-01
1.25818476e-01 1.06822789e+00 -2.52796769e-01 -9.55059052e-01
-2.36241430e-01 -6.80744529e-01 3.87542874e-01 3.31404001e-01
3.03017586e-01 -1.20331991e+00 -1.08693683e+00 1.86536431e-01
-5.79282701e-01 -5.07283926e-01 2.81883061e-01 9.17330757e-02
6.17416382e-01 4.63199526e-01 7.85182655e-01 1.31925628e-01
5.97822905e-01 -6.88311100e-01 -4.98492599e-01 -2.47046694e-01
-4.84118283e-01 -5.35308421e-01 5.73721170e-01 -3.13929975e-01
-1.37462819e+00 -1.43499613e-01 -5.19714057e-01 -4.45060343e-01
-5.39550066e-01 4.84304965e-01 -1.24223545e-01 -1.25500396e-01
6.35250032e-01 2.83665776e-01 -4.75255281e-01 -9.59615886e-01
5.30023873e-02 7.28166640e-01 6.00167036e-01 -4.32402194e-01
8.18461239e-01 8.82385015e-01 2.71400541e-01 -1.14763784e+00
-9.23405409e-01 -7.56830454e-01 -3.92690390e-01 -3.85065228e-01
8.49332631e-01 -1.23543251e+00 -5.71419239e-01 9.38884139e-01
-1.47054672e+00 4.77270782e-02 -1.04293078e-01 1.06562090e+00
-3.34149636e-02 2.44491085e-01 -4.17377919e-01 -1.21570003e+00
-1.81901544e-01 -7.98591673e-01 1.21499908e+00 3.11518848e-01
1.98172510e-01 -1.25846040e+00 4.32541937e-01 1.73453599e-01
6.35236561e-01 2.81601995e-01 3.74762028e-01 -3.88266325e-01
-6.23148024e-01 -1.81394115e-01 -1.08956531e-01 5.22381961e-01
2.22003743e-01 -2.75501221e-01 -1.44692540e+00 2.09479138e-01
6.38574421e-01 3.25259179e-01 6.18604839e-01 1.20474994e+00
7.95911312e-01 4.40445952e-02 -3.11405286e-02 5.88369131e-01
1.57891405e+00 1.20372474e-01 4.68883604e-01 1.05026074e-01
7.82813191e-01 1.46253407e-01 3.39685529e-01 6.09332800e-01
4.58298415e-01 4.77764964e-01 3.89744729e-01 -1.95003241e-01
-8.50180686e-01 1.42470719e-02 2.23526523e-01 1.21229696e+00
-6.12893142e-02 -7.84789562e-01 -8.98272574e-01 7.31098711e-01
-1.41259384e+00 -8.28324080e-01 -7.99109042e-01 2.13704515e+00
6.14843011e-01 -1.48775086e-01 -7.79415220e-02 3.54891360e-01
6.66883230e-01 3.60999376e-01 -8.69361982e-02 -5.25075085e-02
-1.85824737e-01 3.00770372e-01 4.21360940e-01 6.32866144e-01
-9.50257897e-01 2.57575035e-01 6.83964777e+00 9.88873303e-01
-1.09039295e+00 7.54105747e-01 -1.71971414e-02 -5.09826047e-03
-3.70257258e-01 -4.77997780e-01 -8.03361058e-01 5.54460049e-01
1.11412537e+00 3.93432021e-01 3.97090375e-01 5.08342624e-01
8.24302852e-01 -4.46904451e-01 -7.38624871e-01 1.22189558e+00
1.31438524e-01 -1.28372717e+00 -2.90947050e-01 -1.90780729e-01
7.51547992e-01 1.96337700e-01 -1.62149623e-01 -7.46187568e-02
-5.82270510e-02 -5.87245226e-01 8.96983743e-01 9.33432400e-01
2.51532942e-01 -6.19096160e-01 6.81925893e-01 2.92397410e-01
-1.20770085e+00 1.24953210e-01 -3.39117348e-01 -2.67928660e-01
4.51075137e-01 1.66654658e+00 -2.67115384e-01 6.09860122e-01
1.11898601e+00 5.29196620e-01 8.00163448e-02 1.47211421e+00
-2.74401367e-01 1.41653717e+00 -7.65570283e-01 -2.99208499e-02
8.90273377e-02 -2.36410052e-02 1.16535878e+00 1.26863468e+00
8.41117322e-01 7.21450970e-02 -1.98713392e-02 8.55569780e-01
1.47304349e-02 -1.21765062e-01 -4.98685420e-01 4.65626597e-01
7.29024887e-01 1.21387517e+00 -6.32237256e-01 -9.84848216e-02
-4.31633294e-01 4.30561066e-01 -4.45214987e-01 6.36821687e-01
-8.40887010e-01 -5.30722857e-01 7.38606513e-01 7.56779015e-02
3.34903747e-01 -5.49450397e-01 -2.72419304e-01 -7.92813540e-01
1.24056280e-01 -4.80875105e-01 2.32235566e-02 -1.03118885e+00
-1.44806075e+00 2.93038189e-01 1.03498653e-01 -1.44656765e+00
1.11788690e-01 -5.43631256e-01 -1.07303882e+00 8.48257601e-01
-1.81195891e+00 -7.87549376e-01 -5.21990001e-01 4.53151941e-01
5.06676316e-01 -5.35470620e-02 7.23280489e-01 8.68304729e-01
-3.19312066e-01 1.65017545e-01 5.63138604e-01 -1.53403189e-02
5.64769924e-01 -1.22119856e+00 4.82455343e-02 1.11425066e+00
2.79252291e-01 3.94708723e-01 9.79288399e-01 -4.54953969e-01
-1.49514651e+00 -9.41556215e-01 8.44595432e-01 -4.31560963e-01
6.60467207e-01 -3.41616929e-01 -7.37041473e-01 5.04907146e-02
2.02094808e-01 4.37920183e-01 8.16269815e-01 -1.33086056e-01
-8.54548067e-02 -5.36799073e-01 -1.16185641e+00 1.22315384e-01
9.83189464e-01 -5.04701257e-01 -4.87466037e-01 4.36544806e-01
5.20082414e-01 -2.67800927e-01 -8.13592076e-01 3.31771851e-01
3.95554006e-01 -1.29625213e+00 1.48430586e+00 2.28125691e-01
2.01994821e-01 -6.35109186e-01 -4.54341531e-01 -1.38491452e+00
-1.25603482e-01 -6.31360233e-01 -4.15907890e-01 1.58347499e+00
-5.30892126e-02 -7.26821542e-01 2.26591945e-01 -1.11535415e-01
-2.71974266e-01 -2.78007597e-01 -1.22275877e+00 -9.84973729e-01
2.17072740e-02 -1.23644245e+00 7.35344172e-01 6.18757486e-01
-7.65997648e-01 2.45422218e-02 -2.80791879e-01 1.19515824e+00
1.08145535e+00 1.12598678e-02 5.64855099e-01 -1.37830496e+00
-2.96691924e-01 -1.65379524e-01 -2.20950693e-01 -1.19095254e+00
-1.69565335e-01 -4.78079706e-01 4.78201479e-01 -1.85363233e+00
-3.80893856e-01 -4.37030882e-01 -6.43931627e-01 -2.01795802e-01
-1.05007766e-02 3.98095667e-01 -2.45495588e-01 1.74776837e-01
-4.16484237e-01 5.82132459e-01 7.50554204e-01 9.47196037e-02
-1.96023613e-01 2.63928741e-01 -2.75159001e-01 1.26013291e+00
5.64067245e-01 -9.81787443e-01 -2.47302234e-01 -7.62736678e-01
2.67233580e-01 2.01913089e-01 8.23049188e-01 -1.54807591e+00
5.12798309e-01 5.97132482e-02 2.81646788e-01 -9.62801516e-01
3.83198917e-01 -6.79169893e-01 -5.98047599e-02 1.86230004e-01
1.12924194e-02 -4.04669166e-01 3.06089908e-01 8.52888048e-01
-4.74715412e-01 -2.48901799e-01 5.08290231e-01 1.19321615e-01
-3.58554602e-01 1.12561017e-01 -7.67135322e-01 1.27221018e-01
2.29864046e-01 9.85407084e-03 -5.08360751e-02 -8.56483698e-01
-2.75946796e-01 -4.71943825e-01 -4.72723782e-01 -1.70564242e-02
6.87477887e-01 -1.46549511e+00 -8.28085303e-01 -1.80265918e-01
-2.49791875e-01 -1.52356038e-02 4.87220466e-01 9.99581456e-01
-4.78533626e-01 1.64556846e-01 4.00409818e-01 -6.95670307e-01
-8.13003063e-01 2.36822665e-01 7.19102919e-01 1.39289975e-01
-4.63768542e-01 1.12736261e+00 4.69698340e-01 -6.12888396e-01
3.68312150e-01 -5.20617664e-01 -1.80705979e-01 -1.57811597e-01
7.36777186e-01 9.81108725e-01 1.95915878e-01 -5.66753149e-01
-5.17627776e-01 9.11509752e-01 9.55387115e-01 -3.70687932e-01
1.35582674e+00 -5.66633284e-01 -1.66042820e-01 5.32231271e-01
1.27538800e+00 8.87314439e-01 -1.06429577e+00 -2.34609723e-01
-6.03449345e-01 -4.53778535e-01 9.81619239e-01 -9.33601081e-01
-1.25623631e+00 1.20876491e+00 9.77335334e-01 4.91376787e-01
1.32204044e+00 -1.53715760e-01 1.22549474e+00 4.48822796e-01
2.86188394e-01 -8.25237811e-01 5.84846847e-02 3.15361947e-01
8.73520970e-01 -1.02561331e+00 -1.69099063e-01 -3.74354720e-01
2.23751143e-01 1.23523676e+00 1.76139891e-01 -1.24777488e-01
1.45391178e+00 4.49889719e-01 4.24457550e-01 -3.87732267e-01
-4.57745269e-02 -1.63449302e-01 4.15089190e-01 7.70342708e-01
1.57045349e-01 -2.23776013e-01 1.60449758e-01 7.53644884e-01
-8.75218362e-02 -1.58774912e-01 3.84417802e-01 7.77796984e-01
-5.48575997e-01 -5.87024212e-01 -9.43142116e-01 -7.05965534e-02
-6.35891974e-01 -4.00680661e-01 3.05122197e-01 2.98688889e-01
4.38120842e-01 1.61860323e+00 3.27065936e-03 -9.43324342e-02
3.79127860e-01 -1.56458169e-01 2.22944185e-01 -2.83912808e-01
-4.76051867e-01 4.33518648e-01 1.23571977e-01 -4.43697691e-01
-7.86551356e-01 -6.65223002e-01 -1.36330175e+00 -1.18310682e-01
-4.19085860e-01 1.33191779e-01 1.00928724e+00 8.90826344e-01
5.32738209e-01 7.30154514e-01 9.42095101e-01 -9.29954350e-01
-6.69923574e-02 -7.76754320e-01 -8.42918038e-01 -1.92249656e-01
7.64769316e-01 -8.17042828e-01 -9.84156966e-01 -4.70302440e-02] | [15.18129825592041, 5.704602241516113] |
d9927513-96d0-416e-9e81-696356f2533d | reconstructive-sparse-code-transfer-for | 1410.4521 | null | http://arxiv.org/abs/1410.4521v1 | http://arxiv.org/pdf/1410.4521v1.pdf | Reconstructive Sparse Code Transfer for Contour Detection and Semantic Labeling | We frame the task of predicting a semantic labeling as a sparse
reconstruction procedure that applies a target-specific learned transfer
function to a generic deep sparse code representation of an image. This
strategy partitions training into two distinct stages. First, in an
unsupervised manner, we learn a set of generic dictionaries optimized for
sparse coding of image patches. We train a multilayer representation via
recursive sparse dictionary learning on pooled codes output by earlier layers.
Second, we encode all training images with the generic dictionaries and learn a
transfer function that optimizes reconstruction of patches extracted from
annotated ground-truth given the sparse codes of their corresponding image
patches. At test time, we encode a novel image using the generic dictionaries
and then reconstruct using the transfer function. The output reconstruction is
a semantic labeling of the test image.
Applying this strategy to the task of contour detection, we demonstrate
performance competitive with state-of-the-art systems. Unlike almost all prior
work, our approach obviates the need for any form of hand-designed features or
filters. To illustrate general applicability, we also show initial results on
semantic part labeling of human faces.
The effectiveness of our approach opens new avenues for research on deep
sparse representations. Our classifiers utilize this representation in a novel
manner. Rather than acting on nodes in the deepest layer, they attach to nodes
along a slice through multiple layers of the network in order to make
predictions about local patches. Our flexible combination of a generatively
learned sparse representation with discriminatively trained transfer
classifiers extends the notion of sparse reconstruction to encompass arbitrary
semantic labeling tasks. | ['Pietro Perona', 'Stella X. Yu', 'Michael Maire'] | 2014-10-16 | null | null | null | null | ['contour-detection'] | ['computer-vision'] | [ 6.31732762e-01 4.89514202e-01 -2.91132480e-01 -7.03755736e-01
-7.08684385e-01 -6.01075590e-01 5.00096262e-01 -1.68593183e-01
-3.38898529e-03 3.97774756e-01 2.61070669e-01 7.63560683e-02
3.05369139e-01 -8.95094037e-01 -9.87959504e-01 -7.20046580e-01
6.20830022e-02 6.64376855e-01 8.92431363e-02 3.61984298e-02
1.83806047e-01 6.43206418e-01 -1.62288713e+00 8.37018013e-01
2.06003994e-01 1.13221526e+00 3.25888216e-01 3.14624041e-01
-3.46480161e-02 6.67011857e-01 -2.64333099e-01 -2.65754074e-01
2.39357635e-01 -6.06355488e-01 -8.35815609e-01 6.05755568e-01
6.19592786e-01 -1.35104105e-01 -4.20454770e-01 1.06291819e+00
1.34152055e-01 -7.71064870e-03 8.08208644e-01 -6.35268927e-01
-8.07556748e-01 3.43736172e-01 -5.14061391e-01 6.87081041e-03
2.42996871e-01 -1.42766580e-01 1.10261667e+00 -1.11594605e+00
7.47237444e-01 1.11295831e+00 8.83974195e-01 5.53017974e-01
-1.58549678e+00 -3.98142308e-01 1.96633309e-01 -1.62853748e-01
-1.35459363e+00 -7.14989841e-01 8.40626478e-01 -5.27705312e-01
9.53513861e-01 -1.20938368e-01 5.55609286e-01 8.60150516e-01
-1.43267000e-02 6.95039153e-01 8.61365676e-01 -4.45680022e-01
3.61326516e-01 2.12816671e-01 -1.27134919e-01 1.22546661e+00
4.87941429e-02 1.42698556e-01 -4.19696003e-01 -1.17032729e-01
8.89682591e-01 2.49199450e-01 -1.73696861e-01 -7.13194251e-01
-7.70008683e-01 1.23506296e+00 6.87156320e-01 4.09135133e-01
-3.79038423e-01 1.82308674e-01 6.06810525e-02 1.56052366e-01
5.85789979e-01 2.22325414e-01 -4.00918871e-01 5.51838219e-01
-1.23512375e+00 -8.62713754e-02 9.36014175e-01 7.35058784e-01
1.27654994e+00 1.94223166e-01 2.25098222e-03 1.02496672e+00
3.34888488e-01 1.69218168e-01 4.23194706e-01 -1.10129571e+00
-7.02651888e-02 4.23904479e-01 -3.46630901e-01 -9.29492712e-01
-2.32324958e-01 -7.02962697e-01 -6.40335023e-01 2.20872343e-01
1.78154230e-01 -1.18731922e-02 -1.22077024e+00 1.78099608e+00
1.13303050e-01 4.51951385e-01 2.37845834e-02 8.09330702e-01
4.94862139e-01 6.61783576e-01 4.81844954e-02 7.30459690e-02
1.06060147e+00 -8.77135456e-01 -4.97700907e-02 -4.48189497e-01
5.25629282e-01 -7.84410596e-01 6.20864153e-01 3.39859784e-01
-1.06658638e+00 -6.43281877e-01 -8.96298468e-01 -2.15951707e-02
-2.39538655e-01 3.59485179e-01 7.89333105e-01 4.01218265e-01
-1.38989174e+00 5.76918781e-01 -6.93736494e-01 -4.66265827e-01
9.00405824e-01 4.06421363e-01 -4.48043734e-01 -4.16981548e-01
-6.57119572e-01 8.36073637e-01 2.05855817e-01 -1.10430583e-01
-1.35152173e+00 -4.45599586e-01 -1.33917463e+00 2.34810829e-01
-1.72574297e-01 -6.97721541e-01 1.13547409e+00 -1.60175824e+00
-1.20876598e+00 1.26006711e+00 -2.80998886e-01 -3.37036133e-01
-3.39604616e-01 2.64302760e-01 -7.88569972e-02 3.48923326e-01
4.00984317e-01 1.03328371e+00 1.17075920e+00 -1.49665678e+00
-4.44618970e-01 -1.97672322e-01 -1.29515612e-02 -6.81822449e-02
-9.66412723e-02 -3.07322711e-01 -3.61942619e-01 -8.94254565e-01
3.64940763e-01 -8.36905777e-01 -2.43752182e-01 1.74474329e-01
-1.23249136e-01 3.31751294e-02 8.10262442e-01 -5.16864419e-01
6.54075444e-01 -2.41547751e+00 3.73180002e-01 4.40416723e-01
3.24459672e-02 1.52961304e-02 -5.69397628e-01 2.36129001e-01
-3.24406832e-01 -2.45006159e-01 -7.76964664e-01 -6.44607186e-01
-2.39541009e-01 5.21383882e-01 -4.32581991e-01 6.57427192e-01
6.59227490e-01 9.22907054e-01 -8.08309674e-01 -2.24023864e-01
2.86818277e-02 6.26488745e-01 -1.08612037e+00 3.66312236e-01
-4.13650751e-01 3.46175551e-01 -2.94597626e-01 7.55881369e-01
4.00616229e-01 -6.03406072e-01 3.84128571e-01 -2.87439287e-01
1.61600620e-01 3.33908975e-01 -9.40144897e-01 2.25515652e+00
-5.78439116e-01 5.27207613e-01 7.93392882e-02 -1.83388221e+00
8.24866474e-01 2.40813628e-01 4.81381536e-01 -4.03077424e-01
8.55216458e-02 1.88755676e-01 -3.05415988e-01 -3.31998110e-01
-1.16496287e-01 -5.81707120e-01 5.54651991e-02 5.77518940e-01
9.91925836e-01 -1.06509775e-01 -2.08115861e-01 2.01022089e-01
1.06966376e+00 3.44343156e-01 2.64567226e-01 -1.99513525e-01
3.43968600e-01 -3.81448679e-02 3.08941931e-01 4.37312603e-01
2.53434151e-01 9.45964992e-01 4.35607582e-01 -6.23387516e-01
-1.01678109e+00 -1.08710337e+00 -3.52021128e-01 1.28148007e+00
-2.31165588e-01 -2.77194858e-01 -5.93906224e-01 -8.67042899e-01
2.60685086e-01 3.97017479e-01 -1.00189793e+00 -9.22231525e-02
-4.45343345e-01 -5.19065320e-01 1.91293091e-01 5.05334258e-01
1.63986102e-01 -1.31143022e+00 -3.75410974e-01 8.82125720e-02
1.76637456e-01 -9.53540146e-01 -3.37955892e-01 6.72738433e-01
-9.26899970e-01 -9.14327323e-01 -6.21442497e-01 -1.45726717e+00
1.16183829e+00 3.67529094e-01 1.12382627e+00 2.24201724e-01
-4.76225823e-01 6.17399693e-01 -1.99247301e-01 1.34511828e-01
-2.00925425e-01 -3.38215828e-02 -4.37776804e-01 4.00463969e-01
2.66450375e-01 -6.91840410e-01 -4.04880017e-01 -2.87642982e-02
-8.49231660e-01 -1.62070632e-01 6.81496441e-01 1.05880547e+00
8.95757198e-01 -7.16750324e-02 5.03519237e-01 -1.19621921e+00
1.18747225e-03 -9.13561344e-01 -4.90853697e-01 -3.46350074e-02
-1.46524400e-01 3.43576312e-01 3.65625679e-01 -3.56410831e-01
-8.91752481e-01 7.64512420e-01 -2.89022237e-01 -5.40151894e-01
-3.09542209e-01 5.24609566e-01 -4.66637723e-02 -3.55223596e-01
7.07765222e-01 2.84777164e-01 1.26505375e-01 -5.20103395e-01
5.59077561e-01 4.26653475e-01 6.22937799e-01 -6.69959486e-01
7.51198351e-01 6.24506056e-01 -2.18092099e-01 -6.82946503e-01
-1.23688567e+00 -3.88325751e-01 -6.06155396e-01 1.54945925e-01
9.18044806e-01 -1.07319272e+00 -3.28117192e-01 6.95187598e-02
-1.15176773e+00 -4.55766469e-01 -6.79556489e-01 1.45710468e-01
-7.43560314e-01 8.29508081e-02 -5.73543251e-01 -3.98469567e-01
1.82652712e-01 -9.68673587e-01 1.50317693e+00 2.84366794e-02
-2.21297562e-01 -1.13984525e+00 5.58887571e-02 2.44984686e-01
2.71584362e-01 1.24255925e-01 8.76128554e-01 -6.13103628e-01
-7.04854429e-01 -1.62786156e-01 -1.56167299e-01 5.62794864e-01
1.85980216e-01 -5.69040000e-01 -1.18788564e+00 -4.00765121e-01
2.11758405e-01 -7.55905688e-01 1.23503315e+00 3.71507585e-01
1.38467932e+00 -4.25575912e-01 -4.14653391e-01 1.03417635e+00
1.61129344e+00 -1.82023516e-03 4.53163326e-01 2.37263963e-02
6.44766688e-01 6.80219829e-01 3.78985964e-02 2.89189398e-01
1.90830454e-01 4.16630596e-01 3.64000767e-01 -2.30117306e-01
-3.25491846e-01 -4.69382495e-01 3.13360155e-01 4.02886868e-01
2.01641500e-01 1.83902368e-01 -5.90903282e-01 6.90039515e-01
-1.57930493e+00 -9.91772950e-01 6.58782661e-01 1.95761859e+00
8.84893477e-01 -1.11777370e-03 -8.96435827e-02 -2.87039001e-02
5.26099086e-01 3.11158508e-01 -4.92925674e-01 -4.21607256e-01
-4.57893685e-02 9.01822329e-01 2.38818496e-01 7.80857623e-01
-1.08213866e+00 1.27620232e+00 6.77309752e+00 4.71056163e-01
-1.25213408e+00 3.57263297e-01 7.43141294e-01 6.83632866e-02
-5.72173953e-01 2.36055315e-01 -7.76927829e-01 1.58728048e-01
8.27595174e-01 3.89524281e-01 7.22483158e-01 9.52402115e-01
-4.69294935e-01 6.53587133e-02 -1.32240641e+00 7.96924651e-01
4.94968742e-01 -1.65305984e+00 2.47575164e-01 2.71418784e-02
1.06247878e+00 1.66070908e-01 1.98665559e-01 2.17904627e-01
5.24797261e-01 -1.32878876e+00 7.21645117e-01 4.19271708e-01
9.37537253e-01 -3.99682432e-01 4.50187087e-01 2.02899560e-01
-9.62981582e-01 -1.35525107e-01 -6.34157181e-01 -1.85345605e-01
-3.82929407e-02 5.11258364e-01 -7.32914686e-01 -8.13754126e-02
2.90970802e-01 1.14126968e+00 -2.50519961e-01 6.96941912e-01
-3.85436207e-01 6.27936542e-01 -1.67154714e-01 5.49073040e-01
2.97379732e-01 2.60801073e-02 9.05703604e-02 1.39787984e+00
1.96755394e-01 2.57591039e-01 3.48382354e-01 1.00343573e+00
-3.15847874e-01 -1.20519497e-01 -1.02284348e+00 5.03587797e-02
2.61295557e-01 1.34684455e+00 -7.57946193e-01 -3.72356296e-01
-7.17880368e-01 1.12499356e+00 6.35417819e-01 6.02442265e-01
-3.97399276e-01 9.40633714e-02 5.43812931e-01 2.71903872e-01
9.37884271e-01 -1.90713510e-01 -5.08940637e-01 -1.27871788e+00
-3.45185429e-01 -8.98937285e-01 3.21231544e-01 -6.91353381e-01
-1.28691375e+00 6.16181731e-01 -3.12363982e-01 -9.55847383e-01
-2.45490193e-01 -5.58381617e-01 -4.23786938e-01 9.08827901e-01
-1.53508389e+00 -1.36449504e+00 -1.75093263e-01 7.84923494e-01
6.91684186e-01 -5.27677476e-01 1.20568120e+00 8.36572722e-02
-8.56238827e-02 3.02717179e-01 -2.35803306e-01 1.96307167e-01
4.03864473e-01 -1.00413942e+00 1.83026984e-01 6.88092530e-01
7.50812352e-01 5.47695577e-01 1.99144349e-01 -4.73174751e-01
-1.01455069e+00 -1.21157706e+00 8.18398356e-01 -3.34922314e-01
5.17882884e-01 -6.07196271e-01 -9.17781770e-01 1.00899708e+00
8.58003870e-02 6.13277793e-01 8.15365255e-01 1.54749274e-01
-7.25443780e-01 -5.70950732e-02 -1.03867567e+00 -2.98408996e-02
1.07068300e+00 -8.53532195e-01 -7.15091288e-01 4.64576215e-01
4.80053842e-01 -2.08638206e-01 -6.86554730e-01 2.30323911e-01
4.77452248e-01 -8.59350264e-01 9.91693735e-01 -6.93119645e-01
5.26989460e-01 -1.76586062e-01 -5.67069888e-01 -1.18562710e+00
-6.97524309e-01 -1.82800785e-01 8.89061242e-02 7.36410797e-01
3.78026992e-01 -4.84928697e-01 9.94175911e-01 8.35274383e-02
-3.46005619e-01 -8.50252271e-01 -6.98996425e-01 -2.35251606e-01
7.42268190e-02 -1.74445748e-01 1.71484068e-01 1.01183784e+00
-2.05563828e-01 5.20907938e-01 -1.84393063e-01 2.38730088e-01
6.70568645e-01 4.44430113e-01 3.41001570e-01 -1.19867682e+00
-6.03871047e-01 -2.10406825e-01 -5.22033095e-01 -1.21302521e+00
7.30503142e-01 -1.37842250e+00 1.40146568e-01 -1.38741493e+00
3.22191536e-01 -5.80540597e-01 -3.64674777e-01 9.92673814e-01
2.11179540e-01 6.25702024e-01 7.09305331e-02 3.75809163e-01
-4.90695834e-01 4.59018201e-01 1.09379935e+00 -2.75447518e-01
1.20788187e-01 -1.69963151e-01 -9.82142627e-01 7.68993855e-01
5.25987744e-01 -6.90908909e-01 -4.94741261e-01 -6.02653146e-01
-1.08951412e-01 -1.61772817e-02 6.03427172e-01 -9.16822553e-01
1.45034894e-01 4.54349890e-02 7.42806256e-01 5.43700866e-02
3.03698599e-01 -7.61890948e-01 6.94705099e-02 2.95533836e-01
-3.70730013e-01 -4.95156825e-01 9.51811969e-02 5.45551300e-01
-4.11467433e-01 -2.73116261e-01 1.16423547e+00 -3.63772541e-01
-6.96172774e-01 3.08141589e-01 -2.45225072e-01 -1.64918035e-01
8.33108246e-01 -2.81449378e-01 3.05030290e-02 -2.15976894e-01
-1.24046278e+00 -2.37947315e-01 5.82918286e-01 2.26090252e-01
7.27516413e-01 -1.25690937e+00 -5.03506958e-01 8.78122151e-01
-1.08828515e-01 6.43827394e-02 -9.83281583e-02 3.69077682e-01
-2.18010753e-01 3.73199075e-01 -4.29076374e-01 -6.82007551e-01
-5.80801308e-01 6.65657759e-01 5.12194574e-01 1.10409185e-01
-5.89862287e-01 1.10179889e+00 6.99742913e-01 -2.75470614e-01
1.47830397e-01 -5.75586073e-02 -1.42568156e-01 -4.84089963e-02
2.38473430e-01 -4.81214672e-01 8.97373706e-02 -8.60848248e-01
-3.18781018e-01 7.61305273e-01 6.01143464e-02 -6.37577176e-02
1.67531657e+00 1.76754639e-01 -1.28603041e-01 1.72989279e-01
1.64896250e+00 -5.37287071e-02 -1.51106691e+00 -3.12245280e-01
-1.39922053e-01 -3.41028661e-01 3.53059471e-02 -7.30905771e-01
-1.57853639e+00 8.45930815e-01 3.79608482e-01 -1.90935239e-01
1.23021936e+00 4.42539871e-01 5.87281764e-01 1.24475621e-01
3.68296236e-01 -5.54440141e-01 4.13843513e-01 5.07800400e-01
9.20322239e-01 -1.26059926e+00 -2.42198810e-01 -3.27773899e-01
-4.14251655e-01 1.00045025e+00 3.01235408e-01 -6.45081699e-01
1.00780857e+00 3.80550355e-01 -1.30307823e-01 -4.27908987e-01
-6.87084138e-01 -4.47595954e-01 3.15094233e-01 7.21588254e-01
5.72827399e-01 -7.51180649e-02 7.95177296e-02 4.36699718e-01
5.04325926e-02 7.95888454e-02 8.33868608e-02 8.73587489e-01
-8.28696370e-01 -1.22190964e+00 -1.57250047e-01 4.65739101e-01
-2.11395115e-01 -1.90178901e-01 -2.39925355e-01 2.65248805e-01
4.38784093e-01 4.58063096e-01 2.59238183e-01 -3.35196406e-01
5.28973266e-02 1.99903503e-01 5.84102750e-01 -1.47333324e+00
-3.57529283e-01 1.01273879e-01 -1.49368793e-01 -6.73765838e-01
-5.26040971e-01 -7.97565103e-01 -1.26687849e+00 1.54841274e-01
1.13071710e-01 1.54923424e-01 6.08880877e-01 9.49703157e-01
3.34865332e-01 2.68072724e-01 6.21701956e-01 -1.18171644e+00
-3.70911479e-01 -6.21436059e-01 -5.75092554e-01 4.67088997e-01
4.78302479e-01 -8.38096917e-01 -2.88581729e-01 6.13789380e-01] | [9.842270851135254, 1.4905693531036377] |
fbb55201-b020-4a25-ae34-32867d001e7d | image-inpainting-by-patch-propagation-using | null | null | https://ieeexplore.ieee.org/document/5404308 | https://ieeexplore.ieee.org/document/5404308 | Image Inpainting by Patch Propagation Using Patch Sparsity | This paper introduces a novel examplar-based inpainting algorithm through investigating the sparsity of natural image patches. Two novel concepts of sparsity at the patch level are proposed for modeling the patch priority and patch representation, which are two crucial steps for patch propagation in the examplar-based inpainting approach. First, patch structure sparsity is designed to measure the confidence of a patch located at the image structure (e.g., the edge or corner) by the sparseness of its nonzero similarities to the neighboring patches. The patch with larger structure sparsity will be assigned higher priority for further inpainting. Second, it is assumed that the patch to be filled can be represented by the sparse linear combination of candidate patches under the local patch consistency constraint in a framework of sparse representation. Compared with the traditional examplar-based inpainting approach, structure sparsity enables better discrimination of structure and texture, and the patch sparse representation forces the newly inpainted regions to be sharp and consistent with the surrounding textures. Experiments on synthetic and natural images show the advantages of the proposed approach. | ['Zongben Xu', 'Jian Sun'] | 2010-05-01 | null | null | null | journal-2010-5 | ['novel-concepts'] | ['reasoning'] | [ 3.41584355e-01 2.50225455e-01 -5.31851113e-01 7.10905939e-02
-5.91497779e-01 5.81799410e-02 1.01870216e-01 1.37697846e-01
4.42694217e-01 8.55165958e-01 2.55414784e-01 4.32131797e-01
-5.63401952e-02 -8.33211303e-01 -8.53954494e-01 -9.09280419e-01
1.46375328e-01 -2.75574550e-02 2.50253946e-01 -6.26566857e-02
5.12587905e-01 6.06645107e-01 -1.64068329e+00 7.45949149e-01
7.79672682e-01 1.20274210e+00 5.67245126e-01 1.30427569e-01
-2.45643720e-01 8.17359090e-01 -3.60029578e-01 3.67380917e-01
3.39858413e-01 -6.33223653e-01 -5.50799847e-01 6.71305418e-01
7.57099628e-01 -1.83324125e-02 -1.47019029e-01 1.27963662e+00
-1.07565604e-01 -7.46261422e-03 3.14174205e-01 -8.10847342e-01
-5.44646621e-01 1.54920280e-01 -1.12848663e+00 9.88721400e-02
4.67187911e-01 -3.45683068e-01 7.25600898e-01 -1.20318270e+00
8.54850352e-01 1.14827824e+00 5.68412185e-01 2.40980703e-02
-1.19231176e+00 -2.37659991e-01 2.86051333e-01 1.39642686e-01
-1.59082770e+00 -3.07326287e-01 1.38502336e+00 -2.40708396e-01
3.37262750e-01 6.49960279e-01 6.55119240e-01 3.66537660e-01
5.43901742e-01 6.67053938e-01 1.27755141e+00 -6.38861895e-01
3.39141071e-01 9.86659303e-02 5.46903424e-02 7.07416177e-01
1.57914564e-01 8.85035396e-02 -7.67030180e-01 -3.84773523e-01
1.35362601e+00 2.19071299e-01 -4.96413112e-01 -3.67665827e-01
-9.37985361e-01 6.78028464e-01 4.68866110e-01 5.40559828e-01
-8.56163740e-01 -9.83676910e-02 -1.20280787e-01 2.30547026e-01
7.41156816e-01 2.92221963e-01 -7.32363611e-02 3.29438955e-01
-1.35969293e+00 1.84085011e-01 3.58670354e-01 8.37854266e-01
1.23409867e+00 7.86240920e-02 -1.69622079e-01 1.03959811e+00
1.60245880e-01 3.03781956e-01 1.39696300e-01 -1.10636926e+00
2.01240361e-01 6.79039001e-01 1.18061498e-01 -1.67414963e+00
4.37602192e-01 -3.76759052e-01 -9.55976844e-01 1.92089602e-01
-1.46485552e-01 3.94175172e-01 -7.66320407e-01 1.21883738e+00
4.60212916e-01 5.00788093e-01 -1.68594867e-01 8.23131263e-01
7.96236575e-01 1.18376517e+00 -3.61277282e-01 -4.88506794e-01
1.19718337e+00 -9.85358238e-01 -7.60524333e-01 -2.24264383e-01
-1.95359364e-01 -1.11990643e+00 6.15611434e-01 3.49348575e-01
-1.25385714e+00 -8.45859706e-01 -9.31072593e-01 1.52488008e-01
3.19727838e-01 1.30427480e-01 2.26291344e-01 -4.42963615e-02
-8.08601737e-01 4.55556244e-01 -5.75526237e-01 -1.46961764e-01
2.00385869e-01 -2.64711440e-01 -5.02380669e-01 -5.75642169e-01
-6.71362698e-01 4.94012624e-01 9.38171148e-02 1.57599226e-01
-7.64208138e-01 -8.78275573e-01 -8.87969673e-01 2.69556761e-01
3.67133170e-02 -2.63874084e-01 4.46129441e-01 -1.42995119e+00
-8.34571540e-01 6.77595496e-01 -6.56401873e-01 -1.90219358e-01
1.33161545e-01 5.42547496e-04 -1.91392615e-01 6.09466434e-01
4.67786551e-01 6.67452812e-01 1.50811207e+00 -1.75217175e+00
-6.03587985e-01 -9.36598033e-02 -3.31770867e-01 2.88806856e-01
3.12020024e-03 -3.11292678e-01 -4.02109355e-01 -1.24011576e+00
7.94567049e-01 -4.67801124e-01 -2.09413946e-01 3.70884657e-01
-8.36778656e-02 2.05973431e-01 1.29286337e+00 -8.60111952e-01
1.23658836e+00 -2.65216064e+00 1.14347942e-01 6.37679935e-01
1.90391019e-01 -6.66374266e-02 -2.27635801e-01 5.71254790e-01
-2.99787551e-01 -3.85496795e-01 -4.20698255e-01 -1.48470486e-02
-6.57545447e-01 4.13607925e-01 -4.48132455e-01 4.44783151e-01
4.13526237e-01 3.46696138e-01 -7.31690168e-01 -7.08164275e-01
9.44586843e-02 5.52016675e-01 -4.95601475e-01 1.85157672e-01
-4.06628102e-02 4.98439997e-01 -4.66992885e-01 9.30222869e-01
1.04166758e+00 -1.87099390e-02 -4.68535870e-02 -5.78221917e-01
-3.84581417e-01 -1.12931587e-01 -1.47638464e+00 1.36339462e+00
-2.06982702e-01 4.60130930e-01 6.50154650e-01 -1.02870584e+00
1.33467245e+00 2.27701396e-01 5.85723042e-01 -5.86517751e-01
-4.54612970e-01 2.18505234e-01 -4.04370070e-01 -3.18525106e-01
6.30658805e-01 -9.48397219e-02 3.96816194e-01 2.28361756e-01
-2.82294869e-01 -7.95760006e-02 3.17440368e-02 4.34092246e-02
7.46738613e-01 -3.53027396e-02 2.29075298e-01 -7.18076468e-01
5.75067043e-01 9.47222263e-02 8.72781694e-01 4.83071268e-01
1.67493612e-01 9.66022670e-01 2.12856546e-01 -6.33254051e-01
-9.47749257e-01 -7.14748919e-01 -3.08806896e-01 4.47679728e-01
4.69041348e-01 -2.26812497e-01 -5.88150561e-01 -2.60033160e-01
-1.12032361e-01 1.39133692e-01 -9.32755351e-01 4.37330976e-02
-7.46678710e-01 -1.25922456e-01 -3.23270738e-01 1.68806896e-01
7.35987604e-01 -1.11123860e+00 -3.88030350e-01 3.59160990e-01
-4.83777910e-01 -5.58883846e-01 -6.89302981e-01 -2.87266560e-02
-1.03390026e+00 -1.07119107e+00 -1.03665280e+00 -1.37658501e+00
1.31940603e+00 8.11355054e-01 1.05206895e+00 5.58702350e-01
-2.52790511e-01 3.15380961e-01 -7.25883603e-01 3.25884938e-01
-4.16127443e-01 -3.32679063e-01 -4.02173549e-01 5.84965467e-01
-3.02133501e-01 -6.31396532e-01 -6.59336090e-01 4.60347146e-01
-9.84426737e-01 1.56281427e-01 6.76945627e-01 1.10808003e+00
1.42323422e+00 4.95799750e-01 2.08520189e-01 -7.34957278e-01
4.30743784e-01 -5.25655627e-01 -1.94012240e-01 3.20902467e-01
-2.41346359e-01 -1.83135539e-01 3.41308445e-01 -4.17273790e-01
-1.05918920e+00 7.87053555e-02 3.76802795e-02 -5.82385659e-01
-1.27366722e-01 7.40910769e-01 -1.15669318e-01 -3.95224750e-01
4.47097093e-01 7.31101394e-01 1.48420572e-01 -4.67350513e-01
-6.01084195e-02 1.78995907e-01 3.95721197e-01 -8.92937064e-01
5.95245838e-01 6.70210540e-01 5.50354645e-02 -1.22462094e+00
-6.18140340e-01 -3.72386128e-01 -2.89104462e-01 -2.95258462e-01
4.39940006e-01 -7.78377593e-01 1.59797683e-01 2.15000436e-01
-1.00371647e+00 6.31342903e-02 -7.13883400e-01 1.62059546e-01
-4.39378083e-01 8.14454913e-01 -4.39202577e-01 -6.97086692e-01
-1.93267643e-01 -9.33532417e-01 1.06725454e+00 1.43938780e-01
-2.40556225e-01 -8.12331319e-01 1.72419697e-02 3.73196155e-02
3.11707884e-01 4.08071458e-01 1.04054713e+00 4.11724061e-01
-6.94980621e-01 -1.69421911e-01 -1.13430560e-01 4.39284027e-01
4.52674031e-01 4.74880897e-02 -5.44132173e-01 -3.52719247e-01
5.05301535e-01 3.87462191e-02 7.37266243e-01 7.21238136e-01
9.14533317e-01 -5.79866767e-01 -4.64747906e-01 3.76993090e-01
1.55617237e+00 2.06177324e-01 1.08515644e+00 1.79954812e-01
4.09697592e-01 5.35921335e-01 1.13566267e+00 4.49214280e-01
-2.63285905e-01 5.59249818e-01 3.55790854e-01 -4.71393436e-01
-4.02388483e-01 -4.75487798e-01 2.34373227e-01 8.37290108e-01
1.74554899e-01 1.99698254e-01 -2.92236984e-01 6.55871153e-01
-1.81223953e+00 -1.09222007e+00 -1.96267635e-01 2.10339069e+00
7.93116510e-01 -1.11265182e-01 -2.70654291e-01 2.45519638e-01
1.07794082e+00 3.00960034e-01 -2.60065079e-01 -3.40691298e-01
-3.13233525e-01 3.18818450e-01 8.71489719e-02 8.50643516e-01
-7.98952878e-01 6.98685110e-01 6.46179199e+00 1.23695278e+00
-1.02306473e+00 -2.18863532e-01 9.84972715e-01 3.68156403e-01
-4.59276170e-01 3.22675645e-01 -6.86244786e-01 6.09585702e-01
-7.53619820e-02 2.17667818e-01 2.70368010e-01 7.33973682e-01
2.49413580e-01 -6.42632782e-01 -7.38499522e-01 9.29459572e-01
1.99582979e-01 -1.58889687e+00 3.23962510e-01 -1.03384607e-01
1.00477016e+00 -5.82600653e-01 1.99056387e-01 -3.76358092e-01
-4.83029932e-01 -7.73022711e-01 8.09988737e-01 5.93749702e-01
6.52184010e-01 -7.50796795e-01 5.27368188e-01 3.90492052e-01
-1.38305914e+00 2.42590547e-01 -7.38326490e-01 1.28686707e-02
7.52503797e-02 1.10792327e+00 -2.84295797e-01 5.34397483e-01
7.63597667e-01 9.75674450e-01 -2.29077980e-01 9.94302630e-01
-1.79196492e-01 5.78111529e-01 -2.67667145e-01 5.81275821e-01
1.98900029e-01 -4.36424702e-01 6.92179441e-01 8.13239336e-01
5.50605953e-01 3.00733119e-01 4.75842357e-01 1.13134396e+00
3.20648909e-01 2.15994641e-01 -5.14237404e-01 3.47594857e-01
5.78614771e-01 1.13357985e+00 -6.84658408e-01 -3.53955716e-01
-3.08362991e-01 9.09891605e-01 1.03913009e-01 2.89596409e-01
-4.32468593e-01 -1.81803271e-01 2.81019896e-01 7.04307437e-01
7.35438824e-01 -9.80839804e-02 -3.33689570e-01 -9.52065766e-01
2.79273897e-01 -1.04793453e+00 1.25707891e-02 -7.69473851e-01
-1.31009471e+00 7.38440514e-01 -1.11334585e-01 -1.67097223e+00
2.69545227e-01 5.79388160e-03 -8.29825878e-01 1.02786839e+00
-1.52295971e+00 -1.15148985e+00 -2.50979871e-01 7.33052254e-01
6.17624640e-01 3.03764399e-02 6.48495317e-01 5.79810962e-02
-1.90203577e-01 1.54830530e-01 1.02210045e-01 -3.06192786e-01
4.18224931e-01 -4.78188396e-01 -3.44577432e-01 1.03042722e+00
7.30910450e-02 5.58077574e-01 6.39284611e-01 -1.03635168e+00
-1.09926128e+00 -1.09665239e+00 8.73764932e-01 4.38113958e-01
8.59289616e-02 1.80800736e-01 -1.17723525e+00 1.63662001e-01
2.79385209e-01 8.08627009e-02 4.24279571e-01 -3.89327765e-01
-2.43782401e-01 -1.92831114e-01 -1.50506067e+00 2.92251647e-01
4.27777857e-01 -2.48748913e-01 -6.11571252e-01 2.67496705e-01
2.40827695e-01 -1.85869187e-01 -9.45556343e-01 4.25037801e-01
2.89451927e-01 -9.98053968e-01 8.52478027e-01 7.66215697e-02
6.90825105e-01 -6.40676022e-01 -3.37964594e-01 -1.08821487e+00
-8.35285068e-01 -4.14722502e-01 -6.46279827e-02 1.12382996e+00
1.02910712e-01 -3.47644240e-01 9.99109566e-01 1.87134799e-02
-2.42857397e-01 -8.59956682e-01 -1.00250340e+00 -6.85019076e-01
-2.97391504e-01 2.85894096e-01 3.80594552e-01 1.13017762e+00
-8.73917937e-02 -1.61845446e-01 -5.64568758e-01 2.65466630e-01
7.40858316e-01 6.96069181e-01 4.55976993e-01 -1.00386834e+00
-3.92964482e-01 -6.36670440e-02 -3.75375777e-01 -1.17029309e+00
-1.35044783e-01 -6.19692028e-01 1.88396543e-01 -1.47387719e+00
1.86002418e-01 -6.39880240e-01 -1.30984008e-01 4.10472691e-01
-1.34968877e-01 3.82171124e-01 -8.30941573e-02 6.78895593e-01
-1.14450596e-01 5.71407914e-01 1.52649868e+00 -4.22110677e-01
-1.40547737e-01 -2.71148771e-01 -5.64126372e-01 5.33221364e-01
6.11165941e-01 -4.70303684e-01 -3.10007125e-01 -2.80045927e-01
-1.07456982e-01 2.94646651e-01 2.86631107e-01 -9.81477201e-01
7.13739824e-03 -2.99597353e-01 4.31069672e-01 -6.77765310e-01
3.69137645e-01 -1.08501780e+00 5.27366519e-01 5.45804083e-01
-8.46960098e-02 -1.45956259e-02 1.47849053e-01 5.72917879e-01
-8.85381639e-01 -3.21113229e-01 1.22385097e+00 -3.74384314e-01
-7.02421367e-01 2.14541182e-01 -4.02421445e-01 -2.26382926e-01
9.77539003e-01 -7.39661694e-01 1.57563537e-01 -3.43417138e-01
-8.89840841e-01 -2.22684652e-01 6.28273964e-01 9.43829864e-03
1.18361056e+00 -1.52266622e+00 -8.49349976e-01 8.01541984e-01
3.48503143e-02 -7.34464824e-02 4.35487449e-01 7.01086581e-01
-7.51452088e-01 3.55856791e-02 -4.13058639e-01 -7.28815258e-01
-1.32113004e+00 4.38781887e-01 -2.48632468e-02 -2.61486769e-01
-7.68476009e-01 8.01859081e-01 5.05300164e-01 4.20233220e-01
1.04989938e-01 -2.85770714e-01 1.87564287e-02 -4.82842267e-01
5.87898314e-01 3.52847695e-01 -5.84337395e-03 -7.72900999e-01
-1.31799877e-01 8.45544100e-01 -2.06612706e-01 1.14185393e-01
1.38113415e+00 3.66657786e-02 -8.25007021e-01 6.38830140e-02
8.23422074e-01 3.81075680e-01 -1.43363345e+00 -1.05391391e-01
-5.09260476e-01 -9.38174486e-01 1.92456573e-01 -4.44406599e-01
-1.38220739e+00 4.40548927e-01 4.33953226e-01 1.73178494e-01
1.19109142e+00 1.60693040e-03 7.67484426e-01 -1.08665138e-01
6.04922235e-01 -8.77552807e-01 7.23593906e-02 1.23455480e-01
1.45642924e+00 -7.08166242e-01 2.95704573e-01 -1.00287962e+00
-2.30352223e-01 9.14773464e-01 4.93534863e-01 -6.96709275e-01
1.01186192e+00 2.41196036e-01 -2.81901509e-01 -2.94581115e-01
-3.57236803e-01 3.84560674e-01 3.09681714e-01 6.29690707e-01
3.03261250e-01 -8.76954868e-02 -5.67816138e-01 1.69311866e-01
4.15436268e-01 -1.51105285e-01 2.28525236e-01 1.12387347e+00
-8.06746185e-01 -1.14284003e+00 -9.12468672e-01 2.37224340e-01
-1.06590636e-01 -8.51666853e-02 -2.63118386e-01 4.45842803e-01
3.05891067e-01 7.35728025e-01 1.29361004e-01 -5.98533526e-02
1.40725702e-01 -4.37607497e-01 6.19827569e-01 -7.62717664e-01
-2.63181686e-01 5.02548993e-01 -3.57229203e-01 -5.64787388e-01
-5.45351565e-01 -5.56190193e-01 -9.13201392e-01 1.17356861e-02
-2.68120885e-01 4.58256423e-01 4.14390750e-02 4.11831915e-01
6.00062966e-01 1.33569896e-01 9.13009226e-01 -8.46450746e-01
5.93785085e-02 -7.36844540e-01 -1.05803835e+00 3.04630280e-01
9.06962827e-02 -6.37852073e-01 -3.37644488e-01 2.82340944e-01] | [11.075202941894531, -2.0358738899230957] |
0e1aae63-b750-488a-a35c-f64373f74b33 | refining-low-resource-unsupervised | 2205.15544 | null | https://arxiv.org/abs/2205.15544v3 | https://arxiv.org/pdf/2205.15544v3.pdf | Refining Low-Resource Unsupervised Translation by Language Disentanglement of Multilingual Model | Numerous recent work on unsupervised machine translation (UMT) implies that competent unsupervised translations of low-resource and unrelated languages, such as Nepali or Sinhala, are only possible if the model is trained in a massive multilingual environment, where these low-resource languages are mixed with high-resource counterparts. Nonetheless, while the high-resource languages greatly help kick-start the target low-resource translation tasks, the language discrepancy between them may hinder their further improvement. In this work, we propose a simple refinement procedure to separate languages from a pre-trained multilingual UMT model for it to focus on only the target low-resource task. Our method achieves the state of the art in the fully unsupervised translation tasks of English to Nepali, Sinhala, Gujarati, Latvian, Estonian and Kazakh, with BLEU score gains of 3.5, 3.5, 3.3, 4.1, 4.2, and 3.3, respectively. Our codebase is available at https://github.com/nxphi47/refine_unsup_multilingual_mt | ['Ai Ti Aw', 'Wu Kui', 'Shafiq Joty', 'Xuan-Phi Nguyen'] | 2022-05-31 | null | null | null | null | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [-1.51951492e-01 -1.14354484e-01 -3.88935208e-01 -2.72648931e-01
-1.57107496e+00 -9.90433753e-01 7.40708649e-01 -1.57618821e-01
-5.20725727e-01 1.39268756e+00 2.89815575e-01 -9.32455897e-01
3.48556787e-01 -4.56759274e-01 -6.41825080e-01 -5.33098400e-01
5.09053588e-01 1.02992845e+00 -2.57232130e-01 -5.16273916e-01
-1.42622054e-01 1.91694528e-01 -6.59595370e-01 1.58040807e-01
1.44983125e+00 -6.50221575e-03 2.68059134e-01 4.24260557e-01
-1.66025549e-01 4.10757571e-01 -2.66873091e-01 -7.44350970e-01
3.37465614e-01 -6.46145105e-01 -9.76598263e-01 -3.19469213e-01
2.31769696e-01 -1.50958002e-01 5.10781584e-03 9.58347738e-01
5.43828130e-01 -3.31728429e-01 6.27739131e-01 -7.70419538e-01
-7.06447244e-01 1.09846723e+00 -6.14465117e-01 -8.09192061e-02
1.88108504e-01 -2.03334421e-01 9.65791106e-01 -1.41991341e+00
8.92761350e-01 9.92650270e-01 4.07506496e-01 4.21802461e-01
-9.95386183e-01 -7.23686278e-01 -1.71402514e-01 -4.45907228e-02
-1.58639634e+00 -8.29629540e-01 3.46433580e-01 -2.88101971e-01
1.12271035e+00 2.49916434e-01 1.90241352e-01 9.82757032e-01
2.20800176e-01 5.47022939e-01 1.47565043e+00 -7.19356954e-01
-2.97073305e-01 3.87914598e-01 -4.65495706e-01 5.32108784e-01
3.43982965e-01 -3.39510083e-01 -4.95123237e-01 -7.49994814e-02
6.46803916e-01 -2.95004904e-01 -2.90447250e-02 1.42286003e-01
-1.54850590e+00 7.09563017e-01 -6.98958263e-02 6.13449574e-01
-3.92232835e-01 -4.35547471e-01 3.40103149e-01 6.64694846e-01
9.11246181e-01 2.09800586e-01 -8.41704726e-01 -3.58988196e-01
-9.68573630e-01 -1.60915181e-01 6.60789251e-01 1.38085723e+00
1.00386631e+00 1.22647777e-01 3.00597906e-01 1.20876002e+00
1.25407159e-01 1.04320467e+00 3.16145927e-01 -4.25354302e-01
1.04705596e+00 2.13512450e-01 1.67478234e-01 -1.80470333e-01
-6.31096140e-02 -4.63302374e-01 -8.81362379e-01 -3.71864974e-01
4.35198098e-01 -3.97632957e-01 -9.76360202e-01 1.66900778e+00
2.40979716e-01 -5.86676002e-01 5.44770360e-01 9.14253831e-01
5.30283570e-01 8.21210921e-01 -1.31689742e-01 -4.79966074e-01
1.17410922e+00 -1.36662710e+00 -6.67568624e-01 -4.57089931e-01
8.81788313e-01 -1.59338558e+00 1.22110510e+00 -4.58952226e-02
-1.22917485e+00 -3.34811330e-01 -7.26380885e-01 -1.78437695e-01
-2.43572921e-01 3.98774713e-01 4.92667884e-01 4.23569709e-01
-1.19588923e+00 2.77922601e-01 -8.72538924e-01 -6.81687236e-01
3.23766693e-02 2.45597541e-01 -6.28604770e-01 -4.40926462e-01
-1.26626706e+00 1.18925560e+00 3.44087273e-01 -2.80636568e-02
-7.25859046e-01 -1.80322289e-01 -6.68329418e-01 -5.45930803e-01
9.62526202e-02 -3.58224422e-01 1.09813011e+00 -9.90004063e-01
-1.60016763e+00 1.02423716e+00 -3.50210309e-01 4.88712033e-03
6.94942892e-01 -3.80055994e-01 -5.16316772e-01 -1.79137856e-01
6.14364862e-01 4.37656850e-01 2.71626294e-01 -9.53561306e-01
-8.18309665e-01 -3.08951408e-01 -1.88985139e-01 5.87190032e-01
-9.69574526e-02 7.15528846e-01 -7.78873205e-01 -6.79744899e-01
2.90287405e-01 -1.18179977e+00 -2.03665480e-01 -8.77728641e-01
-3.13621104e-01 1.35669485e-01 2.15554982e-01 -1.32913363e+00
1.08684647e+00 -1.84198177e+00 2.89220154e-01 -9.32602361e-02
-3.22116524e-01 4.93346974e-02 -3.12780261e-01 7.80215561e-01
3.28420967e-01 2.14464381e-01 -3.52820426e-01 -1.88363269e-01
-1.98194668e-01 4.02942061e-01 -1.63347591e-02 4.34534669e-01
3.30058604e-01 1.02944839e+00 -9.79958653e-01 -5.19449174e-01
-1.81995496e-01 3.27907026e-01 -1.08021967e-01 -1.66612654e-03
1.97529793e-01 9.58153307e-01 -3.11575502e-01 9.40646589e-01
6.66631639e-01 1.74477711e-01 5.33009589e-01 2.98692703e-01
-5.15876949e-01 8.43550444e-01 -6.25566304e-01 1.85932267e+00
-7.32898235e-01 4.23759818e-01 9.45290923e-02 -5.08460164e-01
8.64388227e-01 5.86665332e-01 2.18966424e-01 -8.05078328e-01
-9.63190049e-02 1.06095982e+00 3.82728636e-01 -1.17912628e-01
7.05722570e-01 -2.65494913e-01 -3.43625128e-01 6.97548687e-01
2.12464556e-01 -6.11069947e-02 4.29661244e-01 1.32656038e-01
6.57636642e-01 3.61732483e-01 3.10825557e-01 -5.91082990e-01
4.33621585e-01 2.93951064e-01 7.36366987e-01 4.36852515e-01
-5.37058935e-02 6.37549341e-01 -7.47973099e-02 -9.96747315e-02
-1.52103162e+00 -1.12397206e+00 -1.25090212e-01 1.23537862e+00
-3.32143396e-01 -4.40144718e-01 -6.29449725e-01 -7.44535625e-01
-5.64469159e-01 5.61115146e-01 7.52650276e-02 3.26672018e-01
-1.09962249e+00 -8.61774206e-01 7.15602517e-01 1.14414759e-01
4.56619442e-01 -9.10935998e-01 2.69002855e-01 3.47796232e-01
-9.05579805e-01 -1.22733867e+00 -6.67842448e-01 1.82754010e-01
-9.92981970e-01 -3.92648727e-01 -1.04811716e+00 -1.13626528e+00
8.20669591e-01 3.04013669e-01 1.12784529e+00 -2.31786683e-01
6.20509863e-01 -2.63446897e-01 -5.28095365e-01 -1.93552509e-01
-7.27165997e-01 6.75973952e-01 4.57230985e-01 -3.70190859e-01
4.41338092e-01 -4.84280050e-01 -1.32717192e-01 3.67867708e-01
-5.62319934e-01 4.45483774e-01 7.85585880e-01 7.73640811e-01
6.16437912e-01 -4.33935523e-01 6.32600665e-01 -1.03853953e+00
4.10635859e-01 -4.17785048e-01 -3.40742499e-01 4.58007425e-01
-5.79315960e-01 -1.71237662e-02 8.09298694e-01 -3.28137517e-01
-1.11804867e+00 -4.88196388e-02 -2.55715162e-01 2.75120050e-01
1.71146635e-02 8.33496928e-01 -3.30741197e-01 1.90733939e-01
5.90027452e-01 4.21838701e-01 -5.14486194e-01 -6.13601148e-01
4.90255684e-01 1.08116198e+00 3.40893626e-01 -7.12450385e-01
1.14196169e+00 1.06838673e-01 -6.25665605e-01 -4.81046915e-01
-5.18087149e-01 -3.97944540e-01 -1.03208268e+00 1.01496138e-01
5.15274823e-01 -1.39993548e+00 4.87312406e-01 3.93892884e-01
-1.14460909e+00 -4.49382395e-01 6.84726760e-02 8.55346501e-01
-3.56806248e-01 1.72363639e-01 -1.02322626e+00 -4.59437877e-01
-6.94221914e-01 -1.27591980e+00 8.90795171e-01 1.17533039e-02
-2.82091588e-01 -1.02414858e+00 2.52661169e-01 5.94333112e-01
4.53401446e-01 -2.23817855e-01 1.02017546e+00 -6.78547442e-01
-3.33894968e-01 9.07169059e-02 -2.14101925e-01 3.40203315e-01
5.89530528e-01 -2.65601426e-01 -5.68783462e-01 -5.38316548e-01
-1.67122737e-01 -4.03489679e-01 2.84486145e-01 -1.26545921e-01
-7.26395622e-02 -3.84334505e-01 2.61921529e-02 4.74564880e-01
1.32720947e+00 1.32562503e-01 4.43032920e-01 3.96237165e-01
7.34519303e-01 4.54538286e-01 8.19758594e-01 -2.00015102e-02
7.22067058e-01 5.69708645e-01 -3.64658237e-01 -3.61483485e-01
-9.72678587e-02 -3.12590688e-01 8.49875987e-01 2.07608747e+00
-5.09575486e-01 1.13787472e-01 -1.25152481e+00 8.36118639e-01
-1.71614945e+00 -3.41120601e-01 -2.57247657e-01 2.41723084e+00
1.48084414e+00 -8.92930478e-02 -1.04260668e-01 -5.39687932e-01
6.39195561e-01 -5.70699312e-02 -2.09503138e-04 -6.48024976e-01
-4.02303666e-01 3.89639109e-01 6.61492288e-01 8.77830267e-01
-8.53957653e-01 1.83149505e+00 5.03336143e+00 8.46164644e-01
-1.25596499e+00 5.93401968e-01 4.76806432e-01 2.49725133e-02
-4.54056054e-01 3.81386667e-01 -7.65052378e-01 2.40062520e-01
1.42770493e+00 -4.19601321e-01 7.54679799e-01 4.57287490e-01
4.11777616e-01 2.40340501e-01 -9.16791916e-01 7.11228490e-01
-1.52129337e-01 -9.75338101e-01 1.08119242e-01 1.75779626e-01
1.23418140e+00 7.50746906e-01 -6.94603100e-02 4.87675607e-01
5.03859758e-01 -8.38515103e-01 6.69309556e-01 -1.25677124e-01
1.26699483e+00 -8.98549020e-01 8.08024108e-01 5.54456651e-01
-9.91932034e-01 7.07594991e-01 -5.53875208e-01 1.62735451e-02
3.06424350e-01 5.06811082e-01 -9.03875172e-01 1.01293063e+00
4.28272247e-01 4.23056096e-01 -3.99382561e-01 3.83839041e-01
-6.55664563e-01 9.29372787e-01 -2.81776935e-01 1.91966414e-01
5.20050943e-01 -5.33100605e-01 4.20899957e-01 1.42511570e+00
6.31237924e-01 -1.19849890e-01 2.80975163e-01 2.78150260e-01
-3.37994933e-01 8.70194793e-01 -6.53035760e-01 -2.01331124e-01
3.66049767e-01 1.20590079e+00 -6.03648484e-01 -4.39749211e-01
-5.42652309e-01 1.36671340e+00 5.57724118e-01 5.00808299e-01
-7.80939758e-01 -3.82110685e-01 5.70765615e-01 -1.31098583e-01
-1.56341996e-02 -5.91242969e-01 -2.61326909e-01 -1.58779538e+00
3.29330891e-01 -1.37811172e+00 1.35178819e-01 -4.54369575e-01
-1.09690082e+00 9.36886370e-01 -3.60932529e-01 -1.35814202e+00
-5.48456490e-01 -3.61310571e-01 -2.67851651e-01 1.46049809e+00
-1.48850465e+00 -1.55515635e+00 6.06938481e-01 4.82333153e-01
7.53169119e-01 -3.72522682e-01 1.08371973e+00 6.77525103e-01
-4.82464164e-01 7.88388371e-01 6.04973912e-01 2.97295004e-01
1.11953318e+00 -8.83906126e-01 7.97621012e-01 1.19687629e+00
4.36598659e-01 8.19974244e-01 4.04696167e-01 -8.37447464e-01
-1.43842089e+00 -1.18869793e+00 1.91537726e+00 -5.74147701e-01
9.14173782e-01 -6.39809489e-01 -6.75538003e-01 9.04551923e-01
5.13661444e-01 -4.68015611e-01 7.74465024e-01 4.20641571e-01
-3.22938025e-01 2.22573638e-01 -7.82531798e-01 9.54720140e-01
9.27425742e-01 -7.58244216e-01 -2.57244378e-01 5.54453909e-01
5.87229490e-01 -3.47356200e-01 -1.06293416e+00 1.94652662e-01
5.36019862e-01 -1.29495904e-01 5.69180906e-01 -6.74382031e-01
5.97727656e-01 -3.38582009e-01 -4.23227489e-01 -1.53284585e+00
-3.53894770e-01 -8.04826856e-01 3.94624174e-01 1.33875358e+00
1.19193149e+00 -7.61616945e-01 4.11112517e-01 1.31947190e-01
-1.81115806e-01 -2.91337460e-01 -9.42437351e-01 -9.17741060e-01
6.68940485e-01 -2.69654989e-01 3.99081439e-01 1.30965972e+00
1.16881862e-01 8.97859514e-01 -7.04760671e-01 1.42220557e-01
4.57626253e-01 1.58434495e-01 8.11545193e-01 -7.57172823e-01
-2.41908267e-01 -8.20595101e-02 2.65552342e-01 -8.80385339e-01
4.77819033e-02 -1.32306969e+00 1.01289399e-01 -1.64383686e+00
4.19662118e-01 -6.77702665e-01 -3.27784806e-01 7.10513949e-01
-2.74922162e-01 6.09402478e-01 1.61857665e-01 6.39642119e-01
-3.26551855e-01 2.93331534e-01 1.21052086e+00 -1.24659665e-01
-3.21444273e-01 -1.50639728e-01 -8.24075043e-01 4.50896204e-01
1.25859284e+00 -8.07530105e-01 -1.02312237e-01 -1.17057538e+00
1.97725162e-01 -1.20720593e-02 -5.53736150e-01 -4.18647557e-01
1.11427277e-01 -4.30743247e-01 7.04077333e-02 -4.66794670e-01
-5.97150885e-02 -4.72068399e-01 2.58700132e-01 2.15481371e-01
3.81552875e-02 4.05844659e-01 2.27800146e-01 -3.06422591e-01
-3.31591427e-01 -1.06204294e-01 5.41711152e-01 -3.17849666e-01
-3.76861960e-01 2.84099370e-01 -4.88014966e-01 1.21654727e-01
3.40160817e-01 1.14256494e-01 -2.62261629e-01 -2.98829108e-01
-4.42012995e-01 1.60874665e-01 6.15414977e-01 5.19296527e-01
2.76416004e-01 -1.51333690e+00 -1.56770372e+00 1.66047528e-01
2.90928870e-01 -2.10185871e-01 -1.42988086e-01 1.11811066e+00
-6.91037416e-01 5.82159579e-01 -3.62915158e-01 -3.71776193e-01
-1.04638875e+00 1.44725114e-01 -1.57237083e-01 -6.99535131e-01
-2.53278136e-01 3.98849607e-01 6.71095774e-03 -1.13868666e+00
-3.61954927e-01 1.20845772e-01 2.54988432e-01 -3.01615894e-01
1.75939038e-01 2.39303820e-02 2.58655667e-01 -1.18034184e+00
-4.92480308e-01 4.25039977e-01 -3.81789625e-01 -7.23972976e-01
1.15074193e+00 -3.99656415e-01 -5.52608430e-01 5.11211276e-01
1.04219401e+00 6.99191272e-01 -4.80298787e-01 -4.14476514e-01
2.09016591e-01 -2.18348369e-01 -3.37134778e-01 -1.00638461e+00
-6.62023723e-01 7.49773085e-01 5.75990826e-02 -5.14769614e-01
1.07379806e+00 9.16069821e-02 8.19925725e-01 4.28696692e-01
8.68910491e-01 -1.22888386e+00 -7.12933660e-01 1.12031603e+00
7.14855075e-01 -1.27640283e+00 -2.18126342e-01 -2.34423950e-01
-6.20166004e-01 7.72878289e-01 2.68126458e-01 2.87315816e-01
1.06503315e-01 2.18245924e-01 6.17995322e-01 3.91642064e-01
-6.50956094e-01 -2.75316447e-01 1.46336824e-01 3.86319727e-01
1.02756691e+00 5.02918422e-01 -8.91628146e-01 4.01328802e-01
-4.46553260e-01 -3.29677969e-01 5.06825268e-01 9.52516437e-01
-1.76858574e-01 -1.74035656e+00 -3.15308183e-01 4.23529707e-02
-9.12725687e-01 -8.13784003e-01 -4.98300105e-01 8.84484529e-01
-8.09647292e-02 9.73096609e-01 -3.15623641e-01 -3.56192499e-01
5.91513962e-02 2.53388017e-01 4.91559565e-01 -7.57081628e-01
-5.86571276e-01 7.20907271e-01 5.30085504e-01 1.55436955e-02
-3.25004369e-01 -5.31540453e-01 -1.05486393e+00 -5.10160565e-01
-2.48347700e-01 5.40261507e-01 7.02902973e-01 8.25109839e-01
1.43538088e-01 -1.91522539e-01 6.61642790e-01 -4.98963892e-01
-1.70070454e-01 -1.32372928e+00 -1.89304799e-01 7.49913827e-02
-7.45964721e-02 1.53204218e-01 -1.39954593e-02 2.18780771e-01] | [11.619808197021484, 10.422966957092285] |
401b24df-11b7-4e48-98bf-08cedf5f41df | method-to-classify-skin-lesions-using | 2008.09418 | null | https://arxiv.org/abs/2008.09418v2 | https://arxiv.org/pdf/2008.09418v2.pdf | Method to Classify Skin Lesions using Dermoscopic images | Skin cancer is the most common cancer in the existing world constituting one-third of the cancer cases. Benign skin cancers are not fatal, can be cured with proper medication. But it is not the same as the malignant skin cancers. In the case of malignant melanoma, in its peak stage, the maximum life expectancy is less than or equal to 5 years. But, it can be cured if detected in early stages. Though there are numerous clinical procedures, the accuracy of diagnosis falls between 49% to 81% and is time-consuming. So, dermoscopy has been brought into the picture. It helped in increasing the accuracy of diagnosis but could not demolish the error-prone behaviour. A quick and less error-prone solution is needed to diagnose this majorly growing skin cancer. This project deals with the usage of deep learning in skin lesion classification. In this project, an automated model for skin lesion classification using dermoscopic images has been developed with CNN(Convolution Neural Networks) as a training model. Convolution neural networks are known for capturing features of an image. So, they are preferred in analyzing medical images to find the characteristics that drive the model towards success. Techniques like data augmentation for tackling class imbalance, segmentation for focusing on the region of interest and 10-fold cross-validation to make the model robust have been brought into the picture. This project also includes usage of certain preprocessing techniques like brightening the images using piece-wise linear transformation function, grayscale conversion of the image, resize the image. This project throws a set of valuable insights on how the accuracy of the model hikes with the bringing of new input strategies, preprocessing techniques. The best accuracy this model could achieve is 0.886. | ['Umarani Jayaraman', 'Subin Sahayam', 'Hemanth Nadipineni', 'Dusa Sai Charan'] | 2020-08-21 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 3.15975517e-01 2.04369858e-01 -2.34541267e-01 -1.39930427e-01
8.61010030e-02 -2.04359114e-01 3.75868618e-01 2.62397021e-01
-4.87697303e-01 5.79978585e-01 -1.44384384e-01 -3.98554236e-01
-2.71683484e-01 -9.32134151e-01 -2.21764907e-01 -9.18128967e-01
2.28907749e-01 1.26541048e-01 2.06568971e-01 -2.78782606e-01
3.06615442e-01 5.75350940e-01 -1.51395953e+00 3.48244101e-01
9.43001568e-01 6.77725434e-01 1.03238501e-01 7.42069125e-01
-4.79606003e-01 5.78761160e-01 -3.62137288e-01 -2.09144905e-01
2.17784837e-01 -4.19743687e-01 -9.49041247e-01 1.17189616e-01
1.58186927e-01 -1.68395028e-01 7.48389661e-02 1.05199015e+00
3.86521965e-01 -3.67087275e-01 8.37344706e-01 -9.72388864e-01
-3.56157243e-01 2.00074404e-01 -7.16844440e-01 2.20920667e-01
2.05335855e-01 7.84934238e-02 8.28204304e-02 -2.73676246e-01
5.63014328e-01 8.10446739e-01 7.69483030e-01 6.40548885e-01
-6.27929866e-01 -3.04996908e-01 -4.11117733e-01 1.26791432e-01
-1.12240887e+00 1.70870408e-01 3.64082277e-01 -3.41066599e-01
8.10130835e-01 6.39242649e-01 9.12536263e-01 6.99655235e-01
4.63009447e-01 2.66242683e-01 1.49953997e+00 -6.26694143e-01
-2.74551492e-02 6.82992578e-01 1.22890756e-01 5.46351969e-01
3.30625236e-01 -5.53698577e-02 8.48909542e-02 3.33236605e-01
7.76811719e-01 2.80953407e-01 -8.82527828e-02 2.83277124e-01
-6.39654815e-01 6.62984490e-01 5.27208090e-01 8.99744034e-01
-4.09361124e-01 -8.27978998e-02 5.61257362e-01 3.58397871e-01
2.03658968e-01 2.26097286e-01 -3.37884039e-01 -6.91106729e-03
-7.88890243e-01 -9.69573930e-02 5.82050443e-01 7.39481822e-02
2.71416008e-01 -1.71505168e-01 2.91521624e-02 6.09254479e-01
4.98858280e-02 2.27524504e-01 9.04068768e-01 -1.98641330e-01
-2.17605457e-01 1.23717713e+00 -2.62653291e-01 -8.13675642e-01
-6.84151113e-01 -3.72673303e-01 -9.88822520e-01 6.72626376e-01
6.98392510e-01 -1.62195072e-01 -1.55185235e+00 1.17303455e+00
4.51021969e-01 -2.40198493e-01 -1.38705969e-01 8.74924242e-01
6.84621274e-01 5.24603963e-01 3.16188037e-01 -3.11084446e-02
1.50958169e+00 -5.56124866e-01 -6.47635520e-01 2.33394474e-01
4.62539554e-01 -8.51802051e-01 8.12095582e-01 6.26340866e-01
-8.27980220e-01 -4.53855008e-01 -9.99560297e-01 1.39136016e-01
-8.45135033e-01 2.43418336e-01 8.44729483e-01 1.10001087e+00
-8.65671694e-01 6.27768159e-01 -7.90643811e-01 -1.11553037e+00
4.94173855e-01 4.77639288e-01 -6.51933432e-01 -9.48533714e-02
-8.63516331e-01 1.26628149e+00 3.46661657e-01 -1.98010802e-02
-5.86215615e-01 -6.17659986e-01 -2.81358838e-01 -2.04203710e-01
1.26534849e-01 -4.76016253e-01 6.79545283e-01 -1.57872617e+00
-1.18059027e+00 1.00958765e+00 1.40568003e-01 -3.39176565e-01
6.30607486e-01 1.32947266e-01 -3.98674637e-01 1.31982669e-01
-3.60443890e-01 4.56049085e-01 5.20939887e-01 -8.70898306e-01
-7.74426520e-01 -4.13145483e-01 -3.38577665e-02 1.59215257e-01
-5.07745564e-01 7.23498389e-02 -7.14231133e-02 -3.30685794e-01
4.04867791e-02 -7.19098091e-01 -3.12959999e-01 5.40329926e-02
-4.35932428e-01 -1.17055029e-01 9.32105601e-01 -9.07181740e-01
1.03870666e+00 -1.90790641e+00 -2.60523379e-01 3.42859596e-01
4.38374281e-02 6.97367013e-01 3.18990856e-01 6.18215799e-01
-2.61679083e-01 4.80696142e-01 -1.52237535e-01 2.56900996e-01
-4.51643497e-01 1.14185914e-01 5.33060133e-01 4.98472333e-01
3.59041661e-01 3.89409572e-01 -3.89907062e-01 -3.48748326e-01
4.87953633e-01 7.20210910e-01 1.42778158e-01 -1.55992240e-01
4.62368280e-02 2.10084230e-01 -2.12604746e-01 7.46972680e-01
8.81647289e-01 -1.59642622e-02 -8.52462947e-02 -3.10405165e-01
-1.05625346e-01 -3.16681981e-01 -1.16267979e+00 1.22583401e+00
-2.79422760e-01 6.61602616e-01 -8.93379301e-02 -1.10304677e+00
7.48593748e-01 4.71819371e-01 4.36595082e-01 -6.24280453e-01
4.81043369e-01 1.60828188e-01 3.29173684e-01 -1.18137085e+00
-3.58595625e-02 -3.34497660e-01 6.34218931e-01 8.95521864e-02
-2.42247775e-01 1.65396705e-01 3.40524465e-01 -1.47419780e-01
8.88734937e-01 -1.17457926e-01 4.02093023e-01 -2.04337686e-01
6.51600242e-01 3.32433254e-01 2.17930451e-01 9.96611342e-02
-2.07907185e-01 4.50083405e-01 5.32766521e-01 -6.20413482e-01
-1.22790480e+00 -6.96585417e-01 -4.99796689e-01 3.69734496e-01
-1.14516929e-01 4.09542918e-01 -1.09598720e+00 -4.55778092e-01
-1.47170529e-01 1.50281340e-01 -8.54720533e-01 -7.84866959e-02
-4.21955204e-03 -1.08278453e+00 6.43678844e-01 2.06281886e-01
8.81958365e-01 -9.68068182e-01 -7.08244085e-01 -7.05097467e-02
5.39376616e-01 -5.48245251e-01 3.41134369e-01 2.53279775e-01
-1.00212777e+00 -1.48056209e+00 -9.22450960e-01 -8.92708421e-01
1.02073300e+00 1.08561195e-01 6.37633145e-01 5.02208591e-01
-1.10151374e+00 -1.00559399e-01 -4.00131226e-01 -6.87613368e-01
-5.38565278e-01 6.28528520e-02 -2.05281243e-01 -1.36781380e-01
6.66316330e-01 -2.96961784e-01 -5.90097368e-01 -6.60125390e-02
-1.14216244e+00 -7.31785521e-02 9.86358285e-01 7.42835045e-01
2.49265939e-01 4.99963790e-01 5.37246466e-01 -1.16888893e+00
7.34441280e-01 -4.70465302e-01 -9.11045149e-02 2.84925431e-01
-6.91981196e-01 -3.37511331e-01 6.37967944e-01 -2.85028815e-01
-9.91580486e-01 1.06683047e-02 -2.59171158e-01 2.69873273e-02
-5.97487509e-01 4.97343212e-01 3.75720531e-01 -2.46419922e-01
9.35482502e-01 -3.30279246e-02 4.21776861e-01 -2.09499791e-01
-1.81332573e-01 8.48519087e-01 7.12331682e-02 2.62928277e-01
4.96619910e-01 4.66371924e-01 2.33366147e-01 -9.92934465e-01
-3.73982191e-01 -4.56078291e-01 -4.57242578e-01 -3.34352106e-01
1.05639660e+00 -4.69258755e-01 -6.02733135e-01 8.23415279e-01
-6.90500915e-01 -1.29206916e-02 -3.52542251e-02 3.58457595e-01
1.38179645e-01 1.46727160e-01 -5.28453231e-01 -1.14785731e+00
-5.33577263e-01 -8.57963622e-01 2.73820728e-01 9.56852198e-01
-1.13698304e-01 -1.17910779e+00 -6.05905205e-02 4.19250607e-01
6.43010437e-01 5.96135080e-01 9.56757307e-01 -5.41138709e-01
-1.38790116e-01 -5.78624845e-01 -3.20580572e-01 5.36132812e-01
3.29294145e-01 5.50912201e-01 -1.06740952e+00 -4.36422676e-02
-1.24277815e-01 -1.75097063e-01 7.92557001e-01 5.23208797e-01
1.08422124e+00 -2.96101570e-02 -4.04420078e-01 4.25642073e-01
1.91101360e+00 6.45099401e-01 9.73466277e-01 3.95016849e-01
3.34404171e-01 8.06765318e-01 5.10980606e-01 -8.20440520e-03
-7.35154524e-02 -2.12587789e-02 6.68446481e-01 -5.91413140e-01
-3.56521815e-01 8.23134705e-02 -2.29727384e-02 3.74312311e-01
-4.81617004e-01 -8.92728753e-03 -8.92904699e-01 5.74976325e-01
-1.12420988e+00 -9.01788473e-01 -6.12648904e-01 2.27164650e+00
6.28077626e-01 2.00910613e-01 2.59326160e-01 6.93721294e-01
6.61886871e-01 -5.24375677e-01 -3.21108192e-01 -8.41148436e-01
-2.19953991e-02 5.09474576e-01 6.96606338e-01 3.09158832e-01
-1.10363925e+00 4.88159448e-01 5.63687944e+00 7.15942144e-01
-1.80242848e+00 -1.06900722e-01 9.61770773e-01 1.80245563e-01
9.66010392e-02 -2.24225938e-01 -4.60397750e-01 5.89493215e-01
7.63555765e-01 2.19872400e-01 1.34092242e-01 6.90096378e-01
3.71512592e-01 -8.10900271e-01 -5.03901362e-01 7.26619124e-01
4.99354303e-02 -1.15695202e+00 -2.22183093e-01 -5.26962150e-03
6.12914979e-01 -3.09194952e-01 1.61851242e-01 -1.55965641e-01
-3.16985905e-01 -1.56376863e+00 -2.13570997e-01 7.11294651e-01
8.15652251e-01 -9.26927567e-01 1.31438768e+00 3.17764193e-01
-5.00999033e-01 -1.84904858e-01 -3.50469977e-01 -1.88251138e-01
-3.32320541e-01 5.89848340e-01 -1.21505129e+00 4.31958735e-01
8.56398106e-01 1.59818962e-01 -7.00217068e-01 1.34271312e+00
1.43625796e-01 5.76001942e-01 -3.04406911e-01 -4.77480471e-01
2.35927910e-01 -2.61995763e-01 1.53501555e-01 1.04277217e+00
3.68870437e-01 -8.93288329e-02 -3.63867372e-01 5.56247652e-01
4.05754179e-01 4.97238159e-01 -6.92282677e-01 -8.12511742e-02
5.39821908e-02 1.54262710e+00 -9.87610579e-01 -2.36406978e-02
-1.22824095e-01 8.37286949e-01 -2.59154707e-01 1.77555960e-02
-5.18467247e-01 -7.04259336e-01 1.87341094e-01 5.88370204e-01
-3.27264100e-01 4.86409694e-01 -5.45288742e-01 -4.48513001e-01
-1.13122284e-01 -7.82599807e-01 2.03351498e-01 -5.52714407e-01
-1.07209408e+00 3.99299413e-01 -2.47626379e-01 -8.28412354e-01
1.29976034e-01 -9.77248073e-01 -1.09780669e+00 1.09175193e+00
-1.31760156e+00 -1.36149335e+00 -7.41022229e-01 4.77860510e-01
3.79451215e-01 -1.34557128e-01 9.65983748e-01 3.35168213e-01
-7.18724251e-01 4.43608165e-01 4.26514409e-02 7.64452815e-02
7.38983154e-01 -1.36077631e+00 -3.86934072e-01 6.06309652e-01
-5.23437858e-01 4.64115858e-01 7.60142267e-01 -5.03308535e-01
-1.05663586e+00 -5.36080658e-01 6.44680262e-01 1.58717573e-01
3.15159678e-01 1.38479874e-01 -7.46766567e-01 9.07266513e-02
5.24547338e-01 -2.61483103e-01 8.13404739e-01 -7.45636001e-02
8.11351240e-02 -3.55146229e-01 -1.67745304e+00 4.35624540e-01
1.17476434e-01 -1.18091911e-01 -4.61913981e-02 4.86592203e-01
5.51128387e-03 -2.21518457e-01 -8.77824128e-01 2.26996958e-01
5.51252484e-01 -1.35121775e+00 6.90124571e-01 -6.03477120e-01
6.53437197e-01 -1.57889202e-01 4.66442585e-01 -1.13646674e+00
-4.47087325e-02 -1.01368323e-01 5.72573721e-01 1.28898942e+00
5.77136576e-01 -5.07638931e-01 1.21983719e+00 5.65738559e-01
2.01691031e-01 -1.20206475e+00 -6.24047399e-01 -2.70486623e-01
2.45241061e-01 1.11604504e-01 9.57608521e-02 1.07816744e+00
-3.43724377e-02 -2.06309304e-01 -4.32474986e-02 -1.20706975e-01
4.52262431e-01 -6.36787176e-01 4.54801947e-01 -1.15840030e+00
4.21894975e-02 -2.96233594e-01 -7.29437709e-01 1.57774597e-01
-7.13195980e-01 -5.43454051e-01 -6.29568756e-01 -1.84207571e+00
2.41550177e-01 -4.67136204e-01 -1.04961999e-01 5.55812359e-01
-1.28348216e-01 2.80068457e-01 -5.46365492e-02 -2.87822068e-01
3.43718648e-01 -4.71623123e-01 1.47551775e+00 -1.26589850e-01
-1.37315899e-01 1.77199468e-01 -7.04655945e-01 7.26663113e-01
1.27095044e+00 -1.65301979e-01 -4.51382965e-01 9.52861756e-02
3.19303691e-01 -6.57052919e-02 1.83623210e-01 -1.23592496e+00
3.40888411e-01 -2.79744893e-01 7.48190880e-01 -4.48001236e-01
1.52506426e-01 -1.00539315e+00 3.29032272e-01 9.07072425e-01
-4.56574596e-02 -1.47051245e-01 1.77004278e-01 1.30265296e-01
-2.28186458e-01 -7.53998041e-01 1.14712739e+00 -5.09335816e-01
-6.29246950e-01 -4.60578315e-02 -4.85298723e-01 -5.25896549e-01
1.60256886e+00 -7.51251042e-01 -3.52020979e-01 -9.36316997e-02
-7.43052840e-01 -4.95249666e-02 4.52968210e-01 6.44454658e-02
2.75508493e-01 -8.24774444e-01 -5.13476551e-01 -4.02443334e-02
-1.03599459e-01 -1.18216937e-02 6.86482728e-01 1.17553198e+00
-1.23759866e+00 2.08157510e-01 -8.14507484e-01 -3.79259825e-01
-1.73939335e+00 4.54389721e-01 7.21176565e-01 -3.04828554e-01
-2.51632780e-01 9.49589670e-01 -5.19409120e-01 -7.39362985e-02
1.20528318e-01 -2.97089338e-01 -6.91780329e-01 -6.67229062e-03
5.91030657e-01 5.20403385e-01 1.67922452e-01 -1.96416989e-01
-1.03808329e-01 6.65887475e-01 -3.01578790e-01 2.67688453e-01
1.27950764e+00 2.08617985e-01 -4.56388086e-01 3.11087340e-01
1.07658994e+00 -4.22474928e-02 -7.07795560e-01 6.44731939e-01
-2.73625463e-01 -4.13993120e-01 1.08827896e-01 -1.36513388e+00
-1.21367705e+00 9.98983026e-01 1.21255684e+00 5.33597767e-01
1.36179781e+00 -6.33227408e-01 4.71214831e-01 4.99163121e-02
-3.41703705e-02 -1.38265848e+00 -2.40071595e-01 1.50765851e-02
4.50564235e-01 -1.48953772e+00 6.25378639e-02 -5.56994140e-01
-6.18364990e-01 1.54267967e+00 6.13105774e-01 -3.81347924e-01
6.72756612e-01 4.40039843e-01 3.53120118e-01 -2.59186715e-01
-4.09435481e-01 -1.67691603e-01 2.67430879e-02 8.14823627e-01
7.67356277e-01 1.60747483e-01 -8.47811639e-01 2.77170628e-01
1.06445877e-02 3.73146683e-01 6.81027651e-01 7.98434615e-01
-6.66629255e-01 -1.22548795e+00 -6.39209449e-01 8.96854997e-01
-1.05392444e+00 1.77077636e-01 -4.98444319e-01 1.16048646e+00
6.70385838e-01 8.54556382e-01 3.68995592e-02 -3.40030819e-01
8.76202658e-02 1.20154373e-01 4.09488142e-01 -1.80327848e-01
-9.29767668e-01 -2.36445442e-01 6.13539033e-02 -1.54706270e-01
-5.34697890e-01 -3.65110576e-01 -1.23633158e+00 -5.12495875e-01
-2.32332736e-01 -9.04829353e-02 1.20378196e+00 7.96009481e-01
-2.07979947e-01 6.82369828e-01 4.07879889e-01 -1.55869380e-01
-3.72622222e-01 -1.22143638e+00 -6.02726877e-01 3.91691178e-01
-6.68673124e-03 -3.45966727e-01 -3.41419429e-01 5.44383265e-02] | [15.611842155456543, -2.983447790145874] |
09a32a4b-b4be-4f68-86e9-1ac258d6f0ff | joint-hand-detection-and-rotation-estimation | 1612.02742 | null | http://arxiv.org/abs/1612.02742v1 | http://arxiv.org/pdf/1612.02742v1.pdf | Joint Hand Detection and Rotation Estimation by Using CNN | Hand detection is essential for many hand related tasks, e.g. parsing hand
pose, understanding gesture, which are extremely useful for robotics and
human-computer interaction. However, hand detection in uncontrolled
environments is challenging due to the flexibility of wrist joint and cluttered
background. We propose a deep learning based approach which detects hands and
calibrates in-plane rotation under supervision at the same time. To guarantee
the recall, we propose a context aware proposal generation algorithm which
significantly outperforms the selective search. We then design a convolutional
neural network(CNN) which handles object rotation explicitly to jointly solve
the object detection and rotation estimation tasks. Experiments show that our
method achieves better results than state-of-the-art detection models on
widely-used benchmarks such as Oxford and Egohands database. We further show
that rotation estimation and classification can mutually benefit each other. | ['Shuo Yang', 'Yinda Zhang', 'Ye Yuan', 'Ping Tan', 'Hongan Wang', 'Xiaoming Deng', 'Liang Chang'] | 2016-12-08 | null | null | null | null | ['hand-detection'] | ['computer-vision'] | [-2.60320365e-01 -4.74422812e-01 -4.32459682e-01 -1.09279059e-01
-3.66727889e-01 -5.85178137e-01 3.13472003e-01 -6.76092625e-01
-5.32963037e-01 3.73203814e-01 1.26756057e-01 -2.80699104e-01
1.10117890e-01 -4.36998188e-01 -6.59810185e-01 -6.41156495e-01
1.71389967e-01 6.23864949e-01 5.37746727e-01 6.23845272e-02
2.76241988e-01 9.96095359e-01 -1.35109532e+00 -5.42723462e-02
1.89059362e-01 8.05614233e-01 5.25121629e-01 8.05236280e-01
1.93394199e-01 4.83768821e-01 -5.37332416e-01 -1.40362501e-01
3.77529860e-01 9.55335274e-02 -7.87941694e-01 1.64072756e-02
4.28802520e-01 -9.85025644e-01 -6.88954175e-01 8.85767639e-01
1.08189964e+00 1.13032959e-01 5.06551087e-01 -1.00024343e+00
-2.70176321e-01 5.99651933e-01 -1.15372777e+00 1.54772758e-01
4.97494996e-01 3.55651706e-01 9.08858299e-01 -1.02171743e+00
8.15109074e-01 1.54056847e+00 3.74406070e-01 4.65474874e-01
-6.69754207e-01 -8.38454425e-01 5.89993060e-01 6.75530359e-02
-1.53293025e+00 -3.80126163e-02 6.84572279e-01 -2.81952977e-01
9.05002594e-01 1.32536441e-01 6.11710668e-01 1.49937034e+00
-2.63507754e-01 1.42833674e+00 6.81995332e-01 -4.78325367e-01
-1.61559001e-01 -4.28675622e-01 8.53836164e-02 8.78974319e-01
3.37563187e-01 4.62117232e-02 -3.90545905e-01 3.13728541e-01
1.44636452e+00 3.76316637e-01 -2.72436708e-01 -4.78261560e-01
-1.56877744e+00 3.97148043e-01 7.01060891e-01 6.49272576e-02
-4.71681595e-01 3.56399447e-01 3.02209675e-01 -1.56520322e-01
-2.19207138e-01 1.97767299e-02 -6.84984088e-01 6.75530359e-02
-5.44828773e-01 5.65590858e-01 6.43346548e-01 1.47478664e+00
-1.14700675e-01 -2.33209282e-01 -6.11196816e-01 6.23806238e-01
5.67833185e-01 6.66660547e-01 -3.71097289e-02 -5.98183274e-01
7.41261661e-01 3.24532032e-01 2.78391302e-01 -5.05937397e-01
-6.71024919e-01 -6.14340246e-01 -7.88264811e-01 1.64809883e-01
7.86406338e-01 2.85794232e-02 -1.15280128e+00 1.51935029e+00
5.04427910e-01 -2.48654991e-01 -4.77578461e-01 1.51178229e+00
9.12111521e-01 6.88918829e-02 1.30817220e-01 4.85278741e-02
1.82660747e+00 -1.23118126e+00 -6.70135677e-01 -2.88888454e-01
9.80140120e-02 -8.89783442e-01 1.14342093e+00 5.99651754e-01
-6.14490032e-01 -6.36327088e-01 -8.91422272e-01 -1.83235884e-01
1.04085542e-01 7.47984707e-01 9.72041667e-01 5.90350330e-01
-2.83266515e-01 3.27776045e-01 -1.06934083e+00 -4.55252588e-01
7.31879711e-01 4.15663242e-01 -1.75699815e-01 7.50133721e-03
-6.71926796e-01 8.10114324e-01 2.76214123e-01 3.91364515e-01
-6.82060301e-01 -1.10866211e-01 -5.57180941e-01 -1.64225604e-02
8.48574102e-01 -6.23927176e-01 1.51891661e+00 -1.16530627e-01
-1.47353625e+00 7.54481673e-01 -8.66765529e-02 4.71365713e-02
1.10209954e+00 -7.47741461e-01 1.53264418e-01 -7.36316070e-02
1.32780522e-02 6.93099916e-01 1.02237701e+00 -9.84649897e-01
-6.37501717e-01 -7.60570109e-01 -7.55647644e-02 1.90385371e-01
-5.06862104e-02 3.12073618e-01 -1.09412825e+00 -9.45555031e-01
5.84272623e-01 -1.03923190e+00 -2.18292281e-01 3.13238531e-01
-8.63461614e-01 -6.18452787e-01 9.73560274e-01 -7.89275587e-01
9.57387149e-01 -1.59715307e+00 2.50317037e-01 3.01095247e-01
1.61366537e-01 2.92798609e-01 -5.49895409e-03 -1.82271108e-01
2.02568993e-01 -3.64952356e-01 2.98963279e-01 -2.94323951e-01
-3.08895521e-02 -4.31568399e-02 -2.93235064e-01 4.82056230e-01
3.38100679e-02 9.68064249e-01 -8.95066202e-01 -6.60039246e-01
3.70812386e-01 6.37603283e-01 -4.17625815e-01 3.15746248e-01
-2.36739054e-01 7.06194937e-01 -6.20898664e-01 1.10228252e+00
5.08713722e-01 -3.26771587e-01 3.35461766e-01 -3.61847758e-01
7.69084916e-02 1.74347863e-01 -1.62900496e+00 1.96105421e+00
-3.35311025e-01 4.40419227e-01 2.12084688e-02 -4.51490939e-01
6.07091367e-01 2.76530832e-01 2.08568648e-02 -2.26729929e-01
5.12430429e-01 9.50150192e-02 2.09652215e-01 -3.50894958e-01
2.90508986e-01 4.86541301e-01 3.48162949e-01 5.72218537e-01
1.39549775e-02 3.78568694e-02 -2.06969365e-01 -9.30019841e-02
7.06085205e-01 5.68984449e-01 3.94453019e-01 1.33841231e-01
2.71384835e-01 -4.75553334e-01 2.81091869e-01 8.85271132e-01
-3.82603586e-01 7.82043874e-01 1.97743475e-01 -6.17510974e-01
-8.49011183e-01 -9.97391284e-01 -1.08673364e-01 1.46631551e+00
2.09556639e-01 -8.87693386e-06 -6.46407187e-01 -9.40651715e-01
3.05468142e-02 2.65796483e-02 -4.40523714e-01 4.66776997e-01
-1.11579669e+00 -4.40963209e-01 3.45173180e-01 1.49518371e+00
4.83378291e-01 -1.37875867e+00 -1.02263081e+00 8.49635676e-02
-9.54985097e-02 -1.28033447e+00 -7.11998343e-01 1.45726129e-01
-6.23741686e-01 -1.16688251e+00 -1.28388309e+00 -9.48062301e-01
4.74988759e-01 4.79417503e-01 8.33431005e-01 1.71813834e-02
-9.14506972e-01 1.79840446e-01 -2.06856057e-01 -5.23361921e-01
2.40618899e-01 6.39237404e-01 2.25777954e-01 -6.20578408e-01
3.97386104e-01 -4.17432129e-01 -8.45114946e-01 4.46661294e-01
-2.31035233e-01 -1.90039173e-01 9.44574416e-01 6.77958310e-01
4.86942559e-01 -5.69157541e-01 2.82950580e-01 -2.78299540e-01
4.17798966e-01 3.22269171e-01 -7.33319819e-01 4.55627888e-01
-2.98006773e-01 2.32571542e-01 -1.10570833e-01 -7.30653822e-01
-1.05515826e+00 6.47558570e-01 -4.53956202e-02 -5.79689205e-01
-1.95331022e-01 -2.90657490e-01 -3.16012025e-01 -4.03597951e-02
5.34208000e-01 -8.04961771e-02 -3.42142075e-01 -7.55681217e-01
4.38897282e-01 8.76305044e-01 7.22646296e-01 -7.20037520e-01
6.16215706e-01 4.08935457e-01 8.02472904e-02 -6.56683743e-01
-7.08160341e-01 -5.58654547e-01 -9.25793290e-01 -1.90587327e-01
7.23597527e-01 -7.38378942e-01 -1.37431800e+00 5.95741689e-01
-1.79296219e+00 -6.85384870e-02 2.61267960e-01 6.31961226e-01
-4.04704213e-01 4.37548071e-01 -5.28453052e-01 -1.22206092e+00
-7.51707733e-01 -1.39456689e+00 1.52233136e+00 2.47717351e-01
-2.36694962e-01 -9.46271345e-02 -1.87372208e-01 7.47924820e-02
-2.10459586e-02 -2.11059488e-03 3.22388619e-01 -4.30132002e-01
-1.01280141e+00 -3.68679017e-01 -6.62185848e-01 -1.26446709e-01
1.37135610e-01 -1.13758571e-01 -1.06852496e+00 -4.35861021e-01
-6.64191425e-01 -3.39334607e-01 9.62211430e-01 4.17481273e-01
1.52128816e+00 3.22968513e-02 -6.88950181e-01 4.88196224e-01
7.41558015e-01 -1.62922814e-02 3.82770330e-01 2.71592349e-01
1.05234945e+00 5.59449673e-01 6.13382578e-01 6.26234233e-01
9.24841911e-02 9.83578861e-01 4.12153363e-01 1.52787670e-01
-3.10070008e-01 -5.31715453e-02 -1.52697742e-01 1.41682491e-01
-8.86274457e-01 1.02135994e-01 -9.33070719e-01 3.37713540e-01
-1.87137246e+00 -8.11949968e-01 6.37210254e-03 2.13318467e+00
8.99936080e-01 1.81714162e-01 5.11585355e-01 2.75480330e-01
8.57638359e-01 4.90692221e-02 -6.73227370e-01 4.46531653e-01
2.19241902e-01 5.88196218e-01 4.81911331e-01 1.75746560e-01
-1.45016515e+00 1.31063700e+00 5.97681713e+00 4.48034823e-01
-9.74786341e-01 6.39332905e-02 2.58792453e-02 -4.43266556e-02
6.75330520e-01 -5.47301292e-01 -1.13016582e+00 4.09928570e-03
-4.07026201e-01 7.76229680e-01 5.46111882e-01 1.39957500e+00
-2.41738990e-01 2.95376897e-01 -1.29937840e+00 1.27368701e+00
-1.01069406e-01 -8.42518985e-01 -1.19519666e-01 -9.19548050e-02
4.50742632e-01 -1.21041730e-01 2.99300700e-02 1.12831183e-01
1.21611416e-01 -9.90896702e-01 8.20648015e-01 2.38868952e-01
9.57092822e-01 -6.33451819e-01 5.95678270e-01 3.65663737e-01
-1.55748630e+00 -1.88309373e-03 -5.16567566e-02 5.65514006e-02
6.92778379e-02 6.64658472e-02 -8.68709505e-01 1.91613242e-01
8.28044355e-01 2.44945422e-01 -2.76360750e-01 9.79050040e-01
-8.02477658e-01 1.07322529e-01 -4.27364081e-01 -2.64241546e-01
-1.36231691e-01 2.90435493e-01 6.38197124e-01 1.33008873e+00
1.09761776e-02 2.29156509e-01 3.07523459e-01 8.29260528e-01
-1.15487874e-01 -1.36137590e-01 -1.84556574e-01 1.58793494e-01
7.42217064e-01 1.09805071e+00 -1.01696765e+00 -2.41576374e-01
-1.69169620e-01 1.29355276e+00 2.77171224e-01 2.63380587e-01
-6.53436005e-01 -7.20459163e-01 6.20410979e-01 -2.00467613e-02
5.77042639e-01 -4.56873924e-01 -3.25706691e-01 -1.26112115e+00
5.98353028e-01 -7.09063530e-01 1.94540113e-01 -4.67110753e-01
-1.03582811e+00 4.02799129e-01 -2.32464060e-01 -1.06994271e+00
-2.69985765e-01 -1.24382007e+00 -4.86632317e-01 7.27869511e-01
-1.29500258e+00 -1.47709477e+00 -8.51909101e-01 5.99104762e-01
7.63936698e-01 -8.49186629e-02 6.23171568e-01 1.24303915e-01
-5.66955388e-01 8.91093433e-01 -6.06983960e-01 8.48113656e-01
8.32166553e-01 -1.19630694e+00 7.40442336e-01 6.46927416e-01
2.73820877e-01 1.02878964e+00 4.33558762e-01 -6.26843989e-01
-1.65115643e+00 -8.06839883e-01 4.03867513e-01 -5.45955718e-01
1.80741265e-01 -3.60563517e-01 -5.64002395e-01 8.10007334e-01
-1.98286518e-01 4.04684901e-01 5.44000119e-02 3.95502061e-01
-7.34975874e-01 2.00060040e-01 -8.98254395e-01 6.19719326e-01
1.79550791e+00 -3.65143836e-01 -6.23593450e-01 6.26305938e-01
5.23496687e-01 -9.70137656e-01 -3.47453088e-01 3.43051642e-01
1.30923569e+00 -6.40156746e-01 1.40888870e+00 -1.00884175e+00
1.08819149e-01 -2.61279434e-01 -6.60941973e-02 -5.75257659e-01
-2.50878125e-01 -5.58526337e-01 -6.19921803e-01 7.13804662e-01
-1.14379413e-01 -3.24715018e-01 1.07808578e+00 2.28156045e-01
4.07708108e-01 -5.98640203e-01 -8.79888952e-01 -9.65078294e-01
-3.11261654e-01 -4.63774115e-01 5.90513170e-01 5.10432243e-01
-1.94491521e-01 4.06136990e-01 -3.48707944e-01 3.13314050e-01
8.65660846e-01 3.08595270e-01 1.05639625e+00 -1.29276204e+00
-3.69800210e-01 -6.64955437e-01 -4.10661578e-01 -1.58869159e+00
4.24197651e-02 -3.30020159e-01 2.74520874e-01 -1.36857474e+00
4.56218004e-01 -3.81285816e-01 -1.96384102e-01 5.39245009e-01
-3.91154170e-01 1.39490113e-01 3.22998047e-01 3.24070722e-01
-5.96270263e-01 1.73174471e-01 1.57035363e+00 -7.36527815e-02
-4.01622236e-01 4.09655988e-01 -2.51752824e-01 8.72518241e-01
4.80313718e-01 -1.69006765e-01 7.93363675e-02 -4.75614190e-01
-1.61714718e-01 3.24045904e-02 5.57248473e-01 -9.31727052e-01
1.91391960e-01 -2.14888304e-01 7.81141877e-01 -1.10470927e+00
1.27750680e-01 -6.39883459e-01 -6.50712550e-01 7.63309419e-01
-1.69138804e-01 -1.53932616e-01 -1.60040006e-01 5.76638341e-01
4.18783367e-01 9.45437700e-02 6.87487662e-01 -1.39191821e-01
-6.06526852e-01 4.78977740e-01 6.92316294e-02 -3.21949899e-01
7.61039972e-01 8.77096355e-02 -1.36497140e-01 -2.46254772e-01
-7.06685066e-01 1.63800508e-01 -6.47308752e-02 7.75491834e-01
5.56832433e-01 -1.25268602e+00 -5.30881107e-01 6.45744205e-02
8.42548981e-02 6.06750250e-01 -1.75305769e-01 7.58510172e-01
-5.25537729e-01 6.68811083e-01 -1.28812060e-01 -8.69268715e-01
-1.42091370e+00 5.90968311e-01 7.30759129e-02 -1.39126629e-01
-8.31479549e-01 1.14026392e+00 4.06840891e-02 -5.01942337e-01
1.05637038e+00 -6.93918109e-01 -3.20730843e-02 -3.00605118e-01
8.36225808e-01 5.69107115e-01 2.17383690e-02 -2.27534279e-01
-6.77860856e-01 7.52523005e-01 -2.68352240e-01 -8.17624256e-02
1.00674832e+00 4.95616108e-01 2.18335003e-01 -2.52343476e-01
7.07479179e-01 -1.93221852e-01 -1.39560819e+00 -4.77964044e-01
-9.72289369e-02 -6.42485499e-01 -9.65381190e-02 -8.88361454e-01
-1.07071519e+00 1.24778819e+00 7.58445382e-01 -5.50440788e-01
7.47865915e-01 2.67851174e-01 6.01437032e-01 1.18280435e+00
5.58257818e-01 -1.12754345e+00 4.42176789e-01 5.01915514e-01
1.19058585e+00 -1.40611970e+00 2.32326239e-01 -6.24854147e-01
-4.29678619e-01 1.30880690e+00 8.86760890e-01 -3.70933563e-02
2.79442996e-01 5.18428445e-01 -4.10706550e-02 3.35571058e-02
1.54387327e-02 -5.33053339e-01 4.21483517e-01 6.10478342e-01
4.86502260e-01 2.67151684e-01 -1.17836662e-01 8.56002271e-01
-5.85596897e-02 9.44872573e-02 -3.45512927e-01 1.31793988e+00
-4.13242519e-01 -1.00004077e+00 -8.10162902e-01 2.82360703e-01
-3.04516464e-01 2.10343823e-01 -4.91153896e-01 9.49151158e-01
-1.16116025e-01 3.25671762e-01 -8.51893350e-02 -3.00607473e-01
4.67520326e-01 -1.59364063e-02 1.23682284e+00 -5.69596469e-01
-2.82180756e-01 3.50068122e-01 -3.42396706e-01 -6.07271612e-01
-1.98981792e-01 -6.18196547e-01 -1.25928891e+00 -1.26960902e-02
-7.05368698e-01 -6.23686969e-01 6.63352370e-01 1.06404400e+00
-2.46078856e-02 6.60098910e-01 1.66407153e-01 -1.33411551e+00
-1.16282570e+00 -1.33443117e+00 -4.25014019e-01 1.40788080e-02
1.95282429e-01 -1.21974552e+00 2.67388642e-01 -2.80078590e-01] | [6.609283447265625, -0.6935527920722961] |
44dda5f7-55c6-4fc6-83cb-0c9e17500984 | accented-speech-recognition-under-the-indian | 2209.03787 | null | https://arxiv.org/abs/2209.03787v4 | https://arxiv.org/pdf/2209.03787v4.pdf | Goodness of Pronunciation Pipelines for OOV Problem | In the following report we propose pipelines for Goodness of Pronunciation (GoP) computation solving OOV problem at testing time using Vocab/Lexicon expansion techniques. The pipeline uses different components of ASR system to quantify accent and automatically evaluate them as scores. We use the posteriors of an ASR model trained on native English speech, along with the phone level boundaries to obtain phone level pronunciation scores. We used this as a baseline pipeline and implemented methods to remove UNK and SPN phonemes in the GoP output by building three pipelines. The Online, Offline and Hybrid pipeline which returns the scores but also can prevent unknown words in the final output. The Online method is based per utterance, Offline method pre-incorporates a set of OOV words for a given data set and the Hybrid method combines the above two ideas to expand the lexicon as well work per utterance. We further provide utilities such as the Phoneme to posterior mappings, GoP scores of each utterance as a vector, and Word boundaries used in the GoP pipeline for use in future research. | ['Ankit Grover'] | 2022-09-08 | null | null | null | null | ['accented-speech-recognition'] | ['speech'] | [ 2.89410166e-02 3.86710137e-01 4.39359486e-01 -4.82080936e-01
-1.28234291e+00 -1.01781404e+00 1.21248491e-01 6.67129681e-02
-3.92602086e-01 6.42904937e-01 7.75541365e-01 -3.62131655e-01
4.71305288e-02 -5.85617959e-01 -4.30263489e-01 -3.00232291e-01
2.07013085e-01 7.30690241e-01 3.06364924e-01 -4.41135585e-01
3.46245438e-01 1.29761830e-01 -1.32440662e+00 2.85338402e-01
6.82269096e-01 9.82345104e-01 3.20622712e-01 1.28763294e+00
-3.74394327e-01 5.77089012e-01 -1.12811852e+00 -4.61015761e-01
2.90627480e-01 -3.47367138e-01 -8.66694391e-01 -3.24601680e-01
2.58937985e-01 1.30669046e-02 3.13858896e-01 1.20929408e+00
8.03734064e-01 4.57647651e-01 5.66296995e-01 -8.11138391e-01
-7.77130902e-01 1.01194739e+00 2.93114185e-01 2.92244434e-01
6.09829009e-01 -1.16244271e-01 9.53489363e-01 -1.11590075e+00
4.30047482e-01 1.28541577e+00 5.55546045e-01 3.72586310e-01
-8.44510019e-01 -6.50857270e-01 -4.29292284e-02 8.86073411e-02
-1.49912143e+00 -5.56620061e-01 5.72691083e-01 -3.77068132e-01
1.79332268e+00 5.16384482e-01 4.77020532e-01 7.47022450e-01
-7.67725557e-02 5.13753414e-01 1.11546075e+00 -8.09182823e-01
3.04528683e-01 5.05611062e-01 6.05773985e-01 5.35161138e-01
-6.86000526e-01 -7.12960735e-02 -6.49063408e-01 8.66567120e-02
4.69974071e-01 -7.46045887e-01 -4.00839327e-03 5.77282310e-01
-9.09670830e-01 7.45216429e-01 -3.20216477e-01 2.99143374e-01
-4.31588411e-01 -6.84556216e-02 2.61584550e-01 5.02124608e-01
2.64254093e-01 4.53786075e-01 -1.09035587e+00 -4.49347466e-01
-9.07769442e-01 -9.96726565e-03 7.83248484e-01 9.91622448e-01
9.13008571e-01 2.12982550e-01 -5.13101697e-01 1.33068526e+00
4.41901118e-01 4.44183469e-01 9.52063143e-01 -8.47149432e-01
5.12010932e-01 3.33451509e-01 9.80676413e-02 -5.10075390e-01
1.20143659e-01 -2.57123142e-01 -6.83472455e-02 -1.15807265e-01
-1.03304461e-01 -3.24046880e-01 -1.40967667e+00 1.84808910e+00
2.59478778e-01 2.64472604e-01 3.57935786e-01 3.57920855e-01
9.38371718e-01 1.16796064e+00 8.28974992e-02 -2.46346444e-01
1.52993500e+00 -1.13031530e+00 -1.14127266e+00 -1.65729254e-01
6.72239244e-01 -1.19506264e+00 1.27933002e+00 6.36168599e-01
-1.23448253e+00 -8.20816755e-01 -1.12214828e+00 4.38267998e-02
-8.29538345e-01 2.15182364e-01 6.33276850e-02 9.92109835e-01
-1.71609259e+00 3.88851345e-01 -4.64781076e-01 -2.80037403e-01
-3.45324665e-01 6.51267350e-01 -3.02437842e-01 5.10119140e-01
-1.59321129e+00 1.08616352e+00 8.20466459e-01 -2.03168228e-01
-8.36013377e-01 -5.44008493e-01 -1.12178922e+00 7.45121837e-02
-6.95923865e-02 -4.32880595e-02 1.42342806e+00 -7.20305026e-01
-2.08841610e+00 7.37186134e-01 -2.58738726e-01 -3.39643508e-01
-4.20862995e-02 -3.76830816e-01 -6.72979355e-01 -3.96165341e-01
-3.53915431e-02 6.57286465e-01 4.39258188e-01 -6.05878770e-01
-9.13614094e-01 -8.33819136e-02 -4.10997868e-01 4.34238613e-01
-1.91650778e-01 3.63387883e-01 -5.05252898e-01 -6.42271817e-01
1.84067190e-01 -7.40778565e-01 1.08023010e-01 -1.24762845e+00
-4.71211821e-01 -7.44794905e-01 4.24477756e-01 -1.36012340e+00
1.58998203e+00 -2.14800811e+00 -6.02886751e-02 4.31196034e-01
-6.13939762e-01 3.15957546e-01 -4.59735766e-02 2.05657929e-01
-8.46087262e-02 3.94733876e-01 -6.89479783e-02 -3.75093192e-01
2.42234275e-01 2.11018488e-01 -3.06673378e-01 -2.14794278e-01
5.04587829e-01 7.18857348e-01 -6.90513253e-01 -5.22419930e-01
4.71228719e-01 5.46604335e-01 -5.18253028e-01 5.68573713e-01
-8.78105536e-02 1.27890795e-01 1.46386459e-01 3.30384344e-01
6.14951313e-01 7.59400785e-01 -1.07378304e-01 1.78098589e-01
-2.23802656e-01 1.04570937e+00 -1.60264087e+00 1.46980691e+00
-5.64514875e-01 3.08381051e-01 -6.18489571e-02 -6.24023378e-01
1.34644699e+00 7.33376086e-01 -3.20556223e-01 -3.88460994e-01
2.04549372e-01 3.72664779e-01 -1.14418522e-01 -3.57714146e-01
7.52034307e-01 -1.05972454e-01 -3.00681859e-01 -4.31874655e-02
8.03281307e-01 -2.54050910e-01 5.75816780e-02 -1.66393351e-02
1.03988492e+00 3.19252133e-01 5.06998539e-01 -2.20702618e-01
8.09148967e-01 -7.44116381e-02 7.22055554e-01 7.60472834e-01
-1.67380363e-01 7.11279213e-01 2.71687776e-01 1.59417287e-01
-8.26909482e-01 -1.33144212e+00 -1.62590131e-01 1.50668097e+00
-3.81210387e-01 -3.67168963e-01 -1.14358115e+00 -4.66212541e-01
-4.00093675e-01 1.22217786e+00 -3.38747174e-01 1.75072640e-01
-7.83197880e-01 -4.27552491e-01 7.60682523e-01 6.44032955e-01
1.72252879e-01 -1.55232704e+00 2.85421461e-01 5.23976803e-01
-2.30338663e-01 -8.57694864e-01 -5.32505929e-01 5.58198094e-01
-3.44304919e-01 -2.75803208e-01 -4.19501871e-01 -1.21757686e+00
1.87929451e-01 -6.67814493e-01 1.00129926e+00 -3.32681715e-01
2.56839782e-01 8.77794921e-02 -6.16023481e-01 -4.47884798e-01
-8.62047851e-01 -9.19409618e-02 1.60615623e-01 -5.22741079e-02
8.21096599e-01 -2.64147252e-01 -1.56581894e-01 8.68913457e-02
-4.41625893e-01 -3.29023182e-01 3.81569952e-01 5.87444603e-01
8.58939230e-01 8.02743733e-02 7.16904461e-01 -8.95087063e-01
7.29550123e-01 -4.23766911e-01 -5.43868899e-01 2.61995465e-01
-3.82897586e-01 2.45511129e-01 5.99337280e-01 -3.44105780e-01
-1.09897423e+00 9.81497467e-02 -1.13957942e+00 -1.00266658e-01
-3.57006878e-01 4.64538366e-01 -5.88318825e-01 4.71995920e-01
4.00068939e-01 1.22667015e-01 -5.39667487e-01 -7.70106316e-01
4.22869414e-01 1.27946794e+00 7.66202152e-01 -3.64072084e-01
3.65111262e-01 -5.99947751e-01 -9.53014970e-01 -7.97823608e-01
-6.44051075e-01 -8.00387442e-01 -6.10287964e-01 -6.94281980e-02
1.11136556e+00 -6.38251781e-01 -3.97157103e-01 2.80959606e-01
-1.35121238e+00 -1.19993627e-01 -2.33157098e-01 5.65356255e-01
-3.16985816e-01 9.87625569e-02 -6.28735900e-01 -1.16834044e+00
-5.89214444e-01 -1.24239123e+00 9.12402213e-01 3.20853144e-01
-5.21742463e-01 -9.98625934e-01 3.81362438e-01 2.23662794e-01
4.42995042e-01 -4.31554228e-01 9.14560854e-01 -1.26205575e+00
-6.65726215e-02 -2.44022474e-01 3.74474972e-01 9.90207791e-01
-7.33580887e-02 -1.70205444e-01 -1.42717409e+00 4.87040915e-02
1.27594203e-01 -1.39822260e-01 5.41924715e-01 4.16687429e-01
7.55635560e-01 -4.02483106e-01 1.72180206e-01 4.80077833e-01
1.26591194e+00 5.46089232e-01 8.81594479e-01 1.38337150e-01
4.65745479e-01 6.16123736e-01 6.94381654e-01 -1.46886026e-02
1.51289895e-01 5.69844782e-01 -1.37534574e-01 2.88666308e-01
-2.47098327e-01 -4.61998194e-01 9.37219739e-01 1.87604570e+00
3.17695051e-01 -5.14048457e-01 -8.54675055e-01 7.69141912e-01
-1.41098011e+00 -5.17588615e-01 3.70398276e-02 2.34524250e+00
1.17714632e+00 2.53619164e-01 -1.83593780e-01 1.82527944e-01
8.61516058e-01 -1.42451316e-01 1.27761960e-01 -1.20564485e+00
-8.71015787e-02 9.21402812e-01 5.10531485e-01 1.28007293e+00
-9.17307317e-01 1.62568021e+00 6.59341621e+00 8.87390792e-01
-8.70319486e-01 4.50403333e-01 3.70831192e-01 3.41295689e-01
-3.87269258e-01 2.00751588e-01 -1.35646152e+00 3.78815055e-01
1.50840807e+00 1.92433015e-01 5.66603780e-01 8.64331841e-01
-4.60320786e-02 9.00559798e-02 -8.87772322e-01 8.97958875e-01
2.04322234e-01 -7.25879788e-01 -1.10148735e-01 -4.09156740e-01
4.92270559e-01 2.71789402e-01 -1.00476980e-01 7.70678401e-01
8.28960896e-01 -9.94664311e-01 8.54015529e-01 1.42132580e-01
8.43661368e-01 -7.89097667e-01 1.12615538e+00 1.47901371e-01
-1.09154153e+00 3.57735276e-01 -4.35530961e-01 -3.30309980e-02
1.53028250e-01 1.02484368e-01 -1.51676202e+00 1.94353968e-01
5.43393731e-01 -1.08723700e-01 -7.42204905e-01 6.86112106e-01
-4.68672961e-01 1.12397468e+00 -5.91407835e-01 -1.96400285e-01
1.67565033e-01 -1.02287382e-01 5.81230223e-01 1.75916708e+00
5.03227115e-01 -1.14892051e-01 -1.75272219e-03 5.63867033e-01
1.34780139e-01 5.41361392e-01 -1.80045575e-01 -2.39943936e-02
1.02580202e+00 1.18636358e+00 -6.68437302e-01 -5.35844207e-01
-7.53162131e-02 1.16930974e+00 4.51136529e-01 2.63086140e-01
-5.87782979e-01 -8.61148238e-01 4.99047130e-01 -3.50414246e-01
1.89139858e-01 7.09659234e-02 -2.14028344e-01 -7.97933877e-01
-1.38725892e-01 -9.12573695e-01 3.78391713e-01 -8.93968940e-01
-9.58082139e-01 8.45347703e-01 -1.78536847e-01 -6.11290336e-01
-5.56281328e-01 -7.46597171e-01 -5.89915574e-01 1.49435508e+00
-1.17520368e+00 -7.11935639e-01 6.62238225e-02 2.49751300e-01
9.81520534e-01 -3.58103871e-01 1.26404071e+00 1.09157294e-01
-4.49943602e-01 8.01309645e-01 -1.84287250e-01 4.33286190e-01
7.86205113e-01 -1.70359802e+00 6.25861764e-01 1.02049601e+00
3.23183537e-01 6.72811151e-01 6.48934901e-01 -9.20194924e-01
-7.80755699e-01 -9.25115228e-01 1.33151412e+00 -1.01701236e+00
5.39896786e-01 -5.50939023e-01 -9.25091684e-01 9.01190937e-01
4.29841071e-01 -4.93078589e-01 7.75429964e-01 2.59124368e-01
7.91507810e-02 6.30723909e-02 -9.93757188e-01 3.45655531e-01
5.49308479e-01 -5.76251090e-01 -1.20445061e+00 1.12046458e-01
1.20744419e+00 -5.25591910e-01 -8.29728425e-01 1.90368533e-01
1.13037415e-01 -5.79433978e-01 6.06293976e-01 -2.61621058e-01
-1.65023088e-01 -5.21541953e-01 -5.64087152e-01 -1.66420496e+00
-3.74433488e-01 -7.76367664e-01 3.10179919e-01 1.95039284e+00
1.06949186e+00 -6.97726190e-01 3.70023578e-01 2.74454772e-01
-4.69818473e-01 -1.96736395e-01 -8.77624631e-01 -5.50461948e-01
-2.42878590e-02 -9.08772469e-01 7.14151382e-01 8.26485753e-01
5.20585524e-03 5.53743601e-01 -1.85798943e-01 4.86974835e-01
7.56552443e-02 -8.41903806e-01 3.39914799e-01 -9.06227529e-01
-6.08004093e-01 -1.65729299e-01 -5.81767380e-01 -8.08637381e-01
-2.26017972e-03 -1.05725861e+00 4.86305565e-01 -1.39780974e+00
-3.11835110e-01 -4.09039974e-01 -7.17174351e-01 5.88917792e-01
-3.40538025e-01 2.15523899e-01 1.46426067e-01 -2.77811527e-01
-4.26854998e-01 4.87142429e-02 3.57135028e-01 2.69628763e-01
-6.40327036e-01 -1.95120722e-01 -5.19693375e-01 6.32200181e-01
9.61285889e-01 -7.43628204e-01 -3.55185986e-01 -4.58408594e-01
1.28802761e-01 -4.44193278e-03 -3.76355350e-01 -9.57627058e-01
1.24036282e-01 1.30243018e-01 3.70087862e-01 -8.69436920e-01
3.14523757e-01 -3.79662335e-01 1.11252896e-01 6.63576722e-02
-3.93394053e-01 2.92616427e-01 2.10409477e-01 6.69818297e-02
-5.01486242e-01 -6.83555722e-01 7.75659263e-01 -1.95971578e-01
-7.20350921e-01 -3.31907123e-01 -4.35772479e-01 1.52266473e-01
6.66068971e-01 -3.67097944e-01 5.42605855e-02 -2.68134922e-01
-1.05995369e+00 1.69269979e-01 8.72818753e-02 6.71561301e-01
5.41970253e-01 -1.20687842e+00 -7.67809391e-01 7.71061599e-01
-1.79491285e-02 -2.48272926e-01 -4.19214442e-02 3.83385777e-01
-8.31732452e-01 4.71851587e-01 -6.65979013e-02 -1.53905228e-01
-1.34792924e+00 2.29592204e-01 3.35264057e-01 -1.79070204e-01
-1.26491651e-01 1.45703316e+00 8.98619518e-02 -9.15229321e-01
5.46420157e-01 -4.08659756e-01 -7.64506221e-01 -8.25247690e-02
5.69749832e-01 1.75410241e-01 4.86755639e-01 -8.77542675e-01
-5.59991181e-01 4.37685281e-01 -3.11452840e-02 -8.75775337e-01
9.76372659e-01 -3.42023432e-01 -4.33753617e-02 7.82559276e-01
1.16695583e+00 6.48232818e-01 -7.00908720e-01 8.88649747e-02
3.34522784e-01 3.99078727e-02 2.06884220e-01 -1.16341138e+00
-3.63058507e-01 7.91565776e-01 7.78716326e-01 2.92770356e-01
9.41301107e-01 5.20510189e-02 7.93339729e-01 3.90614681e-02
5.21189608e-02 -1.68936896e+00 -2.82802701e-01 1.11941361e+00
4.98575240e-01 -8.93060923e-01 -8.02269876e-01 -3.59174281e-01
-9.83421981e-01 7.69895077e-01 3.63598078e-01 -1.34581044e-01
7.39515841e-01 4.85092729e-01 4.08909589e-01 2.86094964e-01
-7.01458871e-01 -2.76942462e-01 3.40214312e-01 8.41125906e-01
7.76269317e-01 4.28310245e-01 -3.76715660e-01 1.01653504e+00
-1.12887907e+00 -4.98091817e-01 4.06712115e-01 6.01653695e-01
-7.67648757e-01 -1.13241029e+00 -5.01974463e-01 1.76403552e-01
-8.98113012e-01 -6.57931149e-01 -6.64128304e-01 3.11643898e-01
3.64694476e-01 1.14032447e+00 1.07074685e-01 -7.17624784e-01
4.67139155e-01 8.56582403e-01 8.60698149e-02 -1.01239979e+00
-8.63239527e-01 2.95702338e-01 3.48412812e-01 -2.29589999e-01
3.81566733e-01 -3.84773403e-01 -1.35324514e+00 2.37112835e-01
-6.12828851e-01 3.89501631e-01 1.04184771e+00 9.35970545e-01
9.81541872e-02 5.63785315e-01 4.38485175e-01 -2.78111547e-01
-4.86575603e-01 -1.54852045e+00 -4.30023879e-01 3.54992449e-01
2.23036423e-01 -2.42773920e-01 -5.62484622e-01 1.42980367e-01] | [14.327410697937012, 6.811738014221191] |
8dee4503-4efb-485a-8be4-51fffc49f8fc | link-prediction-with-contextualized-self | 2201.10069 | null | https://arxiv.org/abs/2201.10069v2 | https://arxiv.org/pdf/2201.10069v2.pdf | Link Prediction with Contextualized Self-Supervision | Link prediction aims to infer the link existence between pairs of nodes in networks/graphs. Despite their wide application, the success of traditional link prediction algorithms is hindered by three major challenges -- link sparsity, node attribute noise and dynamic changes -- that are faced by many real-world networks. To address these challenges, we propose a Contextualized Self-Supervised Learning (CSSL) framework that fully exploits structural context prediction for link prediction. The proposed CSSL framework learns a link encoder to infer the link existence probability from paired node embeddings, which are constructed via a transformation on node attributes. To generate informative node embeddings for link prediction, structural context prediction is leveraged as a self-supervised learning task to boost the link prediction performance. Two types of structural context are investigated, i.e., context nodes collected from random walks vs. context subgraphs. The CSSL framework can be trained in an end-to-end manner, with the learning of model parameters supervised by both the link prediction and self-supervised learning tasks. The proposed CSSL is a generic and flexible framework in the sense that it can handle both attributed and non-attributed networks, and operate under both transductive and inductive link prediction settings. Extensive experiments and ablation studies on seven real-world benchmark networks demonstrate the superior performance of the proposed self-supervision based link prediction algorithm over state-of-the-art baselines, on different types of networks under both transductive and inductive settings. The proposed CSSL also yields competitive performance in terms of its robustness to node attribute noise and scalability over large-scale networks. | ['Philip S. Yu', 'Jie Yin', 'Daokun Zhang'] | 2022-01-25 | null | null | null | null | ['inductive-link-prediction'] | ['graphs'] | [ 3.67462933e-01 4.04595941e-01 -1.02114820e+00 -5.31119883e-01
-1.39055803e-01 -3.61679316e-01 6.06953382e-01 3.93966973e-01
3.50702927e-02 8.24419856e-01 1.60980850e-01 -4.16193575e-01
-4.60701346e-01 -1.16942108e+00 -7.49761045e-01 -4.90422398e-01
-6.02242708e-01 5.70720971e-01 3.96900773e-01 -1.33189902e-01
-4.24219579e-01 1.60205424e-01 -1.04644823e+00 -1.95438161e-01
1.00677657e+00 7.87203431e-01 -1.40143767e-01 3.12334061e-01
-2.72659957e-01 6.59704030e-01 3.67920339e-01 -6.27879500e-01
2.18678400e-01 -9.38960612e-02 -7.73389101e-01 -1.40033528e-01
4.99915242e-01 1.13191232e-02 -7.84815013e-01 8.45834196e-01
2.93464243e-01 -3.30556065e-01 5.15693188e-01 -1.75755453e+00
-7.58109033e-01 1.04943693e+00 -4.94118720e-01 2.91955322e-01
3.69330138e-01 -1.04377035e-03 1.67400086e+00 -8.35719764e-01
7.34987438e-01 1.27878809e+00 1.01770210e+00 2.47688726e-01
-1.82457459e+00 -8.10117841e-01 3.98279011e-01 2.49920934e-01
-1.36273670e+00 -2.74518639e-01 1.10123777e+00 -3.63957316e-01
6.31888032e-01 -7.01143369e-02 5.86884260e-01 1.30453658e+00
-7.81597570e-03 6.27962291e-01 7.80137658e-01 -7.65873343e-02
-4.65275049e-02 1.46633610e-01 2.11745933e-01 9.31556463e-01
4.45595145e-01 1.65031672e-01 -6.46866024e-01 -3.99191886e-01
5.62472641e-01 2.61226892e-01 -2.97506690e-01 -7.46986568e-01
-1.18260193e+00 7.34103501e-01 1.09447432e+00 -7.64334761e-03
-1.63723066e-01 1.70518711e-01 5.64553142e-01 5.09109139e-01
6.49664223e-01 -1.56221524e-01 -5.35762966e-01 4.29129571e-01
-4.38074201e-01 -1.94746658e-01 1.00634825e+00 1.06176782e+00
1.05278766e+00 -2.92035073e-01 -7.27866217e-02 8.06263864e-01
8.48771334e-01 3.62108916e-01 -8.99037719e-02 -2.07144201e-01
7.11421669e-01 9.07476664e-01 -5.51846564e-01 -1.36623406e+00
-3.29337478e-01 -8.24810803e-01 -1.02530253e+00 -3.97595555e-01
1.85648557e-02 -1.33537287e-02 -6.13196969e-01 2.13015723e+00
6.56137764e-01 8.33733737e-01 1.09265722e-01 4.73953873e-01
1.12532914e+00 5.31370997e-01 3.73822808e-01 -3.39638442e-01
8.86038363e-01 -1.07803547e+00 -7.40555763e-01 -2.71240860e-01
8.92943203e-01 -3.22418153e-01 7.78436482e-01 -5.41230142e-01
-7.04953492e-01 -2.87642777e-01 -1.01157665e+00 1.67049557e-01
-3.26140881e-01 -1.69012144e-01 9.35659766e-01 2.42938116e-01
-8.45998049e-01 5.79867244e-01 -6.49284184e-01 -6.40436292e-01
7.25564539e-01 4.22614515e-01 -6.83112979e-01 -2.29684338e-01
-1.49275887e+00 2.90331185e-01 6.47877157e-01 3.20864581e-02
-5.72312474e-01 -9.62872624e-01 -1.04358470e+00 2.67518580e-01
6.92678690e-01 -6.95934534e-01 5.04504442e-01 -7.41304219e-01
-9.36273932e-01 7.82855511e-01 -8.75379443e-02 -4.45118934e-01
3.08425248e-01 -5.97535782e-02 -8.85142148e-01 3.35475542e-02
3.17643344e-01 3.99919003e-01 6.97124183e-01 -1.29443848e+00
-5.44150710e-01 -1.88319474e-01 -6.29782602e-02 1.73512325e-02
-8.14128518e-01 -5.94115436e-01 -7.09991813e-01 -6.48932278e-01
1.92739457e-01 -1.01693654e+00 1.62598174e-02 3.31805646e-01
-7.83853829e-01 -4.76436466e-01 1.20870924e+00 -1.37458339e-01
1.37967491e+00 -1.85316968e+00 4.80009057e-02 6.75625622e-01
5.55646837e-01 3.17840993e-01 -4.91239637e-01 6.31915808e-01
-3.59939396e-01 1.40422374e-01 -1.96992666e-01 -3.68305326e-01
-1.41284153e-01 4.14087087e-01 -9.26750973e-02 2.80337691e-01
3.66531551e-01 1.23453820e+00 -1.30553412e+00 -8.45331192e-01
-1.92069843e-01 4.65181082e-01 -4.34774369e-01 3.61434519e-01
-5.77759705e-02 2.28995442e-01 -5.66121876e-01 7.62116373e-01
5.13857782e-01 -5.40464818e-01 7.61714935e-01 -5.99654317e-01
5.47029376e-01 3.19909990e-01 -9.90250051e-01 1.56984031e+00
-3.53975743e-01 4.80097741e-01 1.15972668e-01 -1.05376840e+00
1.02410722e+00 2.91002125e-01 5.23462951e-01 -2.81911552e-01
-3.02333444e-01 6.50297925e-02 -3.73199210e-03 -5.28115034e-01
-2.62870163e-01 1.60515457e-01 2.48175338e-01 3.46789360e-01
3.14223200e-01 7.00951815e-01 3.71324390e-01 7.93030620e-01
1.49209809e+00 7.47191533e-02 1.96596026e-01 -2.61082411e-01
7.23289073e-01 -2.87985444e-01 9.99669552e-01 3.44168097e-01
-1.13918364e-01 -1.68406352e-01 7.88041234e-01 -4.55576688e-01
-6.21419370e-01 -1.32097352e+00 -1.10448435e-01 1.29926825e+00
3.42820078e-01 -7.38650620e-01 -2.54308134e-02 -1.19674504e+00
4.02813137e-01 6.88331872e-02 -6.82659745e-01 -5.07982969e-01
-4.65054125e-01 -7.43210077e-01 3.03408474e-01 6.31775975e-01
3.86933416e-01 -7.86856771e-01 6.28016472e-01 2.94582546e-01
8.42619762e-02 -1.50792921e+00 -4.19149667e-01 1.95975438e-01
-9.96508956e-01 -1.34530652e+00 2.09780991e-01 -1.19502354e+00
9.01554585e-01 3.57241511e-01 1.36324370e+00 3.51105779e-01
-1.76891506e-01 3.86690587e-01 -3.40502799e-01 2.81241447e-01
-2.14726940e-01 4.88203466e-01 1.46975979e-01 2.63956040e-01
2.08679900e-01 -1.14903200e+00 -6.31085277e-01 5.42159021e-01
-7.35867321e-01 -1.32995322e-02 9.69690025e-01 1.01145768e+00
6.20575428e-01 2.06360385e-01 9.57943141e-01 -1.57118165e+00
5.33298492e-01 -1.04242814e+00 -2.10884243e-01 4.19026762e-01
-1.14537680e+00 1.61085218e-01 6.35349691e-01 -3.06374878e-01
-8.53523552e-01 2.79009975e-02 2.07965910e-01 -1.27629027e-01
2.11152852e-01 8.99300396e-01 -4.76805001e-01 -2.10425973e-01
4.51382250e-01 -1.11539036e-01 1.58675328e-01 -3.98044825e-01
3.71625870e-01 2.61179417e-01 5.07760227e-01 -6.01612270e-01
1.55707514e+00 2.32030436e-01 4.38370079e-01 -3.83090079e-01
-9.84430313e-01 -4.68113184e-01 -7.90510714e-01 -1.06463738e-01
3.20639640e-01 -1.06443083e+00 -4.39682335e-01 -1.97025090e-02
-7.00072646e-01 -2.72160500e-01 3.08460537e-02 1.87680021e-01
2.08477154e-02 3.64946872e-01 -6.09754980e-01 -2.83993512e-01
-5.04292250e-01 -7.84992099e-01 4.73958045e-01 1.04899511e-01
6.48852885e-02 -1.67148316e+00 1.50771022e-01 3.00580591e-01
2.65512556e-01 4.07177210e-01 1.26537395e+00 -7.92451024e-01
-5.91247380e-01 -4.47509110e-01 -5.30970097e-01 1.56314503e-02
5.19206464e-01 7.28014261e-02 -5.61556876e-01 -5.40246487e-01
-1.19069850e+00 -4.92510557e-01 9.38482106e-01 -2.48420328e-01
8.80303442e-01 -4.54086155e-01 -9.88620281e-01 5.91027737e-01
1.53398895e+00 -6.17690444e-01 2.20493048e-01 -6.99498728e-02
1.14149237e+00 4.67842609e-01 5.48843384e-01 1.02074027e-01
7.66911268e-01 5.62371612e-01 8.32393348e-01 -2.16180399e-01
-2.45790973e-01 -8.56107056e-01 1.35326236e-01 7.59019554e-01
1.79382294e-01 -1.90111563e-01 -9.18769777e-01 5.60160160e-01
-2.22249556e+00 -8.68906736e-01 -3.33785027e-01 1.99513292e+00
1.01394379e+00 5.05482852e-01 2.12696284e-01 6.57726005e-02
9.34220731e-01 3.21967810e-01 -7.35499561e-01 2.05161899e-01
8.73048231e-02 -2.38910422e-01 3.35528374e-01 4.64834094e-01
-1.22522509e+00 9.61695492e-01 5.21965599e+00 6.04231477e-01
-8.73248339e-01 5.17963432e-02 4.08193141e-01 2.54889816e-01
-4.42129672e-01 2.33292490e-01 -7.27712512e-01 4.59716499e-01
6.97086155e-01 -4.26373724e-03 2.18760192e-01 8.40916812e-01
4.26308885e-02 5.30195653e-01 -1.34080005e+00 7.07454681e-01
-2.52907962e-01 -1.37438333e+00 2.83744857e-02 3.94236296e-02
8.38697076e-01 1.30945668e-01 -6.87993467e-02 4.57641989e-01
4.73658830e-01 -8.69201481e-01 -5.46259582e-02 2.98222572e-01
7.89993048e-01 -4.97359216e-01 7.72831857e-01 1.70567527e-01
-1.83963215e+00 -1.58031568e-01 -9.09418240e-02 2.18366861e-01
8.74616113e-03 9.22101438e-01 -9.92336750e-01 7.12373376e-01
5.29349804e-01 1.53379202e+00 -6.70993626e-01 8.80872309e-01
-4.39169556e-01 9.76848245e-01 -3.81219327e-01 2.09644526e-01
1.78046510e-01 -8.28666463e-02 5.33614755e-01 1.07602394e+00
-2.55653471e-01 -2.63696462e-01 6.22805715e-01 7.65764952e-01
-6.09289289e-01 1.29427344e-01 -6.84919834e-01 4.23623398e-02
1.13157904e+00 1.68934441e+00 -4.41563517e-01 1.28996104e-01
-6.26815438e-01 5.49222827e-01 9.01872516e-01 4.28122938e-01
-7.41614163e-01 -1.96060464e-01 6.13383889e-01 4.89400297e-01
3.16357791e-01 6.80005029e-02 1.78356990e-01 -9.40177143e-01
7.05215782e-02 -3.19906175e-01 6.67861640e-01 -1.58861279e-01
-1.82610738e+00 5.13739109e-01 -8.58533755e-02 -1.17979968e+00
1.26397252e-01 -3.48664761e-01 -1.10444355e+00 4.26955372e-01
-2.01479912e+00 -1.58799350e+00 -5.30085742e-01 5.51904380e-01
-5.77036617e-03 -2.46072039e-01 7.50962496e-01 4.75180209e-01
-1.00300741e+00 9.67897296e-01 5.29808477e-02 4.64459568e-01
7.64802575e-01 -1.25039017e+00 1.56518444e-01 5.92529356e-01
1.32920071e-01 5.75886667e-01 3.45806867e-01 -8.45185578e-01
-1.57545090e+00 -1.79926825e+00 7.47999132e-01 -4.38641477e-03
1.06131506e+00 -3.34005266e-01 -9.69766378e-01 9.12748158e-01
-1.87447697e-01 9.96331751e-01 1.10411620e+00 5.83682656e-01
-7.48894751e-01 -6.96339667e-01 -1.04837108e+00 4.05050725e-01
1.63115287e+00 -6.70652211e-01 -1.52647585e-01 3.88892382e-01
7.98607409e-01 1.13990687e-01 -1.43991232e+00 7.15404809e-01
3.61262441e-01 -5.57778895e-01 1.22836959e+00 -7.88384855e-01
4.40574527e-01 -3.18247586e-01 1.22827515e-01 -1.33997476e+00
-6.86850309e-01 -6.80137336e-01 -7.04093635e-01 1.78439510e+00
7.42018521e-01 -7.87116945e-01 9.55122352e-01 2.57724106e-01
1.82906106e-01 -1.17696130e+00 -7.98515320e-01 -6.95712268e-01
-3.75523061e-01 -1.04876600e-01 5.57850659e-01 1.23447585e+00
2.86230613e-02 9.51669514e-01 -2.67433614e-01 3.58239561e-01
1.01891422e+00 2.02292487e-01 7.19897926e-01 -1.86039364e+00
-1.17444217e-01 -1.74539626e-01 -7.79990852e-01 -9.54692662e-01
7.33036816e-01 -1.44945872e+00 -3.37835222e-01 -1.52634287e+00
1.85347870e-01 -1.04809332e+00 -5.12664020e-01 7.22550452e-01
-4.82826352e-01 2.53693163e-01 -2.37400934e-01 3.69278282e-01
-9.55482483e-01 8.04707408e-01 9.51632798e-01 -3.14316690e-01
-1.44157976e-01 5.06416298e-02 -5.49913585e-01 6.87494338e-01
5.89240372e-01 -5.08941472e-01 -8.85916710e-01 -2.25518107e-01
5.10107696e-01 -4.53377478e-02 5.48214018e-01 -8.26592684e-01
3.56218159e-01 1.31159127e-01 1.10769995e-01 -4.90031183e-01
4.74214628e-02 -1.25011742e+00 1.83294434e-02 3.88650149e-01
-5.75842500e-01 -2.55731523e-01 -2.63897300e-01 1.44067514e+00
7.77766760e-03 1.85773686e-01 4.39818680e-01 4.39499706e-01
-8.06964338e-01 9.50334728e-01 5.08138895e-01 2.79338926e-01
1.24164641e+00 -4.07842770e-02 -3.22895020e-01 -1.90286458e-01
-8.11070442e-01 7.97071576e-01 1.52367562e-01 4.64828432e-01
5.52324235e-01 -1.88235736e+00 -7.29271650e-01 1.54615790e-01
5.99066377e-01 6.26261234e-02 -1.19698718e-01 9.53296900e-01
-1.09139699e-02 9.48138237e-02 1.48289993e-01 -6.75662160e-01
-1.25653827e+00 5.20524979e-01 2.89374944e-02 -6.36531353e-01
-6.34161472e-01 7.30533063e-01 -2.35500395e-01 -7.04461634e-01
4.29960757e-01 2.13060990e-01 -1.72573075e-01 -1.09687351e-01
6.44465163e-02 2.09040716e-01 -2.38887146e-01 -5.88136673e-01
-3.80268365e-01 4.24003184e-01 -3.32242101e-01 8.12896430e-01
1.51008272e+00 -1.46568447e-01 -2.96899170e-01 3.01554173e-01
1.39030790e+00 -2.20817015e-01 -1.15879130e+00 -1.18345058e+00
1.46337777e-01 -3.62125635e-01 8.44826996e-02 -6.54815435e-01
-1.41155779e+00 3.97829741e-01 2.76992142e-01 3.13416570e-01
7.98574209e-01 2.49093086e-01 8.08798492e-01 5.08733809e-01
1.77494988e-01 -7.35318661e-01 1.45439699e-01 3.01203787e-01
4.27823484e-01 -1.66559815e+00 2.02086762e-01 -1.22842419e+00
-2.15634808e-01 9.54526722e-01 9.25947785e-01 -6.15185276e-02
1.14179456e+00 -1.14279650e-02 -4.34761971e-01 -2.94824153e-01
-1.03599787e+00 -1.95296183e-01 4.08903152e-01 6.52117372e-01
4.85849231e-01 1.01250522e-02 -2.58727651e-02 3.76866490e-01
3.24825406e-01 -1.91280305e-01 -7.19817504e-02 6.94379628e-01
-1.21469505e-01 -1.29260468e+00 1.58488050e-01 7.84518838e-01
6.96472377e-02 -8.10189173e-02 -3.91594589e-01 6.57668531e-01
3.14360522e-02 8.82072628e-01 -9.21217203e-02 -6.17055357e-01
-2.06891298e-02 -1.29485369e-01 -2.46842317e-02 -6.90410674e-01
-2.71627307e-01 -3.84623557e-01 3.65142316e-01 -4.35996026e-01
-5.61684370e-01 -3.26913297e-01 -1.19341123e+00 -4.91513282e-01
-5.47788441e-01 2.17410475e-01 8.18392932e-02 8.67844343e-01
6.14329457e-01 5.06107032e-01 1.02824843e+00 -4.93550926e-01
-3.86596650e-01 -8.17533970e-01 -5.00043035e-01 7.48731673e-01
1.73606455e-01 -8.47960293e-01 -3.34215552e-01 -2.22899988e-01] | [7.293572902679443, 6.36279821395874] |
3be9e53a-a634-4ecf-866b-0933b8abe661 | hum3dil-semi-supervised-multi-modal-3d-human | 2212.07729 | null | https://arxiv.org/abs/2212.07729v1 | https://arxiv.org/pdf/2212.07729v1.pdf | HUM3DIL: Semi-supervised Multi-modal 3D Human Pose Estimation for Autonomous Driving | Autonomous driving is an exciting new industry, posing important research questions. Within the perception module, 3D human pose estimation is an emerging technology, which can enable the autonomous vehicle to perceive and understand the subtle and complex behaviors of pedestrians. While hardware systems and sensors have dramatically improved over the decades -- with cars potentially boasting complex LiDAR and vision systems and with a growing expansion of the available body of dedicated datasets for this newly available information -- not much work has been done to harness these novel signals for the core problem of 3D human pose estimation. Our method, which we coin HUM3DIL (HUMan 3D from Images and LiDAR), efficiently makes use of these complementary signals, in a semi-supervised fashion and outperforms existing methods with a large margin. It is a fast and compact model for onboard deployment. Specifically, we embed LiDAR points into pixel-aligned multi-modal features, which we pass through a sequence of Transformer refinement stages. Quantitative experiments on the Waymo Open Dataset support these claims, where we achieve state-of-the-art results on the task of 3D pose estimation. | ['Cristian Sminchisescu', 'Dragomir Anguelov', 'Yin Zhou', 'Jingwei Ji', 'Alexander Gorban', 'Mihai Zanfir', 'Andrei Zanfir'] | 2022-12-15 | null | null | null | null | ['3d-pose-estimation', '3d-human-pose-estimation'] | ['computer-vision', 'computer-vision'] | [ 5.57716936e-03 -1.34570196e-01 -2.00324878e-01 -6.12072766e-01
-5.94051540e-01 -4.75802779e-01 5.81535697e-01 -2.51232713e-01
-5.69185913e-01 3.34372908e-01 1.02212615e-01 -2.67749727e-01
2.91877508e-01 -6.03552878e-01 -7.70909131e-01 -3.31785977e-01
-8.15909579e-02 7.63475180e-01 6.22235298e-01 -5.26835203e-01
-4.81418567e-03 5.55610716e-01 -2.00462055e+00 -2.17338696e-01
4.56016451e-01 1.17813718e+00 1.14849463e-01 6.71797693e-01
4.18801308e-01 3.31676066e-01 4.58433889e-02 -4.33467567e-01
6.14830077e-01 2.88540304e-01 -2.79232502e-01 2.91276306e-01
8.02702069e-01 -4.64571148e-01 -5.22592425e-01 6.47226870e-01
6.66367114e-01 -2.13502854e-01 3.06537032e-01 -1.64129710e+00
-3.09422873e-02 -8.87540206e-02 -7.13876009e-01 6.05473705e-02
5.38500965e-01 4.97697949e-01 8.98154080e-01 -9.89653170e-01
4.86100912e-01 1.42740119e+00 9.63277996e-01 2.42561072e-01
-1.00004184e+00 -6.12949789e-01 3.07148006e-02 4.63888556e-01
-1.44395900e+00 -4.77481514e-01 6.59340799e-01 -5.71599603e-01
9.60449755e-01 1.96116820e-01 8.86591077e-01 1.17017996e+00
1.61516681e-01 8.39289844e-01 9.64760959e-01 -4.97988760e-02
-3.53739560e-02 2.65368909e-01 1.05610989e-01 8.03504229e-01
4.43800896e-01 5.95826268e-01 -8.38366807e-01 2.21129015e-01
3.22269529e-01 -2.90293079e-02 1.26383558e-01 -9.51073587e-01
-1.27762604e+00 6.84128821e-01 4.79148120e-01 -5.22959232e-01
-1.15847997e-01 1.30448103e-01 3.38517398e-01 1.56255215e-01
2.91276872e-01 -2.54198983e-02 -3.85886580e-01 -4.49223518e-01
-6.54155850e-01 7.52864599e-01 6.89795792e-01 1.32457256e+00
8.93625021e-01 -3.12412620e-01 3.31088781e-01 3.59649092e-01
4.53697175e-01 1.11902201e+00 -1.25743225e-01 -1.06000221e+00
5.72759926e-01 6.75669312e-01 2.61949956e-01 -9.28393126e-01
-7.00159073e-01 -3.18529159e-01 -5.37553072e-01 4.53137964e-01
5.44683695e-01 -3.65019701e-02 -6.02649450e-01 1.36346161e+00
7.00653434e-01 -7.12742731e-02 -4.67077941e-01 1.11821532e+00
5.52427232e-01 2.46440142e-01 -1.61841005e-01 4.12752062e-01
1.51223421e+00 -7.59054184e-01 -1.58266485e-01 -7.54014313e-01
2.41520867e-01 -6.54149532e-01 8.32257330e-01 4.10323620e-01
-7.64110625e-01 -7.68986881e-01 -1.35367286e+00 -4.08385545e-01
-4.65022445e-01 -1.05728306e-01 7.09457994e-01 9.86683846e-01
-8.29904974e-01 1.26291618e-01 -9.08088207e-01 -5.81888855e-01
6.08231962e-01 2.59969085e-01 -5.65231502e-01 -2.56671607e-01
-8.87784958e-01 1.37732530e+00 -4.72026318e-02 1.42396823e-01
-7.08867431e-01 -7.74069905e-01 -1.20668888e+00 -5.34134805e-01
7.11068034e-01 -8.59273851e-01 1.15617609e+00 9.76626426e-02
-1.22095788e+00 1.09165657e+00 -3.21941972e-01 -6.88439846e-01
9.29807961e-01 -6.98105752e-01 -2.04186618e-01 -1.15031786e-01
3.23277682e-01 1.05704582e+00 8.05452704e-01 -1.20661604e+00
-1.09134734e+00 -8.94214809e-01 -1.34731412e-01 1.24538891e-01
9.62034836e-02 -3.06386292e-01 -5.49798965e-01 3.00520081e-02
2.67178148e-01 -1.33755112e+00 -4.67418432e-01 3.22305262e-01
-3.05745572e-01 -1.30019128e-01 9.94996727e-01 -2.92030841e-01
4.68291610e-01 -1.95521986e+00 3.17767845e-03 7.18856528e-02
5.34472585e-01 4.73473594e-02 1.64749682e-01 1.52066380e-01
4.07868773e-01 -3.06762755e-01 -1.78237841e-01 -5.99654019e-01
3.62774700e-01 1.60761371e-01 -2.39802644e-01 7.62344420e-01
3.76039207e-01 1.14543390e+00 -8.71532977e-01 -4.09057677e-01
6.19475484e-01 6.88766360e-01 -5.43084919e-01 -2.52816081e-02
3.42189893e-02 4.15724099e-01 -3.64488035e-01 8.52170825e-01
8.45267832e-01 1.43204764e-01 -2.71436721e-01 -2.26582065e-01
-3.37088913e-01 1.04953721e-01 -1.24918723e+00 1.70829690e+00
-1.55746460e-01 8.32458496e-01 3.54781449e-02 -8.97026956e-01
9.35866952e-01 -1.39021650e-01 5.59835970e-01 -7.59351850e-01
2.61736214e-01 2.92531967e-01 -1.83759391e-01 -4.93416667e-01
7.25861847e-01 1.01132765e-01 -3.88461262e-01 -2.78063072e-03
-1.48186117e-01 -3.33602428e-01 1.18474178e-01 5.68316653e-02
1.09528124e+00 4.67006326e-01 1.52420461e-01 -7.10384995e-02
3.48295301e-01 2.90636450e-01 2.88916558e-01 5.90713441e-01
-6.68371141e-01 4.75468874e-01 1.88765787e-02 -4.81597155e-01
-1.19270504e+00 -1.38754344e+00 -2.55120724e-01 8.55952978e-01
4.36333477e-01 -3.23222816e-01 -4.50840205e-01 -4.72245932e-01
7.12765038e-01 2.38360643e-01 -5.00342071e-01 -9.75658223e-02
-6.35182202e-01 -3.64518493e-01 4.60351408e-01 8.38829517e-01
5.84898889e-01 -5.67999899e-01 -1.34419835e+00 1.72358885e-01
-7.64486119e-02 -1.71754408e+00 -2.57785052e-01 2.31404752e-01
-5.33397615e-01 -1.12466073e+00 -3.53266388e-01 -5.60701191e-01
1.39663458e-01 7.67849088e-01 1.22615302e+00 -3.58651429e-01
-5.72175980e-01 4.78091478e-01 -1.44368008e-01 -7.68023252e-01
7.33419210e-02 -8.22737068e-02 3.40545654e-01 -1.25751734e-01
8.06668341e-01 -6.72285259e-01 -5.78404903e-01 5.78756034e-01
-1.39521062e-02 3.15710381e-02 7.99403548e-01 3.96159887e-01
6.38660848e-01 -2.24165022e-01 1.85655400e-01 -4.51003760e-01
1.09000236e-01 -3.63026410e-01 -6.34832323e-01 -4.97708738e-01
-5.15672505e-01 -7.91089535e-02 1.34831905e-01 -1.01341262e-01
-6.01271093e-01 5.08421957e-01 -1.66285679e-01 -2.84795463e-01
-5.38882375e-01 6.93983287e-02 -3.70532870e-01 -1.69432759e-01
7.27211297e-01 -3.25971916e-02 3.22714150e-01 -4.29539919e-01
7.34071195e-01 6.14475846e-01 1.04641032e+00 -4.45713997e-01
1.19289124e+00 9.86797512e-01 4.94224489e-01 -1.09160221e+00
-8.64541769e-01 -7.30688870e-01 -1.09696805e+00 -3.71878147e-01
8.83102298e-01 -1.20138562e+00 -8.81937802e-01 4.59453702e-01
-9.06790793e-01 -2.84399837e-02 -1.74075201e-01 4.37587172e-01
-6.74092054e-01 2.82344639e-01 -3.41263004e-02 -8.33348393e-01
1.55959144e-01 -1.34203660e+00 1.43789470e+00 -5.02603650e-02
-2.73191690e-01 -4.95588928e-01 -6.20418750e-02 6.90442801e-01
1.80725127e-01 4.45817828e-01 2.84136206e-01 -7.42748752e-02
-8.27574074e-01 -5.34183741e-01 -3.72533113e-01 7.96864405e-02
-2.13106245e-01 -3.85827363e-01 -1.04640222e+00 -1.93407774e-01
-2.78366387e-01 -3.69392723e-01 9.02593791e-01 3.22385281e-01
6.14278436e-01 3.89477134e-01 -5.82426608e-01 5.75979531e-01
8.64945710e-01 -3.16140652e-01 3.50133181e-01 4.36398417e-01
8.64571214e-01 7.89024949e-01 7.34545946e-01 4.49273199e-01
1.12529910e+00 9.77112770e-01 6.67505085e-01 5.96676581e-02
-1.92758873e-01 -4.66749758e-01 2.98739493e-01 4.98530716e-01
-6.71313033e-02 3.40741724e-01 -1.10494745e+00 5.10340154e-01
-1.91897333e+00 -9.19817805e-01 -4.06228423e-01 2.00468564e+00
3.13837498e-01 6.83110178e-01 4.48989391e-01 3.59551013e-01
1.88204274e-01 8.74399170e-02 -7.08873272e-01 3.46619487e-02
8.88716057e-02 7.93048739e-02 8.84446740e-01 4.99063909e-01
-1.29007375e+00 7.46124804e-01 6.22013617e+00 3.22198808e-01
-8.33112061e-01 -1.45973966e-01 9.95593220e-02 -8.22366700e-02
4.91678193e-02 -6.65214807e-02 -1.25198591e+00 3.35471958e-01
8.52569461e-01 7.85078630e-02 1.72880098e-01 1.13154066e+00
2.29306161e-01 -3.37850899e-01 -1.55604708e+00 1.20100033e+00
1.43632919e-01 -1.09766483e+00 -3.55711490e-01 4.10098344e-01
3.09926778e-01 3.28193158e-01 2.37738281e-01 4.71962869e-01
1.59257740e-01 -8.97651851e-01 1.25276804e+00 3.42517048e-01
4.86333132e-01 -8.30627918e-01 5.67377687e-01 7.01866627e-01
-1.52523863e+00 -8.66466984e-02 -2.53867239e-01 -3.77787620e-01
4.26549703e-01 6.42626643e-01 -8.64640951e-01 3.53519559e-01
1.11294842e+00 9.37135816e-01 -9.41514730e-01 1.09083319e+00
-7.65279531e-02 2.58627206e-01 -4.44863349e-01 -2.82499343e-02
3.51551846e-02 -6.55263290e-02 7.31891155e-01 1.05869675e+00
1.34058118e-01 -2.07719952e-01 5.30047834e-01 8.14126313e-01
2.47601271e-01 -4.89117652e-01 -7.98057139e-01 3.50001276e-01
4.60924953e-01 1.24113023e+00 -4.60680515e-01 -1.23378612e-01
-5.19021571e-01 5.93866944e-01 1.16396695e-01 -1.18958168e-01
-9.03860331e-01 -6.92182779e-02 1.05133605e+00 5.69592714e-01
4.15507525e-01 -8.71042550e-01 -5.65745234e-01 -7.85875738e-01
2.74897248e-01 -6.55657709e-01 -1.40046105e-01 -5.99405885e-01
-1.06227601e+00 1.19106412e-01 1.21393159e-01 -1.42792559e+00
-3.35238904e-01 -8.59034657e-01 -2.41359845e-02 5.42064667e-01
-1.77955151e+00 -1.38486016e+00 -6.59670115e-01 5.30651033e-01
6.19901538e-01 -5.97372614e-02 4.34340745e-01 3.93155783e-01
-1.43347800e-01 3.30307871e-01 -4.98824537e-01 -4.12269905e-02
6.23838603e-01 -1.06965041e+00 8.39633644e-01 7.50054121e-01
1.44131139e-01 2.50523746e-01 8.94283235e-01 -5.06655455e-01
-1.97043908e+00 -9.33494687e-01 7.88112581e-01 -1.38284993e+00
5.19877315e-01 -8.37438285e-01 -2.52205878e-01 7.45665550e-01
-2.53311396e-01 1.53958902e-01 3.74551177e-01 1.75022304e-01
-5.39686501e-01 -2.26036340e-01 -9.23646390e-01 5.67349672e-01
1.48821533e+00 -3.48302156e-01 -6.38353765e-01 3.07344124e-02
6.42142653e-01 -5.97855985e-01 -5.04962742e-01 6.42059207e-01
9.25966382e-01 -1.09274268e+00 1.37975538e+00 -1.94120839e-01
9.03050676e-02 -5.81455469e-01 -5.41301966e-01 -9.30982471e-01
-2.13496640e-01 -4.71688688e-01 -2.78480351e-01 6.91132069e-01
1.52460903e-01 -4.85404730e-01 1.15478814e+00 2.78903067e-01
-3.05145264e-01 -6.63866818e-01 -9.95977402e-01 -7.73676038e-01
-2.37916216e-01 -1.02993751e+00 4.93560255e-01 2.32077971e-01
-2.16997966e-01 5.68563640e-01 -3.98456335e-01 3.01738709e-01
1.15617836e+00 -1.06398858e-01 1.51330006e+00 -1.63982832e+00
1.00833885e-01 -3.96986753e-01 -1.08459067e+00 -1.59662676e+00
3.74344154e-03 -6.07825339e-01 2.44672582e-01 -1.25484264e+00
4.62951884e-02 -3.12357038e-01 1.87996954e-01 2.00648904e-01
-9.84547958e-02 5.41638136e-01 2.15741068e-01 -4.34325356e-03
-6.55733228e-01 4.68093872e-01 1.05506122e+00 -2.05525234e-01
1.02827512e-01 1.42902687e-01 -5.97698033e-01 9.13306177e-01
3.06944281e-01 -5.57564087e-02 -2.85769284e-01 -2.90555179e-01
1.63615957e-01 -3.09296608e-01 9.42449093e-01 -1.43882394e+00
3.05388719e-01 -7.34167593e-03 5.74687779e-01 -1.22513103e+00
7.67682970e-01 -1.04885471e+00 -7.97236115e-02 4.50587034e-01
1.62803501e-01 4.55511585e-02 1.53902128e-01 6.77948296e-01
9.85798240e-02 5.83175480e-01 6.80989563e-01 -5.31956693e-03
-1.26592314e+00 4.79092091e-01 -1.42811716e-01 1.78330570e-01
1.13830805e+00 -6.58437312e-01 -1.13599971e-01 -3.77927005e-01
-3.31439555e-01 5.74222088e-01 4.22968090e-01 9.17439222e-01
6.56374812e-01 -1.47831392e+00 -8.16762090e-01 5.20013809e-01
3.81019682e-01 1.76307946e-01 -8.09766632e-03 9.55772102e-01
-2.21231371e-01 5.91909707e-01 -2.25866437e-01 -1.20705950e+00
-1.02001226e+00 2.85291582e-01 1.55374274e-01 2.13895291e-01
-8.42321455e-01 7.57955432e-01 -1.16430476e-01 -6.38091624e-01
1.24776460e-01 -2.63670087e-01 7.64752179e-02 -4.00189608e-02
4.21586812e-01 5.79057932e-01 2.26765096e-01 -1.08979285e+00
-6.21345758e-01 7.71967173e-01 1.05679639e-01 -1.36139005e-01
1.30029321e+00 -3.64976257e-01 3.23603481e-01 4.59477216e-01
1.09506190e+00 3.45568731e-02 -1.73576033e+00 -2.52570719e-01
8.74036402e-02 -6.00470006e-01 -3.98368388e-02 -4.62167472e-01
-6.39203191e-01 1.02610183e+00 8.04850638e-01 1.11798644e-01
3.58767748e-01 3.18832725e-01 9.81819808e-01 3.77983719e-01
8.79306138e-01 -1.03851140e+00 1.11999117e-01 6.30908668e-01
6.35429621e-01 -1.68775547e+00 1.52831659e-01 -5.51490307e-01
-4.84867334e-01 7.50698924e-01 5.99961996e-01 -1.18663266e-01
8.18908513e-01 4.74795103e-01 2.25618798e-02 -2.64383852e-01
-4.89460379e-01 -5.33899188e-01 2.87282646e-01 1.10037851e+00
3.91112752e-02 2.41928414e-01 2.33265743e-01 5.10307968e-01
-8.55571210e-01 -2.41713166e-01 9.12367851e-02 9.72615182e-01
-8.60738993e-01 -9.30185914e-01 -4.68376786e-01 2.20788583e-01
3.74225557e-01 4.25883353e-01 -2.96859473e-01 9.91515815e-01
3.99404705e-01 1.14173651e+00 -5.90941906e-02 -7.63503313e-01
9.39159572e-01 -7.67254159e-02 6.42732382e-01 -4.20810848e-01
9.53564197e-02 -3.12521994e-01 1.16541281e-01 -1.07961023e+00
-3.10967863e-01 -1.10492361e+00 -1.03401351e+00 -5.12785137e-01
-8.91875848e-02 -5.82525551e-01 8.37309718e-01 1.01159453e+00
3.08240682e-01 2.57041156e-01 5.72494030e-01 -1.55765057e+00
-6.30424142e-01 -5.26986301e-01 -3.50598186e-01 3.34310830e-01
4.37279344e-01 -1.09888196e+00 -1.34750992e-01 -4.85759862e-02] | [7.838972568511963, -2.4173150062561035] |
e45a0e05-4f98-4645-868b-6430c31b1498 | all-fragments-count-in-parser-evaluation | null | null | https://aclanthology.org/L14-1324 | https://aclanthology.org/L14-1324.pdf | All Fragments Count in Parser Evaluation | PARSEVAL, the default paradigm for evaluating constituency parsers, calculates parsing success (Precision/Recall) as a function of the number of matching labeled brackets across the test set. Nodes in constituency trees, however, are connected together to reflect important linguistic relations such as predicate-argument and direct-dominance relations between categories. In this paper, we present FREVAL, a generalization of PARSEVAL, where the precision and recall are calculated not only for individual brackets, but also for co-occurring, connected brackets (i.e. fragments). FREVAL fragments precision (FLP) and recall (FLR) interpolate the match across the whole spectrum of fragment sizes ranging from those consisting of individual nodes (labeled brackets) to those consisting of full parse trees. We provide evidence that FREVAL is informative for inspecting relative parser performance by comparing a range of existing parsers. | ["Khalil Sima{'}an", 'Jasmijn Bastings'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['human-parsing'] | ['computer-vision'] | [ 4.76391762e-02 3.82243186e-01 -2.28154719e-01 -6.96648479e-01
-1.19707668e+00 -1.39173269e+00 5.39245009e-01 7.19546676e-01
-3.88401151e-01 7.42426634e-01 5.35511851e-01 -8.14945161e-01
-1.62763879e-01 -9.37925875e-01 -3.37887377e-01 -1.97747037e-01
-1.35598630e-01 3.83254677e-01 6.06154025e-01 -4.31753486e-01
1.91530958e-01 3.20397943e-01 -1.43675542e+00 5.61145246e-01
5.28811455e-01 6.54746234e-01 2.08973885e-01 6.28295600e-01
-5.07977128e-01 5.76205611e-01 -8.46626818e-01 -9.66139972e-01
-1.31573543e-01 -9.09747779e-02 -9.89606440e-01 -8.54083538e-01
4.90921706e-01 1.04610689e-01 3.26732814e-01 1.15481901e+00
1.38052210e-01 -1.73873484e-01 4.21069920e-01 -7.98137486e-01
-2.84897745e-01 1.15856957e+00 -4.04939950e-01 8.51019144e-01
8.50275040e-01 -3.37446958e-01 1.84419298e+00 -6.14042640e-01
7.46547401e-01 1.57094991e+00 7.55524039e-01 1.79279074e-01
-1.35864389e+00 -5.19097269e-01 2.01364994e-01 -4.69348967e-01
-6.30253434e-01 -3.65031481e-01 3.17340314e-01 -5.03534436e-01
1.24838448e+00 4.93974775e-01 2.26073533e-01 5.92997968e-01
3.61576110e-01 2.27240384e-01 1.07930970e+00 -9.79269505e-01
5.11821993e-02 -1.49738878e-01 9.02305722e-01 4.12905902e-01
5.62045872e-01 3.10717195e-01 -3.31184387e-01 -4.41944182e-01
3.67226452e-01 -7.08723962e-01 3.47814672e-02 3.34045082e-01
-9.58415687e-01 1.05069518e+00 1.97578877e-01 8.04804087e-01
-2.19409287e-01 -4.90618870e-02 5.55543065e-01 2.15584189e-01
2.04445779e-01 5.21635413e-01 -8.85377705e-01 -2.50323474e-01
-6.16350532e-01 4.21104401e-01 9.31083798e-01 1.02341580e+00
7.18270600e-01 -4.17832285e-01 -1.19389325e-01 9.49037671e-01
1.59839377e-01 1.61174566e-01 4.01436478e-01 -1.08162475e+00
8.96582127e-01 7.36903846e-01 4.08103913e-02 -6.87313259e-01
-5.43749571e-01 4.68666181e-02 -4.31284606e-02 -1.70392647e-01
8.86970878e-01 -1.39374688e-01 -6.33535147e-01 2.18069100e+00
3.75442892e-01 -9.13239181e-01 2.08067764e-02 3.84265572e-01
1.05642653e+00 5.41188598e-01 1.06138992e+00 -5.65484941e-01
2.06565738e+00 -4.63416129e-02 -5.68657577e-01 -4.67218935e-01
9.35872316e-01 -1.03346682e+00 1.13694334e+00 -5.11834063e-02
-1.24945354e+00 -4.06137437e-01 -8.07618201e-01 -3.32678080e-01
-3.72545213e-01 -3.41540366e-01 9.33199763e-01 8.41749310e-01
-8.85176659e-01 6.14049315e-01 -8.25925350e-01 -1.35381252e-01
-1.55588567e-01 3.16134274e-01 -5.98417759e-01 2.49016345e-01
-1.27296042e+00 9.72222686e-01 5.93681276e-01 -3.24543029e-01
9.81146097e-02 -4.86248285e-01 -1.05601525e+00 2.27676302e-01
-4.72900756e-02 -1.00379281e-01 1.51393938e+00 -5.30623019e-01
-6.31374061e-01 1.28510785e+00 -3.29197556e-01 -6.50319979e-02
-1.39776722e-01 7.14352876e-02 -5.92500806e-01 -5.86666726e-02
6.58599913e-01 5.24432361e-01 -2.06491858e-01 -9.78725135e-01
-1.33386719e+00 -4.86015022e-01 5.67277730e-01 -4.38528843e-02
1.91965073e-01 7.39320695e-01 1.84859544e-01 -3.37726057e-01
8.74538302e-01 -4.91452694e-01 4.07221280e-02 -8.29015136e-01
-2.61622578e-01 -9.43203211e-01 1.76187605e-01 -7.44042575e-01
1.65919304e+00 -2.16276979e+00 -3.81476939e-01 1.22428849e-01
1.91438109e-01 -3.78724486e-02 9.06383395e-02 4.50222075e-01
-5.44455171e-01 5.30549407e-01 -1.46481469e-01 1.55823588e-01
2.00837851e-01 4.53098744e-01 -2.16456875e-01 2.03740388e-01
2.29893193e-01 7.92705953e-01 -9.47525978e-01 -8.00577879e-01
-1.01180390e-01 -4.52443250e-02 -3.75132382e-01 -2.59932399e-01
-3.90278772e-02 -2.35468984e-01 -4.68781888e-01 7.77458310e-01
5.47249198e-01 3.48852098e-01 5.08023322e-01 -8.65160525e-02
-7.33621120e-01 1.46407330e+00 -8.94267917e-01 1.11041832e+00
-3.66245985e-01 3.30803216e-01 1.94321185e-01 -6.94954038e-01
9.28301871e-01 2.68015385e-01 2.22180113e-02 -6.18977368e-01
1.91648915e-01 4.93041396e-01 4.05948102e-01 -1.24098822e-01
8.24988425e-01 -5.15026391e-01 -9.22034383e-01 2.55115598e-01
1.23624057e-01 -1.09024897e-01 9.28282440e-01 1.96367636e-01
1.27334750e+00 1.93674713e-02 9.35454309e-01 -8.44117403e-01
4.53443378e-01 3.59425813e-01 9.35658395e-01 5.39136648e-01
-3.88860941e-01 1.03850991e-01 1.05773222e+00 -4.44799662e-01
-9.74459171e-01 -1.59885979e+00 -8.87665153e-01 1.46313679e+00
-1.08898699e-01 -7.20071793e-01 -6.77056670e-01 -7.51990199e-01
-1.91748768e-01 1.10760677e+00 -4.66787189e-01 2.61089593e-01
-1.09258699e+00 -8.08772564e-01 6.03142500e-01 8.99873376e-01
-1.84151798e-01 -1.27348077e+00 -9.67806339e-01 6.64655983e-01
-4.06928122e-01 -1.07301700e+00 -7.05644414e-02 8.72405469e-01
-7.95202851e-01 -1.40094042e+00 3.14925849e-01 -1.16884947e+00
1.43021032e-01 -1.93658918e-01 1.83385921e+00 1.59764498e-01
4.27135825e-02 -4.05590497e-02 -6.46894455e-01 -1.41421914e-01
-8.66860092e-01 -1.66843668e-01 -2.61167824e-01 -1.17681360e+00
8.57101381e-01 -6.14054561e-01 -6.76363660e-03 7.99575672e-02
-4.13412929e-01 -5.08790493e-01 1.42942145e-01 7.71625519e-01
7.37531960e-01 -1.37023866e-01 4.73494709e-01 -1.27319062e+00
6.70432210e-01 -4.16076064e-01 -6.31995261e-01 4.44458514e-01
-1.67028040e-01 -6.06203498e-03 4.21019316e-01 -1.66591167e-01
-9.26334977e-01 -1.76247969e-01 -4.74929303e-01 6.51390910e-01
-3.74831885e-01 4.26953614e-01 -2.47891128e-01 5.86437762e-01
8.88924956e-01 -3.98512393e-01 -6.46995306e-01 -3.64936799e-01
3.67034316e-01 5.40764749e-01 8.36840212e-01 -1.14465535e+00
2.66983449e-01 -2.13503957e-01 -1.45084888e-01 -2.44345084e-01
-8.13305318e-01 -3.70793700e-01 -7.26614058e-01 8.48128796e-02
7.39347994e-01 -6.09974146e-01 -6.49296343e-01 -4.42308813e-01
-1.32236445e+00 2.03922972e-01 -4.21537161e-01 5.43154478e-01
-3.82608771e-01 3.17476809e-01 -8.96301687e-01 -8.05227757e-01
-3.79724175e-01 -8.76425445e-01 9.77718174e-01 1.94412664e-01
-7.51761317e-01 -9.51202810e-01 1.28399402e-01 -1.74725682e-01
-2.13370889e-01 4.10129815e-01 1.49496472e+00 -9.46488798e-01
2.93010145e-01 -3.49064052e-01 -1.94286793e-01 -5.90919182e-02
1.95820242e-01 6.94214627e-02 -6.50815547e-01 7.86285400e-02
1.16686458e-02 -3.12845670e-02 5.90034902e-01 3.91920388e-01
2.23644316e-01 -1.76765457e-01 -5.64308405e-01 2.65730964e-03
1.37749326e+00 4.41708535e-01 5.22919536e-01 5.21709979e-01
1.29116118e-01 9.59861457e-01 6.24898374e-01 -8.03009421e-02
5.74670374e-01 2.41990760e-01 7.17423931e-02 4.62861896e-01
-3.08502875e-02 -3.83744895e-01 3.05463344e-01 5.99746048e-01
8.03388804e-02 -3.98821570e-02 -1.02848494e+00 6.94038570e-01
-1.15166211e+00 -8.90192986e-01 -5.65576136e-01 2.06280661e+00
1.01062489e+00 7.00475276e-01 1.99110061e-01 5.46617806e-01
1.08616614e+00 4.75941561e-02 1.87726542e-01 -1.00067449e+00
