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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
b0d7c495-c7d7-4e21-8692-81c4cd6872ef | a-recurrent-model-for-collective-entity | null | null | https://ojs.aaai.org/index.php/AAAI/article/view/5367 | https://jqin.gitee.io/files/AAAI2020-zhou.pdf | A Recurrent Model for Collective Entity Linking with Adaptive Features | The vast amount of web data enables us to build knowledge bases with unprecedented quality and coverage. Named Entity Disambiguation (NED) is an important task that automatically resolves ambiguous mentions in free text to correct target entries in the knowledge base. Traditional machine learning based methods for NED were outperformed and made obsolete by the state-of-the-art deep learning based models. However, deep learning models are more complex, requiring large amount of training data and lengthy training and parameter tuning time. In this paper, we revisit traditional machine learning techniques and propose a light-weight, tuneable and time-efficient method without using deep learning or deep learning generated features. We propose novel adaptive features that focus on extracting discriminative features to better model similarities between candidate entities and the mention’s context. We learn a local ranking model based on traditional and the new adaptive features based on the learning-to-rank framework. While arriving at linking decisions individually via the local model, our method also takes into consideration the correlation between decisions by running multiple recurrent global models, which can be deemed as a learned local search method. Our method attains performances comparable to the state-of-the-art deep learning-based methods on NED benchmark datasets while being significantly faster to train. | ['Jianbin Qin', 'Wei Wang', 'Yukai Miao', 'Xiaoling Zhou'] | 2020-04-03 | null | null | null | aaai-2020-4 | ['entity-disambiguation'] | ['natural-language-processing'] | [-3.51066709e-01 7.55498856e-02 -4.73299086e-01 -4.96175081e-01
-9.05394077e-01 -6.81378663e-01 6.89427733e-01 5.27745008e-01
-8.10622513e-01 1.03723216e+00 3.38921756e-01 -2.49697134e-01
-4.37558949e-01 -9.83354449e-01 -5.78350723e-01 -3.78501415e-01
-1.45161599e-01 1.01510227e+00 4.07688618e-01 -4.85116512e-01
2.95621268e-02 5.05594432e-01 -1.40378058e+00 2.22338617e-01
1.01784921e+00 8.66435409e-01 7.10789040e-02 2.00752303e-01
-7.47669876e-01 8.57215405e-01 -6.87500834e-01 -7.27004528e-01
-2.16700155e-02 6.05962873e-02 -1.10837495e+00 -7.55857289e-01
6.91218436e-01 1.11565672e-01 -4.31538850e-01 9.88750279e-01
7.23883033e-01 3.92346263e-01 4.74733233e-01 -7.11910248e-01
-9.02260065e-01 9.05042887e-01 -6.56133816e-02 5.21151066e-01
2.49635100e-01 -4.90395844e-01 1.43572760e+00 -1.01271045e+00
9.74006832e-01 1.00850856e+00 8.04264069e-01 4.73968923e-01
-9.99394536e-01 -5.56978881e-01 4.16143537e-01 6.06253922e-01
-1.29914618e+00 -4.28472668e-01 4.06197459e-01 -3.03085715e-01
1.35307455e+00 1.67217150e-01 2.21942931e-01 1.03833389e+00
-6.56794235e-02 7.41047502e-01 5.89595735e-01 -4.18750376e-01
1.27817646e-01 -4.60174978e-02 3.25559884e-01 7.75771618e-01
5.53247571e-01 -8.03762022e-03 -4.06075388e-01 -3.62393379e-01
5.01203120e-01 1.83059257e-02 -1.28948018e-01 -2.62017041e-01
-1.12744331e+00 8.39912176e-01 7.80811608e-01 7.52593756e-01
-6.05869651e-01 -6.63276389e-03 3.78264129e-01 2.63002545e-01
3.05974633e-01 7.79694617e-01 -1.17217815e+00 1.12412758e-01
-9.86099780e-01 1.83101878e-01 1.03185654e+00 6.91764057e-01
8.65369976e-01 -2.16228068e-01 -4.79654640e-01 9.34987783e-01
1.38151616e-01 1.79323807e-01 7.02122808e-01 -3.73337686e-01
4.53914762e-01 8.90110016e-01 2.10031301e-01 -1.05512393e+00
-8.34296048e-01 -8.15048695e-01 -6.75625861e-01 -2.91993737e-01
2.16790944e-01 -2.02888548e-01 -1.01232827e+00 1.72033238e+00
4.16331112e-01 2.58754432e-01 1.55921012e-01 7.49035060e-01
1.11300528e+00 4.93592590e-01 3.60166103e-01 -9.06547233e-02
1.28164589e+00 -9.77219760e-01 -7.50218511e-01 -1.47603884e-01
8.00460935e-01 -6.23476803e-01 6.11391008e-01 8.55432153e-02
-6.86878502e-01 -2.28338420e-01 -9.11582947e-01 -3.54258984e-01
-1.02968299e+00 4.23471302e-01 8.12841713e-01 3.91700417e-01
-9.66408134e-01 7.10112989e-01 -6.90490782e-01 -3.79818857e-01
2.16983691e-01 6.53129041e-01 -4.45889622e-01 1.71800509e-01
-1.74720764e+00 1.29591429e+00 7.12556958e-01 1.22652501e-01
-5.30011117e-01 -5.97937942e-01 -8.02823901e-01 2.22320139e-01
5.79719722e-01 -7.83230484e-01 1.22230458e+00 -6.86127067e-01
-1.33820581e+00 8.22549045e-01 -2.61684835e-01 -6.97853506e-01
2.90632695e-01 -4.65944350e-01 -7.33287990e-01 -2.35754073e-01
3.59701484e-01 2.46287420e-01 3.55106056e-01 -1.07259357e+00
-8.79163682e-01 -2.77286589e-01 1.43041566e-01 4.53405268e-03
-2.27659494e-01 -9.66675952e-02 -4.58681166e-01 -6.18311644e-01
1.19434714e-01 -7.57514238e-01 -3.10434490e-01 -5.65125227e-01
-5.21264732e-01 -6.93129182e-01 3.95821959e-01 -6.23162508e-01
1.50264823e+00 -1.68327904e+00 -1.03981839e-02 2.78772861e-01
4.16223884e-01 9.31160390e-01 -1.03794068e-01 5.74250460e-01
-6.00489462e-03 8.41011778e-02 1.50098309e-01 -4.64345887e-02
1.72464833e-01 1.09681580e-02 -2.15434104e-01 8.14546496e-02
1.06207676e-01 1.03254461e+00 -1.20271766e+00 -4.50318962e-01
-3.10218912e-02 4.89486217e-01 -2.32901379e-01 1.69970375e-02
-1.79857254e-01 4.42414805e-02 -6.80892646e-01 6.52150929e-01
4.15800333e-01 -4.58975077e-01 5.33799112e-01 -4.16971385e-01
-8.79571810e-02 8.36003482e-01 -1.23628676e+00 1.52053070e+00
-6.05417788e-01 3.59178960e-01 -3.32242548e-01 -1.13083184e+00
1.09957123e+00 3.10051709e-01 2.74400651e-01 -9.21733320e-01
-1.23046912e-01 6.80497527e-01 -2.78662473e-01 -4.18147147e-01
6.65242851e-01 2.02285737e-01 -1.91576764e-01 6.90246001e-02
5.83697379e-01 8.23447227e-01 3.16971242e-01 1.55281246e-01
1.27187121e+00 -1.86140742e-02 5.81848502e-01 -1.50296509e-01
5.05592763e-01 -7.38762096e-02 8.68739724e-01 8.34522784e-01
4.15630192e-02 -3.06975339e-02 5.97443134e-02 -8.53157282e-01
-6.80573523e-01 -8.47912669e-01 -1.51484057e-01 1.54376578e+00
1.99025944e-01 -5.14317155e-01 -4.12457049e-01 -1.06218541e+00
1.84541062e-01 6.03810251e-01 -6.36140883e-01 -1.13822192e-01
-6.92314148e-01 -8.65901470e-01 6.87163115e-01 5.28963447e-01
6.73820198e-01 -1.01286674e+00 -8.75339061e-02 6.19501710e-01
-2.01316535e-01 -9.39233541e-01 -2.34836578e-01 5.13642192e-01
-7.36479163e-01 -1.07796979e+00 -6.40089154e-01 -1.01443994e+00
3.28983039e-01 -8.64743367e-02 1.38178325e+00 -1.09711103e-02
-1.78345725e-01 8.57348144e-02 -3.78944784e-01 -1.20391533e-01
2.68213153e-02 7.29878366e-01 2.63782889e-01 -1.64904624e-01
8.73639941e-01 -4.63610351e-01 -6.20793581e-01 1.37300998e-01
-6.69278741e-01 -4.91878271e-01 9.98601913e-01 1.10832679e+00
6.36469305e-01 -1.98381588e-01 9.02281582e-01 -1.18161356e+00
5.72225392e-01 -6.44039631e-01 -4.32622224e-01 6.34814322e-01
-8.65899146e-01 7.17751682e-01 4.67797756e-01 -2.39445925e-01
-1.17453277e+00 2.77494621e-02 -1.86878577e-01 -2.63818968e-02
1.12957526e-02 9.48367596e-01 -1.10872038e-01 -6.72966093e-02
8.37350190e-01 -7.40473941e-02 -9.87164319e-01 -8.77133489e-01
6.14404619e-01 4.85353082e-01 5.55263698e-01 -4.87878114e-01
6.20009124e-01 1.53689966e-01 -2.86080688e-01 -3.45796406e-01
-1.11007345e+00 -6.56233251e-01 -8.24589372e-01 1.88755497e-01
5.58589697e-01 -8.88675094e-01 -5.28591812e-01 -7.68400077e-03
-1.22229671e+00 1.40880838e-01 -1.54740708e-02 4.59305167e-01
4.72966060e-02 8.99265781e-02 -6.06337368e-01 -4.30194318e-01
-5.50514102e-01 -6.87475681e-01 9.14619803e-01 3.83664489e-01
-1.67171955e-01 -1.12662840e+00 5.08696854e-01 1.32431567e-01
6.33704424e-01 1.10346742e-01 9.31300879e-01 -1.61432612e+00
-5.35900116e-01 -3.37974399e-01 -3.01187813e-01 -1.86952591e-01
1.97270349e-01 -2.97728658e-01 -8.63040745e-01 -5.30192666e-02
-5.71844697e-01 -4.42268662e-02 1.32682931e+00 1.79509759e-01
7.59619474e-01 -5.66617072e-01 -7.44028330e-01 4.94251788e-01
1.51746345e+00 1.35706052e-01 2.80160159e-01 9.59961772e-01
6.03786230e-01 3.07454824e-01 4.86982167e-01 2.30354369e-01
6.92334235e-01 8.01982820e-01 1.47497177e-01 -7.26771727e-02
-3.99197079e-02 -4.03040230e-01 -4.86365594e-02 7.86857784e-01
-2.57744640e-01 -1.73430890e-01 -1.11451077e+00 7.58455634e-01
-2.09880161e+00 -1.07273686e+00 1.32581577e-01 2.16907048e+00
1.15663886e+00 1.12211503e-01 -1.67775556e-01 -2.15221092e-01
9.24238503e-01 1.21995158e-01 -5.68186104e-01 -2.37733170e-01
-2.62212098e-01 4.39753443e-01 6.66799009e-01 4.16764826e-01
-1.51608503e+00 1.33481252e+00 5.76279306e+00 8.57624769e-01
-1.12720382e+00 7.53562897e-02 3.64548683e-01 1.45721175e-02
-3.57039511e-01 -1.36315793e-01 -1.25251698e+00 2.33258888e-01
9.66818869e-01 -2.97147572e-01 1.50723234e-01 1.05786896e+00
-3.12800139e-01 2.67612636e-01 -1.12063253e+00 8.56526494e-01
-9.73150060e-02 -1.66346681e+00 2.45255142e-01 -1.48102418e-01
8.44140172e-01 4.51311290e-01 -1.82708055e-01 6.43800676e-01
7.02305675e-01 -9.21880007e-01 2.79875338e-01 7.62272120e-01
4.46854204e-01 -6.43116772e-01 1.11644030e+00 2.66439319e-02
-1.38378525e+00 -2.34659776e-01 -3.87130439e-01 2.62335628e-01
-8.74514282e-02 9.49391186e-01 -8.76433372e-01 8.43455553e-01
7.37037480e-01 6.12875462e-01 -5.74291229e-01 1.23491693e+00
-3.01044673e-01 3.94940585e-01 -2.33079612e-01 -2.85062194e-01
3.64266008e-01 2.20498681e-01 4.82441425e-01 1.52889621e+00
2.00320944e-01 2.38575526e-02 2.20491663e-01 5.87339163e-01
-5.45588374e-01 2.83308953e-01 -3.77671242e-01 -1.51672810e-02
8.27957094e-01 1.30318987e+00 -5.38239717e-01 -4.94961739e-01
-3.06082278e-01 8.69644046e-01 8.17179859e-01 3.77163738e-01
-5.85433424e-01 -7.78454959e-01 5.81550598e-01 -3.75381172e-01
6.61906064e-01 -1.50781631e-01 -3.33266966e-02 -1.42567503e+00
6.59030862e-03 -5.68690062e-01 7.32761145e-01 -3.81716713e-02
-1.66914082e+00 9.61812377e-01 -3.63818884e-01 -9.37822640e-01
-4.22037691e-01 -6.31287336e-01 -3.69062394e-01 6.12076581e-01
-1.90620065e+00 -1.03613567e+00 -9.63239223e-02 5.57605326e-01
2.33308524e-01 -3.63137662e-01 1.02997184e+00 6.29313350e-01
-5.75311065e-01 8.54692638e-01 5.34405291e-01 6.04812801e-01
1.00621343e+00 -1.40984547e+00 4.09884959e-01 5.60363114e-01
4.71481025e-01 9.17735279e-01 4.67470348e-01 -5.94509602e-01
-1.05900514e+00 -1.13605070e+00 1.66767120e+00 -3.83451819e-01
7.39990652e-01 -2.99422853e-02 -8.72436762e-01 7.34088957e-01
-2.91802380e-02 8.40280429e-02 8.44156981e-01 7.65560389e-01
-5.51185489e-01 -2.53343284e-01 -9.27169621e-01 2.97394931e-01
1.08261621e+00 -6.07027352e-01 -1.17881727e+00 2.85290718e-01
7.04032242e-01 -3.91646951e-01 -8.96543026e-01 5.45254469e-01
4.80341494e-01 -5.68551540e-01 1.02312160e+00 -9.14789081e-01
-3.22630294e-02 -2.80895740e-01 -4.38324064e-02 -1.25431049e+00
-6.00648046e-01 -5.00241637e-01 -3.32076430e-01 1.37186754e+00
9.23579454e-01 -6.16141737e-01 6.08232558e-01 5.97045362e-01
-5.17842397e-02 -8.07818353e-01 -1.06093454e+00 -7.60018528e-01
-7.22443089e-02 7.72865266e-02 7.84117460e-01 1.47842181e+00
1.75491631e-01 6.11657262e-01 -1.02565005e-01 2.98537314e-01
2.29159251e-01 3.43519539e-01 2.87470430e-01 -1.72060323e+00
8.82361904e-02 -4.94347483e-01 -5.39093316e-01 -8.89835000e-01
3.00967395e-01 -1.09666681e+00 -1.42643854e-01 -1.77609110e+00
9.40595008e-03 -6.41405284e-01 -8.92860055e-01 8.96543741e-01
-4.08170909e-01 -2.06592083e-01 -7.67678022e-02 2.80772239e-01
-1.04153335e+00 5.20694137e-01 4.10610706e-01 -3.36739868e-01
-3.35493237e-01 -2.35600062e-02 -6.26472950e-01 6.22245789e-01
5.37149310e-01 -6.43836141e-01 1.09811183e-02 -6.10623002e-01
6.54542506e-01 -3.04435700e-01 -3.07491049e-03 -8.14361393e-01
7.20155478e-01 1.40423313e-01 4.98814553e-01 -5.08348882e-01
5.96416481e-02 -4.95354086e-01 -2.22421914e-01 2.31548265e-01
-5.27722418e-01 -4.57825661e-02 1.21526383e-01 6.96438491e-01
-4.75606173e-01 -1.52293384e-01 2.58708417e-01 -2.59621769e-01
-1.21591485e+00 4.45227891e-01 -1.09165162e-01 2.00967833e-01
6.30577803e-01 2.97145456e-01 -4.13472861e-01 2.71596983e-02
-1.04723620e+00 3.15679014e-01 -5.80087816e-03 5.86409628e-01
3.84488314e-01 -1.48512864e+00 -6.30815744e-01 -2.73458749e-01
1.46997213e-01 -2.09114775e-01 9.89014730e-02 4.61588293e-01
-3.34121406e-01 9.20027614e-01 5.95141873e-02 -1.64338201e-01
-9.12976325e-01 4.77716267e-01 4.97894764e-01 -8.16519678e-01
-4.08959746e-01 1.06208897e+00 -1.94794357e-01 -8.58870506e-01
3.20266455e-01 -1.94246307e-01 -6.81402862e-01 3.99256438e-01
5.42926371e-01 3.75395954e-01 6.04876935e-01 -4.74265784e-01
-7.36438096e-01 3.91684562e-01 -4.15940344e-01 2.45139450e-01
1.47993505e+00 -1.33215589e-02 -9.66661647e-02 2.36783639e-01
1.05108595e+00 4.33285348e-02 -5.00954807e-01 -8.10987055e-01
8.66218328e-01 -6.71228766e-02 1.93055809e-01 -1.12584448e+00
-1.04161060e+00 4.86723483e-01 5.47716975e-01 1.24367975e-01
9.09345567e-01 8.63754004e-02 9.76453424e-01 1.11904943e+00
4.28958327e-01 -1.28120959e+00 -3.21451932e-01 8.54355395e-01
5.16502500e-01 -1.38439488e+00 -1.13800399e-01 -3.11157666e-02
-3.93053770e-01 1.27639663e+00 5.53597867e-01 -8.85235369e-02
7.30437040e-01 -8.85207579e-02 1.55656919e-01 -3.08605254e-01
-7.97097862e-01 -4.99151200e-01 6.56666160e-01 3.89311373e-01
6.12015903e-01 -4.57066372e-02 -5.35149455e-01 9.39265251e-01
-1.02749616e-01 -2.52612103e-02 -2.61711460e-02 7.21210659e-01
-6.66913688e-01 -1.22727966e+00 9.38827619e-02 5.44225693e-01
-6.12817883e-01 -4.99860585e-01 -5.22194922e-01 6.77044690e-01
2.45858327e-01 7.94980288e-01 -1.80054292e-01 -4.29573745e-01
3.61014128e-01 2.77220011e-01 -1.02105431e-01 -5.84077477e-01
-6.61896765e-01 -5.09919167e-01 4.35995638e-01 -6.84447944e-01
-5.93628764e-01 -3.83862585e-01 -1.28287041e+00 -1.02697693e-01
-5.26762486e-01 5.02776980e-01 6.03198707e-01 1.30374193e+00
8.31479788e-01 4.53661531e-01 4.42258418e-01 -5.57874084e-01
-5.04227579e-01 -8.57006907e-01 -3.25630873e-01 1.36260703e-01
4.38551694e-01 -9.25586522e-01 -1.12210512e-02 -3.21150333e-01] | [9.492794036865234, 8.89466381072998] |
c9a6600a-f55b-4098-95e0-62286376fc6c | unsupervised-deep-persistent-monocular-visual | 2011.00341 | null | https://arxiv.org/abs/2011.00341v1 | https://arxiv.org/pdf/2011.00341v1.pdf | Unsupervised Deep Persistent Monocular Visual Odometry and Depth Estimation in Extreme Environments | In recent years, unsupervised deep learning approaches have received significant attention to estimate the depth and visual odometry (VO) from unlabelled monocular image sequences. However, their performance is limited in challenging environments due to perceptual degradation, occlusions and rapid motions. Moreover, the existing unsupervised methods suffer from the lack of scale-consistency constraints across frames, which causes that the VO estimators fail to provide persistent trajectories over long sequences. In this study, we propose an unsupervised monocular deep VO framework that predicts six-degrees-of-freedom pose camera motion and depth map of the scene from unlabelled RGB image sequences. We provide detailed quantitative and qualitative evaluations of the proposed framework on a) a challenging dataset collected during the DARPA Subterranean challenge; and b) the benchmark KITTI and Cityscapes datasets. The proposed approach outperforms both traditional and state-of-the-art unsupervised deep VO methods providing better results for both pose estimation and depth recovery. The presented approach is part of the solution used by the COSTAR team participating at the DARPA Subterranean Challenge. | ['Ali-akbar Agha-mohammadi', 'Benjamin Morrell', 'Angel Santamaria-Navarro', 'Yasin Almalioglu'] | 2020-10-31 | null | null | null | null | ['monocular-visual-odometry'] | ['robots'] | [-2.96953926e-03 -2.86336988e-01 -3.59367132e-02 -3.38346481e-01
-6.72406256e-01 -6.07433915e-01 6.17827237e-01 -2.94450790e-01
-6.16184890e-01 8.55172396e-01 8.44864920e-02 2.99590826e-01
5.31309610e-03 -3.58594537e-01 -7.40005910e-01 -8.70967209e-01
7.46501908e-02 6.63210511e-01 3.83734167e-01 7.57235587e-02
2.67399937e-01 7.00354159e-01 -1.75525343e+00 -1.07022531e-01
4.40978378e-01 9.26336288e-01 4.02209759e-01 8.43926489e-01
2.69228965e-01 1.03816736e+00 -1.52180091e-01 1.80352971e-01
4.40756142e-01 -2.40250185e-01 -6.70190513e-01 4.90896225e-01
8.10698152e-01 -6.61774099e-01 -6.80933237e-01 9.81701553e-01
5.58987379e-01 2.33953327e-01 5.34810185e-01 -1.32034290e+00
1.58910081e-01 -3.39801222e-01 -5.28039336e-01 1.68857798e-01
5.30647218e-01 2.84939557e-01 7.79495955e-01 -9.15507138e-01
1.18360364e+00 1.10742748e+00 6.52855039e-01 2.03763813e-01
-1.03985941e+00 -1.94616169e-01 -4.34063235e-03 4.39833611e-01
-1.37125278e+00 -6.15672350e-01 7.58611143e-01 -7.06503391e-01
1.11380756e+00 -1.33897230e-01 6.18654847e-01 1.05149257e+00
7.83397704e-02 1.03997445e+00 1.07365453e+00 -1.96268335e-01
3.95548791e-01 -1.06830627e-01 -3.50542873e-01 6.07554972e-01
3.17117810e-01 3.87359083e-01 -8.18676174e-01 9.95064899e-02
7.75616348e-01 7.98912495e-02 -2.12952182e-01 -1.24088705e+00
-1.13178241e+00 6.93541348e-01 4.55378830e-01 8.46691057e-02
-4.71233636e-01 3.48302484e-01 2.39341646e-01 7.27989748e-02
3.73002261e-01 4.18865345e-02 -4.38571453e-01 -4.48694408e-01
-1.02481627e+00 4.65819359e-01 4.93863523e-01 1.05473852e+00
9.20683563e-01 2.73121864e-01 4.21328932e-01 4.59354639e-01
4.50512707e-01 5.84515631e-01 3.94835860e-01 -1.25069797e+00
5.39448023e-01 2.81236619e-01 4.47775096e-01 -1.17549574e+00
-6.60831451e-01 -2.73124248e-01 -4.83399808e-01 3.43000293e-01
5.23308873e-01 -1.68759245e-02 -1.00418329e+00 1.43871462e+00
4.43219185e-01 -6.18725009e-02 1.48358047e-01 1.28945410e+00
7.42929399e-01 3.51315677e-01 -4.38478470e-01 -9.68710855e-02
6.08458519e-01 -9.65730488e-01 -7.39153206e-01 -6.12530828e-01
4.72636789e-01 -7.20998526e-01 4.74665612e-01 6.81029737e-01
-8.91835809e-01 -6.41956151e-01 -1.12721312e+00 -2.30079114e-01
-1.25450060e-01 1.67397097e-01 5.67571223e-01 5.59387505e-01
-1.10169137e+00 2.25046858e-01 -1.07659781e+00 -7.60530770e-01
8.63274261e-02 2.39841893e-01 -4.84936088e-01 -4.75704521e-01
-7.46140599e-01 8.00546050e-01 2.60390013e-01 2.83023149e-01
-1.41669190e+00 -1.68577552e-01 -1.14279842e+00 -4.92319673e-01
2.52065778e-01 -4.66229200e-01 1.06520307e+00 -7.66263902e-01
-1.43959308e+00 9.28843439e-01 -1.93195581e-01 -5.79220057e-01
9.92614508e-01 -6.26873851e-01 7.93782845e-02 2.72230178e-01
2.40693782e-02 9.82817054e-01 6.20262027e-01 -1.31635165e+00
-7.33060777e-01 -7.40917981e-01 5.46653010e-02 6.10708594e-01
2.28417754e-01 -6.22976482e-01 -5.37148952e-01 -2.06200078e-01
6.48728788e-01 -1.24593377e+00 -3.58989149e-01 1.04300112e-01
-2.28498936e-01 4.13208753e-01 8.56481671e-01 -5.03927350e-01
5.58473885e-01 -1.89587677e+00 5.28383613e-01 -3.95801783e-01
-2.35935040e-02 8.23031813e-02 2.60578483e-01 3.98879319e-01
2.41713852e-01 -8.02571774e-01 -2.51832068e-01 -6.77050889e-01
-1.37934491e-01 5.66808999e-01 -3.57559440e-03 1.12638319e+00
-3.40163946e-01 5.74921787e-01 -9.97147799e-01 -1.98627487e-01
7.19553053e-01 4.67052788e-01 -3.46446484e-01 2.35569596e-01
-2.78917074e-01 9.80208874e-01 -1.64688855e-01 8.22198927e-01
8.26854706e-01 2.22905055e-01 7.75144920e-02 -5.43824211e-02
-4.25604403e-01 1.88553974e-01 -1.43549526e+00 2.24087429e+00
-8.31650347e-02 9.58406210e-01 9.27111730e-02 -9.57633078e-01
8.65103960e-01 2.28118375e-02 7.69282222e-01 -8.01573992e-01
5.49088046e-02 4.45524961e-01 -4.06545520e-01 -6.16239786e-01
8.05312157e-01 -1.65674433e-01 4.29229923e-02 -1.67050380e-02
4.82375324e-02 -5.34047484e-01 1.71391249e-01 -8.34629592e-03
9.92531598e-01 6.63446307e-01 3.29297334e-01 -4.56113480e-02
6.63130760e-01 1.95523918e-01 6.54301286e-01 4.28612083e-01
-6.54000640e-01 9.95272517e-01 6.08442873e-02 -7.63064742e-01
-1.23553288e+00 -9.53883767e-01 -4.48081568e-02 2.97649145e-01
3.88091624e-01 6.64295331e-02 -4.13812429e-01 -2.82047331e-01
-3.36827263e-02 1.85097456e-01 -4.14336115e-01 2.70887583e-01
-4.72959191e-01 -5.19636273e-01 4.73602444e-01 4.77652133e-01
8.31051052e-01 -9.40887749e-01 -1.12136328e+00 2.62436450e-01
-5.38925886e-01 -1.72462249e+00 4.70465375e-03 1.42463535e-01
-1.03224492e+00 -1.04234469e+00 -8.81688297e-01 -6.22923613e-01
2.25130737e-01 4.66920197e-01 9.06701922e-01 -5.89587331e-01
-3.42190087e-01 5.29965937e-01 -1.99244633e-01 -8.31547305e-02
-1.76620223e-02 2.08307467e-02 2.06266835e-01 2.09130682e-02
3.43636572e-01 -4.30019349e-01 -7.78214276e-01 4.26047951e-01
-8.22857976e-01 -9.12460834e-02 2.57936805e-01 5.26967943e-01
7.90298223e-01 -1.79777920e-01 -5.88761736e-03 -5.39068162e-01
-3.04399103e-01 -2.43202925e-01 -1.10623896e+00 -4.41020250e-01
-3.90053242e-01 -4.59459238e-02 1.39471978e-01 2.20961738e-02
-1.01648796e+00 7.71045148e-01 9.75399371e-03 -5.25496721e-01
-3.78588051e-01 1.81095183e-01 -2.06817567e-01 -2.00320140e-01
6.05817735e-01 3.40363741e-01 -1.92881614e-01 -3.76168787e-01
2.23630622e-01 5.36910594e-01 7.41090655e-01 -1.69830605e-01
7.36505508e-01 1.17844713e+00 2.15174571e-01 -1.16060221e+00
-6.10105753e-01 -1.00777483e+00 -1.04987180e+00 -3.58450472e-01
1.11029434e+00 -1.44462717e+00 -2.85348356e-01 8.89059544e-01
-1.19365895e+00 -4.56350118e-01 8.65412597e-03 6.73692822e-01
-9.87958670e-01 6.58421636e-01 -4.43596691e-01 -1.00198364e+00
3.65313180e-02 -1.23070705e+00 1.27208865e+00 1.72603771e-01
-9.15148016e-03 -8.82165551e-01 5.36728799e-01 5.21321356e-01
8.46829191e-02 7.09285975e-01 1.00278758e-01 2.17342988e-01
-1.06462789e+00 -1.73471883e-01 -3.15445028e-02 2.73836583e-01
-1.20550148e-01 -2.96041042e-01 -1.21702230e+00 -4.91309732e-01
5.64265624e-02 -5.05450189e-01 7.88873672e-01 5.37220180e-01
4.11967456e-01 1.81010768e-01 -3.66778336e-02 1.00157106e+00
1.78114474e+00 2.09498003e-01 8.24224770e-01 8.33136141e-01
1.05602419e+00 6.32931530e-01 9.27203059e-01 5.84342539e-01
6.63928330e-01 9.47847247e-01 9.45772290e-01 1.38321072e-01
-1.54930009e-02 -1.16263136e-01 4.63167876e-01 5.48009574e-01
-9.00208801e-02 -2.39544688e-03 -1.04969168e+00 1.01798296e+00
-1.98215973e+00 -6.73686564e-01 -4.29026961e-01 2.27366447e+00
1.41462758e-01 4.09209803e-02 1.00099415e-01 3.08940142e-01
2.36083567e-01 4.66496646e-01 -6.69381618e-01 -5.01897931e-02
-2.90640086e-01 -3.82039279e-01 8.43028247e-01 7.25494504e-01
-1.11242008e+00 1.01486969e+00 5.76859045e+00 7.99569860e-02
-1.09843516e+00 2.56963193e-01 1.55061871e-01 -1.71085462e-01
1.38441846e-01 6.75631762e-02 -8.07716250e-01 6.81085363e-02
6.83546066e-01 3.73688638e-01 2.67007530e-01 8.99407148e-01
2.25209743e-01 -6.66898966e-01 -1.01217949e+00 1.29469645e+00
1.67840496e-01 -1.05878687e+00 -5.34661651e-01 3.19278359e-01
1.21129954e+00 7.03949273e-01 -2.31706761e-02 -4.09161188e-02
9.03986916e-02 -7.73305655e-01 8.29829752e-01 4.76852566e-01
5.59752643e-01 -8.26578200e-01 8.02046001e-01 5.64006209e-01
-1.17856896e+00 -1.46605745e-01 -6.78988218e-01 -3.02117079e-01
3.51760387e-01 5.14060378e-01 -6.87824965e-01 5.95545471e-01
7.89201617e-01 1.09171057e+00 -5.61845601e-01 1.15234840e+00
-3.63281041e-01 -1.55866919e-02 -3.05792749e-01 3.57287586e-01
6.21007979e-01 -1.75782114e-01 6.61911964e-01 8.76005113e-01
2.45077118e-01 -5.93467690e-02 2.00991586e-01 4.53256339e-01
2.73091733e-01 -3.16723734e-01 -8.92597556e-01 1.94112346e-01
6.55395538e-02 9.82685030e-01 -5.37726104e-01 -2.17964455e-01
-3.54713589e-01 1.12555015e+00 1.50470242e-01 4.89428878e-01
-6.06894970e-01 1.96500108e-01 7.44980097e-01 8.29952881e-02
5.36711335e-01 -7.62772918e-01 -1.99921951e-01 -1.28466940e+00
-1.54987760e-02 -6.59255922e-01 1.81482464e-01 -1.05123353e+00
-6.11725152e-01 6.53626740e-01 8.33044872e-02 -1.56282032e+00
-6.34356081e-01 -5.75407863e-01 7.52080753e-02 4.90423739e-01
-1.67248309e+00 -7.91845620e-01 -8.11441839e-01 6.04940474e-01
7.07572818e-01 -8.11166838e-02 4.22921479e-01 4.87746209e-01
-1.33779094e-01 -6.27120137e-02 3.32257330e-01 -1.75544649e-01
5.78428984e-01 -1.13690543e+00 4.89216775e-01 1.06486320e+00
2.05590904e-01 3.03975325e-02 1.02130604e+00 -4.61300135e-01
-1.53590906e+00 -9.77293551e-01 8.16919386e-01 -5.78087211e-01
3.13750863e-01 -2.38383770e-01 -5.91419518e-01 7.09129095e-01
-8.59721303e-02 2.94918746e-01 -1.84843298e-02 -5.88733673e-01
4.12047058e-02 -1.81465954e-01 -1.16729712e+00 2.18520284e-01
1.03254557e+00 -5.45654595e-01 -2.47177541e-01 2.84637332e-01
3.57751429e-01 -7.78755903e-01 -4.45324063e-01 3.72556239e-01
6.49647355e-01 -1.48889077e+00 1.04053795e+00 4.02932391e-02
3.39101434e-01 -5.23318768e-01 -6.31433010e-01 -1.04507840e+00
1.82882473e-01 -4.28154349e-01 -2.67927498e-01 7.47747123e-01
-2.13250682e-01 -1.95058972e-01 1.21844578e+00 2.16779947e-01
1.55631742e-02 -2.45820269e-01 -1.12404251e+00 -7.34469056e-01
-2.54722476e-01 -6.53828263e-01 5.81968203e-02 5.87303102e-01
-6.51508331e-01 1.86902016e-01 -6.97046757e-01 3.49724144e-01
1.15105486e+00 -1.33221760e-01 1.34161532e+00 -1.03347874e+00
-3.40237141e-01 1.56831607e-01 -1.04866958e+00 -1.38476455e+00
-1.20275348e-01 -2.81782836e-01 4.33099866e-01 -1.62000597e+00
-9.57432240e-02 -8.18946287e-02 1.12733066e-01 -1.95550665e-01
2.97156066e-01 4.61293459e-01 1.87748969e-01 4.31692094e-01
-6.64328635e-01 6.92837536e-01 9.23738182e-01 5.39340712e-02
-8.09263438e-02 -2.76170135e-01 2.40388528e-01 8.78677845e-01
4.60032731e-01 -4.73463565e-01 -6.07027769e-01 -7.92156458e-01
2.58683473e-01 3.87710929e-01 5.20916998e-01 -1.43456113e+00
2.86786795e-01 -1.60690937e-02 5.13272226e-01 -1.23952842e+00
6.57385826e-01 -8.45433652e-01 3.36719036e-01 5.97338974e-01
1.21147119e-01 2.38878235e-01 7.98723772e-02 6.84844255e-01
-3.15765440e-01 -6.43735975e-02 9.17875111e-01 -3.39821070e-01
-1.28336537e+00 3.09811980e-01 -4.23033893e-01 1.36162229e-02
8.28758597e-01 -4.41986620e-01 3.00824661e-02 -6.32198274e-01
-6.08282506e-01 -2.77121812e-02 8.14532161e-01 4.21190441e-01
7.27817595e-01 -1.25604773e+00 -4.51553613e-01 2.85627395e-01
3.59528393e-01 3.50240976e-01 2.97386616e-01 1.00339150e+00
-1.18114734e+00 7.42989779e-01 -5.16490757e-01 -1.23459518e+00
-1.01775503e+00 3.85162592e-01 4.81530458e-01 -2.78905481e-01
-7.01856136e-01 7.13415623e-01 3.10668617e-01 -5.79777420e-01
3.65888059e-01 -3.05718690e-01 -3.01198754e-02 -3.10570132e-02
3.56710225e-01 5.42990267e-01 1.81138381e-01 -1.22641671e+00
-5.99388659e-01 8.75705123e-01 2.88869679e-01 -5.09795666e-01
1.30078292e+00 -8.35995078e-01 3.50649118e-01 6.37557447e-01
1.37554038e+00 -3.32895368e-01 -1.82950139e+00 -2.26476535e-01
2.90458314e-02 -6.38289094e-01 1.50954798e-01 -3.49709213e-01
-1.00506198e+00 1.10753262e+00 1.00963461e+00 -4.78081018e-01
9.26237047e-01 -3.89557660e-01 7.03879535e-01 4.55192119e-01
8.65611851e-01 -1.12548661e+00 1.40249282e-01 7.11025596e-01
5.47992647e-01 -1.42921913e+00 1.30872965e-01 -1.39848381e-01
-7.06358910e-01 1.04379582e+00 4.47240919e-01 -2.40650505e-01
2.01921061e-01 -5.85280284e-02 3.65319997e-01 -2.02462450e-01
-4.52430546e-01 -4.41268265e-01 9.93461683e-02 8.26072633e-01
1.83049932e-01 -1.82701319e-01 -1.76530122e-03 -3.68692994e-01
9.32006985e-02 1.48156673e-01 7.22809076e-01 1.10658813e+00
-4.19368684e-01 -7.17385054e-01 -5.76096654e-01 -4.64483172e-01
-5.00205867e-02 2.51061529e-01 -3.14141542e-01 1.13756490e+00
1.72410876e-01 9.00153577e-01 -9.63965952e-02 -2.28095293e-01
3.77361536e-01 -1.91066325e-01 8.64469469e-01 -3.07765722e-01
-3.61970253e-02 2.71048546e-01 1.26326010e-01 -9.43495750e-01
-7.49883533e-01 -1.05392230e+00 -1.17745125e+00 -1.04846254e-01
2.25427464e-01 -3.48696440e-01 8.73781681e-01 1.18920350e+00
-2.72652619e-02 1.95247039e-01 5.29240727e-01 -1.39713669e+00
-1.58043280e-01 -8.78680944e-01 -6.45313025e-01 4.96977866e-01
6.55474424e-01 -8.90429556e-01 -2.75146753e-01 1.17246628e-01] | [8.423195838928223, -2.2427845001220703] |
9c325236-5c28-4e30-ad8a-504fc2844d47 | communication-efficient-federated-learning-14 | 2202.02580 | null | https://arxiv.org/abs/2202.02580v1 | https://arxiv.org/pdf/2202.02580v1.pdf | Communication Efficient Federated Learning via Ordered ADMM in a Fully Decentralized Setting | The challenge of communication-efficient distributed optimization has attracted attention in recent years. In this paper, a communication efficient algorithm, called ordering-based alternating direction method of multipliers (OADMM) is devised in a general fully decentralized network setting where a worker can only exchange messages with neighbors. Compared to the classical ADMM, a key feature of OADMM is that transmissions are ordered among workers at each iteration such that a worker with the most informative data broadcasts its local variable to neighbors first, and neighbors who have not transmitted yet can update their local variables based on that received transmission. In OADMM, we prohibit workers from transmitting if their current local variables are not sufficiently different from their previously transmitted value. A variant of OADMM, called SOADMM, is proposed where transmissions are ordered but transmissions are never stopped for each node at each iteration. Numerical results demonstrate that given a targeted accuracy, OADMM can significantly reduce the number of communications compared to existing algorithms including ADMM. We also show numerically that SOADMM can accelerate convergence, resulting in communication savings compared to the classical ADMM. | ['Brian M. Sadler', 'Rick S. Blum', 'Yicheng Chen'] | 2022-02-05 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-6.08491674e-02 1.55460835e-01 -1.17465593e-01 -9.80613530e-02
-5.15046477e-01 -2.21818984e-01 4.06202942e-01 5.04956543e-01
-5.85052013e-01 1.13480091e+00 2.16578409e-01 -1.70966104e-01
-5.76441765e-01 -9.84478295e-01 -5.62643409e-01 -1.02246618e+00
-6.17466927e-01 8.99471343e-01 -2.14145809e-01 4.42019068e-02
1.15184173e-01 3.36104333e-01 -5.57661831e-01 -1.35070652e-01
5.67693889e-01 9.24022377e-01 2.15314969e-01 6.10851467e-01
-7.49898329e-02 5.90644956e-01 -6.36999786e-01 -2.21892502e-02
6.83195412e-01 -5.39944053e-01 -1.20712841e+00 3.05423707e-01
-2.15421598e-02 -5.98473430e-01 -3.64814281e-01 9.60468054e-01
5.00057220e-01 2.34878331e-01 2.31159911e-01 -1.36171985e+00
-2.20635533e-01 1.15793836e+00 -9.52128768e-01 -4.25249746e-04
1.74781635e-01 -1.15642935e-01 8.50824237e-01 -6.98123097e-01
7.06511974e-01 1.29679716e+00 7.36836731e-01 1.42576799e-01
-1.56046522e+00 -6.65798962e-01 5.34656167e-01 -1.24027081e-01
-1.47171676e+00 -7.44056702e-02 6.82040453e-01 -1.97252080e-01
4.95505095e-01 4.10375148e-01 7.09634185e-01 3.34835537e-02
4.06104535e-01 4.67682779e-01 1.05525470e+00 -3.99298489e-01
5.83258986e-01 -3.90104741e-01 -1.17644593e-01 6.73425138e-01
3.44993651e-01 -1.88871309e-01 -9.40977454e-01 -6.77048504e-01
5.88167727e-01 2.71101862e-01 -2.30930924e-01 -2.47422785e-01
-1.74615979e+00 1.12515426e+00 5.89814305e-01 1.85506999e-01
-8.75188231e-01 6.07470214e-01 1.01688288e-01 7.61224508e-01
8.40796530e-01 1.95594877e-01 -4.45143729e-01 1.81506723e-01
-1.10185826e+00 3.54180425e-01 8.15673172e-01 7.72091210e-01
1.21591222e+00 -2.24635348e-01 -3.34715366e-01 5.07361650e-01
4.32165861e-01 5.55847943e-01 -5.10665625e-02 -1.47130358e+00
5.55390656e-01 5.57522833e-01 1.52405977e-01 -1.26793373e+00
-6.51239693e-01 -4.61004734e-01 -1.48872054e+00 2.32121378e-01
2.09129319e-01 -7.43998230e-01 -3.95872384e-01 1.63550103e+00
8.45737875e-01 9.08149183e-02 1.02660559e-01 1.20496798e+00
1.66743174e-01 1.01031160e+00 -5.29662311e-01 -8.42271030e-01
7.65189469e-01 -1.13254285e+00 -6.59628391e-01 -1.10378869e-01
7.48821616e-01 -8.63444507e-01 -5.54445721e-02 4.78018165e-01
-1.50994670e+00 3.91565030e-03 -5.96753001e-01 1.83618844e-01
2.68414855e-01 -2.68533468e-01 8.57942164e-01 1.39674902e-01
-1.42970383e+00 6.37215376e-01 -1.06747985e+00 -2.60205381e-02
2.28147298e-01 7.27565527e-01 -1.58960924e-01 -1.46272644e-01
-8.04967582e-01 1.72040403e-01 1.08654290e-01 4.80902582e-01
-1.08371103e+00 -7.88348913e-01 -6.29071057e-01 -7.21868873e-02
3.75898957e-01 -9.87485349e-01 1.28054655e+00 -8.58684361e-01
-1.32608402e+00 1.34730920e-01 -5.70061564e-01 -7.16214538e-01
6.04754627e-01 1.74824014e-01 2.01304942e-01 8.68854150e-02
3.37882429e-01 5.67909241e-01 6.79052591e-01 -1.06642663e+00
-7.85434961e-01 -3.61243814e-01 1.51969329e-01 1.41874611e-01
-5.41963279e-01 -3.91690344e-01 -5.79465747e-01 -4.26532030e-01
4.92197603e-01 -1.11634898e+00 -8.65787089e-01 4.57100004e-01
-4.12146777e-01 -1.13255143e-01 6.37911856e-01 -1.86491162e-01
1.18063915e+00 -1.74737036e+00 5.26734173e-01 6.49549961e-01
8.47011030e-01 -1.27093494e-01 -1.16727568e-01 7.83717692e-01
6.00072980e-01 -1.37563750e-01 -2.96334445e-01 -6.85217023e-01
-1.11547701e-01 5.81918538e-01 1.67279482e-01 9.33197498e-01
-5.45216262e-01 4.62152630e-01 -1.19943964e+00 -5.31362236e-01
-2.31518317e-02 3.84135336e-01 -8.18837702e-01 -8.41985494e-02
-1.13996759e-01 7.56331623e-01 -8.50610733e-01 1.53166056e-01
1.17274690e+00 -6.20782852e-01 5.55640519e-01 1.35462806e-01
-3.36536109e-01 2.17464328e-01 -1.52950275e+00 1.65796113e+00
-6.32847905e-01 5.35657644e-01 1.08605993e+00 -1.21840715e+00
5.92590213e-01 5.36089182e-01 8.00506234e-01 -6.05451107e-01
-1.18007898e-01 2.33522847e-01 -2.80779123e-01 1.11704759e-01
5.09213388e-01 1.93836764e-02 1.07587263e-01 7.25696623e-01
-5.66708446e-01 1.03750728e-01 5.25739670e-01 7.62138188e-01
1.27076471e+00 -7.61678576e-01 8.16357881e-02 -5.00268638e-01
6.46097004e-01 1.14952885e-01 1.03518069e+00 1.08238542e+00
2.50435203e-01 1.68920591e-01 4.00629938e-01 -5.62607944e-01
-6.07074916e-01 -9.96635318e-01 2.27056563e-01 8.74702156e-01
4.08040732e-01 -6.64115548e-01 -4.80607718e-01 -5.06486118e-01
2.73108035e-01 -1.90979227e-01 -3.48459512e-01 4.50795919e-01
-5.66647410e-01 -6.50424302e-01 -6.67292327e-02 2.92159393e-02
7.01391280e-01 -3.92725319e-01 -2.90066719e-01 6.85289860e-01
-1.49665013e-01 -6.64463758e-01 -8.43521059e-01 -2.47357730e-02
-1.13510394e+00 -9.10697222e-01 -8.10495317e-01 -7.59841025e-01
1.27841151e+00 6.49278045e-01 1.05613124e+00 5.87230504e-01
9.81025845e-02 4.91430938e-01 -2.00764239e-01 7.31751472e-02
-3.20201308e-01 2.39967987e-01 -1.07663600e-02 1.89990968e-01
-2.71427810e-01 -5.90408802e-01 -9.72495675e-01 3.12564433e-01
-8.55321884e-01 1.74184114e-01 6.62493646e-01 6.95079088e-01
6.68461442e-01 3.41934294e-01 3.40208203e-01 -1.13480055e+00
6.26537025e-01 -8.25627685e-01 -9.77842093e-01 -4.04096805e-02
-8.01704109e-01 1.58523634e-01 7.98768640e-01 -1.97701737e-01
-8.76558006e-01 9.05003697e-02 2.26859167e-01 -2.78202686e-02
5.37942469e-01 6.96077585e-01 4.07356977e-01 -3.45493227e-01
9.92623195e-02 3.57198790e-02 2.57566571e-01 -5.45891225e-01
2.08909303e-01 3.43946964e-01 1.50499642e-01 -6.23433232e-01
9.14860666e-01 9.38406050e-01 5.63557327e-01 -7.14839339e-01
-5.54862022e-01 -4.04910982e-01 -3.53988588e-01 -1.55381262e-01
3.34561080e-01 -1.04384851e+00 -1.19602466e+00 4.87006575e-01
-1.29906917e+00 -3.37548077e-01 -1.42825380e-01 6.71390772e-01
7.70427585e-02 5.01580417e-01 -7.32566476e-01 -7.27193892e-01
-3.63252521e-01 -1.13490236e+00 6.43253624e-01 -2.36563534e-02
-2.24006161e-01 -1.54231334e+00 6.71571568e-02 1.40916556e-01
8.08542728e-01 2.30044141e-01 2.95550615e-01 -2.04187050e-01
-8.98938239e-01 1.67255148e-01 2.83091236e-02 -7.52966851e-02
1.41243815e-01 -3.68347734e-01 3.87968756e-02 -9.35561359e-01
-4.34105173e-02 -1.36682555e-01 6.60811603e-01 5.59744596e-01
8.14896047e-01 -9.21391129e-01 -7.94462621e-01 6.09050810e-01
1.28446293e+00 -1.98330417e-01 -1.88136458e-01 2.69191802e-01
3.48628610e-01 5.44682294e-02 3.90969008e-01 1.05856526e+00
6.06046617e-01 4.99544352e-01 6.24278426e-01 -3.18059146e-01
1.72728375e-01 2.47742489e-01 2.36902297e-01 1.08297467e+00
1.93927959e-01 -3.84514242e-01 -5.43478489e-01 7.70022392e-01
-2.08936763e+00 -3.67405146e-01 -6.19523525e-01 2.08356452e+00
9.12584424e-01 -2.73611277e-01 -1.91313580e-01 1.37539461e-01
6.91645443e-01 2.91164219e-01 -3.73696625e-01 -3.70890141e-01
1.67051047e-01 -3.24304663e-02 8.98180902e-01 9.33141410e-01
-6.96689725e-01 2.81722635e-01 5.98916197e+00 5.53204834e-01
-8.48615229e-01 4.37384486e-01 4.57939714e-01 -4.23131704e-01
-4.10354942e-01 3.13411176e-01 -8.12194824e-01 5.42128980e-01
5.74300766e-01 -3.72706831e-01 7.60560632e-01 6.09841943e-01
7.10040152e-01 -3.00973684e-01 -9.89224315e-01 6.97705507e-01
-2.32364327e-01 -1.58317423e+00 -8.70638415e-02 5.23109376e-01
1.42802382e+00 -1.03855789e-01 -1.16542630e-01 -3.01523060e-01
7.89082587e-01 -5.59151828e-01 2.60697186e-01 1.56771034e-01
5.71514741e-02 -8.73820782e-01 6.25400007e-01 3.79133224e-01
-1.36625433e+00 5.12857810e-02 -4.91600275e-01 -7.07075059e-01
5.09464204e-01 1.38898695e+00 -6.39470637e-01 7.54418969e-01
5.75138211e-01 6.14552975e-01 2.32813329e-01 1.06383622e+00
-7.27806762e-02 6.68523073e-01 -7.41613448e-01 6.12750500e-02
6.47538364e-01 -4.02900934e-01 7.07089126e-01 8.63067210e-01
3.96646589e-01 1.83806896e-01 8.12641859e-01 6.54605091e-01
-3.23695987e-01 -1.79518074e-01 -1.47724068e-02 4.25470322e-01
9.29996133e-01 1.33201468e+00 -5.95270276e-01 -4.56071854e-01
-3.94598216e-01 9.81449306e-01 2.63727933e-01 5.44224322e-01
-3.81093532e-01 -5.93209192e-02 9.24822748e-01 -9.68009084e-02
2.75203168e-01 -4.45402145e-01 -5.82301952e-02 -7.03627050e-01
4.81490828e-02 -6.42889798e-01 6.51684642e-01 -5.10369092e-02
-1.22044456e+00 1.75799906e-01 -1.81646451e-01 -8.77797902e-01
-2.02854380e-01 1.60049215e-01 -5.83289206e-01 7.62700617e-01
-1.56290495e+00 -8.20626855e-01 -2.05876470e-01 8.56501639e-01
3.12657565e-01 2.05423623e-01 4.55286771e-01 3.13511491e-01
-3.42006296e-01 2.86345661e-01 5.87299347e-01 1.07182331e-01
2.97498435e-01 -1.14341116e+00 -5.83545677e-02 8.67846072e-01
-9.44572315e-02 7.10001767e-01 7.03606427e-01 -6.38132811e-01
-1.73201120e+00 -1.22554755e+00 1.08225286e+00 4.90276426e-01
4.22193557e-01 -4.35293354e-02 -6.16134584e-01 8.55946124e-01
4.15187508e-01 1.91021487e-01 3.54667962e-01 -1.16751142e-01
4.32782412e-01 -5.64371824e-01 -1.08095992e+00 3.35993290e-01
8.11138570e-01 1.27461508e-01 -5.62243955e-03 8.72654557e-01
3.06529522e-01 -5.66376567e-01 -8.55367541e-01 -8.56978893e-02
2.01810729e-02 -7.57368684e-01 8.23697627e-01 -1.33201644e-01
-5.86646870e-02 -6.72680676e-01 1.27454683e-01 -1.57903039e+00
-4.15185153e-01 -1.35962534e+00 -1.25391245e-01 1.07617152e+00
2.81923890e-01 -8.38551283e-01 8.32584798e-01 4.85972404e-01
1.27338380e-01 -6.39308155e-01 -1.36375654e+00 -7.65235603e-01
2.59920526e-02 -1.72581375e-01 5.58972657e-01 8.86592388e-01
-9.93678123e-02 2.42625847e-01 -6.13049030e-01 7.26441383e-01
1.02111804e+00 3.05252045e-01 1.03796375e+00 -1.03225791e+00
-5.09388685e-01 -7.05731660e-02 -5.81395738e-02 -1.69908285e+00
1.12231364e-02 -1.00512445e+00 -8.42626393e-02 -1.94697845e+00
3.01168442e-01 -7.76665449e-01 -1.15263639e-02 3.87405366e-01
1.21862337e-01 3.96107197e-01 2.62905657e-01 2.88476467e-01
-7.83848763e-01 3.07757914e-01 1.36510277e+00 -2.75964379e-01
-2.24118829e-01 -1.78575367e-02 -4.17988628e-01 5.93857169e-01
5.72263300e-01 -8.24219942e-01 -3.36628795e-01 -9.74399269e-01
3.47128093e-01 5.45324802e-01 2.99241811e-01 -6.88825011e-01
7.88311064e-01 -3.19998175e-01 1.28860280e-01 -7.40489483e-01
3.12487125e-01 -1.01921165e+00 5.02871454e-01 1.04399395e+00
-4.34853345e-01 3.44925940e-01 -3.88804257e-01 5.81369638e-01
-1.10166371e-01 7.68133849e-02 5.30016124e-01 -7.22590387e-02
-3.75911266e-01 4.49070632e-01 -7.70566344e-01 -1.00704812e-01
1.11231720e+00 3.50761861e-01 -2.92877108e-02 -6.75107419e-01
-6.04373753e-01 1.06395483e+00 3.03068101e-01 -3.49120498e-01
5.20905197e-01 -1.35869157e+00 -9.64379430e-01 5.74672455e-03
-5.91854990e-01 5.38515031e-01 2.59727806e-01 1.27726388e+00
-4.76764262e-01 2.92144507e-01 5.29495597e-01 -7.02921569e-01
-1.26005292e+00 4.05414373e-01 1.06370322e-01 -5.41438878e-01
-7.44608581e-01 9.30896759e-01 6.17670417e-02 -3.61157835e-01
1.52334392e-01 -2.71231562e-01 5.94555199e-01 -2.43143484e-01
5.55823028e-01 9.34559345e-01 -1.31403267e-01 -2.47525528e-01
-3.98361355e-01 4.41220582e-01 -1.51387841e-01 -2.29656562e-01
1.54889131e+00 -6.11229300e-01 -8.76855552e-01 2.70511322e-02
1.33687782e+00 3.15388590e-01 -1.23752499e+00 -6.21587634e-01
-4.06007499e-01 -7.81313777e-01 5.06003678e-01 -3.41847271e-01
-1.91232371e+00 3.19332212e-01 -5.07465117e-02 2.11917087e-01
1.17902660e+00 -1.05231762e-01 1.07834327e+00 5.79246342e-01
6.81512475e-01 -1.06586277e+00 -1.77817971e-01 3.90748382e-01
4.29065168e-01 -8.09407055e-01 4.65563029e-01 -3.09925497e-01
-1.10850655e-01 1.04172134e+00 1.95258647e-01 -8.20594281e-02
6.55053020e-01 4.09779191e-01 -1.79251164e-01 -5.28600365e-02
-9.49283361e-01 2.77531087e-01 -3.78797680e-01 2.60618210e-01
1.92580462e-01 -6.40343204e-02 -7.74515808e-01 -4.52305861e-02
6.57028193e-03 2.16690637e-02 6.01903081e-01 1.15201437e+00
-3.81093681e-01 -1.39553761e+00 -7.51637101e-01 3.38755965e-01
-4.94084060e-01 7.55811408e-02 -1.47393629e-01 6.38118923e-01
-1.31707331e-02 1.26182842e+00 3.96969497e-01 1.58146679e-01
-2.87727535e-01 -5.69571555e-01 1.33658379e-01 -3.41198444e-01
-6.29020989e-01 1.10347830e-01 -8.38730559e-02 -5.41827977e-01
-6.39180005e-01 -4.21863347e-01 -1.52547073e+00 -1.10434735e+00
-1.64533556e-01 8.62850368e-01 4.60607857e-01 9.69564140e-01
5.65190494e-01 3.45519453e-01 1.15602648e+00 -8.45535755e-01
-4.49530095e-01 -4.93068248e-01 -6.25034988e-01 4.96047474e-02
7.51294017e-01 -3.28985244e-01 -4.25391644e-01 -2.23068997e-01] | [6.244511127471924, 4.948222637176514] |
66cc12d6-6ce8-41e7-8276-1f74395f62f6 | bundletrack-6d-pose-tracking-for-novel | 2108.00516 | null | https://arxiv.org/abs/2108.00516v1 | https://arxiv.org/pdf/2108.00516v1.pdf | BundleTrack: 6D Pose Tracking for Novel Objects without Instance or Category-Level 3D Models | Tracking the 6D pose of objects in video sequences is important for robot manipulation. Most prior efforts, however, often assume that the target object's CAD model, at least at a category-level, is available for offline training or during online template matching. This work proposes BundleTrack, a general framework for 6D pose tracking of novel objects, which does not depend upon 3D models, either at the instance or category-level. It leverages the complementary attributes of recent advances in deep learning for segmentation and robust feature extraction, as well as memory-augmented pose graph optimization for spatiotemporal consistency. This enables long-term, low-drift tracking under various challenging scenarios, including significant occlusions and object motions. Comprehensive experiments given two public benchmarks demonstrate that the proposed approach significantly outperforms state-of-art, category-level 6D tracking or dynamic SLAM methods. When compared against state-of-art methods that rely on an object instance CAD model, comparable performance is achieved, despite the proposed method's reduced information requirements. An efficient implementation in CUDA provides a real-time performance of 10Hz for the entire framework. Code is available at: https://github.com/wenbowen123/BundleTrack | ['Kostas Bekris', 'Bowen Wen'] | 2021-08-01 | null | null | null | null | ['template-matching', 'real-time-visual-tracking', '6d-pose-estimation-using-rgbd', 'video-object-tracking', '3d-object-tracking'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [-1.49889857e-01 -4.36672688e-01 -3.65811586e-01 -1.52983531e-01
-8.05016816e-01 -6.53824687e-01 4.13041621e-01 6.76955730e-02
-3.10870141e-01 2.89028525e-01 -4.60846931e-01 -7.25459754e-02
-7.99509324e-03 -4.07444328e-01 -1.06140673e+00 -4.58932161e-01
-2.11833879e-01 8.87184918e-01 6.74290776e-01 -6.99040517e-02
2.20060825e-01 9.84164596e-01 -1.59952748e+00 -2.96036869e-01
5.08735061e-01 1.17336261e+00 4.11837935e-01 4.82067049e-01
-3.25745232e-02 3.19002420e-01 -4.01153535e-01 -2.14885801e-01
6.58496976e-01 2.22840130e-01 -4.44992065e-01 2.75845826e-01
1.02618122e+00 -3.91211510e-01 -6.60566449e-01 1.15011978e+00
3.71680170e-01 2.22263709e-01 2.74133086e-01 -1.33395851e+00
-6.36442900e-02 7.18296086e-03 -6.37406468e-01 1.32866800e-01
4.91224170e-01 4.86721754e-01 6.40780687e-01 -1.05227184e+00
1.04774666e+00 1.25023174e+00 9.21064496e-01 3.08076024e-01
-1.05082548e+00 -9.29227531e-01 4.05141294e-01 3.19359720e-01
-1.49305701e+00 -4.62225646e-01 7.34177768e-01 -6.42301738e-01
1.11955059e+00 -1.44591825e-02 8.58560741e-01 9.23493505e-01
5.80327451e-01 7.63570547e-01 5.93368828e-01 -2.08563637e-02
-8.22960883e-02 -3.31087887e-01 1.47096589e-01 8.83047998e-01
4.89962101e-01 2.46278569e-01 -7.95608819e-01 -4.46319729e-02
8.93709421e-01 7.79194012e-02 -1.68039784e-01 -1.26567876e+00
-1.55162311e+00 3.75093043e-01 4.83939797e-01 -3.47636803e-03
-4.94744182e-01 5.65428972e-01 5.40665448e-01 2.34507620e-01
3.89581561e-01 9.75667536e-02 -4.46097046e-01 -1.77546561e-01
-7.70640492e-01 3.88361752e-01 7.31984794e-01 1.77366221e+00
7.31472194e-01 1.15421914e-01 -1.31510273e-01 1.27276719e-01
3.91579390e-01 7.93516815e-01 -2.28381269e-02 -9.75367546e-01
4.01298732e-01 5.82915425e-01 3.38964641e-01 -1.17109454e+00
-5.16113281e-01 -5.70486248e-01 -3.41445982e-01 3.42608422e-01
2.71191746e-01 3.19834292e-01 -1.08911526e+00 1.53141248e+00
9.83281851e-01 6.30149007e-01 -2.35679820e-01 1.09885025e+00
9.47883248e-01 2.26109967e-01 -3.09447885e-01 -3.85189243e-02
1.06323779e+00 -1.10065317e+00 -7.27721691e-01 -2.61416912e-01
4.44586903e-01 -8.99759114e-01 5.57278931e-01 1.94154978e-01
-9.89965737e-01 -7.55211294e-01 -1.11324394e+00 -1.64047360e-01
-2.76217967e-01 2.32768357e-02 7.15024769e-01 4.78465289e-01
-9.16415155e-01 5.56789875e-01 -1.26432872e+00 -5.70592582e-01
4.51119959e-01 6.33881509e-01 -5.59539437e-01 -1.31347224e-01
-6.06823206e-01 1.07670152e+00 3.47148389e-01 2.75601327e-01
-1.16464150e+00 -7.72747278e-01 -1.06866920e+00 -3.99675399e-01
8.10159028e-01 -8.59073937e-01 1.29159129e+00 -2.00405747e-01
-1.30111051e+00 9.46277857e-01 -2.17052802e-01 -4.18744445e-01
8.53383899e-01 -6.63721621e-01 9.34875309e-02 7.96335042e-02
1.19020365e-01 5.12777507e-01 9.27591622e-01 -1.25538588e+00
-6.16416097e-01 -6.31675124e-01 2.66348943e-02 2.97197878e-01
2.11206958e-01 -1.87710762e-01 -1.05255294e+00 -3.94198716e-01
6.20621383e-01 -1.29072428e+00 -2.95183122e-01 5.60172081e-01
-1.45678759e-01 -2.16263905e-01 1.27713501e+00 -4.54888880e-01
6.02417231e-01 -2.04640627e+00 2.38305196e-01 -3.58620211e-02
6.09987862e-02 1.47982329e-01 -2.10569166e-02 2.86506623e-01
3.22787523e-01 -5.67867935e-01 8.76605362e-02 -6.67135239e-01
2.37017512e-01 1.37783423e-01 -6.51029423e-02 1.11389112e+00
1.72145702e-02 1.02461791e+00 -1.00694418e+00 -4.78905916e-01
6.35011256e-01 4.78003263e-01 -4.34286147e-01 1.69045508e-01
-3.12770039e-01 7.09402621e-01 -3.70123923e-01 9.59618211e-01
9.21665490e-01 -2.24922612e-01 5.50515056e-02 -5.12452483e-01
-1.86660677e-01 3.08623910e-01 -1.37373018e+00 2.48126626e+00
-8.92920122e-02 4.28525031e-01 3.00982356e-01 -7.33982742e-01
8.74891818e-01 9.67943743e-02 8.07811975e-01 -3.35189641e-01
2.10445881e-01 3.27946573e-01 -2.99743772e-01 -2.07708001e-01
6.52900338e-01 5.87199271e-01 -2.05925897e-01 5.56862652e-02
1.51394680e-01 -3.98655027e-01 1.93018660e-01 1.55251428e-01
1.08604515e+00 8.62719893e-01 1.93565547e-01 -1.30253151e-01
3.16872954e-01 4.33696866e-01 7.05306828e-01 8.50873828e-01
-3.77630591e-01 3.54960531e-01 -2.17602745e-01 -4.05614734e-01
-9.17348981e-01 -1.02419579e+00 -1.17342062e-01 6.83353782e-01
9.86505449e-01 -3.21556985e-01 -1.97772324e-01 -4.92414653e-01
4.91498351e-01 2.14440912e-01 -2.21475661e-01 -1.43966630e-01
-8.74975383e-01 -2.23998591e-01 1.27442300e-01 5.23299813e-01
3.25907052e-01 -6.90294266e-01 -1.06587362e+00 3.84004742e-01
9.25676078e-02 -1.35205638e+00 -4.73350793e-01 1.26032993e-01
-9.94839132e-01 -1.12863684e+00 -3.68289173e-01 -6.47230446e-01
5.26323795e-01 7.33836234e-01 1.05516195e+00 1.95401594e-01
-4.19248968e-01 7.68805206e-01 -3.09789896e-01 -2.93952823e-01
-8.67807940e-02 1.58307627e-02 3.60561639e-01 -3.28890711e-01
3.70378703e-01 -3.61029655e-01 -5.44337749e-01 3.91938359e-01
-3.08601886e-01 -1.03445642e-01 4.89954710e-01 6.73156619e-01
9.29170489e-01 -2.60174334e-01 4.60834652e-02 -3.85569036e-01
-2.75825292e-01 -1.94891274e-01 -1.08856881e+00 6.12735655e-03
-3.72561246e-01 -3.51397812e-01 -1.00776762e-01 -4.69738156e-01
-8.04730237e-01 5.88255763e-01 2.60122791e-02 -1.20664179e+00
-1.46489471e-01 2.25935966e-01 -1.60264313e-01 -6.31827772e-01
2.69101739e-01 1.39089227e-01 1.57422960e-01 -5.77760637e-01
1.98706180e-01 1.83010116e-01 7.65998721e-01 -5.14992774e-01
1.12229693e+00 5.62874496e-01 1.64090499e-01 -4.24689233e-01
-6.66734755e-01 -9.38348234e-01 -1.10451090e+00 -5.48577189e-01
6.81268692e-01 -1.15992749e+00 -7.35424042e-01 6.41806901e-01
-1.36057162e+00 -2.94568121e-01 -4.48568873e-02 6.21355653e-01
-7.73040593e-01 2.73081005e-01 -4.13721025e-01 -6.45766556e-01
-2.90792137e-01 -1.33924270e+00 1.52350652e+00 5.93309924e-02
5.35501651e-02 -5.96060455e-01 -3.34693462e-01 2.87879765e-01
1.17645808e-01 6.38409853e-01 2.30949163e-01 -4.65745419e-01
-1.36913121e+00 -5.06912649e-01 2.15460286e-02 -3.27406436e-01
1.69219390e-01 -9.02592391e-02 -5.60574293e-01 -8.34524632e-01
-3.95334400e-02 -8.93385112e-02 5.84489226e-01 3.05542290e-01
8.10507357e-01 8.12739059e-02 -8.48138034e-01 8.70418847e-01
1.41265821e+00 2.43918598e-01 1.99879944e-01 5.38354158e-01
1.07060623e+00 2.60083407e-01 1.42103100e+00 4.14954454e-01
3.80051076e-01 1.07865000e+00 9.71100092e-01 1.62715957e-01
-1.78704083e-01 -2.12597381e-02 2.52790421e-01 7.06660926e-01
6.77086189e-02 1.41147161e-02 -9.70786512e-01 5.09957612e-01
-2.04625058e+00 -8.46360445e-01 -2.93262869e-01 2.23287034e+00
3.62054467e-01 3.12086135e-01 -5.58594726e-02 -2.84215242e-01
8.14156234e-01 2.42583826e-01 -1.02590156e+00 2.93777794e-01
1.65055603e-01 -9.72759053e-02 8.39294672e-01 3.70372057e-01
-1.37563753e+00 1.17612660e+00 5.18642044e+00 5.41352153e-01
-1.03507328e+00 2.69381315e-01 -2.97656834e-01 -2.52906889e-01
3.08473259e-01 -3.44416536e-02 -1.23470211e+00 7.55523518e-02
3.45095098e-01 -1.67160735e-01 1.21162519e-01 1.01265335e+00
-3.37940380e-02 -9.40977260e-02 -1.31126392e+00 1.12267566e+00
1.20036118e-01 -1.39814460e+00 -2.79511839e-01 9.86059383e-03
5.98425627e-01 3.37179482e-01 -4.94845817e-03 2.02773690e-01
1.04588583e-01 -3.72887492e-01 1.10220277e+00 4.49966997e-01
5.18499970e-01 -6.60239220e-01 7.37167597e-01 3.43891233e-01
-1.63516057e+00 2.07847238e-01 -3.28854442e-01 1.70061976e-01
3.30327809e-01 2.39921004e-01 -9.28324282e-01 8.68134558e-01
1.06314063e+00 1.00424898e+00 -5.27484179e-01 1.34995234e+00
-5.22617437e-03 5.52224927e-02 -5.06482422e-01 3.18584889e-01
2.63448536e-01 6.19375752e-03 1.07835293e+00 1.01907170e+00
3.64103824e-01 -8.27770010e-02 6.76826835e-01 5.89266479e-01
1.73106194e-01 -1.67347729e-01 -6.63147211e-01 1.96847707e-01
8.25358212e-01 1.08019245e+00 -8.37679327e-01 -2.61041462e-01
-3.79070997e-01 8.34917903e-01 1.26259610e-01 1.15956636e-02
-1.11578226e+00 -8.62026140e-02 8.43617201e-01 1.40339816e-02
6.32293820e-01 -9.68753040e-01 -3.58581990e-02 -9.86052871e-01
3.39852124e-01 -7.73655117e-01 2.45412976e-01 -6.47605062e-01
-9.96508539e-01 3.48567218e-01 6.70264214e-02 -1.57759488e+00
-9.73727629e-02 -5.60956478e-01 -6.92053810e-02 5.55906892e-01
-1.44238353e+00 -1.39871001e+00 -7.75888205e-01 6.37000918e-01
7.84830511e-01 2.85796858e-02 3.97798955e-01 3.50719959e-01
-3.39587808e-01 4.08354282e-01 1.31822769e-02 2.04614736e-02
6.86507940e-01 -8.58657360e-01 6.39606178e-01 9.57445323e-01
1.86567679e-01 6.24687374e-01 7.50317633e-01 -9.72768188e-01
-2.32160187e+00 -1.27750278e+00 4.74586546e-01 -6.71138704e-01
5.89655459e-01 -6.77368224e-01 -8.89383614e-01 8.98372233e-01
-1.76073909e-01 2.84482300e-01 4.60280925e-02 -2.41964817e-01
-1.06365509e-01 -5.47058135e-02 -9.81354594e-01 3.79006803e-01
1.68674147e+00 -1.70606151e-01 -5.24410903e-01 6.17916465e-01
1.00747859e+00 -1.46868861e+00 -1.08946979e+00 8.87852132e-01
6.07554972e-01 -7.68510878e-01 1.21583247e+00 -2.93445289e-01
-3.52520734e-01 -8.26413572e-01 -1.63778558e-01 -7.15248644e-01
-2.85329521e-01 -7.57540464e-01 -6.23460710e-01 8.33487630e-01
-9.85569581e-02 -3.97973776e-01 9.97397363e-01 4.16665733e-01
-6.04246438e-01 -6.16808295e-01 -1.14119649e+00 -1.33870685e+00
-5.87505102e-01 -4.77712393e-01 6.34684026e-01 7.03102529e-01
-6.89150631e-01 -1.16294496e-01 -3.64397824e-01 7.24102020e-01
9.34984207e-01 5.15963852e-01 1.44226480e+00 -1.19663835e+00
6.20660074e-02 -2.20995963e-01 -9.17072415e-01 -1.37460470e+00
3.58459473e-01 -8.20814312e-01 3.53602409e-01 -1.51719368e+00
-1.15726762e-01 -6.44706428e-01 -5.47079481e-02 3.86007428e-01
1.03239961e-01 2.92214274e-01 3.30033749e-01 4.19652760e-01
-9.36467707e-01 6.16066515e-01 1.30309975e+00 -2.85430342e-01
-2.04609752e-01 -2.77803652e-02 1.46124706e-01 7.01108575e-01
5.39096415e-01 -6.32873654e-01 -6.82349801e-02 -5.94354868e-01
-3.21457833e-01 6.79643676e-02 7.36681104e-01 -1.37659907e+00
6.20379329e-01 -1.72501430e-01 2.48309433e-01 -1.32555938e+00
8.21635783e-01 -1.14387667e+00 6.19079590e-01 7.84339190e-01
1.92843035e-01 2.49716073e-01 4.63482141e-01 8.52976799e-01
-2.42189579e-02 -6.66851327e-02 5.74508429e-01 -1.79350287e-01
-1.29603481e+00 8.58866274e-01 5.52500151e-02 -3.70281905e-01
1.29276788e+00 -5.31923294e-01 -1.12091333e-01 5.37316538e-02
-6.30543768e-01 3.39896411e-01 8.85449827e-01 8.15750062e-01
6.31360888e-01 -1.38185835e+00 -4.49683815e-01 -4.08048406e-02
2.10748315e-01 4.32265848e-01 1.81168139e-01 1.20055938e+00
-6.12328410e-01 4.78358030e-01 -1.98045626e-01 -1.44369352e+00
-1.29307222e+00 6.42211795e-01 2.58260965e-01 6.38889596e-02
-9.30537939e-01 8.68564248e-01 -1.94912776e-02 -3.64897460e-01
5.51394582e-01 -1.04859896e-01 3.60736668e-01 -1.00414336e-01
1.17611259e-01 3.45751941e-01 2.57922053e-01 -9.40714240e-01
-7.81775951e-01 9.20543551e-01 -4.45432067e-02 2.99478322e-01
1.20697033e+00 -2.55956054e-01 1.98361576e-02 5.06141484e-01
9.53705907e-01 -3.21808517e-01 -1.55778897e+00 -2.98328966e-01
6.66469112e-02 -9.41611707e-01 -8.38833675e-02 -2.48869643e-01
-1.08645558e+00 5.76371431e-01 8.11497927e-01 -4.61394519e-01
6.76533759e-01 8.56631622e-02 8.68397474e-01 6.62837625e-01
1.17233407e+00 -7.64129102e-01 1.61297202e-01 7.21000135e-01
9.48376179e-01 -1.40542841e+00 2.19699666e-01 -6.22237325e-01
-7.19282627e-02 9.72447038e-01 9.71587837e-01 -3.61372292e-01
4.66121048e-01 3.55522841e-01 1.66102983e-02 -3.92458290e-01
-4.71106231e-01 -2.28656515e-01 3.44965488e-01 6.66913927e-01
-1.68244373e-02 -3.61436456e-01 1.56938091e-01 -4.82254922e-02
1.73222814e-02 -2.37989664e-01 5.90255037e-02 1.49888027e+00
-3.36955428e-01 -7.89525688e-01 -4.49284464e-01 2.31555462e-01
-1.02236912e-01 2.89681733e-01 -1.58441648e-01 1.15026796e+00
5.29055968e-02 6.10154867e-01 1.82374775e-01 -3.39495450e-01
4.52023506e-01 -2.23193318e-01 9.51792061e-01 -7.43677735e-01
-5.55724621e-01 1.33331627e-01 4.93156835e-02 -1.16331840e+00
-7.83072531e-01 -1.02593160e+00 -1.25822663e+00 -2.42198333e-01
-6.92214549e-01 -2.29159564e-01 7.88278937e-01 7.50845551e-01
6.04133785e-01 3.40406835e-01 2.08039820e-01 -1.63474500e+00
-5.08609533e-01 -6.32944524e-01 -1.81322828e-01 2.63669282e-01
3.89587283e-01 -1.43384767e+00 -7.40081221e-02 8.83238669e-03] | [6.890987873077393, -2.2801105976104736] |
2a53de3e-78ff-40c9-982a-1d34c7ead1a7 | connectivity-constrained-interactive-panoptic | 2212.06756 | null | https://arxiv.org/abs/2212.06756v1 | https://arxiv.org/pdf/2212.06756v1.pdf | Connectivity-constrained Interactive Panoptic Segmentation | We address interactive panoptic annotation, where one segment all object and stuff regions in an image. We investigate two graph-based segmentation algorithms that both enforce connectivity of each region, with a notable class-aware Integer Linear Programming (ILP) formulation that ensures global optimum. Both algorithms can take RGB, or utilize the feature maps from any DCNN, whether trained on the target dataset or not, as input. We then propose an interactive, scribble-based annotation framework. | ['Thomas Guthier', 'Ismail Ben Ayed', 'Andrea Lodi', 'Bo Tang', 'Ruobing Shen'] | 2022-12-13 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [ 7.01440692e-01 5.87694407e-01 -6.78120792e-01 -4.56600994e-01
-3.60690832e-01 -1.20360076e+00 1.68434054e-01 -5.51389381e-02
-5.89002762e-03 6.21439934e-01 -4.10951316e-01 -3.21283251e-01
-3.02363068e-01 -1.08200848e+00 -6.31852448e-01 -7.11194873e-01
-1.44338414e-01 6.79167747e-01 4.65909809e-01 1.72083348e-01
3.17690581e-01 8.17747891e-01 -1.31854689e+00 2.29376778e-02
1.07523799e+00 1.58514678e+00 -1.37880063e-02 8.29357862e-01
-4.69011158e-01 5.72851598e-01 -3.94925267e-01 -3.88733387e-01
8.57403517e-01 -1.78248644e-01 -1.09940207e+00 6.06157959e-01
6.27960920e-01 -1.17608465e-01 -3.78916338e-02 1.24711335e+00
-2.92024612e-02 1.31572336e-01 6.65672839e-01 -1.37717164e+00
-8.54568779e-01 6.21437311e-01 -9.66086864e-01 -9.39096324e-03
2.40879394e-02 3.10506970e-01 1.22330594e+00 -3.72500271e-01
9.05962646e-01 1.08673513e+00 3.56506854e-01 3.71838719e-01
-1.16763532e+00 -3.10220510e-01 5.62319398e-01 -1.58337608e-01
-1.27571619e+00 2.11449713e-01 8.70134890e-01 -2.43660256e-01
6.73743844e-01 5.52024782e-01 1.40257895e+00 1.77839801e-01
1.46798700e-01 1.02598822e+00 1.14192605e+00 -3.30325812e-01
2.80828387e-01 -1.06222250e-01 2.09953725e-01 9.48186576e-01
1.13774672e-01 -2.02466100e-01 -3.45114350e-01 -1.98821630e-02
9.57994223e-01 -1.79772139e-01 -3.81608903e-01 -5.70917189e-01
-9.26121712e-01 7.82424748e-01 7.89834201e-01 -1.33604392e-01
-2.23213196e-01 4.81556416e-01 1.77895054e-01 -1.04101181e-01
7.64460027e-01 6.66704297e-01 -7.69422650e-01 5.40380836e-01
-1.17529202e+00 -1.13815926e-01 9.37654138e-01 1.14217365e+00
1.10265183e+00 -6.26978278e-02 2.48158630e-02 5.48937201e-01
3.26802701e-01 3.32874954e-01 -2.87529916e-01 -1.28525198e+00
1.68954834e-01 1.16216874e+00 -1.71080217e-01 -9.71647322e-01
-5.18434525e-01 -2.76045531e-01 -5.08489370e-01 3.75506431e-01
2.50644982e-01 -2.17414871e-01 -1.61091471e+00 1.06730163e+00
7.02357113e-01 7.17246979e-02 -4.99590486e-02 1.04776371e+00
8.19396973e-01 6.58205271e-01 1.98445618e-02 -1.64674088e-01
1.07868397e+00 -1.46544552e+00 -6.22384250e-01 -1.86116233e-01
1.37287557e-01 -3.30689639e-01 9.17120576e-01 6.82169974e-01
-8.79469872e-01 -1.52543381e-01 -9.46313739e-01 -2.40729213e-01
-8.67582619e-01 -6.88509792e-02 9.71221924e-01 5.94825864e-01
-1.26940536e+00 4.04613793e-01 -5.56677759e-01 -2.51398921e-01
6.32551253e-01 7.22348213e-01 -5.47394715e-02 1.34042919e-01
-7.61817098e-01 6.36708736e-01 8.01218212e-01 4.24675226e-01
-8.41712594e-01 -5.25212348e-01 -8.08711648e-01 -2.42784366e-01
5.91788650e-01 -5.63964546e-01 7.74665475e-01 -1.38062048e+00
-1.36796713e+00 1.29613042e+00 2.85484552e-01 -3.55276018e-01
3.60004544e-01 -8.60504210e-02 7.03702644e-02 5.16125500e-01
-2.62301356e-01 1.48643291e+00 8.79301906e-01 -1.69467795e+00
-6.35869086e-01 -3.29398036e-01 5.04940927e-01 4.83343989e-01
2.35500722e-03 -1.79934099e-01 -1.01978374e+00 -5.32894254e-01
4.17770237e-01 -8.40381205e-01 -7.06950665e-01 5.46600640e-01
-1.16905785e+00 1.02883605e-02 1.13392878e+00 -6.55336142e-01
1.06877959e+00 -2.02761841e+00 1.88357562e-01 7.30992258e-01
1.88842431e-01 -1.40771851e-01 -4.08621468e-02 -1.56723067e-01
4.10588205e-01 6.08765244e-01 -1.01169622e+00 2.66416725e-02
-1.64324850e-01 7.98942626e-01 -2.45130986e-01 4.61422861e-01
2.65613973e-01 1.05423641e+00 -7.67565548e-01 -9.30059969e-01
2.62629449e-01 1.25791892e-01 -3.13910484e-01 1.62843049e-01
-8.49956870e-01 5.92247248e-01 -6.09376252e-01 1.26419902e+00
1.01114726e+00 -7.87360128e-03 6.13482706e-02 -2.25965455e-01
-3.24592888e-01 -4.79898095e-01 -1.06231737e+00 2.13266325e+00
-3.86226252e-02 6.89939559e-01 4.03540075e-01 -8.83422256e-01
9.37845230e-01 -1.84382066e-01 8.17998707e-01 -4.55782264e-01
1.68483391e-01 1.27157718e-01 -2.69379586e-01 -4.52192217e-01
6.45049930e-01 3.18524212e-01 -7.27805570e-02 3.86689991e-01
1.30395427e-01 -9.35604692e-01 1.41088888e-01 1.73805550e-01
7.56599069e-01 6.00430489e-01 3.26644152e-01 -6.37652993e-01
1.46306485e-01 6.43102646e-01 4.23655450e-01 8.76607895e-01
-2.34146819e-01 1.02241039e+00 6.26726449e-01 -5.33317208e-01
-6.52507186e-01 -9.70040143e-01 -2.74792641e-01 1.16814125e+00
8.61841679e-01 3.00509840e-01 -9.47735310e-01 -9.94267404e-01
1.41633220e-03 4.64214444e-01 -9.75470364e-01 4.46785629e-01
-5.01156747e-01 -7.99358845e-01 3.09741020e-01 5.26434541e-01
5.64263999e-01 -8.65878701e-01 -9.41675127e-01 7.59438053e-02
2.51988411e-01 -9.28278089e-01 -3.04623753e-01 8.02974820e-01
-8.15880358e-01 -1.10638392e+00 -4.24157202e-01 -7.68344164e-01
8.92018199e-01 -2.18085513e-01 1.25596130e+00 2.85275400e-01
-4.88132507e-01 6.49183869e-01 -3.37582290e-01 -2.37899169e-01
3.83849025e-01 1.97405770e-01 -7.33854294e-01 3.17937694e-02
-1.46228418e-01 -3.35222751e-01 -7.47185290e-01 3.16858917e-01
-8.95137906e-01 2.32982457e-01 2.58003771e-01 1.61778882e-01
1.36579728e+00 -1.65615007e-01 -8.12886003e-03 -1.15762055e+00
2.30491638e-01 -3.48711878e-01 -8.54552448e-01 8.71274412e-01
-3.43507260e-01 -2.69628078e-01 1.75651088e-01 -1.93445817e-01
-8.86553049e-01 7.45593727e-01 1.38592854e-01 -5.46374798e-01
-5.66824377e-02 3.32190305e-01 -4.00147259e-01 -6.74096763e-01
4.45064515e-01 -1.84744634e-02 -5.51044881e-01 -1.79519609e-01
1.04020977e+00 3.94769162e-01 1.02989995e+00 -6.18815124e-01
6.71782076e-01 6.84518993e-01 -2.33233739e-02 -2.72719920e-01
-1.06533706e+00 -6.52255177e-01 -1.36376703e+00 -6.91706002e-01
1.27400482e+00 -5.11504054e-01 -2.90006846e-01 2.14703992e-01
-1.12688613e+00 -7.45604455e-01 -6.28918827e-01 -7.89221749e-02
-6.85829520e-01 -1.01980329e-01 -2.08159402e-01 -8.79622936e-01
-3.92682880e-01 -1.11701906e+00 1.20638454e+00 6.53155684e-01
3.32600743e-01 -1.16608393e+00 -1.93144977e-01 1.46568090e-01
-8.03685933e-02 9.50702965e-01 7.34733403e-01 -6.52507126e-01
-9.97748613e-01 -2.52475794e-02 -7.63212085e-01 1.90751940e-01
9.15055349e-03 6.16881490e-01 -9.44202363e-01 2.58146524e-01
-2.86482215e-01 -4.19494480e-01 8.36505294e-01 6.26245081e-01
1.63619137e+00 -6.45589292e-01 -4.72707331e-01 1.01803970e+00
1.96722579e+00 3.67068589e-01 4.48983967e-01 1.32310316e-01
1.10153568e+00 8.33343804e-01 7.11177707e-01 2.12551951e-01
2.82960385e-01 2.16538981e-01 1.07278275e+00 -4.96241242e-01
-1.90298066e-01 5.94575927e-02 -2.47393325e-01 3.78434986e-01
-6.22903369e-02 -8.09274793e-01 -1.06582844e+00 6.58443809e-01
-1.83523142e+00 -1.43458813e-01 -4.73092079e-01 1.47313023e+00
7.80736387e-01 -1.99471489e-01 -1.56425476e-01 -1.37906358e-01
7.60951698e-01 2.67511696e-01 -1.09081590e+00 -7.29818165e-01
-3.32966387e-01 3.52375269e-01 1.05817389e+00 4.33783650e-01
-1.37937546e+00 1.19059193e+00 7.52135420e+00 7.80188978e-01
-1.08293200e+00 6.82898015e-02 1.19392443e+00 1.80347353e-01
-4.96441990e-01 1.57644093e-01 -5.64112365e-01 6.43615127e-02
2.64286846e-01 8.30795318e-02 6.48657620e-01 8.30377162e-01
-2.22751126e-02 -6.85254157e-01 -9.17139590e-01 6.05184436e-01
6.29638880e-02 -1.45042586e+00 1.34798124e-01 3.22992727e-02
1.21940887e+00 -3.30409184e-02 -9.71515179e-02 -3.25243503e-01
8.47317159e-01 -1.12539995e+00 9.30352390e-01 4.42525446e-01
8.87812436e-01 -4.94049847e-01 3.18058014e-01 9.39733535e-02
-1.36181903e+00 7.99120069e-02 -9.37592983e-02 3.38066399e-01
2.44404495e-01 6.07449830e-01 -4.08452421e-01 6.48703456e-01
7.28499413e-01 4.65967894e-01 -7.20329046e-01 1.22551036e+00
-5.22032976e-01 6.98826313e-01 -7.55222797e-01 2.51393139e-01
5.65545619e-01 -5.65077484e-01 3.53796065e-01 1.30730045e+00
-1.33747503e-01 3.55363339e-01 6.81738973e-01 9.14207101e-01
-1.14490412e-01 1.02980159e-01 -4.65719163e-01 1.19974472e-01
9.81446207e-02 1.55983114e+00 -1.75471497e+00 -2.02782646e-01
-1.14034787e-01 1.18609762e+00 1.99911162e-01 5.65990388e-01
-8.77027810e-01 -2.42078379e-01 1.44057572e-01 -2.87123978e-01
2.47672096e-01 -1.88052610e-01 -1.10410166e+00 -7.51572728e-01
-4.06761318e-01 -1.69765890e-01 5.85648417e-01 -1.03140688e+00
-1.05981660e+00 3.83865476e-01 1.67341709e-01 -8.17616522e-01
3.26928109e-01 -6.08351886e-01 -6.14126623e-01 3.86402965e-01
-1.45158064e+00 -1.58708405e+00 -4.17293578e-01 4.38765079e-01
6.00340605e-01 6.32813334e-01 3.30685705e-01 5.92993014e-02
-6.82045519e-01 8.73245113e-03 -2.08842888e-01 1.44650951e-01
-1.90285191e-01 -1.69747365e+00 -9.31140259e-02 8.71417701e-01
2.56122291e-01 8.19514021e-02 3.75025660e-01 -6.82748675e-01
-1.20848739e+00 -1.40170979e+00 5.20790480e-02 -1.87875211e-01
5.85235775e-01 -4.33478028e-01 -6.60230160e-01 8.25909376e-01
4.73885745e-01 5.66587448e-01 3.30608934e-01 -4.29338694e-01
6.14652336e-02 2.03061312e-01 -1.39395201e+00 2.31292114e-01
1.32588577e+00 -2.46940941e-01 -1.71422623e-02 4.92252976e-01
1.22082269e+00 -7.63852358e-01 -9.56877291e-01 4.79849190e-01
4.16083634e-01 -5.75943470e-01 9.11693215e-01 -3.07944179e-01
3.71977687e-01 -4.52632099e-01 -1.98701456e-01 -1.09890318e+00
6.14595190e-02 -5.05770683e-01 -6.05189577e-02 1.10407186e+00
7.03327119e-01 -2.96662182e-01 8.33618402e-01 7.16982663e-01
-3.61197948e-01 -1.04139698e+00 -9.81568933e-01 -6.71738803e-01
-6.92561315e-03 -4.43884581e-01 6.03844285e-01 9.00365889e-01
-4.99084711e-01 -1.92854062e-01 -1.05236381e-01 1.88422814e-01
4.78087485e-01 5.59718192e-01 4.92165446e-01 -1.19192648e+00
1.15214378e-01 -4.50587481e-01 -2.48686776e-01 -9.89976287e-01
4.53939624e-02 -7.78723657e-01 4.16717142e-01 -2.02742982e+00
-2.17596442e-02 -7.30214775e-01 -1.11128703e-01 9.60643709e-01
1.63984418e-01 7.63829410e-01 3.34543400e-02 1.35181263e-01
-8.21053088e-01 1.42742813e-01 1.62938607e+00 -4.25542623e-01
-5.36681116e-01 -1.99424163e-01 -5.13560534e-01 8.12737823e-01
8.08992624e-01 -4.67837512e-01 -4.81936723e-01 -5.63512564e-01
2.58914977e-01 -2.50580192e-01 4.20095503e-01 -7.67290056e-01
2.96976894e-01 -7.42125511e-01 1.59792662e-01 -7.66025245e-01
4.29106578e-02 -1.10021591e+00 2.42626697e-01 2.22286686e-01
-4.30076301e-01 -2.67303079e-01 1.96069464e-01 5.66781342e-01
-6.66961297e-02 -4.33394909e-01 8.80085826e-01 -4.83903378e-01
-1.12431240e+00 6.36935413e-01 -4.06593680e-02 4.11784463e-02
1.44687450e+00 -6.59825623e-01 -3.61066192e-01 1.71165019e-01
-8.83005798e-01 5.53306341e-01 5.66387653e-01 -8.38950947e-02
4.30527538e-01 -7.16569483e-01 -2.06118658e-01 -2.12420180e-01
-1.09584600e-01 7.77090847e-01 -1.52453169e-01 7.70794630e-01
-1.12961650e+00 -3.52608152e-02 -1.97949186e-01 -7.38884032e-01
-7.81125903e-01 4.85085160e-01 5.23125768e-01 -1.72280625e-01
-5.07419527e-01 1.13502622e+00 3.35022122e-01 -7.18401670e-01
2.63709009e-01 -5.34891784e-01 -3.42395633e-01 2.65327424e-01
-2.33006597e-01 2.11657360e-01 -2.43887007e-01 -4.16153610e-01
-3.70760649e-01 8.29455912e-01 5.39586067e-01 -2.76682764e-01
1.10465384e+00 -2.66296089e-01 -4.45736080e-01 3.96787852e-01
8.62724245e-01 -2.38689303e-01 -1.64897263e+00 6.53836057e-02
-1.16195709e-01 -3.90753299e-01 2.65097767e-01 -1.25157309e+00
-1.77668977e+00 6.00867748e-01 5.09700298e-01 5.37907481e-01
1.57107139e+00 2.57105500e-01 5.01397371e-01 9.66229513e-02
4.37287599e-01 -1.27208769e+00 -1.86717540e-01 2.25259453e-01
8.33006561e-01 -8.80887866e-01 4.41183120e-01 -1.14329016e+00
-5.86953878e-01 1.24996853e+00 9.48752284e-01 -2.07707450e-01
7.70382226e-01 5.32761216e-01 -3.04034278e-02 -5.69975972e-01
-2.98946679e-01 -6.95113122e-01 3.47351849e-01 8.49425495e-01
-1.73391938e-01 1.99583426e-01 -1.83651254e-01 1.36496201e-01
7.70827532e-02 -2.65855879e-01 9.89036858e-02 9.44755018e-01
-7.51537442e-01 -4.66541469e-01 -3.06501716e-01 6.65917754e-01
-2.04772905e-01 -1.63090155e-01 -1.00249183e+00 8.40327084e-01
7.71897137e-01 6.53253317e-01 5.68883777e-01 -2.06497848e-01
2.78420909e-03 -3.47177714e-01 4.10619944e-01 -6.73735738e-01
-9.27367032e-01 -7.56048933e-02 -7.96808898e-02 -5.34958601e-01
-9.63862777e-01 -3.47054541e-01 -1.67971027e+00 2.28861660e-01
-7.55351543e-01 -2.04038769e-01 7.14490652e-01 8.40716302e-01
-8.40648338e-02 1.67659253e-01 5.55059731e-01 -9.38647091e-01
7.10661173e-01 -4.15223509e-01 -6.97565377e-01 -2.04829291e-01
2.43950263e-03 -4.38445121e-01 -2.78680742e-01 4.93301511e-01] | [9.582569122314453, 0.3520773649215698] |
7e38f5c8-96ec-464f-b50c-037537599f83 | tracking-everything-everywhere-all-at-once | 2306.05422 | null | https://arxiv.org/abs/2306.05422v1 | https://arxiv.org/pdf/2306.05422v1.pdf | Tracking Everything Everywhere All at Once | We present a new test-time optimization method for estimating dense and long-range motion from a video sequence. Prior optical flow or particle video tracking algorithms typically operate within limited temporal windows, struggling to track through occlusions and maintain global consistency of estimated motion trajectories. We propose a complete and globally consistent motion representation, dubbed OmniMotion, that allows for accurate, full-length motion estimation of every pixel in a video. OmniMotion represents a video using a quasi-3D canonical volume and performs pixel-wise tracking via bijections between local and canonical space. This representation allows us to ensure global consistency, track through occlusions, and model any combination of camera and object motion. Extensive evaluations on the TAP-Vid benchmark and real-world footage show that our approach outperforms prior state-of-the-art methods by a large margin both quantitatively and qualitatively. See our project page for more results: http://omnimotion.github.io/ | ['Noah Snavely', 'Aleksander Holynski', 'Bharath Hariharan', 'Zhengqi Li', 'Ruojin Cai', 'Yen-Yu Chang', 'Qianqian Wang'] | 2023-06-08 | null | null | null | null | ['motion-estimation'] | ['computer-vision'] | [-4.59594578e-01 -8.16437304e-01 -3.27161640e-01 1.52492106e-01
-5.33323467e-01 -7.66468823e-01 6.01469040e-01 -3.85138005e-01
-4.47922021e-01 6.80175006e-01 9.15323049e-02 7.28937984e-02
2.39523500e-01 -3.46687883e-01 -7.93394327e-01 -5.66323459e-01
-4.74332124e-01 3.26152891e-01 6.73378587e-01 2.34927371e-01
8.12676176e-02 3.13389778e-01 -1.34370601e+00 -1.40851438e-01
4.07994837e-01 6.03789568e-01 2.31064230e-01 1.30016398e+00
3.05549324e-01 9.59897459e-01 -1.22331351e-01 -1.80866420e-01
4.26676095e-01 -3.70599300e-01 -6.82202280e-01 1.87609285e-01
1.23634911e+00 -6.01140857e-01 -9.58890140e-01 1.03634548e+00
3.24227512e-01 4.45755303e-01 2.83289820e-01 -1.37853909e+00
-2.82188505e-01 -7.01330453e-02 -6.31984413e-01 7.08023906e-01
7.62255192e-01 5.45295179e-01 6.50337160e-01 -1.06727397e+00
1.37615657e+00 1.30638719e+00 8.14593792e-01 5.44975221e-01
-1.22973239e+00 -4.51412052e-01 3.48180771e-01 2.66530424e-01
-1.47482502e+00 -6.30525112e-01 4.28009957e-01 -7.05505073e-01
8.26129973e-01 3.03414285e-01 1.06334531e+00 9.14625227e-01
5.32417059e-01 9.22718048e-01 4.78881449e-01 1.26297083e-02
2.41049621e-02 -6.07393503e-01 1.79563519e-02 1.09108901e+00
2.36643448e-01 3.22533011e-01 -8.58007193e-01 -2.32627615e-01
1.08907032e+00 -2.17911527e-01 -7.44424760e-01 -8.52098823e-01
-1.72666538e+00 4.08261806e-01 7.27059841e-02 -1.01328036e-02
-2.65295208e-01 7.43195772e-01 3.18642497e-01 -1.66272447e-02
3.99579167e-01 -7.62081891e-02 -9.60527211e-02 -6.26137376e-01
-1.21912456e+00 5.08938968e-01 7.08203137e-01 1.18440557e+00
6.57483041e-01 2.11101577e-01 -1.27465397e-01 1.83831885e-01
4.94237244e-01 7.55832136e-01 1.56730041e-01 -1.80884588e+00
2.88585782e-01 -1.22647703e-01 5.31223297e-01 -1.14146495e+00
-1.39154166e-01 -2.01674983e-01 -5.53047299e-01 3.33949506e-01
6.25094652e-01 -3.00083086e-02 -6.37979865e-01 1.68010855e+00
8.27250540e-01 1.25557888e+00 -3.11544180e-01 1.21115863e+00
5.67822278e-01 7.39309311e-01 -7.33598396e-02 -4.36611712e-01
1.03175414e+00 -1.18282866e+00 -7.66236544e-01 -2.29213223e-01
4.17393178e-01 -7.20191777e-01 5.74911714e-01 2.79888451e-01
-1.38240099e+00 -6.91383064e-01 -9.34544504e-01 -3.38691883e-02
2.05949113e-01 -2.84154296e-01 3.95699412e-01 3.76784712e-01
-1.14881420e+00 7.40384877e-01 -1.52857828e+00 -2.23681569e-01
2.78372020e-01 8.96703899e-02 -5.09693742e-01 -2.23945677e-01
-7.87836611e-01 5.26043713e-01 -2.97553767e-03 9.56501588e-02
-1.15152276e+00 -9.98822927e-01 -1.12227893e+00 -3.63875777e-01
1.81988016e-01 -1.05867553e+00 1.01130521e+00 -3.40220094e-01
-1.12409794e+00 6.58164501e-01 -6.83555126e-01 -4.75322336e-01
9.28899229e-01 -6.38309002e-01 -1.52949318e-01 3.50361407e-01
1.50578439e-01 7.82038748e-01 8.84216130e-01 -1.08716774e+00
-6.41671956e-01 -4.60372642e-02 -1.85378298e-01 2.99903244e-01
2.29241103e-02 -8.17472301e-03 -1.12780523e+00 -6.43430233e-01
8.99130031e-02 -1.26130891e+00 -1.94146425e-01 4.24985111e-01
-1.34142146e-01 1.99755207e-01 1.21944392e+00 -6.54658675e-01
1.46236134e+00 -1.80267239e+00 4.08018142e-01 -1.66150555e-01
4.29910928e-01 1.82037905e-01 -9.68629401e-03 -7.65762553e-02
2.97999501e-01 -2.14650407e-01 -1.14703029e-01 -7.12909460e-01
-2.19390541e-01 8.42368156e-02 -1.14431575e-01 1.17171812e+00
-2.11679041e-01 9.81796801e-01 -1.50751436e+00 -7.23623991e-01
7.78004289e-01 7.93468475e-01 -6.16288364e-01 7.28370845e-02
-1.51405483e-01 8.48505855e-01 -3.34120661e-01 8.03554773e-01
7.60814726e-01 -3.65143508e-01 -5.48524840e-04 -2.70289689e-01
-1.87754646e-01 -1.65806100e-01 -1.50167310e+00 2.18585253e+00
1.35940254e-01 1.04108047e+00 2.35371068e-01 -3.48366320e-01
2.18720049e-01 2.30616868e-01 1.03941667e+00 -2.83910573e-01
3.68619189e-02 -6.57350048e-02 -4.86376852e-01 -3.32076430e-01
7.80611575e-01 5.63924968e-01 3.78135324e-01 2.31728986e-01
-4.05028313e-02 1.27963901e-01 6.73525453e-01 3.71261477e-01
1.22595978e+00 8.41606915e-01 1.92423970e-01 -3.32031131e-01
5.85911572e-01 8.94712508e-02 1.00032604e+00 7.70506680e-01
-7.99277365e-01 9.10603762e-01 -1.44561440e-01 -5.61648607e-01
-1.07771993e+00 -1.23955357e+00 -5.67452423e-02 4.90994483e-01
6.89774454e-01 -7.33389676e-01 -6.01896286e-01 -3.66201818e-01
6.91257119e-02 1.19451798e-01 -4.30974603e-01 3.28847378e-01
-1.09932172e+00 -5.38513541e-01 1.68463826e-01 5.02740920e-01
4.28099632e-01 -6.52646184e-01 -9.24600601e-01 5.43667376e-01
-5.72513163e-01 -1.61152446e+00 -1.02786422e+00 -6.38218462e-01
-9.68231738e-01 -1.18753743e+00 -7.69108176e-01 -4.77517486e-01
3.87669295e-01 6.92742467e-01 1.40897417e+00 3.05926889e-01
-5.26466846e-01 6.93977058e-01 -8.91340449e-02 3.77309710e-01
-2.59539485e-01 -4.86802638e-01 2.76399791e-01 -3.06430273e-02
1.27698481e-01 -4.77646708e-01 -1.07089436e+00 6.02224112e-01
-5.78345954e-01 -3.08351796e-02 -3.57684702e-01 5.35616934e-01
8.16795528e-01 -3.68587404e-01 -3.29635113e-01 -2.81656235e-01
-8.67675617e-02 -4.10129964e-01 -1.00532532e+00 1.25096604e-01
-1.80223048e-01 -1.93407714e-01 3.78411375e-02 -4.94658500e-01
-7.78183460e-01 2.88123578e-01 1.90118060e-01 -1.13421476e+00
2.08103567e-01 -9.19310749e-02 1.86555326e-01 -4.36202317e-01
4.17403728e-01 2.19555199e-01 -8.90043750e-02 -3.41210738e-02
4.71149057e-01 -3.01811486e-01 1.00020397e+00 -5.56070805e-01
1.02895451e+00 1.22631967e+00 1.32674605e-01 -8.14854443e-01
-3.94243598e-01 -9.59978044e-01 -8.32445443e-01 -7.75856018e-01
1.12008107e+00 -1.08572614e+00 -8.89248371e-01 6.15901113e-01
-1.24457753e+00 -5.11800230e-01 -2.19185539e-02 9.40303981e-01
-8.55511308e-01 7.36377478e-01 -7.85049260e-01 -6.14688337e-01
-9.50997770e-02 -1.22045171e+00 1.17261076e+00 1.28107294e-01
-3.87792200e-01 -1.35625792e+00 6.03282571e-01 2.83873994e-02
9.57374275e-02 5.88181019e-01 -2.08892941e-01 4.21924740e-01
-1.28166628e+00 -3.42066586e-02 7.09796101e-02 -2.52030473e-02
-2.91410033e-02 3.35355043e-01 -6.38874292e-01 -6.76304638e-01
-1.84942171e-01 1.27270073e-01 8.63036215e-01 8.35038900e-01
6.64136827e-01 9.09423158e-02 -6.22068524e-01 1.21141720e+00
1.53482759e+00 -2.04970744e-02 6.62754655e-01 4.58745599e-01
1.12964451e+00 8.08500499e-02 9.26507771e-01 4.77725148e-01
3.66453260e-01 9.31355417e-01 4.85672981e-01 3.08728606e-01
-4.86213446e-01 -1.85130775e-01 5.44230223e-01 7.52967298e-01
-3.22703362e-01 -4.67965394e-01 -6.74203813e-01 6.77358270e-01
-2.22711325e+00 -1.40074527e+00 -4.18091625e-01 2.32834053e+00
4.03910249e-01 -8.12206715e-02 1.73591495e-01 -3.49157780e-01
8.03370595e-01 5.14836133e-01 -5.66448689e-01 4.21724617e-01
-2.74022371e-01 -3.37357700e-01 7.32195973e-01 9.84429419e-01
-1.36425900e+00 1.09271920e+00 6.54397821e+00 5.04791200e-01
-7.24590480e-01 2.80348122e-01 2.64887750e-01 -4.30344611e-01
-1.73893142e-02 1.49711356e-01 -1.01710665e+00 4.95353848e-01
7.14612901e-01 -1.02380492e-01 3.47268611e-01 4.43672091e-01
4.44845200e-01 -2.42356643e-01 -1.09820545e+00 1.20657980e+00
3.33271036e-03 -1.80664861e+00 -2.36269072e-01 1.85264766e-01
1.07916248e+00 3.37603956e-01 -1.13196030e-01 -2.43176326e-01
3.20109189e-01 -5.20579398e-01 9.87186134e-01 7.97154903e-01
4.90558714e-01 -2.99345911e-01 1.81126252e-01 3.68743762e-02
-1.83658993e+00 2.84106076e-01 -3.34389538e-01 1.47801831e-01
9.04366732e-01 3.69408965e-01 -6.17929474e-02 3.84653181e-01
8.55769336e-01 1.32569277e+00 -2.84405321e-01 1.44903362e+00
2.43686289e-01 4.24232244e-01 -4.20630306e-01 5.47835588e-01
2.42189422e-01 -5.46424508e-01 1.23083925e+00 1.27804792e+00
4.80226964e-01 7.61807114e-02 3.84618074e-01 6.34376466e-01
1.03738725e-01 -3.17728996e-01 -4.86864388e-01 3.22384477e-01
5.04841387e-01 1.04509473e+00 -7.66917109e-01 -4.87723440e-01
-5.16890109e-01 1.07438838e+00 1.08628377e-01 6.24342978e-01
-1.06292033e+00 2.48180807e-01 1.23212695e+00 1.48043618e-01
4.51162577e-01 -8.45648825e-01 2.17885479e-01 -1.64689600e+00
1.47110611e-01 -4.63362962e-01 3.30603331e-01 -7.94863284e-01
-7.55262196e-01 3.59752893e-01 5.88447750e-02 -1.64648759e+00
-3.68759453e-01 -3.89762342e-01 -5.36510110e-01 5.38633943e-01
-1.37761474e+00 -7.75335073e-01 -6.81437731e-01 8.22817087e-01
6.90035164e-01 1.42209038e-01 2.38909796e-01 4.59919661e-01
-3.53715628e-01 1.02693945e-01 2.38443226e-01 2.51815200e-01
7.63686061e-01 -9.92664158e-01 7.67305732e-01 1.35422337e+00
3.57090980e-01 4.77174520e-01 9.24693763e-01 -8.90571773e-01
-1.78881800e+00 -1.17564166e+00 4.99831468e-01 -1.00167704e+00
6.84983611e-01 -2.29486167e-01 -9.28452253e-01 9.45286870e-01
1.98590770e-01 7.87732303e-01 2.15330701e-02 -5.55148721e-01
-8.67006332e-02 1.97540089e-01 -7.82308936e-01 5.43699086e-01
1.54400325e+00 -3.48335177e-01 -5.69108389e-02 3.40777040e-01
5.94390094e-01 -9.24596429e-01 -8.72528136e-01 1.97990686e-01
7.78770387e-01 -1.11777878e+00 1.37046540e+00 -2.63082504e-01
1.04061849e-01 -7.63043642e-01 -1.53644294e-01 -8.76783669e-01
-3.53218228e-01 -1.22145391e+00 -9.19583917e-01 7.38234162e-01
-3.38842690e-01 -2.87580609e-01 1.06815958e+00 4.85419929e-01
-1.45936571e-02 -4.52419519e-01 -9.23138559e-01 -1.07753360e+00
-2.48821959e-01 -6.34442568e-01 1.87890545e-01 8.81994605e-01
-3.59299511e-01 -3.68332416e-01 -7.12600529e-01 3.68341297e-01
1.33438444e+00 -5.58869466e-02 1.01587784e+00 -7.05777586e-01
-1.99463546e-01 -2.77150333e-01 -8.71143937e-01 -1.62524831e+00
2.98792392e-01 -5.64575255e-01 1.29865706e-01 -1.35391188e+00
1.97254926e-01 -1.02420539e-01 9.60256606e-02 -2.40152478e-01
-3.01519781e-01 5.91541529e-01 4.31790322e-01 4.77414668e-01
-1.07580602e+00 4.74019140e-01 1.42634952e+00 -3.61653827e-02
-1.97245523e-01 -3.38541031e-01 2.10682794e-01 7.72246182e-01
3.36454839e-01 -4.61554915e-01 -3.08758289e-01 -6.32891417e-01
-1.21071510e-01 3.43270868e-01 7.91028917e-01 -1.25727618e+00
5.15678942e-01 -2.82408059e-01 3.92747104e-01 -7.87933707e-01
5.90127468e-01 -5.02227843e-01 5.66728115e-01 5.40938854e-01
1.14647649e-01 4.97900873e-01 2.16447175e-01 8.66196454e-01
-1.25396967e-01 1.49820983e-01 8.36653113e-01 -7.12024420e-02
-1.27380586e+00 8.98886859e-01 -2.18272969e-01 4.73903507e-01
1.03874969e+00 -4.93441612e-01 -3.48941147e-01 -4.98979270e-01
-9.06931460e-01 4.83616710e-01 9.30659950e-01 5.31260967e-01
7.04392314e-01 -1.57851231e+00 -8.33793998e-01 -1.23638511e-02
-1.47051677e-01 -2.68140823e-01 3.62746954e-01 1.04005527e+00
-1.07122219e+00 3.10071290e-01 -1.20313652e-01 -1.14741933e+00
-1.45526874e+00 4.06974882e-01 5.23866117e-01 1.02274828e-02
-1.00015295e+00 8.82568300e-01 2.96960562e-01 3.20778117e-02
1.06967218e-01 -1.88843533e-01 2.09275603e-01 -3.81138921e-01
8.99462640e-01 7.60709703e-01 -3.65758449e-01 -1.14617300e+00
-5.68270028e-01 1.10970294e+00 3.48000407e-01 -2.77606279e-01
7.51763582e-01 -6.26806080e-01 7.01700225e-02 2.78393149e-01
1.45472670e+00 -2.68192198e-02 -2.06927299e+00 -1.62185580e-01
-3.10229033e-01 -1.23333383e+00 3.34687792e-02 1.01675332e-01
-1.12549710e+00 4.58042920e-01 6.28919244e-01 -1.48336560e-01
6.26085520e-01 -1.71707958e-01 9.30033565e-01 4.77230363e-02
5.19589186e-01 -7.40120113e-01 -3.79158519e-02 6.18992746e-01
5.20662069e-01 -1.15014839e+00 2.87094116e-01 -5.02779007e-01
-2.76151538e-01 9.95304286e-01 6.49429321e-01 -4.08929557e-01
6.52648628e-01 2.90975869e-01 -1.38266860e-02 -1.04104308e-02
-8.10569227e-01 1.48876496e-02 4.19899017e-01 6.76012278e-01
2.90011644e-01 -3.47141623e-01 6.85111210e-02 -6.23016715e-01
1.44339129e-01 9.52285975e-02 5.92720211e-01 1.08288670e+00
-2.98410088e-01 -7.84478843e-01 -6.16326630e-01 -2.97307950e-02
-3.34980100e-01 1.03555813e-01 3.44553381e-01 9.06393886e-01
-1.81063846e-01 7.92092562e-01 3.07713777e-01 -1.25144064e-01
-4.54139858e-02 -3.89097720e-01 7.16288686e-01 -9.88732502e-02
-1.16839796e-01 3.38289618e-01 -1.04656462e-02 -1.45552731e+00
-8.89564753e-01 -1.10114253e+00 -1.25766170e+00 -7.64995039e-01
3.31595317e-02 -2.54323389e-02 3.70767295e-01 6.51153862e-01
3.65337372e-01 2.93079108e-01 1.94287837e-01 -1.42554319e+00
-5.53175546e-02 -3.34954351e-01 -3.62855673e-01 5.73831081e-01
7.37700641e-01 -8.09501946e-01 -4.18164164e-01 5.18284738e-01] | [8.524796485900879, -1.807100534439087] |
0365ec85-a2ed-4974-942d-093c6c1eed6a | social-adaptive-module-for-weakly-supervised | 2007.09470 | null | https://arxiv.org/abs/2007.09470v1 | https://arxiv.org/pdf/2007.09470v1.pdf | Social Adaptive Module for Weakly-supervised Group Activity Recognition | This paper presents a new task named weakly-supervised group activity recognition (GAR) which differs from conventional GAR tasks in that only video-level labels are available, yet the important persons within each frame are not provided even in the training data. This eases us to collect and annotate a large-scale NBA dataset and thus raise new challenges to GAR. To mine useful information from weak supervision, we present a key insight that key instances are likely to be related to each other, and thus design a social adaptive module (SAM) to reason about key persons and frames from noisy data. Experiments show significant improvement on the NBA dataset as well as the popular volleyball dataset. In particular, our model trained on video-level annotation achieves comparable accuracy to prior algorithms which required strong labels. | ['Xiangbo Shu', 'Jinhui Tang', 'Rui Yan', 'Qi Tian', 'Lingxi Xie'] | 2020-07-18 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/520_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123530205.pdf | eccv-2020-8 | ['group-activity-recognition'] | ['computer-vision'] | [ 3.15642983e-01 4.68149222e-02 -4.65403438e-01 -5.70657015e-01
-8.85239899e-01 -4.19316262e-01 5.76183617e-01 6.09000884e-02
-6.24947190e-01 9.61232841e-01 7.59545743e-01 2.77459145e-01
-8.40363558e-03 -3.97873402e-01 -8.32639694e-01 -7.12982357e-01
-2.56719828e-01 3.42910022e-01 3.34691525e-01 4.11984622e-02
-3.30566078e-01 5.24424314e-02 -1.46601546e+00 5.20631135e-01
2.49232754e-01 8.79051805e-01 1.16494775e-01 6.35139823e-01
3.71783376e-01 1.33765459e+00 -7.32132375e-01 -3.34030151e-01
1.75088212e-01 -7.57664144e-01 -1.03611314e+00 5.82325220e-01
6.85439587e-01 -2.16558307e-01 -4.11046475e-01 5.98156452e-01
3.81142855e-01 4.11816329e-01 4.50845152e-01 -1.35130799e+00
-8.52287635e-02 7.08904743e-01 -5.91699839e-01 4.19224650e-01
7.69841433e-01 -2.85929795e-02 1.11162519e+00 -7.59312212e-01
8.84227216e-01 1.15255058e+00 8.28087747e-01 6.89790726e-01
-9.04127061e-01 -4.96193588e-01 6.04108691e-01 3.64466608e-01
-1.30261946e+00 -6.72491729e-01 5.48613906e-01 -3.32510144e-01
5.85995436e-01 2.99546063e-01 7.43024409e-01 1.57931447e+00
-6.48021936e-01 1.20040607e+00 8.68138671e-01 -3.38919848e-01
3.12253702e-02 -3.87849838e-01 3.41133833e-01 8.11990321e-01
2.76561707e-01 -6.18200779e-01 -1.10842407e+00 -9.98906642e-02
6.16772473e-01 1.05643123e-01 -1.52194530e-01 -2.83975482e-01
-1.73524308e+00 4.37323421e-01 4.88690399e-02 2.57421464e-01
-3.17218035e-01 3.38162452e-01 4.04481411e-01 1.67610615e-01
4.44966525e-01 3.57583076e-01 -3.08483392e-01 -5.54964840e-01
-7.58415639e-01 3.24543029e-01 7.20320821e-01 7.64141381e-01
8.10480535e-01 -2.38871396e-01 -4.17135447e-01 7.57669091e-01
2.11917564e-01 2.52623618e-01 1.38510317e-01 -1.45300233e+00
5.67483068e-01 6.02803767e-01 2.41310790e-01 -8.75797510e-01
-4.78362173e-01 -3.17220718e-01 -7.15398192e-01 -9.71083194e-02
9.05524194e-01 -1.52294412e-01 -7.71470249e-01 1.79493546e+00
2.95806170e-01 6.36265934e-01 -1.39898047e-01 9.28181410e-01
7.82372594e-01 4.41859275e-01 8.38452354e-02 -3.70058298e-01
1.26382899e+00 -1.33801246e+00 -8.55464578e-01 -2.99574643e-01
6.32259727e-01 -4.05670822e-01 1.04428434e+00 3.54825288e-01
-1.01025236e+00 -8.26690316e-01 -8.53404820e-01 1.05393544e-01
-2.68014167e-02 2.08689913e-01 7.61814117e-01 4.69452769e-01
-8.51842761e-01 4.12793845e-01 -9.99486685e-01 -6.11052454e-01
9.50559616e-01 2.18014374e-01 -6.14836216e-01 -3.29492629e-01
-8.03949118e-01 5.31683326e-01 2.60286599e-01 -3.73544171e-04
-1.18572164e+00 -2.43281603e-01 -1.04297912e+00 -3.14753026e-01
9.36315715e-01 -6.10017955e-01 1.14800668e+00 -1.14592600e+00
-1.06387663e+00 9.38023746e-01 -4.15056229e-01 -6.33645833e-01
5.39676428e-01 -5.04342794e-01 -3.59620571e-01 3.49841565e-01
4.65223134e-01 5.71939409e-01 7.13282824e-01 -1.18132162e+00
-8.69512498e-01 -1.51382551e-01 2.33700767e-01 2.23039806e-01
-4.16942626e-01 4.26510632e-01 -8.39862585e-01 -1.04618645e+00
-1.37163177e-01 -9.72189486e-01 -2.25975230e-01 -1.89656079e-01
-1.71799079e-01 -5.25281370e-01 8.81401896e-01 -6.81577146e-01
1.29765892e+00 -2.11275578e+00 2.47507125e-01 1.53545022e-01
3.04263681e-01 -1.82928871e-02 7.13127851e-02 3.07275087e-01
-2.17486303e-02 -2.19203308e-01 2.66557443e-03 -5.80705523e-01
-2.47563362e-01 6.77262843e-01 7.10321665e-02 6.02446318e-01
2.92444285e-02 9.04156685e-01 -1.33588922e+00 -7.34890223e-01
-1.37895588e-02 3.90322916e-02 -4.48112905e-01 2.39935130e-01
-1.04766481e-01 7.55397797e-01 -4.29297209e-01 1.02400041e+00
6.47738501e-02 -6.30480647e-01 1.87256977e-01 -1.16399378e-01
3.05173844e-01 7.77720436e-02 -1.31585062e+00 1.96189392e+00
1.17147528e-01 7.04701066e-01 6.95590004e-02 -1.31512070e+00
3.41283858e-01 4.31446999e-01 9.36344802e-01 -1.51138633e-01
2.28119064e-02 -1.97188735e-01 -2.24305078e-01 -6.93880856e-01
3.14717650e-01 3.44600499e-01 -3.44715923e-01 5.99377453e-01
3.89491260e-01 6.27654910e-01 5.85995793e-01 6.01709902e-01
1.37620974e+00 4.10119534e-01 2.54978806e-01 -1.92396678e-02
5.17832935e-01 -2.86233313e-02 9.97513175e-01 1.04142785e+00
-4.78800863e-01 7.56768286e-01 4.13071156e-01 -6.51155055e-01
-7.77446508e-01 -7.15761960e-01 3.60964596e-01 1.58849788e+00
4.07590985e-01 -1.11463952e+00 -7.24167347e-01 -1.17975974e+00
-4.29576129e-01 -6.75350521e-03 -7.39337921e-01 1.59932166e-01
-8.04778457e-01 -6.31134570e-01 4.62715000e-01 7.48185217e-01
5.42922378e-01 -9.37067628e-01 -2.37331182e-01 3.72854657e-02
-7.77681231e-01 -1.52323723e+00 -6.39658272e-01 -1.43058404e-01
-4.36554015e-01 -1.44429147e+00 -6.22802377e-01 -7.23410308e-01
7.90269077e-01 4.61327940e-01 1.29828274e+00 1.86553270e-01
-2.15285309e-02 5.89864194e-01 -6.87526703e-01 -3.28625977e-01
-1.41536862e-01 8.34189653e-02 3.44769955e-01 4.72496271e-01
3.78595918e-01 -3.16853195e-01 -3.71671408e-01 6.59215271e-01
-2.87520856e-01 2.02796325e-01 3.68596405e-01 6.39409482e-01
5.80129087e-01 1.74506351e-01 6.69508457e-01 -8.64574552e-01
1.41638845e-01 -4.23993021e-01 3.49475443e-02 2.62202919e-01
-1.16835862e-01 -4.64243352e-01 2.56504208e-01 -4.57726508e-01
-1.09251964e+00 3.20819080e-01 8.25822279e-02 -3.92859757e-01
-3.62238616e-01 2.88674772e-01 -3.18035811e-01 1.95305809e-01
7.05303371e-01 -1.04017258e-01 -2.56914020e-01 -5.94534814e-01
-3.85938361e-02 4.63765055e-01 7.68731296e-01 -7.18162775e-01
7.26371646e-01 7.80880034e-01 -3.80762778e-02 -6.85861647e-01
-1.59229612e+00 -9.74064410e-01 -8.07681143e-01 -7.41625786e-01
9.75496829e-01 -1.43099296e+00 -5.69466114e-01 4.80100513e-01
-7.76773512e-01 -3.72571111e-01 -6.36947572e-01 6.60022855e-01
-6.60776854e-01 5.03861487e-01 -7.03261197e-01 -8.48180473e-01
1.42103627e-01 -7.51351595e-01 1.10302532e+00 9.24089849e-02
-6.91885829e-01 -6.07280254e-01 -1.45771652e-01 1.14260352e+00
-3.46441269e-01 4.46205199e-01 -5.08604459e-02 -6.18354559e-01
-5.90971231e-01 -3.00446600e-01 2.46144030e-02 3.60933512e-01
1.93214968e-01 -2.96203285e-01 -9.20053422e-01 -4.99706715e-01
-4.45020020e-01 -6.60000145e-01 1.03172994e+00 2.47565821e-01
1.33700073e+00 -1.03174917e-01 -4.45988804e-01 3.58875215e-01
5.42969584e-01 -2.88059473e-01 3.98881853e-01 2.52769172e-01
1.05382597e+00 3.90225232e-01 9.69377100e-01 4.35364008e-01
5.39678216e-01 7.21417427e-01 1.18461251e-01 -1.20354697e-01
-4.01866466e-01 -5.06500125e-01 6.66378498e-01 6.99810982e-01
-8.67105305e-01 -1.87031522e-01 -7.00704873e-01 6.61103725e-01
-2.50735831e+00 -1.32732797e+00 -2.75413424e-01 1.87047184e+00
7.24354982e-01 8.80456418e-02 6.44445598e-01 2.49738753e-01
7.23868072e-01 4.09184188e-01 -2.07333699e-01 6.53363705e-01
-4.01032627e-01 -1.43132821e-01 4.56138939e-01 1.59638092e-01
-1.61659193e+00 7.69017279e-01 7.16449642e+00 9.64320779e-01
-1.58030048e-01 4.31796193e-01 6.50038183e-01 -3.99145484e-01
4.05069381e-01 -7.14029968e-02 -8.62232089e-01 5.66960037e-01
8.81341219e-01 1.01231247e-01 2.70986527e-01 8.26913476e-01
3.76241088e-01 -2.67504811e-01 -1.30042219e+00 1.25308204e+00
3.95953417e-01 -1.36091471e+00 -9.80712622e-02 -3.62565788e-03
9.82380390e-01 -4.41685542e-02 -5.49790800e-01 2.26144299e-01
4.33881670e-01 -8.39824975e-01 8.25557351e-01 5.93513727e-01
5.46091080e-01 -5.55041492e-01 8.32952082e-01 4.72565413e-01
-1.26899099e+00 -1.80811509e-01 -1.31463706e-01 -4.25170898e-01
3.54321301e-01 3.74277234e-01 -4.83785450e-01 5.03514946e-01
1.17948866e+00 1.49235082e+00 -7.27731109e-01 8.24670851e-01
-3.16461653e-01 1.03639817e+00 -3.56515795e-01 3.55194926e-01
8.44391733e-02 3.09960674e-02 4.59709585e-01 1.17506015e+00
2.03282237e-01 1.47961229e-01 8.02208185e-01 1.56349882e-01
-5.01930833e-01 -1.11235246e-01 -3.38488907e-01 -1.94931775e-03
1.11256711e-01 1.19966435e+00 -1.01227248e+00 -6.78606808e-01
-8.10903907e-01 9.51772451e-01 1.81252360e-01 3.22986156e-01
-9.22550201e-01 2.40209311e-01 5.45962870e-01 3.53797287e-01
9.86861587e-02 -1.84524745e-01 2.18145952e-01 -1.65217018e+00
7.38571808e-02 -9.98523593e-01 1.13654554e+00 -8.24457943e-01
-1.26066768e+00 1.22915208e-01 7.97670484e-02 -1.35935414e+00
-2.00631633e-01 -2.32514873e-01 -3.98142278e-01 1.97534442e-01
-1.24677384e+00 -1.60823333e+00 -5.48843503e-01 9.87378240e-01
8.57525945e-01 -2.98275858e-01 5.51789939e-01 5.75296223e-01
-6.87199116e-01 3.11493307e-01 -3.96393478e-01 6.18590176e-01
1.08081472e+00 -1.09385693e+00 1.00101836e-01 1.21569335e+00
6.34920895e-01 2.90988505e-01 5.97591937e-01 -7.35725164e-01
-1.09097421e+00 -1.01165771e+00 8.27829421e-01 -9.82145250e-01
5.26438773e-01 -5.51870584e-01 -6.41956508e-01 1.06394529e+00
8.83979201e-02 4.18054909e-01 9.13394511e-01 3.82601589e-01
-1.75210938e-01 -3.93739939e-02 -6.87177181e-01 3.14192683e-01
1.80229354e+00 -5.20399868e-01 -7.52548099e-01 4.78324950e-01
4.73401427e-01 -2.78227895e-01 -6.61183476e-01 3.93953443e-01
2.99748093e-01 -8.13410759e-01 1.11756861e+00 -7.98289061e-01
1.25937372e-01 -6.36903942e-01 -5.16424654e-04 -9.31955874e-01
-2.44317994e-01 -7.39960134e-01 -8.25159490e-01 1.25124741e+00
1.55209154e-01 -4.25866246e-02 1.14653444e+00 6.29774034e-01
-7.22874254e-02 -1.97775945e-01 -7.38641500e-01 -7.87712455e-01
-8.97359431e-01 -5.62609673e-01 2.94901222e-01 1.36712849e+00
1.09015293e-01 3.00316840e-01 -9.56570089e-01 1.14983805e-01
7.73587048e-01 5.79824857e-02 1.03041399e+00 -1.23370647e+00
-3.69239390e-01 4.88698743e-02 -5.82169175e-01 -1.08929002e+00
2.80970365e-01 -6.33732677e-01 6.17187377e-03 -1.44703424e+00
4.44344193e-01 -2.32923374e-01 -5.19845307e-01 8.56370509e-01
-3.85860711e-01 7.08849967e-01 1.08987123e-01 4.19333845e-01
-1.66799188e+00 2.36069009e-01 9.36412752e-01 -1.86053917e-01
-7.62596494e-03 2.60933548e-01 -7.27253854e-01 1.14181447e+00
4.14127320e-01 -3.85737509e-01 -2.49565795e-01 -3.12129676e-01
3.07747722e-01 -2.03512862e-01 4.82603282e-01 -1.22633040e+00
3.49700093e-01 -2.60092229e-01 5.73879361e-01 -5.25940418e-01
3.56274515e-01 -7.45708466e-01 1.28105953e-01 7.74833485e-02
-4.48460370e-01 -3.33395571e-01 -4.49178576e-01 9.91511762e-01
-3.77925724e-01 2.45180186e-02 4.13694382e-01 -2.54175276e-01
-1.08421719e+00 4.71352130e-01 -5.05529642e-01 2.01271549e-01
1.12457836e+00 -3.75773937e-01 -5.04476465e-02 -7.72214890e-01
-1.09027183e+00 4.56021696e-01 5.05303144e-01 5.33353746e-01
1.53670922e-01 -1.51755536e+00 -7.50308871e-01 -8.37599635e-02
4.38613147e-01 1.68117374e-01 3.27841520e-01 8.34088802e-01
-9.98090282e-02 -8.59719813e-02 -1.73781645e-02 -6.43152654e-01
-1.64875066e+00 4.29239601e-01 -1.43756662e-02 -1.92698717e-01
-6.85904205e-01 8.98208141e-01 -3.37188207e-02 9.78201404e-02
5.25702059e-01 -1.21052407e-01 -3.98782104e-01 3.18298042e-01
9.17139292e-01 4.60410267e-01 -2.56530046e-01 -1.00419724e+00
-3.85162354e-01 2.92119980e-01 2.46782318e-01 2.15580672e-01
1.64312029e+00 -4.18598205e-01 2.70601455e-03 5.85986257e-01
8.04039598e-01 7.51091214e-03 -1.63700342e+00 -6.39805973e-01
2.00589180e-01 -7.85645902e-01 -2.20956907e-01 -4.76994395e-01
-9.90454912e-01 3.26488256e-01 1.95300773e-01 9.03111547e-02
9.30558085e-01 4.80823696e-01 6.79718852e-01 6.89102769e-01
4.73129272e-01 -1.49340653e+00 4.47626710e-01 2.56492764e-01
4.26130176e-01 -1.46594048e+00 3.01775455e-01 -4.25571233e-01
-7.62381613e-01 6.85096502e-01 7.36427426e-01 1.34327993e-01
3.54754090e-01 1.94013536e-01 -5.61794965e-03 -2.38218039e-01
-6.38322771e-01 -5.15362501e-01 1.70060396e-01 6.79281771e-01
1.05664685e-01 -2.05021948e-01 -6.14284687e-02 9.17258203e-01
2.42258683e-01 2.75198758e-01 4.78691608e-01 1.20419610e+00
-3.33580822e-01 -1.09615028e+00 -3.85347009e-01 5.17172992e-01
-6.65315747e-01 5.06360829e-01 -2.98452556e-01 4.62269515e-01
3.68948370e-01 1.00073779e+00 -1.06913842e-01 -2.15273410e-01
2.38311723e-01 1.45913303e-01 3.78788382e-01 -6.97674394e-01
-4.11473066e-01 1.33786142e-01 7.50832975e-01 -8.76309991e-01
-1.47285843e+00 -9.62286532e-01 -1.02684879e+00 -2.27166757e-01
-9.09463242e-02 2.03515753e-01 -2.66843252e-02 1.18588066e+00
1.26257658e-01 4.57596213e-01 3.68055016e-01 -1.06153381e+00
3.12890440e-01 -8.21796000e-01 -5.56531072e-01 9.85859632e-01
2.58159637e-01 -6.34526908e-01 -1.86171681e-01 7.91364193e-01] | [8.357389450073242, 0.5873637199401855] |
b918e033-2645-45a9-9ea1-0fc238561e9b | unsupervised-3d-pose-estimation-for | 2109.09166 | null | https://arxiv.org/abs/2109.09166v1 | https://arxiv.org/pdf/2109.09166v1.pdf | Unsupervised 3D Pose Estimation for Hierarchical Dance Video Recognition | Dance experts often view dance as a hierarchy of information, spanning low-level (raw images, image sequences), mid-levels (human poses and bodypart movements), and high-level (dance genre). We propose a Hierarchical Dance Video Recognition framework (HDVR). HDVR estimates 2D pose sequences, tracks dancers, and then simultaneously estimates corresponding 3D poses and 3D-to-2D imaging parameters, without requiring ground truth for 3D poses. Unlike most methods that work on a single person, our tracking works on multiple dancers, under occlusions. From the estimated 3D pose sequence, HDVR extracts body part movements, and therefrom dance genre. The resulting hierarchical dance representation is explainable to experts. To overcome noise and interframe correspondence ambiguities, we enforce spatial and temporal motion smoothness and photometric continuity over time. We use an LSTM network to extract 3D movement subsequences from which we recognize the dance genre. For experiments, we have identified 154 movement types, of 16 body parts, and assembled a new University of Illinois Dance (UID) Dataset, containing 1143 video clips of 9 genres covering 30 hours, annotated with movement and genre labels. Our experimental results demonstrate that our algorithms outperform the state-of-the-art 3D pose estimation methods, which also enhances our dance recognition performance. | ['Narendra Ahuja', 'Xiaodan Hu'] | 2021-09-19 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Hu_Unsupervised_3D_Pose_Estimation_for_Hierarchical_Dance_Video_Recognition_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Hu_Unsupervised_3D_Pose_Estimation_for_Hierarchical_Dance_Video_Recognition_ICCV_2021_paper.pdf | iccv-2021-1 | ['3d-pose-estimation', 'unsupervised-3d-human-pose-estimation', 'weakly-supervised-3d-human-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-2.66164225e-02 -4.13940400e-01 -4.44938570e-01 -4.10206094e-02
-8.02107990e-01 -6.52853131e-01 2.18060404e-01 -7.72566497e-01
-3.86846155e-01 2.78490424e-01 3.41502160e-01 4.73668128e-01
-3.88706289e-03 -5.03717661e-01 -8.45123827e-01 -5.80192864e-01
-2.04015851e-01 7.38528132e-01 2.94278502e-01 -7.17042461e-02
8.02833587e-02 4.88376647e-01 -1.56107771e+00 2.72855550e-01
2.99841106e-01 1.06306648e+00 -3.99620198e-02 1.01928473e+00
7.24031627e-02 7.28905857e-01 -8.28356683e-01 -2.04776719e-01
3.84539872e-01 -6.13713145e-01 -6.25063658e-01 7.14977801e-01
9.21046376e-01 -4.76897985e-01 -6.44719005e-01 8.29459548e-01
4.67385352e-01 4.41799581e-01 4.04616773e-01 -1.10644317e+00
-6.40801847e-01 1.67592332e-01 -8.39767635e-01 1.48881823e-01
8.38087857e-01 1.35099784e-01 7.60748446e-01 -1.00665939e+00
9.49291229e-01 1.46490633e+00 1.04769850e+00 7.95347452e-01
-1.07179844e+00 -7.19231665e-01 1.76204994e-01 1.35225952e-01
-1.41034293e+00 -2.64047563e-01 5.74039936e-01 -4.65659142e-01
7.44491756e-01 2.83583164e-01 1.35571969e+00 1.52058315e+00
7.72847608e-02 9.42994952e-01 6.82111502e-01 -2.95752913e-01
-7.51589388e-02 -1.04640174e+00 -8.81466791e-02 1.05076671e+00
-9.51007232e-02 -9.17896703e-02 -1.09056425e+00 1.51454389e-01
1.36656725e+00 7.44738653e-02 -2.35030189e-01 -3.72763455e-01
-1.44729710e+00 3.57091337e-01 3.94617498e-01 1.49268940e-01
-3.15418601e-01 4.17067438e-01 1.75039962e-01 5.63801229e-02
2.29807481e-01 -4.55940254e-02 -1.65012732e-01 -4.27818775e-01
-1.02120149e+00 2.95433491e-01 7.85315812e-01 1.12948728e+00
3.88664395e-01 2.51749814e-01 -1.02884464e-01 6.92635179e-01
4.21318203e-01 7.08222210e-01 4.79968011e-01 -1.53253984e+00
4.31111425e-01 5.73566318e-01 6.00687452e-02 -1.18362868e+00
-5.86529672e-01 8.68812762e-03 -7.05816984e-01 4.99108694e-02
4.98765320e-01 9.55550373e-02 -1.12448335e+00 1.55540085e+00
5.19319594e-01 3.35496277e-01 -3.91425401e-01 1.50719094e+00
1.01413202e+00 4.38557416e-01 -2.07828879e-01 -9.96454731e-02
1.38758981e+00 -1.03226662e+00 -6.16650105e-01 -2.34222859e-01
6.59362674e-02 -5.79364896e-01 1.04987597e+00 4.99225438e-01
-1.27636540e+00 -8.36845279e-01 -9.56195712e-01 -2.76566297e-01
3.38206589e-01 4.14044619e-01 3.60109031e-01 4.43733692e-01
-7.92124391e-01 5.96096754e-01 -1.30999577e+00 -2.60156095e-01
2.11856827e-01 4.37436730e-01 -4.64739412e-01 1.06144957e-01
-7.34788299e-01 6.28846884e-01 2.02930227e-01 5.26787817e-01
-1.02923906e+00 -3.64943683e-01 -1.01649821e+00 -4.15525585e-01
6.13542438e-01 -8.69955778e-01 1.20909131e+00 -9.29755449e-01
-1.58697665e+00 1.09397209e+00 -1.29034951e-01 -3.39509686e-03
6.83296978e-01 -7.71256566e-01 -2.34351858e-01 1.41523689e-01
2.51199216e-01 6.03074312e-01 8.31261754e-01 -1.25177240e+00
-5.06367922e-01 -5.63901067e-01 -1.78061761e-02 5.80259085e-01
-7.24724457e-02 -1.17474690e-01 -1.40283000e+00 -8.07502568e-01
5.67422807e-01 -1.19381499e+00 1.57830328e-01 2.92865008e-01
-5.74929178e-01 3.85240465e-02 6.59324765e-01 -9.21779454e-01
1.17073786e+00 -2.08355141e+00 9.40573037e-01 6.66416213e-02
1.17244445e-01 -1.13267124e-01 -5.97966053e-02 -6.04551584e-02
4.29539889e-01 -2.77062654e-01 -5.25705963e-02 -6.05506301e-01
7.00949430e-02 6.01520538e-01 7.01535773e-03 6.38896525e-01
-1.49976790e-01 8.17192197e-01 -6.83149338e-01 -5.51500201e-01
1.35814980e-01 5.79296649e-01 -4.98366386e-01 1.64420441e-01
-2.59891838e-01 6.98998928e-01 -5.73615313e-01 1.23836410e+00
6.54070303e-02 -4.17674273e-01 3.02977026e-01 -4.91443574e-01
3.07621118e-02 -9.45542902e-02 -1.52111614e+00 2.42806602e+00
2.12609582e-02 5.75415134e-01 9.77838710e-02 -5.45012534e-01
6.25450611e-01 1.61696315e-01 8.39354813e-01 -4.57195461e-01
3.63556296e-01 2.13898152e-01 -3.17392111e-01 -6.57043517e-01
5.51100850e-01 1.72727779e-01 -3.15080643e-01 3.73947680e-01
2.27288723e-01 2.11386368e-01 3.19005847e-01 -3.40710908e-01
9.25708532e-01 9.86542284e-01 1.47786364e-01 2.23304570e-01
3.02506328e-01 3.06860685e-01 6.52612150e-01 5.68969607e-01
-1.09764002e-01 9.26646292e-01 1.84217349e-01 -5.87751985e-01
-8.14877570e-01 -1.28117990e+00 5.33812046e-01 1.18486536e+00
5.37207961e-01 -4.31158572e-01 -7.55720437e-01 -4.49296564e-01
-2.09165797e-01 -1.56058192e-01 -7.09724009e-01 1.74280345e-01
-1.31373131e+00 -5.75018466e-01 8.94523919e-01 8.11849117e-01
7.58754134e-01 -9.41982687e-01 -7.19521642e-01 8.51330981e-02
-4.69258219e-01 -1.27965581e+00 -1.24373460e+00 -2.51935363e-01
-7.59267271e-01 -1.47047400e+00 -7.86149561e-01 -9.94346082e-01
2.92971343e-01 1.91593379e-01 1.16720891e+00 7.63725862e-02
-5.71096659e-01 5.36372125e-01 -3.08772236e-01 1.49296626e-01
-1.06338710e-01 -2.17177868e-01 4.84569460e-01 -1.47954151e-01
2.72444814e-01 -5.06935656e-01 -5.43247461e-01 6.28551304e-01
-3.46144885e-01 -3.50041757e-03 4.49676692e-01 7.58995414e-01
1.08867061e+00 -5.24228096e-01 -3.81218284e-01 -1.73592791e-01
2.04623610e-01 2.56280512e-01 -3.86990041e-01 3.28122884e-01
2.63956338e-01 -3.38987082e-01 3.98022011e-02 -9.09094334e-01
-8.26336443e-01 4.36649233e-01 -1.69812292e-02 -1.07191908e+00
-2.99916953e-01 -1.84530877e-02 -2.20648021e-01 -2.40278207e-02
6.34915173e-01 2.57934064e-01 -2.54961103e-02 -5.93540251e-01
4.30907905e-01 2.39290252e-01 1.06371486e+00 -7.19218016e-01
6.89448178e-01 5.57980180e-01 6.95832446e-02 -9.03628886e-01
-1.13289511e+00 -4.34027344e-01 -1.15915501e+00 -7.37419009e-01
1.50838256e+00 -1.12031507e+00 -1.19856203e+00 4.47092235e-01
-1.06945431e+00 -4.68323290e-01 -2.36312956e-01 7.39707530e-01
-6.44558907e-01 3.42155755e-01 -1.03082538e+00 -6.13561153e-01
-1.74287021e-01 -1.13525414e+00 1.65024364e+00 1.10963866e-01
-6.45800650e-01 -7.67410219e-01 6.13905564e-02 5.07881880e-01
-5.17582238e-01 7.53893793e-01 3.26790392e-01 1.36512786e-01
-6.52763307e-01 1.05571613e-01 2.38718748e-01 -8.41182172e-02
1.65904775e-01 -2.25657910e-01 -5.48161149e-01 -1.87271267e-01
-1.12329476e-01 -4.67039913e-01 5.54201901e-01 7.38046229e-01
8.44428301e-01 -9.12164897e-02 -1.54491872e-01 9.91294205e-01
8.60691011e-01 1.13609945e-02 2.51363397e-01 5.24691999e-01
1.25892901e+00 3.81159216e-01 6.30149961e-01 3.11694860e-01
4.03017104e-01 1.00654948e+00 3.09977293e-01 9.94798541e-02
-4.70928371e-01 -2.45888308e-01 7.16162920e-01 1.00746059e+00
-9.15951490e-01 -1.58561654e-02 -6.74995780e-01 3.33495624e-02
-1.81077886e+00 -1.04000640e+00 -2.78559893e-01 1.96662879e+00
7.67235041e-01 -4.59543355e-02 6.46429420e-01 -1.86546266e-01
6.73791826e-01 1.18202798e-01 -8.63388300e-01 3.53550225e-01
-3.38927805e-02 -1.43584749e-02 3.78065944e-01 3.45999211e-01
-1.30024064e+00 9.43110168e-01 6.48571110e+00 5.42015791e-01
-8.37843478e-01 5.04183397e-02 2.29437221e-02 -6.22098029e-01
3.23790520e-01 -5.35252929e-01 -8.61977458e-01 8.73704553e-02
1.23572350e-01 5.11611462e-01 5.15665114e-01 6.98511362e-01
8.95853490e-02 2.76609898e-01 -1.29007280e+00 1.41139627e+00
5.99766076e-01 -1.29203582e+00 2.88991127e-02 1.47654399e-01
6.24316096e-01 -1.87002927e-01 -2.18697980e-01 1.74581781e-01
1.80450067e-01 -9.66988683e-01 1.22026277e+00 8.28256309e-01
8.12820017e-01 -4.31080848e-01 1.26484215e-01 4.17357713e-01
-1.57248247e+00 1.32645652e-01 -3.89312324e-03 7.82846138e-02
3.90955150e-01 -2.87183315e-01 -1.34684682e-01 4.98381138e-01
1.14888918e+00 1.07413375e+00 -3.96872848e-01 6.48022830e-01
-2.90854812e-01 1.65049225e-01 -3.91145706e-01 1.97619691e-01
3.63017246e-02 -4.16141927e-01 6.40576005e-01 8.50931764e-01
5.86590588e-01 3.17482203e-01 6.97985291e-01 6.28893912e-01
-1.80146307e-01 -4.09244031e-01 -8.95103589e-02 1.92661919e-02
4.03654367e-01 9.27651882e-01 -1.00397563e+00 -3.00477177e-01
-2.24018633e-01 1.48230290e+00 -8.21224302e-02 3.50004345e-01
-9.79498804e-01 2.30643198e-01 8.22543800e-01 -1.18073612e-01
3.23007792e-01 -7.31336713e-01 1.21149542e-02 -1.33671558e+00
2.06887648e-01 -1.04877794e+00 6.31354570e-01 -8.10611963e-01
-1.25944519e+00 4.10889655e-01 -1.52021060e-02 -1.52427483e+00
-3.61785889e-01 -7.49708891e-01 -1.22931927e-01 2.10252181e-01
-4.87677068e-01 -1.31338811e+00 -6.46062195e-01 7.14050412e-01
1.02247798e+00 3.97675224e-02 7.29675531e-01 4.68568057e-01
-5.43055892e-01 3.61260414e-01 -3.66112500e-01 6.37905657e-01
5.95462143e-01 -1.07718766e+00 3.16494495e-01 5.49026132e-01
8.45232368e-01 4.66908753e-01 4.47909862e-01 -7.96187222e-01
-1.90376318e+00 -8.23577344e-01 3.77664626e-01 -9.28968608e-01
4.82062727e-01 -3.86466116e-01 -6.47137225e-01 8.79505038e-01
-2.55362391e-01 -1.67645529e-01 6.62034214e-01 7.48516945e-03
-1.42846987e-01 2.07844645e-01 -4.42031592e-01 6.95233643e-01
1.83076024e+00 -3.02114815e-01 -8.38891983e-01 2.69054770e-01
4.12388980e-01 -1.28168654e+00 -1.07102168e+00 1.95256189e-01
1.05048645e+00 -8.41118872e-01 1.42585611e+00 -5.51313400e-01
2.46342242e-01 -3.86335045e-01 -2.84106463e-01 -8.09814095e-01
-9.78398398e-02 -4.54633355e-01 -6.31473720e-01 6.70874655e-01
-1.78362593e-01 4.93003160e-01 1.19703305e+00 3.68641615e-01
-3.07283968e-01 -5.13865292e-01 -1.00911236e+00 -8.95799816e-01
-4.19350356e-01 -5.15196443e-01 3.18796188e-01 8.09576094e-01
-5.52919149e-01 2.55746067e-01 -9.19311345e-01 2.43079454e-01
7.26802766e-01 4.68883634e-01 1.18265450e+00 -1.11931622e+00
-5.81589878e-01 -4.61911023e-01 -4.59405720e-01 -1.49233055e+00
8.55578259e-02 -5.33024132e-01 3.49149734e-01 -1.50520563e+00
-4.11125384e-02 2.64262736e-01 2.53551841e-01 5.41903734e-01
1.70729190e-01 9.39180136e-01 3.34943652e-01 5.25586724e-01
-9.51323390e-01 4.88439739e-01 1.57076800e+00 -1.22597143e-01
-5.61273813e-01 -1.44406915e-01 -3.42878141e-02 1.31892097e+00
1.85001850e-01 -4.16659296e-01 -5.35834916e-02 -8.66756856e-01
-1.19155459e-01 3.01697642e-01 6.72914922e-01 -1.14890659e+00
3.07330817e-01 -9.27784890e-02 9.02580261e-01 -8.10966611e-01
9.21007395e-01 -4.42830324e-01 4.96507347e-01 5.98952830e-01
-1.92497954e-01 1.14860669e-01 1.04449093e-02 6.75681770e-01
-6.39337450e-02 2.22985446e-01 2.08520830e-01 -3.85509998e-01
-9.97746110e-01 6.07918441e-01 -3.19026470e-01 2.05546573e-01
7.41010010e-01 -6.32268786e-01 1.94878012e-01 -2.49057487e-01
-1.40533268e+00 2.46048167e-01 4.74961698e-01 6.90080762e-01
6.53223634e-01 -1.66311288e+00 -4.89530474e-01 7.61624426e-02
-6.47853613e-02 8.10201094e-02 3.72180551e-01 9.68278885e-01
-7.97681332e-01 -1.26762941e-01 -2.74906367e-01 -1.09103119e+00
-1.56999993e+00 -2.84215691e-03 3.40422690e-01 1.73829138e-01
-1.11183047e+00 9.47656512e-01 -5.34767136e-02 -3.55224997e-01
6.19975388e-01 -3.61292601e-01 -1.32494748e-01 1.83238104e-01
4.92464244e-01 5.01415908e-01 -3.56320322e-01 -1.20412862e+00
-4.79432911e-01 1.37482047e+00 5.20165741e-01 -1.93166837e-01
1.18079960e+00 -1.16578832e-01 1.27074182e-01 6.65553868e-01
9.07621503e-01 -1.01259790e-01 -1.63700652e+00 -4.29973677e-02
-1.49675369e-01 -6.41922951e-01 -4.75700796e-01 -3.28118205e-01
-1.01932561e+00 7.69470155e-01 4.75546509e-01 -1.46730945e-01
1.13848138e+00 1.06393740e-01 9.75594521e-01 5.76825857e-01
4.55760270e-01 -1.07915974e+00 6.02212965e-01 5.59874594e-01
8.73257101e-01 -9.16764259e-01 1.63384587e-01 -3.67702603e-01
-6.71174824e-01 1.20003676e+00 7.92244673e-01 -3.19633543e-01
2.82267809e-01 8.88838917e-02 1.25400618e-01 -3.53175163e-01
-9.51890796e-02 -3.06255281e-01 7.62438655e-01 5.70577264e-01
3.21359158e-01 -2.46424735e-01 1.22872241e-01 5.84912598e-01
-3.92770708e-01 -7.53330961e-02 1.65035322e-01 8.38922739e-01
-2.60491610e-01 -9.68105495e-01 -7.85249591e-01 -3.46217514e-03
-2.59079605e-01 3.71094137e-01 -6.29720271e-01 1.31046844e+00
4.94252056e-01 5.72311461e-01 2.44088456e-01 -6.09161973e-01
6.71180487e-01 -6.87787756e-02 1.02171910e+00 -5.58586836e-01
-5.14099240e-01 7.58160293e-01 2.07183331e-01 -9.57018256e-01
-9.33143497e-01 -8.72917533e-01 -1.43297398e+00 -1.70750529e-01
8.38650204e-03 -2.64593452e-01 1.80683956e-01 1.09194541e+00
3.58526818e-02 7.45391667e-01 -4.83521149e-02 -1.65222728e+00
-1.25533238e-01 -8.18819940e-01 -5.86406589e-01 5.33024728e-01
3.06450784e-01 -9.99092519e-01 1.85095221e-01 6.29008591e-01] | [7.157917499542236, -0.7518528699874878] |
8f83d3f8-7144-41b7-a573-065dd8dfab93 | learning-what-to-learn-for-video-object | 2003.11540 | null | https://arxiv.org/abs/2003.11540v2 | https://arxiv.org/pdf/2003.11540v2.pdf | Learning What to Learn for Video Object Segmentation | Video object segmentation (VOS) is a highly challenging problem, since the target object is only defined during inference with a given first-frame reference mask. The problem of how to capture and utilize this limited target information remains a fundamental research question. We address this by introducing an end-to-end trainable VOS architecture that integrates a differentiable few-shot learning module. This internal learner is designed to predict a powerful parametric model of the target by minimizing a segmentation error in the first frame. We further go beyond standard few-shot learning techniques by learning what the few-shot learner should learn. This allows us to achieve a rich internal representation of the target in the current frame, significantly increasing the segmentation accuracy of our approach. We perform extensive experiments on multiple benchmarks. Our approach sets a new state-of-the-art on the large-scale YouTube-VOS 2018 dataset by achieving an overall score of 81.5, corresponding to a 2.6% relative improvement over the previous best result. | ['Luc van Gool', 'Martin Danelljan', 'Felix Järemo Lawin', 'Radu Timofte', 'Goutam Bhat', 'Andreas Robinson', 'Michael Felsberg'] | 2020-03-25 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/4440_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123470766.pdf | eccv-2020-8 | ['one-shot-visual-object-segmentation'] | ['computer-vision'] | [ 2.89773464e-01 1.72373146e-01 -6.01336718e-01 -2.88971126e-01
-1.05096436e+00 -3.43131989e-01 2.76234478e-01 -4.30442929e-01
-4.41549242e-01 4.28570271e-01 1.68730672e-02 2.05321819e-01
3.70328277e-01 -4.34281141e-01 -1.01059878e+00 -5.06570935e-01
2.11013392e-01 4.45042312e-01 8.92684639e-01 2.36684620e-01
3.24515365e-02 5.19894622e-02 -1.53617609e+00 2.21154317e-01
6.75130725e-01 1.27384019e+00 4.08090085e-01 6.93615377e-01
-5.54382503e-02 9.14430320e-01 -4.55924094e-01 -2.56776363e-01
4.20633078e-01 -3.90800297e-01 -8.72478068e-01 3.06522250e-01
8.33890438e-01 -5.08698702e-01 -3.85755777e-01 9.62983429e-01
3.57185245e-01 4.48917061e-01 5.69761574e-01 -1.08687246e+00
-3.06697398e-01 5.04264474e-01 -5.47057807e-01 4.73545343e-01
1.11359537e-01 6.40649140e-01 1.11312413e+00 -7.19974697e-01
1.05859256e+00 9.49553728e-01 4.76651311e-01 8.83275390e-01
-1.04518735e+00 -4.27781343e-01 3.47626776e-01 3.95802498e-01
-1.30126059e+00 -5.39496243e-01 7.74352908e-01 -7.04650521e-01
7.57378161e-01 -7.40909129e-02 7.90215909e-01 1.11341608e+00
-1.99298546e-01 1.48881960e+00 7.15150833e-01 -7.79368356e-02
3.88915658e-01 -6.54984862e-02 3.10784996e-01 7.62256324e-01
-7.98851028e-02 -1.85175300e-01 -4.44526732e-01 2.59632736e-01
7.81001687e-01 7.56979510e-02 -3.29955012e-01 -7.39133537e-01
-8.76033425e-01 8.30321550e-01 4.07187045e-01 1.22067459e-01
-3.03529829e-01 5.74760616e-01 3.76407832e-01 -7.05519766e-02
3.91770005e-01 2.17958033e-01 -6.64100468e-01 -4.36036408e-01
-1.50288260e+00 2.36909822e-01 8.41381669e-01 1.04725420e+00
7.43488610e-01 2.59703368e-01 -6.17584109e-01 6.40207052e-01
2.77941167e-01 2.19323844e-01 3.10531348e-01 -1.43369317e+00
2.77693033e-01 2.97500163e-01 1.96140975e-01 -2.50211179e-01
-6.35416731e-02 -2.17388466e-01 -2.69792765e-01 2.34766215e-01
6.19312763e-01 -3.92909795e-01 -1.43081665e+00 1.70715451e+00
6.01661444e-01 1.03932321e+00 -2.04830602e-01 9.87994134e-01
8.96866202e-01 8.36376190e-01 2.58347809e-01 -2.27214828e-01
1.11218202e+00 -1.47843015e+00 -5.51407397e-01 -4.27478135e-01
3.73625904e-01 -3.20619494e-01 1.22370398e+00 6.87123761e-02
-1.12097251e+00 -6.79240882e-01 -9.57781613e-01 -1.15233101e-01
-9.79578644e-02 -2.17807382e-01 4.05715764e-01 5.11439681e-01
-8.45650673e-01 7.60874093e-01 -1.15353954e+00 -1.06696211e-01
1.05880570e+00 1.74806207e-01 4.96077947e-02 -9.17260721e-02
-9.04780149e-01 5.21829903e-01 4.52226579e-01 -4.02372628e-01
-1.30635285e+00 -1.39052498e+00 -1.12932444e+00 1.68592066e-01
1.02124512e+00 -8.36804032e-01 1.43268549e+00 -9.66199040e-01
-1.56936038e+00 8.06370854e-01 -3.51990938e-01 -7.75588334e-01
7.94429898e-01 -3.99790943e-01 1.21972300e-01 4.20008749e-01
2.65380889e-01 9.25985336e-01 1.10986471e+00 -1.15797877e+00
-7.95174181e-01 -5.76259829e-02 1.38601258e-01 -2.33070888e-02
-5.81664033e-02 6.01468561e-03 -1.07624674e+00 -5.68443775e-01
-4.75685686e-01 -8.23112845e-01 -3.07034671e-01 2.08337992e-01
-1.73659548e-01 -3.96959692e-01 9.20275152e-01 -5.04137814e-01
1.04825413e+00 -2.02949286e+00 3.02657515e-01 -3.96856099e-01
3.21931690e-01 5.77493072e-01 -3.39991823e-02 -1.84831873e-01
3.40304852e-01 6.26921132e-02 -5.24021685e-01 -5.64537466e-01
-1.44520728e-02 1.00578219e-01 -3.41199368e-01 4.73984569e-01
2.86123872e-01 1.37244868e+00 -9.17912006e-01 -7.12996125e-01
4.36631441e-01 4.16222334e-01 -5.26284933e-01 5.76200783e-01
-7.35940754e-01 2.89317071e-01 -3.48210394e-01 7.34779596e-01
2.95181453e-01 -4.68359053e-01 -3.31956863e-01 -7.91260973e-03
8.66530836e-02 -1.33917376e-01 -1.14714670e+00 2.05482388e+00
-3.67798395e-02 7.18083024e-01 -7.77058974e-02 -8.94330263e-01
5.30358613e-01 7.59166032e-02 8.16033542e-01 -5.16751111e-01
2.92710215e-01 -1.30710453e-01 -2.34651431e-01 -6.76328361e-01
1.10605113e-01 -2.35803183e-02 6.33761063e-02 8.90473872e-02
5.69817245e-01 6.88529164e-02 3.84718001e-01 2.57189125e-01
9.46728945e-01 5.40967822e-01 1.11840725e-01 -2.11444944e-01
1.90358207e-01 -1.18034603e-02 9.06206369e-01 7.24849045e-01
-6.10261142e-01 9.43168998e-01 6.13367796e-01 -4.43477988e-01
-8.99468541e-01 -9.53471959e-01 7.13928863e-02 1.04660988e+00
4.83303279e-01 -3.57052684e-01 -1.07574975e+00 -8.78063083e-01
-1.63244903e-01 8.07873487e-01 -8.70859861e-01 -3.74845304e-02
-6.58819437e-01 -1.65338352e-01 2.70808160e-01 8.41730773e-01
4.17198211e-01 -1.13351250e+00 -1.04256356e+00 1.90385401e-01
-2.13638157e-01 -1.46704638e+00 -8.17244530e-01 -1.97734967e-01
-7.67681599e-01 -1.27506435e+00 -1.05553174e+00 -6.64932013e-01
1.60460949e-01 2.01899081e-01 1.20567155e+00 -1.17433667e-01
-3.65645170e-01 4.85656857e-01 -1.98551759e-01 -3.61263156e-01
5.35077266e-02 2.36767277e-01 -3.45324755e-01 1.85931101e-01
3.81914675e-01 -2.59266317e-01 -6.90028667e-01 1.46156445e-01
-7.28149831e-01 2.55059060e-02 2.50948638e-01 5.96549392e-01
9.04895842e-01 -5.04717112e-01 4.60510880e-01 -1.07044518e+00
-8.81464854e-02 -6.08482301e-01 -7.02856362e-01 2.94016361e-01
-2.06284344e-01 -1.26261160e-01 4.44443613e-01 -5.73957443e-01
-9.55090940e-01 4.59562361e-01 -1.08579338e-01 -1.07114017e+00
-2.43364304e-01 1.02200277e-01 -1.93703175e-01 -7.38639235e-02
3.32315356e-01 1.28214985e-01 -1.32211030e-01 -4.52333182e-01
5.19104540e-01 2.68915325e-01 6.74407363e-01 -4.35986549e-01
6.52303338e-01 5.82860231e-01 -2.52627999e-01 -8.93905997e-01
-1.40219545e+00 -9.39747751e-01 -8.57755482e-01 -3.84183556e-01
1.08994627e+00 -1.07630825e+00 -4.94791955e-01 3.43457639e-01
-1.04491472e+00 -9.08645630e-01 -7.10632324e-01 1.61448911e-01
-9.34882462e-01 2.67775118e-01 -5.00267088e-01 -6.10753179e-01
-3.40693176e-01 -1.29034662e+00 1.20369709e+00 4.62941140e-01
-2.12673202e-01 -9.49376106e-01 2.28507742e-01 5.20951033e-01
1.79552957e-01 5.03292799e-01 2.29009286e-01 -6.12568498e-01
-8.50225508e-01 1.20452680e-01 -1.57515153e-01 2.83239901e-01
-2.88054556e-01 4.99554835e-02 -9.55564201e-01 -1.59223229e-01
1.50999948e-01 -5.00793457e-01 1.31527817e+00 7.08931088e-01
1.20182157e+00 9.92625877e-02 -3.04282993e-01 1.05700278e+00
1.59620035e+00 -4.33789752e-02 6.38586700e-01 4.18402031e-02
1.02977359e+00 2.57375240e-01 8.64828706e-01 2.97167212e-01
4.43448365e-01 6.84096456e-01 4.45248157e-01 1.42776713e-01
-4.85987365e-01 -2.70652801e-01 2.01458976e-01 2.66534656e-01
2.72668917e-02 -1.04300410e-01 -8.66983116e-01 8.57091069e-01
-2.13476777e+00 -1.06969249e+00 1.81673542e-01 1.88275063e+00
7.95936406e-01 4.45539117e-01 4.02193695e-01 -3.57734561e-01
7.09111035e-01 5.86551368e-01 -1.04701364e+00 1.05784386e-01
2.82530576e-01 2.02864692e-01 4.47875649e-01 4.83875185e-01
-1.46651042e+00 1.51940322e+00 6.40531254e+00 9.46164668e-01
-1.13842571e+00 2.30766982e-01 6.68796122e-01 -3.16323191e-01
7.86606967e-02 -9.15916264e-02 -1.05005383e+00 5.78019142e-01
9.58204329e-01 -1.96172982e-01 2.75997192e-01 1.12208426e+00
2.03243285e-01 -9.22663808e-02 -1.17750990e+00 1.03959310e+00
2.12665170e-01 -1.68978274e+00 -8.33942667e-02 -1.62734002e-01
1.15765178e+00 3.28101009e-01 -1.44012690e-01 6.36968553e-01
7.75490329e-02 -9.86531019e-01 7.88622797e-01 6.83849275e-01
8.80508482e-01 -5.26686132e-01 3.93980831e-01 3.61208349e-01
-1.32393765e+00 1.28506377e-01 -4.29710865e-01 1.20154485e-01
5.37562251e-01 2.00870588e-01 -5.91992259e-01 1.62211642e-01
6.69894278e-01 9.59632516e-01 -3.81749451e-01 1.49250638e+00
-2.46434614e-01 1.00608897e+00 -3.12804580e-01 2.16138780e-01
4.82597470e-01 4.49838303e-02 6.50371373e-01 1.25328732e+00
-8.85544717e-02 2.28922471e-01 4.98859078e-01 1.01791930e+00
-4.78171289e-01 -9.90030989e-02 -3.18652898e-01 5.00866622e-02
1.87507555e-01 1.28204477e+00 -9.63369429e-01 -6.91114783e-01
-5.14870703e-01 9.18970644e-01 3.73160988e-01 3.57987940e-01
-1.13596177e+00 -3.02556396e-01 7.53697991e-01 1.95716009e-01
9.55272973e-01 -1.08877003e-01 -1.89072728e-01 -1.33668387e+00
-2.47596595e-02 -5.68188727e-01 3.37826550e-01 -4.73763168e-01
-1.01050830e+00 3.07435393e-01 4.55452055e-02 -1.01214397e+00
-3.49742234e-01 -3.78876776e-01 -9.16224658e-01 4.51625705e-01
-1.58139300e+00 -1.22405386e+00 -3.67268920e-01 5.21580219e-01
1.25814831e+00 2.92310435e-02 3.44006956e-01 1.28879994e-01
-5.39236724e-01 5.11373520e-01 -7.89421424e-02 3.43401134e-01
5.37532628e-01 -1.40921366e+00 5.13235033e-01 9.75096762e-01
4.07678783e-01 1.32651135e-01 6.00186884e-01 -6.47430122e-01
-1.26020384e+00 -1.32057297e+00 2.48655364e-01 -7.88422823e-01
5.77204704e-01 -3.71389776e-01 -1.06852996e+00 8.82305086e-01
5.95804974e-02 6.81561589e-01 5.17330348e-01 -3.15605193e-01
-3.13343376e-01 1.03051327e-01 -1.05388761e+00 5.25884867e-01
1.15072870e+00 -2.46859133e-01 -7.25270033e-01 1.75281838e-01
1.19194460e+00 -6.31265342e-01 -7.85673201e-01 4.30911213e-01
4.05691594e-01 -7.39183009e-01 1.02006960e+00 -9.28207457e-01
5.72299123e-01 -2.76399940e-01 3.97883169e-02 -1.02202451e+00
-1.27437264e-01 -8.05787921e-01 -8.11430693e-01 1.17226732e+00
1.23913325e-01 1.18520804e-01 1.17695725e+00 9.10211086e-01
-2.08892226e-01 -1.13160586e+00 -8.13622177e-01 -8.44972312e-01
-2.70491047e-03 -5.14176250e-01 2.09815502e-01 6.59764111e-01
-3.15776706e-01 3.81888926e-01 -5.12523234e-01 -1.02102228e-01
9.63812232e-01 2.05069050e-01 8.80696237e-01 -1.10196424e+00
-3.66641343e-01 -3.54157507e-01 -4.97877419e-01 -1.51503873e+00
4.37886000e-01 -7.09644496e-01 2.83797026e-01 -1.62923968e+00
3.04250091e-01 1.08564943e-01 -3.57445538e-01 3.45120251e-01
-4.21505988e-01 3.82147670e-01 5.50536871e-01 7.33030662e-02
-1.30749512e+00 6.59507811e-01 1.27520740e+00 -2.47599274e-01
-3.46617132e-01 1.52546674e-01 -5.20572543e-01 9.71866369e-01
6.21340513e-01 -4.39344674e-01 -4.72903520e-01 -3.90082747e-01
-5.84435582e-01 3.85199301e-02 2.72221297e-01 -1.05035281e+00
4.07969475e-01 -3.98352712e-01 2.64611006e-01 -4.83490586e-01
5.25166810e-01 -5.47847092e-01 -2.52356678e-01 3.39743644e-01
-3.18601221e-01 -7.75164485e-01 6.61079884e-02 7.93627203e-01
-8.18118155e-02 -4.12379235e-01 1.11393476e+00 -1.80292532e-01
-1.38960290e+00 8.48239660e-01 2.70757992e-02 6.90777004e-01
1.53598070e+00 -2.92097300e-01 1.09808855e-01 -1.49254367e-01
-8.41396809e-01 5.30017018e-01 4.47014272e-01 5.15873432e-01
6.80311084e-01 -1.03403437e+00 -5.39837837e-01 -1.22248702e-01
8.37667882e-02 2.58638024e-01 2.96623349e-01 5.82089186e-01
-3.17478240e-01 1.12763375e-01 -1.07721917e-01 -8.27637136e-01
-1.15291858e+00 6.99812710e-01 3.36534292e-01 -1.49055421e-02
-1.01489341e+00 1.01905370e+00 2.01933965e-01 1.65539518e-01
5.16630828e-01 -2.28917614e-01 -2.51690120e-01 4.17771302e-02
8.04483235e-01 4.98461753e-01 -4.92935330e-01 -7.96071589e-01
-1.60127997e-01 6.79342866e-01 -1.01954296e-01 9.95479748e-02
1.41015387e+00 -8.52490664e-02 4.44201291e-01 7.12858915e-01
1.43963265e+00 -3.83369416e-01 -2.13158250e+00 -3.89192075e-01
2.70004757e-02 -6.48070395e-01 -3.24900635e-02 -4.71781522e-01
-1.29849470e+00 8.97608042e-01 4.12742943e-01 -7.77049139e-02
6.42518103e-01 2.55231798e-01 1.21149051e+00 2.69443452e-01
2.67620504e-01 -1.23711288e+00 2.64720291e-01 5.42839348e-01
4.05378997e-01 -1.63337207e+00 -2.77125001e-01 -4.02353197e-01
-9.31165755e-01 1.00010109e+00 7.89033771e-01 -5.59948385e-01
6.74109936e-01 2.00260252e-01 -1.08134016e-01 -2.41889253e-01
-6.51225328e-01 -4.91096616e-01 5.06532192e-01 5.24227560e-01
4.67490777e-02 -1.48500025e-01 -4.36656084e-03 8.25814426e-01
1.84772477e-01 2.97198236e-01 3.49158376e-01 7.27328122e-01
-7.19283700e-01 -6.53069019e-01 3.03292945e-02 4.99512613e-01
-6.70064986e-01 1.71243325e-01 -1.16798863e-01 6.13334417e-01
-2.49750968e-02 6.10871315e-01 4.55585755e-02 -6.82939216e-02
2.33323917e-01 2.50253409e-01 4.13076788e-01 -8.53882730e-01
-1.51726693e-01 1.39873683e-01 -2.63597369e-01 -1.03594530e+00
-5.25322378e-01 -7.40650892e-01 -1.44283700e+00 8.67311209e-02
-1.44878149e-01 -1.35732889e-01 3.21483284e-01 1.21054435e+00
2.27595329e-01 6.67209506e-01 3.55204552e-01 -1.23325205e+00
-4.57351357e-01 -6.74616218e-01 -3.19214314e-01 5.42297423e-01
3.24971348e-01 -6.83054805e-01 -1.54895842e-01 4.12708849e-01] | [9.173091888427734, 0.021023431792855263] |
f91a1cac-6476-40eb-a528-10752a181739 | learning-local-global-contextual-adaptation | 2109.03622 | null | https://arxiv.org/abs/2109.03622v2 | https://arxiv.org/pdf/2109.03622v2.pdf | Learning Local-Global Contextual Adaptation for Multi-Person Pose Estimation | This paper studies the problem of multi-person pose estimation in a bottom-up fashion. With a new and strong observation that the localization issue of the center-offset formulation can be remedied in a local-window search scheme in an ideal situation, we propose a multi-person pose estimation approach, dubbed as LOGO-CAP, by learning the LOcal-GlObal Contextual Adaptation for human Pose. Specifically, our approach learns the keypoint attraction maps (KAMs) from the local keypoints expansion maps (KEMs) in small local windows in the first step, which are subsequently treated as dynamic convolutional kernels on the keypoints-focused global heatmaps for contextual adaptation, achieving accurate multi-person pose estimation. Our method is end-to-end trainable with near real-time inference speed in a single forward pass, obtaining state-of-the-art performance on the COCO keypoint benchmark for bottom-up human pose estimation. With the COCO trained model, our method also outperforms prior arts by a large margin on the challenging OCHuman dataset. | ['Liangpei Zhang', 'Gui-Song Xia', 'Tianfu Wu', 'Nan Xue'] | 2021-09-08 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Xue_Learning_Local-Global_Contextual_Adaptation_for_Multi-Person_Pose_Estimation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Xue_Learning_Local-Global_Contextual_Adaptation_for_Multi-Person_Pose_Estimation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['multi-person-pose-estimation'] | ['computer-vision'] | [-1.20226882e-01 -6.97709396e-02 -2.30609532e-02 -3.39808166e-01
-1.27498233e+00 -4.97637838e-01 4.93669361e-01 4.37197275e-02
-8.41308951e-01 5.78393638e-01 4.59041834e-01 6.85036242e-01
-4.82753180e-02 -1.28554136e-01 -1.04668343e+00 -3.89980972e-01
3.91357467e-02 8.66937339e-01 2.01236159e-01 -3.01781923e-01
-3.30717303e-02 2.70366848e-01 -1.31701946e+00 9.83137451e-03
6.01958692e-01 8.67240071e-01 -9.31057185e-02 8.54774356e-01
8.23899150e-01 2.00635105e-01 -3.56500655e-01 -6.92617953e-01
3.70902449e-01 3.89338955e-02 -6.69216156e-01 -9.63377729e-02
1.31572092e+00 -3.08358461e-01 -4.59818393e-01 6.88060284e-01
9.52596009e-01 3.45555186e-01 3.58378887e-01 -1.19765997e+00
-3.12549353e-01 3.87955159e-02 -7.44665623e-01 -3.17328237e-02
8.16198945e-01 2.28564844e-01 9.32833195e-01 -1.38248456e+00
5.95505118e-01 1.51719892e+00 1.12046742e+00 4.45643693e-01
-1.10061991e+00 -5.38167238e-01 4.38815266e-01 2.24192828e-01
-1.70767784e+00 -3.15440506e-01 4.87300813e-01 -2.77349263e-01
9.14418876e-01 2.07079694e-01 8.37047398e-01 1.21114147e+00
2.72458941e-01 1.05806744e+00 7.94040382e-01 -4.89409983e-01
-1.16543114e-01 -1.20580785e-01 -2.25849241e-01 7.68959284e-01
2.35372514e-01 -7.24776648e-03 -9.41797793e-01 -2.06415117e-01
7.55888760e-01 1.19847348e-02 -1.44055620e-01 -7.61921942e-01
-1.45541430e+00 6.88106298e-01 7.85377979e-01 -2.35299528e-01
-3.65067661e-01 4.21503961e-01 3.30959469e-01 -1.68302283e-01
5.46749055e-01 2.87780404e-01 -6.86476946e-01 -9.73100290e-02
-8.51355612e-01 1.09254634e+00 4.74145770e-01 1.00692308e+00
5.50824165e-01 -6.68475568e-01 -4.25790966e-01 6.51095390e-01
3.04562539e-01 6.84025347e-01 2.86237031e-01 -5.90052187e-01
7.81683981e-01 2.96745747e-01 6.85785294e-01 -9.45151210e-01
-6.69114351e-01 -5.91156721e-01 -4.72693086e-01 -1.28682122e-01
8.89900446e-01 -2.03371093e-01 -9.80565786e-01 1.91466177e+00
9.20219600e-01 4.74340655e-02 -4.47665870e-01 1.11608231e+00
3.56317401e-01 3.67763191e-01 2.88564533e-01 3.16183209e-01
1.75335586e+00 -1.42869997e+00 -4.01474327e-01 -8.75172019e-01
3.75379205e-01 -6.16164744e-01 9.96902227e-01 2.29609862e-01
-9.56148446e-01 -8.03845406e-01 -8.48366618e-01 -2.62051582e-01
-3.68179321e-01 4.69897002e-01 5.72200000e-01 3.80204827e-01
-7.89978087e-01 5.08475423e-01 -9.23535585e-01 -6.96972966e-01
2.08265334e-01 2.88272232e-01 -5.57510614e-01 -7.39623010e-02
-1.17054236e+00 1.00746131e+00 2.59164006e-01 4.54590976e-01
-6.48092866e-01 -9.41063941e-01 -1.00343668e+00 -2.06144303e-01
6.48729146e-01 -1.04276657e+00 1.14647555e+00 -3.76965076e-01
-1.32996333e+00 7.60505021e-01 -3.33142787e-01 -3.60438347e-01
1.00615597e+00 -1.20667374e+00 -6.02816716e-02 2.69412309e-01
4.11284983e-01 8.66274059e-01 9.99283314e-01 -1.07702899e+00
-7.55656183e-01 -6.42794967e-01 -2.52986848e-01 5.64308226e-01
-1.78365976e-01 -1.12412460e-02 -1.04086709e+00 -8.07046056e-01
-2.05455832e-02 -1.45452321e+00 -4.13291365e-01 1.34922519e-01
-5.34961700e-01 -2.70777166e-01 3.46612960e-01 -8.49835217e-01
1.14680541e+00 -1.69798732e+00 5.28027833e-01 2.86170870e-01
1.21676333e-01 -1.56298745e-02 6.75026178e-02 3.11297268e-01
-3.16046514e-02 -4.83820170e-01 3.71330976e-02 -7.69121230e-01
2.78242439e-01 -4.67476875e-01 9.56896041e-03 7.65357077e-01
-2.34430581e-02 1.35575795e+00 -9.39275920e-01 -6.17873073e-01
4.21309620e-01 7.32291043e-01 -6.76919758e-01 2.39020452e-01
1.69995546e-01 5.72688699e-01 -4.36405301e-01 6.04268134e-01
5.32067955e-01 -2.12747633e-01 -1.45414874e-01 -3.42654705e-01
1.37616530e-01 -3.05165946e-01 -1.43484807e+00 2.24355292e+00
-2.40706831e-01 4.00614858e-01 -3.16527724e-01 -3.30123037e-01
3.28606516e-01 1.33606330e-01 4.49977636e-01 -2.38841549e-01
-1.73967108e-01 1.86417215e-02 -5.66481650e-01 -1.60310000e-01
7.86309540e-01 1.97383657e-01 -5.43921947e-01 -6.36787042e-02
2.35968739e-01 3.02238286e-01 -8.94841924e-02 1.35070190e-01
6.40172482e-01 6.96874440e-01 4.29950267e-01 -1.73347160e-01
3.51757944e-01 -1.84849396e-01 3.21946114e-01 7.26813138e-01
-3.91513467e-01 9.19028699e-01 -3.77846546e-02 -6.01546586e-01
-1.21498215e+00 -1.15614522e+00 3.44436429e-02 1.40990222e+00
1.86347827e-01 -7.87141323e-01 -9.67509031e-01 -7.19475031e-01
3.26205552e-01 5.99480867e-02 -8.66163969e-01 2.34296247e-02
-9.16557074e-01 -6.51892304e-01 4.64173079e-01 8.70277345e-01
4.98136103e-01 -6.94170058e-01 -7.39271045e-01 1.49106607e-01
-4.23016995e-01 -1.38091505e+00 -1.24154377e+00 -1.76245496e-01
-4.21080142e-01 -9.53635812e-01 -1.41236138e+00 -7.50034750e-01
6.93315506e-01 -2.93835532e-02 9.98593807e-01 -4.32003498e-01
-3.77994835e-01 6.13183022e-01 -1.66585311e-01 -1.82786077e-01
4.68851686e-01 1.90761134e-01 3.48427981e-01 1.21497065e-01
3.54989320e-01 -1.48177996e-01 -1.15715563e+00 5.26182592e-01
-8.68827477e-02 -8.54954496e-02 6.90918088e-01 8.38948607e-01
7.74480462e-01 -4.39053923e-01 3.05009454e-01 -4.63398725e-01
3.00389856e-01 1.27365142e-01 -3.99646342e-01 4.28405195e-01
-3.94104928e-01 -5.28646708e-02 2.78720230e-01 -4.52230573e-01
-1.02071738e+00 6.11166298e-01 -8.42168629e-02 -3.51986170e-01
-1.71521172e-01 9.03512985e-02 -3.23263675e-01 -3.31979960e-01
8.22831273e-01 5.40816672e-02 -3.74446601e-01 -5.14000177e-01
5.84893525e-01 3.27864230e-01 9.64654446e-01 -7.39206553e-01
1.05566633e+00 5.46255410e-01 -1.02761827e-01 -5.83082855e-01
-1.03502643e+00 -8.62559080e-01 -1.11611342e+00 -3.95564348e-01
1.04550862e+00 -1.44412136e+00 -7.93714464e-01 6.73606932e-01
-9.84105468e-01 -1.92838237e-01 3.33372243e-02 4.00621414e-01
-6.49038911e-01 3.06477487e-01 -5.31616926e-01 -6.86709642e-01
-5.41245937e-01 -9.53332126e-01 1.84230840e+00 2.89359242e-01
-5.50517678e-01 -7.46801078e-01 2.44432017e-01 4.35965836e-01
-2.56901188e-03 6.08491480e-01 1.85727268e-01 -4.06577051e-01
-5.12482226e-01 -9.62232947e-01 4.01539654e-02 -1.73108682e-01
-4.17405516e-01 -5.88267148e-01 -1.00213957e+00 -8.72697830e-01
-5.96489549e-01 -4.03376281e-01 7.23481417e-01 4.96245414e-01
8.05313170e-01 -9.12335217e-02 -7.30340421e-01 7.45282173e-01
1.23320508e+00 -6.94346189e-01 3.26447517e-01 4.90604699e-01
1.07719946e+00 3.48184675e-01 9.14128304e-01 4.49590296e-01
7.14330971e-01 1.30110216e+00 7.78822973e-02 -6.08320534e-02
-2.32278183e-02 -7.08126783e-01 1.58369735e-01 1.04858011e-01
-2.90731072e-01 8.87668654e-02 -7.87082374e-01 5.56643009e-01
-2.28641152e+00 -7.27284908e-01 4.08435911e-01 2.39555120e+00
6.63273096e-01 1.59934714e-01 6.54153407e-01 -3.89295518e-01
8.23431134e-01 1.64419979e-01 -4.79863524e-01 4.21891183e-01
1.44282535e-01 4.38178629e-02 5.81679106e-01 4.74873006e-01
-1.66444564e+00 1.08746004e+00 6.00991774e+00 8.02669406e-01
-6.57616675e-01 1.15848146e-01 3.35953981e-01 -4.16607350e-01
3.60331297e-01 -2.69383997e-01 -1.23847365e+00 2.85976440e-01
5.93041360e-01 3.42043757e-01 2.12296501e-01 1.06514931e+00
3.42303850e-02 -1.78422064e-01 -1.31817293e+00 1.32412851e+00
1.61907166e-01 -8.90467763e-01 -2.04517618e-01 -5.28853713e-03
7.43854225e-01 -2.17522159e-01 9.69054401e-02 2.91258633e-01
-1.47615075e-01 -8.19526851e-01 1.04642332e+00 6.57044530e-01
8.00557435e-01 -1.06746840e+00 5.40339410e-01 1.70350417e-01
-1.60463059e+00 -8.02308023e-02 -2.01939389e-01 2.24214584e-01
2.92366892e-01 1.83573142e-01 -4.96842414e-01 5.79667866e-01
1.00248158e+00 4.75744247e-01 -9.38311219e-01 1.26968789e+00
-3.10038835e-01 4.46739532e-02 -4.23774332e-01 9.21503380e-02
3.00692078e-02 3.56134593e-01 6.42988801e-01 1.35241699e+00
1.48963079e-01 -2.17920825e-01 4.11525458e-01 4.17261571e-01
1.48787707e-01 7.65125230e-02 3.64412181e-02 4.27953869e-01
4.15931314e-01 1.33041036e+00 -5.90126872e-01 -2.27889225e-01
-5.74571490e-02 1.46507072e+00 5.29772103e-01 3.77954572e-01
-1.00445771e+00 -5.02245367e-01 5.73284924e-01 2.99238503e-01
4.04298007e-01 -1.87650859e-01 -2.76657473e-02 -1.23804343e+00
3.72373372e-01 -7.93221235e-01 5.18206179e-01 -6.77483559e-01
-1.18995690e+00 3.05862546e-01 4.17740464e-01 -1.24086583e+00
-5.50665677e-01 -6.36384904e-01 -3.44747066e-01 8.99548233e-01
-1.11846042e+00 -1.82608676e+00 -4.29102659e-01 6.83534741e-01
3.87629896e-01 1.14594579e-01 7.83139586e-01 2.97414303e-01
-3.81007612e-01 1.19027734e+00 -1.62290215e-01 3.00554186e-01
1.32872915e+00 -1.47022748e+00 8.16729844e-01 1.01858866e+00
6.56792596e-02 7.82254577e-01 9.71040308e-01 -7.76860952e-01
-1.28380024e+00 -1.01144826e+00 1.03759837e+00 -1.17268717e+00
2.87420839e-01 -8.13123882e-01 -4.17866081e-01 9.27421629e-01
-1.15662761e-01 2.83164501e-01 3.11162382e-01 5.01366973e-01
-3.95124167e-01 -1.77438243e-03 -9.26666677e-01 7.20037222e-01
1.18306386e+00 -5.18024802e-01 -6.13465846e-01 6.45163238e-01
6.42497420e-01 -1.02161765e+00 -8.32572579e-01 1.51994735e-01
9.96289253e-01 -4.57690537e-01 1.51444137e+00 -5.65985203e-01
1.27491385e-01 -3.82896036e-01 4.86447513e-02 -1.24739802e+00
-5.44625342e-01 -8.23164284e-01 -2.67804623e-01 6.78303242e-01
2.17876837e-01 -2.58774847e-01 9.75244880e-01 6.95209086e-01
1.41340375e-01 -8.09042811e-01 -1.03457654e+00 -5.08748114e-01
-9.92926955e-02 -2.04741269e-01 4.36362028e-01 2.65388042e-01
2.43413281e-02 3.27988178e-01 -8.97894919e-01 3.18335444e-01
1.10373855e+00 -1.51192650e-01 1.26816499e+00 -1.05856967e+00
-4.55607086e-01 -4.06488068e-02 -4.93316859e-01 -1.36064684e+00
-1.05672881e-01 -3.59355032e-01 3.58666569e-01 -1.18938673e+00
3.94651383e-01 1.43525660e-01 -1.58982396e-01 2.29889542e-01
-8.50135684e-01 4.48163569e-01 4.75772411e-01 2.07740411e-01
-1.07641888e+00 4.58713442e-01 1.16639066e+00 4.68777940e-02
-1.13923542e-01 9.57746208e-02 -5.18481433e-01 8.08877110e-01
1.38348296e-01 -3.06643575e-01 -1.87873878e-02 -1.88958690e-01
2.27068901e-01 -2.68636733e-01 8.43567133e-01 -1.33203030e+00
5.32334685e-01 2.17664570e-01 1.00635338e+00 -8.25694978e-01
7.43127763e-01 -5.36799669e-01 -7.79127479e-02 3.33413750e-01
-4.16115671e-01 2.75522590e-01 1.66216329e-01 1.01314831e+00
1.41619727e-01 4.02807564e-01 7.54154027e-01 -1.45595238e-01
-8.71993124e-01 5.47370315e-01 4.49882716e-01 2.55564779e-01
1.00788653e+00 -2.71181405e-01 8.20189640e-02 -4.11043853e-01
-9.70835686e-01 4.28875417e-01 3.93931806e-01 5.72925270e-01
5.31774580e-01 -1.58674324e+00 -6.62951767e-01 4.14202362e-02
4.75511909e-01 1.70369819e-01 5.42662382e-01 9.11334693e-01
-3.16529244e-01 6.24387980e-01 2.29732394e-02 -7.30534434e-01
-1.24472654e+00 4.56379890e-01 5.29352188e-01 -4.58552271e-01
-8.31788123e-01 1.30835652e+00 2.14058623e-01 -4.74787205e-01
3.92656982e-01 -2.21796017e-02 7.49271140e-02 2.69971397e-02
7.21302509e-01 2.88992345e-01 7.47210532e-03 -1.04686654e+00
-7.12008834e-01 1.07753193e+00 -2.01923281e-01 -2.02017054e-01
1.09840786e+00 -3.23557913e-01 4.69377726e-01 1.63398743e-01
1.21779001e+00 1.62996545e-01 -1.72225201e+00 -4.24383968e-01
-1.12199448e-01 -7.24242747e-01 -4.40932304e-01 -9.72871959e-01
-6.16482139e-01 4.62149322e-01 8.48280668e-01 -6.10310435e-01
6.50632262e-01 1.05620451e-01 9.60143030e-01 3.90743554e-01
4.33606654e-01 -1.48378658e+00 2.44021371e-01 1.68438256e-01
1.04344141e+00 -1.31854331e+00 1.91468030e-01 -3.41656595e-01
-7.59144723e-01 8.72199535e-01 7.19852686e-01 -3.19074929e-01
4.34474677e-01 -1.84726000e-01 -8.64273310e-02 -7.23317638e-02
-3.29985946e-01 -1.36496112e-01 8.00805390e-01 5.62089920e-01
2.20816657e-01 1.76995307e-01 1.67373940e-01 6.33285999e-01
-2.99941033e-01 -1.94948748e-01 -2.08665729e-01 8.59692574e-01
-4.18765903e-01 -7.36678720e-01 -7.12464631e-01 -3.01247127e-02
-3.00476909e-01 6.45306557e-02 -3.09378356e-01 1.07318807e+00
7.98155367e-02 5.09170413e-01 -2.05071896e-01 -4.91823554e-01
8.14192533e-01 2.21938789e-01 6.78408504e-01 -3.28405768e-01
-4.55625951e-01 2.90912330e-01 -2.67835669e-02 -9.69777107e-01
-2.57214040e-01 -9.85628724e-01 -9.02339041e-01 -1.35489911e-01
-1.29456148e-01 -6.09019659e-02 3.38211715e-01 1.11360431e+00
3.13955992e-01 3.58760983e-01 -3.80235389e-02 -1.48596680e+00
-6.74920738e-01 -1.08034003e+00 -2.41869241e-01 7.08947837e-01
3.83498371e-01 -7.47156024e-01 -1.97901465e-02 -2.02187359e-01] | [7.14972448348999, -0.8363513946533203] |
060b55f9-92b0-453e-a0c4-7f8d284469ee | note-rcnn-noise-tolerant-ensemble-rcnn-for | 1812.00124 | null | http://arxiv.org/abs/1812.00124v1 | http://arxiv.org/pdf/1812.00124v1.pdf | NOTE-RCNN: NOise Tolerant Ensemble RCNN for Semi-Supervised Object Detection | The labeling cost of large number of bounding boxes is one of the main
challenges for training modern object detectors. To reduce the dependence on
expensive bounding box annotations, we propose a new semi-supervised object
detection formulation, in which a few seed box level annotations and a large
scale of image level annotations are used to train the detector. We adopt a
training-mining framework, which is widely used in weakly supervised object
detection tasks. However, the mining process inherently introduces various
kinds of labelling noises: false negatives, false positives and inaccurate
boundaries, which can be harmful for training the standard object detectors
(e.g. Faster RCNN). We propose a novel NOise Tolerant Ensemble RCNN (NOTE-RCNN)
object detector to handle such noisy labels. Comparing to standard Faster RCNN,
it contains three highlights: an ensemble of two classification heads and a
distillation head to avoid overfitting on noisy labels and improve the mining
precision, masking the negative sample loss in box predictor to avoid the harm
of false negative labels, and training box regression head only on seed
annotations to eliminate the harm from inaccurate boundaries of mined bounding
boxes. We evaluate the methods on ILSVRC 2013 and MSCOCO 2017 dataset; we
observe that the detection accuracy consistently improves as we iterate between
mining and training steps, and state-of-the-art performance is achieved. | ['Li-Jia Li', 'JIyang Gao', 'Ram Nevatia', 'Shengyang Dai', 'Jiang Wang'] | 2018-12-01 | note-rcnn-noise-tolerant-ensemble-rcnn-for-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Gao_NOTE-RCNN_NOise_Tolerant_Ensemble_RCNN_for_Semi-Supervised_Object_Detection_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Gao_NOTE-RCNN_NOise_Tolerant_Ensemble_RCNN_for_Semi-Supervised_Object_Detection_ICCV_2019_paper.pdf | iccv-2019-10 | ['semi-supervised-object-detection'] | ['computer-vision'] | [ 2.12056279e-01 1.44586787e-01 -1.03356726e-01 -5.15993774e-01
-7.15087354e-01 -3.37374926e-01 3.17041963e-01 2.34470651e-01
-6.72620416e-01 6.14872634e-01 -3.02590668e-01 -1.61948770e-01
1.49485424e-01 -7.74959564e-01 -7.07556129e-01 -6.65704191e-01
5.58316670e-02 4.12280887e-01 8.44812274e-01 4.64057773e-02
-9.31576043e-02 3.04442704e-01 -1.81063008e+00 5.70684075e-01
9.26623046e-01 1.40305233e+00 3.17416117e-02 4.77620363e-01
-3.38122725e-01 7.27653623e-01 -8.26386690e-01 -2.90039301e-01
5.60752511e-01 -8.63144249e-02 -4.01091069e-01 1.01551212e-01
1.90274104e-01 -2.83757657e-01 2.36252904e-01 1.18952763e+00
4.66548353e-01 -1.17856348e-02 6.17673516e-01 -1.16083932e+00
-3.26417685e-01 6.62662745e-01 -9.29903388e-01 1.84244588e-01
-2.22884402e-01 1.06381834e-01 7.13333666e-01 -1.05261433e+00
3.61210942e-01 1.08687496e+00 7.99583673e-01 6.41377628e-01
-1.12951744e+00 -1.09829962e+00 3.64253223e-01 -1.09744109e-01
-1.53971839e+00 -2.57073820e-01 4.16641057e-01 -5.07621229e-01
4.88900185e-01 3.44432265e-01 1.97166786e-01 8.71844888e-01
-3.97280067e-01 7.89258957e-01 1.06107259e+00 -4.75632995e-01
3.38257104e-01 6.29992664e-01 4.74130869e-01 5.81733167e-01
6.02158666e-01 1.49500549e-01 -6.40286729e-02 -1.66669771e-01
3.86289656e-01 1.68726176e-01 5.32926098e-02 -3.70441854e-01
-7.85357535e-01 7.61479676e-01 6.31132722e-01 2.20781993e-02
4.43168804e-02 -1.21184684e-01 6.50491297e-01 -1.36331189e-02
5.32483041e-01 2.21432000e-01 -5.08827209e-01 3.66679519e-01
-1.00604141e+00 1.30882457e-01 5.16759038e-01 1.14405537e+00
7.98396170e-01 -1.40953586e-01 -3.69327873e-01 9.75425720e-01
3.04796278e-01 3.57293397e-01 3.08138132e-01 -3.15744728e-01
5.58750570e-01 1.05536306e+00 1.19102538e-01 -6.00409150e-01
-5.27224004e-01 -6.59733891e-01 -9.43512976e-01 3.51135433e-01
2.78402209e-01 -1.82292998e-01 -1.23736906e+00 1.18899012e+00
7.04848230e-01 -2.47458927e-02 -2.68371463e-01 9.31006730e-01
9.90307629e-01 4.51848030e-01 3.08550090e-01 -2.31676951e-01
1.44569147e+00 -8.93089056e-01 -7.21825182e-01 -1.37954116e-01
1.08140588e+00 -4.51571792e-01 8.59127164e-01 3.95144135e-01
-6.91652775e-01 -8.00929904e-01 -1.22411430e+00 1.57029796e-02
-5.88090777e-01 6.05052352e-01 4.51988399e-01 7.84744918e-01
-3.62792790e-01 4.65441525e-01 -6.94031835e-01 7.80793726e-02
1.01237261e+00 4.75035280e-01 -1.50903702e-01 7.60053191e-03
-7.72099853e-01 8.15542579e-01 8.50007534e-01 3.09932888e-01
-6.82878971e-01 -6.37988508e-01 -6.74098969e-01 -1.40339717e-01
9.41397071e-01 1.57982539e-02 1.16518569e+00 -1.01207948e+00
-1.00242591e+00 8.86140287e-01 6.29489198e-02 -6.36940897e-01
8.48941267e-01 -3.97119254e-01 -4.20376509e-01 -4.55971152e-01
1.73302621e-01 6.74055815e-01 8.74702215e-01 -1.34906042e+00
-1.08801937e+00 -4.88578916e-01 -2.83148140e-01 -1.13772951e-01
-3.03465843e-01 2.33601913e-01 -2.87548572e-01 -5.31605661e-01
1.92781001e-01 -6.70708358e-01 -3.85548830e-01 1.22276179e-01
-5.28213382e-01 -4.70287234e-01 1.27081990e+00 -2.32609794e-01
1.33822536e+00 -2.14819241e+00 -5.80031753e-01 3.74447316e-01
3.52452278e-01 5.45865834e-01 2.18870044e-01 -3.12981784e-01
-1.63565516e-01 2.59571761e-01 -3.16619635e-01 -4.33478743e-01
-3.33666265e-01 2.24414513e-01 -3.29614639e-01 5.20194352e-01
5.77092648e-01 6.28726423e-01 -7.54421353e-01 -7.55727291e-01
1.40284508e-01 1.10967711e-01 -4.04588878e-01 1.40061915e-01
-3.34778398e-01 1.68302894e-01 -3.40446681e-01 8.66827548e-01
9.02076364e-01 -2.10392296e-01 -1.97674394e-01 -2.16260076e-01
-1.73621565e-01 6.26775250e-02 -1.77694428e+00 9.73770916e-01
6.96838787e-03 7.55677894e-02 7.59536475e-02 -1.01245987e+00
1.26425266e+00 1.37536988e-01 2.05815002e-01 -3.28659505e-01
4.96308774e-01 2.76285857e-01 9.24749672e-02 -3.45671147e-01
1.48248255e-01 -2.67268848e-02 9.17669907e-02 2.47339129e-01
-6.16957806e-02 1.75052673e-01 3.05367470e-01 -5.77716902e-02
9.71210241e-01 2.74199322e-02 4.49310422e-01 -2.98270792e-01
6.86921716e-01 1.88057959e-01 9.73688662e-01 8.14519882e-01
-4.73673940e-01 5.50616324e-01 3.86319727e-01 -5.58716834e-01
-9.80049908e-01 -7.05793798e-01 -4.45583850e-01 1.41517127e+00
1.98709294e-01 -2.51516432e-01 -6.63697064e-01 -1.27624547e+00
8.62307325e-02 4.09283251e-01 -6.61408007e-01 -1.08765334e-01
-5.98788738e-01 -1.04830503e+00 6.28220975e-01 9.39690709e-01
6.46457911e-01 -9.95586634e-01 -4.78169948e-01 3.82548273e-02
2.47612506e-01 -1.26590335e+00 -1.86514467e-01 7.78538346e-01
-9.11319733e-01 -1.16438270e+00 -3.90547931e-01 -8.00963581e-01
8.77206385e-01 1.40144631e-01 9.01477098e-01 3.88126254e-01
-5.25628686e-01 -4.69316363e-01 -4.28401589e-01 -1.11628377e+00
-3.57996583e-01 -6.26139939e-02 9.68386978e-02 -9.30148643e-04
5.85936487e-01 1.63199473e-02 -5.04209280e-01 6.40805602e-01
-7.50505626e-01 -1.13641955e-02 6.54006600e-01 8.23526382e-01
7.89456308e-01 1.50773555e-01 7.11948931e-01 -1.31413770e+00
4.28013317e-02 -3.22898358e-01 -1.07471728e+00 -2.07133684e-03
-5.77494681e-01 -1.33176073e-01 6.75571918e-01 -7.23566175e-01
-9.48067605e-01 5.75663388e-01 -2.55291946e-02 -3.06015611e-01
-2.76067525e-01 -2.47921273e-01 -2.49032527e-01 -8.11329484e-02
9.76601779e-01 -9.89407226e-02 -3.34954649e-01 -6.10823333e-01
2.69632041e-01 8.88256729e-01 4.40506577e-01 -3.43596518e-01
1.05104673e+00 7.06624627e-01 -1.63790166e-01 -4.88381952e-01
-1.26603341e+00 -9.53461051e-01 -8.64489257e-01 -8.95434991e-02
7.81596541e-01 -9.56221402e-01 -5.13834894e-01 2.08912659e-02
-1.02130032e+00 -6.51642457e-02 -3.96889895e-01 1.85135871e-01
-5.32640293e-02 3.68152671e-02 -3.11270207e-01 -1.16926336e+00
-4.98323202e-01 -1.03563094e+00 1.03263664e+00 3.39685977e-01
8.45239032e-03 -2.38575876e-01 -2.08417684e-01 3.01413536e-01
5.06422929e-02 2.15607390e-01 6.26723707e-01 -1.13270295e+00
-5.20913363e-01 -5.40214181e-01 -6.84320569e-01 7.42524743e-01
-1.31278038e-01 -1.07187778e-01 -1.25100052e+00 -1.23396091e-01
-1.28192484e-01 -4.27432775e-01 1.08122182e+00 5.08312136e-02
1.41390896e+00 -1.74680322e-01 -5.90016127e-01 4.68144119e-01
1.38052738e+00 1.48873717e-01 4.23191935e-01 2.82658339e-01
6.56063616e-01 5.62714398e-01 7.86901474e-01 3.41296822e-01
-1.01999037e-01 3.70321244e-01 4.02973682e-01 -3.00222456e-01
6.55177906e-02 -1.94248065e-01 1.54586568e-01 1.64623201e-01
-5.23097217e-02 1.14721812e-01 -9.86857831e-01 4.64702427e-01
-1.93464434e+00 -5.34530759e-01 -4.89738524e-01 2.14727211e+00
9.22749817e-01 6.06312871e-01 4.10367906e-01 5.40807068e-01
9.56543207e-01 -2.74514586e-01 -5.27198136e-01 -2.61408451e-04
-6.53887391e-02 2.34360695e-01 7.58046508e-01 -7.32073337e-02
-1.66055191e+00 8.62818003e-01 5.62636423e+00 9.19208169e-01
-7.21479535e-01 3.72093648e-01 7.21124828e-01 -1.59710422e-01
5.81299484e-01 -6.68148324e-02 -1.40099180e+00 3.79740536e-01
5.17795801e-01 5.28080642e-01 -2.10264519e-01 1.44212854e+00
1.97636962e-01 -2.46265247e-01 -1.14072883e+00 9.04463828e-01
-4.52656336e-02 -1.24375808e+00 -1.89520270e-01 -1.53907210e-01
8.56434226e-01 -7.89153576e-02 -2.90024668e-01 7.01230645e-01
3.94794792e-01 -8.97971570e-01 8.30450952e-01 -6.69096317e-03
5.79255044e-01 -6.76136792e-01 1.00064623e+00 7.93705583e-01
-1.23951793e+00 -3.82920265e-01 -6.10800266e-01 4.50946316e-02
-1.63803279e-01 9.63499308e-01 -7.89425671e-01 2.71480501e-01
1.01566422e+00 1.68800801e-01 -7.35780180e-01 1.22645450e+00
-1.83790550e-01 5.71758032e-01 -5.21870553e-01 -1.97628081e-01
1.25552118e-01 1.44168049e-01 2.63636261e-01 1.47567570e+00
-2.75619656e-01 1.86538830e-01 5.29964209e-01 9.71387982e-01
-2.64070004e-01 1.91189513e-01 -4.21363711e-01 5.12637198e-01
4.48599368e-01 1.49453509e+00 -1.13260043e+00 -5.48123658e-01
-3.44797403e-01 5.07692993e-01 4.29150581e-01 -4.20564078e-02
-1.06107736e+00 -3.39263916e-01 1.47533134e-01 3.53147060e-01
4.17929471e-01 1.32664323e-01 -8.99721622e-01 -7.18705595e-01
2.03287870e-01 -6.28014028e-01 5.21903217e-01 -2.56604612e-01
-1.20485973e+00 3.44607383e-01 -1.10684708e-01 -1.24182057e+00
5.10173619e-01 -7.06976652e-01 -6.80864096e-01 5.35045087e-01
-1.49965084e+00 -1.02329159e+00 -5.22566974e-01 2.68588305e-01
4.94579583e-01 8.23298767e-02 3.96300703e-01 4.99781102e-01
-9.51435566e-01 7.72289634e-01 -4.09151688e-02 6.35574877e-01
6.91480339e-01 -1.19455671e+00 1.15882047e-01 8.20061982e-01
4.31270190e-02 3.41143340e-01 3.67431194e-01 -7.90553451e-01
-7.85044909e-01 -1.73425674e+00 3.64010364e-01 -5.13073266e-01
4.11086679e-01 -8.27833891e-01 -1.09762466e+00 4.91146475e-01
-7.54052818e-01 6.90605044e-01 4.21483248e-01 2.51458716e-02
-5.04879653e-01 -2.11562574e-01 -1.24147201e+00 2.92319834e-01
1.09967446e+00 -8.04960281e-02 -4.65006799e-01 5.64351201e-01
7.25512505e-01 -4.00725633e-01 -4.51179117e-01 8.49240780e-01
3.38179469e-01 -8.79995883e-01 7.59196818e-01 -5.34722567e-01
1.15068927e-01 -7.28493452e-01 1.25722423e-01 -6.10257089e-01
-1.34223148e-01 -2.37894788e-01 -2.36409351e-01 1.40517139e+00
5.96695900e-01 -3.33808899e-01 1.08930099e+00 4.19544160e-01
-2.79575381e-02 -7.71710098e-01 -8.00223112e-01 -1.06108916e+00
-3.40169430e-01 -6.24301791e-01 3.94517124e-01 6.14239156e-01
-3.26995641e-01 3.93563569e-01 -1.61963239e-01 2.28190705e-01
6.18838608e-01 -1.62808210e-01 9.49464440e-01 -1.47777843e+00
-3.50623317e-02 -1.93165496e-01 -3.58130783e-01 -9.54114556e-01
-2.77847826e-01 -7.96450019e-01 5.62700808e-01 -1.01891935e+00
3.64783168e-01 -8.61798346e-01 -2.24558994e-01 6.64614975e-01
-3.40859890e-01 4.77825791e-01 8.14092308e-02 3.10900152e-01
-9.77124631e-01 3.83083016e-01 9.10463810e-01 1.58145819e-02
-4.06227797e-01 2.94757009e-01 -4.79477644e-01 1.19365513e+00
5.34121573e-01 -9.19053793e-01 -3.74566726e-02 -2.89525557e-02
-2.34579518e-02 -5.61976373e-01 3.81664395e-01 -1.07500815e+00
3.03308487e-01 7.95580596e-02 6.74367428e-01 -1.01739299e+00
-1.21687129e-01 -9.77540612e-01 -4.53085691e-01 4.54194903e-01
-2.00955302e-01 -3.97315562e-01 1.35663629e-01 7.71736026e-01
4.72090803e-02 -3.48363042e-01 1.32000458e+00 -6.40982389e-02
-5.48627973e-01 1.94672644e-01 3.94353010e-02 9.65594649e-02
1.40802240e+00 -4.05896962e-01 -2.17239484e-01 4.30975616e-01
-5.58452368e-01 5.47777116e-01 -2.28086859e-02 2.91410714e-01
4.04282302e-01 -1.20621896e+00 -5.87490976e-01 3.48902315e-01
3.84597212e-01 6.96705341e-01 -1.36433408e-01 7.61050403e-01
-2.83118814e-01 1.58044547e-01 2.95470029e-01 -8.21833849e-01
-1.41202319e+00 6.19695902e-01 2.67566741e-01 -4.09149319e-01
-6.74880862e-01 1.08661377e+00 3.31364363e-01 -4.38085586e-01
7.32489526e-01 -5.08196652e-01 -2.79285192e-01 2.32378729e-02
8.19158792e-01 5.30502319e-01 3.50085765e-01 -3.03654701e-01
-4.00927305e-01 2.60929972e-01 -4.44422990e-01 5.09327769e-01
1.08541811e+00 1.92516416e-01 1.64724812e-01 4.05134827e-01
7.66414404e-01 -2.94940501e-01 -1.18999660e+00 -3.64372015e-01
5.64981997e-01 -1.93409503e-01 9.72873867e-02 -7.63989389e-01
-6.87294483e-01 6.79483473e-01 8.86879385e-01 1.69639528e-01
9.49402869e-01 9.76761281e-02 6.54662311e-01 3.98207158e-01
1.70058668e-01 -1.39534736e+00 6.32339716e-03 4.47854638e-01
6.28579557e-01 -1.64643812e+00 2.11288199e-01 -6.97540164e-01
-5.23047864e-01 8.89433503e-01 1.13265610e+00 -3.26284647e-01
7.79840350e-01 6.38574123e-01 -5.74219823e-02 -9.44321752e-02
-5.49441934e-01 -5.45012355e-01 4.04348463e-01 6.17176056e-01
1.54775307e-01 -2.73147151e-02 -1.95943221e-01 1.08979464e+00
7.30928257e-02 -1.01115018e-01 9.65452492e-02 9.56922412e-01
-8.80639255e-01 -6.94428504e-01 -7.30031371e-01 8.62700582e-01
-5.15006006e-01 5.83659597e-02 -4.15083259e-01 1.01487601e+00
8.58085632e-01 7.79493630e-01 7.88818672e-02 -3.02220851e-01
6.21753216e-01 1.68310344e-01 -3.29348259e-02 -1.12463331e+00
-7.56749451e-01 1.69169419e-02 -6.49382994e-02 -4.87428606e-01
-2.81263977e-01 -2.77119637e-01 -1.51128817e+00 1.35077089e-01
-1.21383381e+00 -8.86057764e-02 5.34359097e-01 1.08073914e+00
-3.65689509e-02 5.53133488e-01 3.79103839e-01 -6.23845637e-01
-8.48658502e-01 -1.23812282e+00 -4.71863568e-01 4.28165942e-01
3.49926412e-01 -8.64078403e-01 -3.59253168e-01 1.08103909e-01] | [9.156103134155273, 1.1982804536819458] |
71d4a067-8033-4fe1-b743-64a464d480f0 | analyzing-multiple-choice-reading-and | 2307.01076 | null | https://arxiv.org/abs/2307.01076v1 | https://arxiv.org/pdf/2307.01076v1.pdf | Analyzing Multiple-Choice Reading and Listening Comprehension Tests | Multiple-choice reading and listening comprehension tests are an important part of language assessment. Content creators for standard educational tests need to carefully curate questions that assess the comprehension abilities of candidates taking the tests. However, recent work has shown that a large number of questions in general multiple-choice reading comprehension datasets can be answered without comprehension, by leveraging world knowledge instead. This work investigates how much of a contextual passage needs to be read in multiple-choice reading based on conversation transcriptions and listening comprehension tests to be able to work out the correct answer. We find that automated reading comprehension systems can perform significantly better than random with partial or even no access to the context passage. These findings offer an approach for content creators to automatically capture the trade-off between comprehension and world knowledge required for their proposed questions. | ['Mark Gales', 'Adian Liusie', 'Vatsal Raina'] | 2023-07-03 | null | null | null | null | ['reading-comprehension'] | ['natural-language-processing'] | [ 4.68980104e-01 5.30401409e-01 1.26486123e-01 -5.40291786e-01
-1.24280894e+00 -1.23976898e+00 3.90952796e-01 8.32332194e-01
-3.87523443e-01 5.36766589e-01 5.38904071e-01 -1.09420836e+00
-5.83719492e-01 -1.04638088e+00 -4.72283244e-01 1.78111225e-01
5.87805092e-01 5.41253507e-01 4.53953832e-01 -6.06002390e-01
5.74951410e-01 -7.64843896e-02 -1.78998566e+00 3.23277175e-01
1.36412168e+00 6.09405935e-01 4.38734263e-01 1.29398799e+00
-5.81877828e-01 1.02594757e+00 -1.05478501e+00 -5.11168778e-01
-1.98965266e-01 -1.03684652e+00 -1.30784285e+00 -3.05312663e-01
8.84217143e-01 -3.11985254e-01 1.87245503e-01 7.83269227e-01
4.33835924e-01 5.69262862e-01 4.92013484e-01 -4.60255682e-01
-3.70800227e-01 9.37126458e-01 3.76364976e-01 5.28234124e-01
1.48296130e+00 9.05335993e-02 1.06268239e+00 -2.95638233e-01
3.88184458e-01 9.20739174e-01 2.25847632e-01 4.50192779e-01
-7.15627849e-01 -3.15105200e-01 -1.37017146e-01 5.22552490e-01
-7.30041981e-01 -3.96572560e-01 5.64057052e-01 -3.74830842e-01
6.32383585e-01 6.22423768e-01 8.89772773e-01 1.01225138e+00
1.50750786e-01 4.74956036e-01 1.42908037e+00 -8.28377545e-01
1.38481766e-01 1.72414303e-01 5.77228725e-01 6.35047376e-01
1.65629759e-01 -5.96445739e-01 -7.40542650e-01 7.43963122e-02
2.27799863e-01 -5.34642339e-01 -4.36636239e-01 3.49187821e-01
-1.04049969e+00 6.08630419e-01 -3.05499971e-01 4.13702548e-01
7.09359646e-02 -5.18330097e-01 8.07672888e-02 9.37794507e-01
-1.64827928e-01 1.28636193e+00 -6.26782298e-01 -1.02997160e+00
-1.06318772e+00 3.72564822e-01 1.52231812e+00 8.42918694e-01
3.25180024e-01 -7.57587194e-01 -6.84241354e-01 7.88604736e-01
8.97094011e-02 6.11815572e-01 7.47738183e-01 -9.41608727e-01
6.45676672e-01 8.51320028e-01 -2.76596006e-02 -9.00643349e-01
-1.15613833e-01 -6.45197928e-01 -1.18123271e-01 -1.27346173e-01
1.11542296e+00 -2.84283608e-01 -4.34204757e-01 1.51042235e+00
1.43167689e-01 -1.36496976e-01 -4.06951010e-02 4.56958503e-01
1.05374765e+00 2.92865008e-01 -7.06742778e-02 -3.73067372e-02
1.75165999e+00 -7.45099783e-01 -8.04948986e-01 -2.13763177e-01
6.89349949e-01 -1.15462196e+00 1.69913340e+00 6.10125184e-01
-1.71650851e+00 -7.30502725e-01 -1.10266793e+00 -3.28076482e-01
-2.16554329e-01 -3.40475470e-01 -1.19532183e-01 1.22043860e+00
-7.50961125e-01 3.54218513e-01 -1.01187631e-01 -1.24057315e-01
-3.18358243e-02 -9.43693146e-02 1.04209431e-01 -5.17156065e-01
-1.26151752e+00 8.92382264e-01 -1.50051937e-01 -3.30277652e-01
-6.54556930e-01 -8.46604824e-01 -8.18609416e-01 4.10002947e-01
6.14764988e-01 -6.22870862e-01 1.89075518e+00 -7.08171189e-01
-1.94406068e+00 7.28906870e-01 -3.69779766e-01 1.65482670e-01
5.00775635e-01 -4.09650594e-01 -5.00872321e-02 3.38756412e-01
1.38039500e-01 1.49653435e-01 3.50047052e-01 -3.05171669e-01
-6.14132345e-01 6.09444007e-02 4.74581391e-01 5.15775502e-01
-3.52256536e-01 -1.12797581e-01 -2.55020142e-01 -4.03890133e-01
2.92867690e-01 -5.29960752e-01 3.01097721e-01 -4.88818496e-01
-3.52190912e-01 -6.14384234e-01 4.25849766e-01 -9.94940758e-01
1.49210227e+00 -1.56114376e+00 -1.84606239e-02 4.52912629e-01
3.09194773e-01 7.11175799e-02 -2.80975491e-01 4.81418401e-01
2.39936039e-01 4.56810415e-01 3.51760596e-01 1.28769830e-01
4.68901604e-01 -3.80751818e-01 -1.60160512e-01 -1.87834024e-01
-3.83000791e-01 8.57595325e-01 -1.10832274e+00 -2.63498366e-01
9.92971435e-02 -1.89151958e-01 -4.81259018e-01 7.79138505e-01
-5.69647372e-01 5.35756707e-01 -3.86400670e-01 5.01555920e-01
1.35776579e-01 -8.69638324e-02 -1.27786934e-01 7.72889793e-01
7.45204017e-02 1.16620624e+00 -9.11799133e-01 1.69443142e+00
-8.64937246e-01 9.27176833e-01 -2.65771031e-01 -4.84631628e-01
7.81427383e-01 3.10621262e-01 -2.62020588e-01 -1.13290310e+00
1.99596003e-01 5.25060482e-02 4.63872343e-01 -9.05083597e-01
3.54612231e-01 9.22721401e-02 -1.45683706e-01 9.89027917e-01
-9.66446251e-02 -3.39092046e-01 6.70403123e-01 2.77908325e-01
1.43647087e+00 -7.72043690e-03 2.70912737e-01 -2.49393821e-01
7.13930845e-01 -2.61208992e-02 -9.25714076e-02 1.35400689e+00
3.65051553e-02 4.79987502e-01 5.13499558e-01 1.20630741e-01
-5.33492267e-01 -8.65700722e-01 3.11932981e-01 1.54262996e+00
-1.75523236e-01 -6.10973418e-01 -1.14127874e+00 -4.57315981e-01
-7.68745303e-01 1.21813965e+00 -2.74882227e-01 -2.86047757e-01
-4.38508958e-01 6.01801835e-02 6.69357598e-01 2.73472399e-01
6.06845498e-01 -8.45412016e-01 -9.09699798e-01 3.63852262e-01
-7.50403523e-01 -9.22668457e-01 -3.47074181e-01 -1.79771826e-01
-4.70508218e-01 -1.06641459e+00 -4.45171148e-01 -8.40578616e-01
4.00177926e-01 2.09965244e-01 1.50689638e+00 6.50025547e-01
2.16560781e-01 1.00235760e+00 -7.23323524e-01 -5.04923999e-01
-6.67359412e-01 4.64988530e-01 -6.72677517e-01 -6.59144878e-01
7.13040233e-01 -5.11938512e-01 -5.43603599e-01 4.24376458e-01
-7.36166000e-01 2.18019888e-01 5.46028435e-01 4.29767638e-01
1.69824257e-01 7.46157253e-03 4.19198900e-01 -7.92728424e-01
1.35884094e+00 -4.62287307e-01 -4.16820675e-01 6.90915883e-01
-2.78669208e-01 6.87332824e-02 6.05895400e-01 -4.48104113e-01
-1.06174290e+00 -6.45424664e-01 -5.79043150e-01 5.15230715e-01
-6.30845070e-01 8.02389979e-01 -3.06335777e-01 -4.40845601e-02
7.34757245e-01 2.82536596e-01 -2.36554667e-01 -3.13270271e-01
5.12819216e-02 4.30187374e-01 3.15805316e-01 -9.02411044e-01
8.02668631e-01 -4.59650934e-01 -2.25890741e-01 -8.59670162e-01
-1.32826602e+00 -6.12335086e-01 -6.21132195e-01 -5.88861465e-01
7.81339943e-01 -5.05250335e-01 -9.72840488e-01 -1.01515418e-03
-8.36635590e-01 -6.07870281e-01 -3.54586035e-01 4.20794994e-01
-4.04125273e-01 1.64764002e-01 -3.52151662e-01 -7.26272225e-01
1.01878501e-01 -1.18689167e+00 5.23454785e-01 6.72653377e-01
-9.64618504e-01 -1.13507450e+00 1.57156855e-01 1.14223397e+00
5.24813771e-01 -2.85457313e-01 1.33424687e+00 -1.19372952e+00
-5.64868093e-01 -3.46467227e-01 4.03256387e-01 1.20489426e-01
-2.69933254e-01 -2.56718487e-01 -9.29851413e-01 5.86501174e-02
2.00833097e-01 -5.08937299e-01 2.99443007e-01 1.19659916e-01
1.02413690e+00 -4.26970780e-01 2.31484681e-01 -1.59218181e-02
9.07079041e-01 -8.12880322e-02 6.80421591e-01 3.32465410e-01
2.24002481e-01 1.06752968e+00 4.91135299e-01 3.73761579e-02
6.28296852e-01 3.11765224e-01 -2.03598350e-01 5.57837367e-01
-2.92243123e-01 -6.90014839e-01 5.16428351e-01 1.29659808e+00
2.96157002e-01 -4.95653123e-01 -1.16240406e+00 6.71423912e-01
-1.11152899e+00 -1.04965532e+00 -4.12510723e-01 2.22580767e+00
1.00423670e+00 3.67877632e-01 -2.60181110e-02 4.36071754e-01
9.97413471e-02 -2.85525352e-01 -1.85999293e-02 -5.67292809e-01
1.90500170e-02 8.14596891e-01 -4.97840233e-02 8.77180219e-01
-2.87342608e-01 5.81902742e-01 6.78343391e+00 8.91057968e-01
-6.00792110e-01 -6.39706850e-02 4.64078188e-01 2.02031001e-01
-7.97973156e-01 1.05062217e-01 -7.94015348e-01 1.96504533e-01
1.47286630e+00 -2.32657492e-01 3.18563968e-01 2.24133849e-01
2.79413879e-01 -6.63334489e-01 -1.17134428e+00 4.59707707e-01
2.89594322e-01 -1.04458332e+00 -2.54192799e-02 -2.75046051e-01
5.99221885e-01 -6.83244109e-01 -8.59496146e-02 4.67567205e-01
2.44083658e-01 -1.34086144e+00 5.40183663e-01 8.11949372e-01
3.85199368e-01 -4.78400797e-01 7.16169953e-01 6.60344243e-01
-7.76985466e-01 -1.50077298e-01 -9.83058810e-02 -7.66495228e-01
-2.09654912e-01 3.07766676e-01 -1.17316711e+00 2.49045908e-01
3.10556889e-01 -3.07275206e-02 -1.13307142e+00 1.40377831e+00
-7.72476852e-01 1.27244127e+00 -2.61255533e-01 -7.84380138e-01
-7.18600228e-02 -2.25329287e-02 4.81129825e-01 8.11934769e-01
7.20397413e-01 4.81335402e-01 2.13542640e-01 6.14254177e-01
-1.66902646e-01 5.38753271e-01 -2.14540303e-01 -4.51860242e-02
6.92467153e-01 7.79350400e-01 -6.37454331e-01 -2.94953734e-01
-4.32312280e-01 6.02729678e-01 1.14711039e-01 2.92849123e-01
-2.27668136e-01 -6.17100358e-01 3.10051113e-01 4.08677518e-01
-1.37314901e-01 -2.58367509e-01 -5.73844194e-01 -8.89300227e-01
2.03961611e-01 -1.47098064e+00 3.90385717e-01 -9.20497298e-01
-9.67375219e-01 1.57378465e-01 -7.36079514e-02 -1.02295351e+00
-2.93719530e-01 -5.63747704e-01 -1.04531193e+00 9.34356511e-01
-1.43112707e+00 -4.05511349e-01 -7.02712357e-01 4.36862558e-01
9.23706055e-01 7.47329593e-02 8.14951599e-01 -9.62326378e-02
-1.68012977e-01 7.48530507e-01 -3.62615675e-01 5.27558364e-02
8.37937057e-01 -1.52254438e+00 -9.74592865e-02 8.37176740e-01
2.31968045e-01 5.66677332e-01 9.08589721e-01 -6.06630504e-01
-1.37013769e+00 -2.43978471e-01 1.41147816e+00 -8.04578722e-01
4.51673210e-01 -1.21673845e-01 -1.02875543e+00 2.55197525e-01
7.56984651e-01 -1.08362949e+00 1.43184412e+00 3.82843822e-01
-1.73292086e-01 2.11370543e-01 -7.55989552e-01 6.48850501e-01
7.21281111e-01 -6.87345922e-01 -1.44267213e+00 3.02026898e-01
6.82679832e-01 -6.06070220e-01 -9.47922766e-01 -1.68010011e-01
5.43239653e-01 -8.18102598e-01 5.64483762e-01 -3.56036633e-01
6.64810598e-01 -6.35660887e-02 1.66183352e-01 -1.25535429e+00
1.23854756e-01 -7.70184636e-01 2.31593966e-01 1.14986229e+00
6.90632463e-01 -2.59995878e-01 3.74014825e-01 9.88982677e-01
6.97549954e-02 -3.25336576e-01 -6.39901638e-01 -2.98735082e-01
4.43888247e-01 -7.26421356e-01 7.14585125e-01 6.61055684e-01
5.07505476e-01 4.99743611e-01 4.05124605e-01 3.75692286e-02
2.51660854e-01 8.69466737e-03 8.11958730e-01 -1.39813519e+00
-3.10342014e-01 -7.72993982e-01 -3.69309708e-02 -1.47902286e+00
1.03097163e-01 -7.15718210e-01 -2.38360837e-02 -1.46705568e+00
-4.42161858e-02 -1.67290273e-03 2.45910034e-01 8.58595446e-02
-5.83061516e-01 -1.86700061e-01 5.99520095e-02 -3.76032829e-01
-8.50762486e-01 -2.10030377e-02 1.61456335e+00 -2.49212477e-02
-1.16184682e-01 3.11630905e-01 -1.08826792e+00 6.30904615e-01
7.44300902e-01 -1.48237690e-01 -7.58645177e-01 -5.60592830e-01
9.13934648e-01 4.41092283e-01 1.45237684e-01 -1.17555249e+00
8.38383794e-01 -3.56658101e-01 4.01481390e-01 -3.81857008e-01
-1.42559677e-01 -4.82741654e-01 -4.79542255e-01 -6.04302064e-02
-1.11543286e+00 4.05302405e-01 1.89627320e-01 2.16903165e-01
-4.82444167e-01 -9.53960955e-01 1.77088261e-01 -3.18005890e-01
-3.28709781e-01 -4.22908932e-01 -1.00173736e+00 4.88504291e-01
6.36148095e-01 -3.29456329e-01 -3.13679099e-01 -1.10662901e+00
-7.48085439e-01 4.00886297e-01 7.33909756e-02 2.48090550e-01
5.61215460e-01 -6.89042747e-01 -8.91046822e-01 8.24571252e-02
-4.71909717e-02 -7.91048929e-02 1.98414758e-01 5.96956730e-01
-6.45106733e-01 5.87710917e-01 -6.94780424e-02 -3.18669617e-01
-1.47788215e+00 -1.48416653e-01 2.27317244e-01 -5.33828735e-01
-1.83375522e-01 1.00720847e+00 -4.97275978e-01 -4.40403849e-01
4.10781443e-01 -5.10123789e-01 -6.14325166e-01 2.67707109e-01
1.13676262e+00 4.10539031e-01 4.15343434e-01 4.48374152e-02
3.76825213e-01 4.46405232e-01 1.29025787e-01 -4.42861736e-01
6.53447509e-01 -4.17812794e-01 1.14696994e-01 5.80340147e-01
7.97986150e-01 7.33329594e-01 -5.10708153e-01 -3.16488743e-02
2.51866281e-01 -3.97471189e-01 -1.18901916e-01 -1.49142075e+00
-1.46960065e-01 1.10733628e+00 1.69905126e-01 5.51358581e-01
1.06826949e+00 -1.45780995e-01 7.58618116e-01 8.97162795e-01
4.51698229e-02 -1.21456409e+00 2.25645140e-01 8.62400651e-01
6.19724810e-01 -1.04745340e+00 -2.62546718e-01 -4.02439058e-01
-1.30737439e-01 1.30493665e+00 9.31014717e-01 4.90744233e-01
2.67248273e-01 -1.51714161e-01 1.09356992e-01 -1.91423118e-01
-8.42223585e-01 -2.27851093e-01 6.55625701e-01 4.06303912e-01
8.63964260e-01 -4.07958254e-02 -4.54791665e-01 8.08483779e-01
-9.86086786e-01 -4.64630038e-01 9.83548224e-01 1.04640639e+00
-7.60690808e-01 -1.29100990e+00 -6.17513299e-01 5.77565312e-01
-3.97104919e-01 -3.91684175e-01 -8.90809953e-01 3.14012915e-01
-2.72412121e-01 1.51897311e+00 -9.27416533e-02 -7.89270028e-02
5.03588140e-01 6.20878160e-01 6.35444105e-01 -9.48176086e-01
-1.07143700e+00 -4.08456266e-01 2.32918069e-01 -2.41271883e-01
-9.05304626e-02 -5.95759213e-01 -8.24550152e-01 -3.13053906e-01
-3.02993685e-01 5.98161876e-01 3.19075286e-01 1.43667364e+00
6.00347063e-03 7.38123298e-01 3.63082178e-02 7.21534565e-02
-5.90722740e-01 -1.15358889e+00 9.81900766e-02 3.60598028e-01
2.97301501e-01 -9.79826599e-03 -2.95451343e-01 -2.21523065e-02] | [11.453063011169434, 8.064873695373535] |
c5d855e3-e91f-4bb7-a60d-cc472a156e86 | abcde-approximating-betweenness-centrality | null | null | https://peerj.com/articles/cs-699/ | https://peerj.com/articles/cs-699.pdf | ABCDE: Approximating Betweenness-Centrality ranking with progressive-DropEdge | Betweenness-centrality is a popular measure in network analysis that aims to describe the importance of nodes in a graph. It accounts for the fraction of shortest paths passing through that node and is a key measure in many applications including community detection and network dismantling. The computation of betweenness-centrality for each node in a graph requires an excessive amount of computing power, especially for large graphs. On the other hand, in many applications, the main interest lies in finding the top-k most important nodes in the graph. Therefore, several approximation algorithms were proposed to solve the problem faster. Some recent approaches propose to use shallow graph convolutional networks to approximate the top-k nodes with the highest betweenness-centrality scores.