-9.61225256e-02 7.84757316e-01 2.58478612e-01 7.01576889e-01
-1.12488627e+00 1.19326425e+00 7.45965767e+00 5.81239641e-01
-4.19191152e-01 1.41785875e-01 4.22726929e-01 2.76507288e-01
-4.78421867e-01 4.39635247e-01 -1.11292899e+00 3.50112528e-01
1.17930555e+00 -7.28614852e-02 -1.52806425e-02 9.64095294e-01
-3.89207751e-01 -4.74077582e-01 -1.28319478e+00 2.24584818e-01
-6.73196733e-01 -7.87839949e-01 -3.73707533e-01 -7.35084340e-02
9.33302790e-02 -1.31528780e-01 -4.74268407e-01 5.19868970e-01
9.24646795e-01 -5.56240380e-01 1.14759457e+00 -4.21397597e-01
9.72077966e-01 -6.07976913e-01 7.84302354e-01 2.28527427e-01
-1.63296044e+00 1.45510724e-02 -4.21079248e-01 -3.65909159e-01
3.76510531e-01 6.94892526e-01 -5.90091884e-01 2.78188944e-01
8.83223653e-01 4.82416339e-03 -2.85567731e-01 3.70916963e-01
-5.17100692e-01 7.30679035e-01 -6.38541818e-01 -5.62979514e-03
1.41377732e-01 -3.61369625e-02 5.14136434e-01 1.59263635e+00
6.56231195e-02 4.43823576e-01 7.65523463e-02 7.68436790e-01
3.23250920e-01 3.87364745e-01 -2.87879974e-01 1.15139917e-01
1.17457211e+00 1.19112575e+00 -1.07565463e+00 -2.58058012e-01
-5.38812578e-01 -1.24538764e-01 7.09434450e-01 -2.60497242e-01
-4.90527868e-01 -3.81928355e-01 6.44776821e-01 6.12872764e-02
3.84739526e-02 -9.91917327e-02 -6.11094892e-01 -5.54883540e-01
2.30722561e-01 -4.63546067e-01 9.79883432e-01 -2.67827094e-01
-1.27799797e+00 7.50697315e-01 4.07673001e-01 -7.25474238e-01
-4.73918468e-01 -7.54420042e-01 -6.46394134e-01 1.09554935e+00
-1.13473856e+00 -8.21051002e-01 3.29358250e-01 1.53186291e-01
2.03768894e-01 2.91692346e-01 1.19872391e+00 7.91053623e-02
-3.34580988e-01 7.49946117e-01 -5.43504357e-01 4.20716643e-01
1.66783124e-01 -1.53459358e+00 7.51930118e-01 8.43355656e-01
1.63172767e-01 1.00539470e+00 8.50693882e-01 -6.82360530e-01
-4.37334955e-01 -5.49486637e-01 1.64970624e+00 -5.49445748e-01
7.29138911e-01 -1.58396274e-01 -6.59845710e-01 9.81013656e-01
-2.34810129e-01 -1.09410450e-01 7.98700511e-01 8.45181942e-01
-7.61423469e-01 2.94391125e-01 -1.41570032e+00 3.20966244e-01
1.11804438e+00 -4.15018082e-01 -1.26112437e+00 1.22508951e-01
7.49699414e-01 -5.03662109e-01 -1.17046010e+00 5.76546073e-01
5.38444579e-01 -1.31833875e+00 7.38120079e-01 -5.91873229e-01
4.71894592e-01 -3.87167819e-02 -6.65154636e-01 -9.98844862e-01
-7.35064209e-01 -1.54018506e-01 3.36351067e-01 1.59291375e+00
8.34185719e-01 -5.66505790e-01 4.37903255e-01 8.10139835e-01
-6.24776423e-01 -3.33328068e-01 -1.29673183e+00 -4.97076452e-01
5.71336389e-01 -5.86636007e-01 7.27562785e-01 8.89467537e-01
6.49925470e-01 3.92823040e-01 7.11390853e-01 1.72956288e-02
3.82554382e-01 2.02169359e-01 -1.05790205e-01 -1.61144590e+00
-2.03605741e-01 -4.33464766e-01 -4.92920905e-01 -5.69346488e-01
1.02390788e-01 -1.03036165e+00 9.18122604e-02 -1.34526527e+00
1.09528163e-02 -9.30454731e-01 1.75643235e-01 5.74044108e-01
-3.99760157e-01 9.63034630e-02 1.80074379e-01 1.65441275e-01
-9.83457863e-02 -4.60010052e-01 8.20775151e-01 3.46694529e-01
-2.58865476e-01 -2.22236767e-01 -1.07311964e+00 1.08637059e+00
8.67184877e-01 -7.91538358e-01 1.24141619e-01 -3.69889319e-01
3.02148610e-01 3.57172549e-01 -9.59057063e-02 -6.40854657e-01
-1.49808004e-01 -1.61701128e-01 2.49213442e-01 -6.64510429e-01
-1.19141124e-01 -3.95905733e-01 2.60952767e-02 2.84970552e-01
-3.42250496e-01 7.07321107e-01 2.52940238e-01 9.30873081e-02
-1.33919656e-01 -6.54114723e-01 5.02946079e-01 -4.50192660e-01
-5.64754486e-01 -4.93929982e-01 -2.39342645e-01 4.40320224e-01
7.14939713e-01 -2.38963246e-01 -6.98769033e-01 4.88197148e-01
-8.49603891e-01 1.22085270e-02 1.27012983e-01 1.38137713e-01
1.06599197e-01 -1.03149402e+00 -6.56960607e-01 -1.21036265e-02
2.41386011e-01 -1.38448358e-01 -3.23690236e-01 1.49524808e-01
-6.00486219e-01 4.62704450e-01 2.15215757e-02 -2.76643842e-01
-1.20934534e+00 4.02036518e-01 1.15934126e-01 -5.43453038e-01
-6.13690674e-01 1.05031562e+00 2.49630421e-01 -4.42570955e-01
-1.44481789e-02 -5.46099424e-01 -4.33499724e-01 1.58133224e-01
5.11395454e-01 1.03437945e-01 9.41271782e-02 -9.23138142e-01
-6.28139973e-01 3.84837478e-01 -8.92063379e-02 -2.16787711e-01
9.88866568e-01 5.91651984e-02 -5.55495799e-01 4.55813944e-01
8.28014553e-01 3.61780047e-01 -5.92600822e-01 -1.02194650e-02
6.12956405e-01 -2.25712225e-01 -2.78621405e-01 -8.83612931e-01
-4.12861407e-01 3.72561693e-01 1.42968863e-01 6.51280046e-01
8.71260464e-01 5.91485202e-01 5.32485068e-01 -7.16039166e-02
7.29332089e-01 -9.32083189e-01 -6.63111448e-01 6.14741743e-01
4.75381106e-01 -8.22430134e-01 -7.88599402e-02 -9.98181760e-01
-2.01540500e-01 1.04761791e+00 3.74756545e-01 -1.39072850e-01
5.38368523e-01 6.92789435e-01 2.75710877e-02 -2.38444746e-01
-7.07542121e-01 -2.27094889e-01 -1.05395161e-01 5.82930923e-01
1.29842854e+00 5.82819641e-01 -1.29670107e+00 1.18158162e+00
-1.20727670e+00 -8.26124847e-01 3.01559478e-01 1.00363290e+00
-8.88660610e-01 -1.44176698e+00 -5.76788723e-01 5.76598644e-01
-9.24735785e-01 -4.47688788e-01 -1.66700691e-01 1.03226769e+00
3.97031635e-01 1.39690173e+00 4.39829022e-01 -3.02004158e-01
6.58244789e-01 2.76429325e-01 5.94417036e-01 -8.56829107e-01
-9.63949025e-01 -6.83445334e-02 7.58428454e-01 -2.50552565e-01
-3.72460723e-01 -1.03866529e+00 -1.56758666e+00 -3.16142768e-01
-7.19691455e-01 4.75944012e-01 5.21106541e-01 9.14994717e-01
-3.88663232e-01 3.23282212e-01 2.10900158e-01 -3.70915204e-01
-6.05954289e-01 -1.24100721e+00 -7.80806363e-01 3.17975163e-01
-7.00360686e-02 -8.47454250e-01 -5.96650422e-01 -1.68569908e-01] | [10.384833335876465, 9.602071762084961] |
969bba10-ac22-4a7c-9355-fd3a5f018619 | who-s-waldo-linking-people-across-text-and | 2108.07253 | null | https://arxiv.org/abs/2108.07253v2 | https://arxiv.org/pdf/2108.07253v2.pdf | Who's Waldo? Linking People Across Text and Images | We present a task and benchmark dataset for person-centric visual grounding, the problem of linking between people named in a caption and people pictured in an image. In contrast to prior work in visual grounding, which is predominantly object-based, our new task masks out the names of people in captions in order to encourage methods trained on such image-caption pairs to focus on contextual cues (such as rich interactions between multiple people), rather than learning associations between names and appearances. To facilitate this task, we introduce a new dataset, Who's Waldo, mined automatically from image-caption data on Wikimedia Commons. We propose a Transformer-based method that outperforms several strong baselines on this task, and are releasing our data to the research community to spur work on contextual models that consider both vision and language. | ['Hadar Averbuch-Elor', 'Noah Snavely', 'Yoav Artzi', 'Apoorv Khandelwal', 'Claire Yuqing Cui'] | 2021-08-16 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Cui_Whos_Waldo_Linking_People_Across_Text_and_Images_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Cui_Whos_Waldo_Linking_People_Across_Text_and_Images_ICCV_2021_paper.pdf | iccv-2021-1 | ['person-centric-visual-grounding'] | ['computer-vision'] | [ 2.22241908e-01 2.66586483e-01 -7.23140389e-02 -6.61908984e-01
-5.67125916e-01 -5.73744535e-01 1.12466359e+00 1.19416364e-01
-4.78156269e-01 6.39141619e-01 9.05042112e-01 -1.48130208e-01
4.17836577e-01 -3.99221659e-01 -1.02356398e+00 -8.29804912e-02
2.47905433e-01 6.97722435e-01 5.67244403e-02 -6.57081679e-02
7.98655301e-02 -1.77571643e-02 -1.50484288e+00 6.82192385e-01
3.41559231e-01 4.48911041e-01 8.72438028e-02 5.40346980e-01
-9.66840610e-02 1.06229997e+00 -2.82811105e-01 -9.86999333e-01
2.07583532e-01 -6.83982730e-01 -1.06970775e+00 6.08058751e-01
1.67440391e+00 -1.08381793e-01 -4.65427101e-01 1.02120912e+00
3.30465347e-01 1.01667963e-01 7.43784904e-01 -1.65627408e+00
-1.35398543e+00 6.40574753e-01 -6.01614356e-01 2.90410727e-01
4.21243012e-01 3.70349258e-01 1.30769801e+00 -8.62805009e-01
9.20485675e-01 1.52809846e+00 5.87685347e-01 8.15428138e-01
-1.42287278e+00 -5.74026525e-01 5.70287824e-01 9.05381143e-02
-1.18111634e+00 -6.26212001e-01 4.03220385e-01 -9.19320226e-01
6.10008299e-01 1.78931192e-01 6.87702835e-01 1.60833669e+00
-5.80554664e-01 8.16203535e-01 1.22246456e+00 -5.53290606e-01
-3.36827457e-01 3.14171672e-01 1.84383065e-01 7.96362102e-01
5.18291652e-01 -3.26270126e-02 -8.39843392e-01 -7.70368427e-02
9.43192065e-01 -1.28852516e-01 -2.52946287e-01 -6.15823567e-01
-1.76560509e+00 7.56983280e-01 8.15367579e-01 2.05469966e-01
-3.13505828e-01 6.49057448e-01 -9.42433178e-02 -1.57569781e-01
4.71435368e-01 6.59714162e-01 -2.75642588e-03 2.66234428e-01
-9.67423677e-01 6.13679349e-01 4.60218996e-01 1.15055573e+00
9.06521082e-01 -4.72062737e-01 -6.69743657e-01 7.06242800e-01
4.46541518e-01 5.77155292e-01 -1.94103867e-01 -9.85386431e-01
6.87165380e-01 7.35430360e-01 5.38945138e-01 -1.10133183e+00
-1.91884503e-01 -1.23249881e-01 -3.88951838e-01 9.83826909e-03
7.27568865e-01 -8.75353292e-02 -1.15278745e+00 2.10302377e+00
2.96755999e-01 4.20858413e-01 -2.67445326e-01 1.08722758e+00
1.29023695e+00 2.31593460e-01 5.11874795e-01 4.46768880e-01
1.90258896e+00 -1.13493335e+00 -5.35862327e-01 -7.84357131e-01
4.45726097e-01 -7.40818024e-01 1.07703316e+00 -2.69073516e-01
-1.09536362e+00 -5.93960464e-01 -5.92709661e-01 -3.28643411e-01
-5.63275576e-01 1.03644125e-01 7.08384156e-01 2.71696657e-01
-1.40019393e+00 1.49301708e-01 -3.73079479e-01 -1.20438123e+00
6.79793537e-01 -1.42336637e-01 -6.97918117e-01 -1.16261229e-01
-9.55053210e-01 1.12039387e+00 1.49519891e-01 -1.09489538e-01
-8.64836752e-01 -6.88151836e-01 -1.06052756e+00 -5.06496727e-01
1.38668865e-01 -1.40517652e+00 1.21030009e+00 -1.42992067e+00
-4.98879999e-01 1.97610390e+00 -3.34420025e-01 -5.10808110e-01
6.29207253e-01 -3.15104067e-01 -4.08536285e-01 1.46543577e-01
5.49925029e-01 1.43975067e+00 7.59485722e-01 -1.70150495e+00
-7.56631792e-01 -2.21093118e-01 4.93971348e-01 3.20948154e-01
-2.22765833e-01 3.78838241e-01 -6.63554668e-01 -4.31122005e-01
-3.28631759e-01 -1.04849148e+00 -1.64007768e-02 2.32768163e-01
-6.96193159e-01 -3.19158286e-01 4.09018993e-01 -7.82065868e-01
6.39887393e-01 -1.81089389e+00 5.20174205e-02 -1.40171483e-01
3.75935674e-01 1.95608828e-02 -4.14516956e-01 4.56967086e-01
-1.07555501e-01 1.50073364e-01 -3.78507301e-02 -7.69666493e-01
1.11404695e-01 1.52453199e-01 -4.29223984e-01 5.57542324e-01
4.31888074e-01 1.31330454e+00 -1.21832430e+00 -8.82873893e-01
1.07590422e-01 3.28357369e-01 -4.01384711e-01 4.30732936e-01
-4.07914430e-01 5.68461657e-01 -1.52274191e-01 5.46182752e-01
4.35754836e-01 -6.40999794e-01 -5.90465553e-02 -2.79620111e-01
3.28638442e-02 1.19916014e-01 -7.74825811e-01 1.65949440e+00
-3.88214141e-01 9.06392455e-01 3.38734575e-02 -5.44209838e-01
5.40454268e-01 3.54699790e-01 1.17174506e-01 -5.84155738e-01
-1.34493604e-01 -2.69278824e-01 -3.82392049e-01 -7.15131342e-01
4.54774529e-01 -5.08018918e-02 -9.63399857e-02 4.48053241e-01
8.21187794e-02 -8.05479437e-02 2.13886350e-01 7.76548803e-01
6.64714515e-01 3.90096396e-01 -8.08282010e-03 -1.81993380e-01
2.10828707e-01 2.39344969e-01 9.31386650e-02 9.89880264e-01
-3.60540122e-01 9.10746574e-01 1.40377238e-01 -6.40587926e-01
-1.33122122e+00 -1.01029241e+00 3.07068527e-02 1.45769370e+00
2.80243874e-01 -5.26406765e-01 -6.13224626e-01 -7.96261489e-01
1.16894051e-01 7.05124617e-01 -1.04831421e+00 4.97163087e-01
-5.90028346e-01 -2.68832386e-01 4.06232506e-01 5.46592534e-01
3.79928112e-01 -1.21159148e+00 -2.62973934e-01 -1.70338139e-01
-5.75258851e-01 -1.52448142e+00 -8.70216608e-01 -4.08235520e-01
-2.64716417e-01 -1.21526694e+00 -8.17333281e-01 -1.05718911e+00
8.75944495e-01 3.54797661e-01 2.01100039e+00 4.08745587e-01
-2.97979295e-01 9.97776568e-01 -1.88033059e-01 -5.95031142e-01
-1.62697583e-01 -1.44430548e-01 -1.02722801e-01 3.18229765e-01
5.78243315e-01 -2.01868162e-01 -7.47882843e-01 1.59751281e-01
-5.74719965e-01 5.81172109e-01 3.66582721e-01 5.84748685e-01
1.26568407e-01 -8.55667889e-01 4.01470847e-02 -1.01351035e+00
1.29017740e-01 -5.19149363e-01 -3.27490091e-01 4.40157771e-01
-1.44118257e-02 8.33637733e-03 -2.47828156e-01 -2.54565656e-01
-9.25886214e-01 1.87146366e-01 3.78955603e-01 -3.49676192e-01
-4.24822658e-01 -4.97304387e-02 6.79083169e-02 1.23497881e-01
7.02319562e-01 -6.63537458e-02 -1.43262357e-01 -3.04219365e-01
9.91434157e-01 4.74482864e-01 1.04498971e+00 -7.49962509e-01
1.01333475e+00 7.24652648e-01 -1.65694371e-01 -6.88674033e-01
-1.45938182e+00 -7.33227730e-01 -6.66718125e-01 -3.74203414e-01
1.42242324e+00 -1.46720946e+00 -4.86220390e-01 5.67005277e-02
-1.59710228e+00 -3.43966842e-01 -6.50125043e-03 1.93009883e-01
-4.94711220e-01 1.36281535e-01 -3.24252367e-01 -6.76166832e-01
-2.65858304e-02 -6.11801505e-01 1.34847713e+00 2.80054182e-01
-5.21937668e-01 -1.18987083e+00 5.04853390e-02 8.42693925e-01
1.84734076e-01 6.86752975e-01 5.07119596e-01 -5.67857087e-01
-8.28333497e-01 8.78076702e-02 -6.86624646e-01 5.19733429e-02
-7.24469647e-02 -1.21513516e-01 -9.69725430e-01 -1.93806276e-01
-6.30278468e-01 -7.04530239e-01 9.84290957e-01 1.15786664e-01
7.10101902e-01 -4.72452968e-01 -7.17591703e-01 4.34897900e-01
1.31468832e+00 -6.93479538e-01 5.23018777e-01 4.31338489e-01
1.04281461e+00 9.97544646e-01 2.40326822e-01 5.98623455e-02
9.39257085e-01 8.38013411e-01 5.33824801e-01 -7.26655960e-01
-7.29515791e-01 -6.56067133e-01 2.58984804e-01 1.55397896e-02
-1.77884758e-01 -2.17053637e-01 -1.10011923e+00 1.09176326e+00
-2.18051338e+00 -1.29003608e+00 -4.35624033e-01 1.88091397e+00
9.24243450e-01 -2.96028435e-01 3.42442244e-01 -9.59020972e-01
1.24585235e+00 2.70598054e-01 -9.63952988e-02 2.68400293e-02
-2.10437834e-01 -1.66048348e-01 3.97443861e-01 4.79162872e-01
-1.41887188e+00 1.14295506e+00 6.63928986e+00 3.54167446e-02
-5.79398155e-01 2.22303286e-01 6.87918007e-01 -1.28021136e-01
-4.11850065e-01 1.40912548e-01 -1.03454196e+00 4.34996784e-01
5.97517550e-01 2.12744579e-01 3.77606153e-01 5.03089070e-01
8.52780715e-02 3.31145562e-02 -1.65710413e+00 1.33440590e+00
5.17640889e-01 -1.34578574e+00 2.52478689e-01 4.61974152e-04
1.08113086e+00 8.52906257e-02 1.74289331e-01 -1.02845751e-01
9.36286628e-01 -1.37993443e+00 1.16971409e+00 6.25416160e-01
5.72941124e-01 -1.81307942e-01 4.94188011e-01 -2.10187510e-01
-9.78955269e-01 1.73396721e-01 -1.84900343e-01 -2.42528841e-01
4.03743416e-01 2.46513069e-01 -9.74257112e-01 1.47500247e-01
7.81229615e-01 9.76923048e-01 -1.19330585e+00 1.09924924e+00
-5.12075365e-01 2.71427423e-01 1.20956533e-01 8.90555158e-02
4.29758489e-01 -1.23785131e-01 3.37884158e-01 1.38567138e+00
1.13337869e-02 -1.30709767e-01 3.26071799e-01 1.16429424e+00
-2.23199859e-01 -3.53346695e-03 -8.92678559e-01 -1.60370991e-01
4.24732089e-01 1.44892871e+00 -5.77807128e-01 -4.52381670e-01
-7.27573693e-01 1.05785894e+00 5.98780870e-01 5.63345969e-01
-8.93549085e-01 1.27803236e-01 7.68764973e-01 6.09334826e-01
2.71335542e-01 -1.72142461e-01 -5.68003394e-02 -1.16967821e+00
-1.01367153e-01 -8.09195399e-01 6.06246412e-01 -1.42946959e+00
-1.82455885e+00 4.77893323e-01 2.38102395e-02 -8.60867321e-01
-9.74378064e-02 -5.62469900e-01 -5.15865505e-01 9.24591959e-01
-1.43738329e+00 -2.05517006e+00 -6.84273183e-01 7.98923314e-01
1.96925879e-01 1.03398807e-01 5.08235455e-01 5.71801327e-03
-2.36091942e-01 3.57896715e-01 -5.32780051e-01 5.27339816e-01
1.19499028e+00 -1.34575379e+00 8.13318491e-01 1.02325070e+00
8.86585355e-01 8.57339203e-01 1.09290445e+00 -6.53390408e-01
-8.59487593e-01 -1.34087110e+00 1.30826509e+00 -1.61921668e+00
7.01682985e-01 -7.42642462e-01 -4.85801488e-01 1.18664134e+00
7.97831833e-01 1.35134652e-01 5.22013009e-01 4.99227077e-01
-9.35923040e-01 3.46985906e-01 -7.04110563e-01 8.54089797e-01
1.54793561e+00 -8.33125532e-01 -9.70750570e-01 7.89940774e-01
6.56170964e-01 -2.38002956e-01 -2.92559981e-01 -7.83774406e-02
2.76781917e-01 -7.40578532e-01 1.35932028e+00 -1.11412382e+00
4.78707612e-01 -4.23870385e-01 -4.81767431e-02 -1.16107631e+00
-4.19535220e-01 -5.41349411e-01 3.04968029e-01 1.67707467e+00
3.67371738e-01 -1.66184947e-01 6.37570798e-01 8.46549273e-01
1.05904184e-01 7.27884769e-02 -5.90084255e-01 -6.04809523e-01
-4.88479994e-02 -9.07824785e-02 4.85583901e-01 1.11732686e+00
-3.07949990e-01 6.84267282e-01 -7.54270196e-01 2.55193472e-01
7.86808908e-01 1.25175282e-01 1.13109064e+00 -1.16795099e+00
-1.69357449e-01 -1.36290520e-01 -3.95913273e-01 -6.87860250e-01
2.77922422e-01 -9.60309982e-01 3.00175160e-01 -2.00850439e+00
8.75460386e-01 -2.49155968e-01 -1.87446594e-01 7.34255791e-01
-5.87718308e-01 8.55702400e-01 2.87735522e-01 1.51893988e-01
-1.19112504e+00 1.81136042e-01 1.25859237e+00 -3.24095607e-01
2.49187052e-01 -5.22819817e-01 -1.03653729e+00 7.03887761e-01
1.90391585e-01 -4.53758687e-01 -2.72876173e-01 -8.12937021e-01
4.18065041e-01 -5.00165164e-01 1.12566984e+00 -8.73654366e-01
2.22503483e-01 -3.27482224e-01 6.18071556e-01 -2.97088414e-01
3.36594433e-01 -5.35908639e-01 1.29538044e-01 2.50064228e-02
-6.17485642e-01 3.45904648e-01 -3.99734676e-02 5.94442010e-01
2.96156090e-02 4.01446223e-02 6.23630881e-01 -6.54828250e-01
-1.02553296e+00 4.30359662e-01 1.02336772e-01 5.51538765e-01
9.10669267e-01 -5.31004667e-02 -7.72548854e-01 -7.09998012e-01
-6.95776701e-01 4.04297233e-01 9.70879793e-01 6.95299685e-01
3.77476543e-01 -1.58531582e+00 -1.08596087e+00 -4.23663199e-01
8.21248591e-01 -3.77621472e-01 1.04618639e-01 6.22430921e-01
-3.91875356e-01 4.61586654e-01 -2.36758351e-01 -5.74608862e-01
-1.40425491e+00 7.54779398e-01 2.85969257e-01 3.53009522e-01
-6.26894414e-01 1.21947455e+00 6.28325582e-01 -2.23171085e-01
1.30209312e-01 1.83060482e-01 -1.90373406e-01 1.47926271e-01
7.30103135e-01 -1.74092337e-01 -4.45418656e-01 -1.22537673e+00
-5.39157867e-01 5.61132789e-01 1.15860924e-01 -3.96158278e-01
1.26298988e+00 -3.68814558e-01 -2.43413180e-01 3.64521891e-01
9.64718699e-01 -3.97163676e-03 -1.38106370e+00 -4.71536070e-01
6.05042800e-02 -6.76916957e-01 -2.63015747e-01 -9.25547481e-01
-6.53960288e-01 8.37587178e-01 4.15787339e-01 4.83699962e-02
3.90029728e-01 7.67569423e-01 4.96720970e-01 1.54258788e-01
2.08711877e-01 -8.72530878e-01 6.65944338e-01 2.54560381e-01
1.04175937e+00 -1.61434495e+00 7.70555809e-02 -3.06219965e-01
-9.11317408e-01 6.55640721e-01 9.93741035e-01 3.55466977e-02
7.28318915e-02 -3.48814189e-01 4.03188944e-01 -4.11805242e-01
-6.80972874e-01 -9.25242722e-01 7.06100225e-01 1.02125120e+00
5.40431798e-01 5.92471249e-02 2.80589849e-01 2.44376615e-01
-1.60518736e-01 -2.33580694e-01 4.32209373e-01 5.83613515e-01
-2.34579161e-01 -1.09870315e+00 -5.27221382e-01 1.95341691e-01
-1.50938913e-01 -4.84464556e-01 -8.64828825e-01 7.97489524e-01
4.44152474e-01 1.02438414e+00 4.86888856e-01 6.61914498e-02
1.54711038e-01 8.17071125e-02 7.25438118e-01 -1.07892382e+00
-4.84969169e-01 -5.10584176e-01 5.15503645e-01 -6.15187466e-01
-8.35093975e-01 -7.47913539e-01 -6.35411799e-01 -8.15735757e-02
1.49453849e-01 -9.39213261e-02 4.31423038e-01 1.01675975e+00
2.37789303e-01 5.82150929e-02 -8.55629370e-02 -6.95798218e-01
1.51891530e-01 -7.95550704e-01 -2.87530124e-01 1.41253054e+00
3.72063041e-01 -7.28955150e-01 -1.58258468e-01 5.92476308e-01] | [10.729677200317383, 1.5873280763626099] |
2ead8087-e1f6-45b0-bbeb-f0cf3a906ef8 | learning-from-context-agnostic-synthetic-data | 2005.14707 | null | https://arxiv.org/abs/2005.14707v3 | https://arxiv.org/pdf/2005.14707v3.pdf | Towards Context-Agnostic Learning Using Synthetic Data | We propose a novel setting for learning, where the input domain is the image of a map defined on the product of two sets, one of which completely determines the labels. We derive a new risk bound for this setting that decomposes into a bias and an error term, and exhibits a surprisingly weak dependence on the true labels. Inspired by these results, we present an algorithm aimed at minimizing the bias term by exploiting the ability to sample from each set independently. We apply our setting to visual classification tasks, where our approach enables us to train classifiers on datasets that consist entirely of a single synthetic example of each class. On several standard benchmarks for real-world image classification, we achieve robust performance in the context-agnostic setting, with good generalization to real world domains, whereas training directly on real world data without our techniques yields classifiers that are brittle to perturbations of the background. | ['Martin Rinard', 'Charles Jin'] | 2020-05-29 | towards-context-agnostic-learning-using | http://proceedings.neurips.cc/paper/2021/hash/dccb1c3a558c50d389c24d69a9856730-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/dccb1c3a558c50d389c24d69a9856730-Paper.pdf | neurips-2021-12 | ['traffic-sign-recognition'] | ['computer-vision'] | [ 6.08279407e-01 1.68489128e-01 -5.34582622e-02 -3.85438561e-01
-8.65519881e-01 -9.43848968e-01 8.76301408e-01 8.75456184e-02
-5.33041656e-01 4.68295008e-01 -2.97127128e-01 -2.01762274e-01
1.94193393e-01 -7.87688196e-01 -1.20591938e+00 -9.39845979e-01
-8.14866200e-02 5.31782568e-01 4.48320985e-01 8.74776989e-02
8.09345841e-02 3.08763266e-01 -1.47230363e+00 2.60330975e-01
5.12775123e-01 1.05251682e+00 -1.52405426e-01 7.39748359e-01
2.66843468e-01 5.83178699e-01 -6.32128894e-01 -3.98261458e-01
6.01244926e-01 -3.54443789e-01 -8.76589775e-01 4.50445473e-01
7.73065925e-01 7.25240782e-02 -1.68175817e-01 1.32310951e+00
2.14597777e-01 -5.25165675e-03 1.08665884e+00 -1.24206996e+00
-5.71462512e-01 3.46311331e-01 -5.44314086e-01 1.06386647e-01
8.86870623e-02 1.68243498e-01 1.07585776e+00 -7.07963884e-01
5.72234213e-01 1.12751067e+00 6.57985389e-01 5.95715165e-01
-1.84826505e+00 -5.19796312e-01 1.87450305e-01 -8.78546685e-02
-1.18970203e+00 -3.17264467e-01 5.06125867e-01 -8.10340166e-01
3.08883786e-01 1.61568418e-01 1.01819851e-01 1.38629854e+00
-5.83037082e-03 6.12011671e-01 1.47609127e+00 -6.48565888e-01
5.53015590e-01 5.61651468e-01 2.23504856e-01 6.65619612e-01
1.44939944e-01 3.01664323e-01 -8.45955312e-02 -1.97802857e-01
4.21363086e-01 -4.58072014e-02 -3.30117732e-01 -9.49757040e-01
-1.17804563e+00 9.52110350e-01 4.96296436e-01 -8.76681209e-02
1.36155680e-01 9.50336456e-02 3.88645828e-01 3.91137332e-01
3.70142788e-01 3.07612926e-01 -2.66760170e-01 4.06517237e-01
-5.16054511e-01 1.84868723e-01 9.18941259e-01 8.74738276e-01
8.66692126e-01 -4.41473931e-01 -1.39511600e-01 6.77441895e-01
-1.37736782e-01 5.64289391e-01 3.63889068e-01 -9.53836739e-01
3.21103990e-01 2.19125822e-01 3.03510636e-01 -7.13356614e-01
-2.02731162e-01 -4.59626377e-01 -6.25417888e-01 5.79533219e-01
9.78801250e-01 -5.98152354e-02 -9.73079681e-01 2.17637467e+00
2.27098152e-01 2.86942571e-01 6.65693507e-02 6.86121285e-01
1.89309284e-01 4.96578962e-01 -8.15635100e-02 8.19363818e-02
1.09053326e+00 -8.88601542e-01 -5.96140958e-02 -3.15690279e-01
6.94260776e-01 -2.40326613e-01 1.37624919e+00 5.23875773e-01
-7.66390443e-01 -3.58331770e-01 -1.18615294e+00 2.16704398e-01
-4.40960974e-01 2.01078728e-01 1.96877524e-01 6.57287836e-01
-9.12137151e-01 5.33864796e-01 -5.96496820e-01 -3.81969869e-01
5.16314209e-01 1.17140077e-01 -5.36459744e-01 -1.47617042e-01
-8.01812649e-01 9.19642329e-01 4.01470631e-01 -3.49746585e-01
-1.29735196e+00 -5.41368484e-01 -7.51153886e-01 -1.68503612e-01
6.08758986e-01 -4.90867913e-01 1.22049975e+00 -1.30788910e+00
-1.10762262e+00 1.22301662e+00 6.14832751e-02 -4.64992315e-01
9.61741626e-01 7.39061683e-02 3.92066464e-02 1.04399361e-01
1.93667561e-01 4.74370420e-01 1.18006206e+00 -1.51105976e+00
-6.04893208e-01 -2.66714752e-01 2.64395058e-01 -1.47179082e-01
-3.15989882e-01 -2.59797424e-01 -1.84060201e-01 -5.76002419e-01
-3.36969942e-01 -1.05236733e+00 -1.73286691e-01 3.30278784e-01
-5.88081777e-01 8.53996724e-02 6.39385045e-01 -2.05628857e-01
5.90460241e-01 -2.36767054e+00 2.39872262e-01 2.21748739e-01
1.49445340e-01 8.92980620e-02 -1.67694911e-01 1.11425541e-01
-2.40976989e-01 1.64884940e-01 -7.13405132e-01 -3.00159514e-01
8.41694772e-02 2.56792784e-01 -5.91105938e-01 7.72313893e-01
3.58091652e-01 6.88548028e-01 -8.84140790e-01 -2.39842132e-01
7.19939941e-04 1.17886476e-01 -4.40953583e-01 3.41633171e-01
-3.25794220e-01 3.01404864e-01 -2.93341726e-01 1.63429633e-01
6.40243411e-01 -4.95799005e-01 1.13668993e-01 2.95371652e-01
4.91027087e-01 -1.12367511e-01 -1.22816575e+00 1.32020187e+00
-5.73027790e-01 5.06736219e-01 1.02701642e-01 -1.34936726e+00
6.74577415e-01 2.71482058e-02 -3.56169045e-02 -3.12003702e-01
4.38430645e-02 -1.51981311e-02 -1.09519519e-01 -4.59478259e-01
-3.06195945e-01 -4.42761093e-01 -9.36724544e-02 6.04041398e-01
2.64890790e-01 3.93479466e-02 3.42566855e-02 8.21679160e-02
1.34740603e+00 -2.01042607e-01 3.76790792e-01 -3.57376754e-01
5.28831184e-01 -3.25819254e-01 3.42573047e-01 1.01473260e+00
-2.46263474e-01 7.58248866e-01 8.93414319e-01 -4.74339902e-01
-1.10672545e+00 -1.41693199e+00 -3.45953166e-01 9.94206131e-01
1.72724351e-01 -1.01708472e-01 -8.44872534e-01 -1.30442107e+00
1.19160108e-01 6.11206472e-01 -1.14223397e+00 -3.59272808e-01
-2.73029059e-01 -8.26872587e-01 4.12129492e-01 3.99601310e-01
2.46549904e-01 -8.05701375e-01 -5.47631800e-01 -1.27917886e-01
1.29749984e-01 -1.28675795e+00 -5.13139188e-01 4.17575240e-01
-5.37826598e-01 -1.28870022e+00 -6.68091655e-01 -7.58841038e-01
7.68081605e-01 -1.66576859e-02 1.25347006e+00 -1.33623916e-03
-3.72170061e-01 6.41447842e-01 -5.20428419e-02 -3.73919308e-01
-6.57276034e-01 7.97793493e-02 -1.44913480e-01 4.64695871e-01
1.21478915e-01 -5.24314404e-01 -3.32762629e-01 4.64462698e-01
-9.45451140e-01 -9.73802581e-02 4.32180762e-01 9.90407407e-01
4.23957825e-01 1.55960610e-02 5.44315517e-01 -1.24060750e+00
1.16565853e-01 -6.53831840e-01 -8.37950289e-01 3.04704040e-01
-4.03819293e-01 3.20186079e-01 8.65383267e-01 -6.49559081e-01
-6.93060935e-01 3.24514687e-01 3.27485412e-01 -4.12366778e-01
-3.84387434e-01 -2.46206708e-02 -3.01926434e-01 -2.30486125e-01
1.02127695e+00 2.32297257e-01 4.14218344e-02 -3.73842955e-01
4.08328742e-01 4.07376379e-01 7.63035655e-01 -8.28372657e-01
1.05738819e+00 6.84195936e-01 1.81989506e-01 -6.32436693e-01
-1.03922153e+00 -1.31931514e-01 -7.30200827e-01 7.28876190e-03
5.70546150e-01 -7.31249809e-01 -5.05226314e-01 3.83324921e-01
-9.69256222e-01 -5.63550115e-01 -3.63063484e-01 1.09560810e-01
-8.55020285e-01 2.94372469e-01 -3.54459167e-01 -6.47131681e-01
3.04090559e-01 -1.21707106e+00 1.05527163e+00 -7.32585043e-02
1.47565663e-01 -1.13039207e+00 1.93891078e-01 6.58750534e-02
6.83966056e-02 3.98228824e-01 1.14641416e+00 -9.57937658e-01
-5.64496219e-01 -2.75598168e-01 -2.61302292e-01 8.73796880e-01
3.44046168e-02 -1.87108696e-01 -1.26857316e+00 -3.97650331e-01
1.40195876e-01 -7.11567342e-01 1.21404290e+00 6.87151775e-02
1.34782672e+00 -4.21567678e-01 -3.67842197e-01 5.20666063e-01
1.60288477e+00 -3.97057980e-02 2.55698115e-01 3.39577317e-01
5.76517940e-01 7.26905167e-01 4.01037067e-01 6.08192720e-02
1.52475378e-02 8.35247219e-01 5.39555967e-01 1.85917746e-02
8.21101516e-02 -1.65704310e-01 3.81252170e-01 -1.49676308e-01
2.64001369e-01 -2.17474207e-01 -8.54234278e-01 5.94403803e-01
-1.74422753e+00 -7.88801134e-01 3.33629400e-01 2.59455490e+00
9.94035602e-01 3.36170852e-01 4.19112414e-01 7.55216628e-02
7.80130506e-01 1.24779440e-01 -6.69208348e-01 -9.63454843e-02
-1.10321883e-02 2.63599545e-01 6.05852723e-01 5.61529458e-01
-1.60447705e+00 5.09905934e-01 7.17026186e+00 6.71216249e-01
-1.24345481e+00 5.80372252e-02 8.90062213e-01 -9.49885696e-02
-1.35147735e-01 -1.52727842e-01 -5.74749947e-01 5.91861784e-01
9.03123438e-01 -1.60210058e-01 4.55500036e-01 9.33299065e-01
-3.06024194e-01 3.17744575e-02 -1.62720644e+00 6.67660296e-01
4.28285562e-02 -1.08363426e+00 -6.07249737e-02 9.13828835e-02
6.51644111e-01 -1.58963315e-02 3.90833765e-01 9.68105197e-02
6.15168154e-01 -1.30622828e+00 7.08195925e-01 1.42382279e-01
7.78565645e-01 -6.07703745e-01 4.67874527e-01 5.65601051e-01
-6.49478376e-01 -1.66679084e-01 -2.26185411e-01 6.20763414e-02
-3.52182865e-01 5.80397010e-01 -9.04781878e-01 1.37092993e-01
5.90862870e-01 7.56291270e-01 -6.57935679e-01 9.43931818e-01
-2.35864863e-01 7.38534093e-01 -3.52676481e-01 3.29500049e-01
1.55490309e-01 -1.22111119e-01 4.76053268e-01 1.49118841e+00
-3.71064432e-02 -3.16302001e-01 3.85358393e-01 7.96635926e-01
-2.60858268e-01 -4.63718139e-02 -1.07142878e+00 3.68411899e-01
2.51778960e-01 1.22151530e+00 -7.17660010e-01 -2.80773014e-01
-5.07675171e-01 8.84491026e-01 7.31904984e-01 4.65857416e-01
-7.98905194e-01 -3.32371771e-01 5.65504968e-01 1.75715595e-01
5.93468130e-01 7.74520636e-02 -2.60466456e-01 -1.15194547e+00
3.38455468e-01 -8.88583362e-01 4.94258285e-01 -4.19535995e-01
-1.59333932e+00 3.91751617e-01 1.66452810e-01 -1.04382968e+00
-2.49907479e-01 -9.82150912e-01 -5.86549640e-01 9.03050244e-01
-1.42787445e+00 -8.47594678e-01 -1.76771611e-01 4.87893820e-01
2.04724476e-01 -1.51896983e-01 7.40198612e-01 3.73785500e-03
-5.12789726e-01 6.47927880e-01 1.94320738e-01 2.96109855e-01
6.89380050e-01 -1.68070197e+00 3.85371804e-01 8.82010520e-01
3.90381545e-01 1.60326406e-01 6.94095492e-01 -3.69253270e-02
-7.60047436e-01 -1.27318823e+00 3.49504948e-01 -8.67735803e-01
7.00323164e-01 -7.92423487e-01 -9.86260176e-01 1.08338010e+00
-1.54393151e-01 5.74754238e-01 4.42491472e-01 -4.09455262e-02
-8.61766934e-01 -1.32772788e-01 -1.25433373e+00 4.27171737e-01
1.03514147e+00 -6.12647474e-01 -5.59534192e-01 4.99645770e-01
4.65933949e-01 -2.66437769e-01 -5.10194778e-01 2.42729917e-01
3.18967491e-01 -9.16845739e-01 9.93904233e-01 -9.23207581e-01
5.35894096e-01 -2.64444619e-01 -3.33813876e-01 -1.46898258e+00
-1.02167979e-01 -4.08803046e-01 1.21708043e-01 9.91181612e-01
5.07174313e-01 -7.21441150e-01 7.09733605e-01 3.32284182e-01
2.88795263e-01 -6.02128685e-01 -9.75228012e-01 -1.00142133e+00
3.08387578e-01 -2.85507649e-01 2.24407345e-01 8.14827800e-01
-2.22023934e-01 4.06898230e-01 -3.28695476e-01 3.02133411e-01
9.17619407e-01 4.75933440e-02 7.66960025e-01 -1.24013674e+00
-6.35569453e-01 -3.34577620e-01 -7.29884028e-01 -8.16858590e-01
5.23911536e-01 -1.01579213e+00 2.31099471e-01 -7.79468775e-01
5.64857543e-01 -5.57563186e-01 -2.73205638e-01 4.81150478e-01
-1.81319743e-01 3.84328932e-01 9.20532048e-02 1.44201398e-01
-5.05368590e-01 3.50196183e-01 9.23948884e-01 -2.88039267e-01
2.43060291e-01 3.30370665e-01 -7.34798908e-01 8.87296021e-01
4.58760053e-01 -6.98079109e-01 -6.62608325e-01 -2.17309505e-01
8.15965608e-02 -3.50970298e-01 6.76286042e-01 -9.88579452e-01
-7.50003308e-02 -3.54357481e-01 2.84084529e-01 1.57920316e-01
1.56858891e-01 -9.13556218e-01 -1.47888869e-01 3.32204968e-01
-8.43921661e-01 -3.32312524e-01 -1.27336266e-03 9.14300442e-01
6.45224154e-02 -4.49133158e-01 1.34919298e+00 -7.71336928e-02
-3.97470176e-01 2.01139286e-01 -5.81595972e-02 5.32117844e-01
1.33272183e+00 1.00428864e-01 -3.34083110e-01 -4.60613877e-01
-9.00646150e-01 1.79465950e-01 7.26451457e-01 1.34748429e-01
1.85185984e-01 -1.26078808e+00 -7.48879731e-01 1.84857339e-01
3.74119550e-01 4.75715473e-02 -2.30534554e-01 5.90078771e-01
-3.66314650e-01 1.46272659e-01 -1.56080797e-01 -8.37475777e-01
-1.13968945e+00 1.07253861e+00 7.21242428e-01 -2.16149285e-01
-6.27015173e-01 6.65713489e-01 7.79197991e-01 -4.15580094e-01
4.88905787e-01 -2.03242034e-01 5.30365817e-02 -8.82958919e-02
6.98194802e-01 1.20479457e-01 7.87170827e-02 -4.74338025e-01
-3.18088025e-01 4.52202886e-01 -2.30615556e-01 -3.14224899e-01
9.84698176e-01 1.03196181e-01 3.62110198e-01 6.76672280e-01
1.52252436e+00 -6.55341074e-02 -1.68932033e+00 -5.24241567e-01
2.18895346e-01 -5.46007574e-01 -2.33153611e-01 -8.65250766e-01
-9.92642522e-01 9.56372440e-01 5.91348350e-01 4.02837127e-01
1.04111385e+00 3.13987225e-01 1.86253130e-01 4.93986964e-01
2.96831608e-01 -8.09890687e-01 3.08386832e-01 3.46296906e-01
7.82779336e-01 -1.55513823e+00 -2.58320898e-01 -2.76872367e-01
-6.87106013e-01 9.90645826e-01 3.30106020e-01 -4.21241105e-01
7.64907777e-01 3.00452590e-01 1.04786515e-01 1.01448186e-01
-9.36501086e-01 -8.40608403e-02 2.34062597e-01 8.27116549e-01
-8.83652419e-02 3.08357328e-02 2.16800213e-01 3.97312403e-01
1.01893194e-01 -1.30754650e-01 5.57859778e-01 7.67060339e-01
-4.44623441e-01 -1.03654146e+00 -2.67257899e-01 3.58908534e-01
-5.51077545e-01 1.77260950e-01 -5.00344694e-01 8.55523765e-01
1.73135288e-02 5.10318041e-01 1.23222955e-01 -2.88576454e-01
3.14054459e-01 2.95632154e-01 5.57480633e-01 -7.55959094e-01
-1.60553411e-01 -2.53746778e-01 -7.06848428e-02 -4.95081425e-01
-2.49351412e-01 -7.65825152e-01 -7.61731684e-01 2.03494318e-02
4.09995988e-02 -2.39542034e-03 4.68933463e-01 9.79935288e-01
1.33857327e-02 7.53526092e-02 1.00957632e+00 -7.94279218e-01
-1.12216461e+00 -7.06627786e-01 -7.40011096e-01 6.99196279e-01
7.95848966e-01 -7.35615849e-01 -7.59869099e-01 2.13854775e-01] | [9.770793914794922, 2.7599313259124756] |
b95f519c-2a23-44a5-a795-743dbb593259 | hate-alert-dravidianlangtech-acl2022 | 2204.12587 | null | https://arxiv.org/abs/2204.12587v1 | https://arxiv.org/pdf/2204.12587v1.pdf | hate-alert@DravidianLangTech-ACL2022: Ensembling Multi-Modalities for Tamil TrollMeme Classification | Social media platforms often act as breeding grounds for various forms of trolling or malicious content targeting users or communities. One way of trolling users is by creating memes, which in most cases unites an image with a short piece of text embedded on top of it. The situation is more complex for multilingual(e.g., Tamil) memes due to the lack of benchmark datasets and models. We explore several models to detect Troll memes in Tamil based on the shared task, "Troll Meme Classification in DravidianLangTech2022" at ACL-2022. We observe while the text-based model MURIL performs better for Non-troll meme classification, the image-based model VGG16 performs better for Troll-meme classification. Further fusing these two modalities help us achieve stable outcomes in both classes. Our fusion model achieved a 0.561 weighted average F1 score and ranked second in this task. | ['Animesh Mukherjee', 'Somnath Banerjee', 'Mithun Das'] | 2022-03-25 | null | https://aclanthology.org/2022.dravidianlangtech-1.8 | https://aclanthology.org/2022.dravidianlangtech-1.8.pdf | dravidianlangtech-acl-2022-5 | ['meme-classification'] | ['natural-language-processing'] | [-3.85233536e-02 -3.00625771e-01 -5.00245579e-02 4.40119714e-01
-8.44415963e-01 -8.41290295e-01 1.38250732e+00 2.94830084e-01
-7.74388373e-01 5.56643844e-01 2.19899118e-01 -1.97283000e-01
5.36772609e-01 -7.27475941e-01 -6.17025793e-01 -6.68450713e-01
8.00044760e-02 2.43366539e-01 5.04357517e-01 -4.58339930e-01
5.85988522e-01 -1.25164121e-01 -9.81710374e-01 7.85310030e-01
5.79623759e-01 5.63106477e-01 5.03553748e-02 7.88203418e-01
-2.31354326e-01 1.20922136e+00 -8.30522418e-01 -6.55011296e-01
2.74633586e-01 -2.88657188e-01 -6.07894123e-01 -1.18978292e-01
1.25987804e+00 -8.84894729e-02 -7.12671816e-01 1.08430433e+00
3.68081450e-01 -2.40705475e-01 9.98275757e-01 -1.07634330e+00
-2.75371850e-01 6.69567823e-01 -9.40773010e-01 4.82822418e-01
4.49169427e-01 4.55116332e-02 8.93439412e-01 -8.42537761e-01
1.06849372e+00 1.52709115e+00 8.67054582e-01 3.59968126e-01
-1.04391634e+00 -8.65156114e-01 -4.21639711e-01 -1.01384871e-01
-1.41105080e+00 -2.40412340e-01 2.85008639e-01 -7.97908902e-01
5.10617673e-01 3.72296929e-01 2.07905471e-01 1.66180599e+00
9.19192672e-01 8.54282975e-01 1.32300079e+00 -3.99254896e-02
-2.98819661e-01 4.32687074e-01 -1.20691722e-02 9.02571440e-01
3.22397292e-01 -3.17950726e-01 -8.00022721e-01 -5.05069137e-01
2.49481395e-01 1.52410078e-03 1.21391714e-01 2.41818115e-01
-1.48963654e+00 1.13076401e+00 4.96620327e-01 5.95461786e-01
9.22283158e-02 5.23736954e-01 7.67861545e-01 5.01813769e-01
1.15918493e+00 5.16014457e-01 2.82920152e-01 3.28211002e-02
-1.23817229e+00 2.25008339e-01 5.47194541e-01 3.28367591e-01
6.89151168e-01 -3.47384959e-01 -1.62904173e-01 9.07339394e-01
4.25865501e-01 9.23304021e-01 1.68315604e-01 -2.20480248e-01
8.45347762e-01 5.73029935e-01 -2.98783425e-02 -1.44739628e+00
-5.44890344e-01 -4.56093431e-01 -6.37755513e-01 1.08940937e-01
4.13094401e-01 -2.09046975e-01 -1.07048965e+00 1.20351565e+00
2.48181149e-01 1.63013339e-01 -4.60672915e-01 6.84305072e-01
7.96255112e-01 9.66352582e-01 1.11582816e-01 3.84357870e-01
1.44051051e+00 -1.04930353e+00 -3.52546006e-01 -2.65331239e-01
1.01040626e+00 -1.04902518e+00 6.39697969e-01 2.47561082e-01
-2.80224174e-01 -1.74997345e-01 -8.94807994e-01 2.56965607e-01
-8.79150569e-01 -1.55393824e-01 2.98407912e-01 9.26829278e-01
-8.82851183e-01 9.02529955e-02 -1.94954425e-01 -8.81256700e-01
4.01290357e-01 -2.20432624e-01 -5.63304722e-01 5.82216866e-02
-1.18252325e+00 9.02787387e-01 1.60682827e-01 -3.59182298e-01
-1.14587760e+00 -4.09791142e-01 -5.32892585e-01 -7.97771037e-01
7.80050308e-02 -2.59482473e-01 7.14418471e-01 -9.00050104e-01
-7.00091541e-01 1.42756128e+00 3.97606611e-01 -6.36360228e-01
1.10436845e+00 -2.39622116e-01 -5.73205948e-01 2.61934310e-01
6.07877314e-01 8.56270611e-01 1.31876159e+00 -1.33991683e+00
-7.53237784e-01 -1.52927727e-01 4.42148112e-02 -8.66372809e-02
-6.83627844e-01 3.10947090e-01 -2.46590599e-01 -7.94712484e-01
-1.15788542e-01 -1.25597370e+00 6.28912821e-02 -5.16749322e-01
-6.39970839e-01 -1.02791287e-01 1.28052199e+00 -8.27524304e-01
1.41474950e+00 -1.96015024e+00 -3.84062342e-02 2.67608166e-01
8.35001588e-01 1.59957469e-01 -2.09116831e-01 9.28116739e-01
3.97551656e-01 5.51083028e-01 1.14891313e-01 -4.69940841e-01
-8.46564025e-02 -3.74610305e-01 -3.95073026e-01 9.64176595e-01
-3.02760422e-01 8.49048734e-01 -9.62204635e-01 -7.20319986e-01
1.36804819e-01 3.94319147e-01 -3.46028775e-01 -3.53471428e-01
-1.01338677e-01 6.23748183e-01 -5.53220451e-01 6.00084662e-01
5.19955873e-01 -1.92737386e-01 -9.46600176e-03 6.31007925e-02
-1.69772774e-01 1.12581007e-01 -5.31080723e-01 1.20849061e+00
-4.79864597e-01 1.32987630e+00 1.76809311e-01 4.73344587e-02
5.94232619e-01 1.52968958e-01 4.13089722e-01 -7.21338868e-01
3.67271334e-01 2.47479498e-01 -2.64457405e-01 -4.43983316e-01
8.77635360e-01 1.40341446e-01 -6.45372510e-01 6.15771890e-01
-3.17961991e-01 -2.04174407e-02 3.00880402e-01 7.61490047e-01
1.58796465e+00 -2.26196200e-01 3.28087285e-02 -3.74217570e-01
3.68365526e-01 2.55329669e-01 -9.97768268e-02 1.16764140e+00
-4.19273406e-01 4.10956651e-01 3.48335713e-01 -6.39804780e-01
-1.14716697e+00 -8.04725587e-01 -9.84553248e-02 1.48607707e+00
1.98367089e-01 -6.78513050e-01 -7.24049032e-01 -1.03355134e+00
-3.24499123e-02 4.22033697e-01 -7.87029922e-01 1.04934879e-01
-4.79126573e-01 -8.60414863e-01 1.24245322e+00 -3.94352406e-01
8.93020332e-01 -8.63736749e-01 -3.54268700e-01 1.33123621e-01
-5.38615465e-01 -1.10395432e+00 -5.14060020e-01 -3.17825407e-01
-4.03421193e-01 -1.00493753e+00 -9.25286949e-01 -4.41007137e-01
4.29339111e-01 3.91447604e-01 1.09643257e+00 3.58331561e-01
-1.62744567e-01 4.12266344e-01 -5.85239172e-01 -2.11937755e-01
-7.48843729e-01 4.54891235e-01 5.91736436e-02 4.15896505e-01
1.03063196e-01 5.52569814e-02 -3.13782603e-01 5.82483828e-01
-9.70700085e-01 1.19506866e-01 4.04859871e-01 3.39887798e-01
-2.19560608e-01 -6.31478761e-05 7.20247701e-02 -1.12476921e+00
4.92829680e-01 -1.06198299e+00 -7.11877570e-02 -7.84835368e-02
-3.82829900e-03 -4.82163697e-01 1.63000911e-01 -5.01861393e-01
-5.30473173e-01 -2.22955465e-01 1.96612805e-01 -1.19251035e-01
1.47871282e-02 2.98131913e-01 3.83136272e-01 -2.48687536e-01
1.01960611e+00 4.99952994e-02 -3.63446504e-01 -3.90551507e-01
8.56016669e-03 1.00177801e+00 -8.68546590e-02 -1.29147798e-01
1.22057581e+00 6.96938396e-01 -2.03769162e-01 -1.25378299e+00
-7.69063056e-01 -7.66599774e-01 -1.93428650e-01 -7.70313978e-01
1.16628087e+00 -9.30393577e-01 -3.99536818e-01 9.23051834e-01
-1.02589548e+00 -3.35766703e-01 6.02507174e-01 9.77863371e-02
-1.02937885e-01 4.56527382e-01 -8.36989880e-01 -7.89623201e-01
-1.88298568e-01 -9.30117846e-01 1.26935875e+00 -3.68559361e-01
-3.80873173e-01 -1.23361087e+00 4.45866615e-01 8.23826551e-01
4.30072248e-01 5.67413151e-01 4.55004483e-01 -6.85686529e-01
-3.25335294e-01 -5.98245084e-01 -4.18632746e-01 -6.49686977e-02
-2.39939228e-01 -2.37619132e-01 -8.96285415e-01 -4.30909187e-01
-4.62914497e-01 -4.47236478e-01 1.31841934e+00 -2.71819979e-01
3.89621854e-01 -2.76680768e-01 -6.52142525e-01 -5.84273301e-02
1.29399157e+00 -3.44460368e-01 7.95760930e-01 5.96476018e-01
1.10478497e+00 4.85920966e-01 4.90038037e-01 3.15025151e-01
3.25461209e-01 7.81797588e-01 6.47352934e-01 -2.68544815e-02
-2.21167535e-01 -3.73134434e-01 9.75144804e-01 4.89741236e-01
1.13953631e-02 -7.09109366e-01 -1.43644309e+00 3.96581441e-01
-1.83501935e+00 -1.11741197e+00 -8.57567906e-01 1.70138001e+00
2.55416900e-01 1.04998916e-01 4.51237589e-01 -6.02544285e-02
9.64896917e-01 5.62326431e-01 1.74956024e-01 -1.44965589e-01
-2.04897717e-01 -4.19978112e-01 1.01517844e+00 4.43158031e-01
-1.58762348e+00 1.19336283e+00 6.04305077e+00 1.50208127e+00
-1.12768936e+00 8.16607952e-01 5.08223653e-01 1.05353281e-01
2.49944255e-02 -2.31568381e-01 -9.62136149e-01 7.99295425e-01
1.10787630e+00 5.55852592e-01 2.92667270e-01 2.55960464e-01
1.66665450e-01 -5.36871195e-01 -5.07887423e-01 9.41562235e-01
4.88100350e-01 -1.47704244e+00 -3.95116657e-02 6.08205497e-01
1.00323236e+00 6.45445406e-01 2.92042017e-01 3.18686843e-01
1.23638541e-01 -1.05973947e+00 1.10692322e+00 3.52924824e-01
5.74203193e-01 -4.11175072e-01 6.21164322e-01 4.54896539e-01
-1.10828745e+00 -5.39949611e-02 -1.32215798e-01 1.71861365e-01
-3.81425731e-02 4.45015252e-01 -8.23062479e-01 1.92745432e-01
6.70001507e-01 9.53868687e-01 -1.14616132e+00 9.31374729e-01
-1.05730347e-01 9.06872451e-01 -3.90438765e-01 -2.27826029e-01
7.10928977e-01 1.62066996e-01 1.09628797e+00 1.66618848e+00
1.72698528e-01 -7.75840700e-01 4.26676452e-01 3.73199016e-01
-7.61782378e-02 2.57409722e-01 -9.19426322e-01 -4.87403482e-01
1.58168390e-01 1.59096742e+00 -1.01421940e+00 -3.33614320e-01
-1.81076482e-01 1.24799144e+00 6.56002481e-03 9.79351904e-03
-1.22338378e+00 -9.64446664e-02 1.36737689e-01 7.13894367e-01
-1.54317990e-01 -4.99138117e-01 -3.24607491e-02 -1.17657709e+00
-5.27753830e-01 -9.98713791e-01 4.21534091e-01 -5.74154735e-01
-1.41663206e+00 6.32365525e-01 -2.09622338e-01 -1.28311551e+00
1.52366757e-01 -7.05091894e-01 -5.04933000e-01 3.75604808e-01
-8.90928209e-01 -1.82149625e+00 -1.33122489e-01 5.21236241e-01
5.42533934e-01 -4.38874990e-01 3.31385702e-01 5.98200262e-01
-4.68532979e-01 4.46379244e-01 3.56980301e-02 3.59676778e-01
9.19959724e-01 -8.76532555e-01 2.44330719e-01 6.04893327e-01
3.74410808e-01 3.79401118e-01 1.01199377e+00 -1.16570354e+00
-1.27748966e+00 -1.32735848e+00 8.71543825e-01 -1.24972522e+00
1.34570158e+00 -7.34080076e-01 -2.69715637e-01 6.70494080e-01
4.33282614e-01 -5.30243158e-01 3.89550358e-01 -4.88854153e-03
-7.84171402e-01 4.40599203e-01 -1.17763340e+00 7.06925631e-01
8.50289524e-01 -7.29764402e-01 -2.71554142e-01 8.32593739e-01
7.55057409e-02 -4.95521650e-02 -7.23783076e-01 -2.22132534e-01
5.24497032e-01 -9.59690988e-01 6.06520474e-01 -2.14471608e-01
7.28533328e-01 -2.93261558e-01 -2.80593842e-01 -1.18428051e+00
-6.73447549e-02 -5.11336625e-01 2.54706353e-01 1.17928898e+00
5.85030496e-01 -7.63854682e-01 7.52237320e-01 -3.06973487e-01
1.42055854e-01 8.28328058e-02 -1.08516204e+00 -7.64817178e-01
2.97227412e-01 -4.98810619e-01 -2.28358239e-01 1.16961050e+00
-1.57915473e-01 6.01049721e-01 -9.56273139e-01 -5.31711243e-02
8.15093040e-01 -6.00420535e-01 1.19348812e+00 -9.84333158e-01
-2.30015032e-02 -3.76770020e-01 -6.90522075e-01 -5.20796478e-01
1.70204103e-01 -1.11021841e+00 -1.84793562e-01 -1.26847780e+00
6.92453802e-01 -4.90026414e-01 5.32106608e-02 4.17723566e-01
2.64282972e-01 1.34739733e+00 4.18795824e-01 5.91619492e-01
-1.14124215e+00 -2.41820112e-01 9.90206718e-01 -5.81429839e-01
1.10890456e-01 -3.83750081e-01 -1.74290538e-01 7.14685142e-01
7.00150430e-01 -8.37754488e-01 2.20612556e-01 -2.15675488e-01
9.27473485e-01 -3.23558360e-01 6.46168351e-01 -1.20088267e+00
1.03475247e-02 -4.12177742e-02 1.88178331e-01 -7.04044223e-01
3.35367799e-01 -6.22517347e-01 2.26833835e-01 7.70038128e-01
-6.93924502e-02 -2.13233605e-02 5.05861565e-02 6.51424825e-01
4.42804396e-02 1.00022471e-02 9.11373615e-01 -1.49333328e-01
-6.65515423e-01 8.24183375e-02 -1.07402384e+00 7.16674551e-02
9.37066019e-01 -2.60373771e-01 -7.45649517e-01 -5.69291651e-01
-5.14825046e-01 -5.17065488e-02 6.47399306e-01 7.38429785e-01
1.76790029e-01 -1.08745515e+00 -1.15457165e+00 -3.31380159e-01
4.36869591e-01 -1.19744515e+00 1.57734409e-01 1.18346572e+00
-6.84592247e-01 3.12960982e-01 -1.33893594e-01 -6.90186441e-01
-1.46343601e+00 9.52081233e-02 3.71044308e-01 -5.67269981e-01
-5.40315807e-01 7.19817519e-01 3.24526608e-01 -4.90439087e-01
-2.74958700e-01 6.64815903e-01 -1.53319016e-01 5.56624353e-01
7.66897798e-01 7.17781603e-01 -1.22282520e-01 -1.42727959e+00
-4.77434844e-01 5.82944691e-01 -2.00943321e-01 -2.30616048e-01
9.37582314e-01 -2.10132033e-01 -3.58879834e-01 6.52551115e-01
1.22047269e+00 6.31871402e-01 -5.57894647e-01 1.25056043e-01
-2.91646402e-02 -4.83650088e-01 1.87888116e-01 -7.40169585e-01
-8.37113500e-01 7.34047174e-01 6.88093662e-01 5.84388077e-01
1.82450593e-01 1.09747887e-01 9.73270476e-01 2.43417948e-01
8.09434772e-01 -1.13147438e+00 5.00486672e-01 7.33898103e-01
8.54673982e-01 -1.23944378e+00 3.03746685e-02 -2.90590733e-01
-4.53935862e-01 9.03245389e-01 3.27520609e-01 -2.94571340e-01
4.10179824e-01 -7.26673827e-02 1.91407040e-01 -5.98072290e-01
-4.45699930e-01 -1.03673980e-01 4.78911608e-01 1.99524507e-01
3.38968575e-01 2.04298005e-01 -4.19331610e-01 -4.04251575e-01
-3.98476385e-02 -7.30326831e-01 6.85362697e-01 8.41531336e-01
-6.92686200e-01 -8.30454588e-01 -8.67576063e-01 6.46716952e-01
-8.51574361e-01 -1.50985241e-01 -1.03600585e+00 9.03693616e-01
4.03583646e-01 1.23896623e+00 7.53127784e-02 -9.84142542e-01
-3.52611244e-01 -2.54083365e-01 2.84758478e-01 -7.17331529e-01
-1.22989118e+00 1.79839414e-02 5.13613641e-01 -3.33329916e-01
-2.72343159e-01 -5.45103967e-01 -8.30490589e-01 -9.60847437e-01
-2.34669521e-01 -1.82081640e-01 7.95389473e-01 8.16438854e-01
4.59232479e-02 3.43243927e-02 5.39258599e-01 -1.02867055e+00
2.47041453e-02 -1.19621754e+00 -6.13540649e-01 5.63306689e-01
1.73912942e-01 -5.42141020e-01 -5.16714036e-01 -2.46151909e-01] | [8.474621772766113, 10.698049545288086] |
4c13782c-45ae-4582-83be-3418ab03a8bc | de-risking-carbon-capture-and-sequestration | 2212.08596 | null | https://arxiv.org/abs/2212.08596v1 | https://arxiv.org/pdf/2212.08596v1.pdf | De-risking Carbon Capture and Sequestration with Explainable CO2 Leakage Detection in Time-lapse Seismic Monitoring Images | With the growing global deployment of carbon capture and sequestration technology to combat climate change, monitoring and detection of potential CO2 leakage through existing or storage induced faults are critical to the safe and long-term viability of the technology. Recent work on time-lapse seismic monitoring of CO2 storage has shown promising results in its ability to monitor the growth of the CO2 plume from surface recorded seismic data. However, due to the low sensitivity of seismic imaging to CO2 concentration, additional developments are required to efficiently interpret the seismic images for leakage. In this work, we introduce a binary classification of time-lapse seismic images to delineate CO2 plumes (leakage) using state-of-the-art deep learning models. Additionally, we localize the leakage region of CO2 plumes by leveraging Class Activation Mapping methods. | ['Felix J. Herrmann', 'Mathias Louboutin', 'Ziyi Yin', 'Abhinav Prakash Gahlot', 'Huseyin Tuna Erdinc'] | 2022-12-16 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [ 2.56422192e-01 -3.04906130e-01 4.07973155e-02 3.48723024e-01
-1.04113233e+00 -7.15535700e-01 9.52776968e-01 2.23986015e-01
-2.43486360e-01 6.06360495e-01 2.82552183e-01 -9.85239625e-01
3.44259478e-03 -8.32489669e-01 -7.21860588e-01 -9.97870088e-01
-5.16860843e-01 1.49229094e-01 2.98217803e-01 6.07644580e-02
4.76193070e-01 1.03418505e+00 -9.31432068e-01 1.37389794e-01
3.68948877e-01 1.17438281e+00 3.21782529e-01 5.75383723e-01
3.08921188e-02 5.65292716e-01 -6.40456140e-01 9.61214185e-01
2.04310596e-01 -2.63072699e-01 -8.18863034e-01 -4.26785111e-01
3.78184058e-02 -3.45851690e-01 -2.08091736e-01 9.67106998e-01
4.04297143e-01 -3.86706889e-02 6.64527357e-01 -5.80287635e-01
-9.86557230e-02 2.76509494e-01 -6.06660604e-01 7.28487134e-01
-2.45504633e-01 3.98681283e-01 4.68892068e-01 -1.09821439e+00
2.03783080e-01 5.57654858e-01 3.93691868e-01 1.18133105e-01
-1.02532291e+00 -6.30020678e-01 -1.01802342e-01 -1.11169584e-01
-1.03983808e+00 -7.01180875e-01 8.23624551e-01 -8.57495427e-01
1.53842950e+00 2.00795606e-01 7.44903624e-01 5.90872407e-01
-1.79243144e-02 2.61660546e-01 1.51276791e+00 -2.80869275e-01
3.94084841e-01 -5.30584335e-01 -4.32100415e-01 3.83321851e-01
2.02474624e-01 4.29822475e-01 -4.59391505e-01 6.07148521e-02
7.12338388e-01 -3.16223890e-01 -2.87004411e-01 2.32552633e-01
-8.13711166e-01 5.43153167e-01 7.46354699e-01 5.43032467e-01
-3.98016512e-01 7.47676313e-01 1.95785195e-01 -2.45335683e-01
8.42507601e-01 6.06702268e-01 -3.42088640e-01 -2.04417795e-01
-1.39383268e+00 1.94221452e-01 1.11135289e-01 7.25808516e-02
5.62617540e-01 5.14378071e-01 3.94852877e-01 1.00028843e-01
6.35656416e-01 1.02978992e+00 1.58326522e-01 -8.64864647e-01
4.62043256e-01 2.81766027e-01 3.24285269e-01 -7.93568254e-01
-8.72855559e-02 -2.56902397e-01 -4.03620511e-01 5.59607029e-01
1.30610794e-01 -5.77286109e-02 -9.39693213e-01 1.08606267e+00
-1.39197677e-01 1.53522342e-01 -5.43973260e-02 6.11434460e-01
-8.85628047e-04 8.61501515e-01 1.30944207e-01 -4.64978740e-02
8.19210947e-01 -5.19852221e-01 -3.93527716e-01 -6.53095722e-01
2.77202368e-01 2.11896896e-01 5.86930871e-01 -3.64679174e-04
-7.08570242e-01 1.61506444e-01 -1.19095433e+00 4.93086576e-01
-8.39741707e-01 -3.35265219e-01 6.92728519e-01 7.96942294e-01
-6.50358915e-01 6.05400264e-01 -1.54450572e+00 4.61843796e-02
8.09653580e-01 7.90654123e-02 -5.41445278e-02 3.84875804e-01
-1.03717148e+00 9.44763899e-01 2.04460829e-01 6.95655704e-01
-1.67304456e+00 -8.66168857e-01 -4.74354833e-01 1.50502175e-01
5.29665239e-02 2.77307063e-01 9.46470499e-01 -5.54413497e-02
-9.98697102e-01 5.13607085e-01 4.01923433e-02 -3.45876545e-01
5.38924992e-01 -1.93027213e-01 -2.84401685e-01 4.91409630e-01
9.80408192e-02 3.88283551e-01 2.64137417e-01 -1.25955892e+00
-4.74399716e-01 -5.05340874e-01 -5.48355579e-01 -6.97310865e-02
-2.95514613e-01 2.97224700e-01 3.45648825e-01 -2.29974657e-01
3.55147235e-02 -6.78144872e-01 -1.37016192e-01 -9.02297795e-02
-1.86982006e-01 2.83506781e-01 1.23579407e+00 -9.66540992e-01
7.84907877e-01 -1.96245050e+00 -1.88590705e-01 1.94003180e-01
-6.94079474e-02 2.33382627e-01 4.71714616e-01 2.88918883e-01
-3.15030068e-02 6.95675790e-01 -9.76021647e-01 1.32739335e-01
-2.98680693e-01 1.42790191e-02 -5.85600436e-01 7.31114924e-01
4.68071133e-01 8.40256631e-01 -1.02999854e+00 1.19324945e-01
1.60644904e-01 3.15145224e-01 2.58283973e-01 1.51117519e-01
-3.14487725e-01 6.61131680e-01 -2.28141829e-01 1.10241723e+00
6.48472309e-01 3.79049480e-02 -2.00392256e-04 2.04777732e-01
-8.49295914e-01 4.31815743e-01 -4.77817118e-01 1.35295510e+00
-1.79848909e-01 8.01440120e-01 4.01892781e-01 -1.17613316e+00
9.32388783e-01 2.52316028e-01 6.94356263e-01 -1.04948723e+00
-4.54147071e-01 8.00283313e-01 -4.02726412e-01 -7.59407520e-01
2.39084288e-01 -6.63328707e-01 2.02243030e-01 3.36414188e-01
-4.01823759e-01 -3.12113225e-01 -5.13192594e-01 -6.17661923e-02
1.02380049e+00 5.13252914e-02 -5.73559403e-01 -5.73967814e-01
2.03918479e-02 1.56007379e-01 -5.80793172e-02 5.05041122e-01
-8.69597718e-02 4.54601616e-01 2.83680022e-01 -2.00868353e-01
-9.30758715e-01 -1.05072606e+00 -1.24651439e-01 4.53823775e-01
-1.29707783e-01 4.37011153e-01 -2.38285452e-01 -4.79841202e-01
2.58670058e-02 6.29159689e-01 -5.12813449e-01 -3.86852771e-02
-6.90835059e-01 -1.25869024e+00 1.07850266e+00 1.10696340e+00
7.09016919e-01 -9.62631226e-01 -1.34190309e+00 3.54305208e-01
-1.13152238e-02 -6.68440700e-01 3.36633891e-01 8.70066643e-01
-1.04596007e+00 -1.22227991e+00 -7.38803267e-01 -2.99966961e-01
5.28580606e-01 2.59975418e-02 6.26215577e-01 -1.17130592e-01
-1.90828413e-01 2.36128286e-01 1.18652426e-01 -4.43143368e-01
-2.49240741e-01 -2.08797455e-01 -2.27129713e-01 -6.04353607e-01
1.07401133e-01 -3.29704434e-01 -7.02391863e-01 7.73851760e-04
-6.93688571e-01 -2.27990985e-01 2.60351509e-01 4.41209644e-01
5.89545548e-01 4.95338261e-01 5.86999476e-01 -3.19750875e-01
6.14232242e-01 -9.15789962e-01 -7.64249384e-01 1.01428755e-01
-9.18123424e-01 -2.32578158e-01 -5.26422262e-02 3.96575928e-02
-1.10348010e+00 -2.25891531e-01 7.24273324e-02 -1.45586893e-01
-3.10081631e-01 1.20253229e+00 1.06913507e-01 -7.21939504e-02
4.27736133e-01 2.09286213e-01 -1.46901369e-01 -5.06122530e-01
-1.44346161e-02 6.27307475e-01 1.09186411e+00 -8.16442311e-01
6.85640037e-01 6.92864358e-01 1.87546000e-01 -1.18354857e+00
-3.35264236e-01 -4.22664136e-01 -4.73638475e-01 -6.28230333e-01
8.77390325e-01 -8.86441350e-01 -1.75113425e-01 7.78611124e-01
-9.34919953e-01 -6.78678393e-01 -6.76803710e-03 2.37379029e-01
-4.41499464e-02 -1.74487926e-04 -3.82637948e-01 -1.39735591e+00
-4.47986782e-01 -1.03416860e+00 8.49671304e-01 1.26460731e-01
3.20123434e-01 -1.05142188e+00 2.75339007e-01 3.29636037e-01
9.73268628e-01 9.00516212e-01 8.61632109e-01 -2.37608761e-01
-7.22310781e-01 -2.68422365e-01 -1.91999346e-01 3.15663069e-01
3.37936103e-01 -3.47496883e-04 -1.33519006e+00 -1.60560563e-01
2.52752960e-01 -2.21300289e-01 1.28961563e+00 3.06233644e-01
1.01506984e+00 -1.96086675e-01 -4.40972835e-01 5.79960346e-01
1.59436846e+00 6.52912378e-01 6.35432839e-01 8.17076921e-01
4.59093392e-01 6.60962343e-01 1.06033534e-01 1.47964507e-01
-2.89860636e-01 -4.59681042e-02 8.91735971e-01 -1.27017871e-01
1.83576450e-03 -2.57421255e-01 3.65278870e-01 4.06643689e-01
-2.43890136e-01 -1.97012387e-02 -1.58180070e+00 1.04441679e+00
-1.39627993e+00 -1.05747163e+00 -4.23937589e-01 2.16865492e+00
4.62701648e-01 2.56148785e-01 -1.11830860e-01 1.26989394e-01
4.30443496e-01 4.30270284e-01 -6.06833518e-01 3.15539762e-02
-4.25272644e-01 1.64242893e-01 8.95381868e-01 4.24825191e-01
-1.05928433e+00 5.54817736e-01 7.14926863e+00 7.52560347e-02
-1.57232726e+00 2.53899693e-01 8.58965516e-01 6.92885369e-02
-7.52734005e-01 1.01609342e-01 -3.71037304e-01 4.54223722e-01
1.22714508e+00 5.51164635e-02 4.50605929e-01 2.98640758e-01
5.44319153e-01 -6.04915023e-01 -5.89468420e-01 2.64491588e-01
-3.93032312e-01 -1.72955060e+00 -6.88832045e-01 1.74765065e-01
6.12921357e-01 7.62763441e-01 -1.55058093e-02 -2.22881913e-01
-7.96358511e-02 -1.14392090e+00 9.63704646e-01 5.16888142e-01
1.15077484e+00 -4.53940153e-01 5.21417022e-01 2.16897383e-01
-1.22812128e+00 -2.29400858e-01 1.03103891e-01 -1.10730588e-01
3.04251015e-01 4.97080058e-01 -7.85379410e-01 1.70049265e-01
8.73370528e-01 3.70332092e-01 -6.97145283e-01 8.68148446e-01
-4.73113745e-01 1.11709225e+00 -8.31044972e-01 4.34267282e-01
4.66634899e-01 -1.23007394e-01 1.52436748e-01 1.23426795e+00
5.62602580e-01 -2.84232013e-03 -1.57257304e-01 1.42815459e+00
1.42297402e-01 -6.83146536e-01 -8.23277652e-01 -6.45664275e-01
5.64809978e-01 7.57302225e-01 -9.48014617e-01 -4.86384369e-02
-1.59007564e-01 4.66421843e-01 -2.73962975e-01 3.66747558e-01
-5.47716975e-01 -3.30048442e-01 3.58217537e-01 5.01337647e-01
1.29774451e-01 -7.31609464e-01 -6.43120110e-01 -7.00130701e-01
1.71235371e-02 -2.08628505e-01 3.70419659e-02 -8.04737866e-01
-7.62071609e-01 -7.57721625e-03 3.37344706e-01 -5.02550066e-01
2.22254857e-01 -6.37920678e-01 -1.25947738e+00 1.21446061e+00
-1.93948400e+00 -1.00728834e+00 -3.40519905e-01 9.56688002e-02
1.30792826e-01 -6.03199527e-02 6.92277312e-01 1.13835871e-01
-4.35338169e-01 -3.36624473e-01 3.72432709e-01 4.15631309e-02
1.08988166e-01 -1.28029895e+00 5.34423113e-01 1.25110769e+00
-3.16660941e-01 4.07837868e-01 4.44801420e-01 -1.33961403e+00
-1.40070868e+00 -1.07659864e+00 3.59853595e-01 -3.98535639e-01
1.05947208e+00 -2.53626913e-01 -1.13118923e+00 6.49773419e-01
1.13699641e-02 2.91610569e-01 3.87924552e-01 -4.13619131e-01
-3.46160442e-01 -6.75163493e-02 -1.21011746e+00 -1.21921320e-02
4.27874088e-01 -1.14317524e+00 -4.01230216e-01 3.52014065e-01
1.92991555e-01 -3.96412611e-02 -6.99552834e-01 5.64397931e-01
5.08222699e-01 -5.57689905e-01 7.74716496e-01 -4.76683795e-01
5.74883342e-01 -4.24596310e-01 -5.22078156e-01 -1.09231126e+00
1.87136400e-02 -1.65682301e-01 -3.54721397e-01 1.22633040e+00
5.94446361e-01 -1.89150050e-01 1.11449373e+00 7.10355937e-01
-4.02713269e-01 -3.99265796e-01 -1.13631344e+00 -9.09983814e-01
6.77101016e-01 -4.04569983e-01 2.39192456e-01 7.48587370e-01
-5.81289195e-02 -4.00882572e-01 3.47645879e-02 7.59468436e-01
5.50077677e-01 -1.84659004e-01 -1.25015542e-01 -9.92746353e-01
-1.75811648e-02 -4.76489872e-01 1.38511822e-01 -4.30024952e-01
-1.73085704e-01 -7.41870701e-01 4.27127779e-01 -1.60214031e+00
8.77658799e-02 -4.42201078e-01 -5.63465297e-01 7.16495812e-01
-2.81535350e-02 1.86659187e-01 3.71003449e-02 3.93870890e-01
1.58056736e-01 6.98280632e-01 4.05548185e-01 -7.72731006e-01
1.00128815e-01 -4.09607857e-01 -6.15266301e-02 4.02432650e-01
1.10567582e+00 -6.01453424e-01 7.44346827e-02 -9.18529391e-01
3.93622696e-01 3.65844786e-01 5.92006564e-01 -1.31397605e+00
2.59368718e-01 -4.18345720e-01 4.51398373e-01 -8.60039055e-01
6.78432211e-02 -8.69069874e-01 3.71343553e-01 7.03586280e-01
-2.06307650e-01 9.18241888e-02 3.28994364e-01 6.35171771e-01
-2.83157766e-01 -2.17818663e-01 8.18730295e-01 -7.08715141e-01
-6.88439727e-01 4.66831028e-02 -9.93808448e-01 -3.60178739e-01
9.16598976e-01 1.50484033e-02 -7.62134790e-01 1.21870272e-01
-5.19622505e-01 1.69777229e-01 4.99007821e-01 1.73125982e-01
7.22663999e-01 -9.26191390e-01 -2.79581934e-01 4.79570515e-02
-1.36498004e-01 2.47638792e-01 3.17077525e-02 6.71528757e-01
-8.73670101e-01 3.33455592e-01 -1.25975117e-01 -3.55796367e-01
-4.83324707e-01 4.04472917e-01 9.08126295e-01 -5.12725413e-02
-4.64469820e-01 9.29312110e-01 -2.92521268e-01 -1.52196456e-02
-1.61732152e-01 -1.38935253e-01 1.67818606e-01 1.81710988e-01
5.75349212e-01 6.63890600e-01 4.58051890e-01 -2.62641311e-01
-5.62528312e-01 2.78131738e-02 5.35126388e-01 -5.56927145e-01
1.76293242e+00 6.31370395e-02 -3.42271566e-01 7.01137900e-01
9.69884872e-01 -1.37891471e-01 -1.54540300e+00 4.99727607e-01
5.64850152e-01 -3.77844214e-01 5.25988221e-01 -1.15490818e+00
-1.41343999e+00 1.29491174e+00 8.70213866e-01 -6.19291291e-02
8.35663676e-01 5.44739924e-02 3.26138496e-01 3.24666381e-01
1.67657420e-01 -9.79921460e-01 9.98088792e-02 3.51343751e-01
8.35247934e-01 -1.06907105e+00 1.36056151e-02 3.41480196e-01
1.32331606e-02 1.24684846e+00 6.20907307e-01 1.40979186e-01
7.07660556e-01 8.26920688e-01 -6.39757961e-02 -8.14931035e-01
-2.16966793e-01 3.05914998e-01 -5.78410149e-01 6.84741080e-01
1.53874606e-01 2.55740851e-01 4.13846403e-01 8.07995945e-02
5.49771845e-01 -3.09540004e-01 4.68252748e-01 1.35578346e+00
-7.38200486e-01 -5.43361247e-01 -5.89816928e-01 3.69707674e-01
-4.15402979e-01 -1.21668689e-01 -4.44711238e-01 2.67204106e-01
-1.72071382e-01 8.20831656e-01 1.42661154e-01 -1.91024970e-02
4.75074463e-02 3.70985478e-01 7.76423514e-02 -4.39708561e-01
-4.70673621e-01 4.31396216e-01 2.02534094e-01 -1.48377702e-01
-5.48057914e-01 -8.68680298e-01 -1.45297587e+00 1.51782200e-01
-4.93696183e-01 2.72442132e-01 1.34490192e+00 8.88995707e-01
9.17504355e-03 5.12878239e-01 6.27320826e-01 -1.19151676e+00
-4.35731262e-01 -1.06517029e+00 -7.45994925e-01 -7.23039582e-02
6.44724309e-01 -5.60366035e-01 -5.03792465e-01 -1.52313009e-01] | [6.823764324188232, 2.62823486328125] |
f5e96ba3-a6b4-4831-9993-89987db0cc71 | on-the-direct-maximization-of-quadratic | 1509.07107 | null | http://arxiv.org/abs/1509.07107v3 | http://arxiv.org/pdf/1509.07107v3.pdf | On The Direct Maximization of Quadratic Weighted Kappa | In recent years, quadratic weighted kappa has been growing in popularity in
the machine learning community as an evaluation metric in domains where the
target labels to be predicted are drawn from integer ratings, usually obtained
from human experts. For example, it was the metric of choice in several recent,
high profile machine learning contests hosted on Kaggle :
https://www.kaggle.com/c/asap-aes , https://www.kaggle.com/c/asap-sas ,
https://www.kaggle.com/c/diabetic-retinopathy-detection . Yet, little is
understood about the nature of this metric, its underlying mathematical
properties, where it fits among other common evaluation metrics such as mean
squared error (MSE) and correlation, or if it can be optimized analytically,
and if so, how. Much of this is due to the cumbersome way that this metric is
commonly defined. In this paper we first derive an equivalent but much simpler,
and more useful, definition for quadratic weighted kappa, and then employ this
alternate form to address the above issues. | ['Derek Justice', 'David Vaughn'] | 2015-09-23 | null | null | null | null | ['diabetic-retinopathy-detection'] | ['medical'] | [-1.48813784e-01 -2.35753935e-02 -3.43830764e-01 -6.19929969e-01
-1.05453479e+00 -7.39740908e-01 1.25664309e-01 5.96340775e-01
-5.60248852e-01 9.24348116e-01 1.42822564e-01 -3.72802854e-01
-5.10656118e-01 -3.73995662e-01 -3.22716266e-01 -8.59823883e-01
-1.40413389e-01 5.35884798e-01 2.17973411e-01 4.49952520e-02
3.55222464e-01 2.02309072e-01 -1.44845355e+00 7.49529377e-02
1.04295659e+00 1.15906751e+00 -1.59452975e-01 5.11611462e-01
3.29577893e-01 4.26972896e-01 -3.42997313e-01 -7.56431937e-01
2.76825368e-01 -2.89225787e-01 -6.32510543e-01 -2.01207653e-01
2.27284700e-01 -1.48699671e-01 1.70336023e-01 1.07271588e+00
4.93320107e-01 -1.13770843e-01 1.00335789e+00 -1.09725320e+00
-5.23156583e-01 3.04707944e-01 -7.90139616e-01 2.83285052e-01
2.55042493e-01 9.29101706e-02 1.21166635e+00 -7.30699539e-01
2.22569257e-01 9.15126681e-01 5.71990907e-01 1.93418100e-01
-1.15222239e+00 -6.70711040e-01 -1.55804664e-01 4.26941574e-01
-1.38390017e+00 -1.38340801e-01 4.64919776e-01 -6.59028471e-01
9.79798436e-02 5.44667184e-01 4.48408395e-01 8.34337652e-01
3.39306235e-01 6.90430343e-01 1.64619470e+00 -5.08864105e-01
1.96808830e-01 4.39566433e-01 3.28133851e-01 6.77567720e-01
3.73609960e-01 2.27523133e-01 -1.79325923e-01 -4.26234335e-01
5.17410338e-01 -2.77931511e-01 -2.55559474e-01 -6.52391970e-01
-9.73111629e-01 1.14694583e+00 3.12161595e-01 7.54087046e-02
-1.67990953e-01 -2.98035175e-01 4.21604276e-01 3.35971951e-01
4.62301642e-01 5.74289262e-01 -4.13449079e-01 -2.77199060e-01
-9.06221092e-01 3.62950414e-01 8.92746687e-01 4.66167808e-01
2.70660400e-01 -4.95308280e-01 1.06532060e-01 1.11359727e+00
2.42394954e-01 2.71572769e-01 2.76707172e-01 -1.29358387e+00
7.90620819e-02 4.30108726e-01 4.12553638e-01 -7.92002559e-01
-6.26152456e-01 -4.13540691e-01 -7.20696926e-01 5.00516176e-01
9.99810874e-01 1.08008618e-02 -5.66315651e-01 1.56307733e+00
1.72956541e-01 -3.74063551e-01 -3.83726418e-01 1.16138816e+00
7.19531476e-01 2.69844711e-01 1.11279055e-01 -2.69751966e-01
1.53095436e+00 -7.62125194e-01 -1.67367831e-01 -1.44935459e-01
4.33427036e-01 -8.79234195e-01 1.16529381e+00 7.38278389e-01
-1.01942849e+00 -2.08578840e-01 -1.01335025e+00 1.57065038e-02
-2.37100959e-01 1.53342053e-01 7.12241113e-01 8.86534631e-01
-8.45505476e-01 4.79311109e-01 -6.35833979e-01 -4.32052255e-01
2.38617539e-01 2.99510628e-01 -3.49131227e-01 -2.83603609e-01
-1.16347063e+00 1.26182508e+00 4.68481854e-02 -1.51497036e-01
-3.50990236e-01 -6.47155881e-01 -5.07835746e-01 -8.66258964e-02
3.37501556e-01 -5.91294706e-01 1.47980559e+00 -8.00203621e-01
-1.26714337e+00 1.09019315e+00 1.64594918e-01 -1.43609852e-01
5.10665715e-01 2.06756443e-02 -4.76848543e-01 2.34490819e-02
3.23843695e-02 3.41778576e-01 4.16894197e-01 -1.04634440e+00
-6.64769709e-01 -5.72329640e-01 1.14423193e-01 -5.79452664e-02
1.81331217e-01 3.69789958e-01 -9.61214006e-02 -5.93314350e-01
-1.15037926e-01 -1.18476033e+00 -3.29506367e-01 2.47593626e-01
-1.95806876e-01 -4.62650418e-01 -1.10065624e-01 -5.29193819e-01
1.31702006e+00 -2.26486802e+00 -3.92476499e-01 6.03037477e-01
4.13807303e-01 2.72858739e-01 1.64055571e-01 4.94690835e-01
-2.25466102e-01 -2.40739882e-02 -3.24701279e-01 2.43708745e-01
-1.27737550e-02 -1.97090387e-01 1.04210116e-01 5.75955927e-01
9.42486990e-03 3.94272894e-01 -1.10252762e+00 -3.87614459e-01
7.64619000e-03 3.79673332e-01 -3.67745221e-01 -2.27248773e-01
3.71222019e-01 2.62787968e-01 -3.61744046e-01 6.51421607e-01
5.52553117e-01 -4.71103996e-01 2.26248682e-01 -2.12317958e-01
-2.58293986e-01 1.37326926e-01 -1.15657866e+00 1.11295342e+00
-5.43736964e-02 7.02777922e-01 6.31246120e-02 -1.09091008e+00
8.60101104e-01 1.35603100e-01 4.78949696e-01 -4.57510322e-01
1.91925749e-01 4.73802567e-01 2.75761545e-01 -3.19446951e-01
3.47894669e-01 -3.74101311e-01 4.78715450e-02 4.08563524e-01
-3.00317943e-01 -1.07628323e-01 3.84252489e-01 -1.02653436e-01
1.07031107e+00 -2.72184730e-01 7.85600662e-01 -4.31445092e-01
3.12673301e-01 1.29452571e-01 3.93226653e-01 6.18063986e-01
-4.84789938e-01 8.43326807e-01 6.56185150e-01 -1.96212396e-01
-9.22462046e-01 -1.30821764e+00 -6.89021766e-01 8.09871495e-01
-4.31259014e-02 -4.04616117e-01 -5.33211231e-01 -4.50442642e-01
2.15834349e-01 7.08882034e-01 -5.77894807e-01 -1.40655145e-01
-2.74834856e-02 -7.98628628e-01 3.12266856e-01 2.91530967e-01
-1.86984483e-02 -6.23326838e-01 -4.24688637e-01 6.06270097e-02
4.12709862e-02 -6.10374212e-01 -4.85778481e-01 1.34281904e-01
-8.01217496e-01 -1.37611318e+00 -8.41898084e-01 -3.82883132e-01
6.62274897e-01 9.11091119e-02 1.24465764e+00 -1.45755649e-01
-9.06467512e-02 2.88972884e-01 -4.98225123e-01 -6.82907462e-01
-2.40369737e-01 -1.72132701e-01 6.02931157e-02 -4.35472699e-03
5.70160091e-01 -3.76323313e-01 -8.86049569e-01 8.02114189e-01
-8.11873019e-01 -3.61421645e-01 9.19417083e-01 8.51463377e-01
4.33929056e-01 -1.40686035e-01 5.54754198e-01 -1.09936452e+00
8.46296608e-01 -3.93290550e-01 -5.91149449e-01 3.63927841e-01
-1.06029844e+00 -2.06122920e-01 2.73951471e-01 -2.51894802e-01
-4.72525686e-01 -3.67343813e-01 -4.23714444e-02 -4.62945253e-02
-5.22595271e-02 8.06055248e-01 3.30356717e-01 -7.15161785e-02
1.00824881e+00 -3.31423968e-01 1.92170173e-01 -4.18949008e-01
1.52706861e-01 9.42594707e-01 3.17297071e-01 -5.16889095e-01
7.79178679e-01 1.77514121e-01 -5.17809018e-02 -5.04426956e-01
-1.03696465e+00 -6.16855025e-01 -2.51157343e-01 -1.50029927e-01
7.39423335e-01 -6.03027582e-01 -8.32592309e-01 2.90904552e-01
-8.63018572e-01 -2.08995998e-01 -1.83821276e-01 9.70920205e-01
-5.84296346e-01 2.71035701e-01 -3.14776599e-01 -7.69440174e-01
-2.73750931e-01 -1.18300807e+00 5.65288246e-01 3.21159065e-01
-6.74940526e-01 -9.44537282e-01 2.10162640e-01 7.61459887e-01
3.56566370e-01 2.79659152e-01 1.02237201e+00 -5.62670767e-01
-3.72885242e-02 -7.32067943e-01 -5.46667755e-01 7.84463584e-01
3.87431234e-02 2.01751813e-01 -8.14919710e-01 -3.07665616e-01
-2.34589353e-01 -2.30594337e-01 5.71212411e-01 6.43599272e-01
1.24181604e+00 -1.65728852e-01 -1.17700763e-01 3.31114262e-01
1.34563684e+00 3.46462190e-01 4.92037117e-01 4.33062643e-01
7.77019188e-02 7.77036309e-01 7.84910381e-01 2.55914509e-01
5.90330541e-01 8.48675013e-01 1.83044404e-01 -1.73711926e-02
2.69112408e-01 1.05829753e-01 1.41815737e-01 7.89630949e-01
-3.16604912e-01 1.53597474e-01 -1.10572898e+00 3.60216796e-01
-1.81031990e+00 -8.63697767e-01 -2.07217693e-01 2.53955245e+00
8.42180967e-01 2.63501734e-01 3.26509923e-01 3.26977313e-01
6.00447476e-01 -5.73892631e-02 -3.47067833e-01 -7.20125973e-01
6.24621697e-02 2.36970395e-01 6.86872721e-01 3.45273733e-01
-9.55643296e-01 1.86739638e-01 5.86984682e+00 1.04683459e+00
-9.90267336e-01 2.35609204e-01 9.05285060e-01 -1.03251815e-01
-3.76124531e-02 1.39872715e-01 -4.05718595e-01 5.74539244e-01
6.74377859e-01 -5.21054626e-01 6.15323298e-02 7.90746808e-01
4.35559869e-01 -4.96691883e-01 -9.83205140e-01 9.50518906e-01
-9.05536860e-02 -8.53394568e-01 -5.37823558e-01 6.01084679e-02
6.17466569e-01 4.33500996e-03 1.25618473e-01 1.67179719e-01
1.81632936e-01 -1.13175893e+00 5.70133865e-01 4.39182252e-01
8.79967213e-01 -5.39598703e-01 8.71579170e-01 7.93106761e-03
-6.22416794e-01 -1.33997262e-01 -3.03260356e-01 -9.84982997e-02
-6.49782876e-03 1.33236289e+00 -6.65816784e-01 4.69691873e-01
6.86309397e-01 2.03148142e-01 -6.10057473e-01 1.73748457e+00
-1.72740251e-01 8.88061583e-01 -2.66208291e-01 -1.81981996e-01
2.10518822e-01 -3.76939058e-01 4.68745470e-01 1.00120413e+00
4.05962288e-01 -1.87852804e-03 -1.11282945e-01 4.60443825e-01
2.50340104e-01 3.50831270e-01 -2.89548129e-01 1.38361782e-01
3.17369699e-01 1.41731727e+00 -5.88219285e-01 8.68578106e-02
-7.04706609e-01 2.76583821e-01 1.75646260e-01 3.12241822e-01
-7.18243301e-01 -2.59218186e-01 5.35318017e-01 3.58368188e-01
-1.11850686e-01 -1.76567603e-02 -3.20265591e-01 -8.92212510e-01
7.11262822e-02 -9.47476864e-01 6.51609004e-01 -8.78847659e-01
-1.65449703e+00 3.40538204e-01 1.59539938e-01 -1.51407385e+00
-9.50771645e-02 -7.41918385e-01 -4.44142908e-01 9.21695352e-01
-1.17663491e+00 -5.72372615e-01 -2.15757027e-01 3.82108778e-01
5.28034233e-02 -4.92681898e-02 6.93069041e-01 4.41224068e-01
-4.66009945e-01 7.32908905e-01 3.82846922e-01 5.96145801e-02
1.17912114e+00 -1.43243706e+00 -2.13788897e-01 2.76791781e-01
-1.57523587e-01 4.83145893e-01 9.54125226e-01 -5.04353866e-02
-7.76190281e-01 -6.92665994e-01 7.05750763e-01 -6.19935274e-01
9.01414871e-01 2.17895806e-01 -7.59840369e-01 2.71347255e-01
-2.96933740e-01 -8.75863656e-02 1.08387411e+00 4.25379843e-01
-4.12913740e-01 -4.59029675e-01 -1.14621246e+00 6.24347925e-01
4.52938616e-01 -2.40175903e-01 -3.46378803e-01 4.86068368e-01
1.18164476e-02 -3.36130947e-01 -1.21433651e+00 4.71883386e-01
9.94961083e-01 -1.18719161e+00 5.70109248e-01 -3.74224395e-01
5.07425189e-01 -3.29993188e-01 -2.04881728e-01 -1.51431620e+00
-3.64279330e-01 -2.43962333e-01 3.97214830e-01 9.92846489e-01
8.62578988e-01 -8.93793523e-01 7.99350083e-01 7.63988674e-01
-1.70100078e-01 -1.37848127e+00 -1.07864130e+00 -8.24371934e-01
1.53394789e-01 -4.10658479e-01 1.52135968e-01 1.03203297e+00
1.82181984e-01 2.34327003e-01 -2.34717086e-01 -1.59900919e-01
7.65677214e-01 -1.01896301e-01 4.87872392e-01 -1.42901802e+00
-2.77283788e-01 -7.92969346e-01 -6.70674801e-01 -5.14200866e-01
-4.31981713e-01 -7.55192339e-01 -2.34240532e-01 -1.39648092e+00
2.39385977e-01 -7.27643967e-01 -6.05072975e-01 1.96369216e-01
-3.39385867e-01 3.86490136e-01 1.84374616e-01 2.56639034e-01
-4.25648093e-01 3.15853953e-02 1.09065855e+00 1.85411111e-01
-8.59938860e-02 4.95225996e-01 -1.04299104e+00 7.20347703e-01
1.03633976e+00 -4.84172523e-01 -3.64783406e-01 -7.89668337e-02
5.51803827e-01 1.17362831e-02 3.77487183e-01 -9.21551645e-01
1.86684459e-01 -3.80831987e-01 2.00550348e-01 -1.32396132e-01
1.27542093e-01 -7.22312450e-01 1.78384870e-01 2.49415517e-01
-2.99653023e-01 2.14429975e-01 -1.22917086e-01 2.20863789e-01
-3.82843941e-01 -4.78153914e-01 9.90292251e-01 -4.55144197e-02
-3.48021209e-01 1.44552439e-01 -1.03689082e-01 9.90438163e-02
1.06613708e+00 -2.70853341e-01 -5.09382606e-01 -4.98977751e-01
-6.09347165e-01 2.41757318e-01 4.50765491e-01 2.10673407e-01
3.35717678e-01 -1.26363087e+00 -7.47034311e-01 -2.59037495e-01
5.19974828e-01 -3.52677882e-01 9.74387825e-02 1.77376842e+00
-5.61635733e-01 4.54357624e-01 -1.55989498e-01 -2.94794679e-01
-1.27346945e+00 2.89208263e-01 2.03174248e-01 -3.59245032e-01
-1.75790071e-01 6.48866773e-01 2.19372675e-01 -2.26560563e-01
9.09205079e-02 -1.06389359e-01 -1.24993354e-01 2.04349488e-01
5.11128128e-01 6.80605173e-01 1.70941547e-01 -2.79808849e-01
-4.76236045e-01 5.99148273e-01 -4.13916893e-02 -6.27701133e-02
1.19461632e+00 -7.40981847e-03 -2.37140372e-01 7.03244805e-01
9.51630831e-01 1.22420684e-01 -7.42771208e-01 -5.04449122e-02
1.12958342e-01 -3.84071380e-01 6.31663809e-03 -1.01679957e+00
-9.83694792e-01 6.10084474e-01 7.87910283e-01 4.84465033e-01
1.05905914e+00 -2.11061575e-02 3.61154467e-01 7.24426731e-02
3.22966516e-01 -1.13831353e+00 -3.73727709e-01 1.43381417e-01
9.39688981e-01 -1.31064820e+00 3.04599702e-01 -4.89122957e-01
-7.89944291e-01 9.54719961e-01 2.01103300e-01 1.84069865e-03
9.86311197e-01 -7.31768683e-02 4.37776387e-01 -2.34789357e-01
-4.79806304e-01 -1.54952258e-01 5.51535845e-01 5.65531373e-01
6.78818107e-01 4.86761183e-01 -8.55140388e-01 6.24165654e-01
-4.92542982e-01 5.00295684e-02 4.23403621e-01 5.65131187e-01
-3.97539318e-01 -1.12663805e+00 -3.24548602e-01 1.11317027e+00
-5.86312711e-01 -3.71162407e-02 -2.34654218e-01 6.14661396e-01
4.92633146e-04 1.11397588e+00 -2.76344955e-01 -3.76779288e-01
3.70365202e-01 7.23001286e-02 3.76997441e-01 -5.24426937e-01
-3.80731136e-01 -3.59244972e-01 2.43413374e-01 -4.98948932e-01
-4.29013640e-01 -7.70881355e-01 -7.09816098e-01 -5.65052986e-01
-3.68681788e-01 6.15110576e-01 7.24029958e-01 7.91046679e-01
-2.46526469e-02 -3.77888009e-02 7.12213755e-01 -4.82949078e-01
-7.60541737e-01 -9.57826078e-01 -9.86407280e-01 2.61406898e-01
2.92478889e-01 -9.07132864e-01 -6.28500044e-01 -4.95294273e-01] | [8.542142868041992, 4.427086353302002] |
90c76ddd-dcc7-42eb-a5c6-3b81c2adb43a | personalised-recommendations-of-sleep | 2208.00033 | null | https://arxiv.org/abs/2208.00033v1 | https://arxiv.org/pdf/2208.00033v1.pdf | Personalised recommendations of sleep behaviour with neural networks using sleep diaries captured in Sleepio | SleepioTM is a digital mobile phone and web platform that uses techniques from cognitive behavioural therapy (CBT) to improve sleep in people with sleep difficulty. As part of this process, Sleepio captures data about the sleep behaviour of the users that have consented to such data being processed. For neural networks, the scale of the data is an opportunity to train meaningful models translatable to actual clinical practice. In collaboration with Big Health, the therapeutics company that created and utilizes Sleepio, we have analysed data from a random sample of 401,174 sleep diaries and built a neural network to model sleep behaviour and sleep quality of each individual in a personalised manner. We demonstrate that this neural network is more accurate than standard statistical methods in predicting the sleep quality of an individual based on his/her behaviour from the last 10 days. We compare model performance in a wide range of hyperparameter settings representing various scenarios. We further show that the neural network can be used to produce personalised recommendations of what sleep habits users should follow to maximise sleep quality, and show that these recommendations are substantially better than the ones generated by standard methods. We finally show that the neural network can explain the recommendation given to each participant and calculate confidence intervals for each prediction, all of which are essential for clinicians to be able to adopt such a tool in clinical practice. | ['Chris Miller', 'Tom Walker', 'Kate Saunders', 'Niall Taylor', 'Jenny Gu', 'Alasdair Henry', 'Maria Liakata', 'Colin Espie', 'Alejo Nevado-Holgado'] | 2022-07-29 | null | null | null | null | ['sleep-quality-prediction'] | ['medical'] | [ 2.46699959e-01 2.96428949e-01 -5.16589940e-01 -6.49657428e-01
-1.80314764e-01 -3.95920351e-02 -1.15238965e-01 1.55038506e-01
-7.09228575e-01 8.22404742e-01 7.80219913e-01 -6.81823075e-01
-4.13293540e-01 -4.43312883e-01 -6.09775819e-03 -6.17182791e-01
3.47118676e-02 6.67760789e-01 -2.02210903e-01 5.16176969e-02
-2.25139428e-02 6.12022430e-02 -1.37128317e+00 1.77977413e-01
8.58263433e-01 8.32701206e-01 1.61399081e-01 7.21287668e-01
3.25035512e-01 4.84553695e-01 -6.04383528e-01 -2.29263939e-02
9.29068998e-02 -4.67440724e-01 -5.73067904e-01 -1.19630583e-01
-6.48468034e-03 -2.18690097e-01 -6.90225735e-02 4.90500748e-01
8.22193027e-01 5.73223889e-01 3.63189548e-01 -1.03503215e+00
-3.43863547e-01 2.30562419e-01 4.20213670e-01 7.51498759e-01
3.90110254e-01 4.97357100e-01 5.75683951e-01 1.39659941e-01
-1.09957233e-01 7.86972880e-01 1.05652916e+00 9.18474674e-01
-1.44845819e+00 -7.06796408e-01 -4.58106011e-01 3.59287888e-01
-1.05205226e+00 -7.66484439e-01 1.65804595e-01 -9.52260792e-02
1.30784738e+00 6.63711071e-01 1.39912045e+00 1.10665178e+00
7.21907616e-01 1.26198381e-01 1.08129203e+00 -2.76636302e-01
7.15340972e-01 2.26057455e-01 2.99342930e-01 1.99925244e-01
-9.03308112e-03 2.15870529e-01 -4.06201690e-01 -3.94812256e-01
3.13530803e-01 6.08138502e-01 -2.27976870e-02 1.03726797e-01
-5.68129301e-01 8.77367437e-01 3.01601887e-01 2.55813360e-01
-5.58321953e-01 -1.00989817e-02 2.01104924e-01 8.08289722e-02
6.92734838e-01 4.43047911e-01 -6.59481347e-01 -6.05327249e-01
-1.29692650e+00 5.59159834e-03 1.03908634e+00 3.43466371e-01
2.15860054e-01 -2.48917654e-01 -4.01278496e-01 1.08116162e+00
3.81129146e-01 4.62787211e-01 1.13678277e+00 -1.36007333e+00
-1.35945007e-01 6.40982270e-01 2.24663883e-01 -8.01193953e-01
-1.12300766e+00 -2.73788989e-01 -7.49830604e-01 -2.22981617e-01
7.01599941e-02 -3.01033199e-01 -5.56739986e-01 1.45137703e+00
-3.25696580e-02 9.33774784e-02 -2.01489300e-01 3.13785493e-01
6.72314882e-01 2.28580490e-01 3.64169896e-01 -6.54420793e-01
1.33014369e+00 -5.84810734e-01 -7.50210285e-01 -7.53627479e-01
6.20827377e-01 -3.17526251e-01 1.06371927e+00 6.28309608e-01
-1.41497076e+00 -2.42390513e-01 -8.35958719e-01 7.47472346e-02
-2.64094025e-01 -2.71292716e-01 6.40910268e-01 1.14555264e+00
-1.52525330e+00 8.62444937e-01 -1.14313507e+00 -9.40931380e-01
5.94930768e-01 1.13061285e+00 2.79793758e-02 1.59748718e-01
-9.54499066e-01 1.06899512e+00 2.48649508e-01 -8.61003995e-02
-2.70189703e-01 -8.62958968e-01 -5.53792357e-01 4.16526198e-01
1.59590855e-01 -1.25945973e+00 1.51899874e+00 -8.10665548e-01
-1.46908379e+00 6.39545977e-01 -3.75563025e-01 -8.88151288e-01
-1.80796295e-01 9.57830772e-02 -6.90257311e-01 -1.82679608e-01
-9.60544869e-03 5.17221987e-01 4.32141870e-01 -1.92602500e-01
-7.50448823e-01 -5.06060719e-01 -2.20836356e-01 1.18473552e-01
-5.58478832e-01 8.69393907e-03 -2.55373001e-01 -9.26894322e-02
-6.25972152e-01 -1.10676527e+00 -5.26998758e-01 -2.43431464e-01
-2.67379582e-01 -3.45897555e-01 1.76169723e-01 -5.17210484e-01
1.57541132e+00 -1.80450463e+00 -3.85702878e-01 1.46617830e-01
4.27884281e-01 3.55447173e-01 2.90273279e-01 3.05155516e-01
6.89332783e-02 2.32481733e-01 1.47105223e-02 -5.84876180e-01
-7.00450689e-02 5.88917613e-01 4.49087709e-01 4.58230615e-01
-4.67067122e-01 7.83040583e-01 -6.14209175e-01 -2.17869014e-01
3.83124888e-01 5.33498406e-01 -9.52632606e-01 2.96624064e-01
1.00442395e-01 2.74186045e-01 -2.72536904e-01 2.32583005e-02
6.19899202e-03 -5.97225368e-01 2.36301795e-01 4.37455446e-01
1.88955382e-01 7.23945379e-01 -3.01574022e-01 1.39420009e+00
-7.35512137e-01 4.51143682e-01 -2.30641931e-01 -6.43387973e-01
6.01541102e-01 2.65779108e-01 6.27229691e-01 -8.94906282e-01
3.06794137e-01 -2.90002704e-01 1.59982190e-01 -7.00179875e-01
1.13620691e-01 -7.99538434e-01 2.45391935e-01 7.63025522e-01
-3.58834356e-01 2.05584601e-01 -8.24899524e-02 -1.16815254e-01
1.58953834e+00 -5.64135373e-01 7.94202507e-01 -2.52752423e-01
1.46989688e-01 -2.99026877e-01 6.07297361e-01 8.06517899e-01
-3.72337461e-01 2.95293450e-01 1.96476802e-01 -5.88012278e-01
-8.29914033e-01 -6.61911190e-01 -1.23624638e-01 1.00702894e+00
-5.88495195e-01 -6.68459952e-01 -7.04962194e-01 -4.56847697e-01
-1.77641720e-01 1.23050094e+00 -1.04329348e+00 -8.25903654e-01
7.37572685e-02 -1.04851508e+00 1.72343463e-01 4.59903628e-01
8.82651061e-02 -1.45946038e+00 -8.71069491e-01 3.19058776e-01
-1.48165986e-01 -5.57095349e-01 -7.06171155e-01 3.98223907e-01
-1.09353423e+00 -1.13778651e+00 -2.78170794e-01 -8.59633386e-02
2.76490688e-01 1.36391968e-01 1.18476021e+00 4.24506009e-01
-1.71796992e-01 5.61432242e-01 -4.85687628e-02 -5.98526955e-01
-4.90061939e-01 8.43523964e-02 4.80209857e-01 -3.91332924e-01
9.70978677e-01 -1.10995364e+00 -1.39367831e+00 3.18668038e-01
-6.64088488e-01 -1.29362002e-01 4.99874294e-01 4.35853094e-01
1.98180780e-01 2.05538601e-01 4.26643014e-01 -9.30952430e-01
1.03289807e+00 -1.00795960e+00 -3.85703407e-02 -2.17424288e-01
-1.50595677e+00 -1.79139063e-01 4.31740701e-01 -3.44554067e-01
-5.93776226e-01 -1.34517401e-01 -3.85591924e-01 4.24526520e-02
-4.65488225e-01 2.54398674e-01 2.21340597e-01 4.63552982e-01
9.07888889e-01 1.59962222e-01 4.53418702e-01 -5.88324666e-01
6.02565333e-02 1.09473431e+00 3.20866406e-01 1.72034338e-01
1.76622961e-02 1.81251079e-01 -1.37846738e-01 -7.93751776e-01
-9.68317747e-01 -6.98092878e-01 -2.29136929e-01 9.94311646e-02
9.99154568e-01 -6.22124672e-01 -1.38282835e+00 -1.63253158e-01
-2.99943149e-01 -1.09356081e+00 -4.29391325e-01 3.48243803e-01
-8.48152936e-01 -1.68093085e-01 -5.54092340e-02 -8.55803788e-01
-8.05090427e-01 -7.77312517e-01 5.63798785e-01 5.28212249e-01
-9.78957415e-01 -1.51310468e+00 4.24385101e-01 4.99278933e-01
7.23438561e-01 -2.29531601e-01 7.86799431e-01 -1.04367936e+00
1.66725203e-01 -1.73482835e-01 3.08937043e-01 4.14335459e-01
6.48763537e-01 -5.21958470e-01 -7.99569130e-01 -2.81525522e-01
4.61140454e-01 1.45512670e-02 2.60980725e-01 1.06183338e+00
1.21407092e+00 -8.33792746e-01 -4.77255285e-01 5.80017567e-01
1.03204525e+00 3.99548769e-01 8.58802497e-01 6.02805555e-01
1.93303421e-01 2.91522622e-01 -4.18910272e-02 3.23355407e-01
7.56877244e-01 8.04751039e-01 1.62422255e-01 8.33937228e-02
1.97978377e-01 1.85919002e-01 4.24385697e-01 3.78865242e-01
3.40273455e-02 -1.41229317e-01 -7.42051840e-01 2.87801355e-01
-1.70852900e+00 -9.49465930e-01 1.61286563e-01 2.33344626e+00
8.39491487e-01 2.69246608e-01 1.00469148e+00 -2.85795834e-02
1.51921153e-01 -4.25288647e-01 -7.55499184e-01 -1.00257158e+00
5.61674118e-01 5.10363400e-01 5.48618436e-01 3.19634438e-01
-3.89400363e-01 4.92723435e-01 7.95697021e+00 5.24643719e-01
-8.05843353e-01 4.05912638e-01 7.44299114e-01 -8.47285032e-01
1.85695127e-01 -3.62823784e-01 -5.61405241e-01 9.60902214e-01
2.43326473e+00 -2.80707181e-01 9.55492795e-01 6.11817539e-01
1.38935995e+00 -4.06311423e-01 -9.31542218e-01 7.53794611e-01
-7.48771057e-02 -1.30369389e+00 -7.15036750e-01 4.52446282e-01
4.16652173e-01 1.49470165e-01 3.18122841e-02 3.95198762e-01
5.71029961e-01 -1.32860601e+00 -6.57270104e-02 1.09192467e+00
5.94328046e-01 -7.57165313e-01 8.41498852e-01 5.19418478e-01
-3.33280951e-01 -5.25023639e-01 -2.84387350e-01 -3.39526594e-01
-1.06982857e-01 4.39609528e-01 -1.44623601e+00 -9.26993713e-02
1.00971842e+00 8.00881565e-01 -8.96730661e-01 1.37177289e+00
1.23233058e-01 1.14715374e+00 -4.73666877e-01 -1.68881238e-01
-1.14561871e-01 -7.63867702e-03 -2.52142977e-02 8.63919079e-01
3.25987160e-01 3.44316453e-01 -3.87864411e-01 7.43604302e-01
2.99296409e-01 -9.69657302e-03 -4.40158993e-01 -9.01764110e-02
3.03269982e-01 1.25696814e+00 -6.13547802e-01 -1.37632683e-01
-3.32767934e-01 6.87808573e-01 1.29602794e-02 4.85289283e-02
-1.89940855e-01 2.45370060e-01 7.43495464e-01 5.33495128e-01
9.37643200e-02 3.37242842e-01 -5.35994947e-01 -4.49526578e-01
-5.56923270e-01 -1.04916024e+00 5.72732687e-01 -1.02253067e+00
-1.27437830e+00 3.42768162e-01 -6.57934248e-02 -9.67625380e-01
-4.87389654e-01 -2.53337890e-01 -1.00129902e+00 8.28338623e-01
-5.66790819e-01 -6.29841685e-01 -1.60822213e-01 4.95205224e-01
4.44512993e-01 -5.80135994e-02 1.02541304e+00 9.39208865e-02
-7.56816328e-01 5.01845121e-01 1.66015774e-01 -6.46970034e-01
6.52157903e-01 -1.45678687e+00 1.93143591e-01 4.73330356e-03
-2.14662075e-01 1.17662489e+00 8.75838816e-01 -6.69834316e-01
-8.32198858e-01 -1.03132534e+00 9.12653685e-01 -8.51862907e-01
3.57602924e-01 -1.53758377e-02 -5.99552631e-01 8.36910605e-01
2.16489375e-01 -5.59480906e-01 1.51342630e+00 5.68525434e-01
7.60333776e-01 -2.84031004e-01 -1.30558538e+00 5.35663903e-01
7.09849656e-01 -2.55344987e-01 -6.26112103e-01 4.96636927e-01
5.46408057e-01 -1.65996671e-01 -9.91893351e-01 -2.58264661e-01
6.13210678e-01 -1.38913786e+00 9.34437990e-01 -7.40933895e-01
2.15499252e-01 3.36008459e-01 2.38230705e-01 -1.54036534e+00
-3.74347061e-01 -8.84949207e-01 -2.90796757e-01 7.57979751e-01
3.32606792e-01 -6.44745290e-01 1.09142315e+00 1.42396522e+00
-1.93224058e-01 -1.00864005e+00 -1.00074625e+00 -4.14692432e-01
5.68149462e-02 -6.14317715e-01 8.34774733e-01 2.65667528e-01
2.98577189e-01 4.61804897e-01 -5.00899434e-01 -2.28884801e-01
1.82793558e-01 -6.89937949e-01 5.04242539e-01 -1.34416771e+00
-3.46353859e-01 -3.49882603e-01 -3.37851435e-01 -5.67311347e-01
-4.64187041e-02 -7.87226677e-01 -1.40179880e-02 -1.77220547e+00
5.26831925e-01 -3.88511926e-01 -5.33733308e-01 8.70290160e-01
-1.99039102e-01 2.85816550e-01 -2.04267338e-01 9.85863879e-02
-5.54695964e-01 3.88736874e-01 8.09101224e-01 1.66471913e-01
-9.01608169e-01 8.55462193e-01 -1.51777434e+00 7.31977522e-01
1.14008987e+00 -7.39519477e-01 -7.12586045e-01 3.86565983e-01
3.73034328e-01 1.26841590e-01 3.90253633e-01 -1.20094204e+00
3.76133710e-01 -1.48748100e-01 5.86948574e-01 -1.70342594e-01
4.55117673e-01 -9.60326910e-01 3.71167779e-01 5.36087632e-01
-4.60767508e-01 1.39057124e-02 4.82726485e-01 6.46437049e-01
8.49444926e-01 -7.51730800e-02 5.21620035e-01 -1.03937231e-01
3.10993828e-02 2.08998591e-01 -9.02490377e-01 -4.21327353e-02
5.42128921e-01 -3.19613099e-01 -2.02557877e-01 -8.63039017e-01
-1.37687445e+00 3.25570017e-01 4.50437576e-01 1.72921613e-01
3.98386627e-01 -1.15749502e+00 -2.40622945e-02 2.87571728e-01
-1.22438625e-01 -4.89156663e-01 4.37164247e-01 1.30588853e+00
1.07646296e-02 7.18791842e-01 -2.18570769e-01 -3.30563009e-01
-1.36424112e+00 5.82439125e-01 6.72844052e-01 -2.16832817e-01
-8.10260773e-01 2.71189421e-01 -2.90192544e-01 -1.74428627e-01
1.47344962e-01 -3.42345327e-01 -2.76867777e-01 -3.17887574e-01
8.47180426e-01 4.74115998e-01 2.27996007e-01 -2.90879309e-01
-2.51533717e-01 -8.49196315e-02 1.68529861e-02 1.09585658e-01
1.73202670e+00 -5.21229684e-01 6.71885256e-03 6.11559212e-01
8.92159700e-01 -1.46603331e-01 -1.25620019e+00 9.67797861e-02
-2.05465332e-01 -1.12399377e-01 3.17998737e-01 -1.19454265e+00
-8.13983202e-01 4.23404008e-01 1.25255442e+00 5.75851262e-01
1.28255117e+00 -8.87508169e-02 9.46848333e-01 3.64656985e-01
-9.14218947e-02 -9.90702212e-01 -1.30770117e-01 -2.34656885e-01
3.76923293e-01 -1.01703560e+00 2.16229439e-01 4.14604276e-01
-6.39781356e-01 8.69479060e-01 1.48271069e-01 1.31088257e-01
9.01518226e-01 -3.44337940e-01 4.55274954e-02 -3.89893979e-01
-9.28025544e-01 1.15346559e-03 4.01280850e-01 9.73990262e-01
1.46996737e-01 2.42514998e-01 -2.31583431e-01 1.00763726e+00
-7.30025351e-01 6.77199125e-01 6.41685724e-01 5.39776206e-01
-5.71090937e-01 -1.32963932e+00 -2.34065175e-01 1.30860710e+00
-5.91328263e-01 -2.86916643e-01 -2.15079010e-01 4.43157226e-01
2.64603049e-01 1.55524170e+00 1.13821462e-01 -5.11285603e-01
3.00373912e-01 3.41280878e-01 8.46504048e-02 -1.00665700e+00
-7.90113807e-01 -9.42642912e-02 2.54951775e-01 -8.06098938e-01
-3.42436612e-01 -7.34401286e-01 -1.04216480e+00 -6.27338648e-01
1.28090560e-01 2.74941951e-01 3.89584899e-01 1.25743449e+00
5.15827060e-01 6.30297720e-01 3.60316902e-01 -8.75012279e-01
-6.56439587e-02 -1.04563391e+00 -7.48424530e-01 1.28667474e-01
4.47663099e-01 -5.05004525e-01 -4.30807680e-01 -1.33682221e-01] | [13.53902816772461, 3.486248731613159] |
cd34ad54-8f62-4200-910a-af8a2f738286 | consistent-structural-relation-learning-for | null | null | http://proceedings.neurips.cc/paper/2020/hash/7504adad8bb96320eb3afdd4df6e1f60-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/7504adad8bb96320eb3afdd4df6e1f60-Paper.pdf | Consistent Structural Relation Learning for Zero-Shot Segmentation | Zero-shot semantic segmentation aims to recognize the semantics of pixels from unseen categories with zero training samples. Previous practice [1] proposed to train the classifiers for unseen categories using the visual features generated from semantic word embeddings. However, the generator is merely learned on the seen categories while no constraint is applied to the unseen categories, leading to poor generalization ability. In this work, we propose a Consistent Structural Relation Learning (CSRL) approach to constrain the generating of unseen visual features by exploiting the structural relations between seen and unseen categories. We observe that different categories are usually with similar relations in either semantic word embedding space or visual feature space. This observation motivates us to harness the similarity of category-level relations on the semantic word embedding space to learn a better visual feature generator. Concretely, by exploring the pair-wise and list-wise structures, we impose the relations of generated visual features to be consistent with their counterparts in the semantic word embedding space. In this way, the relations between seen and unseen categories will be transferred to implicitly constrain the generator to produce relation-consistent unseen visual features. We conduct extensive experiments on Pascal-VOC and Pascal-Context benchmarks. The proposed CSRL significantly outperforms existing state-of-the-art methods by a large margin, resulting in ~7-12% on Pascal-VOC and ~2-5% on Pascal-Context. | ['Yi Yang', 'Yunchao Wei', 'Peike Li'] | 2020-12-01 | null | null | null | neurips-2020-12 | ['zero-shot-segmentation'] | ['computer-vision'] | [ 4.40379977e-01 1.40439555e-01 -1.63250864e-01 -8.07891607e-01
-2.66645044e-01 -6.07870460e-01 7.27033257e-01 1.50792137e-01
-3.58493090e-01 3.73371869e-01 1.94000408e-01 -4.23243567e-02
9.49488059e-02 -8.49992633e-01 -7.63521016e-01 -7.76062191e-01
2.87231773e-01 -4.64999191e-02 3.85677844e-01 -2.14889646e-01
2.43380055e-01 1.08879514e-01 -1.96742725e+00 4.15088236e-01
6.50420606e-01 1.08133471e+00 3.47073913e-01 4.22003090e-01
-4.82601464e-01 5.33498108e-01 -4.32669133e-01 -1.81034848e-01
2.76475698e-01 -6.41621530e-01 -7.02709079e-01 4.32146311e-01
6.55059993e-01 2.72401255e-02 -2.15064928e-01 1.22494280e+00
1.06944188e-01 4.72958028e-01 8.64014089e-01 -1.44402540e+00
-1.00912023e+00 1.46118298e-01 -4.45984006e-01 1.18282117e-01
2.59518445e-01 2.88315922e-01 1.38848519e+00 -9.37527120e-01
7.09934115e-01 1.31602871e+00 1.27391353e-01 8.50109339e-01
-1.31197238e+00 -5.06678641e-01 5.16499281e-01 4.54999745e-01
-1.34487069e+00 -2.06198663e-01 8.63933265e-01 -5.36566019e-01
8.44191432e-01 3.12994450e-01 6.07659101e-01 1.13391280e+00
-8.52563307e-02 6.75264180e-01 1.07358885e+00 -5.19103765e-01
4.30266321e-01 3.40391248e-01 4.65335220e-01 5.96412182e-01
3.55751693e-01 9.31398720e-02 -4.95781243e-01 2.18845323e-01
5.30327260e-01 2.18609691e-01 -3.15877378e-01 -7.37790704e-01
-9.99021113e-01 1.13828230e+00 9.10636365e-01 2.43144572e-01
6.98119327e-02 1.57215759e-01 3.39538157e-01 2.95487847e-02
2.23321468e-01 4.43437248e-01 -1.96710244e-01 4.07316864e-01
-4.43531156e-01 -3.39327962e-03 4.31219131e-01 1.21923912e+00
1.26002800e+00 1.52796302e-02 -3.39563102e-01 8.78980339e-01
5.06039619e-01 3.14346641e-01 6.97003961e-01 -5.34600377e-01
3.89955461e-01 8.58265400e-01 -2.14128271e-01 -1.07567632e+00
3.27009782e-02 -2.19848096e-01 -5.61111629e-01 1.80232376e-01
3.16685826e-01 1.57054320e-01 -1.51730311e+00 1.76560843e+00
3.81008238e-01 4.52978790e-01 3.60461116e-01 1.02725124e+00
1.04692352e+00 7.33988881e-01 3.22018921e-01 1.67477220e-01
1.41271484e+00 -9.55861807e-01 -5.29328585e-01 -7.25015640e-01
5.20741105e-01 -5.75371921e-01 1.33667183e+00 -2.34820798e-01
-3.12687457e-01 -9.44739878e-01 -1.19480085e+00 -2.02586755e-01
-7.71900654e-01 -1.09048195e-01 6.05233073e-01 3.74100357e-01
-6.21681452e-01 4.03121680e-01 -5.90900958e-01 -5.58789074e-01
5.80742180e-01 4.18854542e-02 -2.82874942e-01 -3.18907052e-01
-1.08479309e+00 4.51330751e-01 8.37311149e-01 -3.18145193e-02
-9.78279114e-01 -5.83234251e-01 -1.40014458e+00 3.52304205e-02
4.19958532e-01 -4.85997319e-01 6.89425290e-01 -1.01872110e+00
-9.44066048e-01 1.05005777e+00 -2.48595923e-01 -1.81240425e-01
2.19036434e-02 5.74131534e-02 -4.28049654e-01 9.09764990e-02
2.84401536e-01 1.06171739e+00 8.88213873e-01 -1.65343451e+00
-8.32431376e-01 -2.98135817e-01 6.98461905e-02 3.22989911e-01
-5.38683236e-01 -4.82589185e-01 -3.92189652e-01 -5.54839373e-01
3.41804951e-01 -8.82739842e-01 -1.24549866e-01 2.50805140e-01
-5.43829083e-01 -4.99100536e-01 9.35357273e-01 -1.62330940e-01
7.91752160e-01 -2.49107122e+00 1.84343401e-02 1.16144024e-01
1.44936621e-01 1.04719117e-01 -4.96269941e-01 9.14066359e-02
-7.21305013e-02 2.19653584e-02 -3.55336964e-01 -1.10360384e-01
2.08355989e-02 6.40337169e-01 -5.04506171e-01 2.85289615e-01
5.67383528e-01 9.24904346e-01 -1.23966193e+00 -4.86679196e-01
4.18057829e-01 3.26245248e-01 -4.27725345e-01 5.00339687e-01
-1.93658546e-01 2.00707287e-01 -5.04588604e-01 5.51187754e-01
5.06244302e-01 -9.65060741e-02 1.10336594e-01 -2.70244360e-01
2.00154573e-01 2.89397556e-02 -1.02632308e+00 1.56927633e+00
-4.36687827e-01 6.11836612e-01 -4.70366895e-01 -1.17446339e+00
1.10233438e+00 -1.37979254e-01 -1.23181731e-01 -6.80376351e-01
1.58479035e-01 5.18623255e-02 6.52899072e-02 -5.28016508e-01
2.71436036e-01 -3.39686275e-01 -2.15471193e-01 8.44928697e-02
3.87409806e-01 -2.64153391e-01 2.03625932e-01 2.11107373e-01
6.53563321e-01 6.62503317e-02 2.95377254e-01 -3.13454717e-01
4.96331245e-01 3.73724252e-02 7.09249139e-01 6.49129570e-01
-2.87575871e-01 6.96800947e-01 4.72093910e-01 -3.07464808e-01
-7.62543738e-01 -1.45240688e+00 -2.52548963e-01 1.22517991e+00
7.38640845e-01 -3.03000808e-01 -6.16269112e-01 -9.99892175e-01
4.52592820e-02 1.06373286e+00 -9.12189424e-01 -3.88365000e-01
-2.37563580e-01 -4.93311703e-01 -1.13836611e-02 8.01272273e-01
3.00613880e-01 -1.17205656e+00 -4.95436877e-01 -9.85509679e-02
2.16130748e-01 -1.17376947e+00 -4.25623208e-01 3.46891731e-01
-6.65079474e-01 -1.17400241e+00 -4.12061125e-01 -1.16109300e+00
1.03590381e+00 3.86478186e-01 9.52972412e-01 -1.59088499e-03
-5.10446250e-01 3.04014295e-01 -6.04027092e-01 -1.76210120e-01
-1.14925787e-01 -2.39185825e-01 -2.39504769e-01 3.52080435e-01
7.45486140e-01 -3.19806635e-01 -6.34153545e-01 2.91421592e-01
-1.01450324e+00 1.48511425e-01 4.16724563e-01 1.07734954e+00
8.26097965e-01 -8.51640105e-02 3.76154542e-01 -9.37741995e-01
1.51611060e-01 -5.38896859e-01 -4.04429048e-01 2.62215644e-01
-4.19971973e-01 3.83542359e-01 7.38223553e-01 -5.29242337e-01
-9.74058986e-01 2.95564383e-01 1.74108654e-01 -4.59907085e-01
-4.95423853e-01 5.48322722e-02 -4.18870389e-01 1.59355178e-01
5.89773118e-01 3.37061346e-01 -3.48836124e-01 -1.61993876e-01
8.24832499e-01 6.67887866e-01 5.82703173e-01 -5.38693309e-01
9.50899422e-01 6.21077240e-01 -1.41394109e-01 -8.01163554e-01
-1.11665344e+00 -7.52999842e-01 -6.65582359e-01 6.04563653e-02
1.25893164e+00 -7.89732635e-01 1.17989957e-01 1.00503422e-01
-9.89485562e-01 -9.03589427e-02 -4.56003636e-01 2.34071389e-01
-4.56097960e-01 3.40886921e-01 -1.98613405e-01 -4.69811052e-01
-7.54429474e-02 -1.13005424e+00 1.18454838e+00 3.79272580e-01
-2.41958112e-01 -1.04466665e+00 -1.96262911e-01 1.54630244e-01
-7.28229177e-04 4.21978295e-01 1.18553126e+00 -7.41152406e-01
-5.15393078e-01 -7.54619902e-03 -6.34814143e-01 5.32360911e-01
3.07597667e-01 -1.86689362e-01 -1.09444392e+00 -2.47414649e-01
-8.28332454e-02 -3.66582692e-01 1.12148869e+00 -1.92384329e-02
1.22052360e+00 -9.49978605e-02 -4.09673929e-01 6.93199217e-01
1.78489804e+00 2.16989741e-01 5.64225793e-01 1.01129018e-01
1.07492363e+00 9.20683444e-01 6.94854319e-01 1.36958525e-01
1.39537498e-01 4.03686881e-01 4.86316949e-01 -1.01933545e-02
-3.44147682e-01 -5.66171050e-01 1.39013410e-01 6.71033621e-01
3.27061594e-01 -2.51163810e-01 -7.21018493e-01 8.23980272e-01
-1.74015331e+00 -6.40285671e-01 -1.77856281e-01 2.14235592e+00
7.88091362e-01 1.98563367e-01 -2.75144637e-01 -3.47876325e-02
9.51745868e-01 3.65408897e-01 -5.37666202e-01 -4.94730175e-01
-7.06435964e-02 3.38920474e-01 3.31193388e-01 3.37494791e-01
-1.07332635e+00 1.16087186e+00 5.18783522e+00 9.10589337e-01
-1.13371241e+00 3.06609441e-02 5.38059056e-01 6.31996170e-02
-4.88632470e-01 2.42445946e-01 -9.72107768e-01 4.79013592e-01
3.06737691e-01 -2.16582566e-01 1.65318608e-01 1.06764019e+00
-2.40021393e-01 2.43902374e-02 -1.37940621e+00 9.45394278e-01
3.30348432e-01 -1.06775391e+00 3.85637969e-01 -2.90955938e-02
9.26904559e-01 -2.27622002e-01 6.48001060e-02 4.00051355e-01
1.56064287e-01 -1.05452216e+00 6.88140213e-01 2.85129428e-01
8.68907809e-01 -5.67663074e-01 6.62830234e-01 3.64833735e-02
-1.45389509e+00 -8.34195092e-02 -7.77933657e-01 2.90570147e-02
-4.67225462e-02 3.70475858e-01 -7.53692091e-01 3.67641687e-01
5.97590208e-01 1.02952814e+00 -7.43514955e-01 6.59317493e-01
-6.97928190e-01 5.02339661e-01 1.21864557e-01 -1.87509120e-01
5.76015592e-01 -9.64238793e-02 2.35714555e-01 1.08128858e+00
2.14852944e-01 -1.56361144e-02 2.11103067e-01 1.23180890e+00
-3.26035991e-02 4.41530049e-02 -6.87516391e-01 -7.03092664e-02
3.21873754e-01 1.29690003e+00 -8.23799074e-01 -4.53803927e-01
-5.52809238e-01 1.14596999e+00 3.41446549e-01 5.42844951e-01
-8.09913397e-01 -6.55471742e-01 7.88502038e-01 -2.90397722e-02
5.50579131e-01 -1.77855100e-02 -2.91469932e-01 -1.03948855e+00
-1.39158424e-02 -3.33416760e-01 4.38750267e-01 -6.88860953e-01
-1.71251762e+00 6.92804098e-01 3.30638438e-02 -1.29558158e+00
-1.19557187e-01 -8.84602129e-01 -8.29181552e-01 8.15900862e-01
-1.60026264e+00 -1.17826414e+00 -5.74123979e-01 5.28585374e-01
7.88803577e-01 -6.06673360e-02 7.44791210e-01 -2.67708957e-01
-4.42934722e-01 5.62146366e-01 -3.64414789e-02 2.60164827e-01
4.37303662e-01 -1.44834805e+00 3.36549461e-01 8.80039215e-01
4.45307076e-01 5.95854521e-01 6.24294281e-01 -4.83543307e-01
-9.76605237e-01 -1.48895872e+00 8.09247196e-01 -4.20338809e-01
7.23257065e-01 -6.91446245e-01 -1.10812747e+00 3.97506148e-01
1.81329161e-01 5.99344432e-01 6.89836264e-01 -6.55474961e-02
-7.42931306e-01 -1.44091710e-01 -9.10241842e-01 6.23067796e-01
1.30276597e+00 -6.80506587e-01 -1.05730069e+00 1.57555312e-01
8.91445637e-01 1.38023589e-03 -4.02054518e-01 3.92156363e-01
2.76779085e-01 -7.71216691e-01 8.84030998e-01 -6.66031241e-01
5.51028371e-01 -5.43526530e-01 -4.65772986e-01 -1.41335523e+00
-3.45652252e-01 7.28140399e-02 9.07957405e-02 1.36313093e+00
2.72885978e-01 -5.66058397e-01 6.49234295e-01 3.92179489e-01
-1.95721656e-01 -7.52673090e-01 -7.89700925e-01 -1.02944076e+00
6.97224913e-03 -3.19997698e-01 4.52784210e-01 9.03625250e-01
-3.05788785e-01 5.24784207e-01 5.44770993e-03 1.42686442e-01
6.83391511e-01 2.96316355e-01 4.87057984e-01 -9.89656150e-01
-2.93182641e-01 -3.01814079e-01 -9.59562838e-01 -9.21723604e-01
4.60832059e-01 -1.04792893e+00 4.20050383e-01 -1.54344177e+00
3.07276398e-01 -4.68140811e-01 -6.90860510e-01 6.33825541e-01
-5.04241943e-01 4.57093239e-01 2.19672561e-01 -4.99140508e-02
-6.77643657e-01 7.28649616e-01 1.22022557e+00 -3.20134759e-01
-1.68781672e-02 -5.04074991e-01 -8.45065355e-01 6.73695445e-01
6.87477052e-01 -5.26356280e-01 -8.06153417e-01 -3.87764692e-01
-1.51853949e-01 -6.86294377e-01 5.57748914e-01 -9.44175780e-01
-1.12816975e-01 -4.29810256e-01 3.49023610e-01 -3.14215064e-01
2.46387810e-01 -6.95993662e-01 -3.58892083e-01 2.62197793e-01
-3.97253364e-01 -5.47212303e-01 8.64344910e-02 8.64338994e-01
-3.58997405e-01 -3.72883350e-01 7.79712319e-01 -3.81912291e-02
-1.51187038e+00 1.02806248e-01 6.59016222e-02 3.94731283e-01
1.32803643e+00 -6.13345683e-01 -4.02445287e-01 6.89000040e-02
-6.19631529e-01 2.38209099e-01 4.51587856e-01 7.61102736e-01
8.97215009e-01 -1.41320252e+00 -3.89053226e-01 5.41669488e-01
8.22737217e-01 1.14724241e-01 2.09264770e-01 1.71468854e-01
-2.60857493e-01 2.22465530e-01 -2.90056914e-01 -7.89859176e-01
-1.08646369e+00 8.02028596e-01 1.03493847e-01 2.77390599e-01
-4.15551215e-01 1.15754402e+00 8.72773409e-01 -2.87626952e-01
1.17097035e-01 -4.40534413e-01 -2.33828872e-01 5.95479943e-02
4.07965362e-01 -6.70050532e-02 -3.88543338e-01 -7.50699341e-01
-3.87407988e-01 8.31932366e-01 -1.14000484e-01 1.95179179e-01
1.11059248e+00 -6.58148155e-02 3.44779119e-02 5.20222664e-01
1.63738060e+00 -2.39127979e-01 -1.47253573e+00 -2.11999580e-01
6.14511520e-02 -6.29921854e-01 -7.22095594e-02 -4.50863838e-01
-1.04005611e+00 1.09541118e+00 7.07659483e-01 1.29336491e-01
1.06313622e+00 4.50714260e-01 6.73950911e-01 1.28434852e-01
2.37159699e-01 -9.94880259e-01 3.59519601e-01 3.20379049e-01
7.40021586e-01 -1.34477711e+00 -2.55171180e-01 -8.11149418e-01
-8.41114163e-01 1.01677811e+00 9.28636491e-01 -3.26715410e-01
3.59666198e-01 -2.52727598e-01 1.40125165e-02 -2.23076165e-01
-4.49200571e-01 -6.67341888e-01 7.10044920e-01 8.68049145e-01
2.68752635e-01 1.69484675e-01 -1.87025875e-01 5.44220805e-01
-1.73799172e-02 -5.07746816e-01 3.02135974e-01 7.99015760e-01
-6.47835970e-01 -1.01454949e+00 -9.74194929e-02 4.15038168e-01
2.49347061e-01 -2.19182298e-01 -3.91557008e-01 7.25389242e-01
6.26227379e-01 1.04283559e+00 4.41284031e-01 -3.72568816e-01
2.56123453e-01 6.28283694e-02 3.01058292e-01 -1.20754802e+00
1.25871887e-02 -1.36425197e-01 -1.03746451e-01 -5.11174619e-01
-1.95252046e-01 -3.50352377e-01 -1.53255832e+00 2.97228307e-01
-2.67428845e-01 8.57868232e-03 4.21569347e-01 8.87702763e-01
2.30098054e-01 5.67100585e-01 7.12388098e-01 -5.06030023e-01
-4.53627110e-01 -7.25955427e-01 -6.23530090e-01 1.12979031e+00
1.39262974e-01 -9.49392498e-01 -6.95249021e-01 2.14515209e-01] | [9.913046836853027, 2.0159912109375] |
fee65664-500d-40d6-9767-0afd3944f884 | extractive-summarization-and-dialogue-act | null | null | https://aclanthology.org/W14-4318 | https://aclanthology.org/W14-4318.pdf | Extractive Summarization and Dialogue Act Modeling on Email Threads: An Integrated Probabilistic Approach | null | ['Giuseppe Carenini', 'Tatsuro Oya'] | 2014-06-01 | null | null | null | ws-2014-6 | ['meeting-summarization'] | ['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] | [-7.268677234649658, 3.6840670108795166] |
f0cb3a78-83a2-45e4-9938-7e7b0f5b507b | photogrammetric-point-cloud-segmentation-and | null | null | https://ascelibrary.org/doi/abs/10.1061/(ASCE)ME.1943-5479.0000737 | https://www.researchgate.net/profile/Meida-Chen/publication/339605064_Photogrammetric_Point_Cloud_Segmentation_and_Object_Information_Extraction_for_Creating_Virtual_Environments_and_Simulations/links/5ee1e4d1299bf1faac4aee4e/Photogrammetric-Point-Cloud-Segmentation-and-Object-Information-Extraction-for-Creating-Virtual-Environments-and-Simulations.pdf | Photogrammetric point cloud segmentation and object information extraction for creating virtual environments and simulations | Photogrammetric techniques have dramatically improved over the last few years, enabling the creation of visually compelling three-dimensional (3D) meshes using unmanned aerial vehicle imagery. These high-quality 3D meshes have attracted notice from both academicians and industry practitioners in developing virtual environments and simulations. However, photogrammetric generated point clouds and meshes do not allow both user-level and system-level interaction because they do not contain the semantic information to distinguish between objects. Thus, segmenting generated point clouds and meshes and extracting the associated object information is a necessary step. A framework for point cloud and mesh classification and segmentation is presented in this paper. The proposed framework was designed considering photogrammetric data-quality issues and provides a novel way of extracting object information, including (1) individual tree locations and related features and (2) building footprints. Experiments were conducted to rank different point descriptors and evaluate supervised machine-learning algorithms for segmenting photogrammetric generated point clouds. The proposed framework was validated using data collected at the University of Southern California (USC) and the Muscatatuck Urban Training Center (MUTC). | ['Lucio Soibelman', 'Ryan McAlinden', 'Andrew Feng', 'Meida Chen'] | 2020-03-01 | null | null | null | journal-of-management-in-engineering-2020-3 | ['point-cloud-segmentation'] | ['computer-vision'] | [ 2.62344927e-01 -3.83241475e-01 -3.42609957e-02 -3.18203270e-01
-4.71118957e-01 -5.61342061e-01 5.78359663e-01 7.64083326e-01
-1.70687571e-01 4.86494124e-01 -4.46385205e-01 -2.99049139e-01
-4.17452484e-01 -1.32240760e+00 -5.61506450e-01 -2.19504043e-01
-5.11228204e-01 6.71383142e-01 3.60363573e-01 -3.19744915e-01
3.30306739e-01 1.22398472e+00 -2.24990654e+00 -1.95231497e-01
1.00648522e+00 9.56499755e-01 5.81931531e-01 5.77018678e-01
-2.37049147e-01 -1.59158811e-01 -2.29058102e-01 5.51638901e-02
5.66113472e-01 1.05219863e-01 -6.36570513e-01 5.06321013e-01
8.72544289e-01 -9.02469829e-02 3.19176912e-01 1.17939258e+00
1.39428616e-01 1.35435641e-01 6.72925174e-01 -1.25629747e+00
-1.43970385e-01 -1.88342121e-03 -6.91750348e-01 -3.11946899e-01
9.94304791e-02 -1.66695546e-02 1.05139816e+00 -8.77700925e-01
5.44572473e-01 1.07177401e+00 6.02171242e-01 -1.10425144e-01
-1.17238665e+00 -5.12007177e-01 2.08959337e-02 1.44653350e-01
-1.67111087e+00 9.72377583e-02 9.15972769e-01 -8.25959504e-01
5.34064174e-01 7.46003151e-01 1.17839324e+00 1.85487434e-01
-2.12864056e-01 3.37394506e-01 1.09169364e+00 -6.00644469e-01
3.48008603e-01 6.13421760e-02 6.37079701e-02 6.90190911e-01
4.42399263e-01 1.45953074e-01 1.63030818e-01 -3.46586883e-01
1.15966201e+00 1.53108835e-01 -1.67138502e-01 -8.29234660e-01
-8.68028402e-01 6.86721265e-01 7.90088415e-01 3.28450143e-01
-7.89394557e-01 1.45333391e-02 1.44808024e-01 -9.32144746e-02
5.23077905e-01 3.02512616e-01 -1.47723958e-01 9.92906988e-02
-1.19084895e+00 2.25982726e-01 2.97102034e-01 1.06454706e+00
1.01572144e+00 1.60075724e-01 6.28223598e-01 7.77586281e-01
4.47789371e-01 7.79372215e-01 -1.88518196e-01 -1.00110495e+00
3.84232908e-01 1.17058671e+00 1.99526295e-01 -1.63611472e+00
-3.99619371e-01 -1.38684332e-01 -8.00485373e-01 6.24621868e-01
-2.89262980e-02 1.98743403e-01 -9.22676563e-01 8.51485968e-01
6.00085437e-01 -4.49138656e-02 -3.62121850e-01 9.55821693e-01
7.57948816e-01 6.52947724e-01 2.35406056e-01 1.66951299e-01
1.25322878e+00 -1.31933004e-01 -6.61179423e-02 -1.92728385e-01
4.88120079e-01 -6.06082439e-01 9.01310205e-01 1.99048653e-01
-8.09927821e-01 -6.87050879e-01 -1.18832743e+00 4.12790328e-01
-3.78297776e-01 3.04868937e-01 5.00800967e-01 5.30172467e-01
-7.42972016e-01 7.99399793e-01 -6.70767546e-01 -5.75675547e-01
4.01415795e-01 2.23994479e-01 -3.51811290e-01 1.51780412e-01
-8.07293475e-01 7.42398620e-01 3.05410028e-01 1.66587502e-01
-4.76448357e-01 -4.66907412e-01 -8.99892926e-01 -9.06127244e-02
-7.95135200e-02 -4.40301955e-01 6.13588035e-01 -8.93285871e-01
-8.72377276e-01 1.10716760e+00 3.16083431e-01 -2.19063833e-01
4.38184857e-01 9.62223709e-02 -4.31543767e-01 1.74690455e-01
2.24269390e-01 5.93807697e-01 6.72423482e-01 -1.79310441e+00
-1.02372468e+00 -8.37517738e-01 8.08727592e-02 2.97802836e-01
8.62676054e-02 -3.22901249e-01 -7.88778067e-02 -4.82994169e-01
6.30967796e-01 -9.77325201e-01 -2.18177557e-01 2.15264991e-01
-1.24433234e-01 1.52895749e-01 9.14996862e-01 -6.17625475e-01
8.80338728e-01 -1.95647407e+00 -2.26303488e-02 7.17966318e-01
-5.35858236e-02 3.87032211e-01 9.97272879e-02 4.83350009e-01
-1.72540605e-01 2.08581716e-01 -4.26905423e-01 4.59392294e-02
-2.85367578e-01 2.90792286e-01 -5.97472079e-02 4.77840155e-01
1.97217520e-02 1.39533997e-01 -9.91990805e-01 -6.76365316e-01
1.01873696e+00 4.63310122e-01 -5.82678244e-02 -3.05070430e-01
-1.64255545e-01 3.46934766e-01 -6.23142660e-01 9.24634039e-01
8.46690357e-01 2.39821225e-01 1.58711612e-01 -2.13684693e-01
-3.27870697e-01 -3.72546107e-01 -1.44899321e+00 1.33213687e+00
-6.59139454e-01 5.98518729e-01 2.73515075e-01 -7.00298965e-01
1.40737545e+00 2.52158076e-01 7.52372980e-01 -3.84215832e-01
1.59785479e-01 2.93970674e-01 -4.54353094e-01 -2.14037195e-01
8.18560541e-01 -1.48616612e-01 5.54291531e-02 -1.20609462e-01
-4.01661903e-01 -8.33254695e-01 -1.52073819e-02 -2.10455358e-01
4.10097659e-01 1.43969208e-01 3.79416496e-01 -4.08479720e-01
5.75593352e-01 7.60165930e-01 2.67611533e-01 4.82657459e-03
1.24581702e-01 4.01344031e-01 -2.38798708e-01 -5.65060139e-01
-1.21436286e+00 -1.07082784e+00 -4.10416722e-01 3.71082276e-01
4.97922868e-01 -2.44807169e-01 -5.74674726e-01 -2.07319781e-01
3.10219616e-01 7.22121656e-01 -2.54946947e-01 4.05316681e-01
-3.19147527e-01 -2.62805194e-01 6.48535416e-02 2.76356161e-01
5.89556932e-01 -6.54802322e-01 -1.00165021e+00 1.73618525e-01
-6.32013008e-02 -9.04323936e-01 2.52496630e-01 -4.11521256e-01
-1.41370690e+00 -1.35011387e+00 -4.73327160e-01 -7.62079537e-01
9.43582177e-01 6.08571351e-01 1.01649392e+00 2.80127823e-01
-4.64093924e-01 5.17389476e-01 -5.40358365e-01 -5.76322198e-01
-3.24630082e-01 -9.69142914e-02 -8.89084265e-02 -5.70673980e-02
1.41899213e-01 -7.57096291e-01 -3.62987429e-01 6.02513015e-01
-6.65576041e-01 1.42503470e-01 4.90528673e-01 2.82543272e-01
9.50249970e-01 3.88307989e-01 1.85285985e-01 -5.44342220e-01
1.99186012e-01 -1.93400860e-01 -1.07811439e+00 5.13619781e-02
-2.67538160e-01 -4.53318983e-01 3.23938847e-01 1.74547449e-01
-6.22121811e-01 4.65704083e-01 -1.40813533e-02 -3.41356426e-01
-6.95145726e-01 5.40094495e-01 -2.70285219e-01 -4.07569408e-01
8.08244228e-01 -5.35673797e-02 -1.04094997e-01 -5.82172036e-01
3.36626917e-01 9.07323480e-01 5.33418715e-01 -5.29917538e-01
1.12781572e+00 6.82218552e-01 4.05908346e-01 -1.48300493e+00
-1.87435091e-01 -7.46981502e-01 -1.15330040e+00 -6.75489724e-01
7.02633440e-01 -8.43117595e-01 -3.62290114e-01 2.51270592e-01
-1.04668355e+00 1.49687931e-01 -3.77828509e-01 4.67905581e-01
-5.90914905e-01 4.44348216e-01 1.26087472e-01 -9.41227257e-01
-1.85088769e-01 -1.16023326e+00 1.01659369e+00 5.95075116e-02
-1.81633621e-01 -7.24551380e-01 -8.34724084e-02 3.63964796e-01
-2.32575871e-02 8.32524478e-01 9.70955610e-01 8.44970867e-02
-6.95643961e-01 -5.97431481e-01 -2.14775205e-01 2.90644020e-01
1.69407323e-01 3.53184104e-01 -6.12023652e-01 -4.50167283e-02
-2.87085027e-01 2.73058891e-01 2.22084641e-01 3.92241567e-01
8.24290156e-01 3.77025530e-02 -4.37243074e-01 4.09540206e-01
1.73299325e+00 1.83148906e-01 5.41658342e-01 4.87922668e-01
7.46494710e-01 6.98612273e-01 8.93838704e-01 4.81240779e-01
2.57906824e-01 7.88109660e-01 9.53309178e-01 -3.15581560e-01
1.31261647e-01 -3.86524111e-01 -2.52911508e-01 4.31205124e-01
-7.57407665e-01 3.71477991e-01 -1.30331397e+00 6.26792550e-01
-1.61224020e+00 -9.18987095e-01 -9.17662561e-01 2.51748490e+00
2.77236283e-01 -1.96906611e-01 5.61257005e-02 5.54253161e-01
9.87049043e-01 -6.51305914e-02 -1.78453103e-01 -1.15972236e-01
1.19829789e-01 2.62412995e-01 8.57198656e-01 4.35320526e-01
-1.23258996e+00 8.48443091e-01 4.90618753e+00 6.53072059e-01
-1.03324533e+00 -2.30834901e-01 -1.27874941e-01 4.55069512e-01
-1.85106590e-01 1.73011959e-01 -3.97639871e-01 1.61235869e-01
5.28147995e-01 -2.73625664e-02 6.61252663e-02 1.19103849e+00
4.21362251e-01 -2.93893963e-01 -5.84707379e-01 1.14787066e+00
-2.32854888e-01 -1.31800568e+00 2.24508122e-01 3.69813144e-01
6.82830393e-01 -1.13231458e-01 -2.94282138e-01 -2.20386833e-01
2.93795168e-01 -8.27386320e-01 7.75465548e-01 4.53015417e-01
8.68175089e-01 -8.35445046e-01 5.20311117e-01 4.38643008e-01
-1.62777340e+00 9.55673307e-02 -3.32263261e-01 -1.53993860e-01
6.70980513e-02 4.59077597e-01 -9.16786551e-01 8.94842505e-01
6.56730294e-01 5.91425180e-01 -6.73011303e-01 1.44440043e+00
-7.50912204e-02 3.74603152e-01 -4.53420877e-01 1.59787446e-01
3.38235974e-01 -6.54102623e-01 6.67527318e-01 7.76071846e-01
5.05405188e-01 1.54392511e-01 2.71593153e-01 7.08478451e-01
3.55346560e-01 4.73359078e-01 -7.74131000e-01 1.15017615e-01
7.36258030e-01 1.32521915e+00 -1.04948437e+00 -1.60037965e-01
-1.01163514e-01 4.76090938e-01 -2.06243664e-01 -4.77468744e-02
-5.61357141e-01 -3.54049891e-01 5.36056757e-01 7.27733195e-01
-8.26235712e-02 -7.71471858e-01 -5.28561771e-01 -5.80572128e-01
-1.57753527e-01 -4.11555529e-01 1.03649392e-03 -8.11104178e-01
-9.18264210e-01 4.43994462e-01 3.37795615e-01 -1.67852330e+00
4.45813313e-02 -4.27323580e-01 -3.71290088e-01 9.43122625e-01
-1.17633152e+00 -1.42886806e+00 -8.65914583e-01 2.31080294e-01
5.42377889e-01 1.01829425e-01 1.00819135e+00 1.49498507e-01
-1.73346326e-01 -3.18567485e-01 1.87414780e-01 8.51411223e-02
-9.84864868e-03 -1.17767262e+00 3.00488144e-01 7.21441627e-01
1.44195527e-01 1.91526085e-01 6.58290803e-01 -9.10551369e-01
-1.23818493e+00 -1.27687669e+00 5.60230017e-01 -1.22870095e-01
2.57164836e-01 7.35183880e-02 -7.51279593e-01 4.73568738e-01
-4.20253843e-01 -2.79376417e-01 3.75885636e-01 -8.61586034e-02
1.58642635e-01 -2.99638379e-02 -1.36533988e+00 4.54036802e-01
9.66633797e-01 -1.84524804e-01 -4.85011697e-01 2.72877574e-01
-6.44043535e-02 -2.44941905e-01 -1.04021490e+00 6.22199833e-01
3.85787696e-01 -1.00373638e+00 1.18597019e+00 -5.25302738e-02
1.17531553e-01 -4.69539732e-01 -4.29997385e-01 -1.36095691e+00
-2.57563114e-01 1.31839231e-01 6.30561590e-01 1.03159070e+00
9.34010968e-02 -2.69697815e-01 8.17032278e-01 5.51979184e-01
-2.53107488e-01 -3.31766576e-01 -8.22470963e-01 -9.03422236e-01
-2.18968526e-01 -6.89934433e-01 7.51947224e-01 8.61664236e-01
-3.60255420e-01 7.83448890e-02 8.19146857e-02 6.04112029e-01
8.55816543e-01 4.01628584e-01 1.02580500e+00 -2.13667560e+00
4.52900499e-01 -4.33569461e-01 -1.09481931e+00 -3.91566604e-01
-1.54828858e-02 -9.19586301e-01 -3.09860229e-01 -1.96953332e+00
-4.74697739e-01 -9.97389495e-01 5.58325469e-01 1.64546192e-01
3.10814768e-01 1.87940449e-01 3.02954257e-01 5.06484509e-01
2.48436168e-01 5.30921340e-01 9.55135524e-01 -1.80849627e-01
-2.22707704e-01 2.03799739e-01 -2.14056056e-02 9.79047418e-01
8.07340026e-01 -1.57926515e-01 -3.56215447e-01 -3.92615557e-01
-6.02556951e-02 -6.01730794e-02 8.08045030e-01 -1.34610212e+00
-7.06492811e-02 -3.09714854e-01 3.71023089e-01 -1.04821754e+00
5.92704296e-01 -1.58208752e+00 6.90000772e-01 3.63927335e-01
3.21864486e-01 -1.68628678e-01 2.55878150e-01 4.20036256e-01
-1.16386130e-01 -3.78156990e-01 9.44700301e-01 -3.98030877e-01
-1.05830717e+00 3.57360303e-01 -3.78297083e-02 -4.80249465e-01
1.45035803e+00 -6.48491919e-01 1.60454422e-01 -1.60082564e-01
-7.34241784e-01 2.35966146e-01 1.10401678e+00 1.82151064e-01
8.12881052e-01 -1.30108976e+00 -6.04300737e-01 3.38189751e-01
3.04001689e-01 3.21892828e-01 3.72358382e-01 2.55285919e-01
-1.20550346e+00 2.32685089e-01 -5.62939167e-01 -8.25535178e-01
-1.53659058e+00 2.02251896e-01 1.58453271e-01 3.72707069e-01
-7.00633824e-01 4.10144091e-01 -3.37625921e-01 -5.76643348e-01
-7.35701695e-02 -4.87038195e-01 -4.00379747e-01 2.57606536e-01
5.11479080e-02 5.96004546e-01 2.94189453e-01 -1.30333734e+00
-2.77832299e-01 1.08980370e+00 7.06996977e-01 3.52860540e-02
1.32178342e+00 -1.28448352e-01 3.96917313e-02 4.10061061e-01
7.51931489e-01 -2.58032918e-01 -7.68466830e-01 -1.59774750e-01
2.16379613e-01 -9.94390845e-01 4.46125358e-01 -4.16866064e-01
-9.32373047e-01 1.01900089e+00 7.57275641e-01 4.21476632e-01
9.85610962e-01 -2.19922408e-01 5.85292280e-01 1.03454329e-01
1.08745110e+00 -1.10898447e+00 -6.95804358e-01 6.86203763e-02
1.08782768e+00 -1.02651548e+00 2.58329719e-01 -7.52629101e-01
-5.16615212e-01 1.23975205e+00 2.89344966e-01 -2.27754116e-01
7.29046941e-01 -1.63825586e-01 -3.95449549e-02 -5.17803550e-01
1.06373630e-01 -6.05287790e-01 3.64526242e-01 9.91955817e-01
1.21695876e-01 5.82747340e-01 -2.10778162e-01 -1.00571543e-01
-3.79970968e-01 -9.15017426e-02 3.52234751e-01 1.11020124e+00
-9.44207966e-01 -9.48245764e-01 -9.61623728e-01 6.24478042e-01
3.14494103e-01 3.21534932e-01 -3.75581622e-01 1.04386175e+00
1.19338378e-01 8.16050649e-01 1.93050221e-01 -4.59537119e-01
7.58164227e-01 -3.26161802e-01 3.60351294e-01 -7.45598674e-01
-2.21857950e-01 -1.10523663e-01 8.75101686e-02 -1.71515673e-01
-4.90063965e-01 -6.96286559e-01 -1.22770739e+00 -3.55565220e-01
-4.14831489e-01 3.05956423e-01 1.37334824e+00 4.57609028e-01
3.39095354e-01 1.58130705e-01 7.50769734e-01 -1.50410151e+00
-1.14530541e-01 -6.81057394e-01 -9.39691961e-01 4.49219733e-01
-2.83309221e-01 -9.48309720e-01 -6.88885078e-02 9.80323330e-02] | [8.404561042785645, -2.6539113521575928] |
4e4ab0fc-ec76-442b-9a4a-9e5b6ae2da64 | lightcts-a-lightweight-framework-for | 2302.11974 | null | https://arxiv.org/abs/2302.11974v2 | https://arxiv.org/pdf/2302.11974v2.pdf | LightCTS: A Lightweight Framework for Correlated Time Series Forecasting | Correlated time series (CTS) forecasting plays an essential role in many practical applications, such as traffic management and server load control. Many deep learning models have been proposed to improve the accuracy of CTS forecasting. However, while models have become increasingly complex and computationally intensive, they struggle to improve accuracy. Pursuing a different direction, this study aims instead to enable much more efficient, lightweight models that preserve accuracy while being able to be deployed on resource-constrained devices. To achieve this goal, we characterize popular CTS forecasting models and yield two observations that indicate directions for lightweight CTS forecasting. On this basis, we propose the LightCTS framework that adopts plain stacking of temporal and spatial operators instead of alternate stacking that is much more computationally expensive. Moreover, LightCTS features light temporal and spatial operator modules, called L-TCN and GL-Former, that offer improved computational efficiency without compromising their feature extraction capabilities. LightCTS also encompasses a last-shot compression scheme to reduce redundant temporal features and speed up subsequent computations. Experiments with single-step and multi-step forecasting benchmark datasets show that LightCTS is capable of nearly state-of-the-art accuracy at much reduced computational and storage overheads. | ['Yan Zhao', 'Hua Lu', 'Christian S. Jensen', 'Huan Li', 'Dalin Zhang', 'Zhichen Lai'] | 2023-02-23 | null | null | null | null | ['correlated-time-series-forecasting', 'univariate-time-series-forecasting'] | ['time-series', 'time-series'] | [ 1.21790338e-02 -7.26858675e-01 -3.14633995e-01 -5.65471888e-01
-5.67126274e-01 -1.53337985e-01 6.74556613e-01 -2.06505239e-01
-1.10623598e-01 4.47445452e-01 -7.05294609e-02 -7.52320111e-01
-2.45918453e-01 -6.69099152e-01 -4.09308374e-01 -9.14075971e-01
-2.27162212e-01 1.85298219e-01 3.67169261e-01 -1.77311450e-01
2.95253247e-01 6.80784941e-01 -1.82695007e+00 6.22907341e-01
6.79616272e-01 1.77656829e+00 1.18046351e-01 4.72842783e-01
-3.18836600e-01 1.14578879e+00 -5.17700076e-01 -3.90974820e-01
2.20430955e-01 7.05620721e-02 -3.41091156e-01 -2.02621207e-01
1.24290898e-01 -4.87911552e-01 -4.53999013e-01 2.38060012e-01
5.36577225e-01 9.78427753e-02 2.12390035e-01 -1.64847505e+00
-7.82931373e-02 4.65803266e-01 -5.38828969e-01 4.96557593e-01
-8.88053179e-02 1.72301456e-01 7.13814855e-01 -7.36649871e-01
1.42075449e-01 8.18714678e-01 9.26073372e-01 5.25679206e-04
-8.60796511e-01 -8.93253565e-01 2.67767280e-01 7.84110606e-01
-1.22181332e+00 -6.08155668e-01 8.89536798e-01 9.72510949e-02
1.28220546e+00 5.23482323e-01 5.42419791e-01 9.15814042e-01
5.57361245e-01 7.92620420e-01 8.77451301e-01 -2.31529966e-01
3.86445433e-01 -3.32184404e-01 7.16273859e-02 3.95905763e-01
-6.00937791e-02 -4.95385639e-02 -8.89710844e-01 -3.38229761e-02
3.12638372e-01 4.64501768e-01 8.07742551e-02 -1.61212366e-02
-1.08498383e+00 6.65969551e-01 1.24070019e-01 4.89792675e-01
-7.07128465e-01 2.92707950e-01 7.37763464e-01 4.31520194e-01
8.13654661e-01 -1.37094200e-01 -7.76366174e-01 -7.71395743e-01
-1.24910951e+00 1.85789078e-01 6.19020343e-01 8.87904406e-01
3.66466463e-01 4.71474767e-01 -2.90576041e-01 5.86568713e-01
-3.52985829e-01 5.86779952e-01 5.06905735e-01 -9.30789769e-01
5.66935599e-01 4.33019996e-01 -1.96519718e-01 -1.12856877e+00
-6.45154417e-01 -7.44437575e-01 -1.09885740e+00 -1.79318294e-01
-1.51431635e-01 2.21616462e-01 -6.58536851e-01 1.34828711e+00
2.44352981e-01 7.88011491e-01 -2.84836143e-01 6.88719213e-01
4.42188203e-01 8.05583835e-01 4.76605557e-02 -5.07513225e-01
1.12860370e+00 -9.40373838e-01 -8.15456569e-01 3.59009773e-01
9.61294413e-01 -7.85397947e-01 1.01642203e+00 5.82614899e-01
-1.02739406e+00 -6.79445028e-01 -8.99808109e-01 -1.66848198e-01
-4.56224412e-01 2.10863173e-01 1.11034906e+00 6.46461546e-01
-1.00238490e+00 7.43601799e-01 -1.19794321e+00 -1.40135530e-02
3.98985475e-01 3.52404535e-01 1.82802066e-01 1.40653595e-01
-9.47475255e-01 5.83770633e-01 -1.94587618e-01 1.14720255e-01
-3.08473170e-01 -1.03471589e+00 -3.39540154e-01 2.63011843e-01
4.34632123e-01 -4.07680571e-01 1.47392845e+00 -5.15612483e-01
-1.59994912e+00 8.41544792e-02 -7.04667628e-01 -8.01646054e-01
3.50656778e-01 -2.73421168e-01 -9.80206192e-01 5.86847477e-02
-9.36270058e-02 5.99671081e-02 9.65285659e-01 -5.47582865e-01
-8.77569973e-01 -2.62859076e-01 -2.64208823e-01 -1.60574526e-01
-8.18554282e-01 9.13228020e-02 -6.03906512e-01 -7.34795868e-01
7.96064660e-02 -9.35232282e-01 -8.24129581e-02 -1.38038188e-01
-1.34385169e-01 -4.15122867e-01 1.36939240e+00 -5.09540439e-01
1.78693640e+00 -1.89027429e+00 -4.64903504e-01 1.88835308e-01
3.64861429e-01 5.12704730e-01 3.47972065e-02 6.61596060e-01
2.29630973e-02 -2.00530276e-01 2.85575032e-01 -6.42571628e-01
3.34214792e-02 2.70485371e-01 -9.33688879e-01 3.55237871e-01
6.59341216e-02 9.59532142e-01 -3.81439507e-01 -3.51699919e-01
4.86357242e-01 5.66183746e-01 -4.99131769e-01 -8.95467997e-02
-6.05544262e-02 3.71722549e-01 -3.94056588e-01 7.82690883e-01
6.57288492e-01 -4.80792463e-01 2.19719023e-01 -5.10556996e-01
-4.73175049e-01 5.62532425e-01 -8.47036421e-01 1.19593883e+00
-8.22918057e-01 8.17317307e-01 -4.16584641e-01 -1.29444480e+00
7.96006441e-01 1.96808204e-01 1.18813837e+00 -1.40744007e+00
1.59985006e-01 2.40330279e-01 -3.11796129e-01 -3.69973600e-01
8.80744815e-01 3.31345618e-01 1.14459544e-01 5.35391152e-01
-3.96254659e-01 3.28057945e-01 1.58226520e-01 9.92890671e-02
1.17039084e+00 -4.42931168e-02 -1.28762960e-01 6.47872463e-02
3.53354633e-01 -1.73734248e-01 6.56899035e-01 4.87809867e-01
6.43934123e-03 1.59558311e-01 2.18816623e-01 -9.90554035e-01
-9.07447875e-01 -8.10670197e-01 2.28953920e-02 1.42072225e+00
-1.22437648e-01 -6.81900859e-01 -2.86953270e-01 -2.58827269e-01
-1.81437321e-02 7.15425909e-01 -3.13315392e-01 -8.94607976e-02
-8.96445274e-01 -5.80147624e-01 5.05652308e-01 7.90063620e-01
5.74158847e-01 -5.63775599e-01 -9.88296390e-01 4.21562552e-01
-2.01172724e-01 -1.35502553e+00 -1.58577472e-01 2.38416806e-01
-1.00675225e+00 -6.12044454e-01 -3.90430182e-01 -3.46423149e-01
9.48711857e-02 7.90678561e-01 1.19026732e+00 1.73996612e-01
-6.12867251e-02 -7.45301833e-03 -4.93809700e-01 -5.94189048e-01
2.34650955e-01 2.93176144e-01 1.97047040e-01 6.54768795e-02
4.92862374e-01 -9.98413801e-01 -6.76032960e-01 4.18867022e-01
-9.02149200e-01 1.81006849e-01 5.26228547e-01 5.32188773e-01
5.70854187e-01 2.20987290e-01 4.66776848e-01 -5.54751635e-01
5.05790114e-01 -4.67078298e-01 -5.15014112e-01 3.40897292e-01
-8.94492149e-01 -1.20312288e-01 1.09037101e+00 -5.27621865e-01
-7.98490107e-01 -1.86951771e-01 -2.52264053e-01 -7.12346435e-01
2.51965016e-01 5.26633263e-01 4.45579380e-01 3.35276611e-02
2.26752404e-02 5.12242854e-01 -2.02500641e-01 -6.12436533e-01
2.14661509e-01 7.70131052e-01 4.56664979e-01 -4.90900815e-01
5.82835615e-01 7.48773932e-01 2.80753046e-01 -8.16561341e-01
-6.62331879e-01 -3.11468244e-01 -4.89950359e-01 -3.71501505e-01
2.31374785e-01 -8.45546186e-01 -1.21272027e+00 3.80632192e-01
-9.66143012e-01 -1.30074248e-01 -8.17266479e-02 3.49563986e-01
-4.23431128e-01 2.52514929e-01 -6.09430671e-01 -8.33319843e-01
-5.86339414e-01 -1.03580737e+00 1.31940567e+00 -1.86620042e-01
-2.61560511e-02 -7.49756277e-01 -3.47307622e-01 1.00097798e-01
1.00788689e+00 -3.36387381e-02 9.79365051e-01 -5.67909956e-01
-5.61605811e-01 -2.46999085e-01 -1.54599681e-01 1.06582038e-01
-5.92076965e-02 4.02306719e-03 -9.56872046e-01 -2.83643037e-01
-3.80609259e-02 -1.32632768e-02 6.33917212e-01 4.00522172e-01
1.79738235e+00 -2.65974820e-01 -3.70660365e-01 8.00698340e-01
1.04319453e+00 5.01203179e-01 6.10107303e-01 2.33948424e-01
4.77631986e-01 2.87998050e-01 6.10120118e-01 8.97775531e-01
5.81723392e-01 9.99606848e-01 1.21325523e-01 -9.46188271e-02
-2.10107360e-02 4.93217483e-02 3.44191909e-01 1.54710972e+00
-2.50887871e-01 -4.23090756e-01 -8.60939205e-01 1.93562329e-01
-1.82178140e+00 -1.04424334e+00 -3.36227357e-01 1.93886721e+00
3.84695917e-01 2.62606740e-01 2.06823125e-01 6.98355854e-01
2.31170997e-01 2.87725955e-01 -6.57319784e-01 -4.45674360e-01
-5.09240218e-02 4.67135638e-01 5.18917739e-01 -2.41135091e-01
-7.76763797e-01 7.22083688e-01 6.10995531e+00 1.10991931e+00
-1.83549356e+00 1.41215295e-01 7.97605693e-01 -4.17167127e-01
-7.09618405e-02 -3.05024773e-01 -7.87144244e-01 8.19331348e-01
1.51162601e+00 -6.15509748e-02 3.85118216e-01 9.67537284e-01
6.01245522e-01 1.81720674e-01 -9.06307578e-01 1.25384068e+00
-3.27848881e-01 -1.75685179e+00 5.99764893e-03 1.18030451e-01
4.92938906e-01 6.70685023e-02 3.59506071e-01 5.71493685e-01
-2.25847676e-01 -6.15751922e-01 8.76031220e-01 5.99813700e-01
8.27954888e-01 -8.53833079e-01 8.57923269e-01 2.06920877e-01
-1.76726198e+00 -3.38779002e-01 -2.71116525e-01 -4.38677371e-01
4.04271126e-01 9.68795717e-01 -5.42973578e-01 7.02862918e-01
8.78719330e-01 7.60783017e-01 -5.22042394e-01 6.99838340e-01
5.45686960e-01 9.45454240e-01 -4.46325302e-01 -5.22207171e-02
2.48828784e-01 9.38286260e-02 7.61044696e-02 1.00446868e+00
7.48688877e-01 8.96092728e-02 2.31035367e-01 2.81607062e-01
2.02169076e-01 -1.27976865e-01 -3.10500234e-01 5.67643233e-02
8.27759087e-01 9.92300987e-01 -7.21541643e-01 -6.35373771e-01
-5.85493326e-01 7.13275135e-01 9.63461399e-02 2.51153350e-01
-1.39035106e+00 -1.05814725e-01 7.81900525e-01 1.01034150e-01
5.03577113e-01 -5.41858375e-01 -6.16474032e-01 -1.03728235e+00
2.61542201e-01 -7.01967359e-01 2.22345874e-01 -6.07615471e-01
-1.00235534e+00 7.21371472e-01 -1.32726714e-01 -1.74189508e+00
-4.11480218e-01 -3.23100924e-01 -5.07446706e-01 5.53590417e-01
-1.67448890e+00 -1.24181688e+00 -3.73357594e-01 8.48474801e-01
8.01854730e-01 -2.09331214e-01 5.44522107e-01 7.16908813e-01
-7.35216081e-01 7.03441799e-01 3.96203458e-01 -5.29048204e-01
3.65990698e-01 -5.31365633e-01 6.74872994e-01 7.88735986e-01
1.92211438e-02 5.74353278e-01 5.65782905e-01 -3.64693701e-01
-1.84782231e+00 -1.23628771e+00 9.66575980e-01 1.08928330e-01
6.04080975e-01 -1.97499022e-01 -7.02211738e-01 2.93384880e-01
-1.37391254e-01 1.81916773e-01 5.37312210e-01 5.91650717e-02
-2.80956954e-01 -7.22608387e-01 -7.52680600e-01 5.48587143e-01
1.05280197e+00 -6.51944816e-01 1.29996747e-01 2.96910346e-01
7.45866418e-01 -2.61438578e-01 -8.55673075e-01 4.54549909e-01
7.78002918e-01 -1.18873143e+00 9.58200753e-01 -3.47027600e-01
2.66248256e-01 -1.36376351e-01 -4.19076592e-01 -7.07964480e-01
-3.26353759e-01 -7.17357278e-01 -7.73306251e-01 1.10260797e+00
-3.83234732e-02 -6.68080747e-01 8.62266302e-01 6.30410731e-01
-4.76439208e-01 -1.20619345e+00 -1.18214118e+00 -1.09279275e+00
-4.02495652e-01 -1.20830262e+00 1.06829309e+00 8.68386865e-01
-3.30666602e-01 -9.05042365e-02 -6.20681882e-01 5.05070053e-02
3.81503612e-01 4.16710347e-01 8.47598195e-01 -1.21972227e+00
5.62785044e-02 -3.70482355e-01 -3.05758804e-01 -1.36318696e+00
-1.79498810e-02 -4.87966478e-01 -4.54359591e-01 -1.02835441e+00
-2.79689997e-01 -7.16160774e-01 -5.09022236e-01 5.57768106e-01
3.28170717e-01 3.45599025e-01 2.16605753e-01 5.40885270e-01
-8.64147961e-01 7.00765729e-01 7.51299798e-01 1.39384225e-01
-9.19422284e-02 1.00309469e-01 -2.22620249e-01 4.89092559e-01
9.48989093e-01 -2.55000204e-01 -5.91456711e-01 -5.69484532e-01
-1.01233386e-01 1.86188564e-01 2.94323057e-01 -1.33515489e+00
5.03465712e-01 -1.36216551e-01 2.79329032e-01 -1.19315934e+00
5.80648661e-01 -9.33495462e-01 3.03797662e-01 4.47098792e-01
-8.38708878e-02 6.35127664e-01 2.32208595e-01 3.27848256e-01
-2.27721855e-01 5.32559037e-01 3.77962530e-01 4.38061982e-01
-9.16604340e-01 3.60140711e-01 -3.69764328e-01 -5.83215535e-01
1.00703168e+00 -3.11487556e-01 -3.04210842e-01 -4.73829091e-01
-1.46196052e-01 -9.50165093e-02 -3.68941836e-02 4.25708205e-01
5.21420121e-01 -1.30914617e+00 -2.10124776e-01 2.71878988e-01
-1.04716569e-01 -4.92792636e-01 4.94372517e-01 1.28399527e+00
-3.81654263e-01 1.04169714e+00 1.12851476e-02 -8.48931134e-01
-1.27954149e+00 5.87052107e-01 -1.72240719e-01 -5.78871787e-01
-7.40005195e-01 4.70882297e-01 -2.70535022e-01 -1.70438122e-02
4.52776849e-01 -6.98104322e-01 1.47611082e-01 -4.36456427e-02
7.51532257e-01 8.64495873e-01 7.24437594e-01 -4.62596506e-01
-3.32790732e-01 4.68337566e-01 -5.35450801e-02 2.94892937e-01
1.63943589e+00 -4.02013868e-01 9.38982517e-02 3.38197589e-01
1.32336652e+00 -2.11339086e-01 -1.09266794e+00 -3.31843406e-01
1.53314725e-01 -4.84460950e-01 3.79556030e-01 -5.97837985e-01
-1.47445989e+00 9.00380731e-01 7.40252793e-01 4.59809989e-01
1.71143901e+00 -5.02760053e-01 1.60744095e+00 3.85676861e-01
7.27339685e-01 -1.08027303e+00 -1.19484454e-01 4.81432348e-01
3.97620946e-01 -9.76442695e-01 7.08767995e-02 -3.96349907e-01
-4.75318074e-01 1.23517883e+00 2.42377117e-01 3.67817998e-01
7.17920661e-01 5.90976477e-01 -1.94249883e-01 -1.24826655e-01
-1.32549286e+00 -1.35444984e-01 2.85341531e-01 3.46689165e-01
4.69943225e-01 6.84938878e-02 -3.89908969e-01 5.71942151e-01
-1.91652939e-01 3.13482404e-01 -4.32048179e-02 1.11310637e+00
-9.63807926e-02 -1.01738787e+00 -1.23925991e-01 7.14640439e-01
-4.33828443e-01 -1.00500725e-01 2.62647510e-01 5.58715582e-01
1.14728130e-01 1.16346288e+00 2.34573498e-01 -1.07621741e+00
1.91374436e-01 -8.99190605e-02 -1.42246857e-01 3.51756871e-01
-6.17222726e-01 1.12488896e-01 -2.17412803e-02 -1.08581471e+00
-4.89601821e-01 -4.79181468e-01 -1.05015862e+00 -1.06045949e+00
-1.05693407e-01 -3.60572673e-02 1.01413131e+00 1.12338984e+00
8.79837632e-01 7.77003348e-01 9.88324702e-01 -1.02505112e+00
-4.69194293e-01 -6.92361593e-01 -3.83614987e-01 1.93961591e-01
2.18934134e-01 -7.58144796e-01 4.75514568e-02 5.84717607e-04] | [7.007943153381348, 2.845611095428467] |
7e4d51a4-2980-4b14-b56c-232d68b7e432 | vision-transformer-equipped-with-neural | 2204.02181 | null | https://arxiv.org/abs/2204.02181v1 | https://arxiv.org/pdf/2204.02181v1.pdf | Vision Transformer Equipped with Neural Resizer on Facial Expression Recognition Task | When it comes to wild conditions, Facial Expression Recognition is often challenged with low-quality data and imbalanced, ambiguous labels. This field has much benefited from CNN based approaches; however, CNN models have structural limitation to see the facial regions in distant. As a remedy, Transformer has been introduced to vision fields with global receptive field, but requires adjusting input spatial size to the pretrained models to enjoy their strong inductive bias at hands. We herein raise a question whether using the deterministic interpolation method is enough to feed low-resolution data to Transformer. In this work, we propose a novel training framework, Neural Resizer, to support Transformer by compensating information and downscaling in a data-driven manner trained with loss function balancing the noisiness and imbalance. Experiments show our Neural Resizer with F-PDLS loss function improves the performance with Transformer variants in general and nearly achieves the state-of-the-art performance. | ['Hyeon Yeo', 'Kyungtae Ko', 'Jiho Seo', 'Wei-Jin Park', 'Soyeon Kim', 'Hyeonbin Hwang'] | 2022-04-05 | null | null | null | null | ['facial-expression-recognition'] | ['computer-vision'] | [ 1.87828749e-01 1.92679331e-01 -1.60685137e-01 -7.95219123e-01
-4.78998095e-01 6.79001957e-02 3.19453657e-01 -6.20561659e-01
-2.35932603e-01 6.99443102e-01 2.11633772e-01 1.33748189e-01
1.06655598e-01 -7.68022954e-01 -6.94133461e-01 -8.69487643e-01
4.12705272e-01 5.93307950e-02 -2.54129227e-02 -4.53908116e-01
-1.67349830e-01 6.12961292e-01 -1.59696937e+00 5.39096713e-01
9.25249159e-01 1.55379105e+00 -1.44955069e-01 -2.26287008e-03
-2.89498806e-01 1.05072308e+00 -3.99648279e-01 -7.49944627e-01
4.94790494e-01 -4.18532044e-01 -5.19619167e-01 -1.64005578e-01
7.70559132e-01 -3.41668904e-01 -2.47667462e-01 9.80776966e-01
7.34243572e-01 -3.16661865e-01 4.03100193e-01 -1.52190518e+00
-8.62922549e-01 2.48325869e-01 -8.18745553e-01 -4.42036986e-02
-6.59202263e-02 1.72652751e-01 7.73914158e-01 -1.18706787e+00
3.88452441e-01 1.32606816e+00 9.66953754e-01 9.49314475e-01
-1.24520588e+00 -1.14250469e+00 1.94757327e-01 1.38047114e-01
-1.30981719e+00 -7.00973928e-01 9.16191459e-01 -8.71047750e-02
6.65227473e-01 7.48048425e-02 5.69780707e-01 1.38134205e+00
-1.11309454e-01 7.33402133e-01 1.39833534e+00 -1.57105610e-01
1.88401772e-03 1.27534375e-01 -3.80147547e-01 5.71011901e-01
-3.04355800e-01 -1.41348708e-02 -7.25337088e-01 2.46678755e-01
8.61424267e-01 8.13178122e-02 -2.72780955e-01 -1.62680998e-01
-4.66480315e-01 6.61075294e-01 9.60458994e-01 1.91571042e-01
-3.38739663e-01 7.37666041e-02 4.19107139e-01 3.26573253e-01
7.88524330e-01 1.64976165e-01 -4.86010969e-01 1.36168674e-01
-1.07021046e+00 2.27947697e-01 2.09098533e-01 8.29214334e-01
7.40264058e-01 3.19160789e-01 -4.54067647e-01 1.06950891e+00
8.79093856e-02 3.73281598e-01 3.98295254e-01 -8.60905409e-01
3.84745449e-01 8.12959075e-01 -2.77099282e-01 -9.19150114e-01
-3.75122160e-01 -6.47774518e-01 -1.28024459e+00 6.52351797e-01
5.55892766e-01 8.36211443e-02 -1.04519117e+00 2.13333535e+00
2.12365687e-01 1.08007722e-01 -9.30767059e-02 1.07364583e+00
8.81419659e-01 3.82457256e-01 1.83137178e-01 -4.64628823e-02
1.20283818e+00 -1.04503655e+00 -7.47309268e-01 -5.46086580e-02
2.08768129e-01 -6.00358725e-01 1.33575380e+00 5.32769084e-01
-1.09288096e+00 -6.63810670e-01 -8.26427400e-01 -3.17254215e-01
-2.51804829e-01 1.09199248e-01 5.36245823e-01 5.02607226e-01
-1.16431141e+00 5.88718832e-01 -4.22590792e-01 -3.00883353e-01
1.17126608e+00 3.37204516e-01 -6.21887922e-01 -4.00156938e-02
-1.17880642e+00 8.05951953e-01 -1.60147563e-01 5.06379545e-01
-8.18552554e-01 -8.07727814e-01 -6.60665631e-01 2.33696755e-02
-2.49571316e-02 -5.22487044e-01 9.33539629e-01 -1.76896882e+00
-1.71088910e+00 9.86447752e-01 5.80178248e-03 -3.62537086e-01
7.30657995e-01 -2.29785711e-01 -3.12108546e-01 -1.71165511e-01
-1.40677303e-01 9.50282753e-01 1.10095930e+00 -1.08889294e+00
-2.41464436e-01 -6.16931975e-01 -6.18533790e-02 8.43352973e-02
-7.39445865e-01 1.94912076e-01 -1.98472559e-01 -6.76251054e-01
-1.43440202e-01 -6.71949685e-01 -1.48797810e-01 7.06512928e-01
-1.42948255e-01 -1.13599062e-01 9.41599667e-01 -5.56837916e-01
1.01846087e+00 -2.17811489e+00 -5.72969094e-02 8.57943147e-02
2.64716744e-01 4.89091545e-01 -2.42897034e-01 -9.07606818e-03
-3.50194454e-01 4.96489927e-02 -1.39251038e-01 -4.98243541e-01
3.27343084e-02 2.91085035e-01 -4.47307289e-01 4.47498173e-01
5.74914873e-01 8.96718025e-01 -5.61892629e-01 -4.91262674e-01
-9.88512784e-02 8.26945543e-01 -7.91650534e-01 4.27456051e-01
-1.24569591e-02 5.33383965e-01 -2.99483597e-01 8.88717890e-01
1.27436721e+00 -1.53977022e-01 -1.96840927e-01 -4.93792057e-01
7.08891451e-02 -1.64054975e-01 -7.42694497e-01 1.81578076e+00
-4.53400642e-01 5.90708554e-01 3.90680373e-01 -1.02459455e+00
1.14490879e+00 1.74518257e-01 3.23894888e-01 -1.16043627e+00
1.16271399e-01 1.69045165e-01 -2.73329526e-01 -5.18022001e-01
1.80223718e-01 -5.17688036e-01 1.55493781e-01 -2.33494759e-01
2.28263736e-01 2.25369051e-01 -5.67891240e-01 -2.47446328e-01
8.01071167e-01 5.10293722e-01 -1.24720439e-01 -9.69604477e-02
3.05271298e-01 -4.29489344e-01 9.21427250e-01 2.33401790e-01
-2.80300349e-01 9.67176914e-01 5.96966445e-01 -5.51302016e-01
-1.07653213e+00 -9.37099397e-01 -3.89781743e-01 1.11567402e+00
-2.44839698e-01 -1.74280465e-01 -1.01395893e+00 -6.98929727e-01
-5.11529818e-02 1.33698106e-01 -9.05767024e-01 -2.89412588e-01
-4.78889912e-01 -7.16795266e-01 8.76998067e-01 7.73635805e-01
1.02857327e+00 -1.00559151e+00 -4.30654705e-01 -9.59981680e-02
-1.23672001e-01 -1.02650034e+00 -2.08487734e-01 2.73420274e-01
-6.48028851e-01 -6.07659638e-01 -1.02242541e+00 -5.22076130e-01
7.04133749e-01 -2.66492993e-01 1.07306659e+00 -4.62355800e-02
-1.61051393e-01 -4.16362017e-01 -2.82639354e-01 -6.09305263e-01
2.10212544e-01 1.24604471e-01 -1.12655722e-01 4.83552545e-01
2.61710405e-01 -8.29777241e-01 -9.06835556e-01 4.95893300e-01
-9.77057219e-01 7.44156614e-02 5.40693879e-01 1.04892099e+00
3.41733664e-01 -2.65960276e-01 7.14817584e-01 -6.75556362e-01
3.49590808e-01 -3.48297209e-01 -3.91714483e-01 3.23583074e-02
-6.42401636e-01 -5.90674803e-02 8.01415265e-01 -4.37817693e-01
-1.17456818e+00 -1.03776582e-01 -4.44547832e-01 -8.70346308e-01
-4.69910726e-02 1.35649800e-01 -6.18390620e-01 -2.47905746e-01
7.67130077e-01 -8.69343802e-02 2.80949414e-01 -5.02149940e-01
2.03405529e-01 6.04774535e-01 4.39007312e-01 -5.46043813e-01
5.78296661e-01 6.75967574e-01 1.00653991e-02 -4.08222049e-01
-1.06645989e+00 1.12427220e-01 -4.04948771e-01 -3.56234282e-01
6.59255922e-01 -1.04602230e+00 -7.27261364e-01 8.20928812e-01
-1.01803875e+00 -4.74263340e-01 -4.51731205e-01 -1.07480355e-01
-3.41300249e-01 -1.89833865e-01 -5.35979569e-01 -6.32926762e-01
-3.51764023e-01 -9.49330986e-01 1.11161757e+00 4.01583165e-01
1.09721757e-01 -4.86102879e-01 -1.38882503e-01 3.43161076e-01
9.97134924e-01 5.50793767e-01 4.90183234e-01 -2.65735358e-01
-4.30170536e-01 1.42644554e-01 -6.84955597e-01 9.00225759e-01
1.36712462e-01 -1.01652391e-01 -1.70473742e+00 -1.13382503e-01
-9.44482535e-02 -7.12637424e-01 9.60440278e-01 1.79895222e-01
1.82276249e+00 -4.90795434e-01 -8.03275220e-03 1.12475693e+00
1.42440259e+00 -3.56814712e-01 1.17599881e+00 1.57281071e-01
7.05138028e-01 8.17385137e-01 4.09682572e-01 4.57941115e-01
1.91247344e-01 8.53035331e-01 6.53113782e-01 -5.97000241e-01
-3.03411722e-01 -4.56329346e-01 2.35338524e-01 2.24652261e-01
-2.38673270e-01 -1.93191424e-01 -5.21564603e-01 4.06124413e-01
-1.79774535e+00 -1.01320994e+00 2.57553756e-01 1.86440086e+00
1.06625104e+00 -1.78505778e-02 -4.42735367e-02 3.77175361e-02
3.52981538e-01 2.00155452e-01 -5.76245368e-01 -4.20837313e-01
-4.91509765e-01 5.01252294e-01 4.00909662e-01 5.77386357e-02
-7.44878709e-01 1.02492189e+00 6.10705519e+00 1.26070559e+00
-1.78360593e+00 1.70323402e-01 1.09058344e+00 -2.32261598e-01
-2.63065189e-01 -2.31217042e-01 -6.94505394e-01 6.53541327e-01
7.80872881e-01 4.04377192e-01 2.37834349e-01 8.72398078e-01
1.90012470e-01 2.98318148e-01 -8.39457035e-01 1.17252374e+00
2.92409491e-02 -1.17292321e+00 1.38918400e-01 5.40618189e-02
6.00618422e-01 -1.23065606e-01 5.22061169e-01 3.07964802e-01
-2.60359235e-02 -1.67315805e+00 7.53791451e-01 7.14620531e-01
1.23333728e+00 -6.98791027e-01 7.97561824e-01 3.46069038e-02
-8.58348548e-01 -1.54972434e-01 -4.46437299e-01 -1.81111932e-01
-1.83569789e-01 6.67633951e-01 -6.16385102e-01 3.49369645e-01
1.10959125e+00 6.31949782e-01 -5.22905648e-01 6.56317830e-01
-8.10698122e-02 5.50783455e-01 -2.96019733e-01 2.37920210e-01
1.02002367e-01 -7.69719854e-02 2.39950102e-02 1.02481771e+00
3.22646528e-01 -9.65295266e-03 -2.22404763e-01 9.59778309e-01
-3.73862743e-01 9.16179493e-02 -6.09067321e-01 3.79009962e-01
1.31768450e-01 1.50475895e+00 -1.15923695e-01 4.26278599e-02
-2.18982249e-01 1.00381315e+00 6.19715452e-01 4.00613099e-01
-9.26908851e-01 -2.45429188e-01 9.20380056e-01 6.21306717e-01
1.83851227e-01 2.73944438e-01 -3.55855018e-01 -1.02398324e+00
3.48612756e-01 -9.17865336e-01 1.54859990e-01 -9.98198986e-01
-1.49823546e+00 1.00551081e+00 -2.72034794e-01 -1.23754871e+00
3.09479628e-02 -7.88376510e-01 -5.76102614e-01 9.53729391e-01
-1.89947176e+00 -1.60849857e+00 -7.08688557e-01 8.76700997e-01
1.09976351e-01 -1.59571275e-01 7.84319997e-01 7.60817111e-01
-5.26432574e-01 1.18404067e+00 -2.51242369e-01 1.89416155e-01
1.05335188e+00 -9.27486539e-01 -7.65271932e-02 4.89907801e-01
-3.68464947e-01 3.27587992e-01 5.36410272e-01 2.24313661e-02
-1.02707851e+00 -1.31246412e+00 7.44196475e-01 -3.02704126e-01
2.60925472e-01 -6.64774120e-01 -1.02775156e+00 5.24706483e-01
4.35068697e-01 7.39154696e-01 5.45467973e-01 5.36827073e-02
-7.45962858e-01 -8.72257471e-01 -1.62997389e+00 4.37705219e-01
1.26998484e+00 -4.88393396e-01 -3.28127742e-02 -1.07005708e-01
4.88180459e-01 -3.87678772e-01 -8.31662118e-01 7.05638766e-01
6.68562770e-01 -1.18962586e+00 7.65681684e-01 -5.89700997e-01
6.45081282e-01 -2.15153471e-01 -1.78301409e-01 -1.19896674e+00
-2.81912655e-01 -5.76129019e-01 4.16730523e-01 1.61047864e+00
1.74879998e-01 -6.80496395e-01 1.10966218e+00 4.85799044e-01
-7.05911964e-02 -1.14714408e+00 -1.14436758e+00 -5.62354743e-01
2.27185220e-01 -2.76349396e-01 8.86134326e-01 1.05314565e+00
-2.30545446e-01 2.67430216e-01 -6.25879586e-01 -2.11848602e-01
5.01654029e-01 -1.23070106e-01 7.75809228e-01 -1.10492623e+00
1.36120200e-01 -5.45435727e-01 -4.32044059e-01 -9.59609747e-01
3.15137386e-01 -7.56155729e-01 -1.03821948e-01 -1.05738866e+00
1.88560989e-02 -7.00663090e-01 -4.08281356e-01 8.56014848e-01
1.30735770e-01 7.18572438e-01 1.08312473e-01 -1.20358340e-01
-2.67286807e-01 1.04372704e+00 1.45361006e+00 -1.08845003e-01
2.88844764e-01 -3.15328926e-01 -9.43467021e-01 6.96176171e-01
6.63279355e-01 -1.97444409e-01 -3.78868550e-01 -5.87892234e-01
3.54218572e-01 -2.52501756e-01 6.37356400e-01 -9.70333874e-01
1.67204365e-01 -1.39435381e-01 7.57689357e-01 -9.84369814e-02
5.07556200e-01 -1.06055939e+00 1.52945578e-01 -1.62326172e-01
-3.35589379e-01 3.96108115e-03 2.91207343e-01 1.51795015e-01
-4.46618229e-01 3.41778785e-01 1.21321666e+00 4.43934798e-02
-4.62187827e-01 7.31795192e-01 1.49515510e-01 9.31644738e-02
6.30181789e-01 -4.55561250e-01 -1.95891440e-01 -2.74740785e-01
-4.87721205e-01 1.18768215e-01 5.71446359e-01 4.35188085e-01
7.58662462e-01 -1.65459943e+00 -7.92686403e-01 4.91601557e-01
2.02829674e-01 1.63374081e-01 3.91607016e-01 7.84584105e-01
-2.52693892e-01 -3.51162143e-02 -7.63666630e-01 -5.85407257e-01
-1.07186043e+00 3.64012390e-01 6.19327903e-01 -1.23874038e-01
-3.59217644e-01 1.13114667e+00 4.05498862e-01 -6.74145997e-01
3.70171756e-01 -3.23971182e-01 -1.43220201e-01 8.01653415e-02
5.27106524e-01 -6.09793030e-02 2.76011080e-01 -5.41060746e-01
-3.84065747e-01 6.81413114e-01 -3.34518738e-02 1.87333196e-01
1.41031551e+00 4.91720997e-02 -2.00651214e-01 -3.14879343e-02
1.36696446e+00 -1.68193832e-01 -1.67224252e+00 -3.14911932e-01
-4.59841579e-01 -6.22168005e-01 8.86817947e-02 -8.37548375e-01
-1.66455626e+00 1.06514752e+00 9.37438488e-01 -1.81257218e-01
1.62008119e+00 -1.95087641e-01 7.12652326e-01 -1.45558476e-01
2.85499036e-01 -1.07366753e+00 2.64857918e-01 3.46395791e-01
1.08714736e+00 -1.27159238e+00 -4.59871203e-01 -2.45591238e-01
-6.12934470e-01 9.73783612e-01 1.06940830e+00 -2.42982090e-01
7.83080757e-01 5.40506840e-01 3.02517474e-01 -1.72262490e-01
-7.73149788e-01 2.86138384e-03 1.38458401e-01 6.42551780e-01
4.41080004e-01 -2.55621642e-01 1.04872480e-01 6.63891733e-01
-1.85991585e-01 4.05736357e-01 -1.00922823e-01 4.57157105e-01
-1.69894502e-01 -9.49489236e-01 -1.92700535e-01 2.71027893e-01
-6.37432456e-01 -1.26993939e-01 -1.75020173e-01 6.21588409e-01
6.25806570e-01 5.29195368e-01 1.90141708e-01 -3.65878969e-01
5.56723833e-01 -1.36708841e-01 5.55756509e-01 -8.00654367e-02
-6.72411501e-01 -1.38479471e-01 -1.62648052e-01 -8.55792999e-01
-4.69562024e-01 -2.58485109e-01 -1.00694239e+00 -3.56325448e-01
-1.01588510e-01 -1.07744507e-01 4.66079563e-01 6.83424592e-01
5.19925058e-01 5.26435733e-01 7.15473413e-01 -8.08729053e-01
-6.04694247e-01 -1.12009251e+00 -5.23606420e-01 4.68845993e-01
5.43353438e-01 -7.13491440e-01 -2.68700063e-01 -1.19374879e-01] | [13.550697326660156, 1.5586473941802979] |
d042d0a5-f9a9-4d72-9ccc-87fa07bd8949 | an-empirical-study-of-quantum-dynamics-as-a | 2206.09241 | null | https://arxiv.org/abs/2206.09241v2 | https://arxiv.org/pdf/2206.09241v2.pdf | An Empirical Study of Quantum Dynamics as a Ground State Problem with Neural Quantum States | We consider the Feynman-Kitaev formalism applied to a spin chain described by the transverse field Ising model. This formalism consists of building a Hamiltonian whose ground state encodes the time evolution of the spin chain at discrete time steps. To find this ground state, variational wave functions parameterised by artificial neural networks -- also known as neural quantum states (NQSs) -- are used. Our work focuses on assessing, in the context of the Feynman-Kitaev formalism, two properties of NQSs: expressivity (the possibility that variational parameters can be set to values such that the NQS is faithful to the true ground state of the system) and trainability (the process of reaching said values). We find that the considered NQSs are capable of accurately approximating the true ground state of the system, i.e., they are expressive enough ans\"atze. However, extensive hyperparameter tuning experiments show that, empirically, reaching the set of values for the variational parameters that correctly describe the ground state becomes ever more difficult as the number of time steps increase because the true ground state becomes more entangled, and the probability distribution starts to spread across the Hilbert space canonical basis. | ['Fabio A. González', 'Herbert Vinck-Posada', 'Vladimir Vargas-Calderón'] | 2022-06-18 | null | null | null | null | ['variational-monte-carlo'] | ['miscellaneous'] | [ 1.49507254e-01 2.13528037e-01 -1.48048326e-02 -4.55198586e-02
-3.95247668e-01 -5.41925430e-01 6.55482531e-01 -3.24101478e-01
-4.80084926e-01 1.05305898e+00 -3.31515074e-01 -3.22338611e-01
-3.53911847e-01 -1.04633784e+00 -4.60282177e-01 -1.26968157e+00
-2.08510116e-01 8.11500669e-01 1.35454208e-01 -5.31702578e-01
2.50764072e-01 4.15698111e-01 -1.42800581e+00 -1.98118314e-01
8.82785499e-01 7.80694783e-01 -2.48377338e-01 6.08049989e-01
2.64922172e-01 6.90203071e-01 -4.56025243e-01 -3.67051333e-01
3.28517318e-01 -7.39017785e-01 -1.12584829e+00 -4.85389471e-01
1.78281486e-01 1.70733228e-01 -8.02313805e-01 1.71597648e+00
-1.36364534e-01 4.11515236e-01 7.53039360e-01 -9.85186398e-01
-6.79406345e-01 7.33500540e-01 5.04612803e-01 1.08862266e-01
-1.91005878e-02 6.21277452e-01 1.26879644e+00 -2.01519281e-01
8.99650276e-01 7.24998176e-01 4.89022523e-01 6.69882476e-01
-1.80449665e+00 -3.25054318e-01 -8.72544467e-01 3.47941220e-01
-1.80634189e+00 -3.97872686e-01 7.31592238e-01 -2.98632503e-01
1.10254800e+00 1.40914947e-01 7.92405248e-01 9.27742183e-01
7.88693070e-01 -5.01765013e-02 1.30038619e+00 -6.90941036e-01
6.92270875e-01 -1.08096331e-01 4.21183884e-01 8.22173238e-01
2.74083346e-01 5.86885035e-01 -3.32722247e-01 -2.35906705e-01
7.14875042e-01 -3.87288600e-01 -2.18982473e-01 -4.45917130e-01
-1.30356991e+00 9.50013697e-01 3.08445960e-01 7.29224622e-01
-5.80430984e-01 5.68052828e-01 2.14814901e-01 5.07322013e-01
-3.03080291e-01 8.85762990e-01 2.18196101e-02 -1.94190666e-01
-9.06756222e-01 5.29884934e-01 1.11731279e+00 3.73038977e-01
1.01615369e+00 2.38855369e-02 -2.00037330e-01 -9.06984732e-02
9.49734300e-02 8.19992363e-01 2.01106638e-01 -1.33871651e+00
1.52207315e-02 2.67274648e-01 3.27513903e-01 -5.14903665e-01
-3.00280571e-01 -3.21636528e-01 -1.01562202e+00 1.77881628e-01
6.24425173e-01 -3.41171026e-02 -9.52414632e-01 1.93267500e+00
-1.32842794e-01 -3.95067781e-02 4.96219426e-01 8.44436347e-01
4.14568365e-01 9.34453964e-01 -7.11536482e-02 -5.70067763e-01
1.12908173e+00 -2.70737559e-01 -9.82328951e-01 3.92538160e-02
6.14797950e-01 -2.17095092e-01 7.27301121e-01 2.66065389e-01
-1.33422673e+00 -3.22739810e-01 -1.28012490e+00 3.96690398e-01
-2.40611702e-01 -5.53181410e-01 6.95114136e-01 6.42744958e-01
-1.19571424e+00 1.09628558e+00 -9.37907040e-01 -6.59179166e-02
-3.26062769e-01 6.43691897e-01 -5.21101177e-01 1.99256539e-01
-1.77651966e+00 1.19076288e+00 9.96946096e-01 3.18856776e-01
-8.13295186e-01 2.26751398e-02 -4.21395123e-01 8.81245732e-02
7.05095157e-02 -5.89490235e-01 1.09207320e+00 -8.51957679e-01
-1.63870203e+00 6.56181097e-01 2.84383930e-02 -3.10098290e-01
-8.18111673e-02 8.67508233e-01 -5.51417410e-01 2.42308497e-01
-3.48417521e-01 2.93622255e-01 5.62805593e-01 -8.67597163e-01
2.21764326e-01 -1.26647025e-01 1.58838600e-01 -2.92740613e-01
1.28365234e-01 -4.67389971e-01 2.47734971e-02 2.50821680e-01
4.93115664e-01 -1.38195372e+00 -2.99565613e-01 -8.59932303e-01
-6.23121738e-01 -3.72049391e-01 1.80075943e-01 -1.59584105e-01
1.35232210e+00 -2.01695800e+00 6.03245437e-01 7.44104028e-01
2.58355886e-01 2.81939834e-01 4.79883067e-02 8.37347448e-01
-1.21426947e-01 -8.66437927e-02 -3.27873766e-01 2.88351357e-01
2.39698902e-01 5.86088598e-01 2.77803428e-02 5.12783289e-01
-1.08955570e-01 9.94152606e-01 -7.25534797e-01 -4.29803729e-01
-4.53960337e-02 4.14801121e-01 -7.02391326e-01 -6.89714700e-02
-3.45079511e-01 5.01534402e-01 -4.34795022e-01 -1.58307418e-01
3.89744848e-01 -5.61680496e-01 5.23639917e-01 -3.14025253e-01
-2.75826722e-01 2.64330149e-01 -1.25966871e+00 1.21923828e+00
-4.26429808e-02 4.16312397e-01 -4.31381851e-01 -1.01478648e+00
7.82084227e-01 4.62271780e-01 3.09206218e-01 -7.26551652e-01
3.52696806e-01 4.60785359e-01 6.13840699e-01 -5.09357452e-01
4.99196529e-01 -7.81177282e-01 -3.13842028e-01 4.67894197e-01
4.26680297e-01 -4.36402053e-01 6.32386863e-01 1.27406254e-01
9.39992607e-01 -1.45250618e-01 2.42742822e-01 -5.45936167e-01
4.87971634e-01 -6.32814616e-02 2.12825552e-01 1.15975332e+00
-7.80558214e-02 1.59098506e-01 8.28821898e-01 -5.68442106e-01
-1.48594093e+00 -1.28111780e+00 -5.31133413e-01 1.58615127e-01
2.05768317e-01 -3.21037233e-01 -1.11344814e+00 1.41519075e-02
-3.51537228e-01 8.79123151e-01 -8.22077215e-01 -6.27118170e-01
-4.56277609e-01 -9.14425731e-01 4.84061182e-01 -1.14289299e-01
5.53579211e-01 -1.40967393e+00 -6.37636542e-01 3.52835238e-01
-2.24257395e-01 -8.29391420e-01 3.16694826e-02 5.94378352e-01
-8.38616848e-01 -8.22269917e-01 -2.56430268e-01 -2.40425110e-01
4.51774597e-01 -8.02694678e-01 1.00653327e+00 3.91380452e-02
-3.43078375e-02 -1.29829019e-01 2.67544147e-02 4.97037828e-01
-1.02843487e+00 3.84527333e-02 2.54558891e-01 -3.61815542e-02
3.51725668e-01 -4.98148412e-01 -4.55522388e-01 -9.28338468e-02
-1.14264917e+00 -9.55613926e-02 2.66810179e-01 7.54622757e-01
4.81099904e-01 3.40792477e-01 -3.94484820e-03 -5.97317815e-01
5.43787122e-01 -1.50196970e-01 -7.72642195e-01 2.39172176e-01
-6.05473757e-01 9.30434644e-01 8.01191211e-01 -2.80020952e-01
-7.05619395e-01 -2.49822661e-01 -1.93448886e-01 -6.53488189e-02
-1.59873068e-01 6.24935925e-01 2.99969673e-01 -3.39635640e-01
8.24288368e-01 5.19044101e-01 -8.28845277e-02 -1.22794192e-02
1.23365059e-01 4.56065089e-01 5.30838192e-01 -7.30612218e-01
7.00107336e-01 1.01765305e-01 7.53389180e-01 -5.85983396e-01
-7.77126908e-01 1.33229405e-01 -6.54066682e-01 -1.40757039e-01
9.34902012e-01 -8.91519189e-02 -1.27896452e+00 5.23542523e-01
-1.08981097e+00 -3.18373531e-01 -5.74565291e-01 6.17031157e-01
-8.20076585e-01 3.16524923e-01 -8.67081642e-01 -9.54834938e-01
-1.16856068e-01 -1.44024265e+00 4.79374230e-01 2.42439672e-01
-1.12599388e-01 -1.19047999e+00 6.33835256e-01 -1.02253959e-01
4.34776396e-01 1.19836777e-01 1.33712494e+00 -3.71246159e-01
-6.36080384e-01 -5.60171366e-01 2.63226867e-01 3.40813905e-01
-5.73188543e-01 1.19935200e-01 -6.00445092e-01 -2.22030103e-01
-2.85309926e-02 -1.16064273e-01 5.97936749e-01 4.09396172e-01
4.21420097e-01 -3.49696875e-01 9.30428039e-03 3.83202702e-01
1.71930325e+00 3.79957914e-01 8.60494077e-01 7.79126883e-02
3.31141204e-01 1.38333544e-01 -2.19954789e-01 8.34488571e-02
9.72772762e-02 8.57208848e-01 3.87690365e-01 6.20430827e-01
4.37598228e-01 4.41779802e-03 3.17577869e-01 1.00532091e+00
-7.30008483e-01 -3.40378493e-01 -9.93446112e-01 1.15565240e-01
-1.67130566e+00 -1.38684511e+00 -3.22996557e-01 2.51415825e+00
9.99269903e-01 3.36878031e-01 -2.26818863e-02 1.43313795e-01
4.79991525e-01 1.46390244e-01 -2.96484321e-01 -6.47160888e-01
-1.12937026e-01 5.17932057e-01 3.49871516e-01 7.48668849e-01
-6.70480013e-01 7.30660796e-01 6.89418554e+00 3.91074479e-01
-1.12111962e+00 1.24541238e-01 7.28386715e-02 1.87474072e-01
-2.65820086e-01 3.64288270e-01 -4.77189779e-01 6.92758977e-01
1.52180970e+00 -6.15429804e-02 1.05323124e+00 1.22835241e-01
-8.50577131e-02 -5.87745979e-02 -1.05800402e+00 8.28158736e-01
-3.86134297e-01 -1.44302011e+00 -1.31792873e-01 3.39086533e-01
9.25102293e-01 1.12125576e-01 7.78501853e-04 3.71986181e-01
3.25974405e-01 -1.09981263e+00 8.73965740e-01 1.00613415e+00
7.73496151e-01 -7.04402506e-01 1.04112625e+00 4.03561264e-01
-6.27055287e-01 6.57510944e-03 -2.43688375e-01 -2.95976046e-02
1.73852086e-01 4.10159349e-01 -1.04793303e-01 2.61842608e-01
1.59190029e-01 -1.51774168e-01 -2.19242021e-01 7.55980670e-01
-1.64907277e-02 6.41125381e-01 -4.57952082e-01 -3.41979444e-01
6.11652255e-01 -6.85805619e-01 4.12028044e-01 5.82085788e-01
3.55109602e-01 4.62707311e-01 -3.68396044e-02 1.21441150e+00
1.23907879e-01 -3.35121661e-01 -5.43133199e-01 -4.40298676e-01
2.12690666e-01 8.49226117e-01 -5.82542717e-01 -2.90735930e-01
1.90925971e-01 7.97844112e-01 2.75220156e-01 6.87935948e-01
-5.92692077e-01 -6.61474228e-01 3.10356766e-01 -1.89497069e-01
2.65993953e-01 -3.41601610e-01 2.09598780e-01 -1.30299413e+00
-3.78022760e-01 -6.90157592e-01 -1.62279345e-02 -6.69394612e-01
-9.22257304e-01 8.37233663e-01 1.71372011e-01 -6.07929111e-01
-6.07350349e-01 -6.63523376e-01 -6.39112532e-01 9.15823340e-01
-7.62511849e-01 -6.24579668e-01 2.18118474e-01 4.95824248e-01
-6.55805707e-01 1.09154493e-01 1.17943239e+00 -1.10882729e-01
-5.00443697e-01 2.35342056e-01 4.68999535e-01 -1.30754843e-01
-2.03713924e-01 -1.26504850e+00 1.51072100e-01 6.03607953e-01
1.27069801e-01 1.02560914e+00 1.33484828e+00 -4.04377908e-01
-1.55423021e+00 -2.70856053e-01 1.09486163e+00 -2.60615528e-01
7.77265489e-01 -2.90531516e-01 -9.88952398e-01 5.78720272e-01
1.93316400e-01 1.73230797e-01 3.32756788e-01 1.43296286e-01
-1.38377115e-01 4.59914878e-02 -1.03874660e+00 4.89678800e-01
5.92844486e-01 -8.73758972e-01 -5.02830803e-01 3.61150295e-01
1.31574124e-01 -1.72553524e-01 -1.03199470e+00 1.22959025e-01
5.74862957e-01 -1.13290668e+00 6.12919748e-01 -7.36422896e-01
1.64762661e-01 -1.99325413e-01 -1.85620502e-01 -1.22480345e+00
-6.02564096e-01 -8.95207822e-01 -2.57393688e-01 3.48105729e-01
2.98035651e-01 -7.08553731e-01 6.12114608e-01 5.63485920e-01
1.91852272e-01 -4.35354054e-01 -1.42924035e+00 -8.41970384e-01
2.66264260e-01 -2.23264039e-01 5.92622519e-01 8.91596675e-01
4.04338598e-01 3.39724660e-01 -1.62739247e-01 2.04623610e-01
8.31178904e-01 8.27518851e-02 4.00681011e-02 -1.29752243e+00
-3.98094237e-01 -3.66588116e-01 -6.36775911e-01 -5.11968315e-01
3.22176099e-01 -1.12035429e+00 2.15992928e-01 -1.06580698e+00
3.27664673e-01 -2.24400550e-01 -6.36866987e-01 2.10215181e-01
3.02338660e-01 4.27162439e-01 1.65307581e-01 3.72661173e-01
-6.31769896e-01 3.16428185e-01 1.25838852e+00 1.13815002e-01
-1.04208840e-02 -1.45554334e-01 4.27324735e-02 3.65572989e-01
5.71898520e-01 -5.33897460e-01 -5.67565300e-03 2.02639505e-01
9.00563657e-01 7.19872594e-01 7.62116253e-01 -1.21790254e+00
2.49602243e-01 -2.72290856e-01 -2.69994974e-01 -1.60188839e-01
4.51679617e-01 -5.55312872e-01 7.74150610e-01 7.85383880e-01
-3.98678720e-01 -2.24316359e-01 -3.15308243e-01 1.88532144e-01
-1.15181617e-01 -1.02598357e+00 9.88706768e-01 -1.20850399e-01
-3.74343693e-01 1.84251577e-01 -6.00409389e-01 -9.46382508e-02
6.81728780e-01 -9.60060135e-02 -7.82086924e-02 -2.55835354e-01
-1.06252646e+00 -4.18328434e-01 6.59197211e-01 -4.07353461e-01
5.05533256e-02 -1.33396292e+00 -3.67988706e-01 4.97602165e-01
7.11586326e-02 -6.28392756e-01 3.74655068e-01 8.65217447e-01
-6.66045010e-01 8.32181036e-01 -2.43297100e-01 -4.58403677e-01
-4.20530736e-01 3.23927969e-01 9.09896314e-01 -4.52721477e-01
-4.71400350e-01 4.48513895e-01 -1.64042622e-01 -3.81535172e-01
-4.15986240e-01 -3.01688492e-01 1.08642384e-01 -3.85459870e-01
1.08274110e-01 1.83913842e-01 -3.94227095e-02 -8.14727902e-01
-2.01336756e-01 3.78680050e-01 2.48268694e-01 -3.48938465e-01
1.10546207e+00 2.78999239e-01 -4.98887330e-01 5.22336841e-01
1.19970584e+00 -3.29512209e-01 -9.15663779e-01 -7.94080943e-02
-2.00449467e-01 1.52981937e-01 3.81060600e-01 -5.77605963e-01
-6.94014072e-01 8.10349882e-01 7.01394200e-01 9.34389055e-01
5.15297890e-01 5.62721342e-02 6.25517786e-01 8.90789688e-01
6.03470743e-01 -1.00483179e+00 -3.50274295e-01 8.43166351e-01
2.44437456e-01 -8.10163379e-01 -3.81171644e-01 3.56779575e-01
-4.15418774e-01 1.42012763e+00 1.04559120e-02 -2.29391783e-01
4.31432247e-01 -6.86091557e-02 -3.35986197e-01 -6.29364192e-01
-5.60016513e-01 3.73740606e-02 3.19586128e-01 2.31048331e-01
1.47272021e-01 3.17769885e-01 -1.34562775e-01 1.52462572e-01
-4.35772151e-01 7.16507286e-02 7.87711442e-01 4.52210307e-01
-4.42957312e-01 -1.01710534e+00 -3.99566084e-01 2.97764868e-01
-3.45271587e-01 -1.42442778e-01 1.43575430e-01 6.35283232e-01
1.18910514e-01 6.38476074e-01 1.05786733e-01 -1.58961266e-01
2.30454668e-01 4.49050754e-01 1.05370247e+00 -3.05344820e-01
-2.56904006e-01 -4.60091561e-01 6.65581971e-02 -5.19518018e-01
-3.40453833e-01 -5.07956505e-01 -1.31488001e+00 -7.11398244e-01
-5.91489077e-01 7.75924861e-01 5.18075287e-01 1.34997487e+00
-1.71749756e-01 1.96084961e-01 3.64086062e-01 -6.51710927e-01
-9.16926146e-01 -7.29804873e-01 -9.70976055e-01 3.91098797e-01
4.44118261e-01 -4.39828247e-01 -5.28405547e-01 -1.97620153e-01] | [5.557745456695557, 4.9569854736328125] |
d271938c-896f-4407-ae01-2c748676c1bc | sagc-a68-a-space-access-graph-dataset-for-the | 2307.04515 | null | https://arxiv.org/abs/2307.04515v1 | https://arxiv.org/pdf/2307.04515v1.pdf | SAGC-A68: a space access graph dataset for the classification of spaces and space elements in apartment buildings | The analysis of building models for usable area, building safety, and energy use requires accurate classification data of spaces and space elements. To reduce input model preparation effort and errors, automated classification of spaces and space elements is desirable. A barrier hindering the utilization of Graph Deep Learning (GDL) methods to space function and space element classification is a lack of suitable datasets. To bridge this gap, we introduce a dataset, SAGC-A68, which comprises access graphs automatically generated from 68 digital 3D models of space layouts of apartment buildings. This graph-based dataset is well-suited for developing GDL models for space function and space element classification. To demonstrate the potential of the dataset, we employ it to train and evaluate a graph attention network (GAT) that predicts 22 space function and 6 space element classes. The dataset and code used in the experiment are available online. https://doi.org/10.5281/zenodo.7805872, https://github.com/A2Amir/SAGC-A68. | ['Georg Suter', 'Amir Ziaee'] | 2023-07-10 | null | null | null | null | ['graph-attention', 'classification-1'] | ['graphs', 'methodology'] | [-4.71875072e-02 1.16072923e-01 5.07762358e-02 -4.06512141e-01
-3.93595546e-01 -6.44105673e-01 4.37850654e-01 5.72185695e-01
1.72226742e-01 5.36372483e-01 2.97537833e-01 -1.01771331e+00
-4.04789269e-01 -1.43891633e+00 -7.19625950e-01 -1.87507540e-01
-1.32000670e-01 3.58745873e-01 -1.41927943e-01 -3.75260085e-01
3.72061841e-02 7.23935187e-01 -1.54351592e+00 2.35427663e-01
9.82769787e-01 1.05240059e+00 1.97224826e-01 6.46779776e-01
-9.56751406e-02 7.53499210e-01 -3.20702553e-01 1.83656767e-01
2.36403391e-01 -4.47869390e-01 -8.08928609e-01 -5.69879264e-02
4.16613221e-01 -2.49977380e-01 -6.96204126e-01 5.43451071e-01
5.52925885e-01 5.11474252e-01 7.75800526e-01 -1.61656713e+00
-7.97193348e-01 5.03362060e-01 2.87703097e-01 1.12409860e-01
3.37738782e-01 1.83979303e-01 1.15489137e+00 -5.86908758e-01
2.69566268e-01 9.03751969e-01 7.14778483e-01 1.71832487e-01
-1.15965772e+00 -4.48099971e-01 4.13989276e-02 3.32644284e-01
-1.76526034e+00 -1.72691420e-01 1.09995079e+00 -5.96578538e-01
1.16912568e+00 7.05039501e-01 1.30234611e+00 9.85422552e-01
-1.66986838e-01 7.56224751e-01 8.23143899e-01 -6.06761754e-01
5.62331855e-01 -1.45423472e-01 2.01477677e-01 8.69997442e-01
2.44914934e-01 -3.73376645e-02 -7.99114928e-02 2.57084370e-01
6.96737885e-01 2.28539824e-01 -2.56504178e-01 -4.98943508e-01
-7.65264452e-01 7.79428959e-01 1.05612147e+00 3.40585798e-01
-2.18371764e-01 2.55087435e-01 1.30027145e-01 2.00322196e-02
6.01835907e-01 4.80665773e-01 -2.27260381e-01 -1.90873705e-02
-6.42657876e-01 3.84467334e-01 5.21890759e-01 1.12456501e+00
7.82902062e-01 1.45773619e-01 1.19488202e-01 9.07692671e-01
3.44117522e-01 3.95348340e-01 7.42652640e-02 -4.75760937e-01
5.79326808e-01 1.22693717e+00 -1.36364236e-01 -1.11187053e+00
-7.89869010e-01 -1.19101584e-01 -5.72280943e-01 1.34576093e-02
8.11972693e-02 1.73191518e-01 -1.07565308e+00 1.40185845e+00
8.10419768e-02 -1.52682692e-01 -4.42815483e-01 6.38369739e-01
1.08805263e+00 6.37544513e-01 1.59038201e-01 7.03971267e-01
9.80619490e-01 -7.36063063e-01 -3.87380034e-01 -3.41630369e-01
1.07020211e+00 -1.04397446e-01 1.43831599e+00 -1.90210372e-01
-7.71476328e-01 -6.40946329e-01 -1.26216054e+00 -3.13864142e-01
-1.19456470e+00 1.12557068e-01 6.58966601e-01 5.47188282e-01
-1.14552259e+00 6.10379577e-01 -7.84535527e-01 -5.82858503e-01
9.34928775e-01 1.49754196e-01 -1.29660770e-01 1.97957829e-02
-1.06573701e+00 9.61761892e-01 4.07892644e-01 2.01236635e-01
-6.54795885e-01 -7.43735254e-01 -1.13301408e+00 1.54607981e-01
4.73866090e-02 -5.18871427e-01 9.32585359e-01 -2.68729538e-01
-5.76947987e-01 7.70300388e-01 5.01920044e-01 -4.50868696e-01
2.84751892e-01 -3.22366655e-02 -4.47519571e-01 -3.58823031e-01
1.93725482e-01 4.60856467e-01 8.06569979e-02 -1.13285327e+00
-3.51160765e-01 -5.14184296e-01 2.89864093e-01 -2.01398949e-03
-2.13862315e-01 -6.44574285e-01 4.01001144e-03 -4.08952385e-01
3.53751630e-02 -1.01828349e+00 -5.87931462e-02 -4.40974198e-02
-6.22677267e-01 -7.54727274e-02 6.06587708e-01 -9.37227845e-01
1.47160935e+00 -2.08422637e+00 -3.19345146e-01 4.04174268e-01
2.27680847e-01 1.21714965e-01 -2.14165151e-01 8.24344516e-01
-3.05823237e-01 3.94271195e-01 -3.88500363e-01 -1.55997083e-01
2.99960226e-01 1.12027958e-01 5.38684651e-02 4.30725902e-01
1.20406680e-01 1.12058628e+00 -7.18346715e-01 -1.70565709e-01
8.13593328e-01 6.53358638e-01 -2.93626428e-01 1.54262021e-01
-3.77033740e-01 1.25399128e-01 -5.04649460e-01 7.16087818e-01
4.85500276e-01 -3.80751908e-01 1.64544016e-01 -1.74241319e-01
-8.14170856e-03 5.37025928e-01 -9.25698161e-01 1.48308444e+00
-5.74466825e-01 8.93440783e-01 -4.64084148e-01 -9.56079423e-01
1.03029597e+00 -1.38175875e-01 6.50561452e-01 -1.09298527e+00
3.22529793e-01 3.88626792e-02 -1.99531868e-01 -4.31353837e-01
4.96604264e-01 3.48924935e-01 -2.75396317e-01 3.39865029e-01
4.03064638e-02 -5.29671133e-01 5.01282811e-02 9.73269492e-02
1.27529407e+00 1.73597679e-01 2.27561265e-01 -3.83467704e-01
2.44182110e-01 5.88263012e-02 -7.85871744e-02 2.66929597e-01
-6.35116845e-02 5.37894726e-01 3.20464782e-02 -8.08518231e-01
-1.22669065e+00 -1.03381193e+00 2.03490350e-02 8.75098109e-01
-1.15216605e-01 -7.70334959e-01 -7.88965166e-01 -5.19425750e-01
3.05550873e-01 1.38282514e+00 -6.63178086e-01 -1.91298380e-01
-3.60118002e-01 -3.39821279e-01 3.75498205e-01 7.81110942e-01
2.15380862e-01 -9.29391563e-01 -8.37338448e-01 -7.50457868e-03
-2.05016434e-01 -7.35095024e-01 -1.43949557e-02 3.49669993e-01
-6.42497838e-01 -1.31314659e+00 -4.94163632e-02 -4.57560748e-01
7.23009527e-01 2.63223022e-01 1.29330385e+00 5.17730415e-01
-4.54056442e-01 6.72062218e-01 -4.12567884e-01 -5.02148688e-01
-2.94012606e-01 3.68408561e-01 -1.43218756e-01 -4.46874887e-01
4.59129781e-01 -7.26917446e-01 -5.55789173e-01 1.80852875e-01
-7.46564269e-01 5.20059586e-01 -4.28570323e-02 3.45907003e-01
4.87005979e-01 1.75058469e-01 1.75237983e-01 -4.86088216e-01
4.51183975e-01 -5.25627017e-01 -7.63967991e-01 2.05994412e-01
-8.23779702e-01 -1.81468293e-01 5.70947170e-01 2.67422825e-01
-4.09721643e-01 2.87298989e-02 -1.87347889e-01 1.45504382e-02
-4.00748163e-01 4.85093236e-01 -3.96311611e-01 5.41971996e-02
8.04917634e-01 -6.77876669e-05 -2.87728071e-01 -5.02921104e-01
3.61409783e-01 6.02169216e-01 9.54474062e-02 -4.49295104e-01
8.11799467e-01 1.10406995e-01 8.79727583e-03 -1.03647721e+00
-6.48307085e-01 -1.13286763e-01 -8.24231803e-01 -4.91906255e-01
9.22099411e-01 -7.47710824e-01 -1.93423197e-01 1.84863359e-01
-7.18476176e-01 -1.01649714e+00 -3.15488994e-01 1.18353061e-01
-6.53343558e-01 -2.13200375e-01 -1.52996285e-02 -9.19742465e-01
-2.35557109e-01 -6.81204319e-01 8.89994562e-01 4.30936851e-02
-5.25644302e-01 -1.10124063e+00 -9.84818190e-02 4.48965013e-01
3.40569079e-01 8.75358999e-01 1.13754261e+00 -6.43799901e-01
-9.24476564e-01 -2.97707856e-01 -1.07410945e-01 2.32229024e-01
3.74809980e-01 -8.15217663e-03 -9.47726369e-01 -9.50912163e-02
-4.23238873e-01 -1.93682358e-01 5.01130760e-01 4.11109477e-01
1.65529048e+00 -2.52602309e-01 -3.76026660e-01 6.27835095e-01
1.57675803e+00 4.54438299e-01 6.83943212e-01 5.58260143e-01
1.14856184e+00 4.97200668e-01 4.31270212e-01 4.06189591e-01
6.79718971e-01 6.21226668e-01 6.27689004e-01 -1.60427302e-01
-2.44316444e-01 -6.97908401e-01 -2.29321763e-01 6.21746898e-01
-7.62479454e-02 -4.45561498e-01 -1.63733065e+00 8.55851710e-01
-1.56446111e+00 -7.12961197e-01 -2.86271840e-01 1.78682029e+00
2.10668638e-01 2.30220668e-02 1.82326138e-01 5.49069941e-01
2.92705297e-01 3.67242962e-01 -3.16074580e-01 -1.49277151e-01
1.27482608e-01 1.01982020e-01 5.42640150e-01 4.72764403e-01
-1.13567960e+00 7.07354426e-01 5.02987957e+00 4.11633641e-01
-7.01873302e-01 -2.97794998e-01 6.95983052e-01 -1.57162607e-01
-5.51387250e-01 -1.30380735e-01 -3.48508686e-01 4.79579985e-01
1.26582265e+00 6.42652735e-02 8.92698288e-01 1.10241449e+00
2.36787915e-01 5.30778430e-02 -1.26444829e+00 9.97108817e-01
-1.16493501e-01 -1.66160429e+00 -3.27411145e-01 1.25916988e-01
4.49974477e-01 2.16269687e-01 -1.65355906e-01 3.46436143e-01
2.25324720e-01 -1.11646652e+00 8.25942039e-01 3.77634466e-01
6.89576507e-01 -6.09358847e-01 3.70693117e-01 1.15587249e-01
-1.46124661e+00 -2.09599063e-01 2.07980745e-03 -2.48049662e-01
3.14136595e-02 1.38243943e-01 -1.10988235e+00 5.88577151e-01
8.03819120e-01 5.97733378e-01 -1.19109666e+00 8.54730010e-01
-2.21053630e-01 5.57395816e-01 -3.96508753e-01 -3.22662085e-01
1.92354009e-01 -4.60611761e-01 2.55548730e-02 8.61656725e-01
6.43177569e-01 4.31522168e-02 -8.02562907e-02 1.01027906e+00
9.04398561e-02 -1.93523290e-03 -1.15701044e+00 -4.29368705e-01
7.78767824e-01 9.40768421e-01 -9.21351433e-01 -3.40247639e-02
-2.69573271e-01 5.67695320e-01 4.03400898e-01 3.02738577e-01
-1.08034742e+00 -2.71340579e-01 4.20727402e-01 7.36532807e-01
2.78825194e-01 -4.75038558e-01 -7.19829142e-01 -6.59768999e-01
-4.81088739e-03 -5.84405482e-01 3.08145583e-01 -1.23211586e+00
-9.84237075e-01 5.73022246e-01 1.58501893e-01 -1.09611368e+00
5.04520684e-02 -7.36895978e-01 -5.72578609e-01 8.26789379e-01
-9.15121019e-01 -1.66430748e+00 -8.80936921e-01 2.79220223e-01
3.77268106e-01 -2.37635225e-01 1.00788772e+00 3.26795995e-01
-4.73153323e-01 2.70172060e-01 1.16069037e-04 2.09900245e-01
-2.34923452e-01 -1.37439489e+00 1.09088457e+00 6.53825045e-01
3.00983518e-01 3.17277253e-01 4.48431522e-01 -6.52243555e-01
-1.33840477e+00 -1.47514784e+00 6.95596755e-01 -8.81176531e-01
5.89960754e-01 -9.09731090e-01 -7.61593044e-01 9.80795085e-01
-2.22153813e-01 -4.72433195e-02 8.86792421e-01 8.37352350e-02
6.41006930e-03 -1.16219036e-02 -9.90403295e-01 6.49323225e-01
1.47586334e+00 -9.01407897e-01 -2.08396882e-01 3.90434802e-01
5.91735423e-01 -3.82944465e-01 -9.59354699e-01 2.78073341e-01
2.40457296e-01 -7.56455362e-01 1.03522170e+00 -3.74287367e-01
4.63380218e-01 -3.20199847e-01 -6.47391915e-01 -1.28800213e+00
-6.19522393e-01 8.81888047e-02 -2.15408579e-01 1.04454613e+00
5.40447176e-01 -5.46269953e-01 7.82443583e-01 9.19802666e-01
-5.77713013e-01 -8.40364695e-01 -7.05591023e-01 -6.65970266e-01
-1.68345645e-02 -8.10169935e-01 1.22686410e+00 1.01397860e+00
-3.25881124e-01 3.21372986e-01 1.23680793e-01 9.04908553e-02
2.70322531e-01 1.25618130e-01 7.64148831e-01 -1.21051466e+00
1.51442692e-01 -4.74301934e-01 -5.41986525e-01 -4.53082383e-01
2.77511813e-02 -1.15549040e+00 -2.84958839e-01 -2.50489974e+00
-4.45835292e-01 -7.49077737e-01 -1.82944670e-01 7.77916312e-01
2.90390670e-01 -8.67548436e-02 6.04948364e-02 -2.90400624e-01
-2.94052571e-01 7.70627916e-01 7.85604060e-01 -5.02259791e-01
-2.34312072e-01 -2.45236643e-02 -5.42081118e-01 3.14585567e-01
1.50438952e+00 -1.02421768e-01 -7.59374142e-01 -5.74406207e-01
1.32266730e-01 -3.14315736e-01 6.30539894e-01 -1.24693727e+00
-2.05891758e-01 -1.59144446e-01 5.72123885e-01 -1.05588305e+00
4.82341111e-01 -1.01312089e+00 5.93728006e-01 3.76895219e-01
-2.22088680e-01 1.76840842e-01 5.21788538e-01 2.37031475e-01
1.94921717e-01 4.54821400e-02 3.29447955e-01 -1.39307290e-01
-9.38027859e-01 2.78813481e-01 -2.93476999e-01 -2.49523625e-01
1.02324438e+00 -4.41493183e-01 -4.97403741e-01 -2.30988115e-01
-5.53813398e-01 2.95641541e-01 7.13270426e-01 6.89149320e-01
7.00966358e-01 -1.72559333e+00 -3.26357156e-01 4.28339571e-01
4.54354435e-01 2.78701603e-01 4.60115165e-01 2.34196052e-01
-9.90178585e-01 5.49647689e-01 -4.34654504e-01 -2.71689117e-01
-9.76404428e-01 5.60436904e-01 4.32536393e-01 2.55688757e-01
-8.03982139e-01 7.18133330e-01 -2.10052386e-01 -7.41138637e-01
7.61650689e-03 -7.40462184e-01 -8.12998787e-02 -1.82510957e-01
8.32795650e-02 5.60545266e-01 1.83094054e-01 -5.91052473e-01
-5.02663314e-01 1.55972153e-01 6.21428072e-01 3.12508523e-01
1.67657137e+00 2.41034385e-03 8.77343714e-02 7.02406764e-01
1.13090241e+00 -4.07283545e-01 -8.98746908e-01 2.37263933e-01
1.74960300e-01 -5.65672338e-01 1.33067876e-01 -8.25822115e-01
-7.41733193e-01 7.29934692e-01 9.84851062e-01 7.29432046e-01
9.96144772e-01 -3.32015716e-02 6.59053922e-01 1.12944454e-01
4.17270213e-01 -1.22888923e+00 -1.75787240e-01 2.65493661e-01
1.31040394e+00 -1.10186481e+00 7.92769864e-02 -1.91759303e-01
-2.17107818e-01 7.79828072e-01 7.19907582e-01 -3.50431493e-03
8.27491403e-01 -8.14912617e-02 -3.44687939e-01 -5.50046623e-01
-2.72232920e-01 -2.30774894e-01 4.44326788e-01 8.64102602e-01
4.67327356e-01 6.07833505e-01 2.44002432e-01 4.10072982e-01
-7.34207213e-01 -6.82706609e-02 2.44616330e-01 1.00134206e+00
-1.36066318e-01 -7.95840681e-01 -1.32792860e-01 6.79651916e-01
3.55405092e-01 -2.35331096e-02 -4.77149755e-01 8.65416765e-01
1.35658219e-01 8.48559678e-01 2.06377357e-01 -6.56907141e-01
6.28456712e-01 1.12238243e-01 5.00822902e-01 -5.22496283e-01
-8.75097066e-02 -6.20436311e-01 4.87972498e-01 -5.38203835e-01
-1.36735318e-02 -2.88314879e-01 -1.12006640e+00 -5.86743593e-01
-9.53056142e-02 -1.15758635e-01 8.37971449e-01 4.58640218e-01
4.36020553e-01 1.03900218e+00 4.21230227e-01 -7.98905790e-01
6.07539080e-02 -8.36206615e-01 -6.17383540e-01 4.99528795e-01
2.39861220e-01 -7.61702955e-01 -1.45461693e-01 -2.11681455e-01] | [8.373861312866211, -1.8904662132263184] |
f9262c39-d1d5-47d0-8429-cfcc741aa736 | black-box-adversarial-attacks-in-autonomous | 2101.06092 | null | https://arxiv.org/abs/2101.06092v1 | https://arxiv.org/pdf/2101.06092v1.pdf | Black-box Adversarial Attacks in Autonomous Vehicle Technology | Despite the high quality performance of the deep neural network in real-world applications, they are susceptible to minor perturbations of adversarial attacks. This is mostly undetectable to human vision. The impact of such attacks has become extremely detrimental in autonomous vehicles with real-time "safety" concerns. The black-box adversarial attacks cause drastic misclassification in critical scene elements such as road signs and traffic lights leading the autonomous vehicle to crash into other vehicles or pedestrians. In this paper, we propose a novel query-based attack method called Modified Simple black-box attack (M-SimBA) to overcome the use of a white-box source in transfer based attack method. Also, the issue of late convergence in a Simple black-box attack (SimBA) is addressed by minimizing the loss of the most confused class which is the incorrect class predicted by the model with the highest probability, instead of trying to maximize the loss of the correct class. We evaluate the performance of the proposed approach to the German Traffic Sign Recognition Benchmark (GTSRB) dataset. We show that the proposed model outperforms the existing models like Transfer-based projected gradient descent (T-PGD), SimBA in terms of convergence time, flattening the distribution of confused class probability, and producing adversarial samples with least confidence on the true class. | ['C Krishna Mohan', 'Reshmi Mitra', 'C Vishnu', 'K Naveen Kumar'] | 2021-01-15 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 3.36887762e-02 1.81001499e-01 3.62858742e-01 -2.72743613e-01
-6.36953592e-01 -6.88779473e-01 7.80847669e-01 -3.78318787e-01
-6.96826756e-01 9.66646791e-01 -5.14868736e-01 -7.27483094e-01
1.70865059e-01 -8.84934366e-01 -1.16893935e+00 -9.62480843e-01
1.59766197e-01 2.84703314e-01 8.63532782e-01 -3.27662051e-01
1.09334230e-01 7.87420332e-01 -1.33760309e+00 1.71279743e-01
9.15994525e-01 9.84600902e-01 -4.07540917e-01 6.89311683e-01
2.18519747e-01 7.86504805e-01 -9.36949670e-01 -8.66904795e-01
7.64626622e-01 -3.55047196e-01 -2.42614120e-01 -4.31499779e-01
9.27642941e-01 -6.37426913e-01 -5.64962506e-01 1.50931382e+00
4.95977581e-01 7.60509148e-02 6.13223672e-01 -1.92349684e+00
-7.01467395e-02 -4.16080132e-02 -3.98927659e-01 -7.85515737e-03
-3.20143104e-01 6.76057756e-01 2.01306075e-01 -5.92909992e-01
5.92298269e-01 1.43600798e+00 6.59524560e-01 9.28462684e-01
-8.82095933e-01 -1.08681095e+00 -1.15564413e-01 6.90370381e-01
-1.20031250e+00 -4.44061071e-01 5.12851834e-01 -3.45692128e-01
5.63197017e-01 3.33231211e-01 1.77377909e-01 1.13339257e+00
3.38725239e-01 5.83007216e-01 1.06055725e+00 -1.45065546e-01
2.87909955e-01 3.79715353e-01 1.20767402e-02 6.36064887e-01
3.80890697e-01 8.04081082e-01 -8.80230442e-02 -2.61278361e-01
-8.32657963e-02 -2.81393766e-01 -1.22193776e-01 -4.13619190e-01
-6.17831707e-01 8.17359090e-01 6.17313683e-01 -3.40549588e-01
-4.09898996e-01 3.30971837e-01 4.64954585e-01 3.22310627e-01
1.09028339e-01 -2.50758141e-01 -2.94356674e-01 1.65646791e-01
-6.36829972e-01 4.57271188e-01 7.30726719e-01 6.42569900e-01
5.94808221e-01 5.78577220e-01 -2.22638130e-01 3.59602153e-01
4.15310562e-01 1.03484261e+00 9.72891673e-02 -6.00481391e-01
5.84051609e-01 2.69768119e-01 2.89171964e-01 -9.20919716e-01
-1.43846581e-02 -2.04246998e-01 -5.51032782e-01 1.33852184e+00
7.55916595e-01 -5.59730053e-01 -1.23713303e+00 1.74681723e+00
6.13117278e-01 3.68386328e-01 2.66057670e-01 8.99694979e-01
4.46020931e-01 7.71594763e-01 2.43101195e-01 2.47714594e-01
8.57791245e-01 -7.03777611e-01 -4.43926573e-01 -2.63549298e-01
4.57559109e-01 -6.55098736e-01 5.63627064e-01 2.79011816e-01
-5.95571041e-01 -2.80219495e-01 -1.26518357e+00 5.53626716e-01
-5.38255394e-01 -2.59329230e-01 8.18594992e-02 1.40025055e+00
-6.85110211e-01 4.45541829e-01 -5.87939382e-01 -2.46185079e-01
7.92138636e-01 4.28906620e-01 -4.99605864e-01 -7.43334591e-02
-1.32436407e+00 1.29428697e+00 2.51191735e-01 1.43357635e-01
-1.37030876e+00 -7.67510295e-01 -5.98501086e-01 -2.86719412e-01
3.80879551e-01 -2.43595839e-01 9.38386500e-01 -1.16837621e+00
-1.52531791e+00 6.70346677e-01 2.41552219e-02 -1.03506672e+00
1.18151402e+00 -2.08250612e-01 -4.43585545e-01 -9.11216512e-02
-1.57466084e-01 7.83787906e-01 1.11187649e+00 -1.28715086e+00
-7.85682142e-01 -3.76989275e-01 -7.56973075e-03 -8.21368173e-02
4.22262363e-02 1.64259493e-01 1.53621301e-01 -3.14641297e-01
-5.58963537e-01 -1.13765311e+00 -9.81829762e-02 4.37782049e-01
-4.34171706e-01 4.22506109e-02 1.66138589e+00 -5.70757031e-01
6.34913266e-01 -2.33951473e+00 -5.90162873e-01 1.82733864e-01
-1.89325288e-01 1.08955216e+00 -1.24033950e-01 1.89652845e-01
-7.41445571e-02 -6.37336075e-02 -3.37450922e-01 4.72690687e-02
-4.18254100e-02 4.23376739e-01 -8.16453278e-01 8.32805574e-01
2.61088431e-01 7.36987293e-01 -7.87379563e-01 -8.01409706e-02
3.78664404e-01 4.85713303e-01 -3.79998118e-01 1.05113409e-01
5.85922152e-02 3.06327194e-01 -2.03608632e-01 4.11761045e-01
1.02930605e+00 8.91638994e-01 -5.67479730e-01 -2.49969326e-02
7.87326992e-02 -2.80213300e-02 -1.10807586e+00 4.25673574e-01
-1.82034627e-01 1.03149211e+00 3.26314941e-02 -8.80249739e-01
9.57472146e-01 2.02958718e-01 -6.17312305e-02 -6.95614159e-01
2.11187556e-01 3.28553170e-01 3.65307450e-01 -2.66948014e-01
2.06149861e-01 -1.27790809e-01 1.59378648e-01 5.40471599e-02
-2.77755439e-01 1.69148743e-01 -2.25323156e-01 2.77277023e-01
9.41054285e-01 5.33124246e-02 -1.08663149e-01 -4.06015888e-02
6.27295971e-01 1.41265899e-01 7.03934193e-01 8.30764890e-01
-8.17690253e-01 3.49687189e-01 5.09169161e-01 -4.02347833e-01
-1.18608332e+00 -1.18155634e+00 -7.83830360e-02 6.59531772e-01
2.89699018e-01 4.57566261e-01 -8.88008654e-01 -1.09637630e+00
2.30069324e-01 1.19250333e+00 -5.54649770e-01 -6.64034784e-01
-6.24368489e-01 -6.24250412e-01 1.30893469e+00 2.36856714e-01
8.52989256e-01 -1.03982353e+00 -6.67419553e-01 -1.10734299e-01
1.99146345e-01 -1.13282669e+00 -3.41635346e-01 -1.38921916e-01
-3.71802062e-01 -1.25213671e+00 -5.62361836e-01 -3.78450096e-01
8.06304455e-01 2.65941136e-02 4.96551454e-01 -5.51025569e-02
-3.54703546e-01 1.41926995e-03 -2.31811851e-01 -9.11180794e-01
-9.09757257e-01 -7.77336419e-01 1.38438433e-01 5.31064868e-01
3.55996311e-01 -8.73627663e-02 -5.31030357e-01 6.49827003e-01
-7.71936059e-01 -5.10304987e-01 1.86235815e-01 8.34084749e-01
5.40968664e-02 1.62939522e-02 5.70917845e-01 -6.87481284e-01
3.94839585e-01 -3.79467279e-01 -1.05662847e+00 -5.77380974e-03
-4.05114651e-01 -1.45231664e-01 7.59491086e-01 -6.54167473e-01
-9.67802584e-01 -3.68740968e-02 -9.38657746e-02 -6.90288961e-01
-3.14769030e-01 -2.80344278e-01 -3.60277951e-01 -6.50639594e-01
8.78206968e-01 3.31714123e-01 1.93960175e-01 -1.28385667e-02
2.13881224e-01 6.28844321e-01 6.39850020e-01 -1.37610495e-01
1.27735412e+00 7.00665832e-01 4.08873141e-01 -8.86627674e-01
-2.84623563e-01 1.17081553e-02 -1.02374457e-01 -6.72665477e-01
7.13641047e-01 -5.18741071e-01 -7.20681071e-01 1.11025429e+00
-1.22298765e+00 -2.79038846e-01 -1.88157052e-01 3.78299654e-01
-3.27497542e-01 5.40153742e-01 -1.04915977e-01 -9.91913319e-01
-1.00626983e-01 -1.25004601e+00 5.78062832e-01 3.31281871e-01
2.04004958e-01 -5.53077519e-01 -1.86260611e-01 3.02956730e-01
4.88203019e-01 5.80708206e-01 7.00490296e-01 -9.35040951e-01
-6.23112202e-01 -7.63515294e-01 -2.61755735e-01 7.71852672e-01
-1.61554039e-01 1.28990024e-01 -1.06456840e+00 -3.78060251e-01
-5.38116395e-02 -2.34985486e-01 8.46909881e-01 2.59875029e-01
5.70763052e-01 -4.16923821e-01 -2.63236552e-01 6.73942983e-01
1.32327795e+00 5.56566417e-01 1.04949307e+00 4.74287003e-01
5.58588088e-01 6.13448977e-01 6.89340651e-01 1.13693597e-02
-4.86342087e-02 7.11643040e-01 9.34044302e-01 5.06878123e-02
-2.47073844e-01 -2.26195216e-01 7.22954869e-01 -4.24909830e-01
3.08338642e-01 -4.42648917e-01 -8.02185833e-01 5.00727594e-01
-1.79353714e+00 -1.17582846e+00 -4.19999272e-01 2.57477474e+00
3.45061362e-01 4.58033562e-01 1.11708947e-01 1.78330734e-01
9.25587237e-01 -1.02591038e-01 -6.34490848e-01 -8.14417481e-01
-2.01456428e-01 -1.24196492e-01 1.04502273e+00 7.50710011e-01
-1.23131394e+00 1.05105937e+00 5.50794029e+00 1.06582880e+00
-1.27564836e+00 2.16456771e-01 5.67332029e-01 -7.32183978e-02
2.71293581e-01 7.66879767e-02 -9.45772350e-01 7.04185069e-01
8.74325752e-01 -6.58238381e-02 2.74511784e-01 9.13484812e-01
3.12399089e-01 -1.02969155e-01 -7.87464321e-01 5.48667490e-01
3.40333283e-02 -1.03834796e+00 1.44579709e-01 6.36660531e-02
5.62877655e-01 1.46436796e-01 1.99638590e-01 3.11291724e-01
6.43076301e-01 -8.35700929e-01 8.63030374e-01 2.40645334e-01
5.77092171e-01 -1.00201273e+00 7.41070688e-01 4.27614868e-01
-5.87354183e-01 -1.92202151e-01 -4.49468613e-01 2.87827253e-01
1.58993393e-01 1.18824206e-01 -9.09688294e-01 1.82947800e-01
5.23071051e-01 -9.03790966e-02 -3.50958288e-01 1.15502489e+00
-3.97500515e-01 8.91901255e-01 -5.60957849e-01 -1.22098766e-01
6.38892472e-01 5.78738488e-02 1.31377780e+00 1.05407810e+00
-6.03485405e-02 -2.43210644e-01 -2.55389392e-01 6.90206170e-01
1.64150387e-01 -2.74231344e-01 -8.60672235e-01 4.23911005e-01
3.35298300e-01 8.57395589e-01 -3.28502059e-01 -3.04977268e-01
-1.12823687e-01 8.65209818e-01 -1.19488060e-01 5.10288179e-01
-1.17679656e+00 -6.49863303e-01 7.80222476e-01 3.15574333e-02
4.58105624e-01 1.21881464e-03 -2.70544708e-01 -5.35616279e-01
2.29357675e-01 -9.04775620e-01 2.49035075e-01 -5.75418711e-01
-1.04727018e+00 6.28850579e-01 -4.78177667e-02 -1.60542333e+00
-1.81872323e-01 -7.99328864e-01 -9.22251165e-01 1.04364657e+00
-1.38358641e+00 -1.02508819e+00 2.36370973e-02 8.76801014e-01
4.16779965e-01 -6.69327915e-01 3.36286664e-01 5.25920153e-01
-3.64154220e-01 1.09269524e+00 1.85073078e-01 2.52588093e-01
7.13468373e-01 -7.91233599e-01 4.52067375e-01 1.34048724e+00
-2.33811200e-01 -5.76028414e-02 8.95804167e-01 -7.73055315e-01
-9.41741168e-01 -1.39497197e+00 6.14835382e-01 -4.50580925e-01
5.39151371e-01 -1.03232488e-01 -6.53394341e-01 3.45450789e-01
-9.54388082e-02 2.15990677e-01 -6.57220483e-02 -6.74130440e-01
-6.49626553e-01 -4.17369485e-01 -1.76686859e+00 7.97939003e-01
3.84256244e-01 -2.78216481e-01 -3.66009563e-01 1.79686800e-01
4.45916653e-01 -1.25558212e-01 1.20818779e-01 4.45023566e-01
4.67267334e-01 -1.01951766e+00 8.70484114e-01 -8.84470344e-01
2.57181190e-02 -5.39231479e-01 -9.79963616e-02 -1.24962556e+00
2.26059049e-01 -6.49302483e-01 1.16116874e-01 9.33172286e-01
4.02917325e-01 -1.04074073e+00 8.25548708e-01 6.56700552e-01
-5.18559851e-02 -4.04294223e-01 -1.51215327e+00 -1.05784738e+00
1.81687176e-01 -4.50977415e-01 2.66170412e-01 6.28937781e-01
-6.45701170e-01 -3.53572518e-01 -6.24720097e-01 7.16229975e-01
1.03697371e+00 -6.26765668e-01 1.12629759e+00 -7.01251864e-01
2.54417658e-02 -3.79001290e-01 -1.09914958e+00 -5.07707953e-01
-9.29720178e-02 -4.54927653e-01 2.48788640e-01 -8.63467157e-01
-2.75541276e-01 -2.91351676e-01 -2.40665630e-01 3.23533922e-01
-2.29076415e-01 3.55312496e-01 4.50744390e-01 -6.76863492e-02
-1.15542881e-01 3.86975855e-01 9.36069131e-01 -3.72490078e-01
1.62144825e-01 4.28227574e-01 -3.77029814e-02 7.63419747e-01
8.10849428e-01 -1.08693767e+00 -2.56853044e-01 -1.90595016e-01
-1.70652881e-01 -1.64212331e-01 9.19350445e-01 -1.13760519e+00
3.88504714e-01 -1.43925488e-01 5.27182966e-02 -5.25916755e-01
3.29533905e-01 -1.07225788e+00 3.06997225e-02 9.43400562e-01
-4.08894904e-02 -3.94246042e-01 5.17084897e-01 6.73872411e-01
-5.19671030e-02 -3.41742337e-01 1.35582590e+00 2.37202957e-01
-5.18161416e-01 2.18625605e-01 -5.28268576e-01 -1.47915885e-01
1.59024143e+00 -5.53001106e-01 -4.57154781e-01 -4.95878220e-01
-2.47591212e-01 2.94920027e-01 1.58128753e-01 6.49154782e-01
5.62215269e-01 -1.14374352e+00 -1.02063394e+00 2.66545117e-01
-1.42437175e-01 -3.44873250e-01 2.16581151e-01 4.22549337e-01
-7.54399002e-01 2.37741128e-01 -3.58808696e-01 -4.05349076e-01
-1.48084557e+00 5.04703462e-01 8.30381155e-01 1.32910103e-01
-5.19491494e-01 8.16273034e-01 2.50018626e-01 -3.99223715e-01
4.54644620e-01 3.04556549e-01 8.88207275e-03 -3.69405419e-01
4.47581768e-01 6.61855340e-01 1.77325066e-02 -9.27918494e-01
-4.46565658e-01 1.95774421e-01 -1.12306610e-01 -2.17263147e-01
8.57147217e-01 2.24196091e-01 1.48860008e-01 -1.78936556e-01
1.01761115e+00 -4.20802683e-02 -1.51336074e+00 6.32694811e-02
-3.26615393e-01 -6.31855428e-01 8.58998075e-02 -9.68436360e-01
-1.12728405e+00 9.74151731e-01 1.20790553e+00 -7.94512928e-02
7.68620074e-01 -6.28283620e-01 9.00197685e-01 6.21614695e-01
3.38737875e-01 -8.78258944e-01 -3.54883283e-01 4.12865937e-01
7.24871635e-01 -1.27192211e+00 -4.10702765e-01 -2.25969329e-02
-8.47403705e-01 7.57275105e-01 8.08230519e-01 -4.67332035e-01
7.13285387e-01 1.32478267e-01 4.97347951e-01 2.94524580e-01
-4.99163926e-01 2.03297913e-01 1.04727544e-01 9.70034957e-01
-5.88705420e-01 -3.21067357e-03 -2.05210894e-01 -9.79180913e-03
-8.62078567e-04 -2.07769945e-01 5.32207191e-01 7.67400384e-01
-4.56077754e-01 -8.77865791e-01 -7.55498648e-01 1.96254551e-01
-4.63600397e-01 5.99122085e-02 -4.81826335e-01 7.58543730e-01
2.34070256e-01 9.91436720e-01 -1.22029237e-01 -3.75673383e-01
3.63254726e-01 7.67280236e-02 7.07620159e-02 9.15136337e-02
-5.34209192e-01 -6.32061779e-01 1.71076372e-01 -6.42997205e-01
1.05043113e-01 -4.25047249e-01 -9.12174761e-01 -4.88892525e-01
-3.65545869e-01 -1.38608947e-01 8.85363460e-01 9.97419775e-01
1.58492178e-01 1.89206973e-01 8.62555027e-01 -6.89238012e-01
-9.73641515e-01 -6.45078540e-01 -1.19316377e-01 4.46288317e-01
6.08208776e-01 -6.03800535e-01 -5.93333960e-01 -1.81587771e-01] | [5.414913654327393, 7.826550006866455] |
5bd92525-92ef-4170-8889-ab4b28e3ca3c | topic-propagation-in-conversational-search | 2004.14054 | null | https://arxiv.org/abs/2004.14054v1 | https://arxiv.org/pdf/2004.14054v1.pdf | Topic Propagation in Conversational Search | In a conversational context, a user expresses her multi-faceted information need as a sequence of natural-language questions, i.e., utterances. Starting from a given topic, the conversation evolves through user utterances and system replies. The retrieval of documents relevant to a given utterance in a conversation is challenging due to ambiguity of natural language and to the difficulty of detecting possible topic shifts and semantic relationships among utterances. We adopt the 2019 TREC Conversational Assistant Track (CAsT) framework to experiment with a modular architecture performing: (i) topic-aware utterance rewriting, (ii) retrieval of candidate passages for the rewritten utterances, and (iii) neural-based re-ranking of candidate passages. We present a comprehensive experimental evaluation of the architecture assessed in terms of traditional IR metrics at small cutoffs. Experimental results show the effectiveness of our techniques that achieve an improvement up to 0.28 (+93%) for P@1 and 0.19 (+89.9%) for nDCG@3 w.r.t. the CAsT baseline. | ['C. I. Muntean', 'R. Perego', 'I. Mele', 'F. M. Nardini', 'O. Frieder', 'N. Tonellotto'] | 2020-04-29 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [ 5.60214281e-01 5.17497003e-01 8.99937302e-02 -7.03310311e-01
-1.39777911e+00 -8.48329484e-01 1.16814780e+00 3.21804762e-01
-4.94942516e-01 6.80061996e-01 8.94689202e-01 -1.66922241e-01
-8.24555606e-02 -2.42788255e-01 -3.58467877e-01 -3.18895340e-01
-1.11331113e-01 1.08081651e+00 3.91433477e-01 -5.99080384e-01
4.92877930e-01 9.65628326e-02 -1.48198819e+00 8.85808349e-01
7.52798676e-01 7.31669068e-01 2.83058673e-01 1.10479748e+00
-3.34426045e-01 6.94343090e-01 -9.52946067e-01 -5.51430762e-01
-2.87959993e-01 -6.02528632e-01 -1.48159552e+00 1.31831035e-01
3.09520811e-01 -2.70776749e-01 -1.29687965e-01 7.21872211e-01
3.45933259e-01 5.18001795e-01 6.38038993e-01 -9.66366649e-01
-4.39399540e-01 6.43300056e-01 -1.83198035e-01 2.20574290e-01
9.14578497e-01 -2.72835612e-01 1.20855069e+00 -9.42030430e-01
7.85444617e-01 1.61138713e+00 7.62625039e-02 7.57311046e-01
-8.91844571e-01 -4.80022691e-02 2.62931675e-01 1.32657155e-01
-1.03640723e+00 -7.32251942e-01 3.59770745e-01 -7.38291070e-02
1.33635521e+00 6.56225860e-01 -3.42851132e-02 9.12668884e-01
9.84627828e-02 7.45282292e-01 5.80132961e-01 -5.85253894e-01
2.24741623e-01 4.77280021e-01 6.88255191e-01 2.59102881e-01
-4.61208642e-01 -4.27749068e-01 -6.14845395e-01 -6.24033868e-01
-2.06832677e-01 -3.63025934e-01 -3.16012472e-01 1.52085587e-01
-8.54346216e-01 8.77631843e-01 1.56384513e-01 1.95467547e-01
-4.46735740e-01 -2.14017332e-01 6.37684464e-01 6.94274426e-01
5.85062385e-01 9.44583893e-01 -3.88610810e-01 -3.81576538e-01
-6.50290489e-01 8.74032974e-01 1.31633198e+00 1.20033121e+00
4.70129788e-01 -7.55224645e-01 -5.94682395e-01 1.25477946e+00
2.26351246e-01 3.62603188e-01 4.37657654e-01 -1.35799253e+00
5.43471634e-01 4.40392196e-01 4.44126606e-01 -7.78138876e-01
-3.72340202e-01 -6.32095188e-02 -1.57045215e-01 -6.12778485e-01
2.59101301e-01 -2.14554816e-01 -5.51002324e-01 1.72882187e+00
3.09896886e-01 -5.29103816e-01 4.43894595e-01 4.66504991e-01
1.06988084e+00 1.04154313e+00 -3.59826013e-02 -6.88920975e-01
1.78903461e+00 -1.08826232e+00 -8.63838792e-01 -2.52004713e-01
5.77164590e-01 -1.01078558e+00 1.06327045e+00 2.84627080e-02
-1.37979925e+00 -1.18847527e-01 -6.33477092e-01 -2.36676738e-01
-1.98741361e-01 -3.39170098e-01 9.17463228e-02 4.41396356e-01
-1.28634429e+00 1.30165234e-01 -3.85409504e-01 -6.19926929e-01
-4.98211354e-01 2.41540730e-01 1.23382762e-01 -2.22590327e-01
-1.41304600e+00 9.39381540e-01 2.65226699e-02 -1.02550879e-01
-5.36844254e-01 -5.24145246e-01 -5.61614692e-01 2.38817558e-01
4.67494786e-01 -4.24146235e-01 2.20106411e+00 -5.52241445e-01
-1.56399012e+00 6.51583970e-01 -7.37643421e-01 -5.12409925e-01
1.91959962e-01 -3.87402654e-01 -3.16343516e-01 4.53398883e-01
2.20645294e-01 6.70255303e-01 3.60582709e-01 -1.21532130e+00
-9.62615728e-01 -2.30192304e-01 4.58110243e-01 8.12572837e-01
-1.61711633e-01 5.49983084e-01 -7.43162632e-01 -1.47565931e-01
9.79186371e-02 -1.14173949e+00 -1.37093797e-01 -7.06716537e-01
-2.13886246e-01 -8.60985935e-01 8.62790465e-01 -7.47100532e-01
1.21411502e+00 -1.76160419e+00 -9.83128697e-02 6.94266036e-02
1.23843506e-01 -2.88538709e-02 -3.16369653e-01 8.52509856e-01
3.88919324e-01 1.08508803e-01 -2.63711382e-02 -3.38173091e-01
3.94319184e-02 -1.54591888e-01 -4.54242975e-01 -8.47022608e-02
-2.45032571e-02 7.61790752e-01 -1.02420545e+00 -3.71559173e-01
-1.95046455e-01 1.86550647e-01 -6.47281766e-01 4.73348588e-01
-6.25414014e-01 1.09613471e-01 -5.21437705e-01 2.82506555e-01
1.50980175e-01 -2.53513575e-01 3.61059308e-01 1.46806851e-01
5.96496798e-02 1.01684403e+00 -4.35503483e-01 1.58024800e+00
-4.50158328e-01 7.69982278e-01 7.44118616e-02 -4.43732172e-01
8.09724867e-01 7.36302972e-01 2.62929231e-01 -7.47879684e-01
-1.68243665e-02 2.06974261e-02 -7.04546645e-02 -4.58621740e-01
1.27160978e+00 1.44351766e-01 -5.00675797e-01 1.00739217e+00
-1.32033834e-02 -2.20289931e-01 3.31732213e-01 6.93105102e-01
1.29781401e+00 -5.28808057e-01 2.08554134e-01 -3.45220059e-01
5.45466542e-01 1.94068328e-02 -1.22231347e-02 1.08464003e+00
-7.34869093e-02 4.21644151e-01 6.11463726e-01 -2.29207724e-01
-9.36909437e-01 -7.59360313e-01 2.56264862e-02 1.64107406e+00
4.57047038e-02 -3.13873827e-01 -8.94867778e-01 -5.30775130e-01
-3.24805886e-01 1.16178513e+00 -4.27999228e-01 -1.87140420e-01
-8.04324746e-01 -3.83705378e-01 4.99336869e-01 6.58271983e-02
2.97861814e-01 -1.33895969e+00 -5.03196359e-01 2.63643205e-01
-8.13297331e-01 -1.29282296e+00 -7.59312391e-01 -8.57772380e-02
-6.47237659e-01 -6.79735243e-01 -7.54545271e-01 -8.62038553e-01
3.97475004e-01 5.62556624e-01 1.18566120e+00 9.00143757e-03
2.48672232e-01 5.95624149e-01 -6.72176242e-01 -3.14311713e-01
-8.13454330e-01 2.11844698e-01 -1.46158382e-01 -2.86904901e-01
6.28190756e-01 -1.58888087e-01 -5.29191494e-01 5.60811818e-01
-7.15480089e-01 -2.22906321e-01 2.98908740e-01 7.66969621e-01
-6.48877397e-02 -4.58056152e-01 9.31214273e-01 -1.03225398e+00
1.55997741e+00 -4.86287057e-01 -2.15592578e-01 6.64068997e-01
-7.16996312e-01 -4.99151461e-02 1.46191567e-01 -2.26771653e-01
-1.57336843e+00 -3.06782812e-01 -1.42750945e-02 3.12893927e-01
-1.66066155e-01 5.98178267e-01 2.66102195e-01 4.90172684e-01
9.26275909e-01 1.32881373e-01 -7.98848271e-02 -2.67971933e-01
5.74354172e-01 1.19171703e+00 6.37830615e-01 -7.36984611e-01
1.99490383e-01 -6.74913898e-02 -8.50457549e-01 -9.55304801e-01
-8.86736214e-01 -9.48131621e-01 -2.85805464e-01 -4.70966965e-01
6.55459583e-01 -7.89597631e-01 -8.60055625e-01 -1.11601010e-01
-1.52889967e+00 -7.66786486e-02 -3.04163471e-02 1.59537494e-01
-4.54111338e-01 4.57043618e-01 -7.30840981e-01 -9.96715069e-01
-8.11483800e-01 -1.12214363e+00 1.08677459e+00 3.26269895e-01
-9.16108727e-01 -6.96429431e-01 2.13535741e-01 8.04382682e-01
5.36166131e-01 -2.74316162e-01 1.06105387e+00 -1.19088566e+00
-1.95388094e-01 -4.14408922e-01 -1.77367732e-01 -1.11263379e-01
7.25849643e-02 -2.88375735e-01 -1.09427357e+00 -4.17429268e-01
1.21413350e-01 -5.60262680e-01 4.33015257e-01 1.33332610e-01
6.70534968e-01 -6.67384624e-01 -3.06442618e-01 -4.36891198e-01
6.53234243e-01 7.56411076e-01 4.90130037e-01 1.03981070e-01
-1.26781970e-01 1.10009420e+00 5.76657057e-01 3.76239955e-01
4.52077836e-01 7.97147810e-01 -1.17508285e-01 3.71266514e-01
1.69394463e-01 -2.65211742e-02 2.76884228e-01 7.38811374e-01
2.18566418e-01 -6.34074271e-01 -8.72206211e-01 6.53494775e-01
-1.83708608e+00 -8.62883031e-01 1.75831035e-01 2.04475331e+00
1.10727024e+00 2.08765894e-01 8.77680108e-02 -4.99570966e-01
7.66285658e-01 1.63016394e-01 -5.82640231e-01 -8.21341038e-01
1.91594347e-01 -5.15247770e-02 -1.07303478e-01 7.99908400e-01
-6.77856863e-01 9.28734839e-01 6.44158745e+00 2.64055789e-01
-6.87550604e-01 6.03647232e-02 6.47169530e-01 -2.04430506e-01
-3.79745126e-01 -1.95568353e-01 -9.12622988e-01 1.40261143e-01
1.40943420e+00 -8.12104523e-01 4.31956500e-01 7.15743244e-01
1.80162996e-01 -8.98399204e-02 -1.24574113e+00 4.41664726e-01
4.10163075e-01 -1.30378544e+00 7.49079287e-02 -2.04066709e-01
7.09316611e-01 9.57724452e-02 -2.29791030e-01 7.41836607e-01
5.44609308e-01 -7.26765990e-01 3.39219183e-01 2.11053237e-01
3.59354317e-01 -7.26369381e-01 8.55193019e-01 4.20281261e-01
-6.80504858e-01 1.29270434e-01 -4.05528992e-01 1.11945927e-01
2.95780510e-01 6.51663095e-02 -1.67801273e+00 3.37730587e-01
6.51083112e-01 -1.20950220e-02 -1.68755367e-01 6.80043280e-01
4.48464938e-02 4.33351338e-01 -2.68778890e-01 -6.61281109e-01
5.52447021e-01 1.44402310e-01 6.55155122e-01 1.51739657e+00
-1.20928586e-02 5.85639894e-01 2.15739295e-01 4.74040657e-01
-4.16254461e-01 2.80824751e-01 -4.38475013e-01 8.69987011e-02
8.28627229e-01 1.02772975e+00 -4.55346972e-01 -5.36601186e-01
-2.57868409e-01 9.12490070e-01 2.38096640e-01 5.60692668e-01
-2.86163568e-01 -4.97558951e-01 4.40660268e-01 -2.59280980e-01
1.42671224e-02 2.21640557e-01 1.08437009e-01 -6.77367747e-01
1.87201157e-01 -1.12336600e+00 5.30416310e-01 -6.17554188e-01
-1.11328888e+00 1.03073847e+00 2.83182353e-01 -7.81159222e-01
-1.05373991e+00 3.59700955e-02 -6.65781677e-01 8.97867441e-01
-1.11505461e+00 -4.37270135e-01 -9.21041965e-02 3.24222326e-01
1.26936114e+00 -7.52799213e-02 1.09890473e+00 4.94389571e-02
-1.34173751e-01 6.81384385e-01 2.22915232e-01 -1.89839318e-01
8.28383327e-01 -1.15008199e+00 6.08442008e-01 5.08719683e-01
-2.20525190e-01 9.21074510e-01 1.00757027e+00 -5.17425299e-01
-1.17850053e+00 -8.79686356e-01 1.65227664e+00 -5.04642904e-01
4.67628449e-01 -3.97453129e-01 -1.12963462e+00 6.24653041e-01
7.08256602e-01 -9.48865771e-01 6.80363715e-01 5.34858763e-01
-2.46355921e-01 1.18835740e-01 -1.07910335e+00 8.70289326e-01
5.32864571e-01 -6.87036753e-01 -9.43480611e-01 8.00728381e-01
1.38712943e+00 -5.47059000e-01 -7.38051832e-01 1.54006854e-01
7.08246768e-01 -5.03388882e-01 7.51026332e-01 -7.69336641e-01
4.24850136e-01 2.75948316e-01 -2.19944403e-01 -1.13488269e+00
-1.12567425e-01 -9.67103064e-01 5.60737997e-02 1.04640889e+00
8.51810515e-01 -3.14178884e-01 6.44171596e-01 1.31168807e+00
-3.66838276e-01 -6.18215203e-01 -8.98591638e-01 -3.74722600e-01
4.63083573e-03 -1.66780606e-01 4.49821472e-01 6.17130637e-01
7.18806803e-01 9.77006853e-01 -2.83292770e-01 4.50928919e-02
1.05646372e-01 1.36644036e-01 7.53784478e-01 -9.60674107e-01
-5.10895029e-02 -4.53752875e-01 3.15347135e-01 -1.57758689e+00
1.21646049e-02 -5.46845675e-01 6.20273411e-01 -1.51894367e+00
3.17056835e-01 -4.96540144e-02 9.71083529e-03 6.22835830e-02
-1.94283932e-01 -3.61586571e-01 1.30895227e-01 3.41419667e-01
-9.80820060e-01 5.18179357e-01 8.39064658e-01 -2.00900093e-01
-5.15030026e-01 2.73313642e-01 -8.09609830e-01 3.67159247e-01
7.62634933e-01 -3.53453189e-01 -6.00328624e-01 -3.60056788e-01
-3.92427221e-02 6.93928242e-01 -2.48607546e-01 -3.56016815e-01
5.11568010e-01 -6.83954880e-02 -3.21750134e-01 -8.51622760e-01
5.38746476e-01 -3.10519278e-01 -4.33178961e-01 2.62123078e-01
-1.27846122e+00 2.97106802e-01 9.52228457e-02 5.41723430e-01
-2.83710301e-01 -3.66681963e-01 4.02802080e-01 -1.83682457e-01
-4.52816486e-01 -2.74691522e-01 -7.49482870e-01 2.89567381e-01
5.13442278e-01 2.09814146e-01 -6.99816465e-01 -1.10018039e+00
-5.56869090e-01 6.65788352e-01 -1.83462184e-02 7.55544364e-01
6.32562041e-01 -8.81918371e-01 -9.35400784e-01 -4.44124132e-01
4.08818334e-01 -1.60588861e-01 1.27491117e-01 3.78947526e-01
-1.46245137e-01 1.01886165e+00 2.79468209e-01 -5.34939349e-01
-1.67335451e+00 7.30066001e-02 1.39272556e-01 -5.32064438e-01
-4.59364086e-01 8.86263072e-01 2.93324798e-01 -6.50866866e-01
6.90185845e-01 -1.18717760e-01 -4.38264191e-01 8.83491635e-02
9.55031991e-01 2.08924010e-01 4.05622363e-01 -3.73705894e-01
-1.97748840e-01 -1.66679263e-01 -7.61235297e-01 -6.22430682e-01
9.29556429e-01 -4.96985435e-01 -1.53312504e-01 4.32251036e-01
1.40841913e+00 -3.08343381e-01 -5.28172135e-01 -8.56819630e-01
4.61890161e-01 -1.16508983e-01 -3.13155621e-01 -1.16803670e+00
-1.15783758e-01 3.72475892e-01 2.70695299e-01 6.04563236e-01
8.67875218e-01 3.97212654e-01 8.92039299e-01 1.27932739e+00
1.63819224e-01 -1.02372372e+00 2.28830993e-01 9.34447110e-01
1.38135445e+00 -9.48610365e-01 -1.12854727e-01 -2.76302636e-01
-8.36610734e-01 1.04659450e+00 4.07574743e-01 3.72598261e-01
3.10548574e-01 -1.92691222e-01 2.69909173e-01 -5.37660897e-01
-1.46544302e+00 1.46659672e-01 3.69750410e-01 8.49504247e-02
7.84679949e-01 2.24307869e-02 -4.71941411e-01 1.70275167e-01
-4.87093240e-01 -4.99506712e-01 5.95323920e-01 8.43758345e-01
-8.26104522e-01 -7.31646955e-01 -1.68126196e-01 5.11156440e-01
-3.74298126e-01 -1.12512484e-01 -8.78499687e-01 4.32571471e-01
-1.01775932e+00 1.62061930e+00 1.29225910e-01 -3.05181146e-01
5.66252232e-01 4.47025359e-01 -1.17954046e-01 -8.92156363e-01
-8.94891977e-01 2.14795128e-01 8.11479092e-01 -3.42342347e-01
-3.37529689e-01 -9.37830091e-01 -1.14422321e+00 -1.67832136e-01
-3.56223047e-01 8.66189003e-01 6.81833148e-01 9.80282545e-01
3.92523408e-01 3.19492370e-01 9.09219503e-01 -3.73778313e-01
-7.56778419e-01 -1.39832318e+00 -5.21131568e-02 3.92231405e-01
2.36524150e-01 -1.53390959e-01 -2.30158985e-01 4.26997580e-02] | [12.208694458007812, 7.881113529205322] |
7a0602a7-689b-40d5-acef-0a30b71de452 | unifying-clustered-and-non-stationary-bandits | 2009.02463 | null | https://arxiv.org/abs/2009.02463v1 | https://arxiv.org/pdf/2009.02463v1.pdf | Unifying Clustered and Non-stationary Bandits | Non-stationary bandits and online clustering of bandits lift the restrictive assumptions in contextual bandits and provide solutions to many important real-world scenarios. Though the essence in solving these two problems overlaps considerably, they have been studied independently. In this paper, we connect these two strands of bandit research under the notion of test of homogeneity, which seamlessly addresses change detection for non-stationary bandit and cluster identification for online clustering of bandit in a unified solution framework. Rigorous regret analysis and extensive empirical evaluations demonstrate the value of our proposed solution, especially its flexibility in handling various environment assumptions. | ['Chuanhao Li', 'Qingyun Wu', 'Hongning Wang'] | 2020-09-05 | null | null | null | null | ['online-clustering'] | ['computer-vision'] | [ 9.27764848e-02 -2.16825858e-01 -1.21694219e+00 -2.91012228e-01
-9.14316654e-01 -1.10114920e+00 2.23030910e-01 -3.33233625e-01
-2.12185644e-02 1.22418547e+00 2.03026682e-02 -9.02326405e-01
-9.45373416e-01 -4.81498927e-01 -9.90803003e-01 -9.22652185e-01
-1.07267387e-01 7.15290844e-01 5.74453957e-02 3.55833709e-01
2.19832100e-02 3.21543366e-01 -1.14128613e+00 8.57194960e-02
1.05674362e+00 1.23756373e+00 -1.83094114e-01 5.35803258e-01
-1.42031863e-01 4.78702158e-01 1.13482647e-01 -6.75961852e-01
6.13140523e-01 -5.29987514e-01 -9.87590551e-01 4.09712553e-01
2.21898034e-01 -3.54057342e-01 -3.18339884e-01 1.27810800e+00
2.17939302e-01 2.17930466e-01 1.71883404e-01 -1.24542844e+00
-6.74365580e-01 1.18002594e+00 -8.49375129e-01 5.91135323e-01
-5.30217737e-02 -1.55645162e-01 1.48168814e+00 2.01397195e-01
8.45842287e-02 1.11936438e+00 9.46554661e-01 9.66604874e-02
-1.60905993e+00 -5.92714429e-01 3.97686422e-01 3.44218284e-01
-8.42811227e-01 -3.98018271e-01 6.16109550e-01 -2.55475104e-01
4.04997885e-01 9.59339380e-01 1.28627670e+00 8.68284166e-01
-8.06250095e-01 1.51152635e+00 1.17550504e+00 -4.87968951e-01
4.73563015e-01 4.33746129e-02 7.16446698e-01 1.50131509e-01
6.94631636e-01 4.03226018e-01 -2.37961635e-01 -4.71541166e-01
1.03198576e+00 3.56951952e-01 -1.61963433e-01 -7.33887732e-01
-9.08613265e-01 1.00501657e+00 3.93321395e-01 3.60069200e-02
-5.03087044e-01 5.75259745e-01 4.30171102e-01 3.53160828e-01
9.56803620e-01 -5.40240258e-02 -4.18618143e-01 -1.79481626e-01
-1.10664558e+00 1.00877874e-01 7.22322524e-01 1.08380830e+00
6.57385051e-01 -9.47398171e-02 -3.08155090e-01 8.03009391e-01
-1.70635104e-01 4.58893210e-01 2.57997602e-01 -9.31603312e-01
5.64488590e-01 1.40033439e-01 5.97787023e-01 -6.10226870e-01
-2.54713506e-01 -9.44371164e-01 -5.41580677e-01 -8.14892292e-01
3.01708996e-01 -1.51610985e-01 -7.78208733e-01 1.52818608e+00
7.60562003e-01 7.79642820e-01 -3.89354706e-01 7.59629965e-01
-2.31479287e-01 5.77996135e-01 -3.52215856e-01 -6.82755351e-01
7.67347872e-01 -1.13214278e+00 -7.81910837e-01 -3.34379792e-01
4.50900197e-01 -3.55289817e-01 6.90624297e-01 3.63981910e-02
-1.28010547e+00 2.31469139e-01 -5.09534180e-01 4.10488218e-01
-1.31613892e-02 -5.86848259e-01 1.46269572e+00 1.29424059e+00
-1.15592074e+00 1.50067732e-01 -8.48788559e-01 -6.00966394e-01
5.92228651e-01 4.78358835e-01 2.54946053e-01 -1.49102733e-01
-7.50928879e-01 2.80843109e-01 4.18552727e-01 2.26218194e-01
-4.10655767e-01 -8.99382532e-01 -3.03230226e-01 1.77774444e-01
6.94294930e-01 -8.40779066e-01 1.71936131e+00 -1.54925334e+00
-1.48577976e+00 6.53869808e-01 -3.13243985e-01 -9.72353280e-01
9.68581676e-01 -1.08881174e-02 -1.96567595e-01 -2.74272293e-01
1.06549941e-01 -1.44384980e-01 4.45904911e-01 -9.98269856e-01
-1.13478947e+00 -5.13848484e-01 3.68029922e-02 1.23505645e-01
-2.58411407e-01 -1.53677598e-01 -5.79333365e-01 -6.23302758e-01
4.35234249e-01 -1.13102019e+00 -2.63195425e-01 -6.98852897e-01
-3.96042705e-01 1.83719724e-01 4.68668550e-01 -3.47770095e-01
1.29937971e+00 -1.99161577e+00 -3.47733617e-01 6.78929508e-01
-3.64580184e-01 -1.70126796e-01 8.69814586e-03 4.45771962e-01
-1.86639041e-01 3.52245122e-01 -1.68812826e-01 8.06467980e-02
3.42876911e-01 2.59690315e-01 -7.03724086e-01 9.43488002e-01
-7.24455833e-01 1.18062973e+00 -7.43472040e-01 -1.76588073e-01
1.81903362e-01 -4.72307771e-01 -9.02484357e-01 -3.60621542e-01
-2.69269049e-01 2.57206261e-01 -6.69495523e-01 7.20420480e-01
8.38809073e-01 -5.26223660e-01 5.17370343e-01 3.54834050e-01
-1.12723485e-01 2.04347253e-01 -1.17203796e+00 1.27727342e+00
-7.92394280e-02 3.52960676e-01 2.44335711e-01 -1.77006972e+00
6.32016882e-02 2.58598685e-01 8.24128985e-01 -7.14873135e-01
-2.26863772e-01 9.25990567e-02 -1.74977288e-01 -4.23383653e-01
2.41329923e-01 -3.28603774e-01 1.35951198e-03 6.72200620e-01
-3.83292288e-01 1.58027142e-01 1.06442608e-01 -2.60806382e-01
9.43333387e-01 -3.34793955e-01 4.25207585e-01 -4.04296309e-01
-1.28797516e-01 5.55605702e-02 6.17250681e-01 1.55641091e+00
-4.70408529e-01 -1.95730366e-02 3.12577873e-01 -6.35749221e-01
-8.77554536e-01 -1.04702795e+00 -3.06889355e-01 1.53393078e+00
2.95418739e-01 3.42150718e-01 -5.01767516e-01 -5.72356522e-01
5.80966473e-01 6.60762250e-01 -7.94555843e-01 2.43285879e-01
-2.06750497e-01 -1.10899365e+00 3.17816436e-01 1.94974855e-01
4.40275878e-01 -3.73391092e-01 -2.76958168e-01 3.44360799e-01
-4.49732184e-01 -8.69188845e-01 -3.91319573e-01 2.73606926e-01
-1.22999370e+00 -9.66106236e-01 -4.20758396e-01 -6.32900417e-01
1.65578008e-01 9.14896190e-01 9.24117327e-01 -1.79873794e-01
-2.54165027e-02 7.51924753e-01 -3.13242823e-01 -3.75504434e-01
1.86211899e-01 1.25912681e-01 -2.76895195e-01 3.86747867e-01
2.03430176e-01 -7.96555817e-01 -7.89199173e-01 4.86279309e-01
-1.02867413e+00 -6.40736595e-02 2.56922454e-01 8.38326514e-01
4.98672485e-01 3.41960229e-03 5.98781884e-01 -1.19289052e+00
4.72581685e-01 -6.53511345e-01 -1.00914240e+00 5.59701324e-01
-4.29970980e-01 -2.69578546e-01 -4.16511158e-03 -3.49796116e-01
-9.93768752e-01 -1.68479770e-01 4.47907329e-01 -2.39094570e-01
2.38244891e-01 7.75681436e-01 4.07478333e-01 3.56622711e-02
1.07035205e-01 2.86446452e-01 -3.61103892e-01 -3.23555857e-01
4.95649308e-01 4.79711801e-01 2.91813999e-01 -8.67365956e-01
5.74075520e-01 8.66240323e-01 -3.38285595e-01 -4.39444393e-01
-1.04443014e+00 -7.80058861e-01 -1.54894665e-01 -6.57198834e-04
1.78929955e-01 -8.01357925e-01 -1.05294549e+00 8.34193751e-02
-7.24132001e-01 -6.59009397e-01 -3.65984827e-01 5.59719265e-01
-1.08328390e+00 2.98240781e-01 -4.61467206e-01 -1.62652206e+00
1.27375741e-02 -6.67598903e-01 5.89684308e-01 1.22475214e-01
1.08844303e-01 -1.22015715e+00 3.09246302e-01 6.28825128e-01
1.73259392e-01 8.70008618e-02 7.41926968e-01 -6.54694378e-01
-9.30339873e-01 -1.19378954e-01 -3.45814288e-01 -1.13135867e-01
2.82421082e-01 -3.74018736e-02 -7.69439340e-01 -5.66493869e-01
-1.24894649e-01 -3.49979959e-02 1.05069458e+00 1.33742154e+00
1.27891529e+00 -7.23910391e-01 -6.00305080e-01 8.15340161e-01
1.57014668e+00 5.39433360e-01 3.89345914e-01 5.97347021e-01
6.34038225e-02 3.11042666e-01 5.34673274e-01 7.74195015e-01
1.80971157e-02 5.29019952e-01 5.51475465e-01 -8.78841802e-02
5.86784542e-01 -1.04969449e-01 6.06884947e-03 4.70341772e-01
5.69796041e-02 -1.04098536e-01 -9.09312189e-01 1.03194737e+00
-2.53057337e+00 -1.45776331e+00 -6.76008835e-02 2.30977964e+00
8.27014148e-01 -1.14964448e-01 6.48573935e-01 -1.58929899e-02
1.35389960e+00 6.48616031e-02 -9.90107596e-01 -3.02618742e-01
-2.81767964e-01 -1.22719713e-01 1.19755137e+00 4.06456441e-01
-1.14838576e+00 9.39231575e-01 7.55244064e+00 8.92587185e-01
-8.86163771e-01 3.66934985e-01 1.01773036e+00 -5.97211242e-01
-2.49097809e-01 2.64206529e-01 -3.16497117e-01 5.33001959e-01
7.46037245e-01 -3.57728630e-01 9.64791059e-01 8.81326854e-01
3.89074713e-01 -1.68991849e-01 -1.08503735e+00 1.16704261e+00
-4.96200323e-01 -1.63593912e+00 -3.48882943e-01 1.97416514e-01
1.32559276e+00 3.35598469e-01 4.28939760e-01 7.83718377e-02
1.15850878e+00 -5.57476103e-01 8.78034413e-01 7.02441186e-02
5.73326528e-01 -8.90368760e-01 3.03646713e-01 2.32648358e-01
-7.23605931e-01 -6.34117842e-01 -1.66215360e-01 -3.69399428e-01
-1.38718426e-01 7.17808127e-01 -6.20098412e-01 6.88794971e-01
8.73399198e-01 6.25601649e-01 2.05269769e-01 1.73452759e+00
4.49196607e-01 1.12594855e+00 -5.76020122e-01 4.69930582e-02
7.54530489e-01 -6.32956266e-01 3.17519218e-01 1.04039598e+00
4.03569996e-01 1.72888935e-02 2.04094946e-01 5.66025972e-01
3.44778225e-02 7.38405660e-02 -2.82811075e-01 5.63700758e-02
6.51891470e-01 5.17931044e-01 -7.99547076e-01 -3.51887286e-01
-4.17401671e-01 5.90687931e-01 2.59106398e-01 7.44150817e-01
-1.23657346e+00 3.71327430e-01 9.44792688e-01 -7.97915161e-02
8.31029713e-01 2.53000140e-01 -6.64969563e-01 -1.05855286e+00
-8.12338069e-02 -6.40188694e-01 9.40777004e-01 -2.29331210e-01
-1.33188713e+00 -5.07756412e-01 -2.57836673e-02 -7.26383448e-01
2.35402938e-02 -1.23041838e-01 -5.08824587e-01 2.34645098e-01
-1.78233361e+00 -9.68184769e-01 3.69946510e-01 8.15760851e-01
5.07960916e-01 3.66197556e-01 2.50680208e-01 2.35089928e-01
-1.02417481e+00 5.06438911e-01 1.52302432e+00 -1.84582010e-01
2.69016147e-01 -1.09944892e+00 1.19220451e-01 1.01957929e+00
1.46235272e-01 5.65880537e-01 8.99216831e-01 -4.85949367e-01
-1.49210978e+00 -1.17114782e+00 5.56679592e-02 9.80336890e-02
1.18488896e+00 -1.42552048e-01 -1.74582586e-01 1.44007683e+00
-9.35364962e-02 -5.64397573e-02 8.60763371e-01 7.51950204e-01
-3.31685126e-01 -5.72551548e-01 -9.49951410e-01 5.04115105e-01
1.36604321e+00 -2.21497655e-01 -2.08616838e-01 6.92406654e-01
6.16672456e-01 -1.35263979e-01 -4.89881724e-01 1.32965535e-01
6.21536255e-01 -1.23494661e+00 1.01359904e+00 -9.50125813e-01
-2.87782907e-01 2.69238681e-01 -2.15949371e-01 -1.14585221e+00
-6.27078176e-01 -1.35359275e+00 -6.84141070e-02 1.09797037e+00
2.82239228e-01 -1.15595758e+00 1.00371385e+00 6.06124878e-01
1.20294048e-02 -7.15506747e-02 -1.66670561e+00 -9.94935811e-01
1.53021058e-02 -6.94099009e-01 9.49299812e-01 1.28970563e+00
2.50318557e-01 -2.26406440e-01 -5.88824332e-01 1.82642221e-01
6.11458123e-01 8.13745975e-01 7.83880770e-01 -1.23737407e+00
-4.65971917e-01 -8.41384411e-01 6.51736781e-02 -1.38620710e+00
1.85608119e-01 -7.38839149e-01 -1.86891362e-01 -1.32311380e+00
7.31815636e-01 -7.05930948e-01 -6.86444640e-01 3.53913277e-01
1.94673374e-01 9.30950716e-02 -1.01819806e-01 2.64334917e-01
-9.04570341e-01 2.98068345e-01 7.56707430e-01 -4.07712489e-01
-3.92287195e-01 5.63017011e-01 -9.99507070e-01 2.58303523e-01
7.79159784e-01 -4.98037159e-01 -4.75012660e-01 -4.98793930e-01
5.06204844e-01 4.13905799e-01 3.58569324e-01 -4.66440707e-01
2.20694676e-01 -8.15372169e-01 -1.08721010e-01 -6.45091295e-01
8.23488235e-02 -1.27595270e+00 4.96960849e-01 5.46734035e-01
-3.17338109e-01 -2.50644922e-01 2.04340279e-01 1.25109231e+00
1.95461810e-02 1.01577505e-01 6.71653688e-01 -9.26295370e-02
-2.11406291e-01 4.43077058e-01 -1.86106220e-01 1.91825330e-02
1.27194071e+00 -6.29517496e-01 -3.50477785e-01 -7.65981913e-01
-7.36630082e-01 3.84811848e-01 2.68737853e-01 -1.21359609e-01
-1.49799898e-01 -1.30350900e+00 -5.20198464e-01 -4.34904359e-02
-1.46320626e-01 -4.25513208e-01 3.94177079e-01 1.33654523e+00
-2.63942391e-01 9.79226053e-01 2.52767086e-01 -7.38447905e-01
-8.72119486e-01 9.64374363e-01 3.67455810e-01 -3.34600389e-01
-2.04986602e-01 6.33513153e-01 3.50842297e-01 -1.81444749e-01
3.68856788e-01 -5.18283784e-01 7.17263579e-01 2.29912192e-01
1.56346008e-01 6.16093934e-01 3.28099914e-02 7.49261677e-02
-1.35649458e-01 5.03906943e-02 -5.38277254e-02 -1.39856249e-01
1.77175522e+00 -5.66081107e-01 -3.15632463e-01 5.63015044e-01
8.19430411e-01 -2.32986644e-01 -1.24625707e+00 -5.57251871e-01
4.45628643e-01 -5.47629714e-01 1.24512382e-01 -5.91666818e-01
-9.94787693e-01 2.05706120e-01 3.39136064e-01 1.05894935e+00
9.22652900e-01 6.76132292e-02 5.72634101e-01 4.35378999e-01
3.51222396e-01 -1.40186274e+00 -6.54611647e-01 3.83955449e-01
2.20944196e-01 -1.03038454e+00 -6.45663738e-02 -2.61832416e-01
-7.16855004e-02 6.29079521e-01 -8.36450085e-02 -1.38226122e-01
5.87673306e-01 -1.43145889e-01 -5.17551184e-01 2.24250957e-01
-5.76400518e-01 -4.68243808e-01 -2.13228300e-01 3.30278456e-01
6.03934787e-02 3.16582233e-01 -4.36529279e-01 5.04058778e-01
-1.64277405e-01 -4.25114930e-02 2.51789480e-01 8.00102055e-01
-4.53427851e-01 -6.34677529e-01 -6.88082755e-01 5.49798787e-01
-6.36430144e-01 -1.38161987e-01 -3.26337188e-01 8.42942774e-01
-6.79503679e-01 1.00354779e+00 4.10522252e-01 2.62552589e-01
-1.14002824e-01 1.69951320e-01 4.49446887e-01 -2.84056645e-02
-4.19638693e-01 6.23079062e-01 4.34046835e-02 -2.96011239e-01
-7.83986092e-01 -8.89005065e-01 -3.02972645e-01 -6.90211356e-01
-7.09597349e-01 5.79742074e-01 4.86575991e-01 8.74316275e-01
2.46363580e-01 7.80615136e-02 9.24310863e-01 -5.82284451e-01
-7.84293413e-01 -5.97684920e-01 -1.04049826e+00 2.34819442e-01
6.31127715e-01 -4.50404972e-01 -3.37240338e-01 -3.27877760e-01] | [4.5447869300842285, 3.3008763790130615] |
27aa76b1-da97-41ff-9090-d56e64790076 | mmptrack-large-scale-densely-annotated-multi | 2111.15157 | null | https://arxiv.org/abs/2111.15157v1 | https://arxiv.org/pdf/2111.15157v1.pdf | MMPTRACK: Large-scale Densely Annotated Multi-camera Multiple People Tracking Benchmark | Multi-camera tracking systems are gaining popularity in applications that demand high-quality tracking results, such as frictionless checkout because monocular multi-object tracking (MOT) systems often fail in cluttered and crowded environments due to occlusion. Multiple highly overlapped cameras can significantly alleviate the problem by recovering partial 3D information. However, the cost of creating a high-quality multi-camera tracking dataset with diverse camera settings and backgrounds has limited the dataset scale in this domain. In this paper, we provide a large-scale densely-labeled multi-camera tracking dataset in five different environments with the help of an auto-annotation system. The system uses overlapped and calibrated depth and RGB cameras to build a high-performance 3D tracker that automatically generates the 3D tracking results. The 3D tracking results are projected to each RGB camera view using camera parameters to create 2D tracking results. Then, we manually check and correct the 3D tracking results to ensure the label quality, which is much cheaper than fully manual annotation. We have conducted extensive experiments using two real-time multi-camera trackers and a person re-identification (ReID) model with different settings. This dataset provides a more reliable benchmark of multi-camera, multi-object tracking systems in cluttered and crowded environments. Also, our results demonstrate that adapting the trackers and ReID models on this dataset significantly improves their performance. Our dataset will be publicly released upon the acceptance of this work. | ['Zicheng Liu', 'Jiang Wang', 'Houdong Hu', 'Peng Chu', 'Zhizheng Zhang', 'Chunyu Wang', 'Quanzeng You', 'Xiaotian Han'] | 2021-11-30 | null | null | null | null | ['multiple-people-tracking'] | ['computer-vision'] | [-2.63200402e-01 -8.35862219e-01 2.52850890e-01 -1.40099004e-01
-5.47824502e-01 -9.43264902e-01 1.45072371e-01 -1.92981720e-01
-5.93246460e-01 5.97221911e-01 -8.80782977e-02 1.57239631e-01
5.07825732e-01 -3.52262318e-01 -5.50425172e-01 -6.09097540e-01
4.78111178e-01 6.25034153e-01 6.91113770e-01 2.08433464e-01
-1.84418604e-01 6.27555728e-01 -1.89855671e+00 -2.46756747e-01
3.70175779e-01 4.26143140e-01 2.76469588e-01 1.03066552e+00
1.45111188e-01 1.78691879e-01 -8.64042521e-01 -5.08594215e-01
6.81747615e-01 -6.33259863e-02 -1.40186787e-01 3.00427616e-01
9.48948145e-01 -5.79562783e-01 -7.80337304e-02 1.03565550e+00
6.81989312e-01 8.60631168e-02 7.62666613e-02 -1.74966609e+00
-4.70893741e-01 -3.43795419e-01 -8.75726998e-01 1.00300878e-01
9.57912326e-01 3.48824620e-01 3.83744180e-01 -5.58006823e-01
5.44106722e-01 1.29503334e+00 1.00447989e+00 9.32668030e-01
-8.73020411e-01 -1.02558398e+00 9.28746387e-02 -1.73798993e-01
-1.68384254e+00 -4.16089833e-01 3.54616016e-01 -4.87566680e-01
5.35842121e-01 3.11317027e-01 9.08363938e-01 1.13396180e+00
3.67834419e-02 4.73369926e-01 1.06250167e+00 -3.26772660e-01
-1.35058999e-01 4.17002469e-01 -1.28665902e-02 7.02497303e-01
9.11238611e-01 2.46840596e-01 -3.81416142e-01 -3.11252207e-01
8.78194869e-01 6.26610100e-01 -2.39290208e-01 -4.06971186e-01
-1.38306868e+00 2.41239220e-01 1.17245331e-01 2.29078174e-01
-2.50499602e-02 1.69604629e-01 6.54903129e-02 -3.59089039e-02
3.09550967e-02 -2.80306786e-01 -3.12705815e-01 -1.55348703e-01
-7.73006439e-01 2.54696518e-01 4.94894505e-01 1.55011261e+00
6.25238955e-01 -2.54241019e-01 -6.47554994e-02 5.11503041e-01
5.94635129e-01 1.19636536e+00 2.65026301e-01 -8.96678507e-01
2.45127007e-01 7.86114395e-01 6.50753319e-01 -8.78904045e-01
-5.67814231e-01 -7.06833974e-02 -5.60148716e-01 3.40236813e-01
7.46381819e-01 -3.41814220e-01 -5.49238801e-01 1.42118502e+00
9.00538325e-01 2.73922145e-01 -2.20299006e-01 1.18754077e+00
7.55166471e-01 1.47056088e-01 -1.10850535e-01 -1.42367527e-01
1.46592593e+00 -8.81708860e-01 -8.03441286e-01 -1.31549597e-01
5.12858450e-01 -9.78970587e-01 7.38029182e-01 2.90619582e-01
-8.25626075e-01 -8.37399244e-01 -7.59067714e-01 1.09113596e-01
-3.10983181e-01 2.74680912e-01 2.46754333e-01 1.11893690e+00
-1.13534009e+00 -4.96213101e-02 -7.99715936e-01 -7.36533523e-01
1.25712141e-01 3.91504854e-01 -5.37859499e-01 -3.17453325e-01
-7.18430102e-01 1.00156879e+00 2.48433843e-01 -4.30602245e-02
-7.48565257e-01 -3.51702541e-01 -7.04420567e-01 -4.55001682e-01
3.23539197e-01 -7.41758585e-01 1.07053316e+00 -4.51649904e-01
-1.28878832e+00 1.05811894e+00 -3.40125799e-01 1.34294778e-01
7.30260670e-01 -4.09163982e-01 -5.62981665e-01 -1.15176104e-03
4.31517422e-01 6.12198293e-01 6.01669550e-01 -1.28724003e+00
-9.29732919e-01 -5.38117707e-01 -6.99982271e-02 2.21347481e-01
-2.04000562e-01 3.75507981e-01 -9.73910272e-01 -2.74383605e-01
-2.50425041e-01 -1.30111206e+00 -7.99779817e-02 3.19062293e-01
-2.03896090e-01 1.24405086e-01 9.83183861e-01 -3.61261457e-01
7.62412965e-01 -2.08208561e+00 -3.55010033e-02 -2.94293076e-01
1.57915592e-01 3.32233161e-01 1.59796685e-01 -4.83524241e-02
4.11205828e-01 -3.03526103e-01 4.44497794e-01 -8.20092201e-01
-8.21678862e-02 1.02259442e-01 2.72004575e-01 8.14887166e-01
-4.32420343e-01 6.62133515e-01 -8.84186327e-01 -9.48988676e-01
6.63403153e-01 5.85788012e-01 -2.49399677e-01 3.49279344e-01
1.37398258e-01 7.41779506e-01 -1.17547803e-01 1.11113727e+00
8.93896937e-01 -2.76230752e-01 -1.54849231e-01 -1.33695737e-01
-3.66685092e-01 -8.12427223e-01 -1.90947044e+00 1.65340626e+00
-9.18050930e-02 5.49729764e-01 8.91362503e-02 1.21152818e-01
8.31294119e-01 3.71686459e-01 6.46106184e-01 -2.77645856e-01
2.23101929e-01 -2.38007288e-02 -4.78066921e-01 -1.78861603e-01
9.26108718e-01 1.68078113e-02 -1.21268533e-01 5.31618834e-01
-1.46849200e-01 2.42760152e-01 1.03741914e-01 1.25722811e-01
9.07307684e-01 3.49279344e-01 2.06397951e-01 2.16205195e-01
4.47023720e-01 1.57855883e-01 8.13652456e-01 7.89107382e-01
-7.49921441e-01 6.71521604e-01 -5.97805679e-01 -5.55355668e-01
-1.04420698e+00 -8.88377130e-01 4.17379178e-02 8.83145273e-01
6.77363157e-01 -4.04923290e-01 -4.89011168e-01 -2.86919564e-01
1.13727078e-01 6.41576722e-02 -1.99361175e-01 2.62504876e-01
-4.51690257e-01 -7.38540828e-01 7.86222577e-01 5.77221632e-01
7.02460647e-01 -4.89187092e-01 -9.42887962e-01 3.27260606e-02
-2.89062202e-01 -1.56733501e+00 -7.55018175e-01 -3.41386229e-01
-5.33300400e-01 -1.48602283e+00 -8.47202361e-01 -5.50671279e-01
7.01087415e-01 1.06985021e+00 5.54548979e-01 3.82286459e-01
-3.49227518e-01 8.52168620e-01 -2.61200160e-01 -4.45918381e-01
-4.29098159e-02 -4.26731557e-01 7.31656551e-01 -6.74401224e-02
6.76542401e-01 1.39678776e-01 -3.85804445e-01 9.49246645e-01
-5.26871681e-01 9.68648866e-03 1.52917340e-01 3.00383985e-01
7.00715065e-01 1.13585768e-02 -2.01127008e-02 -3.02573234e-01
1.28337309e-01 2.15293607e-03 -1.25929058e+00 4.94902223e-01
-1.50094584e-01 -5.23528337e-01 3.02784473e-01 -7.01715052e-01
-9.87069070e-01 6.03245556e-01 3.44569832e-01 -7.42251873e-01
-5.25659025e-01 -5.54433227e-01 -2.75856704e-01 -4.77434486e-01
4.74505216e-01 -1.52186747e-03 -3.43627006e-01 -4.58954275e-01
2.48049319e-01 8.26592863e-01 7.34973252e-01 -3.42748851e-01
1.14183700e+00 7.14971662e-01 -8.04151297e-02 -6.25395834e-01
-7.47163653e-01 -9.55608904e-01 -1.08734381e+00 -6.48584068e-01
1.02823055e+00 -1.38465726e+00 -1.12055421e+00 7.20012307e-01
-1.25688660e+00 -3.05165928e-02 1.14386044e-01 6.92107022e-01
-7.22229257e-02 5.76646030e-01 -4.08871591e-01 -1.10588968e+00
-1.91417411e-01 -1.10851812e+00 1.50607920e+00 7.46924162e-01
1.33106962e-01 -8.04751873e-01 2.90229708e-01 5.20204067e-01
2.25689158e-01 4.42566752e-01 -4.49729204e-01 -1.17990345e-01
-9.00140047e-01 -4.44960773e-01 -1.17204569e-01 -2.40144745e-01
1.55509636e-01 1.05020747e-01 -1.09728873e+00 -5.43189228e-01
-2.98505008e-01 -6.19822182e-02 2.39935860e-01 4.02859509e-01
5.67562580e-01 3.52837980e-01 -8.41745019e-01 6.57126963e-01
1.48059583e+00 8.49262774e-02 1.32388502e-01 4.92782325e-01
1.02262223e+00 1.01483718e-01 8.94064426e-01 4.71287429e-01
7.41443276e-01 1.09486222e+00 2.34207019e-01 3.52291055e-02
-2.23367140e-01 -3.55592072e-02 4.76924896e-01 4.96236145e-01
-1.73057139e-01 -2.23730117e-01 -8.53799641e-01 2.71832675e-01
-1.89044666e+00 -1.10689795e+00 -5.94592333e-01 2.36474395e+00
5.16867042e-01 -1.04515351e-01 5.42425573e-01 -2.14127183e-01
1.33343565e+00 -4.17774528e-01 -4.87895668e-01 4.48893875e-01
-2.48802736e-01 -4.89691198e-01 8.15588057e-01 4.25312042e-01
-1.15945232e+00 8.77675951e-01 6.26737833e+00 1.01140760e-01
-6.18616700e-01 3.28411788e-01 -1.94429249e-01 -3.59493047e-01
4.60180491e-01 -2.12994263e-01 -1.74869955e+00 6.65312767e-01
8.65219355e-01 6.64319396e-02 2.41097093e-01 8.05907190e-01
1.14262141e-02 -4.54327941e-01 -9.67839062e-01 1.69025934e+00
2.56568044e-01 -9.43692684e-01 -5.06741583e-01 3.02175552e-01
7.19846129e-01 5.41659333e-02 -4.40469474e-01 3.40876840e-02
5.37439942e-01 -3.75176787e-01 9.04272914e-01 5.46194673e-01
8.62529695e-01 -3.40794116e-01 7.70204186e-01 4.49506432e-01
-1.78951621e+00 1.75790116e-02 -5.49876332e-01 -7.57191516e-03
5.14385879e-01 2.89174169e-01 -5.62395811e-01 5.16256213e-01
1.03817940e+00 5.98526776e-01 -9.53197479e-01 1.55164683e+00
8.24222788e-02 -2.29519501e-01 -4.96038646e-01 -5.74839078e-02
-4.19859022e-01 -4.06092964e-02 4.80772197e-01 1.06321573e+00
3.16024631e-01 1.64857849e-01 5.58204710e-01 4.80887622e-01
3.54216956e-02 -4.59874064e-01 -7.74652183e-01 5.43790400e-01
7.85788894e-01 1.49985814e+00 -7.63810217e-01 -4.08951283e-01
-7.39368975e-01 1.04134738e+00 -4.30393144e-02 3.03446464e-02
-1.23324227e+00 1.08029870e-02 6.37647331e-01 -1.41115338e-01
2.60292858e-01 -4.49957103e-01 9.58120897e-02 -1.61629915e+00
2.51854192e-02 -5.20480096e-01 5.79211354e-01 -1.00545347e+00
-1.06257629e+00 5.69592774e-01 1.34296566e-01 -1.58516884e+00
4.15637158e-02 -5.18590093e-01 -7.04881102e-02 7.80304611e-01
-1.35032928e+00 -1.43154097e+00 -1.10981333e+00 9.97110724e-01
2.27344841e-01 -1.61098525e-01 7.73817539e-01 8.76350701e-01
-7.56491780e-01 6.39177144e-01 -2.66127735e-02 2.66474426e-01
1.14206243e+00 -1.08387268e+00 1.69304043e-01 1.12629735e+00
-4.32085246e-02 5.31421661e-01 4.49941427e-01 -9.63754892e-01
-1.67058718e+00 -1.25095558e+00 4.95494694e-01 -1.30513167e+00
2.19085336e-01 -5.13688624e-01 -5.06883919e-01 1.00846863e+00
-1.34159461e-01 4.54354882e-01 8.58918607e-01 -1.85651183e-01
4.38538603e-02 -5.37202656e-02 -1.31376803e+00 2.10508808e-01
1.17138207e+00 -2.85599262e-01 -3.77861857e-01 3.87269199e-01
6.10433400e-01 -8.55672777e-01 -8.97335231e-01 -1.89629689e-01
7.92095959e-01 -8.06445420e-01 1.04938388e+00 1.28241822e-01
-7.25228012e-01 -1.08878887e+00 -3.32234859e-01 -9.00295436e-01
-2.13138178e-01 -4.58393961e-01 -2.58513018e-02 1.41447711e+00
-2.94803619e-01 -5.36309659e-01 7.16704845e-01 9.09036815e-01
1.20581225e-01 2.53268093e-01 -7.67736733e-01 -9.56121981e-01
-6.22780561e-01 -2.06550390e-01 8.51717889e-01 7.85071254e-01
-4.53936726e-01 9.37831476e-02 -6.91545308e-01 6.52075887e-01
1.17744148e+00 1.53490037e-01 1.53071427e+00 -1.56426549e+00
8.14883132e-03 1.24128416e-01 -4.99207884e-01 -8.58255029e-01
-2.19023362e-01 -4.38994020e-01 -7.75582865e-02 -1.23570359e+00
5.03251970e-01 -4.86381859e-01 3.18686277e-01 3.40250492e-01
-2.33511165e-01 5.20850241e-01 5.12327313e-01 5.63137949e-01
-1.26394999e+00 1.53460488e-01 9.93707120e-01 8.44374299e-02
-1.04328655e-01 -6.68188417e-03 -4.21684355e-01 6.15006506e-01
4.95980054e-01 -5.71982205e-01 1.93963721e-01 -6.56370997e-01
-2.14156300e-01 5.82193397e-03 7.58409083e-01 -1.41021764e+00
7.14529395e-01 -7.39303827e-02 9.46644366e-01 -9.76602912e-01
5.66346765e-01 -1.31620407e+00 7.82700896e-01 5.40658534e-01
3.21146578e-01 4.81435001e-01 2.59162933e-01 5.79936385e-01
2.28634968e-01 1.39158396e-02 7.98377395e-01 -4.50754106e-01
-7.78107285e-01 5.02807021e-01 1.62469354e-02 -1.59875780e-01
1.46151078e+00 -6.44711733e-01 -4.67635423e-01 -6.90728202e-02
-5.34304559e-01 3.95906150e-01 1.14109457e+00 4.43231106e-01
5.80930471e-01 -1.67977214e+00 -4.52392220e-01 3.08939964e-01
1.95502385e-01 9.92252529e-02 2.77127698e-02 7.52758384e-01
-5.24521470e-01 5.49656570e-01 -4.78773147e-01 -1.08287430e+00
-1.85330188e+00 6.31230474e-01 3.26528400e-01 1.25544086e-01
-6.95426762e-01 3.45300525e-01 -2.50658095e-01 -5.21510601e-01
4.00014073e-01 -2.80396342e-01 3.44612002e-02 -2.85845697e-01
1.02606237e+00 6.69332087e-01 -2.51509368e-01 -1.09638870e+00
-8.16663921e-01 1.02467358e+00 2.92253882e-01 -1.26337320e-01
9.69791651e-01 -5.32805979e-01 2.71027207e-01 2.60729373e-01
7.78160274e-01 6.13942146e-02 -1.38088834e+00 -1.63476855e-01
-1.91489860e-01 -1.03034568e+00 -3.15519989e-01 -4.08376127e-01
-9.74413753e-01 4.26102191e-01 1.00699687e+00 -5.76411299e-02
8.23978782e-01 -4.93678860e-02 7.75776803e-01 3.27024400e-01
9.62886572e-01 -1.01170540e+00 7.93599784e-02 2.30760261e-01
1.19630091e-01 -1.58322597e+00 2.03603938e-01 -1.27466410e-01
-6.76223874e-01 9.45596516e-01 1.13766408e+00 3.10790300e-01
2.36958370e-01 4.39598948e-01 2.39190117e-01 -1.82108417e-01
-2.66266018e-01 -3.75237435e-01 -1.09430827e-01 1.02921021e+00
2.72342786e-02 3.17131132e-02 4.32989001e-01 1.18145891e-01
4.24742140e-02 8.48079920e-02 6.41823947e-01 1.06042516e+00
-3.65466416e-01 -1.01844275e+00 -1.32793665e+00 -1.22205429e-01
-1.60942778e-01 6.02810204e-01 -3.70827258e-01 8.89064252e-01
3.57955605e-01 1.19890571e+00 -8.11037943e-02 -3.87692094e-01
4.53816593e-01 -2.30071187e-01 6.97393894e-01 -4.58396673e-01
-7.28861094e-01 1.85350552e-01 -2.80878901e-01 -5.26231050e-01
-9.21473145e-01 -1.00009274e+00 -1.09019029e+00 -6.67194903e-01
-7.55270600e-01 -5.33786379e-02 6.88993216e-01 7.02082396e-01
3.07498783e-01 1.60550237e-01 2.94557631e-01 -1.06411886e+00
-3.50698456e-02 -7.16285050e-01 -3.97584319e-01 5.64797997e-01
3.99233520e-01 -1.01924980e+00 -2.32532956e-02 2.34293684e-01] | [6.7153496742248535, -1.618273377418518] |
18badebe-4c9a-4328-ab6d-6bf5aa5e7c5a | robust-localized-multi-view-subspace | 1705.07777 | null | http://arxiv.org/abs/1705.07777v1 | http://arxiv.org/pdf/1705.07777v1.pdf | Robust Localized Multi-view Subspace Clustering | In multi-view clustering, different views may have different confidence
levels when learning a consensus representation. Existing methods usually
address this by assigning distinctive weights to different views. However, due
to noisy nature of real-world applications, the confidence levels of samples in
the same view may also vary. Thus considering a unified weight for a view may
lead to suboptimal solutions. In this paper, we propose a novel localized
multi-view subspace clustering model that considers the confidence levels of
both views and samples. By assigning weight to each sample under each view
properly, we can obtain a robust consensus representation via fusing the
noiseless structures among views and samples. We further develop a regularizer
on weight parameters based on the convex conjugacy theory, and samples weights
are determined in an adaptive manner. An efficient iterative algorithm is
developed with a convergence guarantee. Experimental results on four benchmarks
demonstrate the correctness and effectiveness of the proposed model. | ['Bao-Gang Hu', 'Ran He', 'Siwei Lyu', 'Yanbo Fan', 'Jian Liang'] | 2017-05-22 | null | null | null | null | ['multi-view-subspace-clustering'] | ['computer-vision'] | [-2.23923460e-01 -5.66226065e-01 -9.42980275e-02 -3.59898508e-01
-6.65625095e-01 -7.22956061e-01 1.95126921e-01 -2.02611983e-01
-5.12057263e-03 3.51302177e-01 3.63308638e-01 3.88584107e-01
-3.36266518e-01 -4.13152516e-01 -4.14440781e-01 -1.10443866e+00
4.04782861e-01 3.34260106e-01 8.24030638e-02 3.16217035e-01
1.30756527e-01 9.85047314e-03 -1.26853788e+00 1.86361045e-01
1.09618163e+00 7.46255398e-01 1.98362827e-01 2.43404403e-01
1.83967799e-01 2.42001727e-01 -4.08084244e-01 -1.18990533e-01
5.42087674e-01 -4.76486444e-01 -2.79798031e-01 8.34906757e-01
4.58838582e-01 -1.00366488e-01 6.76648170e-02 1.63338518e+00
3.79479676e-01 -1.04613891e-02 5.66825688e-01 -1.32913756e+00
-4.33769017e-01 5.62205911e-01 -1.14031303e+00 -4.35296819e-02
1.63087040e-01 -1.58507153e-01 9.50455904e-01 -1.07398117e+00
4.69625533e-01 1.28972781e+00 4.24285293e-01 3.23264897e-01
-1.38637006e+00 -6.91791236e-01 7.63174295e-01 3.04597259e-01
-1.56444764e+00 -4.75469738e-01 1.01531124e+00 -3.68845850e-01
-2.49625936e-01 2.32004017e-01 5.48636019e-01 8.06777060e-01
-6.12628795e-02 8.89328003e-01 1.50525856e+00 6.62410259e-02
2.28584096e-01 1.91229090e-01 1.47920728e-01 4.57185179e-01
7.81500340e-01 -4.19609487e-01 -2.35264033e-01 -4.62286562e-01
5.39579511e-01 3.30607206e-01 -7.69911468e-01 -1.03978419e+00
-1.34244967e+00 6.55882239e-01 1.51495352e-01 2.04877123e-01
-2.49573797e-01 -2.93418646e-01 4.08463538e-01 2.23915979e-01
3.21113139e-01 -5.37787855e-01 -1.95430309e-01 4.11047876e-01
-7.89501131e-01 -6.20414875e-02 7.18149304e-01 1.13814616e+00
6.76812708e-01 -5.37094548e-02 2.02684477e-02 9.36062157e-01
5.44378698e-01 6.55033290e-01 2.43116781e-01 -1.03647482e+00
5.75743616e-01 6.59503877e-01 2.28689238e-01 -1.43120015e+00
-7.27977464e-03 -5.24133682e-01 -1.37652946e+00 2.47746408e-01
4.67627019e-01 -1.68431848e-01 -5.50134897e-01 1.78045988e+00
6.73736572e-01 4.75806773e-01 2.03330696e-01 1.17914486e+00
3.39397311e-01 5.18827319e-01 -5.14182806e-01 -7.09807277e-01
1.08209300e+00 -6.14136934e-01 -7.97890306e-01 3.95308658e-02
-1.10689411e-02 -7.69773066e-01 5.67168415e-01 6.92023098e-01
-9.17836130e-01 -4.17493939e-01 -1.31723976e+00 5.82705736e-01
2.53538013e-01 1.72002330e-01 3.92271653e-02 6.02431893e-01
-7.09012866e-01 2.43889168e-01 -7.48567462e-01 -2.55878896e-01
7.02558383e-02 2.62450904e-01 -4.39503968e-01 -4.03403014e-01
-6.03584111e-01 4.09205258e-01 3.29496324e-01 3.10376227e-01
-6.22222900e-01 -4.22711849e-01 -5.71680009e-01 -5.17421067e-02
5.83640039e-01 -7.62540400e-01 6.98436201e-01 -9.57509160e-01
-1.07203126e+00 5.36717653e-01 -4.60413873e-01 6.31309375e-02
6.44434273e-01 -9.42708403e-02 -6.68217778e-01 1.67092413e-01
3.99777770e-01 -7.88346902e-02 1.04166973e+00 -2.08060217e+00
-8.25650871e-01 -5.23288965e-01 -2.10872024e-01 3.89210790e-01
-2.22216010e-01 -2.13055268e-01 -8.64122927e-01 -3.66327733e-01
7.93899119e-01 -9.21187282e-01 -3.68224680e-01 -1.60940543e-01
-3.60009044e-01 9.35036615e-02 1.03906262e+00 -4.63534266e-01
1.13873446e+00 -2.25792551e+00 6.15850449e-01 6.54074430e-01
4.50784713e-01 -9.15665999e-02 2.25621194e-01 1.49179667e-01
-2.51369458e-02 -1.79898396e-01 -3.18909645e-01 -1.35586053e-01
-6.43448904e-02 2.98813760e-01 -2.13757426e-01 9.47428763e-01
-2.12588966e-01 9.77328345e-02 -9.06214118e-01 -5.85629046e-01
2.76023507e-01 2.75947809e-01 -5.41204751e-01 1.76142633e-01
4.05563444e-01 4.43892509e-01 -6.33347154e-01 5.78923047e-01
1.22216117e+00 -4.62399989e-01 6.74037039e-01 -6.50587797e-01
8.66029933e-02 -6.09028637e-01 -2.15571904e+00 1.51432085e+00
-2.67727852e-01 -9.70171839e-02 7.85094738e-01 -1.17212296e+00
6.68789268e-01 3.85204703e-01 6.23333454e-01 5.54974675e-02
1.12283014e-01 1.70112878e-01 6.22193096e-03 -1.27097189e-01
2.25617424e-01 -1.90278560e-01 6.56445771e-02 4.01350915e-01
-1.58474877e-01 2.78212458e-01 1.09611280e-01 2.62649953e-01
5.25971115e-01 -2.41801292e-01 3.88952225e-01 -4.31616366e-01
9.94921327e-01 -4.82747048e-01 1.16343391e+00 4.80591446e-01
-3.40456039e-01 8.69887292e-01 2.97316074e-01 -1.76572144e-01
-8.65254939e-01 -1.06760752e+00 -5.96499406e-02 5.32225192e-01
5.44669211e-01 -3.88043553e-01 -7.10749507e-01 -7.14555025e-01
-2.06413180e-01 1.72496259e-01 -4.61950272e-01 8.12128361e-04
-4.35437799e-01 -7.75862038e-01 -1.69517413e-01 2.89118916e-01
5.57382047e-01 -1.41061649e-01 -1.25169560e-01 2.77610660e-01
-3.12539279e-01 -1.14619350e+00 -9.03309226e-01 -1.23111360e-01
-9.37861443e-01 -1.38033903e+00 -7.58733034e-01 -7.72740602e-01
1.20682347e+00 9.89562333e-01 7.56437421e-01 1.44438878e-01
3.73935014e-01 6.91060305e-01 -3.09785485e-01 8.33012536e-02
-2.53413588e-01 -3.02280664e-01 4.99359846e-01 6.76815987e-01
1.44534618e-01 -6.48849130e-01 -6.21552289e-01 6.47872984e-01
-9.70541298e-01 -1.59675088e-02 4.33131367e-01 1.03671849e+00
7.72038698e-01 2.47094214e-01 4.25243825e-01 -7.86410809e-01
4.95579898e-01 -5.84162533e-01 -6.81192279e-01 5.89841962e-01
-7.82311320e-01 -7.33374953e-02 8.36676538e-01 -3.52589905e-01
-1.11353600e+00 2.09593758e-01 6.06793165e-01 -9.43695068e-01
-1.27863482e-01 4.11868453e-01 -7.21611202e-01 8.83394778e-02
-3.67133208e-02 3.66329402e-01 1.72050506e-01 -2.60968804e-01
5.28998911e-01 4.72389311e-01 4.70015556e-01 -8.71839762e-01
1.16674495e+00 8.45557988e-01 -1.41913176e-01 -5.29073834e-01
-7.86577344e-01 -6.85814142e-01 -6.63309276e-01 -4.24450427e-01
5.43263257e-01 -1.19318831e+00 -7.12013304e-01 2.94659555e-01
-7.86516547e-01 6.88078046e-01 1.49636179e-01 6.07411206e-01
-3.55164915e-01 9.47874606e-01 -2.56317586e-01 -6.70531332e-01
-1.64008066e-01 -1.22463667e+00 7.08982885e-01 3.11188191e-01
1.88123450e-01 -9.96747255e-01 -3.33849974e-02 3.92727017e-01
-1.53823912e-01 2.74393886e-01 4.48849618e-01 -4.10964131e-01
-5.21944404e-01 -9.73086655e-02 -1.86504975e-01 4.08283412e-01
4.29702252e-01 6.88727126e-02 -6.85768783e-01 -6.66054249e-01
5.17746449e-01 -3.35923920e-04 8.17191720e-01 4.35012728e-01
9.45156574e-01 -3.28141391e-01 -4.06771064e-01 5.27405739e-01
1.60704756e+00 -4.02174294e-02 1.08233444e-01 3.86133529e-02
9.25153613e-01 5.32068372e-01 6.97639823e-01 6.97777033e-01
3.63931268e-01 4.36666816e-01 4.09151107e-01 3.05392087e-01
3.51806313e-01 1.21038690e-01 3.53056401e-01 1.58426976e+00
-1.52951300e-01 -7.63928965e-02 -4.75100011e-01 5.72442472e-01
-2.18585873e+00 -1.21648169e+00 -3.00985783e-01 2.45909643e+00
6.18600070e-01 -1.90663561e-02 9.00952965e-02 2.17432112e-01
1.26874542e+00 2.37820596e-01 -6.37203276e-01 1.73408732e-01
-2.92189062e-01 -5.33697784e-01 3.60827506e-01 2.98490286e-01
-1.08826208e+00 2.88807511e-01 5.81790352e+00 7.64313400e-01
-7.98628092e-01 1.48525671e-03 5.58504999e-01 -7.43381272e-04
-4.80696529e-01 2.13642627e-01 -6.17346942e-01 5.71248889e-01
2.23043457e-01 -5.31129122e-01 5.34418225e-01 7.26694942e-01
2.90970206e-01 4.10071108e-03 -9.10772741e-01 1.15186810e+00
2.60666788e-01 -1.01237428e+00 9.93255600e-02 9.22999308e-02
9.92471874e-01 -2.56091118e-01 -3.14081125e-02 -1.87714174e-01
4.34656024e-01 -4.40156817e-01 6.43506467e-01 4.91082907e-01
4.22872424e-01 -1.10954130e+00 5.13081372e-01 4.76812422e-01
-1.51702070e+00 -8.28650128e-03 -6.54987454e-01 1.95134088e-01
7.35911503e-02 7.34378874e-01 -2.78808266e-01 1.22619712e+00
6.13988757e-01 1.12271523e+00 -3.09471935e-01 9.20919061e-01
5.75638339e-02 4.79616195e-01 -2.53023028e-01 4.72018868e-01
-6.86781630e-02 -9.92971122e-01 7.11002409e-01 7.61560977e-01
4.22606677e-01 2.93002605e-01 6.39219403e-01 6.85503602e-01
1.42598063e-01 1.32151842e-01 -5.24739504e-01 5.20407140e-01
7.33907163e-01 1.50498211e+00 -8.29392552e-01 -4.40302849e-01
-8.50947618e-01 1.01719904e+00 2.52904773e-01 5.17735302e-01
-7.34604061e-01 8.43859687e-02 6.46993756e-01 -3.24694276e-01
5.60187101e-01 -1.82105288e-01 -3.07281315e-01 -1.56175053e+00
2.32889399e-01 -1.04882514e+00 5.98773122e-01 -3.53449315e-01
-1.77013707e+00 1.41675994e-01 -6.78510591e-02 -2.04293466e+00
1.76852137e-01 -2.38871634e-01 -8.01165462e-01 6.66289687e-01
-1.03325331e+00 -8.98798525e-01 -2.92222798e-01 8.20721328e-01
3.07370692e-01 -1.25528917e-01 5.35393417e-01 2.05821574e-01
-7.32958257e-01 3.25701714e-01 5.33251345e-01 2.39059534e-02
8.98856223e-01 -1.24887633e+00 -3.61485660e-01 1.02941084e+00
3.91185582e-02 8.69383991e-01 7.19503105e-01 -4.33064669e-01
-1.55873692e+00 -1.05996156e+00 -1.45573830e-02 -6.85526133e-02
6.62570059e-01 -6.31096363e-02 -1.11611760e+00 6.31527841e-01
4.59829152e-01 1.14032790e-01 9.66347814e-01 1.24339767e-01
-4.42555189e-01 -3.73552859e-01 -1.14259779e+00 5.37110984e-01
7.91393757e-01 -2.93129504e-01 -5.32103479e-01 1.55965745e-01
3.40598047e-01 -2.92456776e-01 -1.08416820e+00 4.41290379e-01
5.02437949e-01 -1.25892782e+00 1.02886689e+00 -2.66069025e-01
2.31887735e-02 -9.86976802e-01 -3.76228303e-01 -1.58446324e+00
-6.32123351e-01 -2.33612731e-01 -1.60145983e-01 1.56985354e+00
-1.33831248e-01 -6.66357279e-01 6.19574249e-01 5.36893010e-01
2.69926768e-02 -3.46288562e-01 -1.02126694e+00 -8.41807365e-01
-1.48139447e-01 -4.88596708e-02 4.68723506e-01 1.29397249e+00
-7.54945129e-02 3.27142537e-01 -6.60474837e-01 8.39926898e-01
1.30468667e+00 6.22502506e-01 7.31412590e-01 -1.35063827e+00
-3.51841271e-01 -3.15968037e-01 -3.00572753e-01 -7.40028322e-01
1.36708036e-01 -7.45711386e-01 3.31804529e-02 -1.42450488e+00
7.99781442e-01 -2.53114283e-01 -6.00535810e-01 -5.53131066e-02
-6.17737472e-01 -1.51663050e-01 2.53811389e-01 4.54524457e-01
-7.86039412e-01 5.45381486e-01 1.20068395e+00 -2.23011777e-01
9.51921865e-02 -8.80472921e-03 -8.31659079e-01 1.02004230e+00
6.68026745e-01 -2.71621883e-01 -5.50883353e-01 -3.05473030e-01
-1.67387396e-01 1.22633316e-01 1.37488902e-01 -1.06211984e+00
4.53573585e-01 -2.68410295e-01 6.02422714e-01 -7.58363187e-01
1.52878568e-01 -1.36101019e+00 4.51427668e-01 3.81919116e-01
1.89370215e-01 -1.81604967e-01 -2.19468668e-01 1.16623700e+00
-3.91474128e-01 3.79253365e-02 9.17891204e-01 -1.34026185e-01
-4.67307955e-01 3.53638321e-01 -9.01484638e-02 8.37567225e-02
1.11097932e+00 -3.63429785e-01 -9.13763121e-02 -4.18101311e-01
-7.16079950e-01 7.20081866e-01 7.44543314e-01 3.45438153e-01
5.84020197e-01 -1.84325171e+00 -7.30503440e-01 1.69504210e-01
2.34175354e-01 1.32284716e-01 4.70157206e-01 9.03693080e-01
3.35195884e-02 -6.70068413e-02 -4.88827676e-02 -9.19520438e-01
-1.52466273e+00 7.20077515e-01 1.29326731e-01 -2.02974509e-02
-3.74349028e-01 3.39130193e-01 4.30249929e-01 -3.61468971e-01
-7.02509470e-03 -4.23773043e-02 -2.73663074e-01 2.67363191e-01
5.22383630e-01 5.95295072e-01 -3.65406007e-01 -8.98291171e-01
-3.51542979e-01 1.03691649e+00 -9.92048308e-02 2.80242204e-03
1.23683071e+00 -5.54940045e-01 -2.35214293e-01 7.39334464e-01
1.01998413e+00 3.36362869e-01 -1.32376003e+00 -5.46202719e-01
-1.81189775e-01 -6.95149720e-01 -1.60766140e-01 -1.09781988e-01
-1.40253723e+00 3.85508180e-01 4.41074610e-01 1.62713587e-01
1.24672031e+00 -1.43110901e-01 4.72561896e-01 1.80524632e-01
5.37768960e-01 -1.17751038e+00 -9.77985188e-02 1.49930194e-01
6.52724147e-01 -1.29326451e+00 2.56264776e-01 -6.72885418e-01
-6.68532372e-01 1.00041521e+00 6.27197087e-01 -2.12061405e-01
6.18899763e-01 -3.04037798e-03 1.24448210e-01 -1.13847919e-01
-6.52269244e-01 1.34135693e-01 3.08594644e-01 3.65450501e-01
1.74866065e-01 2.45912194e-01 -3.98206651e-01 6.16407633e-01
4.54701185e-01 -3.35653782e-01 6.34897590e-01 8.70028257e-01
-4.98635590e-01 -1.11490893e+00 -9.68653917e-01 3.78182173e-01
-3.14618081e-01 3.57163638e-01 -2.24478215e-01 3.62280160e-01
-2.64578010e-03 1.21100569e+00 -2.46990323e-01 -4.33468878e-01
2.45047435e-01 -1.02420419e-01 3.69808942e-01 -5.59744358e-01
-8.41273293e-02 6.66504681e-01 -3.57369572e-01 -3.86071086e-01
-8.85654688e-01 -1.02992284e+00 -1.16469717e+00 -4.33883637e-01
-4.67202395e-01 2.83771604e-01 3.15679871e-02 7.01665163e-01
2.59466290e-01 2.67564178e-01 1.18231249e+00 -6.91730082e-01
-7.49194026e-01 -5.96023917e-01 -1.05381548e+00 5.97634733e-01
3.17305438e-02 -6.53646469e-01 -6.42468691e-01 1.49460137e-01] | [8.211053848266602, 4.61764669418335] |
ab0d9158-3fc4-4093-bcd5-c23c3f3dbf1a | verifying-safety-of-neural-networks-from | 2306.15403 | null | https://arxiv.org/abs/2306.15403v1 | https://arxiv.org/pdf/2306.15403v1.pdf | Verifying Safety of Neural Networks from Topological Perspectives | Neural networks (NNs) are increasingly applied in safety-critical systems such as autonomous vehicles. However, they are fragile and are often ill-behaved. Consequently, their behaviors should undergo rigorous guarantees before deployment in practice. In this paper, we propose a set-boundary reachability method to investigate the safety verification problem of NNs from a topological perspective. Given an NN with an input set and a safe set, the safety verification problem is to determine whether all outputs of the NN resulting from the input set fall within the safe set. In our method, the homeomorphism property and the open map property of NNs are mainly exploited, which establish rigorous guarantees between the boundaries of the input set and the boundaries of the output set. The exploitation of these two properties facilitates reachability computations via extracting subsets of the input set rather than the entire input set, thus controlling the wrapping effect in reachability analysis and facilitating the reduction of computation burdens for safety verification. The homeomorphism property exists in some widely used NNs such as invertible residual networks (i-ResNets) and Neural ordinary differential equations (Neural ODEs), and the open map is a less strict property and easier to satisfy compared with the homeomorphism property. For NNs establishing either of these properties, our set-boundary reachability method only needs to perform reachability analysis on the boundary of the input set. Moreover, for NNs that do not feature these properties with respect to the input set, we explore subsets of the input set for establishing the local homeomorphism property and then abandon these subsets for reachability computations. Finally, some examples demonstrate the performance of the proposed method. | ['Wanwei Liu', 'Wenjing Yang', 'Ji Wang', 'Bai Xue', 'Dejin Ren', 'Zhen Liang'] | 2023-06-27 | null | null | null | null | ['autonomous-vehicles'] | ['computer-vision'] | [ 1.92934334e-01 4.65356171e-01 -1.57715783e-01 -1.22298477e-02
2.01006606e-01 -6.98541939e-01 3.26903552e-01 -1.54925972e-01
-1.67717546e-01 7.48376608e-01 -4.27509040e-01 -8.29660416e-01
-4.91239101e-01 -1.00151205e+00 -1.03590238e+00 -9.33308780e-01
-1.51056796e-01 -1.28122687e-01 4.82017666e-01 -4.00387496e-01
-6.50046319e-02 8.61975431e-01 -1.62454200e+00 -6.44765973e-01
8.86164188e-01 1.16194630e+00 -2.85314232e-01 -1.31628411e-02
2.28076205e-01 5.02433062e-01 -8.90528560e-02 2.78217226e-01
6.81197464e-01 -3.19773078e-01 -7.88150072e-01 -3.66772324e-01
-9.90362093e-02 -2.29725957e-01 -5.87445676e-01 1.43135691e+00
-5.52655496e-02 4.35370415e-01 4.87864196e-01 -1.96464121e+00
-1.93503082e-01 4.31443632e-01 -1.47080407e-01 -1.96738467e-01
-2.17013210e-01 2.88867265e-01 7.25651622e-01 -3.78152072e-01
5.76156974e-01 6.86644018e-01 6.43440783e-01 6.33927703e-01
-1.00043809e+00 -9.38894987e-01 1.76214442e-01 -1.17439061e-01
-1.68589604e+00 -4.43452895e-01 7.60153472e-01 -3.84875864e-01
6.62441432e-01 3.92908067e-01 5.62592387e-01 7.46245980e-01
4.94726121e-01 6.13567829e-02 7.67349720e-01 7.75839016e-02
4.24939781e-01 1.88064545e-01 4.21532750e-01 6.05599463e-01
3.94822478e-01 5.10498583e-01 1.77220240e-01 -1.02591179e-01
8.33473623e-01 1.25615120e-01 -5.09915829e-01 -4.80834574e-01
-1.01691437e+00 7.91937590e-01 5.29170573e-01 2.99912274e-01
1.20129611e-03 1.29023939e-01 5.05190253e-01 4.63147134e-01
-1.23346128e-01 1.43734083e-01 -2.21640691e-01 1.19109765e-01
-4.67598766e-01 4.03357521e-02 7.86657989e-01 1.10582602e+00
7.78788209e-01 2.42620438e-01 2.69102365e-01 -8.69383104e-03
2.44982719e-01 3.98601472e-01 7.15870857e-02 -6.69851363e-01
1.96490422e-01 9.00818348e-01 -4.63473350e-02 -1.14446735e+00
-5.35480618e-01 -1.32597476e-01 -1.28210080e+00 5.06933749e-01
4.93781537e-01 -1.66749313e-01 -6.42853439e-01 2.17353272e+00
5.45708001e-01 3.24826241e-01 4.77409780e-01 7.51731336e-01
3.94643635e-01 7.82384813e-01 -2.86496252e-01 -3.77094001e-01
9.76706564e-01 -3.76619339e-01 -4.23859209e-01 2.23112047e-01
7.97287524e-01 2.26022959e-01 8.55396509e-01 -1.29216745e-01
-7.33143091e-01 -2.57142067e-01 -1.56453884e+00 2.48206764e-01
-3.92146766e-01 -1.67972699e-01 -2.52903141e-02 1.99129254e-01
-1.04270732e+00 8.04014444e-01 -8.00173342e-01 -1.06428415e-01
5.61203398e-02 7.71831572e-01 -5.86284161e-01 4.06330407e-01
-1.66749322e+00 7.76022434e-01 5.03761411e-01 4.36349869e-01
-6.69088900e-01 -6.05381072e-01 -9.21036422e-01 2.49210879e-01
5.72947204e-01 -5.69674596e-02 5.96090376e-01 -8.25338840e-01
-1.38441956e+00 1.67198107e-01 3.40952873e-01 -4.92094070e-01
6.51091039e-01 3.77080113e-01 -4.44779247e-01 6.26003146e-02
-9.85645354e-02 3.09578538e-01 7.99316585e-01 -8.63969743e-01
-2.57675588e-01 -2.19895750e-01 4.41565573e-01 -2.73767650e-01
-5.70259333e-01 -2.25914285e-01 1.23073436e-01 -1.24006532e-01
1.80741593e-01 -1.04263186e+00 -3.60628009e-01 1.77233502e-01
-7.48219311e-01 -1.65315509e-01 1.27364457e+00 -2.84960598e-01
1.19013321e+00 -2.54347801e+00 9.53511298e-02 8.62027287e-01
2.88682640e-01 4.40974802e-01 1.43902391e-01 2.60427594e-01
-1.03061326e-01 2.52996027e-01 -3.65386248e-01 4.00781631e-01
-6.32353127e-02 1.75939843e-01 -5.31626582e-01 9.14159417e-01
2.62627453e-01 5.65003693e-01 -6.24077618e-01 -3.00439745e-01
1.33226037e-01 4.21212971e-01 -3.58412147e-01 -7.91981965e-02
-1.03676297e-01 5.04437506e-01 -6.51448250e-01 5.21839932e-02
5.98030388e-01 2.38421947e-01 -7.89114982e-02 -5.36910631e-02
-4.52323854e-01 -1.07463956e-01 -1.41902888e+00 6.86827004e-01
-3.48444372e-01 4.37945187e-01 2.42062345e-01 -8.44648302e-01
1.11260009e+00 4.19308156e-01 3.95543844e-01 -4.86139119e-01
5.89468658e-01 2.46223450e-01 1.87719882e-01 -1.76536605e-01
9.15627033e-02 -1.26845539e-01 -1.85885102e-01 4.63417232e-01
-3.41816187e-01 2.43109882e-01 -9.52862278e-02 -4.29003239e-02
1.07684767e+00 -2.29370311e-01 2.07598299e-01 -9.17566419e-01
1.03500617e+00 -1.14504144e-01 8.90705526e-01 3.25166255e-01
-2.69334346e-01 1.69160411e-01 9.41190183e-01 -3.99679542e-01
-9.41358864e-01 -7.20186949e-01 -2.42627501e-01 1.02332599e-01
6.67838097e-01 -5.35044298e-02 -8.59713018e-01 -7.16211677e-01
-1.19781539e-01 4.48432773e-01 -6.57926083e-01 -9.05428827e-01
-6.49505556e-01 2.90562034e-01 1.04665375e+00 6.82026327e-01
6.02243543e-01 -9.17539060e-01 -9.57835078e-01 -1.55969277e-01
1.55970573e-01 -9.30589914e-01 -5.06291747e-01 2.01239362e-01
-8.30248773e-01 -1.24840713e+00 -1.82495490e-01 -8.78439128e-01
1.14274907e+00 -4.73244451e-02 2.10956112e-02 3.02765995e-01
1.46509692e-01 -1.13503017e-01 2.27243245e-01 -3.57915387e-02
-5.99808693e-01 -2.13814676e-02 5.53604066e-01 1.60327256e-01
-2.86629915e-01 -6.35416329e-01 -1.46926686e-01 8.92370939e-01
-1.09060657e+00 1.00734634e-02 1.47237048e-01 6.69786513e-01
6.50202751e-01 6.51017725e-01 7.74785757e-01 -3.83121341e-01
3.74083757e-01 -4.67685103e-01 -1.22178423e+00 1.89340487e-01
-5.06652176e-01 3.21891874e-01 1.29185665e+00 -7.76629031e-01
-5.83050311e-01 1.99967831e-01 1.44475430e-01 -8.38535607e-01
3.35120596e-02 3.26571822e-01 -7.24056900e-01 -2.55099654e-01
4.00574774e-01 1.00034304e-01 5.21885931e-01 -3.93213797e-03
-6.70627207e-02 4.67150867e-01 5.08332968e-01 -4.14972305e-01
1.18469107e+00 4.44765151e-01 6.47848487e-01 -7.49646902e-01
8.08221698e-02 -5.05857691e-02 -4.31631118e-01 -3.53417575e-01
4.82337296e-01 -3.60082835e-01 -1.20448518e+00 4.51246470e-01
-1.01131737e+00 -2.32046634e-01 -4.26638186e-01 2.12953165e-01
-4.81839836e-01 5.58382161e-02 -3.38525563e-01 -1.01103473e+00
-3.25130820e-01 -1.44174826e+00 5.34694493e-01 2.93222398e-01
-1.48026809e-01 -8.26861799e-01 -2.64211178e-01 -4.85132217e-01
4.24773604e-01 6.19017303e-01 1.13170433e+00 -8.65853608e-01
-4.04335529e-01 -4.43757743e-01 -1.27838582e-01 3.16911072e-01
-1.67955142e-02 1.71091571e-01 -5.73599517e-01 -3.69264513e-01
2.98067003e-01 2.43911833e-01 3.88399303e-01 6.77136183e-02
6.88876808e-01 -7.28845239e-01 -4.12440568e-01 5.08201540e-01
1.46951056e+00 5.76862991e-01 6.69525266e-01 1.51040956e-01
6.92763090e-01 8.53961051e-01 5.04364133e-01 2.14091893e-02
-1.48714617e-01 2.97647208e-01 6.86387718e-01 2.19975144e-01
5.59439361e-01 -1.33270711e-01 7.12915421e-01 4.39929605e-01
6.24716580e-02 4.31261174e-02 -8.52909029e-01 5.22495031e-01
-1.82494724e+00 -5.74450374e-01 -6.65873662e-02 2.40381694e+00
6.94528878e-01 2.21645519e-01 -2.36980002e-02 5.31557262e-01
1.22113347e+00 -1.11840479e-01 -7.97356606e-01 -5.35623252e-01
1.80271029e-01 -2.76310295e-01 5.60329556e-01 4.55398232e-01
-7.82109380e-01 5.72617769e-01 4.79773045e+00 7.23731875e-01
-1.32229912e+00 -2.09657818e-01 3.67285520e-01 1.56301320e-01
-2.47731730e-01 2.86941856e-01 -7.23778605e-01 3.16065669e-01
9.43617165e-01 -2.56809592e-01 3.75089645e-01 8.24171960e-01
2.09434584e-01 8.69377851e-02 -1.26281190e+00 2.89384484e-01
-4.24832016e-01 -9.98736501e-01 -3.14761579e-01 2.95090437e-01
5.29211640e-01 -3.99644405e-01 2.66277920e-02 1.61147453e-02
-1.66091785e-01 -9.25754189e-01 8.60444427e-01 2.12761492e-01
9.52089548e-01 -1.17837822e+00 6.21163607e-01 4.41137016e-01
-1.35978818e+00 -9.20142084e-02 -1.52322367e-01 -4.19095531e-02
8.18812549e-02 2.84502774e-01 -1.85432106e-01 5.61510384e-01
5.27860463e-01 4.66795623e-01 -7.68261999e-02 6.18236542e-01
-1.27591968e-01 2.14817792e-01 -5.82162917e-01 -4.25481610e-02
4.42609906e-01 -5.38661480e-01 8.50394607e-01 4.13720250e-01
1.46920308e-01 1.33050516e-01 -2.03361958e-01 1.20953107e+00
5.14750108e-02 -1.91345289e-01 -1.18523145e+00 2.21396685e-02
4.78006512e-01 9.80702400e-01 -8.83723259e-01 -5.33568971e-02
-1.32603928e-01 2.00629324e-01 -3.41631249e-02 4.55412120e-01
-1.28104651e+00 -9.41386282e-01 1.03024065e+00 1.02402963e-01
1.10929854e-01 -5.38525246e-02 -4.46948469e-01 -6.25573695e-01
3.17702532e-01 -4.55998659e-01 1.11174332e-02 -3.04824561e-01
-5.61030746e-01 8.61409962e-01 1.83699980e-01 -1.68880308e+00
-2.06174374e-01 -3.83616447e-01 -8.92287016e-01 6.60090268e-01
-1.08476007e+00 -7.45725930e-01 -1.16667271e-01 8.29915404e-01
-2.90044129e-01 1.87289014e-01 4.49643105e-01 1.31443426e-01
-8.91517937e-01 5.87916911e-01 -7.33343372e-03 2.10345149e-01
1.68067496e-02 -4.31979328e-01 7.76866451e-02 1.23983657e+00
-5.74440420e-01 9.69828129e-01 5.73114991e-01 -6.78396583e-01
-1.46512961e+00 -1.44976783e+00 5.83956957e-01 1.07786305e-01
8.02822948e-01 -3.98984283e-01 -1.01884413e+00 7.63218522e-01
-2.50315130e-01 1.86495021e-01 -1.22812800e-01 -5.77670455e-01
-3.25224638e-01 -1.47094101e-01 -1.39693296e+00 9.86970663e-01
1.06391847e+00 -4.13845420e-01 -2.22930342e-01 -2.88622677e-01
9.66880918e-01 -2.67398894e-01 -8.39634180e-01 8.33570242e-01
5.60927391e-01 -5.89156747e-01 6.80210114e-01 -5.27083397e-01
2.40561247e-01 -7.69114196e-01 9.87590700e-02 -9.67600703e-01
1.08178988e-01 -8.02282810e-01 -1.25311077e-01 1.34442008e+00
3.10270727e-01 -1.21813154e+00 3.01911324e-01 8.30708325e-01
-2.75360048e-01 -9.09556925e-01 -1.29914093e+00 -1.34236133e+00
1.26923159e-01 -1.97410986e-01 8.84532154e-01 7.51747549e-01
3.48363221e-01 6.76472764e-03 -8.80625993e-02 4.88748759e-01
2.39002809e-01 4.21050414e-02 2.53193170e-01 -1.32492626e+00
1.83443695e-01 -3.88096243e-01 -6.47967696e-01 -4.21468347e-01
5.41340828e-01 -6.91693664e-01 4.64484274e-01 -1.08513284e+00
-4.18263346e-01 -8.58885884e-01 -3.22848916e-01 8.00277233e-01
4.70558584e-01 -5.30900434e-02 -1.43129444e-02 2.19871938e-01
-5.24577908e-02 5.77244461e-01 1.11128640e+00 -1.25934854e-01
-5.89678109e-01 1.23276673e-01 -3.80376250e-01 6.94246471e-01
9.69736934e-01 -3.64954442e-01 -7.51717746e-01 1.26635119e-01
1.20326325e-01 2.17307791e-01 5.08869410e-01 -1.04812300e+00
4.42671806e-01 -4.50870454e-01 -4.41880107e-01 -2.51932800e-01
8.17664415e-02 -1.40355420e+00 6.48162067e-01 8.13571692e-01
-1.26746029e-01 -2.59350799e-02 3.36509377e-01 5.33585370e-01
-1.29326716e-01 -1.96196556e-01 9.21838403e-01 5.52212715e-01
-4.53795791e-01 3.58243048e-01 -4.88848448e-01 -8.19018334e-02
1.57613516e+00 -4.54720914e-01 -2.74519891e-01 -8.34275782e-02
-4.90697384e-01 4.07139868e-01 4.71709937e-01 4.34909374e-01
5.90662658e-01 -1.35000479e+00 -4.14238162e-02 8.00018966e-01
-3.26077044e-02 4.32193220e-01 1.73704177e-01 1.24186003e+00
-2.98521578e-01 5.15041232e-01 -3.25355411e-01 -4.79655445e-01
-1.02373195e+00 7.65744328e-01 5.58344662e-01 1.60589725e-01
-8.63445938e-01 2.82570451e-01 3.03692430e-01 -3.38862032e-01
3.56739342e-01 -7.36047089e-01 -3.35745931e-01 -1.83233738e-01
3.39496464e-01 6.31486535e-01 -6.60219565e-02 -8.04918349e-01
-5.44724762e-01 4.99710262e-01 3.59691441e-01 -4.06063981e-02