This work presents a deep graph convolutional neural network that outputs a rank score for each node in a given graph. With careful optimization and regularization tricks, including an extended version of DropEdge which is named Progressive-DropEdge, the system achieves better results than the current approaches. Experiments on both real-world and synthetic datasets show that the presented algorithm is an order of magnitude faster in inference and requires several times fewer resources and time to train. | ['Mirakyan Martin'] | 2021-09-06 | null | null | null | peerj-computer-science-2021-9 | ['approximating-betweenness-centrality-ranking'] | ['graphs'] | [-2.90558279e-01 1.85656488e-01 -2.06565578e-02 -2.06883118e-01
-1.15983590e-01 -3.74811798e-01 3.09984118e-01 7.82939017e-01
-3.60892206e-01 6.36750817e-01 -1.29934505e-01 -3.87295425e-01
-5.10146439e-01 -1.32701826e+00 -5.00100076e-01 -6.13963187e-01
-6.65023267e-01 6.73729122e-01 3.89417440e-01 -1.98927253e-01
2.66502440e-01 5.34751892e-01 -7.96067476e-01 -1.06808990e-01
7.52806962e-01 8.06359947e-01 -2.33901199e-02 6.68225110e-01
-1.02506287e-01 8.22145939e-01 -4.81067717e-01 -6.52147412e-01
1.25746042e-01 -1.08327411e-01 -8.57222617e-01 -2.78874457e-01
2.39377007e-01 -2.57838398e-01 -8.10720384e-01 1.10615647e+00
4.79296714e-01 -4.86121736e-02 4.47983146e-01 -1.38135397e+00
-4.18522954e-01 1.07127178e+00 -7.80090213e-01 4.42457676e-01
2.19872728e-01 -2.55226254e-01 1.43389940e+00 -5.19382834e-01
2.52262384e-01 1.16396081e+00 7.92703509e-01 -8.18154365e-02
-1.07418704e+00 -6.65430725e-01 1.31389737e-01 2.55823284e-01
-1.66046453e+00 1.52007237e-01 8.30396354e-01 -2.40242988e-01
7.67720819e-01 2.38935962e-01 7.31507778e-01 3.96322668e-01
1.44344568e-01 4.23542649e-01 4.93604362e-01 1.54423714e-02
3.09226452e-03 -3.00079793e-01 1.70612976e-01 1.03456998e+00
5.61336577e-01 -4.33385253e-01 -1.10274784e-01 -3.41348976e-01
7.65397549e-01 1.19732685e-01 -1.76109120e-01 -2.24948928e-01
-1.21364975e+00 9.26772833e-01 1.13898242e+00 3.49577188e-01
-3.47655565e-01 6.22547328e-01 5.79137921e-01 3.11050683e-01
5.67135155e-01 3.44022989e-01 -1.87236920e-01 2.45148227e-01
-8.53563547e-01 2.37068888e-02 1.00134814e+00 7.77894139e-01
7.99374282e-01 -3.64810139e-01 -2.00690314e-01 4.61473376e-01
1.89827427e-01 1.03206113e-01 -2.22486138e-01 -6.30887985e-01
5.39426208e-01 1.18325782e+00 -3.11218947e-01 -1.70038593e+00
-6.08124852e-01 -7.17014134e-01 -1.57498169e+00 -1.28208444e-01
3.81570727e-01 -2.71902606e-02 -5.60501039e-01 1.59156108e+00
3.27519804e-01 4.68246758e-01 -5.15723109e-01 8.37834418e-01
1.00128722e+00 6.84825957e-01 -8.49572197e-02 7.80700147e-02
1.23984361e+00 -8.58678102e-01 -2.34147713e-01 -7.53529966e-02
4.63883221e-01 -4.07685786e-01 7.13617802e-01 5.36685139e-02
-7.53463030e-01 -1.28967002e-01 -8.11679006e-01 1.01138223e-02
-3.59706581e-01 -1.26138523e-01 9.87971008e-01 3.49915266e-01
-1.26643479e+00 1.04540908e+00 -5.61188281e-01 -3.45001489e-01
3.18774819e-01 4.39964890e-01 -5.35632730e-01 -7.45831653e-02
-1.26142526e+00 3.22190940e-01 5.48620284e-01 4.11961168e-01
-6.94353700e-01 -4.69276577e-01 -7.39646316e-01 6.43787622e-01
5.38875461e-01 -5.62857747e-01 5.72531939e-01 -7.11885035e-01
-8.77701700e-01 5.92408478e-01 3.18825617e-02 -3.13743562e-01
3.41963410e-01 2.23926201e-01 -2.06208393e-01 1.65288806e-01
9.39559117e-02 1.38863519e-01 4.59085733e-01 -5.52645743e-01
-3.66925418e-01 -2.01386377e-01 4.76813704e-01 -1.26662895e-01
-6.14431918e-01 4.60625533e-03 -7.09525108e-01 -4.54844505e-01
2.76191175e-01 -8.10092270e-01 -4.55320150e-01 3.44443228e-03
-7.60641813e-01 -7.36121297e-01 4.17119920e-01 -5.49703717e-01
1.49141359e+00 -1.68640542e+00 1.65481597e-01 6.96865499e-01
1.22234690e+00 2.77052164e-01 -1.93782315e-01 8.12336802e-01
-4.38298248e-02 3.06724429e-01 -9.03993621e-02 1.33722007e-01
-1.98051944e-01 -1.00652233e-01 1.00394733e-01 5.80686271e-01
2.35065192e-01 8.64502370e-01 -1.19595206e+00 -4.91335690e-01
2.69002523e-02 5.11623919e-01 -5.59971988e-01 1.08141012e-01
-5.40551394e-02 2.14011282e-01 -4.96741980e-01 2.76408464e-01
7.64561236e-01 -9.65276062e-01 5.21311939e-01 -1.34392336e-01
1.97640747e-01 2.95680344e-01 -1.27032447e+00 1.14697266e+00
-1.36277705e-01 7.71832526e-01 1.00235157e-01 -1.16092336e+00
7.87494719e-01 1.18143260e-01 5.17649114e-01 -3.86608750e-01
2.28477865e-01 1.60630673e-01 3.00404519e-01 -2.22129956e-01
3.52550745e-01 4.09025222e-01 -5.48434220e-02 5.63305736e-01
-2.19582707e-01 4.96915728e-01 7.36376524e-01 9.41603303e-01
1.63260245e+00 -7.45896041e-01 2.28623614e-01 -3.22142839e-01
7.34855711e-01 -4.92385328e-01 5.46325207e-01 5.11985004e-01
-2.62810849e-02 2.57438272e-01 1.23591709e+00 -6.82844818e-01
-9.76862311e-01 -8.76712441e-01 3.94858688e-01 9.41934288e-01
1.65165827e-01 -6.66662037e-01 -6.57286167e-01 -4.50303704e-01
9.63611305e-02 -1.10064916e-01 -6.11819327e-01 -1.28036097e-01
-5.69336534e-01 -5.72822571e-01 4.91801202e-01 4.86121595e-01
4.40374315e-01 -9.18730497e-01 1.36271462e-01 2.54861683e-01
-3.34198743e-01 -1.08281589e+00 -6.11245871e-01 -3.15064639e-01
-8.50126505e-01 -1.59005761e+00 -6.79124236e-01 -1.00894082e+00
1.10723817e+00 4.23358947e-01 1.35228014e+00 8.99170339e-01
-1.65484741e-01 -1.47222444e-01 -3.53098661e-01 2.47015864e-01
-1.56567842e-01 4.21237916e-01 -1.17886774e-01 2.67154053e-02
2.65661448e-01 -7.96049356e-01 -7.29535997e-01 4.00739312e-02
-6.27537370e-01 -1.87508110e-02 6.92955017e-01 6.93260908e-01
3.14970225e-01 5.05632639e-01 3.82277995e-01 -1.13976586e+00
9.89733398e-01 -5.05269587e-01 -7.12781429e-01 2.34774888e-01
-6.17799282e-01 1.96939453e-01 9.01371777e-01 -1.11006245e-01
-1.57892600e-01 -2.80424386e-01 -2.44895462e-02 -1.70400828e-01
4.29226607e-01 8.73751998e-01 -2.35164468e-03 -1.63424268e-01
2.15060100e-01 9.14914012e-02 -2.39903092e-01 -3.31460863e-01
1.51715860e-01 3.37678701e-01 1.13798074e-01 -3.17439049e-01
8.60536456e-01 3.68188828e-01 4.22147691e-01 -8.17855716e-01
-6.55489564e-01 -6.35154128e-01 -5.31505466e-01 -2.03107163e-01
3.25006664e-01 -7.93764055e-01 -1.46432233e+00 3.02735984e-01
-1.28380978e+00 -8.97363108e-03 3.76955092e-01 2.54033476e-01
1.62314981e-01 6.45291984e-01 -7.68703103e-01 -6.42504215e-01
-6.88187301e-01 -8.78966630e-01 7.31068909e-01 1.46928370e-01
1.11922733e-01 -1.05218136e+00 3.66181396e-02 -2.22147461e-02
4.11946237e-01 4.83390808e-01 1.13102531e+00 -5.65422118e-01
-7.92853057e-01 -5.08728325e-01 -8.73532653e-01 1.17223658e-01
1.37964785e-02 1.07403457e-01 -3.49470615e-01 -6.19053066e-01
-9.59504545e-01 -2.07332112e-02 1.03666186e+00 3.39031786e-01
1.29226077e+00 -4.78284955e-01 -5.03899038e-01 6.49736166e-01
1.42618394e+00 -5.23187995e-01 4.80349541e-01 -1.29337916e-02
9.36921000e-01 4.59784091e-01 1.72385901e-01 4.46162969e-01
5.03655851e-01 2.09297836e-01 7.67431796e-01 -2.09217429e-01
1.67268768e-01 -2.96973914e-01 -4.74609323e-02 1.05148494e+00
-2.22028390e-01 -5.57554483e-01 -8.31858814e-01 4.87049043e-01
-1.81974208e+00 -9.97930586e-01 -4.69384015e-01 2.15557528e+00
5.22009194e-01 3.41514677e-01 1.22926429e-01 2.31794417e-01
1.08427823e+00 2.26834178e-01 -5.10663033e-01 -9.81357619e-02
1.78552508e-01 3.09192725e-02 4.57802922e-01 5.48096478e-01
-9.00107503e-01 7.51038373e-01 5.59424019e+00 7.24355996e-01
-7.66962945e-01 -2.71936834e-01 7.56171048e-01 2.84941822e-01
-1.11307912e-01 1.45262584e-01 -3.07859838e-01 4.04996455e-01
7.02130318e-01 -4.46573049e-01 6.11828506e-01 7.61293709e-01
1.60274357e-01 1.31683365e-01 -9.70516503e-01 7.24150240e-01
-3.34062815e-01 -1.35862994e+00 -6.52692914e-02 1.14907712e-01
5.37038028e-01 2.20439777e-01 -3.43779594e-01 1.56963632e-01
3.99194628e-01 -1.02230430e+00 1.07179865e-01 3.71846974e-01
8.57042909e-01 -9.73983884e-01 8.95911872e-01 3.63336593e-01
-1.74733233e+00 3.40324007e-02 -6.88434303e-01 -2.21374616e-01
-5.29354438e-02 1.15720713e+00 -1.03027046e+00 4.14532661e-01
6.32078111e-01 6.65154874e-01 -6.04843259e-01 1.17578912e+00
-4.14663583e-01 5.69475353e-01 -3.59814763e-01 -4.82251853e-01
2.19794899e-01 -3.37098628e-01 4.92470235e-01 1.07195532e+00
2.05264822e-01 1.35194613e-02 1.97948933e-01 8.16159129e-01
-6.50243521e-01 9.12230276e-03 -4.65771645e-01 -3.75858158e-01
6.84870601e-01 1.70316851e+00 -1.08096159e+00 -2.34763876e-01
-1.83065206e-01 8.58182430e-01 8.71888399e-01 1.58371270e-01
-8.14946651e-01 -9.76322770e-01 4.90318865e-01 5.27706325e-01
4.25627492e-02 -5.14120638e-01 3.16003472e-01 -8.28996420e-01
-4.80312854e-02 -4.88700330e-01 5.05093038e-01 -4.54992265e-01
-1.39055037e+00 7.76486397e-01 -4.56249267e-01 -7.92371333e-01
4.81062420e-02 -6.20472968e-01 -9.83514905e-01 8.47064495e-01
-1.33620930e+00 -8.81277382e-01 -7.12234437e-01 5.75044811e-01
-1.42105058e-01 9.09666047e-02 5.77412009e-01 5.09540141e-01
-6.32432520e-01 5.13330638e-01 1.84703693e-01 7.78775215e-01
3.35812390e-01 -1.37877452e+00 8.82172823e-01 8.11275601e-01
-3.62016186e-02 9.12819862e-01 5.14422953e-01 -7.60661721e-01
-1.29982400e+00 -1.02734327e+00 1.18262970e+00 4.48772795e-02
8.47863734e-01 -4.93778378e-01 -7.37466991e-01 5.07919729e-01
5.33589423e-02 2.74049133e-01 4.74510729e-01 3.44674349e-01
-3.04884344e-01 -3.47925633e-01 -9.26046431e-01 6.10095501e-01
1.04922950e+00 -4.03287977e-01 2.62916684e-01 4.47637528e-01
8.61302257e-01 -1.33285984e-01 -8.17289174e-01 1.61524445e-01
4.66638744e-01 -9.59135413e-01 9.34295118e-01 -5.00441909e-01
3.49963665e-01 -3.87768686e-01 4.50078517e-01 -1.36548638e+00
-9.13479865e-01 -6.04085684e-01 -3.18077505e-01 1.05445194e+00
3.66891325e-01 -5.94521105e-01 1.05228496e+00 1.75222501e-01
4.13178653e-01 -8.78125966e-01 -8.19287539e-01 -3.97334099e-01
-3.61584008e-01 1.65077839e-02 7.43842602e-01 9.46325004e-01
-1.48212118e-02 7.27674603e-01 -1.65987119e-01 4.20977473e-01
8.56635988e-01 1.33883506e-01 8.15232277e-01 -1.76983857e+00
-1.62345283e-02 -5.35533965e-01 -6.92150950e-01 -8.53095889e-01
8.28751549e-02 -1.17463934e+00 -3.34040642e-01 -2.12239933e+00
4.27247405e-01 -5.39733946e-01 -3.34262609e-01 2.30335608e-01
-2.96047598e-01 1.53577536e-01 -1.34251863e-01 -8.37299749e-02
-8.49765837e-01 3.16559941e-01 1.16163623e+00 -4.33677793e-01
-9.76865739e-03 1.94122002e-01 -7.54080594e-01 6.25997365e-01
7.81238258e-01 -5.45294404e-01 -3.05857152e-01 -4.35072243e-01
9.70569730e-01 1.59244630e-02 4.01720464e-01 -8.49220276e-01
6.05395019e-01 3.10792010e-02 2.53046274e-01 -5.71906626e-01
2.42948020e-03 -7.74296880e-01 1.07502915e-01 7.56813884e-01
-2.87967205e-01 3.52313966e-01 -2.18669713e-01 8.75710487e-01
-1.33154854e-01 3.07492334e-02 6.32426739e-01 -1.50576189e-01
-5.05243719e-01 7.21366942e-01 8.32466260e-02 3.71261090e-02
8.81116867e-01 1.20886825e-01 -3.41122121e-01 -6.11709833e-01
-3.80705535e-01 5.80416560e-01 1.89534314e-02 1.96127832e-01
7.70330966e-01 -1.32144845e+00 -1.07368684e+00 -1.12854868e-01
1.69683378e-02 7.54022300e-02 1.02925822e-01 9.08888936e-01
-9.23457563e-01 3.47704560e-01 9.76766273e-02 -3.36560398e-01
-1.34469414e+00 4.24794137e-01 2.86493957e-01 -7.60577619e-01
-6.93536699e-01 1.02771199e+00 -1.17090913e-02 -4.56569582e-01
2.91968137e-01 -9.12704095e-02 -2.94086814e-01 -1.90930843e-01
5.42348027e-01 7.82196701e-01 -2.65815277e-02 -5.17359555e-01
-5.53452075e-01 2.85159320e-01 -1.39246449e-01 5.50187469e-01
1.48943150e+00 6.74331002e-03 -8.02489042e-01 -6.37235567e-02
1.21189785e+00 -3.69763263e-02 -8.17006052e-01 -3.62642735e-01
1.95506424e-01 -4.29203480e-01 2.78213695e-02 -3.07171285e-01
-1.48144138e+00 8.78791273e-01 4.11740541e-02 6.41019404e-01
9.34717953e-01 -5.84959202e-02 8.39329362e-01 6.39959097e-01
4.91560459e-01 -8.88180792e-01 1.46328211e-01 5.67191362e-01
6.47611558e-01 -1.06323242e+00 2.25575179e-01 -5.67586899e-01
-1.10782154e-01 1.16426814e+00 6.01684034e-01 -3.61922801e-01
8.08323264e-01 -4.02417295e-02 -5.87677419e-01 -4.67042446e-01
-6.15366995e-01 -1.55420348e-01 1.91038579e-01 3.01568896e-01
6.93796158e-01 1.87249586e-01 -3.85450959e-01 2.66762525e-01
-3.06690857e-03 -3.72310400e-01 5.50845265e-01 2.66297072e-01
-6.07802570e-01 -9.84731972e-01 5.96927218e-02 7.76861072e-01
-4.63960171e-01 -2.98434466e-01 -6.03462636e-01 5.35851300e-01
-2.73257017e-01 1.00400746e+00 9.24898162e-02 -5.29155910e-01
8.49647224e-02 -6.48510158e-01 2.11045355e-01 -5.11882365e-01
-5.14685810e-01 -4.20738429e-01 -1.38629153e-02 -5.27997017e-01
-1.84251398e-01 1.13122597e-01 -1.34838951e+00 -1.24645901e+00
-5.77762425e-01 4.26512927e-01 3.35672170e-01 5.41221917e-01
4.14181828e-01 7.92791784e-01 7.98565090e-01 -5.46928525e-01
-2.50756443e-01 -1.01097238e+00 -8.02725971e-01 1.79633021e-01
2.82633454e-01 -3.09880912e-01 -3.07361692e-01 -7.20688105e-01] | [7.0471343994140625, 6.1504974365234375] |
afc54c68-f669-41a8-967f-1dedf14ecff6 | weakly-supervised-instance-segmentation-via | 2104.01526 | null | https://arxiv.org/abs/2104.01526v1 | https://arxiv.org/pdf/2104.01526v1.pdf | Weakly-supervised Instance Segmentation via Class-agnostic Learning with Salient Images | Humans have a strong class-agnostic object segmentation ability and can outline boundaries of unknown objects precisely, which motivates us to propose a box-supervised class-agnostic object segmentation (BoxCaseg) based solution for weakly-supervised instance segmentation. The BoxCaseg model is jointly trained using box-supervised images and salient images in a multi-task learning manner. The fine-annotated salient images provide class-agnostic and precise object localization guidance for box-supervised images. The object masks predicted by a pretrained BoxCaseg model are refined via a novel merged and dropped strategy as proxy ground truth to train a Mask R-CNN for weakly-supervised instance segmentation. Only using $7991$ salient images, the weakly-supervised Mask R-CNN is on par with fully-supervised Mask R-CNN on PASCAL VOC and significantly outperforms previous state-of-the-art box-supervised instance segmentation methods on COCO. The source code, pretrained models and datasets are available at \url{https://github.com/hustvl/BoxCaseg}. | ['Wenyu Liu', 'Xiaoxin Chen', 'Longjin Ran', 'Qi Ding', 'Bin Hu', 'Jiapei Feng', 'Xinggang Wang'] | 2021-04-04 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Wang_Weakly-Supervised_Instance_Segmentation_via_Class-Agnostic_Learning_With_Salient_Images_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Wang_Weakly-Supervised_Instance_Segmentation_via_Class-Agnostic_Learning_With_Salient_Images_CVPR_2021_paper.pdf | cvpr-2021-1 | ['box-supervised-instance-segmentation', 'weakly-supervised-instance-segmentation'] | ['computer-vision', 'computer-vision'] | [ 3.10476393e-01 6.70360267e-01 -4.35589880e-01 -7.38932610e-01
-1.19376731e+00 -7.16191232e-01 3.43666136e-01 6.67267740e-02
-3.13123673e-01 5.28453767e-01 -1.12225942e-01 1.65643796e-01
6.92480728e-02 -3.17872584e-01 -9.20503139e-01 -4.51967657e-01
2.10443705e-01 9.63766217e-01 7.28053093e-01 -6.72933832e-02
2.97112435e-01 3.18516999e-01 -1.49067283e+00 5.34010589e-01
9.01909292e-01 9.96954918e-01 5.35340846e-01 6.83075905e-01
-1.71199083e-01 4.57345694e-01 -5.81785262e-01 -1.81295320e-01
4.72415268e-01 -2.81491458e-01 -1.31185400e+00 6.60923183e-01
7.21512616e-01 -2.72401273e-02 2.68343061e-01 1.14594483e+00
3.05114724e-02 8.86251554e-02 5.20690620e-01 -1.08740366e+00
-4.83324975e-01 7.81201065e-01 -8.28570068e-01 2.20854893e-01
-1.61457390e-01 2.76058704e-01 1.20728683e+00 -1.06472242e+00
6.19886875e-01 9.47261035e-01 5.33987582e-01 6.33426368e-01
-1.28160489e+00 -4.46962029e-01 6.58146441e-01 -2.11007908e-01
-1.35633337e+00 -1.04541786e-01 9.09933150e-01 -4.84167635e-01
6.68424129e-01 3.37245375e-01 6.18642390e-01 5.59977472e-01
-4.09194529e-01 1.39602101e+00 1.20716918e+00 -1.68388158e-01
1.65225372e-01 4.27282393e-01 2.78768152e-01 8.49307716e-01
2.68728644e-01 -2.85439696e-02 -2.89171427e-01 2.07675695e-01
8.95191729e-01 1.39523640e-01 -6.44125119e-02 -5.36287487e-01
-1.30807710e+00 6.41578317e-01 8.50666225e-01 2.69749701e-01
-1.19853444e-01 2.20895723e-01 2.18681335e-01 -2.12713614e-01
7.49943674e-01 7.29016006e-01 -9.29675221e-01 3.24526995e-01
-1.30855060e+00 2.63203084e-01 5.81599236e-01 1.26052666e+00
1.15099561e+00 6.13097213e-02 -3.40295225e-01 8.61254632e-01
2.63824046e-01 2.69902050e-01 2.48620465e-01 -1.14229167e+00
3.56223851e-01 9.36006248e-01 7.81773925e-02 -3.81222546e-01
-4.53651279e-01 -8.13649476e-01 -3.94553334e-01 2.50968188e-01
7.10987508e-01 6.93957582e-02 -1.50799215e+00 1.17565084e+00
5.95397115e-01 1.89766549e-02 -1.56381756e-01 1.13400745e+00
1.15037036e+00 5.29248834e-01 1.49820641e-01 4.18098599e-01
1.33051229e+00 -1.72281170e+00 -1.75824538e-01 -7.54686713e-01
4.65401709e-01 -6.75372541e-01 1.26619554e+00 1.63018942e-01
-1.14946735e+00 -8.45641255e-01 -8.22303891e-01 -1.15769669e-01
-5.30310094e-01 4.65333730e-01 6.93248928e-01 3.76802593e-01
-8.20650697e-01 4.82486814e-01 -6.17977917e-01 -5.36243916e-02
1.25889862e+00 2.48588011e-01 -2.59155095e-01 -2.52826083e-02
-4.83804792e-01 5.75287282e-01 8.85559380e-01 6.94639087e-02
-1.35827315e+00 -8.74773085e-01 -1.06539536e+00 -1.05361275e-01
6.95487499e-01 -1.84495538e-01 1.40496671e+00 -1.32093275e+00
-1.10549355e+00 1.60334945e+00 -1.99754741e-02 -4.71881181e-01
5.40456235e-01 -2.92889863e-01 1.01853050e-01 4.43901390e-01
6.42869294e-01 1.31357706e+00 1.20550895e+00 -1.96058464e+00
-7.87629008e-01 -2.37745762e-01 1.66548770e-02 1.14156075e-01
4.38811153e-01 -5.71330404e-03 -6.45200193e-01 -8.99869800e-01
1.96848705e-01 -7.67348289e-01 -6.01504028e-01 2.88276784e-02
-8.66270900e-01 -2.15909243e-01 1.07183325e+00 -5.28440177e-01
7.33909607e-01 -2.02731109e+00 -1.02827244e-01 2.24902276e-02
3.26382399e-01 3.29251617e-01 -2.41706163e-01 -1.87552214e-01
-1.02891102e-01 1.70046091e-01 -9.16403472e-01 -5.69628119e-01
-1.37545899e-01 2.45351210e-01 -2.28804901e-01 3.82202059e-01
6.38497412e-01 1.34745276e+00 -9.72357273e-01 -7.80610621e-01
4.62559253e-01 3.11756264e-02 -4.25257027e-01 5.32613397e-01
-8.23338866e-01 6.89756930e-01 -4.75901812e-01 1.23561764e+00
6.20256484e-01 -6.72316730e-01 -3.36217821e-01 -3.37577239e-02
9.27236769e-03 -2.05800220e-01 -1.04756260e+00 1.77447999e+00
-2.38667261e-02 3.87197763e-01 2.48046413e-01 -1.27445602e+00
9.27065313e-01 3.32658365e-02 3.15964907e-01 -3.36005509e-01
4.03117537e-02 3.48096043e-01 -2.11829454e-01 -2.77528316e-01
3.07555825e-01 9.91123393e-02 -2.44630963e-01 3.37458521e-01
5.26308239e-01 -9.25564229e-01 3.14663738e-01 2.80137062e-01
4.43415165e-01 6.97112739e-01 2.36428142e-01 -6.49785578e-01
4.22700018e-01 4.87371236e-01 6.77118301e-01 8.00058305e-01
-5.11134446e-01 1.29993629e+00 1.58825234e-01 -4.99871701e-01
-9.21703637e-01 -9.79992092e-01 -2.43848011e-01 1.31567729e+00
5.94750941e-01 -8.66749603e-03 -1.03977501e+00 -1.15348339e+00
1.51864678e-01 4.60501254e-01 -8.29424441e-01 3.10828149e-01
-5.30749500e-01 -4.34875458e-01 1.93568692e-01 7.84645975e-01
7.75032520e-01 -1.45896924e+00 -4.72643375e-01 6.30277246e-02
-6.33500610e-03 -1.17167592e+00 -6.53434157e-01 5.78553021e-01
-8.93504262e-01 -1.19903398e+00 -9.34619844e-01 -1.04684985e+00
1.18523538e+00 1.56342849e-01 1.37413812e+00 2.87593871e-01
-6.40872836e-01 4.69435692e-01 -2.87763447e-01 -6.66694939e-01
5.57113774e-02 1.17575422e-01 -4.32785600e-01 5.88727824e-04
1.22802772e-01 -1.54017936e-02 -7.99648941e-01 5.16306818e-01
-8.11129987e-01 3.11836720e-01 5.89242399e-01 6.83772564e-01
1.10669982e+00 -2.93871403e-01 5.81234396e-01 -1.34964526e+00
-2.06410766e-01 -2.46902108e-01 -7.18121529e-01 9.34288055e-02
-2.92418927e-01 -2.91250974e-01 4.36353356e-01 -2.91229963e-01
-1.22585571e+00 5.66222727e-01 -1.07548773e-01 -3.82516116e-01
-6.75403178e-01 -1.31519273e-01 -1.94556400e-01 -1.27034798e-01
6.00590765e-01 4.96111102e-02 -3.64532053e-01 -4.45156157e-01
6.26966357e-01 3.16445678e-01 7.33562768e-01 -6.80980682e-01
1.01825023e+00 7.83262372e-01 -4.77119207e-01 -5.90576231e-01
-1.75625420e+00 -9.46375430e-01 -1.08382595e+00 -3.24893296e-01
1.19902229e+00 -1.01720774e+00 -6.18135147e-02 4.23360616e-01
-9.13421929e-01 -1.08380175e+00 -8.39049399e-01 7.85010383e-02
-8.43329132e-01 -1.16460308e-01 -4.19928133e-01 -5.92639327e-01
-2.65844971e-01 -1.08504963e+00 1.55602849e+00 5.68626225e-01
-2.06186220e-01 -9.22690094e-01 -4.34056371e-01 1.09924471e+00
1.36027545e-01 3.70846450e-01 2.93399781e-01 -9.10525382e-01
-9.45683420e-01 -1.80787221e-01 -5.66007912e-01 4.40587014e-01
-5.00516482e-02 -1.89890057e-01 -1.19342363e+00 -2.69846737e-01
-3.15786511e-01 -6.42144740e-01 1.16269839e+00 6.99663579e-01
1.70781744e+00 -1.95975438e-01 -4.88363802e-01 8.70590091e-01
1.37383568e+00 -2.56584078e-01 1.77687526e-01 2.60385960e-01
9.17104065e-01 6.37070000e-01 1.09716058e+00 2.36092553e-01
2.50885040e-01 2.39380836e-01 6.23982787e-01 -7.34719694e-01
-2.97703832e-01 -2.63536960e-01 -3.26556742e-01 3.21730599e-02
-3.90203833e-03 1.05786785e-01 -9.09335673e-01 1.04436135e+00
-1.83684623e+00 -6.33815587e-01 -1.03687935e-01 1.64390922e+00
1.01167142e+00 4.57731485e-01 3.36596787e-01 -1.14645720e-01
9.01710212e-01 3.33069056e-01 -7.42342830e-01 -1.06364831e-01
-7.69619420e-02 2.70569324e-01 6.77976131e-01 5.11148274e-01
-1.57107604e+00 1.51823294e+00 5.58527040e+00 9.27723885e-01
-7.86944330e-01 2.83085197e-01 1.19242263e+00 2.06388570e-02
-8.19097310e-02 3.41247283e-02 -1.05852163e+00 2.68411070e-01
1.46459967e-01 1.87152699e-01 -7.71632046e-02 1.26594031e+00
-2.64640283e-02 -2.57117212e-01 -9.49565351e-01 7.87523568e-01
1.54274493e-01 -1.57402623e+00 -1.22630186e-01 -3.33085358e-01
1.25691283e+00 2.47783631e-01 1.03994288e-01 3.54629666e-01
4.58520889e-01 -1.04688525e+00 9.38803673e-01 2.04362065e-01
5.69377065e-01 -3.99620384e-01 5.99871933e-01 2.76736170e-01
-1.30421698e+00 -3.84067036e-02 -1.91424400e-01 1.58384308e-01
1.83151871e-01 6.13241851e-01 -6.42784774e-01 3.29206198e-01
1.11711860e+00 7.45323181e-01 -8.00254643e-01 1.08130181e+00
-4.98285443e-01 8.52559209e-01 -2.22272038e-01 2.60562927e-01
5.65011740e-01 -1.49442539e-01 5.74478447e-01 1.42241514e+00
-4.61325943e-01 2.85322040e-01 5.98945618e-01 1.33362091e+00
-1.00063562e-01 -2.22762868e-01 -1.75382838e-01 3.02201081e-02
1.08611412e-01 1.61980438e+00 -1.42773414e+00 -6.94367409e-01
-2.68336654e-01 1.04738915e+00 4.08397347e-01 5.49637258e-01
-6.97509766e-01 -3.33391994e-01 2.90497720e-01 1.99344486e-01
6.05749547e-01 -1.83142424e-02 -7.60270417e-01 -9.51628268e-01
-1.41628221e-01 -6.00666642e-01 4.87553179e-01 -8.46656382e-01
-1.19859970e+00 4.91545647e-01 1.35690331e-01 -1.08410919e+00
2.27892563e-01 -6.23673558e-01 -7.71115720e-01 5.74216783e-01
-1.83499146e+00 -1.63187087e+00 -4.64205205e-01 4.19652075e-01
1.11248577e+00 3.16959731e-02 4.52272803e-01 -2.53906697e-01
-5.72536349e-01 2.21796006e-01 -3.96925896e-01 4.08549637e-01
4.06231552e-01 -1.75111353e+00 4.33180749e-01 8.81675363e-01
2.63930261e-01 3.83488327e-01 4.64445382e-01 -6.68071866e-01
-5.96177459e-01 -1.68598366e+00 2.55323082e-01 -7.60961831e-01
3.08499813e-01 -4.87150192e-01 -9.83649254e-01 8.27973485e-01
1.41202703e-01 7.84371734e-01 5.03830373e-01 -2.36573666e-01
-2.45563120e-01 1.88019071e-02 -1.31122398e+00 3.55822086e-01
1.01281905e+00 -3.41773063e-01 -8.22817981e-01 6.66330338e-01
1.03723550e+00 -6.41974330e-01 -6.98040128e-01 5.89306414e-01
-2.98107658e-02 -7.72613168e-01 1.04561412e+00 -5.13387084e-01
3.48925591e-01 -5.55459023e-01 2.00560018e-01 -8.75330508e-01
-1.07504658e-01 -4.37085837e-01 -1.09811902e-01 1.09961164e+00
8.13364148e-01 -5.39852560e-01 1.19280338e+00 5.25406361e-01
-4.88570273e-01 -9.38519955e-01 -5.95682740e-01 -7.14423716e-01
-6.43540993e-02 -3.84329349e-01 3.66723329e-01 9.42670166e-01
-3.51109713e-01 3.34593875e-04 1.05605178e-01 2.29239017e-01
9.49799657e-01 7.15687752e-01 7.91195989e-01 -1.16081345e+00
-2.38424659e-01 -4.61556762e-01 -6.41344264e-02 -1.21440017e+00
1.80535197e-01 -9.37296987e-01 4.61645603e-01 -1.70197701e+00
3.54784817e-01 -6.70048118e-01 -1.41177356e-01 8.69618952e-01
-4.94239390e-01 8.05605412e-01 2.09972695e-01 1.40484259e-01
-1.03111207e+00 3.39549333e-01 1.60411632e+00 -3.49036157e-01
-3.52156699e-01 2.23064587e-01 -7.57578433e-01 9.48746979e-01
1.00613749e+00 -4.84036446e-01 -1.44478694e-01 -2.53423750e-01
-3.02177280e-01 -4.15981919e-01 6.02928221e-01 -8.88704181e-01
2.22771913e-02 -2.46670648e-01 6.27579629e-01 -8.87680709e-01
3.71576607e-01 -7.31736064e-01 -5.58442533e-01 2.98571646e-01
-3.92329544e-01 -5.76102197e-01 3.25276792e-01 4.54545557e-01
-3.34460467e-01 -3.47045749e-01 1.08534968e+00 -7.31687665e-01
-1.07122266e+00 5.93701780e-01 4.50688340e-02 4.83156115e-01
1.19665241e+00 -5.81845880e-01 -3.09104592e-01 -2.22521406e-02
-9.94027615e-01 5.60277641e-01 4.02147293e-01 3.42693329e-01
7.63802946e-01 -7.75933325e-01 -6.10555053e-01 1.63076192e-01
4.22780573e-01 8.45364988e-01 1.40264317e-01 5.73105872e-01
-4.57937866e-01 3.20136577e-01 2.36157328e-02 -9.81610477e-01
-1.07274938e+00 4.36266214e-01 4.53514129e-01 2.73463540e-02
-6.03708565e-01 1.37586820e+00 5.65538824e-01 -9.56589818e-01
1.67685509e-01 -5.23729742e-01 -5.44121908e-03 -2.59842724e-01
1.97561473e-01 -3.21547128e-02 -2.86453277e-01 -6.21867239e-01
-2.99646437e-01 6.89205170e-01 -1.66613325e-01 1.04817294e-01
1.38132536e+00 -1.05313212e-02 3.94038633e-02 2.90117711e-01
9.88194346e-01 -3.02809387e-01 -1.78723156e+00 -2.71190763e-01
1.63714752e-01 -4.54720438e-01 -7.37188235e-02 -9.44122851e-01
-1.18370926e+00 7.41184652e-01 3.01657468e-01 -6.02164082e-02
9.05607224e-01 6.95949137e-01 4.92392749e-01 9.81303081e-02
3.63329679e-01 -1.38149428e+00 5.23154497e-01 3.72072667e-01
9.49190676e-01 -1.83548725e+00 -6.99683512e-03 -9.01248753e-01
-9.33573663e-01 6.80436552e-01 1.21784365e+00 -3.84864897e-01
6.35651886e-01 1.24398947e-01 2.92047083e-01 -4.28471833e-01
-1.13010325e-01 -7.97456205e-01 5.87923050e-01 7.82698929e-01
1.21407136e-01 1.27858728e-01 1.18663415e-01 6.02632880e-01
-5.08517362e-02 -3.73864800e-01 6.30483478e-02 8.80404174e-01
-6.80570722e-01 -6.73093975e-01 -4.70128149e-01 5.67343771e-01
-4.19439256e-01 7.32900284e-04 -4.65781748e-01 8.97498548e-01
2.53237426e-01 7.37700880e-01 2.39964694e-01 3.14939022e-01
1.46431416e-01 -1.81117564e-01 3.19267213e-01 -1.27859950e+00
-7.75382698e-01 5.25698736e-02 -6.74790666e-02 -7.41515517e-01
-4.82965142e-01 -4.93922085e-01 -1.64624190e+00 6.33261144e-01
-4.49790418e-01 1.33305877e-01 3.75783920e-01 9.93336976e-01
1.02797784e-01 4.69457865e-01 3.89592320e-01 -1.37258637e+00
3.41308266e-02 -8.96069229e-01 -4.70685065e-01 6.73692822e-01
4.15451258e-01 -3.71947587e-01 -4.14914668e-01 2.87925303e-01] | [9.565367698669434, 0.5585986971855164] |
38891514-b022-4acc-853b-23b87ee0310e | iterative-deep-homography-estimation | 2203.15982 | null | https://arxiv.org/abs/2203.15982v1 | https://arxiv.org/pdf/2203.15982v1.pdf | Iterative Deep Homography Estimation | We propose Iterative Homography Network, namely IHN, a new deep homography estimation architecture. Different from previous works that achieve iterative refinement by network cascading or untrainable IC-LK iterator, the iterator of IHN has tied weights and is completely trainable. IHN achieves state-of-the-art accuracy on several datasets including challenging scenes. We propose 2 versions of IHN: (1) IHN for static scenes, (2) IHN-mov for dynamic scenes with moving objects. Both versions can be arranged in 1-scale for efficiency or 2-scale for accuracy. We show that the basic 1-scale IHN already outperforms most of the existing methods. On a variety of datasets, the 2-scale IHN outperforms all competitors by a large gap. We introduce IHN-mov by producing an inlier mask to further improve the estimation accuracy of moving-objects scenes. We experimentally show that the iterative framework of IHN can achieve 95% error reduction while considerably saving network parameters. When processing sequential image pairs, IHN can achieve 32.7 fps, which is about 8x the speed of IC-LK iterator. Source code is available at https://github.com/imdumpl78/IHN. | ['Hui-Liang Shen', 'Zehua Sheng', 'Jianxin Hu', 'Si-Yuan Cao'] | 2022-03-30 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Cao_Iterative_Deep_Homography_Estimation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Cao_Iterative_Deep_Homography_Estimation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['homography-estimation'] | ['computer-vision'] | [-7.58319423e-02 5.51466942e-02 -9.37116891e-02 -7.69020021e-02
-6.86345696e-01 -3.75787437e-01 3.09671909e-01 -7.64346182e-01
-4.23126221e-01 4.76802438e-01 1.68537453e-01 -1.29870266e-01
1.06698632e-01 -6.26926005e-01 -1.00303578e+00 -4.28624213e-01
1.77174121e-01 4.33965981e-01 6.26208544e-01 -1.09501630e-01
3.51232067e-02 3.56192350e-01 -1.04751873e+00 1.44759178e-01
5.75533688e-01 1.03850925e+00 2.40129501e-01 9.29861546e-01
4.86610293e-01 6.66010916e-01 -3.05461705e-01 -4.80933368e-01
8.46383572e-01 -4.53927636e-01 -7.09264219e-01 -5.15674613e-02
8.78450215e-01 -6.57156169e-01 -8.15477192e-01 1.08971989e+00
6.36103570e-01 1.39185667e-01 -1.84070878e-02 -1.37408936e+00
-6.23757005e-01 5.71731746e-01 -5.90138614e-01 -9.33235809e-02
7.29814395e-02 2.57072717e-01 9.57316577e-01 -1.18572807e+00
7.52658188e-01 1.28311825e+00 9.52626169e-01 4.96626019e-01
-1.17654920e+00 -8.71834397e-01 -5.49025722e-02 3.42794150e-01
-1.59352887e+00 -3.52006584e-01 6.02596283e-01 -9.20463279e-02
9.45309222e-01 7.75065459e-03 8.18790078e-01 1.02847254e+00
2.10733801e-01 9.87078786e-01 7.72937000e-01 -6.03823215e-02
-1.67123780e-01 -3.64724010e-01 -1.72085151e-01 7.02277303e-01
1.76033631e-01 1.91435814e-01 -5.28346896e-01 1.83053479e-01
1.40475106e+00 1.21571943e-01 -5.58852017e-01 -4.36168998e-01
-1.65367520e+00 6.80641353e-01 7.47945547e-01 -9.38535333e-02
-7.13237002e-02 6.31401837e-01 2.69128382e-01 2.00451180e-01
2.38695264e-01 4.53943312e-01 -4.08616394e-01 -1.86075583e-01
-8.50065708e-01 6.09186403e-02 7.44608879e-01 1.38499713e+00
7.75201738e-01 -1.49103403e-02 1.69363350e-01 7.60795891e-01
1.16028683e-02 5.04482806e-01 4.10145134e-01 -1.28575242e+00
5.02617955e-01 3.00499678e-01 -2.39484757e-02 -1.20590973e+00
-4.99630421e-01 -5.22023618e-01 -1.05736470e+00 4.74954322e-02
3.30981106e-01 -1.46193013e-01 -1.04702938e+00 1.51446319e+00
3.49602103e-01 5.84115088e-01 -3.31438184e-01 1.11470914e+00
9.53492820e-01 7.31844604e-01 -5.50551474e-01 8.30422416e-02
1.08482540e+00 -1.85449493e+00 -6.85078204e-01 -2.91126221e-01
4.51666176e-01 -9.08631206e-01 1.09995079e+00 3.81059259e-01
-1.28347027e+00 -7.43300080e-01 -1.24651051e+00 -5.98042130e-01
-1.19671822e-01 3.68047029e-01 7.75951326e-01 1.67257160e-01
-1.40865231e+00 7.05223083e-01 -1.03463340e+00 -2.76223212e-01
3.56727570e-01 5.83393872e-01 -3.80434990e-01 1.66687220e-01
-1.01433480e+00 5.76253772e-01 3.79007846e-01 3.62898022e-01
-8.49317670e-01 -9.11435544e-01 -1.02038920e+00 2.02216864e-01
4.58795011e-01 -1.00583160e+00 1.43048680e+00 -5.59248924e-01
-1.61073685e+00 6.25533402e-01 -8.15569609e-02 -5.02861321e-01
1.03643143e+00 -6.46131396e-01 -3.13339472e-01 1.70336828e-01
3.60726640e-02 1.15091062e+00 7.98945010e-01 -1.09467900e+00
-7.16104984e-01 4.72433157e-02 -5.22508696e-02 1.89348295e-01
-2.29846016e-01 -2.70448238e-01 -1.27312911e+00 -7.38308251e-01
4.63269651e-01 -1.19076335e+00 -2.40237176e-01 3.83678883e-01
-5.29252291e-01 1.00320838e-01 1.02940345e+00 -6.00271702e-01
1.28991926e+00 -2.12582564e+00 8.30669329e-02 -1.90682970e-02
5.63192606e-01 3.90584558e-01 -2.29633465e-01 1.97087705e-01
2.08120663e-02 -1.70719370e-01 -9.93030816e-02 -5.79448760e-01
6.46492988e-02 7.96837285e-02 -2.34284788e-01 5.96611559e-01
-2.44681939e-01 1.36306000e+00 -7.45638490e-01 -3.15518528e-01
4.57426459e-01 6.99631393e-01 -6.77721858e-01 1.26175269e-01
2.02315643e-01 1.49809003e-01 1.33303031e-01 5.51267326e-01
8.73779237e-01 -7.37361073e-01 9.09031332e-02 -7.58729219e-01
-9.89890471e-02 1.91686153e-01 -1.34198296e+00 1.80881786e+00
-4.53985244e-01 1.05174696e+00 -2.89377533e-02 -5.31891644e-01
7.46751308e-01 9.36592370e-02 3.61507297e-01 -6.13179266e-01
2.42291898e-01 2.50397950e-01 -9.37083811e-02 -5.32536581e-02
6.44757926e-01 4.06099677e-01 1.10469542e-01 1.64642572e-01
2.92750090e-01 -1.78085908e-01 1.64448529e-01 2.25748122e-01
1.00182533e+00 5.78523315e-02 3.49665225e-01 -2.72247851e-01
3.92214447e-01 -3.50281477e-01 8.37400079e-01 8.43599081e-01
-3.03938508e-01 1.09951937e+00 3.90294671e-01 -8.91491413e-01
-1.23687804e+00 -1.00878584e+00 -6.80641234e-02 7.33467817e-01
6.86132312e-01 -5.46767533e-01 -5.76643229e-01 -5.50788462e-01
5.89466561e-03 1.83911726e-01 -5.87251842e-01 7.16382712e-02
-9.50973332e-01 -4.86442536e-01 3.58300030e-01 7.95884252e-01
1.21342003e+00 -9.55362558e-01 -5.04328370e-01 1.20566480e-01
-1.00900225e-01 -1.51638341e+00 -9.83451605e-01 -2.30312094e-01
-8.38028491e-01 -1.06082284e+00 -7.88699389e-01 -7.57498384e-01
7.39026010e-01 4.75456685e-01 1.22810495e+00 1.92540616e-01
-1.81764558e-01 1.58572495e-01 -1.69073701e-01 -2.28280183e-02
3.36261876e-02 3.98490489e-01 2.05861428e-03 -1.93808690e-01
1.18321208e-02 -7.32579470e-01 -1.11571169e+00 7.37813294e-01
-6.56581163e-01 4.73138094e-01 6.25943422e-01 9.11174238e-01
6.90105557e-01 -1.06461629e-01 -6.78859055e-02 -8.74369800e-01
3.04830968e-02 2.34309435e-02 -9.33803618e-01 7.96575099e-02
-6.23895943e-01 -9.37735140e-02 5.67830682e-01 -5.27931571e-01
-7.50763953e-01 1.96509793e-01 -2.60384560e-01 -8.90986204e-01
1.10875390e-01 -6.69770911e-02 -5.51721714e-02 -5.27355611e-01
2.97138512e-01 1.80152226e-02 -1.44106477e-01 -4.95264620e-01
3.85351539e-01 1.21155545e-01 9.95708048e-01 -1.36900544e-01
1.01239181e+00 7.09582865e-01 -6.37226626e-02 -5.08001804e-01
-8.61542642e-01 -5.48521817e-01 -6.58469260e-01 -1.08252145e-01
8.47400486e-01 -1.10507882e+00 -8.66293311e-01 6.98899388e-01
-1.16527677e+00 -7.88841307e-01 -2.00200185e-01 6.53980136e-01
-5.00778556e-01 3.39185655e-01 -9.83950078e-01 -1.31976634e-01
-5.27734578e-01 -1.37192464e+00 1.34421062e+00 1.44788563e-01
1.18408334e-02 -1.05522037e+00 -8.33491907e-02 1.97698414e-01
3.98303926e-01 4.43782955e-02 1.67548314e-01 -2.95752883e-01
-8.51128161e-01 -8.79741386e-02 -5.23608387e-01 2.43782863e-01
-1.63757920e-01 -2.61339117e-02 -6.30624354e-01 -5.21792591e-01
-9.92339626e-02 -3.42221409e-01 9.17554259e-01 4.84419525e-01
1.18181479e+00 -3.23872715e-01 -1.98036715e-01 1.48678637e+00
1.45287549e+00 2.38423511e-01 9.63947117e-01 5.96603811e-01
1.12418008e+00 -9.02473032e-02 3.72955620e-01 3.52901340e-01
5.54769993e-01 8.82981539e-01 4.62340236e-01 -6.35607064e-01
-3.84451747e-01 -4.05577958e-01 1.49565265e-01 9.28326368e-01
-1.98160812e-01 -1.25116080e-01 -7.94215202e-01 4.42508072e-01
-2.16815066e+00 -8.14999878e-01 -8.35157931e-02 2.01362228e+00
4.54593420e-01 2.74883926e-01 9.66584310e-02 -2.05076933e-01
6.79591596e-01 4.24266815e-01 -6.90646112e-01 5.25212884e-02
-8.43011662e-02 -7.54462332e-02 8.02853823e-01 8.11826169e-01
-1.33080077e+00 1.24064195e+00 6.00796175e+00 7.89189935e-01
-1.02581227e+00 1.05692424e-01 5.12550116e-01 -2.19859660e-01
-2.06497423e-02 2.61570275e-01 -1.14867878e+00 4.15603936e-01
3.24548483e-01 -6.32570833e-02 4.88125533e-01 1.03098977e+00
1.80035178e-02 1.43230125e-01 -1.04267049e+00 1.35461545e+00
1.88554853e-01 -1.55080795e+00 1.40248746e-01 1.78013914e-04
1.09880447e+00 2.47010753e-01 9.00022537e-02 2.89031595e-01
3.35439563e-01 -6.10335290e-01 6.16241157e-01 2.98645318e-01
8.09228957e-01 -5.70323050e-01 8.50611091e-01 7.22330436e-02
-1.52346349e+00 1.63829029e-01 -5.18588781e-01 1.30958840e-01
6.63142562e-01 4.99239832e-01 -4.92532074e-01 4.51201975e-01
1.01092434e+00 9.70152855e-01 -7.05033779e-01 1.17692447e+00
-3.39285135e-01 3.69614273e-01 -4.44835991e-01 3.41271311e-01
4.47309464e-01 -2.43354306e-01 4.41048563e-01 1.20693839e+00
5.19155204e-01 1.86363623e-01 2.45121330e-01 6.85375512e-01
-3.75919014e-01 -2.14216366e-01 -3.35428983e-01 4.03075069e-01
4.99296695e-01 1.36865425e+00 -8.61614764e-01 -6.31140172e-01
-4.39833134e-01 1.37965977e+00 2.68164963e-01 3.66205961e-01
-1.16426146e+00 -6.38484955e-01 5.78016877e-01 -3.55123095e-02
5.58881521e-01 -3.83697718e-01 -8.59638304e-02 -1.40600204e+00
2.33185086e-02 -8.16393554e-01 3.38586897e-01 -8.91358316e-01
-8.81378591e-01 7.38663793e-01 -1.28937781e-01 -1.31300986e+00
1.07729696e-02 -7.34719455e-01 -4.81901407e-01 3.39804798e-01
-1.09916258e+00 -1.12541401e+00 -8.41957271e-01 5.74733019e-01
8.73252690e-01 9.28138047e-02 1.48056597e-01 5.33923805e-01
-6.32559597e-01 8.74714553e-01 3.71923782e-02 4.00059223e-01
8.75999510e-01 -1.11212182e+00 9.57834542e-01 1.01972592e+00
1.63518459e-01 6.13647997e-01 4.89343673e-01 -6.00179613e-01
-1.20864844e+00 -1.16397035e+00 8.28504801e-01 -4.83256876e-01
7.15045512e-01 -6.30975187e-01 -7.74381042e-01 1.01735950e+00
2.98214227e-01 3.88727903e-01 1.07052006e-01 -2.35513374e-01
-3.54052305e-01 -1.67860582e-01 -7.35812724e-01 8.15627754e-01
1.55450606e+00 -4.34700906e-01 -9.49195921e-02 4.74138021e-01
1.02075410e+00 -1.01407361e+00 -8.57499540e-01 7.32258379e-01
7.75151908e-01 -1.36891294e+00 1.00724757e+00 -1.19742893e-01
5.20130754e-01 -4.28994149e-01 1.36806937e-02 -9.64081883e-01
-6.24370515e-01 -9.75529909e-01 -3.95419627e-01 5.38405955e-01
3.90521675e-01 -6.87039196e-01 9.08347368e-01 5.33631921e-01
-2.78843522e-01 -9.93365288e-01 -6.57847285e-01 -1.03689909e+00
-2.54085541e-01 -3.47028404e-01 6.21720493e-01 8.57066512e-01
-3.93206954e-01 4.17050153e-01 -7.72793055e-01 3.13443661e-01
7.05842137e-01 1.72437176e-01 1.22711027e+00 -6.66683018e-01
-5.06721437e-01 -4.68485564e-01 -5.83208561e-01 -1.82920873e+00
-2.55483314e-02 -5.91335058e-01 -1.57087460e-01 -1.43747199e+00
2.83563942e-01 -2.11969968e-02 7.60894492e-02 5.41942894e-01
-1.83756739e-01 7.76980758e-01 4.91091639e-01 4.24914509e-01
-8.42846155e-01 5.63616276e-01 1.56068289e+00 2.27662176e-02
-3.24407607e-01 -2.02205956e-01 -1.65621966e-01 8.73483896e-01
6.89134359e-01 -1.65074646e-01 -4.90302145e-01 -7.26500928e-01
1.01612121e-01 -8.67765471e-02 4.61814910e-01 -1.27060902e+00
5.49038172e-01 2.85684168e-01 5.61015248e-01 -9.87568200e-01
3.62693876e-01 -6.54563963e-01 4.96030509e-01 5.17538667e-01
-1.70665473e-01 3.57697457e-01 1.65582061e-01 2.49294415e-01
-1.99846476e-01 2.59422153e-01 7.76345015e-01 -5.07656410e-02
-8.83831561e-01 7.63891518e-01 2.18276307e-01 3.37417088e-02
8.30075502e-01 -2.94771463e-01 -5.13436735e-01 -7.02593803e-01
-6.19242489e-01 2.46923298e-01 5.66882789e-01 4.68060583e-01
7.00571060e-01 -1.46534944e+00 -3.60341191e-01 2.14131221e-01
-2.57394910e-01 1.47149846e-01 3.59089404e-01 1.09465516e+00
-9.84299600e-01 4.46421474e-01 -9.94695649e-02 -6.27746463e-01
-1.12978792e+00 4.59326208e-01 3.39336962e-01 -2.56635785e-01
-9.83658373e-01 9.96024132e-01 6.35634184e-01 -4.35016513e-01
3.80256385e-01 -3.28166604e-01 2.63160735e-01 -4.17028278e-01
4.05816704e-01 6.50587261e-01 -1.63538694e-01 -6.39510691e-01
-1.27311036e-01 7.78747022e-01 -2.81812727e-01 -1.50165215e-01
1.14587617e+00 -1.06914386e-01 -1.58533543e-01 7.75135756e-02
1.55718112e+00 -2.97150820e-01 -1.65858722e+00 -1.25974104e-01
-4.39128369e-01 -6.25498533e-01 -5.12546189e-02 -3.64297152e-01
-1.50246274e+00 6.94415092e-01 5.21293938e-01 -3.25211942e-01
1.14485633e+00 -1.65996715e-01 1.17174196e+00 5.22719264e-01
4.65829641e-01 -1.03042388e+00 2.42564783e-01 6.33625090e-01
1.06603372e+00 -1.33017063e+00 2.73520947e-01 -4.88442302e-01
-4.84622568e-01 1.07664275e+00 9.93630052e-01 -3.95199925e-01
6.24639392e-01 3.29116374e-01 1.16820663e-01 -2.85997152e-01
-6.26124263e-01 -4.51467521e-02 4.14369524e-01 9.14307982e-02
1.44727677e-01 -2.40093186e-01 -2.31931150e-01 1.68968290e-01
-4.87513900e-01 -2.27266476e-01 3.63684326e-01 6.10076487e-01
-9.58154127e-02 -7.83018649e-01 -2.05156654e-01 1.70869663e-01
-8.22673365e-02 -1.75944149e-01 -3.97662014e-01 1.30758512e+00
-4.74009141e-02 4.13730532e-01 7.09530562e-02 -7.10117638e-01
4.55342591e-01 -4.90765601e-01 2.62025774e-01 -1.08085439e-01
-4.12188590e-01 2.74316609e-01 -1.09371372e-01 -1.30357468e+00
-1.74204543e-01 -3.22216660e-01 -9.92415369e-01 -6.53570056e-01
-3.52514446e-01 -2.12279737e-01 2.78844446e-01 5.46766818e-01
4.62097555e-01 3.60811621e-01 5.27154028e-01 -1.25679326e+00
-2.81932056e-01 -7.80181468e-01 -2.58090824e-01 3.09227020e-01
3.99376363e-01 -5.51544785e-01 -5.62178254e-01 6.03846610e-02] | [8.643437385559082, -2.158951759338379] |
06336d1f-ae9f-4386-a62a-e2396d405ff5 | aspect-and-sentiment-aware-abstractive-review | null | null | https://aclanthology.org/C18-1095 | https://aclanthology.org/C18-1095.pdf | Aspect and Sentiment Aware Abstractive Review Summarization | Review text has been widely studied in traditional tasks such as sentiment analysis and aspect extraction. However, to date, no work is towards the abstractive review summarization that is essential for business organizations and individual consumers to make informed decisions. This work takes the lead to study the aspect/sentiment-aware abstractive review summarization by exploring multi-factor attentions. Specifically, we propose an interactive attention mechanism to interactively learns the representations of context words, sentiment words and aspect words within the reviews, acted as an encoder. The learned sentiment and aspect representations are incorporated into the decoder to generate aspect/sentiment-aware review summaries via an attention fusion network. In addition, the abstractive summarizer is jointly trained with the text categorization task, which helps learn a category-specific text encoder, locating salient aspect information and exploring the variations of style and wording of content with respect to different text categories. The experimental results on a real-life dataset demonstrate that our model achieves impressive results compared to other strong competitors. | ['Qiang Qu', 'Qiao Liu', 'Min Yang', 'Jia Zhu', 'Ying Shen', 'Wei Zhao'] | 2018-08-01 | aspect-and-sentiment-aware-abstractive-review-1 | https://aclanthology.org/C18-1095 | https://aclanthology.org/C18-1095.pdf | coling-2018-8 | ['aspect-extraction'] | ['natural-language-processing'] | [ 5.01982570e-01 1.26102358e-01 -5.59997499e-01 -6.60748959e-01
-1.01016986e+00 -3.15092653e-01 7.14492261e-01 2.81642586e-01
-2.27722988e-01 5.27302623e-01 9.84128594e-01 -1.37724310e-01
4.78213787e-01 -3.74292910e-01 -5.86074293e-01 -4.51896399e-01
5.15996993e-01 1.56885266e-01 -2.70514011e-01 -3.75073582e-01
8.53778839e-01 -9.39588994e-02 -1.40995872e+00 6.17668569e-01
9.85631943e-01 9.32816148e-01 1.83283523e-01 8.33988905e-01
-5.54976165e-01 1.00337744e+00 -1.10000372e+00 -6.68243229e-01
-2.23988533e-01 -6.67545438e-01 -5.40159881e-01 4.74439353e-01
3.92377198e-01 -3.44280928e-01 -1.23616330e-01 1.14068151e+00
4.90347356e-01 2.78475791e-01 9.37444508e-01 -5.51595986e-01
-1.40778553e+00 1.13226306e+00 -1.11593497e+00 2.66110361e-01
1.70301780e-01 -9.06482786e-02 1.58548915e+00 -1.02746677e+00
2.79960603e-01 1.36501002e+00 1.71616122e-01 3.45734954e-01
-4.33870554e-01 -3.47154438e-01 9.24429238e-01 8.34860653e-02
-7.39787281e-01 -3.64842445e-01 9.95500982e-01 -3.43136601e-02
1.29505968e+00 1.52466550e-01 6.62724376e-01 9.38649416e-01
7.00444341e-01 1.30628288e+00 2.95953453e-01 -2.54662037e-01
1.28532574e-01 3.78995925e-01 6.38868749e-01 3.89894605e-01
6.41529620e-01 -9.65644419e-01 -5.89415789e-01 1.69239283e-01
-4.48526740e-02 3.11672747e-01 -2.28932530e-01 1.89227000e-01
-9.73591745e-01 9.54283118e-01 2.83745438e-01 1.48680210e-01
-6.23433828e-01 2.60992050e-01 8.78989697e-01 1.94108710e-01
1.14214408e+00 5.86833954e-01 -6.40148580e-01 1.23609178e-01
-8.19981039e-01 1.61672771e-01 7.88317025e-01 1.22387600e+00
6.51377439e-01 3.87796402e-01 -3.63269448e-01 9.59110975e-01
4.87735778e-01 7.78244078e-01 8.46123219e-01 -3.87137741e-01
7.81530976e-01 1.05384779e+00 -2.37921759e-01 -1.01758051e+00
-2.44528785e-01 -7.52225161e-01 -8.16659093e-01 -3.74368072e-01
-8.47083986e-01 -5.05816996e-01 -9.09277499e-01 1.08952034e+00
-1.52231604e-01 -3.75890166e-01 3.25711668e-01 6.79702997e-01
1.42807043e+00 9.28825200e-01 1.14658065e-01 -3.67048621e-01
1.66458881e+00 -1.55891275e+00 -1.16779077e+00 -7.09596992e-01
5.14792919e-01 -8.99623334e-01 1.06207979e+00 2.10676745e-01
-1.22561944e+00 -5.01752138e-01 -1.50762355e+00 -4.25619185e-01
-4.58337635e-01 4.90036696e-01 5.93591750e-01 1.99131742e-01
-7.17081010e-01 1.89882845e-01 -4.99222785e-01 -2.72520125e-01
6.36982203e-01 2.43136287e-01 -6.50736913e-02 1.81256935e-01
-9.79596972e-01 7.17853844e-01 3.18608060e-02 1.54160142e-01
-5.58328569e-01 -4.65868860e-01 -1.21858311e+00 3.40890706e-01
3.76176685e-01 -1.08500433e+00 1.51470649e+00 -1.43777239e+00
-1.51131642e+00 5.32445133e-01 -5.26014447e-01 -2.63526440e-01
-2.04611361e-01 -6.27332866e-01 -4.01925564e-01 1.90165266e-01
3.74718189e-01 3.27234805e-01 9.68364000e-01 -1.21039534e+00
-6.76349938e-01 -3.67316246e-01 1.33217961e-01 7.00108171e-01
-4.62028831e-01 2.29592353e-01 -4.56346095e-01 -8.75251889e-01
-3.13472927e-01 -5.05798876e-01 -4.73296374e-01 -7.56978810e-01
-7.59530902e-01 -2.74199903e-01 6.22534096e-01 -6.88431561e-01
1.39192331e+00 -1.87546051e+00 2.43442342e-01 -2.34179944e-01
1.67415231e-01 2.36787692e-01 -1.98003456e-01 5.75662613e-01
-1.19314659e-02 3.53621393e-01 -3.52767080e-01 -7.72476375e-01
8.25266317e-02 -2.46185064e-01 -7.82120883e-01 1.25344679e-01
3.78201187e-01 1.31894886e+00 -7.68491983e-01 -2.88957447e-01
-1.50181785e-01 3.52037668e-01 -4.93333042e-01 1.78916052e-01
-4.10274267e-01 -8.54527876e-02 -9.10256326e-01 6.84253097e-01
4.00049448e-01 -3.91886979e-01 -1.70553043e-01 -1.22426018e-01
-3.33378762e-02 6.08870983e-01 -4.75014865e-01 1.50460756e+00
-5.20936012e-01 8.14923584e-01 -2.66717196e-01 -9.03343320e-01
1.09479284e+00 1.61960289e-01 7.65915066e-02 -5.27709186e-01
5.44166565e-01 1.07386455e-01 -2.72342652e-01 -4.19782937e-01
1.38919032e+00 -1.30840138e-01 -5.10703146e-01 6.78715289e-01
1.02779724e-01 -3.74609798e-01 2.33286008e-01 5.59978724e-01
8.96886051e-01 -2.09985003e-01 6.32758975e-01 -9.51310173e-02
7.36676574e-01 -2.66399328e-02 2.59532839e-01 6.02470458e-01
1.93529263e-01 7.42860377e-01 7.85457790e-01 -3.52962226e-01
-7.46760190e-01 -4.96893972e-01 3.40758592e-01 1.20843542e+00
3.18537176e-01 -6.72523916e-01 -6.10730290e-01 -8.59480381e-01
-2.19948769e-01 9.25068498e-01 -8.06342661e-01 -5.06070673e-01
-4.08366829e-01 -6.93530858e-01 -2.10602321e-02 6.71874821e-01
4.53556687e-01 -1.31677306e+00 -2.94882476e-01 1.30143151e-01
6.18432723e-02 -8.21318924e-01 -9.59908068e-01 2.60396540e-01
-6.49979591e-01 -7.30508208e-01 -7.24184513e-01 -9.17373300e-01
7.15664029e-01 5.80841660e-01 1.15443671e+00 -3.77162732e-02
1.05267152e-01 4.77882951e-01 -7.15538383e-01 -1.06232393e+00
-1.76525667e-01 6.29456401e-01 -3.74127656e-01 3.38969529e-02
6.93538368e-01 -3.07032377e-01 -7.49050200e-01 -1.88111991e-01
-9.41902578e-01 -4.50034738e-02 9.55401778e-01 5.78815699e-01
5.60521543e-01 -3.86200339e-01 1.21058202e+00 -1.46166146e+00
1.19709182e+00 -6.40716374e-01 2.30039489e-02 2.36474171e-01
-5.43498158e-01 -1.94245367e-04 8.02123725e-01 -1.44398823e-01
-1.32765019e+00 -4.01990533e-01 -1.06891051e-01 9.95285586e-02
3.61048579e-01 8.28317583e-01 -3.71884137e-01 6.69817567e-01
2.61887491e-01 5.05242229e-01 -3.25645566e-01 -2.72236049e-01
5.62859237e-01 1.00267303e+00 1.07149921e-01 7.84335583e-02
3.61713082e-01 3.90423328e-01 -6.44082606e-01 -8.99040639e-01
-1.47090733e+00 -5.50500631e-01 -3.78754079e-01 3.37616503e-02
9.37542319e-01 -1.20546424e+00 -1.50696218e-01 3.16129714e-01
-1.34063959e+00 2.47860014e-01 -4.59925503e-01 2.27135330e-01
-4.34719145e-01 4.41346943e-01 -3.68378669e-01 -6.69433773e-01
-1.08105338e+00 -1.19897866e+00 1.53476989e+00 6.52459085e-01
-3.55559170e-01 -9.15013909e-01 -6.12381808e-02 3.95978123e-01
3.87631804e-01 -1.65962487e-01 8.64347637e-01 -9.98672426e-01
-4.18319166e-01 -3.80662352e-01 -1.72205031e-01 6.61729455e-01
3.08722496e-01 9.64162648e-02 -8.94379854e-01 1.47584975e-01
1.67479798e-01 -1.91045001e-01 1.42494690e+00 6.14528894e-01
1.07380342e+00 -6.27311468e-01 -1.33148983e-01 3.57008159e-01
1.02223849e+00 4.11186755e-01 5.04503727e-01 1.69447839e-01
9.32405055e-01 5.96751273e-01 6.10448122e-01 4.39143807e-01
7.86809325e-01 -4.96275201e-02 2.54805744e-01 -6.47172630e-02
-1.38135673e-02 -9.89421904e-02 5.13566136e-01 1.53423607e+00
3.85999322e-01 -6.08384967e-01 -2.47285247e-01 7.76803613e-01
-1.79744911e+00 -8.86875689e-01 7.26785045e-03 1.40313685e+00
5.45608342e-01 2.20215827e-01 -3.51917207e-01 -2.05572739e-01
6.82027638e-01 8.45490992e-01 -7.96775460e-01 -1.06962597e+00
-2.59245992e-01 -3.24066393e-02 3.10764223e-01 2.52453029e-01
-1.15908539e+00 9.94346559e-01 5.75550318e+00 5.71820617e-01
-1.04212177e+00 -1.88340142e-01 8.24949503e-01 -3.15192878e-01
-9.42403257e-01 -1.46140426e-01 -8.79746914e-01 2.53606230e-01
6.99428737e-01 -5.91708362e-01 -1.57131568e-01 1.08733106e+00
2.94245362e-01 -4.86179963e-02 -8.14754784e-01 6.17981136e-01
8.58746171e-01 -1.36750841e+00 6.38695836e-01 -2.73771405e-01
1.28617144e+00 -1.20201074e-02 1.66607171e-01 5.91343105e-01
1.52643725e-01 -8.18726778e-01 5.44380128e-01 5.48946857e-01
4.34068263e-01 -1.04513383e+00 1.15804017e+00 -2.48305090e-02
-1.05386817e+00 6.99049383e-02 -5.71135879e-01 -1.67362913e-02
3.59118491e-01 5.46213150e-01 -4.92686689e-01 5.95867395e-01
3.02289009e-01 1.53466558e+00 -5.13059497e-01 5.19226015e-01
-5.84571421e-01 6.35024428e-01 4.61075574e-01 -5.22227407e-01
4.27532285e-01 -2.08010197e-01 6.15794957e-01 1.39203465e+00
3.12524587e-02 -9.21865255e-02 -4.84085120e-02 3.96218240e-01
-5.48016727e-01 5.43371737e-01 -5.05089581e-01 -3.80526096e-01
-2.66503133e-02 1.42394328e+00 -6.44770145e-01 -5.97386360e-01
-6.55833900e-01 9.00004864e-01 1.47853255e-01 5.32026947e-01
-5.63452303e-01 -8.29745591e-01 6.77109897e-01 -3.23424578e-01
7.31777906e-01 2.25506708e-01 -7.92888761e-01 -1.42644417e+00
1.91681340e-01 -9.31421638e-01 2.13059172e-01 -9.05437052e-01
-1.22193885e+00 7.56084204e-01 -3.64764482e-01 -1.06869149e+00
-8.88164863e-02 -3.91629219e-01 -1.14036059e+00 8.82113814e-01
-1.86314654e+00 -1.17534018e+00 7.51632601e-02 -4.46912535e-02
1.46776927e+00 -5.03050387e-01 4.45663899e-01 -9.60519314e-02
-7.72256911e-01 5.38902938e-01 6.94108680e-02 -1.27758384e-01
7.03631520e-01 -1.42440724e+00 6.74633384e-01 8.16959202e-01
-1.13270611e-01 8.73004079e-01 6.55229330e-01 -8.23505878e-01
-1.43036675e+00 -1.27688766e+00 1.07555151e+00 -3.51409018e-01
6.65171027e-01 -1.38731912e-01 -7.80254185e-01 8.29491198e-01
9.78801131e-01 -6.92052364e-01 1.03607941e+00 1.26846939e-01
-2.38447323e-01 -1.59303039e-01 -5.19375026e-01 8.55429947e-01
6.47312641e-01 -4.69275564e-01 -9.90500689e-01 1.33818194e-01
1.37406969e+00 -8.44199732e-02 -3.52209747e-01 2.06896961e-01
4.68128026e-01 -4.41991359e-01 5.18081486e-01 -6.11549079e-01
1.23750746e+00 -2.17848003e-01 -9.56281126e-02 -1.64502048e+00
-1.00495324e-01 -4.33882505e-01 -2.12944463e-01 1.64210331e+00
6.71116590e-01 -4.14800197e-01 4.43878293e-01 1.94560260e-01
-6.33896649e-01 -1.08243251e+00 -3.19907874e-01 1.66009203e-01
-2.86006294e-02 -2.17858717e-01 6.97914779e-01 3.18463564e-01
1.75609201e-01 1.30992389e+00 -4.07006145e-01 -1.31669626e-01
-2.39858925e-02 4.77459908e-01 7.66260087e-01 -7.55237401e-01
-3.81727517e-03 -7.90799081e-01 -8.22777525e-02 -1.42522895e+00
4.26848739e-01 -8.74252796e-01 1.12806179e-01 -2.20864081e+00
5.43407798e-01 3.59441668e-01 -2.40355164e-01 -2.15414353e-03
-8.13336372e-01 -1.12721026e-01 6.05900213e-02 4.33234870e-02
-1.26313746e+00 1.12211108e+00 1.54129791e+00 -6.64826810e-01
-9.65858996e-02 1.10643387e-01 -2.04108214e+00 7.47076452e-01
6.61809921e-01 -3.06195855e-01 -6.31253719e-01 -5.92134118e-01
6.09315932e-01 -2.18413010e-01 -4.23425615e-01 -4.30932522e-01
2.09591761e-01 7.06802905e-02 3.97842377e-01 -1.29286468e+00
7.21852109e-02 -5.26308119e-01 -7.15717614e-01 -6.94399998e-02
-8.77309442e-01 3.06003749e-01 6.11169077e-02 7.42588639e-01
-5.16702116e-01 -4.23768520e-01 2.81418294e-01 -2.66778767e-01
-4.50715601e-01 2.55068332e-01 -6.41324818e-01 2.30117366e-01
6.38670802e-01 4.71456088e-02 -5.60064018e-01 -7.39358366e-01
-1.95177291e-02 5.90699852e-01 2.36794263e-01 6.95376158e-01
6.64341986e-01 -9.48161781e-01 -9.12103355e-01 2.09087897e-02
3.33357871e-01 2.64273703e-01 4.49221611e-01 3.54623079e-01
-2.54552752e-01 5.03527880e-01 3.25246513e-01 -2.96371549e-01
-1.07761145e+00 5.40965497e-01 -1.91170499e-02 -5.43635666e-01
-5.04516065e-01 7.79615819e-01 5.80013454e-01 -2.15121433e-01
8.43436271e-02 -5.00839472e-01 -9.17088389e-01 5.61863065e-01
1.06554985e+00 1.17344648e-01 8.12318996e-02 -5.35745203e-01
-7.10396767e-02 7.20994473e-01 -7.35904813e-01 1.30191520e-01
1.50097406e+00 -5.78016460e-01 -2.13616654e-01 6.58283114e-01
1.42459297e+00 2.06615224e-01 -1.01090896e+00 -1.94283545e-01
9.22954157e-02 1.43973483e-02 4.76519503e-02 -8.01451564e-01
-1.05479801e+00 9.03784692e-01 -3.28736693e-01 1.75256744e-01
1.09181845e+00 9.33698639e-02 9.38067436e-01 5.56433082e-01
-3.44867885e-01 -1.17622638e+00 2.59539962e-01 8.38436425e-01
1.18370390e+00 -1.42430258e+00 3.54094177e-01 -2.94982139e-02
-1.30825734e+00 9.73573208e-01 6.27663553e-01 -4.79955524e-01
4.53314632e-01 8.44090059e-02 2.62797564e-01 -4.49036628e-01
-1.18081093e+00 -1.72258466e-01 5.54691195e-01 2.70876229e-01
6.27178729e-01 1.58453006e-02 -5.70271969e-01 1.23319650e+00
-2.95319915e-01 -5.14906943e-01 8.59700024e-01 6.97175860e-01
-8.13040435e-01 -7.40070641e-01 1.64623648e-01 9.89712715e-01
-9.31951523e-01 -4.40456063e-01 -7.07700968e-01 3.38396043e-01
-4.41737205e-01 9.97712553e-01 1.21168561e-01 -1.79211006e-01
5.72369874e-01 -4.95213009e-02 -2.04305872e-01 -1.24794614e+00
-1.14411426e+00 1.88252568e-01 2.54774958e-01 -5.19146491e-03
-3.72863322e-01 -5.79067945e-01 -1.21500516e+00 6.13490045e-02
-3.45116526e-01 3.99579853e-01 8.79980385e-01 1.06225491e+00
4.87075359e-01 1.20044661e+00 9.26560402e-01 -7.14422643e-01
-4.08895582e-01 -1.16292167e+00 -5.44046104e-01 7.44721442e-02
5.68165898e-01 -1.44799903e-01 -4.24979985e-01 1.64741296e-02] | [11.473560333251953, 6.758363246917725] |
09da6b80-a7c3-4ffa-b060-82497f1f1a80 | efficient-remote-photoplethysmography-with | 2203.10882 | null | https://arxiv.org/abs/2203.10882v2 | https://arxiv.org/pdf/2203.10882v2.pdf | Efficient Remote Photoplethysmography with Temporal Derivative Modules and Time-Shift Invariant Loss | We present a lightweight neural model for remote heart rate estimation focused on the efficient spatio-temporal learning of facial photoplethysmography (PPG) based on i) modelling of PPG dynamics by combinations of multiple convolutional derivatives, and ii) increased flexibility of the model to learn possible offsets between the facial video PPG and the ground truth. PPG dynamics are modelled by a Temporal Derivative Module (TDM) constructed by the incremental aggregation of multiple convolutional derivatives, emulating a Taylor series expansion up to the desired order. Robustness to ground truth offsets is handled by the introduction of TALOS (Temporal Adaptive LOcation Shift), a new temporal loss to train learning-based models. We verify the effectiveness of our model by reporting accuracy and efficiency metrics on the public PURE and UBFC-rPPG datasets. Compared to existing models, our approach shows competitive heart rate estimation accuracy with a much lower number of parameters and lower computational cost. | ['Federico Sukno', 'Adria Ruiz', 'Joaquim Comas'] | 2022-03-21 | null | null | null | null | ['photoplethysmography-ppg', 'heart-rate-estimation'] | ['medical', 'medical'] | [ 8.88703242e-02 3.77482712e-01 -8.86861701e-03 -3.20203990e-01
-5.35993338e-01 -2.95428872e-01 5.65145910e-01 -3.29356611e-01
-3.53600085e-01 5.96941292e-01 1.05370604e-01 -5.69019169e-02
1.52563592e-02 -4.09810990e-01 -7.42108047e-01 -6.59652174e-01
-5.71810126e-01 8.12614337e-02 6.77043572e-02 2.51133651e-01
-1.30218044e-01 7.85187900e-01 -1.16845024e+00 1.00854196e-01
5.07661819e-01 1.43215489e+00 -7.27426827e-01 1.02657151e+00
3.85117114e-01 9.59650397e-01 -2.72909492e-01 -5.37940145e-01
5.57551444e-01 -3.86996418e-01 -6.29215717e-01 -9.38192010e-02
8.85145605e-01 -7.55583763e-01 -3.45407695e-01 2.48928532e-01
9.49409485e-01 -1.30157292e-01 2.52431571e-01 -1.26286089e+00
-2.54613668e-01 5.23492843e-02 -4.01299566e-01 2.09565327e-01
1.38924196e-01 3.32890511e-01 4.45557207e-01 -8.14544380e-01
6.04109526e-01 9.81997788e-01 1.45771611e+00 7.02712059e-01
-1.36551261e+00 -4.13080782e-01 -1.60808399e-01 7.10384920e-02
-1.53253317e+00 -7.14105189e-01 6.03159368e-01 -2.63778478e-01
8.97344708e-01 2.17926025e-01 9.13742721e-01 8.77138197e-01
4.72606048e-02 4.26852614e-01 7.79791653e-01 -2.00002685e-01
1.04931630e-01 3.95119302e-02 -4.81309354e-01 9.70296562e-01
-2.73399025e-01 3.61830175e-01 -6.79249763e-01 -3.17416549e-01
1.01447201e+00 -2.38257959e-01 -2.58559287e-01 -3.61517251e-01
-6.94205403e-01 3.18247318e-01 1.98695675e-01 -5.86973876e-02
-6.85813665e-01 6.52634382e-01 4.65863496e-01 7.87266269e-02
6.07811391e-01 -3.86398584e-02 -5.70086062e-01 -2.32259378e-01
-1.30241930e+00 1.16228387e-01 8.61068666e-01 6.31826520e-01
5.65347254e-01 1.20547801e-01 -2.85487115e-01 5.09965837e-01
3.03234309e-01 4.86085445e-01 2.13500798e-01 -1.41280675e+00
1.37597397e-02 2.50365227e-01 2.03348503e-01 -9.51775432e-01
-6.04799807e-01 -2.42141604e-01 -6.94981515e-01 -1.69949960e-02
7.11666524e-01 -4.61225539e-01 -7.04399824e-01 1.81076682e+00
6.76938057e-01 1.04542327e+00 -2.56434143e-01 6.97054625e-01
6.73162401e-01 2.79597491e-01 1.43309370e-01 -4.20135647e-01
1.07128692e+00 -3.89990538e-01 -4.00590032e-01 2.43089721e-01
7.58975923e-01 -2.37587303e-01 5.30538201e-01 2.40305752e-01
-1.21837521e+00 -4.23782825e-01 -6.01526082e-01 -2.48760566e-01
-1.41306609e-01 2.90445894e-01 6.31602705e-01 8.19127738e-01
-1.49826932e+00 1.17520440e+00 -9.00102735e-01 -2.56049097e-01
6.20041966e-01 5.10711372e-01 -2.46380299e-01 2.93637544e-01
-1.27073812e+00 7.75041640e-01 -7.59070143e-02 5.10603368e-01
-6.29654765e-01 -1.60475683e+00 -9.21498537e-01 -1.26768211e-02
2.65250206e-02 -5.94247460e-01 1.04004264e+00 -1.17754674e+00
-1.89927185e+00 1.02104270e+00 -6.47013634e-02 -7.63397872e-01
1.09803832e+00 -1.28432468e-01 -2.60007083e-01 6.41061008e-01
-4.97599334e-01 8.52199376e-01 1.06106663e+00 -5.40555418e-01
-1.52158558e-01 -8.60293880e-02 -3.33965689e-01 2.59047057e-02
-5.40360995e-02 -1.21353909e-01 -4.00626093e-01 -2.98758000e-01
-2.61684895e-01 -9.06922281e-01 -2.21659485e-02 6.11243486e-01
1.34212062e-01 -6.84004091e-03 6.69013500e-01 -1.08954847e+00
1.20845556e+00 -2.00228286e+00 -3.66390556e-01 1.34433135e-01
3.60066712e-01 5.16812563e-01 -2.04412952e-01 3.51669081e-02
-3.42790902e-01 1.08717211e-01 2.15999093e-02 -5.87189674e-01
-1.20211862e-01 6.50118757e-03 -9.24383327e-02 7.25031435e-01
3.11392814e-01 1.27580225e+00 -9.00900304e-01 -5.61925411e-01
3.52163970e-01 1.01209879e+00 -5.24242103e-01 -1.56597316e-01
3.87328602e-02 3.40845019e-01 2.47575957e-02 6.62974596e-01
7.57018089e-01 -2.00535059e-01 1.73314765e-01 -3.06205958e-01
-1.03727698e-01 1.72889307e-01 -8.69227648e-01 1.51368761e+00
-3.55634660e-01 6.09452426e-01 -2.01117814e-01 -5.28564692e-01
8.18562150e-01 6.90497458e-01 1.02065945e+00 -6.62676871e-01
4.87360284e-02 1.17709385e-02 -2.79440582e-01 -5.73581874e-01
9.31100100e-02 -2.17529908e-01 4.71666545e-01 3.26036394e-01
1.29736885e-01 2.18806997e-01 -1.77167088e-01 -8.80980566e-02
8.80454540e-01 6.46640062e-01 8.63269418e-02 -2.29932666e-01
4.78438675e-01 -5.75209975e-01 5.84706068e-01 6.73621178e-01
-5.62450349e-01 7.28142858e-01 8.90673220e-01 -9.39075768e-01
-1.03708148e+00 -7.64970660e-01 -2.47798070e-01 5.74575722e-01
-3.76200587e-01 -3.22081596e-01 -8.13385844e-01 -6.73721611e-01
2.63037741e-01 2.01577410e-01 -1.03255606e+00 -1.94586799e-01
-7.44128883e-01 -7.52135098e-01 1.10140479e+00 5.70604563e-01
7.23733366e-01 -8.30212593e-01 -8.40384305e-01 1.50753021e-01
-1.40091434e-01 -1.15284109e+00 -5.21441340e-01 -3.26970518e-01
-1.12522948e+00 -1.08481026e+00 -8.45333338e-01 -1.82822660e-01
3.49610686e-01 -3.89855564e-01 1.18078959e+00 7.41581246e-02
-7.44641185e-01 7.50191033e-01 9.90728214e-02 -3.87955040e-01
-2.48079285e-01 -1.58155486e-01 5.66867627e-02 4.02250379e-01
2.92853504e-01 -7.11638510e-01 -1.07957721e+00 6.36094362e-02
-5.85892022e-01 -5.28089963e-02 1.34728372e-01 4.74583536e-01
4.74625558e-01 -6.69412911e-01 4.14324582e-01 -6.05175316e-01
2.16311559e-01 -2.19738469e-01 -9.32247460e-01 1.91714436e-01
-6.70172513e-01 -1.40854850e-01 3.37193578e-01 -5.47314167e-01
-9.03510451e-01 1.66499391e-01 7.52561018e-02 -1.00507498e+00
1.32384405e-01 1.99901193e-01 4.02233034e-01 -4.63982522e-01
7.15431511e-01 3.21416706e-01 4.51042056e-01 -3.36393058e-01
4.30123597e-01 4.17672358e-02 6.89282835e-01 -4.40165579e-01
4.50397581e-01 6.31536663e-01 6.11716390e-01 -9.92884040e-01
-5.34460187e-01 -2.29156375e-01 -9.16643858e-01 -5.76976538e-01
5.78725457e-01 -9.14335310e-01 -1.08309674e+00 6.83519065e-01
-1.03336942e+00 -6.94899678e-01 -4.44758505e-01 5.10166228e-01
-7.10799336e-01 6.13965988e-01 -7.03675270e-01 -9.83290434e-01
-5.56831181e-01 -5.01856565e-01 1.17896831e+00 1.14392623e-01
-1.28274903e-01 -1.28487635e+00 2.32269496e-01 -2.94805709e-02
5.81398189e-01 7.84189403e-01 4.15417254e-01 -1.68809876e-01
-5.06218910e-01 -4.71886784e-01 -1.70791626e-01 5.52477241e-01
-5.84385358e-02 2.30783418e-01 -1.33056402e+00 -1.81184590e-01
-1.14425220e-01 -1.92941397e-01 7.62177229e-01 8.24701607e-01
1.09044707e+00 -5.30021966e-01 -4.80374694e-02 1.10249376e+00
1.25752842e+00 -2.74885982e-01 1.01007068e+00 -1.14385679e-01
5.56823373e-01 4.24671173e-01 2.15667650e-01 7.58613408e-01
2.39937261e-01 5.01126647e-01 4.13252443e-01 -3.58748972e-01
-2.36109108e-01 -1.48447499e-01 2.86186695e-01 -1.24561237e-02
-5.79019189e-01 2.92375952e-01 -6.49094343e-01 3.54552239e-01
-1.61650038e+00 -1.06814945e+00 -9.01596714e-03 2.54815412e+00
7.19419777e-01 -4.55686688e-01 4.48429108e-01 -1.91834688e-01
3.72525871e-01 1.82312191e-01 -5.79876900e-01 -3.11434567e-01
2.17930511e-01 3.16998571e-01 7.37526417e-01 5.78036427e-01
-1.07494295e+00 7.38834321e-01 7.40673161e+00 2.84184605e-01
-1.41796875e+00 -3.67219152e-04 8.44402969e-01 -4.26597744e-01
1.56766340e-01 -1.76883236e-01 -7.18247771e-01 5.52121997e-01
1.20047081e+00 1.32169574e-01 5.17891526e-01 5.85173070e-01
4.92765337e-01 9.00640786e-02 -9.99910712e-01 1.02199566e+00
9.26597640e-02 -1.49614501e+00 -3.79473835e-01 1.16042726e-01
3.74097377e-01 5.41004054e-02 -1.68981068e-02 3.93958241e-02
-1.58306450e-01 -8.91759038e-01 3.31929564e-01 9.02464688e-01
1.15037477e+00 -4.93801594e-01 5.21302760e-01 -1.44527078e-01
-9.75257039e-01 4.24084254e-02 -2.12754637e-01 1.73870042e-01
-1.80605099e-01 3.90894383e-01 -1.00674605e+00 3.30321044e-01
6.16672099e-01 7.40217090e-01 -3.66642028e-01 9.95249808e-01
-7.67153203e-02 5.38543224e-01 -7.51295388e-01 4.02524531e-01
-8.87874216e-02 -7.45286793e-02 3.69807512e-01 1.09690917e+00
2.99209833e-01 1.44481167e-01 -4.30045843e-01 8.25286686e-01
-1.25433564e-01 -7.60169104e-02 -3.98656398e-01 1.51431963e-01
2.53298938e-01 1.33477771e+00 -1.98296428e-01 -6.26524985e-02
-3.73956084e-01 8.87316942e-01 3.99717279e-02 4.15203899e-01
-1.13413990e+00 -1.33472368e-01 7.24726439e-01 5.49443126e-01
2.46261805e-01 6.20298199e-02 -1.50386826e-03 -9.55543220e-01
1.37028143e-01 -3.89811337e-01 7.30226815e-01 -7.86341310e-01
-8.23630154e-01 2.66641915e-01 -1.00377277e-02 -1.08833241e+00
-4.65270251e-01 -4.37136829e-01 -5.84919691e-01 1.10132444e+00
-1.76973546e+00 -1.21341515e+00 -4.30626005e-01 5.20251989e-01
-1.47200346e-01 3.37133259e-01 7.40888178e-01 4.19563442e-01
-4.58493173e-01 9.90855098e-01 -5.13618886e-01 1.33977950e-01
7.33004212e-01 -1.07443321e+00 3.23602796e-01 7.72903204e-01
-4.37995344e-01 3.17873120e-01 4.60012078e-01 -5.55433989e-01
-1.09340477e+00 -1.16571748e+00 1.20747364e+00 -6.46148682e-01
4.95620072e-01 -2.80682474e-01 -8.10531557e-01 7.46075273e-01
-2.76408166e-01 6.90867245e-01 4.07314390e-01 -1.32241294e-01
-3.13097030e-01 -5.31531990e-01 -1.43792260e+00 4.20927405e-01
8.19142997e-01 -4.59786355e-01 1.03315324e-01 2.34133422e-01
3.73285532e-01 -6.42242372e-01 -1.17819941e+00 3.63689333e-01
1.01458061e+00 -1.02670062e+00 9.88251865e-01 -5.71538210e-01
2.76016265e-01 -1.86199024e-01 4.05482233e-01 -7.00719476e-01
-1.04470760e-01 -1.32346380e+00 -5.89078069e-01 8.76044929e-01
6.78206533e-02 -5.72832823e-01 1.06621492e+00 1.09236741e+00
2.73407519e-01 -7.80665934e-01 -1.18415976e+00 -5.66138029e-01
-1.37248442e-01 -3.62783730e-01 4.91610497e-01 8.21451306e-01
-2.62340736e-02 -3.55186880e-01 -8.18078876e-01 2.72061139e-01
6.53213203e-01 -2.94809133e-01 6.21939301e-01 -1.10881746e+00
-3.70352656e-01 -2.61844426e-01 -5.46419978e-01 -7.37786651e-01
-1.14597104e-01 -5.95442235e-01 -1.82852849e-01 -9.66344655e-01
-1.45562544e-01 -1.93219990e-01 -3.05483609e-01 8.82109940e-01
5.40096387e-02 5.42997181e-01 1.61552858e-02 -5.60972504e-02
-2.93392837e-01 3.15026224e-01 1.04495454e+00 5.19062936e-01
-5.24554610e-01 -1.42227057e-02 -2.80743957e-01 5.90371311e-01
6.39212608e-01 -3.58095527e-01 -3.19398403e-01 6.67584175e-03
2.49957159e-01 4.60673124e-01 7.84761310e-01 -9.41049933e-01
1.07225381e-01 8.23828802e-02 7.34105945e-01 -2.23416954e-01
3.62665743e-01 -5.85901618e-01 3.18954676e-01 6.39211297e-01
-2.48851895e-01 -2.11172670e-01 3.89184475e-01 2.12382868e-01
3.40564162e-01 2.82076180e-01 1.07871366e+00 -2.59334203e-02
-3.64527285e-01 7.43716598e-01 -4.24572602e-02 -1.83529004e-01
6.52588964e-01 -4.02059793e-01 -8.70656371e-02 -5.52636504e-01
-8.84534299e-01 2.29768902e-02 2.42071599e-01 1.50288507e-01
3.86521757e-01 -1.25758886e+00 -6.60553038e-01 3.50732774e-01
-1.70879647e-01 -2.87928432e-01 3.75053555e-01 1.50007999e+00
-7.31862843e-01 1.93459138e-01 3.52909206e-03 -5.63153386e-01
-1.26363575e+00 2.63532341e-01 1.31298518e+00 -1.69665426e-01
-8.65246058e-01 6.32052660e-01 -1.10577382e-01 -3.71078134e-01
3.56202304e-01 -3.94993812e-01 1.20097302e-01 -1.53723992e-02
6.13098443e-01 5.71975470e-01 1.60611436e-01 -3.75785589e-01
-3.65605772e-01 4.89395261e-01 3.92081767e-01 1.20441049e-01
1.31139970e+00 -2.15033025e-01 -2.00358406e-02 9.07696337e-02
1.22311234e+00 -2.28683099e-01 -1.93721724e+00 -1.75776154e-01
-2.03677326e-01 -2.76368409e-01 1.06054835e-01 -1.10560739e+00
-1.24306226e+00 7.81664312e-01 8.70040417e-01 -2.08959058e-01
1.26530957e+00 -3.74363840e-01 5.21085024e-01 8.94372985e-02
-3.17129157e-02 -1.08239543e+00 -7.10721388e-02 2.82646507e-01
6.22533739e-01 -8.81140590e-01 1.43666312e-01 -2.74601132e-01
-4.13172454e-01 1.32411230e+00 2.57977515e-01 -1.41621187e-01
7.86645949e-01 2.12862417e-01 3.48936886e-01 -3.88737954e-02
-7.68696427e-01 9.16098729e-02 3.28812212e-01 6.50777817e-01
3.82379740e-01 -4.50046688e-01 -9.70270485e-02 1.24206647e-01
1.85659930e-01 7.69120157e-01 3.03047270e-01 4.88362968e-01
1.37331232e-01 -7.23412037e-01 1.19994201e-01 2.19209462e-01
-6.87348783e-01 -1.61146149e-01 -6.80514798e-02 1.03766072e+00
-4.53446470e-02 3.67351055e-01 2.98538536e-01 -2.41322499e-02
2.08334774e-01 4.44749743e-01 6.22446418e-01 2.68620681e-02
-5.04484177e-01 2.08538864e-02 -4.63584391e-03 -1.04472041e+00
-4.25633520e-01 -8.07833552e-01 -8.71289134e-01 -4.08651382e-01
6.86536133e-02 -2.64620423e-01 5.67569911e-01 8.28656137e-01
7.63441801e-01 1.23065196e-01 7.66637266e-01 -8.46213818e-01
-5.06129801e-01 -7.58817494e-01 -7.00875163e-01 2.93388069e-01
6.58088028e-01 -1.76862329e-01 -5.02902031e-01 1.88103348e-01] | [13.896750450134277, 2.788062572479248] |
011c177e-25f5-41a1-9a15-a22257ed184e | on-the-use-of-learning-based-forecasting | 2211.04798 | null | https://arxiv.org/abs/2211.04798v1 | https://arxiv.org/pdf/2211.04798v1.pdf | On the use of learning-based forecasting methods for ameliorating fashion business processes: A position paper | The fashion industry is one of the most active and competitive markets in the world, manufacturing millions of products and reaching large audiences every year. A plethora of business processes are involved in this large-scale industry, but due to the generally short life-cycle of clothing items, supply-chain management and retailing strategies are crucial for good market performance. Correctly understanding the wants and needs of clients, managing logistic issues and marketing the correct products are high-level problems with a lot of uncertainty associated to them given the number of influencing factors, but most importantly due to the unpredictability often associated with the future. It is therefore straightforward that forecasting methods, which generate predictions of the future, are indispensable in order to ameliorate all the various business processes that deal with the true purpose and meaning of fashion: having a lot of people wear a particular product or style, rendering these items, people and consequently brands fashionable. In this paper, we provide an overview of three concrete forecasting tasks that any fashion company can apply in order to improve their industrial and market impact. We underline advances and issues in all three tasks and argue about their importance and the impact they can have at an industrial level. Finally, we highlight issues and directions of future work, reflecting on how learning-based forecasting methods can further aid the fashion industry. | ['Marco Cristani', 'Matteo Denitto', 'Christian Joppi', 'Geri Skenderi'] | 2022-11-09 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [ 6.39922619e-02 -3.05604577e-01 -3.04602176e-01 -4.94342655e-01
3.06344479e-02 -7.05696225e-01 2.29705483e-01 9.80128050e-02
-1.30594492e-01 5.52069247e-01 1.54971898e-01 -4.66800153e-01
-5.91600597e-01 -9.06753004e-01 -6.02306664e-01 -5.70240259e-01
-1.56256687e-02 6.68321490e-01 -2.44896084e-01 -5.94467342e-01
3.02632421e-01 6.14184320e-01 -1.63926089e+00 4.32735085e-01
4.51797068e-01 1.55069244e+00 3.07024896e-01 4.72772598e-01
-2.41632700e-01 7.45518148e-01 -4.31865484e-01 -9.03161466e-01
1.62210703e-01 -2.04744905e-01 -3.86884719e-01 4.84549642e-01
-1.88962519e-01 -1.66385859e-01 3.43479812e-01 7.05540299e-01
-2.49105208e-02 1.07931435e-01 4.41304415e-01 -1.07365584e+00
-6.36506438e-01 5.38452148e-01 -2.13804945e-01 -1.14662573e-01
3.71890754e-01 1.82724774e-01 9.96693134e-01 -4.70512241e-01
2.99209207e-01 9.70374405e-01 6.31732941e-01 -1.07309058e-01
-1.23712230e+00 -4.05561358e-01 5.95864713e-01 1.65544152e-01
-7.94138849e-01 -2.03045398e-01 7.73477733e-01 -5.56257546e-01
7.40564704e-01 5.74159026e-01 1.05015528e+00 9.72333729e-01
7.07018137e-01 4.66671497e-01 1.14986420e+00 -2.76993155e-01
2.35825941e-01 3.27220261e-01 -1.48972213e-01 -1.54415682e-01
7.73947239e-02 -1.09566590e-02 -1.74151391e-01 2.28025451e-01
6.90410435e-01 3.40813577e-01 2.65442163e-01 -1.95262611e-01
-1.10656941e+00 9.18891847e-01 2.89856315e-01 5.17589748e-01
-7.29986787e-01 1.33335039e-01 2.49217570e-01 6.49938345e-01
3.96562099e-01 6.55890167e-01 -7.33404696e-01 -3.64867121e-01
-7.63243020e-01 4.70827013e-01 1.13460577e+00 7.74167180e-01
3.13687384e-01 -2.20018521e-01 4.25937653e-01 5.83513498e-01
2.55721867e-01 3.39613944e-01 6.66897967e-02 -9.58550394e-01
1.03907892e-02 4.87082273e-01 4.91373926e-01 -1.53706849e+00
-4.86077935e-01 -8.86840999e-01 -7.92574286e-01 -9.56256036e-03
6.53513849e-01 -1.03487112e-01 -5.62771320e-01 1.24675322e+00
7.41367564e-02 -4.94919956e-01 -4.51313287e-01 9.05589283e-01
2.20505908e-01 7.09947765e-01 3.58998030e-01 -4.72904891e-01
1.37198842e+00 -6.71219051e-01 -8.54473472e-01 -5.88830411e-01
3.88046019e-02 -1.43167317e+00 7.95511186e-01 7.22206056e-01
-1.15585685e+00 -6.03055477e-01 -7.10775614e-01 3.25802952e-01
-4.03418273e-01 -1.62083209e-01 1.06639361e+00 4.27338570e-01
-2.90619254e-01 9.92813647e-01 -5.68699181e-01 -3.08108747e-01
2.79407613e-02 4.74816501e-01 1.00211315e-01 -1.72285259e-01
-1.14096606e+00 1.15471017e+00 4.47755605e-02 6.84159696e-01
-7.13299736e-02 -6.99240327e-01 -4.53353703e-01 1.51749011e-02
5.55692971e-01 -4.54535961e-01 1.26323521e+00 -1.07095551e+00
-9.37694848e-01 3.34347069e-01 1.58749297e-01 -2.78596014e-01
8.13243032e-01 -1.17454499e-01 -8.11345160e-01 -6.56136632e-01
-1.18869834e-01 4.22479399e-02 7.10703075e-01 -1.18211758e+00
-1.03721130e+00 -4.04994398e-01 -1.42028630e-01 1.82097834e-02
1.60622820e-01 1.35783583e-01 -3.78346592e-02 -4.45660591e-01
3.13960493e-01 -1.12063181e+00 -6.09541774e-01 -5.16657472e-01
-2.70514339e-02 -7.96909034e-02 5.21224141e-01 -9.35456038e-01
1.19669211e+00 -2.01430011e+00 1.30859256e-01 3.01779866e-01
-2.49974504e-02 -2.36549839e-01 3.64651680e-01 9.29363251e-01
2.22532719e-01 -5.60297966e-02 3.64862621e-01 1.98872194e-01
2.27850080e-01 3.94169569e-01 -4.45836633e-01 1.80500537e-01
1.40087023e-01 8.83868694e-01 -7.74401307e-01 2.90700972e-01
4.82859522e-01 5.70415139e-01 1.89328026e-02 1.89032704e-02
-6.13675714e-01 6.01981461e-01 -5.20682514e-01 7.45410502e-01
3.92781138e-01 -1.92649007e-01 5.98495960e-01 -2.61556566e-01
-2.45717019e-01 -5.95885217e-02 -1.31932724e+00 9.46817338e-01
-7.60364652e-01 1.97722226e-01 1.69848010e-01 -8.73325229e-01
9.46417093e-01 1.63823739e-02 4.88490999e-01 -8.63601685e-01
3.44925612e-01 4.85375404e-01 1.51803359e-01 -4.89474922e-01
7.33588040e-01 -5.69952190e-01 -3.93458635e-01 3.27392638e-01
-4.22337741e-01 -9.66841578e-02 2.80594170e-01 -2.32183129e-01
6.51607811e-01 6.80877268e-02 1.08443156e-01 -1.59311414e-01
2.46640012e-01 1.70065537e-01 6.66475892e-01 2.07745984e-01
-5.47686480e-02 1.94104210e-01 6.39291465e-01 -1.05291545e+00
-1.16646087e+00 -6.35259807e-01 -4.98993136e-02 1.08099341e+00
9.10634771e-02 1.01945281e-01 -8.45155716e-02 -2.50235617e-01
4.29740459e-01 9.72072661e-01 -4.34021264e-01 2.59241741e-02
-4.76845950e-01 -6.52643621e-01 -6.55619979e-01 4.86429453e-01
-2.47119330e-02 -9.32805240e-01 -4.54381287e-01 8.31225455e-01
-3.23568918e-02 -1.09011579e+00 -2.06897318e-01 2.76024342e-01
-6.67135239e-01 -9.03087914e-01 -4.14941132e-01 -6.52567983e-01
3.67212743e-01 1.89359382e-01 1.40995479e+00 -1.99334368e-01
-2.04330176e-01 1.31398782e-01 -2.45839149e-01 -5.87152362e-01
-7.18579948e-01 -1.11712381e-01 3.62022314e-04 9.29321796e-02
5.81732631e-01 -5.71935892e-01 -6.79582357e-01 7.48625696e-01
-5.72274566e-01 1.68001875e-01 7.48501837e-01 4.89414871e-01
5.13363779e-01 8.11256230e-01 8.58978748e-01 -9.46163237e-01
6.74952269e-01 -6.46754563e-01 -5.30642867e-01 2.47477323e-01
-8.22176039e-01 -3.53127539e-01 5.38928628e-01 -6.07673407e-01
-9.30474222e-01 6.16310500e-02 -2.66273290e-01 1.94062084e-01
-2.21273452e-01 4.83080417e-01 -3.95533741e-02 3.15542907e-01
1.17572658e-01 -1.02308668e-01 1.92335695e-01 -5.52895486e-01
4.32756394e-01 5.85202038e-01 2.92915255e-01 -2.84389734e-01
6.73283398e-01 2.52971053e-01 1.17404468e-01 -5.79108059e-01
-8.19515884e-01 -4.46675509e-01 -6.31325245e-01 -7.09100723e-01
6.71541393e-01 -6.55442953e-01 -9.80028927e-01 3.76457095e-01
-8.77074718e-01 -4.56903167e-02 -4.16461647e-01 4.49403793e-01
-4.67942268e-01 -4.78628486e-01 -6.04053795e-01 -9.73629355e-01
5.12397699e-02 -1.25920892e+00 6.17705464e-01 1.41599387e-01
-6.43417239e-01 -1.04777813e+00 -5.73925138e-01 7.69656777e-01
8.46086860e-01 5.17617822e-01 1.16678357e+00 -4.40302014e-01
-7.02640831e-01 -4.81818944e-01 -3.91240008e-02 4.47433800e-01
3.52413625e-01 -1.25141293e-01 -5.84358096e-01 -1.34401664e-01
2.06428811e-01 1.24768019e-01 4.30571437e-01 7.31950819e-01
4.97492880e-01 -2.02302232e-01 -8.62619057e-02 -6.57891631e-02
1.50129950e+00 6.08123064e-01 3.53487432e-01 3.82074565e-01
4.29483056e-01 1.22047722e+00 1.06794238e+00 4.55929071e-01
3.79196227e-01 4.89638299e-01 5.19324064e-01 -2.33677160e-02
1.82419866e-01 -2.40334898e-01 6.50949590e-03 9.43470299e-01
-2.40846694e-01 4.86055203e-02 -4.36958522e-01 3.69916081e-01
-1.96292853e+00 -9.42818701e-01 -3.49521846e-01 2.14494491e+00
4.42770958e-01 4.13219512e-01 4.71927375e-01 3.10482055e-01
4.56489533e-01 -1.05881460e-01 -5.65243959e-01 -8.68799567e-01
7.65787512e-02 -3.09700787e-01 7.31570899e-01 5.94752319e-02
-9.22582567e-01 5.15402794e-01 6.45170212e+00 5.55275202e-01
-1.05887806e+00 -4.08036590e-01 1.11716247e+00 1.00504868e-02
-4.14707154e-01 -4.27053049e-02 -7.90483117e-01 5.60927153e-01
1.00119221e+00 -5.19435890e-02 7.82568336e-01 9.62957919e-01
5.37685990e-01 -1.92267090e-01 -1.12055910e+00 6.76384747e-01
-2.39209607e-01 -1.29850805e+00 -5.22159755e-01 3.71265322e-01
7.31478214e-01 -4.92578924e-01 3.09646159e-01 5.52422963e-02
2.38059536e-01 -1.09311903e+00 1.00659537e+00 6.19083345e-01
1.42284930e-01 -1.02634108e+00 1.00271595e+00 2.59687304e-01
-1.05154777e+00 -5.44654608e-01 -2.07533121e-01 -6.32743359e-01
7.35452414e-01 9.89794910e-01 -3.43003035e-01 2.81772524e-01
6.46243751e-01 3.08094084e-01 -7.17196018e-02 7.92729974e-01
4.85530309e-02 2.73348153e-01 -1.82024956e-01 -2.58788943e-01
1.16289660e-01 -5.75739443e-01 5.31904921e-02 7.71110177e-01
3.75506461e-01 -1.21598616e-01 1.20815933e-01 5.70077598e-01
1.87162071e-01 8.20195600e-02 -3.87069315e-01 -3.62726688e-01
2.04440504e-01 1.30941093e+00 -8.98903787e-01 1.97593912e-01
-5.33653378e-01 6.47525609e-01 -3.25405955e-01 7.86868334e-02
-5.63817203e-01 5.54198883e-02 8.82946730e-01 5.82629502e-01
2.53409177e-01 -4.05682325e-01 -4.17291433e-01 -5.95907152e-01
1.24818869e-01 -1.21225023e+00 -7.07482994e-02 -2.93771058e-01
-1.58551800e+00 3.15115452e-01 -3.30339283e-01 -1.04315686e+00
-3.43368292e-01 -6.48001134e-01 -2.08683029e-01 6.53109193e-01
-1.43226385e+00 -1.53880012e+00 1.37838468e-01 -1.08180061e-01
7.62640357e-01 1.46090612e-01 6.32279634e-01 2.87925392e-01
-1.06689416e-01 -1.35485068e-01 3.46840084e-01 -3.25450927e-01
4.24044877e-01 -1.03808391e+00 4.20177519e-01 2.39272326e-01
-8.56539682e-02 5.67082167e-01 1.03947580e+00 -8.06021273e-01
-1.94631743e+00 -7.70622134e-01 1.37540364e+00 -5.18357337e-01
1.09572268e+00 -4.38813090e-01 -5.22467017e-01 6.91135764e-01
1.60812456e-02 -8.09674561e-01 7.75482953e-01 5.57721436e-01
2.58070022e-01 -4.46798593e-01 -1.03490019e+00 3.77671778e-01
6.18428111e-01 -4.84415516e-02 -3.47460419e-01 4.56193089e-01
6.43960416e-01 -8.86336192e-02 -1.33548343e+00 1.33915797e-01
1.16346014e+00 -1.05516887e+00 7.87096441e-01 -5.14420986e-01
5.13849318e-01 1.50421053e-01 -1.67165160e-01 -1.21406615e+00
-5.58224678e-01 -4.92502093e-01 2.35037833e-01 1.41111863e+00
4.33588535e-01 -5.38302422e-01 7.94750690e-01 1.25585020e+00
2.16434881e-01 -1.03761625e+00 -5.04405141e-01 -6.91204429e-01
-1.14984483e-01 -8.28045309e-01 1.03697300e+00 7.56055295e-01
-3.71806234e-01 3.97844553e-01 -7.78483689e-01 -1.34418935e-01
6.52353406e-01 5.58268070e-01 6.42791033e-01 -1.46058083e+00
-3.32065135e-01 -3.51891786e-01 -4.06014442e-01 -6.98975742e-01
-5.09696841e-01 -2.32092887e-01 -3.15916538e-01 -1.60719371e+00
-1.81459755e-01 -7.33333647e-01 -2.58899540e-01 -7.81132281e-02
3.93153280e-01 2.35852867e-01 4.28426921e-01 1.74335897e-01
-1.99058384e-01 -2.12465242e-01 1.46937561e+00 -3.96281183e-02
-3.04240823e-01 6.62082314e-01 -1.06075108e+00 7.46586263e-01
5.49640656e-01 -5.58479130e-02 -2.87413150e-01 -3.13476562e-01
8.74848485e-01 5.52146882e-03 9.18943584e-02 -5.31446695e-01
1.07527643e-01 -5.91467917e-01 6.95022523e-01 -5.76366067e-01
3.36408854e-01 -1.44386852e+00 1.02428293e+00 4.81725335e-01
-1.80939123e-01 1.89262420e-01 -1.15797631e-01 5.36599696e-01
-7.50184506e-02 -1.40026480e-01 5.03442466e-01 -2.74915308e-01
-7.68831193e-01 -2.25140341e-02 -4.47677970e-01 -6.15382433e-01
1.28091025e+00 -1.17477618e-01 8.34937692e-02 -5.98950088e-01
-1.06599712e+00 3.26237559e-01 3.61427516e-01 7.55452573e-01
1.09095752e-01 -1.25662470e+00 -4.60849136e-01 3.83021295e-01
-2.15791658e-01 -3.40880781e-01 4.56550449e-01 6.77078962e-01
-4.27421153e-01 3.88325334e-01 -3.44775468e-01 -1.33432552e-01
-9.29664314e-01 7.78850555e-01 -2.25301728e-01 -4.31286931e-01
-2.38727435e-01 6.63286746e-01 -1.55046850e-01 -1.45010017e-02
1.15693979e-01 -4.10815179e-01 -2.24522874e-01 4.26751226e-01
4.86368597e-01 5.40495932e-01 1.94226131e-01 -7.25483418e-01
-1.40186459e-01 5.17021716e-01 -4.93475646e-02 4.05046791e-01
1.77927792e+00 -3.04692090e-01 -1.66029930e-01 6.22954249e-01
6.88284695e-01 -1.23391643e-01 -1.18596566e+00 1.46402359e-01
2.36516714e-01 -7.22444057e-01 1.23638280e-01 -1.12665391e+00
-1.16195691e+00 6.11428916e-01 4.00629342e-01 7.88681448e-01
1.18547761e+00 5.99547252e-02 1.12941945e+00 -2.22226486e-01
6.47966146e-01 -1.26394522e+00 -3.39659333e-01 -1.08719446e-01
8.26005578e-01 -1.23349500e+00 9.76817831e-02 -4.19109851e-01
-8.46950710e-01 1.05512416e+00 -4.76797447e-02 6.21553473e-02
7.65345931e-01 2.53049672e-01 1.25109121e-01 -2.32318975e-02
-6.35459185e-01 1.08452857e-01 1.49077624e-01 3.80524725e-01
4.98611003e-01 4.13395196e-01 -5.23794472e-01 9.43444729e-01
-2.85448611e-01 4.24067862e-02 7.05829710e-02 8.36246669e-01
-4.00657386e-01 -1.68571579e+00 -3.91218424e-01 7.92594969e-01
-5.81932843e-01 2.88932830e-01 -1.70538291e-01 7.82300353e-01
3.74012679e-01 1.27790940e+00 5.17823212e-02 -5.29508650e-01
7.33573675e-01 -1.37862757e-01 3.22236091e-01 -2.25966290e-01
-7.52521753e-01 5.39022267e-01 5.11513233e-01 -3.42558295e-01
-6.98304698e-02 -8.72089982e-01 -7.93543458e-01 -9.17890131e-01
-3.06594044e-01 7.86864087e-02 1.26170683e+00 9.95518625e-01
1.03509068e-01 5.97929895e-01 8.16711485e-01 -7.60142803e-01
-8.12897205e-01 -5.61790943e-01 -9.91544485e-01 4.54459190e-01
-6.63885176e-02 -5.74791372e-01 -1.94461554e-01 1.40711501e-01] | [9.281397819519043, 5.8457560539245605] |
ce43e06f-740a-4b14-8be4-59a520072a72 | adapting-unsupervised-syntactic-parsing | null | null | https://aclanthology.org/2021.acl-long.449 | https://aclanthology.org/2021.acl-long.449.pdf | Adapting Unsupervised Syntactic Parsing Methodology for Discourse Dependency Parsing | One of the main bottlenecks in developing discourse dependency parsers is the lack of annotated training data. A potential solution is to utilize abundant unlabeled data by using unsupervised techniques, but there is so far little research in unsupervised discourse dependency parsing. Fortunately, unsupervised syntactic dependency parsing has been studied by decades, which could potentially be adapted for discourse parsing. In this paper, we propose a simple yet effective method to adapt unsupervised syntactic dependency parsing methodology for unsupervised discourse dependency parsing. We apply the method to adapt two state-of-the-art unsupervised syntactic dependency parsing methods. Experimental results demonstrate that our adaptation is effective. Moreover, we extend the adapted methods to the semi-supervised and supervised setting and surprisingly, we find that they outperform previous methods specially designed for supervised discourse parsing. Further analysis shows our adaptations result in superiority not only in parsing accuracy but also in time and space efficiency. | ['Kewei Tu', 'Wenjuan Han', 'Ge Wang', 'Liwen Zhang'] | 2021-08-01 | null | null | null | acl-2021-5 | ['discourse-parsing'] | ['natural-language-processing'] | [ 2.93813556e-01 7.09161162e-01 -4.07932371e-01 -4.41515177e-01
-9.11957979e-01 -7.39649415e-01 4.42446738e-01 3.89906526e-01
-4.40825105e-01 1.00008321e+00 6.30706012e-01 -5.08061886e-01
2.26720765e-01 -6.22847795e-01 -4.08761144e-01 -5.48628449e-01
8.81203562e-02 4.20814902e-01 5.16653538e-01 -1.93594635e-01
2.88998455e-01 4.60564867e-02 -1.30692744e+00 1.47915035e-01
1.00637627e+00 8.95715132e-02 3.50125283e-01 7.08126366e-01
-4.51061606e-01 1.27118218e+00 -7.88459063e-01 -4.05276090e-01
-3.71125311e-01 -8.69273126e-01 -1.36568558e+00 7.59689063e-02
-3.00842971e-01 -1.55754238e-01 2.47674942e-01 8.79749477e-01
2.77333438e-01 1.01465352e-01 3.33329916e-01 -5.17986834e-01
-4.71436203e-01 1.14321733e+00 -1.56920001e-01 3.34233880e-01
4.28878397e-01 -5.25118172e-01 9.58853543e-01 -4.17880148e-01
1.01625633e+00 1.06513393e+00 3.11416358e-01 8.37704599e-01
-8.99395168e-01 2.28069294e-02 4.30774122e-01 1.19114585e-01
-4.97482508e-01 -3.93648505e-01 8.70212734e-01 -2.32310772e-01
1.08224440e+00 2.26578534e-01 2.33238548e-01 1.04040706e+00
-3.35932732e-01 7.02360511e-01 1.24933994e+00 -1.04820645e+00
4.25986975e-01 2.02395186e-01 5.41272581e-01 6.70519769e-01
8.63509476e-02 -3.96730632e-01 -4.14591134e-01 1.78743184e-01
4.79855627e-01 -6.21371925e-01 1.72777195e-02 -3.11668888e-02
-9.81479466e-01 9.36986029e-01 -2.21811552e-02 7.78702915e-01
1.31349385e-01 -2.36804098e-01 7.19012499e-01 3.34108293e-01
8.51236403e-01 4.31836486e-01 -5.33053815e-01 -3.61977845e-01
-5.51131010e-01 -1.01546675e-01 1.29334116e+00 9.17432010e-01
4.57831025e-01 -2.18714193e-01 9.18549970e-02 7.25659907e-01
2.86640286e-01 1.97872251e-01 5.01521289e-01 -1.01234829e+00
7.23058939e-01 7.18576014e-01 -1.34108827e-01 -5.58061242e-01
-6.21990323e-01 3.19980532e-01 -4.33134854e-01 -5.64078726e-02
5.74598849e-01 -4.39264745e-01 -5.82953691e-01 1.60605669e+00
4.96887892e-01 -4.72533822e-01 4.23849791e-01 7.47039735e-01
7.88412631e-01 4.33264077e-01 5.27106106e-01 -7.68553555e-01
1.44448483e+00 -1.17529547e+00 -8.97994459e-01 -1.56710774e-01
1.27536488e+00 -7.62011588e-01 8.51160228e-01 3.63498777e-02
-1.15669358e+00 -3.02777708e-01 -9.22512054e-01 -2.22387537e-01
-3.35522771e-01 1.21007532e-01 9.51738536e-01 1.08197308e+00
-7.44409561e-01 4.47020262e-01 -1.08420002e+00 -6.23052180e-01
1.57160029e-01 4.51446921e-01 -4.07869160e-01 8.41428116e-02
-1.06276727e+00 9.63835895e-01 6.97200835e-01 2.83080768e-02
-1.88203990e-01 5.82970344e-02 -9.84081984e-01 -1.12964986e-02
6.91850841e-01 -3.71920228e-01 1.52247810e+00 -1.23339391e+00
-2.03580785e+00 7.84323454e-01 -4.77675796e-01 -3.71559411e-01
2.69049913e-01 -4.04269814e-01 -6.93486407e-02 2.76177496e-01
2.30524719e-01 1.82539746e-01 3.64737034e-01 -1.15207601e+00
-6.89425886e-01 -3.69720191e-01 3.83587837e-01 3.50624055e-01
-5.03802657e-01 5.92633307e-01 -4.30212140e-01 -4.40058231e-01
1.75950438e-01 -7.91514754e-01 -4.31380481e-01 -8.11315477e-01
-4.20017987e-01 -5.86044252e-01 7.71334291e-01 -7.82852888e-01
1.42811644e+00 -1.92386067e+00 3.59195739e-01 -1.00815184e-01
-1.61388621e-01 3.54790419e-01 2.18862042e-01 6.09118640e-01
1.26217455e-02 4.55847472e-01 -7.07960844e-01 -5.87866247e-01
-1.59937695e-01 7.58651733e-01 -3.25707085e-02 1.86721146e-01
5.09243131e-01 8.17350030e-01 -1.12591445e+00 -8.64013493e-01
1.27267733e-01 9.23963562e-02 -4.73285496e-01 3.70633990e-01
-4.99727994e-01 1.02793312e+00 -8.90460610e-01 6.25090837e-01
3.53652418e-01 -3.54393646e-02 1.02797925e+00 4.10192400e-01
-4.38267618e-01 7.61539817e-01 -6.31721437e-01 1.83203232e+00
-3.51225197e-01 6.17395699e-01 -7.69965425e-02 -1.48337245e+00
7.62998641e-01 4.06456172e-01 1.81345955e-01 -5.97905576e-01
2.31599614e-01 2.40032926e-01 -7.30291605e-02 -9.91331160e-01
6.20136023e-01 -1.51655480e-01 -4.29443449e-01 4.04980153e-01
1.36801466e-01 2.37473875e-01 6.74084127e-01 9.28859711e-02
1.12440979e+00 5.13910651e-01 4.38196093e-01 -3.29332560e-01
8.14532638e-01 4.48204398e-01 4.51412469e-01 4.34225768e-01
-1.09358795e-01 4.93995309e-01 8.93540442e-01 -7.07119927e-02
-8.21699202e-01 -5.70241690e-01 7.26996502e-03 1.53476858e+00
-1.42378315e-01 -5.45986950e-01 -1.24190950e+00 -1.09404290e+00
-6.64485335e-01 6.03453577e-01 -3.98806751e-01 6.09855473e-01
-1.50149262e+00 -8.20450723e-01 4.72909063e-01 7.60412633e-01
3.72248083e-01 -1.40582454e+00 -7.76376724e-01 3.07404041e-01
-1.60263345e-01 -1.41091549e+00 3.47800404e-01 5.04888415e-01
-1.02020538e+00 -1.17521453e+00 -3.64267349e-01 -1.25746214e+00
7.65327692e-01 3.34504023e-02 1.08814538e+00 4.01086599e-01
2.61136919e-01 2.05604509e-01 -1.04445910e+00 -3.38408947e-01
-9.52309668e-01 6.45312369e-01 -2.70196140e-01 -6.58104956e-01
2.98834264e-01 -4.30475742e-01 -1.61978349e-01 -7.48874471e-02
-8.08497488e-01 -2.11554483e-01 4.11752790e-01 7.65291214e-01
3.24302882e-01 -3.77827644e-01 7.71208525e-01 -1.77650273e+00
6.59772694e-01 -3.59613508e-01 -4.55773175e-01 2.56040066e-01
-4.50754255e-01 1.80993110e-01 9.06036913e-01 -3.98569554e-02
-1.72923803e+00 1.81619465e-01 -5.32494724e-01 4.71791774e-01
-4.34464633e-01 8.14466298e-01 -2.56590009e-01 4.48678106e-01
4.69911486e-01 -4.34266359e-01 -3.78667563e-01 -6.22866869e-01
5.26559174e-01 5.48271239e-01 5.13537347e-01 -8.03861558e-01
5.53219676e-01 3.57472688e-01 -7.16385171e-02 -9.85672235e-01
-8.39137673e-01 -4.91051078e-01 -1.13426554e+00 1.68440104e-01
1.28344870e+00 -6.43310308e-01 -2.68082887e-01 1.22981139e-01
-1.31645501e+00 -4.00816917e-01 -2.56341666e-01 4.05074030e-01
-4.74695206e-01 8.54176879e-01 -7.83327103e-01 -9.21686172e-01
-1.47792682e-01 -8.33866477e-01 8.19137216e-01 4.19387251e-01
-3.04761916e-01 -1.35102808e+00 4.45152670e-01 4.10984039e-01
4.97605912e-02 2.63115883e-01 1.08270884e+00 -1.03509712e+00
-3.15083474e-01 2.27872133e-02 7.55510945e-03 3.20236892e-01
2.05886543e-01 1.36340968e-02 -9.56651628e-01 1.20995305e-01
2.33996615e-01 -3.35504740e-01 7.40282595e-01 2.72594422e-01
6.73531115e-01 -1.17622741e-01 -3.07347685e-01 1.96747646e-01
1.07370687e+00 3.56650315e-02 5.97149193e-01 5.71654916e-01
7.62803137e-01 1.06410849e+00 8.35741937e-01 1.84995737e-02
4.78218347e-01 3.77184123e-01 9.37182829e-02 1.94021892e-02
-4.52424251e-02 -7.88072050e-02 4.22958016e-01 1.46713400e+00
-5.25035262e-01 -1.90312639e-01 -8.94282758e-01 7.43727446e-01
-2.19283962e+00 -7.56825507e-01 -4.41324115e-01 1.83496213e+00
8.38672996e-01 2.99076259e-01 1.44426554e-01 2.51456529e-01
6.59358859e-01 7.97636434e-02 -2.32257769e-02 -7.18340755e-01
-1.18588947e-01 6.16474390e-01 2.02111438e-01 4.76666480e-01
-1.38314855e+00 1.22547352e+00 6.38145018e+00 2.06013113e-01
-6.96407557e-01 3.44477534e-01 4.07810509e-01 3.68885309e-01
-1.91378415e-01 5.12639523e-01 -6.60313368e-01 5.41029088e-02
1.31811571e+00 1.32772386e-01 -2.18222275e-01 9.55426812e-01
-2.39890814e-03 -3.90523463e-01 -9.92810488e-01 1.38520777e-01
-9.86585300e-03 -1.20932317e+00 -4.04437244e-01 -3.11058640e-01
5.65811396e-01 -3.04940432e-01 -5.66563547e-01 4.40024495e-01
2.67108053e-01 -8.43357265e-01 2.49098390e-01 -2.07487151e-01
3.37238818e-01 -6.54792547e-01 8.56711388e-01 4.73559469e-01
-1.11776781e+00 3.19421887e-02 -4.06723619e-01 -4.63745564e-01
3.73132795e-01 3.16524893e-01 -6.09311044e-01 8.62072527e-01
5.36162794e-01 4.28176373e-01 -4.07663852e-01 3.58645409e-01
-9.44896877e-01 9.85424459e-01 -2.45136127e-01 -2.73024589e-01
1.89684391e-01 -3.04506451e-01 4.18159604e-01 1.55812120e+00
-1.45029709e-01 2.63786137e-01 2.55759031e-01 2.12515444e-01
1.50556806e-02 5.04235625e-01 -5.03515482e-01 -9.56095457e-02
4.09122199e-01 1.06275833e+00 -1.16243029e+00 -4.30139035e-01
-6.95839524e-01 8.74507368e-01 7.04494655e-01 9.49751884e-02
-6.77872717e-01 -2.54860744e-02 -1.21278718e-01 -3.12650293e-01
5.54075003e-01 -4.41444010e-01 -3.56165439e-01 -1.12856281e+00
2.93273002e-01 -6.43674076e-01 5.34207702e-01 -7.28169233e-02
-1.12406337e+00 7.24057913e-01 3.15887183e-01 -7.98663080e-01
-6.38670027e-01 -7.53972113e-01 -6.12185836e-01 5.57051122e-01
-1.74265492e+00 -1.15882826e+00 -2.37064576e-03 2.00118884e-01
9.04757440e-01 2.71500293e-02 1.30181396e+00 -6.42225817e-02
-8.17023277e-01 3.02451938e-01 1.23865493e-02 4.15442020e-01
6.66588187e-01 -1.39716864e+00 4.19423282e-01 1.06414592e+00
1.91424653e-01 5.51587343e-01 6.19822443e-01 -5.15925646e-01
-1.32846594e+00 -7.10794628e-01 1.27919865e+00 -5.64019442e-01
8.48917186e-01 -3.37889403e-01 -1.11669648e+00 6.60275638e-01
5.97784758e-01 -3.30473632e-01 9.41327512e-01 4.70628560e-01
-2.43373856e-01 5.62931836e-01 -7.60304689e-01 5.01686275e-01
1.03850091e+00 -3.91695857e-01 -1.08350730e+00 3.14010382e-01
8.04329813e-01 -6.61638796e-01 -9.92911100e-01 1.97990611e-01
3.92970651e-01 -1.15667915e+00 4.67335880e-01 -7.05115557e-01
7.47930348e-01 4.51159216e-02 3.60724665e-02 -8.68238211e-01
1.33228257e-01 -7.36253560e-01 -9.10809934e-02 1.66877127e+00
6.62593305e-01 -5.84213018e-01 9.03665364e-01 7.91721702e-01
-3.51287395e-01 -4.56146598e-01 -9.56163585e-01 -6.36815131e-01
4.71514881e-01 -2.18071938e-01 1.03098653e-01 1.00705957e+00
7.66829550e-01 8.36522162e-01 4.38195914e-02 -4.77733323e-03
2.71942914e-01 1.88140213e-01 7.55052984e-01 -1.11285484e+00
-4.13558781e-01 -2.15583235e-01 1.28956214e-01 -9.93997574e-01
5.89658558e-01 -5.98581254e-01 2.78683156e-01 -1.61338603e+00
1.55024111e-01 -4.19315904e-01 1.59905031e-02 3.22904795e-01
-5.22122145e-01 -2.29348987e-01 2.18204170e-01 3.35492402e-01
-5.86362004e-01 2.06131548e-01 8.65200758e-01 1.93874061e-01
-6.59242034e-01 -1.37485012e-01 -6.00339293e-01 9.87786710e-01
1.16349220e+00 -7.68019319e-01 -3.34279329e-01 -4.25327986e-01
1.05987992e-02 9.28324386e-02 -2.74503350e-01 -6.69254839e-01
1.31726444e-01 -5.46908937e-02 -1.53570697e-01 -3.04684192e-01
-1.56502604e-01 -4.97979939e-01 -4.56426263e-01 8.92552584e-02
-2.34143957e-01 1.28078731e-02 7.97340348e-02 4.29322898e-01
-3.91194940e-01 -8.45041990e-01 3.56808782e-01 -3.43044847e-01
-8.54199469e-01 -5.56205988e-01 -6.75288856e-01 1.45404369e-01
9.34637785e-01 8.37090518e-03 -6.23247385e-01 3.30279209e-02
-6.42764330e-01 1.42239198e-01 3.87840480e-01 3.04697424e-01
1.19508699e-01 -5.72463810e-01 -6.35024488e-01 -2.10541978e-01
-1.25615880e-01 4.05828357e-01 -1.43558502e-01 7.28934824e-01
-5.68115652e-01 5.63240647e-01 -5.38201444e-03 -5.30379653e-01
-1.41779637e+00 5.91545165e-01 -4.68631089e-01 -9.07870293e-01
-6.42833054e-01 5.23813426e-01 -1.15066789e-01 -3.16233367e-01
1.33384839e-01 -4.09626484e-01 -5.43416917e-01 6.53270930e-02
3.30234796e-01 1.18206762e-01 -4.91699111e-03 -6.40902877e-01
-4.11605060e-01 3.97835642e-01 -9.08651054e-02 -1.08611800e-01
1.38653493e+00 -2.21173063e-01 -3.30218017e-01 4.76594567e-01
8.35505724e-01 3.18402559e-01 -8.59636903e-01 2.00510934e-01
7.39522874e-01 8.78957212e-02 -4.85998511e-01 -4.11288559e-01
-4.70334917e-01 6.53014421e-01 -6.91727325e-02 7.55344570e-01
1.18986475e+00 3.51729274e-01 4.73847032e-01 7.26837993e-01
2.78147727e-01 -1.23682702e+00 -1.10222787e-01 8.29106152e-01
2.75083274e-01 -1.39520729e+00 8.58545974e-02 -9.73166704e-01
-7.28667498e-01 1.18662000e+00 4.56642628e-01 -2.84862965e-01
3.81898135e-01 4.33371902e-01 2.91378587e-01 -1.74852923e-01
-4.83013630e-01 -5.59292853e-01 -2.62106091e-01 8.04513216e-01
1.03593564e+00 -7.18788430e-02 -9.41986442e-01 7.32585728e-01
-4.32683378e-02 -1.88896701e-01 5.73838949e-01 1.47816503e+00
-3.95056278e-01 -1.91782951e+00 -2.56042540e-01 -2.38145679e-01
-7.56321311e-01 -1.00078411e-01 -6.08399928e-01 1.27933705e+00
-2.99628586e-01 1.27976620e+00 -1.46119595e-01 1.42684177e-01
3.73914719e-01 4.23332393e-01 7.06136525e-01 -9.29770410e-01
-9.28544819e-01 2.25142583e-01 7.92956114e-01 -1.85762942e-01
-1.34228659e+00 -6.56020105e-01 -1.72422302e+00 1.62958279e-01
-4.71825004e-01 3.07735413e-01 5.88327229e-01 1.35625792e+00
2.18425412e-03 5.38927674e-01 5.20959437e-01 -7.38583505e-01
-3.17577511e-01 -1.08142805e+00 -1.13862060e-01 4.04433846e-01
-1.35705499e-02 -4.01944429e-01 -1.50934681e-01 5.22243381e-01] | [10.638288497924805, 9.54377269744873] |
0bc5251b-fb71-470c-8166-669d37ad4b2b | behavioral-analysis-of-vision-and-language | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Yang_Behavioral_Analysis_of_Vision-and-Language_Navigation_Agents_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yang_Behavioral_Analysis_of_Vision-and-Language_Navigation_Agents_CVPR_2023_paper.pdf | Behavioral Analysis of Vision-and-Language Navigation Agents | To be successful, Vision-and-Language Navigation (VLN) agents must be able to ground instructions to actions based on their surroundings. In this work, we develop a methodology to study agent behavior on a skill-specific basis -- examining how well existing agents ground instructions about stopping, turning, and moving towards specified objects or rooms. Our approach is based on generating skill-specific interventions and measuring changes in agent predictions. We present a detailed case study analyzing the behavior of a recent agent and then compare multiple agents in terms of skill-specific competency scores. This analysis suggests that biases from training have lasting effects on agent behavior and that existing models are able to ground simple referring expressions. Our comparisons between models show that skill-specific scores correlate with improvements in overall VLN task performance. | ['Stefan Lee', 'Arjun Majumdar', 'Zijiao Yang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['vision-and-language-navigation'] | ['robots'] | [-1.23131806e-02 1.91119537e-01 3.28563005e-02 -2.58909255e-01
-2.97212124e-01 -5.97541034e-01 1.08410370e+00 1.92656681e-01
-7.63759792e-01 9.28901076e-01 5.43694377e-01 -4.77159798e-01
-2.00507879e-01 -7.45500982e-01 -6.77945793e-01 -2.66214609e-01
-1.87682703e-01 8.39790881e-01 3.68371516e-01 -7.59202838e-01
6.86915815e-01 7.11069167e-01 -1.52959883e+00 7.12390170e-02
8.70014966e-01 -1.72572181e-01 6.33697033e-01 8.44757557e-01
1.19626053e-01 1.45235491e+00 -4.74944770e-01 9.25366674e-03
5.54420911e-02 -6.75043523e-01 -9.09329772e-01 4.51714126e-03
1.97107673e-01 -4.67773855e-01 -1.60508335e-01 9.17025030e-01
1.26738802e-01 5.41998208e-01 5.92169225e-01 -1.34340978e+00
-6.59553468e-01 4.84160572e-01 3.55626673e-01 1.97482437e-01
8.94575417e-01 8.55619967e-01 5.87360144e-01 -2.45483831e-01
8.04234803e-01 1.32542503e+00 6.72123134e-01 8.39783549e-01
-1.29203594e+00 -2.50711113e-01 5.59384942e-01 2.68463939e-01
-8.11758935e-01 -4.13548023e-01 3.90021950e-01 -7.15622783e-01
1.54644144e+00 -1.38396740e-01 8.04626167e-01 1.15944874e+00
2.52039731e-01 4.82549310e-01 1.34616578e+00 -4.95240122e-01
3.70852977e-01 -5.38446940e-03 2.74923444e-01 1.00528538e+00
3.79805654e-01 8.35148573e-01 -6.71143472e-01 2.17626363e-01
1.00629866e+00 -2.86477834e-01 -2.18155682e-01 -7.69840062e-01
-1.37710273e+00 7.20878184e-01 4.89284635e-01 4.12423044e-01
-8.80164266e-01 4.05540854e-01 1.95827529e-01 2.85628498e-01
-2.39600182e-01 1.06782043e+00 -3.05783540e-01 -4.08614069e-01
-2.07952783e-01 5.33067644e-01 6.93120301e-01 5.75324893e-01
8.25287223e-01 3.58446807e-01 -3.38193387e-01 3.97174805e-01
3.54669958e-01 6.51672304e-01 5.91583252e-01 -1.52364933e+00
1.61542207e-01 6.00647151e-01 5.44795334e-01 -9.17735577e-01
-9.24931765e-01 -1.48509040e-01 1.83631182e-01 1.00859332e+00
3.66833568e-01 -1.56403571e-01 -1.00317156e+00 2.15097904e+00
-2.31127560e-01 5.51125854e-02 5.82242012e-01 7.44268596e-01
6.42735422e-01 3.47495049e-01 6.68395638e-01 -1.04877762e-01
8.87341797e-01 -1.27499545e+00 -5.21161377e-01 -9.47922826e-01
1.30173194e+00 -7.53344595e-02 1.34903717e+00 1.10752836e-01
-1.07908106e+00 -7.12126136e-01 -9.15895641e-01 2.63234347e-01
-5.61027646e-01 -5.09100258e-01 7.59953260e-01 6.39903307e-01
-1.55577826e+00 4.97165710e-01 -8.54672015e-01 -9.86305654e-01
-1.26511995e-02 3.13540041e-01 -3.85068029e-01 1.75525099e-01
-7.73499012e-01 1.82630360e+00 4.53405946e-01 -3.02141160e-01
-1.44768870e+00 -1.20042697e-01 -1.25119615e+00 -2.22905740e-01
1.79766595e-01 -8.98170054e-01 1.62208331e+00 -7.49391735e-01
-1.55873597e+00 9.08150911e-01 -2.52243370e-01 -6.42318428e-01
2.36475319e-01 8.31141695e-02 -1.94587529e-01 -2.40985408e-01
4.52809513e-01 7.00323164e-01 1.86245084e-01 -1.79869032e+00
-1.10926962e+00 -3.83763075e-01 6.45927191e-01 6.34678781e-01
3.04046571e-01 -2.63642371e-01 -5.65252006e-02 4.34041768e-02
-1.11763179e-02 -1.09588325e+00 -4.09702241e-01 -7.25577295e-01
1.73494607e-01 -3.05253029e-01 2.66643077e-01 -4.65608597e-01
8.23849559e-01 -1.66010594e+00 2.15538293e-01 1.00124046e-01
6.19766861e-02 7.87065327e-02 -4.29995805e-01 3.73231143e-01
2.47927785e-01 -9.58392695e-02 1.40677750e-01 -1.29353359e-01
1.37103915e-01 5.20561814e-01 -7.21864775e-02 1.58511013e-01
-1.57019764e-01 1.07039607e+00 -1.26008928e+00 -2.37683803e-01
6.35510325e-01 1.04200818e-01 -5.67562819e-01 1.59402788e-01
-2.45313242e-01 7.22131431e-01 -3.28620166e-01 2.68814147e-01
-1.14705883e-01 9.49677154e-02 3.00273299e-01 4.28069681e-01
-4.35886264e-01 4.42969114e-01 -5.10385454e-01 1.58110166e+00
-7.47182131e-01 8.04867506e-01 -1.73669070e-01 -6.15532458e-01
5.67189991e-01 1.86205357e-01 -8.12812299e-02 -1.17848420e+00
-5.87268919e-02 1.25122234e-01 5.05111873e-01 -7.37794399e-01
3.78481030e-01 -5.55734336e-02 -4.22254615e-02 3.21499139e-01
-2.50933677e-01 -5.08243084e-01 4.90376323e-01 -3.65026146e-02
1.21542346e+00 3.39746863e-01 4.56474304e-01 -1.87622160e-01
5.60607135e-01 7.55457580e-01 1.12690046e-01 1.28365266e+00
-4.93923634e-01 -9.87902582e-02 2.08142340e-01 -4.87831831e-01
-8.76140475e-01 -1.05950868e+00 6.81081712e-01 1.46250701e+00
1.96443915e-01 -4.44116518e-02 -8.20880413e-01 -5.15221059e-01
-2.77850538e-01 1.66204643e+00 -9.19131577e-01 -4.00369227e-01
-8.55389476e-01 -2.54621118e-01 2.85365492e-01 6.13897562e-01
4.83674645e-01 -1.86823857e+00 -1.36552215e+00 1.16138473e-01
9.09763277e-02 -8.28239799e-01 1.16447255e-01 1.30713955e-01
-8.54868174e-01 -9.68638361e-01 -1.84762761e-01 -1.05029464e+00
8.15719128e-01 3.17259699e-01 1.34340024e+00 3.93601865e-01
4.48120028e-01 1.01080215e+00 -3.58651370e-01 -5.60010254e-01
-7.48275936e-01 -5.53999245e-02 3.09577346e-01 -9.47943270e-01
4.71326441e-01 -2.99791753e-01 -2.98588544e-01 1.75333634e-01
-2.30501831e-01 7.52320215e-02 6.34510398e-01 5.47248006e-01
-1.65701360e-01 -2.60811388e-01 1.41496256e-01 -7.02795804e-01
1.31743062e+00 -6.10139407e-02 -6.41715944e-01 2.79637486e-01
-1.01909888e+00 1.27938122e-01 1.81922868e-01 -1.59540907e-01
-1.16246998e+00 -1.32283017e-01 1.52986676e-01 3.56608838e-01
-5.60696900e-01 4.03233230e-01 1.83957323e-01 -7.02413246e-02
9.87128496e-01 3.70637625e-01 1.05373487e-01 3.27649206e-01
3.79386663e-01 -9.53279361e-02 5.20518482e-01 -7.49218822e-01
8.15590322e-01 1.97041586e-01 -1.73231229e-01 -6.29794896e-01
-4.70349759e-01 -9.39168334e-02 -3.19602251e-01 -5.79741180e-01
9.71544206e-01 -6.94623172e-01 -1.11908579e+00 3.18480790e-01
-9.58769977e-01 -1.39751339e+00 -2.58452356e-01 6.43078804e-01
-1.09184444e+00 -1.05436146e-01 -3.21496338e-01 -7.24987507e-01
3.80161554e-01 -1.45189106e+00 6.13555074e-01 4.41649139e-01
-7.93718815e-01 -1.24832475e+00 5.63931942e-01 2.12736100e-01
5.42450249e-01 -2.50782251e-01 8.84674251e-01 -7.30481207e-01
-5.25895000e-01 1.38927281e-01 -2.76142880e-02 -2.81607836e-01
6.16008900e-02 -5.92782557e-01 -4.40664232e-01 -2.01883301e-01
-2.09057443e-02 -4.36424494e-01 4.02746439e-01 5.82808614e-01
2.95585006e-01 -1.22640654e-02 -4.61425960e-01 2.45995775e-01
1.31609738e+00 8.57554674e-01 6.08999014e-01 1.19828308e+00
3.90515417e-01 7.18024194e-01 7.34201193e-01 -6.52740449e-02
6.86414897e-01 6.22900426e-01 5.33896208e-01 1.41663387e-01
-1.10390060e-01 -3.12535495e-01 6.39207184e-01 1.30066931e-01
-4.70406771e-01 -3.22337389e-01 -1.20199394e+00 6.12556338e-01
-2.02768564e+00 -1.27328205e+00 -1.20171241e-01 1.89441538e+00
3.38092625e-01 4.01282430e-01 1.14714541e-02 -4.58362371e-01
2.66894668e-01 1.55643180e-01 -6.15062118e-01 -5.86899340e-01
7.89963454e-02 -2.00996578e-01 4.87730026e-01 1.21255338e+00
-5.80977321e-01 1.45011890e+00 7.69848633e+00 -1.19129255e-01
-5.38439989e-01 6.19615056e-02 5.04173189e-02 1.23926438e-01
-1.14039980e-01 -2.67048422e-02 -5.43405235e-01 2.45318413e-02
9.65289354e-01 -2.25823581e-01 7.22328484e-01 7.54107893e-01
5.71563065e-01 -5.79818964e-01 -1.25290406e+00 5.21387458e-01
1.82456806e-01 -1.16927230e+00 7.21146241e-02 -9.24702883e-02
5.62907457e-01 -1.64145846e-02 2.75426358e-01 9.38366294e-01
1.27135873e+00 -1.19119036e+00 7.33163416e-01 7.13824689e-01
-6.02090284e-02 -3.17200154e-01 7.76872337e-01 4.96892929e-01
-5.14177382e-01 -1.47269055e-01 7.14582801e-02 -7.55690396e-01
1.62897632e-01 -7.41506398e-01 -1.34462416e+00 -4.85713519e-02
5.36860526e-01 2.43522450e-01 -4.98817384e-01 6.75906420e-01
-5.15532374e-01 2.28104666e-01 1.60118341e-01 -4.37566310e-01
5.85569143e-01 -3.35352629e-01 3.19624454e-01 9.52999115e-01
1.11116134e-01 3.11226249e-01 3.64141345e-01 5.42505801e-01
6.98647141e-01 -1.49328128e-01 -9.47820127e-01 2.87567705e-01
4.17207628e-01 5.57238102e-01 -7.88921237e-01 -4.04167712e-01
-3.91286701e-01 9.88723576e-01 4.90525335e-01 6.22569621e-01
-7.25347102e-01 2.63912648e-01 9.92163122e-01 -1.11079738e-01
-3.70685086e-02 -5.68606377e-01 -4.78694886e-01 -4.33189154e-01
-4.38666672e-01 -9.51845109e-01 -1.15033753e-01 -1.26302481e+00
-6.63455606e-01 6.00031257e-01 3.36656027e-04 -7.87590683e-01
-6.78471565e-01 -8.42954814e-01 -4.13252920e-01 7.90936530e-01
-1.29783344e+00 -9.95616794e-01 -6.45194054e-01 4.36660677e-01
7.89894462e-01 -5.01684725e-01 1.05196226e+00 -5.32332599e-01
-4.24392611e-01 6.78758323e-02 -3.52278650e-01 8.83322507e-02
3.57198834e-01 -1.12766016e+00 3.00008625e-01 8.16128492e-01
-5.78363724e-02 1.01181793e+00 1.30599749e+00 -8.48449349e-01
-8.46943974e-01 -4.96161789e-01 4.95270073e-01 -1.02735651e+00
4.97267187e-01 2.97445983e-01 -5.13779521e-01 1.28775251e+00
5.24131835e-01 -7.57458508e-01 4.37612295e-01 2.81643182e-01
-2.76994351e-02 3.63044560e-01 -1.01246119e+00 1.22118568e+00
1.62117672e+00 -6.13268077e-01 -1.21382213e+00 3.91829163e-01
4.22208101e-01 -3.67024809e-01 -8.34878311e-02 1.55759662e-01
2.61174053e-01 -1.38991809e+00 1.03874195e+00 -1.04316974e+00
2.74987638e-01 -4.99158323e-01 -2.13836715e-01 -1.72181928e+00
-8.11218262e-01 -2.17311651e-01 4.09251571e-01 4.80127096e-01
6.05006099e-01 -7.92099774e-01 8.41490448e-01 7.67344415e-01
-3.28278959e-01 4.54802392e-03 -5.55418789e-01 -7.50297427e-01
-1.51273832e-01 -4.59970444e-01 4.48355079e-01 7.66099632e-01
2.20875889e-01 3.94321740e-01 7.64363036e-02 5.58679044e-01
3.39046359e-01 -2.33863086e-01 9.89945531e-01 -1.22082436e+00
6.60191923e-02 -8.54839027e-01 -3.89538109e-01 -8.80325437e-01
5.42154729e-01 -6.83094144e-01 3.83661747e-01 -2.29429340e+00
4.34857793e-02 -1.98479086e-01 -7.61724869e-03 4.81827080e-01
5.84289990e-02 -2.85473168e-01 3.67038190e-01 1.11268602e-01
-7.34653473e-01 1.57188654e-01 1.01808584e+00 1.09744415e-01
-4.52605098e-01 -1.41979620e-01 -5.24925351e-01 1.06184161e+00
1.07324409e+00 -1.06712542e-01 -7.65426457e-01 -6.38293207e-01
2.21330434e-01 -8.31353441e-02 6.23223543e-01 -1.40764630e+00
4.50398535e-01 -5.05703151e-01 2.10682005e-01 -2.86297232e-01
4.80936259e-01 -8.74725699e-01 8.09227899e-02 1.09837806e+00
-3.75046581e-01 4.62716341e-01 4.02651906e-01 3.36708546e-01
1.18551426e-01 -4.82458055e-01 3.33134115e-01 -6.31409466e-01
-1.64264679e+00 -4.12590086e-01 -1.08942413e+00 -1.24986231e-01
1.29160452e+00 -5.80754220e-01 -5.90430796e-01 -7.62967408e-01
-9.52663958e-01 4.71404076e-01 8.25493872e-01 3.83811265e-01
5.84385514e-01 -1.08638453e+00 -5.67932487e-01 6.67019468e-03
1.94391847e-01 -5.49258471e-01 3.46485972e-02 5.53794742e-01
-9.08075511e-01 9.02151227e-01 -6.15055203e-01 -2.69789636e-01
-1.08020222e+00 5.41799366e-01 7.37246156e-01 -2.58338362e-01
-2.38364905e-01 9.69812155e-01 3.45804036e-01 -7.36588359e-01
1.63099691e-01 -2.83694714e-01 -5.89967430e-01 -1.97748020e-01
2.56773233e-01 1.73663631e-01 -3.67425293e-01 -6.25968039e-01
-3.46886158e-01 5.53811312e-01 -2.84414589e-02 -5.98015368e-01
1.13187921e+00 -1.26526505e-01 4.26277556e-02 5.10289252e-01
3.09937328e-01 -1.41834855e-01 -1.43576992e+00 1.45770937e-01
1.52678162e-01 -3.05320203e-01 -1.54140040e-01 -1.12333012e+00
-3.55711222e-01 3.25363219e-01 7.89577544e-01 2.48375714e-01
6.53570294e-01 1.44966826e-01 7.05489069e-02 8.75187695e-01
8.85259688e-01 -1.14992821e+00 2.86033094e-01 9.76073265e-01
8.81661355e-01 -1.35990274e+00 -8.91087353e-02 3.12199481e-02
-8.26551497e-01 6.36507273e-01 1.16297841e+00 -2.56772432e-02
-5.29752299e-03 -1.48113251e-01 4.26700622e-01 -4.36183900e-01
-6.68839633e-01 -6.23089790e-01 -2.06908032e-01 1.51788402e+00
2.35504463e-01 2.23535120e-01 -1.20744646e-01 6.41016115e-04
-5.43596864e-01 -1.13147952e-01 7.61226177e-01 9.99094665e-01
-8.94840360e-01 -7.27648199e-01 -5.48658669e-01 2.29381800e-01
4.76937264e-01 7.51795173e-02 -5.36234200e-01 1.13224471e+00
4.65298668e-02 1.10731280e+00 1.34794593e-01 -3.90492469e-01
7.86495447e-01 2.72010982e-01 6.24697566e-01 -8.34062397e-01
-4.79955763e-01 -6.08540595e-01 4.76115495e-01 -7.18971848e-01
-7.12275326e-01 -1.04679132e+00 -1.69381964e+00 -3.07458103e-01
3.75133485e-01 -1.92642938e-02 3.35554034e-01 1.05042744e+00
-2.31757425e-02 9.30908799e-01 -1.75650775e-01 -9.81777430e-01
-2.40152389e-01 -8.76229823e-01 1.74317863e-02 4.68414128e-01
4.20007527e-01 -8.51471424e-01 -2.77531117e-01 -2.63600707e-01] | [4.28153133392334, 0.8116440773010254] |
b1578979-b3e3-4376-92a1-54f6d9beb2ab | active-learning-driven-surrogate-modeling-for | 2306.06174 | null | https://arxiv.org/abs/2306.06174v1 | https://arxiv.org/pdf/2306.06174v1.pdf | Active-Learning-Driven Surrogate Modeling for Efficient Simulation of Parametric Nonlinear Systems | When repeated evaluations for varying parameter configurations of a high-fidelity physical model are required, surrogate modeling techniques based on model order reduction are desired. In absence of the governing equations describing the dynamics, we need to construct the parametric reduced-order surrogate model in a non-intrusive fashion. In this setting, the usual residual-based error estimate for optimal parameter sampling associated with the reduced basis method is not directly available. Our work provides a non-intrusive optimality criterion to efficiently populate the parameter snapshots, thereby, enabling us to effectively construct a parametric surrogate model. We consider separate parameter-specific proper orthogonal decomposition (POD) subspaces and propose an active-learning-driven surrogate model using kernel-based shallow neural networks, abbreviated as ActLearn-POD-KSNN surrogate model. To demonstrate the validity of our proposed ideas, we present numerical experiments using two physical models, namely Burgers' equation and shallow water equations. Both the models have mixed -- convective and diffusive -- effects within their respective parameter domains, with each of them dominating in certain regions. The proposed ActLearn-POD-KSNN surrogate model efficiently predicts the solution at new parameter locations, even for a setting with multiple interacting shock profiles. | ['Peter Benner', 'Lihong Feng', 'Harshit Kapadia'] | 2023-06-09 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [-1.38343140e-01 -2.26107508e-01 1.12163253e-01 3.48204076e-01
-7.76857615e-01 -3.56235951e-01 4.06991988e-01 8.48262832e-02
-2.45288372e-01 1.11396420e+00 -1.44874500e-02 -3.04996073e-01
-5.34891009e-01 -5.76632857e-01 -5.31481922e-01 -1.32289231e+00
-3.07735682e-01 3.83416921e-01 -6.23820797e-02 -2.24356279e-01
2.09125251e-01 7.42527485e-01 -1.17440653e+00 -2.80492812e-01
1.39621615e+00 8.12241018e-01 -2.54744321e-01 8.25433850e-01
1.29629493e-01 5.44924617e-01 -9.40458477e-03 3.50791663e-01
4.69198048e-01 -4.46174949e-01 -3.61517340e-01 -5.47007136e-02
-1.28738275e-02 -3.61113161e-01 -3.21854621e-01 8.34242642e-01
4.76805627e-01 9.20989633e-01 1.13977075e+00 -8.77663076e-01
-4.76605266e-01 5.28941527e-02 -3.98954928e-01 2.21894875e-01
-1.32645145e-01 5.34519196e-01 4.57242280e-01 -1.22839034e+00
3.25819790e-01 9.42237198e-01 1.09931421e+00 2.15765148e-01
-1.49704719e+00 -4.10993308e-01 -7.86731243e-02 -2.55552888e-01
-1.48653793e+00 -4.96088445e-01 1.15116704e+00 -7.27963448e-01
7.21941590e-01 3.25199753e-01 7.77730823e-01 1.04810381e+00
2.67964482e-01 3.45631540e-01 1.20961463e+00 -3.19667727e-01
5.76860428e-01 1.10470079e-01 5.11359215e-01 5.38484514e-01
6.16860867e-01 5.31706691e-01 -2.27220267e-01 -1.08606780e+00
9.68100131e-01 -1.89915448e-02 -5.56119382e-01 -8.39755177e-01
-7.62215078e-01 8.38225186e-01 -1.03793770e-01 -1.03814593e-02
-6.90117002e-01 -1.14702649e-01 1.90642908e-01 7.15972409e-02
8.05874586e-01 7.45866239e-01 -2.46071577e-01 -5.44476733e-02
-1.17018509e+00 4.23905849e-01 1.00859725e+00 3.63622129e-01
6.30993366e-01 6.60749912e-01 1.03840970e-01 6.55417383e-01
4.58930880e-01 6.52733386e-01 2.29780048e-01 -1.15842307e+00
6.89980015e-02 3.88125211e-01 8.80052805e-01 -8.59077513e-01
-2.25085109e-01 -5.74223876e-01 -1.24241984e+00 4.60945398e-01
3.47362250e-01 -5.08209288e-01 -5.90520203e-01 1.44019473e+00
8.75996470e-01 6.64652944e-01 4.33699012e-01 9.74088728e-01
4.52678382e-01 1.01110041e+00 -3.32771569e-01 -7.49931812e-01
6.68545485e-01 -8.77189875e-01 -5.55262804e-01 2.80076295e-01
5.65040886e-01 -4.05414313e-01 9.88085330e-01 4.06174511e-01
-1.17460358e+00 -3.50732297e-01 -1.21243203e+00 5.24217010e-01
-3.82256024e-02 -8.84774253e-02 2.81496465e-01 2.47089356e-01
-8.20915461e-01 1.03003836e+00 -1.08302486e+00 -9.34967250e-02
-1.47435457e-01 5.70580252e-02 -2.06728756e-01 4.58586484e-01
-9.95427668e-01 7.37550199e-01 2.15481455e-03 4.18470681e-01
-1.23906481e+00 -1.06184661e+00 -5.45155227e-01 1.58334356e-02
-1.97724570e-02 -8.53268743e-01 9.68982458e-01 -4.80444342e-01
-1.73815858e+00 6.82333037e-02 -3.66907790e-02 -8.23334381e-02
6.33270025e-01 -3.21693271e-01 -1.68460891e-01 3.14810991e-01
-2.04637289e-01 -3.09427857e-01 1.07465911e+00 -1.69799769e+00
3.10212165e-01 -2.36710310e-02 -1.72827259e-01 4.40362483e-01
-3.64972264e-01 -4.35154289e-01 1.71603903e-01 -5.10093689e-01
1.88922659e-01 -1.04052627e+00 -6.62225008e-01 4.03403789e-02
-2.69317985e-01 2.43427023e-01 8.28359723e-01 -6.89600229e-01
1.22484887e+00 -1.94824064e+00 2.43108302e-01 1.63633198e-01
2.14509904e-01 5.24494827e-01 2.96669036e-01 7.63883829e-01
-2.06837282e-01 -5.09503298e-02 -7.90905654e-01 -4.60917175e-01
-3.01927805e-01 1.12235002e-01 -7.56886482e-01 7.32100666e-01
1.76960841e-01 7.82593638e-02 -8.10325027e-01 -2.87477374e-01
1.85350582e-01 6.54548168e-01 -6.26727045e-01 4.96186405e-01
1.16525836e-01 8.94570529e-01 -6.78717256e-01 3.62139136e-01
9.41054523e-01 -2.17751488e-01 -2.87500173e-01 -1.25262395e-01
-5.18526137e-01 -3.48605245e-01 -1.60054374e+00 1.05668843e+00
-6.67843223e-01 2.16185689e-01 7.34391689e-01 -1.41505361e+00
9.30891871e-01 4.12314773e-01 5.97761333e-01 -5.90021089e-02
1.30989440e-02 3.88720959e-01 -2.14359254e-01 -7.08554804e-01
3.98704410e-01 -3.79446298e-01 1.63449004e-01 4.05765057e-01
-1.93971485e-01 -3.40366922e-02 -9.56268534e-02 1.01272754e-01
9.49364543e-01 2.03223526e-01 4.05533165e-01 -7.47988284e-01
5.97776771e-01 1.51101604e-01 4.73704994e-01 7.71789491e-01
-1.97176382e-01 5.38728535e-01 3.29157650e-01 -4.78128225e-01
-1.12240613e+00 -8.78188193e-01 -5.47415197e-01 4.65408057e-01
1.23666435e-01 1.78471640e-01 -5.19137383e-01 1.95399225e-02
7.51868188e-02 5.98588467e-01 -6.58761084e-01 -3.63289803e-01
-6.87029123e-01 -1.23970091e+00 6.19926572e-01 2.18561649e-01
4.15556371e-01 -7.72410393e-01 -6.08594894e-01 4.04414415e-01
1.87997848e-01 -5.17575622e-01 3.81134041e-02 5.44293106e-01
-1.32610834e+00 -1.08412004e+00 -9.33192134e-01 -3.74738455e-01
6.89037383e-01 -1.18296854e-01 7.07172096e-01 -9.89463404e-02
-7.82886744e-02 3.64006847e-01 9.37817094e-04 2.47465789e-01
-3.84839267e-01 -3.04642677e-01 4.66900200e-01 3.04010302e-01
-4.34372485e-01 -9.11221623e-01 -6.47286952e-01 3.67178880e-02
-8.31765711e-01 -1.88150272e-01 1.64819047e-01 1.12126422e+00
4.13524389e-01 -1.95181385e-01 4.73120302e-01 -4.96524960e-01
9.22524989e-01 -7.87206948e-01 -7.74049997e-01 2.27499194e-02
-5.00112236e-01 2.03533649e-01 1.11757374e+00 -8.37567449e-01
-1.34998643e+00 -1.25429451e-01 1.51974857e-01 -7.99521625e-01
-8.88069794e-02 5.53716183e-01 2.00173721e-01 -3.30189735e-01
8.06068122e-01 4.16904151e-01 -1.46385550e-01 -6.50914967e-01
2.58894898e-02 4.90866482e-01 4.39839900e-01 -1.14402044e+00
9.47127879e-01 5.73359549e-01 2.31290206e-01 -1.26483309e+00
-5.31116247e-01 -5.04864275e-01 -4.40741628e-01 -1.34055808e-01
4.38939303e-01 -6.15289330e-01 -6.44064248e-01 6.46684766e-01
-1.13966751e+00 -5.07797599e-01 -5.11780083e-01 8.11391413e-01
-5.19534647e-01 5.95320165e-01 -7.69477010e-01 -1.51095366e+00
-4.26351309e-01 -9.86130953e-01 5.59736788e-01 1.99188456e-01
-4.29288000e-02 -1.33651197e+00 9.79110897e-01 -9.74944904e-02
5.19302666e-01 6.21506929e-01 8.39474201e-01 -4.78737861e-01
-9.49872434e-02 -6.82194978e-02 1.94355994e-01 3.24025333e-01
-1.16159305e-01 4.25835490e-01 -8.83065701e-01 -3.89349848e-01
6.40502930e-01 -2.40944609e-01 7.12073922e-01 6.18165314e-01
6.40110195e-01 -6.31605923e-01 -4.59123284e-01 1.07062292e+00
1.51535726e+00 3.34158212e-01 6.40935451e-02 -8.91100764e-02
5.94201624e-01 3.74809802e-01 2.25154638e-01 9.14040089e-01
-2.25934163e-01 2.24342465e-01 1.64964646e-01 -2.51777381e-01
4.19087529e-01 6.32140189e-02 3.30377311e-01 1.16245043e+00
-2.98328370e-01 -9.39652771e-02 -1.08758128e+00 5.88198543e-01
-1.80339706e+00 -6.81891084e-01 -3.10009420e-01 2.36295795e+00
8.13990891e-01 -1.13467053e-01 -1.33099481e-01 3.67039219e-02
6.21026158e-01 9.82908085e-02 -9.24840450e-01 -4.79796916e-01
-1.18222798e-03 6.25642985e-02 2.25678489e-01 8.68115366e-01
-1.00969589e+00 2.49843240e-01 6.49664354e+00 5.12619376e-01
-1.29605460e+00 8.98115039e-02 4.51309890e-01 5.68676852e-02
-7.85400197e-02 2.95802623e-01 -7.20822334e-01 3.71774167e-01
9.57941473e-01 -1.47466645e-01 3.24866503e-01 9.17366922e-01
7.82987535e-01 -1.48438334e-01 -6.62301600e-01 8.21763813e-01
-2.12204531e-01 -1.24442708e+00 -9.13010538e-02 1.53702702e-02
9.36154842e-01 -3.23078781e-01 -8.92656073e-02 2.83493310e-01
2.71614611e-01 -7.82318413e-01 4.16968971e-01 9.92397666e-01
5.72328031e-01 -5.21153510e-01 4.42627281e-01 6.74471080e-01
-1.08774638e+00 1.39956055e-02 -3.60147387e-01 -2.86175877e-01
3.16966623e-01 6.56375289e-01 -2.79019386e-01 3.18279386e-01
4.43917185e-01 5.37168622e-01 3.50688323e-02 1.22367287e+00
4.70561832e-01 9.25422311e-01 -6.79226100e-01 1.31047830e-01
2.87814617e-01 -8.56578529e-01 1.12844467e+00 9.26014662e-01
7.30338573e-01 5.47042906e-01 1.79176942e-01 1.05919492e+00
5.16149521e-01 1.33357838e-01 -6.70567095e-01 -1.91080235e-02
3.78925264e-01 1.25172579e+00 -6.72254920e-01 -3.51864785e-01
-3.12616187e-03 5.72950602e-01 1.56546995e-01 9.23668981e-01
-8.18591952e-01 -2.34734297e-01 7.40152359e-01 1.73706949e-01
2.21964791e-01 -5.43468356e-01 -1.75210848e-01 -1.40971982e+00
-1.90265656e-01 -5.89424789e-01 1.68795168e-01 -6.61690891e-01
-1.43747699e+00 5.31357288e-01 4.91994828e-01 -1.44165385e+00
-1.85886517e-01 -5.06667078e-01 -1.08882177e+00 1.07168901e+00
-1.27615762e+00 -7.11342216e-01 -2.73750693e-01 5.55267036e-01
1.22788541e-01 -4.91239876e-02 9.32469249e-01 2.25978605e-02
-9.31222737e-01 1.46886483e-01 1.08594871e+00 -2.19988674e-01
5.02450466e-01 -9.18330610e-01 -1.86279908e-01 9.11724329e-01
-7.04331398e-01 9.38863754e-01 1.11450005e+00 -7.78323829e-01
-1.57319522e+00 -7.31186688e-01 1.12772807e-01 1.31390486e-02
7.78123081e-01 -1.02813214e-01 -1.35673738e+00 2.25498796e-01
1.01436168e-01 4.64029670e-01 5.17768145e-01 -1.33838117e-01
7.75587112e-02 -8.02342221e-02 -1.11364520e+00 5.42473197e-01
6.15654171e-01 -3.24498534e-01 -6.87435627e-01 1.70203760e-01
4.17793930e-01 -1.08653232e-01 -9.12382305e-01 5.97896814e-01
3.11432183e-01 -7.81909287e-01 1.06367207e+00 -8.38207066e-01
1.73072755e-01 -2.65185714e-01 -4.18782141e-03 -1.44347644e+00
-3.62658232e-01 -1.09715819e+00 -7.13065267e-01 9.53595638e-01
6.28912523e-02 -8.54660332e-01 5.65384805e-01 6.74194813e-01
-3.44612777e-01 -1.06773591e+00 -1.09027135e+00 -8.08737934e-01
4.93220478e-01 -2.30745882e-01 1.92647189e-01 1.19217527e+00
2.74049602e-02 3.66093107e-02 -4.52868491e-01 6.13534808e-01
1.02300394e+00 5.98954894e-02 6.07167959e-01 -1.23771286e+00
-5.17277420e-01 -1.79506406e-01 3.72217536e-01 -8.82624209e-01
1.17650926e-01 -3.08411360e-01 1.40988693e-01 -1.28830767e+00
-8.76769125e-02 -6.46951377e-01 -2.10481480e-01 -2.99708080e-02
-1.44116268e-01 -9.92941707e-02 -3.29849362e-01 5.42246699e-01
-1.21855596e-02 1.10138822e+00 1.12318623e+00 1.47250101e-01
-5.77248275e-01 -1.02495685e-01 -1.18740447e-01 8.66000295e-01
5.54657340e-01 -4.41790432e-01 -5.30685246e-01 4.25494611e-02
4.08693179e-02 7.21936882e-01 4.88949597e-01 -1.25243652e+00
2.93345064e-01 -4.46210951e-01 2.39152327e-01 -4.84193504e-01
6.12131953e-01 -5.38326919e-01 3.27905983e-01 3.88821304e-01
-2.28715807e-01 -5.54685108e-02 3.61651421e-01 6.24247313e-01
-1.53480530e-01 -3.89751315e-01 1.14584923e+00 -9.66221914e-02
-1.73919916e-01 2.53101170e-01 -6.98983133e-01 2.43120983e-01
8.60582113e-01 -4.26776111e-01 -4.05737981e-02 -2.62309134e-01
-9.30868685e-01 -1.51947826e-01 5.71659625e-01 -4.81736153e-01
4.65207219e-01 -1.26785195e+00 -5.95133066e-01 4.00125176e-01
-1.92977563e-01 9.98966843e-02 5.63479781e-01 1.11453724e+00
-7.52781034e-01 9.32289809e-02 -9.51812938e-02 -4.89056170e-01
-5.64769804e-01 4.67993617e-01 6.94447637e-01 -4.19868946e-01
-5.90572715e-01 5.55323184e-01 3.91060472e-01 -6.48334086e-01
-1.57195628e-01 -4.10136402e-01 -1.84105083e-01 8.02205279e-02
1.62447199e-01 8.26529860e-01 -6.20821118e-02 -5.69971323e-01
4.55429517e-02 5.98686159e-01 5.56329966e-01 -2.41757408e-01
1.41913497e+00 -3.58972959e-02 2.25453917e-03 6.83430374e-01
1.25743747e+00 4.15550023e-02 -1.62317252e+00 -8.25256854e-02
-4.81688440e-01 -4.38254699e-02 2.70148933e-01 -1.98508441e-01
-6.28313065e-01 8.52189779e-01 4.69898790e-01 3.16732168e-01
7.81963766e-01 -6.99583769e-01 7.33147740e-01 4.86264646e-01
-2.36089583e-02 -9.11231518e-01 -2.47879922e-02 6.53507233e-01
8.77606452e-01 -9.28865612e-01 7.74495378e-02 -5.71536645e-02
-2.90231735e-01 1.29399157e+00 6.60552561e-01 -5.42031765e-01
9.48201776e-01 4.77534801e-01 -1.16053417e-01 1.50329126e-02
-7.25141764e-01 4.37384486e-01 2.15760857e-01 2.13120863e-01
7.65256062e-02 -2.92997181e-01 -1.89547271e-01 5.95939815e-01
3.32657784e-01 -1.45060986e-01 5.78308821e-01 8.17468762e-01
-2.89934933e-01 -6.10651910e-01 -8.30491185e-01 2.67611146e-01
3.06593883e-03 -8.95084441e-02 1.60980850e-01 7.45640814e-01
-3.27496052e-01 6.22247338e-01 -1.50574163e-01 7.47389048e-02
3.29427011e-02 1.85042471e-01 5.01517542e-02 -3.38819772e-01
-3.28342527e-01 2.37230211e-01 -3.67355347e-03 -2.01648444e-01
-2.52827644e-01 -7.04444408e-01 -1.15180528e+00 -2.25004360e-01
-3.50542337e-01 7.30852246e-01 5.50205223e-02 9.03245747e-01
2.17296243e-01 1.52640060e-01 6.78611279e-01 -1.27244294e+00
-9.00078177e-01 -1.07408082e+00 -7.53742814e-01 1.32951811e-01
6.83149993e-01 -9.85820532e-01 -1.15997732e+00 -1.53837875e-01] | [6.504958152770996, 3.4567759037017822] |
2761d020-45a0-4d60-a180-40d510f65465 | gamnet-robust-feature-matching-via-graph | null | null | https://dl.acm.org/doi/abs/10.1145/3474085.3475669 | https://dl.acm.org/doi/pdf/10.1145/3474085.3475669 | GAMnet: Robust Feature Matching via Graph Adversarial-Matching Network | Recently, deep graph matching (GM) methods have gained increasing attention. These methods integrate graph nodes¡¯s embedding, node/edges¡¯s affinity learning and final correspondence solver together in an end-to-end manner. For deep graph matching problem, one main issue is how to generate consensus node's embeddings for both source and target graphs that best serve graph matching tasks. In addition, it is also challenging to incorporate the discrete one-to-one matching constraints into the differentiable correspondence solver in deep matching network. To address these issues, we propose a novel Graph Adversarial Matching Network (GAMnet) for graph matching problem. GAMnet integrates graph adversarial embedding and graph matching simultaneously in a unified end-to-end network which aims to adaptively learn distribution consistent and domain invariant embeddings for GM tasks. Also, GAMnet exploits sparse GM optimization as correspondence solver which is differentiable and can also incorporate discrete one-to-one matching constraints approximately in natural in the final matching prediction. Experimental results on three public benchmarks demonstrate the effectiveness and benefits of the proposed GAMnet. | ['Bin Luo', 'Jin Tang', 'Ziyan Zhang', 'Pengfei Sun', 'Bo Jiang'] | 2021-10-17 | null | null | null | mm-2021-10 | ['graph-matching'] | ['graphs'] | [-2.23402351e-01 2.82877356e-01 -2.54959557e-02 -3.36283058e-01
-6.53095782e-01 -4.94970292e-01 3.43096644e-01 -8.94502848e-02
-6.57390729e-02 3.40747058e-01 -9.03037749e-03 -9.21006873e-02
-9.65913087e-02 -1.06069863e+00 -8.21431577e-01 -4.33078706e-01
1.07521834e-02 8.09517205e-01 1.23969555e-01 -2.84323305e-01
-2.00161740e-01 4.48887557e-01 -8.33826840e-01 -1.16309449e-01
1.22130370e+00 8.20820928e-01 8.27611983e-02 5.08423150e-01
-1.54417530e-01 5.72578013e-01 -1.66788667e-01 -7.78769851e-01
7.54098773e-01 -2.68922478e-01 -5.57762444e-01 -8.83770883e-02
9.30706143e-01 -3.93983364e-01 -1.04534400e+00 1.42407000e+00
7.35678017e-01 3.75268906e-01 4.83093470e-01 -1.75358665e+00
-1.11376417e+00 5.19536734e-01 -6.02598131e-01 -3.12901288e-02
2.18986660e-01 5.60140610e-01 1.26632440e+00 -6.91576958e-01
6.21849895e-01 1.49425840e+00 7.74112582e-01 5.41302621e-01
-1.16343653e+00 -9.61952507e-01 3.78594548e-01 1.28240556e-01
-1.36277580e+00 8.21172670e-02 1.11203277e+00 -4.02837485e-01
9.15936410e-01 -5.24201579e-02 5.53140819e-01 7.84739017e-01
2.84697205e-01 6.21472597e-01 4.40692484e-01 1.39177814e-01
-5.97633794e-02 -2.85255834e-02 -1.68691337e-01 9.08791602e-01
1.28714919e-01 2.95478821e-01 -7.33700693e-02 -3.05533260e-01
7.89552212e-01 3.66931617e-01 -1.82095803e-02 -4.92645830e-01
-8.75625968e-01 1.01710761e+00 1.16820610e+00 -1.03778616e-01
-2.86660939e-01 5.72529852e-01 4.15805042e-01 5.17240167e-01
4.35487777e-01 3.32298875e-01 -1.05420016e-01 5.11372566e-01
-5.09476483e-01 4.27845240e-01 6.05591834e-01 1.32854199e+00
1.10572624e+00 2.15859115e-01 -2.79129922e-01 6.94926679e-01
5.68983912e-01 5.82904518e-01 2.04927474e-01 -6.66774929e-01
7.15291321e-01 1.04758644e+00 -3.70879263e-01 -1.45560539e+00
-2.40366250e-01 -6.22573018e-01 -1.25731599e+00 -5.44421710e-02
1.69777974e-01 2.24745441e-02 -1.08040524e+00 1.86619675e+00
5.99229753e-01 7.94630885e-01 -8.69669542e-02 1.04427135e+00
1.06526577e+00 5.98631382e-01 1.53104112e-01 5.63288867e-01
1.13864422e+00 -1.31368923e+00 -3.46933395e-01 -5.55047512e-01
6.32483780e-01 -5.43985367e-01 8.86388600e-01 -3.63572121e-01
-9.72963691e-01 -4.94225234e-01 -9.49162424e-01 -2.92869687e-01
-3.60675484e-01 -3.53610128e-01 7.03890204e-01 1.91065192e-01
-1.01252794e+00 8.00871253e-01 -6.78511441e-01 -1.08301401e-01
6.04062080e-01 7.14868307e-01 -5.54399490e-01 -4.23664600e-01
-1.54529607e+00 5.75888336e-01 1.87944993e-01 1.52680933e-01
-8.13935816e-01 -1.07381737e+00 -1.23004377e+00 1.68152824e-01
1.92356884e-01 -1.19383729e+00 7.94481695e-01 -9.75740671e-01
-1.15523350e+00 1.01760137e+00 1.71086013e-01 -2.93183565e-01
5.40062845e-01 3.06601703e-01 -4.10118103e-01 1.90171488e-02
8.38816985e-02 6.16326332e-01 7.32321799e-01 -8.91126275e-01
-2.16201037e-01 -5.85943639e-01 2.13769838e-01 2.59920448e-01
-2.25201055e-01 -1.57441795e-01 -5.61969280e-01 -7.41003990e-01
8.45489651e-03 -9.99929428e-01 -3.91132832e-01 3.55160594e-01
-3.18002731e-01 -3.54586959e-01 6.84379816e-01 -8.28857780e-01
1.03972113e+00 -2.02321434e+00 3.97664368e-01 4.65733111e-01
6.68117940e-01 1.97843045e-01 -6.26162946e-01 5.63568771e-01
-1.97550908e-01 -9.25082788e-02 -3.22280169e-01 -3.99501592e-01
4.62149888e-01 1.40377074e-01 -9.36367810e-02 7.64996886e-01
3.77251357e-01 1.43688738e+00 -1.06027651e+00 -4.45181489e-01
2.57164717e-01 5.15631557e-01 -8.80341470e-01 5.72204351e-01
-1.72731727e-01 3.42393041e-01 -4.87490982e-01 5.63350737e-01
1.16596591e+00 -3.70172203e-01 7.19359443e-02 -3.13222855e-01
5.90507865e-01 -4.83004823e-02 -1.16901517e+00 1.87718439e+00
-4.30606723e-01 2.64897287e-01 3.12220752e-01 -1.11473143e+00
1.06690288e+00 -9.78822471e-04 3.39033842e-01 -6.97855830e-01
1.11144349e-01 2.70972908e-01 -5.89233413e-02 8.70642737e-02
2.20945150e-01 -1.51569247e-01 -2.21235707e-01 2.07593724e-01
8.77955928e-02 -7.41365626e-02 -1.83064729e-01 4.33767498e-01
1.25043523e+00 -1.89441025e-01 5.41739948e-02 -3.57589930e-01
7.05366611e-01 -1.05271347e-01 8.50302875e-01 4.00938809e-01
-1.86729074e-01 6.35466099e-01 5.08647978e-01 -5.50125360e-01
-9.73161161e-01 -1.10350573e+00 4.24412161e-01 8.28042865e-01
5.70562303e-01 -1.85384825e-01 -6.14240944e-01 -8.27232778e-01
4.76707339e-01 1.52221501e-01 -6.59679949e-01 -6.19147599e-01
-8.13539326e-01 -2.35417157e-01 4.21844274e-01 5.94894230e-01
5.24622858e-01 -9.98721957e-01 5.30803800e-01 3.37961644e-01
2.61262119e-01 -1.07247841e+00 -1.35445333e+00 -3.28761607e-01
-4.74137723e-01 -1.12098932e+00 -7.03193963e-01 -1.16270411e+00
1.03947723e+00 2.90019840e-01 1.14325798e+00 5.05361080e-01
-3.19768190e-01 2.57796645e-01 -6.55091554e-02 1.37304693e-01
-3.96145910e-01 3.81931543e-01 -2.47155726e-01 3.90648749e-03
2.96030402e-01 -8.61030340e-01 -8.85658562e-01 5.06003022e-01
-8.93838942e-01 -9.98027623e-02 4.14093018e-01 9.75399017e-01
5.02772331e-01 -3.26484650e-01 5.85202098e-01 -1.08360076e+00
7.15348482e-01 -6.74100399e-01 -9.09538090e-01 3.57496738e-01
-7.01494217e-01 1.58508122e-01 1.04776049e+00 -4.11301166e-01
-2.77515799e-01 -1.18555212e-02 -2.76754379e-01 -1.25421715e+00
4.72070456e-01 3.97529006e-01 -5.93238592e-01 -4.86912400e-01
4.53341842e-01 1.64326772e-01 2.62529314e-01 -1.95422590e-01
5.60648203e-01 1.19638525e-01 7.23829567e-01 -5.89528024e-01
1.50161695e+00 2.74519920e-01 6.01186678e-02 -2.11364061e-01
-3.77947718e-01 -6.00130022e-01 -3.21732670e-01 1.39261201e-01
7.34575272e-01 -1.14036059e+00 -7.63163269e-01 5.81429005e-01
-1.04081154e+00 -3.21320146e-01 -1.04933187e-01 2.36937448e-01
-3.75355363e-01 6.05518401e-01 -4.56612706e-01 -1.22766040e-01
-8.61972809e-01 -1.17732179e+00 1.06308496e+00 3.40158731e-01
6.19202778e-02 -1.43079209e+00 1.31987303e-01 1.65334389e-01
6.17231131e-01 5.22048771e-01 6.88020647e-01 -5.86214066e-01
-9.08608913e-01 -4.29421961e-01 -6.70597672e-01 1.62337810e-01
1.33612692e-01 -2.43369877e-01 -4.56325442e-01 -7.30970323e-01
-5.32808065e-01 -8.49151537e-02 7.72634149e-01 1.19236313e-01
9.89091158e-01 -4.35573846e-01 -4.10349160e-01 1.39509106e+00
1.60338247e+00 -2.37201899e-01 5.98406017e-01 1.17302248e-02
1.25924206e+00 3.50242078e-01 5.02954781e-01 1.26993164e-01
6.69983804e-01 6.86389565e-01 6.27732098e-01 -2.95758635e-01
-2.77618527e-01 -6.80707932e-01 2.01225638e-01 7.81918347e-01
5.86877167e-01 -1.65728763e-01 -7.76823759e-01 5.64116657e-01
-2.08741403e+00 -7.83883750e-01 2.41054092e-02 1.96031988e+00
3.57159317e-01 -1.69277027e-01 9.94024575e-02 -4.83211279e-01
1.08139658e+00 3.39675337e-01 -9.08825099e-01 -1.95596159e-01
9.51031595e-02 3.08078647e-01 6.45879686e-01 6.14730954e-01
-8.88163090e-01 1.11991131e+00 4.71948862e+00 8.85974884e-01
-1.09489286e+00 4.65517975e-02 3.38130891e-01 4.09707099e-01
-9.30346251e-01 3.26975942e-01 -5.45704782e-01 6.10365570e-01
5.04623115e-01 -5.76325476e-01 8.00943017e-01 8.88845921e-01
-2.64208049e-01 6.61589742e-01 -1.14846408e+00 1.12117219e+00
-1.40920550e-01 -1.53572202e+00 9.71221775e-02 4.61620614e-02
7.81856775e-01 1.03648469e-01 -1.63649488e-02 5.09368598e-01
6.97462201e-01 -1.14002895e+00 2.56901979e-01 2.41044939e-01
8.00337136e-01 -8.65632117e-01 5.63488722e-01 1.31851166e-01
-1.81036997e+00 1.92993626e-01 -6.68046951e-01 2.49224052e-01
1.26567900e-01 2.36314386e-01 -8.85526299e-01 9.06197608e-01
2.92352110e-01 1.01224065e+00 -3.59373391e-01 1.02844167e+00
-1.45353213e-01 -9.69599583e-04 -2.11246192e-01 1.48117661e-01
5.79468429e-01 -5.36070585e-01 5.48083127e-01 8.73612642e-01
1.86332077e-01 -1.62920251e-01 5.32816231e-01 1.35143244e+00
-5.31396568e-01 -7.98354447e-02 -7.03149974e-01 -1.53927818e-01
6.82766378e-01 1.38914335e+00 -2.62349457e-01 -1.19211087e-02
-3.98163378e-01 1.18763936e+00 7.04169273e-01 2.97262669e-01
-9.92501378e-01 -6.28969133e-01 1.07637501e+00 2.54023999e-01
1.15946464e-01 -2.73183268e-02 1.44661337e-01 -1.21695817e+00
4.39010933e-02 -1.05526733e+00 5.47764540e-01 -5.18345594e-01
-1.88371253e+00 6.24211311e-01 -4.03643101e-01 -1.12945175e+00
-1.98144261e-02 -3.79182190e-01 -1.00740957e+00 1.30614054e+00
-1.71457386e+00 -1.57730544e+00 -6.77009463e-01 9.40679848e-01
-3.56155634e-02 -2.66513526e-01 6.17617249e-01 6.82844818e-01
-6.95740581e-01 1.15024626e+00 -3.68471071e-02 5.50219119e-01
7.37652898e-01 -1.34962153e+00 1.20689118e+00 8.13917160e-01
-2.28404239e-01 4.79426295e-01 2.92729557e-01 -7.09572434e-01
-1.78754580e+00 -1.55917478e+00 5.81197679e-01 -1.81294575e-01
8.31893563e-01 -5.25400519e-01 -1.07336152e+00 9.23017919e-01
4.63233143e-02 6.33844852e-01 1.65855780e-01 -2.15384796e-01
-5.88790596e-01 -4.30691093e-01 -1.38894892e+00 5.17345369e-01
1.38297129e+00 -7.69945145e-01 -1.43152162e-01 5.92045665e-01
1.03522384e+00 -7.90123045e-01 -1.09580493e+00 3.53344202e-01
1.58595845e-01 -4.39493716e-01 1.15616775e+00 -8.45259428e-01
1.70490667e-01 -3.66219193e-01 -4.90345396e-02 -1.44591212e+00
-4.90712851e-01 -1.05648541e+00 1.20346896e-01 1.17357600e+00
2.65431315e-01 -1.14846385e+00 1.04335856e+00 8.15314829e-01
-2.61782050e-01 -6.96865559e-01 -8.55795145e-01 -7.51031995e-01
3.73404711e-01 1.24387138e-01 1.04383934e+00 1.13810694e+00
-3.93458188e-01 2.24088043e-01 -5.58083951e-01 4.05200481e-01
9.81286466e-01 2.51759440e-01 1.24156582e+00 -1.10496604e+00
-3.79199833e-01 -4.54863757e-01 -8.32641721e-01 -1.04234195e+00
7.54872859e-01 -1.49453950e+00 -2.24837437e-01 -1.72908497e+00
7.74892643e-02 -6.54495180e-01 -1.03308253e-01 2.96455503e-01
-4.27625448e-01 -1.53420523e-01 1.36324644e-01 -1.17306083e-01
-3.36921990e-01 8.32230091e-01 1.45122302e+00 -4.97310907e-01
1.01030447e-01 -1.42800376e-01 -8.18506658e-01 1.98361859e-01
5.46018958e-01 -5.40532351e-01 -5.96069038e-01 -6.12760067e-01
2.15174124e-01 1.27890170e-01 4.59851086e-01 -5.89184582e-01
6.82811677e-01 -1.26149192e-01 -8.66196305e-02 -3.47779661e-01
1.26008749e-01 -1.09324849e+00 4.83522773e-01 4.62083340e-01
-1.06005557e-03 3.58029485e-01 2.34625190e-01 8.07374179e-01
-2.52482563e-01 2.29903921e-01 1.05231392e+00 -9.29960981e-02
-6.75267041e-01 1.30152678e+00 6.50943637e-01 5.26201844e-01
1.09534359e+00 -2.13023394e-01 -4.04890776e-01 -2.82766223e-01
-5.37439048e-01 7.47793913e-01 6.45625114e-01 5.48616648e-01
6.39572918e-01 -1.80737484e+00 -9.45574224e-01 2.91529626e-01
1.88586190e-01 3.81978303e-01 4.43542272e-01 6.43489122e-01
-8.44534636e-01 -1.28378674e-01 -7.69269839e-02 -4.12419528e-01
-9.44923878e-01 6.21607602e-01 7.78772414e-01 -6.28474653e-01
-7.26262629e-01 9.76714492e-01 4.56953883e-01 -1.01898324e+00
2.54004836e-01 -7.73529038e-02 4.37797993e-01 -6.04098558e-01
2.00831845e-01 1.98819473e-01 -1.59396037e-01 -4.98996407e-01
-5.43224275e-01 6.72427773e-01 -2.58680075e-01 5.97971201e-01
1.27756548e+00 1.68216258e-01 -2.41051018e-01 -3.38106066e-01
1.71008277e+00 -2.50571102e-01 -1.23028052e+00 -4.81444806e-01
-3.34886193e-01 -6.93668783e-01 -1.17802516e-01 -2.37372071e-01
-1.57085192e+00 7.75921345e-01 2.92515606e-01 -1.51003987e-01
9.01172459e-01 -2.55973130e-01 1.25110006e+00 3.00758127e-02
1.72935978e-01 -7.50698686e-01 5.32909892e-02 2.04347372e-01
9.32276964e-01 -1.34782493e+00 -7.79767036e-02 -6.47327960e-01
-3.87794673e-01 8.15196574e-01 1.02967048e+00 -9.02539253e-01
8.94109130e-01 2.62890081e-03 -1.07495882e-01 -3.58531922e-01
-6.12907112e-01 7.70480372e-03 6.93717182e-01 7.60050654e-01
4.91611473e-02 4.72885296e-02 5.81872612e-02 4.13449734e-01
1.30971791e-02 -4.37417626e-01 -1.31847188e-01 3.67708355e-01
-6.86729848e-02 -1.31362271e+00 9.71950293e-02 3.74840379e-01
-1.01546511e-01 -1.41218036e-01 -5.05892277e-01 8.42005253e-01
-3.36248070e-01 5.33896267e-01 -3.24285999e-02 -5.51862836e-01
4.85725939e-01 -4.91675138e-01 4.20095891e-01 -4.96862262e-01
-9.74530697e-01 -3.87328565e-01 -2.77559429e-01 -6.66961372e-01
1.61857143e-01 -1.62646338e-01 -1.41822088e+00 -8.64055097e-01
-3.27835679e-01 1.56370327e-01 1.94298983e-01 5.29625773e-01
7.23274410e-01 6.40206695e-01 8.00923645e-01 -6.29314959e-01
-8.74989629e-01 -6.96323395e-01 -6.36482000e-01 8.05757225e-01
6.44925311e-02 -4.26733762e-01 -5.77675343e-01 -6.42969310e-01] | [7.115074157714844, 6.40888786315918] |
511e0aae-3e70-4f10-ad29-b9db41d7bf97 | development-of-a-hand-pose-recognition-system-1 | null | null | https://ieeexplore.ieee.org/document/8853573 | https://dennishnf.bitbucket.io/research/2019%20-%20intercon%202019%20-%20hand%20pose%20on%20embedded%20computer%20using%20ai.pdf | Development of a hand pose recognition system on an embedded computer using Artificial Intelligence | The recognition of hand gestures is a very interesting research topic due to the growing demand in recent years in robotics, virtual reality, autonomous driving systems, human-machine interfaces and in other new technologies. Despite several approaches for a robust recognition system, gesture recognition based on visual perception has many advantages over devices such as sensors, or electronic gloves. This paper describes the implementation of a visual-based recognition system on a embedded computer for 10 hand poses recognition. Hand detection is achieved using a tracking algorithm and classification by a light convolutional neural network. Results show an accuracy of 94.50%, a low power consumption and a near real-time response. Thereby, the proposed system could be applied in a large range of applications, from robotics to entertainment. | ['Dennis Núñez-Fernández'] | 2019-10-03 | null | null | null | ieee-international-conference-on-electronics | ['hand-detection'] | ['computer-vision'] | [-2.42874157e-02 -5.12766123e-01 -3.27964127e-01 -1.97705805e-01
2.06516832e-01 -3.79261851e-01 5.59279680e-01 -5.53862035e-01
-8.09532583e-01 4.83990997e-01 -2.57615298e-01 -3.46607149e-01
1.30190188e-02 -6.36495352e-01 -6.02670573e-02 -7.07461417e-01
2.92328924e-01 1.43107310e-01 6.02543831e-01 8.41582939e-02
4.31416392e-01 1.15660453e+00 -1.78595138e+00 -1.74413234e-01
6.12805411e-02 1.06805384e+00 4.73746598e-01 8.14753711e-01
-6.56351820e-02 4.06468600e-01 -3.70045274e-01 -3.91068347e-02
2.95601994e-01 -5.11989705e-02 -2.65003592e-01 1.27641838e-02
1.29058689e-01 -6.33158326e-01 -5.74200213e-01 9.10969138e-01
9.21020389e-01 9.39605087e-02 6.00990057e-01 -1.27219939e+00
-3.01928997e-01 1.33690819e-01 -5.45985520e-01 -9.03312489e-02
6.16169453e-01 5.62648952e-01 3.96333814e-01 -8.94763589e-01
5.93685329e-01 1.15616894e+00 2.61929303e-01 8.68373454e-01
-6.09694362e-01 -7.55134583e-01 -1.38003796e-01 4.94132191e-01
-1.41763949e+00 9.65934247e-03 6.07924700e-01 -6.24044776e-01
1.16935050e+00 2.64843762e-01 6.03118598e-01 1.15290189e+00
2.45232910e-01 7.18487680e-01 1.11795259e+00 -5.67776024e-01
2.20434070e-01 1.78105667e-01 4.50742394e-01 6.32515848e-01
3.92986089e-01 3.14381152e-01 -2.63568401e-01 2.53454715e-01
1.08635640e+00 6.96523070e-01 -2.01434031e-01 -3.41464221e-01
-1.10632658e+00 3.10261667e-01 6.29246891e-01 5.20948470e-01
-6.11187577e-01 2.86424547e-01 2.02178672e-01 7.65297711e-02
-4.16281164e-01 -1.98830396e-01 -2.63172537e-01 -3.25844496e-01
-4.45936739e-01 1.33706450e-01 8.26996326e-01 9.43332314e-01
1.31011531e-02 2.00201064e-01 3.90508063e-02 4.30449724e-01
8.48041356e-01 7.92114258e-01 5.69479227e-01 -3.45776081e-01
3.44306678e-01 7.06852913e-01 1.77867666e-01 -6.61668479e-01
-7.59387732e-01 2.78505176e-01 -8.12920690e-01 1.05807793e+00
6.09028876e-01 -1.21335618e-01 -1.14989328e+00 8.86292219e-01
1.97037175e-01 -2.57292539e-01 -3.04610699e-01 1.49404371e+00
9.88428354e-01 2.70664245e-01 1.39987007e-01 1.42159730e-01
1.61150074e+00 -6.07668757e-01 -8.07728827e-01 -2.43287347e-02
-7.29497820e-02 -8.09038639e-01 8.69830608e-01 8.62611055e-01
-4.05508429e-01 -6.19040310e-01 -1.06133938e+00 1.99788794e-01
-4.85834599e-01 3.60315025e-01 7.03344703e-01 1.29580832e+00
-5.70853591e-01 2.82027900e-01 -1.06175029e+00 -1.14215410e+00
2.30871573e-01 7.41111517e-01 -1.48087755e-01 1.83707669e-01
-4.35445428e-01 9.83643413e-01 4.76161420e-01 2.53122896e-01
-3.66364032e-01 3.03751916e-01 -5.18232644e-01 -1.67308170e-02
3.92384827e-02 -1.26516253e-01 9.70350444e-01 -5.08969963e-01
-1.94801760e+00 7.12571383e-01 -1.17411226e-01 -1.34541720e-01
6.71461225e-01 -2.14289233e-01 -5.27365088e-01 -9.55368578e-02
-7.07258463e-01 2.74187744e-01 9.07579660e-01 -6.39296949e-01
-7.50805795e-01 -7.09640980e-01 -2.56184310e-01 -2.77675483e-02
-3.60663980e-01 3.38674426e-01 -2.55517215e-01 -2.54335284e-01
2.76397109e-01 -1.03818154e+00 -4.04765736e-03 3.57041061e-01
-1.79216504e-01 -5.79874396e-01 1.20282173e+00 -5.23350418e-01
8.13201666e-01 -1.87529409e+00 -3.39402780e-02 3.80725175e-01
-5.29316925e-02 8.36240530e-01 7.99565464e-02 2.80234724e-01
3.58068258e-01 -3.72570276e-01 2.86882013e-01 3.71968180e-01
-6.88674375e-02 9.85569879e-02 -1.37136415e-01 4.47726727e-01
-1.01274371e-01 8.77894402e-01 -6.79853022e-01 -1.28619462e-01
1.16622531e+00 7.83051670e-01 4.60656732e-02 2.27421671e-01
1.57142401e-01 5.32710791e-01 -4.59158093e-01 9.12461758e-01
4.14265543e-01 1.47819623e-01 1.80100188e-01 5.51974438e-02
-3.01158488e-01 -1.23208268e-02 -1.51855505e+00 1.25754249e+00
-3.62462491e-01 1.00887036e+00 2.41956085e-01 -6.42895401e-01
1.14964068e+00 5.89098096e-01 1.85544193e-01 -6.81450427e-01
6.90570951e-01 4.48965013e-01 1.40197337e-01 -9.56576645e-01
3.38327587e-01 1.80842489e-01 5.14925838e-01 4.50527877e-01
-1.93703309e-01 9.28198397e-02 -1.06028073e-01 -5.39890766e-01
9.94304001e-01 3.62167120e-01 3.93167824e-01 2.81799406e-01
5.92263818e-01 -5.25113344e-02 -4.84722890e-02 4.75567997e-01
-3.49440724e-01 5.25134444e-01 -5.08742094e-01 -7.00804472e-01
-7.56625593e-01 -7.46279180e-01 -1.26168504e-01 7.24799871e-01
3.03598762e-01 2.84781039e-01 -4.51416939e-01 -3.58937502e-01
2.08777443e-01 4.42914069e-02 -9.95517373e-02 2.76655734e-01
-8.28142285e-01 -2.04386786e-01 3.91930848e-01 9.77752566e-01
6.96930349e-01 -1.62489200e+00 -1.40017593e+00 2.25079775e-01
5.96745789e-01 -1.14743972e+00 1.13025866e-01 1.41661510e-01
-8.44751596e-01 -1.08833063e+00 -1.12683737e+00 -1.09070194e+00