1.11025858e+00 3.59554626e-02 -2.88138539e-01 9.76882204e-02
1.46512997e+00 -2.94523418e-01 -1.23150659e+00 1.47547245e-01
-2.71831781e-01 9.76667553e-02 -1.14332467e-01 2.72433534e-02
-1.49753714e+00 5.95683157e-01 7.92528912e-02 2.75714397e-01
1.19642484e+00 -1.95612386e-01 8.86412919e-01 4.45335299e-01
4.67917502e-01 -8.90452921e-01 -6.04637384e-01 6.82231069e-01
6.81042254e-01 -4.15753037e-01 -5.29610872e-01 -4.98686522e-01
-2.85876423e-01 1.05334580e+00 8.96129370e-01 -3.58835131e-01
8.94314468e-01 5.30657351e-01 -4.28497583e-01 8.18395168e-02
-4.92837161e-01 1.74721062e-01 2.73734368e-02 2.84884393e-01
-6.43526733e-01 -2.39456266e-01 -2.06465602e-01 8.65828812e-01
-3.26608941e-02 -1.57563701e-01 6.30427480e-01 9.77823615e-01
-3.75782073e-01 -4.95196283e-01 -2.19830617e-01 2.59100437e-01
1.28683686e-01 1.18123740e-01 -3.31647724e-01 1.01762843e+00
-9.28426103e-04 8.37145865e-01 4.95482609e-02 -8.35157692e-01
4.27695811e-01 5.81551529e-02 -9.84881539e-03 -1.68067321e-01
-4.30998921e-01 -4.32367474e-01 2.37447210e-03 -6.41634703e-01
-3.29971388e-02 -3.88960660e-01 -1.99387360e+00 -6.30516052e-01
-8.53778183e-01 2.56892443e-01 4.31767374e-01 1.15167356e+00
4.49671358e-01 4.58960414e-01 7.59633601e-01 -3.48456532e-01
-7.43866026e-01 -3.82543355e-01 -6.50167346e-01 -1.74642041e-01
6.40481293e-01 -7.34978914e-01 -7.17609704e-01 -2.30787039e-01] | [6.09644079208374, 7.645134925842285] |
8d80db8f-cf3a-4780-ac07-592332bc52f2 | change-detection-in-vhr-imagery-with-severe | null | null | https://ieeexplore.ieee.org/document/9740657 | https://ieeexplore.ieee.org/document/9740657 | Change Detection in VHR Imagery With Severe Co-Registration Errors Using Deep Learning: A Comparative Study | Change detection (CD) through Earth observation techniques can offer very significant information for monitoring tasks in a time-efficient manner. Very high-resolution (VHR) images can display objects in fine detail, thus making it possible to rapidly perceive isolated changes. However, this is a challenging task because of the increased within-class variance and geometric registration errors caused by different satellite view directions and angles. Lately, deep learning (DL) CD methods have proven very appealing for the CD problem because of their flexibility to combine and process different types of information along with the increased availability of higher processing power systems. Even though previous research has developed several notable DL methodologies, it has mostly focused on images with minor co-registration errors. Based on that, the goal of this study is to evaluate the performance of five state-of-the-art DL CD methods, two unsupervised and three supervised, on VHR images with severe co-registration errors. The methods are implemented on four urban European areas of versatile morphology. In addition, before applying the CD process, four popular automatic co-registration methods were evaluated because of the importance of this pre-processing step for the successful output of the CD problem. It was shown that phase correlation used on the Fourier-Mellin Transform produced the most satisfactory co-registration results and STANet detected building-related changes most successfully. Its success can be attributed to its particular attention mechanism and its training dataset. The rest of the co-registration and CD methods showed low performance. | ['Vassilia Karathanassi', 'Viktoria Kristollari'] | 2022-03-24 | null | null | null | ieee-access-2022-3 | ['change-detection-for-remote-sensing-images'] | ['miscellaneous'] | [-3.61233577e-02 -5.06369889e-01 3.74375165e-01 -2.40210593e-01
-6.57242119e-01 -2.01244831e-01 9.32018101e-01 3.15097868e-01
-6.61314487e-01 6.03415668e-01 -1.64652139e-01 9.03883204e-02
-4.35960531e-01 -1.15392208e+00 -1.95762679e-01 -8.86385798e-01
-5.07580340e-01 5.54967344e-01 2.91381419e-01 -6.72955573e-01
-4.48319428e-02 1.13349426e+00 -1.99057221e+00 -6.43651560e-02
9.08361316e-01 6.98096991e-01 5.96307516e-01 4.52352643e-01
1.55171871e-01 3.16669106e-01 -4.24331397e-01 1.64412498e-01
1.97309196e-01 -1.65904880e-01 -4.77048248e-01 1.48829641e-02
6.48783624e-01 -1.36587560e-01 -6.26176000e-02 9.63319600e-01
8.67963135e-01 2.65854806e-01 6.04662538e-01 -6.48408711e-01
-2.38912508e-01 9.50443074e-02 -5.72657585e-01 4.40950722e-01
5.26930213e-01 -8.85442942e-02 8.31205070e-01 -9.23604488e-01
7.97736287e-01 8.83404732e-01 9.16920424e-01 -6.76293746e-02
-1.50709057e+00 -5.50495028e-01 -3.09001446e-01 4.66436595e-01
-1.44785023e+00 -2.89087981e-01 8.08264315e-01 -6.77809060e-01
1.11798334e+00 3.67396832e-01 8.05893719e-01 4.45205748e-01
2.15192825e-01 2.75396645e-01 1.57628846e+00 -4.38368142e-01
9.59008411e-02 -7.88854584e-02 2.39352603e-02 3.02406073e-01
2.89543927e-01 2.44927928e-01 -1.24399737e-01 1.51856586e-01
6.41310453e-01 -1.50559127e-01 -5.48886836e-01 -2.17565283e-01
-1.00527763e+00 8.11530113e-01 6.92970574e-01 9.06176388e-01
-5.25416553e-01 -2.29623288e-01 6.85981140e-02 5.07987797e-01
7.12134600e-01 4.72503066e-01 -3.21230948e-01 1.60105392e-01
-1.16166937e+00 1.63438499e-01 4.48991746e-01 8.35058987e-02
9.33359325e-01 3.20856094e-01 2.61814117e-01 8.95485938e-01
3.46901536e-01 9.24756825e-01 4.37736928e-01 -3.21257114e-01
1.44122899e-01 3.78372341e-01 -1.77827403e-01 -1.53513169e+00
-8.16943884e-01 -4.79331940e-01 -1.14880204e+00 6.08562589e-01
2.69202560e-01 2.45738178e-01 -8.51406455e-01 1.24640167e+00
2.47900829e-01 -2.02279508e-01 7.03325123e-02 8.84644687e-01
1.02943623e+00 7.04160511e-01 -1.91162705e-01 -1.44515097e-01
1.14930964e+00 -1.59198567e-01 -8.26640189e-01 -2.42971271e-01
3.38826984e-01 -8.52587104e-01 8.43687057e-01 2.58125693e-01
-5.60709417e-01 -7.30548203e-01 -1.15276432e+00 3.16945881e-01
-5.97288311e-01 2.31509954e-01 4.49549675e-01 4.29534614e-01
-1.14726043e+00 8.01802516e-01 -9.27187085e-01 -6.34977639e-01
1.46166176e-01 3.23783070e-01 -6.37967110e-01 5.44797517e-02
-1.11455655e+00 1.18981087e+00 2.41611242e-01 4.31292981e-01
-3.05102378e-01 -5.05292952e-01 -8.06514204e-01 -1.51994228e-01
-5.37842289e-02 -2.14500204e-01 5.32163978e-01 -7.93279111e-01
-1.35113168e+00 1.21126664e+00 1.50727406e-01 -4.71831679e-01
6.80547297e-01 -1.49017617e-01 -7.07840621e-01 2.06306711e-01
2.31168512e-02 2.02239484e-01 7.47173131e-01 -1.28209674e+00
-5.40094852e-01 -4.72692102e-01 -2.95922607e-01 7.10622445e-02
3.51164863e-02 -3.55280861e-02 -1.22283071e-01 -5.74993968e-01
6.01978481e-01 -8.98514330e-01 -1.28770277e-01 -2.05495596e-01
1.73037261e-01 2.46317759e-01 8.79013598e-01 -1.10591829e+00
1.03000617e+00 -2.18169880e+00 7.22345188e-02 3.87893260e-01
3.24152373e-02 4.72291142e-01 1.54208079e-01 4.44158524e-01
-2.89774954e-01 -1.52286991e-01 -4.38685268e-01 -6.55667111e-02
-1.81970477e-01 1.78149879e-01 -6.19770177e-02 8.99396539e-01
7.96092115e-03 3.41056347e-01 -6.19022012e-01 -2.83249110e-01
8.09719443e-01 7.77434409e-01 -7.98516870e-02 -2.86860857e-02
1.23698883e-01 5.79027057e-01 5.33050857e-02 4.19229567e-01
9.55043733e-01 2.87350893e-01 -4.76479009e-02 -4.43353385e-01
-5.82560301e-01 -8.48411769e-02 -1.64561570e+00 1.07690644e+00
-4.83632237e-01 1.16106129e+00 4.17914204e-02 -1.02207971e+00
1.34275115e+00 3.20535004e-01 5.20771861e-01 -1.19756842e+00
-1.54824555e-01 5.21096349e-01 -3.03342636e-03 -5.82040966e-01
6.04235590e-01 -2.58511424e-01 3.13968241e-01 -1.39128659e-02
-2.70078897e-01 -4.35099036e-01 2.91892625e-02 -3.07535529e-01
4.98341113e-01 1.15696020e-01 5.25772035e-01 -3.05408925e-01
6.98719203e-01 1.23446863e-02 3.60802978e-01 4.11636204e-01
8.86374861e-02 8.52540970e-01 -2.58410484e-01 -6.15250826e-01
-8.38930666e-01 -9.36424673e-01 -5.45983732e-01 4.14644152e-01
1.04257219e-01 6.10587671e-02 -1.08708084e-01 6.80713058e-02
-1.38725042e-01 5.46338081e-01 -4.28149253e-01 2.61754334e-01
-6.31959736e-01 -1.07496750e+00 3.48970234e-01 1.60973758e-01
9.41250741e-01 -1.04408813e+00 -8.98669481e-01 4.51085508e-01
-3.60900313e-01 -9.72166121e-01 5.27397454e-01 1.41901240e-01
-1.06159139e+00 -1.05022585e+00 -6.42777681e-01 -6.24548793e-01
3.47090364e-01 3.35300386e-01 9.98625994e-01 -1.53512642e-01
-5.04563510e-01 2.87756830e-01 -3.67387742e-01 -3.86372358e-02
-4.17027980e-01 -1.63244501e-01 -5.96288629e-02 -1.26134381e-02
1.58567354e-01 -7.68455327e-01 -2.86932617e-01 2.96153843e-01
-9.04820859e-01 -2.76365608e-01 5.38460791e-01 7.01834202e-01
7.44520485e-01 3.43122602e-01 5.42116761e-02 -6.28536522e-01
2.76427299e-01 -2.92375889e-02 -7.34217942e-01 9.15730149e-02
-5.51875412e-01 -1.26075789e-01 3.46033335e-01 -1.33366480e-01
-1.02919137e+00 -7.64159411e-02 -4.80477631e-01 1.14748299e-01
-4.61531639e-01 7.57195175e-01 -1.46909848e-01 -3.49023849e-01
8.27371478e-01 1.90141410e-01 -4.70297821e-02 -5.17856956e-01
-8.05143043e-02 5.57221651e-01 5.42287707e-01 -1.63855210e-01
1.03419697e+00 8.11994791e-01 1.44448295e-01 -1.66055536e+00
-3.59410793e-01 -7.41169393e-01 -9.41187978e-01 -3.99536192e-01
9.71764445e-01 -1.06980562e+00 -8.21913630e-02 8.51779044e-01
-8.06574345e-01 -2.65029103e-01 -2.66295582e-01 5.83136618e-01
-2.46424228e-01 4.73019272e-01 -2.15361580e-01 -6.23051107e-01
-2.67348170e-01 -1.02192092e+00 8.13534021e-01 1.71248272e-01
4.52341102e-02 -1.14014912e+00 4.69658822e-01 1.59576368e-02
6.76728070e-01 8.43051970e-01 5.48376203e-01 -1.74291268e-01
-3.81990999e-01 -2.83581495e-01 -6.54420555e-02 2.20900834e-01
3.02810311e-01 2.03526184e-01 -1.08821452e+00 -5.27248681e-01
-3.84597145e-02 2.84281522e-01 8.72517347e-01 5.61590791e-01
3.92519861e-01 1.90734759e-01 -2.00298369e-01 8.16405714e-01
1.88446879e+00 8.10512751e-02 9.72392261e-01 1.03123677e+00
5.07794023e-01 4.92016166e-01 6.98696375e-01 3.95906776e-01
1.78483263e-01 1.22919261e+00 4.54690725e-01 -5.90799928e-01
-3.39881003e-01 3.74964654e-01 3.03959638e-01 7.21842945e-01
-6.14349663e-01 1.71901748e-01 -1.11593187e+00 4.69633996e-01
-1.51852858e+00 -1.34525990e+00 -8.59291673e-01 2.30679822e+00
3.83067518e-01 1.91573985e-02 -3.10269088e-01 5.51204622e-01
5.36918938e-01 5.11962175e-01 4.72509116e-02 -1.05621859e-01
-7.01073527e-01 3.98631901e-01 5.24102271e-01 4.48093414e-01
-1.31411862e+00 5.75807810e-01 5.40570641e+00 6.60424232e-01
-1.52609193e+00 4.76463176e-02 -4.00000736e-02 4.86416847e-01
-2.71152742e-02 -2.52642214e-01 -8.03053617e-01 2.10434064e-01
6.20773494e-01 2.48687059e-01 1.29562840e-01 5.55719078e-01
4.54130769e-01 -5.51823437e-01 -5.12803435e-01 1.19551075e+00
9.90169346e-02 -1.09768307e+00 -1.26783535e-01 1.24217026e-01
7.96294570e-01 3.27259302e-01 -1.05643809e-01 -2.21066177e-02
-9.99010280e-02 -8.62257421e-01 5.50484478e-01 6.34117484e-01
6.75953090e-01 -7.35773325e-01 1.00627482e+00 2.48294454e-02
-1.51453352e+00 5.04219308e-02 -3.81299943e-01 -2.80482143e-01
2.08290964e-01 9.46415365e-01 -6.02701902e-01 9.71698642e-01
9.44660366e-01 7.53407836e-01 -8.60087931e-01 1.27414787e+00
-4.22696024e-01 3.63303423e-01 -6.30726039e-01 4.17797595e-01
1.47699863e-01 -4.64739919e-01 7.19469190e-01 1.43292701e+00
5.37928700e-01 3.80313322e-02 2.61060297e-02 5.67923129e-01
5.91246068e-01 2.52298504e-01 -7.37087846e-01 4.06902194e-01
2.09091902e-01 1.36974049e+00 -9.71985877e-01 -2.86002997e-02
-2.86137313e-01 7.32298970e-01 -6.65816218e-02 1.46005020e-01
-5.86870313e-01 -2.98589170e-01 5.22046387e-01 4.93000984e-01
3.83311331e-01 -5.70249021e-01 -1.95828378e-01 -9.47056055e-01
-5.47549278e-02 -7.81764030e-01 4.05876398e-01 -8.37446451e-01
-9.01308298e-01 7.12485194e-01 1.01051219e-02 -1.43758845e+00
-4.48711738e-02 -3.82488221e-01 -4.86675680e-01 8.58133018e-01
-1.80903256e+00 -1.19377315e+00 -7.88479090e-01 5.57369888e-01
2.79641837e-01 -7.29114339e-02 8.48261356e-01 6.11934543e-01
-2.66714275e-01 5.80555871e-02 5.62980056e-01 -6.87258318e-02
6.81492209e-01 -1.34190857e+00 1.76695243e-01 1.05494225e+00
1.68894976e-01 7.16135129e-02 9.21986520e-01 -5.30784607e-01
-8.52845967e-01 -8.80581498e-01 9.87901330e-01 -6.32356713e-03
3.99405658e-01 6.29622117e-02 -1.19233191e+00 4.78504270e-01
-1.47917300e-01 -8.68403986e-02 4.71563697e-01 -1.35060504e-01
5.75337522e-02 -3.21712106e-01 -1.19460917e+00 2.22023591e-01
4.19914335e-01 -5.42791545e-01 -7.74339139e-01 4.61492948e-02
-2.27523055e-02 -4.37560439e-01 -9.37463284e-01 6.17819846e-01
3.83343369e-01 -1.27522635e+00 1.20832443e+00 3.55109334e-01
2.19968885e-01 -6.41361296e-01 -2.54409164e-01 -1.37236404e+00
-4.50587988e-01 9.29239094e-02 4.62293565e-01 1.28082156e+00
1.36818200e-01 -8.25872779e-01 1.35707319e-01 1.13836445e-01
-1.20213300e-01 -3.76926512e-02 -9.60272908e-01 -7.42465734e-01
-3.69176000e-01 -4.14436936e-01 3.27659339e-01 1.28638220e+00
-6.16358399e-01 9.06272680e-02 -2.03424484e-01 6.47728145e-01
6.22230709e-01 4.15623039e-01 8.85021925e-01 -1.85619450e+00
-2.41263151e-01 -3.59567583e-01 -6.82586968e-01 -3.58639538e-01
-2.55088538e-01 -7.85067022e-01 1.19917020e-01 -1.85210514e+00
-2.53034085e-01 -5.57888150e-01 1.79968297e-01 2.61763811e-01
3.54006141e-02 4.98048544e-01 1.49331689e-01 4.91465926e-01
9.24896598e-02 5.05715191e-01 1.00643730e+00 -7.60767013e-02
-4.54853743e-01 1.65888131e-01 2.03097731e-01 7.78778374e-01
6.90393090e-01 -3.64861876e-01 1.76420406e-01 -3.96672487e-01
3.31738889e-01 -2.42280826e-01 5.23458004e-01 -1.47997069e+00
9.47959870e-02 3.08080286e-01 5.00108063e-01 -9.18624699e-01
2.34129697e-01 -1.00406194e+00 8.75903130e-01 6.21436417e-01
2.84386367e-01 4.01004925e-02 2.32539132e-01 2.18645617e-01
-6.58598423e-01 -1.82866216e-01 1.29912984e+00 -1.05635174e-01
-1.23648179e+00 -2.56159957e-02 -5.56679249e-01 -2.87021220e-01
8.91257942e-01 -5.35553753e-01 -6.05552383e-02 -2.52699018e-01
-9.24876392e-01 -2.48512045e-01 5.61078668e-01 7.15734661e-02
4.43349302e-01 -1.08704364e+00 -7.80447066e-01 3.28539371e-01
6.62604943e-02 7.85314143e-02 4.54933763e-01 1.04748130e+00
-1.01379216e+00 1.46880016e-01 -5.77376068e-01 -9.00524080e-01
-1.51614630e+00 7.98852593e-02 6.17215931e-01 -4.22959059e-01
-1.02486825e+00 4.39548641e-01 -2.94642001e-01 -5.22676110e-01
-1.18209876e-01 -3.00031602e-01 -7.27661967e-01 6.63384199e-01
4.46318567e-01 5.96791148e-01 5.58413684e-01 -1.04918873e+00
-4.33535963e-01 1.17783582e+00 3.65855992e-01 -1.45892546e-01
1.78326190e+00 -1.77299723e-01 -1.79413646e-01 4.16739136e-01
9.87142980e-01 2.41623130e-02 -9.09628630e-01 -2.23676339e-01
1.92048382e-02 -6.03581011e-01 4.45549935e-01 -5.58896303e-01
-1.15279579e+00 9.25095677e-01 1.23702121e+00 3.09721082e-01
1.19999313e+00 -3.66334766e-01 3.58069271e-01 2.58662730e-01
3.95688355e-01 -1.18636191e+00 -1.62367031e-01 4.72231537e-01
8.50087643e-01 -1.32758582e+00 4.96789783e-01 -2.50556618e-01
-1.98407501e-01 1.38430452e+00 1.39568880e-01 -4.36726697e-02
7.44209051e-01 1.67604312e-01 1.37478188e-01 -4.53162014e-01
1.48938388e-01 -5.96790910e-01 2.46861279e-01 8.99670243e-01
4.17755634e-01 1.61872268e-01 -5.13200104e-01 -7.94090703e-02
-2.80941576e-01 -1.25090823e-01 5.91173410e-01 8.59089494e-01
-6.16697669e-01 -7.25655079e-01 -8.36013734e-01 1.94157138e-01
-3.12700599e-01 1.76613390e-01 2.43893154e-02 1.24360049e+00
8.88920650e-02 7.89585948e-01 1.48190111e-01 -2.13404030e-01
5.77726126e-01 -3.35699499e-01 3.87980372e-01 -2.40860343e-01
-6.38458848e-01 7.57729709e-02 1.30495787e-01 -4.42173332e-01
-9.72510934e-01 -9.57831621e-01 -1.05797482e+00 -2.32446998e-01
-4.07286674e-01 -5.19393608e-02 8.09879720e-01 9.08120215e-01
-1.31013781e-01 5.47351241e-01 6.94370031e-01 -1.21510816e+00
-1.50984809e-01 -1.01621687e+00 -8.55464101e-01 4.79092926e-01
3.39699835e-01 -8.17969203e-01 -5.19599199e-01 4.96379733e-02] | [9.781009674072266, -1.8983839750289917] |
e4f5689b-f0c0-40d7-afe4-3b4cb867ea00 | verifying-fairness-in-quantum-machine | 2207.11173 | null | https://arxiv.org/abs/2207.11173v1 | https://arxiv.org/pdf/2207.11173v1.pdf | Verifying Fairness in Quantum Machine Learning | Due to the beyond-classical capability of quantum computing, quantum machine learning is applied independently or embedded in classical models for decision making, especially in the field of finance. Fairness and other ethical issues are often one of the main concerns in decision making. In this work, we define a formal framework for the fairness verification and analysis of quantum machine learning decision models, where we adopt one of the most popular notions of fairness in the literature based on the intuition -- any two similar individuals must be treated similarly and are thus unbiased. We show that quantum noise can improve fairness and develop an algorithm to check whether a (noisy) quantum machine learning model is fair. In particular, this algorithm can find bias kernels of quantum data (encoding individuals) during checking. These bias kernels generate infinitely many bias pairs for investigating the unfairness of the model. Our algorithm is designed based on a highly efficient data structure -- Tensor Networks -- and implemented on Google's TensorFlow Quantum. The utility and effectiveness of our algorithm are confirmed by the experimental results, including income prediction and credit scoring on real-world data, for a class of random (noisy) quantum decision models with 27 qubits ($2^{27}$-dimensional state space) tripling ($2^{18}$ times more than) that of the state-of-the-art algorithms for verifying quantum machine learning models. | ['Mingsheng Ying', 'Wang Fang', 'Ji Guan'] | 2022-07-22 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [-5.10529941e-03 1.96146563e-01 -1.43631101e-01 -6.12438202e-01
-3.03879797e-01 -3.27518970e-01 2.52005965e-01 3.91733050e-01
-6.01965725e-01 9.03585970e-01 -3.80284965e-01 -6.27138078e-01
-3.51696849e-01 -1.25501370e+00 -4.55360264e-01 -8.17150354e-01
-2.48731464e-01 5.64376414e-01 -3.43544066e-01 -4.12525296e-01
6.04229093e-01 1.75728038e-01 -1.45142722e+00 1.91433951e-02
1.02626061e+00 1.06033623e+00 -1.02101409e+00 6.16137266e-01
2.59904206e-01 9.65391576e-01 -1.04844503e-01 -1.48956513e+00
5.72844803e-01 -5.54100096e-01 -1.12370241e+00 -5.46196043e-01
2.83801764e-01 -4.79118943e-01 -6.05111957e-01 1.84341276e+00
4.88993317e-01 1.70874242e-02 6.03823304e-01 -1.45701528e+00
-8.45180988e-01 8.71314168e-01 -1.03247352e-02 1.68075889e-01
-4.18384299e-02 2.43033603e-01 1.53046262e+00 -2.88947970e-02
3.23768318e-01 1.08075905e+00 4.18990314e-01 7.87651122e-01
-1.37779057e+00 -9.32891846e-01 -8.10302019e-01 5.94100833e-01
-1.51139212e+00 -5.18600106e-01 3.48491669e-01 -4.35451627e-01
6.75173283e-01 2.82513410e-01 4.77508754e-01 7.27517009e-01
3.91867250e-01 2.01623678e-01 1.68012905e+00 -3.79558742e-01
6.69160008e-01 2.53225058e-01 6.05444372e-01 9.26054418e-01
6.40338242e-01 8.88845980e-01 -6.26494348e-01 -3.97163987e-01
3.73777390e-01 -1.81368917e-01 2.98856914e-01 -9.67049077e-02
-7.97080159e-01 1.15230250e+00 4.11784321e-01 -6.91743521e-03
-1.93843275e-01 4.41277474e-01 4.34446514e-01 6.32713854e-01
2.23526105e-01 3.47868264e-01 -5.62479421e-02 -3.19310069e-01
-6.13217473e-01 5.75253606e-01 8.21025133e-01 6.14550114e-01
1.02087700e+00 -3.36145014e-01 -3.74158144e-01 -1.04214214e-01
1.93209350e-01 8.25445533e-01 5.88397309e-02 -1.17794538e+00
3.11795950e-01 4.67889994e-01 1.77415445e-01 -1.00674379e+00
-2.75493950e-01 -4.05655891e-01 -1.28271246e+00 2.07194775e-01
6.37193799e-01 2.71083694e-02 -7.69094229e-02 1.73041141e+00
2.73323148e-01 2.65986519e-03 1.28308773e-01 9.80932355e-01
4.63403463e-01 1.15530625e-01 -7.53188804e-02 -1.92338690e-01
1.26896560e+00 -2.61454493e-01 -7.02045143e-01 3.85082930e-01
1.00260186e+00 -4.00863379e-01 6.74476922e-01 3.54794562e-01
-8.77615869e-01 -1.73907131e-01 -9.65875506e-01 -1.14133902e-01
-1.97790384e-01 -4.89923239e-01 1.21291935e+00 1.52734792e+00
-6.68029487e-01 1.30501604e+00 -4.71114367e-01 1.79040685e-01
7.21717536e-01 6.14619851e-01 -5.21098554e-01 -8.23577344e-02
-1.72378802e+00 7.84991741e-01 5.15596010e-02 -2.16361452e-02
-5.98482847e-01 -3.35000753e-01 -5.06084323e-01 2.03630835e-01
7.51057416e-02 -8.86940360e-01 1.07805550e+00 -8.17906916e-01
-1.56904006e+00 1.20029032e+00 7.59617016e-02 -6.96706057e-01
5.84217072e-01 4.13188308e-01 -4.79452789e-01 -6.27033263e-02
8.54266062e-02 -4.18249965e-02 6.43173873e-01 -3.00644785e-01
-4.59733456e-01 -7.57140160e-01 3.27572167e-01 -2.59910136e-01
-2.21412152e-01 2.18953922e-01 5.44019341e-01 8.73915479e-02
-4.61554155e-02 -1.02254641e+00 -2.98207343e-01 -2.83146828e-01
-6.33666337e-01 -2.72525638e-01 -2.97167987e-01 -2.47260690e-01
1.41024625e+00 -1.92743528e+00 -5.91634177e-02 6.25411868e-01
4.25126344e-01 1.46374349e-02 2.14224651e-01 2.52780877e-02
-6.50261194e-02 4.20459300e-01 -2.27779090e-01 -2.90079594e-01
7.78616250e-01 5.23669273e-02 -1.40570402e-01 1.03888738e+00
-2.03747302e-02 8.44655156e-01 -8.19775939e-01 -4.70100522e-01
-1.17670968e-01 -2.15282217e-01 -7.72420049e-01 -8.19423527e-04
3.69668067e-01 2.70729274e-01 -3.96968395e-01 6.52715445e-01
1.02991498e+00 -2.46868163e-01 1.39998794e-01 -8.13450664e-03
4.54282090e-02 6.89267397e-01 -1.43216980e+00 1.32449639e+00
1.82652622e-01 1.36344805e-01 -2.96610504e-01 -1.07391310e+00
6.15605593e-01 1.41073734e-01 3.30398567e-02 -1.09755182e+00
5.32205164e-01 3.68126631e-01 5.62762082e-01 -3.82918715e-01
5.43045819e-01 -7.74605691e-01 -4.16434258e-01 9.15865183e-01
-3.16033959e-02 -1.09643981e-01 2.28930995e-01 1.65557489e-01
1.03990102e+00 -3.79482955e-01 2.21264750e-01 -3.92163992e-01
5.28259039e-01 -2.67140865e-01 7.47794390e-01 1.06685603e+00
-1.00261819e+00 4.77497280e-02 9.05189157e-01 -6.16040349e-01
-1.21988153e+00 -9.17710781e-01 -6.00142360e-01 8.58329058e-01
1.80065379e-01 -4.89926398e-01 -8.49119902e-01 -4.38428402e-01
3.20757419e-01 6.25983298e-01 -8.66618633e-01 -3.22948307e-01
3.33482809e-02 -1.12362993e+00 1.02072334e+00 -1.89086399e-03
5.58367550e-01 -4.49053556e-01 -3.87821048e-01 -1.25948206e-01
-5.64871021e-02 -8.38635266e-01 2.17697799e-01 2.04436436e-01
-7.52617359e-01 -1.11267471e+00 9.45018604e-02 1.98197171e-01
5.05955696e-01 -2.95425087e-01 1.14005518e+00 2.96337038e-01
-1.70928285e-01 -2.91545987e-01 -1.96061000e-01 -2.35721231e-01
-8.01111341e-01 -2.90698260e-01 6.32798314e-01 8.80182981e-02
8.50081742e-01 -3.12017441e-01 -5.41629314e-01 4.12372835e-02
-9.17302430e-01 -1.30419120e-01 3.24431360e-01 1.08596170e+00
2.21408412e-01 3.47156703e-01 8.18853155e-02 -1.00303757e+00
5.59311211e-01 -2.13320598e-01 -9.58678961e-01 2.02189803e-01
-8.78547966e-01 3.80037457e-01 6.89811409e-01 2.81447440e-01
-4.37517524e-01 -4.54935372e-01 1.32818863e-01 9.01759788e-02
1.65613159e-03 3.90662879e-01 1.33825913e-01 -4.32971776e-01
7.15539575e-01 -3.17517787e-01 5.88379577e-02 -1.41732236e-02
4.33606148e-01 8.48879516e-01 2.13372961e-01 -8.52151215e-01
8.54184091e-01 4.82716650e-01 8.65510464e-01 -2.53317475e-01
-7.39933014e-01 2.17985976e-02 -6.34792268e-01 1.19421154e-01
6.49596274e-01 -6.42617464e-01 -1.74686968e+00 4.93729115e-01
-8.76086831e-01 7.27884248e-02 4.57473332e-03 6.18076205e-01
-3.42419893e-01 5.37218869e-01 -8.88680518e-01 -1.33466089e+00
-3.41875583e-01 -1.15873349e+00 5.95747411e-01 2.75561392e-01
1.15059599e-01 -8.35067868e-01 5.82007952e-02 6.08204544e-01
2.96748668e-01 -8.08027536e-02 1.07704651e+00 -5.78008890e-01
-5.60727656e-01 -3.89464110e-01 -3.92616481e-01 4.61173743e-01
-3.92461479e-01 7.35334381e-02 -1.19862986e+00 -1.92347974e-01
-1.74430922e-01 -5.43323934e-01 6.55218780e-01 1.83472931e-02
1.01276565e+00 -2.29758546e-01 3.74356389e-01 4.27896380e-01
1.26328599e+00 -4.95618820e-01 7.55936623e-01 1.95945442e-01
2.90964574e-01 6.36465311e-01 4.52475637e-01 8.71303797e-01
5.06611466e-01 6.21353328e-01 6.66071057e-01 3.30712497e-01
7.36826181e-01 -1.67344399e-02 3.72780532e-01 6.63880885e-01
-4.87658024e-01 4.18600202e-01 -7.49857545e-01 8.51475447e-02
-1.68474948e+00 -1.37792194e+00 -4.59161520e-01 2.64321303e+00
9.50202465e-01 1.24705024e-01 1.14467435e-01 3.18018883e-01
6.39261782e-01 -3.40124726e-01 -2.53057808e-01 -9.15049791e-01
-2.91464865e-01 7.93398976e-01 8.85323763e-01 3.68756652e-01
-1.29011953e+00 7.54255056e-01 5.63899136e+00 7.61867344e-01
-8.95450234e-01 2.11875692e-01 8.46792459e-01 -4.92717586e-02
-3.51105601e-01 3.08067381e-01 -5.64901233e-01 4.98656303e-01
1.17523313e+00 -6.32882714e-01 8.22170794e-01 5.71534514e-01
1.79111154e-03 -1.34696245e-01 -1.43619633e+00 1.09462357e+00
-4.42119628e-01 -1.31274557e+00 -3.59346539e-01 2.80802131e-01
8.78529549e-01 -1.31321877e-01 3.78610730e-01 4.45168734e-01
4.69101101e-01 -1.32919228e+00 7.70624816e-01 4.90370393e-01
8.53664696e-01 -9.08193946e-01 1.23728716e+00 3.28925043e-01
-4.34005201e-01 -2.63199925e-01 -8.76088977e-01 -7.13762760e-01
-2.96609342e-01 8.91780734e-01 -2.17317209e-01 8.93643737e-01
7.42156982e-01 3.80369097e-01 -5.39891958e-01 7.15772927e-01
-2.96178013e-01 4.81083810e-01 -2.53937900e-01 -3.82137924e-01
1.99576199e-01 -5.75330377e-01 6.78861216e-02 6.87531710e-01
1.67616695e-01 2.86778599e-01 -4.66862768e-01 1.35168469e+00
-3.66408557e-01 1.24291010e-01 -5.26178479e-01 -3.21212620e-01
3.18325490e-01 1.04297471e+00 -2.21839070e-01 -3.25142056e-01
-1.28446355e-01 8.59025419e-01 1.82054952e-01 -2.04187736e-01
-8.98695409e-01 -3.95012915e-01 8.52614582e-01 -3.28585684e-01
-2.19412014e-01 1.91049576e-01 -6.67493343e-01 -1.52361083e+00
-8.20713565e-02 -8.27510834e-01 3.35983425e-01 -1.42314315e-01
-1.81903183e+00 1.32930830e-01 -5.71753204e-01 -1.05048323e+00
-7.37812221e-02 -8.54263723e-01 -4.07135248e-01 1.15166414e+00
-1.59772837e+00 -6.60544574e-01 1.72344372e-01 6.89884067e-01
-9.42232370e-01 -2.09820881e-01 1.37314212e+00 5.67955196e-01
-7.83493340e-01 9.71136630e-01 3.18810493e-01 3.72586787e-01
5.76271057e-01 -1.22677505e+00 1.96566269e-01 8.97922516e-01
2.15304382e-02 9.54865932e-01 7.01114893e-01 -2.86811769e-01
-1.51080847e+00 -6.53522253e-01 1.35248542e+00 -4.72561032e-01
9.74497676e-01 -3.58607441e-01 -4.38811541e-01 2.24737510e-01
-2.39403427e-01 2.57477582e-01 1.25987041e+00 3.99401665e-01
-6.58090174e-01 -3.63235921e-01 -1.44100475e+00 2.83354104e-01
9.07690465e-01 -1.13293159e+00 -3.40290546e-01 3.55811775e-01
1.96507603e-01 -1.81747660e-01 -9.24828947e-01 1.05858587e-01
7.97623336e-01 -1.61988759e+00 6.59365475e-01 -1.28124702e+00
6.96353018e-01 -2.09801331e-01 -5.17351568e-01 -9.11431849e-01
-4.19995397e-01 -8.40177417e-01 1.87629133e-01 8.14925492e-01
2.71100998e-01 -7.83202827e-01 8.22657347e-01 1.13977110e+00
5.27250588e-01 -2.24869326e-01 -1.59265399e+00 -6.52945757e-01
5.98853171e-01 -6.08223677e-01 9.88248587e-01 1.14114642e+00
4.24168944e-01 5.95786572e-02 -4.55082357e-01 2.09213614e-01
1.16629219e+00 3.07123095e-01 7.36474335e-01 -1.41095185e+00
-5.06442249e-01 -6.72329545e-01 -1.01292229e+00 -3.46994519e-01
3.67301285e-01 -1.17309654e+00 -5.56340635e-01 -5.46149075e-01
6.18525743e-01 -7.23733723e-01 -5.48891187e-01 2.17045724e-01
-2.23332778e-01 2.47862801e-01 2.08263323e-01 1.33108720e-01
-6.79328680e-01 5.80643594e-01 9.42852378e-01 -1.59583390e-01
4.93557423e-01 1.78442374e-01 -7.98486114e-01 3.17594588e-01
4.17908221e-01 -6.98754251e-01 2.81505555e-01 -5.55797257e-02
8.74009490e-01 6.84309006e-03 6.52649105e-01 -8.07276666e-01
1.80288881e-01 -2.32286066e-01 -1.57930881e-01 1.13671295e-01
-6.95031509e-02 -5.34667075e-01 -3.48320566e-02 8.48928273e-01
-4.25478876e-01 -2.00200260e-01 -4.49699134e-01 2.80072838e-01
1.84430555e-02 -4.01326060e-01 7.89090097e-01 7.89576173e-02
-4.28077370e-01 4.98690128e-01 1.67865053e-01 -3.81237753e-02
7.73955643e-01 4.75221574e-01 -8.74925017e-01 -1.87636361e-01
-4.44768548e-01 6.40081018e-02 5.67035973e-01 -2.32572511e-01
2.28033677e-01 -1.38847661e+00 -8.58540595e-01 2.40441695e-01
3.89786065e-01 -6.52980268e-01 4.11968589e-01 1.05847669e+00
-5.98629773e-01 5.77528179e-01 -3.73853862e-01 -3.10344726e-01
-8.75971556e-01 4.75829005e-01 6.08267725e-01 -3.02504659e-01
1.17864691e-01 9.42528903e-01 -2.23415732e-01 -7.08316982e-01
-9.30896252e-02 -1.48329705e-01 2.86863565e-01 -2.95820624e-01
5.53722262e-01 6.41234338e-01 1.89400658e-01 -7.33742654e-01
-4.31337029e-01 2.28424773e-01 4.18598145e-01 -5.08089215e-02
1.01545072e+00 -1.18621951e-02 -7.91504979e-01 3.61580074e-01
9.42941666e-01 -2.46134758e-01 -4.82546329e-01 -2.74227589e-01
-3.65867801e-02 -6.57084346e-01 9.19734389e-02 -5.97276211e-01
-6.19902194e-01 1.32359350e+00 7.63334930e-01 5.38580716e-01
6.50893629e-01 -3.93616676e-01 5.02187252e-01 6.98862970e-01
9.43503499e-01 -1.32731009e+00 -5.29860318e-01 4.48154181e-01
-1.40871316e-01 -1.48067546e+00 6.74489811e-02 -1.87927380e-01
-4.61552680e-01 1.19653630e+00 -2.82052290e-02 -4.19440866e-03
6.46308243e-01 -2.66447991e-01 -1.53217375e-01 -1.53495222e-01
-5.91666102e-01 -2.12037131e-01 5.81578054e-02 3.08610320e-01
3.66050601e-01 8.93080175e-01 -4.51598853e-01 1.07109320e+00
-4.47171628e-01 2.61095673e-01 8.10310006e-01 4.82093692e-01
-1.16641104e-01 -1.40993118e+00 -3.34606737e-01 4.52862442e-01
-7.59638667e-01 -3.02992880e-01 -8.30226466e-02 9.79340896e-02
3.47211242e-01 1.20215762e+00 -1.44065917e-01 -6.47282839e-01
5.32449409e-02 3.87972780e-02 6.49009883e-01 -1.98084950e-01
-6.80390418e-01 -6.40752733e-01 8.98157433e-02 -7.57833481e-01
-6.82107329e-01 -6.30758941e-01 -1.12495410e+00 -1.58372390e+00
-4.52201128e-01 3.66996616e-01 6.53188288e-01 1.12955606e+00
2.15424180e-01 3.32847498e-02 9.76885438e-01 -2.31735393e-01
-1.42587090e+00 -6.43172324e-01 -1.32545412e+00 7.46010840e-01
1.07939495e-02 -6.68696523e-01 -4.99361187e-01 -4.19208437e-01] | [5.593581199645996, 5.019985198974609] |
9286e713-4c6f-4fc6-9830-8b65df1a4b4d | a-joint-sentiment-target-stance-model-for | null | null | https://aclanthology.org/C16-1250 | https://aclanthology.org/C16-1250.pdf | A Joint Sentiment-Target-Stance Model for Stance Classification in Tweets | Classifying the stance expressed in online microblogging social media is an emerging problem in opinion mining. We propose a probabilistic approach to stance classification in tweets, which models stance, target of stance, and sentiment of tweet, jointly. Instead of simply conjoining the sentiment or target variables as extra variables to the feature space, we use a novel formulation to incorporate three-way interactions among sentiment-stance-input variables and three-way interactions among target-stance-input variables. The proposed specification intuitively aims to discriminate sentiment features from target features for stance classification. In addition, regularizing a single stance classifier, which handles all targets, acts as a soft weight-sharing among them. We demonstrate that discriminative training of this model achieves the state-of-the-art results in supervised stance classification, and its generative training obtains competitive results in the weakly supervised setting. | ['Javid Ebrahimi', 'Dejing Dou', 'Daniel Lowd'] | 2016-12-01 | a-joint-sentiment-target-stance-model-for-1 | https://aclanthology.org/C16-1250 | https://aclanthology.org/C16-1250.pdf | coling-2016-12 | ['subjectivity-analysis'] | ['natural-language-processing'] | [ 2.83059269e-01 2.31721938e-01 -5.96684039e-01 -8.94527853e-01
-1.00418532e+00 -5.19168258e-01 9.52985227e-01 2.66045809e-01
-3.68418664e-01 6.29458547e-01 4.14119840e-01 -1.61991924e-01
3.68348777e-01 -1.10566413e+00 -5.78336298e-01 -9.56646740e-01
1.70255482e-01 6.57682300e-01 7.44899884e-02 -5.29206038e-01
1.08828187e-01 -4.47069645e-01 -1.44921863e+00 5.45976937e-01
6.83213115e-01 1.27936208e+00 -4.28158045e-01 4.50596660e-01
-1.43801212e-01 9.83048797e-01 -6.37535632e-01 -6.50563121e-01
2.37635020e-02 -3.77556443e-01 -6.58976078e-01 3.23715985e-01
7.39456117e-02 3.17689389e-01 1.64898351e-01 8.66486788e-01
1.00633629e-01 -1.41215235e-01 9.90198791e-01 -1.19623160e+00
-5.51560640e-01 1.11630714e+00 -8.20714474e-01 9.85880047e-02
2.18640268e-01 -4.28710848e-01 1.78984809e+00 -8.45938742e-01
4.52823490e-01 1.15756416e+00 5.84144115e-01 -2.85413414e-02
-1.24324274e+00 -4.90782470e-01 5.94928861e-01 -2.21269920e-01
-5.63619912e-01 -1.65376127e-01 1.15726471e+00 -7.62736559e-01
3.54755938e-01 2.43622124e-01 7.93955147e-01 1.35313296e+00
3.90176475e-01 1.20757651e+00 1.20909715e+00 -2.74161935e-01
1.22913271e-01 5.77216268e-01 1.08869839e+00 3.37115467e-01
2.08560601e-01 -2.75827646e-01 -9.16112602e-01 -6.04857624e-01
-1.99994117e-01 -1.64259262e-02 6.82692602e-02 -9.36844125e-02
-8.69850636e-01 1.58338392e+00 2.11814180e-01 7.16684163e-02
-3.10993344e-01 -1.46000206e-01 5.19984543e-01 5.22503316e-01
1.44880557e+00 2.45659098e-01 -7.64559686e-01 5.48251905e-02
-8.28164399e-01 3.89881670e-01 1.00600529e+00 6.62810683e-01
7.98274279e-01 -2.52147883e-01 -1.85870424e-01 7.16896534e-01
3.77789468e-01 7.44272292e-01 5.13835728e-01 -1.88766539e-01
5.49712420e-01 8.68988931e-01 -5.13409637e-02 -9.51037049e-01
-5.40368199e-01 -7.71350265e-01 -4.60414439e-01 -2.54001945e-01
2.20878869e-01 -4.45011109e-01 -6.98040485e-01 1.68130982e+00
5.83059371e-01 -4.50253725e-01 6.22623004e-02 4.64410305e-01
6.21698678e-01 6.24086797e-01 -1.88064188e-01 -4.81948167e-01
1.52916586e+00 -8.64508569e-01 -4.84549075e-01 -5.47792673e-01
4.87985611e-01 -8.18886518e-01 8.97237003e-01 2.56216556e-01
-9.19433594e-01 -1.47562653e-01 -9.77857947e-01 1.47223175e-01
-5.52942932e-01 -1.90198217e-02 7.52561212e-01 5.52301109e-01
-4.58612144e-01 2.91703671e-01 -5.91411591e-01 1.71726853e-01
3.91064316e-01 3.10835123e-01 6.57835305e-02 3.13517898e-01
-1.41095865e+00 6.63950026e-01 7.12166503e-02 -1.48366913e-01
-1.89362615e-01 -7.14104533e-01 -9.88178968e-01 -3.10486287e-01
3.89061451e-01 -6.10394895e-01 1.31478381e+00 -9.80003595e-01
-1.43566453e+00 1.14683580e+00 -5.73257446e-01 -4.15146828e-01
4.97628003e-01 -1.98528096e-01 -3.61028492e-01 -3.24404091e-01
5.54456115e-01 -4.22119312e-02 1.10062504e+00 -1.08446181e+00
-8.02568853e-01 -3.42293650e-01 2.33865440e-01 1.25404909e-01
-5.15850067e-01 9.98628363e-02 -1.22083575e-01 -6.16956055e-01
3.74016106e-01 -1.04261649e+00 -1.47899732e-01 -5.94081938e-01
-6.04472399e-01 -5.19988418e-01 8.43811572e-01 -7.50821158e-02
1.07838047e+00 -1.75209129e+00 2.36360654e-01 5.43713987e-01
3.55154008e-01 -2.34864324e-01 2.63916999e-01 1.86156645e-01
2.21850108e-02 -6.46701735e-03 -5.47931567e-02 -7.04516053e-01
2.48409510e-01 1.36636466e-01 -5.61338842e-01 6.56179607e-01
1.90664306e-01 9.20309246e-01 -8.47207725e-01 -4.26481396e-01
-3.00619692e-01 3.30388606e-01 -7.46879458e-01 -2.31914464e-02
-3.43843281e-01 9.53094736e-02 -7.20180333e-01 5.29472411e-01
5.23156464e-01 -6.28754437e-01 4.42293435e-01 -1.58691049e-01
2.82281060e-02 8.72273445e-01 -7.53400683e-01 7.75211513e-01
-4.60035563e-01 3.59202415e-01 -6.32511377e-02 -1.14738095e+00
8.50730836e-01 2.63984650e-01 7.03620374e-01 -3.99755418e-01
6.27526164e-01 2.05594137e-01 -1.16496585e-01 -1.24023929e-01
4.22738910e-01 -3.94373417e-01 -8.61688733e-01 9.81178522e-01
-7.51899257e-02 1.01066209e-01 3.36396456e-01 1.35594204e-01
6.15373909e-01 -1.28614470e-01 2.27349609e-01 -3.81362915e-01
4.64256823e-01 -3.88354324e-02 5.56758523e-01 5.36148787e-01
2.86746114e-01 3.68260890e-01 9.95531440e-01 -2.95274019e-01
-5.28400898e-01 -6.91023171e-01 -2.67787158e-01 1.72424531e+00
-1.30934626e-01 -5.56343257e-01 -4.63241786e-01 -1.11690068e+00
1.31590575e-01 2.83589274e-01 -1.17415226e+00 5.57952896e-02
-6.03814363e-01 -1.40209544e+00 -5.95035555e-04 4.85699594e-01
9.99056101e-02 -8.51048946e-01 -3.17961484e-01 1.47835359e-01
-3.97554398e-01 -9.44978654e-01 -5.52128136e-01 8.98266435e-01
-3.87646347e-01 -1.05797386e+00 -3.03323895e-01 -6.89833522e-01
5.59823930e-01 6.09661005e-02 1.23216975e+00 -1.61333665e-01
6.17202222e-01 -2.84575194e-01 -5.64366221e-01 -6.81526899e-01
-2.10988775e-01 4.08490717e-01 2.70983782e-02 3.31968874e-01
4.03407514e-01 -5.42409599e-01 -2.59629965e-01 1.11184172e-01
-6.44347847e-01 -1.46883607e-01 1.86302483e-01 1.12982714e+00
3.95992815e-01 -1.82106361e-01 5.55375695e-01 -1.87779582e+00
5.67913055e-01 -8.18789482e-01 -3.11469436e-01 -1.62996799e-01
-6.59146190e-01 -6.61433069e-03 4.95542377e-01 -5.44401765e-01
-7.11723387e-01 -2.69467562e-01 -2.82419175e-01 2.41625339e-01
3.24461758e-01 1.08114183e+00 -7.73517117e-02 3.15087885e-01
5.49849212e-01 -1.90650462e-03 -1.07909590e-01 -2.91778266e-01
2.20803618e-01 7.77072072e-01 -3.74031626e-02 -4.96325344e-01
9.73057628e-01 6.20795488e-01 -4.05684978e-01 -7.10607469e-01
-2.01561666e+00 -7.38574743e-01 -5.17936468e-01 7.22046643e-02
6.94057703e-01 -1.01553655e+00 -4.58968729e-01 5.36726594e-01
-9.34164524e-01 -4.94806580e-02 -1.59605771e-01 1.72327876e-01
-4.61873949e-01 -4.28887159e-02 -6.99862838e-01 -7.76786149e-01
-4.68273044e-01 -1.01379085e+00 1.17352855e+00 -2.18475774e-01
-5.59644341e-01 -1.19468510e+00 1.91194385e-01 5.27488649e-01
2.24143133e-01 3.87049645e-01 7.78946340e-01 -9.84026074e-01
1.91603705e-01 -4.67346102e-01 8.69293138e-02 3.28564793e-01
2.40916640e-01 -1.12574264e-01 -1.04603982e+00 -1.30879804e-01
4.28276181e-01 -5.72266519e-01 1.22667253e+00 5.08494914e-01
6.02459788e-01 -5.43465078e-01 -2.97319859e-01 3.60351235e-01
8.63627911e-01 -2.73922592e-01 -5.46281375e-02 4.40784574e-01
5.20452619e-01 6.77789032e-01 8.80493879e-01 4.44243520e-01
7.43065536e-01 5.66984117e-01 2.14533493e-01 -2.72387862e-01
5.24988711e-01 -5.63734537e-03 6.41742647e-01 8.07011724e-01
2.24496394e-01 -2.87329704e-01 -5.61661959e-01 3.47733498e-01
-2.04351354e+00 -9.32624936e-01 -3.70083183e-01 1.72058511e+00
1.17520440e+00 7.88639963e-01 5.42431414e-01 4.73349512e-01
4.23698515e-01 6.24394178e-01 -4.00303602e-01 -1.63770735e-01
-4.48773980e-01 2.61192799e-01 3.27686429e-01 8.09873044e-01
-1.61624551e+00 8.37671757e-01 6.52941132e+00 4.45288032e-01
-1.38513625e+00 3.23308229e-01 6.42018199e-01 -1.80301920e-01
-5.85988522e-01 -1.22388732e-02 -1.16695487e+00 5.23947835e-01
7.25878179e-01 -2.92738825e-01 -2.17699721e-01 9.88556087e-01
2.17578650e-01 4.75201979e-02 -1.10780966e+00 3.56740892e-01
2.85321712e-01 -1.44491136e+00 -3.36900093e-02 2.45249718e-01
9.18328702e-01 1.68483883e-01 3.00468266e-01 4.00894821e-01
4.79382187e-01 -5.64949453e-01 1.18578827e+00 4.83127795e-02
3.23058546e-01 -7.62928307e-01 9.42707896e-01 3.86541605e-01
-9.74469543e-01 -1.21525653e-01 8.72238539e-03 -3.04163337e-01
2.71159321e-01 1.16140473e+00 -5.14652073e-01 3.33758324e-01
4.62189555e-01 1.13391328e+00 -2.61334717e-01 1.75750434e-01
-3.46601218e-01 9.16675866e-01 -2.84795702e-01 -2.46034980e-01
3.57891589e-01 -7.65024945e-02 3.83925050e-01 1.17251766e+00
-3.32007200e-01 -1.60178900e-01 6.13429725e-01 4.85096186e-01
-1.41276240e-01 1.98244810e-01 -3.91416073e-01 -6.53626695e-02
3.86180729e-02 1.31182694e+00 -6.30297720e-01 -6.09564364e-01
-4.58239317e-01 4.64774430e-01 3.72309715e-01 1.97008535e-01
-8.64419639e-01 1.47574067e-01 8.91206741e-01 2.44187534e-01
4.86905962e-01 8.68943185e-02 -5.42118371e-01 -1.43469238e+00
2.48012915e-01 -1.00714898e+00 4.86109972e-01 -1.36347309e-01
-1.59382653e+00 5.68937898e-01 -1.19389080e-01 -1.18105996e+00
-4.38503504e-01 -6.88791275e-01 -8.45470846e-01 8.11354816e-01
-1.83041537e+00 -1.54707479e+00 2.43583947e-01 5.03945291e-01
5.02482772e-01 7.56072924e-02 7.53098309e-01 4.70883287e-02
-4.78063345e-01 6.16934836e-01 -6.01214692e-02 2.28043884e-01
6.85394466e-01 -1.35323858e+00 2.86135048e-01 4.12874281e-01
4.74533485e-03 7.30585039e-01 9.92001712e-01 -5.13941407e-01
-1.01267231e+00 -1.17916012e+00 1.43424189e+00 -6.82921946e-01
1.22475183e+00 -7.16260970e-01 -4.21150774e-01 9.18982983e-01
6.88153058e-02 -3.59663576e-01 1.27871501e+00 8.73654783e-01
-6.85826004e-01 1.93795428e-01 -7.22737193e-01 3.29989105e-01
6.39852166e-01 -6.12386286e-01 -7.04025805e-01 6.62057042e-01
6.21576190e-01 -3.24520797e-01 -6.37168288e-01 5.04725203e-02
5.86017430e-01 -7.14678943e-01 7.58431911e-01 -7.61934400e-01
8.17925334e-01 -3.66446190e-02 -3.84590626e-01 -1.37714517e+00
-1.83297083e-01 -3.63250196e-01 -3.17611217e-01 1.13082290e+00
1.06440938e+00 -8.38185966e-01 1.06771469e+00 2.07296684e-01
-1.08795680e-01 -1.16568923e+00 -5.31327963e-01 -1.88677847e-01
4.08789009e-01 -4.15938318e-01 3.12617928e-01 9.34891164e-01
5.60713649e-01 1.02864063e+00 -5.55706143e-01 -2.71475799e-02
5.87260544e-01 8.46430480e-01 6.66984618e-01 -1.66858089e+00
-6.47306800e-01 -4.73417878e-01 -2.03572344e-02 -1.19264531e+00
5.65764070e-01 -1.02794898e+00 4.61062454e-02 -1.31009734e+00
2.92468458e-01 -6.28980577e-01 -1.75005108e-01 4.58743811e-01
-4.05793250e-01 5.81473410e-01 -1.25094622e-01 1.68204546e-01
-6.36792362e-01 5.48445642e-01 9.04448509e-01 -4.47481066e-01
-1.39646992e-01 5.99123776e-01 -1.26032698e+00 1.09877002e+00
6.24700248e-01 -7.99443781e-01 -3.38059485e-01 -5.04861884e-02
6.68965518e-01 -2.31004730e-01 -6.74688742e-02 -2.51605451e-01
1.38055250e-01 -2.02262372e-01 2.25276008e-01 -1.03533924e+00
7.38552749e-01 -3.94463032e-01 -5.23150921e-01 1.14231415e-01
-4.76822436e-01 -1.85430512e-01 -5.46480954e-01 5.48373342e-01
-4.40687835e-01 6.87832981e-02 5.99130988e-01 -6.86675608e-02
1.41628802e-01 2.48374149e-01 -4.95720476e-01 3.64571929e-01
6.73885942e-01 -1.68935880e-02 -4.28340286e-01 -3.98601115e-01
-7.98282623e-01 4.00126874e-01 2.33843312e-01 4.24096018e-01
1.21289000e-01 -1.12950194e+00 -8.55732620e-01 2.37101331e-01
2.33627900e-01 -1.27981648e-01 -2.57561117e-01 1.20395792e+00
4.45763499e-01 2.33319134e-01 4.59261984e-01 -6.27928913e-01
-1.36399412e+00 1.29445165e-01 2.95552146e-02 -7.62708485e-01
-2.76455849e-01 1.09881735e+00 2.24706709e-01 -5.80200791e-01
-3.27412337e-02 -9.82622430e-02 -4.71207350e-01 9.47019756e-01
3.48594904e-01 -1.88277483e-01 3.33819628e-01 -8.70935857e-01
-2.98223227e-01 4.74645466e-01 -3.76036018e-01 -1.67660028e-01
1.52693641e+00 -6.89406395e-02 -2.16006070e-01 1.03325737e+00
1.40236962e+00 3.92827243e-01 -7.09327638e-01 -5.34908891e-01
1.02088146e-01 -8.03771988e-02 -7.10808486e-02 -3.76717478e-01
-8.95255804e-01 4.83833760e-01 -3.48765075e-01 6.79989338e-01
5.79837561e-01 3.99184585e-01 7.09636390e-01 2.98407376e-01
2.49528393e-01 -1.06833935e+00 2.80644387e-01 1.11026192e+00
5.14400482e-01 -1.50336456e+00 1.25111742e-02 -6.40720069e-01
-7.67891169e-01 7.44958818e-01 1.48782879e-01 -3.85518551e-01
1.24684227e+00 4.86054868e-01 4.03861880e-01 -4.52806085e-01
-9.95589077e-01 -1.29381567e-01 4.12417442e-01 1.58663958e-01
9.65455472e-01 1.97727248e-01 -6.87135279e-01 9.82660890e-01
-8.29122245e-01 -5.65352261e-01 3.52576345e-01 1.07449830e+00
-5.30595005e-01 -1.33934999e+00 -1.44291505e-01 8.82720590e-01
-7.86783636e-01 -2.87361443e-01 -5.21671832e-01 5.27309716e-01
7.22263977e-02 1.20385432e+00 4.72319126e-02 -5.95635176e-01
2.63214171e-01 1.91627219e-01 -9.86648127e-02 -9.84871387e-01
-9.66206074e-01 3.30417186e-01 5.23180366e-01 -2.44404227e-01
-6.56069577e-01 -9.86992478e-01 -9.42181110e-01 -4.14348930e-01
-6.15410566e-01 7.22013712e-01 4.94983256e-01 1.39971066e+00
-1.16432540e-01 5.43098390e-01 1.06450844e+00 -9.00335193e-01
-7.93752968e-01 -1.07488799e+00 -6.37805700e-01 4.43459302e-01
6.32608294e-01 -7.68137574e-01 -6.46496534e-01 9.52781439e-02] | [8.816258430480957, 9.641390800476074] |
a72a2696-a7d8-4d0e-b6b7-a767d91e0266 | otseq2set-an-optimal-transport-enhanced | 2210.14523 | null | https://arxiv.org/abs/2210.14523v2 | https://arxiv.org/pdf/2210.14523v2.pdf | OTSeq2Set: An Optimal Transport Enhanced Sequence-to-Set Model for Extreme Multi-label Text Classification | Extreme multi-label text classification (XMTC) is the task of finding the most relevant subset labels from an extremely large-scale label collection. Recently, some deep learning models have achieved state-of-the-art results in XMTC tasks. These models commonly predict scores for all labels by a fully connected layer as the last layer of the model. However, such models can't predict a relatively complete and variable-length label subset for each document, because they select positive labels relevant to the document by a fixed threshold or take top k labels in descending order of scores. A less popular type of deep learning models called sequence-to-sequence (Seq2Seq) focus on predicting variable-length positive labels in sequence style. However, the labels in XMTC tasks are essentially an unordered set rather than an ordered sequence, the default order of labels restrains Seq2Seq models in training. To address this limitation in Seq2Seq, we propose an autoregressive sequence-to-set model for XMTC tasks named OTSeq2Set. Our model generates predictions in student-forcing scheme and is trained by a loss function based on bipartite matching which enables permutation-invariance. Meanwhile, we use the optimal transport distance as a measurement to force the model to focus on the closest labels in semantic label space. Experiments show that OTSeq2Set outperforms other competitive baselines on 4 benchmark datasets. Especially, on the Wikipedia dataset with 31k labels, it outperforms the state-of-the-art Seq2Seq method by 16.34% in micro-F1 score. The code is available at https://github.com/caojie54/OTSeq2Set. | ['Yin Zhang', 'Jie Cao'] | 2022-10-26 | null | null | null | null | ['multi-label-text-classification', 'multi-label-text-classification'] | ['methodology', 'natural-language-processing'] | [ 5.65023661e-01 -2.84958720e-01 -4.16480660e-01 -7.07279146e-01
-1.15854371e+00 -7.10289359e-01 2.49932304e-01 -2.31478095e-01
-5.48397720e-01 8.35963249e-01 8.93623531e-02 -2.50779450e-01
-1.08565643e-01 -6.54581368e-01 -6.46018982e-01 -8.67452383e-01
4.41888779e-01 8.81026685e-01 1.75727338e-01 -1.41918287e-01
3.52917254e-01 -1.71319738e-01 -1.52013063e+00 8.68694901e-01
8.48766267e-01 9.36748266e-01 2.78030068e-01 5.36004424e-01
-4.73627567e-01 7.78114676e-01 -4.46734309e-01 -3.87154073e-01
2.60390311e-01 -6.40606701e-01 -1.10937238e+00 -4.93532091e-01
6.54322088e-01 3.27753741e-03 -2.08248328e-02 1.18293953e+00
7.88739204e-01 3.40373032e-02 9.03498173e-01 -1.39462483e+00
-6.94187403e-01 8.79192114e-01 -7.73075998e-01 -2.63726205e-01
2.09573552e-01 -4.78188656e-02 1.55538440e+00 -9.40060616e-01
6.29441977e-01 1.29392886e+00 6.98839962e-01 7.86491632e-01
-1.11622369e+00 -1.07409811e+00 7.81245977e-02 2.62588978e-01
-1.33325887e+00 1.14744697e-02 4.20819491e-01 -3.78756136e-01
6.82355464e-01 5.27926266e-01 -5.93307391e-02 1.13847780e+00
8.47421810e-02 1.02424300e+00 1.28253567e+00 -3.81351650e-01
-4.99942638e-02 -1.32805079e-01 4.57618624e-01 6.70077264e-01
-3.60173523e-01 -1.63328439e-01 -5.03813744e-01 -2.87240058e-01
8.66273418e-02 7.43582696e-02 -3.50223146e-02 -6.37677461e-02
-1.43935442e+00 8.82444620e-01 3.16713542e-01 1.52765781e-01
1.02886446e-01 3.12476695e-01 8.64508986e-01 4.68673527e-01
7.13120699e-01 2.53378361e-01 -1.06756818e+00 1.15265856e-02
-7.03807771e-01 1.39895782e-01 5.35618246e-01 1.08484685e+00
8.01959097e-01 -5.32055974e-01 -7.00458884e-01 1.29775250e+00
2.81502813e-01 5.09814143e-01 7.07139552e-01 -7.90926933e-01
7.15802789e-01 6.31791055e-01 -3.53026718e-01 -3.90017807e-01
-5.70257246e-01 -4.44085985e-01 -9.04008567e-01 -2.10646689e-01
4.13497120e-01 -2.46858642e-01 -9.01971877e-01 1.84632266e+00
2.99135000e-01 2.32233912e-01 -8.33734795e-02 8.54477823e-01
1.02895832e+00 8.36300850e-01 2.76610434e-01 -1.75942302e-01
1.38944161e+00 -1.32012713e+00 -4.98807907e-01 -2.89735328e-02
1.31129229e+00 -5.55316985e-01 1.32070196e+00 3.70695084e-01
-5.69156229e-01 -5.79993069e-01 -5.32068193e-01 -2.61528641e-01
-4.38824415e-01 2.54903525e-01 3.26072544e-01 4.25518930e-01
-1.07916367e+00 5.94010413e-01 -1.54064387e-01 -1.73471600e-01
3.04767579e-01 4.45937306e-01 -2.34566003e-01 -2.33149067e-01
-1.53615248e+00 4.91527438e-01 6.08011007e-01 -1.24192074e-01
-6.51388764e-01 -7.96889305e-01 -5.86662948e-01 1.39450446e-01
3.93356353e-01 -4.53534186e-01 1.34300947e+00 -1.02009082e+00
-1.42611897e+00 1.09901965e+00 -1.95506692e-01 1.32942433e-02
5.03092706e-01 1.39001803e-02 -3.07779014e-01 -1.80731103e-01
4.08557981e-01 1.15610099e+00 4.44314480e-01 -1.02889788e+00
-9.55247045e-01 -2.06800416e-01 -3.39726120e-01 2.05630928e-01
-4.60630715e-01 3.36485118e-01 -2.68292457e-01 -5.42808115e-01
-3.21488529e-02 -1.28975105e+00 -1.26941040e-01 -3.63449186e-01
-5.74753106e-01 -9.38590825e-01 5.54608762e-01 -4.07546997e-01
1.28979421e+00 -2.00554204e+00 7.13699758e-02 -8.96462053e-02
-8.45067799e-02 2.12895975e-01 -4.94991213e-01 5.44527411e-01
-2.28856593e-01 2.11905807e-01 -2.25295961e-01 -3.86549264e-01
2.86694318e-01 -8.89035687e-02 -2.97747016e-01 3.38035047e-01
-5.83008826e-02 9.55654860e-01 -1.18796074e+00 -6.68869436e-01
-2.34897926e-01 8.76391381e-02 -3.76141161e-01 -1.29457591e-02
-5.63251853e-01 4.36926723e-01 -4.73032355e-01 4.81317431e-01
7.24828422e-01 -6.04538023e-01 2.93159157e-01 1.25507578e-01
8.75208974e-02 3.77150923e-01 -8.52918923e-01 1.85324132e+00
-4.57541764e-01 3.06259423e-01 -6.22219026e-01 -8.54472756e-01
1.03220284e+00 4.02287871e-01 5.38409114e-01 -7.15378046e-01
1.23106569e-01 4.82399166e-01 -1.50094122e-01 -4.99355972e-01
4.32181656e-01 -2.22421780e-01 -4.02159393e-01 6.68410540e-01
-8.01291689e-02 3.28527510e-01 3.27994704e-01 4.42617051e-02
1.07142854e+00 3.36318821e-01 5.50511666e-02 -1.79875955e-01
6.44883811e-01 -8.42942223e-02 8.72710824e-01 7.25925982e-01
-1.18036300e-01 7.79208064e-01 5.89620888e-01 -3.71456265e-01
-9.55058277e-01 -5.71765184e-01 -2.08666041e-01 1.84329772e+00
1.21011290e-05 -2.94291794e-01 -4.91095066e-01 -1.46213973e+00
7.10419863e-02 8.31533015e-01 -5.41502237e-01 -9.65465531e-02
-6.03322864e-01 -8.72853577e-01 7.42000699e-01 3.85139912e-01
4.21190381e-01 -1.34784472e+00 2.10578114e-01 1.12649128e-01
-5.54341495e-01 -7.63464272e-01 -9.81201172e-01 4.10823017e-01
-5.59216797e-01 -9.15573120e-01 -7.79401660e-01 -1.14634383e+00
6.43139482e-01 3.98725756e-02 1.05996335e+00 -2.11864989e-03
1.74401090e-01 -3.07268530e-01 -6.38367593e-01 -2.53192514e-01
-4.63221610e-01 5.25351167e-01 -2.08882570e-01 1.29353628e-01
7.39261687e-01 -5.65091334e-02 -4.51739252e-01 6.01586998e-01
-7.35751212e-01 2.20644906e-01 4.40433949e-01 1.09605849e+00
6.70897663e-01 -7.82715827e-02 1.09708703e+00 -1.28695130e+00
3.91771108e-01 -7.33703673e-01 -3.48336935e-01 4.74367112e-01
-6.21501505e-01 8.84125754e-02 1.15703940e+00 -4.58347946e-01
-8.82400155e-01 3.05232227e-01 -4.13434893e-01 -1.50822520e-01
-1.90604448e-01 3.59563828e-01 -9.85092372e-02 2.34925419e-01
4.68686610e-01 3.52550775e-01 -2.53510982e-01 -5.08501291e-01
2.81857014e-01 1.09764004e+00 3.02334726e-02 -5.41108131e-01
2.91864574e-01 1.04598798e-01 6.40807450e-02 -8.20795074e-03
-1.38937187e+00 -7.69843698e-01 -7.43538797e-01 -1.44084081e-01
8.05939972e-01 -7.98216820e-01 -9.46625829e-01 5.94598472e-01
-1.01930809e+00 -6.19733572e-01 1.22577846e-01 2.85146922e-01
-5.09244978e-01 2.79835343e-01 -8.89535964e-01 -4.52155888e-01
-6.02147281e-01 -1.17443156e+00 1.35189235e+00 -4.93219793e-02
-8.07586238e-02 -8.76423120e-01 1.89933062e-01 5.89016914e-01
3.70838828e-02 -1.71542630e-01 1.03576338e+00 -1.08225513e+00
-2.86783576e-01 6.50619343e-02 -3.33981335e-01 3.55140895e-01
-1.14907376e-01 -2.68042713e-01 -1.02029121e+00 -4.59415793e-01
-4.21930015e-01 -8.24645460e-01 1.16518867e+00 2.81861126e-01
1.56883609e+00 -2.50682253e-02 -5.66210926e-01 5.53540468e-01
1.42900610e+00 1.14534646e-01 3.99224341e-01 3.97095442e-01
9.82094705e-01 6.45039260e-01 9.80749428e-01 1.56572059e-01
5.40173531e-01 7.40731716e-01 3.82132769e-01 1.45659000e-01
-1.14025183e-01 -3.60673666e-01 3.73657256e-01 1.01018608e+00
6.49750888e-01 -6.46521747e-01 -1.00209963e+00 3.87036562e-01
-1.96045208e+00 -7.98388004e-01 -5.55640817e-01 2.06354785e+00
1.17756629e+00 -4.43993285e-02 -4.66232747e-02 -2.61316374e-02
1.13326383e+00 2.58241445e-02 -7.35205233e-01 -3.51595134e-01
-5.12283333e-02 -1.23075776e-01 6.30818248e-01 2.20020577e-01
-1.28076470e+00 1.05407214e+00 5.28011036e+00 1.51780593e+00
-9.17597115e-01 3.09752405e-01 8.64250243e-01 -3.53167653e-02
-4.06188518e-01 -7.21449628e-02 -1.32749069e+00 9.34885144e-01
1.09701693e+00 3.63768011e-01 1.08522557e-01 5.36140919e-01
-2.62178425e-02 3.21406871e-01 -1.06845498e+00 7.61574328e-01
9.58060473e-02 -9.09241617e-01 -7.39409635e-03 8.25887695e-02
1.04250479e+00 2.15285316e-01 2.38130480e-01 7.30953157e-01
6.56051219e-01 -1.02689540e+00 6.11950040e-01 3.21978003e-01
1.23460162e+00 -8.08771908e-01 8.44994307e-01 6.56755745e-01
-1.03339195e+00 -3.71692806e-01 -6.35042727e-01 1.78197250e-01
1.50074407e-01 6.70485437e-01 -9.17814374e-01 5.53703785e-01
4.63278919e-01 8.31926048e-01 -4.74637955e-01 8.98312747e-01
-3.38858485e-01 8.87519062e-01 -4.06367928e-02 -3.88708740e-01
4.63710606e-01 -1.35603726e-01 2.73354854e-02 1.39595699e+00
4.07323718e-01 -5.48957996e-02 4.07209575e-01 4.02278423e-01
-4.31160510e-01 4.20994073e-01 -2.96501428e-01 3.60196650e-01
5.83608508e-01 1.39183795e+00 -6.99900270e-01 -3.16595405e-01
-5.36090255e-01 9.50307429e-01 6.70055330e-01 1.37355119e-01
-9.41473067e-01 -4.88079906e-01 3.50623965e-01 -3.85514736e-01
2.71028399e-01 3.25261414e-01 -3.29103291e-01 -8.88497651e-01
-2.03487724e-01 -9.84079599e-01 8.06876361e-01 -6.93598032e-01
-1.56798077e+00 4.23421890e-01 -4.09699202e-01 -1.46073008e+00
5.48718730e-03 -5.72632253e-01 -1.80889562e-01 8.85411978e-01
-1.53521359e+00 -1.23643100e+00 -2.55976785e-02 1.80579588e-01
7.14789748e-01 -1.54542655e-01 7.95882285e-01 4.59795088e-01
-6.01806700e-01 9.63193119e-01 8.11962485e-01 2.02164963e-01
1.32708907e+00 -1.45325959e+00 5.49612522e-01 2.94865578e-01
2.49898992e-02 1.99769527e-01 2.13903382e-01 -6.64368033e-01
-8.58844995e-01 -1.52564597e+00 1.33458710e+00 -6.50780141e-01
4.22529042e-01 -4.75300133e-01 -8.63493443e-01 7.83831060e-01
4.92817797e-02 8.57337341e-02 9.12207782e-01 7.77644292e-02
-5.03793657e-01 -7.83363804e-02 -8.64270747e-01 4.45717692e-01
1.33790958e+00 -4.58546579e-01 -2.09638581e-01 7.66168833e-01
1.03579235e+00 -2.79387861e-01 -8.65666747e-01 4.34209466e-01
4.96348888e-01 -6.37284338e-01 6.93950593e-01 -6.04872584e-01
6.75486088e-01 -3.08649510e-01 2.00543925e-02 -1.55200601e+00
-6.87605917e-01 -1.69413015e-01 4.15162951e-01 1.37277591e+00
7.47036874e-01 -6.81889832e-01 8.88728321e-01 1.30262761e-03
-3.74140173e-01 -1.03315294e+00 -7.26157427e-01 -7.59573817e-01
3.89057606e-01 -1.78958908e-01 7.75991380e-01 1.26536238e+00
-5.05656488e-02 5.63860595e-01 -4.88599777e-01 -1.57913014e-01
4.79450852e-01 3.26095223e-01 4.84076381e-01 -1.35877109e+00
-1.19817324e-01 -4.04462010e-01 7.68171921e-02 -1.30799985e+00
8.50896597e-01 -1.69092703e+00 2.89813966e-01 -1.48672605e+00
6.99955523e-01 -8.19856167e-01 -7.12321162e-01 8.00663948e-01
-5.31742871e-01 3.32735270e-01 -2.18399707e-02 2.49472111e-01
-1.03145719e+00 5.85121632e-01 1.40868771e+00 -2.19712898e-01
1.09752588e-01 -1.08104706e-01 -6.27280116e-01 2.53369987e-01
9.66357946e-01 -9.56879735e-01 -4.04445291e-01 -4.75151330e-01
5.46119094e-01 -8.70962162e-03 -1.14633217e-01 -5.69507360e-01
3.62464011e-01 -3.28248411e-01 1.78527057e-01 -7.68330336e-01
-2.12803651e-02 -6.14388764e-01 1.32505134e-01 4.30997014e-01
-1.07616973e+00 -1.21727258e-01 -2.41508797e-01 4.18150872e-01
-1.87608786e-02 -6.69850409e-01 6.13488734e-01 1.57037862e-02
-8.12070787e-01 5.27524531e-01 -8.76770094e-02 3.46117496e-01
9.24641728e-01 9.71160904e-02 -6.89885974e-01 3.78357694e-02
-3.36903989e-01 6.83965623e-01 2.35175684e-01 5.03365278e-01
3.07645768e-01 -1.58538425e+00 -1.07768595e+00 7.94962514e-03
4.33240622e-01 5.55365458e-02 3.70473057e-01 6.92969978e-01
-2.72227764e-01 6.44201696e-01 5.68840541e-02 -6.16755843e-01
-1.40062916e+00 5.78435659e-01 1.82113707e-01 -5.11467516e-01
-3.39548320e-01 1.15875435e+00 4.51385289e-01 -1.39007151e+00
1.13569535e-01 6.88252300e-02 -4.33569700e-01 1.66376084e-01
4.09257054e-01 3.82091075e-01 -9.29229613e-03 -6.06661081e-01
-3.20731759e-01 6.58264756e-01 -3.71769786e-01 1.94177032e-01
9.57288146e-01 -7.37347901e-02 -3.23849797e-01 5.27734101e-01
1.70965183e+00 -3.59627604e-01 -1.14585412e+00 -4.61548418e-01
2.04429671e-01 -2.21030965e-01 -2.26835355e-01 -9.86125171e-01
-9.27943587e-01 7.65749454e-01 3.79897177e-01 2.34413035e-02
9.07974958e-01 -8.43430758e-02 1.09542096e+00 3.27403188e-01
3.10478240e-01 -1.21125948e+00 1.73806608e-01 9.31980550e-01
5.36880493e-01 -1.37395632e+00 -4.96020526e-01 -4.09877867e-01
-7.76850581e-01 9.15074170e-01 8.37662280e-01 3.20279956e-01
4.86867994e-01 -2.54139323e-02 2.76682109e-01 -2.49741357e-02
-1.05255818e+00 -1.19280286e-01 9.63974074e-02 2.19466314e-01
8.05089176e-01 1.85659468e-01 -5.70596457e-01 3.20323467e-01
-1.14735857e-01 -1.23146579e-01 2.51604348e-01 4.86094981e-01
-5.32428145e-01 -1.31072223e+00 -1.89513668e-01 7.63471127e-01
-5.90076745e-01 -3.12727153e-01 -2.11368740e-01 2.50978559e-01
1.88586593e-01 9.71268237e-01 -1.33950993e-01 -4.70665962e-01
1.70679107e-01 4.68007505e-01 4.39872965e-03 -8.32360744e-01
-7.80939877e-01 -2.20716581e-01 1.07085250e-01 -1.88908815e-01
-5.39646149e-02 -7.45790064e-01 -1.50496018e+00 -1.70928895e-01
-5.54560781e-01 1.85073733e-01 4.99355525e-01 8.50094795e-01
3.50136518e-01 6.43954813e-01 8.91893148e-01 -3.06223094e-01
-9.09781694e-01 -1.18465579e+00 -6.68859065e-01 5.46993494e-01
5.14346547e-02 -4.12784100e-01 -2.37725660e-01 -2.48188883e-01] | [9.603764533996582, 4.413486957550049] |
e2d1a57c-f7a1-4049-b0b3-a9a8767a1bcd | nilc-at-sr20-exploring-pre-trained-models-in | null | null | https://aclanthology.org/2020.msr-1.6 | https://aclanthology.org/2020.msr-1.6.pdf | NILC at SR’20: Exploring Pre-Trained Models in Surface Realisation | This paper describes the submission by the NILC Computational Linguistics research group of the University of S ̃ao Paulo/Brazil to the English Track 2 (closed sub-track) at the Surface Realisation Shared Task 2020. The success of the current pre-trained models like BERT or GPT-2 in several tasks is well-known, however, this is not the case for data-to-text generation tasks and just recently some initiatives focused on it. This way, we explore how a pre-trained model (GPT-2) performs on the UD-to-text generation task. In general, the achieved results were poor, but there are some interesting ideas to explore. Among the learned lessons we may note that it is necessary to study strategies to represent UD inputs and to introduce structural knowledge into these pre-trained models. | ['Thiago Pardo', 'Marco Antonio Sobrevilla Cabezudo'] | null | null | null | null | msr-coling-2020-12 | ['data-to-text-generation'] | ['natural-language-processing'] | [ 1.82083011e-01 1.10501528e+00 -1.19966224e-01 -2.30167419e-01
-7.46741951e-01 -3.65518689e-01 1.25443590e+00 1.38359785e-01
-3.42506319e-01 1.06472659e+00 9.19020057e-01 -6.09636128e-01
1.26624927e-01 -7.85722792e-01 -4.46962208e-01 -3.88097942e-01
1.81685418e-01 1.09017670e+00 5.34200259e-02 -7.51758754e-01
3.80852163e-01 2.43053094e-01 -1.48106480e+00 5.74348867e-01
1.05980849e+00 1.82557106e-01 4.18066353e-01 6.63573265e-01
-3.59866500e-01 7.78429031e-01 -7.90903866e-01 -7.42272913e-01
-9.74806473e-02 -7.77942896e-01 -1.36740744e+00 -3.84885877e-01
1.48476623e-02 -3.23635489e-02 3.75227898e-01 6.22651339e-01
6.97787702e-01 1.29024968e-01 8.08098614e-01 -1.00818932e+00
-8.53563726e-01 1.48504794e+00 -4.40553669e-03 2.29748771e-01
5.67604184e-01 9.95140430e-03 1.03514659e+00 -8.72222364e-01
1.18009329e+00 1.36464047e+00 4.83263284e-01 9.48938847e-01
-1.09631908e+00 -3.50087553e-01 -1.66772991e-01 2.44137436e-01
-1.03891742e+00 -6.52984083e-01 4.80014950e-01 -4.60583657e-01
1.44821775e+00 1.11530438e-01 4.17568266e-01 1.34001315e+00
-3.38231921e-02 9.37765062e-01 1.42732370e+00 -8.87566924e-01
-3.42811607e-02 4.67413723e-01 -1.87387928e-01 2.30068564e-01
2.46046349e-01 6.99836835e-02 -5.79719365e-01 8.67936537e-02
4.04268086e-01 -1.10683990e+00 -1.62574932e-01 2.26316944e-01
-1.27514291e+00 1.07366586e+00 7.00886995e-02 1.00543034e+00
-3.12002152e-01 -6.12554252e-02 5.89053452e-01 3.21439445e-01
7.31762946e-01 6.96117938e-01 -4.64110255e-01 -7.32007861e-01
-1.07325304e+00 6.37815952e-01 9.98414874e-01 9.96875644e-01
5.01502573e-01 1.32428482e-01 -4.32419449e-01 8.43470991e-01
4.17872876e-01 1.12038404e-01 7.05944061e-01 -6.14320219e-01
7.13444829e-01 4.28155065e-01 -2.66320677e-03 -7.53602386e-01
-3.37784857e-01 -6.28272742e-02 -6.03698075e-01 -3.05017471e-01
4.98461217e-01 -5.18630624e-01 -8.06411088e-01 1.41337252e+00
-4.46864739e-02 -3.48151982e-01 3.96328300e-01 5.37204444e-01
9.57968056e-01 8.14716756e-01 1.41417265e-01 -3.40206116e-01
1.12408352e+00 -7.76213169e-01 -1.06401002e+00 -4.50052470e-01
1.12359262e+00 -1.14191735e+00 9.01366353e-01 1.26596034e-01
-1.45567536e+00 -7.19889283e-01 -8.11179996e-01 -1.28145203e-01
-7.43422210e-01 2.91099876e-01 4.74284053e-01 9.03659642e-01
-1.23640919e+00 5.60938358e-01 -7.25843966e-01 -8.09583306e-01
-1.12363398e-01 -1.00311100e-01 -6.82399347e-02 1.82133459e-03
-1.52664459e+00 1.49911988e+00 8.26167524e-01 1.00576885e-01
-6.26106560e-01 -3.28388929e-01 -8.32498968e-01 -4.01918858e-01
1.01995818e-01 -5.07875323e-01 1.41399181e+00 -6.52871668e-01
-1.57620513e+00 1.19975746e+00 -3.16191822e-01 -6.53589547e-01
6.74073458e-01 1.28260292e-02 -3.32496464e-01 -4.14010823e-01
2.71023482e-01 8.72650027e-01 2.92398125e-01 -1.33068633e+00
-6.83061779e-01 -1.74970925e-01 -3.95642640e-03 2.17559204e-01
1.07220160e-02 4.59304452e-01 2.59911001e-01 -7.59099364e-01
-2.28210226e-01 -9.24292564e-01 -1.52647033e-01 -9.11317289e-01
-2.90203333e-01 -7.80751765e-01 4.49343354e-01 -9.67585742e-01
1.34373307e+00 -1.57388580e+00 3.38741988e-01 -2.18656912e-01
-4.62974548e-01 4.98032600e-01 -8.96386132e-02 1.10656166e+00
-2.10575350e-02 5.51401913e-01 -1.54580832e-01 -5.64261079e-01
2.92251647e-01 5.14737725e-01 -4.75633889e-01 -2.52100587e-01
4.54701155e-01 1.07282090e+00 -1.07583225e+00 -3.67563516e-01
2.25627497e-02 3.50245267e-01 -3.24922293e-01 1.84315801e-01
-4.42099929e-01 3.98528725e-01 -2.54012913e-01 -2.15917500e-03
5.43676674e-01 3.27668875e-01 2.79609442e-01 1.96771026e-01
-5.45151412e-01 9.81832445e-01 -1.09702659e+00 1.60703313e+00
-4.06134784e-01 6.82355106e-01 -4.16910797e-01 -1.01864159e+00
9.82888401e-01 8.62716973e-01 2.95398422e-02 -6.54052854e-01
3.75889428e-02 7.86488533e-01 3.55870426e-01 -5.04164994e-01
1.12586045e+00 -4.53505576e-01 -1.00207455e-01 7.48879790e-01
3.50005835e-01 -5.89807212e-01 6.01102114e-01 1.60524771e-01
5.07167280e-01 5.59860528e-01 4.85897005e-01 -4.00351852e-01
4.81732339e-01 3.56825978e-01 2.05125347e-01 5.30808330e-01
6.79205135e-02 8.04154098e-01 3.98319900e-01 -1.19390793e-01
-1.00227857e+00 -4.50524330e-01 -1.47892699e-01 9.73808348e-01
-5.47968507e-01 -6.14352882e-01 -1.04967344e+00 -6.03076220e-01
-6.38965607e-01 1.46734047e+00 -6.80869937e-01 1.53896704e-01
-7.73875237e-01 -7.44043946e-01 9.45913374e-01 2.97236711e-01
2.01343015e-01 -1.75449491e+00 -4.54951733e-01 5.89194298e-01
-4.54381824e-01 -1.01888287e+00 -4.07601967e-02 1.52550653e-01
-7.93427169e-01 -4.68013078e-01 -7.93563008e-01 -9.04721022e-01
2.50966191e-01 -2.03926265e-01 1.20578635e+00 -3.22748311e-02
2.15241343e-01 1.59348231e-02 -8.15770090e-01 -9.87535119e-01
-1.25497687e+00 6.35838151e-01 -4.19687301e-01 -3.60212058e-01
3.39482337e-01 -2.24396661e-01 5.19128032e-02 -4.74795066e-02
-8.00470233e-01 4.54065889e-01 3.41098309e-01 7.28828490e-01
1.38526753e-01 -3.29803050e-01 9.03132319e-01 -1.11657846e+00
1.09283280e+00 -3.70986044e-01 4.58135158e-02 1.24487817e-01
-5.24345756e-01 4.06130403e-02 3.66956532e-01 4.91567887e-02
-1.41941226e+00 -3.52827728e-01 -5.66189587e-01 4.13198441e-01
-3.05865198e-01 8.61826956e-01 -1.29900090e-02 5.52179813e-01
7.85498261e-01 1.97536260e-01 -3.49440649e-02 -3.98260772e-01
6.99999630e-01 9.30144191e-01 1.23268768e-01 -7.81280339e-01
6.23982906e-01 1.04600564e-02 -4.94273990e-01 -8.40941727e-01
-7.44020820e-01 -8.75512287e-02 -7.27330387e-01 -1.59116492e-01
9.44108129e-01 -8.59204710e-01 2.89001614e-01 2.88830906e-01
-1.69754004e+00 -6.58748388e-01 -6.29615366e-01 1.27614588e-01
-6.68448925e-01 -5.20069301e-02 -4.92923439e-01 -8.84766340e-01
-3.84790540e-01 -1.16580510e+00 1.08269882e+00 1.96205094e-01
-7.19851732e-01 -1.58965445e+00 3.40181619e-01 6.59169674e-01
6.85582519e-01 1.03441924e-01 1.13003445e+00 -7.96405971e-01
-1.17185660e-01 1.12875424e-01 -4.10304479e-02 3.58955234e-01
5.85315749e-02 8.27743486e-02 -9.73723650e-01 -7.80046210e-02
-5.14590144e-02 -4.31470782e-01 3.81417811e-01 1.24240525e-01
3.87251824e-01 -3.18414003e-01 -2.43874118e-01 -1.68868184e-01
1.09714365e+00 3.49401161e-02 9.41921711e-01 3.89766544e-01
4.05267209e-01 9.46004748e-01 7.55254269e-01 9.92443413e-02
9.88492250e-01 8.78855705e-01 -5.49911754e-04 1.05858840e-01
-6.06350482e-01 -4.41739678e-01 6.34320438e-01 1.27882063e+00
-4.98459816e-01 -3.31308216e-01 -1.06501222e+00 9.87119079e-01
-1.89609861e+00 -1.11083865e+00 -7.94427991e-01 1.87493539e+00
1.13822663e+00 1.41301349e-01 1.62129775e-01 1.75255444e-02
5.24543703e-01 1.82045057e-01 4.06682611e-01 -1.11251760e+00
-5.07958531e-01 5.13946533e-01 -2.03074753e-01 6.65054798e-01
-5.23731768e-01 1.30051184e+00 6.61179543e+00 8.81873846e-01
-1.04213381e+00 4.48477209e-01 4.46613103e-01 2.67534167e-01
-6.87514961e-01 3.24197292e-01 -9.88351285e-01 2.24946812e-01
1.36250782e+00 -5.72699189e-01 3.07875305e-01 5.10730982e-01
4.89204258e-01 -5.04067726e-02 -9.03533936e-01 2.22853079e-01
4.01736110e-01 -1.27822971e+00 2.64617234e-01 2.22108155e-01
1.05902565e+00 2.44719639e-01 -3.60570014e-01 7.08431602e-01
5.29704928e-01 -1.09262300e+00 8.62689495e-01 2.97970563e-01
4.50587243e-01 -6.12724960e-01 9.45976198e-01 5.47629893e-01
-7.70759344e-01 3.71253490e-01 -6.42310798e-01 -3.04314286e-01
5.09349823e-01 4.97902960e-01 -1.31885934e+00 9.53585386e-01
1.66621208e-01 5.54680109e-01 -7.22099245e-01 8.08025718e-01
-6.03763402e-01 9.06737983e-01 1.16693303e-01 -2.72125721e-01
4.31137770e-01 -1.14840738e-01 6.48722589e-01 1.46248317e+00
4.90379184e-01 -2.33733743e-01 -2.12081909e-01 7.99341857e-01
1.29746929e-01 5.34376085e-01 -8.22138131e-01 -1.26285806e-01
1.46288365e-01 1.21658897e+00 -5.71325541e-01 -3.25717688e-01
-3.47961009e-01 8.89778972e-01 4.22741413e-01 8.11449960e-02
-3.75382900e-01 -1.22861624e-01 2.08844796e-01 2.26944640e-01
1.36099279e-01 -1.59246206e-01 -3.83057535e-01 -8.39052200e-01
-2.05622062e-01 -1.03810692e+00 2.40366131e-01 -9.30205464e-01
-1.29169464e+00 8.97163093e-01 2.62810171e-01 -6.53133392e-01
-1.01264441e+00 -4.56890404e-01 -6.52608097e-01 1.14548564e+00
-1.46984267e+00 -1.43294203e+00 2.26133049e-01 2.40953162e-01
5.78275442e-01 -2.11840928e-01 1.19667804e+00 -8.77801701e-02
-2.73021758e-01 4.11846191e-01 -6.81792423e-02 -7.67505378e-04
6.87715590e-01 -1.42241180e+00 8.67849648e-01 8.49207401e-01
3.50395650e-01 5.77631474e-01 1.00769544e+00 -6.59037232e-01
-8.00229073e-01 -8.82555544e-01 2.03896713e+00 -6.23708427e-01
7.94395506e-01 -2.35791296e-01 -7.35556424e-01 8.44337881e-01
1.09737599e+00 -6.84374034e-01 5.18583119e-01 9.44484025e-02
1.96308881e-01 4.65478718e-01 -8.22487652e-01 7.43646383e-01
9.27140832e-01 -3.71102661e-01 -1.14061725e+00 5.07046282e-01
7.42470741e-01 -5.87952793e-01 -9.06588674e-01 -4.79513500e-03
1.52354062e-01 -8.46477330e-01 3.93393934e-01 -5.47771811e-01
6.42273724e-01 -2.15657763e-02 1.77559480e-02 -1.92132425e+00
1.21851429e-01 -9.23012912e-01 2.49397904e-01 1.71362805e+00
9.24634576e-01 -7.64669657e-01 5.09618342e-01 3.42595875e-01
-4.53232616e-01 -4.15928543e-01 -1.05706203e+00 -5.83265126e-01
7.60848105e-01 -7.05269217e-01 5.43013334e-01 1.06021404e+00
3.00864935e-01 6.57108665e-01 -1.30885080e-01 -4.97077614e-01
6.37693256e-02 -2.42045224e-01 8.62353325e-01 -1.12219894e+00
-1.31699041e-01 -5.42888105e-01 -5.16401455e-02 -5.62537074e-01
2.14580238e-01 -1.24732840e+00 8.25105608e-02 -2.16210246e+00
-2.68241376e-01 -5.74209392e-01 4.23603088e-01 4.40041214e-01
-3.26053441e-01 6.34051636e-02 4.89610881e-01 1.01680033e-01
7.13789985e-02 3.87721211e-01 1.46559834e+00 2.09694818e-01
-3.48340809e-01 2.64536440e-02 -9.05636549e-01 2.05416635e-01
9.05201554e-01 -5.19228578e-01 -3.88658047e-01 -6.80654943e-01
3.09756577e-01 -7.96969905e-02 2.68380303e-04 -6.77888870e-01
-2.68874262e-02 -1.27167553e-01 -3.11704367e-01 -4.80796218e-01
3.86678763e-02 -1.60636693e-01 2.27422372e-01 3.22330177e-01
-3.15853685e-01 1.97264954e-01 4.16381061e-01 -3.06623191e-01
-3.20455909e-01 -8.17866385e-01 4.29790258e-01 -4.04860735e-01
-3.27177674e-01 -2.59067148e-01 -8.11145604e-01 3.50290030e-01
8.42104375e-01 -2.35685945e-01 -3.35085601e-01 -5.39694190e-01
-4.99446630e-01 4.67647202e-02 1.88035667e-01 6.07361555e-01
1.76689312e-01 -1.17025995e+00 -1.28652787e+00 -6.60530999e-02
2.63465911e-01 -2.77728271e-02 1.63516914e-03 7.11500823e-01
-4.92827207e-01 9.32916105e-01 -1.63869068e-01 -2.58572727e-01
-8.34061146e-01 -1.00338263e-02 2.78337508e-01 -7.21687973e-01
-2.24273771e-01 6.42097175e-01 -5.20071805e-01 -8.33519220e-01
-2.78398335e-01 -9.20573696e-02 -5.92184186e-01 4.62220728e-01
4.47708309e-01 3.14178675e-01 4.02696371e-01 -9.11147237e-01
-1.25208152e-02 5.65398559e-02 -7.98707902e-02 -6.76880360e-01
1.38765132e+00 -6.76454380e-02 -2.72269487e-01 6.06985569e-01
7.66959012e-01 -1.97380707e-02 -4.97309029e-01 -6.54519796e-02
3.64835531e-01 1.59272239e-01 -4.36982512e-02 -9.61629272e-01
-5.78871012e-01 1.24130905e+00 7.34036863e-02 5.29071450e-01
6.49496436e-01 -5.36149479e-02 4.73636925e-01 1.01557583e-01
3.86537403e-01 -1.34156775e+00 -2.43133992e-01 9.89362895e-01
1.31954122e+00 -9.42019880e-01 -2.66530216e-01 -2.61227280e-01
-1.02933943e+00 1.09178734e+00 5.43245196e-01 4.23062630e-02
3.70962858e-01 2.63511717e-01 1.75741866e-01 -7.16931149e-02
-1.04460967e+00 -4.70468789e-01 1.48442313e-01 1.05651748e+00
1.16623700e+00 1.88118845e-01 -1.00270033e+00 3.31241757e-01
-8.09838295e-01 5.82639575e-02 1.00366902e+00 6.67085886e-01
-1.64921746e-01 -1.76011050e+00 -1.81291386e-01 2.24395677e-01
-3.81113887e-01 -4.81665701e-01 -7.21444190e-01 1.02534425e+00
1.81487828e-01 1.04642069e+00 1.84644870e-02 -1.30027607e-01
4.29515868e-01 4.68259007e-01 4.67025518e-01 -1.23756635e+00
-1.23930788e+00 -6.82874247e-02 6.64522767e-01 -1.70440711e-02
-6.41420603e-01 -1.00716114e+00 -1.06462264e+00 -2.87598610e-01
-2.36372769e-01 3.78444165e-01 8.74392271e-01 1.07176602e+00
1.81744620e-01 4.02204603e-01 1.87667057e-01 -9.10800219e-01
-2.82426387e-01 -1.53578126e+00 -1.95272371e-01 1.98229566e-01
-2.77974069e-01 -1.97363630e-01 -2.15383947e-01 1.61141440e-01] | [11.369269371032715, 9.177227973937988] |
dd019e5b-8e3e-4282-8a8e-c6177e8abdc3 | ghq-grouped-hybrid-q-learning-for | 2303.01070 | null | https://arxiv.org/abs/2303.01070v1 | https://arxiv.org/pdf/2303.01070v1.pdf | GHQ: Grouped Hybrid Q Learning for Heterogeneous Cooperative Multi-agent Reinforcement Learning | Previous deep multi-agent reinforcement learning (MARL) algorithms have achieved impressive results, typically in homogeneous scenarios. However, heterogeneous scenarios are also very common and usually harder to solve. In this paper, we mainly discuss cooperative heterogeneous MARL problems in Starcraft Multi-Agent Challenges (SMAC) environment. We firstly define and describe the heterogeneous problems in SMAC. In order to comprehensively reveal and study the problem, we make new maps added to the original SMAC maps. We find that baseline algorithms fail to perform well in those heterogeneous maps. To address this issue, we propose the Grouped Individual-Global-Max Consistency (GIGM) and a novel MARL algorithm, Grouped Hybrid Q Learning (GHQ). GHQ separates agents into several groups and keeps individual parameters for each group, along with a novel hybrid structure for factorization. To enhance coordination between groups, we maximize the Inter-group Mutual Information (IGMI) between groups' trajectories. Experiments on original and new heterogeneous maps show the fabulous performance of GHQ compared to other state-of-the-art algorithms. | ['Kai Lv', 'Sheng Han', 'Xiangsen Wang', 'Youfang Lin', 'Xiaoyang Yu'] | 2023-03-02 | null | null | null | null | ['smac-1', 'starcraft', 'smac'] | ['playing-games', 'playing-games', 'playing-games'] | [-6.56167746e-01 -1.94647044e-01 -1.99373409e-01 2.61648446e-02
-8.38785410e-01 -6.37820542e-01 5.76359272e-01 2.82527566e-01
-4.29530293e-01 1.11235058e+00 1.82975963e-01 1.98401749e-01
-6.82420135e-01 -6.84606373e-01 -6.54991806e-01 -1.01051664e+00
-6.90878868e-01 8.46049666e-01 3.36767286e-01 -8.19249332e-01
1.34948641e-01 -6.41603246e-02 -1.09498692e+00 3.02314192e-01
1.09585106e+00 5.66633821e-01 5.43818295e-01 6.83113635e-01
2.00353235e-01 1.39024031e+00 -6.64418280e-01 -3.35187674e-01
5.43505847e-01 -5.11116743e-01 -8.32829952e-01 -3.96158919e-02
-7.71851316e-02 -2.74872094e-01 -9.43130925e-02 9.41676855e-01
6.82915986e-01 5.78730643e-01 4.50848311e-01 -1.91130614e+00
-4.26307112e-01 8.68521214e-01 -7.11762011e-01 1.03540771e-01
1.74812689e-01 1.86427668e-01 1.17098069e+00 -3.37951750e-01
7.67390668e-01 1.39526427e+00 5.68886459e-01 4.25354362e-01
-9.04495180e-01 -4.85293329e-01 8.06298256e-01 3.63277555e-01
-1.01434243e+00 1.05164111e-01 6.50693297e-01 -3.01530808e-01
8.76446187e-01 -1.70708578e-02 7.21282542e-01 1.02758276e+00
4.70601946e-01 1.17547023e+00 1.07765687e+00 4.43386585e-02
6.88564897e-01 -2.99040079e-01 -2.15333745e-01 8.90397489e-01
9.43859443e-02 1.70103475e-01 -7.48360157e-01 -2.51348406e-01
6.11957788e-01 -1.03291526e-01 9.90031660e-02 -7.84735382e-01
-1.55749571e+00 1.17222750e+00 4.63386893e-01 3.53607163e-02
-7.43899822e-01 3.84787709e-01 4.37154770e-01 5.64101458e-01
3.66277397e-01 6.91416383e-01 -3.82508427e-01 -2.67714173e-01
-4.95354563e-01 9.82657313e-01 6.46351814e-01 8.34877610e-01
7.46257544e-01 3.14091481e-02 -3.68976705e-02 7.26290762e-01
2.16911107e-01 3.58579338e-01 4.06956643e-01 -1.38229609e+00
7.16346025e-01 3.72641057e-01 3.04534525e-01 -1.11060023e+00
-8.31500530e-01 -6.50231123e-01 -8.69825780e-01 2.83816904e-01
3.10468107e-01 -6.49095774e-01 -2.20175475e-01 1.78475285e+00
6.22636974e-01 2.12093964e-01 4.14416522e-01 1.30698621e+00
5.17113447e-01 7.64567196e-01 1.30117740e-02 -1.22604489e-01
7.87121236e-01 -1.62324572e+00 -7.43741691e-01 -1.12423822e-01
7.10396111e-01 -3.97974044e-01 6.06354237e-01 3.50894690e-01
-1.07220733e+00 -4.55892980e-01 -1.03756166e+00 5.01262724e-01
-2.73728669e-01 -3.70714009e-01 9.45541978e-01 2.26866260e-01
-1.12058306e+00 5.97921789e-01 -9.01194870e-01 -2.23606348e-01
1.36274010e-01 4.03992385e-01 -2.00307399e-01 9.13262293e-02
-1.37948847e+00 7.76261806e-01 4.07150507e-01 -1.34836495e-01
-1.42050874e+00 -6.36465967e-01 -7.32935548e-01 -3.47437799e-01
9.34086740e-01 -8.65940273e-01 1.36044133e+00 -1.02270186e+00
-1.74585116e+00 1.94508042e-02 3.29260916e-01 -6.03642106e-01
5.30916572e-01 -2.06002414e-01 -1.03948444e-01 2.35320866e-01
3.01951408e-01 8.27338696e-01 6.60466135e-01 -1.53217077e+00
-1.12673151e+00 -1.30950466e-01 4.66651320e-01 8.05686593e-01
7.64820427e-02 -1.46260604e-01 -2.12986216e-01 -6.06886089e-01
-3.49403560e-01 -1.11521971e+00 -6.39266014e-01 -5.65444767e-01
-6.97979657e-03 -4.37552541e-01 4.87743258e-01 -4.26562488e-01
9.90769625e-01 -1.74671650e+00 8.58771443e-01 3.78922522e-02
4.87755328e-01 -1.02525763e-03 -6.54711306e-01 8.20445478e-01
5.55133104e-01 -2.15862080e-01 6.91030174e-02 -5.33009410e-01
2.88018346e-01 5.03697753e-01 1.56853765e-01 6.07544243e-01
5.19980937e-02 9.71580267e-01 -1.40163898e+00 -4.61966962e-01
3.14357951e-02 -9.22147557e-02 -6.71917558e-01 2.04529226e-01
-4.68590021e-01 5.57294428e-01 -6.40464664e-01 6.99756563e-01
5.74890077e-01 -1.96737230e-01 3.50300372e-01 1.68129250e-01
-1.50645599e-01 -3.06242794e-01 -1.40828145e+00 1.93500519e+00
-1.95861280e-01 1.54113904e-01 4.14478272e-01 -1.12982535e+00
7.16821373e-01 1.56333461e-01 1.01083112e+00 -8.41028273e-01
-1.13178395e-01 4.11205478e-02 1.46906957e-01 -3.23491722e-01
7.99640238e-01 5.43988459e-02 -2.71813273e-01 5.70516765e-01
1.59459323e-01 2.09668535e-03 5.41775942e-01 4.97545779e-01
1.11299300e+00 1.99046999e-01 1.35024235e-01 -4.35467333e-01
2.54984558e-01 1.26381993e-01 8.78969491e-01 1.04235375e+00
-6.90617859e-01 3.51237357e-01 6.40052080e-01 -6.62427247e-01
-8.00204992e-01 -1.01821601e+00 8.18915069e-01 1.32827318e+00
7.92848468e-01 -6.31688714e-01 -6.29954338e-01 -9.21207905e-01
2.01021209e-01 1.94501191e-01 -7.45994329e-01 7.81271234e-02
-6.15439713e-01 -1.16751516e+00 1.38258889e-01 3.30023646e-01
6.25527084e-01 -1.07211757e+00 -6.48561239e-01 4.62294281e-01
-2.52354085e-01 -9.09887075e-01 -3.95874977e-01 3.69662531e-02
-4.04312760e-01 -1.05742705e+00 -7.19839215e-01 -7.06287861e-01
1.69839323e-01 4.50866789e-01 1.22778153e+00 -9.78544354e-02
1.84592783e-01 4.12193120e-01 -1.01924038e+00 -2.37309963e-01
-1.03698552e-01 5.18680274e-01 2.46500462e-01 -3.75713669e-02
-1.91331282e-01 -2.68984109e-01 -5.25587678e-01 4.62549537e-01
-7.80975044e-01 1.54213995e-01 6.16830707e-01 9.48720694e-01
2.56655484e-01 1.52631611e-01 1.00553870e+00 -5.68054914e-01
8.30044389e-01 -7.33333647e-01 -4.69352484e-01 2.61838228e-01
-4.23743606e-01 2.50589922e-02 7.77688682e-01 -4.99551088e-01
-8.13443005e-01 -1.95994452e-01 1.96036324e-01 -2.26467311e-01
2.41760254e-01 7.22766280e-01 1.22227080e-01 -5.19815758e-02
3.68275434e-01 -8.91400650e-02 2.34213397e-01 -6.28207400e-02
3.32355887e-01 1.98140964e-01 1.07832566e-01 -8.46182644e-01
6.51166797e-01 3.24420303e-01 -1.32048264e-01 -4.72826958e-01
-6.77246511e-01 -3.95836741e-01 -3.69898260e-01 -5.65778375e-01
7.52784550e-01 -1.10466635e+00 -1.14107990e+00 6.00004017e-01
-8.77773881e-01 -9.46569383e-01 -8.76987651e-02 7.35308290e-01
-8.96326840e-01 7.46345073e-02 -6.91609383e-01 -8.02462041e-01
-8.01463500e-02 -1.31872940e+00 8.88949037e-01 3.39654237e-01
2.77292341e-01 -1.13378072e+00 6.60521984e-01 4.10632312e-01
5.12327015e-01 4.00731802e-01 5.33104002e-01 -4.98138845e-01
-5.85708916e-01 4.70007032e-01 1.30236536e-01 -1.34667531e-01
-1.16012804e-01 -3.28916371e-01 -2.83184528e-01 -9.09974515e-01
-5.41641712e-01 -5.93991339e-01 7.60619700e-01 4.12791938e-01
5.30538380e-01 -1.94598541e-01 -1.30739510e-01 3.98869365e-01
1.38486969e+00 4.75197077e-01 2.14400351e-01 9.35599566e-01
4.91386056e-01 5.28753817e-01 1.00886834e+00 7.67260373e-01
9.30803716e-01 7.16187537e-01 8.76860857e-01 8.99309665e-03
3.01332504e-01 -6.49205148e-02 8.15896988e-01 1.02176225e+00
-2.61311501e-01 -4.89399940e-01 -6.83988273e-01 5.17330348e-01
-2.69284081e+00 -9.54221845e-01 1.65455237e-01 1.63240099e+00
4.73522007e-01 -9.45785195e-02 5.99732399e-01 -4.52270925e-01
6.20024800e-01 4.34294790e-01 -5.33190668e-01 -1.00166395e-01
-4.98706132e-01 -2.55052745e-01 5.16813934e-01 6.54071152e-01
-1.28242373e+00 1.05181289e+00 5.69524288e+00 8.32307994e-01
-6.03551328e-01 3.14683795e-01 2.91525662e-01 -3.44864279e-01
-1.30137518e-01 -6.21302575e-02 -5.99028409e-01 4.31644291e-01
6.03929877e-01 -7.90086463e-02 7.84421325e-01 6.31358981e-01
1.62205279e-01 -1.46990612e-01 -7.79146850e-01 9.62497234e-01
-7.12100193e-02 -1.31533754e+00 -1.34391367e-01 -1.34448260e-02
1.24421501e+00 2.83310115e-01 6.62360294e-03 8.24903548e-01
1.03888595e+00 -6.80192232e-01 8.55576038e-01 3.81436110e-01
-1.88497990e-01 -1.14844859e+00 1.00294781e+00 4.59650367e-01
-1.26875162e+00 -4.50921267e-01 -5.70149481e-01 -2.13181317e-01
2.06580609e-01 6.52140081e-02 -3.70056123e-01 1.19078577e+00
7.60032475e-01 9.54222143e-01 -5.16638219e-01 9.64357615e-01
1.02890551e-01 8.01323429e-02 4.81437513e-04 -4.08802778e-01
9.26345885e-01 -5.77069402e-01 6.24086261e-01 6.93457425e-01
6.50671348e-02 -2.35693697e-02 8.88811946e-01 3.67393047e-01
1.69409111e-01 1.31448761e-01 -3.09080958e-01 -4.56460461e-04
3.69664222e-01 1.38963521e+00 -5.51556230e-01 -3.09557259e-01
-4.53757584e-01 9.26210403e-01 7.30929434e-01 3.24595898e-01
-9.13408041e-01 -7.66086802e-02 8.20429325e-01 -4.33111221e-01
1.48301139e-01 -5.56987047e-01 2.62711287e-01 -1.26741672e+00
-2.06639439e-01 -1.20379627e+00 4.84353036e-01 -3.25635701e-01
-1.49562585e+00 5.02490044e-01 -4.81040329e-02 -1.31391954e+00
-4.88810688e-01 -2.62282908e-01 -4.10935700e-01 2.66398013e-01
-1.61348808e+00 -1.31887496e+00 -1.29503965e-01 7.62202203e-01
7.44739711e-01 -7.08116770e-01 5.72757602e-01 3.20847452e-01
-7.22839892e-01 3.04732114e-01 5.57543039e-01 -6.51634783e-02
8.47182035e-01 -1.48944998e+00 2.38427967e-01 5.37097275e-01
-1.64963052e-01 2.14519978e-01 6.37926877e-01 -7.86020875e-01
-1.72123325e+00 -1.18193781e+00 1.09798618e-01 -1.27403304e-01
9.29568768e-01 -3.28712761e-01 -4.37830925e-01 5.12248158e-01
5.98453462e-01 -2.38106981e-01 4.69699711e-01 2.68898219e-01
1.70860328e-02 -1.77708238e-01 -8.00845027e-01 6.18855178e-01
8.08057368e-01 -8.41897167e-03 -2.87031740e-01 4.96376961e-01
8.38898420e-01 -2.99311578e-01 -9.18616414e-01 2.58699507e-01
4.59317029e-01 -9.77456093e-01 8.39151382e-01 -8.00811887e-01
2.83306062e-01 -5.10024309e-01 -2.95998067e-01 -2.05471921e+00
-6.10530674e-01 -8.46670568e-01 -1.22690275e-01 8.49452853e-01
3.42044055e-01 -5.05846143e-01 8.86765659e-01 2.85235350e-03
-3.84490132e-01 -7.43684113e-01 -7.19951093e-01 -9.52515543e-01
3.37724239e-01 4.35277186e-02 6.30177379e-01 1.08704114e+00
1.31113753e-01 2.59505540e-01 -8.10576499e-01 3.06806803e-01
8.08071613e-01 2.96212047e-01 9.11924601e-01 -9.79466617e-01
-6.46015167e-01 -2.82655597e-01 -7.01290295e-02 -6.38976753e-01
2.16250911e-01 -7.37423718e-01 9.30301379e-03 -1.62113357e+00
3.00681651e-01 -6.54320598e-01 -2.38158494e-01 2.73399115e-01
-1.70839846e-01 1.27509544e-02 8.04851532e-01 2.79867858e-01
-1.60316217e+00 9.02226508e-01 1.47327733e+00 -2.70862341e-01
-4.50331181e-01 -1.35385603e-01 -4.43268210e-01 3.16965640e-01
1.14435577e+00 -1.95653677e-01 -4.92155254e-01 -6.16247535e-01
4.79949325e-01 2.69444406e-01 1.35659873e-01 -9.88395572e-01
4.76222813e-01 -6.20162904e-01 -2.18584146e-02 -6.02541625e-01
3.06813180e-01 -4.66339111e-01 1.29458487e-01 6.43322170e-01
-4.10649776e-01 5.20308316e-01 -1.89407900e-01 5.71173489e-01
-3.96773130e-01 -6.46543056e-02 7.55972266e-01 -3.77482146e-01
-9.04360831e-01 3.32990676e-01 -6.00131214e-01 1.49275556e-01
1.42554843e+00 5.18752098e-01 -6.50159538e-01 -6.35809720e-01
-5.96786857e-01 1.07731283e+00 3.40543658e-01 5.09065568e-01
5.37584007e-01 -1.45897055e+00 -9.23678219e-01 -2.12433949e-01
-5.15890867e-02 -7.20320782e-03 6.70132160e-01 1.02836633e+00
-5.97207367e-01 1.62530512e-01 -6.07248962e-01 -3.22431207e-01
-8.92932057e-01 7.46297717e-01 4.56970692e-01 -6.86019123e-01
-4.78798002e-01 5.68320692e-01 9.61559862e-02 -7.01332152e-01
1.42672762e-01 -7.36826882e-02 -2.23507702e-01 2.68116355e-01
4.80257154e-01 5.80942333e-01 -1.67281643e-01 -4.71016079e-01
-1.61196887e-01 3.97124767e-01 -1.98005334e-01 -1.96783900e-01
1.45832598e+00 -3.04811090e-01 2.62235105e-02 3.08892787e-01
8.46878231e-01 -3.55312258e-01 -1.53885269e+00 -1.36307463e-01
-1.53482288e-01 -1.53528124e-01 -1.57074764e-01 -7.35437036e-01
-1.30178154e+00 4.34176683e-01 4.11920816e-01 4.14248258e-01
7.52189517e-01 -6.23823218e-02 7.30321109e-01 3.91049266e-01
7.21648335e-01 -1.52811122e+00 3.83340657e-01 7.77329683e-01
7.99508810e-01 -1.21889448e+00 -3.47431526e-02 7.15660080e-02
-1.29251945e+00 8.28921795e-01 9.22384143e-01 -3.70605618e-01
3.64715576e-01 1.78884372e-01 1.10337645e-01 -2.54465401e-01
-1.24813735e+00 -4.75073665e-01 -2.43907690e-01 8.07496250e-01
-1.43812045e-01 2.06399053e-01 -2.71845490e-01 5.05664051e-01
-1.24649614e-01 -3.87426615e-01 5.85115850e-01 1.12171626e+00
-4.50388432e-01 -1.15396786e+00 -3.42751712e-01 2.42062584e-01
-1.85198128e-01 4.43799764e-01 -2.25981668e-01 1.00931573e+00
1.09513372e-01 1.14677536e+00 -9.35601944e-04 -5.59080899e-01
1.50979027e-01 -5.73895395e-01 3.57861698e-01 -2.35166010e-02
-9.37077641e-01 1.64951578e-01 -5.75711280e-02 -7.58635879e-01
-7.87247062e-01 -7.39282131e-01 -1.18335462e+00 -4.66602743e-01
9.32583213e-02 5.47929764e-01 5.23958683e-01 8.49536717e-01
2.93194532e-01 6.86557710e-01 7.20993817e-01 -9.88583386e-01
-5.96300662e-01 -8.23873401e-01 -8.15602303e-01 3.66994441e-01
4.56653863e-01 -1.07597578e+00 -1.85452372e-01 -3.96249354e-01] | [3.723856210708618, 1.9335036277770996] |
f90818fd-6612-4ffc-b53e-d0bdcb42e501 | semantic-clustering-of-pivot-paraphrases | null | null | https://aclanthology.org/L14-1401 | https://aclanthology.org/L14-1401.pdf | Semantic Clustering of Pivot Paraphrases | Paraphrases extracted from parallel corpora by the pivot method (Bannard and Callison-Burch, 2005) constitute a valuable resource for multilingual NLP applications. In this study, we analyse the semantics of unigram pivot paraphrases and use a graph-based sense induction approach to unveil hidden sense distinctions in the paraphrase sets. The comparison of the acquired senses to gold data from the Lexical Substitution shared task (McCarthy and Navigli, 2007) demonstrates that sense distinctions exist in the paraphrase sets and highlights the need for a disambiguation step in applications using this resource. | ['Emilia Verzeni', 'Diana McCarthy', 'Marianna Apidianaki'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['multilingual-nlp'] | ['natural-language-processing'] | [ 2.08167389e-01 -1.61341894e-02 -6.12859070e-01 -2.19166905e-01
-5.55778563e-01 -1.14067650e+00 8.08789313e-01 8.65910232e-01
-5.28596640e-01 9.89794791e-01 6.24689579e-01 -6.27014935e-01
-5.33849120e-01 -5.92487514e-01 -3.00679952e-01 -2.64168948e-01
4.39799219e-01 6.99493170e-01 2.24055946e-01 -9.42394018e-01
8.83508623e-01 4.62569773e-01 -1.31821716e+00 7.78110743e-01
1.12378407e+00 1.41348347e-01 5.51402628e-01 2.67498076e-01
-5.95811188e-01 7.59514809e-01 -6.11250877e-01 -8.32693279e-01
2.17531443e-01 -4.50058997e-01 -1.21385300e+00 -3.70291442e-01
6.42191529e-01 7.31497943e-01 8.31199214e-02 1.25295758e+00
3.61218840e-01 8.08147341e-02 3.51509750e-01 -1.16169667e+00
-5.86702228e-01 1.08848250e+00 -3.37240458e-01 7.69046605e-01
1.04139328e+00 -3.39065909e-01 1.65066218e+00 -9.77651119e-01
1.32365823e+00 1.14168572e+00 6.02649987e-01 7.62238130e-02
-1.31425619e+00 -5.29681444e-01 -2.24956393e-01 7.46665061e-01
-1.21142471e+00 -1.51320994e-01 8.08301568e-01 -4.69686240e-01
1.50587928e+00 3.78826350e-01 8.12460959e-01 1.06751764e+00
4.36729223e-01 6.69624686e-01 1.37055910e+00 -1.06093836e+00
8.84512737e-02 2.24344313e-01 5.09348631e-01 2.18357772e-01
4.20478106e-01 -2.43417937e-02 -7.32801437e-01 -5.54287434e-01
6.56463921e-01 -2.49436483e-01 -1.88020542e-01 -7.23058462e-01
-1.29108036e+00 1.01280963e+00 3.77183288e-01 1.08629787e+00
-3.36837888e-01 -5.58005273e-01 8.71358693e-01 1.05420828e+00
4.27558124e-01 1.07987297e+00 -4.61820066e-01 -1.65395483e-01
-1.06126094e+00 2.14352921e-01 1.07396603e+00 1.16064358e+00
6.40322685e-01 -5.00705123e-01 2.45551363e-01 1.19663787e+00
6.23924099e-02 2.66266853e-01 1.03200376e+00 -6.11563444e-01
6.34828329e-01 8.25815976e-01 -5.52047603e-02 -1.14736605e+00
-1.77849859e-01 -1.21569760e-01 -2.70728141e-01 -5.02523065e-01
1.22816607e-01 2.24715516e-01 -4.27377045e-01 1.48777354e+00
8.55339244e-02 -2.17109621e-01 3.66920173e-01 6.56724989e-01
7.96456158e-01 1.66015595e-01 1.11257717e-01 -3.64796668e-01
1.46182048e+00 -6.50513053e-01 -4.69104707e-01 -4.05858099e-01
7.56754100e-01 -1.26699030e+00 1.33553648e+00 2.17026368e-01
-8.10168505e-01 -5.40594518e-01 -1.22626960e+00 -6.37420416e-02
-4.99424845e-01 -1.78991571e-01 6.52509928e-01 4.88405943e-01
-8.61427963e-01 5.52108169e-01 -2.58766770e-01 -1.06271839e+00
-1.47107095e-01 7.70914704e-02 -5.29040515e-01 -2.39433646e-01
-1.53982282e+00 1.36590219e+00 7.92107403e-01 -2.99882233e-01
9.07385796e-02 -6.39653563e-01 -8.34604323e-01 -3.21500391e-01
3.93102944e-01 -6.80881381e-01 8.07439864e-01 -9.80161667e-01
-9.99226511e-01 1.58963895e+00 -1.16161995e-01 -4.60344017e-01
1.08435027e-01 -8.96357298e-02 -6.60935879e-01 8.48286971e-02
6.32185698e-01 1.23647586e-01 3.51886302e-01 -8.43714237e-01
-6.98687673e-01 -4.12068188e-01 3.17280412e-01 4.94516522e-01
1.91965714e-01 4.04901952e-01 1.12557337e-01 -8.87610197e-01
2.83159941e-01 -1.10456586e+00 2.45879274e-02 -9.20840085e-01
-2.51031965e-01 -2.89603114e-01 1.09015234e-01 -6.47697747e-01
1.27007508e+00 -1.89200807e+00 4.48990285e-01 4.27830368e-01
2.06889778e-01 8.82549956e-02 -2.45912805e-01 1.28678715e+00
-4.82966095e-01 9.92346480e-02 -2.74733961e-01 3.84016007e-01
-8.37844759e-02 5.75476050e-01 -4.27937895e-01 7.67853558e-02
-3.26627344e-01 9.66818631e-01 -1.39312184e+00 -5.87743998e-01
3.48785877e-01 -1.53763711e-01 -2.17028484e-01 -1.94579035e-01
9.86201763e-02 8.34316984e-02 -1.91665828e-01 2.18147218e-01
1.93552047e-01 3.01183239e-02 7.80724108e-01 -1.11156657e-01
-1.91903934e-01 9.88649011e-01 -7.58907259e-01 1.91847527e+00
-9.40910161e-01 6.35319889e-01 -4.15942699e-01 -8.32405984e-01
9.07961130e-01 2.31052414e-01 2.07610995e-01 -7.76130617e-01
-1.19977064e-01 5.68695903e-01 2.83173144e-01 -4.68621522e-01
8.77308428e-01 -5.45669198e-01 -4.56987649e-01 3.44258875e-01
2.05163971e-01 -5.59429407e-01 6.99910581e-01 4.18636948e-01
8.91430616e-01 -5.65805361e-02 1.17071319e+00 -8.55445087e-01
8.26953530e-01 8.25380802e-01 4.89786178e-01 6.30210817e-01
1.68733206e-02 2.75026590e-01 3.93778771e-01 -4.56091881e-01
-1.10525608e+00 -1.24535859e+00 -3.06802779e-01 1.08694756e+00
2.38552004e-01 -7.46703267e-01 -3.31976563e-01 -5.21667600e-01
8.93697888e-03 1.04309666e+00 -4.70130503e-01 4.27034497e-02
-7.52493083e-01 -3.69555026e-01 6.09496713e-01 2.04444915e-01
1.90386698e-01 -1.16916001e+00 -4.29758161e-01 4.70178246e-01
-5.09940207e-01 -9.74764705e-01 -1.75461993e-01 1.74019247e-01
-6.95996761e-01 -1.35574508e+00 -2.59129107e-01 -9.53051209e-01
1.28169820e-01 3.31377715e-01 1.45936239e+00 -9.05966572e-03
-2.15320453e-01 5.18741831e-02 -6.75439537e-01 -2.36352175e-01
-9.19151366e-01 3.49676937e-01 -1.13881361e-02 -8.09207678e-01
1.01352727e+00 -7.58926630e-01 -3.75704691e-02 1.37946576e-01
-7.66389310e-01 -2.20062271e-01 3.99772048e-01 1.02651179e+00
4.32193428e-01 -4.52425957e-01 5.77155530e-01 -1.18337119e+00
1.32085168e+00 -6.66809857e-01 -3.79467219e-01 5.72513640e-01
-6.82469130e-01 2.10882232e-01 3.65435660e-01 -1.12542607e-01
-8.93271863e-01 -3.75891387e-01 -9.82224941e-02 1.73153952e-01
-1.82793528e-01 7.22200513e-01 2.13511121e-02 -1.68360621e-01
1.06344318e+00 1.72379643e-01 -3.31500977e-01 -3.29811096e-01
6.79214239e-01 8.02408278e-01 3.95524770e-01 -7.62873650e-01
3.69221121e-01 7.79791400e-02 -2.21852154e-01 -1.04159880e+00
-7.28906989e-01 -1.07514167e+00 -1.00534046e+00 2.08083943e-01
5.27871668e-01 -8.85282755e-01 9.94066969e-02 -1.91456869e-01
-1.20445573e+00 1.11465745e-01 -5.26791155e-01 3.69398594e-01
-6.13702834e-01 8.11229169e-01 -6.29162073e-01 -7.35391825e-02
-2.92378753e-01 -7.67957091e-01 8.72423410e-01 -2.19506696e-01
-1.36385858e+00 -1.34117544e+00 6.81888700e-01 4.59186375e-01
-2.58346461e-02 -6.15915023e-02 1.32483733e+00 -1.25401187e+00
1.97372749e-01 1.41699016e-01 5.08778682e-03 -3.01382001e-02
3.51709604e-01 -2.46876627e-01 -3.13949198e-01 -3.89958695e-02
1.92105412e-01 -2.03029260e-01 4.44367170e-01 -1.54652566e-01
-6.62479922e-02 -3.52662295e-01 -2.53663272e-01 3.81726623e-02
1.42449903e+00 7.96976909e-02 4.79475170e-01 7.41687536e-01
5.45482039e-01 6.58006370e-01 6.80901170e-01 2.55826898e-02
1.49114326e-01 5.71130633e-01 -5.23499787e-01 5.47284722e-01
-1.06664419e-01 -4.16121185e-01 9.09632221e-02 1.38989365e+00
1.50913671e-01 -2.98191812e-02 -1.11867249e+00 8.87707829e-01
-2.01381612e+00 -9.41745162e-01 -4.30347323e-01 1.94814897e+00
9.97682691e-01 1.57140493e-01 1.93296243e-02 1.33367687e-01
7.07483470e-01 3.62729520e-01 1.49430737e-01 -9.27779853e-01
-3.99884909e-01 6.70270383e-01 3.33587646e-01 8.05918753e-01
-3.85353565e-01 1.52092230e+00 6.50103903e+00 9.22300041e-01
-6.57267511e-01 6.92152083e-02 -3.85657251e-01 3.25399846e-01
-5.95771134e-01 3.57977092e-01 -3.06345493e-01 4.12459821e-01
6.41820729e-01 -5.83473325e-01 3.26563746e-01 6.43545210e-01
-1.53292373e-01 -1.91052809e-01 -1.05321705e+00 9.31368291e-01
2.70383805e-01 -1.25100064e+00 1.88834459e-01 -4.23521340e-01
7.54534721e-01 5.39470375e-01 -6.51674032e-01 3.92212756e-02
7.17280567e-01 -4.76585537e-01 3.38589877e-01 -4.08311374e-02
2.78532743e-01 -6.93252504e-01 7.74513960e-01 2.65654922e-01
-1.01898360e+00 1.51893288e-01 -7.00715899e-01 -3.80759925e-01
1.88156754e-01 5.39080262e-01 -8.53454292e-01 8.02156091e-01
1.33603320e-01 7.10181296e-01 -6.25435770e-01 7.47987092e-01
-8.79504383e-01 3.79316658e-01 -6.04988821e-02 -2.88951486e-01
1.42219633e-01 -4.98035848e-01 9.38135445e-01 1.35324836e+00
8.78287107e-03 -1.94050372e-01 9.76857170e-02 5.89254737e-01
2.05497593e-01 6.84021473e-01 -9.34224725e-01 -1.14963636e-01
8.08636606e-01 9.87929583e-01 -8.81293893e-01 -4.29415315e-01
-3.37539136e-01 1.24422872e+00 2.73064673e-01 2.22662166e-01
-1.81291595e-01 -5.47571599e-01 7.24106252e-01 -1.67788401e-01
4.03257906e-02 -2.09041357e-01 -4.29199547e-01 -1.41006315e+00
4.30523530e-02 -1.01180136e+00 5.65966010e-01 -8.18453670e-01
-1.88394988e+00 6.61397457e-01 3.95941794e-01 -9.50741529e-01
-6.55412555e-01 -4.86326128e-01 -6.36646092e-01 1.01989484e+00
-1.02226412e+00 -1.13432825e+00 7.16101304e-02 5.65953791e-01
6.57039523e-01 -2.41977319e-01 7.89023280e-01 4.60583158e-02
1.06902942e-01 2.82047123e-01 -5.25908843e-02 -1.06940821e-01
7.88261294e-01 -1.25487649e+00 9.17322934e-01 6.73694968e-01
8.60259175e-01 1.26723695e+00 9.54039156e-01 -9.67321575e-01
-8.76824677e-01 -3.59138489e-01 1.67991638e+00 -6.11114860e-01
1.40606070e+00 -1.30011693e-01 -8.95360291e-01 6.71429873e-01
7.90730357e-01 -7.56445229e-01 1.03669798e+00 3.72907996e-01
-6.64496660e-01 3.91836256e-01 -1.06280446e+00 7.38469183e-01
1.37840831e+00 -1.06329727e+00 -1.83021319e+00 4.25935984e-01
5.63881695e-01 -1.94672421e-02 -9.19044077e-01 1.76482871e-01
3.50110859e-01 -9.90704298e-01 9.42867041e-01 -9.24827039e-01
4.13442224e-01 2.38872692e-02 -2.68278301e-01 -1.74303555e+00
-3.85488778e-01 -4.98169363e-01 7.56529391e-01 1.19073927e+00
5.79383373e-01 -6.72849715e-01 4.61630404e-01 1.87623411e-01
-8.42978656e-02 -1.41095683e-01 -1.08905172e+00 -8.46242189e-01
3.90362531e-01 -2.64187664e-01 3.87435317e-01 1.60003877e+00
1.17449236e+00 9.64378476e-01 1.60519734e-01 -4.79044616e-01
4.20863599e-01 4.31693017e-01 4.51100200e-01 -1.19053245e+00
-2.96454847e-01 -4.08825368e-01 -6.17261767e-01 -7.00921357e-01
5.93537509e-01 -1.47371471e+00 -2.95162350e-01 -1.59343350e+00
1.57639071e-01 -1.50306731e-01 -2.42569298e-01 1.94441542e-01
-1.19425669e-01 1.38836585e-05 3.23872834e-01 5.89883626e-01
-3.23601544e-01 -1.42939538e-02 7.71218061e-01 -3.82347144e-02
-3.17921907e-01 -3.76782179e-01 -8.56248319e-01 7.63738453e-01
8.19388211e-01 -6.86559141e-01 -6.42542005e-01 -2.56363243e-01
8.00267696e-01 3.77399325e-02 9.70579535e-02 -5.05594909e-01
3.92982066e-01 -6.06228411e-01 -3.01415294e-01 -1.87812924e-01
-1.20707087e-01 -6.73232913e-01 2.88886368e-01 5.05128562e-01
-4.87362891e-01 7.34171331e-01 2.19487488e-01 2.65251160e-01
-4.74042684e-01 -7.13187516e-01 3.82880807e-01 -3.86395901e-01
-1.13591146e+00 -5.97481072e-01 -4.37529027e-01 5.26715636e-01
7.68181384e-01 -3.59610379e-01 -1.86141610e-01 4.90377769e-02
-5.70555568e-01 -3.74348350e-02 8.58942866e-01 7.31780648e-01
3.60375285e-01 -1.39617062e+00 -5.89508653e-01 6.77354932e-02
6.42223835e-01 -7.22185850e-01 -8.84560198e-02 5.73768198e-01
-5.84146023e-01 4.20895964e-01 -6.90378845e-01 -2.01985717e-01
-1.41854131e+00 6.76838636e-01 -1.81777284e-01 -5.04600585e-01
-3.46626103e-01 6.81151867e-01 -1.91658229e-01 -5.03472209e-01
-7.15993822e-01 -2.31869951e-01 -2.60254383e-01 3.07506174e-01
7.93463886e-02 3.34453791e-01 2.76222885e-01 -7.97565639e-01
-6.26112223e-01 3.99217308e-01 -2.31782734e-01 -3.19206208e-01
1.17114103e+00 -3.05788159e-01 -5.59082389e-01 7.62437761e-01
1.13840568e+00 4.97899055e-01 2.42076486e-01 -2.61474580e-01
8.04277658e-01 -6.49437129e-01 -5.74143767e-01 -7.43835330e-01
-3.07101816e-01 4.18687582e-01 -9.33712050e-02 3.57708603e-01
6.35571122e-01 1.17866725e-01 5.86425960e-01 5.07808387e-01
7.07074881e-01 -1.22044933e+00 -5.23215711e-01 8.13520789e-01
1.06456697e+00 -8.86131823e-01 -5.63076837e-03 -8.26484263e-01
-8.34643781e-01 9.49362695e-01 1.09082572e-01 -3.15410048e-01
3.41497600e-01 1.59710944e-02 1.62268564e-01 -4.32540804e-01
-5.22650242e-01 -2.35086322e-01 3.70990187e-01 6.07952833e-01
6.99624240e-01 2.82690208e-02 -1.45982516e+00 3.72400939e-01
-7.21213281e-01 -1.60652816e-01 6.55200541e-01 9.52562451e-01
-2.08322927e-01 -1.75250244e+00 -4.18493859e-02 3.92014205e-01
-2.37460241e-01 -8.03724587e-01 -1.29086995e+00 1.16991520e+00
-5.03191017e-02 9.75934207e-01 -2.26184830e-01 -2.98750371e-01
4.88090992e-01 3.77537996e-01 8.35453272e-01 -1.12361848e+00
-1.15948308e+00 -3.03540260e-01 6.60230100e-01 -3.99188608e-01
-6.74996674e-01 -5.17304659e-01 -9.85503435e-01 -2.66974956e-01
-2.72145659e-01 6.71830237e-01 3.69050324e-01 1.18716180e+00
3.43093485e-01 5.22140115e-02 1.69762850e-01 -1.76408455e-01
-4.78617758e-01 -1.00797081e+00 -5.51204801e-01 9.14531410e-01
-3.63253027e-01 -6.18761778e-01 -3.45271498e-01 -6.52707070e-02] | [10.71280288696289, 9.524370193481445] |
60c082dd-6110-478b-997f-97754d8ed46a | end-to-end-automatic-speech-translation-of | 1802.04200 | null | http://arxiv.org/abs/1802.04200v1 | http://arxiv.org/pdf/1802.04200v1.pdf | End-to-End Automatic Speech Translation of Audiobooks | We investigate end-to-end speech-to-text translation on a corpus of
audiobooks specifically augmented for this task. Previous works investigated
the extreme case where source language transcription is not available during
learning nor decoding, but we also study a midway case where source language
transcription is available at training time only. In this case, a single model
is trained to decode source speech into target text in a single pass.
Experimental results show that it is possible to train compact and efficient
end-to-end speech translation models in this setup. We also distribute the
corpus and hope that our speech translation baseline on this corpus will be
challenged in the future. | ['Alexandre Bérard', 'Laurent Besacier', 'Ali Can Kocabiyikoglu', 'Olivier Pietquin'] | 2018-02-12 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 3.98794800e-01 4.18387949e-01 9.84089524e-02 -3.22793901e-01
-1.53484392e+00 -7.52440691e-01 5.45614898e-01 -2.43377805e-01
-3.61580372e-01 8.54815423e-01 4.11831468e-01 -7.44828463e-01
5.51863134e-01 -1.84803993e-01 -7.70977974e-01 -4.01154548e-01
2.28554577e-01 7.42200434e-01 9.81556401e-02 -3.15671921e-01
-2.71701664e-01 1.56441644e-01 -9.39854145e-01 5.99526882e-01
6.43274546e-01 3.00531089e-01 4.62542802e-01 1.11796772e+00
-9.84185562e-02 4.73611534e-01 -8.54408205e-01 -5.59929669e-01
2.43695885e-01 -7.66691387e-01 -9.44459498e-01 2.96345890e-01
3.65204394e-01 -2.15469539e-01 -2.59267241e-01 7.55044878e-01
9.11326706e-01 -2.90232133e-02 4.95090336e-01 -8.12942624e-01
-3.44634205e-01 8.49895775e-01 -8.20432156e-02 2.43824214e-01
6.28618777e-01 -7.93174580e-02 7.31044173e-01 -1.00052750e+00
5.45842409e-01 1.01136744e+00 9.84348804e-02 5.60425043e-01
-1.11313140e+00 -4.30676162e-01 -1.96486432e-03 -7.94254094e-02
-1.26749957e+00 -1.25883389e+00 5.05199850e-01 -5.97668923e-02
1.27870190e+00 3.00763309e-01 3.44151288e-01 1.49123049e+00
6.48069754e-02 8.38688731e-01 8.19074094e-01 -8.43992651e-01
1.34839371e-01 2.25099891e-01 -4.47986692e-01 3.50274444e-01
-4.73165154e-01 -4.85002398e-02 -8.70866239e-01 1.64619371e-01
4.45958793e-01 -7.28804052e-01 -3.79031569e-01 2.85141110e-01
-1.59058833e+00 5.72248995e-01 -2.73541391e-01 3.76431644e-01
-3.09444040e-01 -2.76159286e-01 5.32609761e-01 1.06067801e+00
8.50824833e-01 3.28545421e-01 -3.84359866e-01 -7.68151879e-01
-1.36230493e+00 -1.64212510e-02 1.19427264e+00 1.27281547e+00
1.96104273e-01 2.30689302e-01 -9.09068957e-02 1.01472831e+00
-7.17713758e-02 7.28912711e-01 6.09778881e-01 -8.94254208e-01
1.18438077e+00 -3.80711436e-01 1.15380809e-01 -1.22796282e-01
-1.42367527e-01 -6.08129084e-01 -4.26175892e-01 -1.68106988e-01
3.99644405e-01 -6.44090712e-01 -8.04975450e-01 1.49847484e+00
8.86514783e-02 1.55458897e-01 5.01063585e-01 8.00427973e-01
5.66035450e-01 9.49866414e-01 -5.15528619e-01 -5.75799823e-01
9.44974184e-01 -1.28995502e+00 -9.46117818e-01 -4.19363528e-01
7.52996802e-01 -1.48398459e+00 1.23164165e+00 3.64354789e-01
-1.49021661e+00 -3.75333369e-01 -9.58422899e-01 9.55186784e-02
5.37072010e-02 4.54673439e-01 3.75451297e-02 7.71802843e-01
-1.35077047e+00 3.57344359e-01 -1.01438892e+00 -5.24299204e-01
-2.68324941e-01 4.38682079e-01 -2.65412718e-01 -3.73158343e-02
-1.14462757e+00 9.06527162e-01 -2.34644543e-02 6.14663102e-02
-8.39878201e-01 1.90327615e-02 -5.97472668e-01 4.14317697e-02
2.29901418e-01 -5.89447618e-01 1.98574829e+00 -1.20244384e+00
-2.07184076e+00 6.76921844e-01 -5.55817604e-01 -3.45053017e-01
5.72399378e-01 -9.40538570e-03 -4.70981389e-01 1.90396994e-01
3.52404229e-02 3.57358217e-01 8.29895139e-01 -9.08032596e-01
-5.07426858e-01 -1.23800844e-01 -5.05003668e-02 5.99686444e-01
-3.57965320e-01 5.09449780e-01 -4.51791704e-01 -6.35369718e-01
-4.78355847e-02 -1.20592749e+00 2.32590914e-01 -6.06948733e-01
-3.55057240e-01 3.60846557e-02 5.26680708e-01 -1.12549222e+00
1.15682399e+00 -1.94914865e+00 2.97232896e-01 -2.50384063e-02
-4.80590165e-01 3.09225678e-01 -4.04730022e-01 9.27574694e-01
5.42218648e-02 4.85233590e-02 -1.90485954e-01 -1.00346529e+00
-7.93009475e-02 1.16940230e-01 -5.54284811e-01 3.38260114e-01
5.01673296e-02 6.33842647e-01 -5.75250685e-01 -4.75723833e-01
-8.84563401e-02 4.59531754e-01 -3.65773797e-01 4.28124994e-01
-7.28548840e-02 7.13827372e-01 -1.21492870e-01 4.09967691e-01
1.06167160e-01 2.53503859e-01 1.21775661e-02 6.14528537e-01
-1.92885846e-01 1.09471631e+00 -8.90286267e-01 1.97999775e+00
-1.00605810e+00 9.97446895e-01 3.62921357e-01 -8.24241757e-01
7.84419179e-01 1.04443359e+00 2.68610734e-02 -6.07979000e-01
-4.91065830e-02 4.86110419e-01 1.46171808e-01 -4.49562758e-01
5.27452409e-01 -2.71370351e-01 6.57071397e-02 6.13468885e-01
1.19811095e-01 -3.99621665e-01 1.46538526e-01 4.89102192e-02
1.00934732e+00 1.07651778e-01 -1.14184797e-01 1.38884902e-01
2.04884514e-01 1.57251321e-02 9.89073049e-03 4.90101308e-01
5.67606427e-02 1.00827932e+00 1.71878725e-01 3.21699560e-01
-1.24310875e+00 -9.56454396e-01 2.43941948e-01 1.26275146e+00
-5.29689193e-01 -4.67711776e-01 -1.05236769e+00 -4.82612669e-01
-9.29209471e-01 1.01363015e+00 1.15310840e-01 1.71869591e-01
-8.50209951e-01 -2.53723353e-01 7.48613775e-01 2.27260709e-01
1.25569001e-01 -9.12132084e-01 7.15350360e-02 3.94133419e-01
-6.37490213e-01 -1.34463429e+00 -9.16196764e-01 2.48980626e-01
-9.53016877e-01 -3.74024987e-01 -1.02856338e+00 -1.20140767e+00
3.76132727e-01 2.74861395e-01 9.88065302e-01 -1.89213142e-01
6.46287978e-01 2.91630924e-01 -5.51979363e-01 -3.81238163e-01
-1.09013128e+00 6.03839755e-01 3.21763381e-02 -1.08001865e-01
3.74549925e-02 -6.72835588e-01 -1.32975176e-01 3.86972100e-01
-5.49499154e-01 2.55943209e-01 5.13217866e-01 7.17475951e-01
1.97400212e-01 -3.19359452e-01 8.07699561e-01 -3.92039925e-01
7.68956602e-01 -3.79924983e-01 -3.71575743e-01 2.19760090e-01
-1.97668597e-01 -1.68370515e-01 9.16480839e-01 -4.87882495e-01
-9.94655013e-01 1.67775184e-01 -6.06160879e-01 -3.42861563e-01
-2.15484947e-01 6.30802631e-01 -2.98189700e-01 2.22114652e-01
5.76910496e-01 5.25857329e-01 -9.14448127e-02 -4.51691210e-01
3.06277666e-02 1.50507295e+00 4.96638864e-01 -5.13047934e-01
5.68609834e-01 -1.76234499e-01 -4.95695442e-01 -1.04997492e+00
-2.95977831e-01 -4.21655566e-01 -5.80854952e-01 -9.12738666e-02
5.61769962e-01 -1.17106462e+00 -8.75868127e-02 1.09394349e-01
-1.56181061e+00 -6.56427503e-01 -5.55824600e-02 8.11995924e-01
-9.23903823e-01 2.34956786e-01 -5.76391578e-01 -8.52175176e-01
-2.18617529e-01 -1.29402387e+00 1.32777321e+00 -2.99436033e-01
-3.07235301e-01 -1.11689806e+00 9.48079377e-02 4.26971078e-01
3.26615900e-01 -4.86080348e-01 4.22489613e-01 -8.67752671e-01
-3.67835313e-01 -1.66708767e-01 3.79272789e-01 4.04907465e-01
1.16558626e-01 -7.87320286e-02 -1.03362644e+00 -4.93398875e-01
1.41404524e-01 -3.32826912e-01 3.61796260e-01 1.98588416e-01
3.39344531e-01 -5.22123814e-01 5.60437180e-02 2.53188848e-01
7.39327908e-01 2.87594527e-01 5.62566698e-01 9.27460864e-02
5.36099613e-01 6.86118901e-01 4.83805686e-01 -8.32728967e-02
4.54552710e-01 1.12096643e+00 -1.43118709e-01 -2.19479740e-01
-3.69903028e-01 -3.64507139e-01 1.02890337e+00 1.57225144e+00
1.77014217e-01 -7.07188249e-01 -1.02765572e+00 8.69446874e-01
-1.58554268e+00 -7.65409768e-01 -6.48100451e-02 2.24075508e+00
1.03880143e+00 1.79817677e-01 3.78163606e-01 2.02978879e-01
7.86867082e-01 -6.29146695e-02 6.90829083e-02 -5.95172226e-01
9.16201845e-02 2.78759718e-01 4.24560547e-01 9.12473857e-01
-6.98934197e-01 1.09641802e+00 6.72708702e+00 7.49012589e-01
-1.36288309e+00 6.23793304e-01 3.70448291e-01 -4.92702514e-01
-1.89792231e-01 9.46757495e-02 -6.68905735e-01 4.85259056e-01
1.75676465e+00 -1.20164007e-01 6.90415025e-01 4.44972783e-01
6.14026904e-01 1.08657509e-01 -1.19045877e+00 8.80723536e-01
-5.03333621e-02 -8.91572893e-01 -3.21478963e-01 5.07987812e-02
5.87726533e-01 3.42991650e-01 -1.36613563e-01 3.28251004e-01
-5.58642969e-02 -8.11578095e-01 9.05735135e-01 -1.10448383e-01
1.04019654e+00 -7.10390806e-01 5.48544526e-01 7.92654514e-01
-9.35235083e-01 3.19326401e-01 -1.36119947e-01 -2.22354501e-01
5.13034165e-01 2.93374956e-01 -1.57215571e+00 6.46746457e-01
-2.66699698e-02 3.89656872e-01 -3.34952444e-01 8.94082904e-01
-3.02432179e-01 1.25954235e+00 -4.15750116e-01 -4.94432561e-02
2.82912046e-01 1.28581505e-02 7.52675295e-01 1.49311674e+00
9.30110872e-01 -2.38109276e-01 2.30928108e-01 2.81491339e-01
-3.71384807e-02 3.76949936e-01 -7.86387563e-01 -1.77710876e-01
4.08996403e-01 6.96721554e-01 -6.51538789e-01 -2.47671202e-01
-4.73663956e-01 1.43182635e+00 1.10482536e-01 5.86239815e-01
-4.98162419e-01 -4.26977366e-01 2.99561411e-01 1.81454137e-01
2.40009591e-01 -6.50351107e-01 -2.67152727e-01 -1.35901535e+00
1.23575777e-01 -1.10065138e+00 -4.01260145e-02 -7.67006814e-01
-7.87040293e-01 9.11589265e-01 -1.89281017e-01 -1.23906434e+00
-9.20031250e-01 -1.12653933e-01 -5.78133941e-01 1.13919532e+00
-1.31142235e+00 -1.15054715e+00 4.69356149e-01 5.23946166e-01
1.15841508e+00 -3.78973871e-01 1.10872614e+00 4.52225089e-01
-3.63139659e-01 7.78890550e-01 2.89656788e-01 1.10601343e-01
9.86862063e-01 -1.13826191e+00 8.00689816e-01 1.01648903e+00
6.02879226e-01 5.82115114e-01 1.10205865e+00 -5.78793347e-01
-1.29767835e+00 -9.83617187e-01 1.52150297e+00 -3.98609042e-01
4.64723945e-01 -8.03887665e-01 -7.65964866e-01 8.08354378e-01
8.22266221e-01 -3.68102968e-01 7.29982197e-01 1.21105604e-01
2.01671571e-01 7.72637650e-02 -7.36301124e-01 7.06861615e-01
9.29234147e-01 -9.14740086e-01 -5.84887803e-01 5.64184666e-01
1.05913591e+00 -7.21506834e-01 -6.10738277e-01 -1.73009306e-01
3.00805360e-01 -5.62152863e-01 3.88306230e-01 -2.88752526e-01
2.95882881e-01 -1.81790322e-01 -2.69093961e-01 -1.74416697e+00
4.28656131e-01 -1.35298097e+00 1.25589207e-01 1.32694972e+00
9.76070464e-01 -4.74615306e-01 5.13764739e-01 1.91653043e-01
-6.04002178e-01 -3.65960687e-01 -1.33633471e+00 -9.80028450e-01
1.97343126e-01 -4.85422671e-01 4.12715018e-01 6.15186691e-01
3.27821702e-01 7.64561594e-01 -4.06762809e-01 2.28110284e-01
9.46274251e-02 -2.19868466e-01 8.02916765e-01 -5.16643822e-01
-6.42120898e-01 -5.47150001e-02 6.43700734e-02 -1.35952711e+00
4.14162010e-01 -1.04761553e+00 2.92937994e-01 -1.57109964e+00
-3.19647402e-01 -8.14004093e-02 3.06759626e-01 1.97373539e-01
6.11472689e-02 1.71878815e-01 3.37509781e-01 1.23366930e-01
-2.90060639e-01 5.34025848e-01 1.24243188e+00 1.69922262e-02
-4.06161129e-01 4.93794680e-01 -4.12438244e-01 8.37775022e-02
8.78735602e-01 -7.28524566e-01 -6.85228884e-01 -6.24223769e-01
-1.98113307e-01 6.80776000e-01 -9.75212380e-02 -8.71306181e-01
2.36811578e-01 -5.11527993e-03 -1.63408786e-01 -3.20352525e-01
5.75738966e-01 -6.44317448e-01 -1.59629360e-02 -7.68124983e-02
-4.76228714e-01 1.74146056e-01 2.29078367e-01 7.05780461e-02
-5.23886085e-01 -4.32569683e-01 4.28348541e-01 4.45846617e-02
1.58935010e-01 -6.10671453e-02 -7.18544602e-01 -2.23508757e-02
5.41265726e-01 -6.43805787e-02 -3.29347141e-02 -1.03581047e+00
-9.37822163e-01 -1.59696892e-01 4.58355397e-01 4.37506527e-01
3.82323205e-01 -1.22275317e+00 -1.16974509e+00 2.34137058e-01
-1.59029692e-01 -2.40926713e-01 -3.22830856e-01 9.80161488e-01
-2.13596955e-01 6.26689434e-01 2.58183032e-01 -5.51903009e-01
-1.51132178e+00 1.76587746e-01 2.80804545e-01 -1.18686311e-01
-3.91843647e-01 6.29323006e-01 -1.84925571e-01 -2.94435591e-01
3.54229361e-01 -1.22904100e-01 2.17969581e-01 -1.59474790e-01
4.83708143e-01 3.34701538e-02 5.16097724e-01 -7.18289912e-01
-1.07679553e-02 2.40285426e-01 1.89322401e-02 -1.06137455e+00
1.11187541e+00 -5.37721634e-01 3.05891901e-01 7.84555197e-01
1.19685698e+00 4.45456356e-01 -1.08526444e+00 -1.47453740e-01
-1.00439556e-01 -2.26506367e-01 4.63790819e-03 -7.96444416e-01
-6.36985600e-01 1.08135867e+00 2.95825541e-01 1.73472568e-01
1.23803329e+00 -1.35266110e-01 8.75961125e-01 6.13896310e-01
2.92597324e-01 -1.21668446e+00 -1.95509210e-01 8.76967132e-01
1.08024263e+00 -1.01210141e+00 -5.69652319e-01 -2.95614272e-01
-8.08444321e-01 1.15417325e+00 1.80724129e-01 3.14318120e-01
1.08824939e-01 5.77602088e-01 3.91950935e-01 4.06107843e-01
-9.68898535e-01 -2.11914480e-01 1.79725826e-01 5.80845118e-01
8.92115474e-01 5.16225435e-02 -1.24059811e-01 2.96799093e-01
-7.48431385e-01 -1.16067305e-01 7.56857276e-01 9.75625336e-01
-2.51260370e-01 -1.59686816e+00 -5.02656937e-01 -8.50640461e-02
-7.81840980e-01 -3.59664679e-01 -6.51927888e-01 3.66777450e-01
-4.17285383e-01 1.55388057e+00 -1.51254520e-01 -3.08587134e-01
4.14208889e-01 5.35893679e-01 5.70235610e-01 -1.00587499e+00
-7.57315218e-01 7.50597298e-01 6.72275186e-01 -1.87892035e-01
-1.98042482e-01 -6.53395414e-01 -1.17047417e+00 -1.51706964e-01
-3.67994756e-01 3.63510311e-01 1.06802964e+00 9.57481802e-01
2.81370521e-01 6.27587795e-01 6.70626879e-01 -9.26077604e-01
-5.44917941e-01 -1.24486792e+00 -3.59781653e-01 -3.68035227e-01
7.86928535e-01 7.14260265e-02 -2.62666136e-01 4.52788889e-01] | [14.542235374450684, 7.043604850769043] |
e89b62f2-278f-4031-bc49-59853718d71c | progressively-parsing-interactional-objects | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Ni_Progressively_Parsing_Interactional_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Ni_Progressively_Parsing_Interactional_CVPR_2016_paper.pdf | Progressively Parsing Interactional Objects for Fine Grained Action Detection | Fine grained video action analysis often requires reliable detection and tracking of various interacting objects and human body parts, denoted as interactional object parsing. However, most of the previous methods based on either independent or joint object detection might suffer from high model complexity and challenging image content, e.g., illumination/pose/appearance/scale variation, motion, occlusion etc. In this work, we propose an end-to-end system based on recursive neural network to perform frame by frame interactional object parsing, which can alleviate the difficulty through a incremental manner. Our key innovation is that: instead of jointly outputting all object detections at once, for each frame, we use a set of long-short term memory (LSTM) nodes to incrementally refine the detections. After passing each LSTM node, more object detections are consolidated and thus more contextual information could be utilized to determine more difficult object detections. Extensive experiments on two benchmark fine grained activity datasets demonstrate that our proposed algorithm achieves better interacting object detection performance, which in turn boosts the action recognition performance over the state-of-the-art. | ['Bingbing Ni', 'Shenghua Gao', 'Xiaokang Yang'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['fine-grained-action-detection', 'action-analysis'] | ['computer-vision', 'computer-vision'] | [ 3.53314489e-01 -1.35131210e-01 6.30211905e-02 -2.98553795e-01
-5.59325755e-01 -3.38610768e-01 4.13911968e-01 -9.22369584e-02
-5.02344549e-01 6.81647003e-01 9.77322236e-02 2.50768006e-01
1.22737288e-01 -4.49921131e-01 -8.29700947e-01 -7.18749583e-01
1.53469786e-01 1.48185700e-01 8.72715950e-01 2.29814202e-01
1.88061148e-01 3.20258528e-01 -1.48959363e+00 2.85904109e-01
5.23126125e-01 1.13829613e+00 3.62344772e-01 6.91154242e-01
1.22071907e-03 1.14925289e+00 -5.59616625e-01 -6.74732625e-02
1.88366733e-02 -2.27028161e-01 -6.10660136e-01 4.84600753e-01
4.32836503e-01 -8.41498375e-01 -2.21473232e-01 1.12801754e+00
4.59425509e-01 3.05853754e-01 3.21444154e-01 -1.07467330e+00
-3.44457954e-01 4.32525069e-01 -9.16479409e-01 4.98651803e-01
3.06597501e-01 3.28954309e-01 5.74209929e-01 -9.89566326e-01
2.91073620e-01 1.52557969e+00 3.65600377e-01 5.89859903e-01
-7.63913572e-01 -7.61033535e-01 8.27394903e-01 4.63804781e-01
-1.17533445e+00 -5.33253670e-01 7.49710739e-01 -2.70289093e-01
8.35224628e-01 -8.14092010e-02 6.27371132e-01 9.61552739e-01
2.14505047e-02 1.14242184e+00 6.32163644e-01 -2.39769623e-01
1.20930806e-01 -3.35667551e-01 5.00288904e-02 9.91892517e-01
2.41170004e-01 -2.41033316e-01 -3.89079511e-01 9.30410549e-02
1.00659382e+00 3.98069561e-01 -9.36211199e-02 -4.28374231e-01
-1.32754409e+00 3.35412055e-01 3.81988585e-01 2.31545001e-01
-6.25895679e-01 4.87053216e-01 5.62733293e-01 -1.73160955e-01
1.44032389e-01 -3.59354615e-01 -5.82121849e-01 -2.48731568e-01
-6.54375434e-01 1.37185842e-01 4.28390682e-01 6.97444797e-01
6.16046906e-01 -3.47763151e-02 -4.77248222e-01 7.47832954e-01
5.20365953e-01 3.39267641e-01 4.16802377e-01 -8.52936387e-01
6.87094212e-01 7.90244222e-01 2.54760742e-01 -9.96594429e-01
-4.16725457e-01 -4.63674247e-01 -6.63772404e-01 2.76012361e-01
3.43801886e-01 -2.38540858e-01 -1.03666532e+00 1.69282007e+00
7.12812245e-01 4.50032085e-01 -1.38456360e-01 1.10572624e+00
7.03185916e-01 4.56351101e-01 5.23118615e-01 -1.87768787e-01
1.51197076e+00 -1.38569570e+00 -6.79721117e-01 -4.86230463e-01
4.95249242e-01 -6.53489232e-01 7.71858394e-01 2.83969700e-01
-1.16523087e+00 -1.05117416e+00 -9.29103673e-01 -1.42110484e-02
-1.06189370e-01 4.05991167e-01 6.13792062e-01 3.58914435e-01
-6.21883929e-01 3.44780713e-01 -1.28868711e+00 -2.08254784e-01
7.41407990e-01 4.48718458e-01 -3.78764063e-01 3.89142223e-02
-7.23921955e-01 4.87133861e-01 6.60646081e-01 6.65269256e-01
-1.13127148e+00 -3.55034806e-02 -8.38281512e-01 7.11794868e-02
7.98785686e-01 -6.86237097e-01 1.36959350e+00 -1.09634519e+00
-1.43926573e+00 4.45615798e-01 -2.76408792e-01 -3.86446655e-01
5.03525734e-01 -5.50237536e-01 -1.24396540e-01 8.15057531e-02
8.20958689e-02 7.06534684e-01 8.91683936e-01 -8.85215640e-01
-1.08218765e+00 -6.19809449e-01 2.62949526e-01 3.14769626e-01
-1.89833313e-01 2.43838653e-01 -9.28315222e-01 -5.57837009e-01
2.29566604e-01 -9.06169593e-01 -2.55163759e-01 1.47275731e-01
-2.17663467e-01 -5.75860322e-01 1.05343270e+00 -6.00565851e-01
1.11168182e+00 -2.16380429e+00 2.11824775e-01 -3.42207402e-01
1.94543257e-01 4.37626183e-01 -1.47879407e-01 -4.64330253e-04
9.24908295e-02 -2.21317127e-01 4.74330485e-02 -4.32288319e-01
-1.35817811e-01 1.37299120e-01 1.64770141e-01 5.12612283e-01
3.40467721e-01 1.01220226e+00 -8.04544687e-01 -8.06974471e-01
6.35195613e-01 4.13966566e-01 -4.18547362e-01 9.50096473e-02
-3.54994088e-01 5.41277647e-01 -8.26243579e-01 8.55699301e-01
5.60601950e-01 -3.45683932e-01 -2.07320765e-01 -3.41955900e-01
1.56136975e-02 -2.20257595e-01 -1.49242616e+00 1.79940486e+00
-1.39534101e-01 2.43770868e-01 2.06421062e-01 -1.08448577e+00
6.39415801e-01 2.24094048e-01 4.12995815e-01 -7.00285196e-01
2.58696020e-01 -1.16575241e-01 1.98238716e-01 -7.24435508e-01
1.76423341e-01 2.73596555e-01 -1.41237108e-02 2.48770118e-01
-1.01546913e-01 7.83513188e-01 1.86875507e-01 -9.22429711e-02
1.03726947e+00 5.06732583e-01 3.22693169e-01 2.55824447e-01
8.87323201e-01 -2.90816933e-01 8.61529350e-01 6.79782212e-01
-7.11402476e-01 2.73358524e-01 9.39556286e-02 -6.03658319e-01
-6.00659609e-01 -9.24986482e-01 2.92020917e-01 1.50792778e+00
4.47654694e-01 -1.22809887e-01 -8.97974730e-01 -7.90327370e-01
-3.31142902e-01 2.01829955e-01 -5.78939915e-01 -1.11088581e-01
-9.46633995e-01 -5.72617590e-01 3.71567726e-01 9.24673080e-01
7.83055782e-01 -1.54208004e+00 -1.01687706e+00 5.79988480e-01
-1.60050318e-01 -1.22035205e+00 -4.75681484e-01 -5.83350770e-02
-9.56765294e-01 -1.10950863e+00 -6.33270860e-01 -7.86511660e-01
6.67776287e-01 4.27446842e-01 7.64288187e-01 1.35861292e-01
-4.08824384e-01 1.37314141e-01 -4.22874391e-01 -2.23120049e-01
8.00483674e-02 -2.45362937e-01 -3.95945497e-02 2.77145565e-01
3.54713827e-01 -3.55617881e-01 -1.02583778e+00 4.21236604e-01
-7.58167028e-01 1.77575022e-01 8.77801895e-01 6.07475102e-01
7.93195784e-01 1.63406074e-01 4.77496743e-01 -6.44570589e-01
2.08898351e-01 -8.00869614e-02 -4.73463476e-01 4.13137138e-01
6.29492551e-02 -6.63350895e-02 3.83681923e-01 -5.79245329e-01
-1.39568377e+00 4.56392646e-01 -4.32279706e-02 -5.89979053e-01
-4.73492920e-01 1.30544946e-01 -3.57153386e-01 -3.70478407e-02
2.69307375e-01 3.67331386e-01 -4.56685781e-01 -5.70546329e-01
3.42607230e-01 5.42013943e-01 6.42974198e-01 -4.06701058e-01
5.62981606e-01 4.90637839e-01 -2.38498434e-01 -3.65778327e-01
-8.32935274e-01 -3.41631949e-01 -7.97252595e-01 -3.85793477e-01
1.27657211e+00 -1.01483536e+00 -8.62036407e-01 7.72384048e-01
-1.20355010e+00 -1.57798260e-01 1.55503228e-01 5.65557003e-01
-3.24882001e-01 4.13695157e-01 -5.65667570e-01 -9.76197362e-01
-4.59639668e-01 -1.19276643e+00 1.50861073e+00 5.47515094e-01
9.63762775e-02 -6.11794651e-01 -3.04355264e-01 4.98153359e-01
1.20496057e-01 3.15305382e-01 4.68694866e-01 -2.82345980e-01
-7.78148592e-01 -3.69504154e-01 -4.35842991e-01 1.64633989e-01
2.70720422e-01 -9.60614160e-02 -7.19584823e-01 -2.40727574e-01
6.11307565e-03 -4.03456658e-01 7.59498298e-01 4.27640051e-01
1.29270697e+00 -2.55100310e-01 -4.93878245e-01 2.61688441e-01
1.07362974e+00 4.14029002e-01 3.70681077e-01 1.42373115e-01
1.08857703e+00 4.10457015e-01 9.89534080e-01 6.10464454e-01
3.81620198e-01 7.09501028e-01 4.82750386e-01 -2.08068252e-01
-7.32197315e-02 -1.20447792e-01 3.37131441e-01 3.11200291e-01
-2.09249526e-01 -8.20733160e-02 -4.86247540e-01 3.24101090e-01
-2.22238374e+00 -9.61271286e-01 2.30421983e-02 1.89689577e+00
4.45886910e-01 4.08437639e-01 2.50153512e-01 1.50841614e-02
8.94717693e-01 1.50040850e-01 -9.24945593e-01 1.83814660e-01
2.15125263e-01 -7.09428862e-02 3.71875823e-01 1.04258619e-01
-1.25488174e+00 1.25225341e+00 5.11046028e+00 7.79973567e-01
-9.99128640e-01 1.27878234e-01 5.79806268e-01 -2.35440165e-01
4.73314375e-01 -2.40176514e-01 -9.35793877e-01 4.50057805e-01
3.19518238e-01 3.59874934e-01 2.42142543e-01 9.88334477e-01
4.70124513e-01 -2.66086400e-01 -1.07931840e+00 1.13828039e+00
5.35742119e-02 -9.33562815e-01 -7.51097947e-02 -1.78809509e-01
4.55003768e-01 -3.91627587e-02 -4.62862462e-01 4.29397792e-01
1.36644080e-01 -6.14651144e-01 8.15620542e-01 5.75338662e-01
1.21832177e-01 -6.01391315e-01 6.02673590e-01 5.51088512e-01
-1.61675704e+00 -2.38510028e-01 -3.59474093e-01 -2.31456339e-01
3.50188285e-01 2.20360637e-01 -3.53293985e-01 3.13700378e-01
8.71136367e-01 7.65366673e-01 -7.22103119e-01 9.93610680e-01
-9.74691585e-02 3.38631511e-01 -2.68904924e-01 -4.51695956e-02
3.35411489e-01 1.20073393e-01 3.35016310e-01 1.03639233e+00
-9.87471342e-02 4.01835114e-01 7.48935699e-01 4.58122969e-01
1.11454010e-01 -1.62659913e-01 -2.68622637e-02 1.02097340e-01
3.26782823e-01 1.42315352e+00 -1.05199385e+00 -6.03360474e-01
-6.23999953e-01 1.16911077e+00 4.73082125e-01 3.39203805e-01
-1.11968672e+00 -1.95850119e-01 3.15609574e-01 -3.35535146e-02
7.19149828e-01 -3.27984005e-01 1.68236181e-01 -1.20882654e+00
3.64657402e-01 -8.63207281e-01 5.95004141e-01 -6.94521189e-01
-9.60412323e-01 3.88223410e-01 -2.11417004e-01 -1.08977544e+00
-2.20540851e-01 -4.65927720e-01 -5.66329241e-01 4.04840708e-01
-1.11196196e+00 -1.33542943e+00 -6.65751696e-01 8.19784641e-01
1.05289781e+00 2.52107263e-01 4.53473985e-01 5.02460003e-01
-9.32931602e-01 4.47580308e-01 -4.20524687e-01 4.20919329e-01
4.51408565e-01 -1.04338968e+00 3.32063556e-01 9.40471172e-01
1.43235609e-01 3.90012294e-01 3.59057069e-01 -7.20555007e-01
-1.34184849e+00 -1.18693018e+00 3.08376342e-01 -2.92515934e-01
4.72421318e-01 -3.61257881e-01 -6.82080865e-01 6.68440342e-01
-2.01428056e-01 3.54900509e-01 2.30186656e-01 -5.24684712e-02
-4.10128850e-03 -2.13184372e-01 -9.41805840e-01 6.14565969e-01
1.42884958e+00 -1.81443602e-01 -5.31077206e-01 4.01601076e-01
6.68297231e-01 -5.01116812e-01 -6.05014801e-01 5.74147761e-01
7.41636693e-01 -9.28973496e-01 9.35187459e-01 -6.21428013e-01
4.62108515e-02 -7.71117568e-01 -9.72195193e-02 -4.66526181e-01
-4.97810185e-01 -2.06856892e-01 -4.11575973e-01 1.25842607e+00
-4.17548269e-02 -3.43335509e-01 9.49515641e-01 7.43623912e-01
-7.04211071e-02 -8.77925098e-01 -9.17458892e-01 -5.22436619e-01
-5.90762019e-01 -3.52776408e-01 3.11226845e-01 3.37541580e-01
-3.69143397e-01 4.05019611e-01 -4.88844275e-01 2.44070902e-01
6.08170390e-01 1.91910356e-01 9.73006368e-01 -9.99847114e-01
-5.02384841e-01 -2.65780717e-01 -6.05001152e-01 -1.33479238e+00
-1.42820597e-01 -1.45832330e-01 3.58561009e-01 -1.60875058e+00
4.02531743e-01 -2.38565028e-01 -5.11361539e-01 7.06714213e-01
-5.41997552e-01 4.05010104e-01 1.76306248e-01 1.36252955e-01
-1.41615474e+00 5.17289400e-01 1.32241297e+00 -7.71611184e-02
-3.05847913e-01 9.58317742e-02 -4.45139647e-01 1.07364500e+00
6.03800118e-01 -4.27837431e-01 -3.18924487e-01 -6.06835544e-01
-4.44608152e-01 2.76505262e-01 6.55518949e-01 -1.32555878e+00
3.41775477e-01 -1.40838161e-01 8.10103059e-01 -8.71829093e-01
4.68689471e-01 -7.14136302e-01 1.78247206e-02 6.62690163e-01
-7.89281055e-02 -1.21524185e-01 2.43172035e-01 8.94476771e-01
-1.55310437e-01 -7.73867443e-02 5.70161700e-01 -4.15050715e-01
-1.05531108e+00 5.23423374e-01 -1.80084363e-01 -1.50437489e-01
1.23578453e+00 -4.60407853e-01 -4.72886302e-03 2.80765002e-03
-7.90258884e-01 2.61370927e-01 1.62367821e-01 6.97563708e-01
3.97394329e-01 -1.26391172e+00 -5.78720033e-01 4.29821247e-03
9.29272100e-02 2.22587958e-01 6.33167565e-01 9.96130943e-01
-1.05859883e-01 3.75525177e-01 -1.76960185e-01 -7.79936314e-01
-1.55682361e+00 6.09249830e-01 3.36276919e-01 -2.04551831e-01
-7.32226491e-01 1.07822978e+00 5.84523678e-01 1.13365717e-01
5.21901906e-01 -4.34033990e-01 -3.14604819e-01 -5.14972173e-02
6.90164447e-01 2.94453353e-01 -2.70258814e-01 -6.75979078e-01
-6.10472798e-01 9.06468630e-01 -3.59746665e-01 6.43258989e-02
1.09119570e+00 -2.60949939e-01 1.59030497e-01 2.41879404e-01
9.25820291e-01 -3.58806908e-01 -1.70460737e+00 -3.97812366e-01
-1.60741851e-01 -5.19172370e-01 1.27025845e-03 -5.98723829e-01
-1.20379317e+00 7.80580521e-01 9.14503634e-01 3.13327797e-02
1.28976262e+00 4.54634950e-02 8.80048692e-01 5.25511920e-01
3.43788385e-01 -1.08526134e+00 2.81116754e-01 2.14929625e-01
5.83017826e-01 -1.38199675e+00 -6.20051892e-03 -2.97604561e-01
-4.92459923e-01 9.41130400e-01 9.70835447e-01 -1.53452858e-01
3.58048171e-01 1.43134281e-01 -4.16619219e-02 -1.55982122e-01
-5.18032968e-01 -2.47197315e-01 2.03560725e-01 2.88963556e-01
2.32740134e-01 -1.77542299e-01 -2.78073490e-01 6.37119293e-01
3.67222607e-01 2.50618607e-01 -7.43283052e-03 1.08087718e+00
-7.63187468e-01 -9.61005747e-01 -2.83022076e-01 4.44708705e-01
-5.98876655e-01 1.39090478e-01 -2.86734611e-01 5.38799107e-01
4.95574117e-01 1.00442433e+00 5.49704991e-02 -1.63245738e-01
3.46603394e-01 -1.47009060e-01 5.86331367e-01 -5.69905341e-01
-3.64787936e-01 3.51662397e-01 3.50319548e-04 -7.55241513e-01
-6.63821697e-01 -8.04808378e-01 -1.60675812e+00 6.81632832e-02
-6.04098558e-01 -2.70210683e-01 4.65140015e-01 1.32935572e+00
3.23743671e-01 8.16980958e-01 2.16286361e-01 -1.11274207e+00
-3.01174492e-01 -1.17093778e+00 -1.99540913e-01 3.22627842e-01
3.12843084e-01 -9.59282517e-01 6.44165426e-02 2.87055880e-01] | [8.091127395629883, 0.3083561360836029] |
57ba6eda-ba64-46a1-8519-a7df16b4c9ac | generalization-in-adaptive-data-analysis-and | 1506.02629 | null | http://arxiv.org/abs/1506.02629v2 | http://arxiv.org/pdf/1506.02629v2.pdf | Generalization in Adaptive Data Analysis and Holdout Reuse | Overfitting is the bane of data analysts, even when data are plentiful.
Formal approaches to understanding this problem focus on statistical inference
and generalization of individual analysis procedures. Yet the practice of data
analysis is an inherently interactive and adaptive process: new analyses and
hypotheses are proposed after seeing the results of previous ones, parameters
are tuned on the basis of obtained results, and datasets are shared and reused.
An investigation of this gap has recently been initiated by the authors in
(Dwork et al., 2014), where we focused on the problem of estimating
expectations of adaptively chosen functions.
In this paper, we give a simple and practical method for reusing a holdout
(or testing) set to validate the accuracy of hypotheses produced by a learning
algorithm operating on a training set. Reusing a holdout set adaptively
multiple times can easily lead to overfitting to the holdout set itself. We
give an algorithm that enables the validation of a large number of adaptively
chosen hypotheses, while provably avoiding overfitting. We illustrate the
advantages of our algorithm over the standard use of the holdout set via a
simple synthetic experiment.
We also formalize and address the general problem of data reuse in adaptive
data analysis. We show how the differential-privacy based approach given in
(Dwork et al., 2014) is applicable much more broadly to adaptive data analysis.
We then show that a simple approach based on description length can also be
used to give guarantees of statistical validity in adaptive settings. Finally,
we demonstrate that these incomparable approaches can be unified via the notion
of approximate max-information that we introduce. | ['Moritz Hardt', 'Cynthia Dwork', 'Vitaly Feldman', 'Toniann Pitassi', 'Aaron Roth', 'Omer Reingold'] | 2015-06-08 | generalization-in-adaptive-data-analysis-and-1 | http://papers.nips.cc/paper/5993-generalization-in-adaptive-data-analysis-and-holdout-reuse | http://papers.nips.cc/paper/5993-generalization-in-adaptive-data-analysis-and-holdout-reuse.pdf | neurips-2015-12 | ['holdout-set'] | ['computer-vision'] | [ 4.07011151e-01 6.78533539e-02 -1.24751389e-01 -3.78357023e-01
-7.57020354e-01 -8.65030944e-01 3.37133616e-01 4.47255552e-01
-6.88473105e-01 8.51466775e-01 -3.63828182e-01 -4.61262494e-01
-5.16845226e-01 -7.89238274e-01 -7.78551996e-01 -8.56976986e-01
-3.89452010e-01 3.12758476e-01 6.92865327e-02 8.44665840e-02
3.34415048e-01 5.85008979e-01 -1.68148351e+00 6.62504211e-02
6.15487099e-01 8.12757969e-01 -4.60457593e-01 8.08409870e-01
3.95133018e-01 1.56554222e-01 -4.39769119e-01 -5.08447647e-01
7.34258175e-01 -4.45170283e-01 -8.84656906e-01 2.98345864e-01
2.62049973e-01 -2.59961963e-01 1.50147453e-01 1.11291540e+00
5.31984925e-01 3.39483619e-01 4.08998549e-01 -1.68242657e+00
-1.30082652e-01 7.01996386e-01 -2.83810973e-01 -6.78026676e-02
2.77871907e-01 6.04384281e-02 1.04622531e+00 -6.54884517e-01
3.49964261e-01 8.70787740e-01 7.04523146e-01 3.87393385e-01
-1.73443186e+00 -7.73528695e-01 -4.79783304e-02 -1.07549012e-01
-1.57518744e+00 -5.24017394e-01 3.94667059e-01 -3.90701562e-01
3.26868057e-01 7.97529995e-01 5.51317692e-01 7.76019692e-01
-1.85428649e-01 7.13794708e-01 1.09340429e+00 -5.23409605e-01
6.30203605e-01 5.10223031e-01 3.42970341e-01 5.12458026e-01
5.67105353e-01 2.11425170e-01 -2.36292988e-01 -7.50111997e-01
4.94599670e-01 1.38377458e-01 -2.89053172e-01 -8.10876906e-01
-7.33042777e-01 1.10248435e+00 -8.11806917e-02 2.42187530e-01
-9.04881656e-02 -1.28386855e-01 4.95755792e-01 6.82094157e-01
3.57372493e-01 5.84193468e-01 -6.98007345e-01 3.43069918e-02
-8.42070997e-01 5.31832635e-01 1.03221071e+00 9.04960275e-01
6.39446318e-01 -4.75947708e-01 -8.72391686e-02 5.08738816e-01
-1.01511432e-02 1.72547087e-01 3.00708592e-01 -9.63218153e-01
1.16632842e-01 1.90384597e-01 2.87031114e-01 -7.60528207e-01
-3.24004561e-01 -1.88397869e-01 -7.01873899e-01 3.01130861e-01
7.53288865e-01 -3.38546187e-01 -3.05989206e-01 1.89400268e+00
4.26472604e-01 -1.15073755e-01 -4.96113673e-02 4.12758827e-01
-1.01768292e-01 2.03488380e-01 -9.39220190e-02 -6.79639459e-01
9.42589939e-01 -2.45118693e-01 -5.56799710e-01 4.60158497e-01
7.84582138e-01 -2.28781849e-01 1.07291687e+00 6.90991998e-01
-1.16497469e+00 -1.65960953e-01 -9.42512810e-01 1.48763865e-01
-3.41205925e-01 -4.32943851e-01 6.28364384e-01 1.14173794e+00
-9.82640326e-01 7.79789567e-01 -6.67541146e-01 -4.49224204e-01
6.37640238e-01 6.49606228e-01 -3.04068178e-01 3.18336695e-01
-1.02756417e+00 4.34954971e-01 4.41138655e-01 -1.63466245e-01
-5.68759561e-01 -8.02519441e-01 -5.48415720e-01 1.37795731e-01
7.33435214e-01 -5.34337521e-01 1.20069909e+00 -1.01899719e+00
-1.16616428e+00 8.22417438e-01 7.84149393e-03 -6.57858312e-01
9.90099490e-01 -5.39651550e-02 -1.63863808e-01 -2.28889778e-01
-1.98998749e-01 -9.95121226e-02 6.33997679e-01 -1.09011829e+00
-7.09085822e-01 -6.08181894e-01 2.97591332e-02 -3.95913303e-01
-4.24251497e-01 -2.99656913e-02 -1.12765886e-01 -5.78798354e-01
-1.72338590e-01 -9.76075351e-01 -5.65593660e-01 1.55269936e-01
-5.55725038e-01 5.52208647e-02 5.13346732e-01 -8.56434628e-02
1.55873895e+00 -2.24879527e+00 -1.93390593e-01 8.25375855e-01
2.21465692e-01 9.34271365e-02 1.96827903e-01 5.34880280e-01
-2.73733318e-01 5.52365780e-01 -7.28511512e-01 -1.66014075e-01
1.80946633e-01 6.81139603e-02 -3.47907752e-01 8.84683132e-01
-2.20396206e-01 4.84160990e-01 -6.62039757e-01 -2.21778959e-01
7.50089958e-02 1.68788880e-02 -8.98382127e-01 7.10229874e-02
3.88074555e-02 2.28123799e-01 -4.57521409e-01 3.33096355e-01
7.74441838e-01 -1.49239480e-01 3.28720093e-01 9.14350450e-02
-2.24133477e-01 -1.61171511e-01 -1.51624250e+00 1.13375127e+00
-1.99318796e-01 3.73454630e-01 -3.62612307e-02 -1.21088183e+00
6.20566070e-01 2.43525326e-01 4.81138855e-01 -2.72102654e-02
2.22961292e-01 3.26785892e-01 -5.85632101e-02 -3.96505862e-01
2.60011822e-01 -3.02884370e-01 -2.45054573e-01 7.08428323e-01
-3.07421178e-01 7.26716518e-02 -4.08384874e-02 -2.52183646e-01
1.06564760e+00 -2.59784549e-01 8.70649338e-01 -4.43862528e-01
6.24306083e-01 -2.87252784e-01 3.80553931e-01 1.23100340e+00
-1.51354104e-01 3.59981328e-01 6.46918714e-01 -3.32993507e-01
-1.13885295e+00 -1.01101792e+00 -6.46735609e-01 9.98918414e-01
-2.91864276e-01 -4.45907503e-01 -7.10853040e-01 -8.74066174e-01
2.98984617e-01 8.49012792e-01 -9.80731189e-01 -5.29074669e-02
-1.28145978e-01 -9.15645361e-01 4.66710985e-01 1.92292795e-01
3.16341162e-01 -5.47209024e-01 -8.12727630e-01 3.79468612e-02
4.10999566e-01 -6.09764457e-01 -2.97766268e-01 3.70096147e-01
-9.25640166e-01 -1.07973051e+00 -2.82077134e-01 -2.10823759e-01
8.05427730e-01 -3.38003710e-02 7.29389787e-01 2.47502267e-01
-1.53473437e-01 5.60371041e-01 -2.06483811e-01 -6.29738390e-01
-6.21069610e-01 -6.71136305e-02 1.79473564e-01 2.12525502e-01
4.03955638e-01 -7.45057762e-01 -3.48254889e-01 3.30796033e-01
-1.27935588e+00 -4.09242123e-01 1.95073158e-01 7.83358872e-01
5.85758984e-01 2.51779556e-01 5.06557107e-01 -1.53363562e+00
5.87716520e-01 -7.26642013e-01 -1.10453367e+00 3.51761103e-01
-9.82322156e-01 3.25356036e-01 7.18533814e-01 -4.50193435e-01
-5.37447512e-01 2.00071260e-01 1.07620113e-01 -3.79897326e-01
2.79199816e-02 2.82878608e-01 -3.83074194e-01 -2.02877983e-01
5.89594185e-01 3.18793580e-02 3.23150873e-01 -5.09198666e-01
2.96362072e-01 7.22955585e-01 2.62762338e-01 -5.90930760e-01
6.82391286e-01 4.01542276e-01 2.99195796e-01 -5.67321956e-01
-5.24980366e-01 -1.87564120e-01 -6.56181037e-01 1.12346694e-01
2.78702527e-01 -3.44709545e-01 -1.11372674e+00 2.16362283e-01
-6.04496896e-01 -3.27494532e-01 -7.58184016e-01 3.47142160e-01
-7.20674217e-01 4.03179497e-01 1.55025944e-01 -1.16648972e+00
3.90502326e-02 -1.03153419e+00 6.28328562e-01 -2.42668957e-01
-4.07624513e-01 -1.19482553e+00 3.81736793e-02 -9.18910578e-02
1.58680514e-01 4.10562009e-01 7.95096338e-01 -1.46476460e+00
-3.50575387e-01 -5.06024837e-01 2.30955660e-01 1.85974509e-01
2.28458926e-01 1.09028660e-01 -1.23384511e+00 -5.81574023e-01
3.18621606e-01 -3.18673737e-02 6.00252748e-01 3.48093241e-01
1.85126066e+00 -8.38207781e-01 -2.68253148e-01 5.46518028e-01
1.52899301e+00 4.25674953e-02 3.99048328e-01 3.64730000e-01
1.38452977e-01 7.30887830e-01 5.59808314e-01 9.69658017e-01
-2.61675869e-03 4.99817640e-01 1.92647681e-01 1.94677532e-01
8.88557613e-01 -7.24647865e-02 2.35836312e-01 3.84571403e-02
6.73545003e-02 -3.19838107e-01 -5.42210817e-01 4.22557414e-01
-1.81012607e+00 -1.07009280e+00 -8.32212642e-02 3.20388007e+00
1.00412691e+00 5.08371629e-02 5.82209349e-01 4.05099928e-01
6.58370554e-01 -3.95048767e-01 -6.05074406e-01 -6.16369724e-01
1.01472186e-02 2.76956260e-01 1.01622212e+00 5.95025659e-01
-1.00128865e+00 2.77632713e-01 6.98183107e+00 8.16053033e-01
-6.76720738e-01 -1.26602083e-01 8.71428668e-01 -8.96112323e-02
-6.02273643e-01 2.46465266e-01 -4.87708956e-01 4.18814778e-01
1.20183897e+00 -7.90490627e-01 4.78632808e-01 8.30648541e-01
1.29188448e-01 -2.49434009e-01 -1.61128712e+00 7.06618249e-01
-1.81116492e-01 -1.03977156e+00 -2.14233592e-01 4.55183625e-01
5.61114907e-01 -5.28428674e-01 2.57602185e-01 1.05404653e-01
5.26122570e-01 -1.09735084e+00 3.57387602e-01 4.65638280e-01
4.70708251e-01 -8.86969686e-01 5.53341508e-01 4.62946653e-01
-6.72481775e-01 -3.85116577e-01 -1.81442425e-01 -1.00448713e-01
-3.90349388e-01 6.43979490e-01 -7.44855702e-01 6.37012064e-01
5.38482785e-01 3.70354593e-01 -4.41301227e-01 1.15163195e+00
2.64689416e-01 6.48496330e-01 -6.49804771e-01 -6.85659423e-02
-1.67632297e-01 -2.43022725e-01 6.39828980e-01 9.95378196e-01
3.20967466e-01 -4.25329618e-03 -2.08070874e-02 7.00287819e-01
5.86022027e-02 3.60874981e-01 -7.70341456e-01 1.83904678e-01
5.77305436e-01 8.57304335e-01 -4.63412017e-01 -1.79048121e-01
-4.24578160e-01 6.55475080e-01 1.05483912e-01 1.52843654e-01
-5.35542488e-01 -1.83616444e-01 5.27443290e-01 3.49998698e-02
3.07085335e-01 9.85908806e-02 -3.42959464e-01 -8.61534595e-01
-9.88524780e-03 -1.06753910e+00 1.10660088e+00 -1.40618637e-01
-1.46635962e+00 1.02188572e-01 4.59743172e-01 -1.29288590e+00
-3.37314546e-01 -4.20718819e-01 -2.54416496e-01 7.49395847e-01
-8.27877283e-01 -4.80474591e-01 9.83727798e-02 5.27321637e-01
1.92226201e-01 6.64673820e-02 9.11900282e-01 1.47331674e-02
-5.39493680e-01 1.08498693e+00 4.18268830e-01 -1.13700256e-01
4.53665465e-01 -1.26235402e+00 1.02580599e-01 8.65072727e-01
5.04449904e-02 8.41256201e-01 1.04627919e+00 -3.95109117e-01
-1.33042622e+00 -9.88622665e-01 6.29377365e-01 -5.55920422e-01
7.55570948e-01 -4.62537974e-01 -1.03534186e+00 9.01545942e-01
-1.83151171e-01 8.58841389e-02 9.62917387e-01 2.45842919e-01
-1.28826737e-01 -2.21290022e-01 -1.63843358e+00 6.77909970e-01
8.01413357e-01 -2.83127457e-01 -1.69875741e-01 1.85051724e-01
3.99553418e-01 -1.28165647e-01 -1.01061690e+00 3.46795917e-01
7.41294503e-01 -1.10809493e+00 6.57104075e-01 -9.59954083e-01
-2.10431740e-01 -8.72864351e-02 -3.11465263e-01 -8.75547230e-01
6.66225478e-02 -1.17895472e+00 3.94105986e-02 1.24543738e+00
3.90330315e-01 -1.03391957e+00 5.40571213e-01 1.18964946e+00
5.78679860e-01 -6.22438312e-01 -9.67673898e-01 -8.58658016e-01
2.51684964e-01 -7.15192795e-01 7.96373129e-01 1.01061797e+00
1.50811136e-01 -2.15285972e-01 -2.97354043e-01 2.87475169e-01
6.72590733e-01 2.42594369e-02 1.05378854e+00 -1.32517576e+00
-4.37545687e-01 -4.13092822e-01 -5.43266177e-01 -6.91264987e-01
-4.17516865e-02 -9.43159103e-01 -1.32320940e-01 -5.04346371e-01
2.48476163e-01 -6.40253961e-01 -3.74917060e-01 4.41894382e-01
-1.41247451e-01 -5.64658381e-02 3.32069173e-02 1.27320677e-01
-2.52508432e-01 1.42257825e-01 8.45943689e-01 3.89901668e-01
-4.07890230e-01 6.56278908e-01 -7.96073079e-01 5.74805379e-01
8.17155838e-01 -5.31958103e-01 -5.01959801e-01 4.21601474e-01
3.31384480e-01 -3.20977531e-02 6.32884204e-01 -6.65821850e-01
1.83793291e-01 -2.23126993e-01 1.61477387e-01 -1.87238201e-01
-1.59877419e-01 -1.23246610e+00 4.71742690e-01 5.10196567e-01
-8.54572654e-01 -1.96356270e-02 2.23606974e-01 7.00619340e-01
3.16566467e-01 -5.09525180e-01 7.46880352e-01 8.29705968e-02
-8.26387554e-02 4.57701355e-01 -3.88918191e-01 1.53009623e-01
1.25643528e+00 -2.35138103e-01 1.61020398e-01 -4.31564391e-01
-8.89625072e-01 2.92996973e-01 6.64110363e-01 -1.35752469e-01
2.85447031e-01 -9.80445027e-01 -6.60454512e-01 4.77556020e-01
1.10732727e-01 -2.60034949e-01 4.33021076e-02 1.02153695e+00
1.45980017e-03 -1.96852796e-02 1.53253645e-01 -3.79001707e-01
-1.37815785e+00 8.87980998e-01 3.64727676e-01 -3.84581015e-02
-5.10935009e-01 5.85685313e-01 1.87361702e-01 -1.72911108e-01
2.04945624e-01 -3.62216562e-01 2.76059151e-01 7.67147243e-02
5.88055551e-01 4.90581512e-01 -4.45754565e-02 -1.07021004e-01
-2.74424285e-01 1.27308786e-01 -1.52839825e-01 -2.31194183e-01
1.41773236e+00 -3.80426317e-01 -4.22818996e-02 9.09675300e-01
1.34859204e+00 4.74330395e-01 -1.08401287e+00 -2.59422123e-01
8.35640207e-02 -6.43122077e-01 -2.59418279e-01 -4.69797194e-01
-8.15630674e-01 5.18280923e-01 6.86549246e-01 7.82479465e-01
1.42547023e+00 -1.68710828e-01 1.10644326e-01 5.03700316e-01
3.52275699e-01 -9.76390123e-01 -6.65794849e-01 -1.11714378e-01
6.80592358e-01 -1.09975386e+00 3.25939745e-01 -2.83966988e-01
-4.81761426e-01 1.22196770e+00 3.59474421e-02 5.77003174e-02
8.64284098e-01 2.60721236e-01 -4.58181679e-01 -4.27033119e-02
-7.17067122e-01 -1.76907461e-02 5.56446165e-02 4.37989354e-01
7.65089020e-02 -8.49544704e-02 -5.28951466e-01 7.73007035e-01
-2.00095758e-01 1.79426178e-01 6.80632770e-01 1.08526444e+00
-2.11330369e-01 -1.33019805e+00 -4.74544376e-01 5.76674104e-01
-6.67103589e-01 -8.79473761e-02 -4.21181381e-01 1.12479246e+00
-6.95109516e-02 7.84120142e-01 5.66404089e-02 -2.36868352e-01
1.96198285e-01 2.20133469e-01 3.88119131e-01 -3.66807431e-01
-5.40699780e-01 -1.53724760e-01 -2.49263719e-02 -4.40488219e-01
-5.13435006e-01 -1.18659186e+00 -9.26901340e-01 -5.40406466e-01
-3.68482322e-01 4.10303175e-01 3.19257736e-01 8.96494091e-01
1.60875350e-01 -2.71789283e-02 1.17913914e+00 -7.65225440e-02
-9.18779910e-01 -5.06946683e-01 -9.64210391e-01 3.12544256e-01
4.93417233e-01 -3.23451579e-01 -7.93459117e-01 4.93059643e-02] | [7.552873134613037, 4.570426940917969] |
059204a7-9b5e-480d-b296-68bfafb4cbf6 | monitoring-and-improving-personalized-sleep | 2211.12778 | null | https://arxiv.org/abs/2211.12778v1 | https://arxiv.org/pdf/2211.12778v1.pdf | Monitoring and Improving Personalized Sleep Quality from Long-Term Lifelogs | Sleep plays a vital role in our physical, cognitive, and psychological well-being. Despite its importance, long-term monitoring of personalized sleep quality (SQ) in real-world contexts is still challenging. Many sleep researches are still developing clinically and far from accessible to the general public. Fortunately, wearables and IoT devices provide the potential to explore the sleep insights from multimodal data, and have been used in some SQ researches. However, most of these studies analyze the sleep related data and present the results in a delayed manner (i.e., today's SQ obtained from last night's data), it is sill difficult for individuals to know how their sleep will be before they go to bed and how they can proactively improve it. To this end, this paper proposes a computational framework to monitor the individual SQ based on both the objective and subjective data from multiple sources, and moves a step further towards providing the personalized feedback to improve the SQ in a data-driven manner. The feedback is implemented by referring the insights from the PMData dataset based on the discovered patterns between life events and different levels of SQ. The deep learning based personal SQ model (PerSQ), using the long-term heterogeneous data and considering the carry-over effect, achieves higher prediction performance compared with baseline models. A case study also shows reasonable results for an individual to monitor and improve the SQ in the future. | ['Koji Zettsu', 'Minh-Son Dao', 'Wenbin Gan'] | 2022-11-23 | null | null | null | null | ['sleep-quality-prediction'] | ['medical'] | [-2.06867427e-01 -1.87126338e-01 -4.68747735e-01 -5.67839026e-01
-3.23918313e-01 2.27401257e-01 -9.75317359e-02 3.60681623e-01
-3.84998202e-01 1.03996980e+00 6.04037642e-01 1.37883931e-01
-4.51989144e-01 -9.28153574e-01 2.26827506e-02 -7.27726340e-01
3.52644362e-02 1.03564754e-01 4.37342599e-02 -3.19101483e-01
-5.81616089e-02 -7.32388571e-02 -1.64210594e+00 -1.03050068e-01
1.28671622e+00 9.76853609e-01 3.01272899e-01 1.78963810e-01
1.21789344e-01 2.92458653e-01 -6.77959502e-01 -2.53748447e-01
-1.78471878e-01 -4.83408868e-01 -5.15924156e-01 5.96167240e-03
-1.70325458e-01 -1.76545687e-03 -1.64235216e-02 7.60366559e-01
8.86195421e-01 4.14080083e-01 -2.38252804e-01 -1.23269653e+00
-8.64360273e-01 1.52685627e-01 9.84698981e-02 6.27993762e-01
5.44763923e-01 3.26668799e-01 8.70195389e-01 -2.51843989e-01
-1.90415487e-01 8.13082516e-01 5.78841746e-01 6.24060869e-01
-1.13816679e+00 -4.70418960e-01 7.10182711e-02 7.49654353e-01
-1.22160721e+00 -5.77137887e-01 7.40892768e-01 2.67272890e-02
8.89058709e-01 5.06707251e-01 1.33267343e+00 1.22854912e+00
5.65008581e-01 3.83828878e-01 1.19850612e+00 -4.28718999e-02
3.20743382e-01 1.53129503e-01 3.07296276e-01 4.39884275e-01
2.14496106e-01 -2.20974833e-02 -9.62441862e-01 6.46216720e-02
-2.41049323e-02 6.44862235e-01 -9.85233709e-02 2.52613753e-01
-9.12245870e-01 4.95526105e-01 5.03098905e-01 5.82208574e-01
-5.12044430e-01 -3.28809828e-01 8.51276442e-02 3.56273949e-02
8.60127866e-01 2.49276653e-01 -5.32919705e-01 -4.46509987e-01
-1.22087824e+00 -5.24002761e-02 4.65998471e-01 3.59666467e-01
8.55432510e-01 -1.83866486e-01 -5.28019488e-01 8.77790451e-01
4.20909673e-01 6.54789329e-01 9.23533916e-01 -8.52249026e-01
1.76742390e-01 1.06261933e+00 1.09293893e-01 -1.21264458e+00
-1.01274967e+00 -8.84064615e-01 -1.02099633e+00 -3.98330510e-01
-1.37211457e-01 -1.50398791e-01 -2.33685583e-01 1.68443775e+00
1.85576066e-01 3.01013529e-01 -1.69182941e-01 7.02598691e-01
1.19279134e+00 3.25342864e-01 1.39747024e-01 -6.47837460e-01
1.48604774e+00 -7.89600015e-01 -1.13912642e+00 -6.62820518e-01
2.20686033e-01 -1.80551559e-01 1.39812768e+00 5.60246289e-01
-1.14801276e+00 -8.83035243e-01 -1.08621967e+00 -1.74941272e-02
-4.09795254e-01 9.98462886e-02 4.70370173e-01 8.48067164e-01
-1.39827371e+00 6.88331783e-01 -1.27527332e+00 -8.44166636e-01
4.94597852e-01 6.47768021e-01 -7.72913173e-03 -3.33439521e-02
-1.30849874e+00 7.17542291e-01 1.15512080e-01 1.77643657e-01
-5.18307924e-01 -5.07735908e-01 -5.44089615e-01 2.26172134e-01
3.40164661e-01 -1.24140847e+00 8.71708989e-01 -5.51357985e-01
-1.46295607e+00 4.47327554e-01 -5.78632712e-01 -2.92778432e-01
-1.34417817e-01 -2.75158674e-01 -9.86793816e-01 -1.86907589e-01
3.46921355e-01 2.54845351e-01 4.00223374e-01 -6.36128247e-01
-6.48039997e-01 -8.76647174e-01 3.80049273e-02 1.67190060e-01
-8.46635044e-01 -1.98875189e-01 -2.91008770e-01 -2.77934875e-02
-2.24405766e-01 -7.12498724e-01 -2.36217827e-01 -2.90000975e-01
-4.00998026e-01 -4.30059344e-01 3.77844572e-01 -5.35408318e-01
1.80456841e+00 -2.02323413e+00 -1.63888127e-01 -2.98593313e-01
3.39689016e-01 3.45118344e-01 2.46175498e-01 4.62621719e-01
5.27739227e-01 1.17014118e-01 3.09892204e-02 -1.00180852e+00
-6.27498180e-02 6.46222651e-01 5.33046246e-01 3.02143544e-01
-1.88433155e-01 1.05281472e+00 -9.65557575e-01 -4.36421901e-01
2.83535212e-01 3.61644953e-01 -4.54021990e-01 3.83512467e-01
1.59047440e-01 7.10748494e-01 -4.09643799e-01 7.01654136e-01
3.74473095e-01 -5.54912090e-01 -7.35866949e-02 -5.25886565e-02
-2.89468527e-01 4.56218332e-01 -6.34824514e-01 1.74822915e+00
-5.23184597e-01 3.01554501e-01 -1.41045228e-01 -9.15304482e-01
9.85687256e-01 6.99427947e-02 5.81836402e-01 -1.24507999e+00
-2.20872648e-02 -2.53357291e-01 -9.19409320e-02 -8.92069340e-01
4.31029856e-01 -4.06053811e-01 6.55081272e-02 3.29232216e-01
-3.32484454e-01 7.73694873e-01 2.36106277e-01 -1.39249071e-01
1.17245567e+00 -2.29204044e-01 5.91385961e-01 -7.92141184e-02
6.69148028e-01 -3.98825407e-01 9.66324627e-01 4.61274415e-01
-6.04280591e-01 2.55707383e-01 -1.06286965e-02 -3.26005697e-01
-2.15352356e-01 -1.07252002e+00 -1.41377106e-01 1.11163890e+00
1.94133624e-01 -8.37592781e-01 -3.49295020e-01 -4.09964055e-01
-1.84461221e-01 5.67775488e-01 -6.36017561e-01 -7.01007783e-01
9.90850851e-02 -1.15578175e+00 9.98360589e-02 4.23979104e-01
9.74385142e-01 -1.08396757e+00 -5.57263017e-01 3.41679543e-01
-5.74693084e-01 -7.37594903e-01 -3.09432834e-01 -1.49870396e-01
-9.92636800e-01 -8.34903479e-01 -1.96823955e-01 -1.12767257e-01
2.59980142e-01 4.10597622e-01 1.26158977e+00 2.90386736e-01
3.22854482e-02 4.13905084e-01 -3.89156133e-01 -4.98949826e-01
2.16436535e-01 3.16476762e-01 4.24320608e-01 2.37967536e-01
6.71024501e-01 -1.08872926e+00 -1.35415208e+00 3.77259493e-01
-5.52769601e-01 -2.52903193e-01 6.02239311e-01 4.28633749e-01
6.65614903e-01 3.81611824e-01 9.33641553e-01 -4.01561886e-01
7.54986465e-01 -9.09208894e-01 1.17085926e-01 6.43846169e-02
-1.41293013e+00 -2.91917652e-01 5.53478479e-01 -1.04241781e-02
-8.16723049e-01 -4.52886224e-01 -5.26279926e-01 2.62296826e-01
-2.37735137e-01 3.99506420e-01 -3.97734672e-01 5.22988677e-01
4.29132044e-01 4.81620461e-01 -6.53227568e-02 -6.53373301e-01
-1.09880604e-01 7.83314764e-01 3.15502256e-01 -1.23582035e-01
4.33453172e-01 5.64798594e-01 -2.85352990e-02 -7.65379786e-01
-1.38725019e+00 -6.32318079e-01 -4.64445412e-01 -2.01669604e-01
1.01290834e+00 -8.78397942e-01 -1.11742544e+00 1.45276174e-01
-4.58935857e-01 -3.02732110e-01 -2.69487113e-01 3.20006311e-01
-2.29390889e-01 2.61719614e-01 -2.82784514e-02 -8.37206960e-01
-8.12632263e-01 -8.42361212e-01 8.75644922e-01 7.39778340e-01
-3.16850930e-01 -1.29384744e+00 4.12360132e-01 7.38305271e-01
6.42241478e-01 -2.47170590e-02 5.53092182e-01 -4.09085959e-01
-3.71588707e-01 4.94204164e-02 5.53612374e-02 4.95235920e-01
7.59582281e-01 -5.44903159e-01 -8.96260619e-01 -3.60699803e-01
2.17970371e-01 7.10556209e-02 4.50016707e-01 5.23426354e-01
1.04892194e+00 -3.40802103e-01 -2.39846975e-01 5.52592933e-01
1.14849114e+00 1.82533592e-01 7.49456823e-01 5.23266673e-01
4.55025256e-01 2.31423452e-01 5.51828444e-01 6.82295263e-01
1.22502673e+00 6.09055400e-01 6.08539283e-01 -6.65006265e-02
-3.04113068e-02 -5.05376831e-02 5.90487421e-01 8.96315455e-01
-1.89895764e-01 -2.93963760e-01 -5.11165082e-01 4.15640473e-01
-1.84772432e+00 -1.01258695e+00 -2.56646097e-01 2.28488946e+00
5.21143138e-01 2.15950832e-01 4.59960282e-01 2.69728512e-01
1.47678241e-01 2.63095438e-01 -7.29904950e-01 -3.28269273e-01
-3.68509553e-02 2.84589708e-01 -6.67210529e-03 -2.27231402e-02
-6.95693314e-01 3.21890146e-01 6.01135254e+00 3.51951987e-01
-1.00374806e+00 6.54855847e-01 4.87448692e-01 -5.56668997e-01
-1.11766599e-01 -1.53702170e-01 -9.22536194e-01 1.06363642e+00
1.61182535e+00 -1.34758070e-01 6.85406148e-01 4.82644737e-01
1.16209424e+00 -4.80459839e-01 -8.19393635e-01 1.17264962e+00
1.06967606e-01 -9.14647043e-01 -6.72906637e-01 2.98161209e-01
3.70801210e-01 7.12605715e-02 -3.42027694e-02 4.74065632e-01
-2.04767227e-01 -8.29925835e-01 9.94943008e-02 1.30868495e+00
4.05294269e-01 -5.80242634e-01 8.51051986e-01 5.88869095e-01
-1.10767949e+00 -4.75101352e-01 -4.09738690e-01 -5.52796304e-01
1.97573215e-01 1.04083526e+00 -6.97508872e-01 7.21496463e-01
1.01271224e+00 1.23694873e+00 -1.13932550e+00 1.07767546e+00
-2.24787846e-01 7.70258069e-01 -1.46913612e-02 -1.48000509e-01
-4.26947355e-01 -2.97540575e-01 8.64897594e-02 6.41185164e-01
6.14686608e-01 2.24262737e-02 1.41499028e-01 8.89706671e-01
7.02173859e-02 5.98009080e-02 -3.04469377e-01 -5.81467804e-03
2.37295926e-01 1.50128496e+00 -3.36891413e-01 -9.62651893e-02
-6.19243741e-01 8.08125257e-01 1.38998866e-01 1.82565637e-02
-4.69279379e-01 1.84918478e-01 8.27509999e-01 4.33638722e-01
-1.87712744e-01 -1.42250761e-01 -4.36839461e-01 -9.94436562e-01
5.95096350e-02 -4.82429981e-01 5.52317858e-01 -7.18148232e-01
-1.50610816e+00 2.52716660e-01 -2.87107587e-01 -1.10538375e+00
1.86078414e-01 5.65768369e-02 -9.51581419e-01 6.90521479e-01
-1.51372218e+00 -8.87778580e-01 -6.59809530e-01 6.55252695e-01
4.27274019e-01 1.82464033e-01 8.08482826e-01 6.70164883e-01
-1.12671721e+00 6.01546943e-01 -1.48503864e-02 -5.46239316e-01
7.43681788e-01 -1.22724891e+00 -1.28035948e-01 6.26913607e-01
-1.55834809e-01 7.85095155e-01 4.74872589e-01 -5.99130452e-01
-1.33744073e+00 -1.09900439e+00 1.11401498e+00 -6.53286159e-01
2.89087981e-01 -1.24945268e-01 -6.93764508e-01 3.82482946e-01
2.49972701e-01 -2.21664473e-01 1.27668178e+00 7.10178673e-01
7.42695451e-01 -1.02867866e+00 -1.15964949e+00 2.92485237e-01
1.19493425e+00 -2.87376255e-01 -4.17240769e-01 2.26275653e-01
6.16573334e-01 2.14733496e-01 -1.06000447e+00 1.80210143e-01
4.40890878e-01 -1.52187085e+00 9.40512836e-01 -2.44761616e-01
5.41950762e-02 -2.81056672e-01 -1.70760043e-02 -1.44798040e+00
-4.20151979e-01 -4.12605435e-01 -2.32632712e-01 1.48691034e+00
-7.97663927e-02 -6.83640599e-01 7.24136889e-01 8.58154178e-01
-4.22844142e-01 -9.38709855e-01 -9.96540368e-01 -5.48204780e-01
-5.27403235e-01 -4.90559429e-01 1.03065217e+00 3.70745629e-01
2.42967010e-02 5.50165653e-01 -6.18873775e-01 2.50934154e-01
3.22414488e-01 5.20785116e-02 6.69601738e-01 -1.46196711e+00
-7.50156268e-02 9.54897492e-04 -3.92136157e-01 -7.77981281e-01
-1.06975757e-01 -8.03129792e-01 -2.75321156e-01 -2.12628865e+00
4.96821880e-01 -1.67903766e-01 -8.93964469e-01 4.78847980e-01
-4.59549010e-01 1.50434017e-01 -1.09474011e-01 -2.34976374e-02
-1.14242315e+00 9.35533762e-01 1.26030660e+00 -9.08159837e-02
-5.88468492e-01 5.68356335e-01 -1.25475144e+00 5.17306149e-01
1.09597671e+00 -3.64627570e-01 -6.91833973e-01 -1.14314094e-01
4.38510358e-01 2.24065378e-01 3.67868483e-01 -1.41162813e+00
2.96930373e-01 -1.81355700e-01 4.40978199e-01 -6.74187064e-01
6.98621809e-01 -6.99596584e-01 2.16850251e-01 4.30574805e-01
8.28043818e-02 2.19919175e-01 -1.54290646e-01 7.21214533e-01
1.20375134e-01 1.77564695e-01 5.08351028e-01 -2.23322794e-01
-4.81720924e-01 5.95214307e-01 -2.11075604e-01 -5.36448089e-03
6.12382352e-01 -1.83077574e-01 -4.26886708e-01 -4.74145979e-01
-1.20358384e+00 4.64158922e-01 2.08275780e-01 2.81663150e-01
6.52656555e-01 -1.37272000e+00 -2.04371467e-01 3.26961368e-01
2.06747189e-01 -3.24080884e-01 7.00196505e-01 1.40075302e+00
3.31862867e-01 4.12542701e-01 -2.37110808e-01 -5.62788010e-01
-1.15375388e+00 5.93788385e-01 3.02002132e-01 -3.61466616e-01
-3.91488045e-01 3.94841999e-01 -2.69312620e-01 -3.51734161e-02
1.01939194e-01 -3.81201565e-01 -6.26030803e-01 1.18074685e-01
7.59662509e-01 7.64550626e-01 2.76688308e-01 -4.29206342e-01
-5.06044567e-01 2.71384776e-01 5.06521702e-01 4.11145449e-01
1.53067386e+00 -9.42292333e-01 -2.12457255e-01 9.27300453e-01
9.11730707e-01 6.17147349e-02 -9.09458518e-01 -2.03986883e-01
-3.82999033e-02 -5.22525072e-01 -4.40615602e-03 -8.37341428e-01
-1.15002418e+00 8.75747442e-01 1.24769855e+00 5.53262115e-01
1.47446477e+00 1.04375564e-01 1.22004759e+00 2.26084396e-01
3.99424851e-01 -1.05448997e+00 2.40932524e-01 -9.77788419e-02
3.77744317e-01 -1.44273829e+00 2.72584587e-01 3.19588274e-01
-4.97467250e-01 6.98346078e-01 5.12746215e-01 2.28637591e-01
7.93886423e-01 -4.14552540e-01 -7.13649988e-02 -2.68743366e-01
-8.31045568e-01 -5.17468870e-01 3.09213519e-01 8.27581704e-01
3.22666824e-01 3.78393620e-01 -5.27720392e-01 1.17085457e+00
-4.03111935e-01 3.18175256e-01 3.04546237e-01 5.75116277e-01
-6.59674168e-01 -1.57556820e+00 -2.12122872e-01 9.12460566e-01
-5.34671068e-01 -9.73401126e-03 -8.83920714e-02 2.69915313e-01
7.75742531e-01 1.78596330e+00 -2.62768656e-01 -6.26821041e-01
5.54200292e-01 2.27608338e-01 -9.30281132e-02 -8.65713716e-01
-3.97232324e-01 -2.44999066e-01 1.49442032e-01 -8.38127077e-01
-9.46454108e-01 -8.05979252e-01 -1.10813773e+00 -4.42597717e-01
2.52728462e-01 1.07221901e-01 3.79339635e-01 1.27915609e+00
6.84469879e-01 8.78320456e-01 7.59587348e-01 -5.69126010e-01
2.30699498e-02 -9.77052271e-01 -7.34804988e-01 1.70997381e-01
4.39166546e-01 -7.91336060e-01 -1.28812686e-01 -3.56452525e-01] | [13.585501670837402, 3.4223227500915527] |
f5034b72-5ef7-41ee-a24c-b933d25f5011 | revisiting-lstm-networks-for-semi-supervised-1 | 2009.04007 | null | https://arxiv.org/abs/2009.04007v1 | https://arxiv.org/pdf/2009.04007v1.pdf | Revisiting LSTM Networks for Semi-Supervised Text Classification via Mixed Objective Function | In this paper, we study bidirectional LSTM network for the task of text classification using both supervised and semi-supervised approaches. Several prior works have suggested that either complex pretraining schemes using unsupervised methods such as language modeling (Dai and Le 2015; Miyato, Dai, and Goodfellow 2016) or complicated models (Johnson and Zhang 2017) are necessary to achieve a high classification accuracy. However, we develop a training strategy that allows even a simple BiLSTM model, when trained with cross-entropy loss, to achieve competitive results compared with more complex approaches. Furthermore, in addition to cross-entropy loss, by using a combination of entropy minimization, adversarial, and virtual adversarial losses for both labeled and unlabeled data, we report state-of-the-art results for text classification task on several benchmark datasets. In particular, on the ACL-IMDB sentiment analysis and AG-News topic classification datasets, our method outperforms current approaches by a substantial margin. We also show the generality of the mixed objective function by improving the performance on relation extraction task. | ['Manzil Zaheer', 'Ruslan Salakhutdinov', 'Devendra Singh Sachan'] | 2020-09-08 | revisiting-lstm-networks-for-semi-supervised | https://www.semanticscholar.org/paper/Revisiting-LSTM-Networks-for-Semi-Supervised-Text-Sachan-Petuum/c3f89364aecd661eb032840d2fe3efd0f6d1698c | https://www.aaai.org/Papers/AAAI/2019/AAAI-SachanD.7236.pdf | aaai-2019-2019-2 | ['semi-supervised-text-classification-1'] | ['natural-language-processing'] | [ 1.19204640e-01 3.65337193e-01 -3.78085554e-01 -6.32250905e-01
-9.14822936e-01 -3.65321606e-01 7.81426847e-01 3.22247654e-01
-6.41238034e-01 8.77319455e-01 1.07258342e-01 -4.64982599e-01
2.39616230e-01 -7.00351417e-01 -6.42521799e-01 -4.28986460e-01
2.11188197e-01 4.02926087e-01 -1.36859016e-02 -2.44121522e-01
-2.91437488e-02 7.24400431e-02 -1.12707460e+00 4.10395354e-01
8.77048075e-01 1.36389697e+00 -3.88111353e-01 4.56970572e-01
-2.12628737e-01 1.34757924e+00 -4.04505372e-01 -9.82554018e-01
-2.62754038e-02 -3.28124493e-01 -1.25548232e+00 -1.09818622e-01
1.72646999e-01 -2.84884304e-01 -2.78377682e-01 8.24628592e-01
3.48817796e-01 1.95479676e-01 8.54362965e-01 -1.11447608e+00
-5.13078928e-01 1.11443377e+00 -3.30608040e-01 -1.82988852e-01
1.13704680e-02 -4.00530457e-01 1.31783795e+00 -8.67687464e-01
4.81447160e-01 9.86435652e-01 7.99880266e-01 3.89808148e-01
-1.30502009e+00 -6.36715353e-01 2.91259080e-01 3.22936237e-01
-9.74634469e-01 -4.66875106e-01 8.80655885e-01 -3.88936549e-01
1.28074014e+00 2.19279993e-02 1.93572149e-01 1.36360908e+00
2.45166421e-01 1.04200661e+00 9.88580525e-01 -6.88631475e-01
-1.91773120e-02 4.21216458e-01 4.57453012e-01 5.81063390e-01
-4.01176922e-02 -1.58127919e-01 -4.60373461e-01 -1.24097869e-01
-7.77758956e-02 -2.29318276e-01 -2.26484925e-01 -3.38400453e-01
-9.86657977e-01 1.01977491e+00 3.85791719e-01 5.03509820e-01
-7.01492131e-02 -4.46137674e-02 9.62527394e-01 3.80555779e-01
1.18908525e+00 3.57840508e-01 -7.60371983e-01 -2.75235306e-02
-1.16214371e+00 -2.64141187e-02 1.09473038e+00 8.16825747e-01
5.12188077e-01 6.77927583e-02 -2.43059173e-01 1.00949323e+00
2.02486798e-01 5.85540384e-02 6.82504654e-01 -4.94101495e-01
7.97583342e-01 4.83680338e-01 -3.48729938e-01 -8.56431723e-01
-4.59859192e-01 -5.17803848e-01 -1.24002934e+00 -8.74918476e-02
3.62668037e-01 -3.85713696e-01 -7.70044684e-01 1.81817257e+00
-6.87951595e-02 -1.59763977e-01 3.42989266e-01 4.48390275e-01
9.41549122e-01 6.35839880e-01 1.82110265e-01 -2.76856065e-01
1.11717606e+00 -1.31334734e+00 -9.98656571e-01 -1.85895681e-01
1.04764283e+00 -5.94090700e-01 9.64041889e-01 4.25328046e-01
-1.06558239e+00 -3.59507561e-01 -1.17242479e+00 -4.15634841e-01
-6.94618225e-01 3.90955091e-01 5.76947272e-01 5.72442532e-01
-8.38331223e-01 6.71260238e-01 -9.04944420e-01 -4.97234523e-01
3.65233243e-01 2.74884522e-01 -2.72455394e-01 2.85540789e-01
-1.57729292e+00 1.18542349e+00 4.90949720e-01 1.74006775e-01
-4.65195209e-01 -4.49595124e-01 -8.90581548e-01 1.58395097e-01
2.87882358e-01 -4.67663586e-01 1.25622559e+00 -9.41067874e-01
-1.83854818e+00 8.30029309e-01 7.53822327e-02 -1.08900499e+00
5.95852196e-01 -4.59013909e-01 -1.03150971e-01 -1.23789303e-01
-3.06438208e-01 5.59760630e-01 4.66006488e-01 -1.12166250e+00
-3.48604977e-01 -3.30453545e-01 5.45857009e-03 1.36790335e-01
-8.39766324e-01 1.22927584e-01 9.76940021e-02 -7.21974611e-01
-2.27306366e-01 -7.03890741e-01 -1.23746488e-02 -3.48077357e-01
-8.45335603e-01 -3.96348327e-01 7.98662484e-01 -9.61251080e-01
1.05037391e+00 -1.95100963e+00 1.32913232e-01 -1.70307815e-01
2.97612697e-02 3.35108817e-01 1.80045113e-01 3.10951293e-01
-8.00687820e-03 4.17229116e-01 -2.99643189e-01 -8.64786506e-01
1.55942038e-01 -1.82536046e-03 -6.30641878e-01 2.62623549e-01
2.07071498e-01 1.00608265e+00 -4.26812142e-01 -5.61702073e-01
-4.23787832e-02 5.91731250e-01 -3.25444371e-01 1.19769931e-01
-2.02651009e-01 2.41714582e-01 -2.00702772e-01 3.37187022e-01
2.37848192e-01 -3.01571220e-01 2.24152431e-01 -1.78128466e-01
2.57001728e-01 9.79165912e-01 -6.95919693e-01 1.48059344e+00
-7.41306365e-01 9.10778642e-01 -8.35180357e-02 -1.55382276e+00
8.02440047e-01 6.19432390e-01 2.56651461e-01 -2.91963786e-01
3.94295275e-01 1.63846031e-01 -1.87716708e-01 -1.24902859e-01
5.19987166e-01 -1.66601896e-01 -6.19991980e-02 5.23171902e-01
4.22600836e-01 -2.71106474e-02 6.62206188e-02 1.71465158e-01
7.04197109e-01 2.98371196e-01 2.05950215e-01 -1.93280637e-01
5.24146914e-01 -1.51777983e-01 3.31410557e-01 6.34084463e-01
-6.28567711e-02 3.52552831e-01 5.98695278e-01 -2.53313065e-01
-9.53478158e-01 -4.91347402e-01 -3.97989392e-01 1.28881240e+00
-3.15943211e-01 -3.31983238e-01 -7.50076890e-01 -1.12934160e+00
-2.05134317e-01 9.63447928e-01 -6.55377090e-01 -3.07862252e-01
-3.52041692e-01 -1.05703592e+00 8.29586565e-01 4.93236154e-01
8.51354182e-01 -1.00780356e+00 -1.29116714e-01 2.50869572e-01
-3.90817940e-01 -1.31712723e+00 -3.39172006e-01 7.88037717e-01
-9.35141742e-01 -5.52638948e-01 -5.89496195e-01 -9.71917152e-01
3.18009466e-01 -3.13497156e-01 1.06721413e+00 -2.03277126e-01
1.83973938e-01 1.60615221e-02 -4.25077707e-01 -3.17180455e-01
-4.67585206e-01 6.97853506e-01 -2.65064299e-01 -4.36803102e-02
2.99781501e-01 -4.30789918e-01 -1.67055205e-01 4.82547171e-02
-6.77416503e-01 1.92951560e-01 3.33916575e-01 1.22858524e+00
3.51749241e-01 5.11057712e-02 7.14882553e-01 -1.10828876e+00
6.40281200e-01 -3.23720336e-01 -2.39567131e-01 3.54972899e-01
-8.50328505e-01 1.09284393e-01 9.01362240e-01 -4.72780526e-01
-1.17306125e+00 -3.04046392e-01 -1.94202572e-01 -7.23832324e-02
-8.18032846e-02 7.47613072e-01 -1.30775079e-01 1.38579294e-01
5.65122008e-01 2.77120531e-01 -1.36484683e-01 -4.46220964e-01
3.93457025e-01 9.38050449e-01 -6.44795075e-02 -3.26479703e-01
5.63482404e-01 3.11850578e-01 -1.60446346e-01 -4.26601648e-01
-1.43694007e+00 7.86678120e-03 -8.26672137e-01 2.02395499e-01
9.33654904e-01 -8.39381993e-01 -5.43410897e-01 5.82587004e-01
-1.19013047e+00 -4.51751500e-01 -9.06710029e-02 5.67412078e-01
-3.57329160e-01 3.54367882e-01 -1.16106844e+00 -8.12194824e-01
-7.49808669e-01 -1.09479606e+00 8.07283938e-01 -2.02427894e-01
-2.71763712e-01 -1.44187117e+00 -1.84852839e-01 4.88892525e-01
5.04933536e-01 8.98881257e-02 1.05626917e+00 -1.26202333e+00
4.13685217e-02 -2.67332494e-01 -5.78980520e-02 8.69929790e-01
2.68912185e-02 2.12184358e-02 -1.16084015e+00 -2.53836483e-01
1.72783747e-01 -8.78814101e-01 1.28582609e+00 1.77049980e-01
1.18471980e+00 -4.57214057e-01 -2.37170145e-01 5.37947297e-01
9.93478775e-01 4.85164486e-02 4.64352816e-01 6.67131066e-01
8.15794349e-01 8.13241422e-01 4.33014333e-01 2.74980851e-02
6.66856825e-01 6.72841251e-01 -1.65159851e-02 -2.25646585e-01
1.14686742e-01 -2.04460248e-01 3.93932879e-01 1.08243537e+00
2.05321729e-01 -5.61709762e-01 -9.15297031e-01 3.02572042e-01
-1.88050842e+00 -7.63484180e-01 -5.80808558e-02 2.03517842e+00
1.17273259e+00 4.47242945e-01 3.98055092e-02 4.06892091e-01
5.28727710e-01 2.57894367e-01 -3.82526159e-01 -6.31918550e-01
-2.73388416e-01 6.65475726e-02 4.28363502e-01 4.49920952e-01
-1.49588001e+00 1.11989427e+00 6.08274794e+00 8.79533708e-01
-1.19297051e+00 2.84927040e-01 1.17804718e+00 1.33526884e-02
-8.92380476e-02 -7.28629455e-02 -9.45539236e-01 4.02880788e-01
1.15335536e+00 1.54073797e-02 1.50344416e-01 7.41385818e-01
-2.97515541e-01 1.17162101e-01 -1.05392516e+00 5.33077180e-01
2.50462353e-01 -9.09193575e-01 -1.87711406e-03 -8.17519948e-02
6.08485579e-01 2.05622874e-02 8.10119435e-02 6.15844369e-01
4.19846863e-01 -1.00570762e+00 5.66588283e-01 1.30460098e-01
4.32492465e-01 -7.68976629e-01 1.04145300e+00 4.64019358e-01
-7.56757200e-01 2.08412141e-01 -2.49313354e-01 6.93396665e-03
1.04851965e-02 8.24095726e-01 -5.25256455e-01 7.69441009e-01
6.50546372e-01 8.71211588e-01 -4.94185507e-01 3.86733592e-01
-3.01670313e-01 1.01056862e+00 -2.49611095e-01 -1.75975546e-01
1.26049533e-01 -1.30423367e-01 2.73052007e-01 1.25342655e+00
-7.41384849e-02 -4.12172705e-01 3.91929746e-02 6.53339148e-01
-4.34834361e-01 4.59334224e-01 -4.97666121e-01 -1.36375099e-01
9.67546478e-02 1.32467675e+00 -5.59154391e-01 -4.48477417e-01
-4.69172329e-01 7.95580804e-01 6.64179504e-01 4.99405026e-01
-7.24744081e-01 -5.46477258e-01 1.56791106e-01 -2.91397244e-01
3.05558652e-01 -1.00578971e-01 -5.92296183e-01 -1.39637601e+00
1.25536233e-01 -7.32529283e-01 3.96007836e-01 -4.51722831e-01
-1.39938009e+00 9.76466000e-01 -1.08633988e-01 -9.71612513e-01
-5.58112144e-01 -5.71540117e-01 -5.30493677e-01 7.69656301e-01
-1.68955350e+00 -1.52191103e+00 8.85939300e-02 2.81976372e-01
3.99724573e-01 -3.60959619e-01 8.70476723e-01 3.21915150e-01
-8.24452221e-01 9.11120832e-01 3.06807309e-01 3.48067820e-01
7.88554728e-01 -1.24539852e+00 3.76928031e-01 5.98265290e-01
2.71168858e-01 2.72171736e-01 4.49293524e-01 -4.09398377e-01
-9.26631987e-01 -1.11437511e+00 1.16034973e+00 -4.48458135e-01
8.46337855e-01 -7.92500854e-01 -1.12218142e+00 1.10444653e+00
5.64627409e-01 -1.70709893e-01 8.99151385e-01 3.07323396e-01
-4.48249310e-01 -7.10161179e-02 -1.00781560e+00 4.53127384e-01
5.71512818e-01 -5.39794564e-01 -5.87976038e-01 4.71926808e-01
8.08246076e-01 -2.25372210e-01 -1.04788196e+00 6.63662791e-01
4.01713580e-01 -7.93985665e-01 6.14690423e-01 -8.18011999e-01
7.72677124e-01 2.95624018e-01 -3.82709764e-02 -1.25289035e+00
1.51116133e-01 -4.21876490e-01 -1.36949956e-01 1.71529770e+00
8.37960780e-01 -7.92902708e-01 5.66470087e-01 4.12888616e-01
6.89194649e-02 -1.02060032e+00 -8.55929375e-01 -6.52556777e-01
7.22355664e-01 -5.37246585e-01 1.52542859e-01 1.08083570e+00
3.39580745e-01 6.52313650e-01 -4.68900353e-01 -4.35720563e-01
5.21271467e-01 6.49050474e-02 5.60203493e-01 -1.14701962e+00
-1.34160176e-01 -6.24390781e-01 -1.50647879e-01 -1.01525307e+00
8.91768396e-01 -1.12608385e+00 3.15503217e-02 -1.50195599e+00
2.69589543e-01 -1.62806913e-01 -3.97882074e-01 7.08508372e-01
-1.86902016e-01 1.24545470e-01 4.71473634e-02 2.08967090e-01
-5.68861187e-01 8.72075677e-01 9.23512876e-01 -2.98850626e-01
-7.65562132e-02 -6.76778778e-02 -6.00571513e-01 6.51011646e-01
9.29139555e-01 -4.68961626e-01 -1.46103740e-01 -3.42775017e-01
2.23175153e-01 -1.37962118e-01 1.95151642e-01 -6.33523524e-01
2.25686386e-01 3.70469004e-01 2.87631869e-01 -5.22404611e-01
3.65804464e-01 -5.93107402e-01 -5.00996888e-01 3.85339886e-01
-9.20497000e-01 -4.35972899e-01 2.70187497e-01 3.52715254e-01
-5.93479633e-01 -2.12335005e-01 7.22598374e-01 2.38269866e-01
1.43996775e-01 3.67011433e-03 -4.48522955e-01 1.32085308e-01
6.89770997e-01 2.52915263e-01 -4.24325168e-01 -6.85956717e-01
-6.22814298e-01 2.95127183e-01 4.92776707e-02 3.79817635e-01
1.70571402e-01 -1.24089634e+00 -7.51014113e-01 7.83842504e-02
-1.38224423e-01 -1.13937296e-02 -1.07744478e-01 1.03083539e+00
-6.56614676e-02 6.93435669e-01 2.67276287e-01 -3.33776385e-01
-1.04105902e+00 3.96960884e-01 4.69322562e-01 -8.34720194e-01
-3.41850638e-01 7.42976844e-01 1.19798198e-01 -7.57719934e-01
5.88625491e-01 -3.72968167e-01 -3.56087685e-01 2.97950983e-01
1.39728278e-01 1.79347575e-01 3.88607085e-01 -4.69569415e-01
-2.95841515e-01 2.19934627e-01 -4.54763025e-01 -1.29615292e-01
1.23721218e+00 -1.20828614e-01 -6.22871555e-02 8.34187329e-01
1.44484985e+00 -3.44108433e-01 -7.87778318e-01 -5.44946611e-01
2.39877343e-01 2.67098069e-01 3.47809017e-01 -9.10410106e-01
-1.07833087e+00 1.26383924e+00 3.11276913e-01 5.54377019e-01
1.04714549e+00 -1.18704394e-01 8.59839618e-01 6.88712120e-01
4.83918451e-02 -1.10954082e+00 -1.47622541e-01 8.68842065e-01
6.95639431e-01 -1.41254532e+00 -1.47851974e-01 -3.32960635e-01
-7.60172784e-01 1.04950273e+00 5.15566647e-01 1.55640855e-01
8.61638308e-01 3.71956527e-01 3.47512998e-02 2.13760048e-01
-9.61674213e-01 -4.47989143e-02 4.42953944e-01 4.13762033e-02
9.32498276e-01 -2.65292138e-01 -4.04547989e-01 8.55137765e-01
-3.59318942e-01 -1.14710480e-01 2.77132839e-01 7.74092793e-01
-5.22771291e-02 -1.15801167e+00 5.78171946e-02 6.31618917e-01
-1.03348303e+00 -4.17681485e-01 -5.09146214e-01 8.79783154e-01
-4.47937727e-01 1.01949680e+00 -2.04308294e-02 -5.17649114e-01
1.37261242e-01 5.56088507e-01 1.77621827e-01 -5.82000256e-01
-8.94122243e-01 1.91030409e-02 4.49193120e-01 4.38966118e-02
-5.14875531e-01 -4.44773793e-01 -1.05628407e+00 -1.40992686e-01
-7.73153782e-01 2.48934522e-01 7.59628594e-01 1.29650342e+00
2.30957925e-01 6.09060228e-01 5.78681707e-01 -5.71641445e-01
-7.04800069e-01 -1.60768545e+00 -4.58671004e-01 3.84704173e-01
2.23334417e-01 -5.30151367e-01 -7.23379672e-01 1.56156272e-01] | [10.628454208374023, 8.137564659118652] |
d3757d30-2efd-45e2-89b0-2340048e588f | pixor-real-time-3d-object-detection-from | 1902.06326 | null | http://arxiv.org/abs/1902.06326v3 | http://arxiv.org/pdf/1902.06326v3.pdf | PIXOR: Real-time 3D Object Detection from Point Clouds | We address the problem of real-time 3D object detection from point clouds in
the context of autonomous driving. Computation speed is critical as detection
is a necessary component for safety. Existing approaches are, however,
expensive in computation due to high dimensionality of point clouds. We utilize
the 3D data more efficiently by representing the scene from the Bird's Eye View
(BEV), and propose PIXOR, a proposal-free, single-stage detector that outputs
oriented 3D object estimates decoded from pixel-wise neural network
predictions. The input representation, network architecture, and model
optimization are especially designed to balance high accuracy and real-time
efficiency. We validate PIXOR on two datasets: the KITTI BEV object detection
benchmark, and a large-scale 3D vehicle detection benchmark. In both datasets
we show that the proposed detector surpasses other state-of-the-art methods
notably in terms of Average Precision (AP), while still runs at >28 FPS. | ['Bin Yang', 'Raquel Urtasun', 'Wenjie Luo'] | 2019-02-17 | pixor-real-time-3d-object-detection-from-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Yang_PIXOR_Real-Time_3D_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Yang_PIXOR_Real-Time_3D_CVPR_2018_paper.pdf | cvpr-2018-6 | ['birds-eye-view-object-detection'] | ['computer-vision'] | [-3.18215191e-02 -2.57118702e-01 6.58419132e-02 -4.00004297e-01
-6.31916106e-01 -5.70665836e-01 5.99920332e-01 1.18487336e-01
-7.74446130e-01 -1.56537622e-01 -5.76292157e-01 -6.71926200e-01
4.04389709e-01 -6.42145038e-01 -1.03259361e+00 -3.20823789e-01
6.37894049e-02 5.48676670e-01 8.97582650e-01 -2.02158839e-01
3.08123648e-01 1.09348261e+00 -2.04493952e+00 -1.45819649e-01
4.90399063e-01 1.38469410e+00 3.47943693e-01 7.65819907e-01
1.29342303e-01 2.87893623e-01 -1.93550378e-01 -1.38934314e-01
6.35200560e-01 4.07213688e-01 -4.10026386e-02 4.75476794e-02
8.06760669e-01 -5.57155490e-01 -2.42174596e-01 1.01578808e+00
2.75448084e-01 -5.63739911e-02 5.45622408e-01 -1.55935609e+00
1.22337379e-01 -3.65933120e-01 -7.28067279e-01 2.98324138e-01
-3.38960513e-02 5.28619766e-01 6.18964434e-01 -1.11645877e+00
4.08423305e-01 1.42373908e+00 6.51871741e-01 5.06991148e-01
-1.17378557e+00 -7.04922616e-01 2.36556619e-01 2.17758551e-01
-1.41620183e+00 -5.47508538e-01 5.80009103e-01 -7.17685342e-01
1.40316176e+00 9.52315778e-02 5.77846587e-01 5.82000077e-01
2.90425152e-01 8.32147896e-01 8.52573574e-01 -1.16950624e-01
3.38381052e-01 2.67163604e-01 3.07690591e-01 6.81258380e-01
5.10270715e-01 6.76591754e-01 -4.66320515e-01 8.64328146e-02
4.24899489e-01 4.14222367e-02 1.31978408e-01 -8.84342074e-01
-1.13125467e+00 7.36510277e-01 5.39311409e-01 -5.21919250e-01
-3.61926079e-01 4.21279609e-01 2.28722230e-01 4.22779061e-02
4.31496769e-01 -3.23338509e-02 -3.85114372e-01 -6.30169213e-02
-6.93908989e-01 4.92576987e-01 4.11366701e-01 1.12965810e+00
8.13596845e-01 -2.92460658e-02 9.52394083e-02 2.33744040e-01
5.83938777e-01 9.72275496e-01 -1.85117736e-01 -9.76817131e-01
5.01985252e-01 7.73826957e-01 4.17244583e-01 -7.18447924e-01
-6.23035133e-01 -2.60619223e-01 -3.68623912e-01 1.10815752e+00
1.92870095e-01 2.17589602e-01 -9.77987647e-01 1.07797098e+00
7.22838402e-01 1.16769329e-01 3.94336134e-02 1.24484622e+00
9.47584748e-01 7.54869401e-01 -6.67017326e-02 3.44122797e-01
1.41123021e+00 -7.67605841e-01 -8.26831385e-02 -8.06960106e-01
5.35232782e-01 -4.15633917e-01 6.63728833e-01 2.58690089e-01
-8.96469891e-01 -6.91628218e-01 -1.20876014e+00 -2.42709711e-01
-3.56885105e-01 3.90191883e-01 2.05438659e-01 5.06733000e-01
-8.43252182e-01 1.18144967e-01 -1.08769500e+00 -3.81096900e-01
5.27178228e-01 3.94336849e-01 -3.62854242e-01 7.50227273e-02
-4.05895770e-01 1.25690663e+00 2.93941408e-01 -2.49533337e-02
-9.67061818e-01 -8.48794878e-01 -8.10929716e-01 -2.53921710e-02
4.33463097e-01 -4.96610045e-01 1.42305899e+00 -1.25770941e-01
-1.29062486e+00 1.18620658e+00 -2.64469504e-01 -9.03366268e-01
7.00701416e-01 -4.35024768e-01 -9.09193084e-02 -4.60452260e-03
-1.12854861e-01 1.20124626e+00 9.16665733e-01 -1.27291834e+00
-1.23101580e+00 -5.99429905e-01 -9.46813524e-02 1.45706788e-01
1.89810440e-01 7.19118956e-03 -6.32892430e-01 2.09629983e-01
2.72352725e-01 -1.20695782e+00 -3.47018003e-01 5.73985100e-01
-2.11725101e-01 -3.06745708e-01 1.16679716e+00 -1.27548143e-01
3.20556760e-01 -2.32056999e+00 -3.56516093e-01 -1.62312329e-01
4.21909332e-01 3.94137681e-01 7.92817101e-02 -4.64139991e-02
2.58185595e-01 -4.70113009e-01 -7.39953667e-02 -5.84131002e-01
2.02732041e-01 1.80227876e-01 -5.02623796e-01 7.45119572e-01
4.76407439e-01 8.28503013e-01 -6.25788629e-01 -3.14886570e-01
7.28136063e-01 5.32881439e-01 -5.02952218e-01 1.35279670e-01
-2.92919725e-01 6.95028231e-02 -4.02167767e-01 6.57517493e-01
9.29377496e-01 3.57806459e-02 -5.68164229e-01 -9.33018327e-02
-5.93298972e-01 2.58631587e-01 -1.03596306e+00 1.28483784e+00
-1.45427361e-01 9.98818278e-01 2.10329995e-01 -7.45551825e-01
1.18721628e+00 -1.78669170e-01 1.88043013e-01 -7.33856499e-01
3.41715097e-01 2.15602860e-01 -2.88273424e-01 -2.44826034e-01
8.71499240e-01 2.02129751e-01 2.51154718e-03 -4.72227251e-03
-4.95962501e-01 -5.05343854e-01 1.74744315e-02 5.41997403e-02
1.10569537e+00 8.21743384e-02 1.33803710e-01 -2.04794019e-01
1.99543551e-01 5.41939139e-01 3.36535633e-01 6.64358258e-01
-5.31135798e-01 5.13386905e-01 4.28831607e-01 -5.49637377e-01
-1.11553574e+00 -1.06343544e+00 -2.48108372e-01 7.06618667e-01
6.95470572e-01 -9.03899819e-02 -4.73991483e-01 -5.42890310e-01
6.23907506e-01 8.12961459e-01 -3.10263962e-01 -1.08221978e-01
-5.11052907e-01 -3.66651058e-01 9.28210840e-02 6.60724044e-01
1.67698190e-01 -5.47224998e-01 -1.69734454e+00 1.81991160e-01
4.02141422e-01 -1.74188077e+00 -3.11202127e-02 2.84878582e-01
-8.65851045e-01 -9.95762527e-01 -1.43937185e-01 -3.61459792e-01
5.78012764e-01 9.12493169e-01 1.15345669e+00 -1.57846436e-01
-4.41968560e-01 3.59180808e-01 -7.42200315e-02 -1.06736970e+00
-4.41575050e-01 -2.39229113e-01 2.45895073e-01 -2.66850054e-01
8.52745891e-01 -1.31826326e-01 -6.12902224e-01 4.75126028e-01
-4.23795938e-01 1.79416344e-01 6.51304483e-01 2.55370319e-01
7.63051987e-01 -2.25915909e-01 -7.62762427e-02 -2.26274803e-01
-2.76588257e-02 -2.69575119e-02 -1.77447999e+00 -4.67525750e-01
-6.23948216e-01 -1.28427982e-01 1.22133054e-01 -8.85519087e-02
-5.12379408e-01 8.08459044e-01 -4.18685265e-02 -8.90621006e-01
-3.62153679e-01 -2.09596515e-01 2.03552797e-01 -4.25783902e-01
7.85998106e-01 4.26766798e-02 2.18974262e-01 -4.61180568e-01
3.22525293e-01 5.27787805e-01 7.43900418e-01 2.09370870e-02
9.71891999e-01 8.10955405e-01 3.43337089e-01 -8.22651505e-01
-5.02078593e-01 -8.08148801e-01 -7.74162352e-01 -4.50595886e-01
9.09890771e-01 -1.29135919e+00 -1.10344827e+00 3.01727176e-01
-1.52183962e+00 -1.01953998e-01 -1.25022024e-01 5.70102036e-01
-5.75762451e-01 -4.50299382e-02 -1.59269258e-01 -1.08485603e+00
-2.46136844e-01 -1.37393188e+00 1.59360945e+00 2.79757474e-02
2.89292991e-01 -1.29534870e-01 -1.94249868e-01 1.57535300e-01
1.81391209e-01 2.82999873e-01 4.77328509e-01 -1.56575233e-01
-1.14125478e+00 -6.24240935e-01 -7.07732320e-01 1.25833958e-01
-5.99857986e-01 1.04771890e-01 -1.14824045e+00 -1.42185926e-01
-3.03557605e-01 -1.52387451e-02 1.08551025e+00 3.68703812e-01
8.13514113e-01 2.05103591e-01 -7.39962161e-01 5.48574150e-01
1.50430453e+00 1.14864055e-02 2.68907905e-01 4.09335047e-01
5.80643713e-01 5.45619667e-01 8.08772087e-01 3.62636685e-01
4.68406767e-01 8.20364594e-01 1.04495597e+00 -3.43148485e-02
-6.74376711e-02 -7.31919780e-02 4.14148420e-01 9.01828334e-02
1.55019253e-01 -5.54669797e-02 -1.20997429e+00 6.68533385e-01
-1.88147175e+00 -7.18990922e-01 -4.74883497e-01 2.21535754e+00
3.38089205e-02 6.80524886e-01 3.54605079e-01 6.86634704e-02
4.24980551e-01 -8.13149512e-02 -7.70327628e-01 -1.60506994e-01
2.35406399e-01 -3.62930685e-01 1.17971420e+00 3.14096808e-01
-1.28477633e+00 8.46699953e-01 5.97123718e+00 3.50349993e-01
-1.26234841e+00 1.38302967e-01 2.25490153e-01 -4.61547166e-01
2.96087354e-01 -2.15713650e-01 -1.50766516e+00 2.57351726e-01
9.79390621e-01 1.83698311e-02 1.09319203e-02 1.36958194e+00
3.18447262e-01 -1.98743433e-01 -1.18806195e+00 1.21458244e+00
-1.39577910e-01 -1.29223573e+00 -4.70068485e-01 1.24581151e-01
2.60193229e-01 9.42768455e-01 -2.54496560e-02 1.53543621e-01
3.06792468e-01 -7.54240334e-01 1.27607989e+00 1.83913082e-01
5.80590785e-01 -6.51619494e-01 4.91116047e-01 8.08682144e-01
-1.25091660e+00 -9.04542357e-02 -6.19016826e-01 -7.25839660e-02
2.61895269e-01 4.58795190e-01 -1.04588997e+00 -1.23966970e-01
1.04764140e+00 6.13444030e-01 -6.05286360e-01 1.23377764e+00
-5.88233881e-02 3.50538313e-01 -8.05452228e-01 -2.62149692e-01
4.83886331e-01 3.66434865e-02 8.43152463e-01 1.11801994e+00
4.03443962e-01 1.32381842e-01 2.39601284e-01 1.06075585e+00
2.62572438e-01 -3.85596305e-01 -8.78229141e-01 4.20121998e-01
4.66154814e-01 1.31285632e+00 -6.68567359e-01 -2.00583234e-01
-5.28947413e-01 3.11429411e-01 1.63777709e-01 1.43988617e-02
-8.94696712e-01 -6.42392337e-02 1.13338602e+00 4.51096654e-01
7.65335381e-01 -7.01539457e-01 -3.75611126e-01 -6.11371636e-01
2.82252371e-01 -2.35221058e-01 -2.74588376e-01 -8.78633738e-01
-6.94256723e-01 5.60535252e-01 1.19257130e-01 -1.65962708e+00
-1.10330321e-01 -1.15752971e+00 -2.25146204e-01 6.37507439e-01
-2.06402969e+00 -1.01007628e+00 -6.38919353e-01 1.37940362e-01
5.08569062e-01 2.15838607e-02 2.80782104e-01 2.80390680e-01
-4.20717865e-01 2.16351792e-01 -2.62871031e-02 -1.95817888e-01
2.25652412e-01 -8.66141021e-01 1.08355200e+00 9.78459358e-01
-5.86055126e-03 -1.00827821e-01 7.66258180e-01 -3.75573516e-01
-2.03977346e+00 -1.36666405e+00 6.14929974e-01 -9.00516868e-01
3.75875533e-01 -6.71287775e-01 -7.65996039e-01 5.03726304e-01
-3.38172674e-01 4.32811379e-01 -1.68514758e-01 -4.39384609e-01
-2.70046592e-01 -2.92089581e-01 -1.10401595e+00 4.44117188e-01
1.10570204e+00 -1.68381244e-01 -3.86748642e-01 2.45523855e-01
8.72339487e-01 -8.59859288e-01 -2.52854198e-01 5.17056406e-01
4.84357655e-01 -1.05125618e+00 1.19643497e+00 -2.72760868e-01
4.55259234e-02 -7.06398070e-01 -2.98800796e-01 -7.89314747e-01
-3.64600182e-01 -2.96808720e-01 -1.88816622e-01 5.55508673e-01
3.87345254e-01 -6.88242435e-01 1.09518552e+00 6.25070989e-01
-4.64709491e-01 -3.71197313e-01 -1.23978233e+00 -1.02594030e+00
-4.32277799e-01 -9.47081566e-01 6.32668912e-01 9.99874920e-02
-5.68748534e-01 3.65562439e-01 -7.39052594e-02 6.76929355e-01
9.12261963e-01 2.97010332e-01 1.25318432e+00 -1.46118724e+00
2.07079977e-01 -5.60134947e-01 -1.02752805e+00 -1.37268841e+00
-3.66375335e-02 -5.75365901e-01 5.06085396e-01 -1.27806520e+00
-2.46684298e-01 -5.49972713e-01 1.20136328e-02 3.39143306e-01
5.92794269e-02 3.39334548e-01 4.30413842e-01 2.90847421e-01
-5.96373677e-01 4.12696928e-01 7.29097247e-01 -1.43410474e-01
-1.24813318e-01 1.06380209e-01 -1.65819183e-01 8.00173342e-01
5.46647787e-01 -5.37207603e-01 -1.05771668e-01 -7.26965070e-01
1.80846304e-02 -7.82620609e-02 8.69991422e-01 -1.44693696e+00
4.63372290e-01 -1.29199117e-01 2.59332955e-01 -1.43759096e+00
8.30694616e-01 -1.12827742e+00 -1.47861630e-01 6.04157984e-01
1.29940599e-01 1.58061385e-01 5.89983881e-01 6.02953672e-01
1.02793835e-01 1.69219986e-01 1.05007422e+00 1.27108037e-01
-1.28710103e+00 4.46919829e-01 -2.26582155e-01 -3.09519053e-01
1.33542669e+00 -4.37335998e-01 -3.60108197e-01 1.44808680e-01
2.36248448e-02 3.59757304e-01 6.28438890e-01 7.49046564e-01
8.07434022e-01 -1.13925314e+00 -8.07626963e-01 5.60848176e-01
5.81117332e-01 4.58286107e-01 -1.06013343e-01 7.39713490e-01
-7.07518935e-01 7.79228151e-01 -5.94426654e-02 -1.38627112e+00
-1.37042356e+00 7.51348913e-01 3.38973075e-01 4.05299395e-01
-1.02026772e+00 7.58767307e-01 4.31647032e-01 -3.62528712e-01
3.89291137e-01 -6.00987911e-01 1.25606149e-01 -4.28360820e-01
5.15491605e-01 2.51771748e-01 3.42688560e-01 -7.42191494e-01
-5.79201341e-01 8.00388396e-01 1.10101543e-01 1.69557974e-01
1.10566592e+00 -4.78389636e-02 3.98388803e-01 1.96410462e-01
9.92266655e-01 -4.90480453e-01 -1.74120617e+00 -2.09203705e-01
-1.34889022e-01 -6.57484710e-01 4.95821476e-01 -3.11846972e-01
-8.42271447e-01 1.02922308e+00 1.11095071e+00 5.66683114e-02
6.60294175e-01 2.04919264e-01 7.49770820e-01 6.82135165e-01
4.60449636e-01 -8.29658866e-01 -3.68775487e-01 5.90504706e-01
6.15336299e-01 -1.64293277e+00 1.35527030e-01 -4.58105594e-01
-3.66578609e-01 8.51907909e-01 8.02683532e-01 -2.82487184e-01
4.64850843e-01 4.61050540e-01 1.59091562e-01 -3.37716609e-01
-8.99000585e-01 -3.52490574e-01 3.77206653e-01 5.72864592e-01
-3.46103936e-01 3.79039608e-02 3.96033078e-01 -9.00899526e-04
-1.01183340e-01 -1.01330027e-01 1.69854045e-01 9.74316537e-01
-9.56637919e-01 -3.76964569e-01 -5.60653448e-01 3.15285116e-01
1.13696054e-01 2.56306499e-01 -1.62092552e-01 8.67317736e-01
2.04389077e-02 8.51867199e-01 4.68275279e-01 -3.55336845e-01
7.13549197e-01 -2.64972657e-01 2.87371874e-01 -3.54924738e-01
-1.33258775e-01 -2.20661700e-01 -1.35199189e-01 -8.67783546e-01
-1.21971235e-01 -7.35050261e-01 -1.19404006e+00 -2.30035976e-01
-3.65785122e-01 -3.54526043e-01 1.41506922e+00 7.76920497e-01
7.56170273e-01 2.26695493e-01 5.88090241e-01 -1.37780917e+00
-5.52013099e-01 -4.94809091e-01 -2.93966055e-01 4.57853936e-02
7.23850846e-01 -7.99769759e-01 -1.74746484e-01 -1.24386296e-01] | [7.73106575012207, -2.5710713863372803] |
f8003472-2f78-4dd0-8bb7-9ed90003d0d3 | multimodal-emotion-recognition-on-ravdess | null | null | https://www.mdpi.com/1424-8220/21/22/7665/htm | https://www.mdpi.com/1424-8220/21/22/7665/pdf | Multimodal Emotion Recognition on RAVDESS Dataset Using Transfer Learning | Emotion Recognition is attracting the attention of the research community due to the multiple areas where it can be applied, such as in healthcare or in road safety systems. In this paper, we propose a multimodal emotion recognition system that relies on speech and facial information. For the speech-based modality, we evaluated several transfer-learning techniques, more specifically, embedding extraction and Fine-Tuning. The best accuracy results were achieved when we fine-tuned the CNN-14 of the PANNs framework, confirming that the training was more robust when it did not start from scratch and the tasks were similar. Regarding the facial emotion recognizers, we propose a framework that consists of a pre-trained Spatial Transformer Network on saliency maps and facial images followed by a bi-LSTM with an attention mechanism. The error analysis reported that the frame-based systems could present some problems when they were used directly to solve a video-based task despite the domain adaptation, which opens a new line of research to discover new ways to correct this mismatch and take advantage of the embedded knowledge of these pre-trained models. Finally, from the combination of these two modalities with a late fusion strategy, we achieved 80.08% accuracy on the RAVDESS dataset on a subject-wise 5-CV evaluation, classifying eight emotions. The results revealed that these modalities carry relevant information to detect users’ emotional state and their combination enables improvement of system performance. | ['Fernando Fernández-Martínez', 'Juan M. Montero', 'Ricardo Kleinlein', 'Zoraida Callejas', 'David Griol', 'Cristina Luna-Jiménez'] | 2021-11-18 | null | null | null | sensors-2021-11 | ['facial-emotion-recognition', 'multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'computer-vision', 'speech'] | [ 1.59453437e-01 4.30372134e-02 1.67973742e-01 -3.53712201e-01
-4.57624793e-01 -6.36903644e-02 6.01551056e-01 -1.54241338e-01
-6.95351779e-01 5.86217642e-01 9.11740586e-02 1.63695827e-01
-3.08845639e-02 -5.16202152e-01 -6.93812609e-01 -7.43574798e-01
1.49458677e-01 -1.97846934e-01 2.49362633e-01 -5.45946956e-01
3.54015708e-01 4.82767433e-01 -2.18168688e+00 5.13406694e-01
8.00235927e-01 1.33237374e+00 2.50321925e-01 4.09062147e-01
-1.42425179e-01 6.71314240e-01 -4.97151464e-01 -4.87216592e-01
-3.05489115e-02 -3.68953288e-01 -6.05625033e-01 -1.64502934e-01
1.26105502e-01 -7.19371960e-02 -6.52354434e-02 9.05438125e-01
7.14307606e-01 1.96354687e-01 4.40611869e-01 -1.41126871e+00
-4.01251644e-01 6.23914823e-02 -3.40671718e-01 8.20890665e-02
4.88265187e-01 -3.15017030e-02 4.74199057e-01 -9.67948973e-01
4.60996538e-01 1.18465221e+00 5.36844671e-01 5.47670662e-01
-7.88329959e-01 -6.15781903e-01 1.64129375e-03 9.22019303e-01
-1.27385247e+00 -5.27915120e-01 8.70243490e-01 -4.42387670e-01
1.04123127e+00 9.09471437e-02 7.13450193e-01 1.52766001e+00
1.42354280e-01 4.83773381e-01 1.34741712e+00 -5.54911792e-01
8.69505703e-02 7.24345148e-01 -7.53516629e-02 5.43897331e-01
-2.61367917e-01 1.26129672e-01 -5.95288098e-01 2.08662450e-01
5.15833795e-01 3.02852690e-02 -2.08720952e-01 -1.52030453e-01
-9.79349434e-01 7.37477720e-01 5.77615380e-01 1.03058875e+00
-6.42611027e-01 -2.28449896e-01 6.75261796e-01 2.77369320e-01
5.18154263e-01 2.96305150e-01 -4.50917870e-01 -1.51566297e-01
-1.02017379e+00 -3.58206213e-01 4.97365236e-01 3.07608604e-01
7.54516900e-01 4.57855538e-02 -3.06811541e-01 8.07860494e-01
2.27502242e-01 4.22679991e-01 7.41533279e-01 -4.72031295e-01
1.44755751e-01 6.92781448e-01 -2.11290151e-01 -1.39668536e+00
-5.23486257e-01 -3.23075086e-01 -7.16086924e-01 3.44778448e-01
1.85534075e-01 -1.79420277e-01 -7.80104995e-01 1.88481009e+00
2.58177549e-01 4.71667618e-01 2.13365212e-01 1.10634744e+00
8.64802480e-01 6.43435836e-01 3.97776634e-01 -1.60633266e-01
1.48018408e+00 -9.82805967e-01 -1.02578294e+00 1.99181363e-02
5.02390504e-01 -7.24964023e-01 9.90087390e-01 5.41453660e-01
-8.85290921e-01 -8.14456820e-01 -1.13725054e+00 1.93149820e-01
-8.79486382e-01 4.80888456e-01 3.03794891e-01 7.91884899e-01
-1.30304086e+00 6.45907104e-01 -4.69237864e-01 -7.65363395e-01
2.92959124e-01 2.67239720e-01 -6.04177177e-01 2.78428853e-01
-1.43764317e+00 1.36925185e+00 3.53488654e-01 5.21105111e-01
-6.48125112e-01 -2.43543908e-01 -6.35716379e-01 2.84432620e-01
2.19427481e-01 -5.07409275e-01 6.98193431e-01 -1.62655234e+00
-1.84766269e+00 7.59921432e-01 -2.09973305e-01 -3.28838885e-01
3.42353791e-01 -2.67182559e-01 -8.33615065e-01 3.63859534e-01
-1.17172025e-01 5.82492054e-01 1.08977306e+00 -8.89639676e-01
-4.96167719e-01 -4.92638588e-01 -1.54488236e-01 1.19364023e-01
-7.80707479e-01 1.70679659e-01 -1.73410580e-01 -3.70373726e-01
-2.99415648e-01 -7.54866123e-01 1.96686447e-01 -2.89516598e-01
-4.71501052e-02 -1.26664788e-01 9.17460620e-01 -8.68623734e-01
1.15661585e+00 -2.28961968e+00 3.38334054e-01 2.99230099e-01
-1.56509921e-01 6.87006354e-01 -2.39127919e-01 3.75592560e-01
-4.25355852e-01 5.53768687e-02 -2.89262906e-02 -3.28077853e-01
-1.75654426e-01 -2.57045150e-01 -4.03234288e-02 3.94232452e-01
4.73045975e-01 6.69594169e-01 -5.20979106e-01 -4.09643114e-01
5.10978997e-01 8.84000659e-01 -3.22169274e-01 4.40166295e-01
1.61426798e-01 5.01284122e-01 -2.23204806e-01 4.25063401e-01
7.26200700e-01 2.05701962e-01 -1.89849794e-01 -6.40488803e-01
-1.44212276e-01 -2.72467196e-01 -1.02590203e+00 1.95757866e+00
-6.28470838e-01 6.85305834e-01 1.54607713e-01 -1.27345800e+00
1.07649136e+00 5.56942105e-01 4.94559824e-01 -1.07625568e+00
6.17242098e-01 1.87395394e-01 -2.07853183e-01 -1.00980294e+00
2.58797973e-01 -1.35271072e-01 2.15212584e-01 -1.11784935e-01
4.91292626e-01 2.93611050e-01 -2.32715607e-02 -1.68967307e-01
8.96857023e-01 2.38715023e-01 4.94010523e-02 -2.71551125e-02
1.11155570e+00 -2.63154596e-01 1.77578509e-01 2.73197919e-01
-2.75333136e-01 4.75812078e-01 2.47580215e-01 -3.50957870e-01
-6.09594703e-01 -6.19377136e-01 -3.46603096e-02 1.10573626e+00
-3.07164853e-03 -1.97484687e-01 -8.73044491e-01 -6.91853464e-01
-4.79321271e-01 6.28411412e-01 -7.68506408e-01 -5.84235013e-01
-2.90139075e-02 -5.09511769e-01 4.61843967e-01 1.92039073e-01
6.69575512e-01 -1.24667847e+00 -9.08702433e-01 9.82929841e-02
-2.66306221e-01 -1.24613428e+00 7.95469433e-02 6.08761143e-03
-6.44123673e-01 -1.07789600e+00 -9.60182130e-01 -6.63061559e-01
3.41095358e-01 -4.49841283e-02 6.31859779e-01 6.09138124e-02
-2.46772438e-01 6.07416630e-01 -6.72279894e-01 -4.59088266e-01
-1.53041795e-01 1.19088404e-01 -9.07952785e-02 6.51926517e-01
4.95633721e-01 -5.13781548e-01 -5.13343394e-01 2.55014807e-01
-8.28232288e-01 -6.64361864e-02 7.36611247e-01 7.59258389e-01
3.11120041e-02 -2.83391535e-01 7.27231562e-01 -3.22857857e-01
6.74697578e-01 -4.59411293e-01 -1.87520742e-01 4.54549611e-01
-4.61328506e-01 -7.07065687e-02 4.08557504e-01 -4.09909666e-01
-1.36030149e+00 3.30399238e-02 -4.44130898e-01 -6.40523076e-01
-4.51029539e-01 2.74969161e-01 -2.70044446e-01 -6.19100966e-02
5.96321464e-01 8.24901611e-02 1.93751737e-01 -1.65570259e-01
1.49070263e-01 8.84491205e-01 1.27516896e-01 -2.40162075e-01
2.43277669e-01 3.07272404e-01 -1.79358736e-01 -9.59120452e-01
-3.78624856e-01 -3.14178407e-01 -4.40745741e-01 -6.45503521e-01
1.23728943e+00 -8.85269821e-01 -8.71227443e-01 5.96427143e-01
-1.40666580e+00 9.35154036e-02 -3.27703282e-02 7.05895305e-01
-4.57324982e-01 2.20354006e-01 -3.40847969e-01 -9.53822672e-01
-3.72385442e-01 -1.25210989e+00 1.02383876e+00 4.87526625e-01
7.30916709e-02 -7.50391543e-01 6.34462312e-02 2.49179587e-01
7.17806518e-01 1.32766232e-01 5.85436404e-01 -7.46033907e-01
-5.48098721e-02 2.28302293e-02 -3.22441101e-01 5.89243114e-01
-2.88708750e-02 3.09461653e-02 -1.44842970e+00 -9.82615575e-02
1.58032566e-01 -3.27522963e-01 8.09521139e-01 1.69498488e-01
1.01833260e+00 -1.15944244e-01 -1.59396172e-01 3.52652192e-01
1.33781803e+00 2.82298207e-01 1.04893470e+00 3.81222785e-01
4.54683006e-01 9.71856952e-01 6.22624695e-01 4.53437120e-01
2.11171746e-01 8.78955901e-01 6.28485858e-01 -4.36002135e-01
7.23246066e-03 -3.42179909e-02 5.26398718e-01 5.91089547e-01
-3.02430332e-01 5.87295294e-02 -5.20694613e-01 4.41389382e-01
-1.92190433e+00 -1.11147308e+00 1.85075309e-02 2.09549260e+00
4.24284041e-01 -1.88386023e-01 1.65453181e-01 1.94996938e-01
8.34486961e-01 -6.57121390e-02 -2.88011760e-01 -8.30669165e-01
-2.43997827e-01 3.60792875e-01 -8.93578283e-04 1.42848060e-01
-1.04129970e+00 8.90194654e-01 5.45616961e+00 8.94752085e-01
-1.72644973e+00 3.47970039e-01 5.53507030e-01 -1.30380113e-02
3.37089635e-02 -4.27035689e-01 -4.85859692e-01 5.74103713e-01
1.40704954e+00 3.36875081e-01 3.19452047e-01 6.95125461e-01
2.80476540e-01 -2.30439693e-01 -7.38801897e-01 1.26360774e+00
4.60833430e-01 -8.91373038e-01 -2.44099408e-01 -2.42777720e-01
2.31200099e-01 -2.22408831e-01 2.77806986e-02 5.30308425e-01
-7.48770773e-01 -9.60737884e-01 6.02306545e-01 9.56945956e-01
4.95489806e-01 -6.20554149e-01 1.03192115e+00 1.74597293e-01
-9.23586786e-01 -1.37911096e-01 -1.66753545e-01 -4.52998988e-02
1.44509524e-01 4.52450693e-01 -7.41532922e-01 6.14929199e-01
1.06833696e+00 6.60574317e-01 -7.03309298e-01 8.98449659e-01
-4.71993610e-02 3.45197588e-01 -2.08021000e-01 -1.81882739e-01
-4.17274283e-03 -2.00782746e-01 3.18509698e-01 1.33295977e+00
6.82762325e-01 -1.28717437e-01 -4.45801675e-01 8.95152509e-01
3.25731754e-01 4.66747552e-01 -6.72620118e-01 6.09655418e-02
-8.58992636e-02 1.69041336e+00 -5.76893270e-01 -2.42486820e-01
-4.84115541e-01 1.13240767e+00 1.93814263e-01 3.64996701e-01
-1.27145386e+00 -5.32728493e-01 4.86654073e-01 -5.97838536e-02
3.58706057e-01 1.12742692e-01 6.57453611e-02 -1.21732950e+00
1.71485394e-02 -6.06011569e-01 9.92420018e-02 -1.13846159e+00
-1.03338623e+00 9.95851576e-01 -5.34458570e-02 -1.12902176e+00
-2.59829104e-01 -7.79673219e-01 -4.71043825e-01 7.23705769e-01
-1.64216125e+00 -1.17657208e+00 -4.58880633e-01 8.40494871e-01
2.21755713e-01 -2.72838831e-01 9.40606654e-01 6.87103033e-01
-6.76442742e-01 6.12276316e-01 -3.08273733e-01 -1.56722248e-01
9.22524810e-01 -7.40195453e-01 -5.43813527e-01 7.98631847e-01
-4.26389240e-02 2.24147171e-01 5.92700481e-01 -2.06946343e-01
-1.15294945e+00 -6.98467553e-01 8.71680140e-01 6.53913543e-02
2.42684305e-01 -2.44511098e-01 -1.00569141e+00 1.72737002e-01
6.96757913e-01 -6.32736832e-02 6.45827472e-01 -1.09105460e-01
-1.30964085e-01 -3.11779767e-01 -1.42877817e+00 3.24929655e-01
7.47090578e-01 -5.27170360e-01 -7.07186639e-01 -1.48030877e-01
4.63671505e-01 -6.40725344e-02 -7.96993315e-01 5.46402991e-01
5.73378205e-01 -1.39973414e+00 7.65237033e-01 -4.06395763e-01
4.58316714e-01 -2.13058844e-01 -8.31752717e-02 -1.41420698e+00
-1.15083754e-01 -7.11246952e-02 1.31064773e-01 1.51596558e+00
3.03900659e-01 -6.41727507e-01 3.48216414e-01 5.12008429e-01
-4.60928343e-02 -6.21983051e-01 -9.57852304e-01 -3.87445331e-01
-4.42102909e-01 -3.28868598e-01 4.05978888e-01 7.89414704e-01
1.01026364e-01 4.64681894e-01 -4.37742740e-01 -7.54382089e-02
-9.46892984e-03 -3.10154349e-01 5.35195887e-01 -1.22263706e+00
4.56126556e-02 -4.86186117e-01 -6.28708303e-01 -2.67664760e-01
4.60274309e-01 -6.72722876e-01 -1.29647642e-01 -1.22931516e+00
-8.91344845e-02 -4.48765159e-02 -5.41125536e-01 6.37700558e-01
6.93642199e-02 3.25875819e-01 2.99141556e-01 -3.17484707e-01
-5.66477656e-01 9.69340503e-01 9.76196051e-01 -6.41405508e-02
-1.96353704e-01 -2.13790819e-01 -6.09725595e-01 5.98438621e-01
6.85751140e-01 -1.90873116e-01 -1.66266367e-01 -1.46679416e-01
2.15526223e-01 -2.23557968e-02 5.09143114e-01 -1.33232343e+00
2.83668667e-01 2.29980797e-01 5.08339226e-01 -2.88883269e-01
6.27751708e-01 -1.40009379e+00 1.49050891e-01 3.01774681e-01
-1.10544495e-01 -4.06842157e-02 5.47169745e-01 2.63904810e-01
-5.39068997e-01 -1.78570986e-01 7.71539092e-01 1.80679277e-01
-1.04101861e+00 -1.31497011e-01 -5.45673251e-01 -4.97967184e-01
1.33297563e+00 -3.70667934e-01 -1.49138242e-01 -3.22014898e-01
-9.53838646e-01 -1.39858797e-01 3.47196609e-02 7.04318225e-01
6.48336112e-01 -1.27172744e+00 -4.78692055e-01 3.53858978e-01
1.48565710e-01 -7.55639791e-01 7.84156621e-01 1.18020427e+00
2.37280401e-04 3.62166792e-01 -8.32928121e-01 -6.15625620e-01
-1.26071942e+00 7.25839913e-01 5.21771967e-01 -2.52882242e-02
-3.39047387e-02 5.26040137e-01 -1.59458682e-01 -2.55335510e-01
2.88863331e-01 -1.82172254e-01 -7.71960855e-01 6.12622917e-01
5.05224764e-01 4.64783579e-01 4.76275116e-01 -9.36451375e-01
-6.14265978e-01 8.14742684e-01 3.54909569e-01 -2.51760364e-01
1.40908635e+00 -2.36621648e-01 -4.62488234e-02 4.13895667e-01
1.35569751e+00 -3.13518047e-01 -8.62056851e-01 -2.88608074e-02
-2.12835744e-01 -2.51956403e-01 1.63719371e-01 -9.91915941e-01
-1.30757499e+00 1.30120564e+00 1.19580472e+00 8.76822174e-02
1.63514471e+00 -2.38681972e-01 4.26010877e-01 6.79774582e-02
2.96124697e-01 -1.22180128e+00 1.58740476e-01 4.80435967e-01
1.03178096e+00 -1.36040223e+00 -5.57734787e-01 -1.29003659e-01
-8.86811554e-01 1.34288800e+00 5.29901564e-01 4.54500376e-04
7.65125632e-01 -1.28985524e-01 -6.55713752e-02 -8.95239711e-02
-5.16262829e-01 -5.20117462e-01 4.93645012e-01 5.33805728e-01
4.10448074e-01 -2.23931193e-01 -5.03443599e-01 1.12613225e+00
8.40079337e-02 4.61640060e-01 2.22527474e-01 5.44786155e-01
-2.39961922e-01 -1.00116122e+00 -4.75699395e-01 1.03046730e-01
-3.72326732e-01 1.29443347e-01 -1.79271951e-01 6.78523302e-01
4.48484689e-01 1.15568614e+00 -1.51879624e-01 -7.93176591e-01
5.84306479e-01 4.81166512e-01 5.14120936e-01 -2.44090855e-01
-8.49702358e-01 -3.74687389e-02 -1.27683520e-01 -8.94044399e-01
-8.17537904e-01 -5.42661369e-01 -1.01543546e+00 2.29048426e-03
-3.45775068e-01 9.14599150e-02 9.85103726e-01 1.15687561e+00
7.11550951e-01 6.91032946e-01 6.77777410e-01 -1.10118139e+00
-4.14185189e-02 -1.07842445e+00 -4.38077003e-01 4.64822829e-01
1.83376521e-01 -9.28602040e-01 -3.39438021e-01 -9.30972397e-02] | [13.352835655212402, 5.053073883056641] |
9abfc183-840a-4992-8d94-38c43e8290e6 | abusive-language-detection-in-online | 1905.07894 | null | https://arxiv.org/abs/1905.07894v1 | https://arxiv.org/pdf/1905.07894v1.pdf | Abusive Language Detection in Online Conversations by Combining Content-and Graph-based Features | In recent years, online social networks have allowed worldwide users to meet and discuss. As guarantors of these communities, the administrators of these platforms must prevent users from adopting inappropriate behaviors. This verification task, mainly done by humans, is more and more difficult due to the ever growing amount of messages to check. Methods have been proposed to automatize this moderation process, mainly by providing approaches based on the textual content of the exchanged messages. Recent work has also shown that characteristics derived from the structure of conversations, in the form of conversational graphs, can help detecting these abusive messages. In this paper, we propose to take advantage of both sources of information by proposing fusion methods integrating content-and graph-based features. Our experiments on raw chat logs show that the content of the messages, but also of their dynamics within a conversation contain partially complementary information, allowing performance improvements on an abusive message classification task with a final F-measure of 93.26%. | ['Georges Linarès', 'Noé Cecillon', 'Vincent Labatut', 'Richard Dufour'] | 2019-05-20 | null | null | null | null | ['abuse-detection'] | ['natural-language-processing'] | [-2.76719462e-02 1.93001151e-01 -1.69121116e-01 -3.77019167e-01
-4.04388338e-01 -6.38063908e-01 9.90681171e-01 6.88328505e-01
-1.91879869e-01 5.96718907e-01 4.28811103e-01 -2.07742646e-01
3.62807475e-02 -7.46787727e-01 1.49027064e-01 -2.67656296e-01
-2.44717777e-01 2.71698982e-01 4.25303698e-01 -4.93943930e-01
4.84756052e-01 5.16216457e-01 -1.25480783e+00 4.02266175e-01
7.41643786e-01 8.11986029e-01 -1.71240911e-01 6.64356768e-01
-4.12593216e-01 9.73727167e-01 -9.25355852e-01 -4.80829388e-01
-1.26032708e-02 -6.37958407e-01 -8.07956159e-01 3.90550733e-01
1.25656247e-01 -2.66976386e-01 -2.41405353e-01 1.01355863e+00
-2.96734646e-02 -1.05153203e-01 3.69315177e-01 -1.50513959e+00
1.01016626e-01 9.05315816e-01 -4.77967292e-01 3.30585361e-01
7.93129027e-01 -5.43498062e-02 1.21502733e+00 -4.32640314e-01
7.74181843e-01 1.09962273e+00 6.05285823e-01 1.93163231e-01
-1.27291167e+00 -6.51457727e-01 -4.26861756e-02 2.05961078e-01
-1.07778418e+00 -4.34367597e-01 9.11334574e-01 -5.80119729e-01
6.72311187e-01 5.86804271e-01 5.48900127e-01 1.05368435e+00
-4.77222428e-02 2.97910571e-01 1.02410352e+00 -5.27363777e-01
-6.24526180e-02 6.15778387e-01 4.50216025e-01 6.70534611e-01
2.67877966e-01 -6.93325281e-01 -4.67240781e-01 -7.40591884e-01
1.13665715e-01 -1.61094833e-02 -3.02822500e-01 3.43589000e-02
-7.22119629e-01 9.94810820e-01 2.80425310e-01 1.02519393e+00
-1.55103356e-01 -4.28491294e-01 5.66579342e-01 6.45127177e-01
7.25536406e-01 4.79811639e-01 9.40087587e-02 -4.37644720e-01
-6.45673692e-01 4.68119122e-02 1.40813088e+00 4.70955610e-01
8.08950245e-01 -4.18096691e-01 1.31846920e-01 7.55348921e-01
2.85125256e-01 1.05398931e-01 3.09350729e-01 -5.72709560e-01
5.84709227e-01 1.32423186e+00 -5.06994836e-02 -1.56343639e+00
-5.50909698e-01 -1.93064705e-01 -7.54457176e-01 -1.38395250e-01
6.38631999e-01 -2.85047024e-01 -1.93900302e-01 1.48305881e+00
4.67232764e-01 -2.23390624e-01 -6.35936797e-01 4.31276977e-01
4.58464712e-01 2.85062164e-01 -2.35693350e-01 -4.29986835e-01
1.20012259e+00 -5.09708107e-01 -8.83237660e-01 1.88591197e-01
6.78073108e-01 -8.15016389e-01 6.54177845e-01 3.90748560e-01
-8.60564709e-01 -1.28907710e-01 -1.09304190e+00 3.08222681e-01
-3.58667672e-01 -5.25211930e-01 2.44463101e-01 9.71461833e-01
-8.56569707e-01 7.39518642e-01 -4.50263828e-01 -4.24078524e-01
2.39731222e-01 2.48720095e-01 -4.80249554e-01 3.29169065e-01
-1.11942852e+00 8.45285237e-01 -7.47836754e-02 -8.80510956e-02
-2.34653860e-01 -1.77997857e-01 -6.32638216e-01 2.08875194e-01
4.89639014e-01 -4.69982997e-03 1.12723434e+00 -1.05972052e+00
-1.22554159e+00 7.38700867e-01 1.00159265e-01 -4.20786500e-01
7.70451546e-01 3.18340883e-02 -5.26765108e-01 3.58376920e-01
-9.17121768e-02 -2.02324480e-01 8.83552730e-01 -9.96190131e-01
-5.95439732e-01 -4.96881276e-01 1.89030945e-01 -2.57548511e-01
-8.15282226e-01 4.82831746e-01 -6.32051528e-02 -2.18506172e-01
-3.80236991e-02 -1.12020671e+00 2.45808333e-04 -1.67319670e-01
-5.57375193e-01 -4.17359173e-01 1.12937677e+00 -8.23506355e-01
1.63441265e+00 -2.03345108e+00 4.28545512e-02 6.60270691e-01
8.29331398e-01 3.32204103e-01 2.22987488e-01 8.53697181e-01
2.76084810e-01 4.97023493e-01 -3.96665558e-02 -6.12062633e-01
-7.03574196e-02 1.07645923e-02 -6.78426996e-02 6.10737920e-01
-1.33010656e-01 3.53070080e-01 -7.73052037e-01 -4.99667913e-01
5.76300872e-03 3.83162826e-01 -2.27973789e-01 2.44202211e-01
3.96834277e-02 5.11503160e-01 -4.37530279e-01 2.15203688e-01
2.96796829e-01 -4.12529409e-01 7.43679643e-01 3.81277442e-01
-3.10246972e-03 7.08613515e-01 -7.75869668e-01 1.00507677e+00
-3.62732232e-01 1.03521752e+00 5.63138962e-01 -7.50330269e-01
8.44798028e-01 4.79078680e-01 4.67541993e-01 -1.94744959e-01
5.48141956e-01 1.13717904e-02 3.13357860e-01 -3.32710505e-01
4.00555074e-01 2.40496382e-01 -1.23152070e-01 8.25263500e-01
-1.94335192e-01 1.49933353e-01 4.66190487e-01 7.34822154e-01
1.37203825e+00 -6.04937136e-01 5.99163830e-01 7.50941876e-03
8.27528656e-01 -1.96855664e-01 2.80903488e-01 6.82487488e-01
-3.75154972e-01 2.06636786e-01 1.05212653e+00 -1.35011733e-01
-8.90592754e-01 -4.22053754e-01 9.90097672e-02 8.67662251e-01
-2.18394920e-02 -8.88560832e-01 -8.82227540e-01 -9.48590815e-01
1.94764793e-01 4.00293082e-01 -6.03142798e-01 -3.53757255e-02
-6.62190676e-01 -4.70727205e-01 4.38985169e-01 -6.45624846e-02
2.53671885e-01 -8.03488612e-01 -2.92338312e-01 4.10583109e-01
-4.52841848e-01 -1.25277317e+00 -3.75804245e-01 -3.10147166e-01
-7.14965403e-01 -1.41707969e+00 -8.48303810e-02 -1.08760871e-01
3.84478211e-01 4.59196717e-01 8.95216167e-01 8.16518843e-01
-2.19218642e-01 2.57751912e-01 -6.63883209e-01 -1.02597661e-01
-1.02171361e+00 1.77466199e-01 -7.75168687e-02 5.67970991e-01
2.12564155e-01 -9.58884120e-01 -2.24162310e-01 5.32461107e-01
-8.34434092e-01 -2.15724513e-01 1.86321586e-01 4.31374609e-01
-4.50758606e-01 -1.62017301e-01 6.85766697e-01 -1.33930552e+00
9.73941624e-01 -8.11457276e-01 -2.85791188e-01 -1.43902982e-02
-6.92932427e-01 -3.22078377e-01 6.54896021e-01 -3.45920146e-01
-8.72989118e-01 -4.40398395e-01 4.76725996e-02 9.61011872e-02
-1.63141400e-01 4.21523184e-01 3.35953757e-02 -1.17751911e-01
6.71114802e-01 -9.57806036e-02 4.85034227e-01 -5.54709673e-01
-2.10345481e-02 1.16084194e+00 -1.24205396e-01 2.02990137e-02
7.24898517e-01 4.60860908e-01 -1.15212686e-01 -1.19654620e+00
-6.56008422e-01 -9.65099335e-01 -4.88765687e-01 -6.02530599e-01
5.06315947e-01 -3.91478717e-01 -9.85306561e-01 4.16233152e-01
-1.11857808e+00 1.23662405e-01 3.96982402e-01 1.63584501e-01
-5.29877171e-02 8.80810201e-01 -8.76672089e-01 -1.08929145e+00
-2.48963609e-01 -6.89143538e-01 4.36436176e-01 -7.24004209e-02
-7.37510860e-01 -1.05219686e+00 2.01370701e-01 7.80425847e-01
6.61776483e-01 3.75610739e-01 7.29529798e-01 -1.19798112e+00
-3.29841614e-01 -6.87461078e-01 -1.21446028e-01 3.02341610e-01
3.80270630e-01 2.29071409e-01 -9.52167273e-01 -1.72557235e-01
1.47328511e-01 7.26148486e-02 4.97637004e-01 -3.20975542e-01
6.45770609e-01 -6.26118720e-01 -3.48874748e-01 -3.20216656e-01
9.01169300e-01 -3.20237763e-02 4.03045654e-01 7.17187822e-02
4.86791849e-01 8.37105393e-01 5.86133189e-02 7.25343049e-01
2.08906129e-01 7.23710954e-01 5.95131338e-01 3.90088916e-01
-9.85793956e-03 -6.32451549e-02 4.44575489e-01 9.31980371e-01
-9.72474832e-03 -3.48546386e-01 -8.63916099e-01 2.83635288e-01
-1.79060280e+00 -1.05017388e+00 -5.12627602e-01 1.95297909e+00
7.14861512e-01 5.66571176e-01 5.83128095e-01 6.07738972e-01
9.37622547e-01 2.29179189e-01 4.06099111e-02 -3.46033365e-01
2.27435425e-01 -1.85790837e-01 2.14084312e-01 6.69144690e-01
-8.39909017e-01 2.16773540e-01 5.61340380e+00 3.52730930e-01
-1.00641382e+00 1.10961378e-01 4.58588809e-01 4.47702594e-02
-6.99625239e-02 5.97627200e-02 -5.68270922e-01 6.09172463e-01
1.12573051e+00 -2.65204579e-01 3.54976654e-01 6.84323609e-01
3.32609832e-01 -2.70340741e-01 -1.01483417e+00 7.01664686e-01
2.69389361e-01 -1.18134594e+00 -3.48705530e-01 2.85638869e-01
3.84137750e-01 -1.02414265e-01 -3.86343122e-01 4.81501222e-02
1.37914360e-01 -6.07642531e-01 3.44575077e-01 3.57815415e-01
7.03421384e-02 -5.35274744e-01 7.78619409e-01 6.32782340e-01
-9.49662507e-01 -1.07133225e-01 3.15245330e-01 -4.35601443e-01
2.15587854e-01 8.47834826e-01 -1.23462844e+00 4.90693629e-01
4.45420504e-01 7.87912190e-01 -5.94951630e-01 9.58942413e-01
-1.99393094e-01 7.69805372e-01 -3.28783333e-01 -4.29359585e-01
6.02459013e-02 -2.60806948e-01 7.35405564e-01 1.21555567e+00
-6.30993024e-02 -1.43064603e-01 2.66931415e-01 5.99083662e-01
-2.29841799e-01 2.36569151e-01 -5.56457698e-01 -3.90341789e-01
3.60290140e-01 1.58696997e+00 -7.91814089e-01 -4.40810621e-01
-4.43687350e-01 5.67423463e-01 4.48490918e-01 -1.22497693e-01
-5.05838454e-01 -3.37692887e-01 6.65497780e-01 5.11810720e-01
-1.36841580e-01 -4.50297445e-01 -1.03090413e-01 -1.00675046e+00
1.15570948e-01 -1.04377306e+00 3.63842309e-01 -1.19046599e-01
-1.41514111e+00 7.43119478e-01 -2.24839598e-01 -9.25809443e-01
-3.41633707e-01 -1.75680101e-01 -7.07529545e-01 5.61532319e-01
-1.13406849e+00 -6.49203479e-01 -4.21197861e-01 3.99584234e-01
2.74283558e-01 6.77681118e-02 5.60335100e-01 2.69065201e-01
-6.03207707e-01 3.68872583e-01 -1.65796518e-01 5.06379426e-01
6.26963615e-01 -1.12249184e+00 6.40411451e-02 5.46007037e-01
2.81468540e-01 6.31282210e-01 8.99228871e-01 -6.85316145e-01
-1.19232440e+00 -5.35144746e-01 1.31726265e+00 -5.72758257e-01
1.10242605e+00 -6.13216341e-01 -1.12052524e+00 4.53741193e-01
3.40666115e-01 -5.26822984e-01 8.25613379e-01 4.58546311e-01
-6.05033159e-01 5.23861200e-02 -1.14312863e+00 4.35634613e-01
8.97748768e-01 -7.29231358e-01 -5.68644226e-01 4.48092312e-01
2.21263036e-01 5.56244068e-02 -6.88954234e-01 -8.43113959e-02
4.52320725e-01 -1.34432459e+00 2.62538165e-01 -5.08704662e-01
2.17765659e-01 5.73327877e-02 2.94271260e-01 -1.07926440e+00
7.05529451e-02 -9.45507109e-01 -8.10370892e-02 1.22138989e+00
4.63510871e-01 -7.38838255e-01 7.04348505e-01 5.93850493e-01
3.34989399e-01 -2.18855098e-01 -8.44882965e-01 -5.78857005e-01
-5.15126050e-01 -2.38116071e-01 2.10770607e-01 1.13578546e+00
8.74641240e-01 6.27710640e-01 -4.70128715e-01 -2.10042760e-01
2.80415952e-01 4.07693274e-02 9.47295606e-01 -1.67067051e+00
-2.28764266e-01 -6.80752635e-01 -6.19099140e-01 -4.51725751e-01
3.55672799e-02 -7.75138855e-01 -2.17781141e-01 -1.21893871e+00
2.54914880e-01 -3.86626720e-01 1.51930347e-01 2.74182677e-01
5.29906228e-02 1.17565691e-01 2.77797222e-01 4.12890673e-01
-6.09568417e-01 1.27675414e-01 7.87206233e-01 -4.61189710e-02
-2.85511285e-01 1.17134564e-01 -4.40486312e-01 7.94848859e-01
9.25394475e-01 -4.37958270e-01 -1.50854185e-01 4.26963419e-01
4.84689981e-01 1.31682038e-01 -3.60865816e-02 -8.66565645e-01
7.04902411e-02 3.27211507e-02 -2.60243446e-01 -1.96894825e-01
4.09602821e-01 -8.81053209e-01 1.16678402e-02 6.17492855e-01
-4.08331156e-01 5.61764240e-02 -2.14834571e-01 8.02749693e-01
-3.77696753e-01 -1.79310232e-01 6.92396045e-01 -1.13990335e-02
9.63021815e-02 -1.38461322e-01 -7.26767898e-01 -1.33073673e-01
1.05239475e+00 7.11807460e-02 -4.31872964e-01 -1.08020079e+00
-8.03674519e-01 9.41563621e-02 3.60946774e-01 5.30609906e-01
2.02192262e-01 -7.27413476e-01 -7.26171851e-01 -9.65366587e-02
-8.31637606e-02 -7.54499018e-01 1.34438239e-02 1.20463800e+00
-3.31805944e-01 1.58918083e-01 -9.38050896e-02 -3.95981491e-01
-1.97129297e+00 3.76877695e-01 1.92499012e-01 -3.94883513e-01
-4.38608438e-01 3.40043277e-01 -4.86894101e-01 4.73831221e-02
2.67734796e-01 -6.15291782e-02 -4.53140497e-01 6.64288521e-01
8.60696733e-01 5.78534901e-01 2.86158085e-01 -7.21648753e-01
-2.72161245e-01 -2.25738585e-01 -3.53161126e-01 -5.35625452e-03
1.19840121e+00 -2.73950875e-01 -5.58680654e-01 5.05911171e-01
1.17751706e+00 5.19976258e-01 -4.73017722e-01 -3.18287432e-01
5.05909503e-01 -7.44476616e-01 -2.24687651e-01 -4.64701295e-01
-7.93764830e-01 6.03992820e-01 -2.09003329e-01 1.40259111e+00
6.29200220e-01 -3.92203331e-02 7.44024158e-01 2.03329712e-01
2.89355069e-01 -8.93071234e-01 1.82768121e-01 3.76788169e-01
7.64806151e-01 -1.33222365e+00 3.95691656e-02 -8.54769170e-01
-3.37098539e-01 1.18042159e+00 2.64536172e-01 1.47039741e-02
8.43903780e-01 -6.02522073e-03 -5.09028416e-03 -1.68271631e-01
-8.06424379e-01 2.43034530e-02 1.08443953e-01 8.77174288e-02
4.73496974e-01 1.66481398e-02 -7.59281218e-01 2.43281469e-01
4.31761630e-02 -4.33965027e-01 1.00955653e+00 8.74807000e-01
-7.05574512e-01 -1.38511407e+00 -3.55674207e-01 5.94697297e-01
-5.76488495e-01 1.92337036e-01 -1.26442325e+00 7.97761202e-01
-3.48038614e-01 1.60138619e+00 -1.14501230e-01 -5.19013643e-01
5.15631251e-02 2.65223593e-01 1.04904264e-01 -6.46161318e-01
-1.22746801e+00 -2.42164612e-01 7.34554350e-01 -5.01074791e-01
-5.22172093e-01 -7.45790780e-01 -9.30693328e-01 -7.94094384e-01
-6.06973469e-01 5.34956634e-01 7.52425671e-01 1.02138412e+00
3.97537827e-01 1.44717649e-01 1.05184138e+00 -5.72038591e-01
-5.77659190e-01 -1.23469710e+00 -5.62609315e-01 8.47914398e-01
4.12791342e-01 -5.51259816e-01 -8.65868092e-01 -2.78499126e-01] | [8.41421127319336, 10.241193771362305] |
30cce135-762b-4a6a-9a05-bcca35d40c03 | attributed-grammars-for-joint-estimation-of | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Park_Attributed_Grammars_for_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Park_Attributed_Grammars_for_ICCV_2015_paper.pdf | Attributed Grammars for Joint Estimation of Human Attributes, Part and Pose | In this paper, we are interested in developing compositional models to explicit representing pose, parts and attributes and tackling the tasks of attribute recognition, pose estimation and part localization jointly. This is different from the recent trend of using CNN-based approaches for training and testing on these tasks separately with a large amount of data. Conventional attribute models typically use a large number of region-based attribute classifiers on parts of pre-trained pose estimator without explicitly detecting the object or its parts, or considering the correlations between attributes. In contrast, our approach jointly represents both the object parts and their semantic attributes within a unified compositional hierarchy. We apply our attributed grammar model to the task of human parsing by simultaneously performing part localization and attribute recognition. We show our modeling helps performance improvements on pose-estimation task and also outperforms on other existing methods on attribute prediction task. | ['Se-Young Park', 'Song-Chun Zhu'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['human-parsing'] | ['computer-vision'] | [ 4.00655001e-01 4.63274658e-01 -2.08807006e-01 -8.74646604e-01
-1.01480675e+00 -5.37156343e-01 5.81770420e-01 3.10564220e-01
-1.53378397e-01 2.85306364e-01 1.40433922e-01 1.85923427e-01
1.25412598e-01 -5.61990440e-01 -1.04307437e+00 -3.34154755e-01
-7.20092505e-02 1.47552907e+00 3.20424408e-01 9.32369083e-02
1.32651543e-02 4.53025550e-01 -1.86537004e+00 2.53552973e-01
5.77440321e-01 1.42215800e+00 2.08700016e-01 7.25494325e-01
-2.20161051e-01 5.89237094e-01 -4.44605500e-01 -6.27467573e-01
1.07368268e-01 -1.75425902e-01 -1.00382733e+00 3.87835085e-01
1.14301074e+00 -3.72077495e-01 3.16783756e-01 6.69576764e-01
2.26555496e-01 -1.19152769e-01 7.77858555e-01 -1.31073225e+00
-2.95181364e-01 4.95818198e-01 -4.51141715e-01 -4.88948703e-01
5.16285837e-01 -4.12214637e-01 1.33180773e+00 -7.42477775e-01
6.24172270e-01 1.44654453e+00 8.00293446e-01 5.51819384e-01
-1.38066256e+00 -3.97139966e-01 7.04681516e-01 7.66865611e-02
-1.33330047e+00 -2.68712014e-01 7.78921545e-01 -5.13844132e-01
9.43730712e-01 1.70029730e-01 6.43516421e-01 1.10058129e+00
-3.37226927e-01 1.15557837e+00 8.32242548e-01 -4.47302818e-01
6.61615133e-02 -6.29340261e-02 1.14995994e-01 7.91548133e-01
1.32789791e-01 -2.32398763e-01 -5.23746252e-01 -1.70949265e-01
6.24699712e-01 -1.81643218e-01 3.18237633e-01 -9.84638393e-01
-1.25989783e+00 6.15685284e-01 2.39166334e-01 -3.89382094e-01
-3.65418434e-01 4.49649155e-01 3.74404639e-01 -2.16557875e-01
5.43840110e-01 3.01417887e-01 -1.16484225e+00 2.13264108e-01
-7.15156257e-01 4.01127756e-01 8.63920152e-01 1.49947357e+00
8.36159885e-01 -4.46499795e-01 2.35980563e-02 8.86611044e-01
3.88392329e-01 3.59245718e-01 -4.46766466e-02 -1.11842227e+00
4.11784708e-01 8.34285915e-01 -1.09917492e-01 -3.60752374e-01
-4.95691568e-01 -4.60757822e-01 -1.13859162e-01 1.24592356e-01
5.41884243e-01 2.65638083e-01 -1.18643129e+00 1.90141594e+00
6.92978561e-01 -1.63174570e-01 -3.51986617e-01 7.42468596e-01
6.82359934e-01 3.21955644e-02 7.05049038e-01 2.87252069e-01
1.78152144e+00 -9.89660978e-01 -4.91815239e-01 -6.21038198e-01
6.80028319e-01 -7.67590880e-01 8.93324196e-01 3.62000823e-01
-9.91311908e-01 -7.02235043e-01 -9.72360015e-01 -4.80614394e-01
-4.34530318e-01 5.46490908e-01 1.01761699e+00 5.35104692e-01
-6.72006667e-01 5.79215705e-01 -7.94876218e-01 -3.99629772e-01
6.34734094e-01 7.46282101e-01 -4.99989569e-01 -8.39330629e-02
-6.57265723e-01 1.01528490e+00 1.88264474e-01 -9.95987579e-02
-9.06966686e-01 -4.28766370e-01 -1.23511755e+00 -1.84995040e-01
5.24600625e-01 -9.86649215e-01 1.39603114e+00 -9.80048835e-01
-1.00718319e+00 1.12630522e+00 -2.37984702e-01 -5.88722005e-02
3.84054512e-01 -8.59110296e-01 7.65246525e-02 -2.11060479e-01
2.20245630e-01 1.15959144e+00 7.11291730e-01 -1.26895881e+00
-7.43147075e-01 -9.66276467e-01 1.06536508e-01 3.20914775e-01
7.23762512e-02 -6.30143881e-02 -5.88757932e-01 -4.78304088e-01
7.34529555e-01 -1.03510237e+00 -1.93067491e-01 2.02950746e-01
-6.97566152e-01 -3.92022759e-01 7.47754812e-01 -5.55977881e-01
5.91304302e-01 -1.99367893e+00 2.20022127e-01 -6.82709962e-02
1.29705712e-01 -3.37519258e-01 -1.58692732e-01 -5.06815836e-02
1.08953170e-01 -9.04737487e-02 -2.55935222e-01 -7.00790465e-01
2.36766905e-01 3.53674352e-01 -7.16518462e-02 3.56993675e-01
3.96708459e-01 9.10642982e-01 -6.65618956e-01 -9.68749166e-01
1.34412661e-01 5.65253198e-01 -5.04735589e-01 2.10383669e-01
-6.18486702e-01 4.61047471e-01 -6.85757995e-01 1.21139288e+00
6.81612372e-01 -9.95114595e-02 3.18241239e-01 -3.29581231e-01
4.45956171e-01 6.02237165e-01 -1.23869419e+00 1.82013118e+00
-4.77717459e-01 2.10193723e-01 1.65259704e-01 -6.60819888e-01
8.07190120e-01 1.76450774e-01 6.66175008e-01 -2.89237648e-01
1.12529002e-01 1.98985577e-01 -1.50224492e-01 -1.06680200e-01
2.36972734e-01 -2.28482895e-02 -4.75395739e-01 2.59153932e-01
3.33968759e-01 -7.12929815e-02 -4.83516790e-02 9.71055217e-03
8.62740219e-01 1.14514220e+00 4.08172965e-01 -1.05357818e-01
4.17095125e-01 2.97498435e-01 5.40764689e-01 4.60190445e-01
-1.63034111e-01 8.58089447e-01 4.23659295e-01 -6.49033248e-01
-1.14686823e+00 -1.07475376e+00 -2.62768805e-01 1.75638342e+00
1.71511963e-01 -4.81609195e-01 -9.01590586e-01 -1.07672644e+00
2.41604030e-01 6.02827728e-01 -8.04313362e-01 2.86686927e-01
-7.11253941e-01 -4.08961833e-01 3.08723062e-01 1.15027237e+00
1.96478754e-01 -1.11005366e+00 -4.49421078e-01 2.05541894e-01
-2.30349392e-01 -1.40195203e+00 -3.95222038e-01 5.90932965e-01
-9.56793904e-01 -1.10041308e+00 -1.61819756e-01 -9.01728511e-01
6.90863967e-01 -3.37385416e-01 1.66646743e+00 1.15261711e-01
-4.14364338e-01 4.61836785e-01 -2.46429399e-01 -7.77811289e-01
-1.18086062e-01 3.94258767e-01 -1.62248164e-01 -1.61873043e-01
4.05493379e-01 -4.70902830e-01 -4.40284371e-01 5.10715663e-01
-3.18594545e-01 1.21318795e-01 1.03490078e+00 3.84378284e-01
9.70628083e-01 -7.81084597e-01 1.94634587e-01 -1.20833993e+00
-1.83058411e-01 5.12473471e-02 -3.57934266e-01 4.15411085e-01
-4.08863187e-01 3.35912615e-01 2.08126277e-01 -5.39107740e-01
-9.66699898e-01 1.01738060e+00 -1.60154060e-01 -1.83267519e-01
-6.58823907e-01 -2.14215547e-01 -8.38546395e-01 2.04695076e-01
2.53449798e-01 5.89337200e-03 5.85538186e-02 -6.89627290e-01
4.17749166e-01 2.74291068e-01 5.03459752e-01 -9.05850768e-01
5.33138692e-01 4.09370095e-01 4.01185244e-01 -3.27898204e-01
-1.19155228e+00 -5.55495739e-01 -1.32137525e+00 7.58168027e-02
9.98368680e-01 -1.01984537e+00 -8.15869749e-01 2.78028756e-01
-1.27698743e+00 -4.76170257e-02 -2.68787533e-01 3.97309422e-01
-9.18050349e-01 1.84744462e-01 -5.95193148e-01 -7.34999120e-01
-9.11995918e-02 -1.14589763e+00 1.87211311e+00 -2.85054773e-01
-5.36698222e-01 -5.91978312e-01 -1.05032928e-01 6.31995022e-01
1.55999418e-02 1.97867990e-01 9.18030798e-01 -1.01463950e+00
-7.91011274e-01 -5.07290900e-01 -5.73332980e-02 5.48957400e-02
-7.16384724e-02 -3.12217116e-01 -1.15218282e+00 2.09675178e-01
-3.13573837e-01 -5.41134417e-01 6.66704297e-01 3.95724744e-01
1.28230512e+00 2.80961003e-02 -6.79955721e-01 6.40933752e-01
1.11053240e+00 -1.15875766e-01 4.01319772e-01 3.01775783e-01
1.00282085e+00 1.06402564e+00 9.76989210e-01 2.24930778e-01
6.42525852e-01 9.34647083e-01 7.34840214e-01 2.11292449e-02
-2.34051317e-01 -3.65235358e-01 2.57801898e-02 6.80273697e-02
-2.22035408e-01 -1.09637879e-01 -7.19970345e-01 3.35259080e-01
-1.76078951e+00 -4.84044641e-01 -3.33765745e-01 2.13081002e+00
5.49885631e-01 2.55212002e-02 3.79953325e-01 -2.62159765e-01
7.63588786e-01 -2.75267929e-01 -4.62748140e-01 -1.69639245e-01
1.04473330e-01 1.18704349e-01 5.41962326e-01 3.30218256e-01
-1.66674232e+00 1.06447065e+00 6.90677738e+00 5.05929828e-01
-3.78255129e-01 -6.46469742e-02 7.51402020e-01 2.60655140e-03
-4.86304611e-02 2.04970837e-01 -1.15214312e+00 -5.37279819e-04
6.62536263e-01 8.65828991e-01 8.94405320e-02 1.27713442e+00
-6.16178870e-01 -2.03447461e-01 -1.70737493e+00 5.92607558e-01
1.78720802e-01 -6.83666170e-01 1.17311083e-01 1.62650928e-01
1.79863885e-01 -3.82958770e-01 -1.76143169e-01 3.45010251e-01
3.63225937e-01 -1.06536245e+00 1.04411340e+00 2.83318847e-01
6.36290312e-01 -6.42365992e-01 7.21714139e-01 2.48619989e-01
-1.42745411e+00 6.74467236e-02 -1.92050368e-01 3.73075828e-02
-3.43868956e-02 9.07568187e-02 -8.35519493e-01 3.43389988e-01
6.39088154e-01 3.15802783e-01 -8.10984254e-01 7.17737854e-01
-4.81602132e-01 3.52832675e-01 -3.11039269e-01 2.04179168e-01
-2.27126598e-01 -4.58596012e-04 2.34425083e-01 1.03173399e+00
6.89237192e-02 -8.06165412e-02 5.28738439e-01 6.98417306e-01
5.98441586e-02 7.93019012e-02 -1.82938799e-01 2.45695442e-01
4.63026196e-01 1.39083147e+00 -8.63338292e-01 -3.96464884e-01
-4.89468127e-01 9.22161818e-01 5.17540514e-01 -1.45583734e-01
-9.18150425e-01 4.09178138e-02 6.71454012e-01 3.86332661e-01
6.66693866e-01 -1.72481105e-01 -6.08016610e-01 -6.23337448e-01
1.20872237e-01 -6.31954193e-01 5.05057454e-01 -7.23800361e-01
-1.15489173e+00 6.21773899e-01 1.74863279e-01 -1.15707994e+00
-3.33413422e-01 -9.18562651e-01 -1.57511309e-01 7.03653216e-01
-1.17066717e+00 -2.09116054e+00 -2.17944354e-01 1.89267620e-01
6.18053734e-01 1.13000363e-01 1.05217028e+00 4.32244092e-02
-3.01726878e-01 5.24143934e-01 -5.87154806e-01 2.98763424e-01
7.20647335e-01 -1.52892709e+00 5.75361013e-01 2.82610595e-01
3.75320584e-01 5.34411073e-01 8.99124503e-01 -8.48295927e-01
-1.18272686e+00 -8.59866083e-01 1.06380796e+00 -1.18733668e+00
2.39732161e-01 -9.30541873e-01 -5.80108345e-01 1.05012524e+00
-1.84945971e-01 2.51107275e-01 5.13682365e-01 7.00056970e-01
-6.89464569e-01 6.67894334e-02 -1.00848758e+00 3.71392146e-02
1.56359863e+00 -3.45990866e-01 -4.08967704e-01 3.24489236e-01
5.53286731e-01 -4.03859079e-01 -8.81428480e-01 7.30531752e-01
8.12482953e-01 -7.24879682e-01 1.31441009e+00 -6.72699749e-01
3.81168246e-01 -1.78384706e-01 -5.10756850e-01 -8.06802511e-01
-3.03645521e-01 1.58058658e-01 -2.24378884e-01 1.23647094e+00
5.55077553e-01 -7.20335543e-02 1.28202212e+00 6.62247777e-01
-3.20611715e-01 -9.10553694e-01 -6.39854550e-01 -7.12983906e-01
-2.19191059e-01 -5.00978529e-01 6.73680305e-01 5.03513873e-01
-5.05872846e-01 5.69852173e-01 -1.04474738e-01 3.15269113e-01
8.81469548e-01 3.46365631e-01 7.93470740e-01 -1.58295512e+00
-4.02329743e-01 -2.31719568e-01 -6.81453884e-01 -9.57143009e-01
3.90788883e-01 -6.42190158e-01 4.38595295e-01 -1.54133737e+00
4.26855773e-01 -4.60574210e-01 1.97341871e-02 7.52463818e-01
-2.57530779e-01 2.88468093e-01 -9.96007994e-02 2.32196972e-01
-9.21646416e-01 3.00529420e-01 1.14699173e+00 -1.79158345e-01
2.36762926e-01 1.70849368e-01 -5.88918447e-01 9.47129667e-01
4.81414706e-01 -5.88017881e-01 -3.33406121e-01 -3.34261060e-01
2.84119155e-02 -1.03164583e-01 5.52404821e-01 -1.02900088e+00
-5.64779937e-02 5.21899387e-02 9.66228247e-01 -9.09515023e-01
7.77285039e-01 -9.25015211e-01 -4.10216802e-04 2.55363137e-01
-3.25957537e-01 1.23911932e-01 1.12404726e-01 6.41798854e-01
-1.54560089e-01 -1.11384779e-01 4.08500433e-01 -4.11426693e-01
-9.48162735e-01 3.23657095e-01 3.07234284e-02 -1.21739857e-01
1.05445051e+00 -4.92333680e-01 3.99703793e-02 -2.30638787e-01
-1.37449443e+00 4.01707664e-02 5.74932814e-01 2.82449663e-01
5.08401811e-01 -1.18334639e+00 -5.33785284e-01 1.24585278e-01
4.28811193e-01 2.44794950e-01 1.96340936e-03 6.93236947e-01
-3.45785320e-01 1.89568624e-01 -2.61878192e-01 -8.56287837e-01
-1.55226564e+00 6.82101727e-01 1.51623115e-01 -1.29821226e-01
-3.03719521e-01 1.15639198e+00 5.29117286e-01 -9.52271581e-01
4.58479851e-01 2.82581765e-02 -1.14560686e-01 7.20531270e-02
-6.77226335e-02 2.31207237e-01 1.87235221e-01 -8.84671390e-01
-6.93820477e-01 7.56429732e-01 -1.89134479e-01 4.07763273e-02
1.20012867e+00 -2.63080765e-02 -1.89146012e-01 2.87887424e-01
1.07095361e+00 4.99666668e-02 -1.42013061e+00 -1.63824096e-01
2.85897195e-01 -1.13982432e-01 -2.91554719e-01 -9.24312770e-01
-8.18715453e-01 8.17665040e-01 4.47718680e-01 -3.97134542e-01
8.94246817e-01 7.77366698e-01 5.38847446e-01 3.32993716e-01
5.57812274e-01 -1.20445275e+00 1.64018005e-01 3.32136661e-01
6.68184280e-01 -1.20217705e+00 3.76117408e-01 -1.08604562e+00
-8.95205140e-01 9.41455722e-01 1.10644364e+00 -5.16581573e-02
3.84713203e-01 5.31784356e-01 -3.78046930e-02 -3.36087227e-01
-6.68289065e-01 -5.06266654e-01 8.02115738e-01 7.01022744e-01
8.16354036e-01 2.88395524e-01 1.69225499e-01 5.51301301e-01
-3.31363618e-01 -6.46342218e-01 -1.04065120e-01 1.12992358e+00
-6.02844536e-01 -1.56642723e+00 -3.85443211e-01 4.92350280e-01
-4.59024191e-01 -3.87760065e-02 -8.42539787e-01 1.06907225e+00
4.97994691e-01 4.29105490e-01 2.74273604e-01 -4.51406211e-01
3.07802826e-01 4.98516291e-01 9.10459459e-01 -8.89155388e-01
-4.42174971e-01 2.59124309e-01 4.25706953e-01 -6.54222548e-01
-3.84454131e-01 -1.13136160e+00 -1.22070253e+00 3.14125419e-01
-3.59999269e-01 -2.58484423e-01 8.42906773e-01 1.08253372e+00
2.00968593e-01 4.33452189e-01 8.48376378e-02 -9.03715253e-01
-5.47275126e-01 -1.09490323e+00 -4.73581403e-01 6.85798347e-01
6.99895918e-02 -9.76627886e-01 5.21110334e-02 1.88585185e-02] | [8.227133750915527, -0.21894672513008118] |
0f0daff5-8434-448d-a00f-77e1b665555e | interval-load-forecasting-for-individual | 2306.03010 | null | https://arxiv.org/abs/2306.03010v1 | https://arxiv.org/pdf/2306.03010v1.pdf | Interval Load Forecasting for Individual Households in the Presence of Electric Vehicle Charging | The transition to Electric Vehicles (EV) in place of traditional internal combustion engines is increasing societal demand for electricity. The ability to integrate the additional demand from EV charging into forecasting electricity demand is critical for maintaining the reliability of electricity generation and distribution. Load forecasting studies typically exclude households with home EV charging, focusing on offices, schools, and public charging stations. Moreover, they provide point forecasts which do not offer information about prediction uncertainty. Consequently, this paper proposes the Long Short-Term Memory Bayesian Neural Networks (LSTM-BNNs) for household load forecasting in presence of EV charging. The approach takes advantage of the LSTM model to capture the time dependencies and uses the dropout layer with Bayesian inference to generate prediction intervals. Results show that the proposed LSTM-BNNs achieve accuracy similar to point forecasts with the advantage of prediction intervals. Moreover, the impact of lockdowns related to the COVID-19 pandemic on the load forecasting model is examined, and the analysis shows that there is no major change in the model performance as, for the considered households, the randomness of the EV charging outweighs the change due to pandemic. | ['Syed Mir', 'Katarina Grolinger', 'Mohamed Ahmed T. A. Elgalhud', 'Raiden Skala'] | 2023-06-05 | null | null | null | null | ['bayesian-inference', 'load-forecasting', 'prediction-intervals'] | ['methodology', 'miscellaneous', 'miscellaneous'] | [-2.57708848e-01 8.01298991e-02 -2.33811930e-01 -3.91334265e-01
-3.69908988e-01 -2.54484385e-01 8.77966881e-01 2.24061206e-01
-3.07750732e-01 9.16936398e-01 1.57808349e-01 -6.63968384e-01
-4.68538314e-01 -1.10176349e+00 -7.30134130e-01 -9.03787792e-01
-8.47504586e-02 8.66474867e-01 -2.41601244e-01 -7.96799455e-03
-1.75616458e-01 4.73832279e-01 -1.46351290e+00 2.15625595e-02
9.90225255e-01 1.08280587e+00 5.09833038e-01 2.29310483e-01
-2.26573586e-01 4.08595592e-01 -6.95504010e-01 4.38703448e-02
-2.43494362e-01 3.10705572e-01 -3.95750821e-01 -5.66207886e-01
-6.49997175e-01 -8.20683062e-01 -1.93282723e-01 8.13442707e-01
5.99832237e-01 2.73395300e-01 7.97763109e-01 -1.41086566e+00
-2.34081209e-01 1.12046933e+00 5.35406768e-02 6.25954494e-02
-1.13386385e-01 2.68411003e-02 3.95531267e-01 -4.02551711e-01
-2.52048783e-02 9.39868033e-01 7.71462798e-01 5.63827902e-02
-1.17007983e+00 -6.34732723e-01 8.00379366e-02 6.31517351e-01
-1.36664760e+00 -1.34444758e-01 8.26126635e-01 -6.42510474e-01
1.44306362e+00 1.27840266e-01 7.68618584e-01 1.20055532e+00
5.57344913e-01 3.20465863e-01 9.13064301e-01 -1.77574262e-01
5.63922286e-01 4.27821070e-01 3.36931407e-01 -5.33432722e-01
2.92381257e-01 3.98742199e-01 2.06722260e-01 -1.89259380e-01
-2.71837205e-01 2.12013274e-01 5.92138357e-02 4.03273731e-01
-4.49496388e-01 7.02904463e-01 6.03445694e-02 8.41753721e-01
-8.43228102e-01 1.07531980e-01 5.72354496e-01 -7.92044625e-02
7.88059175e-01 -3.40335608e-01 -4.59052682e-01 -1.37057766e-01
-1.25796604e+00 -2.39016488e-02 7.11406171e-01 7.35991657e-01
3.34804744e-01 6.88895583e-01 -3.11871767e-01 2.92750180e-01
2.37388358e-01 1.02487600e+00 4.50017959e-01 -4.73070383e-01
2.18467191e-01 5.38091250e-02 4.84848350e-01 -8.79802585e-01
-6.78255558e-01 -6.08201265e-01 -1.16613913e+00 -1.95690289e-01
-2.04205155e-01 -4.44922924e-01 -5.79058409e-01 1.64847112e+00
-1.92363888e-01 1.61361098e-01 1.20169170e-01 6.61887601e-02
5.84765255e-01 9.79742885e-01 3.78002882e-01 -2.55979776e-01
8.91894817e-01 -2.60489345e-01 -1.21522224e+00 2.02992395e-01
6.19070888e-01 -2.22775981e-01 1.81133240e-01 2.76146203e-01
-9.71089482e-01 -6.96606755e-01 -8.74011099e-01 4.68210727e-01
-9.02171016e-01 -1.08296648e-01 1.41982600e-01 6.56223297e-01
-1.06156301e+00 8.39619339e-01 -8.20423722e-01 -8.85967463e-02
1.57217253e-02 3.91539633e-01 3.02297771e-01 2.59317100e-01
-1.75869834e+00 1.29838324e+00 6.94674015e-01 6.44810796e-01
-7.24323034e-01 -7.79563785e-01 -6.75136745e-01 4.53884691e-01
-2.65590608e-01 -3.43597203e-01 1.20346737e+00 -5.79016864e-01
-1.28500783e+00 -2.53209203e-01 -3.90105456e-01 -8.67982507e-01
6.54559374e-01 3.55153114e-01 -7.28034019e-01 -2.98728138e-01
-2.76860684e-01 4.05758381e-01 3.85731995e-01 -9.96235967e-01
-6.95165217e-01 -2.49906138e-01 -5.26130617e-01 -3.31108600e-01
-3.15051675e-01 -4.36122268e-01 6.78767741e-01 -2.22735018e-01
-6.61034167e-01 -7.09051311e-01 1.25029773e-01 -1.32092571e+00
-3.78292650e-01 -8.21734071e-01 1.00981116e+00 -1.19612014e+00
1.28212011e+00 -1.90407419e+00 -5.18765509e-01 4.82526004e-01
-4.30059344e-01 -8.63102265e-03 2.36419559e-01 8.09332311e-01
-2.08740771e-01 5.21298870e-02 1.47279233e-01 -4.56866056e-01
5.39050102e-01 5.52098095e-01 -6.31992221e-01 4.08951432e-01
-3.46241236e-01 9.47177827e-01 -6.18934691e-01 9.92137343e-02
7.74204493e-01 8.73399019e-01 1.39810801e-01 -9.67015252e-02
-3.42291355e-01 3.54150951e-01 -1.70623273e-01 -4.76235673e-02
1.00981939e+00 1.91912483e-02 2.93499500e-01 -1.30609959e-01
-5.51426828e-01 2.62969583e-01 -6.89524055e-01 1.03029120e+00
-6.56178117e-01 6.91543698e-01 -3.43321651e-01 -1.28850055e+00
8.49683106e-01 7.11942971e-01 5.75321436e-01 -9.71961021e-01
3.10249001e-01 2.87628293e-01 4.67838310e-02 -2.02826455e-01
3.05989981e-01 -3.31533924e-02 2.27952436e-01 2.48530000e-01
-8.77386555e-02 1.44953415e-01 1.28465071e-01 -3.93966138e-01
4.16026264e-01 -1.82246432e-01 -4.01432425e-01 -5.58772564e-01
2.83732295e-01 -3.42574447e-01 4.56267118e-01 5.61763644e-01
4.71898735e-01 -3.49653542e-01 1.89973637e-01 -3.04362118e-01
-1.05308926e+00 -8.94662440e-01 -5.31661689e-01 5.95070541e-01
-4.02482063e-01 2.67514199e-01 -6.72066450e-01 1.14422366e-01
9.60700586e-02 1.95542705e+00 -4.54754233e-01 -1.60834804e-01
-2.41781130e-01 -9.32156324e-01 2.66109377e-01 6.61712408e-01
3.13319713e-01 -8.87530684e-01 -7.52089798e-01 5.44147551e-01
-2.97414750e-01 -1.00847757e+00 3.10569823e-01 2.92869925e-01
-8.21904480e-01 -5.50839245e-01 -7.99131334e-01 -3.20556432e-01
3.44303340e-01 -3.03505510e-01 9.34880793e-01 -4.44416702e-01
2.55847156e-01 3.67128760e-01 1.62803844e-01 -7.84695804e-01
-5.52024484e-01 3.71119797e-01 5.79166189e-02 -3.54747772e-01
6.70247912e-01 -6.88486516e-01 -4.91912872e-01 9.53509123e-04
-7.57943511e-01 7.35415742e-02 5.73257059e-02 5.34504652e-01
1.07746355e-01 7.20022976e-01 1.31432140e+00 -3.76073480e-01
6.75219774e-01 -8.22257042e-01 -9.56513405e-01 1.11590773e-01
-1.25678992e+00 -1.28054857e-01 7.06827462e-01 -1.57591417e-01
-1.19204056e+00 -3.95137966e-01 -4.88873452e-01 -2.06854403e-01
-4.22637165e-01 5.06235003e-01 -8.57945681e-02 6.31587327e-01
-4.05550808e-01 4.00215328e-01 -3.68735671e-01 -5.91521680e-01
-2.46865172e-02 7.21144497e-01 1.70016199e-01 -8.40680450e-02
4.19093162e-01 2.96941787e-01 5.38397208e-02 -7.96583891e-01
-1.64037034e-01 3.23282517e-02 -3.81501049e-01 -4.69940573e-01
9.60736513e-01 -8.04864645e-01 -1.21941900e+00 6.46374106e-01
-1.37863731e+00 -4.43173528e-01 -2.58810222e-01 6.15535915e-01
-4.08176959e-01 -1.12980939e-01 -5.43399990e-01 -1.30952215e+00
-4.22343582e-01 -1.16570485e+00 3.38320762e-01 1.45267218e-01
-1.41007543e-01 -1.55803025e+00 1.98271708e-04 -7.48767853e-02
9.84540820e-01 2.41654485e-01 1.44855964e+00 -9.49077427e-01
-1.17356487e-01 -3.21457088e-01 9.13765207e-02 5.68895161e-01
5.75940348e-02 -1.57284483e-01 -1.21320868e+00 -5.07717371e-01
3.64211425e-02 4.25979227e-01 4.36383158e-01 8.06288242e-01
1.12436903e+00 -5.72260857e-01 -3.98917317e-01 7.54441246e-02
1.43791115e+00 7.50903964e-01 5.77675283e-01 1.42867714e-01
1.77605540e-01 7.59609938e-01 6.07605241e-02 6.57585859e-01
7.82479882e-01 3.10027480e-01 5.03795147e-01 -1.30307451e-01
5.41272938e-01 -2.75970191e-01 2.05859035e-01 8.67720127e-01
1.07890770e-01 -5.08073449e-01 -8.92479718e-01 7.45063663e-01
-1.59228969e+00 -1.10595691e+00 -2.22721338e-01 2.38685346e+00
1.95041850e-01 3.64970714e-01 5.64772263e-02 3.37272316e-01
7.71588624e-01 -1.48448823e-02 -4.79630053e-01 -8.35350692e-01
-9.00619105e-02 -2.87624467e-02 8.71254385e-01 5.63796043e-01
-4.34831381e-01 7.94819817e-02 6.08646536e+00 8.62090290e-01
-1.11485648e+00 3.40769678e-01 8.52107525e-01 2.16886669e-01
-7.01781690e-01 -3.52603078e-01 -7.94599712e-01 1.16193295e+00
1.78162050e+00 -3.14982802e-01 3.89491498e-01 4.58260238e-01
1.08482218e+00 -1.80496156e-01 -9.66080248e-01 5.51793396e-01
-3.52732807e-01 -9.88318920e-01 -1.18975885e-01 3.05400729e-01
6.40297472e-01 3.58910173e-01 1.73632100e-01 3.74475777e-01
2.08795577e-01 -1.12669504e+00 6.51621819e-01 1.10170197e+00
4.09386605e-01 -1.34005201e+00 1.36657584e+00 6.69555902e-01
-9.57602501e-01 -3.53922933e-01 -1.91660047e-01 -1.10812150e-02
6.64283097e-01 8.43958497e-01 -9.52231050e-01 6.28111064e-01
7.02566087e-01 2.90012479e-01 -2.81866603e-02 6.16151094e-01
1.82018310e-01 8.57672989e-01 -5.82146645e-01 -1.01960115e-02
2.46521994e-01 -3.74241352e-01 3.06767106e-01 1.00176442e+00
7.81311691e-01 -3.52764785e-01 -9.45789441e-02 1.04186392e+00
2.58546084e-01 -3.23292375e-01 -7.17113435e-01 4.72942889e-02
5.73306859e-01 7.30663776e-01 -3.14633280e-01 -4.75355119e-01
-3.48916292e-01 4.00728852e-01 -5.66037893e-01 7.13049412e-01
-9.17734981e-01 -3.55641544e-01 1.86810240e-01 3.23870063e-01
5.07085621e-01 7.97634125e-02 -1.31340012e-01 -3.14415902e-01
-2.98630625e-01 -6.51246235e-02 7.97476172e-02 -8.43693256e-01
-1.11972380e+00 3.83312821e-01 4.70107228e-01 -6.29458785e-01
-1.07450461e+00 -2.49647126e-01 -8.50416362e-01 1.12827623e+00
-1.74653316e+00 -1.05160391e+00 1.21250898e-01 2.56321520e-01
4.04897809e-01 2.23576292e-01 7.62920797e-01 5.77970922e-01
-4.50915247e-01 2.10164756e-01 9.46660519e-01 -1.46328345e-01
2.13338640e-02 -9.80518162e-01 1.96307451e-01 4.08454031e-01
-7.97946990e-01 6.61632195e-02 9.48234975e-01 -7.57262707e-01
-1.01816380e+00 -1.12716424e+00 1.22970176e+00 4.22273763e-02
4.46966946e-01 -3.49707037e-01 -8.71593118e-01 7.04518795e-01
7.70050466e-01 -7.17304587e-01 5.11865616e-01 -2.42605582e-01
3.40297252e-01 -3.68088692e-01 -1.26966190e+00 -5.43345846e-02
8.50688219e-02 -4.05431569e-01 -4.28723574e-01 8.27162564e-02
5.24080336e-01 3.38529617e-01 -1.18960023e+00 6.61739588e-01
5.41819990e-01 -7.67612934e-01 6.73642755e-01 -1.83102056e-01
-2.72602230e-01 1.67041883e-01 -1.91224784e-01 -1.49797249e+00
-4.17081803e-01 -1.77231878e-01 -2.11791828e-01 1.61264277e+00
4.15906280e-01 -1.21369267e+00 2.13239208e-01 9.78012800e-01
-6.28003627e-02 -3.51473153e-01 -1.37652326e+00 -9.03588951e-01
4.27726001e-01 -6.08421445e-01 1.26883817e+00 6.79284632e-01
-1.33829623e-01 -9.30233821e-02 -3.40037018e-01 1.68340728e-01
7.12418616e-01 -1.86829288e-02 5.70523888e-02 -1.43817377e+00
5.52044325e-02 -4.35290754e-01 3.76256835e-03 -4.05272245e-01
5.42077482e-01 -6.04174137e-01 -1.07793607e-01 -1.88238370e+00
1.23987243e-01 -1.80266961e-01 -6.10651493e-01 2.09280521e-01
5.64435482e-01 -3.45367998e-01 1.20000690e-01 -2.76012093e-01
1.10392019e-01 8.65854204e-01 2.68254042e-01 -3.35672349e-01
7.74075463e-02 4.83973414e-01 1.23070896e-01 7.01356053e-01
1.07473719e+00 -5.33698320e-01 -5.21518230e-01 -2.83162713e-01
4.18464005e-01 1.70973927e-01 3.76753688e-01 -9.74539280e-01
2.29430825e-01 1.60440996e-01 5.54433584e-01 -1.45794773e+00
1.38819888e-01 -1.25478542e+00 8.31820309e-01 8.50845456e-01
6.16483800e-02 5.13635159e-01 4.67262268e-01 4.58110094e-01
5.93851022e-02 -3.01668197e-01 3.62338364e-01 1.79652378e-01
-2.36421689e-01 1.47149796e-02 -1.13856924e+00 -4.66070473e-01
8.20314348e-01 -8.73670056e-02 -1.25489473e-01 -5.78624547e-01
-7.43644536e-01 5.16523480e-01 2.15957444e-02 4.79689807e-01
1.08783491e-01 -1.15108371e+00 -5.29146612e-01 8.91029611e-02
-5.49432814e-01 -2.44958609e-01 6.74250007e-01 7.95520425e-01
1.79378495e-01 1.04527962e+00 1.43976778e-01 -6.13537192e-01
-5.71454883e-01 6.96442068e-01 4.84184474e-01 -2.85916388e-01
-3.48746508e-01 -1.38659582e-01 -1.97588786e-01 -4.07520175e-01
3.53501916e-01 -6.44208789e-01 -3.90451401e-01 5.10921657e-01
2.36058727e-01 1.07450581e+00 4.08898056e-01 -5.64147234e-01
-1.49063677e-01 -5.26663885e-02 5.20411193e-01 -6.53409213e-02
1.45839846e+00 -4.03411090e-01 -1.38831019e-01 9.62687969e-01
9.35464025e-01 -6.50833666e-01 -8.00995052e-01 3.10554564e-01
1.14549540e-01 1.92694604e-01 4.45181012e-01 -1.00636435e+00
-8.25383484e-01 1.10768330e+00 1.01121950e+00 5.48229396e-01
9.68095124e-01 -4.60810363e-01 9.09903824e-01 1.37282416e-01
3.11260819e-01 -1.35849559e+00 -1.19631743e+00 4.63280827e-01
4.32442904e-01 -7.47796178e-01 -4.89354461e-01 6.76416814e-01
1.48648061e-02 9.71145093e-01 -4.06822823e-02 2.89942086e-01
1.04956222e+00 4.23005819e-01 -5.57222962e-01 2.48485610e-01
-8.31993639e-01 2.07844585e-01 3.50164734e-02 4.93867636e-01
5.82606765e-03 5.89829922e-01 -2.10377514e-01 9.30683613e-01
-1.31145254e-01 3.28116775e-01 3.78807336e-01 4.35903728e-01
-2.45425761e-01 -6.93587482e-01 -2.94495374e-01 5.29622793e-01
-4.59043026e-01 -8.33199993e-02 5.57083845e-01 6.72580719e-01
2.56483257e-01 1.14109957e+00 5.64645827e-01 -1.57814119e-02
2.89635569e-01 4.31546330e-01 -6.25538081e-02 1.92603827e-01
-5.09416401e-01 -1.86919477e-02 -6.88502565e-02 1.28582001e-01
-1.84318386e-02 -5.89462101e-01 -1.24774790e+00 -7.69064784e-01
-3.69382232e-01 6.30415618e-01 1.17532694e+00 1.23311472e+00
3.89080316e-01 7.97097981e-01 7.98472226e-01 -9.86532807e-01
-7.53598511e-01 -1.33042717e+00 -7.22642183e-01 -8.94416943e-02
4.07749891e-01 -4.58296150e-01 -6.99059844e-01 -5.17088175e-01] | [6.175178050994873, 2.820765733718872] |
ec05b45d-4a71-4322-8892-e6620942696c | topology-aware-correlations-between-relations | 2103.03642 | null | https://arxiv.org/abs/2103.03642v1 | https://arxiv.org/pdf/2103.03642v1.pdf | Topology-Aware Correlations Between Relations for Inductive Link Prediction in Knowledge Graphs | Inductive link prediction -- where entities during training and inference stages can be different -- has been shown to be promising for completing continuously evolving knowledge graphs. Existing models of inductive reasoning mainly focus on predicting missing links by learning logical rules. However, many existing approaches do not take into account semantic correlations between relations, which are commonly seen in real-world knowledge graphs. To address this challenge, we propose a novel inductive reasoning approach, namely TACT, which can effectively exploit Topology-Aware CorrelaTions between relations in an entity-independent manner. TACT is inspired by the observation that the semantic correlation between two relations is highly correlated to their topological structure in knowledge graphs. Specifically, we categorize all relation pairs into several topological patterns, and then propose a Relational Correlation Network (RCN) to learn the importance of the different patterns for inductive link prediction. Experiments demonstrate that TACT can effectively model semantic correlations between relations, and significantly outperforms existing state-of-the-art methods on benchmark datasets for the inductive link prediction task. | ['Jie Wang', 'Feng Wu', 'Huarui He', 'Jiajun Chen'] | 2021-03-05 | null | null | null | null | ['inductive-link-prediction'] | ['graphs'] | [-5.79395220e-02 5.08128762e-01 -7.55935788e-01 -2.17809647e-01
2.22758397e-01 -5.24635971e-01 6.34061873e-01 5.83763540e-01
3.38146865e-01 7.61748374e-01 2.44604498e-01 -7.05363512e-01
-8.85410368e-01 -1.52289069e+00 -7.59430528e-01 1.28298290e-02
-6.40335917e-01 9.01914418e-01 6.56904697e-01 -4.40370649e-01
1.17218576e-01 2.64927864e-01 -1.27968740e+00 4.05107230e-01
1.18395174e+00 6.55960739e-01 -3.31733763e-01 1.79335505e-01
-5.55328131e-01 1.49487376e+00 -1.09220177e-01 -8.59768152e-01
-1.00466862e-01 -3.55875403e-01 -1.45797610e+00 -6.23871386e-01
2.19679445e-01 1.37227014e-01 -6.28048480e-01 8.64211619e-01
-1.12864502e-01 1.10522220e-02 7.14697421e-01 -1.54488099e+00
-7.24604905e-01 1.15850019e+00 -3.69280487e-01 3.58579308e-01
5.43017149e-01 -5.35396039e-01 1.63202274e+00 -6.48186207e-01
1.08076096e+00 1.31337261e+00 8.71240437e-01 -3.09115574e-02
-1.27999079e+00 -5.44181168e-01 4.57975596e-01 9.40384567e-01
-1.39166617e+00 -1.94816086e-02 1.00346684e+00 -5.10531068e-01
9.24813628e-01 1.13798842e-01 7.91226268e-01 6.68662906e-01
-2.86657475e-02 6.59061193e-01 8.61392438e-01 -4.06587988e-01
-1.15947962e-01 -8.59697461e-02 2.55844980e-01 7.96207428e-01
5.09190202e-01 -2.06274360e-01 -7.61276126e-01 7.73696378e-02
6.46495998e-01 -8.38549361e-02 -2.05730572e-01 -7.71405339e-01
-1.27937627e+00 4.99685943e-01 1.04382348e+00 3.67220283e-01
-1.18617378e-01 2.85549909e-02 3.37751061e-01 3.92711014e-01
4.31755692e-01 5.96209109e-01 -6.17542326e-01 1.61721051e-01
-2.99997836e-01 1.75447427e-02 1.04672837e+00 1.07887781e+00
9.04170334e-01 -7.69068718e-01 4.56886105e-02 7.94764698e-01
3.88865799e-01 -6.16335608e-02 -1.43015221e-01 -5.94566703e-01
6.76877439e-01 1.28850484e+00 -2.35762358e-01 -1.47434616e+00
-4.47221041e-01 -4.68156159e-01 -7.06312537e-01 -3.88311625e-01
3.73692542e-01 2.69212067e-01 -4.69892025e-01 1.61090446e+00
4.95054305e-01 4.50576842e-01 1.44411281e-01 4.95067030e-01
1.04524755e+00 2.58290023e-01 1.38128936e-01 -1.08966455e-01
1.10128891e+00 -9.46908474e-01 -6.56388819e-01 -3.06631848e-02
9.90077794e-01 -3.71386647e-01 5.69027543e-01 -2.51812756e-01
-9.10573363e-01 -2.91978687e-01 -8.75405967e-01 -1.31635547e-01
-8.41202974e-01 -3.88996840e-01 1.20028877e+00 -1.46444496e-02
-8.09906423e-01 6.81782663e-01 -4.83504444e-01 -5.20905793e-01
5.38191319e-01 3.17968458e-01 -4.08974171e-01 -1.24603324e-01
-1.71401894e+00 1.22808945e+00 8.70688677e-01 1.24685258e-01
-6.58558905e-02 -9.45733666e-01 -8.47007036e-01 1.51929870e-01
8.96516800e-01 -9.43911016e-01 5.46079814e-01 -4.99938339e-01
-9.31196690e-01 7.44692564e-01 -1.83031335e-01 -4.23781037e-01
2.20768496e-01 -4.53698561e-02 -7.53246129e-01 1.36890545e-01
1.07230641e-01 4.56056803e-01 6.56549484e-02 -1.51458800e+00
-5.84388554e-01 -1.81099683e-01 4.92657214e-01 2.22816765e-01
-4.39670533e-01 -4.35315758e-01 -4.97135073e-01 -3.74832511e-01
5.05699694e-01 -6.88347042e-01 5.88332722e-03 -2.34569125e-02
-8.25241506e-01 -5.63651502e-01 8.32402229e-01 -4.02238250e-01
1.44409847e+00 -1.46278489e+00 4.50087786e-01 7.31416106e-01
5.56507528e-01 1.12671502e-01 1.34712771e-01 5.72769523e-01
-1.35931417e-01 1.18668199e-01 1.42753810e-01 4.01806563e-01
-8.29661489e-02 4.71879840e-01 -4.35972720e-01 -2.42446169e-01
1.61197275e-01 1.25723565e+00 -1.29499269e+00 -8.94815445e-01
-1.69490092e-02 1.61406934e-01 -3.67652625e-01 4.23404500e-02
-4.70970303e-01 3.06104273e-01 -5.54914117e-01 5.52853703e-01
4.71426398e-01 -6.46276057e-01 1.03699923e+00 -5.49535513e-01
3.84615481e-01 7.09182084e-01 -8.39792371e-01 1.46131587e+00
-2.81351596e-01 5.23682594e-01 -6.57873809e-01 -1.24648798e+00
1.14202499e+00 -3.88056673e-02 5.77065587e-01 -6.66786611e-01
-2.68879741e-01 1.97619841e-01 2.35493734e-01 -5.60341358e-01
3.03335100e-01 5.24454154e-02 3.61631632e-01 1.94316849e-01
-3.23660299e-02 1.16611794e-01 4.40125257e-01 6.54862344e-01
1.28878355e+00 3.02414209e-01 1.92565247e-01 -2.20591173e-01
6.60236120e-01 1.11302204e-01 6.82002544e-01 3.56663287e-01
1.93810493e-01 -2.29413375e-01 9.44955647e-01 -6.51782572e-01
-6.39951706e-01 -1.35088801e+00 1.33689065e-02 8.39601636e-01
7.51026452e-01 -8.45248938e-01 -1.52362689e-01 -1.07445097e+00
3.29054415e-01 4.79933888e-01 -7.45257795e-01 -2.92994291e-01
-6.30501032e-01 -3.22723240e-01 2.71718442e-01 7.48408496e-01
4.65492964e-01 -9.40542996e-01 2.28418544e-01 -1.49758086e-02
-4.43273097e-01 -1.44519162e+00 2.21324787e-01 3.81917716e-03
-1.04300737e+00 -1.68621504e+00 1.49498299e-01 -8.92368197e-01
9.41111505e-01 1.41151980e-01 1.64569116e+00 6.34965181e-01
-3.87663767e-02 2.16375545e-01 -4.56873327e-01 8.38959124e-03
-1.28062576e-01 4.39392954e-01 -1.39481366e-01 -2.44141698e-01
5.28971195e-01 -1.03458405e+00 -5.02328217e-01 5.03942788e-01
-4.22911108e-01 4.24543500e-01 4.77347553e-01 6.67918205e-01
4.31660116e-01 6.65216684e-01 6.39169157e-01 -1.34752381e+00
3.82266611e-01 -6.90736711e-01 -8.92345682e-02 8.19955587e-01
-8.84700298e-01 2.96020597e-01 4.34231281e-01 -1.18545793e-01
-1.05533051e+00 -4.99400914e-01 2.92472273e-01 4.25368622e-02
2.29238600e-01 1.09830594e+00 -1.95506379e-01 -2.34024554e-01
3.57910901e-01 -1.57913908e-01 -3.80711973e-01 -1.38046190e-01
4.71976310e-01 -3.36296484e-02 3.89168143e-01 -9.82175410e-01
1.13814592e+00 4.87224251e-01 5.70945323e-01 -2.09795326e-01
-1.23228455e+00 -4.24888641e-01 -1.17776656e+00 -1.70856982e-01
4.55434114e-01 -7.85889566e-01 -7.71791399e-01 -6.69424757e-02
-1.04446232e+00 -2.34549955e-01 -1.42244369e-01 3.56131703e-01
-2.88115948e-01 2.02440470e-01 -5.59114158e-01 -4.97163415e-01
4.58899923e-02 -4.88822222e-01 3.72077674e-01 1.79413542e-01
-3.63348156e-01 -1.63781416e+00 6.93288073e-02 4.57702935e-01
1.75134256e-01 3.12965870e-01 1.61569393e+00 -6.66760087e-01
-9.37055945e-01 3.43462229e-02 -6.12490058e-01 -4.64687884e-01
2.93365955e-01 1.87715888e-02 -2.68482268e-01 3.15362662e-01
-1.09043109e+00 -1.99531332e-01 8.74069870e-01 -1.88439161e-01
1.07337940e+00 -2.23904505e-01 -1.00150561e+00 3.46311182e-01
1.16653609e+00 -4.53993641e-02 8.05396914e-01 4.44088817e-01
1.08580828e+00 8.45861077e-01 6.89892113e-01 -1.42608687e-01
1.06979454e+00 8.06867957e-01 4.19437021e-01 -8.53163525e-02
-8.16089883e-02 -8.46037030e-01 -2.26427928e-01 9.87281621e-01
-5.30536652e-01 7.85012990e-02 -1.03668582e+00 6.61629677e-01
-2.25740981e+00 -9.68270659e-01 -5.92836261e-01 1.78799009e+00
1.11263716e+00 4.43407744e-01 -4.32110652e-02 1.93351671e-01
6.52464032e-01 -1.86387021e-02 -3.97136897e-01 4.57543619e-02
-1.58487082e-01 1.30403355e-01 9.99211594e-02 4.31496680e-01
-8.93775582e-01 1.14077103e+00 5.99294567e+00 5.36204517e-01
-4.47955906e-01 -1.92102417e-01 3.48746091e-01 4.23567414e-01
-6.83579743e-01 5.65544903e-01 -5.55540502e-01 1.92747682e-01
4.81203020e-01 -2.61781484e-01 2.56158262e-01 6.06970787e-01
-3.70279759e-01 2.10113898e-02 -1.25910223e+00 5.14271677e-01
-1.87690735e-01 -1.63484299e+00 3.47167909e-01 1.24653867e-02
8.72477889e-01 -5.44131219e-01 -3.04514170e-01 3.41575742e-01
6.95692420e-01 -1.02240610e+00 1.68146983e-01 8.94485772e-01
3.48226517e-01 -7.35799432e-01 8.16874206e-01 -8.11987743e-03
-1.71591735e+00 7.39859119e-02 -2.66722858e-01 -1.81682155e-01
2.31767241e-02 9.00925338e-01 -8.80672872e-01 1.17287469e+00
7.00595975e-01 1.28589749e+00 -8.14031959e-01 8.69255543e-01
-7.41384089e-01 4.40272957e-01 -7.46281296e-02 6.09385520e-02
-3.25500786e-01 -1.26777634e-01 2.17485160e-01 8.88520777e-01
1.08743787e-01 7.57870674e-02 4.34119888e-02 8.27277482e-01
-3.53876442e-01 -8.36981088e-03 -7.37126708e-01 7.75977271e-03
8.67814541e-01 1.01263249e+00 -9.24870133e-01 -3.00600201e-01
-4.66580123e-01 6.19905770e-01 8.75100791e-01 3.11488181e-01
-7.50349402e-01 -2.53097266e-01 4.92057234e-01 2.53889471e-01
2.88933396e-01 -2.42957175e-01 -4.04004723e-01 -1.13667083e+00
1.48176029e-01 -2.36240819e-01 6.97487235e-01 -7.41687357e-01
-1.44272518e+00 3.30500543e-01 9.04601365e-02 -9.54353154e-01
8.77831802e-02 -4.77456897e-01 -6.76898658e-01 4.89085764e-01
-1.73297071e+00 -1.45672250e+00 -4.38312739e-01 5.53492785e-01
-1.27096787e-01 -1.13979233e-02 7.77012229e-01 1.96477264e-01
-4.60119694e-01 4.13485855e-01 -3.31354946e-01 3.38720322e-01
4.62850839e-01 -1.36713719e+00 3.38519394e-01 4.80031639e-01
4.52586263e-01 8.48471582e-01 4.13153887e-01 -8.85749698e-01
-1.22606182e+00 -1.13563371e+00 1.27425301e+00 -5.73016644e-01
1.15032458e+00 -2.88924985e-02 -1.12572885e+00 9.35878396e-01
-8.61601308e-02 3.59120548e-01 8.71893585e-01 1.03450382e+00
-9.33782220e-01 -1.58290207e-01 -5.40191531e-01 6.63260937e-01
1.92089128e+00 -6.19319260e-01 -7.89285123e-01 2.96389729e-01
7.89465487e-01 -2.05129489e-01 -1.35366571e+00 9.02472138e-01
4.99995023e-01 -9.30975735e-01 1.41110241e+00 -8.13963652e-01
8.37827086e-01 -5.21927238e-01 2.29993194e-01 -1.35253775e+00
-4.71844703e-01 -2.73927450e-01 -8.50393891e-01 1.27770114e+00
7.97728240e-01 -6.35437131e-01 8.96987677e-01 2.60011047e-01
3.33727390e-01 -9.13670301e-01 -5.24746418e-01 -8.10492992e-01
-1.40568018e-01 -9.28011984e-02 7.21263885e-01 1.65667725e+00
6.53708935e-01 7.60548651e-01 -1.89539209e-01 2.74644285e-01
5.45421302e-01 5.19019544e-01 6.99071229e-01 -1.94065750e+00
-1.19042218e-01 -5.00896573e-01 -8.60160351e-01 -8.76175821e-01
4.63134974e-01 -1.16456687e+00 -3.67774993e-01 -2.06221128e+00
2.78202325e-01 -9.93766904e-01 -3.98889571e-01 5.59664071e-01
-5.17267168e-01 3.22187692e-02 -8.80845711e-02 3.13084215e-01
-1.03662026e+00 5.71340442e-01 1.32279384e+00 -2.82308519e-01
-8.25530663e-02 -2.26660758e-01 -6.97823703e-01 7.65457273e-01
5.35710335e-01 -1.29576668e-01 -8.50165665e-01 -1.97530761e-01
1.05969107e+00 -8.54554996e-02 3.73958766e-01 -7.63961256e-01
5.96792340e-01 -3.55631322e-01 1.88333467e-01 -5.22398055e-01
-1.46721760e-02 -1.05426776e+00 4.17903185e-01 3.01014125e-01
-4.38640773e-01 -3.22248697e-01 -1.10718161e-01 7.44218767e-01
-3.88365597e-01 1.07532538e-01 2.32144240e-02 1.57604754e-01
-9.46182907e-01 3.17919761e-01 4.61915225e-01 1.55751824e-01
1.01058280e+00 9.38434619e-03 -8.63528967e-01 -1.87977493e-01
-8.20769370e-01 5.82100451e-01 1.87737659e-01 6.19104385e-01
5.37827194e-01 -1.64572251e+00 -3.60744387e-01 -3.90523762e-01
4.23181295e-01 2.75895864e-01 1.08942732e-01 1.00041866e+00
-3.04093689e-01 3.63955021e-01 -2.78962273e-02 -2.60169327e-01
-1.28985107e+00 6.85642898e-01 2.74311244e-01 -8.48864675e-01
-7.22789168e-01 7.23262787e-01 -6.45114556e-02 -5.81450164e-01
-3.62670682e-02 -1.25232458e-01 -4.56449777e-01 3.88468914e-02
3.20972763e-02 2.76642352e-01 -1.23187415e-01 -4.08077329e-01
-4.71427053e-01 7.63497829e-01 -2.35677928e-01 5.73538482e-01
1.30942321e+00 1.91904642e-02 -5.93800128e-01 5.44203103e-01
7.75038183e-01 1.16630923e-02 -7.38258302e-01 -7.56221950e-01
6.33310735e-01 -3.39389682e-01 -4.37353373e-01 -8.45770538e-01
-1.02411950e+00 5.44560254e-01 -3.35470021e-01 5.93564212e-01
8.69042218e-01 4.31848109e-01 6.93385720e-01 5.85887372e-01
5.89116573e-01 -9.60943282e-01 2.59239644e-01 7.49474227e-01
6.52736843e-01 -1.21735883e+00 3.82940233e-01 -1.54300034e+00
-3.25259775e-01 1.17356205e+00 1.02755916e+00 1.92128345e-01
8.56926024e-01 -2.13324055e-02 -5.05786061e-01 -6.47197306e-01
-9.16551113e-01 -4.25529450e-01 5.56529880e-01 7.86768556e-01
4.99424905e-01 2.20332414e-01 -2.41681695e-01 1.71259567e-01
-2.35791728e-01 -1.34974182e-01 -4.66972627e-02 8.50351453e-01
-2.64656782e-01 -1.39413941e+00 6.38068467e-02 5.32976031e-01
1.53390795e-01 -1.50204390e-01 -8.28288913e-01 7.16900647e-01
2.40598470e-01 7.26602197e-01 1.16425946e-01 -7.37768531e-01
3.42014939e-01 5.03786653e-02 6.23116016e-01 -4.28911000e-01
-1.10405311e-01 -7.38351524e-01 5.03235579e-01 -3.81547928e-01
-6.90050840e-01 -3.30694348e-01 -1.46877444e+00 -5.83119392e-01
-3.17622125e-01 2.97626257e-01 1.31184086e-01 1.26206827e+00
4.83764172e-01 9.01666880e-01 3.18454057e-01 -6.35474101e-02
3.60389113e-01 -7.01788008e-01 -4.38950479e-01 6.23933434e-01
-2.35288784e-01 -1.20852172e+00 -1.08834706e-01 -3.09332237e-02] | [8.914011001586914, 7.936952114105225] |
7d035292-7210-454b-8cf6-b111b8ab55ec | multi-task-sub-band-network-for-deep-residual | 2303.06404 | null | https://arxiv.org/abs/2303.06404v1 | https://arxiv.org/pdf/2303.06404v1.pdf | Multi-Task Sub-Band Network For Deep Residual Echo Suppression | This paper introduces the SWANT team entry to the ICASSP 2023 AEC Challenge. We submit a system that cascades a linear filter with a neural post-filter. Particularly, we adopt sub-band processing to handle full-band signals and shape the network with multi-task learning, where dual signal voice activity detection (DSVAD) and echo estimation are adopted as auxiliary tasks. Moreover, we particularly improve the time frequency convolution module (TFCM) to increase the receptive field using small convolution kernels. Finally, our system has ranked 4th in ICASSP 2023 AEC Challenge Non-personalized track. | ['Yang Li', 'Yukai Ju', 'Jindong Li', 'Zhaoxia Li', 'Dawei Luo', 'Jiayao Sun'] | 2023-03-11 | null | null | null | null | ['activity-detection'] | ['computer-vision'] | [-1.98401421e-01 -4.69292641e-01 3.25216979e-01 -2.93767691e-01
-1.17507279e+00 -7.17536509e-01 5.00592291e-01 -6.09152019e-01
-6.79989338e-01 4.16220605e-01 6.57818556e-01 -1.89850330e-01
-1.68836266e-01 -1.43058561e-02 -6.30908132e-01 -4.76233393e-01
-3.69447649e-01 -2.53759682e-01 1.53910816e-01 -4.61750180e-02
-1.15511857e-01 4.67183411e-01 -1.10328257e+00 7.68067598e-01
8.53673220e-01 1.16699123e+00 1.10339515e-01 1.09643877e+00
2.85466820e-01 2.34612778e-01 -8.14682245e-01 -1.95493370e-01
1.76695958e-01 -2.19326600e-01 -5.48548639e-01 -3.94956887e-01
5.99631727e-01 -7.04907402e-02 -3.47807288e-01 9.20076847e-01
1.13983011e+00 3.37468296e-01 4.05261785e-01 -1.02229750e+00
-3.48549724e-01 7.60327458e-01 -3.23597997e-01 6.21324837e-01
2.69949973e-01 2.44622409e-01 7.71673441e-01 -1.27090132e+00
1.06219454e-02 1.21579826e+00 1.08962870e+00 4.87802446e-01
-9.95083690e-01 -9.28253233e-01 1.85962647e-01 5.80887675e-01
-1.11169839e+00 -7.71260560e-01 7.36158371e-01 -2.07226202e-01
1.21755981e+00 3.67655277e-01 5.32528996e-01 1.65643930e+00
3.54814436e-03 8.48746896e-01 1.10755575e+00 3.18788812e-02
6.11221511e-03 -1.65371656e-01 -3.05266865e-02 2.30679557e-01
-4.90801364e-01 2.61094213e-01 -8.32609475e-01 -1.45805627e-01
4.56053406e-01 -2.64301360e-01 -4.92889702e-01 7.69131303e-01
-1.33536720e+00 3.94486904e-01 2.00518057e-01 3.48142207e-01
-7.61659622e-01 3.29703093e-01 4.62042570e-01 5.23802340e-01
5.37672877e-01 3.30258459e-01 -7.58491814e-01 -3.50758970e-01
-1.10523450e+00 3.09989393e-01 7.46516287e-01 6.92899227e-01
1.20521903e-01 4.84741271e-01 -5.43844044e-01 1.16286552e+00
3.39684129e-01 6.20716631e-01 5.84102571e-01 -1.12620533e+00
3.95174891e-01 -3.89896423e-01 -9.38546062e-02 -5.48967719e-01
-7.03776002e-01 -1.09542406e+00 -8.58023643e-01 -2.68792808e-01
3.78885753e-02 -6.81627214e-01 -7.04551220e-01 1.62366498e+00
7.84746632e-02 6.64098740e-01 -1.01610973e-01 1.04699862e+00
8.16296339e-01 9.58191872e-01 5.86343892e-02 -3.81272763e-01
1.48690546e+00 -1.23433197e+00 -1.16282928e+00 -2.57054657e-01
-2.57624425e-02 -9.46722209e-01 7.72690892e-01 8.63580644e-01
-1.17365873e+00 -9.75905240e-01 -9.05655146e-01 3.02824587e-01
-3.73749197e-01 4.08501208e-01 4.96057481e-01 8.12972844e-01
-9.09449637e-01 5.67429662e-01 -6.23064756e-01 1.53375626e-01
1.36973029e-02 2.26770714e-01 -4.23866101e-02 3.11973274e-01
-1.52887440e+00 8.84714305e-01 1.51753590e-01 4.06989276e-01
-9.91975904e-01 -1.10431015e+00 -4.77840215e-01 1.06926352e-01
7.93858916e-02 -3.98390502e-01 1.25190747e+00 -6.02206826e-01
-1.88364470e+00 2.61989623e-01 3.62390168e-02 -7.47089446e-01
2.77725220e-01 -5.84181547e-01 -1.24726439e+00 1.21588998e-01
-3.63155454e-01 5.35916269e-01 1.40506077e+00 -6.24232531e-01
-8.00313652e-01 2.00968340e-01 -3.39855075e-01 1.18530281e-01
-3.70128423e-01 3.48944306e-01 -3.02310735e-01 -1.11069512e+00
-1.35822982e-01 -6.44516110e-01 3.63210887e-02 -5.09481251e-01
-1.06331095e-01 -3.71399701e-01 8.65090907e-01 -1.10666108e+00
1.50106514e+00 -2.55741119e+00 -7.22598881e-02 1.22818574e-01
1.95841983e-01 4.53134060e-01 -5.20036340e-01 1.63150847e-01
-3.93000007e-01 -1.32357731e-01 9.59765539e-02 -5.31550407e-01
1.23697966e-01 -3.47339481e-01 -4.59046423e-01 4.83027250e-01
2.83582747e-01 5.58928490e-01 -4.42488492e-01 2.74368399e-03
4.59263213e-02 7.10751534e-01 -7.10160017e-01 3.77071500e-02
2.71664739e-01 2.94489652e-01 -7.43941665e-02 4.06043470e-01
9.99694228e-01 7.58385882e-02 -3.80882293e-01 -6.34068787e-01
-5.34710884e-01 5.85153997e-01 -1.21882927e+00 1.86973512e+00
-6.81624055e-01 5.88833570e-01 6.27247512e-01 -8.04881811e-01
6.86289132e-01 7.32361317e-01 6.20877624e-01 -4.84610975e-01
-5.02435043e-02 3.57911646e-01 3.71715188e-01 -4.17320579e-01
2.29283407e-01 -1.66936070e-02 7.47132450e-02 1.89118966e-01
7.16987550e-01 -1.27466872e-01 -2.83570588e-01 -1.69424608e-01
1.10372651e+00 7.26279691e-02 -1.52258098e-01 -3.61357093e-01
7.95112669e-01 -5.89527667e-01 6.91597283e-01 7.77642012e-01
-3.61330539e-01 8.67971003e-01 -1.20058037e-01 -2.08382413e-01
-6.12782121e-01 -1.07155716e+00 -1.22206636e-01 1.43557453e+00
-6.07391477e-01 -5.53966045e-01 -7.79093087e-01 -4.73701656e-01
-5.07517867e-02 5.57618976e-01 -3.85188282e-01 -1.83979750e-01
-7.35507369e-01 -7.19077647e-01 1.05712044e+00 4.97016162e-01
5.98916948e-01 -1.09893334e+00 -1.14067845e-01 6.57481372e-01
-4.67502385e-01 -1.13731706e+00 -1.30081224e+00 5.75232148e-01
-3.94464076e-01 -6.30083919e-01 -9.49434638e-01 -6.61290228e-01
-3.66751671e-01 4.28267159e-02 4.34491277e-01 -5.49356639e-01
-7.12282732e-02 4.98654276e-01 -3.59740138e-01 -4.99703199e-01
1.32028237e-01 1.34050563e-01 4.11010951e-01 4.00413811e-01
1.18781582e-01 -6.87393665e-01 -7.35239744e-01 1.30421460e-01
-3.23121756e-01 -1.35364592e-01 5.68364441e-01 6.98457539e-01
2.66562372e-01 -1.28997907e-01 9.09220576e-01 -2.71138877e-01
1.17495108e+00 -2.15948686e-01 -3.81512374e-01 1.42676249e-01
-4.23947006e-01 -3.60288739e-01 6.48247540e-01 -9.23547685e-01
-1.32831609e+00 1.08794086e-01 -7.43089557e-01 -5.26916206e-01
1.18579678e-01 2.28047147e-01 -7.77501753e-03 -9.34019908e-02
5.22280455e-01 4.13586587e-01 -3.19702268e-01 -7.40987599e-01
2.60096610e-01 1.00837445e+00 8.69757593e-01 -3.87029946e-01
7.95578480e-01 8.20175931e-02 -4.43685263e-01 -8.93374860e-01
-8.89714658e-01 -6.11872971e-01 -2.19389364e-01 -4.30795014e-01
1.02080309e+00 -1.34277582e+00 -1.12355614e+00 7.90098011e-01
-1.47199714e+00 -1.75886780e-01 5.49537539e-02 1.15631402e+00
-4.13431823e-01 -1.97012443e-02 -7.63792992e-01 -8.01236749e-01
-9.71987486e-01 -8.32685471e-01 8.86812687e-01 2.85415381e-01
-1.46660239e-01 -6.29365206e-01 2.36457467e-01 8.00532997e-02
1.09005022e+00 -4.56942707e-01 3.12219203e-01 -7.53704727e-01
-1.08367175e-01 7.11806118e-02 1.00268207e-01 8.39901507e-01
-1.37859985e-01 -2.91414708e-01 -1.61697555e+00 -3.40719283e-01
3.14641684e-01 -7.77818114e-02 1.00330865e+00 8.02271962e-01
1.49551761e+00 -1.35659322e-01 1.10930569e-01 1.02993309e+00
6.37613356e-01 3.32014918e-01 5.11049807e-01 -9.81136337e-02
3.39735121e-01 3.56483012e-01 9.87240374e-02 4.22413468e-01
1.76863208e-01 5.32975435e-01 -6.15576506e-02 -7.57833943e-02
-5.30555904e-01 3.23419110e-03 6.58330381e-01 1.48242068e+00
-2.37248495e-01 -3.26973140e-01 -5.52646041e-01 3.08487087e-01
-1.43318272e+00 -9.64587510e-01 -2.59823233e-01 1.75249660e+00
8.27021658e-01 2.62875576e-02 1.56211793e-01 -1.38911515e-01
7.37196445e-01 2.98661649e-01 -3.79097342e-01 -2.25544691e-01
-2.88200736e-01 7.76607215e-01 4.59581912e-01 5.00612855e-01
-1.35498405e+00 6.72030389e-01 7.06657934e+00 1.23600960e+00
-1.20637667e+00 5.64590931e-01 2.82110926e-02 -3.59877944e-01
-9.21414942e-02 -4.62747395e-01 -8.10513377e-01 5.98939955e-01
1.47825778e+00 -7.59570524e-02 7.51660943e-01 4.88831878e-01
2.54999936e-01 3.74629378e-01 -6.89647377e-01 1.44376445e+00
2.65425779e-02 -1.04667008e+00 -6.03947520e-01 3.03841084e-02
2.86018074e-01 4.81102169e-01 2.21120745e-01 8.10411811e-01
-1.82019994e-01 -8.61808598e-01 8.66818666e-01 6.89011991e-01
7.08172917e-01 -7.76182890e-01 4.28650081e-01 5.65143488e-02
-1.41522598e+00 -3.20249826e-01 -1.09392725e-01 4.30813849e-01
2.97296166e-01 4.68547493e-01 -3.86551857e-01 4.55792338e-01
8.76158774e-01 4.04666632e-01 -2.36645922e-01 1.49326062e+00
-3.38779166e-02 1.07176769e+00 -4.28219348e-01 6.96169734e-02
1.71281084e-01 1.81398973e-01 7.62053490e-01 1.66732562e+00
4.64980453e-01 1.63750753e-01 1.18406932e-03 6.02064371e-01
-2.91290730e-01 -1.62837982e-01 1.39777148e-02 -2.81575229e-02
5.23981929e-01 1.47999275e+00 -1.11061528e-01 -1.88850462e-01
-5.49627423e-01 1.11791182e+00 -2.12277249e-01 6.12239420e-01
-9.65843022e-01 -7.16982245e-01 7.88272023e-01 -5.37077904e-01
2.46163920e-01 -1.99692219e-01 -4.56850380e-02 -8.55183184e-01
-2.76582301e-01 -9.75171149e-01 8.33758935e-02 -6.84364915e-01
-1.33565438e+00 7.17118561e-01 -3.75760257e-01 -1.32491219e+00
-2.71143336e-02 -7.01088727e-01 -9.18640256e-01 1.28072810e+00
-1.35400009e+00 -8.51723611e-01 6.35451823e-02 8.34763169e-01
7.38404274e-01 -4.10411179e-01 8.70394349e-01 1.10362756e+00
-3.81706774e-01 6.84648454e-01 -1.63889229e-01 -1.28873631e-01
9.80475366e-01 -1.10144246e+00 7.02472746e-01 8.34757626e-01
6.98368251e-02 6.55282199e-01 4.65631247e-01 -5.37489414e-01
-1.34884393e+00 -1.12258828e+00 8.08118820e-01 -3.12407333e-02
8.30544889e-01 -5.91066420e-01 -8.74989152e-01 4.01025653e-01
5.92028379e-01 2.05308422e-01 6.33946598e-01 2.50471592e-01
-1.73739031e-01 -5.02973497e-01 -6.90671921e-01 1.90339059e-01
8.80445242e-01 -1.09048975e+00 -7.01177120e-01 1.16747141e-01
9.86281872e-01 -4.83526289e-01 -8.83438468e-01 4.48671281e-01
8.18539143e-01 -3.13151270e-01 1.16147065e+00 -5.05293250e-01
-2.68980145e-01 -1.82696030e-01 -1.69213757e-01 -1.72835457e+00
-5.61900735e-01 -1.19882309e+00 -3.01199734e-01 1.20948839e+00
5.66981554e-01 -5.09004474e-01 2.03912631e-01 -2.15629920e-01
-7.66144693e-01 -2.42854238e-01 -9.05014932e-01 -7.95201480e-01
-2.17551380e-01 -8.28996658e-01 2.34697863e-01 6.27048552e-01
-1.82120040e-01 2.62030005e-01 -8.77535701e-01 1.47202134e-01
4.10416126e-01 -3.74840856e-01 5.47859706e-02 -1.13505232e+00
-5.73690414e-01 -3.77759308e-01 4.58136916e-01 -1.13221681e+00
-1.16162039e-01 -7.84419119e-01 1.69791520e-01 -9.32270944e-01
-4.19211894e-01 1.85196221e-01 -7.26437688e-01 2.16556817e-01
-2.55023185e-02 1.76883399e-01 3.61988515e-01 -2.62658596e-01
-5.74652612e-01 5.82142889e-01 1.06040871e+00 -1.15017541e-01
-2.39682138e-01 4.55306709e-01 -4.52899545e-01 6.76586688e-01
6.73971653e-01 -4.30860400e-01 -1.60957798e-01 -3.98199975e-01
-5.69589809e-03 9.71705690e-02 2.90998429e-01 -1.15924656e+00
6.09067202e-01 2.85350502e-01 5.99099815e-01 -1.02247798e+00
6.32983208e-01 -6.40270770e-01 2.05741391e-01 4.89472449e-01
-3.86472911e-01 -1.06267612e-02 5.14541984e-01 2.49423280e-01
-2.43750498e-01 2.76501337e-03 6.89988792e-01 1.36689976e-01
-4.66382056e-01 3.08862388e-01 -7.42347479e-01 -8.40520039e-02
3.21064085e-01 3.48234415e-01 -3.45171839e-01 -2.16738135e-01
-1.12140572e+00 4.58158553e-01 -9.53788340e-01 6.39791310e-01
7.25797892e-01 -1.56204462e+00 -8.66571903e-01 3.06036979e-01
-3.12600970e-01 -7.96894729e-01 7.06411898e-01 1.26638949e+00
6.24987781e-02 3.56056601e-01 -1.96479440e-01 -3.23783934e-01
-1.00830281e+00 3.96621041e-02 5.79493105e-01 1.19540073e-01
-5.48835218e-01 1.14769709e+00 -7.89486617e-02 -4.94260937e-01
7.35706508e-01 -2.77303785e-01 -2.72750348e-01 3.58854473e-01
8.72018158e-01 5.26103914e-01 3.35725725e-01 -2.14713797e-01
-7.05707967e-01 2.75863647e-01 1.58055499e-02 -4.17421371e-01
1.58688200e+00 -1.28868341e-01 1.46365643e-01 2.66746253e-01
1.31221521e+00 3.02191257e-01 -1.28705537e+00 -1.59120321e-01
-3.84175926e-02 1.20577000e-01 5.47293186e-01 -1.22080004e+00
-1.02379870e+00 1.02457583e+00 1.08899009e+00 1.38426036e-01
1.40499735e+00 -3.71585727e-01 8.88086915e-01 4.45686042e-01
-1.26114696e-01 -1.38760686e+00 7.38715380e-02 9.52871859e-01
1.31400371e+00 -6.81486130e-01 -3.86025071e-01 3.59117314e-02
-5.14564455e-01 1.34834039e+00 4.14773226e-01 -2.38209307e-01
1.05470800e+00 6.32315457e-01 1.11995004e-02 1.05172945e-02
-8.01597714e-01 -2.50228792e-01 7.39897013e-01 6.44082129e-01
4.57800508e-01 1.01350425e-02 -4.14689392e-01 1.14493370e+00
-3.47036779e-01 -3.13260645e-01 1.80351157e-02 1.33235529e-01
-3.92304957e-01 -8.03380191e-01 -3.55984122e-01 3.03545743e-01
-7.67580032e-01 -4.88754511e-01 -8.37927312e-02 2.99000531e-01
2.22527936e-01 9.20998275e-01 -8.35587271e-03 -7.53548682e-01
6.00906968e-01 3.40366811e-01 3.31989318e-01 -2.45522931e-01
-1.27739084e+00 9.13705766e-01 2.39531040e-01 -7.30182409e-01
-2.69654632e-01 -7.33340383e-01 -1.02606940e+00 -5.49761467e-02
-9.03004631e-02 1.27063900e-01 9.40647244e-01 8.61796200e-01
3.70712280e-01 1.14732063e+00 7.46978521e-01 -8.43849182e-01
-6.52629793e-01 -1.47337711e+00 -6.08954012e-01 -1.66546583e-01
7.81630814e-01 -3.59453708e-01 -3.87772262e-01 -1.28873615e-02] | [14.963545799255371, 5.912715911865234] |
faaf8a4c-5473-4d8c-9dc2-9482a204205c | depac-a-corpus-for-depression-and-anxiety-1 | 2306.12443 | null | https://arxiv.org/abs/2306.12443v1 | https://arxiv.org/pdf/2306.12443v1.pdf | DEPAC: a Corpus for Depression and Anxiety Detection from Speech | Mental distress like depression and anxiety contribute to the largest proportion of the global burden of diseases. Automated diagnosis systems of such disorders, empowered by recent innovations in Artificial Intelligence, can pave the way to reduce the sufferings of the affected individuals. Development of such systems requires information-rich and balanced corpora. In this work, we introduce a novel mental distress analysis audio dataset DEPAC, labeled based on established thresholds on depression and anxiety standard screening tools. This large dataset comprises multiple speech tasks per individual, as well as relevant demographic information. Alongside, we present a feature set consisting of hand-curated acoustic and linguistic features, which were found effective in identifying signs of mental illnesses in human speech. Finally, we justify the quality and effectiveness of our proposed audio corpus and feature set in predicting depression severity by comparing the performance of baseline machine learning models built on this dataset with baseline models trained on other well-known depression corpora. | ['Jekaterina Novikova', 'Brian Diep', 'Malikeh Ehghaghi', 'Mashrura Tasnim'] | 2023-06-20 | depac-a-corpus-for-depression-and-anxiety | https://aclanthology.org/2022.clpsych-1.1 | https://aclanthology.org/2022.clpsych-1.1.pdf | naacl-clpsych-2022-7 | ['anxiety-detection'] | ['medical'] | [ 1.53569892e-01 3.74769956e-01 1.99306443e-01 -5.78879893e-01
-1.29943609e+00 -1.50354400e-01 3.03454965e-01 4.99415666e-01
-3.05232137e-01 4.57540691e-01 6.65929496e-01 2.93333173e-01
-3.72257054e-01 -5.50555587e-01 3.63929957e-01 -5.04783213e-01
-7.43338987e-02 5.52712262e-01 -2.26511553e-01 -2.60066241e-01
4.17544730e-02 3.13481867e-01 -1.51232553e+00 5.31296611e-01
7.73147345e-01 1.02012086e+00 -4.71126996e-02 5.20521998e-01
7.68542066e-02 7.15351224e-01 -8.17090452e-01 -5.98122120e-01
-5.93965173e-01 -2.81676531e-01 -9.36117530e-01 -8.89428053e-03
2.55090035e-02 -3.84310246e-01 -8.64606053e-02 7.09368169e-01
1.13834035e+00 -2.93745965e-01 7.66664088e-01 -9.72577095e-01
-2.99091250e-01 6.62041783e-01 -2.61923820e-01 3.49219531e-01
8.29583704e-01 -2.42196992e-01 7.06459403e-01 -6.09875500e-01
3.82172137e-01 1.24616337e+00 9.64746356e-01 6.77325487e-01
-1.06087649e+00 -6.86327040e-01 -3.24884713e-01 3.36045206e-01
-1.12330449e+00 -8.55869591e-01 8.51563156e-01 -5.93380749e-01
1.14781201e+00 8.11920986e-02 6.14670038e-01 1.33620071e+00
-2.25147307e-01 3.30005616e-01 4.85379547e-01 -4.35903579e-01
2.07627758e-01 2.83274710e-01 1.42446995e-01 4.61720288e-01
1.03031673e-01 -5.13239801e-01 -5.24099708e-01 -5.61267078e-01
-1.27165452e-01 -2.71642923e-01 -9.83490050e-03 1.14149906e-01
-8.45701993e-01 7.72743583e-01 -2.68351585e-01 6.87910438e-01
-7.05673397e-01 -1.28038913e-01 7.63753831e-01 1.17086515e-01
8.79137099e-01 2.43854687e-01 -5.57073057e-01 -6.41951919e-01
-8.48840594e-01 8.19355715e-03 5.50233722e-01 1.65379107e-01
2.30663523e-01 -7.78454496e-03 -1.22130424e-01 1.46304882e+00
2.85258293e-01 5.58565676e-01 7.91995287e-01 -8.95433664e-01
4.78949606e-01 8.58670413e-01 -2.80378968e-01 -1.33529913e+00
-9.41084027e-01 -2.41016179e-01 -9.09034669e-01 -5.92079461e-01
-3.92811783e-02 -4.66302484e-01 -3.38607430e-01 1.79995286e+00
3.39182287e-01 -1.71332374e-01 1.84073105e-01 3.28518212e-01
8.36942852e-01 3.61681223e-01 -4.98377904e-02 -3.81399989e-01
1.18834591e+00 -1.01941377e-01 -9.02664363e-01 -1.77535132e-01
8.32952023e-01 -4.69626397e-01 8.00671637e-01 7.66798794e-01
-1.11612499e+00 -3.62890065e-01 -6.80767357e-01 1.60550535e-01
-2.36921370e-01 3.26914608e-01 6.00432217e-01 1.23117447e+00
-1.13882375e+00 2.72981405e-01 -8.09166670e-01 -6.07907712e-01
4.62673396e-01 5.54259598e-01 -6.14126682e-01 1.09485820e-01
-1.12663257e+00 6.44810200e-01 2.07119249e-02 -1.06621971e-02
-6.40407205e-01 -5.54022372e-01 -7.68754721e-01 -3.48738767e-02
-1.05057336e-01 -2.90402204e-01 1.24643695e+00 -7.13940799e-01
-1.12209654e+00 1.15984094e+00 -2.03506276e-03 -4.83125716e-01