3.44704062e-01 3.49720240e-01 2.71809608e-01 1.61015823e-01
-7.21173406e-01 5.08548737e-01 -4.25328732e-01 -6.56398773e-01
2.68103871e-02 1.71200648e-01 1.67018816e-01 -1.02433890e-01
9.90109265e-01 -3.54399562e-01 -7.15703726e-01 3.44852567e-01
-3.53785545e-01 -5.07788777e-01 6.92999303e-01 7.55297780e-01
1.24506816e-01 -5.91676414e-01 2.17470437e-01 -6.47402629e-02
6.33038580e-01 6.83156177e-02 -8.67608488e-01 3.14116359e-01
-7.20090806e-01 -8.17719921e-02 3.00408632e-01 -7.32035875e-01
-9.24209297e-01 6.26282334e-01 -3.08773786e-01 -2.98862845e-01
-6.31910026e-01 7.88650811e-02 -1.57426417e-01 -4.31529373e-01
7.00322211e-01 2.40307644e-01 1.56046301e-01 -5.81979692e-01
2.33438909e-01 1.58548558e+00 5.06087422e-01 3.83764319e-02
5.83246708e-01 2.75289416e-01 5.55015281e-02 -1.29028440e+00
3.03181320e-01 -6.95969641e-01 -8.80384147e-01 -5.88789403e-01
8.32599998e-01 -4.81935322e-01 -1.41756046e+00 8.36416781e-01
-1.53729737e+00 -8.10251012e-02 1.68985933e-01 1.06836641e+00
-2.07306609e-01 2.40594089e-01 -3.77084136e-01 -1.37271237e+00
-6.39257193e-01 -1.13472080e+00 7.56006777e-01 6.65509522e-01
-3.44802737e-01 -4.61367667e-01 -1.50911525e-01 1.12990730e-01
4.10290897e-01 -7.28862286e-02 2.14791149e-01 -3.88242215e-01
-8.59494627e-01 -6.20193005e-01 -4.62177157e-01 -2.12874152e-02
3.51461858e-01 1.15564913e-01 -1.07049716e+00 -2.25350544e-01
-3.69082451e-01 -2.43186414e-01 5.90730071e-01 2.14198992e-01
8.69243205e-01 5.95319569e-02 -6.84185088e-01 3.63361448e-01
1.48530471e+00 9.60854709e-01 7.62658179e-01 3.26328367e-01
7.22983837e-01 3.11404228e-01 3.51650536e-01 5.61773360e-01
-1.03582993e-01 8.39618921e-01 4.69007313e-01 2.18260095e-01
-2.47975439e-01 2.35909894e-01 1.53381824e-01 5.18456876e-01
-7.43869722e-01 -2.12508827e-01 -1.05339551e+00 2.91304618e-01
-1.81843436e+00 -9.15229440e-01 -3.88865650e-01 2.38668752e+00
2.76044846e-01 -1.51007444e-01 2.22541854e-01 5.70152760e-01
7.73415446e-01 -1.78203672e-01 -6.23106062e-01 -5.87448299e-01
3.14457715e-01 6.46212876e-01 6.40079141e-01 2.17442364e-01
-1.03276575e+00 8.67792010e-01 6.35058928e+00 2.85986930e-01
-1.72547030e+00 -4.87949373e-03 -3.45749050e-01 1.74059778e-01
6.49596572e-01 -7.03467667e-01 -8.15473139e-01 4.14315879e-01
2.42951676e-01 3.25292021e-01 6.66515529e-01 1.06744814e+00
1.18721649e-02 -9.99652520e-02 -1.01436818e+00 1.46615922e+00
1.38729945e-01 -8.19160640e-01 -9.65407491e-02 9.43829864e-02
1.68441921e-01 -1.09719941e-02 -1.54887035e-01 -7.18197227e-02
-1.15752973e-01 -1.06237042e+00 5.46930373e-01 2.98438281e-01
8.39581847e-01 -5.90400755e-01 8.45891058e-01 3.86738718e-01
-1.33965898e+00 -2.19672695e-01 -2.65088767e-01 -7.08559930e-01
5.60610071e-02 -1.55218234e-02 -8.94009233e-01 -7.22724050e-02
7.71781981e-01 2.77196139e-01 -3.53708081e-02 1.27990127e+00
-3.70905131e-01 3.23361218e-01 -4.07472312e-01 -1.02719450e+00
-1.30341470e-01 -1.86132595e-01 4.02626395e-01 1.41567433e+00
2.44580597e-01 4.54238296e-01 2.95850891e-03 5.47278225e-01
2.73842424e-01 1.01552330e-01 -8.27885389e-01 -4.25433181e-02
2.58389950e-01 1.26818585e+00 -1.02676082e+00 -3.22618723e-01
-5.58501899e-01 1.14111328e+00 -2.06432804e-01 2.69056022e-01
-4.85578775e-01 -1.04255652e+00 5.66982865e-01 -1.31727457e-01
3.28850776e-01 -7.45976925e-01 -6.22394919e-01 -9.05263543e-01
3.37905884e-01 -5.34752548e-01 -1.29351512e-01 -6.71637118e-01
-6.59571886e-01 5.39734304e-01 -5.53948939e-01 -1.50244415e+00
-4.40530479e-01 -1.53145051e+00 -4.73757803e-01 9.26666558e-01
-1.20314133e+00 -9.85776663e-01 -8.95149887e-01 5.72735012e-01
6.07109785e-01 -5.76061249e-01 9.67440188e-01 2.29441360e-01
-3.86490107e-01 4.18818742e-01 -1.01087745e-02 5.44207454e-01
4.79477525e-01 -8.55439126e-01 2.67141491e-01 7.31724620e-01
1.82268277e-01 5.18161893e-01 2.53523260e-01 -4.79149610e-01
-1.80032730e+00 -3.94143879e-01 5.80636322e-01 -3.07569504e-01
2.63683140e-01 -1.58773035e-01 -5.65824568e-01 4.54705894e-01
2.47526303e-01 1.23488396e-01 6.37481287e-02 -2.61995196e-01
-2.55502582e-01 -1.13625027e-01 -1.27817178e+00 4.72137362e-01
1.09804118e+00 -4.34556872e-01 -5.80019414e-01 1.02601692e-01
-1.33539647e-01 -4.86012399e-01 -5.15024245e-01 1.80356175e-01
1.53027570e+00 -7.16270566e-01 7.65861154e-01 -3.89601171e-01
-1.76289722e-01 -2.96393096e-01 -7.05811083e-02 -7.78710961e-01
-2.85142571e-01 -3.07741821e-01 -1.29679099e-01 9.26838040e-01
-5.69671243e-02 -7.94431031e-01 1.04865837e+00 8.59825492e-01
4.72773254e-01 -2.65201807e-01 -1.02384436e+00 -9.57187772e-01
-5.74664533e-01 -4.24484164e-01 4.06001478e-01 4.47715610e-01
4.64127183e-01 1.99861065e-01 -2.23051354e-01 1.09442137e-02
4.67739284e-01 1.05528925e-02 7.11563587e-01 -1.51866925e+00
6.50760010e-02 -6.67463064e-01 -1.12198830e+00 -1.14340401e+00
-3.06535333e-01 -4.95728761e-01 7.32993707e-02 -1.62567568e+00
-3.98586020e-02 -2.75840402e-01 -1.80312157e-01 4.97629344e-01
2.89035141e-01 4.71952111e-01 3.05439889e-01 1.09331816e-01
-1.21361054e-01 2.09288150e-02 9.47891951e-01 -3.23989242e-01
-4.64882314e-01 1.75306767e-01 3.66652638e-01 6.19783163e-01
8.88800621e-01 -7.57508278e-02 1.89098701e-01 -2.47046590e-01
-2.26928100e-01 -2.61194706e-02 4.73316431e-01 -1.15570188e+00
5.79376400e-01 -1.87566802e-01 5.75878143e-01 -5.23508370e-01
3.49685967e-01 -1.31895375e+00 8.87419581e-02 1.08690369e+00
1.77141145e-01 -2.13786930e-01 -7.34312162e-02 1.84794500e-01
-3.83353345e-02 -2.92841285e-01 5.63966274e-01 -1.02013029e-01
-1.29496229e+00 -4.51143309e-02 -6.87157929e-01 -9.02127922e-01
1.19079542e+00 -8.68444383e-01 -2.18865395e-01 -2.41451547e-03
-4.96954769e-01 -1.93714365e-01 1.68121517e-01 8.00380409e-01
9.31957245e-01 -1.29569435e+00 -1.13588728e-01 4.51263338e-01
9.78508741e-02 -3.00600916e-01 -2.98051745e-01 5.05902648e-01
-8.34195971e-01 8.23953032e-01 -8.08952510e-01 -7.14256048e-01
-1.71626878e+00 4.01943028e-01 1.44882232e-01 4.03195500e-01
-7.49966085e-01 4.04776752e-01 -6.29539669e-01 -2.63659537e-01
9.63194191e-01 -4.29128140e-01 -5.35292149e-01 -1.83484897e-01
8.01524878e-01 7.73990154e-01 1.79300740e-01 -5.53808987e-01
-7.18527913e-01 1.08148134e+00 3.63978595e-01 -3.29702765e-01
1.02856469e+00 3.07786942e-01 2.41910979e-01 4.19759631e-01
7.20244765e-01 -3.43288988e-01 -9.80638444e-01 4.71771462e-03
1.40002057e-01 -6.01292014e-01 -2.60040462e-02 -1.02793920e+00
-1.05417824e+00 1.20310497e+00 1.43125486e+00 3.38875726e-02
7.75484920e-01 -2.38461122e-01 5.07731438e-01 9.16553199e-01
9.66844678e-01 -1.12295330e+00 -1.84678301e-01 5.54918408e-01
9.56923544e-01 -1.43852603e+00 -1.80571210e-02 -2.51102716e-01
-3.51191580e-01 1.52537632e+00 5.77984929e-01 -1.65763438e-01
8.09575140e-01 4.73390847e-01 5.52775443e-01 4.72828373e-02
-3.71440537e-02 -6.41541779e-01 2.84269005e-01 9.62788463e-01
6.06870353e-01 3.67825866e-01 -7.40086854e-01 3.22284192e-01
1.75824746e-01 5.83511174e-01 1.68351322e-01 1.13373482e+00
-5.42482495e-01 -1.04663265e+00 -7.06172645e-01 4.25109565e-01
-3.13845932e-01 4.06463563e-01 -4.13134754e-01 7.68608510e-01
6.75498275e-03 1.00224435e+00 1.34163186e-01 -8.21472108e-01
5.05319834e-01 2.59775162e-01 7.86181629e-01 -3.32930982e-01
-5.93344688e-01 -1.08732790e-01 -3.77196193e-01 -5.10534644e-01
-4.77725804e-01 -4.73313749e-01 -1.47241342e+00 -1.19298622e-01
-4.83693898e-01 -6.13750339e-01 1.25913942e+00 9.71658647e-01
1.67335168e-01 2.64951587e-01 3.44060242e-01 -1.21629691e+00
-4.96494621e-01 -1.23376846e+00 -7.62945950e-01 7.17269853e-02
1.06176518e-01 -8.33309531e-01 1.64828315e-01 -1.44899026e-01] | [6.494182586669922, -0.2504022717475891] |
c21d182d-22d9-4fd3-8a8d-bf07669256f1 | affordances-in-grounded-language-learning | null | null | https://aclanthology.org/W18-2806 | https://aclanthology.org/W18-2806.pdf | Affordances in Grounded Language Learning | We present a novel methodology involving mappings between different modes of semantic representation. We propose distributional semantic models as a mechanism for representing the kind of world knowledge inherent in the system of abstract symbols characteristic of a sophisticated community of language users. Then, motivated by insight from ecological psychology, we describe a model approximating affordances, by which we mean a language learner{'}s direct perception of opportunities for action in an environment. We present a preliminary experiment involving mapping between these two representational modalities, and propose that our methodology can become the basis for a cognitively inspired model of grounded language learning. | ['Kyungtae Lim', 'Stephen McGregor'] | 2018-07-01 | null | null | null | ws-2018-7 | ['grounded-language-learning'] | ['natural-language-processing'] | [-1.41592026e-02 3.13579649e-01 -8.19685087e-02 -3.27281892e-01
2.42398754e-01 -6.85219407e-01 1.22327042e+00 4.70876217e-01
-5.22158384e-01 1.55028775e-01 7.57773519e-01 -6.85261846e-01
-5.73009670e-01 -1.24048698e+00 -6.08806133e-01 -9.75348353e-02
-4.18532528e-02 1.54450119e-01 2.48143658e-01 -8.03887904e-01
9.00613666e-01 7.61570036e-01 -1.94087493e+00 3.22736681e-01
9.57207084e-01 2.69839376e-01 7.66653419e-01 1.62148803e-01
-7.07199752e-01 1.24195337e+00 -3.81512463e-01 1.56369671e-01
-7.21195936e-02 -5.53089559e-01 -1.08564901e+00 -1.96125478e-01
3.56571674e-01 -2.76511580e-01 -5.34386873e-01 1.19295359e+00
1.40971199e-01 6.81900978e-01 6.86292112e-01 -7.97972202e-01
-1.68877661e+00 7.79486656e-01 2.60542214e-01 3.29370648e-01
1.12816644e+00 -1.64546803e-01 6.71762347e-01 -8.32865059e-01
8.49033594e-01 1.61857820e+00 4.32190388e-01 7.43749201e-01
-1.31798697e+00 -1.67793214e-01 4.21539694e-01 2.45705321e-01
-1.40596735e+00 -4.95111674e-01 7.18842089e-01 -7.27483451e-01
1.12476254e+00 1.82240561e-01 1.14169192e+00 7.87873387e-01
-3.67902480e-02 5.14419079e-01 1.31329870e+00 -1.08067989e+00
5.28934658e-01 5.91554284e-01 3.09933543e-01 6.71828985e-01
3.03324521e-01 4.39247876e-01 -1.11373723e+00 -1.48147807e-01
1.27346492e+00 6.72431663e-02 -1.23400234e-01 -6.81657612e-01
-1.21420002e+00 6.17330194e-01 7.28616476e-01 1.01855302e+00
-2.58037239e-01 2.73900390e-01 -1.29402444e-01 3.34784925e-01
1.16088048e-01 8.95727158e-01 1.08257726e-01 -3.31990309e-02
-3.11460972e-01 2.90161014e-01 6.39676630e-01 9.06097054e-01
5.70352137e-01 -5.34058735e-02 6.78112730e-02 8.65419686e-01
1.05167544e+00 5.17319798e-01 7.17095554e-01 -1.06824517e+00
-2.37727150e-01 7.01619148e-01 6.48873672e-02 -8.94106507e-01
-7.87755176e-02 -6.97912201e-02 4.21976477e-01 2.45567992e-01
1.32501200e-01 5.12911975e-01 -7.55024076e-01 1.92382598e+00
7.54577443e-02 -2.57035065e-02 2.53017098e-01 8.60581219e-01
6.78580523e-01 4.29964781e-01 7.30294585e-01 7.86059424e-02
1.34547877e+00 -3.02442014e-01 -5.79752982e-01 -1.49268985e-01
8.46072018e-01 -1.76373914e-01 1.49890196e+00 4.90596890e-02
-9.80311275e-01 -4.13856626e-01 -9.22570765e-01 -3.44112486e-01
-1.00725007e+00 -4.99204993e-01 1.14744985e+00 7.00459182e-01
-1.64891887e+00 5.31869352e-01 -5.14470100e-01 -1.09589982e+00
2.83518106e-01 -2.08856449e-01 -1.48827329e-01 1.65077433e-01
-1.23169971e+00 1.34843111e+00 6.35035276e-01 -4.47526015e-02
-8.84363174e-01 -4.53987449e-01 -1.01619697e+00 -1.00229569e-01
3.95868830e-02 -9.33686435e-01 1.17550397e+00 -9.70559359e-01
-1.12990439e+00 1.31667209e+00 -1.21116318e-01 7.55537003e-02
-6.81833252e-02 -2.57210642e-01 -5.66767991e-01 9.40052196e-02
2.38988891e-01 5.19930661e-01 2.36159548e-01 -1.66675031e+00
-1.88594654e-01 -5.58897138e-01 6.62730992e-01 6.39914513e-01
-3.52267861e-01 1.19175442e-01 3.13442975e-01 -6.90358222e-01
4.52919126e-01 -5.14826953e-01 -2.69275218e-01 3.04782182e-01
1.92514434e-01 -5.48117757e-01 1.18970245e-01 -2.79397994e-01
1.18497324e+00 -2.34629750e+00 6.14620522e-02 4.73788232e-01
2.89181650e-01 -4.38940600e-02 -1.61440715e-01 1.13439071e+00
-4.87549379e-02 5.09110868e-01 1.01473674e-01 2.39977226e-01
4.03060138e-01 3.50204855e-01 -5.75302303e-01 1.53641582e-01
-2.84480810e-01 7.15501964e-01 -1.35660875e+00 -3.83425921e-01
3.81780118e-01 3.64281267e-01 -4.98319924e-01 3.47154498e-01
-1.97720796e-01 2.38225281e-01 -8.18591535e-01 3.37087333e-01
2.57702857e-01 9.34835672e-02 2.64762729e-01 4.67476755e-01
-4.97191727e-01 7.10188091e-01 -1.07064450e+00 2.33673811e+00
-4.94075477e-01 5.28588414e-01 -4.63784248e-01 -7.40141034e-01
7.28837967e-01 2.58421093e-01 -3.25917542e-01 -7.64365852e-01
-3.20244133e-02 1.74499288e-01 1.25657678e-01 -7.50333011e-01
4.53522146e-01 -7.61183679e-01 6.30502105e-02 6.33462906e-01
3.30919810e-02 -4.54837739e-01 4.40495042e-03 4.69454408e-01
6.06995523e-01 6.48500204e-01 5.02080619e-01 -1.06952989e+00
3.37510616e-01 -2.84974929e-02 -7.47703835e-02 8.80225122e-01
-6.42232299e-02 -1.43269002e-01 2.49906182e-02 -5.69594681e-01
-6.23108149e-01 -1.55514729e+00 -4.58608150e-01 1.57104385e+00
4.47868019e-01 -3.34644794e-01 -6.98736012e-01 2.59978563e-01
-1.44608200e-01 1.45312989e+00 -7.05443203e-01 -3.52623403e-01
-2.25456908e-01 -1.65192001e-02 3.70366782e-01 5.89156866e-01
7.25790635e-02 -1.50729322e+00 -1.23048973e+00 1.00274511e-01
2.39117637e-01 -3.59327823e-01 1.64203942e-01 1.02423914e-01
-9.99918699e-01 -5.75645983e-01 -8.63532647e-02 -9.38872278e-01
5.98513663e-01 5.11891305e-01 1.35482013e+00 6.54670238e-01
-1.81879759e-01 9.71285224e-01 -4.72221553e-01 -6.86108768e-01
-4.17315245e-01 -6.35835946e-01 1.74451604e-01 -5.99980354e-01
7.34215677e-01 -8.27123880e-01 -3.32544237e-01 -4.45786864e-01
-1.21363378e+00 1.10645071e-01 4.41505536e-02 1.58156380e-01
1.55306742e-01 -3.49601299e-01 4.27204072e-01 -8.24768543e-01
1.01131332e+00 -8.05452466e-01 -1.55762196e-01 3.91007155e-01
-2.94722587e-01 2.65697479e-01 2.22461671e-02 -3.41229945e-01
-1.47646248e+00 -5.56956589e-01 9.25591886e-02 1.20565221e-01
-5.78509152e-01 7.12775588e-01 -1.46160915e-01 -9.38045457e-02
1.05417931e+00 5.65726042e-01 2.78513525e-02 -5.68937421e-01
9.37652051e-01 5.05421519e-01 2.23531559e-01 -1.34476769e+00
5.74738562e-01 3.66775572e-01 -1.23548202e-01 -1.19898820e+00
-7.03495800e-01 -3.87186378e-01 -7.60290682e-01 -1.71584025e-01
8.95440221e-01 -8.87269914e-01 -7.03425586e-01 5.28083295e-02
-1.07862997e+00 -3.02492082e-01 -8.55799615e-01 5.68176508e-01
-8.89612675e-01 1.50881514e-01 -2.26270258e-01 -9.50426996e-01
4.45337832e-01 -4.57872242e-01 9.03233528e-01 1.93709135e-01
-5.35197616e-01 -1.56541276e+00 3.61425877e-01 -1.26757577e-01
5.15240252e-01 8.84861201e-02 1.18639028e+00 -8.25764120e-01
-6.20467544e-01 3.00299644e-01 4.71770428e-02 -1.39720336e-01
1.24993816e-01 -2.93972403e-01 -9.62853849e-01 1.91391587e-01
3.22928041e-01 -3.72491717e-01 3.92577201e-01 -1.52285799e-01
1.17330217e+00 -7.86117017e-02 -3.63146007e-01 2.85193443e-01
1.86135662e+00 6.09661698e-01 6.03826880e-01 4.90508378e-01
5.25736272e-01 8.49803388e-01 3.38426918e-01 1.98516980e-01
6.79934442e-01 3.54946166e-01 -1.28753111e-01 2.16194585e-01
2.99015995e-02 -8.10129166e-01 1.41252100e-01 6.73218787e-01
-4.26337302e-01 1.10597484e-01 -1.16041601e+00 7.08749950e-01
-1.58600628e+00 -1.30661595e+00 1.42793283e-01 2.11180592e+00
9.46996331e-01 -4.20789808e-01 -1.36531219e-01 -3.56372982e-01
3.62553567e-01 2.04804540e-01 -3.44381154e-01 -8.19925666e-01
1.35971650e-01 3.00258756e-01 -1.80420801e-01 7.08129883e-01
-3.65104318e-01 1.17044806e+00 8.13684845e+00 3.89326066e-01
-5.64284742e-01 3.40847857e-02 -1.51023522e-01 3.18075329e-01
-1.23335540e+00 1.61603853e-01 -8.60656798e-02 4.97904532e-02
1.05727112e+00 -4.42292899e-01 7.45532870e-01 5.17335534e-01
8.41111541e-02 -1.67661384e-01 -1.41914713e+00 5.60605705e-01
1.93173259e-01 -1.29455566e+00 7.40163624e-01 9.34101343e-02
2.68697709e-01 -3.05011451e-01 -1.14682429e-01 1.21190637e-01
6.23452544e-01 -1.22974205e+00 1.25946379e+00 8.66666496e-01
4.85300153e-01 -1.28278866e-01 -5.64698540e-02 6.00806832e-01
-1.03478956e+00 -2.02094361e-01 -3.93051863e-01 -6.91832066e-01
-1.16129331e-01 -2.85136074e-01 -2.01497868e-01 2.33624533e-01
5.27284443e-01 4.71029907e-01 -4.71870154e-01 6.59619272e-01
-2.79931277e-01 1.74802959e-01 -1.35666564e-01 -4.19018894e-01
-2.14697029e-02 -3.23355973e-01 4.34825659e-01 9.63471889e-01
5.13275027e-01 6.50654137e-01 1.32640406e-01 1.31180227e+00
4.28043723e-01 3.53533328e-01 -1.35329342e+00 -1.80605590e-01
1.02830005e+00 4.93937761e-01 -7.74707317e-01 -5.07477283e-01
-4.10853028e-01 6.26842856e-01 4.90445435e-01 4.51013744e-01
-1.91535532e-01 -8.59853774e-02 7.40022242e-01 2.80390143e-01
-2.85806090e-01 -3.72285247e-01 -3.66565734e-01 -1.37569237e+00
-3.54433924e-01 -4.98825967e-01 -5.83570302e-02 -1.21575749e+00
-1.08073008e+00 1.28452286e-01 4.46823925e-01 -7.22500205e-01
-1.49531975e-01 -9.27142084e-01 -5.01967907e-01 1.10571027e+00
-1.03795338e+00 -1.13266540e+00 -6.65159374e-02 6.18107617e-01
4.25172925e-01 -9.88343954e-02 1.27251184e+00 -4.62499082e-01
2.21938759e-01 -2.02909812e-01 -2.84182906e-01 -3.27419043e-01
1.04444548e-02 -1.51075077e+00 4.76482928e-01 6.35337591e-01
5.14490306e-01 1.62099731e+00 8.43805313e-01 -4.59595054e-01
-1.47614527e+00 -2.21105620e-01 1.12192106e+00 -1.07898748e+00
8.10866177e-01 -3.78790885e-01 -1.15606332e+00 9.73444402e-01
3.79863501e-01 -2.94839263e-01 1.29530966e+00 2.87039489e-01
-5.14470279e-01 4.56499189e-01 -1.16834903e+00 7.74892211e-01
1.96865129e+00 -1.21955025e+00 -1.78396094e+00 2.12800547e-01
8.67672384e-01 -7.67335668e-02 -7.29486108e-01 -2.93699890e-01
5.28551102e-01 -9.59838629e-01 1.19760716e+00 -7.46667862e-01
9.03685093e-02 -4.77236390e-01 -4.76766109e-01 -1.42273009e+00
-8.11456323e-01 -1.76539034e-01 2.38517821e-01 9.97912586e-01
-6.48078844e-02 -9.49832320e-01 4.70387824e-02 6.95122242e-01
-3.05249304e-01 -8.10607150e-02 -7.95881510e-01 -5.39702713e-01
5.53485155e-01 -6.11567378e-01 8.66637528e-01 1.16397357e+00
7.12345779e-01 6.88037574e-02 6.47778094e-01 -1.51059464e-01
4.80631948e-01 -1.03813425e-01 5.48520796e-02 -1.70724046e+00
-1.53200299e-01 -4.73415285e-01 -4.46090013e-01 -9.37406480e-01
3.80227417e-01 -1.20526016e+00 2.31631752e-02 -1.80742860e+00
1.42137527e-01 -4.63404745e-01 -5.42945921e-01 2.54729569e-01
2.59196281e-01 -1.48996606e-01 5.21884799e-01 2.81616807e-01
-2.89778709e-01 4.44174409e-01 1.13599217e+00 4.56598818e-01
-3.07460397e-01 -8.74590576e-01 -1.44733202e+00 1.07164812e+00
5.57158709e-01 -1.50197357e-01 -1.09917951e+00 -6.93375587e-01
7.15803027e-01 -1.00786567e-01 6.66923165e-01 -7.73931444e-01
2.79804710e-02 -8.43084157e-01 3.16487730e-01 9.96071547e-02
4.09906637e-03 -1.15292156e+00 1.71136558e-01 4.41667199e-01
-6.49730504e-01 -5.17918356e-02 3.42139453e-01 4.92543191e-01
7.64542399e-03 -3.62103701e-01 2.09092185e-01 -4.77080107e-01
-1.43370759e+00 -4.13296044e-01 -8.11391473e-01 -6.30674660e-02
9.43933070e-01 -6.15912616e-01 -4.33617473e-01 1.41375735e-01
-8.99132609e-01 -1.89932302e-01 8.77847850e-01 7.21788168e-01
8.18804443e-01 -1.42657340e+00 -7.56313726e-02 3.92279804e-01
4.96755838e-01 -4.54896361e-01 -4.98380214e-02 -4.25242186e-02
-8.22502255e-01 3.14303130e-01 -4.88106519e-01 -8.74412581e-02
-4.63155627e-01 7.01581120e-01 5.29652655e-01 9.14116383e-01
-6.15033388e-01 9.31888640e-01 5.14442503e-01 -5.75326085e-01
-2.71341443e-01 -3.21986735e-01 -6.39396131e-01 6.98649604e-03
5.73730171e-01 1.77284092e-01 -6.26586199e-01 -8.51985037e-01
-5.60498238e-01 6.11073196e-01 4.09363598e-01 -4.97324854e-01
1.09168863e+00 -3.41782898e-01 -6.24252200e-01 1.28841591e+00
6.38860762e-01 -4.72777244e-03 -7.14985430e-01 -2.95506328e-01
1.56416491e-01 -8.71232092e-01 -2.32874542e-01 -8.24250340e-01
1.93047468e-02 8.95069003e-01 7.21930146e-01 4.61167097e-01
8.67761075e-01 5.62171638e-01 -2.19854310e-01 7.15943158e-01
7.98518896e-01 -9.88210082e-01 -6.18287399e-02 4.07153994e-01
1.17593312e+00 -6.46410167e-01 -1.87328815e-01 -3.76428038e-01
-3.18448991e-01 1.03126717e+00 5.11568248e-01 -2.58236200e-01
8.73446703e-01 -1.49189830e-01 1.57807201e-01 -5.73068321e-01
-7.63742387e-01 -4.22093987e-01 2.77611554e-01 1.14680111e+00
1.01505673e+00 3.55727762e-01 -5.38959026e-01 7.81150609e-02
-5.65343976e-01 2.19308496e-01 5.89000821e-01 1.23939657e+00
-9.91636753e-01 -8.87499452e-01 -2.66954690e-01 6.83295205e-02
7.66763696e-03 -2.28449360e-01 -3.95114481e-01 7.33345389e-01
3.36695760e-01 8.55569839e-01 2.60263920e-01 3.68657410e-02
5.11820853e-01 3.82925063e-01 8.65498543e-01 -1.21118009e+00
-1.91417783e-01 -4.22082394e-01 -1.46601181e-02 -7.90783167e-01
-8.53858888e-01 -6.90006554e-01 -1.70965350e+00 -1.68426841e-01
2.43414700e-01 2.00188011e-01 5.00379503e-01 1.07024419e+00
2.17253622e-02 2.94436276e-01 1.12080805e-01 -6.42685056e-01
-3.98502648e-01 -8.23573291e-01 -8.98524821e-01 5.80603480e-01
1.60872236e-01 -8.50907445e-01 -1.72250837e-01 2.54516214e-01] | [9.350427627563477, 6.896485328674316] |
40950347-708a-4e77-8812-790c05460826 | pix2struct-screenshot-parsing-as-pretraining | 2210.03347 | null | https://arxiv.org/abs/2210.03347v2 | https://arxiv.org/pdf/2210.03347v2.pdf | Pix2Struct: Screenshot Parsing as Pretraining for Visual Language Understanding | Visually-situated language is ubiquitous -- sources range from textbooks with diagrams to web pages with images and tables, to mobile apps with buttons and forms. Perhaps due to this diversity, previous work has typically relied on domain-specific recipes with limited sharing of the underlying data, model architectures, and objectives. We present Pix2Struct, a pretrained image-to-text model for purely visual language understanding, which can be finetuned on tasks containing visually-situated language. Pix2Struct is pretrained by learning to parse masked screenshots of web pages into simplified HTML. The web, with its richness of visual elements cleanly reflected in the HTML structure, provides a large source of pretraining data well suited to the diversity of downstream tasks. Intuitively, this objective subsumes common pretraining signals such as OCR, language modeling, image captioning. In addition to the novel pretraining strategy, we introduce a variable-resolution input representation and a more flexible integration of language and vision inputs, where language prompts such as questions are rendered directly on top of the input image. For the first time, we show that a single pretrained model can achieve state-of-the-art results in six out of nine tasks across four domains: documents, illustrations, user interfaces, and natural images. | ['Kristina Toutanova', 'Ming-Wei Chang', 'Peter Shaw', 'Urvashi Khandelwal', 'Julian Eisenschlos', 'Fangyu Liu', 'Hexiang Hu', 'Iulia Turc', 'Mandar Joshi', 'Kenton Lee'] | 2022-10-07 | null | null | null | null | ['chart-question-answering', 'chart-question-answering'] | ['computer-code', 'computer-vision'] | [ 5.76329827e-01 2.36634701e-01 -3.45081389e-02 -3.29403311e-01
-7.55927920e-01 -1.00525987e+00 8.94918025e-01 -1.27500847e-01
-1.91230521e-01 2.29837820e-01 4.31405157e-01 -7.47272730e-01
2.18080640e-01 -4.41256166e-01 -1.12282813e+00 -1.46694228e-01
2.98855096e-01 2.21978799e-01 -5.52558117e-02 -1.47821561e-01
2.24573687e-01 2.86514789e-01 -1.58982015e+00 1.14477861e+00
6.71405554e-01 8.86736810e-01 6.83060169e-01 9.57522631e-01
-6.61800742e-01 8.52148414e-01 -4.39737797e-01 -4.88946468e-01
1.08501218e-01 -2.62910515e-01 -7.69141197e-01 3.04646879e-01
9.87272441e-01 -4.81196135e-01 -2.14625210e-01 7.59567082e-01
2.23580673e-01 -1.54256210e-01 6.05569005e-01 -1.15269339e+00
-1.63910103e+00 6.34188771e-01 -5.46020269e-01 -2.52391338e-01
5.87395728e-01 5.99659503e-01 1.12621844e+00 -1.23797560e+00
7.63524175e-01 1.54207110e+00 5.11995673e-01 7.82291234e-01
-1.61211336e+00 -2.00504094e-01 5.00791311e-01 -1.47729844e-01
-9.89687562e-01 -3.38613153e-01 4.78055745e-01 -7.05118358e-01
1.07799518e+00 2.99140900e-01 3.40828031e-01 1.64467847e+00
-2.04687938e-01 1.05015361e+00 1.25441706e+00 -7.14393675e-01
-1.39156446e-01 5.60584903e-01 1.37382410e-02 8.83936107e-01
6.03765696e-02 -5.13971485e-02 -6.08239472e-01 3.20880502e-01
1.00661099e+00 1.07859552e-01 -2.94464052e-01 -4.71386790e-01
-1.17495286e+00 3.68572563e-01 4.09898847e-01 -3.63861881e-02
2.95021515e-02 1.44539475e-01 9.39132422e-02 1.45004451e-01
1.72166094e-01 5.32477856e-01 -5.07318497e-01 1.49132814e-02
-7.42035270e-01 2.52955873e-03 7.35889256e-01 1.16413915e+00
7.35715091e-01 3.65090966e-02 -3.15650284e-01 8.79986703e-01
3.76323551e-01 6.74920857e-01 2.89669067e-01 -9.09038544e-01
8.79353106e-01 6.60711765e-01 3.13977860e-02 -6.54006302e-01
-2.50060350e-01 -2.19986737e-01 -5.56748450e-01 6.15575731e-01
6.25914752e-01 -3.31314132e-02 -1.38012588e+00 1.68320394e+00
-3.06819052e-01 -5.10245085e-01 -4.15317295e-03 8.15225899e-01
1.06099045e+00 6.98331952e-01 3.95770401e-01 4.79239732e-01
1.66746855e+00 -1.16619992e+00 -4.74571496e-01 -7.38467813e-01
2.88388759e-01 -7.22534537e-01 2.15185332e+00 5.16909480e-01
-1.18727827e+00 -7.21996188e-01 -8.92793119e-01 -8.76502216e-01
-8.42360675e-01 3.57195407e-01 4.22689319e-01 3.19350272e-01
-1.38026524e+00 3.97796810e-01 -4.68534082e-01 -6.33164167e-01
4.81996983e-01 -1.23422347e-01 -5.22950649e-01 -2.02962145e-01
-7.09918320e-01 8.14057469e-01 1.03380000e-02 -1.25521734e-01
-6.95483744e-01 -9.11580920e-01 -9.85974252e-01 2.98164696e-01
5.28748393e-01 -9.60185766e-01 1.52952683e+00 -1.28686404e+00
-1.41007924e+00 1.20417809e+00 -4.54618484e-02 -2.42460042e-01
5.52337110e-01 -4.30211633e-01 -1.82071134e-01 1.53136209e-01
-5.59641048e-02 1.15575683e+00 9.74515438e-01 -1.62306297e+00
-3.01935077e-01 -2.07119152e-01 2.18228579e-01 2.21648380e-01
-3.41538757e-01 -1.28811955e-01 -9.38737631e-01 -6.68971658e-01
-3.72836828e-01 -6.08945668e-01 1.14099249e-01 3.89013678e-01
-5.62855303e-01 9.82003883e-02 7.89341569e-01 -9.29698408e-01
1.01252842e+00 -2.12846828e+00 1.40631974e-01 -9.14226696e-02
2.34709218e-01 5.69076426e-02 -5.62045932e-01 6.13107264e-01
-1.98419571e-01 5.93275905e-01 -3.60143371e-02 -5.98615110e-01
3.82146597e-01 -3.41783240e-02 -5.23470759e-01 -2.48693660e-01
4.44679469e-01 1.27921474e+00 -5.78830123e-01 -3.85682821e-01
2.85660833e-01 5.57733774e-01 -5.02189696e-01 2.81598836e-01
-6.84911370e-01 -4.09540255e-03 -2.19465722e-03 6.10064089e-01
4.72394347e-01 -7.39954591e-01 5.52844070e-02 -3.06089759e-01
-1.21639967e-01 2.94718832e-01 -7.94490099e-01 1.94768441e+00
-6.96262300e-01 1.00876296e+00 2.87743658e-01 -3.66245955e-01
5.77426732e-01 2.82463636e-02 -4.10308577e-02 -1.05803478e+00
-2.54693091e-01 -2.34414071e-01 -4.62349385e-01 -6.70353234e-01
5.70593536e-01 2.80709356e-01 -2.73947269e-02 5.26887715e-01
8.86317939e-02 -8.05954486e-02 8.73215348e-02 4.63296413e-01
8.94043446e-01 5.21531761e-01 7.86881149e-02 1.11514777e-01
1.29043639e-01 4.86447737e-02 -3.27449501e-01 9.93473232e-01
1.80804878e-01 1.13291991e+00 5.61034143e-01 -2.76588291e-01
-1.23569882e+00 -1.41756761e+00 2.14092776e-01 1.58991206e+00
-1.63765743e-01 -6.12154007e-01 -7.78923810e-01 -5.67489147e-01
9.06678513e-02 7.96891689e-01 -6.10802591e-01 1.72366679e-01
-3.50116223e-01 -7.42272511e-02 3.00030708e-01 6.52484894e-01
2.74675816e-01 -1.30817354e+00 -5.21223545e-01 -2.58736908e-01
1.17594145e-01 -1.31377208e+00 -5.32982171e-01 2.75147736e-01
-7.94481039e-01 -1.05189168e+00 -8.84352207e-01 -1.03326523e+00
6.96724534e-01 3.51904094e-01 1.65019584e+00 -1.30633954e-02
-2.94128954e-01 8.87499034e-01 -6.91893399e-02 -3.25212240e-01
-4.03987139e-01 -4.91108336e-02 -6.02820218e-01 -1.63077220e-01
1.65840030e-01 -2.85666525e-01 -5.84267378e-01 -1.34839267e-01
-1.17240429e+00 8.17468822e-01 1.01232982e+00 6.30265832e-01
4.56430167e-01 -6.56370938e-01 -4.95558567e-02 -1.08386993e+00
8.10957789e-01 -3.01687509e-01 -4.37462062e-01 5.76830685e-01
-3.69840235e-01 3.28310519e-01 6.33304238e-01 -5.31116307e-01
-1.07475078e+00 1.55179307e-01 1.45550385e-01 -4.07075614e-01
-5.01886606e-01 3.80097806e-01 -2.68210143e-01 2.52301246e-01
7.86350191e-01 1.45411491e-01 -1.55081460e-02 -7.95802891e-01
1.09857130e+00 4.85175729e-01 9.37635720e-01 -5.93958974e-01
8.21952820e-01 3.53937119e-01 -4.05338585e-01 -8.92964244e-01
-6.14432991e-01 -1.46919072e-01 -4.95466650e-01 2.38273423e-02
9.53739882e-01 -9.28591967e-01 -6.09680057e-01 2.35061452e-01
-1.33969736e+00 -8.18153083e-01 -2.24590376e-01 -1.42646834e-01
-5.39326966e-01 2.24481836e-01 -5.69967330e-01 -6.47959292e-01
-2.15282887e-01 -9.24988270e-01 1.31452918e+00 2.03675434e-01
-2.29378968e-01 -9.23256636e-01 -3.45758021e-01 2.94756085e-01
3.60692054e-01 1.90034524e-01 1.33421636e+00 -4.74582553e-01
-8.10854256e-01 1.80458531e-01 -7.51596928e-01 2.37752825e-01
8.32534675e-03 1.56572714e-01 -1.13541675e+00 5.39038368e-02
-5.88645756e-01 -6.23398006e-01 9.87496495e-01 1.85712487e-01
1.52928162e+00 -5.13730645e-01 -3.41676980e-01 7.02105701e-01
1.44237161e+00 8.63754284e-03 6.62203968e-01 4.19724673e-01
1.00345778e+00 6.94317639e-01 -1.37708500e-01 5.55568859e-02
5.81814885e-01 5.84622145e-01 3.39505315e-01 -6.31264329e-01
-5.67713320e-01 -7.70777166e-01 4.76553828e-01 2.30637893e-01
4.13504958e-01 -4.51003522e-01 -9.98081625e-01 3.67753327e-01
-1.57574582e+00 -8.85772645e-01 2.02216372e-01 1.96953344e+00
9.46332097e-01 1.52571693e-01 -7.44187087e-03 -3.37110668e-01
4.33034688e-01 1.77397847e-01 -5.79088449e-01 -5.93650460e-01
-2.24462941e-01 2.02363655e-02 3.53979439e-01 6.01877987e-01
-9.57889676e-01 9.66509879e-01 6.84403324e+00 4.56941396e-01
-1.24569619e+00 -2.30216235e-01 6.38768315e-01 -1.53270394e-01
-6.76387489e-01 -2.55866170e-01 -6.99226081e-01 3.36873084e-01
7.37105191e-01 2.51581699e-01 7.22558975e-01 8.18991244e-01
3.83262217e-01 -3.21487756e-03 -1.52957094e+00 1.21418285e+00
1.97373301e-01 -1.54596806e+00 6.00188494e-01 -1.04466811e-01
5.00372589e-01 2.09065173e-02 5.91685593e-01 4.24550653e-01
4.51672912e-01 -1.52091837e+00 1.03651166e+00 5.90084016e-01
1.29928350e+00 -1.06590383e-01 -1.06439479e-01 1.34142995e-01
-8.80540967e-01 -5.49402423e-02 7.33743981e-02 -1.56586677e-01
6.70598894e-02 8.90247002e-02 -9.14218068e-01 3.37143242e-02
7.91977167e-01 5.95046818e-01 -1.09944594e+00 7.96204865e-01
-2.42044821e-01 2.71199495e-01 -9.39530805e-02 -9.20397695e-03
1.82389006e-01 1.16028473e-01 1.74496755e-01 1.65538990e+00
1.35222107e-01 -1.84519798e-01 4.44291271e-02 1.19267035e+00
-2.71436095e-01 1.49439070e-02 -8.75829220e-01 -4.45038289e-01
3.03126365e-01 1.22026014e+00 -4.60562617e-01 -3.03695202e-01
-7.94629753e-01 1.17593563e+00 4.62818503e-01 9.57665861e-01
-5.79551339e-01 -4.41886961e-01 6.33041203e-01 4.61312324e-01
4.39601243e-01 -4.14802641e-01 -6.24841332e-01 -1.10939264e+00
1.08163729e-01 -1.25934088e+00 1.29458174e-01 -1.67201304e+00
-1.20017862e+00 5.91496289e-01 -2.35481411e-02 -8.72822821e-01
-3.31650585e-01 -1.28869712e+00 -4.94190693e-01 1.03815818e+00
-1.40195930e+00 -1.40431535e+00 -5.30988872e-01 5.51330924e-01
7.96873152e-01 -6.23726379e-03 7.27698803e-01 -1.51427286e-02
-2.63286918e-01 4.81246561e-01 4.37270328e-02 3.18097532e-01
8.41452718e-01 -1.60040724e+00 7.74429023e-01 8.24453890e-01
5.23337185e-01 7.34311521e-01 5.42174280e-01 -3.79495412e-01
-1.64060974e+00 -8.19407284e-01 6.35124743e-01 -9.00389850e-01
6.30780041e-01 -9.08668220e-01 -9.00794804e-01 8.01115930e-01
6.81262851e-01 -2.68840611e-01 4.53966409e-01 1.60998985e-01
-9.43074405e-01 1.52019933e-02 -6.16476238e-01 1.16418719e+00
1.12098062e+00 -1.08427608e+00 -6.27383411e-01 3.10944766e-01
9.40957844e-01 -4.61134881e-01 -2.06989467e-01 1.21960661e-03
7.44472802e-01 -8.46235454e-01 1.24038422e+00 -8.84467244e-01
7.95210302e-01 -1.85040444e-01 -4.92991768e-02 -1.10970461e+00
-2.62370974e-01 -7.11380422e-01 -3.21297944e-02 1.29012847e+00
7.54269123e-01 -1.20258428e-01 6.51529253e-01 6.75540328e-01
-9.27842185e-02 -3.79401267e-01 -2.11225674e-01 -3.81396651e-01
7.18746260e-02 -4.79394376e-01 2.88176745e-01 4.91562873e-01
-4.40016128e-02 6.43550158e-01 -1.84995860e-01 -6.49205819e-02
3.78957659e-01 2.25121379e-02 9.04135466e-01 -9.89979029e-01
-4.72808212e-01 -7.77393460e-01 2.82463461e-01 -1.48712707e+00
-1.37440369e-01 -8.95559549e-01 5.49609140e-02 -1.91710639e+00
3.24924827e-01 1.37544228e-02 1.03826329e-01 5.81951499e-01
-1.14498697e-01 1.36981741e-01 6.46219194e-01 3.52686457e-02
-6.52542949e-01 7.24723272e-04 1.40227151e+00 -3.73194665e-01
-2.71260858e-01 -4.31631893e-01 -1.18272829e+00 7.96644568e-01
4.43373710e-01 1.30060032e-01 -6.30448818e-01 -1.14067280e+00
3.19003016e-01 -1.88866436e-01 7.43526340e-01 -7.83611357e-01
1.11777864e-01 -1.47952199e-01 7.21378088e-01 -4.38014925e-01
3.39798778e-01 -7.44811237e-01 -1.19481087e-01 -4.97526489e-02
-8.43116462e-01 3.47891062e-01 5.51110685e-01 5.92359006e-01
-1.99666638e-02 8.50810632e-02 3.86186451e-01 -4.39769566e-01
-9.84239221e-01 -8.59524757e-02 -2.92959213e-01 2.44721547e-01
4.32842940e-01 -3.85123610e-01 -7.94997752e-01 -5.72303355e-01
-8.18074584e-01 1.30319387e-01 6.60528660e-01 8.48465800e-01
6.43854260e-01 -1.12861204e+00 -3.83505434e-01 2.35799074e-01
2.97433287e-01 -2.07467973e-02 9.36871916e-02 2.65340537e-01
-4.39203858e-01 6.52752697e-01 -2.46236503e-01 -7.13464677e-01
-1.19609499e+00 8.15884352e-01 9.63716432e-02 -9.57056284e-02
-7.10793734e-01 6.46199405e-01 8.74792159e-01 -2.93453991e-01
5.49319744e-01 -8.62790823e-01 1.41831525e-02 -9.17083099e-02
6.76460207e-01 -2.51586467e-01 -7.20574260e-02 -1.52224675e-01
-1.17909573e-01 6.34791136e-01 7.64971748e-02 -3.46446693e-01
1.11782789e+00 -4.46367353e-01 2.98970670e-01 4.42209840e-01
1.11505866e+00 9.57215130e-02 -1.73793340e+00 -1.72849819e-01
1.40192121e-01 -1.42784566e-01 -3.06879520e-01 -1.51587737e+00
-6.30111516e-01 1.34626293e+00 3.18335325e-01 1.47300780e-01
1.04666829e+00 1.42421618e-01 3.96359146e-01 4.50095475e-01
-1.02692842e-01 -9.26978886e-01 5.65429568e-01 5.09385943e-01
1.29071546e+00 -1.39346015e+00 -3.34880173e-01 -2.43455291e-01
-9.16174412e-01 1.14648235e+00 7.57821620e-01 2.56698489e-01
1.26739174e-01 4.93436426e-01 2.89910376e-01 -6.34899770e-04
-9.93527532e-01 -3.51065695e-01 7.05643773e-01 9.61554229e-01
5.59498608e-01 -2.19886199e-01 4.80516523e-01 8.41066420e-01
-1.66506395e-01 -1.78280950e-01 3.65559608e-01 5.89414239e-01
-4.34835315e-01 -8.71795714e-01 -3.78208458e-01 1.61363870e-01
-2.18732625e-01 -6.33380771e-01 -7.41529942e-01 1.00140846e+00
-9.57170576e-02 6.53703809e-01 2.26487383e-01 -1.58186004e-01
4.06084985e-01 3.54285121e-01 4.54080820e-01 -7.81588197e-01
-4.66197819e-01 -3.66269164e-02 8.38404447e-02 -6.84429705e-01
1.18162349e-01 -2.71908283e-01 -9.95051563e-01 -5.14622740e-02
4.82315242e-01 -3.87102127e-01 7.46228993e-01 6.95325255e-01
6.51496351e-01 6.10067904e-01 -1.47625014e-01 -1.04207218e+00
-3.96212399e-01 -7.23587871e-01 -1.60562202e-01 7.25746155e-01
5.61932743e-01 -2.69676298e-01 -1.12371221e-01 5.61133802e-01] | [11.009029388427734, 1.8727617263793945] |
b2d27483-cb58-42b9-b713-91c71bd8630c | multi-view-non-negative-matrix-factorization | 2201.04726 | null | https://arxiv.org/abs/2201.04726v1 | https://arxiv.org/pdf/2201.04726v1.pdf | Multi-View Non-negative Matrix Factorization Discriminant Learning via Cross Entropy Loss | Multi-view learning accomplishes the task objectives of classification by leverag-ing the relationships between different views of the same object. Most existing methods usually focus on consistency and complementarity between multiple views. But not all of this information is useful for classification tasks. Instead, it is the specific discriminating information that plays an important role. Zhong Zhang et al. explore the discriminative and non-discriminative information exist-ing in common and view-specific parts among different views via joint non-negative matrix factorization. In this paper, we improve this algorithm on this ba-sis by using the cross entropy loss function to constrain the objective function better. At last, we implement better classification effect than original on the same data sets and show its superiority over many state-of-the-art algorithms. | ['Xionglin Luo', 'Run-kun Lu', 'Yuan-Fang Wang', 'Jian-wei Liu'] | 2022-01-08 | null | null | null | null | ['multi-view-learning'] | ['computer-vision'] | [-3.67256939e-01 -5.60250878e-01 -4.96177912e-01 -4.51054007e-01
-4.23528939e-01 -5.73216498e-01 2.88919121e-01 -2.88094670e-01
-9.80830044e-02 4.79686975e-01 4.18620378e-01 2.68896490e-01
-2.63779402e-01 -5.88571012e-01 -2.76642829e-01 -8.21734488e-01
2.20416874e-01 2.68898934e-01 1.52870387e-01 -1.98277518e-01
2.19117537e-01 2.55589426e-01 -1.39630246e+00 7.79590487e-01
7.11137831e-01 9.24498737e-01 3.36976737e-01 1.18311919e-01
1.54798955e-01 6.13488317e-01 -1.45542562e-01 -4.69755352e-01
4.29792374e-01 -4.79371369e-01 -4.69561517e-01 4.03588206e-01
4.58438814e-01 -1.34758338e-01 -3.01899850e-01 1.20153737e+00
4.12672162e-01 2.06955262e-02 6.56274617e-01 -1.21848857e+00
-8.82696986e-01 2.55657911e-01 -1.09401751e+00 3.30243886e-01
1.57464325e-01 -3.44500571e-01 1.31118083e+00 -1.04869986e+00
4.71572876e-01 1.27220607e+00 3.63831252e-01 3.03205848e-01
-1.09287775e+00 -5.66249609e-01 4.98724043e-01 6.30139410e-01
-1.07883286e+00 -5.41235544e-02 1.13162553e+00 -5.93906164e-01
3.81723344e-01 4.63841587e-01 7.16251254e-01 9.46196139e-01
4.49644655e-01 1.11557388e+00 1.45247829e+00 -1.96904898e-01
-3.44701976e-01 3.04343432e-01 1.71959147e-01 6.62189364e-01
1.81216955e-01 -9.58879441e-02 -4.42606300e-01 -3.21820267e-02
5.08561075e-01 5.41760862e-01 -4.63097960e-01 -7.38754988e-01
-1.31206715e+00 9.05622065e-01 3.50457877e-01 5.01396537e-01
-2.39823356e-01 -5.49474955e-01 4.65035439e-01 1.98501021e-01
6.00606203e-01 1.82707816e-01 -6.39579654e-01 1.72302276e-01
-5.83574593e-01 -1.76531941e-01 4.52845126e-01 6.64680600e-01
6.62264884e-01 -1.53565168e-01 -2.12820396e-02 1.01864874e+00
4.94435757e-01 3.05464834e-01 4.89863098e-01 -5.96822023e-01
5.98217666e-01 9.37249899e-01 -2.63113767e-01 -1.22021198e+00
-3.06591779e-01 -6.90401018e-01 -1.20282519e+00 1.45174367e-02
8.42784252e-03 1.62356384e-02 -6.84644818e-01 1.52699995e+00
3.78107965e-01 -1.31495252e-01 -2.27028038e-02 1.05645096e+00
9.66098905e-01 5.19144833e-01 -3.34776670e-01 -4.07022804e-01
1.31975436e+00 -1.32000768e+00 -8.58198941e-01 -2.73177326e-01
1.50160044e-01 -9.51505899e-01 7.96953619e-01 5.05676866e-01
-7.44275808e-01 -6.71146154e-01 -1.05796063e+00 1.55312493e-01
-2.20304489e-01 4.90831107e-01 9.43726838e-01 2.79677659e-01
-6.13891721e-01 2.33440235e-01 -5.39876699e-01 -1.10212795e-01
3.38263392e-01 2.17083797e-01 -8.98653626e-01 -3.53179276e-01
-8.57363760e-01 7.40243375e-01 3.77124876e-01 1.53991684e-01
-4.45333034e-01 -4.68904316e-01 -8.45448434e-01 -1.79419234e-01
5.58551431e-01 -6.47305071e-01 6.89983010e-01 -1.10596049e+00
-1.17376494e+00 8.02379072e-01 -1.55363619e-01 2.43318021e-01
5.38957059e-01 -2.81355590e-01 -5.81121624e-01 9.21781138e-02
1.66677356e-01 1.55406818e-01 7.38907874e-01 -1.53416514e+00
-6.31147146e-01 -7.56340683e-01 3.77397835e-01 4.14560914e-01
-4.62288916e-01 -1.43708408e-01 -8.08161974e-01 -7.92569816e-01
5.81038356e-01 -1.05687582e+00 -2.55463459e-02 -1.04738697e-01
-3.46757323e-01 -1.93886712e-01 1.10983896e+00 -7.15119600e-01
1.12998283e+00 -2.16993737e+00 5.13835073e-01 -4.65706475e-02
3.43123138e-01 1.50237307e-01 -8.61418322e-02 4.35514241e-01
-2.81693667e-01 -2.24496145e-02 1.14951059e-01 -7.42078647e-02
-4.57452595e-01 2.08086357e-01 -3.05808708e-02 6.43905818e-01
-1.23004057e-01 7.00514019e-01 -6.86773479e-01 -5.24276197e-01
2.47841269e-01 3.06511760e-01 -5.98554194e-01 3.21726322e-01
2.67387837e-01 6.41667187e-01 -6.66004121e-01 6.31423295e-01
8.22965026e-01 -4.68969792e-01 2.91550308e-01 -7.27925718e-01
1.41641378e-01 -2.49418139e-01 -1.28391683e+00 1.83091521e+00
-4.97576088e-01 2.77475148e-01 1.58267021e-01 -1.39295614e+00
7.05133617e-01 1.92426488e-01 8.43181968e-01 -4.69036043e-01
-6.73510358e-02 -1.15563251e-01 1.48620203e-01 -6.77921176e-01
1.44564435e-01 -3.76887947e-01 1.63446784e-01 3.37865502e-01
1.07221834e-01 2.65609026e-01 4.01296392e-02 1.64097920e-01
2.78962523e-01 1.23259813e-01 6.96479440e-01 -2.15888500e-01
7.79300570e-01 -3.53907853e-01 9.55835640e-01 3.15691680e-01
-9.46884006e-02 8.30875456e-01 3.48629773e-01 -4.06342238e-01
-8.52372944e-01 -7.96081364e-01 -2.32502177e-01 9.15624321e-01
4.39527482e-01 -5.41819811e-01 -1.83252767e-01 -1.23062360e+00
1.13672495e-01 2.24787876e-01 -9.63195622e-01 -1.16825886e-01
-3.39252621e-01 -8.03688347e-01 -3.07999998e-01 6.35506153e-01
7.24601388e-01 -3.25703830e-01 -1.65166140e-01 -1.14230193e-01
-2.87194848e-01 -9.82329369e-01 -6.13295078e-01 3.04647125e-02
-1.00595093e+00 -1.19899905e+00 -6.07829928e-01 -5.32692909e-01
7.80702055e-01 9.75575507e-01 1.16061378e+00 1.56596545e-02
2.05539502e-02 4.33131605e-01 -6.04781985e-01 -2.68904507e-01
1.69102773e-01 -6.32429048e-02 9.29506123e-02 2.85636604e-01
2.57359058e-01 -6.65798545e-01 -5.57298303e-01 7.45843887e-01
-6.73802793e-01 3.57161343e-01 6.90181851e-01 1.13442206e+00
8.45331013e-01 1.68601558e-01 3.27285826e-01 -9.53531206e-01
4.18389551e-02 -4.02578473e-01 -1.99294269e-01 6.73870265e-01
-5.60587049e-01 -5.31129353e-02 6.56186461e-01 -2.17345804e-01
-1.02494836e+00 -8.06865469e-03 1.59498781e-01 -7.79967368e-01
1.46983176e-01 4.59199965e-01 -6.96001828e-01 -1.91418212e-02
6.82791695e-02 3.72337341e-01 3.13221999e-02 -5.61575890e-01
2.70643055e-01 4.49707419e-01 -1.85946152e-01 -2.92423427e-01
6.46326423e-01 7.37389505e-01 1.28200427e-01 -5.03608346e-01
-1.42360151e+00 -7.69669235e-01 -9.34749961e-01 -2.06742302e-01
9.13322330e-01 -1.11732662e+00 -5.42532504e-01 2.58716106e-01
-7.97062039e-01 6.04867578e-01 6.60247430e-02 8.76330793e-01
-2.60440916e-01 5.80877066e-01 -3.78074288e-01 -5.81165612e-01
1.52215259e-02 -1.21824706e+00 1.04593587e+00 1.67635217e-01
4.78111655e-01 -9.77746308e-01 6.89312443e-02 7.88944364e-01
2.41384003e-02 5.19090593e-02 6.06943727e-01 -7.12247670e-01
-4.37849253e-01 -1.46439686e-01 -3.09491783e-01 6.68617129e-01
4.90168452e-01 2.53268685e-02 -9.01118219e-01 -5.88982642e-01
4.60064471e-01 -1.77064523e-01 1.16209435e+00 3.80622059e-01
1.26335371e+00 -1.83088064e-01 -3.90155256e-01 6.90506041e-01
1.58249724e+00 3.66043270e-01 3.51775616e-01 2.47201174e-01
1.11956501e+00 6.69135153e-01 8.10180664e-01 3.37181956e-01
4.36315030e-01 7.17104077e-01 4.16378081e-01 -2.27066558e-02
2.08126426e-01 -8.85901228e-02 2.27696672e-01 1.31249452e+00
-3.32087576e-01 -7.37885982e-02 -6.43219173e-01 3.10001522e-01
-2.15246892e+00 -1.24870372e+00 -1.08003303e-01 1.89811647e+00
3.30676347e-01 -3.65637988e-02 5.02567850e-02 -1.28797248e-01
8.50172281e-01 5.45173526e-01 -4.92700458e-01 5.21170273e-02
-3.22458059e-01 -6.05130136e-01 1.00673258e-01 1.22610420e-01
-1.55035245e+00 4.68535691e-01 5.84926939e+00 9.18220222e-01
-1.11102700e+00 1.58805832e-01 7.94541538e-01 -6.33588955e-02
-4.52803344e-01 -6.56986833e-02 -7.42150009e-01 4.22653019e-01
7.69858882e-02 5.12170978e-02 3.18691581e-01 9.62070227e-01
-1.38614979e-03 -2.07141206e-01 -9.85617697e-01 1.35410738e+00
6.70366287e-01 -1.09686565e+00 1.00148104e-01 1.02015190e-01
8.94035816e-01 -2.31566489e-01 6.68226779e-02 2.44647011e-01
-1.18185259e-01 -8.45478117e-01 5.44426858e-01 5.96988559e-01
4.15551186e-01 -8.01298916e-01 8.27912748e-01 5.28735995e-01
-1.46835399e+00 -8.86442438e-02 -4.31304544e-01 8.25848430e-02
2.68508866e-02 7.75394678e-01 -1.64674893e-01 1.23149204e+00
5.76778948e-01 1.44721043e+00 -7.89150357e-01 8.25228512e-01
-2.08136350e-01 2.55997866e-01 8.10846239e-02 3.25150490e-01
6.72629401e-02 -6.19741976e-01 6.33535385e-01 8.15722108e-01
9.96190310e-02 2.56087720e-01 5.55722237e-01 3.40726048e-01
2.20239177e-01 2.23763019e-01 -6.98902965e-01 2.12359607e-01
-5.44476584e-02 1.50791764e+00 -7.04224288e-01 -2.42826656e-01
-9.01051939e-01 9.79198396e-01 3.47923368e-01 1.36531681e-01
-7.95518816e-01 2.20704988e-01 5.21151602e-01 -2.75429070e-01
5.14987826e-01 -1.61544919e-01 -3.40143412e-01 -1.84586108e+00
3.50764513e-01 -1.06679571e+00 6.79826856e-01 -5.77870309e-01
-1.73187387e+00 4.62196589e-01 -3.47435251e-02 -1.85833669e+00
1.07701711e-01 -6.09493613e-01 -4.37653452e-01 8.27118278e-01
-1.41289127e+00 -1.51503205e+00 -2.08894610e-01 6.33040726e-01
6.91850960e-01 -5.03881216e-01 5.63502729e-01 3.59387308e-01
-7.23900914e-01 3.90008032e-01 4.29060876e-01 1.19366452e-01
8.23490977e-01 -1.22151458e+00 -2.11793363e-01 7.66281724e-01
3.99083823e-01 7.00218379e-01 3.65582198e-01 -6.14856184e-01
-1.54715347e+00 -7.87793875e-01 3.49240839e-01 -4.33236718e-01
4.83044833e-01 -1.19217329e-01 -7.89189041e-01 5.87390661e-01
2.86304086e-01 3.03664058e-01 1.05729342e+00 4.16991323e-01
-5.34820855e-01 -3.33890855e-01 -7.93581903e-01 2.08292246e-01
9.06622946e-01 -6.71542048e-01 -6.06347501e-01 2.21472710e-01
3.89679581e-01 -3.11463296e-01 -7.82179058e-01 5.57061017e-01
7.47491300e-01 -1.40407574e+00 1.01757264e+00 -1.00523472e+00
6.99934185e-01 -3.73922318e-01 -5.71178138e-01 -1.45873284e+00
-6.21925533e-01 5.56951948e-02 -5.04316092e-02 1.38387823e+00
1.75650090e-01 -5.00881910e-01 7.02684164e-01 1.91259414e-01
1.15873784e-01 -1.16464698e+00 -6.31062210e-01 -7.32885957e-01
-1.34505495e-01 -1.18725061e-01 2.80608803e-01 1.30440784e+00
-2.42535412e-01 6.67976916e-01 -7.00753510e-01 8.76522884e-02
6.36881232e-01 8.76240015e-01 6.89334929e-01 -1.25006735e+00
-5.42179406e-01 -5.76998442e-02 -5.72889328e-01 -7.93158531e-01
2.49670386e-01 -9.14959848e-01 -4.50035721e-01 -1.57605922e+00
9.84360635e-01 -1.61403462e-01 -5.51301897e-01 3.08041334e-01
-6.74604952e-01 7.53916502e-02 4.64807570e-01 4.99879956e-01
-8.06185544e-01 7.60931551e-01 1.75628388e+00 -1.20703615e-01
1.57608196e-01 7.99933225e-02 -8.23057353e-01 8.71835947e-01
5.77978373e-01 -1.99366093e-01 -5.69086015e-01 -3.45240980e-01
8.15361366e-02 1.59284219e-01 1.15248255e-01 -7.55454242e-01
-7.30532259e-02 -4.71564770e-01 6.75163627e-01 -7.63341308e-01
3.18144768e-01 -1.16697001e+00 2.38742039e-01 2.80978531e-01
-3.74413095e-02 3.96158546e-01 -1.57043070e-01 9.33747470e-01
-6.90455914e-01 7.49286860e-02 8.76172245e-01 -3.50434810e-01
-8.53730619e-01 4.05502796e-01 1.94232225e-01 7.11076856e-02
1.23403656e+00 -9.21855643e-02 -2.01206669e-01 -4.08110231e-01
-7.31073916e-01 1.52622625e-01 3.87078851e-01 6.04277492e-01
5.25402486e-01 -1.64120018e+00 -6.73795402e-01 2.15371549e-01
2.76560098e-01 -3.17204714e-01 6.56770647e-01 1.03673375e+00
-3.94301824e-02 3.19000393e-01 -4.32563931e-01 -6.71209276e-01
-1.74526036e+00 7.42485225e-01 1.96657911e-01 -3.62924784e-01
-3.56046706e-01 8.81049156e-01 8.55390012e-01 -3.34886760e-01
-1.55423880e-01 1.81530669e-01 -6.48379385e-01 5.22530735e-01
4.50259328e-01 3.75898421e-01 -7.53529370e-02 -9.18986857e-01
-5.12401462e-01 9.07022119e-01 -2.76598662e-01 1.97544962e-01
1.49099922e+00 -1.88483983e-01 -3.79048795e-01 7.42066801e-01
1.50656307e+00 2.62660831e-01 -1.17583036e+00 -2.48941109e-01
-4.15274411e-01 -9.05600905e-01 1.06972463e-01 -7.22038925e-01
-1.67935956e+00 8.85383427e-01 6.22047722e-01 1.94146618e-01
1.19789863e+00 2.41631940e-02 3.70374769e-01 1.48002654e-01
4.39021498e-01 -8.35886598e-01 2.30978861e-01 4.44601625e-01
9.81999815e-01 -1.73870969e+00 4.73812938e-01 -6.15295053e-01
-1.09849966e+00 1.10244036e+00 8.62758577e-01 -1.14720717e-01
8.83231759e-01 -1.28632650e-01 5.33405244e-02 -3.20749491e-01
-7.02898920e-01 -8.10175389e-02 8.03124607e-01 3.35036576e-01
5.47176778e-01 2.26408660e-01 -5.22986472e-01 5.07804513e-01
2.98265100e-01 -2.70822167e-01 7.23777711e-02 7.81595767e-01
-1.10804304e-01 -1.28351259e+00 -3.22263747e-01 7.06235468e-01
-6.77306235e-01 3.41007672e-02 -4.36677575e-01 7.76697278e-01
3.15709263e-01 8.07558477e-01 -2.51458049e-01 -7.86815464e-01
2.02715397e-01 -2.06369951e-01 4.69120026e-01 -5.95999002e-01
-4.37646419e-01 3.85981172e-01 -6.15869351e-02 -5.16761124e-01
-8.43232751e-01 -1.02114975e+00 -5.59464395e-01 -4.39322069e-02
-6.18478358e-01 1.45056203e-01 3.48645955e-01 9.29603696e-01
3.16793859e-01 4.81356263e-01 1.12827802e+00 -5.58350861e-01
-4.66605276e-01 -6.91572547e-01 -8.97718906e-01 5.06591141e-01
3.25836092e-01 -9.38588560e-01 -3.40427727e-01 -1.02954432e-01] | [8.485538482666016, 4.5124192237854] |
24a654af-de1d-49f2-8238-b24788bb5605 | counterfactual-recipe-generation-exploring | 2210.11431 | null | https://arxiv.org/abs/2210.11431v1 | https://arxiv.org/pdf/2210.11431v1.pdf | Counterfactual Recipe Generation: Exploring Compositional Generalization in a Realistic Scenario | People can acquire knowledge in an unsupervised manner by reading, and compose the knowledge to make novel combinations. In this paper, we investigate whether pretrained language models can perform compositional generalization in a realistic setting: recipe generation. We design the counterfactual recipe generation task, which asks models to modify a base recipe according to the change of an ingredient. This task requires compositional generalization at two levels: the surface level of incorporating the new ingredient into the base recipe, and the deeper level of adjusting actions related to the changing ingredient. We collect a large-scale recipe dataset in Chinese for models to learn culinary knowledge, and a subset of action-level fine-grained annotations for evaluation. We finetune pretrained language models on the recipe corpus, and use unsupervised counterfactual generation methods to generate modified recipes. Results show that existing models have difficulties in modifying the ingredients while preserving the original text style, and often miss actions that need to be adjusted. Although pretrained language models can generate fluent recipe texts, they fail to truly learn and use the culinary knowledge in a compositional way. Code and data are available at https://github.com/xxxiaol/counterfactual-recipe-generation. | ['Dongyan Zhao', 'Chengang Hu', 'Jizhi Tang', 'Yansong Feng', 'Xiao Liu'] | 2022-10-20 | null | null | null | null | ['recipe-generation'] | ['miscellaneous'] | [ 4.01004761e-01 3.01621765e-01 -2.52408832e-01 -4.31711972e-01
-5.34358263e-01 -1.05159247e+00 8.11864793e-01 8.71635824e-02
-2.09723473e-01 8.97461772e-01 9.13554311e-01 -2.48893321e-01
3.35040808e-01 -1.23361981e+00 -1.27819717e+00 -3.08589995e-01
3.96019548e-01 4.48713869e-01 -3.28046918e-01 -6.13698304e-01
1.07907735e-01 -1.98015831e-02 -1.17392898e+00 9.36694384e-01
1.12481689e+00 1.80371311e-02 3.27576756e-01 7.14502752e-01
-3.87009084e-01 9.46170092e-01 -4.82930034e-01 -7.67978549e-01
3.92744064e-01 -9.75085795e-01 -7.49041438e-01 7.25041926e-02
4.74243104e-01 -5.33709705e-01 -2.01878071e-01 9.32587802e-01
2.47380272e-01 3.43696982e-01 8.30039024e-01 -8.47397387e-01
-1.63241959e+00 1.81344342e+00 9.79853645e-02 -2.82009602e-01
3.68574888e-01 6.58535838e-01 1.11916733e+00 -5.29720068e-01
4.87057239e-01 1.36895418e+00 6.34275496e-01 1.11791968e+00
-1.59818566e+00 -6.44926965e-01 4.99608666e-01 -2.31152385e-01
-8.76938283e-01 -2.93365568e-01 8.91866267e-01 -4.20340538e-01
9.47464108e-01 8.84510055e-02 7.74367392e-01 1.23001683e+00
-7.95073286e-02 1.04769683e+00 1.19125533e+00 -3.97591710e-01
1.38561308e-01 -7.36775920e-02 -2.02843860e-01 6.85772777e-01
3.67848933e-01 4.12806362e-01 -5.26655316e-01 1.08401723e-01
8.68015528e-01 -7.72768036e-02 -2.38431588e-01 -1.19212702e-01
-1.68767202e+00 9.84214962e-01 5.80655754e-01 3.42918158e-01
-6.59495413e-01 5.83192468e-01 2.76009262e-01 4.07500982e-01
3.83066446e-01 1.42820871e+00 -9.93022382e-01 2.23836049e-01
-6.90218925e-01 7.30202496e-01 8.54587793e-01 9.45146680e-01
6.75946712e-01 1.30826637e-01 -5.61673462e-01 8.40766847e-01
3.99900861e-02 7.68218696e-01 5.93434870e-01 -1.07727766e+00
5.68689167e-01 4.39964294e-01 3.98349434e-01 -5.52256823e-01
-1.14825755e-01 -2.09549610e-02 -7.37689018e-01 -1.43319130e-01
5.21615565e-01 -5.69914579e-01 -7.94386983e-01 2.27481627e+00
2.10250199e-01 1.35296732e-01 1.85540840e-01 6.60682201e-01
7.27166712e-01 7.21224904e-01 5.01604497e-01 5.92908822e-02
9.98650372e-01 -1.15957618e+00 -6.61336243e-01 -1.99943662e-01
9.82314885e-01 -7.01128006e-01 1.43948722e+00 1.43572316e-01
-1.23380578e+00 -7.48276532e-01 -9.46184576e-01 -1.38221130e-01
-6.00964487e-01 2.16604948e-01 9.46156502e-01 4.95242149e-01
-1.01315856e+00 9.71833229e-01 -5.28219581e-01 -7.34841973e-02
1.93067864e-01 -8.39882493e-02 3.37423757e-02 -1.82430655e-01
-1.62630713e+00 8.91954303e-01 9.68629241e-01 -2.67796367e-02
-1.00902438e+00 -1.25254929e+00 -1.31377399e+00 -1.69269934e-01
5.37328422e-01 -1.14206100e+00 1.81066334e+00 -1.39239168e+00
-1.85355306e+00 6.07639968e-01 3.22588235e-01 -2.95529276e-01
3.28074336e-01 -2.22805738e-01 -3.95107597e-01 -3.72163206e-01
3.66177797e-01 1.13653946e+00 5.88073611e-01 -1.20598245e+00
-4.86291170e-01 -9.09985974e-02 5.96823275e-01 4.42487508e-01
1.89050272e-01 -3.34379852e-01 2.39745393e-01 -1.27371311e+00
-4.07655329e-01 -9.55967128e-01 -4.17698115e-01 -3.85846168e-01
-4.54664350e-01 -2.66526520e-01 -1.20699614e-01 -8.23104441e-01
1.14698088e+00 -1.73550940e+00 1.13623656e-01 -3.61162603e-01
-2.05365241e-01 1.75809041e-01 -5.17797470e-01 3.82836550e-01
-1.73335910e-01 5.33530056e-01 -3.91820431e-01 -6.93555847e-02
5.51051557e-01 1.03140481e-01 -4.56575483e-01 -2.08676115e-01
2.55758405e-01 1.38659775e+00 -1.44338095e+00 3.02292034e-02
2.85497725e-01 7.15855928e-03 -1.04837608e+00 2.87874401e-01
-1.16533625e+00 2.41192982e-01 -3.82588059e-01 1.38681978e-01
2.86911964e-01 -3.95083427e-03 2.99548328e-01 -2.97838241e-01
-1.33527979e-01 9.02400196e-01 -1.10400558e+00 2.05341911e+00
-7.35189021e-01 3.25738937e-02 -4.13892657e-01 -5.86025953e-01
5.48438251e-01 3.68171453e-01 -7.98273608e-02 -4.81528521e-01
-2.59977520e-01 1.95729598e-01 3.31484154e-02 -4.53876197e-01
6.65121198e-01 -7.17421651e-01 -5.27930796e-01 9.22990918e-01
-2.72238739e-02 -7.18113422e-01 5.75116456e-01 9.55450460e-02
5.90174437e-01 7.72851646e-01 5.99138975e-01 -3.44760925e-01
2.78063267e-01 3.40425521e-01 2.89171904e-01 7.99512863e-01
2.94597358e-01 1.55395359e-01 9.57050622e-02 -5.40258348e-01
-1.21792877e+00 -1.27698243e+00 5.44388056e-01 1.45786631e+00
-3.01435381e-01 -2.67342031e-01 -7.67275393e-01 -8.18644106e-01
-1.35282222e-02 1.79109442e+00 -7.82353520e-01 -3.90564948e-01
-8.28097403e-01 -7.95604527e-01 5.69168091e-01 5.79798520e-01
8.21811855e-01 -1.44312823e+00 -2.39654690e-01 4.18660253e-01
-4.64901507e-01 -4.86913353e-01 -1.23037529e+00 -2.56895840e-01
-8.37689042e-01 -9.63642657e-01 -4.19627607e-01 -8.35105717e-01
5.29093087e-01 -6.45143315e-02 1.64737415e+00 -2.59675435e-03
2.50070155e-01 2.90378600e-01 -2.47416064e-01 -4.97588098e-01
-1.16127002e+00 -3.81649882e-02 -1.14013232e-01 -3.43681663e-01
4.62308079e-01 -4.19222027e-01 -4.01934236e-01 -6.58682585e-02
-1.06151652e+00 8.61377358e-01 4.80126888e-01 6.95212126e-01
4.46101636e-01 7.67771676e-02 7.75072455e-01 -1.16290689e+00
7.02044785e-01 -4.82127786e-01 -2.33204037e-01 4.22211468e-01
-2.63136119e-01 6.19294465e-01 1.06719327e+00 -7.49591887e-01
-1.50812972e+00 5.73111549e-02 2.08440367e-02 1.49511144e-01
-3.94403398e-01 4.88617092e-01 -2.80539274e-01 9.22768772e-01
8.88160169e-01 3.35554183e-01 -4.04143870e-01 -5.69508433e-01
1.37669361e+00 2.12940767e-01 5.53352535e-01 -1.16883361e+00
6.72444403e-01 7.85617828e-02 -5.23821831e-01 -3.22778046e-01
-1.25516081e+00 1.41385064e-01 -6.73229516e-01 1.68964118e-01
9.95879531e-01 -9.27500308e-01 -3.71442050e-01 4.70181972e-01
-1.19196486e+00 -1.31078708e+00 -6.61143243e-01 3.80432814e-01
-7.25470960e-01 -4.53506075e-02 -7.74521232e-01 -2.48609617e-01
-2.17509776e-01 -6.82156146e-01 7.47500360e-01 1.16750635e-01
-4.87564743e-01 -1.19077337e+00 3.25118601e-01 2.56985396e-01
3.18995833e-01 3.56599808e-01 1.29414654e+00 -4.95592177e-01
-5.09756744e-01 3.43774676e-01 1.48203298e-01 4.40465122e-01
6.26464784e-01 -1.38702482e-01 -4.14996952e-01 1.64003104e-01
-1.39377832e-01 -4.92694587e-01 8.47274661e-01 2.63203651e-01
8.44046295e-01 -8.61987650e-01 1.14478819e-01 3.74109358e-01
1.22021353e+00 6.06683679e-02 3.93296748e-01 -5.35376407e-02
8.56862903e-01 5.53408861e-01 1.81625307e-01 1.49658434e-02
7.00355053e-01 5.09405375e-01 -1.69009387e-01 2.14684740e-01
-6.08504653e-01 -1.07488096e+00 8.23970616e-01 6.93139970e-01
1.76253654e-02 -9.40943584e-02 -4.78370786e-01 9.16413367e-01
-1.69807053e+00 -1.35801375e+00 6.27379864e-02 1.97363830e+00
1.78793943e+00 -2.27696419e-01 2.44535692e-02 -4.07865435e-01
7.11991489e-01 1.11835226e-01 -6.88698351e-01 -4.86916780e-01
-6.23341324e-03 2.34736040e-01 3.23257685e-01 7.76301682e-01
-1.01551545e+00 1.39179850e+00 6.03625870e+00 5.76457798e-01
-7.96939433e-01 7.88767934e-02 5.28395832e-01 -1.93098977e-01
-9.16521311e-01 -1.86466664e-01 -4.78782535e-01 4.45101976e-01
8.55522454e-01 -1.92628786e-01 1.16361558e+00 3.91985446e-01
3.86789471e-01 3.24152887e-01 -1.62331939e+00 5.28385520e-01
7.86984637e-02 -1.61749840e+00 7.61243224e-01 -5.58718681e-01
1.38411951e+00 -2.01254860e-01 -1.04097761e-01 7.65814960e-01
1.24521768e+00 -1.02358711e+00 1.10440278e+00 5.25100768e-01
5.46443939e-01 -4.74229485e-01 1.80755943e-01 4.37482685e-01
-1.13966668e+00 1.09037429e-01 -1.87332913e-01 -5.21091282e-01
2.32176900e-01 2.35639438e-01 -7.42185950e-01 3.55498672e-01
1.22524954e-01 7.28675663e-01 -6.09203458e-01 2.46212721e-01
-1.04226041e+00 6.83146656e-01 -8.32170546e-02 -2.35916317e-01
2.70900249e-01 -2.35890791e-01 2.25370407e-01 1.15597796e+00
3.65289509e-01 2.99563974e-01 1.49955541e-01 1.67741036e+00
-4.52062339e-01 1.29671574e-01 -6.38610363e-01 -5.25060117e-01
2.77910441e-01 7.46292233e-01 -2.83161372e-01 -7.52972364e-01
-3.83024484e-01 1.23513949e+00 3.22191834e-01 5.32562613e-01
-7.81493545e-01 1.10887662e-01 6.75643563e-01 -1.12545975e-01
2.59706736e-01 -4.02114242e-02 -4.40337062e-01 -1.46946108e+00
-3.28141898e-01 -1.26813126e+00 3.59231025e-01 -1.04651988e+00
-1.63392341e+00 -6.03487790e-02 2.46899515e-01 -4.88319397e-01
-4.64727581e-01 -5.80618024e-01 -6.25165939e-01 8.28025281e-01
-1.05600941e+00 -1.43962908e+00 2.82241791e-01 3.90631378e-01
9.20342088e-01 9.27886441e-02 9.71761048e-01 -8.65636393e-02
-2.48513624e-01 3.35372776e-01 -2.67632127e-01 3.36688787e-01
6.88774049e-01 -1.58402956e+00 7.62557447e-01 9.64827955e-01
3.08025330e-01 1.07519197e+00 8.03909063e-01 -1.10199332e+00
-9.31580544e-01 -1.54601741e+00 1.25619221e+00 -9.31780100e-01
5.76142132e-01 -3.53122473e-01 -5.93785167e-01 1.06311977e+00
4.29990679e-01 -6.85814023e-01 8.10664535e-01 2.14818157e-02
-6.61390662e-01 3.03384423e-01 -9.55288529e-01 1.29475474e+00
1.38702154e+00 -6.53031886e-01 -1.30682707e+00 3.20621401e-01
1.33120155e+00 -1.24794453e-01 -7.30350137e-01 1.01450399e-01
3.40958029e-01 -5.60501814e-01 8.61788869e-01 -1.00528884e+00
9.09638405e-01 -5.49330831e-01 -1.18428402e-01 -2.14778471e+00
-5.17744422e-01 -7.42359936e-01 -1.91887859e-02 9.67093170e-01
9.44732428e-01 -3.52481246e-01 3.15343857e-01 9.35209453e-01
-3.74369919e-01 -2.51474321e-01 2.84420475e-02 -6.19010985e-01
5.15124738e-01 -3.47916067e-01 1.32948709e+00 1.22831655e+00
7.70652946e-03 5.55583060e-01 -3.21543366e-01 -1.17125317e-01
4.33289945e-01 4.16609615e-01 7.33745754e-01 -4.55077350e-01
-7.78843105e-01 -7.37331212e-01 7.39159048e-01 -1.09132135e+00
2.82509983e-01 -1.39214277e+00 1.68491662e-01 -1.80100250e+00
2.51240343e-01 -1.59469336e-01 -7.29899406e-02 7.04545557e-01
-6.82716668e-01 3.18016298e-02 4.28377718e-01 -2.85565525e-01
-3.83907855e-01 5.28712451e-01 1.61923170e+00 -4.23185706e-01
-5.29168189e-01 -1.23310044e-01 -1.18546939e+00 7.28790462e-01
1.12717712e+00 -3.65899801e-01 -6.32732987e-01 -7.61132419e-01
5.90017736e-01 -3.78696859e-01 4.32724386e-01 -3.56262505e-01
-2.02642754e-01 -6.39636874e-01 3.38215858e-01 -9.59481597e-02
-2.35028446e-01 -3.51019055e-01 4.85903800e-01 6.67467713e-01
-1.10526919e+00 1.05512217e-01 2.93169767e-01 2.25822955e-01
2.41928071e-01 -2.17142001e-01 4.57933664e-01 -7.47812688e-01
-7.25657284e-01 9.85403433e-02 -2.59891331e-01 2.29494318e-01
5.81614316e-01 2.72415161e-01 -3.73008728e-01 -2.14768887e-01
-6.96587265e-01 1.48425475e-01 6.67800784e-01 7.42958605e-01
1.19379967e-01 -1.75688803e+00 -1.15673745e+00 -5.84382340e-02
9.24220532e-02 -1.01275168e-01 3.89878452e-01 2.33381167e-01
-5.08805275e-01 3.31010461e-01 -9.20376703e-02 2.45395541e-01
-3.65836829e-01 1.12731135e+00 5.77458203e-01 -4.31409389e-01
-2.71093935e-01 6.74596012e-01 5.54213285e-01 -1.09834540e+00
-4.60494131e-01 -9.10175204e-01 4.09544222e-02 -2.43184671e-01
6.66257501e-01 1.83746055e-01 -5.44665992e-01 -1.57733798e-01
2.52790958e-01 1.98712841e-01 2.07755044e-01 -1.88270882e-01
1.15323591e+00 -1.69940859e-01 -2.49566451e-01 6.60097480e-01
7.02743530e-01 6.01457246e-02 -1.35763931e+00 -2.41626114e-01
-1.13967940e-01 -9.72391367e-02 -3.44210893e-01 -1.49529314e+00
-6.24120176e-01 3.84716660e-01 -2.49267757e-01 -8.29453114e-03
8.18736672e-01 -7.77210593e-02 8.82511139e-01 5.85937858e-01
2.18060926e-01 -1.23047817e+00 8.43399316e-02 5.34721792e-01
1.24021661e+00 -9.72162426e-01 -3.31874430e-01 -1.08849622e-01
-7.90289342e-01 6.60722136e-01 6.95544720e-01 -2.44517207e-01
3.82124424e-01 -2.15932503e-02 8.45699534e-02 -6.72096107e-03
-8.08157861e-01 -8.61273333e-02 3.57450664e-01 5.63198030e-01
7.80983984e-01 7.74898648e-01 -2.53603816e-01 8.88661683e-01
-7.06353366e-01 1.62595749e-01 4.60928559e-01 2.69980311e-01
-1.16179802e-01 -1.13492465e+00 -2.05848902e-01 4.09422040e-01
-1.21808223e-01 -6.85332537e-01 -5.65370917e-01 6.68437958e-01
5.47291577e-01 9.04327154e-01 -2.12252483e-01 -4.98062260e-02
4.40589070e-01 4.36665088e-01 8.37555647e-01 -1.08477902e+00
-7.96158075e-01 -3.34123194e-01 3.83901507e-01 -4.51330960e-01
-6.23055458e-01 -6.48368597e-01 -1.40308857e+00 -4.54619825e-01
3.41288567e-01 1.59177646e-01 2.00393617e-01 1.05458558e+00
1.56764477e-01 6.44676864e-01 3.31087679e-01 -7.92944491e-01
-7.92834878e-01 -1.10047066e+00 -3.48203421e-01 9.50454116e-01
9.36126336e-02 -8.67021382e-02 -1.63138747e-01 6.21249318e-01] | [11.512299537658691, 4.579752445220947] |
4a6841cb-af14-4106-b912-3558ea470827 | mutual-learning-to-adapt-for-joint-human | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Xuecheng_Nie_Mutual_Learning_to_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Xuecheng_Nie_Mutual_Learning_to_ECCV_2018_paper.pdf | Mutual Learning to Adapt for Joint Human Parsing and Pose Estimation | This paper presents a novel Mutual Learning to Adapt model (MuLA) for joint human parsing and pose estimation. It effectively exploits mutual benefits from both tasks and simultaneously boosts their performance. Different from existing post-processing or multi-task learning based methods, MuLA predicts dynamic task-specific model parameters via recurrently leveraging guidance information from its parallel tasks. Thus MuLA can fast adapt parsing and pose models to provide more powerful representations by incorporating information from their counterparts, giving more robust and accurate results. MuLA is implemented with convolutional neural networks and end-to-end trainable. Comprehensive experiments on benchmarks LIP and extended PASCAL-Person-Part demonstrate the effectiveness of the proposed MuLA model with superior performance to well established baselines. | ['Xuecheng Nie', 'Jiashi Feng', 'Shuicheng Yan'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['human-parsing'] | ['computer-vision'] | [ 4.91678715e-03 -4.90952358e-02 -3.85514766e-01 -5.88519812e-01
-1.58544886e+00 -2.47091755e-01 2.57077962e-01 -2.90539652e-01
-4.37214136e-01 5.05708933e-01 6.03694260e-01 2.14056909e-01
3.53653610e-01 -2.74789065e-01 -8.14170361e-01 -3.93474340e-01
2.08208948e-01 5.20621419e-01 3.59645098e-01 1.22782700e-01
-2.67619491e-01 2.15665251e-02 -1.33388817e+00 4.90717858e-01
7.13742971e-01 8.54093969e-01 4.67415571e-01 8.91342282e-01
4.31434475e-02 5.13826132e-01 -4.33824658e-01 -6.25281274e-01
1.38743699e-01 -2.28559077e-02 -9.08262730e-01 -1.48267865e-01
7.17214644e-01 -4.25616592e-01 -1.50500655e-01 7.01075196e-01
1.20669556e+00 2.41713360e-01 5.12728989e-01 -8.05040777e-01
-5.20757198e-01 5.34769952e-01 -8.75735879e-01 -1.40653193e-01
6.22428656e-01 3.84773880e-01 9.77653325e-01 -9.76381242e-01
4.19038504e-01 1.79993284e+00 1.06812143e+00 1.08130980e+00
-1.01619327e+00 -7.56461620e-01 4.39540744e-01 2.13493481e-02
-9.90826130e-01 -7.34741867e-01 8.92631710e-01 3.14472727e-02
1.10726285e+00 -3.38307656e-02 4.29936588e-01 1.67605293e+00
9.00309682e-02 1.61525667e+00 8.68057251e-01 -3.67955506e-01
-3.44614565e-01 -4.17157978e-01 1.51146695e-01 8.39858830e-01
-5.47770560e-02 -5.66188544e-02 -7.18267679e-01 -6.03422187e-02
5.03021717e-01 -4.34505939e-01 3.26895751e-02 -4.46108192e-01
-1.10098243e+00 4.16823238e-01 2.46572882e-01 -2.46386021e-01
-5.01069307e-01 5.05252182e-01 7.65829265e-01 -2.15011522e-01
3.75609338e-01 -1.07003599e-01 -9.24159169e-01 -2.52032280e-01
-5.83317637e-01 3.48695725e-01 4.23635274e-01 1.02116525e+00
3.43391538e-01 1.32069185e-01 -4.78060991e-01 1.12881005e+00
4.87578481e-01 6.07356369e-01 6.67094827e-01 -9.14271295e-01
6.09475672e-01 2.47582674e-01 -1.28149360e-01 -6.76388204e-01
-6.84908628e-01 -2.56466717e-01 -6.08516276e-01 5.60652167e-02
4.75166261e-01 -3.04965556e-01 -1.00410581e+00 2.09941959e+00
5.11546552e-01 9.47707817e-02 1.40318215e-01 5.75158834e-01
1.01018322e+00 4.40938294e-01 8.23867321e-01 9.23812911e-02
1.37332189e+00 -1.44950175e+00 -8.43428016e-01 -6.02743983e-01
5.35767913e-01 -1.01913750e+00 1.07051158e+00 3.68898183e-01
-1.15504956e+00 -1.10227430e+00 -6.63143039e-01 -3.17529380e-01
-1.88676044e-01 4.66860533e-01 9.25645828e-01 4.77249205e-01
-1.05273294e+00 3.71767491e-01 -8.87111425e-01 -2.49694929e-01
6.92963183e-01 4.03969884e-01 -4.82301712e-01 -5.59298741e-03
-8.26144993e-01 6.01557732e-01 3.34133476e-01 1.59028113e-01
-6.24721050e-01 -5.29854000e-01 -1.25526929e+00 -2.66182035e-01
4.44340169e-01 -1.21581078e+00 1.67396808e+00 -6.98674917e-01
-2.18026948e+00 8.48144829e-01 -4.31537956e-01 -4.48553234e-01
7.10177839e-01 -1.00683820e+00 -2.60640346e-02 -1.31624803e-01
8.75715390e-02 1.26106393e+00 8.86184812e-01 -1.02446485e+00
-5.71637809e-01 -4.19454575e-01 -3.30571711e-01 2.85001189e-01
-1.71310171e-01 1.50752291e-01 -8.62955689e-01 -7.28502393e-01
-1.72139276e-02 -9.01728153e-01 -2.23708242e-01 4.26421762e-02
-4.57814962e-01 -8.14396918e-01 9.85938489e-01 -8.59259784e-01
9.17806864e-01 -1.79294407e+00 2.85499394e-01 -5.59171677e-01
6.10220693e-02 5.68905175e-01 -5.00367820e-01 2.57271007e-02
5.28956801e-02 -2.18913913e-01 1.00044481e-01 -1.10788715e+00
3.44771594e-02 1.86704323e-01 1.82189092e-01 3.55934203e-01
1.26206860e-01 1.40427184e+00 -8.04034293e-01 -7.44817853e-01
3.69922400e-01 6.91919267e-01 -5.88407516e-01 5.19589126e-01
-3.03752780e-01 5.57352960e-01 -5.34639120e-01 8.48469496e-01
6.60332084e-01 -6.35795966e-02 2.74026003e-02 -5.26063979e-01
3.10123473e-01 1.40528366e-01 -9.23532248e-01 2.14971995e+00
-8.08257282e-01 4.28081483e-01 3.39670070e-02 -7.32880712e-01
8.03489864e-01 3.53415787e-01 4.50453311e-01 -6.92769051e-01
1.44836500e-01 -5.97209036e-02 -4.10658926e-01 -6.11888885e-01
3.42626572e-01 3.54618251e-01 -2.97351152e-01 1.56688988e-01
4.71150368e-01 1.08009845e-01 -3.56871694e-01 -3.46583612e-02
7.12365985e-01 8.47100139e-01 6.13028347e-01 -1.77220777e-01
6.65725231e-01 -7.13849962e-01 9.75732386e-01 4.21867937e-01
-8.23167384e-01 6.03249133e-01 -5.39303198e-02 -5.48898637e-01
-6.35655284e-01 -9.99420464e-01 9.23998281e-02 1.78012061e+00
-1.46415859e-01 -5.59449136e-01 -8.46586645e-01 -1.03268898e+00
3.66612780e-03 2.85455763e-01 -8.03931832e-01 -4.01748940e-02
-8.88499498e-01 -5.44231653e-01 5.65789282e-01 1.06804264e+00
8.13919604e-01 -1.14620805e+00 -4.48896497e-01 3.12262714e-01
-6.19352758e-01 -1.47283101e+00 -8.58858526e-01 -1.55037582e-01
-5.95044494e-01 -1.01071930e+00 -9.98145700e-01 -1.01460958e+00
5.91977239e-02 1.06043458e-01 1.22779548e+00 -3.22815955e-01
-3.04064870e-01 5.77368617e-01 -1.80071503e-01 -5.68351090e-01
-3.48605931e-01 3.05639416e-01 1.47938162e-01 5.58153316e-02
2.27384835e-01 -4.31304067e-01 -6.84215307e-01 2.63974488e-01
-1.59838825e-01 3.48387867e-01 7.07361937e-01 7.04212844e-01
9.17412758e-01 -6.63753390e-01 8.87788773e-01 -7.03502655e-01
5.08875489e-01 -1.02920141e-02 -2.43160620e-01 4.59359467e-01
-4.46378291e-01 1.61925718e-01 2.35228300e-01 -6.22231662e-01
-1.47693264e+00 6.39566660e-01 -4.90021974e-01 -3.93116087e-01
-2.98203498e-01 -1.68195099e-01 -7.45961964e-01 -2.32966375e-02
2.71911651e-01 1.00789927e-02 -8.54750723e-02 -7.96962023e-01
7.32347190e-01 6.44047320e-01 1.06723106e+00 -8.39392662e-01
4.95711923e-01 1.51153266e-01 -1.16518743e-01 -5.46834290e-01
-1.11890018e+00 -6.72674358e-01 -1.01607525e+00 -4.11219358e-01
1.17021191e+00 -1.15274215e+00 -8.67761791e-01 1.08291888e+00
-1.47178960e+00 -5.38293004e-01 2.76045525e-03 3.41916621e-01
-7.98304558e-01 5.09279668e-01 -6.89052343e-01 -7.62784064e-01
-7.97951281e-01 -1.14345324e+00 1.67470944e+00 4.49516386e-01
-2.17941761e-01 -1.15601230e+00 1.55044600e-01 7.54185081e-01
2.89300263e-01 2.28194863e-01 4.86906350e-01 -6.19377196e-01
-1.35760143e-01 -2.16165677e-01 -1.14179499e-01 3.03150922e-01
2.29271695e-01 -1.61601126e-01 -1.19746494e+00 -3.00889879e-01
-5.23744881e-01 -8.20279241e-01 1.05307281e+00 7.10661054e-01
1.24630117e+00 -2.95484960e-01 -5.20949483e-01 8.40519905e-01
8.14307570e-01 -3.42301637e-01 3.41599256e-01 1.10985167e-01
8.99745941e-01 5.22678018e-01 4.78934079e-01 2.71724343e-01
8.59323084e-01 1.05143225e+00 2.94816881e-01 -2.02207044e-01
-7.12353170e-01 -6.02920353e-01 5.61965108e-01 9.70444441e-01
-3.09287440e-02 2.22116262e-02 -7.05045044e-01 3.32665354e-01
-1.96687734e+00 -7.57786632e-01 1.06020615e-01 1.88456416e+00
8.97690952e-01 7.41170645e-02 5.78482747e-01 -2.45629519e-01
6.14436269e-01 3.52169305e-01 -6.75601363e-01 -3.01929891e-01
-1.18768461e-01 3.40929478e-01 2.29136482e-01 5.64828396e-01
-1.50023770e+00 1.52029419e+00 6.65628672e+00 8.55429471e-01
-7.23849595e-01 3.06608588e-01 5.49817204e-01 5.83752021e-02
1.58002362e-01 -4.42864180e-01 -1.06542087e+00 3.04471292e-02
7.21208692e-01 1.70852780e-01 2.52099745e-02 1.06195438e+00
9.75776315e-02 1.76012024e-01 -1.05531538e+00 1.15467346e+00
2.67913073e-01 -8.96429002e-01 2.80892402e-01 -2.73111641e-01
5.61978042e-01 6.17693588e-02 3.55794340e-01 7.23587871e-01
4.85381156e-01 -7.95036614e-01 6.19116247e-01 4.94194478e-01
7.07407057e-01 -7.76352108e-01 5.51039279e-01 1.94445372e-01
-1.73645043e+00 -1.09770641e-01 -2.11807922e-01 2.26480693e-01
3.86080921e-01 -7.91654363e-02 -7.67120659e-01 6.33841932e-01
7.21764743e-01 7.87096739e-01 -7.82347023e-01 9.05830026e-01
-2.69834012e-01 4.79196608e-01 -1.61431521e-01 1.67381644e-01
2.03650475e-01 3.47845525e-01 6.05736673e-01 1.62357736e+00
-1.78411722e-01 -1.71267509e-01 5.11274219e-01 3.73391896e-01
-2.36024141e-01 3.70133370e-01 -4.11668956e-01 4.03326601e-01
2.89828539e-01 1.38482988e+00 -2.44988278e-01 -2.77762055e-01
-4.53950197e-01 1.24325335e+00 6.37140989e-01 2.48399526e-01
-7.93919623e-01 -6.76871762e-02 8.01127791e-01 -2.22598910e-01
2.55719662e-01 -2.31909543e-01 -9.31823701e-02 -8.36571276e-01
-6.74398988e-02 -9.48337793e-01 6.05777264e-01 -6.59507215e-01
-1.30636764e+00 7.19144046e-01 5.01859263e-02 -8.16781521e-01
-6.70748174e-01 -7.48652399e-01 -6.24311149e-01 6.64752305e-01
-1.46647239e+00 -2.14755034e+00 -1.67637169e-01 7.56002247e-01
1.22630870e+00 -3.15242022e-01 1.08177412e+00 1.75696146e-02
-7.84080446e-01 1.34851336e+00 -4.25848752e-01 4.16647643e-01
9.48338091e-01 -1.34082961e+00 7.62395680e-01 7.60545135e-01
2.52861828e-01 2.93745726e-01 4.31278259e-01 -5.97771466e-01
-1.23296380e+00 -1.17205667e+00 6.15383446e-01 -5.58494747e-01
3.22888821e-01 -4.03300554e-01 -8.03586066e-01 7.85904527e-01
3.94849420e-01 2.09392384e-01 5.85553527e-01 4.81524944e-01
-7.23923862e-01 -1.44282565e-01 -8.80058825e-01 5.30096650e-01
1.31281626e+00 -5.10702193e-01 -5.07282376e-01 4.26796615e-01
8.71380627e-01 -7.63491929e-01 -6.18439436e-01 7.11778462e-01
9.06921983e-01 -6.98666275e-01 1.30410540e+00 -8.41738820e-01
-4.60484363e-02 2.94910669e-01 -1.59463242e-01 -1.13164663e+00
-3.00972730e-01 -1.10134363e+00 -4.27242368e-01 1.32115686e+00
2.82648921e-01 -3.38289768e-01 7.36019015e-01 4.16461349e-01
-1.64575905e-01 -9.46419001e-01 -1.16126478e+00 -6.93494499e-01
3.82699460e-01 -5.71384847e-01 5.53802788e-01 4.53207672e-01
-4.26508427e-01 5.89497268e-01 -8.47106040e-01 1.95322856e-01
9.46945369e-01 -1.67236909e-01 1.20360482e+00 -9.99272645e-01
-4.91684049e-01 -4.35391247e-01 -3.42163950e-01 -1.39022768e+00
7.63479471e-01 -7.73554087e-01 2.78707683e-01 -1.35774565e+00
4.54624921e-01 -2.28837445e-01 -2.97166795e-01 8.62280011e-01
-7.10369587e-01 1.32932037e-01 2.89027214e-01 -8.65151659e-02
-9.65161145e-01 6.85287952e-01 1.30727589e+00 -1.62810624e-01
-3.45970333e-01 4.96886343e-01 -4.66376305e-01 1.07989764e+00
6.89566731e-01 -2.09034219e-01 -2.57172287e-01 -6.80260837e-01
-2.86296278e-01 -1.10043939e-02 2.82980680e-01 -1.24378455e+00
2.25638643e-01 9.29446816e-02 4.88272548e-01 -7.61086345e-01
5.83592117e-01 -3.16037387e-01 -3.99234116e-01 7.88949728e-02
-4.03155208e-01 1.19680293e-01 5.71720302e-01 8.64939094e-01
1.04474708e-01 1.76962107e-01 8.30646515e-01 -1.49951631e-03
-9.74562287e-01 5.40234983e-01 1.13672152e-01 9.32013914e-02
7.82608449e-01 -9.18693170e-02 1.37162700e-01 -4.49880332e-01
-8.66857052e-01 2.70310283e-01 -1.36430219e-01 7.79235184e-01
6.08107746e-01 -1.46273601e+00 -9.07504439e-01 -2.25599054e-02
1.54201418e-01 5.80594465e-02 4.48628813e-01 4.80762303e-01
2.83262059e-02 4.03146237e-01 -1.37521952e-01 -8.50626469e-01
-1.76298392e+00 4.05122876e-01 4.66086596e-01 -4.87033278e-01
-8.31829548e-01 1.17792308e+00 3.31321776e-01 -7.41672337e-01
6.78226948e-01 -2.06328005e-01 -2.18664363e-01 -1.82686269e-01
7.09649503e-01 2.25417033e-01 -2.52060384e-01 -8.77506673e-01
-3.50162894e-01 8.25872362e-01 -3.17564964e-01 5.31835016e-03
1.25235105e+00 -2.56699651e-01 4.87591416e-01 2.44319588e-01
1.26085818e+00 8.82314071e-02 -1.62612951e+00 -4.78610247e-01
-1.69680595e-01 -2.03515887e-01 -4.66963947e-02 -1.04443741e+00
-1.12690580e+00 1.11144376e+00 7.18972385e-01 -8.50629807e-01
1.07748508e+00 1.73549727e-01 1.12327528e+00 4.46298152e-01
3.52299988e-01 -1.21571386e+00 6.49942040e-01 3.72446328e-01
1.01591039e+00 -1.36473179e+00 -1.53767318e-01 -4.71738100e-01
-8.00641060e-01 1.09534383e+00 9.04621780e-01 3.18514049e-01
3.88885736e-01 2.89978117e-01 3.50559950e-01 3.39385986e-01
-6.25891805e-01 -3.60293955e-01 5.51812768e-01 1.14912319e+00
4.86328065e-01 1.83585837e-01 5.53698838e-03 1.10096395e+00
-9.43714529e-02 -9.64259636e-03 -3.63105267e-01 6.07453227e-01
-1.79316849e-01 -1.40846002e+00 -1.70761645e-01 -1.33484587e-01
-5.17691314e-01 1.42435074e-01 -1.72331706e-01 9.22263265e-01
6.92610629e-03 7.54179478e-01 -2.80271649e-01 -4.93259162e-01
3.77255708e-01 3.75885874e-01 6.67748451e-01 -3.73316467e-01
-5.21210134e-01 2.81704277e-01 1.12845488e-01 -9.86537278e-01
-3.52110565e-01 -9.84368622e-01 -1.10266411e+00 1.41641395e-02
-3.05603236e-01 -4.46192294e-01 4.21300650e-01 1.00416422e+00
4.97344166e-01 7.18167245e-01 4.43280607e-01 -1.19742966e+00
-7.11562216e-01 -1.32681298e+00 1.95693433e-01 3.72951597e-01
1.76452234e-01 -8.45531225e-01 1.85173392e-01 -1.78965330e-02] | [7.554295539855957, -0.5628314018249512] |
075d1b17-6b9a-497e-9005-785fdaa846cb | experimental-analysis-regarding-the-influence | 2211.05507 | null | https://arxiv.org/abs/2211.05507v1 | https://arxiv.org/pdf/2211.05507v1.pdf | Experimental analysis regarding the influence of iris segmentation on the recognition rate | In this study the authors will look at the detection and segmentation of the iris and its influence on the overall performance of the iris-biometric tool chain. The authors will examine whether the segmentation accuracy, based on conformance with a ground truth, can serve as a predictor for the overall performance of the iris-biometric tool chain. That is: If the segmentation accuracy is improved will this always improve the overall performance? Furthermore, the authors will systematically evaluate the influence of segmentation parameters, pupillary and limbic boundary and normalisation centre (based on Daugman's rubbersheet model), on the rest of the iris-biometric tool chain. The authors will investigate if accurately finding these parameters is important and how consistency, that is, extracting the same exact region of the iris during segmenting, influences the overall performance. | ['Andreas Uhl', 'Josef Bigun', 'Fernando Alonso-Fernandez', 'Heinz Hofbauer'] | 2022-11-10 | null | null | null | null | ['iris-segmentation'] | ['medical'] | [-7.26394430e-02 1.11717626e-01 -1.04596511e-01 -4.97571640e-02
1.59828156e-01 -5.66312790e-01 3.01954240e-01 1.54494241e-01
-4.08267498e-01 1.67290837e-01 2.85349488e-01 -5.04763484e-01
-4.04740572e-01 -5.21429777e-01 -2.67215967e-01 -6.73935592e-01
4.22497690e-01 3.96646559e-01 -7.11931363e-02 1.60249829e-01
6.15388870e-01 8.22407961e-01 -1.56772888e+00 -2.87126184e-01
9.03929770e-01 5.75046301e-01 -4.99728620e-01 6.34261906e-01
2.66620159e-01 1.68766394e-01 -7.32155502e-01 -5.57181895e-01
4.17773962e-01 -5.02084851e-01 -8.79576564e-01 3.36578995e-01
7.12362707e-01 -3.28379154e-01 5.06789684e-01 8.15197587e-01
5.51731348e-01 -3.55762959e-01 6.24686718e-01 -3.84516358e-01
3.08952071e-02 2.73316562e-01 -4.16363508e-01 -2.24997085e-02
3.25636953e-01 5.07866800e-01 5.09272218e-01 -3.05188715e-01
6.78163350e-01 7.29115784e-01 6.60253763e-01 6.02316931e-02
-1.39884007e+00 -4.67418551e-01 -4.05218154e-01 -1.63458958e-01
-1.40673280e+00 -4.23530638e-01 3.45278233e-01 -8.25859964e-01
3.41729432e-01 1.89274877e-01 1.06877613e+00 1.63130552e-01
3.11629534e-01 1.98484180e-04 1.43932736e+00 -7.88712025e-01
-5.71721122e-02 5.80916941e-01 1.14250354e-01 4.52970862e-01
6.45570576e-01 5.42116046e-01 -7.04212561e-02 1.79561988e-01
7.97672212e-01 -4.98678803e-01 -2.63411403e-01 -1.54743835e-01
-6.36622012e-01 4.16432351e-01 4.45272028e-02 8.45434308e-01
-3.05608064e-01 -2.70617813e-01 1.24125041e-01 1.84593365e-01
5.97582385e-02 8.35707128e-01 8.28862749e-03 -4.39055711e-01
-1.33245754e+00 -4.10556942e-02 6.39216185e-01 1.40329003e-01
5.05060136e-01 -2.57950455e-01 -2.51969725e-01 3.83180231e-01
5.46957970e-01 4.85772312e-01 2.15361148e-01 -5.30579150e-01
-1.29472569e-01 1.20947051e+00 1.02657408e-01 -6.95920348e-01
-6.07542634e-01 -5.57936788e-01 -2.78614759e-01 4.23480272e-01
9.60709453e-01 -4.98772949e-01 -9.55310047e-01 7.91123807e-01
4.74038035e-01 -1.65935740e-01 -3.56421888e-01 9.51446354e-01
7.37433195e-01 -6.08786680e-02 8.02525040e-03 -9.92206782e-02
1.30506289e+00 -3.42661202e-01 -5.84941089e-01 2.66965091e-01
7.06093431e-01 -1.48638952e+00 9.07575786e-01 3.06043327e-01
-1.00587714e+00 -6.68302059e-01 -8.14433277e-01 2.31587216e-01
-1.65362194e-01 6.32128239e-01 1.54473990e-01 1.35306239e+00
-6.48271739e-01 5.21624923e-01 -5.05092978e-01 -4.43581998e-01
-2.66483068e-01 7.59368956e-01 -3.64884108e-01 4.48704839e-01
-4.84271020e-01 1.26093805e+00 1.55191585e-01 1.64704576e-01
1.93173438e-01 -4.94625628e-01 -5.60370684e-01 -2.48134315e-01
-2.24523485e-01 -5.45185924e-01 8.35550845e-01 -9.09191370e-01
-1.69437528e+00 1.20102620e+00 -1.38480201e-01 -2.37850711e-01
5.31441689e-01 9.06173140e-02 -2.57418871e-01 9.06547531e-02
-2.38291070e-01 2.41343424e-01 6.73844039e-01 -9.96977925e-01
-6.20954931e-01 -8.42547596e-01 -2.21223801e-01 -1.01754561e-01
3.51492465e-01 2.78473258e-01 -3.31129670e-01 -2.24260017e-01
1.60744444e-01 -1.23819315e+00 2.19195381e-01 -4.17981416e-01
-4.21589881e-01 -1.19265787e-01 3.67574930e-01 -8.69481444e-01
1.22939515e+00 -2.26120138e+00 -3.28433245e-01 8.71527970e-01
-1.10896619e-03 8.25342596e-01 1.56887531e-01 2.08357528e-01
-9.54398885e-02 1.12100177e-01 1.99450538e-01 1.80693865e-01
-2.19290584e-01 -2.24718705e-01 2.98509687e-01 8.61604273e-01
-1.65695906e-01 8.04803014e-01 -2.60716915e-01 -4.20494676e-01
6.72532499e-01 5.51093519e-01 -1.14904389e-01 5.53149730e-02
2.24107116e-01 6.59070253e-01 -7.55174682e-02 6.51404023e-01
6.48268163e-01 2.05463871e-01 9.91926268e-02 -5.10505512e-02
-4.88346934e-01 6.15185052e-02 -1.19214308e+00 9.17275131e-01
-2.52752602e-01 8.30643177e-01 -1.43154427e-01 -3.25758129e-01
1.23488581e+00 4.67128724e-01 4.04516518e-01 -6.38567150e-01
4.09313738e-01 4.55753952e-01 6.99545503e-01 -5.03299475e-01
3.30283076e-01 -2.28270486e-01 9.00791466e-01 3.71430874e-01
-3.59111845e-01 -1.08736783e-01 1.88158005e-01 -4.56447363e-01
7.45328963e-02 2.71163106e-01 2.20632136e-01 -4.20275152e-01
7.67443717e-01 1.74779575e-02 1.44745976e-01 2.85527974e-01
-1.62618816e-01 6.32851660e-01 7.94168591e-01 -2.45236978e-01
-7.89209604e-01 -4.38908368e-01 -7.64011204e-01 1.68068707e-01
1.33620024e-01 -1.17342152e-01 -1.16092920e+00 -3.66808325e-01
9.90984663e-02 4.92525518e-01 -6.09186649e-01 8.80809501e-02
-6.48446456e-02 -7.49605298e-01 1.63046539e-01 -1.09671198e-01
2.15539530e-01 -6.16653800e-01 -8.92894447e-01 -9.65017080e-02
3.89798969e-01 -8.26645374e-01 -2.47246653e-01 -7.40059674e-01
-1.06537819e+00 -1.58489883e+00 -6.49433553e-01 -1.87449321e-01
9.68787074e-01 -1.76254392e-01 8.42893422e-01 5.81642270e-01
-3.01105529e-01 4.30451363e-01 -3.60592067e-01 -6.68203294e-01
-5.86931348e-01 1.31155252e-01 -2.57546157e-01 3.95239830e-01
6.40580654e-01 -2.45170929e-02 -7.70751238e-01 5.40707529e-01
-6.44863069e-01 -3.17084312e-01 3.88649315e-01 2.17251331e-01
4.36447144e-01 -3.07702199e-02 -1.78664222e-01 -9.60213125e-01
5.36029220e-01 2.96055913e-01 -8.09930503e-01 3.20788622e-01
-1.26092434e+00 -2.46109627e-02 2.94020642e-02 -1.57935813e-01
-4.96768624e-01 -1.08112052e-01 -1.19116090e-01 -1.71450945e-03
-4.44408983e-01 9.64249223e-02 3.52759153e-01 -8.32209766e-01
7.20337272e-01 -1.50046363e-01 4.91294444e-01 -5.04748762e-01
-1.18697479e-01 7.23612547e-01 2.62771338e-01 -3.59833151e-01
6.27302170e-01 3.08745295e-01 4.41603512e-01 -8.63003671e-01
-1.87492892e-01 -6.89870298e-01 -8.09667349e-01 -4.72488999e-01
7.11376846e-01 -2.62693226e-01 -9.94812131e-01 5.35092652e-01
-8.56898487e-01 -3.58589143e-02 -3.30457389e-01 9.38876212e-01
-1.98414192e-01 3.68086636e-01 4.10197265e-02 -9.37690616e-01
-3.57089013e-01 -1.75671256e+00 8.36315930e-01 6.46890163e-01
-3.47410023e-01 -9.59785163e-01 1.10068977e-01 7.10222721e-01
3.12508643e-01 4.36278343e-01 7.08695531e-01 -4.25721765e-01
-3.32799822e-01 -4.34177876e-01 -1.78108454e-01 3.33669275e-01
1.85373537e-02 8.15501213e-01 -8.27632964e-01 -1.70860484e-01
-1.18528530e-02 6.23654008e-01 3.57012480e-01 8.11278224e-01
4.62866038e-01 -3.80631164e-02 -8.34883302e-02 6.87258303e-01
1.56769145e+00 1.05554648e-01 8.53661239e-01 4.35476065e-01
2.39109010e-01 1.12478423e+00 6.00633442e-01 -2.34838109e-02
-3.44787687e-02 7.44550169e-01 1.97021052e-01 -4.90985841e-01
-4.78343397e-01 1.45458980e-02 -5.91765419e-02 2.51358777e-01
-6.56111300e-01 3.74022752e-01 -1.09216309e+00 3.93611968e-01
-1.09683681e+00 -3.83966655e-01 -4.82397109e-01 2.47211909e+00
4.48709011e-01 4.24575731e-02 6.29126728e-01 3.20587605e-01
6.27677500e-01 -3.87691468e-01 2.25575753e-02 -8.71112227e-01
1.93145335e-01 3.05324167e-01 8.19674075e-01 7.18231201e-01
-4.85907733e-01 5.35933614e-01 6.99915934e+00 3.30237329e-01
-1.54358935e+00 -3.94162923e-01 6.57437623e-01 -5.31809703e-02
1.30367309e-01 3.14780921e-01 -9.51464713e-01 6.12952709e-01
8.11737597e-01 1.74339965e-01 2.10268348e-01 1.32288113e-01
6.37031019e-01 -7.47527897e-01 -7.03688502e-01 7.03279495e-01
-5.87653257e-02 -1.13904059e+00 -3.34133297e-01 6.00984573e-01
5.69577634e-01 -4.57107455e-01 8.99882764e-02 -5.98915935e-01
-3.49827707e-01 -1.18878818e+00 2.19969809e-01 9.52942073e-01
9.22191978e-01 -4.78009790e-01 1.36591423e+00 -8.68635327e-02
-5.69408953e-01 2.33630240e-01 2.93806076e-01 1.42280802e-01
-2.07566425e-01 4.50439543e-01 -1.23489952e+00 3.57342422e-01
1.14291079e-01 1.83826182e-02 -8.95503998e-01 1.48228574e+00
-2.12813139e-01 7.89047360e-01 -2.11380988e-01 6.65135905e-02
-4.30745147e-02 -7.51766026e-01 7.72364795e-01 9.39744294e-01
-1.35960773e-01 -1.78077877e-01 -5.95164478e-01 9.00558174e-01
5.50515115e-01 6.96830213e-01 -2.77586013e-01 -5.67095466e-02
2.33699873e-01 8.05647790e-01 -7.81742752e-01 2.01733544e-01
-4.96431649e-01 1.56424701e-01 -5.75534940e-01 3.73220295e-01
-8.11173320e-02 -2.45727584e-01 4.74089801e-01 9.83714938e-01
1.67994332e-02 2.81030089e-01 -8.38481247e-01 -4.22660977e-01
5.91178127e-02 -1.03905451e+00 7.45760947e-02 -4.31480914e-01
-3.93902153e-01 2.38048300e-01 -3.08603108e-01 -1.09584570e+00
-2.07534924e-01 -6.58854842e-01 -6.21703446e-01 1.23676944e+00
-9.99761343e-01 -1.43767321e+00 1.95884774e-03 3.48818153e-02
-3.66773516e-01 -2.31197789e-01 5.30197144e-01 -1.27529308e-01
-6.62576199e-01 9.39379871e-01 7.05541670e-02 3.62794161e-01
5.86749017e-01 -1.19803834e+00 -1.64330974e-02 9.14096296e-01
1.13855496e-01 8.84482861e-01 7.03713953e-01 -5.72294950e-01
-9.69583929e-01 -2.70574361e-01 8.50519240e-01 -5.21655500e-01
2.04344466e-01 5.40011704e-01 -4.46115017e-01 3.74484897e-01
-3.33371125e-02 -5.38465500e-01 7.64758646e-01 5.19671619e-01
8.54037553e-02 -2.55992919e-01 -1.28765380e+00 4.25676435e-01
4.14481871e-02 -4.81815100e-01 -5.33542454e-01 -1.22277126e-01
-1.66222379e-01 -6.16901398e-01 -1.32205987e+00 5.19123614e-01
1.04081631e+00 -1.38427866e+00 5.91855645e-01 -1.09459952e-01
-3.99297290e-02 -3.78404319e-01 7.68100142e-01 -7.45430768e-01
-7.60800689e-02 -6.63889527e-01 4.03879315e-01 1.08131468e+00
5.78080356e-01 -9.45093513e-01 8.89165342e-01 8.80990505e-01
5.62737226e-01 -8.32098842e-01 -8.86918306e-01 -2.90625691e-01
9.48921964e-03 5.64510264e-02 8.02485108e-01 5.90090752e-01
-4.02038470e-02 -3.81997749e-02 1.25997096e-01 9.56552699e-02
3.60613555e-01 1.88354790e-01 1.13330448e+00 -1.46515882e+00
-2.01433171e-02 -7.16055930e-01 -1.01971877e+00 -2.31030270e-01
-4.50585783e-01 -3.36080670e-01 -7.78410852e-01 -1.14732099e+00
-3.63121897e-01 -2.45610878e-01 1.57259792e-01 1.00133702e-01
-9.95301083e-02 1.98045313e-01 3.65512490e-01 2.93476075e-01
3.60734046e-01 -6.17850482e-01 1.43365073e+00 3.52522045e-01
-6.49423301e-01 4.87086415e-01 -5.95772207e-01 4.90211248e-01
5.43640494e-01 -2.35281944e-01 6.05765777e-03 4.01512310e-02
1.48264691e-01 9.54792202e-02 3.07657719e-01 -9.94971752e-01
2.40157023e-01 -7.98386987e-03 1.81339830e-01 -3.26444000e-01
-1.15688853e-01 -9.37108755e-01 4.53989804e-01 6.97378218e-01
-4.55653407e-02 -2.81057447e-01 3.02642107e-01 -1.98238328e-01
-1.47790879e-01 -7.55986392e-01 1.01469505e+00 1.28776416e-01
-5.11743687e-02 -1.52228683e-01 -7.60337710e-02 -2.89117485e-01
9.34569418e-01 -9.99136269e-01 -3.36601362e-02 -4.00554948e-02
-8.32144737e-01 -6.50694072e-02 1.14363480e+00 4.45871763e-02
-1.45292237e-01 -4.71610516e-01 -8.15292835e-01 5.31893492e-01
6.62844628e-03 -2.59693474e-01 -9.75237861e-02 1.35699260e+00
-9.61859047e-01 6.55568063e-01 -1.42766252e-01 -6.30953074e-01
-1.73010111e+00 1.13126568e-01 9.93940890e-01 1.59728602e-02
-2.42337525e-01 4.69880432e-01 -5.08749843e-01 3.43200341e-02
2.29521602e-01 -2.75337100e-01 -3.88953328e-01 1.22457802e-01
3.01873565e-01 5.33949852e-01 1.83110282e-01 -7.64811039e-01
-1.70286596e-02 1.18034852e+00 6.45808354e-02 -5.75727178e-03
7.17756331e-01 -7.22262412e-02 -5.94020188e-01 2.81251907e-01
5.04036427e-01 5.83875954e-01 -5.73612213e-01 -1.18729724e-02
-5.43807782e-02 -8.45696390e-01 2.97951967e-01 -9.43219423e-01
-9.90260661e-01 6.43882394e-01 1.07832074e+00 2.03651056e-01
1.12268364e+00 -4.66242909e-01 1.88742891e-01 -5.75541019e-01
-1.59174297e-02 -1.24471641e+00 -9.35839236e-01 -2.43594468e-01
4.53709543e-01 -1.04169989e+00 4.56769288e-01 -3.70328128e-01
-4.43784058e-01 1.27010071e+00 2.24387616e-01 1.27350092e-01
7.30311155e-01 -1.85524061e-01 2.97490448e-01 -3.21873635e-01
1.80096626e-01 -5.20591438e-01 9.46345687e-01 6.06958747e-01
8.08709264e-01 4.34198290e-01 -1.14368534e+00 4.51144651e-02
-5.35656750e-01 3.01713824e-01 4.41690177e-01 2.28006825e-01
-2.59177536e-01 -1.31361413e+00 -8.35849881e-01 6.45666599e-01
-6.87394857e-01 1.18479699e-01 -7.03717470e-01 9.36041296e-01
8.27032030e-01 9.63932812e-01 8.42995271e-02 -1.93723068e-01
3.89431387e-01 9.59861204e-02 4.77608144e-01 -4.82341647e-01
-1.10854721e+00 2.35684097e-01 -4.80342470e-03 -2.10817993e-01
-5.59860289e-01 -8.48043025e-01 -7.61164248e-01 -5.79926908e-01
-6.37515426e-01 3.62155378e-01 1.02602708e+00 1.14968371e+00
1.55483052e-01 -1.23853378e-01 3.81018370e-01 9.48903114e-02
-3.05242836e-01 -1.03366435e+00 -6.83465004e-01 1.49453327e-01
4.49831843e-01 -3.07907701e-01 -4.01321083e-01 -2.25904956e-01] | [3.748403787612915, -3.6240618228912354] |
1ed691c8-12d8-4f3e-9206-f587d196b18f | gradient-scarcity-with-bilevel-optimization | 2303.13964 | null | https://arxiv.org/abs/2303.13964v1 | https://arxiv.org/pdf/2303.13964v1.pdf | Gradient scarcity with Bilevel Optimization for Graph Learning | A common issue in graph learning under the semi-supervised setting is referred to as gradient scarcity. That is, learning graphs by minimizing a loss on a subset of nodes causes edges between unlabelled nodes that are far from labelled ones to receive zero gradients. The phenomenon was first described when optimizing the graph and the weights of a Graph Neural Network (GCN) with a joint optimization algorithm. In this work, we give a precise mathematical characterization of this phenomenon, and prove that it also emerges in bilevel optimization, where additional dependency exists between the parameters of the problem. While for GCNs gradient scarcity occurs due to their finite receptive field, we show that it also occurs with the Laplacian regularization model, in the sense that gradients amplitude decreases exponentially with distance to labelled nodes. To alleviate this issue, we study several solutions: we propose to resort to latent graph learning using a Graph-to-Graph model (G2G), graph regularization to impose a prior structure on the graph, or optimizing on a larger graph than the original one with a reduced diameter. Our experiments on synthetic and real datasets validate our analysis and prove the efficiency of the proposed solutions. | ['Nicolas Keriven', 'Samuel Vaiter', 'Hashem Ghanem'] | 2023-03-24 | null | null | null | null | ['bilevel-optimization'] | ['methodology'] | [ 3.65266472e-01 8.17339838e-01 -8.08529109e-02 -1.00309409e-01
2.96986960e-02 -5.11558235e-01 3.11680228e-01 5.01932144e-01
-3.35622311e-01 7.05971360e-01 -1.41500741e-01 -1.82314187e-01
-4.24864203e-01 -8.52877975e-01 -9.70870197e-01 -9.90622878e-01
-3.50439042e-01 3.30688775e-01 7.35785142e-02 -1.28667608e-01
1.76493898e-01 5.30718863e-01 -1.04531097e+00 -5.83697796e-01
7.67471910e-01 4.60325897e-01 1.20247409e-01 5.83476722e-01
1.42799895e-02 7.68169463e-01 -3.63223404e-01 -3.15469056e-01
5.01947820e-01 -4.03005272e-01 -7.38692939e-01 6.05539620e-01
4.84228909e-01 3.69659215e-01 -3.03755045e-01 1.39678323e+00
2.39045635e-01 7.68737495e-02 7.79096901e-01 -1.42439032e+00
-7.74467170e-01 6.64978027e-01 -7.78235674e-01 5.76877259e-02
-4.46137190e-02 -2.59905875e-01 1.20365047e+00 -4.67708200e-01
7.88415611e-01 9.99938071e-01 6.45692468e-01 4.52798218e-01
-1.49250937e+00 4.83363643e-02 3.84915769e-01 -2.24612191e-01
-1.37345588e+00 -1.40868381e-01 1.13656759e+00 -4.53784019e-01
5.15924692e-01 6.02452494e-02 5.47481835e-01 6.13022447e-01
1.82963893e-01 4.23734337e-01 9.66680646e-01 -6.63966417e-01
2.30046153e-01 3.72247458e-01 3.12171340e-01 1.13611066e+00
5.60785234e-01 -3.06796432e-01 -1.44749627e-01 -5.38472608e-02
6.71517968e-01 -1.26300126e-01 -5.56065619e-01 -9.48582351e-01
-6.81709886e-01 1.06343091e+00 7.17388093e-01 4.35363770e-01
-2.97564358e-01 9.26416367e-02 1.11578502e-01 4.90041316e-01
7.03950882e-01 3.80499184e-01 -4.70945500e-02 6.36143327e-01
-5.48080444e-01 -2.64402360e-01 1.11697996e+00 8.56658697e-01
1.18900156e+00 3.91523838e-02 2.55247146e-01 6.07791364e-01
4.03144062e-01 1.62747428e-01 1.97504133e-01 -5.37654281e-01
5.68182170e-01 7.53392816e-01 -2.87335366e-01 -1.45301378e+00
-5.50613284e-01 -7.45142996e-01 -1.04532635e+00 2.11499110e-01
7.92913616e-01 -1.09628491e-01 -7.69886732e-01 1.97195160e+00
1.64597645e-01 1.69394404e-01 -5.68936206e-02 9.86326694e-01
3.90369356e-01 4.70686287e-01 -6.24967553e-02 -5.43546200e-01
7.01720953e-01 -9.48470354e-01 -6.48264349e-01 -3.34280193e-01
9.05447960e-01 -2.93030769e-01 1.15251195e+00 2.30983749e-01
-7.63279021e-01 -9.03151035e-02 -1.14113235e+00 2.70933330e-01
-5.17102420e-01 -5.53400256e-02 4.23453331e-01 7.27318704e-01
-1.33259058e+00 8.50265086e-01 -5.70550740e-01 -6.22103333e-01
1.36024192e-01 4.31936175e-01 -4.24421519e-01 5.64912856e-02
-9.05085266e-01 7.53381312e-01 3.82999808e-01 3.24558884e-01
-6.09686732e-01 -2.26150975e-01 -9.13224638e-01 3.11620766e-03
6.05919242e-01 -4.25878912e-01 3.49618286e-01 -1.25582719e+00
-1.23861003e+00 1.06347561e+00 2.13608474e-01 -4.96563166e-01
5.70446253e-01 1.45605519e-01 4.23635216e-03 1.28816843e-01
-6.42115250e-02 3.07905525e-01 9.95916069e-01 -1.37719941e+00
1.20448209e-01 -5.95979095e-01 1.85197234e-01 2.26685762e-01
-6.15053356e-01 -5.80199420e-01 -3.15941215e-01 -2.89445966e-01
3.47995013e-01 -1.05513096e+00 -4.51815635e-01 -2.64093131e-01
-6.87973440e-01 -4.65212986e-02 7.42411911e-01 -4.09679949e-01
1.18171918e+00 -1.98250699e+00 5.99412382e-01 7.42731929e-01
7.27981091e-01 -1.29878605e-02 -2.27926165e-01 4.85542864e-01
-2.94366390e-01 2.23681003e-01 -4.94666755e-01 -3.01412255e-01
-7.29444101e-02 3.72979015e-01 5.19016199e-02 1.06984973e+00
-4.76997569e-02 7.30819523e-01 -9.06529784e-01 -5.12990832e-01
6.64923489e-02 4.90622997e-01 -3.88804883e-01 -4.11771648e-02
-3.86529714e-02 2.84723431e-01 -5.07097602e-01 2.12318614e-01
7.10513949e-01 -6.82590842e-01 5.09905338e-01 -1.98217034e-02
-2.21880469e-02 -1.15667738e-01 -1.39521861e+00 1.31354070e+00
-3.83673012e-01 4.90163475e-01 5.18448710e-01 -1.59530318e+00
9.17579234e-01 2.22239457e-02 5.32066345e-01 -1.85406789e-01
1.69340577e-02 1.48813367e-01 -1.14231713e-01 -3.92985672e-01
4.00841683e-02 -4.64818813e-03 1.77737162e-01 3.75202030e-01
6.42081723e-02 1.11452192e-01 5.50470173e-01 4.73016560e-01
1.20374119e+00 -2.03700528e-01 2.41594017e-01 -6.08388424e-01
6.33933306e-01 -3.16533893e-01 2.92841107e-01 8.18008363e-01
5.03168181e-02 3.50498378e-01 1.02134049e+00 -1.85722977e-01
-8.52299154e-01 -8.16647232e-01 -1.67173631e-02 8.98144305e-01
2.26816893e-01 -3.88280988e-01 -9.34384406e-01 -9.09210324e-01
-8.63697156e-02 1.66301116e-01 -8.31360638e-01 -2.28960976e-01
-4.88543719e-01 -9.37173486e-01 1.26740932e-01 -4.83497418e-02
3.03893358e-01 -8.50267410e-01 -3.85941625e-01 5.22426553e-02
1.66088283e-01 -9.55599546e-01 -3.22480232e-01 4.68285918e-01
-9.39167202e-01 -1.19614947e+00 -7.71151543e-01 -9.15639758e-01
1.26121974e+00 4.58684303e-02 1.17498887e+00 4.41089839e-01
-1.69700265e-01 5.12460172e-01 -3.08185071e-01 5.27315512e-02
-3.06950659e-01 3.91669661e-01 -9.51293334e-02 2.77111471e-01
-2.12631717e-01 -8.53069246e-01 -3.41198236e-01 1.68086737e-01
-1.04522157e+00 -1.68956518e-01 6.41077161e-01 6.38713956e-01
5.31570554e-01 2.13979244e-01 3.87812495e-01 -1.36501062e+00
8.07997048e-01 -4.68183190e-01 -8.28001618e-01 4.32555377e-01
-1.03959715e+00 5.17934561e-01 8.03381622e-01 -3.07692826e-01
-4.38632280e-01 3.27174574e-01 3.35242599e-01 -3.88237029e-01
2.21331939e-01 5.92960238e-01 -2.33031660e-01 -5.17524004e-01
6.11481547e-01 7.29836673e-02 6.93712011e-02 -3.06197196e-01
4.74474967e-01 1.56584144e-01 1.43276110e-01 -2.16258958e-01
9.66084182e-01 5.24845064e-01 5.60335100e-01 -1.00820482e+00
-6.94829762e-01 -4.45267737e-01 -6.43987536e-01 -3.42259467e-01
7.46543109e-01 -2.83249378e-01 -6.78409219e-01 9.00969282e-02
-8.55854273e-01 -4.14695948e-01 -2.99502015e-01 4.58099306e-01
-4.68170106e-01 6.38239622e-01 -5.19072890e-01 -8.42465222e-01
-7.02905059e-02 -8.60777736e-01 7.49877274e-01 5.37270084e-02
3.57385695e-01 -1.61997294e+00 2.12882534e-01 -1.99771196e-01
1.77330062e-01 4.27377015e-01 9.78729010e-01 -4.69479471e-01
-3.97786558e-01 -1.86318725e-01 -2.83157438e-01 3.01024914e-01
1.50352046e-01 -2.27755949e-01 -5.46340227e-01 -5.21201193e-01
1.25798553e-01 -1.39642701e-01 1.05459523e+00 5.55271327e-01
8.90476763e-01 -3.93983036e-01 -3.34960699e-01 7.01452792e-01
1.88849247e+00 -4.39897060e-01 2.34458730e-01 3.13900486e-02
9.77085471e-01 7.49863267e-01 -3.53488624e-02 1.70781985e-01
1.27630273e-03 5.55696845e-01 7.85095096e-01 -3.58672231e-01
-2.93146614e-02 -2.08538696e-01 1.12433128e-01 8.83857667e-01
-1.37800783e-01 -5.40756524e-01 -7.34917819e-01 2.92656898e-01
-1.92743063e+00 -4.08242345e-01 -5.03171802e-01 2.39495969e+00
4.21353072e-01 2.64206320e-01 1.00000814e-01 1.84486121e-01
1.00836587e+00 3.60930800e-01 -4.31760937e-01 -2.85042465e-01
-4.62396383e-01 -1.23349711e-01 1.04080224e+00 9.80590343e-01
-9.82097030e-01 8.26617837e-01 6.14283895e+00 5.44584215e-01
-1.20294893e+00 1.13650143e-01 6.18789434e-01 3.47470969e-01
-3.86735380e-01 2.53143907e-01 -4.59810585e-01 1.89018741e-01
4.43536162e-01 -1.02044836e-01 7.12937772e-01 7.17265368e-01
7.64431953e-02 -6.04702570e-02 -8.10313404e-01 7.23179519e-01
2.04228908e-01 -9.69923794e-01 -8.95348713e-02 2.99388438e-01
6.60055935e-01 -1.71118438e-01 -3.97795774e-02 -7.29541406e-02
4.01306078e-02 -9.50338185e-01 2.05648929e-01 3.87492239e-01
4.38396960e-01 -6.62972391e-01 4.59444225e-01 4.17627811e-01
-1.26904023e+00 1.47524491e-01 -5.24573028e-01 -7.40478933e-02
-1.01819895e-02 9.32770431e-01 -7.56311893e-01 5.83746791e-01
1.44085661e-02 7.15489089e-01 -8.42445672e-01 9.22912836e-01
-3.31043929e-01 6.31488085e-01 -4.56527829e-01 -9.50200707e-02
3.65093052e-01 -8.23543787e-01 8.29985261e-01 1.04728794e+00
-1.91248208e-02 -3.77167642e-01 1.75787657e-01 9.51066792e-01
-1.72161549e-01 5.78643143e-01 -1.01363301e+00 -1.42864600e-01
-1.13057211e-01 1.45652366e+00 -1.37219369e+00 -4.85941321e-02
-3.83848190e-01 9.96895075e-01 7.08730578e-01 6.49819314e-01
-5.76838553e-01 -3.88112903e-01 -1.84679151e-01 3.38641524e-01
3.38435248e-02 -1.76056176e-01 -6.75604641e-02 -1.12064660e+00
1.59387335e-01 -2.91188926e-01 3.45630497e-01 -3.58277112e-01
-1.19865584e+00 6.25089645e-01 -6.40282109e-02 -9.17661190e-01
2.61175558e-02 -6.01911843e-01 -5.78236222e-01 6.47641540e-01
-1.46435726e+00 -9.30020571e-01 -1.70922384e-01 6.91926122e-01
-2.10565701e-02 1.25658110e-01 4.63720322e-01 3.20894688e-01
-5.90107977e-01 3.58445078e-01 -1.23682804e-02 7.75850704e-03
2.79466629e-01 -1.54334021e+00 -1.45941973e-01 9.16036904e-01
3.47670197e-01 5.18709958e-01 8.55850458e-01 -6.46067739e-01
-1.40847135e+00 -9.53898072e-01 8.96806240e-01 -1.09457269e-01
7.09850013e-01 -5.25665045e-01 -8.76561105e-01 7.21682131e-01
2.89437294e-01 2.56721616e-01 2.50360370e-01 1.55434474e-01
-4.01629135e-02 2.16644928e-02 -9.21883762e-01 4.39172417e-01
1.17661333e+00 -2.97136784e-01 -2.06703261e-01 7.28257835e-01
4.56304222e-01 1.15246162e-01 -5.41826427e-01 2.45308444e-01
4.85892966e-02 -9.44626987e-01 6.92040205e-01 -5.13183713e-01
6.65597469e-02 -2.87903547e-01 1.57907411e-01 -1.49171734e+00
-2.03025758e-01 -7.61683762e-01 -4.62458382e-04 9.71449852e-01
5.49152672e-01 -8.10395598e-01 9.92284060e-01 1.90067530e-01
1.28259450e-01 -6.79535031e-01 -7.76651025e-01 -8.49916637e-01
-1.38524817e-02 1.21406145e-01 -5.82086220e-02 1.10012686e+00
2.34065093e-02 6.37337089e-01 -3.67729217e-01 2.61373788e-01
9.64814663e-01 3.13052651e-03 4.57211226e-01 -1.42157638e+00
-3.32815886e-01 -5.17511129e-01 -5.07758200e-01 -9.76181686e-01
4.21545058e-01 -1.17292511e+00 -1.05244547e-01 -1.49794912e+00
-9.48527735e-03 -4.20436949e-01 -1.99631244e-01 2.23172948e-01
5.64088449e-02 1.75662592e-01 6.16764575e-02 1.43767640e-01
-6.97452724e-01 3.15071464e-01 1.21029890e+00 -5.17528914e-02
-4.73144472e-01 1.71240777e-01 -5.10727704e-01 8.45814526e-01
6.76352322e-01 -6.91557407e-01 -6.09042823e-01 -2.45858938e-01
7.75420308e-01 -6.34640502e-03 2.72601694e-01 -7.78779805e-01
3.29128355e-01 2.00796068e-01 -1.19740352e-01 -7.72660449e-02
4.44289111e-02 -8.59497726e-01 -2.98348628e-02 5.41403830e-01
-5.14324665e-01 -1.16178274e-01 -3.09465230e-01 7.78719127e-01
-2.48883683e-02 -6.61564469e-01 8.05830657e-01 -1.64094791e-01
-2.72799164e-01 2.03600138e-01 -2.23854214e-01 1.16141140e-01
8.79387915e-01 -8.33009556e-02 -8.40137601e-02 -6.34150505e-01
-1.07037866e+00 2.25275651e-01 4.68427002e-01 -7.19816312e-02
4.44121331e-01 -1.11956978e+00 -5.36023855e-01 1.62889332e-01
-3.23492615e-03 -1.41348138e-01 -8.17964748e-02 1.21072483e+00
-5.53235054e-01 1.91679865e-01 1.22731887e-02 -5.53733766e-01
-1.09466243e+00 6.93872094e-01 4.21330512e-01 -5.38202643e-01
-5.97372532e-01 8.17833662e-01 2.63213485e-01 -4.86489028e-01
4.00689840e-01 -1.82017118e-01 -3.03689450e-01 -4.00622608e-03
-1.78001255e-01 3.56356055e-01 -5.43853566e-02 -6.88188136e-01
-1.91024795e-01 7.41286039e-01 1.61751553e-01 6.73873872e-02
1.29538906e+00 -3.04653049e-01 -3.56684744e-01 4.88249958e-01
1.43320084e+00 2.71871626e-01 -1.16423798e+00 -2.67995268e-01
1.34048611e-01 -1.59244463e-01 1.60179242e-01 -5.25578335e-02
-1.37765098e+00 6.30100906e-01 4.09742564e-01 9.83428836e-01
9.95426953e-01 2.60003567e-01 1.67191207e-01 4.49321002e-01
-4.80883522e-03 -1.09074509e+00 5.10603711e-02 3.15734923e-01
7.38162398e-01 -1.08595896e+00 6.58057258e-02 -7.43681550e-01
-3.48757774e-01 1.15358591e+00 3.96301538e-01 -5.71985483e-01
8.86674762e-01 -2.67275665e-02 -2.03071594e-01 -4.49515194e-01
-3.57784629e-01 -4.11982983e-01 1.30944580e-01 4.09180731e-01
3.06337953e-01 -1.15393205e-02 -5.69183946e-01 -1.34399727e-01
4.79629673e-02 -3.11949939e-01 6.23273909e-01 7.14611292e-01
-2.91776031e-01 -1.10220218e+00 -5.17764948e-02 3.11916411e-01
-3.30091983e-01 -6.97168857e-02 -7.44189203e-01 9.42260861e-01
-3.21027972e-02 8.12034547e-01 -1.39679924e-01 -4.43173721e-02
1.16040796e-01 -1.85190678e-01 5.71640491e-01 -6.24055088e-01
-1.74506158e-01 1.16163559e-01 -1.50764763e-01 -3.36167008e-01
-7.40511417e-01 -3.15395653e-01 -1.09661782e+00 -3.79612148e-02
-6.12098277e-01 5.17791927e-01 7.60348320e-01 8.10109794e-01
3.37016471e-02 3.94610316e-01 6.87571108e-01 -6.19308770e-01
-4.25350934e-01 -7.85326600e-01 -1.14727521e+00 4.79617864e-01
3.59299898e-01 -4.46431577e-01 -1.02725565e+00 -1.24328628e-01] | [6.990577220916748, 5.475196361541748] |
00273fc3-98b8-4c90-b610-bc897f7dcace | abcnet-real-time-scene-text-spotting-with | 2002.10200 | null | https://arxiv.org/abs/2002.10200v2 | https://arxiv.org/pdf/2002.10200v2.pdf | ABCNet: Real-time Scene Text Spotting with Adaptive Bezier-Curve Network | Scene text detection and recognition has received increasing research attention. Existing methods can be roughly categorized into two groups: character-based and segmentation-based. These methods either are costly for character annotation or need to maintain a complex pipeline, which is often not suitable for real-time applications. Here we address the problem by proposing the Adaptive Bezier-Curve Network (ABCNet). Our contributions are three-fold: 1) For the first time, we adaptively fit arbitrarily-shaped text by a parameterized Bezier curve. 2) We design a novel BezierAlign layer for extracting accurate convolution features of a text instance with arbitrary shapes, significantly improving the precision compared with previous methods. 3) Compared with standard bounding box detection, our Bezier curve detection introduces negligible computation overhead, resulting in superiority of our method in both efficiency and accuracy. Experiments on arbitrarily-shaped benchmark datasets, namely Total-Text and CTW1500, demonstrate that ABCNet achieves state-of-the-art accuracy, meanwhile significantly improving the speed. In particular, on Total-Text, our realtime version is over 10 times faster than recent state-of-the-art methods with a competitive recognition accuracy. Code is available at https://tinyurl.com/AdelaiDet | ['Chunhua Shen', 'Yuliang Liu', 'Lianwen Jin', 'Hao Chen', 'Tong He', 'Liangwei Wang'] | 2020-02-24 | abcnet-real-time-scene-text-spotting-with-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Liu_ABCNet_Real-Time_Scene_Text_Spotting_With_Adaptive_Bezier-Curve_Network_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Liu_ABCNet_Real-Time_Scene_Text_Spotting_With_Adaptive_Bezier-Curve_Network_CVPR_2020_paper.pdf | cvpr-2020-6 | ['text-spotting'] | ['computer-vision'] | [ 1.28707707e-01 -6.76416397e-01 6.12945855e-02 -2.33533427e-01
-6.22044504e-01 -6.84385061e-01 3.06326717e-01 2.01350585e-01
-5.97563326e-01 1.40649840e-01 -2.16275796e-01 -4.92829025e-01
3.23960692e-01 -8.72214615e-01 -6.20624125e-01 -3.96382660e-01
5.39338768e-01 3.22187096e-01 7.45782495e-01 -1.07211128e-01
5.83774805e-01 5.21282911e-01 -1.10378885e+00 1.95468888e-01
1.11084032e+00 1.14886451e+00 -3.64185274e-02 7.27951944e-01
-5.01404226e-01 3.50375891e-01 -5.34747422e-01 -5.91502011e-01
3.29890221e-01 -5.69359064e-02 -4.67723727e-01 5.84010333e-02
6.02222204e-01 -6.17832899e-01 -4.54002947e-01 1.04853392e+00
5.14521837e-01 -1.46700978e-01 4.26309675e-01 -8.26583803e-01
-4.28280085e-01 6.54857755e-01 -8.13738585e-01 1.63971968e-02
2.16135755e-01 7.25902840e-02 9.49226499e-01 -9.93870616e-01
9.82365310e-02 1.05149186e+00 1.02310789e+00 3.83411676e-01
-9.96711314e-01 -8.32678795e-01 2.94513375e-01 -8.04281514e-03
-1.64581549e+00 -3.30087721e-01 6.03795886e-01 -3.76762748e-01
7.05288351e-01 3.63487840e-01 6.61616921e-01 6.83012128e-01
-2.20538631e-01 1.25087047e+00 6.85517430e-01 -3.63317639e-01
4.61775996e-02 -2.63503611e-01 2.28984341e-01 9.25872564e-01
1.87128320e-01 -5.76916218e-01 -2.26299226e-01 -7.97754899e-02
1.10693026e+00 1.81717575e-01 -3.35299969e-01 -1.04368024e-01
-1.38677680e+00 6.21170282e-01 2.42568135e-01 1.91331238e-01
2.50192136e-02 3.14351410e-01 6.18325412e-01 -8.50866213e-02
4.27377552e-01 9.37595516e-02 -1.88525453e-01 -3.57648671e-01
-1.06617379e+00 9.46359485e-02 8.12358797e-01 1.14877427e+00
3.86371702e-01 1.46409452e-01 -8.27284753e-02 1.02302945e+00
1.43093407e-01 5.03299952e-01 3.64275783e-01 -2.55810201e-01
7.85084248e-01 8.38295400e-01 -1.73148885e-02 -9.84174967e-01
-2.77938217e-01 -1.11214951e-01 -1.02883863e+00 -1.37923032e-01
7.42975593e-01 -1.31752640e-01 -8.68679225e-01 8.47943842e-01
2.67405719e-01 1.37607977e-01 -3.36637765e-01 8.83351266e-01
8.12272131e-01 7.51465619e-01 -2.54711270e-01 2.64271170e-01
1.52106631e+00 -1.12039876e+00 -5.35749614e-01 1.85276810e-02
5.43646932e-01 -9.98452127e-01 1.27365720e+00 6.80156827e-01
-8.46357942e-01 -4.18575406e-01 -1.12975073e+00 -1.00257404e-01
-1.76729441e-01 5.28664947e-01 5.51767528e-01 8.81871283e-01
-7.56432950e-01 4.61439848e-01 -9.43035781e-01 -2.39602223e-01
6.84093416e-01 3.32899779e-01 1.60078257e-01 9.84176621e-02
-6.88318014e-01 2.01301888e-01 1.10663317e-01 1.50609627e-01
-3.93355250e-01 -5.44670463e-01 -6.37899816e-01 8.08832943e-02
6.10074759e-01 -1.72913462e-01 1.45041907e+00 -8.00085604e-01
-1.60972929e+00 5.74601471e-01 -1.56733409e-01 -1.89897627e-01
1.06189001e+00 -4.74053919e-01 -3.29027236e-01 2.92893231e-01
-1.53408095e-01 3.33165348e-01 8.20612669e-01 -8.28084052e-01
-7.30307639e-01 -1.36148944e-01 -1.64524719e-01 4.70799766e-02
-7.46534348e-01 2.99000591e-01 -1.14317560e+00 -1.02540600e+00
3.25474709e-01 -6.37151957e-01 -5.50140887e-02 5.26240349e-01
-6.20878816e-01 -4.62351292e-01 1.06288469e+00 -5.37011087e-01
1.45368755e+00 -2.05495548e+00 -4.45271522e-01 -4.99826260e-02
3.25266212e-01 6.21051431e-01 2.50559688e-01 1.93132237e-01
2.78430969e-01 3.80031884e-01 -2.28150651e-01 -3.57738972e-01
4.51762974e-02 -3.15588802e-01 -4.40329790e-01 6.20832860e-01
2.04579458e-01 8.69461119e-01 -5.13445199e-01 -5.80754578e-01
2.66587555e-01 4.15767401e-01 -3.84213269e-01 -6.97115436e-02
-3.09761673e-01 7.95026571e-02 -7.15443373e-01 9.31681216e-01
7.91484654e-01 -2.64419109e-01 -1.06955461e-01 -2.38300890e-01
-4.03893739e-01 1.86873332e-01 -1.26712012e+00 1.41227913e+00
-4.30119969e-02 8.28404188e-01 -5.84997274e-02 -7.91390717e-01
1.15204227e+00 5.74675761e-02 2.48462066e-01 -3.58250350e-01
3.77090156e-01 3.61982882e-01 -1.85788408e-01 -3.26605082e-01
7.61581600e-01 3.57103169e-01 -6.13531284e-02 3.42526257e-01
-6.25728488e-01 -3.01488619e-02 9.43243802e-02 9.91621912e-02
8.74872923e-01 2.33529538e-01 1.20737456e-01 -2.81577736e-01
5.10241508e-01 -1.88575145e-02 5.76494277e-01 7.06021249e-01
-2.07875684e-01 8.37599218e-01 6.75494850e-01 -5.94881892e-01
-1.13885474e+00 -6.53030276e-01 -3.54801565e-01 9.58521485e-01
3.23947012e-01 -5.51503897e-01 -1.06491899e+00 -4.77675378e-01
-3.27178687e-02 3.29243809e-01 -3.25258136e-01 5.38163245e-01
-1.03436649e+00 -7.85804272e-01 9.72413659e-01 9.35861290e-01
9.63149011e-01 -6.07891083e-01 -7.63125539e-01 1.76848829e-01
3.99718396e-02 -1.32333601e+00 -1.04707372e+00 -2.16224849e-01
-1.04276872e+00 -7.92073429e-01 -9.80022848e-01 -9.21753824e-01
7.76630580e-01 3.81837577e-01 6.31904244e-01 4.36021030e-01
-4.31554407e-01 6.78544268e-02 -5.57328343e-01 -1.92966253e-01
1.92645006e-02 9.10876617e-02 -3.44453990e-01 1.59488693e-01
4.02182072e-01 -2.10000694e-01 -7.77685344e-01 6.41990364e-01
-8.95292044e-01 3.47431630e-01 4.58711386e-01 6.79205656e-01
6.96206868e-01 3.03316694e-02 1.06453963e-01 -7.48720109e-01
4.08904403e-01 5.88160641e-02 -8.01857054e-01 1.54253706e-01
-4.66137946e-01 -3.41753542e-01 8.89048100e-01 -7.19030499e-01
-8.00949574e-01 3.40304852e-01 -2.06639633e-01 -3.35162342e-01
-1.56699717e-01 1.49336964e-01 -9.63925421e-02 -1.94572985e-01
3.87753636e-01 5.14873624e-01 -3.41732234e-01 -6.20279729e-01
6.15831204e-02 9.46903527e-01 5.52426994e-01 -8.30046177e-01
6.63068593e-01 6.86534941e-01 -2.75391519e-01 -9.84507978e-01
-4.75372702e-01 -5.08171201e-01 -8.24145734e-01 -1.01743266e-01
6.81313396e-01 -6.85247362e-01 -9.47503626e-01 1.10677624e+00
-1.27213728e+00 -5.56728065e-01 2.71499395e-01 1.40136182e-01
-1.50289714e-01 8.77800941e-01 -9.17802751e-01 -8.09019566e-01
-7.17443466e-01 -1.25105703e+00 1.16784263e+00 4.00503188e-01
1.85318500e-01 -7.04270184e-01 -5.18665493e-01 2.19046131e-01
3.52754653e-01 1.90895259e-01 6.96880639e-01 -6.32959008e-01
-7.39113033e-01 -4.60604072e-01 -6.82329774e-01 6.46315739e-02
-1.56702369e-01 4.16424870e-01 -7.39236534e-01 -2.51939565e-01
-4.41658020e-01 -1.35858268e-01 8.16874802e-01 5.95655590e-02
1.62316465e+00 -9.46872681e-02 -3.48214060e-01 9.50387061e-01
1.38173783e+00 2.52078980e-01 6.43718183e-01 4.13694322e-01
9.59908664e-01 2.14323312e-01 5.88909507e-01 7.97896087e-01
3.18724453e-01 7.64400721e-01 2.01050296e-01 -1.46851629e-01
6.78232014e-02 -2.23941371e-01 1.21471822e-01 8.44712198e-01
-3.43481041e-02 -3.91740948e-01 -1.20663476e+00 4.57684219e-01
-1.89567804e+00 -5.23479104e-01 -7.60534167e-01 2.12844968e+00
7.22024322e-01 3.26696664e-01 3.25572520e-01 2.37233683e-01
1.02208114e+00 4.65041539e-03 -7.09307909e-01 -1.91511020e-01
-8.86831731e-02 1.67278484e-01 6.57932162e-01 1.57677621e-01
-1.30618918e+00 1.13730192e+00 5.43006659e+00 1.21140623e+00
-1.26607072e+00 -8.17165524e-02 6.73725307e-01 9.50265154e-02
1.54209420e-01 -5.48042022e-02 -1.24823093e+00 4.79141831e-01
3.28824341e-01 -4.78412062e-02 2.66008914e-01 8.06884825e-01
7.68197849e-02 5.42624742e-02 -8.52244437e-01 1.10394669e+00
1.88511953e-01 -1.24630940e+00 3.86281721e-02 -1.80216461e-01
4.60945994e-01 -3.52316797e-02 -1.00001432e-01 1.32348567e-01
1.04704581e-01 -8.81294429e-01 1.00861847e+00 1.96840256e-01
1.19591010e+00 -6.39845490e-01 4.53891724e-01 2.77435988e-01
-1.61344814e+00 2.28082940e-01 -5.15180886e-01 7.84690678e-02
2.60629691e-02 5.82778752e-01 -3.55353534e-01 2.97574073e-01
7.29520202e-01 6.57259583e-01 -5.28575659e-01 1.37016964e+00
-1.19222544e-01 8.21228862e-01 -4.66456503e-01 -5.84015608e-01
3.12182188e-01 -2.28529751e-01 3.92823577e-01 1.57457614e+00
1.87568143e-01 2.03261554e-01 3.50627154e-01 8.78456533e-01
-2.43400812e-01 4.82080936e-01 4.88476008e-02 -2.44662482e-02
5.50589025e-01 1.30663061e+00 -1.35033345e+00 -5.17037988e-01
-5.01801372e-01 1.21832705e+00 7.91445673e-02 6.55193552e-02
-1.00657701e+00 -1.13662159e+00 1.95676997e-01 7.27903470e-02
6.54649317e-01 -4.14470077e-01 -6.42772615e-01 -1.30050898e+00
3.46260279e-01 -8.38184357e-01 2.48562783e-01 -4.29348409e-01
-1.04542446e+00 6.10621750e-01 -4.29129303e-01 -1.27566421e+00
5.94745517e-01 -9.67854798e-01 -6.80332243e-01 7.29464769e-01
-1.33764970e+00 -1.14542615e+00 -6.39383852e-01 4.12702650e-01
8.58676434e-01 5.00333831e-02 5.15079737e-01 3.47825021e-01
-9.35245454e-01 1.01621342e+00 3.21161360e-01 8.60791981e-01
6.86493635e-01 -1.06816745e+00 1.04858446e+00 8.63246679e-01
-9.32744220e-02 5.50933123e-01 1.09936871e-01 -7.17799246e-01
-1.58952296e+00 -1.11933875e+00 3.77181917e-01 -2.94229388e-01
8.10235143e-01 -7.02661753e-01 -1.18397164e+00 4.11320657e-01
-2.10253879e-01 1.24421380e-01 4.15599406e-01 -3.07810813e-01
-6.95668876e-01 -3.00825853e-02 -1.06197202e+00 8.58190536e-01
9.67848003e-01 -1.94459260e-01 -2.14998543e-01 3.16849053e-01
6.36784136e-01 -7.75466442e-01 -7.71728039e-01 1.49331957e-01
7.82410800e-01 -8.70766044e-01 6.74829125e-01 -1.53829083e-01
4.60730761e-01 -4.22340870e-01 -5.69684803e-02 -5.95686495e-01
-1.42788157e-01 -8.81667435e-01 -3.28151584e-02 1.28080368e+00
2.51965582e-01 -8.76944244e-01 9.50209558e-01 3.58074039e-01
-1.54955775e-01 -1.12255383e+00 -7.02653348e-01 -8.39369535e-01
3.35137993e-01 -6.52747929e-01 7.04567671e-01 8.57516348e-01
-7.66242296e-02 -1.02087269e-02 -1.70421734e-01 3.74765657e-02
4.74641979e-01 1.36913121e-01 8.94197583e-01 -1.12147653e+00
-2.19686657e-01 -9.44834888e-01 -3.78570050e-01 -1.90298486e+00
-2.93028533e-01 -6.17911816e-01 1.11197442e-01 -1.34800029e+00
2.15048358e-01 -5.99668741e-01 3.06688339e-01 6.57749891e-01
-2.51210034e-01 3.47643614e-01 1.46399170e-01 4.25089419e-01
-6.14156425e-01 3.61094177e-01 1.19928622e+00 -5.36905639e-02
-8.36587176e-02 -1.20265976e-01 -4.76544321e-01 9.55824852e-01
8.78557920e-01 -2.63133854e-01 1.93084583e-01 -7.26069808e-01
1.37096122e-01 -1.62770882e-01 2.46708468e-01 -9.84152853e-01
4.10669714e-01 -2.18634680e-02 3.48184675e-01 -9.08560514e-01
7.65474066e-02 -5.31908751e-01 -4.58806127e-01 4.55295444e-01
-1.13043569e-01 1.00749666e-02 4.25142586e-01 4.85959768e-01
1.60495818e-01 -5.46670735e-01 8.97863507e-01 2.12388456e-01
-5.56745589e-01 3.89118820e-01 -3.62493187e-01 2.79542729e-02
8.62248778e-01 -4.68072683e-01 -4.56919819e-01 -1.76840410e-01
7.41994381e-02 1.89041942e-01 6.85553491e-01 1.88419059e-01
6.56750023e-01 -9.82671440e-01 -7.45405197e-01 8.22522268e-02
4.71219828e-04 3.84994447e-01 5.73040638e-03 9.31809962e-01
-1.17773187e+00 4.39226061e-01 3.99672508e-01 -8.55569422e-01
-1.28308034e+00 2.71174639e-01 3.22870493e-01 -1.08387209e-02
-1.01980317e+00 7.83678710e-01 7.50858709e-02 -2.77025282e-01
4.39389199e-01 -4.46803749e-01 1.43269256e-01 -2.40504459e-01
6.28906250e-01 4.97960031e-01 -4.47268710e-02 -4.12585258e-01
-1.83917180e-01 1.01821852e+00 -3.74098301e-01 1.62767693e-01
1.12170172e+00 2.67924160e-01 -6.33654445e-02 2.47683480e-01
8.18608940e-01 2.33558193e-01 -1.44945908e+00 -3.55690449e-01
-9.06187519e-02 -6.56150162e-01 -2.58314852e-02 -6.35498703e-01
-1.01078188e+00 1.07836580e+00 4.05757606e-01 2.19838858e-01
9.95418310e-01 -4.12395865e-01 1.21939158e+00 6.06459439e-01
1.94559604e-01 -1.10658443e+00 1.78011894e-01 6.45864069e-01
6.24255896e-01 -1.01129603e+00 1.20103180e-01 -7.74393618e-01
-4.27833110e-01 1.60976803e+00 7.49506772e-01 -1.88917667e-01
4.03250664e-01 5.99526703e-01 -2.31605452e-02 3.45083736e-02
-2.42545068e-01 6.16755933e-02 2.41856590e-01 9.28705409e-02
6.10471010e-01 -8.10783065e-04 -2.92874873e-01 6.91925824e-01
-2.56480044e-03 -2.47998044e-01 5.17182112e-01 1.05390775e+00
-5.47642469e-01 -9.23195779e-01 -6.46483421e-01 5.77731848e-01
-6.51366889e-01 -3.68880987e-01 -4.57888454e-01 7.70447016e-01
-2.94073433e-01 7.32055068e-01 2.20582873e-01 -2.92431951e-01
4.49343681e-01 -1.03788204e-01 2.47029424e-01 -4.43584919e-01
-5.62755644e-01 4.11661714e-01 -1.50743857e-01 -1.92656979e-01
1.05412908e-01 -6.72413111e-01 -1.47632194e+00 -5.23148835e-01
-7.25495338e-01 -2.50899464e-01 7.07508743e-01 6.48229182e-01
2.90860087e-01 3.64362925e-01 4.95203406e-01 -7.80053496e-01
-5.99140227e-01 -8.21326494e-01 -3.01560402e-01 7.06403255e-02
-3.92784104e-02 -3.15593362e-01 -2.25726619e-01 1.66824982e-01] | [12.047886848449707, 2.2474300861358643] |
99f15c65-397f-422a-8577-1f17c25ea168 | an-incremental-negative-sequence-admittance | 2205.02962 | null | https://arxiv.org/abs/2205.02962v1 | https://arxiv.org/pdf/2205.02962v1.pdf | An Incremental Negative Sequence Admittance Method for Fault Detection in Inverter-Based Microgrids | Superimposed sequence quantities have been relied on for years to provide pure fault components for various applications in protection schemes. In more recent times, they have been employed in various solutions to the challenges introduced by integration of distributed generation in distribution systems. To improve the overall reliability of protection schemes deployed in such active distribution networks, this paper evaluates a new approach for detecting unbalanced faults using negative sequence quantities. A proposed superimposed negative sequence admittance element is validated using simulation tests on a modelled community microgrid. The performance of the method is tested for various fault scenarios in MATLAB/Simulink. | ['Aleksandar Dimitrovski', 'Subash Pokharel', 'Kwasi Opoku'] | 2022-05-05 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [-2.00483993e-01 -1.74928367e-01 3.48692656e-01 -2.81076059e-02
1.89765003e-02 -9.91100848e-01 5.02262473e-01 5.61055303e-01
4.21009153e-01 1.32532084e+00 -2.20869318e-01 -2.66079068e-01
-6.56830668e-01 -8.40276539e-01 1.32241398e-01 -8.20969045e-01
-6.70450211e-01 1.56634197e-01 3.84978265e-01 -7.58847713e-01
3.17663103e-01 1.13025510e+00 -1.21437931e+00 -1.03355609e-01
1.09671128e+00 5.41104436e-01 -2.15008080e-01 3.00955921e-01
5.49298704e-01 6.07549250e-01 -1.88043666e+00 2.47220956e-02
1.99675292e-01 -6.11432552e-01 -4.66755092e-01 6.51364475e-02
-8.63468707e-01 -3.36082339e-01 -2.06703246e-01 1.31005573e+00
6.87539399e-01 1.13267578e-01 9.35698390e-01 -1.99070966e+00
1.00971498e-02 8.20296586e-01 -6.06342614e-01 8.01693320e-01
7.18863964e-01 -7.87237138e-02 4.99038965e-01 -3.71763200e-01
5.14690764e-02 7.83064187e-01 4.65544850e-01 -2.34183803e-01
-1.26102960e+00 -3.55425119e-01 -3.36337537e-01 5.10097444e-01
-1.38309574e+00 3.79772812e-01 1.09691668e+00 -1.86358824e-01
1.36775577e+00 4.88453299e-01 7.78652430e-01 4.55438383e-02
5.75482428e-01 3.62454087e-01 9.26550269e-01 -3.71416807e-01
3.37245256e-01 8.84096697e-03 1.48633406e-01 -1.14009738e-01
8.20340931e-01 -2.64465600e-01 1.91624001e-01 -5.37343383e-01
3.44427496e-01 -4.60665971e-01 -9.44337785e-01 4.59497096e-03
-3.91718477e-01 7.01587319e-01 3.53383601e-01 1.10721529e+00
-1.38336465e-01 -3.11644703e-01 4.61338341e-01 6.53653145e-01
5.52074969e-01 4.06708479e-01 -2.41219729e-01 -6.62059337e-02
-6.58609927e-01 3.45321178e-01 1.02230215e+00 7.62540281e-01
-9.59160086e-03 9.13905263e-01 6.26334131e-01 5.26783705e-01
2.11563468e-01 4.33169812e-01 4.15418535e-01 -1.32846326e-01
-1.68648716e-02 8.10274482e-01 2.05491886e-01 -8.47716630e-01
-6.79646611e-01 -9.04128373e-01 -8.62096012e-01 6.64169133e-01
-1.95439532e-02 -4.20361906e-01 -2.30888903e-01 9.33875799e-01
2.31000319e-01 -2.65127122e-01 7.97846913e-02 5.43129206e-01
-1.47282675e-01 8.43936443e-01 -4.63107526e-01 -3.38275313e-01
9.93142009e-01 -1.56220481e-01 -9.54422414e-01 7.70197988e-01
4.39975381e-01 -9.93600309e-01 -1.26714677e-01 9.90425706e-01
-1.01141632e+00 -1.61584869e-01 -1.59527838e+00 8.17622066e-01
-4.66604084e-01 -1.67519465e-01 4.86830622e-03 1.13323021e+00
-1.17261004e+00 7.26963580e-01 -3.12264204e-01 -3.08994889e-01
-9.04949084e-02 4.60981548e-01 -2.34853357e-01 5.05645096e-01
-1.55029464e+00 1.44595957e+00 4.49884564e-01 4.93220180e-01
-5.29713333e-01 -4.44804102e-01 -5.66832960e-01 2.31241316e-01
-4.82131280e-02 -1.25476494e-01 9.51353133e-01 -8.82251382e-01
-1.06907868e+00 -9.32350662e-03 5.59808195e-01 -8.55644464e-01
4.82620180e-01 1.95070088e-01 -1.11929059e+00 5.92231572e-01
-9.82642546e-02 -6.97570920e-01 3.27387393e-01 -1.03280973e+00
-3.69265884e-01 -6.60997778e-02 9.57755148e-02 1.63399838e-02
1.69927459e-02 2.59916723e-01 1.19746995e+00 -9.11439955e-01
1.82875320e-01 -1.25717908e-01 -2.69915760e-01 -5.11129916e-01
-2.40904287e-01 -2.87094682e-01 1.10988986e+00 -7.01723039e-01
1.02103925e+00 -1.66467571e+00 4.78772894e-02 9.83882606e-01
-3.08775514e-01 3.99960726e-01 6.34458601e-01 1.57966149e+00
-4.86693561e-01 -2.02871084e-01 -3.75147909e-01 3.93147200e-01
-1.83913317e-02 4.17683542e-01 -3.74853790e-01 1.06227970e+00
3.15102518e-01 9.37204659e-02 -7.61262059e-01 1.79503798e-01
4.15496439e-01 6.01045787e-01 9.25156027e-02 1.22050770e-01
3.57044220e-01 1.69503450e-01 -1.68311536e-01 5.08102179e-01
1.01781261e+00 3.26790661e-01 3.14194500e-01 -1.93504333e-01
-2.61806071e-01 -5.22992527e-03 -1.61040044e+00 7.75543272e-01
2.29375362e-02 5.14954865e-01 2.45025009e-01 -1.69128227e+00
1.03097737e+00 8.79091322e-01 5.95290601e-01 -5.95267236e-01
2.06621483e-01 3.81583035e-01 5.23862958e-01 -2.46115416e-01
1.85666010e-01 -1.72587156e-01 2.07510084e-01 3.89824241e-01
4.89347100e-01 -3.87006819e-01 7.54571378e-01 3.18685055e-01
7.21092999e-01 -1.85473636e-02 6.68487668e-01 -1.02941525e+00
1.24491692e+00 -5.07149994e-02 5.24775684e-01 -1.42742753e-01
-7.10626319e-02 8.28492567e-02 7.97586322e-01 2.42941082e-02
-8.84513140e-01 -9.82357204e-01 -5.68394780e-01 -3.87915224e-01
1.46135151e-01 7.80721977e-02 -4.88348544e-01 -3.69035214e-01
1.28903672e-01 8.52602243e-01 -1.28588140e-01 -2.23155141e-01
-6.45344913e-01 -1.08308721e+00 7.84466922e-01 4.91849422e-01
3.64490420e-01 -5.23538291e-01 -6.52899623e-01 6.81297123e-01
4.48621392e-01 -5.30058742e-01 4.89057690e-01 3.72059345e-01
-7.46529281e-01 -1.43566406e+00 -9.00195062e-01 -6.58520222e-01
8.85969639e-01 1.52018778e-02 9.67296958e-01 6.91132963e-01
-5.54616630e-01 1.44891709e-01 -6.10657394e-01 -5.13763689e-02
-5.84869862e-01 -2.74060369e-01 1.23136081e-01 -3.54728818e-01
-4.24078256e-01 -1.00089276e+00 -4.38493699e-01 5.13337553e-01
-1.16347921e+00 -6.92734182e-01 9.55274776e-02 6.95059478e-01
-6.29474998e-01 9.32508230e-01 1.59469438e+00 -3.62349540e-01
9.31079626e-01 -8.18831682e-01 -1.12060344e+00 1.10863581e-01
-4.89940345e-01 -3.62440497e-01 1.07525873e+00 -4.68991660e-02
-9.50128019e-01 -8.03109467e-01 -6.45343542e-01 4.98064071e-01
-1.34526119e-01 6.18604898e-01 -3.86225283e-01 -6.67915404e-01
3.04668814e-01 6.06458746e-02 -1.86743159e-02 -5.39451838e-01
-2.63947129e-01 4.97482896e-01 3.53908151e-01 -3.55357260e-01
1.21533000e+00 1.70285106e-01 4.77254182e-01 -7.89475739e-01
2.14150101e-01 -3.80395710e-01 -4.13120747e-01 -4.32036877e-01
-1.32602215e-01 -5.68734229e-01 -8.66490841e-01 8.73349547e-01
-1.00386202e+00 2.25349128e-01 2.06234269e-02 2.72211462e-01
1.20970279e-01 7.31264591e-01 -6.06560409e-01 -1.03787494e+00
-4.28458929e-01 -1.06137216e+00 2.01279745e-01 3.06383967e-01
1.33508801e-01 -1.54120111e+00 3.69026251e-02 -6.87352300e-01
6.42443717e-01 6.92605019e-01 8.60350132e-01 -7.57996261e-01
-9.20510814e-02 -4.74670142e-01 2.98770130e-01 8.78287256e-01
4.57108527e-01 2.54937887e-01 -5.54360807e-01 -6.34555042e-01
4.45199758e-01 3.18361372e-01 -2.82458365e-01 -2.91136026e-01
1.20782256e-01 -9.56320390e-02 -1.80203170e-01 7.32672727e-03
2.10953522e+00 7.61381924e-01 4.77032632e-01 3.77593070e-01
-1.92241386e-01 2.59717256e-01 6.32512271e-01 5.38747013e-01
-1.45083908e-02 2.78423935e-01 5.31159401e-01 -8.99933353e-02
3.75358492e-01 6.55870676e-01 1.63849205e-01 5.37979603e-01
1.03818238e-01 -6.39170468e-01 -7.15202153e-01 7.32004762e-01
-1.46743882e+00 -8.70846272e-01 -6.48661494e-01 1.66158271e+00
3.78301740e-01 4.70278412e-01 1.79570280e-02 1.40368533e+00
7.10709393e-01 -1.31745219e-01 -3.21638994e-02 -6.46021307e-01
-8.06520760e-01 4.04464215e-01 2.83039361e-01 4.48486805e-01
-5.12451231e-01 -5.89632273e-01 5.54067087e+00 5.78657448e-01
-1.11723340e+00 -1.91994935e-01 1.79972768e-01 6.25443816e-01
-1.61329210e-01 2.46925019e-02 -1.51611999e-01 6.89817667e-01
9.59872961e-01 -1.03980958e+00 -3.13541740e-01 4.32421476e-01
2.60966748e-01 -8.22711229e-01 -4.73532915e-01 2.19407693e-01
1.29637644e-01 -6.47554457e-01 -1.87451318e-02 -2.67746925e-01
9.25521970e-01 -3.46838325e-01 -4.75410700e-01 -2.75273919e-01
-1.01980448e-01 -5.34746408e-01 5.67010701e-01 1.28973424e-01
-2.13758890e-02 -1.37461174e+00 1.33918083e+00 2.92329013e-01
-1.02370322e+00 -1.25960827e-01 -2.27498025e-01 -3.90989095e-01
9.51135397e-01 8.25659573e-01 -9.07632053e-01 1.70792127e+00
4.18659180e-01 2.37756476e-01 -3.18777204e-01 1.69722486e+00
-4.79939729e-01 8.21946383e-01 -4.43330914e-01 1.26668304e-01
9.24266502e-02 -3.78763735e-01 5.75491011e-01 6.64900064e-01
3.94526750e-01 -4.19350415e-02 -9.23693255e-02 3.66835982e-01
4.33808297e-01 -4.62064259e-02 -5.23930967e-01 4.94638264e-01
4.46457595e-01 1.13630104e+00 -9.72098589e-01 -4.59208757e-01
-3.08234006e-01 4.80636358e-01 -6.39943838e-01 5.20509779e-01
-6.92362785e-01 -9.07803893e-01 5.29706359e-01 1.49636552e-01
-3.13280463e-01 -3.50097179e-01 -7.59510621e-02 -9.08084214e-01
2.81968918e-02 -5.79338610e-01 4.16153044e-01 -7.78315008e-01
-1.24900532e+00 9.06090617e-01 5.36254168e-01 -1.53872848e+00
-7.45046616e-01 -3.44041497e-01 -9.61803794e-01 1.52233148e+00
-1.37459528e+00 -7.20170021e-01 1.61250338e-01 4.49415416e-01
3.04141134e-01 -1.13115042e-01 7.79153168e-01 6.21245563e-01
-4.93456841e-01 3.04749981e-02 3.81953031e-01 -4.68673697e-03
1.08817779e-01 -1.35487163e+00 1.41403258e-01 1.42561066e+00
-4.58848774e-01 3.00606042e-01 1.17497003e+00 -5.92991590e-01
-1.03043306e+00 -2.51083612e-01 6.16010249e-01 6.87447429e-01
8.31677735e-01 -1.78569093e-01 -1.03371406e+00 4.69473928e-01
1.25180483e+00 -7.93932676e-02 5.82446754e-01 -9.36104178e-01
2.75129288e-01 -1.45052090e-01 -2.04655361e+00 3.25039029e-02
-1.09170601e-01 -1.74592063e-01 -8.44480455e-01 1.13257878e-01
-4.53871265e-02 -4.99782152e-02 -1.18559897e+00 4.44860578e-01
-4.02011387e-02 -9.62635756e-01 7.06269383e-01 7.65852556e-02
-5.22126615e-01 -8.30906630e-01 3.27871770e-01 -1.96111739e+00
2.68530305e-02 -9.95522797e-01 6.90388680e-02 1.36400008e+00
8.19236934e-02 -1.52283585e+00 2.64607091e-02 -3.75163674e-01
-1.99370325e-01 -6.15152657e-01 -1.17136109e+00 -6.04134798e-01
3.92002538e-02 3.14229608e-01 9.37755287e-01 1.04385793e+00
6.06333673e-01 -1.59941494e-01 3.23725551e-01 7.18058586e-01
7.99685001e-01 -8.33758116e-02 -6.80424497e-02 -1.26105559e+00
-9.52211320e-02 -3.56106371e-01 -9.96678114e-01 5.40763279e-03
-1.23286769e-01 -3.21242571e-01 -6.00348830e-01 -1.77037585e+00
-8.87708604e-01 -1.32246584e-01 -2.98350722e-01 1.27898812e-01
-3.93829271e-02 4.07600015e-01 -2.02384949e-01 -1.84231430e-01
4.85056907e-01 4.21013534e-01 2.42060527e-01 -1.57484319e-02
6.81769967e-01 1.70897529e-01 1.89822793e-01 4.63504314e-01
1.26883292e+00 -1.00407921e-01 -8.39868426e-01 6.21683821e-02
1.76702812e-01 3.47188741e-01 -2.82239076e-03 -1.58467484e+00
1.37943745e-01 3.14634800e-01 4.42455649e-01 -9.27353144e-01
-2.05693379e-01 -1.26139975e+00 6.07831776e-01 9.73474503e-01
6.07508659e-01 8.62673581e-01 2.67135769e-01 -1.30281970e-01
-6.39216423e-01 -5.42388797e-01 5.56429267e-01 2.36287594e-01
-4.14964169e-01 -5.51383436e-01 -9.72026944e-01 -5.39611638e-01
1.50195158e+00 -2.39288434e-01 -5.61912596e-01 -1.85619384e-01
-6.29834771e-01 5.34745455e-01 5.26351869e-01 1.20236911e-01
4.84382421e-01 -8.83837163e-01 -6.35162771e-01 2.87972897e-01
-4.32795614e-01 -6.04451299e-01 2.13057294e-01 8.28317106e-01
-1.08249402e+00 3.41069788e-01 -7.28838205e-01 -2.92796582e-01
-1.23460150e+00 3.42952341e-01 8.35170031e-01 1.81685053e-02
-3.97441208e-01 3.15139085e-01 -6.31275594e-01 2.76744962e-01
-2.35537693e-01 -1.97443843e-01 -6.19378150e-01 2.77405739e-01
4.67074066e-01 8.89217913e-01 7.16757178e-01 -5.74856937e-01
-3.20293099e-01 2.24032383e-02 4.60364789e-01 -1.94591627e-01
1.53569186e+00 -2.35411584e-01 -7.18129158e-01 3.81267160e-01
8.09962869e-01 2.84769796e-02 -7.40751743e-01 6.43727362e-01
2.03102216e-01 -3.59100670e-01 -4.40990716e-01 -1.03494644e+00
-1.05608618e+00 3.61975580e-01 3.01307946e-01 1.35782146e+00
1.42293167e+00 -8.01751018e-01 1.23652451e-01 -6.42775893e-02
1.05977511e+00 -5.41818678e-01 -6.77486479e-01 -2.39207856e-02
7.59058952e-01 -2.67252713e-01 3.01753044e-01 -7.10743010e-01
7.88122043e-02 1.43266428e+00 1.35550991e-01 -6.42932415e-01
9.15477693e-01 9.21858728e-01 3.06384463e-04 1.69645809e-02
-6.21501207e-01 2.59017557e-01 -1.95093632e-01 9.00228620e-01
7.76272953e-01 -2.09279984e-01 -1.30007505e+00 -3.22126113e-02
7.85584152e-02 -2.24712670e-01 1.35615814e+00 1.46726918e+00
-2.97497988e-01 -1.56505334e+00 -7.04009056e-01 3.01257163e-01
-9.46716547e-01 4.79708254e-01 1.88084587e-01 1.00633001e+00
5.10169193e-03 1.27309072e+00 -1.30400479e-01 2.46027082e-01
6.24572814e-01 -1.80076912e-01 4.25264925e-01 -6.18447810e-02
-1.18595767e+00 -3.41844857e-02 2.25353539e-01 2.16747671e-01
-3.36903006e-01 -5.40017366e-01 -1.55637503e+00 5.67908073e-03
-6.47800982e-01 1.42573130e+00 7.23921537e-01 6.11749291e-01
-2.26143450e-01 8.69162679e-01 1.04500496e+00 -6.92855060e-01
-8.46961379e-01 -1.24190438e+00 -1.08969712e+00 -8.20999593e-02
1.55427068e-01 -7.95605123e-01 -7.03719676e-01 -5.05930841e-01] | [5.9120402336120605, 2.5434539318084717] |
51bf8564-76f1-4fad-a7ab-2c1781d1aaaa | scanpaths-in-reading-are-informative-about | null | null | https://aclanthology.org/W12-4904 | https://aclanthology.org/W12-4904.pdf | Scanpaths in reading are informative about sentence processing | null | ['Reinhold Kliegl', 'Titus von der Malsburg', 'Shravan Vasishth'] | 2012-12-01 | scanpaths-in-reading-are-informative-about-1 | https://aclanthology.org/W12-4904 | https://aclanthology.org/W12-4904.pdf | ws-2012-12 | ['human-parsing'] | ['computer-vision'] | [-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] | [-7.341312408447266, 3.7039685249328613] |
5587994e-dd04-4bb8-b1b0-ae47f3688589 | robust-object-detection-with-multi-input | 2111.13065 | null | https://arxiv.org/abs/2111.13065v1 | https://arxiv.org/pdf/2111.13065v1.pdf | Robust Object Detection with Multi-input Multi-output Faster R-CNN | Recent years have seen impressive progress in visual recognition on many benchmarks, however, generalization to the real-world in out-of-distribution setting remains a significant challenge. A state-of-the-art method for robust visual recognition is model ensembling. however, recently it was shown that similarly competitive results could be achieved with a much smaller cost, by using multi-input multi-output architecture (MIMO). In this work, a generalization of the MIMO approach is applied to the task of object detection using the general-purpose Faster R-CNN model. It was shown that using the MIMO framework allows building strong feature representation and obtains very competitive accuracy when using just two input/output pairs. Furthermore, it adds just 0.5\% additional model parameters and increases the inference time by 15.9\% when compared to the standard Faster R-CNN. It also works comparably to, or outperforms the Deep Ensemble approach in terms of model accuracy, robustness to out-of-distribution setting, and uncertainty calibration when the same number of predictions is used. This work opens up avenues for applying the MIMO approach in other high-level tasks such as semantic segmentation and depth estimation. | ['Andrzej Czyzewski', 'Sebastian Cygert'] | 2021-11-25 | null | null | null | null | ['robust-object-detection'] | ['computer-vision'] | [ 2.58036494e-01 1.69180200e-01 -3.43825333e-02 -5.33787012e-01
-1.20455372e+00 -4.07580316e-01 6.82723403e-01 1.75453827e-01
-6.45102918e-01 5.92880130e-01 -4.65629250e-01 -1.50246456e-01
3.23775448e-02 -6.16518140e-01 -1.17470086e+00 -8.05383801e-01
1.61524087e-01 6.83477402e-01 4.48085964e-01 -8.08010518e-04
2.08081484e-01 6.81664050e-01 -1.91391218e+00 2.20068038e-01
3.06701362e-01 1.48645627e+00 -1.79836079e-01 8.95715058e-01
1.63440540e-01 4.96564090e-01 -6.27815306e-01 -4.25458133e-01
3.61225814e-01 2.14717075e-01 -5.73109746e-01 -1.63357735e-01
8.83151829e-01 -3.25187057e-01 -3.32355529e-01 7.17048883e-01
7.03244269e-01 2.18663529e-01 7.22011566e-01 -1.04107797e+00
-4.06646073e-01 4.10486490e-01 -7.29782820e-01 1.34548739e-01
-1.00193225e-01 9.16643515e-02 7.01965570e-01 -7.20924079e-01
4.06270921e-01 1.27352822e+00 8.34235489e-01 6.02952600e-01
-1.40141404e+00 -6.72752261e-01 1.45788491e-01 6.20403998e-02
-1.33868909e+00 -2.37722427e-01 1.97005078e-01 -5.07440031e-01
1.57997239e+00 3.16576511e-02 2.89754480e-01 1.24953759e+00
1.37854189e-01 9.32035625e-01 1.23973358e+00 -3.29615086e-01
2.87124425e-01 1.84004113e-01 3.78676325e-01 5.33960581e-01
1.49146080e-01 3.67204189e-01 -3.63055021e-01 1.32808283e-01
7.64436066e-01 -4.19693679e-01 2.07587808e-01 -3.41207236e-01
-9.64787424e-01 7.84154952e-01 6.37286782e-01 1.24392055e-01
1.96931332e-01 6.57535970e-01 6.10049546e-01 5.06970957e-02
6.65313482e-01 2.68341005e-01 -5.99467993e-01 -1.87825426e-01
-1.12080920e+00 2.52610385e-01 5.19395232e-01 8.59248161e-01
5.05426705e-01 1.92338958e-01 -9.20758322e-02 7.79287577e-01
3.67282361e-01 4.94442046e-01 3.02204758e-01 -9.52477515e-01
4.20397758e-01 2.82327324e-01 -3.28088626e-02 -6.18353426e-01
-7.12880552e-01 -6.83868945e-01 -1.02530682e+00 7.35113025e-01
7.62271702e-01 3.63601036e-02 -1.51333988e+00 1.57834053e+00
1.24831766e-01 1.00709550e-01 4.05239910e-02 6.31552875e-01
8.19653690e-01 5.23288310e-01 -2.24569067e-01 4.03863519e-01
1.12158620e+00 -7.47215271e-01 -1.34960487e-01 -2.23581821e-01
2.26703227e-01 -4.97754425e-01 6.15922391e-01 7.90998101e-01
-9.56465006e-01 -6.87736392e-01 -1.23778915e+00 -1.72542632e-01
-6.19153082e-01 3.81236583e-01 6.83467090e-01 8.25395226e-01
-1.02275872e+00 8.81671548e-01 -9.50389981e-01 -2.87639558e-01
6.75629735e-01 7.90388107e-01 -5.06519735e-01 -4.06205133e-02
-8.30045044e-01 1.04261363e+00 4.88948703e-01 1.77729324e-01
-1.09949827e+00 -7.28900790e-01 -1.08110428e+00 -2.10867658e-01
3.25360775e-01 -6.03864133e-01 1.33480597e+00 -8.87243271e-01
-1.59287965e+00 9.14574802e-01 2.38910969e-02 -8.17343116e-01
8.77376139e-01 -3.46037716e-01 -8.53525475e-02 -2.64483392e-01
-2.68245369e-01 1.06052661e+00 8.96078169e-01 -1.15099156e+00
-4.30809051e-01 -6.20374739e-01 6.96295425e-02 -1.66052014e-01
2.46185720e-01 7.58050233e-02 -6.51261032e-01 -5.40370345e-01
-3.37000680e-03 -9.25399244e-01 -2.79259175e-01 1.95781663e-01
-3.14732999e-01 -8.39138106e-02 3.73565495e-01 -4.96786803e-01
7.39413202e-01 -1.91278660e+00 2.63439834e-01 2.89037406e-01
5.21196313e-02 2.63368309e-01 -1.23244023e-03 5.44414036e-02
-8.29786509e-02 7.88804367e-02 -4.26953614e-01 -6.55658484e-01
-3.56275477e-02 3.10169548e-01 -1.43481329e-01 5.27230978e-01
3.81109297e-01 8.80236924e-01 -4.96577263e-01 -2.18358889e-01
5.24269104e-01 6.23340011e-01 -2.12278873e-01 -7.96642229e-02
-2.68666714e-01 4.33042854e-01 8.09626728e-02 5.70178747e-01
6.39528871e-01 -2.89796680e-01 -2.46834472e-01 -2.19615661e-02
8.02285746e-02 -2.22408459e-01 -1.30177498e+00 1.62630534e+00
-6.99741066e-01 1.01389670e+00 -1.85133591e-01 -1.14190090e+00
1.00380838e+00 9.02147815e-02 1.19260818e-01 -5.88143945e-01
3.60679299e-01 2.92671233e-01 8.63608867e-02 -2.64058113e-02
3.66192043e-01 -6.75381795e-02 -3.48792858e-02 -1.28679559e-01
2.80941695e-01 -3.44342440e-01 4.48125377e-02 -1.37858063e-01
8.69069636e-01 3.51978987e-01 1.90736964e-01 -1.23453893e-01
3.80454749e-01 -2.13749573e-01 2.78095424e-01 9.01325583e-01
-1.05951063e-01 1.00618494e+00 4.09244239e-01 -3.42806190e-01
-1.18928063e+00 -9.99264002e-01 -6.18523419e-01 8.19978595e-01
1.19599549e-03 -6.96610287e-02 -7.91967034e-01 -6.13418460e-01
4.17754740e-01 5.93879819e-01 -1.00812423e+00 2.18600426e-02
-3.80789310e-01 -9.93088424e-01 9.09362555e-01 1.17920852e+00
4.68569010e-01 -7.68139005e-01 -6.77748382e-01 6.39780313e-02
3.93176615e-01 -1.24567485e+00 1.62821069e-01 4.42014635e-01
-8.38508725e-01 -8.81262064e-01 -1.01471114e+00 -2.06757605e-01
3.86484563e-01 -2.88451105e-01 1.15282845e+00 -2.98743308e-01
-5.68369091e-01 3.42594087e-01 -6.42289519e-02 -6.33139789e-01
-1.98167101e-01 -8.70619491e-02 -4.75843772e-02 -1.89460441e-01
2.66395926e-01 -1.53521836e-01 -3.96417797e-01 3.77176106e-01
-8.76038492e-01 -4.94092591e-02 5.07478952e-01 1.09576654e+00
7.25515246e-01 -2.20620915e-01 4.17161286e-01 -6.84710085e-01
-1.08860202e-01 -3.37032288e-01 -9.56696332e-01 1.70885056e-01
-6.75693393e-01 3.88039172e-01 4.35605675e-01 -2.92591035e-01
-9.74953294e-01 2.50641078e-01 -4.29881543e-01 -6.05548084e-01
-2.66411036e-01 2.00461801e-02 2.54586339e-02 -2.48828202e-01
7.54771829e-01 1.91802979e-02 -3.89595516e-02 -4.48571235e-01
4.54983383e-01 5.56381583e-01 4.66545701e-01 -3.53808284e-01
5.03001630e-01 6.01819336e-01 3.92120689e-01 -7.29352772e-01
-6.55879736e-01 -4.97352719e-01 -6.83155954e-01 -7.61071071e-02
8.70391726e-01 -9.73697722e-01 -6.76385999e-01 8.68515134e-01
-1.19294119e+00 -5.28591216e-01 -2.73554921e-01 4.07202452e-01
-8.04139555e-01 3.24842393e-01 -3.55180681e-01 -1.00541055e+00
-1.42987266e-01 -1.33840871e+00 1.34768724e+00 2.17230588e-01
1.27275646e-01 -7.82480359e-01 -5.47493771e-02 4.33242470e-01
3.71856630e-01 3.61193597e-01 5.96530020e-01 -6.41585350e-01
-5.03449976e-01 -4.08596873e-01 -4.91587311e-01 7.95090079e-01
-4.12496448e-01 2.54166543e-01 -1.42689025e+00 -2.52536595e-01
-5.19105077e-01 -5.68610311e-01 1.37256825e+00 7.59325802e-01
1.47446024e+00 3.04905564e-01 -2.71878630e-01 7.32736349e-01
1.48968709e+00 3.93876852e-03 7.14716554e-01 3.03067535e-01
6.13689780e-01 5.94515502e-01 5.77760935e-01 2.51926273e-01
1.79918632e-01 9.01735246e-01 6.31843328e-01 -1.27738416e-01
-1.77464858e-01 5.35215717e-03 2.65748203e-01 8.77404884e-02
-2.07771823e-01 -2.80309498e-01 -1.19158733e+00 4.08836007e-01
-1.98559332e+00 -6.77710474e-01 -1.70635164e-01 2.37556148e+00
3.80879670e-01 4.49698836e-01 1.40034989e-01 1.14373632e-01
5.20093679e-01 -2.24520132e-01 -7.90114284e-01 -5.15199900e-01
-2.47167647e-01 4.11748767e-01 9.57930863e-01 4.33078617e-01
-1.25157773e+00 6.89212143e-01 6.75290394e+00 1.13402164e+00
-1.15954483e+00 1.04211979e-01 1.08212936e+00 -1.97254941e-01
9.51131955e-02 -3.09691817e-01 -1.14952111e+00 2.67794877e-01
1.12823486e+00 3.13144475e-01 1.73218563e-01 1.08652139e+00
-2.71522701e-01 -5.08264244e-01 -1.30930126e+00 1.35346448e+00
2.87528664e-01 -1.28265131e+00 -2.54823357e-01 -1.75696369e-02
7.76964247e-01 5.26222885e-01 2.31935725e-01 3.41341436e-01
1.96428195e-01 -1.50474370e+00 6.96054637e-01 6.65457070e-01
9.24437225e-01 -8.24022412e-01 1.05865622e+00 4.53250557e-01
-1.03567421e+00 -7.19678309e-03 -4.92423594e-01 7.91919902e-02
-1.08883396e-01 5.51584065e-01 -5.66875100e-01 6.42998934e-01
1.03424466e+00 5.76816618e-01 -7.47960865e-01 1.14253092e+00
1.68974057e-01 2.39561751e-01 -7.17642128e-01 1.49630204e-01
1.78296909e-01 2.14472011e-01 2.64079183e-01 1.42530870e+00
1.89846426e-01 -3.50301266e-01 -2.28894860e-01 6.96580350e-01
2.45439517e-03 -1.46224663e-01 -6.30324423e-01 3.21210861e-01
-1.57410160e-01 1.17779946e+00 -8.40452075e-01 -4.56437141e-01
-2.03299448e-01 8.75838220e-01 4.42678005e-01 8.81611090e-03
-9.34851527e-01 -2.68465012e-01 4.89879727e-01 -1.30568296e-01
5.66828966e-01 -6.23863842e-03 -4.16194469e-01 -9.91708994e-01
7.40924329e-02 -6.99722111e-01 2.94073790e-01 -7.44517624e-01
-1.39979041e+00 6.63990438e-01 2.25585908e-01 -1.00883567e+00
-3.91802102e-01 -1.34081709e+00 -3.04592013e-01 8.65706444e-01
-1.46771598e+00 -1.19073617e+00 -3.48740309e-01 2.34179810e-01
5.50600886e-01 -1.25493735e-01 7.83128202e-01 1.39497831e-01
-6.26644909e-01 8.30563009e-01 3.81363869e-01 -8.60609338e-02
5.62399030e-01 -1.64219558e+00 5.56045413e-01 7.06852376e-01
3.94046456e-01 1.77475706e-01 6.49637878e-01 -3.44794571e-01
-1.02272832e+00 -1.02915609e+00 3.01989913e-01 -7.82182157e-01
5.09052098e-01 -5.37152886e-01 -7.44164646e-01 6.06258631e-01
2.86620529e-03 3.58286828e-01 4.81752068e-01 1.41509145e-01
-6.78542435e-01 -1.97995201e-01 -1.29857564e+00 2.10486799e-01
8.19506705e-01 -5.41988313e-01 -3.43594283e-01 2.13150784e-01
3.81247163e-01 -7.47556031e-01 -1.03626049e+00 7.73794711e-01
8.40820312e-01 -1.22476947e+00 1.01626146e+00 -4.64487970e-01
4.01467592e-01 -1.89572647e-01 -3.61254305e-01 -1.07421947e+00
4.74656969e-02 -2.69987553e-01 -2.64308125e-01 1.09255910e+00
5.93796313e-01 -5.57682097e-01 8.21207285e-01 5.73597193e-01
-9.37447399e-02 -1.03286302e+00 -1.15379548e+00 -1.02524674e+00
4.65184957e-01 -1.00745642e+00 2.31097087e-01 7.35289082e-02
-5.27610004e-01 1.34278864e-01 -3.71312410e-01 1.57422349e-01
6.70014977e-01 -2.17449129e-01 8.84038925e-01 -1.26389945e+00
-6.67105913e-01 -5.90222895e-01 -8.39863658e-01 -1.06254828e+00
1.59066871e-01 -7.32383370e-01 1.58020273e-01 -1.31406093e+00
7.80133754e-02 -4.47764844e-01 -4.47465718e-01 3.99629086e-01
1.53403938e-01 5.41217744e-01 2.72840112e-01 -9.67804044e-02
-4.20526236e-01 3.85834754e-01 7.56529152e-01 -2.90120035e-01
2.34214813e-01 3.08002681e-01 -2.25596845e-01 7.43050277e-01
7.41683125e-01 -3.00041825e-01 -1.88350946e-01 -3.34225535e-01
1.67379484e-01 -3.24356146e-02 6.62173092e-01 -1.25547945e+00
2.34844685e-02 2.91324288e-01 7.68196404e-01 -6.75448716e-01
7.38350868e-01 -8.12289178e-01 1.07722491e-01 2.37566978e-01
-2.50304222e-01 -1.02532066e-01 8.74218345e-01 7.17588365e-01
-1.31501332e-01 -3.05685043e-01 1.11991549e+00 -4.97338213e-02
-7.74325192e-01 1.08809218e-01 -1.41508609e-01 -1.43824115e-01
1.19361722e+00 -5.21040857e-01 -2.30220512e-01 -1.34806380e-01
-9.29165006e-01 -6.72506616e-02 3.60733330e-01 4.29073393e-01
4.17125642e-01 -8.90790403e-01 -5.88375866e-01 9.15711299e-02
3.01707268e-01 2.50591367e-01 4.25739169e-01 7.87334204e-01
-4.59323138e-01 6.06202662e-01 -1.69577077e-01 -1.15181518e+00