-1.76869612e-02 -2.54406869e-01 -7.99346268e-01 6.71932578e-01
2.16908470e-01 5.03804743e-01 5.66474319e-01 -5.54969311e-01
-5.58873355e-01 -7.61209905e-01 -4.69610244e-01 -7.59831518e-02
-1.01636946e+00 4.79852974e-01 -1.53373471e-02 -4.08531010e-01
-4.58713956e-02 -5.49764872e-01 -9.15142428e-03 -1.85580537e-01
-4.28927630e-01 -1.93913803e-01 6.51354492e-01 -1.02104533e+00
1.48458350e+00 -2.40881634e+00 -4.76300530e-02 1.71669215e-01
2.05979481e-01 3.71399760e-01 -1.82730518e-02 6.14743948e-01
-2.05869079e-01 9.99683887e-02 -9.77520347e-02 -5.58603942e-01
1.01748720e-01 7.68482834e-02 4.49459665e-02 4.80273515e-01
6.21106923e-01 1.96434721e-01 -6.15881979e-01 -3.91028821e-01
8.50281939e-02 6.85339212e-01 -8.69965613e-01 2.71361470e-01
3.49503845e-01 2.08071187e-01 -4.52599317e-01 7.36739457e-01
3.95397305e-01 2.33060300e-01 1.13203466e-01 8.34441558e-02
1.99781969e-01 6.70382440e-01 -8.82431209e-01 1.37432289e+00
-2.49658495e-01 2.99749255e-01 1.85012668e-01 -1.44059110e+00
1.06563270e+00 8.15738499e-01 6.83280110e-01 -5.53640902e-01
3.54848891e-01 1.62458986e-01 1.25517756e-01 -9.72252786e-01
-4.68418971e-02 -1.37100995e-01 -2.84194767e-01 1.60060540e-01
3.59605700e-01 -2.46047139e-01 2.27371261e-01 1.66501299e-01
1.35796821e+00 -7.21659362e-01 4.92486149e-01 3.86784002e-02
6.77477360e-01 -3.91290575e-01 5.09362400e-01 2.88009077e-01
-6.59449339e-01 6.46277905e-01 8.07453692e-01 -1.34540930e-01
-5.28558910e-01 -5.57015836e-01 -4.03523773e-01 8.97975981e-01
-9.39540327e-01 -4.92572010e-01 -9.56538379e-01 -3.12344104e-01
-1.85626507e-01 4.08602893e-01 -6.48088515e-01 -5.34267187e-01
6.74726442e-02 -1.17094338e+00 1.17562306e+00 2.54258156e-01
2.70633072e-01 -1.07190120e+00 -4.52747136e-01 2.72499859e-01
-4.66058582e-01 -1.13752162e+00 3.17307681e-01 2.37072766e-01
-6.14877462e-01 -9.39013720e-01 -6.29976749e-01 -7.20242918e-01
5.16391322e-02 -2.49270737e-01 9.53613758e-01 -1.28726155e-01
-5.51918387e-01 6.94227457e-01 -5.86009383e-01 -6.11803949e-01
-5.36498964e-01 1.57874346e-01 3.17361861e-01 1.38241231e-01
5.15670598e-01 -8.95162463e-01 -3.06150675e-01 -2.34646380e-01
-6.51780009e-01 -4.33071971e-01 3.47398907e-01 3.12656522e-01
1.18371613e-01 9.89983678e-02 1.29562652e+00 -6.49658680e-01
8.57836962e-01 -7.09918916e-01 3.15837771e-01 -1.75115958e-01
-2.42358446e-01 -4.07647580e-01 1.67196214e-01 -2.77134925e-01
-9.30157006e-01 4.15353104e-02 -8.85783315e-01 2.82401830e-01
-7.00450599e-01 5.63125789e-01 -3.18354219e-01 3.67435664e-01
6.07224584e-01 -1.63063765e-01 -1.13878965e-01 -7.15397596e-01
-3.30314934e-02 1.35063767e+00 5.94025314e-01 -1.75396219e-01
-2.34327130e-02 2.38807842e-01 -2.46022955e-01 -1.33403659e+00
-7.74963796e-01 -7.28556156e-01 -5.99787533e-01 -1.96255893e-02
8.04340661e-01 -8.89324129e-01 -2.85758644e-01 8.55598390e-01
-1.01474106e+00 2.44694665e-01 4.91528064e-02 4.88613009e-01
-3.64168674e-01 1.60199016e-01 -5.56208968e-01 -1.22706592e+00
-5.87731004e-01 -6.44978344e-01 1.30445659e+00 -2.84842223e-01
-7.93852925e-01 -8.65400076e-01 6.87162399e-01 5.31656504e-01
2.21356913e-01 3.90516996e-01 1.08420968e+00 -1.03352928e+00
8.59563708e-01 -4.85954463e-01 3.57832275e-02 6.21671796e-01
3.17673832e-01 8.88172090e-02 -1.26923776e+00 6.54060915e-02
1.76409841e-01 -7.64213502e-01 7.30796397e-01 4.67169821e-01
7.38839447e-01 -1.24270350e-01 -2.24233851e-01 2.25191843e-02
8.63760173e-01 3.57733935e-01 4.75340694e-01 -4.65870947e-02
3.69638056e-01 8.36073339e-01 2.71390676e-01 9.30806339e-01
4.51415777e-01 5.46911001e-01 3.09056759e-01 -8.95597413e-02
4.32905694e-03 2.05957592e-01 3.61444384e-01 1.06511807e+00
1.86445743e-01 -3.58937122e-02 -1.24490166e+00 9.09054160e-01
-1.25465822e+00 -9.52989578e-01 -1.20184563e-01 1.79286933e+00
9.79017675e-01 1.04219858e-02 6.18719637e-01 1.06654668e+00
4.54229504e-01 -2.17802167e-01 -6.02702238e-02 -5.98182201e-01
-5.82841188e-02 3.04509342e-01 -5.15687704e-01 7.39729553e-02
-1.12163448e+00 3.99466783e-01 6.55997658e+00 4.87063348e-01
-1.01980543e+00 1.64888188e-01 7.31909156e-01 -2.42038712e-01
1.31367177e-01 -8.41406226e-01 -5.54946840e-01 3.44307631e-01
1.66876781e+00 1.07267544e-01 1.01904236e-01 6.62305474e-01
5.94035566e-01 -5.11877835e-02 -9.16766405e-01 9.88092721e-01
3.14995110e-01 -5.62608242e-01 -2.78928131e-01 -1.15935072e-01
2.85574853e-01 -1.02188848e-01 1.68910891e-01 1.71947002e-01
-2.55498648e-01 -7.97229230e-01 4.04033035e-01 4.00706470e-01
7.32656002e-01 -1.14573836e+00 1.06355166e+00 1.97483510e-01
-7.12529421e-01 -3.34089071e-01 -1.10577345e-01 -2.31899098e-01
-1.84652191e-02 9.47998881e-01 -1.07029259e+00 2.52512664e-01
8.66420269e-01 8.91519547e-01 -6.88313663e-01 7.82327533e-01
1.54596418e-01 8.86910319e-01 -2.34291881e-01 5.19584231e-02
-1.39630377e-01 1.82987630e-01 1.19268328e-01 1.42679501e+00
5.04247665e-01 1.34081468e-01 -2.48981640e-01 3.97483647e-01
1.93614796e-01 3.91777664e-01 -6.97326899e-01 -5.54125488e-01
3.28167319e-01 1.27463675e+00 -4.11992222e-01 -1.87246993e-01
-3.38956147e-01 7.36381114e-01 2.44722217e-01 -1.94225192e-01
-5.24944365e-01 -5.19697607e-01 6.31941319e-01 -4.56359275e-02
-9.10399780e-02 3.59115422e-01 -1.18544787e-01 -8.61321568e-01
-2.09005907e-01 -1.12726045e+00 3.73177797e-01 -5.78966439e-01
-1.13351715e+00 5.90079784e-01 -2.40091667e-01 -8.07051241e-01
-6.28279865e-01 -3.47966015e-01 -4.15996760e-01 6.21035695e-01
-9.79857385e-01 -8.91005456e-01 7.57998030e-04 5.99463820e-01
3.94854754e-01 -2.66313732e-01 1.41701651e+00 8.34271431e-01
-9.74140286e-01 4.63706970e-01 -1.12801790e-01 4.79609668e-02
7.67862439e-01 -9.30375457e-01 -2.81357509e-03 3.25631738e-01
-6.97382167e-02 2.07414448e-01 5.34203947e-01 -4.38986570e-01
-8.57488811e-01 -9.60830092e-01 1.26741040e+00 -4.14821208e-01
7.44366705e-01 -3.64954054e-01 -8.99443924e-01 2.04127848e-01
-8.07486027e-02 -5.89454174e-01 1.11860991e+00 2.13870585e-01
-6.75545782e-02 -1.90879866e-01 -1.56961966e+00 -8.44435617e-02
6.87517166e-01 -7.56739318e-01 -6.73989415e-01 3.61865431e-01
4.62968528e-01 1.70343429e-01 -8.38302255e-01 3.97479236e-01
5.86646616e-01 -1.17324388e+00 1.06415677e+00 -5.32217443e-01
6.49003565e-01 6.43303156e-01 -2.91171789e-01 -1.35332477e+00
-1.98513031e-01 -3.75281662e-01 1.41847301e-02 1.82792437e+00
3.62831295e-01 -3.55714619e-01 4.30630684e-01 4.74359989e-01
-7.40552694e-02 -5.99710643e-01 -9.27617371e-01 -8.78190473e-02
-3.73063348e-02 -7.78512597e-01 2.97001541e-01 7.81854391e-01
6.12886429e-01 4.49346364e-01 -1.55359313e-01 1.24115668e-01
6.44880608e-02 -7.02594578e-01 2.35526547e-01 -1.65005541e+00
-3.37634757e-02 -1.95180088e-01 -7.35526443e-01 3.52246523e-01
2.51125753e-01 -5.74607491e-01 -2.59298295e-01 -1.51517987e+00
3.92424554e-01 -9.80798081e-02 -1.91558406e-01 7.38378406e-01
-2.93555055e-02 4.43510324e-01 -3.20549369e-01 -3.57831061e-01
-3.28653902e-01 3.83721143e-01 4.61608320e-01 -1.26569331e-01
-2.85371691e-01 1.66616127e-01 -9.69041705e-01 1.07164288e+00
1.07185006e+00 -5.19088924e-01 -5.94922423e-01 -1.41894922e-01
2.25660339e-01 1.88967228e-01 3.96912098e-02 -1.16947842e+00
-2.51994967e-01 1.23728171e-01 2.46273801e-01 -4.58377063e-01
8.98086488e-01 -5.31278312e-01 -1.91872567e-01 5.59691310e-01
-3.82523537e-01 -6.60997331e-02 2.44505852e-01 1.49678022e-01
-3.06909412e-01 -2.09492803e-01 7.06193566e-01 2.12706700e-01
-1.34621516e-01 -1.96015388e-01 -8.14054608e-01 1.74901575e-01
6.16297841e-01 2.52616853e-01 -9.99896973e-02 -6.84131145e-01
-8.35086644e-01 -1.89428508e-01 -2.85237372e-01 2.72881955e-01
5.26199579e-01 -9.06024098e-01 -7.07493365e-01 1.70479298e-01
1.01300575e-01 -5.89992642e-01 3.52117836e-01 1.12062681e+00
-1.41575620e-01 6.18328452e-01 -2.58995384e-01 -1.57588601e-01
-1.65003455e+00 3.90542835e-01 2.19696820e-01 -1.81728736e-01
-3.63070905e-01 9.35068846e-01 -2.00003996e-01 -4.39835370e-01
6.77734733e-01 -5.82009494e-01 -7.44845390e-01 6.43532515e-01
9.79355693e-01 5.45127690e-01 4.70648438e-01 -7.00172246e-01
-6.37982249e-01 1.87301368e-01 1.03712946e-01 -2.68804014e-01
1.63597941e+00 -2.14578032e-01 -2.29441285e-01 5.15952647e-01
1.03618371e+00 1.22216955e-01 -1.66839600e-01 2.44891956e-01
2.06922963e-01 1.55841723e-01 2.23357514e-01 -9.15292084e-01
-8.42388511e-01 1.23866892e+00 9.22799170e-01 6.12828255e-01
1.40622115e+00 9.62476805e-03 6.13688707e-01 4.49885875e-01
4.35334183e-02 -8.70657802e-01 2.00505167e-01 2.23882094e-01
7.71669626e-01 -1.12178349e+00 -3.75944287e-01 -3.00582111e-01
-6.21625125e-01 8.58261108e-01 1.02328800e-01 9.48507860e-02
6.66599214e-01 3.76417905e-01 2.68896610e-01 -2.35640109e-01
-9.16177571e-01 -2.51947105e-01 1.93619847e-01 1.12638223e+00
6.55811787e-01 -1.85654853e-02 -3.65455061e-01 1.48225605e+00
-4.04175967e-01 -6.90286383e-02 3.58324945e-01 6.04622960e-01
-6.01051807e-01 -1.12018907e+00 -3.92125130e-01 6.49972260e-01
-1.05554426e+00 -6.64710701e-02 -9.59508121e-01 4.13241655e-01
3.52599800e-01 1.37517583e+00 -1.06531464e-01 -3.50335360e-01
4.05549109e-01 3.32397163e-01 1.55666724e-01 -8.02248716e-01
-6.47399068e-01 1.64768204e-01 6.15572512e-01 -3.33543032e-01
-6.52351201e-01 -9.12875831e-01 -1.01807344e+00 1.00080818e-01
5.58954431e-04 -2.24099960e-02 5.88822126e-01 1.00527918e+00
3.78991246e-01 7.14855909e-01 5.01447976e-01 -7.67004550e-01
-4.20018226e-01 -1.30084133e+00 -8.50808084e-01 1.56315461e-01
4.74212557e-01 -5.09479523e-01 -3.49191576e-01 1.17688067e-01] | [13.734728813171387, 4.928059101104736] |
6f5c2bf7-f932-4366-931d-e1fd851c5aa3 | the-devil-is-in-the-channels-mutual-channel | 2002.04264 | null | https://arxiv.org/abs/2002.04264v3 | https://arxiv.org/pdf/2002.04264v3.pdf | The Devil is in the Channels: Mutual-Channel Loss for Fine-Grained Image Classification | Key for solving fine-grained image categorization is finding discriminate and local regions that correspond to subtle visual traits. Great strides have been made, with complex networks designed specifically to learn part-level discriminate feature representations. In this paper, we show it is possible to cultivate subtle details without the need for overly complicated network designs or training mechanisms -- a single loss is all it takes. The main trick lies with how we delve into individual feature channels early on, as opposed to the convention of starting from a consolidated feature map. The proposed loss function, termed as mutual-channel loss (MC-Loss), consists of two channel-specific components: a discriminality component and a diversity component. The discriminality component forces all feature channels belonging to the same class to be discriminative, through a novel channel-wise attention mechanism. The diversity component additionally constraints channels so that they become mutually exclusive on spatial-wise. The end result is therefore a set of feature channels that each reflects different locally discriminative regions for a specific class. The MC-Loss can be trained end-to-end, without the need for any bounding-box/part annotations, and yields highly discriminative regions during inference. Experimental results show our MC-Loss when implemented on top of common base networks can achieve state-of-the-art performance on all four fine-grained categorization datasets (CUB-Birds, FGVC-Aircraft, Flowers-102, and Stanford-Cars). Ablative studies further demonstrate the superiority of MC-Loss when compared with other recently proposed general-purpose losses for visual classification, on two different base networks. Code available at https://github.com/dongliangchang/Mutual-Channel-Loss | ['Yi-Zhe Song', 'Yifeng Ding', 'Jiyang Xie', 'Ayan Kumar Bhunia', 'Zhanyu Ma', 'Xiaoxu Li', 'Dongliang Chang', 'Ming Wu', 'Jun Guo'] | 2020-02-11 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [ 1.54977337e-01 -2.61647433e-01 -1.91145942e-01 -6.48953199e-01
-7.02190220e-01 -6.87362850e-01 4.78254616e-01 1.26220852e-01
-4.41955298e-01 6.20531023e-01 2.33194344e-02 -1.88335568e-01
-1.80750623e-01 -6.67775154e-01 -7.54136264e-01 -7.25340664e-01
-1.83202505e-01 2.39442587e-02 2.24078983e-01 -7.46816257e-03
1.49300992e-01 7.15976298e-01 -1.55444515e+00 5.50283194e-01
5.11609972e-01 1.53565037e+00 1.08363040e-01 5.22471011e-01
4.05982435e-02 4.54673260e-01 -2.70112783e-01 -2.82198459e-01
4.77125734e-01 -3.68008792e-01 -7.49970675e-01 -4.25333343e-03
8.19094896e-01 -1.57166451e-01 -1.81298882e-01 9.75136161e-01
4.06781882e-01 -6.19111210e-02 9.06787694e-01 -1.22723246e+00
-7.51142323e-01 3.24741960e-01 -7.54426777e-01 2.21667275e-01
-1.62612185e-01 1.64829612e-01 1.24737620e+00 -8.22009087e-01
3.61405075e-01 1.22975779e+00 8.77752066e-01 5.33212245e-01
-1.51849520e+00 -7.85214782e-01 4.67377394e-01 3.58841382e-02
-1.51988947e+00 -3.03486049e-01 5.43311179e-01 -5.65317154e-01
8.65604222e-01 2.44011968e-01 4.76064682e-01 8.82368088e-01
3.09316784e-01 5.34984231e-01 1.22095025e+00 -3.04265469e-01
5.53954989e-02 4.74313013e-02 1.70310810e-01 9.95832264e-01
1.70807213e-01 2.08610401e-01 -3.09747785e-01 9.02073681e-02
8.08021367e-01 1.50309265e-01 -1.87016100e-01 -7.73789465e-01
-8.88131142e-01 8.75435114e-01 8.90194297e-01 2.38590240e-01
-3.37525383e-02 3.52511585e-01 4.20032412e-01 3.04538041e-01
4.20290381e-01 3.07966888e-01 -5.62538028e-01 9.40760300e-02
-8.79143655e-01 2.37391233e-01 5.06842136e-01 7.10052550e-01
9.75254834e-01 -2.73468226e-01 -4.41385657e-01 8.19577634e-01
2.45106310e-01 6.83336183e-02 3.20972353e-01 -8.27642262e-01
2.52941728e-01 5.55198133e-01 -2.34492168e-01 -9.91938412e-01
-4.49976265e-01 -8.69106710e-01 -9.21055734e-01 4.84079659e-01
6.04765296e-01 2.90628988e-02 -1.09108448e+00 2.07398510e+00
1.75745234e-01 -2.08423033e-01 -6.13425016e-01 1.06285059e+00
5.58370829e-01 3.02426606e-01 2.10546494e-01 4.01485056e-01
1.35266078e+00 -9.12229538e-01 -1.39662148e-02 -2.15429768e-01
3.40607673e-01 -4.70447332e-01 1.23499894e+00 3.23732793e-01
-8.78591895e-01 -7.03410327e-01 -1.14363420e+00 -1.73752442e-01
-6.42199218e-01 1.03540257e-01 8.60986173e-01 5.45132339e-01
-1.10358357e+00 6.16180360e-01 -5.04822969e-01 -1.76282927e-01
8.62271726e-01 3.03902656e-01 -5.34241915e-01 5.49753429e-03
-7.45430350e-01 6.49057865e-01 9.16335508e-02 7.20701367e-02
-1.11587298e+00 -8.22739422e-01 -8.33435118e-01 3.27532172e-01
1.40144199e-01 -7.65014827e-01 9.63001430e-01 -1.23211336e+00
-1.04281080e+00 1.00335658e+00 -1.38817325e-01 -3.66413713e-01
5.20957589e-01 -2.58538295e-02 -8.56589377e-02 5.24404682e-02
3.39019299e-01 1.01253402e+00 9.34966683e-01 -1.37523258e+00
-6.64149344e-01 -3.96966070e-01 1.96312249e-01 -5.88064417e-02
-1.50565192e-01 -1.64722636e-01 -4.81887519e-01 -8.29685926e-01
-1.34727299e-01 -7.48621225e-01 -1.87818140e-01 4.99526203e-01
-5.50495327e-01 -2.21295357e-01 5.06931543e-01 -3.20251346e-01
9.75718915e-01 -2.27434134e+00 7.29566514e-02 3.17857623e-01
2.83264965e-01 2.81983986e-02 -3.16481233e-01 2.33129755e-01
-2.01173782e-01 2.44018704e-01 -4.04053003e-01 -1.86790690e-01
1.50125846e-01 -1.35265142e-01 -3.59399796e-01 5.52614629e-01
5.37323177e-01 9.48386133e-01 -5.36846697e-01 -2.33833462e-01
1.82816982e-01 5.15457749e-01 -7.74480104e-01 -1.40600026e-01
-5.84550165e-02 3.11474472e-01 -3.86718392e-01 7.25549042e-01
7.09224880e-01 -3.41677576e-01 -5.49748056e-02 -2.94137686e-01
-9.83111933e-02 2.69493777e-02 -1.00274920e+00 1.84488833e+00
-5.02994359e-01 4.81963754e-01 2.66929269e-01 -1.11912167e+00
6.61881268e-01 -2.25257531e-01 -8.61796737e-02 -8.51669669e-01
2.59000212e-01 4.78670485e-02 6.24111518e-02 -2.03195028e-02
1.71843708e-01 -3.45928699e-01 -4.22181904e-01 1.89702570e-01
3.25451165e-01 1.80172533e-01 6.32347539e-02 8.34383667e-02
8.65285695e-01 1.39736280e-01 2.73896277e-01 -4.34819251e-01
3.59891981e-01 -1.36505336e-01 5.39119482e-01 7.97971547e-01
-3.65198702e-01 6.78272843e-01 6.28055990e-01 -3.52886856e-01
-8.06291342e-01 -1.10378146e+00 -4.16053504e-01 1.42447770e+00
2.31746659e-01 -1.65420771e-01 -5.66608131e-01 -8.33613813e-01
4.89424884e-01 2.98403740e-01 -1.12470734e+00 -2.70045191e-01
-2.38719359e-01 -4.03170079e-01 6.81655586e-01 7.72504747e-01
4.98553097e-01 -9.53786135e-01 -6.12222373e-01 -6.34320974e-02
2.24486604e-01 -6.93745196e-01 -5.10307729e-01 7.23362088e-01
-6.34379923e-01 -9.51308668e-01 -6.73513591e-01 -8.40547740e-01
6.57718420e-01 3.24076951e-01 1.05561316e+00 7.56047443e-02
-7.42604673e-01 1.33155197e-01 -1.73300266e-01 -1.71719298e-01
2.60751933e-01 7.24586323e-02 -2.47953370e-01 1.08434461e-01
1.80802241e-01 -5.85254014e-01 -7.93647766e-01 3.95210236e-01
-6.78288221e-01 -1.95203692e-01 7.39451230e-01 1.08422577e+00
8.08499396e-01 -4.04509753e-02 6.36614382e-01 -7.47648060e-01
3.26067418e-01 -3.37959021e-01 -4.61669236e-01 1.70208052e-01
-3.32923204e-01 -1.36967804e-02 8.04928839e-01 -2.32898459e-01
-7.51418173e-01 3.18288691e-02 -1.73891470e-01 -4.92970109e-01
-5.10386884e-01 2.75694251e-01 -1.86816737e-01 -1.87488511e-01
5.67013621e-01 1.39834598e-01 -1.27655461e-01 -5.45265794e-01
6.07117295e-01 4.64361459e-01 4.40981686e-01 -6.34234488e-01
7.32014596e-01 5.90705276e-01 7.06660450e-02 -6.65030241e-01
-9.05311346e-01 -4.53870595e-01 -6.80617511e-01 8.39382187e-02
9.68439102e-01 -9.70285594e-01 -7.34793901e-01 5.39815009e-01
-7.57047474e-01 -5.56081235e-01 -3.65739226e-01 1.13653071e-01
-3.90886515e-01 1.71130463e-01 -5.49570143e-01 -3.37589920e-01
-1.24896958e-01 -9.56390560e-01 1.05903935e+00 3.29168767e-01
-3.33120450e-02 -7.69297242e-01 -1.38388336e-01 1.53814718e-01
5.09099543e-01 1.67594686e-01 1.11432636e+00 -3.47032040e-01
-5.22276282e-01 -1.91960573e-01 -7.96885729e-01 4.55404371e-01
-4.81628068e-02 -1.79124832e-01 -1.19436765e+00 -4.81592864e-01
-5.00461936e-01 -7.32532680e-01 1.43774569e+00 4.68660474e-01
1.78097534e+00 -1.60483137e-01 -3.17012280e-01 1.07456470e+00
1.60842085e+00 4.40006293e-02 3.80669862e-01 2.23221600e-01
7.60633528e-01 5.08896768e-01 3.32290292e-01 3.13938409e-01
2.64247119e-01 7.16149151e-01 5.07726252e-01 -3.03237081e-01
-3.81758869e-01 -2.69251496e-01 -6.18198961e-02 7.88831413e-02
1.24776430e-01 -2.20716089e-01 -4.64484036e-01 6.39323533e-01
-1.60346639e+00 -9.05452371e-01 2.55451024e-01 2.02462482e+00
7.79342651e-01 1.46243110e-01 1.69298351e-01 -8.06058943e-02
4.27731037e-01 2.39143088e-01 -7.26746500e-01 -5.94664335e-01
-1.13058299e-01 3.94779086e-01 4.52288181e-01 4.19905216e-01
-1.39297330e+00 9.22240973e-01 5.27248383e+00 1.12964332e+00
-1.39402497e+00 1.15180060e-01 8.88824642e-01 -1.37373656e-01
-1.54112011e-01 -1.54516369e-01 -8.72191668e-01 2.49151945e-01
4.44833547e-01 3.30727100e-01 4.41693217e-01 8.54375780e-01
-1.51068032e-01 -2.10459456e-02 -1.07243645e+00 8.19030941e-01
-3.00239213e-02 -1.25806379e+00 9.73579139e-02 -8.93298129e-04
4.19368178e-01 3.68966558e-03 2.75139213e-01 3.34137470e-01
8.93616080e-02 -1.26176107e+00 1.15375555e+00 2.91310519e-01
1.21663094e+00 -7.19899714e-01 4.49140728e-01 7.64857084e-02
-1.34010732e+00 -2.42284864e-01 -5.71304500e-01 -1.00282736e-01
-2.15749353e-01 5.16614556e-01 -4.29884404e-01 3.17568719e-01
9.17603254e-01 5.42643964e-01 -6.22375071e-01 1.07353926e+00
-1.02044083e-01 4.65918094e-01 1.04768332e-02 5.42837381e-02
5.02495885e-01 1.84822515e-01 2.51031995e-01 1.51910985e+00
-3.71376909e-02 -2.79046655e-01 2.35274464e-01 1.07878315e+00
-1.46081433e-01 -1.84890091e-01 -4.47256297e-01 1.10963851e-01
2.17546806e-01 1.57366836e+00 -8.10672939e-01 -7.11997077e-02
-4.12450343e-01 1.14579535e+00 7.14357615e-01 3.24104905e-01
-8.61392796e-01 -6.34757102e-01 9.38576460e-01 1.44591987e-01
8.40702653e-01 2.63113081e-02 -5.09495020e-01 -9.36910808e-01
-3.67175974e-02 -6.90317869e-01 3.91094714e-01 -4.29370105e-01
-1.68489301e+00 6.46438539e-01 -2.15943277e-01 -1.03772044e+00
1.76310301e-01 -8.50468397e-01 -5.89017451e-01 1.05172384e+00
-1.65992010e+00 -1.42620981e+00 -4.19839472e-01 7.72823155e-01
3.78333390e-01 9.63628739e-02 7.84118593e-01 3.60872477e-01
-5.05708992e-01 1.02504158e+00 -6.62625358e-02 2.78121054e-01
7.39816248e-01 -1.29749596e+00 1.04703926e-01 7.70506084e-01
1.10722810e-01 6.66835785e-01 2.47766718e-01 -2.20648825e-01
-9.08742189e-01 -1.25351286e+00 7.26295948e-01 -2.84608155e-01
4.89480793e-01 -7.49891281e-01 -7.31246769e-01 5.13606966e-01
-2.59234086e-02 3.96864533e-01 5.76342940e-01 2.67840594e-01
-9.58739221e-01 -3.31769496e-01 -1.17973137e+00 3.08035016e-01
1.20476496e+00 -6.48449183e-01 -3.07796121e-01 -2.51693483e-02
5.48920572e-01 2.45708041e-02 -6.46100938e-01 4.08077985e-01
8.73890460e-01 -1.17239749e+00 1.03571117e+00 -7.39272535e-01
4.57086474e-01 -4.40677375e-01 -3.05362016e-01 -1.23487854e+00
-9.26924765e-01 -1.17816731e-01 3.24497104e-01 1.18266952e+00
5.38918078e-01 -6.80069447e-01 5.80790102e-01 1.64972961e-01
-3.26507390e-01 -1.02945888e+00 -9.05782521e-01 -8.40341389e-01
3.73991132e-01 -3.95907581e-01 3.98600876e-01 7.91441202e-01
-2.61816502e-01 2.80355960e-01 -1.25933424e-01 5.25515527e-02
5.53067088e-01 4.29371327e-01 5.95663130e-01 -1.17439377e+00
-5.22937357e-01 -8.76630187e-01 -4.14820194e-01 -1.09352911e+00
-4.74973656e-02 -1.05103862e+00 1.58445776e-01 -1.34549415e+00
5.03004313e-01 -7.79579520e-01 -6.07382536e-01 8.56332183e-01
-3.61625515e-02 6.87313616e-01 3.40230644e-01 -9.34690516e-03
-6.73275709e-01 4.51113611e-01 1.24810278e+00 -2.58308381e-01
-4.42906097e-02 -2.24868193e-01 -1.13108015e+00 6.56733155e-01
6.68725610e-01 -3.58952403e-01 -2.67782956e-01 -2.55398422e-01
-1.00586846e-01 -3.49323750e-01 8.55390549e-01 -9.83790815e-01
-9.90638603e-03 9.93229225e-02 9.27949369e-01 -2.78181195e-01
3.64930630e-01 -7.01146543e-01 -5.27410656e-02 4.11307782e-01
-4.47443038e-01 -1.38750389e-01 3.89375955e-01 5.96239686e-01
-2.00890243e-01 1.20127223e-01 1.15164173e+00 -1.83829237e-02
-1.00573993e+00 3.68647486e-01 -6.87611178e-02 2.01664165e-01
1.08087778e+00 -3.32683474e-01 -4.58058864e-01 -1.44337401e-01
-5.83019733e-01 4.66505252e-03 5.13616145e-01 4.75883961e-01
3.50456446e-01 -1.22660816e+00 -6.07686639e-01 4.53368306e-01
3.36180478e-01 -2.13909760e-01 4.80487227e-01 7.57323325e-01
-3.24791551e-01 5.55496156e-01 -5.29211521e-01 -6.70401216e-01
-1.06926715e+00 5.94294190e-01 4.68999863e-01 -3.04435462e-01
-4.76647824e-01 1.38164485e+00 9.32631493e-01 -4.15390372e-01
3.26567352e-01 -5.04603803e-01 4.30042706e-02 7.07143471e-02
3.95582974e-01 2.56592557e-02 6.03685714e-02 -5.68514705e-01
-6.31915331e-01 7.91962862e-01 -9.78063047e-02 1.49438843e-01
1.24979603e+00 -5.67514747e-02 -9.57749598e-03 1.98971033e-01
1.39846730e+00 -4.88535166e-02 -1.54088545e+00 -4.70904447e-02
-4.01729316e-01 -5.34564316e-01 1.17146559e-01 -1.32031834e+00
-1.18056822e+00 1.07108116e+00 8.43064249e-01 1.09234542e-01
1.26434219e+00 2.87616640e-01 4.18898940e-01 -9.22096521e-02
3.90615672e-01 -6.58467889e-01 -9.51244533e-02 5.37057936e-01
8.94829988e-01 -1.11399925e+00 -2.58268446e-01 -2.98990846e-01
-5.17946720e-01 8.25018108e-01 6.65921688e-01 -3.01525712e-01
7.71331012e-01 1.31349683e-01 -5.38947023e-02 -2.08094344e-01
-6.60487056e-01 -4.01863366e-01 6.06651902e-01 5.63883603e-01
4.91145581e-01 2.68071175e-01 -7.31479749e-02 7.84813285e-01
4.47208248e-02 -1.30498335e-01 -1.56216562e-01 6.95578337e-01
-4.77902204e-01 -9.00895536e-01 -5.36813289e-02 7.25971997e-01
-5.09567022e-01 -1.60921276e-01 -4.09832746e-01 9.17251050e-01
6.45917296e-01 7.29183376e-01 3.56568456e-01 -4.72429246e-01
2.32121810e-01 -1.00168340e-01 6.28922880e-01 -5.33025146e-01
-6.74741983e-01 -6.18498959e-02 1.32402023e-02 -8.21727037e-01
-1.40058577e-01 -5.05357921e-01 -1.07417357e+00 -2.62918234e-01
-2.28274912e-01 -1.20582744e-01 5.06782770e-01 5.78360975e-01
4.90760148e-01 6.04132235e-01 4.97386426e-01 -1.06370020e+00
-4.76363301e-01 -7.48192847e-01 -7.29761183e-01 3.58301431e-01
5.10238647e-01 -8.01881552e-01 -3.44917774e-01 -2.98719317e-01] | [9.56346321105957, 2.0897889137268066] |
77e67339-bf4f-4eb9-8a2c-97c88f8efaaa | active-learning-for-interactive-relation | null | null | https://aclanthology.org/2021.ranlp-main.101 | https://aclanthology.org/2021.ranlp-main.101.pdf | Active Learning for Interactive Relation Extraction in a French Newspaper’s Articles | Relation extraction is a subtask of natural langage processing that has seen many improvements in recent years, with the advent of complex pre-trained architectures. Many of these state-of-the-art approaches are tested against benchmarks with labelled sentences containing tagged entities, and require important pre-training and fine-tuning on task-specific data. However, in a real use-case scenario such as in a newspaper company mostly dedicated to local information, relations are of varied, highly specific type, with virtually no annotated data for such relations, and many entities co-occur in a sentence without being related. We question the use of supervised state-of-the-art models in such a context, where resources such as time, computing power and human annotators are limited. To adapt to these constraints, we experiment with an active-learning based relation extraction pipeline, consisting of a binary LSTM-based lightweight model for detecting the relations that do exist, and a state-of-the-art model for relation classification. We compare several choices for classification models in this scenario, from basic word embedding averaging, to graph neural networks and Bert-based ones, as well as several active learning acquisition strategies, in order to find the most cost-efficient yet accurate approach in our French largest daily newspaper company’s use case. | ['Pascale Sébillot', 'Guillaume Gravier', 'Michel Le Nouy', 'Cyrielle Mallart'] | null | null | https://aclanthology.org/2021.ranlp-1.101 | https://aclanthology.org/2021.ranlp-1.101.pdf | ranlp-2021-9 | ['relation-classification'] | ['natural-language-processing'] | [ 2.22199187e-01 8.13865423e-01 -3.81730050e-01 -3.39626998e-01
-5.97471654e-01 -4.14030463e-01 8.32939386e-01 9.91079628e-01
-8.76467586e-01 1.01419938e+00 1.95662111e-01 -4.23879653e-01
-3.01622540e-01 -1.15556300e+00 -4.41387117e-01 -3.22998196e-01
-5.05557775e-01 9.57710624e-01 5.23345053e-01 -6.10290825e-01
-1.50875270e-01 4.71330702e-01 -1.18923473e+00 3.03534180e-01
6.23937488e-01 9.40828204e-01 2.51538083e-02 6.90651536e-01
-4.69728023e-01 9.65351820e-01 -5.72862387e-01 -7.76104271e-01
-4.34130840e-02 -7.89481029e-02 -1.34357524e+00 -1.78913459e-01
-4.47720140e-02 2.80041486e-01 -3.36052030e-01 8.50018322e-01
3.70098770e-01 2.13691860e-01 5.46119273e-01 -7.52111554e-01
-4.46385622e-01 1.10791683e+00 -1.06572807e-02 4.96360540e-01
5.60708225e-01 -1.99838206e-01 1.32612085e+00 -6.88235819e-01
1.04273665e+00 8.81632984e-01 6.50058210e-01 2.93128490e-01
-1.23471534e+00 -1.06366239e-01 3.47663939e-01 4.88091439e-01
-1.07016993e+00 -7.21530735e-01 6.11460686e-01 -3.59506279e-01
1.78581023e+00 1.97812557e-01 4.89884913e-01 1.22867191e+00
7.16848182e-04 5.65146863e-01 7.34472156e-01 -9.17639792e-01
2.13878289e-01 5.06693065e-01 4.41049784e-01 5.61651468e-01
5.19405186e-01 -2.77152956e-01 -6.20732486e-01 -1.77196965e-01
3.02000642e-01 -4.05833453e-01 -3.13873082e-01 -4.62257445e-01
-1.07135212e+00 7.39549458e-01 2.59901166e-01 1.10303140e+00
-5.05103588e-01 -4.13985163e-01 5.32024920e-01 4.79752958e-01
9.77476001e-01 8.13465059e-01 -9.05464649e-01 -8.32986981e-02
-9.09619808e-01 1.24568082e-01 1.49701858e+00 9.35115397e-01
6.85985386e-01 -5.02519548e-01 -7.27772564e-02 8.51875186e-01
2.63315052e-01 -2.98542351e-01 3.79159600e-01 3.35412584e-02
9.53575909e-01 9.95081663e-01 -2.70367235e-01 -8.90711308e-01
-6.55102849e-01 -5.34225643e-01 -7.55825162e-01 1.03235524e-02
4.71979052e-01 -3.21862042e-01 -6.90747559e-01 1.30460989e+00
3.80237401e-01 -2.56992817e-01 2.03748867e-01 3.83877814e-01
9.65914965e-01 3.33312720e-01 9.82995108e-02 -5.04941344e-01
1.36726677e+00 -7.76624322e-01 -9.19424951e-01 -6.01960480e-01
1.01200914e+00 -6.07099354e-01 6.05747402e-01 1.66642204e-01
-8.55507910e-01 -2.32097462e-01 -1.16851425e+00 -1.31686658e-01
-1.17564595e+00 -1.71824902e-01 9.13647473e-01 5.72690725e-01
-8.04302216e-01 8.48304391e-01 -7.99974203e-01 -8.83703709e-01
4.87928480e-01 5.01638472e-01 -7.59055436e-01 1.62374407e-01
-1.78772807e+00 1.54676843e+00 1.01496875e+00 2.63443738e-01
-2.23990828e-01 -2.99264878e-01 -1.13999474e+00 1.74955532e-01
1.00426376e+00 -4.63727951e-01 1.01850319e+00 -6.86241448e-01
-1.25432789e+00 1.02916443e+00 5.85172176e-02 -8.32158923e-01
2.47024879e-01 -2.35651284e-01 -7.08946228e-01 -2.70981342e-01
-2.61975020e-01 -5.69844097e-02 4.08447146e-01 -1.03059304e+00
-5.87873936e-01 -2.87717909e-01 4.74874139e-01 1.42276853e-01
-6.05456948e-01 3.45253676e-01 -6.12643659e-02 -2.48629093e-01
-2.25032210e-01 -5.59423864e-01 -3.63251299e-01 -4.90371764e-01
-4.62259501e-01 -5.74055731e-01 8.08404982e-01 -6.33199275e-01
1.37304902e+00 -1.77665770e+00 7.71153346e-02 8.32847729e-02
2.79598862e-01 6.05311632e-01 9.71375480e-02 8.73075128e-01
-2.07041830e-01 1.61577180e-01 -1.76159650e-01 -2.81897396e-01
1.46119846e-02 2.83597618e-01 9.23263952e-02 6.71645403e-02
7.06890225e-01 8.87858391e-01 -1.09934783e+00 -7.83812702e-01
-1.90726612e-02 3.30719918e-01 2.01540381e-01 3.13711315e-01
-2.87032574e-01 1.41191021e-01 -3.36370617e-01 5.01046240e-01
6.83752894e-02 -1.29840508e-01 6.02422774e-01 -7.59040341e-02
-3.56262922e-02 7.83828676e-01 -1.11857355e+00 1.70059204e+00
-6.29989743e-01 5.82594454e-01 -1.29942060e-01 -1.37015998e+00
1.00327206e+00 6.47888899e-01 4.51166570e-01 -5.28329670e-01
2.12722123e-01 3.04832965e-01 3.11445355e-01 -7.01990604e-01
5.90875566e-01 1.82345018e-01 2.19037876e-01 1.06637515e-01
6.42574966e-01 8.82788002e-02 7.58544564e-01 1.54972866e-01
1.61866879e+00 1.93285912e-01 9.80088174e-01 -3.11253611e-02
6.48273826e-01 -6.22480921e-03 4.27162230e-01 4.91290182e-01
5.99580724e-03 3.04068387e-01 7.35962451e-01 -3.72772574e-01
-7.88733184e-01 -2.85250157e-01 -1.74854379e-02 1.05343997e+00
-3.53938669e-01 -5.42762399e-01 -4.90632921e-01 -1.11529672e+00
-4.62983400e-01 7.28757918e-01 -4.90266025e-01 5.66159151e-02
-8.00805151e-01 -1.01521957e+00 3.63858759e-01 1.29873782e-01
2.44909763e-01 -1.39954114e+00 -4.66419548e-01 6.79911792e-01
1.69626415e-01 -1.39578295e+00 3.56947929e-01 9.70447481e-01
-6.78116500e-01 -1.09171498e+00 -4.38385218e-01 -8.12241733e-01
3.26658905e-01 -5.45561194e-01 1.85760880e+00 -6.70711473e-02
-7.40707815e-02 7.90136214e-03 -8.16608012e-01 -6.03159547e-01
-4.07913327e-01 7.79235780e-01 -3.43672663e-01 -3.08362041e-02
6.25869095e-01 -5.01577020e-01 4.55517769e-02 -2.81768471e-01
-6.74501598e-01 -3.33132327e-01 7.30083764e-01 7.54269063e-01
1.61731184e-01 2.54571646e-01 3.19162190e-01 -1.50862193e+00
8.33435357e-01 -5.10591924e-01 -2.93543696e-01 4.89606142e-01
-6.50135875e-01 1.48434818e-01 6.05800509e-01 -5.26182055e-01
-8.33016932e-01 1.55934259e-01 -2.60664493e-01 4.31440562e-01
-3.75206262e-01 9.48755980e-01 -4.31285411e-01 1.02601036e-01
9.46231365e-01 -2.51531512e-01 -4.75267768e-01 -3.32739800e-01
3.95297885e-01 6.33449137e-01 2.05417693e-01 -3.00919533e-01
7.65627384e-01 -9.39272717e-02 -6.37651309e-02 -7.97205150e-01
-1.15993631e+00 -4.91057336e-01 -1.10879576e+00 7.43637905e-02
8.25848341e-01 -5.25741279e-01 -2.49567285e-01 1.85747832e-01
-1.37018168e+00 -4.80357736e-01 -5.49516499e-01 5.34478009e-01
-2.58637100e-01 9.33689177e-02 -6.39966130e-01 -9.26121473e-01
-4.26201046e-01 -7.19258845e-01 6.60037577e-01 1.32715195e-01
-5.85958242e-01 -1.36370647e+00 3.73260766e-01 1.38328448e-01
4.28932250e-01 2.99429864e-01 1.11886466e+00 -1.35030723e+00
-3.27756345e-01 -5.42837143e-01 -4.54050303e-02 1.87250793e-01
1.33276939e-01 -1.06446020e-01 -9.87003624e-01 -2.96056978e-02
-2.61660159e-01 -4.40173537e-01 6.80094242e-01 -2.02344283e-01
5.78316271e-01 -2.91945040e-01 -6.23185396e-01 -4.24441434e-02
1.25801182e+00 1.30033612e-01 4.87763196e-01 3.89378101e-01
7.17529416e-01 1.04669833e+00 5.98961830e-01 -1.00445032e-01
4.99868751e-01 7.23662972e-01 2.94359297e-01 -1.91561684e-01
-5.61432689e-02 2.22560167e-01 1.65823489e-01 9.93819535e-01
-4.47340786e-01 -5.70107162e-01 -1.20256770e+00 7.11886704e-01
-2.13378239e+00 -8.45158696e-01 -2.26658225e-01 2.22387838e+00
1.24033666e+00 8.28822017e-01 1.81911774e-02 5.99879682e-01
5.48948467e-01 3.62026721e-01 -1.70930997e-02 -5.63930809e-01
-1.28295675e-01 5.18254936e-01 5.36982894e-01 3.89675558e-01
-1.48209512e+00 9.12321270e-01 5.25792599e+00 9.29686010e-01
-9.03997421e-01 3.15808415e-01 3.45934391e-01 2.80700356e-01
5.37761487e-02 2.09602326e-01 -7.72838533e-01 1.43471271e-01
1.29039097e+00 1.22572899e-01 5.25079258e-02 7.68430710e-01
-1.36709899e-01 -1.91226870e-01 -1.49007893e+00 6.71183586e-01
4.03388664e-02 -1.20371425e+00 -5.21321356e-01 9.42094326e-02
1.66804418e-01 6.70681149e-02 -6.59719229e-01 5.79433084e-01
3.57945919e-01 -9.45405662e-01 3.88812721e-01 5.63475549e-01
4.09984559e-01 -4.06641364e-01 1.26303577e+00 4.11591917e-01
-1.26720977e+00 1.77933425e-01 -2.85000980e-01 -2.33256042e-01
2.68285573e-01 1.10000741e+00 -8.47960293e-01 9.60274339e-01
4.51742828e-01 5.58870554e-01 -6.52460039e-01 7.10124671e-01
-4.17396873e-01 5.00082433e-01 -3.17669809e-01 -4.76280779e-01
9.46550965e-02 1.05277993e-01 4.56896842e-01 1.55565310e+00
-1.23822287e-01 -1.15271017e-01 6.18012398e-02 5.02481997e-01
-3.84218693e-02 4.20568526e-01 -8.50681365e-01 -3.04740787e-01
3.12380314e-01 1.58874929e+00 -9.58854198e-01 -3.48240018e-01
-6.52948558e-01 7.21563339e-01 8.50445092e-01 1.42146632e-01
-4.26722735e-01 -7.19064713e-01 9.35133174e-02 1.45865604e-01
2.64172822e-01 -3.10383528e-01 -9.60151851e-02 -1.17388785e+00
1.51728660e-01 -7.40142405e-01 3.91646147e-01 -1.65825844e-01
-1.31293118e+00 1.08651829e+00 3.45060639e-02 -8.11949313e-01
-5.73680162e-01 -6.33726239e-01 -4.66480374e-01 6.46817446e-01
-1.37901902e+00 -1.27834761e+00 5.35185933e-02 2.55800247e-01
3.51351440e-01 -4.51382697e-01 1.30424702e+00 6.24829412e-01
-5.25639117e-01 3.87998313e-01 -3.76502812e-01 4.67797011e-01
5.70356846e-01 -1.52623165e+00 6.42724693e-01 8.90131235e-01
8.23176563e-01 5.81511796e-01 7.54926741e-01 -4.98945564e-01
-1.12644029e+00 -8.07607949e-01 1.78373468e+00 -5.32269478e-01
1.14794219e+00 -8.73667240e-01 -1.01983583e+00 6.15638137e-01
4.83255744e-01 2.43345544e-01 5.51980436e-01 8.64619136e-01
-1.23337731e-01 -4.42318432e-02 -8.67301404e-01 4.21080112e-01
1.20534277e+00 -6.14804447e-01 -8.78180563e-01 6.03741944e-01
5.83347678e-01 -3.81807804e-01 -1.09214163e+00 5.39947629e-01
8.11196789e-02 -7.36275017e-01 7.36775517e-01 -6.11186445e-01
3.15390050e-01 1.10139810e-01 3.08858156e-01 -1.37962663e+00
-2.40639508e-01 -6.10063851e-01 -5.74995220e-01 1.64833772e+00
1.22158623e+00 -6.24385655e-01 8.02663326e-01 4.60340768e-01
6.62727654e-02 -1.20423484e+00 -8.11066389e-01 -7.52921402e-01
-2.55915999e-01 -3.46161187e-01 3.02829683e-01 1.08252501e+00
2.74157345e-01 9.54558432e-01 -1.15852710e-02 -1.48605347e-01
3.01459478e-03 -2.64777184e-01 6.21418953e-01 -1.55420768e+00
-3.76913846e-01 -2.66099960e-01 -6.79559052e-01 -3.91018480e-01
3.56477737e-01 -6.84569120e-01 2.04891525e-02 -2.04265189e+00
-2.35669613e-01 -4.43776518e-01 -2.33130798e-01 6.75748706e-01
-6.01227470e-02 -9.24923718e-02 -9.74617749e-02 -1.22665152e-01
-5.54750204e-01 1.63600862e-01 6.67854965e-01 -3.98756504e-01
-2.27723151e-01 3.09638400e-02 -4.48359758e-01 7.41901457e-01
5.19542158e-01 -6.76247299e-01 -3.47471178e-01 -1.48303434e-01
7.70284712e-01 -2.53511727e-01 -8.81458223e-02 -8.86600077e-01
3.92645210e-01 1.11993253e-01 1.18945308e-01 -2.19853804e-01
3.52616966e-01 -1.01720941e+00 -1.30379051e-01 2.30606332e-01
-3.96524638e-01 -2.28579819e-01 -2.08117098e-01 2.69729197e-01
-5.74122727e-01 -6.61857724e-01 3.93529177e-01 -3.69800359e-01
-7.36660063e-01 3.20461333e-01 -1.99814081e-01 1.81709319e-01
9.83910620e-01 -1.44683495e-01 -3.24970752e-01 -1.56370163e-01
-8.52860689e-01 -8.11132565e-02 -1.47202373e-01 4.42024767e-01
8.60803649e-02 -9.40771282e-01 -8.36340964e-01 -1.86912656e-01
3.90907139e-01 3.72893989e-01 -3.08509737e-01 7.81104803e-01
-4.62422997e-01 3.88417155e-01 1.56453684e-01 4.10628654e-02
-1.31064832e+00 6.90100491e-01 1.21028386e-01 -1.13266706e+00
-5.68387270e-01 8.62081945e-01 -6.25950694e-01 -3.68964791e-01
2.22712636e-01 -5.11315227e-01 -8.64269018e-01 7.29161382e-01
1.95647940e-01 -2.10803784e-02 8.18763852e-01 -4.44217473e-01
-5.16054749e-01 1.44967511e-01 -1.48161709e-01 -6.41368032e-02
1.45561302e+00 1.15438379e-01 -2.72949278e-01 8.20515335e-01
9.85484779e-01 6.32294193e-02 -4.12369728e-01 -4.99323428e-01
8.80265057e-01 1.05451122e-01 -1.27274707e-01 -5.86129487e-01
-7.81468391e-01 5.50201178e-01 2.14707419e-01 1.07128644e+00
7.97229409e-01 2.09396318e-01 5.53102911e-01 7.62468517e-01
5.76828897e-01 -1.22734666e+00 -5.07331848e-01 8.27342689e-01
7.48645306e-01 -1.37333155e+00 4.28812176e-01 -6.66626453e-01
-4.79542971e-01 1.15433753e+00 2.10438177e-01 -8.25623944e-02
7.47124612e-01 4.64781553e-01 1.75893605e-01 -3.57361972e-01
-9.28817630e-01 -7.20327556e-01 4.09931928e-01 7.34781861e-01
9.37668204e-01 -5.39498143e-02 -6.55974209e-01 2.71342665e-01
-2.83827871e-01 -2.23069459e-01 2.30413601e-01 1.01271832e+00
-8.99941400e-02 -1.51857734e+00 1.11771509e-01 5.63588977e-01
-6.21648550e-01 -3.80750924e-01 -7.98951805e-01 9.85733688e-01
3.82349819e-01 1.10769331e+00 9.72655602e-03 -2.81625569e-01
5.18381119e-01 3.78256619e-01 5.03269613e-01 -1.19697142e+00
-1.03032291e+00 -3.18253934e-01 1.01394451e+00 -3.40184927e-01
-8.57095182e-01 -4.72554058e-01 -8.80343258e-01 1.53079391e-01
-7.74738133e-01 1.00752190e-01 5.55534184e-01 1.38287282e+00
-4.88205962e-02 7.79850662e-01 1.14928082e-01 -7.27970421e-01
-1.18806444e-01 -1.46734047e+00 -5.16936898e-01 2.93379873e-01
-2.11662315e-02 -5.87177932e-01 -4.74236123e-02 -9.54979286e-02] | [9.310953140258789, 8.75106143951416] |
301a4aa1-42bb-40cb-bf6f-3f914a8f91a7 | reinforcement-learning-for-predicting-traffic | 2212.04677 | null | https://arxiv.org/abs/2212.04677v1 | https://arxiv.org/pdf/2212.04677v1.pdf | Reinforcement Learning for Predicting Traffic Accidents | As the demand for autonomous driving increases, it is paramount to ensure safety. Early accident prediction using deep learning methods for driving safety has recently gained much attention. In this task, early accident prediction and a point prediction of where the drivers should look are determined, with the dashcam video as input. We propose to exploit the double actors and regularized critics (DARC) method, for the first time, on this accident forecasting platform. We derive inspiration from DARC since it is currently a state-of-the-art reinforcement learning (RL) model on continuous action space suitable for accident anticipation. Results show that by utilizing DARC, we can make predictions 5\% earlier on average while improving in multiple metrics of precision compared to existing methods. The results imply that using our RL-based problem formulation could significantly increase the safety of autonomous driving. | ['Dongsoo Har', 'TaeYoung Kim', 'Praveen Kumar Rajendran', 'Injoon Cho'] | 2022-12-09 | null | null | null | null | ['accident-anticipation'] | ['computer-vision'] | [-9.77775455e-02 5.38129508e-01 -3.32792372e-01 -2.69773901e-01
-9.76292610e-01 7.31903389e-02 5.82611978e-01 3.30231078e-02
-7.72364974e-01 5.86481988e-01 2.20414400e-01 -6.85923398e-01
-3.17343771e-01 -5.82980454e-01 -7.14677513e-01 -4.33169246e-01
-1.83468938e-01 1.76724747e-01 5.15759051e-01 -6.61978066e-01
1.22063674e-01 6.13440216e-01 -1.92972279e+00 2.40830630e-01
4.84522730e-01 1.01511073e+00 1.06164902e-01 7.36341119e-01
5.07160544e-01 1.33148670e+00 -2.25996181e-01 -3.74306262e-01
5.33300519e-01 -1.38276681e-01 -6.15919173e-01 -3.54682028e-01
-1.16603673e-01 -5.00475466e-01 -6.88590467e-01 3.26210737e-01
4.15529013e-01 3.73130053e-01 3.23926985e-01 -1.54283357e+00
1.67067721e-01 3.14830482e-01 -2.01885507e-01 1.78379506e-01
2.54873633e-01 6.37894452e-01 6.25959873e-01 -3.48497152e-01
2.98223317e-01 9.35614586e-01 5.89387953e-01 9.13078368e-01
-7.28822708e-01 -4.53455657e-01 1.98550954e-01 8.30492496e-01
-1.02257025e+00 -4.79806006e-01 8.49682331e-01 -4.07636315e-01
1.19180059e+00 1.76335946e-01 8.53904605e-01 1.18286657e+00
5.48039675e-01 1.05616415e+00 8.13001156e-01 -3.46497566e-01
4.69530135e-01 -1.84893608e-01 -1.96768213e-02 5.10562658e-01
-1.42431669e-02 1.02157843e+00 -4.03733373e-01 3.50771040e-01
1.05467193e-01 1.29570678e-01 4.43813205e-01 -1.98029980e-01
-9.09774184e-01 9.73091424e-01 3.00997972e-01 -2.19938546e-01
-7.97431350e-01 6.23955786e-01 5.77762485e-01 8.31815898e-02
3.51012945e-01 2.56637424e-01 -1.47078544e-01 -6.91518426e-01
-7.16781855e-01 6.65773153e-01 5.09637952e-01 6.58320367e-01
5.22308946e-01 1.24811269e-01 -4.59592313e-01 3.68435383e-01
2.38228634e-01 4.73037988e-01 1.81772321e-01 -1.45538294e+00
3.15181583e-01 4.90070134e-01 3.17046016e-01 -6.78056896e-01
-7.44622707e-01 -2.12466419e-01 -3.49768102e-01 8.63013804e-01
1.52538076e-01 -4.73133057e-01 -5.83267629e-01 1.34358895e+00
1.35505021e-01 3.95400107e-01 2.55042315e-01 7.66834795e-01
2.09466040e-01 5.82885087e-01 4.64307398e-01 -1.33848295e-01
1.11741686e+00 -9.53708172e-01 -6.67375147e-01 -5.03989875e-01
8.21908832e-01 -2.92582333e-01 6.07439578e-01 6.89954281e-01
-9.19646323e-01 -8.03263605e-01 -9.78374660e-01 1.20140277e-01
-2.69712359e-01 -4.67332406e-03 4.63949710e-01 5.91023028e-01
-9.49120164e-01 4.08114433e-01 -1.03839254e+00 -1.00833364e-01
3.30650926e-01 4.11023229e-01 -1.86868086e-01 2.42767911e-02
-1.28503752e+00 1.33091688e+00 3.65993708e-01 1.01424292e-01
-1.07962561e+00 -5.28676927e-01 -7.94826269e-01 -1.69621259e-01
6.16630435e-01 -2.48041093e-01 1.58995855e+00 -5.27900100e-01
-1.63885057e+00 3.42594773e-01 -7.75486529e-02 -1.31886435e+00
7.93898046e-01 -6.06951654e-01 -7.55028546e-01 2.19635218e-02
-1.23169407e-01 9.44976807e-01 9.06262457e-01 -1.12542403e+00
-1.17696190e+00 -1.84210744e-02 2.49157146e-01 -2.97523998e-02
1.47848204e-01 -5.00227436e-02 -9.81494188e-02 -1.88047308e-02
-8.75347137e-01 -1.20060778e+00 -6.45798683e-01 -8.59145597e-02
1.48002103e-01 -6.37644172e-01 9.12445605e-01 -7.67773211e-01
1.53309357e+00 -2.17585206e+00 -3.70175540e-01 6.09145425e-02
8.05325583e-02 7.64917076e-01 -6.84403628e-02 5.98211050e-01
-1.50459437e-02 -5.75117588e-01 2.69333702e-02 -3.85491073e-01
1.36610255e-01 5.48524916e-01 -4.46642041e-01 2.58568168e-01
5.33936977e-01 1.01928818e+00 -9.54824626e-01 -3.57511967e-01
7.35365391e-01 3.80102575e-01 -5.56768358e-01 4.26560938e-01
-1.30872071e-01 4.38011169e-01 -3.63850892e-01 5.56246899e-02
2.29288578e-01 5.64086020e-01 -2.18762264e-01 3.49618882e-01
-4.67037678e-01 -6.09329529e-02 -7.79836118e-01 1.26994753e+00
-6.16330385e-01 8.65259767e-01 -4.25266743e-01 -1.15070343e+00
1.01481199e+00 4.44855303e-01 5.86802959e-01 -1.21125722e+00
1.21366754e-02 1.75452102e-02 9.84951016e-03 -9.36223090e-01
4.57452327e-01 9.66300517e-02 -2.54852474e-01 6.81010261e-02
-4.97446328e-01 2.20458478e-01 2.41393119e-01 -5.66964559e-02
1.39442039e+00 2.85096288e-01 6.60558566e-02 5.17908894e-02
5.12719810e-01 2.52169073e-01 3.81689161e-01 7.22035646e-01
-7.46750593e-01 1.02255791e-01 5.93619347e-01 -9.80705678e-01
-7.84051001e-01 -6.70736134e-01 3.65262032e-01 1.05759025e+00
-1.80438757e-01 -1.48161247e-01 -9.33414459e-01 -7.85712779e-01
1.18342981e-01 1.40452528e+00 -5.92492580e-01 -6.43473446e-01
-8.38502765e-01 -1.68941677e-01 3.70013118e-01 7.54560709e-01
2.03954905e-01 -1.36186886e+00 -1.45379126e+00 4.64130193e-01
1.20930225e-01 -1.28885067e+00 1.22067802e-01 1.63084447e-01
-3.26823056e-01 -1.04680777e+00 -4.26491261e-01 -3.79449159e-01
2.91549861e-01 6.30111620e-02 9.25539255e-01 -3.34923789e-02
-3.10859114e-01 6.92658365e-01 -4.15109724e-01 -8.30340266e-01
-6.48919582e-01 -1.53993547e-01 1.02900147e-01 1.73803434e-01
5.69339275e-01 -8.89625102e-02 -9.11408722e-01 1.82439834e-01
-4.36302632e-01 2.67848372e-01 7.78679192e-01 4.24019873e-01
3.74000520e-01 -9.33759362e-02 1.01510739e+00 -5.54707170e-01
6.64212286e-01 -4.89778370e-01 -7.93166518e-01 2.75176354e-02
-8.60022068e-01 2.81470984e-01 6.05412781e-01 -1.74967367e-02
-1.09610868e+00 5.56100190e-01 -5.29487014e-01 -3.43731344e-01
-5.28600872e-01 2.32906148e-01 4.92657602e-01 1.17692210e-01
6.52917206e-01 6.86220126e-03 3.24837267e-01 -4.10589539e-02
2.85102397e-01 5.96698463e-01 4.56377685e-01 -8.78805444e-02
3.49661797e-01 2.89608628e-01 3.75672579e-01 -4.36755747e-01
-8.63547325e-01 -5.74866295e-01 -5.57372093e-01 -8.64564240e-01
1.08530521e+00 -1.05481505e+00 -1.43306732e+00 1.64193451e-01
-1.09430659e+00 -8.03325474e-01 -3.96451950e-01 6.46414697e-01
-1.03544009e+00 2.30211280e-02 -1.56221583e-01 -1.41029823e+00
2.66541056e-02 -1.18163633e+00 1.09022605e+00 -7.59788156e-02
-1.30170301e-01 -7.77203262e-01 1.76074237e-01 4.98690128e-01
3.79744172e-01 4.71037358e-01 3.96872461e-01 -4.92503256e-01
-4.28935021e-01 -5.67588329e-01 5.61086135e-03 3.82659793e-01
-3.91765058e-01 -2.11490542e-01 -9.21549737e-01 -2.79224757e-02
-4.04570907e-01 -1.29430115e-01 8.80773127e-01 3.92232567e-01
1.17204177e+00 -6.30254894e-02 -3.17128181e-01 5.51576726e-02
1.11456656e+00 5.14223099e-01 7.84934521e-01 5.31605482e-01
3.96173328e-01 8.65594387e-01 1.05739319e+00 6.96810663e-01
7.44102418e-01 7.84494042e-01 7.42571771e-01 -9.59125385e-02
-1.84372053e-01 -3.23570400e-01 6.73516333e-01 3.12554955e-01
-3.02877665e-01 -1.80602409e-02 -1.14414692e+00 5.86538255e-01
-2.46903801e+00 -1.18676674e+00 -1.08115666e-01 2.18488932e+00
3.26713651e-01 5.84984779e-01 5.18277645e-01 2.87771314e-01
1.93824366e-01 -6.19576015e-02 -6.49688125e-01 -7.52913892e-01
5.25745451e-01 -8.80123395e-03 8.83291543e-01 5.41206598e-01
-1.01242018e+00 8.30356658e-01 6.39406967e+00 7.75353551e-01
-9.45322275e-01 4.50352840e-02 6.85302079e-01 -2.31423587e-01
2.92326063e-01 -2.85247900e-02 -9.01994467e-01 4.55230743e-01
1.61746979e+00 8.68777409e-02 5.81250608e-01 9.54320312e-01
8.97949278e-01 -4.43068892e-01 -9.34923112e-01 8.36596608e-01
-1.42924577e-01 -1.05541074e+00 -5.67352295e-01 -4.43945602e-02
4.34392661e-01 -2.17081998e-02 2.73094177e-02 8.51910353e-01
3.30843687e-01 -9.35304523e-01 7.65080631e-01 1.06176507e+00
3.01460236e-01 -1.30746448e+00 7.43305564e-01 6.37393296e-01
-1.00128579e+00 -6.87727094e-01 -1.18997179e-01 -3.79727274e-01
4.96823847e-01 2.76404142e-01 -9.20183837e-01 2.75895953e-01
4.52306181e-01 7.05804110e-01 -6.37460530e-01 9.76889551e-01
-2.58465022e-01 9.39093530e-01 1.22158211e-02 -2.19270602e-01
5.28865933e-01 2.92867154e-01 2.67677337e-01 1.13148880e+00
2.70995438e-01 6.19664378e-02 2.19891101e-01 2.47503012e-01
5.21160066e-01 -4.54256803e-01 -9.28796470e-01 4.12524790e-01
2.88977381e-03 8.99559796e-01 -3.07429135e-01 -1.31558076e-01
-5.16732633e-01 5.27909040e-01 3.31662744e-01 2.21586108e-01
-1.10788155e+00 -3.03950965e-01 7.63562858e-01 2.23312512e-01
2.95252144e-01 -3.09732467e-01 -2.13978216e-01 -2.63629287e-01
-9.90953520e-02 -4.64648038e-01 1.30888700e-01 -7.39762664e-01
-5.69036067e-01 6.03388250e-01 2.33439013e-01 -1.49285150e+00
-6.89394712e-01 -7.61940420e-01 -6.66337729e-01 4.43015575e-01
-1.93177402e+00 -9.33285952e-01 -1.14137447e-03 6.05383992e-01
8.17299366e-01 -4.04862165e-01 4.73395288e-01 9.46400687e-02
-6.86027586e-01 4.58941162e-01 -6.36294559e-02 -2.07924992e-01
3.56960893e-01 -9.38632846e-01 5.44840813e-01 8.58363152e-01
-3.08743238e-01 -2.10601240e-01 8.05681527e-01 -5.76093733e-01
-1.47847188e+00 -1.26364875e+00 9.92151082e-01 -6.63248420e-01
4.66352195e-01 -1.06565468e-01 -4.66877043e-01 4.64098155e-01
3.10330570e-01 -9.11012525e-04 3.13251287e-01 -2.83722758e-01
2.14760527e-01 -4.56488729e-01 -9.36572492e-01 6.46765471e-01
7.78280735e-01 -3.68149281e-01 -3.71690840e-01 3.50734025e-01
6.67524636e-01 -2.44118273e-01 -5.91356337e-01 2.74097830e-01
4.67216015e-01 -1.15460849e+00 8.58248293e-01 -6.25837982e-01
3.77582878e-01 -1.34385854e-01 2.31191903e-01 -1.23484886e+00
-3.77538651e-01 -8.03334296e-01 -4.34489787e-01 5.13819873e-01
4.56997067e-01 -5.03614187e-01 8.13574493e-01 1.04189074e+00
-5.95378458e-01 -9.47314322e-01 -1.22415960e+00 -8.49642873e-01
-2.53906220e-01 -1.18014038e+00 5.86995423e-01 -3.21313925e-02
9.69492719e-02 7.96197355e-02 -8.04616392e-01 7.44021013e-02
3.92770499e-01 -4.67517376e-01 7.04226851e-01 -1.14030838e+00
-4.54197153e-02 -3.62893462e-01 -6.23746216e-01 -6.74878836e-01
4.26761955e-01 -4.97055978e-01 4.61424649e-01 -1.50103903e+00
-4.68018860e-01 -2.55953670e-01 -5.58761299e-01 7.76874542e-01
-1.20084643e-01 -1.48958564e-01 2.31764242e-01 -2.06377774e-01
-8.97425830e-01 4.86912996e-01 9.58105385e-01 9.98745486e-02
-5.06583154e-02 4.31632072e-01 -3.68373901e-01 4.11099792e-01
1.00906754e+00 -4.21877503e-01 -4.32611108e-01 -1.66249663e-01
2.80848563e-01 5.08211553e-01 6.72979474e-01 -1.49543011e+00
4.90102172e-01 -2.89116085e-01 4.60282415e-02 -5.38305938e-01
6.19634449e-01 -1.16954625e+00 -3.86558361e-02 8.75630438e-01
-6.90260887e-01 2.20039099e-01 4.55826730e-01 9.44921076e-01
-2.57588252e-02 -5.74708953e-02 4.89202291e-01 3.20668846e-01
-1.43829310e+00 1.67109817e-01 -8.92570615e-01 -2.42700234e-01
1.57135808e+00 -3.00729983e-02 -8.48038569e-02 -4.59386319e-01
-5.84966421e-01 4.68212634e-01 -3.54446948e-01 7.25427330e-01
8.11058283e-01 -1.24155974e+00 -7.76534796e-01 3.81829709e-01
2.75163859e-01 -3.72824430e-01 3.17270339e-01 9.51768458e-01
-3.79709244e-01 6.43719971e-01 -3.35546017e-01 -2.04927340e-01
-8.71911347e-01 9.07301605e-01 7.98259750e-02 -3.86260271e-01
-6.63475215e-01 4.18849170e-01 -2.94634372e-01 -1.31188139e-01
3.37809175e-01 -1.11337543e-01 -5.03767073e-01 -2.91160882e-01
4.71512586e-01 7.13582456e-01 2.13127121e-01 -6.13823712e-01
-3.27154428e-01 1.69471845e-01 1.41060144e-01 -1.33126304e-01
1.21332216e+00 4.32579331e-02 6.03197038e-01 3.87454510e-01
8.45708489e-01 -4.52532917e-01 -1.94343626e+00 3.21433604e-01
2.36294717e-01 -1.64675012e-01 3.56088459e-01 -1.03735101e+00
-9.14373517e-01 1.01637447e+00 1.05917788e+00 3.08169484e-01
9.60915446e-01 -2.57732153e-01 1.17156112e+00 4.87120509e-01
4.30837631e-01 -1.37829840e+00 -2.22587362e-01 4.21252072e-01
8.40677440e-01 -1.56808543e+00 -3.65133286e-01 2.00938299e-01
-1.21826899e+00 1.13188446e+00 6.53555810e-01 -2.91207403e-01
7.24439323e-01 1.03014216e-01 1.96113229e-01 -2.87762154e-02
-1.14664066e+00 -5.44813812e-01 3.44566137e-01 7.30785966e-01
-9.07054520e-04 3.78576010e-01 -1.81187391e-01 3.74723315e-01
1.07402183e-01 1.62722751e-01 2.71572471e-01 9.36655104e-01
-6.69877827e-01 -1.06896794e+00 -7.09252711e-03 3.93851578e-01
-1.52171671e-01 3.49019885e-01 9.66978818e-02 6.83387160e-01
2.02079967e-01 1.26405215e+00 3.80190276e-02 -5.79309106e-01
7.27056503e-01 2.06148643e-02 -8.41894075e-02 -6.78313226e-02
-5.41925192e-01 -5.32232225e-01 2.46667549e-01 -8.53403449e-01
-2.99194515e-01 -7.38808572e-01 -1.31907403e+00 -9.96098965e-02
1.84172019e-01 -3.00406534e-02 8.01025748e-01 1.16286469e+00
7.33460248e-01 7.28052795e-01 8.64361882e-01 -9.35520709e-01
-5.61873078e-01 -4.00283992e-01 -9.04613137e-02 3.97240520e-02
6.29343688e-01 -7.35557556e-01 -1.91727243e-02 -2.15208847e-02] | [5.673739910125732, 1.0263299942016602] |
7d52c281-a3f7-4ce4-b97c-7b193cf00857 | a-processing-framework-to-access-large | 2303.07442 | null | https://arxiv.org/abs/2303.07442v1 | https://arxiv.org/pdf/2303.07442v1.pdf | A processing framework to access large quantities of whispered speech found in ASMR | Whispering is a ubiquitous mode of communication that humans use daily. Despite this, whispered speech has been poorly served by existing speech technology due to a shortage of resources and processing methodology. To remedy this, this paper provides a processing framework that enables access to large and unique data of high-quality whispered speech. We obtain the data from recordings submitted to online platforms as part of the ASMR media-cultural phenomenon. We describe our processing pipeline and a method for improved whispered activity detection (WAD) in the ASMR data. To efficiently obtain labelled, clean whispered speech, we complement the automatic WAD by using Edyson, a bulk audio-annotation tool with human-in-the-loop. We also tackle a problem particular to ASMR: separation of whisper from other acoustic triggers present in the genre. We show that the proposed WAD and the efficient labelling allows to build extensively augmented data and train a classifier that extracts clean whisper segments from ASMR audio. Our large and growing dataset enables whisper-capable, data-driven speech technology and linguistic analysis. It also opens opportunities in e.g. HCI as a resource that may elicit emotional, psychological and neuro-physiological responses in the listener. | ['Zofia Malisz', 'Gustav Eje Henter', 'Pablo Perez Zarazaga'] | 2023-03-13 | null | null | null | null | ['activity-detection'] | ['computer-vision'] | [ 3.22308034e-01 1.58852175e-01 5.92199147e-01 -2.10391477e-01
-1.26877213e+00 -5.87044716e-01 1.33604899e-01 2.85874993e-01
-2.77493268e-01 4.41643268e-01 5.38260341e-01 1.77784473e-01
1.22693390e-01 -1.01745613e-01 -3.93181980e-01 -6.30594850e-01
-3.66503932e-02 -8.22504684e-02 4.87493217e-01 -4.77694720e-01
8.40817317e-02 4.78447825e-01 -2.14057446e+00 6.38301432e-01
2.91091800e-01 1.01215971e+00 4.93113697e-01 1.19436848e+00
2.25510243e-02 4.81488615e-01 -1.09722042e+00 3.04093417e-02
-1.31436884e-01 -3.52198541e-01 -5.76030314e-01 -2.30126292e-01
4.63226922e-02 1.69167578e-01 1.96425512e-01 7.42112160e-01
1.09627485e+00 2.35696778e-01 -5.95048862e-03 -1.12742400e+00
4.26206663e-02 8.15330267e-01 -2.24100336e-01 3.42056245e-01
9.44020212e-01 -6.07898086e-02 7.26289868e-01 -9.03667927e-01
4.64800358e-01 1.04277802e+00 7.38661349e-01 6.09777749e-01
-1.23519039e+00 -6.30366862e-01 -4.98430640e-01 3.63013172e-03
-1.42008424e+00 -1.28735757e+00 9.02262568e-01 -3.84051621e-01
1.29171669e+00 8.03066790e-01 4.91016686e-01 1.69592130e+00
-4.81877059e-01 6.53186440e-01 9.29286659e-01 -6.90581560e-01
5.07358730e-01 1.78078964e-01 1.78427026e-01 -1.34246638e-02
-5.37027597e-01 -9.76117551e-02 -1.26376176e+00 -2.27199912e-01
2.28819296e-01 -6.09143913e-01 -4.58710134e-01 5.95945895e-01
-1.13465512e+00 2.67332703e-01 -3.83192152e-01 6.20561242e-01
-5.58857918e-01 -5.39763719e-02 8.25630188e-01 4.42587495e-01
5.47103941e-01 5.30195773e-01 -4.86565471e-01 -8.59983802e-01
-1.29443598e+00 1.03692062e-01 9.80206490e-01 6.69300377e-01
3.27080667e-01 2.34509408e-01 -9.94758159e-02 1.43750679e+00
1.26591325e-01 6.54246330e-01 5.50866961e-01 -8.78861368e-01
1.47863388e-01 -4.82383780e-02 8.33090991e-02 -8.87458265e-01
-5.76051176e-01 -2.90229898e-02 -3.05738717e-01 -8.19546953e-02
7.96952844e-02 -1.55900687e-01 -4.54036415e-01 1.51444769e+00
3.73177022e-01 1.83722883e-01 1.63843464e-02 9.14605141e-01
9.00368094e-01 7.69216716e-01 -2.96789318e-01 -5.38719833e-01
1.70877516e+00 -6.13304913e-01 -1.29197311e+00 -2.11916002e-03
3.19335282e-01 -9.79206860e-01 1.40174866e+00 1.05365276e+00
-1.11728346e+00 -4.62163031e-01 -9.30091500e-01 -8.46573859e-02
-3.84445548e-01 -1.35628164e-01 9.79150161e-02 1.18960202e+00
-1.14177585e+00 3.38394850e-01 -6.19344175e-01 -5.43572187e-01
1.02570064e-01 2.12549478e-01 -4.33458686e-01 6.21282816e-01
-1.25072861e+00 5.07891417e-01 7.13533536e-02 -1.07196733e-01
-6.86344862e-01 -8.44745040e-01 -7.35270917e-01 -1.72969431e-01
3.53307426e-01 1.01177379e-01 1.65806079e+00 -7.00683355e-01
-1.88793063e+00 9.66833472e-01 -1.31575301e-01 -5.87876916e-01
7.95429870e-02 -4.17662621e-01 -1.04707897e+00 3.92737240e-01
-6.25057667e-02 1.75537258e-01 1.05270898e+00 -8.24245036e-01
-3.85056466e-01 -2.17011377e-01 -8.21045995e-01 -6.39611632e-02
-3.48451048e-01 8.42197061e-01 -1.61787033e-01 -8.11262310e-01
-2.12858185e-01 -4.84590977e-01 2.56381482e-01 -5.58315217e-01
-5.22090495e-01 -8.61364827e-02 7.52512693e-01 -9.86852407e-01
1.48367524e+00 -2.49138212e+00 -1.51415721e-01 1.60202459e-01
1.39493614e-01 4.80890632e-01 1.51235655e-01 7.52487481e-01
-1.46703765e-01 9.69613120e-02 6.17705882e-02 -4.90537494e-01
1.43098190e-01 -1.69842422e-01 -4.57016647e-01 2.96172976e-01
1.36951104e-01 4.55229849e-01 -9.02814150e-01 -3.12616169e-01
2.17305690e-01 6.57684803e-01 -3.17331403e-01 6.24151289e-01
2.10964411e-01 3.70802253e-01 3.24126445e-02 6.28520489e-01
3.27276766e-01 7.41449714e-01 -5.65141022e-01 -2.35699818e-01
-7.60580897e-01 6.22336030e-01 -1.15790975e+00 1.98694181e+00
-6.77352428e-01 7.83012807e-01 7.05243528e-01 -5.70325613e-01
1.27367115e+00 8.71463358e-01 2.80933797e-01 -5.98275959e-01
3.39304507e-01 6.33869708e-01 -1.76111221e-01 -1.05813897e+00
5.76738954e-01 -1.51619375e-01 -1.79113105e-01 3.04626107e-01
2.55685121e-01 -5.87165296e-01 -8.15404505e-02 -1.05365291e-02
1.51537430e+00 -2.58902282e-01 4.00348663e-01 -2.80532926e-01
5.62724352e-01 -4.71687227e-01 3.12514454e-02 5.46160817e-01
-5.12906373e-01 8.06501329e-01 7.98808336e-02 3.72054949e-02
-8.51142049e-01 -8.94382179e-01 -8.11646059e-02 1.56773019e+00
-5.89304328e-01 -5.90912580e-01 -1.03574574e+00 1.27353236e-01
-5.06486058e-01 8.76572371e-01 -4.04773623e-01 -1.37174100e-01
-1.18588902e-01 -3.26786369e-01 1.20299935e+00 5.78351542e-02
-7.43696913e-02 -1.49023294e+00 -7.61441529e-01 5.94316363e-01
-3.94669592e-01 -1.24983251e+00 -4.57344562e-01 5.63388407e-01
-8.29024799e-03 -3.51400167e-01 -5.28447568e-01 -4.73033786e-01
-3.70499939e-01 -9.06569697e-03 7.08680391e-01 -5.60308874e-01
-6.15020275e-01 5.66292346e-01 -7.47491419e-01 -9.36295927e-01
-9.09501016e-01 -5.27659804e-02 3.54647547e-01 2.66765863e-01
4.10686523e-01 -9.75770831e-01 -3.02488834e-01 1.88870072e-01
-5.54493368e-01 -1.87438130e-01 5.86224049e-02 7.00276196e-02
2.90406108e-01 -4.56773750e-02 1.21364212e+00 -4.11019355e-01
1.11195970e+00 -4.74621594e-01 -1.19219668e-01 -2.38237023e-01
1.04189590e-01 -4.59626079e-01 5.58873296e-01 -5.49958110e-01
-9.11130548e-01 9.31463689e-02 -8.57391179e-01 -3.52047324e-01
-7.71413207e-01 1.34364828e-01 -2.70641953e-01 2.97455013e-01
1.05233753e+00 1.70678452e-01 3.93706327e-03 -6.68863118e-01
1.09707884e-01 1.68302798e+00 1.07275426e+00 -2.08713651e-01
2.71302223e-01 2.43458435e-01 -6.23706043e-01 -1.74162674e+00
-7.02765644e-01 -9.82710481e-01 -2.54122466e-01 -4.67492074e-01
8.67224932e-01 -8.14833403e-01 -7.01820970e-01 6.55211389e-01
-1.18262851e+00 -3.17828774e-01 -4.62309718e-01 3.54756892e-01
-6.45879567e-01 2.62266606e-01 -4.59932685e-01 -1.76129735e+00
-7.94954181e-01 -7.15166390e-01 1.34201717e+00 -6.58461303e-02
-8.70372653e-01 -3.99507344e-01 4.03278679e-01 4.15590644e-01
5.41530252e-01 1.20998867e-01 8.41127243e-03 -1.06699443e+00
5.39767206e-01 -1.90496385e-01 3.74541402e-01 4.15818542e-01
1.85712099e-01 2.56475415e-02 -1.97406268e+00 3.00938457e-01
3.74047130e-01 -6.03876829e-01 4.23015982e-01 2.45991141e-01
9.46761906e-01 -1.18033096e-01 9.87298861e-02 1.42973319e-01
5.18245995e-01 3.18587184e-01 6.69965267e-01 9.47785527e-02
3.42370301e-01 9.44094479e-01 5.74363768e-01 7.08735704e-01
-2.00563818e-01 7.19290257e-01 1.43903911e-01 6.59380406e-02
-1.53182358e-01 -5.70940822e-02 7.13025630e-01 1.25210071e+00
5.02845459e-02 -1.62577420e-01 -8.91807318e-01 5.99441707e-01
-1.43350136e+00 -1.04525721e+00 -5.24292707e-01 2.10440874e+00
1.12004149e+00 -5.85045293e-02 4.71273243e-01 9.44775283e-01
8.66727471e-01 -1.06656902e-01 5.46002574e-02 -9.65517938e-01
-1.11042634e-01 4.89658475e-01 8.29769224e-02 5.15116334e-01
-1.03498662e+00 6.80749536e-01 5.99338913e+00 1.04352212e+00
-1.18537974e+00 2.86249518e-01 2.17335648e-03 -4.49299186e-01
-4.82974462e-02 -8.60115469e-01 -5.57953238e-01 4.25332099e-01
1.83513248e+00 5.31638376e-02 8.32826972e-01 6.46657467e-01
1.09983408e+00 -6.95314556e-02 -1.08051920e+00 1.37170076e+00
1.88425392e-01 -9.65532303e-01 -7.80942857e-01 -2.02239200e-01
2.05672868e-02 2.11278394e-01 -2.18450561e-01 1.01582266e-01
-3.67293030e-01 -9.29444551e-01 1.20259047e+00 3.37999791e-01
8.56863856e-01 -6.96430683e-01 4.28462684e-01 5.06114960e-01
-1.15245295e+00 3.28507461e-02 1.56885147e-01 1.35875801e-02
2.83763766e-01 8.25459957e-01 -1.17903733e+00 3.07757825e-01
9.04091716e-01 1.98584080e-01 -2.42317572e-01 9.12240624e-01
6.71652406e-02 1.00128174e+00 -5.11435747e-01 -6.92197755e-02
-3.21986705e-01 3.66568595e-01 8.61167729e-01 1.83966374e+00
4.43937480e-01 -5.03377058e-02 -1.85607269e-01 7.21403837e-01
2.12005079e-01 3.59758407e-01 -6.25700116e-01 -2.95942158e-01
8.17244172e-01 1.50860381e+00 -7.24847257e-01 3.42044979e-02
-1.35591865e-01 9.22176421e-01 -1.58875257e-01 -3.14945094e-02
-7.01104343e-01 -6.55605257e-01 4.78381604e-01 -1.44985197e-02
-1.48967385e-01 -5.75625449e-02 -3.00645735e-03 -6.92704380e-01
-3.20769399e-02 -1.04376316e+00 2.52868142e-02 -1.02592194e+00
-1.17692626e+00 9.16721165e-01 -2.95769840e-01 -9.39104736e-01
-2.82269865e-01 -3.52930129e-01 -4.75107521e-01 9.46029067e-01
-1.01997924e+00 -8.52814317e-01 -2.23495871e-01 6.08597159e-01
9.08173978e-01 -1.17035061e-01 1.23795807e+00 6.31331742e-01
-5.17889202e-01 2.64074951e-01 -4.43018019e-01 -3.34850103e-01
8.76354873e-01 -1.12653148e+00 4.36849415e-01 5.81154346e-01
1.88382640e-01 5.35369337e-01 1.21081924e+00 -2.47866362e-01
-1.35948718e+00 -9.34474707e-01 9.12629902e-01 -4.37377602e-01
1.08291626e+00 -1.06458175e+00 -1.05227482e+00 -2.23763306e-02
2.25003719e-01 -2.14113563e-01 1.05281210e+00 -9.29652154e-03
-2.56411791e-01 -1.64384976e-01 -1.01365578e+00 2.87737995e-01
7.76880682e-01 -1.04960608e+00 -9.00701880e-01 -3.62760611e-02
8.88338327e-01 -1.17899306e-01 -6.54941201e-01 -1.68652534e-01
5.41546464e-01 -1.01297808e+00 6.70618594e-01 -1.28788903e-01
-2.97820382e-02 -2.23946884e-01 -2.28131384e-01 -1.37214875e+00
3.38445812e-01 -1.60058355e+00 -1.46614015e-01 1.95135021e+00
4.10794288e-01 -2.86097765e-01 8.88072997e-02 4.97196972e-01
-6.28490031e-01 -5.47097959e-02 -1.11263192e+00 -7.97193706e-01
-6.66101873e-01 -1.28677070e+00 3.63493741e-01 5.73138297e-01
5.77236712e-01 3.95578802e-01 -4.71065402e-01 1.92980379e-01
1.25287741e-01 -5.09768009e-01 5.38333595e-01 -1.14249659e+00
-2.17271566e-01 -1.78684726e-01 -3.07101756e-01 -2.85769373e-01
-5.29292263e-02 -6.01148307e-01 6.57742918e-01 -1.04195189e+00
-4.17434007e-01 -2.42624022e-02 -2.49668926e-01 3.78103256e-01
2.61267990e-01 3.11170280e-01 -6.36671409e-02 -1.76525772e-01
-5.02743542e-01 2.52796382e-01 5.85962355e-01 2.32723609e-01
-6.02964163e-01 7.95853958e-02 -4.32621598e-01 7.47856796e-01
7.71940231e-01 -4.87345278e-01 -4.76053715e-01 1.40392944e-01
2.90454537e-01 7.65500739e-02 4.01526719e-01 -1.12410879e+00
1.46279320e-01 2.69761413e-01 -3.68381560e-01 -5.40962577e-01
7.25505114e-01 -6.87659800e-01 1.11925013e-01 -3.96105200e-02
-5.00926733e-01 -4.31618571e-01 2.87795067e-01 1.61519215e-01
-9.84959677e-02 -1.91554308e-01 7.28832245e-01 3.64200324e-02
-2.33797282e-01 -3.66519690e-01 -9.99720395e-01 1.84620723e-01
5.58001280e-01 -1.56765506e-01 -1.82628542e-01 -3.54878247e-01
-8.16843867e-01 -3.21264893e-01 -1.70594361e-02 5.49173176e-01
3.94543380e-01 -8.76580000e-01 -5.68541825e-01 3.91906679e-01
1.84536919e-01 -3.09160620e-01 4.25181866e-01 1.03530097e+00
-2.39731401e-01 1.69291854e-01 -9.74167660e-02 -3.80176097e-01
-1.50463259e+00 4.61760819e-01 2.87285954e-01 4.82439101e-01
-6.72048092e-01 9.91516471e-01 -2.12824404e-01 -5.65857776e-02
6.41304433e-01 -3.76276612e-01 -3.66719544e-01 6.05687201e-01
1.07225776e+00 6.12922013e-01 7.34382451e-01 -6.48196995e-01
-5.21657944e-01 -1.50800779e-01 4.22251850e-01 -7.76312888e-01
1.37263894e+00 -5.11739969e-01 6.03803620e-03 1.39889205e+00
1.10013521e+00 6.12515450e-01 -7.22008049e-01 4.53091592e-01
2.08885819e-01 -8.31443891e-02 3.02546144e-01 -7.96846211e-01
-4.25978005e-01 1.10676682e+00 5.47085047e-01 9.37593579e-01
1.21281219e+00 1.01434633e-01 8.41143191e-01 3.33706528e-01
3.17710847e-01 -1.54706597e+00 -8.55266452e-02 3.98710221e-01
1.28892374e+00 -7.68582761e-01 -7.18112230e-01 -4.95895922e-01
-6.81240916e-01 1.06156027e+00 -2.09410042e-02 2.27987558e-01
6.28045201e-01 9.59617138e-01 4.78956550e-01 -1.07075192e-01
-8.76147032e-01 -3.09002191e-01 8.42915028e-02 1.06225681e+00
5.88273346e-01 -3.29053439e-02 -2.87717972e-02 1.35400093e+00
-8.36592257e-01 1.83316525e-02 5.12871087e-01 6.91398680e-01
-5.99593878e-01 -7.91449249e-01 -7.42326796e-01 8.13886449e-02
-8.43629479e-01 -1.89804167e-01 -8.07880044e-01 8.46419409e-02
2.84722280e-02 1.59711540e+00 -1.95866704e-01 -5.31136036e-01
6.51778758e-01 5.67153990e-01 1.83880314e-01 -7.69956112e-01
-1.16816342e+00 8.08876514e-01 6.03876472e-01 -6.41877890e-01
-3.43566984e-01 -8.13874841e-01 -1.35950792e+00 2.99375206e-01
-2.66409039e-01 2.05075681e-01 9.67743516e-01 7.54760206e-01
4.59287196e-01 8.69260311e-01 6.80181324e-01 -1.06849694e+00
-1.11704864e-01 -1.18144596e+00 -8.98615777e-01 2.57443219e-01
6.01459503e-01 -3.11511666e-01 -5.11402011e-01 3.43450397e-01] | [15.084329605102539, 5.292364597320557] |
f6518b3d-76c3-481c-9843-0f8dc80c7824 | diagnostic-benchmark-and-iterative-inpainting | 2304.06671 | null | https://arxiv.org/abs/2304.06671v2 | https://arxiv.org/pdf/2304.06671v2.pdf | Diagnostic Benchmark and Iterative Inpainting for Layout-Guided Image Generation | Spatial control is a core capability in controllable image generation. Advancements in layout-guided image generation have shown promising results on in-distribution (ID) datasets with similar spatial configurations. However, it is unclear how these models perform when facing out-of-distribution (OOD) samples with arbitrary, unseen layouts. In this paper, we propose LayoutBench, a diagnostic benchmark for layout-guided image generation that examines four categories of spatial control skills: number, position, size, and shape. We benchmark two recent representative layout-guided image generation methods and observe that the good ID layout control may not generalize well to arbitrary layouts in the wild (e.g., objects at the boundary). Next, we propose IterInpaint, a new baseline that generates foreground and background regions in a step-by-step manner via inpainting, demonstrating stronger generalizability than existing models on OOD layouts in LayoutBench. We perform quantitative and qualitative evaluation and fine-grained analysis on the four LayoutBench skills to pinpoint the weaknesses of existing models. Lastly, we show comprehensive ablation studies on IterInpaint, including training task ratio, crop&paste vs. repaint, and generation order. Project website: https://layoutbench.github.io | ['Mohit Bansal', 'Lijuan Wang', 'Zhe Gan', 'Zhengyuan Yang', 'Linjie Li', 'Jaemin Cho'] | 2023-04-13 | null | null | null | null | ['layout-to-image-generation'] | ['computer-vision'] | [ 4.22113717e-01 4.60444055e-02 3.42276767e-02 -6.98753148e-02
-6.93692565e-01 -9.78536427e-01 8.16194475e-01 -3.76310758e-02
5.40551618e-02 7.20492363e-01 2.81142980e-01 -5.23242235e-01
2.15535588e-03 -7.43820012e-01 -9.31533515e-01 -5.02517939e-01
7.93687925e-02 3.79626125e-01 1.17198460e-01 -1.62063688e-01
5.73724926e-01 6.54256880e-01 -1.53314435e+00 6.46362901e-01
1.08642924e+00 5.56280851e-01 4.26294953e-01 1.09528673e+00
-1.45844728e-01 9.06548917e-01 -1.32138634e+00 -8.28162432e-02
3.57040912e-01 -6.16304159e-01 -6.30497813e-01 2.59342998e-01
1.13623416e+00 -5.36357999e-01 -2.61494160e-01 5.90630293e-01
8.60113919e-01 -1.30900934e-01 8.88112783e-01 -1.57718468e+00
-1.28690755e+00 5.22545338e-01 -9.07600820e-01 2.08597377e-01
4.20325935e-01 6.95311785e-01 6.42565548e-01 -8.44688416e-01
1.12691379e+00 1.38137269e+00 5.51435351e-01 5.86954653e-01
-1.61782146e+00 -5.57340741e-01 3.98552865e-01 -4.31019515e-01
-1.17143524e+00 -2.38271683e-01 5.02680242e-01 -5.58193326e-01
7.91935325e-01 3.44641119e-01 5.09929717e-01 1.49553704e+00
1.06893204e-01 9.66380119e-01 1.27099168e+00 -5.53260744e-01
1.23961322e-01 4.03766893e-02 -3.95421952e-01 5.01493990e-01
4.06447679e-01 7.53173977e-02 -6.75016999e-01 1.53171137e-01
1.10740101e+00 -5.28371036e-01 -3.55311453e-01 -5.66593409e-01
-1.50641263e+00 5.13233781e-01 4.56250578e-01 -6.26594052e-02
1.32656991e-01 4.40159082e-01 1.91111192e-01 4.37025540e-02
5.84379613e-01 1.10512495e+00 -3.16473305e-01 -1.92686841e-01
-1.13087511e+00 8.67138147e-01 6.68446720e-01 1.68912792e+00
6.21610343e-01 8.03234056e-02 -1.00358355e+00 6.60062730e-01
-2.13225633e-01 7.94449806e-01 -3.40159908e-02 -1.24172127e+00
6.52863801e-01 2.70244837e-01 2.99730867e-01 -1.00902510e+00
-4.66853827e-02 -2.51636744e-01 -8.93443584e-01 4.03550416e-01
5.63884556e-01 -3.47201854e-01 -1.18577945e+00 1.55426776e+00
8.63803625e-02 -9.77891609e-02 -4.99689281e-01 7.11896420e-01
8.34113955e-01 6.01212680e-01 -3.07976510e-02 3.76796514e-01
1.08288276e+00 -1.19826901e+00 -5.73425174e-01 -3.85551453e-01
4.88369405e-01 -1.05502987e+00 1.57112241e+00 3.46170843e-01
-1.25381470e+00 -6.67699754e-01 -9.05778825e-01 -2.37099692e-01
-5.13493598e-01 3.75059247e-01 5.49968243e-01 6.31060719e-01
-1.36604726e+00 5.49677074e-01 -2.15094730e-01 -2.97531873e-01
5.17579079e-01 -7.57230967e-02 -5.65508269e-02 -2.60321219e-02
-6.74705982e-01 4.78477418e-01 1.55469254e-01 -2.64290720e-01
-9.67693567e-01 -1.27574527e+00 -8.07355225e-01 -1.85021266e-01
1.70823008e-01 -8.14713895e-01 1.28970528e+00 -6.03585780e-01
-9.97140110e-01 9.42343295e-01 -1.04106270e-01 -2.15631217e-01
8.04964066e-01 -2.61433989e-01 2.50693053e-01 -3.15224230e-02
4.69013691e-01 1.39428842e+00 1.08481693e+00 -1.75945091e+00
-5.75673938e-01 1.55524105e-01 1.07578181e-01 1.66457579e-01
-9.83534753e-02 -5.24777830e-01 -4.42523092e-01 -1.00966132e+00
-3.12966466e-01 -7.25135267e-01 -5.97198084e-02 9.91702601e-02
-1.01105678e+00 1.12811640e-01 9.08641577e-01 -4.01475281e-01
1.46002126e+00 -2.21818256e+00 1.41175771e-02 -9.47485026e-03
2.20730171e-01 -5.00598066e-02 -5.41875899e-01 6.52133346e-01
-1.60782233e-01 7.47084916e-01 -8.89584981e-03 -4.69640315e-01
2.63066888e-01 -2.02263996e-01 -4.89969879e-01 -1.24815911e-01
4.78830576e-01 1.12518728e+00 -9.82988656e-01 -6.31200373e-01
2.47587353e-01 2.20297575e-01 -7.85819113e-01 2.78467238e-01
-4.52442735e-01 3.34304243e-01 -5.26366234e-02 8.40477943e-01
8.39520574e-01 -1.11889914e-01 -1.35436863e-01 -2.30047941e-01
-3.02873373e-01 1.59744604e-03 -8.62903357e-01 1.82688081e+00
-3.95271331e-01 9.90082502e-01 -1.72568843e-01 -1.31314039e-01
6.99769139e-01 -2.85967618e-01 -4.21970822e-02 -1.06017125e+00
-2.44006708e-01 9.72678587e-02 -1.85008213e-01 -3.30967635e-01
9.91929233e-01 5.94432950e-01 -2.15219796e-01 4.74981815e-01
-2.04610944e-01 -7.39695609e-01 6.29811227e-01 6.24703050e-01
1.17332268e+00 5.17121971e-01 -9.27582234e-02 -4.61557955e-01
-3.41212213e-01 2.05514669e-01 -7.49719366e-02 1.26485848e+00
4.98392843e-02 1.31525850e+00 9.28874731e-01 -6.46455288e-02
-1.52682054e+00 -1.33014250e+00 5.30809164e-02 9.49312329e-01
1.02996349e-01 -4.82152045e-01 -1.19233584e+00 -7.12042928e-01
1.15682721e-01 8.95850301e-01 -9.53635812e-01 4.33458276e-02
-5.61814606e-01 -4.19455916e-01 6.96661115e-01 4.96817380e-01
4.59208816e-01 -1.33951354e+00 -7.19055116e-01 -2.00936109e-01
-1.00307852e-01 -9.20201302e-01 -8.55929971e-01 9.84190181e-02
-6.12651289e-01 -8.56416702e-01 -1.09548771e+00 -7.99565315e-01
1.10702419e+00 5.21001041e-01 1.70457685e+00 1.36883780e-01
-5.62764049e-01 4.61341918e-01 -2.86156446e-01 -6.04862332e-01
-3.85721952e-01 5.24576128e-01 -5.80387354e-01 -5.93564689e-01
-3.20065707e-01 -3.50572109e-01 -1.02834046e+00 3.17621082e-01
-9.88799095e-01 3.81426096e-01 6.44864261e-01 7.99082398e-01
5.04932046e-01 -5.43479696e-02 2.68764079e-01 -1.09233546e+00
1.03580606e+00 -2.95428991e-01 -4.08008456e-01 1.42041907e-01
-2.55847991e-01 9.10462067e-02 3.41434002e-01 -4.83877808e-01
-1.17239618e+00 -4.19207036e-01 2.25668550e-01 -4.35989141e-01
-4.13453966e-01 -1.81157827e-01 -2.03080460e-01 3.07090372e-01
8.39329362e-01 1.94535732e-01 -2.57349700e-01 -2.86432683e-01
7.14240670e-01 2.55662560e-01 4.69500631e-01 -1.10058808e+00
1.08456469e+00 2.83762395e-01 -2.55231678e-01 -6.57405615e-01
-7.26584256e-01 4.48851287e-02 -5.82212985e-01 -5.56568727e-02
8.23403418e-01 -7.14805841e-01 -4.04423356e-01 6.13208532e-01
-1.23035157e+00 -1.31089675e+00 -6.43964112e-01 -4.40595716e-01
-5.38432777e-01 -9.11570787e-02 -4.66310114e-01 -6.25066578e-01
-1.09655969e-02 -1.04505575e+00 1.75584221e+00 1.87093690e-01
-5.48436105e-01 -8.65370095e-01 -2.00671971e-01 -1.91093367e-02
4.59706783e-01 6.18677676e-01 1.07415867e+00 1.73079833e-01
-1.03551888e+00 2.83851385e-01 -6.21489227e-01 1.04629189e-01
2.72472084e-01 5.43566525e-01 -7.48429596e-01 -1.81688681e-01
-8.52745891e-01 -2.69699842e-01 7.16255128e-01 5.77323198e-01
1.43605959e+00 -2.25826889e-01 -5.32041132e-01 7.49775827e-01
1.23581254e+00 1.66450351e-01 9.40218866e-01 5.14427125e-01
7.93949127e-01 6.39599442e-01 9.77805853e-01 4.98760641e-01
2.07056016e-01 4.26358700e-01 2.86169916e-01 -5.40366948e-01
-8.49365711e-01 -7.59853601e-01 3.90451658e-03 9.79567096e-02
2.70195663e-01 -8.82251263e-01 -7.94155955e-01 7.41440237e-01
-1.53909791e+00 -9.20604050e-01 -2.72366256e-01 1.99581265e+00
8.91766191e-01 1.80586785e-01 9.25701186e-02 -6.33999780e-02
5.35545111e-01 2.27485135e-01 -4.32074875e-01 -5.17109036e-01
-2.97289968e-01 2.85426110e-01 4.96623009e-01 2.85888791e-01
-9.87029791e-01 1.12756860e+00 6.69898748e+00 1.07866275e+00
-8.80136549e-01 -2.16363043e-01 1.26635075e+00 -2.86227286e-01
-7.46667445e-01 -2.25352153e-01 -9.05321658e-01 5.79574764e-01
2.05115378e-01 1.74584627e-01 3.61973554e-01 7.30397820e-01
4.32875663e-01 -4.49760914e-01 -1.17046559e+00 8.67878318e-01
1.05923750e-01 -1.60174346e+00 2.55665004e-01 2.05599308e-01
1.29746222e+00 -6.25234246e-01 6.32251978e-01 2.23301783e-01
3.02198559e-01 -1.18872106e+00 1.26569057e+00 4.31424856e-01
1.27836204e+00 -7.08200812e-01 1.81215689e-01 2.33152605e-04
-9.62651193e-01 9.28259566e-02 -1.90550163e-01 1.89668804e-01
1.01615518e-01 6.13674998e-01 -9.66201186e-01 1.99617967e-01
8.72698426e-01 4.45743293e-01 -1.10430193e+00 8.50291967e-01
-3.42198521e-01 3.45141500e-01 -3.87679003e-02 1.76187642e-02
2.84132302e-01 -3.25772092e-02 2.36294180e-01 1.37571120e+00
5.59230328e-01 -3.26596439e-01 -2.40754828e-01 1.41233945e+00
-7.99421668e-02 -2.19711825e-01 -9.69880998e-01 -6.20869137e-02
6.87937260e-01 1.05423522e+00 -9.75427806e-01 -2.03810558e-01
1.93170726e-01 1.24081743e+00 2.86826402e-01 7.29948819e-01
-1.21454453e+00 -5.48671722e-01 6.46498322e-01 6.76745296e-01
3.85395080e-01 -3.55825841e-01 -5.90704978e-01 -6.88315749e-01
6.68934546e-03 -1.10082150e+00 -2.27720112e-01 -1.22387588e+00
-1.18886888e+00 3.12975854e-01 3.30646902e-01 -1.12936127e+00
2.69122049e-02 -5.24003863e-01 -5.85333109e-01 8.00930560e-01
-1.20287347e+00 -1.21546638e+00 -7.54459262e-01 1.45493910e-01
7.99077868e-01 1.75736651e-01 4.52448010e-01 2.03900740e-01
-4.97298837e-01 7.48138845e-01 -1.91697404e-01 1.29448116e-01
1.17569065e+00 -1.72856557e+00 1.05344427e+00 7.50391662e-01
-1.05381049e-01 7.64666319e-01 6.96748078e-01 -7.49600053e-01
-1.26328397e+00 -1.26766288e+00 5.02486169e-01 -7.49176502e-01
2.28374913e-01 -8.84249985e-01 -3.47315967e-01 5.39682746e-01
7.01683044e-01 -2.01928198e-01 8.09234530e-02 -5.49298041e-02
-1.91380680e-01 4.03370969e-02 -1.01729119e+00 1.15923023e+00
1.63825798e+00 -1.49955034e-01 9.85336453e-02 3.46915245e-01
8.82010400e-01 -7.60682583e-01 -4.76122767e-01 1.09056406e-01
4.98643756e-01 -1.49480999e+00 1.21116841e+00 -2.35763505e-01
9.16319191e-01 -4.44710195e-01 1.39618933e-01 -1.61183691e+00
-3.18843275e-01 -8.19750249e-01 9.80641097e-02 1.60620856e+00
4.95622873e-01 -2.68882513e-01 8.46479118e-01 2.55335420e-01
-2.78538078e-01 -7.87045240e-01 -1.00783966e-01 -7.81039178e-01
1.49482667e-01 -1.28799379e-01 7.73617148e-01 5.85515678e-01
-4.76459086e-01 2.05900654e-01 -2.73660898e-01 -1.98914260e-01
5.64962447e-01 1.20968781e-02 1.34521401e+00 -5.14616549e-01
-2.18580499e-01 -5.71171284e-01 1.01676084e-01 -1.41155040e+00
-1.25305474e-01 -5.11920929e-01 1.80204511e-01 -1.87733757e+00
7.36964494e-02 -6.72878385e-01 4.64454442e-01 2.40115091e-01
-4.42195565e-01 4.73871589e-01 4.47115123e-01 -6.72282204e-02
-4.95587081e-01 3.20832789e-01 2.00300837e+00 -2.86128432e-01
-1.90966859e-01 -4.09867823e-01 -9.25532043e-01 2.05772802e-01
9.16879952e-01 -1.60721079e-01 -8.85554433e-01 -8.01010787e-01
1.92329973e-01 -3.78165483e-01 5.38405418e-01 -1.04581559e+00
-2.28311196e-01 -1.86756343e-01 6.58349931e-01 -7.51384437e-01
1.09376490e-01 -2.64781207e-01 -3.45334457e-03 1.07836477e-01
-4.97071981e-01 3.81022364e-01 4.41592366e-01 2.71585137e-01
1.09240294e-01 1.30974744e-02 4.95287448e-01 -1.87744915e-01
-6.67739391e-01 2.79426068e-01 -3.71334255e-01 5.80788732e-01
1.03025568e+00 -5.95323741e-01 -8.07549536e-01 -3.92057538e-01
-4.15530622e-01 7.55704939e-02 8.20207953e-01 6.16527379e-01
6.79239452e-01 -1.36635756e+00 -6.42538249e-01 1.91010833e-01
2.57729977e-01 3.58367562e-01 3.57029617e-01 5.30205190e-01
-9.83380854e-01 1.99435025e-01 -7.26442263e-02 -7.55100787e-01
-1.01537716e+00 5.83323538e-01 1.53979093e-01 -2.80998617e-01
-2.76697487e-01 1.09986532e+00 7.74323642e-01 -3.90337557e-01
1.92768797e-01 -6.62912309e-01 6.02543890e-01 2.95953471e-02
4.16862100e-01 3.39852005e-01 -2.56010383e-01 7.21159801e-02
-3.04230563e-02 4.57026333e-01 2.73574255e-02 -2.54491299e-01
8.05543065e-01 -8.61660913e-02 4.87449951e-02 2.08305389e-01
8.10273588e-01 1.88895315e-01 -1.68433130e+00 5.92982531e-01
-2.47695461e-01 -8.61430168e-01 -5.01533747e-01 -1.01122701e+00
-8.68926108e-01 9.60705400e-01 3.00249219e-01 3.11303556e-01
1.05041432e+00 -1.49880588e-01 5.95623851e-01 -1.97617769e-01
3.11954588e-01 -9.58666801e-01 6.55314326e-01 3.46573114e-01
1.31538045e+00 -9.47270870e-01 -8.23258385e-02 -5.82974732e-01
-6.94996834e-01 7.70154357e-01 1.16852129e+00 -7.72920251e-02
2.66462505e-01 8.05873513e-01 1.59040421e-01 -5.15777036e-04
-7.52276242e-01 -1.01627722e-01 1.61413804e-01 1.14507675e+00
6.65414035e-01 -5.41395210e-02 1.54940218e-01 -1.90156654e-01
-4.28614706e-01 -3.75676006e-01 5.47561347e-01 9.97470498e-01
-1.23372324e-01 -1.10026872e+00 -4.95354205e-01 5.03783226e-01
6.14347160e-02 -4.52896357e-01 -6.90533280e-01 1.16645741e+00
4.18046445e-01 7.95680821e-01 2.73772240e-01 -6.03204146e-02
5.24680674e-01 -4.04032439e-01 8.37105036e-01 -8.00514758e-01
-5.38934052e-01 1.67318895e-01 1.30474612e-01 -6.09310329e-01
8.63654688e-02 -5.40518761e-01 -8.50840628e-01 -4.48503166e-01
-3.47084478e-02 -2.79391378e-01 3.31782669e-01 1.99820951e-01
7.99755573e-01 1.05319190e+00 3.29660237e-01 -1.23294950e+00
-1.19418532e-01 -8.75283778e-01 -4.95074809e-01 6.00399613e-01
2.28688359e-01 -5.16999900e-01 -2.69845039e-01 2.31128097e-01] | [11.369099617004395, -0.2305523306131363] |
cb9eade9-ae22-4751-9426-a779c1429523 | instanceformer-an-online-video-instance | 2208.10547 | null | https://arxiv.org/abs/2208.10547v1 | https://arxiv.org/pdf/2208.10547v1.pdf | InstanceFormer: An Online Video Instance Segmentation Framework | Recent transformer-based offline video instance segmentation (VIS) approaches achieve encouraging results and significantly outperform online approaches. However, their reliance on the whole video and the immense computational complexity caused by full Spatio-temporal attention limit them in real-life applications such as processing lengthy videos. In this paper, we propose a single-stage transformer-based efficient online VIS framework named InstanceFormer, which is especially suitable for long and challenging videos. We propose three novel components to model short-term and long-term dependency and temporal coherence. First, we propagate the representation, location, and semantic information of prior instances to model short-term changes. Second, we propose a novel memory cross-attention in the decoder, which allows the network to look into earlier instances within a certain temporal window. Finally, we employ a temporal contrastive loss to impose coherence in the representation of an instance across all frames. Memory attention and temporal coherence are particularly beneficial to long-range dependency modeling, including challenging scenarios like occlusion. The proposed InstanceFormer outperforms previous online benchmark methods by a large margin across multiple datasets. Most importantly, InstanceFormer surpasses offline approaches for challenging and long datasets such as YouTube-VIS-2021 and OVIS. Code is available at https://github.com/rajatkoner08/InstanceFormer. | ['Volker Tresp', 'Thomas Seidl', 'Matthias Schubert', 'Sahand Sharifzadeh', 'Suprosanna Shit', 'Tanveer Hannan', 'Rajat Koner'] | 2022-08-22 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [-7.81214014e-02 -3.07544589e-01 -3.99280518e-01 -3.44956130e-01
-7.49047279e-01 -4.44452226e-01 4.47378308e-01 2.08018851e-02
-5.44481874e-01 5.13506532e-01 2.38736704e-01 -1.20170685e-02
-2.39656419e-02 -4.62381840e-01 -1.00032556e+00 -4.47481632e-01
-1.90053582e-01 8.95371586e-02 6.35611832e-01 9.10036191e-02
1.19352721e-01 1.98537141e-01 -1.34042764e+00 4.27115709e-01
7.71562457e-01 1.09088480e+00 6.30774498e-01 5.79425037e-01
7.16518462e-02 8.77020478e-01 -2.64853776e-01 -2.55939037e-01
8.73997658e-02 -3.00315529e-01 -6.91780090e-01 1.94417968e-01
5.88026345e-01 -5.56369364e-01 -8.92373800e-01 8.00237358e-01
5.17541170e-01 4.28813845e-01 3.64863761e-02 -1.10744584e+00
-4.37775671e-01 5.42481720e-01 -7.85956264e-01 7.57541895e-01
3.45385969e-01 4.27314401e-01 9.64472353e-01 -1.07052147e+00
6.97542906e-01 1.07333434e+00 4.68853474e-01 3.44271004e-01
-7.63777912e-01 -6.15442693e-01 9.48062658e-01 8.60894680e-01
-1.31329083e+00 -5.70628166e-01 6.78352058e-01 -3.79477561e-01
1.02214098e+00 1.85481429e-01 7.55927622e-01 1.08341873e+00
9.04929489e-02 1.33270264e+00 7.67821550e-01 1.13372453e-01
-2.24700812e-02 -3.43509316e-01 1.05790263e-02 6.23830020e-01
-3.26367408e-01 -9.38870292e-03 -8.00460935e-01 3.58434230e-01
7.27942109e-01 2.14968249e-01 -4.68426406e-01 -7.82930329e-02
-1.29074728e+00 3.93241316e-01 5.16134977e-01 1.99845150e-01
-4.44492400e-01 3.36671740e-01 6.01969302e-01 1.09892987e-01
6.40201032e-01 -5.30830882e-02 -5.11995077e-01 -3.83315474e-01
-1.24960661e+00 -1.76623370e-02 3.41511816e-01 1.07109761e+00
4.35075909e-01 -8.58732164e-02 -4.99854773e-01 8.64492297e-01
3.96737158e-01 2.28940830e-01 4.08180654e-01 -9.56740975e-01
6.69864476e-01 2.35906214e-01 -5.61832078e-02 -9.89848852e-01
-2.34241158e-01 -7.10826755e-01 -5.38868487e-01 -3.70124012e-01
2.55060136e-01 1.64764658e-01 -9.09682989e-01 1.66893315e+00
4.88652200e-01 9.34925735e-01 -4.10384238e-01 1.11735070e+00
8.73598278e-01 8.89885664e-01 1.07106462e-01 -3.59377146e-01
1.18898261e+00 -1.42208815e+00 -7.01622427e-01 -4.00961250e-01
4.03467834e-01 -6.82759106e-01 1.04634809e+00 3.27418059e-01
-1.30857897e+00 -5.70469856e-01 -7.30120718e-01 -3.14036846e-01
-6.59619048e-02 7.46295378e-02 4.69831198e-01 -3.79910469e-02
-8.78491342e-01 4.77824807e-01 -1.09139001e+00 -1.89467028e-01
6.61406517e-01 2.68855691e-01 -7.41058588e-02 -3.83419216e-01
-1.11510241e+00 4.37563926e-01 1.42059460e-01 4.56860453e-01
-1.08849490e+00 -7.87933588e-01 -9.31579232e-01 3.30842994e-02
6.90699577e-01 -4.34807897e-01 1.21531236e+00 -9.28518236e-01
-1.23378301e+00 5.62390268e-01 -6.07056320e-01 -5.69687247e-01
7.24339366e-01 -5.35330415e-01 -2.19492376e-01 3.51619750e-01
4.92155291e-02 6.68841541e-01 8.35862756e-01 -9.34460342e-01
-6.40544653e-01 -2.23241583e-01 4.04494137e-01 4.03951854e-01
-3.89990300e-01 2.72982568e-03 -1.32804203e+00 -8.73105943e-01
5.52673303e-02 -8.12777460e-01 -1.43669486e-01 2.08902597e-01
-2.25422025e-01 -7.00672045e-02 9.39518929e-01 -9.20576215e-01
1.49944830e+00 -2.30405855e+00 2.74688512e-01 -1.54136851e-01
2.19891340e-01 3.20967585e-01 -3.03995013e-01 2.14831322e-01
1.07716538e-01 -1.31967649e-01 -3.00251823e-02 -5.32315850e-01
-9.62640643e-02 2.44728208e-01 -2.82752931e-01 4.75408256e-01
6.88524172e-02 1.04824603e+00 -9.83672857e-01 -4.90493327e-01
4.09552723e-01 7.27950335e-01 -7.62980342e-01 1.10488743e-01
-3.45836282e-01 6.17944360e-01 -3.29345316e-01 5.19110858e-01
5.72091460e-01 -3.90046358e-01 -4.73191328e-02 -4.06262219e-01
-1.70022950e-01 3.28567564e-01 -9.37606037e-01 1.96334910e+00
-6.04992688e-01 9.54983771e-01 -6.04184046e-02 -9.05191123e-01
2.87745953e-01 3.16602379e-01 6.38328373e-01 -1.08335483e+00
5.55670336e-02 -1.13573400e-02 -3.13765854e-01 -6.42896771e-01
4.56314832e-01 3.61051619e-01 2.18823135e-01 1.92728505e-01
-3.19604464e-02 5.28519928e-01 5.13642490e-01 4.80736405e-01
9.77152526e-01 4.92214233e-01 -6.13304339e-02 9.46922675e-02
4.80340689e-01 -3.21013004e-01 9.43808734e-01 4.56915647e-01
-3.94284666e-01 8.22952747e-01 3.85649949e-01 -3.28162879e-01
-6.35264277e-01 -8.26408088e-01 6.73429593e-02 1.16266620e+00
6.78861380e-01 -4.68140781e-01 -5.61341584e-01 -6.96983159e-01
-1.83346495e-01 4.04395938e-01 -4.35944945e-01 -1.18192345e-01
-7.90408075e-01 -5.13844252e-01 1.16332799e-01 7.11576402e-01
5.62068045e-01 -9.22777295e-01 -4.50293064e-01 4.45224971e-01
-6.16581798e-01 -1.58073211e+00 -1.05263269e+00 -3.00241023e-01
-9.55333352e-01 -1.08889151e+00 -8.51903617e-01 -6.14216924e-01
5.52733719e-01 4.13334042e-01 9.77595925e-01 1.49846226e-01
-2.47616068e-01 4.24918085e-01 -4.63840395e-01 1.37241244e-01
3.60455602e-01 3.90968472e-02 -2.55994290e-01 2.20242530e-01
1.60327539e-01 -4.39228356e-01 -1.03655803e+00 4.48818505e-01
-8.40504646e-01 1.85793161e-01 3.71996403e-01 7.42215216e-01
7.99593270e-01 -1.24865383e-01 4.71683711e-01 -6.76965177e-01
1.19903482e-01 -6.45626605e-01 -4.39395487e-01 2.82729149e-01
-2.06899434e-01 -3.49082619e-01 4.72133845e-01 -5.48189819e-01
-1.05722880e+00 -1.24451190e-01 -2.27185875e-01 -6.80753171e-01
1.61150753e-01 5.31107962e-01 -1.44675016e-01 1.51002750e-01
-1.09614618e-03 3.42133194e-01 -3.96980345e-01 -5.41546881e-01
1.31131917e-01 2.48281017e-01 5.18721342e-01 -4.95450735e-01
4.33881998e-01 5.37834108e-01 -4.88887072e-01 -6.61893547e-01
-1.01325846e+00 -5.77087164e-01 -5.73140144e-01 -4.66616541e-01
7.59026110e-01 -1.21751976e+00 -5.80129445e-01 5.93794167e-01
-1.11586118e+00 -7.76208341e-01 -1.74049288e-01 5.88611603e-01
-4.78074640e-01 4.60832447e-01 -8.03425550e-01 -4.32095349e-01
-1.59243912e-01 -1.20646226e+00 9.90358949e-01 2.94574142e-01
2.60795485e-02 -1.02194750e+00 -2.48109996e-01 5.46448052e-01
1.93733081e-01 1.75528526e-02 2.45358840e-01 -1.46522745e-01
-9.91212726e-01 -1.68622192e-02 -3.56321514e-01 2.09471360e-01
-1.31690726e-01 6.44887090e-02 -7.15852916e-01 -4.80860621e-01
-2.25434303e-01 4.03580116e-03 1.14521301e+00 6.27036631e-01
1.50857842e+00 -3.49528998e-01 -3.43762785e-01 8.27554107e-01
1.26557875e+00 2.76177198e-01 6.20042503e-01 2.14049250e-01
8.83015513e-01 3.23583275e-01 1.05288041e+00 5.42487025e-01
6.76349998e-01 8.47309649e-01 5.01702249e-01 -1.10506721e-01
-4.19065565e-01 -1.79204538e-01 4.67875868e-01 1.04860210e+00
-7.50543922e-02 -5.03358424e-01 -6.62370384e-01 8.64410937e-01
-2.11405039e+00 -1.04375517e+00 -3.54737118e-02 2.16096067e+00
6.47318959e-01 2.18891487e-01 5.31262085e-02 -1.58403829e-01
6.70124590e-01 4.93726045e-01 -7.15955973e-01 -1.80135760e-02
2.01368369e-02 -2.32538749e-02 2.93715775e-01 6.53537095e-01
-1.11396098e+00 9.81556296e-01 5.10300112e+00 8.49699318e-01
-1.20373225e+00 5.60224414e-01 7.56753206e-01 -7.87459016e-01
-1.81536406e-01 -8.14159885e-02 -5.71421444e-01 8.47490132e-01
6.92211330e-01 6.20663632e-04 3.35479945e-01 4.71092612e-01
6.26881778e-01 -3.01077753e-01 -1.08910143e+00 1.01639640e+00
1.21852510e-01 -1.31822145e+00 -9.93172452e-02 -1.04856446e-01
7.13301718e-01 2.51249284e-01 1.75287113e-01 1.47709846e-01
-2.92758048e-01 -6.61724687e-01 9.63564515e-01 5.39840698e-01
6.44936800e-01 -7.48258948e-01 5.07425129e-01 1.21207304e-01
-1.55793917e+00 -1.28001839e-01 -6.92348331e-02 1.31869167e-01
5.11421621e-01 7.43803561e-01 -4.38000299e-02 4.22577292e-01
7.78094709e-01 1.25659168e+00 -3.80476743e-01 1.35802889e+00
-2.63361961e-01 6.67484462e-01 -3.56451660e-01 4.67154473e-01
4.37511712e-01 -1.56596210e-02 5.32473207e-01 1.29461825e+00
2.64835268e-01 2.65476644e-01 2.59787828e-01 4.01965290e-01
-9.19257551e-02 -6.12926334e-02 -2.03295082e-01 6.54180795e-02
5.01492620e-01 9.56056476e-01 -6.96069479e-01 -4.73915428e-01
-6.57289982e-01 1.20694876e+00 2.09868819e-01 6.15880668e-01
-1.41645098e+00 -3.79380696e-02 6.22399688e-01 1.59080341e-01
6.23030126e-01 -4.27423269e-01 -6.50182217e-02 -1.32478750e+00
4.32043165e-01 -6.99499309e-01 4.88494068e-01 -6.86611235e-01
-7.88673818e-01 6.21221483e-01 -1.76444173e-01 -1.26665807e+00
1.20626293e-01 -1.99757069e-01 -5.67415237e-01 4.18133080e-01
-1.79079843e+00 -1.07919180e+00 -4.66733575e-01 7.11267173e-01
1.12256420e+00 2.37511814e-01 4.37122397e-02 9.49738324e-01
-1.00518525e+00 6.11447930e-01 -5.17852865e-02 6.48995563e-02
7.50676692e-01 -7.80719042e-01 2.54775435e-01 1.04402673e+00
2.48046249e-01 4.54439849e-01 4.65548873e-01 -5.75412393e-01
-1.29294693e+00 -1.28638303e+00 8.22197020e-01 -2.16040313e-01
5.42495489e-01 -2.44712770e-01 -9.20872509e-01 7.70263016e-01
1.20008089e-01 3.97115409e-01 3.12666178e-01 -1.24733992e-01
-3.07472199e-01 -1.90491423e-01 -7.66183972e-01 5.92799067e-01
1.32018483e+00 -6.16435766e-01 -1.63067907e-01 4.41148251e-01
7.66617954e-01 -6.11730754e-01 -8.18444014e-01 3.91154557e-01
6.72963917e-01 -8.73620272e-01 9.31301534e-01 -2.34117463e-01
6.02195501e-01 -3.78501207e-01 8.19157586e-02 -8.74919415e-01
-2.34920278e-01 -8.17815959e-01 -6.26871765e-01 1.17042577e+00
2.44318247e-01 -4.88083571e-01 6.30117178e-01 5.34005582e-01
-2.49327600e-01 -1.08520281e+00 -9.05162811e-01 -7.28129745e-01
-3.41035753e-01 -6.93668485e-01 3.08827072e-01 8.19100976e-01
-1.48008108e-01 -3.80370952e-02 -5.41788638e-01 2.84062952e-01
5.95749557e-01 2.20618322e-01 4.19080585e-01 -5.50578356e-01
-4.22670841e-01 -2.37329841e-01 -3.63435388e-01 -1.75892854e+00
1.49435364e-02 -6.42985165e-01 2.00365465e-02 -1.61052918e+00
3.43957752e-01 -4.60222602e-01 -5.41494429e-01 3.88331622e-01
-3.31264555e-01 5.05549192e-01 3.81522864e-01 2.79484481e-01
-1.10622275e+00 7.37558663e-01 1.35480344e+00 -1.17368035e-01
-1.71460435e-01 -1.14070274e-01 -2.46083573e-01 6.56401336e-01
6.59874737e-01 -4.19359952e-01 -5.64288318e-01 -8.61186624e-01
5.09192422e-02 1.45855591e-01 5.41140616e-01 -9.25246477e-01
5.28406799e-01 -1.39577642e-01 1.64251477e-01 -7.63120055e-01
5.02006352e-01 -6.76004112e-01 9.73014235e-02 2.87458509e-01
-2.47704342e-01 1.01805225e-01 2.25349814e-01 7.50067651e-01
-4.71886426e-01 1.59977391e-01 6.78554833e-01 1.50667792e-02
-1.16973925e+00 8.70922267e-01 -1.83761656e-01 2.97868520e-01
1.19766700e+00 -3.54439288e-01 -2.22941294e-01 -3.46569210e-01
-9.56773102e-01 6.93304896e-01 2.74385929e-01 7.14860141e-01
7.90628850e-01 -1.15027761e+00 -5.53094804e-01 6.05341531e-02
-4.62488607e-02 1.06800482e-01 8.58170509e-01 1.41318679e+00
-3.68764997e-01 3.12060803e-01 2.86465418e-02 -7.50664294e-01
-1.43598306e+00 4.88916874e-01 2.07152680e-01 -8.11065584e-02
-8.72247159e-01 1.11720741e+00 6.15053058e-01 1.47156760e-01
4.20134634e-01 -2.84636647e-01 -3.34661789e-02 1.47864386e-01
6.01252973e-01 2.61115521e-01 -1.84170023e-01 -7.91440487e-01
-3.90629470e-01 5.86290061e-01 -2.51296490e-01 9.80608910e-02
1.30119669e+00 -5.01483500e-01 8.52182508e-02 3.00184190e-01
1.27069044e+00 -2.36469805e-01 -1.96664131e+00 -4.56099778e-01
-2.03609154e-01 -7.68348992e-01 7.97392204e-02 -4.94063437e-01
-1.55441260e+00 9.13748443e-01 4.06264991e-01 -3.07128668e-01
1.33199763e+00 3.67488153e-02 1.23942828e+00 -1.52917847e-01
2.60205477e-01 -1.21940124e+00 2.36255869e-01 5.50101340e-01
8.33591938e-01 -1.19664133e+00 4.18347381e-02 -5.29433608e-01
-4.94153172e-01 8.51339638e-01 7.04310000e-01 6.47085699e-06
5.83001614e-01 1.73544735e-01 -1.78860784e-01 -9.34721753e-02
-1.07657242e+00 -1.64413214e-01 4.40897614e-01 1.77231282e-01
3.77897650e-01 -2.00851843e-01 -3.30486774e-01 3.26863676e-01
3.64745289e-01 1.34989277e-01 3.25290889e-01 7.73419619e-01
-2.39333704e-01 -8.13644290e-01 -5.83511172e-03 3.20165396e-01
-6.01943135e-01 -1.89919844e-01 1.77062005e-01 5.32840550e-01
1.63349628e-01 7.81111717e-01 2.06165820e-01 -2.15479195e-01
1.99952871e-01 -3.61229211e-01 4.89620060e-01 -4.54360366e-01
-4.49013710e-01 3.64646286e-01 8.64780173e-02 -1.12717330e+00
-5.64282775e-01 -8.41684163e-01 -1.30427980e+00 -2.70902485e-01
-1.92074984e-01 -1.81642428e-01 4.32833523e-01 9.77252424e-01
5.47818780e-01 7.93419480e-01 5.36173522e-01 -9.97587860e-01
3.77988163e-03 -6.29499912e-01 -2.33850330e-01 3.28985482e-01
4.11537856e-01 -6.86830282e-01 -2.04880357e-01 2.30222762e-01] | [9.287270545959473, 0.10547833144664764] |
6e5406ce-9755-4753-8af3-74244f82d021 | deca-deep-viewpoint-equivariant-human-pose | 2108.08557 | null | https://arxiv.org/abs/2108.08557v1 | https://arxiv.org/pdf/2108.08557v1.pdf | DECA: Deep viewpoint-Equivariant human pose estimation using Capsule Autoencoders | Human Pose Estimation (HPE) aims at retrieving the 3D position of human joints from images or videos. We show that current 3D HPE methods suffer a lack of viewpoint equivariance, namely they tend to fail or perform poorly when dealing with viewpoints unseen at training time. Deep learning methods often rely on either scale-invariant, translation-invariant, or rotation-invariant operations, such as max-pooling. However, the adoption of such procedures does not necessarily improve viewpoint generalization, rather leading to more data-dependent methods. To tackle this issue, we propose a novel capsule autoencoder network with fast Variational Bayes capsule routing, named DECA. By modeling each joint as a capsule entity, combined with the routing algorithm, our approach can preserve the joints' hierarchical and geometrical structure in the feature space, independently from the viewpoint. By achieving viewpoint equivariance, we drastically reduce the network data dependency at training time, resulting in an improved ability to generalize for unseen viewpoints. In the experimental validation, we outperform other methods on depth images from both seen and unseen viewpoints, both top-view, and front-view. In the RGB domain, the same network gives state-of-the-art results on the challenging viewpoint transfer task, also establishing a new framework for top-view HPE. The code can be found at https://github.com/mmlab-cv/DECA. | ['Nicola Conci', 'Piotr Bródka', 'Niccolò Bisagno', 'Nicola Garau'] | 2021-08-19 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Garau_DECA_Deep_Viewpoint-Equivariant_Human_Pose_Estimation_Using_Capsule_Autoencoders_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Garau_DECA_Deep_Viewpoint-Equivariant_Human_Pose_Estimation_Using_Capsule_Autoencoders_ICCV_2021_paper.pdf | iccv-2021-1 | ['monocular-3d-human-pose-estimation'] | ['computer-vision'] | [-4.08536822e-01 -1.74858257e-01 1.17986582e-01 -2.11864948e-01
-6.62335575e-01 -4.99863416e-01 3.94013435e-01 -3.72981608e-01
-4.23335761e-01 4.24708068e-01 1.81119129e-01 2.91131020e-01
-1.10139094e-01 -5.91326118e-01 -8.70243371e-01 -8.14337611e-01
9.58347544e-02 4.55158770e-01 1.48025602e-01 -3.31761576e-02
-2.09996536e-01 6.03239834e-01 -1.34572852e+00 1.19339041e-01
2.37256020e-01 9.96288836e-01 4.32673469e-02 3.87742817e-01
5.61820984e-01 3.96421850e-01 -4.34058368e-01 -4.10962135e-01
3.51019949e-01 -2.02508301e-01 -5.58522999e-01 2.99257100e-01
4.09695029e-01 -4.72173840e-01 -6.18727863e-01 9.77576613e-01
7.45135844e-01 1.61922984e-02 4.75835741e-01 -1.12252557e+00
-4.62545186e-01 -1.16413511e-01 -2.75658607e-01 -1.81703135e-01
4.02248114e-01 -1.44885167e-01 9.03802276e-01 -1.06530786e+00
9.18080032e-01 1.11320150e+00 6.36957467e-01 5.91101050e-01
-1.04342330e+00 -1.70179024e-01 1.00726753e-01 2.50283360e-01
-1.51455736e+00 -2.72414893e-01 1.03767192e+00 -3.56350094e-01
6.52050138e-01 2.00705409e-01 9.00675535e-01 1.44823563e+00
4.59339947e-01 9.62990165e-01 1.06403828e+00 -1.56295046e-01
2.22824514e-01 -6.61296472e-02 -4.00276721e-01 7.48491645e-01
2.13899940e-01 7.64637068e-03 -5.83811522e-01 1.26829609e-01
1.10742557e+00 2.95267552e-01 -7.09475636e-01 -1.04116201e+00
-1.39781809e+00 7.56146848e-01 7.49077797e-01 2.66605794e-01
-3.06786925e-01 7.24528432e-02 2.76296884e-01 2.34238774e-01
4.48464304e-01 2.65886247e-01 -6.48251116e-01 1.34566754e-01
-4.42899048e-01 2.47488454e-01 7.42422342e-01 9.21982169e-01
4.47867423e-01 -1.52846947e-01 3.77838537e-02 5.45675039e-01
5.51328301e-01 3.74034405e-01 5.01226664e-01 -9.80357349e-01
2.66024888e-01 4.74121243e-01 7.19689727e-02 -9.97960567e-01
-6.73004389e-01 -7.29613364e-01 -8.63071680e-01 3.27966183e-01
4.52423722e-01 5.55275865e-02 -9.41233337e-01 1.81822050e+00
4.95784163e-01 -2.05702797e-01 -8.87594372e-02 1.22909951e+00
8.40525687e-01 2.06609026e-01 -4.21640098e-01 1.56873197e-03
1.57339358e+00 -9.73999739e-01 -5.18799603e-01 -1.64823636e-01
2.90021092e-01 -7.81824887e-01 7.87110925e-01 7.17204273e-01
-9.40753937e-01 -5.05569398e-01 -1.05068839e+00 -3.70606571e-01
-3.10687572e-01 3.12777609e-01 4.05202657e-01 3.48289400e-01
-8.81308138e-01 6.13860369e-01 -1.36781573e+00 -1.63827538e-01
2.31917232e-01 3.79971772e-01 -7.85534859e-01 -2.06496775e-01
-1.04879189e+00 8.38590980e-01 9.61867124e-02 4.36315507e-01
-8.00606132e-01 -4.12135392e-01 -1.08316958e+00 -9.00545809e-03
5.20613194e-01 -1.13401592e+00 9.62393641e-01 -5.49797893e-01
-1.68633032e+00 7.92877495e-01 4.62781033e-03 -9.94880497e-02
8.60161364e-01 -4.98561800e-01 -8.29748437e-02 3.67253423e-01
-7.01505840e-02 6.47940814e-01 9.97322619e-01 -1.26626050e+00
-6.65919334e-02 -7.92112350e-01 2.26446360e-01 3.19673687e-01
-2.29609519e-01 -4.43479151e-01 -8.09691012e-01 -7.31249750e-01
5.54123521e-01 -1.20269871e+00 -4.64570485e-02 4.19038534e-01
-4.36615467e-01 -1.09979399e-01 5.59666038e-01 -8.36041749e-01
6.66509330e-01 -2.23520470e+00 7.86880255e-01 -8.24985132e-02
2.70757258e-01 -9.61026698e-02 2.07042977e-01 2.00336173e-01
-4.75508720e-02 -3.20390671e-01 -9.80125833e-03 -6.84476435e-01
1.66107323e-02 3.14350069e-01 2.23647580e-01 8.93656611e-01
5.97916096e-02 8.13623607e-01 -7.65839338e-01 -4.11043346e-01
4.69112605e-01 9.45663631e-01 -7.66813934e-01 1.37173623e-01
-6.06674366e-02 8.95836294e-01 -3.59993309e-01 7.04504192e-01
6.31905258e-01 -4.60947633e-01 1.45628959e-01 -5.07008970e-01
1.73573583e-01 3.58257554e-02 -1.24049926e+00 2.26425862e+00
-5.33060491e-01 4.10809010e-01 1.66924179e-01 -9.22780752e-01
5.43050885e-01 5.31192124e-01 5.83781600e-01 -5.61596036e-01
2.68599838e-01 3.27215552e-01 -1.98797613e-01 -4.68635052e-01
9.71696153e-03 -1.74528524e-01 -8.21010694e-02 -2.01976717e-01
3.59998554e-01 2.47099474e-02 -2.42453977e-01 -2.41075084e-01
6.74034417e-01 4.78578955e-01 2.32499719e-01 4.01550308e-02
4.51994866e-01 -3.98448497e-01 6.14340246e-01 2.68063009e-01
-1.00783542e-01 9.72182155e-01 3.69126558e-01 -5.73088944e-01
-9.10706818e-01 -1.37339866e+00 -2.62434602e-01 4.60010588e-01
2.64384955e-01 -2.65581489e-01 -6.11831129e-01 -6.45459890e-01
-1.06610954e-02 2.78895646e-01 -6.84274077e-01 -1.44845247e-01
-6.03959203e-01 -4.77334052e-01 1.66574847e-02 5.41799307e-01
4.69306976e-01 -7.35607147e-01 -6.90207899e-01 1.68429613e-01
-2.59192199e-01 -1.38420630e+00 -3.39150757e-01 3.63449706e-03
-9.57916379e-01 -1.07990360e+00 -1.20793474e+00 -6.61168039e-01
6.83578193e-01 1.81214899e-01 9.13235188e-01 -2.25636780e-01
-2.28244290e-01 4.64924037e-01 -2.64079779e-01 -5.61899394e-02
5.55456616e-02 4.36027497e-02 1.70679837e-01 9.02647153e-02
9.45619941e-02 -7.54738927e-01 -9.54260945e-01 5.03961921e-01
-7.40170479e-01 -1.26099557e-01 7.16314852e-01 9.21735108e-01
6.88991904e-01 -1.42221764e-01 4.86954376e-02 -2.22793683e-01
-3.72709613e-03 -8.91694427e-02 -5.65845907e-01 1.85437158e-01
-1.98906437e-01 2.09415574e-02 6.37546062e-01 -3.06698412e-01
-7.42110968e-01 2.22211182e-01 -3.01747590e-01 -8.74312997e-01
-1.73259348e-01 3.70623499e-01 -3.85988832e-01 -1.21814243e-01
4.59495425e-01 2.76514232e-01 1.97207004e-01 -7.16612101e-01
2.36324355e-01 1.51259452e-01 3.54721129e-01 -3.56621444e-01
8.50174606e-01 8.98413360e-01 1.39946803e-01 -7.56429374e-01
-7.03398585e-01 -4.03628439e-01 -8.44993293e-01 -2.18646154e-01
1.18849397e+00 -1.25576818e+00 -7.67412126e-01 5.05227625e-01
-1.01141524e+00 5.27077529e-04 2.94089243e-02 9.87063050e-01
-6.52218282e-01 5.44893444e-01 -6.91689432e-01 -2.34349564e-01
-2.20614880e-01 -1.39474869e+00 1.40393710e+00 2.26659700e-03
1.02767197e-03 -9.27013516e-01 -4.42056060e-02 4.50406641e-01
1.23462610e-01 5.54323196e-01 6.65946960e-01 -4.70206678e-01
-8.16563845e-01 -4.24134433e-01 1.70934021e-01 5.59443533e-01
-5.00452891e-03 -3.85140330e-01 -9.09393191e-01 -6.64955080e-01
3.22770357e-01 -2.43998006e-01 7.35738456e-01 5.05760312e-01
9.36610401e-01 -9.98375788e-02 -1.36572912e-01 8.12973022e-01
1.38213551e+00 -2.22102419e-01 5.06050050e-01 3.43881220e-01
8.07748497e-01 4.94453460e-01 3.70806128e-01 4.21302229e-01
3.53350550e-01 9.94181335e-01 7.03243196e-01 -8.40184540e-02
-4.98503149e-02 -1.35334179e-01 3.28852504e-01 9.38514948e-01
-3.90299708e-01 -1.29804030e-01 -6.05733335e-01 3.74132365e-01
-1.64415765e+00 -5.16779125e-01 1.50612503e-01 2.32913637e+00
5.26810765e-01 7.32802302e-02 3.14314775e-02 3.04079382e-03
2.94582784e-01 1.44712076e-01 -4.64079767e-01 1.94216624e-01
1.00641556e-01 1.09724522e-01 2.32367292e-01 4.51830953e-01
-1.24611640e+00 6.48530841e-01 4.49852610e+00 5.15025556e-01
-1.28950000e+00 3.11154455e-01 9.63720754e-02 -2.60028452e-01
-5.14896624e-02 -3.32254291e-01 -5.01046419e-01 2.37876713e-01
3.42693329e-01 4.28270280e-01 2.64862955e-01 9.59750891e-01
-1.42703950e-01 1.80831879e-01 -1.28643489e+00 1.07550180e+00
2.94444054e-01 -9.33359027e-01 -6.57162517e-02 2.75364250e-01
4.28489655e-01 8.46209303e-02 4.69148383e-02 1.85695320e-01
-3.22117150e-01 -6.56712294e-01 7.81361222e-01 4.83578920e-01
4.23483372e-01 -5.69198430e-01 8.70461404e-01 2.08660722e-01
-1.00699449e+00 1.32260188e-01 -4.20053393e-01 1.99383646e-01
1.84022218e-01 4.99065340e-01 -3.79002690e-01 9.43610311e-01
8.90412331e-01 7.77203262e-01 -4.84831661e-01 1.05099702e+00
-5.79749823e-01 -1.44268557e-01 -4.07133579e-01 1.74736917e-01
2.13760082e-02 -5.61077856e-02 7.86175907e-01 6.10348403e-01
4.63741958e-01 -1.74793482e-01 1.50668416e-02 6.67619109e-01
2.13085115e-02 -1.92843571e-01 -4.57095474e-01 4.37664747e-01
-1.93843413e-02 1.16857827e+00 -6.79416001e-01 -2.36308482e-02
-4.87204731e-01 1.39947450e+00 1.29308462e-01 4.94888365e-01
-9.30203438e-01 -7.49960020e-02 5.55076420e-01 -1.31583298e-02
7.16325939e-01 -5.08239686e-01 9.48753729e-02 -1.70757306e+00
5.56959093e-01 -8.57669771e-01 2.29723379e-01 -7.58120954e-01
-1.08334482e+00 6.22614622e-01 4.78370562e-02 -1.70315874e+00
-1.54040381e-01 -1.11507010e+00 -7.21703768e-02 3.55400950e-01
-1.28298354e+00 -1.23119104e+00 -4.40342009e-01 8.22089970e-01
4.07036543e-01 1.27880067e-01 8.28798771e-01 4.72792387e-01
-3.47000510e-01 5.07166624e-01 -1.14199636e-03 3.49894673e-01
9.63386059e-01 -1.25377023e+00 1.07188947e-01 5.84652305e-01
2.89004087e-01 7.96086013e-01 5.37519097e-01 -2.63158679e-01
-1.67943847e+00 -8.64450037e-01 6.55648351e-01 -6.42108381e-01
3.33544582e-01 -6.60504103e-01 -6.83470309e-01 8.44490290e-01
2.32099388e-02 3.40055108e-01 5.85962057e-01 9.42235887e-02
-3.45096231e-01 -9.50360447e-02 -9.10111547e-01 5.61299443e-01
1.04310167e+00 -5.43696225e-01 -6.78054154e-01 4.68199044e-01
6.25071228e-01 -7.55520523e-01 -1.30698192e+00 5.29934049e-01
7.96220899e-01 -1.08526421e+00 1.30115366e+00 -3.48800987e-01
5.17690897e-01 -3.53836566e-01 -2.78427780e-01 -1.20534277e+00
-2.48239785e-01 -2.81626195e-01 -2.44200826e-01 5.77898920e-01
8.09494331e-02 -6.24980152e-01 7.64902234e-01 2.33413279e-01
-3.27340156e-01 -1.03040099e+00 -1.34209156e+00 -1.06218767e+00
3.10519245e-02 -1.47460073e-01 1.63396761e-01 7.83983111e-01
-4.32100058e-01 2.51778185e-01 -5.11881709e-01 5.50362587e-01
7.69511878e-01 2.22659871e-01 8.47286582e-01 -1.17353475e+00
-6.18392050e-01 -2.09864989e-01 -7.24927723e-01 -1.17244911e+00
-5.68838008e-02 -7.34282315e-01 -7.67012611e-02 -1.49550104e+00
-8.06062147e-02 2.19999537e-01 -1.97074100e-01 3.59987438e-01
1.04531147e-01 4.19927627e-01 2.22001851e-01 2.06108496e-01
-5.55194676e-01 8.04882228e-01 1.61338758e+00 1.51987076e-02
-4.77789380e-02 1.23791330e-01 -2.45904505e-01 9.23875928e-01
5.46107531e-01 -5.13326943e-01 -2.09508374e-01 -6.37833536e-01
2.11758405e-01 1.77504390e-01 8.05421889e-01 -1.07051802e+00
1.59419775e-01 3.62593085e-01 8.16994071e-01 -4.80668634e-01
7.93614566e-01 -9.42749679e-01 2.18766749e-01 6.97478890e-01
1.74970672e-01 -1.35799162e-02 -3.20882164e-02 7.10585952e-01
-3.06480378e-01 4.97850925e-02 6.57595694e-01 -3.32649142e-01
-4.89019662e-01 5.68287194e-01 -4.73831929e-02 -1.00243308e-01
7.99485266e-01 -1.38663664e-01 -1.96623914e-02 -3.71150494e-01
-1.27153683e+00 -3.66295427e-02 6.51324987e-01 4.61336315e-01
6.84595823e-01 -1.43055809e+00 -5.80808461e-01 3.51986557e-01
2.04563260e-01 3.07931006e-01 5.33792377e-01 1.26228237e+00
-5.83798707e-01 3.76733065e-01 -2.65306205e-01 -9.28038299e-01
-1.02598917e+00 7.50940144e-01 4.77844983e-01 -2.46317074e-01
-8.33178639e-01 8.75036061e-01 3.49227101e-01 -5.71409225e-01
2.82376170e-01 -4.93342549e-01 6.11279011e-02 -1.51206180e-02
1.43839806e-01 6.41333163e-02 2.68528581e-01 -5.48574626e-01
-5.21769822e-01 9.40548420e-01 5.70255145e-02 -4.19050232e-02
1.34367132e+00 2.87791472e-02 1.40410278e-03 3.40702564e-01
1.60253680e+00 2.26371121e-02 -1.37280715e+00 -2.12648943e-01
-3.82683039e-01 -5.29277205e-01 7.29047209e-02 -4.59907442e-01
-1.31127739e+00 1.05666316e+00 7.73086071e-01 -1.31459609e-01
1.04238868e+00 1.94720715e-01 6.65243983e-01 4.70604718e-01
5.94657958e-01 -8.62635911e-01 2.58676887e-01 1.87235638e-01
1.35928857e+00 -1.38806200e+00 2.76274711e-01 -4.06748652e-01
-3.35329235e-01 1.22783709e+00 4.14666206e-01 -3.05239648e-01
8.17882001e-01 -1.76884189e-01 4.42768559e-02 -3.56459260e-01
-3.48295122e-01 2.10077632e-02 6.08706534e-01 4.23575819e-01
2.40807340e-01 2.37850938e-02 -1.13878384e-01 4.29646820e-01
-1.79303274e-01 -1.73603445e-01 1.73097327e-01 9.40589428e-01
-2.91033182e-02 -1.00752819e+00 -3.26587528e-01 -2.14235634e-01
-4.82995272e-01 2.28926331e-01 -2.51275629e-01 1.28644073e+00
5.24632037e-02 5.10225475e-01 -7.71948844e-02 -3.26773703e-01
5.43641090e-01 -2.81071160e-02 8.70308399e-01 -3.73616129e-01
-1.59102052e-01 6.42041683e-01 -1.38919607e-01 -8.98535430e-01
-5.98162889e-01 -6.27739668e-01 -9.35097098e-01 -6.44975081e-02
-2.43227750e-01 -2.02157438e-01 5.87235749e-01 8.24086905e-01
2.91204810e-01 5.77929914e-01 5.06776392e-01 -1.12012768e+00
-7.27153480e-01 -7.64343977e-01 -5.65885484e-01 5.42422295e-01
4.70618427e-01 -1.05107784e+00 -5.04503012e-01 -2.42438748e-01] | [6.961808204650879, -1.0654551982879639] |
e3783ef2-5d55-4658-8874-dceed9566434 | graph-graph-similarity-network | null | null | https://openreview.net/forum?id=R3a2G2tSf3c | https://openreview.net/pdf?id=R3a2G2tSf3c | Graph-Graph Similarity Network | Graph classification aims to predict the class label for an entire graph. Recently, Graph Neural Networks (GNNs)-based approaches become an essential strand to learn low-dimensional continuous embeddings of the entire graphs for graph label prediction. While GNNs explicitly aggregate the neighborhood information and implicitly capture the topological structure for graph representation, they ignore the relationships among graphs. In this paper, we propose a Graph-Graph Similarity Network to tackle the graph classification problem by constructing a SuperGraph through learning the relationships among graphs. Each node in the SuperGraph represents an input graph, and the weights of edges denote the similarity between graphs. By this means, the graph classification is then transformed into a classical node classification problem. Specifically, we employ an Adversarial Autoencoder to align embeddings of all the graphs to a same distribution. After the alignment, we design the Graph-Graph Similarity Network to learn the similarity between graphs, which function as the adjacency matrix of the SuperGraph. By running node classification algorithms on the SuperGraph, we can predict the labels of graphs. Experiments on five widely used benchmarks under a fair setting demonstrate the effectiveness of our method. | ['Hongfu Liu', 'Pengyu Hong', 'Han Yue'] | 2021-01-01 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [-6.95722848e-02 4.64104950e-01 -2.47447267e-01 -5.11122048e-01
2.11101964e-01 -5.83930552e-01 3.73208553e-01 3.87222886e-01
8.61105025e-02 1.42023712e-01 1.87844709e-02 -2.07729146e-01
-7.11750314e-02 -1.42341459e+00 -6.02726400e-01 -6.80799425e-01
-1.60945848e-01 3.74320328e-01 -3.25902477e-02 -1.05293477e-02
-1.37301683e-01 3.39974463e-01 -8.78059864e-01 -2.34184086e-01
6.91223383e-01 9.51344788e-01 -2.74547357e-02 4.90820438e-01
-2.57110655e-01 9.42922473e-01 -3.43794227e-01 -4.87480044e-01
1.57851070e-01 -5.30226946e-01 -7.22622395e-01 2.40706310e-01
4.45473254e-01 -1.33395903e-02 -1.19720209e+00 1.52260613e+00
2.48362720e-01 9.14362669e-02 7.40579009e-01 -1.70878994e+00
-1.25634444e+00 8.30792487e-01 -3.52423221e-01 3.45748775e-02
2.82945454e-01 -7.61619136e-02 1.45426893e+00 -3.43485236e-01
4.74575788e-01 1.16266680e+00 6.39104903e-01 3.58362079e-01
-1.25093472e+00 -5.57445407e-01 4.17652726e-01 1.67614520e-01
-1.50583875e+00 2.29902461e-01 1.20616567e+00 -5.76508343e-01
3.83989215e-01 1.33059800e-01 9.46675897e-01 8.61303091e-01
1.38840005e-01 4.24476832e-01 4.76966858e-01 -8.51158947e-02
1.11735463e-01 -1.96229026e-01 3.65796953e-01 1.14158583e+00
2.79745072e-01 -8.88498202e-02 2.49308079e-01 -3.02955601e-02
7.03651011e-01 4.91613746e-01 -3.58676851e-01 -6.48661077e-01
-9.45618093e-01 1.08578527e+00 1.23722494e+00 3.10120106e-01
-3.28904331e-01 2.83735901e-01 6.21433794e-01 4.97493804e-01
6.38346970e-01 1.60854936e-01 3.86081487e-02 6.73168421e-01
-1.06197916e-01 -1.89259171e-01 9.02429521e-01 9.75158930e-01
1.00190961e+00 6.30919859e-02 -5.18752187e-02 6.59480572e-01
5.21667957e-01 3.79343599e-01 4.62885231e-01 -5.19192934e-01
4.08955097e-01 1.16840661e+00 -6.03761852e-01 -1.96558988e+00
-4.26035225e-01 -5.07866681e-01 -1.27793050e+00 -2.59388596e-01
2.48951957e-01 6.94116801e-02 -7.67094672e-01 1.78932703e+00
3.90816808e-01 7.13961840e-01 -1.84511971e-02 8.41420650e-01
1.04479134e+00 6.56643391e-01 4.27825004e-02 2.42060259e-01
1.05695581e+00 -9.96109664e-01 -5.94722807e-01 -3.24154168e-01
7.88096488e-01 -8.82185157e-03 8.26609850e-01 -2.53899008e-01
-5.04778147e-01 -4.49785411e-01 -1.18033314e+00 1.04678407e-01
-4.05783266e-01 -2.61651427e-01 6.27743661e-01 2.53670245e-01
-1.01918936e+00 7.59969294e-01 -7.39337444e-01 -4.31713343e-01
3.73180360e-01 3.88517201e-01 -6.23425543e-01 -2.71209031e-01
-1.26565814e+00 4.36986417e-01 8.00334394e-01 1.53915152e-01
-5.53326786e-01 -2.81001031e-01 -1.37019050e+00 3.03532273e-01
2.11556628e-01 -5.13072252e-01 5.79715014e-01 -1.17197728e+00
-9.25786316e-01 1.05868173e+00 4.21717465e-01 -2.98741728e-01
-2.25873776e-02 5.78477919e-01 -6.92746937e-01 1.04280993e-01
7.16849118e-02 3.17311525e-01 7.24497318e-01 -1.06937599e+00
-4.64060545e-01 -3.95944595e-01 3.11181307e-01 7.04701319e-02
-6.50356114e-01 -6.02847457e-01 -4.38127548e-01 -6.35334492e-01
3.15198123e-01 -9.45150733e-01 -1.14138603e-01 -1.38545886e-01
-7.13908792e-01 -3.05649012e-01 7.77404547e-01 -7.30798781e-01
1.43713200e+00 -2.31336474e+00 4.62605387e-01 5.11163831e-01
9.49426174e-01 -5.60759231e-02 -5.52620232e-01 4.01971787e-01
-4.17485595e-01 1.69251971e-02 -2.21009344e-01 1.18465140e-01
1.12150356e-01 4.02805805e-01 -1.24622963e-01 8.15603316e-01
1.02101244e-01 1.18958056e+00 -1.24010777e+00 -4.19618815e-01
8.70100185e-02 3.53120893e-01 -4.23452288e-01 6.02035522e-01
-1.82964757e-01 2.08157256e-01 -7.26153910e-01 2.32213378e-01
6.67015731e-01 -6.61530674e-01 5.02971172e-01 -3.74152303e-01
6.63571000e-01 -2.97670346e-02 -9.41573858e-01 1.52334034e+00
-3.01400512e-01 4.33425784e-01 -8.45225155e-02 -1.66259420e+00
1.19967794e+00 -1.18404157e-01 4.69052702e-01 -4.66389328e-01
2.47082293e-01 -5.24525754e-02 1.98372066e-01 -3.10339212e-01
-1.17972277e-01 9.15856566e-03 -2.11117774e-01 5.52830338e-01
2.08662823e-01 1.42813399e-01 -3.32439430e-02 4.67068493e-01
1.29745781e+00 -4.19758230e-01 2.83250511e-01 -1.43671826e-01
7.31701612e-01 -4.67846304e-01 5.36330283e-01 2.50202715e-01
-1.60160735e-01 2.73502797e-01 7.96890140e-01 -8.56716931e-01
-8.28987062e-01 -1.12865376e+00 3.27293009e-01 8.96341920e-01
4.73214060e-01 -5.66427708e-01 -8.05923879e-01 -1.12970543e+00
2.63146788e-01 1.21694043e-01 -8.87547016e-01 -8.44131470e-01
-5.38017988e-01 -5.17139614e-01 2.64060318e-01 3.82781744e-01
1.93287358e-01 -1.02606618e+00 3.77893001e-01 5.10472357e-02
8.43394175e-03 -1.09815001e+00 -8.15926552e-01 -5.81076462e-03
-6.01536512e-01 -1.44179106e+00 -4.17697906e-01 -1.61710930e+00
1.04180896e+00 1.40999883e-01 1.10283089e+00 6.29905283e-01
4.69322950e-02 3.77502203e-01 -3.40286374e-01 2.51924872e-01
-5.56853592e-01 2.24029675e-01 -9.32298452e-02 5.25666296e-01
2.52929151e-01 -9.35822070e-01 -5.32331705e-01 2.44788930e-01
-8.26139152e-01 8.88885707e-02 3.14918131e-01 9.54268456e-01
7.39527225e-01 2.58660674e-01 5.13229787e-01 -1.37954497e+00
7.48671353e-01 -8.74810874e-01 -4.97918814e-01 3.22997391e-01
-6.16933942e-01 2.79807597e-01 1.08353341e+00 -4.38874215e-01
-1.78334936e-01 -2.42335349e-02 9.03289393e-02 -6.45094097e-01
1.17681183e-01 8.67242098e-01 -6.21799946e-01 -2.47385666e-01
2.71937728e-01 3.18905264e-01 1.27563894e-01 -1.57373592e-01
6.24346018e-01 4.27451789e-01 5.60793221e-01 -2.24066630e-01
1.07305276e+00 1.90772310e-01 2.11456552e-01 -3.88962001e-01
-8.98084819e-01 -1.40876055e-01 -6.03483677e-01 -4.92004082e-02
8.86074543e-01 -6.30625963e-01 -7.17597544e-01 3.42387885e-01
-8.94272089e-01 -3.40988815e-01 -8.61628428e-02 3.23664844e-01
-2.84317762e-01 5.32265186e-01 -7.87620902e-01 -9.56137478e-02
-1.65327355e-01 -8.26563120e-01 9.34670627e-01 2.27261364e-01
1.37501493e-01 -1.60809720e+00 3.25965613e-01 -8.57241377e-02
4.57543209e-02 6.25980616e-01 1.28367245e+00 -1.03055918e+00
-4.59915757e-01 -5.14507473e-01 -4.74696189e-01 3.59180927e-01
4.83170986e-01 -2.04822838e-01 -4.44822282e-01 -5.54323912e-01
-2.60675341e-01 -1.19033732e-01 6.30824387e-01 1.53858826e-01
1.35757196e+00 -5.37669122e-01 -6.35919154e-01 9.59442914e-01
1.38067925e+00 -3.42833549e-02 2.27805868e-01 1.30518386e-03
1.43466318e+00 4.94989008e-01 1.15176767e-01 9.43756998e-02
5.80256820e-01 3.16662550e-01 8.15458596e-01 -6.45749494e-02
6.60562590e-02 -8.18306208e-01 4.06080596e-02 1.42429864e+00
2.78511584e-01 -5.49566686e-01 -9.02364254e-01 1.28398940e-01
-1.72871315e+00 -5.76690018e-01 -1.22969493e-01 1.92982280e+00
2.26707995e-01 4.01701443e-02 7.84505624e-03 -7.00894818e-02
1.38901150e+00 6.46204352e-01 -7.55835176e-01 -5.69293536e-02
9.44789052e-02 -1.31175369e-01 4.01867568e-01 5.08880913e-01
-1.19429231e+00 9.13709939e-01 5.28126669e+00 4.98188138e-01
-1.10014260e+00 -1.12592742e-01 6.63631558e-01 5.52997887e-01
-5.17652810e-01 1.20572627e-01 -4.85911630e-02 7.39628255e-01
8.33320737e-01 -4.35598344e-01 7.98960924e-01 8.50675821e-01
-4.66525137e-01 8.37035239e-01 -1.25496137e+00 1.03048933e+00
2.11046129e-01 -1.05644357e+00 2.36398101e-01 8.46426263e-02
6.05915189e-01 -4.15395275e-02 -2.94576809e-02 4.98692155e-01
5.53319871e-01 -1.28216481e+00 2.43722796e-01 6.00095272e-01
7.88460732e-01 -7.96961427e-01 7.67903626e-01 2.55572587e-01
-1.75521314e+00 -1.14834368e-01 -6.22518241e-01 -5.16048633e-02
-4.71438356e-02 4.28644806e-01 -6.01649284e-01 7.59579062e-01
2.63570279e-01 1.35802066e+00 -6.40120685e-01 6.45572245e-01
-4.33735102e-01 4.34619725e-01 1.39499202e-01 1.92127731e-02
2.47442707e-01 -7.84765422e-01 2.75892138e-01 7.78292716e-01
1.14838272e-01 6.91138208e-02 7.23967195e-01 8.61983240e-01
-6.98718309e-01 9.93303284e-02 -9.44072247e-01 -5.33129692e-01
5.84451199e-01 1.38371515e+00 -7.45056868e-01 -1.46368608e-01
-5.84955871e-01 1.27234828e+00 1.05045581e+00 3.37904721e-01
-8.42363834e-01 -6.11811578e-01 6.15446448e-01 -7.02992827e-02
9.00526196e-02 2.98418533e-02 2.97609866e-01 -1.17393279e+00
-1.45249553e-02 -7.31935084e-01 7.72578955e-01 -6.02693856e-01
-1.72757256e+00 8.40880394e-01 -4.48198050e-01 -1.25045645e+00
1.26523077e-01 -6.39738679e-01 -9.28220510e-01 6.87585652e-01
-1.10061765e+00 -1.18752253e+00 -6.88423693e-01 6.50921226e-01
-2.20715106e-01 -3.24154913e-01 8.70631576e-01 3.09017152e-01
-7.03993976e-01 6.34350836e-01 1.78916827e-01 8.84502530e-01
1.94005951e-01 -1.46345758e+00 6.47245049e-01 5.56401610e-01
4.20229048e-01 4.02324617e-01 1.94308013e-01 -6.58973455e-01
-1.50869572e+00 -1.53074884e+00 5.11539042e-01 -1.82058334e-01
1.11177540e+00 -6.46298468e-01 -1.10878181e+00 1.01631784e+00
-1.69086471e-01 8.85744572e-01 6.22284174e-01 -4.01949622e-02
-5.80082417e-01 -2.21825287e-01 -9.40860093e-01 4.43776578e-01
1.37342834e+00 -9.90314424e-01 -3.21648717e-01 5.30015290e-01
1.18089187e+00 -3.71657521e-01 -1.19051456e+00 1.05032280e-01
2.95994133e-01 -5.75302839e-01 9.17265892e-01 -9.20820415e-01
5.28541744e-01 -2.17377707e-01 -9.29590017e-02 -1.78990173e+00
-6.22392774e-01 -8.93861055e-02 -2.01648206e-01 1.23318529e+00
1.44786006e-02 -8.95535171e-01 9.28802609e-01 2.49527127e-01
5.65182744e-03 -7.24730730e-01 -6.00790143e-01 -4.76402164e-01
-1.18856490e-01 6.02858663e-02 9.93781447e-01 1.53340769e+00
-1.41862765e-01 6.96020305e-01 -1.83200046e-01 4.72988456e-01
6.67363763e-01 3.44415754e-01 6.34885430e-01 -1.48959053e+00
-2.22102046e-01 -5.38431883e-01 -1.25731885e+00 -8.31706285e-01
9.93147314e-01 -1.71145236e+00 -3.18906754e-01 -1.61976552e+00
2.19850406e-01 -3.89113486e-01 -6.02976739e-01 3.46884966e-01
-3.28494936e-01 -1.06322862e-01 -4.25901860e-02 -1.30484864e-01
-6.61835253e-01 7.87471175e-01 1.34867609e+00 -6.81236506e-01
5.58360405e-02 -1.81159154e-01 -7.17871308e-01 6.41564846e-01
5.43590724e-01 -4.86890346e-01 -7.45722711e-01 -5.61662078e-01
1.99575841e-01 2.77034659e-02 3.67055327e-01 -7.59074807e-01
3.09468687e-01 -3.25066969e-02 1.09036788e-01 -7.35946223e-02
-1.09667040e-01 -1.07845736e+00 3.48976731e-01 5.33771992e-01
-4.53062892e-01 3.79320793e-02 -3.64365101e-01 1.09058511e+00
-3.77454638e-01 3.21019953e-03 6.38979197e-01 1.04711346e-01
-6.38053775e-01 1.12111890e+00 3.44927758e-01 3.10360134e-01
1.03615916e+00 4.58115013e-03 -2.96813428e-01 -4.51659411e-01
-8.84751141e-01 4.77929085e-01 5.79497516e-01 4.07852381e-01
6.41735554e-01 -1.79322135e+00 -5.79303384e-01 2.89612383e-01
3.65725070e-01 1.93876177e-01 1.70837432e-01 4.46202219e-01
-6.16462827e-01 -1.85178593e-02 -2.38529265e-01 -4.45108920e-01
-1.00697184e+00 8.11082661e-01 6.80779099e-01 -3.43718201e-01
-8.44980180e-01 7.59060919e-01 6.42364085e-01 -9.43365216e-01
1.49108216e-01 -1.12347320e-01 -3.91151547e-01 -1.82633355e-01
1.92297757e-01 1.06844887e-01 -2.78122127e-01 -8.54163349e-01
-2.29164317e-01 4.42075551e-01 -2.01005246e-02 6.27509415e-01
1.31236911e+00 -4.54853661e-03 -4.90416735e-01 3.72878432e-01
1.78358114e+00 -2.80705959e-01 -1.11244118e+00 -4.38919276e-01
-1.10824192e-02 -3.90498370e-01 -2.07203049e-02 -8.52138624e-02
-1.67906535e+00 8.23409081e-01 2.73072213e-01 6.46474004e-01
9.84596729e-01 2.64006406e-01 8.12522471e-01 3.41340095e-01
9.97433513e-02 -4.35355306e-01 7.29893818e-02 4.23478901e-01
6.42088175e-01 -1.19656217e+00 -2.28100821e-01 -5.21587193e-01
-3.80533278e-01 1.08121026e+00 7.68546820e-01 -5.96224606e-01
9.34197783e-01 -2.85181284e-01 -3.04356247e-01 -5.18574178e-01
-2.72195727e-01 6.69907331e-02 3.41543972e-01 7.34831691e-01
5.49780242e-02 4.39755499e-01 1.01247333e-01 7.91500688e-01
-1.47087485e-01 -5.29070318e-01 2.74722785e-01 3.79597455e-01
-2.22116470e-01 -9.95554090e-01 1.92006931e-01 7.14067578e-01
-1.25478078e-02 1.74927320e-02 -5.50431073e-01 6.44919991e-01
-1.25579640e-01 5.38216770e-01 2.52751052e-01 -9.69022334e-01
2.92125046e-01 -2.63722032e-01 1.79685131e-01 -8.37338924e-01
-2.11085752e-01 -6.36901319e-01 -2.63483107e-01 -4.79082316e-01
-5.56012541e-02 -8.83712098e-02 -1.30723870e+00 -5.14368892e-01
-2.49363348e-01 4.12963688e-01 1.51211604e-01 6.92542553e-01
3.84014994e-01 6.31197989e-01 1.05179644e+00 -4.81037199e-01
-4.34776127e-01 -7.70644248e-01 -1.12217808e+00 9.82201159e-01
2.26776034e-01 -5.82309663e-01 -5.71579754e-01 -2.29651868e-01] | [7.189666271209717, 6.2856950759887695] |
31ae3d0e-568b-41c8-9aab-56f14eee14e7 | skin-deep-unlearning-artefact-and-instrument | 2109.09818 | null | https://arxiv.org/abs/2109.09818v7 | https://arxiv.org/pdf/2109.09818v7.pdf | Skin Deep Unlearning: Artefact and Instrument Debiasing in the Context of Melanoma Classification | Convolutional Neural Networks have demonstrated dermatologist-level performance in the classification of melanoma from skin lesion images, but prediction irregularities due to biases seen within the training data are an issue that should be addressed before widespread deployment is possible. In this work, we robustly remove bias and spurious variation from an automated melanoma classification pipeline using two leading bias unlearning techniques. We show that the biases introduced by surgical markings and rulers presented in previous studies can be reasonably mitigated using these bias removal methods. We also demonstrate the generalisation benefits of unlearning spurious variation relating to the imaging instrument used to capture lesion images. Our experimental results provide evidence that the effects of each of the aforementioned biases are notably reduced, with different debiasing techniques excelling at different tasks. | ['Peter J. Bevan', 'Amir Atapour-Abarghouei'] | 2021-09-20 | null | null | null | null | ['skin-cancer-classification'] | ['medical'] | [ 8.26781213e-01 9.35202464e-02 -9.91405025e-02 -5.23616552e-01
-8.95527482e-01 -5.25806129e-01 5.81545293e-01 1.19320333e-01
-7.70109475e-01 7.98112035e-01 2.80098796e-01 -3.91675949e-01
-3.51498604e-01 -3.70268703e-01 -5.75782418e-01 -9.40843642e-01
3.65914702e-01 -8.70474428e-02 -7.57041425e-02 2.77901273e-02
4.95393485e-01 7.98912704e-01 -1.21984839e+00 3.31359625e-01
8.34083200e-01 7.10535169e-01 -4.50737625e-01 8.12502086e-01
8.29597786e-02 7.86323130e-01 -9.12916720e-01 -2.11909294e-01
4.01633173e-01 -9.84641537e-02 -5.05530298e-01 1.51581079e-01
1.19502676e+00 -6.60139501e-01 -1.09055355e-01 8.91981184e-01
7.64353991e-01 -4.71439868e-01 8.16913903e-01 -7.39563465e-01
-4.58828032e-01 2.56577253e-01 -5.61377466e-01 2.26861969e-01
-2.72131294e-01 8.14212799e-01 3.97282720e-01 -4.35854524e-01
7.16485798e-01 7.61359990e-01 1.33276701e+00 7.58802116e-01
-1.43911386e+00 -6.93789363e-01 -1.63842261e-01 -4.51007128e-01
-1.17697310e+00 -6.65268421e-01 4.31745261e-01 -7.27397621e-01
7.09788442e-01 4.69971836e-01 6.36174798e-01 1.39292085e+00
6.68024838e-01 2.83298194e-01 1.50076437e+00 -4.77396131e-01
2.70342771e-02 4.92975742e-01 3.78727257e-01 7.04385698e-01
6.46236598e-01 3.57864738e-01 -2.83618301e-01 -2.09936202e-01
8.42006564e-01 -2.13318065e-01 -2.73762763e-01 -1.77956492e-01
-8.35183024e-01 6.56277001e-01 6.10639632e-01 2.51876473e-01
-3.30904573e-01 3.25674057e-01 5.06367683e-01 6.31508306e-02
4.20542955e-01 9.79403019e-01 -2.82454342e-01 2.40868330e-01
-1.18857348e+00 -1.69934079e-01 4.20908183e-01 5.50797284e-01
3.15293431e-01 1.84056640e-01 -4.80409771e-01 8.90179217e-01
-2.02472627e-01 3.57382149e-01 4.58484858e-01 -7.68004358e-01
-3.29062670e-01 5.43169975e-01 -5.28863855e-02 -9.93591309e-01
-7.57752359e-01 -6.48505747e-01 -8.73893797e-01 6.05640352e-01
6.87950015e-01 -3.70983541e-01 -1.53445160e+00 1.42318237e+00
5.43619785e-03 -2.60144055e-01 -1.46910369e-01 7.31003821e-01
8.77640069e-01 -4.70138103e-01 5.43326080e-01 -3.98390414e-03
1.15026891e+00 -5.53873003e-01 -6.51984572e-01 -4.04162318e-01
4.83622700e-01 -8.17103267e-01 7.83403993e-01 4.76328313e-01
-7.90475249e-01 -2.36189693e-01 -1.22746694e+00 6.28584437e-03
-1.18593670e-01 3.20828944e-01 7.97326863e-01 9.76601064e-01
-9.87004220e-01 6.87163711e-01 -8.35284114e-01 -5.29758036e-01
9.58665609e-01 5.64682901e-01 -5.24472892e-01 7.25630252e-03
-7.43877888e-01 1.10157514e+00 5.02474122e-02 3.71641248e-01
-8.38246107e-01 -1.22952747e+00 -5.35809696e-01 -4.89432245e-01
1.84039563e-01 -5.60523927e-01 9.16379571e-01 -1.34382915e+00
-1.03059864e+00 1.22481942e+00 -3.63297947e-02 -4.49388504e-01
8.78660321e-01 -5.30899540e-02 -3.53897601e-01 -3.77386734e-02
-1.44187108e-01 6.90549433e-01 1.25644946e+00 -1.25330722e+00
-4.34515476e-01 -3.14711511e-01 -4.00960267e-01 -2.65373826e-01
-3.57375860e-01 -8.21821913e-02 7.33317360e-02 -7.28954256e-01
-2.36198768e-01 -1.10948479e+00 -3.57402444e-01 3.74920875e-01
-3.63584399e-01 5.01009345e-01 3.47561419e-01 -1.07844186e+00
9.68210518e-01 -2.14940882e+00 -4.01435077e-01 3.32934886e-01
3.57659817e-01 5.06397963e-01 -3.06432784e-01 4.84558493e-02
-1.93688527e-01 5.00876427e-01 -1.85015574e-01 1.13204688e-01
-3.74508172e-01 -4.65109050e-02 1.12035580e-01 7.72303402e-01
5.54612577e-01 8.36645305e-01 -5.81770062e-01 -2.44812772e-01
3.52172494e-01 6.86414957e-01 -1.74980268e-01 -1.44575447e-01
2.72985250e-01 3.43856364e-01 1.71203613e-01 7.51374662e-01
6.86844885e-01 -3.75903882e-02 1.35573417e-01 -4.31467891e-01
1.91297323e-01 6.91878870e-02 -6.74882591e-01 1.35557306e+00
-5.29863477e-01 8.89730930e-01 2.70463794e-01 -1.74999237e-01
7.59014130e-01 1.24210984e-01 2.73303479e-01 -4.07158047e-01
1.29363343e-01 3.43961060e-01 5.78757763e-01 -5.80708265e-01
3.40029866e-01 -6.22790039e-01 4.32160109e-01 1.07722692e-01
6.13058135e-02 -2.96760619e-01 -3.29386562e-01 -1.94799989e-01
1.15871465e+00 -3.25666338e-01 3.96795064e-01 -5.77652693e-01
1.88554287e-01 4.02256280e-01 5.10459185e-01 1.03379500e+00
-7.43983686e-01 9.23750281e-01 5.07115960e-01 -4.25827861e-01
-1.05908632e+00 -8.65212023e-01 -6.54143989e-01 4.49747384e-01
-4.48044509e-01 1.83757856e-01 -6.44105554e-01 -1.01728106e+00
3.35268766e-01 5.16666710e-01 -1.31010640e+00 -3.40603888e-01
-3.47163528e-01 -1.24115419e+00 9.16146934e-01 6.73121810e-01
1.17001094e-01 -7.01975942e-01 -7.94469833e-01 9.07993689e-02
5.04301727e-01 -6.94392860e-01 -1.77384377e-01 5.15729129e-01
-7.16799617e-01 -1.30015039e+00 -8.20639551e-01 -4.75927353e-01
1.05981112e+00 7.91333020e-02 9.27624524e-01 4.54244316e-01
-1.07853901e+00 4.29865599e-01 2.19084974e-03 -8.59174371e-01
-7.35506833e-01 1.96156174e-01 -2.52607763e-01 -1.88390151e-01
6.00327373e-01 -2.26292573e-02 -7.51417875e-01 6.51428998e-02
-1.10173213e+00 -3.52804512e-01 9.14686143e-01 1.20306146e+00
2.96943128e-01 3.64815071e-02 4.09728020e-01 -1.27226210e+00
8.08304429e-01 -1.08381324e-01 -2.26287454e-01 -1.31412502e-02
-8.09745729e-01 -9.11715701e-02 2.08170637e-01 -3.90792996e-01
-1.05785477e+00 2.26806663e-02 -1.21808633e-01 -9.11041498e-02
-8.54049981e-01 1.09234966e-01 3.50189596e-01 -7.39291608e-01
1.19658661e+00 -2.94434041e-01 6.29412472e-01 6.36064187e-02
-2.44975671e-01 6.09759748e-01 2.71338493e-01 -8.86059031e-02
5.62450826e-01 7.96730995e-01 1.98116273e-01 -8.12847018e-01
-9.67658043e-01 -2.62563169e-01 -5.65463185e-01 -1.95947021e-01
6.85228050e-01 -7.58090794e-01 -9.57742035e-02 7.22884059e-01
-7.18298376e-01 -4.23137724e-01 -2.86909550e-01 2.65034378e-01
7.18409643e-02 1.95670962e-01 -4.58153218e-01 -7.46996105e-01
-4.42598760e-01 -1.28366399e+00 8.69627237e-01 4.31530863e-01
-7.89163828e-01 -1.25521922e+00 -1.21138841e-01 3.55026901e-01
7.66544878e-01 5.97296953e-01 8.43651175e-01 -7.16709077e-01
1.57506578e-02 -5.05491614e-01 -2.05881104e-01 7.25420475e-01
6.40956283e-01 4.55638468e-01 -1.39944482e+00 -5.75377464e-01
-1.00825943e-01 -3.59922290e-01 1.08374476e+00 7.76254177e-01
1.07462537e+00 -7.07203196e-03 -2.51835436e-01 6.76783979e-01
1.58861732e+00 -1.22564472e-01 7.50306368e-01 4.81949985e-01
6.78811669e-01 8.56552243e-01 1.70473740e-01 -1.27436429e-01
-3.84498507e-01 2.13096887e-01 3.09223950e-01 -6.37420177e-01
-6.99758410e-01 8.10849145e-02 -7.60609359e-02 2.35620141e-02
-6.96434220e-03 2.70343363e-01 -9.87585425e-01 7.52164960e-01
-1.23975945e+00 -7.24068463e-01 -4.50175285e-01 2.20526719e+00
7.82951772e-01 2.44322151e-01 8.65402911e-03 -6.85818344e-02
5.61151445e-01 -6.83998596e-03 -7.59998024e-01 -4.08679605e-01
-2.67059237e-01 3.44916314e-01 1.07616162e+00 4.81590897e-01
-1.04336751e+00 6.60444915e-01 7.95664072e+00 5.84376574e-01
-1.47799170e+00 -2.98996747e-01 7.19455779e-01 -3.76210332e-01
-1.37230456e-01 -3.71875644e-01 -5.69477081e-01 2.02253491e-01
6.19082451e-01 3.46736491e-01 1.82015374e-01 4.15511310e-01
1.02327496e-01 -3.45426112e-01 -1.02688301e+00 6.04791999e-01
3.11176687e-01 -1.21776533e+00 -5.08822538e-02 1.64407492e-01
7.23968148e-01 -1.39199823e-01 5.10530531e-01 -1.83468774e-01
2.49036282e-01 -1.56705248e+00 1.40947685e-01 6.65149331e-01
8.72170806e-01 -6.09403670e-01 1.11299503e+00 -2.91007161e-01
-6.14048429e-02 -2.20620409e-02 -2.30738431e-01 1.17169462e-01
-5.19962430e-01 9.45873618e-01 -1.49214089e+00 2.96849370e-01
5.16760528e-01 3.61005962e-01 -1.16812670e+00 1.04358673e+00
-3.73008728e-01 8.65075290e-01 -7.17707127e-02 -8.03537294e-02
1.18463360e-01 2.48608992e-01 4.23187166e-01 1.30806565e+00
-4.55746539e-02 -6.36246264e-01 -7.31643200e-01 7.08141208e-01
2.24145457e-01 -7.02945739e-02 -5.46132267e-01 4.75606918e-02
1.31335318e-01 1.63491464e+00 -6.16335154e-01 5.47573939e-02
-9.31710675e-02 7.03603804e-01 1.59699827e-01 2.12355316e-01
-4.41592097e-01 -2.72373050e-01 7.00753272e-01 3.40304166e-01
1.52352169e-01 2.92223781e-01 -7.96689510e-01 -8.43355238e-01
5.43560646e-02 -1.34314442e+00 5.16088545e-01 -5.64809024e-01
-1.51863813e+00 3.10847431e-01 -4.71138209e-01 -8.00302029e-01
2.21302599e-01 -9.72882450e-01 -4.78984565e-01 9.59282041e-01
-1.59266138e+00 -1.23959780e+00 -5.14442027e-01 2.01928139e-01
1.38130158e-01 -1.82770833e-01 8.57205093e-01 -2.09795997e-01
-5.61020076e-01 8.46130371e-01 -7.15391487e-02 2.90214747e-01
1.41828096e+00 -1.32012129e+00 -5.11796288e-02 7.63016522e-01
-2.09070221e-01 9.84802246e-01 6.67603612e-01 -6.59758031e-01
-9.24188852e-01 -9.73780096e-01 2.98257738e-01 -7.66213179e-01
6.27046049e-01 -3.92324142e-02 -9.59876359e-01 6.27084255e-01
3.12553138e-01 -2.12499112e-01 1.21912825e+00 4.23383743e-01
-4.60972935e-01 -2.17820346e-01 -1.44276774e+00 7.19534993e-01
6.74488604e-01 -3.58613253e-01 -4.78769124e-01 2.28092223e-01
-9.67027694e-02 -4.68228400e-01 -8.59901488e-01 5.74459016e-01
7.04597831e-01 -9.48793650e-01 7.37987101e-01 -8.68371308e-01
7.34795630e-01 6.37839288e-02 3.34106654e-01 -1.58223796e+00
-4.59419459e-01 -2.57215530e-01 6.21536672e-01 1.14321756e+00
5.92958927e-01 -6.41671956e-01 9.78677452e-01 8.92041802e-01
1.08198129e-01 -4.63017434e-01 -8.25525701e-01 -4.61693108e-01
7.52884686e-01 -1.38330519e-01 2.71911889e-01 9.53893244e-01
-3.83158863e-01 -1.92398652e-02 -3.30914587e-01 1.40313715e-01
6.26933932e-01 -5.66288710e-01 7.20470130e-01 -1.05019104e+00
-1.23227887e-01 -5.95589042e-01 -4.90488619e-01 1.10514805e-01
-2.06112699e-03 -8.08542550e-01 1.56807482e-01 -1.18484628e+00
2.56933481e-01 -3.74808162e-01 -4.01122421e-01 4.75320399e-01
-5.83243906e-01 4.47840244e-01 1.82069317e-02 1.18566252e-01
1.90931380e-01 -3.31986666e-01 1.26971531e+00 -3.57901156e-01
7.68997222e-02 -3.38393182e-01 -1.15515506e+00 8.08180928e-01
9.91106987e-01 -5.73932052e-01 -2.72569992e-02 -6.17929399e-01
6.56304508e-02 -6.81049943e-01 7.90905774e-01 -1.16980970e+00
1.24500938e-01 -5.41513823e-02 9.55599368e-01 7.83415809e-02
-6.63793981e-02 -9.23436046e-01 1.49461448e-01 7.66982436e-01
-6.29461646e-01 -4.82955158e-01 7.51823246e-01 3.77043754e-01
1.67762935e-01 -3.66416186e-01 1.22057784e+00 -1.82038262e-01
-2.89754897e-01 -8.62404630e-02 -6.34684503e-01 -1.25425160e-01
1.00965118e+00 -1.49681628e-01 -5.86283267e-01 -1.08706236e-01
-6.96245551e-01 -1.51540205e-01 6.96544588e-01 2.04858035e-01
2.31264517e-01 -9.06297743e-01 -7.74025083e-01 2.22455993e-01
2.15932712e-01 -2.59061813e-01 5.30220151e-01 1.04867411e+00
-6.41672432e-01 3.20630640e-01 -5.42926311e-01 -6.65564775e-01
-1.46767306e+00 2.82703042e-01 7.38348842e-01 -2.98323810e-01
-3.27885449e-01 9.30149019e-01 -2.85734504e-01 -5.39913714e-01
1.60887808e-01 -8.24094489e-02 1.69465423e-01 -4.59885933e-02
4.68083829e-01 3.75301480e-01 5.51241159e-01 -6.50738999e-02
-3.39004487e-01 3.75581712e-01 -7.02448606e-01 1.36432827e-01
1.24910843e+00 3.56452912e-01 4.40772921e-02 6.40511513e-02
8.08296919e-01 1.53267473e-01 -1.28503704e+00 2.73105681e-01
-2.03179494e-01 -5.04763186e-01 3.38418841e-01 -1.50094211e+00
-9.93139029e-01 9.19994831e-01 9.36027110e-01 6.12795353e-02
1.01220679e+00 -6.30295277e-01 2.07497180e-01 -1.18742203e-02
1.55504458e-02 -1.26436532e+00 -2.55293876e-01 -1.35136947e-01
8.62654924e-01 -1.36269963e+00 4.19092596e-01 -4.12666678e-01
-6.19908154e-01 1.12152123e+00 6.49643660e-01 -3.02598685e-01
3.49844128e-01 6.55878246e-01 7.66745090e-01 -1.43973276e-01
-3.98322165e-01 -1.29994331e-02 3.41510028e-01 1.09872103e+00
5.14071584e-01 -7.85171539e-02 -5.49624860e-01 1.33553982e-01
3.27527821e-02 1.89002141e-01 7.71179676e-01 1.00938106e+00
3.16951796e-02 -8.86073947e-01 -5.53288877e-01 1.07227159e+00
-6.64765120e-01 -1.94030017e-01 -9.57553029e-01 1.20667017e+00
2.19287157e-01 7.23262489e-01 5.43871224e-02 -2.00229421e-01
2.48240754e-01 1.25629500e-01 5.31966448e-01 -6.05194390e-01
-1.15708101e+00 -1.25781754e-02 3.03362638e-01 -5.19881010e-01
-4.19791967e-01 -7.34061301e-01 -5.40059447e-01 1.17413945e-01
-4.59367216e-01 -4.76444602e-01 6.29127026e-01 6.75720870e-01
3.88840944e-01 9.72175658e-01 6.41643927e-02 -3.04513991e-01
-7.61960864e-01 -9.80146527e-01 -5.34910798e-01 4.82629985e-01
6.22202396e-01 -5.51877320e-01 -8.32013428e-01 1.50025524e-02] | [15.59247875213623, -2.924461841583252] |
9d17f11c-7917-4581-938e-7b12ac7956f3 | emotion-recognition-in-speech-using-cross | 1808.05561 | null | http://arxiv.org/abs/1808.05561v1 | http://arxiv.org/pdf/1808.05561v1.pdf | Emotion Recognition in Speech using Cross-Modal Transfer in the Wild | Obtaining large, human labelled speech datasets to train models for emotion
recognition is a notoriously challenging task, hindered by annotation cost and
label ambiguity. In this work, we consider the task of learning embeddings for
speech classification without access to any form of labelled audio. We base our
approach on a simple hypothesis: that the emotional content of speech
correlates with the facial expression of the speaker. By exploiting this
relationship, we show that annotations of expression can be transferred from
the visual domain (faces) to the speech domain (voices) through cross-modal
distillation. We make the following contributions: (i) we develop a strong
teacher network for facial emotion recognition that achieves the state of the
art on a standard benchmark; (ii) we use the teacher to train a student, tabula
rasa, to learn representations (embeddings) for speech emotion recognition
without access to labelled audio data; and (iii) we show that the speech
emotion embedding can be used for speech emotion recognition on external
benchmark datasets. Code, models and data are available. | ['Arsha Nagrani', 'Andrew Zisserman', 'Andrea Vedaldi', 'Samuel Albanie'] | 2018-08-16 | null | null | null | null | ['facial-emotion-recognition'] | ['computer-vision'] | [ 3.54467630e-01 4.87739503e-01 1.01823568e-01 -7.40072727e-01
-7.88405180e-01 -5.69255650e-01 5.48526645e-01 -2.06096917e-01
-3.64634097e-01 4.44010675e-01 1.95410222e-01 -1.20564796e-01
3.37382823e-01 -1.57357231e-01 -6.20639205e-01 -6.30732536e-01
-1.29148141e-01 3.68455976e-01 -3.88601243e-01 -2.19256163e-01
-3.18865687e-01 6.53346479e-01 -1.73387885e+00 5.10707140e-01
2.47372285e-01 1.62174535e+00 -4.35489535e-01 8.36022317e-01
-1.06292181e-01 1.12505579e+00 -5.11934638e-01 -4.81488436e-01
-1.61404893e-01 -3.69713962e-01 -1.19226933e+00 3.21875423e-01
3.55138421e-01 -1.71875358e-01 -6.73045143e-02 8.32614720e-01
5.83978772e-01 1.79520085e-01 9.58999038e-01 -1.60912168e+00
-6.07825875e-01 2.16375366e-01 -2.28502095e-01 -8.34166110e-02
2.39224464e-01 -2.90935576e-01 9.79801297e-01 -1.19790518e+00
6.89928770e-01 1.19741297e+00 5.55897057e-01 1.04434884e+00
-1.15581596e+00 -7.29875684e-01 1.66716635e-01 2.51915962e-01
-1.33078611e+00 -1.00111151e+00 8.89745057e-01 -4.57325518e-01
1.14034581e+00 2.08930910e-01 5.67300856e-01 1.54888272e+00
-6.16178513e-01 9.99331832e-01 1.01094115e+00 -5.75582802e-01
1.82471722e-01 4.73878175e-01 -8.05612281e-02 7.67343402e-01
-9.23153818e-01 1.27059281e-01 -7.16502011e-01 4.03900519e-02
8.41122419e-02 -5.70081592e-01 -2.55284876e-01 -3.26741546e-01
-6.49419010e-01 9.37520921e-01 3.06199491e-01 2.83768475e-01
-3.07235569e-01 1.66104853e-01 7.26564467e-01 6.65572822e-01
7.90475667e-01 2.41624162e-01 -6.87423110e-01 -5.15248001e-01
-6.65594935e-01 -3.91256899e-01 9.56247985e-01 5.54999411e-01
5.67006528e-01 3.42254370e-01 2.28311956e-01 1.23681283e+00
3.99285465e-01 3.86041105e-01 5.15326321e-01 -9.46802497e-01
-3.89351174e-02 2.28607431e-01 -2.84302086e-01 -8.23605478e-01
-2.86442757e-01 -3.14649455e-02 -5.54244459e-01 2.93332249e-01
1.76413149e-01 -4.83106136e-01 -6.68620110e-01 1.93978548e+00
3.10867667e-01 4.17828470e-01 4.43729609e-01 6.71971560e-01
1.06856692e+00 7.48330534e-01 2.80767560e-01 -2.24276364e-01
1.47551537e+00 -9.66994166e-01 -7.34036386e-01 -1.55062199e-01
6.73785567e-01 -6.16123140e-01 8.35076988e-01 5.06924689e-01
-8.78077030e-01 -3.97230029e-01 -6.86945081e-01 -1.11332484e-01
-6.68754518e-01 3.36060435e-01 6.01617634e-01 7.39094198e-01
-1.19235337e+00 1.72731638e-01 -7.10736513e-01 -5.22771299e-01
6.78077996e-01 4.09179628e-01 -8.31002951e-01 2.30083495e-01
-1.08019602e+00 9.07426000e-01 5.69516048e-02 1.04535431e-01
-1.12671971e+00 -5.40239692e-01 -1.20993853e+00 1.71784461e-01
4.60761376e-02 -2.41715252e-01 1.42814219e+00 -2.04686117e+00
-1.77934182e+00 1.21911037e+00 -1.00696936e-01 -1.35741681e-01
1.06798247e-01 3.06778271e-02 -4.85740811e-01 4.75445330e-01
-4.50537652e-01 1.04085851e+00 1.13127828e+00 -1.18244922e+00
-4.14827555e-01 -4.21938211e-01 -2.28123948e-01 1.00776376e-02
-7.53676951e-01 4.76978660e-01 -2.54172254e-02 -4.56248760e-01
-5.29156029e-01 -8.53497028e-01 2.00904116e-01 1.91289738e-01
-1.35958076e-01 -4.92275864e-01 1.09650171e+00 -4.89624649e-01
7.27490067e-01 -2.56169868e+00 2.13423461e-01 1.05812170e-01
-2.61720028e-02 3.73909354e-01 -5.03529549e-01 1.69949397e-01
-6.42580569e-01 -1.92252528e-02 -1.14256002e-01 -7.53339887e-01
3.54188532e-01 3.06762099e-01 -5.85684061e-01 3.98522466e-01
7.97287107e-01 8.81643414e-01 -6.57108545e-01 -4.40041244e-01
-4.94657364e-03 9.52617705e-01 -5.36272585e-01 5.49432755e-01
2.59204246e-02 3.31929922e-01 -1.57178342e-02 6.06229246e-01
2.62920022e-01 7.16625601e-02 1.25147536e-01 -2.41126031e-01
1.67319566e-01 3.19776744e-01 -9.39209938e-01 1.45536721e+00
-7.37397432e-01 8.32389593e-01 5.54583073e-01 -1.33830333e+00
1.03589332e+00 9.67627048e-01 4.46786493e-01 -3.91797006e-01
3.46039087e-01 1.23370305e-01 -2.87157893e-01 -6.78276718e-01
1.87209360e-02 -6.24687493e-01 -1.18000500e-01 5.49721181e-01
8.42354596e-01 -1.77747652e-01 -1.93874329e-01 -8.30571651e-02
1.00728893e+00 -1.06314324e-01 8.24427307e-02 9.01349187e-02
5.69899976e-01 -3.75208020e-01 2.05131605e-01 4.56511676e-02
-4.45601463e-01 1.36104241e-01 8.16924155e-01 -4.36092526e-01
-6.99123263e-01 -9.13425207e-01 -2.54349202e-01 1.74793530e+00
-7.09106147e-01 -3.00062954e-01 -7.65140653e-01 -9.14090574e-01
-1.73837826e-01 3.88670027e-01 -8.40657651e-01 -2.36632213e-01
-6.36334643e-02 -1.78977832e-01 8.88764441e-01 7.77802408e-01
-6.87471703e-02 -1.37256837e+00 -3.66815865e-01 -2.07200065e-01
1.66307781e-02 -1.32123590e+00 -2.36231789e-01 6.06137395e-01
-2.51070797e-01 -7.66322672e-01 -6.43600345e-01 -1.04130971e+00
6.05470777e-01 -4.05981660e-01 1.12915361e+00 -2.70812307e-02
-3.26914907e-01 1.11447263e+00 -3.84044021e-01 -6.94247782e-01
-3.79664898e-01 -9.55732018e-02 1.83294602e-02 3.78156573e-01
4.90765452e-01 -6.05270982e-01 -1.07694156e-01 1.77233934e-01
-8.31402719e-01 -2.81323045e-01 1.13333382e-01 9.28415477e-01
2.74088264e-01 -2.40795940e-01 8.23393166e-01 -5.05847335e-01
5.05101800e-01 -3.53458941e-01 -3.40982020e-01 3.85552719e-02