-1.35793817e+00 3.44130158e-01 6.15607142e-01 -3.15244317e-01
-4.63201702e-01 1.19213104e+00 1.96455881e-01 -3.81022334e-01
5.05052507e-01 -6.29325330e-01 7.20228162e-03 3.75001393e-02
5.07445991e-01 5.08060157e-01 4.39629376e-01 -6.23259783e-01
-5.05597770e-01 8.83987248e-01 -6.69878647e-02 -1.42950594e-01
1.24395895e+00 2.01611042e-01 3.30532491e-01 7.82244503e-01
1.13878059e+00 -4.56520498e-01 -1.66722894e+00 9.06543881e-02
-1.06371008e-01 -3.86107743e-01 2.17158660e-01 -1.03354478e+00
-1.11156654e+00 1.28219533e+00 9.82546985e-01 -9.14501995e-02
8.75293851e-01 1.17400981e-01 1.06212728e-01 4.72970933e-01
4.50394452e-01 -1.12412643e+00 -1.49952382e-01 3.95953357e-01
9.84855711e-01 -1.73840106e+00 -9.33777988e-02 -2.29861066e-01
-6.70883060e-01 1.22641265e+00 8.21180761e-01 -3.19235951e-01
7.03105986e-01 4.13572758e-01 -2.88907606e-02 4.35944907e-02
-6.46026254e-01 -2.62127787e-01 5.75050414e-01 8.04782152e-01
3.47132206e-01 -3.82544361e-02 3.86618793e-01 2.42673114e-01
5.84847108e-02 -8.66230354e-02 3.53420377e-01 6.39461398e-01
-2.05640659e-01 -9.68806565e-01 -2.07697570e-01 6.74506783e-01
-4.70258713e-01 -7.44556263e-02 -1.20791666e-01 1.00085580e+00
1.84823737e-01 8.08625638e-01 3.19487184e-01 -2.35533446e-01
2.78215736e-01 1.10469297e-01 8.70395124e-01 -5.89655757e-01
-7.78209627e-01 -2.02366069e-01 -5.74357472e-02 -8.12157929e-01
-2.98897594e-01 -5.96638858e-01 -1.17774475e+00 -4.39888313e-02
-6.41314924e-01 -4.24098969e-01 9.97280836e-01 9.08390582e-01
3.05471450e-01 5.92966914e-01 1.07065588e-01 -1.18204832e+00
-8.56243670e-01 -1.10974586e+00 -4.40692574e-01 7.84244156e-04
4.42357570e-01 -9.75515902e-01 -3.16741377e-01 -3.07130367e-01] | [8.755318641662598, 1.937585473060608] |
09977628-0cbf-45c5-8441-46e457b2fdf1 | emergency-response-person-localization-and | 2305.15795 | null | https://arxiv.org/abs/2305.15795v1 | https://arxiv.org/pdf/2305.15795v1.pdf | Emergency Response Person Localization and Vital Sign Estimation Using a Semi-Autonomous Robot Mounted SFCW Radar | The large number and scale of natural and man-made disasters have led to an urgent demand for technologies that enhance the safety and efficiency of search and rescue teams. Semi-autonomous rescue robots are beneficial, especially when searching inaccessible terrains, or dangerous environments, such as collapsed infrastructures. For search and rescue missions in degraded visual conditions or non-line of sight scenarios, radar-based approaches may contribute to acquire valuable, and otherwise unavailable information. This article presents a complete signal processing chain for radar-based multi-person detection, 2D-MUSIC localization and breathing frequency estimation. The proposed method shows promising results on a challenging emergency response dataset that we collected using a semi-autonomous robot equipped with a commercially available through-wall radar system. The dataset is composed of 62 scenarios of various difficulty levels with up to five persons captured in different postures, angles and ranges including wooden and stone obstacles that block the radar line of sight. Ground truth data for reference locations, respiration, electrocardiogram, and acceleration signals are included. The full emergency response benchmark data set as well as all codes to reproduce our results, are publicly available at https://doi.org/10.21227/4bzd-jm32. | ['Michael Muma', 'Abdelhak M. Zoubir', 'Oskar von Stryk', 'Stefan Fabian', 'Ibrahim Kakouche', 'Christian Eckrich', 'Christian A. Schroth'] | 2023-05-25 | null | null | null | null | ['human-detection'] | ['computer-vision'] | [ 3.66776794e-01 -4.54744637e-01 5.80886483e-01 -1.75622016e-01
-7.00135648e-01 -4.43842590e-01 1.57736182e-01 9.94375944e-02
-6.76425874e-01 9.34619784e-01 3.44702631e-01 9.20545831e-02
-6.66832507e-01 -5.85384488e-01 5.42686842e-02 -9.05560911e-01
-5.47799110e-01 6.45598531e-01 -1.28293306e-01 -4.90988493e-01
-3.69093232e-02 7.86010742e-01 -1.33848882e+00 1.92851685e-02
7.64455914e-01 7.77405024e-01 3.14243704e-01 7.23304451e-01
8.71994734e-01 2.42567763e-01 -5.75528622e-01 9.63083759e-04
5.75440407e-01 -3.27535309e-02 -1.57379001e-01 -3.30279022e-01
-5.84572367e-02 -4.87456053e-01 -6.16109312e-01 6.02340817e-01
1.27486622e+00 2.19426826e-01 6.05265260e-01 -1.15046227e+00
4.82316595e-03 7.85393566e-02 -6.55080557e-01 5.77586472e-01
8.37522924e-01 2.66076356e-01 1.09345190e-01 -7.66622007e-01
2.71289110e-01 8.44395638e-01 1.11442161e+00 2.88087517e-01
-7.92463362e-01 -4.45640981e-01 -5.82991779e-01 5.90989947e-01
-1.79271758e+00 -4.64416683e-01 4.34685171e-01 -3.36242527e-01
7.62304187e-01 2.00805202e-01 5.24909556e-01 1.40321422e+00
2.90823251e-01 -2.90862788e-02 9.81234670e-01 -4.56576832e-02
1.60941616e-01 -4.33233589e-01 -6.19378053e-02 3.67244422e-01
8.77286434e-01 3.08428615e-01 -6.78918302e-01 -4.66937184e-01
3.22694331e-01 3.15354079e-01 -8.69144797e-01 -1.43182725e-01
-1.47224808e+00 5.93556523e-01 2.86423624e-01 2.30301842e-01
-1.07406807e+00 -2.51269072e-01 2.84621626e-01 1.66151032e-01
-5.59334829e-02 8.03987607e-02 -2.50289440e-01 -4.70919609e-01
-6.71967506e-01 3.07597935e-01 7.77484715e-01 5.38088202e-01
2.14622587e-01 3.15220833e-01 -2.06260048e-02 5.93675315e-01
1.52510196e-01 1.35553062e+00 2.64613032e-01 -7.67465830e-01
6.42682254e-01 2.03753814e-01 4.85378593e-01 -1.53972447e+00
-1.19838440e+00 -6.06255412e-01 -1.35219991e+00 -4.50643264e-02
2.04953060e-01 -4.57499266e-01 -4.85898733e-01 1.36215937e+00
4.88248974e-01 7.65849650e-02 2.09221289e-01 1.39819062e+00
8.39986086e-01 2.87948012e-01 -2.97764152e-01 -4.54103202e-01
1.43615508e+00 -1.65870473e-01 -7.50005960e-01 -4.15644526e-01
2.44585961e-01 -4.02489811e-01 6.22935414e-01 7.11888194e-01
-5.65248668e-01 -3.16689759e-01 -9.09945071e-01 6.34080946e-01
3.26714255e-02 3.25777539e-04 1.86839715e-01 6.29981697e-01
-7.07062125e-01 2.64156371e-01 -7.85642862e-01 -6.20813787e-01
1.02197386e-01 4.96843942e-02 -3.81024122e-01 -6.01999819e-01
-1.41635680e+00 1.09978151e+00 -1.79943249e-01 6.46741033e-01
-8.00644279e-01 -3.01926911e-01 -8.44257653e-01 -2.84382343e-01
1.91928744e-01 -7.12210834e-01 5.88887453e-01 2.81075478e-01
-6.44194186e-01 4.57662970e-01 3.08472335e-01 -5.73600471e-01
5.54009140e-01 -4.59921449e-01 -7.31798828e-01 6.16957247e-01
2.93121159e-01 -2.05740728e-03 7.33724952e-01 -8.21171403e-01
-2.36267760e-01 -7.12436378e-01 -5.85760713e-01 4.96827275e-01
9.00688395e-02 7.89599586e-03 5.03164887e-01 -5.37025452e-01
2.84353226e-01 -8.31446111e-01 -2.42796645e-01 -3.86159778e-01
-3.02048087e-01 5.52879512e-01 4.58134502e-01 -8.50214303e-01
8.79029214e-01 -1.93238008e+00 -2.76280139e-02 1.96270689e-01
-2.06942931e-01 8.41783732e-02 1.51342139e-01 8.09289992e-01
2.33614996e-01 -3.94652635e-01 -5.62219977e-01 8.57264027e-02
-2.13202581e-01 -6.25420781e-03 -2.01267287e-01 1.03005862e+00
-4.55506235e-01 3.88665557e-01 -5.85405707e-01 -1.38728246e-01
2.44696900e-01 8.02202165e-01 1.26944352e-02 1.58191496e-03
8.06768179e-01 8.23250234e-01 -4.52786088e-01 1.01694894e+00
7.47129202e-01 3.89135122e-01 -4.60545808e-01 -1.75662175e-01
-1.96486294e-01 -1.55469686e-01 -1.58559299e+00 1.52645051e+00
3.15053426e-02 4.88781065e-01 4.72759008e-01 -9.21640515e-01
9.88040745e-01 5.01674712e-01 7.06457973e-01 -7.64080167e-01
1.10292941e-01 1.17165312e-01 -2.52212077e-01 -9.51660156e-01
4.78599310e-01 -2.97390133e-01 -4.48511422e-01 3.88624460e-01
-4.88574296e-01 -1.42901883e-01 6.88947663e-02 -8.12464859e-03
1.78443313e+00 -2.60282278e-01 4.86401856e-01 -5.69587946e-02
2.17008337e-01 1.79595485e-01 6.27697766e-01 9.51195776e-01
-5.51075816e-01 6.30771220e-01 -4.62785155e-01 -6.18368328e-01
-4.01709855e-01 -1.29462123e+00 -2.89209962e-01 4.59551990e-01
2.55594015e-01 -3.51263955e-02 -2.10212857e-01 1.36528000e-01
-8.31227973e-02 5.69446862e-01 -2.13233232e-01 -2.84726024e-01
-5.73929310e-01 -9.27203119e-01 9.97348309e-01 1.08095475e-01
7.60014176e-01 -1.21182525e+00 -1.51387668e+00 2.26798952e-01
-9.58177626e-01 -1.29239738e+00 1.10043973e-01 7.99666718e-02
-6.53125286e-01 -1.24778223e+00 -7.44307935e-01 -1.92540944e-01
3.62708360e-01 8.04810345e-01 6.68238521e-01 -2.88815182e-02
-1.11205542e+00 8.26166272e-01 -5.06447554e-01 -3.07623923e-01
3.42663676e-01 -4.52301145e-01 9.11209702e-01 -1.90445960e-01
1.73962981e-01 -7.60042071e-01 -7.54482329e-01 5.58102787e-01
-3.43279630e-01 -4.60414290e-01 6.02419794e-01 5.63121438e-01
1.87212467e-01 2.40747303e-01 6.61144435e-01 1.89228058e-01
7.11116076e-01 -6.74188852e-01 -1.51772052e-01 -1.64227024e-01
-9.76877138e-02 -5.32073438e-01 3.56014222e-01 -2.07880646e-01
-7.10469484e-01 -8.86768103e-02 2.01714441e-01 -6.44617528e-03
-6.90784931e-01 5.08667529e-01 7.86480531e-02 2.29509979e-01
1.07181883e+00 2.96161115e-01 -3.32213223e-01 -2.27371156e-01
-1.22664355e-01 9.13470984e-01 1.08689809e+00 -5.84075600e-02
1.00216830e+00 7.27962196e-01 1.38485894e-01 -1.37614226e+00
-4.02846456e-01 -7.35998929e-01 -5.86857498e-01 -4.62533981e-01
7.59046674e-01 -1.05002427e+00 -7.10998833e-01 4.66729432e-01
-9.64501023e-01 -1.22345477e-01 1.98361352e-02 1.01030529e+00
-4.53650951e-01 5.91395974e-01 -3.00685525e-01 -1.23523355e+00
-7.20765829e-01 -3.90615791e-01 8.73779356e-01 2.02864617e-01
-3.28598320e-01 -3.26306671e-01 2.98204392e-01 6.67033434e-01
6.26549363e-01 7.80686378e-01 1.29675105e-01 -3.23324949e-01
-1.47049874e-01 -3.83757681e-01 2.16619477e-01 -3.05655092e-01
1.92983761e-01 -7.93879986e-01 -6.10990703e-01 -5.72807014e-01
4.16171700e-01 -3.94919366e-01 6.14990532e-01 6.42553270e-01
1.23983368e-01 -1.54420242e-01 -3.99470389e-01 6.22713089e-01
9.67629611e-01 1.48915440e-01 5.53890109e-01 3.84010255e-01
2.33033359e-01 7.37053752e-01 1.00694311e+00 1.26181102e+00
4.12841469e-01 6.80451274e-01 5.40871561e-01 3.15593064e-01
3.07338715e-01 3.75780523e-01 3.10570300e-01 3.97553474e-01
-4.31513399e-01 -2.85359025e-01 -1.18130136e+00 4.75784302e-01
-1.40666354e+00 -1.43177056e+00 -4.76408809e-01 2.23184156e+00
1.36674076e-01 -2.11231336e-01 -1.33203626e-01 5.59130967e-01
6.09568357e-01 -1.73957452e-01 -4.41873521e-01 7.76215047e-02
-1.34700269e-01 -1.89463630e-01 5.26597023e-01 3.19203138e-01
-1.04604888e+00 2.15520173e-01 5.03776646e+00 1.42663732e-01
-7.92268574e-01 2.33244244e-02 -4.03324336e-01 -2.25785986e-01
2.84712911e-01 -5.09320199e-01 -3.88028532e-01 2.06898198e-01
9.25658464e-01 2.24532813e-01 4.37269956e-01 3.94312799e-01
6.28502965e-01 -6.36825442e-01 -3.93533409e-01 1.34494412e+00
2.28529200e-01 -4.83582765e-01 -7.82377481e-01 -1.55244041e-02
-1.57047305e-02 2.58432835e-01 -2.79786676e-01 1.07715316e-01
-1.70521393e-01 -8.08861792e-01 4.46411610e-01 8.47110868e-01
7.10596502e-01 -6.99769616e-01 8.69649887e-01 7.50729024e-01
-1.21718919e+00 -2.66148657e-01 -3.58804196e-01 -3.53331417e-01
8.30501258e-01 6.98037982e-01 -7.47462511e-01 7.68881083e-01
1.18710041e+00 2.50580251e-01 -2.36733660e-01 1.39748609e+00
-3.63097459e-01 4.56461072e-01 -6.97500288e-01 2.96975344e-01
-2.98667669e-01 -1.00325823e-01 1.02012062e+00 9.23164010e-01
8.52797151e-01 7.86882997e-01 1.51869461e-01 2.59806633e-01
7.27932632e-01 -3.38903934e-01 -1.00130057e+00 4.58375603e-01
6.85086966e-01 1.32510209e+00 -7.35183477e-01 1.06444716e-01
-8.57514888e-03 9.02907670e-01 -4.06256467e-01 2.95459062e-01
-1.00523734e+00 -6.58921003e-01 3.63314420e-01 1.67500705e-01
7.59405922e-03 -7.24271476e-01 -1.78684801e-01 -9.43471134e-01
8.38906318e-02 -8.04572284e-01 5.41318715e-01 -9.50161815e-01
-9.17012930e-01 8.45368505e-01 2.90003061e-01 -1.55357909e+00
-2.93022484e-01 -1.77239805e-01 -3.80743742e-01 5.81175923e-01
-1.06101251e+00 -7.04979062e-01 -1.01601195e+00 8.86027932e-01
3.65699589e-01 -3.61537158e-01 1.10642457e+00 5.15841305e-01
-2.77355403e-01 -9.30361822e-02 -8.05740580e-02 -2.36935183e-01
6.95203722e-01 -4.50885624e-01 -4.32694145e-02 8.40306520e-01
-2.09197819e-01 3.25677425e-01 1.16711414e+00 -1.05672610e+00
-1.60094738e+00 -7.68472672e-01 6.65604174e-01 -8.48526582e-02
2.69303113e-01 -2.39920214e-01 -5.26516020e-01 2.35942021e-01
-1.78481713e-01 -3.76648568e-02 6.09531462e-01 -3.37481678e-01
4.21402097e-01 3.33402045e-02 -1.45381200e+00 3.65901679e-01
1.05840409e+00 2.18172111e-02 -7.91313469e-01 3.39860052e-01
-2.72285603e-02 -3.14266920e-01 -4.72723961e-01 6.54632688e-01
6.43854260e-01 -9.53633904e-01 1.27666998e+00 7.68932030e-02
-4.83004659e-01 -2.56476134e-01 -4.95880723e-01 -1.29756570e+00
-3.25876564e-01 -6.08335555e-01 5.71753122e-02 6.33623779e-01
-1.54548362e-01 -6.70332789e-01 4.37563479e-01 1.43125638e-01
-2.04024911e-01 -3.40659171e-01 -1.30788970e+00 -8.17064583e-01
-8.85534048e-01 -5.86116850e-01 -5.40665761e-02 6.21492684e-01
3.57645787e-02 2.00183257e-01 -7.18802691e-01 6.54825568e-01
1.01616216e+00 -8.16429127e-03 5.45262039e-01 -1.54641211e+00
-8.15634802e-02 3.51023823e-01 -6.19961143e-01 -4.27548528e-01
-4.09434855e-01 -4.91872013e-01 2.53697544e-01 -1.87032032e+00
-9.45054144e-02 -4.45180506e-01 2.31455058e-01 4.88793761e-01
2.65530229e-01 4.25480098e-01 1.26050301e-02 3.11633348e-01
-3.91413838e-01 7.25397408e-01 6.50380611e-01 -9.55460966e-02
-1.90652549e-01 3.77009779e-01 -4.35774893e-01 9.42287803e-01
9.60359752e-01 -8.23347270e-01 -2.78451651e-01 -3.56593311e-01
1.48777694e-01 6.07110798e-01 7.31649697e-01 -1.68867552e+00
5.35034180e-01 -1.36587620e-01 5.15286326e-01 -7.79421628e-01
8.91548693e-01 -8.76027465e-01 4.32059675e-01 9.79253292e-01
2.07586035e-01 3.98902953e-01 8.23102891e-02 7.04799473e-01
1.15651615e-01 -1.98986322e-01 4.52547133e-01 8.33968967e-02
-4.33473945e-01 1.22708462e-01 -7.96140611e-01 1.66916966e-01
9.97257292e-01 -3.83574575e-01 -4.63096440e-01 -7.60539532e-01
-8.39100182e-01 3.28703851e-01 5.05689159e-03 2.14991003e-01
1.14745665e+00 -1.15634930e+00 -1.11051595e+00 2.27238521e-01
-6.13962635e-02 -2.35154599e-01 7.50504255e-01 1.43096554e+00
-4.52461809e-01 2.96098530e-01 -5.96986234e-01 -4.93826181e-01
-1.29862988e+00 2.24871770e-01 1.67379543e-01 3.81241925e-02
-9.69555497e-01 4.31206733e-01 -4.20566946e-01 -3.17663282e-01
2.35602289e-01 -1.05431780e-01 -2.53621310e-01 1.50985867e-01
7.84996212e-01 9.89276707e-01 1.54992238e-01 -9.18099463e-01
-1.00126314e+00 7.63856053e-01 6.95732057e-01 -4.12511498e-01
1.44419813e+00 -4.93846238e-01 2.61395007e-01 2.07850203e-01
3.34987491e-01 2.31175184e-01 -7.85649657e-01 1.61429252e-02
-2.06915557e-01 -3.38886172e-01 -2.09817231e-01 -9.30449724e-01
-5.98056018e-01 8.33821058e-01 8.94102633e-01 8.20958838e-02
1.35250485e+00 -2.40047336e-01 6.16166532e-01 8.41141582e-01
1.04308081e+00 -7.27851689e-01 -1.40681371e-01 4.94068921e-01
1.14737177e+00 -1.03216529e+00 2.94241726e-01 -8.21776018e-02
-8.54897380e-01 9.02659595e-01 1.48075134e-01 5.42423762e-02
4.13228333e-01 4.33497906e-01 2.22790986e-01 -2.78383285e-01
-3.35753351e-01 -3.87491763e-01 -2.85992622e-01 1.24051964e+00
-1.28482446e-01 7.53806606e-02 -2.57303357e-01 9.59274530e-01
-2.82931507e-01 -2.86274910e-01 7.36002564e-01 1.22192907e+00
-7.13248730e-01 -2.79623061e-01 -1.04955149e+00 3.70136768e-01
-2.66188532e-01 1.23986982e-01 -1.46307871e-01 6.87839210e-01
8.16758871e-02 1.45638669e+00 -3.31172854e-01 -3.80539179e-01
6.28269017e-01 -2.05009311e-01 3.15761387e-01 -1.09176300e-01
-2.75082469e-01 -2.01774046e-01 2.78748751e-01 -5.47988176e-01
-3.81117076e-01 -9.60349560e-01 -1.48688102e+00 -1.03993744e-01
-1.46871125e-02 1.87263846e-01 7.95273900e-01 6.72390521e-01
4.48726386e-01 3.75974000e-01 4.74231988e-01 -1.13561833e+00
-6.01770341e-01 -1.26209128e+00 -7.85778105e-01 7.67408982e-02
3.43085706e-01 -1.00519574e+00 -4.40712869e-01 -2.23120421e-01] | [6.817476272583008, 0.5689353346824646] |
9d430bca-99de-4b19-b592-758735dbc554 | neighborhood-matching-network-for-entity | 2005.05607 | null | https://arxiv.org/abs/2005.05607v1 | https://arxiv.org/pdf/2005.05607v1.pdf | Neighborhood Matching Network for Entity Alignment | Structural heterogeneity between knowledge graphs is an outstanding challenge for entity alignment. This paper presents Neighborhood Matching Network (NMN), a novel entity alignment framework for tackling the structural heterogeneity challenge. NMN estimates the similarities between entities to capture both the topological structure and the neighborhood difference. It provides two innovative components for better learning representations for entity alignment. It first uses a novel graph sampling method to distill a discriminative neighborhood for each entity. It then adopts a cross-graph neighborhood matching module to jointly encode the neighborhood difference for a given entity pair. Such strategies allow NMN to effectively construct matching-oriented entity representations while ignoring noisy neighbors that have a negative impact on the alignment task. Extensive experiments performed on three entity alignment datasets show that NMN can well estimate the neighborhood similarity in more tough cases and significantly outperforms 12 previous state-of-the-art methods. | ['Zheng Wang', 'Xiao Liu', 'Dongyan Zhao', 'Yuting Wu', 'Yansong Feng'] | 2020-05-12 | neighborhood-matching-network-for-entity-1 | https://aclanthology.org/2020.acl-main.578 | https://aclanthology.org/2020.acl-main.578.pdf | acl-2020-6 | ['graph-sampling'] | ['graphs'] | [-1.20362826e-01 2.73807943e-01 -6.72747016e-01 -2.57763982e-01
-5.48896253e-01 -5.61225772e-01 5.28482676e-01 6.88189447e-01
2.44956203e-02 4.29018587e-01 5.24776340e-01 5.35740778e-02
-2.95391142e-01 -1.21264875e+00 -6.42609000e-01 -3.59279543e-01
-4.69407350e-01 6.62719667e-01 1.17505908e-01 -3.63410980e-01
2.89772093e-01 4.34136629e-01 -1.04699290e+00 1.63414076e-01
1.14430094e+00 4.67483163e-01 1.49142727e-01 1.48556411e-01
-4.95745808e-01 6.33067548e-01 -3.57727766e-01 -4.37600911e-01
3.43408942e-01 -1.96674198e-01 -1.02622485e+00 -4.57870752e-01
1.10251546e+00 2.53028393e-01 -7.45887160e-01 1.22320950e+00
7.05774844e-01 2.16895446e-01 8.13825667e-01 -1.30891144e+00
-9.81052816e-01 9.50144589e-01 -5.88634431e-01 5.81706047e-01
6.95828736e-01 -3.23733002e-01 1.70383167e+00 -8.67823899e-01
1.03412521e+00 1.19850242e+00 1.05641568e+00 -3.83896977e-02
-1.31169808e+00 -6.20030701e-01 2.46790051e-01 4.71718103e-01
-1.71051168e+00 -9.27173868e-02 8.81994188e-01 -3.51623148e-01
1.06399214e+00 1.29885718e-01 7.39864290e-01 1.03838456e+00
6.78786561e-02 6.70423567e-01 5.70492268e-01 -1.97100505e-01
-9.43876728e-02 -5.15488923e-01 2.82444715e-01 6.76348746e-01
6.01016343e-01 -3.66818666e-01 -5.23143768e-01 -4.25563753e-01
7.76808202e-01 -7.30863661e-02 -2.64023364e-01 -8.08910728e-01
-1.44350839e+00 4.74371791e-01 9.73149478e-01 6.19370461e-01
-2.41070837e-01 2.12961465e-01 2.74410814e-01 4.29379314e-01
3.75039339e-01 7.69019961e-01 -2.51239866e-01 1.93932593e-01
-5.92036188e-01 2.99751818e-01 7.50672638e-01 1.16890407e+00
1.06995499e+00 -2.58810401e-01 -3.20866525e-01 8.36256504e-01
2.09397018e-01 3.85284632e-01 1.87191799e-01 -6.15377188e-01
1.02282095e+00 1.11065865e+00 -1.14035159e-01 -1.94732916e+00
-5.83446145e-01 -6.74118459e-01 -9.96305227e-01 -4.77376431e-01
2.83228546e-01 3.94480020e-01 -5.37971437e-01 1.90645850e+00
5.56371272e-01 8.88279021e-01 -4.67797816e-02 5.02798557e-01
1.34657776e+00 4.48711157e-01 1.76038258e-02 3.06085140e-01
1.12990367e+00 -1.16248429e+00 -5.70688963e-01 -2.21426070e-01
1.07125199e+00 -6.43062413e-01 6.95318937e-01 -6.46270216e-01
-9.56980109e-01 -5.13964653e-01 -1.13664186e+00 -1.31843984e-01
-5.88617802e-01 -7.31919780e-02 6.68524146e-01 2.57122725e-01
-1.18043327e+00 8.82576823e-01 -6.44761741e-01 -4.08226013e-01
3.11815739e-01 4.28816646e-01 -7.55879819e-01 -1.26815781e-01
-1.40200007e+00 9.52798903e-01 4.81530786e-01 8.54835585e-02
1.29826710e-01 -9.35515761e-01 -1.36452591e+00 2.73529619e-01
1.95086882e-01 -1.00337172e+00 4.16349828e-01 -4.61715162e-01
-7.92794645e-01 9.71969664e-01 -3.21114868e-01 -3.10676664e-01
2.27547616e-01 -7.45689124e-02 -6.94098592e-01 -2.34674305e-01
4.59391385e-01 5.67294300e-01 3.33686411e-01 -9.89756227e-01
-2.82209933e-01 -2.62687653e-01 -4.82690856e-02 4.36744124e-01
-2.30247095e-01 -4.97398078e-01 -6.06707990e-01 -9.22483385e-01
5.03046930e-01 -9.41885352e-01 -4.17699516e-01 -3.69334400e-01
-5.97351134e-01 -4.33022350e-01 5.06088972e-01 -4.40143079e-01
1.61211598e+00 -1.79759371e+00 4.58400249e-01 7.07969010e-01
6.33515656e-01 1.44522712e-01 -7.61457920e-01 8.33579183e-01
-3.17652583e-01 1.07841767e-01 -1.38621135e-02 -2.04847872e-01
1.55427363e-02 -6.43857270e-02 -7.98361152e-02 5.40598631e-01
2.24521965e-01 1.38973784e+00 -1.23260307e+00 -6.80149496e-01
-9.83043313e-02 3.54025036e-01 -3.92959207e-01 1.25993326e-01
1.83827519e-01 1.69860601e-01 -4.29330558e-01 7.12703168e-01
9.06671405e-01 -5.44531167e-01 7.49896348e-01 -7.15814233e-01
3.03684175e-01 4.03832465e-01 -1.50847423e+00 1.97453189e+00
-1.97937936e-01 6.00908637e-01 -3.27422291e-01 -1.01080370e+00
1.12707376e+00 -1.66380420e-01 5.90409398e-01 -6.90834284e-01
-1.85603991e-01 2.13301495e-01 -1.33045211e-01 -2.15609998e-01
7.90630877e-01 7.38525033e-01 -2.23590836e-01 2.36858726e-01
1.88907251e-01 3.78960401e-01 2.03079581e-01 4.88887459e-01
1.27403450e+00 -1.82655573e-01 6.78942263e-01 -4.63634163e-01
5.62576175e-01 -1.22286826e-01 8.55549216e-01 7.25634754e-01
-1.00177839e-01 4.77852523e-01 2.68452644e-01 -7.72678554e-01
-1.00266623e+00 -1.36962342e+00 6.92792013e-02 7.69876838e-01
7.66218603e-01 -9.53704119e-01 -3.98531437e-01 -8.38595331e-01
3.61178190e-01 1.55841634e-01 -6.93432152e-01 -3.58975381e-01
-9.94020820e-01 -4.99360263e-01 4.32649195e-01 8.01436961e-01
3.97511423e-01 -7.46227920e-01 5.74714184e-01 1.24812201e-01
-5.85898399e-01 -1.04344475e+00 -8.61405194e-01 -2.78797776e-01
-9.11202848e-01 -1.21282864e+00 -2.95297563e-01 -1.27187657e+00
9.19377685e-01 4.76179898e-01 1.60347438e+00 3.93022358e-01
-2.19491377e-01 2.45765746e-01 -1.77829474e-01 3.34445536e-01
-2.16246173e-01 7.31350064e-01 4.27210480e-02 -2.87288547e-01
6.29220963e-01 -8.69919062e-01 -5.41304469e-01 4.12942171e-01
-4.25659627e-01 -2.55371660e-01 6.68223023e-01 8.80206525e-01
7.60830224e-01 -2.58698523e-01 5.17336428e-01 -1.04997349e+00
6.00284815e-01 -8.17233920e-01 -3.14864784e-01 6.33791506e-01
-6.56430125e-01 2.77627289e-01 3.70059490e-01 -4.40830618e-01
-3.94254208e-01 -1.01969592e-01 3.11608106e-01 -3.25623184e-01
2.33348623e-01 6.20886683e-01 -2.60834545e-01 -3.02646339e-01
4.97473449e-01 6.20014369e-02 -3.60296339e-01 -3.78919750e-01
6.37373269e-01 2.11790890e-01 7.39970207e-01 -6.18911803e-01
1.11627162e+00 2.50474215e-01 2.47858122e-01 -4.27920967e-01
-6.39289618e-01 -8.97665739e-01 -1.07532406e+00 2.29376420e-01
4.48156595e-01 -1.18162525e+00 -4.03063774e-01 1.20613761e-01
-1.03668511e+00 1.25485465e-01 -1.48728117e-01 3.48569453e-01
-3.03818285e-01 6.73902452e-01 -4.41570610e-01 -3.70323434e-02
-3.61274391e-01 -7.23422825e-01 9.56259131e-01 2.55653858e-01
-5.27263641e-01 -1.38990915e+00 6.93291783e-01 2.93107003e-01
3.98990005e-01 3.37862313e-01 1.04784286e+00 -1.15488434e+00
-7.71136880e-01 -2.45757490e-01 -3.28063607e-01 -4.03875887e-01
3.92999977e-01 -4.87823598e-02 -3.42740983e-01 -5.11229515e-01
-9.38788772e-01 1.69130445e-01 9.38984931e-01 1.31034359e-01
6.68028712e-01 -3.54872912e-01 -8.81169140e-01 8.53244662e-01
1.28043926e+00 -4.91736978e-01 7.20968127e-01 4.63196099e-01
1.20576155e+00 3.71883601e-01 6.63719237e-01 6.04619384e-02
8.99397492e-01 9.56593394e-01 2.90793568e-01 -1.68727726e-01
-1.71609953e-01 -5.98623455e-01 9.50269401e-02 1.24744308e+00
-8.47442150e-02 -2.50563830e-01 -1.02566957e+00 8.67980421e-01
-2.28276539e+00 -1.17167830e+00 -2.23325923e-01 1.90296984e+00
6.02658331e-01 -2.22830564e-01 -4.28654701e-02 -4.48167831e-01
1.24927688e+00 6.73239231e-01 -2.78803915e-01 2.40137596e-02
-4.87397343e-01 2.53104810e-02 4.21463996e-01 3.82687330e-01
-1.28579021e+00 1.10411930e+00 6.62813425e+00 8.97014618e-01
-4.62356657e-01 -2.02564701e-01 1.06806010e-01 4.55032051e-01
-6.61332667e-01 7.11708516e-02 -7.82449007e-01 4.60063964e-01
5.02428114e-01 -3.23159456e-01 1.57043681e-01 9.11911786e-01
-4.00060803e-01 4.31511730e-01 -1.12403786e+00 1.03301644e+00
-3.50163318e-02 -1.84176886e+00 4.19169396e-01 -4.03699800e-02
1.12591302e+00 1.96285799e-01 -1.41222745e-01 2.11496323e-01
3.85122657e-01 -1.12262523e+00 7.95025229e-02 6.83354795e-01
4.57477808e-01 -7.88564861e-01 9.83647704e-01 -9.94204730e-02
-2.18676329e+00 1.52723581e-01 -6.97986960e-01 2.90974736e-01
5.23929484e-02 6.19899213e-01 -8.77706707e-01 1.03016627e+00
5.83743215e-01 1.24446702e+00 -8.19037974e-01 1.15266550e+00
-1.08505726e-01 4.98357676e-02 -2.45352685e-01 2.66400337e-01
1.48719683e-01 -6.07587099e-01 7.23265529e-01 1.34538555e+00
3.70520085e-01 -3.60495061e-01 4.12254810e-01 8.24595213e-01
-5.70954025e-01 4.05018359e-01 -1.00373578e+00 -4.34139408e-02
1.34650898e+00 1.27603734e+00 -6.04535818e-01 -1.26712382e-01
-4.48023260e-01 1.14766181e+00 1.15209770e+00 1.70422509e-01
-6.45249605e-01 -4.45255995e-01 9.69488084e-01 -7.81220347e-02
4.41957414e-01 -2.86090761e-01 -7.35373646e-02 -1.33037150e+00
1.54939562e-01 -8.11099529e-01 7.07299888e-01 -3.73337418e-01
-1.62742960e+00 5.12567937e-01 -1.57746777e-01 -1.31551731e+00
-1.13665886e-01 -2.56421655e-01 -1.00330424e+00 6.72881663e-01
-1.39498174e+00 -1.38968158e+00 -4.90650326e-01 4.39992130e-01
4.29829992e-02 -3.29298109e-01 8.61126125e-01 3.99188668e-01
-6.19428098e-01 9.92961049e-01 2.48376533e-01 5.04919648e-01
8.22645962e-01 -1.46744561e+00 1.08980227e+00 5.88071883e-01
7.10425973e-01 9.29171085e-01 3.11941057e-01 -9.37568426e-01
-1.21030593e+00 -1.29501998e+00 1.28950882e+00 -6.67758644e-01
7.51541853e-01 -2.01662317e-01 -1.05263758e+00 7.29617715e-01
3.80754955e-02 3.20807695e-01 8.21820974e-01 3.70862007e-01
-9.13226068e-01 -6.32924736e-02 -8.94412100e-01 6.93271697e-01
1.84290254e+00 -9.10991013e-01 -7.61302531e-01 1.01127163e-01
7.38196194e-01 -4.44925129e-01 -1.41634500e+00 8.12164128e-01
4.17249769e-01 -7.76865542e-01 1.28319526e+00 -1.00434566e+00
-8.82409979e-03 -5.71084619e-01 -2.06417307e-01 -1.47810054e+00
-7.88917601e-01 -6.54949725e-01 -5.59809983e-01 1.49088526e+00
6.36507273e-01 -5.64669073e-01 9.26267326e-01 2.10051626e-01
1.33445248e-01 -8.37313414e-01 -7.75504947e-01 -9.64362264e-01
-1.04650915e-01 2.86445618e-01 9.78993773e-01 1.76299345e+00
1.10896334e-01 3.67941260e-01 -1.65717080e-01 6.08332038e-01
5.81981182e-01 5.29399097e-01 8.89673233e-01 -1.40976954e+00
1.16333358e-01 -5.15595078e-01 -1.16544580e+00 -1.10739803e+00
6.54868126e-01 -1.37057936e+00 -5.58579266e-01 -1.69965339e+00
3.92404556e-01 -4.13751423e-01 -5.44861078e-01 1.66986868e-01
-6.22405589e-01 -8.28899257e-03 -6.23824000e-02 4.42137837e-01
-8.92120123e-01 5.01839459e-01 9.21459198e-01 -5.17208636e-01
-1.29456908e-01 -1.61354125e-01 -6.02745056e-01 3.86494219e-01
6.62657797e-01 -4.27871615e-01 -4.31813240e-01 -3.30176502e-01
5.54947257e-01 -4.70039338e-01 8.48927274e-02 -8.58719110e-01
6.45074010e-01 -5.72507493e-02 3.29639047e-01 -6.96633637e-01
-1.14954643e-01 -7.20136940e-01 5.20303190e-01 2.38010451e-01
-2.59684533e-01 6.75559163e-01 -2.20259368e-01 1.05729008e+00
-4.94353533e-01 1.00397564e-01 4.08581853e-01 8.90824422e-02
-9.10693765e-01 5.52648783e-01 3.24196517e-01 4.00886238e-01
9.45268571e-01 -1.83528125e-01 -5.38809001e-01 -2.61022896e-01
-7.03305662e-01 5.67351699e-01 7.56303310e-01 5.49357593e-01
4.87598836e-01 -1.95241308e+00 -7.33087659e-01 -3.26485984e-04
4.23419833e-01 1.90361198e-02 2.09145293e-01 7.69943833e-01
-4.05699372e-01 1.20142817e-01 -1.63400009e-01 -4.79119003e-01
-1.27746797e+00 4.24241513e-01 4.13756698e-01 -8.22768450e-01
-7.08808959e-01 8.61701369e-01 -2.99180690e-02 -8.99763167e-01
1.29473522e-01 -3.35652456e-02 -3.92621428e-01 1.22390717e-01
1.98658317e-01 6.48703456e-01 -4.14669653e-03 -9.22135174e-01
-6.14594162e-01 9.03938770e-01 -2.97684342e-01 7.20871031e-01
1.23152840e+00 -2.20848784e-01 -3.56157720e-01 1.14990123e-01
1.28892159e+00 3.09270859e-01 -5.78909397e-01 -7.62922347e-01
5.40713966e-01 -5.24051845e-01 -3.63878340e-01 -1.88572213e-01
-1.10145283e+00 2.60361344e-01 2.41357654e-01 7.83893242e-02
5.61710000e-01 1.81949168e-01 8.56403530e-01 7.50930727e-01
2.66877323e-01 -1.09429121e+00 1.29015207e-01 5.34254611e-01
6.50350869e-01 -1.17454064e+00 3.05786967e-01 -9.37075019e-01
-3.93317819e-01 1.05544364e+00 8.81071389e-01 -3.51378083e-01
7.31044710e-01 -9.48716998e-02 -9.95382220e-02 -5.33353269e-01
-4.53191340e-01 -2.38729089e-01 9.27522004e-01 7.90375173e-01
4.38346863e-01 -7.25527434e-03 -2.55452454e-01 3.43425930e-01
-1.30585507e-01 -8.81916583e-01 7.06890821e-02 6.23103082e-01
-2.14996904e-01 -1.35668552e+00 2.07615986e-01 3.27301651e-01
-6.20326102e-02 -2.97468454e-01 -8.68005991e-01 9.23110604e-01
-3.12568545e-02 3.70138437e-01 2.66148925e-01 -5.98991454e-01
3.45905662e-01 -1.52097210e-01 3.76522779e-01 -5.12510180e-01
-5.94185054e-01 -5.39636970e-01 3.56902272e-01 -9.15420294e-01
-5.64865112e-01 -4.15796697e-01 -1.09934163e+00 -5.33711374e-01
-4.87073034e-01 2.21076176e-01 8.43683705e-02 7.10071027e-01
1.00382650e+00 2.43294805e-01 6.88811779e-01 -6.55500650e-01
-3.34545791e-01 -7.61033952e-01 -5.70168138e-01 8.06274414e-01
3.87360230e-02 -7.54711688e-01 -3.01047891e-01 -5.68670332e-01] | [8.705121994018555, 7.913496971130371] |
5b3155be-39d4-4861-9ab7-85987e31aa55 | vector-quantized-input-contextualized-soft | 2205.11024 | null | https://arxiv.org/abs/2205.11024v2 | https://arxiv.org/pdf/2205.11024v2.pdf | Vector-Quantized Input-Contextualized Soft Prompts for Natural Language Understanding | Prompt Tuning has been largely successful as a parameter-efficient method of conditioning large-scale pre-trained language models to perform downstream tasks. Thus far, soft prompt tuning learns a fixed set of task-specific continuous vectors, i.e., soft tokens that remain static across the task samples. A fixed prompt, however, may not generalize well to the diverse kinds of inputs the task comprises. In order to address this, we propose Vector-quantized Input-contextualized Prompts (VIP) as an extension to the soft prompt tuning framework. VIP particularly focuses on two aspects -- contextual prompts that learns input-specific contextualization of the soft prompt tokens through a small-scale sentence encoder and quantized prompts that maps the contextualized prompts to a set of learnable codebook vectors through a Vector quantization network. On various language understanding tasks like SuperGLUE, QA, Relation classification, NER and NLI, VIP outperforms the soft prompt tuning (PT) baseline by an average margin of 1.19%. Further, our generalization studies show that VIP learns more robust prompt representations, surpassing PT by a margin of 0.6% - 5.3% on Out-of-domain QA and NLI tasks respectively, and by 0.75% on Multi-Task setup over 4 tasks spanning across 12 domains. | ['Soujanya Poria', 'Steven C. H. Hoi', 'Amrita Saha', 'Rishabh Bhardwaj'] | 2022-05-23 | null | null | null | null | ['relation-classification'] | ['natural-language-processing'] | [ 3.88154149e-01 3.28985065e-01 -3.46755683e-01 -8.61684680e-01
-1.20057523e+00 -8.77371311e-01 8.39935899e-01 3.68074656e-01
-5.61019838e-01 7.52430022e-01 5.64881146e-01 -3.95231396e-01
1.50606290e-01 -5.03171563e-01 -8.41458678e-01 -4.07517672e-01
2.88479358e-01 6.01409674e-01 1.36939555e-01 -5.06745636e-01
1.77332833e-01 -1.32613406e-01 -9.20523465e-01 6.17469192e-01
9.66162264e-01 8.38283718e-01 3.49350095e-01 7.25387335e-01
-3.71164203e-01 5.60380161e-01 -8.41242313e-01 -2.58098841e-01
-1.29747391e-01 -7.31612816e-02 -1.31676912e+00 -4.37321573e-01
5.37935436e-01 -3.33274692e-01 -1.17582567e-01 7.78133571e-01
4.28301603e-01 3.41939420e-01 8.35629821e-01 -1.08214247e+00
-1.37260330e+00 1.07052219e+00 9.79268774e-02 3.48524749e-01
4.33668375e-01 1.86613098e-01 1.29084563e+00 -9.14555788e-01
5.84426582e-01 1.35192263e+00 4.65683311e-01 7.75546670e-01
-1.29506278e+00 -5.67536175e-01 3.19650352e-01 2.78616808e-02
-1.19721711e+00 -3.48885387e-01 4.17053163e-01 -3.77016932e-01
1.46096492e+00 -9.19272378e-02 -2.30157644e-01 1.42385018e+00
2.67408967e-01 8.34857821e-01 1.02572584e+00 -3.58445674e-01
2.96413720e-01 1.25338316e-01 4.71504927e-01 1.80837572e-01
-1.57353371e-01 6.48126602e-02 -7.61426747e-01 -1.07749812e-01
5.83767712e-01 -1.75842106e-01 -2.97245979e-01 3.32498401e-01
-1.34796298e+00 1.01700974e+00 4.06332463e-01 3.33894819e-01
-5.21121740e-01 7.96597674e-02 6.65503919e-01 5.07555664e-01
3.69865835e-01 6.27923608e-01 -1.17681170e+00 -3.16939384e-01
-6.69704258e-01 2.41739497e-01 7.07449675e-01 1.08673453e+00
9.63141799e-01 1.39473841e-01 -8.68695438e-01 8.94824684e-01
2.39151448e-01 5.69993436e-01 5.90647101e-01 -7.48793483e-01
8.23567092e-01 4.69970793e-01 3.72035317e-02 -3.71302783e-01
-2.56784886e-01 -3.24377626e-01 -8.45003307e-01 -4.79605138e-01
2.72959828e-01 -3.10882509e-01 -1.08858228e+00 2.16478515e+00
7.96675161e-02 6.23839647e-02 3.94224614e-01 7.28896141e-01
1.13115764e+00 1.04001904e+00 7.29728401e-01 -1.00953570e-02
1.30938721e+00 -9.77777302e-01 -7.54916310e-01 -3.82671475e-01
8.91817629e-01 -7.22901762e-01 1.87670159e+00 2.49125198e-01
-9.52082098e-01 -8.74150753e-01 -8.61084461e-01 -4.92742658e-01
-5.76252103e-01 1.08011439e-01 7.02715278e-01 4.20364171e-01
-1.21788669e+00 2.56091028e-01 -4.98453647e-01 -3.80897224e-01
3.01717311e-01 2.45700687e-01 -3.20863694e-01 -1.98816091e-01
-1.39451337e+00 7.14481175e-01 5.80040097e-01 -4.86894131e-01
-1.12351787e+00 -1.28282273e+00 -9.33520198e-01 3.64853293e-01
3.34028095e-01 -5.97211421e-01 1.69717693e+00 -7.16063559e-01
-1.74796820e+00 7.87913382e-01 -3.73307467e-01 -5.22780418e-01
-4.05617692e-02 -6.42651916e-01 -2.94205517e-01 -1.33877276e-02
2.77167171e-01 9.43279207e-01 7.04952896e-01 -1.02617836e+00
-4.96837586e-01 1.88033413e-02 3.90156955e-01 2.61207879e-01
-3.44901621e-01 1.38571113e-01 -2.12941870e-01 -6.10286891e-01
-3.14737320e-01 -6.18383884e-01 -8.14865157e-02 -4.94089395e-01
-3.77140820e-01 -9.64699686e-01 5.07792473e-01 -5.25463343e-01
1.07157290e+00 -2.08180714e+00 1.12840876e-01 -3.17486912e-01
-7.56485667e-03 4.24020112e-01 -4.66470480e-01 5.60369790e-01
4.68518771e-02 2.29189366e-01 -2.09907681e-01 -4.23792034e-01
3.36436510e-01 5.79118788e-01 -7.70587325e-01 -1.32783100e-01
5.36369681e-01 1.25397885e+00 -1.02386105e+00 -1.91598430e-01
3.90414782e-02 3.55474621e-01 -6.28056407e-01 6.47817016e-01
-6.24680519e-01 4.40649748e-01 -3.14410895e-01 4.79936153e-01
3.51708680e-01 -3.57247949e-01 -9.84090045e-02 -1.43851444e-01
2.09413528e-01 8.36690247e-01 -8.22982252e-01 2.11065578e+00
-6.91369951e-01 3.80948365e-01 5.24715707e-02 -1.17049277e+00
9.48409081e-01 7.18848884e-01 -4.00853232e-02 -9.16655421e-01
-1.08710080e-02 -2.93498822e-02 -3.33779275e-01 -2.83735365e-01
9.29757655e-01 -2.15355918e-01 -6.26734257e-01 5.23848534e-01
6.48532271e-01 -5.49596660e-02 -5.46382293e-02 6.43542707e-01
9.73751724e-01 -6.73617944e-02 3.58696818e-01 -3.17292273e-01
4.67444539e-01 -6.38805553e-02 5.81962109e-01 8.37330461e-01
-3.15825611e-01 4.13336277e-01 3.64424080e-01 -3.18159908e-02
-7.78409660e-01 -1.28214335e+00 -1.27278334e-02 1.97833824e+00
-1.10620350e-01 -6.75609052e-01 -5.89608192e-01 -7.57865429e-01
4.27273810e-02 9.47986007e-01 -5.36059499e-01 -2.72443622e-01
-5.47101498e-01 -3.08572412e-01 6.09527826e-01 9.07393634e-01
3.19999903e-01 -1.22945654e+00 -2.09807441e-01 3.34184796e-01
-2.93155819e-01 -1.32978725e+00 -6.06925964e-01 7.16291070e-01
-6.57682836e-01 -6.14485025e-01 -4.48212117e-01 -8.72381389e-01
4.33818907e-01 -1.36381136e-02 1.44701827e+00 -2.19692200e-01
2.06427023e-01 4.88470584e-01 -6.92152977e-01 -2.45532811e-01
-4.88878101e-01 4.36807096e-01 7.73021206e-02 -3.20750862e-01
4.58633095e-01 -2.07765058e-01 -3.31204861e-01 7.97246322e-02
-8.91836286e-01 -1.68651313e-01 5.99702597e-01 1.23962271e+00
6.54193878e-01 -5.03920972e-01 1.02647042e+00 -1.11787879e+00
1.04093885e+00 -6.93690419e-01 -1.23211160e-01 4.40476269e-01
-4.46589828e-01 2.32265964e-01 9.21679974e-01 -7.19448268e-01
-1.27257359e+00 -3.07910562e-01 -2.45035455e-01 -5.61074950e-02
-4.52479124e-01 5.36775947e-01 -2.24677786e-01 3.96824479e-01
8.85179579e-01 1.37964159e-01 -6.54697955e-01 -5.48585117e-01
1.04407656e+00 8.96577179e-01 8.40607285e-01 -1.03996623e+00
6.06852829e-01 -4.50404696e-02 -6.67836249e-01 -6.03358567e-01
-1.07045925e+00 -5.99409223e-01 -5.69387317e-01 4.44213241e-01
1.09906578e+00 -9.18751836e-01 -5.10593593e-01 7.58442506e-02
-1.26633561e+00 -9.84042823e-01 -3.67110252e-01 2.25774512e-01
-5.00603855e-01 9.53931808e-02 -8.33944976e-01 -3.68602157e-01
-7.41789162e-01 -1.00086403e+00 1.28866124e+00 2.89189279e-01
-5.46388447e-01 -1.14443862e+00 -1.77580714e-02 3.69949341e-01
6.99136019e-01 -2.54072011e-01 1.15963137e+00 -9.76378918e-01
-3.53464365e-01 2.78154492e-01 -2.35724762e-01 5.91603875e-01
6.72943369e-02 -3.68126392e-01 -1.15591371e+00 -3.20493400e-01
-2.94276386e-01 -9.26592529e-01 6.82283938e-01 2.48810500e-01
1.05776083e+00 -3.27544719e-01 -2.88074583e-01 6.50436223e-01
1.28988969e+00 9.64362249e-02 1.95854858e-01 6.66511282e-02
5.46366274e-01 2.04088122e-01 7.91433871e-01 1.83201164e-01
5.72987974e-01 4.55017358e-01 1.73435360e-01 8.05858448e-02
-2.97649622e-01 -5.04959643e-01 5.00139713e-01 1.01608288e+00
4.66595620e-01 -2.51292914e-01 -8.83333802e-01 8.06727171e-01
-1.54347408e+00 -4.57702994e-01 2.19849348e-01 1.77906454e+00
1.42861521e+00 2.03954995e-01 -2.80079067e-01 -1.54790103e-01
2.75067180e-01 8.69980305e-02 -4.59047467e-01 -9.07330692e-01
-6.19772859e-02 6.12719715e-01 2.05412030e-01 6.79467738e-01
-9.86503065e-01 1.58008420e+00 5.78924894e+00 7.52805829e-01
-1.34645855e+00 4.63362306e-01 3.52302402e-01 1.79313287e-01
-5.42862415e-01 7.94599429e-02 -1.11420047e+00 3.58322471e-01
1.33049726e+00 -2.67769903e-01 3.40941161e-01 9.08502400e-01
-5.64049333e-02 1.48924023e-01 -1.47342789e+00 8.47488344e-01
-1.48399815e-01 -1.36533368e+00 1.74374461e-01 -4.28497076e-01
8.49436224e-01 2.95335442e-01 1.51108727e-01 1.11750686e+00
6.00401282e-01 -1.31371140e+00 6.05237663e-01 1.02173507e-01
1.14949155e+00 -3.82675916e-01 6.54893577e-01 3.95375162e-01
-9.58460927e-01 1.23939728e-02 -6.00243330e-01 -1.53751045e-01
2.96300441e-01 3.28792393e-01 -1.34840775e+00 4.71321315e-01
6.35903180e-01 4.96111572e-01 -4.32074904e-01 4.07250226e-01
-6.40746236e-01 1.25952160e+00 -2.49227136e-01 -2.23744288e-02
5.31725228e-01 3.03857297e-01 1.16840646e-01 1.58611047e+00
-6.29336759e-02 3.30071211e-01 3.19209278e-01 8.64968181e-01
-3.80305171e-01 1.63040489e-01 -2.66429871e-01 -1.81509942e-01
8.60222340e-01 1.07808006e+00 -1.02428056e-01 -6.72535479e-01
-4.48594868e-01 1.03561997e+00 5.41629612e-01 6.37678206e-01
-4.01907653e-01 -4.28664058e-01 7.65986443e-01 -3.70432287e-01
3.56071621e-01 -3.77426654e-01 -4.55851138e-01 -1.06053400e+00
-6.37359768e-02 -8.12370181e-01 3.74160379e-01 -7.01841056e-01
-1.49651778e+00 7.12448955e-01 2.41841208e-02 -5.47399223e-01
-4.70319092e-01 -5.71241140e-01 -8.53925943e-01 1.25781417e+00
-1.81833553e+00 -1.13316500e+00 -1.69353977e-01 7.80780435e-01
8.23738873e-01 -1.88005671e-01 1.23436272e+00 6.57158941e-02
-3.63027871e-01 9.05637920e-01 -6.04076730e-03 1.61535442e-01
1.21622944e+00 -1.62521207e+00 5.47356129e-01 6.25495851e-01
1.11296915e-01 9.28120971e-01 6.14549994e-01 -5.42994499e-01
-1.22895074e+00 -1.20673442e+00 1.30168748e+00 -6.42511666e-01
6.09781504e-01 -5.54571688e-01 -1.28280580e+00 1.04545164e+00
6.26889944e-01 2.38971207e-02 8.00667107e-01 4.34796661e-01
-5.88836908e-01 -1.01435140e-01 -9.65516925e-01 3.46647739e-01
9.34085786e-01 -9.19168293e-01 -1.23971093e+00 4.20330942e-01
1.37208843e+00 -3.71118695e-01 -1.20077479e+00 2.78831303e-01
-7.70020783e-02 -3.63465935e-01 9.58328843e-01 -9.33870852e-01
2.44486988e-01 7.64489695e-02 -4.12961483e-01 -1.59107435e+00
-4.04411405e-01 -6.57377243e-01 5.86877763e-02 1.66130304e+00
4.46909308e-01 -5.79933524e-01 5.41861057e-01 4.13346201e-01
-5.16154706e-01 -7.68513143e-01 -7.86339164e-01 -7.51151621e-01
6.71762645e-01 -6.19320869e-01 7.48932481e-01 9.56205249e-01
7.83683062e-02 9.48279798e-01 4.69956696e-02 2.06752673e-01
3.31279159e-01 -1.26340678e-02 7.04537153e-01 -9.86581564e-01
-2.87076980e-01 -1.89778104e-01 1.72476545e-01 -1.64979243e+00
4.11423743e-01 -1.33215463e+00 2.12491170e-01 -1.51123476e+00
1.52708320e-02 -7.24980950e-01 -4.13852155e-01 9.18634713e-01
-6.16984367e-01 -2.63420343e-01 3.02678328e-02 1.97456628e-02
-7.03019023e-01 7.33861864e-01 1.08862305e+00 -2.32905492e-01
-2.21125379e-01 -3.35045457e-01 -9.44924533e-01 2.61713773e-01
8.65487337e-01 -3.51511389e-01 -7.16107666e-01 -7.98621416e-01
1.36441454e-01 3.50548066e-02 8.84465724e-02 -4.51991796e-01
2.39886373e-01 -2.55067468e-01 9.72758681e-02 -6.44476652e-01
2.86527008e-01 -2.84983397e-01 -7.17470765e-01 -3.24290805e-02
-8.30511034e-01 -1.57017082e-01 2.29120493e-01 3.71838748e-01
-3.57152939e-01 -7.44313896e-02 5.17229378e-01 -3.26471440e-02
-1.04310918e+00 2.72531122e-01 -2.14041159e-01 7.63425767e-01
5.12369037e-01 3.64326164e-02 -6.08000159e-01 -2.41427004e-01
-5.89287400e-01 4.33567256e-01 -2.52820645e-02 7.28215218e-01
5.28169453e-01 -1.20136595e+00 -9.00034606e-01 2.28195757e-01
3.44786137e-01 3.94620836e-01 1.86139762e-01 4.17496741e-01
2.42767930e-01 9.71173584e-01 -5.39096855e-02 -8.33730221e-01
-1.00462461e+00 3.66508514e-01 -3.35716754e-02 -3.20388317e-01
-4.74389821e-01 1.30649734e+00 5.60067594e-01 -7.89776504e-01
3.36824775e-01 -9.04703975e-01 -6.72433712e-03 -2.64197588e-01
5.67187011e-01 -1.68320864e-01 1.22631185e-01 -3.86132210e-01
-2.89275855e-01 3.69296759e-01 -3.72448891e-01 -4.43285070e-02
1.29370666e+00 -1.10814974e-01 2.15406120e-01 4.90406364e-01
1.40271235e+00 -2.68886417e-01 -1.35582662e+00 -7.82551408e-01
1.88057572e-01 5.45222424e-02 -6.94040656e-02 -1.23641431e+00
-4.71784443e-01 1.08154762e+00 1.63130224e-01 -1.70512532e-03
9.26155627e-01 4.67691541e-01 1.01103389e+00 5.14475286e-01
2.82946169e-01 -9.71089780e-01 2.98084706e-01 1.01143146e+00
1.06021678e+00 -1.29200399e+00 -6.69963121e-01 -2.56473005e-01
-9.44324076e-01 8.83376718e-01 6.65433288e-01 2.36761030e-02
4.06280011e-01 1.84415475e-01 2.32108966e-01 -2.55496129e-02
-1.14944744e+00 2.39402279e-02 3.23079854e-01 8.47161591e-01
6.98436022e-01 4.73052174e-01 -7.89005086e-02 1.18553507e+00
-4.25281018e-01 -8.99260864e-02 2.26523936e-01 7.59668529e-01
-6.03092372e-01 -1.20998669e+00 -1.89612702e-01 3.38969171e-01
-3.72550875e-01 -4.64769989e-01 -2.15962425e-01 4.42570984e-01
3.68996449e-02 1.14073467e+00 -3.88923660e-02 -2.27690518e-01
3.33614469e-01 4.59813327e-01 2.36358106e-01 -1.13465762e+00
-8.53951633e-01 -3.54457855e-01 5.11655286e-02 -6.64962709e-01
-3.74762341e-02 -5.07797837e-01 -1.60853779e+00 1.54845655e-01
-8.15221369e-02 2.67502725e-01 5.18136263e-01 1.14315403e+00
3.53850931e-01 6.53336525e-01 2.54087538e-01 -5.25391102e-01
-1.13703144e+00 -1.36738825e+00 -1.56408891e-01 6.06848300e-01
3.92953515e-01 -4.71864492e-01 -2.11417854e-01 2.33103976e-01] | [10.88086223602295, 8.355691909790039] |
a2e7eaba-477c-4327-a1b3-314412180704 | computational-model-discovery-with | 2001.00008 | null | https://arxiv.org/abs/2001.00008v1 | https://arxiv.org/pdf/2001.00008v1.pdf | Computational model discovery with reinforcement learning | The motivation of this study is to leverage recent breakthroughs in artificial intelligence research to unlock novel solutions to important scientific problems encountered in computational science. To address the human intelligence limitations in discovering reduced-order models, we propose to supplement human thinking with artificial intelligence. Our three-pronged strategy consists of learning (i) models expressed in analytical form, (ii) which are evaluated a posteriori, and iii) using exclusively integral quantities from the reference solution as prior knowledge. In point (i), we pursue interpretable models expressed symbolically as opposed to black-box neural networks, the latter only being used during learning to efficiently parameterize the large search space of possible models. In point (ii), learned models are dynamically evaluated a posteriori in the computational solver instead of based on a priori information from preprocessed high-fidelity data, thereby accounting for the specificity of the solver at hand such as its numerics. Finally in point (iii), the exploration of new models is solely guided by predefined integral quantities, e.g., averaged quantities of engineering interest in Reynolds-averaged or large-eddy simulations (LES). We use a coupled deep reinforcement learning framework and computational solver to concurrently achieve these objectives. The combination of reinforcement learning with objectives (i), (ii) and (iii) differentiate our work from previous modeling attempts based on machine learning. In this report, we provide a high-level description of the model discovery framework with reinforcement learning. The method is detailed for the application of discovering missing terms in differential equations. An elementary instantiation of the method is described that discovers missing terms in the Burgers' equation. | ['Adrián Lozano-Durán', 'Maxime Bassenne'] | 2019-12-29 | null | null | null | null | ['model-discovery'] | ['miscellaneous'] | [ 2.45307580e-01 2.25387961e-01 4.33204919e-02 5.88489622e-02
-5.16715109e-01 -5.56620836e-01 6.62759721e-01 1.87864453e-01
-3.55260164e-01 1.17224157e+00 -4.06301409e-01 -7.38521457e-01
-6.50646031e-01 -8.91274095e-01 -8.71740758e-01 -6.41260386e-01
-2.01013267e-01 6.59886599e-01 -3.99413794e-01 -3.93632591e-01
5.26584804e-01 1.01858294e+00 -1.59216964e+00 5.51074706e-02
1.16394901e+00 1.07089615e+00 -1.41003624e-01 6.92561150e-01
-1.28116667e-01 1.08517444e+00 -2.29227468e-01 1.86151087e-01
2.95810908e-01 -4.89917874e-01 -1.07558143e+00 3.79692800e-02
-1.85048863e-01 -2.24446431e-01 1.32824510e-01 8.99140596e-01
2.47789416e-02 4.29458380e-01 6.00714862e-01 -1.03065693e+00
-2.03348342e-02 1.73402593e-01 -1.09225482e-01 3.99322100e-02
2.14546412e-01 6.11856997e-01 8.39247167e-01 -7.08938539e-01
4.91046518e-01 8.69854987e-01 8.56445670e-01 3.60180199e-01
-1.35583961e+00 -2.71851271e-01 8.52223393e-03 1.06644787e-01
-1.30630648e+00 -2.37517104e-01 9.56179976e-01 -7.83629656e-01
1.27363241e+00 3.82937580e-01 9.21138167e-01 5.13571143e-01
3.56797606e-01 4.06978309e-01 1.28011394e+00 -6.94807231e-01
7.71677315e-01 2.98577160e-01 6.56598955e-02 6.07260287e-01
1.97735161e-01 7.24624872e-01 -3.23745579e-01 -3.03213686e-01
8.91416252e-01 -5.18475056e-01 -2.89963707e-02 -5.29404461e-01
-9.71181333e-01 1.07371473e+00 1.12017162e-01 1.86210379e-01
-6.75173998e-01 6.64081946e-02 2.79358268e-01 2.77374029e-01
3.86771441e-01 1.32429826e+00 -8.17685306e-01 -2.79013872e-01
-1.05243909e+00 5.86695135e-01 1.01839316e+00 3.77426624e-01
1.05621672e+00 4.82317626e-01 3.22809428e-01 4.60366845e-01
3.05242479e-01 2.12240964e-01 3.12931478e-01 -1.24506474e+00
-4.74866554e-02 5.10984838e-01 5.56064665e-01 -7.48781383e-01
-4.27524507e-01 -6.35516644e-01 -5.97067237e-01 6.89193070e-01
3.64072561e-01 -5.15887856e-01 -6.14618897e-01 1.40002716e+00
5.53962708e-01 3.05968881e-01 2.42660150e-01 9.03891623e-01
4.05004293e-01 7.49714077e-01 -7.86548331e-02 -4.25778955e-01
9.26049232e-01 -8.29868197e-01 -3.68641853e-01 2.06089560e-02
7.92494059e-01 -4.52826500e-01 7.97179341e-01 6.52245939e-01
-1.28261733e+00 -7.50753939e-01 -1.07789922e+00 2.41799340e-01
-4.31375325e-01 -1.09709844e-01 8.34828138e-01 4.29166049e-01
-1.16832697e+00 1.30242956e+00 -9.34713483e-01 1.24960557e-01
-1.75476559e-02 4.27854896e-01 1.84089337e-02 5.71521044e-01
-1.31736314e+00 1.07781422e+00 2.81213582e-01 2.79472709e-01
-7.48687387e-01 -1.01951087e+00 -8.22054029e-01 1.22640297e-01
4.73449767e-01 -8.00582767e-01 1.44576252e+00 -1.29872286e+00
-1.94542336e+00 5.26454806e-01 -1.67107567e-01 -4.45593625e-01
7.41917074e-01 -1.19659998e-01 -1.97628081e-01 3.67558509e-01
-6.37472346e-02 3.69997144e-01 7.54465938e-01 -1.58472085e+00
-2.80244857e-01 1.57840356e-01 3.29759985e-01 7.01554418e-02
2.97115356e-01 -3.21571618e-01 3.08462113e-01 -5.66966176e-01
-1.40741905e-02 -7.91645586e-01 -4.72426772e-01 5.97364083e-02
8.53262469e-03 -4.99270372e-02 5.29629350e-01 -6.49554014e-01
1.01663196e+00 -1.59345496e+00 3.90312731e-01 4.26036745e-01
9.07110721e-02 2.87423819e-01 2.80108243e-01 7.09282339e-01
-3.64922911e-01 1.58114642e-01 -6.31572723e-01 -7.18963519e-02
-9.95978490e-02 3.09967548e-01 -6.10415280e-01 1.73766151e-01
5.72326660e-01 8.97108495e-01 -1.05842054e+00 -2.95656949e-01
4.32935148e-01 3.14191490e-01 -7.58277297e-01 2.16829494e-01
-5.33536136e-01 9.91578698e-01 -7.34624267e-01 3.56752127e-01
5.37093103e-01 -3.22902739e-01 1.74120978e-01 -2.07757309e-01
-6.43592119e-01 3.73429477e-01 -1.47247076e+00 1.25884366e+00
-7.34939337e-01 1.42394766e-01 3.41930330e-01 -1.57288992e+00
7.38894820e-01 1.89068794e-01 6.76139295e-01 -5.52662909e-01
-4.19239290e-02 3.09666097e-01 6.86269328e-02 -6.05078340e-01
2.85450041e-01 -6.31934762e-01 2.86936849e-01 4.94413584e-01
1.76882818e-02 -7.08790362e-01 2.66883701e-01 -1.86486393e-01
8.20002198e-01 6.55883491e-01 4.43756759e-01 -7.10985422e-01
8.87849271e-01 3.72623622e-01 4.22489464e-01 9.16817129e-01
7.68448040e-02 1.35127500e-01 7.01697886e-01 -7.84347832e-01
-1.12787342e+00 -7.87866294e-01 -3.51607442e-01 5.48132479e-01
-1.31996110e-01 -2.00755075e-01 -5.89857578e-01 -1.60518125e-01
5.68543114e-02 1.03644061e+00 -6.15337014e-01 -7.29435682e-02
-9.04035687e-01 -6.41592562e-01 1.92166552e-01 4.84669179e-01
2.96604455e-01 -1.22521544e+00 -1.17206037e+00 4.75422889e-01
3.57576311e-01 -6.97743416e-01 4.97954398e-01 4.70787704e-01
-1.28011501e+00 -1.12979817e+00 -2.72527874e-01 -3.18497717e-01
4.88089621e-01 -5.06382823e-01 1.19961596e+00 2.72743464e-01
-3.13116640e-01 6.06375098e-01 -9.90403295e-02 -8.05315450e-02
-4.56726342e-01 -2.44905904e-01 8.53368938e-02 -4.89406697e-02
-2.01576069e-01 -5.48455298e-01 -2.38752946e-01 -1.42237410e-01
-9.77126360e-01 2.90770859e-01 3.83800477e-01 1.11948478e+00
5.62720895e-01 2.79752403e-01 6.10528231e-01 -6.78700089e-01
5.41792750e-01 -4.76130396e-01 -1.01541531e+00 1.08464353e-01
-6.97459817e-01 4.55048889e-01 1.04767621e+00 -3.48639607e-01
-1.12745678e+00 -8.49773437e-02 -1.86320260e-01 -4.34089452e-01
-2.65734643e-01 8.13418627e-01 1.03193127e-01 -2.98562586e-01
3.67891908e-01 1.91087395e-01 2.30950296e-01 -5.15012383e-01
1.43825427e-01 1.87359408e-01 3.11102837e-01 -1.40356505e+00
6.26295030e-01 2.92258501e-01 4.23227072e-01 -8.11852634e-01
-5.73763549e-01 1.77289278e-03 -7.52616704e-01 -1.90495506e-01
5.64959526e-01 -3.88640255e-01 -1.02197206e+00 4.30876672e-01
-1.09429693e+00 -5.16633391e-01 -7.31344581e-01 6.61800802e-01
-8.84138227e-01 7.27622136e-02 -4.71407861e-01 -1.16097808e+00
-1.87258571e-02 -1.37591052e+00 9.66080666e-01 2.33732507e-01
-4.51797128e-01 -1.35232711e+00 1.65524334e-01 5.41981347e-02
4.41163212e-01 4.39972878e-01 1.16219735e+00 -4.96390224e-01
-4.55763310e-01 -3.67237185e-03 5.51937968e-02 2.78611153e-01
-1.02107912e-01 1.99343979e-01 -9.58634496e-01 -9.19441730e-02
4.73142505e-01 -3.17285776e-01 3.99834305e-01 2.41942599e-01
1.26042998e+00 -4.89485353e-01 -1.46032199e-01 6.27368629e-01
1.44435668e+00 4.48238879e-01 2.99845517e-01 3.86382610e-01
3.63059253e-01 6.79003179e-01 3.08874011e-01 6.00816667e-01
5.53598776e-02 4.35776711e-01 1.72367752e-01 -1.50975939e-02
3.70335639e-01 3.22723798e-02 -4.35345583e-02 6.63504303e-01
-4.10665989e-01 4.07957882e-01 -1.17936110e+00 2.96467602e-01
-1.69120502e+00 -8.17963123e-01 2.34795004e-01 2.15715647e+00
9.98965085e-01 3.31513315e-01 -1.07350573e-01 1.10849716e-01
2.61208564e-01 -1.80743858e-01 -8.90385807e-01 -8.55108976e-01
2.54007429e-01 6.73443079e-01 8.40815604e-02 9.89043176e-01
-8.89220297e-01 6.07355773e-01 6.14034462e+00 4.39205855e-01
-1.33318758e+00 -3.46210927e-01 6.68063343e-01 1.05888084e-01
-3.32757086e-01 3.12862307e-01 -5.60200393e-01 1.89814731e-01
1.17210507e+00 -1.64205864e-01 8.63842964e-01 7.61953056e-01
4.15086716e-01 -2.75314242e-01 -1.25670445e+00 4.35336649e-01
-3.19704175e-01 -1.56504667e+00 1.99678689e-02 -1.64053127e-01
6.92853212e-01 -3.71480137e-01 -1.95819199e-01 5.69659650e-01
2.25629032e-01 -1.28998101e+00 7.64499068e-01 9.14449811e-01
6.16940975e-01 -6.93056285e-01 4.36480492e-01 6.13092482e-01
-8.87041509e-01 -4.72644232e-02 1.03386037e-01 -7.10431993e-01
7.10066482e-02 4.67723131e-01 -5.59164941e-01 7.59093702e-01
3.60712111e-01 6.84863150e-01 -9.85151157e-02 6.82086885e-01
-1.67261697e-02 7.14773893e-01 -5.09988666e-01 -3.90325710e-02
5.58870316e-01 -6.22669578e-01 5.71429014e-01 8.11557889e-01
1.28525123e-01 4.12699342e-01 1.32697597e-01 1.48721635e+00
4.82653737e-01 -1.99058965e-01 -3.66490901e-01 -9.71207246e-02
4.65178043e-02 1.08894145e+00 -6.97807431e-01 -4.95997071e-01
-2.25322872e-01 1.77106798e-01 2.44082749e-01 6.60517752e-01
-7.16394365e-01 -1.85619190e-01 6.01122081e-01 1.58547089e-01
3.47433597e-01 -4.47284609e-01 -6.39816225e-01 -1.13455617e+00
-7.27315843e-02 -1.16764545e+00 5.56279952e-03 -7.55337775e-01
-1.18065381e+00 3.20591807e-01 4.03518230e-01 -9.46675241e-01
-6.65425003e-01 -8.34113896e-01 -5.95724344e-01 1.18903792e+00
-1.58219123e+00 -7.11794555e-01 1.11168370e-01 1.72304899e-01
3.28479707e-01 -1.24551440e-02 9.36950326e-01 -2.06137806e-01
-3.45201015e-01 -2.70385109e-02 3.26800257e-01 -1.49773985e-01
5.71713434e-04 -1.13252592e+00 1.32026821e-01 3.99521232e-01
-2.73668438e-01 8.40917706e-01 8.39734614e-01 -8.32896769e-01
-1.54189193e+00 -6.19792581e-01 6.10734165e-01 -9.93723944e-02
9.06275749e-01 6.79643033e-03 -1.18238676e+00 4.15851802e-01
-3.97317484e-02 2.43912023e-02 2.06987306e-01 1.10099278e-02
1.38057902e-01 -9.19225290e-02 -1.14999712e+00 4.67241019e-01
3.31721157e-01 -3.58095974e-01 -6.55115485e-01 1.59604579e-01
3.53934258e-01 -5.50268888e-01 -8.96830976e-01 7.65233099e-01
4.39132601e-01 -7.57473826e-01 1.09279585e+00 -1.07296145e+00
6.48914814e-01 -2.86101162e-01 1.73873484e-01 -1.16890740e+00
-1.09999115e-02 -9.49303806e-01 -5.42408407e-01 5.51388144e-01
1.77005038e-01 -8.73609364e-01 4.92708027e-01 8.71488631e-01
-1.97719991e-01 -1.23017132e+00 -1.06239712e+00 -6.61413908e-01
7.01744795e-01 -4.02044982e-01 4.71167028e-01 1.08778775e+00
-6.14280365e-02 -7.57806227e-02 -1.11080810e-01 1.73174456e-01
5.70926249e-01 4.52201366e-01 4.32181686e-01 -1.12235820e+00
-6.63391232e-01 -5.39517820e-01 1.15187213e-01 -8.34789813e-01
3.57418746e-01 -6.09345019e-01 -1.26794949e-01 -1.16988432e+00
-3.95690084e-01 -5.74699819e-01 -2.14182392e-01 3.78973395e-01
4.33622301e-02 -2.09991246e-01 -2.03449428e-02 1.19845711e-01
-1.32693723e-01 5.92881680e-01 1.20213687e+00 1.20034844e-01
-3.40820462e-01 -7.03463629e-02 -3.71291459e-01 9.43345249e-01
9.11629915e-01 -8.08628872e-02 -2.55373418e-01 -1.24203883e-01
4.26697791e-01 5.21339953e-01 8.48575890e-01 -1.09476376e+00
1.89712495e-01 -5.83820164e-01 5.32967329e-01 -2.78480738e-01
2.23342299e-01 -6.19012535e-01 7.23446086e-02 6.07160211e-01
-4.14789230e-01 -1.09398618e-01 6.24064088e-01 1.00692131e-01
-9.62142944e-02 -5.29780746e-01 8.57225776e-01 -4.15900797e-01
-5.94166577e-01 -2.87184745e-01 -5.51202059e-01 1.06799275e-01
8.34546268e-01 -1.82883248e-01 1.28629684e-01 -2.28365973e-01
-8.77076209e-01 1.35658130e-01 3.28843653e-01 -9.41289738e-02
2.29250297e-01 -8.23387146e-01 -4.02230710e-01 4.33101147e-01
-4.47349787e-01 2.41143540e-01 1.34498596e-01 6.95885181e-01
-8.29472065e-01 4.48687583e-01 -2.05637574e-01 -3.66013408e-01
-5.01956046e-01 6.03171527e-01 8.09517682e-01 -5.89137018e-01
-3.83903772e-01 6.09612465e-01 1.75284930e-02 -6.17960393e-01
-1.87653080e-01 -4.32428688e-01 -9.88443941e-03 -1.87115133e-01
1.95507661e-01 5.34030020e-01 2.12582946e-01 -4.19147074e-01
-1.90349594e-01 5.38195431e-01 2.31721431e-01 -1.47815585e-01
1.48451376e+00 1.99111462e-01 -2.05024153e-01 6.21874213e-01
9.14552093e-01 -1.94739491e-01 -1.43280053e+00 -2.00595483e-02
1.89871285e-02 -9.41508543e-03 2.24086225e-01 -1.00055325e+00
-6.97860837e-01 9.51430500e-01 1.35546133e-01 1.92021295e-01
9.39525187e-01 -4.13962424e-01 5.33669233e-01 6.12470984e-01
2.61457354e-01 -1.24717009e+00 -1.80230796e-01 6.43605888e-01
9.72503126e-01 -1.13911653e+00 2.93181419e-01 1.69625902e-03
-3.41336638e-01 1.51844513e+00 6.94956243e-01 -3.16704631e-01
7.54578352e-01 5.56310773e-01 -6.59806281e-02 -9.88574997e-02
-6.26426697e-01 1.73722729e-01 4.10165429e-01 1.60652906e-01
3.38197738e-01 -3.60236585e-01 -3.43259782e-01 5.90639889e-01
-1.92365259e-01 1.77298650e-01 2.92316198e-01 1.21805191e+00
-3.09548080e-01 -1.01361263e+00 -5.74850559e-01 3.05510283e-01
-1.25763655e-01 -1.23526447e-01 -1.02620702e-02 8.72035682e-01
1.39340863e-01 7.48242736e-01 -2.15847651e-03 1.41346633e-01
7.18187168e-02 2.89147407e-01 4.40195620e-01 -4.04019564e-01
-6.79548264e-01 -7.74496654e-03 1.40170334e-02 -3.90664399e-01
-3.51141065e-01 -6.01604879e-01 -1.44037271e+00 -1.24264531e-01
2.91237645e-02 5.40615976e-01 5.64101279e-01 1.30283260e+00
2.71678299e-01 7.35445201e-01 3.66624802e-01 -1.28669488e+00
-8.75456572e-01 -5.65036833e-01 -3.51451486e-01 6.97334781e-02
5.67373216e-01 -9.43777382e-01 -4.34112996e-01 1.37764797e-01] | [6.4554643630981445, 3.468336820602417] |
b4a75a0a-f400-49dd-8110-7eb88db84217 | uadb-unsupervised-anomaly-detection-booster | 2306.01997 | null | https://arxiv.org/abs/2306.01997v1 | https://arxiv.org/pdf/2306.01997v1.pdf | UADB: Unsupervised Anomaly Detection Booster | Unsupervised Anomaly Detection (UAD) is a key data mining problem owing to its wide real-world applications. Due to the complete absence of supervision signals, UAD methods rely on implicit assumptions about anomalous patterns (e.g., scattered/sparsely/densely clustered) to detect anomalies. However, real-world data are complex and vary significantly across different domains. No single assumption can describe such complexity and be valid in all scenarios. This is also confirmed by recent research that shows no UAD method is omnipotent. Based on above observations, instead of searching for a magic universal winner assumption, we seek to design a general UAD Booster (UADB) that empowers any UAD models with adaptability to different data. This is a challenging task given the heterogeneous model structures and assumptions adopted by existing UAD methods. To achieve this, we dive deep into the UAD problem and find that compared to normal data, anomalies (i) lack clear structure/pattern in feature space, thus (ii) harder to learn by model without a suitable assumption, and finally, leads to (iii) high variance between different learners. In light of these findings, we propose to (i) distill the knowledge of the source UAD model to an imitation learner (booster) that holds no data assumption, then (ii) exploit the variance between them to perform automatic correction, and thus (iii) improve the booster over the original UAD model. We use a neural network as the booster for its strong expressive power as a universal approximator and ability to perform flexible post-hoc tuning. Note that UADB is a model-agnostic framework that can enhance heterogeneous UAD models in a unified way. Extensive experiments on over 80 tabular datasets demonstrate the effectiveness of UADB. | ['Jiang Bian', 'Yi Chang', 'Huishuai Zhang', 'Xiaofan Gui', 'Shun Zheng', 'Wei Cao', 'Xinyi Shen', 'Zhining Liu', 'Hangting Ye'] | 2023-06-03 | null | null | null | null | ['unsupervised-anomaly-detection'] | ['methodology'] | [ 2.20981110e-02 -1.58136442e-01 -1.27470121e-01 -1.38302505e-01
-3.22636455e-01 -3.35124195e-01 5.74933827e-01 2.48966515e-01
-5.99850267e-02 5.08702874e-01 -1.94069251e-01 -6.14699244e-01
-3.19482893e-01 -8.18970323e-01 -7.89378166e-01 -6.86357319e-01
-3.60476494e-01 4.46721107e-01 4.43758786e-01 -3.84101987e-01
6.51003420e-02 5.28905094e-01 -1.60382295e+00 -4.57335226e-02
1.19921231e+00 1.08493221e+00 -7.37658814e-02 2.16693789e-01
-2.01127067e-01 7.80908585e-01 -5.67085862e-01 -1.90238118e-01
4.91663486e-01 -4.48302805e-01 -5.39475381e-01 4.48290072e-02
1.11067571e-01 -2.02610567e-01 -2.86253631e-01 1.05985022e+00
2.57774115e-01 -1.07706888e-02 6.17259562e-01 -1.53814137e+00
-5.34515858e-01 5.59505403e-01 -6.38037801e-01 4.63168979e-01
1.71062052e-01 1.71491653e-01 9.80988204e-01 -7.65566289e-01
3.97316933e-01 1.00194001e+00 7.74202228e-01 5.82671046e-01
-1.21633041e+00 -5.89237750e-01 6.61420166e-01 1.80228308e-01
-1.27092528e+00 -7.92373568e-02 1.00320375e+00 -2.82583773e-01
7.45890141e-01 2.80799329e-01 4.00950730e-01 1.54172277e+00
8.46793689e-03 8.22564423e-01 1.03004968e+00 -3.44721764e-01
4.86994892e-01 1.61334544e-01 2.38724306e-01 5.49795032e-01
5.42975962e-01 9.98105109e-02 -3.63908201e-01 -3.41213763e-01
4.18308854e-01 3.25538099e-01 -1.94630623e-01 -6.32771373e-01
-8.19707096e-01 7.02319682e-01 3.95205706e-01 4.57186073e-01
-4.08060282e-01 -5.17838895e-01 4.80401576e-01 7.28129745e-01
2.85362989e-01 7.30394423e-01 -5.90640724e-01 -1.21574633e-01
-6.14221454e-01 2.72533119e-01 7.59004176e-01 6.76475108e-01
7.38007247e-01 2.70880133e-01 1.36959344e-01 8.24275732e-01
4.94920723e-02 3.31537455e-01 8.16624641e-01 -2.87546277e-01
3.17794234e-01 1.01299596e+00 -7.02159330e-02 -1.15208209e+00
-5.64901412e-01 -6.76261723e-01 -1.18555200e+00 3.01370591e-01
5.50445557e-01 -1.15190595e-01 -7.92910516e-01 1.85580063e+00
5.43168962e-01 5.72943568e-01 6.46905601e-02 7.39470363e-01
2.73102522e-01 3.18905115e-01 -2.49562219e-01 -2.04553992e-01
9.20156538e-01 -6.00042820e-01 -4.41207588e-01 -2.66464770e-01
8.27561617e-01 -2.44219542e-01 1.22463393e+00 6.42153919e-01
-4.58331645e-01 -4.15851563e-01 -1.27487600e+00 5.65032601e-01
-5.64254403e-01 -4.98257220e-01 6.49001837e-01 5.16787589e-01
-5.96516073e-01 5.02347708e-01 -9.38894033e-01 -4.73706335e-01
4.18161958e-01 1.37644723e-01 -3.64955217e-01 -2.28589028e-02
-1.36071980e+00 8.80293906e-01 5.83583355e-01 -9.17481445e-03
-8.58824015e-01 -7.13987887e-01 -6.51346147e-01 -1.83943793e-01
6.97567761e-01 -5.36035717e-01 9.59233820e-01 -1.06989765e+00
-1.08857298e+00 4.64536399e-01 7.33013898e-02 -8.62434983e-01
7.54087269e-01 -1.12296887e-01 -7.15297222e-01 -1.72235966e-01
-3.03615332e-02 -1.85401514e-01 9.73902822e-01 -1.45662153e+00
-8.23304713e-01 -3.99497122e-01 -9.14991796e-02 -1.45730786e-02
-5.48869550e-01 -2.73996204e-01 -1.73352078e-01 -7.46267915e-01
2.83402175e-01 -7.63972163e-01 -2.18106464e-01 -2.42894366e-01
-3.86246622e-01 -2.53920376e-01 1.21009088e+00 -3.96185935e-01
1.61334348e+00 -2.20958543e+00 -9.06217992e-02 6.75099730e-01
3.73115420e-01 3.95538211e-01 2.47368012e-02 3.04667503e-01
-1.87469065e-01 -5.07825390e-02 -6.57907963e-01 -1.32718295e-01
7.71140605e-02 7.15689301e-01 -5.88334024e-01 3.10624570e-01
3.55799556e-01 5.09915292e-01 -1.03655922e+00 -9.46570188e-02
1.02224953e-01 -3.76863331e-02 -6.63431942e-01 4.24764931e-01
-1.28089324e-01 3.58116210e-01 -6.02154315e-01 9.24531937e-01
6.87469721e-01 -2.05643684e-01 1.81978226e-01 1.05941303e-01
-1.66387819e-02 2.77109407e-02 -1.60231912e+00 1.15359676e+00
-3.10762554e-01 1.82848781e-01 -1.32596225e-01 -1.63737369e+00
1.21275711e+00 1.52076989e-01 6.27185404e-01 -8.07140350e-01
-7.35215098e-02 5.74839950e-01 2.20192730e-01 -4.76921618e-01
1.17547110e-01 7.64181912e-02 -3.78395966e-03 5.02409875e-01
1.52236581e-01 3.47651809e-01 4.22838666e-02 1.26641586e-01
1.49224436e+00 -4.07430008e-02 3.56398582e-01 -1.59125194e-01
6.05996370e-01 -3.42618257e-01 8.77644002e-01 1.17876244e+00
-3.14861029e-01 5.91692090e-01 6.79336786e-01 -8.40402424e-01
-8.05254638e-01 -1.18119466e+00 -3.28440487e-01 9.68951523e-01
1.39270231e-01 -2.53432363e-01 -4.05055791e-01 -1.06447554e+00
1.85543820e-01 7.62128115e-01 -8.45734417e-01 -5.34696877e-01
-4.99466687e-01 -1.18252337e+00 5.90728521e-01 3.59331459e-01
4.22761351e-01 -1.01115012e+00 -5.55990279e-01 2.85831690e-01
3.46149057e-02 -9.52440917e-01 -1.72207933e-02 6.37568772e-01
-8.67791712e-01 -1.16224766e+00 -1.13150172e-01 -3.81019205e-01
6.05328500e-01 1.56133398e-01 1.21460128e+00 2.42082596e-01
-1.19615898e-01 3.66835088e-01 -6.24214232e-01 -7.14660704e-01
-5.98848104e-01 4.18395288e-02 6.20230854e-01 2.84675777e-01
6.97935998e-01 -9.28477705e-01 -2.17613995e-01 4.49782521e-01
-1.19129658e+00 -4.98337477e-01 6.29792750e-01 1.04880846e+00
4.89524096e-01 3.39837343e-01 8.91528785e-01 -1.09379971e+00
6.37716174e-01 -8.96508098e-01 -5.66154063e-01 1.40671864e-01
-1.08371389e+00 4.44927253e-02 8.41205716e-01 -5.39421320e-01
-8.14998090e-01 -3.02608848e-01 -5.53223900e-02 -7.80701518e-01
-3.93465728e-01 5.99281311e-01 -3.50891590e-01 2.12378219e-01
9.20264363e-01 3.47393095e-01 1.18952766e-01 -5.95730007e-01
-8.80548358e-02 5.01476884e-01 6.12207830e-01 -8.11444581e-01
1.22365236e+00 4.18147057e-01 -1.81519434e-01 -8.24780703e-01
-9.25331116e-01 -3.29998344e-01 -7.27721334e-01 1.15233086e-01
3.65902483e-01 -7.11028039e-01 -3.06547582e-01 4.94124472e-01
-6.01599395e-01 -2.85809636e-01 -2.77681977e-01 2.79090405e-01
-2.40812063e-01 3.85933697e-01 -2.25730494e-01 -7.92590678e-01
-1.65151596e-01 -8.08388293e-01 5.04279971e-01 9.06874388e-02
-2.46594265e-01 -1.14147854e+00 1.51570946e-01 -1.01547003e-01
4.09388602e-01 4.62382764e-01 1.02628982e+00 -1.27298820e+00
-1.44589826e-01 -3.62870723e-01 -3.23304534e-02 5.46250761e-01
4.40465361e-01 9.67887696e-03 -1.04300952e+00 -2.77364194e-01
1.65628329e-01 -1.34930804e-01 6.95624709e-01 1.56674497e-02
1.43504429e+00 -3.38386565e-01 -1.62596524e-01 5.79336047e-01
1.16527283e+00 1.09028779e-01 3.97855550e-01 7.34430134e-01
6.19665980e-01 3.62275600e-01 5.86149454e-01 5.99647641e-01
3.52579683e-01 5.92254043e-01 8.32249820e-01 -1.69417262e-01
3.19409102e-01 -4.42108154e-01 5.34962118e-01 7.71981239e-01
-3.44308726e-02 4.08685803e-02 -1.05801892e+00 5.80613852e-01
-1.99083948e+00 -1.09984827e+00 -5.01940073e-03 2.31416154e+00
6.85644746e-01 5.76056838e-01 3.50925148e-01 4.19832259e-01
4.38863248e-01 1.77727696e-02 -8.14263642e-01 -5.05984962e-01
-3.22412550e-01 -3.94126996e-02 1.77596882e-01 -8.34277943e-02
-1.14270318e+00 5.74901640e-01 5.19985533e+00 7.09950864e-01
-1.18350554e+00 -2.17002854e-01 4.22064334e-01 2.41211146e-01
-2.39797756e-01 -1.86631113e-01 -4.37014669e-01 6.16439939e-01
7.60010242e-01 -8.95669684e-02 3.95033062e-01 9.03958261e-01
-1.11775063e-01 3.13190848e-01 -1.23292446e+00 8.00920963e-01
-7.24975243e-02 -9.24956203e-01 2.17252627e-01 2.19706278e-02
7.41769195e-01 1.42223328e-01 1.32732745e-03 6.74201548e-01
3.68877470e-01 -8.87367606e-01 5.06327689e-01 3.15644562e-01
1.75414696e-01 -7.59152412e-01 9.01906729e-01 6.75364077e-01
-8.24308217e-01 -4.78634655e-01 -4.47062761e-01 -1.11968718e-01
-2.71351010e-01 6.50288880e-01 -8.21306705e-01 8.89861286e-01
1.03276491e+00 7.39837408e-01 -6.71918929e-01 9.01454806e-01
-3.54831517e-02 9.61208284e-01 -4.86705422e-01 3.39780986e-01
4.00732785e-01 -2.53716558e-01 8.30287457e-01 1.08027363e+00
4.33096230e-01 -2.18552753e-01 3.17438930e-01 6.75071836e-01
3.78346303e-04 6.03660084e-02 -9.26477015e-01 3.13876450e-01
4.67005968e-01 1.00928438e+00 -3.94024223e-01 -3.66118997e-02
-7.35748053e-01 6.76872075e-01 3.82406831e-01 3.69210243e-01
-6.73587263e-01 -9.55941901e-02 8.26546729e-01 1.64978996e-01
2.77523994e-01 3.29187214e-02 -2.97885835e-01 -1.43319058e+00
2.32398972e-01 -1.32483268e+00 8.17782044e-01 -9.21342820e-02
-1.91663897e+00 6.83430851e-01 -1.48751676e-01 -1.65730739e+00
-1.02382019e-01 -6.35073602e-01 -8.40865195e-01 4.80654389e-01
-1.64974749e+00 -9.14886296e-01 -2.79405683e-01 9.32382464e-01
3.84237230e-01 -3.67414653e-01 8.89870644e-01 3.36805642e-01
-8.66942346e-01 7.85324991e-01 1.51236102e-01 2.56316423e-01
7.60768473e-01 -1.49781251e+00 1.35669440e-01 1.12127042e+00
2.24264607e-01 6.40325904e-01 7.73409784e-01 -4.73616749e-01
-1.31971848e+00 -1.25608003e+00 4.38327223e-01 -7.43299186e-01
9.43247080e-01 -2.88900584e-01 -1.58911550e+00 6.59118712e-01
-2.25285634e-01 2.16664344e-01 6.89873040e-01 4.15266126e-01
-5.84374547e-01 -3.70820522e-01 -1.12826979e+00 6.28829002e-01
1.03707004e+00 -2.11914062e-01 -8.30104291e-01 -1.19108837e-02
4.85702544e-01 -2.77967721e-01 -6.64745629e-01 7.63585806e-01
2.07579046e-01 -1.09516203e+00 8.07174027e-01 -8.83581102e-01
2.42467210e-01 -4.04605299e-01 -2.30092645e-01 -1.33458149e+00
-1.43039793e-01 -4.34755057e-01 -5.96529186e-01 1.30685306e+00
4.50127542e-01 -1.01581025e+00 3.06211263e-01 5.94357312e-01
-2.59235442e-01 -8.65676403e-01 -1.00335789e+00 -1.09460366e+00
7.24331364e-02 -7.54144907e-01 8.52519810e-01 1.36188531e+00
-8.28613415e-02 -7.56758451e-02 -5.49265862e-01 5.17922878e-01
5.44498265e-01 1.86279088e-01 9.91922855e-01 -1.68852723e+00
-2.26702794e-01 -5.34644723e-01 -4.57509398e-01 -6.98704243e-01
1.91115960e-02 -8.97075891e-01 -1.64662108e-01 -9.26831186e-01
-2.04781801e-01 -8.43410194e-01 -8.74513566e-01 7.05961347e-01
-3.77775073e-01 -1.73993409e-02 -1.43294409e-01 2.68333137e-01
-6.52667403e-01 5.64419389e-01 7.55632401e-01 -5.45745566e-02
-4.48693097e-01 2.83229917e-01 -9.04843271e-01 9.20716703e-01
7.86332369e-01 -4.31347966e-01 -3.77359867e-01 -8.67178962e-02
2.07364529e-01 -4.28615212e-01 3.38239759e-01 -1.05822265e+00
-6.20279089e-03 -1.51404455e-01 3.24489832e-01 -1.91192284e-01
-2.56726384e-01 -1.11068416e+00 -2.59839773e-01 5.04062116e-01
2.22949516e-02 1.07543796e-01 9.24747884e-02 6.82879865e-01
-3.76399398e-01 -1.15182273e-01 6.53714240e-01 -3.28192189e-02
-8.51519287e-01 4.23261881e-01 -2.27621302e-01 2.51356065e-01
1.01050448e+00 -2.94439018e-01 -1.30279198e-01 -2.68611521e-01
-4.45996761e-01 4.49815184e-01 4.16316003e-01 7.99665451e-01
3.37548345e-01 -1.31675375e+00 -6.55737877e-01 6.30893886e-01
4.44689214e-01 3.18989903e-01 1.15271404e-01 9.29293275e-01
-6.33372068e-02 -3.58003862e-02 -1.30765378e-01 -7.42604852e-01
-7.99076378e-01 8.59119654e-01 4.16835725e-01 -4.67345059e-01
-8.22076797e-01 3.41618657e-01 1.16623491e-01 -6.73970163e-01
2.48502448e-01 -3.87951173e-02 -1.60152242e-01 1.80666938e-01
5.48577487e-01 1.77695870e-01 3.07969481e-01 -2.55440712e-01
-3.35521340e-01 2.12572128e-01 -2.01833844e-01 6.14500046e-01
1.55452538e+00 -5.16338758e-02 1.36671588e-01 5.96551776e-01
7.16338336e-01 -1.05052544e-02 -1.23904347e+00 -5.85456133e-01
3.97627652e-01 -4.00551736e-01 -4.04528737e-01 -6.61036491e-01
-9.57171500e-01 7.00761795e-01 5.09033442e-01 6.80169463e-01
1.30036688e+00 -1.34585053e-01 4.43801552e-01 4.86856431e-01
2.32771710e-01 -1.19603169e+00 1.42782718e-01 5.88320076e-01
6.90621316e-01 -1.42016363e+00 -2.49025494e-01 1.78665623e-01
-7.93769836e-01 1.16420293e+00 1.00959945e+00 -1.92228332e-01
7.42798269e-01 1.61251992e-01 2.50650078e-01 -1.20840706e-01
-8.26127470e-01 -2.32543468e-01 1.59929022e-01 6.84288323e-01
-5.27816601e-02 -1.65802881e-01 -3.83149944e-02 7.85868406e-01
-2.65937075e-02 -2.97811210e-01 4.41079974e-01 8.13763738e-01
-3.27487409e-01 -1.05130899e+00 -3.32386613e-01 6.41097784e-01
-3.32061589e-01 3.10612708e-01 -1.56291664e-01 9.10587549e-01
2.38384902e-01 8.13444018e-01 2.02319678e-02 -3.81380558e-01
5.54257810e-01 3.83978546e-01 -1.60587862e-01 -2.62074798e-01
-3.46194327e-01 -2.57012516e-01 -4.05901909e-01 -5.87891281e-01
-9.55001041e-02 -6.07085168e-01 -1.01135468e+00 -1.02701537e-01
-1.13624446e-01 2.14831784e-01 2.27740958e-01 1.09480488e+00
3.65097672e-01 6.10916495e-01 1.00773227e+00 -3.46236765e-01
-8.98216665e-01 -8.19537461e-01 -6.08188868e-01 8.00545514e-01
5.05069554e-01 -9.21756625e-01 -7.13202357e-01 -3.17926258e-01] | [7.592746257781982, 2.4586031436920166] |
16872667-1f53-4253-8010-fe681d23b706 | a-neural-graph-based-local-coherence-model | null | null | https://aclanthology.org/2021.findings-emnlp.199 | https://aclanthology.org/2021.findings-emnlp.199.pdf | A Neural Graph-based Local Coherence Model | Entity grids and entity graphs are two frameworks for modeling local coherence. These frameworks represent entity relations between sentences and then extract features from such representations to encode coherence. The benefits of convolutional neural models for extracting informative features from entity grids have been recently studied. In this work, we study the benefits of Relational Graph Convolutional Networks (RGCN) to encode entity graphs for measuring local coherence. We evaluate our neural graph-based model for two benchmark coherence evaluation tasks: sentence ordering (SO) and summary coherence rating (SCR). The results show that our neural graph-based model consistently outperforms the neural grid-based model for both tasks. Our model performs competitively with a strong baseline coherence model, while our model uses 50% fewer parameters. Our work defines a new, efficient, and effective baseline for local coherence modeling. | ['Iryna Gurevych', 'Leonardo F. R. Ribeiro', 'Mohsen Mesgar'] | null | null | null | null | findings-emnlp-2021-11 | ['sentence-ordering', 'coherence-evaluation'] | ['natural-language-processing', 'natural-language-processing'] | [-3.25838715e-01 6.11128092e-01 -4.00443852e-01 -5.27596235e-01
-8.21302295e-01 -2.32000008e-01 6.97036564e-01 5.56684613e-01
-1.54550731e-01 5.48150897e-01 1.39275789e+00 4.94723916e-02
-7.82401487e-02 -1.07890379e+00 -5.09023130e-01 -1.41498566e-01
-4.87947196e-01 2.67396659e-01 3.32217030e-02 -3.45694780e-01
4.13909554e-01 1.92822590e-01 -8.48295987e-01 4.99216348e-01
8.52174461e-01 5.20417750e-01 -1.35219902e-01 8.94273162e-01
-2.94384789e-02 1.64894760e+00 -8.26696992e-01 -2.67602056e-01
-4.64866906e-01 -4.07646418e-01 -1.37115419e+00 -3.75188291e-01
6.19724751e-01 -3.64330977e-01 -1.02600121e+00 7.27692544e-01
5.34417391e-01 1.72360256e-01 4.64924037e-01 -8.91133547e-01
-9.23702598e-01 1.42599154e+00 -3.11839491e-01 6.18999362e-01
5.36650419e-01 -4.07859823e-03 1.81246912e+00 -6.07210875e-01
9.50990975e-01 1.18337238e+00 9.31901634e-01 -2.84790881e-02
-1.24088562e+00 -2.42191032e-01 2.77074613e-02 2.44208962e-01
-1.37913346e+00 -6.12931371e-01 7.90783942e-01 -2.85866231e-01
1.81250930e+00 2.65446097e-01 8.17488909e-01 1.04368925e+00
4.90708411e-01 7.94975579e-01 5.61385095e-01 -2.83838082e-02
-1.10357285e-01 -5.90777993e-01 9.55712318e-01 6.51328743e-01
4.00853038e-01 -2.32678622e-01 -9.00571704e-01 -5.68533465e-02
3.91585648e-01 -5.46002150e-01 -3.32622319e-01 2.34174594e-01
-1.22352135e+00 8.80136192e-01 1.12412953e+00 6.38387620e-01
-3.45468670e-01 7.38617659e-01 5.57567775e-01 3.45753878e-01
9.32816505e-01 9.72718418e-01 5.46577200e-02 -2.06794247e-01
-1.02786708e+00 4.18389380e-01 1.02359951e+00 1.03649890e+00
6.26245737e-01 -1.34810954e-01 -8.61307681e-01 6.78494155e-01
1.53225049e-01 1.33899957e-01 5.83187677e-02 -1.06663120e+00
7.15597808e-01 7.42243171e-01 -3.32704604e-01 -1.74186897e+00
-9.70771909e-01 -7.96559572e-01 -9.99538720e-01 -8.29817414e-01
-1.84853628e-01 1.34656027e-01 -8.55743960e-02 1.86440039e+00
-2.41490185e-01 1.37518868e-01 -9.19727609e-02 4.78043824e-01
1.62369967e+00 6.96506739e-01 -9.25029218e-02 -8.74210820e-02
1.03168058e+00 -1.14718449e+00 -1.00689077e+00 -1.41115949e-01
1.13888717e+00 -4.49785143e-01 1.06283116e+00 -1.22794800e-01
-1.29502177e+00 -4.46905881e-01 -1.15445852e+00 -5.69722354e-01
-2.37000659e-02 -4.49918099e-02 9.21132505e-01 9.80513636e-03
-1.83161259e+00 9.03803766e-01 -9.38533902e-01 -4.77210909e-01
4.47173715e-01 2.31570452e-01 -3.07870984e-01 2.05290079e-01
-1.38888216e+00 9.67106044e-01 4.49224889e-01 -3.18078771e-02
-3.67007107e-01 -6.01505637e-01 -1.02134681e+00 6.15047157e-01
-6.79143751e-03 -1.15714705e+00 9.25721705e-01 -2.66649723e-01
-1.11484277e+00 6.37733161e-01 -1.60103828e-01 -8.58199716e-01
-2.94495612e-01 -2.96162724e-01 -2.45790482e-01 1.34559661e-01
1.86953053e-01 5.84719419e-01 -1.37148842e-01 -7.24250972e-01
1.45355329e-01 4.54999926e-03 2.76600897e-01 2.59596258e-01
-4.80008900e-01 -5.04447334e-03 -4.78926957e-01 -6.03455245e-01
-6.34397194e-02 -6.18773758e-01 -6.40645027e-02 -8.46813500e-01
-9.84784007e-01 -7.11808562e-01 3.26411694e-01 -6.23171806e-01
2.06514859e+00 -1.70383286e+00 1.98462561e-01 -1.53342202e-01
9.28560197e-01 -3.36779892e-01 -6.49233580e-01 9.31777418e-01
1.50391817e-01 5.50997734e-01 9.16541591e-02 -4.42887783e-01
4.36620265e-02 -1.57809511e-01 -2.42684618e-01 1.01767786e-01
3.74957770e-01 1.57723367e+00 -8.82655263e-01 -5.97025514e-01
-2.28641376e-01 5.10623574e-01 -9.68988359e-01 3.30545098e-01
-1.24143787e-01 6.97491467e-02 -2.09844038e-01 1.44778222e-01
3.60625952e-01 -1.01014805e+00 6.24742627e-01 -5.75636148e-01
2.65303731e-01 1.07310617e+00 -4.27722186e-01 1.93928194e+00
-4.02728856e-01 1.19280469e+00 -6.75400853e-01 -5.97677469e-01
7.81664491e-01 1.42766103e-01 4.21674550e-01 -9.60872769e-01
-9.66810957e-02 -2.47058630e-01 1.22966565e-01 -2.25444496e-01
1.14933074e+00 6.83868289e-01 -4.63543594e-01 7.13932395e-01
3.21694613e-01 -1.99336633e-01 3.32425147e-01 1.02407742e+00
1.84571576e+00 -2.41772041e-01 4.89609510e-01 -5.86038649e-01
1.18982151e-01 -3.02183986e-01 2.75491983e-01 7.93868840e-01
7.18509853e-02 5.13330162e-01 1.08497608e+00 -4.53786165e-01
-9.45930302e-01 -9.96054649e-01 2.18743831e-01 8.25021327e-01
6.80176243e-02 -1.39257014e+00 -6.54874086e-01 -5.58506012e-01
-2.27737606e-01 6.41935766e-01 -6.89859927e-01 -3.19268823e-01
-9.98745263e-01 -4.71889973e-01 6.57824636e-01 7.08455324e-01
4.31847185e-01 -1.01878214e+00 -1.18741289e-01 1.63437188e-01
-6.93675518e-01 -1.39878023e+00 -5.38430870e-01 -2.41673678e-01
-7.79137492e-01 -8.77413332e-01 1.68176755e-01 -4.05814052e-01
2.09044486e-01 3.23939919e-01 2.22259402e+00 5.09828806e-01
1.80905879e-01 1.87368721e-01 -2.87020624e-01 2.81818062e-01
-2.58959025e-01 6.81893766e-01 -2.56471872e-01 -7.06710219e-01
2.28380233e-01 -7.68215716e-01 -7.13487566e-01 -2.96705425e-01
-6.26170099e-01 2.97643989e-01 2.77426988e-01 7.77869642e-01
2.76708335e-01 -3.24853659e-01 8.00202787e-01 -1.22268999e+00
1.36781204e+00 -5.58483243e-01 -1.16697580e-01 1.17843047e-01
-5.20705640e-01 2.86124438e-01 3.39113325e-01 4.03942913e-01
-5.83387077e-01 -7.72863925e-01 -8.59070495e-02 8.60202461e-02
3.47886980e-01 1.23039949e+00 1.61148995e-01 3.99854064e-01
8.24699104e-01 -3.99301171e-01 -5.99673510e-01 3.71522978e-02
7.52550185e-01 3.10584128e-01 6.19254529e-01 -4.76711541e-01
3.43921751e-01 7.70308077e-02 -1.63200572e-02 -5.85982323e-01
-1.14960611e+00 -3.05989057e-01 -7.41499305e-01 4.24102619e-02
9.24065650e-01 -1.26605201e+00 -3.47205520e-01 4.89709191e-02
-1.65193367e+00 -4.59313452e-01 -3.01416039e-01 1.64280310e-01
-4.08471912e-01 3.95508707e-01 -9.99146819e-01 -2.44311869e-01
-8.07720065e-01 -8.21117878e-01 1.21695817e+00 2.84036435e-02
-6.94109559e-01 -1.36939347e+00 6.04157031e-01 2.40253955e-01
7.42557168e-01 6.16276681e-01 8.60857546e-01 -4.94917989e-01
-9.65450943e-01 1.49139222e-02 -4.33068246e-01 -1.46224156e-01
-4.89334986e-02 2.34068096e-01 -8.91206622e-01 -1.70326769e-01
-6.14577949e-01 -5.61354756e-01 1.26985335e+00 7.16480255e-01
1.01071429e+00 -6.77453101e-01 -3.05660546e-01 6.00260377e-01
1.26765239e+00 -6.25201643e-01 7.54157484e-01 1.15861937e-01
8.24859262e-01 2.81570077e-01 1.06715880e-01 4.43095148e-01
1.00951588e+00 7.39672422e-01 1.42247230e-01 -5.08383326e-02
-6.21884227e-01 -1.07407011e-01 2.75562435e-01 1.73316216e+00
-7.61934370e-02 -6.70817673e-01 -1.05589569e+00 5.16129255e-01
-2.07573485e+00 -1.23670602e+00 -2.85635084e-01 1.45908248e+00
1.07882166e+00 2.52246946e-01 -1.67682245e-02 -4.34747010e-01
3.83400232e-01 8.50773811e-01 3.87380756e-02 -5.38412213e-01
-4.22723770e-01 3.42294842e-01 -1.91159472e-02 8.07802916e-01
-1.20135522e+00 1.11684549e+00 6.64023590e+00 5.54600239e-01
-6.45522773e-01 -2.88987774e-02 6.93565190e-01 -1.88816175e-01
-8.66395175e-01 4.42329310e-02 -4.76048619e-01 1.68772653e-01
1.38145149e+00 -3.53420019e-01 2.76294947e-01 3.57599914e-01
1.54016793e-01 -7.44056702e-03 -1.39347243e+00 7.32939065e-01
-1.83882825e-02 -2.05801225e+00 -5.40528409e-02 -2.20650174e-02
9.54025447e-01 6.13633394e-01 -1.16784923e-01 3.86429071e-01
9.02288139e-01 -1.38021147e+00 3.47738415e-01 1.02217376e+00
5.84045529e-01 -7.19702661e-01 9.74298596e-01 -1.02347352e-01
-1.46798217e+00 4.54816878e-01 -5.25832474e-01 -9.12605897e-02
1.03429750e-01 9.93488014e-01 -6.74174428e-01 7.70087421e-01
5.20273328e-01 1.30897200e+00 -1.05878186e+00 6.09373689e-01
-2.59492815e-01 7.42333829e-01 -7.74741080e-03 -1.38331622e-01
1.08530194e-01 1.90190282e-02 4.82752085e-01 1.64617920e+00
9.04491842e-02 -1.67268381e-01 3.21410336e-02 1.41425121e+00
-5.78525007e-01 9.40341875e-03 -9.72849607e-01 -4.48669016e-01
8.74574602e-01 1.37205756e+00 -4.38369751e-01 -2.49852881e-01
-2.77829021e-01 7.98942387e-01 1.23639309e+00 2.16113165e-01
-7.94947922e-01 -4.23914850e-01 3.81376624e-01 -2.60721236e-01
-3.58430408e-02 -4.86374736e-01 -6.63659096e-01 -1.38564801e+00
-2.39962831e-01 -5.42591751e-01 3.69521350e-01 -7.19927728e-01
-1.34873092e+00 1.00378871e+00 -2.35228371e-02 -8.48191679e-01
-6.09461486e-01 2.70800740e-01 -8.72677147e-01 8.26424599e-01
-1.24738073e+00 -1.29472387e+00 -5.38752198e-01 2.56242096e-01
9.11745429e-02 -5.42688034e-02 9.09788728e-01 -1.39176279e-01
-5.24173856e-01 5.50550282e-01 -3.14060211e-01 4.03935224e-01
4.36432779e-01 -1.54122651e+00 1.20781803e+00 8.37682784e-01
5.26341021e-01 1.10670567e+00 5.49363375e-01 -6.25769198e-01
-1.07650781e+00 -1.25306976e+00 1.58536100e+00 -7.23541200e-01
8.21635902e-01 -3.76379281e-01 -7.83785939e-01 7.71616280e-01
1.05326056e+00 9.02166590e-03 8.37153018e-01 1.03824890e+00
-5.92340171e-01 1.03707775e-01 -3.26600760e-01 5.84864974e-01
1.33728325e+00 -8.81273210e-01 -3.92472804e-01 5.45660555e-01
1.35374415e+00 -2.90654033e-01 -1.43143606e+00 4.38366175e-01
1.70875177e-01 -1.22877431e+00 8.62118721e-01 -5.65132141e-01
1.05337715e+00 5.47484756e-02 -1.22836158e-01 -1.46271276e+00
-8.78614962e-01 -4.61383343e-01 -6.78018034e-01 1.16382194e+00
4.97954488e-01 -1.92994907e-01 5.02519011e-01 1.58922255e-01
-4.69687194e-01 -9.97794151e-01 -6.61120892e-01 -4.45494056e-01
2.70157725e-01 -9.06930491e-02 6.17239296e-01 1.10482299e+00
3.95745665e-01 1.18316293e+00 -1.32345855e-01 -1.33852154e-01
1.62208304e-01 3.63024563e-01 7.72464216e-01 -1.03256011e+00
-3.94577950e-01 -8.79841685e-01 -3.14238906e-01 -9.05387998e-01
5.24669945e-01 -1.31277192e+00 -3.45063418e-01 -2.30578327e+00
6.61886394e-01 -1.25853375e-01 -2.03310475e-01 2.31137767e-01
-5.51662743e-01 -2.38593202e-02 2.03571975e-01 2.40222797e-01
-1.16167116e+00 6.90102339e-01 1.04603100e+00 -3.94322366e-01
-4.18276787e-02 -4.61801589e-01 -8.74677122e-01 1.89555719e-01
8.05255055e-01 -1.81915328e-01 -5.95370054e-01 -7.98822820e-01
9.53895748e-01 1.54812902e-01 1.29291818e-01 -9.56303835e-01
4.49786574e-01 2.69182384e-01 1.01506323e-01 -8.44061136e-01
1.95707306e-02 2.70336438e-02 1.33872759e-02 1.06569566e-01
-8.79045784e-01 3.85282367e-01 -2.41656862e-02 5.19278109e-01
-4.72390622e-01 3.48861694e-01 3.82159621e-01 -3.06577934e-03
-3.61833423e-01 2.87256509e-01 1.80703495e-02 4.89554435e-01
1.67475328e-01 4.34464425e-01 -1.22973871e+00 -8.50248754e-01
-6.56702757e-01 9.72523168e-02 1.99743390e-01 3.26411515e-01
7.68440545e-01 -1.68601859e+00 -9.94199038e-01 -4.22183633e-01
1.19811893e-01 -1.42750263e-01 1.54879272e-01 1.06534600e+00
-6.76453173e-01 8.20204377e-01 1.21254504e-01 -5.66295207e-01
-1.00854719e+00 3.24075848e-01 2.74250329e-01 -1.02682567e+00
-6.79376960e-01 9.61512208e-01 2.07356006e-01 -4.92811024e-01
-1.04849674e-01 -7.36278355e-01 -4.81373012e-01 -1.89768925e-01
2.01162517e-01 2.75301009e-01 1.56769082e-01 -5.19689798e-01
-3.65037352e-01 2.51863629e-01 -1.25936583e-01 1.32632434e-01
1.40585482e+00 -1.87657371e-01 -6.61913097e-01 4.29791331e-01
1.45428121e+00 1.79959118e-01 -5.49310565e-01 -4.33175981e-01
4.49858963e-01 -1.20408852e-02 9.52415317e-02 -4.96720940e-01
-8.88553500e-01 8.20298553e-01 -1.71158016e-01 6.68278992e-01
8.87540221e-01 3.29587996e-01 7.21599460e-01 4.99683887e-01
1.23950750e-01 -9.52885151e-01 4.98752326e-01 1.06980920e+00
1.25902081e+00 -1.07904088e+00 3.08694839e-01 -3.51736665e-01
-6.13423884e-01 1.17027926e+00 6.13513947e-01 -5.39467692e-01
5.81934512e-01 3.35365146e-01 -4.83094841e-01 -7.97673583e-01
-1.42449808e+00 -2.11023375e-01 9.45110798e-01 2.81854481e-01
1.24005830e+00 3.27318907e-01 -4.35734153e-01 7.32863188e-01
-6.29502058e-01 -3.28062415e-01 6.23125672e-01 4.75644827e-01
-4.50571656e-01 -7.76397347e-01 5.78104794e-01 6.18922353e-01
-1.21925287e-01 -8.66194546e-01 -7.67966688e-01 7.29512453e-01
-5.69248080e-01 1.07864857e+00 3.68977249e-01 -9.41527963e-01
8.20545331e-02 -3.62900019e-01 4.55565721e-01 -7.96821415e-01
-8.54596972e-01 -2.32516490e-02 9.27398145e-01 -1.11822188e+00
-6.54557288e-01 -4.15518224e-01 -1.21724212e+00 -7.22461402e-01
-3.53033960e-01 4.13749367e-02 -3.15122008e-02 6.25637591e-01
7.24827409e-01 1.09930217e+00 6.37830675e-01 -3.37357581e-01
-1.34569332e-01 -1.32649624e+00 -2.66208082e-01 2.85932481e-01
2.00679064e-01 -2.05908164e-01 2.66180243e-02 -4.39843893e-01] | [11.791855812072754, 9.163168907165527] |
419dc9ae-92e9-41b9-9d20-97bdd1e7352e | on-a-built-in-conflict-between-deep-learning | 2208.11633 | null | https://arxiv.org/abs/2208.11633v1 | https://arxiv.org/pdf/2208.11633v1.pdf | On a Built-in Conflict between Deep Learning and Systematic Generalization | In this paper, we hypothesize that internal function sharing is one of the reasons to weaken o.o.d. or systematic generalization in deep learning for classification tasks. Under equivalent prediction, a model partitions an input space into multiple parts separated by boundaries. The function sharing prefers to reuse boundaries, leading to fewer parts for new outputs, which conflicts with systematic generalization. We show such phenomena in standard deep learning models, such as fully connected, convolutional, residual networks, LSTMs, and (Vision) Transformers. We hope this study provides novel insights into systematic generalization and forms a basis for new research directions. | ['Yuanpeng Li'] | 2022-08-24 | null | null | null | null | ['systematic-generalization'] | ['reasoning'] | [ 1.95652574e-01 4.84799266e-01 -3.70455086e-01 -4.41150278e-01
3.64424139e-01 -5.19014835e-01 2.03082204e-01 -3.79465938e-01
-1.59948200e-01 7.82635868e-01 7.28273168e-02 -5.29481351e-01
-2.96052188e-01 -1.01309180e+00 -8.91740799e-01 -7.30175257e-01
-1.34487990e-02 -7.51986355e-02 2.02197760e-01 -6.82058781e-02
3.19580757e-03 6.49162114e-01 -1.42669237e+00 5.05481899e-01
7.39851534e-01 1.21175635e+00 4.03586209e-01 1.28161773e-01
-1.32478029e-01 8.82148206e-01 -5.84047735e-01 -4.05549377e-01
2.30090261e-01 -2.66816944e-01 -1.12042820e+00 -1.20961741e-01
2.41988346e-01 -2.53289163e-01 -8.34261298e-01 1.12648880e+00
1.39634479e-02 3.65049094e-01 6.95241868e-01 -1.28664339e+00
-1.47052169e+00 1.14574468e+00 3.09359469e-02 8.00255537e-02
-3.15582633e-01 1.43113092e-01 8.74281764e-01 -7.96879530e-01
1.96875423e-01 1.35015380e+00 9.07261431e-01 8.39092672e-01
-1.19249272e+00 -5.00660717e-01 5.59082150e-01 -1.14224181e-01
-1.25353956e+00 -2.25715488e-01 5.29708982e-01 -2.26707995e-01
1.03965521e+00 2.22832516e-01 7.52480626e-01 1.25623965e+00
5.80337167e-01 9.09338117e-01 9.10735726e-01 -1.38178989e-01
4.07123193e-02 3.00969601e-01 4.45957392e-01 6.29541516e-01
6.17831528e-01 1.03635877e-01 -2.30358794e-01 2.14018643e-01
1.08856714e+00 6.16094708e-01 -6.57280147e-01 -2.32341096e-01
-7.65508056e-01 4.47584301e-01 9.44260716e-01 5.11008620e-01
-2.30424955e-01 3.65696609e-01 2.96788037e-01 7.07217336e-01
3.21088433e-01 6.82523251e-01 -9.40040052e-01 3.19880009e-01
-5.66335976e-01 -1.08189911e-01 5.57063103e-01 1.02048647e+00
1.03195238e+00 3.85834962e-01 -7.72944242e-02 7.27566600e-01
-1.53250406e-02 1.33617103e-01 1.08174241e+00 -1.12094557e+00
2.14331478e-01 7.98496604e-01 -4.05967593e-01 -6.58578992e-01
-5.43437600e-01 -1.13757753e+00 -1.03013432e+00 -1.99789613e-01
1.62943173e-03 -1.16287261e-01 -9.04223979e-01 1.97501290e+00
-5.27801871e-01 -1.06728747e-01 1.13924205e-01 4.83747303e-01
6.00700200e-01 5.31010449e-01 -7.71493241e-02 -4.67268899e-02
8.61623585e-01 -1.17016768e+00 -6.21647894e-01 -2.13097110e-01
9.96918142e-01 -1.39446050e-01 1.15025270e+00 6.02987468e-01
-1.06741679e+00 -1.09592509e+00 -9.45987403e-01 1.13824653e-02
-6.54085100e-01 4.43039797e-02 9.25426245e-01 7.81504989e-01
-1.49025321e+00 9.99334335e-01 -5.53295076e-01 -4.28572804e-01
5.51436365e-01 4.83511299e-01 -3.80906463e-01 5.38199022e-02
-1.33550930e+00 1.04669535e+00 7.43719876e-01 2.97969908e-01
-9.88112271e-01 -5.27337015e-01 -6.91464365e-01 3.68376374e-01
8.41962770e-02 -8.28005910e-01 1.33455181e+00 -1.20328903e+00
-1.37642479e+00 5.25019288e-01 -3.48370895e-02 -7.98936486e-01
-2.79200561e-02 -4.14325297e-01 -3.93802315e-01 -4.24328178e-01
-2.30966732e-01 8.11775982e-01 6.94886327e-01 -1.18173516e+00
-4.54901576e-01 -5.84611893e-01 2.18996942e-01 1.87759299e-03
-8.71942520e-01 -5.72725177e-01 1.71360657e-01 -6.02851093e-01
3.83892775e-01 -7.74671257e-01 -1.86417505e-01 -2.37100840e-01
-3.85244459e-01 -4.69537586e-01 7.28877127e-01 -1.84554264e-01
1.44122386e+00 -2.27781153e+00 1.89459354e-01 -8.54877234e-02
4.73657340e-01 3.96135822e-02 -6.98929429e-02 2.58746952e-01
-3.57630730e-01 5.60095072e-01 1.42683357e-01 -1.43258423e-01
-2.71006599e-02 5.46266735e-01 -7.78725147e-01 4.01929498e-01
-1.83972418e-02 1.12929904e+00 -5.74363291e-01 1.90758169e-01
-1.57303691e-01 1.46256283e-01 -3.34506422e-01 -6.68061823e-02
-1.05565779e-01 8.79015848e-02 -3.56424868e-01 5.36413789e-01
7.41026163e-01 -4.01059866e-01 -1.59432199e-02 -4.90619205e-02
-3.23732793e-01 6.07139468e-01 -6.73186302e-01 1.41068745e+00
-4.80417788e-01 7.28228867e-01 -4.03265983e-01 -1.44371665e+00
1.10936213e+00 2.93033510e-01 -2.27817103e-01 -7.30195999e-01
1.05671354e-01 3.84880424e-01 3.59021991e-01 -1.26189590e-01
4.34906691e-01 -2.03084201e-02 1.66607320e-01 1.65743589e-01
3.96519601e-01 3.08708668e-01 -4.21940804e-01 -2.11321905e-01
9.67888951e-01 8.46054927e-02 9.61043537e-02 -5.15668094e-01
3.26209277e-01 -4.41922486e-01 6.68053567e-01 9.58320320e-01
-1.28253967e-01 2.51582474e-01 4.81366038e-01 -7.54523218e-01
-7.72061408e-01 -1.09852099e+00 -2.94343442e-01 1.15581691e+00
9.20765176e-02 -2.29150176e-01 -5.95135510e-01 -6.28882945e-01
2.94866543e-02 6.23512566e-01 -8.92191648e-01 -9.31935549e-01
-5.76197743e-01 -3.32367182e-01 9.02282715e-01 9.79781806e-01
8.68475735e-01 -1.01304567e+00 -4.42005932e-01 7.40306303e-02
3.54056031e-01 -6.42822504e-01 -6.77287281e-02 8.02002132e-01
-1.64829075e+00 -6.29383028e-01 -7.35996544e-01 -1.12473047e+00
6.89027846e-01 4.16153938e-01 9.29882884e-01 5.86034767e-02
3.79239976e-01 2.72362512e-02 -1.56039357e-01 1.71372795e-03
1.37204956e-02 3.82485539e-01 3.93369496e-01 -3.90570551e-01
1.44235715e-01 -5.97988248e-01 -5.66321254e-01 5.10577857e-01
-8.62628043e-01 -8.25662836e-02 6.72844291e-01 9.64514673e-01
3.03415924e-01 3.25538337e-01 7.26648927e-01 -8.77717674e-01
8.41849446e-01 -5.08378088e-01 -2.39775687e-01 4.48720694e-01
-6.73032820e-01 3.01766574e-01 1.15383089e+00 -5.34751713e-01
-1.13444531e+00 -3.83365184e-01 1.96846381e-01 -6.00226402e-01
-2.21787710e-02 5.95510602e-01 -2.53766209e-01 -1.72434539e-01
7.10419416e-01 6.72155142e-01 -4.13775861e-01 -7.77265429e-01
3.06354254e-01 5.60050189e-01 1.54331908e-01 -4.77085620e-01
3.01592827e-01 7.41667077e-02 -2.07003489e-01 -6.36758566e-01
-7.35065043e-01 1.37389228e-01 -4.82215315e-01 6.76783323e-02
5.60904145e-01 -7.78766274e-01 -6.81187868e-01 4.69958931e-01
-1.21648562e+00 -6.31369650e-01 -4.46221471e-01 3.74280155e-01
-4.20590997e-01 7.83556178e-02 -7.59182274e-01 -7.54445970e-01
-2.06125930e-01 -1.04133546e+00 3.81956905e-01 4.70115662e-01
-1.69696704e-01 -1.23727131e+00 -1.90185517e-01 -1.69434801e-01
5.91212630e-01 -4.68241334e-01 1.13024461e+00 -8.64161670e-01
-5.52025735e-01 1.85445979e-01 -8.41055661e-02 9.15822268e-01
1.55650511e-01 -5.28143207e-03 -1.25889182e+00 -1.52266830e-01
2.50578225e-01 -2.81628639e-01 1.41271675e+00 4.05071169e-01
1.92170048e+00 -5.10377645e-01 -5.93885958e-01 8.71447980e-01
9.83363032e-01 5.64986885e-01 7.76376188e-01 2.33865738e-01
4.48679715e-01 5.28664768e-01 -4.55811806e-02 2.48322263e-01
1.49552435e-01 1.29876554e-01 3.15939486e-01 -1.10767715e-01
-1.62100128e-03 -4.02742624e-01 5.29293776e-01 6.02812707e-01
2.92538628e-02 -3.16903263e-01 -7.34456360e-01 5.59982777e-01
-1.82344317e+00 -7.36384928e-01 1.82189330e-01 2.11813188e+00
7.07603633e-01 3.16576302e-01 -3.21613580e-01 2.20103696e-01
8.30147564e-01 -6.50853291e-02 -7.19599664e-01 -8.03386867e-01
-2.59159416e-01 4.11549628e-01 6.35332286e-01 1.97490692e-01
-6.89310551e-01 1.10005903e+00 7.20904779e+00 7.78377414e-01
-1.22212899e+00 3.21692489e-02 9.06514883e-01 3.42359722e-01
-4.78769779e-01 3.98092121e-02 -8.19047987e-01 2.48980746e-01
9.02159989e-01 -2.15360131e-02 4.12907422e-01 1.08743405e+00
-2.24823147e-01 3.17588568e-01 -1.42670214e+00 6.45904899e-01
-3.88533562e-01 -1.38902318e+00 4.55024600e-01 1.24146171e-01
7.81305730e-01 6.34176508e-02 5.22783458e-01 7.60922730e-01
3.49884212e-01 -1.32625902e+00 6.59205556e-01 6.69911623e-01
6.02773130e-01 -4.52892035e-01 6.87397361e-01 4.80747491e-01
-9.19175327e-01 -3.44899237e-01 -7.37074614e-01 -6.44786358e-01
-6.04172111e-01 4.42956656e-01 -4.62100327e-01 3.19726586e-01
8.34230661e-01 7.90839016e-01 -4.89657849e-01 8.17860365e-01
-2.65178531e-01 5.03904402e-01 -1.85011551e-02 -2.21174769e-02
2.75048882e-01 9.02323052e-02 -1.02282241e-01 9.85927761e-01
3.49778086e-01 -1.42405361e-01 2.88082706e-03 1.17618430e+00
-4.53533262e-01 -3.73604715e-01 -9.04612720e-01 -1.25864431e-01
4.99200195e-01 8.58693957e-01 -5.54879904e-01 -3.80808413e-01
-4.27094430e-01 9.90777791e-01 6.02577269e-01 7.99552381e-01
-7.13710487e-01 -4.93852109e-01 6.26809180e-01 1.12044113e-02
2.65852839e-01 -1.03957623e-01 -7.13835478e-01 -1.16792822e+00
-1.75628409e-01 -3.18816602e-01 1.97056666e-01 -5.65669179e-01
-1.11777461e+00 7.73992777e-01 -1.79835334e-01 -9.58191037e-01
1.78696692e-01 -9.90822196e-01 -8.30912352e-01 7.94924557e-01
-1.34635448e+00 -8.23889136e-01 4.15109433e-02 6.59002423e-01
3.82040203e-01 -3.55375379e-01 8.80738795e-01 -1.59347616e-02
-9.03541028e-01 8.24728608e-01 5.61357029e-02 5.24315648e-02
2.64909834e-01 -1.01893270e+00 5.07135987e-01 5.35527527e-01
-9.91550274e-03 1.25817490e+00 3.28615308e-01 -3.21473837e-01
-1.15077150e+00 -1.12465692e+00 7.84960091e-01 -8.83954242e-02
5.47447145e-01 -2.84712493e-01 -1.15751362e+00 1.18343973e+00
2.83086270e-01 7.41311014e-02 7.72161543e-01 3.23907256e-01
-4.45001185e-01 -3.63232523e-01 -9.68334258e-01 5.36285460e-01
1.42876387e+00 -4.91780341e-01 -7.96447039e-01 -3.96725722e-02
1.22334516e+00 1.02814630e-01 -8.77848029e-01 5.97657859e-01
5.41895270e-01 -1.00364888e+00 6.10799551e-01 -8.43420386e-01
3.25850338e-01 1.48363590e-01 -2.88156211e-01 -1.48812580e+00
-7.78597772e-01 -3.69159669e-01 -2.65784174e-01 8.73340249e-01
6.17968023e-01 -1.23718262e+00 4.56252515e-01 6.05644643e-01
-5.69023073e-01 -1.02600825e+00 -7.88017392e-01 -1.18760860e+00
5.26922405e-01 -3.37469906e-01 6.69431269e-01 8.44011366e-01
1.31322727e-01 2.39894569e-01 8.59934464e-03 -8.57031569e-02
-4.46906276e-02 8.38438720e-02 3.05287361e-01 -1.30472660e+00
-2.55267769e-01 -9.14348900e-01 -3.57055455e-01 -1.72470415e+00
3.24801832e-01 -9.19565916e-01 -5.33682346e-01 -1.49972951e+00
1.26041472e-01 -5.28564036e-01 -7.99613535e-01 8.64701390e-01
4.07540411e-01 -8.02288391e-03 -8.20888951e-02 3.07924688e-01
-5.79363883e-01 7.49765158e-01 1.42006540e+00 -1.61672726e-01
-2.27924198e-01 2.11403072e-01 -1.07340765e+00 7.79782295e-01
1.37671065e+00 -2.83295214e-01 -5.67382395e-01 -6.61086380e-01
3.83142591e-01 -3.03147405e-01 3.31779063e-01 -1.01215863e+00
1.18833721e-01 -1.03209242e-01 7.56993651e-01 -2.33057708e-01
7.62466490e-02 -8.52933943e-01 -1.81534573e-01 8.52990210e-01
-6.16021156e-01 1.89108953e-01 3.90379101e-01 2.98280090e-01
-1.66962057e-01 -3.37103754e-01 4.57303941e-01 -4.40320015e-01
-7.07925916e-01 2.53825456e-01 -5.46181321e-01 -1.57839820e-01
7.18307734e-01 -4.94637549e-01 -5.75442851e-01 -1.57724962e-01
-9.41278100e-01 1.72509983e-01 2.44556546e-01 5.85658610e-01
7.34834969e-01 -1.41720438e+00 -1.21168777e-01 3.61888289e-01
-9.41097885e-02 -8.18318129e-02 2.35020474e-01 6.35544658e-01
-1.51664555e-01 8.88649702e-01 -3.40379030e-01 -3.39136660e-01
-5.88829815e-01 6.08160079e-01 7.21035719e-01 9.78887752e-02
-5.03893197e-01 1.24011302e+00 1.00283730e+00 -5.50343990e-01
4.88739610e-01 -1.01331651e+00 -1.35095522e-01 -2.73624539e-01
2.41481170e-01 4.26142305e-01 2.45948024e-02 -3.33441165e-03
-3.09976429e-01 2.11597845e-01 -3.71204853e-01 3.29433829e-01
1.06457567e+00 7.87696168e-02 -2.92188823e-01 8.14181864e-01
1.19906628e+00 -5.65415382e-01 -1.16030276e+00 -3.23258638e-01
-1.26510173e-01 -6.99818581e-02 8.43268186e-02 -7.90104985e-01
-1.22101390e+00 1.18867016e+00 3.51058155e-01 5.77043712e-01
1.13003588e+00 1.50465176e-01 6.31398082e-01 8.02409291e-01
3.71692747e-01 -1.03643072e+00 5.20341992e-02 8.30316305e-01
9.27146077e-01 -7.12414682e-01 -4.79591131e-01 6.95766881e-02
-4.60401773e-01 1.34038651e+00 1.10220194e+00 -2.02412501e-01
7.19884515e-01 1.41737998e-01 -5.54891348e-01 5.87067381e-02
-1.12218428e+00 -1.58852097e-04 9.67370421e-02 4.79082584e-01
7.14431107e-01 2.24064231e-01 -6.34602085e-02 9.94327605e-01
-3.32707465e-01 -2.44699437e-02 3.69558901e-01 5.13622940e-01
-7.03168273e-01 -9.93964255e-01 -2.30429769e-02 3.58621776e-01
-1.89785749e-01 -2.99635828e-01 -4.84720141e-01 7.93545544e-01
5.50096571e-01 5.54631591e-01 2.42500558e-01 -8.21271956e-01
1.41871795e-01 2.82938600e-01 7.03277707e-01 -6.03192031e-01
-6.56607985e-01 -3.68697703e-01 -4.21466410e-01 -3.44739258e-01
-1.04920939e-02 -1.50874615e-01 -1.39368653e+00 -3.23206395e-01
-4.63204354e-01 -3.93045880e-02 1.91423595e-02 7.39655375e-01
4.57066923e-01 8.53332281e-01 5.91712654e-01 -3.52144837e-01
-6.42510474e-01 -1.12941086e+00 -6.20124102e-01 -7.25842118e-02
3.80524039e-01 -6.92008078e-01 -3.27225715e-01 -3.39347571e-01] | [8.827778816223145, 3.2615926265716553] |
d230aaa3-a13f-4188-9d9a-cb6c9bc1d1de | deep-variation-prior-joint-image-denoising | 2209.09214 | null | https://arxiv.org/abs/2209.09214v1 | https://arxiv.org/pdf/2209.09214v1.pdf | Deep Variation Prior: Joint Image Denoising and Noise Variance Estimation without Clean Data | With recent deep learning based approaches showing promising results in removing noise from images, the best denoising performance has been reported in a supervised learning setup that requires a large set of paired noisy images and ground truth for training. The strong data requirement can be mitigated by unsupervised learning techniques, however, accurate modelling of images or noise variance is still crucial for high-quality solutions. The learning problem is ill-posed for unknown noise distributions. This paper investigates the tasks of image denoising and noise variance estimation in a single, joint learning framework. To address the ill-posedness of the problem, we present deep variation prior (DVP), which states that the variation of a properly learnt denoiser with respect to the change of noise satisfies some smoothness properties, as a key criterion for good denoisers. Building upon DVP, an unsupervised deep learning framework, that simultaneously learns a denoiser and estimates noise variances, is developed. Our method does not require any clean training images or an external step of noise estimation, and instead, approximates the minimum mean squared error denoisers using only a set of noisy images. With the two underlying tasks being considered in a single framework, we allow them to be optimised for each other. The experimental results show a denoising quality comparable to that of supervised learning and accurate noise variance estimates. | ['Rihuan Ke'] | 2022-09-19 | null | null | null | null | ['noise-estimation'] | ['medical'] | [ 4.25033003e-01 -8.57319310e-02 4.32265729e-01 -3.72737199e-01
-1.25627446e+00 -2.39176661e-01 5.91730475e-01 -2.11302742e-01
-4.58362311e-01 7.25633562e-01 5.69660440e-02 1.97012946e-01
-5.15064418e-01 -6.50329590e-01 -7.65892565e-01 -1.41164446e+00
1.12211294e-01 1.45316571e-01 -1.44954711e-01 -9.72827822e-02
-9.60492715e-02 2.68175811e-01 -1.57013500e+00 -7.18372911e-02
8.71804118e-01 1.17601824e+00 4.63771105e-01 7.75487781e-01
7.40893977e-03 6.76508129e-01 -5.73632002e-01 -1.92533463e-01
5.79866290e-01 -6.53851151e-01 -3.90435398e-01 5.19141734e-01
4.78221744e-01 -1.11177854e-01 -3.02517653e-01 1.42490077e+00
7.57999778e-01 1.94705501e-01 6.53446972e-01 -7.71461189e-01
-3.93079370e-01 4.14728373e-01 -5.18830836e-01 1.13665469e-01
-1.92316622e-01 2.64060110e-01 6.68396533e-01 -5.75347006e-01
2.94422477e-01 1.00737774e+00 1.00909364e+00 4.61327970e-01
-1.63684154e+00 -2.56715417e-01 -2.11276561e-01 -5.66659719e-02
-1.30270779e+00 -7.30925739e-01 9.26425934e-01 -4.82202619e-01
2.80848950e-01 2.17266873e-01 3.34306300e-01 1.12633336e+00
1.97511286e-01 3.77397835e-01 1.34836185e+00 -4.85206991e-01
5.41920722e-01 -1.63857684e-01 -3.02542955e-01 1.34782299e-01
1.56246513e-01 1.35960966e-01 -3.44218999e-01 1.90210670e-01
9.81604755e-01 -2.96456754e-01 -5.29456437e-01 -3.97687525e-01
-8.73909533e-01 7.44633198e-01 4.48575467e-01 4.73721027e-01
-7.18437374e-01 2.98275262e-01 2.98023552e-01 5.18451631e-01
6.26525760e-01 2.55303800e-01 -3.90955061e-01 1.46745350e-02
-1.29637468e+00 1.19734250e-01 6.87351763e-01 4.80864465e-01
6.86274707e-01 3.19048077e-01 -1.02269582e-01 9.21684623e-01
3.17567348e-01 7.01898932e-01 3.33767593e-01 -1.27891386e+00
1.71495438e-01 -6.77650049e-02 1.28922135e-01 -1.19449568e+00
-1.99802909e-02 -7.21599519e-01 -1.53219211e+00 6.26638889e-01
4.35614794e-01 -1.83057770e-01 -1.18503702e+00 1.72390842e+00
1.97766989e-01 3.39212298e-01 5.81827238e-02 9.12423730e-01
5.12162030e-01 5.48255861e-01 -1.48442373e-01 -6.65720880e-01
8.67612839e-01 -5.69109261e-01 -1.21238053e+00 -3.78685683e-01
1.75990295e-02 -9.08490777e-01 5.62230051e-01 8.09710979e-01
-1.28744411e+00 -7.66806126e-01 -9.49488580e-01 8.35661143e-02
2.45416793e-03 -4.79132719e-02 3.80624644e-02 6.75841987e-01
-1.30975246e+00 9.16983485e-01 -1.02419949e+00 -1.70611702e-02
5.81313133e-01 3.54354739e-01 -4.50467736e-01 -1.52784243e-01
-1.01327384e+00 9.24558818e-01 2.19783232e-01 6.97757244e-01
-1.10034680e+00 -4.41819668e-01 -9.04603660e-01 -3.31548527e-02
2.74951041e-01 -8.36909711e-01 8.97327244e-01 -1.29658198e+00
-1.67952192e+00 6.55437946e-01 -6.72260821e-02 -5.18058002e-01
9.30921793e-01 -4.21037883e-01 -2.25282162e-01 1.56092703e-01
1.54234357e-02 1.10806182e-01 1.62477291e+00 -1.70765769e+00
-1.74476847e-01 -3.13264549e-01 -5.11746407e-01 -1.03495717e-02
-2.41242703e-02 -2.28016734e-01 -3.97848278e-01 -8.22308838e-01
4.09530640e-01 -4.60964978e-01 -6.21963322e-01 -6.70413971e-02
-1.43670157e-01 3.39869380e-01 6.77362502e-01 -9.54131722e-01
9.07755911e-01 -2.11560369e+00 4.36792612e-01 3.97975087e-01
2.54982412e-01 3.38263541e-01 -1.96211874e-01 1.30757540e-01
-2.85559624e-01 -2.84663457e-02 -7.77301788e-01 -6.19153678e-01
-1.82097271e-01 4.84481931e-01 -4.74767163e-02 8.34678769e-01
2.85586774e-01 6.47047222e-01 -8.46038997e-01 -1.68101028e-01
5.38012564e-01 8.95199776e-01 -2.05542535e-01 4.18726414e-01
2.78829802e-02 8.74016583e-01 -1.27821326e-01 1.95724908e-02
1.01818478e+00 6.54130057e-02 1.57777056e-01 -4.94030148e-01
8.74899700e-02 -2.47902200e-01 -1.67461777e+00 1.63569236e+00
-7.02906251e-01 6.32418871e-01 8.15066695e-01 -1.50518024e+00
9.37822044e-01 4.61764961e-01 4.52232242e-01 -5.92986465e-01
1.94203481e-01 4.44629252e-01 -3.27940285e-01 -7.48313069e-01
-2.25242272e-01 -5.64092219e-01 2.58518189e-01 -4.37100194e-02
3.80849510e-01 -3.65755588e-01 1.25427306e-01 -1.51179969e-01
1.11862636e+00 1.04582563e-01 2.75687754e-01 -3.53636324e-01
5.94475389e-01 -4.53444630e-01 6.65841460e-01 9.85916734e-01
-1.27971834e-02 1.08963561e+00 2.46399701e-01 -2.39229694e-01
-1.14301562e+00 -9.44695532e-01 -2.10584506e-01 5.20139933e-01
2.95839552e-03 -3.15425806e-02 -1.05822492e+00 -3.27510983e-01
-3.27901870e-01 3.96270931e-01 -5.45754313e-01 4.59767058e-02
-7.15385497e-01 -9.62246895e-01 2.37088114e-01 2.03100577e-01
7.10227013e-01 -6.18821204e-01 -3.27431351e-01 1.80751458e-01
-2.32023805e-01 -1.05622506e+00 -1.76981285e-01 5.76166332e-01
-9.34767962e-01 -9.96616423e-01 -8.89277458e-01 -7.65438676e-01
6.64732099e-01 1.05526805e-01 1.17979741e+00 7.81631023e-02
-1.38980895e-01 4.35449421e-01 -1.27352893e-01 -1.23374723e-01
-5.77626705e-01 -4.95705843e-01 8.14568922e-02 2.82893866e-01
-1.57460764e-01 -1.08075738e+00 -5.91984570e-01 2.81535506e-01
-1.18125892e+00 -4.38352078e-01 7.50382721e-01 1.16852009e+00
7.54232526e-01 7.81458259e-01 4.18329567e-01 -6.18498862e-01
5.84886491e-01 -3.49840492e-01 -6.09984994e-01 -6.41280562e-02
-4.04186547e-01 1.39144778e-01 7.30943382e-01 -2.52305031e-01
-1.17096865e+00 2.79374301e-01 -3.84047627e-01 -5.34542203e-01
-4.88536626e-01 4.73217696e-01 -5.58132112e-01 2.59833876e-02
8.74379218e-01 3.41564894e-01 3.09961826e-01 -7.85083294e-01
4.73609358e-01 5.08307397e-01 8.68283570e-01 -2.91539431e-01
1.07605457e+00 5.74063718e-01 1.17663808e-01 -1.07860053e+00
-8.51255834e-01 -5.72808444e-01 -8.17271531e-01 -2.30685130e-01
8.61060441e-01 -1.01339173e+00 -3.72664124e-01 8.02273631e-01
-1.04688549e+00 -3.81988764e-01 -4.48619753e-01 3.77770960e-01
-7.70802677e-01 6.42196953e-01 -3.88529450e-01 -1.03691113e+00
-1.84448868e-01 -1.19167078e+00 1.04642320e+00 -2.29866132e-02
6.60030842e-02 -1.38529468e+00 -3.71561735e-03 3.64160120e-01
6.15493655e-01 4.69029427e-01 4.33003604e-01 -2.98098266e-01
-3.80546629e-01 -1.72625065e-01 4.09036912e-02 1.18957818e+00
3.74221444e-01 -4.00424808e-01 -1.09357214e+00 -3.26442182e-01
1.01678920e+00 -1.36471957e-01 1.17764330e+00 8.61153662e-01
1.01067924e+00 -2.28823721e-01 1.55237392e-01 7.37942457e-01
1.78814244e+00 -3.09968919e-01 8.21433187e-01 4.19914126e-01
6.16940439e-01 4.38054085e-01 3.46549861e-02 2.61739075e-01
-2.42596269e-01 3.84209365e-01 6.64201558e-01 -2.31947079e-01
-4.02493149e-01 2.24163562e-01 1.79174483e-01 9.23368335e-01
-2.24077061e-01 -1.93030357e-01 -7.27261662e-01 6.63019538e-01
-1.79342341e+00 -1.04744685e+00 -4.36047137e-01 2.28046727e+00
1.07006919e+00 -4.02420713e-03 -2.35291138e-01 5.46784103e-01
7.46311784e-01 2.38052696e-01 -3.66239667e-01 2.38968641e-03
-3.65601033e-01 3.56728315e-01 5.10036707e-01 7.83020198e-01
-1.37553322e+00 4.02369469e-01 6.70226145e+00 9.99212742e-01
-9.92587209e-01 1.15219519e-01 6.17320359e-01 1.38154045e-01
-1.63583502e-01 -2.42480889e-01 -2.30009690e-01 5.18424273e-01
8.10003042e-01 3.30417037e-01 6.60453856e-01 5.18063784e-01
6.71898425e-01 -1.73822299e-01 -8.61452401e-01 1.20580292e+00
-6.03871793e-03 -9.39587533e-01 -2.21258685e-01 -4.89140488e-02
9.49291766e-01 -1.14231974e-01 9.52906683e-02 -8.47930089e-02
3.13694090e-01 -1.20048416e+00 4.68753546e-01 1.02378786e+00
3.16718310e-01 -6.45451665e-01 1.08630514e+00 5.30345261e-01
-6.72631741e-01 2.37317141e-02 -3.49836439e-01 -1.01942189e-01
3.45878720e-01 1.39089501e+00 -9.44714807e-03 7.32727826e-01
7.83168554e-01 7.54123986e-01 -1.99806675e-01 1.04569912e+00
-5.06359398e-01 7.15272009e-01 -3.71449977e-01 8.11761141e-01
-4.50451300e-02 -5.31817257e-01 6.80851281e-01 1.06380129e+00
3.88728172e-01 4.15626094e-02 4.52380329e-02 7.70318627e-01
-2.77699176e-02 -2.44976297e-01 -2.38402277e-01 4.12540704e-01
-7.93566406e-02 1.24869406e+00 -6.52298748e-01 -1.28351226e-01
-9.20161605e-02 1.02613127e+00 1.60715630e-04 6.17428482e-01
-4.62156445e-01 -9.62938145e-02 6.36986494e-01 5.55704869e-02
6.33106411e-01 -1.93832651e-01 -6.81373596e-01 -1.07859647e+00
2.49891862e-01 -1.06685078e+00 1.28047355e-03 -5.88463485e-01
-1.68465221e+00 5.44550955e-01 -2.47643918e-01 -1.15429759e+00
-1.99159369e-01 -4.57415044e-01 -7.34231770e-01 1.06824768e+00
-1.67234242e+00 -8.81203473e-01 -4.04811770e-01 4.75804150e-01
5.05199254e-01 3.29768248e-02 6.71256781e-01 3.93349558e-01
-5.03789246e-01 1.77449912e-01 6.87934339e-01 -2.28023604e-02
6.00125909e-01 -1.47832489e+00 6.35595545e-02 1.19475949e+00
-7.05205230e-03 4.79917079e-01 1.24284542e+00 -6.36583388e-01
-1.28482568e+00 -9.59461093e-01 5.33048272e-01 -3.17279041e-01
6.52228653e-01 -1.76420897e-01 -1.19493461e+00 3.61838788e-01
4.35421467e-01 3.77395391e-01 3.32570642e-01 -2.34716505e-01
1.61087718e-02 -3.19751352e-01 -1.17066193e+00 1.37254611e-01
7.88604140e-01 -3.51824939e-01 -5.86518168e-01 2.04949453e-01
1.71690807e-01 -4.17229295e-01 -8.99282753e-01 3.90158147e-01
6.30189106e-02 -1.11629748e+00 1.03396082e+00 -1.74100339e-01
4.76915479e-01 -4.51948076e-01 -1.59982666e-01 -1.61287141e+00
-2.84233481e-01 -7.69883811e-01 -2.05843076e-01 1.32403839e+00
4.66526933e-02 -2.61313885e-01 5.10347962e-01 5.08068919e-01
-1.24602839e-02 -4.17966634e-01 -9.42852795e-01 -8.44222188e-01
2.25420240e-02 -5.65068781e-01 6.08543381e-02 8.82817984e-01
-7.40100801e-01 8.64845291e-02 -6.64961219e-01 5.08728504e-01
1.23701584e+00 -3.91802281e-01 7.74906218e-01 -1.12556612e+00
-3.58395278e-01 -3.28459680e-01 -5.11888623e-01 -9.94776905e-01
1.69195216e-02 -4.54461008e-01 6.36087418e-01 -1.74608004e+00
-1.25188529e-01 -1.00183286e-01 -1.63678929e-01 1.12264246e-01
-3.37356001e-01 3.93471211e-01 -4.62689996e-02 2.58165121e-01
-3.12855244e-01 7.22086906e-01 9.74475384e-01 -2.25981891e-01
-6.79536760e-02 1.60208821e-01 -5.03804088e-01 1.05010557e+00
6.80155694e-01 -5.18317521e-01 -3.93779427e-01 -7.71346033e-01
1.87050581e-01 -1.22629412e-01 5.20259023e-01 -1.09686410e+00
3.24366629e-01 1.57288432e-01 4.96595353e-01 -1.13633908e-01
2.31329232e-01 -1.07459867e+00 3.36487859e-01 1.57118484e-01
-2.52567619e-01 -5.38999081e-01 -2.39056349e-02 7.83944011e-01
-5.54406881e-01 -4.88781124e-01 1.10360050e+00 -2.84850568e-01
-5.66027403e-01 9.06029996e-03 -4.13624585e-01 -5.17609380e-02
6.02194011e-01 -2.23478079e-01 3.72659534e-01 -5.83769798e-01
-1.18021095e+00 -4.78819432e-03 2.73249626e-01 -5.57441860e-02
5.08705378e-01 -1.18269455e+00 -9.82522130e-01 1.54912829e-01
-4.64770228e-01 3.58988822e-01 4.78469014e-01 8.80300283e-01
-3.75776798e-01 -3.72018874e-01 7.53645077e-02 -7.59515405e-01
-9.68986213e-01 5.18950284e-01 6.30558610e-01 -2.51507372e-01
-6.69293761e-01 8.17659557e-01 -5.70958294e-02 -2.56902695e-01
2.96904802e-01 -4.35322709e-02 -1.09927356e-01 2.96103340e-02
7.02289343e-01 4.57168251e-01 3.28942716e-01 -7.05035269e-01
6.09032772e-02 7.13862717e-01 3.74520600e-01 -8.92727375e-02
1.73730886e+00 -5.56328535e-01 -3.37012470e-01 2.89881229e-01
1.29404378e+00 7.22501951e-04 -1.60704863e+00 -3.38840157e-01
5.46059199e-02 -5.19552052e-01 7.32831955e-01 -6.29647553e-01
-1.50514889e+00 6.67037368e-01 8.24421942e-01 3.60104382e-01
1.60800958e+00 -2.65318602e-01 6.65838003e-01 5.41014373e-02
-3.17678712e-02 -1.16122043e+00 -1.52026268e-03 2.90869981e-01
9.21471179e-01 -1.48965502e+00 9.35899988e-02 -3.79254669e-01
-2.14291647e-01 1.03525341e+00 6.75278604e-02 -4.04146731e-01
7.68619776e-01 5.05173683e-01 4.09648389e-01 -1.30499557e-01
-1.91667929e-01 -3.57676238e-01 3.33785653e-01 7.42792666e-01
1.85795948e-01 -3.64614010e-01 -1.93224266e-01 4.08033609e-01
7.78345689e-02 3.66112441e-02 3.14433455e-01 7.34813631e-01
-4.75117147e-01 -9.86676216e-01 -7.23463714e-01 2.23957002e-01
-5.58764040e-01 -5.57468869e-02 1.04685359e-01 3.92550111e-01
3.67255151e-01 1.17678165e+00 -1.34116411e-01 4.36864644e-02
4.28472817e-01 -1.56684339e-01 3.49873990e-01 -2.32909292e-01
-4.23826754e-01 5.15202403e-01 -1.97569922e-01 -4.40807462e-01
-8.74567568e-01 -4.18099880e-01 -6.85648859e-01 -7.27861226e-02
-3.94378752e-01 1.11681931e-01 7.29220569e-01 1.23456848e+00
-6.48026615e-02 6.72366619e-01 7.32824564e-01 -1.19245303e+00
-7.46827185e-01 -8.91718566e-01 -7.05776572e-01 6.85207427e-01
6.87890112e-01 -3.60612690e-01 -8.74287426e-01 4.23537135e-01] | [11.5934476852417, -2.3936054706573486] |
b1529c6b-a4b5-47dc-b821-676b7d62f7b1 | dual-path-learning-for-domain-adaptation-of | 2108.06337 | null | https://arxiv.org/abs/2108.06337v1 | https://arxiv.org/pdf/2108.06337v1.pdf | Dual Path Learning for Domain Adaptation of Semantic Segmentation | Domain adaptation for semantic segmentation enables to alleviate the need for large-scale pixel-wise annotations. Recently, self-supervised learning (SSL) with a combination of image-to-image translation shows great effectiveness in adaptive segmentation. The most common practice is to perform SSL along with image translation to well align a single domain (the source or target). However, in this single-domain paradigm, unavoidable visual inconsistency raised by image translation may affect subsequent learning. In this paper, based on the observation that domain adaptation frameworks performed in the source and target domain are almost complementary in terms of image translation and SSL, we propose a novel dual path learning (DPL) framework to alleviate visual inconsistency. Concretely, DPL contains two complementary and interactive single-domain adaptation pipelines aligned in source and target domain respectively. The inference of DPL is extremely simple, only one segmentation model in the target domain is employed. Novel technologies such as dual path image translation and dual path adaptive segmentation are proposed to make two paths promote each other in an interactive manner. Experiments on GTA5$\rightarrow$Cityscapes and SYNTHIA$\rightarrow$Cityscapes scenarios demonstrate the superiority of our DPL model over the state-of-the-art methods. The code and models are available at: \url{https://github.com/royee182/DPL} | ['Wenqiang Zhang', 'Fang Wen', 'Dong Chen', 'Jianmin Bao', 'Fangyun Wei', 'Yiting Cheng'] | 2021-08-13 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Cheng_Dual_Path_Learning_for_Domain_Adaptation_of_Semantic_Segmentation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Cheng_Dual_Path_Learning_for_Domain_Adaptation_of_Semantic_Segmentation_ICCV_2021_paper.pdf | iccv-2021-1 | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 4.01825488e-01 2.66356796e-01 -2.27804974e-01 -3.90186250e-01
-9.35312927e-01 -5.93940794e-01 3.88403594e-01 -1.54646561e-01
-3.23999554e-01 6.19628668e-01 -2.34226152e-01 -3.14757615e-01
2.27549180e-01 -7.45868206e-01 -9.16116059e-01 -5.08237660e-01
5.17317533e-01 4.72130775e-01 8.00073504e-01 -2.74459362e-01
3.82832475e-02 2.43115425e-01 -1.17849946e+00 3.40236336e-01
1.24054039e+00 6.82396173e-01 6.11668110e-01 5.08145094e-01
-5.36006391e-01 2.21213937e-01 -2.68938005e-01 -4.62408751e-01
4.87872958e-01 -6.31304741e-01 -8.76662374e-01 3.06313515e-01
5.00427663e-01 -2.00336710e-01 2.74442863e-02 1.19940209e+00
5.79059541e-01 -1.02960423e-01 4.66219485e-01 -1.46443808e+00
-4.58076537e-01 3.75795811e-01 -9.77540731e-01 9.41214338e-02
2.55282491e-01 4.05967683e-01 7.45565772e-01 -8.97000015e-01
8.44150007e-01 1.08480775e+00 5.78214109e-01 4.67715114e-01
-1.20673525e+00 -8.73386621e-01 4.67970371e-01 2.79557705e-01
-1.31652772e+00 -2.48821735e-01 9.30971861e-01 -3.50391597e-01
5.68552375e-01 -8.30355063e-02 6.88480139e-01 1.17071962e+00
-2.56162621e-02 9.40170050e-01 1.25042474e+00 -6.50101244e-01
7.14027584e-02 3.35382551e-01 -2.63303250e-01 6.66509390e-01
4.26783897e-02 -4.95725796e-02 -5.16017318e-01 3.49145114e-01
9.47052717e-01 -2.52229095e-01 -8.73801261e-02 -5.05013704e-01
-1.19254148e+00 4.74458575e-01 2.92369545e-01 2.83788413e-01
-1.00345269e-01 -3.63058262e-02 3.73676628e-01 2.00652033e-01
5.50796390e-01 4.14941981e-02 -7.15537608e-01 9.70639437e-02
-9.64189947e-01 7.38059580e-02 3.67922902e-01 1.25077784e+00
8.90741765e-01 1.56856701e-02 3.99177335e-03 9.10946548e-01
3.20615202e-01 6.42611980e-01 3.62327307e-01 -9.74702716e-01
5.86923361e-01 7.17738867e-01 2.17131879e-02 -8.30103755e-01
-3.83365154e-01 -4.23205167e-01 -6.68932855e-01 3.10870051e-01
6.88465416e-01 -7.30457008e-02 -1.04942656e+00 1.64871657e+00
6.84951365e-01 2.95333266e-01 2.72358954e-03 7.56716549e-01
7.35061526e-01 6.15226626e-01 2.85562038e-01 -2.72565428e-03
1.31136549e+00 -1.27384806e+00 -5.00477552e-01 -5.17086565e-01
5.37859201e-01 -9.51470256e-01 1.59173918e+00 2.15664446e-01
-1.16797400e+00 -7.27918983e-01 -8.43214989e-01 -1.72693372e-01
-4.76772249e-01 1.52820706e-01 8.76257792e-02 4.12783295e-01
-9.53665316e-01 2.24568918e-01 -7.19796658e-01 -5.44319153e-01
4.96274114e-01 1.81495845e-01 -2.79735506e-01 -1.15235418e-01
-1.20932984e+00 7.02590942e-01 5.73058844e-01 -1.57274649e-01
-6.80874228e-01 -6.05391383e-01 -8.02776933e-01 -2.64427602e-01
5.55991769e-01 -7.00202346e-01 1.33791041e+00 -1.40342295e+00
-1.44423676e+00 1.17320395e+00 -6.91640377e-02 -2.95895875e-01
9.11134481e-01 -2.13708460e-01 -4.19616699e-01 3.35104316e-01
4.95872170e-01 1.10734832e+00 9.63886023e-01 -1.35387647e+00
-9.45045233e-01 -3.62859368e-01 -8.16391557e-02 4.92490441e-01
-1.78344563e-01 -2.00802326e-01 -1.11494982e+00 -6.57541692e-01
1.95582122e-01 -9.25540924e-01 -2.32265443e-01 2.23754242e-01
-3.19980413e-01 -8.10704008e-02 9.82781291e-01 -7.81529129e-01
1.09554231e+00 -2.06955671e+00 -1.00966796e-01 4.79749590e-02
-1.22973070e-01 3.90127063e-01 -2.99364746e-01 6.79445565e-02
-5.29059060e-02 -1.81917865e-02 -6.14186823e-01 -4.05988157e-01
-7.78617337e-02 2.87939280e-01 -6.41014874e-02 2.90065855e-01
1.04385018e-01 1.00242329e+00 -9.45612371e-01 -1.03274119e+00
3.94740909e-01 2.61696339e-01 -4.12206173e-01 1.66392431e-01
-5.22561312e-01 9.08905864e-01 -4.49578822e-01 8.08431506e-01
9.28421259e-01 -2.00092256e-01 1.28835574e-01 -2.69091487e-01
-1.48345485e-01 -7.68494904e-02 -1.32308936e+00 2.13327575e+00
-4.03948367e-01 3.51070642e-01 2.75314692e-02 -9.47817802e-01
8.44816864e-01 1.67810708e-01 4.50030744e-01 -1.12457538e+00
-1.48610890e-01 3.86198878e-01 -3.92352730e-01 -5.94071269e-01
2.75971144e-01 7.44474083e-02 -1.45227507e-01 2.36017540e-01
-3.04128975e-03 -4.09510195e-01 2.85073429e-01 1.41234100e-01
5.60716629e-01 8.40916753e-01 3.96231532e-01 -8.42020214e-02
6.78677261e-01 5.19902468e-01 8.41320336e-01 5.06095469e-01
-5.30658126e-01 8.73242617e-01 3.84914696e-01 -2.20159531e-01
-1.17527664e+00 -1.21666992e+00 8.10982659e-02 1.12903225e+00
6.62482142e-01 -1.87797397e-02 -1.01745856e+00 -9.03002620e-01
-4.09245521e-01 7.52007723e-01 -3.02169740e-01 1.21402413e-01
-6.09842479e-01 -4.79567856e-01 5.27302504e-01 5.78804314e-01
1.13254571e+00 -1.02643490e+00 -5.63749552e-01 6.27194792e-02
-5.48446119e-01 -1.42249262e+00 -6.77046657e-01 -1.84274375e-01
-9.22534943e-01 -9.71755981e-01 -9.26383257e-01 -9.94918823e-01
7.55032420e-01 3.16742688e-01 1.23762560e+00 -1.97450504e-01
-2.83457376e-02 4.44288224e-01 -3.58242959e-01 -2.10892305e-01
-4.40663993e-01 1.87945902e-01 -3.14316392e-01 -9.28013623e-02
2.42107734e-01 -5.48447967e-01 -9.57477152e-01 6.26791716e-01
-1.00534165e+00 5.81552088e-01 6.97531700e-01 5.99478722e-01
1.05076456e+00 -2.52986085e-02 4.29227799e-01 -1.17861462e+00
1.95782259e-01 -3.19147438e-01 -6.26367390e-01 4.66350228e-01
-6.80355668e-01 -2.31310472e-01 4.53697830e-01 -3.28072101e-01
-1.53305197e+00 4.66580868e-01 -1.43099040e-01 -4.12561595e-01
-5.67073822e-01 1.09661929e-01 -4.54551071e-01 2.12945268e-01
6.98633492e-01 3.74320924e-01 -1.61485791e-01 -3.44102830e-01
6.66994512e-01 5.58432341e-01 7.54831970e-01 -4.60711539e-01
7.91789472e-01 6.11037493e-01 -3.39245319e-01 -5.30426860e-01
-6.01025343e-01 -4.87524778e-01 -1.06173718e+00 -3.79863471e-01
1.04276991e+00 -9.89337504e-01 -8.03578496e-02 6.30640626e-01
-1.07320678e+00 -7.89478540e-01 -2.82731175e-01 1.54249266e-01
-6.25575483e-01 4.22613263e-01 -2.10979044e-01 -3.65592480e-01
-3.31779331e-01 -1.19622195e+00 1.17896676e+00 4.65531498e-01
-1.64179936e-01 -1.07574415e+00 -1.56582594e-01 5.97574413e-01
1.22252636e-01 9.93929431e-02 7.10880637e-01 -4.28581119e-01
-5.93540072e-01 1.59970075e-01 -5.45584917e-01 3.17523599e-01
9.35476869e-02 -6.14568442e-02 -7.93176115e-01 -9.30393264e-02
-2.71376193e-01 5.35055585e-02 4.69978005e-01 4.44146842e-01
1.03362858e+00 7.64383450e-02 -4.19595122e-01 5.88907957e-01
1.60058999e+00 2.58862793e-01 6.96428597e-01 6.22454762e-01
7.87568510e-01 6.22135580e-01 1.01860213e+00 2.54232436e-01
5.86036384e-01 8.31349134e-01 2.93029904e-01 -5.81408262e-01
-6.37461424e-01 -4.21923220e-01 3.67516398e-01 3.38581920e-01
2.24349633e-01 -1.42474741e-01 -1.09624720e+00 6.62055492e-01
-1.83978796e+00 -5.85637927e-01 -4.21280980e-01 2.13455892e+00
8.51403236e-01 1.76290572e-01 1.80241406e-01 -1.65712059e-01
7.88080692e-01 -7.19228089e-02 -6.73981428e-01 -2.31858283e-01
-3.40685636e-01 -8.98590405e-03 5.98175585e-01 5.13780475e-01
-1.10485339e+00 1.28941298e+00 5.22667122e+00 1.12344217e+00
-1.15079212e+00 4.58620816e-01 8.45045805e-01 1.67718470e-01
-2.96933293e-01 5.62731214e-02 -8.10046256e-01 5.50948262e-01
4.01222706e-01 6.44584894e-02 2.90092826e-02 8.14225733e-01
5.13793886e-01 -4.12094742e-01 -6.91969633e-01 9.03156519e-01
-1.83287133e-02 -1.13074446e+00 -4.99508530e-02 -3.07406843e-01
9.11414862e-01 6.57863682e-03 1.39405578e-01 1.55456796e-01
2.77303696e-01 -4.55675960e-01 7.96769500e-01 1.32056192e-01
1.12052023e+00 -4.09675658e-01 3.73177618e-01 2.91765869e-01
-1.35088897e+00 7.22326115e-02 -1.06956474e-02 3.81092012e-01
3.16468507e-01 5.10788679e-01 -8.92761230e-01 6.49948418e-01
1.03571260e+00 6.23618245e-01 -6.81445479e-01 7.00152159e-01
-5.83715439e-01 4.57956851e-01 -1.26304388e-01 6.22764349e-01
1.81292713e-01 -3.96819919e-01 6.01387978e-01 1.35499454e+00
2.85790205e-01 2.09881756e-02 2.54172683e-01 8.90718460e-01
1.35916159e-01 2.04597473e-01 -5.37289560e-01 3.31644893e-01
4.27494705e-01 1.17821860e+00 -1.17273533e+00 -5.52532375e-01
-5.90815663e-01 1.45418537e+00 -2.16617528e-03 5.36164045e-01
-1.11004722e+00 -1.49222434e-01 3.68142396e-01 4.06421512e-01
2.29113951e-01 -2.48364240e-01 -7.27623224e-01 -1.03847539e+00
6.27198592e-02 -7.68771350e-01 5.73036730e-01 -1.02646339e+00
-1.03953218e+00 5.43412745e-01 1.71399042e-01 -1.57151306e+00
4.24681045e-02 -2.83322752e-01 -4.54601079e-01 7.69501567e-01
-1.74860525e+00 -1.48562300e+00 -5.46179235e-01 9.45251703e-01
1.14636683e+00 8.39974284e-02 4.72764552e-01 4.58735138e-01
-7.07289755e-01 5.77412546e-01 1.99427083e-02 7.93471746e-03
9.64528382e-01 -1.21272933e+00 4.22214180e-01 1.23504364e+00
3.28307301e-02 3.57244872e-02 5.50714314e-01 -8.19316745e-01
-7.69759595e-01 -1.29299772e+00 6.15414917e-01 -1.88452780e-01
3.94931048e-01 -1.64955005e-01 -9.17880774e-01 6.04873419e-01
3.35289836e-01 1.21769961e-02 3.33083779e-01 -5.58265507e-01
-1.43181160e-01 -2.61659950e-01 -1.23656523e+00 7.92558670e-01
1.22447789e+00 -2.67314285e-01 -2.69639760e-01 3.56396258e-01
7.86054373e-01 -6.92656338e-01 -6.57226145e-01 3.57317001e-01
3.45184684e-01 -1.03666270e+00 1.07906783e+00 1.02096107e-02
6.34145021e-01 -6.39879465e-01 -2.40229778e-02 -1.03878486e+00
-2.54863407e-02 -4.35193181e-01 2.83768058e-01 1.41413128e+00
5.23276210e-01 -6.75890803e-01 8.91989589e-01 5.54114163e-01
-2.77522445e-01 -4.81739312e-01 -7.78560340e-01 -8.01022232e-01
1.09357476e-01 -5.06645858e-01 3.55169445e-01 8.81159544e-01
-3.46512854e-01 2.03890935e-01 -1.00775272e-01 3.69733214e-01
6.27625465e-01 2.33669817e-01 9.78785396e-01 -8.28152180e-01
-2.38570154e-01 -4.95796561e-01 6.72429502e-02 -1.09799469e+00
-1.35256760e-02 -8.05725873e-01 4.40548807e-02 -1.74376404e+00
1.27928387e-02 -6.93756819e-01 -1.67818129e-01 6.09731138e-01
-1.95260480e-01 5.10658801e-01 1.93703860e-01 2.63015658e-01
-5.97400963e-01 2.23992288e-01 1.50750136e+00 -1.22005329e-01
-4.65168804e-01 -4.86023091e-02 -5.63998461e-01 8.00743818e-01
1.01284158e+00 -3.84059012e-01 -6.35048747e-01 -6.33173108e-01
-8.41998160e-02 -1.27712667e-01 3.50147545e-01 -9.26086664e-01
2.55273968e-01 -2.04423070e-01 3.08346510e-01 -5.48734128e-01
1.01779252e-01 -8.62421334e-01 2.11159885e-01 3.79520088e-01
-1.15876198e-01 -7.12469816e-02 4.35930133e-01 5.22204578e-01
-3.13004643e-01 -1.19842134e-01 1.08308458e+00 -3.88655394e-01
-1.38482356e+00 2.87273437e-01 5.88920377e-02 8.64611343e-02
1.28708959e+00 -6.57377303e-01 -5.79852201e-02 -1.74425870e-01
-8.62597108e-01 3.24491560e-01 6.52505338e-01 4.27187979e-01
5.48616946e-01 -1.02624893e+00 -6.44960940e-01 1.66986302e-01
2.87013233e-01 3.52436066e-01 6.27299607e-01 1.10020936e+00
-6.54458225e-01 1.33415326e-01 -3.91791612e-01 -8.93601835e-01
-1.33203638e+00 3.58177155e-01 3.89948726e-01 -3.06723833e-01
-5.73884308e-01 8.82810593e-01 5.32432020e-01 -5.81298530e-01
-3.18432897e-02 -6.44809902e-02 6.86894655e-02 -3.56012173e-02
1.55405238e-01 2.66469091e-01 4.88718562e-02 -6.48399055e-01
-2.80417025e-01 8.78954530e-01 -3.38696055e-02 -3.13343734e-01
1.01672697e+00 -7.74146795e-01 8.97771269e-02 1.63875103e-01
8.88726473e-01 -2.55952895e-01 -1.68305647e+00 -4.61044669e-01
-1.59174860e-01 -5.36958694e-01 -2.65303105e-01 -1.01692498e+00
-1.15227187e+00 9.49648976e-01 9.01768386e-01 -1.98401868e-01
1.64372933e+00 1.03754826e-01 9.16932940e-01 -3.74417633e-01
2.96671450e-01 -1.41882682e+00 1.79141730e-01 2.15128005e-01
7.00381815e-01 -1.49588060e+00 -1.23242341e-01 -6.69488430e-01
-1.03063154e+00 8.98074269e-01 9.84224439e-01 1.78301707e-01
2.44997203e-01 1.43329978e-01 3.07953417e-01 9.09235775e-02
-2.36022294e-01 -3.35976988e-01 1.58648327e-01 8.49579215e-01
2.61221290e-01 -3.59834023e-02 -3.40636134e-01 2.84975737e-01
1.42923310e-01 8.30436125e-02 3.31825793e-01 7.30769813e-01
-2.15159684e-01 -1.35962546e+00 -5.64826906e-01 -5.60340211e-02
-1.99498549e-01 -3.16474028e-02 -1.44174159e-01 9.42237854e-01
5.14676809e-01 8.30395103e-01 6.54623136e-02 -1.15178928e-01
3.84362429e-01 2.43392531e-02 3.11791748e-01 -5.04139066e-01
-2.89845288e-01 2.97689140e-01 -1.68582097e-01 -6.33443952e-01
-5.26361048e-01 -6.44490421e-01 -1.51457787e+00 1.57252159e-02
-1.94085836e-02 -3.61512661e-01 5.98814428e-01 8.33183289e-01
4.71797049e-01 4.73549515e-01 3.99673313e-01 -7.20904648e-01
1.17083840e-01 -6.43141091e-01 -3.90919477e-01 5.21744788e-01
-9.57184136e-02 -5.37119150e-01 8.03992599e-02 5.03534496e-01] | [9.653471946716309, 1.3098806142807007] |
d5192b83-7c4f-4e82-898f-c1337da4b22f | fastlr-non-autoregressive-lipreading-model | 2008.02516 | null | https://arxiv.org/abs/2008.02516v4 | https://arxiv.org/pdf/2008.02516v4.pdf | FastLR: Non-Autoregressive Lipreading Model with Integrate-and-Fire | Lipreading is an impressive technique and there has been a definite improvement of accuracy in recent years. However, existing methods for lipreading mainly build on autoregressive (AR) model, which generate target tokens one by one and suffer from high inference latency. To breakthrough this constraint, we propose FastLR, a non-autoregressive (NAR) lipreading model which generates all target tokens simultaneously. NAR lipreading is a challenging task that has many difficulties: 1) the discrepancy of sequence lengths between source and target makes it difficult to estimate the length of the output sequence; 2) the conditionally independent behavior of NAR generation lacks the correlation across time which leads to a poor approximation of target distribution; 3) the feature representation ability of encoder can be weak due to lack of effective alignment mechanism; and 4) the removal of AR language model exacerbates the inherent ambiguity problem of lipreading. Thus, in this paper, we introduce three methods to reduce the gap between FastLR and AR model: 1) to address challenges 1 and 2, we leverage integrate-and-fire (I\&F) module to model the correspondence between source video frames and output text sequence. 2) To tackle challenge 3, we add an auxiliary connectionist temporal classification (CTC) decoder to the top of the encoder and optimize it with extra CTC loss. We also add an auxiliary autoregressive decoder to help the feature extraction of encoder. 3) To overcome challenge 4, we propose a novel Noisy Parallel Decoding (NPD) for I\&F and bring Byte-Pair Encoding (BPE) into lipreading. Our experiments exhibit that FastLR achieves the speedup up to 10.97$\times$ comparing with state-of-the-art lipreading model with slight WER absolute increase of 1.5\% and 5.5\% on GRID and LRS2 lipreading datasets respectively, which demonstrates the effectiveness of our proposed method. | ['Nicholas Jing Yuan', 'Yi Ren', 'Jinglin Liu', 'Chen Zhang', 'Zhou Zhao', 'Baoxing Huai'] | 2020-08-06 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [ 4.41989273e-01 -9.52187553e-02 -2.86031276e-01 -6.87796846e-02
-1.17516255e+00 -3.00326288e-01 4.02205080e-01 -3.77666146e-01
-2.27866188e-01 6.32066607e-01 5.84335744e-01 -3.46875429e-01
3.39733958e-01 -2.17875287e-01 -7.31248498e-01 -6.17709696e-01
4.69145387e-01 -1.24225751e-01 2.79122859e-01 -6.65276796e-02
2.44691283e-01 -4.88176756e-02 -1.48738086e+00 2.71926105e-01
9.39039588e-01 1.08425379e+00 6.73668563e-01 9.76715863e-01
-3.15544665e-01 5.91645598e-01 -4.84749168e-01 -2.51272380e-01
1.62262619e-01 -5.64824224e-01 -4.57867414e-01 -1.74752623e-01
-2.44829822e-02 -5.42040586e-01 -5.69021463e-01 9.95414972e-01
1.03653145e+00 -1.15673430e-02 7.82531023e-01 -1.28880441e+00
-5.78116477e-01 5.20250797e-01 -8.17279935e-01 5.29016228e-03
4.13924932e-01 3.98787677e-01 6.78995848e-01 -9.56703126e-01
1.35295376e-01 1.35129631e+00 6.32717490e-01 7.52377987e-01
-7.20952749e-01 -7.79720843e-01 -4.68334071e-02 4.27974582e-01
-1.53525758e+00 -1.27894771e+00 8.03145289e-01 -1.04585864e-01
9.64438081e-01 1.77200451e-01 2.14483634e-01 1.18214202e+00
3.89957428e-02 9.25411105e-01 9.89762366e-01 -4.76528615e-01
-9.06233117e-03 2.16877949e-03 -1.62482873e-01 5.01197398e-01
-4.40677494e-01 3.17598321e-02 -8.15866292e-01 2.25920230e-01
4.20365334e-01 -3.09892952e-01 -4.15139139e-01 3.45392048e-01
-1.01963997e+00 5.05856454e-01 -1.71765313e-01 -9.93124768e-02
-2.44705036e-01 2.56213814e-01 6.12222254e-01 7.56655857e-02
1.38237998e-01 -4.09580380e-01 -3.64897698e-01 -6.88887894e-01
-9.67303038e-01 -1.81511119e-01 4.73883450e-01 1.07193494e+00
4.05809373e-01 3.85569662e-01 -2.20108092e-01 1.40082324e+00
5.86220026e-01 7.98240244e-01 8.51310194e-01 -7.27177441e-01
9.62559700e-01 4.93725054e-02 -1.86113685e-01 -6.10220551e-01
1.79453418e-02 -4.45282348e-02 -9.68191981e-01 -3.16834338e-02
2.35980675e-01 -9.91276205e-02 -1.02471685e+00 1.84173274e+00
-6.51407242e-03 4.21169072e-01 3.74742806e-01 7.26395428e-01
8.16196382e-01 1.01730525e+00 -1.56438984e-02 -5.72618008e-01
1.35824347e+00 -1.09667766e+00 -9.83723640e-01 -1.09335706e-01
3.35673928e-01 -1.27265286e+00 1.27908885e+00 1.76155910e-01
-1.30360818e+00 -7.95596123e-01 -9.99542654e-01 -3.33591193e-01
-7.18019456e-02 4.78264958e-01 8.22432637e-02 6.57674491e-01
-9.67237830e-01 1.04926378e-01 -5.73031425e-01 -1.27778545e-01
1.28101677e-01 4.38778311e-01 -1.89527690e-01 -1.59053549e-01
-1.21106839e+00 5.71648359e-01 9.06876028e-02 3.41100633e-01
-6.71400785e-01 -4.59246129e-01 -9.20594692e-01 3.67453955e-02
3.47203255e-01 -3.91539663e-01 1.19038677e+00 -8.66604924e-01
-2.11867046e+00 3.54015350e-01 -8.87310684e-01 -2.95194924e-01
5.30520082e-01 -1.47647724e-01 -5.03293753e-01 2.58368012e-02
-1.69090014e-02 7.94029534e-01 1.04157174e+00 -1.03951299e+00
-5.76604843e-01 -1.70426201e-02 -2.52647012e-01 4.82124865e-01
-1.05766393e-01 1.54874444e-01 -6.84433281e-01 -7.38666236e-01
8.89178813e-02 -8.62525702e-01 3.75152349e-01 -2.37260520e-01
-4.98513520e-01 -3.29891503e-01 9.66467381e-01 -9.97384667e-01
1.41073895e+00 -2.25847483e+00 -1.26126483e-01 -2.49585614e-01
-2.01192856e-01 5.77714086e-01 -2.18966812e-01 3.91633183e-01
6.36387616e-03 2.71954477e-01 -9.65599194e-02 -6.26175642e-01
3.19330208e-02 1.45384043e-01 -5.08113444e-01 2.97203273e-01
7.02122152e-02 8.52160871e-01 -5.52654922e-01 -7.53893435e-01
2.52259135e-01 6.28863096e-01 -4.66926932e-01 3.49486589e-01
-4.62827533e-02 1.90842211e-01 -9.00910944e-02 6.77726090e-01
8.81845236e-01 1.84096038e-01 -8.62040743e-02 -3.16854626e-01
-1.64982378e-01 6.27427757e-01 -1.24798095e+00 1.63895106e+00
-6.96478069e-01 7.29468107e-01 -3.96483615e-02 -7.61440635e-01
8.09003532e-01 7.03488529e-01 1.85044646e-01 -6.10451162e-01
2.79684663e-02 2.10095912e-01 3.30459583e-03 -8.40743840e-01
4.26120728e-01 -3.98422666e-02 2.38377526e-01 4.01341319e-01
-1.12662658e-01 -6.77022757e-03 -1.92071348e-01 -9.13549140e-02
8.07429194e-01 2.76072055e-01 1.21609069e-01 2.10070506e-01
9.27920401e-01 -7.45541036e-01 7.62651742e-01 5.52383542e-01
-3.29121977e-01 8.09209824e-01 4.74175960e-01 1.57653213e-01
-9.97795820e-01 -1.08349288e+00 1.87433019e-01 1.02495110e+00
8.27560723e-02 -4.11311507e-01 -8.31584156e-01 -3.08322817e-01
-4.41135675e-01 4.29809958e-01 -1.32166803e-01 1.97642315e-02
-6.92282081e-01 -4.88082051e-01 1.06824291e+00 5.32959938e-01
9.60121155e-01 -9.68043745e-01 -1.49117172e-01 1.86036646e-01
-8.63899171e-01 -1.28666866e+00 -1.02085495e+00 -2.28591233e-01
-4.67505276e-01 -6.27956033e-01 -8.74556959e-01 -8.42595339e-01
3.13970387e-01 3.81623685e-01 5.30340910e-01 -1.46016432e-02
2.79139951e-02 -6.10093400e-02 -3.41104835e-01 -4.00606543e-01
-6.06315076e-01 1.08367264e-01 2.77206123e-01 9.89814475e-02
2.78494120e-01 -4.43971038e-01 -5.96626163e-01 3.58541161e-01
-6.88619554e-01 2.44360238e-01 6.68587208e-01 9.80775833e-01
3.98843914e-01 5.80298491e-02 8.01213086e-01 -2.40581547e-04
7.12812543e-01 -3.82165253e-01 -4.24707443e-01 2.40731969e-01
-5.37229180e-01 3.83889116e-02 7.12678432e-01 -6.30962908e-01
-1.09880877e+00 2.91556977e-02 -5.94769120e-01 -5.62873662e-01
3.65591981e-02 8.60306248e-02 -3.93157095e-01 2.90271640e-01
7.70160854e-02 8.00271749e-01 2.74000943e-01 -6.11291468e-01
1.21544749e-01 1.44195342e+00 5.84093869e-01 -3.17104101e-01
5.18952191e-01 6.96141124e-02 -2.22204804e-01 -1.09320593e+00
-1.63993850e-01 -4.53562558e-01 -2.27736205e-01 -9.31370631e-02
8.02013278e-01 -1.06420541e+00 -1.12394965e+00 7.76602030e-01
-1.43490815e+00 -4.17731225e-01 1.27552658e-01 6.61307573e-01
-7.44614303e-01 7.51856208e-01 -6.59210026e-01 -1.23096848e+00
-5.32920182e-01 -1.55750632e+00 1.13523686e+00 2.42045969e-01
6.03698045e-02 -4.51103956e-01 -3.91658127e-01 7.02607632e-01
4.33378249e-01 -4.97545063e-01 7.94187248e-01 -1.58307835e-01
-4.75758046e-01 1.94208562e-01 -4.18314427e-01 6.47908449e-01
2.17737257e-01 -7.36555532e-02 -1.18283284e+00 -6.37616292e-02
1.86617896e-02 -3.65978748e-01 6.40511096e-01 3.79342377e-01
1.01974142e+00 -5.35107315e-01 -1.66902363e-01 5.31778812e-01
1.23611999e+00 3.25471073e-01 1.00129068e+00 -2.09679991e-01
4.89545256e-01 4.73843217e-01 5.86293221e-01 2.97327131e-01
5.78215480e-01 1.03574800e+00 2.49099862e-02 -6.10640384e-02
-6.15668654e-01 -7.39819169e-01 9.06498849e-01 1.36488473e+00
-3.51891555e-02 -5.04121542e-01 -6.64021313e-01 4.74355638e-01
-1.73717439e+00 -9.30551112e-01 -2.06192154e-02 2.17964745e+00
1.02581453e+00 1.01593487e-01 2.86599901e-02 5.00439584e-01
8.84463429e-01 2.07685053e-01 -4.44779038e-01 -5.73525548e-01
1.38214203e-02 1.49218276e-01 5.03151655e-01 7.81759858e-01
-6.18224084e-01 1.10556316e+00 5.63263750e+00 1.33383071e+00
-1.37760377e+00 2.05264986e-01 5.64677775e-01 9.71397907e-02
1.67554561e-02 -1.44004449e-01 -1.12236261e+00 1.04786444e+00
1.01962781e+00 3.70110534e-02 5.44606984e-01 4.90198910e-01
7.69214511e-01 -1.72332197e-01 -8.58666778e-01 1.34574366e+00
2.62980491e-01 -9.68550324e-01 -4.66404557e-02 6.82814568e-02
1.96839795e-01 -5.59151061e-02 1.53000921e-01 3.59247416e-01
-2.30423629e-01 -1.14180493e+00 7.04571307e-01 3.88465255e-01
1.30992556e+00 -6.93928659e-01 7.11028755e-01 5.43552220e-01
-1.58558071e+00 -4.19253893e-02 -4.44116294e-01 1.88502014e-01
3.40543479e-01 1.13786504e-01 -1.03143859e+00 3.05682540e-01
4.32481289e-01 2.97274351e-01 -2.80827153e-02 7.76931465e-01
-2.52162308e-01 7.35673666e-01 -3.89815956e-01 5.91704771e-02
6.24411879e-03 3.31978723e-02 6.33851290e-01 1.25512743e+00
4.04081166e-01 1.04662746e-01 -2.59882629e-01 5.88017941e-01
-1.25922829e-01 1.09225892e-01 -5.07502437e-01 1.76692337e-01
5.27609348e-01 7.53054142e-01 -1.16969511e-01 -1.65999308e-01
-2.54618108e-01 1.16860521e+00 -2.35483974e-01 4.80699927e-01
-1.08154833e+00 -5.25405109e-01 6.69060767e-01 5.57321012e-02
2.51487345e-01 -2.84774691e-01 -1.06794439e-01 -1.04562318e+00
2.57438034e-01 -1.06198525e+00 -1.54740766e-01 -1.02211356e+00
-7.19340861e-01 5.06346345e-01 -3.76138121e-01 -1.24704111e+00
-5.11457503e-01 -1.23086184e-01 -4.81187224e-01 1.16856003e+00
-1.95833659e+00 -1.08863342e+00 -1.69279844e-01 7.62300789e-01
1.22434759e+00 -1.54984459e-01 6.47625685e-01 4.09279823e-01
-4.96304780e-01 1.08460820e+00 2.14345828e-02 1.92573905e-01
8.28047395e-01 -7.26099849e-01 2.47811645e-01 8.57189119e-01
-1.72015667e-01 4.38175231e-01 4.92664009e-01 -4.65260416e-01
-1.56493545e+00 -9.01350439e-01 1.27898014e+00 9.20980573e-02
2.50944465e-01 -5.57989359e-01 -7.31334269e-01 2.91735560e-01
1.96591213e-01 -1.58595994e-01 5.75029612e-01 -3.96762222e-01
-3.09202701e-01 -3.92277151e-01 -1.03027236e+00 7.42019236e-01
9.84072447e-01 -8.30501258e-01 -4.74509805e-01 6.41743243e-02
8.37506950e-01 -3.00609976e-01 -5.90710700e-01 3.36165518e-01
7.37717748e-01 -9.04256225e-01 8.10532093e-01 -2.15465650e-02
3.29043627e-01 -3.25807184e-01 -3.57454717e-01 -8.62689137e-01
2.45345876e-01 -1.12419713e+00 -1.26027450e-01 1.59730637e+00
3.47452730e-01 -5.57072341e-01 5.49171746e-01 2.40640312e-01
-1.77588314e-01 -7.79675484e-01 -1.15479302e+00 -7.06773102e-01
-6.01569787e-02 -5.78373253e-01 5.41861951e-01 2.26305053e-01
1.30174637e-01 4.16128814e-01 -7.42755771e-01 8.58249366e-02
2.86286831e-01 -3.90213341e-01 7.28797853e-01 -5.06567359e-01
-3.62288117e-01 -1.68846726e-01 -1.72374249e-01 -1.85661626e+00
2.01140016e-01 -4.77438152e-01 5.29143333e-01 -1.26850736e+00
6.43572509e-02 -5.49160779e-01 -3.54687944e-02 4.13680196e-01
-2.23719060e-01 2.75161862e-01 3.86950135e-01 3.11605781e-01
-1.55183405e-01 7.01468229e-01 9.64988947e-01 1.50080957e-02
-2.76952058e-01 1.11805096e-01 -3.79848480e-01 5.45024455e-01
7.81348467e-01 -1.83215559e-01 -6.54475689e-01 -4.97243971e-01
-1.02538019e-01 4.99654174e-01 7.49258548e-02 -9.17327225e-01
3.85588467e-01 5.39901592e-02 1.00265495e-01 -6.20570600e-01
8.58701587e-01 -4.90593910e-01 -1.95783511e-01 1.48968533e-01
-2.18139380e-01 7.75173604e-02 1.70484424e-01 5.43563962e-01
-2.13449925e-01 -2.16413915e-01 9.26526904e-01 1.93750322e-01
-4.32624340e-01 1.19204611e-01 -4.75366175e-01 5.80252782e-02
6.23362005e-01 -3.71152192e-01 -2.66051859e-01 -6.85527921e-01
-2.79750168e-01 3.69797871e-02 2.19004974e-01 4.88688439e-01
7.40801692e-01 -1.10614347e+00 -7.56914496e-01 4.04511601e-01
-2.82643974e-01 -1.78316608e-01 2.75664091e-01 8.96604717e-01
-1.63236007e-01 6.11371398e-01 1.22786365e-01 -5.60583234e-01
-1.56010544e+00 3.07690203e-01 2.35945359e-01 -1.69841215e-01
-4.84680146e-01 6.71346128e-01 1.33601651e-01 4.47704867e-02
3.78161311e-01 -8.99568573e-02 -1.21649086e-01 -1.41986921e-01
4.69659030e-01 4.45133805e-01 -1.43816710e-01 -9.44234610e-01
-3.02283734e-01 7.13625908e-01 -3.51767540e-02 -4.42411155e-01
9.65368867e-01 -6.50400043e-01 2.62223929e-01 3.08863819e-01
1.38520598e+00 2.81741410e-01 -1.30951083e+00 8.60927254e-02
-4.63762820e-01 -3.06319714e-01 -4.57546301e-02 -7.40209699e-01
-7.66018271e-01 1.30666459e+00 6.15092814e-01 -2.44762227e-01
1.17470431e+00 -3.44332516e-01 1.21446979e+00 -1.50318399e-01
-9.94145572e-02 -1.33869398e+00 -4.04692031e-02 5.43584287e-01
8.37435544e-01 -1.15379441e+00 -3.93103123e-01 -5.64714789e-01
-6.97244048e-01 1.11786032e+00 2.57117331e-01 4.28929448e-01
3.46901327e-01 4.23708737e-01 1.60273492e-01 6.02906704e-01
-8.30195427e-01 -1.58373669e-01 2.32628826e-02 6.53162658e-01
3.40048790e-01 -1.50962397e-01 -3.47234756e-01 4.64274168e-01
-2.63760358e-01 4.07869786e-01 4.27882493e-01 5.06648958e-01
-1.71684191e-01 -1.16515779e+00 -2.96775371e-01 6.74652979e-02
-6.85994089e-01 -5.60197294e-01 1.60781637e-01 4.32052761e-01
1.82671130e-01 1.37732863e+00 -5.44714481e-02 -4.27125156e-01
7.97023473e-04 3.90516460e-01 1.56685144e-01 -1.69511259e-01
-1.99843884e-01 5.03064990e-01 4.44366457e-03 -3.42475116e-01
-1.86139287e-03 -4.81450498e-01 -1.27556026e+00 -2.77575493e-01
-4.87702847e-01 3.23281833e-03 1.06838810e+00 9.79021549e-01
4.94783431e-01 2.97382355e-01 7.44708061e-01 -5.41740477e-01
-7.89112985e-01 -1.14038217e+00 -3.02883595e-01 -5.48077631e-04
4.73535985e-01 -4.83817548e-01 -5.05022645e-01 3.03830206e-01] | [14.330039978027344, 4.992620468139648] |
cb0e4ca4-83cf-4ebd-98f9-79a86b2d987d | long-term-feature-banks-for-detailed-video | 1812.05038 | null | http://arxiv.org/abs/1812.05038v2 | http://arxiv.org/pdf/1812.05038v2.pdf | Long-Term Feature Banks for Detailed Video Understanding | To understand the world, we humans constantly need to relate the present to
the past, and put events in context. In this paper, we enable existing video
models to do the same. We propose a long-term feature bank---supportive
information extracted over the entire span of a video---to augment
state-of-the-art video models that otherwise would only view short clips of 2-5
seconds. Our experiments demonstrate that augmenting 3D convolutional networks
with a long-term feature bank yields state-of-the-art results on three
challenging video datasets: AVA, EPIC-Kitchens, and Charades. | ['Philipp Krähenbühl', 'Chao-yuan Wu', 'Haoqi Fan', 'Ross Girshick', 'Kaiming He', 'Christoph Feichtenhofer'] | 2018-12-12 | long-term-feature-banks-for-detailed-video-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Wu_Long-Term_Feature_Banks_for_Detailed_Video_Understanding_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Wu_Long-Term_Feature_Banks_for_Detailed_Video_Understanding_CVPR_2019_paper.pdf | cvpr-2019-6 | ['egocentric-activity-recognition'] | ['computer-vision'] | [-4.58700895e-01 -3.32430869e-01 -2.29956776e-01 -5.54184318e-01
-8.38493556e-02 -6.59558296e-01 8.54365706e-01 -1.73397779e-01