-9.11192745e-02 -1.44635633e-01 4.42628235e-01 -5.36129236e-01
-9.90832865e-01 5.56004345e-01 -6.10200226e-01 -6.70162320e-01
-7.10833132e-01 4.00824517e-01 -2.44633600e-01 -2.08233461e-01
3.09686929e-01 -1.76233128e-01 3.18803936e-01 -2.62089312e-01
6.12706780e-01 1.08621049e+00 5.72634280e-01 -7.00365841e-01
1.80241048e-01 3.67226958e-01 -1.36039585e-01 -1.12345541e+00
-8.11220706e-01 -2.97103226e-01 -7.09973574e-01 -3.59583765e-01
1.14678717e+00 -1.05144942e+00 -9.27708864e-01 1.49011299e-01
-1.22728074e+00 -5.38184226e-01 -2.97921032e-01 4.23173457e-01
-8.45703423e-01 1.00052550e-01 -7.42264092e-01 -1.12663019e+00
-1.61020279e-01 -9.46780384e-01 1.28920925e+00 -4.57050018e-02
-5.45748293e-01 -1.13140404e+00 1.47086427e-01 3.60341102e-01
2.19869807e-01 1.33572847e-01 7.90323138e-01 -9.48070765e-01
2.02290386e-01 -1.07727975e-01 -2.29823083e-01 7.23948300e-01
-5.93710393e-02 2.17705175e-01 -1.69283307e+00 2.41710749e-02
-1.42201647e-01 -1.33174706e+00 7.57169366e-01 -3.58982496e-02
1.27196360e+00 -1.37943655e-01 7.08834082e-02 6.14952207e-01
6.85742140e-01 -1.32955350e-02 4.37483132e-01 -6.91782907e-02
4.36051697e-01 1.16974378e+00 2.45430902e-01 4.63755399e-01
2.99883932e-01 7.26341009e-01 3.11140150e-01 -3.64550710e-01
7.97357485e-02 -1.31761432e-01 6.61746800e-01 7.83981681e-01
1.88839510e-01 -7.16943070e-02 -7.76776433e-01 6.96632445e-01
-1.50271928e+00 -9.61805522e-01 3.11194003e-01 1.55722368e+00
9.05809224e-01 -5.38908958e-01 3.11790437e-01 1.69874549e-01
2.17513815e-01 7.53231421e-02 -1.53795376e-01 -9.77763414e-01
2.10097939e-01 6.29981935e-01 -2.42843017e-01 5.71907401e-01
-1.38613105e+00 1.03827429e+00 6.18727446e+00 5.23042500e-01
-1.46335816e+00 2.70446569e-01 6.67853236e-01 -2.17813805e-01
2.86494344e-02 -5.68465471e-01 -4.92969245e-01 5.20324335e-02
1.43626213e+00 2.33255029e-01 4.64735627e-01 1.01219571e+00
-2.08570629e-01 5.15461445e-01 -1.29244065e+00 1.14729881e+00
3.73289049e-01 -7.53629386e-01 -2.76164621e-01 -1.48726732e-01
3.18714023e-01 4.40627337e-02 2.28902102e-01 6.78145409e-01
3.41144055e-01 -1.60791254e+00 5.32558501e-01 2.42399991e-01
8.96776319e-01 -9.08367038e-01 6.70979381e-01 1.78127646e-01
-1.03230464e+00 -9.78675559e-02 -1.46205842e-01 1.01331405e-01
1.48835585e-01 7.93328062e-02 -8.95345330e-01 1.69190377e-01
6.87125564e-01 7.47823596e-01 -2.97041953e-01 2.01559976e-01
-4.78999615e-01 7.14283228e-01 -2.46944532e-01 1.24418333e-01
1.91706687e-01 1.74494594e-01 1.83807071e-02 1.48644197e+00
2.76084870e-01 1.66193768e-02 7.34304339e-02 6.07376993e-01
-4.72132415e-01 2.43033454e-01 -7.96478033e-01 -5.79404294e-01
5.65684773e-02 1.47940469e+00 -3.24420959e-01 -2.35674217e-01
-5.08520842e-01 1.10415781e+00 6.96032584e-01 3.37747097e-01
-5.55015445e-01 -3.24098706e-01 9.63093102e-01 -4.33084190e-01
6.74694777e-01 2.14998171e-01 2.74065912e-01 -9.85237658e-01
-2.74962097e-01 -1.17559624e+00 4.51822251e-01 -9.36885118e-01
-1.42233026e+00 9.52414155e-01 -4.29135174e-01 -7.53273368e-01
-4.82277900e-01 -1.06968451e+00 -4.50455844e-01 7.58942008e-01
-1.56614208e+00 -1.20169842e+00 -1.17661111e-01 7.58191764e-01
2.79403895e-01 -3.28454584e-01 1.49440861e+00 4.41971123e-01
-5.24371564e-01 7.32833743e-01 -4.76629853e-01 5.32687306e-01
8.38792622e-01 -1.18181229e+00 -1.22590132e-01 1.14992894e-01
6.29604816e-01 1.92308456e-01 4.77561116e-01 6.22955598e-02
-1.27469456e+00 -9.36782420e-01 1.13870358e+00 -5.63402891e-01
9.38943207e-01 -7.29412377e-01 -8.36741507e-01 8.78537118e-01
4.65736836e-01 5.96224785e-01 1.37248814e+00 4.19112623e-01
-6.95008218e-01 -7.93359503e-02 -1.09714317e+00 1.10316433e-01
6.27395272e-01 -1.23682392e+00 -4.86009568e-01 1.30777463e-01
5.03515959e-01 -1.56165063e-01 -9.64371026e-01 2.20901310e-01
7.37295270e-01 -6.25330329e-01 8.32308352e-01 -1.22242999e+00
6.52739167e-01 8.78231749e-02 -3.84735823e-01 -1.49800730e+00
2.66208857e-01 -4.87641484e-01 1.30184377e-02 1.51162076e+00
5.60503840e-01 -4.69461888e-01 4.50424463e-01 5.23282051e-01
-5.61757684e-02 -8.39445829e-01 -1.23149681e+00 -5.29910922e-01
8.77607092e-02 -7.53013611e-01 3.45475256e-01 1.24144053e+00
2.94966996e-01 7.37231910e-01 -3.51712823e-01 1.59288421e-01
-3.81244197e-02 -1.48550838e-01 7.54340172e-01 -1.21540964e+00
-2.72754997e-01 -4.11242157e-01 -6.40212238e-01 -7.67172515e-01
1.17403758e+00 -1.03954029e+00 2.15865687e-01 -9.29314435e-01
-1.32802784e-01 3.96729968e-02 -4.03715342e-01 7.97988296e-01
3.50104839e-01 5.03237247e-01 6.83715120e-02 -4.86190081e-01
-7.30825543e-01 8.96634519e-01 6.98089480e-01 -1.96640775e-01
2.70640373e-01 -2.13608190e-01 -6.13950729e-01 8.77008140e-01
3.87828678e-01 -4.86251235e-01 -3.44353378e-01 -2.19536439e-01
2.61577129e-01 1.29740357e-01 4.46879476e-01 -6.02642953e-01
-4.42090631e-02 3.37776721e-01 3.98126483e-01 1.42264925e-02
7.69315302e-01 -1.22213161e+00 -3.45956773e-01 -8.09542313e-02
-7.03105867e-01 -1.31905824e-01 6.12712681e-01 3.61777246e-01
-5.58143079e-01 -4.13199335e-01 8.33269536e-01 3.65000635e-01
-6.24822497e-01 2.14511380e-01 -5.41204393e-01 3.78524363e-02
9.27889943e-01 2.87358373e-01 -1.21289968e-01 -7.61299074e-01
-1.13896847e+00 1.15359209e-01 1.78806320e-01 4.44806993e-01
4.80127186e-01 -1.53766441e+00 -4.67267901e-01 4.21767533e-01
3.66016120e-01 -5.39453506e-01 1.25273257e-01 9.51163173e-01
9.77828056e-02 2.70470858e-01 -1.42003223e-01 -5.23743987e-01
-1.80184162e+00 5.27294695e-01 5.27398944e-01 1.44639209e-01
-6.42455667e-02 1.26157558e+00 2.42115930e-01 -9.29440320e-01
6.10101879e-01 7.93140754e-03 -2.26514846e-01 5.48675597e-01
8.06143284e-01 -2.82980174e-01 6.30394816e-02 -1.03950047e+00
-4.01506454e-01 3.25351030e-01 2.06479937e-01 -2.28704542e-01
1.64804661e+00 6.09729961e-02 -9.61601958e-02 5.84916770e-01
1.67540967e+00 -1.65390596e-01 -1.09204113e+00 -5.58948517e-02
2.30434397e-03 -7.74030387e-02 1.52217969e-01 -8.44494641e-01
-1.12520742e+00 1.39826143e+00 8.00374687e-01 2.82963574e-01
1.29353213e+00 3.96825939e-01 5.03715515e-01 4.48824346e-01
-1.89722627e-01 -1.13792861e+00 3.18587095e-01 7.10200012e-01
1.01066017e+00 -1.28383207e+00 -6.36076331e-01 -1.71256468e-01
-1.04683471e+00 1.26189446e+00 3.83385569e-01 1.41590670e-01
7.64313042e-01 4.24350590e-01 5.36878467e-01 -4.82707709e-01
-1.23064649e+00 -3.32782477e-01 3.59886020e-01 9.25141573e-01
7.97091722e-01 -1.06028803e-01 4.64972734e-01 8.05410504e-01
-2.21231371e-01 -4.88333553e-02 8.62216800e-02 5.83402812e-01
-2.36170456e-01 -1.08455825e+00 -1.25147521e-01 -9.36270431e-02
-6.75402403e-01 -1.36492411e-02 -9.60592091e-01 5.37754357e-01
6.16079494e-02 1.00020897e+00 3.52087319e-01 -1.27950147e-01
3.88722062e-01 8.28898132e-01 4.86848533e-01 -6.52709842e-01
-5.30477226e-01 2.48221401e-02 5.11501193e-01 -5.52641571e-01
-5.23706377e-01 -4.65045601e-01 -1.07271314e+00 1.54616877e-01
-4.93183993e-02 2.79182315e-01 8.09778094e-01 1.00766277e+00
4.03587699e-01 5.42867780e-01 6.75677180e-01 -9.02630448e-01
-4.50889736e-01 -9.65697527e-01 -6.64048791e-01 4.06741977e-01
5.87707162e-01 -6.68476224e-01 -6.22738063e-01 3.30707014e-01] | [13.550518989562988, 5.5705766677856445] |
f88e372c-af3a-4923-a923-cfe1cf99d341 | improving-pre-trained-weights-through-meta | 2212.09447 | null | https://arxiv.org/abs/2212.09447v1 | https://arxiv.org/pdf/2212.09447v1.pdf | Improving Pre-Trained Weights Through Meta-Heuristics Fine-Tuning | Machine Learning algorithms have been extensively researched throughout the last decade, leading to unprecedented advances in a broad range of applications, such as image classification and reconstruction, object recognition, and text categorization. Nonetheless, most Machine Learning algorithms are trained via derivative-based optimizers, such as the Stochastic Gradient Descent, leading to possible local optimum entrapments and inhibiting them from achieving proper performances. A bio-inspired alternative to traditional optimization techniques, denoted as meta-heuristic, has received significant attention due to its simplicity and ability to avoid local optimums imprisonment. In this work, we propose to use meta-heuristic techniques to fine-tune pre-trained weights, exploring additional regions of the search space, and improving their effectiveness. The experimental evaluation comprises two classification tasks (image and text) and is assessed under four literature datasets. Experimental results show nature-inspired algorithms' capacity in exploring the neighborhood of pre-trained weights, achieving superior results than their counterpart pre-trained architectures. Additionally, a thorough analysis of distinct architectures, such as Multi-Layer Perceptron and Recurrent Neural Networks, attempts to visualize and provide more precise insights into the most critical weights to be fine-tuned in the learning process. | ['Claudio F. G. dos Santos', 'João Paulo Papa', 'Mateus Roder', 'Gustavo H. de Rosa'] | 2022-12-19 | null | null | null | null | ['text-categorization'] | ['natural-language-processing'] | [ 3.25170428e-01 -1.35353431e-01 -2.86861449e-01 -1.78221881e-01
-7.89477900e-02 -9.49689224e-02 4.69585776e-01 2.72357911e-01
-7.11301684e-01 7.49691784e-01 -3.26581746e-02 -3.58157098e-01
-5.42686760e-01 -5.95350206e-01 -3.54217380e-01 -8.74783754e-01
4.81843725e-02 1.51312038e-01 2.24846359e-02 -2.78993577e-01
9.53157544e-01 5.89700162e-01 -1.75949347e+00 1.39268786e-01
9.68514502e-01 9.39388275e-01 2.84088075e-01 3.55742782e-01
-4.00687963e-01 5.21833301e-01 -5.82644522e-01 -5.75940311e-01
-5.91605566e-02 -3.92908394e-01 -6.35622144e-01 -1.45463675e-01
-2.45059654e-01 3.47673565e-01 1.86998583e-02 1.07762408e+00
6.16285622e-01 3.02304655e-01 5.53052008e-01 -7.59133577e-01
-8.67883086e-01 6.59731150e-01 -4.12836045e-01 4.38524365e-01
1.39661297e-01 1.84292591e-03 9.49343741e-01 -6.98390245e-01
1.89745665e-01 1.13036966e+00 6.42067730e-01 6.67855918e-01
-1.07929504e+00 -3.83555293e-01 2.09362805e-01 6.33965671e-01
-1.50576484e+00 -3.63830417e-01 1.20372725e+00 -1.15322776e-01
1.14061153e+00 4.02652115e-01 3.88277322e-01 9.54466403e-01
4.38387215e-01 8.35129499e-01 9.61260676e-01 -8.44301045e-01
4.65400130e-01 4.36452687e-01 -7.59016201e-02 6.90001786e-01
1.00210473e-01 2.35005617e-02 -7.05994487e-01 -6.87501356e-02
4.07925546e-01 -3.98941115e-02 -2.81853735e-01 -3.31837863e-01
-8.76539588e-01 8.53345990e-01 6.28431320e-01 7.05117941e-01
-5.92760265e-01 -2.41222322e-01 3.57411832e-01 1.20096348e-01
3.42107296e-01 1.03810871e+00 -4.24363732e-01 -1.44360095e-01
-7.44905353e-01 -2.04980463e-01 5.68769336e-01 2.42446572e-01
6.60466373e-01 3.48070264e-01 -6.41354546e-02 1.05865717e+00
1.01694189e-01 -7.41782933e-02 1.17700469e+00 -4.06579912e-01
4.27782953e-01 1.03273320e+00 -6.07623979e-02 -1.31035650e+00
-5.26512742e-01 -8.39867175e-01 -1.01901734e+00 2.34560579e-01
1.33946419e-01 -7.44168088e-02 -7.80202985e-01 1.23250806e+00
2.24361107e-01 3.61432806e-02 1.05077349e-01 9.91246045e-01
4.50784117e-01 5.30599356e-01 -5.79756014e-02 -2.56141514e-01
1.09120393e+00 -1.02451146e+00 -5.34649432e-01 -3.52618515e-01
4.58609313e-01 -6.44204497e-01 1.09369826e+00 4.70184773e-01
-8.55742872e-01 -5.26658773e-01 -1.20233989e+00 4.31971669e-01
-7.20448136e-01 2.46614143e-01 5.79028964e-01 8.37357938e-01
-7.79524982e-01 1.11068881e+00 -7.55965114e-01 -8.75492916e-02
3.92515302e-01 5.48610568e-01 1.18503891e-01 3.27971160e-01
-1.15782702e+00 1.14170110e+00 7.41834104e-01 5.00881732e-01
-3.31740201e-01 -3.36140454e-01 -4.17808205e-01 1.80066288e-01
2.86547393e-01 -3.76386315e-01 6.70026064e-01 -1.15627515e+00
-1.92147410e+00 7.43048370e-01 -8.72788429e-02 -6.86791480e-01
3.63056153e-01 -1.29922420e-01 -4.67338651e-01 -1.28084332e-01
-4.72103626e-01 4.24838156e-01 1.28032553e+00 -8.49989474e-01
-4.56707984e-01 -3.14505935e-01 -1.83746874e-01 3.10758412e-01
-9.40018356e-01 4.25850414e-02 -1.20136149e-01 -7.61637747e-01
2.57228315e-01 -7.04913855e-01 -4.17882472e-01 -3.98975044e-01
-6.03685141e-01 -3.14806968e-01 6.13751471e-01 -2.08974421e-01
1.60988593e+00 -1.92960548e+00 4.12580937e-01 2.85790175e-01
-1.02043167e-01 6.57962203e-01 -1.21284001e-01 4.23133403e-01
-8.30062702e-02 1.79072857e-01 -1.55402452e-01 -2.51394540e-01
-2.41576910e-01 1.65918827e-01 -3.03549320e-01 4.64903831e-01
2.71035075e-01 7.99522758e-01 -4.94662464e-01 -3.45568657e-01
4.31322128e-01 5.05407929e-01 -3.49429607e-01 -6.53647259e-02
-1.88986540e-01 2.61730462e-01 -7.71503985e-01 6.67348027e-01
1.84311464e-01 -2.60329515e-01 3.45960632e-02 -1.07537732e-01
-3.08399647e-01 1.27596200e-01 -1.02613103e+00 1.26378226e+00
-5.74963391e-01 6.46802545e-01 -2.79817313e-01 -1.34529495e+00
1.24877346e+00 5.71875423e-02 3.71283978e-01 -9.20720816e-01
3.24954987e-01 2.23744392e-01 1.86892170e-02 -5.15326798e-01
5.47592282e-01 2.29205549e-01 4.34942812e-01 4.28275734e-01
-1.98122218e-01 2.02748284e-01 2.18362540e-01 -4.48044777e-01
6.57856762e-01 -2.76001077e-02 2.07690477e-01 -2.40487322e-01
1.08615720e+00 -1.54946297e-01 1.99989825e-01 6.64653420e-01
-1.76090553e-01 4.88423795e-01 -2.34185662e-02 -7.36928701e-01
-9.60204363e-01 -4.78769809e-01 -3.26357961e-01 1.17998254e+00
1.68592766e-01 -1.76239997e-01 -7.43903041e-01 -3.80019516e-01
-2.69568086e-01 7.80534744e-01 -6.76376700e-01 -5.54172754e-01
-6.36996984e-01 -1.16382980e+00 5.25156856e-01 1.98323414e-01
6.73044026e-01 -1.27988636e+00 -1.08257508e+00 4.79736388e-01
1.79253429e-01 -6.14967227e-01 6.96596131e-02 3.98246199e-01
-1.27294850e+00 -7.53500342e-01 -8.50292444e-01 -5.62228799e-01
6.51726961e-01 -1.13507072e-02 8.08696687e-01 4.21840608e-01
-5.57563663e-01 1.33960675e-02 -3.57747257e-01 -4.42732990e-01
-3.64382327e-01 5.08688211e-01 4.94606718e-02 2.26882398e-01
4.12034601e-01 -5.63168526e-01 -4.96169716e-01 5.03450811e-01
-7.17700422e-01 -1.68336064e-01 8.50806296e-01 1.04376912e+00
4.36618716e-01 2.81446815e-01 6.19762421e-01 -4.66314286e-01
9.57947910e-01 -2.66515911e-01 -6.15145087e-01 5.17002583e-01
-1.04451072e+00 4.95632499e-01 1.00704014e+00 -6.87361777e-01
-8.48121405e-01 -1.95169136e-01 -1.44192830e-01 -4.07198280e-01
-4.96441498e-02 5.83234251e-01 2.09927544e-01 -2.79422879e-01
9.20357406e-01 5.34989655e-01 -3.45816463e-02 -5.87147653e-01
1.23701617e-01 5.82554996e-01 2.81008393e-01 -3.57806832e-01
5.16408205e-01 2.16823950e-01 -7.99993202e-02 -8.47490311e-01
-7.38598406e-01 -4.30582941e-01 -4.21415836e-01 -2.96666443e-01
4.83539641e-01 -2.01325417e-01 -7.29736209e-01 4.12101358e-01
-9.39961612e-01 1.16661444e-01 -6.52332753e-02 4.66238588e-01
-2.56275803e-01 2.01890439e-01 -2.18965501e-01 -1.10789490e+00
-6.18794560e-01 -1.19107461e+00 6.21037662e-01 6.31959736e-01
-2.24587977e-01 -1.22910202e+00 -1.06175587e-01 1.28438935e-01
6.83778644e-01 -9.15227830e-02 1.10533369e+00 -8.01079452e-01
-2.56994694e-01 -1.82370663e-01 6.42173067e-02 3.34411055e-01
4.86467518e-02 6.23935014e-02 -9.54037666e-01 -2.29264706e-01
1.14849284e-01 -1.64345846e-01 7.24259377e-01 4.66326028e-01
1.26645255e+00 -4.26857293e-01 -4.52407062e-01 5.53828835e-01
1.21658194e+00 4.20423061e-01 7.13590980e-01 8.96283448e-01
2.73012102e-01 6.04899764e-01 3.58741879e-01 5.88439643e-01
-3.38946790e-01 5.55940509e-01 6.63550079e-01 4.87698540e-02
3.15413505e-01 -3.57477553e-02 -4.10032757e-02 6.37337506e-01
-2.31940076e-01 -6.36107177e-02 -7.67405689e-01 3.31694573e-01
-1.82431448e+00 -8.65989506e-01 2.29895592e-01 2.03200817e+00
7.90098131e-01 4.47699100e-01 -1.15645677e-01 5.06031752e-01
8.86046469e-01 2.29538590e-01 -9.06138539e-01 -6.56077325e-01
-2.43007600e-01 1.72476903e-01 4.12771910e-01 1.88495159e-01
-1.02659714e+00 8.06163132e-01 6.54446316e+00 1.01586807e+00
-1.51497388e+00 -3.58318299e-01 7.32723832e-01 -1.68423593e-01
-1.72762960e-01 -2.96673447e-01 -8.02553952e-01 4.53220367e-01
8.08403611e-01 -4.57836352e-02 7.56166577e-01 1.01726890e+00
3.20465595e-01 6.99923486e-02 -6.38974488e-01 1.11370444e+00
5.22592701e-02 -1.56829417e+00 1.55188128e-01 -2.47912273e-01
7.81706274e-01 -1.04799554e-01 3.11810791e-01 1.57291442e-01
-2.05143318e-01 -1.17063868e+00 7.40534008e-01 4.94282722e-01
2.51272589e-01 -1.01163137e+00 6.21383607e-01 4.64940578e-01
-6.83724463e-01 -4.39205170e-01 -5.70424259e-01 -7.88238570e-02
-2.59779632e-01 7.12956667e-01 -7.79618323e-01 5.00184417e-01
7.61244357e-01 4.53011394e-01 -6.53467059e-01 1.13968146e+00
-1.58932190e-02 4.87323910e-01 -3.03031802e-01 -8.41151476e-01
4.23100978e-01 -4.59260762e-01 6.03001475e-01 8.86025190e-01
4.86935070e-03 -1.49417654e-01 -1.98088422e-01 8.59897733e-01
4.62418944e-02 3.82525474e-01 -2.12016836e-01 -2.08232120e-01
3.41209590e-01 9.77175176e-01 -8.91982019e-01 3.83731537e-02
5.76358326e-02 7.90670574e-01 3.82329494e-01 3.86676341e-01
-7.61128008e-01 -7.24405825e-01 4.62021381e-01 -2.50962257e-01
1.48021638e-01 -9.91320238e-02 -7.13539243e-01 -8.32621455e-01
-7.28675723e-03 -8.02846909e-01 3.09819400e-01 -4.88558501e-01
-8.32863510e-01 8.71129870e-01 -1.90655097e-01 -9.25856411e-01
-4.02431607e-01 -7.11207688e-01 -6.50016189e-01 8.00458550e-01
-1.41328251e+00 -6.59198999e-01 9.60255787e-02 4.34436470e-01
7.03066111e-01 -5.90143919e-01 7.09478974e-01 1.61895484e-01
-1.00480914e+00 6.96998537e-01 4.29340094e-01 -2.72952259e-01
2.33922571e-01 -7.21503139e-01 -3.15315723e-02 7.07692266e-01
3.03430021e-01 6.59397840e-01 9.10701811e-01 -2.16804728e-01
-1.51215053e+00 -5.91895938e-01 7.09673703e-01 1.24190375e-01
5.94931960e-01 3.09026451e-03 -1.07428932e+00 1.79161485e-02
-2.34879237e-02 -4.88021195e-01 3.09941202e-01 1.57193393e-01
-5.73811959e-03 -2.37039745e-01 -9.68029916e-01 9.57657754e-01
7.27645338e-01 -3.29601496e-01 -4.81808603e-01 2.17302740e-01
2.37172589e-01 -2.68669397e-01 -4.97773468e-01 4.15446877e-01
6.19095862e-01 -1.12236643e+00 1.05097282e+00 -6.73597634e-01
3.31886858e-01 8.12222995e-03 1.36788860e-01 -1.28709900e+00
-2.54945248e-01 -6.76413119e-01 -2.29154408e-01 1.06175435e+00
5.88924527e-01 -7.60793805e-01 1.09027123e+00 5.07777452e-01
-3.90673503e-02 -1.39149499e+00 -1.07542956e+00 -6.46997452e-01
-1.59441292e-01 -3.05626899e-01 5.57228208e-01 7.78351605e-01
-1.37722656e-01 1.93063810e-01 -2.99507976e-01 -7.51771927e-02
3.52830261e-01 1.19228318e-01 3.15256745e-01 -1.00157797e+00
-2.14785382e-01 -1.17536318e+00 -3.57012689e-01 -8.53844941e-01
8.30876380e-02 -5.76049685e-01 -9.94279906e-02 -1.11128163e+00
-3.09209496e-01 -5.00448167e-01 -4.47723448e-01 4.09899592e-01
-1.82058200e-01 1.93402350e-01 -4.93638963e-02 4.42604244e-01
-4.26418394e-01 7.18588293e-01 9.59623635e-01 -2.55883098e-01
-4.98709649e-01 3.49309564e-01 -5.99509418e-01 8.23339880e-01
9.66774940e-01 -5.18278003e-01 -2.93161392e-01 -3.45689088e-01
3.61214221e-01 -3.99850279e-01 7.70328566e-02 -1.08198094e+00
3.13207895e-01 -2.42680475e-01 3.97291690e-01 -3.70432466e-01
2.33791292e-01 -6.53384626e-01 -3.59704494e-02 7.12038159e-01
-5.93091190e-01 9.68862791e-03 2.42941156e-01 4.45745975e-01
-2.77160972e-01 -9.01504755e-01 9.27340925e-01 -4.51529622e-02
-7.63160706e-01 -2.04646870e-01 -4.97282475e-01 -2.50129580e-01
9.45609510e-01 -5.41284561e-01 1.19455889e-01 -7.07537979e-02
-5.09023964e-01 5.48193306e-02 6.15692250e-02 5.68814099e-01
7.18179047e-01 -8.40463281e-01 -4.27173465e-01 2.05922693e-01
-1.63410947e-01 -3.97614777e-01 1.86028898e-01 6.34089351e-01
-4.24211770e-01 6.21019900e-01 -1.23629317e-01 -6.88512802e-01
-1.10403931e+00 6.34893656e-01 5.40717900e-01 -4.54383522e-01
-4.39805210e-01 8.17341208e-01 -4.50155616e-01 -1.86176464e-01
5.79381824e-01 -1.54709771e-01 -4.60716248e-01 -2.88630817e-02
5.07780969e-01 6.53681695e-01 3.79690558e-01 -2.20164001e-01
-4.71214116e-01 5.63907266e-01 -1.87480673e-01 2.63035685e-01
1.23536146e+00 -1.37081534e-01 -3.82394157e-02 2.29676262e-01
1.15679884e+00 -3.38532954e-01 -9.77336049e-01 -2.22705781e-01
5.09066224e-01 -2.14355171e-01 2.15846390e-01 -8.51726353e-01
-9.88141358e-01 8.23429227e-01 6.61089182e-01 4.97874767e-01
1.34226489e+00 -5.36117792e-01 4.74954665e-01 9.14874494e-01
2.37033650e-01 -1.39890349e+00 2.86889076e-01 4.57752287e-01
8.03813338e-01 -1.07707214e+00 5.83444349e-02 2.08250821e-01
-4.73741412e-01 1.42092645e+00 4.33490068e-01 -1.48153394e-01
4.22290295e-01 3.98975275e-02 3.12749334e-02 3.11725233e-02
-6.91261292e-01 1.02947824e-01 4.51400071e-01 2.64635026e-01
3.65340233e-01 -3.28844130e-01 -4.46097881e-01 2.84299880e-01
-1.66602209e-01 -2.11029708e-01 1.17194392e-01 1.02240705e+00
-5.76640010e-01 -9.41706955e-01 -5.59747159e-01 5.66519141e-01
-6.85108840e-01 -9.74823162e-02 -3.42634887e-01 5.30206144e-01
-1.78227246e-01 8.76415014e-01 -1.39560610e-01 -2.99935669e-01
2.52498358e-01 1.41013592e-01 2.39816338e-01 -1.64807484e-01
-9.07816947e-01 -2.18960851e-01 -1.86103597e-01 -2.53642797e-01
-4.68042105e-01 -2.83294201e-01 -1.17994237e+00 -2.15384886e-01
-7.11529076e-01 4.01879191e-01 1.10750175e+00 1.15967119e+00
3.71812761e-01 5.57781100e-01 7.78922975e-01 -1.06751144e+00
-8.35258901e-01 -8.60919297e-01 -1.33572608e-01 8.33506659e-02
1.20074615e-01 -6.86525464e-01 -3.43290895e-01 -2.54409581e-01] | [8.275918006896973, 3.291749954223633] |
b711cafb-936c-4cc9-9fb0-650eca32259f | linear-cross-lingual-mapping-of-sentence | 2305.14256 | null | https://arxiv.org/abs/2305.14256v1 | https://arxiv.org/pdf/2305.14256v1.pdf | Linear Cross-Lingual Mapping of Sentence Embeddings | Semantics of a sentence is defined with much less ambiguity than semantics of a single word, and it should be better preserved by translation to another language. If multilingual sentence embeddings intend to represent sentence semantics, then the similarity between embeddings of any two sentences must be invariant with respect to translation. Based on this suggestion, we consider a simple linear cross-lingual mapping as a possible improvement of the multilingual embeddings. We also consider deviation from orthogonality conditions as a measure of deficiency of the embeddings. | ['John Bohannon', 'Fumika Isono', 'Oleg Vasilyev'] | 2023-05-23 | null | null | null | null | ['sentence-embeddings', 'sentence-embeddings'] | ['methodology', 'natural-language-processing'] | [-2.05092758e-01 1.70120150e-01 -2.44411066e-01 -6.82791233e-01
-2.76001781e-01 -7.95623124e-01 9.32111919e-01 4.97707039e-01
-6.68230236e-01 8.68689001e-01 8.29166889e-01 -3.63456637e-01
4.18899842e-02 -7.07861125e-01 -5.48035502e-01 -3.56776506e-01
2.59190351e-01 2.87950248e-01 1.82451308e-01 -5.95210195e-01
6.32312223e-02 4.22102064e-01 -9.96870577e-01 1.30203769e-01
6.18111134e-01 2.57868886e-01 2.66219378e-01 3.44737828e-01
-4.92429197e-01 4.16680783e-01 -4.40623522e-01 -5.48594654e-01
1.95034012e-01 -4.46285337e-01 -1.02766025e+00 -1.57138944e-01
4.51951206e-01 2.86841691e-01 -4.09195840e-01 1.42574000e+00
3.12365621e-01 -2.44982541e-02 5.60963273e-01 -8.46749008e-01
-1.27318931e+00 6.24473929e-01 9.10264347e-03 3.14758956e-01
6.78679585e-01 -4.73233223e-01 1.28536975e+00 -9.02086914e-01
9.56193984e-01 1.18907750e+00 5.85605383e-01 3.22154731e-01
-1.29754865e+00 1.18735671e-01 7.72462487e-02 5.65928370e-02
-1.36959481e+00 -2.05366075e-01 6.78792834e-01 -2.72372305e-01
9.01247084e-01 2.82535553e-01 6.22696042e-01 1.00205982e+00
6.99322522e-01 8.57127234e-02 1.21363819e+00 -6.42004490e-01
-1.11476876e-01 7.67960787e-01 2.72779047e-01 4.79298860e-01
5.45227587e-01 -3.36756796e-01 -3.75787258e-01 -5.94803542e-02
2.10812524e-01 -2.31081903e-01 -2.37552375e-01 -6.76950037e-01
-1.38665414e+00 1.11384642e+00 2.82394111e-01 9.06423450e-01
-1.58949450e-01 -7.17529748e-03 5.67470312e-01 6.56170547e-01
4.67935473e-01 6.76734447e-01 -6.58547521e-01 -4.82869484e-02
-5.33137083e-01 7.75138810e-02 6.71407700e-01 7.42483616e-01
8.81373465e-01 -6.54069632e-02 2.27055401e-01 6.69800639e-01
3.12670261e-01 6.27835989e-01 8.64550173e-01 -3.01011860e-01
7.43441880e-02 5.59885204e-01 -1.28798708e-01 -1.17775750e+00
-4.28250343e-01 -1.93609372e-01 -2.62085646e-01 -4.13666070e-01
1.63043931e-01 8.58170167e-02 -2.82059312e-01 2.01373100e+00
2.20940962e-01 -1.74398303e-01 4.52324510e-01 8.65776896e-01
5.42864084e-01 6.58128500e-01 -1.37396902e-02 -3.48221481e-01
1.61881018e+00 -5.60377359e-01 -9.63141441e-01 -2.00167328e-01
9.73722875e-01 -8.72253418e-01 1.32754529e+00 -2.03135461e-01
-7.73895860e-01 -3.55247885e-01 -1.24882221e+00 -3.24436039e-01
-9.00698960e-01 -2.69706279e-01 5.74750006e-01 8.43596637e-01
-1.10853624e+00 3.94257605e-01 -5.24796665e-01 -8.43267143e-01
-6.30025506e-01 1.85389861e-01 -8.77051771e-01 2.67855585e-01
-1.66872537e+00 1.61234975e+00 5.59236526e-01 -2.03977004e-02
-3.00665889e-02 -4.23445225e-01 -1.17578125e+00 -2.37023011e-01
-3.34570050e-01 -5.80623090e-01 7.95631647e-01 -9.24411297e-01
-1.04067433e+00 1.09699380e+00 -2.05959022e-01 -4.00938243e-01
8.52323547e-02 9.89676267e-03 -9.63491142e-01 5.71474880e-02
4.79627967e-01 3.89456540e-01 6.42076075e-01 -1.02450633e+00
-3.43459755e-01 -5.80735028e-01 1.93334863e-01 2.75260091e-01
-6.86232924e-01 2.40799785e-01 7.95015767e-02 -4.82887030e-01
3.00442189e-01 -9.69747126e-01 1.00538477e-01 -3.16747129e-01
-2.32914034e-02 -1.96465030e-01 6.82364464e-01 -7.23637998e-01
1.28554273e+00 -2.06054783e+00 2.00369284e-01 2.65977215e-02
-3.11677545e-01 -8.32073689e-02 -3.51413697e-01 8.78217340e-01
-4.66084093e-01 3.05950820e-01 -1.75337359e-01 -2.01034546e-01
2.86497712e-01 4.94995832e-01 -4.26792234e-01 9.06991482e-01
2.03469828e-01 7.79945314e-01 -9.29202020e-01 -5.21625996e-01
7.51254009e-03 5.00439107e-01 -2.93547422e-01 -3.44452858e-01
1.24198288e-01 -3.60624678e-02 -4.23212081e-01 1.57873794e-01
4.92974371e-01 3.15271854e-01 5.52132785e-01 -1.54889897e-01
-1.35801092e-01 6.65445924e-01 -8.52111459e-01 1.89396489e+00
-5.44538796e-01 6.85964465e-01 -4.71579850e-01 -1.04077637e+00
1.16982567e+00 5.13524413e-01 3.92362475e-01 -8.03447127e-01
-9.48790908e-02 3.91549408e-01 1.60073295e-01 -4.05519396e-01
9.47751641e-01 -7.51814365e-01 -5.05235732e-01 5.83600938e-01
1.71828985e-01 -1.68051571e-01 1.57215253e-01 1.28700472e-02
6.89686596e-01 1.31110456e-02 4.91333246e-01 -8.98195326e-01
6.37176514e-01 -6.93899840e-02 4.92761731e-01 2.82023549e-01
-2.21406505e-01 4.21434164e-01 3.94772291e-01 -4.15727079e-01
-1.18183124e+00 -1.25474548e+00 -7.22687662e-01 7.79083490e-01
1.02841705e-01 -5.36773205e-01 -6.74892247e-01 -8.67111027e-01
-2.05027293e-02 9.05330300e-01 -6.56044483e-01 -1.83439687e-01
-5.44869423e-01 -6.70132339e-01 6.06021643e-01 2.99238771e-01
-9.91052985e-02 -7.57250488e-01 -4.00325805e-01 4.63844053e-02
3.31000090e-02 -1.09232318e+00 -5.81172168e-01 2.21849501e-01
-6.89278483e-01 -8.65972936e-01 -4.88281280e-01 -1.22153378e+00
4.84211832e-01 1.99876547e-01 9.98955905e-01 -1.66250780e-01
3.27742100e-01 3.74422818e-01 -5.24349093e-01 -1.52018415e-02
-6.73963487e-01 2.19457652e-02 6.97064400e-01 -1.27162278e-01
7.78000236e-01 -4.88147616e-01 -7.57241100e-02 -1.67549178e-01
-1.12286496e+00 -6.08910263e-01 7.44895265e-02 8.24744225e-01
4.49319243e-01 -3.55221704e-02 6.15280211e-01 -7.74780810e-01
1.02287316e+00 -3.37405503e-01 -1.57817692e-01 4.57581222e-01
-6.97095871e-01 4.41813529e-01 7.75999486e-01 -1.26281545e-01
-7.72690713e-01 -2.35751748e-01 -7.54595967e-04 -7.66810961e-03
2.64372285e-02 7.17479885e-01 -1.46876097e-01 6.93623647e-02
4.95837301e-01 3.48482221e-01 1.19597465e-01 -3.32171947e-01
6.79476142e-01 6.23982370e-01 1.20104052e-01 -5.06549001e-01
6.85461998e-01 2.79176921e-01 -2.13575318e-01 -1.17714417e+00
-5.97019970e-01 -3.63401473e-01 -8.79834175e-01 2.32778534e-01
1.04999721e+00 -7.46236086e-01 -1.23710781e-01 -2.37128064e-01
-1.45393085e+00 5.85372448e-01 -4.87567872e-01 8.62736642e-01
-3.72240156e-01 6.04658544e-01 -3.16520751e-01 -2.48114139e-01
2.90249269e-02 -1.02476799e+00 6.79583192e-01 -1.90739363e-01
-5.51140130e-01 -1.74094999e+00 4.52596009e-01 -8.80810022e-02
1.49330258e-01 -1.20586582e-01 1.07846582e+00 -1.14675856e+00
2.84405828e-01 -4.33660001e-01 1.41532170e-02 6.00678384e-01
5.25303900e-01 -3.06356302e-03 -7.01704144e-01 -4.14543480e-01
5.48839808e-01 6.49486110e-02 6.24007046e-01 1.17234901e-01
-7.11160246e-03 -2.31219381e-01 3.07175126e-02 3.17284286e-01
1.58086610e+00 6.08839579e-02 4.64693367e-01 6.22769296e-01
5.92009723e-01 7.76851177e-01 3.76054317e-01 3.67644504e-02
4.08367842e-01 6.21893406e-01 -1.11281864e-01 6.25441521e-02
2.71309111e-02 -4.00399387e-01 7.36525655e-01 1.69013572e+00
5.20660222e-01 1.69821575e-01 -7.91621089e-01 6.84987426e-01
-1.50746775e+00 -8.50962222e-01 -3.61673176e-01 2.08839369e+00
6.20803893e-01 1.62168015e-02 -7.94116706e-02 -1.17156699e-01
7.21527874e-01 5.11593580e-01 1.24868013e-01 -1.22763324e+00
-5.04902482e-01 5.10235578e-02 5.16488731e-01 8.34498644e-01
-9.12175536e-01 1.00746644e+00 7.93116903e+00 4.30933297e-01
-1.26914692e+00 4.11291897e-01 -9.24111754e-02 3.66478860e-01
-1.00147152e+00 4.24248338e-01 -5.06524205e-01 4.61624950e-01
9.62996304e-01 -6.07066214e-01 5.80803975e-02 5.03202558e-01
2.05633923e-01 1.63507521e-01 -1.07271099e+00 4.82080251e-01
4.28362370e-01 -1.13442838e+00 3.43287855e-01 3.03734504e-02
5.89495242e-01 -7.98532441e-02 -4.49870080e-02 1.22083016e-01
-2.05406815e-01 -8.94116998e-01 7.63627529e-01 5.05654573e-01
4.03046191e-01 -8.83971870e-01 1.00664413e+00 7.95601159e-02
-1.11421072e+00 3.35525125e-01 -7.64074206e-01 -1.58699483e-01
2.49948621e-01 2.65074253e-01 -5.94214499e-01 6.79981172e-01
1.23364285e-01 7.04210281e-01 -7.85808086e-01 2.54906923e-01
-3.00439864e-01 1.58955544e-01 -5.70041724e-02 -2.74293751e-01
3.27626556e-01 -7.19026744e-01 6.56304181e-01 1.36721253e+00
4.62968320e-01 -5.92018723e-01 -3.62839364e-02 7.68379807e-01
2.14814663e-01 5.92826545e-01 -1.18509626e+00 -4.09486145e-01
4.47752953e-01 7.98449874e-01 -3.87475938e-01 -2.85894424e-01
-8.18064630e-01 1.25525141e+00 1.77851722e-01 2.00907901e-01
-7.79656351e-01 -4.04783726e-01 1.07080936e+00 -6.00270834e-03
8.31095800e-02 -3.04841042e-01 -3.68698418e-01 -1.37201810e+00
3.68081868e-01 -5.85169852e-01 5.66317886e-02 -6.32938445e-01
-1.23085642e+00 7.68002212e-01 -1.29190564e-01 -1.17147946e+00
-1.98994249e-01 -7.25865126e-01 -3.33097249e-01 1.02087855e+00
-1.28432059e+00 -1.18574405e+00 8.14180732e-01 5.35380960e-01
2.02773824e-01 -2.85647303e-01 1.38009632e+00 1.62230819e-01
-6.37794733e-02 5.25480807e-01 8.61489922e-02 -8.52517784e-03
7.40027249e-01 -1.39355981e+00 2.76472121e-01 9.94244993e-01
6.84686542e-01 1.13318276e+00 1.12297654e+00 -5.17304838e-01
-1.48274612e+00 -8.52823555e-01 1.91913629e+00 -5.74800253e-01
1.18605149e+00 -3.25989068e-01 -9.32062268e-01 8.41663837e-01
7.41999567e-01 -2.16322601e-01 8.06579411e-01 5.48781455e-01
-5.29500842e-01 -1.11195788e-01 -8.87108684e-01 6.40465081e-01
6.38359547e-01 -1.27084577e+00 -1.34645295e+00 4.08586115e-01
1.20403695e+00 2.62158424e-01 -1.28138113e+00 1.26439989e-01
4.34100449e-01 -7.71114826e-01 8.13893020e-01 -7.85098016e-01
9.69095156e-02 -3.83116424e-01 -6.66480362e-01 -1.50621223e+00
-3.28627855e-01 -1.41531259e-01 7.18139172e-01 1.09015071e+00
5.84551752e-01 -8.86694431e-01 2.95177311e-01 2.57761747e-01
-2.08875105e-01 -2.15591773e-01 -1.17494309e+00 -1.24132848e+00
7.62752116e-01 -4.34796333e-01 7.22491503e-01 1.34256673e+00
7.08237171e-01 7.10196912e-01 -2.14385837e-01 6.06486239e-02
1.61142558e-01 -5.69675267e-02 2.74982601e-01 -1.00023770e+00
-7.60923401e-02 -3.66705537e-01 -9.27382052e-01 -7.49711275e-01
7.78244317e-01 -1.38147068e+00 -3.06080520e-01 -1.47756505e+00
4.55295369e-02 1.11338779e-01 -3.93509954e-01 2.85497475e-02
1.56722412e-01 2.15773821e-01 1.73912570e-01 -1.26518775e-02
-1.76135883e-01 7.98831761e-01 1.18998420e+00 -1.28406724e-02
1.54119745e-01 -4.96330082e-01 -7.36909330e-01 7.26432860e-01
7.23870456e-01 -5.82848430e-01 -4.11882997e-01 -5.57080448e-01
3.58837396e-01 -7.57083669e-02 -6.90309703e-02 -5.55095017e-01
-8.74565318e-02 -1.82319909e-01 -8.46697241e-02 -4.47914861e-02
3.14068258e-01 -8.84194076e-01 1.07064508e-01 4.03669447e-01
-4.68571752e-01 8.09354722e-01 -3.15584391e-02 3.48264784e-01
-6.51610970e-01 -5.78192472e-01 5.46779573e-01 3.50019708e-03
-7.38324523e-01 -1.95919096e-01 -5.11303127e-01 -5.57377599e-02
8.34409356e-01 -4.82363641e-01 -2.44108066e-01 -7.25582801e-03
-7.12990046e-01 -1.73729554e-01 9.26412880e-01 8.10642898e-01
4.79863614e-01 -1.75901747e+00 -7.18050361e-01 2.37734884e-01
2.94343859e-01 -1.00007570e+00 -1.51865840e-01 9.36764240e-01
-5.97835302e-01 8.26644599e-01 -1.47512376e-01 -2.15940028e-01
-1.09535980e+00 7.00504959e-01 2.05867812e-01 1.90284491e-01
-4.87466782e-01 3.04200619e-01 -5.79676852e-02 -9.56935287e-01
-3.83410364e-01 -3.55856627e-01 -1.24171637e-01 2.26216421e-01
2.95162648e-01 -4.29348089e-02 1.09444313e-01 -1.35561800e+00
-7.98773289e-01 6.30925417e-01 1.69289574e-01 -3.25641453e-01
1.00432515e+00 -3.98302972e-01 -5.76097608e-01 1.02866292e+00
1.66289866e+00 4.68911648e-01 -2.78030843e-01 -1.13688320e-01
4.20524925e-01 -4.71400052e-01 -2.15594307e-01 -2.22973868e-01
-5.36024868e-01 7.64361441e-01 3.49250764e-01 4.97231960e-01
6.62755311e-01 -2.91155558e-02 4.96835589e-01 3.62030417e-01
2.60609925e-01 -1.25971866e+00 -3.75453055e-01 7.53540695e-01
9.71855521e-01 -1.05505288e+00 -8.17429349e-02 -6.35690913e-02
-8.09399366e-01 1.26750505e+00 5.40321469e-02 -3.75931948e-01
8.04504454e-01 -8.74355808e-02 2.56700754e-01 -1.23950243e-01
-6.05167627e-01 -1.60123557e-01 3.01151067e-01 4.82409686e-01
8.96612465e-01 3.95174652e-01 -1.16419291e+00 1.58830658e-01
-5.58081448e-01 -4.96283829e-01 6.95532084e-01 6.79888725e-01
-3.85010868e-01 -1.38555360e+00 -2.23505184e-01 -3.54077071e-02
-2.91370928e-01 -3.63568813e-01 -4.30443257e-01 8.29055190e-01
2.09867463e-01 8.63006771e-01 1.90236390e-01 -3.28454792e-01
3.59884232e-01 3.65684211e-01 4.18281466e-01 -8.24252725e-01
-4.42401499e-01 -5.67672886e-02 1.91172644e-01 -2.62056708e-01
-5.66350877e-01 -7.95830131e-01 -1.25378776e+00 -3.40323716e-01
-1.81152835e-01 5.26905537e-01 8.74519646e-01 1.00462079e+00
1.49972066e-01 2.28271142e-01 4.76540655e-01 6.01751078e-03
-7.75218844e-01 -9.55999970e-01 -1.05713415e+00 6.37369871e-01
1.18347660e-01 -4.26818699e-01 -4.46040273e-01 -1.15921609e-01] | [11.006617546081543, 9.90994930267334] |
da7b47d7-a4ac-4792-bd0c-82514a2dc0c7 | clulex-at-semeval-2021-task-1-a-simple-system | null | null | https://aclanthology.org/2021.semeval-1.81 | https://aclanthology.org/2021.semeval-1.81.pdf | CLULEX at SemEval-2021 Task 1: A Simple System Goes a Long Way | This paper presents the system we submitted to the first Lexical Complexity Prediction (LCP) Shared Task 2021. The Shared Task provides participants with a new English dataset that includes context of the target word. We participate in the single-word complexity prediction sub-task and focus on feature engineering. Our best system is trained on linguistic features and word embeddings (Pearson{'}s score of 0.7942). We demonstrate, however, that a simpler feature set achieves comparable results and submit a model trained on 36 linguistic features (Pearson{'}s score of 0.7925). | ['Elin Askl{\\"o}v', "H{\\'e}ctor Hern{\\'a}ndez", 'Mironas Bitinis', 'Sinan Tang', 'Peter Kolb', 'Greta Smolenska'] | 2021-08-01 | null | null | null | semeval-2021 | ['lexical-complexity-prediction'] | ['natural-language-processing'] | [-3.42899412e-01 1.25531539e-01 -1.98997214e-01 -3.85139227e-01
-9.58420157e-01 -4.59013492e-01 4.73416299e-01 6.26704454e-01
-1.14233470e+00 5.48737526e-01 5.15253127e-01 -4.15293217e-01
2.35288593e-04 -4.22315925e-01 -2.03677252e-01 1.75803602e-01
-2.81370997e-01 4.20569420e-01 1.95222050e-01 -4.64109808e-01
6.17808580e-01 2.00285196e-01 -1.69941139e+00 5.65572798e-01
6.65268600e-01 8.92488122e-01 5.86816609e-01 9.43533838e-01
-2.46967107e-01 2.88425833e-01 -5.15495002e-01 -4.30798441e-01
1.16979487e-01 8.09383243e-02 -1.04770863e+00 -8.14056396e-01
2.41355330e-01 1.86376452e-01 4.22212742e-02 7.02195764e-01
5.20947337e-01 3.54808092e-01 7.99912989e-01 -1.03838825e+00
-8.27895761e-01 8.29898536e-01 1.55797735e-01 6.15606070e-01
8.87916267e-01 7.22555816e-02 1.74770892e+00 -1.49558306e+00
8.34535837e-01 1.08427393e+00 1.01464355e+00 4.70097601e-01
-1.13171685e+00 -5.38940549e-01 2.49496520e-01 5.89234889e-01
-1.64475179e+00 -4.54932541e-01 4.37825531e-01 -4.08605605e-01
2.23473167e+00 2.12092832e-01 6.47424817e-01 9.79144037e-01
3.48514169e-01 6.48772299e-01 1.00895107e+00 -8.56896400e-01
-1.39963984e-01 3.25024515e-01 4.79532212e-01 7.79365242e-01
7.36065358e-02 5.45412526e-02 -7.63234973e-01 -5.28520113e-03
-8.80426466e-02 -6.72012568e-01 3.19825374e-02 3.15926105e-01
-1.14206672e+00 9.36415792e-01 -9.64655168e-03 6.37841702e-01
-4.47091311e-02 -1.42234161e-01 4.64294493e-01 7.55872846e-01
4.82571125e-01 8.70379031e-01 -1.09357548e+00 -5.22497833e-01
-6.16243005e-01 3.75207245e-01 1.18744349e+00 1.01575303e+00
6.51068330e-01 -3.78706634e-01 -2.57685169e-04 1.03380024e+00
3.85191441e-01 1.02302574e-01 9.85084176e-01 -5.17310977e-01
4.79078203e-01 4.90151942e-01 4.06511314e-02 -8.03239703e-01
-9.35530484e-01 8.95674452e-02 -2.20041052e-02 -2.39389330e-01
2.66786695e-01 -1.35808527e-01 -3.17027181e-01 1.68981469e+00
-2.47891009e-01 -4.10932988e-01 3.84496897e-02 4.90250021e-01
1.03003359e+00 5.82292736e-01 4.41787601e-01 -3.04968327e-01
1.35861838e+00 -8.60870421e-01 -7.27495193e-01 -1.11581594e-01
1.32431614e+00 -1.27702200e+00 1.55152631e+00 4.90272015e-01
-1.10602903e+00 -7.66622066e-01 -1.12670958e+00 -3.97978246e-01
-9.46280241e-01 -1.49494231e-01 4.99708891e-01 8.43481243e-01
-1.21310925e+00 7.92322397e-01 -5.10287881e-01 -5.72749794e-01
-1.53537408e-01 3.78564894e-01 -5.73011398e-01 1.38162225e-01
-1.52119219e+00 1.61335444e+00 3.42602491e-01 -5.57545125e-01
-1.74755558e-01 -8.33001912e-01 -8.48749340e-01 -5.19801788e-02
-4.50351574e-02 -3.66219491e-01 1.02758610e+00 -6.29694760e-01
-1.39493132e+00 9.94174898e-01 -2.34751716e-01 -1.45823881e-01
6.25814125e-02 -3.91646862e-01 -8.16004217e-01 -3.85158718e-01
1.85874656e-01 6.49643660e-01 1.45959467e-01 -6.86890125e-01
-7.48063385e-01 5.41281588e-02 -6.05125949e-02 2.79883027e-01
-7.35779941e-01 5.57943285e-01 -9.55682024e-02 -7.92168856e-01
-4.25519317e-01 -6.48060441e-01 -1.50560439e-01 -3.11978817e-01
-1.39472634e-01 -8.69566023e-01 9.23632979e-02 -7.83813000e-01
1.90396678e+00 -1.88709760e+00 1.77735224e-01 1.16823465e-01
7.41837025e-02 1.87541932e-01 -5.66717088e-01 8.01493049e-01
-9.41793025e-02 6.49429560e-01 2.31134802e-01 -6.35458767e-01
4.27467287e-01 4.73760953e-03 1.96539402e-01 3.27560753e-02
3.93360227e-01 9.06784415e-01 -7.90007234e-01 -5.61069310e-01
6.63776770e-02 2.21871346e-01 -9.91733372e-01 7.10779727e-02
-3.00538559e-02 -3.25458556e-01 -1.21326283e-01 3.13762814e-01
2.55443156e-01 2.27324262e-01 2.92847991e-01 -2.29763150e-01
-4.78948981e-01 6.02957428e-01 -1.01665843e+00 1.73518288e+00
-8.76926243e-01 6.64252162e-01 -6.18693471e-01 -5.19234478e-01
9.07880664e-01 1.42943040e-01 2.86471128e-01 -7.12273479e-01
2.12917417e-01 3.78938079e-01 3.27636957e-01 -4.94282991e-01
8.32676232e-01 -1.01639731e-02 -5.54657876e-01 5.55301368e-01
3.96590680e-01 -2.19398752e-01 3.97434920e-01 -3.37719396e-02
1.28729761e+00 4.03612014e-03 7.26459563e-01 -9.35590625e-01
6.75815821e-01 -2.19030812e-01 3.33744764e-01 3.93747181e-01
-5.77432156e-01 3.06208998e-01 1.92299277e-01 -7.60105073e-01
-1.24002111e+00 -8.79409313e-01 -3.13054413e-01 1.64679086e+00
-3.56784612e-01 -1.38526249e+00 -4.26509619e-01 -3.55353534e-01
-8.07320923e-02 8.03500593e-01 -6.94041312e-01 -1.30544037e-01
-5.84104955e-01 -3.45881641e-01 7.35161185e-01 6.47952378e-01
-1.47181913e-01 -1.24371874e+00 -5.36319196e-01 4.63985771e-01
-2.77040333e-01 -1.09516072e+00 -7.47627676e-01 5.02523720e-01
-3.51832479e-01 -8.81127119e-01 -7.53026605e-02 -1.23426616e+00
-8.38515237e-02 -3.38088334e-01 1.34462988e+00 2.56411970e-01
-5.87718129e-01 3.00265104e-01 -7.88844943e-01 -4.54525650e-01
-1.44474998e-01 4.15154010e-01 4.60498422e-01 -7.74066150e-01
1.04213226e+00 -2.83892363e-01 -1.26484424e-01 2.37110034e-02
-3.89706880e-01 -1.22034438e-01 2.79573232e-01 8.53951812e-01
4.32587922e-01 -3.06952745e-01 5.90180874e-01 -6.62507951e-01
1.11472058e+00 -3.60929042e-01 -2.54906565e-01 3.80670339e-01
-7.47321725e-01 9.25257131e-02 6.18904054e-01 -4.22736228e-01
-6.02327406e-01 1.49249405e-01 -5.92185497e-01 3.41723263e-01
-1.66321441e-03 6.85779452e-01 -1.55234635e-01 1.34431526e-01
7.22998738e-01 -6.97124749e-02 -3.06837469e-01 -6.91260219e-01
5.56682944e-01 5.71541548e-01 8.61353129e-02 -5.79347253e-01
4.74048913e-01 -5.45675218e-01 -3.73143733e-01 -1.02395105e+00
-7.44025946e-01 -4.25278962e-01 -9.45897341e-01 2.70527810e-01
7.81230092e-01 -7.93490350e-01 -5.95023632e-01 6.44200593e-02
-1.26291573e+00 -2.08122641e-01 -2.13768512e-01 6.67573214e-01
-7.03890204e-01 4.33186620e-01 -5.45348644e-01 -5.61814487e-01
-3.25525701e-01 -7.97789633e-01 7.53486574e-01 -3.98886234e-01
-1.06469929e+00 -1.24214661e+00 3.97733718e-01 1.48351525e-03
5.90618551e-01 -1.57157376e-01 1.25255668e+00 -1.13478816e+00
1.73259318e-01 -1.34075075e-01 -5.18474579e-02 4.20039415e-01
1.47214308e-01 1.31754845e-01 -8.32557738e-01 3.59939374e-02
-3.25846702e-01 -4.46318328e-01 7.35685170e-01 1.61700472e-01
1.05272973e+00 -2.03151815e-02 -2.12629493e-02 4.17020291e-01
1.25059140e+00 -2.05048531e-01 2.11452171e-01 4.87163901e-01
4.16650891e-01 7.56956339e-01 5.69703519e-01 4.74211574e-01
8.10271144e-01 7.69959986e-01 -1.83964029e-01 5.98235011e-01
-2.93799371e-01 -2.08240971e-01 6.78475261e-01 1.46302450e+00
5.60204759e-02 -5.70160784e-02 -1.22328174e+00 5.68331480e-01
-1.53953993e+00 -6.22480392e-01 -1.96050882e-01 1.71361864e+00
1.18191779e+00 3.57040584e-01 1.97195753e-01 3.26819122e-01
4.14181888e-01 -6.10992610e-02 -6.59915432e-03 -1.23284757e+00
-2.29843453e-01 7.74541318e-01 1.10518940e-01 1.18927944e+00
-1.00744748e+00 1.35542178e+00 7.24618578e+00 9.32956398e-01
-3.58335823e-01 1.70871958e-01 -4.89141680e-02 -2.17536371e-02
-3.75658661e-01 -1.42137796e-01 -1.09535909e+00 4.46471423e-01
1.58109021e+00 -8.86540592e-01 3.82154703e-01 7.76347876e-01
-1.35319546e-01 -7.75778666e-02 -1.32760346e+00 8.18904519e-01
2.80023485e-01 -9.33208346e-01 1.08191945e-01 -9.13154483e-02
4.10514891e-01 2.22249597e-01 5.00146113e-03 8.09831619e-01
1.01415284e-01 -1.25414312e+00 5.77145278e-01 6.51712596e-01
9.30876195e-01 -9.71672893e-01 7.77709544e-01 2.71231532e-01
-1.53919721e+00 1.22361183e-02 -4.15778697e-01 -5.86636007e-01
1.17127553e-01 6.27689064e-01 -6.13172412e-01 2.36219734e-01
5.69272518e-01 5.84639370e-01 -8.63276064e-01 5.94367027e-01
-1.84840217e-01 3.40578943e-01 -4.11425143e-01 -6.72546506e-01
-4.67731133e-02 3.75856876e-01 4.13770318e-01 1.77750230e+00
5.95413335e-02 6.08552597e-04 3.79019588e-01 2.78346032e-01
-3.45492214e-02 8.41888249e-01 -6.17151260e-01 -6.83079511e-02
6.77043140e-01 1.14201212e+00 -3.72318923e-01 -2.94262528e-01
-5.65105498e-01 7.51283109e-01 7.43998110e-01 -3.62196475e-01
-5.58004200e-01 -9.26292896e-01 7.35798597e-01 -2.20659688e-01
2.41139457e-01 -4.00899529e-01 -5.07354319e-01 -1.15490699e+00
-1.40613760e-04 -5.53937554e-01 1.84560061e-01 -2.72791862e-01
-1.77405846e+00 8.91772032e-01 -5.12537844e-02 -7.73634732e-01
-1.90610126e-01 -1.22107136e+00 -4.04654890e-01 1.29783535e+00
-1.53592348e+00 -1.07008135e+00 1.86697971e-02 4.60009187e-01
6.59694195e-01 -4.69474673e-01 1.50943506e+00 2.41594866e-01
-1.54049039e-01 1.02071333e+00 -2.66018003e-01 -2.10943833e-01
8.61357868e-01 -1.43522978e+00 6.77168727e-01 1.55943066e-01
1.04284711e-01 6.93003416e-01 6.07028306e-01 -5.69313228e-01
-1.06329775e+00 -8.32288504e-01 2.28574896e+00 -1.07064462e+00
1.18750966e+00 -2.86742538e-01 -6.40520751e-01 4.76249516e-01
2.41707772e-01 -2.17875078e-01 1.42052305e+00 6.37010992e-01
-4.90510166e-01 3.18537623e-01 -9.59773242e-01 4.47067469e-01
1.25912404e+00 -8.32787693e-01 -1.14804244e+00 4.19297040e-01
9.84836519e-01 6.03154674e-03 -1.23546684e+00 1.55905515e-01
8.15038264e-01 -5.85910678e-01 6.74654365e-01 -9.03415084e-01
4.36631173e-01 1.07137606e-01 -7.73788512e-01 -1.40960968e+00
-4.89172131e-01 -3.36107939e-01 -4.84360941e-02 8.67179930e-01
8.50244701e-01 -4.16331679e-01 3.65099132e-01 4.55553621e-01
-3.34521830e-01 -9.33513165e-01 -1.09933650e+00 -1.05217874e+00
7.78566360e-01 -9.21191037e-01 3.62437844e-01 9.21666145e-01
7.41130173e-01 4.83771682e-01 1.03354208e-01 -2.95004219e-01
4.50451970e-02 -3.60945731e-01 2.78393388e-01 -1.32315695e+00
-4.89507392e-02 -6.26217902e-01 -4.90237087e-01 -7.38574445e-01
4.62920457e-01 -1.33032358e+00 -1.71737969e-01 -1.27307153e+00
-3.54765989e-02 -3.10824782e-01 -6.14396393e-01 5.05040824e-01
-2.93759435e-01 1.14437163e-01 2.49478459e-01 -2.29353800e-01
-6.30598187e-01 4.84753400e-01 6.34211898e-01 1.52973652e-01
-1.42121062e-01 -5.23121893e-01 -6.82153881e-01 7.41216481e-01
1.07900476e+00 -3.82461667e-01 -1.00648247e-01 -6.20793223e-01
6.55239463e-01 -5.90905309e-01 -2.72557139e-01 -9.20559406e-01
2.49740362e-01 -5.81659153e-02 2.41397828e-01 -5.86214781e-01
3.49171728e-01 -5.28361380e-01 -3.12586635e-01 6.58045352e-01
-5.00453770e-01 6.66094363e-01 4.73698854e-01 -9.92475823e-02
-6.45850301e-02 -5.06432474e-01 4.42399293e-01 -9.79286209e-02
-8.33296537e-01 -6.43078461e-02 -7.48488367e-01 2.45903641e-01
7.62506723e-01 -8.21919888e-02 -2.96485778e-02 1.40788272e-01
-1.11684096e+00 2.24082351e-01 2.56345153e-01 7.95577407e-01
6.23532116e-01 -1.48180771e+00 -6.41134620e-01 2.53744304e-01
5.19506693e-01 -8.44085932e-01 -2.86070079e-01 4.25974071e-01
-5.07534504e-01 6.82554007e-01 -2.73603320e-01 7.54791945e-02
-1.23237073e+00 4.72205371e-01 -8.91976655e-02 -4.50603962e-01
-3.46205562e-01 1.18981564e+00 -7.03652024e-01 -6.97951794e-01
8.45243931e-02 -4.74774897e-01 -4.06697810e-01 4.46987361e-01
7.36924112e-01 5.30041635e-01 3.17784220e-01 -6.99431419e-01
-6.17737532e-01 8.19981992e-01 -2.76114732e-01 -1.70971125e-01
1.53553975e+00 -2.19271006e-03 -5.65377027e-02 7.58925259e-01
1.61438525e+00 1.35698333e-01 -3.12643737e-01 -3.35598350e-01
6.14584446e-01 -2.43717030e-01 -1.21446541e-02 -9.29127038e-01
-2.30531842e-01 7.29684114e-01 6.80209696e-01 7.79330209e-02
6.05614007e-01 -1.29309461e-01 7.44843364e-01 9.33570981e-01
3.58832568e-01 -1.74876034e+00 8.59916359e-02 1.28670883e+00
8.71879518e-01 -1.23374772e+00 7.19513223e-02 -3.26368421e-01
-6.07474923e-01 1.35443389e+00 6.71336412e-01 -3.33624125e-01
1.20995104e+00 2.23850146e-01 -5.76961450e-02 1.72314269e-03
-1.35042000e+00 -3.02472651e-01 4.33008641e-01 8.53369594e-01
1.02675223e+00 3.22258621e-01 -9.48794663e-01 1.34039640e+00
-8.24393511e-01 -2.60998964e-01 3.26703936e-01 9.62390363e-01
-7.42060661e-01 -1.57835078e+00 3.02055240e-01 5.37432313e-01
-2.49630645e-01 -7.04617083e-01 -4.71122831e-01 8.65827084e-01
3.36701870e-01 9.63263273e-01 -1.11149913e-02 -6.31048262e-01
6.94621325e-01 6.18944764e-01 5.15010297e-01 -1.03326225e+00
-8.53316009e-01 -4.22832340e-01 5.52769005e-01 -8.03008020e-01
-6.17501214e-02 -6.87938333e-01 -1.15568686e+00 -5.35025954e-01
-1.85604259e-01 1.09324291e-01 7.63051808e-01 7.38333344e-01
6.24765046e-02 3.01480323e-01 6.90364480e-01 -8.93866062e-01
-3.95113736e-01 -1.23618877e+00 -4.59027261e-01 2.91516930e-01
-4.97210100e-02 -4.65084255e-01 -4.80528265e-01 -2.01024115e-02] | [10.629242897033691, 10.442846298217773] |
b3bd0a92-7d17-470e-9e27-81b446b669a6 | hybrid-model-for-single-stage-multi-person | 2305.01167 | null | https://arxiv.org/abs/2305.01167v2 | https://arxiv.org/pdf/2305.01167v2.pdf | Hybrid model for Single-Stage Multi-Person Pose Estimation | In general, human pose estimation methods are categorized into two approaches according to their architectures: regression (i.e., heatmap-free) and heatmap-based methods. The former one directly estimates precise coordinates of each keypoint using convolutional and fully-connected layers. Although this approach is able to detect overlapped and dense keypoints, unexpected results can be obtained by non-existent keypoints in a scene. On the other hand, the latter one is able to filter the non-existent ones out by utilizing predicted heatmaps for each keypoint. Nevertheless, it suffers from quantization error when obtaining the keypoint coordinates from its heatmaps. In addition, unlike the regression one, it is difficult to distinguish densely placed keypoints in an image. To this end, we propose a hybrid model for single-stage multi-person pose estimation, named HybridPose, which mutually overcomes each drawback of both approaches by maximizing their strengths. Furthermore, we introduce self-correlation loss to inject spatial dependencies between keypoint coordinates and their visibility. Therefore, HybridPose is capable of not only detecting densely placed keypoints, but also filtering the non-existent keypoints in an image. Experimental results demonstrate that proposed HybridPose exhibits the keypoints visibility without performance degradation in terms of the pose estimation accuracy. | ['Jungho Lee', 'Dowoo Kwon', 'Lanying Jin', 'Wonhyeok Im', 'Jungpyo Kim', 'Hyotae Lee', 'Bosang Kim', 'Jonghyun Kim'] | 2023-05-02 | null | null | null | null | ['multi-person-pose-estimation'] | ['computer-vision'] | [-3.38761687e-01 -1.76178187e-01 2.30298638e-01 -1.32205546e-01
-5.04028320e-01 -2.89501429e-01 4.15577173e-01 3.87339503e-01
-5.84954023e-01 6.15747988e-01 -2.10520044e-01 4.41280812e-01
-1.29917994e-01 -9.74042296e-01 -7.16568112e-01 -6.20539844e-01
1.10786706e-01 4.06987369e-01 5.59863627e-01 -2.10734382e-01
1.42539591e-01 5.04769623e-01 -1.65513146e+00 -1.64708674e-01
1.02939665e+00 1.04750764e+00 2.41474435e-01 2.39619017e-01
-5.49377594e-03 2.14174375e-01 -6.48089230e-01 -4.79240000e-01
1.77767351e-01 -1.74280927e-01 -1.18039940e-02 -2.62204885e-01
7.48207033e-01 -2.23314077e-01 -4.00713027e-01 1.14282298e+00
4.73830730e-01 -1.09508283e-01 3.99606019e-01 -1.48106158e+00
-4.06159252e-01 4.29762863e-02 -9.83553231e-01 7.83499926e-02
5.14473081e-01 1.23454720e-01 7.91887045e-01 -1.01677287e+00
4.76698756e-01 1.30786645e+00 8.35148275e-01 2.63334271e-02
-1.04184067e+00 -7.99598217e-01 2.65985519e-01 2.90897697e-01
-2.00382280e+00 -1.62814647e-01 8.48813057e-01 -3.59520614e-01
4.75575387e-01 4.06988889e-01 8.83778989e-01 6.90794766e-01
3.16045046e-01 8.17486107e-01 1.07326412e+00 -2.26383403e-01
9.06173959e-02 1.36650875e-01 5.21369725e-02 8.72838080e-01
3.27399164e-01 2.98517384e-02 -6.08062744e-01 -1.40020967e-01
9.20180202e-01 1.80979252e-01 -5.87184608e-01 -7.17049778e-01
-1.46812165e+00 3.89701217e-01 9.25115943e-01 2.24509671e-01
-5.39022326e-01 -8.72752909e-03 2.36751623e-02 -2.13186324e-01
1.03351548e-01 1.77250892e-01 -1.13513350e-01 -1.86111555e-02
-1.10234261e+00 5.43013573e-01 1.74358875e-01 1.02885032e+00
1.00480044e+00 -3.67894471e-01 -3.26517314e-01 4.47436810e-01
3.34147155e-01 5.88658988e-01 3.53860945e-01 -2.85911947e-01
7.89317727e-01 9.46592391e-01 4.72645819e-01 -1.88293755e+00
-7.76987195e-01 -5.78637838e-01 -9.10768270e-01 2.19693691e-01
5.57944000e-01 -3.51076946e-02 -8.90748441e-01 1.59766293e+00
5.00952959e-01 1.45940572e-01 -4.39309657e-01 1.30508399e+00
8.90186727e-01 5.79298496e-01 1.96037069e-02 -1.49269924e-01
1.43606472e+00 -8.34832311e-01 -7.01166570e-01 -2.59651929e-01
-1.86629745e-03 -5.60473680e-01 8.05953443e-01 4.31724072e-01
-8.51792216e-01 -8.65421236e-01 -1.17120492e+00 1.15682304e-01
-4.04740751e-01 5.16264975e-01 4.65457916e-01 6.00212216e-01
-8.96712303e-01 3.60382527e-01 -8.53660345e-01 -2.44214669e-01
1.18239783e-01 2.81538516e-01 -4.25187349e-01 6.56904578e-02
-1.12543714e+00 7.99137354e-01 3.68425250e-01 3.72405559e-01
-3.90647262e-01 -6.31578088e-01 -8.66976976e-01 1.83719784e-01
3.75284404e-01 -6.09076858e-01 7.17694700e-01 -6.32847786e-01
-1.01314378e+00 4.17662859e-01 -2.77763873e-01 -1.04864702e-01
9.48670268e-01 -5.45165181e-01 -4.77579832e-01 1.58746153e-01
1.92576796e-01 5.66279650e-01 9.34963942e-01 -1.33004570e+00
-8.89845371e-01 -6.54259384e-01 -7.50051364e-02 3.69794518e-01
-3.91391188e-01 -2.72298396e-01 -9.25858855e-01 -5.62046170e-01
6.30243719e-01 -8.16107631e-01 -7.39749297e-02 3.10661048e-01
-8.95931184e-01 -2.23919824e-01 6.61000133e-01 -4.61827129e-01
1.16094542e+00 -1.93957472e+00 9.97503027e-02 3.64007950e-01
4.03399587e-01 1.44285798e-01 2.09569380e-01 3.66128296e-01
2.05661088e-01 -2.25075543e-01 1.51580602e-01 -3.37274432e-01
-7.80617148e-02 -2.36425251e-01 1.49079144e-01 7.18761384e-01
1.58649430e-01 7.70754695e-01 -8.76103580e-01 -6.74687803e-01
6.32365346e-01 8.25052798e-01 -3.02341729e-01 -5.84165100e-03
2.11801067e-01 4.34609860e-01 -4.10490781e-01 6.65872037e-01
1.09189844e+00 3.43665779e-02 -8.34058076e-02 -5.79990625e-01
-3.62836868e-01 -2.58431017e-01 -1.71591938e+00 1.71635485e+00
-1.90656304e-01 2.75987476e-01 -1.34332636e-02 -4.55869734e-01
9.07719672e-01 8.22126195e-02 4.45606977e-01 -6.71987414e-01
-4.84432466e-02 2.01133758e-01 -3.81219298e-01 -1.47428870e-01
9.06572998e-01 1.01299696e-01 -5.97850308e-02 -9.78055671e-02
-1.67767391e-01 4.99787658e-01 1.21988147e-01 -1.81810074e-02
7.16142416e-01 1.20796435e-01 2.96540350e-01 2.73760445e-02
7.08613753e-01 -1.01438656e-01 7.64112532e-01 6.79667532e-01
-3.06054682e-01 8.86150718e-01 2.32189849e-01 -4.39000398e-01
-8.11251581e-01 -1.19878006e+00 -1.70894966e-01 6.20356679e-01
9.42650318e-01 -7.97116756e-01 -6.58220947e-01 -6.20256841e-01
1.05872154e-01 1.25559226e-01 -5.72480619e-01 2.23255679e-02
-6.77453637e-01 -6.21031940e-01 2.92332500e-01 4.22402978e-01
8.16571891e-01 -6.83859169e-01 -9.70910966e-01 5.78240715e-02
-5.00133276e-01 -9.23492432e-01 -3.84687930e-01 -2.27633089e-01
-3.67636859e-01 -1.25404692e+00 -1.19686496e+00 -5.52005589e-01
8.84013414e-01 5.72856128e-01 8.77121449e-01 8.02016407e-02
-1.22437976e-01 1.81438640e-01 -3.38534951e-01 -1.65858254e-01
3.24492544e-01 -1.70944049e-03 1.22861251e-01 2.98252940e-01
5.52533448e-01 -3.41634542e-01 -1.12317693e+00 5.52036285e-01
-3.28554094e-01 7.14860484e-02 8.74791682e-01 6.17788434e-01
7.96624184e-01 4.09298807e-01 1.54705524e-01 -2.17279896e-01
3.39989603e-01 -2.29031742e-01 -5.17822981e-01 3.18180203e-01
-4.19755876e-01 -1.82930633e-01 4.89797920e-01 -2.09068552e-01
-7.34548390e-01 4.46776599e-01 -7.43678510e-02 -3.40745986e-01
-2.73006201e-01 1.94910496e-01 -2.43254974e-01 -1.60086378e-01
4.87325251e-01 4.41899925e-01 -2.28111759e-01 -5.73045671e-01
3.59463751e-01 4.89085525e-01 7.30440080e-01 -2.41329432e-01
1.12061083e+00 6.74351513e-01 -1.62075926e-02 -8.81137490e-01
-3.07308257e-01 -6.53019309e-01 -8.21671426e-01 -4.80896622e-01
1.02921975e+00 -1.00265741e+00 -8.33632588e-01 6.19553983e-01
-1.18055916e+00 5.44914961e-01 1.74577385e-01 4.52627212e-01
-1.95862249e-01 5.62138498e-01 -2.41881058e-01 -9.38977420e-01
-1.40950471e-01 -1.18617952e+00 1.28647268e+00 5.57986259e-01
-7.32953325e-02 -3.72651815e-01 -7.23377466e-02 -1.35088097e-02
4.55819219e-02 5.66607833e-01 4.35083419e-01 -2.67977715e-01
-7.02089965e-01 -6.60228610e-01 -3.92984033e-01 -2.17461020e-01
-6.18809536e-02 -1.69658437e-01 -8.44695032e-01 -4.73271787e-01
-3.73940170e-01 -9.46032349e-03 6.80621803e-01 3.82782221e-01
7.83526123e-01 -8.04302022e-02 -6.02695763e-01 5.54358721e-01
1.43500698e+00 -4.87229787e-02 6.75721586e-01 6.55245543e-01
9.08250093e-01 5.65072000e-01 8.77315640e-01 4.17978853e-01
5.85586369e-01 1.12327170e+00 5.45091212e-01 -3.58815610e-01
9.52156037e-02 -5.61257720e-01 6.84028640e-02 2.14744166e-01
-2.64789969e-01 -3.80469263e-02 -7.77148187e-01 3.22134972e-01
-2.00600171e+00 -7.51656234e-01 -4.75295573e-01 2.40393639e+00
2.80450046e-01 2.74797231e-01 3.16586554e-01 2.51408160e-01
9.75630164e-01 2.09508181e-01 -5.35830140e-01 3.67465049e-01
-1.34404510e-01 -7.95852840e-02 5.48886538e-01 8.28515142e-02
-1.41029525e+00 6.24036193e-01 4.94659472e+00 7.32512951e-01
-8.33004594e-01 -7.05912411e-02 2.61066437e-01 -3.22621614e-02
4.87530492e-02 -2.12843791e-01 -9.02895689e-01 6.81390703e-01
5.41751087e-02 2.53336310e-01 -8.01351592e-02 9.15038705e-01
9.38849971e-02 -3.76314908e-01 -7.02820361e-01 1.40557039e+00
1.24667160e-01 -8.58306944e-01 -1.91889837e-01 -1.21312328e-01
3.77455711e-01 -5.17574608e-01 -7.81271886e-03 1.35016903e-01
-3.17890108e-01 -8.36131752e-01 1.07904232e+00 5.71007252e-01
5.85158706e-01 -1.09151971e+00 7.42434025e-01 5.15560389e-01
-1.58075249e+00 -1.17461033e-01 -4.03141290e-01 -2.13265829e-02
1.68910369e-01 6.23246431e-01 -3.52518976e-01 7.88627207e-01
9.39944029e-01 4.47555423e-01 -8.59585047e-01 1.60544598e+00
-4.76568699e-01 -1.91704348e-01 -4.49088395e-01 -4.42991890e-02
-2.85974937e-03 -9.44626033e-02 6.42247975e-01 1.06119645e+00
5.09998858e-01 -2.03860208e-01 4.43560034e-01 9.88844395e-01
4.43398982e-01 2.62495190e-01 -1.95124522e-01 6.30595922e-01
7.18973994e-01 1.40005481e+00 -9.65058625e-01 -2.23659873e-01
-2.80457973e-01 1.26688373e+00 3.25219482e-01 3.90093088e-01
-9.12914991e-01 -7.64140904e-01 5.60413599e-01 4.84023631e-01
3.43044758e-01 -3.79868597e-01 -2.16831088e-01 -1.09495676e+00
4.80737954e-01 -5.56568861e-01 2.41833612e-01 -7.41376460e-01
-8.58452678e-01 6.45080209e-01 -2.64621433e-02 -1.48661900e+00
-1.60942152e-02 -4.09948438e-01 -3.83280694e-01 9.68839824e-01
-1.42272902e+00 -1.27267671e+00 -8.73632133e-01 6.84881806e-01
2.27387264e-01 2.70539075e-01 4.72733647e-01 3.70694906e-01
-6.83850765e-01 8.32615793e-01 1.65898446e-02 1.85114563e-01
7.96203017e-01 -1.35689187e+00 4.26229179e-01 1.01016712e+00
1.48198664e-01 7.63698101e-01 7.11906910e-01 -8.85483921e-01
-1.26483798e+00 -8.13184738e-01 7.66009212e-01 -3.37965608e-01
1.10701047e-01 -5.55936456e-01 -7.32758284e-01 3.15003544e-01
-3.42201829e-01 -4.40967977e-02 1.77136645e-01 1.77656397e-01
-2.43766606e-01 -1.82918429e-01 -1.04003239e+00 6.28051341e-01
8.52364957e-01 -9.33312103e-02 -5.14864266e-01 1.58739090e-01
2.53027529e-01 -4.97069895e-01 -6.75873399e-01 3.82392317e-01
6.66784108e-01 -1.48577917e+00 1.32073963e+00 2.13990659e-01
1.46911412e-01 -7.76067078e-01 2.62601852e-01 -1.20041251e+00
-5.95344186e-01 -2.81784564e-01 -1.42091483e-01 1.24252820e+00
-1.98897049e-02 -5.09491563e-01 9.39897835e-01 4.05237019e-01
7.56597519e-02 -6.92122996e-01 -1.10229838e+00 -7.45033026e-01
-3.23578119e-01 -2.19092265e-01 7.31596529e-01 5.77921391e-01
-3.64252133e-03 7.70068020e-02 -6.11922920e-01 5.58729887e-01
7.63512492e-01 2.09781915e-01 9.64425743e-01 -1.30327284e+00
-6.30494282e-02 -3.01998347e-01 -9.30166841e-01 -1.39797318e+00
-4.71628904e-01 -3.30824822e-01 1.99981794e-01 -1.46858966e+00
1.16112210e-01 -7.05886900e-01 -3.27910155e-01 2.73241639e-01
-4.66415346e-01 5.34948409e-01 3.05155635e-01 4.17292178e-01
-6.83849394e-01 4.30466264e-01 1.12051952e+00 -2.99683716e-02
-3.90454173e-01 -1.34267667e-02 -4.53852594e-01 8.29184353e-01
5.82701802e-01 -2.10655689e-01 -1.77937433e-01 -1.78983718e-01
2.70902038e-01 -1.57227546e-01 8.97269130e-01 -1.60459781e+00
4.84486938e-01 7.44554624e-02 9.21800613e-01 -1.19803119e+00
5.57511687e-01 -7.52906561e-01 2.42423058e-01 4.49785650e-01
1.00854345e-01 1.66177720e-01 -2.66027339e-02 8.59017611e-01
-1.43423915e-01 9.98078510e-02 7.61132360e-01 -2.39944860e-01
-8.76104414e-01 2.64504939e-01 1.38623668e-02 -2.93853015e-01
1.19754434e+00 -4.26969379e-01 -2.30817974e-01 -2.40524352e-01
-4.43127751e-01 2.18925506e-01 5.25194287e-01 5.14057577e-01
8.28449845e-01 -1.49302840e+00 -4.89952236e-01 3.69163096e-01
2.64229268e-01 5.84517270e-02 5.51151752e-01 1.13792872e+00
-4.76122022e-01 4.23266113e-01 -1.80870578e-01 -7.63679802e-01
-1.32635903e+00 6.71419322e-01 2.65086979e-01 -8.93128142e-02
-7.80373991e-01 7.71882951e-01 2.35927492e-01 -1.63087264e-01
3.41628700e-01 -2.35246584e-01 -4.09154087e-01 1.55861706e-01
8.35177422e-01 3.42921615e-01 6.39998093e-02 -1.09708583e+00
-6.48225129e-01 1.03249609e+00 6.29901513e-03 1.86931804e-01
8.94445300e-01 -3.00242871e-01 1.41482368e-01 3.21449578e-01
1.05868089e+00 1.57306150e-01 -1.28359473e+00 -2.23346382e-01
-3.18815410e-01 -8.66861045e-01 -1.16386905e-01 -5.68499088e-01
-9.98102367e-01 7.76855409e-01 9.00696397e-01 1.49398401e-01
1.04448676e+00 -2.83582836e-01 9.26625371e-01 9.49513763e-02
6.23576522e-01 -1.12057328e+00 -5.81298061e-02 -6.34449497e-02
7.37649143e-01 -1.20817089e+00 3.08923781e-01 -6.29924357e-01
-4.35487449e-01 1.17606843e+00 8.44779491e-01 -1.29028842e-01
4.16078955e-01 -1.63007259e-01 3.66697460e-02 -4.27023560e-01
1.29887030e-01 -3.08985174e-01 4.82232779e-01 7.37585306e-01
8.69799629e-02 4.36284915e-02 -3.66921544e-01 4.51210111e-01
-3.82037073e-01 -3.23525906e-01 -2.13826522e-02 7.51388609e-01
-5.51067889e-01 -6.46143436e-01 -9.22573686e-01 1.99694097e-01
-1.23400003e-01 9.75267291e-02 -3.42936397e-01 9.39569950e-01
4.82494801e-01 9.44724262e-01 1.20408893e-01 -7.15300024e-01
6.37248933e-01 -1.89756081e-01 2.34705135e-01 -1.93973988e-01
-4.65514958e-01 1.49601117e-01 -2.69945920e-01 -6.73664868e-01
-2.21296951e-01 -5.03360629e-01 -1.09683132e+00 -1.74372524e-01
-5.00602484e-01 1.06352188e-01 5.75197756e-01 6.17886126e-01
2.89201409e-01 3.93272191e-01 3.86004031e-01 -1.07053483e+00
-2.59749800e-01 -7.05273926e-01 -5.75453162e-01 4.07085419e-01
3.33188802e-01 -1.07841194e+00 -1.41213983e-01 -4.42425311e-01] | [7.096185684204102, -0.8242926597595215] |
8b6c800d-8835-4f16-86ec-a4b7b35e984b | gesture-recognition-with-keypoint-and-radar | 2302.09998 | null | https://arxiv.org/abs/2302.09998v1 | https://arxiv.org/pdf/2302.09998v1.pdf | Gesture Recognition with Keypoint and Radar Stream Fusion for Automated Vehicles | We present a joint camera and radar approach to enable autonomous vehicles to understand and react to human gestures in everyday traffic. Initially, we process the radar data with a PointNet followed by a spatio-temporal multilayer perceptron (stMLP). Independently, the human body pose is extracted from the camera frame and processed with a separate stMLP network. We propose a fusion neural network for both modalities, including an auxiliary loss for each modality. In our experiments with a collected dataset, we show the advantages of gesture recognition with two modalities. Motivated by adverse weather conditions, we also demonstrate promising performance when one of the sensors lacks functionality. | ['Vasileios Belagiannis', 'Klaus Dietmayer', 'Christian Waldschmidt', 'Nicolai Kern', 'Adrian Holzbock'] | 2023-02-20 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 4.57410008e-01 -8.66121799e-02 -1.75928071e-01 -6.30453110e-01
-4.53325599e-01 -2.58154094e-01 1.03746092e+00 -6.92602575e-01
-9.35459614e-01 4.05855626e-01 -6.60646558e-02 -1.54052198e-01
1.86053574e-01 -5.49278617e-01 -7.63562560e-01 -6.05839431e-01
2.15827953e-02 3.56287569e-01 4.48875993e-01 -6.88208640e-02
-8.60210657e-02 9.26887095e-01 -1.60589552e+00 1.22034021e-01
1.00385651e-01 1.15170169e+00 6.54056892e-02 1.17192888e+00
3.52822334e-01 9.66777444e-01 -3.48946631e-01 -2.09904686e-01
6.06159091e-01 2.32515894e-02 -2.54988819e-01 1.00925311e-01
6.13291621e-01 -7.52198935e-01 -1.04139376e+00 6.92134678e-01
3.17058653e-01 3.73048723e-01 5.28221071e-01 -1.61219370e+00
5.83367869e-02 -3.16052362e-02 -4.59486067e-01 -1.35740578e-01
4.03621167e-01 5.24237335e-01 6.11250162e-01 -7.09191799e-01
6.44841492e-01 1.35234118e+00 5.91284096e-01 7.56936371e-01
-7.45004773e-01 -6.56814277e-01 1.61043987e-01 3.98820400e-01
-1.16222239e+00 -7.03591228e-01 8.87180269e-01 -2.01090172e-01
1.03358984e+00 7.85280764e-02 4.81682330e-01 1.47787976e+00
1.68312803e-01 8.54136169e-01 7.62084723e-01 6.90830424e-02
2.43040815e-01 -4.56488460e-01 1.33625656e-01 6.21063113e-01
1.26559690e-01 7.49704957e-01 -7.11886585e-01 9.43382084e-02
6.29655659e-01 5.34745455e-01 5.27037047e-02 -4.92750496e-01
-1.29207993e+00 3.67874831e-01 5.91243446e-01 -1.20821431e-01
-6.46925807e-01 6.96693063e-01 1.02549963e-01 2.33009994e-01
1.41888848e-02 -2.75781989e-01 -1.14138022e-01 -1.12704270e-01
-9.31320786e-01 -1.50920367e-02 7.17755318e-01 1.02630579e+00
5.44912994e-01 -3.48574333e-02 9.93771404e-02 1.39277354e-01
8.16896021e-01 1.19605958e+00 7.03051910e-02 -1.11223102e+00
7.40041912e-01 2.22520769e-01 2.71689177e-01 -8.98493707e-01
-6.65625811e-01 2.02629998e-01 -8.26804340e-01 7.48269379e-01
3.60113114e-01 -4.37663674e-01 -1.20367825e+00 1.49909842e+00
2.48084173e-01 6.20598674e-01 3.71608168e-01 1.33765841e+00
7.91859388e-01 4.64389592e-01 2.39369228e-01 1.16227232e-01
1.38417995e+00 -8.50015223e-01 -6.95515215e-01 -5.88666916e-01
5.17619587e-02 -2.20585421e-01 1.87857613e-01 2.77431339e-01
-7.86082327e-01 -7.57004440e-01 -1.22414911e+00 -2.78783077e-03
-3.45824093e-01 3.39525193e-01 2.76342243e-01 4.37130541e-01
-8.86546910e-01 1.91569835e-01 -1.52516234e+00 -6.03905976e-01
1.28284469e-01 4.40818548e-01 -5.29499292e-01 2.55379416e-02
-1.07159281e+00 1.28394711e+00 1.22583516e-01 5.73390186e-01
-7.19249308e-01 -2.77825333e-02 -1.12019837e+00 -3.54687512e-01
2.17753187e-01 -7.68564641e-01 1.15745628e+00 -5.34940839e-01
-1.91300786e+00 5.33341289e-01 -3.63381863e-01 -7.62716353e-01
7.36942947e-01 -5.33490300e-01 -5.72231233e-01 3.78272533e-01
-2.57594198e-01 9.59820688e-01 1.12826467e+00 -1.24695468e+00
-8.42813432e-01 -4.86408353e-01 -6.17528446e-02 3.46669108e-01
2.97038049e-01 1.58302262e-02 -7.53549337e-01 -1.95295885e-01
3.10676217e-01 -1.12765384e+00 -3.20429325e-01 1.85226619e-01
-2.12700307e-01 8.74485150e-02 1.07753265e+00 -5.67500532e-01
6.23752654e-01 -2.11829662e+00 2.47576043e-01 4.70949054e-01
2.01732099e-01 1.93102673e-01 -1.84251711e-01 4.88606319e-02
2.83605486e-01 -4.56739843e-01 -7.67726451e-02 -6.84897602e-01
3.98407355e-02 5.88415444e-01 -4.43391025e-01 7.69564271e-01
2.39139259e-01 9.78874505e-01 -6.71722889e-01 -3.45058531e-01
5.91013849e-01 7.77131855e-01 -1.54958650e-01 4.13531840e-01
-1.07667483e-02 6.48086071e-01 -5.61612368e-01 8.33431959e-01
4.68552262e-01 1.17338665e-01 6.58221496e-03 -2.68552750e-01
-1.61396503e-01 -7.79668167e-02 -1.09794188e+00 1.50166678e+00
-3.32706481e-01 8.64964247e-01 4.25935179e-01 -6.25674784e-01
9.14800346e-01 3.68950665e-01 4.49955970e-01 -8.18736851e-01
2.71352708e-01 -6.75391406e-02 -3.42433363e-01 -7.01649070e-01
6.01528108e-01 -1.53209835e-01 -2.44258210e-01 1.96271405e-01
4.17084545e-02 1.07544987e-02 -1.36222392e-01 -2.87177503e-01
1.43973577e+00 3.41513753e-01 1.42917141e-01 6.50762022e-01
3.88200969e-01 5.84854968e-02 4.06562239e-01 9.00837481e-01
-5.66120565e-01 5.84468544e-01 -1.03297755e-01 -6.90650523e-01
-7.02166677e-01 -1.33459783e+00 3.15113634e-01 1.06622422e+00
4.47639078e-01 1.34055644e-01 -1.02733769e-01 -6.29016280e-01
1.17844157e-01 4.31497216e-01 -3.41903269e-01 -2.27439348e-02
-1.03950202e+00 -1.56830639e-01 9.28881288e-01 1.05808890e+00
7.21352875e-01 -9.49994087e-01 -1.41827798e+00 -8.63846093e-02
1.76999401e-02 -1.55323350e+00 -1.29289389e-01 1.20156147e-01
-6.14873469e-01 -9.69155788e-01 -5.92449009e-01 -4.58320886e-01
4.07106429e-01 3.14631492e-01 5.39782465e-01 -5.26482798e-02
6.81128651e-02 6.78813517e-01 -2.16369390e-01 -3.84275168e-01
-6.26824722e-02 -2.20654070e-01 3.57543647e-01 2.07871541e-01
4.10122067e-01 -5.16307950e-01 -4.15394932e-01 2.36291349e-01
-4.78599429e-01 -1.88217405e-02 8.18297088e-01 4.24086422e-01
1.60458609e-01 -3.91825050e-01 -1.87933862e-01 -1.06867209e-01
7.42467269e-02 -1.10608637e-01 -7.61167645e-01 -1.60186645e-02
9.37879086e-02 -1.05598830e-02 1.73472092e-01 -4.48688060e-01
-1.11379313e+00 8.31640720e-01 -1.22361975e-02 -9.35551703e-01
-8.19966733e-01 2.64973164e-01 -2.08187044e-01 -6.75837100e-02
3.10329586e-01 -2.57820096e-02 3.03956032e-01 -3.69492173e-01
6.61351025e-01 6.30610764e-01 1.06875980e+00 -2.04121783e-01
1.04112589e+00 8.60090256e-01 1.95773989e-01 -1.09072638e+00
-3.57045680e-01 -6.33005679e-01 -1.00020194e+00 -5.92621207e-01
1.19788480e+00 -1.03105605e+00 -1.24069285e+00 5.04727304e-01
-1.39971673e+00 -3.20742816e-01 -1.15275972e-01 1.05825222e+00
-7.43033230e-01 1.41409010e-01 -5.15830934e-01 -1.02891862e+00
-5.57971597e-02 -1.02130687e+00 1.49814320e+00 2.17476308e-01
-1.12725489e-01 -6.35131240e-01 2.33020067e-01 2.72147834e-01
2.84097731e-01 2.78047115e-01 -9.99732614e-02 -6.04907632e-01
-7.55416632e-01 -4.10419703e-01 -2.35006556e-01 -3.01379673e-02
-2.16259792e-01 -1.42200381e-01 -1.05785072e+00 -1.97250947e-01
-1.26689166e-01 -2.57529199e-01 1.18688798e+00 3.15791786e-01
4.26644921e-01 -7.45340958e-02 -6.93882406e-01 7.03029037e-01
9.23872709e-01 2.86881328e-01 3.94636869e-01 2.27784425e-01
8.67540061e-01 4.85897869e-01 4.10543054e-01 2.66596526e-01
5.50863087e-01 6.76074803e-01 4.78935003e-01 -2.43897140e-02
-1.02053002e-01 -1.41288206e-01 8.46566379e-01 2.74322867e-01
-3.89314145e-01 -1.69044197e-01 -1.04817998e+00 1.65382177e-01
-2.06275344e+00 -1.03819525e+00 9.91268605e-02 1.77810979e+00
7.99637064e-02 1.56496644e-01 6.34816512e-02 -2.33440638e-01
6.48777544e-01 4.54233676e-01 -6.15416050e-01 -3.31260115e-02
7.66716823e-02 -7.70996697e-03 9.28641856e-01 6.11799836e-01
-1.50400853e+00 9.26612258e-01 7.17911577e+00 4.92818952e-02
-1.26234341e+00 -1.13569893e-01 -1.47991419e-01 -4.19377118e-01
4.33329493e-01 -1.98200911e-01 -7.78203905e-01 1.44996300e-01
1.06026673e+00 4.48139012e-01 2.49504551e-01 6.99747980e-01
1.39997229e-01 -1.11371413e-01 -1.32313907e+00 1.01855373e+00
2.79919177e-01 -1.04400682e+00 -3.20747733e-01 -1.42206606e-02
9.20517296e-02 5.77519596e-01 -1.38515428e-01 3.84052038e-01
3.81174833e-01 -9.11113262e-01 8.60607266e-01 1.04224789e+00
6.65244043e-01 -3.80958855e-01 6.32125616e-01 5.26375175e-01
-1.40033293e+00 -8.19326565e-02 -7.22860359e-03 -2.60050088e-01
5.87248266e-01 -1.56558715e-02 -5.92031896e-01 4.70762521e-01
6.71042025e-01 7.32432663e-01 -3.48215371e-01 8.28049660e-01
-3.99901748e-01 3.91134918e-01 -7.55295813e-01 7.60525614e-02
1.83560804e-01 -1.13433272e-01 9.86589134e-01 1.18119371e+00
2.16631055e-01 4.07535762e-01 4.23608005e-01 6.42093956e-01
2.08748713e-01 -8.29138100e-01 -9.72505450e-01 3.80875707e-01
3.35898399e-01 1.10388660e+00 -4.62317288e-01 -3.86539251e-01
-5.75544477e-01 9.16228533e-01 -5.08406349e-02 7.19423532e-01
-9.11067903e-01 -1.56694397e-01 8.06508183e-01 -3.79774868e-01
4.47129548e-01 -6.92483723e-01 -4.06995177e-01 -1.06519330e+00
1.05984375e-01 -1.71721354e-01 2.16324031e-01 -1.02336717e+00
-1.08129334e+00 4.64809805e-01 1.88692212e-01 -1.57470381e+00
-4.89788741e-01 -8.51999700e-01 -6.13736033e-01 7.14309335e-01
-1.50952983e+00 -1.48236895e+00 -3.34654987e-01 6.63805723e-01
3.99049640e-01 -2.31777459e-01 3.01647484e-01 1.17483400e-01
-4.48066741e-01 1.46169081e-01 -4.35014367e-01 5.18034101e-01
5.42622209e-01 -8.70971203e-01 4.97137457e-01 9.84914005e-01
-4.13464680e-02 4.15202141e-01 7.03973591e-01 -8.87909770e-01
-1.83597529e+00 -1.17896473e+00 7.25414038e-01 -6.23812556e-01
6.29585445e-01 -2.85273403e-01 -4.23750281e-01 1.01128304e+00
1.27040967e-01 2.78911710e-01 1.04768649e-01 -1.78256840e-01
-8.84315968e-02 -1.32763296e-01 -9.89247859e-01 6.47354603e-01
9.76798117e-01 -5.64317107e-01 -1.11000288e+00 -1.52822450e-01
5.14232993e-01 -6.52265906e-01 -4.78953153e-01 5.53677082e-01
9.73769844e-01 -5.14085054e-01 9.97637153e-01 -5.86372495e-01
-1.63512930e-01 -5.28630495e-01 -4.51509565e-01 -9.05003667e-01
-1.04256682e-01 -4.06928122e-01 -4.37933624e-01 5.15370786e-01
1.73507422e-01 -4.50655192e-01 1.05387747e+00 8.79994571e-01
7.12438449e-02 -2.71847937e-02 -1.30569553e+00 -7.51541078e-01
-4.28914189e-01 -1.02520704e+00 3.15433145e-01 3.26760352e-01
-3.46401148e-02 3.75224084e-01 -7.45007992e-01 6.81910872e-01
7.52596080e-01 -5.22511862e-02 1.02247465e+00 -1.11497092e+00
-4.78374287e-02 -2.32901499e-01 -7.20388830e-01 -1.45864809e+00
2.33575761e-01 -3.63566488e-01 6.82852149e-01 -1.35876656e+00
-2.20997423e-01 1.95845470e-01 -6.39369413e-02 4.72557694e-01
3.28131616e-01 3.48964483e-01 5.44190168e-01 2.54483491e-01
-8.67392957e-01 5.29510617e-01 8.06628227e-01 -2.82304049e-01
-2.65424669e-01 2.06671953e-01 1.48649752e-01 1.04464602e+00
5.85362613e-01 -1.83383927e-01 4.43754978e-02 -5.13383746e-01
-1.98837250e-01 5.09054899e-01 9.63515520e-01 -1.43351042e+00
9.79802191e-01 -3.16386729e-01 5.82651556e-01 -1.04604805e+00
8.50228131e-01 -1.38105536e+00 9.94935557e-02 4.86990690e-01
-1.83267906e-01 1.05980925e-01 1.13301329e-01 8.36257339e-01
-7.02282488e-02 5.06553411e-01 4.04044509e-01 3.15237284e-01
-1.00883400e+00 1.81057081e-01 -7.12451458e-01 -6.93928301e-01
9.95267391e-01 -4.89998490e-01 -2.38893181e-01 -4.27870870e-01
-9.07980680e-01 5.33250630e-01 2.07207441e-01 7.88932502e-01
9.55529869e-01 -1.48293853e+00 -4.37206894e-01 6.19665027e-01
3.32151316e-02 -1.36304408e-01 6.87723607e-02 1.02550626e+00
-3.28518301e-01 5.82287014e-01 -3.84556353e-01 -9.19517398e-01
-1.26480699e+00 4.16092306e-01 4.09457833e-01 1.48299992e-01
-8.52859676e-01 3.43504339e-01 -2.37118825e-01 -5.33504188e-01
6.02091014e-01 -5.16109765e-01 -3.17818284e-01 -1.13855466e-01
5.11555612e-01 4.76810157e-01 -9.51337069e-02 -1.21815968e+00
-5.78329146e-01 6.90683782e-01 4.17082548e-01 -8.00500929e-01
1.14883626e+00 -2.08750799e-01 3.76611501e-01 5.06472409e-01
1.04487681e+00 -4.51467276e-01 -1.81501305e+00 -3.39071631e-01
-1.21479951e-01 -1.44973561e-01 1.27046788e-02 -5.63663661e-01
-1.00342381e+00 8.03099334e-01 7.08946407e-01 -2.29034841e-01
1.03936839e+00 -7.03253672e-02 7.67867982e-01 1.02094281e+00
4.37262446e-01 -1.00497770e+00 -3.02971780e-01 9.73506868e-01
7.94181406e-01 -1.37804186e+00 -1.30533203e-01 2.83574332e-02
-6.64878011e-01 1.12415922e+00 4.50001687e-01 -2.22989932e-01
8.38347912e-01 5.88098705e-01 2.61442661e-01 -1.73824251e-01
-7.66192794e-01 -6.69373751e-01 4.66296732e-01 7.00821757e-01
-1.66026071e-01 8.14002752e-02 2.98373103e-01 3.60140115e-01
-4.89587262e-02 1.45787612e-01 -5.19481488e-02 1.26296484e+00
-5.19652426e-01 -6.49273813e-01 -7.15343535e-01 1.21567547e-01
2.29964316e-01 3.62506241e-01 -6.25516415e-01 8.67026865e-01
2.07451619e-02 1.00725269e+00 3.92593861e-01 -8.34482849e-01
6.09491110e-01 1.76467389e-01 4.53274459e-01 -9.07962844e-02
-2.95209199e-01 5.19322455e-02 1.37727425e-01 -1.15532148e+00
-6.84684813e-01 -8.04451466e-01 -1.42932844e+00 -1.29836902e-01
1.82550251e-01 -4.40178037e-01 6.38060093e-01 1.27735722e+00
1.63994253e-01 3.46475691e-01 3.50357860e-01 -1.51012540e+00
-4.26119089e-01 -8.33393693e-01 -2.52942622e-01 1.28887504e-01
8.94173443e-01 -7.68772840e-01 -1.38398603e-01 3.07965785e-01] | [7.259193420410156, -0.15876542031764984] |
0464bc6b-6a75-4120-ae6a-70368ec9f9b6 | less-is-more-surgical-phase-recognition-from | 2202.08199 | null | https://arxiv.org/abs/2202.08199v2 | https://arxiv.org/pdf/2202.08199v2.pdf | Less is More: Surgical Phase Recognition from Timestamp Supervision | Surgical phase recognition is a fundamental task in computer-assisted surgery systems. Most existing works are under the supervision of expensive and time-consuming full annotations, which require the surgeons to repeat watching videos to find the precise start and end time for a surgical phase. In this paper, we introduce timestamp supervision for surgical phase recognition to train the models with timestamp annotations, where the surgeons are asked to identify only a single timestamp within the temporal boundary of a phase. This annotation can significantly reduce the manual annotation cost compared to the full annotations. To make full use of such timestamp supervisions, we propose a novel method called uncertainty-aware temporal diffusion (UATD) to generate trustworthy pseudo labels for training. Our proposed UATD is motivated by the property of surgical videos, i.e., the phases are long events consisting of consecutive frames. To be specific, UATD diffuses the single labelled timestamp to its corresponding high confident ( i.e., low uncertainty) neighbour frames in an iterative way. Our study uncovers unique insights of surgical phase recognition with timestamp supervisions: 1) timestamp annotation can reduce 74% annotation time compared with the full annotation, and surgeons tend to annotate those timestamps near the middle of phases; 2) extensive experiments demonstrate that our method can achieve competitive results compared with full supervision methods, while reducing manual annotation cost; 3) less is more in surgical phase recognition, i.e., less but discriminative pseudo labels outperform full but containing ambiguous frames; 4) the proposed UATD can be used as a plug and play method to clean ambiguous labels near boundaries between phases, and improve the performance of the current surgical phase recognition methods. | ['Xiaowei Xu', 'Jian Zhuang', 'Xinjian Yan', 'Xinpeng Ding', 'Xiaomeng Li', 'Wei Zhao', 'Zixun Wang'] | 2022-02-16 | null | null | null | null | ['surgical-phase-recognition'] | ['computer-vision'] | [ 4.05269057e-01 2.98135638e-01 -6.74502909e-01 -3.51630419e-01
-9.43055391e-01 -6.22639716e-01 1.40850827e-01 3.40651840e-01
-4.87013817e-01 4.76761103e-01 2.38297805e-01 -2.97048867e-01
-3.24281931e-01 -2.87448496e-01 -6.05100572e-01 -1.10352993e+00
-1.60154730e-01 3.95749897e-01 3.46598804e-01 2.78044432e-01
1.41868755e-01 1.45586342e-01 -1.02260709e+00 1.62600487e-01
7.41293728e-01 1.16987956e+00 1.96074188e-01 2.69106984e-01
1.22461937e-01 7.35635161e-01 -4.79326665e-01 6.74137175e-02
4.32810813e-01 -3.51277381e-01 -6.40845180e-01 1.05197005e-01
-1.16186932e-01 -3.75317335e-01 -3.70773911e-01 9.87752080e-01
3.42446893e-01 -1.07424743e-01 3.01014930e-01 -1.09824216e+00
1.36980116e-02 8.27767909e-01 -5.57452679e-01 1.71729773e-01
1.84936628e-01 1.18442282e-01 6.34125054e-01 -4.50658083e-01
8.27856243e-01 6.48128986e-01 7.32027352e-01 5.53855360e-01
-8.90591502e-01 -9.65351522e-01 3.24959725e-01 2.91793831e-02
-1.42305839e+00 -1.99752733e-01 7.47339666e-01 -5.56442440e-01
7.99839050e-02 1.53629810e-01 7.94813633e-01 1.12933326e+00
4.13824111e-01 8.51804614e-01 1.01688600e+00 -1.71717808e-01
2.90206730e-01 -1.02776371e-01 -9.72376615e-02 1.03665161e+00
1.45410389e-01 1.89372495e-01 -6.20963693e-01 -8.00736621e-02
8.35123301e-01 4.28829014e-01 -5.19489288e-01 -2.66028613e-01
-1.83404672e+00 5.89848459e-01 4.74272460e-01 3.08283746e-01
-2.12270066e-01 1.03469506e-01 6.13612413e-01 -8.46198499e-02
2.24425316e-01 4.14828897e-01 -1.70271367e-01 -4.21293288e-01
-1.07231212e+00 -4.25800651e-01 7.43022978e-01 1.17825341e+00
6.29124522e-01 -4.22702521e-01 -3.89954299e-01 3.64756882e-01
9.09032896e-02 1.13200068e-01 7.59663105e-01 -1.15640128e+00
7.91839659e-02 5.32333434e-01 1.05877981e-01 -7.08269000e-01
-4.68379110e-01 -3.56386453e-01 -9.82287765e-01 -2.37177283e-01
3.79233450e-01 -6.43886998e-02 -1.04484046e+00 1.58168459e+00
3.10470074e-01 7.07940578e-01 -1.27162859e-01 9.50875103e-01
5.69611967e-01 2.78485328e-01 -2.52525866e-01 -7.91437209e-01
1.49098396e+00 -8.95761967e-01 -1.01403248e+00 -1.57976881e-01
9.20774341e-01 -7.16580331e-01 6.69033647e-01 4.53720629e-01
-6.02762520e-01 -2.42519408e-01 -1.07490563e+00 3.34056318e-01
3.63556892e-02 3.02888334e-01 7.73118496e-01 3.26796889e-01
-6.30100548e-01 6.89165831e-01 -1.37093282e+00 1.85320169e-01
4.24614936e-01 5.26029289e-01 -4.92451102e-01 -2.24331394e-01
-1.04412341e+00 4.63921934e-01 4.76593107e-01 4.15585905e-01
-1.10738575e+00 -7.35092580e-01 -1.06877542e+00 -3.60313863e-01
9.26762521e-01 -4.27070171e-01 1.36337626e+00 -8.16554189e-01
-1.26015568e+00 7.86417842e-01 -3.57362956e-01 -4.41462606e-01
5.92299938e-01 -2.10810378e-01 -7.50860497e-02 3.82801384e-01
1.83751091e-01 5.36246896e-01 7.99983084e-01 -1.17335343e+00
-6.35528326e-01 -1.92094028e-01 -5.29255196e-02 7.86772445e-02
-4.25842106e-01 -4.85836118e-01 -8.00552964e-01 -8.61087143e-01
4.73529339e-01 -1.37818766e+00 -4.09887254e-01 3.16609770e-01
-4.47503924e-01 -6.15418032e-02 5.90008736e-01 -5.49439371e-01
1.57980359e+00 -2.35603166e+00 1.58382282e-02 3.18909049e-01
4.36241955e-01 1.47951469e-02 4.92392510e-01 1.34429276e-01
-8.82038288e-03 9.75368023e-02 -2.29040727e-01 -4.50969130e-01
-3.77722770e-01 7.05376208e-01 -1.79330915e-01 6.06736660e-01
-2.48914719e-01 4.60969478e-01 -1.42125130e+00 -1.00632739e+00
8.03768635e-02 1.45706102e-01 -3.91351044e-01 2.36329988e-01
1.26361558e-02 8.75723302e-01 -5.05112112e-01 7.29904830e-01
2.67163485e-01 -4.17195797e-01 2.29165971e-01 -4.63413298e-01
-8.68339390e-02 1.40168130e-01 -1.04382920e+00 2.16368508e+00
-4.47339952e-01 4.30888116e-01 -1.58603355e-01 -7.27493584e-01
7.08975136e-01 6.05406821e-01 1.05369818e+00 -2.38073632e-01
2.49989167e-01 4.95366067e-01 -1.11015297e-01 -4.47886735e-01
1.84960112e-01 -2.25384664e-02 -1.32042527e-01 2.78270423e-01
-1.86068788e-01 9.17188078e-02 1.67935342e-01 2.74304748e-01
1.22622001e+00 8.37419853e-02 4.23040122e-01 3.34578678e-02
3.64096880e-01 2.28924938e-02 1.15253437e+00 6.65625513e-01
-5.53603530e-01 6.70230210e-01 6.31556869e-01 -4.36962664e-01
-4.64386076e-01 -8.41721237e-01 -8.06065202e-02 4.02287304e-01
6.78693056e-01 -6.07303441e-01 -5.40187359e-01 -1.12999344e+00
-2.34759569e-01 1.87962040e-01 -8.45590234e-01 -3.48101407e-01
-8.19170535e-01 -2.68149078e-01 2.55204082e-01 6.14402592e-01
1.15097441e-01 -7.12248206e-01 -6.49051845e-01 1.53357506e-01
-4.99042302e-01 -1.26203024e+00 -8.88847411e-01 3.20332944e-01
-1.07286143e+00 -1.22524440e+00 -7.66010344e-01 -8.24312508e-01
1.03499794e+00 3.59399140e-01 6.80240035e-01 5.05797304e-02
-3.15662056e-01 2.01887086e-01 -5.45043290e-01 -3.97231191e-01
-2.90653318e-01 -9.54274833e-02 1.11143641e-01 6.18286356e-02
1.01152845e-01 -3.12182724e-01 -1.04451191e+00 7.19167709e-01
-9.57204163e-01 2.55509824e-01 7.52524734e-01 1.13472021e+00
1.01755190e+00 -5.98448068e-02 1.51433796e-01 -1.07629871e+00
5.95965609e-02 -4.12255168e-01 -5.28147399e-01 1.35088995e-01
-8.37067723e-01 9.30057690e-02 4.82355088e-01 -6.31783724e-01
-6.76924467e-01 4.11141545e-01 2.17065707e-01 -9.26530600e-01
1.09860957e-01 5.01288056e-01 4.90223706e-01 -2.81835645e-02
4.20860827e-01 1.37531951e-01 2.38885313e-01 -2.02620462e-01
-5.12410402e-02 4.74803090e-01 6.41107142e-01 -4.38117146e-01
6.23216629e-01 7.44136870e-01 1.77619141e-02 -1.83522537e-01
-1.17132008e+00 -8.63326609e-01 -4.97019351e-01 -3.71927112e-01
6.87199295e-01 -9.00419891e-01 -8.43007803e-01 2.38633662e-01
-8.46450150e-01 -3.50857079e-02 -3.15053940e-01 8.31204116e-01
-5.53588390e-01 6.73693120e-01 -7.38365054e-01 -5.31974614e-01
-2.98803151e-01 -1.46053755e+00 1.29464817e+00 3.95952195e-01
-2.46124953e-01 -8.48533869e-01 -3.85789573e-02 3.46879274e-01
-9.76746976e-02 4.96258825e-01 4.94102269e-01 -7.19640851e-01
-7.51394629e-01 -4.99483377e-01 6.45134673e-02 1.46727666e-01
3.91423374e-01 -1.35247678e-01 -6.54234767e-01 -3.11275482e-01
1.42316610e-01 7.67156556e-02 5.94067931e-01 6.20208561e-01
1.42530131e+00 -3.26734185e-01 -7.94086337e-01 7.58826554e-01
1.09353852e+00 3.30684006e-01 3.43762398e-01 4.09098029e-01
5.39204955e-01 3.80682409e-01 1.36406910e+00 7.23630369e-01
2.68395334e-01 5.79411387e-01 4.82171893e-01 -8.23727399e-02
1.00696780e-01 -2.14278758e-01 2.97371835e-01 1.06210136e+00
3.21002118e-02 8.44302773e-03 -9.34334755e-01 6.31067753e-01
-2.11270905e+00 -5.40892541e-01 2.54251927e-01 2.46331310e+00
1.25235724e+00 2.87170082e-01 -2.35454783e-01 1.96403578e-01
6.38227642e-01 2.20370322e-01 -4.54927713e-01 3.16124201e-01
5.05084574e-01 -4.59805429e-02 7.67613471e-01 2.47578129e-01
-1.13569224e+00 5.49386978e-01 5.17314720e+00 9.75706100e-01
-1.14288390e+00 1.52969241e-01 5.28414369e-01 -2.93438643e-01
7.42423013e-02 1.20746270e-01 -7.27638900e-01 7.17044652e-01
6.87253118e-01 -4.84159328e-02 1.55703817e-02 8.97138357e-01
3.19161057e-01 -2.73572773e-01 -1.38721251e+00 1.05760300e+00
2.25943569e-02 -1.39865863e+00 -2.70221144e-01 7.33542815e-02
5.94946444e-01 -3.47791553e-01 -1.36863291e-01 -6.83829980e-03
9.63183865e-02 -8.55558217e-01 5.50545216e-01 5.58903515e-01
8.62102985e-01 -3.90133351e-01 1.10419798e+00 4.31910545e-01
-1.24980342e+00 -1.19121246e-01 -2.85543874e-02 5.31640947e-01
3.16577733e-01 6.99744284e-01 -1.19605577e+00 8.05703402e-01
6.06516063e-01 1.07929504e+00 -2.05227479e-01 1.25041354e+00
-2.58334458e-01 5.68476915e-01 -4.89165336e-01 2.56147742e-01
1.63554996e-01 -5.47977835e-02 5.32320678e-01 1.02495480e+00
5.82829237e-01 1.81848139e-01 5.22644699e-01 2.10538864e-01
1.02508180e-01 -1.73710525e-01 -3.87759000e-01 -4.75228950e-02
5.95553458e-01 1.21185935e+00 -9.34659481e-01 -2.65567213e-01
-1.68333396e-01 8.03833485e-01 -2.09048852e-01 4.18038368e-02
-1.03775299e+00 -2.44963646e-01 2.66154826e-01 2.75309514e-02
4.76567782e-02 -2.53515452e-01 -3.45713556e-01 -1.03622937e+00
2.88152933e-01 -4.70014006e-01 7.25807548e-01 -5.54512680e-01
-8.01141798e-01 4.83987302e-01 -9.05605257e-02 -2.10596275e+00
-2.06790894e-01 -3.10242474e-01 -3.36105049e-01 2.58412957e-01
-1.52470887e+00 -9.38457072e-01 -7.27937102e-01 6.57750487e-01
5.38671076e-01 3.63941580e-01 7.38737524e-01 3.37005079e-01
-4.51700419e-01 5.29643655e-01 -1.49442464e-01 2.64125526e-01
1.19403052e+00 -1.19898665e+00 -4.78679746e-01 7.15485096e-01
-3.54749919e-03 7.16893733e-01 5.30587375e-01 -7.58601487e-01
-1.31349659e+00 -8.87518764e-01 5.89715421e-01 -2.73169994e-01
6.14613354e-01 9.24789067e-03 -6.41280055e-01 8.27995896e-01
-2.89406151e-01 1.43441916e-01 9.32092309e-01 -1.78953826e-01
-9.47971940e-02 -4.63196903e-01 -8.13626885e-01 4.98504460e-01
9.44566250e-01 -2.98752606e-01 -5.66938102e-01 6.22517705e-01
8.83623719e-01 -1.00114346e+00 -1.19789863e+00 6.51845753e-01
5.46661675e-01 -6.96458876e-01 7.59149194e-01 1.36801019e-01
3.94374847e-01 -5.40040791e-01 3.92565995e-01 -9.36406493e-01
1.19402438e-01 -8.89324129e-01 -2.33758777e-01 1.00083625e+00
1.70587033e-01 -4.87027973e-01 8.02546561e-01 5.13087928e-01
-5.96441925e-01 -9.90380108e-01 -1.22016788e+00 -6.99834645e-01
-6.17371440e-01 -2.45108575e-01 1.07999377e-01 9.42858636e-01
1.79660052e-01 -4.09207940e-01 -3.81414473e-01 3.27486843e-01
4.33436960e-01 1.49570763e-01 3.49508435e-01 -1.12599921e+00
-1.41222432e-01 -1.61132351e-01 -3.75078976e-01 -9.98950362e-01
-2.28498369e-01 -7.13493407e-01 6.21036232e-01 -1.34331310e+00
1.11351073e-01 -7.43253410e-01 -6.51296496e-01 6.94895208e-01
-2.09168985e-01 8.53845254e-02 9.49934497e-03 6.83605254e-01
-7.15232909e-01 2.46917307e-01 1.41092038e+00 -1.27209857e-01
-1.15111746e-01 3.38137805e-01 -3.75931144e-01 8.14928114e-01
3.33584785e-01 -8.79152954e-01 -4.62302059e-01 -1.72495097e-01
1.44802243e-01 4.09199625e-01 7.38435835e-02 -1.00163507e+00
6.45985186e-01 -9.66861323e-02 3.54054086e-02 -4.33784604e-01
2.60156512e-01 -1.15839267e+00 2.66527772e-01 8.24066699e-01
-1.69639289e-01 -2.02293560e-01 2.52771340e-02 9.69705403e-01
-5.36784053e-01 -3.21839750e-01 5.63811719e-01 -1.80016115e-01
-7.18739629e-01 5.39624631e-01 -1.73015520e-01 -7.67310858e-02
1.33438361e+00 -3.65315378e-01 -1.60569146e-01 -1.66003481e-01
-9.05665517e-01 4.15698230e-01 4.35304999e-01 3.98896992e-01
5.70756376e-01 -1.15159154e+00 -2.50290364e-01 -7.36320838e-02
1.89850450e-01 6.34082735e-01 3.93279016e-01 1.50364995e+00
-5.57000160e-01 2.93177634e-01 2.14273304e-01 -1.10013676e+00
-1.28779745e+00 7.06477642e-01 8.26490670e-02 -5.12592077e-01
-8.80629599e-01 7.95215964e-01 4.52925622e-01 5.99525124e-02
4.93900299e-01 -5.33240855e-01 -1.51068136e-01 1.30516768e-01
3.73200446e-01 -2.99233627e-02 -2.30067479e-03 -1.17469057e-01
-4.17220920e-01 7.00881660e-01 -3.26264828e-01 1.56158715e-01
9.85796511e-01 -2.66286302e-02 1.17259741e-01 5.31866968e-01
1.02707577e+00 2.93017346e-02 -1.42982864e+00 -3.67392272e-01
4.53873575e-02 -5.27605593e-01 -1.50595866e-02 -6.45870566e-01
-1.21624351e+00 5.44617534e-01 6.48464262e-01 -1.38848454e-01
1.24173570e+00 -7.66211972e-02 1.10812271e+00 9.69847739e-02
5.04452467e-01 -1.00580895e+00 5.79514876e-02 9.78170484e-02
4.25227553e-01 -1.33566582e+00 6.30855635e-02 -6.59339547e-01
-8.25258255e-01 1.16109359e+00 4.54458803e-01 2.12090999e-01
6.30485594e-01 3.51135701e-01 1.81998983e-01 -2.10369274e-01
-5.64995766e-01 1.42741516e-01 2.46928513e-01 1.80554137e-01
2.06092447e-01 7.36250207e-02 -4.11079347e-01 6.22608840e-01
4.91748415e-02 4.19981450e-01 5.13037086e-01 1.19684315e+00
-1.43256202e-01 -8.57914329e-01 -1.35560602e-01 3.67129296e-01
-5.43523669e-01 -2.98413169e-02 2.85610110e-01 5.87814033e-01
1.05837658e-01 7.95880020e-01 -1.67811543e-01 -3.31452072e-01
1.38907716e-01 -1.40098736e-01 1.96381602e-02 -6.82152152e-01
-5.58888435e-01 3.55313748e-01 -2.66149603e-02 -9.62389767e-01
-5.49412489e-01 -4.48194534e-01 -1.53844905e+00 1.81006208e-01
-5.60635030e-01 3.76466304e-01 5.76178432e-01 9.25014377e-01
3.51291001e-01 8.38528574e-01 7.52788663e-01 -7.81171262e-01
-4.50009167e-01 -5.91546297e-01 -3.77731115e-01 3.27247888e-01
6.22667849e-01 -8.79320025e-01 -6.57215178e-01 3.38578850e-01] | [14.126240730285645, -3.2573118209838867] |
c389c000-5ec0-46d0-a714-1915e2a1abd3 | a-comprehensive-survey-on-pretrained | 2302.09419 | null | https://arxiv.org/abs/2302.09419v3 | https://arxiv.org/pdf/2302.09419v3.pdf | A Comprehensive Survey on Pretrained Foundation Models: A History from BERT to ChatGPT | Pretrained Foundation Models (PFMs) are regarded as the foundation for various downstream tasks with different data modalities. A PFM (e.g., BERT, ChatGPT, and GPT-4) is trained on large-scale data which provides a reasonable parameter initialization for a wide range of downstream applications. BERT learns bidirectional encoder representations from Transformers, which are trained on large datasets as contextual language models. Similarly, the generative pretrained transformer (GPT) method employs Transformers as the feature extractor and is trained using an autoregressive paradigm on large datasets. Recently, ChatGPT shows promising success on large language models, which applies an autoregressive language model with zero shot or few shot prompting. The remarkable achievements of PFM have brought significant breakthroughs to various fields of AI. Numerous studies have proposed different methods, raising the demand for an updated survey. This study provides a comprehensive review of recent research advancements, challenges, and opportunities for PFMs in text, image, graph, as well as other data modalities. The review covers the basic components and existing pretraining methods used in natural language processing, computer vision, and graph learning. Additionally, it explores advanced PFMs used for different data modalities and unified PFMs that consider data quality and quantity. The review also discusses research related to the fundamentals of PFMs, such as model efficiency and compression, security, and privacy. Finally, the study provides key implications, future research directions, challenges, and open problems in the field of PFMs. Overall, this survey aims to shed light on the research of the PFMs on scalability, security, logical reasoning ability, cross-domain learning ability, and the user-friendly interactive ability for artificial general intelligence. | ['Lichao Sun', 'Philip S. Yu', 'Jian Pei', 'Caiming Xiong', 'Pengtao Xie', 'Ziwei Liu', 'Jia Wu', 'JianXin Li', 'Hao Peng', 'Lifang He', 'Qiben Yan', 'Cheng Ji', 'Kai Zhang', 'Guangjing Wang', 'Yixin Liu', 'Jun Yu', 'Chen Li', 'Qian Li', 'Ce Zhou'] | 2023-02-18 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [ 3.30604911e-01 1.25361338e-01 -5.89849889e-01 -2.41858289e-01
-2.66292870e-01 -2.50094175e-01 5.79022288e-01 2.37642210e-02
-1.40386447e-01 2.90672958e-01 2.52756894e-01 -3.42483729e-01
-1.40157312e-01 -1.25101197e+00 -6.52008295e-01 -6.83222294e-01
-4.21434045e-02 5.24642527e-01 -1.03355564e-01 -1.55998379e-01
1.74365476e-01 3.02795440e-01 -1.26100349e+00 5.21841526e-01
6.70378923e-01 1.33726752e+00 2.47015879e-01 5.64776480e-01
-2.65545636e-01 1.31919670e+00 -4.92794603e-01 -1.12327600e+00
7.17758462e-02 -1.01099335e-01 -7.16831207e-01 8.79685059e-02
1.62407011e-01 -3.36139232e-01 -8.22788298e-01 9.43711042e-01
4.59117234e-01 7.79551128e-03 7.15666592e-01 -1.76370919e+00
-1.44306958e+00 8.53493333e-01 -4.49547946e-01 -8.49888101e-02
4.03385907e-01 4.00552973e-02 1.16805935e+00 -1.01679766e+00
4.69469517e-01 1.54667413e+00 6.04010224e-01 5.58123648e-01
-7.35087872e-01 -6.49890661e-01 2.26066530e-01 7.25190639e-01
-1.33047307e+00 -3.51786226e-01 6.39919758e-01 -1.36859238e-01
1.28020382e+00 -1.20314071e-02 6.06007278e-01 1.41079593e+00
4.90071267e-01 1.23879862e+00 9.65732276e-01 -5.77317834e-01
6.41335323e-02 1.24791026e-01 5.71526699e-02 1.11708140e+00
7.04300404e-02 -3.88440676e-02 -1.12091279e+00 -2.17570662e-01
8.30379546e-01 1.41694888e-01 -6.21568188e-02 -3.33461881e-01
-1.01880932e+00 1.10347199e+00 4.24489170e-01 1.48356840e-01
-1.77388728e-01 -1.86582480e-03 5.61951876e-01 5.96054912e-01
2.96726733e-01 -3.35758477e-02 -2.99834669e-01 1.16906814e-01
-4.56246465e-01 -1.52535692e-01 8.16685736e-01 1.33637142e+00
3.71099800e-01 4.63859826e-01 -2.43199542e-01 8.73916626e-01
4.78005975e-01 7.07368016e-01 4.61638004e-01 -6.12800181e-01
5.78811228e-01 6.29939795e-01 -8.30027342e-01 -1.17832732e+00
-6.05257191e-02 -1.71146438e-01 -1.34555721e+00 -2.68184274e-01
-2.49666646e-01 1.49529636e-01 -1.09200859e+00 1.57329655e+00
2.39146799e-02 2.73258299e-01 3.61213297e-01 3.94109786e-01
1.56729090e+00 8.19227040e-01 2.79919505e-01 -1.11087866e-01
1.46610773e+00 -1.08324921e+00 -8.77453148e-01 -4.30628777e-01
4.55169111e-01 -5.58858693e-01 1.22231317e+00 4.93011117e-01
-9.60345745e-01 -4.31003690e-01 -8.19341183e-01 -4.64188159e-01
-6.78026199e-01 2.10590407e-01 9.78086829e-01 6.98801696e-01
-1.18563163e+00 2.41535977e-01 -7.76020646e-01 -5.66717625e-01
7.19623506e-01 3.04722637e-01 -3.48073512e-01 -5.06774902e-01
-1.51303101e+00 8.63974392e-01 4.25974756e-01 -7.11059123e-02
-1.22245717e+00 -3.05276871e-01 -1.09349704e+00 1.29515916e-01
4.75302428e-01 -1.01855063e+00 7.49578714e-01 -5.87128401e-01
-1.64798307e+00 9.40047681e-01 -9.13366750e-02 -8.38948548e-01
1.07569575e-01 -1.39138564e-01 -5.38086534e-01 2.16531768e-01
-1.53391570e-01 5.59766412e-01 1.23097610e+00 -7.19177783e-01
-3.24389458e-01 -2.80423582e-01 1.65338472e-01 2.77108073e-01
-7.38026738e-01 2.75451362e-01 -6.79853797e-01 -5.94995499e-01
-3.19985628e-01 -5.03844142e-01 -5.92132770e-02 -2.37539820e-02
-2.85398036e-01 -4.66386110e-01 9.65715051e-01 -5.02463698e-01
1.40725195e+00 -2.38889384e+00 2.11427271e-01 1.14883497e-01
3.98066908e-01 1.82061136e-01 -3.53307009e-01 7.92116106e-01
3.73549722e-02 1.37020275e-01 1.54335694e-02 -2.65393347e-01
-1.15281492e-02 4.54998404e-01 -8.21818829e-01 1.99759796e-01
6.89330399e-02 1.38068724e+00 -6.57434583e-01 -6.36514544e-01
3.42423618e-01 5.45498967e-01 -3.77376109e-01 3.15112114e-01
-3.53643037e-02 4.91798334e-02 -5.65463305e-01 8.86460066e-01
4.12267923e-01 -6.73997819e-01 3.95946950e-02 -3.12047988e-01
4.59988296e-01 3.82805010e-03 -8.17188680e-01 1.63284516e+00
-2.30495557e-01 4.29328948e-01 -1.08137704e-01 -1.41684806e+00
1.03796351e+00 4.19394583e-01 3.20276648e-01 -5.86991608e-01
2.62033224e-01 -3.52501944e-02 -2.87407577e-01 -5.86170316e-01
4.97988641e-01 -8.40270966e-02 -1.46482199e-01 2.43216634e-01
4.43152726e-01 -4.03399676e-01 1.81494817e-01 5.91262579e-01
9.88156259e-01 -2.14615792e-01 5.23309886e-01 2.59858668e-01
6.08252168e-01 -2.04353049e-01 2.58167893e-01 7.29730248e-01
-8.58376473e-02 2.10467324e-01 4.34196740e-01 -4.37639028e-01
-5.93547046e-01 -1.08183300e+00 2.04317033e-01 1.34770250e+00
1.52950019e-01 -1.14553559e+00 -3.96451980e-01 -5.41302800e-01
-4.43226807e-02 5.35297990e-01 -4.61109191e-01 -6.48803055e-01
-1.61380231e-01 -8.29542339e-01 7.94985533e-01 8.33239496e-01
1.00186467e+00 -1.19506001e+00 -1.25735551e-01 -1.87394157e-01
-3.42737168e-01 -1.48866200e+00 1.09192103e-01 -1.11840412e-01
-9.20308590e-01 -9.16745424e-01 -3.37402999e-01 -9.23580289e-01
6.20050609e-01 2.28781506e-01 1.21102047e+00 9.29834396e-02
-8.54352862e-02 9.01869833e-01 -4.79755014e-01 -6.92220032e-01
-1.96020931e-01 -1.63206279e-01 1.09824046e-01 -1.21764667e-01
5.55998087e-01 -4.13949639e-01 -2.24977303e-02 1.20003417e-01
-8.48986566e-01 6.30543008e-02 7.12238550e-01 1.20146537e+00
6.55367076e-01 3.02206665e-01 3.17032337e-01 -8.80504489e-01
1.01858616e+00 -3.83548349e-01 -3.80345225e-01 7.59041607e-01
-7.62705564e-01 -8.86802673e-02 6.96000993e-01 -4.46829259e-01
-1.14878917e+00 -3.83983791e-01 -9.76404399e-02 -6.51704609e-01
1.83892697e-01 8.89379025e-01 -2.21772641e-01 -2.98215985e-01
5.96369505e-01 4.25876409e-01 -1.57046944e-01 -1.93833873e-01
6.27812684e-01 7.13868916e-01 3.82861704e-01 -8.29218209e-01
6.89992964e-01 4.75206584e-01 -4.22127619e-02 -8.79819214e-01
-7.50366151e-01 -1.23362117e-01 -3.86846870e-01 5.52203320e-03
7.62934864e-01 -1.01212037e+00 -5.79555392e-01 7.09879458e-01
-8.02888036e-01 -2.49680355e-01 -3.85206819e-01 4.55853432e-01
-5.80604792e-01 5.40344119e-01 -1.06487215e+00 -7.68529952e-01
-8.17109048e-01 -1.02547729e+00 1.04731834e+00 2.39356309e-01
1.60255536e-01 -1.20242071e+00 -4.96791869e-01 6.22131705e-01
3.71143222e-01 -2.16180697e-01 1.10975802e+00 -5.55406213e-01
-6.08458161e-01 -1.94079146e-01 -2.17275932e-01 3.98764938e-01
-1.18641451e-01 2.23572887e-02 -9.54141378e-01 -3.45112473e-01
9.87818167e-02 -8.52226257e-01 8.84401858e-01 4.66161102e-01
1.40054607e+00 -2.82967865e-01 -3.30985546e-01 7.80279815e-01
1.14370680e+00 1.55706882e-01 6.39152765e-01 -1.34114558e-02
6.75250709e-01 3.32449764e-01 5.82699478e-01 3.90260726e-01
4.80841368e-01 -7.94560835e-02 3.37630361e-01 5.17384931e-02
-4.93195429e-02 -5.58025479e-01 6.61106646e-01 1.09267914e+00
-3.59513700e-01 -6.53035045e-01 -8.10907006e-01 2.20847696e-01
-1.81189489e+00 -8.98199141e-01 1.88336402e-01 1.91291320e+00
4.56802458e-01 3.65217514e-02 -4.04809833e-01 6.57620281e-02
5.90171456e-01 3.49267095e-01 -6.74048603e-01 -3.69387358e-01
-4.07723427e-01 4.13435131e-01 3.29516053e-01 9.28218812e-02
-9.90820408e-01 1.33088803e+00 7.25747633e+00 1.07282841e+00
-1.21840215e+00 4.70548943e-02 6.47727489e-01 2.35632837e-01
-1.48501202e-01 -4.05038297e-02 -7.35051751e-01 -7.87319019e-02
1.05801547e+00 -5.67048252e-01 7.75211275e-01 1.05597270e+00
-3.38827997e-01 1.23374328e-01 -1.08566296e+00 1.46474111e+00
5.25605440e-01 -1.47260129e+00 4.75970387e-01 -1.10627403e-02
4.79954213e-01 3.58196795e-01 4.90646720e-01 5.86999178e-01
5.29126585e-01 -1.14919364e+00 5.26899517e-01 2.47128740e-01
1.09655249e+00 -4.64605629e-01 6.03794932e-01 3.71473640e-01
-1.52128720e+00 -3.89552504e-01 -7.42403686e-01 -1.30427122e-01
3.05986926e-02 3.81009638e-01 -6.23498201e-01 8.32176030e-01
8.25733423e-01 1.19146919e+00 -4.92178589e-01 5.47519624e-01
-5.57838857e-01 5.41986942e-01 -2.92188853e-01 7.57413823e-03
-9.86242592e-02 -2.29944512e-01 1.55073822e-01 9.72160697e-01
2.59098887e-01 9.42131355e-02 3.12781662e-01 5.90903640e-01
-2.94115365e-01 1.86066374e-01 -8.53735566e-01 -5.88132918e-01
5.79222620e-01 1.13823998e+00 -4.47732151e-01 -6.02245510e-01
-7.23108470e-01 7.43025661e-01 3.89108539e-01 4.67919707e-01
-7.30068147e-01 -3.95343564e-02 1.90861627e-01 -3.73876810e-01
-6.60235137e-02 -1.97492093e-01 -2.19568491e-01 -1.65433812e+00
-2.70122498e-01 -1.08612680e+00 1.25723362e+00 -9.67972875e-01
-1.75301051e+00 5.92839539e-01 2.58596957e-01 -1.06433511e+00
-1.89645007e-01 -9.04049695e-01 -3.84976089e-01 5.97450793e-01
-1.46653700e+00 -1.62276411e+00 -2.56655782e-01 1.26102388e+00
5.25200129e-01 -5.75661063e-01 1.07263839e+00 8.39153975e-02
-5.47515631e-01 6.00661755e-01 -2.21165761e-01 3.93261254e-01
6.55845881e-01 -7.19696462e-01 2.65727699e-01 8.42773974e-01
4.01706815e-01 8.42972159e-01 1.52672127e-01 -6.64721668e-01
-1.91089177e+00 -1.02785480e+00 7.90911078e-01 -2.31388524e-01
7.41935730e-01 -4.03159469e-01 -7.62663245e-01 1.11410570e+00
3.82750452e-01 8.13475773e-02 7.85619259e-01 6.98165596e-02
-4.99691039e-01 -2.59937465e-01 -1.18761134e+00 5.52541673e-01
1.18420923e+00 -8.99625838e-01 -4.78386551e-01 5.30485034e-01
6.60659969e-01 -5.63067138e-01 -9.42245305e-01 3.33394200e-01
4.07688677e-01 -6.88252389e-01 1.14498019e+00 -6.26616299e-01
3.72706890e-01 2.53500849e-01 -2.28321880e-01 -9.87248302e-01
-4.82809573e-01 -6.36848390e-01 -6.07207596e-01 1.11994326e+00
-3.15597691e-02 -8.30573082e-01 6.89432085e-01 7.45487988e-01
-2.75933053e-02 -7.62299657e-01 -7.67559707e-01 -6.71816707e-01
-9.61875692e-02 -7.78309941e-01 6.27864003e-01 8.89625132e-01
1.11083664e-01 6.08979940e-01 -6.44821942e-01 -2.34285314e-02
5.18684566e-01 1.39539391e-01 7.86733508e-01 -9.96208131e-01
-1.72224984e-01 -2.33798698e-01 -4.58378613e-01 -1.37056553e+00
2.80579478e-01 -1.10915649e+00 -5.15201628e-01 -1.70070744e+00
2.42057279e-01 -7.14400038e-03 -2.91373253e-01 7.77826846e-01
1.60059094e-01 -5.22988327e-02 2.67338663e-01 1.35615230e-01
-4.02222633e-01 6.70715153e-01 1.34754276e+00 -4.78636920e-01
1.52343944e-01 -1.07548617e-01 -7.70492971e-01 7.82386065e-01
6.51560128e-01 -6.10465668e-02 -1.01623881e+00 -6.54951096e-01
4.24179107e-01 -8.67284834e-02 2.62599230e-01 -5.92280567e-01
4.71929967e-01 -3.62800926e-01 3.49560797e-01 -5.12486100e-01
5.86032987e-01 -1.00763178e+00 -7.51367733e-02 4.55263525e-01
-2.02150166e-01 2.12826699e-01 -5.26466034e-02 6.38706088e-01
-3.61933917e-01 -7.88312405e-02 4.45422769e-01 -4.53265309e-01
-1.18732870e+00 7.31786489e-01 -3.44763488e-01 -5.39333411e-02
9.82894123e-01 -1.11288063e-01 -5.00920177e-01 -6.48014486e-01
-5.84332883e-01 2.79908538e-01 -6.13843836e-02 5.24216175e-01
1.14393103e+00 -1.33358431e+00 -6.02821112e-01 5.71028829e-01
1.80268541e-01 -9.53399390e-02 2.21260160e-01 7.93361008e-01
-8.69895369e-02 4.11754221e-01 -9.92717519e-02 -5.10492861e-01
-1.38704622e+00 1.04108179e+00 3.65390368e-02 -4.73164797e-01
-5.22070825e-01 1.09710443e+00 4.52939570e-01 -2.45472044e-01
4.36772853e-01 -4.32589114e-01 -2.80906707e-01 -1.56562373e-01
5.30102789e-01 1.60933346e-01 -1.29482269e-01 -5.77278674e-01
-3.61455351e-01 4.35429633e-01 -1.79845523e-02 2.30267927e-01
1.38914120e+00 -1.21039942e-01 -3.49675417e-01 3.44096690e-01
8.86947632e-01 -2.58744627e-01 -6.36736572e-01 -5.97622752e-01
-3.89997333e-01 -4.08546478e-01 -4.15834822e-02 -6.86384261e-01
-1.31946397e+00 1.24232197e+00 2.18495622e-01 9.27897394e-02
1.50654185e+00 1.26838788e-01 7.36077130e-01 5.95156372e-01
7.42912829e-01 -1.00233376e+00 3.75009507e-01 8.08876097e-01
1.04239917e+00 -1.06941330e+00 1.28317431e-01 -5.97775698e-01
-9.47573543e-01 1.34874201e+00 6.65850580e-01 1.24919236e-01
8.41998994e-01 4.27023351e-01 -2.75625914e-01 -3.24307531e-01
-9.86379087e-01 8.18298459e-02 2.60130823e-01 9.85793889e-01
2.07299292e-01 -2.86186971e-02 -8.48759711e-02 6.06496811e-01
-3.66183758e-01 3.19803387e-01 1.47450089e-01 1.04533231e+00
-1.43526290e-02 -1.21589971e+00 -3.00813764e-01 6.13481343e-01
-2.69648135e-01 -4.51343268e-01 -5.04631817e-01 4.85020250e-01
-3.19460072e-02 1.05074573e+00 -2.95707971e-01 -6.94247186e-01
1.12234496e-01 -8.48081633e-02 4.62967098e-01 -6.36050224e-01
-3.78066570e-01 -8.13798308e-02 6.98259026e-02 -4.39704418e-01
-5.23046196e-01 -2.86776274e-01 -1.16445708e+00 -5.97937286e-01
-3.16430092e-01 7.50359567e-03 3.15435439e-01 8.00914347e-01
4.59914893e-01 3.21156591e-01 1.75172240e-01 -2.96148181e-01
-4.23946083e-01 -8.40921342e-01 -5.40852249e-01 2.50818640e-01
-8.02438855e-02 -5.64566970e-01 -5.38271479e-02 2.84783155e-01] | [10.535195350646973, 1.7084239721298218] |
5cb3f3e2-62c9-47cc-8d54-013439d8671c | graph-based-blind-image-deblurring-from-a | 1802.07929 | null | http://arxiv.org/abs/1802.07929v1 | http://arxiv.org/pdf/1802.07929v1.pdf | Graph-Based Blind Image Deblurring From a Single Photograph | Blind image deblurring, i.e., deblurring without knowledge of the blur
kernel, is a highly ill-posed problem. The problem can be solved in two parts:
i) estimate a blur kernel from the blurry image, and ii) given estimated blur
kernel, de-convolve blurry input to restore the target image. In this paper, we
propose a graph-based blind image deblurring algorithm by interpreting an image
patch as a signal on a weighted graph. Specifically, we first argue that a
skeleton image---a proxy that retains the strong gradients of the target but
smooths out the details---can be used to accurately estimate the blur kernel
and has a unique bi-modal edge weight distribution. Then, we design a
reweighted graph total variation (RGTV) prior that can efficiently promote a
bi-modal edge weight distribution given a blurry patch. Further, to analyze
RGTV in the graph frequency domain, we introduce a new weight function to
represent RGTV as a graph $l_1$-Laplacian regularizer. This leads to a graph
spectral filtering interpretation of the prior with desirable properties,
including robustness to noise and blur, strong piecewise smooth (PWS) filtering
and sharpness promotion. Minimizing a blind image deblurring objective with
RGTV results in a non-convex non-differentiable optimization problem. We
leverage the new graph spectral interpretation for RGTV to design an efficient
algorithm that solves for the skeleton image and the blur kernel alternately.
Specifically for Gaussian blur, we propose a further speedup strategy for blind
Gaussian deblurring using accelerated graph spectral filtering. Finally, with
the computed blur kernel, recent non-blind image deblurring algorithms can be
applied to restore the target image. Experimental results demonstrate that our
algorithm successfully restores latent sharp images and outperforms
state-of-the-art methods quantitatively and qualitatively. | ['Xian-Ming Liu', 'Yuanchao Bai', 'Wen Gao', 'Gene Cheung'] | 2018-02-22 | null | null | null | null | ['blind-image-deblurring'] | ['computer-vision'] | [ 4.17453736e-01 -2.57018864e-01 1.69202685e-01 1.03409335e-01
-4.99093294e-01 -5.27980804e-01 7.63608590e-02 -7.40347207e-01
1.19740196e-01 6.97697043e-01 5.99968493e-01 -6.02102317e-02
-3.88182133e-01 -3.68504107e-01 -7.39229262e-01 -9.55646038e-01
1.38369575e-01 -3.27758759e-01 -1.44618422e-01 5.76778613e-02
2.40434796e-01 2.90041894e-01 -9.73514557e-01 -2.72307158e-01
1.39476132e+00 7.73960292e-01 4.79230106e-01 8.79928410e-01
2.71367729e-01 9.75526690e-01 -3.19007576e-01 -2.18524352e-01
3.04639131e-01 -7.80188918e-01 -8.33636880e-01 6.46250844e-01
6.73878908e-01 -4.04953599e-01 -8.00882995e-01 1.72794104e+00
3.15970451e-01 2.41539001e-01 7.26340175e-01 -7.50629365e-01
-1.54359460e+00 3.09309810e-01 -1.15146816e+00 2.32641786e-01
2.66163915e-01 1.94487259e-01 6.85361743e-01 -9.79531586e-01
3.86285782e-01 1.06533730e+00 7.70184040e-01 3.94766867e-01
-1.55777586e+00 -2.92452663e-01 -1.42956242e-01 2.78722286e-01
-1.51199162e+00 -5.33375084e-01 1.12860548e+00 -4.99859095e-01
2.42916182e-01 5.99979401e-01 4.29070920e-01 6.84315264e-01
3.07185441e-01 4.93236572e-01 1.32611477e+00 -3.03532451e-01
1.36753052e-01 -4.25447285e-01 2.83360392e-01 9.13849652e-01
4.86171663e-01 5.03293127e-02 -3.43217522e-01 -2.15189770e-01
1.13252747e+00 1.23557627e-01 -1.36130178e+00 -3.53577614e-01
-1.20899558e+00 4.68413681e-01 7.02683389e-01 3.07575196e-01
-4.22235459e-01 3.63634139e-01 -2.12657467e-01 2.23181888e-01
6.89179659e-01 5.12498498e-01 -8.42674300e-02 2.47031271e-01
-1.08785272e+00 -2.85164136e-02 6.31713152e-01 7.59957016e-01
1.06352890e+00 8.33258107e-02 -4.46321636e-01 1.04221070e+00
4.59142208e-01 7.32452452e-01 2.93755591e-01 -1.10939240e+00
9.60394666e-02 1.98333815e-01 4.00920391e-01 -1.01847494e+00
1.02984346e-01 -3.93682748e-01 -1.44358110e+00 8.78743231e-02
4.59092259e-01 -1.04808144e-01 -1.13773453e+00 1.71133590e+00
4.23783883e-02 8.19108903e-01 -9.51083750e-02 1.45637357e+00
4.86670822e-01 5.56969404e-01 -4.79014486e-01 -5.42066932e-01
1.43845403e+00 -9.86371040e-01 -1.01482952e+00 -3.83649766e-01
-2.60541618e-01 -9.11231697e-01 1.00290442e+00 1.44452065e-01
-1.32672894e+00 -4.99299824e-01 -9.09608126e-01 -2.47555524e-01
2.62029767e-01 8.34438875e-02 2.23889768e-01 5.01310587e-01
-1.47052944e+00 4.67610985e-01 -6.58243597e-01 -2.61399835e-01
1.78344354e-01 8.49417783e-03 -9.48213488e-02 -5.84854901e-01
-1.02582240e+00 9.30752039e-01 -1.45866960e-01 3.02687258e-01
-8.23663056e-01 -8.53535175e-01 -9.42056358e-01 4.97311540e-02
2.21076071e-01 -1.14947784e+00 7.91562557e-01 -1.01696861e+00
-1.47723186e+00 6.50163770e-01 -4.35516208e-01 -1.40434802e-01
4.51652139e-01 -2.58935809e-01 -4.09538299e-01 2.73166895e-01
1.12251095e-01 -7.46256113e-02 1.76918173e+00 -1.59438336e+00
-4.68374826e-02 -2.12459877e-01 -3.10405731e-01 2.72671103e-01
-4.33885939e-02 -1.66956991e-01 -6.54255629e-01 -1.25071919e+00
2.78334051e-01 -6.87256992e-01 -1.66658893e-01 -1.71966538e-01
-4.56565917e-01 4.79314357e-01 7.36016631e-01 -1.15286899e+00
1.40414691e+00 -2.26134634e+00 6.07141197e-01 2.77854130e-02
7.30373323e-01 8.66140649e-02 -1.40982002e-01 1.34932756e-01
-3.15938979e-01 -1.69458851e-01 -8.57310236e-01 -1.77676499e-01
-3.33390832e-01 -1.27764267e-03 -4.25000995e-01 9.82582092e-01
-5.32716960e-02 1.09633160e+00 -1.08706737e+00 -1.41061861e-02
2.37330526e-01 6.60229683e-01 -4.22099531e-01 4.89887446e-01
2.28849411e-01 5.10982752e-01 -1.51099876e-01 4.45855975e-01
1.31029832e+00 -5.73299646e-01 -1.69005454e-01 -7.81805575e-01
-4.27365899e-02 -6.20607853e-01 -1.08278298e+00 1.76565909e+00
-2.95405477e-01 6.21073782e-01 7.41876781e-01 -8.78575981e-01
5.07437468e-01 6.32829368e-02 2.99474150e-01 4.92995307e-02
-4.62630130e-02 2.34058663e-01 -4.23922092e-01 -6.70969486e-01
4.45879400e-01 -2.43008256e-01 3.04736227e-01 2.58391619e-01
-2.54499335e-02 -2.98500746e-01 -2.67203182e-01 1.93204105e-01
1.25453711e+00 -2.11738452e-01 3.36149968e-02 -5.93334377e-01
6.60302401e-01 -3.69138777e-01 1.32594004e-01 7.47149169e-01
-1.97028071e-01 1.08787632e+00 -9.89136621e-02 2.46693622e-02
-7.39372492e-01 -1.16800284e+00 2.62187552e-02 6.49417400e-01
7.81148016e-01 1.01361610e-01 -1.15168488e+00 -3.92099082e-01
-3.18773985e-02 5.37140429e-01 -5.01551986e-01 -3.17002058e-01
-4.52874869e-01 -9.36687291e-01 1.34909406e-01 -1.83528531e-02
8.32375467e-01 -4.10825104e-01 1.74355116e-02 7.26698041e-02
-4.77126628e-01 -9.06820834e-01 -1.54984283e+00 -1.99628174e-01
-6.09910011e-01 -1.12798631e+00 -1.46029067e+00 -1.01047015e+00
1.01305962e+00 1.06984067e+00 8.42973709e-01 9.50380811e-04
-1.51709184e-01 7.73311198e-01 2.94528133e-03 5.20282328e-01
-3.17376852e-01 -6.79686725e-01 1.74534805e-02 7.21610427e-01
-3.14018667e-01 -6.49218261e-01 -9.81464624e-01 3.86459380e-01
-1.06026733e+00 2.84971058e-01 6.10003710e-01 1.02246571e+00
2.55383343e-01 3.53796899e-01 3.20489347e-01 -4.88653809e-01
1.05798352e+00 -3.40734601e-01 -4.72837299e-01 3.88367802e-01
-7.37883806e-01 2.05021963e-01 4.68321413e-01 -6.43969476e-01
-1.29003596e+00 -1.36104628e-01 5.37560821e-01 -9.64786351e-01
1.35060295e-01 5.85004210e-01 -7.04711378e-02 -5.36899447e-01
7.78677642e-01 4.70037103e-01 2.58419514e-01 -6.73942626e-01
9.66335237e-01 6.51021063e-01 1.01641452e+00 -3.82513434e-01
1.32756531e+00 5.24583757e-01 -5.35491928e-02 -9.99059856e-01
-6.69972718e-01 -7.28042305e-01 -9.26818848e-02 -2.86868155e-01
8.79037023e-01 -1.01330817e+00 -6.01556301e-01 9.17941630e-01
-1.19706655e+00 -3.70610207e-01 -1.60491556e-01 3.91707003e-01
-4.54398304e-01 1.02757847e+00 -1.03393292e+00 -6.99326634e-01
-2.73177266e-01 -1.05358756e+00 7.79986262e-01 3.38273883e-01
2.36391842e-01 -1.15777719e+00 1.04050646e-02 4.28279907e-01
6.89716816e-01 -2.95605399e-02 5.10453165e-01 4.95239407e-01
-7.85789013e-01 1.17596827e-01 -8.56138647e-01 4.86699462e-01
6.71479225e-01 -5.78421354e-01 -9.85905111e-01 -5.85453689e-01
5.82619846e-01 2.60773450e-01 1.13209808e+00 9.43909585e-01
9.66368675e-01 -7.56148160e-01 -1.78268597e-01 1.00451207e+00
1.45702529e+00 -3.20210934e-01 6.10958576e-01 -1.38694510e-01
1.19274867e+00 2.08754703e-01 3.33272703e-02 1.38832256e-01
1.57751516e-01 5.45305014e-01 2.67979980e-01 -3.24642301e-01
-8.35301936e-01 -8.11006427e-02 4.59101439e-01 8.48930120e-01
-2.53492028e-01 -4.03289460e-02 -5.39606392e-01 6.60789490e-01
-1.97020376e+00 -8.46910417e-01 -4.23424304e-01 2.14454484e+00
1.13921785e+00 -5.80641329e-01 -3.28650117e-01 -2.44269833e-01
1.20002437e+00 4.00569618e-01 -5.78573644e-01 3.70907873e-01
-2.35036805e-01 1.02136046e-01 8.57256055e-01 1.15457356e+00
-9.81070280e-01 8.02135825e-01 5.89167166e+00 9.66092944e-01
-9.95168626e-01 1.87806129e-01 5.26588678e-01 4.57372576e-01
-4.15206254e-01 3.01559810e-02 5.55373058e-02 6.60564721e-01
4.72339869e-01 -5.07691085e-01 1.48274016e+00 3.86859983e-01
4.36068356e-01 -2.87725590e-02 -6.68753207e-01 1.47224474e+00
1.92316264e-01 -1.24140286e+00 -1.07998319e-01 4.08211984e-02
8.61499488e-01 -1.78026676e-01 1.38478503e-01 -3.34930032e-01
4.40429956e-01 -9.64758337e-01 7.07289517e-01 8.12946022e-01
1.04995179e+00 -2.45959222e-01 3.15139532e-01 1.14978969e-01
-1.13706529e+00 1.74011186e-01 -2.96784878e-01 8.67515281e-02
1.99445739e-01 1.05556571e+00 -2.26619586e-01 7.76951194e-01
5.92826426e-01 9.91905689e-01 -3.57934117e-01 1.20674253e+00
-2.73750067e-01 5.59188247e-01 2.90723860e-01 6.56397402e-01
-2.62024999e-01 -8.05491865e-01 9.98849571e-01 1.25376701e+00
5.62936723e-01 4.22708601e-01 -1.75706968e-01 1.29161954e+00
-2.35542521e-01 -4.21394855e-01 -3.31223786e-01 2.96036988e-01
1.18285246e-01 1.21973395e+00 -3.86437416e-01 -2.37357363e-01
-3.60164762e-01 1.76676559e+00 4.25869636e-02 1.23792171e+00
-7.64034152e-01 -3.48363668e-01 8.84523749e-01 -3.00832670e-02
2.10568681e-01 -2.34010920e-01 -3.39279056e-01 -1.74600446e+00
-1.32535756e-01 -7.81384230e-01 4.57881950e-02 -1.19964254e+00
-1.70031846e+00 4.47836339e-01 -2.16500804e-01 -1.10238755e+00
3.72128338e-01 -3.50974619e-01 -6.20459020e-01 1.39367080e+00
-1.76139128e+00 -1.14585948e+00 -6.90589070e-01 8.70534718e-01
4.50130701e-01 3.03864092e-01 2.15180397e-01 1.49011895e-01
-5.24592936e-01 9.82522890e-02 2.21282795e-01 -7.81312063e-02
8.90692592e-01 -1.42198288e+00 2.56489009e-01 1.39944243e+00
-3.05423617e-01 7.97793031e-01 9.31014001e-01 -8.50269377e-01
-1.82274127e+00 -1.25064826e+00 3.09267968e-01 -1.87232897e-01
9.66777623e-01 1.47052452e-01 -1.22310925e+00 5.50282240e-01
5.51897228e-01 3.00565273e-01 1.96128972e-02 -5.45459986e-01
-2.27317572e-01 -7.76702911e-02 -1.10706580e+00 5.77593803e-01
1.06189048e+00 -7.76446998e-01 -6.20978773e-01 2.56775141e-01
8.26926589e-01 -5.55676520e-01 -6.53568864e-01 2.42947251e-01
9.24616605e-02 -8.02777648e-01 1.26306331e+00 -1.01307735e-01
2.38343269e-01 -8.02716494e-01 1.10430993e-01 -1.78903961e+00
-9.34489906e-01 -1.35993373e+00 -5.32233119e-01 1.07648468e+00
-4.40771207e-02 -7.60783195e-01 2.96751559e-01 6.18104696e-01
-2.98948467e-01 -1.83347493e-01 -6.45720780e-01 -7.22343504e-01
-4.88166660e-01 -5.37269935e-02 1.71584845e-01 1.21379054e+00
-8.13901275e-02 2.06638366e-01 -7.88617909e-01 6.24588549e-01
1.31128550e+00 8.69623795e-02 3.00992817e-01 -7.03214407e-01
-3.89031082e-01 -4.50074852e-01 -3.25847347e-03 -1.58602834e+00
9.88745876e-03 -7.49550700e-01 3.56020153e-01 -1.62383211e+00
3.96116793e-01 1.89781431e-02 -1.80722579e-01 4.30423617e-02
-7.53079116e-01 2.82925367e-01 -3.88665721e-02 6.63970470e-01
-1.36894584e-01 4.41910177e-01 1.64616895e+00 -3.55698407e-01
-1.68645024e-01 -1.42445505e-01 -8.66875052e-01 5.86656690e-01
2.34700248e-01 -3.16015892e-02 -3.75358641e-01 -4.87270743e-01
-1.03191443e-01 2.54711539e-01 6.05618954e-01 -6.75464511e-01
3.52879107e-01 -3.15101892e-01 1.70492917e-01 1.42886400e-01
1.03361972e-01 -7.57287443e-01 4.71473277e-01 2.51153886e-01
-1.26411375e-02 -5.90291798e-01 -6.96252510e-02 9.98605251e-01
-1.48101300e-01 -9.04363021e-02 1.03185773e+00 4.68527637e-02
-4.11907792e-01 3.91630650e-01 -2.83626944e-01 4.54338267e-02
6.15778029e-01 -1.58555999e-01 -6.46775424e-01 -6.14371359e-01
-6.52266681e-01 -1.80573136e-01 8.17511320e-01 1.08848512e-01
7.52336442e-01 -1.36996329e+00 -7.34223425e-01 1.67854801e-01
-2.76264876e-01 -2.35216916e-01 4.79994148e-01 1.14196765e+00
-5.88334203e-01 -1.86084479e-01 1.85826853e-01 -3.10988069e-01
-9.93377864e-01 8.62555504e-01 5.42678833e-01 4.95954305e-02
-5.96869946e-01 9.56670642e-01 7.23607719e-01 3.41438323e-01
-7.09644556e-02 -4.97665048e-01 1.66930005e-01 -3.48332196e-01
5.90105176e-01 5.41075170e-01 -2.09712476e-01 -7.88674653e-01
-1.43785655e-01 9.98514354e-01 3.43484610e-01 -8.74047279e-02
1.11220443e+00 -7.73356020e-01 -6.76747620e-01 -1.77027792e-01
1.34231794e+00 4.72495705e-01 -1.61213386e+00 -2.93568194e-01
-2.43456364e-01 -8.22959423e-01 6.91977561e-01 -6.23400927e-01
-1.34694052e+00 3.39849740e-01 5.29294848e-01 6.15956485e-01
1.55435729e+00 4.59337272e-02 7.30124354e-01 -4.82854724e-01
3.51647958e-02 -3.94121975e-01 2.23960862e-01 2.03047201e-01
1.31967521e+00 -8.61888826e-01 -2.93927360e-02 -6.81507647e-01
-3.60248655e-01 1.03183985e+00 5.78408241e-02 -2.02680588e-01
6.23106360e-01 -9.26944390e-02 -4.08016779e-02 -2.71611571e-01
7.13278577e-02 -2.43810013e-01 8.57426286e-01 4.95286971e-01
1.99407771e-01 5.46361059e-02 -1.04172051e-01 2.94877648e-01
3.13120276e-01 1.88750938e-01 6.50342822e-01 3.87609512e-01
-4.81462985e-01 -5.79601407e-01 -9.20594037e-01 1.52995273e-01
-2.82697856e-01 -4.43143070e-01 -1.43011034e-01 -8.67615342e-02
-1.56417340e-01 1.24080074e+00 -4.06941295e-01 -2.32525572e-01
1.49397776e-01 -3.79032135e-01 5.74491978e-01 -2.90139616e-01
1.51756406e-02 4.78899449e-01 -3.42326134e-01 -4.98831898e-01
-4.46938068e-01 -3.11300099e-01 -6.87261701e-01 -4.20185030e-01
-5.67197800e-01 1.30490288e-01 1.64649591e-01 5.51908970e-01
4.10907745e-01 4.30350363e-01 6.60198331e-01 -1.10392129e+00
-3.61200094e-01 -9.19891894e-01 -9.33520913e-01 4.96881306e-01
1.03633893e+00 -3.57997060e-01 -9.97596264e-01 5.75105309e-01] | [11.613428115844727, -2.7813332080841064] |
54736440-3be4-47ba-a1f6-a7d766efa935 | a-survey-of-label-efficient-deep-learning-for | 2305.19812 | null | https://arxiv.org/abs/2305.19812v1 | https://arxiv.org/pdf/2305.19812v1.pdf | A Survey of Label-Efficient Deep Learning for 3D Point Clouds | In the past decade, deep neural networks have achieved significant progress in point cloud learning. However, collecting large-scale precisely-annotated training data is extremely laborious and expensive, which hinders the scalability of existing point cloud datasets and poses a bottleneck for efficient exploration of point cloud data in various tasks and applications. Label-efficient learning offers a promising solution by enabling effective deep network training with much-reduced annotation efforts. This paper presents the first comprehensive survey of label-efficient learning of point clouds. We address three critical questions in this emerging research field: i) the importance and urgency of label-efficient learning in point cloud processing, ii) the subfields it encompasses, and iii) the progress achieved in this area. To achieve this, we propose a taxonomy that organizes label-efficient learning methods based on the data prerequisites provided by different types of labels. We categorize four typical label-efficient learning approaches that significantly reduce point cloud annotation efforts: data augmentation, domain transfer learning, weakly-supervised learning, and pretrained foundation models. For each approach, we outline the problem setup and provide an extensive literature review that showcases relevant progress and challenges. Finally, we share insights into current research challenges and potential future directions. A project associated with this survey has been built at \url{https://github.com/xiaoaoran/3D_label_efficient_learning}. | ['Shijian Lu', 'Ling Shao', 'Xiaoqin Zhang', 'Aoran Xiao'] | 2023-05-31 | null | null | null | null | ['efficient-exploration'] | ['methodology'] | [ 4.06354805e-03 6.77761436e-03 -5.73499739e-01 -5.23806334e-01
-9.59393919e-01 -6.64373219e-01 2.68983364e-01 3.22785228e-01
-3.31644088e-01 4.60050672e-01 -3.46834302e-01 -3.00765663e-01
-1.25484064e-01 -7.08590269e-01 -9.68553543e-01 -4.27430183e-01
-2.36589760e-01 9.24021602e-01 3.23201194e-02 2.42247522e-01
3.15210879e-01 1.13687491e+00 -1.90069115e+00 1.20901495e-01
6.88167512e-01 1.18874955e+00 3.30460489e-01 1.87562913e-01
-5.65117776e-01 3.77967238e-01 -4.82154518e-01 -1.22911043e-01
4.40273076e-01 2.87065417e-01 -1.02510428e+00 1.08454444e-01
7.25946128e-01 -3.48152786e-01 -1.33489326e-01 8.70229244e-01
4.95796472e-01 1.06814288e-01 6.87474370e-01 -1.51812351e+00
-3.96703601e-01 1.04129642e-01 -5.15182018e-01 7.93345049e-02
-2.43208691e-01 -1.42226666e-01 9.09556985e-01 -1.22472501e+00
5.07216334e-01 6.76764429e-01 9.11408544e-01 5.78921974e-01
-7.47952223e-01 -1.01983488e+00 1.46102965e-01 1.05833091e-01
-1.62817633e+00 -1.20268986e-01 1.05091059e+00 -8.02817702e-01
9.30709958e-01 1.10548899e-01 8.76302421e-01 6.50629222e-01
-4.37270075e-01 8.90079558e-01 8.01238596e-01 -3.80077749e-01
3.06472361e-01 -2.18606205e-03 4.41374555e-02 6.31221533e-01
4.51327348e-03 2.24488795e-01 -4.01825011e-01 -2.39094347e-01
9.17133808e-01 2.10633427e-01 1.68602392e-01 -5.87848067e-01
-1.02920365e+00 8.36163402e-01 8.07485640e-01 2.06295952e-01
-1.24653317e-01 2.94931382e-01 4.24993455e-01 9.66725405e-03
9.26155448e-01 5.27279377e-01 -7.42445290e-01 1.19617678e-01
-1.08633316e+00 4.05424446e-01 2.46470198e-01 1.52188218e+00
1.16130221e+00 4.95689884e-02 2.52156705e-01 8.87445390e-01
3.49360764e-01 4.77381021e-01 -1.83858961e-01 -1.18582785e+00
3.02713901e-01 6.85330451e-01 1.62449270e-01 -5.65601349e-01
-5.79175055e-01 -5.96184194e-01 -6.83112621e-01 5.64169347e-01
8.57294500e-02 4.35869806e-02 -9.11778152e-01 1.33870089e+00
4.27335232e-01 5.29101849e-01 -5.08301675e-01 7.61328757e-01
1.17674148e+00 5.23976862e-01 4.40609813e-01 1.21313214e-01
1.11724174e+00 -1.02155137e+00 -2.99238473e-01 -2.74910867e-01
7.20033884e-01 -5.98656952e-01 1.23731840e+00 8.55783448e-02
-9.67411578e-01 -6.76839769e-01 -9.72689331e-01 -2.07246453e-01
-6.14569843e-01 2.80909777e-01 9.05028999e-01 2.86880404e-01
-1.09287393e+00 7.07111895e-01 -9.16218698e-01 -2.88323432e-01
1.03496575e+00 7.17696428e-01 -1.59385875e-01 -1.16044849e-01
-7.06752777e-01 6.60263836e-01 3.49673152e-01 -2.86962837e-01
-8.29547405e-01 -1.07911944e+00 -6.10837519e-01 -2.43172094e-01
1.21362075e-01 -7.41487741e-01 1.51488793e+00 -6.26721382e-01
-1.07324135e+00 1.47855341e+00 -1.11973792e-01 -1.16336219e-01
1.21021494e-01 -3.17609936e-01 4.26360704e-02 7.83473179e-02
3.32083493e-01 1.23383546e+00 5.62065721e-01 -1.65414155e+00
-1.00849545e+00 -4.11578685e-01 -1.53446663e-02 3.24024320e-01
-3.52463365e-01 1.32135719e-01 -7.04243600e-01 -5.12732387e-01
5.73232286e-02 -1.12758267e+00 -3.62922609e-01 4.57645386e-01
6.74388278e-03 -8.13100338e-01 8.70131969e-01 -6.31806627e-02
6.96125150e-01 -2.16948128e+00 -3.35098863e-01 2.17734836e-02
4.56862450e-01 2.68260300e-01 -5.79433627e-02 3.79487008e-01
-1.99641943e-01 2.31145099e-01 -4.04329091e-01 -7.20732510e-01
-4.52557467e-02 2.94050634e-01 -4.58198488e-01 4.99827027e-01
1.24415927e-01 1.01307321e+00 -9.50453639e-01 -6.56328022e-01
6.60417020e-01 4.28806305e-01 -3.10635507e-01 2.68610775e-01
-4.41485703e-01 8.79561186e-01 -3.33387583e-01 1.16395617e+00
8.33478034e-01 -6.05590582e-01 -5.47589600e-01 -1.87082723e-01
-3.10469359e-01 1.08782962e-01 -8.23896527e-01 1.88581455e+00
-5.09285986e-01 6.82993054e-01 -8.12103748e-02 -9.32485282e-01
1.04996288e+00 2.08044320e-01 1.10506546e+00 -2.98360616e-01
9.58878994e-02 4.97203529e-01 -7.08027005e-01 -3.72852296e-01
3.63459975e-01 -3.13312709e-01 4.30404432e-02 2.27566108e-01
8.26511011e-02 -4.54317153e-01 -2.46188954e-01 -2.47954786e-01
8.49122822e-01 3.42962682e-01 9.39935967e-02 -1.14300065e-01
2.65767127e-01 5.77007890e-01 3.01618576e-01 6.76276624e-01
-4.14804190e-01 7.40425110e-01 -3.20878088e-01 -9.08235908e-01
-1.23790324e+00 -9.34087276e-01 -4.26581740e-01 1.19891608e+00
3.41053158e-01 -3.55112731e-01 -3.45768690e-01 -6.59496486e-01
2.47850701e-01 3.96047562e-01 -4.54567760e-01 2.33711585e-01
-5.94215393e-01 -5.29676557e-01 4.85770911e-01 8.85841608e-01
4.84925717e-01 -1.13655412e+00 -3.18008035e-01 -2.56926596e-01
-2.75807500e-01 -1.19665849e+00 3.53317380e-01 1.80700079e-01
-1.36688030e+00 -9.07982230e-01 -6.57684088e-01 -1.08440328e+00
7.19588041e-01 5.75983763e-01 1.62675083e+00 3.90476346e-01
-5.79788499e-02 6.04949832e-01 -4.40465599e-01 -9.49158967e-01
2.94451505e-01 3.83698255e-01 5.38520291e-02 -6.64564192e-01
8.56333435e-01 -6.66305661e-01 -5.97140849e-01 1.47608116e-01
-6.77481830e-01 -1.60430834e-01 6.76107526e-01 3.18135381e-01
1.17901909e+00 -2.13848464e-02 3.00343484e-01 -7.66861200e-01
2.49703571e-01 -6.25181496e-01 -7.64125168e-01 -3.47537473e-02
-4.94572580e-01 -4.51660156e-01 1.91295490e-01 7.73449913e-02
-7.30486989e-01 4.92177516e-01 -6.25957489e-01 -8.22174907e-01
-7.63286829e-01 2.95665979e-01 2.86243204e-02 -6.71145558e-01
6.99476957e-01 -1.80792987e-01 -2.92903781e-01 -5.99771142e-01
3.98189932e-01 6.48926198e-01 5.65432906e-01 -9.79787529e-01
9.89675879e-01 8.93860161e-01 1.43190891e-01 -1.00028706e+00
-1.17985761e+00 -9.35165107e-01 -1.12811601e+00 -4.95392591e-01
9.08432305e-01 -1.09574735e+00 -5.89453936e-01 2.74956077e-01
-1.36593676e+00 -3.91062886e-01 -6.27012789e-01 3.38653773e-01
-6.52010381e-01 2.37118304e-01 -5.32210648e-01 -6.83468044e-01
-4.32953119e-01 -1.10087335e+00 1.50469971e+00 8.73047858e-02
6.63038939e-02 -1.03838658e+00 1.33603066e-01 3.93363327e-01
1.45301282e-01 4.82212514e-01 8.26999485e-01 -4.41990793e-01
-9.91569042e-01 -4.42578614e-01 -3.84745389e-01 2.48471290e-01
-1.26615942e-01 -1.22582309e-01 -1.24542665e+00 -2.88560987e-01
-2.46841833e-01 -5.72268903e-01 5.56989849e-01 5.15302837e-01
1.71724439e+00 3.60296428e-01 -7.05245554e-01 9.02844846e-01
1.54484379e+00 -3.98631617e-02 2.41721362e-01 6.13019586e-01
9.98810649e-01 6.08289301e-01 9.79659259e-01 2.83628941e-01
4.40311998e-01 5.27871668e-01 8.70735407e-01 -3.54297459e-01
-1.53346911e-01 -2.24104956e-01 -3.73875618e-01 8.68720114e-01
-5.37992656e-01 -5.14500104e-02 -1.39062715e+00 5.97780585e-01
-1.71202564e+00 -2.92513698e-01 -3.48778248e-01 1.96418309e+00
3.80735695e-01 -1.13225549e-01 1.84738394e-02 -1.37744799e-01
5.47928631e-01 7.00484738e-02 -6.86699092e-01 2.34188721e-01
9.16586742e-02 4.43052560e-01 5.64531326e-01 2.69326657e-01
-1.65056002e+00 1.26974761e+00 6.74436283e+00 7.69431293e-01
-1.03878784e+00 3.87130082e-01 2.89637893e-01 9.79940780e-03
3.45403552e-02 -8.91174376e-02 -8.26349199e-01 2.94495642e-01
6.79743946e-01 2.06336275e-01 -4.41731606e-03 1.58634424e+00
-6.70794472e-02 2.58370131e-01 -1.13268292e+00 1.42383373e+00
-1.48744687e-01 -1.78340304e+00 -1.06696710e-01 2.56055236e-01
7.65747249e-01 9.11657393e-01 8.87077600e-02 2.92072862e-01
2.84326911e-01 -8.40269923e-01 5.68896770e-01 1.84443638e-01
1.17176604e+00 -7.52009928e-01 8.95309567e-01 3.74866962e-01
-1.38859773e+00 -2.79892180e-02 -7.19152808e-01 -2.08833069e-01
1.25083432e-01 6.79960132e-01 -5.98700643e-01 4.41017687e-01
1.14878511e+00 9.28562939e-01 -3.84395212e-01 1.41420341e+00
-3.43649834e-01 5.37202716e-01 -4.69913989e-01 3.83626133e-01
3.54845554e-01 -1.95423692e-01 3.73094589e-01 1.19282496e+00
3.39294314e-01 1.68447882e-01 4.09919679e-01 7.95265675e-01
-2.67597824e-01 5.03427275e-02 -8.63430560e-01 1.95807472e-01
8.64359200e-01 1.16016304e+00 -8.41273785e-01 -2.78470069e-01
-5.54380953e-01 4.91730154e-01 5.92009664e-01 1.41793981e-01
-5.65541804e-01 -7.15160593e-02 8.45090151e-01 2.20704630e-01
-1.21412754e-01 -5.98247886e-01 -8.49112034e-01 -7.78583765e-01
-1.34064093e-01 -3.19715977e-01 3.20968032e-01 -8.67693305e-01
-1.49926269e+00 5.70428729e-01 2.69749045e-01 -1.69972384e+00
2.45983183e-01 -8.04888189e-01 -3.34728688e-01 6.43069685e-01
-1.88404894e+00 -1.55982196e+00 -8.08646560e-01 3.92878026e-01
7.07863986e-01 -1.42343745e-01 8.63674760e-01 7.39354908e-01
-2.22515926e-01 3.64787966e-01 1.30814314e-02 1.05489694e-01
5.50648093e-01 -1.13859892e+00 6.34788871e-01 3.63146454e-01
1.61827341e-01 3.93910110e-01 1.75664097e-01 -7.09972024e-01
-8.97664428e-01 -1.44385493e+00 6.86359942e-01 -8.25160623e-01
2.86223412e-01 -4.53725964e-01 -8.93694103e-01 8.70394230e-01
-1.06211610e-01 2.55935609e-01 1.07083917e+00 2.39075020e-01
-9.76012871e-02 1.04202695e-01 -9.83705521e-01 1.73420787e-01
1.32578170e+00 -5.32297015e-01 -4.05865103e-01 8.71614039e-01
7.69853473e-01 -5.53431988e-01 -1.01891220e+00 8.46964717e-01
1.63999379e-01 -7.97779024e-01 1.25297403e+00 -4.32693601e-01
4.12148148e-01 -3.16798240e-01 -7.62737617e-02 -8.89743209e-01
-6.48294210e-01 6.85221255e-02 -1.97496235e-01 9.89277422e-01
1.58275262e-01 -1.97238445e-01 1.53906500e+00 6.06551945e-01
-8.09703648e-01 -9.69785452e-01 -7.72840381e-01 -8.05657029e-01
4.07965511e-01 -9.20870364e-01 8.13626766e-01 1.22872972e+00
-4.43869054e-01 2.02702910e-01 -9.40881893e-02 3.16798031e-01
9.71263111e-01 2.08989933e-01 9.12229538e-01 -1.80418873e+00
4.35650110e-01 -3.09169859e-01 -2.87015140e-01 -1.33859336e+00
3.87791544e-01 -1.10162079e+00 -5.41892648e-03 -1.88042295e+00
-3.46501656e-02 -1.23769987e+00 -6.20760769e-02 7.19303668e-01
1.82753980e-01 5.50974309e-01 1.21201679e-01 8.55478644e-01
-6.98224783e-01 4.66364712e-01 9.82472241e-01 -1.93022430e-01
-3.33522856e-02 3.44115943e-01 -3.11641872e-01 9.09954250e-01
1.07264411e+00 -4.94187623e-01 -4.49258775e-01 -8.90275657e-01
3.52436900e-01 -5.61475754e-01 3.16981375e-01 -1.23771191e+00
3.30286950e-01 -8.90991315e-02 3.55927497e-01 -1.24912572e+00
5.01835525e-01 -1.19117177e+00 -9.38133672e-02 6.03984483e-02
-8.36451873e-02 7.85505772e-02 4.18986350e-01 5.01461446e-01
-1.85283229e-01 -3.19363445e-01 8.32977295e-01 -5.40417969e-01
-1.21284091e+00 8.33450079e-01 -7.04667205e-03 -1.94326177e-01
1.01441002e+00 -3.18664223e-01 5.08306501e-03 -1.74146574e-02
-7.27157354e-01 2.90295452e-01 5.62888741e-01 4.76846755e-01
6.98926866e-01 -1.49912155e+00 -5.43733478e-01 -2.10381881e-03
4.60658044e-01 7.08031535e-01 2.88284779e-01 3.56980830e-01
-7.15592802e-01 4.23539162e-01 -1.84696466e-01 -1.07658350e+00
-1.05442142e+00 5.90387404e-01 2.64601707e-01 2.92261522e-02
-7.47859240e-01 1.15317810e+00 1.63159356e-01 -9.62265849e-01
4.43252027e-01 -1.03179298e-01 -2.03265712e-01 -3.12889487e-01
1.19249374e-01 3.83190781e-01 4.47285473e-01 -5.96050084e-01
-4.51912075e-01 1.05587327e+00 2.14830354e-01 3.15700591e-01
1.44225276e+00 7.83300474e-02 -2.46229723e-01 4.82446909e-01
1.23349047e+00 -4.92340207e-01 -1.22290814e+00 -3.99858594e-01
1.60116389e-01 -5.46245217e-01 1.79984167e-01 -4.86784071e-01
-1.05664003e+00 1.10272574e+00 7.77706206e-01 -1.61371216e-01
9.70316231e-01 5.83510578e-01 6.87801957e-01 4.26686585e-01
6.58259213e-01 -9.86692429e-01 1.06570840e-01 6.65730774e-01
7.79624581e-01 -1.62529540e+00 1.28284395e-01 -7.01462507e-01
-2.25780994e-01 9.38485742e-01 8.61936212e-01 -2.06119061e-01
1.17829180e+00 2.60089815e-01 1.86655194e-01 -8.24481130e-01
-2.07231715e-01 -3.63031656e-01 2.41977498e-01 9.35113013e-01
2.94163465e-01 1.24967303e-02 2.11865038e-01 1.97816253e-01
-3.55859667e-01 8.53937417e-02 -1.86416641e-01 1.13978124e+00
-5.99225700e-01 -1.12819088e+00 -3.83495986e-01 4.75380421e-01
3.35661769e-02 6.41519204e-02 -4.84003305e-01 8.41039598e-01
6.63305104e-01 6.76943064e-01 4.80016053e-01 -5.41949987e-01
3.88169169e-01 -3.97881791e-02 3.14184487e-01 -9.45288360e-01
-3.00711036e-01 -7.46827498e-02 -2.50195473e-01 -4.03517276e-01
-8.27534080e-01 -5.52439749e-01 -1.33996749e+00 -2.57103115e-01
-3.14833462e-01 1.86821297e-01 1.04383183e+00 8.88774514e-01
5.10911524e-01 3.37697297e-01 4.16129678e-01 -1.50222135e+00
1.32961869e-02 -7.35199392e-01 -3.53384584e-01 1.41004622e-01
3.39662105e-01 -1.03230846e+00 -2.20837072e-01 6.36716783e-02] | [8.030072212219238, -3.2365293502807617] |
cb09133f-bb4f-4d6f-8b5d-05d8add1f130 | bottom-up-parsing-via-sequence-labeling | null | null | https://openreview.net/forum?id=CMXc3nK27q1 | https://openreview.net/pdf?id=CMXc3nK27q1 | Bottom Up Parsing via Sequence Labeling | We translate the sequence labeling framework, first introduced for top-down discourse parsing by Koto et al. (2021), to bottom-up discourse parsing. We introduce a novel parser that is not constrained by parsing direction (left-to-right or otherwise), and is conditioned on previous parsing decisions. We describe the unique training requirements of a (directionally) unconstrained parser and explore two different training procedures. Additionally, we introduce a novel dynamic oracle for unconstrained bottom-up parsing. Our proposed parser achieves state-of-the-art performance amongst bottom-up RST parsers. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['discourse-parsing'] | ['natural-language-processing'] | [ 6.81917369e-01 1.17826593e+00 -1.97053865e-01 -6.03934407e-01
-1.21742690e+00 -1.07202637e+00 6.95305049e-01 1.93011791e-01
-2.41983473e-01 8.92449260e-01 7.06118882e-01 -9.97275531e-01
3.50846052e-01 -8.91859293e-01 -7.19324470e-01 -2.89096802e-01
-4.25295904e-02 4.79099989e-01 6.43186569e-01 -3.54467005e-01
2.63656765e-01 1.64953351e-01 -9.29242074e-01 7.17125177e-01
3.86148006e-01 3.37902606e-01 3.43679070e-01 1.40536046e+00
-5.13153076e-01 1.10717273e+00 -5.80465078e-01 -5.38676918e-01
-2.37757117e-01 -6.23583734e-01 -1.56740117e+00 1.17446743e-01
8.80296901e-02 -2.80444235e-01 -1.67586897e-02 8.92985821e-01
1.12884358e-01 -6.94629923e-02 1.87706694e-01 -4.35027927e-01
-5.65693378e-01 1.04520643e+00 -1.53604582e-01 2.79299706e-01
8.31693769e-01 -4.32174891e-01 1.31446993e+00 -5.65270305e-01
1.05073404e+00 1.36649215e+00 3.11059117e-01 8.87701988e-01
-1.00138795e+00 2.28424340e-01 9.42924857e-01 -2.09122181e-01
-3.76258105e-01 -3.72585386e-01 6.81651950e-01 -4.27032143e-01
1.29108191e+00 4.62001413e-01 4.89602685e-02 1.00642061e+00
-9.31046531e-02 1.03798831e+00 1.21702433e+00 -9.91446614e-01
2.24533081e-01 -3.16638917e-01 6.27377093e-01 7.36945152e-01
-5.64087890e-02 -4.21550214e-01 -4.80536431e-01 -1.50857016e-01
4.03776199e-01 -8.90564740e-01 -7.68337697e-02 -9.19650048e-02
-9.90892410e-01 8.63215029e-01 -1.45826608e-01 3.39476973e-01
-5.81269041e-02 -1.13044223e-02 6.74669981e-01 3.27370375e-01
3.09483230e-01 4.17413354e-01 -6.67944551e-01 -4.53905970e-01
-5.26825309e-01 -3.54530178e-02 1.28676069e+00 1.07084298e+00
3.78699511e-01 -1.21499494e-01 -2.88183868e-01 5.84466338e-01
3.33005577e-01 9.34467912e-02 4.83804733e-01 -1.22498417e+00
1.01256585e+00 2.12266296e-01 1.54018492e-01 -3.78900141e-01
-6.14338577e-01 3.23974527e-02 -3.17856729e-01 -2.54760087e-01
5.43712139e-01 -4.74960566e-01 -8.15792143e-01 1.92398894e+00
4.04798716e-01 -5.27519345e-01 3.66145372e-01 5.31527936e-01
9.17949498e-01 6.48758411e-01 3.20349574e-01 -6.85671568e-01
1.60984564e+00 -1.36973608e+00 -9.26886082e-01 -5.56112885e-01
1.00812721e+00 -5.25029302e-01 8.70161295e-01 2.17162043e-01
-1.56187737e+00 -3.24749827e-01 -9.42484319e-01 -3.67659986e-01
-9.18232501e-02 5.42841256e-02 5.36807299e-01 9.43196714e-01
-1.27421081e+00 5.49410939e-01 -1.09922183e+00 -2.87960261e-01
4.07351851e-02 1.77325159e-01 -3.46438855e-01 2.74219602e-01
-1.24020219e+00 7.41780698e-01 5.83982706e-01 1.54397711e-01
-5.18462896e-01 2.24549964e-01 -1.13508570e+00 -1.36647135e-01
8.50030303e-01 -6.69417322e-01 1.98500657e+00 -5.41002393e-01
-2.28831577e+00 1.09210026e+00 -6.84545577e-01 -4.51129705e-01
4.99732494e-01 -6.56125665e-01 1.40718394e-03 3.78808945e-01
3.94305587e-01 3.67319256e-01 4.60861683e-01 -1.26849866e+00
-6.10796511e-01 -2.03468129e-01 7.69358754e-01 5.48577428e-01
2.47801945e-01 5.78651607e-01 -4.27819937e-01 -5.08105516e-01
6.41879737e-01 -7.77409971e-01 -3.37152153e-01 -7.25525320e-01
-9.14564669e-01 -2.70781964e-01 4.68697488e-01 -1.02852619e+00
1.51931143e+00 -1.73855722e+00 4.11615282e-01 -3.98821652e-01
-8.28262866e-02 5.27432654e-03 8.16595834e-03 7.52154827e-01
-2.19504803e-01 4.76457208e-01 -6.97433949e-01 -6.62233293e-01
4.04112972e-02 4.76265132e-01 -2.73561954e-01 3.72420326e-02
3.94058049e-01 8.37040901e-01 -1.13145053e+00 -5.23665607e-01
5.41550247e-03 -2.98264176e-01 -5.32546878e-01 6.06035352e-01
-5.42213559e-01 6.33341014e-01 -5.73435128e-01 7.23777950e-01
2.71378309e-01 -2.16896474e-01 1.13052320e+00 4.03274775e-01
-4.19493109e-01 9.65790212e-01 -9.36624587e-01 1.61527014e+00
-3.67317259e-01 4.00221378e-01 4.60783571e-01 -8.78989160e-01
5.73920369e-01 5.50124347e-01 -2.63065785e-01 -3.10679227e-01
1.37620449e-01 1.65712804e-01 -9.34994668e-02 -5.92930257e-01
7.68432140e-01 -3.41103494e-01 -6.68853641e-01 3.89001369e-01
2.47905198e-02 1.37417823e-01 6.25147998e-01 1.73868194e-01
1.13965797e+00 6.51548326e-01 8.97599995e-01 -2.40017414e-01
9.00759995e-01 1.40763447e-01 5.84363163e-01 9.15998518e-01
-1.51681945e-01 6.80947661e-01 1.22269630e+00 -8.93931761e-02
-8.69953692e-01 -6.93829894e-01 -3.49265635e-02 1.86176038e+00
-2.16573223e-01 -6.19087458e-01 -1.07116699e+00 -9.08295691e-01
-6.86600149e-01 8.22451115e-01 -6.62684262e-01 6.81947589e-01
-1.57305408e+00 -5.80867648e-01 6.93857610e-01 7.36340106e-01
3.43749613e-01 -1.15135479e+00 -8.09283018e-01 5.13878167e-01
-2.28557169e-01 -1.42040682e+00 -4.75902809e-03 5.06736934e-01
-9.68574822e-01 -1.11506796e+00 -4.87819642e-01 -1.16448414e+00
5.62188745e-01 -3.07601154e-01 1.19384718e+00 -3.71207483e-02
7.02505171e-01 2.59224147e-01 -6.32669091e-01 6.42070696e-02
-1.01726365e+00 2.79377043e-01 -5.24051905e-01 -5.20090878e-01
-1.74549952e-01 -3.98474149e-02 -3.56426998e-03 -1.72392413e-01
-4.88545746e-01 3.60968024e-01 3.20520669e-01 9.58975911e-01
5.03820479e-01 -4.65219319e-01 3.18237603e-01 -1.85910237e+00
7.70939350e-01 -2.12825343e-01 -3.15752804e-01 4.97311503e-01
7.57170767e-02 2.01786861e-01 9.00636077e-01 1.08020954e-01
-1.41223955e+00 2.27834862e-02 -7.33735204e-01 5.01129150e-01
-4.07857388e-01 4.95532334e-01 -4.84330297e-01 3.84465516e-01
1.76457047e-01 -1.09916463e-01 -7.64944732e-01 -7.23879516e-01
8.39790642e-01 4.66533452e-01 6.25086367e-01 -1.03211677e+00
3.37758750e-01 1.33051232e-01 -4.40167896e-02 -7.72232115e-01
-9.30323839e-01 -2.53141522e-01 -1.18659472e+00 2.16452450e-01
1.26306283e+00 -6.38405561e-01 -4.84554708e-01 3.20580214e-01
-1.69782424e+00 -5.43706954e-01 -3.35456371e-01 -1.21518351e-01
-7.35873044e-01 7.17818201e-01 -1.03629172e+00 -1.02141547e+00
-3.25801253e-01 -1.14940321e+00 1.10287356e+00 1.35181516e-01
-3.16064954e-01 -1.19064105e+00 1.63870946e-01 1.90678298e-01
-1.04791448e-01 4.03637230e-01 1.05908287e+00 -9.89892125e-01
-2.22245201e-01 2.39714339e-01 -1.15423702e-01 2.05858663e-01
-1.36050448e-01 -2.14700118e-01 -9.13815022e-01 -3.99911441e-02
2.35904962e-01 -2.32151896e-01 6.53419673e-01 2.60969281e-01
7.80195415e-01 -4.71477538e-01 -3.52839828e-01 2.90201515e-01
1.08951247e+00 2.50724286e-01 1.78584933e-01 8.16256821e-01
7.16460943e-01 8.14545631e-01 6.09096289e-01 3.19349505e-02
6.94410443e-01 3.02825540e-01 2.12327719e-01 2.57295638e-01
5.91282807e-02 -4.42148060e-01 4.60038811e-01 1.11009610e+00
-9.42501500e-02 -5.19155681e-01 -8.75555396e-01 6.52033627e-01
-1.94859731e+00 -5.96572161e-01 -3.59297395e-01 1.74512935e+00
9.11708951e-01 5.34023345e-01 4.12702858e-02 4.00392972e-02
8.96998644e-01 6.98398948e-01 -1.09526232e-01 -1.19813299e+00
-1.28856018e-01 3.81684750e-01 3.80459428e-01 1.01788187e+00
-1.46273923e+00 1.18944979e+00 7.50068760e+00 4.38338518e-02
-6.22963071e-01 2.73243546e-01 5.33014953e-01 3.88798922e-01
-2.83374071e-01 3.59952241e-01 -9.55641210e-01 -5.30903526e-02
1.26422346e+00 -6.03215247e-02 -3.62403654e-02 9.77483273e-01
-1.11785740e-01 -5.20045161e-02 -1.31541455e+00 1.19295284e-01
-3.08201835e-02 -1.32799459e+00 -1.31163314e-01 -4.10462886e-01
4.38947380e-01 -2.28615165e-01 -5.74819505e-01 5.13184905e-01
5.83187997e-01 -7.52253771e-01 8.70589972e-01 -1.46380797e-01
6.28068745e-01 -3.23368192e-01 6.78185463e-01 7.03942120e-01
-1.38868022e+00 -4.41820100e-02 -7.72726443e-03 -4.17527109e-01
9.43589211e-01 1.56951293e-01 -7.59267032e-01 8.38360727e-01
2.76762247e-01 3.01418096e-01 -1.88451529e-01 7.65642300e-02
-1.10167348e+00 8.44170034e-01 -7.69352317e-02 -9.00787041e-02
4.30798113e-01 -6.19169883e-03 7.46583581e-01 1.67056525e+00
-1.48028418e-01 7.39816725e-01 4.81626868e-01 1.22619674e-01
-2.39687547e-01 3.07824987e-04 -2.28367522e-01 6.33657724e-02
5.66767156e-01 8.66541624e-01 -8.07718873e-01 -6.80664480e-01
-5.34757435e-01 8.98534119e-01 6.06206834e-01 1.35232091e-01
-3.73068631e-01 -1.48928553e-01 5.73878661e-02 -1.10139057e-01
4.93548006e-01 -4.13983315e-01 -3.88129652e-01 -1.01657939e+00
7.99270794e-02 -7.16255248e-01 6.12455606e-01 -4.68958527e-01
-9.94987071e-01 8.82237017e-01 2.11120307e-01 -6.10654831e-01
-5.26471913e-01 -9.48450208e-01 -7.25349307e-01 7.98890173e-01
-1.74541962e+00 -9.94741380e-01 3.82724166e-01 -1.66786164e-01
1.01470149e+00 2.96575487e-01 1.16385448e+00 -2.28807628e-01
-6.39804721e-01 2.11410269e-01 -1.96490705e-01 4.60823745e-01
1.75775304e-01 -1.75238025e+00 1.03788984e+00 1.26221991e+00
-2.70895600e-01 8.91223073e-01 7.46148586e-01 -6.90107942e-01
-1.20320189e+00 -8.08601320e-01 1.43991208e+00 -4.44495827e-01
1.00122654e+00 -3.86662960e-01 -8.74064207e-01 1.24500036e+00
5.68744361e-01 -4.72684920e-01 5.97820699e-01 2.97361642e-01
-7.97785725e-03 8.53294790e-01 -9.13097918e-01 3.75797182e-01
1.13642204e+00 -5.86439073e-01 -1.32981098e+00 3.27749878e-01
1.07805610e+00 -1.25936091e+00 -8.17551851e-01 3.21845561e-01
3.29782039e-01 -8.50143135e-01 6.35878503e-01 -8.69898498e-01
5.42867601e-01 6.07027039e-02 -3.15777749e-01 -7.48552442e-01
-2.45995954e-01 -1.00656796e+00 -3.37932736e-01 1.15336442e+00
6.37630105e-01 -5.36183000e-01 8.38044226e-01 6.48721635e-01
-6.20778680e-01 -8.68253350e-01 -9.15191531e-01 -5.53036988e-01
5.78289092e-01 -3.87595236e-01 2.53503114e-01 6.78370178e-01
5.23424566e-01 7.00775146e-01 -1.11053512e-01 3.91850173e-01
1.93013847e-01 3.31051946e-01 4.84039187e-01 -8.16139340e-01
-7.69794941e-01 -3.50773394e-01 2.39782289e-01 -1.88575006e+00
8.80256891e-02 -7.24910557e-01 4.21717614e-01 -1.93955338e+00
1.46358386e-02 -8.15702081e-02 1.04895055e-01 3.98595423e-01
-4.01285052e-01 -2.12496415e-01 4.98191506e-01 1.92194413e-02
-6.32018149e-01 1.39955252e-01 1.18046725e+00 3.02722305e-01
-2.44278893e-01 4.75977268e-03 -8.15362215e-01 1.05076444e+00
8.77281129e-01 -5.33071101e-01 -2.07601130e-01 -7.26684451e-01
2.33875722e-01 7.27160871e-01 -4.20403630e-01 -4.41659480e-01
1.11876681e-01 -1.87843099e-01 -3.71062547e-01 -7.08396077e-01
9.22660828e-02 4.02645394e-02 -5.56574166e-01 2.97872633e-01
-5.60149491e-01 4.22112375e-01 5.25964200e-02 3.96299392e-01
-3.05579245e-01 -8.12121451e-01 3.39374840e-01 -6.51076257e-01
-7.76548803e-01 -4.64439273e-01 -5.43368518e-01 4.88275051e-01
8.86015356e-01 -1.00568503e-01 -6.47383988e-01 1.30129643e-02
-1.33515310e+00 2.43997574e-01 2.20745653e-01 1.16309524e-01
-3.07044224e-03 -5.98691285e-01 -5.00464976e-01 -1.61311835e-01
-4.22478110e-01 4.50036645e-01 -8.27886090e-02 3.46757561e-01
-7.90412664e-01 7.38790452e-01 2.58705795e-01 -3.82552415e-01
-1.50377786e+00 3.29357803e-01 2.05864781e-03 -9.34739530e-01
-6.92310452e-01 1.16924298e+00 2.94033140e-01 -6.93225682e-01
-5.80797195e-02 -5.19175708e-01 -4.61169302e-01 -7.05236346e-02
3.55359197e-01 1.04696356e-01 -5.88373952e-02 -6.17528379e-01
-5.76174021e-01 4.74839389e-01 -5.98486811e-02 -5.13809443e-01
1.15821600e+00 -3.03381592e-01 -1.62374780e-01 7.33780086e-01
7.25235462e-01 2.47917771e-01 -9.92273390e-01 -3.79525386e-02
5.93963623e-01 -2.15500537e-02 -4.82795894e-01 -7.11752832e-01
-8.91098231e-02 1.04367161e+00 -4.04145896e-01 8.14802706e-01
7.78402150e-01 1.81527197e-01 7.43638456e-01 5.77998877e-01
3.26527357e-01 -1.14555502e+00 -2.01985478e-01 1.08817148e+00
6.69664741e-01 -1.00215006e+00 8.16502422e-02 -8.96862090e-01
-6.70055032e-01 1.37955534e+00 4.00471896e-01 2.25250479e-02
2.93574989e-01 4.34714556e-01 1.74147151e-02 -5.30082434e-02
-9.22420681e-01 -1.96643010e-01 -2.59849131e-01 4.89687085e-01
9.84729886e-01 2.50537902e-01 -6.98465884e-01 7.22770095e-01
-4.20423448e-01 -2.84506559e-01 8.05946231e-01 1.60213304e+00
-7.35907018e-01 -1.60033953e+00 -2.79986680e-01 -1.10546499e-01
-6.55969799e-01 -9.50524509e-02 -5.23525834e-01 1.02341187e+00
-2.87175387e-01 1.09590209e+00 -1.33732378e-01 1.28446072e-01
5.58238745e-01 4.77149844e-01 4.59247112e-01 -1.15093815e+00
-6.15413427e-01 1.14526525e-01 1.14547706e+00 -2.35479280e-01
-6.75911963e-01 -9.52176273e-01 -1.59076488e+00 2.14412529e-02
-1.77117556e-01 1.46426037e-01 4.32123035e-01 1.22881043e+00
-1.71561211e-01 8.89058888e-01 4.73364085e-01 -7.05390096e-01
-4.61163998e-01 -8.04363191e-01 4.25795317e-02 5.11677414e-02
3.05245966e-01 -1.74203604e-01 -5.83534837e-02 2.25822061e-01] | [10.702310562133789, 9.490985870361328] |
1b2b775a-4985-4a7b-87c9-de887b553f89 | runtime-construction-of-large-scale-spiking | 2306.09855 | null | https://arxiv.org/abs/2306.09855v1 | https://arxiv.org/pdf/2306.09855v1.pdf | Runtime Construction of Large-Scale Spiking Neuronal Network Models on GPU Devices | Simulation speed matters for neuroscientific research: this includes not only how quickly the simulated model time of a large-scale spiking neuronal network progresses, but also how long it takes to instantiate the network model in computer memory. On the hardware side, acceleration via highly parallel GPUs is being increasingly utilized. On the software side, code generation approaches ensure highly optimized code, at the expense of repeated code regeneration and recompilation after modifications to the network model. Aiming for a greater flexibility with respect to iterative model changes, here we propose a new method for creating network connections interactively, dynamically, and directly in GPU memory through a set of commonly used high-level connection rules. We validate the simulation performance with both consumer and data center GPUs on two neuroscientifically relevant models: a cortical microcircuit of about 77,000 leaky-integrate-and-fire neuron models and 300 million static synapses, and a two-population network recurrently connected using a variety of connection rules. With our proposed ad hoc network instantiation, both network construction and simulation times are comparable or shorter than those obtained with other state-of-the-art simulation technologies, while still meeting the flexibility demands of explorative network modeling. | ['Johanna Senk', 'Abigail Morrison', 'Pier Stanislao Paolucci', 'Viviana Fanti', 'Jonas Stapmanns', 'Elena Pastorelli', 'Gianmarco Tiddia', 'Jose Villamar', 'Bruno Golosio'] | 2023-06-16 | null | null | null | null | ['code-generation'] | ['computer-code'] | [-3.56871751e-03 -1.59005493e-01 6.70560241e-01 1.11554250e-01
4.48014438e-01 -7.29200900e-01 5.72186530e-01 1.81275666e-01
-9.07747209e-01 8.36496234e-01 -5.88504493e-01 -5.21983504e-01
2.59552784e-02 -9.69904304e-01 -6.18440032e-01 -6.35523915e-01
-3.83492768e-01 8.17321181e-01 6.66404307e-01 -4.39604580e-01
2.04432458e-01 9.14151371e-01 -1.96254885e+00 -1.51544079e-01
6.24070168e-01 6.04179025e-01 3.57091933e-01 6.05325639e-01
-3.32684852e-02 2.25423411e-01 -4.17865694e-01 -2.00651467e-01
1.97326809e-01 -3.59840691e-01 -3.03410113e-01 -2.61094719e-01
-2.84568489e-01 1.66145697e-01 -2.81895101e-01 9.05507267e-01
5.59353650e-01 -1.59861133e-01 2.69371152e-01 -1.23133373e+00
2.01903805e-01 8.64355445e-01 -2.13486820e-01 4.17371720e-01
-1.91341698e-01 6.17774546e-01 3.09886456e-01 -4.16876078e-01
9.84160542e-01 8.86877179e-01 7.37043500e-01 4.16506141e-01
-1.77631700e+00 -1.07107484e+00 1.19547360e-01 -1.57496363e-01
-1.62603974e+00 -3.07909846e-01 3.10819596e-01 -4.36923891e-01
1.47328973e+00 1.94645897e-01 1.73095167e+00 9.30372775e-01
5.03688991e-01 -8.47470164e-02 8.93149614e-01 3.21404710e-02
8.26924562e-01 -1.11114644e-01 6.73918948e-02 4.87744451e-01
6.58846378e-01 3.34232002e-02 -3.72223377e-01 -5.18652022e-01
1.51338315e+00 -4.45336968e-01 -5.57330949e-03 -1.46123052e-01
-1.36514318e+00 4.38582063e-01 2.27480054e-01 5.14957607e-01
-3.69796723e-01 5.84799588e-01 5.76423347e-01 2.31508777e-01
-5.98701462e-02 5.92551708e-01 -3.24462652e-01 -3.29110831e-01
-1.02441919e+00 6.73977554e-01 1.18615294e+00 8.75814617e-01
6.96183801e-01 2.05514312e-01 3.32498729e-01 4.53541279e-01
1.15237661e-01 2.39359170e-01 4.33095217e-01 -9.36831951e-01
-1.61249593e-01 6.12294555e-01 -2.86176145e-01 -7.33381867e-01
-8.03174973e-01 -7.95366526e-01 -1.08042991e+00 5.65853238e-01
3.25026870e-01 -2.40150616e-01 -6.22864544e-01 1.69808698e+00
3.04676950e-01 2.80852526e-01 -2.84697086e-01 5.13342559e-01
3.74428779e-01 4.98026520e-01 1.59513488e-01 -2.70281643e-01
1.37282610e+00 -5.43084264e-01 -4.45716023e-01 -1.70796271e-02
6.18617713e-01 -5.13182104e-01 6.00941956e-01 4.30308998e-01
-1.44194758e+00 -1.85710609e-01 -9.95063901e-01 2.78809041e-01
-4.12977070e-01 -2.71514356e-01 9.06159759e-01 6.56915784e-01
-1.53848732e+00 7.13339925e-01 -1.14287233e+00 -5.97962558e-01
3.31712306e-01 7.68990159e-01 -1.24544553e-01 5.96274972e-01
-7.94198394e-01 9.71975982e-01 2.94938058e-01 4.46518473e-02
-7.77473152e-01 -8.12163830e-01 -7.84583241e-02 3.82381588e-01
-2.85256147e-01 -1.20709491e+00 9.44031775e-01 -8.33968818e-01
-1.65436864e+00 7.17210293e-01 9.93743818e-03 -7.26528764e-01
6.82058632e-01 6.46317601e-01 -1.83677927e-01 -1.09670773e-01
-3.94975871e-01 8.51172388e-01 2.37292081e-01 -1.07687020e+00
-1.05940670e-01 -5.81024960e-02 -8.00137520e-02 1.49722844e-01
-2.44007502e-02 1.96597680e-01 -3.80102545e-01 -5.86827636e-01
1.57107636e-01 -1.13474500e+00 -6.75019979e-01 5.53123236e-01
-2.17323586e-01 1.91738784e-01 4.32878792e-01 -1.15660138e-01
8.86050701e-01 -1.87343252e+00 2.42582127e-01 6.82217538e-01
3.04861456e-01 3.46668035e-01 -2.40190513e-02 5.60216367e-01
-1.33909911e-01 1.02422461e-02 -3.77295494e-01 -1.75262123e-01
-4.54797804e-01 1.93887666e-01 3.54933329e-02 2.51988024e-01
-1.29295513e-01 7.31356859e-01 -6.35187745e-01 -2.33274594e-01
-1.75623611e-01 6.33288443e-01 -9.62186456e-01 -2.16354564e-01
-4.27772440e-02 6.06586456e-01 -1.61073208e-01 1.72872290e-01
6.18212640e-01 -3.42058450e-01 3.32066834e-01 2.99682375e-02
-5.34553766e-01 8.90555531e-02 -1.21096015e+00 1.68219745e+00
-4.25940126e-01 5.71423233e-01 1.92058191e-01 -6.74457014e-01
9.79136169e-01 1.86920479e-01 2.75148362e-01 -5.66485226e-01
4.34060156e-01 4.12967175e-01 8.19592535e-01 7.89535567e-02
1.44763619e-01 2.34631255e-01 3.46709579e-01 7.56385624e-01
1.29345819e-01 -3.41179788e-01 4.64542001e-01 1.81230694e-01
1.43536747e+00 -5.48866726e-02 1.59299299e-01 -9.00563121e-01
1.62564248e-01 1.12293981e-01 3.93436491e-01 7.55375028e-01
3.89162712e-02 3.13799918e-01 6.71298146e-01 -3.98064345e-01
-1.27234483e+00 -1.11257052e+00 -3.87630731e-01 4.89217341e-01
-6.15719659e-03 -2.98995346e-01 -1.19171190e+00 5.66317439e-01
-1.03567615e-01 4.23186392e-01 -4.82088506e-01 -7.79197067e-02
-7.15195119e-01 -9.93255138e-01 8.96618068e-01 -1.02998354e-02
5.62437594e-01 -1.36776471e+00 -1.15601981e+00 6.76698267e-01
3.94958466e-01 -7.68478215e-01 1.57603443e-01 4.95568752e-01
-1.31499171e+00 -8.65566254e-01 -4.24014717e-01 -7.75915980e-01
9.20130789e-01 -2.59421375e-02 8.10300946e-01 5.28466046e-01
-6.63551569e-01 -3.18487972e-01 1.92864552e-01 -1.09379113e-01
-3.10065061e-01 2.45243698e-01 1.22784548e-01 -3.63565743e-01
-2.45692015e-01 -1.44053316e+00 -6.29800081e-01 2.11519524e-01
-8.58511090e-01 6.30475461e-01 2.83356577e-01 5.84319830e-01
6.83633089e-01 -6.35344535e-02 5.05915999e-01 -8.41486812e-01
7.14612424e-01 -4.06806022e-01 -9.71695840e-01 -2.44679935e-02
-5.22950292e-01 1.03678621e-01 6.91819072e-01 -7.24084377e-01
-6.77465439e-01 2.10835546e-01 -3.00513804e-01 -6.80717733e-03
1.25236511e-01 4.09543455e-01 2.94601262e-01 -5.77059031e-01
8.63078177e-01 3.75148475e-01 2.05540255e-01 -1.27714738e-01
8.19521099e-02 -1.46419525e-01 3.51278991e-01 -5.05353630e-01
5.96393526e-01 6.30474806e-01 8.88969079e-02 -7.34754920e-01
6.68218136e-01 2.88118750e-01 -5.40441215e-01 -4.31677938e-01
2.39977866e-01 -7.09134161e-01 -1.12838185e+00 1.01407385e+00
-1.40021420e+00 -7.15448320e-01 -2.43001357e-01 1.92179009e-01
-6.17969394e-01 -1.81467742e-01 -8.60886753e-01 -4.62159306e-01
-4.21579629e-01 -1.21587133e+00 4.01331484e-01 2.47649685e-01
-5.21262825e-01 -8.49195838e-01 1.87989801e-01 -5.87649584e-01
8.65653515e-01 1.39123425e-01 1.15098631e+00 -2.15044290e-01
-6.37806594e-01 1.77222509e-02 -2.00407848e-01 -5.10454774e-01
-4.18669790e-01 5.31607091e-01 -7.06562996e-01 -3.25331926e-01
-3.54551487e-02 1.73440158e-01 4.68768567e-01 4.07260805e-01
1.13407826e+00 -7.64243156e-02 -7.19666421e-01 7.05206394e-01
1.40367138e+00 2.63634682e-01 7.78582275e-01 3.80209237e-01
3.81934941e-01 5.29269993e-01 -2.80906767e-01 5.87521017e-01
1.26556098e-01 6.66986287e-01 4.31486756e-01 -5.53844050e-02
5.92224710e-02 1.42720863e-01 -9.76091847e-02 1.01740515e+00
-2.44875848e-01 4.83275726e-02 -1.14062798e+00 2.56062210e-01
-1.97044432e+00 -8.36280584e-01 -2.15716764e-01 2.13326931e+00
1.01883984e+00 3.59168857e-01 1.25132248e-01 4.57258299e-02
7.27131367e-01 -4.36834395e-01 -8.08155656e-01 -5.26555002e-01
-3.59158367e-01 3.88236791e-01 4.39118862e-01 2.36800015e-01
-2.96432555e-01 8.94364536e-01 7.06357861e+00 7.16299415e-01
-1.42204130e+00 1.70191646e-01 5.32264829e-01 -5.71754336e-01
-2.97767222e-01 2.73308456e-01 -7.19854116e-01 4.75455254e-01
1.30787265e+00 -4.46348965e-01 9.13230419e-01 4.64221239e-01
4.93574858e-01 -3.96619350e-01 -9.43679631e-01 1.16242230e+00
-5.06969392e-01 -1.95887792e+00 6.06985502e-02 3.04885685e-01
6.12459779e-01 3.91039610e-01 -2.65909255e-01 -1.08426109e-01
3.37350637e-01 -8.01310062e-01 1.01937938e+00 5.65893054e-01
4.45373088e-01 -7.54795015e-01 2.92586684e-01 3.93851817e-01
-1.20275080e+00 2.46136904e-01 -2.97000140e-01 -4.42310601e-01
2.58031130e-01 7.49849558e-01 -2.49461129e-01 -3.20163637e-01
7.46909976e-01 2.53987104e-01 -5.03289104e-01 1.56061542e+00
2.97322512e-01 3.47913772e-01 -8.64735961e-01 -4.52087700e-01
-4.45436276e-02 -2.55902410e-01 4.16613579e-01 1.16220260e+00
3.94483566e-01 1.50047079e-01 -6.57182515e-01 1.35715723e+00
2.28927955e-02 -3.18309106e-02 -5.72583497e-01 2.30152249e-01
8.80081713e-01 1.39711165e+00 -1.49169791e+00 -3.01536620e-01
1.10352777e-01 5.28948843e-01 4.81232941e-01 4.95109826e-01
-8.40000153e-01 -4.25444305e-01 9.03211594e-01 3.22023004e-01
6.47624582e-02 -5.62929809e-01 -7.34312952e-01 -9.35592294e-01
-2.47164100e-01 -5.79694450e-01 -5.45599937e-01 -7.30173051e-01
-7.68319964e-01 9.24435735e-01 -2.18128577e-01 -8.74598384e-01
-1.56133100e-01 -3.00893039e-01 -8.44230711e-01 8.65050793e-01
-8.38688016e-01 -6.41177654e-01 -1.19492710e-01 6.52783513e-01
2.03431342e-02 -1.17260635e-01 1.03763032e+00 4.33338523e-01
-6.22509241e-01 4.29054350e-01 1.32463515e-01 -3.25166434e-01
-7.04598501e-02 -3.63272935e-01 9.91538644e-01 5.45433581e-01
-2.61272639e-01 9.01753604e-01 8.94441426e-01 -6.44445181e-01
-1.41066873e+00 -8.64227891e-01 6.99596822e-01 3.35692018e-01
8.54240954e-01 -9.28822041e-01 -1.08600175e+00 3.91523212e-01
1.08315669e-01 5.41323386e-02 2.72802800e-01 -2.44461343e-01
-6.33836389e-02 4.66338582e-02 -1.29085171e+00 1.18925619e+00
1.41563344e+00 -2.36187682e-01 1.77670851e-01 1.25378057e-01
2.38431185e-01 -2.04445243e-01 -6.55252278e-01 -7.21189752e-02
6.48109376e-01 -8.77148509e-01 7.24387050e-01 4.73508723e-02
-1.03855446e-01 -3.83908331e-01 4.33365881e-01 -1.16025376e+00
-3.26600373e-01 -8.73185396e-01 2.34284818e-01 8.45103979e-01
4.53095376e-01 -1.11134362e+00 7.18079627e-01 5.59342623e-01
2.13482771e-02 -7.70251811e-01 -1.29989898e+00 -7.44529307e-01
1.16472870e-01 -2.23939359e-01 5.70292592e-01 7.63370216e-01
3.01039875e-01 4.91786413e-02 4.36528355e-01 -1.51277900e-01
5.11513710e-01 -3.07648271e-01 5.68026721e-01 -1.42289972e+00
-3.37097496e-01 -9.89592731e-01 -6.10609055e-01 -5.69302499e-01
-9.88932475e-02 -8.06206286e-01 -2.63035804e-01 -1.22515047e+00
4.61450629e-02 -8.23679030e-01 2.66339421e-01 4.46123570e-01
5.11685073e-01 1.58259451e-01 1.34185821e-01 4.61049825e-01
-9.54445824e-02 3.14024895e-01 1.07809281e+00 3.31690729e-01
-2.57550776e-01 -3.39657485e-01 -1.85225442e-01 7.09452510e-01
7.35531747e-01 -6.83654904e-01 -3.39732170e-01 -4.75238591e-01
6.43663943e-01 -1.00237634e-02 4.19297695e-01 -1.58552444e+00
7.91224539e-01 1.05875365e-01 4.80509400e-02 -2.07291394e-01
3.67285192e-01 -6.47215784e-01 1.04019964e+00 1.09183383e+00
-1.46890387e-01 4.83261108e-01 7.01657176e-01 2.86673933e-01
1.91436395e-01 -9.52141285e-02 9.30885255e-01 -1.06214665e-01
-2.48889938e-01 2.08906189e-01 -1.30077052e+00 -3.04413050e-01
1.07581449e+00 -3.87619317e-01 -4.96110260e-01 8.90857447e-03
-9.39314127e-01 -2.18666121e-02 7.01020837e-01 1.01965263e-01
3.69802237e-01 -9.20222700e-01 -3.73248935e-01 4.11943227e-01
-2.88484037e-01 -2.09205538e-01 2.13681906e-01 7.83043265e-01
-1.20863414e+00 8.35427418e-02 -9.03768659e-01 -5.46591222e-01
-8.56596529e-01 7.87179768e-02 6.40454948e-01 -1.27550766e-01
-5.83971083e-01 6.53397143e-01 2.60454929e-03 -3.58134925e-01
-3.42022143e-02 -5.05231202e-01 4.08293605e-02 -1.44432113e-01
2.56209493e-01 1.73913166e-01 3.27599317e-01 -2.43145540e-01
-2.34652773e-01 3.99920970e-01 9.91091505e-02 -4.37001854e-01
1.28618312e+00 8.07833821e-02 -5.29955328e-01 4.35780674e-01
6.59336269e-01 -6.24355137e-01 -1.23641622e+00 2.02728480e-01
-2.67210603e-01 1.89157456e-01 -9.96455327e-02 -6.61425114e-01
-1.19095135e+00 6.94177926e-01 5.71076334e-01 1.38478950e-01
8.28841031e-01 -3.74642164e-01 5.43958724e-01 5.54730475e-01
1.08196485e+00 -8.42440844e-01 -3.63056034e-01 6.93429112e-01
7.21808493e-01 -3.20450753e-01 -1.16025589e-01 -3.81224126e-01
-1.38633803e-01 1.09796333e+00 5.86367905e-01 -7.28635192e-01
9.85504389e-01 9.22107399e-01 -2.97906071e-01 -1.40427515e-01
-1.04133236e+00 1.74172014e-01 -6.99919641e-01 6.21769130e-01
2.26570264e-01 -9.68213677e-02 -3.86942059e-01 3.57695341e-01
-4.69637692e-01 2.53781587e-01 7.24524677e-01 8.75929236e-01
-3.99358571e-01 -9.34916675e-01 5.45525998e-02 5.95129728e-01
1.45698607e-01 -3.59827727e-01 8.69203266e-03 7.60002971e-01
-2.15308573e-02 4.61524904e-01 5.02494335e-01 -6.57022744e-02
1.57064274e-01 -2.91581452e-01 6.64055467e-01 -5.17059028e-01
-1.47149932e+00 -2.17293687e-02 8.48080516e-02 -4.82753038e-01
2.67283618e-03 -7.32461214e-01 -1.75028396e+00 -9.04352427e-01
-1.99441433e-01 -3.58386710e-02 1.01373291e+00 7.56225944e-01
7.23860085e-01 6.92350030e-01 -8.77921376e-03 -1.33164954e+00
2.22817108e-01 -5.48093379e-01 -7.56160140e-01 -1.32244959e-01
-3.58634233e-01 -6.14084542e-01 -1.80886999e-01 -1.39223129e-01] | [8.134110450744629, 2.6895735263824463] |
aec43131-86d8-44f7-989d-7d0844f0a53c | combining-graph-degeneracy-and-submodularity | null | null | https://aclanthology.org/W17-4507 | https://aclanthology.org/W17-4507.pdf | Combining Graph Degeneracy and Submodularity for Unsupervised Extractive Summarization | We present a fully unsupervised, extractive text summarization system that leverages a submodularity framework introduced by past research. The framework allows summaries to be generated in a greedy way while preserving near-optimal performance guarantees. Our main contribution is the novel coverage reward term of the objective function optimized by the greedy algorithm. This component builds on the graph-of-words representation of text and the k-core decomposition algorithm to assign meaningful scores to words. We evaluate our approach on the AMI and ICSI meeting speech corpora, and on the DUC2001 news corpus. We reach state-of-the-art performance on all datasets. Results indicate that our method is particularly well-suited to the meeting domain. | ['Antoine Tixier', 'Polykarpos Meladianos', 'Michalis Vazirgiannis'] | 2017-09-01 | null | null | null | ws-2017-9 | ['unsupervised-extractive-summarization', 'extractive-document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.85605061e-01 5.10376871e-01 -4.45658714e-01 -2.42935687e-01
-1.37315547e+00 -7.84950912e-01 6.12969995e-01 4.84947473e-01
-3.28578532e-01 7.52418756e-01 1.20671678e+00 -2.40953565e-02
-2.61937052e-01 -2.87131846e-01 -4.69370246e-01 -5.58743000e-01
-2.40194649e-01 6.91766620e-01 -1.57056659e-01 -2.99549133e-01
5.82822144e-01 1.25628293e-01 -1.24026537e+00 5.03298819e-01
9.64565516e-01 4.88840103e-01 1.33274257e-01 9.55884755e-01
4.06709760e-02 5.58102489e-01 -6.82112813e-01 -2.34411523e-01
1.32387757e-01 -3.75919431e-01 -1.04683757e+00 4.84108239e-01
4.15253848e-01 -1.82119250e-01 -4.42356765e-01 7.51774490e-01
5.70417285e-01 3.48398805e-01 8.19561422e-01 -9.28593278e-01
-2.99289078e-01 1.06083965e+00 -4.82187688e-01 4.28390384e-01
5.49632967e-01 -2.38311708e-01 1.95534372e+00 -7.06364930e-01
7.79279470e-01 1.19235015e+00 2.64105171e-01 3.35034221e-01
-1.31943083e+00 1.05580151e-01 4.17624474e-01 -2.33046524e-02
-1.02163708e+00 -7.62224317e-01 5.24415493e-01 -1.50074298e-03
1.45839953e+00 5.89500725e-01 5.77894688e-01 8.23444009e-01
2.63439626e-01 1.37449026e+00 4.57064092e-01 -5.77581882e-01
3.37554067e-01 -2.05193639e-01 4.47083682e-01 4.82269406e-01
6.17597699e-01 -4.59342390e-01 -8.54781330e-01 -5.28531134e-01
-3.92206050e-02 -4.09292698e-01 -4.82058674e-01 -3.31726223e-01
-1.29467797e+00 9.28611577e-01 -1.03915438e-01 2.89705426e-01
-7.75850594e-01 2.67386168e-01 5.35425723e-01 2.26685598e-01
1.06843293e+00 6.68211937e-01 -3.64789695e-01 -4.04731274e-01
-1.33008277e+00 4.76744026e-01 1.29532409e+00 9.25080836e-01
2.92879343e-01 4.88883890e-02 -6.11338794e-01 7.93949962e-01
5.73645830e-02 3.94423544e-01 2.24096864e-01 -1.03686082e+00
8.18324149e-01 3.89183313e-01 4.56424914e-02 -8.29205036e-01
-4.71490473e-01 -6.28126800e-01 -6.15028679e-01 -6.90654993e-01
-1.80273324e-01 -3.09195906e-01 -6.59360170e-01 1.64319420e+00
1.38501868e-01 -1.53191119e-01 3.45029354e-01 5.71129024e-01
6.18718624e-01 8.89405131e-01 -4.66203153e-01 -7.60178983e-01
1.19199646e+00 -1.16144049e+00 -9.04448509e-01 -2.54673630e-01
8.36314917e-01 -5.18741369e-01 7.10057020e-01 6.36044621e-01
-1.42780864e+00 2.48614401e-01 -1.06517410e+00 -4.82119806e-02
2.45166272e-01 1.16747454e-01 4.38981295e-01 5.14878631e-01
-1.35250938e+00 4.44413126e-01 -6.62316263e-01 -5.02486110e-01
3.18446785e-01 2.59880573e-01 -1.00244910e-01 -1.79996490e-02
-6.76219285e-01 6.87710881e-01 6.49204731e-01 -3.96571279e-01
-7.43182600e-01 -6.63869739e-01 -7.58117437e-01 3.16245884e-01
6.75940454e-01 -9.84171271e-01 1.52241421e+00 -5.86912453e-01
-1.35896862e+00 6.91635728e-01 -2.84216583e-01 -8.85694325e-01
3.35677028e-01 -3.92483264e-01 -9.03476328e-02 6.18374348e-01
7.73360580e-02 3.90041798e-01 5.57267308e-01 -1.16367948e+00
-6.84683144e-01 -2.47347280e-01 -2.46901497e-01 5.65851748e-01
-7.49695659e-01 9.42008942e-03 -4.45811719e-01 -8.46126080e-01
3.50127518e-02 -8.60550582e-01 -3.82414043e-01 -9.04006124e-01
-8.11334670e-01 -3.29518557e-01 3.76911253e-01 -9.34509277e-01
1.79812872e+00 -1.73479974e+00 6.51331723e-01 2.34670177e-01
3.19589615e-01 -1.71695441e-01 -2.51738727e-01 1.12325609e+00
2.06695452e-01 5.37678897e-02 -3.82146955e-01 -7.04230368e-01
1.49346024e-01 1.76514402e-01 -4.77086335e-01 6.13276064e-01
-1.31306693e-01 8.44518304e-01 -7.44549870e-01 -4.35267389e-01
-2.67403185e-01 -1.73601747e-01 -8.52841735e-01 1.31760463e-02
-5.18272638e-01 -3.41186881e-01 -5.13648033e-01 4.77722168e-01
2.35248283e-01 -1.54171169e-01 3.77164602e-01 7.35390931e-03
-1.10459626e-01 5.93367279e-01 -1.07059979e+00 2.05879211e+00
-1.36561662e-01 5.63062966e-01 2.58231163e-01 -1.16555810e+00
5.77325046e-01 3.51454914e-01 6.74881220e-01 -9.10231471e-03
1.59910142e-01 5.47040924e-02 -3.94026279e-01 -2.95345336e-01
1.20161355e+00 1.58377081e-01 -3.70346695e-01 7.80967116e-01
9.38834324e-02 -1.65362522e-01 6.31306112e-01 9.51947749e-01
1.41484499e+00 -3.83769363e-01 5.65114856e-01 -7.53880024e-01
3.06525767e-01 2.52970099e-01 2.62696773e-01 9.03895199e-01
1.87366560e-01 7.55389631e-01 7.96307325e-01 -4.19280417e-02
-1.10388935e+00 -8.44035983e-01 1.92987502e-01 1.22635055e+00
-2.90349662e-01 -8.01160395e-01 -9.07010615e-01 -7.12414503e-01
-1.26861393e-01 1.22741473e+00 -5.24503171e-01 -1.27378613e-01
-5.29833257e-01 -7.32572019e-01 4.79844511e-01 3.44376147e-01
-3.60792242e-02 -7.75840700e-01 -4.51926410e-01 3.46130580e-01
-6.04521334e-01 -1.29442167e+00 -1.01094866e+00 1.38692677e-01
-9.84485507e-01 -7.83865750e-01 -7.29320288e-01 -7.02349246e-01
5.62618554e-01 5.00556409e-01 1.32832980e+00 -2.19610423e-01
-1.15679771e-01 9.77742493e-01 -5.39595842e-01 -3.80733371e-01
-5.88834584e-01 5.40762424e-01 8.05805847e-02 8.41763839e-02
4.12241109e-02 -4.02741373e-01 -3.87807757e-01 -3.21086526e-01
-1.01647580e+00 -3.40528972e-02 5.11740565e-01 7.00132310e-01
4.90609258e-01 -2.43508250e-01 8.38704586e-01 -8.68564785e-01
1.09363902e+00 -4.88169074e-01 -1.76654801e-01 2.62447029e-01
-7.37843037e-01 4.03601348e-01 3.67842197e-01 -8.49440787e-03
-8.90909553e-01 -6.37849048e-02 2.81408895e-02 3.68703268e-02
3.09519708e-01 7.83535063e-01 -8.36222917e-02 6.19835138e-01
6.71589494e-01 4.26310033e-01 -1.40183911e-01 -3.69391710e-01
5.92795551e-01 9.81042683e-01 5.63006282e-01 -5.63354790e-01
4.30883110e-01 6.26012564e-01 -8.92963186e-02 -1.25857949e+00
-9.14413750e-01 -9.39578772e-01 -2.26470783e-01 -5.47405295e-02
6.55519426e-01 -9.90196705e-01 -2.85269588e-01 1.09956441e-02
-1.22112906e+00 -3.02618295e-02 -6.50193512e-01 3.29429477e-01
-7.30310082e-01 7.93624341e-01 -5.58128893e-01 -8.55993688e-01
-1.04030573e+00 -5.49641192e-01 1.36817729e+00 -6.67621195e-02
-3.29619527e-01 -8.63574982e-01 5.11744320e-01 3.37251186e-01
7.36144260e-02 1.35404885e-01 7.90699840e-01 -1.24541926e+00
-2.90581435e-01 -2.64838845e-01 1.28922820e-01 2.58237213e-01
-1.34375080e-01 -1.75224796e-01 -7.21534491e-01 -5.96797347e-01
-1.32793680e-01 -2.14698732e-01 1.55599117e+00 6.73513472e-01
6.93153441e-01 -7.59612083e-01 -3.18715513e-01 2.13476703e-01
1.24295485e+00 -3.71661544e-01 2.90639579e-01 2.40189046e-01
5.17849922e-01 6.87762141e-01 4.65029657e-01 8.50749671e-01
5.85138559e-01 5.53196728e-01 2.34374002e-01 4.72383589e-01
1.32133320e-01 -1.98235914e-01 5.83891630e-01 1.29120338e+00
1.85678825e-01 -8.91178131e-01 -7.22616076e-01 1.06934083e+00
-2.28197312e+00 -1.07245374e+00 -2.01437145e-01 2.02053547e+00
7.27051198e-01 1.12412207e-01 4.43615079e-01 1.25536159e-01
5.41145802e-01 4.96538460e-01 -1.63677707e-01 -7.68802404e-01
-2.83679396e-01 7.11298734e-02 5.15618503e-01 6.14047527e-01
-9.07064736e-01 8.93918395e-01 6.83963203e+00 8.47434998e-01
-3.93219203e-01 7.62104765e-02 4.62745786e-01 -6.69007897e-01
-6.93427384e-01 -8.05641264e-02 -6.71353281e-01 -1.39371934e-03
1.19500184e+00 -8.41275632e-01 4.83444422e-01 6.58466458e-01
4.09387946e-01 -7.14918599e-02 -1.06065500e+00 7.32291400e-01
4.87117052e-01 -1.69425488e+00 1.17494531e-01 2.98962146e-01
1.08252728e+00 2.04220772e-01 -1.65976077e-01 9.66903269e-02
4.32335466e-01 -6.77698731e-01 5.71669638e-01 2.03643993e-01
6.06075466e-01 -1.00644588e+00 5.03912330e-01 5.08177817e-01
-9.59286392e-01 -8.22427124e-02 -3.06099772e-01 6.43592179e-02
3.33959341e-01 8.69306266e-01 -1.11573529e+00 9.14272428e-01
3.14001203e-01 7.59818494e-01 -2.14238584e-01 7.77948678e-01
-1.34870246e-01 8.96593630e-01 -3.92888725e-01 -3.20842534e-01
3.20455611e-01 -7.96231255e-02 1.20384562e+00 1.69696319e+00
1.79260403e-01 1.35704756e-01 3.93694222e-01 3.17587078e-01
-5.06682158e-01 4.76477146e-01 -6.66131496e-01 -1.40247419e-01
3.66782606e-01 1.21669602e+00 -5.58495522e-01 -4.20241296e-01
-6.37764335e-02 8.80562782e-01 2.46832177e-01 3.11377525e-01
-2.80006409e-01 -3.43254000e-01 3.08243632e-01 -1.21591635e-01
6.67506278e-01 -2.39458546e-01 -5.47589660e-01 -1.29190230e+00
2.64812648e-01 -9.84698713e-01 6.62420154e-01 -2.58935004e-01
-9.54859436e-01 4.17928636e-01 2.58274049e-01 -8.31927598e-01
-6.14560068e-01 -8.42784196e-02 -6.77790701e-01 4.33417499e-01
-1.38329804e+00 -7.44349718e-01 8.05128962e-02 2.04387888e-01
8.53495657e-01 -1.96685106e-01 6.72598243e-01 -2.08152384e-01
-5.01402438e-01 5.73194087e-01 6.57395124e-01 -4.23467189e-01
4.63062197e-01 -1.49263728e+00 6.00027323e-01 1.05130219e+00
3.60382795e-01 2.10649028e-01 1.13623714e+00 -5.87973654e-01
-1.77200270e+00 -8.93689513e-01 1.18038476e+00 -3.03961515e-01
6.26251876e-01 -2.54569948e-01 -4.90538388e-01 6.79400682e-01
6.37861133e-01 -6.62075043e-01 6.62905395e-01 2.07407296e-01
-1.31971195e-01 1.47824734e-01 -9.28968012e-01 6.16208434e-01
1.07263553e+00 -6.93126172e-02 -1.02194333e+00 6.57449305e-01
8.82870674e-01 -4.08034116e-01 -4.86379206e-01 8.61421749e-02
2.51651525e-01 -5.64532161e-01 5.54165483e-01 -7.42389262e-01
7.64035881e-01 2.07820728e-01 -4.08353984e-01 -1.63020277e+00
-3.36736679e-01 -1.15298772e+00 -6.02209449e-01 1.15448248e+00
6.34976625e-01 -4.29474443e-01 7.18820870e-01 2.89346784e-01
-5.15397847e-01 -7.49936223e-01 -1.07344747e+00 -8.04727435e-01
-6.42954484e-02 -3.47835630e-01 1.48964822e-01 4.65910554e-01
4.27346975e-01 8.17050040e-01 -4.99246210e-01 -5.60739264e-02
8.47930908e-01 2.17411146e-01 7.70061970e-01 -1.16370547e+00
-5.17559826e-01 -5.51987112e-01 -1.83481872e-02 -1.26910067e+00
2.64990538e-01 -1.12882328e+00 1.56158388e-01 -1.92532074e+00
4.92554486e-01 3.99412602e-01 -6.60905093e-02 2.44323105e-01
-1.40896678e-01 -2.01213792e-01 2.96633393e-01 8.58478546e-02
-1.01732790e+00 7.37390876e-01 7.48106062e-01 -2.76561081e-01
-4.71644551e-01 -3.47540565e-02 -1.23559463e+00 3.66000473e-01
8.00416231e-01 -6.08451724e-01 -3.09663504e-01 -4.31009054e-01
2.56689161e-01 1.78196877e-01 -1.63796991e-01 -4.62866008e-01
3.78888100e-01 -8.25132709e-03 -3.74278784e-01 -9.19525564e-01
1.31623074e-01 -3.25859189e-01 -4.26841468e-01 3.66233945e-01
-6.49009049e-01 8.55210572e-02 1.67611063e-01 9.35238361e-01
-9.32266638e-02 -3.89881790e-01 3.71176809e-01 1.33016944e-01
-1.48214936e-01 3.52242240e-03 -4.47348803e-01 6.51405513e-01
8.07422698e-01 3.03469896e-02 -3.58520716e-01 -7.04842269e-01
-2.79706717e-01 5.58629274e-01 2.72589356e-01 2.87437856e-01
5.33083677e-01 -1.01354229e+00 -1.51070428e+00 -3.39329362e-01
1.32416308e-01 -1.54734701e-01 1.53533906e-01 8.54609132e-01
-2.62225538e-01 6.98597670e-01 3.80658597e-01 -4.67525154e-01
-1.34411359e+00 4.11561728e-01 -1.96919724e-01 -8.70208681e-01
-8.41256261e-01 4.99750286e-01 -1.51528800e-02 -2.05370076e-02
2.99918085e-01 -3.53251845e-01 -1.76300794e-01 3.65592003e-01
5.64652681e-01 7.12463915e-01 1.85763836e-01 -3.79798532e-01
-3.61141354e-01 2.15010643e-02 -3.90565187e-01 -6.33556366e-01
1.70250726e+00 -2.97299713e-01 -1.11589827e-01 2.87769943e-01
1.18455279e+00 3.04310501e-01 -7.35324383e-01 -3.03512216e-01
2.41382733e-01 -1.67675351e-03 3.04501146e-01 -5.61232984e-01
-6.58613205e-01 2.21261680e-01 -3.19984496e-01 6.22942030e-01
1.04356730e+00 1.04948089e-01 9.05447066e-01 7.58881986e-01
3.14443335e-02 -1.47851515e+00 9.19471160e-02 6.33780777e-01
1.07757497e+00 -7.21817553e-01 4.82266665e-01 -2.23782405e-01
-8.63736808e-01 9.93174613e-01 -2.03561291e-01 -1.95689201e-01
2.64235020e-01 2.02750534e-01 -4.45556968e-01 -2.74249643e-01
-1.29633605e+00 -5.94675131e-02 5.16015589e-01 1.43874258e-01
2.66019583e-01 8.19004029e-02 -7.29228079e-01 7.28878260e-01
-2.24177256e-01 -3.73032451e-01 9.32421327e-01 9.86072004e-01
-8.23525906e-01 -9.06343460e-01 -2.66139954e-01 7.17091084e-01
-6.60243928e-01 -3.50582361e-01 -6.85631514e-01 3.72973114e-01
-9.86762404e-01 1.38446832e+00 -1.79130524e-01 -2.70545721e-01
4.94524837e-01 1.64826944e-01 2.83926904e-01 -7.78969347e-01
-8.04515541e-01 3.72843653e-01 6.80841506e-01 -3.63054484e-01
-2.58724928e-01 -1.04074514e+00 -1.28352427e+00 -3.65706146e-01
-3.70080650e-01 4.17478472e-01 6.11538410e-01 9.06410158e-01
5.96907079e-01 6.64583385e-01 8.55197906e-01 -8.11723471e-01
-1.09555507e+00 -9.85562265e-01 -6.73254371e-01 1.16266951e-01
3.21216971e-01 4.96530645e-02 -4.52628732e-01 -6.12816662e-02] | [12.512043952941895, 9.501816749572754] |
79fa3610-3fa4-4d15-8f86-ec117c2107a7 | mathematical-model-for-transmission-dynamics | 2304.01975 | null | https://arxiv.org/abs/2304.01975v1 | https://arxiv.org/pdf/2304.01975v1.pdf | Mathematical Model for Transmission Dynamics of Tuberculosis in Burundi | Tuberculosis (TB) is among the main public health challenges in Burundi. The literature lacks mathematical models for key parameter estimates of TB transmission dynamics in Burundi. In this paper, the supectible-exposed-infected-recovered (SEIR) model is used to investigate the transmission dynamics of tuberculosis in Burundi. Using the next generation method, we calculated the basic reproduction number, R0. The model is demonstrated to have a disease-free equilibrium (DEF) that is locally and globally asymptotically stable. When the corresponding reproduction threshold quantity approaches unity, the model enters an endemic equilibrium (EE). That means, the disease can be controlled through different interventions in Burundi. A sensitivity analysis of the model parameters was also investigated. It shows that the progression rate from latent to becoming infectious had the highest positive sensitivity, which means that R0 increases and decreases proportionally with an increase and a decrease of that progression rate. | ['David Niyukuri', 'Paterne Gahungu', 'Kelly Joelle Gatore Sinigirira', 'Steve Sibomanaa'] | 2023-04-04 | null | null | null | null | ['unity'] | ['computer-vision'] | [ 2.66212583e-01 -2.67532200e-01 -2.47050494e-01 3.94588441e-01
-6.74252287e-02 -1.03722796e-01 5.67247510e-01 -8.70841295e-02
-1.27548471e-01 1.18983734e+00 -1.50583178e-01 -5.74788570e-01
-6.70791805e-01 -7.50864744e-01 -1.30581006e-01 -1.30069005e+00
-4.10869658e-01 8.44608486e-01 2.34530807e-01 -3.13966069e-03
-1.78073987e-01 5.22412062e-01 -7.74477959e-01 -2.95995027e-01
9.71276104e-01 1.65172527e-03 5.84708512e-01 7.13193238e-01
2.20519662e-01 4.46726501e-01 -8.68605673e-01 2.09108561e-01
-3.43017802e-02 -5.97723663e-01 -6.40794337e-01 -2.28645802e-01
-7.18267381e-01 -6.01422787e-01 -1.45896345e-01 4.53583360e-01
4.61005479e-01 -6.99943900e-01 1.22352004e+00 -1.16843319e+00
-4.31400359e-01 1.06216788e-01 -7.74393797e-01 5.19025087e-01
-1.12393714e-01 8.21825489e-02 7.83213675e-02 -4.23056066e-01
6.63042724e-01 1.46624374e+00 7.31985688e-01 6.25535190e-01
-1.23525167e+00 -5.52163005e-01 -6.27925754e-01 -3.64639759e-01
-1.44629157e+00 9.77153704e-02 -9.34829786e-02 -6.76564336e-01
9.28694308e-01 2.73019373e-01 1.06003726e+00 6.30876064e-01
1.07518411e+00 1.03218019e-01 1.20195580e+00 -3.77082020e-01
-1.89671576e-01 6.62339665e-03 7.22205825e-03 2.42301688e-01
1.06558108e+00 -4.64731362e-03 2.79526204e-01 -6.88228905e-01
1.43019474e+00 4.08106893e-01 -4.35706191e-02 2.57538855e-01
-9.45424974e-01 9.67260778e-01 2.91814506e-02 6.61270797e-01
-6.44441545e-01 3.17198575e-01 -1.24595381e-01 3.68282497e-01
4.65660095e-01 -1.02171153e-01 -2.93982387e-01 1.86690047e-01
-4.65062082e-01 3.37672308e-02 4.29154247e-01 2.70018339e-01
3.04411829e-01 -7.79732019e-02 -4.04615164e-01 6.71008229e-01
5.12621164e-01 1.45631254e+00 -1.49477631e-01 -7.22096562e-01
3.02669793e-01 5.22018969e-01 3.13377202e-01 -6.55674279e-01
-1.31396428e-01 -2.56982505e-01 -1.09997034e+00 -5.56421459e-01
4.44411516e-01 -2.59244561e-01 -6.59896910e-01 1.42549431e+00
4.34510678e-01 -2.79505134e-01 9.55052823e-02 8.14086869e-02
1.18426479e-01 1.19309330e+00 2.15666413e-01 -1.07895911e+00
1.32302511e+00 -2.84666687e-01 -1.03681648e+00 2.16478363e-01
4.92277294e-01 -5.31163216e-01 3.36803734e-01 -5.19537702e-02
-7.95080245e-01 1.71681628e-01 -3.45839649e-01 8.97008061e-01
1.17033236e-01 -3.99225920e-01 1.77604035e-01 6.92173958e-01
-1.09093785e+00 2.63403654e-01 -1.18302393e+00 -8.38411450e-01
1.61482394e-01 4.92807031e-01 -6.41109496e-02 -1.67441204e-01
-1.35155272e+00 8.22010696e-01 -8.18923116e-02 2.43323315e-02
-7.24629998e-01 -6.48419201e-01 -1.43246323e-01 -2.37155512e-01
-1.16520330e-01 -1.06837440e+00 7.57001936e-01 -2.35157907e-01
-8.00199926e-01 5.97276986e-01 -2.26253569e-01 -4.80245203e-02
6.78555012e-01 7.44149387e-02 -4.33848351e-02 2.81764895e-01
2.82286108e-01 9.64406207e-02 3.69098991e-01 -1.20220649e+00
-4.06370431e-01 -7.62495935e-01 -5.02514660e-01 4.46581431e-02
-2.23147288e-01 6.47267580e-01 6.99268701e-03 -4.29939628e-01
-1.20075881e-01 -8.65509987e-01 -3.28688353e-01 -4.09842849e-01
-9.24414396e-02 -4.28467661e-01 8.25715542e-01 -7.92952120e-01
1.57917607e+00 -1.75093567e+00 -1.77794933e-01 1.40231743e-01
1.21113732e-01 2.37899587e-01 2.39934772e-01 1.05250978e+00
2.77613640e-01 2.39249140e-01 -1.90711632e-01 6.20940387e-01
-5.46022117e-01 3.01282108e-01 1.01424195e-01 6.97860777e-01
3.98296416e-01 6.49429798e-01 -1.07431722e+00 -7.31640160e-01
-6.75680041e-02 9.10461903e-01 -4.78266925e-01 1.99365646e-01
4.59752828e-01 7.17923284e-01 -8.62738669e-01 4.35512751e-01
7.52488136e-01 -4.89468008e-01 5.48600435e-01 4.31142509e-01
-3.95391405e-01 -1.39523774e-01 -7.05752075e-01 8.36020708e-03
-3.25316846e-01 1.90974414e-01 -9.90368798e-03 -7.08358824e-01
8.97202671e-01 7.63431489e-01 8.06571066e-01 -3.14864278e-01
-7.11914077e-02 4.89864886e-01 3.53333026e-01 -5.77404916e-01
-4.89294454e-02 -1.38582960e-01 1.66538671e-01 7.66812384e-01
-4.53047186e-01 4.36432511e-02 1.08792171e-01 6.28112704e-02
1.06818950e+00 -3.80090356e-01 5.59814930e-01 -6.66215897e-01
4.27364647e-01 1.21119075e-01 2.61430144e-01 7.24975586e-01
-1.73640496e-04 -2.29666591e-01 5.08640885e-01 1.65609196e-02
-1.02461720e+00 -1.32293284e+00 -6.50555491e-01 2.29903415e-01
-9.59313959e-02 2.75321156e-01 -7.23662198e-01 1.77980646e-01
-1.79446712e-01 3.20160240e-01 -3.99092823e-01 -1.55699670e-01
-8.19555640e-01 -1.38598251e+00 4.00825411e-01 -1.49917886e-01
6.41422153e-01 -1.10640728e+00 -6.96978509e-01 3.24358165e-01
-3.73391360e-01 -1.20434888e-01 -1.15921080e-01 2.28186511e-02
-1.39859188e+00 -1.15095210e+00 -1.46269262e+00 -5.56853533e-01
9.10715222e-01 2.73989379e-01 5.94668806e-01 2.75802642e-01
-5.01374424e-01 4.41983134e-01 2.18017563e-01 -1.57336727e-01
-9.28921819e-01 -2.72188902e-01 1.63469389e-01 -5.18150985e-01
1.05161980e-01 -9.68573987e-02 -9.37114120e-01 5.77354550e-01
-8.33704948e-01 -2.94130296e-01 3.22905779e-01 7.98749328e-01
4.73449230e-01 4.00201857e-01 9.76397455e-01 -4.92378980e-01
7.96645105e-01 -8.19402635e-01 -4.19912845e-01 5.48201859e-01
-5.59439003e-01 -3.44082415e-01 3.80057469e-02 -5.36027193e-01
-1.10124111e+00 -3.29792082e-01 3.31141412e-01 2.38940790e-01
2.14860201e-01 2.33542830e-01 2.36822963e-01 5.20176411e-01
5.05628847e-02 3.14204484e-01 1.68697178e-01 -3.76570374e-01
-4.34950918e-01 1.31842279e+00 -8.65687057e-02 -1.54533982e-01
5.11072516e-01 3.89411956e-01 2.87527740e-01 -1.12078428e+00
-1.26564026e-01 -4.63680178e-01 -3.41035128e-01 -4.45390165e-01
6.55798614e-01 -7.37931967e-01 -9.31444585e-01 1.07483411e+00
-1.01022077e+00 -6.09822333e-01 2.75118407e-02 8.01180840e-01
-5.72485209e-01 7.57346861e-03 -7.79891014e-01 -1.43407798e+00
-5.10434031e-01 -7.09301233e-01 6.67867780e-01 2.26539536e-03
-1.74555451e-01 -1.37201524e+00 7.34326601e-01 2.54923642e-01
6.96650028e-01 2.66374230e-01 1.07233608e+00 -2.04781651e-01
-4.23047572e-01 6.05335273e-02 -2.99921036e-01 -7.53098130e-02
8.37300241e-01 3.75237793e-01 -1.10291187e-02 -5.44990003e-01
5.85087180e-01 1.88087583e-01 4.48795408e-01 8.96562517e-01
7.77993798e-02 -8.87397408e-01 -7.90275991e-01 4.22434621e-02
1.63733792e+00 8.32646132e-01 7.94849515e-01 2.88118482e-01
1.36074319e-01 7.57551372e-01 6.18560135e-01 5.15001655e-01
2.50960439e-01 4.31906730e-01 2.08461676e-02 -1.58676863e-01
1.70136198e-01 -8.55883434e-02 3.28809470e-01 9.99052703e-01
-5.71802735e-01 -4.65946704e-01 -1.30903709e+00 6.78291023e-01
-1.44787049e+00 -1.02757299e+00 -6.78342998e-01 2.40434504e+00
1.01594138e+00 -3.29980612e-01 1.60604760e-01 -1.27499670e-01
9.98634040e-01 -5.40471554e-01 -1.60601437e-01 -3.64265800e-01
1.71918616e-01 9.96451229e-02 5.26692867e-01 5.53720832e-01
-4.33931112e-01 4.09732044e-01 7.65903664e+00 7.76029229e-01
-9.76044476e-01 1.77969232e-01 7.53410161e-01 5.29687464e-01
-1.39377341e-01 -1.16008237e-01 -7.54931688e-01 4.45103437e-01
1.23931992e+00 -2.28185743e-01 1.49970546e-01 -5.16621955e-02
8.61493826e-01 -5.12022674e-01 -3.84003639e-01 2.38683566e-01
-5.34455121e-01 -7.17315853e-01 2.27087811e-01 6.62888944e-01
8.26305270e-01 -1.57930493e-01 -1.35625243e-01 -1.81057870e-01
3.57899487e-01 -7.47855246e-01 -4.01201248e-01 7.19193161e-01
9.63983655e-01 -9.08270001e-01 9.98560905e-01 6.38827682e-01
-9.97997165e-01 2.29092352e-02 -2.38381505e-01 6.76438734e-02
2.69572586e-01 4.08257455e-01 -1.02687180e+00 4.03137773e-01
4.38816249e-01 2.04944625e-01 -1.79834768e-01 9.75928843e-01
-1.19984992e-01 8.01949382e-01 -4.37757552e-01 -1.22960232e-01
-7.24365562e-02 -4.21240360e-01 3.82927507e-01 7.86889970e-01
6.67474449e-01 1.03041820e-01 -3.55476707e-01 7.98525453e-01
6.81200266e-01 3.84026505e-02 -7.80331433e-01 -1.94170102e-01
5.66961765e-01 4.93522853e-01 -7.78113067e-01 -1.49655253e-01
1.91407695e-01 7.10588574e-01 -5.34312487e-01 4.05515343e-01
-6.71708107e-01 -2.11153045e-01 6.37752265e-02 5.45733929e-01
3.85605991e-02 4.81811576e-02 4.57377404e-01 -6.21211588e-01
-5.82201004e-01 -2.92707413e-01 3.11275721e-01 -5.82772076e-01
-1.01537693e+00 3.32636684e-01 6.24066174e-01 -6.31600559e-01
-4.13972348e-01 -1.08003557e-01 -7.46937156e-01 9.80292618e-01
-9.64866579e-01 -9.13634717e-01 3.71059507e-01 6.33603454e-01
8.39860290e-02 3.13208431e-01 7.17720091e-01 2.45914117e-01
-6.55412376e-01 1.40267819e-01 8.98466408e-01 -4.09942716e-01
1.19177632e-01 -5.98653913e-01 -1.42204449e-01 3.00990373e-01
-1.14832151e+00 1.09253645e+00 5.54539919e-01 -1.38406229e+00
-1.04322076e+00 -1.07639790e+00 1.48381066e+00 -3.73820476e-02
9.39831793e-01 -1.53154051e-02 -5.50591767e-01 4.22022790e-01
1.52001202e-01 -7.01651752e-01 3.77992094e-01 -7.27034807e-01
4.52434361e-01 3.13722640e-02 -1.39678264e+00 6.22548223e-01
4.79632646e-01 6.29499406e-02 -2.71916360e-01 5.92684209e-01
5.06053627e-01 5.25481761e-01 -1.16772664e+00 4.11294878e-01
5.07323503e-01 -4.18420285e-01 1.24371135e+00 -3.10023397e-01
1.65291112e-02 1.25717431e-01 1.43477708e-01 -7.92749226e-01
-3.33219558e-01 -4.72597122e-01 1.40371442e-01 1.14556789e+00
2.43962388e-02 -1.09764898e+00 6.49374798e-02 -2.08829626e-01
8.06340635e-01 -9.13829863e-01 -1.24881482e+00 -1.19719982e+00
2.87872404e-01 6.20960951e-01 3.49569112e-01 6.06121659e-01
-3.86290312e-01 2.43826985e-01 -4.97366905e-01 4.73996103e-02
6.78801060e-01 -3.73633623e-01 4.84813035e-01 -1.19785368e+00
4.46985252e-02 1.61694273e-01 5.08006439e-02 -5.42962968e-01
-7.49574602e-01 -3.39948177e-01 -2.12128371e-01 -2.01596546e+00
1.06671333e+00 -6.99482560e-01 -1.16471812e-01 2.11852744e-01
-1.72962189e-01 7.67170340e-02 -6.33430555e-02 9.66470122e-01
1.08880527e-01 3.16874385e-01 1.54593682e+00 9.90712177e-03
-6.29593670e-01 5.81681430e-01 -1.04133256e-01 3.40344340e-01
1.15924346e+00 -1.00554347e+00 -3.79593223e-01 8.11867707e-04
2.07009520e-02 6.40368998e-01 4.40440327e-01 -3.16201597e-01
-3.76460463e-01 -7.07157552e-01 -1.47546023e-01 -1.05476367e+00
6.38173074e-02 -8.36436152e-01 1.08134341e+00 1.73999131e+00
3.64041120e-01 -3.42704728e-02 -1.83168545e-01 4.35906619e-01
5.37245452e-01 -3.31628442e-01 7.52662003e-01 1.86326236e-01
5.93588233e-01 1.30475312e-01 -1.15427959e+00 -1.23004302e-01
1.16844797e+00 -4.32327032e-01 -4.89171445e-01 -1.13622956e-01
-5.66403508e-01 1.41895488e-01 3.22942853e-01 -2.50531703e-01
3.53928119e-01 -1.27327633e+00 -1.00480306e+00 -1.75084144e-01
-3.86801034e-01 -3.54577422e-01 2.68273652e-01 1.22929060e+00
-9.04230177e-01 7.98591495e-01 -1.39706850e-01 -8.80047619e-01
-1.39389098e+00 3.67886037e-01 5.90265751e-01 -9.49687064e-01
-1.37817785e-01 -6.12990139e-03 3.74411970e-01 -3.75061899e-01
-2.59649098e-01 2.51243740e-01 -2.33609661e-01 -5.22210971e-02
4.56703335e-01 9.24861550e-01 -5.88544667e-01 -7.79003620e-01
-6.33794725e-01 6.61265612e-01 3.24531168e-01 8.27515274e-02
1.47381794e+00 -4.32401359e-01 -7.91186571e-01 6.38837814e-01
8.09592068e-01 -1.15418576e-01 -8.00410390e-01 1.90096229e-01
-2.59203941e-01 -3.25004041e-01 -3.89740795e-01 -6.40843868e-01
-6.98497355e-01 5.65233946e-01 8.14105213e-01 4.18315917e-01
8.22731674e-01 2.01672137e-01 2.25032672e-01 2.68642157e-01
3.15542489e-01 -3.89523000e-01 8.21618736e-02 3.21664929e-01
8.21545362e-01 -6.72428071e-01 -1.06578227e-02 -4.27178264e-01
-8.10569972e-02 5.53631544e-01 2.19785243e-01 7.91872218e-02
7.87830949e-01 1.26592413e-01 -2.11535707e-01 -2.77571499e-01
-8.55195224e-01 2.70954460e-01 -3.05963099e-01 9.26107705e-01
4.51170266e-01 2.20480412e-01 -1.09615278e+00 6.50793537e-02
3.39640498e-01 1.19952694e-01 4.37908471e-01 7.31188893e-01
-6.40168250e-01 -1.04732955e+00 -9.48463619e-01 5.58179915e-01
-7.75996327e-01 -1.44086346e-01 -4.03475523e-01 1.27129078e+00
-1.03602268e-01 1.07324636e+00 5.86218953e-01 4.71625268e-01
-1.41340181e-01 -3.62535477e-01 5.09414852e-01 -3.05659562e-01
-4.80205804e-01 5.26272297e-01 -3.83024722e-01 3.26094061e-01
-6.81520939e-01 -5.19213855e-01 -1.29343498e+00 -6.40388191e-01
-8.72054756e-01 3.84924680e-01 4.38760489e-01 5.27570784e-01
-3.39120254e-02 3.59378874e-01 1.06391728e+00 -7.55158141e-02
-4.01960790e-01 -1.03410494e+00 -7.87212610e-01 -1.32425696e-01
2.31125951e-01 -4.08296525e-01 -4.73227531e-01 -1.32347152e-01] | [5.924240589141846, 4.413984775543213] |
0c19ccd5-4bed-4be9-8765-4581c5ee29df | deciphering-the-interaction-of-genetic-and | 2210.11323 | null | https://arxiv.org/abs/2210.11323v2 | https://arxiv.org/pdf/2210.11323v2.pdf | Bottom-up data integration in polymer models of chromatin organisation | Cellular functions crucially depend on the precise execution of complex biochemical reactions taking place on the chromatin fiber in the tightly packed environment of the cell nucleus. Despite the availability of large data sets probing this process from multiple angles, we still lack a bottom-up framework which can incorporate the sequence-specific nature of biochemistry in a unified model of 3D chromatin dynamics. Here we propose SEMPER (Sequence Enhanced Magnetic PolymER), a novel stochastic polymer model which naturally incorporates observational data about sequence-driven biochemical processes, such as binding of transcription factor proteins, in a 3D model of chromatin structure. By introducing a new algorithm for approximate Bayesian inference, we discuss how to estimate in a robust manner the relative importance of biochemical vs. polymer signals in the determination of the chromatin epigenetic states which is leading to a significant revision of the interpretation of previous models. Furthermore we show that, without additional input from the genome 3D structure, our model can predict with reasonable accuracy some notable and non trivial conformational features of chromatin folding within the nucleus. Our work highlights the importance of introducing physically realistic statistical models for predicting chromatin states from epigenetic data, and opens the way to a new class of more systematic approaches to interpret epigenomic data. | ['Guido Sanguinetti', 'Angelo Rosa', 'Alex Chen Yi Zhang'] | 2022-10-20 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [ 4.81273502e-01 -3.44163179e-02 -3.64839770e-02 -1.12668268e-01
-3.80779743e-01 -8.33593607e-01 7.09181786e-01 3.49549711e-01
-4.66172636e-01 8.37343931e-01 4.73341912e-01 -4.22152966e-01
-3.20866466e-01 -6.58577383e-01 -9.40702975e-01 -1.24319446e+00
2.36125588e-01 9.84916985e-01 5.09061933e-01 4.14408781e-02
2.79902965e-01 6.85116112e-01 -1.15089786e+00 -1.71381772e-01
4.24872249e-01 4.01903361e-01 3.02294642e-01 1.06671667e+00
-4.18513715e-02 9.31385010e-02 -2.53738035e-02 -3.11810404e-01
-1.57590955e-01 -5.88749945e-01 -5.56421757e-01 7.15536326e-02
-2.40299068e-02 2.15748817e-01 -8.21421891e-02 7.71318197e-01
6.25569642e-01 -6.66697323e-02 9.94412661e-01 -3.30088794e-01
-4.30600256e-01 2.09583029e-01 -2.49833226e-01 2.18160838e-01
2.16353297e-01 2.67512381e-01 7.66637206e-01 -6.42608166e-01
1.10218894e+00 1.10452735e+00 5.65125883e-01 2.76624948e-01
-1.70877278e+00 1.56221077e-01 -3.58846277e-01 -1.72436029e-01
-1.09999704e+00 -3.19313496e-01 4.94402409e-01 -6.18560672e-01
7.03742623e-01 4.42645818e-01 8.16747606e-01 1.00232935e+00
7.23035276e-01 -5.21137491e-02 1.46195829e+00 -4.69920605e-01
5.01910746e-01 -5.42349339e-01 3.73606801e-01 6.03829265e-01
5.60137093e-01 5.10734431e-02 -6.67075455e-01 -4.61359859e-01
4.28805321e-01 -6.28952235e-02 -1.04956783e-01 -5.61403632e-01
-1.34212947e+00 2.69675732e-01 -3.38040769e-01 3.61755818e-01
-1.31902948e-01 1.14671979e-02 3.49161446e-01 -3.44065249e-01
1.01841189e-01 2.16930315e-01 -6.09199286e-01 -3.70278269e-01
-8.30471158e-01 2.91594356e-01 7.13918567e-01 2.99831748e-01
6.89546287e-01 -7.09531307e-01 1.13301352e-01 1.78415686e-01
5.99971056e-01 7.96491861e-01 2.39824820e-02 -1.13431597e+00
-2.51339376e-01 1.79298788e-01 4.15417522e-01 -8.41796637e-01
-5.12481153e-01 -3.54134679e-01 -6.07706547e-01 -1.90930992e-01
9.34425771e-01 -7.12713152e-02 -6.70867920e-01 1.68951941e+00
5.96292138e-01 3.29637453e-02 -1.24728553e-01 6.41038835e-01
1.37081474e-01 5.99677920e-01 -1.12641491e-02 -5.32600939e-01
1.53915024e+00 -1.47612274e-01 -5.78387558e-01 2.26821125e-01
3.58249843e-01 -5.51945686e-01 6.87152624e-01 4.31339324e-01
-1.07034850e+00 -1.37015581e-01 -9.03465867e-01 -1.50994644e-01
-2.83626974e-01 -5.65370560e-01 6.44394338e-01 7.74812281e-01
-8.85379732e-01 7.91050434e-01 -1.30881333e+00 -9.19858158e-01
-1.86733887e-01 3.52245986e-01 -3.80266786e-01 4.61083800e-02
-8.59578550e-01 9.97190893e-01 3.50586176e-01 3.75561535e-01
-5.89049160e-01 -3.09744626e-01 -3.22591215e-01 1.23467967e-01
1.24533005e-01 -8.01842093e-01 9.41487491e-01 -7.68213123e-02
-1.55953729e+00 8.14917445e-01 -3.70514333e-01 4.00063097e-02
9.26574618e-02 1.50391102e-01 9.60653946e-02 1.13100924e-01
-4.31371003e-01 2.67660707e-01 7.11840689e-02 -1.00866890e+00
2.19694629e-01 -5.69966018e-01 -5.88576138e-01 -2.47958899e-02
3.04296285e-01 -8.06435868e-02 -2.31928289e-01 -9.56219658e-02
3.78672332e-01 -1.22534454e+00 -3.56662422e-01 -4.99822199e-02
-3.97233129e-01 1.12903215e-01 3.38985801e-01 -4.21246320e-01
6.71159029e-01 -2.08380914e+00 7.22216070e-01 3.72840047e-01
1.43549979e-01 2.06983089e-01 2.89075613e-01 9.76920366e-01
-2.74129696e-02 1.16116859e-01 -3.76502395e-01 1.56739980e-01
2.09644333e-01 2.39107236e-01 -2.06385881e-01 7.24265158e-01
5.90357333e-02 8.43376338e-01 -8.42979372e-01 -1.88504010e-01
9.41106007e-02 7.01289654e-01 -5.40679097e-01 -5.43987425e-03
-5.08480072e-01 6.56772614e-01 -6.43249333e-01 3.57713938e-01
7.29334950e-01 -4.07962084e-01 1.16521943e+00 2.13773534e-01
-4.13833559e-01 2.72238284e-01 -1.11922681e+00 1.46461999e+00
3.90967846e-01 2.90648967e-01 1.30642578e-01 -7.37703621e-01
7.75038719e-01 2.65424013e-01 2.04536244e-01 -2.20314369e-01
4.63372841e-02 2.02080160e-01 9.80066061e-02 -3.48807186e-01
2.08399191e-01 -6.15966916e-01 -9.76038650e-02 5.02454102e-01
-8.92749615e-03 3.00172046e-02 3.94307859e-02 1.62129596e-01
1.10616899e+00 5.24618506e-01 2.85562396e-01 -6.97262347e-01
2.50049859e-01 -2.55137563e-01 6.54615879e-01 7.03475177e-01
-2.25168481e-01 6.75524950e-01 1.09015214e+00 -3.53084683e-01
-1.37265551e+00 -1.13774323e+00 -3.50825995e-01 7.71961749e-01
1.76415965e-02 -4.97231752e-01 -9.40761209e-01 -1.05208144e-01
9.02532041e-03 1.47837207e-01 -6.43330336e-01 -2.00258926e-01
-4.19316858e-01 -1.70626891e+00 5.28463125e-01 9.24796090e-02
-3.11496109e-01 -4.65814024e-01 -2.19581142e-01 2.51141638e-01
4.06810716e-02 -7.48864055e-01 -2.35685334e-01 5.62263370e-01
-9.14003789e-01 -1.21502686e+00 -4.22589302e-01 -2.25806370e-01
6.11125052e-01 4.57326211e-02 5.75377941e-01 4.78394479e-02
-1.06047623e-01 2.36305296e-01 5.08677177e-02 -1.06057100e-01
-6.90694571e-01 -4.64270934e-02 2.42737159e-01 -2.69567162e-01
2.93714732e-01 -8.46343517e-01 -5.65103471e-01 1.23254813e-01
-1.01047373e+00 -3.25273573e-02 3.96323174e-01 5.75422585e-01
6.31001353e-01 -1.39243111e-01 8.88191611e-02 -9.11981404e-01
1.72465503e-01 -4.61022258e-01 -8.08528364e-01 6.69869483e-02
-5.52995741e-01 6.63895726e-01 4.36466157e-01 4.06885371e-02
-1.17199636e+00 1.69710413e-01 -4.44064826e-01 6.18427813e-01
-3.17909479e-01 5.03419876e-01 -4.04334247e-01 1.27220303e-01
2.57660538e-01 4.90577161e-01 2.62564957e-01 -6.16196394e-01
1.06283695e-01 2.15762049e-01 3.93278718e-01 -9.07936633e-01
5.41203856e-01 7.51767576e-01 7.91680753e-01 -9.91219580e-01
-6.41173661e-01 -1.73418716e-01 -8.89777064e-01 -8.67257938e-02
8.98607016e-01 -4.71898407e-01 -1.23617387e+00 5.44266820e-01
-8.76911819e-01 -3.28693300e-01 7.04413056e-02 5.14331579e-01
-7.07819104e-01 8.49714816e-01 -8.93797517e-01 -9.23362195e-01
2.44911179e-01 -1.16633809e+00 1.21076524e+00 1.64158959e-02
-2.09332034e-01 -8.61752987e-01 1.03571999e+00 3.52384180e-01
2.04046026e-01 3.03478926e-01 1.21246827e+00 -3.08519274e-01
-7.16548204e-01 2.06342876e-01 1.66887298e-01 -2.39697814e-01
-2.21538812e-01 7.25682497e-01 -8.75241280e-01 1.12794982e-02
-1.41377389e-01 9.29447860e-02 9.57149565e-01 5.01274705e-01
5.14356315e-01 1.99053884e-01 -5.67454815e-01 3.25223356e-01
1.47747457e+00 -5.83721362e-02 7.60776043e-01 8.37494805e-02
3.36108416e-01 6.34039760e-01 4.29154664e-01 3.01365614e-01
1.41009986e-01 6.82021558e-01 2.02117488e-01 3.32817495e-01
-8.41408223e-03 -2.73391664e-01 4.13604170e-01 1.02022660e+00
-2.57972807e-01 -3.72119009e-01 -1.08112001e+00 8.29776004e-02
-1.73175573e+00 -1.07858956e+00 -5.44653237e-01 2.31561756e+00
1.04567039e+00 3.89552355e-01 1.53631762e-01 -1.42363897e-02
8.14217031e-01 2.18891501e-02 -2.48104513e-01 -3.19781452e-01
-2.35427901e-01 2.07349956e-01 4.95673031e-01 8.94918442e-01
-6.21651709e-01 3.17796558e-01 7.67257547e+00 4.11867082e-01
-1.08633685e+00 -1.15837948e-02 3.79235715e-01 1.56515643e-01
-4.86536592e-01 4.76935238e-01 -1.01419556e+00 6.92076862e-01
1.44458246e+00 2.21864671e-01 2.33943731e-01 9.12479982e-02
6.26405954e-01 -4.99787986e-01 -8.75281155e-01 2.53225684e-01
-4.92154658e-01 -1.59316707e+00 -3.64598125e-01 6.03033006e-01
5.85877955e-01 1.22143030e-01 -2.05313832e-01 -4.46550727e-01
1.86152771e-01 -6.95231736e-01 4.68758106e-01 1.26038253e+00
3.80394459e-01 -5.18988132e-01 6.29948676e-01 6.53307021e-01
-6.12769365e-01 3.99986088e-01 -3.70570987e-01 -2.54234493e-01
1.94398105e-01 1.20310295e+00 -8.55152845e-01 1.57182559e-01
1.72234491e-01 2.49698356e-01 -2.65234977e-01 6.61154807e-01
1.48348451e-01 7.07304835e-01 -4.87375259e-01 -2.15665668e-01
-9.31792185e-02 -6.15734518e-01 4.76066619e-01 1.23406732e+00
3.60880792e-01 1.97051018e-01 -3.48075420e-01 5.43743730e-01
8.32416341e-02 -2.65570372e-01 -4.03381079e-01 -4.29688692e-01
1.80373520e-01 1.18442714e+00 -1.20363903e+00 4.20527309e-02
-1.51865020e-01 5.93946099e-01 4.65123862e-01 1.18139006e-01
-5.32142997e-01 6.27320111e-02 6.54485464e-01 3.34917039e-01
4.78968292e-01 -8.20562005e-01 1.02361061e-01 -1.01598263e+00
-1.05538815e-01 -5.33662915e-01 -5.99054014e-03 -6.26030684e-01
-7.68653035e-01 -3.94963443e-01 -4.03459907e-01 -2.22596556e-01
-3.40572260e-02 -7.66091764e-01 -2.87250549e-01 7.25898504e-01
-1.16706264e+00 -5.75850487e-01 3.58702570e-01 -3.35752696e-01
-3.42659473e-01 7.50724494e-01 8.69952023e-01 -6.64374083e-02
-6.14274144e-01 -4.21907842e-01 9.97639239e-01 -3.20527405e-01
9.44738269e-01 -1.49687541e+00 4.50802594e-01 6.42043054e-01
-2.69085735e-01 9.81637836e-01 1.25444353e+00 -9.08607781e-01
-1.93749797e+00 -7.39820421e-01 1.09558582e+00 -1.05261958e+00
6.13312721e-01 -8.34887028e-01 -8.28351676e-01 5.70816457e-01
1.17898710e-01 -6.86115474e-02 1.14421701e+00 2.04411134e-01
-3.61792408e-02 1.55122533e-01 -9.45234597e-01 4.52265799e-01
7.76654005e-01 -4.97834682e-01 -6.31497145e-01 3.31684440e-01
4.74462986e-01 -1.38117224e-01 -1.17432821e+00 1.55445591e-01
8.04761946e-01 -1.10709488e+00 6.55249059e-01 -5.79752862e-01
4.93338071e-02 -6.24100566e-01 -1.53139576e-01 -8.23939264e-01
-5.21908820e-01 -9.10128772e-01 -1.19198933e-01 9.55765605e-01
1.29734814e-01 -4.28706467e-01 7.33633041e-01 5.26336432e-01
7.16727600e-02 -5.39899886e-01 -9.77650583e-01 -2.94436425e-01
1.86018795e-01 -8.01077783e-02 2.87997395e-01 7.51782417e-01
2.74035543e-01 2.79692024e-01 -1.82810456e-01 1.88229993e-01
6.16843760e-01 -1.21498138e-01 4.44227427e-01 -1.34583735e+00
-5.49577892e-01 7.45267347e-02 -3.91792893e-01 -7.86836863e-01
-8.83339271e-02 -6.57229245e-01 1.06587380e-01 -1.00407755e+00
8.97514105e-01 4.99919616e-03 -2.26201788e-01 -2.36989245e-01
-3.31975007e-03 1.78264946e-01 -1.57112345e-01 -3.65483239e-02
-6.94425642e-01 3.46640319e-01 1.15905690e+00 3.83227229e-01
3.08325022e-01 -4.40236092e-01 -7.34106958e-01 5.19814610e-01
5.11373580e-01 -6.90717220e-01 8.34998712e-02 5.92510663e-02
9.67111468e-01 1.64290562e-01 2.25765526e-01 -5.80691874e-01
4.24175337e-02 -2.37692058e-01 5.54306984e-01 -4.04561728e-01
2.80502856e-01 -6.13107443e-01 6.94524825e-01 6.20671749e-01
-3.66930038e-01 -1.18589215e-01 -1.89853795e-02 9.48590159e-01
4.51142430e-01 -2.71299422e-01 8.43550980e-01 -3.08988988e-01
6.41576201e-02 -7.23222941e-02 -1.18894219e+00 -1.05908819e-01
7.24383473e-01 1.21407760e-02 -3.92353147e-01 1.86644405e-01
-1.14676130e+00 -3.64944607e-01 1.08701205e+00 -3.27055186e-01
-1.27499714e-01 -9.15438533e-01 -2.74688691e-01 8.63091871e-02
-2.90375322e-01 -3.11652988e-01 3.37173373e-01 9.42288637e-01
-7.60456979e-01 8.65453899e-01 -4.81748357e-02 -6.62229717e-01
-1.05682969e+00 6.30403280e-01 3.48011166e-01 -2.57850498e-01
-1.24608807e-01 1.75334796e-01 3.43427323e-02 -5.40833294e-01
-4.79187399e-01 -2.36669883e-01 2.71239787e-01 -2.29375631e-01
3.91670585e-01 4.21878040e-01 8.84317234e-02 -5.98800480e-01
-3.62761438e-01 5.93017936e-01 3.19215003e-04 -2.14211643e-02
1.37997890e+00 -3.75049680e-01 -6.42121017e-01 7.30729699e-01
7.15700984e-01 4.30635065e-01 -1.51853061e+00 1.40181497e-01
2.69499570e-01 -1.20697498e-01 -1.75077647e-01 -5.71235061e-01
-2.29464710e-01 6.70015991e-01 2.42265731e-01 7.42186159e-02
6.10415459e-01 2.20369264e-01 4.38811451e-01 5.96786320e-01
2.69414276e-01 -1.02314258e+00 -2.70750970e-01 4.04303372e-01
1.69766322e-01 -6.16097331e-01 8.23683515e-02 -3.29950005e-01
1.40895382e-01 1.11331964e+00 -5.93356155e-02 7.95891508e-02
4.75502580e-01 6.00755095e-01 -2.98327059e-01 -1.43756330e-01
-1.09432721e+00 9.57636535e-02 -2.03216493e-01 5.42997122e-01
7.67763853e-01 2.27520037e-02 -5.84264278e-01 2.68854588e-01
2.79029578e-01 -7.65725672e-02 8.37139606e-01 1.03566432e+00
-8.79805684e-01 -1.62929046e+00 -5.28161228e-01 9.13107023e-02
-6.88090265e-01 6.57798424e-02 -5.13674557e-01 5.26948631e-01
-2.86940504e-02 4.59914178e-01 -1.05956919e-01 1.99653372e-01
-1.39068916e-01 6.80029333e-01 9.24927413e-01 -4.82920200e-01
-3.69427800e-02 3.67182851e-01 9.09173116e-02 -2.79896468e-01
-4.63715941e-01 -1.22182679e+00 -1.23053229e+00 -5.03020406e-01
-2.68975347e-01 5.22850394e-01 7.27378547e-01 1.14861715e+00
6.51836395e-01 2.54172891e-01 2.70618886e-01 -7.95110345e-01
-4.33725059e-01 -3.98042172e-01 -1.02006686e+00 -4.50115763e-02
3.13313782e-01 -5.86044371e-01 -1.70389459e-01 5.81243495e-03] | [5.0018792152404785, 5.193822860717773] |
0b1b39c0-884b-432c-b713-e34f9b59c27a | how-to-address-monotonicity-for-model-risk | 2305.00799 | null | https://arxiv.org/abs/2305.00799v1 | https://arxiv.org/pdf/2305.00799v1.pdf | How to address monotonicity for model risk management? | In this paper, we study the problem of establishing the accountability and fairness of transparent machine learning models through monotonicity. Although there have been numerous studies on individual monotonicity, pairwise monotonicity is often overlooked in the existing literature. This paper studies transparent neural networks in the presence of three types of monotonicity: individual monotonicity, weak pairwise monotonicity, and strong pairwise monotonicity. As a means of achieving monotonicity while maintaining transparency, we propose the monotonic groves of neural additive models. As a result of empirical examples, we demonstrate that monotonicity is often violated in practice and that monotonic groves of neural additive models are transparent, accountable, and fair. | ['Weicheng Ye', 'Dangxing Chen'] | 2023-04-28 | null | null | null | null | ['additive-models'] | ['methodology'] | [ 1.60733029e-01 4.41548139e-01 -5.59423804e-01 -8.25715721e-01
-4.46074493e-02 -5.37941098e-01 2.71138847e-01 -3.36167872e-01
-4.79296535e-01 1.16406345e+00 2.48936757e-01 -6.08435154e-01
-4.59098995e-01 -4.87816632e-01 -8.89011919e-01 -4.08928931e-01
1.25483144e-02 -3.33405770e-02 -2.60540009e-01 1.23089448e-01
2.57560015e-01 1.91581547e-01 -1.20182967e+00 5.12573779e-01
1.21369016e+00 8.78249288e-01 -8.36719453e-01 4.93862867e-01
1.63820758e-01 1.45162559e+00 -3.08298409e-01 -9.37859595e-01
7.94857323e-01 -4.86409009e-01 -6.82587743e-01 -3.98911268e-01
1.04909575e+00 -8.85373414e-01 9.66188014e-02 1.23936534e+00
1.44805625e-01 -3.31368208e-01 6.07015967e-01 -1.84363282e+00
-1.39864051e+00 1.19745982e+00 -8.11301410e-01 5.56767210e-02
-3.34053725e-01 1.49714306e-01 1.54842472e+00 -2.21605659e-01
1.96023852e-01 1.22582400e+00 8.70332897e-01 1.02540433e+00
-1.21682465e+00 -1.05286169e+00 4.72206801e-01 -2.33942226e-01
-7.12965906e-01 -7.11367309e-01 5.61770439e-01 -5.59942067e-01
5.52485287e-01 5.41581810e-01 5.44114590e-01 8.95067394e-01
3.39312665e-02 6.23818338e-01 1.49513245e+00 -1.63609967e-01
-1.79299656e-02 2.57900834e-01 6.24727905e-01 5.98877013e-01
1.12401330e+00 4.81889158e-01 -8.11839759e-01 -3.72411758e-01
8.41925859e-01 -2.13125318e-01 -3.35894406e-01 -5.30516624e-01
-6.73578918e-01 7.71818519e-01 5.60740292e-01 -1.24531880e-01
-2.85358399e-01 5.68736136e-01 4.15669143e-01 7.44618952e-01
6.02387071e-01 5.18514216e-01 -5.02597213e-01 1.29448429e-01
-8.09449434e-01 8.01948458e-02 7.86780477e-01 7.38376915e-01
4.77261305e-01 3.47590625e-01 -1.25607789e-01 3.55756670e-01
3.80701721e-01 2.14825451e-01 8.42386037e-02 -1.55418193e+00
5.57387233e-01 6.22915804e-01 4.82532501e-01 -7.62472808e-01
-2.58192748e-01 -3.93793494e-01 -8.30213249e-01 5.03714085e-01
6.51185572e-01 -2.43346840e-01 -4.58691269e-01 2.19791770e+00
1.27089113e-01 -3.65211010e-01 1.86911836e-01 8.39742303e-01
5.57619750e-01 1.74311444e-01 1.51133031e-01 -4.12991226e-01
5.36035836e-01 -8.02881777e-01 -6.81610227e-01 -1.38360914e-02
5.73048532e-01 -1.74218416e-01 1.16056108e+00 2.49258950e-01
-1.46247137e+00 1.19551718e-01 -7.65504599e-01 -3.58965516e-01
1.49859026e-01 -2.85319477e-01 9.42243099e-01 1.09870231e+00
-1.27294683e+00 4.12616789e-01 -4.82740492e-01 4.45846394e-02
8.15615714e-01 5.17594934e-01 -5.14181733e-01 1.88528284e-01
-1.05373180e+00 1.01411879e+00 1.08498679e-02 4.10989434e-01
-4.20151591e-01 -8.29516947e-01 -6.42542779e-01 2.66255230e-01
1.40800625e-01 -9.50791895e-01 1.47312176e+00 -1.81023562e+00
-1.25752294e+00 8.94923747e-01 1.29963279e-01 -6.65548801e-01
1.14566374e+00 -4.36514407e-01 2.18684107e-01 -3.29764545e-01
-1.95914924e-01 5.47468662e-01 4.37455207e-01 -1.57607973e+00
-4.99353826e-01 -5.16955137e-01 5.15334964e-01 3.25281352e-01
-8.24318290e-01 2.97775865e-01 6.11136079e-01 -9.66867432e-02
-1.52774036e-01 -7.05160558e-01 7.47335851e-02 4.38650876e-01
-5.82303047e-01 -2.24296446e-03 3.86246026e-01 -5.40347517e-01
1.13862967e+00 -2.25129175e+00 -3.15304518e-01 2.02400908e-01
7.11874604e-01 -1.67489037e-01 6.99631274e-02 -2.89163083e-01
-6.11572713e-02 6.19536042e-01 -5.10229588e-01 -2.05697104e-01
4.18585032e-01 1.12285621e-01 -2.96101660e-01 6.21960282e-01
-7.23490044e-02 1.04347908e+00 -4.70860869e-01 -4.34516698e-01
-6.69625521e-01 1.79245085e-01 -8.02383244e-01 -9.39953700e-02
2.20735818e-01 1.43510327e-02 -2.88329627e-02 5.89670360e-01
7.07198262e-01 -8.47110525e-02 4.22974348e-01 1.14532396e-01
-3.43414158e-01 5.56706011e-01 -8.41632783e-01 4.27179784e-01
8.51682648e-02 7.43556142e-01 3.86836648e-01 -7.27479637e-01
4.34314191e-01 4.16711062e-01 2.29941353e-01 -8.22931945e-01
3.12368479e-02 3.75683933e-01 5.81101298e-01 -3.37720603e-01
5.99379301e-01 -5.06635189e-01 1.30547285e-01 1.11126375e+00
-6.29705489e-01 3.65256667e-02 -1.66207943e-02 8.28109607e-02
6.08353972e-01 1.51467398e-01 5.60766757e-01 -3.65902245e-01
-1.30453259e-01 -2.19990894e-01 9.10233080e-01 9.17980969e-01
-8.49381804e-01 3.68011177e-01 9.89030540e-01 -5.56176066e-01
-1.05976999e+00 -9.30184603e-01 4.99023013e-02 1.29769683e+00
-2.02013433e-01 3.11645567e-01 -8.14897060e-01 -6.66777492e-01
3.87123346e-01 5.32050371e-01 -9.56822932e-01 -1.12024635e-01
-4.47305322e-01 -7.39777088e-01 8.49999547e-01 6.77037001e-01
5.96984982e-01 -9.23097432e-01 -9.79106963e-01 -5.04567683e-01
1.34882361e-01 -6.42776608e-01 -6.79371953e-01 1.70688763e-01
-9.91385758e-01 -1.01117551e+00 -3.41863960e-01 -4.05204773e-01
8.87291193e-01 3.84977013e-01 1.05427063e+00 5.18687367e-01
5.37985325e-01 2.21329942e-01 3.00817937e-01 -7.59867370e-01
-3.79274905e-01 -3.21442455e-01 2.28010893e-01 -1.33677050e-01
8.39396268e-02 -3.95699352e-01 -3.16454560e-01 9.41793770e-02
-7.07316875e-01 1.81459829e-01 5.23960710e-01 7.84342945e-01
9.23630297e-02 -4.69125867e-01 7.07363307e-01 -1.50722349e+00
9.40190673e-01 -2.22398326e-01 -6.32441580e-01 2.44730368e-01
-9.99432027e-01 -3.05001289e-01 4.60764259e-01 -3.93480986e-01
-1.10613430e+00 -2.26449028e-01 7.75405169e-01 -2.20595933e-02
2.95803666e-01 6.04042530e-01 -1.75359368e-01 -3.03687721e-01
8.82139564e-01 -3.03421676e-01 4.61900495e-02 3.25984545e-02
4.67551589e-01 6.88112020e-01 2.38153458e-01 -6.50156736e-01
7.03645945e-01 4.68238831e-01 -2.87572950e-01 -3.07547659e-01
-1.01566899e+00 3.11917871e-01 -5.30148566e-01 -2.30449706e-01
3.89937848e-01 -6.46699071e-01 -8.00341249e-01 3.33880782e-01
-9.54194188e-01 -6.45284235e-01 -3.56659234e-01 3.73470247e-01
-5.56215882e-01 4.42793250e-01 -7.81616688e-01 -1.16349018e+00
-4.55653101e-01 -6.80444360e-01 1.95750836e-02 1.26908347e-01
-4.95759875e-01 -7.33641982e-01 -6.15788437e-02 4.21651632e-01
6.38011634e-01 3.92625704e-02 1.02317786e+00 -4.45342213e-01
-6.28541648e-01 -8.22668895e-02 -4.50134397e-01 5.58238328e-01
2.25396425e-01 5.01828611e-01 -1.11963689e+00 -2.52706668e-04
8.03135782e-02 -4.92201477e-01 8.97359550e-01 6.36548162e-01
1.00099707e+00 -1.21351135e+00 2.88513452e-01 5.11235297e-01
1.04796731e+00 -4.21486571e-02 3.97449732e-01 3.28747153e-01
6.08378530e-01 8.97648573e-01 7.55983070e-02 8.26627612e-02
6.24583542e-01 -2.71091387e-02 7.54460037e-01 -1.74338624e-01
3.77910286e-01 -5.07289264e-03 4.46911067e-01 4.35938358e-01
-3.91249686e-01 1.42858118e-01 -8.22789967e-01 4.54089463e-01
-1.98033261e+00 -1.42374301e+00 -3.60970706e-01 2.33200502e+00
1.22111666e+00 3.58201772e-01 3.40945840e-01 1.00238770e-01
7.73994327e-01 -1.45314127e-01 -5.65976977e-01 -1.11211801e+00
-1.99244067e-01 -6.15984321e-01 6.78922415e-01 6.10382438e-01
-8.74283850e-01 5.98353207e-01 7.55106878e+00 -2.33381242e-01
-9.42116499e-01 1.24775216e-01 9.96634543e-01 -5.63373864e-01
-8.26856732e-01 -3.58207859e-02 -2.46757209e-01 3.17399234e-01
5.56849003e-01 -3.20139140e-01 3.85801196e-01 7.24867105e-01
3.39199275e-01 7.96219483e-02 -1.45834613e+00 4.58641171e-01
-1.73787579e-01 -1.03257406e+00 3.26818600e-02 2.07845777e-01
1.15367186e+00 -4.37187329e-02 6.16631031e-01 -4.91606779e-02
8.83440912e-01 -1.27459562e+00 1.19541454e+00 3.72583210e-01
6.52040541e-01 -5.93400598e-01 6.74183428e-01 3.75759721e-01
-3.03132921e-01 -5.67465127e-01 -3.12852532e-01 -5.42024076e-01
-3.07540506e-01 6.49018109e-01 -9.60000828e-02 1.62726268e-01
6.46578372e-01 4.24327672e-01 -1.11005455e-01 1.11574554e+00
-4.21567470e-01 7.71155953e-01 -1.98485941e-01 1.73481733e-01
1.77194685e-01 -2.12321311e-01 -9.03990120e-03 9.42613363e-01
-1.16586804e-01 -3.62997279e-02 -4.10887688e-01 1.22541630e+00
-6.25623047e-01 -1.28734708e-01 -7.36016214e-01 -8.61667320e-02
3.31149548e-01 8.81371140e-01 -1.59438968e-01 -2.30467722e-01
-3.74816507e-01 4.63178843e-01 5.95974028e-01 4.46144283e-01
-8.47160876e-01 2.66793638e-01 6.79670393e-01 -1.66999653e-01
-3.48057628e-01 4.12859255e-03 -1.16857934e+00 -9.65535522e-01
3.34985346e-01 -8.76693606e-01 5.24962842e-01 -6.26885176e-01
-1.47400415e+00 -5.54960109e-02 -2.21583933e-01 -8.96979690e-01
2.05298036e-01 -4.27856535e-01 -6.91680670e-01 7.73169756e-01
-1.75072932e+00 -1.18337381e+00 -1.59602121e-01 3.25770050e-01
-2.72185683e-01 2.25654081e-01 6.38279378e-01 2.08449960e-01
-6.07452035e-01 8.14194977e-01 -1.88628938e-02 3.47175896e-02
8.31081450e-01 -1.23653567e+00 8.18562508e-02 9.04310822e-01
-5.90171330e-02 1.21684027e+00 6.62749052e-01 -4.97074723e-01
-7.63606608e-01 -8.35372329e-01 9.89723563e-01 -5.13408840e-01
6.89101994e-01 -1.31040916e-01 -8.93863499e-01 1.21958077e+00
3.73604983e-01 3.59283276e-02 9.35208976e-01 2.94002652e-01
-1.05493474e+00 -2.72691160e-01 -1.34953249e+00 6.95949614e-01
1.14472127e+00 -6.03781581e-01 -2.61290401e-01 2.54772529e-02
7.24306166e-01 -1.93817332e-01 -3.37423295e-01 3.21273744e-01
1.17040908e+00 -1.29703486e+00 4.08519745e-01 -1.04115164e+00
7.06159413e-01 -8.65344796e-03 -1.11036584e-01 -1.03712833e+00
-3.61497343e-01 -5.91890574e-01 -1.97118908e-01 1.03181863e+00
8.61380935e-01 -1.06976771e+00 8.48465800e-01 1.36323500e+00
-4.59043775e-03 -5.22338629e-01 -1.15029645e+00 -7.50127435e-01
5.68807304e-01 2.18667910e-02 6.09839976e-01 1.38190162e+00
1.84566557e-01 -1.68484464e-01 -5.67666709e-01 -3.27545032e-02
6.23364747e-01 1.92333281e-01 6.61735535e-01 -1.38859439e+00
-3.80210169e-02 -1.02459157e+00 -3.29008773e-02 -6.22157753e-01
2.92658329e-01 -6.41411722e-01 2.83069909e-02 -1.61653352e+00
8.67357433e-01 -4.42917764e-01 -6.89182818e-01 8.14597964e-01
-1.85076311e-01 2.94244409e-01 2.29975775e-01 3.10314775e-01
-4.77051437e-01 6.69763312e-02 9.90139127e-01 -2.69015193e-01
-9.55495834e-02 1.21603318e-01 -1.62875664e+00 7.02507496e-01
1.15169024e+00 -4.71341908e-01 -3.63160998e-01 -8.47764254e-01
7.82932758e-01 -5.72906256e-01 5.23993015e-01 -1.87294155e-01
1.16374455e-01 -6.92255616e-01 2.53100656e-02 1.48912981e-01
3.15783359e-02 -9.40336585e-01 -9.31948647e-02 5.40974438e-01
-8.18036914e-01 1.84659243e-01 -2.85310671e-02 1.77739739e-01
1.43683210e-01 6.75787553e-02 9.19663072e-01 -2.24000841e-01
-7.24092731e-03 1.14638813e-01 -3.35554481e-02 6.06525168e-02
7.02432752e-01 -5.10615110e-01 -8.76411319e-01 -5.20537972e-01
-1.82820275e-01 3.18961591e-01 7.67691910e-01 -6.46981150e-02
4.22341675e-01 -9.73370850e-01 -7.34304965e-01 -1.74269840e-01
-1.78404525e-01 -2.07389161e-01 -1.00549668e-01 1.22560513e+00
-2.78789788e-01 3.34433526e-01 -5.49604356e-01 -7.20132096e-03
-1.35880125e+00 2.77109593e-01 7.89260745e-01 2.23827213e-02
-1.39807418e-01 9.29881334e-01 5.81007123e-01 -5.99540591e-01
3.54712844e-01 -4.44612294e-01 4.89610508e-02 -1.13067098e-01
1.86150819e-01 3.03753227e-01 -4.64224160e-01 -4.36135232e-01
-2.36349985e-01 1.85070947e-01 1.25531303e-02 -1.17293730e-01
1.32178473e+00 -5.76306656e-02 -5.54007530e-01 7.29966223e-01
3.80311906e-01 2.55809426e-01 -1.41702807e+00 1.05483383e-01
8.21581185e-02 -3.75812173e-01 -3.43695015e-01 -9.82764482e-01
-1.02054429e+00 8.83967578e-01 3.56514789e-02 5.13679981e-01
8.93751562e-01 -4.90680963e-01 2.05917917e-02 3.76774311e-01
6.87252954e-02 -1.07056296e+00 -3.64939421e-01 3.92940402e-01
7.43045986e-01 -1.18837285e+00 1.79826215e-01 -2.88756102e-01
-7.50509083e-01 6.82028115e-01 1.02261066e+00 2.06335336e-01
1.81690425e-01 2.33494803e-01 2.67194599e-01 3.34158242e-02
-9.25065815e-01 3.06498766e-01 -2.64161061e-02 5.33903480e-01
7.52474248e-01 3.81905407e-01 -5.85230291e-01 7.62253761e-01
-4.08582568e-01 9.58698243e-02 9.93754804e-01 7.16411889e-01
-4.96899992e-01 -5.45920610e-01 -1.94226578e-01 7.27771521e-01
-8.16233397e-01 -3.96982938e-01 -1.04394293e+00 9.05416071e-01
-1.39010265e-01 7.32752860e-01 -1.59234125e-02 7.91234672e-02
1.36491925e-01 1.91064216e-02 4.68951732e-01 -4.40099835e-01
-9.30813789e-01 -5.92100382e-01 2.52189368e-01 -3.28397751e-01
-6.74998820e-01 -6.02236569e-01 -1.00632060e+00 -5.95371425e-01
-4.37613666e-01 3.23507329e-03 3.19878727e-01 7.20886528e-01
3.80501486e-02 5.61124235e-02 4.75871056e-01 -7.31297210e-02
-1.03504264e+00 -8.24465454e-01 -5.61481774e-01 2.64033854e-01
8.96151185e-01 -3.29560310e-01 -5.89878857e-01 -1.74620224e-03] | [8.876046180725098, 5.3567423820495605] |
99ef5eca-d50d-4bf1-910e-80a301b430ce | climart-a-benchmark-dataset-for-emulating | 2111.14671 | null | https://arxiv.org/abs/2111.14671v1 | https://arxiv.org/pdf/2111.14671v1.pdf | ClimART: A Benchmark Dataset for Emulating Atmospheric Radiative Transfer in Weather and Climate Models | Numerical simulations of Earth's weather and climate require substantial amounts of computation. This has led to a growing interest in replacing subroutines that explicitly compute physical processes with approximate machine learning (ML) methods that are fast at inference time. Within weather and climate models, atmospheric radiative transfer (RT) calculations are especially expensive. This has made them a popular target for neural network-based emulators. However, prior work is hard to compare due to the lack of a comprehensive dataset and standardized best practices for ML benchmarking. To fill this gap, we build a large dataset, ClimART, with more than \emph{10 million samples from present, pre-industrial, and future climate conditions}, based on the Canadian Earth System Model. ClimART poses several methodological challenges for the ML community, such as multiple out-of-distribution test sets, underlying domain physics, and a trade-off between accuracy and inference speed. We also present several novel baselines that indicate shortcomings of datasets and network architectures used in prior work. Download instructions, baselines, and code are available at: https://github.com/RolnickLab/climart | ['David Rolnick', 'Howard Barker', 'Jason N. S. Cole', 'Venkatesh Ramesh', 'Salva Rühling Cachay'] | 2021-11-29 | null | null | null | null | ['physical-simulations'] | ['miscellaneous'] | [-4.14622277e-01 -5.36354125e-01 -4.94292267e-02 -4.50833678e-01
-3.53677005e-01 -7.12129593e-01 1.00115037e+00 1.27652004e-01
-2.67406166e-01 1.20039463e+00 -1.65222064e-02 -8.56827199e-01
2.01991901e-01 -1.06953454e+00 -7.70210505e-01 -5.57232738e-01
-1.89059392e-01 5.16486645e-01 -9.08458084e-02 -6.05009645e-02
-2.91583557e-02 6.04311943e-01 -1.57566321e+00 3.58819515e-02
1.08060884e+00 6.56277657e-01 1.62968799e-01 8.77071679e-01
-2.59454936e-01 6.35420680e-01 -4.30957198e-01 4.84524332e-02
1.60996392e-01 -5.01495540e-01 -4.58159328e-01 -7.62094498e-01
6.65215135e-01 -3.23600531e-01 -3.37810874e-01 9.19495165e-01
6.83728874e-01 4.63312596e-01 8.23853612e-01 -1.22942591e+00
-3.82174373e-01 4.26148146e-01 -1.72405928e-01 3.24521899e-01
-3.44137102e-01 5.34255028e-01 7.98031926e-01 -8.17925811e-01
4.64854866e-01 1.17212021e+00 1.02430475e+00 2.63817787e-01
-1.44754159e+00 -9.33034778e-01 -2.15607900e-02 -9.46260318e-02
-1.46906197e+00 -4.89660680e-01 1.79048508e-01 -7.26547897e-01
1.19776475e+00 3.80355865e-01 7.76912570e-01 1.02667272e+00
2.40035966e-01 2.56135076e-01 1.39758039e+00 -1.78652346e-01
5.85746467e-01 2.86202908e-01 -6.71753660e-02 4.30016220e-01
2.13350445e-01 5.76330185e-01 -2.99124748e-01 -4.61959958e-01
8.56496394e-01 -3.02579075e-01 -2.24751860e-01 3.04994918e-02
-9.80326474e-01 8.26248467e-01 4.27532971e-01 5.03631961e-03
-3.70395660e-01 4.59298581e-01 2.22373977e-01 2.51113057e-01
1.06118572e+00 4.00750726e-01 -9.64325070e-01 -3.19056451e-01
-1.24746072e+00 8.78781676e-01 1.18853474e+00 4.70123053e-01
8.37164879e-01 3.81498277e-01 4.24086750e-01 8.45328331e-01
3.84811133e-01 9.50372458e-01 -8.93350616e-02 -1.27840734e+00
1.78845376e-01 7.15009123e-02 3.52282465e-01 -1.02812409e+00
-3.70190352e-01 -6.23759568e-01 -1.15822482e+00 4.81101871e-01
5.26945114e-01 -7.50070989e-01 -6.36211455e-01 1.67246473e+00
4.82164502e-01 4.49750125e-01 -1.17950961e-01 7.56651580e-01
6.17885768e-01 1.24694991e+00 2.55130619e-01 5.55606931e-02
8.92091215e-01 -7.22035468e-01 -3.50036383e-01 -3.42688799e-01
5.69144368e-01 -7.69865394e-01 9.20382380e-01 3.82098183e-02
-1.02425015e+00 -4.41523403e-01 -7.32051969e-01 1.65545821e-01
-8.83667052e-01 -1.17130503e-01 9.55789506e-01 5.68499386e-01
-1.21465623e+00 8.74834776e-01 -1.13415372e+00 -4.33078289e-01
1.23526722e-01 -6.98219687e-02 1.89509496e-01 1.98735818e-01
-1.45725691e+00 1.22638404e+00 3.38237345e-01 3.47693384e-01
-7.07760632e-01 -1.22139764e+00 -6.71004057e-01 -9.93482023e-02
-2.06070364e-01 -7.30378389e-01 1.30880404e+00 -9.03157592e-01
-1.34122849e+00 5.55235505e-01 -2.32710525e-01 -4.19805229e-01
4.90687937e-01 -1.31844416e-01 -4.95058984e-01 -3.26242417e-01
-2.05727041e-01 6.49259090e-01 2.21607015e-01 -1.13815463e+00
-4.01029855e-01 -2.84552108e-02 -2.84459502e-01 1.12043023e-01
6.90046549e-02 1.07482493e-01 -1.03748716e-01 -6.17496967e-01
-2.50862241e-01 -1.09161115e+00 -3.70520622e-01 2.53171384e-01
-6.35515004e-02 1.12897661e-02 3.92529786e-01 -9.22741234e-01
9.36393261e-01 -1.82815266e+00 -2.04143494e-01 1.83359683e-01
5.27962074e-02 2.45289937e-01 -3.95366959e-02 5.52176952e-01
-1.33079365e-01 4.27833349e-01 -5.70587933e-01 -3.24769527e-01
8.11564326e-02 2.71003544e-01 -5.52114308e-01 5.91552734e-01
1.43919632e-01 6.37041867e-01 -8.39999318e-01 -3.36571395e-01
4.76189107e-01 6.12259865e-01 -4.66912895e-01 8.17816108e-02
-7.45887578e-01 7.17333794e-01 -1.90717682e-01 5.49028516e-01
9.29273963e-01 -1.79096326e-01 3.97185050e-02 1.88880652e-01
-6.59967780e-01 6.72733068e-01 -1.23314488e+00 1.27030444e+00
-6.67374551e-01 7.65647054e-01 3.70130628e-01 -6.60643339e-01
7.33876586e-01 1.82597816e-01 1.90337896e-01 -5.61323106e-01
-1.83352664e-01 3.73462707e-01 -9.96087715e-02 -1.10922366e-01
5.99135160e-01 -1.46773115e-01 3.77129436e-01 6.49386585e-01
-2.60565042e-01 -7.22192109e-01 6.78177103e-02 -1.30716458e-01
6.22838616e-01 6.25625551e-01 3.93087938e-02 -7.06583023e-01
-9.08476189e-02 3.70891362e-01 6.41380131e-01 7.91166961e-01
-7.05176443e-02 3.99025202e-01 3.21525633e-01 -6.47611976e-01
-1.18825614e+00 -1.09964848e+00 -4.87837672e-01 1.01605725e+00
-4.89221513e-01 -1.47973642e-01 -4.74360466e-01 1.18835002e-01
3.31012309e-01 1.10211265e+00 -3.57325166e-01 2.21613824e-01
-4.68347520e-01 -1.30020714e+00 8.28176796e-01 3.25822026e-01
4.88225162e-01 -9.87591743e-01 -3.63801539e-01 1.16306193e-01
-2.60342091e-01 -7.90924311e-01 2.42151424e-01 1.53879687e-01
-1.16957283e+00 -7.47357190e-01 -4.88041192e-01 -2.08386257e-01
1.90050080e-01 -2.72974730e-01 1.71451819e+00 5.76170981e-02
-1.42970204e-01 -3.39426585e-02 2.37394616e-01 -6.64547980e-01
-6.19569063e-01 4.58742678e-01 1.83326393e-01 -8.10015321e-01
1.20719500e-01 -8.42393398e-01 -6.04523540e-01 9.55608040e-02
-6.43390119e-01 2.20553339e-01 2.47810259e-01 4.16393310e-01
3.48244101e-01 4.86008227e-02 4.35090691e-01 -6.17000997e-01
3.65716338e-01 -1.03473389e+00 -1.19427490e+00 -5.81871439e-03
-7.54394531e-01 -1.02777794e-01 5.06242752e-01 -1.18696347e-01
-1.21566534e+00 -2.24098563e-01 -1.43965140e-01 -1.19100530e-02
-3.61771196e-01 7.94451356e-01 4.74251777e-01 8.20318460e-02
9.32758391e-01 1.91847190e-01 4.77093160e-02 -5.49352884e-01
4.24790114e-01 5.67350447e-01 4.41376895e-01 -1.05882931e+00
8.20752680e-01 2.33736768e-01 5.01515158e-02 -1.13557208e+00
-6.59912765e-01 -1.33308722e-02 -4.58633780e-01 -3.06124955e-01
4.86231148e-01 -1.31425250e+00 -3.61299694e-01 8.01463604e-01
-1.09243989e+00 -1.13388145e+00 -1.42780691e-02 7.61000812e-01
-2.43271589e-01 -1.18092395e-01 -7.69930601e-01 -9.68385994e-01
-4.47329074e-01 -8.55179250e-01 4.82351899e-01 2.44695455e-01
-4.36509281e-01 -1.40418875e+00 4.88036007e-01 -2.68158197e-01
1.03067231e+00 5.82471013e-01 9.53890026e-01 -4.53924686e-02
-5.71527719e-01 1.77932680e-01 -3.63877296e-01 3.98334652e-01
1.32557461e-02 6.07339978e-01 -1.30458105e+00 -2.00133130e-01
-2.59021699e-01 -3.71818572e-01 9.42957759e-01 6.33302689e-01
1.12121558e+00 -2.62346447e-01 -3.06042910e-01 8.36523712e-01
1.44954550e+00 -2.36552179e-01 4.06324506e-01 1.92506924e-01
4.52620625e-01 5.57076454e-01 1.16338119e-01 4.62899178e-01
4.30080712e-01 1.36087969e-01 2.98547417e-01 -3.85895938e-01
1.19402438e-01 1.05424553e-01 2.55054861e-01 8.44783425e-01
-1.39129221e-01 -1.85320899e-01 -1.42049932e+00 5.21288335e-01
-1.60224771e+00 -1.04456353e+00 -5.01893520e-01 2.24477816e+00
1.34249508e+00 -4.94412184e-02 -1.72384188e-01 -4.54819381e-01
4.16694015e-01 3.73889565e-01 -5.61463237e-01 -4.78111655e-01
-1.19213603e-01 4.27960277e-01 6.66207194e-01 6.59342706e-01
-1.12968588e+00 8.73771966e-01 6.93685818e+00 5.88648975e-01
-1.29311395e+00 1.28252178e-01 9.89374757e-01 -2.35745221e-01
-2.29755342e-01 7.71022215e-02 -9.17802215e-01 4.66357768e-01
1.77466798e+00 -3.15158926e-02 8.49870443e-01 7.99025416e-01
6.71046615e-01 -3.58749658e-01 -8.82671714e-01 4.11436409e-01
-5.14863193e-01 -1.52419710e+00 -3.52820873e-01 9.66296624e-03
9.71807122e-01 1.03631830e+00 -9.98099744e-02 6.10663712e-01
9.31453228e-01 -1.15753305e+00 6.02587700e-01 6.95691466e-01
8.07188869e-01 -6.39329195e-01 4.56272036e-01 3.86804312e-01
-1.17867708e+00 4.68342394e-01 -5.90970933e-01 -5.96670866e-01
-3.67297977e-02 1.08618939e+00 -4.37936842e-01 3.14985901e-01
8.71409833e-01 5.51997960e-01 -3.19281757e-01 9.98368382e-01
3.62772942e-02 1.16600406e+00 -8.23312938e-01 4.51046461e-03
2.59021848e-01 -3.39409709e-01 1.90741077e-01 1.38119948e+00
3.36487144e-01 -6.92414567e-02 1.42041922e-01 1.35852635e+00
-2.45297793e-02 -1.84382871e-01 -7.39413261e-01 -3.52562815e-02
8.03928316e-01 1.29197502e+00 -4.65056121e-01 -3.84985119e-01
-3.49583954e-01 4.70652729e-01 3.58948469e-01 5.38725019e-01
-8.71777654e-01 -6.30464032e-02 1.40253830e+00 3.08278687e-02
-1.45301744e-01 -6.89334452e-01 -3.70479941e-01 -9.10113513e-01
-4.00605202e-01 -9.34570730e-01 7.93216601e-02 -8.87936831e-01
-1.26236963e+00 1.01853870e-01 2.02814922e-01 -6.67997837e-01
-3.08114111e-01 -5.39740741e-01 -7.77263582e-01 1.60169804e+00
-1.74148512e+00 -7.63539433e-01 -3.74785602e-01 7.35866837e-03
2.35258684e-01 2.96951622e-01 1.07219636e+00 3.21525723e-01
-5.71197569e-01 -4.18497957e-02 7.57798791e-01 -7.04275593e-02
7.58593678e-01 -1.27572715e+00 1.05264032e+00 6.10374749e-01
-3.88968617e-01 7.94618726e-01 8.67471695e-01 -7.25791633e-01
-1.17856669e+00 -1.49186015e+00 8.32511306e-01 -5.69224536e-01
8.30300331e-01 -2.78927565e-01 -1.08707249e+00 7.64545202e-01
2.28446022e-01 1.38868883e-01 4.80318308e-01 2.88998336e-01
-2.20661342e-01 4.57361666e-03 -9.40203309e-01 6.08937442e-01
5.87029994e-01 -6.07788503e-01 -1.11061670e-01 5.25227189e-01
4.23184574e-01 -5.11873364e-01 -1.06586981e+00 3.52549106e-01
6.34908438e-01 -8.08673322e-01 9.72413182e-01 -3.86673301e-01
6.21196926e-01 -2.51714557e-01 -1.25218570e-01 -1.64461756e+00
-9.32304263e-02 -3.06997061e-01 -1.65396795e-01 1.32293165e+00
5.15308797e-01 -9.48170483e-01 4.24833804e-01 8.06107998e-01
-3.96929123e-02 -5.29064476e-01 -7.02692330e-01 -7.17311442e-01
9.04968023e-01 -6.18662119e-01 6.16566598e-01 1.33631957e+00
-5.69984853e-01 -1.96850300e-01 -1.82816043e-01 4.11993593e-01
6.67277932e-01 1.95137933e-01 8.12905967e-01 -1.48155117e+00
-2.65811145e-01 -5.98666489e-01 4.77050662e-01 -7.46211886e-01
1.93820760e-01 -8.14287603e-01 1.76771447e-01 -1.45323825e+00
-6.18736632e-02 -8.83232176e-01 -1.54955059e-01 5.34521997e-01
6.94883913e-02 5.89054376e-02 -2.36547738e-01 3.03817809e-01
1.45430833e-01 6.95371330e-01 8.30587804e-01 1.93000570e-01
-8.38560089e-02 -1.83510840e-01 -9.90404114e-02 7.77283967e-01
1.38813925e+00 -5.82011282e-01 -1.16400182e-01 -6.76486790e-01
4.55700070e-01 1.45303300e-02 6.75384641e-01 -1.18440378e+00
1.42016886e-02 -6.33650184e-01 6.63016975e-01 -8.26409638e-01
3.60295415e-01 -3.70741367e-01 3.97195309e-01 4.67000008e-01
-1.68287233e-01 3.88215542e-01 7.66564786e-01 2.21095651e-01
1.63824007e-01 4.34848368e-02 8.85806739e-01 -4.50929701e-01
-4.50664163e-01 2.50824809e-01 -7.68105865e-01 2.78648823e-01
4.25423682e-01 3.31007242e-01 -5.13642490e-01 -2.43000910e-01
-3.30992550e-01 2.91124672e-01 6.52366817e-01 1.32064745e-01
-1.44255966e-01 -9.15126562e-01 -9.91051018e-01 1.03828162e-02
-3.14220577e-01 1.98541299e-01 4.93926869e-04 5.87826610e-01
-9.98287261e-01 3.13516438e-01 3.19071747e-02 -3.67074370e-01
-7.06987560e-01 -1.65726572e-01 8.53933752e-01 6.36547478e-03
-4.88937914e-01 6.51336014e-01 -8.94475654e-02 -1.37842262e+00
-1.39362156e-01 -6.08651996e-01 3.64548832e-01 -8.65090042e-02
4.04558301e-01 5.59771538e-01 -1.71879735e-02 -2.19795093e-01
-1.03585497e-01 7.77430013e-02 5.73970020e-01 -2.17818007e-01
1.31216466e+00 -8.27211048e-03 -4.87041384e-01 9.22361135e-01
7.85969377e-01 -2.77674705e-01 -1.23952413e+00 -1.87539935e-01
-2.20352501e-01 -1.97998971e-01 4.39210713e-01 -9.57301915e-01
-9.54371393e-01 8.39208901e-01 6.75635815e-01 1.09263584e-01
5.86693883e-01 -4.93602514e-01 5.10242701e-01 6.49641037e-01
-4.68776980e-03 -1.06616056e+00 -7.13790774e-01 9.27609682e-01
7.91365743e-01 -1.11046529e+00 2.72707433e-01 2.65111327e-02
3.18715461e-02 8.86344433e-01 6.23661458e-01 1.03595041e-01
1.06604731e+00 4.81855065e-01 1.02023423e-01 -9.95657369e-02
-9.74725306e-01 8.39951858e-02 -1.77885786e-01 2.91375041e-01
7.13674843e-01 1.90425485e-01 6.44023642e-02 -6.36175200e-02
-4.09988403e-01 1.52430832e-01 2.96808362e-01 8.63699377e-01
-4.58049238e-01 -7.45723963e-01 -5.38277268e-01 5.89459658e-01
-3.77970576e-01 -6.05187654e-01 -7.49312863e-02 8.01100552e-01
-1.08722001e-01 8.13726664e-01 3.42920810e-01 7.24433511e-02
-2.47819439e-01 3.61238539e-01 7.09657967e-02 -4.24982160e-01
-5.28122783e-01 -4.88746047e-01 3.77167493e-01 -4.35900688e-01
-2.10499272e-01 -8.01714599e-01 -1.05934167e+00 -1.11505997e+00
-7.76252449e-02 1.50289580e-01 1.09424102e+00 7.54335463e-01
4.64673102e-01 4.11827296e-01 2.00680271e-01 -1.29082406e+00
-5.24923623e-01 -1.20225978e+00 -4.57330376e-01 -1.43968463e-01
1.88261792e-01 -4.94116694e-01 -5.68248808e-01 -1.15268463e-02] | [6.553930282592773, 3.0260226726531982] |
f0e0ea05-7cb2-4468-a148-61093f2d159d | a-deep-learning-accelerated-data-assimilation | 2105.09468 | null | https://arxiv.org/abs/2105.09468v2 | https://arxiv.org/pdf/2105.09468v2.pdf | A Deep Learning-Accelerated Data Assimilation and Forecasting Workflow for Commercial-Scale Geologic Carbon Storage | Fast assimilation of monitoring data to update forecasts of pressure buildup and carbon dioxide (CO2) plume migration under geologic uncertainties is a challenging problem in geologic carbon storage. The high computational cost of data assimilation with a high-dimensional parameter space impedes fast decision-making for commercial-scale reservoir management. We propose to leverage physical understandings of porous medium flow behavior with deep learning techniques to develop a fast history matching-reservoir response forecasting workflow. Applying an Ensemble Smoother Multiple Data Assimilation framework, the workflow updates geologic properties and predicts reservoir performance with quantified uncertainty from pressure history and CO2 plumes interpreted through seismic inversion. As the most computationally expensive component in such a workflow is reservoir simulation, we developed surrogate models to predict dynamic pressure and CO2 plume extents under multi-well injection. The surrogate models employ deep convolutional neural networks, specifically, a wide residual network and a residual U-Net. The workflow is validated against a flat three-dimensional reservoir model representative of a clastic shelf depositional environment. Intelligent treatments are applied to bridge between quantities in a true-3D reservoir model and those in a single-layer reservoir model. The workflow can complete history matching and reservoir forecasting with uncertainty quantification in less than one hour on a mainstream personal workstation. | ['Joseph P. Morris', 'Matthew Burton-Kelly', 'Nicholas A. Azzolina', 'François Hamon', 'Xin Ju', 'Jize Zhang', 'Christopher S. Sherman', 'Pengcheng Fu', 'Hewei Tang'] | 2021-05-09 | null | null | null | null | ['seismic-inversion'] | ['miscellaneous'] | [-4.04733062e-01 -4.55487549e-01 3.13517898e-01 2.43313715e-01
-6.46487117e-01 -6.07158363e-01 7.40960956e-01 3.93587768e-01
-2.59871125e-01 1.09544444e+00 3.15197766e-01 -8.23286951e-01
1.34144247e-01 -1.12499297e+00 -6.18524790e-01 -6.30872428e-01
-6.94889665e-01 6.79713607e-01 2.13266071e-02 -5.59252501e-01
2.04504311e-01 8.12202930e-01 -1.47039521e+00 -2.17211738e-01
9.36319649e-01 1.12258291e+00 -8.28507766e-02 7.51313150e-01
-8.80117193e-02 5.04427910e-01 -2.97687292e-01 7.80402124e-01
6.03798389e-01 -2.22109899e-01 -2.31160775e-01 -3.82655472e-01
4.58062142e-02 -4.35482174e-01 -8.17841738e-02 7.16515899e-01
6.74331605e-01 3.37444782e-01 7.10442364e-01 -4.50887740e-01
-1.31818697e-01 5.35904706e-01 -2.01162890e-01 5.08343101e-01
-3.82484972e-01 7.26321697e-01 4.98251230e-01 -1.13044477e+00
1.21406630e-01 9.10781026e-01 1.02986610e+00 -4.19146419e-02
-1.18458855e+00 -4.60894078e-01 -5.26518375e-02 -7.95653760e-01
-1.23693538e+00 -6.82598412e-01 3.01664680e-01 -1.25171924e+00
1.31333709e+00 5.20721152e-02 1.06838524e+00 3.58112067e-01
3.92949194e-01 -6.72214404e-02 1.26126254e+00 1.59840062e-01
4.39000398e-01 -9.33447778e-02 -3.82937402e-01 3.09143811e-01
5.46307266e-01 9.38356280e-01 -3.15686762e-01 -1.88365623e-01
8.53026032e-01 1.11178428e-01 -2.62234271e-01 4.65875924e-01
-5.75192273e-01 8.73077214e-01 4.34617162e-01 7.33289868e-02
-5.48303187e-01 4.94530141e-01 2.08595335e-01 3.23335052e-01
1.09590328e+00 7.88494408e-01 -6.06075644e-01 -9.92817134e-02
-1.49725544e+00 7.54584789e-01 1.03248298e+00 5.34968078e-01
8.87185574e-01 9.75425363e-01 8.86197984e-02 3.27115089e-01
9.00827169e-01 1.40765691e+00 2.36927211e-01 -9.39141631e-01
5.44554442e-02 2.47633696e-01 5.44868529e-01 -4.51839745e-01
-2.51763552e-01 -2.36175835e-01 -1.03938878e+00 7.39303172e-01
7.08663985e-02 -6.31612003e-01 -1.03482747e+00 9.63239491e-01
7.29197934e-02 3.56292814e-01 3.03195983e-01 7.27184296e-01
4.57040459e-01 7.80404627e-01 7.57640675e-02 -1.46356374e-01
1.22466731e+00 -6.36763275e-01 -6.24145091e-01 -4.64852482e-01
3.43138278e-01 -8.86356607e-02 4.44392622e-01 -4.47895437e-01
-1.14784014e+00 -1.99045464e-01 -1.23857284e+00 3.10854495e-01
-9.79908466e-01 -4.19286102e-01 7.71210074e-01 3.54071945e-01
-1.07454526e+00 1.51840389e+00 -1.30165064e+00 1.73703954e-01
2.75061820e-02 8.10468197e-02 2.26394385e-01 7.13033020e-01
-1.63482559e+00 1.17916882e+00 3.20876837e-01 8.59858394e-01
-1.52228916e+00 -1.41712737e+00 -1.01765108e+00 2.29475573e-01
-3.41455966e-01 -5.40153742e-01 1.29176593e+00 1.17024064e-01
-1.64064705e+00 8.04621577e-02 -2.14805920e-02 -5.85566163e-01
9.10647511e-01 -5.12720197e-02 -5.98439693e-01 -2.13652715e-01
-2.14690659e-02 -5.88198472e-03 5.39842129e-01 -1.20634675e+00
-3.59677911e-01 -1.05874263e-01 -7.48203695e-01 2.60664791e-01
3.32278371e-01 -1.81564376e-01 1.25987321e-01 -3.05503964e-01
1.16418421e-01 -6.39050782e-01 -7.66674459e-01 -1.30379453e-01
2.48317584e-01 4.47170287e-01 5.51898003e-01 -9.76731122e-01
9.98179734e-01 -1.57096553e+00 -2.06154764e-01 3.79855394e-01
-2.04933472e-02 4.54982854e-02 4.10513699e-01 5.70935309e-01
1.50160134e-01 5.72488725e-01 -9.14013624e-01 -4.84438062e-01
3.38980369e-02 4.29352760e-01 -6.15700543e-01 5.49810410e-01
6.88396275e-01 7.35346138e-01 -1.04752266e+00 8.97441506e-02
4.52298343e-01 3.50717753e-01 -2.31077880e-01 5.14770746e-01
-6.73223019e-01 8.98054779e-01 7.06211329e-02 8.37879717e-01
1.17146957e+00 -2.96110302e-01 -1.53619081e-01 4.43834066e-01
-9.40403700e-01 2.02008098e-01 -1.18972266e+00 1.41062868e+00
-6.12638116e-01 3.84532064e-01 5.89671910e-01 -4.14119214e-01
1.10471082e+00 1.66952297e-01 3.08417261e-01 -7.70057917e-01
-1.67041913e-01 9.31585670e-01 4.15551895e-03 -6.67455435e-01
9.62174177e-01 -8.22238207e-01 2.25572348e-01 4.84992504e-01
-3.84102404e-01 -6.92856729e-01 4.76490967e-02 -1.49740115e-01
5.68596244e-01 2.52892077e-01 -4.20206219e-01 -8.66934359e-01
2.67957985e-01 6.81552812e-02 4.39317495e-01 6.91407859e-01
8.89755562e-02 5.63208699e-01 2.25781366e-01 -3.66209120e-01
-1.32583845e+00 -8.79694223e-01 -4.64118451e-01 6.79220140e-01
-1.47228286e-01 2.15853184e-01 3.37876409e-01 3.13188970e-01
8.68598342e-01 4.30506527e-01 -7.11143136e-01 1.15678675e-01
-5.32470822e-01 -1.11922801e+00 7.28484154e-01 7.78939545e-01
4.57145423e-01 -9.94897127e-01 -5.08456171e-01 7.75487840e-01
6.48836851e-01 -4.14749205e-01 2.15404421e-01 4.42058355e-01
-1.21431911e+00 -8.03237140e-01 -6.24049902e-01 6.05305051e-03
3.39715689e-01 -6.02950156e-01 1.15726900e+00 -9.30973887e-02
1.24273822e-01 -5.92110045e-02 2.52789408e-01 -5.87573230e-01
-8.46276164e-01 -2.32495174e-01 3.43806624e-01 -5.70744634e-01
-5.41535765e-02 -6.57222748e-01 -8.73122156e-01 -1.27171800e-02
-7.13853061e-01 -2.12530524e-01 1.09476700e-01 5.21007717e-01
4.19779390e-01 -1.80923194e-01 5.79460502e-01 -5.87399781e-01
7.67251611e-01 -1.16345835e+00 -1.34917355e+00 1.06698394e-01
-1.14496613e+00 1.83540195e-01 3.39345098e-01 1.57244995e-01
-1.32292569e+00 -2.67679632e-01 -1.25766695e-01 -5.12904882e-01
2.90395200e-01 1.39716554e+00 6.11105502e-01 5.53965047e-02
6.40127063e-01 3.41437519e-01 3.38208288e-01 -6.28823817e-01
-6.34857267e-02 5.74242592e-01 5.52768588e-01 -8.15083206e-01
7.83487499e-01 3.49572927e-01 1.04750477e-01 -1.01363015e+00
-1.67636409e-01 -3.85080963e-01 -3.77614468e-01 -4.76740599e-01
3.91573370e-01 -1.54525399e+00 -7.56462097e-01 8.84813011e-01
-8.10083747e-01 -8.77275527e-01 -3.33791107e-01 4.78321940e-01
6.46621138e-02 -6.21845573e-02 -6.16144240e-01 -1.42360783e+00
-7.39441752e-01 -8.61564994e-01 8.04938495e-01 3.41750085e-01
3.32633376e-01 -1.41999340e+00 6.01018488e-01 -5.33297777e-01
1.40218961e+00 8.48622143e-01 3.06376070e-01 -1.84867993e-01
-6.19287133e-01 -1.64579265e-02 -1.84139729e-01 3.06241214e-01
1.21973924e-01 3.29439729e-01 -1.32173300e+00 -2.64726043e-01
1.09234497e-01 -3.88813876e-02 1.36290252e+00 3.58401537e-01
7.00758696e-01 -2.93229759e-01 -1.10740155e-01 9.07415807e-01
1.71108544e+00 3.00741881e-01 3.35034370e-01 2.69466519e-01
3.43698889e-01 2.15628594e-01 -1.65429324e-01 8.97675753e-01
-1.23371761e-02 -5.02333641e-01 5.87975085e-01 8.43188614e-02
1.10317409e-01 -4.27558981e-02 3.68036836e-01 8.47603023e-01
-4.37859267e-01 3.69938537e-02 -1.57741487e+00 8.49894583e-01
-1.49397743e+00 -9.80123043e-01 -1.31943390e-01 2.18607593e+00
8.84916008e-01 -1.35921482e-02 -4.77683038e-01 -5.11079371e-01
4.71216738e-01 5.94681919e-01 -9.51473534e-01 -3.78035098e-01
-2.08617687e-01 3.12854111e-01 1.11137438e+00 8.00386190e-01
-8.84812534e-01 5.70565403e-01 6.59331560e+00 -4.25717562e-01
-1.50492489e+00 1.57407001e-01 3.91303748e-01 1.82182014e-01
-7.55236208e-01 5.55781797e-02 -1.01720071e+00 4.40351963e-01
1.74847484e+00 -4.65770453e-01 5.63844919e-01 6.07226074e-01
7.01575637e-01 -3.53881538e-01 -8.36118996e-01 2.88248122e-01
-5.52279115e-01 -2.20770574e+00 -5.00270844e-01 3.75917777e-02
8.33874941e-01 9.83959913e-01 -2.09067836e-01 5.42918444e-01
7.98110068e-01 -1.09938467e+00 7.44749665e-01 1.16208541e+00
1.25040627e+00 -1.11969672e-01 6.49855971e-01 -1.31227802e-02
-1.49069846e+00 -6.30575195e-02 -1.63739577e-01 -6.13985300e-01
5.41741252e-01 5.52449405e-01 -8.88144970e-01 4.12543833e-01
7.40057111e-01 9.42581356e-01 -1.21555172e-01 1.10180831e+00
1.25917390e-01 5.81391573e-01 -9.38608706e-01 3.43663245e-01
2.41758093e-01 -5.33707440e-01 3.82886648e-01 1.09948337e+00
8.39206040e-01 1.72338918e-01 -1.16863072e-01 1.25413370e+00
-3.71604003e-02 -3.40676099e-01 -8.37004900e-01 -4.46848243e-01
5.82044840e-01 8.20052683e-01 -1.30773395e-01 -5.09201586e-01
1.52884945e-01 2.17172205e-01 -2.96064496e-01 3.17939997e-01
-4.64102954e-01 -2.00066358e-01 1.18019891e+00 3.12420845e-01
2.01666355e-01 -3.58010024e-01 -4.58367258e-01 -1.00389707e+00
-2.40375444e-01 -1.20786391e-01 8.40073898e-02 -8.35137129e-01
-1.37271488e+00 2.66610473e-01 5.10661714e-02 -1.30012047e+00
-3.95000756e-01 -6.77181602e-01 -1.15345919e+00 1.92362773e+00
-2.35761571e+00 -4.69633341e-01 -6.31462038e-01 -3.30919735e-02
-8.63318890e-02 -1.07303135e-01 1.04303586e+00 2.17248842e-01
-5.46807289e-01 -3.42975646e-01 1.07549298e+00 -1.27032667e-03
3.05847794e-01 -1.54443359e+00 7.38927186e-01 9.22553301e-01
-1.32418263e+00 6.98640525e-01 7.89144695e-01 -1.32453251e+00
-1.67629933e+00 -1.48025930e+00 4.60102051e-01 1.43322125e-02
1.29798055e+00 7.49398693e-02 -1.43646216e+00 5.46378553e-01
2.90083081e-01 8.21759999e-01 3.38054866e-01 -3.32955837e-01
4.46539335e-02 -1.65170163e-01 -1.15320408e+00 3.71418372e-02
3.14107627e-01 -6.92879558e-01 -4.36241418e-01 3.35574508e-01
4.60790217e-01 -8.76360655e-01 -1.59329212e+00 6.74993813e-01
6.48463190e-01 -3.32091153e-01 3.30534935e-01 -8.39868426e-01
6.64051473e-01 -5.95747888e-01 -1.69317693e-01 -1.67573488e+00
8.07557628e-02 -7.30010271e-01 -5.06833792e-01 8.12526107e-01
5.64298034e-01 -7.91606307e-01 5.10398865e-01 8.89891982e-01
-4.71650541e-01 -3.30849886e-01 -9.20713484e-01 -6.78112566e-01
8.32269669e-01 -4.12061125e-01 9.21769798e-01 7.03328848e-01
-2.57826328e-01 -2.43889511e-01 -2.42250636e-01 7.91545808e-01
6.14807248e-01 5.90282232e-02 3.61787856e-01 -1.67161429e+00
9.39936116e-02 -5.47056079e-01 4.68564272e-01 -3.03197771e-01
-2.65403718e-01 -7.44427562e-01 4.52204764e-01 -1.54008436e+00
-6.48923397e-01 -8.95450115e-01 -4.39580470e-01 2.81738639e-01
6.73154742e-02 -2.43286595e-01 -2.34829821e-02 4.96901393e-01
4.99283969e-01 1.13374650e+00 9.57509577e-01 -3.13231409e-01
-5.07224977e-01 -1.39370844e-01 1.21399775e-01 3.96696925e-01
8.14199388e-01 -4.04267818e-01 2.84606576e-01 -7.80509412e-01
5.41446149e-01 7.24220216e-01 1.46684915e-01 -1.45392966e+00
3.45432311e-01 -2.91880399e-01 6.25007629e-01 -8.56001794e-01
1.63462833e-01 -6.99600160e-01 6.66683078e-01 7.04362273e-01
1.43147171e-01 1.18049048e-01 9.00880992e-01 5.45897484e-01
-3.91044319e-01 -5.08456230e-02 8.26219976e-01 -7.69546151e-01
-8.62490535e-01 7.35043168e-01 -4.75675315e-01 -1.35488346e-01
7.49432981e-01 -6.29290054e-03 -7.27554977e-01 2.14726999e-01
-8.73749375e-01 8.93809319e-01 5.34415543e-01 2.64174305e-02
5.82623243e-01 -8.83501530e-01 -9.46663320e-01 6.14104807e-01
-4.74119112e-02 6.41400516e-01 3.93296719e-01 4.98430669e-01
-1.18011618e+00 4.44091782e-02 -5.97696155e-02 -5.63289106e-01
-1.31839156e-01 -6.80587962e-02 1.33698606e+00 -2.23937839e-01
-4.88094866e-01 7.85727561e-01 -7.07605243e-01 -4.74052578e-01
-4.38985407e-01 -4.86360550e-01 -8.79333168e-02 6.25592649e-01
5.69977283e-01 3.58930647e-01 2.59704173e-01 -3.05932552e-01
-7.13265389e-02 1.10685788e-01 4.78472382e-01 -2.70774573e-01
1.81721592e+00 1.17685115e-02 -4.24941301e-01 9.70770001e-01
1.00174999e+00 -3.42502117e-01 -1.77493548e+00 -1.73452988e-01
8.03728700e-02 -2.29203016e-01 4.58153814e-01 -9.78801548e-01
-1.39491427e+00 8.70383322e-01 6.87657595e-01 1.29030198e-01
5.11830389e-01 -5.09527266e-01 3.94909739e-01 4.89125699e-01
-2.88656980e-01 -9.25094604e-01 -5.07534206e-01 1.08417940e+00
1.25455189e+00 -1.42485523e+00 3.35778117e-01 8.01356554e-01
-2.85634488e-01 1.24108100e+00 7.11070597e-01 -6.73339963e-02
1.40645266e+00 6.73998415e-01 2.31590286e-01 -5.68848312e-01
-6.08338892e-01 2.08165869e-02 -3.75454307e-01 -1.20503522e-01
2.16717087e-02 1.57123506e-01 4.29360360e-01 5.15794046e-02
8.64235982e-02 1.89563781e-01 6.64357781e-01 1.13549328e+00
-3.73514622e-01 -4.33531940e-01 -5.49981833e-01 7.24749446e-01
-1.66100428e-01 -3.55294257e-01 7.25952685e-01 4.15536702e-01
-1.41794741e-01 3.73180449e-01 5.43261230e-01 1.27968058e-01
1.38024747e-01 2.67397135e-01 -3.61614674e-01 -6.65194571e-01
-8.75023305e-01 7.69694448e-02 4.84462939e-02 -1.19143620e-01
-2.36950442e-01 -7.89823472e-01 -1.26721764e+00 -6.09693229e-01
-5.09602018e-02 5.51898479e-01 1.09456885e+00 5.78538239e-01
2.65075922e-01 6.85260415e-01 7.36769438e-01 -1.72403526e+00
-9.24723685e-01 -1.64433217e+00 -1.16749430e+00 -2.39653274e-01
8.62051070e-01 -7.04367280e-01 -9.43001568e-01 -3.65207314e-01] | [6.482611656188965, 3.0959951877593994] |
e5727e22-1ce5-4d59-8722-3fca105df6f8 | prompt-engineering-for-transformer-based | 2305.16330 | null | https://arxiv.org/abs/2305.16330v1 | https://arxiv.org/pdf/2305.16330v1.pdf | Prompt Engineering for Transformer-based Chemical Similarity Search Identifies Structurally Distinct Functional Analogues | Chemical similarity searches are widely used in-silico methods for identifying new drug-like molecules. These methods have historically relied on structure-based comparisons to compute molecular similarity. Here, we use a chemical language model to create a vector-based chemical search. We extend implementations by creating a prompt engineering strategy that utilizes two different chemical string representation algorithms: one for the query and the other for the database. We explore this method by reviewing the search results from five drug-like query molecules (penicillin G, nirmatrelvir, zidovudine, lysergic acid diethylamide, and fentanyl) and three dye-like query molecules (acid blue 25, avobenzone, and 2-diphenylaminocarbazole). We find that this novel method identifies molecules that are functionally similar to the query, indicated by the associated patent literature, and that many of these molecules are structurally distinct from the query, making them unlikely to be found with traditional chemical similarity search methods. This method may aid in the discovery of novel structural classes of molecules that achieve target functionality. | ['Andrew D. Ellington', 'Claus O. Wilke', 'Aaron L. Feller', 'Clayton W. Kosonocky'] | 2023-05-17 | null | null | null | null | ['prompt-engineering'] | ['natural-language-processing'] | [ 6.23754978e-01 -1.78468332e-01 -4.32920009e-01 4.00276445e-02
-7.75885403e-01 -1.39059079e+00 5.54152966e-01 7.36120582e-01
-2.64413297e-01 1.26274788e+00 -1.43675217e-02 -1.01361763e+00
-1.46231994e-01 -6.13738954e-01 -2.76426286e-01 -5.32069325e-01
-1.73441898e-02 3.33438367e-01 4.03165460e-01 -3.15173179e-01
8.15712273e-01 1.00616682e+00 -1.03702056e+00 2.59287894e-01
9.48842347e-01 4.58785892e-01 1.43029511e-01 2.17344835e-01
-1.44119009e-01 3.94011259e-01 -7.86965668e-01 -4.23806429e-01
5.15510514e-02 -7.12947845e-01 -8.00869465e-01 -7.48119652e-01
2.80405253e-01 3.92465368e-02 2.03222305e-01 6.81091607e-01
7.17571199e-01 1.97435111e-01 1.02929080e+00 -5.76787233e-01
-4.95041877e-01 -3.29645723e-02 -1.35615453e-01 1.26363039e-01
1.07578707e+00 2.10532859e-01 8.32206190e-01 -1.04544544e+00
8.62503111e-01 1.09064507e+00 6.68184340e-01 4.07047778e-01
-1.18140769e+00 -6.52885854e-01 -3.15130562e-01 9.57775116e-02
-1.53236818e+00 -3.59028697e-01 2.77764440e-01 -6.30397916e-01
1.54817045e+00 4.98191655e-01 5.09812653e-01 7.36859858e-01
6.36599958e-01 3.35734859e-02 7.82010376e-01 -3.97454411e-01
5.87651968e-01 1.44669116e-01 -3.77520502e-01 6.92280531e-01
4.77955997e-01 3.38083476e-01 -3.24513614e-01 -1.15096068e+00
3.26777577e-01 -5.68475574e-02 -3.30934584e-01 -4.74577367e-01
-7.28958130e-01 1.08302546e+00 2.65430093e-01 4.15162981e-01
-4.45404083e-01 -1.12785764e-01 2.10939988e-01 1.92667153e-02
-4.71776575e-02 1.49717259e+00 -5.47675669e-01 7.23578781e-02
-5.00156879e-01 2.98651546e-01 1.07447159e+00 4.20087665e-01
6.06072128e-01 -8.14255774e-02 2.24999681e-01 3.39322418e-01
3.05811495e-01 -1.68198925e-02 2.52831489e-01 -5.41794956e-01
-9.76565573e-03 3.14368635e-01 3.88395309e-01 -7.40788639e-01
-2.82919079e-01 1.01440184e-01 5.80888763e-02 2.65714884e-01
3.03489864e-01 -1.56274095e-01 -7.82953143e-01 1.36721480e+00
2.65258640e-01 -8.26480389e-02 3.91114652e-01 2.90124148e-01
7.27804422e-01 5.40867090e-01 5.29691696e-01 -6.56351507e-01
1.03274262e+00 -5.66309988e-01 -2.75235027e-01 4.74496454e-01
8.26571286e-01 -1.15084589e+00 5.94170094e-01 5.56158304e-01
-8.50640595e-01 -1.83830753e-01 -1.32045615e+00 5.33693552e-01
-8.20010841e-01 -4.67407852e-01 8.41140747e-01 1.04065371e+00
-6.28779292e-01 1.15901399e+00 -5.02156377e-01 -2.58288831e-01
4.07043956e-02 6.52421176e-01 -3.81170660e-01 -4.80870530e-02
-1.15831649e+00 1.25462675e+00 5.59834957e-01 -5.45519769e-01
-6.49281621e-01 -6.83817565e-01 -5.99929571e-01 -9.56607014e-02
1.24413252e-01 -7.09185839e-01 9.59601343e-01 -2.92080313e-01
-1.53254616e+00 4.28964406e-01 -2.70931214e-01 -2.30904475e-01
-8.23593587e-02 4.29452926e-01 -5.44816911e-01 2.63437688e-01
1.99968129e-01 3.50717604e-01 2.74172187e-01 -8.86645257e-01
-4.18690354e-01 -3.13419402e-01 -6.40768232e-03 3.05443197e-01
1.90234542e-01 3.49440455e-01 9.29714665e-02 -5.93732417e-01
-1.62261263e-01 -6.97022736e-01 -6.20940328e-01 -2.17699692e-01
-3.13261598e-01 -3.00849468e-01 3.49204868e-01 -1.60992369e-01
1.11741078e+00 -1.65552616e+00 -1.96335196e-01 8.90100420e-01
-6.49877489e-02 4.70381707e-01 -2.86103666e-01 1.26714802e+00
-5.94967425e-01 6.84742510e-01 2.51183156e-02 9.63149130e-01
-4.65535790e-01 -3.84725690e-01 -3.50803226e-01 4.54192281e-01
1.81539342e-01 6.63598776e-01 -1.15783358e+00 -5.41305989e-02
-1.64968744e-02 3.82798344e-01 -4.72247392e-01 -1.78069472e-01
-4.69551086e-01 3.69444609e-01 -9.37255442e-01 1.00810695e+00
3.63361120e-01 -5.94626516e-02 6.42157435e-01 -2.91754395e-01
-2.67654628e-01 4.90741134e-01 -7.23338842e-01 1.29289353e+00
9.12950188e-02 -3.05224750e-02 -6.93328023e-01 -3.98875684e-01
9.63900447e-01 4.88093168e-01 5.03389359e-01 -6.73566580e-01
-1.37383118e-01 5.48372507e-01 2.73320436e-01 -2.08827659e-01
-7.52252117e-02 -4.14221197e-01 4.10577059e-01 2.25590512e-01
-3.43566447e-01 -4.04457241e-01 5.03777742e-01 -1.30058546e-03
1.43577504e+00 1.09238297e-01 7.11875081e-01 -1.71735406e-01
4.53270972e-01 3.52068782e-01 4.71142590e-01 6.43163145e-01
2.14141592e-01 -2.05513346e-03 1.26976207e-01 -4.13418740e-01
-9.05670702e-01 -8.67154837e-01 -2.76090741e-01 7.17191756e-01
1.35698244e-01 -8.23293328e-01 -3.02649051e-01 -3.27370256e-01
2.49476895e-01 5.29723167e-01 -3.37018758e-01 -2.39658102e-01
-1.98954299e-01 -5.02143204e-01 5.78627884e-01 5.91757894e-03
-4.71471511e-02 -5.56187689e-01 -3.95704091e-01 8.40569735e-01
6.19084835e-01 -2.09619492e-01 -2.41610020e-01 4.53495026e-01
-6.75132811e-01 -1.36700714e+00 -6.97660804e-01 -6.53773963e-01
1.69366553e-01 5.20235859e-02 7.28896856e-01 3.81811373e-02
-5.61041117e-01 5.02015352e-02 -2.94202745e-01 -6.38214648e-01
-5.79064369e-01 -3.89074177e-01 2.57315904e-01 -6.90911412e-01
4.73125219e-01 -4.69830304e-01 -7.14738905e-01 3.51294160e-01
-5.10266542e-01 -7.54740655e-01 6.13332331e-01 8.71015310e-01
7.60511041e-01 -9.37302411e-02 6.86078548e-01 -9.08437490e-01
1.13880932e+00 -6.75376356e-01 -4.23936844e-01 5.74713349e-01
-9.55485582e-01 1.51411057e-01 3.39785069e-01 -8.44281673e-01
-5.22804081e-01 5.35465300e-01 -2.57195622e-01 1.26675218e-01
8.09697732e-02 1.04692090e+00 -1.59820348e-01 -7.66973257e-01
1.05259025e+00 1.35887906e-01 3.28835249e-02 -5.62564790e-01
1.56591877e-01 3.53861690e-01 3.07576060e-01 -6.20066762e-01
5.43295741e-01 1.03576943e-01 1.04553208e-01 -7.48779774e-01
2.08373338e-01 -4.72453237e-01 1.88535050e-01 5.55191457e-01
7.20616817e-01 -7.02728927e-01 -1.19893861e+00 -3.02796572e-01
-1.24148309e+00 1.79470092e-01 3.58271673e-02 6.82687521e-01
-3.45206022e-01 6.60127997e-01 -2.84758747e-01 -4.78588134e-01
-3.55566293e-01 -1.20812893e+00 5.94676554e-01 3.13494429e-02
-7.89512873e-01 -7.36338496e-01 5.93452096e-01 5.06747253e-02
2.24702582e-01 6.19600654e-01 1.56458938e+00 -1.25752115e+00
-5.05921245e-01 -2.45951161e-01 1.79769441e-01 -3.21365595e-01
5.41710496e-01 2.52162069e-01 -2.91380554e-01 -3.71582866e-01
-1.71564579e-01 -2.27127075e-01 4.78507489e-01 2.00437561e-01
5.60928404e-01 -3.85133803e-01 -9.53192174e-01 4.86671299e-01
1.35748887e+00 1.59631014e+00 8.39009583e-01 3.55084866e-01
2.93604195e-01 3.37381363e-01 6.07849061e-01 3.62702906e-01
-5.20575464e-01 6.78892851e-01 2.13201538e-01 -2.61792213e-01
4.86476779e-01 -2.92219698e-01 -9.89911482e-02 -4.36199069e-01
-9.97599065e-02 -5.43012679e-01 -7.93938577e-01 1.97623655e-01
-1.29014838e+00 -1.13386691e+00 -2.10484155e-02 2.22757077e+00
1.23014975e+00 -1.58834755e-01 2.10514352e-01 -2.91063756e-01
4.01583493e-01 -4.17019963e-01 -9.07844961e-01 -7.16802955e-01
-1.40576661e-01 1.08962095e+00 6.25986934e-01 4.91398633e-01
-8.06491137e-01 9.33593214e-01 7.93823814e+00 1.04853880e+00
-9.92195010e-01 -6.00945771e-01 2.75340617e-01 3.70721608e-01
-6.47652507e-01 6.51018560e-01 -5.99443197e-01 3.13285589e-01
9.22510386e-01 -6.90410316e-01 1.02046862e-01 7.37543344e-01
3.25058907e-01 -2.58141011e-01 -1.40518641e+00 8.75019133e-01
-3.05933744e-01 -2.20504928e+00 4.47775275e-01 1.68938532e-01
6.74049377e-01 -5.12662351e-01 -1.48270965e-01 -2.93742418e-01
1.85650408e-01 -1.31329393e+00 1.33776143e-01 2.50047833e-01
8.07692289e-01 -8.95135641e-01 1.99897632e-01 -2.38275275e-01
-9.31478202e-01 -4.03743684e-02 -3.50319445e-01 1.69696823e-01
-5.09270839e-02 7.31182918e-02 -1.06578350e+00 3.57694387e-01
1.92302346e-01 4.03925896e-01 -2.53118396e-01 1.43388760e+00
3.86936739e-02 1.20428458e-01 -3.15940022e-01 -4.60230023e-01
2.53751636e-01 -2.71554619e-01 3.51002872e-01 7.84527659e-01
1.43213913e-01 4.13941354e-01 3.61650735e-01 7.08344758e-01
1.32788077e-01 4.85371351e-01 -9.15447533e-01 -6.21441543e-01
6.74367666e-01 5.29086113e-01 -5.70529997e-01 -3.86051983e-01
-3.25762302e-01 8.85226905e-01 -5.19165218e-01 6.40903413e-01
-2.57889748e-01 -1.01173329e+00 8.25578749e-01 3.77368294e-02
3.13864022e-01 1.19732663e-01 3.18997115e-01 -3.78303170e-01
-5.59059203e-01 -1.39991784e+00 4.15724546e-01 -6.47108555e-01
-1.14727855e+00 4.74903524e-01 8.65557492e-02 -1.05166638e+00
-3.81090224e-01 -5.90335846e-01 -7.25106120e-01 1.35446644e+00
-7.35956967e-01 -5.38724899e-01 4.98279154e-01 4.70126390e-01
1.16173796e-01 -4.51234043e-01 1.25284028e+00 5.02249338e-02
4.25134972e-02 4.02917653e-01 4.94715005e-01 -7.09653795e-01
9.39627945e-01 -6.53834879e-01 2.10479662e-01 9.29927453e-02
-2.53202170e-01 1.35705554e+00 8.91197205e-01 -1.05101120e+00
-1.21745467e+00 -4.95494515e-01 7.27163970e-01 -4.43817079e-01
6.19081020e-01 2.04054907e-01 -6.80707872e-01 9.04894993e-02
3.17263454e-02 -9.27400589e-01 1.51442158e+00 -3.17632049e-01
-5.84394574e-01 6.27419889e-01 -1.40772474e+00 7.57141232e-01
7.07960486e-01 -6.08249545e-01 -4.63344008e-01 8.34398210e-01
7.11966574e-01 -1.72590390e-01 -1.01595795e+00 5.04110813e-01
8.35783362e-01 -4.97243106e-01 1.19974351e+00 -1.17325044e+00
-1.19984932e-01 -7.87049472e-01 -3.37913856e-02 -1.01005352e+00
-3.23501736e-01 -9.34472919e-01 3.94256324e-01 4.08755064e-01
7.93068409e-01 -8.91316116e-01 7.88656294e-01 6.28513396e-01
1.27260581e-01 -9.07939255e-01 -8.65607381e-01 -1.22908413e+00
1.48898035e-01 6.50994956e-01 6.28854692e-01 1.06165648e+00
5.73207319e-01 8.62540066e-01 1.55497137e-02 -4.65278238e-01
-5.98485731e-02 2.09134240e-02 2.41375893e-01 -1.33980954e+00
-4.95525122e-01 -4.92064595e-01 -6.34183407e-01 -6.72597110e-01
-2.50276804e-01 -7.53428996e-01 -6.53763950e-01 -1.38037002e+00
1.85589194e-01 -2.52931267e-01 3.81472893e-02 5.43085039e-01
3.26263577e-01 -1.82556987e-01 -3.40618283e-01 2.49118790e-01
1.46460801e-01 -3.44374515e-02 6.67743683e-01 -4.15400565e-01
-7.22416639e-01 -2.25489531e-02 -1.09709477e+00 4.28864717e-01
6.70560777e-01 -5.05328417e-01 -6.48801327e-01 4.85746741e-01
4.55026865e-01 3.98423746e-02 -2.73514479e-01 -3.23053420e-01
3.49789672e-02 -7.37615824e-01 4.47487563e-01 -4.69274938e-01
2.12329924e-01 -5.39543688e-01 8.65012825e-01 1.11773133e+00
-1.13452226e-01 1.12144522e-01 3.94927561e-01 6.57557786e-01
-6.94399849e-02 -5.14676332e-01 7.06238866e-01 -3.24108899e-01
-5.41806042e-01 2.64687955e-01 -8.46773624e-01 -5.24332345e-01
9.94992971e-01 -9.46141899e-01 -3.95113409e-01 -2.68604815e-01
-6.90373003e-01 -2.39665285e-01 7.19694912e-01 1.01706490e-01
8.92168701e-01 -1.01020622e+00 -3.76934230e-01 -3.15688938e-01
5.88927925e-01 -8.07499170e-01 -4.38528091e-01 4.17798489e-01
-9.43987250e-01 6.43083632e-01 -1.26702532e-01 -3.28524970e-02
-1.54764545e+00 1.02207208e+00 5.27013004e-01 8.54325965e-02
2.67222703e-01 8.89224648e-01 2.20080376e-01 1.02706894e-01
1.63665414e-01 2.29979176e-02 -5.23611158e-02 -1.24800220e-01
4.62297022e-01 6.54448643e-02 1.07321076e-01 -1.49267972e-01
-9.04619098e-01 6.45407021e-01 -1.79494277e-01 2.00636119e-01
1.07385254e+00 6.76238596e-01 -1.21427059e-01 -2.88916290e-01
1.40658653e+00 1.50235161e-01 -1.83122829e-01 1.27291039e-01
2.88089842e-01 -5.61334252e-01 -4.35435116e-01 -1.04574144e+00
-3.57225239e-02 3.10863912e-01 8.63429606e-01 -3.76765490e-01
5.61268866e-01 -5.07541224e-02 3.20934296e-01 8.58316898e-01
3.54135871e-01 -8.06695104e-01 1.69678196e-01 1.70009956e-01
7.62115538e-01 -6.34537637e-01 4.65516478e-01 -5.51029623e-01
-5.44569418e-02 1.05856538e+00 2.20126137e-01 4.88702208e-01
2.61719584e-01 -2.47311011e-01 -3.09876680e-01 -5.45363724e-01
-7.49497712e-01 -1.31831080e-01 2.04823956e-01 7.45619357e-01
7.24772215e-01 -4.04618233e-01 -1.17895305e+00 -2.81613804e-02
3.37779462e-01 -1.90710202e-01 2.73401052e-01 1.42975223e+00
-5.29605329e-01 -1.94734848e+00 -2.53151745e-01 1.78481758e-01
-5.68545759e-01 -7.07493722e-01 -1.42341971e+00 6.68155789e-01
-7.05271289e-02 1.03697264e+00 -4.13672417e-01 -2.99510002e-01
5.07022738e-01 1.29365236e-01 6.04860246e-01 -7.58248210e-01
-6.75524235e-01 4.76632297e-01 5.24746358e-01 -2.73328155e-01
-1.92910969e-01 -1.04420856e-01 -1.27022266e+00 -2.11322308e-01
-7.90960848e-01 7.55676985e-01 7.87384450e-01 6.10878646e-01
6.31737173e-01 -1.70275137e-01 7.15799034e-01 -4.09889847e-01
-4.05669898e-01 -4.05982137e-01 -4.89863932e-01 -1.20515074e-03
-1.31770149e-01 -6.14579439e-01 -9.05376375e-02 8.16639066e-02] | [4.9503889083862305, 5.748427867889404] |
2572c1b2-c828-492a-b72d-c05f8047c15e | differentially-private-sliced-inverse | 2306.06324 | null | https://arxiv.org/abs/2306.06324v1 | https://arxiv.org/pdf/2306.06324v1.pdf | Differentially private sliced inverse regression in the federated paradigm | We extend the celebrated sliced inverse regression to address the challenges of decentralized data, prioritizing privacy and communication efficiency. Our approach, federated sliced inverse regression (FSIR), facilitates collaborative estimation of the sufficient dimension reduction subspace among multiple clients, solely sharing local estimates to protect sensitive datasets from exposure. To guard against potential adversary attacks, FSIR further employs diverse perturbation strategies, including a novel multivariate Gaussian mechanism that guarantees differential privacy at a low cost of statistical accuracy. Additionally, FSIR naturally incorporates a collaborative variable screening step, enabling effective handling of high-dimensional client data. Theoretical properties of FSIR are established for both low-dimensional and high-dimensional settings, supported by extensive numerical experiments and real data analysis. | ['Xin Chen', 'Jiarui Zhang', 'Shuaida He'] | 2023-06-10 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [-1.18893810e-01 -9.05322134e-02 -3.86632383e-01 -2.81898528e-01
-1.16687155e+00 -9.72568035e-01 1.96859062e-01 -1.02160200e-01
-2.44139746e-01 8.19040000e-01 1.62663177e-01 -4.33728039e-01
-4.35638517e-01 -6.80750847e-01 -5.46227813e-01 -1.21064579e+00
-5.15415609e-01 3.52382362e-01 -4.70488042e-01 3.85012925e-02
1.51362851e-01 5.72322845e-01 -6.57968700e-01 -1.48208532e-02
5.77768683e-01 1.13861847e+00 -4.23527658e-01 4.10232604e-01
3.22097868e-01 3.49240661e-01 -4.20221090e-01 -7.67762899e-01
7.36157238e-01 7.26154819e-02 -5.46688557e-01 -1.75843433e-01
-2.08045542e-01 -6.93063498e-01 -4.64339286e-01 9.68266308e-01
6.93558395e-01 -9.35793146e-02 4.18883085e-01 -1.65271854e+00
-7.58651197e-01 6.41365647e-01 -9.13282335e-01 -1.70245464e-03
3.90967041e-01 -2.03877594e-02 9.24398303e-01 -8.24210286e-01
4.64114249e-01 1.13907671e+00 8.26959968e-01 4.62993532e-01
-1.79768550e+00 -1.09353673e+00 2.74897534e-02 -5.31936176e-02
-1.77544641e+00 -4.85715687e-01 6.30482793e-01 -1.62705943e-01
3.61893326e-01 6.19771659e-01 -1.41677931e-01 8.70438576e-01
-1.29051760e-01 5.93075514e-01 9.43814754e-01 7.21948966e-02
4.36926603e-01 5.83533049e-01 -1.00379623e-01 1.88022032e-01
3.34700227e-01 7.65100718e-02 -6.29982471e-01 -1.30789733e+00
5.90509653e-01 2.36856163e-01 -6.37365460e-01 -9.20688212e-01
-6.47965848e-01 9.75867212e-01 -1.43091502e-02 -4.72428203e-01
-5.05494416e-01 -7.69819170e-02 3.66545439e-01 5.55138052e-01
4.70998973e-01 -2.95934260e-01 -7.75770545e-01 2.77090162e-01
-3.42372835e-01 2.81188279e-01 1.04766750e+00 1.37673521e+00
6.52921319e-01 -4.61268008e-01 -2.80789286e-01 4.70010549e-01
4.05694306e-01 7.24503219e-01 -2.58668870e-01 -9.70006406e-01
9.53314126e-01 1.72216654e-01 3.42718214e-01 -1.00270343e+00
-1.17474131e-01 -2.72984177e-01 -1.17661285e+00 6.67283684e-02
2.69247770e-01 -6.01149917e-01 5.07315025e-02 1.99897885e+00
8.36889386e-01 -5.81805408e-03 2.53305078e-01 9.91977394e-01
2.36151018e-03 2.76262105e-01 -1.06129825e-01 -7.02630997e-01
9.24151182e-01 -2.31317475e-01 -6.39444172e-01 5.19351542e-01
7.11328089e-01 7.29701445e-02 4.88622457e-01 4.92368102e-01
-9.17991817e-01 5.65970778e-01 -5.93694270e-01 1.07009366e-01
5.95498756e-02 -4.43026990e-01 6.84539080e-01 1.09411180e+00
-9.22014594e-01 2.59233177e-01 -5.37797332e-01 1.84385449e-01
8.68985653e-01 7.88839161e-01 -6.44048452e-01 -1.31069273e-01
-1.09713066e+00 -2.01129578e-02 -5.15958488e-01 -2.01787114e-01
-5.04193246e-01 -9.77797031e-01 -5.02575278e-01 8.39164481e-02
6.29977703e-01 -6.58450305e-01 9.47018385e-01 -1.80662304e-01
-1.11982441e+00 5.57816029e-01 -2.29463428e-01 -6.68718636e-01
9.74394083e-01 -5.08956462e-02 -1.94128692e-01 1.90335438e-01
-9.08838138e-02 -5.46348751e-01 6.73030734e-01 -1.21335340e+00
-3.03585202e-01 -1.27353883e+00 -2.63369709e-01 9.15547982e-02
-6.12907648e-01 2.05499664e-01 -4.06861693e-01 -4.60145175e-01
1.07088864e-01 -7.62451291e-01 -7.38347352e-01 5.83129287e-01
-4.87191647e-01 1.45880282e-01 9.73136425e-01 -5.59467018e-01
1.32907259e+00 -2.09889150e+00 2.23540254e-02 8.50368440e-01
5.81967175e-01 -2.29520738e-01 1.06527574e-01 5.88892996e-01
1.10829942e-01 2.56649405e-01 -4.33379635e-02 -6.55114949e-01
9.72156376e-02 -7.65627101e-02 -6.43514872e-01 1.16265428e+00
-5.16256809e-01 4.62239623e-01 -5.07423460e-01 -3.56300861e-01
-1.90110162e-01 6.63389087e-01 -7.88334489e-01 1.86011568e-01
1.67097285e-01 4.93711561e-01 -1.06562483e+00 6.43894315e-01
1.30192709e+00 -2.72724301e-01 4.07311678e-01 -3.46463220e-03
2.11721152e-01 -1.00776494e-01 -1.38595510e+00 1.19592476e+00
-3.56147468e-01 -1.30001828e-01 9.83038425e-01 -7.91474819e-01
7.74939895e-01 5.02485931e-01 9.15626764e-01 -4.15600836e-01
2.89675504e-01 -4.97491844e-02 -9.54400122e-01 -2.21221253e-01
1.79050550e-01 2.03372911e-01 -3.79840225e-01 8.97655070e-01
-5.57381451e-01 4.78905231e-01 -6.67656541e-01 5.34074664e-01
1.07400918e+00 -4.98593807e-01 3.11527193e-01 -2.59771287e-01
4.64682966e-01 -3.84887755e-01 7.70878673e-01 8.35246205e-01
-3.04361910e-01 4.73127395e-01 6.24301136e-01 -1.73992053e-01
-6.60240710e-01 -1.08166993e+00 -2.95135409e-01 1.07990527e+00
3.02574873e-01 -4.85542297e-01 -5.06874263e-01 -8.20392728e-01
7.82345295e-01 5.15189052e-01 -6.55136287e-01 1.32846534e-01
-9.03831646e-02 -8.79826009e-01 3.37819427e-01 6.16770945e-02
9.21427235e-02 -2.70541608e-02 -3.70878587e-03 -2.34155860e-02
6.98322207e-02 -7.96080053e-01 -8.29680264e-01 1.92423746e-01
-5.39715290e-01 -1.06632769e+00 -3.44236583e-01 -4.88480292e-02
6.03209436e-01 7.47500956e-01 6.52338684e-01 -2.18893901e-01
-2.98853308e-01 6.57791138e-01 -2.24111639e-02 -5.46895303e-02
-1.65734798e-01 -3.92464459e-01 3.23393434e-01 4.14070010e-01
4.51466769e-01 -9.72168565e-01 -8.74957144e-01 4.88336980e-01
-6.98500395e-01 -6.82982087e-01 2.37988666e-01 5.29546797e-01
7.42637396e-01 -6.50660992e-02 4.62560594e-01 -1.04659188e+00
7.34663486e-01 -1.11781895e+00 -7.62698948e-01 3.19883406e-01
-8.96803975e-01 4.74122167e-02 8.65664482e-01 -5.20549119e-01
-8.49427938e-01 -5.21903932e-02 3.82376194e-01 -6.62981093e-01
2.53771931e-01 6.50731334e-03 -4.59384680e-01 -3.67909074e-01
6.34019732e-01 2.25362509e-01 2.56212026e-01 -7.33147323e-01
5.61894715e-01 1.06060052e+00 4.26663131e-01 -5.74290395e-01
1.14265323e+00 9.36973333e-01 2.68371016e-01 -4.13955390e-01
-3.20197999e-01 -3.01451713e-01 -2.86149949e-01 2.67771095e-01
2.98921585e-01 -1.05925369e+00 -1.51059771e+00 3.44493359e-01
-7.40864456e-01 9.46596079e-03 -5.58735579e-02 2.51854122e-01
-5.27042449e-01 6.03944659e-01 -5.58925688e-01 -1.26477730e+00
-6.54590547e-01 -7.73897767e-01 7.93361247e-01 -1.18701883e-01
1.83606103e-01 -8.40797484e-01 3.41403410e-02 4.93705839e-01
5.57228804e-01 2.57193118e-01 6.40586793e-01 -9.17595506e-01
-8.48008156e-01 -4.65637147e-01 -1.83779135e-01 -1.11840568e-01
1.24312686e-02 -2.79175431e-01 -6.93184137e-01 -7.03354657e-01
1.65340677e-01 -3.08178782e-01 2.82112479e-01 1.33077368e-01
1.59788609e+00 -1.02371883e+00 -4.39949602e-01 1.00063312e+00
1.37760961e+00 -4.57616597e-01 3.82911325e-01 7.89335147e-02
2.79893100e-01 4.54409599e-01 5.88766992e-01 1.54341006e+00
5.77488363e-01 6.12614989e-01 4.96880770e-01 -1.78567637e-02
7.53273487e-01 -1.59816951e-01 1.61599353e-01 5.29468767e-02
6.18696868e-01 -2.54646778e-01 -4.11094964e-01 3.21818620e-01
-1.65257537e+00 -8.46624494e-01 -2.14614064e-01 2.67478228e+00
8.10479403e-01 -5.95016658e-01 3.34906310e-01 -3.84488478e-02
8.58150780e-01 6.31115139e-02 -9.32997227e-01 -2.13867724e-01
-1.84755579e-01 -1.96967825e-01 1.26821375e+00 9.32009816e-02
-9.68808770e-01 3.19587976e-01 6.51184893e+00 7.18156278e-01
-5.94828784e-01 1.96121573e-01 8.29158664e-01 -4.62394416e-01
-5.28761446e-01 -1.32922277e-01 -7.52764523e-01 3.79984498e-01
7.19143391e-01 -7.00070083e-01 9.06663835e-01 1.21460080e+00
2.14645535e-01 2.87744850e-01 -8.33380461e-01 9.85682666e-01
-1.24672092e-01 -1.34133375e+00 -3.57850730e-01 5.06647348e-01
5.94914436e-01 8.75116512e-02 4.20370013e-01 7.99194723e-02
7.25546002e-01 -6.88115656e-01 7.86619931e-02 2.85866886e-01
9.73552048e-01 -1.13770592e+00 1.90656304e-01 3.27163011e-01
-9.58975017e-01 -5.79835236e-01 -2.65440345e-01 4.31126505e-01
1.03835426e-01 3.57888311e-01 -5.70242226e-01 5.33083677e-01
8.69900286e-01 1.46276280e-01 -2.68592462e-02 7.05949306e-01
3.50039154e-01 2.09054098e-01 -7.19311953e-01 2.25111201e-01
-3.89592558e-01 -5.12335420e-01 7.40937710e-01 9.22410190e-01
1.11658603e-01 3.76303643e-01 1.04749523e-01 7.39626408e-01
-1.25120968e-01 3.10132235e-01 -5.09279788e-01 3.07702035e-01
1.17601085e+00 1.15053248e+00 2.23340437e-01 1.67453200e-01
-2.70963579e-01 9.82586205e-01 3.33576709e-01 5.32695115e-01
-5.50791323e-01 -1.83396801e-01 1.46225631e+00 -2.19104588e-01
4.36646640e-01 -1.89345822e-01 -6.36391997e-01 -1.01291096e+00
9.26222503e-02 -7.92917669e-01 8.51792872e-01 -1.10810213e-02
-1.56233454e+00 8.22835416e-02 -2.98011720e-01 -1.11232674e+00
-2.82697618e-01 9.66833606e-02 -4.36230808e-01 9.08518255e-01
-1.06621540e+00 -8.31839859e-01 4.46024127e-02 1.18389022e+00
-5.22503316e-01 -8.58292580e-02 9.19834912e-01 3.32625657e-01
-7.72702456e-01 1.41622055e+00 8.25318396e-01 -6.95632473e-02
6.64463818e-01 -9.41748977e-01 2.69564986e-01 5.05031049e-01
-2.90999532e-01 8.24271142e-01 7.92086244e-01 -5.44391215e-01
-2.25689793e+00 -1.19836009e+00 2.19606206e-01 -3.42282325e-01
6.42920792e-01 -6.73099637e-01 -9.64585483e-01 5.61035454e-01
-6.40874147e-01 3.60136688e-01 1.16685367e+00 8.29711854e-02
-8.75132024e-01 -4.11169559e-01 -2.06202793e+00 5.99323630e-01
9.82979715e-01 -6.79988444e-01 3.80227447e-01 5.22639692e-01
9.68188107e-01 -9.92816291e-04 -1.29348421e+00 1.60262868e-01
5.76598942e-01 -9.04220402e-01 1.09405744e+00 -7.44057536e-01
-4.15536016e-01 6.78531006e-02 -6.00578368e-01 -6.02219820e-01
-1.83846667e-01 -1.45563889e+00 -7.08038867e-01 1.50101936e+00
2.81639308e-01 -9.71583545e-01 1.20445251e+00 1.48351789e+00
8.27893853e-01 -6.27579272e-01 -1.23787236e+00 -6.71029150e-01
1.11229271e-02 -4.56584692e-01 8.04915190e-01 9.21888888e-01
7.64563680e-02 -2.02124298e-01 -7.72263706e-01 7.21286476e-01
1.38014269e+00 6.33642524e-02 1.31451416e+00 -9.16106582e-01
-5.51311851e-01 3.96269001e-02 -3.00218701e-01 -9.70947087e-01
1.52433828e-01 -6.48039341e-01 -5.60578108e-01 -7.69132674e-01
3.28320533e-01 -7.03610599e-01 -3.45859349e-01 1.85511619e-01
7.34344125e-02 1.95103154e-01 -1.14428073e-01 2.59917319e-01
-7.23947406e-01 7.79218376e-01 6.72800064e-01 3.78799975e-01
-3.60729188e-01 6.54507458e-01 -1.13805115e+00 -5.90408817e-02
5.82404196e-01 -6.47646129e-01 -3.62152606e-01 -4.39030640e-02
-1.22769224e-02 5.13676941e-01 2.02328697e-01 -2.25225389e-01
3.66336644e-01 -4.61093187e-01 7.82878473e-02 -8.00113559e-01
4.16449666e-01 -1.23887706e+00 3.35551322e-01 3.09802324e-01
-3.26199323e-01 -4.37994093e-01 -4.13543582e-01 1.12771571e+00
4.99177158e-01 5.77643692e-01 8.32599699e-01 5.98527014e-01
2.02884927e-01 9.47588027e-01 -7.20766112e-02 -1.28420174e-01
1.38999534e+00 1.70708857e-02 -2.99219608e-01 -8.51655364e-01
-4.34083223e-01 7.52785802e-01 6.37906849e-01 8.66490155e-02
4.06929016e-01 -9.82471049e-01 -8.07023108e-01 5.33962607e-01
4.51123156e-03 -1.32214442e-01 5.35710335e-01 1.05710566e+00
2.21310362e-01 1.83872879e-01 4.51785654e-01 -3.54176372e-01
-1.38363492e+00 1.11291015e+00 1.33847058e-01 8.10494572e-02
-6.99162304e-01 7.54479587e-01 9.59049538e-02 -4.54423100e-01
7.16830611e-01 5.39293885e-01 3.25527489e-01 -1.61526263e-01
9.69928026e-01 5.37458539e-01 -1.27669975e-01 -4.55416977e-01
-3.87120515e-01 2.66733598e-02 -2.36362845e-01 -1.01426452e-01
1.56818891e+00 -9.32494402e-01 -2.08423540e-01 -2.24034414e-01
1.61650026e+00 4.04464811e-01 -1.35949600e+00 -9.07730639e-01
-1.14548970e-02 -8.80400181e-01 1.09913625e-01 -4.24521863e-01
-1.25519717e+00 2.76178837e-01 4.28360194e-01 7.24245086e-02
1.26926184e+00 -3.46759148e-02 6.73386335e-01 3.91626149e-01
9.30724919e-01 -8.01055253e-01 -5.82740903e-01 -2.63515897e-02
6.14143610e-01 -8.43178034e-01 1.29820168e-01 -4.43266779e-01
-7.94021726e-01 8.05269182e-01 1.61783978e-01 -9.81758833e-02
8.86439323e-01 6.83677197e-01 -2.03236595e-01 1.39337957e-01
-1.18500257e+00 4.09741521e-01 -3.06271583e-01 7.38244534e-01
-2.63667673e-01 -1.46603491e-02 -2.48829588e-01 1.32833946e+00
2.04652771e-01 -3.16241831e-01 4.05737132e-01 7.50838518e-01
1.91616267e-02 -1.16966200e+00 -8.06679010e-01 5.63642204e-01
-7.79760957e-01 2.03883126e-01 -4.17102844e-01 3.67392272e-01
-6.96017444e-01 9.79149044e-01 -1.37909725e-01 -4.57666308e-01
-7.45710433e-02 -2.65113235e-01 -1.30256444e-01 5.74266724e-02
-6.36649370e-01 4.05236986e-03 -2.52174705e-01 -9.80660558e-01
2.64753759e-01 -8.93110216e-01 -1.00666082e+00 -7.35192418e-01
-3.30291182e-01 3.62805337e-01 9.84916627e-01 5.18022120e-01
8.74649465e-01 -4.37068909e-01 1.80455339e+00 -4.83911932e-01
-1.64156699e+00 -2.97921687e-01 -1.08026004e+00 2.65089482e-01
5.88410795e-01 -3.36278915e-01 -8.56310070e-01 -6.20115399e-01] | [5.929462909698486, 6.471562385559082] |
9f18a663-b8a4-429c-9983-42a6b224d21a | upb-at-semeval-2020-task-6-pretrained | 2009.05603 | null | https://arxiv.org/abs/2009.05603v2 | https://arxiv.org/pdf/2009.05603v2.pdf | UPB at SemEval-2020 Task 6: Pretrained Language Models for Definition Extraction | This work presents our contribution in the context of the 6th task of SemEval-2020: Extracting Definitions from Free Text in Textbooks (DeftEval). This competition consists of three subtasks with different levels of granularity: (1) classification of sentences as definitional or non-definitional,(2) labeling of definitional sentences, and (3) relation classification. We use various pretrained language models (i.e., BERT, XLNet, RoBERTa, SciBERT, and ALBERT) to solve each of the three subtasks of the competition. Specifically, for each language model variant, we experiment by both freezing its weights and fine-tuning them. We also explore a multi-task architecture that was trained to jointly predict the outputs for the second and the third subtasks. Our best performing model evaluated on the DeftEval dataset obtains the 32nd place for the first subtask and the 37th place for the second subtask. The code is available for further research at: https://github.com/avramandrei/DeftEval. | ['Costin-Gabriel Chiru', 'Dumitru-Clementin Cercel', 'Andrei-Marius Avram'] | 2020-09-11 | null | https://aclanthology.org/2020.semeval-1.97 | https://aclanthology.org/2020.semeval-1.97.pdf | semeval-2020 | ['definition-extraction'] | ['natural-language-processing'] | [ 6.22411482e-02 3.93910199e-01 -1.62091270e-01 -4.58797514e-01
-8.29753995e-01 -7.91532815e-01 1.11023176e+00 1.92224830e-01
-5.89602113e-01 1.04596996e+00 8.44672397e-02 -7.68382847e-01
-1.01546757e-01 -6.83579028e-01 -6.69365704e-01 -2.73228496e-01
1.80369034e-01 6.53976738e-01 1.42744839e-01 -2.85035133e-01
9.66548547e-02 1.64014369e-01 -1.18667257e+00 5.68757713e-01
6.91701353e-01 9.43781674e-01 1.61908001e-01 9.21239078e-01
-2.91093558e-01 9.65067208e-01 -6.39070630e-01 -7.13692009e-01
-4.56661265e-03 -2.20984802e-01 -1.33709514e+00 -2.81945080e-01
3.66918266e-01 1.98620081e-01 -1.08948082e-01 9.14593697e-01
1.58522084e-01 2.37066358e-01 6.06009841e-01 -1.16816831e+00
-4.32751536e-01 1.18433535e+00 -2.34179005e-01 2.78058261e-01
4.51204687e-01 -2.48780362e-02 1.26553226e+00 -9.89439964e-01
6.21862292e-01 1.26687229e+00 4.36231971e-01 5.72218835e-01
-1.04896009e+00 -6.98926628e-01 3.83325666e-01 3.16026658e-01
-1.37795520e+00 -5.02133906e-01 4.84289974e-01 -7.14417398e-01
1.36644781e+00 3.01500976e-01 4.71889287e-01 1.25340629e+00
3.42059374e-01 8.41000438e-01 1.15638447e+00 -5.75434566e-01
1.03710525e-01 5.05506918e-02 8.91062021e-01 6.20395005e-01
1.22043170e-01 -1.21827275e-01 -2.87214756e-01 3.24012749e-02
4.43471670e-01 -4.79340285e-01 -1.96654201e-01 1.40475228e-01
-1.17305350e+00 7.94078946e-01 1.90026835e-02 5.15915513e-01
-2.95760244e-01 2.51137406e-01 4.42297906e-01 4.45534587e-01
5.43817639e-01 7.21432388e-01 -9.95904326e-01 -1.55584753e-01
-9.40252483e-01 5.14257193e-01 1.20900452e+00 1.19629455e+00
5.98097622e-01 -4.94076014e-01 -4.51656193e-01 7.14191616e-01
1.93273589e-01 4.04401403e-03 4.57307428e-01 -6.93005919e-01
6.74884737e-01 3.51016521e-01 1.66174084e-01 -3.91594499e-01
-6.97040915e-01 -5.80009937e-01 -4.96535927e-01 -3.86085063e-01
5.46029866e-01 -5.53298891e-01 -1.06377101e+00 1.66844380e+00
6.38622567e-02 2.62472510e-01 -1.23770542e-01 6.04776978e-01
1.42765236e+00 5.46673179e-01 4.49971646e-01 -7.49932006e-02
1.56161821e+00 -1.32852948e+00 -7.78924048e-01 -3.46106112e-01
8.05397749e-01 -8.90763998e-01 7.96876073e-01 3.33702296e-01
-9.20674860e-01 -5.16348839e-01 -1.16634989e+00 -3.86891693e-01
-7.83270895e-01 2.83330560e-01 5.65760672e-01 1.83839947e-01
-1.05721676e+00 3.19710076e-01 -8.57809246e-01 -5.69998771e-02
3.68916355e-02 1.64127231e-01 -8.61265510e-02 2.71484286e-01
-1.77649200e+00 1.17270660e+00 5.50608218e-01 -5.81715144e-02
-7.49966204e-01 -7.03341305e-01 -1.00048542e+00 2.10468527e-02
5.51276684e-01 -5.83347917e-01 1.43700516e+00 -6.96644247e-01
-1.19233978e+00 1.26695299e+00 -9.18458924e-02 -7.57185876e-01
3.24806392e-01 -5.16570091e-01 -2.67467529e-01 -4.71918762e-01
3.89462501e-01 2.70840824e-01 2.75170445e-01 -1.07087576e+00
-6.46718383e-01 -4.16969433e-02 4.20856714e-01 1.95099358e-02
2.61629254e-01 1.01950914e-01 -6.28602564e-01 -5.60996056e-01
-3.42116505e-01 -7.23799407e-01 -1.42313600e-01 -7.10618734e-01
-8.63816500e-01 -1.07896996e+00 1.21207856e-01 -7.48284101e-01
1.57291043e+00 -1.85943925e+00 3.26981932e-01 -3.44622731e-02
4.87191021e-01 4.38550264e-01 -3.21379155e-02 4.19102073e-01
-3.58402073e-01 3.38849366e-01 -4.48114723e-02 -6.14594996e-01
2.83903092e-01 9.69775915e-02 -1.16089910e-01 -9.90149900e-02
9.58174840e-02 1.03446984e+00 -1.03673911e+00 -3.59224617e-01
2.20797032e-01 3.26386511e-01 6.49021240e-03 1.65782005e-01
-6.40094995e-01 2.28834450e-01 -5.41379333e-01 2.61688501e-01
3.28583300e-01 -4.19822663e-01 3.08552474e-01 7.29588494e-02
-1.31034195e-01 1.15645516e+00 -1.09491408e+00 1.64645851e+00
-7.88424194e-01 6.82219386e-01 -5.16698807e-02 -8.67855251e-01
6.53605044e-01 4.97671396e-01 2.20985577e-01 -4.34004337e-01
1.10648781e-01 2.47479171e-01 3.98618206e-02 -5.11910856e-01
5.09481251e-01 5.18451035e-02 -3.51980746e-01 3.68443698e-01
4.57345605e-01 -3.28659475e-01 7.40604997e-01 3.13593954e-01
1.14127707e+00 3.55367184e-01 6.16930068e-01 -3.92199993e-01
9.18988585e-01 -2.61519969e-01 4.58615333e-01 8.29496026e-01
-6.24841191e-02 3.83061647e-01 7.27792859e-01 -6.40914440e-01
-5.37925601e-01 -1.16293585e+00 -2.90176123e-01 1.15360951e+00
-3.92573252e-02 -8.68271828e-01 -6.63084507e-01 -8.49911034e-01
-4.14111502e-02 1.37337625e+00 -6.26832783e-01 2.18611896e-01
-6.66911662e-01 -4.48253810e-01 6.83920860e-01 2.38303259e-01
5.98299921e-01 -1.21209371e+00 -4.95224595e-01 1.87067941e-01
-3.68411928e-01 -1.23226595e+00 -4.05436903e-01 5.70430815e-01
-9.24090669e-02 -1.25295711e+00 -2.72906810e-01 -7.63359845e-01
4.78375219e-02 -4.53183562e-01 1.69888330e+00 1.55153751e-01
1.48166925e-01 1.45391256e-01 -4.64444458e-01 -6.18919492e-01
-2.77254850e-01 5.39314449e-01 -1.99492708e-01 -4.00136411e-01
7.49112308e-01 -2.88901746e-01 -6.52000755e-02 -2.54585862e-01
-6.24920547e-01 3.17924529e-01 3.29328567e-01 6.61797881e-01
5.18812716e-01 -7.88926482e-02 5.46406209e-01 -1.37776315e+00
8.46785009e-01 -4.83414024e-01 -5.94692528e-01 4.83728915e-01
-6.17793858e-01 1.58754904e-02 6.12994432e-01 -1.21372089e-01
-9.08443511e-01 -7.09662437e-02 -4.36420530e-01 1.65052325e-01
-2.32300073e-01 7.23195612e-01 -1.48608744e-01 4.85445976e-01
5.47564149e-01 9.47936531e-03 -6.65309370e-01 -4.48825687e-01
5.54415047e-01 6.18363798e-01 3.13460261e-01 -7.58238018e-01
7.86693394e-01 -3.07024717e-01 -3.10612977e-01 -6.88137114e-01
-1.58729267e+00 -4.83984560e-01 -9.27250266e-01 -1.13157414e-01
1.10227358e+00 -8.95614982e-01 -2.80975819e-01 3.82376194e-01
-1.57270336e+00 -6.84812427e-01 -2.40896836e-01 3.09999973e-01
-2.76348054e-01 1.40240088e-01 -7.15588927e-01 -7.05969453e-01
-5.02821207e-01 -1.00283909e+00 9.09127235e-01 2.21610218e-01
-5.20265579e-01 -1.32569897e+00 1.49840675e-02 6.37708902e-01
2.67888814e-01 2.16951445e-01 1.09611785e+00 -1.02727151e+00
-2.77903318e-01 -6.17216080e-02 -1.73452035e-01 2.45815724e-01
1.37656368e-02 -4.83594984e-02 -7.90346503e-01 8.08427110e-02
-1.98853746e-01 -6.34684503e-01 1.17395794e+00 3.16120505e-01
1.12602890e+00 -2.47988552e-01 -3.78188074e-01 4.92315084e-01
1.12800193e+00 7.34364316e-02 4.60148156e-01 5.30923903e-01
6.03990316e-01 4.55673754e-01 5.22548914e-01 1.03394121e-01
8.31674278e-01 6.64123952e-01 2.32476648e-02 -1.33995146e-01
-2.49636203e-01 -1.52867824e-01 1.48110256e-01 8.15384984e-01
7.96049368e-03 -4.11637962e-01 -1.28922248e+00 6.68206990e-01
-1.94023812e+00 -6.69253051e-01 -2.90707499e-01 1.86638629e+00
1.26557469e+00 4.82439876e-01 -1.46722579e-02 -9.17256325e-02
4.64775801e-01 3.57084304e-01 -2.00192943e-01 -7.37341046e-01
5.25346817e-03 6.74205959e-01 1.21625423e-01 8.01119804e-01
-1.60089731e+00 1.31123471e+00 5.31471968e+00 8.56746852e-01
-7.96081245e-01 2.39617869e-01 5.20086050e-01 1.12287803e-02
-2.63832510e-01 5.69827505e-04 -1.03357005e+00 4.39480513e-01
1.08708167e+00 -2.45201454e-01 2.68262178e-01 4.73058283e-01
-7.44819120e-02 -4.50892523e-02 -1.25430155e+00 5.68185687e-01
2.21659010e-03 -1.20003617e+00 1.03413299e-01 -3.80459934e-01
5.55127680e-01 3.61315697e-01 -3.78649533e-01 8.55850399e-01
5.64575732e-01 -1.23933458e+00 7.11075306e-01 5.49388766e-01
6.40373707e-01 -4.06842142e-01 1.08934486e+00 3.81266534e-01
-1.14299202e+00 3.68996352e-01 2.40356172e-03 -3.87727678e-01
1.40370071e-01 8.15693736e-01 -5.58144033e-01 7.89577663e-01
2.58707643e-01 7.45596886e-01 -4.89109755e-01 6.63737655e-01
-1.01559687e+00 7.91129112e-01 -1.63042367e-01 -3.43585849e-01
2.22657859e-01 -7.06638396e-02 5.53317487e-01 1.66933167e+00
-2.41361603e-01 1.35223463e-01 2.77866215e-01 1.01816130e+00
-3.92535836e-01 -7.33177587e-02 -2.52811223e-01 3.35514802e-03
6.15501463e-01 1.32392108e+00 -3.89277518e-01 -4.49641198e-01
-4.23876256e-01 5.35920382e-01 6.98311150e-01 5.80061197e-01
-8.11953902e-01 -5.20632148e-01 3.37029546e-01 -3.45131829e-02
1.38987809e-01 -2.86717087e-01 -6.53606176e-01 -1.39816725e+00
-1.87960844e-02 -6.25016689e-01 4.82965946e-01 -5.38251817e-01
-1.22427893e+00 8.07888985e-01 9.90702733e-02 -4.17362154e-01
-4.32686031e-01 -7.25625098e-01 -5.21747589e-01 1.10754728e+00
-1.46634305e+00 -1.16477048e+00 -1.26053265e-03 2.58044153e-01
7.20573843e-01 -2.22562589e-02 1.07169819e+00 3.78724694e-01
-6.59995615e-01 6.79328263e-01 -3.33388567e-01 5.47667801e-01
5.42760491e-01 -1.69440567e+00 5.62857151e-01 8.15452516e-01
1.60847738e-01 6.82599843e-01 7.98779070e-01 -4.74869192e-01
-7.92831898e-01 -1.21381152e+00 1.71222055e+00 -6.88963354e-01
9.83322501e-01 -5.39531291e-01 -6.37830973e-01 1.08736980e+00
2.65576035e-01 -4.95605975e-01 6.31201565e-01 6.84509337e-01
-4.06489223e-01 1.96817994e-01 -6.70491517e-01 4.14722890e-01
7.48155951e-01 -4.86508757e-01 -9.28522944e-01 7.00235665e-01
8.63573074e-01 -7.49847233e-01 -9.43577051e-01 2.37076610e-01
4.61245835e-01 -4.50730652e-01 6.66956425e-01 -6.91742122e-01
8.58038127e-01 -9.20310840e-02 2.87157972e-03 -1.14639032e+00
-3.38227302e-01 -4.83917683e-01 -3.12052757e-01 1.31830478e+00
1.18965387e+00 -4.20759887e-01 2.84438193e-01 5.70554614e-01
-2.00944170e-02 -1.13816166e+00 -1.06375134e+00 -6.68496788e-01
6.83287084e-01 -6.72598481e-01 2.74898231e-01 9.18989956e-01
-8.75873640e-02 1.15759671e+00 -1.09610148e-01 -1.05889328e-01
4.02372658e-01 1.70391262e-01 6.79152548e-01 -8.91350687e-01
-3.32065105e-01 -6.09766841e-01 -9.05697495e-02 -1.21628261e+00
5.09833574e-01 -1.32102048e+00 -2.18228027e-02 -1.91169989e+00
2.04654813e-01 -1.81983113e-01 -4.01426196e-01 7.62671471e-01
-4.55123603e-01 -1.60031229e-01 1.86866149e-01 -8.04225057e-02
-9.53913510e-01 2.81052202e-01 9.57631528e-01 -2.66929060e-01
-1.19832776e-01 1.92659616e-01 -6.96341991e-01 5.63630283e-01
8.58991802e-01 -3.10759634e-01 -1.08905062e-01 -6.99731767e-01
2.50528008e-01 -9.80132148e-02 2.33998239e-01 -5.20661652e-01
1.42447487e-01 -9.73886177e-02 1.28306851e-01 -5.64037859e-01
2.69902915e-01 -1.82607606e-01 -2.87488848e-01 2.54137933e-01
-6.30957663e-01 1.30079687e-02 1.15429461e-01 2.69977804e-02
-1.53060168e-01 -2.87579536e-01 5.70101857e-01 -1.44749910e-01
-6.20380580e-01 7.42685199e-02 -4.09401000e-01 4.46401387e-01
7.52422571e-01 4.49566960e-01 -3.93563151e-01 -2.53700078e-01
-9.42824066e-01 8.09098840e-01 -1.52526736e-01 6.40959740e-01
3.32290381e-01 -9.35852587e-01 -1.08218992e+00 -1.51176423e-01
-7.55117228e-03 3.33858699e-01 -2.72377580e-01 7.00572968e-01
-2.97626793e-01 8.80777299e-01 3.80472451e-01 -2.59796619e-01
-1.05685222e+00 2.55813777e-01 6.73796773e-01 -1.13743079e+00
-3.71343136e-01 1.28678489e+00 1.09278522e-01 -8.31494451e-01
1.99479356e-01 -2.99866140e-01 -5.17065346e-01 -4.65914495e-02
4.03748274e-01 1.86416328e-01 8.83322507e-02 -4.69733983e-01
-5.60284913e-01 2.57686257e-01 -4.12203878e-01 -1.67203899e-02
1.33553302e+00 3.39020826e-02 -3.46583307e-01 7.71247923e-01
1.01771748e+00 -2.03531548e-01 -7.06524670e-01 -3.92747194e-01
5.67968309e-01 2.86154926e-01 6.13004044e-02 -1.40515101e+00
-6.13656342e-01 6.48358703e-01 -1.53373823e-01 3.44111145e-01
9.87417877e-01 1.62760094e-01 7.94703603e-01 3.17859590e-01
5.20163476e-01 -1.14011335e+00 -4.33338791e-01 1.24873972e+00
1.09119940e+00 -1.05225635e+00 3.64491567e-02 -4.61975336e-01
-2.99968302e-01 1.16518402e+00 5.67995608e-01 -1.33935675e-01
8.53009641e-01 3.39436859e-01 4.64850590e-02 -4.54847246e-01
-1.32098174e+00 -3.33693504e-01 6.22729063e-01 2.24605560e-01
1.11680233e+00 5.22334158e-01 -9.00126040e-01 1.12162232e+00
-5.84837437e-01 -7.79793859e-02 3.51965576e-01 5.02000570e-01
2.98735127e-02 -1.15290058e+00 1.85244292e-01 7.37550437e-01
-6.23452842e-01 -4.94293779e-01 -8.52590978e-01 1.03207481e+00
3.72713298e-01 1.31917918e+00 -3.15626651e-01 -4.95138437e-01
3.92430604e-01 1.95253983e-01 3.21801215e-01 -1.13559377e+00
-9.24233615e-01 -1.21199533e-01 7.10142016e-01 -4.28698689e-01
-2.53406525e-01 -4.96763051e-01 -1.39968586e+00 -1.55713186e-01
-2.29550198e-01 3.99615854e-01 4.16582137e-01 1.38642681e+00
-1.50075853e-01 9.13026810e-01 9.09089670e-02 -3.22994828e-01
-5.66376328e-01 -1.33912516e+00 -3.60257804e-01 4.92314398e-01
3.36095780e-01 -6.43735349e-01 -5.08190691e-01 -1.03972383e-01] | [9.8046875, 9.01154613494873] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.