-3.98458660e-01 3.25737089e-01 6.10163808e-01 -1.06468111e-01
3.06201488e-01 -7.08041012e-01 -8.28899086e-01 2.78896131e-02
-6.63094699e-01 -1.70669809e-01 4.23846960e-01 -4.52963352e-01
6.92529380e-02 4.42343891e-01 -1.45910132e+00 8.08128357e-01
-5.72291687e-02 1.07865739e+00 6.97647110e-02 8.94441247e-01
2.50361919e-01 1.43929195e+00 -2.26177230e-01 -3.25715274e-01
2.08584383e-01 -2.95726776e-01 -7.62128890e-01 1.41793326e-01
6.06990576e-01 -8.99870098e-01 -1.31205499e+00 4.85891819e-01
-8.93431753e-02 2.82148391e-01 3.91192943e-01 -1.20825005e+00
-8.67968261e-01 2.49788925e-01 -2.23051012e-01 7.28051841e-01
7.99764574e-01 4.47714508e-01 8.37140918e-01 -7.65881598e-01
1.07583082e+00 1.35511982e+00 8.13545108e-01 5.45128286e-01
-6.39788032e-01 -4.21876222e-01 6.02326453e-01 7.18332171e-01
-1.03010678e+00 -4.88858551e-01 7.23455191e-01 -7.60346413e-01
1.60967624e+00 -1.07234523e-01 1.23685312e+00 1.73485565e+00
3.85178655e-01 8.31083298e-01 3.58199596e-01 1.36361029e-02
-1.54303268e-01 -4.34873164e-01 8.75937864e-02 7.84555852e-01
-4.27602418e-02 2.59662002e-01 -1.07004797e+00 3.56679380e-01
9.76752758e-01 2.45026901e-01 -2.35830396e-01 -2.37423956e-01
-1.51488125e+00 4.72051382e-01 3.99479210e-01 2.63480335e-01
-6.25575066e-01 5.34452021e-01 5.70233107e-01 3.17127734e-01
5.87601900e-01 2.66831070e-01 -4.06479180e-01 -6.29469812e-01
-9.83632386e-01 4.81193960e-01 7.47874260e-01 9.70170677e-01
4.60838020e-01 1.98691655e-02 7.89654478e-02 6.84735999e-02
1.15823634e-01 1.75972268e-01 8.59891251e-02 -1.22806716e+00
4.46252674e-01 3.43297482e-01 2.49760941e-01 -1.13898325e+00
-3.98501664e-01 -1.38957396e-01 -4.49787289e-01 1.82203087e-03
1.68448463e-01 -1.62768513e-01 -9.81990337e-01 1.74709475e+00
5.23059927e-02 8.16528201e-01 3.82944234e-02 9.10800755e-01
9.36576784e-01 9.35008347e-01 1.84405614e-02 -1.62303951e-02
1.10302246e+00 -1.14770091e+00 -5.99553287e-01 -3.81030619e-01
3.85949969e-01 -2.36677930e-01 7.70760655e-01 3.35281789e-01
-1.22313941e+00 -1.06098270e+00 -1.19382298e+00 -1.95374101e-01
-4.15023088e-01 -1.66954577e-01 7.27936029e-01 -1.18913494e-01
-1.16270423e+00 8.37758780e-01 -1.13021564e+00 -7.49293864e-01
3.76407683e-01 -1.64117649e-01 -8.67177248e-01 -2.13138163e-02
-1.20496440e+00 1.04172611e+00 4.06319231e-01 1.86853483e-02
-1.37546670e+00 -7.81215549e-01 -1.01118147e+00 2.19956130e-01
3.72810662e-01 -8.04481864e-01 1.30929196e+00 -8.45737696e-01
-9.99266863e-01 7.88906455e-01 -2.49821424e-01 -6.66523099e-01
3.94474387e-01 -9.27073300e-01 -6.11822665e-01 5.55003166e-01
-1.02783404e-01 6.86108232e-01 6.56004369e-01 -9.22652543e-01
-6.81729376e-01 -1.90124542e-01 6.76301241e-01 8.27183109e-03
-2.69945264e-01 2.76933044e-01 -7.53943682e-01 -6.59123063e-01
-1.72167718e-01 -8.04211974e-01 -1.29183486e-01 2.40458205e-01
2.70868778e-01 -2.81147659e-02 1.08248389e+00 -8.05950999e-01
1.28814065e+00 -2.30303526e+00 4.75442618e-01 -4.47895110e-01
2.97001123e-01 1.34234786e-01 -2.89857447e-01 5.92798889e-01
-1.54608905e-01 1.29182041e-01 5.12641311e-01 -3.82651538e-01
-1.06238119e-01 1.75826356e-01 -3.81689757e-01 4.28793013e-01
3.35966051e-01 8.95392656e-01 -1.09268141e+00 -3.00820112e-01
4.24638033e-01 6.51916683e-01 -6.03375733e-01 4.56211567e-01
-3.56473744e-01 2.83095837e-01 -2.10911095e-01 5.77384472e-01
1.69528529e-01 -3.70001078e-01 1.12042978e-01 -2.99356133e-01
-1.61049604e-01 1.20829657e-01 -6.06138647e-01 2.31032205e+00
-2.06869766e-02 1.28473759e+00 -3.71719003e-01 -8.15955758e-01
5.80102146e-01 5.36145389e-01 6.16010487e-01 -7.30836451e-01
2.89284512e-02 -2.94205874e-01 -3.10375094e-01 -8.65921199e-01
4.81973559e-01 1.66345268e-01 -5.51065169e-02 -3.81470062e-02
6.11759782e-01 2.21610978e-01 4.45181757e-01 5.06168723e-01
1.26177406e+00 7.08451152e-01 2.81042099e-01 -1.40324365e-02
3.73662353e-01 -9.38035250e-02 7.07368612e-01 5.88583291e-01
-4.29743767e-01 6.24035001e-01 6.13921344e-01 -1.25897324e+00
-1.08706236e+00 -1.04458249e+00 5.53115129e-01 1.08257174e+00
8.96247923e-02 -1.02146053e+00 -4.95502174e-01 -8.53743970e-01
-1.11120574e-01 6.57259941e-01 -9.73130226e-01 -2.43230350e-02
-8.32513690e-01 -3.39379758e-02 4.59048927e-01 1.08446789e+00
5.04613876e-01 -9.47061598e-01 -1.04863262e+00 3.32957387e-01
-2.10326999e-01 -1.50383711e+00 -3.22638005e-01 -1.77246794e-01
-6.15671337e-01 -1.19600630e+00 -3.92649502e-01 -5.37721634e-01
-7.93669838e-03 4.04480517e-01 1.55928159e+00 -3.87363546e-02
-1.94929540e-01 6.26873255e-01 -6.00595355e-01 -3.90836954e-01
5.22012785e-02 -3.46867770e-01 1.60847738e-01 -2.41598740e-01
5.56557953e-01 -8.43674064e-01 -6.07902646e-01 9.14751962e-02
-6.18617058e-01 2.53348380e-01 1.15906239e-01 5.11016607e-01
2.57652223e-01 -2.73655057e-01 3.42639476e-01 -4.20659244e-01
-1.51797622e-01 -7.89231300e-01 -2.94200748e-01 2.76959300e-01
2.18463749e-01 -2.99724638e-01 5.41474879e-01 -5.94052136e-01
-1.05168450e+00 -5.66741638e-02 7.19126090e-02 -1.06252491e+00
-3.72350752e-01 7.24807262e-01 6.48062751e-02 3.75307053e-01
5.63885689e-01 -2.61596479e-02 -4.01114553e-01 -4.44050699e-01
5.23560166e-01 1.34910956e-01 1.04448617e+00 -4.16799873e-01
5.94958425e-01 6.69175148e-01 3.72696668e-02 -7.22593725e-01
-1.05371952e+00 -5.00693321e-01 -1.03399396e+00 -7.29401469e-01
1.12816131e+00 -1.42913210e+00 -4.94764447e-01 5.11314154e-01
-1.37489069e+00 -3.56696218e-01 -1.44915506e-01 5.83571553e-01
-7.11416066e-01 2.47542232e-01 -7.59472966e-01 -4.27177757e-01
1.16000794e-01 -8.02488267e-01 6.71556890e-01 3.37571353e-01
-3.02540988e-01 -9.96043205e-01 2.05662966e-01 -7.85714574e-03
2.65657634e-01 7.61314213e-01 4.85410988e-01 -4.12862778e-01
-6.35329723e-01 -1.30071640e-01 -1.89949438e-01 1.35428578e-01
3.71197350e-02 1.94576368e-01 -9.06280518e-01 -2.93591946e-01
-1.31006420e-01 -3.91210467e-01 1.12671876e+00 3.03780347e-01
1.14739382e+00 -1.25193954e-01 -3.83437127e-01 8.85697126e-01
1.24161112e+00 3.42612803e-01 7.09272802e-01 4.41436768e-01
7.24661052e-01 3.04321796e-01 6.61909640e-01 7.05320776e-01
5.54730296e-01 4.01801437e-01 6.28797591e-01 1.91270530e-01
-3.47420663e-01 -4.71170366e-01 5.44823289e-01 6.45494699e-01
-5.04743814e-01 -5.77928364e-01 -8.78776968e-01 8.74137282e-01
-2.19812369e+00 -1.49900782e+00 2.15279624e-01 1.37276721e+00
3.98010433e-01 4.14404333e-01 9.72378328e-02 -3.31144929e-01
2.18794301e-01 8.16890776e-01 -4.94172543e-01 -4.98127267e-02
-7.09775165e-02 -1.96920276e-01 6.21319748e-02 1.28697053e-01
-1.43149316e+00 9.81073201e-01 7.27050877e+00 4.92398925e-02
-1.14159739e+00 -9.24344286e-02 4.19514567e-01 -5.97894371e-01
8.91397744e-02 1.20298311e-01 -5.51956058e-01 2.78175801e-01
1.34643185e+00 -2.36090466e-01 3.08820665e-01 8.81655633e-01
2.58919984e-01 -4.95166183e-02 -1.78267837e+00 1.15479636e+00
3.71225297e-01 -1.74703217e+00 1.58952609e-01 -3.59613299e-02
5.07119775e-01 1.64074749e-01 -5.33408709e-02 3.80112588e-01
4.78471108e-02 -9.04574335e-01 1.24777520e+00 9.02558267e-01
7.54247189e-01 -5.10959208e-01 4.94823098e-01 1.50452882e-01
-1.46187794e+00 -1.81642801e-01 -1.55467436e-01 -4.27216977e-01
6.54663324e-01 1.46083757e-01 -4.54383641e-01 6.84653759e-01
1.05734110e+00 1.59827125e+00 -6.77254558e-01 7.73108721e-01
3.24633531e-02 4.52848673e-01 2.80911643e-02 4.16790128e-01
5.47150016e-01 2.74549901e-01 5.07993639e-01 1.39163578e+00
5.90265632e-01 5.15203416e-01 1.82581648e-01 4.37193274e-01
-8.27176496e-02 -6.35215104e-01 -9.51166451e-01 -2.90358067e-01
1.73905820e-01 7.80839980e-01 -3.47757429e-01 -5.69182396e-01
-9.27163541e-01 9.70965385e-01 3.37784171e-01 3.06620926e-01
-1.11272478e+00 1.97797455e-02 9.71613586e-01 1.60851717e-01
6.37379646e-01 -6.55396104e-01 3.39568555e-01 -1.59893799e+00
6.85859821e-04 -7.35266387e-01 4.33559060e-01 -1.23880625e+00
-1.24080551e+00 6.64657116e-01 2.84828305e-01 -1.28190744e+00
-4.99959826e-01 -7.30685234e-01 -6.54717982e-01 3.27593088e-01
-1.37394416e+00 -1.44310176e+00 -5.95420480e-01 6.77497685e-01
9.82225955e-01 -1.84694201e-01 6.51981354e-01 1.88929796e-01
-3.40080947e-01 2.54491791e-02 -2.81446248e-01 4.64997679e-01
6.41602278e-01 -8.51162016e-01 1.01005888e+00 1.03737187e+00
3.45714331e-01 4.62061793e-01 8.09628427e-01 -6.90875351e-01
-1.59751844e+00 -8.97812307e-01 6.79226279e-01 -7.23078310e-01
8.71579707e-01 -2.11767599e-01 -7.51863420e-01 1.22973776e+00
5.85785270e-01 2.88561344e-01 5.92680216e-01 1.96928173e-01
-8.34444821e-01 -6.68395236e-02 -6.71238303e-01 5.19813120e-01
1.50182581e+00 -8.03576648e-01 -9.70112205e-01 -2.71446677e-03
9.92406905e-01 -3.17099661e-01 -8.89066279e-01 3.61665517e-01
1.06378293e+00 -1.13677466e+00 1.11922908e+00 -1.26967299e+00
8.19606066e-01 -1.63103074e-01 -5.23365140e-01 -1.26314116e+00
-6.29585922e-01 -6.37342632e-01 -8.05211604e-01 8.88433278e-01
-5.18878317e-03 1.67415738e-01 5.82391620e-01 4.29950297e-01
-3.10696721e-01 -4.54342693e-01 -6.61874235e-01 -7.99582481e-01
-2.27285624e-01 -6.14787638e-01 6.60699904e-01 9.94437456e-01
1.80359289e-01 1.54194593e-01 -6.96751893e-01 2.45882064e-01
3.97462904e-01 -2.13278040e-01 7.59879410e-01 -1.23084772e+00
-2.06299067e-01 -1.37752965e-01 -8.34031105e-01 -9.88836348e-01
2.73015887e-01 -4.37879190e-02 -3.27118427e-01 -1.56323683e+00
3.33540529e-01 3.21946889e-01 -5.69338381e-01 3.46649319e-01
6.25108406e-02 -5.28345890e-02 5.65335155e-01 -9.53659415e-02
-1.20329905e+00 4.48700756e-01 9.31357741e-01 2.70661525e-03
2.12066382e-01 -6.39238298e-01 -2.71794945e-01 9.99346852e-01
4.16255742e-01 -5.38673326e-02 -5.31998277e-01 -9.98233199e-01
3.44774723e-01 4.24844056e-01 6.12138391e-01 -1.43845117e+00
9.41464752e-02 -4.47499692e-01 8.31446469e-01 -9.20625567e-01
7.54306734e-01 -7.65997589e-01 2.74257004e-01 1.42790288e-01
-2.91153282e-01 5.20790935e-01 3.95275950e-01 8.30737054e-01
-4.13555413e-01 3.30847591e-01 2.85061002e-01 -5.42759120e-01
-1.50159407e+00 4.46053803e-01 -2.69889027e-01 -1.04356535e-01
1.15136993e+00 3.27873006e-02 -5.75786412e-01 -7.29888260e-01
-8.34823668e-01 3.23007524e-01 6.75802350e-01 9.50240374e-01
7.36647487e-01 -1.44863331e+00 -7.28770971e-01 5.38828485e-02
1.24655895e-01 -3.67354602e-01 5.98603845e-01 3.24078143e-01
-6.98247969e-01 5.53034723e-01 -6.33680522e-01 -5.88689387e-01
-1.16642320e+00 9.68722284e-01 2.16193035e-01 -4.22301143e-02
-8.33708107e-01 8.39947581e-01 3.22933495e-01 2.71556497e-01
3.95597339e-01 -4.25038010e-01 -2.89303958e-01 2.20703751e-01
8.99482727e-01 8.45485032e-02 -2.82177478e-01 -8.15768540e-01
-5.31473935e-01 3.60845923e-01 7.75053818e-03 -6.72697872e-02
1.84832263e+00 -3.06320488e-01 2.42294103e-01 7.05816627e-01
1.14100873e+00 -6.20700836e-01 -1.87322354e+00 -1.47594705e-01
-1.92098230e-01 -7.25454867e-01 3.50614749e-02 -7.21249819e-01
-9.55576539e-01 9.84759688e-01 3.47988278e-01 9.27259699e-02
1.11117506e+00 -5.51949302e-03 7.76830673e-01 6.62760735e-01
4.90436554e-01 -8.34950149e-01 4.03945595e-01 6.87898159e-01
9.90305126e-01 -1.18511820e+00 1.03029199e-01 3.10350973e-02
-7.83787727e-01 1.35732734e+00 8.21560860e-01 -2.61255711e-01
9.13071573e-01 5.02127595e-02 -1.92337736e-01 -3.72960746e-01
-1.54529691e+00 -3.09548061e-02 3.35993230e-01 4.88738656e-01
3.57501745e-01 -2.41148308e-01 3.97894174e-01 6.11349940e-01
1.62647009e-01 4.99323577e-01 7.37357616e-01 1.28409219e+00
-2.82075584e-01 -6.15856409e-01 8.49332940e-03 4.76887375e-02
-4.45789337e-01 1.45575225e-01 -1.76047564e-01 9.29218054e-01
2.05176145e-01 1.05888665e+00 3.44704330e-01 -7.11277008e-01
4.19417024e-01 1.69375181e-01 5.78193009e-01 -4.17416900e-01
-4.65014428e-01 7.17133060e-02 4.50596988e-01 -1.20327103e+00
-8.88624132e-01 -8.02860200e-01 -8.40503216e-01 -5.85924208e-01
2.56246030e-01 -3.36991340e-01 2.98825532e-01 8.89172018e-01
4.63391036e-01 6.09887540e-01 3.25965315e-01 -1.33505845e+00
2.21206285e-02 -9.16677177e-01 -3.46319526e-01 5.33453524e-01
7.90522039e-01 -7.66045630e-01 -1.19370632e-01 6.14957869e-01] | [8.505716323852539, 0.569106936454773] |
d0ed48b7-35f0-4562-856f-681e08080504 | rethinking-self-supervision-objectives-for-1 | 2110.07198 | null | https://arxiv.org/abs/2110.07198v2 | https://arxiv.org/pdf/2110.07198v2.pdf | Rethinking Self-Supervision Objectives for Generalizable Coherence Modeling | Given the claims of improved text generation quality across various pre-trained neural models, we consider the coherence evaluation of machine generated text to be one of the principal applications of coherence models that needs to be investigated. Prior work in neural coherence modeling has primarily focused on devising new architectures for solving the permuted document task. We instead use a basic model architecture and show significant improvements over state of the art within the same training regime. We then design a harder self-supervision objective by increasing the ratio of negative samples within a contrastive learning setup, and enhance the model further through automatic hard negative mining coupled with a large global negative queue encoded by a momentum encoder. We show empirically that increasing the density of negative samples improves the basic model, and using a global negative queue further improves and stabilizes the model while training with hard negative samples. We evaluate the coherence model on task-independent test sets that resemble real-world applications and show significant improvements in coherence evaluations of downstream tasks. | ['Xiang Lin', 'Shafiq Joty', 'Prathyusha Jwalapuram'] | 2021-10-14 | rethinking-self-supervision-objectives-for | https://aclanthology.org/2022.acl-long.418 | https://aclanthology.org/2022.acl-long.418.pdf | acl-2022-5 | ['coherence-evaluation'] | ['natural-language-processing'] | [ 3.26858550e-01 4.24743146e-01 -2.04715937e-01 -3.18454891e-01
-8.81531477e-01 -4.84715849e-01 1.14316308e+00 7.52028450e-02
-5.90231895e-01 9.50093448e-01 7.27029324e-01 -3.39590311e-01
-4.65829708e-02 -5.24434507e-01 -5.77145219e-01 -4.89032656e-01
-1.43572643e-01 8.13598454e-01 2.13226125e-01 -3.13196629e-01
3.56899559e-01 -1.51444286e-01 -1.33092618e+00 5.36966205e-01
8.27308834e-01 4.59356815e-01 1.71237573e-01 8.53187382e-01
2.41115212e-01 9.52207148e-01 -8.97755802e-01 -4.65107143e-01
2.39758924e-01 -6.33181632e-01 -1.07309699e+00 -3.00158653e-03
3.84763151e-01 -1.98485345e-01 -1.42140845e-02 8.20405483e-01
5.66463947e-01 2.63868451e-01 6.93419337e-01 -9.41408575e-01
-6.59989893e-01 1.21526992e+00 -3.89790863e-01 5.15993953e-01
-1.04704924e-01 2.38269255e-01 1.47936106e+00 -6.80231333e-01
6.62552416e-01 1.19278443e+00 6.18461430e-01 5.16026318e-01
-1.58272874e+00 -4.91276115e-01 1.04960747e-01 -5.33771701e-02
-8.59909475e-01 -7.05105484e-01 3.97457659e-01 -3.38477194e-01
1.31867838e+00 2.23879248e-01 5.92612207e-01 1.19153905e+00
2.53797621e-01 8.56957853e-01 7.58339465e-01 -5.61726630e-01
1.63607538e-01 4.20458913e-02 4.18441921e-01 4.37808067e-01
3.96799326e-01 -5.55372238e-03 -7.09311008e-01 -2.84961388e-02
3.77447546e-01 -6.07782960e-01 -3.00900131e-01 -1.83224455e-01
-1.26620972e+00 9.07406628e-01 1.37138054e-01 4.28574234e-01
-1.65544868e-01 3.61581564e-01 3.68494362e-01 4.51000124e-01
9.11017001e-01 1.09357357e+00 -4.66583878e-01 -3.33806425e-01
-1.29460633e+00 5.19527197e-01 8.79987180e-01 7.00656593e-01
5.69464445e-01 -1.68615878e-02 -7.00485289e-01 8.66120815e-01
1.95835680e-01 1.92061380e-01 7.78053224e-01 -1.13588452e+00
7.24027038e-01 3.06240231e-01 1.65617224e-02 -5.95993042e-01
-3.88375312e-01 -9.12729919e-01 -6.95276976e-01 2.58287061e-02
4.75721240e-01 -3.70243907e-01 -8.00226510e-01 2.06299424e+00
-1.47764683e-01 -2.17358738e-01 2.47651726e-01 5.99984169e-01
2.79456079e-01 6.51302338e-01 -1.02405354e-01 -3.01855713e-01
8.94847691e-01 -1.31842935e+00 -6.59066200e-01 -4.53271866e-01
1.05077195e+00 -7.15698421e-01 1.30384636e+00 4.75462884e-01
-1.37513840e+00 -2.91491628e-01 -1.19269264e+00 -2.56195694e-01
8.75652879e-02 1.65605202e-01 6.41394138e-01 3.29397738e-01
-1.39958525e+00 9.22918558e-01 -9.64421451e-01 -6.61341473e-02
4.79202151e-01 3.06039870e-01 2.28294954e-02 8.78162012e-02
-1.17073095e+00 8.86066616e-01 3.56529206e-01 -2.31335144e-02
-5.89334726e-01 -5.73348045e-01 -6.51360095e-01 2.72191435e-01
1.79049581e-01 -1.22342205e+00 1.61546946e+00 -9.71362829e-01
-1.45305014e+00 6.39714718e-01 -2.08150581e-01 -6.43949747e-01
4.99264300e-01 -4.80009854e-01 5.38551137e-02 -1.34182274e-01
2.59884655e-01 8.05488229e-01 6.29948854e-01 -9.98873353e-01
-4.71953183e-01 -1.24295458e-01 -1.57611802e-01 3.29415977e-01
-6.34411097e-01 -1.45332351e-01 -2.10907623e-01 -6.50318086e-01
-2.59497970e-01 -9.65439558e-01 -2.05080941e-01 -6.35459960e-01
-4.64414418e-01 -4.61799204e-01 3.46681654e-01 -3.81373256e-01
1.48182344e+00 -1.73513949e+00 2.58671165e-01 1.41862974e-01
3.05717349e-01 4.20537069e-02 -4.62883830e-01 3.08096319e-01
-2.72746347e-02 3.48934501e-01 -2.39270523e-01 -7.86406815e-01
9.71318129e-03 -1.74874872e-01 -4.67852324e-01 4.69171293e-02
4.56291229e-01 1.11736357e+00 -9.77886021e-01 -3.08485687e-01
-4.42987680e-01 7.16824681e-02 -1.04676783e+00 1.03768207e-01
-5.74561238e-01 2.87558679e-02 -1.25286832e-01 -7.55395666e-02
1.49316013e-01 -8.40194941e-01 3.33530635e-01 1.64066583e-01
2.81991046e-02 8.69573534e-01 -8.39030802e-01 1.56321836e+00
-2.71740854e-01 8.94671798e-01 -1.25675932e-01 -8.64486933e-01
4.69889313e-01 3.20174038e-01 2.63554547e-02 -7.93648303e-01
-5.71321435e-02 3.93582806e-02 5.91732264e-01 -2.42574289e-01
9.81757164e-01 -2.55314201e-01 1.56612188e-01 1.12263978e+00
9.94286686e-02 -1.61999196e-01 5.13718724e-01 5.85570097e-01
1.34029257e+00 4.01070863e-02 -9.69701335e-02 -4.44669515e-01
1.34121114e-02 -2.32643168e-02 2.06562802e-01 9.37932312e-01
2.77517363e-02 7.17777729e-01 8.20368707e-01 -1.09057292e-01
-1.34904969e+00 -7.83659160e-01 1.83596108e-02 1.40320528e+00
-2.97656864e-01 -7.24163651e-01 -7.66940057e-01 -6.42775536e-01
-1.79051831e-01 9.03251469e-01 -6.35593235e-01 -1.08462781e-01
-7.86877036e-01 -1.24765897e+00 5.24762094e-01 4.26841885e-01
1.60298273e-01 -1.13042223e+00 -5.48401058e-01 2.09505990e-01
-3.81692499e-01 -8.07559133e-01 -3.23264599e-01 5.51006079e-01
-9.82995629e-01 -6.61569118e-01 -5.49648941e-01 -5.02984226e-01
3.44765365e-01 1.97840527e-01 1.71416354e+00 1.58037171e-01
1.62251770e-01 -1.34636790e-01 -1.35287553e-01 -2.41491348e-01
-6.32084131e-01 7.51930892e-01 -8.32456350e-02 -4.69399124e-01
8.10672063e-03 -5.89014769e-01 -3.98517013e-01 -1.93868116e-01
-8.62892568e-01 4.14413124e-01 6.22900605e-01 1.19719326e+00
4.19730656e-02 -2.08421350e-01 9.00447428e-01 -1.10906971e+00
1.31630588e+00 -4.58786994e-01 -3.25800300e-01 4.86627035e-02
-9.61442351e-01 7.00206578e-01 4.81159806e-01 -3.19386810e-01
-9.75854158e-01 -5.67556739e-01 5.94860129e-02 -1.96274444e-02
3.70023847e-01 7.07867920e-01 2.18067840e-01 8.15006137e-01
9.71717238e-01 -7.33261229e-03 1.39191508e-01 -2.92439517e-02
5.62443137e-01 3.89014661e-01 3.58868480e-01 -5.54235935e-01
5.80438673e-01 1.42934695e-01 -1.81821898e-01 -4.73274559e-01
-1.11611056e+00 -1.49595454e-01 -3.40129822e-01 1.96666092e-01
5.91931820e-01 -1.03568828e+00 -2.51858741e-01 2.25929067e-01
-1.18062973e+00 -1.09499669e+00 -3.67147386e-01 3.05011451e-01
-5.47136486e-01 1.73989847e-01 -6.89004302e-01 -5.76692641e-01
-6.90857112e-01 -8.86240482e-01 1.02115202e+00 -1.46412119e-01
-8.13458562e-01 -1.16047192e+00 5.04078448e-01 2.70425797e-01
6.36432588e-01 -3.14411402e-01 1.12332189e+00 -7.43990719e-01
-6.95542216e-01 4.47497927e-02 4.79817390e-03 4.50247943e-01
-1.08323924e-01 -1.04140872e-02 -1.05213189e+00 -1.06917769e-01
-1.47414580e-01 -6.67537570e-01 1.40773022e+00 3.73478025e-01
8.04258823e-01 -4.06716049e-01 -2.83604562e-01 3.25165123e-01
9.14735913e-01 -4.02071685e-01 5.46806753e-01 4.64212120e-01
3.30397040e-01 5.63468516e-01 2.13521481e-01 3.96406293e-01
3.77652377e-01 4.26845998e-01 1.15893573e-01 -9.16113630e-02
-1.27520755e-01 -3.28797624e-02 4.14218932e-01 7.83668220e-01
1.33934751e-01 -6.27660513e-01 -9.17860150e-01 7.19916344e-01
-2.04871035e+00 -1.36216760e+00 7.42765656e-03 1.93639731e+00
1.49627304e+00 5.83568037e-01 3.92473228e-02 6.92731291e-02
3.19975257e-01 3.62914264e-01 -1.89064384e-01 -2.49850273e-01
-1.54328257e-01 5.10510743e-01 7.51884952e-02 8.26295018e-01
-9.72855926e-01 1.00666523e+00 7.19603300e+00 6.71080291e-01
-1.11341929e+00 -5.47568724e-02 9.33949292e-01 -6.81726098e-01
-7.76665449e-01 1.19763091e-02 -7.63808012e-01 3.97001445e-01
1.21998119e+00 -1.64168760e-01 3.62115175e-01 4.88264471e-01
3.64027262e-01 -3.44536036e-01 -1.58828926e+00 4.53192800e-01
6.97164759e-02 -1.48516655e+00 -1.37616089e-02 1.06639557e-01
1.06405151e+00 4.69408035e-01 1.16447568e-01 4.53555852e-01
7.67884612e-01 -1.14819849e+00 8.46229494e-01 4.33893770e-01
4.00197655e-01 -4.87586111e-01 7.39113986e-01 5.81611574e-01
-4.03332591e-01 -6.37057349e-02 -2.01539919e-01 -3.55856240e-01
2.18769565e-01 7.77691424e-01 -1.06569624e+00 9.40292254e-02
4.10351962e-01 7.46341467e-01 -9.02536273e-01 6.87793374e-01
-1.68163866e-01 1.01191580e+00 -3.23458105e-01 -1.32244498e-01
1.93583176e-01 7.75321946e-02 3.40888619e-01 1.37743473e+00
9.93591696e-02 -3.08713317e-01 -1.74535096e-01 1.07772732e+00
-3.38964164e-01 -7.99240321e-02 -5.00348866e-01 -1.61616325e-01
3.74021798e-01 1.21828473e+00 -4.89267945e-01 -5.83488762e-01
-6.69430643e-02 7.71862507e-01 7.45333731e-01 3.10077637e-01
-5.29563189e-01 -1.52168602e-01 2.96118647e-01 1.82527587e-01
3.20808381e-01 -2.61131138e-01 -7.88529932e-01 -1.17911935e+00
7.03755021e-02 -8.83567691e-01 3.61596160e-02 -6.45296633e-01
-1.08770883e+00 5.71889639e-01 -4.81854156e-02 -7.79603958e-01
-9.03737545e-01 -1.10671781e-01 -7.37709105e-01 1.03006816e+00
-1.47665966e+00 -7.51430571e-01 1.28162820e-02 7.11111873e-02
4.38147783e-01 -1.75152972e-01 6.87843680e-01 -1.17214881e-01
-7.06727147e-01 6.62567437e-01 1.49812654e-01 -1.22321554e-01
8.81642342e-01 -1.37249458e+00 6.68008864e-01 9.27631617e-01
2.14639083e-01 1.01702142e+00 8.10770333e-01 -6.71237886e-01
-7.55973756e-01 -1.09568524e+00 1.25757051e+00 -7.72697091e-01
6.76925480e-01 -4.88284647e-01 -7.09530175e-01 7.39428401e-01
6.52362704e-01 -4.23594594e-01 7.03998446e-01 4.85210121e-01
-2.57902831e-01 1.62250191e-01 -4.50461507e-01 6.61043823e-01
1.04078758e+00 -3.67341191e-01 -5.48304737e-01 4.47579920e-01
8.24238956e-01 -1.79276213e-01 -4.76215541e-01 2.84306496e-01
4.84615862e-01 -9.84808326e-01 5.51313102e-01 -6.92393661e-01
1.09427392e+00 -2.06903307e-04 1.31384701e-01 -1.51295483e+00
-4.78431284e-01 -5.78621864e-01 -2.51108557e-01 9.67634141e-01
8.76186907e-01 -3.25227112e-01 9.32998955e-01 4.60363060e-01
-4.89605874e-01 -9.45184469e-01 -5.77129126e-01 -5.34216583e-01
4.80862826e-01 -2.23895088e-01 5.59116118e-02 7.28432775e-01
2.22947612e-01 1.19137967e+00 -2.39286438e-01 -3.55382353e-01
2.55361557e-01 -5.07166330e-03 8.02072823e-01 -1.12829447e+00
-7.93396652e-01 -7.32523143e-01 3.20826113e-01 -1.33207870e+00
2.92969286e-01 -1.05847788e+00 2.37235233e-01 -1.60495913e+00
5.86773813e-01 -4.11523759e-01 -5.92859462e-03 4.85323071e-01
-5.37928164e-01 3.38764101e-01 3.95751633e-02 4.19687837e-01
-8.09917748e-01 6.98257685e-01 1.05921996e+00 -5.67100197e-02
-2.21069887e-01 -7.97103345e-02 -1.19511032e+00 3.26667696e-01
7.47192085e-01 -3.66922379e-01 -7.18419611e-01 -8.44169080e-01
6.91335678e-01 -2.44866163e-01 1.06066167e-01 -8.81775379e-01
2.49771163e-01 1.66324452e-02 4.16019619e-01 -4.52394456e-01
1.73181891e-01 -3.14582326e-02 -3.30005497e-01 3.90087456e-01
-9.73501325e-01 3.58472317e-01 1.87350705e-01 3.97924811e-01
-3.79424472e-03 -2.05483243e-01 6.62077427e-01 -1.42471164e-01
5.28977178e-02 -2.87347853e-01 -4.15633261e-01 4.42419440e-01
4.59021926e-01 7.48289824e-02 -6.69398487e-01 -5.06789446e-01
-4.57865953e-01 2.50010103e-01 3.52974921e-01 3.12726796e-01
1.69656679e-01 -1.10211849e+00 -9.10112739e-01 1.72046646e-01
-1.36627808e-01 6.84163943e-02 -2.57669568e-01 7.42620528e-01
-2.15894610e-01 6.74893260e-01 2.19476625e-01 -6.00122750e-01
-1.15884411e+00 9.23460871e-02 3.04284543e-01 -8.78555238e-01
-2.26473451e-01 8.58579040e-01 7.89430067e-02 -1.82735682e-01
1.92945585e-01 -3.84801924e-01 -1.41766533e-01 8.66810754e-02
3.60771060e-01 2.07472712e-01 3.08871806e-01 -1.17139749e-01
1.41660348e-02 -1.58833712e-01 -4.33905602e-01 -6.74536765e-01
1.41870379e+00 2.02472806e-02 -5.05318604e-02 4.22482967e-01
9.86242831e-01 4.51352680e-03 -1.19529998e+00 -2.32679635e-01
1.21718459e-01 3.88268679e-02 -2.70735249e-02 -8.60453308e-01
-5.83500922e-01 8.41727495e-01 1.39710888e-01 2.60016680e-01
6.89353585e-01 5.93369007e-02 5.33898234e-01 6.54928803e-01
-1.01379804e-01 -1.20425630e+00 7.22809911e-01 9.71546292e-01
7.93217361e-01 -1.15767801e+00 1.16627827e-01 9.81462225e-02
-5.39770842e-01 8.71112168e-01 5.84161937e-01 -1.07949868e-01
4.94477093e-01 4.32307929e-01 -1.15793720e-01 -1.22672811e-01
-1.55956900e+00 9.14937705e-02 2.25592151e-01 1.25432059e-01
9.75896776e-01 -7.70885572e-02 -4.27645475e-01 6.04725659e-01
-7.99270213e-01 -6.01797067e-02 7.00735092e-01 7.94498861e-01
-6.43179536e-01 -1.19317937e+00 1.07002944e-01 7.06328928e-01
-5.36601126e-01 -7.27141678e-01 -3.75764042e-01 5.69154203e-01
-2.11261630e-01 9.38962817e-01 4.66169685e-01 -4.08363938e-01
-1.30097046e-01 3.71784806e-01 6.27083898e-01 -1.08857894e+00
-7.66398251e-01 1.13845080e-01 3.87707204e-01 -2.18809888e-01
-2.53089815e-01 -7.55832851e-01 -1.08374465e+00 -3.15941960e-01
-5.44860840e-01 1.98020667e-01 3.25792700e-01 1.01804054e+00
5.26578605e-01 5.83537340e-01 4.52082396e-01 -6.93645954e-01
-1.08099079e+00 -1.48620737e+00 -1.14077523e-01 3.85265768e-01
4.70931888e-01 -1.58615172e-01 -3.29946786e-01 1.85577154e-01] | [11.617504119873047, 8.962964057922363] |
eac7594f-6b43-47a9-8798-2a0335007460 | jointly-harnessing-prior-structures-and | 2207.03714 | null | https://arxiv.org/abs/2207.03714v1 | https://arxiv.org/pdf/2207.03714v1.pdf | Jointly Harnessing Prior Structures and Temporal Consistency for Sign Language Video Generation | Sign language is the window for people differently-abled to express their feelings as well as emotions. However, it remains challenging for people to learn sign language in a short time. To address this real-world challenge, in this work, we study the motion transfer system, which can transfer the user photo to the sign language video of specific words. In particular, the appearance content of the output video comes from the provided user image, while the motion of the video is extracted from the specified tutorial video. We observe two primary limitations in adopting the state-of-the-art motion transfer methods to sign language generation:(1) Existing motion transfer works ignore the prior geometrical knowledge of the human body. (2) The previous image animation methods only take image pairs as input in the training stage, which could not fully exploit the temporal information within videos. In an attempt to address the above-mentioned limitations, we propose Structure-aware Temporal Consistency Network (STCNet) to jointly optimize the prior structure of human with the temporal consistency for sign language video generation. There are two main contributions in this paper. (1) We harness a fine-grained skeleton detector to provide prior knowledge of the body keypoints. In this way, we ensure the keypoint movement in a valid range and make the model become more explainable and robust. (2) We introduce two cycle-consistency losses, i.e., short-term cycle loss and long-term cycle loss, which are conducted to assure the continuity of the generated video. We optimize the two losses and keypoint detector network in an end-to-end manner. | ['Yi Yang', 'Bang Zhang', 'Xiaohan Wang', 'Zhedong Zheng', 'Yucheng Suo'] | 2022-07-08 | null | null | null | null | ['image-animation'] | ['computer-vision'] | [ 8.72156303e-03 -1.91721305e-01 -3.19103122e-01 -2.19291553e-01
-3.81478667e-01 -3.70964825e-01 5.31519473e-01 -9.83912706e-01
-4.04848814e-01 6.64208353e-01 3.30963463e-01 1.37035698e-01
1.56903550e-01 -4.52445120e-01 -7.36459374e-01 -6.41708076e-01
2.83091784e-01 -1.62994668e-01 4.37567264e-01 -2.12141559e-01
1.14811830e-01 3.72152984e-01 -1.56969988e+00 2.21846282e-01
9.76822436e-01 6.70714498e-01 3.55750263e-01 6.46660745e-01
-3.33293565e-02 9.76370692e-01 -3.42159063e-01 -1.57154471e-01
2.58412778e-01 -9.17321026e-01 -7.12004066e-01 2.82474548e-01
5.16494453e-01 -7.13499486e-01 -7.85622954e-01 9.19927418e-01
6.64061666e-01 3.69185895e-01 5.19252181e-01 -1.53562438e+00
-3.69037569e-01 1.16290957e-01 -6.41417742e-01 -1.14084557e-01
3.51941705e-01 7.01476336e-01 6.58586979e-01 -7.06084669e-01
1.07094729e+00 1.29942095e+00 3.80647093e-01 1.09799063e+00
-6.21609211e-01 -8.10237050e-01 4.46566135e-01 4.00868595e-01
-1.13444042e+00 -3.89097869e-01 9.22972322e-01 -3.97093624e-01
5.23552656e-01 3.02729756e-02 1.13899541e+00 1.20698166e+00
2.29883064e-02 1.08790410e+00 8.80560100e-01 -4.61261481e-01
-6.05641119e-02 -3.19031119e-01 -2.26552308e-01 8.12861800e-01
-8.80859494e-02 2.10771024e-01 -7.40061760e-01 2.52565175e-01
1.24492931e+00 -9.02450532e-02 -7.16741562e-01 -5.06557167e-01
-1.30401874e+00 5.23712277e-01 3.15817326e-01 3.22883010e-01
-3.35524082e-01 4.85810190e-01 2.01086357e-01 1.95200324e-01
-2.46566199e-02 -1.96253136e-01 -2.22915739e-01 -3.79882574e-01
-8.61535251e-01 2.19122469e-01 4.87778991e-01 1.00982487e+00
3.84466290e-01 2.28881463e-01 -2.83420950e-01 4.67959046e-01
5.95592320e-01 5.05278111e-01 5.73787272e-01 -9.36609924e-01
6.38479710e-01 2.80693233e-01 1.02260731e-01 -7.85670400e-01
-7.22191930e-02 -2.95110047e-02 -7.92638242e-01 5.56113422e-01
6.29830122e-01 -2.94617176e-01 -1.14519477e+00 2.02436137e+00
3.96885365e-01 5.45278072e-01 -4.47340170e-03 1.41336989e+00
6.81239069e-01 5.56231976e-01 7.27718323e-02 -1.39564186e-01
1.27797282e+00 -1.21742249e+00 -8.22441041e-01 -7.92024881e-02
3.83351564e-01 -7.94041693e-01 1.23618567e+00 2.08112702e-01
-1.24154127e+00 -6.06682837e-01 -9.02800798e-01 -1.42742723e-01
2.25388408e-01 1.97282314e-01 4.79815036e-01 3.01691860e-01
-6.66154981e-01 4.90694910e-01 -1.17626560e+00 -4.02296990e-01
1.08298540e-01 1.88139439e-01 -3.00148159e-01 -1.07606098e-01
-1.19421327e+00 7.02634275e-01 1.33634657e-01 3.67999792e-01
-5.38892627e-01 -7.13394642e-01 -9.01058674e-01 -2.76034445e-01
2.51956135e-01 -1.13787162e+00 1.33436215e+00 -1.40677023e+00
-1.92805934e+00 7.54877806e-01 -3.23320121e-01 -8.19818825e-02
1.07858717e+00 -3.42607319e-01 -1.39355913e-01 2.60028481e-01
-1.74296632e-01 9.06033576e-01 8.83257389e-01 -1.22510564e+00
-7.56287575e-01 8.30592737e-02 -1.47257419e-02 4.12583977e-01
-1.11870058e-01 -1.25138741e-02 -9.51854348e-01 -1.02189112e+00
-2.65427046e-02 -1.07616007e+00 -3.43710035e-02 5.07164121e-01
-1.18043542e-01 -4.37923940e-03 9.64307308e-01 -9.01343882e-01
1.17047405e+00 -2.00469303e+00 4.00839061e-01 1.45447418e-01
-1.40435949e-01 3.75957191e-01 -2.47129634e-01 2.14239448e-01
8.16632584e-02 -1.67871401e-01 -2.47884691e-01 -2.86810637e-01
-8.37981030e-02 2.96333134e-01 -3.39442730e-01 3.04258853e-01
1.32493570e-01 9.74456787e-01 -9.47375655e-01 -6.25444174e-01
3.47757965e-01 7.33924448e-01 -5.51316381e-01 4.19652075e-01
-1.76533043e-01 8.44320416e-01 -6.03873014e-01 4.74281073e-01
4.72006917e-01 -5.30777732e-03 7.59662874e-03 -5.03692269e-01
-2.63889637e-02 -4.93932106e-02 -1.33969617e+00 1.98247921e+00
-3.45081955e-01 6.03075743e-01 2.77535096e-02 -5.46716213e-01
4.43823993e-01 5.21859467e-01 6.35377586e-01 -6.23811901e-01
8.39083195e-02 2.37411082e-01 -1.93928760e-02 -7.36741066e-01
1.80327415e-01 -1.39525220e-01 2.21065804e-01 3.84017199e-01
-1.55998275e-01 -9.91786271e-02 1.56003937e-01 -5.41305467e-02
6.20767593e-01 6.55657768e-01 1.19327195e-01 2.47989938e-01
6.26727581e-01 -8.99186432e-02 7.08258629e-01 1.55657098e-01
-2.34215617e-01 9.06827629e-01 2.72882819e-01 -2.40756691e-01
-9.79280055e-01 -9.09466445e-01 4.32957590e-01 6.50267839e-01
3.50994587e-01 -6.67080879e-02 -7.16405869e-01 -5.87822199e-01
-2.64100462e-01 3.82907748e-01 -3.21149349e-01 -1.72636956e-01
-1.06617355e+00 -2.06123069e-01 4.69170451e-01 8.79572451e-01
7.96661198e-01 -1.22433746e+00 -9.14620996e-01 4.66629229e-02
-5.88267028e-01 -1.15779090e+00 -1.24328780e+00 -7.26814508e-01
-7.25416064e-01 -1.01361227e+00 -1.19074690e+00 -1.01345587e+00
8.60788822e-01 2.56266326e-01 4.98467147e-01 2.12959334e-01
-2.48113871e-01 5.23626506e-01 -4.81814951e-01 4.47441228e-02
-2.78421253e-01 -2.29750738e-01 -5.81346676e-02 1.66119263e-01
-3.61811593e-02 -6.26535654e-01 -8.48695159e-01 4.47650701e-01
-9.67027128e-01 4.86902952e-01 6.24904573e-01 9.66826081e-01
5.13721824e-01 -2.24031806e-01 1.96624592e-01 -9.25537944e-02
3.66415143e-01 4.06426638e-02 -4.29033279e-01 3.72203499e-01
-1.96476862e-01 1.65026575e-01 3.87938052e-01 -8.56819272e-01
-1.17419398e+00 3.72954041e-01 -8.57297406e-02 -6.67659223e-01
1.00542150e-01 2.39843294e-01 -3.79432231e-01 -1.43535897e-01
1.91387132e-01 3.73146683e-01 2.55236804e-01 -2.49985844e-01
5.43170929e-01 4.91794735e-01 8.33926976e-01 -5.36241114e-01
8.96315694e-01 5.44719279e-01 5.33368960e-02 -7.67290652e-01
-4.64552343e-01 -4.83767569e-01 -7.54965961e-01 -4.97162700e-01
1.07341766e+00 -8.57764423e-01 -7.87843764e-01 7.74251580e-01
-1.24401975e+00 -5.70145488e-01 -1.64621025e-01 8.67700100e-01
-8.67742181e-01 8.30516279e-01 -6.03711426e-01 -7.63864338e-01
-4.63053584e-01 -1.00675571e+00 9.35630023e-01 3.72271448e-01
-2.31937662e-01 -6.90342069e-01 9.07683466e-03 3.27323169e-01
1.20893747e-01 4.96198446e-01 5.60246468e-01 3.01912010e-01
-8.77112091e-01 9.41125304e-02 -1.73023969e-01 3.57051313e-01
1.46479651e-01 1.22818641e-01 -5.99413872e-01 -3.40278268e-01
-3.61628622e-01 -3.10800880e-01 8.44766855e-01 5.15480161e-01
8.12686086e-01 -2.75024772e-01 -9.31051970e-02 7.52884746e-01
1.07405615e+00 2.87956178e-01 8.69799316e-01 2.53463149e-01
8.72280419e-01 7.04653442e-01 7.60709524e-01 2.53545940e-01
4.85385180e-01 9.32291508e-01 -1.20349173e-02 -1.16673961e-01
-6.50800645e-01 -5.58893383e-01 7.47705936e-01 8.40020955e-01
-6.20875418e-01 -2.14644849e-01 -6.10315204e-01 5.37295461e-01
-2.20775723e+00 -1.10339022e+00 -7.69047514e-02 2.19754338e+00
8.02831829e-01 -1.65084630e-01 1.56302616e-01 -4.25085649e-02
6.99597239e-01 5.49100665e-03 -6.52374268e-01 2.54800647e-01
-4.05605063e-02 -3.87891382e-02 2.45529607e-01 6.09299958e-01
-9.50316906e-01 1.08591080e+00 5.08250904e+00 5.61198771e-01
-1.37860560e+00 -1.62029162e-01 4.98646535e-02 -2.02934131e-01
-1.62785575e-01 8.32016692e-02 -5.18837512e-01 4.65647578e-01
3.00798357e-01 -1.50151640e-01 3.44551444e-01 5.64208388e-01
5.80288112e-01 2.06928462e-01 -1.00512934e+00 1.08888078e+00
8.35704505e-02 -9.17351902e-01 1.70968354e-01 4.92502600e-02
7.21924841e-01 -4.00338203e-01 -9.29555148e-02 5.48657030e-02
-1.61535181e-02 -8.27734649e-01 1.04826748e+00 8.76084208e-01
9.75234687e-01 -6.35262847e-01 4.91051555e-01 3.59026700e-01
-1.49848282e+00 1.29000977e-01 8.43185112e-02 1.32522553e-01
7.28518665e-01 3.52182724e-02 -2.82819241e-01 5.56381047e-01
4.23659712e-01 6.83926105e-01 -3.70072350e-02 1.17298567e+00
-5.86294353e-01 5.33998430e-01 -2.95291126e-01 1.33630022e-01
2.67644167e-01 -1.95394844e-01 6.35719419e-01 1.01839745e+00
5.58648944e-01 2.64832646e-01 1.72674224e-01 7.94617355e-01
2.46736556e-01 3.69147062e-02 -2.49747634e-01 8.77146423e-02
1.76332340e-01 8.28269064e-01 -4.26879883e-01 -1.71763837e-01
-4.25335199e-01 1.47708404e+00 -1.40874803e-01 5.39286137e-01
-8.63186359e-01 -3.26332033e-01 6.22780919e-01 2.15740502e-01
3.95578384e-01 -3.98238033e-01 7.67325759e-02 -1.51971579e+00
4.41228509e-01 -8.24788392e-01 2.75882781e-01 -9.35032904e-01
-1.12619150e+00 2.61049390e-01 4.62136278e-03 -1.65593147e+00
-6.24139071e-01 -4.79684442e-01 -7.40264893e-01 8.42197299e-01
-1.64814377e+00 -1.64145947e+00 -6.85489714e-01 7.78516650e-01
5.29790223e-01 5.97182997e-02 3.86601448e-01 3.30028594e-01
-2.70687908e-01 7.23820329e-01 -3.78908008e-01 4.22649771e-01
9.19363499e-01 -8.24672461e-01 3.66184622e-01 1.03426123e+00
-5.14958985e-02 4.60758984e-01 4.81588155e-01 -6.83637261e-01
-1.18629444e+00 -9.07623112e-01 8.22964191e-01 -2.11247087e-01
3.89753044e-01 -4.68047336e-02 -8.47990632e-01 4.92657095e-01
-1.26268072e-02 2.98917703e-02 9.16156769e-02 -6.67022586e-01
-1.30920485e-01 2.88794544e-02 -8.80286038e-01 9.43411648e-01
1.39440882e+00 -3.38914812e-01 -6.07022703e-01 -5.25990874e-02
5.75199604e-01 -6.91866040e-01 -6.43650115e-01 3.64626855e-01
1.06105101e+00 -6.21583521e-01 9.23199296e-01 -7.02037156e-01
5.62489688e-01 -5.78315854e-01 1.50146559e-01 -1.02581429e+00
-5.01168519e-02 -9.15430963e-01 -3.07283312e-01 1.26805305e+00
-1.35034360e-02 -3.15724403e-01 8.44953775e-01 7.13852048e-01
-7.71307386e-03 -6.66330159e-01 -9.91126060e-01 -8.09953332e-01
3.15466220e-03 -2.77923971e-01 3.50297838e-01 8.08362722e-01
-2.78290641e-02 1.37088344e-01 -9.10430253e-01 2.77776942e-02
6.03830636e-01 1.44379318e-01 1.10422146e+00 -7.91334569e-01
-3.19837034e-01 -4.58554924e-01 -3.79173130e-01 -1.52699053e+00
1.86898232e-01 -4.82116461e-01 1.65202439e-01 -1.64089429e+00
1.27962828e-01 -8.73010010e-02 3.70502099e-02 4.57196474e-01
-2.65362263e-01 -7.04291984e-02 5.17601669e-01 3.85732204e-01
-7.21931234e-02 7.80059338e-01 1.91620803e+00 1.14467919e-01
-4.27861035e-01 1.57642290e-01 -1.52075469e-01 8.38798285e-01
6.32992387e-01 -1.33832365e-01 -6.21083856e-01 -4.70901847e-01
-1.83313236e-01 3.18441391e-01 7.94230521e-01 -8.89579654e-01
3.91049296e-01 -5.36493957e-01 1.32064223e-01 -5.33704162e-01
3.48373622e-01 -7.32917011e-01 7.41665512e-02 6.62570536e-01
-1.39368832e-01 -8.85131359e-02 -5.44888861e-02 5.60338557e-01
-2.84512877e-01 -5.09857126e-02 9.20419037e-01 -6.84131309e-02
-1.00826180e+00 4.57354605e-01 -2.38077894e-01 5.31884283e-02
1.00017154e+00 -4.72072631e-01 -8.00611302e-02 -7.60608256e-01
-4.19400632e-01 3.58103067e-01 4.94971544e-01 5.67422271e-01
7.31787026e-01 -1.56628454e+00 -7.22867668e-01 1.17002219e-01
2.68204026e-02 7.86429569e-02 5.49924791e-01 8.65391016e-01
-6.30452871e-01 2.24289864e-01 -4.70286518e-01 -4.46221143e-01
-1.46518850e+00 2.73220122e-01 4.02495772e-01 -6.91716075e-02
-1.09626317e+00 5.56989372e-01 2.54923403e-01 -2.00438902e-01
4.02406722e-01 -5.71589172e-01 1.85589828e-02 -2.61843860e-01
5.30232668e-01 2.43001297e-01 -5.79282522e-01 -8.33990216e-01
-2.16747537e-01 1.17691875e+00 2.13445470e-01 -4.72323060e-01
1.01217997e+00 -1.37495264e-01 2.15973184e-01 1.46110207e-01
9.69313204e-01 -4.05915156e-02 -1.67995679e+00 -7.69583359e-02
-4.97413039e-01 -5.56991041e-01 -1.20356150e-01 -6.97205365e-01
-1.08944571e+00 9.13978159e-01 6.00063920e-01 -7.43148983e-01
1.25499582e+00 -3.40217680e-01 1.40666962e+00 9.13848057e-02
2.65299290e-01 -1.29317498e+00 3.17451417e-01 4.05999482e-01
9.97658074e-01 -9.44906533e-01 -1.23464167e-01 -4.76111114e-01
-9.57392275e-01 1.15256870e+00 8.29370379e-01 1.80187970e-01
3.51605773e-01 -1.36277257e-02 2.64666647e-01 1.20457977e-01
-5.68381131e-01 -3.71463090e-01 6.96413994e-01 6.37591004e-01
2.72416323e-01 -2.89875507e-01 -5.40971518e-01 5.39113998e-01
-7.84534290e-02 4.50017571e-01 4.17165250e-01 1.01649535e+00
-2.32812494e-01 -1.19628537e+00 -3.60040247e-01 -2.47772500e-01
-4.56906930e-02 1.89479500e-01 -2.91439116e-01 8.93066287e-01
2.57078499e-01 5.93795240e-01 -1.63692266e-01 -2.37579942e-01
5.74516058e-01 3.57056372e-02 7.42348611e-01 -2.12191954e-01
-2.17074260e-01 1.98179260e-01 -5.94645888e-02 -5.87027967e-01
-7.03015506e-01 -6.01909220e-01 -1.67191195e+00 -8.94355327e-02
-1.13561109e-01 -2.22848609e-01 3.63993347e-01 8.67605329e-01
1.56403318e-01 4.39645112e-01 3.16280663e-01 -1.01905107e+00
-5.35670877e-01 -8.52430761e-01 -3.37701619e-01 7.02144682e-01
3.16273868e-01 -6.36702240e-01 -2.09182918e-01 5.85061729e-01] | [10.936177253723145, -0.8680644035339355] |
8cdad5f2-b3a0-414e-8525-f7a94900b1d3 | gt-gan-general-purpose-time-series-synthesis | 2210.02040 | null | https://arxiv.org/abs/2210.02040v3 | https://arxiv.org/pdf/2210.02040v3.pdf | GT-GAN: General Purpose Time Series Synthesis with Generative Adversarial Networks | Time series synthesis is an important research topic in the field of deep learning, which can be used for data augmentation. Time series data types can be broadly classified into regular or irregular. However, there are no existing generative models that show good performance for both types without any model changes. Therefore, we present a general purpose model capable of synthesizing regular and irregular time series data. To our knowledge, we are the first designing a general purpose time series synthesis model, which is one of the most challenging settings for time series synthesis. To this end, we design a generative adversarial network-based method, where many related techniques are carefully integrated into a single framework, ranging from neural ordinary/controlled differential equations to continuous time-flow processes. Our method outperforms all existing methods. | ['Noseong Park', 'Seunghyeon Cho', 'Haryong Song', 'Jeonghak Kim', 'Jinsung Jeon'] | 2022-10-05 | null | null | null | null | ['irregular-time-series'] | ['time-series'] | [ 1.38634536e-02 -2.96863705e-01 6.83634058e-02 1.42735660e-01
-3.54099810e-01 -5.75774074e-01 7.44826555e-01 -4.77897942e-01
8.83352086e-02 6.89708829e-01 -1.32791653e-01 -5.64758182e-01
9.99901071e-02 -1.06682444e+00 -6.74251437e-01 -8.38523448e-01
-8.21009874e-02 4.21083719e-02 -4.93533611e-02 -6.49180293e-01
-3.45376655e-02 6.47928357e-01 -1.18086874e+00 -1.79653332e-01
7.90018201e-01 8.97009432e-01 -4.69115466e-01 6.20071352e-01
2.31395755e-02 6.62680030e-01 -8.64699364e-01 -1.61949098e-01
3.68897051e-01 -7.68254817e-01 -3.25647473e-01 -1.61596000e-01
-1.44657299e-01 -4.73084509e-01 -8.69074702e-01 9.09149170e-01
6.92537427e-01 3.63705337e-01 6.39909029e-01 -1.76521695e+00
-9.89821732e-01 7.03472793e-01 -2.90136158e-01 3.28453392e-01
2.92162802e-02 3.11281651e-01 4.48633730e-01 -4.88599002e-01
2.91151106e-01 1.09629786e+00 1.00027716e+00 6.48676038e-01
-1.08370495e+00 -8.28165472e-01 2.66106039e-01 4.25439999e-02
-9.52960491e-01 -1.28542051e-01 1.49200892e+00 -3.62822801e-01
7.00362682e-01 3.41053784e-01 8.77239645e-01 1.35017979e+00
3.26524645e-01 8.26833665e-01 1.27501404e+00 -1.24885999e-01
3.80363762e-01 -5.32085001e-01 -1.54556438e-01 2.26251036e-01
-2.48059779e-01 2.68855840e-01 6.37514144e-02 -7.37377405e-02
1.26422560e+00 5.31531990e-01 -1.35883838e-01 6.95819631e-02
-1.46054924e+00 8.68999183e-01 2.82664508e-01 6.43179238e-01
-2.92867273e-01 4.57797915e-01 5.77230990e-01 8.13971996e-01
6.35553837e-01 3.62416238e-01 -3.17993969e-01 -2.06441298e-01
-7.82256663e-01 5.37337959e-01 6.72814965e-01 9.80863452e-01
3.23940426e-01 9.19053376e-01 -3.06049973e-01 4.55318749e-01
-1.15066782e-01 5.76509178e-01 9.74221230e-01 -5.80947042e-01
2.21112758e-01 3.29536289e-01 -1.91453204e-01 -1.19204497e+00
-3.65342706e-01 -3.87124121e-01 -1.34736204e+00 -7.07648695e-02
1.54712573e-01 -6.03287280e-01 -1.03459311e+00 1.90783477e+00
1.78665876e-01 9.33277786e-01 9.63600129e-02 6.03240907e-01
8.30673397e-01 1.03571117e+00 -2.30158001e-01 -6.11324489e-01
8.91981244e-01 -8.49665225e-01 -1.12307596e+00 2.22611204e-01
3.62211689e-02 -7.41982996e-01 7.51846492e-01 1.52964294e-01
-1.14137173e+00 -7.62585580e-01 -1.03126192e+00 3.71587053e-02
-5.92106640e-01 -2.77521342e-01 9.58309829e-01 4.29051578e-01
-9.75177646e-01 8.68203759e-01 -1.23027980e+00 -9.31092501e-02
1.55756786e-01 1.13919549e-01 -2.48661055e-03 4.64425266e-01
-1.55093312e+00 5.67258060e-01 8.30650851e-02 2.00033307e-01
-9.78092492e-01 -8.15189600e-01 -8.41514766e-01 -2.45664790e-01
1.00243650e-01 -6.60662949e-01 1.47260988e+00 -8.68743241e-01
-1.74252617e+00 4.14963931e-01 7.87851587e-03 -7.64555216e-01
5.72754085e-01 7.61699444e-03 -9.51702237e-01 -2.58088589e-01
-8.37841779e-02 1.12651773e-01 1.09909284e+00 -8.15023661e-01
-5.91160096e-02 -1.07302114e-01 2.70218179e-02 -2.99272686e-01
-3.38823825e-01 -1.01956408e-02 1.25825182e-01 -1.56870365e+00
8.92788470e-02 -8.33408058e-01 -4.48898405e-01 -1.54173121e-01
-4.43100065e-01 -2.81479716e-01 1.17114353e+00 -5.86872041e-01
1.40407193e+00 -1.93136072e+00 5.00084870e-02 -2.99265590e-02
1.60731897e-01 3.06186527e-01 -7.43005145e-03 7.76767552e-01
-4.48365718e-01 2.43598744e-01 -3.56989503e-01 -2.36988023e-01
1.54425055e-01 2.33734518e-01 -1.06810093e+00 4.35143113e-01
3.52828294e-01 1.24837339e+00 -8.31753075e-01 -3.22242193e-02
2.83282310e-01 5.12442291e-01 -3.09991062e-01 2.58380264e-01
-3.36040556e-01 8.31066012e-01 -4.95995045e-01 5.21864653e-01
5.57207286e-01 9.27419681e-03 -4.39388573e-01 -1.18611090e-01
-3.29860091e-01 1.55923799e-01 -1.00775015e+00 1.55176902e+00
-4.63980645e-01 4.54842001e-01 -4.87079829e-01 -1.24979126e+00
1.18718743e+00 6.82868779e-01 8.79853070e-01 -4.44252610e-01
3.68844360e-01 2.81305462e-01 9.84585807e-02 -4.64805812e-01
2.45218158e-01 -3.69467378e-01 -1.95998833e-01 5.77400684e-01
-2.02131256e-01 -4.91065770e-01 7.36509860e-02 -3.57562989e-01
9.91155982e-01 3.98927666e-02 1.79345444e-01 2.00314328e-01
3.18878621e-01 -1.95000961e-01 5.84266126e-01 4.74255741e-01
-9.64697897e-02 8.07492495e-01 4.73605037e-01 -7.33583093e-01
-1.42858076e+00 -8.87601197e-01 3.67844105e-02 7.01777935e-01
-1.06122375e-01 -9.42227170e-02 -5.73285758e-01 -2.40645632e-01
-1.41834110e-01 4.83933806e-01 -7.44467437e-01 -4.09631193e-01
-9.99115586e-01 -8.36153865e-01 8.09417963e-01 8.64737093e-01
5.29683709e-01 -1.53504455e+00 1.25702685e-02 5.74492633e-01
-2.10968070e-02 -7.91518569e-01 -7.51135528e-01 1.28207043e-01
-1.14524472e+00 -6.25399768e-01 -1.03919840e+00 -9.76004779e-01
3.81240666e-01 -9.07852128e-03 8.79742563e-01 -1.37445509e-01
-2.29529850e-02 8.63684267e-02 -3.61778110e-01 -7.40014315e-01
-6.24372184e-01 1.84480380e-02 1.25823915e-01 1.64982781e-01
1.42247468e-01 -1.22088313e+00 -5.79875290e-01 2.70139784e-01
-1.30725420e+00 -1.41246635e-02 1.83236122e-01 7.25012422e-01
7.77955592e-01 2.97439307e-01 9.31418180e-01 -6.70308650e-01
1.02661622e+00 -6.01118445e-01 -4.43550557e-01 -1.33239955e-01
-2.73112714e-01 -1.71510249e-01 1.29798222e+00 -1.09135425e+00
-6.18486822e-01 -2.31461942e-01 -5.45330703e-01 -7.89692342e-01
-1.33847743e-01 5.70344090e-01 3.80620547e-02 -9.20019373e-02
4.63641047e-01 6.15892887e-01 4.32253703e-02 -3.82044762e-01
3.26891690e-01 3.85240555e-01 6.72656119e-01 -5.21851718e-01
1.10199380e+00 6.28857791e-01 1.02616236e-01 -5.62849641e-01
-5.53707302e-01 2.10717060e-02 -4.26723003e-01 -9.37374458e-02
5.19602656e-01 -5.79951346e-01 -4.62532371e-01 1.02375495e+00
-1.09433722e+00 -5.45546770e-01 -6.03931010e-01 2.97820151e-01
-9.30559397e-01 2.33968586e-01 -8.72828543e-01 -6.88983977e-01
-4.97514278e-01 -8.98843229e-01 1.09207034e+00 2.99128294e-01
1.74835585e-02 -1.21354496e+00 3.84118110e-01 -4.41441864e-01
8.53510737e-01 1.03930867e+00 5.28015673e-01 -6.89158142e-01
-3.42315465e-01 -4.10321653e-01 3.59160423e-01 4.33345884e-01
2.49549717e-01 3.04912418e-01 -7.42885292e-01 -2.83588469e-01
5.82101762e-01 3.31878103e-02 5.18286526e-01 3.79357964e-01
1.53806293e+00 -3.79645288e-01 -2.04029247e-01 9.66470361e-01
1.15065646e+00 8.24546814e-01 1.04118705e+00 2.39026949e-01
7.35990763e-01 -6.99885339e-02 1.93157434e-01 4.39188302e-01
1.87740788e-01 3.50539476e-01 2.67093867e-01 -2.65338004e-01
2.04925358e-01 -3.40366244e-01 3.82073075e-01 1.15488708e+00
-3.06275845e-01 -3.09288651e-01 -7.77087986e-01 5.16251147e-01
-1.67393374e+00 -1.38362205e+00 -1.58768907e-01 1.70432854e+00
8.36505115e-01 5.01173735e-02 2.92716354e-01 6.89299703e-01
7.34108925e-01 3.90631258e-01 -8.64432633e-01 -2.28158116e-01
-4.98229265e-02 5.96475840e-01 8.71331841e-02 -8.78256783e-02
-1.38230479e+00 7.36415982e-01 6.97942877e+00 7.02632308e-01
-1.65821171e+00 -2.08807230e-01 4.55044091e-01 3.86533946e-01
-4.05059874e-01 -3.07617337e-01 -3.19702178e-01 7.58825719e-01
9.42400217e-01 -6.67998016e-01 4.73780692e-01 8.72936070e-01
2.65482157e-01 9.64338243e-01 -9.91060197e-01 1.02086663e+00
-1.08259447e-01 -1.18736696e+00 1.37798423e-02 -1.35939747e-01
9.75161016e-01 -3.81040186e-01 3.51968169e-01 5.84586918e-01
2.11038440e-01 -1.14113104e+00 4.38617468e-01 7.17785478e-01
7.34270632e-01 -7.16343045e-01 5.55397332e-01 2.45728374e-01
-1.32901192e+00 1.64824605e-01 -1.33801578e-02 -2.23209009e-01
3.86676550e-01 4.70726937e-01 -1.43928647e-01 4.80381995e-01
3.85314345e-01 1.10666597e+00 -9.36308876e-02 1.06569755e+00
4.67739664e-02 8.48867774e-01 -2.30637044e-01 -4.94263247e-02
1.00071818e-01 -1.76387474e-01 6.62257969e-01 7.73292542e-01
7.38643825e-01 2.13520765e-01 2.21205056e-01 9.33418691e-01
-2.92130075e-02 -6.56769127e-02 -8.47687423e-01 -3.88638467e-01
3.70409071e-01 9.79772985e-01 -6.53091490e-01 -2.62022614e-01
-4.12690639e-01 8.76433551e-01 -2.40063742e-01 4.08255249e-01
-1.29135418e+00 -6.59727514e-01 5.77672124e-01 -9.78929084e-03
-4.77390422e-04 -4.84110802e-01 -1.39310107e-01 -1.51360679e+00
7.98098370e-02 -1.07993317e+00 3.08537453e-01 -8.27525318e-01
-1.72606146e+00 7.83580840e-01 -1.25068828e-01 -1.77482700e+00
-4.29957747e-01 -4.58930194e-01 -1.01530862e+00 9.13136184e-01
-1.28584242e+00 -1.13382745e+00 -3.65598381e-01 9.04447317e-01
5.25884509e-01 -3.39174569e-01 6.87575936e-01 3.19702297e-01
-5.14995992e-01 5.51229119e-01 1.97572470e-01 3.06472600e-01
3.91666561e-01 -1.22715712e+00 1.04005575e+00 1.11349046e+00
-1.91496134e-01 8.68825853e-01 6.43685758e-01 -5.89874566e-01
-1.51928079e+00 -1.47373223e+00 3.61561954e-01 -1.48672119e-01
1.04228222e+00 -2.07882375e-01 -1.26217830e+00 8.90998840e-01
2.52944708e-01 3.03360429e-02 4.73502964e-01 -5.19845128e-01
-2.07300950e-02 -9.83974859e-02 -1.05395949e+00 8.93568158e-01
1.01665485e+00 -3.98696840e-01 -6.31746173e-01 4.03699130e-01
1.15419590e+00 -7.03194916e-01 -1.15643442e+00 6.19883120e-01
1.80178687e-01 -6.23960078e-01 1.17655849e+00 -6.49184883e-01
6.17657185e-01 -2.80324370e-01 1.76395327e-01 -1.54495168e+00
-2.01386482e-01 -1.35567653e+00 -5.72228432e-01 1.31243491e+00
-5.33265956e-02 -1.01533413e+00 5.06207526e-01 1.85926422e-01
-4.38163429e-01 -6.84535205e-01 -5.90054333e-01 -9.52277720e-01
5.36630213e-01 -4.69872594e-01 1.15229523e+00 1.26280785e+00
-3.86210322e-01 2.70092115e-02 -5.93543291e-01 -2.23029759e-02
2.76246428e-01 3.99778157e-01 9.12159622e-01 -1.10033214e+00
-1.97959021e-01 -7.42284358e-01 -3.14023674e-01 -9.35902596e-01
3.61054949e-02 -7.03841448e-01 1.48048978e-02 -1.27304661e+00
-5.83540797e-01 -4.65114594e-01 -1.97757855e-01 2.13751078e-01
-1.79078341e-01 4.07125168e-02 -1.09094828e-01 2.45221943e-01
1.11891769e-01 8.52920711e-01 1.65892947e+00 4.49059568e-02
-2.73180962e-01 3.30519527e-01 -6.56310380e-01 6.14308059e-01
1.30653596e+00 -1.63806543e-01 -6.22108877e-01 -1.79822817e-01
-3.98028418e-02 4.01996344e-01 4.69139189e-01 -1.18253076e+00
1.09068871e-01 -5.68005323e-01 8.34345967e-02 -6.96108699e-01
1.21653974e-01 -8.26690018e-01 5.73121965e-01 4.82720584e-01
-2.92327136e-01 5.97831547e-01 2.81505197e-01 4.27269459e-01
-4.30615634e-01 2.37575069e-01 6.80404663e-01 -6.45783544e-02
-4.52060997e-01 9.99167621e-01 -2.65671402e-01 1.33176774e-01
1.04146361e+00 6.16219230e-02 -3.81335348e-01 -6.46526515e-01
-7.62908638e-01 -1.21568315e-01 1.04542747e-01 6.37846529e-01
4.76948947e-01 -1.85700417e+00 -6.09357417e-01 4.22546536e-01
-2.63935983e-01 6.35240600e-02 3.14301372e-01 7.02049613e-01
-5.31984448e-01 2.46807635e-01 -3.44183058e-01 -3.92284870e-01
-5.54651916e-01 1.16925931e+00 5.80289602e-01 -2.35372201e-01
-8.33792984e-01 2.11157680e-01 7.44696483e-02 -3.67726237e-01
6.76993132e-02 -6.36732638e-01 -2.00164661e-01 -2.06259847e-01
5.97688258e-01 3.63209486e-01 -8.07345435e-02 -2.98156172e-01
8.00821334e-02 6.69247091e-01 3.20134431e-01 -7.05721527e-02
1.52471530e+00 3.20111603e-01 -2.09574327e-01 7.69753158e-01
1.28592896e+00 -2.32427254e-01 -1.01772058e+00 -1.52515396e-01
-6.19908631e-01 1.80131663e-02 -3.79225403e-01 -2.22748980e-01
-1.39641929e+00 8.52748215e-01 3.01066518e-01 1.07178104e+00
1.57529747e+00 -4.71204221e-01 1.30593848e+00 1.59046814e-01
3.36510479e-01 -6.16152048e-01 1.73366547e-01 6.53190196e-01
1.16472805e+00 -8.81207585e-01 -5.20255029e-01 -2.34645620e-01
-3.29360217e-01 1.08216560e+00 6.59824252e-01 -8.18677187e-01
7.66025424e-01 5.44909835e-01 -1.35615602e-01 9.54202786e-02
-6.74876511e-01 -1.40207216e-01 2.56134808e-01 6.76715732e-01
5.56086302e-01 -4.82833497e-02 -4.77458954e-01 6.43113554e-01
-5.35696685e-01 1.95650265e-01 4.06163871e-01 7.83164501e-01
6.21629171e-02 -1.04681861e+00 -4.14863348e-01 3.82678419e-01
-5.40444970e-01 4.43872735e-02 8.91870912e-03 9.32572484e-01
-7.01252967e-02 6.28373086e-01 1.07046895e-01 -4.62629944e-01
4.36568141e-01 5.50918095e-02 1.91952005e-01 -1.82611227e-01
-7.14954674e-01 1.96705848e-01 -4.21244591e-01 -1.98381662e-01
-5.28274596e-01 -6.25174999e-01 -1.04228342e+00 -6.33842766e-01
5.38446493e-02 -1.17338449e-01 3.15700531e-01 7.26957858e-01
1.63023680e-01 9.73821402e-01 8.94340217e-01 -8.17142904e-01
-3.37713182e-01 -8.66809070e-01 -3.38065803e-01 5.53895295e-01
6.36273563e-01 -4.72159266e-01 -3.79537106e-01 2.48593569e-01] | [7.021917819976807, 3.081449270248413] |
003a1f6e-db86-4a35-8362-efaf1cfdd30c | pepe-personalized-post-editing-model | 2209.10139 | null | https://arxiv.org/abs/2209.10139v2 | https://arxiv.org/pdf/2209.10139v2.pdf | PePe: Personalized Post-editing Model utilizing User-generated Post-edits | Incorporating personal preference is crucial in advanced machine translation tasks. Despite the recent advancement of machine translation, it remains a demanding task to properly reflect personal style. In this paper, we introduce a personalized automatic post-editing framework to address this challenge, which effectively generates sentences considering distinct personal behaviors. To build this framework, we first collect post-editing data that connotes the user preference from a live machine translation system. Specifically, real-world users enter source sentences for translation and edit the machine-translated outputs according to the user's preferred style. We then propose a model that combines a discriminator module and user-specific parameters on the APE framework. Experimental results show that the proposed method outperforms other baseline models on four different metrics (i.e., BLEU, TER, YiSi-1, and human evaluation). | ['Jaegul Choo', 'Cheonbok Park', 'Yunwon Tae', 'Taehee Kim', 'Jihyeon Lee'] | 2022-09-21 | null | null | null | null | ['automatic-post-editing', 'automatic-post-editing'] | ['computer-vision', 'natural-language-processing'] | [ 3.78574610e-01 -4.01772171e-01 -3.70497078e-01 -6.71924591e-01
-1.14392030e+00 -7.87492216e-01 7.97916830e-01 -2.77320266e-01
-4.24768478e-01 8.72892678e-01 2.87632734e-01 -4.53989923e-01
3.06859016e-01 -2.51026362e-01 -3.45152259e-01 -2.78909862e-01
8.18514287e-01 7.96679914e-01 -3.28198403e-01 -4.91528809e-01
4.50952113e-01 -4.47537862e-02 -8.23188901e-01 4.95258033e-01
1.63305521e+00 6.64831161e-01 3.38567168e-01 6.17388308e-01
1.07217260e-01 5.85578196e-02 -5.89608371e-01 -1.13251996e+00
1.28471151e-01 -8.01844120e-01 -6.43248260e-01 -9.73911881e-02
2.32669428e-01 -1.78888366e-01 -4.64693159e-02 1.05675888e+00
8.09814990e-01 3.03336661e-02 7.85964310e-01 -8.36205482e-01
-1.14902747e+00 8.29567432e-01 -1.98962867e-01 2.54186928e-01
4.25084770e-01 2.45766729e-01 9.92672265e-01 -1.21050429e+00
3.97957176e-01 1.08616579e+00 3.71878028e-01 7.10581720e-01
-1.30201650e+00 -3.90381575e-01 1.17196646e-02 6.18621670e-02
-1.20919013e+00 -6.96887434e-01 6.22020721e-01 -1.79398775e-01
5.91210425e-01 5.84851027e-01 1.40161857e-01 1.86293268e+00
4.81854588e-01 8.06769133e-01 1.39736903e+00 -4.55772161e-01
1.03257872e-01 5.59681654e-01 -8.26245472e-02 1.04136132e-01
-1.14360787e-01 -4.01764154e-01 -5.84974349e-01 -1.71834260e-01
4.19327140e-01 -1.23350784e-01 -1.91093266e-01 6.09948933e-02
-1.61730361e+00 4.40756887e-01 -2.07294896e-01 3.06613445e-01
-3.83700997e-01 -4.30603653e-01 3.54709297e-01 6.51981711e-01
7.12757707e-01 5.73926806e-01 -5.92341661e-01 -6.59180462e-01
-8.79670024e-01 2.74505988e-02 7.89629340e-01 1.31778896e+00
4.83328432e-01 -6.17180049e-01 -9.50342655e-01 1.29646730e+00
9.68695655e-02 8.09456468e-01 7.79328108e-01 -6.33056939e-01
1.03221893e+00 3.44957173e-01 4.45522785e-01 -7.42240191e-01
1.42696202e-01 -7.04254866e-01 -8.56521726e-01 -7.34330475e-01
5.00279367e-02 -3.16725224e-01 -5.99044979e-01 1.70525551e+00
-1.06529675e-01 -2.24406183e-01 -5.00419773e-02 9.54082251e-01
2.48176605e-01 6.96841657e-01 -6.47399575e-02 -5.80790401e-01
1.12959802e+00 -1.38383806e+00 -8.54313493e-01 -3.23595524e-01
4.44356173e-01 -1.17978513e+00 1.66283858e+00 2.22996503e-01
-1.08227408e+00 -5.28305888e-01 -7.41011679e-01 5.30855206e-04
3.80320735e-02 7.01572776e-01 3.44661385e-01 8.15675259e-01
-9.40648615e-01 7.84606397e-01 -6.14170909e-01 -6.50754690e-01
5.42155048e-03 2.18161479e-01 -3.51675376e-02 1.90370634e-01
-1.29116201e+00 9.33172822e-01 2.48570666e-02 1.24730147e-01
-4.72356051e-01 -4.63302255e-01 -3.00517410e-01 -3.42719108e-02
1.78998709e-02 -1.11146665e+00 1.65343475e+00 -1.09362912e+00
-2.29611039e+00 7.83461332e-01 -4.26725715e-01 -3.10054962e-02
1.03593874e+00 -5.47505736e-01 -5.32139003e-01 -5.24903238e-01
1.17506333e-01 1.31665409e-01 7.72639275e-01 -9.31451380e-01
-6.00526690e-01 -2.00760767e-01 -5.84806129e-02 7.04692841e-01
-7.00824201e-01 2.44654328e-01 -8.91956031e-01 -8.46541643e-01
-2.50866443e-01 -1.22154355e+00 -1.44676298e-01 -6.52705610e-01
-5.58659315e-01 -3.64755481e-01 1.64956644e-01 -8.09505999e-01
1.65468907e+00 -2.01187372e+00 5.13813794e-01 1.06209114e-01
-2.81507492e-01 3.39275926e-01 -3.24839145e-01 6.04695797e-01
5.36940455e-01 1.55724436e-01 -3.04928422e-01 -8.56767654e-01
2.13695273e-01 7.28668198e-02 -2.19624668e-01 1.46272564e-02
-1.01229109e-01 9.43686962e-01 -1.03864849e+00 -5.24414539e-01
-1.36056885e-01 3.51146013e-01 -5.35995603e-01 5.07288456e-01
-1.14655025e-01 6.65897012e-01 -5.80664814e-01 4.52347606e-01
5.04610479e-01 -1.65796191e-01 3.49050939e-01 5.90642765e-02
1.24165295e-02 5.68251371e-01 -6.36302888e-01 2.00476933e+00
-7.07627356e-01 3.23386788e-01 -2.53118128e-01 -3.67634207e-01
8.86340141e-01 3.22294563e-01 -3.65109965e-02 -7.51063943e-01
1.96121663e-01 5.34095943e-01 -1.92033574e-01 -3.84632915e-01
8.36337507e-01 1.72672361e-01 -1.93385199e-01 8.09957206e-01
-1.42228872e-01 1.85226426e-01 2.73924440e-01 1.94314703e-01
6.49194896e-01 2.85702258e-01 1.53068230e-01 -3.65495682e-01
7.22719312e-01 -3.56031507e-01 5.98386526e-01 7.37034738e-01
-3.67381126e-01 7.98035920e-01 1.36333793e-01 -1.03244036e-01
-9.58029687e-01 -1.01628077e+00 1.20414905e-01 1.44155419e+00
6.42452464e-02 -5.28259635e-01 -1.14034343e+00 -8.33773613e-01
-3.95079762e-01 8.71866643e-01 -3.35738450e-01 -1.70115054e-01
-6.75219178e-01 -8.88270080e-01 3.39460015e-01 2.25047946e-01
4.87074643e-01 -8.05313408e-01 5.67617938e-02 2.12195948e-01
-1.01447523e+00 -1.05553174e+00 -1.29607832e+00 -4.83060271e-01
-8.26143801e-01 -1.77718788e-01 -8.90794575e-01 -7.16249883e-01
5.69135427e-01 3.15920025e-01 1.31791461e+00 -1.16589822e-01
3.94567043e-01 -6.78361729e-02 -5.29331565e-01 -1.96776897e-01
-6.05561674e-01 8.95008326e-01 4.61745888e-01 2.75077254e-01
4.68785793e-01 -7.30277419e-01 -7.95283020e-01 6.23946011e-01
-4.57909226e-01 4.41731840e-01 9.18484688e-01 7.78350949e-01
4.61589068e-01 -7.03479111e-01 4.41896647e-01 -9.84830320e-01
1.32241333e+00 -3.88745904e-01 1.10725135e-01 7.26954162e-01
-8.98752868e-01 -5.94492815e-02 8.78944993e-01 -4.59895164e-01
-1.19412196e+00 -4.10978466e-01 3.01021873e-03 5.07664979e-02
1.30171284e-01 4.15063769e-01 -4.17867303e-01 3.55879575e-01
6.59165144e-01 5.61179996e-01 -3.20337832e-01 -6.68022037e-01
3.85007858e-01 1.37183046e+00 4.99708861e-01 -8.91779244e-01
7.91408896e-01 -2.08425716e-01 -7.72205591e-01 -3.63352090e-01
-7.63612211e-01 -1.06729224e-01 -6.64624929e-01 7.72091597e-02
4.75972384e-01 -6.98179424e-01 -3.33674490e-01 3.62119228e-01
-1.33353221e+00 -2.43385941e-01 4.08549279e-01 4.47798550e-01
-6.97563112e-01 5.22659540e-01 -7.74645448e-01 -6.44233465e-01
-8.74697328e-01 -1.33072364e+00 1.02812040e+00 1.43823043e-01
-6.58482671e-01 -8.25932264e-01 2.93807715e-01 6.96440399e-01
6.91132009e-01 -3.40273052e-01 7.14750051e-01 -5.68989336e-01
-2.72720188e-01 8.47645253e-02 -1.38186783e-01 3.09248418e-01
4.05721456e-01 -1.45420292e-02 -7.19085991e-01 -3.57220531e-01
-8.92959833e-02 -7.81838447e-02 3.24754149e-01 -2.80017436e-01
1.11433876e+00 -5.53837955e-01 -2.07282722e-01 5.98700523e-01
8.16884160e-01 -8.79407376e-02 5.91155469e-01 3.57268929e-01
5.07490277e-01 4.01072770e-01 6.10894263e-01 3.39520574e-01
5.69588780e-01 1.07740092e+00 -3.97107899e-01 2.33008057e-01
8.71333033e-02 -5.47639310e-01 8.11682522e-01 1.44351935e+00
-3.21693838e-01 -5.33627927e-01 -5.82564056e-01 2.51369804e-01
-1.99594736e+00 -8.17485631e-01 7.55431876e-02 2.18338633e+00
1.30204797e+00 6.21013418e-02 4.55827918e-03 -3.82889599e-01
8.27478170e-01 -4.79841456e-02 -4.80565101e-01 -7.63879001e-01
-8.63448381e-02 8.67907181e-02 2.95821011e-01 5.77563107e-01
-7.66169965e-01 1.17863643e+00 5.93432236e+00 7.94426918e-01
-1.06909049e+00 2.50935346e-01 7.27835655e-01 -8.35474804e-02
-5.22059858e-01 -1.82676092e-01 -6.91533923e-01 1.06334412e+00
1.14811623e+00 -5.40160656e-01 8.42528403e-01 5.22593200e-01
4.85541224e-01 4.93665963e-01 -1.46143556e+00 1.03935122e+00
2.94829875e-01 -7.51601875e-01 4.05116528e-01 -1.77277494e-02
8.20243239e-01 -1.22392721e-01 4.58512336e-01 5.09568989e-01
1.64609507e-01 -7.20593095e-01 6.52902067e-01 4.88391131e-01
9.77455795e-01 -5.72302461e-01 5.35288692e-01 6.12217546e-01
-4.25514907e-01 1.90913901e-01 -3.46845359e-01 -3.58235347e-03
2.85316050e-01 6.03407741e-01 -7.95422256e-01 7.60182500e-01
3.79727870e-01 6.72644734e-01 -5.35267651e-01 4.15861964e-01
-3.22627455e-01 6.92468584e-01 -1.89685896e-02 -1.26133457e-01
-3.09053827e-02 -5.98787189e-01 4.32288885e-01 1.46901774e+00
6.90971971e-01 -7.11667091e-02 -9.99602154e-02 6.66043282e-01
-4.54240233e-01 5.85508704e-01 -1.88648656e-01 -2.91120350e-01
6.63473427e-01 1.41648316e+00 -1.05921291e-01 -3.56760919e-01
-1.87683895e-01 1.92257512e+00 5.52883685e-01 5.22373736e-01
-8.02166283e-01 -1.76714703e-01 9.41205800e-01 -2.72003412e-01
-1.88355491e-01 -2.87447035e-01 -5.56116164e-01 -1.67235601e+00
4.12442058e-01 -1.39715087e+00 8.22805613e-02 -4.67454314e-01
-1.40913522e+00 8.35657775e-01 -5.34051895e-01 -1.35624361e+00
-3.48939210e-01 -3.71981174e-01 -7.50453949e-01 1.20892906e+00
-1.16642296e+00 -1.11024964e+00 3.81832682e-02 2.53247499e-01
8.09487224e-01 -2.81202823e-01 8.23077857e-01 5.06642520e-01
-8.99842560e-01 1.43148291e+00 3.72716069e-01 -4.43567187e-02
1.42299843e+00 -1.02745569e+00 8.87384892e-01 8.75683308e-01
1.37883171e-01 1.18923926e+00 8.88946831e-01 -6.36481881e-01
-1.51860368e+00 -1.10606265e+00 1.67220569e+00 -1.08062530e+00
4.35241252e-01 -4.58635896e-01 -4.39640939e-01 6.92867398e-01
5.33579230e-01 -5.85856318e-01 9.34923172e-01 2.44636491e-01
-9.92439687e-02 -1.31292284e-01 -9.42691326e-01 1.08923924e+00
1.43928468e+00 -6.47195101e-01 -5.30958176e-01 4.27046925e-01
6.03114367e-01 -4.86702204e-01 -8.14384937e-01 2.76678830e-01
6.94169283e-01 -7.25993514e-01 6.36249542e-01 -7.75346041e-01
6.12776875e-01 -9.26655009e-02 -1.75601795e-01 -1.96874499e+00
-5.55582464e-01 -1.19835377e+00 -6.13872595e-02 1.33885038e+00
8.96492481e-01 -6.27684951e-01 4.60609704e-01 7.35043883e-01
-1.81042984e-01 -7.87115872e-01 -6.28903449e-01 -6.98560834e-01
1.75763935e-01 1.67086348e-02 7.61653125e-01 8.87567282e-01
1.94619492e-01 8.34353447e-01 -8.14121783e-01 -8.19594041e-02
5.18455267e-01 1.73617452e-01 9.92345035e-01 -9.11460638e-01
-6.22291684e-01 -5.60709417e-01 3.03200841e-01 -1.40717947e+00
9.78630334e-02 -1.00511491e+00 -4.42475304e-02 -1.42610490e+00
5.49989164e-01 -3.65872264e-01 -5.04843056e-01 -5.85617535e-02
-8.55934024e-01 2.55299091e-01 3.66642252e-02 5.46286225e-01
-6.81851268e-01 6.38672769e-01 1.29080963e+00 -7.81832561e-02
-2.68374324e-01 2.34717309e-01 -1.02520335e+00 1.67679146e-01
1.00054479e+00 -2.20420361e-01 -3.50180447e-01 -1.07144856e+00
3.94884467e-01 -1.21325783e-01 -3.34363371e-01 -6.35942161e-01
1.13524690e-01 -4.58785504e-01 2.01356217e-01 -2.28503257e-01
5.22720739e-02 -4.03936088e-01 2.05962043e-02 4.60830666e-02
-6.72178626e-01 5.42811036e-01 -4.37073648e-01 3.70681316e-01
1.29099727e-01 2.19638377e-01 4.87137109e-01 5.18352985e-02
-5.19562662e-02 4.07495946e-01 -1.54122442e-01 3.07443980e-02
6.66185737e-01 1.50364503e-01 -2.09913507e-01 -4.33252722e-01
-3.80656689e-01 3.66220236e-01 6.52934074e-01 8.81874025e-01
2.73925275e-01 -1.55396771e+00 -1.18414247e+00 3.55605558e-02
4.78846908e-01 -6.79042935e-01 1.02053083e-01 9.87631142e-01
-1.48073107e-01 3.34242135e-01 -1.38845593e-01 -3.79032940e-01
-1.11718881e+00 2.44953632e-01 1.03081405e-01 -3.63904715e-01
-2.03742072e-01 6.74899638e-01 -1.31565273e-01 -6.63126230e-01
-5.51685765e-02 -9.30203218e-03 -2.31644381e-02 -2.46344730e-01
6.41622305e-01 3.47512156e-01 2.99766183e-01 -5.87625802e-01
-1.38986200e-01 1.92887932e-01 -4.80003446e-01 -4.13717538e-01
9.01889086e-01 -7.03525186e-01 -1.91693544e-01 5.38988709e-01
9.46911216e-01 1.78173661e-01 -8.80055606e-01 -6.43809676e-01
-1.64288506e-02 -6.52322650e-01 -4.49885637e-01 -1.14983082e+00
-4.38657403e-01 8.38811755e-01 4.44568157e-01 -2.81328917e-01
8.10451686e-01 -3.65354568e-01 1.29903877e+00 5.31623065e-01
5.73300600e-01 -1.51527083e+00 -1.26377881e-01 8.54908407e-01
8.33283007e-01 -1.35019100e+00 -3.92536849e-01 -2.48612583e-01
-8.40067148e-01 7.90217757e-01 6.13413453e-01 4.62918580e-01
7.10543245e-02 -4.29573476e-01 5.84687114e-01 6.34928167e-01
-8.46194029e-01 3.03572536e-01 4.20836598e-01 3.02494854e-01
1.00256479e+00 5.49078107e-01 -1.05812275e+00 8.70148063e-01
-5.18171191e-01 8.07192363e-03 2.03315020e-01 5.07807016e-01
-1.88864455e-01 -1.67537928e+00 -1.04909474e-02 2.56109238e-01
-4.76214528e-01 -3.30260515e-01 -7.65672207e-01 -2.15643048e-01
-3.70308161e-01 1.14508259e+00 -2.84777433e-01 -7.17054307e-01
4.19717610e-01 2.61483639e-01 4.73006576e-01 -6.06337965e-01
-7.21251070e-01 -3.25885385e-01 2.25079551e-01 -3.60874414e-01
-2.24275030e-02 -5.86240530e-01 -4.51537997e-01 -5.68678319e-01
1.00356311e-01 4.70104873e-01 6.46852195e-01 8.74345243e-01
7.89093733e-01 8.28988031e-02 1.20461202e+00 -7.28490651e-01
-1.10707009e+00 -1.40210974e+00 8.27426463e-02 7.05728412e-01
-6.69277133e-03 -1.60873324e-01 -1.48576289e-01 1.79125182e-02] | [11.690176010131836, 10.173251152038574] |
3a09b35b-3587-452d-be6d-b622880f2cc9 | generating-dataset-for-large-scale-3d-facial | 2109.08043 | null | https://arxiv.org/abs/2109.08043v1 | https://arxiv.org/pdf/2109.08043v1.pdf | Generating Dataset For Large-scale 3D Facial Emotion Recognition | The tremendous development in deep learning has led facial expression recognition (FER) to receive much attention in the past few years. Although 3D FER has an inherent edge over its 2D counterpart, work on 2D images has dominated the field. The main reason for the slow development of 3D FER is the unavailability of large training and large test datasets. Recognition accuracies have already saturated on existing 3D emotion recognition datasets due to their small gallery sizes. Unlike 2D photographs, 3D facial scans are not easy to collect, causing a bottleneck in the development of deep 3D FER networks and datasets. In this work, we propose a method for generating a large dataset of 3D faces with labeled emotions. We also develop a deep convolutional neural network(CNN) for 3D FER trained on 624,000 3D facial scans. The test data comprises 208,000 3D facial scans. | ['Syed Zulqarnain Gilani', 'Faizan Farooq Khan'] | 2021-09-16 | null | null | null | null | ['facial-emotion-recognition'] | ['computer-vision'] | [-1.56413913e-01 1.15177426e-02 1.31143764e-01 -8.81245911e-01
-4.27290052e-01 -1.51310429e-01 3.30421656e-01 -6.13332570e-01
-1.41487703e-01 3.98104459e-01 -4.33519706e-02 -8.94805416e-02
4.46273804e-01 -6.73503816e-01 -4.32012886e-01 -4.24055934e-01
-1.06918104e-01 1.71293959e-01 -3.61682385e-01 -2.45653108e-01
-1.95391282e-01 1.06320834e+00 -1.82184529e+00 6.04351051e-02
2.13598236e-01 1.71021390e+00 -4.79970753e-01 3.69334012e-01
-3.97510558e-01 5.32244503e-01 -7.37510204e-01 -6.58786952e-01
3.57032984e-01 -6.23473227e-01 -5.40758193e-01 4.99610722e-01
6.82746410e-01 -4.30646330e-01 -3.42006594e-01 8.17241013e-01
8.89522135e-01 -8.17860290e-02 6.11111522e-01 -1.68040502e+00
-6.20536566e-01 -2.33546168e-01 -8.76439452e-01 -9.17122290e-02
4.17363465e-01 -2.03127071e-01 5.93733370e-01 -1.16227162e+00
5.74010253e-01 1.46106815e+00 7.66413987e-01 1.03916550e+00
-8.61771405e-01 -9.03718054e-01 -4.86637652e-01 -1.95381701e-01
-1.38677204e+00 -6.30117238e-01 9.44930851e-01 -2.63808906e-01
9.82699513e-01 -1.80614471e-01 9.52540874e-01 1.36396563e+00
-1.33315355e-01 9.71230388e-01 1.41046739e+00 -3.87510300e-01
1.76182464e-02 -1.29105733e-03 -2.61440158e-01 8.78547907e-01
-7.10781068e-02 -5.56509346e-02 -5.46269953e-01 -6.10874929e-02
9.30852711e-01 -7.44680390e-02 1.04531623e-01 -1.26940817e-01
-2.28593633e-01 7.75654197e-01 1.88322812e-01 7.10694715e-02
-2.22010344e-01 -7.26017132e-02 5.89363098e-01 5.83902836e-01
9.40912664e-01 2.27912799e-01 -3.16279233e-01 -4.09698397e-01
-8.65229309e-01 2.86768138e-01 6.17613673e-01 6.98391855e-01
7.49253392e-01 4.99276280e-01 3.62973660e-01 1.22471440e+00
7.37616271e-02 6.90625787e-01 2.20984474e-01 -7.69471049e-01
-2.74926480e-02 9.18169439e-01 -2.17587143e-01 -1.12892532e+00
-3.42389107e-01 6.22444414e-02 -9.63804305e-01 5.00398636e-01
2.19865054e-01 -3.01207185e-01 -9.91392910e-01 1.72663426e+00
3.51535052e-01 -2.21039504e-01 5.87851834e-03 9.57826018e-01
1.03875339e+00 5.58598220e-01 -4.51245010e-02 1.22896798e-01
8.26660812e-01 -5.85038900e-01 -4.84675527e-01 8.20028707e-02
5.84541678e-01 -7.30076849e-01 8.64656866e-01 3.85473549e-01
-1.04848099e+00 -5.33533573e-01 -6.76248193e-01 -2.61662658e-02
-4.55207944e-01 5.42850979e-02 9.44036245e-01 8.93818378e-01
-1.21367764e+00 1.57830343e-01 -3.34612548e-01 -4.67281759e-01
1.13337743e+00 4.76359874e-01 -1.01055205e+00 -2.72959918e-01
-1.08907831e+00 8.49699497e-01 -1.95366919e-01 3.89741451e-01
-8.54167938e-01 -4.54565614e-01 -9.39948678e-01 -2.01610476e-01
-7.21163973e-02 -1.53930649e-01 1.15240848e+00 -1.39878178e+00
-1.50389254e+00 1.66108823e+00 -2.63672862e-02 1.22511119e-01
4.00893658e-01 -5.08757308e-02 -6.55406058e-01 1.14713669e-01
-1.46939293e-01 9.61605489e-01 9.74660397e-01 -1.04001021e+00
-2.64250398e-01 -6.99925303e-01 2.41666306e-02 -1.90933794e-01
-3.51770192e-01 1.77386552e-01 -4.52963740e-01 -3.00780803e-01
-1.27438381e-01 -9.31276739e-01 7.50340447e-02 2.42613524e-01
-1.60178170e-01 -4.58807081e-01 1.09835708e+00 -3.42301399e-01
6.25675380e-01 -2.35431170e+00 -2.29431614e-01 1.29827885e-02
1.31530091e-01 4.77394640e-01 -3.32490951e-01 2.71050017e-02
-3.82545471e-01 2.51985013e-01 9.43488330e-02 -5.54209590e-01
-4.45441790e-02 2.47473091e-01 -9.73457545e-02 3.60628217e-01
8.88053715e-01 7.59634793e-01 -6.26428962e-01 -3.93622935e-01
-3.23917158e-02 7.61441171e-01 -6.00576162e-01 4.16014224e-01
3.93858775e-02 1.79010928e-01 -5.15416920e-01 1.14629710e+00
1.01201749e+00 -1.45911500e-01 -3.14338654e-01 -8.58828872e-02
1.08405277e-01 -2.39059612e-01 -6.59372449e-01 1.53019965e+00
-3.74960035e-01 7.56761968e-01 2.18086451e-01 -1.11085987e+00
1.54389155e+00 4.91061121e-01 8.37010086e-01 -8.61314833e-01
4.85769838e-01 3.51383030e-01 -3.24378580e-01 -6.12694025e-01
2.75221944e-01 -5.88382304e-01 -2.22919926e-01 5.46810389e-01
3.83378208e-01 -4.28040594e-01 -2.59474982e-02 -2.24670291e-01
8.84126604e-01 1.37883380e-01 -1.39729947e-01 1.76102802e-01
3.51959348e-01 -6.71326220e-02 5.20724177e-01 -4.27838042e-02
-4.50101286e-01 4.56674993e-01 8.64242733e-01 -9.48040783e-01
-1.00527418e+00 -7.27854550e-01 -2.96425253e-01 6.14275515e-01
-2.34414473e-01 -2.73262650e-01 -6.99238718e-01 -9.28290188e-01
1.56483695e-01 -6.62016217e-03 -9.79611933e-01 -2.12633297e-01
-2.51141667e-01 -6.02036297e-01 9.99251842e-01 4.30240721e-01
9.14416194e-01 -1.03714621e+00 -5.34056127e-01 -2.06597582e-01
8.24473351e-02 -1.05353713e+00 3.61806266e-02 8.43395963e-02
-7.04118729e-01 -1.16369069e+00 -8.36748719e-01 -6.94777906e-01
7.56879032e-01 1.83632523e-01 1.24116361e+00 -4.81067747e-02
-4.73826379e-01 3.48166674e-01 -3.90675843e-01 -8.78801286e-01
-2.54181266e-01 -1.77330613e-01 2.77388453e-01 2.32802574e-02
8.92668664e-01 -4.26212549e-01 -2.58428991e-01 4.25736457e-01
-8.67596805e-01 -8.33513290e-02 5.04476547e-01 9.26443338e-01
5.23865104e-01 -4.97145131e-02 6.62838519e-01 -5.48726857e-01
5.04833162e-01 -3.90173286e-01 -4.29728121e-01 -3.85264978e-02
-2.53195792e-01 -2.77718663e-01 3.33850890e-01 -3.02594870e-01
-1.14769745e+00 -1.07650615e-01 -6.13539636e-01 -7.88551390e-01
-5.01929343e-01 3.53300095e-01 -1.86724916e-01 -1.13942526e-01
5.15662014e-01 -3.54760885e-01 4.77174014e-01 -4.55685198e-01
-9.10922438e-02 9.52212870e-01 2.12417189e-02 -5.06849289e-01
4.06396151e-01 3.27555895e-01 2.68628031e-01 -1.12625957e+00
-1.00287259e+00 5.35681192e-03 -5.51126957e-01 -6.09622598e-01
7.95501113e-01 -1.02998865e+00 -6.55881643e-01 8.44008744e-01
-8.72766137e-01 -2.98801690e-01 -2.77596682e-01 1.00589693e-01
-4.45182055e-01 -1.25145316e-01 -5.61733186e-01 -9.48815048e-01
-2.34237254e-01 -1.20730686e+00 1.42263710e+00 2.63619661e-01
-2.59988397e-01 -7.15856194e-01 -4.66073714e-02 4.89964187e-01
4.11410362e-01 7.44373679e-01 6.68121278e-01 -1.90919489e-01
-6.66127726e-02 -6.23930633e-01 -3.88987243e-01 8.51055145e-01
2.63615698e-01 1.87649533e-01 -1.30407739e+00 7.77720660e-02
-2.04676911e-01 -1.26976418e+00 5.21697998e-01 1.36655137e-01
1.37198424e+00 4.03861135e-01 7.99695030e-02 5.99805772e-01
1.08108068e+00 1.99778661e-01 6.97550058e-01 -7.41853938e-02
2.91882515e-01 7.10421503e-01 4.92634147e-01 5.79999149e-01
6.06894977e-02 4.82622981e-01 4.97003973e-01 -5.04400909e-01
-5.20434603e-02 -2.68708587e-01 2.12480560e-01 6.23489439e-01
-3.96295518e-01 -8.87590721e-02 -9.84166682e-01 2.40507185e-01
-1.14593315e+00 -8.36120784e-01 2.34746471e-01 1.78475416e+00
6.07591212e-01 -2.16859326e-01 1.67509228e-01 1.74269766e-01
4.82189685e-01 1.34778470e-01 -6.16835475e-01 -7.77454019e-01
-4.15266871e-01 5.55593014e-01 -2.01354668e-01 -1.30199179e-01
-1.05860615e+00 9.19672668e-01 6.62544203e+00 6.29611135e-01
-1.72939610e+00 -2.84084409e-01 1.20769870e+00 -2.57905573e-01
5.19903339e-02 -7.28297710e-01 -8.46651137e-01 1.09184824e-01
8.10227573e-01 1.98105767e-01 -7.70550817e-02 1.23858738e+00
1.37277856e-01 6.43056352e-04 -8.76333475e-01 1.47533047e+00
2.89071411e-01 -1.08977139e+00 2.17637926e-01 2.56838828e-01
7.68391550e-01 1.29702222e-02 7.80572966e-02 4.11851346e-01
2.93558985e-02 -1.51351297e+00 4.63679433e-01 2.78479874e-01
1.51503086e+00 -9.46973205e-01 8.44294190e-01 -8.19230527e-02
-6.93562031e-01 1.21022619e-01 -4.72484767e-01 -1.56932354e-01
7.23356456e-02 5.55654287e-01 -5.90166390e-01 1.06442638e-01
1.05905449e+00 8.71825159e-01 -2.37033352e-01 4.99425560e-01
-1.10510781e-01 4.86858279e-01 -3.58679891e-01 -4.76456918e-02
3.65628272e-01 -2.40662202e-01 4.26478162e-02 9.37306106e-01
5.82751751e-01 1.27311125e-01 -1.29991308e-01 5.58069170e-01
-6.26399398e-01 -9.68203600e-03 -9.40249264e-01 -5.07682860e-01
-1.22313043e-02 1.41737878e+00 -3.52680624e-01 -5.37843294e-02
-7.35676169e-01 9.25446153e-01 3.08920681e-01 2.53856093e-01
-7.13145792e-01 -2.97015280e-01 1.14772213e+00 -1.52370473e-02
4.24259268e-02 -1.93862706e-01 1.87025353e-01 -9.74719226e-01
-2.93654144e-01 -9.74361002e-01 1.69197246e-01 -9.78820145e-01
-1.55152845e+00 8.20450604e-01 -3.38261485e-01 -8.83703649e-01
-1.65337294e-01 -1.04296613e+00 -2.89406210e-01 8.37903082e-01
-1.49843633e+00 -1.12856638e+00 -6.16091788e-01 8.59813273e-01
3.00016731e-01 -2.85541952e-01 1.18709493e+00 4.96100396e-01
-6.12070143e-01 6.82728827e-01 -3.86788577e-01 4.47103411e-01
8.77775252e-01 -7.93498516e-01 1.57341450e-01 1.43131241e-01
-2.90700998e-02 2.19865113e-01 9.25149918e-02 -1.26393795e-01
-1.60952461e+00 -9.09815669e-01 8.10832381e-01 -3.04665178e-01
2.84895658e-01 -5.78197420e-01 -5.50716877e-01 4.80564743e-01
-1.78136706e-01 3.00546229e-01 1.09618640e+00 1.00033887e-01
-5.41635990e-01 -2.64010012e-01 -1.38878942e+00 4.12993520e-01
1.14383733e+00 -6.03477895e-01 -1.32974610e-01 -7.54775926e-02
9.12165642e-03 -2.99756378e-01 -1.03355908e+00 5.69418013e-01
8.29678714e-01 -1.09431064e+00 6.44972026e-01 -6.59531116e-01
7.35532582e-01 1.69470608e-01 -2.24978238e-01 -1.27566457e+00
1.03895612e-01 -4.72117990e-01 1.02768205e-01 1.00288320e+00
1.67855456e-01 -3.61112595e-01 1.15481150e+00 7.94740200e-01
-5.02757058e-02 -1.09852207e+00 -9.33134675e-01 -5.33586502e-01
1.08887069e-01 -4.44696009e-01 8.06571126e-01 1.03686476e+00
-2.96843439e-01 4.06416684e-01 -3.76371205e-01 -4.42904770e-01
2.70061880e-01 2.49982044e-01 1.06458712e+00 -1.39411891e+00
3.43053639e-01 -4.16616738e-01 -8.95621240e-01 -9.18238044e-01
5.77196062e-01 -8.03299427e-01 -2.88700104e-01 -1.02205110e+00
1.77895814e-01 -5.61502934e-01 -4.09898460e-02 8.71726990e-01
3.11104804e-01 7.90049613e-01 -9.95891392e-02 -2.96599150e-01
-2.36382917e-01 9.12269592e-01 1.54122126e+00 -4.90134284e-02
1.81896016e-01 -3.20918858e-01 -5.69756985e-01 5.83648920e-01
8.34004879e-01 2.91590616e-02 -4.15371895e-01 -4.98524964e-01
1.81528449e-01 -3.84930745e-02 2.25933492e-01 -7.99538493e-01
-4.53227073e-01 9.58576128e-02 9.47024226e-01 -4.14067060e-01
6.83455527e-01 -8.15466523e-01 3.44598591e-02 -1.94558516e-01
-4.56169881e-02 5.18820062e-02 5.31204283e-01 -2.85595329e-03
-7.16969669e-01 2.16368362e-01 8.95710886e-01 -1.77134186e-01
-8.87018561e-01 8.61769736e-01 -3.32923174e-01 6.10994233e-04
8.99295986e-01 -5.56861281e-01 8.52441788e-02 -4.38768923e-01
-5.54138005e-01 -8.03855807e-02 5.84156930e-01 6.84367478e-01
8.42475712e-01 -1.53282332e+00 -4.65548515e-01 6.30259216e-01
1.78405017e-01 2.41024047e-02 3.19698840e-01 4.90805089e-01
-3.84401768e-01 2.19477877e-01 -8.29214334e-01 -4.86312091e-01
-1.41340458e+00 6.81053707e-03 5.42889416e-01 1.56800434e-01
-2.80076444e-01 1.07515800e+00 3.74365486e-02 -5.81893623e-01
1.02254152e-01 1.55270830e-01 -2.19020452e-02 1.43061504e-01
6.09353244e-01 6.54721335e-02 -7.10796099e-04 -9.79817986e-01
-2.79478997e-01 5.89856386e-01 1.88364357e-01 1.91719577e-01
1.82308948e+00 2.62740344e-01 -1.74031690e-01 2.49826387e-01
1.63288915e+00 -3.03422362e-01 -1.20363951e+00 1.39498353e-01
-3.58274668e-01 -6.45647883e-01 2.40118504e-01 -5.26734591e-01
-1.65251160e+00 1.16727376e+00 5.42758763e-01 2.75227483e-02
1.41994786e+00 9.55694690e-02 8.82298410e-01 2.05185294e-01
5.51854134e-01 -9.60847259e-01 2.58336723e-01 6.84145331e-01
8.83139431e-01 -1.40540278e+00 -3.07850391e-01 -2.82525569e-01
-5.08476853e-01 1.40172875e+00 8.30924630e-01 -1.03757326e-02
8.62527907e-01 3.74864161e-01 3.52763385e-01 -6.04021847e-01
-4.99042928e-01 -2.42713951e-02 -6.30872175e-02 5.90848923e-01
7.50159383e-01 -2.16263503e-01 2.97433436e-01 4.70761418e-01
-1.35563686e-01 4.59708840e-01 1.79960340e-01 9.17519927e-01
-1.63186044e-02 -1.19895267e+00 -5.90730757e-02 4.64933157e-01
-6.00464940e-01 4.25968349e-01 -7.02392757e-01 9.95563507e-01
2.92640477e-01 7.70161152e-01 2.84906507e-01 -4.92699921e-01
4.91662651e-01 4.53952461e-01 7.92419672e-01 -2.74048150e-01
-1.42098948e-01 -1.83959052e-01 2.91325450e-01 -6.82454526e-01
-6.89009488e-01 -5.20240307e-01 -9.68392253e-01 -5.19916654e-01
-5.29691540e-02 7.42473602e-02 7.75679469e-01 5.54127455e-01
6.56778753e-01 -1.39587089e-01 8.66973817e-01 -9.50267434e-01
-1.56590208e-01 -8.55088592e-01 -1.00432968e+00 6.20165288e-01
1.54986575e-01 -1.00759470e+00 -3.01788718e-01 9.19867605e-02] | [13.544835090637207, 1.5708718299865723] |
b9ecadd8-2c7d-47f7-916c-9e72c706c203 | identifying-adversarial-sentences-by | 1912.08981 | null | https://arxiv.org/abs/1912.08981v1 | https://arxiv.org/pdf/1912.08981v1.pdf | Identifying Adversarial Sentences by Analyzing Text Complexity | Attackers create adversarial text to deceive both human perception and the current AI systems to perform malicious purposes such as spam product reviews and fake political posts. We investigate the difference between the adversarial and the original text to prevent the risk. We prove that the text written by a human is more coherent and fluent. Moreover, the human can express the idea through the flexible text with modern words while a machine focuses on optimizing the generated text by the simple and common words. We also suggest a method to identify the adversarial text by extracting the features related to our findings. The proposed method achieves high performance with 82.0% of accuracy and 18.4% of equal error rate, which is better than the existing methods whose the best accuracy is 77.0% corresponding to the error rate 22.8%. | ['Hoang-Quoc Nguyen-Son', 'Tran Phuong Thao', 'Shinsaku Kiyomoto', 'Seira Hidano'] | 2019-12-19 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [ 8.6202279e-02 4.0093738e-01 5.0214939e-02 -1.0088666e-01
-2.9380581e-01 -1.0484366e+00 1.0576190e+00 -2.0062317e-01
-2.6924777e-01 6.6025990e-01 1.1266548e-01 -3.0535805e-01
6.5060925e-01 -9.5087874e-01 -5.2247226e-01 -4.5609483e-01
5.0884646e-01 8.7011628e-02 -1.4039434e-01 -6.4584398e-01
6.8483621e-01 7.1108460e-02 -7.8139895e-01 2.7747500e-01
1.1690345e+00 7.3842710e-01 -2.8970578e-01 6.2284720e-01
-2.5549161e-01 8.7985963e-01 -1.2123570e+00 -1.3267215e+00
6.5181792e-01 -3.6245576e-01 -6.0722226e-01 -1.6583608e-01
1.3121544e-01 -5.0896442e-01 -4.0716416e-01 1.8535126e+00
1.9863003e-01 -2.0403758e-01 6.5589142e-01 -1.2168467e+00
-1.3655878e+00 7.4855232e-01 -6.9863629e-01 -2.8050482e-01
4.5312327e-01 2.3101965e-01 5.7751942e-01 -6.1522788e-01
4.5076632e-01 1.4938725e+00 3.7749320e-01 8.4091067e-01
-6.5501839e-01 -1.0778588e+00 1.9309342e-01 -1.3245443e-01
-1.1872290e+00 -2.0796290e-01 7.9852092e-01 -5.4068959e-01
2.4949025e-01 4.3089071e-01 3.3140129e-01 1.4633803e+00
8.5564297e-01 4.8960415e-01 1.4726686e+00 -1.9976985e-01
-2.9965440e-02 8.2928067e-01 1.5836097e-01 6.7705268e-01
5.2251977e-01 1.7345843e-01 -2.2189696e-01 -5.6619626e-01
1.8254150e-01 6.5110505e-02 -2.3929922e-01 5.8289659e-01
-8.4566677e-01 1.3150359e+00 3.0432212e-01 2.2099689e-01
-3.2315424e-01 -1.4940044e-01 3.6034125e-01 3.2585695e-01
5.1470482e-01 7.8321826e-01 -3.2418147e-01 8.1615388e-02
-5.5165088e-01 2.6042980e-01 1.1441320e+00 9.3706727e-01
2.8180796e-01 3.7812027e-01 1.6300267e-01 4.4387919e-01
5.9390056e-01 1.1541579e+00 8.5431230e-01 -5.2597111e-01
4.4577226e-01 6.6498113e-01 3.6195734e-01 -1.7669280e+00
3.0302355e-01 -3.3634445e-01 -8.5922939e-01 2.7852380e-01
2.1871884e-01 -4.2972651e-01 -7.7706683e-01 1.4193923e+00
2.7156204e-02 -3.9355239e-01 2.9765201e-01 8.0539560e-01
6.0241884e-01 1.0889076e+00 1.5049110e-01 -1.5793447e-01
1.4038033e+00 -9.6205443e-01 -1.1507730e+00 -4.7394148e-01
4.1445732e-01 -1.1762633e+00 1.0640770e+00 6.4573210e-01
-7.1345401e-01 -2.7614215e-01 -1.3008718e+00 2.1766631e-01
-7.5501645e-01 4.1632030e-02 3.4498271e-01 1.1717588e+00
-4.3475318e-01 5.1254755e-01 -6.6574670e-02 6.1429322e-02
4.5491853e-01 1.0419920e-01 -2.7566895e-01 2.8837833e-01
-1.8637964e+00 1.0998312e+00 2.6560882e-01 1.2035160e-01
-6.6138822e-01 -4.3229669e-01 -7.0084178e-01 -3.9466199e-01
2.6507056e-01 -4.1922298e-01 1.0806531e+00 -1.5633589e+00
-1.6929102e+00 9.8957229e-01 8.3947070e-02 -6.0858440e-01
1.0105972e+00 -4.7793624e-01 -7.1937829e-01 -3.2557745e-02
3.3196028e-02 5.8809292e-02 1.3412036e+00 -1.2366953e+00
-2.1285470e-01 -3.3724385e-01 1.3122854e-01 1.7069920e-03
-4.7571883e-01 7.0672587e-02 9.8347068e-02 -1.0279409e+00
-2.5520197e-01 -9.8813111e-01 -1.3194126e-01 -1.9622302e-01
-8.2535577e-01 -5.7911240e-02 9.7108132e-01 -8.7716663e-01
1.2800409e+00 -1.7560133e+00 -2.4563098e-01 2.7303511e-01
6.3624734e-01 7.6239014e-01 3.5119212e-01 4.5506707e-01
-3.7224863e-02 9.0055656e-01 -5.1692266e-02 1.5134220e-01
1.2659810e-01 -1.9349648e-01 -7.8716993e-01 4.9788934e-01
6.5995224e-02 9.7816241e-01 -9.1144556e-01 -3.1825325e-01
-1.1660905e-02 8.1441991e-02 -1.8396546e-01 2.4418773e-01
2.3106501e-02 4.1939504e-02 -9.4615281e-01 4.9799508e-01
9.7290146e-01 -1.0864901e-01 1.9750054e-01 7.6408878e-02
2.5284460e-01 1.2379035e-01 -7.3501420e-01 6.4529496e-01
-3.0592331e-01 6.0016221e-01 6.7341670e-02 -3.5968834e-01
1.1166736e+00 -1.1492950e-02 -2.3752058e-01 -3.3911744e-01
5.3522962e-01 1.5895642e-01 -1.8733816e-02 -6.3011634e-01
6.6844106e-01 -3.4391654e-01 -3.9254579e-01 5.5112410e-01
-6.1114860e-01 -2.1061899e-01 -4.3081564e-01 5.7510036e-01
9.3734723e-01 -2.5808066e-01 5.4102111e-01 -3.8469878e-01
8.3317792e-01 -7.0241846e-02 1.8097964e-01 9.0309232e-01
-4.9681005e-01 1.1378131e-02 6.1901689e-01 -2.7820158e-01
-9.4249594e-01 -8.4497529e-01 1.6470371e-01 6.8412632e-01
4.3799764e-01 -4.4340491e-01 -8.2972980e-01 -1.1476679e+00
2.1083141e-03 1.0749525e+00 -7.4705303e-01 -4.6525702e-01
-4.4762376e-01 -5.7012236e-01 8.1105363e-01 -3.2041602e-02
8.5631394e-01 -8.9212054e-01 7.3852800e-02 -1.2426196e-01
-2.0949772e-01 -9.5802295e-01 -5.8936507e-01 -5.7296526e-01
-2.9200584e-01 -9.4713175e-01 -3.8256046e-01 -4.4391420e-01
8.6157000e-01 8.8699304e-02 7.3718208e-01 2.2237954e-01
2.3629345e-01 -2.4057445e-01 -4.6725345e-01 -7.4964768e-01
-1.2250990e+00 -6.0117766e-02 1.2090824e-01 1.3432364e-01
4.2659152e-01 -3.7683059e-02 -3.6493838e-01 2.6199460e-01
-1.1934965e+00 8.2411796e-02 5.6475496e-01 7.7684194e-01
-1.6397679e-01 1.5883422e-01 6.5531343e-01 -1.3723522e+00
1.0790021e+00 -6.4280772e-01 -4.6964073e-01 4.5428932e-02
-1.0308076e+00 -3.4530892e-03 1.1261915e+00 -7.6023197e-01
-1.0940170e+00 -2.5640687e-01 -8.4855191e-02 -1.7696209e-02
9.2732400e-02 1.8595135e-01 -2.1882465e-01 -2.4725972e-01
8.4187287e-01 2.4258007e-01 1.5050319e-01 -1.5655957e-02
5.3060883e-01 1.0261916e+00 2.0450476e-01 -2.2867356e-01
1.4215991e+00 3.5200566e-01 -3.2805195e-01 -3.9980128e-01
-8.1806350e-01 1.4452799e-01 -7.0742570e-02 -2.9900181e-01
7.3137927e-01 -4.6127430e-01 -8.8526833e-01 7.2329134e-01
-1.4757257e+00 3.1260499e-01 3.6096558e-01 3.6749531e-02
-1.8563937e-02 8.2438064e-01 -7.4446529e-01 -9.0738726e-01
-7.8183889e-01 -9.9557996e-01 7.1325165e-01 1.6550165e-01
-4.7493529e-01 -9.6909052e-01 -3.0612740e-01 5.7081729e-01
4.9541560e-01 5.1662987e-01 7.4211723e-01 -1.1856989e+00
-2.8017518e-01 -7.3745298e-01 -4.3280900e-02 4.7839853e-01
2.4047500e-01 2.6093578e-01 -8.5958880e-01 1.1370089e-01
5.1236057e-01 -1.5575108e-01 5.1600718e-01 -3.2386369e-01
7.8978682e-01 -1.1002465e+00 -1.3710661e-01 9.2196435e-02
1.1544253e+00 3.8216287e-01 8.3081418e-01 3.1919944e-01
6.5238971e-01 7.5011456e-01 6.5336764e-01 2.8739527e-01
1.0282923e-02 2.1566069e-01 5.1786894e-01 2.5712705e-01
4.7505316e-01 -7.2916716e-01 6.2888879e-01 8.2220954e-01
6.0353482e-01 -6.3591367e-01 -7.3519504e-01 1.5618283e-02
-1.7044320e+00 -1.0297482e+00 -4.7221497e-01 1.8763095e+00
9.0225422e-01 7.0669132e-01 -4.6753687e-01 -1.0492023e-01
1.2002330e+00 3.6530855e-01 -4.3555298e-01 -9.1052073e-01
-1.1726482e-01 9.4683440e-03 9.3025154e-01 8.3194685e-01
-1.0907122e+00 1.4310786e+00 6.6020904e+00 1.0736501e+00
-1.0203454e+00 -8.5176766e-02 7.5350964e-01 2.0017491e-01
-4.5366293e-01 -7.5960509e-02 -7.4576306e-01 8.7701964e-01
9.8884720e-01 -7.3425311e-01 2.5832048e-01 9.8091078e-01
2.9742515e-01 2.1989600e-01 -5.4714507e-01 7.4813503e-01
4.1103396e-01 -1.1582501e+00 5.0975192e-01 7.5879194e-02
8.0072582e-01 -6.8469936e-01 4.6807459e-01 2.0760632e-01
5.1941341e-01 -1.2967244e+00 7.7939540e-01 6.1161840e-01
3.7307489e-01 -6.9807589e-01 1.0002328e+00 6.4776427e-01
-4.9893284e-01 2.2728008e-01 -1.9785580e-01 -2.1861571e-01
2.9216681e-02 5.7599062e-01 -7.4968785e-01 3.8183868e-01
1.4706895e-01 5.0930989e-01 -5.5589336e-01 4.3339785e-02
-4.9555376e-01 5.9704125e-01 -9.9119380e-02 -7.4006468e-01
4.1335440e-01 -3.6006707e-01 7.8978360e-01 9.7332180e-01
-7.8126043e-02 9.4771283e-03 4.8850223e-02 1.0692990e+00
-4.5866504e-01 3.8214821e-01 -1.0108007e+00 -5.1867658e-01
4.7317705e-01 1.2142817e+00 -2.1741250e-01 -4.2801365e-01
6.0107531e-03 1.1600367e+00 -4.4681128e-02 1.5483849e-01
-1.0481170e+00 -8.1590939e-01 3.6628798e-01 5.1705744e-02
-2.0884340e-01 8.7608099e-02 -5.8602971e-01 -1.0594574e+00
2.1538112e-01 -1.3696505e+00 -1.1607886e-01 -8.6025852e-01
-1.7375927e+00 7.5576818e-01 -4.0789831e-01 -1.1418338e+00
-2.8579608e-01 -7.3817509e-01 -6.7191404e-01 1.0434121e+00
-1.0860003e+00 -1.0851427e+00 5.5878121e-02 3.4944952e-01
4.0002912e-01 -4.4097242e-01 7.2165173e-01 -2.0118311e-01
-4.9595323e-01 8.1606871e-01 -1.1726264e-02 4.0116051e-01
8.6254090e-01 -9.9560946e-01 6.7521584e-01 9.4865096e-01
-4.3406898e-01 1.0729607e+00 1.0401211e+00 -1.0959522e+00
-1.3886713e+00 -9.4613749e-01 1.1087848e+00 -7.7282697e-01
1.1554790e+00 -4.1389287e-01 -7.5316322e-01 4.6257308e-01
3.3383536e-01 -5.8425009e-01 6.1514556e-01 -5.5104595e-01
-6.2914097e-01 1.6669804e-01 -1.5938940e+00 1.0622404e+00
5.2140963e-01 -5.1441872e-01 -8.6236250e-01 5.7590848e-01
9.7644371e-01 -3.7503415e-01 -4.3328345e-01 -6.4478569e-02
6.2085068e-01 -7.4070907e-01 5.4848611e-01 -7.6080167e-01
9.6248889e-01 -2.5740555e-01 2.6038850e-02 -1.1936295e+00
-1.9804908e-01 -1.0688184e+00 -4.4286437e-02 1.1469030e+00
5.1529670e-01 -1.1821548e+00 6.3272732e-01 8.2323635e-01
3.8981593e-01 -5.1614082e-01 -4.1581261e-01 -4.6675220e-01
3.3041802e-01 -1.2893674e-01 4.2275342e-01 1.2061793e+00
-2.1201912e-01 3.4090644e-01 -7.7110064e-01 1.8816945e-01
4.4964713e-01 -5.2066497e-03 8.0758500e-01 -9.6155304e-01
2.2507517e-02 -4.8464838e-01 -3.0408579e-01 -1.0549982e+00
5.4026157e-01 -7.4245322e-01 -2.3067310e-01 -8.2276744e-01
-1.6225573e-02 6.9002211e-02 1.2544997e-01 9.3453050e-02
-5.0083554e-01 1.3334931e-01 2.1504983e-01 2.8374249e-01
-1.8565309e-01 4.4375756e-01 1.4305807e+00 -2.8203118e-01
6.1022330e-02 1.1690574e-01 -1.4458354e+00 1.1116289e+00
1.1263593e+00 -7.5382310e-01 -2.4698788e-01 1.0785329e-01
3.8885850e-01 -3.6195329e-01 1.9878113e-01 -1.7315556e-01
4.7790397e-02 -4.8059008e-01 3.4714827e-01 -2.4338856e-01
-3.3594765e-02 -7.2903430e-01 -1.9110554e-01 8.3979160e-01
-5.6736547e-01 2.3989847e-01 -1.2810509e-01 8.6931342e-01
4.4023857e-02 -5.5517548e-01 9.4297874e-01 -3.9223805e-01
-2.4292390e-01 -2.0853562e-02 -5.6584489e-01 1.8366356e-01
1.1588567e+00 5.6970309e-02 -8.6863655e-01 -8.7176639e-01
-3.1615961e-01 7.6016784e-02 4.1760236e-01 6.8055880e-01
6.2047917e-01 -1.3018988e+00 -8.0349058e-01 4.4906449e-02
-7.9054892e-02 -8.8634467e-01 -1.2933587e-01 8.9688338e-02
-8.2208937e-01 2.9752550e-01 -4.6638981e-02 1.3332620e-01
-1.2688078e+00 5.7201731e-01 4.1889167e-01 -3.6731693e-01
-1.7636459e-01 4.7603419e-01 2.1865867e-01 -2.0632076e-01
-2.6500219e-01 3.6210951e-01 -2.3698516e-01 -2.0443611e-01
6.4441419e-01 3.7205380e-01 -3.0355963e-01 -8.8804454e-01
-2.1417168e-01 2.8797373e-01 -5.4975760e-01 -8.3561487e-02
5.2343267e-01 -3.2441944e-02 -4.4667619e-01 1.7119500e-01
1.3196951e+00 7.6728928e-01 -3.9246354e-01 -7.1826406e-02
-2.8261846e-01 -7.6820344e-01 -2.0876342e-01 -1.0910302e+00
-7.0303261e-01 5.2445072e-01 2.3011211e-01 8.3955628e-01
5.9301925e-01 -4.7252768e-01 1.0611143e+00 4.3022278e-01
6.6574097e-02 -9.2846990e-01 1.7951798e-01 4.6235073e-01
1.1210934e+00 -1.3071930e+00 2.7273241e-01 -6.1511493e-01
-1.0669473e+00 1.0594367e+00 7.2559834e-01 -4.8202410e-01
7.1039021e-01 -3.3129333e-04 3.9334810e-01 -1.2487503e-01
-6.2454897e-01 4.4093296e-01 4.4494739e-01 5.7192826e-01
1.7897330e-01 2.1075812e-01 -8.1774902e-01 9.4340032e-01
-7.6751316e-01 -6.3723481e-01 9.3252748e-01 6.1804497e-01
-3.8907024e-01 -9.7688282e-01 -4.1789433e-01 4.3123651e-01
-1.1257761e+00 -2.7112839e-01 -1.1176236e+00 6.5337515e-01
-8.3101287e-02 1.3831550e+00 -5.0216687e-01 -8.0073500e-01
7.7940121e-02 2.0336773e-01 -2.0516495e-01 -4.7627011e-01
-1.0597955e+00 -2.3714611e-01 2.2720821e-01 -3.0360249e-01
-7.4564524e-02 -6.0101724e-03 -1.1369982e+00 -1.0098699e+00
-4.9000764e-01 1.5102731e-01 7.0083195e-01 9.5095384e-01
2.0725718e-01 9.4602942e-02 1.1885684e+00 -6.0341260e-03
-1.0503834e+00 -9.0005422e-01 -4.5975494e-01 7.7984089e-01
1.1406469e-01 -2.2944678e-01 -9.1025877e-01 4.5494683e-02] | [6.071565628051758, 8.12752914428711] |
48379ef5-a71b-49a9-ab40-8b0cf8979dc0 | virtual-multi-modality-self-supervised-1 | 2110.03278 | null | https://arxiv.org/abs/2110.03278v2 | https://arxiv.org/pdf/2110.03278v2.pdf | Virtual Multi-Modality Self-Supervised Foreground Matting for Human-Object Interaction | Most existing human matting algorithms tried to separate pure human-only foreground from the background. In this paper, we propose a Virtual Multi-modality Foreground Matting (VMFM) method to learn human-object interactive foreground (human and objects interacted with him or her) from a raw RGB image. The VMFM method requires no additional inputs, e.g. trimap or known background. We reformulate foreground matting as a self-supervised multi-modality problem: factor each input image into estimated depth map, segmentation mask, and interaction heatmap using three auto-encoders. In order to fully utilize the characteristics of each modality, we first train a dual encoder-to-decoder network to estimate the same alpha matte. Then we introduce a self-supervised method: Complementary Learning(CL) to predict deviation probability map and exchange reliable gradients across modalities without label. We conducted extensive experiments to analyze the effectiveness of each modality and the significance of different components in complementary learning. We demonstrate that our model outperforms the state-of-the-art methods. | ['Yandong Guo', 'Ziwen Li', 'Cheng Lu', 'Han Huang', 'Bo Xu'] | 2021-10-07 | virtual-multi-modality-self-supervised | http://openaccess.thecvf.com//content/ICCV2021/html/Xu_Virtual_Multi-Modality_Self-Supervised_Foreground_Matting_for_Human-Object_Interaction_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Xu_Virtual_Multi-Modality_Self-Supervised_Foreground_Matting_for_Human-Object_Interaction_ICCV_2021_paper.pdf | iccv-2021-1 | ['image-matting'] | ['computer-vision'] | [ 7.72235751e-01 2.11894318e-01 -2.26699817e-03 -4.30105120e-01
-9.32583570e-01 -3.73519033e-01 5.45891881e-01 -4.46718842e-01
-2.91592270e-01 5.87318480e-01 -1.14431627e-01 -2.20262393e-01
4.93685842e-01 -5.35884142e-01 -1.27838957e+00 -8.73709798e-01
4.81782854e-01 5.84596515e-01 6.55106723e-01 3.69625598e-01
2.45627798e-02 2.74197608e-02 -1.33794653e+00 8.03593993e-01
9.29296613e-01 8.20120156e-01 4.73543525e-01 8.65365386e-01
-4.02413011e-01 1.22829688e+00 -3.80190462e-01 -4.94213045e-01
3.05695713e-01 -8.55456114e-01 -8.01796913e-01 5.10152817e-01
5.81512749e-01 -5.12831092e-01 -4.08585578e-01 9.48792756e-01
2.15153262e-01 -7.04355389e-02 6.71054780e-01 -1.47084439e+00
-4.67354357e-01 6.61011755e-01 -1.04282165e+00 6.96972832e-02
1.63645655e-01 3.53897810e-01 5.10150194e-01 -8.53383839e-01
7.27944970e-01 1.26506174e+00 2.74604499e-01 6.56640053e-01
-1.17117786e+00 -3.59369934e-01 5.36593199e-01 4.53323275e-02
-1.16082275e+00 -2.10552156e-01 8.12906504e-01 -4.98632461e-01
3.97159368e-01 1.29539087e-01 7.29872048e-01 1.11643422e+00
1.99062407e-01 1.35156620e+00 1.30724597e+00 -3.87255281e-01
2.18066603e-01 2.34344646e-01 -1.46252945e-01 1.01371205e+00
7.71561638e-02 -3.16009820e-01 -8.06542397e-01 2.18782984e-02
9.94066834e-01 4.48503010e-02 -2.45180354e-01 -6.20546997e-01
-1.60815287e+00 2.07936451e-01 3.87735754e-01 2.54062098e-02
-2.68715918e-01 4.22633111e-01 2.00311039e-02 -1.49289548e-01
4.38590556e-01 -1.56773150e-01 -3.11246663e-01 -2.48865280e-02
-1.20074522e+00 -7.93112963e-02 5.67975104e-01 1.00758910e+00
9.85557675e-01 1.60847083e-02 -1.44264027e-01 3.73140424e-01
4.82502639e-01 8.80057395e-01 1.73777029e-01 -1.15048099e+00
4.87428963e-01 5.83366513e-01 7.50331730e-02 -4.83709574e-01
-1.37398943e-01 -2.73065343e-02 -7.32675552e-01 2.66365290e-01
5.37746012e-01 -1.28643259e-01 -1.49779391e+00 1.56553733e+00
5.04037321e-01 4.43945676e-01 -3.34447473e-02 1.02171481e+00
6.32567346e-01 7.32373893e-01 -1.70619152e-02 -2.25963574e-02
9.21224475e-01 -1.42770028e+00 -7.22122788e-01 -5.02066612e-01
1.77686498e-01 -5.65226555e-01 1.08549666e+00 4.46629524e-01
-1.39695847e+00 -5.66896141e-01 -9.10315633e-01 -2.34597996e-01
-2.08111435e-01 2.03194305e-01 6.25226974e-01 5.89094281e-01
-8.35511446e-01 5.38668692e-01 -1.25643682e+00 4.78683412e-02
6.52071238e-01 1.99635848e-01 -1.94949239e-01 -1.23087846e-01
-7.46902585e-01 7.81401813e-01 3.61238480e-01 9.97035727e-02
-1.56421268e+00 -6.63671851e-01 -1.00278282e+00 -2.83889592e-01
4.01599735e-01 -9.53191698e-01 1.10897887e+00 -1.60810721e+00
-1.44731700e+00 1.09243798e+00 -6.05837822e-01 -1.44511953e-01
7.21076012e-01 -3.92590433e-01 8.34008586e-03 3.27861756e-01
1.81984380e-01 8.75621855e-01 1.12053645e+00 -1.90727174e+00
-6.95725322e-01 -2.22437918e-01 3.43849836e-03 2.97936797e-01
1.07026406e-01 -2.71432340e-01 -8.41527581e-01 -4.78568226e-01
2.96136975e-01 -8.51281464e-01 -2.62700263e-02 1.85148597e-01
-6.34580672e-01 3.46544147e-01 7.29114175e-01 -8.37815702e-01
8.66616189e-01 -2.16011548e+00 8.03119600e-01 1.36324093e-01
4.45349187e-01 -1.11045621e-01 -7.26976618e-02 -1.12263896e-01
2.65849262e-01 -1.95135549e-01 -5.27344048e-01 -7.84902871e-01
-1.61036789e-01 4.28774446e-01 -7.82964826e-02 6.17988348e-01
2.36583933e-01 1.08550823e+00 -9.55812693e-01 -9.26129162e-01
4.19540584e-01 6.59691513e-01 -2.20261067e-01 4.70629066e-01
-6.06566370e-01 6.77942038e-01 -3.53892408e-02 9.72203732e-01
8.03073168e-01 -5.86347401e-01 1.85287282e-01 -4.17682022e-01
1.42356500e-01 -1.73327848e-01 -1.13821888e+00 1.93359315e+00
-2.41571292e-01 6.14141226e-01 2.26572603e-01 -5.19492507e-01
2.86529690e-01 4.35177833e-02 4.14826602e-01 -4.50914830e-01
2.36549377e-01 1.85210645e-01 -1.96577758e-01 -4.70995843e-01
3.00632715e-01 3.42622772e-02 2.57815987e-01 5.75338840e-01
2.44984582e-01 2.38899216e-02 -5.69009371e-02 4.68815476e-01
9.35166180e-01 5.55995226e-01 -2.16047734e-01 1.65164113e-01
2.96025127e-01 -1.54102147e-01 5.11183500e-01 7.01019645e-01
-2.51414061e-01 9.63070750e-01 4.48814362e-01 -6.29832298e-02
-9.89505827e-01 -1.46005177e+00 3.07439625e-01 1.21115959e+00
7.07173884e-01 -6.03035539e-02 -9.38965440e-01 -7.03091502e-01
-8.43225420e-02 5.89480042e-01 -8.79946828e-01 8.58917981e-02
-5.38123846e-01 -6.51782811e-01 4.36655372e-01 6.68809175e-01
7.17034221e-01 -9.56983268e-01 -6.68967903e-01 -3.24399233e-01
-3.81478727e-01 -1.17382789e+00 -4.85715568e-01 4.27027106e-01
-6.71340168e-01 -9.77923930e-01 -9.51133549e-01 -7.34262466e-01
8.23763847e-01 3.97638768e-01 9.84064162e-01 -1.26283586e-01
-8.17465261e-02 5.24704218e-01 -5.58335930e-02 -1.01252168e-01
-3.42025757e-01 -2.81269429e-03 -2.18901485e-01 3.56404990e-01
1.35335058e-01 -4.08241570e-01 -5.79558194e-01 1.18931748e-01
-1.00272453e+00 8.48834813e-01 8.72329712e-01 5.99593878e-01
8.31128716e-01 -3.04079682e-01 2.83551645e-02 -1.15131652e+00
-9.94781703e-02 -4.16779876e-01 -3.59043956e-01 6.36141777e-01
-1.46719337e-01 8.61831680e-02 -3.61371366e-03 -5.53073943e-01
-1.29861081e+00 5.68963170e-01 1.65537521e-01 -8.11854303e-01
-2.00780824e-01 5.22583686e-02 -6.31778657e-01 -1.82142109e-01
3.07030261e-01 4.49725121e-01 -2.25137919e-01 -3.09732050e-01
7.75981367e-01 4.66977060e-01 7.92635679e-01 -5.05447686e-01
8.03906083e-01 7.42239177e-01 -1.78435922e-01 -5.28985262e-01
-9.90775228e-01 -3.82054448e-01 -9.80404019e-01 -5.87688863e-01
1.36241364e+00 -1.08491349e+00 -5.84386945e-01 7.20345140e-01
-1.14514887e+00 -9.38659787e-01 -4.44527306e-02 5.90537749e-02
-5.04860878e-01 3.82559031e-01 -8.42828512e-01 -9.61941838e-01
-6.85706213e-02 -1.10260880e+00 1.33981502e+00 2.94648737e-01
2.65077472e-01 -9.51670468e-01 -1.86693326e-01 7.61993706e-01
2.20765635e-01 3.51586848e-01 4.78014886e-01 -1.24427602e-01
-1.08656859e+00 3.59309137e-01 -4.35446918e-01 2.35318884e-01
2.89456546e-01 -8.73741508e-02 -1.26517785e+00 -1.96537394e-02
-1.47235319e-01 -3.39288890e-01 1.21597743e+00 4.52678233e-01
1.19155955e+00 2.61174869e-02 -4.76084828e-01 8.67371500e-01
1.23815596e+00 1.14034139e-01 6.96992695e-01 3.30874801e-01
1.42652917e+00 2.82582670e-01 4.42203969e-01 1.12387918e-01
5.17743587e-01 2.20320374e-01 3.65505695e-01 -4.65783834e-01
-2.91082382e-01 -3.12594742e-01 5.03904462e-01 5.65998137e-01
-3.98940742e-02 -3.02981198e-01 -8.45156789e-01 3.44133943e-01
-2.00285912e+00 -7.95660257e-01 -2.58445479e-02 1.83863318e+00
9.94153023e-01 3.23373049e-01 9.76586640e-02 -1.42399162e-01
6.93952143e-01 2.26680823e-02 -8.29363704e-01 2.20484003e-01
-4.19600934e-01 -9.97883379e-02 5.57177067e-01 6.77629292e-01
-1.22733188e+00 1.16012704e+00 6.09981728e+00 4.86610383e-01
-9.86239076e-01 3.40813279e-01 8.60895932e-01 -1.74277395e-01
-5.61908960e-01 -1.08859882e-01 -5.32993078e-01 5.67614436e-01
6.08483553e-01 3.75357121e-01 6.43379688e-01 7.92064130e-01
-1.86182424e-01 -5.29602110e-01 -1.19255185e+00 1.12143421e+00
4.18079317e-01 -1.18382359e+00 2.94224937e-02 -1.32773697e-01
9.39505935e-01 -9.95991305e-02 5.20043746e-02 1.66230276e-02
1.38703883e-01 -8.43831539e-01 1.00551069e+00 7.09941328e-01
6.73772037e-01 -3.16832185e-01 4.79585588e-01 1.88500509e-01
-1.15770876e+00 2.32647568e-01 -3.17287408e-02 1.83040351e-01
5.31244278e-01 6.17579281e-01 -3.75773728e-01 4.91763383e-01
5.39630949e-01 6.17111921e-01 -7.69560397e-01 6.54628575e-01
-2.53585279e-01 5.14254928e-01 -1.71507120e-01 3.33349317e-01
-4.69548665e-02 -1.91949457e-01 2.17490658e-01 1.24514830e+00
-1.55639544e-01 3.29413600e-02 2.66936392e-01 1.05619490e+00
-1.18802385e-02 -3.67205232e-01 -1.94054991e-01 -1.29612699e-01
2.17473164e-01 1.15854311e+00 -1.08359063e+00 -6.93343282e-01
-5.08878529e-01 1.84820342e+00 1.34183630e-01 6.42417312e-01
-1.17471099e+00 2.80143823e-02 2.86201090e-01 1.54802442e-01
3.35190445e-01 -2.18590602e-01 -5.83883226e-01 -1.40471876e+00
-7.90287927e-02 -8.02957952e-01 1.03948705e-01 -1.04443264e+00
-1.14114606e+00 4.59328860e-01 5.78111000e-02 -9.03509736e-01
7.86081329e-02 -6.25501633e-01 -3.02853048e-01 7.70000458e-01
-1.31145251e+00 -1.72516716e+00 -6.70609176e-01 6.68434322e-01
4.92581606e-01 6.78336769e-02 2.44362190e-01 4.47106212e-01
-6.76983476e-01 3.81186128e-01 -2.16791406e-01 1.80814669e-01
8.09337378e-01 -1.54588771e+00 3.07940751e-01 1.08271682e+00
1.00351319e-01 3.45710069e-01 4.36935484e-01 -9.51250672e-01
-1.56676424e+00 -1.07879210e+00 3.14797014e-01 -6.96488500e-01
3.57687801e-01 -5.56993425e-01 -8.84956419e-01 1.00033891e+00
4.25285190e-01 8.57067704e-02 4.82735217e-01 -2.90372014e-01
-1.57349229e-01 7.12193698e-02 -8.77122760e-01 6.16981983e-01
1.01210105e+00 -5.77125490e-01 -3.29852730e-01 1.79431751e-01
8.44424486e-01 -8.79642725e-01 -4.53634351e-01 2.42045268e-01
4.38043058e-01 -9.84181821e-01 9.18034911e-01 -2.17440262e-01
6.64289057e-01 -6.79657400e-01 -2.82939017e-01 -8.99264097e-01
-4.24965471e-02 -2.96303600e-01 -6.25048161e-01 1.13074875e+00
3.22379023e-01 -2.00430050e-01 8.59562516e-01 7.59788632e-01
-8.41354281e-02 -7.35454679e-01 -6.34367824e-01 -2.59981990e-01
-1.48451254e-01 -3.38755608e-01 2.24846601e-01 7.48826563e-01
-3.36963862e-01 2.68241197e-01 -6.98729694e-01 3.73151213e-01
9.08800960e-01 3.17784518e-01 8.89372349e-01 -6.87040806e-01
-5.70172250e-01 -1.24744304e-01 -8.19927454e-02 -1.13983083e+00
4.45605442e-02 -6.34812117e-01 3.58046025e-01 -1.64219904e+00
7.13415802e-01 -1.34554788e-01 -2.41883531e-01 5.35410404e-01
-5.66067636e-01 2.86975622e-01 2.45744258e-01 4.32714447e-02
-1.07689309e+00 5.01902759e-01 1.35768557e+00 -3.16854894e-01
-1.23777628e-01 -1.36274740e-01 -3.83491367e-01 6.74178421e-01
5.51343143e-01 -3.23142886e-01 -4.36011165e-01 -6.11455441e-01
4.76845615e-02 1.02865390e-01 5.85143089e-01 -9.10958111e-01
2.58309573e-01 -4.04464155e-01 8.36122155e-01 -6.94864094e-01
5.39469242e-01 -6.28273308e-01 1.84058428e-01 2.19276413e-01
-1.90840408e-01 -1.36589736e-01 1.90322563e-01 6.24012530e-01
1.00541137e-01 1.12480847e-02 6.09444559e-01 -2.39998043e-01
-7.66395271e-01 1.26942396e-01 -3.58021468e-01 1.75885446e-02
9.62832868e-01 -1.51120946e-01 -2.42585927e-01 -2.79984742e-01
-6.54178679e-01 2.12132737e-01 8.34527433e-01 1.50101826e-01
7.84535229e-01 -1.09165120e+00 -3.12078089e-01 1.11148641e-01
-2.31069222e-01 2.32671663e-01 2.74152637e-01 8.93369198e-01
-6.77406907e-01 -2.09965706e-01 -2.07260728e-01 -8.26700628e-01
-1.19603217e+00 5.93547761e-01 5.56540847e-01 1.65948197e-01
-6.73786163e-01 1.02371943e+00 5.26352346e-01 -1.92583859e-01
5.30438781e-01 -3.62351239e-01 5.59879661e-01 -2.65476674e-01
5.49788356e-01 2.54717886e-01 -4.07437444e-01 -4.99408275e-01
-3.77617598e-01 3.80491823e-01 6.80088922e-02 -5.58156371e-01
9.01365340e-01 -3.44361693e-01 -1.66775659e-01 1.01167214e+00
9.22457159e-01 -2.04180017e-01 -1.73643601e+00 -1.41952783e-01
-3.32711339e-01 -6.07996047e-01 8.83289147e-03 -7.91025817e-01
-1.29771912e+00 9.99330461e-01 7.21216679e-01 -3.13693762e-01
1.22752345e+00 1.99196249e-01 8.70818019e-01 4.98821363e-02
2.13824421e-01 -9.93438423e-01 5.37057698e-01 1.27206087e-01
3.42183381e-01 -1.57767344e+00 2.46589966e-02 -5.32119274e-01
-1.17911017e+00 8.26043785e-01 1.02704823e+00 2.02057779e-01
5.11879265e-01 5.30247808e-01 4.42998201e-01 -1.44689932e-01
-5.43656826e-01 -4.12214637e-01 2.96805769e-01 4.79417622e-01
3.50677907e-01 2.79901437e-02 3.87688577e-01 4.51629639e-01
3.41960102e-01 -1.86802577e-02 3.54796857e-01 1.13323247e+00
-4.27324086e-01 -8.27274084e-01 -3.81142080e-01 3.27031523e-01
-1.44616589e-01 2.55522616e-02 -5.49720168e-01 4.15352672e-01
4.23616976e-01 6.88305438e-01 4.81023081e-02 -5.33562541e-01
-1.49269819e-01 1.12011507e-01 8.77266943e-01 -4.78834420e-01
-4.36545819e-01 3.13096702e-01 -3.76306355e-01 -5.86659431e-01
-7.24268079e-01 -6.43861711e-01 -1.30087829e+00 -2.14485377e-01
-2.32520849e-01 -3.70921433e-01 5.11182427e-01 1.10287356e+00
7.46719837e-02 9.22020495e-01 1.50071859e-01 -1.32575822e+00
2.24823058e-01 -8.60860705e-01 -4.53868628e-01 4.68950272e-01
4.57618326e-01 -6.28928661e-01 -3.53192091e-01 6.32636368e-01] | [10.6184720993042, -0.8966217637062073] |
f04d0123-2bb6-4547-bd16-190736aed9be | accounting-for-variations-in-speech-emotion | 2109.04316 | null | https://arxiv.org/abs/2109.04316v1 | https://arxiv.org/pdf/2109.04316v1.pdf | Accounting for Variations in Speech Emotion Recognition with Nonparametric Hierarchical Neural Network | In recent years, deep-learning-based speech emotion recognition models have outperformed classical machine learning models. Previously, neural network designs, such as Multitask Learning, have accounted for variations in emotional expressions due to demographic and contextual factors. However, existing models face a few constraints: 1) they rely on a clear definition of domains (e.g. gender, noise condition, etc.) and the availability of domain labels; 2) they often attempt to learn domain-invariant features while emotion expressions can be domain-specific. In the present study, we propose the Nonparametric Hierarchical Neural Network (NHNN), a lightweight hierarchical neural network model based on Bayesian nonparametric clustering. In comparison to Multitask Learning approaches, the proposed model does not require domain/task labels. In our experiments, the NHNN models generally outperform the models with similar levels of complexity and state-of-the-art models in within-corpus and cross-corpus tests. Through clustering analysis, we show that the NHNN models are able to learn group-specific features and bridge the performance gap between groups. | ['Emily Mower Provost', 'Amrit Romana', 'Lance Ying'] | 2021-09-09 | null | null | null | null | ['cross-corpus'] | ['computer-vision'] | [-2.39530846e-01 -1.17440417e-01 -3.23302478e-01 -9.40522611e-01
-7.22645879e-01 -2.17897117e-01 4.94438976e-01 1.86886430e-01
-5.65494895e-01 6.83789134e-01 2.37227678e-01 1.55459568e-01
-3.74641269e-01 -2.27038756e-01 -2.07937390e-01 -6.21843874e-01
-7.27417544e-02 6.41177893e-01 -1.87440708e-01 4.08348292e-02
-1.86189171e-02 1.09259516e-01 -1.82823372e+00 3.34821314e-01
9.54406202e-01 1.41312456e+00 -3.88003625e-02 1.45839885e-01
-3.51490080e-01 7.61985838e-01 -8.67766619e-01 -3.21927786e-01
-2.55722582e-01 -2.80370504e-01 -6.75058007e-01 7.18672760e-03
1.78921055e-02 -2.30499338e-02 3.13856006e-02 9.76722479e-01
7.61862814e-01 4.99744952e-01 1.11494315e+00 -1.55817175e+00
-7.99761653e-01 6.91654265e-01 -3.45851958e-01 -2.78651386e-01
-3.36084701e-02 -6.33511364e-01 9.24265921e-01 -9.56310093e-01
3.11716437e-01 1.26648390e+00 8.36789787e-01 8.31171811e-01
-1.20211196e+00 -9.04561341e-01 3.12472641e-01 3.97288740e-01
-1.44528961e+00 -5.14807045e-01 1.14432585e+00 -5.50985634e-01
8.56022716e-01 5.14895543e-02 5.47106043e-02 1.74073637e+00
-2.00312197e-01 7.91873872e-01 1.47212160e+00 -4.95558619e-01
5.96329570e-01 3.46724480e-01 2.49045283e-01 3.29517961e-01
-3.18878978e-01 -1.26370177e-01 -7.71279693e-01 -3.71547043e-01
3.30519587e-01 -1.93446457e-01 1.06567733e-01 -5.01950562e-01
-7.03051150e-01 1.19420910e+00 -1.59960032e-01 6.85973585e-01
-4.08506691e-01 -2.01213121e-01 8.64894390e-01 2.15209514e-01
9.42428827e-01 8.25510100e-02 -8.32441211e-01 -4.16124254e-01
-1.06525946e+00 -1.43185839e-01 9.97466981e-01 6.98884785e-01
8.55114400e-01 1.92916691e-01 -7.72102252e-02 1.62169147e+00
2.07703456e-01 -5.30849863e-03 9.53888893e-01 -7.22338438e-01
1.42485723e-01 2.37069249e-01 -2.74662584e-01 -9.07507718e-01
-6.64266765e-01 -1.51182741e-01 -1.14438832e+00 -1.78165555e-01
2.23162025e-01 -5.01481593e-01 -7.97233880e-01 2.04873919e+00
-6.97265496e-04 8.82641003e-02 1.86808899e-01 5.60143828e-01
1.09439635e+00 2.83057719e-01 5.21744549e-01 -3.69055390e-01
1.09690392e+00 -8.30618262e-01 -1.09466088e+00 -1.96743712e-01
5.83320200e-01 -4.39656496e-01 1.03421664e+00 6.55398011e-01
-8.74667287e-01 -7.45573401e-01 -6.39442801e-01 1.29202619e-01
-6.77245557e-01 3.63014758e-01 7.88531601e-01 1.31925201e+00
-1.05536318e+00 2.14307487e-01 -5.38118958e-01 -3.91083211e-01
1.73482820e-01 5.41413903e-01 -3.82094085e-01 2.50884026e-01
-1.49274313e+00 7.19827294e-01 5.53328216e-01 3.63413431e-02
-7.95669198e-01 -3.78995776e-01 -9.00905848e-01 2.11880967e-01
2.74905980e-01 -4.54597414e-01 1.19871294e+00 -1.55844223e+00
-2.03407288e+00 7.84473181e-01 -1.95760190e-01 -3.15566659e-01
1.52980417e-01 -1.14705995e-01 -6.18481338e-01 5.82937337e-02
-1.20185018e-01 6.38427317e-01 8.16024423e-01 -1.30779839e+00
-4.01016384e-01 -3.78176033e-01 -5.07102847e-01 1.27956206e-02
-8.98166716e-01 5.05512178e-01 -2.97958136e-01 -7.00400293e-01
-1.56418711e-01 -7.44132936e-01 -1.27380997e-01 -5.29858053e-01
-1.82069749e-01 -7.29639888e-01 9.69880939e-01 -4.33006823e-01
1.14488089e+00 -2.32443333e+00 1.69700176e-01 2.75500387e-01
8.21043912e-04 7.79953897e-02 6.09819889e-02 2.21766084e-01
-1.46160528e-01 3.03365201e-01 -5.72523586e-02 -9.23937559e-01
4.28270161e-01 4.03752834e-01 4.81441915e-02 3.32609892e-01
6.17784522e-02 6.64223611e-01 -4.53283995e-01 -6.91691697e-01
9.71601158e-02 4.47869003e-01 -4.18874115e-01 2.16346711e-01
1.81840211e-02 2.52501458e-01 -2.07601666e-01 4.95603621e-01
5.23896635e-01 6.28432408e-02 4.69829589e-01 5.31968251e-02
6.99692890e-02 3.28376144e-02 -1.09100926e+00 1.53697979e+00
-6.13089621e-01 5.22109091e-01 4.81084138e-01 -1.63370991e+00
1.35777092e+00 6.18406117e-01 5.92432439e-01 -4.68976945e-01
2.93215543e-01 1.24860778e-01 -5.48653640e-02 -3.27331364e-01
4.04788703e-01 -4.27134782e-01 -5.96393943e-01 1.60583615e-01
5.28312325e-01 5.46858050e-02 -2.48674527e-01 -3.34265620e-01
6.78767502e-01 -2.03393906e-01 3.71671438e-01 -3.63647670e-01
4.16058391e-01 -5.37495196e-01 1.11413181e+00 7.61793554e-01
-5.26022077e-01 4.10766482e-01 5.79933047e-01 -2.21140131e-01
-5.59957087e-01 -9.41402614e-01 -2.70858735e-01 1.78500271e+00
-2.79271543e-01 -2.16665447e-01 -6.67538881e-01 -7.21326947e-01
-2.15787664e-01 5.70155084e-01 -8.88353646e-01 -1.74599215e-01
5.12051247e-02 -8.00996959e-01 9.08883452e-01 5.65162599e-01
4.91531998e-01 -1.09148264e+00 -1.34305686e-01 1.13862157e-01
-3.58684301e-01 -1.28161538e+00 1.16002388e-01 6.87291443e-01
-5.99920273e-01 -3.99772048e-01 -6.00521743e-01 -9.51446652e-01
2.68037677e-01 -3.57310534e-01 1.19228530e+00 -3.07742804e-01
1.14947468e-01 5.20766675e-01 -6.79450691e-01 -6.66903436e-01
-2.22892672e-01 2.67963260e-01 3.80852193e-01 4.45173651e-01
7.92734921e-01 -7.99441516e-01 -1.54948952e-02 4.13388401e-01
-7.87153125e-01 -3.36718589e-01 5.77553570e-01 1.14179051e+00
2.94401556e-01 3.70332718e-01 1.18997574e+00 -7.41435587e-01
1.13166904e+00 -6.45701647e-01 7.52457767e-04 2.80240804e-01
-3.75778258e-01 -1.93064138e-01 4.71986949e-01 -7.38333762e-01
-1.29031587e+00 3.88082042e-02 -1.80059299e-01 -6.19560063e-01
-8.82320404e-01 7.06974745e-01 -3.77887368e-01 1.19598478e-01
4.68567103e-01 2.39343904e-02 -9.08397138e-02 -4.41895306e-01
2.20310628e-01 1.00476801e+00 1.98223174e-01 -1.02156675e+00
2.06083909e-01 1.21597067e-01 -4.59672600e-01 -1.01265764e+00
-7.38013566e-01 -5.53811431e-01 -8.32259655e-01 -3.17251533e-01
1.08818364e+00 -1.06316972e+00 -7.34686732e-01 5.44861615e-01
-1.01261663e+00 -5.30560315e-01 1.09791048e-01 6.64669096e-01
-6.64521039e-01 4.26419616e-01 -6.08053684e-01 -1.29399419e+00
-6.21165931e-02 -9.77686942e-01 9.17115450e-01 -7.62902666e-03
-4.83340561e-01 -1.30163145e+00 -1.70274436e-01 3.15731108e-01
4.63724047e-01 1.75659359e-01 1.03608143e+00 -1.26061654e+00
5.24806082e-01 2.84452766e-01 1.50497239e-02 6.97234631e-01
1.11365035e-01 -1.33118957e-01 -1.20838702e+00 4.87017930e-02
1.51078522e-01 -8.96674216e-01 8.44681203e-01 6.78358912e-01
1.63887084e+00 5.91878174e-03 -1.27876043e-01 3.02812696e-01
8.73872936e-01 2.85976022e-01 4.00767535e-01 1.93082035e-01
3.67812485e-01 1.16538787e+00 4.50699627e-01 6.37125373e-01
6.08630598e-01 5.89920998e-01 1.75653681e-01 -1.48838565e-01
4.55778986e-01 1.46829471e-01 3.61385375e-01 9.88216758e-01
-4.19280957e-03 -3.76150711e-03 -1.07522070e+00 5.98003507e-01
-2.07933307e+00 -8.62579465e-01 1.14194982e-01 1.83057582e+00
7.75325954e-01 -4.44451384e-02 4.53542203e-01 4.54232469e-03
7.55917490e-01 6.83136806e-02 -4.51945454e-01 -7.10558534e-01
-3.47714603e-01 2.96362579e-01 7.94330686e-02 1.56250536e-01
-1.37979460e+00 9.68056202e-01 6.44037914e+00 1.18407536e+00
-1.05232882e+00 3.68627757e-01 9.08437014e-01 -6.34629726e-02
-7.13892281e-03 -6.28476560e-01 -7.13807940e-01 3.85312438e-01
1.25221944e+00 1.28030747e-01 1.71815574e-01 1.11836052e+00
2.63747689e-03 5.91650233e-02 -9.12762940e-01 1.15623271e+00
3.41451347e-01 -6.64971411e-01 -3.52117240e-01 -5.14984550e-03
5.80235064e-01 -1.38751015e-01 2.54295588e-01 1.07825089e+00
4.54791129e-01 -1.21768939e+00 5.38331449e-01 3.53417337e-01
5.50802231e-01 -1.12252951e+00 8.46126914e-01 4.99414831e-01
-7.03441620e-01 -3.32409590e-01 -2.63028413e-01 -1.55555010e-01
-3.05456053e-02 6.34713352e-01 -4.88971025e-01 6.38211608e-01
1.15931094e+00 3.01430136e-01 -3.08743060e-01 4.59430635e-01
-3.09163891e-02 8.28162551e-01 -1.45185143e-01 -2.84022540e-01
2.66966283e-01 -2.47069910e-01 -6.61544204e-02 1.45006657e+00
3.95389736e-01 -1.88019186e-01 3.60496253e-01 5.63946784e-01
-7.98299462e-02 5.37509322e-01 -4.09206241e-01 -1.97704528e-02
4.79460686e-01 1.32916534e+00 -5.41679442e-01 -2.10698336e-01
-6.39688194e-01 9.16845798e-01 6.27681017e-01 4.99942511e-01
-8.10362101e-01 -3.95815432e-01 7.94995785e-01 -4.30147469e-01
4.67832863e-01 -3.00941914e-01 -3.50443244e-01 -9.52809930e-01
-2.22969264e-01 -9.59971309e-01 3.83617967e-01 -5.04147828e-01
-1.82713008e+00 5.82853436e-01 6.53531924e-02 -8.22164595e-01
-5.16752779e-01 -7.61044145e-01 -3.43102515e-01 5.78251839e-01
-1.29410005e+00 -1.22348237e+00 -7.41764382e-02 8.40089262e-01
4.16817129e-01 -5.25427580e-01 1.23735392e+00 4.39989209e-01
-6.09025717e-01 7.63517499e-01 2.85590321e-01 4.68474507e-01
1.08109951e+00 -1.23678768e+00 -3.27513665e-01 2.22737893e-01
1.47786245e-01 5.99741817e-01 4.89993036e-01 -2.89076865e-01
-6.38415873e-01 -8.65669727e-01 8.70188653e-01 -1.19830929e-01
6.84621930e-01 -1.01534557e+00 -9.34421599e-01 5.49734056e-01
3.48953605e-01 -1.09829873e-01 1.60749006e+00 1.03705931e+00
-6.03387356e-01 -2.60161757e-02 -1.14644992e+00 2.93542176e-01
7.03749299e-01 -6.86959863e-01 -5.92587411e-01 8.35535526e-02
4.73166466e-01 7.25684613e-02 -1.12754929e+00 4.38017964e-01
6.16801500e-01 -1.25353920e+00 7.32136607e-01 -6.80659056e-01
2.25859091e-01 2.77796805e-01 -3.12738001e-01 -1.56675017e+00
-3.88824880e-01 -4.83460933e-01 -2.80501433e-02 1.60942888e+00
4.33836222e-01 -4.90246534e-01 7.49309540e-01 8.13276291e-01
-1.80172667e-01 -6.68810070e-01 -1.32625437e+00 -9.06971395e-01
4.38161850e-01 -6.90093040e-01 4.73537773e-01 1.31304574e+00
2.03198224e-01 4.73416388e-01 -6.16676211e-01 1.03933942e-02
3.06562245e-01 -2.21996889e-01 5.72782576e-01 -1.80057442e+00
-1.90332949e-01 -6.43947363e-01 -2.03874663e-01 -6.75054252e-01
1.13445163e+00 -5.17186821e-01 9.33656842e-02 -1.18606329e+00
1.61097154e-01 -3.58537763e-01 -5.06636620e-01 6.45707488e-01
-6.25723824e-02 -8.33906000e-04 -1.24685682e-01 -1.70027286e-01
-8.13984632e-01 9.17491615e-01 2.45688587e-01 6.89132735e-02
-3.80433589e-01 -9.85736549e-02 -7.57557631e-01 8.37605357e-01
1.14083064e+00 -4.12483394e-01 -3.40547591e-01 -1.01089366e-01
-4.23583575e-02 -9.80070308e-02 9.23585594e-02 -9.02391136e-01
3.68008256e-01 1.05617684e-03 4.46640790e-01 -4.49894190e-01
7.07119882e-01 -8.62313330e-01 -1.75682515e-01 -1.74791455e-01
-3.27146173e-01 -2.61300504e-01 3.11622083e-01 5.13217330e-01
-6.50764942e-01 -2.07722366e-01 5.13474166e-01 8.73978809e-02
-7.96888053e-01 1.03343718e-01 -8.74921083e-01 -3.04468386e-02
9.23469782e-01 -2.62472570e-01 -3.14806402e-02 -6.66844904e-01
-9.29312468e-01 4.93453406e-02 -1.56347126e-01 7.42802560e-01
3.68458450e-01 -1.42252898e+00 -5.53136587e-01 4.80423309e-02
1.90319657e-01 -3.72041702e-01 6.34740531e-01 8.38874221e-01
5.22743523e-01 3.61837655e-01 -2.29284212e-01 -5.34183860e-01
-1.50036013e+00 4.69510436e-01 2.68424720e-01 -4.06846762e-01
2.99870104e-01 7.92010009e-01 3.47188562e-01 -1.05172133e+00
5.06689787e-01 -1.45359427e-01 -3.40887964e-01 6.54783189e-01
7.93392658e-02 2.26899803e-01 1.24337422e-02 -7.18555629e-01
-3.83809894e-01 3.38924170e-01 7.38858059e-02 -2.99701959e-01
1.43883801e+00 -6.28926232e-02 -1.30175501e-01 8.65854323e-01
1.18530619e+00 -3.05655777e-01 -8.83384645e-01 -3.03127557e-01
3.13920140e-01 1.00004204e-01 2.13209003e-01 -8.52077007e-01
-7.72452712e-01 9.89520907e-01 3.93227696e-01 2.30604261e-01
1.41077375e+00 1.49380326e-01 2.18284369e-01 3.24475259e-01
2.42453992e-01 -1.80040169e+00 2.26180404e-01 7.83537924e-01
6.23773813e-01 -1.33313501e+00 -3.95140558e-01 -2.64844537e-01
-9.39506471e-01 1.05637562e+00 6.97750330e-01 4.57848668e-01
9.03276324e-01 1.94521800e-01 3.13769847e-01 1.33755570e-02
-7.60335565e-01 -2.88698524e-01 2.94212669e-01 9.92075503e-01
5.94331563e-01 2.07216129e-01 -1.60077766e-01 1.33601892e+00
-2.56072015e-01 -1.70061767e-01 7.70015782e-03 5.08879900e-01
-2.89610445e-01 -1.22950721e+00 -3.84517521e-01 2.69189388e-01
-6.32789493e-01 2.70106038e-03 -5.95150769e-01 8.46362472e-01
2.31178731e-01 1.34319484e+00 8.36276412e-02 -4.71659094e-01
2.90554821e-01 6.51169538e-01 1.36038870e-01 -5.46157181e-01
-7.70344496e-01 1.59369901e-01 3.83757859e-01 -2.09934115e-01
-8.06106329e-01 -8.08858633e-01 -8.77577960e-01 1.56512216e-01
-2.47465327e-01 4.58859712e-01 7.35296965e-01 9.27619636e-01
4.64273334e-01 3.91783625e-01 5.10601342e-01 -8.70990157e-01
-3.50587338e-01 -1.35028839e+00 -1.00038791e+00 4.46853846e-01
1.85199053e-04 -9.50148046e-01 -4.42830831e-01 -4.67731617e-02] | [13.39714527130127, 5.701663970947266] |
fb46f250-b293-42c1-94c3-251f36875bd3 | scene-recognition-based-on-dnn-and-game | 1912.01293 | null | https://arxiv.org/abs/1912.01293v4 | https://arxiv.org/pdf/1912.01293v4.pdf | Scene recognition based on DNN and game theory with its applications in human-robot interaction | Scene recognition model based on the DNN and game theory with its applications in human-robot interaction is proposed in this paper. The use of deep learning methods in the field of scene recognition is still in its infancy, but has become an important trend in the future. As the innovative idea of the paper, we propose the following novelties. (1) In this paper, the image registration problem is transformed into a problem of minimum energy in Markov Random Field to finalize the image pre-processing task. Game theory is used to find the optimal. (2) We select neighboring homogeneous sample features and the neighboring heterogeneous sample features for the extracted sample features to build a triple and modify the traditional neural network to propose the novel DNN for scene understanding. (3) The robot control is well combined to guide the robot vision for multiple tasks. The experiment is then conducted to validate the overall performance. | ['G. H. Chen', 'D. Z. Zhao', 'W. Z. Wang', 'R. Q. Wang', 'D. S. Luo'] | 2019-12-03 | null | null | null | null | ['scene-recognition'] | ['computer-vision'] | [ 2.50977725e-01 -1.60848737e-01 1.66384764e-02 -3.19965392e-01
1.52466968e-01 6.52840361e-02 3.91304910e-01 -3.09404403e-01
-7.82341361e-01 4.98414576e-01 -1.32123098e-01 1.21356942e-01
-4.47253138e-01 -8.96044970e-01 -3.03564250e-01 -9.51746166e-01
2.65691191e-01 2.52323091e-01 3.19947302e-01 -3.16365272e-01
6.52966917e-01 4.69308436e-01 -1.41822994e+00 -2.23332360e-01
8.82000983e-01 1.08548546e+00 9.82818127e-01 3.16896260e-01
-2.20611840e-01 1.21806157e+00 -2.37742469e-01 1.77893519e-01
4.71745878e-01 -6.48452580e-01 -8.14964950e-01 2.83630967e-01
-3.87847573e-01 -1.72066197e-01 -3.13667685e-01 1.44743729e+00
6.51299894e-01 5.71899652e-01 6.33900046e-01 -1.14579022e+00
-4.46332127e-01 1.59500510e-01 -7.55624533e-01 3.20190974e-02
7.05174506e-02 9.08398703e-02 3.89804810e-01 -6.31212354e-01
5.54674923e-01 1.31465399e+00 2.70192355e-01 5.73327839e-01
-5.32344639e-01 -5.57656407e-01 1.97563265e-02 7.84523487e-01
-1.52719164e+00 -1.56188294e-01 1.08674526e+00 -3.08667064e-01
7.81246960e-01 -1.61678717e-01 7.48545468e-01 6.67098403e-01
3.37059408e-01 7.23003805e-01 1.01648200e+00 -5.44835448e-01
2.69544601e-01 -9.50656980e-02 2.61239201e-01 7.09469855e-01
2.62565017e-01 2.63734251e-01 8.75167269e-03 5.45194149e-01
1.09173822e+00 3.48865390e-01 -9.47007462e-02 -4.64109898e-01
-1.11012626e+00 1.03035128e+00 6.72313273e-01 7.11008728e-01
-5.80090284e-01 -9.08605456e-02 3.93402070e-01 -3.67170200e-03
-3.32325585e-02 2.82713830e-01 -4.73177582e-02 -8.48358497e-02
-5.25786340e-01 2.35246271e-01 5.27231038e-01 6.51741385e-01
9.23664808e-01 2.34032094e-01 1.62431717e-01 9.06163990e-01
4.17070836e-01 3.52106124e-01 5.73824406e-01 -1.01499867e+00
2.23867655e-01 7.41630912e-01 2.12042406e-02 -1.45644641e+00
-6.35950685e-01 -2.06175700e-01 -1.25179362e+00 4.70449656e-01
1.09128833e-01 -3.98480207e-01 -9.41844702e-01 1.32676423e+00
3.30240160e-01 3.43867689e-01 3.83006305e-01 1.07716918e+00
8.25925291e-01 8.57153773e-01 -5.12620173e-02 -2.94242114e-01
1.24734509e+00 -1.05507338e+00 -8.79510105e-01 -2.09032804e-01
4.68033761e-01 -5.63266575e-01 7.21040905e-01 4.28758234e-01
-5.27646720e-01 -9.34612811e-01 -1.12508631e+00 1.19736791e-01
-4.39641625e-01 1.24967448e-01 7.48644829e-01 4.01259124e-01
-1.03550529e+00 2.84578860e-01 -6.52039587e-01 -7.75965929e-01
1.10093258e-01 6.15412533e-01 -3.12894911e-01 -2.09902361e-01
-1.11089301e+00 1.11689448e+00 9.74167228e-01 5.16987085e-01
-6.01861298e-01 2.01674700e-01 -1.00362802e+00 -1.26542673e-01
4.32961553e-01 -6.50547445e-01 1.12341702e+00 -1.21704972e+00
-1.74968112e+00 6.71840191e-01 -5.45613393e-02 -2.96486080e-01
1.54563099e-01 4.11488488e-02 -2.45261222e-01 1.88757572e-02
9.63961612e-03 7.40027070e-01 3.60486537e-01 -1.23484981e+00
-8.68750632e-01 -2.78980315e-01 9.49611589e-02 8.08791935e-01
-1.11184157e-01 4.64624278e-02 -3.88574809e-01 -3.64276052e-01
2.55862296e-01 -6.09143019e-01 -7.48793185e-01 -4.30677116e-01
-2.28416070e-01 -4.05935138e-01 8.85150969e-01 -3.45930696e-01
7.71870494e-01 -2.03431082e+00 1.85804799e-01 2.03002110e-01
8.52432027e-02 1.82731718e-01 -4.81312275e-02 1.57856256e-01
3.30608264e-02 -3.17448407e-01 -3.12065005e-01 -7.34657096e-03
-2.55663484e-01 2.71152884e-01 2.45077550e-01 3.43308836e-01
-2.42553968e-02 7.30715752e-01 -8.12146962e-01 -7.91149497e-01
7.75917351e-01 9.07495320e-02 -3.75614405e-01 1.73941404e-01
1.31421462e-01 5.88378847e-01 -6.86790287e-01 2.08309785e-01
9.42237318e-01 3.37182321e-02 -5.73746711e-02 -3.29223633e-01
-3.28036755e-01 -3.19811225e-01 -1.52194285e+00 1.64066613e+00
-1.39977574e-01 4.97646153e-01 2.58257121e-01 -1.54143071e+00
1.26845086e+00 -6.94060847e-02 6.27620697e-01 -9.40862834e-01
6.43419027e-01 1.23685421e-02 2.55346060e-01 -8.03409159e-01
5.89210808e-01 -1.98388621e-01 -8.43347609e-02 7.53755495e-02
2.52492595e-02 -1.82733729e-01 2.96992511e-02 -2.26351634e-01
7.75824249e-01 7.21628517e-02 5.52083194e-01 -2.07552224e-01
7.30382919e-01 3.64536166e-01 7.87553251e-01 5.64709842e-01
-3.84172380e-01 2.52756417e-01 2.93018483e-02 -3.99283141e-01
-8.09320927e-01 -6.73922956e-01 2.43935689e-01 4.00444746e-01
8.80011141e-01 2.98299700e-01 -8.81766796e-01 -4.58935976e-01
-6.64153278e-01 5.47969103e-01 -2.75220275e-01 -3.17979157e-01
-4.84579027e-01 -6.26766145e-01 9.79618654e-02 3.78340423e-01
1.40114987e+00 -1.73161709e+00 -7.49627054e-01 1.61442980e-01
-1.22824863e-01 -8.92900586e-01 -8.63574147e-02 2.57147253e-01
-6.16629303e-01 -1.08601534e+00 -6.85214221e-01 -1.46030223e+00
6.42920792e-01 5.63748181e-01 3.75128299e-01 -2.25807335e-02
-1.12740502e-01 3.10025454e-01 -5.92677236e-01 -4.06779259e-01
-5.92602752e-02 -1.28471583e-01 9.97982994e-02 9.39621404e-02
4.89940405e-01 -4.74242657e-01 -4.21881646e-01 2.10253090e-01
-8.06798100e-01 2.15019077e-01 6.96886361e-01 8.62249553e-01
5.79429984e-01 6.56544566e-01 3.57701391e-01 -5.06542802e-01
7.19812036e-01 -3.12880427e-01 -7.22559810e-01 -2.84753181e-02
-3.24964583e-01 -1.45368904e-01 7.35981584e-01 -3.13131422e-01
-1.34263384e+00 3.94648254e-01 -2.01032877e-01 -2.43621916e-01
-4.79839474e-01 7.21997917e-01 -4.74954247e-01 -2.66185671e-01
1.96722552e-01 5.23048460e-01 1.25686988e-01 -3.13134134e-01
2.39278346e-01 8.01725626e-01 4.84252840e-01 -1.10929646e-01
4.71028715e-01 3.08345139e-01 6.35122210e-02 -9.31249976e-01
-3.31040323e-01 -5.27140439e-01 -7.67005265e-01 -4.71615523e-01
1.31297517e+00 -6.14707708e-01 -1.19173205e+00 8.45507205e-01
-1.44946218e+00 -4.85240698e-01 -1.56248257e-01 9.24404442e-01
-7.03831255e-01 6.07634008e-01 -4.20649439e-01 -9.57512617e-01
-1.72775418e-01 -1.24900866e+00 6.07543528e-01 7.79117167e-01
2.78959394e-01 -8.57209682e-01 1.05046868e-01 1.13839723e-01
-9.45916399e-03 -4.83730286e-02 5.11757195e-01 -5.80451667e-01
-6.69183612e-01 5.28259650e-02 -4.24072772e-01 4.05958533e-01
2.17495337e-01 -2.18578994e-01 -7.04534769e-01 1.75510243e-01
6.20923936e-01 -1.50469348e-01 7.89983988e-01 6.49687648e-01
9.43659008e-01 1.08576017e-02 -3.59226614e-01 4.88169283e-01
1.66534865e+00 1.10028601e+00 9.61493552e-01 7.15271473e-01
8.22088599e-01 8.39962721e-01 1.02924061e+00 3.07743102e-01
3.68421346e-01 5.32974958e-01 6.07175946e-01 -1.17005087e-01
2.50355750e-01 -8.31743032e-02 3.84613335e-01 1.06166613e+00
1.99183574e-04 -2.07418784e-01 -9.21045542e-01 3.79416883e-01
-2.18584681e+00 -1.19762993e+00 -1.86659247e-01 1.82927942e+00
1.56079605e-01 1.14488140e-01 -2.02143803e-01 -3.35070640e-02
9.77748454e-01 2.36992478e-01 -4.59168077e-01 -3.20382088e-01
-9.59858373e-02 -2.85894629e-02 4.08630461e-01 4.57901627e-01
-1.16390693e+00 1.19357836e+00 5.45607710e+00 1.17414272e+00
-1.30182457e+00 -1.64515465e-01 4.69836861e-01 5.36502957e-01
3.08958769e-01 -1.07580656e-02 -6.43616080e-01 4.65214342e-01
3.27014953e-01 -1.59001611e-02 5.80040097e-01 9.81952548e-01
5.94078958e-01 -5.14047384e-01 -4.51641530e-01 1.53574574e+00
1.40387625e-01 -1.11302245e+00 9.42679960e-03 5.20971306e-02
5.32107592e-01 -2.99735945e-02 -9.12443101e-02 2.88336098e-01
4.74431038e-01 -9.19544935e-01 5.43692827e-01 8.01851153e-01
2.42836744e-01 -9.23769057e-01 1.13847518e+00 7.26224303e-01
-1.24303067e+00 -2.95680255e-01 -7.60898411e-01 -4.30433482e-01
2.34194189e-01 3.56343299e-01 -6.64586604e-01 8.98952067e-01
5.37996590e-01 9.82992113e-01 -3.19970876e-01 1.33495879e+00
-4.57290784e-02 1.92711845e-01 -2.76361972e-01 -5.05679846e-01
3.86196822e-01 -6.37596250e-01 3.58066380e-01 8.53093445e-01
3.62811983e-01 1.61240801e-01 3.92404705e-01 9.36834991e-01
2.46346936e-01 3.36596161e-01 -8.77463400e-01 3.15515310e-01
2.97581583e-01 1.37478924e+00 -1.15554368e+00 -3.47940236e-01
-2.53609836e-01 1.17758965e+00 7.84586444e-02 2.84463108e-01
-6.64515615e-01 -8.62344503e-01 1.07254587e-01 -4.43570167e-01
2.67787814e-01 -3.44737381e-01 -4.30239886e-01 -9.46131945e-01
-6.80883750e-02 -6.16849840e-01 1.36663929e-01 -8.79938781e-01
-1.08926439e+00 6.10960245e-01 1.08719403e-02 -1.36441958e+00
-1.64649740e-01 -8.06519032e-01 -9.37995374e-01 8.57463717e-01
-1.34735036e+00 -9.21655297e-01 -5.28759718e-01 8.14403951e-01
6.83715463e-01 -2.99249142e-01 6.31314635e-01 2.60743797e-01
-6.78797960e-01 -1.43327098e-02 3.25778872e-01 3.56029868e-01
2.35806003e-01 -8.04341674e-01 3.55698764e-02 9.99413788e-01
-2.47607112e-01 3.40389788e-01 4.16761100e-01 -7.85305977e-01
-1.05692756e+00 -1.00029862e+00 5.99428892e-01 2.25536749e-01
3.81142467e-01 -1.20754600e-01 -5.16167939e-01 2.36218959e-01
3.93083423e-01 -3.29882622e-01 8.50966647e-02 -4.36644763e-01
5.11312187e-01 -2.91928828e-01 -1.17510486e+00 7.97061622e-01
8.10122311e-01 -2.06772700e-01 -6.15808904e-01 6.01534992e-02
6.61749840e-01 -1.84583247e-01 -4.39674079e-01 3.66510451e-01
2.51205385e-01 -9.02777374e-01 6.28203988e-01 -4.61523160e-02
9.72092617e-03 -6.19281054e-01 -1.06267340e-01 -1.13050020e+00
-5.37713587e-01 -4.22648132e-01 5.99237204e-01 1.25282383e+00
-1.24250419e-01 -6.21153831e-01 1.05584383e+00 3.17488015e-01
-2.06003681e-01 -6.01830304e-01 -8.89307201e-01 -6.38633847e-01
-3.47417027e-01 -4.57917273e-01 1.96599543e-01 7.25225985e-01
-3.17244381e-02 5.55098653e-01 -3.58202547e-01 -1.11437485e-01
5.56454539e-01 -7.65179247e-02 7.45302141e-01 -1.24432778e+00
-2.48849462e-03 -3.78944427e-01 -6.20919168e-01 -1.43300223e+00
2.55866200e-01 -5.67165256e-01 4.77239966e-01 -1.91459715e+00
4.11014169e-01 -4.18848246e-01 -3.91692877e-01 1.63832992e-01
3.38070430e-02 -2.33700365e-01 3.85038167e-01 1.49327353e-01
-7.78142631e-01 9.18866396e-01 1.44809282e+00 -2.54816294e-01
-2.72022575e-01 2.14399725e-01 -6.29335344e-01 7.24404693e-01
1.06157684e+00 -2.45298862e-01 -5.96280456e-01 -1.06142476e-01
-2.49274641e-01 1.52715713e-01 3.53602886e-01 -1.34048235e+00
5.71636558e-01 -4.21628326e-01 3.46684396e-01 -7.30855882e-01
3.99466485e-01 -1.22544837e+00 -2.32191816e-01 6.64837241e-01
1.26382157e-01 6.57199696e-02 6.09560497e-02 6.66825593e-01
-3.77515763e-01 -5.95141947e-01 8.27352345e-01 -3.21349531e-01
-1.49192858e+00 2.04497159e-01 -6.07925177e-01 -3.12927634e-01
1.27929521e+00 -8.45046639e-01 7.79420286e-02 -4.53789055e-01
-3.64585757e-01 2.56094158e-01 2.47642100e-01 2.78326690e-01
8.78006995e-01 -1.29964995e+00 -1.91027984e-01 7.48393536e-02
-2.67483592e-01 2.13818580e-01 6.56885207e-01 6.82148993e-01
-8.26465905e-01 3.66963595e-01 -6.54937685e-01 -5.04726410e-01
-1.13630033e+00 7.69830704e-01 4.96554255e-01 -3.33246529e-01
-3.29963207e-01 5.78232229e-01 3.91349763e-01 -4.89964992e-01
2.47882485e-01 -1.83916837e-01 -8.68753493e-01 -1.67852134e-01
1.80725038e-01 3.05465788e-01 -3.94467741e-01 -8.95562530e-01
-2.90204048e-01 7.97990322e-01 7.13950098e-02 -1.67415932e-01
1.31765699e+00 -4.37466919e-01 -4.40239161e-01 3.32569152e-01
1.23440194e+00 -4.98867482e-01 -1.02030504e+00 -6.71138987e-02
8.63158703e-02 -2.14270856e-02 9.33980793e-02 -4.64450657e-01
-1.14930713e+00 9.89754200e-01 1.03386056e+00 -8.15973580e-02
1.27617872e+00 -3.74160796e-01 8.38394761e-01 6.06417537e-01
6.26362801e-01 -1.13215029e+00 -5.45288622e-02 9.63743091e-01
5.02935946e-01 -1.22409415e+00 -3.26454997e-01 -2.36678764e-01
-7.80607641e-01 1.26388228e+00 9.78808522e-01 -4.11199242e-01
8.03150892e-01 1.49813756e-01 3.06040403e-02 -2.85496861e-01
-1.05845138e-01 -5.42714953e-01 -7.95513093e-02 7.97810912e-01
-2.15330366e-02 -2.22279310e-01 -5.20298183e-01 7.44615078e-01
-3.37183080e-03 1.27524614e-01 4.84521300e-01 8.67896557e-01
-8.40222001e-01 -1.01972711e+00 -3.24498713e-01 3.33839506e-01
-1.11924253e-01 9.26802084e-02 -1.15774430e-01 6.98331356e-01
5.27981818e-01 1.18673921e+00 6.03981502e-02 -7.71446288e-01
3.49622935e-01 -4.25101489e-01 2.98299283e-01 -6.24366105e-01
-3.06404620e-01 -4.09355722e-02 -3.97965670e-01 -4.33394194e-01
-7.56599367e-01 -3.93870443e-01 -1.55584836e+00 -6.51689991e-02
-3.95542115e-01 2.59326071e-01 5.90673208e-01 9.92356598e-01
1.81614414e-01 4.92908210e-01 5.89864731e-01 -1.13896227e+00
-1.65695548e-01 -1.04539919e+00 -8.63771379e-01 2.23020941e-01
-1.39068132e-02 -1.00070965e+00 1.19613288e-02 -3.31643708e-02] | [9.53616714477539, -0.45816493034362793] |
f81bc3fe-a9bc-4a2b-96bd-1e42de600eda | road-aware-monocular-structure-from-motion | 2112.08635 | null | https://arxiv.org/abs/2112.08635v1 | https://arxiv.org/pdf/2112.08635v1.pdf | Road-aware Monocular Structure from Motion and Homography Estimation | Structure from motion (SFM) and ground plane homography estimation are critical to autonomous driving and other robotics applications. Recently, much progress has been made in using deep neural networks for SFM and homography estimation respectively. However, directly applying existing methods for ground plane homography estimation may fail because the road is often a small part of the scene. Besides, the performances of deep SFM approaches are still inferior to traditional methods. In this paper, we propose a method that learns to solve both problems in an end-to-end manner, improving performance on both. The proposed networks consist of a Depth-CNN, a Pose-CNN and a Ground-CNN. The Depth-CNN and Pose-CNN estimate dense depth map and ego-motion respectively, solving SFM, while the Pose-CNN and Ground-CNN followed by a homography layer solve the ground plane estimation problem. By enforcing coherency between SFM and homography estimation results, the whole network can be trained end to end using photometric loss and homography loss without any groundtruth except the road segmentation provided by an off-the-shelf segmenter. Comprehensive experiments are conducted on KITTI benchmark to demonstrate promising results compared with various state-of-the-art approaches. | ['Qian Zhang', 'Jiao Lu', 'Jiaxin Zhang', 'Teng Chen', 'Wei Sui'] | 2021-12-16 | null | null | null | null | ['road-segementation', 'homography-estimation'] | ['computer-vision', 'computer-vision'] | [ 6.21680766e-02 3.72443795e-01 -9.72111002e-02 -5.29573679e-01
-6.75251484e-01 -2.81074464e-01 4.97215152e-01 -3.98504972e-01
-5.69388330e-01 4.39187646e-01 -4.27036136e-01 -1.93311628e-02
2.72264063e-01 -9.93260682e-01 -1.25325906e+00 -5.29064298e-01
3.70187193e-01 8.67503941e-01 5.88928401e-01 -1.90212637e-01
4.19429958e-01 3.86908114e-01 -1.52826834e+00 -4.92362499e-01
9.36954379e-01 1.20414960e+00 3.12742651e-01 4.52045619e-01
1.86897486e-01 4.87699211e-01 -1.20298132e-01 -5.60838103e-01
5.86033404e-01 -1.01732284e-01 -5.75376749e-01 8.46687704e-02
1.09625459e+00 -7.54285455e-01 -7.76280761e-01 1.22690678e+00
3.87949973e-01 4.19431746e-01 3.53375584e-01 -1.26974905e+00
7.25605711e-02 -1.56194791e-02 -8.35179269e-01 -3.31627429e-01
9.91777554e-02 1.58044696e-01 7.05190480e-01 -7.38859892e-01
7.09875762e-01 1.43631649e+00 7.74275839e-01 2.14187816e-01
-8.18425238e-01 -6.55874074e-01 1.79148875e-02 3.09849471e-01
-1.33847761e+00 -2.18156025e-01 9.63501453e-01 -3.52379829e-01
1.06841052e+00 -3.81932884e-01 6.70810044e-01 5.56858182e-01
3.90012383e-01 8.07412982e-01 5.33194005e-01 6.75729290e-02
-7.67773613e-02 -1.30454466e-01 -1.49447680e-01 1.01116049e+00
2.85140693e-01 3.38455617e-01 -4.05745208e-01 4.34959859e-01
9.99495685e-01 3.53471749e-02 -1.20040819e-01 -8.43792737e-01
-1.23251128e+00 7.56287873e-01 7.74454653e-01 -4.57136750e-01
-2.75089145e-01 6.07060730e-01 2.87158817e-01 -3.31018604e-02
3.66996884e-01 1.03715152e-01 -9.90408212e-02 -3.53896394e-02
-1.06424499e+00 4.64348674e-01 6.91661060e-01 1.29992616e+00
1.44398153e+00 -1.31766684e-02 2.94396520e-01 5.65057218e-01
3.78000200e-01 7.46441960e-01 2.84160227e-01 -1.41168499e+00
8.15296054e-01 5.21005571e-01 -2.34991368e-02 -1.17796814e+00
-4.26924139e-01 -4.65989530e-01 -8.31840396e-01 3.04804355e-01
4.06709343e-01 -5.05331233e-02 -1.05570793e+00 1.44527864e+00
5.31267047e-01 4.93851840e-01 -5.15278801e-02 1.16215515e+00
5.42235017e-01 5.36667407e-01 -4.69343781e-01 4.35734421e-01
1.19820344e+00 -1.50004661e+00 -5.83290398e-01 -8.75131190e-01
6.28858089e-01 -8.11659455e-01 6.72666788e-01 1.94650441e-01
-1.11389136e+00 -6.17075145e-01 -1.43475890e+00 -8.02072942e-01
-2.09620863e-01 2.81370014e-01 3.73351932e-01 1.66244358e-01
-9.63373840e-01 6.03204966e-01 -9.33446884e-01 -2.34840408e-01
3.81608158e-01 4.18779135e-01 -4.45711017e-01 -2.14754969e-01
-1.07125759e+00 1.04768848e+00 5.51228881e-01 4.24145013e-01
-9.74950254e-01 -4.94789392e-01 -1.10483968e+00 -1.42485932e-01
3.99359405e-01 -1.09474254e+00 1.24994087e+00 -4.42426354e-01
-1.75302863e+00 1.00229895e+00 -9.89538021e-05 -7.18228459e-01
9.70976830e-01 -4.74825233e-01 2.02298358e-01 1.46200761e-01
2.65688688e-01 1.03338397e+00 6.83283269e-01 -1.12405622e+00
-1.03700578e+00 -5.60223281e-01 2.74320040e-02 4.13690984e-01
3.00019830e-01 -7.70520806e-01 -8.50828648e-01 -5.17416745e-03
6.47981524e-01 -1.14462149e+00 -3.39892894e-01 1.36235908e-01
-6.43281341e-01 1.27938747e-01 1.02733338e+00 -6.94373071e-01
6.44307077e-01 -1.89750314e+00 7.89128095e-02 4.31603715e-02
2.14857966e-01 1.54471830e-01 -9.36361626e-02 1.03176646e-01
2.95866668e-01 -4.62940663e-01 -4.20694321e-01 -8.21470737e-01
1.01270482e-01 3.50513548e-01 -2.73855746e-01 8.96666169e-01
4.60082814e-02 9.76947904e-01 -8.15066934e-01 -3.73944998e-01
6.45303190e-01 6.19886577e-01 -5.40630937e-01 1.52765736e-01
-1.97966695e-01 4.62712675e-01 -3.11627239e-01 4.45922971e-01
1.09879994e+00 8.14912170e-02 -2.85361618e-01 -2.55139440e-01
-3.54922652e-01 3.04754466e-01 -1.14272404e+00 2.08772397e+00
-5.43005228e-01 8.33342910e-01 2.36755281e-04 -8.81075084e-01
9.88070786e-01 -2.03395009e-01 1.39706239e-01 -8.22097659e-01
4.66113091e-01 4.55696791e-01 -1.80050820e-01 -2.57780850e-01
6.85021281e-01 1.36312813e-01 1.13687433e-01 -2.05863297e-01
-2.87552141e-02 -7.61431813e-01 1.12593181e-01 -5.60405897e-03
6.33584380e-01 3.85771453e-01 8.10904801e-03 3.91138792e-02
6.59626603e-01 2.40069807e-01 6.85483098e-01 2.58594304e-01
-1.09878466e-01 7.97048211e-01 3.70970726e-01 -4.89642471e-01
-1.26402318e+00 -9.66058195e-01 3.74659598e-02 2.71781147e-01
9.31438923e-01 -1.42266918e-02 -8.91734779e-01 -3.46205413e-01
1.06984854e-01 3.88499856e-01 -3.45692426e-01 -2.40264058e-01
-7.94758737e-01 -1.82990581e-01 4.83218104e-01 6.54844046e-01
1.30088806e+00 -6.26906753e-01 -8.50855470e-01 3.44659328e-01
-2.96685189e-01 -1.68388999e+00 -6.00171149e-01 -3.77375409e-02
-9.33709264e-01 -1.00783241e+00 -6.64515197e-01 -8.24655354e-01
5.03201783e-01 5.94750643e-01 9.29095745e-01 -2.34099224e-01
6.43471861e-03 -2.22174436e-01 1.91740617e-01 -2.49763787e-01
-5.61335795e-02 4.07734364e-01 -2.91092753e-01 -4.42300774e-02
1.55263901e-01 -6.82449460e-01 -9.40348864e-01 6.42763853e-01
-4.45617288e-01 2.65641093e-01 7.59645104e-01 4.88394409e-01
8.74945223e-01 -1.26493305e-01 -1.12545177e-01 -7.78072059e-01
-2.95579255e-01 -1.36553571e-01 -1.22768009e+00 -3.93482596e-01
-5.59167862e-01 -9.26044211e-02 3.53515208e-01 7.98347518e-02
-8.94167721e-01 6.33302450e-01 -3.24719965e-01 -7.96536803e-01
-1.60707068e-02 1.59773409e-01 -4.15948510e-01 -2.86231220e-01
2.71517307e-01 2.11971045e-01 1.90484688e-01 -2.05056578e-01
3.08577091e-01 2.26202235e-01 1.10190439e+00 -1.64665468e-02
1.21291757e+00 9.03160393e-01 3.43094110e-01 -7.22704470e-01
-9.00492549e-01 -7.47297525e-01 -8.98677468e-01 -2.25030512e-01
1.21166050e+00 -1.25606728e+00 -7.60639429e-01 7.27618873e-01
-1.34161830e+00 -3.66911590e-01 1.45903826e-01 6.19466960e-01
-8.86163056e-01 4.97218549e-01 -2.38329217e-01 -4.48053002e-01
-2.61464864e-01 -1.41672826e+00 1.45793068e+00 2.62639374e-01
1.42480507e-01 -8.21683824e-01 3.18196346e-03 6.27306581e-01
1.13420591e-01 4.72644359e-01 3.37157130e-01 1.30974352e-01
-1.23012042e+00 -5.07316053e-01 -4.82858419e-01 2.28275582e-01
-3.55754286e-01 -1.44329771e-01 -1.09145510e+00 -1.38036311e-01
-8.83456320e-02 -1.13909535e-01 1.10203946e+00 5.65756738e-01
6.84674799e-01 2.09085215e-02 -3.55768889e-01 1.33219326e+00
1.47550309e+00 1.10111170e-01 9.37784612e-01 7.73212850e-01
1.36173809e+00 7.86481023e-01 8.80297661e-01 1.14879012e-01
8.40164781e-01 7.54376292e-01 1.02715850e+00 -2.54463971e-01
-1.98059693e-01 -5.11561692e-01 2.40201458e-01 5.29172003e-01
1.11774243e-01 -1.43734619e-01 -8.91647935e-01 6.04604065e-01
-2.03610921e+00 -6.01520181e-01 -5.35791934e-01 2.24097538e+00
3.60762149e-01 2.89093524e-01 -8.61522481e-02 -5.85758768e-04
5.39872408e-01 3.95309031e-01 -8.76562417e-01 -8.73255208e-02
-1.00970991e-01 -4.67356890e-01 1.05098689e+00 7.29201376e-01
-1.26853001e+00 1.40068305e+00 4.81046486e+00 7.53749132e-01
-1.22503018e+00 -5.69474623e-02 3.90646577e-01 1.95504755e-01
-5.91240004e-02 1.98091552e-01 -1.11833370e+00 8.22870210e-02
5.63382089e-01 2.14064017e-01 2.92731822e-01 1.10772312e+00
1.10154442e-01 -3.28729361e-01 -9.92180526e-01 1.30420160e+00
1.32126932e-03 -1.22862399e+00 -1.66361868e-01 1.76341474e-01
9.01686072e-01 4.85631615e-01 -5.80353029e-02 6.89750984e-02
7.74164274e-02 -6.80663109e-01 8.80707979e-01 3.10424626e-01
7.01828361e-01 -9.98532534e-01 8.88612866e-01 5.53232729e-01
-1.43684065e+00 2.21210778e-01 -5.70165575e-01 -3.26521806e-02
4.81388271e-01 7.10025251e-01 -5.07821620e-01 7.98132360e-01
6.00169361e-01 9.53674555e-01 -2.72865623e-01 1.07484257e+00
-5.02851903e-01 -4.27675620e-02 -4.37209606e-01 4.11174685e-01
6.96695268e-01 -5.35173893e-01 6.10140622e-01 8.97711694e-01
3.10550094e-01 -2.30093375e-01 1.38781086e-01 9.52637434e-01
-7.42680058e-02 -2.37425536e-01 -7.06265688e-01 3.21235776e-01
3.33863944e-01 1.28079462e+00 -7.47377396e-01 -4.38163579e-01
-3.23639780e-01 1.07099390e+00 1.54568344e-01 2.47605562e-01
-8.82312655e-01 -6.09005094e-01 7.23675966e-01 4.17737395e-01
1.14539020e-01 -4.71754313e-01 -3.66142094e-01 -1.09007895e+00
1.14825301e-01 -4.22243506e-01 -2.04611450e-01 -6.78449512e-01
-7.16700315e-01 4.96076345e-01 -2.69519836e-02 -1.33678317e+00
-3.48988891e-01 -4.80161309e-01 -5.05614877e-01 9.40746665e-01
-1.95321059e+00 -1.03560853e+00 -9.15187001e-01 3.88555765e-01
6.37276351e-01 1.18120283e-01 3.06684747e-02 4.31913048e-01
-6.08984411e-01 3.56658489e-01 -4.45200764e-02 7.47063234e-02
6.04900241e-01 -1.13315129e+00 9.32588637e-01 9.61611032e-01
-2.98732191e-01 3.09988499e-01 6.04692042e-01 -6.22272551e-01
-1.38953221e+00 -1.38194108e+00 8.27969313e-01 -2.47978806e-01
3.48929763e-01 -3.73770177e-01 -7.79560030e-01 6.96113348e-01
2.79736407e-02 3.80148776e-02 -3.26386094e-01 -5.34367859e-01
-3.59603092e-02 -1.89652681e-01 -9.09643710e-01 4.89354789e-01
1.14061201e+00 -4.74945188e-01 -2.04982683e-01 1.59046933e-01
7.90552795e-01 -1.09807444e+00 -4.35392559e-01 4.01532441e-01
6.19203150e-01 -1.03439450e+00 1.03621387e+00 2.99976498e-01
6.33143127e-01 -6.46621823e-01 -1.18364923e-01 -1.05096042e+00
3.74321528e-02 -6.58716619e-01 -1.97995938e-02 8.27148080e-01
1.66268259e-01 -7.45181203e-01 1.26638842e+00 3.37105274e-01
-7.19041169e-01 -6.32687271e-01 -8.62683535e-01 -8.48513842e-01
-1.42864704e-01 -3.47144395e-01 5.55004776e-01 6.63391292e-01
-7.23925292e-01 4.97405231e-01 -4.21911716e-01 4.53865200e-01
9.88849163e-01 7.84836709e-02 1.36246657e+00 -1.07495451e+00
-7.69588277e-02 -4.58076715e-01 -7.71994174e-01 -1.76641667e+00
2.53535777e-01 -6.31564319e-01 4.97584283e-01 -1.59610760e+00
-3.53678852e-01 3.93752344e-02 4.56163704e-01 9.09468755e-02
1.13646366e-01 2.95812160e-01 -3.80724557e-02 2.66332597e-01
-2.87201643e-01 7.71098614e-01 1.63879013e+00 -2.40797684e-01
-2.35069960e-01 -1.74383838e-02 4.31574509e-02 1.06442237e+00
5.96150517e-01 -3.33724380e-01 -7.39353478e-01 -7.84214675e-01
1.85469866e-01 2.56537378e-01 5.92186511e-01 -1.38351929e+00
6.03639126e-01 1.91899747e-01 2.21025929e-01 -1.35299063e+00
6.32907391e-01 -6.88453317e-01 -3.81065719e-02 5.46325088e-01
1.54281184e-01 -5.73978908e-02 1.17465347e-01 5.80708563e-01
-4.63200092e-01 -2.51755193e-02 1.05605578e+00 -9.43511128e-02
-9.99353707e-01 7.87280917e-01 4.08057533e-02 4.57430221e-02
8.75242889e-01 -6.49486363e-01 -2.02054128e-01 -3.85515958e-01
-2.15967491e-01 5.10789514e-01 5.73643744e-01 3.13991547e-01
7.97469020e-01 -1.15462601e+00 -4.32774872e-01 2.90475965e-01
3.28842103e-02 9.83167171e-01 3.20212305e-01 1.03461707e+00
-1.24277127e+00 4.61657017e-01 -2.21206963e-01 -8.25245023e-01
-8.17549527e-01 2.04535663e-01 6.50485575e-01 -1.03246920e-01
-8.53593588e-01 8.84336770e-01 6.38360322e-01 -6.81003034e-01
3.69948626e-01 -4.52153593e-01 3.90346199e-02 -1.74251974e-01
1.66773036e-01 6.50738418e-01 4.46017608e-02 -8.61752510e-01
-2.87174076e-01 1.25677204e+00 3.14628519e-03 -2.15982959e-01
1.03041959e+00 -4.26695853e-01 5.54981753e-02 1.31706551e-01
1.68018687e+00 -5.63822091e-01 -1.69418669e+00 -4.30721976e-02
-3.23095053e-01 -5.04084706e-01 3.87453765e-01 -1.62809417e-01
-1.36306822e+00 1.39696908e+00 4.79093373e-01 -4.79402751e-01
7.89967239e-01 -4.57318991e-01 1.36476505e+00 5.73057652e-01
4.09816086e-01 -1.24282038e+00 -5.86532317e-02 8.22811604e-01
6.58694565e-01 -1.45644999e+00 -4.48981263e-02 -6.92672551e-01
-4.51343983e-01 1.16502750e+00 8.14210653e-01 -5.50550818e-01
4.55212116e-01 -7.59601146e-02 1.37155339e-01 -2.38017768e-01
-2.84324974e-01 -2.56713867e-01 2.69847363e-01 4.48253274e-01
-1.18048899e-01 -1.98793858e-01 3.53430696e-02 -9.38681364e-02
-6.30015969e-01 -1.37197152e-01 4.54773903e-01 6.78522885e-01
-5.53256810e-01 -8.19649041e-01 -2.29068294e-01 4.17615883e-02
5.93330264e-02 2.08414018e-01 -1.84377626e-01 1.14351642e+00
2.19784930e-01 5.78901291e-01 3.19611311e-01 -5.15316129e-01
6.21726453e-01 -4.75620896e-01 4.44010615e-01 -3.86909902e-01
-8.22729990e-02 1.78738385e-01 1.41453564e-01 -9.92544711e-01
-1.08752541e-01 -4.67144877e-01 -1.56169689e+00 -5.60216069e-01
-2.16449842e-01 -4.39193040e-01 8.22221994e-01 9.36586320e-01
7.31735379e-02 4.35648113e-01 5.89559972e-01 -1.27067339e+00
-3.28670263e-01 -7.59434879e-01 -5.31361759e-01 5.50526232e-02
4.54573274e-01 -7.46478856e-01 -3.49474609e-01 -8.14445242e-02] | [8.24503231048584, -2.1238863468170166] |
e161c837-e104-4089-bec0-2114231ca1e4 | collocation-polarity-disambiguation-using-web | null | null | https://aclanthology.org/D12-1015 | https://aclanthology.org/D12-1015.pdf | Collocation Polarity Disambiguation Using Web-based Pseudo Contexts | null | ['Ting Liu', 'Bing Qin', 'Yanyan Zhao'] | 2012-07-01 | null | null | null | emnlp-2012-7 | ['subjectivity-analysis'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.3288679122924805, 3.753092050552368] |
aefb5c4a-429f-4b9d-8191-90e31f5bd100 | visual-spatial-reasoning | 2205.00363 | null | https://arxiv.org/abs/2205.00363v3 | https://arxiv.org/pdf/2205.00363v3.pdf | Visual Spatial Reasoning | Spatial relations are a basic part of human cognition. However, they are expressed in natural language in a variety of ways, and previous work has suggested that current vision-and-language models (VLMs) struggle to capture relational information. In this paper, we present Visual Spatial Reasoning (VSR), a dataset containing more than 10k natural text-image pairs with 66 types of spatial relations in English (such as: under, in front of, and facing). While using a seemingly simple annotation format, we show how the dataset includes challenging linguistic phenomena, such as varying reference frames. We demonstrate a large gap between human and model performance: the human ceiling is above 95%, while state-of-the-art models only achieve around 70%. We observe that VLMs' by-relation performances have little correlation with the number of training examples and the tested models are in general incapable of recognising relations concerning the orientations of objects. | ['Nigel Collier', 'Guy Emerson', 'Fangyu Liu'] | 2022-04-30 | null | null | null | null | ['visual-entailment'] | ['reasoning'] | [-1.00489408e-01 1.29761904e-01 -2.44769212e-02 -4.01216418e-01
-3.45711529e-01 -7.07167804e-01 1.19360662e+00 1.15980700e-01
-5.26086450e-01 4.58227992e-01 1.97285250e-01 -6.08394086e-01
-2.23400474e-01 -4.53530610e-01 -6.42675519e-01 -3.47308755e-01
5.98860085e-02 5.37753046e-01 6.05762422e-01 -4.93935406e-01
4.18519318e-01 6.44633412e-01 -1.59781671e+00 5.45473099e-01
5.28102815e-01 9.63951349e-01 2.80901343e-01 6.32107198e-01
-5.17841168e-02 1.53157127e+00 -6.01693451e-01 -3.75520289e-01
-4.25462723e-02 -3.15601021e-01 -9.61118579e-01 -9.58668441e-02
7.86489427e-01 -5.98137602e-02 -7.70753741e-01 9.19527292e-01
2.88814545e-01 2.60894686e-01 8.04824293e-01 -1.25913703e+00
-1.13256264e+00 2.86424458e-01 -4.32945341e-01 4.51866299e-01
8.75147462e-01 1.50414228e-01 1.03219438e+00 -9.78940725e-01
7.44836450e-01 1.32652235e+00 5.67801118e-01 3.13895345e-01
-1.29448080e+00 -1.34058267e-01 1.96007952e-01 7.30587661e-01
-1.69079673e+00 -5.54980934e-01 5.39508641e-01 -6.30903661e-01
1.37592053e+00 4.27917153e-01 5.89041889e-01 8.98661494e-01
2.60833763e-02 6.35395229e-01 1.43248618e+00 -7.16217041e-01
1.77277476e-01 -1.51091799e-01 9.75971594e-02 5.78978837e-01
2.19268650e-02 -1.17773019e-01 -7.76639700e-01 2.33914793e-01
1.08126450e+00 -3.40584517e-01 -2.66057372e-01 -5.32772124e-01
-1.67405570e+00 3.45859081e-01 7.93111086e-01 5.19051015e-01
-1.01458274e-01 1.33977935e-01 -3.60233821e-02 1.42044574e-01
-7.06184432e-02 5.50776601e-01 -1.26481384e-01 -8.77508000e-02
-6.45679295e-01 4.01931524e-01 4.54511434e-01 1.23635411e+00
5.69495559e-01 -2.40636453e-01 -2.88830757e-01 9.01585519e-01
1.97933003e-01 4.93216872e-01 3.63625556e-01 -1.12682307e+00
5.89428723e-01 6.51494265e-01 2.73565084e-01 -1.45257711e+00
-7.02963889e-01 -8.20114650e-03 -7.44082272e-01 1.47848383e-01
8.06313753e-01 5.05778849e-01 -7.86831260e-01 1.55076289e+00
-9.03559383e-03 -2.46826127e-01 8.82263556e-02 1.03502786e+00
1.11925828e+00 3.18096906e-01 2.38353610e-01 7.96665624e-02
1.56634367e+00 -1.02920222e+00 -7.71189451e-01 -7.38730192e-01
6.92436039e-01 -7.09097147e-01 1.38301682e+00 2.07961366e-01
-1.20101035e+00 -6.98724985e-01 -8.27596784e-01 -6.55993760e-01
-7.60742843e-01 1.75791115e-01 6.42979264e-01 2.70417392e-01
-1.30450416e+00 1.67006418e-01 -5.17759085e-01 -5.68700016e-01
4.31339920e-01 6.71540573e-02 -7.45953143e-01 -1.86196521e-01
-9.12082493e-01 1.51390338e+00 3.40785325e-01 1.64954752e-01
-3.11714202e-01 -3.39447230e-01 -1.06155682e+00 -2.07097247e-01
4.89699304e-01 -5.13549387e-01 1.25884604e+00 -4.90658522e-01
-8.46750975e-01 1.36344206e+00 -3.23129505e-01 -3.72050107e-01
6.52273834e-01 -2.07108915e-01 -5.49183309e-01 1.89265609e-01
1.55601665e-01 8.84963512e-01 3.55976015e-01 -1.37717283e+00
-5.84936976e-01 -3.94866943e-01 5.06669641e-01 3.18281382e-01
3.33257526e-01 7.32121021e-02 -5.58458984e-01 -5.54180920e-01
5.60822010e-01 -9.84458864e-01 1.34342664e-03 3.85005981e-01
-4.03103471e-01 -4.34095055e-01 5.14790893e-01 -5.80544949e-01
1.09614646e+00 -2.18492770e+00 1.09315388e-01 -1.19908573e-02
1.47962064e-01 2.02055797e-01 -2.06701290e-02 3.53142500e-01
-1.56101674e-01 6.04561381e-02 -5.98355308e-02 -1.59302548e-01
1.80370525e-01 6.32183611e-01 -4.88208294e-01 5.28213322e-01
-3.80289741e-03 1.20722473e+00 -1.06531072e+00 -7.64730871e-01
4.92554069e-01 4.78127152e-01 -2.23399207e-01 4.07022797e-03
-1.49714142e-01 4.09126431e-01 8.94225296e-03 4.42838758e-01
4.56400514e-01 -3.71622056e-01 2.09960699e-01 -1.89747080e-01
-1.72247440e-01 3.91238660e-01 -1.10854626e+00 1.81394660e+00
-1.84596822e-01 1.18804479e+00 -5.05173922e-01 -6.79494500e-01
6.32178962e-01 5.84516115e-02 -1.35385931e-01 -1.39355028e+00
-7.44786337e-02 -8.39481726e-02 3.44404787e-01 -7.06104219e-01
5.59163272e-01 2.93026298e-01 1.74434781e-01 6.79181889e-02
-2.13360921e-01 -2.38557711e-01 3.83594066e-01 1.61490023e-01
7.54190207e-01 2.67228365e-01 6.82738364e-01 -3.42856288e-01
5.10723650e-01 3.92952971e-02 4.57135364e-02 7.71875381e-01
-2.89412946e-01 7.25043714e-01 7.18486249e-01 -5.99160194e-01
-1.00376892e+00 -1.08575749e+00 -1.08568750e-01 1.07778382e+00
4.91987497e-01 -6.56408429e-01 -5.23593247e-01 -2.06053287e-01
-2.96226889e-01 1.06296182e+00 -8.56341958e-01 1.04873203e-01
-4.86347079e-01 -2.20495760e-01 6.14908099e-01 8.37650418e-01
7.01817691e-01 -1.16220772e+00 -1.20414674e+00 -3.49571377e-01
-2.49849021e-01 -1.77679253e+00 6.33438900e-02 -5.72396591e-02
-4.55487281e-01 -1.22412550e+00 -4.19414729e-01 -8.08440983e-01
7.03665137e-01 4.29404765e-01 1.53032446e+00 1.53972551e-01
-9.57300067e-02 3.82185251e-01 -2.74469972e-01 -2.18833044e-01
-2.03718722e-01 -3.61746550e-01 -2.24597499e-01 -3.26268941e-01
3.16836238e-01 -3.62087965e-01 -6.30029559e-01 4.79894042e-01
-7.39331126e-01 4.61021900e-01 3.53178978e-01 6.50178850e-01
5.03118992e-01 -1.72343165e-01 -1.45560741e-01 -5.68983078e-01
3.37043136e-01 -3.66646424e-02 -4.22780782e-01 5.33992887e-01
-2.38455936e-01 9.74373370e-02 1.99049324e-01 -4.99893904e-01
-9.30195272e-01 -4.92531247e-02 2.07474560e-01 -2.04963118e-01
-5.56801617e-01 2.08713770e-01 4.79651466e-02 -9.23024192e-02
9.77671981e-01 2.22727016e-01 -2.76207179e-01 -4.30850498e-02
6.57534480e-01 2.86646783e-01 8.33911777e-01 -6.27675533e-01
4.54896361e-01 5.46826839e-01 3.38679790e-01 -1.01303625e+00
-9.43265617e-01 -3.70228708e-01 -1.00199425e+00 -2.43496940e-01
9.33633208e-01 -9.09955859e-01 -8.03359926e-01 2.72603303e-01
-1.37714159e+00 -6.56264365e-01 -1.18897535e-01 3.69026661e-01
-7.07797408e-01 1.93108365e-01 -2.93493301e-01 -6.94171727e-01
4.35060322e-01 -1.00524008e+00 1.15663517e+00 5.20301908e-02
-5.43505490e-01 -1.04963481e+00 -2.67319828e-01 4.40940112e-01
2.14116499e-01 2.10778624e-01 1.07845187e+00 -3.61842453e-01
-5.08921027e-01 1.48459584e-01 -6.14435673e-01 -1.60805315e-01
5.92138842e-02 -1.86991349e-01 -9.59024191e-01 2.81699508e-01
-2.84571975e-01 -3.67522717e-01 5.44974327e-01 1.13532990e-01
1.08706224e+00 -3.87699977e-02 -2.88666785e-01 3.50251824e-01
1.18448663e+00 3.39190930e-01 1.13453591e+00 4.73931789e-01
9.97653604e-01 8.09160352e-01 5.38762093e-01 2.30694227e-02
8.53770137e-01 9.29942489e-01 4.08947945e-01 3.35989743e-02
-3.07995707e-01 -1.52829036e-01 -1.20741203e-01 2.46122777e-01
-4.93683338e-01 -6.94286972e-02 -1.57493186e+00 4.41226929e-01
-1.96934271e+00 -1.10607445e+00 -4.39989865e-01 1.91590691e+00
8.12271297e-01 5.35211004e-02 -6.61204234e-02 2.87815422e-01
3.28665823e-01 3.67336273e-01 -2.29932100e-01 -2.29412004e-01
-4.94925678e-01 -1.26566976e-01 2.79835165e-01 5.73974073e-01
-9.73075807e-01 1.29687691e+00 7.19916487e+00 5.77031553e-01
-9.25078034e-01 -2.00600609e-01 3.60272110e-01 1.90395549e-01
-1.09924294e-01 -2.19575725e-02 -3.60949725e-01 -2.59480141e-02
5.20083785e-01 3.63101572e-01 6.25044286e-01 5.02155960e-01
-8.67784247e-02 -5.40358484e-01 -1.39829755e+00 1.58704543e+00
3.64437371e-01 -1.14644611e+00 -8.61214940e-03 -1.92921042e-01
4.70520735e-01 -1.74143314e-01 1.33001462e-01 1.70786083e-01
1.69048995e-01 -1.42962193e+00 1.19252980e+00 7.58696258e-01
6.92873001e-01 -2.73651779e-01 2.51451135e-01 5.10239661e-01
-1.02377057e+00 3.78667042e-02 -2.79251009e-01 -5.07928252e-01
-1.72308981e-02 5.82410283e-02 -5.62359631e-01 3.38057816e-01
9.29313719e-01 6.07291341e-01 -1.31215048e+00 7.28284061e-01
-5.42893946e-01 -3.59808542e-02 -1.58070862e-01 -8.39652345e-02
1.40106827e-01 -3.69154252e-02 1.22444645e-01 1.06354022e+00
5.50898910e-02 3.35887074e-01 -2.24906415e-01 8.32926631e-01
4.22712743e-01 -9.96558368e-02 -8.41552556e-01 2.12735131e-01
4.16281164e-01 7.75071740e-01 -8.86765301e-01 -1.60774961e-01
-6.75823271e-01 8.84668410e-01 7.08836973e-01 6.89982533e-01
-9.15007114e-01 1.13321483e-01 4.30275172e-01 2.39810765e-01
2.40464509e-01 -6.54250205e-01 -4.47354794e-01 -1.02451634e+00
3.02598625e-01 -7.53259897e-01 1.54072762e-01 -1.57662082e+00
-1.03935528e+00 5.48696220e-01 1.57168493e-01 -9.94714499e-01
-2.83413887e-01 -1.05352628e+00 3.06114461e-03 6.99524224e-01
-1.36667514e+00 -1.32932055e+00 -6.41599655e-01 6.76219344e-01
4.50784743e-01 -1.50664439e-02 8.97257984e-01 -5.76233901e-02
2.90149041e-02 3.26666296e-01 -4.33153570e-01 3.79291981e-01
5.28632224e-01 -1.36447048e+00 5.78725457e-01 6.07290685e-01
7.29376316e-01 8.51631343e-01 7.71619081e-01 -2.60740072e-01
-1.09711421e+00 -6.03306234e-01 1.21421635e+00 -1.01272130e+00
6.57553911e-01 -4.41964179e-01 -8.28968644e-01 9.76896465e-01
3.28155458e-01 5.39049864e-01 4.63875055e-01 2.54112691e-01
-9.71493423e-01 2.39874765e-01 -7.92984724e-01 1.29728794e+00
1.58434999e+00 -1.06478322e+00 -1.02667773e+00 3.89040470e-01
2.76611239e-01 -8.78774166e-01 -4.39818323e-01 2.74027735e-01
4.51094568e-01 -1.37687385e+00 1.38642621e+00 -6.94842100e-01
4.96837765e-01 -5.15003920e-01 -5.55999577e-01 -1.02522635e+00
-5.20372450e-01 -1.17058411e-01 -2.16133431e-01 8.58915567e-01
2.19271794e-01 -5.53741992e-01 2.92787910e-01 7.62466192e-01
9.58097801e-02 -6.80285752e-01 -1.05354321e+00 -7.49631166e-01
-2.75401231e-02 -5.98963022e-01 2.31097028e-01 1.01491082e+00
3.10142115e-02 4.45963621e-01 3.27447429e-02 2.81513512e-01
1.82245582e-01 -7.72305727e-02 7.02973664e-01 -1.06224585e+00
-8.23305696e-02 -7.03572392e-01 -7.99584746e-01 -1.15923202e+00
3.12900871e-01 -5.50159395e-01 -1.22029729e-01 -2.00107670e+00
-6.36198446e-02 -3.07648450e-01 1.18626468e-01 5.31852782e-01
5.50669953e-02 2.54847109e-01 3.81609887e-01 2.80137241e-01
-8.42715323e-01 1.93929747e-01 1.52585924e+00 -1.67364106e-01
1.10871293e-01 -5.15125215e-01 -6.24503851e-01 1.00754666e+00
4.58318144e-01 4.77753878e-02 -6.17481947e-01 -8.48125160e-01
5.77470601e-01 -3.24716456e-02 1.09764445e+00 -1.07071304e+00
4.91553992e-01 -3.53906155e-01 5.89885890e-01 -5.34594655e-01
5.49467683e-01 -9.17692423e-01 1.26736164e-01 5.34471087e-02
-5.22952259e-01 4.65447336e-01 2.67148435e-01 3.43069524e-01
-1.60056010e-01 1.28353372e-01 5.30221522e-01 -5.46094626e-02
-1.25820351e+00 -2.22139239e-01 -8.90609771e-02 3.98459494e-01
1.07667494e+00 -4.05607700e-01 -7.62432933e-01 -3.54384154e-01
-7.32811391e-01 7.98525438e-02 4.90296155e-01 5.97045839e-01
6.01102591e-01 -1.46129692e+00 -3.34260643e-01 -7.55146816e-02
5.15105486e-01 2.32753158e-01 3.41445535e-01 8.30421209e-01
-8.05629790e-01 6.08406782e-01 -2.13052973e-01 -8.50746214e-01
-1.10638273e+00 7.59940386e-01 4.79861259e-01 2.37586230e-01
-8.19998384e-01 7.36727297e-01 2.92540848e-01 -2.09299639e-01
1.96077615e-01 -5.43886781e-01 -4.01329875e-01 1.96742732e-02
5.00978589e-01 1.41283363e-01 -9.19566974e-02 -1.20582533e+00
-6.17011249e-01 9.00146186e-01 3.10232580e-01 -2.13822275e-01
8.71445477e-01 -4.73008007e-02 -1.26557454e-01 7.71076322e-01
8.02358568e-01 -5.72203919e-02 -1.23475850e+00 -3.74470085e-01
1.22767828e-01 -6.23538435e-01 -2.79494226e-01 -7.11648464e-01
-4.98626858e-01 1.00783181e+00 4.99588370e-01 3.14487934e-01
9.13631856e-01 4.09024894e-01 2.83865593e-02 6.55559957e-01
5.08844972e-01 -1.05227876e+00 1.47208974e-01 7.66866446e-01
1.20928323e+00 -1.44734395e+00 7.81856552e-02 -5.41693151e-01
-9.18398857e-01 8.18242550e-01 5.79210460e-01 1.42634153e-01
3.76423270e-01 1.55523317e-02 1.84798270e-01 -2.79104799e-01
-5.14954269e-01 -5.31857789e-01 6.33757114e-01 9.72650051e-01
4.83623981e-01 1.26770929e-01 4.83478606e-02 2.11446434e-01
-6.51873469e-01 -2.97787040e-01 2.64977217e-01 9.56903040e-01
-1.15334935e-01 -4.41191047e-01 -3.30269635e-01 -3.93422469e-02
-2.51992699e-02 -2.05131292e-01 -5.20985007e-01 1.04482019e+00
2.21581221e-01 9.48278189e-01 3.93721521e-01 -4.08551060e-02
6.46583617e-01 -1.03657477e-01 9.85036254e-01 -4.08672601e-01
-7.35859498e-02 -4.87428397e-01 9.27051082e-02 -7.53312230e-01
-7.67083704e-01 -6.11893892e-01 -1.47561693e+00 -1.20640822e-01
2.46558949e-01 -3.77234757e-01 1.86073929e-01 1.22517979e+00
4.61145081e-02 5.54161131e-01 -4.27717775e-01 -7.20304191e-01
-2.27688961e-02 -8.31707060e-01 -4.51920182e-01 7.45866835e-01
4.59869623e-01 -1.02943659e+00 -1.68138489e-01 2.74183363e-01] | [10.462747573852539, 1.7850518226623535] |
3c6d525b-f229-4349-a0ae-cd537eaa76af | knowledge-augmented-reasoning-distillation | 2305.18395 | null | https://arxiv.org/abs/2305.18395v1 | https://arxiv.org/pdf/2305.18395v1.pdf | Knowledge-Augmented Reasoning Distillation for Small Language Models in Knowledge-Intensive Tasks | Large Language Models (LLMs) have shown promising performance in knowledge-intensive reasoning tasks that require a compound understanding of knowledge. However, deployment of the LLMs in real-world applications can be challenging due to their high computational requirements and concerns on data privacy. Previous studies have focused on building task-specific small language models (LMs) by fine-tuning them with labeled data or distilling LLMs. However, these approaches are ill-suited for knowledge-intensive reasoning tasks due to the limited capacity of small LMs in memorizing the knowledge required. Motivated by our theoretical analysis on memorization, we propose Knowledge-Augmented Reasoning Distillation (KARD), a novel method that fine-tunes small LMs to generate rationales with augmented knowledge retrieved from an external knowledge base. Moreover, we further propose a neural reranker to obtain documents relevant to rationale generation. We empirically show that KARD significantly improves the performance of small T5 and Flan-T5 models on the challenging knowledge-intensive reasoning datasets, namely MedQA-USMLE and StrategyQA. Notably, our method makes the 250M models achieve superior performance against the fine-tuned 3B models, having 12 times larger parameters, on both MedQA-USMLE and StrategyQA benchmarks. | ['Sung Ju Hwang', 'Kenji Kawaguchi', 'Jinheon Baek', 'Seanie Lee', 'Minki Kang'] | 2023-05-28 | null | null | null | null | ['memorization', 'strategyqa'] | ['natural-language-processing', 'reasoning'] | [ 2.04950646e-01 5.82922697e-01 -4.20856446e-01 -2.78864175e-01
-1.08221030e+00 -4.90806848e-01 5.60992062e-01 2.23205924e-01
-6.28588498e-01 1.04462361e+00 4.93399709e-01 -5.19230783e-01
-5.88569880e-01 -8.80008101e-01 -7.13673353e-01 -2.78153241e-01
2.28449807e-01 8.90814424e-01 1.05898477e-01 -3.67455065e-01
1.80608869e-01 1.71065733e-01 -1.05151308e+00 6.27725303e-01
1.27537572e+00 8.80560338e-01 -1.98113903e-01 4.81725961e-01
-3.73450488e-01 1.37650669e+00 -5.67188680e-01 -1.14728296e+00
3.08864027e-01 -1.46026984e-01 -1.23018754e+00 -5.47888219e-01
6.14091873e-01 -4.17758018e-01 -3.88470352e-01 8.76955748e-01
4.38162386e-01 4.95195985e-01 7.27918983e-01 -1.20319092e+00
-1.09893012e+00 1.29029429e+00 -4.30476159e-01 5.57819717e-02
3.86186212e-01 -3.10690284e-01 1.16575682e+00 -8.16274226e-01
5.76473355e-01 1.52744973e+00 5.07639945e-01 8.15038264e-01
-9.66794908e-01 -8.65525067e-01 3.60486388e-01 5.51297724e-01
-1.36925113e+00 -4.06612962e-01 3.92948478e-01 -9.66877304e-03
9.64780211e-01 4.75783050e-01 1.54441148e-01 1.03682065e+00
-2.07937986e-01 1.21459198e+00 9.23811555e-01 -4.76824909e-01
3.81811380e-01 2.67777115e-01 3.40127081e-01 7.84514070e-01
5.20843744e-01 -3.38023812e-01 -9.02397215e-01 -4.68829036e-01
5.24584651e-01 -1.15072981e-01 -1.97260395e-01 -3.89280111e-01
-1.05073130e+00 9.33266580e-01 4.21437144e-01 -2.00092467e-03
-4.15236294e-01 7.18785077e-02 3.10482174e-01 5.07928193e-01
1.08296581e-01 8.69782567e-01 -7.28228569e-01 1.64533094e-01
-9.44288731e-01 3.41814846e-01 8.91183794e-01 1.03784096e+00
4.06521291e-01 -3.71646315e-01 -8.29406500e-01 9.28341627e-01
2.35801756e-01 5.99997699e-01 6.74655795e-01 -9.81402278e-01
9.05139327e-01 8.02194655e-01 1.72541007e-01 -7.38594115e-01
-3.83248091e-01 -3.42854351e-01 -9.29398537e-01 -3.87987792e-01
3.56618732e-01 -4.47747521e-02 -9.19769406e-01 1.90014601e+00
5.07963300e-02 -3.34677026e-02 7.42527664e-01 5.94850421e-01
9.83622313e-01 2.91292071e-01 2.70466536e-01 -1.25899315e-01
1.32619345e+00 -1.18604600e+00 -7.05851376e-01 -3.37953806e-01
7.96503723e-01 -3.85394037e-01 1.19856703e+00 3.61943752e-01
-1.09483290e+00 -1.24624625e-01 -7.17735708e-01 -3.83809835e-01
-4.79637861e-01 6.17028475e-02 9.90894318e-01 7.13083088e-01
-9.45185781e-01 1.22915111e-01 -4.57029074e-01 -1.55428752e-01
7.33242035e-01 3.30593854e-01 -2.16162458e-01 -4.36963409e-01
-1.72308230e+00 1.02282894e+00 7.08440661e-01 1.60756007e-01
-6.37919426e-01 -8.65715086e-01 -8.44129801e-01 2.71530569e-01
9.93310511e-01 -1.12073469e+00 1.30390441e+00 -1.24930657e-01
-1.43559539e+00 5.58057427e-01 -4.58700582e-02 -8.27454686e-01
8.33027422e-01 -5.00136197e-01 -3.41131359e-01 -2.21903138e-02
1.00928731e-02 6.01004779e-01 7.65890419e-01 -9.87381458e-01
-5.91206372e-01 -6.06693365e-02 5.16334057e-01 2.34971195e-01
-6.46943450e-01 -4.72099125e-01 -6.49819911e-01 -6.77563667e-01
-1.19539022e-01 -6.90847456e-01 -4.40895617e-01 -3.73659641e-01
-6.47698998e-01 -2.54969448e-01 3.52101773e-01 -4.99575526e-01
1.45524859e+00 -1.66415071e+00 2.91309715e-03 5.77591836e-01
2.73598999e-01 8.01914930e-01 -4.24263179e-01 3.26408625e-01
3.78221780e-01 1.18788175e-01 6.09718589e-03 -7.46298358e-02
2.47619957e-01 3.93943131e-01 -8.45606506e-01 -4.44535017e-01
-1.88247398e-01 1.35380185e+00 -1.00438058e+00 -6.36197805e-01
-1.21954344e-01 4.80448566e-02 -7.53871381e-01 8.40273499e-03
-5.26784062e-01 -6.86963201e-02 -8.55208755e-01 5.61657071e-01
5.72374642e-01 -5.23148179e-01 3.51635009e-01 -1.07756995e-01
6.87742293e-01 1.05966844e-01 -1.09491992e+00 1.65034187e+00
-6.96432590e-01 1.25074804e-01 -3.01838964e-01 -6.02000058e-01
6.31108940e-01 1.46563828e-01 2.76578199e-02 -8.63136411e-01
-1.39279246e-01 5.93864284e-02 -2.86940157e-01 -5.34383118e-01
7.33408272e-01 -4.19712737e-02 -7.77612180e-02 6.26559556e-01
-3.16309452e-01 -5.71837313e-02 4.04152989e-01 7.38281608e-01
1.38037288e+00 -3.79135370e-01 3.02048057e-01 2.39798203e-01
7.42824435e-01 -8.97728652e-02 3.77778947e-01 1.30546129e+00
1.81681156e-01 5.56998253e-02 3.21179301e-01 -2.70310819e-01
-2.07486823e-01 -1.06853378e+00 2.44128227e-01 1.29948127e+00
-6.94058463e-03 -7.79241264e-01 -4.13353562e-01 -1.12184584e+00
4.03344691e-01 1.21785307e+00 -4.90955234e-01 -4.91303533e-01
-3.54863554e-01 -8.27647865e-01 9.17476296e-01 6.84667349e-01
6.49909079e-01 -1.04126120e+00 -2.93497801e-01 6.93669394e-02
-4.69226509e-01 -1.13379455e+00 -3.07323694e-01 2.40793601e-02
-6.76803231e-01 -1.32256937e+00 -4.22028035e-01 -3.51447165e-01
9.95091498e-01 1.18735030e-01 1.22150588e+00 2.52378378e-02
-3.59341204e-02 5.50927043e-01 -2.84715056e-01 -4.51741874e-01
-3.68136883e-01 2.89728165e-01 -2.49816310e-02 -3.82509977e-01
5.42787671e-01 -3.34587574e-01 -3.48286241e-01 1.78359106e-01
-1.14784467e+00 2.36309450e-02 1.11287951e+00 9.12753105e-01
5.01387417e-01 3.39060694e-01 8.22504282e-01 -1.35139549e+00
1.15349376e+00 -3.62818450e-01 -4.42040056e-01 9.89648640e-01
-9.10959244e-01 5.79968691e-01 5.45280516e-01 -4.06930447e-01
-1.48437142e+00 -3.35912615e-01 1.24567032e-01 -3.48212540e-01
3.98562759e-01 7.82870948e-01 -6.36235476e-02 5.53424843e-02
6.90740883e-01 -5.45250904e-03 -4.41594839e-01 -3.59419227e-01
8.76732945e-01 4.92675960e-01 5.94638824e-01 -9.45913911e-01
9.71513748e-01 3.59914541e-01 -1.24010257e-01 -2.99412102e-01
-1.47548676e+00 -1.68383896e-01 -2.62308687e-01 3.45781773e-01
5.00001431e-01 -1.03512871e+00 -9.17921901e-01 1.25470415e-01
-8.63123059e-01 -3.29374284e-01 -4.10925359e-01 3.81334305e-01
-4.92594004e-01 4.22393054e-01 -5.47693312e-01 -7.55526125e-01
-7.60887682e-01 -7.23104417e-01 7.26616681e-01 9.76783857e-02
-4.35517073e-01 -1.03254580e+00 -9.96366069e-02 1.11203730e+00
4.91543621e-01 -3.14092487e-01 1.47723055e+00 -1.09086359e+00
-7.52736092e-01 -1.30595952e-01 -2.74788916e-01 1.54339045e-01
6.77305311e-02 -6.03157818e-01 -8.95204902e-01 -9.30686072e-02
-4.41528499e-01 -8.84396672e-01 1.27317345e+00 1.31132573e-01
1.42403889e+00 -5.66465378e-01 -3.06666642e-01 3.02175611e-01
9.01682675e-01 -6.61473274e-02 3.75589103e-01 4.35177654e-01
6.20171189e-01 2.43520901e-01 9.34583008e-01 5.18811643e-01
7.77962506e-01 3.24628830e-01 1.78450033e-01 2.44871631e-01
4.44539972e-02 -5.79554856e-01 2.76380837e-01 5.93035996e-01
-1.01139136e-01 -3.30739230e-01 -8.06576610e-01 4.70180959e-01
-2.26498437e+00 -8.58516395e-01 2.15893909e-01 2.12047815e+00
1.33159626e+00 2.20682949e-01 -4.08412874e-01 -1.43257439e-01
1.83761895e-01 1.13263372e-02 -8.42049539e-01 -2.46085197e-01
-2.25671843e-01 2.92241424e-01 4.51504111e-01 5.10972619e-01
-8.41231108e-01 1.14795256e+00 6.39917088e+00 1.14980066e+00
-4.33867335e-01 -5.96472472e-02 3.18892181e-01 -5.27813673e-01
-8.17966998e-01 -2.24039465e-01 -9.32267666e-01 -4.04762477e-03
7.84761786e-01 -4.09735173e-01 5.51990807e-01 8.45691323e-01
-3.82922560e-01 -1.34885624e-01 -1.30042589e+00 9.32172716e-01
1.46280974e-01 -1.62795389e+00 7.74043143e-01 -3.58226120e-01
9.97103333e-01 -4.03444052e-01 2.20041081e-01 9.17501450e-01
9.24788535e-01 -1.11432791e+00 4.68221903e-01 6.40071332e-01
4.68543321e-01 -7.36150563e-01 8.35573614e-01 5.03544629e-01
-8.08730900e-01 -4.00191754e-01 -5.16925991e-01 1.32213593e-01
3.30741778e-02 6.20975852e-01 -1.16358662e+00 7.86257744e-01
4.46751207e-01 2.69133985e-01 -8.65447879e-01 5.85125864e-01
-6.21900439e-01 4.17658180e-01 -1.29599050e-01 -1.36832641e-02
1.61161646e-01 2.52327383e-01 9.48559772e-03 8.92186880e-01
5.49328364e-02 8.57086182e-02 1.14402622e-01 7.37438321e-01
-5.88761091e-01 5.87934218e-02 -1.73849612e-01 -3.76410574e-01
7.27973282e-01 1.10080409e+00 -2.26749063e-01 -4.74292547e-01
-2.20363230e-01 6.75831735e-01 8.02094698e-01 3.79687101e-01
-6.73086166e-01 -2.00319305e-01 5.39428115e-01 -2.97791064e-01
1.54632539e-01 1.54853314e-01 -7.77303874e-02 -1.25237370e+00
1.29879087e-01 -1.06863928e+00 1.19466126e+00 -8.33583295e-01
-1.55097353e+00 4.85248119e-01 1.83835626e-01 -6.31619275e-01
-5.32406509e-01 -3.50863963e-01 -1.59787968e-01 6.69371009e-01
-1.87967598e+00 -1.31338394e+00 -9.64467302e-02 8.41867387e-01
2.79903471e-01 -3.46902668e-01 9.76141214e-01 2.11903974e-01
-5.32342017e-01 1.06566882e+00 1.08420499e-01 7.82003105e-02
9.88757551e-01 -1.16433597e+00 2.05913857e-01 5.88982880e-01
2.39668801e-01 1.15265238e+00 3.68344456e-01 -8.26050639e-01
-1.48168910e+00 -1.27096248e+00 1.03951931e+00 -8.59686971e-01
6.18145943e-01 -1.49246708e-01 -9.11961496e-01 1.03418660e+00
-1.01494089e-01 -2.28802681e-01 9.95208740e-01 4.15667892e-01
-5.59725404e-01 2.41443142e-02 -1.27155113e+00 8.08574557e-01
1.12760162e+00 -6.51122451e-01 -1.09143674e+00 4.37842101e-01
7.26154625e-01 -5.64096451e-01 -8.51390421e-01 2.67591715e-01
4.42799270e-01 -5.34499764e-01 1.26955068e+00 -1.11157620e+00
4.03533816e-01 -9.13191736e-02 -2.51061153e-02 -1.17135620e+00
-2.39067137e-01 -6.55748129e-01 -6.16575420e-01 1.26830018e+00
6.14873528e-01 -5.84917903e-01 9.09525037e-01 1.23573828e+00
4.96944398e-01 -7.13496745e-01 -6.95324183e-01 -7.59487569e-01
-1.22039825e-01 -4.26320255e-01 9.31938410e-01 9.93963420e-01
4.43808772e-02 1.89860612e-01 -4.76988852e-01 2.78760105e-01
4.60884988e-01 3.79663050e-01 7.24887729e-01 -1.15139019e+00
-1.74942970e-01 -2.69787580e-01 1.83682188e-01 -9.74367142e-01
5.29526234e-01 -1.21636605e+00 -2.74002612e-01 -1.86896229e+00
4.07253146e-01 -5.83496332e-01 -6.87340915e-01 1.05394351e+00
-5.31257033e-01 -1.22949168e-01 1.62639201e-01 -1.04732364e-01
-1.09741187e+00 6.36178553e-01 1.20037222e+00 -3.36317211e-01
-2.30475768e-01 7.25097805e-02 -1.15404940e+00 7.63482451e-01
5.33011794e-01 -3.98690343e-01 -1.06216180e+00 -4.68017876e-01
7.11492181e-01 3.27712707e-02 2.74766862e-01 -6.67854190e-01
5.69701910e-01 -3.27182084e-01 3.44151050e-01 -7.20009744e-01
3.47240955e-01 -7.72589564e-01 -9.40790847e-02 3.82074296e-01
-1.12074494e+00 -2.17934832e-01 1.91959620e-01 7.24070132e-01
-1.06908470e-01 -3.06993812e-01 1.45363271e-01 -2.28719085e-01
-8.95856619e-01 2.88289160e-01 -9.84591022e-02 4.84799773e-01
7.34817803e-01 2.08434418e-01 -6.92812681e-01 -4.72819239e-01
-5.22283852e-01 7.39175916e-01 -1.07033387e-01 4.57069755e-01
6.89404547e-01 -1.38546133e+00 -5.05602717e-01 8.66591409e-02
2.68469095e-01 2.32669950e-01 4.09496844e-01 5.69831073e-01
3.08169872e-02 9.25974071e-01 -1.11858733e-01 1.96561649e-01
-1.19638050e+00 7.43430853e-01 -7.43078515e-02 -9.00864422e-01
-2.62104839e-01 1.20449698e+00 1.98227808e-01 -6.82858109e-01
5.39856434e-01 -5.89317262e-01 -2.41330788e-01 8.04006308e-02
6.69757247e-01 5.73497355e-01 8.38374794e-02 1.63244382e-01
-3.77829015e-01 1.28577175e-02 -6.20318651e-01 1.30872563e-01
1.18652916e+00 -3.91442934e-03 -1.02572501e-01 -1.99844390e-01
4.93601412e-01 3.34239215e-01 -7.61793554e-01 -8.40372384e-01
3.20559263e-01 -4.50056225e-01 -1.26513496e-01 -1.34881496e+00
-9.97131705e-01 4.56973284e-01 -9.15016383e-02 -3.21199238e-01
1.13199353e+00 -4.20024544e-02 8.60583186e-01 1.36966276e+00
4.60171133e-01 -1.10822761e+00 2.64371365e-01 6.28384173e-01
9.04109478e-01 -1.13329935e+00 1.29215389e-01 -3.06738973e-01
-8.78686965e-01 8.84130776e-01 8.61231506e-01 8.18198800e-01
3.25688034e-01 -1.65658202e-02 1.73483446e-01 -2.28839070e-01
-1.02253854e+00 -9.30178910e-02 5.07644892e-01 4.04316813e-01
2.96846423e-02 1.17515922e-02 -6.60847425e-02 1.11605060e+00
-3.89442146e-01 3.84815425e-01 3.28662902e-01 1.04742384e+00
-1.75324932e-01 -1.18840253e+00 -3.00334364e-01 6.89471006e-01
-1.51486129e-01 -3.89930844e-01 -6.00134730e-01 6.06070876e-01
-1.67409286e-01 1.06306565e+00 -4.52267915e-01 -2.78164506e-01
3.49334806e-01 3.15136582e-01 3.97376508e-01 -4.85094517e-01
-5.72927535e-01 -7.79479444e-01 3.05160820e-01 -5.63207984e-01
-3.00202012e-01 -2.06196368e-01 -1.45338678e+00 -3.54101390e-01
-1.71254590e-01 3.81167173e-01 4.58598211e-02 1.07996047e+00
5.50710261e-01 5.48174679e-01 -8.65401104e-02 1.12197347e-01
-1.03604746e+00 -6.50453389e-01 -4.74546999e-01 3.58712792e-01
6.53699785e-02 -5.61948419e-01 1.26435012e-01 -3.20350051e-01] | [10.739303588867188, 7.984165191650391] |
b183a6c7-d5d1-4023-8a51-6ef1c030c2e7 | monolingual-versus-multilingual-bertology-for | 2108.13741 | null | https://arxiv.org/abs/2108.13741v3 | https://arxiv.org/pdf/2108.13741v3.pdf | Monolingual versus Multilingual BERTology for Vietnamese Extractive Multi-Document Summarization | Recent researches have demonstrated that BERT shows potential in a wide range of natural language processing tasks. It is adopted as an encoder for many state-of-the-art automatic summarizing systems, which achieve excellent performance. However, so far, there is not much work done for Vietnamese. In this paper, we showcase how BERT can be implemented for extractive text summarization in Vietnamese on multi-document. We introduce a novel comparison between different multilingual and monolingual BERT models. The experiment results indicate that monolingual models produce promising results compared to other multilingual models and previous text summarizing models for Vietnamese. | ['Huy Quoc To', 'Anh Gia-Tuan Nguyen', 'Ngan Luu-Thuy Nguyen', 'Kiet Van Nguyen'] | 2021-08-31 | null | null | null | null | ['extractive-document-summarization'] | ['natural-language-processing'] | [-2.34348968e-01 1.97961062e-01 -3.32519233e-01 -2.01924041e-01
-1.08358908e+00 -3.93140942e-01 8.43909442e-01 5.37995458e-01
-6.67319477e-01 1.21722353e+00 1.07394660e+00 -2.39077225e-01
3.36151451e-01 -3.41567516e-01 -4.42519665e-01 -1.88076839e-01
1.51981503e-01 5.41645706e-01 2.16313288e-01 -7.32891262e-01
5.87323904e-01 2.73835450e-01 -9.86908853e-01 5.60081959e-01
1.40567636e+00 -1.05390608e-01 4.45854962e-01 8.68997216e-01
-3.87936115e-01 1.02913606e+00 -1.47552538e+00 -8.65441680e-01
-3.61535132e-01 -7.06267774e-01 -9.52884257e-01 -2.92216718e-01
7.46347964e-01 -3.15386623e-01 -3.10044616e-01 1.03752387e+00
8.30252290e-01 -1.52801827e-01 8.14408898e-01 -8.39886010e-01
-8.39850128e-01 1.34567952e+00 -4.83917654e-01 3.64389777e-01
7.06871867e-01 -4.74110842e-01 1.13091683e+00 -7.24797249e-01
9.89150465e-01 1.27468610e+00 4.27672118e-01 4.66959000e-01
-5.56820512e-01 -4.52564359e-01 -4.95562181e-02 4.20455396e-01
-9.49366987e-01 -6.10263526e-01 7.70845234e-01 1.32501394e-01
1.52422118e+00 3.82871360e-01 7.19930768e-01 9.70905960e-01
1.10775971e+00 1.30882943e+00 7.36908674e-01 -5.15467525e-01
-3.16489875e-01 1.26320064e-01 2.45166168e-01 8.06144357e-01
4.21844006e-01 -5.74915469e-01 -6.25362217e-01 1.47338942e-01
1.81680351e-01 -6.06902957e-01 -7.51033649e-02 5.02930582e-01
-1.33529615e+00 8.98508549e-01 8.58618841e-02 6.21190250e-01
-4.53751385e-01 -3.09503581e-02 1.16965187e+00 4.35967267e-01
7.63962150e-01 5.49362719e-01 -1.33584931e-01 -1.83952361e-01
-1.21613288e+00 3.88945013e-01 1.10027766e+00 1.26209903e+00
3.73901874e-01 4.43351299e-01 -3.22071254e-01 1.01296329e+00
-1.51742905e-01 6.95816994e-01 7.82082081e-01 -7.30964661e-01
9.63796437e-01 2.16745928e-01 -3.27957511e-01 -7.70121634e-01
-2.96350449e-01 -2.65393943e-01 -1.10730898e+00 -4.67868209e-01
-3.39552581e-01 -3.31635296e-01 -3.95702779e-01 1.09638488e+00
-2.50243217e-01 -6.26184881e-01 4.99723256e-01 3.86340857e-01
1.30981636e+00 1.40843868e+00 -8.22558329e-02 -8.00458848e-01
1.09244180e+00 -1.48778284e+00 -1.38543737e+00 -1.81736082e-01
6.17095768e-01 -1.29838085e+00 7.53686428e-01 5.50406158e-01
-1.46765256e+00 -6.21634841e-01 -1.18237042e+00 -6.04055643e-01
-3.91536176e-01 4.59228486e-01 6.29155457e-01 4.68526691e-01
-1.08223748e+00 3.08899701e-01 -5.47967076e-01 -7.64035046e-01
-1.06105655e-01 -1.25601087e-02 -4.07229632e-01 9.03546717e-03
-1.22761798e+00 1.27923000e+00 8.53717208e-01 8.75735134e-02
-6.14571095e-01 -2.24765539e-02 -8.52527201e-01 1.80422127e-01
3.70469481e-01 -7.90900826e-01 1.56533229e+00 -5.83639085e-01
-1.70498371e+00 4.11251307e-01 -3.91821384e-01 -7.38958240e-01
2.94181705e-01 -5.29814363e-01 -6.65098667e-01 3.44706148e-01
4.46252793e-01 6.49180532e-01 5.12744248e-01 -1.00940955e+00
-6.95910037e-01 -4.25318144e-02 -1.39122754e-01 4.95544851e-01
-4.51942265e-01 4.81638253e-01 -3.68089408e-01 -7.15248048e-01
-6.07339144e-01 -3.61292899e-01 -5.10834679e-02 -8.30215037e-01
-6.64645374e-01 -6.92686558e-01 8.02168846e-01 -1.20522928e+00
1.77426779e+00 -1.62377179e+00 2.13898510e-01 -4.67207968e-01
-1.53266847e-01 6.65348411e-01 -2.99099505e-01 1.47799015e+00
3.48897606e-01 3.45618010e-01 -2.02829972e-01 -3.09417993e-01
3.30583230e-02 2.98733979e-01 -4.60515529e-01 1.02593526e-01
1.14379987e-01 1.09636664e+00 -8.67238343e-01 -1.20543993e+00
1.34716675e-01 1.56388968e-01 -3.30471814e-01 9.85870510e-02
-1.75553769e-01 -1.09913610e-02 -4.95933950e-01 4.73395824e-01
4.08625573e-01 4.96266335e-01 2.65760511e-01 6.07227609e-02
-6.10921085e-01 4.71203387e-01 -5.82243562e-01 2.01686311e+00
-3.12405944e-01 9.04644668e-01 -2.81014107e-02 -8.20556402e-01
8.39290202e-01 5.84902942e-01 1.11193009e-01 -4.72014189e-01
1.14979364e-01 5.03569543e-01 4.67353985e-02 -7.03870356e-01
1.51693606e+00 -1.22324983e-02 -3.98904324e-01 3.41803342e-01
3.25199693e-01 -7.56207705e-01 1.15637279e+00 8.10051858e-01
8.15275311e-01 -4.66292575e-02 1.03437245e+00 -3.34328681e-01
9.10144389e-01 4.86236870e-01 3.41843188e-01 7.18999445e-01
-8.98833722e-02 4.94930774e-01 5.49941897e-01 8.78315121e-02
-1.28867066e+00 -8.41241956e-01 1.51925921e-01 8.77293646e-01
-2.92433977e-01 -9.86220956e-01 -9.66171026e-01 -6.85417771e-01
-3.90366435e-01 1.03226995e+00 -1.06792100e-01 2.24598393e-01
-1.11984611e+00 -4.82473433e-01 1.08174241e+00 2.95827180e-01
8.43780756e-01 -1.38396597e+00 -2.88116962e-01 5.22292495e-01
-6.07709646e-01 -1.07579958e+00 -6.04138196e-01 -1.56504914e-01
-8.49585950e-01 -5.71851850e-01 -8.59725296e-01 -1.22723651e+00
3.92898209e-02 3.61008227e-01 1.09429979e+00 -2.28078812e-02
8.96356553e-02 9.90735590e-02 -6.54507220e-01 -8.95509601e-01
-9.76257026e-01 7.68990278e-01 -2.20328927e-01 -7.62488246e-01
7.46825114e-02 -7.96747357e-02 -1.93636678e-02 -4.70019549e-01
-1.03739285e+00 1.84778288e-01 8.33680511e-01 7.19747722e-01
7.89781436e-02 -6.96653575e-02 1.01809895e+00 -1.17281938e+00
1.48821211e+00 -2.87120998e-01 -4.97896783e-02 6.30002916e-01
-2.98724025e-01 1.89013079e-01 1.16944253e+00 -4.57879715e-02
-1.48959112e+00 -5.00906169e-01 -7.33394325e-01 5.84116995e-01
1.71644807e-01 1.10166144e+00 6.57851323e-02 5.29015303e-01
4.58309770e-01 3.06218535e-01 -3.26848894e-01 -4.36013967e-01
5.15492618e-01 1.08080959e+00 5.60557723e-01 -3.89658123e-01
3.65184098e-01 6.79743811e-02 -3.74600738e-01 -1.42841208e+00
-8.21212411e-01 -6.23540223e-01 -7.42691398e-01 -2.23431304e-01
9.38388288e-01 -8.09838593e-01 -6.18905351e-02 1.98784694e-01
-1.68844247e+00 1.49174646e-01 -2.35327467e-01 4.76804405e-01
-3.94600332e-01 8.09511125e-01 -8.36129189e-01 -3.98579359e-01
-1.07478070e+00 -9.02261376e-01 1.18435717e+00 1.91340521e-01
-2.96305060e-01 -1.10134745e+00 3.79801989e-01 2.67920375e-01
2.77510136e-01 -1.27672583e-01 9.94372845e-01 -6.14944935e-01
7.98753947e-02 -1.64909482e-01 6.02201894e-02 4.98452127e-01
-9.08756554e-02 3.85574341e-01 -4.08367008e-01 -2.06304401e-01
1.51122808e-02 -3.12259257e-01 1.07011104e+00 3.08642417e-01
5.43200672e-01 -5.92858076e-01 2.24838946e-02 3.83881368e-02
1.22474062e+00 2.52147049e-01 7.20526576e-01 2.00006038e-01
6.04794919e-01 5.00309467e-01 7.92049289e-01 2.52084285e-01
6.01918399e-01 2.73745984e-01 -4.97238673e-02 -2.18943860e-02
-3.91361833e-01 -4.14856493e-01 8.39415967e-01 2.22503090e+00
6.77891225e-02 -7.79693961e-01 -6.96891546e-01 5.08202553e-01
-1.90656459e+00 -1.28769338e+00 -4.77240235e-01 1.60168517e+00
8.88991594e-01 -4.28396761e-02 -6.96920976e-02 -1.00101121e-01
5.42606890e-01 6.80602133e-01 -3.04471310e-02 -1.30221832e+00
-4.43486661e-01 2.26408213e-01 2.19768584e-01 5.56363046e-01
-8.59454036e-01 1.36964881e+00 6.77479076e+00 9.05964077e-01
-9.28877950e-01 1.73400819e-01 -2.05076206e-02 1.33672744e-01
-2.38987789e-01 -6.85943291e-02 -8.03003848e-01 -9.27913710e-02
1.15705609e+00 -9.50313807e-01 9.64222476e-03 4.69720691e-01
3.37297738e-01 -2.06717834e-01 -8.26115429e-01 6.24963939e-01
8.29661429e-01 -1.36617506e+00 6.54016793e-01 -2.67150164e-01
9.92096901e-01 1.16626076e-01 -4.65172261e-01 5.98156035e-01
1.55863658e-01 -6.11728013e-01 7.41195619e-01 3.03775162e-01
5.61543882e-01 -1.06119335e+00 1.01837623e+00 7.25314260e-01
-1.08250785e+00 1.54273227e-01 -9.15082872e-01 1.24305181e-01
5.89713991e-01 4.32515889e-01 -8.63666832e-01 1.32808506e+00
1.05543680e-01 1.03474700e+00 -7.29659259e-01 9.89105940e-01
-6.73648059e-01 6.76082671e-01 1.97972074e-01 -5.36018729e-01
5.42454958e-01 -3.06881011e-01 7.06178963e-01 1.88210034e+00
4.41077709e-01 -4.91081290e-02 1.64176807e-01 2.75545847e-02
-2.54279912e-01 9.41760719e-01 -8.01278651e-01 -2.94162154e-01
5.25650345e-02 9.02194798e-01 -3.48302156e-01 -1.06806803e+00
-4.00555700e-01 1.21765292e+00 5.24985135e-01 3.01590621e-01
-6.41786695e-01 -8.11919808e-01 -2.95801908e-01 -4.42858040e-01
4.70616333e-02 -5.93403518e-01 -9.80933979e-02 -1.54671037e+00
-1.20914795e-01 -1.17178738e+00 2.80956358e-01 -8.16893816e-01
-9.32506680e-01 7.13562548e-01 3.39186221e-01 -1.04433274e+00
-6.03407741e-01 -2.77677566e-01 -8.20612788e-01 5.27753770e-01
-1.43475509e+00 -1.08339536e+00 3.48289847e-01 3.07842582e-01
1.38699126e+00 -3.81892622e-01 9.73207831e-01 1.81619376e-01
-6.48246527e-01 1.84831262e-01 3.84461254e-01 -4.44205366e-02
1.03086364e+00 -1.40697110e+00 5.22265732e-01 1.25898278e+00
3.38616282e-01 6.37951195e-01 8.31960261e-01 -8.59674096e-01
-1.43060863e+00 -9.40757751e-01 1.94473720e+00 1.05360612e-01
7.91214645e-01 -8.66957605e-02 -7.51773655e-01 7.33838797e-01
1.51099813e+00 -1.04161966e+00 4.78465974e-01 8.38564783e-02
6.48473278e-02 -6.27362654e-02 -6.73276365e-01 8.17323208e-01
6.81557119e-01 -2.88916767e-01 -1.16694582e+00 6.15563393e-01
7.35170722e-01 -1.81732744e-01 -9.21088815e-01 1.91091727e-02
4.10528690e-01 -8.03018451e-01 3.66703719e-01 -3.01356375e-01
8.41223121e-01 -1.56833194e-02 3.42240483e-01 -1.91455853e+00
-5.13563864e-02 -6.95370257e-01 -2.79839411e-02 1.51194537e+00
3.99457157e-01 -5.98791957e-01 1.13567919e-01 -4.66406792e-01
-8.27825665e-01 -2.27570817e-01 -7.66080201e-01 -7.91617334e-01
5.48650861e-01 3.56138162e-02 4.16363120e-01 5.90196073e-01
4.55655843e-01 1.18041646e+00 -5.50267577e-01 -3.89371336e-01
2.72325546e-01 6.88128695e-02 8.31273377e-01 -8.92410874e-01
1.53618380e-01 -5.63580751e-01 6.57824725e-02 -1.21376717e+00
5.62219024e-01 -1.07440603e+00 3.89635526e-02 -2.42721701e+00
4.37480569e-01 4.91452932e-01 3.57791066e-01 1.32776603e-01
-2.57691145e-01 -1.68967336e-01 3.89902741e-01 1.76491335e-01
-7.52615094e-01 7.64426947e-01 1.46836269e+00 -4.23939407e-01
-9.09734592e-02 -3.60985041e-01 -7.51732171e-01 3.95153522e-01
1.20884538e+00 -3.95159036e-01 -3.73075575e-01 -7.58567214e-01
9.96309966e-02 3.56404930e-01 -4.82474446e-01 -1.00839889e+00
4.44526315e-01 8.90749022e-02 -3.47410738e-02 -1.21736622e+00
3.53440866e-02 -2.79037654e-01 -2.13128492e-01 4.40285921e-01
-4.59195405e-01 7.45072961e-01 1.23541228e-01 5.70899285e-02
-9.24512029e-01 -7.43802369e-01 2.08064258e-01 -2.78874785e-01
-6.52445495e-01 -6.21830970e-02 -9.63734508e-01 1.14908673e-01
6.04422510e-01 2.13347003e-01 -7.81601727e-01 -5.31260967e-01
7.99383596e-02 2.22897366e-01 7.18679279e-02 3.82442266e-01
7.00611889e-01 -8.54791760e-01 -1.35277987e+00 -1.73105240e-01
-3.35796811e-02 -5.00828624e-01 -8.00632909e-02 6.01670623e-01
-1.05421507e+00 1.05618548e+00 -4.52484906e-01 -1.95859790e-01
-1.46662855e+00 4.64379221e-01 -2.57876366e-01 -7.78572261e-01
-6.84779108e-01 1.07323468e-01 -4.06176299e-01 -3.91560376e-01
-6.93264157e-02 -2.70450205e-01 -5.56924880e-01 3.31170321e-01
5.67071557e-01 5.93320131e-01 1.56923473e-01 -7.95274019e-01
3.42863165e-02 2.40351990e-01 -4.98197466e-01 -4.29646432e-01
1.15830231e+00 -2.07115024e-01 -6.08324111e-01 6.95103943e-01
1.19465709e+00 4.46892887e-01 -1.81034505e-01 1.04132503e-01
1.33261055e-01 3.97917023e-03 -3.24598551e-01 -6.94830179e-01
-4.25346166e-01 1.04355419e+00 -2.11897448e-01 4.30132985e-01
1.03055000e+00 -2.72299945e-01 1.04376650e+00 1.00261557e+00
2.76015729e-01 -1.51867986e+00 -1.83277488e-01 1.08824241e+00
1.29797721e+00 -1.11954081e+00 5.16072989e-01 -3.65814418e-01
-9.77622092e-01 1.39735687e+00 2.17111975e-01 -1.65971220e-01
3.61187384e-02 2.87848979e-01 1.22362509e-01 -1.73211098e-01
-9.59457755e-01 -3.58057544e-02 1.29321605e-01 2.80815899e-01
1.02474201e+00 2.44746432e-01 -1.36176109e+00 -3.43970358e-02
-7.87642121e-01 -2.94552892e-01 1.03501189e+00 9.33210075e-01
-7.52059698e-01 -1.34518194e+00 -1.35980353e-01 4.69426602e-01
-7.55992532e-01 -4.62045938e-01 -7.00313210e-01 1.21853828e+00
-4.08537745e-01 1.17785740e+00 -3.47587466e-01 5.37526347e-02
4.16203886e-01 6.04838841e-02 5.68584085e-01 -8.94837737e-01
-1.20958114e+00 2.99679965e-01 7.30133533e-01 4.17607501e-02
-6.04000032e-01 -7.60254264e-01 -1.12573922e+00 -4.80850935e-01
-2.05274791e-01 7.11504638e-01 7.13568509e-01 7.30468929e-01
-1.01842195e-01 6.17850661e-01 3.59827608e-01 -6.73417091e-01
-5.60624361e-01 -1.43749011e+00 -5.13591349e-01 -1.91999197e-01
1.91818163e-01 2.21076801e-01 -4.38738465e-02 1.23607963e-01] | [12.386953353881836, 9.534976959228516] |
4efe39fe-626c-4c8e-a3a2-672a9426052c | hdrunet-single-image-hdr-reconstruction-with | 2105.13084 | null | https://arxiv.org/abs/2105.13084v2 | https://arxiv.org/pdf/2105.13084v2.pdf | HDRUNet: Single Image HDR Reconstruction with Denoising and Dequantization | Most consumer-grade digital cameras can only capture a limited range of luminance in real-world scenes due to sensor constraints. Besides, noise and quantization errors are often introduced in the imaging process. In order to obtain high dynamic range (HDR) images with excellent visual quality, the most common solution is to combine multiple images with different exposures. However, it is not always feasible to obtain multiple images of the same scene and most HDR reconstruction methods ignore the noise and quantization loss. In this work, we propose a novel learning-based approach using a spatially dynamic encoder-decoder network, HDRUNet, to learn an end-to-end mapping for single image HDR reconstruction with denoising and dequantization. The network consists of a UNet-style base network to make full use of the hierarchical multi-scale information, a condition network to perform pattern-specific modulation and a weighting network for selectively retaining information. Moreover, we propose a Tanh_L1 loss function to balance the impact of over-exposed values and well-exposed values on the network learning. Our method achieves the state-of-the-art performance in quantitative comparisons and visual quality. The proposed HDRUNet model won the second place in the single frame track of NITRE2021 High Dynamic Range Challenge. | ['Chao Dong', 'Yu Qiao', 'Zhengwen Zhang', 'Yihao Liu', 'Xiangyu Chen'] | 2021-05-27 | null | null | null | null | ['hdr-reconstruction'] | ['computer-vision'] | [ 5.70865333e-01 -4.82855529e-01 1.06758587e-01 -4.93512034e-01
-8.18153381e-01 -1.44814819e-01 1.65653244e-01 -2.94800609e-01
-6.07219636e-01 6.66225374e-01 -1.02910702e-03 -3.36410291e-02
-8.00282732e-02 -7.90654957e-01 -8.04659843e-01 -8.46865594e-01
2.99674898e-01 -2.20987350e-01 4.06211525e-01 -2.26394013e-01
-3.29365358e-02 4.34835464e-01 -1.59510219e+00 3.15270007e-01
9.01491523e-01 1.28406608e+00 7.28341579e-01 6.68965220e-01
3.26965332e-01 1.05379868e+00 -3.43906909e-01 -2.71408200e-01
6.20080888e-01 -4.91272569e-01 -1.43351436e-01 9.80505943e-02
7.21010983e-01 -7.36318171e-01 -8.16213906e-01 1.21252537e+00
9.33620393e-01 7.75940670e-03 3.01553011e-01 -8.51432145e-01
-5.87558806e-01 2.68027812e-01 -7.57980049e-01 2.29740277e-01
1.10030174e-01 5.58667898e-01 7.97896147e-01 -7.08908617e-01
6.04396582e-01 9.56420898e-01 5.14075398e-01 4.36619014e-01
-1.21571934e+00 -8.07865024e-01 -2.44770095e-01 5.15219748e-01
-1.40035427e+00 -5.92050970e-01 7.59089351e-01 -1.30566791e-01
7.69044518e-01 -8.63524061e-03 5.60951591e-01 8.31066608e-01
3.88963372e-01 3.85819435e-01 1.42446339e+00 -1.94972858e-01
1.66846812e-02 -2.10284367e-01 -3.34419340e-01 4.96250600e-01
-2.09942497e-02 3.75477850e-01 -6.69355989e-01 5.03773510e-01
9.79341567e-01 -9.48350504e-03 -6.28295898e-01 -1.04817018e-01
-1.05732489e+00 3.40282857e-01 6.57058537e-01 8.63929316e-02
-2.40772769e-01 2.81459808e-01 2.50516117e-01 4.50721532e-01
3.35849494e-01 1.65715396e-01 -2.45576233e-01 -2.80198026e-02
-1.09638619e+00 -5.12759313e-02 1.87183768e-01 7.32043743e-01
7.22144425e-01 6.01058230e-02 -3.94315898e-01 1.00832987e+00
2.65063822e-01 5.72279751e-01 1.66591257e-01 -1.15319824e+00
5.19476712e-01 1.11658514e-01 8.94502774e-02 -8.33378255e-01
-1.94566518e-01 -4.24641490e-01 -1.30870950e+00 5.71178138e-01
2.53078520e-01 8.32198039e-02 -9.22526121e-01 1.64398766e+00
-2.80566122e-02 1.78289399e-01 -1.24602571e-01 1.25871730e+00
6.51689053e-01 8.81866574e-01 -9.36958119e-02 -3.67526531e-01
1.13011372e+00 -7.59960055e-01 -9.08549786e-01 -1.31370127e-01
-7.96876177e-02 -8.09847057e-01 1.04276097e+00 6.73865199e-01
-1.34381711e+00 -8.47075820e-01 -1.41400599e+00 -4.75714862e-01
6.67867586e-02 1.41117305e-01 1.89977780e-01 5.43370843e-01
-1.01034284e+00 6.99773788e-01 -5.48791230e-01 1.09167144e-01
4.48575884e-01 3.35838288e-01 -1.96536615e-01 -6.36012793e-01
-1.21810055e+00 7.87393749e-01 3.39475155e-01 2.94011950e-01
-1.03181267e+00 -7.90630460e-01 -6.41469538e-01 -3.87288295e-02
4.03989255e-01 -7.03346670e-01 6.84801459e-01 -8.12128067e-01
-1.71169710e+00 6.63731456e-01 1.21592775e-01 -4.76001769e-01
6.77839100e-01 -1.49659082e-01 -4.88738716e-01 2.98911572e-01
-2.69646794e-01 6.59540594e-01 9.96121585e-01 -1.17268419e+00
-6.09834731e-01 -2.82661557e-01 5.61054796e-02 3.57283622e-01
-2.01151237e-01 -3.98056349e-03 -6.90962017e-01 -6.07534587e-01
6.62910519e-04 -6.24729216e-01 1.53873526e-02 4.70983505e-01
-1.79355279e-01 3.71077567e-01 9.62922752e-01 -9.00555909e-01
1.20503724e+00 -2.26434755e+00 7.62435198e-02 -1.82085171e-01
3.52693886e-01 3.93368304e-01 -2.94038147e-01 -8.14181045e-02
-3.64084095e-02 -2.87275702e-01 -3.60653758e-01 -3.05747896e-01
-2.99862802e-01 8.58759787e-03 -1.63861275e-01 6.72925413e-01
6.47339150e-02 5.74418128e-01 -8.00649166e-01 -4.46495622e-01
7.50073731e-01 8.57087135e-01 -3.99415046e-01 3.97572935e-01
-1.25892058e-01 6.02073193e-01 1.90645736e-02 4.02542800e-01
9.84417796e-01 -1.25691682e-01 7.35738277e-02 -8.17934990e-01
-2.95072734e-01 5.14390357e-02 -1.22304142e+00 1.85578084e+00
-6.81751013e-01 7.97962964e-01 2.93900818e-01 -5.45410395e-01
8.78156483e-01 1.19150497e-01 4.29438770e-01 -1.18002653e+00
6.10776432e-02 4.08032715e-01 -2.90809982e-02 -4.08434182e-01
4.47627634e-01 -2.65502959e-01 2.19907790e-01 -9.83017683e-02
-8.32856297e-02 -2.84667939e-01 1.42364740e-01 -1.29039168e-01
1.10850930e+00 8.31674263e-02 9.09171849e-02 1.09406762e-01
4.76114571e-01 -6.85882449e-01 6.96421802e-01 5.05143225e-01
-1.33088559e-01 1.11670768e+00 2.74646878e-01 -2.53489196e-01
-1.52353883e+00 -1.20806277e+00 -2.57483363e-01 7.75482535e-01
2.80588716e-01 -1.15661420e-01 -3.84562254e-01 -1.19602293e-01
-4.43137169e-01 5.37120700e-01 -1.27146646e-01 -2.42299795e-01
-5.77235639e-01 -7.71693349e-01 4.24755752e-01 1.45835727e-01
1.17903209e+00 -7.81767309e-01 -7.00930417e-01 1.53050542e-01
-3.16822201e-01 -1.32860136e+00 -5.97466469e-01 3.39848489e-01
-5.03487587e-01 -8.49257469e-01 -7.31332779e-01 -5.47748685e-01
1.44148707e-01 3.06341827e-01 1.09838283e+00 -1.88151076e-01
-3.62315625e-01 -1.29216447e-01 -2.31622100e-01 1.15551651e-01
-4.19805586e-01 -1.14810444e-01 -2.60721534e-01 1.03862368e-01
-1.46661505e-01 -6.36040211e-01 -1.04197586e+00 3.78371209e-01
-1.16511345e+00 3.41124773e-01 7.08889782e-01 1.01291621e+00
7.75825143e-01 5.06322443e-01 4.36082661e-01 -4.99525398e-01
1.79021358e-01 -1.10451177e-01 -8.86313856e-01 2.02151716e-01
-6.03139818e-01 -1.52494282e-01 7.81401157e-01 -3.06756079e-01
-1.20886219e+00 1.64112568e-01 -3.88146073e-01 -6.45492315e-01
1.66443676e-01 2.72768382e-02 -5.20093799e-01 -4.04245317e-01
4.27412122e-01 3.85059118e-01 -5.01747131e-02 -2.77793735e-01
3.23603541e-01 7.65024960e-01 7.38570988e-01 -1.67059183e-01
8.49691451e-01 4.64510798e-01 2.29743674e-01 -9.16453302e-01
-7.78418303e-01 -2.87405372e-01 -2.39687338e-01 -4.78770822e-01
1.01125348e+00 -1.38470268e+00 -6.22196972e-01 9.49685514e-01
-9.87218142e-01 -5.40297568e-01 -3.06077212e-01 3.58442575e-01
-5.55699289e-01 3.91801089e-01 -7.85087943e-01 -5.10172307e-01
-2.84952909e-01 -1.34230316e+00 1.01418543e+00 2.92099744e-01
5.83475947e-01 -4.82889444e-01 -2.70265698e-01 3.64870489e-01
7.35467553e-01 1.78923577e-01 7.82920659e-01 3.74542773e-01
-1.01397443e+00 2.26563781e-01 -7.05460489e-01 8.72533441e-01
8.86140615e-02 -2.88848221e-01 -1.00518537e+00 -5.27471066e-01
1.67484567e-01 -4.22534585e-01 1.11878848e+00 5.89446425e-01
1.43792677e+00 -6.05628714e-02 2.14076862e-01 1.13542807e+00
1.92896724e+00 2.00949684e-01 1.38686752e+00 1.70861095e-01
9.37171698e-01 2.49806836e-01 4.63982463e-01 4.75584298e-01
1.83640167e-01 1.03753853e+00 4.51226592e-01 -2.26620764e-01
-7.07379460e-01 -1.38373539e-01 4.58452642e-01 5.69895566e-01
5.99398650e-02 -5.32034934e-01 -5.82264841e-01 1.84929773e-01
-1.37786353e+00 -9.67669368e-01 -1.00743972e-01 2.29804564e+00
1.06993127e+00 1.69177204e-01 -2.64312208e-01 1.69360235e-01
6.29337907e-01 3.98177296e-01 -6.63316011e-01 1.63681060e-01
-6.18025005e-01 1.58928156e-01 8.89798105e-01 4.03114766e-01
-9.35699582e-01 5.48810482e-01 5.58782959e+00 1.20170927e+00
-1.30607235e+00 1.76182806e-01 9.58161414e-01 -4.23511505e-01
-1.39275759e-01 -2.45599747e-01 -7.59190261e-01 6.94884956e-01
9.48248982e-01 2.11221099e-01 6.80192828e-01 3.14764291e-01
5.35636246e-01 -2.52787441e-01 -8.48138392e-01 1.31173563e+00
1.59014631e-02 -1.25963581e+00 -1.41496524e-01 5.87771228e-03
8.24191034e-01 7.56280795e-02 4.35946107e-01 -1.11413635e-01
-8.58959407e-02 -1.06031954e+00 7.50018418e-01 6.60131693e-01
1.36182761e+00 -6.90602124e-01 6.22144818e-01 5.46127409e-02
-1.15261018e+00 -2.85706878e-01 -4.75070387e-01 2.69546926e-01
3.72829407e-01 9.70763981e-01 -1.22310400e-01 5.18728673e-01
9.46883798e-01 8.60441983e-01 -5.56046128e-01 1.15697098e+00
-2.66568214e-01 3.09108883e-01 -2.18073860e-01 5.57799280e-01
-9.74011645e-02 -2.63364315e-01 4.92968440e-01 9.76535439e-01
5.06865263e-01 1.47569776e-01 2.11876836e-02 6.52779877e-01
-3.53727490e-01 -3.84006649e-01 -4.08189327e-01 4.38347816e-01
3.10214162e-01 1.10174561e+00 -3.65525782e-01 -7.72587582e-02
-4.78761286e-01 1.24587047e+00 -6.46633580e-02 2.66852140e-01
-9.10048664e-01 -3.47481906e-01 5.02278507e-01 4.10560459e-01
3.93831789e-01 -2.03763232e-01 -2.05173716e-01 -1.09090805e+00
1.32646441e-01 -1.10730016e+00 -1.57882944e-02 -1.12283766e+00
-1.21540380e+00 5.52240551e-01 -1.23236291e-01 -1.38842535e+00
2.12727740e-01 -5.38182735e-01 -1.36426598e-01 9.47156370e-01
-1.96793950e+00 -8.74416351e-01 -5.66538036e-01 6.29853606e-01
5.79788566e-01 -5.95628358e-02 1.37767702e-01 9.87595022e-01
-5.52970290e-01 4.44717109e-01 3.30989867e-01 -1.42803431e-01
1.02684426e+00 -1.00504625e+00 -1.37781780e-02 1.03155279e+00
-3.85344774e-01 1.39127836e-01 7.30949640e-01 -5.03178835e-01
-1.37257540e+00 -1.30051136e+00 3.84022236e-01 1.94028601e-01
3.34071606e-01 -3.14995170e-01 -1.01988220e+00 1.62881836e-01
2.78913409e-01 4.37866151e-01 2.48127002e-02 -5.64837337e-01
-3.59202892e-01 -7.81033576e-01 -1.23824990e+00 4.60966289e-01
9.71990108e-01 -6.72908008e-01 1.09643035e-01 9.52078104e-02
8.33787680e-01 -4.99823511e-01 -9.31753397e-01 4.64151323e-01
4.90063280e-01 -1.34998751e+00 1.15250206e+00 4.60689098e-01
7.06685007e-01 -7.94263721e-01 -3.60204011e-01 -1.18004072e+00
-1.38060883e-01 -6.44908130e-01 -1.91030502e-02 1.18909061e+00
3.19330804e-02 -3.33682388e-01 5.14635980e-01 3.14435452e-01
-1.23344921e-01 -6.37590766e-01 -1.14937854e+00 -7.17053175e-01
-3.07935745e-01 -2.40806773e-01 2.97816426e-01 4.23587501e-01
-7.85170138e-01 3.09023410e-01 -9.49580073e-01 9.23440680e-02
1.06755304e+00 -2.31106818e-01 4.22389001e-01 -7.87048340e-01
-4.96522695e-01 -1.40909538e-01 -4.70012516e-01 -1.25062811e+00
-1.61115304e-01 -4.73005623e-01 3.43778700e-01 -1.35266829e+00
2.78156191e-01 -4.02307332e-01 -3.52322638e-01 -1.80066749e-02
-1.09512685e-03 6.23009861e-01 3.69332582e-01 1.27330005e-01
-7.57841945e-01 7.02596426e-01 1.40377963e+00 -1.36220202e-01
-7.85322860e-02 -3.21800143e-01 -3.29747975e-01 4.86614019e-01
6.06252849e-01 -3.53309184e-01 -4.43666846e-01 -7.19802082e-01
3.73266429e-01 2.88446367e-01 6.21163726e-01 -1.35912979e+00
3.36320013e-01 6.96888268e-02 6.45151675e-01 -4.91536736e-01
3.81672412e-01 -1.01049054e+00 3.88049155e-01 1.87732652e-01
-3.39223623e-01 -1.46762922e-01 9.65556689e-03 6.26785696e-01
-2.94551462e-01 6.46633729e-02 1.48219848e+00 7.55726993e-02
-7.90851474e-01 4.61243391e-01 2.89471122e-03 -4.13149074e-02
7.58001208e-01 -1.50714323e-01 -5.13011813e-01 -3.35671514e-01
-2.41026014e-01 -2.54096407e-02 6.99858785e-01 1.96199358e-01
9.77703691e-01 -1.18792725e+00 -8.01023006e-01 1.54716030e-01
-9.52734426e-02 -4.86900024e-02 7.68766701e-01 6.50311470e-01
-6.64427400e-01 -8.29912499e-02 -5.79125643e-01 -4.60412562e-01
-1.10668683e+00 3.84075135e-01 6.15427971e-01 -3.56759399e-01
-7.44209051e-01 6.63185000e-01 5.58512323e-02 1.14477903e-01
1.55085504e-01 -8.84486660e-02 2.13609599e-02 -2.36132532e-01
6.97659671e-01 4.29019809e-01 5.54828458e-02 -5.53484797e-01
1.07580442e-02 6.79806054e-01 1.86468139e-02 -1.03863038e-01
1.33370054e+00 -5.37512422e-01 -7.68507496e-02 4.06335950e-01
1.41362774e+00 -2.89068550e-01 -1.86070609e+00 -1.80000037e-01
-6.31167829e-01 -7.53299475e-01 6.12730861e-01 -9.26174164e-01
-1.45575869e+00 8.64470065e-01 1.23974895e+00 -1.63499743e-01
1.83185053e+00 -4.06862050e-01 1.04940701e+00 1.51746184e-01
4.04688686e-01 -1.20111084e+00 1.90211862e-01 2.32655585e-01
7.28438675e-01 -1.32169521e+00 1.71776786e-01 -3.34557772e-01
-4.33913022e-01 1.06294250e+00 5.14517188e-01 4.32287268e-02
4.96828347e-01 4.50687349e-01 -4.64156494e-02 1.74371034e-01
-7.27851927e-01 -5.02205156e-02 1.61289603e-01 5.81753314e-01
2.92245865e-01 -2.75845528e-01 -1.27455875e-01 -9.47846286e-03
1.71813056e-01 3.05889547e-01 5.48657954e-01 4.18313950e-01
-3.57132584e-01 -8.54908705e-01 -2.21081838e-01 4.60359424e-01
-5.59642315e-01 -2.83134609e-01 4.60405290e-01 4.35561925e-01
4.61103708e-01 1.02063155e+00 -3.43830064e-02 -4.63253498e-01
4.51286107e-01 -4.77474093e-01 6.96161091e-01 -1.57782137e-01
-4.28450018e-01 1.43695056e-01 -1.02747455e-01 -7.15614319e-01
-5.30768514e-01 -3.38414967e-01 -8.55939865e-01 -4.35314208e-01
-5.65717183e-02 -6.09776318e-01 6.21903181e-01 5.85673988e-01
1.63487747e-01 8.30617368e-01 7.32961953e-01 -8.02686572e-01
-4.49291915e-01 -7.06802726e-01 -9.17333305e-01 3.78401220e-01
5.42702138e-01 -3.68326068e-01 -3.89772564e-01 1.54755861e-01] | [10.844559669494629, -2.2987563610076904] |
f69ee596-28f7-4d3e-ad1a-64912c29f99a | 3d-shape-retrieval-basing-on-representatives | 1810.09008 | null | http://arxiv.org/abs/1810.09008v3 | http://arxiv.org/pdf/1810.09008v3.pdf | 3D shape retrieval basing on representatives of classes | In this paper, we present an improvement of our proposed technique for 3D
shape retrieval in classified databases [2] which is based on representatives
of classes. Instead of systematically matching the object-query with all 3D
models of the database, our idea presented in [2] consist, for a classified
database, to represent each class by one representative that is used to orient
the retrieval process to the right class (the class excepted to contain 3D
models similar to the query). In order to increase the chance to fall in the
right class, our idea in this work is to represent each class by more than one
representative. In this case, instead of using only one representative to
decide which is the right class we use a set of representatives this will
contribute certainly to improving the relevance of retrieval results. The
obtained experimental results show that the relevance is significantly
improved. | ['M. Benjelloun', 'E. W. Dadi', 'E. M. Daoudi'] | 2018-10-21 | null | null | null | null | ['3d-shape-retrieval'] | ['computer-vision'] | [ 2.62454629e-01 1.50330648e-01 -4.58661765e-02 -2.22436726e-01
-4.87713099e-01 -4.10257697e-01 5.50828159e-01 5.78701675e-01
-3.19679916e-01 6.24726951e-01 -1.43933177e-01 -6.44086674e-02
-3.31724405e-01 -1.15270174e+00 -2.67813474e-01 -5.97663403e-01
3.35235149e-02 1.10768116e+00 8.33808005e-01 -2.89298266e-01
5.97293258e-01 1.18723238e+00 -2.01034594e+00 3.06759238e-01
4.52966958e-01 8.69073749e-01 1.48386598e-01 3.01526129e-01
-4.73628789e-01 2.69948304e-01 -7.08674908e-01 -2.09583417e-01
4.06895220e-01 -6.18946373e-01 -1.01585853e+00 4.67418253e-01
2.25763768e-01 -9.81807262e-02 4.15616095e-01 9.74335134e-01
4.03264076e-01 2.25463256e-01 1.22195876e+00 -7.68986881e-01
2.46621177e-01 6.42630532e-02 -2.55143255e-01 -9.59139988e-02
5.98773897e-01 -6.03325486e-01 9.02281821e-01 -6.66244447e-01
1.10718274e+00 1.04790735e+00 1.51008144e-01 3.66270393e-01
-1.14807165e+00 -3.08796823e-01 2.69511305e-02 1.07813336e-01
-1.57930338e+00 -3.57901901e-01 8.76807690e-01 -2.69799352e-01
4.67334539e-01 4.45341885e-01 7.68938780e-01 -3.06167435e-02
-2.40717441e-01 5.44949770e-01 1.08717382e+00 -1.00648844e+00
3.79816711e-01 4.50165391e-01 4.97079641e-01 3.73051316e-01
5.23434043e-01 -2.86445320e-01 2.26014592e-02 -4.39603806e-01
3.41276973e-01 -1.44370452e-01 -2.76273102e-01 -9.72098053e-01
-4.34487998e-01 5.45884013e-01 2.95535237e-01 9.39404130e-01
-5.57479441e-01 -1.83685645e-01 2.80874491e-01 8.88133869e-02
4.06766921e-01 4.56503719e-01 -1.32525548e-01 2.56629676e-01
-8.89506817e-01 4.91522402e-01 1.02758253e+00 7.85700142e-01
9.27757382e-01 -4.75936115e-01 1.82263851e-01 8.44934165e-01
1.23743437e-01 3.45401227e-01 2.79526860e-01 -3.24166775e-01
3.07644248e-01 1.07450294e+00 2.60495484e-01 -1.08311284e+00
-2.68499851e-01 -4.36772496e-01 -2.65627712e-01 4.78550106e-01
4.38994914e-01 5.52652657e-01 -8.39895725e-01 1.31697714e+00
5.73692620e-01 -5.05219996e-01 2.40486622e-01 8.53558004e-01
5.29205382e-01 4.53982294e-01 -1.38121203e-01 -3.47512811e-01
1.19118667e+00 -3.47898006e-01 -3.39200974e-01 3.73434931e-01
7.04918802e-01 -1.14109790e+00 5.59535503e-01 5.88428855e-01
-8.46359491e-01 -7.22767234e-01 -1.12850380e+00 3.97349864e-01
-6.04970515e-01 2.54705071e-01 -9.69682783e-02 5.66738546e-01
-7.71429241e-01 6.29633546e-01 -3.51282805e-01 -6.78574562e-01
-1.36647820e-01 4.44684267e-01 -4.36225295e-01 -1.37875855e-01
-8.95853817e-01 1.02864850e+00 6.30814373e-01 -1.36593297e-01
-3.58131170e-01 6.46859109e-02 -4.14672464e-01 -2.00101337e-03
5.09935975e-01 -4.12433565e-01 8.20560455e-01 -9.52139020e-01
-8.61849606e-01 1.18157589e+00 -3.58048171e-01 -2.53870189e-01
5.11175990e-01 2.54054427e-01 -6.08230270e-02 4.98790205e-01
-1.14364319e-01 4.98134255e-01 7.81466961e-01 -1.82340503e+00
-6.08090401e-01 -7.22536027e-01 1.96606535e-02 1.71610311e-01
-2.98582077e-01 -3.48608345e-02 -7.49187529e-01 -2.78256953e-01
7.66594350e-01 -9.96386349e-01 -2.81866137e-02 -2.12802976e-01
-6.63333982e-02 -4.30196911e-01 9.76962090e-01 -4.43816453e-01
1.05111086e+00 -1.95045936e+00 4.73204032e-02 7.36846745e-01
9.56763476e-02 4.65217412e-01 1.36768281e-01 6.84196413e-01
-5.60273975e-02 6.51185140e-02 -1.17294401e-01 -1.97321713e-01
-2.05054387e-01 2.71100760e-01 -2.18562797e-01 2.92403251e-01
-9.14735906e-03 4.13234532e-02 -5.70952952e-01 -7.92694628e-01
2.67202049e-01 2.31397241e-01 -3.43784869e-01 4.25537974e-02
-2.55171329e-01 4.73065749e-02 -1.01010489e+00 2.53120154e-01
1.04599416e+00 2.62231737e-01 1.98326439e-01 -2.74347097e-01
-1.71213791e-01 2.18384206e-01 -1.73971784e+00 9.47880685e-01
-3.12950224e-01 -1.52048498e-01 -1.11432001e-01 -1.16743779e+00
1.59676623e+00 3.78296852e-01 5.36429226e-01 -5.21606863e-01
-6.25214027e-03 6.20670974e-01 -2.17707828e-01 -3.26304972e-01
5.31647623e-01 5.15296981e-02 1.92825273e-01 4.74196941e-01
-2.83695191e-01 -3.81905079e-01 4.48365301e-01 -7.00635016e-02
4.81455773e-01 2.79037412e-02 3.63285929e-01 -2.11496502e-01
9.94268179e-01 1.20039262e-01 1.78244635e-01 7.01543391e-01
2.19425142e-01 5.18863678e-01 4.89557922e-01 -4.11485851e-01
-1.06776416e+00 -6.42545700e-01 -1.24175675e-01 5.65229595e-01
2.37034217e-01 -3.39773327e-01 -5.56412280e-01 -7.18931079e-01
-1.15538423e-03 4.15943563e-01 -4.81507599e-01 1.33928899e-02
-6.77551448e-01 -4.44609642e-01 2.94604599e-02 -2.77873855e-02
5.18697321e-01 -9.20112669e-01 -9.18138444e-01 1.53656214e-01
-5.24516180e-02 -3.08403403e-01 3.25864367e-02 3.24146569e-01
-1.19884264e+00 -1.29793072e+00 -9.73349810e-01 -6.02684140e-01
8.43966126e-01 4.28782701e-01 1.02442789e+00 4.95889544e-01
-8.31493288e-02 3.95664543e-01 -8.53520691e-01 -3.18994582e-01
-7.95857430e-01 2.35741928e-01 -3.14127117e-01 -5.88891059e-02
1.96662188e-01 -2.32535034e-01 -2.48595640e-01 4.75685209e-01
-1.11319745e+00 -5.08031905e-01 4.56186771e-01 4.42328453e-01
7.40407884e-01 2.29371384e-01 3.30214292e-01 -9.81563866e-01
3.54302704e-01 -3.95992398e-02 -9.18792367e-01 3.82190853e-01
-7.11007476e-01 2.15434521e-01 6.29657507e-01 -1.66828036e-01
-6.50558829e-01 2.05000833e-01 -1.76184535e-01 -1.45085752e-01
-4.85558689e-01 2.91791469e-01 -3.46908927e-01 -1.83508605e-01
4.70329553e-01 3.06049466e-01 3.98744419e-02 -7.40580380e-01
-2.69939210e-02 8.01925242e-01 -5.36902957e-02 -4.69463766e-01
6.65890455e-01 2.08736435e-01 5.40012062e-01 -8.28485131e-01
-1.26258090e-01 -9.07550573e-01 -7.18760967e-01 -3.39900821e-01
5.89317799e-01 -2.70929992e-01 -4.07615453e-01 8.40602741e-02
-1.20849395e+00 5.66006660e-01 -2.26957917e-01 3.58647108e-01
-5.38454354e-01 5.10457695e-01 1.26395658e-01 -1.17634046e+00
-2.08872855e-01 -1.19819355e+00 1.00804341e+00 6.63711056e-02
-1.02713011e-01 -4.53779727e-01 9.43718180e-02 1.12154074e-01
1.08130120e-01 -1.67161936e-03 1.20157230e+00 -1.37877929e+00
-6.38704181e-01 -8.90715837e-01 2.85685491e-02 3.11919957e-01
2.09816888e-01 -3.42339501e-02 -6.99628174e-01 -1.98589265e-01
1.09495804e-01 8.75448436e-02 7.40121722e-01 -4.85513508e-02
9.19584095e-01 1.88859954e-01 -7.19043255e-01 -1.19959913e-01
1.49844694e+00 7.57612109e-01 7.22579479e-01 3.50253373e-01
1.33075854e-02 7.47373581e-01 9.70435679e-01 1.92588061e-01
-1.69037327e-01 1.12505579e+00 4.30245578e-01 -5.50967641e-02
-1.45907164e-01 1.35442436e-01 -3.02746266e-01 4.01483685e-01
-1.71531528e-01 -3.88739794e-01 -8.24358284e-01 4.73263383e-01
-1.50274968e+00 -8.34632397e-01 -2.08794907e-01 2.72212338e+00
3.60504597e-01 2.92152852e-01 3.82908970e-01 5.88299930e-01
8.59480858e-01 -6.67304965e-04 1.02784216e-01 -4.90303427e-01
-2.85831839e-02 1.61378145e-01 3.74706611e-02 4.77256566e-01
-8.64260852e-01 4.61282164e-01 5.51340199e+00 1.03479075e+00
-1.00400019e+00 -3.54237676e-01 2.82384127e-01 4.61511046e-01
-3.21902245e-01 2.84000814e-01 -1.09005976e+00 1.82079658e-01
2.64988810e-01 -2.09687296e-02 1.96183100e-02 6.92006052e-01
-2.68588588e-02 -7.07869291e-01 -1.04062557e+00 4.63120461e-01
2.10656941e-01 -8.98513973e-01 5.27535498e-01 3.50613832e-01
4.69646990e-01 -7.75064409e-01 -3.76465231e-01 -1.72517061e-01
-4.52197105e-01 -3.86784941e-01 5.71867883e-01 6.70604050e-01
2.33974174e-01 -9.82422113e-01 9.74091470e-01 5.87555707e-01
-1.28016317e+00 3.02124709e-01 -4.95933831e-01 3.74053776e-01
-1.07369579e-01 4.65796500e-01 -1.26890504e+00 1.00126767e+00
4.04089093e-01 7.13549778e-02 -3.72080594e-01 1.41232312e+00
2.92665303e-01 2.05505416e-01 -2.79315978e-01 -3.74424011e-01
1.40422508e-02 -2.47767732e-01 7.11241901e-01 1.05079710e+00
3.58275682e-01 2.76381999e-01 2.83289522e-01 4.72274899e-01
2.58830994e-01 7.39083886e-01 -9.46667731e-01 4.11111176e-01
4.24566388e-01 8.27604711e-01 -9.20052171e-01 -6.40526474e-01
-1.22459151e-01 7.13040829e-01 4.45559919e-02 8.41651559e-02
-2.03648642e-01 -7.23632097e-01 -8.16160738e-02 4.43118721e-01
6.47889078e-01 -5.02549373e-02 1.56079561e-01 -3.49562079e-01
9.23950747e-02 -5.73236942e-01 5.77671587e-01 -6.82305932e-01
-8.79817426e-01 8.76949131e-01 4.88097042e-01 -1.62436283e+00
-5.59395790e-01 -5.09131730e-01 -2.05250397e-01 1.29949462e+00
-1.21292031e+00 -9.10223901e-01 -6.95887953e-02 4.75183100e-01
1.57286510e-01 -5.95699959e-02 8.85116398e-01 2.38952339e-01
2.02184334e-01 1.11519590e-01 1.37303859e-01 -1.29938468e-01
5.75312734e-01 -9.64187682e-01 -3.60945970e-01 3.92345399e-01
1.88910484e-01 5.36678493e-01 9.26803827e-01 -7.45386720e-01
-8.51461291e-01 -4.97165829e-01 1.35935688e+00 1.75216243e-01
2.22204700e-01 1.99420266e-02 -9.32006598e-01 -5.60531057e-02
-1.74305663e-01 -3.20353925e-01 2.72189319e-01 -1.00417741e-01
-1.13431759e-01 -5.41106403e-01 -1.30868733e+00 2.79355347e-01
6.16726577e-01 -4.18799579e-01 -8.79329026e-01 1.53213039e-01
1.11988395e-01 -5.78835197e-02 -6.82955801e-01 5.68498075e-01
3.99887949e-01 -9.79511559e-01 8.80393863e-01 -2.94825107e-01
-1.24780603e-01 -5.99445999e-01 -7.06319883e-02 -1.14257634e+00
1.59217238e-01 1.16837464e-01 4.19325203e-01 1.03017366e+00
1.69885278e-01 -6.43730998e-01 9.32911813e-01 2.37015098e-01
-3.80938803e-03 -5.59575379e-01 -1.06145680e+00 -8.02414715e-01
-1.90944567e-01 6.61971793e-02 7.25762784e-01 4.34696048e-01
-3.92411232e-01 -4.60736454e-02 1.03381919e-02 -4.02375963e-03
4.01421815e-01 9.18899536e-01 8.56777787e-01 -1.74336493e+00
-2.44561762e-01 -3.18953127e-01 -6.42503798e-01 -9.96460140e-01
-1.63172990e-01 -7.13445127e-01 4.66485173e-02 -1.45442283e+00
7.45479017e-02 -7.85647511e-01 -2.31039301e-01 2.03048483e-01
2.62693405e-01 3.14266562e-01 4.68662798e-01 3.45395029e-01
-3.49459201e-01 8.78347233e-02 1.25783706e+00 -2.25555450e-02
-2.75803715e-01 5.32761335e-01 -1.66828617e-01 4.37916249e-01
5.19318342e-01 -5.82280874e-01 -2.84254223e-01 6.87822476e-02
-1.23892777e-01 2.60994256e-01 1.80558085e-01 -9.54475999e-01
8.94070789e-02 5.14553934e-02 4.44732487e-01 -1.10842299e+00
5.22456706e-01 -1.27243507e+00 4.47611034e-01 7.62581170e-01
-1.90381512e-01 -6.12168349e-02 1.30550101e-01 2.43564978e-01
-2.44629338e-01 -1.12605691e+00 6.11797512e-01 -3.43960404e-01
-5.58413982e-01 -1.93492055e-01 -4.48947370e-01 -5.44839144e-01
8.18835080e-01 -6.26869261e-01 -9.04863793e-03 -4.18321699e-01
-1.16528332e+00 -1.81997895e-01 6.28082335e-01 -1.75390899e-01
4.40011501e-01 -1.14572012e+00 -3.49873364e-01 -8.72047991e-02
4.92083430e-01 -2.61400551e-01 -1.06411740e-01 4.82308656e-01
-5.23551166e-01 7.74724483e-01 -1.75319001e-01 -5.55880845e-01
-1.57109654e+00 7.18026996e-01 3.26637328e-01 -3.94495577e-01
-3.09264362e-01 3.29791158e-01 -1.38109624e-01 -6.73013851e-02
1.19526304e-01 -5.60167544e-02 -7.03651667e-01 3.38036925e-01
1.67839378e-01 2.71354884e-01 6.92724228e-01 -7.16960669e-01
-3.66481632e-01 1.02549732e+00 3.94104831e-02 -2.12176532e-01
1.13458109e+00 1.03874095e-01 -3.55354846e-01 3.83587480e-01
1.27589655e+00 5.38032651e-01 -4.28560734e-01 -1.35747477e-01
3.30315232e-01 -5.82509935e-01 -1.33427978e-01 -5.90593457e-01
-7.94265985e-01 5.11141062e-01 7.23787308e-01 5.80616415e-01
1.45237613e+00 2.21164137e-01 1.60451636e-01 6.63750529e-01
8.16654682e-01 -8.65732789e-01 -1.92856967e-01 3.08097303e-01
8.88433933e-01 -6.93625271e-01 2.42041826e-01 -6.19449437e-01
-3.35510671e-01 1.39638245e+00 2.84484208e-01 -2.52750307e-01
6.62940919e-01 -1.51384115e-01 -8.73021185e-02 -3.25798899e-01
-2.85234153e-01 -5.89916706e-01 5.23010015e-01 4.15920794e-01
2.94865727e-01 -2.51514286e-01 -1.13805127e+00 1.48986131e-02
3.08508098e-01 3.95076722e-02 3.17237526e-01 1.11117625e+00
-8.60824347e-01 -1.77041125e+00 -7.99648583e-01 3.35188240e-01
-1.63215455e-02 3.19939196e-01 -6.35684967e-01 1.10667717e+00
2.93889463e-01 7.98349023e-01 1.17105775e-01 -9.06004310e-02
7.32374609e-01 2.59404033e-01 6.04667246e-01 -5.34368575e-01
-5.45327544e-01 4.48904902e-01 1.91235378e-01 -1.35692865e-01
-6.13302588e-01 -6.39219165e-01 -1.16119981e+00 1.62340611e-01
-5.91103256e-01 6.47187471e-01 7.87677169e-01 6.94923878e-01
-5.36220670e-02 -8.29114541e-02 8.58971298e-01 -6.68695629e-01
-5.77392578e-01 -7.15221882e-01 -5.92317939e-01 4.72504705e-01
8.41398835e-02 -1.04265463e+00 -2.71163046e-01 -2.33999684e-01] | [8.974051475524902, 2.1923165321350098] |
fb9f46db-f944-4d12-9de9-5cdaec200d6a | mpool-motif-based-graph-pooling | 2303.03654 | null | https://arxiv.org/abs/2303.03654v1 | https://arxiv.org/pdf/2303.03654v1.pdf | MPool: Motif-Based Graph Pooling | Graph Neural networks (GNNs) have recently become a powerful technique for many graph-related tasks including graph classification. Current GNN models apply different graph pooling methods that reduce the number of nodes and edges to learn the higher-order structure of the graph in a hierarchical way. All these methods primarily rely on the one-hop neighborhood. However, they do not consider the higher- order structure of the graph. In this work, we propose a multi-channel Motif-based Graph Pooling method named (MPool) captures the higher-order graph structure with motif and local and global graph structure with a combination of selection and clustering-based pooling operations. As the first channel, we develop node selection-based graph pooling by designing a node ranking model considering the motif adjacency of nodes. As the second channel, we develop cluster-based graph pooling by designing a spectral clustering model using motif adjacency. As the final layer, the result of each channel is aggregated into the final graph representation. We perform extensive experiments on eight benchmark datasets and show that our proposed method shows better accuracy than the baseline methods for graph classification tasks. | ['Esra Akbas', 'Max Khanov', 'Muhammad Ifte Khairul Islam'] | 2023-03-07 | null | null | null | null | ['graph-classification'] | ['graphs'] | [-1.02157585e-01 -5.34436032e-02 -2.99585521e-01 -2.73153603e-01
-9.52865109e-02 -5.20069003e-01 4.87971246e-01 5.39726436e-01
-2.51392215e-01 2.50278443e-01 8.42284113e-02 -3.01003695e-01
-8.74943361e-02 -1.37820327e+00 -7.84296334e-01 -7.54164994e-01
-4.00833756e-01 -2.87890639e-02 5.45765340e-01 -4.96315360e-02
1.60601720e-01 4.57162052e-01 -1.01110268e+00 4.59817916e-01
6.50785565e-01 7.91097760e-01 3.49499166e-01 4.11411971e-01
-3.07439506e-01 6.25467300e-01 -2.99283713e-01 -5.76538630e-02
1.37147695e-01 -3.51893157e-01 -4.57735837e-01 1.08152308e-01
4.33492929e-01 -3.26378085e-02 -6.86819494e-01 1.33695412e+00
4.32117969e-01 2.34193698e-01 4.28680867e-01 -1.08424747e+00
-7.64796913e-01 1.00228274e+00 -6.00983083e-01 1.87796563e-01
1.69068530e-01 -1.17474329e-02 1.64791703e+00 -6.91281855e-01
6.92732215e-01 1.34028029e+00 6.30399108e-01 1.07900843e-01
-1.15927792e+00 -5.08790076e-01 5.73845088e-01 3.07151079e-01
-1.54411507e+00 3.23692232e-01 1.24375820e+00 -3.07142049e-01
1.16473794e+00 8.69943723e-02 8.54679286e-01 5.30071080e-01
3.15229982e-01 5.58788121e-01 6.11534297e-01 -1.12816915e-01
4.32572439e-02 -4.82913822e-01 5.90310752e-01 1.29844463e+00
2.75708705e-01 -4.83490527e-01 -3.42775643e-01 -1.09245934e-01
8.14130247e-01 3.39220554e-01 -2.28059024e-01 -2.53663868e-01
-1.07376325e+00 9.25989509e-01 1.47082937e+00 4.99315530e-01
-4.22378331e-01 4.59680051e-01 3.87747616e-01 2.51405180e-01
3.85836929e-01 3.00392240e-01 1.10698342e-01 7.59420812e-01
-6.65492833e-01 3.13385539e-02 6.92229569e-01 9.17419910e-01
1.24484181e+00 -2.82897979e-01 -5.06187260e-01 8.55342090e-01
5.07759929e-01 -1.83926687e-01 4.16529477e-01 -1.06864870e-01
6.52618527e-01 1.40603745e+00 -6.20454133e-01 -1.62910521e+00
-7.19145775e-01 -6.63355708e-01 -1.08459520e+00 -2.66371906e-01
3.32825147e-02 1.72697276e-01 -1.14740157e+00 1.70650649e+00
1.21316776e-01 4.73799646e-01 -3.80584300e-01 8.39418709e-01
1.26385498e+00 7.72674441e-01 3.28567326e-02 1.38991684e-01
1.33819687e+00 -1.23535025e+00 -4.25909191e-01 -1.09386705e-01
7.33433723e-01 -2.11658269e-01 1.00195527e+00 1.30957533e-02
-7.73260951e-01 -4.84159142e-01 -1.20058501e+00 -1.95671469e-01
-6.46285653e-01 -1.14536351e-02 8.58470619e-01 3.16556364e-01
-1.36253905e+00 8.20946574e-01 -7.63816416e-01 -5.02506614e-01
4.24612105e-01 3.78811836e-01 -5.17843723e-01 -1.84313580e-01
-1.14449394e+00 2.20242545e-01 5.74450493e-01 4.51827794e-01
-4.29321378e-01 -2.49876052e-01 -1.13894272e+00 5.03551841e-01
2.80886948e-01 -4.62881178e-01 3.50507587e-01 -5.29122233e-01
-1.04801273e+00 6.68050349e-01 -1.74292430e-01 -2.95554787e-01
-1.10140182e-01 4.05941308e-01 -1.42048329e-01 2.92340934e-01
-1.06854044e-01 5.59459984e-01 4.50965017e-01 -9.35196757e-01
-1.81664720e-01 -4.09159571e-01 1.57119989e-01 2.82684982e-01
-3.83298635e-01 -3.26403141e-01 -7.49632180e-01 -6.45058095e-01
6.28954470e-01 -7.33502388e-01 -4.22335118e-01 -2.80309767e-01
-7.32314408e-01 -5.52687347e-01 6.73056126e-01 -4.27710116e-01
1.67703736e+00 -2.17423368e+00 4.08452332e-01 6.01266861e-01
7.77460635e-01 3.08621349e-03 -2.78559625e-01 7.82118499e-01
-8.42953101e-02 3.96700323e-01 -1.36283591e-01 -2.87332296e-01
7.08114505e-02 5.74234538e-02 -5.05845323e-02 3.50927353e-01
3.30472529e-01 1.27049339e+00 -8.40456843e-01 -4.71817762e-01
1.51970610e-01 3.58485907e-01 -7.73411632e-01 2.43701250e-03
-2.44913414e-01 3.00731715e-02 -5.86004257e-01 3.83947402e-01
7.69944727e-01 -7.73672163e-01 5.31110823e-01 -4.45598871e-01
9.70450118e-02 2.59575427e-01 -1.00028110e+00 1.72518146e+00
-1.36804298e-01 3.64425600e-01 -1.06472664e-01 -1.10680330e+00
9.64864075e-01 -1.79357901e-01 2.10369140e-01 -3.51152569e-01
7.54795820e-02 2.04484053e-02 3.00340652e-01 -1.98544115e-01
2.46352598e-01 2.07782224e-01 -1.94510192e-01 3.47670019e-01
3.18900675e-01 3.61217707e-01 3.31070572e-01 6.40507400e-01
1.59555519e+00 -2.71687746e-01 1.29799530e-01 -4.02258366e-01
5.18400311e-01 -4.48565215e-01 3.85761201e-01 6.60978198e-01
3.35213658e-03 7.36574173e-01 1.04354727e+00 -6.65888667e-01
-6.04737818e-01 -9.08107221e-01 3.23404044e-01 9.86614347e-01
2.38523290e-01 -7.99293637e-01 -7.51787186e-01 -6.39881432e-01
2.46781332e-04 -1.08836092e-01 -7.19896615e-01 -3.28817368e-01
-6.51286006e-01 -1.08533347e+00 2.65486658e-01 3.29380661e-01
6.56473219e-01 -1.27211738e+00 5.70345931e-02 3.04737061e-01
-5.01509309e-02 -1.12652826e+00 -8.28919530e-01 1.50582522e-01
-8.13128352e-01 -1.01408732e+00 -4.32100922e-01 -1.25448334e+00
1.03975511e+00 3.01828444e-01 8.82833064e-01 7.88398445e-01
-1.19969271e-01 -6.13978691e-02 -5.04129589e-01 2.79124022e-01
3.26215178e-01 5.06188869e-01 -5.37552238e-01 3.81886691e-01
2.52582312e-01 -8.28035474e-01 -5.61657071e-01 -9.11833066e-03
-7.43771613e-01 1.20728590e-01 6.97014809e-01 7.59359360e-01
6.74213946e-01 2.30542526e-01 3.44993055e-01 -1.13404346e+00
9.10116851e-01 -4.62388098e-01 -5.64213395e-01 3.33630890e-01
-2.22446322e-01 2.92912334e-01 8.42335045e-01 -2.01275840e-01
-2.80421764e-01 4.31497879e-02 4.07944694e-02 -3.76064748e-01
2.54838198e-01 1.04647136e+00 -5.21178126e-01 -2.19732136e-01
2.65059084e-01 2.36793026e-01 -1.96731478e-01 -5.15959620e-01
2.89206654e-01 2.17207104e-01 1.27620727e-01 -3.35262835e-01
6.43714249e-01 3.47019881e-01 3.06208879e-01 -7.54849911e-01
-3.78609717e-01 -3.92352790e-01 -6.13169134e-01 -1.98213235e-01
1.05690503e+00 -6.98878586e-01 -7.88888395e-01 6.35106087e-01
-1.13159597e+00 -4.11433071e-01 4.48818952e-01 2.74547040e-01
6.22918792e-02 6.20519698e-01 -8.90289009e-01 -4.52986181e-01
-3.33192527e-01 -1.21541762e+00 1.01210332e+00 1.44756526e-01
2.69160658e-01 -1.01345181e+00 -1.78962037e-01 -2.08786532e-01
1.49965063e-01 4.80962753e-01 1.30783141e+00 -6.15017891e-01
-1.02190673e+00 -2.43383855e-01 -6.93663955e-01 4.47241962e-02
9.00377408e-02 -1.72449544e-01 -4.52629149e-01 -3.80958557e-01
-6.13192379e-01 8.88254121e-02 1.44342959e+00 4.23001558e-01
1.34723067e+00 -2.55666882e-01 -4.83708054e-01 7.72008836e-01
1.69742775e+00 -2.24883333e-01 4.49512661e-01 -1.05847992e-01
1.45486724e+00 4.90073144e-01 -1.74966976e-01 1.77732371e-02
5.64548492e-01 4.61680621e-01 5.42569220e-01 -1.19413927e-01
-1.52222544e-01 -6.05917513e-01 2.20893949e-01 1.15454745e+00
-5.98307066e-02 -5.37403047e-01 -7.71620631e-01 3.72435510e-01
-2.12989473e+00 -7.56563783e-01 -4.78448957e-01 1.92317104e+00
2.71551937e-01 2.84468472e-01 9.37233716e-02 -1.74038053e-01
1.10352600e+00 6.88199878e-01 -3.93794060e-01 -2.55443752e-01
-6.78757653e-02 1.60767451e-01 5.28365612e-01 5.88869512e-01
-1.08236229e+00 1.16522992e+00 5.29574490e+00 6.89174950e-01
-1.00782788e+00 -1.40907645e-01 4.39118773e-01 3.24620754e-01
-4.38318819e-01 -1.54015319e-02 -4.75770772e-01 5.57694316e-01
4.29609925e-01 7.45513961e-02 6.20857358e-01 5.52738130e-01
-1.45276189e-01 1.75610781e-01 -9.21933830e-01 9.28460240e-01
-6.84875771e-02 -1.43415129e+00 5.20395398e-01 2.20898315e-01
4.92681801e-01 2.14467183e-01 -3.28507274e-01 2.05252782e-01
2.41455570e-01 -1.14056146e+00 3.87728751e-01 4.95865196e-01
3.99284303e-01 -7.32841909e-01 7.86199450e-01 1.43769950e-01
-1.99130726e+00 -6.55693486e-02 -5.07266164e-01 -9.89611968e-02
2.73013245e-02 6.44310176e-01 -3.80364329e-01 9.24498737e-01
6.22952342e-01 1.10045838e+00 -6.97350621e-01 9.89348829e-01
-3.84445935e-01 4.76217330e-01 -3.11464131e-01 -3.79395932e-01
5.99880815e-01 -4.55475420e-01 4.02276099e-01 1.11649930e+00
9.73744765e-02 5.45625202e-02 5.62283099e-01 9.42370296e-01
-5.45237899e-01 2.75373429e-01 -6.20149612e-01 -3.99581343e-01
3.28481823e-01 1.50575089e+00 -1.27141643e+00 -1.03408746e-01
-4.76135015e-01 9.30438340e-01 9.29276645e-01 5.43418169e-01
-5.57499766e-01 -8.86190534e-01 3.17100674e-01 7.48924762e-02
4.23556358e-01 -4.62679088e-01 1.04963839e-01 -1.12037182e+00
1.64298251e-01 -4.32984918e-01 5.17038524e-01 -4.79515821e-01
-1.42855561e+00 7.28124022e-01 -1.70646027e-01 -7.37987697e-01
4.08321828e-01 -7.91429996e-01 -1.02979207e+00 9.12152946e-01
-1.30813146e+00 -1.38077712e+00 -4.55234319e-01 5.50237715e-01
-4.00220491e-02 -4.52631898e-02 5.55486381e-01 3.76701802e-01
-7.38431811e-01 5.34102023e-01 -4.34462219e-01 5.50862372e-01
2.23293304e-01 -1.31028104e+00 5.54665983e-01 8.44513893e-01
2.60891825e-01 8.86955798e-01 1.49627373e-01 -8.16722989e-01
-1.45123422e+00 -1.32381928e+00 7.81146705e-01 5.85854128e-02
6.67650044e-01 -1.00591934e+00 -1.07138026e+00 6.64119601e-01
1.35990409e-02 3.32513481e-01 4.64727610e-01 2.72885531e-01
-4.44746822e-01 -1.99430704e-01 -8.94930959e-01 5.97313583e-01
1.40474689e+00 -6.32940352e-01 -7.59658441e-02 2.67380297e-01
1.10167396e+00 -1.84459761e-01 -8.14747393e-01 2.84405679e-01
4.12652969e-01 -7.27865100e-01 6.27811730e-01 -4.31652516e-01
2.37131551e-01 -5.81817746e-01 7.88054313e-04 -1.20133197e+00
-7.51374066e-01 -5.38348377e-01 2.30958372e-01 9.59132195e-01
5.68444550e-01 -9.15022075e-01 7.85434783e-01 -1.37093395e-01
-2.02258855e-01 -9.74461734e-01 -6.97006345e-01 -4.61281151e-01
-2.77743071e-01 -1.28980368e-01 7.52510190e-01 1.03113973e+00
1.07305720e-01 6.59552932e-01 -7.16197491e-02 4.33814675e-01
5.19577265e-01 3.02743167e-01 5.98082602e-01 -1.12466443e+00
-2.43703872e-01 -7.88389087e-01 -9.96626079e-01 -1.09780347e+00
2.56907940e-01 -1.50424600e+00 -1.35546580e-01 -2.07583928e+00
4.18450356e-01 -1.28956988e-01 -7.19651520e-01 5.89369893e-01
-3.21164876e-01 2.37373129e-01 8.07289034e-02 -7.12574413e-03
-7.77409554e-01 4.29930925e-01 1.18781602e+00 -2.63413429e-01
-2.59331256e-01 -4.26835686e-01 -6.09619021e-01 4.31485593e-01
7.19748080e-01 -2.54864961e-01 -4.70736206e-01 -5.69090486e-01
4.77244049e-01 -3.39834064e-01 5.45330405e-01 -9.26550806e-01
4.77237701e-01 2.31726259e-01 1.00847147e-01 -5.18587232e-01
1.44809913e-02 -5.52066505e-01 2.19349265e-02 4.96281534e-01
-1.69621587e-01 8.62699226e-02 8.38700682e-02 8.00869167e-01
-3.60003799e-01 1.37685075e-01 6.50874972e-01 -2.29196072e-01
-6.31656051e-01 8.22428644e-01 7.88039416e-02 -3.79874021e-01
6.98688567e-01 -7.39977509e-03 -4.39996302e-01 -3.20960790e-01
-8.64305973e-01 4.48709965e-01 3.63918990e-01 3.66371542e-01
6.85611427e-01 -1.53767037e+00 -5.02159178e-01 2.85534471e-01
1.24798335e-01 1.05602913e-01 3.32401484e-01 7.42550015e-01
-6.29508555e-01 1.37435839e-01 -1.33330867e-01 -4.66682971e-01
-1.18690479e+00 4.88586694e-01 3.81993741e-01 -6.59606636e-01
-6.17696762e-01 1.05642295e+00 5.23294926e-01 -5.18152595e-01
5.37211969e-02 -5.97817838e-01 -3.44253689e-01 -1.43532619e-01
1.35965616e-01 2.21055560e-02 -4.45272438e-02 -6.39613032e-01
-3.96318287e-01 5.84156334e-01 -1.53579608e-01 2.00719491e-01
1.29856873e+00 2.12857157e-01 -8.16564679e-01 3.93937171e-01
1.37340438e+00 -2.40249291e-01 -8.92660558e-01 -2.37797141e-01
8.81126300e-02 -1.34284168e-01 3.88723873e-02 -1.98206902e-01
-1.26295388e+00 9.88399863e-01 4.21930775e-02 6.16766453e-01
1.11645329e+00 1.97876409e-01 6.63078904e-01 3.51061344e-01
2.90011883e-01 -6.31583214e-01 8.24583247e-02 4.60449755e-01
9.23853397e-01 -9.49553788e-01 -2.88465451e-02 -7.22375035e-01
-5.82832769e-02 9.47381020e-01 6.26262546e-01 -6.59338534e-01
1.02242732e+00 -3.35544258e-01 -5.20226598e-01 -5.60064912e-01
-5.33263326e-01 -4.24077034e-01 6.64635003e-01 2.71396011e-01
4.05188888e-01 3.43217909e-01 -4.77603227e-01 7.42942810e-01
-9.45023727e-03 -4.64575440e-01 1.46910101e-01 5.81355631e-01
-5.61001956e-01 -1.26825392e+00 3.60620946e-01 7.16824114e-01
-2.79933244e-01 -3.93465251e-01 -7.27251172e-01 6.02617562e-01
-3.79479602e-02 7.77910590e-01 1.37602001e-01 -8.05113256e-01
2.03385487e-01 -1.98303893e-01 3.66049469e-01 -1.14918435e+00
-6.48161054e-01 7.27181211e-02 -1.26456499e-01 -6.86674654e-01
-1.27062827e-01 -2.91640699e-01 -1.49414289e+00 -1.90898284e-01
-3.35749418e-01 7.74082094e-02 3.39299053e-01 6.73782229e-01
6.51703835e-01 6.70081794e-01 5.18675923e-01 -9.26275671e-01
1.31526917e-01 -1.00256550e+00 -8.46238136e-01 3.57106209e-01
3.32979470e-01 -4.66654062e-01 -3.11926782e-01 -5.81592083e-01] | [7.106286525726318, 6.309908866882324] |
3e394370-427b-452b-8425-fc083657ab93 | contrastive-inverse-regression-for-dimension | 2305.12287 | null | https://arxiv.org/abs/2305.12287v1 | https://arxiv.org/pdf/2305.12287v1.pdf | Contrastive inverse regression for dimension reduction | Supervised dimension reduction (SDR) has been a topic of growing interest in data science, as it enables the reduction of high-dimensional covariates while preserving the functional relation with certain response variables of interest. However, existing SDR methods are not suitable for analyzing datasets collected from case-control studies. In this setting, the goal is to learn and exploit the low-dimensional structure unique to or enriched by the case group, also known as the foreground group. While some unsupervised techniques such as the contrastive latent variable model and its variants have been developed for this purpose, they fail to preserve the functional relationship between the dimension-reduced covariates and the response variable. In this paper, we propose a supervised dimension reduction method called contrastive inverse regression (CIR) specifically designed for the contrastive setting. CIR introduces an optimization problem defined on the Stiefel manifold with a non-standard loss function. We prove the convergence of CIR to a local optimum using a gradient descent-based algorithm, and our numerical study empirically demonstrates the improved performance over competing methods for high-dimensional data. | ['Didong Li', 'Hengrui Luo', 'Sam Hawke'] | 2023-05-20 | null | null | null | null | ['dimensionality-reduction'] | ['methodology'] | [ 3.18933815e-01 -7.58701861e-02 -4.11352217e-01 -3.00717324e-01
-7.18094885e-01 -2.87867814e-01 5.02047896e-01 1.78517565e-01
-3.06672692e-01 6.71783149e-01 4.51771408e-01 3.67015898e-02
-7.52562404e-01 -6.02095664e-01 -3.20915490e-01 -1.21801281e+00
-2.29386747e-01 5.44706464e-01 -5.44723272e-01 1.21511715e-02
-1.81782451e-02 4.48240131e-01 -1.25326002e+00 -3.24202567e-01
1.03106546e+00 3.67043525e-01 -1.19837604e-01 1.36784300e-01
3.66097003e-01 1.37658656e-01 -1.50457889e-01 -8.11666325e-02
4.52382237e-01 -5.19277930e-01 -5.40353119e-01 2.02216730e-01
3.21873337e-01 9.51938629e-02 -2.77304649e-01 9.33599949e-01
5.14012992e-01 2.93811798e-01 9.56834137e-01 -1.27824485e+00
-3.69263291e-01 1.82366207e-01 -7.51912534e-01 -2.58220900e-02
-1.28784791e-01 -4.59233463e-01 1.11816192e+00 -8.85036767e-01
7.63787866e-01 1.25747740e+00 5.16409278e-01 5.15732229e-01
-1.92380214e+00 -5.19689083e-01 -6.97297081e-02 -1.82406530e-01
-1.33056164e+00 -3.23900640e-01 1.12062001e+00 -8.15133631e-01
2.82967001e-01 4.20260191e-01 2.85423338e-01 9.00339782e-01
4.20982437e-03 5.05030274e-01 1.22045815e+00 -3.11340123e-01
3.26452374e-01 9.61312801e-02 3.81857425e-01 4.67280000e-01
3.96997094e-01 1.59990579e-01 -4.09532338e-01 -5.17241895e-01
6.46872759e-01 2.10516512e-01 -2.20037326e-01 -9.95156229e-01
-1.01890540e+00 1.24049377e+00 6.45969659e-02 2.04973534e-01
-5.14838636e-01 -2.77262181e-01 2.63805419e-01 2.25171372e-01
1.04363263e+00 3.47233593e-01 -2.46751383e-01 2.02812865e-01
-9.04702008e-01 2.47104347e-01 5.51025808e-01 4.84399229e-01
6.54865563e-01 -1.46793872e-01 -9.61794332e-02 8.34699988e-01
3.17240924e-01 3.28533381e-01 2.80907333e-01 -8.26856494e-01
5.63913226e-01 8.30207825e-01 3.69391367e-02 -1.30673277e+00
-6.37939751e-01 -6.25517249e-01 -1.43362796e+00 1.00452490e-01
3.73211533e-01 -8.19926932e-02 -2.95590669e-01 2.11462283e+00
6.80739224e-01 1.60263345e-01 1.00130506e-01 8.54010344e-01
1.68969929e-01 3.57136309e-01 9.17014387e-03 -7.02380359e-01
1.06594265e+00 -4.70381141e-01 -7.91055024e-01 -5.59139904e-03
7.78634906e-01 -3.89610171e-01 8.79108071e-01 3.54851544e-01
-8.42382371e-01 -2.31818795e-01 -8.21563601e-01 -1.60946056e-01
-1.64856184e-02 -1.96513850e-02 6.44424856e-01 6.17392361e-01
-6.28186524e-01 5.19063711e-01 -7.68769085e-01 -1.26294762e-01
4.94846225e-01 4.39662874e-01 -6.76907480e-01 -8.43592137e-02
-1.09417129e+00 4.84906822e-01 -1.82771519e-01 2.67065130e-02
-6.52011216e-01 -9.08420682e-01 -8.85020673e-01 2.98782066e-02
4.68921870e-01 -6.70658052e-01 3.70635480e-01 -6.65065646e-01
-1.17232716e+00 9.26491022e-01 -2.29240164e-01 -4.62165952e-01
6.70561194e-01 -2.27907062e-01 -1.28035173e-01 2.26193443e-02
1.91764712e-01 3.25000938e-03 1.03559411e+00 -1.05439639e+00
-2.70044148e-01 -9.21686292e-01 -1.80684969e-01 -2.00425982e-02
-6.19713962e-01 -1.24433935e-01 -1.13592483e-01 -1.05329669e+00
2.74606556e-01 -1.02083921e+00 -4.26185369e-01 -1.11656012e-02
-3.52385461e-01 -1.22241400e-01 7.70440340e-01 -7.54939795e-01
1.42706513e+00 -2.42422533e+00 7.78164268e-01 2.64983684e-01
6.31620169e-01 2.08231201e-03 3.48383635e-02 4.52891707e-01
-4.30918723e-01 -6.88534379e-02 -7.10936368e-01 -5.95154464e-01
-1.32060274e-01 -6.86321929e-02 -1.79775789e-01 1.19054270e+00
4.76156129e-04 5.37422717e-01 -8.82015705e-01 -3.96800220e-01
1.58875152e-01 6.20628238e-01 -7.71014094e-01 1.45043671e-01
2.09189475e-01 6.74575508e-01 -4.68214393e-01 2.19820023e-01
6.61209524e-01 -2.10289925e-01 2.94721454e-01 -4.82085608e-02
-1.16866522e-01 -7.82143772e-02 -1.25154781e+00 1.55569255e+00
-3.16958696e-01 5.30196309e-01 2.53998846e-01 -1.33458710e+00
1.06995058e+00 1.82413459e-01 1.06855059e+00 -4.11714733e-01
-2.30529860e-01 1.02404088e-01 -1.46857172e-01 -4.10369158e-01
2.23746091e-01 -5.28397799e-01 -1.50341824e-01 3.55967700e-01
-3.82038742e-01 4.67628479e-01 3.46865579e-02 1.65979519e-01
9.52533424e-01 -1.59087867e-01 5.28825104e-01 -5.30643344e-01
5.61009824e-01 -1.49268985e-01 8.66908729e-01 3.82572621e-01
6.65694699e-02 4.66212511e-01 8.28100741e-01 -5.97197041e-02
-9.63768482e-01 -9.40578341e-01 -6.42948031e-01 6.74021244e-01
-1.56978499e-02 -1.79996580e-01 -5.85872650e-01 -5.98155200e-01
2.77998209e-01 6.42697871e-01 -8.44066083e-01 -4.91002679e-01
-5.61074793e-01 -1.11267078e+00 1.94975268e-02 1.72596350e-01
2.72731602e-01 -3.50951344e-01 -3.08886647e-01 2.29042366e-01
-3.05025671e-02 -6.40697360e-01 -4.20200080e-01 7.44966492e-02
-1.06814432e+00 -1.15972948e+00 -8.18107665e-01 -5.68262994e-01
8.92748713e-01 2.85303712e-01 7.92776227e-01 -3.94701421e-01
-3.01510096e-01 3.14679384e-01 -1.23285256e-01 -7.32341558e-02
-2.62961119e-01 4.31425348e-02 2.82105654e-01 5.72047293e-01
3.44220191e-01 -5.55625081e-01 -6.29526377e-01 4.78687108e-01
-1.12856519e+00 2.17000395e-02 2.82186121e-01 1.00007033e+00
8.19402456e-01 3.21547896e-01 7.08896697e-01 -1.16326141e+00
5.22372723e-01 -7.69216657e-01 -7.61419356e-01 2.58544926e-02
-8.56386840e-01 2.92686433e-01 4.89953876e-01 -5.11166096e-01
-9.18540418e-01 1.50631085e-01 3.58589292e-01 -3.50642204e-01
1.26701340e-01 5.63736320e-01 -6.25604331e-01 2.75112480e-01
4.16159749e-01 -1.67908445e-02 2.56294608e-01 -9.17676508e-01
1.17158800e-01 6.92902982e-01 1.59741655e-01 -2.54986078e-01
9.19525445e-01 9.16002274e-01 5.91690361e-01 -8.07173312e-01
-7.91518807e-01 -6.67559564e-01 -7.54128993e-01 2.36630037e-01
7.38130689e-01 -6.48813307e-01 -5.07866621e-01 3.30925196e-01
-5.00761092e-01 -1.34333745e-01 -4.47153628e-01 8.18194091e-01
-6.56989217e-01 3.99020761e-01 -2.09987357e-01 -6.96338534e-01
-1.51684642e-01 -9.91350591e-01 1.04698133e+00 -3.58734369e-01
-2.07116798e-01 -1.32295167e+00 6.27457500e-01 2.92308360e-01
2.90354132e-03 6.55742109e-01 1.25132692e+00 -5.94417810e-01
-4.46133539e-02 -4.93989557e-01 6.34532869e-02 3.78141463e-01
5.33839643e-01 -3.89652193e-01 -6.38936222e-01 -5.16550064e-01
4.52916324e-01 2.54550606e-01 8.53831947e-01 7.22371221e-01
1.02355576e+00 -3.61038655e-01 -4.48975235e-01 7.77729571e-01
1.40546894e+00 -1.22321308e-01 3.08041543e-01 1.13192186e-01
7.31950521e-01 9.78605270e-01 6.03869498e-01 5.06452262e-01
2.28353187e-01 1.02940071e+00 1.46338284e-01 -3.84910613e-01
1.62100047e-01 -7.08868802e-02 7.03687742e-02 5.59685528e-01
1.02457151e-01 5.38414083e-02 -7.64080405e-01 4.54516858e-01
-1.99759996e+00 -7.44436741e-01 -4.24649000e-01 2.54739165e+00
7.11813450e-01 -1.30150750e-01 2.87743360e-01 2.90404648e-01
7.90221870e-01 8.37341025e-02 -6.48557007e-01 -1.63778570e-02
-3.01131845e-01 -1.12092495e-01 3.54048878e-01 5.20153761e-01
-1.19745541e+00 4.88937318e-01 5.68712568e+00 6.17043793e-01
-7.46632159e-01 1.15601622e-01 6.20901763e-01 -1.13924876e-01
-1.91492781e-01 -6.65336177e-02 -7.70216346e-01 3.32025528e-01
6.92982554e-01 -3.91103685e-01 3.46014768e-01 7.33498633e-01
7.08900869e-01 1.10025652e-01 -1.15937662e+00 9.91762519e-01
9.18680504e-02 -8.31360877e-01 -2.13819191e-01 6.11907959e-01
9.09881353e-01 -6.77729905e-01 3.16049010e-01 -4.35415003e-03
-1.42047241e-01 -9.50489521e-01 1.10870905e-01 6.23375952e-01
8.25562179e-01 -9.12676692e-01 4.18951273e-01 2.13279232e-01
-8.42281103e-01 -5.94839044e-02 -4.10249650e-01 7.41401762e-02
8.60610008e-02 9.21800315e-01 -5.18913984e-01 3.88325006e-01
2.09850907e-01 1.01930916e+00 -3.67719054e-01 8.37474585e-01
1.54226601e-01 5.78895509e-01 -1.13447092e-01 5.48952878e-01
-4.99609485e-02 -8.03717792e-01 9.29722309e-01 8.35468054e-01
9.93751362e-02 8.72184522e-03 -1.23977453e-01 7.79969811e-01
-3.69920023e-02 3.57079893e-01 -7.29567766e-01 -1.14896214e-02
1.81759953e-01 1.14158857e+00 -5.29029608e-01 -7.19143674e-02
-3.85759354e-01 7.66881883e-01 1.35171100e-01 3.39198351e-01
-4.67080086e-01 -1.93275899e-01 9.18051004e-01 2.44795039e-01
4.89279814e-03 -2.39558533e-01 -4.13919300e-01 -1.22482455e+00
1.02533780e-01 -8.86417806e-01 6.37769043e-01 3.10329758e-02
-1.34015143e+00 1.07064441e-01 1.83104381e-01 -1.30898368e+00
-2.99951553e-01 -3.07199478e-01 -1.32251203e-01 9.13923204e-01
-1.19920778e+00 -8.64030600e-01 -2.74087340e-02 6.70918703e-01
3.06055427e-01 -2.32119769e-01 8.04157794e-01 4.32470679e-01
-8.84141207e-01 5.11821270e-01 8.02385747e-01 -1.32275313e-01
7.19810903e-01 -1.34316850e+00 -1.19864672e-01 7.84411669e-01
-2.21931428e-01 6.65723681e-01 8.33585978e-01 -8.66377294e-01
-1.50023639e+00 -1.14329898e+00 7.37775445e-01 -2.35437945e-01
5.35108328e-01 -6.18893325e-01 -1.04447782e+00 6.32715166e-01
-5.37367642e-01 -1.26519492e-02 8.66422772e-01 2.81775236e-01
-2.37654686e-01 -3.63490045e-01 -1.16238034e+00 6.89002991e-01
9.83130395e-01 -3.88496161e-01 -3.08898777e-01 3.93767893e-01
5.16089499e-01 1.01059042e-01 -1.03791368e+00 3.89852256e-01
3.80066186e-01 -5.85332811e-01 1.18188381e+00 -9.35600579e-01
4.28673118e-01 -2.36819033e-02 -2.16071725e-01 -1.22231722e+00
-4.23797816e-01 -6.74492657e-01 -2.69744813e-01 1.38331497e+00
3.59886289e-02 -7.48225451e-01 8.05152595e-01 7.26462901e-01
2.22398058e-01 -5.84417641e-01 -1.10008347e+00 -8.58882844e-01
1.96558505e-01 -1.24372095e-01 2.95815974e-01 1.15023804e+00
-1.41340047e-01 4.71352518e-01 -6.13124788e-01 1.08464248e-01
1.13018882e+00 2.74463207e-01 7.15404391e-01 -1.49086618e+00
-2.69832611e-01 -2.34298438e-01 -4.42702532e-01 -6.20947242e-01
3.62328470e-01 -1.04485357e+00 -3.11452657e-01 -1.20859516e+00
4.41728443e-01 -5.48310041e-01 -2.74437398e-01 2.75463972e-04
-2.71764219e-01 -5.35363369e-02 -1.12206869e-01 3.22092384e-01
-2.46954039e-01 7.99825132e-01 1.06477118e+00 -8.03421140e-02
-6.80603862e-01 2.85539836e-01 -7.86985755e-01 5.27025521e-01
6.84872568e-01 -8.31245005e-01 -4.60569263e-01 6.47950992e-02
2.21991822e-01 1.69251516e-01 3.73422503e-01 -5.57647645e-01
-1.40665740e-01 -2.42279083e-01 7.71590620e-02 -3.93582940e-01
3.14168155e-01 -9.50943291e-01 3.88856322e-01 4.78366822e-01
-6.38598919e-01 -2.06885517e-01 -3.21610510e-01 7.31601179e-01
-1.72260687e-01 -7.92264268e-02 8.77652645e-01 3.93089265e-01
1.23296620e-03 3.70263726e-01 -1.05628841e-01 1.62827954e-01
1.05797529e+00 8.30551609e-02 -5.84118068e-02 -1.35379687e-01
-7.48220921e-01 6.45012185e-02 4.21925336e-01 3.16728175e-01
5.46270907e-01 -1.41266036e+00 -9.94546115e-01 2.94069082e-01
6.96700737e-02 -6.88204095e-02 2.92989016e-01 1.15557265e+00
3.43128256e-02 5.34869671e-01 6.53593093e-02 -4.37306046e-01
-1.39589441e+00 8.71211886e-01 1.90142229e-01 -4.21955824e-01
-9.33760524e-01 3.24805230e-01 7.57627845e-01 -2.48356357e-01
1.00230798e-01 7.81036317e-02 -3.61446112e-01 4.68092352e-01
4.47853684e-01 8.91751826e-01 -1.14314575e-02 -7.64811456e-01
-2.33060062e-01 5.25956631e-01 5.97127527e-02 -3.60939652e-02
1.76098120e+00 -3.74869704e-01 -2.95676440e-01 4.40295100e-01
1.73721230e+00 1.46391109e-01 -1.35871577e+00 -4.70217973e-01
2.05803797e-01 -5.30734777e-01 2.36860856e-01 -1.22445337e-01
-1.05759537e+00 6.89265072e-01 6.72838211e-01 2.69816756e-01
1.25030911e+00 -3.33340317e-02 3.18298072e-01 1.07847946e-02
9.03473273e-02 -9.84188974e-01 -1.36463111e-02 -1.45426570e-02
9.36467111e-01 -1.10339344e+00 2.25176048e-02 -4.21400130e-01
-4.32634920e-01 7.84824669e-01 1.40941039e-01 -2.28586286e-01
8.24303150e-01 -1.21193215e-01 -2.26552546e-01 -2.70720661e-01
-4.81772095e-01 -4.10757288e-02 4.24578547e-01 4.73267943e-01
2.39681810e-01 3.00364822e-01 -6.83877170e-01 5.39752066e-01
-5.09798042e-02 -2.66425580e-01 3.51977557e-01 6.27471685e-01
7.86998793e-02 -1.16878450e+00 -5.27980983e-01 7.52862036e-01
-4.80861872e-01 1.10222273e-01 -2.37535760e-01 7.82579780e-01
-1.48846939e-01 8.17371249e-01 4.59296852e-02 5.28499577e-03
3.06615502e-01 -5.15282489e-02 1.43493921e-01 -6.29238665e-01
-2.13991478e-01 4.54087079e-01 -2.24057600e-01 -4.58830595e-01
-5.36006629e-01 -1.19534695e+00 -8.83903921e-01 -7.79270232e-02
-2.49439105e-01 2.21100196e-01 6.00331068e-01 8.77989888e-01
2.72700608e-01 2.99245834e-01 1.14509046e+00 -4.35034305e-01
-5.87689817e-01 -6.77891016e-01 -1.12669349e+00 7.03364789e-01
6.19385302e-01 -8.02738607e-01 -5.79248428e-01 5.11367954e-02] | [7.597624778747559, 4.387448310852051] |
461d5489-8a00-4d48-81d4-d7776d41758f | sim-semantic-aware-instance-mask-generation | 2303.08578 | null | https://arxiv.org/abs/2303.08578v1 | https://arxiv.org/pdf/2303.08578v1.pdf | SIM: Semantic-aware Instance Mask Generation for Box-Supervised Instance Segmentation | Weakly supervised instance segmentation using only bounding box annotations has recently attracted much research attention. Most of the current efforts leverage low-level image features as extra supervision without explicitly exploiting the high-level semantic information of the objects, which will become ineffective when the foreground objects have similar appearances to the background or other objects nearby. We propose a new box-supervised instance segmentation approach by developing a Semantic-aware Instance Mask (SIM) generation paradigm. Instead of heavily relying on local pair-wise affinities among neighboring pixels, we construct a group of category-wise feature centroids as prototypes to identify foreground objects and assign them semantic-level pseudo labels. Considering that the semantic-aware prototypes cannot distinguish different instances of the same semantics, we propose a self-correction mechanism to rectify the falsely activated regions while enhancing the correct ones. Furthermore, to handle the occlusions between objects, we tailor the Copy-Paste operation for the weakly-supervised instance segmentation task to augment challenging training data. Extensive experimental results demonstrate the superiority of our proposed SIM approach over other state-of-the-art methods. The source code: https://github.com/lslrh/SIM. | ['Lei Zhang', 'Liyi Chen', 'Shuai Li', 'Yabin Zhang', 'Chenhang He', 'Ruihuang Li'] | 2023-03-14 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_SIM_Semantic-Aware_Instance_Mask_Generation_for_Box-Supervised_Instance_Segmentation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_SIM_Semantic-Aware_Instance_Mask_Generation_for_Box-Supervised_Instance_Segmentation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['box-supervised-instance-segmentation', 'weakly-supervised-instance-segmentation'] | ['computer-vision', 'computer-vision'] | [ 6.02809131e-01 3.85777146e-01 -3.11866373e-01 -7.16376603e-01
-6.76509321e-01 -4.04193252e-01 4.22706991e-01 3.90632153e-01
-4.14366901e-01 5.28827965e-01 -1.56708613e-01 2.92024408e-02
1.42517492e-01 -6.92512751e-01 -9.31869149e-01 -9.29816544e-01
3.38817209e-01 4.42964047e-01 8.70442152e-01 1.65022403e-01
2.27072164e-01 1.59282550e-01 -1.59502268e+00 4.46589351e-01
1.14870608e+00 1.06939363e+00 4.11708444e-01 1.58032507e-01
-4.43953574e-01 4.86576766e-01 -4.83122766e-01 -2.39925250e-01
3.64573926e-01 -4.30689633e-01 -7.00027585e-01 4.73665774e-01
6.28523111e-01 -6.47051856e-02 2.87464336e-02 1.35723603e+00
-1.21101495e-02 1.24645352e-01 3.90823007e-01 -1.16246092e+00
-4.75691617e-01 3.75494778e-01 -1.05313897e+00 2.61792928e-01
-9.55534801e-02 1.77137554e-01 9.24216986e-01 -9.75118935e-01
4.75438416e-01 1.08856511e+00 3.83672625e-01 4.69529897e-01
-1.18304563e+00 -6.91418052e-01 7.98076272e-01 2.39066437e-01
-1.51367843e+00 -2.97729790e-01 9.53237236e-01 -2.06876189e-01
1.72607824e-01 2.36167222e-01 5.46997368e-01 6.18121445e-01
-5.90511978e-01 1.10516000e+00 1.03885114e+00 -2.87963986e-01
3.30978930e-01 4.29125398e-01 3.27632993e-01 7.55459964e-01
3.13483238e-01 -4.56215709e-01 -2.56917626e-01 -1.08960196e-01
7.17095554e-01 3.78116995e-01 -3.46912771e-01 -6.91033244e-01
-1.26485682e+00 5.45974493e-01 7.16557562e-01 4.22957897e-01
-2.47479722e-01 1.31683592e-02 1.07939035e-01 -3.38519961e-01
6.41668737e-01 8.12189430e-02 -7.28614509e-01 3.82497728e-01
-1.07399738e+00 7.03117624e-02 2.67965287e-01 9.10521686e-01
1.23842406e+00 -4.16340888e-01 -2.57665306e-01 7.88280606e-01
2.91710943e-01 1.36621252e-01 3.63626391e-01 -7.58735001e-01
4.89488155e-01 1.16186893e+00 8.51093903e-02 -9.06208813e-01
-1.13426998e-01 -6.84629023e-01 -4.67566818e-01 -1.48547098e-01
6.31124735e-01 1.27931282e-01 -1.21610057e+00 1.47327030e+00
8.67553174e-01 6.59655869e-01 -1.05672978e-01 1.10989678e+00
7.31360853e-01 6.53229773e-01 1.49844676e-01 -3.66612077e-02
1.36095428e+00 -1.34777439e+00 -4.88780975e-01 -5.02196908e-01
6.13655329e-01 -6.02920771e-01 1.28113163e+00 -5.97662181e-02
-9.22185957e-01 -5.50398886e-01 -9.05663848e-01 7.27792084e-02
-3.29778403e-01 3.28171104e-01 6.29258573e-01 4.21428889e-01
-5.90896547e-01 3.39268386e-01 -8.31752241e-01 -1.07828289e-01
8.85408163e-01 2.25593224e-01 -1.62116662e-01 -1.55869648e-01
-7.62274683e-01 1.66056797e-01 7.03654885e-01 2.00352266e-01
-7.97940671e-01 -7.22199082e-01 -7.81236112e-01 -8.90187994e-02
9.34098959e-01 -3.18163484e-01 9.72618043e-01 -1.53312528e+00
-1.02506661e+00 1.06948221e+00 -4.19204980e-01 -3.13580297e-02
4.93157893e-01 -1.67847574e-01 -1.90095622e-02 2.18796015e-01
4.11468774e-01 7.58915067e-01 1.04358530e+00 -1.69810319e+00
-9.89843130e-01 -5.27580440e-01 6.23458549e-02 2.40638167e-01
-3.39638829e-01 -2.19173267e-01 -8.69309366e-01 -7.09693789e-01
6.41615570e-01 -8.42097104e-01 -3.16739351e-01 2.03605950e-01
-6.21513963e-01 -1.95752203e-01 1.12960505e+00 -3.23900729e-01
1.04469168e+00 -2.30464792e+00 -6.36878163e-02 3.27299654e-01
9.22854170e-02 3.12532037e-01 -3.87871526e-02 -1.81733325e-01
1.54336601e-01 -2.04640236e-02 -7.64784396e-01 -2.73599446e-01
-2.24728182e-01 3.29695433e-01 -1.52595744e-01 4.49559987e-01
3.96867752e-01 7.77055204e-01 -1.11804616e+00 -7.40538001e-01
2.10205033e-01 3.10675919e-01 -4.82950777e-01 2.15228215e-01
-5.14467359e-01 6.96994543e-01 -6.90884173e-01 8.01262736e-01
8.08532834e-01 -4.60647076e-01 -5.06078675e-02 -2.99353212e-01
1.38206899e-01 7.09199458e-02 -1.34414744e+00 1.71353316e+00
-1.50896743e-01 8.95507485e-02 2.09829956e-01 -1.13620937e+00
7.24145889e-01 -1.51719242e-01 2.30995461e-01 -3.26772779e-01
3.11740804e-02 3.35481256e-01 -9.68369618e-02 -3.99433970e-01
2.30692536e-01 4.12700279e-03 5.83407879e-02 2.37466440e-01
-1.37100771e-01 1.06453113e-01 1.92590013e-01 3.42970073e-01
7.48959899e-01 4.33688670e-01 -1.05144624e-02 -1.79544538e-01
7.31981635e-01 5.84941134e-02 1.00731170e+00 6.90081596e-01
-3.20953250e-01 9.13693607e-01 2.98900217e-01 -1.30335182e-01
-6.50669336e-01 -1.05940449e+00 -2.35608652e-01 1.33571935e+00
7.07873642e-01 -1.77558586e-01 -1.10244858e+00 -1.01294219e+00
-3.36410152e-03 3.71751577e-01 -6.82634294e-01 -2.09226017e-03
-5.21485269e-01 -8.36656332e-01 2.32046410e-01 6.78202808e-01
6.83375657e-01 -8.67904544e-01 -4.27416116e-01 1.04792379e-02
-2.55700946e-01 -1.17504859e+00 -5.80797434e-01 2.32106477e-01
-9.60063815e-01 -1.02485096e+00 -9.08585906e-01 -1.04094398e+00
1.26912463e+00 5.34263253e-01 7.20262527e-01 4.85099733e-01
-3.18063736e-01 -2.52400991e-02 -3.95742983e-01 -9.71262306e-02
-5.93908690e-02 -5.58105931e-02 -2.40754291e-01 5.57897210e-01
2.63973117e-01 -2.92431563e-01 -9.13355231e-01 5.00284016e-01
-9.48689640e-01 2.08164558e-01 6.62242234e-01 7.33608842e-01
1.01129425e+00 2.42014304e-02 4.64778215e-01 -1.31675231e+00
-1.43008530e-01 -4.01655257e-01 -4.40430641e-01 2.38543347e-01
-1.57098159e-01 -6.82170242e-02 3.87379199e-01 -5.03975809e-01
-1.22775221e+00 4.19638842e-01 2.36283690e-01 -5.22790670e-01
-5.09458601e-01 -1.01487478e-02 -5.47975659e-01 1.93390056e-01
1.97850674e-01 2.87036449e-01 -5.12301266e-01 -5.49221694e-01
5.25871217e-01 6.00519300e-01 5.91745734e-01 -6.24824345e-01
8.65372419e-01 9.07343328e-01 -4.19878393e-01 -6.69654429e-01
-1.22888434e+00 -8.19240391e-01 -8.22708130e-01 -8.07526857e-02
9.64293659e-01 -9.52530265e-01 -1.04950421e-01 4.34875607e-01
-9.83623385e-01 -3.98263156e-01 -3.56834412e-01 1.08739041e-01
-1.65235192e-01 3.32672119e-01 -3.71742517e-01 -6.81434631e-01
4.32433523e-02 -1.17207468e+00 1.45109832e+00 5.17252564e-01
3.87469791e-02 -6.16026819e-01 -3.92774761e-01 7.67229736e-01
-3.60144787e-02 1.45913824e-01 8.52805674e-01 -7.83088565e-01
-8.58170867e-01 -2.38536149e-01 -6.56893671e-01 2.92120308e-01
2.00997010e-01 -9.84359607e-02 -1.17863238e+00 -6.56610355e-02
-2.35059693e-01 -6.85773045e-02 1.04492605e+00 2.47002289e-01
1.52106225e+00 -3.52445602e-01 -6.88951313e-01 6.01420105e-01
1.18912828e+00 -1.95127372e-02 1.74042836e-01 1.86862722e-01
1.22762835e+00 8.77712429e-01 9.19269264e-01 2.90995896e-01
2.94717342e-01 6.36637449e-01 4.77673143e-01 -4.09136623e-01
-3.48097794e-02 -3.44202071e-01 -1.37611234e-04 3.09376329e-01
1.80071697e-01 -3.41065452e-02 -7.33309329e-01 7.66303003e-01
-2.01908994e+00 -5.95243156e-01 -4.04721320e-01 2.16409135e+00
8.91506970e-01 3.59385878e-01 8.06580707e-02 8.83206576e-02
1.15906727e+00 1.50514469e-01 -7.36382246e-01 1.80866748e-01
-3.48690338e-02 5.51325008e-02 4.62377101e-01 3.08045447e-01
-1.30621541e+00 1.15258205e+00 4.09083366e+00 1.01152456e+00
-8.51471484e-01 3.60271275e-01 1.04485369e+00 -7.19004795e-02
-2.97206849e-01 2.04613969e-01 -7.84592450e-01 6.45819783e-01
1.49165824e-01 3.57388377e-01 -3.74045596e-02 9.21905875e-01
2.76143849e-01 -4.04940605e-01 -9.83316422e-01 6.40772283e-01
1.29480645e-01 -1.03919792e+00 3.01221367e-02 -2.73604453e-01
1.04795146e+00 -2.79919833e-01 -3.25819775e-02 6.68349043e-02
-1.35971829e-01 -5.74712753e-01 8.57794940e-01 2.09059462e-01
3.61196518e-01 -5.89729190e-01 6.58034265e-01 3.78993601e-01
-1.35066903e+00 -7.74819553e-02 -4.17307794e-01 4.21541855e-02
-1.12526072e-03 9.32645857e-01 -5.74363530e-01 2.16079623e-01
8.61069262e-01 6.58448815e-01 -7.61296093e-01 9.59037006e-01
-4.57889676e-01 6.87211096e-01 -3.96760672e-01 2.50129849e-01
4.09771979e-01 -2.55511552e-01 3.77204001e-01 1.10345697e+00
-1.94531471e-01 1.38023436e-01 4.60278958e-01 1.12580764e+00
-8.08259472e-02 2.32622787e-01 -1.11078419e-01 2.16835588e-01
4.28635120e-01 1.42683744e+00 -1.37829840e+00 -5.92193425e-01
-5.14827371e-01 1.32758391e+00 2.49553874e-01 4.35707748e-01
-8.92539561e-01 -2.96760172e-01 5.45064032e-01 3.20799261e-01
4.92582202e-01 1.34666488e-01 -5.00982940e-01 -1.09894907e+00
4.68952864e-01 -4.44665164e-01 3.41244876e-01 -6.66998565e-01
-1.13063645e+00 2.06834599e-01 -8.95952210e-02 -1.14972651e+00
4.39277142e-01 -2.73213744e-01 -8.32955778e-01 4.86256331e-01
-1.64332938e+00 -1.22393417e+00 -5.84803522e-01 4.44428444e-01
6.86116338e-01 4.18497890e-01 1.97102323e-01 2.76575595e-01
-8.18267703e-01 4.42960531e-01 -4.17050645e-02 2.61587650e-01
5.66418469e-01 -1.26388824e+00 8.48563109e-03 9.53260958e-01
7.53554404e-02 6.82586312e-01 4.64745462e-01 -6.84272110e-01
-7.96493292e-01 -1.38985932e+00 5.30524611e-01 -2.96952814e-01
4.72790778e-01 -5.51103175e-01 -1.26470876e+00 4.21998024e-01
-2.22554296e-01 5.40141165e-01 4.05115575e-01 -1.98781073e-01
-2.54201472e-01 -2.02431992e-01 -1.11938095e+00 5.68576694e-01
1.26274669e+00 -3.02554399e-01 -5.09218037e-01 4.10949141e-01
8.04017246e-01 -2.17266455e-01 -3.72741431e-01 6.22360408e-01
1.01830691e-01 -9.35442567e-01 9.39939141e-01 -2.78329074e-01
3.54118735e-01 -9.01934564e-01 7.65173286e-02 -8.59788656e-01
2.89806724e-02 -1.52375430e-01 2.75557432e-02 1.56355262e+00
4.48839456e-01 -4.72515613e-01 1.08737051e+00 5.35307705e-01
-2.24865377e-01 -8.61468434e-01 -7.86536157e-01 -6.90373003e-01
-2.63832688e-01 -2.15243340e-01 5.09724975e-01 1.02052212e+00
-2.81212181e-01 8.17542803e-03 1.05251662e-01 4.67163205e-01
6.92749381e-01 5.73800027e-01 7.48978972e-01 -1.09190488e+00
-3.85619193e-01 -3.25880378e-01 -4.70864803e-01 -1.11086881e+00
1.35344058e-01 -9.40573573e-01 4.94321525e-01 -1.40141809e+00
4.81061548e-01 -8.63320947e-01 -5.06282270e-01 7.78391540e-01
-6.99412525e-01 6.42819345e-01 4.72980142e-02 3.36901784e-01
-9.87492144e-01 5.63344836e-01 1.24518013e+00 -2.28012741e-01
-2.80528516e-01 -2.77304910e-02 -6.73802912e-01 1.01478064e+00
7.78070688e-01 -6.22051895e-01 -3.80636275e-01 -2.60931075e-01
-1.96909457e-01 -4.55855697e-01 6.38692617e-01 -9.67762768e-01
2.69826408e-02 -2.32250229e-01 4.22109216e-01 -4.98064935e-01
9.03939232e-02 -9.55618441e-01 -2.11323887e-01 2.80966073e-01
-3.61583114e-01 -4.50253993e-01 8.83560032e-02 8.92948031e-01
-1.55848488e-01 -3.61725301e-01 9.34604049e-01 -2.78456420e-01
-8.90507400e-01 3.06110293e-01 -4.58702259e-02 2.89601654e-01
1.36976624e+00 -3.45694631e-01 -2.74108201e-01 1.56344876e-01
-6.25527084e-01 2.75385857e-01 7.47851193e-01 4.26483184e-01
5.24578750e-01 -8.23240995e-01 -4.46847498e-01 2.93514311e-01
5.21525562e-01 6.71285093e-01 2.61096328e-01 9.15733099e-01
-1.97578803e-01 -3.75397466e-02 2.40665779e-01 -8.67267311e-01
-1.30394721e+00 4.58053380e-01 2.72865266e-01 1.51846856e-01
-6.34551108e-01 1.11343229e+00 9.34099972e-01 -3.27837437e-01
2.62520462e-01 -4.69385356e-01 9.14866552e-02 -8.78999010e-02
4.29917276e-01 8.85846317e-02 -4.55164462e-02 -7.77358592e-01
-5.78142047e-01 5.67557096e-01 -3.73064578e-01 2.01226026e-01
1.18121517e+00 -1.73514754e-01 -7.53449947e-02 4.17982429e-01
1.03065431e+00 -9.53985080e-02 -1.50894332e+00 -4.11658198e-01
1.94263116e-01 -6.76091731e-01 1.23469852e-01 -6.66078627e-01
-1.44162834e+00 8.27239215e-01 6.13309085e-01 -2.59045631e-01
1.07232821e+00 4.42239910e-01 8.21871698e-01 5.86693287e-02
2.47233197e-01 -1.16755497e+00 2.12640896e-01 4.50953916e-02
3.26664537e-01 -1.41944277e+00 -1.20943375e-01 -9.66002405e-01
-5.91441572e-01 6.55666113e-01 9.67122793e-01 -1.62817404e-01
3.66844565e-01 5.64051643e-02 1.72663659e-01 -1.21338323e-01
-2.73013294e-01 -5.63442469e-01 2.66854256e-01 5.31047106e-01
3.18394214e-01 -2.34986655e-02 -2.49139205e-01 7.11887658e-01
4.78793681e-01 -2.80151278e-01 2.06009299e-01 1.08526862e+00
-6.16823494e-01 -1.00589240e+00 -4.32040542e-01 4.92787182e-01
-3.26949060e-01 -1.64926186e-01 -4.60110992e-01 5.45653760e-01
4.95837390e-01 8.78289282e-01 2.35750914e-01 -7.21223745e-03
1.89710438e-01 9.60510597e-02 2.49948636e-01 -8.55713248e-01
-4.82023358e-01 2.70148218e-01 -3.46860319e-01 -4.67156917e-01
-6.18100345e-01 -7.22640395e-01 -1.81297457e+00 3.66875380e-01
-5.68248987e-01 2.20113825e-02 4.10576135e-01 9.83040988e-01
3.00162494e-01 5.45086265e-01 4.95103389e-01 -9.78702903e-01
-1.15922624e-02 -6.97439790e-01 -4.57166851e-01 7.72732615e-01
3.08316886e-01 -7.78704345e-01 -3.69673163e-01 2.81749398e-01] | [9.538690567016602, 0.6253294944763184] |
3137ae03-7336-4e0c-82e4-8e7429d37e90 | morphological-sampling-theorem-and-its | 2305.13279 | null | https://arxiv.org/abs/2305.13279v1 | https://arxiv.org/pdf/2305.13279v1.pdf | Morphological Sampling Theorem and its Extension to Grey-value Images | Sampling is a basic operation in image processing. In classic literature, a morphological sampling theorem has been established, which shows how sampling interacts by morphological operations with image reconstruction. Many aspects of morphological sampling have been investigated for binary images, but only some of them have been explored for grey-value imagery. With this paper, we make a step towards completion of this open matter. By relying on the umbra notion, we show how to transfer classic theorems in binary morphology about the interaction of sampling with the fundamental morphological operations dilation, erosion, opening and closing, to the grey-value setting. In doing this we also extend the theory relating the morphological operations and corresponding reconstructions to use of non-flat structuring elements. We illustrate the theoretical developments at hand of examples. | ['Michael Breuß', 'Vivek Sridhar'] | 2023-05-22 | null | null | null | null | ['image-reconstruction'] | ['computer-vision'] | [ 9.15000379e-01 2.62513727e-01 3.91017824e-01 -2.02252775e-01
3.55053246e-02 -5.34348905e-01 7.51364231e-01 2.45332301e-01
-6.01936460e-01 5.91794491e-01 -2.54601657e-01 -3.82938534e-01
-3.50148767e-01 -1.09611380e+00 -3.41199815e-01 -9.52268541e-01
-3.48136246e-01 8.54044929e-02 3.65006626e-01 -4.00182694e-01
4.37983394e-01 9.45682526e-01 -1.43776524e+00 1.91312850e-01
6.24569833e-01 9.71546233e-01 2.45047845e-02 9.43243861e-01
-3.32085848e-01 3.77915502e-01 -6.75222278e-01 -6.00895703e-01
4.41933215e-01 -6.62221014e-01 -8.76900494e-01 7.24079549e-01
2.37980619e-01 1.64117869e-02 2.07787856e-01 1.46050501e+00
3.37755501e-01 2.02657506e-01 8.29001546e-01 -8.55910540e-01
-8.30519438e-01 4.23878998e-01 -6.62779033e-01 1.93472862e-01
9.40537080e-02 -2.53987521e-01 6.18526399e-01 -1.01152444e+00
7.47484028e-01 1.25596714e+00 7.89303184e-01 2.28271380e-01
-1.45631313e+00 2.86829978e-01 -4.60146129e-01 2.75247067e-01
-1.16780794e+00 -2.97446221e-01 6.23890877e-01 -2.48433262e-01
5.22569597e-01 6.54387057e-01 8.95124853e-01 6.57389611e-02
1.18478291e-01 5.73777258e-01 1.50067210e+00 -1.07512367e+00
2.04961106e-01 1.93405420e-01 1.62503630e-01 6.44109011e-01
4.54226881e-01 5.14570735e-02 -2.97116488e-02 8.97521079e-02
1.19259799e+00 -1.70199186e-01 -3.62744808e-01 -3.68978024e-01
-1.05374312e+00 7.60396898e-01 3.30277503e-01 5.48165441e-01
-2.21447363e-01 -1.55813307e-01 4.04316410e-02 6.48927152e-01
4.71821666e-01 3.87950808e-01 5.45412451e-02 4.55342501e-01
-1.02932537e+00 -1.86687484e-01 9.86556530e-01 8.66589904e-01
8.87519002e-01 -2.54377481e-02 2.69237787e-01 5.24406910e-01
-3.54952179e-02 4.62014049e-01 1.15931667e-01 -1.01029873e+00
-1.77831650e-01 3.73012632e-01 -1.05216503e-01 -9.89617467e-01
-2.30339199e-01 -9.43754986e-02 -8.51424456e-01 8.73919129e-01
6.93150222e-01 2.65287101e-01 -6.70839429e-01 1.12645769e+00
3.40730488e-01 -2.29509309e-01 5.75061142e-02 3.15848917e-01
3.99449795e-01 4.40061629e-01 -2.78944790e-01 -5.18754482e-01
1.40064144e+00 -5.08884966e-01 -9.22416329e-01 4.52665389e-01
-3.90529893e-02 -9.11223054e-01 1.03446937e+00 8.18556905e-01
-1.40190482e+00 -4.15345103e-01 -1.11063981e+00 -1.64973125e-01
-6.59827292e-01 9.97086167e-02 5.91723859e-01 9.47648883e-01
-1.14722645e+00 9.27366197e-01 -6.54990673e-01 -7.88815916e-01
2.74646342e-01 2.52063990e-01 -1.85585484e-01 3.62191170e-01
-7.32374489e-01 1.05114388e+00 2.62498170e-01 4.39733475e-01
-3.68493229e-01 -1.33576840e-01 -6.32794738e-01 -2.32919082e-01
3.39785993e-01 -5.39660513e-01 1.00709736e+00 -1.01790679e+00
-1.45706964e+00 1.13497007e+00 -2.24505160e-02 -6.21612251e-01
8.07652891e-01 1.40982896e-01 -1.84012964e-01 5.98691881e-01
-3.66118729e-01 4.12157059e-01 1.07079923e+00 -1.25332427e+00
-4.05309916e-01 -4.79337245e-01 2.03765914e-01 -1.11917906e-01
-1.45000920e-01 3.33657116e-02 1.26718611e-01 -7.37412095e-01
4.11305249e-01 -3.70307148e-01 -3.15452099e-01 3.78571153e-01
-1.02707431e-01 9.24564600e-02 6.81649983e-01 -4.49237108e-01
1.04830337e+00 -2.24406552e+00 2.10452586e-01 3.50207329e-01
-4.83416654e-02 3.57342809e-02 2.93305784e-01 5.57797790e-01
-9.18828547e-02 2.53306657e-01 -8.88255417e-01 -2.07431912e-01
-4.27351370e-02 2.67567307e-01 -4.65189517e-01 8.25425148e-01
3.52885067e-01 6.40401721e-01 -5.83195925e-01 -7.68212616e-01
4.55827892e-01 6.20285213e-01 -2.85815150e-01 -4.76205856e-01
1.17756963e-01 1.34639993e-01 -1.28110066e-01 6.52948916e-01
8.85318458e-01 2.51616389e-01 -2.55216029e-03 -1.68042094e-01
-4.93602782e-01 -2.85434902e-01 -1.49661076e+00 1.24141419e+00
-1.43487737e-01 6.18165195e-01 7.03146219e-01 -1.20766819e+00
9.70639288e-01 6.43569976e-02 1.89015657e-01 -1.93215221e-01
2.82333285e-01 2.43514270e-01 -8.77096727e-02 -5.03845751e-01
6.37444019e-01 -6.40987337e-01 3.46955091e-01 2.85482705e-01
-2.27486402e-01 -8.16245675e-01 4.33277696e-01 -1.77048847e-01
7.46561289e-01 2.06204951e-02 9.01479602e-01 -5.88097990e-01
8.79755139e-01 1.30076811e-01 -1.01110958e-01 6.14003301e-01
-3.86624150e-02 6.40683293e-01 4.88005191e-01 -1.53979033e-01
-1.03961384e+00 -1.26095545e+00 -7.77667224e-01 6.20307148e-01
2.51490980e-01 -5.80485240e-02 -1.08135438e+00 -3.15595008e-02
-2.20711425e-01 4.55463499e-01 -8.92147064e-01 2.64465570e-01
-6.69978201e-01 -1.06657457e+00 4.05959845e-01 1.82370514e-01
6.78801119e-01 -1.26470292e+00 -1.18686235e+00 2.10900694e-01
3.43710899e-01 -7.98219383e-01 3.02722696e-02 3.27388585e-01
-1.45317566e+00 -1.00767744e+00 -7.64917672e-01 -8.04319024e-01
1.04117584e+00 2.13145584e-01 1.01745021e+00 4.89279568e-01
-4.35372144e-01 4.70809639e-01 -4.90484834e-01 -4.27065670e-01
-5.61854899e-01 -2.78314769e-01 -3.78700644e-01 1.84626102e-01
-5.86218499e-02 -7.59911716e-01 -2.14099735e-01 1.81231514e-01
-1.63985574e+00 -2.29264498e-01 7.75933385e-01 7.10245013e-01
7.30391681e-01 4.44664240e-01 -3.02603822e-02 -8.34756315e-01
4.56943393e-01 2.56387144e-01 -5.11746645e-01 2.61758447e-01
-3.59564334e-01 4.90840413e-02 4.73598093e-01 -4.01743531e-01
-9.04139102e-01 2.16685813e-02 -1.67135075e-01 8.03787410e-02
-1.13632753e-01 3.26494187e-01 -2.95951545e-01 -4.14374977e-01
4.56624389e-01 3.06412131e-01 2.87499785e-01 -5.92866838e-01
5.96640170e-01 4.92493063e-01 7.70369768e-01 -3.31017107e-01
8.97230744e-01 1.28764975e+00 5.18527389e-01 -1.45148921e+00
-2.24409834e-01 -2.18598410e-01 -1.00731528e+00 -9.97516513e-02
8.43239009e-01 -7.27354884e-02 -4.33330864e-01 4.63313192e-01
-1.00885963e+00 -2.86471337e-01 -9.96394515e-01 1.77090645e-01
-9.49974537e-01 7.14587867e-01 -5.70622563e-01 -1.14344168e+00
2.16679787e-03 -9.86178637e-01 7.08294511e-01 1.23477161e-01
2.52549917e-01 -1.17896152e+00 -2.26369619e-01 -2.50859767e-01
3.53959590e-01 3.14216882e-01 9.26269710e-01 -1.20244406e-01
-3.87807161e-01 1.24761546e-02 -7.26393312e-02 3.72344285e-01
3.71372327e-02 2.37418398e-01 -7.56360650e-01 -3.83233614e-02
9.22310054e-01 4.16984856e-01 1.03154206e+00 5.14783859e-01
9.01784837e-01 -2.31910832e-02 -6.39947429e-02 7.58941472e-01
1.79998899e+00 1.46778941e-01 1.07097948e+00 6.28938198e-01
7.80569091e-02 9.10327375e-01 5.87398767e-01 2.31379047e-01
-2.65571892e-01 4.43881661e-01 4.60920900e-01 -3.10592860e-01
-2.70374060e-01 2.86112875e-01 1.88350409e-01 5.74731052e-01
-4.26848561e-01 1.61520496e-01 -4.47221637e-01 4.10375595e-01
-1.14817119e+00 -1.08592010e+00 -5.93766868e-01 2.39829826e+00
8.77770722e-01 1.24801733e-01 6.79402724e-02 7.19740927e-01
1.04858303e+00 3.72238792e-02 5.63835688e-02 -6.08057916e-01
-6.11361563e-01 7.02569306e-01 6.81293845e-01 9.72567499e-01
-1.05344403e+00 5.50358772e-01 6.88810539e+00 8.81238043e-01
-8.91587913e-01 1.84849665e-01 3.77150267e-01 3.86931777e-01
-3.54702681e-01 1.49434552e-01 -4.46133792e-01 1.31845996e-01
3.92654479e-01 -6.77265525e-02 4.95027721e-01 4.77732450e-01
2.00065792e-01 -6.35177851e-01 -7.37776995e-01 6.24376357e-01
1.54317588e-01 -8.25843990e-01 2.75047481e-01 2.67833441e-01
5.17713726e-01 -9.30840075e-01 9.49537903e-02 -4.62943733e-01
-1.15035713e-01 -1.07770622e+00 8.39870870e-01 5.69011629e-01
7.04960644e-01 -5.23338795e-01 3.86884302e-01 1.87722534e-01
-1.26263034e+00 -5.70653006e-03 -5.91718614e-01 -2.49022037e-01
4.58938748e-01 7.66617119e-01 -5.75083971e-01 6.04823828e-01
1.91094294e-01 3.94465089e-01 -6.88721597e-01 1.38324976e+00
-3.33708048e-01 2.38023996e-01 -4.82239902e-01 -9.10478011e-02
6.18850663e-02 -9.13375556e-01 5.85300624e-01 1.29444706e+00
3.83134395e-01 1.64161831e-01 -4.40507323e-01 1.00257754e+00
4.68699723e-01 2.71427512e-01 -5.87200820e-01 2.14017555e-01
1.25531569e-01 1.18149233e+00 -1.55031466e+00 -5.27616441e-01
-2.44867042e-01 9.63685095e-01 -3.84545058e-01 1.33342389e-02
-6.05530798e-01 -7.81166673e-01 2.35134155e-01 2.81991214e-01
5.28958261e-01 -3.98202360e-01 -8.93540621e-01 -7.28772163e-01
-6.55094311e-02 -5.88346839e-01 1.13693140e-01 -5.71916997e-01
-8.94611061e-01 4.71429557e-01 3.59357357e-01 -1.30492282e+00
1.59226581e-01 -8.47236991e-01 -5.96708000e-01 8.54200363e-01
-1.31555045e+00 -7.62563169e-01 -2.18286030e-02 4.90371406e-01
5.69499969e-01 3.47362697e-01 6.67419553e-01 4.55455445e-02
-2.00842023e-01 -8.61660764e-02 7.09349960e-02 -2.10140515e-02
1.93976507e-01 -1.60824108e+00 2.70492703e-01 1.27113175e+00
1.97712719e-01 8.68550003e-01 1.09116268e+00 -4.15444255e-01
-1.36361682e+00 -4.03066397e-01 8.36995423e-01 -1.80991963e-01
6.19565308e-01 -1.15218069e-02 -8.87331188e-01 4.65799183e-01
4.97243702e-01 -1.06434032e-01 2.12100208e-01 -6.13193274e-01
1.41902372e-01 -2.65163090e-02 -1.43702257e+00 5.74862599e-01
1.01430273e+00 -2.37023741e-01 -1.07070839e+00 -5.31423464e-02
1.88738182e-01 -4.62248512e-02 -8.60182941e-01 3.50477725e-01
4.98513013e-01 -1.27989328e+00 1.43730354e+00 -1.73972785e-01
1.75129011e-01 -5.67562580e-01 -7.37191141e-02 -9.64571893e-01
4.61465418e-02 -7.10420370e-01 3.47704083e-01 9.97467041e-01
1.50051758e-01 -8.30000222e-01 4.33115751e-01 6.13145158e-02
-8.29471871e-02 -6.83940411e-01 -1.02488017e+00 -7.37356722e-01
2.13137984e-01 -1.54528663e-01 3.25767756e-01 6.18896902e-01
9.44176223e-03 -6.22653514e-02 -6.95936009e-02 -6.68891072e-02
9.27264154e-01 6.68060109e-02 4.66670513e-01 -1.40936923e+00
-4.35422421e-01 -8.64406586e-01 -5.19014716e-01 -9.60695267e-01
-4.34867412e-01 -7.34585941e-01 -5.26666455e-02 -1.51994693e+00
-1.56279922e-01 -1.98455334e-01 2.54161835e-01 -1.46463439e-01
-2.47578342e-02 5.46179056e-01 2.98993796e-01 3.36866766e-01
-1.11151692e-02 9.92907137e-02 1.53946447e+00 1.84300676e-04
-1.48279071e-01 1.22565910e-01 -5.00130951e-01 1.02379262e+00
7.14254856e-01 -2.94175178e-01 -1.70359105e-01 -2.09329247e-01
3.12329471e-01 -2.03346923e-01 4.86073315e-01 -9.65285957e-01
1.75335541e-01 -1.28798783e-02 3.36672813e-01 -5.50483048e-01
4.10115629e-01 -9.74432647e-01 3.22723895e-01 6.75853193e-01
-3.19359303e-01 3.44534606e-01 -2.12150902e-01 4.13036942e-01
-2.13958248e-01 -1.10272920e+00 1.01377177e+00 -5.34263849e-01
-6.79838479e-01 -2.29324222e-01 -4.21793133e-01 -2.09811568e-01
9.75272179e-01 -8.52034628e-01 -4.23764251e-02 -6.54402003e-02
-1.34944165e+00 -5.39658725e-01 9.34822559e-01 -6.61348045e-01
5.88934898e-01 -9.95604455e-01 -4.81654495e-01 3.18069577e-01
-6.28753543e-01 -1.11267455e-01 -8.52331519e-02 1.48420477e+00
-1.09089267e+00 -7.10086524e-02 -3.66749078e-01 -4.41876292e-01
-1.51712799e+00 7.59755433e-01 3.54493916e-01 3.40294056e-02
-7.32320130e-01 6.57390594e-01 2.38116030e-02 1.44542545e-01
1.60884857e-01 -7.05331922e-01 -2.04138160e-01 3.69500995e-01
7.39900112e-01 6.88804269e-01 7.65736923e-02 -5.06695688e-01
-6.53965920e-02 1.04519308e+00 5.73225915e-01 -6.37310445e-01
1.30915618e+00 -4.29123163e-01 -6.71449959e-01 6.74500287e-01
9.15145576e-01 3.87303591e-01 -9.72545564e-01 3.51491198e-02
3.25847194e-02 -6.47797346e-01 -2.38900870e-01 -3.37308884e-01
-8.54814589e-01 1.16658711e+00 3.47038388e-01 1.25236356e+00
1.65942061e+00 -1.02072522e-01 1.87570482e-01 2.08275333e-01
4.92364079e-01 -8.62371087e-01 -3.31870794e-01 2.79853374e-01
1.05903423e+00 -6.46938324e-01 2.54442215e-01 -9.76894975e-01
-2.09912226e-01 1.70190036e+00 -3.48183364e-01 -4.03309554e-01
6.68057501e-01 6.66856289e-01 -6.13116771e-02 -1.15409426e-01
-1.31195202e-01 -6.13762796e-01 -2.08310574e-01 7.66220808e-01
2.90441036e-01 8.00107718e-02 -1.00438714e+00 -1.69067562e-01
-3.54792506e-01 -4.83743399e-02 9.04625773e-01 1.00834703e+00
-8.72115791e-01 -1.28527832e+00 -9.48333859e-01 2.24626184e-01
-4.65601504e-01 -3.48788828e-01 -5.21890461e-01 1.11375773e+00
2.70636111e-01 6.81441486e-01 4.51440141e-02 3.44917364e-02
3.58272374e-01 -2.76268907e-02 1.12610853e+00 -4.38483149e-01
-6.65433407e-01 1.78339154e-01 -2.34433889e-01 4.59441496e-03
-8.88470829e-01 -1.09119964e+00 -1.05869699e+00 -3.06625813e-01
-3.40371132e-01 1.63790643e-01 6.27839565e-01 6.73381329e-01
-4.30628151e-01 3.66593242e-01 3.11684579e-01 -9.50485349e-01
-1.02811253e+00 -1.04028571e+00 -1.09593213e+00 2.39853427e-01
2.93425024e-01 -4.94519293e-01 -6.62721038e-01 4.27694887e-01] | [11.114587783813477, -2.4598209857940674] |
4e9c90b5-b53c-4d4c-9bd7-b96e5bf44c0f | unsupervised-context-aware-sentence | 2206.03281 | null | https://arxiv.org/abs/2206.03281v1 | https://arxiv.org/pdf/2206.03281v1.pdf | Unsupervised Context Aware Sentence Representation Pretraining for Multi-lingual Dense Retrieval | Recent research demonstrates the effectiveness of using pretrained language models (PLM) to improve dense retrieval and multilingual dense retrieval. In this work, we present a simple but effective monolingual pretraining task called contrastive context prediction~(CCP) to learn sentence representation by modeling sentence level contextual relation. By pushing the embedding of sentences in a local context closer and pushing random negative samples away, different languages could form isomorphic structure, then sentence pairs in two different languages will be automatically aligned. Our experiments show that model collapse and information leakage are very easy to happen during contrastive training of language model, but language-specific memory bank and asymmetric batch normalization operation play an essential role in preventing collapsing and information leakage, respectively. Besides, a post-processing for sentence embedding is also very effective to achieve better retrieval performance. On the multilingual sentence retrieval task Tatoeba, our model achieves new SOTA results among methods without using bilingual data. Our model also shows larger gain on Tatoeba when transferring between non-English pairs. On two multi-lingual query-passage retrieval tasks, XOR Retrieve and Mr.TYDI, our model even achieves two SOTA results in both zero-shot and supervised setting among all pretraining models using bilingual data. | ['Daxin Jiang', 'Ming Gong', 'Nan Duan', 'Linjun Shou', 'Houxing Ren', 'Yaobo Liang', 'Ning Wu'] | 2022-06-07 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [-7.80194402e-02 -3.31527919e-01 -2.59756207e-01 -4.57556427e-01
-1.40651572e+00 -4.98323411e-01 7.01414585e-01 2.56855816e-01
-9.78819609e-01 7.92619109e-01 6.22261047e-01 -3.49787325e-01
2.22884163e-01 -5.67429960e-01 -6.79680765e-01 -3.63383979e-01
9.62044150e-02 5.40234685e-01 1.09016612e-01 -7.24426389e-01
2.41285101e-01 2.04420224e-01 -8.50067914e-01 7.30213821e-01
1.06483519e+00 3.42373163e-01 6.62978292e-01 4.52451676e-01
-6.43581331e-01 6.39768958e-01 -5.40788114e-01 -5.46570420e-01
2.52280354e-01 -3.70025337e-01 -9.96425331e-01 -5.84527016e-01
7.07600534e-01 -2.22295538e-01 -4.93462354e-01 1.16289294e+00
1.05201542e+00 1.46407992e-01 6.39665127e-01 -6.63934827e-01
-9.85989213e-01 1.08509815e+00 -6.23564363e-01 5.37984490e-01
4.30041671e-01 -1.35214597e-01 1.18179417e+00 -1.30832267e+00
7.61453748e-01 1.42671847e+00 4.67387706e-01 4.51301694e-01
-1.20263243e+00 -9.31743562e-01 1.24646351e-01 3.50825548e-01
-1.57734096e+00 -6.22140765e-01 5.58962584e-01 1.58947706e-01
1.61571145e+00 5.16365588e-01 3.14838737e-01 1.17290604e+00
4.01822805e-01 1.12983191e+00 1.11961293e+00 -7.34675765e-01
-3.53835285e-01 6.41098022e-01 2.89030224e-01 3.83150131e-01
-1.54427812e-01 -4.49823663e-02 -5.87565422e-01 -5.20639680e-02
1.76604465e-01 5.13093639e-03 -3.19379985e-01 1.74861580e-01
-1.21698773e+00 7.41438389e-01 6.02560580e-01 7.01102257e-01
5.65129565e-03 -1.34783462e-01 7.86343873e-01 1.00079596e+00
5.43175399e-01 4.16120380e-01 -4.41434801e-01 3.14793020e-01
-9.97240424e-01 7.94720724e-02 5.50110221e-01 1.10419571e+00
7.53831565e-01 -3.30129743e-01 -5.06756663e-01 1.17750728e+00
1.27235040e-01 1.01437378e+00 9.11809742e-01 -3.11875373e-01
1.03443730e+00 2.85205990e-01 -3.52635711e-01 -1.00295746e+00
-1.22669429e-01 -4.01544750e-01 -9.81257856e-01 -7.88058996e-01
-1.45772010e-01 2.32113704e-01 -6.60884798e-01 1.86212552e+00
-2.52348721e-01 -6.88153058e-02 4.69223887e-01 8.09670568e-01
9.74717438e-01 9.70719099e-01 2.68300205e-01 -4.09558475e-01
1.32324636e+00 -1.14934921e+00 -1.02517295e+00 -5.05226552e-01
1.10777676e+00 -1.26950657e+00 1.51464319e+00 -1.82503253e-01
-1.16487050e+00 -4.22632337e-01 -9.32684064e-01 -6.82975650e-01
-6.72548294e-01 -2.24219915e-02 5.66183269e-01 2.47153252e-01
-1.09543502e+00 2.13615268e-01 -5.00343144e-01 -5.05379319e-01
1.95185952e-02 2.17873380e-01 -5.61301887e-01 -5.35499752e-01
-1.81941402e+00 1.22881114e+00 3.60102594e-01 8.79610777e-02
-6.11275494e-01 -5.97415090e-01 -9.09145594e-01 1.77209750e-01
9.69767720e-02 -5.59769869e-01 7.80978501e-01 -6.81043625e-01
-1.11241674e+00 1.00463617e+00 -4.02441740e-01 -4.99526024e-01
6.56270310e-02 -3.93986613e-01 -5.42998552e-01 2.72839330e-02
2.58906037e-01 9.57823157e-01 5.12240648e-01 -1.00251257e+00
-1.63187072e-01 -2.19587728e-01 -1.44629210e-01 6.91806436e-01
-7.15143144e-01 4.12209421e-01 -6.10149324e-01 -7.47829020e-01
7.36777112e-02 -6.23968780e-01 1.82644669e-02 -6.03400767e-01
-4.10633594e-01 -3.88178885e-01 5.21799266e-01 -9.63319778e-01
1.62877357e+00 -2.07515836e+00 2.20422104e-01 1.64317414e-02
-1.93304390e-01 2.44054273e-01 -7.09793568e-01 7.64072835e-01
1.83902346e-02 3.29676241e-01 -2.01438278e-01 -6.01058781e-01
1.06441677e-02 1.10134199e-01 -8.08000505e-01 5.79725169e-02
1.79191753e-01 1.24388540e+00 -8.39681089e-01 -7.70022452e-01
-1.66020021e-01 3.06443810e-01 -6.14928365e-01 1.97845310e-01
2.82873124e-01 5.76426871e-02 -7.58094117e-02 5.56726515e-01
6.97927773e-01 1.92169875e-01 2.90246785e-01 -1.38359636e-01
3.65312338e-01 6.68625653e-01 -6.94818079e-01 2.22127652e+00
-8.43317807e-01 6.06334805e-01 -4.96633984e-02 -8.68837714e-01
6.86902583e-01 3.55422139e-01 -1.97703287e-01 -1.27365661e+00
2.02108081e-02 2.92298734e-01 -7.90678337e-02 -5.38460910e-01
1.01117742e+00 -3.93914551e-01 -4.50068384e-01 5.88804841e-01
3.07335645e-01 -1.56677455e-01 1.38093960e-02 7.43818164e-01
8.52394581e-01 -3.77437264e-01 -2.08491413e-03 -5.00667870e-01
7.33550668e-01 -3.05172235e-01 4.72721606e-01 1.00080693e+00
-6.56997859e-02 4.22626406e-01 4.59866486e-02 5.84343374e-02
-9.65902090e-01 -1.14846420e+00 -1.62668630e-01 1.33654666e+00
1.90506443e-01 -5.71729779e-01 -4.33549225e-01 -6.65870011e-01
-1.81367531e-01 7.66283274e-01 -1.10130012e-01 -6.41484976e-01
-8.85702670e-01 -8.28552485e-01 6.32796645e-01 2.33304560e-01
5.36941826e-01 -1.23460054e+00 3.25938135e-01 1.12690195e-01
-5.96387863e-01 -1.07705045e+00 -9.89162803e-01 2.00812310e-01
-9.27206576e-01 -4.46109623e-01 -5.44619679e-01 -1.23028505e+00
6.70691907e-01 4.58176047e-01 1.39779186e+00 1.33486256e-01
-2.09074557e-01 2.21565172e-01 -2.73963749e-01 -8.49670470e-02
-2.70003915e-01 4.15540397e-01 2.37046599e-01 -5.07377207e-01
7.42075026e-01 -4.26685542e-01 -5.43564916e-01 -1.22759409e-01
-1.01982248e+00 -1.54660985e-01 6.59139812e-01 1.15224063e+00
5.03612399e-01 -3.42568696e-01 7.42833197e-01 -8.39811087e-01
1.08616650e+00 -5.89350224e-01 -1.80347636e-01 7.58389175e-01
-5.54978609e-01 3.04707944e-01 6.05049431e-01 -4.36128706e-01
-1.05209732e+00 -6.85208797e-01 -1.51665658e-01 -2.00699568e-01
3.44972998e-01 5.96405864e-01 -1.73713148e-01 2.76284516e-01
5.69001198e-01 4.66715008e-01 -2.56175935e-01 -6.60126567e-01
4.92243201e-01 8.21619809e-01 1.15646750e-01 -5.48335016e-01
6.48897290e-01 4.96176556e-02 -5.89362681e-01 -6.07563615e-01
-6.87569439e-01 -4.94997054e-01 -5.39720893e-01 2.52657354e-01
7.01818764e-01 -1.21743178e+00 -1.32399425e-01 1.25321731e-01
-1.22083342e+00 -5.51298186e-02 -4.54848967e-02 5.38167596e-01
1.32317454e-01 4.54682678e-01 -1.01451194e+00 -5.19714594e-01
-9.25484836e-01 -1.15493464e+00 1.05024934e+00 -9.30150822e-02
2.86420733e-02 -1.06094480e+00 1.93722695e-01 2.74733782e-01
6.58468068e-01 -7.79679358e-01 1.14062381e+00 -8.76284420e-01
-5.42228639e-01 -2.08941370e-01 -1.28966451e-01 5.77777386e-01
-5.77097014e-02 -5.32736659e-01 -8.43742073e-01 -8.53252649e-01
6.95907995e-02 -6.99678302e-01 1.19328177e+00 -5.11040390e-02
8.44112635e-01 -2.68066764e-01 -2.84099430e-01 3.19685042e-01
1.46024406e+00 -9.34890509e-02 6.08512282e-01 8.69761184e-02
5.13770342e-01 5.66064239e-01 5.93083680e-01 -4.67836559e-01
4.80073810e-01 7.02775717e-01 -2.98059493e-01 -1.39039010e-01
-3.43808234e-01 -4.28815007e-01 8.06131661e-01 2.00589132e+00
6.38827980e-01 -1.64941445e-01 -6.77482247e-01 6.48515224e-01
-1.55596888e+00 -9.44983840e-01 1.71700358e-01 2.23182988e+00
1.40974295e+00 7.52525181e-02 -6.50529981e-01 -4.50325817e-01
4.42964762e-01 2.12696150e-01 -1.48931906e-01 -5.83747327e-01
-6.30101800e-01 4.49581265e-01 2.44774282e-01 1.07456112e+00
-8.47917497e-01 1.54762661e+00 6.12269402e+00 1.23433626e+00
-1.13975966e+00 4.52920794e-01 4.91885483e-01 -4.45599347e-01
-7.14187682e-01 8.79970789e-02 -9.57932353e-01 3.12274665e-01
9.73613501e-01 -2.27086768e-01 4.04247344e-01 3.96865845e-01
-1.40728027e-01 -1.79140925e-01 -1.05852401e+00 1.06523204e+00
5.20646155e-01 -1.13336813e+00 4.74215776e-01 -3.06330532e-01
6.31291986e-01 3.60163182e-01 7.37194344e-02 1.01790535e+00
-3.02487276e-02 -9.13032830e-01 1.97397694e-01 5.62876344e-01
6.95987880e-01 -7.77969897e-01 9.90133464e-01 4.76648271e-01
-8.95054162e-01 1.85091347e-01 -8.20007205e-01 1.65560201e-01
1.50005788e-01 4.12161320e-01 -7.45909572e-01 5.74476540e-01
6.39360189e-01 5.36796749e-01 -9.40775573e-01 5.50264001e-01
-1.89107969e-01 4.66074109e-01 -3.00574571e-01 -3.23792882e-02
1.43318981e-01 -1.86072513e-01 5.67646503e-01 1.65582776e+00
1.38421610e-01 -1.27213359e-01 5.59984557e-02 5.40407240e-01
-4.96315777e-01 6.41357422e-01 -9.59443867e-01 1.55075431e-01
6.37389064e-01 1.06297624e+00 -1.46031260e-01 -6.42544270e-01
-3.17556322e-01 1.29674244e+00 7.77707398e-01 5.71426570e-01
-4.66527730e-01 -5.52567244e-01 4.41367179e-01 -3.21777642e-01
-1.22112177e-01 -7.79509991e-02 -5.49195036e-02 -1.80646157e+00
2.69231528e-01 -9.78875339e-01 3.36812824e-01 -6.43067420e-01
-1.52220809e+00 7.54231811e-01 -3.62692177e-02 -1.11335659e+00
-1.49517208e-01 -6.09592013e-02 -4.12573338e-01 1.24869668e+00
-1.74302971e+00 -1.10755754e+00 4.98235762e-01 6.87307537e-01
6.77659035e-01 -4.64535862e-01 1.10012245e+00 7.63605356e-01
-5.39806604e-01 1.05354369e+00 2.48666018e-01 2.14086562e-01
1.37494922e+00 -1.09287322e+00 1.27046257e-02 9.68960226e-01
4.80209380e-01 1.21540999e+00 2.91997671e-01 -5.65664709e-01
-1.39263070e+00 -9.11073089e-01 1.73080814e+00 -5.60868621e-01
6.48548841e-01 -7.97672451e-01 -1.22901475e+00 7.12084889e-01
8.87562215e-01 -3.43647450e-01 8.70808303e-01 3.84417593e-01
-4.68433112e-01 -3.89250457e-01 -8.38493884e-01 8.72750223e-01
8.94089520e-01 -1.36439526e+00 -1.15352237e+00 6.16335332e-01
1.11467016e+00 -1.29241601e-01 -8.35892081e-01 4.40074533e-01
1.07294127e-01 -4.99257445e-01 1.10642636e+00 -8.90326440e-01
3.80660951e-01 5.31928730e-04 -3.31680179e-01 -1.26868010e+00
-1.50552720e-01 -2.44074687e-01 1.73095107e-01 1.37834632e+00
5.51728189e-01 -5.92430651e-01 3.68946940e-02 2.76463062e-01
-4.96219285e-02 -5.65012038e-01 -1.11294842e+00 -6.78062975e-01
6.10177755e-01 -3.26803446e-01 3.34235668e-01 1.25463271e+00
1.93639606e-01 9.99123991e-01 -3.46576184e-01 -2.51215789e-02
3.43828022e-01 6.09301478e-02 4.44982320e-01 -4.21511233e-01
-9.33439881e-02 -3.39279890e-01 7.12467283e-02 -1.27188706e+00
5.56240737e-01 -1.76483893e+00 -1.26688197e-01 -1.24349689e+00
7.16098666e-01 -3.75352323e-01 -6.48104250e-01 4.12388772e-01
-6.16430342e-01 1.90518871e-01 3.69809061e-01 2.71557093e-01
-7.55243003e-01 8.34846199e-01 1.18781829e+00 -3.56615394e-01
-6.92042559e-02 -5.73812783e-01 -5.39982736e-01 1.08528636e-01
5.67666173e-01 -6.65567517e-01 -4.41796958e-01 -9.60063756e-01
3.79937649e-01 8.47091824e-02 -9.49195586e-03 -4.70969021e-01
3.20886135e-01 3.59036475e-01 2.15802297e-01 -7.31717169e-01
2.64123857e-01 -8.43177736e-01 -4.11416769e-01 4.13706332e-01
-6.90657973e-01 3.50800306e-01 4.41719592e-01 3.72734606e-01
-6.68103456e-01 -3.72083992e-01 3.11338186e-01 -2.27359340e-01
-5.82826376e-01 7.65018761e-02 -2.75014609e-01 2.53677189e-01
2.40976691e-01 5.06488204e-01 -4.26654488e-01 -3.33692163e-01
-7.08727896e-01 6.44230485e-01 8.52146931e-03 7.88160682e-01
7.18604326e-01 -1.65098941e+00 -8.87326956e-01 2.77965754e-01
1.89522415e-01 -2.57146597e-01 3.10893267e-01 8.75017762e-01
-1.55645803e-01 7.89013326e-01 1.66008189e-01 -5.03950119e-01
-1.32632780e+00 5.04183650e-01 1.79124787e-01 -7.99862146e-01
-3.27049345e-01 1.00920200e+00 1.85186088e-01 -1.03982949e+00
2.36344233e-01 -1.76479504e-01 -7.43240416e-02 1.82859749e-01
6.14866316e-01 -2.50339508e-01 2.53145248e-01 -4.99781102e-01
-3.77656162e-01 4.47017878e-01 -8.11398625e-01 -2.34834656e-01
1.14356494e+00 -3.60115141e-01 -6.99979722e-01 4.68207240e-01
1.80917668e+00 2.94596165e-01 -2.40772098e-01 -7.41753876e-01
1.28205165e-01 -2.92745769e-01 1.10086627e-01 -9.42050576e-01
-7.71874666e-01 1.13406920e+00 7.22784877e-01 -2.82541215e-01
1.18219280e+00 6.52379468e-02 1.12144983e+00 8.67912710e-01
4.58853215e-01 -1.27840340e+00 6.75245672e-02 9.02049184e-01
1.15065014e+00 -1.56173658e+00 -3.69551070e-02 9.12602898e-03
-6.52373493e-01 6.70574665e-01 5.76239169e-01 -7.66935665e-03
5.73643446e-01 -1.06911054e-02 2.41888493e-01 3.59917209e-02
-8.64137888e-01 4.44670022e-02 4.47451800e-01 4.12947945e-02
5.66955984e-01 1.28447516e-02 -7.39813685e-01 5.35398245e-01
-2.76860714e-01 -5.07552803e-01 4.22168039e-02 9.52107668e-01
-2.19385639e-01 -1.46515858e+00 -1.45852804e-01 3.74612510e-01
-4.75396097e-01 -1.03640795e+00 -3.40944976e-01 6.71074569e-01
-4.34366792e-01 7.04623878e-01 1.28244683e-01 -2.44024947e-01
3.04343939e-01 5.17796636e-01 3.95767361e-01 -7.14765549e-01
-9.59897399e-01 1.98391259e-01 8.15023035e-02 -4.93050337e-01
-1.72686085e-01 -3.89001250e-01 -1.04137218e+00 -2.18262464e-01
-4.17104036e-01 2.93229789e-01 4.51245040e-01 9.20877755e-01
4.51289982e-01 3.07704329e-01 6.60858333e-01 -2.33287930e-01
-5.90874910e-01 -1.46770763e+00 -3.42953712e-01 5.57975829e-01
1.98680669e-01 -5.45229129e-02 -4.43438172e-01 -3.90167743e-01] | [11.350184440612793, 9.793774604797363] |
2d772dca-c2db-42d9-91c4-750d3c263cbd | counting-the-uncountable-deep-semantic | 1809.07091 | null | http://arxiv.org/abs/1809.07091v2 | http://arxiv.org/pdf/1809.07091v2.pdf | Counting the uncountable: deep semantic density estimation from Space | We propose a new method to count objects of specific categories that are
significantly smaller than the ground sampling distance of a satellite image.
This task is hard due to the cluttered nature of scenes where different object
categories occur. Target objects can be partially occluded, vary in appearance
within the same class and look alike to different categories. Since traditional
object detection is infeasible due to the small size of objects with respect to
the pixel size, we cast object counting as a density estimation problem. To
distinguish objects of different classes, our approach combines density
estimation with semantic segmentation in an end-to-end learnable convolutional
neural network (CNN). Experiments show that deep semantic density estimation
can robustly count objects of various classes in cluttered scenes. Experiments
also suggest that we need specific CNN architectures in remote sensing instead
of blindly applying existing ones from computer vision. | ['Jan D. Wegner', 'Andres C. Rodriguez'] | 2018-09-19 | null | null | null | null | ['object-counting'] | ['computer-vision'] | [ 1.13943480e-01 -3.10729116e-01 3.11052442e-01 -4.70534295e-01
-4.98266786e-01 -7.29502857e-01 3.40197921e-01 9.09911876e-04
-7.57894278e-01 8.52407277e-01 -3.36826921e-01 -2.18916371e-01
-1.55450776e-02 -1.19210255e+00 -7.84585118e-01 -7.42024481e-01
-1.36563763e-01 1.01459634e+00 5.50656915e-01 6.44623876e-01
-3.64204124e-02 6.48523748e-01 -1.57178164e+00 4.68151160e-02
8.90677392e-01 1.01754069e+00 6.02500558e-01 9.43346977e-01
-4.71550256e-01 6.62603915e-01 -8.30304682e-01 -5.53974435e-02
2.76041865e-01 -3.65568250e-01 -7.93320000e-01 3.92790079e-01
9.33583617e-01 -8.28058064e-01 1.56190395e-02 1.82266831e+00
3.82759608e-02 1.25746667e-01 1.14663827e+00 -1.04635835e+00
-6.59229398e-01 2.73887396e-01 -8.42678666e-01 4.10806507e-01
-1.68576181e-01 -2.17560619e-01 6.66578948e-01 -7.15155661e-01
-2.64846440e-02 1.40020895e+00 6.63553596e-01 4.97537911e-01
-1.11715353e+00 -4.98888344e-01 3.91424477e-01 1.50954220e-02
-1.68041849e+00 5.37182530e-03 1.45146564e-01 -6.35491014e-01
5.89052916e-01 2.83875644e-01 6.02457345e-01 5.25954306e-01
-3.43460619e-01 6.10265553e-01 1.04966736e+00 -2.01687608e-02
7.00758815e-01 1.46807507e-01 8.68807659e-02 4.76834834e-01
8.35335314e-01 -3.48770857e-01 2.11382344e-01 -1.72533432e-03
9.24965918e-01 1.92874774e-01 -9.61999521e-02 -3.60269815e-01
-1.13726854e+00 8.76708150e-01 6.64811790e-01 1.52667630e-02
-1.13826729e-01 3.25611293e-01 -8.89489725e-02 -2.16725379e-01
5.64640343e-01 -9.65509191e-03 -4.92248803e-01 2.03622848e-01
-1.31765532e+00 2.24414393e-01 9.00250256e-01 8.46287966e-01
1.02809799e+00 -6.93657398e-02 -1.03703938e-01 7.97736824e-01
1.98467791e-01 9.66692328e-01 -2.88503543e-02 -9.32708383e-01
2.01700553e-01 5.38613021e-01 3.65352958e-01 -6.73762381e-01
-1.36169970e-01 -2.26060167e-01 -8.87010932e-01 6.19550347e-01
9.49461639e-01 -1.77193612e-01 -1.30583167e+00 1.31936991e+00
4.19679582e-01 1.17093816e-01 -2.45037466e-01 1.15442026e+00
5.92458546e-01 7.63163388e-01 2.30962321e-01 1.31955355e-01
1.58115590e+00 -4.78256136e-01 -1.41879842e-01 -7.29272902e-01
2.53996313e-01 -4.05011743e-01 9.36445951e-01 3.51183712e-01
-6.43648684e-01 -2.14717478e-01 -7.44809628e-01 1.49654582e-01
-4.28856701e-01 2.73654848e-01 7.59719372e-01 8.07734251e-01
-7.20638692e-01 3.98774713e-01 -1.00438607e+00 -4.81857359e-01
1.03452086e+00 2.10221708e-01 -9.37149748e-02 -1.52983323e-01
-3.97010028e-01 6.07894063e-01 6.19466484e-01 8.28152299e-02
-1.14026046e+00 -4.85502034e-01 -7.74584234e-01 2.89618254e-01
2.35074297e-01 -6.57143474e-01 9.96886492e-01 -1.17219508e+00
-6.46956027e-01 9.09056246e-01 -1.04116417e-01 -3.59816968e-01
5.67494750e-01 -2.57671446e-01 5.23705371e-02 1.44806892e-01
2.66460985e-01 6.99244618e-01 8.59118998e-01 -1.25199187e+00
-9.98913765e-01 -4.86556649e-01 1.25999615e-01 -2.06394261e-03
-1.22524396e-01 -1.94357336e-02 -2.29923408e-02 -3.13687623e-01
2.57485390e-01 -7.43080914e-01 -2.38917008e-01 5.54302037e-01
-3.97120118e-01 -3.70086953e-02 1.18383086e+00 -4.76235211e-01
5.18394291e-01 -1.94468403e+00 -9.70068797e-02 -1.48234308e-01
2.88848579e-01 8.41597095e-02 7.44104804e-03 -2.72480488e-01
3.76974136e-01 1.11025181e-02 -6.76350892e-01 -5.34564592e-02
-1.53751075e-01 4.23482150e-01 -2.08635166e-01 7.06492424e-01
3.96870464e-01 4.62324888e-01 -1.03128433e+00 -6.37007415e-01
3.85222137e-01 6.42213821e-01 -2.19371438e-01 2.59602547e-01
-4.70138907e-01 4.24160600e-01 -2.97686189e-01 1.11527812e+00
1.40066898e+00 -3.17863971e-01 -1.75633952e-01 -2.32487507e-02
6.50807917e-02 2.73875371e-02 -1.55923951e+00 1.06174135e+00
-3.70449573e-01 5.69620669e-01 2.64398575e-01 -1.02962875e+00
6.37465417e-01 -3.99008304e-01 -1.06054820e-01 -1.15336478e-01
2.40622595e-01 2.63997495e-01 -5.62595539e-02 -1.94297254e-01
5.88115215e-01 -2.87660986e-01 7.31904209e-02 1.77794203e-01
1.53137922e-01 -5.08196592e-01 3.62010300e-01 3.26352455e-02
8.83704305e-01 -1.52426690e-01 2.73988694e-01 -2.04519585e-01
1.96403526e-02 7.06214532e-02 4.44646418e-01 1.09536922e+00
-3.81469220e-01 9.03785884e-01 4.06873912e-01 -3.67519051e-01
-1.02375901e+00 -1.54673314e+00 -5.90879142e-01 1.08508503e+00
3.07619721e-01 4.16667998e-01 -7.98143268e-01 -6.95106447e-01
7.02379867e-02 3.62220407e-01 -5.87856174e-01 3.92863274e-01
-1.89245135e-01 -1.29461312e+00 2.46130511e-01 7.31773019e-01
7.51736283e-01 -7.80051708e-01 -6.61217928e-01 1.12311533e-02
-1.44526914e-01 -1.34303713e+00 -2.48736396e-01 4.18909758e-01
-1.00561070e+00 -1.17962670e+00 -1.04539073e+00 -7.26103663e-01
7.55952895e-01 6.08076930e-01 1.22082460e+00 -1.20823242e-01
-6.29997075e-01 3.14315885e-01 -2.65138119e-01 -6.26416981e-01
-1.94619209e-01 -1.13174170e-01 -1.36761591e-01 -5.02221175e-02
7.85977423e-01 -5.19286454e-01 -7.74345517e-01 1.75274923e-01
-8.57766926e-01 -4.20605391e-01 5.45106351e-01 5.47456086e-01
5.17204404e-01 4.68049169e-01 6.56813905e-02 -6.81165636e-01
2.25985255e-02 -6.28684044e-01 -1.07222021e+00 2.53456563e-01
-1.18359933e-02 -1.53629258e-01 1.65026098e-01 -6.41118348e-01
-8.19236755e-01 3.43113005e-01 3.89596701e-01 -2.46801078e-01
-4.71566141e-01 1.68145299e-01 -2.71533817e-01 4.55738753e-02
5.05559087e-01 1.56547531e-01 -1.90525100e-01 -3.97756606e-01
2.75849164e-01 6.82945430e-01 5.51428556e-01 -3.28386903e-01
8.63664687e-01 8.93864453e-01 -1.13927417e-01 -9.40630913e-01
-1.10768378e+00 -6.58943951e-01 -7.05218434e-01 -1.43473789e-01
9.97825623e-01 -1.35041916e+00 -5.50055981e-01 6.66825056e-01
-1.10518265e+00 -3.22994798e-01 -2.32584700e-01 7.05822766e-01
-1.99944258e-01 4.79363471e-01 -4.74016607e-01 -1.21145058e+00
-1.61759317e-01 -7.00083017e-01 1.29639137e+00 6.45198345e-01
2.22251549e-01 -8.82047057e-01 -2.53495932e-01 9.08122584e-02
3.13035458e-01 4.60724756e-02 5.72486579e-01 -4.42175120e-01
-9.09193516e-01 -2.48648494e-01 -9.30369914e-01 6.45863235e-01
1.74377218e-01 1.54749408e-01 -1.04911494e+00 -1.47520408e-01
-1.75126612e-01 -4.07854766e-01 1.32550728e+00 8.66587937e-01
1.43786621e+00 -2.07566649e-01 -3.95152867e-01 7.36718476e-01
1.63372874e+00 -2.10472688e-01 7.67244399e-01 2.18437776e-01
8.27356398e-01 6.11629367e-01 4.71963227e-01 5.48667312e-01
1.50292739e-01 5.42628244e-02 8.00215662e-01 -1.42612875e-01
-1.69702962e-01 1.58439353e-01 4.58246469e-02 -3.69143635e-02
-1.20108709e-01 -1.38654187e-01 -8.41239274e-01 8.29260945e-01
-1.48287261e+00 -1.19187605e+00 -4.19344455e-01 2.27334452e+00
6.24518991e-01 -1.13143876e-01 2.21099317e-01 -9.07440037e-02
1.00308692e+00 -6.81411102e-02 -5.18778563e-01 1.20649293e-01
-2.31402099e-01 1.55507609e-01 1.02206004e+00 4.64228988e-01
-1.51747406e+00 7.26342797e-01 6.24876547e+00 8.83654594e-01
-8.30434978e-01 -6.41848072e-02 8.79576564e-01 -2.65432410e-02
1.88887835e-01 -7.26729212e-03 -9.59108770e-01 5.93366265e-01
4.44068551e-01 2.39623800e-01 2.11965173e-01 1.21584928e+00
-1.44037336e-01 -7.81961977e-01 -8.39090347e-01 1.03082180e+00
-2.17102543e-01 -8.57675910e-01 9.21542868e-02 8.33425671e-02
5.21415830e-01 2.50487506e-01 -7.54074007e-03 1.56091392e-01
6.81231499e-01 -1.27368677e+00 7.37971485e-01 3.19998413e-01
6.72263086e-01 -5.59743643e-01 7.14255691e-01 5.16375244e-01
-1.24685419e+00 -5.98880323e-03 -1.20263374e+00 -3.82763863e-01
4.20341901e-02 1.15019619e+00 -7.69074380e-01 -1.91455215e-01
1.07121658e+00 3.59385908e-01 -5.68188310e-01 1.50014234e+00
-8.06433633e-02 5.06084800e-01 -5.94811738e-01 -8.35125521e-02
2.03570962e-01 -4.35974181e-01 3.40722084e-01 1.21847880e+00
5.65493107e-01 -1.34772003e-01 1.53769881e-01 1.42099464e+00
-2.56760959e-02 -4.73294884e-01 -4.29078937e-01 -1.16058372e-01
5.46038330e-01 1.48182881e+00 -1.27373433e+00 -4.32712555e-01
-4.28689927e-01 1.13396442e+00 1.88968658e-01 2.91397870e-01
-8.91965449e-01 -3.17990124e-01 5.98520935e-01 1.98900670e-01
5.93357861e-01 -3.29726577e-01 -1.57625467e-01 -1.34895742e+00
-2.95932796e-02 -2.36111537e-01 2.89880157e-01 -7.84839749e-01
-1.73203838e+00 4.03159827e-01 1.61414355e-01 -1.28996158e+00
2.96745598e-01 -9.41415131e-01 -5.04269481e-01 9.70577896e-01
-1.51693130e+00 -1.09308362e+00 -7.72340834e-01 5.03948510e-01
4.04211879e-01 1.37760624e-01 8.06839645e-01 2.84433037e-01
-1.93377241e-01 4.67018504e-03 3.69488567e-01 5.47837853e-01
2.15087622e-01 -1.53606784e+00 1.29262373e-01 9.70661104e-01
2.22574070e-01 1.33956566e-01 4.82291371e-01 -5.12871146e-01
-6.96654558e-01 -1.42822576e+00 4.29939419e-01 -4.83123064e-01
7.21660316e-01 -5.38409233e-01 -1.03420126e+00 3.88150305e-01
-3.68779391e-01 6.38376951e-01 4.66652274e-01 -2.75817085e-02
-4.51095581e-01 -9.46925604e-04 -1.22933054e+00 1.35872841e-01
7.45375216e-01 -4.52759862e-01 -3.85816962e-01 5.36738753e-01
3.77785414e-01 -2.08711013e-01 -2.85679013e-01 2.54479766e-01
3.52010965e-01 -1.00993657e+00 1.12882531e+00 -2.43170753e-01
2.83707768e-01 -4.89372432e-01 -3.46839577e-01 -1.04691148e+00
-1.37668386e-01 3.94403666e-01 3.22754495e-02 1.12540162e+00
1.23374850e-01 -2.81042457e-01 1.04131877e+00 6.08744383e-01
2.64001280e-01 1.59151256e-01 -1.02707052e+00 -1.20260644e+00
3.11864793e-01 -1.77156150e-01 4.48852450e-01 9.07433033e-01
-7.21254051e-01 4.67759511e-03 -6.81394711e-02 7.43538320e-01
8.72981191e-01 2.00188875e-01 7.37434566e-01 -1.71478415e+00
-3.30287665e-01 -2.81546235e-01 -6.43581033e-01 -1.00582063e+00
-7.06801489e-02 -5.22850931e-01 4.63632643e-01 -1.82033265e+00
7.29061544e-01 -5.46884656e-01 1.19034171e-01 2.19132036e-01
-3.52042913e-01 5.06591439e-01 -3.46377864e-02 2.59615093e-01
-4.99056458e-01 3.05212200e-01 9.30226028e-01 -6.37172580e-01
4.13845554e-02 1.58497334e-01 -3.22259754e-01 9.15809333e-01
6.84065700e-01 -7.94775188e-01 -1.83389276e-01 -6.93404257e-01
1.41888335e-01 -2.15896025e-01 9.76449072e-01 -1.36096299e+00
-5.70677035e-02 -4.29519832e-01 7.59382546e-01 -7.43636072e-01
2.67825544e-01 -9.23342705e-01 -1.78893030e-01 4.61619288e-01
3.40896875e-01 -7.78916419e-01 2.74683148e-01 7.38591313e-01
-8.30679312e-02 -5.29478788e-01 1.11333263e+00 -5.55810809e-01
-7.86966145e-01 3.87229383e-01 -5.06587327e-01 6.28798679e-02
8.08766782e-01 -4.93406415e-01 -3.64671558e-01 -1.06808327e-01
-7.09357858e-01 -1.95517346e-01 3.97304267e-01 -1.36243865e-01
3.69909227e-01 -9.01261210e-01 -8.31198454e-01 -2.63622463e-01
1.93579882e-01 5.30405581e-01 9.12813172e-02 5.73620081e-01
-9.03057694e-01 -7.25331157e-02 -1.36461183e-01 -9.96472001e-01
-1.13048303e+00 3.53447646e-01 7.27646232e-01 2.12594017e-01
-4.53923702e-01 1.27984238e+00 6.89907968e-01 -4.84465122e-01
2.14819536e-01 -6.26649141e-01 -3.95293720e-02 1.47472480e-02
8.15030098e-01 3.06768090e-01 -1.33271381e-01 -3.19337249e-01
-5.28005421e-01 4.86201555e-01 2.58334307e-03 1.76982313e-01
1.27876878e+00 -1.64441586e-01 -3.25285971e-01 4.80944544e-01
8.73005033e-01 -2.14642510e-01 -1.58537745e+00 -2.33613491e-01
-1.30664036e-01 -9.14993167e-01 1.39327481e-01 -7.18038440e-01
-1.11973548e+00 1.00499690e+00 8.92435968e-01 2.12690845e-01
1.05508876e+00 3.99222881e-01 3.19026351e-01 5.39752305e-01
4.14140940e-01 -1.04801130e+00 -2.28081737e-02 6.70616925e-01
3.07444185e-01 -1.72097778e+00 2.29577720e-01 -4.26706016e-01
-2.68071204e-01 1.20514226e+00 5.98502815e-01 -2.67917573e-01
6.77324891e-01 3.89838189e-01 -1.45024329e-01 -1.62146270e-01
2.94574872e-02 -8.09437990e-01 2.34902397e-01 1.11143744e+00
2.83752948e-01 2.36978367e-01 1.15419924e-01 2.50617325e-01
1.37374282e-01 -1.68502405e-01 8.16885352e-01 7.92568862e-01
-1.04962933e+00 -4.95696127e-01 -7.54174292e-01 6.67626441e-01
-3.71938705e-01 -3.49294782e-01 -2.55071223e-01 7.61892855e-01
2.80732781e-01 8.54039252e-01 7.01930583e-01 1.96459293e-01
-1.10564627e-01 -2.93484628e-01 6.68816507e-01 -7.58632541e-01
1.56670690e-01 -2.63974164e-02 -2.21681461e-01 -1.04148574e-01
-6.85774028e-01 -5.71230352e-01 -1.17628539e+00 -4.44825917e-01
-4.09460098e-01 -2.87862033e-01 6.67930424e-01 8.27892780e-01
-3.02625656e-01 3.04100096e-01 1.65621072e-01 -8.98243248e-01
-4.11302090e-01 -1.18333864e+00 -1.30127168e+00 1.79918125e-01
3.72693241e-01 -7.65953541e-01 -7.41422832e-01 2.49214441e-01] | [8.836224555969238, -0.2695405185222626] |
e8d9eb93-a16a-4441-a752-09c03850500e | data-driven-bilateral-generalized-two | 2306.07045 | null | https://arxiv.org/abs/2306.07045v1 | https://arxiv.org/pdf/2306.07045v1.pdf | Data-Driven Bilateral Generalized Two-Dimensional Quaternion Principal Component Analysis with Application to Color Face Recognition | A new data-driven bilateral generalized two-dimensional quaternion principal component analysis (BiG2DQPCA) is presented to extract the features of matrix samples from both row and column directions. This general framework directly works on the 2D color images without vectorizing and well preserves the spatial and color information, which makes it flexible to fit various real-world applications. A generalized ridge regression model of BiG2DQPCA is firstly proposed with orthogonality constrains on aimed features. Applying the deflation technique and the framework of minorization-maximization, a new quaternion optimization algorithm is proposed to compute the optimal features of BiG2DQPCA and a closed-form solution is obtained at each iteration. A new approach based on BiG2DQPCA is presented for color face recognition and image reconstruction with a new data-driven weighting technique. Sufficient numerical experiments are implemented on practical color face databases and indicate the superiority of BiG2DQPCA over the state-of-the-art methods in terms of recognition accuracies and rates of image reconstruction. | ['Yong Zhang', 'Dun-Wei Gong', 'Zhi-Gang Jia', 'Mei-Xiang Zhao'] | 2023-06-12 | null | null | null | null | ['face-recognition', 'image-reconstruction'] | ['computer-vision', 'computer-vision'] | [-2.71980733e-01 -5.10235429e-01 -9.12932083e-02 -2.50323892e-01
-2.98022002e-01 3.43516842e-02 4.18535233e-01 -5.79413414e-01
-5.96657634e-01 3.70820194e-01 -2.76035309e-01 -6.86641112e-02
-1.09613098e-01 -5.77418387e-01 -2.17366070e-01 -1.03841913e+00
4.96203713e-02 1.59624472e-01 -3.74037355e-01 -2.28945494e-01
4.45617318e-01 1.01761186e+00 -1.26589894e+00 -3.90914977e-01
5.16266823e-01 1.01657927e+00 -6.07162341e-02 4.27772552e-01
-2.26559024e-02 1.44843340e-01 -2.42356926e-01 -7.24732280e-01
6.04744196e-01 -4.31313246e-01 -1.43630505e-01 6.19132936e-01
3.04561079e-01 -2.92508245e-01 -3.37743849e-01 1.15315354e+00
6.23718023e-01 -7.30016604e-02 6.35659933e-01 -1.42239511e+00
-8.96751940e-01 -3.05764437e-01 -1.18647122e+00 -6.88079968e-02
3.59278888e-01 -2.64930308e-01 7.97361434e-01 -1.50109279e+00
5.98151326e-01 1.47274840e+00 2.86628991e-01 2.98234373e-01
-1.38254035e+00 -7.23507345e-01 -1.28320411e-01 5.74983001e-01
-1.69535005e+00 -4.98054847e-02 1.04978836e+00 -3.81970316e-01
5.78990221e-01 4.59687203e-01 9.40389216e-01 5.25395989e-01
2.24309161e-01 7.24701643e-01 1.30420351e+00 -4.27220136e-01
4.66847755e-02 7.44307712e-02 3.10290605e-02 1.05505049e+00
4.47971016e-01 3.25880833e-02 -8.31903458e-01 -1.65111750e-01
1.04852557e+00 1.13831230e-01 -2.01020986e-01 -9.28906262e-01
-1.33780968e+00 8.78802598e-01 1.29542425e-01 8.22388679e-02
-6.23651922e-01 -1.96753681e-01 -3.96590866e-02 9.52891037e-02
2.85255313e-01 2.97106300e-02 -6.64989650e-02 -7.68603981e-02
-8.36943567e-01 -8.46180916e-02 7.31013775e-01 9.42996264e-01
1.07650506e+00 5.23062944e-01 2.80566722e-01 8.49298835e-01
7.33361185e-01 1.27364910e+00 5.41029632e-01 -8.50747347e-01
2.33795568e-01 6.82746232e-01 -7.12786987e-02 -1.62433219e+00
-3.84450465e-01 -1.42643869e-01 -1.11345983e+00 4.55997050e-01
1.86383933e-01 -1.11726644e-02 -6.98238194e-01 1.22876990e+00
5.13760686e-01 -3.12231965e-02 -1.89187210e-02 1.18692434e+00
5.51191688e-01 8.50396335e-01 -3.66998315e-01 -6.74021542e-01
1.30275762e+00 -5.22163212e-01 -9.74773586e-01 2.60021746e-01
-3.77647169e-02 -9.06798780e-01 6.98723555e-01 7.54960954e-01
-9.06596780e-01 -5.81964254e-01 -1.27858710e+00 3.66631709e-02
-2.08764210e-01 6.14536881e-01 6.25680327e-01 1.06727386e+00
-8.40931535e-01 3.17129999e-01 -9.06121373e-01 -8.17812458e-02
-1.59873590e-01 4.64258522e-01 -8.70204151e-01 -5.42447791e-02
-5.69224000e-01 7.60890603e-01 1.03514157e-01 7.58371115e-01
-3.19256485e-01 -2.70442158e-01 -8.52862239e-01 -4.04431254e-01
2.49979924e-02 -1.28883809e-01 5.05174518e-01 -7.48968363e-01
-1.97949517e+00 7.79800117e-01 -3.05263579e-01 4.07586724e-01
4.37592983e-01 -1.34087235e-01 -5.92777967e-01 4.23745573e-01
-9.21245068e-02 3.73071641e-01 1.53040910e+00 -1.00201726e+00
-4.28632647e-01 -7.11972654e-01 -6.01610184e-01 2.57947594e-01
-6.29180014e-01 -1.84941739e-01 -1.00629699e+00 -6.23894393e-01
9.20406461e-01 -9.87788260e-01 -3.44980061e-02 1.50107801e-01
-2.66360223e-01 -4.80871275e-03 7.79636979e-01 -8.52596998e-01
1.09344530e+00 -2.31606507e+00 8.01827133e-01 7.23494112e-01
-4.89442274e-02 5.89878149e-02 -3.28851163e-01 4.54206288e-01
-3.01870883e-01 -4.04078871e-01 -1.74909577e-01 -2.85933524e-01
4.42569293e-02 2.00319424e-01 -1.06941447e-01 9.94044363e-01
3.84897202e-01 2.78582126e-01 -5.38392007e-01 -4.16807324e-01
3.11542243e-01 8.58789861e-01 -4.59283054e-01 2.87676185e-01
7.80605078e-01 5.82583807e-02 -4.33386058e-01 8.42112362e-01
1.32780361e+00 2.01284274e-01 1.58294454e-01 -7.61591315e-01
-1.83785662e-01 -5.48751414e-01 -1.85463583e+00 1.27278733e+00
-2.06844747e-01 3.77878100e-01 3.05481762e-01 -8.66787016e-01
1.06213796e+00 1.53806061e-01 7.51975000e-01 -6.64411485e-01
5.79611473e-02 1.70160100e-01 -2.83049941e-01 -2.71047145e-01
3.16377878e-01 -1.33559972e-01 3.81893665e-01 3.70282382e-01
2.66898513e-01 -2.45481193e-01 3.80476981e-01 -4.13654074e-02
1.65804178e-01 2.58299410e-01 5.20001233e-01 -4.02900070e-01
1.02779448e+00 -3.30050468e-01 7.82534957e-01 -5.72887175e-02
-1.24033704e-01 6.53877139e-01 5.53155303e-01 -4.31055188e-01
-8.80500674e-01 -8.29388320e-01 -2.81294256e-01 6.65029526e-01
1.18185200e-01 -3.17256480e-01 -5.79136491e-01 -3.45292985e-01
2.13015884e-01 1.74655348e-01 -4.89447027e-01 1.97087899e-01
-3.95179003e-01 -1.22086620e+00 4.08785343e-02 9.62633863e-02
6.17705643e-01 -4.57467407e-01 -3.99322927e-01 7.38680661e-02
1.65751621e-01 -9.19614315e-01 -3.75927716e-01 -2.55107254e-01
-8.89819682e-01 -1.18701255e+00 -1.11654902e+00 -6.98587120e-01
1.05725217e+00 5.21076679e-01 5.00896871e-01 -1.47350237e-01
-4.71398443e-01 6.55036867e-01 -2.28601173e-01 -2.98290029e-02
2.70468537e-02 -5.77979326e-01 3.52365345e-01 7.50952542e-01
5.12794614e-01 -3.65320593e-01 -4.80927765e-01 3.97252113e-01
-7.61190414e-01 -5.56680895e-02 1.00940788e+00 1.15747619e+00
6.88815713e-01 -7.31305182e-02 1.08233087e-01 -4.05892491e-01
6.74049497e-01 1.85724087e-02 -1.05743992e+00 2.23749802e-01
-8.58092844e-01 1.88474163e-01 4.26010609e-01 -2.72099614e-01
-9.94989216e-01 3.36873829e-01 -4.82840240e-02 -3.85125428e-01
3.42298806e-01 4.38371450e-01 -3.81742924e-01 -5.56346238e-01
2.08783492e-01 5.55616796e-01 5.00676692e-01 -6.14800811e-01
4.96867299e-01 5.84972560e-01 2.81514078e-01 -3.08246285e-01
1.21478987e+00 5.94405293e-01 4.88161057e-01 -1.39938581e+00
9.33427662e-02 -5.76530278e-01 -9.33749378e-01 -1.76707581e-01
6.64319515e-01 -8.65848422e-01 -1.16274047e+00 8.66846621e-01
-8.23176444e-01 4.73620147e-01 6.09453134e-02 8.45945477e-01
-3.76937360e-01 9.37473178e-01 -6.41422570e-01 -8.07226658e-01
-2.64810652e-01 -1.31649506e+00 7.95902908e-01 1.90036759e-01
5.37670016e-01 -6.22276843e-01 -5.71909472e-02 6.30306602e-02
-4.24554534e-02 -1.87091812e-01 7.15336204e-01 -8.16288069e-02
-4.46764410e-01 -2.85597414e-01 -1.86930567e-01 6.33868575e-01
1.77583992e-01 4.07664061e-01 -5.27612090e-01 -5.71182013e-01
1.70365885e-01 -6.99986070e-02 4.38452303e-01 8.18196759e-02
6.36771262e-01 -2.35401556e-01 1.35031894e-01 8.73401225e-01
1.55217767e+00 1.97863981e-01 5.32547116e-01 3.78208190e-01
7.28935182e-01 4.07356411e-01 6.73382878e-01 7.20718324e-01
1.77559346e-01 7.02538252e-01 2.98561037e-01 -1.56701088e-01
2.39116848e-01 2.66686827e-02 4.46199477e-01 1.22895670e+00
-3.95893037e-01 5.60529232e-01 -4.71881896e-01 2.19432116e-01
-1.53937411e+00 -7.27188408e-01 -2.58887820e-02 2.36005163e+00
6.87892795e-01 -3.90629679e-01 -1.45442958e-03 3.93835098e-01
5.59448779e-01 1.30292758e-01 -4.05830085e-01 -3.22262973e-01
-2.53827155e-01 9.79070067e-02 4.34783340e-01 4.79232967e-01
-1.05197155e+00 6.15254819e-01 6.07445717e+00 8.55509639e-01
-1.38373005e+00 -2.18163610e-01 2.59965271e-01 4.95859116e-01
-8.95529687e-02 -2.18997553e-01 -8.31259251e-01 -1.14964498e-02
2.56790280e-01 -1.00585975e-01 6.17702305e-01 8.17425370e-01
-6.83154315e-02 -5.62683493e-03 -6.55169368e-01 1.67896605e+00
5.18712759e-01 -7.53352344e-01 4.10018146e-01 1.83537617e-01
5.58204114e-01 -5.85813463e-01 3.89160097e-01 -1.98491693e-01
-3.52694601e-01 -5.89384317e-01 5.09136319e-01 5.63588202e-01
7.33351111e-01 -8.14168155e-01 4.87041980e-01 -1.12187788e-01
-1.01354933e+00 1.84963725e-03 -7.44765580e-01 1.74615949e-01
-9.15778503e-02 4.63548362e-01 -7.33868539e-01 6.85183167e-01
4.23426390e-01 8.50035906e-01 -6.75854683e-01 8.84444773e-01
-3.18160802e-01 3.49209398e-01 -4.48237687e-01 -1.75305411e-01
-2.21962538e-02 -1.21233046e+00 6.25634909e-01 9.65544462e-01
6.23993158e-01 1.91618472e-01 -3.63866866e-01 4.70878601e-01
3.66102308e-01 7.20889449e-01 -2.81167150e-01 -2.50936151e-01
2.30732501e-01 1.64963913e+00 -5.34455717e-01 -7.21135512e-02
-5.79637051e-01 1.15453410e+00 -1.11460909e-01 6.03810847e-01
-3.43584806e-01 -4.48185503e-01 6.42760575e-01 -3.59890103e-01
5.00299156e-01 -7.21177101e-01 -4.12180647e-02 -1.40941381e+00
1.21585876e-02 -1.00602770e+00 2.68458575e-01 -5.46601176e-01
-1.13240337e+00 5.69277823e-01 -1.20584905e-01 -1.51791489e+00
-3.74027759e-01 -1.24956572e+00 -1.02015451e-01 9.35685456e-01
-1.56826544e+00 -1.20666957e+00 -1.73470140e-01 1.00993860e+00
1.54438158e-02 -6.21863127e-01 9.37656641e-01 3.07226598e-01
-8.63955319e-01 7.71839499e-01 5.59116483e-01 1.32032052e-01
6.66712523e-01 -1.16073227e+00 -2.21318305e-01 1.10854530e+00
2.70468444e-01 8.21649551e-01 6.96373820e-01 -4.10730422e-01
-2.46851277e+00 -4.39230740e-01 5.49209535e-01 2.55330652e-01
4.69682127e-01 -1.21516354e-01 -5.40152013e-01 4.36274737e-01
1.09381631e-01 -2.15194263e-02 7.80577421e-01 -8.90892670e-02
-4.88113761e-01 -7.77518868e-01 -9.91711676e-01 6.13724113e-01
3.70401680e-01 -4.56286520e-01 -3.06485802e-01 1.91139773e-01
-1.47788227e-01 -1.28035828e-01 -1.05646288e+00 8.31169114e-02
8.05656970e-01 -9.75443304e-01 1.08691835e+00 -3.66535872e-01
-1.84785232e-01 -6.41904533e-01 -4.31914777e-01 -1.22451496e+00
-4.92957741e-01 -7.28578150e-01 1.00520235e-02 1.00181103e+00
1.15829244e-01 -8.76513183e-01 5.41669667e-01 3.98997754e-01
4.15518433e-01 -4.16658849e-01 -1.25517762e+00 -6.31706417e-01
-3.66585672e-01 -1.93162054e-01 3.93817544e-01 6.10087395e-01
1.78311765e-01 1.14631876e-01 -6.67331815e-01 3.68519157e-01
1.13795304e+00 2.71406233e-01 8.63371134e-01 -8.97729397e-01
6.45832643e-02 -2.28142336e-01 -1.02778113e+00 -9.84471142e-01
-6.07701913e-02 -4.10367727e-01 -2.56997705e-01 -9.74115610e-01
1.13036849e-01 5.38914166e-02 -2.98404127e-01 2.21981540e-01
-3.03585809e-02 6.29035234e-01 4.78443235e-01 2.53019691e-01
1.13078319e-01 8.41159344e-01 1.43188894e+00 -1.93135172e-01
-7.79030323e-02 -6.98205084e-02 -2.52000034e-01 4.68710661e-01
2.15552971e-01 1.51206395e-02 -3.41619581e-01 -1.93231493e-01
1.20836064e-01 7.04508126e-02 2.59962846e-02 -9.81988609e-01
1.03116326e-01 -1.42271176e-01 7.26411700e-01 -7.10377812e-01
7.89255619e-01 -1.06714392e+00 1.72912672e-01 4.55513775e-01
3.26839000e-01 3.10079962e-01 -8.85396302e-02 4.39957410e-01
-5.13069272e-01 -2.75440123e-02 7.47797728e-01 2.41602197e-01
-9.86331344e-01 3.74726951e-01 -3.91751349e-01 -6.76503778e-01
1.05919147e+00 -3.84892702e-01 2.56049126e-01 -4.18491751e-01
-6.76604211e-01 -3.51147830e-01 2.59373367e-01 1.65914237e-01
1.13461733e+00 -1.75313342e+00 -7.66904831e-01 1.05182874e+00
9.90927499e-03 -6.83918834e-01 1.96612597e-01 9.63368058e-01
-8.47173870e-01 4.96687382e-01 -7.76852250e-01 -6.27403796e-01
-1.34290051e+00 5.95788836e-01 7.96653479e-02 1.64759755e-01
-3.22940648e-01 6.14587605e-01 -3.56366858e-02 -3.10867906e-01
2.28514280e-02 9.38856415e-03 -4.46393669e-01 3.65224510e-01
4.91238832e-01 5.80754697e-01 6.44894540e-02 -1.25827563e+00
-4.08993274e-01 1.13019991e+00 2.07305074e-01 -5.09410977e-01
1.39991832e+00 -1.70930922e-01 -5.81423163e-01 1.82068095e-01
1.48672688e+00 1.56217620e-01 -1.14115763e+00 -2.72117602e-03
-2.15849727e-01 -8.78866911e-01 9.95617360e-02 -2.46158212e-01
-1.44285607e+00 1.16627645e+00 9.21131968e-01 -9.50260982e-02
1.41546202e+00 -8.96796048e-01 2.34786302e-01 5.52833140e-01
4.48078990e-01 -1.12298501e+00 6.78778514e-02 1.07372336e-01
1.28465176e+00 -1.09202397e+00 4.52180803e-01 -5.11405528e-01
-7.47142792e-01 1.66882312e+00 4.49659497e-01 -2.16940805e-01
8.94974589e-01 -2.73515403e-01 3.00781101e-01 1.26882806e-01
-1.05952241e-01 3.87686566e-02 5.17518759e-01 5.02806485e-01
2.41750181e-01 5.70724420e-02 -8.26518059e-01 1.90928549e-01
7.05980957e-02 -2.03517064e-01 4.41728085e-01 8.46262991e-01
-1.05636962e-01 -1.25744879e+00 -8.95927966e-01 -3.86776105e-02
-1.24526054e-01 2.54600197e-01 -5.14478609e-02 9.18471038e-01
-2.24196270e-01 8.16719234e-01 -2.08414808e-01 -4.48249340e-01
3.00297827e-01 -1.10649474e-01 6.08873069e-01 2.89890226e-02
2.27693453e-01 6.00901425e-01 -3.41782212e-01 -8.12376916e-01
-5.11613548e-01 -7.46374130e-01 -1.03837907e+00 -2.67902195e-01
-2.83491284e-01 4.28409785e-01 1.10964715e+00 3.85886014e-01
3.29199225e-01 -2.47695848e-01 1.03702569e+00 -8.82579982e-01
-7.73334026e-01 -7.97375977e-01 -1.25043762e+00 4.47796434e-01
3.02418649e-01 -7.12802708e-01 -2.44251385e-01 1.21674024e-01] | [10.7764310836792, -1.6159309148788452] |
28d7d6e1-a869-4234-b5a8-9f83543d5bb8 | improving-few-shot-performance-of-language | 2212.02216 | null | https://arxiv.org/abs/2212.02216v1 | https://arxiv.org/pdf/2212.02216v1.pdf | Improving Few-Shot Performance of Language Models via Nearest Neighbor Calibration | Pre-trained language models (PLMs) have exhibited remarkable few-shot learning capabilities when provided a few examples in a natural language prompt as demonstrations of test instances, i.e., in-context learning. However, the performance of in-context learning is susceptible to the choice of prompt format, training examples and the ordering of the training examples. In this paper, we propose a novel nearest-neighbor calibration framework for in-context learning to ease this issue. It is inspired by a phenomenon that the in-context learning paradigm produces incorrect labels when inferring training instances, which provides a useful supervised signal to calibrate predictions. Thus, our method directly augments the predictions with a $k$-nearest-neighbor ($k$NN) classifier over a datastore of cached few-shot instance representations obtained by PLMs and their corresponding labels. Then adaptive neighbor selection and feature regularization modules are introduced to make full use of a few support instances to reduce the $k$NN retrieval noise. Experiments on various few-shot text classification tasks demonstrate that our method significantly improves in-context learning, while even achieving comparable performance with state-of-the-art tuning-based approaches in some sentiment analysis tasks. | ['Xu Cheng', 'Zhirui Zhang', 'Meixi Chen', 'Feng Nie'] | 2022-12-05 | null | null | null | null | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 2.40064681e-01 -2.56333232e-01 -6.05386913e-01 -9.33181167e-01
-1.04703653e+00 -1.93831265e-01 5.66411972e-01 4.18493718e-01
-5.07869363e-01 5.37746549e-01 1.29634887e-01 -2.18897566e-01
-1.62394002e-01 -9.26533222e-01 -5.98539114e-01 -4.99978453e-01
4.18537766e-01 3.44417959e-01 4.65570807e-01 -3.90656203e-01
5.29241741e-01 -1.06178209e-01 -2.06562519e+00 6.49906993e-01
7.56237030e-01 1.18141615e+00 2.02349603e-01 5.20515084e-01
-8.27559233e-01 8.54060590e-01 -6.04232609e-01 -3.79043341e-01
-3.49826440e-02 -3.00048709e-01 -4.75922376e-01 -1.65973946e-01
3.76419395e-01 -2.05227107e-01 -1.38312578e-02 8.93313348e-01
4.63147342e-01 6.83116972e-01 7.09303439e-01 -1.17996252e+00
-7.97105610e-01 7.20357835e-01 -1.71321690e-01 4.39027160e-01
3.62918437e-01 1.21291578e-01 1.17534077e+00 -1.43037307e+00
4.91223305e-01 9.30896997e-01 6.13826275e-01 6.29181623e-01
-1.05751824e+00 -7.44839609e-01 1.40597552e-01 4.04076844e-01
-1.30323195e+00 -5.78467309e-01 8.30753922e-01 -3.48312408e-01
1.17157423e+00 4.41571400e-02 3.39175105e-01 1.24391341e+00
5.46150748e-03 7.44514048e-01 8.26778471e-01 -9.04465556e-01
8.24785531e-01 6.04093373e-01 7.85368741e-01 4.14440274e-01
-8.94107819e-02 -2.29099728e-02 -8.33146930e-01 -5.12036085e-01
1.59305573e-01 4.40292746e-01 -4.54540327e-02 -3.93208504e-01
-6.91177130e-01 1.10749352e+00 2.66060680e-01 4.91007030e-01
-7.22427526e-03 -1.90991297e-01 6.28460586e-01 4.02835488e-01
6.13946974e-01 3.91291171e-01 -6.21295989e-01 -1.82920396e-01
-8.46534073e-01 -4.76003699e-02 7.98727214e-01 1.15751302e+00
1.04007030e+00 -4.67539728e-02 -5.06575167e-01 1.25574136e+00
-1.32344246e-01 1.98066592e-01 1.15719426e+00 -4.65708196e-01
3.68427575e-01 6.74150527e-01 7.13674054e-02 -6.42315388e-01
-8.41673091e-02 -1.51474029e-01 -2.84551084e-01 -1.89237759e-01
8.66253227e-02 -7.30034187e-02 -6.56663477e-01 1.59392107e+00
4.55955505e-01 6.11359119e-01 2.08886459e-01 7.01059341e-01
6.74648285e-01 7.90901482e-01 3.98270816e-01 -3.81727159e-01
1.26170647e+00 -1.02808440e+00 -6.39728069e-01 -4.50316519e-01
1.04661286e+00 -5.59786081e-01 1.80250525e+00 1.57061890e-01
-4.30042058e-01 -7.16205597e-01 -1.05918837e+00 3.01233698e-02
-7.87188351e-01 -2.55761743e-01 5.21709800e-01 5.25652766e-01
-4.35542017e-01 7.97610402e-01 -3.00632298e-01 -3.30920488e-01
1.38159886e-01 8.19882751e-02 4.26757857e-02 -2.11939216e-01
-1.44781041e+00 6.24814630e-01 5.02974212e-01 -6.49210632e-01
-5.43985784e-01 -7.86768258e-01 -1.03697038e+00 2.13269636e-01
5.67788184e-01 -1.85059100e-01 1.48384714e+00 -8.92448246e-01
-1.35577714e+00 6.33507013e-01 -2.12048873e-01 -3.64902914e-01
-2.45436989e-02 -3.02917987e-01 -7.66692698e-01 -7.53524825e-02
1.34765074e-01 3.61200392e-01 1.05132127e+00 -8.06522906e-01
-7.76969671e-01 -2.64566541e-01 -1.18841141e-01 1.36938289e-01
-7.65546262e-01 -7.78320655e-02 -2.83102661e-01 -5.36389768e-01
-1.04897246e-01 -6.24193132e-01 -2.13304579e-01 -2.23835438e-01
-6.29927451e-03 -5.71243763e-01 8.50741804e-01 1.53051347e-01
1.57618105e+00 -2.37251782e+00 -6.02266371e-01 9.92244855e-02
-2.17085838e-01 3.73018771e-01 -1.87597796e-01 4.62360382e-01
-4.94655184e-02 -1.59581244e-01 8.96203965e-02 -2.24297345e-01
9.10205953e-03 2.89343834e-01 -8.66553664e-01 4.13371511e-02
8.80746990e-02 7.84084201e-01 -1.14753127e+00 -5.71636677e-01
2.78265417e-01 2.44104400e-01 -4.39144969e-01 5.13968110e-01
-3.51984441e-01 -9.32344422e-02 -5.19210696e-01 6.64717257e-01
3.47608477e-02 -5.03919840e-01 1.00282900e-01 -4.02802462e-03
1.58149049e-01 2.88787514e-01 -1.16722596e+00 1.71993351e+00
-6.59486949e-01 2.24718004e-01 -6.97366297e-01 -1.13840663e+00
1.22068501e+00 2.60102689e-01 1.80739924e-01 -7.16213822e-01
1.85374901e-01 2.45048046e-01 -2.81447291e-01 -5.32642663e-01
5.37685394e-01 -1.60531133e-01 -3.33167106e-01 5.79777956e-01
3.77521157e-01 1.99899753e-03 2.14950219e-01 2.71047473e-01
9.41579461e-01 -2.18180567e-02 7.07031250e-01 -8.66611525e-02
4.37748462e-01 7.77826458e-02 6.00260437e-01 1.13448799e+00
-3.28981578e-01 4.66388613e-01 2.03112811e-01 -6.09251082e-01
-9.81674075e-01 -7.21483290e-01 -4.04814690e-01 1.93735063e+00
2.15135664e-02 -7.20805407e-01 -6.41707480e-01 -8.38833094e-01
1.10768475e-01 1.30851579e+00 -7.76165247e-01 -5.12819707e-01
-2.21915245e-01 -5.67764044e-01 1.89677789e-03 7.09865689e-01
1.49258431e-02 -1.16419113e+00 -5.39662838e-01 4.34038818e-01
1.93152010e-01 -8.50341737e-01 -5.92885613e-01 6.70526564e-01
-8.76320601e-01 -1.03931355e+00 -3.20051461e-01 -8.23380113e-01
5.31145513e-01 4.67114061e-01 1.04096973e+00 -2.96472777e-02
-3.52370799e-01 4.14436042e-01 -7.00626850e-01 -4.61552024e-01
-2.07857221e-01 -7.23599792e-02 3.29035968e-01 2.71018967e-02
1.30758667e+00 -4.58218396e-01 -4.91414458e-01 2.97045469e-01
-8.23569417e-01 -3.77764583e-01 4.10326213e-01 1.30190766e+00
8.69964063e-01 -2.23781228e-01 7.98255324e-01 -1.36732173e+00
7.87526369e-01 -6.86972082e-01 -3.92033607e-01 5.03832579e-01
-1.15036523e+00 8.48323628e-02 9.69040453e-01 -7.96772122e-01
-8.75165582e-01 1.42284750e-03 2.43193626e-01 -6.58938110e-01
-2.11787239e-01 4.66909409e-01 9.02756900e-02 1.76182762e-01
1.15169692e+00 3.31641734e-01 -3.25096965e-01 -3.47887099e-01
4.10205543e-01 1.07589805e+00 2.10311547e-01 -6.72654927e-01
4.07103568e-01 -2.03795414e-02 -5.65930128e-01 -7.07824886e-01
-1.46639991e+00 -9.58916008e-01 -4.71550673e-01 2.72300672e-02
4.32282537e-01 -8.79225552e-01 -1.61018834e-01 -3.93562578e-02
-6.17016733e-01 -2.89449543e-01 -5.97364008e-01 3.46719354e-01
-5.78839719e-01 6.17648661e-02 -5.07455349e-01 -8.91305566e-01
-4.65437323e-01 -9.24795687e-01 8.87510061e-01 3.40668947e-01
-3.47228646e-01 -7.80344188e-01 2.22633675e-01 1.35214955e-01
3.98070008e-01 -4.71320361e-01 1.14663208e+00 -1.35422528e+00
-2.21276488e-02 -5.68673849e-01 7.41472319e-02 1.07314974e-01
2.38640811e-02 -2.52566397e-01 -1.34429717e+00 -2.15440750e-01
1.46842659e-01 -7.32093811e-01 6.38254941e-01 1.14902541e-01
1.34543860e+00 -3.02952558e-01 -3.59193295e-01 4.40014839e-01
1.41743863e+00 2.16147289e-01 2.38512531e-01 2.70813167e-01
2.21002460e-01 4.53334183e-01 1.16583526e+00 7.30168521e-01
-1.08461320e-01 4.90038663e-01 -1.31821945e-01 5.85895360e-01
1.79498687e-01 -5.51373303e-01 1.32209748e-01 8.27538133e-01
5.98612726e-01 7.46887177e-02 -9.30590689e-01 5.21145761e-01
-1.97888601e+00 -9.35595870e-01 4.71916765e-01 2.23304987e+00
1.21430993e+00 3.61945659e-01 -2.11326510e-01 9.81288627e-02
8.90032351e-01 1.92539871e-01 -7.48582482e-01 -3.84334952e-01
1.93857223e-01 3.66786957e-01 6.82470128e-02 2.27251083e-01
-1.09214389e+00 1.01737261e+00 5.75084352e+00 1.11622739e+00
-9.63868380e-01 2.38183394e-01 6.82491899e-01 -2.47157738e-01
-1.89923093e-01 1.99812017e-02 -1.26355124e+00 6.15811229e-01
1.32130349e+00 -2.57655323e-01 1.63548052e-01 1.71548343e+00
-2.26560935e-01 -5.20475134e-02 -1.28877616e+00 1.08403468e+00
2.58428752e-01 -1.42927825e+00 1.14529654e-01 -5.09493053e-01
7.03201175e-01 -8.22974965e-02 4.38417383e-02 1.03335059e+00
1.69962332e-01 -4.65362579e-01 2.82259047e-01 3.51044059e-01
9.84559834e-01 -8.57899010e-01 5.65985024e-01 6.43567383e-01
-9.62444663e-01 -5.17382085e-01 -9.01763201e-01 -5.69447167e-02
-1.67689607e-01 6.70650542e-01 -1.05513680e+00 -1.39409557e-01
8.13056827e-01 4.97069031e-01 -5.28178155e-01 7.01824427e-01
-3.88296172e-02 6.49648249e-01 1.64339710e-02 -6.01684332e-01
2.02237532e-01 8.33866671e-02 9.67260152e-02 1.18472350e+00
2.38276392e-01 5.18573880e-01 2.56130099e-01 6.29084051e-01
-1.85842142e-02 6.00614369e-01 -7.63448179e-01 4.86840904e-02
8.15615773e-01 1.21964562e+00 -3.82164747e-01 -7.87995219e-01
-6.82045698e-01 6.30726159e-01 6.95970654e-01 2.09601060e-01
-5.17040193e-01 -7.72103608e-01 4.98554051e-01 2.10888609e-02
4.00065064e-01 2.25623325e-01 -6.34356886e-02 -1.30591202e+00
-1.52516454e-01 -5.93725860e-01 5.63437343e-01 -7.77522266e-01
-1.73516011e+00 4.04127389e-01 -8.05042610e-02 -1.46244180e+00
-5.28468609e-01 -6.05523050e-01 -9.23989892e-01 5.48487484e-01
-1.47329116e+00 -6.59999609e-01 -1.44654483e-01 7.10308552e-01
8.87755930e-01 -4.31903660e-01 1.18941641e+00 -5.71109727e-02
-4.51767206e-01 8.49975348e-01 3.46268326e-01 1.32732615e-01
1.21946275e+00 -1.04353166e+00 2.11122200e-01 4.38721687e-01
4.32507604e-01 8.42255116e-01 4.56101567e-01 -6.21547997e-01
-1.27216697e+00 -1.11003208e+00 1.04858649e+00 -3.67670298e-01
9.17148471e-01 -3.47792923e-01 -1.16253328e+00 5.67961335e-01
-2.30505064e-01 6.37964070e-01 1.45223260e+00 3.20968688e-01
-7.15280592e-01 -4.05456096e-01 -9.76311505e-01 6.20288312e-01
7.23026395e-01 -8.37373853e-01 -1.07116389e+00 4.82754797e-01
1.07777154e+00 -1.00121036e-01 -5.66232979e-01 1.95800379e-01
3.61880749e-01 -7.75220215e-01 7.16404140e-01 -9.04860020e-01
3.78821403e-01 8.66248235e-02 -5.43655753e-01 -1.10249341e+00
-2.41546452e-01 -1.85288280e-01 -4.39336270e-01 1.08791995e+00
4.71800357e-01 -2.77101934e-01 7.69990265e-01 9.70754325e-01
-4.01430354e-02 -7.93228924e-01 -7.69809306e-01 -6.81977093e-01
-1.80665702e-01 -6.88273668e-01 5.26493013e-01 1.13686764e+00
5.89641988e-01 5.70778310e-01 -2.41418377e-01 -2.10348770e-01
5.16469896e-01 3.47630233e-01 7.07613707e-01 -1.23762178e+00
-3.56956244e-01 7.06369663e-03 -1.89978257e-01 -8.62589419e-01
3.02675605e-01 -7.86543250e-01 2.14939222e-01 -6.43331826e-01
2.59845525e-01 -6.06887400e-01 -7.56907165e-01 5.85849404e-01
-4.79782730e-01 -2.38714181e-02 -7.67901763e-02 2.74345756e-01
-9.64329660e-01 4.51569051e-01 6.94196522e-01 -2.11185500e-01
-4.02253687e-01 1.07021399e-01 -5.98994970e-01 6.01550460e-01
5.85361838e-01 -6.06542528e-01 -6.00346923e-01 -3.50830443e-02
-1.05168130e-02 1.36137784e-01 -1.29690289e-01 -9.90573227e-01
6.63335085e-01 -3.25633854e-01 4.76927221e-01 -5.17852128e-01
3.70357126e-01 -8.22975695e-01 -5.42218089e-01 5.71970791e-02
-8.58193159e-01 -3.47718120e-01 -9.06892419e-02 9.25134480e-01
-2.29654983e-01 -7.58104861e-01 7.16537237e-01 -2.23640651e-01
-1.26280642e+00 2.12807313e-01 -5.33918217e-02 3.37756336e-01
9.84753072e-01 -9.46514234e-02 -3.22193444e-01 -1.88961059e-01
-6.83510602e-01 3.13415341e-02 2.55287349e-01 5.71263850e-01
8.20477784e-01 -1.34635317e+00 -1.65352672e-01 4.46750224e-01
9.04108286e-01 -2.63986677e-01 2.56877691e-01 3.75346154e-01
3.18174571e-01 3.81809980e-01 2.13251457e-01 -6.34400785e-01
-8.15338969e-01 9.83181834e-01 -1.42092794e-01 -1.90843809e-02
-6.32053256e-01 8.14427137e-01 -5.85949644e-02 -4.50597227e-01
4.31476414e-01 -9.00669396e-02 -2.87920922e-01 1.87996060e-01
1.09778368e+00 6.83424175e-02 1.91633105e-01 -1.03895113e-01
-9.61216614e-02 3.38335723e-01 -4.92234886e-01 1.33636817e-01
1.18705249e+00 -7.22211003e-02 3.77599716e-01 1.04692590e+00
1.28178895e+00 -2.99158037e-01 -1.32890368e+00 -8.62887621e-01
2.62499511e-01 -4.76683915e-01 -8.07933789e-03 -8.07501495e-01
-6.34845376e-01 8.15172434e-01 4.88194615e-01 1.27945310e-02
8.31461728e-01 -6.36591315e-02 8.06872845e-01 9.64258790e-01
6.37761354e-01 -1.60473752e+00 2.68093169e-01 6.47227705e-01
2.87186116e-01 -1.61306727e+00 -1.20813601e-01 -2.99007259e-02
-8.36593449e-01 1.21743655e+00 9.32544887e-01 -1.49739176e-01
9.29312050e-01 1.34538099e-01 1.47964761e-01 4.78543900e-02
-1.18145573e+00 -9.17087868e-03 1.62813336e-01 4.16925400e-01
4.42722470e-01 -2.10697979e-01 -1.07081942e-01 1.14025223e+00
2.48712525e-02 5.29614203e-02 1.36544555e-01 1.22564721e+00
-1.03317893e+00 -9.48475182e-01 -1.22751541e-01 6.41251862e-01
-2.19613835e-01 -4.21165049e-01 -2.88080405e-02 5.03163338e-01
-8.38502124e-02 9.22295630e-01 1.84832171e-01 -4.97941047e-01
3.87742043e-01 9.13522780e-01 -1.51994899e-01 -1.13141072e+00
-5.65807343e-01 -8.54703560e-02 -2.38976941e-01 -5.80072939e-01
-1.32774040e-01 -3.46992254e-01 -1.16405213e+00 1.19561933e-01
-6.41822577e-01 5.14936388e-01 5.38228810e-01 1.02770853e+00
3.23172599e-01 2.06939533e-01 8.97218645e-01 -5.93897641e-01
-1.17745471e+00 -9.76929724e-01 -8.02849591e-01 6.17948472e-01
8.09697062e-02 -7.39626706e-01 -6.79194033e-01 -1.19475417e-01] | [10.169939994812012, 3.6057944297790527] |
fd4a3bf5-e839-4219-b846-c580598ebf44 | u2-former-a-nested-u-shaped-transformer-for | 2112.02279 | null | https://arxiv.org/abs/2112.02279v2 | https://arxiv.org/pdf/2112.02279v2.pdf | U2-Former: A Nested U-shaped Transformer for Image Restoration | While Transformer has achieved remarkable performance in various high-level vision tasks, it is still challenging to exploit the full potential of Transformer in image restoration. The crux lies in the limited depth of applying Transformer in the typical encoder-decoder framework for image restoration, resulting from heavy self-attention computation load and inefficient communications across different depth (scales) of layers. In this paper, we present a deep and effective Transformer-based network for image restoration, termed as U2-Former, which is able to employ Transformer as the core operation to perform image restoration in a deep encoding and decoding space. Specifically, it leverages the nested U-shaped structure to facilitate the interactions across different layers with different scales of feature maps. Furthermore, we optimize the computational efficiency for the basic Transformer block by introducing a feature-filtering mechanism to compress the token representation. Apart from the typical supervision ways for image restoration, our U2-Former also performs contrastive learning in multiple aspects to further decouple the noise component from the background image. Extensive experiments on various image restoration tasks, including reflection removal, rain streak removal and dehazing respectively, demonstrate the effectiveness of the proposed U2-Former. | ['Guangming Lu', 'Jinxing Li', 'Wenjie Pei', 'Xin Feng', 'Haobo Ji'] | 2021-12-04 | null | null | null | null | ['image-dehazing', 'reflection-removal'] | ['computer-vision', 'computer-vision'] | [ 4.26253825e-01 -1.62893727e-01 4.00664389e-01 -1.74926728e-01
-7.25635111e-01 2.15817653e-02 4.45964634e-01 -1.63710237e-01
-1.34948462e-01 3.84720653e-01 4.79579419e-01 -4.15727854e-01
-1.53478533e-01 -8.34982991e-01 -7.77722776e-01 -1.02709520e+00
9.01779681e-02 -4.69578177e-01 2.94486344e-01 -4.23335373e-01
2.72669762e-01 4.02070969e-01 -1.61427534e+00 3.51628363e-01
1.16267765e+00 1.22416389e+00 5.80981493e-01 4.79322404e-01
-7.87565410e-02 1.08165514e+00 -4.86980528e-01 -3.81322652e-01
3.68192911e-01 -4.08876508e-01 -4.33368593e-01 4.78860468e-01
2.71862149e-01 -6.57410443e-01 -8.11087489e-01 1.24710786e+00
6.54920995e-01 9.15706679e-02 2.55713254e-01 -8.86449397e-01
-9.61463571e-01 3.58581662e-01 -6.92805529e-01 4.12608504e-01
1.92630868e-02 9.25893188e-02 6.99571371e-01 -9.25743997e-01
2.64516789e-02 1.42144215e+00 5.75921476e-01 1.81387931e-01
-8.84064555e-01 -5.62489629e-01 3.45091641e-01 5.91113985e-01
-1.22102523e+00 -5.51301181e-01 7.41918087e-01 -9.85035598e-02
9.03172255e-01 1.15657039e-01 5.71307778e-01 5.89056730e-01
2.33014658e-01 8.59535396e-01 1.34307754e+00 -2.91390896e-01
-2.91376989e-02 -2.33459845e-01 3.31688374e-02 6.55192196e-01
-1.27824306e-01 1.65544488e-02 -5.35521567e-01 4.00425464e-01
8.57341707e-01 2.80379981e-01 -6.69030428e-01 1.00111663e-01
-9.88243163e-01 5.40255964e-01 8.23077440e-01 -4.77646962e-02
-2.55495012e-01 2.75334474e-02 5.26208341e-01 5.56172490e-01
6.31648839e-01 -1.09887578e-01 -3.08083028e-01 3.09358954e-01
-7.02913940e-01 -1.32070214e-01 3.61449480e-01 7.28443205e-01
8.84698689e-01 1.94128156e-01 -4.09505367e-01 1.01034749e+00
2.70257711e-01 2.89138228e-01 4.95586693e-01 -7.51991212e-01
4.70157772e-01 5.06521106e-01 -1.18512258e-01 -8.89182866e-01
-2.78619584e-02 -6.27771735e-01 -1.47678840e+00 4.61247206e-01
-1.30013600e-01 2.79071331e-01 -1.13998246e+00 1.42688942e+00
3.07923913e-01 5.02731800e-01 2.48977855e-01 9.83460784e-01
8.89423013e-01 8.01516831e-01 -2.13791758e-01 -2.31411234e-01
1.52564323e+00 -1.32550335e+00 -6.70325041e-01 -3.26708376e-01
5.68501800e-02 -8.52633834e-01 1.08635616e+00 4.09087360e-01
-1.09637821e+00 -6.94107890e-01 -1.18116021e+00 -6.65916502e-01
-7.45367259e-02 1.35522991e-01 5.97794712e-01 3.47558826e-01
-1.14158821e+00 5.59553921e-01 -7.64239252e-01 -8.48695040e-02
6.65520012e-01 2.61805773e-01 -3.66489410e-01 -4.77118254e-01
-1.13482857e+00 7.75698721e-01 5.86456656e-02 7.66147852e-01
-1.04694223e+00 -5.28355122e-01 -1.02443874e+00 3.28470111e-01
6.34047017e-02 -6.61340296e-01 7.84729004e-01 -7.95394003e-01
-1.58775508e+00 5.91544151e-01 -3.09920758e-01 -4.73014563e-01
4.68926311e-01 -3.13134968e-01 -2.37745926e-01 1.20145015e-01
-1.12106450e-01 3.28771800e-01 1.18939853e+00 -1.13150036e+00
-6.14461422e-01 -4.14335579e-01 2.12746993e-01 4.57359701e-01
-4.19860691e-01 -3.20627317e-02 -7.10260212e-01 -8.96721542e-01
3.17404240e-01 -3.99475813e-01 -2.77308285e-01 1.40316337e-01
-3.60216677e-01 -2.45820992e-02 9.03679729e-01 -9.97152686e-01
1.03381550e+00 -2.41673183e+00 2.52075046e-01 -2.03038409e-01
4.11156833e-01 4.49242026e-01 -2.75878966e-01 4.24238712e-01
-7.29957744e-02 -1.96507499e-01 -4.53438520e-01 -5.98205805e-01
-9.00852159e-02 4.19817507e-01 -6.53280854e-01 6.09162271e-01
2.99313933e-01 7.79823184e-01 -6.12495959e-01 -2.86670744e-01
5.35859525e-01 9.43243444e-01 -5.25336683e-01 3.94278705e-01
1.21052377e-01 4.18248534e-01 -3.03583175e-01 7.17412353e-01
1.16784823e+00 -1.53242528e-01 -1.50588840e-01 -5.98618627e-01
-2.28418469e-01 3.91253769e-01 -9.70337451e-01 1.64940012e+00
-6.63410008e-01 6.01365805e-01 4.86349821e-01 -1.33284163e+00
8.29893231e-01 9.57524031e-02 8.30543414e-02 -1.06425214e+00
4.02159663e-03 1.30033433e-01 -2.11010396e-01 -5.87886870e-01
4.27205116e-01 -1.16904378e-01 3.75091672e-01 1.76476449e-01
-1.34357423e-01 1.93528041e-01 2.27787755e-02 2.50364076e-02
1.06281853e+00 1.74770638e-01 1.47423476e-01 -2.46327579e-01
8.20339322e-01 -5.28683007e-01 6.84080601e-01 6.14963710e-01
-1.50485262e-01 7.94922471e-01 3.18785906e-01 -4.43176001e-01
-8.78153265e-01 -1.06292772e+00 -1.15796588e-01 9.35357332e-01
4.86094385e-01 -3.18699330e-01 -5.73149920e-01 -3.45622927e-01
-3.98033112e-01 1.30414948e-01 -3.94167423e-01 -2.41860256e-01
-6.37342632e-01 -1.13093019e+00 3.41573596e-01 2.70384908e-01
1.07103407e+00 -1.02865064e+00 -4.34360504e-01 1.32953316e-01
-3.89220417e-01 -1.14744020e+00 -4.23946291e-01 3.24113995e-01
-8.05658937e-01 -1.04959822e+00 -5.76461017e-01 -9.92285490e-01
7.04617500e-01 9.16132331e-01 8.30863059e-01 3.20038170e-01
-3.37257266e-01 -9.89456400e-02 -4.78994489e-01 -1.14451898e-02
-9.67622548e-03 -2.86348134e-01 -4.22764510e-01 2.49495819e-01
-5.30945621e-02 -9.91377652e-01 -9.93121624e-01 1.95852414e-01
-1.25041366e+00 2.56859958e-01 7.84186184e-01 1.10998905e+00
3.89152706e-01 5.36576748e-01 4.00407046e-01 -5.60444176e-01
5.27125776e-01 -2.72631735e-01 -5.16670465e-01 2.40054414e-01
-4.09076989e-01 2.79860478e-02 7.64722347e-01 -1.45536348e-01
-1.14057481e+00 -2.91900635e-01 -4.58571047e-01 -4.06580508e-01
2.21485168e-01 4.21251565e-01 -4.23046023e-01 -1.84996173e-01
1.26135856e-01 6.95653975e-01 7.09215999e-02 -7.40177333e-01
1.87986329e-01 6.37869298e-01 7.64122963e-01 -2.70815164e-01
1.04498112e+00 5.40494621e-01 -1.31205767e-01 -8.13247561e-01
-9.69121993e-01 -1.54120386e-01 -3.27964127e-01 5.94863184e-02
6.76728725e-01 -1.34499502e+00 -9.23876405e-01 9.57961738e-01
-1.14090192e+00 -2.80984581e-01 -1.68761432e-01 1.84205964e-01
-8.14978257e-02 7.38873541e-01 -8.91324520e-01 -6.14301205e-01
-6.10305786e-01 -1.49401593e+00 1.17090499e+00 3.33280921e-01
9.09759760e-01 -8.07980835e-01 -4.44067180e-01 4.33363587e-01
6.38779461e-01 -1.29608005e-01 1.01053798e+00 1.77051947e-02
-9.76888180e-01 1.25781327e-01 -8.47899735e-01 8.29403877e-01
2.40855232e-01 -5.21445513e-01 -1.12284315e+00 -5.13072312e-01
3.52142662e-01 -2.80092120e-01 1.39800739e+00 1.92646161e-01
1.48650789e+00 -3.34752291e-01 4.72668223e-02 1.15658569e+00
1.50908780e+00 -9.19585824e-02 1.16811240e+00 5.93560278e-01
8.59651923e-01 4.27407473e-01 2.85360187e-01 3.53299379e-01
6.02951169e-01 4.94944096e-01 7.57439971e-01 -5.14952898e-01
-4.58152533e-01 -1.59413636e-01 5.92903674e-01 1.07967472e+00
1.32714942e-01 -1.77900061e-01 -3.68810534e-01 4.70073104e-01
-1.73718846e+00 -7.55170166e-01 1.04689494e-01 2.11563301e+00
7.66737878e-01 -1.37119368e-01 -5.99641800e-01 3.35247189e-01
5.86396575e-01 5.07968605e-01 -6.04439616e-01 -5.27611077e-02
-3.32009554e-01 2.47013971e-01 4.20819700e-01 6.91730380e-01
-1.18679273e+00 8.01882088e-01 5.91148663e+00 1.02006710e+00
-1.02835822e+00 9.30132344e-02 7.19698370e-01 1.36193544e-01
-2.00746983e-01 5.17587028e-02 -5.61528862e-01 4.60841119e-01
3.74700993e-01 1.64794311e-01 7.02732623e-01 3.34664166e-01
2.42479563e-01 -9.96758137e-03 -6.98277593e-01 1.03991318e+00
2.08400890e-01 -1.12914526e+00 2.60046721e-01 -1.02572016e-01
4.15088058e-01 6.03485480e-02 9.94407088e-02 2.69199282e-01
1.36219621e-01 -1.11424220e+00 5.54358840e-01 2.51666456e-01
7.24145710e-01 -7.58817136e-01 6.67112947e-01 2.82559484e-01
-1.23933196e+00 -2.46500134e-01 -7.27728963e-01 -1.92463651e-01
1.21959224e-01 9.24131989e-01 -1.90816611e-01 7.70202160e-01
1.02693975e+00 1.08229053e+00 -5.15312433e-01 9.06983554e-01
-3.79938900e-01 3.84242624e-01 -1.24587134e-01 7.72932172e-01
1.52094781e-01 -4.86494511e-01 4.41305965e-01 1.20300281e+00
4.43517476e-01 8.86002705e-02 1.59259662e-01 4.08257365e-01
-8.84051248e-02 -2.80043542e-01 -2.71496147e-01 4.90382165e-01
3.06978792e-01 1.26318002e+00 -3.95339072e-01 -1.30550250e-01
-4.95174408e-01 1.29230869e+00 2.57678002e-01 4.84134316e-01
-7.43802130e-01 -5.23814917e-01 9.53904331e-01 -9.28314961e-03
5.88579774e-01 -1.65544018e-01 -1.26970381e-01 -1.32882917e+00
3.50308657e-01 -1.05283415e+00 1.03350073e-01 -9.02635992e-01
-1.27898765e+00 9.68241930e-01 -3.62069845e-01 -1.28830218e+00
4.98850465e-01 -6.22057855e-01 -6.28359020e-01 1.06242061e+00
-2.30115557e+00 -1.41953063e+00 -6.30961001e-01 9.35406089e-01
6.59014165e-01 -1.53437220e-02 3.77789289e-01 6.05329633e-01
-8.93589556e-01 5.12623429e-01 2.18801200e-01 7.20252916e-02
6.54384017e-01 -1.02772641e+00 4.21002239e-01 1.35344255e+00
-2.01530516e-01 5.71211994e-01 5.40472746e-01 -3.12779993e-01
-1.60717082e+00 -1.19051433e+00 5.40705562e-01 3.00236166e-01
6.24779344e-01 -3.40131313e-01 -1.03498662e+00 5.34860790e-01
4.08264577e-01 2.97462314e-01 2.63023198e-01 -2.98529476e-01
-5.89722574e-01 -5.06906569e-01 -8.87139022e-01 6.14805460e-01
1.06805491e+00 -7.01266944e-01 -4.05472070e-01 4.45970684e-01
8.87008548e-01 -5.26803434e-01 -6.73948288e-01 4.34578091e-01
2.02712446e-01 -1.28220606e+00 1.37826192e+00 -2.99601313e-02
7.69648969e-01 -6.31259322e-01 -4.02252764e-01 -1.20194185e+00
-4.09718812e-01 -6.52599812e-01 -8.13848302e-02 1.34811044e+00
-2.58432835e-01 -8.15244257e-01 3.07549000e-01 -5.32035865e-02
-4.91641521e-01 -7.36199498e-01 -1.11377287e+00 -3.50205570e-01
-2.43436188e-01 -2.25383028e-01 5.75279415e-01 6.09177411e-01
-5.45944870e-01 2.59024054e-01 -8.22572291e-01 5.18939078e-01
9.12350953e-01 2.52329737e-01 5.42073369e-01 -8.77667129e-01
-3.93913209e-01 -2.34708011e-01 -3.52299035e-01 -1.64183676e+00
2.76653608e-03 -8.94098341e-01 2.61191785e-01 -1.77171183e+00
3.00965488e-01 -2.68540978e-01 -4.62205797e-01 5.24142206e-01
-4.56956953e-01 4.49292153e-01 2.85933435e-01 3.64890516e-01
-4.15029913e-01 9.91824865e-01 1.43040276e+00 -2.25622609e-01
1.28608972e-01 -1.68817222e-01 -9.62641358e-01 6.43633187e-01
4.11889017e-01 -1.99850559e-01 -5.32686651e-01 -1.08320785e+00
5.15544489e-02 5.54606952e-02 7.64631867e-01 -7.82029569e-01
4.30681497e-01 1.01781800e-01 3.43570799e-01 -4.63875324e-01
3.56804550e-01 -7.01925337e-01 -1.74021333e-01 2.51741290e-01
-2.15068776e-02 1.14505254e-02 7.83732533e-02 6.92332029e-01
-6.40931308e-01 9.02909860e-02 9.59042609e-01 -1.30139306e-01
-8.47059727e-01 4.32177752e-01 -3.16059202e-01 -1.90547585e-01
5.41301072e-01 -2.88829386e-01 -5.75435698e-01 -1.52677029e-01
-4.50864851e-01 2.36198172e-01 3.52108419e-01 2.60690153e-01
9.03105557e-01 -1.11546230e+00 -8.26207161e-01 5.21762729e-01
-1.90808654e-01 2.28147015e-01 6.89243913e-01 9.58585620e-01
-5.95506370e-01 -5.60251921e-02 -2.97286451e-01 -4.93548095e-01
-1.12911940e+00 4.63116497e-01 4.46299821e-01 -4.73564148e-01
-1.37039292e+00 8.77792418e-01 7.55743444e-01 -1.82657376e-01
3.55158776e-01 -1.49897754e-01 -1.58191815e-01 -1.63434207e-01
9.24152255e-01 2.70958781e-01 1.26629531e-01 -4.27201003e-01
-1.84692517e-01 5.05128205e-01 -2.62601346e-01 1.73874572e-01
1.56915283e+00 -6.16041303e-01 -6.78320467e-01 -5.41185811e-02
1.14500010e+00 -1.82466835e-01 -1.57166362e+00 -4.68024373e-01
-4.45164263e-01 -6.95495427e-01 5.00263333e-01 -5.81413507e-01
-1.58042836e+00 1.11936224e+00 6.86616659e-01 9.17998403e-02
1.88376009e+00 -5.49386859e-01 1.05604970e+00 2.27854982e-01
1.66300476e-01 -6.26497805e-01 1.21000432e-03 5.00145555e-01
9.13223088e-01 -1.15905833e+00 1.62627101e-01 -5.11810005e-01
-3.41921628e-01 9.70539868e-01 4.59659845e-01 -2.25693360e-01
6.47161067e-01 4.77825016e-01 6.93621263e-02 -1.23120472e-02
-6.65052235e-01 -2.09182873e-01 -8.35380703e-03 5.93610942e-01
2.90557414e-01 -3.00575763e-01 -5.75402342e-02 2.83396035e-01
8.81805569e-02 -2.33267527e-02 6.16209507e-01 7.03592658e-01
-5.23677349e-01 -1.03522730e+00 -3.87797654e-01 2.61923641e-01
-4.45171088e-01 -5.53500116e-01 3.03995788e-01 3.94205958e-01
1.01066075e-01 1.11482561e+00 -1.59228057e-01 -3.03637356e-01
2.85272658e-01 -4.93685931e-01 3.74482542e-01 -2.67003775e-01
-6.73356116e-01 2.22061932e-01 -2.69685268e-01 -6.03567183e-01
-4.98412579e-01 -2.47124374e-01 -8.66103292e-01 -3.21626931e-01
-5.75452037e-02 -1.35120302e-02 4.07093793e-01 1.00496686e+00
3.72114003e-01 8.66619051e-01 7.73692131e-01 -9.61096287e-01
-5.52132010e-01 -9.65678215e-01 -6.38139129e-01 2.19757661e-01
8.96400392e-01 -4.66537535e-01 -4.88324195e-01 1.62198301e-02] | [11.124495506286621, -2.3519446849823] |
ffe5c744-4fd8-4e7c-b96b-35d222edf50e | two-stream-3d-semantic-scene-completion | 1804.03550 | null | https://arxiv.org/abs/1804.03550v4 | https://arxiv.org/pdf/1804.03550v4.pdf | Two Stream 3D Semantic Scene Completion | Inferring the 3D geometry and the semantic meaning of surfaces, which are occluded, is a very challenging task. Recently, a first end-to-end learning approach has been proposed that completes a scene from a single depth image. The approach voxelizes the scene and predicts for each voxel if it is occupied and, if it is occupied, the semantic class label. In this work, we propose a two stream approach that leverages depth information and semantic information, which is inferred from the RGB image, for this task. The approach constructs an incomplete 3D semantic tensor, which uses a compact three-channel encoding for the inferred semantic information, and uses a 3D CNN to infer the complete 3D semantic tensor. In our experimental evaluation, we show that the proposed two stream approach substantially outperforms the state-of-the-art for semantic scene completion. | ['Yueh-Tung Chen', 'Johann Sawatzky', 'Martin Garbade', 'Juergen Gall'] | 2018-04-10 | null | null | null | null | ['3d-semantic-scene-completion'] | ['computer-vision'] | [ 4.84700024e-01 3.18141639e-01 2.51334999e-02 -6.41060174e-01
-7.46199250e-01 -4.69174147e-01 4.02051896e-01 2.75070399e-01
-1.55242831e-01 1.62041470e-01 1.58988968e-01 -1.01586908e-01
1.35626927e-01 -9.90418017e-01 -9.90243196e-01 -3.29573184e-01
1.50197700e-01 6.63122356e-01 4.07069474e-01 2.65507042e-01
4.41933870e-01 4.20281947e-01 -1.79849756e+00 5.66336751e-01
5.64408064e-01 1.59557748e+00 4.11848783e-01 6.34084404e-01
-6.30684316e-01 6.79864764e-01 -4.39525768e-02 -2.24925932e-02
2.84563214e-01 1.72776133e-01 -9.70508933e-01 4.54503179e-01
7.51244366e-01 -5.43313146e-01 -3.69902283e-01 8.11454475e-01
-1.09843634e-01 -5.30627146e-02 5.79783738e-01 -1.02490199e+00
-2.92968061e-02 -5.82379140e-02 -4.21769083e-01 -3.48903775e-01
5.37002385e-01 -6.24442138e-02 1.05266666e+00 -1.15023375e+00
6.21830761e-01 1.34598565e+00 4.48421866e-01 1.23048879e-01
-1.11664486e+00 -2.99338490e-01 4.67135310e-01 7.99681526e-03
-1.21321607e+00 1.56700108e-02 1.10575211e+00 -6.04057729e-01
9.62964177e-01 -5.24540953e-02 1.07564759e+00 6.33877337e-01
-3.58855367e-01 9.79955077e-01 1.16183734e+00 -2.83516198e-01
7.63856471e-01 -3.84454191e-01 4.38203514e-02 8.80967259e-01
-1.66710131e-02 -1.77851081e-01 -7.60030746e-01 -7.59848803e-02
8.53588462e-01 2.62962461e-01 1.94982979e-02 -8.44990611e-01
-1.11443293e+00 5.23708642e-01 7.21512198e-01 -2.74626762e-01
-5.23626685e-01 5.40109992e-01 2.15976104e-01 -1.89306796e-01
9.32776332e-01 1.40033802e-02 -6.20917499e-01 -1.96226224e-01
-8.03828597e-01 3.68694186e-01 8.42822909e-01 9.49705422e-01
1.28136957e+00 -4.52884465e-01 2.86129028e-01 5.78529298e-01
4.60103303e-01 4.94554847e-01 -3.22673440e-01 -1.27921569e+00
5.45583844e-01 9.58276212e-01 1.00097261e-01 -6.71389639e-01
-3.61648321e-01 -4.10405695e-02 -3.28054637e-01 4.19426441e-01
2.80044734e-01 4.49445784e-01 -1.22030628e+00 1.28711343e+00
8.50235224e-01 6.56631529e-01 5.10284454e-02 1.02437460e+00
7.81449497e-01 3.24566960e-01 4.48698644e-03 5.29381573e-01
1.43553638e+00 -9.27712500e-01 -1.42721996e-01 -3.68498206e-01
3.08779806e-01 -6.10117018e-01 1.10053110e+00 5.10990322e-01
-8.89058352e-01 -3.66216570e-01 -1.06736159e+00 -6.93336964e-01
-2.94846147e-01 -7.09749013e-02 1.08225763e+00 2.70236760e-01
-9.83703613e-01 6.07694268e-01 -1.19527400e+00 -3.35508853e-01
6.48370802e-01 2.30298303e-02 -3.49365264e-01 -7.22049594e-01
-6.32448912e-01 3.39301676e-01 3.55296105e-01 -1.28138274e-01
-1.20679057e+00 -8.22498679e-01 -1.14914322e+00 -3.68978456e-02
3.96394312e-01 -9.39233840e-01 1.14394701e+00 -6.40056372e-01
-1.39903712e+00 1.01007032e+00 -4.12725270e-01 -8.96824431e-03
2.47893587e-01 -3.22257370e-01 4.54738081e-01 5.17419159e-01
2.47531235e-01 8.33611965e-01 8.52587640e-01 -1.71820474e+00
-6.60375357e-01 -1.03936076e+00 5.23086190e-01 4.93085563e-01
4.68477681e-02 -6.42923355e-01 -7.16156602e-01 -1.78947866e-01
8.95203710e-01 -8.47681999e-01 -2.44951636e-01 5.80433547e-01
-6.14290535e-01 -1.53851151e-01 6.46685719e-01 -4.81610656e-01
4.24651474e-01 -2.23100543e+00 2.54168838e-01 2.97235608e-01
3.51810157e-01 -4.38023567e-01 8.18801578e-03 2.06021041e-01
3.13852280e-01 -1.02721505e-01 -6.87807024e-01 -1.03299236e+00
1.88771248e-01 4.27230448e-01 -4.46019739e-01 4.02748227e-01
1.88813761e-01 6.66711152e-01 -1.09526777e+00 -2.10054010e-01
5.94416440e-01 9.22648370e-01 -8.51122022e-01 4.38722610e-01
-6.43911600e-01 6.45480633e-01 -7.62475848e-01 9.23384488e-01
1.04894948e+00 -2.96105325e-01 -5.45091145e-02 -3.10431033e-01
-1.48885444e-01 5.10042727e-01 -9.60570216e-01 2.66082096e+00
-6.12187982e-01 3.12602311e-01 9.37227905e-02 -9.03052509e-01
7.66989112e-01 9.60234180e-02 6.32065594e-01 -5.24520814e-01
3.01288045e-03 2.69734234e-01 -9.94366884e-01 -4.48531657e-01
5.29501796e-01 -2.78868787e-02 -1.07815638e-01 2.01331481e-01
-1.60932302e-01 -8.81869018e-01 -3.44549328e-01 2.27093652e-01
1.29529369e+00 7.23972797e-01 -1.55483812e-01 1.69403367e-02
3.39242697e-01 3.41386914e-01 3.02284151e-01 3.86089832e-01
2.42510319e-01 7.05291927e-01 6.21761858e-01 -4.75449830e-01
-1.05183434e+00 -1.47749317e+00 1.60672646e-02 6.58229411e-01
5.84961712e-01 -3.88054907e-01 -8.31957519e-01 -4.76762086e-01
3.23897511e-01 4.15904671e-01 -5.87478161e-01 1.01668686e-01
-3.04787278e-01 -8.13840106e-02 1.51018247e-01 5.41185081e-01
6.09836400e-01 -4.73009795e-01 -9.33286965e-01 -3.18219960e-02
-3.08532119e-01 -1.65612268e+00 7.50167295e-02 1.24144144e-01
-1.19033504e+00 -1.19688356e+00 -3.34199905e-01 -5.95876515e-01
8.54152501e-01 5.44232905e-01 1.02670515e+00 6.44222125e-02
-3.30081403e-01 5.41435182e-01 -3.92537296e-01 -1.51505992e-01
1.81847677e-01 -7.22759068e-02 -4.30477738e-01 2.40087807e-01
1.71913341e-01 -7.44899511e-01 -9.06782627e-01 -8.75483230e-02
-8.98687482e-01 5.50570786e-01 3.23764086e-01 2.89156049e-01
1.24322999e+00 -1.23670503e-01 -1.49743274e-01 -9.35066342e-01
-3.36894870e-01 -3.84549886e-01 -5.63310027e-01 -2.24027604e-01
-2.15230912e-01 2.15589643e-01 3.25409174e-01 1.54524639e-01
-1.13983476e+00 6.26272500e-01 -2.08766490e-01 -7.84455776e-01
-4.95370030e-01 3.89347941e-01 -2.94951022e-01 1.68061145e-02
3.79452370e-02 -1.81490667e-02 -2.96535641e-01 -9.63587940e-01
6.63716137e-01 5.16813576e-01 6.52706087e-01 -7.52129197e-01
6.59037888e-01 1.28145123e+00 3.48657548e-01 -7.51351595e-01
-1.40468168e+00 -8.68135333e-01 -9.38526154e-01 -1.91282302e-01
1.21738362e+00 -1.36917007e+00 -5.41604042e-01 6.22819841e-01
-1.46355307e+00 -6.56745493e-01 -2.50379145e-01 4.10881490e-01
-8.32157373e-01 2.62053043e-01 -4.28085327e-01 -6.61513865e-01
3.19731422e-02 -1.08620310e+00 1.89992118e+00 -1.67272404e-01
1.26957268e-01 -7.63693929e-01 -1.50139228e-01 6.61491394e-01
1.09918319e-01 5.51510513e-01 9.61603701e-01 6.56713173e-02
-1.18737507e+00 -5.00353947e-02 -6.17545426e-01 8.86734873e-02
-8.29652995e-02 -4.72947150e-01 -1.40754116e+00 1.24863490e-01
-9.35038775e-02 -3.78425479e-01 1.01613915e+00 2.05353022e-01
1.53160346e+00 1.25329465e-01 -1.53326482e-01 1.07829785e+00
1.65855598e+00 -3.45246583e-01 6.27571046e-01 1.14311151e-01
1.24694860e+00 7.46980786e-01 7.29718924e-01 6.56667113e-01
1.07472730e+00 4.99719262e-01 1.08074570e+00 -1.63276657e-01
-3.86607230e-01 -5.17460883e-01 -3.82687822e-02 5.66866159e-01
-2.61506590e-04 4.28052694e-02 -8.08134317e-01 4.08133656e-01
-1.79103851e+00 -3.56140435e-01 -2.27074414e-01 2.11040211e+00
5.56825995e-01 8.82938951e-02 -4.65541244e-01 1.69058889e-01
2.75997192e-01 3.54757249e-01 -8.53053987e-01 -1.06277496e-01
1.08662575e-01 4.93517578e-01 5.87695897e-01 7.77119458e-01
-9.63169336e-01 1.24663854e+00 5.70869493e+00 3.48718226e-01
-9.82880712e-01 1.79671779e-01 6.15311444e-01 -2.13689096e-02
-6.21405065e-01 4.24409449e-01 -4.74516749e-01 -3.18321027e-02
4.53381479e-01 4.13780093e-01 5.45677185e-01 1.00743675e+00
1.68006018e-01 -5.71845174e-01 -1.32188141e+00 1.20652282e+00
2.53163457e-01 -1.19791734e+00 6.20692261e-02 1.27427399e-01
7.01423585e-01 2.24101648e-01 -1.84023544e-01 -1.06608100e-01
1.43835083e-01 -8.02546799e-01 1.07598579e+00 5.59735060e-01
1.00624609e+00 -4.97775435e-01 2.65452564e-01 2.09672123e-01
-1.56655633e+00 1.29103899e-01 -2.55643904e-01 -3.82432401e-01
2.48745516e-01 9.98315334e-01 -5.64221501e-01 3.84787321e-01
7.82977343e-01 1.00548470e+00 -4.36604410e-01 8.24963152e-01
-5.47634780e-01 3.74988228e-01 -4.07934189e-01 4.54344243e-01
2.40588635e-01 -1.53878853e-01 3.65466893e-01 7.52417326e-01
2.19427958e-01 2.21299693e-01 4.73257512e-01 1.08888161e+00
-1.07744209e-01 -2.20811501e-01 -5.37933111e-01 2.20826566e-01
4.19506878e-01 1.14123356e+00 -8.71027708e-01 -4.91912127e-01
-4.42026168e-01 1.50311422e+00 4.27792490e-01 4.63587850e-01
-5.57308376e-01 -1.39562160e-01 9.32030320e-01 1.51483312e-01
4.48235840e-01 -5.54032981e-01 -8.90012443e-01 -1.13762629e+00
2.87851810e-01 5.28748631e-02 1.56210199e-01 -1.09213412e+00
-8.98992717e-01 2.23146081e-01 -1.25497617e-02 -9.84272540e-01
1.86550155e-01 -7.97965407e-01 -2.09858268e-01 7.26496279e-01
-2.03374934e+00 -1.36780131e+00 -9.56653953e-01 6.45483911e-01
6.49279296e-01 5.66595018e-01 6.41636431e-01 1.22032627e-01
-2.42204480e-02 -2.09286109e-01 -3.23393881e-01 -1.36660896e-02
3.29089202e-02 -1.32119155e+00 5.62463105e-01 3.97571683e-01
-1.54666975e-01 1.39974922e-01 3.50448400e-01 -5.92868447e-01
-1.87268806e+00 -1.24212897e+00 7.30778873e-01 -4.25598264e-01
2.87687689e-01 -6.83998406e-01 -7.10366309e-01 5.78287661e-01
-5.50045013e-01 3.33505094e-01 3.43739092e-01 -1.27886623e-01
-7.57901132e-01 9.76477191e-02 -1.17521834e+00 3.45029742e-01
1.58467484e+00 -8.84104133e-01 -2.81702101e-01 3.07730615e-01
1.19269049e+00 -7.28677809e-01 -8.86655271e-01 3.44757617e-01
3.99560750e-01 -1.07535410e+00 1.26603460e+00 -1.39396697e-01
7.14581013e-01 -4.46875125e-01 -7.22314477e-01 -1.04517567e+00
1.91168740e-01 2.74963081e-02 -1.21718667e-01 6.64703131e-01
4.23779413e-02 -2.27628186e-01 1.17890716e+00 6.52114868e-01
-5.66751301e-01 -8.00734699e-01 -1.14119971e+00 -4.20367032e-01
-2.56417096e-01 -9.61608529e-01 7.32815802e-01 5.74386358e-01
-4.11217481e-01 1.59717232e-01 6.71498403e-02 4.62297708e-01
1.04379201e+00 4.20777828e-01 9.04759228e-01 -1.36472118e+00
-1.07135020e-01 -7.97301978e-02 -7.03276873e-01 -1.50804389e+00
3.38416934e-01 -1.02539790e+00 2.12980583e-01 -1.94138443e+00
5.76431677e-02 -5.96026003e-01 2.63262372e-02 4.95584577e-01
8.14396963e-02 4.06062305e-01 6.35576108e-03 2.75970940e-02
-7.17530072e-01 7.45893240e-01 1.14285672e+00 -1.76949352e-01
-6.55135587e-02 -5.29263914e-01 -3.78318787e-01 8.36541235e-01
4.52650130e-01 -3.01679432e-01 -3.30765665e-01 -8.71034026e-01
2.80476958e-01 8.60955939e-02 8.73978794e-01 -9.85737205e-01
-5.34410216e-02 -1.76070899e-01 2.56010741e-01 -8.90590250e-01
9.94537532e-01 -9.73873138e-01 -1.62485942e-01 2.32956216e-01
-1.35527253e-01 -4.06081259e-01 -6.62031695e-02 7.61930406e-01
-1.43948883e-01 2.44339347e-01 5.18593907e-01 -3.71763051e-01
-9.96199489e-01 7.02049375e-01 2.91140825e-01 3.71866226e-02
7.58639455e-01 -3.23265254e-01 1.46290481e-01 -3.21413875e-02
-6.17736578e-01 1.24096312e-01 8.45552146e-01 3.66160929e-01
1.04138196e+00 -1.27142692e+00 -4.83802766e-01 2.67372161e-01
4.39783186e-01 8.78030241e-01 3.14670593e-01 4.99001950e-01
-7.15800107e-01 1.70734420e-01 1.39649019e-01 -1.00830948e+00
-8.88452888e-01 3.40639085e-01 9.06488672e-02 1.36759564e-01
-1.04106188e+00 7.91411459e-01 5.48632026e-01 -5.28254807e-01
2.83328503e-01 -6.84332013e-01 2.81348169e-01 -2.80888736e-01
4.46982950e-01 1.74042210e-01 1.72612876e-01 -6.34707272e-01
-3.16221476e-01 1.01171398e+00 3.68414104e-01 -3.30620021e-01
1.39054489e+00 -2.68916190e-01 -3.41111064e-01 6.72383726e-01
1.41230309e+00 -4.09772307e-01 -1.74523771e+00 -4.31956202e-01
-1.36080235e-01 -1.00539601e+00 2.77284354e-01 -5.09508908e-01
-1.02120817e+00 1.10674608e+00 2.54377306e-01 -2.28686124e-01
1.00033617e+00 4.83525842e-01 9.85833287e-01 1.12334698e-01
7.86845624e-01 -8.95134985e-01 3.44736010e-01 6.93636417e-01
6.69494331e-01 -1.20601344e+00 -3.99038456e-02 -1.13650286e+00
-2.20408395e-01 1.04881787e+00 4.43475842e-01 -3.13243985e-01
7.36108124e-01 -2.45274920e-02 -2.47864336e-01 -6.72211945e-01
-4.71396774e-01 -3.46574336e-01 9.52775329e-02 5.79895139e-01
7.68647417e-02 2.53831267e-01 3.58212382e-01 2.01157346e-01
-1.99603125e-01 -2.04768684e-02 2.61243641e-01 1.04513180e+00
-6.13943815e-01 -7.78670490e-01 -4.00264144e-01 1.78120315e-01
8.42600390e-02 -1.19895242e-01 -4.00004834e-01 2.10025594e-01
3.18707287e-01 8.16190660e-01 4.74864483e-01 -3.53749633e-01
3.84256691e-01 -3.62102687e-02 7.81816602e-01 -8.56302857e-01
6.68261349e-02 -1.18782796e-01 6.53967038e-02 -1.22759008e+00
-4.07596171e-01 -6.37078285e-01 -1.80442178e+00 1.31841093e-01
2.51261920e-01 -3.95817757e-01 1.30397439e+00 9.93427455e-01
3.03818107e-01 4.38852817e-01 6.48981988e-01 -1.37348068e+00
-6.18381463e-02 -4.87585574e-01 -6.06378913e-01 6.45253062e-01
4.37118560e-01 -9.70775783e-01 -2.92301327e-01 6.17611632e-02] | [8.432531356811523, -2.895667552947998] |
15948948-a938-4bf5-87c7-395b2d3c3bc3 | mde-multi-distance-embeddings-for-link | 1905.10702 | null | https://arxiv.org/abs/1905.10702v8 | https://arxiv.org/pdf/1905.10702v8.pdf | MDE: Multiple Distance Embeddings for Link Prediction in Knowledge Graphs | Over the past decade, knowledge graphs became popular for capturing structured domain knowledge. Relational learning models enable the prediction of missing links inside knowledge graphs. More specifically, latent distance approaches model the relationships among entities via a distance between latent representations. Translating embedding models (e.g., TransE) are among the most popular latent distance approaches which use one distance function to learn multiple relation patterns. However, they are mostly inefficient in capturing symmetric relations since the representation vector norm for all the symmetric relations becomes equal to zero. They also lose information when learning relations with reflexive patterns since they become symmetric and transitive. We propose the Multiple Distance Embedding model (MDE) that addresses these limitations and a framework to collaboratively combine variant latent distance-based terms. Our solution is based on two principles: 1) we use a limit-based loss instead of a margin ranking loss and, 2) by learning independent embedding vectors for each of the terms we can collectively train and predict using contradicting distance terms. We further demonstrate that MDE allows modeling relations with (anti)symmetry, inversion, and composition patterns. We propose MDE as a neural network model that allows us to map non-linear relations between the embedding vectors and the expected output of the score function. Our empirical results show that MDE performs competitively to state-of-the-art embedding models on several benchmark datasets. | ['Jens Lehmann', 'Hamed Shariat Yazdi', 'Damien Graux', 'Afshin Sadeghi'] | 2019-05-25 | null | null | null | null | ['relational-pattern-learning'] | ['knowledge-base'] | [-4.26908135e-02 5.02343655e-01 -7.48951316e-01 -2.26523712e-01
-1.54371023e-01 -6.20504200e-01 7.53651977e-01 3.99218023e-01
-1.20512426e-01 3.45393986e-01 3.66597086e-01 -4.42259669e-01
-8.13533843e-01 -1.17305851e+00 -6.30669236e-01 -3.56683105e-01
-4.69620019e-01 6.82146072e-01 1.12063684e-01 -2.92156696e-01
-6.01833984e-02 4.25456613e-01 -1.29263377e+00 3.53201121e-01
1.07998407e+00 8.33998621e-01 -3.42805445e-01 1.50112301e-01
-2.29537681e-01 1.02923369e+00 -2.38456503e-01 -9.12952483e-01
1.84901640e-01 -3.38723846e-02 -9.63783562e-01 -5.07480621e-01
3.03764313e-01 -9.03509110e-02 -9.37852085e-01 9.10068333e-01
7.55817667e-02 5.28898612e-02 8.94470155e-01 -1.73045218e+00
-1.39321566e+00 8.97426903e-01 -2.91768879e-01 -5.02605774e-02
4.76618588e-01 -5.72812557e-01 1.91297638e+00 -1.20945263e+00
7.07029045e-01 1.15841269e+00 8.05883288e-01 9.90787297e-02
-1.56961691e+00 -5.54277003e-01 1.84173733e-01 6.79791331e-01
-1.58943391e+00 -6.77080527e-02 9.41565752e-01 -5.75250149e-01
1.24603713e+00 2.22130671e-01 5.10647774e-01 9.54116344e-01
1.13897428e-01 6.46934748e-01 5.87321818e-01 -4.85908031e-01
-1.30749837e-01 1.77517995e-01 3.40318382e-01 7.95632124e-01
4.99008566e-01 -9.63211246e-03 -6.82408988e-01 -3.16136211e-01
4.88556087e-01 2.29787812e-01 -3.14557910e-01 -9.85733867e-01
-1.26767159e+00 1.07962894e+00 7.90864825e-01 2.85223633e-01
-1.62472025e-01 1.39724299e-01 1.59380317e-01 5.79065859e-01
5.76939285e-01 5.97541451e-01 -3.75255585e-01 2.19159007e-01
-3.71809363e-01 6.06277771e-03 8.43048871e-01 8.97662103e-01
8.28636408e-01 -4.68816757e-01 -1.65613219e-01 8.50924909e-01
5.07356584e-01 -8.55650082e-02 2.73282647e-01 -7.30622530e-01
8.42980921e-01 1.19037974e+00 -8.12793002e-02 -1.61207497e+00
-2.10307315e-01 -3.92693132e-01 -7.79260218e-01 -1.68942481e-01
1.03410341e-01 2.96451330e-01 -5.22548676e-01 1.81780076e+00
6.37421161e-02 1.37452558e-01 2.32737899e-01 4.81081516e-01
7.67462134e-01 6.29686296e-01 -1.33350834e-01 -9.53036249e-02
1.02635539e+00 -9.87698555e-01 -8.29976022e-01 -6.03642575e-02
9.68429267e-01 -3.91918838e-01 9.28658068e-01 -9.42532271e-02
-9.32475269e-01 -2.64061809e-01 -1.16322565e+00 -2.49330923e-01
-8.51278603e-01 -1.77447811e-01 9.66803789e-01 3.06187868e-01
-1.06057560e+00 7.41731942e-01 -8.05554271e-01 -1.06961243e-01
1.53049484e-01 4.82895404e-01 -6.74954534e-01 -2.17072144e-02
-1.66158605e+00 1.09278643e+00 5.39526045e-01 9.52889770e-02
-1.63420364e-01 -8.75257134e-01 -1.14497817e+00 4.02250379e-01
6.36338234e-01 -6.35219574e-01 3.21241707e-01 -3.47153634e-01
-1.15975130e+00 7.67053485e-01 1.07090376e-01 -3.59695762e-01
2.90818691e-01 -3.09978396e-01 -5.98397553e-01 -5.35788089e-02
7.34011307e-02 3.17934930e-01 2.04473168e-01 -1.20787537e+00
-2.56039828e-01 -2.40294859e-01 3.63456279e-01 8.71649534e-02
-7.36186326e-01 -2.46797800e-01 -3.37064743e-01 -6.02262616e-01
3.32806170e-01 -9.76704836e-01 2.50385076e-01 2.59903401e-01
-6.39132738e-01 -6.35195136e-01 7.94779122e-01 -4.02536064e-01
1.62808037e+00 -1.86905956e+00 5.84936380e-01 6.00003123e-01
6.03592813e-01 3.53750020e-01 -2.67986566e-01 7.93717146e-01
-3.92888635e-01 2.57771134e-01 -8.58469084e-02 -4.92409244e-02
3.23513001e-01 3.50654691e-01 -4.12589818e-01 1.56828746e-01
2.39366189e-01 1.29756200e+00 -1.00485921e+00 -3.36458653e-01
-7.62202293e-02 7.31003642e-01 -6.24341309e-01 1.01455644e-01
6.75869584e-02 -4.03546363e-01 -6.31438866e-02 4.49980795e-01
3.69719684e-01 -6.14669383e-01 7.64109492e-01 -5.26900232e-01
1.50225118e-01 4.96063888e-01 -1.32099473e+00 1.50567544e+00
-4.06305313e-01 5.31622291e-01 -4.89141792e-01 -1.13041759e+00
9.62489486e-01 2.59559929e-01 5.79185486e-01 -3.55830103e-01
-1.89769477e-01 1.60724953e-01 2.84261815e-02 -1.80048674e-01
3.86255562e-01 1.66830063e-01 5.37067093e-02 3.98032069e-01
2.49213487e-01 4.59500253e-01 1.37097120e-01 5.95063746e-01
1.25346518e+00 -7.79141579e-03 2.95384556e-01 3.53694521e-02
3.31748068e-01 -4.33117151e-01 6.09154463e-01 4.97585654e-01
1.32492319e-01 3.03209350e-02 1.03392673e+00 -4.95434195e-01
-8.34701359e-01 -1.41658211e+00 5.72627634e-02 8.57658744e-01
2.32791066e-01 -9.07811105e-01 1.30141422e-01 -9.82129693e-01
4.45634395e-01 5.65277934e-01 -8.71482849e-01 -6.22615635e-01
-4.44936872e-01 -4.55663979e-01 3.67333025e-01 7.63460338e-01
9.82950777e-02 -6.70233965e-01 2.66799927e-01 9.13489610e-02
-1.29324287e-01 -8.96295428e-01 -3.28985751e-01 3.75582367e-01
-6.70728505e-01 -1.08976614e+00 -2.60563850e-01 -8.64045739e-01
6.85308218e-01 2.16659084e-01 1.07104981e+00 1.72315195e-01
-5.85428439e-02 2.72897601e-01 -4.33532506e-01 1.51205078e-01
-1.37905568e-01 1.64023221e-01 2.67344326e-01 6.68790638e-02
5.44757843e-01 -9.06629324e-01 -5.31910121e-01 3.69943321e-01
-8.22029531e-01 -8.32561702e-02 5.48691571e-01 1.04211247e+00
5.01238525e-01 3.03849597e-02 6.02630019e-01 -1.00964177e+00
8.56275558e-01 -6.73410892e-01 -4.72122073e-01 9.14574802e-01
-1.08510172e+00 4.56212491e-01 4.78223383e-01 -5.10995924e-01
-5.58570445e-01 -2.32567415e-01 4.03739929e-01 -6.45598471e-01
4.88949001e-01 9.11347747e-01 -1.99082255e-01 3.60351875e-02
5.20054579e-01 -3.61003801e-02 -1.88015193e-01 -2.42443368e-01
8.47591937e-01 3.84297460e-01 1.72419548e-01 -5.58959067e-01
1.07362449e+00 2.25996748e-01 2.66775012e-01 -2.99164832e-01
-9.40448940e-01 -4.54969049e-01 -8.34631801e-01 2.66339511e-01
6.88742757e-01 -6.98387921e-01 -8.13723028e-01 -2.21878484e-01
-1.17382705e+00 1.66697893e-02 -3.41044217e-01 6.44303679e-01
-3.48914146e-01 4.94190097e-01 -7.02985048e-01 -5.94749033e-01
-1.07481703e-01 -7.30605185e-01 5.98429203e-01 -2.20955819e-01
-3.94517869e-01 -1.30863142e+00 3.15251857e-01 1.85030043e-01
2.87785232e-01 1.93678081e-01 1.55982614e+00 -7.71997452e-01
-6.68502510e-01 -2.43372872e-01 -3.81088585e-01 1.76053867e-01
3.62789750e-01 -3.03184590e-03 -4.82432574e-01 -5.16918488e-02
-7.09071398e-01 -3.56363684e-01 1.01761627e+00 -1.32498488e-01
9.29184675e-01 -6.18258893e-01 -5.98916531e-01 7.40619361e-01
1.37601531e+00 1.59363765e-02 5.90298653e-01 3.20852637e-01
1.04968774e+00 7.32923269e-01 2.85613477e-01 -2.89148688e-02
6.64935529e-01 9.11641717e-01 2.27881119e-01 1.91670001e-01
7.49552622e-02 -6.42907262e-01 2.48100579e-01 1.20652306e+00
-2.53681928e-01 -1.75588921e-01 -1.01588190e+00 6.24867141e-01
-2.07753539e+00 -1.18312776e+00 -4.95360121e-02 2.10794282e+00
1.04709470e+00 2.57116228e-01 -1.33140013e-01 2.14058265e-01
6.03705883e-01 2.69388974e-01 -3.40171129e-01 -3.76189202e-01
-1.82701379e-01 1.61837146e-01 2.67798960e-01 7.07428336e-01
-1.03111410e+00 8.53045285e-01 5.97615480e+00 5.81017494e-01
-6.38407528e-01 3.15433443e-02 -5.10880835e-02 1.37924716e-01
-8.30866218e-01 2.19866887e-01 -6.55821741e-01 1.87642589e-01
5.42852998e-01 -3.98917407e-01 3.58357191e-01 7.04824865e-01
-5.87742865e-01 5.63717663e-01 -1.49785089e+00 9.82105017e-01
5.77766486e-02 -1.49931920e+00 4.34861213e-01 1.91048741e-01
9.35436368e-01 -1.67349428e-01 1.87078238e-01 4.87942576e-01
4.65544343e-01 -1.07435966e+00 2.46164009e-01 6.24231637e-01
8.19958270e-01 -8.24121773e-01 6.94634974e-01 4.38659154e-02
-1.39566672e+00 -1.71176821e-01 -4.24067646e-01 3.98733243e-02
-5.74858300e-02 4.99846876e-01 -6.23776793e-01 8.74632418e-01
2.55218685e-01 9.58878398e-01 -5.17723799e-01 7.01473057e-01
-6.08759701e-01 2.87011474e-01 -1.39745116e-01 1.07144088e-01
2.81927586e-02 -3.50912392e-01 3.94360632e-01 8.65676045e-01
2.85520434e-01 -2.31975436e-01 -7.50424480e-03 1.04358006e+00
-3.94397378e-01 4.05151136e-02 -9.87855792e-01 -1.83999702e-01
7.58298874e-01 9.71724391e-01 -4.25106496e-01 -7.48879742e-03
-7.47263968e-01 9.02438521e-01 9.06239629e-01 4.44521248e-01
-6.26784801e-01 -6.40747726e-01 7.56479502e-01 -4.71063377e-03
3.00645232e-01 -2.77455896e-01 -4.27565910e-02 -1.39038253e+00
3.02256703e-01 -3.60041320e-01 6.26261532e-01 -4.66898292e-01
-1.69296777e+00 5.14086246e-01 1.10381626e-01 -1.15144038e+00
-1.58428371e-01 -6.90620363e-01 -3.92592251e-01 6.95299864e-01
-1.70572114e+00 -1.28737211e+00 9.76767857e-03 4.92461503e-01
-1.98713735e-01 -3.02864611e-01 1.18931556e+00 4.48597193e-01
-4.45849806e-01 9.04724360e-01 2.11461067e-01 4.69060183e-01
5.70894778e-01 -1.39647639e+00 2.31405601e-01 4.61637408e-01
6.08004272e-01 1.05353701e+00 3.12007487e-01 -6.80891156e-01
-1.21963346e+00 -1.02846348e+00 1.47190249e+00 -6.84864581e-01
1.03276050e+00 -5.03378391e-01 -1.04921377e+00 1.14058661e+00
-1.00724645e-01 2.79949456e-01 9.98704374e-01 8.30282688e-01
-9.52873230e-01 -2.69724697e-01 -5.29571772e-01 5.97534657e-01
1.27145147e+00 -9.91022348e-01 -9.43820238e-01 2.55757928e-01
9.19622540e-01 1.45334348e-01 -1.25338686e+00 6.87108397e-01
7.29443908e-01 -7.13304639e-01 1.23806798e+00 -8.98360610e-01
5.60151517e-01 -2.70024389e-01 -3.00086498e-01 -1.35382676e+00
-6.04309797e-01 -3.79224747e-01 -9.10495698e-01 1.22597194e+00
8.02513242e-01 -7.88349330e-01 8.51838827e-01 6.12059176e-01
2.20400035e-01 -1.24489701e+00 -7.77264297e-01 -1.15362251e+00
3.38680707e-02 -2.86667142e-02 5.82896411e-01 1.56506646e+00
4.34621155e-01 6.29822135e-01 -4.21884388e-01 1.61742941e-01
4.93412375e-01 2.97568470e-01 4.46100593e-01 -1.61996472e+00
-3.31551939e-01 -4.39616114e-01 -8.75953972e-01 -9.20974553e-01
4.66969192e-01 -1.49381626e+00 -6.82812631e-01 -1.90971911e+00
2.24852264e-01 -5.95772743e-01 -6.40584648e-01 7.62877524e-01
-1.87996790e-01 -1.31571308e-01 -1.16951717e-02 3.50408643e-01
-7.27114677e-01 8.83655190e-01 8.64118695e-01 -3.65044266e-01
-2.65479594e-01 -2.56369919e-01 -6.73657894e-01 6.33351982e-01
5.47504723e-01 -6.81339145e-01 -8.19808900e-01 -3.41909468e-01
1.01357710e+00 -1.79870501e-01 3.41351509e-01 -5.52317679e-01
5.62876105e-01 -7.52513409e-02 1.38264224e-01 -4.57019120e-01
4.43985254e-01 -9.46964741e-01 1.08602330e-01 2.84126043e-01
-5.93020797e-01 -1.19878776e-01 -3.52503628e-01 8.12274456e-01
-4.51514453e-01 8.69060308e-02 1.31269753e-01 3.30796778e-01
-5.59229672e-01 4.03535128e-01 3.12847406e-01 8.87008607e-02
9.69940722e-01 -6.95565045e-02 -6.12874269e-01 -2.72440016e-01
-8.47755492e-01 2.59171128e-01 1.23973146e-01 6.92187786e-01
7.65360057e-01 -1.96066213e+00 -5.45471430e-01 3.70391421e-02
5.36836147e-01 -1.67664573e-01 -7.92486146e-02 7.34091282e-01
-2.11949036e-01 5.65831065e-01 -7.57658482e-02 -1.98784783e-01
-1.20487058e+00 7.67581761e-01 1.58186153e-01 -7.34352648e-01
-5.92800856e-01 9.46258664e-01 1.08342551e-01 -7.52871454e-01
3.59647304e-01 -2.47596219e-01 -2.56082296e-01 3.02684426e-01
2.13212445e-01 3.56373399e-01 -1.04561150e-01 -5.90470016e-01
-5.65267742e-01 5.74907899e-01 -1.76780343e-01 1.44512996e-01
1.50126719e+00 2.33463719e-01 -2.39373296e-01 4.87338901e-01
1.48337221e+00 4.65168320e-02 -8.03497016e-01 -7.81833231e-01
3.04831564e-01 -4.19125289e-01 -1.34874836e-01 -5.88539660e-01
-7.93679178e-01 7.39028335e-01 6.99128658e-02 3.58800501e-01
7.27285922e-01 3.17864001e-01 4.48673218e-01 6.94970071e-01
2.14229599e-01 -9.32052433e-01 3.26193094e-01 5.59642255e-01
8.65071416e-01 -1.02939332e+00 2.04749882e-01 -6.49306297e-01
-4.84507978e-01 1.05466580e+00 5.69250882e-01 -6.62719924e-03
9.92448628e-01 -5.02467304e-02 -3.17464828e-01 -4.88616347e-01
-9.69231665e-01 -1.80953622e-01 8.12022924e-01 5.50500512e-01
4.87417847e-01 1.61908656e-01 -3.30861151e-01 5.12426138e-01
-1.37284592e-01 -2.85558224e-01 4.90838960e-02 7.11407185e-01
-6.77382797e-02 -1.49511802e+00 1.44568756e-01 5.18072784e-01
-9.47834328e-02 -1.45435184e-01 -5.34084380e-01 7.59347677e-01
1.76560089e-01 7.92803407e-01 -2.57084537e-02 -7.40626395e-01
4.35253918e-01 2.21310392e-01 3.55717510e-01 -7.32620180e-01
-7.08772019e-02 -6.10559464e-01 9.73292217e-02 -4.32733089e-01
-3.47462028e-01 -2.41067335e-01 -1.03510809e+00 -3.45262140e-01
-5.87959766e-01 3.29095930e-01 3.37805510e-01 7.13302433e-01
4.86298263e-01 4.42552447e-01 6.22796953e-01 -1.26373678e-01
-4.96630162e-01 -6.67724967e-01 -7.18434155e-01 6.49398506e-01
5.57195805e-02 -1.04731846e+00 -3.75764996e-01 -3.00206572e-01] | [8.722485542297363, 7.80568790435791] |
8b4ebd5b-6157-4bf5-baa8-860ca055354f | commonly-uncommon-semantic-sparsity-in | 1612.00901 | null | http://arxiv.org/abs/1612.00901v1 | http://arxiv.org/pdf/1612.00901v1.pdf | Commonly Uncommon: Semantic Sparsity in Situation Recognition | Semantic sparsity is a common challenge in structured visual classification
problems; when the output space is complex, the vast majority of the possible
predictions are rarely, if ever, seen in the training set. This paper studies
semantic sparsity in situation recognition, the task of producing structured
summaries of what is happening in images, including activities, objects and the
roles objects play within the activity. For this problem, we find empirically
that most object-role combinations are rare, and current state-of-the-art
models significantly underperform in this sparse data regime. We avoid many
such errors by (1) introducing a novel tensor composition function that learns
to share examples across role-noun combinations and (2) semantically augmenting
our training data with automatically gathered examples of rarely observed
outputs using web data. When integrated within a complete CRF-based structured
prediction model, the tensor-based approach outperforms existing state of the
art by a relative improvement of 2.11% and 4.40% on top-5 verb and noun-role
accuracy, respectively. Adding 5 million images with our semantic augmentation
techniques gives further relative improvements of 6.23% and 9.57% on top-5 verb
and noun-role accuracy. | ['Vicente Ordonez', 'Luke Zettlemoyer', 'Ali Farhadi', 'Mark Yatskar'] | 2016-12-03 | commonly-uncommon-semantic-sparsity-in-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Yatskar_Commonly_Uncommon_Semantic_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Yatskar_Commonly_Uncommon_Semantic_CVPR_2017_paper.pdf | cvpr-2017-7 | ['grounded-situation-recognition', 'situation-recognition'] | ['computer-vision', 'computer-vision'] | [ 6.60076201e-01 1.75843403e-01 -4.36620414e-01 -5.33503473e-01
-7.77597308e-01 -5.93063414e-01 8.47949922e-01 2.40843371e-01
-3.63170713e-01 6.93659365e-01 7.84928799e-01 2.83777881e-02
2.40146428e-01 -3.86219144e-01 -1.06799400e+00 -4.90680963e-01
-7.37137999e-03 5.96675634e-01 2.55104929e-01 2.13319659e-02
1.08448677e-01 5.22061959e-02 -1.88157868e+00 9.82328832e-01
6.82368040e-01 1.23198986e+00 3.97111177e-01 4.62623179e-01
-3.76099437e-01 1.42936158e+00 -4.59402502e-01 -6.08716547e-01
2.58692265e-01 -1.25391632e-01 -7.50496030e-01 7.37020791e-01
9.35934603e-01 -1.54327095e-01 -3.66313875e-01 8.69893610e-01
-6.92394003e-02 1.34842694e-01 5.02439916e-01 -1.42991161e+00
-6.34811282e-01 7.10770488e-01 -4.18517113e-01 1.24177016e-01
4.18051034e-01 1.09390557e-01 1.30895746e+00 -1.27194762e+00
9.02834535e-01 1.28029191e+00 4.45331275e-01 3.47185463e-01
-1.26187992e+00 -4.83216017e-01 4.21027154e-01 4.87029195e-01
-1.19809055e+00 -5.14853239e-01 5.94152927e-01 -5.71919322e-01
1.40139377e+00 1.89964220e-01 6.87445879e-01 1.17902827e+00
-2.80218691e-01 1.09775102e+00 1.00891161e+00 -2.94137865e-01
1.95509598e-01 6.65746778e-02 7.81649351e-03 6.91154242e-01
2.94958055e-01 -3.93492669e-01 -8.82408857e-01 -2.50206828e-01
5.67496419e-01 1.84941530e-01 7.39777461e-02 -3.33947837e-01
-1.58559930e+00 6.03212774e-01 3.06208313e-01 2.39518702e-01
-3.34290445e-01 4.48028862e-01 4.18706864e-01 8.44175667e-02
5.56111217e-01 3.78131777e-01 -4.82764900e-01 -1.92817420e-01
-9.32227910e-01 4.74271566e-01 8.19733024e-01 1.19937897e+00
7.01005220e-01 2.18153015e-01 -2.97281504e-01 1.00151324e+00
-5.34671992e-02 4.91790205e-01 3.21531713e-01 -1.17191732e+00
7.70031810e-01 7.78183699e-01 2.02899247e-01 -9.00325239e-01
7.68681569e-03 -4.71322268e-01 -5.57965040e-01 -3.18481803e-01
5.77832699e-01 2.81213224e-01 -1.22904038e+00 1.52961195e+00
3.00396740e-01 1.62840798e-01 -2.49599237e-02 8.91609550e-01
7.67432094e-01 8.08611751e-01 6.08245492e-01 -1.45172819e-01
1.63332999e+00 -1.15515399e+00 -7.48053610e-01 -6.25006557e-01
5.53855062e-01 -7.24527776e-01 1.29715121e+00 3.18283767e-01
-8.56413007e-01 -4.04030561e-01 -7.45743096e-01 -2.27142870e-01
-2.26538122e-01 1.67800248e-01 9.56735611e-01 1.89006865e-01
-8.98624420e-01 3.51421565e-01 -5.78700423e-01 -4.18586940e-01
8.60451341e-01 1.30890608e-01 -6.24621749e-01 -5.94619751e-01
-6.90381765e-01 7.36105025e-01 3.86726946e-01 -3.40024948e-01
-1.23009777e+00 -9.06740427e-01 -1.04951620e+00 1.54572144e-01
8.85602653e-01 -6.08091116e-01 1.30017900e+00 -1.09329832e+00
-5.78689873e-01 8.95217836e-01 -5.85397780e-01 -6.36943102e-01
2.33428195e-01 -2.96622306e-01 -1.84674725e-01 3.40437949e-01
5.21771610e-01 1.01493919e+00 8.76295030e-01 -1.42538571e+00
-7.69403994e-01 -3.00177723e-01 5.83033785e-02 6.73868716e-01
-4.04827207e-01 1.87331676e-01 -6.40541911e-01 -8.13296735e-01
4.69074734e-02 -7.67336249e-01 -4.21929121e-01 2.26504192e-01
-3.52456003e-01 -3.94549370e-01 6.83788776e-01 -7.22087026e-01
7.87668049e-01 -2.04010296e+00 -1.34040937e-01 -1.27722189e-01
2.37908617e-01 -1.10486291e-01 -1.26232207e-01 3.24135959e-01
4.76042554e-02 1.14881299e-01 -3.58414531e-01 -6.61950052e-01
-1.27668545e-01 5.16876400e-01 -4.97055739e-01 8.80387500e-02
2.70267874e-01 8.53310347e-01 -1.01852858e+00 -6.58080041e-01
1.12508535e-01 1.33666739e-01 -4.82603043e-01 8.86696205e-02
-6.13326430e-01 2.00504616e-01 -2.50276983e-01 9.38824654e-01
1.90945357e-01 -7.62252390e-01 4.98610973e-01 -2.85722345e-01
1.19265772e-01 4.60783213e-01 -9.38993096e-01 1.76946807e+00
-3.73198539e-01 6.44299746e-01 -2.27174193e-01 -1.22022629e+00
5.85152924e-01 3.15472811e-01 5.50539434e-01 -6.51049852e-01
-2.61388332e-01 -8.83029192e-04 -3.79247107e-02 -5.32943785e-01
5.83006561e-01 -2.90607184e-01 -1.44600019e-01 2.25932881e-01
3.90163362e-01 -7.18029216e-02 4.21142668e-01 5.52194834e-01
1.24822223e+00 6.39526248e-02 3.78634036e-01 1.11651635e-02
8.83677676e-02 5.48723638e-01 5.36114752e-01 9.03721333e-01
3.47411595e-02 5.97780466e-01 6.75595641e-01 -6.45595551e-01
-1.42897999e+00 -9.98619497e-01 2.24255890e-01 1.40501976e+00
4.28990088e-02 -5.21627665e-01 -3.86183947e-01 -9.12382782e-01
1.86134622e-01 6.84393525e-01 -4.69389290e-01 3.08673829e-01
-6.24558985e-01 -5.19389570e-01 3.09059173e-01 8.16017866e-01
3.86745095e-01 -1.09728396e+00 -2.59805948e-01 1.84956908e-01
-4.56251472e-01 -1.81701231e+00 -3.32896560e-01 2.59512216e-01
-9.85756159e-01 -1.01454294e+00 -4.05564755e-01 -8.97176683e-01
7.75187254e-01 3.83595049e-01 1.45510805e+00 9.06631052e-02
-2.04436138e-01 3.79870534e-01 -4.51901764e-01 -3.81073028e-01
-8.28705952e-02 -3.40210766e-01 -1.44491076e-01 2.36893684e-01
2.88098991e-01 -4.03332651e-01 -3.29548061e-01 2.27813482e-01
-6.68607354e-01 3.79968107e-01 7.16439247e-01 8.47048938e-01
7.97899246e-01 -8.58216360e-02 2.86266088e-01 -1.03944051e+00
2.79004611e-02 -6.40709043e-01 -2.61706591e-01 1.98026091e-01
-4.72241670e-01 1.85023504e-03 5.80713272e-01 -6.64312065e-01
-1.19695950e+00 5.70581973e-01 2.15148896e-01 -7.21255720e-01
-4.01797682e-01 3.83585811e-01 -8.68864581e-02 2.27957323e-01
7.35884190e-01 5.25785327e-01 -2.47521579e-01 -5.24915874e-01
4.21340883e-01 3.33605379e-01 5.24955511e-01 -4.97725189e-01
6.27727330e-01 8.10275018e-01 -2.59969890e-01 -9.87624049e-01
-1.48878562e+00 -8.45518649e-01 -3.28752190e-01 -2.36012384e-01
8.00424755e-01 -1.38504493e+00 -2.40757599e-01 2.07614988e-01
-1.17233002e+00 -3.27045351e-01 -5.42060435e-01 2.47115552e-01
-5.33471584e-01 4.22590733e-01 -4.09910858e-01 -7.24526286e-01
-8.75512809e-02 -1.01931119e+00 1.27445936e+00 -2.42639631e-01
-2.03838065e-01 -6.55879796e-01 -5.58846354e-01 9.95688260e-01
1.57062054e-01 1.67621121e-01 9.02603090e-01 -8.18700910e-01
-8.03886652e-01 -3.94016206e-02 -3.74411821e-01 3.05034339e-01
-7.32051507e-02 -6.33633673e-01 -8.16151083e-01 9.42059010e-02
-1.29427120e-01 -6.36988997e-01 9.27790582e-01 2.38402084e-01
1.59221673e+00 -6.49925292e-01 -3.33213389e-01 1.01283237e-01
1.38808382e+00 4.73821573e-02 5.37945330e-01 -7.36579970e-02
1.10649192e+00 8.05023611e-01 7.62149811e-01 4.52309787e-01
6.62157238e-01 6.72920644e-01 5.35338223e-01 2.32805554e-02
-3.36617529e-01 -4.89977151e-01 2.26553351e-01 3.74676347e-01
-4.28086556e-02 -2.59178489e-01 -9.81316507e-01 1.09948754e+00
-2.06830812e+00 -1.19590938e+00 -2.23051995e-01 1.88988721e+00
9.29310143e-01 2.24141017e-01 1.90014571e-01 7.11166188e-02
7.14466393e-01 4.16270465e-01 -5.37302077e-01 1.38261214e-01
-2.16575503e-01 1.28494501e-01 6.36643767e-01 1.49242833e-01
-1.27594960e+00 1.11291909e+00 6.20858240e+00 1.02264440e+00
-6.46499217e-01 3.93906534e-01 7.32689500e-01 -2.41134956e-01
-3.91654998e-01 2.59125948e-01 -8.53901029e-01 3.03772181e-01
6.80978000e-01 1.02615766e-01 4.16771650e-01 1.18288028e+00
1.27778947e-01 -3.64552140e-01 -1.25334501e+00 1.30339622e+00
5.08367896e-01 -1.62648642e+00 4.32550281e-01 2.96483338e-02
8.45486581e-01 -3.48712094e-02 -5.00545055e-02 2.47011676e-01
3.56257945e-01 -1.28221369e+00 1.13244188e+00 2.58873254e-01
8.95424068e-01 -1.88817695e-01 4.81006175e-01 4.17374194e-01
-1.19484830e+00 -2.91552633e-01 -2.58843452e-01 -3.09398472e-01
1.71937019e-01 7.15772510e-01 -1.00423026e+00 1.19444065e-01
7.99570024e-01 1.15014744e+00 -5.09909928e-01 9.58416522e-01
-3.49010140e-01 9.33358431e-01 -3.87618899e-01 -1.29532844e-01
2.97577590e-01 3.55266690e-01 4.68986928e-01 1.12888920e+00
2.16789637e-02 1.27300352e-01 4.60582286e-01 6.75612450e-01
-3.54967237e-01 3.29994336e-02 -8.16151857e-01 -1.48533553e-01
4.71826345e-01 1.15083611e+00 -7.77226210e-01 -8.73134673e-01
-6.98137343e-01 8.15175831e-01 3.50766271e-01 3.22522432e-01
-6.99857593e-01 1.12778828e-01 6.20715261e-01 3.79121572e-01
5.48048139e-01 -2.08416000e-01 -5.57114959e-01 -1.33220351e+00
3.55188400e-01 -8.73040676e-01 4.85967845e-01 -1.03523576e+00
-1.44616044e+00 1.94762930e-01 2.15051800e-01 -1.24818635e+00
-2.50987820e-02 -5.23364723e-01 -1.65712729e-01 4.61117834e-01
-1.33661902e+00 -1.40205133e+00 -3.72762352e-01 6.04684591e-01
1.14839721e+00 -1.74907893e-01 6.21027291e-01 1.93891183e-01
2.97576957e-03 2.76224732e-01 -3.65804166e-01 4.34779339e-02
3.33582193e-01 -1.31361759e+00 3.40413332e-01 7.97220051e-01
5.46919048e-01 3.12254608e-01 6.27569318e-01 -8.44213068e-01
-1.23156786e+00 -1.33516669e+00 1.31003988e+00 -7.77309775e-01
8.05802464e-01 -7.87202239e-01 -7.64443278e-01 8.08469534e-01
4.77460399e-02 4.14321452e-01 6.78329349e-01 1.63506180e-01
-6.92619443e-01 4.12523448e-02 -1.09271491e+00 6.80289090e-01
1.57955253e+00 -6.71041787e-01 -7.62437522e-01 7.69618809e-01
9.65564549e-01 -2.90907234e-01 -6.47928178e-01 1.59627974e-01
5.07115066e-01 -6.04302168e-01 1.04673576e+00 -7.55511522e-01
8.62195671e-01 -2.81370372e-01 -5.71054697e-01 -8.94611418e-01
-2.21390232e-01 -2.45173320e-01 -4.62093294e-01 8.09080362e-01
4.42636758e-01 4.94903070e-04 1.17426825e+00 4.36682105e-01
-2.04523653e-01 -8.04673076e-01 -7.75590301e-01 -7.50727713e-01
-4.35244113e-01 -8.37915719e-01 1.81761459e-01 9.40922141e-01
-2.36268163e-01 4.19805080e-01 -6.69781506e-01 6.53055981e-02
7.78787017e-01 3.73609550e-02 5.64777851e-01 -9.50994194e-01
-2.53221691e-01 5.80781847e-02 -5.08909523e-01 -1.09937835e+00
1.95587233e-01 -9.86114323e-01 -7.15990439e-02 -1.80946779e+00
6.85417354e-01 -3.79860878e-01 4.07677330e-02 1.01235509e+00
-6.41265959e-02 4.02217805e-01 3.24814290e-01 5.36438346e-01
-1.03698182e+00 5.32795191e-01 1.27122915e+00 -3.01212698e-01
1.12950593e-01 -4.14070219e-01 -7.86071420e-01 8.03467393e-01
4.25424993e-01 -7.26761162e-01 -4.89559829e-01 -5.78954697e-01
4.04602587e-02 2.62190588e-02 7.58785009e-01 -9.65836883e-01
1.69245452e-02 -3.11260790e-01 5.50020337e-01 -5.84878147e-01
6.14411652e-01 -8.90371382e-01 1.41625358e-02 1.85401067e-01
-5.62488079e-01 -2.01111451e-01 -5.33812977e-02 9.55062270e-01
-3.46275032e-01 5.15593402e-02 3.37302148e-01 -5.05946159e-01
-1.11671114e+00 2.66433477e-01 -4.75871354e-01 2.69089967e-01
1.05753410e+00 -4.21657652e-01 -3.06664079e-01 -4.01774496e-01
-9.21022892e-01 2.75386095e-01 3.18213433e-01 4.86734986e-01
7.19604969e-01 -1.25241792e+00 -5.43380678e-01 -1.61829948e-01
4.47083861e-01 1.62077665e-01 3.34051371e-01 6.62175536e-01
-1.48441374e-01 2.93433845e-01 8.00212771e-02 -7.66329706e-01
-1.32552254e+00 3.76832962e-01 1.67328175e-02 -4.00408626e-01
-5.40056169e-01 9.35139775e-01 3.21319968e-01 -1.79230914e-01
3.76211584e-01 -3.56706738e-01 -1.69247225e-01 1.62896346e-02
3.83427650e-01 1.80508364e-02 -4.66274534e-04 -7.02602804e-01
-4.27152693e-01 2.11245552e-01 -6.45463765e-02 -1.12542994e-01
1.52840912e+00 2.87843049e-01 1.05000138e-02 4.65338677e-01
9.65786755e-01 -2.92958438e-01 -1.29459023e+00 -4.81532633e-01
1.46573275e-01 -6.26709700e-01 -2.01140687e-01 -9.96159494e-01
-7.72727013e-01 6.27245009e-01 1.16418362e-01 -2.25982629e-02
9.73214447e-01 5.50971150e-01 6.23396754e-01 5.58402777e-01
4.57934558e-01 -1.01185966e+00 5.53010106e-01 4.83329922e-01
9.88290370e-01 -1.43697226e+00 1.10391706e-01 -8.69706392e-01
-1.14412093e+00 5.17129064e-01 7.42067933e-01 -1.38364106e-01
4.67606455e-01 9.98489335e-02 -3.78724754e-01 -4.47892070e-01
-1.14841914e+00 -2.63718188e-01 3.51840943e-01 6.62173927e-01
3.22989583e-01 1.31158963e-01 1.43564316e-02 4.26441491e-01
-4.49495837e-02 -1.77564129e-01 3.28816831e-01 1.11144662e+00
-4.19674128e-01 -7.95646131e-01 -2.94500828e-01 9.90827262e-01
-5.46169996e-01 -4.13908362e-01 -2.58528680e-01 5.31355679e-01
2.92988628e-01 8.46753120e-01 2.98528254e-01 -2.10838646e-01
2.78040349e-01 3.02673012e-01 3.83794725e-01 -1.17705476e+00
-5.09681880e-01 -3.35528143e-02 6.17183864e-01 -8.66961002e-01
-7.11362720e-01 -8.53278041e-01 -1.16342783e+00 1.40665248e-01
5.84056117e-02 -1.98699504e-01 6.46444559e-01 1.12856352e+00
3.31729978e-01 2.27585405e-01 1.57250538e-01 -6.99608207e-01
-5.22912562e-01 -8.48964512e-01 -3.77517939e-01 7.91420639e-01
8.46546516e-02 -7.56580770e-01 -2.84458220e-01 6.49180949e-01] | [10.342270851135254, 1.4233859777450562] |
a1d54ab5-a424-4fff-853e-0de8e52a0594 | evaluation-discrepancy-discovery-a-sentence | 2101.09079 | null | https://arxiv.org/abs/2101.09079v1 | https://arxiv.org/pdf/2101.09079v1.pdf | Evaluation Discrepancy Discovery: A Sentence Compression Case-study | Reliable evaluation protocols are of utmost importance for reproducible NLP research. In this work, we show that sometimes neither metric nor conventional human evaluation is sufficient to draw conclusions about system performance. Using sentence compression as an example task, we demonstrate how a system can game a well-established dataset to achieve state-of-the-art results. In contrast with the results reported in previous work that showed correlation between human judgements and metric scores, our manual analysis of state-of-the-art system outputs demonstrates that high metric scores may only indicate a better fit to the data, but not better outputs, as perceived by humans. | ['Yevgeniy Puzikov'] | 2021-01-22 | null | null | null | null | ['sentence-compression'] | ['natural-language-processing'] | [ 1.72642738e-01 2.79862314e-01 1.65587753e-01 -6.19541109e-01
-1.07265389e+00 -8.62845063e-01 8.01404119e-01 8.20922911e-01
-8.88446927e-01 9.37336326e-01 3.26782048e-01 -4.17692184e-01
-1.76836491e-01 -4.69269037e-01 -4.20206577e-01 -1.34500951e-01
3.00860792e-01 7.37865329e-01 3.26837271e-01 -3.06771338e-01
6.58314586e-01 3.37900937e-01 -1.48124599e+00 6.65209651e-01
7.29836464e-01 6.20602489e-01 1.19155094e-01 8.98809314e-01
-1.59667641e-01 7.02446342e-01 -1.12778759e+00 -8.18818152e-01
1.75699126e-02 -4.95663673e-01 -1.06548381e+00 -6.01020396e-01
3.92875791e-01 -5.83273359e-02 1.11766934e-01 9.66991186e-01
6.16316795e-01 -9.09857303e-02 7.30822682e-01 -9.85715330e-01
-7.07100093e-01 8.40792775e-01 1.09939024e-01 3.66766155e-01
9.07615066e-01 2.82456249e-01 1.12508106e+00 -5.86230516e-01
8.39948952e-01 1.03447056e+00 6.63986742e-01 2.37055093e-01
-1.38920891e+00 -3.93628895e-01 -4.21727985e-01 1.79585904e-01
-1.34614444e+00 -6.86855912e-01 1.60382763e-01 -2.89878905e-01
1.39553952e+00 4.43071067e-01 4.97567594e-01 8.85014653e-01
2.33669765e-02 3.92852813e-01 1.06088686e+00 -7.20644593e-01
1.82157055e-01 4.48939919e-01 8.35876018e-02 2.72004008e-01
6.51149213e-01 -1.93328887e-01 -5.66738844e-01 -1.55796587e-01
4.41776961e-02 -6.86551809e-01 -4.75268900e-01 6.04943596e-02
-1.34426689e+00 5.41980267e-01 1.06498599e-01 8.70420873e-01
-4.51651096e-01 -1.19489245e-01 4.64321375e-01 4.98776108e-01
2.95131862e-01 1.09205115e+00 -5.55995643e-01 -9.16492105e-01
-1.38531125e+00 5.10911942e-01 1.35721421e+00 6.47827506e-01
2.04099268e-01 -5.00955939e-01 -2.93360770e-01 6.63807154e-01
1.48641199e-01 2.93155730e-01 5.89881361e-01 -1.25586510e+00
6.14998341e-01 6.21742368e-01 5.04272878e-01 -1.09252119e+00
-2.32259899e-01 -3.44863087e-01 -2.92960167e-01 1.73137337e-01
7.46903300e-01 1.40616149e-01 -5.40866971e-01 1.58919561e+00
-3.90725911e-01 -6.54820263e-01 3.26019108e-01 7.30200052e-01
6.54455900e-01 4.62915182e-01 5.65164201e-02 -5.24603367e-01
1.04234183e+00 -5.26304424e-01 -8.38508427e-01 -3.38410288e-02
9.43249226e-01 -1.16430295e+00 1.50164652e+00 6.19464278e-01
-1.34981167e+00 -4.81185734e-01 -1.13476789e+00 -1.04073189e-01
-5.15743494e-01 -2.35511139e-02 3.45142126e-01 6.75903022e-01
-1.08415103e+00 9.79262948e-01 -6.87512994e-01 -8.52853298e-01
3.60936373e-02 -7.89242331e-03 -3.18786621e-01 5.76816238e-02
-1.07849097e+00 1.48527455e+00 5.49972117e-01 -6.13181703e-02
-4.17835116e-01 -5.13699651e-01 -4.42503095e-01 1.18796572e-01
2.47018173e-01 -5.32510638e-01 1.66361260e+00 -2.65250742e-01
-1.25280869e+00 9.41221118e-01 -2.16340661e-01 -7.36504316e-01
7.07464039e-01 -1.14206731e-01 -5.48753858e-01 2.53472149e-01
-7.40804747e-02 5.60688496e-01 1.16401818e-02 -1.11726475e+00
-5.07945180e-01 -7.19076544e-02 2.22710986e-02 3.49972606e-01
-9.93571281e-02 3.62802267e-01 -1.12130538e-01 -2.29656443e-01
5.26344515e-02 -5.84240913e-01 5.94706610e-02 -1.83767289e-01
-1.69103891e-01 -3.68325651e-01 3.43656540e-01 -8.13551545e-01
1.64526033e+00 -1.76271856e+00 -1.25534058e-01 5.39592244e-02
7.92012066e-02 4.42406565e-01 1.35638922e-01 7.98696041e-01
2.82287419e-01 6.39605224e-01 -2.52189755e-01 -3.77759218e-01
3.19780231e-01 1.39724553e-01 -9.75014120e-02 3.45019817e-01
1.88280493e-01 7.42180586e-01 -1.06932330e+00 -7.54481375e-01
7.43895769e-02 2.73503095e-01 -2.38020137e-01 2.14932725e-01
-7.59614110e-02 -2.82941367e-02 -1.52926622e-02 2.65722483e-01
3.12472701e-01 -1.87172800e-01 1.01446368e-01 -2.75420733e-02
-1.33243054e-01 7.25721657e-01 -9.93257642e-01 1.59475565e+00
-9.14059132e-02 8.94621670e-01 -2.46245265e-01 -7.16491044e-01
8.48850846e-01 4.60660636e-01 -5.32998852e-02 -8.18724871e-01
1.37317628e-01 3.53515863e-01 4.50873524e-01 -6.47884130e-01
8.91348660e-01 7.50762410e-03 1.07724546e-03 6.93552077e-01
5.70859686e-02 -6.28715336e-01 7.17082858e-01 3.41357231e-01
1.11416817e+00 -1.88832939e-01 4.54967797e-01 -3.42471361e-01
4.61409509e-01 1.84643820e-01 2.85453022e-01 8.64404023e-01
-3.28075707e-01 7.80553460e-01 6.20113194e-01 -6.92231804e-02
-1.27926850e+00 -9.55790877e-01 -2.21541896e-01 8.94246578e-01
-2.65731752e-01 -8.12487483e-01 -1.08490038e+00 -5.70120037e-01
-2.52293169e-01 1.25814760e+00 -3.83681744e-01 1.37212485e-01
-1.94221199e-01 -5.20762742e-01 1.05062699e+00 3.17919821e-01
9.58320946e-02 -1.20747447e+00 -9.62454677e-01 2.17771307e-01
-4.39490497e-01 -1.17185640e+00 3.28085721e-02 -3.33696976e-02
-7.00889468e-01 -9.13830578e-01 -6.37730896e-01 -5.42315125e-01
5.07884204e-01 -4.24671471e-02 1.42032897e+00 2.98669279e-01
7.57256299e-02 2.41281897e-01 -5.05174935e-01 -5.36068678e-01
-9.99381483e-01 9.46867689e-02 1.02052286e-01 -9.73921061e-01
7.99854696e-01 -2.71539509e-01 -5.74716210e-01 2.15075552e-01
-1.03587186e+00 -1.90711856e-01 6.67543352e-01 5.28020263e-01
3.41827452e-01 -8.96810666e-02 4.91262436e-01 -9.62842524e-01
1.40460706e+00 -1.16320275e-01 -6.09492101e-02 5.44477820e-01
-1.12116146e+00 8.22834224e-02 5.63978016e-01 -9.62794200e-02
-7.70880163e-01 -3.78678352e-01 -1.98856741e-01 1.60827026e-01
-3.24936956e-01 7.94701695e-01 1.81461185e-01 1.98890582e-01
1.16470659e+00 8.65137577e-02 -1.71365812e-02 -4.33395863e-01
3.14023256e-01 8.91785502e-01 5.71854591e-01 -3.93971086e-01
3.39036405e-01 -7.98610672e-02 -4.69646275e-01 -6.87887847e-01
-5.59187710e-01 -4.24134701e-01 -6.69246376e-01 -8.80393684e-02
5.28024673e-01 -4.44431543e-01 -6.42025352e-01 1.09467946e-01
-1.25492024e+00 -2.08784133e-01 -4.25018847e-01 5.29239416e-01
-6.73098683e-01 4.52167600e-01 -4.55443621e-01 -8.19167793e-01
-4.97025847e-01 -1.04190469e+00 8.12740326e-01 7.87874162e-02
-9.55990672e-01 -8.87348533e-01 3.58553618e-01 5.03698051e-01
5.56644201e-01 -5.42280003e-02 6.07231379e-01 -9.57995772e-01
6.63742945e-02 -5.10168552e-01 -1.76729590e-01 4.86739248e-01
-1.16915718e-01 3.33689302e-01 -9.77847695e-01 -1.97384194e-01
1.41794980e-01 -5.22836030e-01 3.91972244e-01 -5.35291433e-02
8.34578872e-01 -3.02894980e-01 -1.70518532e-01 -7.14119822e-02
1.30414581e+00 -5.80265447e-02 8.76294374e-01 5.41862965e-01
9.99032259e-02 6.42421246e-01 7.47985125e-01 1.72346398e-01
2.83955544e-01 6.27621949e-01 -1.55380219e-01 2.56377369e-01
-2.46140733e-02 -5.23384094e-01 1.53018355e-01 8.98459911e-01
3.48567218e-03 -5.92905879e-01 -1.09454203e+00 5.05928159e-01
-1.76918840e+00 -8.84174109e-01 -9.63895246e-02 2.23667526e+00
9.64308679e-01 5.62971294e-01 1.08866215e-01 5.14343321e-01
5.15411079e-01 3.63775864e-02 -8.14371929e-03 -8.68733227e-01
-6.68171346e-02 5.90302199e-02 4.49727505e-01 3.97422999e-01
-4.83742774e-01 6.84153020e-01 7.90818405e+00 7.62429476e-01
-8.08497906e-01 2.44159214e-02 3.82108629e-01 -1.83786586e-01
-3.32495987e-01 1.02566928e-01 -4.56548929e-01 6.82028592e-01
1.53393495e+00 -6.32630467e-01 4.17011380e-01 4.45180535e-01
3.03489774e-01 -1.54685155e-01 -1.33681440e+00 8.55776250e-01
1.23841912e-01 -1.11621392e+00 -1.11120954e-01 1.10190390e-02
2.74439931e-01 1.47364289e-03 -2.45711461e-01 3.20960790e-01
2.26815745e-01 -1.07381225e+00 7.78628588e-01 5.36796451e-01
5.82163811e-01 -4.67341095e-01 1.11286569e+00 7.00634480e-01
-4.62759286e-01 3.26396823e-01 -5.92988908e-01 -3.25153798e-01
4.21805441e-01 5.58830857e-01 -1.30194712e+00 4.33979750e-01
5.48227549e-01 1.28070101e-01 -7.95159698e-01 1.09234023e+00
-4.56395924e-01 7.18435526e-01 -6.24362826e-01 -5.05252123e-01
1.18569192e-02 2.30523765e-01 4.91460919e-01 1.86594605e+00
2.25030273e-01 2.74819247e-02 -3.48374605e-01 7.67210722e-01
-2.80041508e-02 5.02636731e-01 -8.14689934e-01 -4.00955826e-01
6.24967575e-01 1.09724545e+00 -7.64973223e-01 -4.87482756e-01
-1.01210594e-01 7.27172494e-01 5.27297854e-01 1.20094053e-01
-4.94376421e-01 -4.96088624e-01 1.72658235e-01 2.62595695e-02
9.94929522e-02 -3.58277649e-01 -6.24842823e-01 -1.01722240e+00
2.57990539e-01 -9.52739596e-01 4.06355858e-01 -9.03768837e-01
-1.07153344e+00 6.68149710e-01 1.17606692e-01 -1.21178818e+00
-5.42976499e-01 -4.80072081e-01 -4.06312406e-01 9.23458993e-01
-1.02585399e+00 -3.97857904e-01 -1.46311373e-01 -1.46519512e-01
1.17855251e-01 6.89022392e-02 1.17709875e+00 -4.30489890e-02
-1.08967550e-01 6.14977002e-01 4.31648716e-02 -9.47670639e-02
8.47322226e-01 -1.35664296e+00 4.55713421e-01 9.43345308e-01
5.48670411e-01 8.36217105e-01 1.35682178e+00 -4.99879718e-01
-1.06391048e+00 -3.34856808e-01 1.36249173e+00 -7.82837212e-01
5.86822271e-01 1.09061720e-02 -1.02356267e+00 2.31477693e-01
4.53884959e-01 -4.58408862e-01 8.10418963e-01 2.37630308e-01
-3.04974794e-01 1.29423484e-01 -1.32676125e+00 4.58596289e-01
9.74260092e-01 -5.48200786e-01 -1.21842670e+00 4.15609092e-01
6.20615244e-01 -6.05535328e-01 -1.05094194e+00 3.54994357e-01
6.90453649e-01 -9.90435183e-01 6.08614743e-01 -5.49232900e-01
5.43847919e-01 -2.53535330e-01 -2.94223964e-01 -1.33025229e+00
-1.12587638e-01 -2.87150651e-01 1.03033058e-01 1.28279471e+00
9.41466331e-01 -3.05566072e-01 7.09905386e-01 9.43241000e-01
4.85724658e-02 -6.36058331e-01 -8.06521177e-01 -8.29054296e-01
1.72913700e-01 -8.61920357e-01 4.26892459e-01 8.59396756e-01
5.02566218e-01 4.44461077e-01 2.45230526e-01 -2.37925649e-01
2.55276591e-01 -2.13715330e-01 7.56955862e-01 -1.03586507e+00
-1.58599421e-01 -6.92481637e-01 -7.63959408e-01 -6.21326566e-01
-1.06445178e-01 -8.43240499e-01 2.49480188e-01 -1.89550996e+00
1.72128707e-01 -2.05118418e-01 -4.77947176e-01 3.48214269e-01
-2.31342204e-02 3.75122488e-01 4.40901995e-01 2.61985421e-01
-7.86801696e-01 1.07945897e-01 9.16617036e-01 1.80699527e-01
-1.69431224e-01 -3.01866770e-01 -1.17483282e+00 4.82250482e-01
7.89636850e-01 -6.36215091e-01 -3.94940376e-01 -5.31052351e-01
4.90236789e-01 -3.55802417e-01 -7.38292560e-02 -1.32891011e+00
4.91583288e-01 4.33852747e-02 2.60917664e-01 -4.20253217e-01
1.57468051e-01 -6.95479751e-01 5.84625639e-02 4.94730145e-01
-6.97040319e-01 2.39829779e-01 2.42911220e-01 2.29784369e-01
-3.27760309e-01 -6.69114172e-01 3.62005591e-01 -2.20362306e-01
-3.60884547e-01 -4.93094414e-01 -1.72595501e-01 1.63519233e-01
7.28839517e-01 -2.16850817e-01 -7.48511255e-01 -5.11579454e-01
-4.55255747e-01 2.81112432e-01 8.00645709e-01 2.73182303e-01
4.86982226e-01 -8.87637377e-01 -1.07993972e+00 -1.63641065e-01
3.04802477e-01 -2.85993338e-01 -1.62451029e-01 6.42524838e-01
-9.30960536e-01 7.18215227e-01 -8.87622982e-02 -7.18102336e-01
-1.25077820e+00 3.26382875e-01 -4.60559092e-02 -3.91596049e-01
-4.20712203e-01 4.63557214e-01 -6.31184936e-01 -4.10911620e-01
2.93402970e-01 -4.82997477e-01 -2.18012333e-01 1.94806471e-01
6.89613223e-01 2.77518481e-01 5.34989715e-01 -3.76331300e-01
-2.63353378e-01 1.57432944e-01 -2.11153373e-01 -7.06748545e-01
1.21218967e+00 -4.12648320e-02 1.20084837e-01 8.41671467e-01
9.74945843e-01 2.02400405e-02 -4.99137223e-01 -1.96816381e-02
8.11157972e-02 -6.23329461e-01 -8.51349831e-02 -1.28524911e+00
-2.14543164e-01 9.00497556e-01 2.70259470e-01 7.24312127e-01
8.33887935e-01 -1.76165476e-01 6.11961961e-01 8.45445454e-01
5.18480241e-01 -1.27975893e+00 -3.34796309e-01 4.23934907e-01
1.06053483e+00 -1.17752814e+00 2.64901549e-01 -1.55000806e-01
-9.22287345e-01 1.02177477e+00 2.19937459e-01 1.90776169e-01
2.71516860e-01 1.39210045e-01 2.41257697e-01 1.58037432e-02
-1.01603460e+00 2.32689418e-02 1.42452598e-01 4.61502969e-01
8.74519229e-01 8.11299533e-02 -9.99552846e-01 5.65285087e-01
-9.88031685e-01 4.13754791e-01 6.25537872e-01 1.06223130e+00
-5.93945384e-01 -1.22876906e+00 -1.66645482e-01 6.53160393e-01
-7.38964140e-01 -9.31865051e-02 -6.58284903e-01 9.72440481e-01
-3.00939441e-01 1.13606465e+00 -1.41579285e-02 -3.92033488e-01
6.75539613e-01 3.99538755e-01 6.57739937e-01 -7.03279376e-01
-7.84290314e-01 -4.88427877e-01 5.69323540e-01 -4.56410587e-01
-4.55104500e-01 -8.19533467e-01 -1.02594864e+00 -5.40774465e-01
-1.67138457e-01 5.00341892e-01 9.18227732e-01 9.32085276e-01
3.29918772e-01 1.45175323e-01 2.85931397e-03 -4.06640768e-01
-7.50713825e-01 -1.53653133e+00 -2.62594640e-01 6.75541341e-01
-5.74823804e-02 -2.02684700e-01 -5.06458104e-01 1.26484722e-01] | [11.695161819458008, 9.034539222717285] |
0c8ee844-7f2b-4f13-bacd-5b36ef5ed094 | distributional-reinforcement-learning-with-3 | 2007.12354 | null | https://arxiv.org/abs/2007.12354v3 | https://arxiv.org/pdf/2007.12354v3.pdf | Distributional Reinforcement Learning via Moment Matching | We consider the problem of learning a set of probability distributions from the empirical Bellman dynamics in distributional reinforcement learning (RL), a class of state-of-the-art methods that estimate the distribution, as opposed to only the expectation, of the total return. We formulate a method that learns a finite set of statistics from each return distribution via neural networks, as in (Bellemare, Dabney, and Munos 2017; Dabney et al. 2018b). Existing distributional RL methods however constrain the learned statistics to \emph{predefined} functional forms of the return distribution which is both restrictive in representation and difficult in maintaining the predefined statistics. Instead, we learn \emph{unrestricted} statistics, i.e., deterministic (pseudo-)samples, of the return distribution by leveraging a technique from hypothesis testing known as maximum mean discrepancy (MMD), which leads to a simpler objective amenable to backpropagation. Our method can be interpreted as implicitly matching all orders of moments between a return distribution and its Bellman target. We establish sufficient conditions for the contraction of the distributional Bellman operator and provide finite-sample analysis for the deterministic samples in distribution approximation. Experiments on the suite of Atari games show that our method outperforms the standard distributional RL baselines and sets a new record in the Atari games for non-distributed agents. | ['Sunil Gupta', 'Svetha Venkatesh', 'Thanh Tang Nguyen'] | 2020-07-24 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [-4.35865074e-01 1.65418044e-01 -2.89155841e-01 -1.13276146e-01
-1.20420933e+00 -9.85398054e-01 6.25946939e-01 -4.45898622e-02
-7.41575301e-01 1.06319630e+00 5.04144952e-02 -4.63976383e-01
-5.47097802e-01 -7.59798586e-01 -1.01852930e+00 -1.12730813e+00
-5.94645798e-01 9.88664091e-01 -3.88122797e-01 -3.13069373e-01
5.52393943e-02 3.22832823e-01 -1.40927958e+00 -1.24761701e-01
6.60525382e-01 1.07468843e+00 1.23929173e-01 7.88471460e-01
7.61756748e-02 8.73004496e-01 -8.65407050e-01 -2.34610111e-01
4.13609654e-01 -4.90201950e-01 -4.71683204e-01 -3.73447537e-01
2.26311669e-01 -6.24619305e-01 -2.46744096e-01 1.43660212e+00
4.69612777e-01 4.47110981e-01 1.00111783e+00 -1.36890733e+00
-8.82479966e-01 1.04733777e+00 -6.02222145e-01 2.53493518e-01
9.22670439e-02 1.14496827e-01 1.27987897e+00 -5.08719325e-01
2.88470536e-01 1.21187222e+00 6.27115667e-01 6.83428168e-01
-1.45787919e+00 -4.99489039e-01 1.63013518e-01 -2.22915873e-01
-1.23471582e+00 -1.18517056e-01 3.44991833e-01 -4.95094448e-01
7.07168818e-01 8.07176679e-02 5.59022069e-01 1.09429097e+00
5.70128411e-02 8.42843413e-01 1.23513973e+00 -2.42470428e-01
6.34959757e-01 -3.11391503e-02 -1.45455211e-01 5.36329448e-01
2.23320290e-01 6.14935398e-01 -2.10563749e-01 -4.49982822e-01
8.37933421e-01 -1.19021952e-01 -3.30030113e-01 -5.73703289e-01
-6.70020461e-01 1.26030314e+00 4.19141464e-02 -7.41679296e-02
-5.73671579e-01 6.16269112e-01 2.67667264e-01 3.95784736e-01
3.60819042e-01 2.76426166e-01 -5.31205058e-01 -3.37067336e-01
-6.73633397e-01 8.79623890e-01 1.16094768e+00 8.55236411e-01
7.56043971e-01 2.03445241e-01 -3.58425260e-01 2.82606184e-01
3.08042258e-01 1.13559639e+00 3.92236292e-01 -1.15803742e+00
5.08358359e-01 -2.49016866e-01 6.30970955e-01 -3.45522434e-01
-2.19950110e-01 -3.84684354e-01 -5.84529519e-01 3.00803035e-01
1.13736391e+00 -6.03449583e-01 -6.53305352e-01 2.35680652e+00
8.21598247e-02 1.52669996e-01 2.57102847e-01 9.01521683e-01
5.16179688e-02 6.47795618e-01 -1.81619808e-01 -3.03425401e-01
7.00614035e-01 -4.61045444e-01 -4.13575798e-01 1.17262147e-01
4.14494038e-01 -2.34826729e-01 1.37974131e+00 4.26931083e-01
-1.05582952e+00 -6.05575219e-02 -7.42695808e-01 4.33576196e-01
-7.94327036e-02 -2.22009525e-01 5.92221856e-01 6.24497652e-01
-1.24700177e+00 7.14206100e-01 -7.79566586e-01 6.10516109e-02
4.36019391e-01 4.41315919e-01 2.10518925e-03 3.03273171e-01
-9.44629848e-01 8.06690812e-01 2.32007071e-01 -1.91301212e-03
-1.52859950e+00 -9.17135954e-01 -8.94848585e-01 6.90294504e-02
5.21147788e-01 -1.66961193e-01 1.69927239e+00 -1.16239727e+00
-1.85801482e+00 4.46437359e-01 3.18764865e-01 -8.90855193e-01
5.01941502e-01 -1.85039952e-01 2.19719470e-01 3.51685099e-03
-2.37578712e-02 4.55943555e-01 8.68732035e-01 -1.17301416e+00
-6.22723043e-01 -2.26196244e-01 3.25077891e-01 9.65794548e-02
-1.72127727e-02 -4.03966546e-01 4.12441820e-01 -4.07468736e-01
-6.70324206e-01 -8.34737957e-01 -4.68232296e-02 -3.89813811e-01
-3.38595390e-01 -4.63728458e-01 2.81024486e-01 -3.55008572e-01
9.08505321e-01 -1.98273420e+00 3.83309007e-01 4.09826040e-01
4.56144102e-02 -1.96901426e-01 -3.87922794e-01 4.06931221e-01
8.50372165e-02 4.66074701e-03 -3.11837643e-01 -3.19490045e-01
6.90036952e-01 5.60549140e-01 -9.61307645e-01 8.97098064e-01
-1.17086738e-01 7.04081774e-01 -1.08183670e+00 8.62545818e-02
-1.92020550e-01 2.24602252e-01 -8.50721598e-01 4.65967417e-01
-7.52104461e-01 3.77719194e-01 -1.97354540e-01 8.12690035e-02
5.51700473e-01 2.28197068e-01 1.30363241e-01 2.76912361e-01
-1.89460054e-01 1.88931108e-01 -9.98685956e-01 1.47166085e+00
-4.30923790e-01 9.15728882e-02 1.19003519e-01 -1.24704933e+00
7.60639369e-01 -2.37627067e-02 5.44966459e-01 -3.44663858e-01
3.48057300e-01 2.60819942e-01 2.53403544e-01 -2.80837327e-01
3.62998277e-01 -5.84195733e-01 -3.80799592e-01 7.58786619e-01
1.70051128e-01 -2.37403363e-01 3.16625506e-01 -1.26514241e-01
8.02014351e-01 6.01920247e-01 1.38925731e-01 -7.12186575e-01
1.46663114e-01 -2.50921249e-01 3.12502116e-01 1.21209872e+00
-2.33624831e-01 2.07469225e-01 9.16669846e-01 -2.08704829e-01
-1.11466885e+00 -1.68604827e+00 2.91434079e-02 1.11473620e+00
-2.08572090e-01 -2.06061900e-01 -8.07643652e-01 -8.81104887e-01
3.46370101e-01 1.07847261e+00 -9.14980650e-01 -3.09589922e-01
-3.14131677e-01 -6.47513866e-01 8.47506344e-01 3.86085033e-01
1.75060868e-01 -1.10297573e+00 -9.14327860e-01 -6.49588481e-02
7.13111460e-02 -5.07772744e-01 -8.08413208e-01 6.17552638e-01
-4.02098924e-01 -9.34382379e-01 -6.77092075e-01 -4.36179101e-01
3.04190010e-01 -4.12027389e-01 1.16466486e+00 -5.25909007e-01
1.44880831e-01 8.64276946e-01 1.44521892e-01 -5.73934674e-01
-4.42059249e-01 -2.24948958e-01 3.27101618e-01 -2.05041096e-01
2.77756780e-01 -6.36382520e-01 -3.06624800e-01 -2.06099927e-01
-9.47129250e-01 -7.09268987e-01 3.96656960e-01 8.76381218e-01
8.66368115e-01 1.53797626e-01 6.90585196e-01 -6.11990511e-01
1.04934120e+00 -7.75135994e-01 -1.07098997e+00 1.04925126e-01
-2.49661624e-01 5.43383777e-01 1.08650112e+00 -8.02932024e-01
-8.24065924e-01 -3.36093545e-01 1.81140214e-01 -6.72726452e-01
2.20429758e-03 5.83925545e-01 -6.71979859e-02 5.08770406e-01
5.44266641e-01 3.29715669e-01 3.67967606e-01 -1.62940413e-01
7.00198650e-01 3.87954444e-01 7.07617700e-01 -1.27040672e+00
6.86136305e-01 4.22661632e-01 2.98804399e-02 -5.24982512e-01
-1.27932644e+00 -9.38110948e-02 -2.99889058e-01 -4.28737104e-02
6.57218635e-01 -5.02619147e-01 -1.31166923e+00 2.67350048e-01
-8.72678757e-01 -1.06702626e+00 -1.07894969e+00 5.26535153e-01
-1.35284376e+00 3.53059247e-02 -4.05725211e-01 -1.33664143e+00
-2.97011379e-02 -8.53214085e-01 9.56749082e-01 9.42925587e-02
5.59633188e-02 -1.16508925e+00 6.05241954e-01 -4.64593977e-01
4.45244938e-01 3.36765721e-02 1.08864260e+00 -8.16191733e-01
-1.36376709e-01 2.00175017e-01 2.54607767e-01 5.35403788e-01
-9.07630995e-02 -1.60606325e-01 -6.44965529e-01 -4.71753806e-01
2.98808426e-01 -7.76445210e-01 8.51728141e-01 7.80156076e-01
1.12348545e+00 -8.59244347e-01 2.19637364e-01 5.46006024e-01
1.32627249e+00 1.37992010e-01 3.72090667e-01 3.21630239e-01
2.27071539e-01 4.01323825e-01 5.09869218e-01 9.59841311e-01
4.64584261e-01 1.84357896e-01 4.47254568e-01 5.00238776e-01
4.05408442e-01 -6.57159388e-01 8.55914235e-01 2.88835198e-01
1.48379177e-01 -2.38520756e-01 -5.92199624e-01 3.82365078e-01
-1.98202789e+00 -1.42797601e+00 6.89578950e-01 2.65485883e+00
1.18048584e+00 -2.60808188e-02 7.40259051e-01 -4.19758618e-01
6.41166568e-01 -1.28070787e-01 -9.16256785e-01 -4.98910099e-01
-1.03643015e-01 5.40151298e-01 6.63632214e-01 7.44445622e-01
-9.34855342e-01 6.94728196e-01 6.46967268e+00 8.50305557e-01
-9.48805153e-01 -6.78230003e-02 4.99792427e-01 -1.71090901e-01
-5.64821899e-01 -1.49368063e-01 -6.54003322e-01 4.43723947e-01
1.02776778e+00 -3.62897336e-01 1.06797457e+00 8.64619792e-01
8.55518952e-02 -6.01172186e-02 -1.37333393e+00 8.36281717e-01
9.69662610e-03 -1.02653503e+00 -1.80841804e-01 4.65335369e-01
7.17072845e-01 8.18155482e-02 5.41979134e-01 7.00289786e-01
1.09651184e+00 -1.30646443e+00 1.23376811e+00 6.96049213e-01
7.08506465e-01 -1.14051318e+00 7.68382609e-01 6.19101822e-01
-6.86453223e-01 -2.88322538e-01 -5.70029914e-01 -1.71471670e-01
-3.45218539e-01 1.72916710e-01 -3.91926706e-01 7.58898109e-02
6.63452208e-01 4.90646243e-01 2.33960748e-01 6.45255625e-01
-2.09279031e-01 6.97469652e-01 -6.32182956e-01 -2.72968739e-01
4.29335028e-01 -5.47613680e-01 5.44261336e-01 8.30103278e-01
3.31110358e-01 -1.42507821e-01 5.50746143e-01 1.23320699e+00
-1.63030252e-01 -2.36545637e-01 -7.72327125e-01 -2.88312256e-01
5.82061112e-01 9.97335017e-01 -3.25724036e-01 3.15352343e-02
-6.29041670e-03 4.88190264e-01 8.96593153e-01 5.64185679e-01
-1.05589497e+00 -1.63569763e-01 7.73129702e-01 -3.16142887e-01
6.42536461e-01 -3.17146897e-01 2.59089082e-01 -1.00086725e+00
-1.55374661e-01 -9.33566391e-01 6.13178313e-01 -1.53978586e-01
-1.77131522e+00 2.75890440e-01 3.95851344e-01 -8.96268725e-01
-7.02966094e-01 -6.18848562e-01 -6.33114517e-01 7.71885812e-01
-1.40275848e+00 -8.18222880e-01 4.76860791e-01 7.80934334e-01
-1.78708993e-02 -3.19214940e-01 8.45314026e-01 -2.08185047e-01
-3.68581682e-01 5.04117489e-01 4.76612478e-01 -6.11750707e-02
3.11026007e-01 -1.79722893e+00 -2.93161303e-01 4.24792230e-01
2.79022396e-01 4.73847449e-01 9.02264893e-01 -4.40016985e-01
-1.64153826e+00 -8.69956076e-01 9.88385156e-02 -4.42239851e-01
1.20419264e+00 -4.22052324e-01 -3.64383638e-01 1.00367689e+00
2.28298783e-01 -6.21978492e-02 5.56064069e-01 -1.83463395e-02
-4.82320368e-01 -1.58648029e-01 -1.20090425e+00 6.16556883e-01
5.58424890e-01 -3.21598589e-01 -4.30736601e-01 1.15633942e-01
5.10125995e-01 -2.89494157e-01 -5.81292272e-01 -4.68391702e-02
4.24196035e-01 -8.99660528e-01 6.93628430e-01 -9.69226539e-01
2.18068630e-01 -3.39246631e-01 -5.42191446e-01 -1.67355335e+00
1.16074339e-01 -1.05776107e+00 -3.69673729e-01 1.11884189e+00
2.04590574e-01 -6.64571822e-01 5.03708184e-01 2.42622465e-01
1.50896898e-02 -8.26494992e-01 -1.18281925e+00 -9.79692042e-01
8.75176132e-01 -4.80989069e-01 5.85548222e-01 3.69601578e-01
1.97777107e-01 4.25298065e-02 -3.05671334e-01 1.83116570e-01
8.74740958e-01 3.32313269e-01 6.12121940e-01 -7.67614722e-01
-7.51358390e-01 -6.84528291e-01 -4.45826314e-02 -1.13282335e+00
8.72693658e-01 -1.13368738e+00 5.12858450e-01 -1.14849162e+00
2.27813825e-01 -5.65547705e-01 -3.42034429e-01 3.17736506e-01
3.64924669e-01 -2.34286726e-01 2.03493699e-01 5.95080815e-02
-7.19662488e-01 9.59517241e-01 9.80883241e-01 -5.42634130e-02
-1.94585443e-01 1.24822900e-01 -7.25185573e-01 7.79197216e-01
8.67666960e-01 -4.62410510e-01 -6.34576917e-01 -1.31429568e-01
3.09316725e-01 2.23851487e-01 4.87942040e-01 -4.79313672e-01
1.22629404e-02 -5.00447750e-01 1.25382543e-01 -3.70071948e-01
1.00165524e-01 -5.04543960e-01 -4.51943278e-01 5.23169458e-01
-7.45653510e-01 2.46674061e-01 1.74560383e-01 6.79015160e-01
2.16741160e-01 -4.38842654e-01 7.94771433e-01 -1.92097910e-02
1.65125297e-03 4.40175384e-01 -5.44557452e-01 8.70755494e-01
8.14523637e-01 4.36634272e-01 -3.09530228e-01 -8.24546278e-01
-4.05097902e-01 2.83739269e-01 5.25826551e-02 -2.26814866e-01
5.30714214e-01 -1.20120215e+00 -7.41397440e-01 3.22512165e-02
-3.53190422e-01 5.68852685e-02 -1.78584248e-01 8.57635736e-01
-9.58537683e-02 1.16086967e-01 -9.98578593e-02 -2.62052000e-01
-2.51563460e-01 6.15285099e-01 7.33756125e-01 -3.70899498e-01
-3.86965662e-01 6.84625387e-01 1.69101089e-01 -6.12754047e-01
4.58095461e-01 -4.80221093e-01 1.14914320e-01 -5.43396622e-02
4.97768521e-01 2.56342947e-01 -5.82003951e-01 -3.13702762e-01
-1.01006061e-01 1.57367945e-01 1.01740867e-01 -6.49106503e-01
1.42249787e+00 -9.73141659e-03 -7.58948028e-02 7.05262303e-01
9.81738091e-01 2.09230334e-01 -1.78473759e+00 2.55024340e-02
2.59083994e-02 -1.54809967e-01 -2.85259694e-01 -8.81648242e-01
-1.03098571e+00 7.87284672e-01 2.96596020e-01 3.83689791e-01
7.60118842e-01 2.58881748e-01 4.72821474e-01 6.60411656e-01
4.69647199e-01 -1.01143277e+00 2.36491319e-02 7.25679159e-01
9.19881523e-01 -7.67295718e-01 -3.82546961e-01 5.63556850e-01
-7.43623197e-01 1.17491043e+00 4.91980374e-01 -6.71217024e-01
7.97127962e-01 7.62111723e-01 -1.65859327e-01 2.70695258e-02
-8.05300117e-01 -4.44134891e-01 2.98871081e-02 9.22240496e-01
4.36415114e-02 2.00426087e-01 2.89591160e-02 9.70618606e-01
-6.55737936e-01 -3.46534044e-01 6.70499802e-01 6.00832403e-01
-4.10683811e-01 -8.36726904e-01 -1.27007246e-01 3.53561640e-01
-5.11171341e-01 -2.01967508e-01 -7.29268119e-02 9.73017454e-01
-1.35905132e-01 8.65771770e-01 3.61345977e-01 -1.30526247e-02
1.06266588e-01 -1.04628049e-01 9.35440063e-01 -4.80311811e-01
-2.80398518e-01 8.67463872e-02 -3.01566958e-01 -3.18877339e-01
-2.28206038e-01 -8.64936650e-01 -1.47917438e+00 -4.35065955e-01
-1.34683907e-01 5.54703712e-01 3.58748555e-01 1.14478004e+00
-1.53638914e-01 -1.28109818e-02 5.98865569e-01 -8.49400759e-01
-1.83264601e+00 -8.08258533e-01 -9.99581039e-01 3.44125271e-01
5.96465051e-01 -6.13330901e-01 -6.78476930e-01 -3.20724219e-01] | [4.099051475524902, 2.588942766189575] |
54e102fb-5bce-4226-8618-97b21f4a46f9 | nonlinear-residual-echo-suppression-based-on | 2005.07631 | null | https://arxiv.org/abs/2005.07631v1 | https://arxiv.org/pdf/2005.07631v1.pdf | Nonlinear Residual Echo Suppression Based on Multi-stream Conv-TasNet | Acoustic echo cannot be entirely removed by linear adaptive filters due to the nonlinear relationship between the echo and far-end signal. Usually a post processing module is required to further suppress the echo. In this paper, we propose a residual echo suppression method based on the modification of fully convolutional time-domain audio separation network (Conv-TasNet). Both the residual signal of the linear acoustic echo cancellation system, and the output of the adaptive filter are adopted to form multiple streams for the Conv-TasNet, resulting in more effective echo suppression while keeping a lower latency of the whole system. Simulation results validate the efficacy of the proposed method in both single-talk and double-talk situations. | ['Kai Chen', 'Jing Lu', 'Teng Xiang', 'Hongsheng Chen'] | 2020-05-15 | null | null | null | null | ['acoustic-echo-cancellation', 'acoustic-echo-cancellation'] | ['medical', 'speech'] | [ 1.09767802e-01 -5.10663271e-01 7.96993792e-01 -2.18698457e-01
-4.21864957e-01 -3.59272450e-01 9.66763124e-02 -2.14320138e-01
-6.66722476e-01 1.79245248e-01 3.52388740e-01 -9.51417089e-02
-1.05674885e-01 -2.52430946e-01 -2.15392381e-01 -7.41509259e-01
-7.94921517e-02 -6.55096233e-01 3.93244714e-01 -2.74980307e-01
-4.57089767e-02 1.95633978e-01 -1.19928265e+00 2.20810294e-01
9.67283249e-01 1.13082159e+00 2.16387764e-01 8.98778141e-01
-1.51958242e-01 5.57765901e-01 -8.72924566e-01 2.65498850e-02
3.37216109e-01 -5.42804956e-01 3.97679925e-01 -3.54268700e-01
4.13802385e-01 -6.17011189e-01 -8.64081740e-01 1.13729322e+00
1.00754297e+00 4.67860579e-01 2.26724803e-01 -8.66458178e-01
2.69200414e-01 8.36042762e-01 -3.75718474e-01 4.05923158e-01
-5.32005019e-02 -3.92240770e-02 5.25718689e-01 -1.14129162e+00
-1.73200309e-01 1.27198613e+00 9.67244446e-01 2.22539872e-01
-9.63870943e-01 -1.49021626e+00 1.02170650e-02 2.56426483e-01
-1.44509065e+00 -1.05173445e+00 8.88717771e-01 1.37026697e-01
8.19360495e-01 2.38851964e-01 6.87058628e-01 6.15208030e-01
3.39462847e-01 5.54562390e-01 8.49135160e-01 -3.33016008e-01
-2.01683175e-02 -2.40612864e-01 2.65766978e-01 3.00686061e-01
-2.08746582e-01 3.56759846e-01 -7.31271267e-01 -2.05385178e-01
5.08326352e-01 -9.34705064e-02 -4.99288589e-01 2.35909000e-01
-6.90906227e-01 2.14547753e-01 3.86377305e-01 4.15182561e-01
-2.70037413e-01 1.79632485e-01 6.76515102e-01 5.69560170e-01
3.72030109e-01 -1.40152127e-01 -6.06946610e-02 7.19106048e-02
-1.33540893e+00 7.41338218e-03 7.39541769e-01 6.22244954e-01
2.20575392e-01 9.10806715e-01 2.44102366e-02 8.75234425e-01
3.94711822e-01 7.66354203e-01 5.17828226e-01 -5.62516153e-01
3.43630791e-01 -2.41775900e-01 -8.09818730e-02 -1.05775166e+00
-4.23275948e-01 -8.65780175e-01 -1.15992904e+00 1.33159995e-01
1.93890780e-01 -6.01257741e-01 -9.61899817e-01 1.59562218e+00
3.66498888e-01 8.30471635e-01 2.61126131e-01 1.25062203e+00
7.40145802e-01 1.02196753e+00 -1.32062301e-01 -3.29516917e-01
9.48905408e-01 -9.39466596e-01 -1.21041429e+00 -3.26188505e-01
2.28557184e-01 -1.14994156e+00 3.12906355e-01 4.87621188e-01
-9.89392698e-01 -7.47698843e-01 -1.42392290e+00 2.08497941e-01
2.58411795e-01 2.05735043e-01 -7.13273212e-02 5.89179099e-01
-8.39851499e-01 4.31220055e-01 -6.87909245e-01 3.54513079e-01
-3.01128775e-01 3.24545026e-01 -3.84944156e-02 -5.96761797e-03
-1.33861661e+00 2.36577213e-01 3.32867689e-02 9.21683371e-01
-7.70413280e-01 -7.62422562e-01 -5.68046987e-01 4.89122033e-01
3.72769773e-01 -2.25180954e-01 1.29784071e+00 -1.10593498e+00
-1.55800927e+00 -2.28309304e-01 -1.85460702e-01 -4.62852687e-01
5.10354698e-01 -6.28147423e-01 -1.05849338e+00 -4.64980491e-03
-2.81693459e-01 1.18857093e-01 1.32740092e+00 -7.90733635e-01
-7.45476007e-01 6.34674588e-03 -4.56167519e-01 3.00219655e-01
-3.68748873e-01 -4.90956455e-02 -3.39136690e-01 -8.83915544e-01
2.80128360e-01 -8.21929336e-01 -2.98068672e-01 -3.33376408e-01
-3.08788180e-01 3.53152275e-01 9.29333448e-01 -9.13877606e-01
1.50071824e+00 -2.98156834e+00 -1.86733544e-01 2.83500284e-01
1.87150702e-01 8.01590502e-01 -4.54993427e-01 3.31456035e-01
-2.19254613e-01 -3.62791985e-01 -2.25326613e-01 -2.63808101e-01
-3.14516991e-01 -3.48470777e-01 -5.05394101e-01 5.26993454e-01
-4.41921968e-03 1.98782861e-01 -6.32292211e-01 -2.56522149e-01
2.92790174e-01 7.61835992e-01 -1.95431888e-01 3.96259457e-01
5.94445407e-01 2.82372683e-01 -1.64275348e-01 -2.54804902e-02
1.17634606e+00 8.13201010e-01 1.19357869e-01 -3.83891195e-01
-4.95980740e-01 3.48043919e-01 -1.74020064e+00 1.36925077e+00
-5.83737791e-01 8.69214714e-01 1.07040548e+00 -5.79377532e-01
1.08676064e+00 5.56617677e-01 3.35852593e-01 -1.07855296e+00
2.97195315e-01 4.98452961e-01 5.57034254e-01 -2.32443660e-01
3.87314439e-01 -3.93097132e-01 2.15749264e-01 8.63621086e-02
-3.27264294e-02 1.56717017e-01 -1.93431690e-01 7.70608112e-02
1.14736378e+00 -2.82688707e-01 -2.58529961e-01 -1.45519450e-01
7.90778399e-01 -5.53041041e-01 8.24722886e-01 6.45387411e-01
-2.53612697e-01 2.96144307e-01 -1.95644021e-01 -4.29866128e-02
-7.02116728e-01 -8.82838309e-01 2.03157112e-01 1.09443069e+00
1.50423333e-01 -4.08633679e-01 -7.29063809e-01 -1.39371097e-01
-9.93128121e-02 4.34268415e-01 7.70502165e-02 -4.72646862e-01
-1.00487685e+00 -1.66992992e-01 1.03177083e+00 3.12515199e-01
5.85659266e-01 -8.74658644e-01 -3.42947811e-01 7.19060242e-01
-3.04692328e-01 -1.02432227e+00 -9.31269348e-01 3.41937244e-01
-7.54156530e-01 -5.69937468e-01 -7.02944100e-01 -6.68085635e-01
1.62759647e-01 7.52190530e-01 1.58666074e-01 2.03011855e-01
2.83852257e-02 2.00545751e-02 -1.22015081e-01 -4.13557917e-01
-2.92305171e-01 -1.90834999e-01 2.42993727e-01 3.43640208e-01
-3.02398615e-02 -6.74378932e-01 -6.92966163e-01 3.33846927e-01
-1.04624414e+00 -3.06086600e-01 6.06161952e-01 7.20362127e-01
-5.65064065e-02 5.31188071e-01 5.88866413e-01 -1.66008249e-01
9.78566587e-01 -8.18031803e-02 -6.01548374e-01 -1.67835534e-01
-1.87178403e-01 -1.54717416e-01 9.73382175e-01 -7.80855238e-01
-1.47853625e+00 2.52398867e-02 -4.86500293e-01 -6.31992936e-01
2.24954695e-01 4.69144583e-01 -1.01460971e-01 -4.99706492e-02
1.31944418e-01 2.84370601e-01 -1.10364556e-01 -5.96244812e-01
1.59740463e-01 9.36833799e-01 6.33168221e-01 2.51536489e-01
9.40327823e-01 2.63368189e-01 -8.55404139e-02 -1.16228426e+00
-2.82165438e-01 -5.36081970e-01 -3.28453392e-01 -4.24185485e-01
4.08451051e-01 -1.03008366e+00 -7.59885609e-01 8.15673351e-01
-1.22452796e+00 -9.81295854e-02 1.92401394e-01 1.10164690e+00
2.24083036e-01 3.63722563e-01 -8.47335994e-01 -1.10349596e+00
-8.61662567e-01 -9.05339003e-01 5.74155033e-01 3.25661004e-01
3.49664204e-02 -4.62047130e-01 1.48507580e-01 -2.73888588e-01
8.75607431e-01 -5.26061594e-01 4.50222731e-01 -4.32822466e-01
-3.23227882e-01 -2.08122730e-01 -7.32159168e-02 5.23730338e-01
-5.28951436e-02 -7.77608156e-02 -1.08629203e+00 -4.39631820e-01
4.60337430e-01 2.05039203e-01 1.01385534e+00 6.33806825e-01
3.56913805e-01 -7.44081512e-02 -1.31444573e-01 5.93857467e-01
1.15764809e+00 4.64248449e-01 6.44219100e-01 -1.79804325e-01
4.82544988e-01 4.65005845e-01 7.16534615e-01 3.94568980e-01
-3.49301547e-01 4.91452545e-01 -1.56027507e-02 -2.89163679e-01
-4.16859388e-01 -1.73477307e-02 6.87847197e-01 1.68451619e+00
4.72085714e-01 -4.30022627e-01 -4.52253610e-01 3.59499127e-01
-1.64522541e+00 -7.39578903e-01 -5.92398584e-01 2.26686740e+00
4.71695513e-01 1.59378156e-01 -2.45787188e-01 3.50420356e-01
8.49561632e-01 2.12757647e-01 -5.31302691e-01 -5.54449975e-01
6.39333278e-02 4.20230389e-01 5.14888406e-01 7.13422179e-01
-8.22337806e-01 6.03141904e-01 6.13931847e+00 1.11926198e+00
-1.46393931e+00 1.30512014e-01 -8.38622749e-02 -2.22139910e-01
1.31244212e-01 -1.37198359e-01 -5.21380544e-01 2.05412462e-01
1.22251165e+00 -9.42488685e-02 4.99084204e-01 2.71304220e-01
4.28906918e-01 7.78513476e-02 -6.65671051e-01 9.38655794e-01
-8.88087749e-02 -3.79173487e-01 -4.85069543e-01 -2.91086972e-01
-6.00415803e-02 -7.08034188e-02 1.76901054e-02 4.19441342e-01
-2.05784902e-01 -5.37752092e-01 1.03492653e+00 4.97241378e-01
6.32873297e-01 -9.74382162e-01 5.64050555e-01 3.22566807e-01
-1.79282629e+00 -2.82093227e-01 -3.52720737e-01 -5.22051752e-02
3.55568707e-01 8.96447539e-01 -4.73275900e-01 4.80283111e-01
6.17547691e-01 2.21996397e-01 6.54680878e-02 1.30521989e+00
-1.10325888e-02 9.97969449e-01 -5.94266057e-01 2.14058533e-01
3.77469003e-01 -1.32381484e-01 1.11463571e+00 1.59834623e+00
4.77979243e-01 4.69883472e-01 6.84911106e-03 3.83188069e-01
7.51511604e-02 3.75348441e-02 -3.08723420e-01 6.43042028e-02
4.81401861e-01 1.33791161e+00 -4.29740399e-01 -1.82837412e-01
-2.47139335e-01 8.15589488e-01 -2.16673940e-01 5.28393269e-01
-9.03780699e-01 -1.13384855e+00 6.01405799e-01 -1.47222802e-01
3.64715695e-01 -5.39178610e-01 1.04985915e-01 -7.14524806e-01
-4.70554903e-02 -9.10178244e-01 -2.91481949e-02 -6.60123348e-01
-8.12476575e-01 7.96568036e-01 -5.68888485e-01 -1.37370360e+00
4.84694056e-02 -2.06078216e-01 -7.04013526e-01 1.08640563e+00
-1.34992135e+00 -4.20855880e-01 -2.25444153e-01 6.07723176e-01
4.64047313e-01 1.91255473e-02 4.96948689e-01 9.47477162e-01
-4.84613001e-01 7.36026585e-01 2.74431527e-01 1.07932396e-01
1.10296500e+00 -6.99217975e-01 2.26245105e-01 1.12886024e+00
-1.78209975e-01 5.89820921e-01 8.11054826e-01 -7.42895484e-01
-1.20123827e+00 -8.72419596e-01 5.77309132e-01 6.43669903e-01
4.33474153e-01 -7.11290240e-01 -1.18704462e+00 2.59075575e-02
5.93440056e-01 -2.75394857e-01 4.87857550e-01 -2.74986982e-01
-2.64817536e-01 -6.71869695e-01 -6.81749403e-01 6.43474579e-01
6.91596210e-01 -5.09114504e-01 -4.24316198e-01 -3.81493151e-01
8.89725804e-01 -3.57717544e-01 -3.99326146e-01 3.57286870e-01
8.64212573e-01 -8.70740354e-01 8.55938673e-01 -1.30102392e-02
-3.27726454e-02 -4.57741231e-01 -2.27523744e-02 -1.52470601e+00
-4.83980268e-01 -8.23907912e-01 2.28178799e-01 1.44756031e+00
1.28613412e-01 -6.02609694e-01 3.17073643e-01 2.60565221e-01
-4.31782246e-01 -8.04737285e-02 -9.65993881e-01 -7.66234696e-01
-4.79953945e-01 -5.41901052e-01 3.17249894e-01 4.51662481e-01
-1.44388646e-01 4.69743192e-01 -7.27452159e-01 4.98420388e-01
6.38560951e-01 -3.78713787e-01 4.99284178e-01 -8.30157876e-01
-2.94731110e-01 -2.68680960e-01 -7.99081922e-02 -1.35231948e+00
-5.23850285e-02 -4.80374336e-01 6.53507352e-01 -1.02249169e+00
-3.94229621e-01 -3.01663101e-01 -6.58408046e-01 -1.95905324e-02
-2.41897941e-01 1.07638068e-01 4.49145764e-01 -8.78443941e-02
-2.53009021e-01 7.89155543e-01 9.08276975e-01 -5.46754263e-02
-5.52726626e-01 1.29676983e-01 -1.03318617e-01 6.99466348e-01
5.78637362e-01 -8.47654343e-01 -3.88571709e-01 -6.28640354e-01
-2.32331485e-01 5.24892151e-01 2.22839024e-02 -1.40470803e+00
6.97813392e-01 3.50197166e-01 2.54483730e-01 -7.83723474e-01
5.92877150e-01 -1.19181359e+00 2.50225633e-01 5.82937837e-01
-3.49707425e-01 -2.24181294e-01 6.51676714e-01 6.37703001e-01
-5.38554192e-01 -9.44636986e-02 7.59248614e-01 2.53987432e-01
-2.49202177e-01 -1.22466929e-01 -8.83745790e-01 -3.46475720e-01
3.91974181e-01 6.92143440e-02 -8.75775963e-02 -6.64676011e-01
-6.31544113e-01 1.90150633e-01 -5.28380632e-01 3.47096235e-01
8.62078011e-01 -1.15337157e+00 -7.41079628e-01 3.60335976e-01
-5.29533505e-01 -4.71249133e-01 8.57066095e-01 1.06177771e+00
-1.35304108e-01 4.46818560e-01 -1.06155351e-01 -3.75407577e-01
-1.60971498e+00 3.05133611e-01 6.70055330e-01 2.73640081e-02
-8.20012689e-01 6.89293504e-01 4.11607653e-01 -6.79165572e-02
5.59512556e-01 -2.37703547e-01 -9.27807242e-02 -1.84662730e-01
7.79567122e-01 6.69646204e-01 8.89910907e-02 -6.10038340e-01
-3.74125481e-01 4.40032840e-01 5.47273383e-02 -5.84470570e-01
1.19506192e+00 -4.64823455e-01 -3.72643061e-02 3.73265594e-01
1.16599107e+00 5.98322749e-01 -9.32907343e-01 -4.60638523e-01
-4.88369495e-01 -4.69331563e-01 6.16250098e-01 -5.77438891e-01
-1.25697100e+00 1.13838995e+00 1.09790039e+00 3.35608460e-02
1.52245069e+00 -9.21551585e-01 1.35699415e+00 1.28162473e-01
-8.29581022e-02 -9.68453348e-01 7.47891143e-02 5.45778930e-01
7.25721598e-01 -4.51660752e-01 -2.72175372e-01 -2.39002123e-01
-2.97208637e-01 1.25568557e+00 5.41250765e-01 -1.76877975e-01
7.88770020e-01 8.19471359e-01 4.52320814e-01 2.24954277e-01
-5.44275820e-01 1.63583588e-02 2.12792635e-01 1.77987635e-01
3.42784047e-01 -2.47859821e-01 -3.98129225e-01 8.73792231e-01
1.58508286e-01 -2.09346652e-01 3.76486808e-01 7.86999822e-01
-6.03757083e-01 -9.25524354e-01 -7.69238353e-01 2.21278280e-01
-5.21763504e-01 -3.12149763e-01 -4.07138497e-01 2.02941209e-01
-1.19069405e-01 1.44847322e+00 1.03543177e-01 -7.61438906e-01
7.02967644e-01 1.15302309e-01 3.74127999e-02 -4.19888310e-02
-9.62141812e-01 1.10862494e+00 2.44413428e-02 -4.03886020e-01
-1.02338463e-01 -2.82995135e-01 -1.46670580e+00 -2.84841835e-01
-8.29460144e-01 2.40679398e-01 7.06135035e-01 7.67268002e-01
3.74648303e-01 1.12755132e+00 7.51537025e-01 -7.11215496e-01
-5.38582087e-01 -1.14671826e+00 -8.32183957e-01 4.82631940e-03
7.46974349e-01 -1.87039748e-01 -5.10680914e-01 -3.21296811e-01] | [15.037907600402832, 5.9467315673828125] |
ca4e56b7-9607-4d92-9517-af79ba7aa0e5 | automatic-discovery-of-interpretable-planning | 2005.11730 | null | https://arxiv.org/abs/2005.11730v3 | https://arxiv.org/pdf/2005.11730v3.pdf | Automatic Discovery of Interpretable Planning Strategies | When making decisions, people often overlook critical information or are overly swayed by irrelevant information. A common approach to mitigate these biases is to provide decision-makers, especially professionals such as medical doctors, with decision aids, such as decision trees and flowcharts. Designing effective decision aids is a difficult problem. We propose that recently developed reinforcement learning methods for discovering clever heuristics for good decision-making can be partially leveraged to assist human experts in this design process. One of the biggest remaining obstacles to leveraging the aforementioned methods is that the policies they learn are opaque to people. To solve this problem, we introduce AI-Interpret: a general method for transforming idiosyncratic policies into simple and interpretable descriptions. Our algorithm combines recent advances in imitation learning and program induction with a new clustering method for identifying a large subset of demonstrations that can be accurately described by a simple, high-performing decision rule. We evaluate our new algorithm and employ it to translate information-acquisition policies discovered through metalevel reinforcement learning. The results of large behavioral experiments showed that prividing the decision rules generated by AI-Interpret as flowcharts significantly improved people's planning strategies and decisions across three diferent classes of sequential decision problems. Moreover, another experiment revealed that this approach is significantly more effective than training people by giving them performance feedback. Finally, a series of ablation studies confirmed that AI-Interpret is critical to the discovery of interpretable decision rules. We conclude that the methods and findings presented herein are an important step towards leveraging automatic strategy discovery to improve human decision-making. | ['Julian Skirzyński', 'Falk Lieder', 'Frederic Becker'] | 2020-05-24 | null | null | null | null | ['program-induction'] | ['computer-code'] | [ 1.64264828e-01 2.58482546e-01 -5.70567250e-01 -3.08391511e-01
-5.86787939e-01 -7.50919461e-01 4.10149813e-01 1.05809592e-01
-5.42268932e-01 9.12691534e-01 1.96287811e-01 -7.27379858e-01
-1.76861510e-01 -3.43751162e-01 -4.49642539e-01 -2.72635937e-01
1.44786969e-01 8.28649342e-01 -9.67254192e-02 -1.79620340e-01
6.23968065e-01 2.84595579e-01 -1.50753009e+00 2.39275947e-01
1.28952789e+00 2.42281929e-01 2.22178504e-01 6.87871516e-01
1.71425179e-01 1.02382886e+00 -6.31836951e-01 -2.15098828e-01
1.65436372e-01 -6.00173771e-01 -7.47035325e-01 2.89635174e-03
1.15008064e-01 -9.89054382e-01 4.32087146e-02 8.36639404e-01
1.99140742e-01 1.80850372e-01 7.73933232e-01 -1.24925685e+00
-6.03827536e-01 8.92356634e-01 -2.13225216e-01 7.21839294e-02
6.38411880e-01 8.27073216e-01 9.30114269e-01 -7.73199946e-02
5.45306802e-01 1.42259622e+00 4.69635487e-01 9.79475856e-01
-1.62472749e+00 -7.29871333e-01 4.42000568e-01 2.21480072e-01
-9.75997090e-01 -5.27518928e-01 6.56889260e-01 -7.96383381e-01
9.55832601e-01 1.01774044e-01 9.27113652e-01 1.56226790e+00
1.38063028e-01 1.04041123e+00 1.56502330e+00 -4.36682314e-01
5.66539526e-01 4.13141608e-01 2.22372860e-01 9.02222812e-01
6.34132743e-01 6.93874121e-01 -4.46404994e-01 -3.38273764e-01
9.32692885e-01 3.42878923e-02 -5.46413995e-02 -4.49192077e-01
-1.14597058e+00 9.56056058e-01 8.75429213e-02 2.83738345e-01
-3.37257713e-01 2.81314254e-01 -3.09497248e-02 5.06055415e-01
-3.52310181e-01 1.28012884e+00 -3.77879262e-01 -5.28476536e-01
-4.47448045e-01 7.27334201e-01 9.55686390e-01 7.81816006e-01
6.13071442e-01 3.70161086e-02 -4.20151711e-01 3.89666408e-01
3.48448932e-01 6.23342335e-01 4.57463771e-01 -1.55317712e+00
2.21449777e-01 7.23045290e-01 7.07274258e-01 -7.84227729e-01
-5.38719416e-01 -1.05304360e-01 -5.13667203e-02 6.56561971e-01
6.57386780e-01 -5.57567120e-01 -7.72120357e-01 1.79859650e+00
5.32255955e-02 -5.31829536e-01 -8.73301476e-02 8.65980804e-01
-8.42031464e-02 5.91817051e-02 4.10294056e-01 -5.49131259e-02
1.22609782e+00 -6.51772797e-01 -4.59026217e-01 -3.20478052e-01
7.92780459e-01 -2.35024840e-01 1.35156024e+00 4.21291798e-01
-7.00029016e-01 -4.17592257e-01 -9.97545719e-01 4.18815941e-01
6.02247678e-02 7.14945514e-03 1.01205909e+00 8.23866367e-01
-8.29385459e-01 8.41817200e-01 -8.67919326e-01 -4.70257252e-01
6.19636834e-01 4.70790893e-01 1.93045914e-01 1.47525415e-01
-8.84085596e-01 9.57622409e-01 2.10309878e-01 -3.69037688e-01
-1.28432024e+00 -3.61028552e-01 -6.48002267e-01 -1.06127153e-03
7.75045455e-01 -9.73152697e-01 1.81779277e+00 -1.25916564e+00
-1.75484657e+00 4.91873115e-01 -6.12797476e-02 -4.29122925e-01
5.75136185e-01 -2.78094321e-01 1.02698959e-01 1.44420609e-01
3.77133012e-01 6.59597933e-01 1.16599655e+00 -1.22909319e+00
-7.69402921e-01 -3.50691199e-01 2.33885422e-01 3.67792875e-01
-6.55307919e-02 3.29688266e-02 3.27762127e-01 -3.86449188e-01
-5.97543299e-01 -1.11876905e+00 -4.17971253e-01 -2.80683011e-01
-4.82192427e-01 -5.53071439e-01 3.22971821e-01 -4.97254103e-01
1.09455347e+00 -1.77330470e+00 1.16405018e-01 3.20026308e-01
4.72836673e-01 2.17378482e-01 4.51519340e-02 2.35534862e-01
3.10864300e-01 4.70151305e-01 2.17209458e-02 -2.32092589e-02
4.00316119e-01 1.46088228e-01 -2.72805065e-01 8.16200152e-02
-1.66044887e-02 9.07792211e-01 -1.04850566e+00 -1.85721949e-01
2.53853232e-01 -1.03997909e-01 -7.36063242e-01 4.73999888e-01
-4.51818556e-01 8.60076070e-01 -8.64482164e-01 5.71687639e-01
-1.72060937e-01 -2.85735458e-01 4.12692577e-01 6.70613050e-01
-4.98730727e-02 5.12453496e-01 -8.21699619e-01 1.22306371e+00
-2.36339331e-01 3.76933873e-01 -1.46218345e-01 -7.44723737e-01
4.83018160e-01 2.40163039e-02 9.81977060e-02 -6.05516315e-01
1.30303934e-01 1.25401244e-01 4.70998704e-01 -7.95150578e-01
6.74696118e-02 -1.14731938e-01 -1.90764934e-01 9.29904580e-01
-3.11150044e-01 -2.27604270e-01 3.36909406e-02 1.20315045e-01
1.23998165e+00 3.76744509e-01 6.33052707e-01 -2.12697208e-01
-4.93681841e-02 4.55801636e-01 7.01883078e-01 1.35788131e+00
-5.32828450e-01 -3.64270918e-02 6.68318152e-01 -4.68489170e-01
-9.50408936e-01 -8.05803597e-01 4.08587366e-01 1.19957805e+00
-2.63470739e-01 -3.35442692e-01 -9.26061213e-01 -8.66639137e-01
3.46525431e-01 1.31546390e+00 -6.39659762e-01 -3.20363343e-01
-4.34534609e-01 -1.60609379e-01 5.22883952e-01 5.65482795e-01
5.78217864e-01 -1.10318708e+00 -1.23576236e+00 1.46501362e-01
-3.57900485e-02 -7.67505169e-01 -4.74464178e-01 5.16081862e-02
-8.13248813e-01 -1.20384526e+00 -4.08812284e-01 -5.82709074e-01
8.87204170e-01 2.99898207e-01 8.68155420e-01 4.83153582e-01
-1.07740415e-02 6.87873304e-01 -2.53365755e-01 -5.79522014e-01
-7.47731626e-01 -7.71390498e-02 4.46091712e-01 -5.38818181e-01
6.21658266e-01 -4.76576716e-01 -5.16189933e-01 4.22365397e-01
-1.47264659e-01 2.40442500e-01 7.96656966e-01 7.99072027e-01
3.51910130e-03 8.14585686e-02 5.18710077e-01 -1.09650004e+00
1.23611569e+00 -1.28496468e-01 -6.41915679e-01 2.23966882e-01
-1.08128560e+00 5.86739063e-01 6.63117826e-01 -5.98529518e-01
-1.14798605e+00 5.67171350e-02 3.27714354e-01 -7.77976662e-02
-5.77473938e-01 1.70766249e-01 1.21064469e-01 7.00532719e-02
9.77673590e-01 4.91840504e-02 6.03564419e-02 -2.24170595e-01
2.27522314e-01 5.52268982e-01 2.48247571e-02 -1.11052585e+00
7.21556783e-01 1.12587735e-02 -4.68792558e-01 -6.08370543e-01
-5.29380500e-01 -9.06155556e-02 -2.13397741e-01 -1.40174210e-01
8.86245251e-01 -6.11047626e-01 -1.29910684e+00 2.77468741e-01
-8.65463138e-01 -1.00273240e+00 -8.17692950e-02 4.24001396e-01
-8.78537834e-01 1.82490095e-01 -4.29516286e-01 -9.93589401e-01
6.06191978e-02 -1.30090761e+00 6.42245471e-01 3.81000489e-01
-1.11260533e+00 -7.10840940e-01 8.99061859e-02 5.98467290e-01
4.10678118e-01 -1.23474337e-01 1.34045565e+00 -9.64714944e-01
-7.10174322e-01 2.05634370e-01 2.10282817e-01 -2.04400569e-02
1.21279493e-01 -7.55703524e-02 -6.36277020e-01 -9.55283716e-02
-4.94269617e-02 -6.39317334e-01 4.64842319e-01 3.57882321e-01
9.26918864e-01 -6.23168886e-01 -4.55411226e-01 2.57712394e-01
7.29046404e-01 5.94464719e-01 1.32388383e-01 3.92885596e-01
4.80418742e-01 7.07329631e-01 4.55096453e-01 3.37554634e-01
6.53476000e-01 7.90277243e-01 -2.70077497e-01 3.49300325e-01
1.24336027e-01 -7.83000648e-01 6.58445299e-01 2.21472420e-02
-3.21774453e-01 2.67507076e-01 -9.53790069e-01 2.71568567e-01
-2.10436130e+00 -1.12876368e+00 5.32166600e-01 2.30785203e+00
8.94579411e-01 3.47431630e-01 5.86857617e-01 -2.01284230e-01
4.40037847e-01 -4.65707451e-01 -9.35031176e-01 -4.35902059e-01
4.75803733e-01 -6.63264915e-02 3.27682078e-01 5.58053732e-01
-7.82290936e-01 1.06682098e+00 7.12538671e+00 2.18586370e-01
-7.50353932e-01 -2.34399706e-01 6.28166735e-01 7.68365711e-02
-4.41033989e-01 5.33374995e-02 -8.26207399e-01 3.73887181e-01
9.98421073e-01 -4.27101441e-02 8.93126488e-01 1.06980300e+00
5.85308313e-01 -4.24891353e-01 -1.59543014e+00 8.05505216e-01
-2.87355423e-01 -1.10818601e+00 -7.02903494e-02 7.21820742e-02
6.51441455e-01 -6.53524935e-01 1.41046852e-01 7.38892555e-01
1.35127342e+00 -1.25671422e+00 6.93177164e-01 2.64576077e-01
4.00043458e-01 -5.28028548e-01 3.23161811e-01 5.96119165e-01
-4.51000571e-01 -5.55760980e-01 -1.09644584e-01 -5.43508470e-01
-3.62724990e-01 -1.22957334e-01 -1.36534584e+00 -3.14016134e-01
3.83496135e-01 4.97713208e-01 -4.49420482e-01 6.27346516e-01
-9.77161705e-01 7.40114987e-01 -7.22269416e-02 -5.59235275e-01
-3.50809619e-02 1.06079124e-01 5.63420296e-01 7.56698310e-01
-2.50773579e-02 2.67545879e-01 3.15446675e-01 1.37115955e+00
2.81086922e-01 -3.10475826e-01 -8.28675687e-01 -4.34814751e-01
6.51168108e-01 7.73433208e-01 -6.76675558e-01 -3.12363386e-01
-9.95112732e-02 7.74902821e-01 5.49664497e-01 6.95841491e-01
-5.21817982e-01 -9.18407291e-02 7.60002494e-01 -7.39322156e-02
2.22656295e-01 -3.05774719e-01 -6.57612920e-01 -1.07053459e+00
-2.43626580e-01 -1.53887761e+00 1.48928270e-01 -5.06232142e-01
-9.90433514e-01 5.65534160e-02 2.82465518e-01 -7.77725339e-01
-7.32068062e-01 -5.31886756e-01 -5.50283372e-01 5.72456479e-01
-1.15162599e+00 -7.25024223e-01 -7.17370212e-02 3.90165329e-01
6.96786046e-01 -3.71926516e-01 7.30301857e-01 -4.70908970e-01
-7.84339070e-01 5.81183791e-01 -3.41289550e-01 3.32207382e-02
6.03870630e-01 -1.33803904e+00 1.49143562e-01 6.06482863e-01
-9.77628604e-02 1.22040355e+00 7.58977473e-01 -8.37528825e-01
-1.34637249e+00 -4.50077325e-01 4.48089778e-01 -7.68107295e-01
4.02380198e-01 -2.46458188e-01 -5.90515137e-01 9.07608986e-01
4.62563708e-02 -1.00164974e+00 7.49873817e-01 4.05496657e-01
-2.51547724e-01 2.24691883e-01 -1.08455884e+00 1.05074906e+00
1.07532263e+00 -4.99409288e-01 -1.13405454e+00 1.23314023e-01
6.97079837e-01 -8.77418146e-02 -3.69447529e-01 -2.13978365e-01
5.55879414e-01 -8.69771302e-01 6.52857065e-01 -9.93314564e-01
4.88526344e-01 -2.26742566e-01 2.40327343e-01 -1.57394445e+00
-5.01840651e-01 -1.25441837e+00 -1.09502085e-01 7.01944292e-01
5.22952795e-01 -5.63030779e-01 6.60801232e-01 1.20823026e+00
1.32263228e-01 -3.42752904e-01 -4.10113633e-01 -5.49537182e-01
-7.16126710e-03 -3.24081182e-01 4.54752713e-01 8.32675219e-01
4.44714040e-01 4.71695334e-01 -2.66433626e-01 -2.65285492e-01
7.79833257e-01 1.00666925e-01 1.01487935e+00 -1.15545189e+00
-7.34262884e-01 -7.42683649e-01 1.14295408e-01 -1.14785886e+00
2.28447199e-01 -6.89486027e-01 1.88693851e-01 -1.34325540e+00
2.29683951e-01 -6.30040228e-01 -4.72598039e-02 6.68321669e-01
-3.11063707e-01 -8.43001366e-01 3.21991205e-01 3.40038985e-01
-5.84061265e-01 2.71917194e-01 1.24094534e+00 -6.87482134e-02
-6.98656499e-01 4.19759542e-01 -1.42288899e+00 1.03212833e+00
1.02510810e+00 -5.16310394e-01 -5.83711326e-01 -2.10736826e-01
2.29118779e-01 8.24026912e-02 4.81624246e-01 -9.81530428e-01
2.39614695e-01 -6.89852893e-01 4.78366017e-01 5.50263599e-02
-9.80910882e-02 -6.77308142e-01 -2.47394323e-01 7.83449829e-01
-7.02378690e-01 5.31753562e-02 1.88218027e-01 6.17428184e-01
5.90827644e-01 -1.11126333e-01 4.61079061e-01 -4.04580653e-01
-9.55084562e-01 -3.48143280e-01 -1.13451779e+00 3.72156873e-02
1.16823602e+00 -1.69587508e-01 -4.05066490e-01 -7.91714191e-01
-6.13758147e-01 4.71725702e-01 5.97973704e-01 1.92258492e-01
6.83397055e-01 -7.81558752e-01 -4.80136305e-01 1.83438763e-01
-3.84453535e-02 -3.23747605e-01 -1.83308437e-01 8.00334215e-01
-2.36816168e-01 6.34865642e-01 -4.74580437e-01 -2.82646358e-01
-1.09869373e+00 4.67031121e-01 4.02539700e-01 -3.17139953e-01
-5.70734382e-01 5.80475450e-01 1.35416448e-01 -4.09902185e-01
3.78108621e-01 -5.99815249e-01 -9.40643996e-02 -2.90381521e-01
6.81178331e-01 2.58768976e-01 -6.03208423e-01 8.30196142e-02
-2.77941167e-01 2.42422804e-01 -3.78070265e-01 -3.28547269e-01
1.15373051e+00 -3.54896337e-02 4.00905311e-01 3.06338280e-01
2.81505555e-01 -6.53123409e-02 -1.62883675e+00 2.29050312e-02
1.78158447e-01 -4.41732287e-01 -3.82317603e-01 -1.27143240e+00
-2.07902730e-01 6.69460356e-01 2.76958436e-01 9.70730837e-03
6.26827002e-01 -1.30045548e-01 3.30875516e-01 1.11503339e+00
7.65698731e-01 -1.38781905e+00 4.46889877e-01 1.91859215e-01
5.55707932e-01 -1.42241299e+00 5.43345734e-02 1.35728698e-02
-1.01614499e+00 9.97497439e-01 8.86384189e-01 5.28487973e-02
-1.39422268e-02 -1.30363464e-01 3.26374322e-02 -2.00557247e-01
-8.87224495e-01 -2.00184271e-01 1.81692958e-04 1.06014872e+00
1.33381233e-01 4.97208595e-01 -3.28219235e-02 7.59748459e-01
-3.07556570e-01 3.05698663e-01 6.80548966e-01 1.05305707e+00
-7.10187256e-01 -1.10560620e+00 -4.87334251e-01 4.55584794e-01
-1.79297328e-01 2.43674651e-01 -7.52540112e-01 8.58602762e-01
-2.06213564e-01 1.13142490e+00 -2.24139094e-01 -6.40248179e-01
1.76869452e-01 2.29189888e-01 5.52032471e-01 -7.15706706e-01
-5.51720023e-01 -3.07472587e-01 2.74734914e-01 -7.25380361e-01
-1.84882596e-01 -9.33207393e-01 -1.31701970e+00 -3.88725698e-01
3.16761672e-01 1.39384149e-02 1.91708684e-01 1.10175109e+00
4.02946383e-01 4.87730891e-01 3.24596196e-01 -4.32753742e-01
-1.13391113e+00 -6.37104034e-01 -1.73339203e-01 4.16842520e-01
3.63947034e-01 -8.90003324e-01 -1.86267883e-01 -1.70144975e-01] | [4.214734077453613, 1.7548210620880127] |
0fc9add8-b05a-4bcb-b267-489765963b6b | self-supervised-3d-face-reconstruction-via-1 | 2110.04800 | null | https://arxiv.org/abs/2110.04800v1 | https://arxiv.org/pdf/2110.04800v1.pdf | Self-Supervised 3D Face Reconstruction via Conditional Estimation | We present a conditional estimation (CEST) framework to learn 3D facial parameters from 2D single-view images by self-supervised training from videos. CEST is based on the process of analysis by synthesis, where the 3D facial parameters (shape, reflectance, viewpoint, and illumination) are estimated from the face image, and then recombined to reconstruct the 2D face image. In order to learn semantically meaningful 3D facial parameters without explicit access to their labels, CEST couples the estimation of different 3D facial parameters by taking their statistical dependency into account. Specifically, the estimation of any 3D facial parameter is not only conditioned on the given image, but also on the facial parameters that have already been derived. Moreover, the reflectance symmetry and consistency among the video frames are adopted to improve the disentanglement of facial parameters. Together with a novel strategy for incorporating the reflectance symmetry and consistency, CEST can be efficiently trained with in-the-wild video clips. Both qualitative and quantitative experiments demonstrate the effectiveness of CEST. | ['Rita Singh', 'Bhiksha Raj', 'Weiyang Liu', 'Yandong Wen'] | 2021-10-10 | self-supervised-3d-face-reconstruction-via | http://openaccess.thecvf.com//content/ICCV2021/html/Wen_Self-Supervised_3D_Face_Reconstruction_via_Conditional_Estimation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Wen_Self-Supervised_3D_Face_Reconstruction_via_Conditional_Estimation_ICCV_2021_paper.pdf | iccv-2021-1 | ['3d-face-reconstruction', 'face-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 9.51384846e-03 7.44119063e-02 -6.44075945e-02 -6.92261159e-01
-6.18621707e-01 -4.45815921e-01 4.63713437e-01 -5.27627170e-01
-2.35006034e-01 2.20098808e-01 2.52505720e-01 4.18378174e-01
8.20821710e-03 -3.20806026e-01 -6.25007689e-01 -9.94518638e-01
1.81161940e-01 3.06127220e-01 -3.40840518e-01 1.34463802e-01
-1.44322470e-01 8.18363190e-01 -1.72674239e+00 1.98356081e-02
4.24281657e-01 1.08723891e+00 -5.19196652e-02 4.09734994e-01
7.07419142e-02 4.98210013e-01 -2.28649795e-01 -4.42756891e-01
3.10682893e-01 -3.71584505e-01 -2.72864252e-01 8.88829231e-01
6.59896612e-01 -6.75361335e-01 -9.48825479e-02 9.49697912e-01
3.21376383e-01 7.31684864e-02 9.28771377e-01 -1.26897848e+00
-2.76910901e-01 -2.41981402e-01 -5.37779510e-01 -5.26647568e-01
4.08748984e-01 3.89708541e-02 9.17865217e-01 -1.22909164e+00
6.73013210e-01 1.29596293e+00 3.78636360e-01 6.05700493e-01
-1.27510130e+00 -7.61990666e-01 2.25710541e-01 7.90987238e-02
-1.57401597e+00 -9.53894079e-01 1.18903923e+00 -6.42276943e-01
1.60014004e-01 -7.44822621e-02 8.10967922e-01 9.58297372e-01
-3.25130314e-01 6.63856089e-01 9.24061418e-01 -4.47621167e-01
1.84246555e-01 1.61541939e-01 -4.48942900e-01 1.05535185e+00
-2.00033799e-01 7.87872821e-02 -5.84444165e-01 -5.54965585e-02
1.01218283e+00 -9.06602740e-02 -2.17164278e-01 -7.28662550e-01
-9.74875987e-01 5.71488202e-01 5.38517646e-02 -1.27227321e-01
-3.13385367e-01 -1.86706364e-01 2.28632260e-02 1.18708815e-02
5.80009639e-01 -1.43826529e-01 -4.62218940e-01 2.77101815e-01
-8.10198426e-01 3.56434472e-02 5.19905627e-01 1.10694802e+00
1.22756827e+00 6.17992878e-02 1.02770418e-01 8.73217344e-01
6.34107053e-01 8.39596570e-01 1.88014973e-02 -1.36567342e+00
1.79659963e-01 5.90870261e-01 6.24930039e-02 -1.17326558e+00
-2.20551223e-01 3.12779546e-02 -5.78634560e-01 1.44411385e-01
3.47387582e-01 8.13208595e-02 -7.61066437e-01 1.91362572e+00
8.14197361e-01 2.05079287e-01 -1.01146899e-01 8.82536948e-01
8.64307523e-01 3.21847856e-01 -1.30649969e-01 -4.80664611e-01
1.09665537e+00 -6.33895218e-01 -7.92080820e-01 2.57224888e-02
2.94384360e-01 -7.21383095e-01 5.80591023e-01 2.62244761e-01
-1.11296344e+00 -6.53669178e-01 -7.63979495e-01 -2.00658485e-01
1.94411069e-01 6.61169767e-01 2.24862263e-01 5.87262928e-01
-7.77049184e-01 2.20421568e-01 -8.43680263e-01 -6.86498433e-02
3.46815139e-01 3.73607039e-01 -7.92030215e-01 -2.00051725e-01
-6.24916852e-01 5.58638632e-01 -1.10729318e-02 5.59299827e-01
-9.70995665e-01 -5.15746772e-01 -1.07820654e+00 -1.01775922e-01
4.64560717e-01 -6.18510902e-01 9.58477914e-01 -1.19382954e+00
-1.94941676e+00 1.15089989e+00 -4.67777729e-01 4.34355170e-01
5.19644380e-01 -8.03069770e-02 -2.45718360e-01 6.11916304e-01
-1.68624744e-02 6.72903061e-01 1.32987452e+00 -1.30273318e+00
-1.23790957e-01 -6.00641847e-01 -1.38053328e-01 2.71179736e-01
-2.63109118e-01 9.33408886e-02 -8.90674770e-01 -3.48917812e-01
3.40155303e-01 -9.14213777e-01 1.74438983e-01 6.50347650e-01
-2.92395473e-01 1.45865560e-01 4.96304423e-01 -9.00360167e-01
5.35446644e-01 -2.36418509e+00 4.93028760e-01 3.20013463e-01
1.36942357e-01 -8.94627068e-03 -3.59613001e-01 -3.64065403e-03
-1.23676881e-01 -2.15794936e-01 1.24399015e-03 -6.83197260e-01
-3.62576209e-02 2.35204935e-01 1.09989941e-01 7.78363824e-01
4.13872242e-01 6.37928903e-01 -6.81913614e-01 -7.64762402e-01
4.38303232e-01 8.43533278e-01 -7.59159744e-01 6.02345109e-01
-1.51589051e-01 8.50411952e-01 -6.22577250e-01 6.62183523e-01
8.79637241e-01 -1.06912114e-01 3.35426450e-01 -7.20833302e-01
1.54170468e-01 -8.76769125e-02 -1.08985102e+00 1.60370135e+00
-4.52468991e-01 2.61325985e-01 3.75429958e-01 -9.99166489e-01
9.06508923e-01 5.08401215e-01 7.30252683e-01 -4.27295625e-01
2.93431699e-01 1.44491762e-01 -4.75975066e-01 -9.09748316e-01
-8.55887830e-02 -3.25261235e-01 2.73353577e-01 3.59141141e-01
4.76920068e-01 -2.16117516e-01 -7.92977586e-02 -8.96351635e-02
2.04491019e-01 5.95368564e-01 2.14041680e-01 -5.20509891e-02
8.05450141e-01 -8.08222055e-01 7.30942667e-01 -2.38836735e-01
1.77772362e-02 7.36224294e-01 5.70121467e-01 -2.81613886e-01
-1.00868213e+00 -1.02815866e+00 -8.32232460e-02 5.99623859e-01
-1.25811681e-01 -3.99534464e-01 -8.90444696e-01 -6.89366281e-01
-1.44379064e-01 1.00904599e-01 -7.20100343e-01 3.65547948e-02
-3.83612841e-01 -3.80928874e-01 9.00578946e-02 3.10916692e-01
4.46158797e-01 -5.55017710e-01 -1.51561588e-01 -2.07246721e-01
-3.10606837e-01 -1.61594665e+00 -7.77306437e-01 -4.28513318e-01
-7.57265210e-01 -1.18268156e+00 -5.10149539e-01 -5.68571627e-01
1.20534945e+00 3.08877558e-01 6.93144619e-01 2.06214577e-01
-1.89722449e-01 6.07727706e-01 -6.17823005e-02 9.49551119e-04
-3.75229537e-01 -4.53959763e-01 9.00188982e-02 8.24380577e-01
-1.19410746e-01 -7.89327621e-01 -5.21146894e-01 6.23406172e-01
-7.79340267e-01 3.37548375e-01 3.24552745e-01 6.54483438e-01
7.36056566e-01 -9.50720087e-02 1.18579790e-01 -7.24992156e-01
-3.97416621e-01 -1.74120501e-01 -6.81995869e-01 4.05303717e-01
-7.19481409e-02 -1.16209416e-02 3.44328374e-01 -5.08188367e-01
-1.40007675e+00 4.79086280e-01 -1.54826835e-01 -8.47874999e-01
-2.49920666e-01 2.20029458e-01 -8.28139246e-01 -4.75192107e-02
1.61540642e-01 2.93848254e-02 3.38819891e-01 -4.74210620e-01
3.68551016e-01 4.61280853e-01 3.88839006e-01 -7.32727945e-01
7.40140855e-01 8.56231689e-01 1.55150101e-01 -9.82709050e-01
-9.17294204e-01 -2.14920878e-01 -1.17657065e+00 -4.38082069e-01
9.37177658e-01 -1.23966241e+00 -6.85265064e-01 6.64381146e-01
-9.96454775e-01 -1.55809551e-01 -8.36896375e-02 6.29932761e-01
-6.47354424e-01 4.68951911e-01 -3.06582600e-01 -9.31412399e-01
1.00728557e-01 -1.25398088e+00 1.35572731e+00 5.33531457e-02
7.01192543e-02 -1.05925608e+00 -2.59151757e-01 6.11018479e-01
-1.38016760e-01 1.88242882e-01 8.95225227e-01 -1.37329876e-01
-6.81124568e-01 -1.05163522e-01 -1.50912151e-01 6.05065227e-01
4.10245210e-01 3.87541354e-01 -1.23905647e+00 -1.62417620e-01
1.76666737e-01 -1.35972634e-01 3.87788087e-01 4.29781377e-01
9.41303968e-01 -3.14134568e-01 4.54990752e-02 9.65378582e-01
1.02735817e+00 -1.76049352e-01 3.23944151e-01 -2.49321014e-01
8.19780529e-01 1.01595724e+00 4.85469103e-01 4.60040092e-01
4.94761080e-01 9.03289497e-01 4.33074594e-01 -7.60781020e-03
-2.49437779e-01 -4.82599109e-01 5.04025817e-01 8.30917835e-01
-2.83357680e-01 1.30293250e-01 -3.13028663e-01 9.28171873e-02
-1.43964577e+00 -7.01035857e-01 3.15210611e-01 2.28364420e+00
7.84929872e-01 -4.35951412e-01 9.70443860e-02 -6.71498403e-02
7.35597849e-01 8.32434073e-02 -6.26515567e-01 2.82792449e-01
-1.51335880e-01 -5.13987429e-02 -4.95967045e-02 6.08411431e-01
-7.61660337e-01 7.16781259e-01 6.15832424e+00 5.48419833e-01
-1.20567501e+00 -1.22077994e-01 5.59312224e-01 -1.02067739e-01
-5.83508074e-01 1.41325980e-01 -8.74871135e-01 2.93148100e-01
4.04284298e-01 2.70663649e-01 4.89416838e-01 5.61534882e-01
2.09736899e-01 2.55034547e-02 -1.11989450e+00 1.04117513e+00
3.96042705e-01 -9.26346481e-01 2.34220728e-01 1.79070503e-01
5.97659945e-01 -5.43016970e-01 7.66421035e-02 -1.18338943e-01
-2.99568921e-01 -7.38267958e-01 9.19766247e-01 7.42230475e-01
1.06783497e+00 -5.79497576e-01 3.81911874e-01 2.87261814e-01
-1.02696300e+00 1.60017759e-01 -1.47949710e-01 3.77589971e-01
7.36837164e-02 6.00861549e-01 -6.18828952e-01 6.09594464e-01
6.47942007e-01 1.09907305e+00 -3.58496934e-01 4.47625875e-01
-5.68431139e-01 2.94502139e-01 -2.78838962e-01 5.31778872e-01
-4.80159312e-01 -6.56209409e-01 4.51154083e-01 6.80457652e-01
3.92666459e-01 2.16098472e-01 -1.32797472e-02 9.03168082e-01
-6.14145882e-02 2.24486023e-01 -2.17286676e-01 -3.17862332e-02
2.72181183e-01 1.53530741e+00 -4.70555604e-01 -9.14663821e-02
-4.42350626e-01 7.91338444e-01 1.89541563e-01 5.73061049e-01
-6.40051663e-01 3.55214059e-01 8.07799876e-01 2.10385397e-01
5.44862628e-01 -2.36922055e-01 2.39982769e-01 -1.48031211e+00
1.52162358e-01 -7.47562289e-01 7.23850131e-02 -9.71659303e-01
-1.22590947e+00 5.25685191e-01 2.42110729e-01 -1.24595618e+00
-3.48326892e-01 -7.27723479e-01 -3.13374043e-01 5.98551810e-01
-1.71312523e+00 -1.38102329e+00 -3.19506347e-01 8.45278382e-01
1.29161432e-01 -1.38058573e-01 8.59999239e-01 2.00999871e-01
-8.59701753e-01 6.48179054e-01 -1.66423276e-01 1.57317445e-01
8.70413840e-01 -6.95247114e-01 -1.39035121e-01 4.85535800e-01
2.73654368e-02 5.35235643e-01 4.14253801e-01 -2.84768552e-01
-1.51315975e+00 -8.42164397e-01 7.23379135e-01 -2.93625891e-01
2.83091545e-01 -6.41486287e-01 -7.10633218e-01 7.00937510e-01
-3.31548035e-01 5.05385280e-01 7.97476947e-01 -2.32686717e-02
-7.87575841e-01 -4.69994426e-01 -1.05499315e+00 4.19203550e-01
1.04368305e+00 -7.97666132e-01 -1.34271339e-01 1.39821172e-01
2.83366561e-01 -4.52439338e-01 -9.51593459e-01 2.79044867e-01
9.48542714e-01 -1.11183286e+00 1.01678467e+00 -3.90986979e-01
2.60824442e-01 -3.22859764e-01 -3.66269827e-01 -9.97085631e-01
9.99150202e-02 -7.26949334e-01 8.89073834e-02 1.40988004e+00
-2.39086114e-02 -4.91081923e-01 6.61034763e-01 7.36505926e-01
1.96450040e-01 -4.16973531e-01 -9.20045733e-01 -5.18139064e-01
-2.36203626e-01 -3.16971958e-01 7.29902923e-01 7.19691277e-01
-3.61589283e-01 3.38057399e-01 -5.95283508e-01 3.73781383e-01
8.05387735e-01 3.15619558e-01 1.02098858e+00 -1.30264008e+00
-7.25448132e-02 -9.67231765e-03 -4.87901449e-01 -9.94368136e-01
6.56334758e-01 -6.83100939e-01 -3.37448306e-02 -7.84520566e-01
3.33003193e-01 -2.77584940e-01 7.89674222e-02 4.13563371e-01
-6.36824593e-03 2.86165506e-01 9.96502414e-02 1.99931219e-01
-3.94175738e-01 8.52995396e-01 1.51170921e+00 2.17789561e-01
6.09022379e-02 4.88004684e-02 -3.90652120e-01 8.95299852e-01
2.83617467e-01 -2.65308350e-01 -4.32173461e-01 -4.88308430e-01
1.79732874e-01 4.49444532e-01 4.82348621e-01 -4.64135915e-01
-2.61385813e-02 -3.09526265e-01 5.74072063e-01 -3.02552819e-01
7.44493365e-01 -9.09376025e-01 2.54435599e-01 -1.34852961e-01
-2.83754498e-01 -2.84684181e-01 -1.65692605e-02 5.64308345e-01
-1.59915179e-01 -1.24304257e-01 7.87311435e-01 8.60609785e-02
-2.85100698e-01 6.10683084e-01 -6.14919215e-02 -1.60655051e-01
7.44439363e-01 -2.73759991e-01 2.78731555e-01 -5.44911504e-01
-8.72312009e-01 -5.30772982e-03 6.69009268e-01 1.69485375e-01
6.80225074e-01 -1.38764167e+00 -6.82227015e-01 9.70901132e-01
1.55669793e-01 2.76542127e-01 6.22215688e-01 8.84611428e-01
-1.73055500e-01 -2.99926195e-02 -2.82666415e-01 -8.03335786e-01
-1.44745207e+00 4.42497700e-01 5.17127872e-01 3.41507435e-01
-4.20862496e-01 5.40247262e-01 5.94291985e-01 -5.83283901e-01
2.06982508e-01 -8.54522660e-02 -1.28907025e-01 1.95690975e-01
4.99965072e-01 4.70304266e-02 2.53444770e-03 -1.15004909e+00
-2.02428594e-01 1.17833090e+00 1.60503373e-01 -1.92220017e-01
1.49483311e+00 -4.36765343e-01 -1.52812943e-01 3.47997248e-01
1.59508669e+00 2.53749758e-01 -1.70348907e+00 -5.65782964e-01
-4.32101220e-01 -7.36727893e-01 7.38689154e-02 -3.85393828e-01
-1.56447375e+00 8.92011762e-01 2.44377643e-01 -5.22907376e-01
1.30382812e+00 1.63547210e-02 3.95470619e-01 4.94074030e-03
1.85519263e-01 -8.96197557e-01 3.28917503e-01 1.30906790e-01
9.04821873e-01 -1.18170810e+00 2.16183737e-01 -7.15352595e-01
-4.86783564e-01 1.40987217e+00 3.17372113e-01 5.44787459e-02
9.34756577e-01 3.90852764e-02 9.11565125e-02 -1.18164800e-01
-4.97753918e-01 1.66533911e-03 6.33311033e-01 6.50609434e-01
1.94596782e-01 -1.63150445e-01 3.76896441e-01 4.06482846e-01
6.78089410e-02 -1.01753183e-01 1.94318444e-01 3.76087248e-01
1.07521959e-01 -1.06878757e+00 -2.38841504e-01 -2.12916836e-01
-6.44441918e-02 3.37016463e-01 -9.69355851e-02 7.04384029e-01
1.66176274e-01 6.79133654e-01 9.53882486e-02 -2.68051893e-01
2.76674360e-01 2.88111288e-02 1.04035568e+00 -6.05432391e-01
1.57522410e-01 5.63150406e-01 -1.17841944e-01 -6.03723168e-01
-9.14866567e-01 -9.17872369e-01 -1.03614652e+00 -7.96562135e-02
-4.58134770e-01 4.58247922e-02 8.03717852e-01 1.04335177e+00
2.78235257e-01 -1.32538006e-01 1.14507878e+00 -1.15254045e+00
-3.18152189e-01 -6.81552231e-01 -7.88933218e-01 3.95659178e-01
6.66767597e-01 -9.89443898e-01 -7.56205201e-01 3.92502189e-01] | [13.06727409362793, 0.032123927026987076] |
6734c61d-feee-4cfb-96f9-489c2f206fee | template-nerf-towards-modeling-dense-shape | 2111.04237 | null | https://arxiv.org/abs/2111.04237v1 | https://arxiv.org/pdf/2111.04237v1.pdf | Template NeRF: Towards Modeling Dense Shape Correspondences from Category-Specific Object Images | We present neural radiance fields (NeRF) with templates, dubbed Template-NeRF, for modeling appearance and geometry and generating dense shape correspondences simultaneously among objects of the same category from only multi-view posed images, without the need of either 3D supervision or ground-truth correspondence knowledge. The learned dense correspondences can be readily used for various image-based tasks such as keypoint detection, part segmentation, and texture transfer that previously require specific model designs. Our method can also accommodate annotation transfer in a one or few-shot manner, given only one or a few instances of the category. Using periodic activation and feature-wise linear modulation (FiLM) conditioning, we introduce deep implicit templates on 3D data into the 3D-aware image synthesis pipeline NeRF. By representing object instances within the same category as shape and appearance variation of a shared NeRF template, our proposed method can achieve dense shape correspondences reasoning on images for a wide range of object classes. We demonstrate the results and applications on both synthetic and real-world data with competitive results compared with other methods based on 3D information. | ['Qingfu Zhang', 'Xi Lin', 'Zhiyuan Yang', 'Jianfei Guo'] | 2021-11-08 | null | null | null | null | ['3d-aware-image-synthesis'] | ['computer-vision'] | [ 5.82682610e-01 1.71903983e-01 1.78423002e-01 -6.59482598e-01
-9.02093470e-01 -7.44476974e-01 7.21155822e-01 -2.71895349e-01
9.70017985e-02 2.52026320e-01 -1.16098881e-01 1.51698872e-01
-7.94569626e-02 -7.98865080e-01 -1.24487174e+00 -6.34993434e-01
4.72575754e-01 5.28484821e-01 1.79641306e-01 -1.55522704e-01
8.65446627e-02 1.02497900e+00 -1.69669747e+00 3.12212795e-01
5.91435373e-01 1.30182040e+00 4.14304346e-01 3.37170035e-01
-1.18682727e-01 3.78642410e-01 -1.40154704e-01 -2.07318053e-01
8.14029038e-01 -2.96812624e-01 -4.65312302e-01 6.85397029e-01
1.05314481e+00 -1.98248684e-01 -2.16253668e-01 8.88086915e-01
4.02504206e-01 2.58649230e-01 8.71495485e-01 -1.04046631e+00
-6.72390759e-01 -1.15215234e-01 -7.32840359e-01 -3.24642956e-01
2.12866858e-01 2.21714213e-01 7.10586607e-01 -1.22788858e+00
6.17772937e-01 1.26730275e+00 5.77712834e-01 5.48480332e-01
-1.39991248e+00 -5.38196087e-01 2.78191149e-01 -5.67669645e-02
-1.40724790e+00 -5.43350697e-01 9.60453451e-01 -4.36345458e-01
9.05964553e-01 2.78805673e-01 5.11273742e-01 8.04930687e-01
-3.09663098e-02 5.64103484e-01 1.12927353e+00 -3.47488314e-01
1.97730869e-01 1.86381519e-01 -2.87608027e-01 7.92719245e-01
-2.00806335e-01 1.64394543e-01 -3.93007725e-01 -1.30055308e-01
1.24024713e+00 1.20955156e-02 -3.90374362e-01 -6.36204302e-01
-1.48233593e+00 6.44840181e-01 7.21374214e-01 -1.40168309e-01
-4.50300604e-01 3.47129107e-01 -2.31987730e-01 3.20264313e-04
6.14728987e-01 4.32647198e-01 -6.36060476e-01 6.43301845e-01
-6.13992929e-01 1.81517079e-01 2.88891405e-01 1.31254792e+00
1.16954494e+00 1.52212903e-01 -2.17325389e-01 8.58227253e-01
2.65298069e-01 9.40247238e-01 8.66657123e-02 -9.97179866e-01
7.18059987e-02 6.07177556e-01 2.57904708e-01 -8.30773473e-01
-3.08456749e-01 -3.37779284e-01 -8.62498045e-01 2.69920558e-01
1.26676783e-01 1.71500921e-01 -1.45189524e+00 1.83902061e+00
7.55542159e-01 3.28364223e-01 -9.40781161e-02 1.15733683e+00
8.99084330e-01 7.29884267e-01 -3.82444590e-01 6.10866547e-02
1.31555176e+00 -8.64122450e-01 -2.45519191e-01 -3.56447786e-01
2.06489578e-01 -9.25579965e-01 9.88592863e-01 9.70983654e-02
-1.06947041e+00 -5.43983757e-01 -7.00430989e-01 -1.90812185e-01
-1.99753284e-01 1.10475779e-01 7.38919795e-01 4.04365778e-01
-8.52753639e-01 3.01033556e-01 -5.78431010e-01 -2.09960729e-01
6.32486045e-01 3.98327619e-01 -4.82179999e-01 -3.14310700e-01
-7.19892561e-01 6.84797764e-01 -3.86294723e-02 2.54999548e-01
-1.11965287e+00 -1.00666749e+00 -1.05333221e+00 -1.44215986e-01
2.88886726e-01 -1.13476586e+00 9.85862315e-01 -1.05149901e+00
-1.49134874e+00 1.18904507e+00 -1.13581091e-01 1.83730870e-02
1.92889839e-01 -6.97585270e-02 -4.77128811e-02 -3.63363116e-03
2.65781116e-02 1.02443504e+00 1.12480092e+00 -1.39412713e+00
-3.33113164e-01 -4.84506935e-01 1.53613687e-01 4.32397038e-01
1.51951760e-01 -1.75830960e-01 -3.84798318e-01 -5.94925106e-01
4.51577455e-01 -7.70162046e-01 -3.04369152e-01 5.95723391e-01
-4.08397555e-01 6.33338764e-02 7.62437880e-01 -2.35605568e-01
1.58834904e-01 -2.00912619e+00 2.52882153e-01 2.40684956e-01
-1.19127661e-01 -1.47982061e-01 -4.79823291e-01 -5.84763736e-02
-9.37566608e-02 -1.89088032e-01 -4.90911037e-01 -1.52606070e-01
-1.41047249e-02 1.45598399e-02 -3.27296823e-01 7.08551168e-01
5.89530647e-01 1.08255506e+00 -7.09521651e-01 -1.55146509e-01
4.44726437e-01 5.15833676e-01 -5.89165807e-01 4.45638269e-01
-6.92482769e-01 6.85071051e-01 -4.73890126e-01 6.01598918e-01
9.11436081e-01 -3.91949534e-01 -2.58967370e-01 -8.26328993e-01
1.13495691e-02 2.02681664e-02 -1.30179012e+00 2.00270557e+00
-8.96541834e-01 3.17241251e-01 1.24891937e-01 -8.97726953e-01
1.11792660e+00 6.63070828e-02 4.50544417e-01 -6.56186759e-01
1.81675814e-02 1.36553988e-01 -1.93393230e-01 -2.60567784e-01
2.85743654e-01 -1.86685845e-01 -1.08307168e-01 5.24942696e-01
2.39053354e-01 -7.20461428e-01 -4.86837775e-01 -1.32102862e-01
6.76801324e-01 3.62554789e-01 1.31772280e-01 -2.06466407e-01
4.29930001e-01 -1.05132177e-01 5.01697540e-01 6.09403610e-01
4.51394320e-01 1.09869683e+00 -2.36080080e-01 -3.58087420e-01
-1.28660297e+00 -1.13464642e+00 -3.43928933e-01 6.79388642e-01
5.03129482e-01 1.88343272e-01 -4.09000576e-01 -4.72036183e-01
1.14554048e-01 4.94684756e-01 -5.64356863e-01 2.58614570e-02
-4.82399881e-01 -5.07835627e-01 1.09160051e-01 2.88905323e-01
5.40428102e-01 -9.22230661e-01 -6.73594058e-01 -4.77019995e-02
5.69578707e-02 -1.33867335e+00 -8.17808628e-01 1.41955450e-01
-6.38177872e-01 -9.69118834e-01 -8.96853685e-01 -7.94420660e-01
1.01440978e+00 6.56648219e-01 1.08863199e+00 -6.44294871e-03
-6.31809890e-01 7.12289989e-01 -4.50749919e-02 -1.96293026e-01
-3.40760648e-01 -1.90126047e-01 -1.64674535e-01 4.45357114e-01
-1.18976884e-01 -6.31628454e-01 -6.68411314e-01 6.42811060e-01
-8.69689107e-01 6.10745132e-01 5.31900942e-01 9.19370234e-01
9.52650905e-01 -3.51690441e-01 4.01036382e-01 -7.53193557e-01
-1.71005547e-01 -2.70867974e-01 -9.49094713e-01 3.17019343e-01
-2.78066009e-01 1.15262836e-01 3.91893566e-01 -5.11402190e-01
-1.24511337e+00 6.04265809e-01 1.04787327e-01 -8.99495363e-01
-5.56472957e-01 -9.39505994e-02 -5.09062409e-01 -2.87109554e-01
7.67339170e-01 2.65457392e-01 -7.07943290e-02 -4.67397362e-01
6.57575965e-01 4.29314792e-01 5.70956290e-01 -6.20093763e-01
9.51252043e-01 6.67835772e-01 2.66430259e-01 -8.09514046e-01
-1.09434688e+00 -4.82940018e-01 -8.14740896e-01 -2.03089789e-01
1.05633521e+00 -1.20470095e+00 -3.84702325e-01 6.59543455e-01
-1.39578366e+00 -4.45984721e-01 -3.19895178e-01 2.07024097e-01
-8.13380003e-01 -1.02306588e-03 -1.60478294e-01 -6.21395051e-01
-2.54255414e-01 -1.14147651e+00 1.76357031e+00 2.66353935e-01
3.00703317e-01 -8.49345446e-01 -3.12585324e-01 3.68817866e-01
2.15198547e-01 4.94564950e-01 1.03815985e+00 -1.22553311e-01
-1.22578442e+00 9.84153599e-02 -3.37983102e-01 3.17002505e-01
4.36653107e-01 -3.36076200e-01 -1.20450389e+00 -1.30842224e-01
-7.39665627e-02 -4.36769396e-01 5.64161479e-01 3.75346512e-01
1.36982739e+00 -1.64285630e-01 -3.17323118e-01 8.96588802e-01
1.47969186e+00 -1.66887879e-01 4.69045043e-01 -1.43437460e-01
1.15404427e+00 7.74463713e-01 6.22367442e-01 2.34739617e-01
2.37292200e-01 1.19331467e+00 6.37375474e-01 -3.99371833e-01
-3.59468669e-01 -2.77506799e-01 1.56715646e-01 3.82367820e-01
1.80490986e-01 -9.65398625e-02 -6.21183753e-01 5.43651760e-01
-1.64551735e+00 -7.00771391e-01 -3.05920001e-03 2.20346165e+00
7.09592879e-01 -2.35594988e-01 -1.99195534e-01 -4.29829210e-01
7.92769015e-01 -5.94999082e-02 -9.64818954e-01 8.34681615e-02
-1.58688873e-01 3.79150629e-01 5.84176898e-01 5.32289743e-01
-9.45530117e-01 9.88743901e-01 5.46409893e+00 7.25620568e-01
-1.28132248e+00 -1.33279916e-02 7.61093736e-01 1.00232534e-01
-8.69246423e-01 2.38054004e-02 -7.36574590e-01 -1.39763936e-01
3.55194151e-01 7.14899153e-02 5.68058312e-01 6.69709325e-01
1.07010245e-01 8.16335157e-03 -1.40214694e+00 1.21349442e+00
2.18858689e-01 -1.61751235e+00 1.85222358e-01 4.45897169e-02
1.04271281e+00 -4.55397591e-02 6.14986829e-02 -2.25329384e-01
4.14081998e-02 -1.12190270e+00 9.54683065e-01 6.97997510e-01
1.11746514e+00 -5.08739710e-01 2.15375587e-01 3.67762536e-01
-1.13284314e+00 2.51112938e-01 -4.37773436e-01 4.95882601e-01
1.08105563e-01 4.97988671e-01 -8.17499459e-01 7.54592419e-01
5.73558748e-01 7.54696667e-01 -2.34137252e-01 8.78105283e-01
-8.91983956e-02 7.93486312e-02 -5.55083394e-01 2.78414935e-01
4.74645160e-02 -2.50231475e-01 5.43999076e-01 6.92103863e-01
4.40645874e-01 3.22140932e-01 1.25937805e-01 1.34044564e+00
-1.40233889e-01 -7.61201605e-02 -6.97375536e-01 3.49896371e-01
4.56285089e-01 1.48123372e+00 -7.11486340e-01 1.12523846e-02
-4.00167555e-01 1.22076440e+00 8.36400092e-02 2.83610135e-01
-7.62627542e-01 -1.26840755e-01 6.78009510e-01 3.88079047e-01
5.74308336e-01 -2.13919014e-01 -2.40166947e-01 -1.10094833e+00
-4.04427908e-02 -6.62136853e-01 -4.98368070e-02 -1.29036295e+00
-1.51238477e+00 5.86115241e-01 2.65535951e-01 -1.36019647e+00
-1.39772341e-01 -6.86537683e-01 -4.45439905e-01 1.02192616e+00
-1.52349985e+00 -1.48727953e+00 -6.18260264e-01 6.99813306e-01
5.48176527e-01 -1.17189392e-01 9.54642832e-01 1.26385927e-01
-9.28768963e-02 5.07601500e-01 -2.58355528e-01 -1.74310669e-01
6.19611502e-01 -1.00829911e+00 7.16411114e-01 4.67470050e-01
3.88830483e-01 1.89907208e-01 3.36057574e-01 -3.72238219e-01
-1.64958727e+00 -1.58630729e+00 1.86802372e-01 -5.34447074e-01
8.69050249e-02 -8.34124148e-01 -7.97837734e-01 4.35951531e-01
-5.34912571e-02 3.12584907e-01 3.48139882e-01 -3.79187346e-01
-4.11340743e-01 -8.89851674e-02 -1.40662432e+00 3.63113552e-01
1.40351439e+00 -5.33368051e-01 -2.56261140e-01 5.88560760e-01
8.41091275e-01 -8.42801511e-01 -9.14755344e-01 5.64308345e-01
3.11015368e-01 -7.71938562e-01 1.15098286e+00 -4.85129207e-01
3.36112201e-01 -5.97224474e-01 -4.80866909e-01 -1.33098376e+00
-2.71698445e-01 -5.61288118e-01 2.50342786e-01 9.62854087e-01
3.53663951e-01 -3.22535366e-01 6.34100139e-01 5.04764795e-01
-4.51672941e-01 -5.55321395e-01 -8.24198365e-01 -5.52551329e-01
-4.89813723e-02 -4.19351846e-01 8.94120455e-01 9.37477946e-01
-9.70370233e-01 3.76271755e-01 -2.57156849e-01 5.78440487e-01
8.09363723e-01 7.32705057e-01 1.05611801e+00 -1.09107792e+00
-4.53538775e-01 -6.12013265e-02 -4.95258391e-01 -1.40219426e+00
3.49171668e-01 -1.21769428e+00 3.32048506e-01 -1.42603576e+00
1.00596108e-01 -7.46215701e-01 1.62122637e-01 6.04425430e-01
1.74561635e-01 5.22724152e-01 7.68919811e-02 9.56299454e-02
-1.51587725e-01 9.01675105e-01 1.59999859e+00 -2.17196658e-01
3.86534035e-02 -1.39673948e-01 -6.68461323e-01 5.76443374e-01
3.70415211e-01 -2.71001041e-01 -5.16396403e-01 -6.61370039e-01
8.32278877e-02 5.01269065e-02 7.24499226e-01 -6.65248334e-01
2.53133308e-02 -4.51107144e-01 5.88728666e-01 -6.02761149e-01
6.33276224e-01 -1.08624184e+00 5.32935441e-01 -2.09563360e-01
-2.37568736e-01 -3.48314762e-01 3.50999027e-01 6.20629609e-01
1.29658058e-01 3.25105749e-02 1.09451044e+00 -2.34909862e-01
-7.61119366e-01 7.00404048e-01 2.40332395e-01 -1.11552671e-01
1.12563396e+00 -3.02917629e-01 -1.39724165e-01 -3.33083928e-01
-5.89646161e-01 9.91089717e-02 7.29013622e-01 5.02789736e-01
8.49895239e-01 -1.68588090e+00 -7.70499468e-01 6.05866373e-01
4.78571832e-01 7.34002769e-01 4.82008666e-01 4.04317886e-01
-2.24069774e-01 1.96296662e-01 -2.68468589e-01 -1.16518760e+00
-1.00326335e+00 3.05652499e-01 9.19869602e-01 3.74670953e-01
-8.25034499e-01 7.92202055e-01 1.00395274e+00 -9.89058077e-01
-1.66908707e-02 -2.30435461e-01 2.85118133e-01 -4.93614197e-01
2.95620352e-01 -2.80464143e-01 2.16783598e-01 -7.65338480e-01
-3.19928259e-01 1.09618473e+00 8.00015032e-02 -7.04953894e-02
1.45252216e+00 -4.34478670e-02 5.26553057e-02 3.66183281e-01
1.12910724e+00 -3.20031881e-01 -1.66013944e+00 -3.82356852e-01
-5.11595607e-01 -8.01562726e-01 2.54398763e-01 -7.46045649e-01
-1.24285686e+00 7.63395488e-01 8.09444368e-01 -4.59324837e-01
1.08459830e+00 3.66115391e-01 5.58569372e-01 2.65800446e-01
7.27044463e-01 -6.78467751e-01 1.25406861e-01 2.88452566e-01
1.19164574e+00 -1.14938068e+00 -8.95810425e-02 -6.90417647e-01
-3.53756607e-01 9.90417361e-01 7.22884119e-01 -1.85649380e-01
7.56405413e-01 8.73527676e-02 -9.80297551e-02 -3.15847427e-01
-7.27072120e-01 -1.18510067e-01 6.75070703e-01 7.78872490e-01
2.66283363e-01 -5.40178493e-02 4.63859111e-01 3.20113599e-01
1.08768366e-01 -3.27137083e-01 4.59240079e-01 6.13299608e-01
-2.22818971e-01 -9.62559402e-01 -3.90836686e-01 4.17683154e-01
4.89319116e-02 -8.60846490e-02 -2.29482055e-01 5.81723750e-01
1.10663220e-01 3.96970123e-01 4.29065406e-01 -2.32600361e-01
5.60886562e-01 -1.02458283e-01 1.08890176e+00 -9.92178857e-01
-3.26707989e-01 3.06465328e-01 -6.80910498e-02 -5.44694245e-01
-6.70609474e-01 -6.71581984e-01 -1.16796410e+00 8.67354795e-02
-3.96898806e-01 -4.01754886e-01 6.71777308e-01 7.62347162e-01
5.31531394e-01 3.44462216e-01 1.06123114e+00 -1.34060133e+00
-4.01018918e-01 -6.98405504e-01 -4.94933575e-01 5.56189299e-01
3.06081086e-01 -7.77321160e-01 -2.77357221e-01 2.06581265e-01] | [8.610577583312988, -3.146728754043579] |
8cfca4b8-8260-4531-9c0a-b46eeb23f95c | musical-instrument-recognition-using-their | 1705.04971 | null | http://arxiv.org/abs/1705.04971v1 | http://arxiv.org/pdf/1705.04971v1.pdf | Musical Instrument Recognition Using Their Distinctive Characteristics in Artificial Neural Networks | In this study an Artificial Neural Network was trained to classify musical
instruments, using audio samples transformed to the frequency domain. Different
features of the sound, in both time and frequency domain, were analyzed and
compared in relation to how much information that could be derived from that
limited data. The study concluded that in comparison with the base experiment,
that had an accuracy of 93.5%, using the attack only resulted in 80.2% and the
initial 100 Hz in 64.2%. | ['Babak Toghiani-Rizi', 'Marcus Windmark'] | 2017-05-14 | null | null | null | null | ['instrument-recognition'] | ['audio'] | [ 2.03396022e-01 9.10448208e-02 2.86413699e-01 2.11258322e-01
-1.47629753e-01 -7.07093835e-01 5.62796295e-02 9.49746892e-02
-4.83725727e-01 8.17659140e-01 -4.41927388e-02 -5.80329895e-02
-4.93215382e-01 -9.37818408e-01 -7.31480122e-02 -5.37577868e-01
-3.47984731e-01 2.73328245e-04 -3.98643985e-02 -2.78060704e-01
2.99885482e-01 5.25941312e-01 -1.98670745e+00 3.45557153e-01
2.27189019e-01 1.03188336e+00 -2.49774143e-01 8.34382415e-01
8.25057775e-02 5.69730520e-01 -1.52280354e+00 1.31661296e-01
5.11136234e-01 -4.07448202e-01 -6.16370440e-01 -5.13633609e-01
4.31101061e-02 -3.15378696e-01 -1.45252094e-01 9.87529933e-01
7.31101215e-01 3.50367486e-01 7.08725572e-01 -1.00243914e+00
-1.40897736e-01 1.08447814e+00 -1.61650013e-02 1.46093413e-01
7.22727776e-01 -2.86815614e-01 5.72078526e-01 -6.24302030e-01
4.90400672e-01 7.53613234e-01 8.72086406e-01 2.10631669e-01
-1.19922400e+00 -1.10981929e+00 -7.46452093e-01 4.34343554e-02
-1.33389854e+00 -2.57557184e-02 1.21831191e+00 -5.83145022e-01
9.73631144e-01 3.53141874e-01 1.00868952e+00 7.04668283e-01
9.26960260e-02 5.11431210e-02 1.31221080e+00 -1.17983270e+00
7.70897344e-02 2.00201184e-01 3.75083685e-01 1.02820329e-01
1.09332688e-01 4.91580278e-01 -3.04939121e-01 -2.42890000e-01
7.80119896e-01 -5.78531802e-01 -3.13178539e-01 5.89461684e-01
-8.40068400e-01 8.68115485e-01 1.00187160e-01 1.03873849e+00
-6.18321419e-01 -1.69973373e-01 6.90629363e-01 6.79765642e-01
8.21211562e-02 1.10630596e+00 -6.00681007e-01 -5.76993346e-01
-1.01734960e+00 2.37264514e-01 1.12356412e+00 6.61421567e-02
2.38329783e-01 8.45500469e-01 5.26294827e-01 7.36682594e-01
-1.92651078e-01 3.05598319e-01 8.64767969e-01 -9.92105782e-01
-3.10522951e-02 3.18526000e-01 -7.12698996e-02 -1.10916817e+00
-6.57140076e-01 -5.14128268e-01 -7.53502727e-01 6.27317786e-01
7.38170445e-01 -6.26171470e-01 -6.16437614e-01 1.56599569e+00
4.29420769e-02 -7.52302408e-02 3.24140728e-01 6.82123065e-01
7.81813085e-01 7.57338703e-01 -4.45328832e-01 -2.68004954e-01
1.00225079e+00 -3.85237962e-01 -1.06302416e+00 4.02520061e-01
1.91152189e-02 -1.19786406e+00 8.73168349e-01 1.26026118e+00
-9.30919290e-01 -1.13690877e+00 -1.40714490e+00 5.28215706e-01
-4.18558419e-01 9.48620439e-02 3.59491050e-01 9.28187072e-01
-8.38284731e-01 1.13696325e+00 -3.07761818e-01 -5.24337590e-03
-4.36408430e-01 4.16099638e-01 -2.07850531e-01 1.06598234e+00
-1.54510438e+00 6.11227334e-01 9.40497756e-01 -2.07839847e-01
-2.31516361e-01 -7.16255486e-01 -2.33499154e-01 -1.56109408e-02
-2.09266976e-01 1.28798932e-01 1.34748054e+00 -9.18168783e-01
-1.89370358e+00 2.58409858e-01 4.04318571e-01 -5.21594644e-01
1.74905837e-01 -2.18227774e-01 -8.86210978e-01 5.80971003e-01
-2.91914701e-01 6.02417290e-02 6.70600355e-01 -8.49762917e-01
-7.17461169e-01 -1.62045106e-01 -9.22558457e-02 -3.99579763e-01
-2.01799810e-01 -6.16219826e-02 4.63476151e-01 -9.92093146e-01
2.32152596e-01 -8.07283044e-01 4.10961896e-01 -6.50629163e-01
-2.85996765e-01 -2.11177841e-02 6.70180321e-01 -1.12005293e+00
1.37579930e+00 -2.06382322e+00 5.93500808e-02 6.34443045e-01
-2.51810670e-01 3.62150043e-01 2.97637701e-01 8.00152123e-01
-6.03628576e-01 4.71815765e-02 1.16012052e-01 3.88789713e-01
-2.38162786e-01 -1.98138803e-01 -4.22996819e-01 -8.48994311e-03
-1.63767338e-01 -1.19257510e-01 -5.18556595e-01 -3.53796855e-02
4.17883277e-01 5.49448669e-01 -2.72550911e-01 1.21410362e-01
3.55964661e-01 2.55829841e-01 -1.68554321e-01 4.98538405e-01
4.61904615e-01 8.05118144e-01 9.77747794e-03 -4.12515283e-01
-2.00591698e-01 1.49989709e-01 -1.41193306e+00 1.16095388e+00
-5.95409870e-01 1.08448613e+00 -8.13724548e-02 -1.09993255e+00
1.40631902e+00 9.94510353e-01 6.12553298e-01 -5.28426945e-01
4.75076765e-01 4.84535575e-01 5.46241999e-01 -6.66269302e-01
3.75622600e-01 -2.74987161e-01 3.95841748e-02 4.34252918e-01
2.92326599e-01 -4.69266057e-01 2.05602124e-01 -5.43164313e-01
7.14129984e-01 2.35388637e-01 3.06120157e-01 -1.45586744e-01
4.22139615e-01 -6.77166954e-02 1.15509897e-01 6.03316128e-01
3.15664932e-02 2.63612539e-01 3.26758444e-01 -5.70968688e-01
-9.85193729e-01 -5.72118521e-01 -2.62684971e-01 5.95833838e-01
-5.27932763e-01 -2.57148236e-01 -9.48642313e-01 8.51415023e-02
-1.59102529e-01 7.79197335e-01 -5.00222206e-01 -3.60890329e-01
-5.19617021e-01 -2.19038025e-01 1.12118316e+00 1.47359326e-01
2.66145110e-01 -1.43903947e+00 -8.80139530e-01 4.56391454e-01
1.90975685e-02 -7.54266918e-01 2.74003595e-01 5.12876570e-01
-1.15741968e+00 -9.36371088e-01 -4.67671543e-01 -5.13235986e-01
-1.32423356e-01 -3.96151304e-01 8.08694601e-01 -1.10170223e-01
-4.01396781e-01 5.19486442e-02 -6.25541747e-01 -1.15154767e+00
-6.58813477e-01 -5.06120501e-03 2.42922083e-01 -3.29234660e-01
3.93044025e-01 -1.16606593e+00 -2.20894456e-01 -1.00299597e-01
-7.42443562e-01 -7.04732001e-01 3.82622153e-01 6.95487738e-01
-7.83591866e-02 5.74445546e-01 8.17516625e-01 -4.00739312e-01
1.03667951e+00 -6.11448400e-02 -3.10883045e-01 -4.33716655e-01
-5.53943336e-01 -2.23993942e-01 9.24268007e-01 -9.19999540e-01
-5.21300256e-01 -1.38173178e-01 -3.48630488e-01 -3.32175404e-01
-4.55251753e-01 4.86677021e-01 3.58363926e-01 -1.64407745e-01
9.35038030e-01 2.23565921e-02 3.15253854e-01 -7.12965310e-01
-2.48616695e-01 8.91155720e-01 6.10395849e-01 -5.95548749e-01
5.53207457e-01 -1.71388492e-01 4.34952118e-02 -1.15179729e+00
-4.40855712e-01 -2.27646112e-01 -7.03848958e-01 -5.27331829e-01
5.00517130e-01 -3.89638036e-01 -7.97990918e-01 6.26361787e-01
-1.09767342e+00 6.55263588e-02 -7.02323556e-01 1.30828428e+00
-4.93899941e-01 3.87580432e-02 -6.26199424e-01 -1.32959151e+00
-5.65215886e-01 -7.76597977e-01 3.07743013e-01 2.47401848e-01
-8.82883966e-01 -7.04537272e-01 2.43371516e-01 -9.48698521e-02
3.50138456e-01 4.51666832e-01 7.83167660e-01 -9.09702778e-01
5.42312622e-01 -7.35213280e-01 4.34886038e-01 7.40635276e-01
4.03794736e-01 3.35468322e-01 -1.46371782e+00 -6.81690276e-02
4.20593172e-01 -2.53609478e-01 3.23318630e-01 3.31314504e-01
8.31150413e-01 -2.49817446e-01 2.98052728e-01 2.67110407e-01
1.40101182e+00 1.14769673e+00 6.78276896e-01 5.81725419e-01
4.13862616e-02 4.32497114e-01 5.60024917e-01 4.09733593e-01
-7.20237434e-01 5.18754303e-01 2.11616710e-01 1.78632718e-02
-2.84762681e-02 -6.98012263e-02 1.46688744e-01 9.44459617e-01
-7.33924031e-01 5.67151383e-02 -6.30343616e-01 1.96034357e-01
-1.08937740e+00 -1.14311779e+00 3.18312906e-02 2.10954261e+00
8.70606542e-01 4.99813318e-01 5.17363369e-01 1.54035056e+00
7.50186026e-01 -1.33833900e-01 -6.53949156e-02 -8.84619653e-01
1.68699056e-01 9.19125736e-01 1.10391088e-01 4.21881855e-01
-9.54563320e-01 8.41841251e-02 7.34512043e+00 7.28387713e-01
-1.48941994e+00 -4.51137662e-01 -2.34344244e-01 -3.07874903e-02
4.13252175e-01 -3.57084185e-01 -2.44156614e-01 5.41069150e-01
1.36701751e+00 -2.59612113e-01 4.67540890e-01 6.85521424e-01
1.23615436e-01 6.94743097e-02 -5.35938084e-01 9.22448695e-01
-1.58834532e-01 -6.61529362e-01 -8.40322208e-03 -2.60250811e-02
3.13010603e-01 -6.55038714e-01 1.67107657e-02 3.51398736e-01
-4.75907475e-01 -9.34848845e-01 9.06624198e-01 8.37922215e-01
5.23950100e-01 -1.09361219e+00 1.16325867e+00 4.78951126e-01
-1.06481075e+00 -1.82268053e-01 -1.33265369e-02 -7.72570729e-01
-1.85359180e-01 5.26968956e-01 -1.12193918e+00 7.67256439e-01
7.41218567e-01 6.38168678e-02 -9.01448017e-04 9.86038208e-01
2.59650387e-02 1.19048631e+00 -6.75558329e-01 -2.38005713e-01
-8.21676925e-02 -3.97769928e-01 6.56946123e-01 1.06673229e+00
7.64358580e-01 -1.28811762e-01 -2.26043567e-01 5.57350338e-01
5.37721753e-01 3.40641111e-01 -4.75504667e-01 -1.83352917e-01
8.47199440e-01 9.49862838e-01 -6.14401937e-01 -1.64393842e-01
1.47870313e-02 4.12795216e-01 -4.21857148e-01 1.12504207e-01
-5.26853621e-01 -1.42007589e+00 1.66762009e-01 -5.32265753e-02
1.19086178e-02 1.24870114e-01 -1.41330704e-01 -4.48088974e-01
-6.01748899e-02 -9.37357724e-01 3.18667948e-01 -8.21865499e-01
-8.95489752e-01 8.75672340e-01 1.66615441e-01 -1.46630013e+00
-8.85661900e-01 -9.19165015e-01 -6.12505853e-01 1.25170171e+00
-5.23756504e-01 -6.76178277e-01 1.77924573e-01 1.91387460e-01
2.28321388e-01 -6.36213899e-01 1.39324164e+00 3.87143791e-01
-6.89475313e-02 5.06174743e-01 1.11009970e-01 1.92188025e-01
4.78311092e-01 -1.30655813e+00 -2.73439258e-01 2.77307063e-01
1.53701231e-01 5.60937345e-01 1.08386445e+00 -2.14117572e-01
-7.10751772e-01 -5.00994205e-01 6.66633368e-01 1.53686836e-01
5.69775224e-01 1.01021081e-01 -9.52677846e-01 1.54128119e-01
3.76662284e-01 -4.84227419e-01 7.94502437e-01 9.14319083e-02
-9.11346376e-02 -9.05412808e-02 -1.35346103e+00 1.39205351e-01
1.39070258e-01 -6.88812017e-01 -1.00883651e+00 -1.90448180e-01
6.43332481e-01 -2.50955641e-01 -1.41953599e+00 4.03577328e-01
9.96561825e-01 -1.20661592e+00 1.11006474e+00 -4.32546765e-01
2.84850508e-01 -1.32854924e-01 -1.40480921e-01 -1.33187461e+00
-2.07826704e-01 -4.58455414e-01 1.75623726e-02 1.16702569e+00
4.92670774e-01 -6.34121835e-01 6.11729860e-01 -2.48744730e-02
3.37177478e-02 -3.96871895e-01 -9.36815500e-01 -7.68515527e-01
1.38705730e-01 -6.67069197e-01 5.67335904e-01 8.32246125e-01
2.59873360e-01 2.08722934e-01 -2.84955114e-01 -1.15555733e-01
3.09545726e-01 -9.44365188e-02 4.61521685e-01 -1.74994481e+00
-3.07944804e-01 -4.11419690e-01 -9.02222157e-01 1.88872591e-01
5.50579876e-02 -4.77579355e-01 -3.28953713e-01 -9.58577156e-01
-4.61981505e-01 -6.10238612e-02 -5.74576080e-01 4.59194839e-01
2.98189998e-01 5.71056008e-01 3.26210707e-01 8.31320211e-02
9.13668215e-01 -1.18139461e-01 8.69455993e-01 -9.93403345e-02
-5.30963540e-01 4.79089558e-01 -1.81499422e-01 1.13461626e+00
1.19129312e+00 -3.42960626e-01 -3.93856406e-01 3.36796880e-01
3.74820121e-02 2.83910006e-01 1.29364744e-01 -1.56740320e+00
-1.56511813e-01 2.42236286e-01 5.68945527e-01 -7.33196676e-01
6.35690033e-01 -1.12426686e+00 6.43748283e-01 8.80152047e-01
-3.26708794e-01 -2.51336068e-01 6.51299059e-01 -9.25954506e-02
-6.28945231e-01 -7.48698056e-01 5.39696097e-01 7.64021501e-02
-2.90278137e-01 -7.31441855e-01 -5.23601890e-01 -4.53489393e-01
7.21914291e-01 -6.62478387e-01 2.54217654e-01 -6.12545013e-01
-1.24729443e+00 -9.08740580e-01 -6.14550486e-02 6.42380416e-02
2.04367742e-01 -1.36649644e+00 -4.85992342e-01 1.96951851e-01
-3.63880515e-01 -8.16310048e-01 4.53864127e-01 3.34261924e-01
-6.53703272e-01 5.56755066e-01 -8.43832433e-01 -1.23114884e-01
-1.46449792e+00 3.76729548e-01 4.62991744e-01 8.61272365e-02
-5.61958015e-01 4.51990217e-01 -7.67199457e-01 -2.88367063e-01
3.69278580e-01 -4.10074711e-01 -8.30948174e-01 4.40033197e-01
5.69101810e-01 6.06076539e-01 3.22208017e-01 -4.66662735e-01
-4.84719798e-02 9.66272831e-01 6.06450319e-01 -7.45963216e-01
1.23203969e+00 5.19315064e-01 -6.48859795e-03 1.09350562e+00
1.06598198e+00 5.58276296e-01 -3.46113920e-01 2.37912536e-01
-1.91664979e-01 -4.01593268e-01 9.58320126e-02 -9.75113988e-01
-9.24882591e-01 8.00135434e-01 9.41473007e-01 9.98902619e-01
1.46879828e+00 -7.18016207e-01 3.76484305e-01 5.67303002e-01
3.72679174e-01 -1.17777443e+00 -3.74459535e-01 4.02422756e-01
1.04290032e+00 -3.33378196e-01 5.34953699e-02 -1.86090484e-01
-8.15305635e-02 1.85237706e+00 1.71471015e-01 -5.12797773e-01
8.55047464e-01 5.44838130e-01 2.34080762e-01 1.64653331e-01
-3.72045428e-01 3.05435658e-01 2.61353850e-01 8.08183789e-01
8.42135787e-01 -1.12531915e-01 -5.20671666e-01 1.13736188e+00
-1.07140529e+00 4.73173499e-01 6.95521414e-01 9.44001317e-01
-5.64117193e-01 -1.01998794e+00 -1.08294392e+00 4.60295022e-01
-9.73551810e-01 2.62455732e-01 -4.76133168e-01 1.34181595e+00
5.41118205e-01 1.10924959e+00 2.32665822e-01 -9.33637321e-01
6.65474892e-01 3.64997506e-01 5.60036600e-01 -1.85389102e-01
-9.16369796e-01 3.06790471e-01 2.84027249e-01 -7.71707669e-02
-3.94140840e-01 -4.14083958e-01 -1.21750665e+00 -6.38510473e-03
-5.14988899e-01 7.63944447e-01 1.02548444e+00 4.63125974e-01
-2.06560746e-01 1.07494211e+00 6.42798424e-01 -1.00906944e+00
-6.26294613e-01 -1.55535698e+00 -1.04394674e+00 1.73718616e-01
3.78357291e-01 -5.27932227e-01 -8.09122562e-01 1.75426021e-01] | [15.7870512008667, 5.28786039352417] |
3c2e931b-61b2-4be1-91d8-71367ec777cc | real-time-speech-enhancement-with-dynamic | 2302.10377 | null | https://arxiv.org/abs/2302.10377v1 | https://arxiv.org/pdf/2302.10377v1.pdf | Real-time speech enhancement with dynamic attention span | For real-time speech enhancement (SE) including noise suppression, dereverberation and acoustic echo cancellation, the time-variance of the audio signals becomes a severe challenge. The causality and memory usage limit that only the historical information can be used for the system to capture the time-variant characteristics. We propose to adaptively change the receptive field according to the input signal in deep neural network based SE model. Specifically, in an encoder-decoder framework, a dynamic attention span mechanism is introduced to all the attention modules for controlling the size of historical content used for processing the current frame. Experimental results verify that this dynamic mechanism can better track time-variant factors and capture speech-related characteristics, benefiting to both interference removing and speech quality retaining. | ['Yan Lu', 'Yuan Zhang', 'Xiulian Peng', 'Yuan Zhou', 'Chengyu Zheng'] | 2023-02-21 | null | null | null | null | ['acoustic-echo-cancellation', 'speech-enhancement', 'acoustic-echo-cancellation'] | ['medical', 'speech', 'speech'] | [ 1.00908950e-01 -4.08505172e-01 2.23582566e-01 -1.46549001e-01
-3.91893893e-01 -1.03224352e-01 1.51086867e-01 -1.85362265e-01
-5.23390114e-01 3.57919633e-01 4.41759706e-01 -1.08282574e-01
-2.84429103e-01 -3.85282755e-01 -3.29873711e-01 -7.51488864e-01
-4.71265391e-02 -6.78223193e-01 2.95852631e-01 -2.53969669e-01
5.13060205e-02 4.33507591e-01 -1.55980349e+00 1.25295743e-01
9.29314494e-01 1.19901562e+00 8.98787558e-01 9.14116025e-01
1.89017102e-01 5.95925212e-01 -8.30623329e-01 -2.98375059e-02
8.37244466e-02 -4.59891349e-01 8.70316848e-02 3.12621258e-02
6.11125939e-02 -4.62429643e-01 -8.82942080e-01 1.27387774e+00
1.03050053e+00 5.63628256e-01 2.18169048e-01 -6.99815750e-01
-5.23027956e-01 6.18807495e-01 -2.73150861e-01 8.55356872e-01
1.10359285e-02 8.42493325e-02 4.95043039e-01 -9.93405938e-01
1.46881253e-01 1.18993223e+00 5.80088079e-01 4.40757334e-01
-6.25669420e-01 -7.06011415e-01 4.12456006e-01 7.75169015e-01
-1.12882602e+00 -1.00565875e+00 9.84494448e-01 7.41291195e-02
1.03283286e+00 3.11404228e-01 7.66544163e-01 8.01525712e-01
2.76337892e-01 7.37877309e-01 5.63180685e-01 -4.74476427e-01
5.91389388e-02 -2.66666800e-01 -6.28393143e-02 1.13711273e-02
-4.57283974e-01 4.00720626e-01 -7.63813555e-01 3.32323909e-01
6.83342397e-01 -2.11836830e-01 -7.34867573e-01 4.07971174e-01
-7.99144387e-01 2.26636156e-01 2.37371475e-01 5.05145192e-01
-6.33587182e-01 1.69816226e-01 6.83045328e-01 3.34390789e-01
3.74485970e-01 2.38822982e-01 -6.67934537e-01 -4.05914217e-01
-9.42752063e-01 -1.45294428e-01 1.39602527e-01 7.34001100e-01
7.45728686e-02 8.37280750e-01 -4.46440935e-01 1.01818120e+00
1.67619392e-01 6.39151633e-01 7.99311042e-01 -8.94559741e-01
3.28807920e-01 -2.92146295e-01 7.11342469e-02 -1.00616360e+00
-2.73497671e-01 -9.20243859e-01 -1.00373721e+00 -4.79254872e-03
4.75078411e-02 -3.49260598e-01 -9.03365374e-01 1.90265989e+00
3.01246047e-01 5.27490973e-01 -1.24198824e-01 9.42770362e-01
4.59233642e-01 8.86585772e-01 8.50511938e-02 -8.21975052e-01
1.36872721e+00 -8.19784939e-01 -1.49006450e+00 -3.47033888e-01
-2.13334307e-01 -8.08127165e-01 7.12205887e-01 4.27852839e-01
-1.35598874e+00 -1.00550652e+00 -1.28280795e+00 3.85981165e-02
2.17159428e-02 1.43393278e-01 1.07063852e-01 5.27105629e-01
-1.02099931e+00 6.32703424e-01 -8.34539592e-01 3.74928921e-01
8.40252563e-02 2.40240753e-01 1.01611033e-01 2.54360020e-01
-1.60440326e+00 6.23982012e-01 1.87760904e-01 5.79132617e-01
-7.32980371e-01 -7.72746742e-01 -6.96070015e-01 5.30728817e-01
2.11972460e-01 -3.95431906e-01 1.25596821e+00 -1.13022757e+00
-1.70628166e+00 1.24853119e-01 -3.86020988e-01 -4.62652683e-01
2.58576900e-01 -3.97664070e-01 -1.04712212e+00 1.98309928e-01
-4.48183894e-01 1.97900236e-01 1.47891712e+00 -6.55211210e-01
-7.57454813e-01 -1.48306072e-01 -2.99172193e-01 4.42501336e-01
-7.03581154e-01 1.12162299e-01 -7.49481261e-01 -1.17345667e+00
1.97265387e-01 -4.78710949e-01 -9.55170244e-02 -1.45092785e-01
7.31114894e-02 3.00207794e-01 8.67606401e-01 -1.26959705e+00
1.77510655e+00 -2.69370198e+00 1.22619644e-01 -2.07205936e-01
2.77138129e-02 8.40777397e-01 -2.49565780e-01 -7.60033901e-04
-3.21228616e-02 -2.64730245e-01 1.29009470e-01 -1.99474663e-01
-1.68047637e-01 -2.05955371e-01 -3.18599850e-01 3.63161057e-01
1.16946548e-01 5.57242632e-01 -8.01929832e-01 -2.61789292e-01
3.69185060e-01 8.85347366e-01 -4.72928375e-01 3.69091600e-01
1.86647505e-01 4.51605320e-01 -9.42154601e-02 6.73728287e-02
8.34658325e-01 3.98133487e-01 8.09079036e-02 -4.68851179e-01
-1.57495722e-01 3.84820670e-01 -1.17471743e+00 1.46552324e+00
-7.74798036e-01 9.50004756e-01 7.10113525e-01 -4.79139447e-01
8.24831367e-01 5.99027514e-01 1.88286975e-01 -1.25480926e+00
2.41726577e-01 5.07857315e-02 3.12241316e-01 -4.65162814e-01
4.47455436e-01 1.69220995e-02 4.48312998e-01 1.90650858e-03
-2.69351769e-02 3.32811564e-01 -8.66355002e-02 -3.26521069e-01
8.65190625e-01 -2.90922999e-01 1.41383335e-01 6.74608070e-03
7.38219321e-01 -8.47924948e-01 9.44613218e-01 4.92522895e-01
-4.93851423e-01 3.59394014e-01 -4.22983505e-02 5.76925837e-03
-1.11183906e+00 -8.56327116e-01 -7.14424849e-02 1.47691560e+00
8.79270118e-03 -1.87173709e-01 -8.78194392e-01 8.05491954e-03
-4.84979153e-01 8.10242176e-01 -1.71314552e-01 -5.85904241e-01
-9.42397773e-01 -2.77372599e-01 3.80428493e-01 4.60836589e-01
4.22058910e-01 -1.10837948e+00 -5.98389149e-01 7.62034059e-01
-2.32077897e-01 -9.08612072e-01 -1.07339716e+00 2.74663180e-01
-7.17268407e-01 -4.65724856e-01 -7.19589055e-01 -7.18971491e-01
2.93103397e-01 3.82728487e-01 5.51020026e-01 -1.66318920e-02
-2.15345453e-02 9.34619755e-02 -3.68805826e-01 -4.61684763e-01
-3.72144461e-01 -2.37439170e-01 1.70309357e-02 2.05276489e-01
-4.78052441e-03 -9.00586247e-01 -9.35462236e-01 2.50710905e-01
-8.92845690e-01 4.23415974e-02 4.51192617e-01 9.07479644e-01
2.81015068e-01 4.99516994e-01 8.47668231e-01 -1.49768800e-01
9.23957407e-01 -1.89546198e-01 -6.77829206e-01 2.37342604e-02
-3.07679892e-01 -2.21482649e-01 8.59957576e-01 -1.04472733e+00
-1.55360997e+00 -2.80417979e-01 -4.38814282e-01 -6.53167963e-01
2.69313723e-01 3.91085714e-01 -6.24304354e-01 3.47495109e-01
1.67062894e-01 6.09492719e-01 -2.55414784e-01 -5.14663517e-01
1.25151128e-01 9.27891016e-01 7.12967813e-01 3.21333408e-02
4.93082523e-01 1.66575208e-01 -3.38558704e-01 -8.63428950e-01
-4.64671582e-01 -2.84224451e-01 -3.23350847e-01 -4.38384086e-01
6.14377737e-01 -9.20448840e-01 -5.87693512e-01 8.23038697e-01
-1.28427529e+00 -1.46407276e-01 -8.35020989e-02 8.00776541e-01
-2.85589606e-01 2.63711512e-01 -7.54002094e-01 -1.15065241e+00
-5.95421970e-01 -1.09669554e+00 5.94387650e-01 3.54566902e-01
8.67467001e-02 -6.33834183e-01 -1.48861766e-01 -2.24738568e-01
8.55585098e-01 -4.97288197e-01 8.13535333e-01 -2.30902359e-01
-3.79921257e-01 -7.40719736e-02 6.87145442e-02 5.25777221e-01
1.70643330e-01 -1.91924408e-01 -1.31504917e+00 -2.38227844e-01
6.25833333e-01 4.49546814e-01 6.30981028e-01 7.59068489e-01
1.18296623e+00 -4.11172539e-01 -2.44720578e-02 7.42429018e-01
9.78922427e-01 9.34436142e-01 9.70813215e-01 -6.52840137e-02
4.70975637e-01 3.90140265e-01 6.69680178e-01 6.05083644e-01
-2.71749273e-02 7.69599319e-01 1.55479580e-01 -1.02996111e-01
-5.30638874e-01 -2.07338095e-01 3.74610394e-01 1.51767147e+00
2.97214508e-01 -4.12282169e-01 -4.87813264e-01 6.76050425e-01
-1.48689556e+00 -1.12611592e+00 2.36663133e-01 2.26473880e+00
9.57815945e-01 1.59175694e-01 -4.79030102e-01 4.05328512e-01
9.81257319e-01 3.31941426e-01 -7.90946245e-01 -4.05773222e-01
-2.20812991e-01 2.16793269e-01 2.60141701e-01 5.20346284e-01
-7.60138929e-01 6.49855852e-01 6.07804251e+00 1.30375850e+00
-1.39219356e+00 2.80956030e-01 5.07630825e-01 -2.66847521e-01
-2.80989241e-02 -4.47748721e-01 -4.61265773e-01 5.53999722e-01
1.30895007e+00 -4.38548267e-01 8.14448237e-01 4.65250760e-01
8.30823421e-01 1.77366287e-01 -6.91324770e-01 1.11852765e+00
-2.01728210e-01 -8.91499341e-01 -3.88745993e-01 -2.43263215e-01
4.78700668e-01 -1.91927731e-01 3.03897768e-01 2.90880442e-01
-3.66710454e-01 -7.79811442e-01 1.07069492e+00 8.13731253e-01
8.42503309e-01 -8.31140459e-01 3.38320047e-01 2.79231429e-01
-1.49333858e+00 -4.40944970e-01 -3.70577753e-01 1.27233282e-01
3.05986941e-01 7.02183068e-01 -4.33157742e-01 2.96367943e-01
6.11474216e-01 4.15655345e-01 -5.03383130e-02 1.13198817e+00
-1.31702378e-01 6.03341460e-01 -9.81437117e-02 1.70438275e-01
-1.23639695e-01 3.34256887e-02 9.04009283e-01 1.25946939e+00
5.14866889e-01 4.15419817e-01 -4.19158489e-01 3.70457828e-01
5.02732731e-02 -1.62103176e-01 -4.29610945e-02 1.58173040e-01
8.55175912e-01 8.44563425e-01 -2.49472350e-01 -1.39256299e-01
-2.33882546e-01 8.34353507e-01 6.15502559e-02 6.54437840e-01
-8.77862334e-01 -7.94542313e-01 9.01324153e-01 -2.13059355e-02
7.76019275e-01 -2.84745812e-01 -3.36355641e-02 -8.28829765e-01
1.05226882e-01 -9.81931686e-01 -7.82780349e-02 -1.06679785e+00
-8.14635217e-01 8.72464776e-01 -3.81935060e-01 -1.23688316e+00
-2.22914502e-01 -3.31354886e-01 -6.13147199e-01 1.06621265e+00
-1.42554200e+00 -5.75547814e-01 1.15668461e-01 6.31909311e-01
8.86444688e-01 -1.23653403e-02 4.86206651e-01 8.19701254e-01
-6.26534224e-01 6.89401865e-01 1.84908926e-01 -1.85890853e-01
5.86109281e-01 -7.27731407e-01 2.24693432e-01 1.18600798e+00
-1.20739862e-01 5.10981321e-01 9.03166711e-01 -6.62316144e-01
-1.30261946e+00 -1.01873553e+00 7.03417242e-01 3.33223730e-01
5.02331078e-01 -3.39946151e-01 -1.21706378e+00 1.20289102e-01
2.76141256e-01 -1.04757413e-01 3.12748730e-01 -3.04726422e-01
-7.17548057e-02 -5.67905664e-01 -7.03366876e-01 6.98971987e-01
8.86947632e-01 -8.67670298e-01 -4.49275404e-01 -2.30422691e-01
9.41431046e-01 -5.93263030e-01 -5.21022379e-01 2.43993357e-01
5.34410894e-01 -8.54260266e-01 9.06027913e-01 -2.68462181e-01
6.30925372e-02 -4.20005500e-01 -1.31708458e-01 -1.44239199e+00
-4.78752851e-01 -9.35004175e-01 -3.14053804e-01 1.37549031e+00
1.62002057e-01 -3.29071403e-01 1.42595038e-01 4.05012161e-01
-4.32837129e-01 -3.33268225e-01 -1.11476827e+00 -5.57714760e-01
-3.72138381e-01 -7.88990676e-01 7.05463052e-01 4.97249663e-01
-2.01724648e-01 7.13392496e-02 -7.44630754e-01 5.56312740e-01
1.96005046e-01 -3.37037355e-01 9.34697688e-02 -5.93881547e-01
-3.88187468e-01 -4.61536407e-01 -1.47216231e-01 -1.37725914e+00
-5.61114773e-03 -2.37779573e-01 3.18837166e-01 -1.15294516e+00
-2.60664254e-01 7.32145011e-02 -6.74201131e-01 -1.51010737e-01
-5.23426533e-01 -2.55560666e-01 4.07577842e-01 -1.25192225e-01
-4.58007216e-01 1.05865073e+00 1.37881875e+00 -1.34325996e-01
-3.30415368e-01 1.95256040e-01 -2.50881881e-01 7.10813582e-01
5.25199473e-01 -4.32669580e-01 -5.83337724e-01 -6.38650119e-01
-5.38469106e-02 6.75542891e-01 4.70913388e-02 -1.17048609e+00
5.16438305e-01 -5.90920374e-02 3.68451208e-01 -6.62790656e-01
4.69243169e-01 -1.10453165e+00 3.32010716e-01 3.91362220e-01
-4.29205537e-01 -3.68914045e-02 6.02510452e-01 7.22965181e-01
-5.07148445e-01 -2.47132465e-01 9.28035259e-01 2.16829583e-01
-5.90450108e-01 3.50395292e-01 -6.53120399e-01 -1.42887771e-01
4.60235387e-01 -8.60199928e-02 4.06025201e-02 -6.22287989e-01
-7.28511214e-01 -9.66750830e-02 -2.96673954e-01 3.86205435e-01
7.68108785e-01 -1.36631072e+00 -7.18046129e-01 3.07485551e-01
-3.96491736e-01 -4.91081655e-01 1.01038373e+00 6.29376769e-01
-4.26207259e-02 2.84274489e-01 -1.91423208e-01 -1.84889928e-01
-1.17677963e+00 6.78062379e-01 7.04555929e-01 -2.94359643e-02
-5.83448172e-01 8.24323833e-01 3.69945377e-01 4.21899706e-01
6.24082148e-01 -2.41327763e-01 -4.39400643e-01 -4.62010838e-02
1.05457366e+00 5.21629632e-01 1.77592129e-01 -5.63334882e-01
-5.98691255e-02 3.14698786e-01 5.52774668e-02 -3.45505655e-01
1.21939659e+00 -8.03692043e-01 1.66534811e-01 4.50429171e-01
1.06473470e+00 2.80673951e-01 -1.60500383e+00 -4.00494576e-01
-2.93077737e-01 -5.43799877e-01 7.65080154e-01 -8.04646432e-01
-1.37906718e+00 1.03397131e+00 1.03530777e+00 1.31332159e-01
1.89875972e+00 -7.11222529e-01 9.89171982e-01 -1.04868494e-01
-4.83372388e-03 -1.17567980e+00 2.23196335e-02 6.93999708e-01
1.10878944e+00 -6.19576156e-01 -3.36270630e-01 -1.68208499e-02
-3.12295049e-01 1.19149697e+00 5.07061005e-01 1.97894648e-01
8.09185028e-01 5.83418667e-01 1.47666737e-01 3.77462476e-01
-9.35976863e-01 -8.27742293e-02 3.51788998e-01 7.92915046e-01
1.69697374e-01 -2.04619035e-01 -8.25169086e-02 8.33731055e-01
-1.31400302e-01 -3.20931792e-01 1.94793329e-01 4.69970554e-01
-5.51453948e-01 -8.13718081e-01 -6.01761460e-01 2.32846886e-01
-5.41228235e-01 -4.67443913e-01 2.24796295e-01 1.80462539e-01
1.16350099e-01 1.19970942e+00 2.79240191e-01 -3.54912043e-01
5.55286884e-01 -7.29357302e-02 1.88471064e-01 -8.07651877e-02
-7.42611527e-01 8.01118791e-01 -1.24780729e-01 -2.67519593e-01
-1.11849420e-01 -4.63391423e-01 -1.16074824e+00 -2.28342950e-01
-6.06960833e-01 -9.57234949e-02 7.21613884e-01 7.04794526e-01
4.78976637e-01 1.33092523e+00 7.68499553e-01 -7.53922760e-01
-4.84847009e-01 -1.28010333e+00 -7.13305473e-01 1.94841444e-01
8.00887406e-01 -4.28572655e-01 -4.38274533e-01 6.99959993e-02] | [15.006919860839844, 5.924919605255127] |
93921dd8-19cc-40b4-b6a2-3e43fa50ad91 | fingpt-open-source-financial-large-language | 2306.06031 | null | https://arxiv.org/abs/2306.06031v1 | https://arxiv.org/pdf/2306.06031v1.pdf | FinGPT: Open-Source Financial Large Language Models | Large language models (LLMs) have shown the potential of revolutionizing natural language processing tasks in diverse domains, sparking great interest in finance. Accessing high-quality financial data is the first challenge for financial LLMs (FinLLMs). While proprietary models like BloombergGPT have taken advantage of their unique data accumulation, such privileged access calls for an open-source alternative to democratize Internet-scale financial data. In this paper, we present an open-source large language model, FinGPT, for the finance sector. Unlike proprietary models, FinGPT takes a data-centric approach, providing researchers and practitioners with accessible and transparent resources to develop their FinLLMs. We highlight the importance of an automatic data curation pipeline and the lightweight low-rank adaptation technique in building FinGPT. Furthermore, we showcase several potential applications as stepping stones for users, such as robo-advising, algorithmic trading, and low-code development. Through collaborative efforts within the open-source AI4Finance community, FinGPT aims to stimulate innovation, democratize FinLLMs, and unlock new opportunities in open finance. Two associated code repos are \url{https://github.com/AI4Finance-Foundation/FinGPT} and \url{https://github.com/AI4Finance-Foundation/FinNLP} | ['Christina Dan Wang', 'Xiao-Yang Liu', 'Hongyang Yang'] | 2023-06-09 | null | null | null | null | ['algorithmic-trading'] | ['time-series'] | [-8.49996865e-01 -2.79493760e-02 -2.59268403e-01 -4.06511784e-01
-1.04410350e+00 -7.96447337e-01 7.29839087e-01 2.13516861e-01
-5.69465160e-01 3.76194715e-01 5.50012410e-01 -6.72755659e-01
-7.74474740e-02 -8.39833677e-01 -5.28161764e-01 8.03378895e-02
-2.18516856e-01 7.69068539e-01 -1.19851179e-01 -8.56392980e-02
4.33678031e-01 5.67606509e-01 -1.03814363e+00 5.78307450e-01
5.69579899e-01 8.20454597e-01 5.25205061e-02 4.14018422e-01
-5.62472641e-01 1.04843962e+00 -1.87154919e-01 -8.78686249e-01
7.37198114e-01 9.97841433e-02 -8.33061635e-01 -3.75666171e-01
2.69599885e-01 -2.39089116e-01 -6.27447590e-02 9.07480896e-01
3.68523777e-01 -5.60483821e-02 1.41201258e-01 -1.07935870e+00
-7.07941234e-01 8.35745454e-01 -6.11108482e-01 3.56821477e-01
1.58104196e-01 4.40697551e-01 1.62515700e+00 -1.22613239e+00
7.09166765e-01 1.17662311e+00 5.46475947e-01 1.66958854e-01
-1.00871849e+00 -8.33927214e-01 5.79068996e-02 -5.51981926e-02
-1.06220484e+00 -6.80916846e-01 4.88599300e-01 -6.35530651e-01
1.11407959e+00 3.00710976e-01 4.93927956e-01 8.11037123e-01
-2.46880688e-02 1.02508807e+00 1.20501995e+00 -2.07940191e-01
8.42270777e-02 2.61623889e-01 3.18743020e-01 6.54677868e-01
2.92839497e-01 -1.54465154e-01 -8.20761144e-01 -5.32127380e-01
8.65637481e-01 1.74932163e-02 1.90294996e-01 -6.25029672e-04
-1.32693219e+00 1.08455265e+00 1.19571418e-01 5.66199958e-01
-3.73261243e-01 5.39393127e-02 2.53719240e-01 5.06243229e-01
7.16834962e-01 7.03662574e-01 -7.21625805e-01 -4.84429538e-01
-1.11553907e+00 5.35592556e-01 1.18542254e+00 5.36912620e-01
5.48552692e-01 -1.02918871e-01 2.87203997e-01 8.44943702e-01
4.46312606e-01 3.35385591e-01 5.84878027e-01 -1.11385119e+00
7.00680554e-01 7.37450778e-01 1.21301204e-01 -8.48483860e-01
-3.78942549e-01 -3.84882003e-01 -5.25945485e-01 2.49054536e-01
7.05682635e-01 6.17515743e-02 -8.88890252e-02 1.24002635e+00
2.41832793e-01 -3.02136779e-01 4.47114781e-02 9.08443332e-01
4.44608986e-01 5.75088441e-01 -9.78297591e-02 6.94023669e-02
1.23681450e+00 -8.59506071e-01 -1.63435400e-01 -2.12401658e-01
7.83997416e-01 -1.07094598e+00 1.16706455e+00 6.84987605e-01
-1.18042827e+00 -2.50106961e-01 -5.73668718e-01 -2.55844653e-01
-5.11302531e-01 -3.18004847e-01 1.13334954e+00 3.78093481e-01
-8.99272025e-01 6.46250546e-01 -9.03453946e-01 -1.23429060e-01
5.35989046e-01 1.62538528e-01 -3.42400372e-01 5.07004894e-02
-1.17789602e+00 6.24679148e-01 1.34998366e-01 -1.69454321e-01
-5.41471243e-01 -8.04596484e-01 -4.16668385e-01 -1.60481960e-01
2.90750861e-01 -4.60063338e-01 1.23379385e+00 -7.25614548e-01
-1.12566996e+00 1.02552044e+00 1.96960881e-01 -6.49019241e-01
8.16677511e-01 -5.23540974e-01 -4.42587614e-01 -1.07749477e-01
2.95956016e-01 2.43030459e-01 4.09310102e-01 -3.70254576e-01
-4.87752348e-01 -2.69952506e-01 -5.06994501e-02 -2.83265591e-01
-6.13217235e-01 5.60016930e-01 -5.31370997e-01 -8.92588735e-01
-2.26421714e-01 -6.10379279e-01 -3.53064507e-01 -1.65018141e-01
-1.76029727e-01 -1.38944224e-01 4.18578506e-01 -8.33344460e-01
1.40026808e+00 -2.09464788e+00 -2.13436961e-01 2.99656719e-01
4.38046873e-01 2.34906614e-01 -1.18788213e-01 7.12369263e-01
1.57295726e-02 4.19200301e-01 -3.71422083e-03 -3.93543214e-01
2.97327727e-01 -1.07925646e-01 -4.18738991e-01 2.12169334e-01
-1.59891434e-02 1.04863966e+00 -7.79076695e-01 -3.60812724e-01
-2.89410856e-02 1.79962143e-01 -7.40535617e-01 5.20735933e-03
-4.10844058e-01 2.61409551e-01 -7.68897414e-01 9.44480360e-01
4.88432556e-01 -6.38679981e-01 1.68850496e-01 1.20064259e-01
-6.20034158e-01 3.74988884e-01 -1.33113587e+00 1.51337528e+00
-4.90682364e-01 2.20346451e-01 3.84891361e-01 -5.11722028e-01
8.52865636e-01 8.44298825e-02 5.59373677e-01 -6.69490755e-01
-1.14052750e-01 6.90744340e-01 -2.15802789e-01 -3.11373979e-01
5.22542775e-01 -2.59139121e-01 -2.02812806e-01 1.10279846e+00
-2.66631663e-01 1.09363183e-01 4.99823749e-01 4.52845603e-01
1.03437328e+00 2.48108998e-01 1.92918748e-01 -2.01720983e-01
1.53768569e-01 -2.72320770e-02 5.58714032e-01 5.28336108e-01
1.17428591e-02 3.40242565e-01 4.81010944e-01 -6.55483961e-01
-9.95884717e-01 -9.32207823e-01 -3.18745673e-01 1.10179937e+00
-8.58277738e-01 -7.58228898e-01 -3.68503243e-01 -6.31045043e-01
5.85595131e-01 5.17197549e-01 -5.46220131e-03 7.24463105e-01
-5.11563301e-01 -9.57282186e-01 6.67038858e-01 1.26268879e-01
6.16410553e-01 -8.90881360e-01 -5.83929121e-01 4.14562315e-01
-2.79259056e-01 -9.44653332e-01 -3.83161068e-01 -5.35770766e-02
-7.62492478e-01 -1.00981438e+00 -6.66226268e-01 -2.53100127e-01
1.63261145e-01 -1.96288019e-01 1.22545457e+00 -9.09538269e-02
-3.08573961e-01 1.05902739e-01 -1.27587184e-01 -2.91468084e-01
-3.65369648e-01 2.20936745e-01 1.35952413e-01 1.46817081e-02
6.34139001e-01 -5.57357013e-01 -4.29567724e-01 -5.99665241e-03
-8.82894993e-01 1.44727558e-01 4.66333479e-01 3.98455411e-01
5.59206545e-01 -1.73555911e-01 6.88908815e-01 -1.29875910e+00
7.89462447e-01 -8.03838730e-01 -1.06583083e+00 1.41830876e-01
-8.72629166e-01 -1.48800937e-02 5.83388031e-01 -5.67706041e-02
-6.89049661e-01 -3.71968836e-01 -1.60775289e-01 -3.41974944e-01
1.76248685e-01 1.06403863e+00 3.16631608e-02 2.44980037e-01
5.63080907e-01 -2.48794496e-01 6.15198165e-02 -1.21110833e+00
4.41701591e-01 8.49870443e-01 2.21718937e-01 -6.66591823e-01
8.19771230e-01 3.87714893e-01 -3.07077557e-01 -5.33324480e-01
-8.85129154e-01 -4.73201782e-01 -4.04235989e-01 7.69379586e-02
6.32506907e-01 -9.97115314e-01 -5.98369300e-01 4.94685471e-01
-6.93648040e-01 -5.09056270e-01 -3.62034827e-01 4.26187694e-01
-7.98888952e-02 4.13175464e-01 -1.00912380e+00 -5.99431932e-01
-4.67138886e-01 -6.92583144e-01 5.29588640e-01 -5.29578030e-02
-5.81176460e-01 -1.24387646e+00 1.66801184e-01 1.05300939e+00
5.89650035e-01 2.86988646e-01 7.68848360e-01 -1.19068170e+00
-9.35954154e-01 -3.90594751e-01 -2.27608815e-01 3.85268986e-01
-6.25307038e-02 -1.47841983e-02 -8.93930554e-01 -3.31530780e-01
-6.08733632e-02 -4.21518385e-01 6.50530338e-01 -5.74359484e-02
8.64849925e-01 -5.54904044e-01 5.46052977e-02 5.65440953e-01
1.22911298e+00 -3.00411701e-01 8.72958675e-02 8.33833039e-01
7.12240696e-01 6.70090616e-01 5.61806917e-01 8.89562190e-01
5.12718379e-01 4.04951394e-01 -5.69883883e-02 -5.61750717e-02
2.14692399e-01 -2.17962980e-01 6.39996469e-01 1.11868441e+00
-1.38657419e-02 1.89952880e-01 -1.48670924e+00 3.15211743e-01
-1.87913680e+00 -8.89616191e-01 -2.90438324e-01 2.21100354e+00
1.07184541e+00 4.97387290e-01 9.27798599e-02 -3.44926000e-01
9.38956365e-02 3.26959550e-01 -6.21317923e-01 -3.52774173e-01
-3.31210196e-01 4.06589746e-01 6.48732781e-01 6.57715261e-01
-8.77863109e-01 1.10171223e+00 5.12966537e+00 9.26735282e-01
-9.67334628e-01 3.48142445e-01 9.57455277e-01 -7.04934895e-01
-4.59594518e-01 2.44332135e-01 -1.01748288e+00 4.38462645e-01
1.38662946e+00 -6.08654380e-01 5.82741916e-01 8.96850705e-01
5.98228395e-01 8.28840509e-02 -9.50567067e-01 6.63898408e-01
-2.92710423e-01 -1.77434957e+00 3.45757641e-02 5.10822177e-01
4.48986053e-01 7.68826544e-01 1.38999015e-01 1.56431869e-01
7.12270319e-01 -9.64676619e-01 8.47660482e-01 7.27361441e-01
7.57467330e-01 -6.06687069e-01 5.28572321e-01 2.68486530e-01
-1.15800405e+00 -2.22084522e-01 -2.20524654e-01 -3.48605484e-01
3.64904761e-01 9.34341669e-01 -4.48134005e-01 5.10740519e-01
7.77519643e-01 9.86534178e-01 -7.00022399e-01 6.44320846e-01
2.37571802e-02 7.21504688e-01 -2.73362905e-01 3.22532952e-01
2.17168927e-01 -3.00542235e-01 4.48037714e-01 1.14887750e+00
1.52936298e-02 -3.70752113e-03 3.71176064e-01 9.08837914e-01
-2.97403932e-01 3.36212784e-01 -3.12273890e-01 -7.86851943e-01
3.33891869e-01 1.39189315e+00 -6.32839382e-01 -2.86720783e-01
-7.36584485e-01 5.28456867e-01 3.79142314e-01 1.91956148e-01
-5.06493092e-01 -2.92212933e-01 6.98426008e-01 6.10815823e-01
3.54587026e-02 -3.66830379e-01 -3.37028682e-01 -1.58718443e+00
1.40345970e-03 -1.23892581e+00 7.74210930e-01 -5.35450280e-01
-1.53604126e+00 4.54699486e-01 -1.06745824e-01 -8.88654649e-01
-3.22931439e-01 -5.82627356e-01 -3.33987355e-01 1.07323420e+00
-1.51302803e+00 -1.21553540e+00 2.64665931e-01 5.77630520e-01
3.48138899e-01 -6.37605071e-01 6.14003718e-01 5.04868805e-01
-5.49997926e-01 3.69929552e-01 3.60312074e-01 3.03940266e-01
8.54491472e-01 -9.97989118e-01 7.93404281e-01 7.14675725e-01
4.79688376e-01 1.10427976e+00 3.95343095e-01 -6.89135849e-01
-1.34822595e+00 -1.00324702e+00 1.27493036e+00 -8.94881487e-01
1.43494499e+00 -5.07633865e-01 -9.08956409e-01 8.42301309e-01
-6.89379722e-02 -2.43680686e-01 9.63703454e-01 3.38415205e-01
-6.48039341e-01 -1.20405480e-01 -8.74428868e-01 3.38596731e-01
5.98208785e-01 -7.21795499e-01 -5.97451448e-01 6.27960980e-01
5.42403281e-01 1.03949361e-01 -1.13597405e+00 -7.25981072e-02
4.84224916e-01 -1.09708881e+00 8.73106122e-01 -4.77515668e-01
5.63172400e-01 9.75369737e-02 -1.01955757e-01 -5.27008057e-01
-3.27743083e-01 -1.00827038e+00 -1.19827934e-01 1.44197559e+00
5.52961767e-01 -9.98512805e-01 8.20215285e-01 1.08211660e+00
1.72115132e-01 -7.76823759e-01 -8.76453221e-01 -7.33828247e-01
4.96064872e-01 -8.12732935e-01 6.10438466e-01 1.19742739e+00
1.98602557e-01 1.19646318e-01 -2.42118523e-01 -1.80474579e-01
6.55380070e-01 4.00523961e-01 8.51604104e-01 -1.27854669e+00
-8.68762314e-01 -5.83906114e-01 7.92966187e-02 -8.00878108e-01
3.83565091e-02 -1.19638479e+00 -1.00890303e+00 -1.42703521e+00
2.85411239e-01 -6.90428078e-01 -3.57806921e-01 8.00302267e-01
3.86954516e-01 2.34089226e-01 4.79281485e-01 8.53178263e-01
-5.93109190e-01 -6.71493337e-02 8.27413261e-01 1.41800508e-01
-1.40767470e-01 -3.64954285e-02 -1.02699983e+00 7.39661574e-01
7.60224521e-01 -3.90011340e-01 9.25510153e-02 -4.71764237e-01
6.78845942e-01 -8.87891650e-02 4.19582218e-01 -6.00015759e-01
2.17667267e-01 -2.47902900e-01 1.06027156e-01 -3.24972749e-01
6.03527501e-02 -4.40610677e-01 2.41996601e-01 4.48941588e-01
-2.45085135e-01 3.02094758e-01 1.42609015e-01 9.67747867e-02
-3.44253212e-01 -2.17450932e-01 5.04046559e-01 -8.23252141e-01
-5.42879403e-01 3.79987925e-01 -2.26578444e-01 4.42699373e-01
6.87295973e-01 1.28733888e-01 -2.82061368e-01 -1.51573896e-01
-6.17453039e-01 3.59287053e-01 5.31178653e-01 4.54145908e-01
2.74074465e-01 -9.47417140e-01 -9.29071069e-01 2.93623120e-01
-1.11161582e-01 -1.83920488e-01 -8.47004279e-02 8.36462915e-01
-6.63551688e-01 6.42746329e-01 6.97505698e-02 1.89472809e-01
-9.81627762e-01 2.20622078e-01 7.27668256e-02 -7.65122831e-01
-5.12737751e-01 1.00019085e+00 9.39330959e-04 -5.15429974e-01
-1.87385678e-02 -3.40204477e-01 7.38334060e-02 3.71289819e-01
7.15446472e-01 2.67821074e-01 -1.53251272e-03 -4.57571089e-01
-2.00765625e-01 4.77324761e-02 -4.21522468e-01 -4.33928609e-01
1.89594805e+00 2.01224908e-02 -4.25416321e-01 7.24180043e-01
8.94925117e-01 5.59458196e-01 -1.11303401e+00 -2.53308237e-01
6.88114762e-01 -6.07431650e-01 5.88745624e-02 -9.53554034e-01
-1.07921135e+00 6.33859754e-01 9.00358930e-02 1.41152143e-01
7.35905051e-01 9.04755294e-02 8.97938490e-01 3.84465963e-01
6.19613469e-01 -1.10719860e+00 7.74417818e-02 6.26201034e-01
8.60382497e-01 -1.12525034e+00 2.63812870e-01 2.04232901e-01
-6.92350626e-01 1.04352450e+00 1.38220534e-01 1.12291716e-01
7.70651698e-01 4.52318192e-01 3.13255042e-01 -4.25918221e-01
-9.74178731e-01 2.40448609e-01 -9.83298048e-02 -1.55992471e-02
8.62504482e-01 -8.71173367e-02 -2.07456291e-01 1.07405567e+00
-5.81958532e-01 1.49652779e-01 4.18881148e-01 1.00090981e+00
-3.08729500e-01 -1.69474471e+00 -3.95343244e-01 7.33937562e-01
-1.01430631e+00 -6.62048757e-01 -4.13235307e-01 7.32539177e-01
-1.47040650e-01 6.84906006e-01 -4.87147197e-02 2.88631357e-02
1.79046214e-01 2.36659408e-01 -2.33401313e-01 -7.82239974e-01
-1.04494250e+00 2.09508598e-01 2.51143187e-01 -7.60006547e-01
-1.39095709e-01 -1.11168933e+00 -1.21358192e+00 -6.72249556e-01
1.29476860e-01 4.83343780e-01 5.11625051e-01 6.40338421e-01
5.85353494e-01 -1.44442037e-01 2.78571993e-01 -4.58819658e-01
-6.50858581e-01 -7.48152792e-01 -6.80259466e-01 1.69198573e-01
-2.68155318e-02 -2.29211047e-01 -4.36784983e-01 1.20000355e-01] | [10.637198448181152, 8.35982608795166] |
2672a54a-cae0-4386-901d-2cc2adffd766 | a-template-independent-approach-for | null | null | https://www.ital-ia2023.it/workshop/ai-per-la-finanza-ed-il-commercio | https://www.ital-ia2023.it/submission/17/paper | A template-independent approach for information extraction in real estate documents | Business corporations manage tons of unstructured data daily, such as PDFs and websites. Recent advances in the deep
learning field help find insight from this unstructured information. New models leverage the power of the Transformer
architecture to accomplish natural language understanding tasks on these data, jointly using the raw image and its text
content or directly the image without OCR. We propose an extraction pipeline that employs question-answering models
to get insight from unstructured data, allowing fast and efficient information retrieval from different sources. We show an
application of this technique to a specific set of documents and how we can scale this infrastructure to different types of
records. Our solution can effectively handle large document corpora robustly, helping corporations exploit all the power
coming from their data. | ['Ignazio Gallo', 'Stefano Taverni', 'Gabriele Destro', 'Nicola Landro'] | 2023-05-30 | null | null | null | ital-ia-2023-5 | ['optical-character-recognition', 'retrieval', 'question-answering', 'information-retrieval', 'natural-language-understanding'] | ['computer-vision', 'methodology', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-1.43259078e-01 2.33408332e-01 -1.87469289e-01 -1.47106752e-01
-1.20809746e+00 -1.15663946e+00 6.57391846e-01 3.92483205e-01
-2.83295631e-01 2.29556799e-01 4.42183346e-01 -5.38307548e-01
-1.17557265e-01 -1.02672029e+00 -6.36052847e-01 -8.98300037e-02
8.82138386e-02 7.52960920e-01 1.44984797e-01 -5.91369905e-02
6.43773556e-01 5.11738539e-01 -9.86651599e-01 7.48930812e-01
4.21152532e-01 1.05885625e+00 5.18548489e-01 6.26415908e-01
-9.33448136e-01 1.33515012e+00 -5.35993159e-01 -9.35502231e-01
5.40163934e-01 5.90802610e-01 -1.14792991e+00 1.27767364e-03
8.05646002e-01 -9.23263133e-01 -3.58439863e-01 8.91370475e-01
-1.54167768e-02 -5.58183670e-01 2.81162143e-01 -7.79662371e-01
-1.10273206e+00 7.24322975e-01 -4.83253747e-01 5.05218029e-01
1.20160766e-01 1.85454711e-01 1.50531471e+00 -1.04425621e+00
9.50505137e-01 1.21480989e+00 5.29239535e-01 -3.82840708e-02
-8.74866486e-01 -3.46466839e-01 -1.68712959e-02 3.08906585e-01
-7.68483639e-01 -4.30937350e-01 4.74447727e-01 -5.38179994e-01
1.46512198e+00 -1.26082256e-01 5.15144527e-01 8.95364165e-01
1.16355635e-01 1.16034925e+00 8.75099182e-01 -2.12295771e-01
-1.94737718e-01 2.98774302e-01 4.17420775e-01 8.83525968e-01
3.52848828e-01 -3.58147860e-01 -5.12072623e-01 3.18672329e-01
6.41150773e-01 5.54523289e-01 1.43681332e-01 6.26175776e-02
-1.22498810e+00 8.35369527e-01 5.37714362e-01 3.38338822e-01
-2.44105473e-01 1.63446665e-01 5.38787603e-01 4.95029211e-01
2.26467893e-01 8.44175696e-01 -6.68519676e-01 -2.96899408e-01
-8.09158444e-01 2.82411184e-02 1.13064742e+00 1.26785398e+00
8.92433941e-01 -1.88574061e-01 1.26473799e-01 4.82407451e-01
3.04502398e-01 5.02090991e-01 3.27809095e-01 -1.08841479e+00
1.20772755e+00 7.99846351e-01 -5.96547872e-02 -1.15415359e+00
-4.95497882e-02 -3.21838886e-01 -4.80111629e-01 -1.92835599e-01
4.70465720e-01 3.73542994e-01 -1.05203283e+00 8.76473129e-01
-7.27121234e-02 -6.39900625e-01 -1.45930514e-01 5.36294460e-01
8.36729288e-01 7.85663605e-01 -7.56530464e-02 4.59740371e-01
1.59434557e+00 -1.20765626e+00 -6.82478189e-01 -5.17927408e-01
3.02783251e-01 -6.53800309e-01 9.75043774e-01 7.07836151e-01
-9.26184595e-01 -3.96963567e-01 -1.01149714e+00 -9.39928770e-01
-9.41389441e-01 -2.99220443e-01 5.83437920e-01 1.51345402e-01
-1.01666391e+00 6.25020921e-01 -8.20742786e-01 -2.12699860e-01
1.02918065e+00 3.37505847e-01 -4.35664356e-01 -4.78389561e-01
-8.56943667e-01 7.50027716e-01 2.89075851e-01 -7.24163055e-02
-8.16179276e-01 -8.51607144e-01 -8.27793717e-01 5.86177707e-01
6.03336573e-01 -6.16382897e-01 1.64667475e+00 -6.97219193e-01
-9.29501653e-01 7.67843127e-01 3.15988995e-02 -5.04748344e-01
1.49868533e-01 -7.44544387e-01 3.93172689e-02 6.34375393e-01
2.23002881e-01 4.17487174e-01 1.05188656e+00 -1.04263878e+00
-5.19215822e-01 -7.16393650e-01 3.80060151e-02 -4.54630256e-01
-8.99647653e-01 1.61318690e-01 -6.71436071e-01 -5.21589220e-01
-1.99834615e-01 -4.06345397e-01 -6.57904297e-02 -1.20717622e-01
-6.51695073e-01 -1.72824755e-01 1.00307345e+00 -1.14682102e+00
8.57822597e-01 -1.95030391e+00 -2.81235367e-01 1.47472665e-01
8.29630852e-01 1.00210324e-01 -3.28278691e-01 6.81214988e-01
2.67605960e-01 5.35290956e-01 2.07765415e-01 -1.67371884e-01
3.70955944e-01 7.56993890e-02 -8.50631595e-01 -9.37780738e-02
6.92593694e-01 1.48542821e+00 -8.71481121e-01 -5.77522755e-01
-8.50238651e-02 2.71443754e-01 -4.19261366e-01 2.92761534e-01
-4.62604821e-01 -4.38652188e-02 -7.80397594e-01 9.66854095e-01
4.40935075e-01 -8.59616756e-01 2.64962882e-01 -1.84414208e-01
-1.07144326e-01 7.16017306e-01 -6.11521244e-01 1.58154440e+00
-7.10804582e-01 1.06911707e+00 4.39518064e-01 -9.81850684e-01
8.81131768e-01 1.00280054e-01 4.94677842e-01 -9.87358987e-01
-1.19132072e-01 9.22378078e-02 -4.95843619e-01 -7.12588608e-01
6.84240282e-01 -1.11634575e-01 -1.01883814e-01 7.30496168e-01
3.16066891e-01 -2.81615198e-01 3.22036058e-01 4.48277652e-01
1.52866364e+00 -1.86575890e-01 1.40986115e-01 7.21660852e-02
-1.10012345e-01 3.28727990e-01 -2.55844947e-02 9.60341334e-01
1.23555586e-01 4.51082259e-01 4.75053251e-01 -9.59239483e-01
-1.60047472e+00 -1.03527534e+00 1.10702910e-01 1.06563032e+00
-3.45686167e-01 -7.36431181e-01 -4.06007797e-01 -7.82027543e-01
3.76375407e-01 2.58764863e-01 -1.75159037e-01 3.62329602e-01
-6.24297082e-01 -1.25254601e-01 1.55675501e-01 7.36037612e-01
5.28552353e-01 -1.10437775e+00 -4.43236411e-01 2.63901293e-01
-1.34095162e-01 -1.37246168e+00 -1.34262428e-01 4.43644464e-01
-1.03416979e+00 -9.30158079e-01 -3.92327607e-01 -7.02272832e-01
3.62676919e-01 1.37732387e-01 1.77237380e+00 5.25020547e-02
-5.18904388e-01 5.89936912e-01 -1.67319983e-01 -4.84634668e-01
-2.93533564e-01 4.95579988e-01 -4.57432508e-01 -3.70679498e-01
6.63183391e-01 -4.67592239e-01 -3.57471377e-01 -2.83409387e-01
-1.15177512e+00 -1.06418394e-01 8.37711334e-01 5.61672926e-01
2.26017922e-01 2.39138246e-01 3.68201554e-01 -1.03230679e+00
5.92399359e-01 -7.98871815e-01 -7.95378804e-01 3.28999698e-01
-7.12818325e-01 3.78467441e-01 8.43737245e-01 -3.72298583e-02
-8.98281217e-01 -3.29928845e-01 -2.32120994e-02 -4.64506149e-01
1.74105857e-02 6.18911326e-01 -1.18800789e-01 3.82118791e-01
2.27767915e-01 7.43827522e-02 -2.74616688e-01 -9.16381419e-01
6.28040612e-01 8.56977820e-01 7.66210198e-01 -5.73262513e-01
1.08877838e+00 6.22613966e-01 -4.39008087e-01 -3.75934631e-01
-1.22778106e+00 -6.08885765e-01 -7.79658377e-01 2.58015096e-01
7.47724950e-01 -9.43030953e-01 -6.32116258e-01 1.60526589e-01
-1.34612489e+00 1.37002831e-02 -5.40257215e-01 -6.03640452e-02
-2.25567251e-01 2.73521096e-01 -9.06045854e-01 -3.11440319e-01
-8.49111915e-01 -7.60870039e-01 1.45258951e+00 2.54403114e-01
4.41406742e-02 -8.66912544e-01 9.35853869e-02 9.86454844e-01
5.78175068e-01 -3.48489463e-01 1.12565875e+00 -1.11744905e+00
-1.52850747e+00 -3.93979937e-01 -6.97775483e-01 4.68332231e-01
4.63455692e-02 -1.91775769e-01 -9.98356044e-01 -2.64282137e-01
8.93690363e-02 -6.24214768e-01 1.10471654e+00 -3.28465372e-01
1.49672449e+00 -8.72600079e-01 -2.14089841e-01 6.51737392e-01
1.66603732e+00 -1.16904624e-01 7.21787572e-01 9.02822375e-01
1.03679836e+00 5.79685032e-01 4.96021584e-02 3.54370892e-01
5.40953875e-01 7.46074645e-03 6.68167770e-01 5.42284874e-03
1.65746853e-01 -4.06031728e-01 1.49934351e-01 1.11545110e+00
2.44925186e-01 -2.82350093e-01 -1.20692909e+00 9.22665417e-01
-1.38741493e+00 -1.20018435e+00 3.98454070e-01 1.56450593e+00
8.17428350e-01 3.33769739e-01 -3.79182905e-01 -1.84145376e-01
2.59313494e-01 2.55942404e-01 -7.11632550e-01 -4.63345557e-01
1.25167161e-01 4.82259244e-01 5.63202500e-01 7.10206032e-02
-1.02315450e+00 6.34713531e-01 7.19375515e+00 5.07931352e-01
-9.36542571e-01 -1.85258716e-01 6.14078462e-01 -2.00424135e-01
-8.26963782e-01 -8.35343227e-02 -9.46595371e-01 2.58019418e-01
1.12139118e+00 -1.08880185e-01 3.89262974e-01 1.14763725e+00
-2.37990946e-01 2.71002471e-01 -1.31637621e+00 8.84567499e-01
-1.04420461e-01 -1.91662705e+00 2.55766392e-01 3.06844920e-01
4.93474275e-01 3.84205788e-01 3.62202823e-01 3.70392978e-01
7.84319401e-01 -1.25124812e+00 5.00802934e-01 6.12264574e-01
6.94439888e-01 -3.50151420e-01 5.65936804e-01 1.16159245e-01
-9.91625845e-01 -6.07502460e-01 -2.55986482e-01 1.77563429e-01
-8.44594687e-02 8.13592076e-01 -1.51627660e+00 3.11971694e-01
9.34022248e-01 1.00392079e+00 -9.39766288e-01 7.14218378e-01
-1.02816299e-01 4.02625978e-01 -3.63276213e-01 -1.70642231e-02
3.72661918e-01 -1.32254630e-01 2.68357784e-01 1.35970330e+00
4.08967972e-01 -1.96677685e-01 3.35761681e-02 1.13154185e+00
-5.62948883e-01 3.07457633e-02 -8.88046265e-01 -7.91553497e-01
4.68008995e-01 1.58540964e+00 -6.52903378e-01 -4.99212950e-01
-9.05762792e-01 6.30115986e-01 7.14047670e-01 1.92599773e-01
-1.06534049e-01 -6.65337622e-01 3.28245848e-01 2.99659610e-01
9.16942179e-01 -4.50561315e-01 -2.77747691e-01 -1.26209676e+00
1.95774600e-01 -1.27324152e+00 4.54187006e-01 -1.18881845e+00
-1.50141501e+00 3.53099465e-01 -3.45265001e-01 -7.68345952e-01
-3.00806344e-01 -9.05951142e-01 -3.13588649e-01 4.54569012e-01
-1.83893621e+00 -1.11160839e+00 -1.44363657e-01 5.58885455e-01
8.65848064e-01 -4.57420558e-01 5.13271809e-01 2.42905080e-01
-3.31573129e-01 2.31872834e-02 4.40270990e-01 8.33360553e-01
5.64194262e-01 -1.74647307e+00 7.28735149e-01 5.25101006e-01
6.22070372e-01 9.38437760e-01 1.45286031e-03 -3.92953336e-01
-1.97985756e+00 -9.12794530e-01 8.33867192e-01 -1.00005317e+00
1.20434606e+00 -5.22583902e-01 -7.65136838e-01 1.01933289e+00
6.24983668e-01 -1.74425751e-01 8.23424935e-01 2.36124367e-01
-9.91483092e-01 -2.82002807e-01 -8.47771168e-01 3.08641851e-01
6.21716857e-01 -1.02814543e+00 -1.02492976e+00 4.02896434e-01
1.18942630e+00 2.18941383e-02 -1.15606713e+00 -7.57220713e-03
4.16292161e-01 -7.40277290e-01 1.10281682e+00 -9.50959504e-01
1.23209631e+00 1.91263989e-01 -1.37058392e-01 -7.86435962e-01
-3.60150099e-01 -6.80771112e-01 -5.20498037e-01 1.24195862e+00
3.69983584e-01 -1.95796207e-01 6.09018862e-01 7.00559556e-01
7.84442201e-02 -4.88297760e-01 -5.30041754e-01 -3.71320039e-01
1.98824346e-01 -5.25853872e-01 7.40760565e-01 8.35297048e-01
8.55225697e-03 7.34914005e-01 5.09708375e-02 -1.08150959e-01
6.60839736e-01 6.60080492e-01 8.03698123e-01 -1.32616746e+00
-2.31863692e-01 -1.59656703e-01 -2.22594708e-01 -1.09743917e+00
-1.00581506e-02 -1.02569294e+00 -2.36260980e-01 -1.84294987e+00
5.74194074e-01 -2.61468887e-01 -1.65442124e-01 8.73931408e-01
1.37195572e-01 2.13213310e-01 5.50834417e-01 4.54255104e-01
-9.20525730e-01 1.21586204e-01 1.37441885e+00 -6.49004579e-01
4.40508008e-01 -5.36182046e-01 -1.25776446e+00 6.95598602e-01
4.72820193e-01 -4.84027267e-01 -1.52483687e-01 -1.04789472e+00
8.60568583e-01 -1.65420711e-01 3.72540236e-01 -6.25775039e-01
3.06685746e-01 1.11403890e-01 7.08325326e-01 -9.20706093e-01
-6.29002079e-02 -1.13659108e+00 -4.31455314e-01 7.79126436e-02
-4.06453252e-01 1.36501804e-01 2.69893229e-01 5.62353015e-01
-3.87948930e-01 -3.31194967e-01 2.45061204e-01 -8.22132707e-01
-6.79151118e-01 3.61459315e-01 -4.01339501e-01 3.96187544e-01
4.75120872e-01 1.71554625e-01 -9.19754744e-01 -4.47056621e-01
-5.49263358e-01 2.93901563e-01 3.81580174e-01 6.12389505e-01
4.92698312e-01 -1.10168862e+00 -5.61308682e-01 1.88479275e-01
-4.12356034e-02 2.53940612e-01 -2.82975703e-01 3.93026829e-01
-6.05532408e-01 7.96915829e-01 -2.14414507e-01 -4.57285821e-01
-6.67570233e-01 1.09019601e+00 -3.19781452e-02 -1.00954294e+00
-7.01656342e-01 5.59993684e-01 1.29005149e-01 -3.92782986e-01
7.41480365e-02 -5.55009067e-01 -1.42073959e-01 3.27713430e-01
7.19958186e-01 8.45359787e-02 2.29950503e-01 9.98700634e-02
-1.87577590e-01 5.08860111e-01 -6.79756284e-01 -4.20839228e-02
1.94420373e+00 -1.73841819e-01 -4.63226795e-01 1.31561488e-01
1.70862651e+00 2.79270947e-01 -9.73132730e-01 -5.68023324e-01
5.62627256e-01 -5.08032262e-01 -1.56703200e-02 -7.48182356e-01
-1.39727783e+00 1.16711330e+00 5.80369234e-02 6.53957903e-01
1.00342286e+00 3.04451883e-01 1.12263095e+00 1.17842174e+00
2.23561510e-01 -1.12476325e+00 5.84038913e-01 7.21340537e-01
6.09017730e-01 -1.49417579e+00 1.62221372e-01 1.12355925e-01
-3.27784032e-01 1.52305734e+00 2.99100310e-01 4.36615162e-02
5.24037778e-01 8.18733931e-01 1.20208979e-01 -7.85636127e-01
-1.00724936e+00 -3.95168141e-02 6.09232001e-02 6.75158381e-01
1.14020698e-01 -5.63631415e-01 7.53565013e-01 6.48498774e-01
-3.38243902e-01 5.43824248e-02 4.07677084e-01 1.00393760e+00
-5.74737966e-01 -9.51663435e-01 -4.51670051e-01 7.17984319e-01
-1.01123464e+00 -6.27042651e-01 -4.92960185e-01 5.12330294e-01
-4.45910424e-01 6.70862854e-01 3.80726457e-01 -1.36550635e-01
7.70851374e-02 1.91748098e-01 1.42485708e-01 -8.29440176e-01
-6.60149217e-01 -1.84715033e-01 1.18529707e-01 -7.46686697e-01
-6.73487633e-02 -4.71071333e-01 -1.12557662e+00 -2.47056320e-01
1.31889418e-01 -9.19217616e-02 7.40445495e-01 8.78650904e-01
5.55897176e-01 3.28305751e-01 6.94078505e-01 -4.51537549e-01
-7.49583781e-01 -7.46971011e-01 -2.88738251e-01 3.43298376e-01
7.26293027e-01 7.25074634e-02 -3.18548292e-01 4.95408893e-01] | [9.990751266479492, 7.943849086761475] |
1603d66d-f2f6-4bcc-b088-5d92ff01ad93 | quantification-of-disaggregation-difficulty | 2101.07191 | null | https://arxiv.org/abs/2101.07191v1 | https://arxiv.org/pdf/2101.07191v1.pdf | Quantification of Disaggregation Difficulty with Respect to the Number of Meters | A promising approach toward efficient energy management is non-intrusive load monitoring (NILM), that is to extract the consumption profiles of appliances within a residence by analyzing the aggregated consumption signal. Among efficient NILM methods are event-based algorithms in which events of the aggregated signal are detected and classified in accordance with the appliances causing them. The large number of appliances and the presence of appliances with close consumption values are known to limit the performance of event-based NILM methods. To tackle these challenges, one could enhance the feature space which in turn results in extra hardware costs, installation complexity, and concerns regarding the consumer's comfort and privacy. This has led to the emergence of an alternative approach, namely semi-intrusive load monitoring (SILM), where appliances are partitioned into blocks and the consumption of each block is monitored via separate power meters. While a greater number of meters can result in more accurate disaggregation, it increases the monetary cost of load monitoring, indicating a trade-off that represents an important gap in this field. In this paper, we take a comprehensive approach to close this gap by establishing a so-called notion of "disaggregation difficulty metric (DDM)," which quantifies how difficult it is to monitor the events of any given group of appliances based on both their power values and the consumer's usage behavior. Thus, DDM in essence quantifies how much is expected to be gained in terms of disaggregation accuracy of a generic event-based algorithm by installing meters on the blocks of any partition of the appliances. Experimental results based on the REDD dataset illustrate the practicality of the proposed approach in addressing the aforementioned trade-off. | ['Sadegh Bolouki', 'Mohammad T H Beheshti', 'Elnaz Azizi'] | 2021-01-18 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [ 1.65009201e-01 1.06658451e-01 -2.35770661e-02 -2.09415689e-01
-5.77329457e-01 -6.10635459e-01 5.18804610e-01 6.47032678e-01
-1.67402670e-01 7.10688353e-01 2.91734990e-02 -8.83247256e-02
-3.67609829e-01 -1.10080075e+00 -2.88084477e-01 -9.71876323e-01
-5.24750724e-02 9.34388936e-02 -5.55796064e-02 1.15640350e-01
-1.74004316e-01 5.47335804e-01 -1.47950900e+00 -2.50177801e-01
9.91218865e-01 1.38376784e+00 -6.02576882e-02 8.60820338e-02
9.27512050e-02 5.89432955e-01 -7.58262634e-01 -1.57167375e-01
1.61136508e-01 -5.03328323e-01 -4.30441380e-01 1.68089941e-01
-4.78881687e-01 -3.71620417e-01 1.38393417e-01 1.10992467e+00
3.39414984e-01 1.99527532e-01 5.47003925e-01 -1.68017685e+00
2.94729829e-01 7.50635087e-01 -6.29486263e-01 1.09453194e-01
5.45139670e-01 6.59585297e-02 9.88067329e-01 5.86932600e-02
-3.23366001e-02 7.63340056e-01 4.17164296e-01 -4.90513928e-02
-1.61174130e+00 -6.92217469e-01 1.04893036e-01 3.68360072e-01
-1.58071136e+00 -2.64972687e-01 1.13110840e+00 -2.95141399e-01
8.33228827e-01 9.52108085e-01 6.19663954e-01 7.99915612e-01
-5.39522767e-02 5.15364766e-01 1.16685641e+00 -4.26144183e-01
8.98464859e-01 2.82845318e-01 1.77736044e-01 1.71852689e-02
6.36462092e-01 2.38268841e-02 -1.37579024e-01 -5.01542330e-01
-1.02063462e-01 9.85852778e-02 -2.84550965e-01 -2.87770867e-01
-6.81900322e-01 7.15879977e-01 8.98834243e-02 5.59740663e-01
-5.98528087e-01 -1.83503658e-01 5.11355758e-01 -2.84759831e-02
5.00119388e-01 1.93620890e-01 -3.58638376e-01 -1.60929605e-01
-1.16202486e+00 -6.59469441e-02 8.54740620e-01 6.62796199e-01
8.52385402e-01 -1.21417671e-01 1.04399063e-01 3.02158326e-01
2.30435625e-01 2.43917480e-01 4.12768155e-01 -6.43302023e-01
3.15952122e-01 1.07169163e+00 1.73638374e-01 -1.01168156e+00
-7.17175007e-01 -1.36498436e-01 -1.03658342e+00 3.00774306e-01
2.69705921e-01 -2.65524328e-01 -1.57924771e-01 1.90096533e+00
5.59533417e-01 4.29649949e-02 -1.03860155e-01 4.21008885e-01
3.96203175e-02 6.65560722e-01 3.26577485e-01 -8.45278382e-01
1.48846972e+00 -1.33779913e-01 -9.26030636e-01 2.59633690e-01
3.44649166e-01 -4.24636036e-01 6.71586812e-01 1.88352987e-01
-8.65713835e-01 -1.89511374e-01 -1.19886637e+00 4.85047907e-01
-7.28249490e-01 -3.08428556e-01 3.45381290e-01 9.33371961e-01
-5.39386153e-01 5.66749156e-01 -1.05242646e+00 -4.40633416e-01
2.27923185e-01 4.76993889e-01 9.15367156e-02 4.13466901e-01
-9.39703226e-01 7.99703300e-01 6.19763911e-01 -1.18948989e-01
-3.71950835e-01 -7.43070364e-01 -7.19788849e-01 4.15234953e-01
4.17868733e-01 -1.86173424e-01 9.11072791e-01 -5.37052274e-01
-1.30814648e+00 2.54545510e-01 1.10105099e-02 -5.68026721e-01
7.23637760e-01 1.90452352e-01 -7.69791245e-01 4.61912230e-02
5.56012392e-02 -1.58850029e-01 5.38084567e-01 -1.13653314e+00
-9.97728348e-01 -5.63899755e-01 3.35426256e-02 -1.02341071e-01
-4.18605566e-01 -1.86540142e-01 2.99882740e-01 -3.41732681e-01
-2.39274308e-01 -7.94977307e-01 -4.60032821e-02 -5.59996843e-01
-6.31730974e-01 -4.55738485e-01 1.10008574e+00 -5.46525061e-01
1.50632095e+00 -2.26711273e+00 -3.56899172e-01 4.74347264e-01
6.95026591e-02 5.14780022e-02 3.74694616e-01 6.06347442e-01
-5.14742024e-02 -1.17995583e-01 -3.05175781e-01 -4.07234162e-01
4.13956136e-01 2.61787534e-01 -3.12326699e-01 7.71581411e-01
-1.50412366e-01 6.25211358e-01 -9.34003055e-01 -3.09246987e-01
9.24782693e-01 4.84378904e-01 -3.93659174e-02 2.25228280e-01
-3.22803676e-01 4.81792450e-01 -3.76013339e-01 5.57234526e-01
6.26531303e-01 -3.08543467e-03 3.98450911e-01 -4.27859783e-01
-1.90009043e-01 3.52855980e-01 -1.47137487e+00 1.00976121e+00
-5.86205482e-01 3.00879836e-01 1.59250051e-02 -1.10651672e+00
6.03081107e-01 7.24360466e-01 9.68905449e-01 -7.05901861e-01
3.09761643e-01 1.36461407e-01 -4.61439967e-01 -2.02220589e-01
2.68716872e-01 1.41799361e-01 -4.35042650e-01 5.31494260e-01
-4.61828679e-01 3.18223566e-01 4.12538499e-01 -3.28239858e-01
1.22750568e+00 -2.38226518e-01 8.84912550e-01 -4.93494183e-01
5.93965113e-01 -3.73598009e-01 6.37866378e-01 2.69154996e-01
-2.50252098e-01 -2.47472480e-01 2.90599078e-01 -2.61428595e-01
-5.46233714e-01 -9.78604019e-01 -4.26207520e-02 7.88193524e-01
1.35011747e-01 -3.90889674e-01 -8.39519322e-01 -5.60349941e-01
-1.02206901e-01 1.26255906e+00 -4.59735513e-01 -1.76019952e-01
-4.31975156e-01 -1.18214118e+00 1.56904981e-01 1.10894248e-01
5.50966501e-01 -7.70948768e-01 -1.37936437e+00 4.97538149e-01
-4.24755901e-01 -1.11169291e+00 -1.56718761e-01 5.66547394e-01
-5.58041990e-01 -1.08379817e+00 -7.26216435e-02 -5.04321605e-03
6.89294398e-01 2.36323960e-02 1.05041504e+00 -3.84095758e-01
-1.67348504e-01 2.37402722e-01 -3.93196523e-01 -5.93314886e-01
-4.30109560e-01 2.15799332e-01 1.08991235e-01 4.90306348e-01
7.16918468e-01 -1.17305696e+00 -6.85804367e-01 3.38868856e-01
-9.52335596e-01 -3.44877541e-01 2.62904793e-01 9.84734371e-02
4.16805476e-01 1.02186465e+00 6.58728480e-01 -5.35733759e-01
5.01923442e-01 -8.00025225e-01 -1.14617205e+00 1.16566911e-01
-1.05353761e+00 -1.08723506e-01 9.79144096e-01 -3.81571442e-01
-7.91926622e-01 2.57294416e-01 1.58201069e-01 1.05245300e-01
-4.16784734e-01 -2.07919348e-02 -6.30946636e-01 2.47838333e-01
8.64972696e-02 3.41080397e-01 -5.98899841e-01 -5.55501401e-01
1.83983877e-01 6.21374011e-01 5.19302666e-01 -3.12034756e-01
8.49064469e-01 5.56991816e-01 2.39764228e-01 -8.41461658e-01
-4.47867602e-01 -4.49766725e-01 -3.24018359e-01 -1.13504931e-01
8.82645607e-01 -5.57325304e-01 -1.30018544e+00 1.96775451e-01
-8.16431999e-01 1.24331549e-01 -7.20600069e-01 2.36091226e-01
-4.13194597e-01 3.55806500e-01 -5.69356829e-02 -1.32212937e+00
-2.89727151e-01 -1.07102787e+00 8.69077682e-01 2.81169683e-01
-7.80399203e-01 -8.04283500e-01 4.05697711e-02 9.96744409e-02
3.60459536e-01 8.91218603e-01 1.03138208e+00 -6.29683137e-01
-5.07291853e-01 -5.01867533e-01 5.92735894e-02 1.54627234e-01
6.61666632e-01 -3.56159538e-01 -1.06850076e+00 -5.07151484e-01
4.91675913e-01 2.80367255e-01 3.41475047e-02 2.86957204e-01
1.02165270e+00 -5.74888051e-01 -4.17255759e-01 3.30529243e-01
1.85134649e+00 6.49227440e-01 4.47220087e-01 1.63656592e-01
3.41561466e-01 4.98039722e-01 2.75568783e-01 8.75469446e-01
4.54068422e-01 8.58624220e-01 5.64301074e-01 -1.13940835e-01
3.88418287e-01 -1.85217634e-01 3.22986066e-01 5.45497417e-01
1.06419891e-01 -3.93569946e-01 -2.94134617e-01 4.76071835e-01
-1.82967281e+00 -9.54898596e-01 1.38928667e-01 2.25291252e+00
5.82304001e-01 6.83969557e-02 5.27468503e-01 1.02528155e+00
5.79382300e-01 1.37322024e-01 -5.94021201e-01 -3.26828688e-01
1.58645272e-01 -1.72937755e-02 6.52088165e-01 3.12904954e-01
-8.07596564e-01 -1.33282825e-01 5.48379135e+00 8.82719040e-01
-8.50919724e-01 1.96713492e-01 5.95968068e-01 -8.18375647e-02
8.36387873e-02 -2.02911958e-01 -6.82270110e-01 1.22001719e+00
1.35051739e+00 -4.88234073e-01 5.34141839e-01 7.52134502e-01
7.28836179e-01 -5.96701562e-01 -1.43241286e+00 9.89013314e-01
-3.51418704e-01 -7.01890409e-01 -4.75728363e-01 5.36955178e-01
5.41646481e-01 -2.35676095e-01 -3.76692563e-01 -4.89396639e-02
4.38989609e-01 -5.67655325e-01 5.40135026e-01 3.18186104e-01
2.40514636e-01 -1.06067502e+00 6.63960218e-01 5.11297882e-01
-1.62457585e+00 -2.72684038e-01 5.24647012e-02 -7.84998611e-02
4.05268312e-01 1.04468715e+00 -6.22693598e-01 7.74236023e-01
7.09188104e-01 -7.13968873e-02 -2.70377010e-01 7.19632208e-01
6.08808920e-02 6.52464569e-01 -8.31451714e-01 8.81523341e-02
-2.07375184e-01 -5.10388315e-01 3.66961747e-01 1.00792456e+00
3.68589908e-01 4.48624790e-02 1.87232941e-01 8.91979098e-01
9.12363082e-02 -4.44942638e-02 -4.19789761e-01 3.23996276e-01
6.54416442e-01 1.47326028e+00 -8.49149466e-01 -3.15344155e-01
-4.75172371e-01 8.80334973e-01 -1.26641259e-01 1.70980319e-01
-8.68671894e-01 -1.64385617e-01 6.17552519e-01 2.15465695e-01
2.91823357e-01 1.15927905e-01 -3.81534100e-01 -7.29707181e-01
2.44541094e-01 -5.67683995e-01 6.31178975e-01 -1.28070071e-01
-1.23768294e+00 3.46303940e-01 2.69776523e-01 -1.22128415e+00
-4.98741925e-01 1.08897962e-01 -6.27822816e-01 4.94450748e-01
-1.47593570e+00 -8.42147231e-01 -4.18514013e-01 6.77242160e-01
3.47214699e-01 4.68167096e-01 9.69287276e-01 5.36376417e-01
-6.91581666e-01 6.03439629e-01 2.84262359e-01 -1.50006294e-01
-1.33228332e-01 -1.29225242e+00 -2.46955454e-02 8.67829263e-01
-4.14664075e-02 -1.93073213e-01 9.80330467e-01 -3.90989542e-01
-1.15619636e+00 -1.22762668e+00 8.69578600e-01 -3.66388291e-01
5.94475389e-01 -5.36346555e-01 -5.80293298e-01 5.03448904e-01
1.65154859e-01 -2.40579650e-01 8.60004365e-01 -2.91426808e-01
1.06080003e-01 -5.71422100e-01 -1.72388673e+00 3.47985178e-01
5.61045110e-01 -4.53888446e-01 -4.28369254e-01 1.47920653e-01
3.69211733e-01 3.64611387e-01 -1.14925945e+00 3.49597037e-01
1.71920180e-01 -1.11475110e+00 6.51524484e-01 2.82745212e-01
-3.96686554e-01 -6.15964890e-01 -5.17039835e-01 -1.23527336e+00
-1.54319167e-01 -8.06415141e-01 -5.58437407e-01 1.85655475e+00
-8.97311699e-03 -9.32026744e-01 5.01789033e-01 9.06386316e-01
4.68948275e-01 -4.96184349e-01 -1.32195735e+00 -9.15367723e-01
-3.90801430e-01 -4.35264796e-01 1.28593707e+00 8.94459426e-01
3.29238623e-01 9.93163735e-02 -1.30039454e-01 5.17483354e-01
9.04316068e-01 2.02101201e-01 4.41583961e-01 -1.30752254e+00
-1.52530208e-01 -3.35407436e-01 -5.82560480e-01 -3.51237923e-01
-4.13008817e-02 -4.37560230e-01 -1.77013114e-01 -1.14995658e+00
1.93781331e-02 -2.00613081e-01 -4.80293244e-01 3.72298837e-01
1.78837627e-01 1.43016726e-01 1.03661947e-01 5.97397722e-02
-4.67279404e-01 3.45487535e-01 9.85024944e-02 -1.88128620e-01
-4.64145988e-01 2.75986314e-01 -4.72895265e-01 8.87806654e-01
9.16683078e-01 -4.27584767e-01 -4.80229467e-01 2.75670081e-01
2.87518073e-02 -6.20078817e-02 2.57996380e-01 -1.22035611e+00
1.76106393e-01 -1.77692801e-01 2.41443291e-01 -8.21189940e-01
2.76089460e-01 -1.69508696e+00 8.97092223e-01 5.76002538e-01
9.81499776e-02 1.61231458e-01 -4.45761047e-02 4.49143231e-01
3.93343484e-03 4.15704809e-02 8.13981235e-01 -3.26337181e-02
-3.64271849e-01 -1.99319255e-02 -4.65215266e-01 -4.32455957e-01
1.51736248e+00 -1.99009374e-01 -6.16090782e-02 -1.94288209e-01
-5.04707336e-01 1.43088505e-01 3.68192852e-01 1.24321118e-01
-2.78279543e-01 -1.31944740e+00 -2.50679374e-01 8.60733315e-02
4.90485318e-02 -1.59326091e-01 1.10615529e-02 7.75212586e-01
1.26842409e-01 3.91608804e-01 2.32864857e-01 -4.39621925e-01
-1.14383340e+00 8.48416388e-01 2.40790978e-01 -6.52759612e-01
-5.57748377e-01 -1.57564208e-01 9.60472822e-02 2.50166804e-01
2.39457026e-01 -6.08839631e-01 4.12493050e-02 5.85888684e-01
5.64284265e-01 1.11960733e+00 3.05132270e-01 -5.89414775e-01
-5.30897319e-01 3.58214855e-01 2.61641562e-01 2.58575529e-01
1.21292019e+00 -6.29263282e-01 -1.44434750e-01 5.04388869e-01
1.28835201e+00 7.62447417e-02 -1.12110794e+00 1.17005883e-02
3.40946525e-01 -7.39774257e-02 6.74036741e-02 -8.13592136e-01
-1.10491812e+00 1.41222745e-01 8.76044810e-01 1.09971285e+00
1.77148938e+00 -7.82835260e-02 1.03214145e+00 -8.24914500e-02
6.26326323e-01 -1.37787879e+00 -7.05274045e-01 -4.72391039e-01
2.83279926e-01 -9.96030092e-01 5.53743467e-02 -3.93990576e-01
1.60009623e-01 6.04083598e-01 5.17310128e-02 1.52135447e-01
7.39996433e-01 4.51157242e-01 -3.03759247e-01 -1.52720317e-01
-2.28721410e-01 -1.50473699e-01 -2.44114876e-01 4.31400627e-01
-1.95208371e-01 6.44393981e-01 -3.92457038e-01 6.70207322e-01
-3.20069820e-01 5.34182191e-02 2.71259546e-01 9.93226349e-01
-1.32263675e-01 -9.12478745e-01 -5.59461832e-01 4.49434251e-01
-4.73788291e-01 4.43364590e-01 2.37646438e-02 6.18132591e-01
4.45115298e-01 1.56513798e+00 4.26653586e-03 -2.33331889e-01
5.86646676e-01 1.14817366e-01 4.06401344e-02 -3.63752828e-03
-3.50309938e-01 2.50800163e-03 -3.50715108e-02 -6.62346065e-01
-4.52314258e-01 -6.15048826e-01 -1.00121927e+00 -5.87215483e-01
-3.49326491e-01 2.50284940e-01 6.46372736e-01 1.11251020e+00
2.16606811e-01 5.83110511e-01 1.29086602e+00 -7.76240885e-01
-6.95937812e-01 -7.44640768e-01 -9.00783956e-01 5.48139215e-01
4.41231519e-01 -4.81082708e-01 -7.29410172e-01 -1.03740737e-01] | [5.97746467590332, 2.581223249435425] |
be4ca019-a2c7-4d91-b8e2-419445a6201f | modified-splice-and-its-extension-to-non | 1307.4048 | null | http://arxiv.org/abs/1307.4048v1 | http://arxiv.org/pdf/1307.4048v1.pdf | Modified SPLICE and its Extension to Non-Stereo Data for Noise Robust Speech Recognition | In this paper, a modification to the training process of the popular SPLICE
algorithm has been proposed for noise robust speech recognition. The
modification is based on feature correlations, and enables this stereo-based
algorithm to improve the performance in all noise conditions, especially in
unseen cases. Further, the modified framework is extended to work for
non-stereo datasets where clean and noisy training utterances, but not stereo
counterparts, are required. Finally, an MLLR-based computationally efficient
run-time noise adaptation method in SPLICE framework has been proposed. The
modified SPLICE shows 8.6% absolute improvement over SPLICE in Test C of
Aurora-2 database, and 2.93% overall. Non-stereo method shows 10.37% and 6.93%
absolute improvements over Aurora-2 and Aurora-4 baseline models respectively.
Run-time adaptation shows 9.89% absolute improvement in modified framework as
compared to SPLICE for Test C, and 4.96% overall w.r.t. standard MLLR
adaptation on HMMs. | ['D. S. Pavan Kumar', 'S. Umesh', 'N. Vishnu Prasad', 'Vikas Joshi'] | 2013-07-15 | null | null | null | null | ['robust-speech-recognition'] | ['speech'] | [ 7.65056685e-02 -6.66854950e-03 6.00708485e-01 -6.60175622e-01
-1.82000303e+00 -7.18270361e-01 6.85804784e-01 -5.72544694e-01
-4.87577647e-01 6.20541811e-01 5.26114881e-01 -3.04993868e-01
2.00127527e-01 -3.45445514e-01 -5.28672099e-01 -1.05705631e+00
2.02799425e-01 1.47020057e-01 2.17468798e-01 -1.83923423e-01
-4.84610237e-02 5.02259970e-01 -1.55140448e+00 3.25627118e-01
3.14154416e-01 8.44887495e-01 2.34188646e-01 1.23057044e+00
2.14366108e-01 1.18499890e-01 -7.33575702e-01 -1.39217198e-01
4.13225830e-01 -4.29259896e-01 -7.28862822e-01 1.06353007e-01
9.18077469e-01 2.06598043e-01 -6.24721229e-01 8.14569890e-01
1.27824378e+00 7.29622185e-01 3.81025463e-01 -3.16417694e-01
-2.15295237e-02 4.00779396e-01 -3.94828111e-01 4.19351518e-01
3.71725589e-01 -7.18110725e-02 6.82244956e-01 -1.07901454e+00
2.34904766e-01 1.45852220e+00 7.47217476e-01 6.55002058e-01
-1.09049571e+00 -6.91869318e-01 -2.63298213e-01 2.50667870e-01
-1.38155770e+00 -8.79900038e-01 3.00856113e-01 5.15235439e-02
1.52840805e+00 8.40171695e-01 4.18885529e-01 1.30639923e+00
-1.75448567e-01 6.99540138e-01 1.56940961e+00 -6.73514605e-01
3.28076363e-01 -2.13643894e-01 1.06194682e-01 6.92179576e-02
-2.68408090e-01 4.39494550e-01 -8.23687315e-01 -9.59349331e-03
1.75727114e-01 -8.12626004e-01 -2.21138254e-01 1.93409026e-01
-6.84417844e-01 3.25471818e-01 -3.12711648e-03 2.81829238e-01
1.28470913e-01 -2.08839238e-01 6.51405752e-01 3.82510662e-01
4.82107222e-01 4.32430627e-03 -5.53927243e-01 -8.34922552e-01
-1.15027952e+00 1.78268880e-01 8.14890921e-01 9.67680216e-01
4.52108204e-01 7.17091978e-01 -9.94282812e-02 1.53390551e+00
4.58867997e-01 1.00840855e+00 5.46078801e-01 -7.58314431e-01
4.97700363e-01 -2.14615792e-01 -2.50688851e-01 -3.91616747e-02
-3.55898649e-01 -1.03196931e+00 -6.71689153e-01 4.35471069e-03
-2.20295355e-01 -2.42935531e-02 -1.47916532e+00 1.61538076e+00
2.86248982e-01 3.51407707e-01 6.03189707e-01 5.85540831e-01
7.73363411e-01 8.94824386e-01 -2.78476179e-01 -4.19811040e-01
9.54066515e-01 -9.31767344e-01 -8.90985072e-01 -2.23855466e-01
5.35008669e-01 -1.60100794e+00 1.01296020e+00 8.37558329e-01
-1.06023419e+00 -7.16208756e-01 -1.15289927e+00 1.26488596e-01
-2.23500267e-01 -1.44874919e-02 -1.03162274e-01 1.28241467e+00
-1.28116906e+00 7.96997473e-02 -6.07840300e-01 -6.83950305e-01
-3.22695613e-01 4.41594064e-01 -3.05758774e-01 -1.07526556e-01
-9.54727054e-01 7.17016459e-01 2.29348511e-01 3.13666373e-01
-1.02385116e+00 -1.38452267e-02 -8.19878876e-01 -3.46429080e-01
2.36055076e-01 -2.87391394e-01 1.88569880e+00 -4.45704877e-01
-1.99535811e+00 5.42882442e-01 -7.00537622e-01 -6.96711302e-01
3.20358783e-01 -5.56688249e-01 -1.05310011e+00 -2.08960518e-01
-1.52144358e-01 3.65766376e-01 5.46033978e-01 -1.18893135e+00
-6.92229033e-01 -3.67885262e-01 -4.49458152e-01 6.68683708e-01
2.68684655e-01 2.20698282e-01 -8.19230735e-01 -7.66027570e-01
5.00469744e-01 -1.05666387e+00 -3.28463912e-02 -8.12450469e-01
-2.99489200e-01 1.75407782e-01 1.34389138e+00 -9.50147629e-01
1.02627945e+00 -2.24589157e+00 -2.03023434e-01 3.18613976e-01
-6.98092163e-01 7.91708589e-01 -2.54697800e-02 4.39615041e-01
-1.54652521e-01 -2.14696407e-01 -3.11903358e-01 -5.37885964e-01
-2.88709641e-01 6.46574438e-01 -1.04899958e-01 3.17369998e-01
-3.01726729e-01 1.65289059e-01 -3.10991585e-01 -1.24177761e-01
4.63661879e-01 5.95074236e-01 -3.68555248e-01 2.86654741e-01
4.51949328e-01 3.84823799e-01 6.02112599e-02 6.56193197e-01
1.08591759e+00 1.04689610e+00 -1.55893803e-01 -2.79881358e-01
-3.58457923e-01 5.34324944e-01 -1.43358815e+00 1.61139727e+00
-8.77097845e-01 8.19988787e-01 2.96616584e-01 -9.03839409e-01
1.27140474e+00 8.15378785e-01 -3.46642286e-01 -8.84555459e-01
-2.62515426e-01 4.96215314e-01 4.89214659e-02 -5.42498052e-01
4.16502208e-01 -3.22551101e-01 1.12124376e-01 -1.59935534e-01
2.94547737e-01 -5.26679635e-01 -2.01602075e-02 1.07549913e-02
1.01817584e+00 1.78167015e-01 3.85298848e-01 -2.78010070e-01
9.52820241e-01 -7.02953696e-01 7.39015877e-01 9.14048135e-01
-4.31599468e-01 1.05134833e+00 -2.86171466e-01 -8.16814750e-02
-9.21468437e-01 -1.22450435e+00 -3.77667010e-01 8.48813474e-01
-3.78406852e-01 -4.45239216e-01 -8.50647569e-01 -6.04109228e-01
-5.59897482e-01 9.72976387e-01 -1.90610990e-01 6.63252473e-02
-7.51481652e-01 -8.13266456e-01 9.51045036e-01 5.54843783e-01
7.63884604e-01 -7.26689696e-01 2.37119496e-01 2.89829165e-01
-3.26297492e-01 -1.25360167e+00 -3.95526856e-01 6.55306518e-01
-7.82309234e-01 -4.48527396e-01 -3.54198873e-01 -8.16566885e-01
1.63607910e-01 4.56237137e-01 6.69672906e-01 -4.24180716e-01
-1.11201540e-01 4.25538838e-01 -6.19596601e-01 -3.49724740e-01
-4.79077667e-01 -3.15518171e-01 3.95649105e-01 -4.70374748e-02
4.48352277e-01 -5.66379786e-01 -3.07050914e-01 6.39494359e-01
-7.27349222e-01 -3.92063558e-01 7.03918457e-01 1.13450551e+00
6.31962359e-01 -3.43215317e-01 4.89139557e-01 -4.74375397e-01
1.23106971e-01 6.74015209e-02 -5.18525481e-01 2.65713446e-02
-4.90561992e-01 8.21628720e-02 3.72873813e-01 1.23531274e-01
-1.61761022e+00 2.32268080e-01 -9.68421280e-01 1.27363548e-01
-5.15155673e-01 1.21575341e-01 -4.72461849e-01 -9.16941166e-02
8.50725055e-01 4.10290271e-01 -3.62702906e-01 -1.01428807e+00
1.44123852e-01 1.55145311e+00 8.25165808e-01 -3.69816035e-01
8.72100890e-01 2.37741888e-01 -2.12415516e-01 -1.43626201e+00
-7.19381034e-01 -1.18272758e+00 -5.79473853e-01 -7.23546371e-02
5.01915514e-01 -1.09592867e+00 1.79001540e-02 6.44977927e-01
-7.90095329e-01 -8.80583450e-02 -1.75280705e-01 7.97703683e-01
-7.37908006e-01 5.32276690e-01 -3.74370515e-01 -1.24437189e+00
-4.42829549e-01 -1.40009248e+00 1.13890183e+00 1.30260438e-01
-4.14731503e-02 -5.25291502e-01 3.11104476e-01 7.82808304e-01
3.50493133e-01 -4.34243321e-01 2.72675186e-01 -8.14406276e-01
-1.46696955e-01 -1.89147145e-01 6.97397068e-02 9.79316831e-01
1.65339574e-01 -5.33866212e-02 -1.66197467e+00 -4.98712689e-01
2.40932375e-01 -1.01652727e-01 8.57569039e-01 3.38534147e-01
7.44602084e-01 -1.09255679e-01 -6.30270913e-02 6.91361308e-01
1.22651505e+00 7.86059439e-01 7.55407393e-01 3.96521419e-01
1.97964832e-01 9.14967060e-02 6.49722934e-01 1.55232206e-01
-2.70977467e-01 1.10610747e+00 5.41958213e-02 -3.16066705e-02
-6.98296189e-01 -5.58364391e-03 7.12691367e-01 1.44349408e+00
1.72293261e-01 -2.21550241e-01 -9.20854032e-01 6.27509475e-01
-1.42274284e+00 -9.40171361e-01 -3.62289280e-01 2.44748664e+00
7.06865311e-01 2.06670418e-01 -5.30728735e-02 5.23052812e-01
6.41123533e-01 5.41146994e-02 -1.02372341e-01 -9.77477312e-01
-5.39798319e-01 5.63404441e-01 4.42352146e-01 1.03406084e+00
-1.15065777e+00 7.80119956e-01 6.27368498e+00 1.18010724e+00
-9.13028181e-01 3.18056881e-01 4.51765209e-01 -1.61731929e-01
1.21742733e-01 1.53282940e-01 -9.98889387e-01 -3.82393934e-02
1.48805320e+00 1.56076208e-01 3.04404378e-01 7.36692309e-01
7.26237833e-01 -3.02663207e-01 -5.75286984e-01 1.20102251e+00
3.47686976e-01 -6.67920053e-01 -1.99006930e-01 -4.23219614e-02
7.64306068e-01 3.24494332e-01 9.46256891e-03 4.27332848e-01
-1.28571335e-02 -7.85512686e-01 6.36212587e-01 4.19913605e-02
5.64916909e-01 -1.04186618e+00 1.17054892e+00 1.12708621e-01
-1.18742096e+00 1.83875591e-01 -3.87651950e-01 2.01318964e-01
2.11231202e-01 4.29027706e-01 -1.17318773e+00 7.78129935e-01
1.13051033e+00 9.61203203e-02 -4.36366975e-01 1.35521901e+00
-1.42361537e-01 1.31312060e+00 -6.19096696e-01 2.37923607e-01
4.39123988e-01 8.64998158e-03 9.56801414e-01 1.88673913e+00
6.17640376e-01 -1.80638880e-01 -2.81936646e-01 -3.07406545e-01
2.85929799e-01 2.81068057e-01 -7.04815924e-01 6.59151793e-01
3.92201602e-01 1.19743419e+00 -1.53257638e-01 -3.25951099e-01
-4.63710666e-01 9.34083521e-01 -8.28759521e-02 5.20639002e-01
-4.97222602e-01 -4.67022151e-01 4.98062819e-01 -3.07499200e-01
5.54219067e-01 -3.58437985e-01 -1.20672174e-01 -6.30172431e-01
9.50111970e-02 -1.36330974e+00 3.78445685e-01 -9.08564806e-01
-8.32498252e-01 9.73482370e-01 -1.23883471e-01 -1.24155343e+00
-3.35735768e-01 -5.53973734e-01 -6.26941264e-01 1.00962818e+00
-9.25487399e-01 -1.16668344e+00 1.14246262e-02 3.30168515e-01
1.18257570e+00 -4.71687675e-01 1.03672552e+00 3.07974666e-01
-6.26959264e-01 9.21443105e-01 7.54142880e-01 -3.18599761e-01
8.90548825e-01 -1.34086442e+00 6.67131543e-01 1.39081454e+00
6.09071016e-01 5.07332385e-01 1.08187151e+00 -3.61835539e-01
-1.19613957e+00 -1.00377190e+00 9.74912524e-01 -9.06256288e-02
3.03402871e-01 -5.73067546e-01 -7.81560779e-01 5.71767390e-01
5.98025918e-01 -1.84206665e-01 7.81547010e-01 2.51124203e-01
-4.95570987e-01 -4.93964076e-01 -1.19596803e+00 5.70443332e-01
9.94266212e-01 -8.04907382e-01 -7.23366141e-01 1.91344053e-01
4.19009209e-01 -6.27173126e-01 -5.83554029e-01 5.99541128e-01
3.89348298e-01 -1.09177232e+00 7.22837448e-01 -5.92759922e-02
-7.86823273e-01 -5.91333508e-01 -9.17241693e-01 -1.26342094e+00
1.15409799e-01 -1.31557930e+00 1.60649195e-01 1.47873867e+00
7.72179723e-01 -4.85532701e-01 5.13329387e-01 -4.91750613e-02
-8.53048921e-01 -3.10812026e-01 -1.59316063e+00 -1.13699520e+00
-2.15164319e-01 -1.18360794e+00 1.39758810e-01 1.64141968e-01
-4.24236178e-01 3.80661219e-01 -5.37313521e-01 3.80936354e-01
4.81327116e-01 -3.77191067e-01 8.46061647e-01 -5.25740027e-01
-6.96559012e-01 3.01029563e-01 -7.33345389e-01 -1.20613241e+00
-2.92968839e-01 -5.08847892e-01 3.98456216e-01 -1.12270892e+00
6.08655550e-02 -1.97489578e-02 -3.57411534e-01 2.13441923e-01
4.44929153e-02 5.13761520e-01 -1.54994905e-01 -1.00179605e-01
-1.70888096e-01 5.71331263e-01 6.73099637e-01 -1.25698328e-01
-8.62324759e-02 5.76476872e-01 -2.47495115e-01 5.18165112e-01
1.12836397e+00 -4.12744194e-01 -3.76620352e-01 -2.30725542e-01
-3.74417663e-01 -9.75033343e-02 6.92096818e-03 -1.42344642e+00
1.94933079e-02 3.83624256e-01 1.59502953e-01 -9.56621289e-01
8.14938486e-01 -5.43705285e-01 1.71484388e-02 2.14207619e-01
-2.00239979e-02 1.67150766e-01 5.80246806e-01 7.34294653e-01
-7.15642214e-01 -4.14467275e-01 1.06442499e+00 1.30864039e-01
-6.56776428e-01 -3.26356590e-01 -7.51945078e-01 -6.83848262e-02
5.79805970e-01 -4.82617110e-01 3.85037102e-02 -6.10294521e-01
-1.25600076e+00 -4.11208302e-01 -1.21035211e-01 4.60722804e-01
3.97681832e-01 -1.13635182e+00 -7.74346352e-01 3.10613036e-01
3.81829701e-02 -2.32285976e-01 4.33038682e-01 7.72493958e-01
-3.58229011e-01 6.24951005e-01 3.52229476e-01 -7.06413567e-01
-1.96699727e+00 -9.79210064e-02 5.46182752e-01 2.11893588e-01
-4.27480102e-01 1.18286431e+00 4.78579365e-02 -6.84638262e-01
5.68709373e-01 -3.14761966e-01 1.65407762e-01 -4.42103922e-01
3.89197320e-01 5.40174365e-01 8.92307281e-01 -1.06386650e+00
-3.21404427e-01 6.70286775e-01 -1.72105655e-01 -9.00066793e-01
1.13593113e+00 -3.75009745e-01 2.11178809e-01 5.67304671e-01
1.14431548e+00 4.89299148e-01 -8.66485536e-01 -3.16294372e-01
-1.96604114e-02 -3.80995095e-01 2.90545970e-01 -1.22578502e+00
-6.08440638e-01 1.01421022e+00 1.37662339e+00 -2.09219560e-01
1.42205834e+00 -2.46466711e-01 4.46152270e-01 6.65631354e-01
1.18658252e-01 -1.32344615e+00 -3.11249733e-01 7.32627153e-01
1.08488309e+00 -8.57449472e-01 -3.21016610e-01 -4.22734559e-01
-3.84041935e-01 1.13427186e+00 6.68148100e-01 5.54443598e-01
4.67686504e-01 6.33631647e-01 6.37959421e-01 3.08465213e-01
-7.16976285e-01 -3.19174409e-01 2.57204741e-01 8.48932862e-01
7.13740110e-01 1.29707828e-01 -3.95033509e-01 1.08572312e-01
-4.53953922e-01 -7.02163577e-01 3.87373626e-01 9.74820077e-01
-6.23706579e-01 -1.46471024e+00 -7.33274996e-01 8.59853402e-02
-6.15548968e-01 -3.10212493e-01 -4.11928684e-01 7.16773391e-01
-3.18536200e-02 1.36109424e+00 -3.24864239e-01 -4.73990738e-01
6.02627635e-01 5.25987387e-01 1.37126938e-01 -5.13607144e-01
-6.98102951e-01 8.35514486e-01 7.46331751e-01 -3.59218150e-01
-3.00246119e-01 -9.44500685e-01 -9.99929309e-01 1.49819121e-01
-6.17184281e-01 4.56564218e-01 9.62566793e-01 9.32087660e-01
1.75604805e-01 2.86340058e-01 7.02833474e-01 -3.70971501e-01
-6.58966064e-01 -1.41241145e+00 -4.47379619e-01 1.29409477e-01
2.79305905e-01 -3.43072534e-01 -5.79703331e-01 -1.53398998e-02] | [14.83006763458252, 5.941051959991455] |
d2b36af0-5f2e-45f9-9876-ec9d7bc6bc26 | single-image-reflection-removal-with-1 | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zheng_Single_Image_Reflection_Removal_With_Absorption_Effect_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zheng_Single_Image_Reflection_Removal_With_Absorption_Effect_CVPR_2021_paper.pdf | Single Image Reflection Removal With Absorption Effect | In this paper, we consider the absorption effect for the problem of single image reflection removal. We show that the absorption effect can be numerically approximated by the average of refractive amplitude coefficient map. We then reformulate the image formation model and propose a two-step solution that explicitly takes the absorption effect into account. The first step estimates the absorption effect from a reflection-contaminated image, while the second step recovers the transmission image by taking a reflection-contaminated image and the estimated absorption effect as the input. Experimental results on four public datasets show that our two-step solution not only successfully removes reflection artifact, but also faithfully restores the intensity distortion caused by the absorption effect. Our ablation studies further demonstrate that our method achieves superior performance on the recovery of overall intensity and has good model generalization capacity. The code is available at https://github.com/q-zh/absorption. | ['Alex C. Kot', 'Ling-Yu Duan', 'Xudong Jiang', 'Jinnan Chen', 'Boxin Shi', 'Qian Zheng'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['reflection-removal'] | ['computer-vision'] | [ 8.48084688e-01 -1.32214561e-01 4.09844905e-01 -1.63480252e-01
-8.87629747e-01 -3.36974710e-01 1.72977760e-01 -4.28561121e-01
-3.22946250e-01 5.47200143e-01 1.63169857e-02 -1.95634156e-01
1.62358195e-01 -8.27994645e-01 -8.74857783e-01 -1.24163508e+00
2.92606115e-01 -2.40010377e-02 5.66214584e-02 2.90585905e-02
3.33578080e-01 3.50681633e-01 -1.30233908e+00 1.71918541e-01
1.34894550e+00 9.11345363e-01 3.49888891e-01 6.23054564e-01
8.61808881e-02 9.73825872e-01 -6.65200710e-01 -4.62813443e-03
5.52600324e-01 -5.76009631e-01 -5.85159957e-01 2.85316080e-01
4.31552261e-01 -7.81633317e-01 -3.99922371e-01 1.16906428e+00
3.92220557e-01 -2.26336699e-02 5.74861050e-01 -8.17592919e-01
-7.42753685e-01 7.36004440e-03 -1.03299451e+00 1.55395940e-02
1.44823074e-01 2.13922903e-01 5.07585526e-01 -8.00664783e-01
4.61034417e-01 7.58386314e-01 5.33797979e-01 4.37161237e-01
-1.31282735e+00 -8.09749126e-01 -2.98731059e-01 5.96684776e-02
-1.41599035e+00 -3.64321768e-01 7.17294872e-01 -2.11639151e-01
7.13732719e-01 2.28722125e-01 5.72106481e-01 3.27531695e-01
1.97369978e-01 6.81804001e-01 1.37813878e+00 -5.91959178e-01
-2.48313457e-01 5.15831970e-02 2.39838839e-01 7.38459468e-01
2.09505096e-01 1.73012599e-01 -1.41011164e-01 -3.77940871e-02
7.56887734e-01 6.61196560e-02 -8.71759593e-01 1.35958180e-01
-7.63499856e-01 3.32906157e-01 6.05899513e-01 -5.13216946e-04
-3.37893486e-01 -5.91158075e-03 -3.33071679e-01 3.65879595e-01
7.31376410e-01 4.43429351e-01 -7.82149136e-02 4.13824737e-01
-8.69877577e-01 8.06080624e-02 3.98702264e-01 6.96837842e-01
1.10716772e+00 8.90037511e-03 -9.22386050e-02 1.02222967e+00
1.14179924e-02 1.07522225e+00 -2.84231827e-02 -1.15136743e+00
2.21218154e-01 3.36146295e-01 4.12497550e-01 -4.79679406e-01
-3.04018706e-01 -2.96869874e-01 -9.60921884e-01 4.02610093e-01
5.14065564e-01 -1.17167942e-01 -1.19744349e+00 1.62558186e+00
2.64459223e-01 3.81714314e-01 1.99314415e-01 8.71658921e-01
9.18990612e-01 8.28644395e-01 -3.82604659e-01 -5.52164078e-01
1.17954540e+00 -1.07464981e+00 -7.35666573e-01 -1.24363720e-01
2.66447872e-01 -9.80580330e-01 6.49564028e-01 5.20020843e-01
-1.47890961e+00 -3.02708268e-01 -1.06361496e+00 -2.86554366e-01
2.48320952e-01 3.15367907e-01 2.87852526e-01 3.87474179e-01
-1.00755203e+00 5.46102583e-01 -5.53939342e-01 1.77583508e-02
3.83544326e-01 2.49644265e-01 -9.20248777e-02 -5.25620282e-01
-8.49463344e-01 7.04285741e-01 -1.35961652e-01 1.76115900e-01
-5.86260021e-01 -9.76625741e-01 -4.69606519e-01 -7.51906708e-02
2.47000426e-01 -7.06842780e-01 1.02411330e+00 -1.18561089e+00
-1.47859967e+00 8.67773592e-01 -6.66118324e-01 -2.01514184e-01
4.71906811e-01 -2.11880907e-01 -6.75027892e-02 4.33925807e-01
-2.95659870e-01 1.91392183e-01 9.39219356e-01 -1.66102612e+00
-3.02121341e-01 -2.47661218e-01 -9.02953967e-02 3.32830697e-01
-5.24791256e-02 -1.34103848e-02 -7.14838326e-01 -1.83172494e-01
2.65634567e-01 -8.97311270e-01 2.02204064e-02 4.29921448e-02
-6.10360444e-01 5.76502025e-01 5.34786701e-01 -8.08250248e-01
9.58681464e-01 -2.18494296e+00 -7.41571710e-02 2.29887530e-01
3.25626522e-01 4.49979424e-01 -3.94403219e-01 5.45757473e-01
-1.93428308e-01 -8.69918019e-02 -7.95896232e-01 -4.67326432e-01
-6.47701800e-01 -2.44453773e-01 -3.74966681e-01 5.21527350e-01
-1.86462831e-02 8.14406991e-01 -5.35294175e-01 -1.55929655e-01
2.23077163e-01 8.92761827e-01 -4.31975096e-01 4.22535062e-01
1.18155688e-01 7.09374249e-01 -1.83754757e-01 4.58577573e-01
1.48430574e+00 -4.19928253e-01 2.53444910e-01 -6.84937716e-01
-2.49485433e-01 -6.98220432e-02 -8.11080992e-01 1.09419179e+00
-3.93406659e-01 4.49636281e-01 3.06413114e-01 -6.32338881e-01
6.80722237e-01 8.45205188e-02 6.78162932e-01 -1.06260359e+00
4.03694808e-02 1.78019255e-01 -1.26861809e-02 -4.88388747e-01
2.14326963e-01 -3.75121862e-01 6.13690495e-01 6.06158435e-01
-3.34169507e-01 -4.00233924e-01 -1.38636455e-01 3.00851643e-01
9.22533154e-01 8.07631910e-02 8.73569772e-02 1.70418937e-02
4.37343001e-01 -1.01926975e-01 3.78544748e-01 7.47802138e-01
6.17924668e-02 1.05177689e+00 1.88428625e-01 -1.63481116e-01
-1.02617741e+00 -1.17617595e+00 -1.31467700e-01 4.01310384e-01
5.42364061e-01 -8.80339518e-02 -9.55197871e-01 -1.21755883e-01
-1.13164127e-01 5.46240151e-01 -4.22498316e-01 -2.20192745e-01
-5.36804438e-01 -1.28021955e+00 1.98533237e-01 1.06884828e-02
9.86347556e-01 -8.71229708e-01 -2.54076958e-01 -2.28596300e-01
-7.74134696e-01 -1.12001503e+00 -3.37360770e-01 -2.55913794e-01
-6.46636486e-01 -1.20463192e+00 -8.24105322e-01 -4.87646967e-01
1.01593852e+00 9.41956460e-01 8.93561721e-01 8.31373334e-01
-6.07936561e-01 4.53739971e-01 -8.62899125e-02 -4.44522619e-01
-3.69188666e-01 -6.26724660e-01 -4.33169514e-01 1.84677079e-01
3.61299217e-02 -4.62998182e-01 -1.01446044e+00 3.01119387e-01
-1.01933897e+00 4.25413460e-01 6.15791559e-01 5.94142020e-01
6.77948594e-01 2.15050295e-01 1.99905351e-01 -1.01373470e+00
3.42786998e-01 -9.54423621e-02 -7.32571304e-01 1.01707503e-01
-6.74406767e-01 -1.45254031e-01 4.20809060e-01 -2.33579695e-01
-1.58366215e+00 -7.46722659e-03 -1.92948997e-01 -3.04955006e-01
1.25172073e-02 1.91188127e-01 7.33172670e-02 -2.88086891e-01
4.47368026e-01 4.64474618e-01 2.78745890e-01 -4.73435670e-01
8.20894688e-02 5.24400413e-01 5.52578628e-01 -3.75538588e-01
9.82847750e-01 1.01219797e+00 1.84355974e-01 -1.19401503e+00
-9.64079142e-01 -5.14779866e-01 -2.98297077e-01 -2.43225098e-01
6.06749892e-01 -9.53410566e-01 -6.86216176e-01 1.00419271e+00
-1.13555908e+00 -7.35657513e-01 -8.30116719e-02 3.48813772e-01
-3.34645152e-01 6.26280904e-01 -7.51486301e-01 -8.52121055e-01
-5.64345956e-01 -1.01745307e+00 8.79492640e-01 3.07741046e-01
5.59234560e-01 -7.68584788e-01 2.92369351e-02 6.14453137e-01
4.74387676e-01 4.61491421e-02 8.90889347e-01 2.41565123e-01
-1.04948461e+00 -6.50504651e-03 -7.07133889e-01 5.63664675e-01
2.76059955e-01 -1.71453375e-02 -1.29940701e+00 -3.49154085e-01
3.23599756e-01 -4.52229232e-02 1.31339169e+00 8.07495117e-01
1.03502142e+00 -5.76517992e-02 -1.14381894e-01 1.00700307e+00
1.78319812e+00 -2.70271227e-02 1.33370614e+00 7.00043812e-02
7.18572617e-01 5.73155165e-01 5.75753331e-01 2.12612361e-01
1.37603387e-01 5.71012497e-01 5.00390232e-01 -7.45732725e-01
-5.95372975e-01 1.90729603e-01 1.75619036e-01 5.83379626e-01
-4.84903872e-01 -4.22304213e-01 -5.66303492e-01 2.57737219e-01
-1.62924373e+00 -1.13948333e+00 -6.12577438e-01 2.52094269e+00
8.68875623e-01 -3.23611915e-01 -1.35198578e-01 -1.55445784e-01
4.51331258e-01 -1.12056546e-02 -5.93658924e-01 3.16093909e-03
-2.29546636e-01 2.65169650e-01 7.86709785e-01 1.08036768e+00
-7.25407839e-01 6.07237339e-01 7.01517296e+00 4.35814649e-01
-1.15721393e+00 4.90108840e-02 4.91559595e-01 -1.56303242e-01
-4.48752522e-01 -1.92372084e-01 -4.90735024e-01 3.51812690e-01
6.54346824e-01 9.48537607e-03 6.45767212e-01 -9.70506668e-02
5.45724332e-01 -4.53923225e-01 -5.77146947e-01 7.98302352e-01
4.13617082e-02 -9.48932886e-01 7.08329454e-02 2.08004460e-01
7.17425406e-01 1.62031159e-01 1.99251994e-01 -3.88163775e-01
5.77397048e-02 -8.44262660e-01 2.56627083e-01 7.97243059e-01
1.04824233e+00 -5.28030634e-01 4.18927699e-01 3.13757330e-01
-8.13473284e-01 5.67876063e-02 -4.47319031e-01 1.05790924e-02
1.24365158e-01 1.01702142e+00 -5.38266420e-01 5.88435829e-01
5.91809213e-01 6.94280386e-01 -2.87564635e-01 1.16895127e+00
-6.01469517e-01 6.97409093e-01 -2.77481109e-01 7.05742061e-01
-4.50169832e-01 -8.36773396e-01 4.70045030e-01 1.05604029e+00
3.05195272e-01 4.61796403e-01 -1.70287415e-01 9.78411138e-01
-2.17908382e-01 -1.18642248e-01 -3.99348676e-01 4.55667257e-01
5.54739945e-02 1.33782947e+00 -5.06587029e-01 -2.70401508e-01
-5.02806008e-01 1.20057261e+00 2.07916170e-01 9.64666367e-01
-7.74778008e-01 -3.75175148e-01 7.20302284e-01 2.20523953e-01
1.94800407e-01 5.09827584e-02 -2.95435730e-02 -1.27956009e+00
1.11379258e-01 -5.69207489e-01 -1.25315581e-02 -1.33607066e+00
-8.93673003e-01 2.73240387e-01 -9.79966596e-02 -1.03127015e+00
3.31450492e-01 -5.44849753e-01 -7.08309829e-01 1.17565668e+00
-2.30389142e+00 -1.01070499e+00 -7.61923969e-01 5.85005045e-01
1.35350421e-01 6.04807675e-01 6.39545500e-01 4.25591111e-01
-5.95052540e-01 2.06174791e-01 3.82616997e-01 -4.85887676e-02
9.03388739e-01 -1.06842935e+00 1.79963317e-02 9.49238598e-01
-3.03866595e-01 6.37928486e-01 7.49963641e-01 -5.15369952e-01
-1.30687106e+00 -8.81689548e-01 4.20974016e-01 -1.48485571e-01
2.43436456e-01 -1.26360804e-01 -1.17724204e+00 6.81660414e-01
3.89228314e-01 -3.06966044e-02 5.22267520e-01 -4.63415921e-01
-5.44901848e-01 -2.69681960e-01 -1.31961524e+00 3.53403896e-01
8.71645093e-01 -3.53486001e-01 5.53411320e-02 5.43196321e-01
4.50484484e-01 -4.38632756e-01 -5.81543148e-01 2.95106292e-01
6.45171881e-01 -1.15130532e+00 1.29217100e+00 6.92567751e-02
6.90944910e-01 -3.32250148e-01 4.65104170e-02 -1.30504394e+00
-2.93420374e-01 -5.50931215e-01 1.64454535e-01 8.69310677e-01
4.67037499e-01 -9.75369573e-01 3.88654470e-01 5.85242033e-01
5.33018634e-02 -4.76096988e-01 -3.84531081e-01 -6.13445699e-01
1.26609802e-01 -1.74610056e-02 2.82703221e-01 6.17463171e-01
-4.39141154e-01 6.18773140e-02 -6.31174088e-01 5.46227574e-01
1.16070402e+00 4.57830906e-01 7.51819968e-01 -1.01632893e+00
-2.73135751e-01 1.50531465e-02 2.65694767e-01 -1.29565704e+00
-1.45530760e-01 -5.95475614e-01 2.85650015e-01 -1.77466094e+00
6.31771147e-01 -2.72316605e-01 -2.34634399e-01 3.97547364e-01
-3.92238826e-01 7.05111802e-01 1.32322252e-01 5.75456381e-01
-2.13099182e-01 2.32346714e-01 1.67999232e+00 -6.42687306e-02
-1.87310100e-01 -3.07719428e-02 -9.94690120e-01 6.41112626e-01
9.45958257e-01 -3.69703054e-01 -4.48584080e-01 -5.87852478e-01
1.74911797e-01 -1.66515768e-01 5.35759151e-01 -7.80769944e-01
-4.42082398e-02 -3.03004086e-01 2.83347398e-01 -1.96084514e-01
6.72481894e-01 -7.45365858e-01 1.48338571e-01 3.69106114e-01
-1.72058903e-02 -7.41389036e-01 2.91918337e-01 3.89972925e-01
6.67246953e-02 -2.49338835e-01 1.17801499e+00 -1.84990197e-01
-2.04333425e-01 1.74753413e-01 -2.19345778e-01 -1.46157712e-01
4.95848149e-01 -4.83659506e-02 -8.22608590e-01 -6.03612959e-01
-3.20693135e-01 2.71095261e-02 7.59166658e-01 -2.35126585e-01
7.54459023e-01 -8.41327369e-01 -8.87446821e-01 2.36857742e-01
-1.11846693e-01 -1.58492818e-01 4.41201001e-01 1.19442213e+00
-5.51268518e-01 -1.42292857e-01 2.12032929e-01 -4.95565087e-01
-1.74279916e+00 4.34832990e-01 7.63495803e-01 -1.63632423e-01
-8.37549567e-01 7.63933778e-01 7.23087370e-01 -1.45986617e-01
-2.42850497e-01 -1.00293644e-01 1.30187646e-01 -4.56037968e-01
9.90396202e-01 2.94005841e-01 4.86456417e-03 -6.15291774e-01
-6.11063316e-02 1.10172093e+00 -1.01677448e-01 3.74980196e-02
1.34372139e+00 -5.25937438e-01 -2.77487695e-01 5.14076576e-02
1.39093471e+00 2.69699275e-01 -1.36734104e+00 -3.77156347e-01
-9.22827244e-01 -6.55871868e-01 3.17151755e-01 -1.03701317e+00
-1.39576662e+00 8.29470277e-01 4.59691465e-01 1.63699120e-01
1.70797312e+00 -2.04119235e-01 8.49769354e-01 1.34584367e-01
3.26238163e-02 -7.99238086e-01 6.83482364e-02 1.52362272e-01
8.54869962e-01 -1.15863633e+00 4.55715477e-01 -1.01485133e+00
-3.25855792e-01 8.97125244e-01 3.82662505e-01 -2.42615789e-02
5.87110639e-01 4.56367791e-01 4.40128714e-01 -3.07706058e-01
-4.02047724e-01 -2.21117750e-01 1.26001239e-01 7.02627063e-01
3.54985505e-01 -2.24578246e-01 -1.09046489e-01 -9.73969027e-02
2.98844397e-01 1.66515514e-01 7.97465086e-01 7.15396523e-01
-5.24764299e-01 -8.93243492e-01 -4.46566075e-01 3.49123389e-01
-3.58706176e-01 -3.38363528e-01 -3.51409644e-01 4.01787966e-01
-7.57508054e-02 1.09361577e+00 8.51139575e-02 -9.17325616e-02
3.75002503e-01 -1.80229977e-01 7.93119550e-01 -4.34549361e-01
-2.98822671e-01 3.71814817e-01 -1.67341873e-01 -5.73675632e-01
-7.85175264e-01 -3.18289459e-01 -1.42047286e+00 -3.96537393e-01
-4.27733272e-01 -1.12674199e-01 6.01471305e-01 6.26546860e-01
4.08777356e-01 4.79633301e-01 7.77426481e-01 -8.06526780e-01
-1.01644650e-01 -7.61888444e-01 -7.43970275e-01 6.37736917e-01
8.69399190e-01 -3.69779646e-01 -8.84750545e-01 1.14535458e-01] | [10.577493667602539, -2.6907458305358887] |
9defa53d-84d3-4a29-8cf4-6fe25079639b | fair-and-optimal-classification-via | 2211.01528 | null | https://arxiv.org/abs/2211.01528v3 | https://arxiv.org/pdf/2211.01528v3.pdf | Fair and Optimal Classification via Post-Processing | To mitigate the bias exhibited by machine learning models, fairness criteria can be integrated into the training process to ensure fair treatment across all demographics, but it often comes at the expense of model performance. Understanding such tradeoffs, therefore, underlies the design of fair algorithms. To this end, this paper provides a complete characterization of the inherent tradeoff of demographic parity on classification problems, under the most general multi-group, multi-class, and noisy setting. Specifically, we show that the minimum error rate achievable by randomized and attribute-aware fair classifiers is given by the optimal value of a Wasserstein-barycenter problem. On the practical side, our findings lead to a simple post-processing algorithm that derives fair classifiers from score functions, which yields the optimal fair classifier when the score is Bayes optimal. We provide suboptimality analysis and sample complexity for our algorithm, and demonstrate its effectiveness on benchmark datasets. | ['Han Zhao', 'Lang Yin', 'Ruicheng Xian'] | 2022-11-03 | null | null | null | null | ['classification'] | ['methodology'] | [ 2.98076123e-01 2.47131452e-01 -6.66588604e-01 -8.50961566e-01
-8.96854043e-01 -4.93767112e-01 3.00051033e-01 3.84513319e-01
-5.83363116e-01 9.17957008e-01 1.27593502e-01 -5.53081632e-01
-4.36785907e-01 -6.51649058e-01 -4.74004298e-01 -7.38829970e-01
1.22578003e-01 3.08623314e-01 -6.95983827e-01 2.95693398e-01
3.56436342e-01 2.51976311e-01 -1.43327880e+00 -1.12329535e-01
1.09694088e+00 1.09662664e+00 -4.78942871e-01 4.96923178e-01
2.78908700e-01 6.32135332e-01 -3.34622473e-01 -1.09239817e+00
5.34298539e-01 -4.16624457e-01 -7.04537690e-01 -3.96378450e-02
6.69032693e-01 -6.24724567e-01 9.92982239e-02 1.16403878e+00
5.25989771e-01 7.14504942e-02 1.17668700e+00 -1.70446730e+00
-5.86360276e-01 8.80302191e-01 -5.20644486e-01 -3.73306684e-02
-1.25471875e-01 -1.37805417e-01 1.65185845e+00 -5.30973732e-01
1.81693152e-01 1.43730617e+00 7.53490090e-01 6.96165442e-01
-1.45490646e+00 -8.73630047e-01 3.56646292e-02 -1.07672870e-01
-1.21522880e+00 -6.27726197e-01 3.25748980e-01 -4.92334485e-01
1.28859282e-01 5.30285537e-01 4.10900712e-01 1.06411004e+00
2.11452529e-01 7.03536928e-01 1.32028747e+00 -3.68294656e-01
5.62606156e-01 9.61176977e-02 5.58036208e-01 4.67046350e-01
9.28890526e-01 1.49157241e-01 -3.83061051e-01 -6.56459510e-01
3.00179332e-01 -5.31304069e-02 -1.17294781e-01 -5.51060617e-01
-3.70299041e-01 1.04122138e+00 1.11109249e-01 -4.47095335e-01
-1.96969047e-01 2.77062297e-01 2.81471878e-01 3.11873883e-01
6.10617399e-01 3.63128155e-01 -2.85557956e-01 7.19091743e-02
-9.23227012e-01 5.70794582e-01 8.24960887e-01 7.57292509e-01
4.41978544e-01 -3.52166235e-01 -4.26458836e-01 7.18088329e-01
1.32829756e-01 5.15054047e-01 -1.39921859e-01 -1.30228829e+00
4.05687600e-01 2.72259057e-01 4.34801370e-01 -7.97157288e-01
-3.60261261e-01 -6.92532241e-01 -8.85380685e-01 -9.95045975e-02
8.75990152e-01 -3.33601803e-01 -3.67012411e-01 2.24764633e+00
4.61038113e-01 -1.73676118e-01 -1.42947450e-01 8.13881934e-01
-1.71516445e-02 -7.44894668e-02 5.90713263e-01 -3.94897223e-01
1.30403495e+00 -3.53199303e-01 -5.13373137e-01 -2.28861079e-01
7.54488826e-01 -3.95875126e-01 1.11416268e+00 1.53070256e-01
-9.67427075e-01 5.17771803e-02 -9.45600748e-01 -1.14781789e-01
2.04100221e-01 -1.35099366e-01 7.08666623e-01 1.27860892e+00
-5.32692015e-01 6.86359763e-01 -4.83847886e-01 -4.02720094e-01
8.28748286e-01 1.94995925e-01 -2.77619623e-02 1.66005101e-02
-8.75783205e-01 7.62860477e-01 -2.05889270e-01 -1.53921142e-01
-6.22804523e-01 -1.15095818e+00 -5.28987467e-01 4.26054060e-01
2.74864197e-01 -9.49211180e-01 1.43289721e+00 -1.19889295e+00
-8.51365626e-01 7.99199879e-01 1.15584238e-02 -5.24497747e-01
1.06546974e+00 -9.04736146e-02 3.20928097e-01 -2.69330680e-01
2.18185306e-01 2.41519585e-01 5.55971205e-01 -1.14331436e+00
-6.96711838e-01 -5.39550364e-01 1.52719513e-01 3.06010127e-01
-6.52371526e-01 1.43993989e-01 3.85814577e-01 -5.59672177e-01
-1.57123402e-01 -8.63950968e-01 -3.58660519e-01 2.21854702e-01
-4.73994583e-01 -2.08011374e-01 -2.66682267e-01 -3.92869622e-01
1.18937433e+00 -2.08549404e+00 -3.29946756e-01 4.15686399e-01
2.51887262e-01 -3.92722905e-01 8.03990941e-03 2.80339252e-02
1.11947134e-01 4.63425368e-01 -2.08333552e-01 -5.39546013e-01
4.65753317e-01 -4.72151325e-04 -3.14018011e-01 8.30625534e-01
-1.56854570e-01 5.73063433e-01 -6.44852698e-01 -5.83283663e-01
-2.46515527e-01 -1.19079508e-01 -8.98404479e-01 -1.02918437e-02
3.38820040e-01 -1.72817811e-01 -4.30945337e-01 5.09360015e-01
8.44142377e-01 -2.42730632e-01 4.21175271e-01 1.49849746e-02
2.16310978e-01 2.47948036e-01 -1.06739855e+00 8.19762766e-01
-2.34267160e-01 2.35656649e-01 4.33715820e-01 -1.15787971e+00
4.98918265e-01 -7.67403319e-02 4.89642769e-01 -4.36809570e-01
3.81409377e-01 2.40949228e-01 -3.26418020e-02 -7.65922219e-02
5.20205498e-01 -5.76139152e-01 -3.91534060e-01 7.37706959e-01
-3.27898234e-01 3.20831209e-01 -1.12446837e-01 2.02432973e-03
8.32435668e-01 -4.16918397e-01 5.16974568e-01 -7.63370752e-01
-1.03537224e-01 -2.74729341e-01 7.89697766e-01 1.32143843e+00
-6.92563593e-01 4.75122988e-01 8.03163946e-01 -2.36713365e-01
-1.04748869e+00 -1.03785551e+00 -5.24564862e-01 1.26102209e+00
-1.40467370e-02 -1.00080691e-01 -9.09992754e-01 -6.55494273e-01
5.78890979e-01 7.06497371e-01 -9.53270197e-01 -2.79792905e-01
2.29940414e-02 -1.30484283e+00 6.07695043e-01 4.31353569e-01
2.30499104e-01 -9.33585465e-02 -7.33830214e-01 -2.45742992e-01
-1.64103448e-01 -7.92597651e-01 -5.71238458e-01 1.13528535e-01
-8.74279559e-01 -1.13868439e+00 -6.43469691e-01 -1.91980675e-01
6.18852079e-01 1.72683865e-01 1.05888486e+00 1.17591657e-01
-2.21379008e-02 2.53795922e-01 1.52998231e-02 -5.74977398e-01
-2.02391997e-01 2.65691310e-01 2.30177030e-01 -2.22138390e-02
4.20443535e-01 -3.78802627e-01 -6.51506722e-01 2.52221465e-01
-4.39155281e-01 -2.64633596e-02 3.59209448e-01 9.08024073e-01
3.60057741e-01 -2.83060670e-01 8.33061874e-01 -1.07147098e+00
7.73001373e-01 -5.36462247e-01 -5.99192441e-01 5.44381797e-01
-9.60215986e-01 1.31336585e-01 4.79942292e-01 -3.22048366e-01
-9.61504936e-01 -1.99012756e-01 2.00004101e-01 3.88809703e-02
3.02759916e-01 3.31065238e-01 -3.42612058e-01 1.58456098e-02
8.44524503e-01 -3.67181122e-01 9.41150784e-02 -3.81854117e-01
3.88507307e-01 1.03671563e+00 2.80825734e-01 -1.07158363e+00
4.70182151e-01 3.65519881e-01 2.30932489e-01 -3.84065062e-01
-1.39282608e+00 -2.09184624e-02 -2.79121399e-01 -8.23226646e-02
3.24484169e-01 -8.59537244e-01 -1.04402030e+00 1.93974569e-01
-5.79702020e-01 -4.47079927e-01 -3.12825620e-01 3.98851007e-01
-5.93014956e-01 2.03772947e-01 -4.14094239e-01 -1.36398184e+00
-3.91101807e-01 -7.37677157e-01 7.39636123e-01 5.36022484e-02
-3.82137418e-01 -7.61609614e-01 -2.82102048e-01 6.18905067e-01
1.21495612e-01 1.28462449e-01 1.21686125e+00 -7.52562821e-01
-3.02445829e-01 -2.47627974e-01 -4.25851852e-01 2.18778238e-01
-7.80654997e-02 -3.00764311e-02 -1.13907862e+00 -3.33966196e-01
-2.61870116e-01 -4.26975429e-01 7.69095540e-01 7.31051087e-01
1.48554695e+00 -5.81095457e-01 -1.28407300e-01 6.17161632e-01
1.34455955e+00 -3.08677465e-01 8.53291377e-02 6.80868700e-02
3.97717148e-01 8.64757299e-01 6.32663786e-01 9.38246250e-01
6.77933872e-01 3.39550465e-01 1.71627015e-01 2.16191351e-01
2.79857785e-01 -3.34852248e-01 3.29635963e-02 2.20552951e-01
1.45297959e-01 -1.37502745e-01 -7.64685690e-01 3.96256745e-01
-1.77222049e+00 -1.09000123e+00 1.15711272e-01 2.67835522e+00
8.47838342e-01 8.02117437e-02 3.88744324e-01 2.87089437e-01
7.05137730e-01 -1.73491761e-01 -7.31477737e-01 -5.83230138e-01
-1.87550206e-03 4.86565046e-02 9.50705409e-01 5.85407794e-01
-1.04422665e+00 3.51013720e-01 7.32102728e+00 8.71923089e-01
-5.39321184e-01 6.47501051e-02 1.25030887e+00 -4.64930654e-01
-4.93340552e-01 -8.88728499e-02 -6.75544798e-01 4.84093755e-01
9.73557591e-01 -8.05109859e-01 4.59980756e-01 9.76617336e-01
1.98591754e-01 -7.24676028e-02 -1.42506647e+00 7.77920902e-01
-1.41710445e-01 -8.71733129e-01 -1.48270547e-01 3.65448415e-01
7.04207361e-01 -4.53412324e-01 2.86004454e-01 2.32881904e-01
7.17142820e-01 -9.96629298e-01 8.72273564e-01 7.13291168e-02
8.08130801e-01 -1.08795667e+00 6.25099540e-01 3.21115106e-01
-6.08790815e-01 -6.05639517e-01 -4.64382559e-01 -3.41888875e-01
-2.55311906e-01 7.98300445e-01 -4.27128375e-01 3.64356816e-01
4.88552809e-01 3.10639024e-01 -3.98893028e-01 9.82692480e-01
2.82794386e-01 6.40405774e-01 -3.90167207e-01 -9.62148011e-02
-2.16478333e-01 -9.23504680e-02 2.59405617e-02 9.45861042e-01
3.64942312e-01 2.85898983e-01 -5.03169782e-02 7.27309167e-01
-4.78437662e-01 3.27845365e-01 -2.84678251e-01 1.92886770e-01
9.59094465e-01 1.12249815e+00 -5.27562797e-01 -1.59211934e-01
-1.53831407e-01 3.82122785e-01 4.45964932e-01 2.65498757e-01
-6.76362038e-01 -2.17554569e-01 1.11183000e+00 1.01585783e-01
-2.89970458e-01 2.80030042e-01 -1.15354788e+00 -9.71627176e-01
-3.78324762e-02 -8.55356812e-01 8.40536833e-01 -2.42102444e-01
-1.80278563e+00 -3.15767676e-01 1.17185652e-01 -7.97518730e-01
2.57684998e-02 -5.60925901e-01 -3.81167263e-01 7.52120495e-01
-1.32744539e+00 -7.46643245e-01 -1.84527799e-01 2.55782008e-01
-9.68424529e-02 1.75036699e-01 6.41756237e-01 4.26617473e-01
-7.43992746e-01 1.34663975e+00 3.59194398e-01 3.02896556e-02
8.36188793e-01 -1.25880396e+00 1.02108298e-02 6.99252546e-01
-4.64991480e-01 5.70198834e-01 8.84343505e-01 -4.03673530e-01
-1.05536747e+00 -1.00299597e+00 9.72432733e-01 -4.92057890e-01
5.09384215e-01 -3.88056368e-01 -4.48813051e-01 6.09648705e-01
-3.31263691e-01 -9.64110419e-02 1.06394243e+00 6.64242983e-01
-4.50290471e-01 -4.34958071e-01 -1.52396584e+00 8.43477666e-01
1.21759462e+00 -4.60074216e-01 -8.72819051e-02 1.21062972e-01
2.91517347e-01 -1.16778992e-01 -9.02026594e-01 2.92449117e-01
1.04303789e+00 -9.57049012e-01 8.51706982e-01 -1.02395272e+00
5.80538809e-01 4.09684569e-01 -5.16536593e-01 -1.18402600e+00
-2.11388007e-01 -5.31120777e-01 2.55149931e-01 1.23392200e+00
3.70348662e-01 -7.05693841e-01 7.91336060e-01 1.46633542e+00
2.80508667e-01 -7.57408082e-01 -1.07960916e+00 -5.90849817e-01
7.55208433e-01 -4.52351838e-01 7.91847646e-01 9.13958609e-01
2.21851375e-02 5.38295545e-02 -5.60076058e-01 -5.28794788e-02
1.28728926e+00 2.05225199e-01 6.71531022e-01 -1.58058918e+00
-1.52324677e-01 -6.11238360e-01 -1.39023021e-01 -6.82572603e-01
4.91523325e-01 -8.06361079e-01 -4.97402251e-02 -9.75826740e-01
8.59557927e-01 -9.08866227e-01 -2.69420087e-01 4.25465018e-01
-5.09585679e-01 -1.62297599e-02 3.12684536e-01 1.03023268e-01
-3.77707273e-01 3.88354957e-01 9.50616002e-01 2.15637181e-02
1.52484268e-01 1.75817609e-01 -1.48933899e+00 5.11258602e-01
8.39966297e-01 -6.87506914e-01 -3.95131409e-01 -1.72968760e-01
2.86283523e-01 -5.63466959e-02 4.12711680e-01 -4.72080618e-01
2.98778899e-02 -7.00806558e-01 4.35135484e-01 -1.15347197e-02
1.85458794e-01 -6.43577099e-01 -9.35385972e-02 4.46460336e-01
-1.03755546e+00 -1.16158091e-01 -3.52695763e-01 5.00617564e-01
4.79002059e-01 -1.85482398e-01 1.13961005e+00 2.11087242e-01
3.01387519e-01 3.68082374e-01 -1.32678613e-01 5.36171794e-01
9.41627085e-01 6.94288760e-02 -5.53115845e-01 -4.21763241e-01
-2.65716940e-01 3.82885277e-01 5.95560372e-01 -9.62803792e-03
-2.70653167e-03 -1.42819214e+00 -7.89659619e-01 4.40257601e-02
1.22184418e-01 -4.84258384e-01 1.41386256e-01 6.91829979e-01
-6.50168434e-02 6.30987855e-03 -1.86459601e-01 -1.59560248e-01
-1.28926623e+00 2.82007098e-01 6.01615012e-01 4.58256453e-02
1.05704255e-02 6.90693855e-01 2.27226004e-01 -2.47583613e-01
3.75446945e-01 -1.13003649e-01 2.09116802e-01 2.34544292e-01
4.51090872e-01 7.25805998e-01 2.07693670e-02 -2.97788292e-01
-3.42048228e-01 2.15668961e-01 8.78776684e-02 -2.39899904e-01
9.90919113e-01 -2.66400278e-01 7.00581213e-03 2.11738750e-01
1.01546824e+00 -1.65873900e-01 -1.28448868e+00 -7.82646015e-02
5.26452214e-02 -7.78841436e-01 -7.71030188e-02 -5.62343240e-01
-8.98926616e-01 7.16702163e-01 4.86527592e-01 2.67064780e-01
8.91791046e-01 -3.36075991e-01 2.69139767e-01 3.44196945e-01
2.28778720e-01 -1.24144590e+00 -4.52330202e-01 3.12140435e-02
2.89865375e-01 -1.25030637e+00 2.89667070e-01 -2.68474787e-01
-6.22467875e-01 6.54847801e-01 5.15859425e-01 3.86346504e-02
8.09274614e-01 2.00811803e-01 -6.69038966e-02 4.10459757e-01
-8.69396865e-01 1.08595401e-01 7.92201161e-02 3.88141513e-01
4.21183497e-01 6.31185174e-01 -7.27881789e-01 1.13194573e+00
-5.41704178e-01 -5.20507284e-02 5.79364240e-01 5.94289482e-01
-3.62492651e-01 -9.23343182e-01 -2.70024598e-01 9.62253809e-01
-9.24650729e-01 1.46176936e-02 -3.33531410e-01 4.26590711e-01
-1.00416772e-01 1.08771300e+00 3.25346142e-01 -1.43624529e-01
7.15206563e-02 5.46024814e-02 5.88095546e-01 -2.54730016e-01
-4.71324295e-01 -3.47310960e-01 1.55353695e-01 -3.95721883e-01
-1.80517137e-01 -9.17621732e-01 -6.38799071e-01 -1.02072835e+00
-1.99954569e-01 1.95218980e-01 3.27479839e-01 8.16317856e-01
2.10742399e-01 -1.51719436e-01 8.91120851e-01 -4.25430477e-01
-1.59775841e+00 -6.89837635e-01 -7.38399148e-01 5.60125172e-01
3.71768981e-01 -6.21424496e-01 -6.26797616e-01 -3.60613197e-01] | [8.857048988342285, 5.271857738494873] |
e0a726bb-37bc-4b62-837c-66049b5091aa | learning-clip-guided-visual-text-fusion | 2304.10091 | null | https://arxiv.org/abs/2304.10091v1 | https://arxiv.org/pdf/2304.10091v1.pdf | Learning CLIP Guided Visual-Text Fusion Transformer for Video-based Pedestrian Attribute Recognition | Existing pedestrian attribute recognition (PAR) algorithms are mainly developed based on a static image. However, the performance is not reliable for images with challenging factors, such as heavy occlusion, motion blur, etc. In this work, we propose to understand human attributes using video frames that can make full use of temporal information. Specifically, we formulate the video-based PAR as a vision-language fusion problem and adopt pre-trained big models CLIP to extract the feature embeddings of given video frames. To better utilize the semantic information, we take the attribute list as another input and transform the attribute words/phrase into the corresponding sentence via split, expand, and prompt. Then, the text encoder of CLIP is utilized for language embedding. The averaged visual tokens and text tokens are concatenated and fed into a fusion Transformer for multi-modal interactive learning. The enhanced tokens will be fed into a classification head for pedestrian attribute prediction. Extensive experiments on a large-scale video-based PAR dataset fully validated the effectiveness of our proposed framework. | ['Xiao Wang', 'Xiaohao Wu', 'Zihan Yang', 'Jiandong Jin', 'Jun Zhu'] | 2023-04-20 | null | null | null | null | ['pedestrian-attribute-recognition'] | ['computer-vision'] | [ 1.02909297e-01 -5.78465939e-01 -1.92967087e-01 -7.88815558e-01
-7.58293569e-01 -1.64714083e-01 5.13806701e-01 1.30138814e-01
-4.72296655e-01 6.62701845e-01 4.15456414e-01 1.16207696e-01
4.24893469e-01 -6.98299408e-01 -7.19865322e-01 -7.85773277e-01
4.99908537e-01 -5.27216829e-02 2.11150974e-01 1.90056220e-01
-8.74895416e-03 -1.67420413e-02 -1.67114198e+00 5.61443388e-01
7.69981921e-01 1.39398956e+00 3.82689655e-01 4.69800681e-01
-3.82009357e-01 1.05282426e+00 -1.94971934e-01 -6.79698944e-01
-1.20284855e-01 -1.68691933e-01 -3.41482460e-01 4.49657708e-01
3.54211599e-01 -7.75720358e-01 -6.68385625e-01 1.00738251e+00
5.51891506e-01 1.77359521e-01 4.91721928e-01 -1.45966029e+00
-7.30620146e-01 4.55426365e-01 -3.95149916e-01 7.59079158e-02
6.68908656e-01 2.26338878e-01 9.24301088e-01 -1.17410612e+00
5.05300701e-01 1.50332582e+00 3.32676679e-01 4.20692712e-01
-8.01565528e-01 -7.31190503e-01 5.62416613e-01 1.03389049e+00
-1.44606471e+00 -5.94091058e-01 8.80688131e-01 -3.56961489e-01
5.65588832e-01 1.46563813e-01 9.44571972e-01 1.26543844e+00
-2.21302938e-02 1.50203824e+00 8.12847137e-01 -9.63336751e-02
-1.41586244e-01 3.02047282e-01 7.62270018e-02 5.28278708e-01
-1.10812299e-01 -3.23662668e-01 -5.53800166e-01 1.81956142e-01
3.13952565e-01 5.18275976e-01 -1.66944459e-01 -2.28357553e-01
-1.42835438e+00 6.88522935e-01 2.52112776e-01 -1.46663383e-01
-6.61549449e-01 -1.13107368e-01 7.50804782e-01 5.13249300e-02
4.51979160e-01 -5.05216062e-01 -2.68757463e-01 -1.48380801e-01
-7.57511258e-01 1.48375288e-01 3.92270625e-01 1.15329361e+00
8.28562260e-01 -1.89099964e-02 -3.60439390e-01 9.58531260e-01
5.81168473e-01 6.73190117e-01 2.41661668e-01 -8.07800114e-01
5.46015322e-01 6.03455067e-01 -1.24336325e-01 -1.12711680e+00
2.92305052e-02 -3.15913074e-02 -8.03231478e-01 -2.53049850e-01
5.20658903e-02 -4.89447564e-02 -7.44843602e-01 1.47623670e+00
4.18308228e-01 3.27099651e-01 1.44646898e-01 1.13550317e+00
9.58429337e-01 9.57327902e-01 7.47492611e-01 -2.51232922e-01
1.73504674e+00 -1.10213065e+00 -7.19233334e-01 -2.95163602e-01
4.12590206e-01 -8.91086757e-01 9.19984162e-01 1.61450505e-01
-8.11654985e-01 -8.71951699e-01 -9.66086268e-01 -2.27487519e-01
-2.75448591e-01 4.98991787e-01 3.91723394e-01 2.74085522e-01
-5.40821493e-01 -1.74471363e-01 -5.90979040e-01 -3.84138823e-01
6.45615399e-01 1.62264973e-01 -5.09173095e-01 -4.80386287e-01
-1.30141222e+00 5.21615684e-01 5.18854678e-01 2.09434152e-01
-1.07994330e+00 -3.66604656e-01 -1.18633270e+00 -1.11630969e-01
3.45461011e-01 -8.06628823e-01 9.54379082e-01 -1.15393913e+00
-1.16027927e+00 7.22742975e-01 -5.04133463e-01 -3.91871721e-01
4.31864470e-01 -4.99881208e-01 -4.26118761e-01 2.60198146e-01
3.01922470e-01 7.85189271e-01 7.20773876e-01 -9.97522533e-01
-1.01437819e+00 -5.46702266e-01 1.76043540e-01 4.65418696e-01
-6.42336190e-01 1.99936226e-01 -6.87209249e-01 -5.58423936e-01
-1.52207345e-01 -7.30137348e-01 -1.51909841e-02 4.00481001e-02
-3.74262154e-01 -2.85415828e-01 9.72164989e-01 -9.81502235e-01
1.07410836e+00 -2.41051126e+00 5.59229776e-02 -1.74628437e-01
8.46088231e-02 1.08722247e-01 -6.95125163e-02 5.64785562e-02
1.09336808e-01 -1.91574052e-01 1.45131573e-01 -3.08501214e-01
-9.65180248e-02 8.13520402e-02 -1.28979787e-01 2.85634279e-01
2.22664833e-01 6.93361640e-01 -7.96812475e-01 -1.26419199e+00
6.18516684e-01 6.26047850e-01 -4.01695162e-01 4.30429876e-01
-2.53679365e-01 3.14519674e-01 -9.83931005e-01 8.18737626e-01
7.20593810e-01 1.23125121e-01 -4.89196301e-01 -6.71535790e-01
-1.51872441e-01 -2.15345189e-01 -1.07502234e+00 1.74813282e+00
-3.25580537e-01 6.47935033e-01 -1.04231134e-01 -1.04664481e+00
9.74608779e-01 2.94347227e-01 6.03128374e-01 -4.80402917e-01
1.08952790e-01 -6.01264983e-02 -3.81170332e-01 -1.02688456e+00
5.21133959e-01 1.95785433e-01 -4.32435125e-02 5.50090075e-02
1.14147522e-01 4.57703263e-01 1.42649472e-01 2.52676487e-01
5.97669125e-01 3.97818506e-01 1.62818402e-01 2.62491494e-01
1.16508758e+00 -1.18416600e-01 6.87275231e-01 2.04672948e-01
-4.42575276e-01 4.76897299e-01 1.71683401e-01 -7.08317041e-01
-1.31358266e+00 -9.65637028e-01 1.44587420e-02 1.37461936e+00
3.61056209e-01 -4.89641845e-01 -6.74847186e-01 -5.94422460e-01
-1.83191404e-01 6.14569962e-01 -3.40153545e-01 -1.63324118e-01
-4.05155212e-01 -5.10805726e-01 2.94969290e-01 6.40787423e-01
9.99951005e-01 -9.77359116e-01 -3.15560371e-01 3.28659594e-01
-6.81577504e-01 -1.49974418e+00 -6.50761783e-01 -5.25033176e-01
-2.14507192e-01 -8.17481339e-01 -8.51803839e-01 -9.69013095e-01
4.31387007e-01 3.31513762e-01 6.19742990e-01 -9.30401012e-02
-6.74137920e-02 4.33561593e-01 -5.60257196e-01 -2.64197022e-01
1.50204813e-02 -1.53639287e-01 1.88080698e-01 7.37421334e-01
7.82001078e-01 -3.81498605e-01 -7.67424405e-01 2.23915547e-01
-4.84117299e-01 4.39269125e-01 8.09334755e-01 8.68521690e-01
7.00159848e-01 -2.80431539e-01 3.63207817e-01 -2.88972735e-01
2.33168900e-01 -4.46057320e-01 -9.37146544e-02 3.30731183e-01
1.50728682e-02 -9.05746520e-02 6.22534275e-01 -6.05356812e-01
-1.38433778e+00 3.18712294e-01 -2.78380007e-01 -5.62507272e-01
-4.49867100e-01 3.33425879e-01 -7.80848026e-01 3.15449893e-01
-1.00170381e-01 6.65036798e-01 -1.12427771e-01 -2.87523210e-01
4.28021371e-01 1.06762052e+00 6.49635673e-01 -6.62591279e-01
4.23819393e-01 2.66937703e-01 -3.64068657e-01 -8.74373734e-01
-9.53426957e-01 -3.32685471e-01 -6.19818926e-01 -5.51436484e-01
1.41906345e+00 -1.29021370e+00 -8.19210589e-01 6.78452373e-01
-1.27347040e+00 3.04585218e-01 1.16390884e-01 6.64844692e-01
-6.19203925e-01 5.88843107e-01 -5.38523436e-01 -8.48441601e-01
-3.12933326e-01 -1.35991395e+00 1.06275606e+00 5.93265533e-01
2.96278447e-01 -6.41649544e-01 -5.11751354e-01 6.81564093e-01
1.74406901e-01 4.90229987e-02 6.67767286e-01 -8.49978089e-01
-6.79068565e-01 -2.74384260e-01 -5.73558092e-01 3.82334292e-01
-9.89061818e-02 1.19113155e-01 -9.52072203e-01 1.27669983e-02
-2.38158271e-01 -4.88816291e-01 8.74172270e-01 3.58120501e-01
1.40083015e+00 -3.96784782e-01 -3.96440238e-01 5.78684568e-01
1.17213655e+00 1.64312601e-01 5.50986588e-01 4.96410549e-01
9.81353819e-01 5.04199982e-01 1.14269960e+00 6.37563467e-01
9.22674894e-01 5.34383833e-01 2.99691290e-01 2.11158454e-01
3.55777927e-02 -3.54929507e-01 6.17079496e-01 9.49120820e-01
-4.38751243e-02 -1.66855454e-01 -6.91990912e-01 4.34827685e-01
-1.98035884e+00 -1.23521781e+00 1.15117207e-01 1.88019192e+00
6.29206300e-01 1.58203840e-01 -3.92240845e-02 -4.95022424e-02
9.54468966e-01 2.98181593e-01 -4.89988267e-01 8.39163363e-03
-3.06172490e-01 -5.86208522e-01 8.87774229e-02 3.70096862e-02
-1.43456435e+00 9.32818115e-01 4.51562738e+00 9.26625490e-01
-9.00646508e-01 -4.94249500e-02 7.66994298e-01 8.67960677e-02
-1.92731485e-01 5.53431734e-02 -8.73640239e-01 7.03315794e-01
7.70075083e-01 -2.73450911e-01 6.71009868e-02 9.78014112e-01
3.23567361e-01 6.00366890e-02 -1.05636871e+00 1.25148249e+00
3.27661812e-01 -8.87758434e-01 5.48554540e-01 -1.72213256e-01
1.35426298e-01 -3.11242551e-01 4.66059782e-02 4.18319404e-01
-8.89154822e-02 -7.87824810e-01 7.94033170e-01 7.96123505e-01
6.32092476e-01 -7.45269001e-01 9.05018389e-01 1.87265366e-01
-1.68098247e+00 -2.46243328e-01 -5.79803228e-01 7.98711330e-02
3.63602877e-01 2.64750689e-01 -3.52420479e-01 6.61030829e-01
7.05766022e-01 1.21802580e+00 -6.02866173e-01 1.00927222e+00
-1.34328594e-02 3.72554421e-01 -2.31528178e-01 -5.06047010e-02
1.26342684e-01 -2.26054549e-01 3.54965359e-01 1.20935547e+00
4.03292865e-01 3.47273350e-01 6.41035557e-01 4.00185645e-01
1.67792179e-02 4.26370472e-01 -5.12917757e-01 7.01833218e-02
6.35427535e-01 1.34038043e+00 -8.43611658e-02 -6.46572173e-01
-9.95779097e-01 1.10042286e+00 3.27817127e-02 4.06990021e-01
-9.88144994e-01 -2.79137373e-01 6.88026547e-01 -2.14184687e-01
3.42999876e-01 -6.82065636e-02 9.79759544e-02 -1.42417383e+00
1.77336901e-01 -7.35690892e-01 5.16340494e-01 -1.01934242e+00
-1.16461658e+00 4.32049781e-01 -6.82099089e-02 -1.41116107e+00
-1.09980896e-01 -4.64575529e-01 -4.88435835e-01 6.28996372e-01
-1.40425527e+00 -1.84945381e+00 -7.21591651e-01 9.82565284e-01
1.04863024e+00 -3.60518128e-01 7.05882430e-01 6.66750312e-01
-7.60696769e-01 5.97637057e-01 -1.57476127e-01 4.81614172e-01
8.85871708e-01 -9.21267867e-01 -3.01728696e-02 7.90352762e-01
-2.05416903e-01 1.57003596e-01 6.96776986e-01 -5.35436571e-01
-1.44720995e+00 -1.27690458e+00 7.39450216e-01 -1.59465849e-01
5.81443012e-01 -1.80509165e-01 -8.22148919e-01 6.86230421e-01
2.56197691e-01 1.50190473e-01 6.78064942e-01 -3.40854585e-01
-3.01823527e-01 -4.46551621e-01 -8.01437736e-01 4.36043799e-01
8.86480153e-01 -5.79703867e-01 -5.86283028e-01 1.92199439e-01
8.56680453e-01 -2.02740148e-01 -9.78446841e-01 3.49232465e-01
7.11827755e-01 -6.11608505e-01 1.23325551e+00 -6.30948544e-01
6.32801235e-01 -2.27039769e-01 -5.96186697e-01 -1.02184904e+00
-1.56810477e-01 7.15654939e-02 3.25706720e-01 1.77990365e+00
-6.75347541e-03 -2.04249859e-01 5.34991860e-01 6.01489604e-01
-1.61791369e-01 -5.36645412e-01 -7.69358397e-01 -2.97266960e-01
-4.62587744e-01 -4.56237048e-01 6.03429258e-01 6.71588004e-01
-1.57456368e-01 4.99279380e-01 -6.34690523e-01 1.05384469e-01
8.88990760e-01 4.24770825e-02 8.40850294e-01 -9.03123617e-01
2.25186989e-01 -3.67187262e-02 -8.18103671e-01 -1.13523769e+00
4.07747179e-01 -5.31496644e-01 4.89462749e-04 -1.49822521e+00
6.26130283e-01 -2.19858855e-01 -4.12923902e-01 3.23731631e-01
-4.65328485e-01 1.19517878e-01 3.17557603e-01 2.69371986e-01
-9.87495959e-01 9.81269062e-01 1.07455134e+00 -5.95647335e-01
2.16474563e-01 -1.42562583e-01 -3.06116045e-01 8.74750018e-01
6.14082694e-01 4.35355790e-02 -2.51939833e-01 -5.26797414e-01
-2.50064343e-01 2.44048208e-01 5.47913849e-01 -1.20142531e+00
3.85850996e-01 -3.34578633e-01 8.08743060e-01 -6.86503470e-01
6.18560255e-01 -9.30020213e-01 -1.56092718e-01 1.67014048e-01
-4.36239988e-01 2.00891912e-01 -2.82044243e-02 6.71505928e-01
-5.09097397e-01 1.70503139e-01 4.66503084e-01 7.40775689e-02
-1.29860890e+00 7.48502433e-01 -2.16769695e-01 -1.33905679e-01
1.38561094e+00 -2.10898265e-01 -2.98621804e-01 -5.35777152e-01
-8.22586298e-01 5.73168516e-01 4.00189638e-01 5.65989256e-01
1.03168881e+00 -1.79505444e+00 -1.03740513e+00 1.86862573e-01
3.82859260e-01 -2.18991756e-01 7.41357148e-01 7.63621449e-01
-2.67996043e-01 1.39076352e-01 -5.89500487e-01 -7.06837177e-01
-1.58571625e+00 8.47166896e-01 3.13606933e-02 3.03752393e-01
-5.52247286e-01 5.85494041e-01 3.14988464e-01 1.21122204e-01
1.71528369e-01 2.41974473e-01 -4.86791968e-01 2.88722008e-01
7.70231664e-01 -1.49969542e-02 -5.07944047e-01 -1.20366871e+00
-5.13173938e-01 6.64584577e-01 -2.81160086e-01 -4.90446985e-02
1.16573763e+00 -5.74096739e-01 -2.52299793e-02 3.95629138e-01
1.29426277e+00 -3.53071690e-01 -1.29712546e+00 -6.10890627e-01
-4.51493055e-01 -5.46129286e-01 -2.32609600e-01 -1.83334857e-01
-1.17165101e+00 1.20835829e+00 7.41832137e-01 -2.34367892e-01
1.19169724e+00 3.66541445e-02 1.14512062e+00 3.54734749e-01
1.54784441e-01 -1.01170695e+00 6.51581362e-02 1.22446932e-01
4.64768052e-01 -1.25416434e+00 -7.44649172e-02 -2.95424968e-01
-9.78125989e-01 1.14485347e+00 8.54480743e-01 1.84038356e-01
5.47027767e-01 -3.93099599e-02 2.35959701e-03 3.67381960e-01
-6.97947562e-01 -2.34959871e-01 5.99424951e-02 7.12431192e-01
2.37802133e-01 8.16679522e-02 -6.15807548e-02 8.60267520e-01
-5.96253201e-02 1.78685173e-01 3.48236948e-01 6.28502250e-01
-6.31518781e-01 -8.87780130e-01 -3.21373314e-01 4.79287058e-01
-3.35638344e-01 -5.27773798e-02 1.26483545e-01 2.77624786e-01
3.16292942e-01 9.77277040e-01 -2.88247131e-03 -7.47109115e-01
2.34900370e-01 3.23106885e-01 2.08975881e-01 -1.01684384e-01
-5.11953793e-02 -3.10733337e-02 2.10738152e-01 -5.08464336e-01
-6.48853302e-01 -9.52128232e-01 -1.03682184e+00 -3.76483440e-01
8.86192247e-02 1.76748149e-02 3.72481316e-01 1.13955760e+00
1.50388047e-01 4.49743867e-01 5.80797195e-01 -6.97619379e-01
-2.67227918e-01 -9.39701319e-01 -1.47568494e-01 8.16170394e-01
3.41623962e-01 -7.06086695e-01 -1.18138805e-01 5.34511149e-01] | [14.325281143188477, 0.9915963411331177] |
bfd5fba3-9b35-4500-b42c-6f96a60139b7 | statistical-learning-machines-from-atr-to-dna | 1906.10019 | null | https://arxiv.org/abs/1906.10019v4 | https://arxiv.org/pdf/1906.10019v4.pdf | Machine Learning Construction: implications to cybersecurity | Statistical learning is the process of estimating an unknown probabilistic input-output relationship of a system using a limited number of observations. A statistical learning machine (SLM) is the algorithm, function, model, or rule, that learns such a process; and machine learning (ML) is the conventional name of this field. ML and its applications are ubiquitous in the modern world. Systems such as Automatic target recognition (ATR) in military applications, computer aided diagnosis (CAD) in medical imaging, DNA microarrays in genomics, optical character recognition (OCR), speech recognition (SR), spam email filtering, stock market prediction, etc., are few examples and applications for ML; diverse fields but one theory. In particular, ML has gained a lot of attention in the field of cyberphysical security, especially in the last decade. It is of great importance to this field to design detection algorithms that have the capability of learning from security data to be able to hunt threats, achieve better monitoring, master the complexity of the threat intelligence feeds, and achieve timely remediation of security incidents. The field of ML can be decomposed into two basic subfields: \textit{construction} and \textit{assessment}. We mean by \textit{construction} designing or inventing an appropriate algorithm that learns from the input data and achieves a good performance according to some optimality criterion. We mean by \textit{assessment} attributing some performance measures to the constructed ML algorithm, along with their estimators, to objectively assess this algorithm. \textit{Construction} and \textit{assessment} of a ML algorithm require familiarity with different other fields: probability, statistics, matrix theory, optimization, algorithms, and programming, among others.f | ['Waleed A. Yousef'] | 2019-06-24 | null | null | null | null | ['stock-market-prediction'] | ['time-series'] | [ 4.80800509e-01 -1.58615783e-01 -1.17152810e-01 -7.06087947e-02
-5.13285756e-01 -7.43890464e-01 7.57253230e-01 4.27518040e-01
-2.05370620e-01 5.50975144e-01 -4.47920054e-01 -7.89772272e-01
-7.31640995e-01 -7.74188221e-01 -5.67268312e-01 -7.83794224e-01
-3.85052651e-01 3.61937046e-01 1.69120833e-01 -9.84353870e-02
7.31674612e-01 8.77126992e-01 -1.28002703e+00 -6.34592324e-02
6.03303909e-01 1.47969830e+00 1.06311999e-02 7.24204302e-01
3.11547443e-02 7.97965348e-01 -5.40135205e-01 -2.73830384e-01
2.67302543e-01 -3.10648441e-01 -6.13348186e-01 -1.07685134e-01
-3.30843180e-01 1.82736561e-01 -9.47779864e-02 1.26739800e+00
8.98749903e-02 -1.59949630e-01 1.07516658e+00 -1.36012173e+00
-9.36431959e-02 1.81706101e-01 -4.48842168e-01 1.57989591e-01
2.69014925e-01 7.50088990e-02 7.31605291e-01 -6.48548305e-01
1.45128220e-01 1.17727172e+00 3.96164536e-01 3.16905856e-01
-1.18606687e+00 -6.55485988e-01 -6.11025952e-02 1.78589657e-01
-1.22673738e+00 -2.13881377e-02 6.41318798e-01 -7.70138741e-01
4.27381724e-01 4.95382875e-01 7.70002231e-02 9.03300285e-01
6.06536686e-01 8.70912373e-01 1.17392838e+00 -5.21111786e-01
5.72834313e-01 3.03562433e-01 1.92015812e-01 6.30945563e-01
2.51004219e-01 5.16786039e-01 -4.13737476e-01 -3.79860580e-01
5.10375261e-01 1.33604571e-01 -7.44186267e-02 -6.01926520e-02
-8.52845430e-01 9.54987645e-01 7.57306367e-02 4.77724522e-01
-3.71415317e-01 -8.54982808e-02 3.29209298e-01 4.97749299e-01
1.61003485e-01 8.27473342e-01 -5.37352860e-01 -3.65787670e-02
-6.23550355e-01 -7.73954764e-02 9.50441480e-01 4.31654125e-01
5.97120881e-01 2.86918610e-01 3.19611698e-01 4.13807780e-01
1.16248518e-01 6.76000774e-01 4.20999289e-01 -5.77050507e-01
1.88651100e-01 5.84982216e-01 4.00662571e-02 -9.40972269e-01
-5.81165612e-01 -4.78953868e-01 -1.13053536e+00 3.98201257e-01
2.73105860e-01 -1.57503903e-01 -4.50109720e-01 1.46747911e+00
3.46930251e-02 3.50313991e-01 7.91813433e-02 4.94043261e-01
3.47603798e-01 9.17751789e-01 -7.07543194e-02 -5.77280462e-01
1.08673584e+00 3.72886122e-03 -2.84662247e-01 -1.90425158e-01
3.88887584e-01 -7.13136733e-01 6.42870486e-01 1.04957223e+00
-4.56243902e-01 -3.90575171e-01 -1.01620066e+00 9.47878242e-01
-4.34295565e-01 7.34999403e-02 5.62468767e-01 8.26337695e-01
-4.26263660e-01 6.48280442e-01 -6.43974900e-01 -1.54823124e-01
3.19437943e-02 5.14933646e-01 -2.40275711e-01 1.71789333e-01
-1.12224746e+00 1.04918694e+00 4.71575350e-01 7.48247579e-02
-1.00044656e+00 -4.16666180e-01 -7.13584602e-01 -6.97456002e-02
5.87115288e-01 -1.53017074e-01 7.86226809e-01 -8.24149847e-01
-1.23686242e+00 5.61509550e-01 1.43886164e-01 -4.75650162e-01
1.22596174e-01 -1.89484403e-01 -6.47717893e-01 -2.19739169e-01
-1.77394077e-01 -8.71975794e-02 1.16068757e+00 -1.15840471e+00
-5.68377614e-01 -6.19282246e-01 -3.85704160e-01 -3.79901558e-01
-3.45497668e-01 2.73348600e-01 4.15385664e-01 -5.71746230e-01
1.71990469e-01 -7.58381248e-01 -2.65654981e-01 -4.11039680e-01
-7.44251788e-01 -3.18252742e-01 1.02418876e+00 -5.22056937e-01
1.14887857e+00 -2.16057801e+00 2.40543276e-01 7.08064497e-01
-1.00343227e-01 4.61359739e-01 9.04786661e-02 7.23780870e-01
-2.59775013e-01 4.00679410e-02 -5.33878267e-01 2.78960407e-01
-1.64237782e-01 -1.03413211e-02 -6.29369557e-01 6.35403275e-01
2.61536062e-01 3.59276503e-01 -6.91082001e-01 -9.71299335e-02
2.74867475e-01 -1.16241544e-01 -8.24545473e-02 3.70083004e-01
-2.99892217e-01 4.57665503e-01 -8.26305270e-01 8.08752596e-01
1.99567392e-01 -1.31892692e-02 1.13938995e-01 -9.32845846e-02
-1.33355424e-01 -2.30079144e-01 -1.39160538e+00 6.42015755e-01
-1.70686141e-01 5.16852498e-01 2.87455022e-02 -1.52300787e+00
1.31470382e+00 2.99134463e-01 6.98964775e-01 -4.53422248e-01
4.47872043e-01 2.63612896e-01 6.55121431e-02 -5.92261255e-01
3.34216617e-02 -3.07975709e-02 -3.00706565e-01 6.20654106e-01
-2.79858083e-01 -1.86482057e-01 -1.15383342e-01 -7.05378801e-02
1.18117654e+00 -3.78913432e-01 5.26436627e-01 -2.51837164e-01
9.84099269e-01 -9.63743031e-02 4.22787786e-01 7.79037297e-01
-8.36540684e-02 -1.72140732e-01 4.66025651e-01 -6.04944646e-01
-6.64301336e-01 -1.15913832e+00 -1.35234252e-01 7.54418433e-01
-7.24217072e-02 -5.81906363e-02 -4.84112114e-01 -4.03030127e-01
-4.72550541e-02 6.37722790e-01 -1.79195747e-01 -3.98939282e-01
-4.59791005e-01 -9.52799261e-01 5.45936882e-01 1.70802027e-01
3.96661162e-01 -1.22130096e+00 -5.06065130e-01 3.32918048e-01
2.61431932e-01 -9.64719772e-01 1.64052039e-01 4.15645391e-01
-8.01030695e-01 -1.35406256e+00 8.59528184e-02 -5.74923396e-01
5.62268376e-01 -1.96843803e-01 7.92509139e-01 1.95395052e-01
-4.93415564e-01 3.50976229e-01 -3.74953866e-01 -8.75263572e-01
-7.93782711e-01 -2.40742207e-01 6.98332608e-01 3.27605605e-01
2.57984549e-01 -4.73288655e-01 -2.51271874e-02 3.37376952e-01
-1.18472397e+00 -5.14352918e-01 9.36695755e-01 8.53782415e-01
3.75577480e-01 6.58145249e-01 5.83177328e-01 -7.76614130e-01
6.92165434e-01 -4.04056400e-01 -1.02555442e+00 4.89182651e-01
-5.38988948e-01 1.01972245e-01 7.96850145e-01 -4.66164857e-01
-4.19930339e-01 3.66686867e-03 -4.47650068e-02 -2.70402104e-01
-2.84539849e-01 6.59542561e-01 -4.21217829e-01 -5.62618747e-02
7.91061819e-01 7.06306517e-01 1.46039128e-01 -3.92154098e-01
4.22319584e-02 7.65902102e-01 5.62872350e-01 -6.89225733e-01
1.08891964e+00 9.81097221e-02 5.84797204e-01 -1.09270704e+00
-8.15762043e-01 -4.76324230e-01 -5.42210937e-01 -4.43757445e-01
6.01001620e-01 -6.55944124e-02 -1.15549302e+00 5.28545439e-01
-9.03480053e-01 -9.71090794e-02 1.09534629e-01 5.59893191e-01
-4.42339420e-01 3.02127302e-01 -4.31220800e-01 -1.46309447e+00
-2.85043150e-01 -1.11734605e+00 7.69505978e-01 2.24240571e-02
-1.08661480e-01 -1.00505531e+00 -1.78285450e-01 1.87648252e-01
1.61336154e-01 3.68211091e-01 1.12475216e+00 -1.10117018e+00
-5.19995153e-01 -6.94628596e-01 4.18824404e-02 7.26821482e-01
1.93858296e-01 -1.35763576e-02 -8.60229671e-01 -3.48444164e-01
4.29735363e-01 -2.06623584e-01 6.49547637e-01 2.47130468e-01
1.47164762e+00 -5.20959675e-01 -2.93605000e-01 1.60530880e-01
1.27172875e+00 7.05476463e-01 5.42044759e-01 3.26183170e-01
3.23139161e-01 8.86906922e-01 7.20932364e-01 3.20291191e-01
-3.48958492e-01 4.31708366e-01 5.83334148e-01 1.70791090e-01
6.84736788e-01 -5.49357273e-02 7.23364592e-01 6.41876161e-01
1.87275395e-01 3.08919908e-03 -9.03144538e-01 -1.49167389e-01
-1.64210880e+00 -1.19311702e+00 3.17777283e-02 2.51003385e+00
5.77485681e-01 5.11010647e-01 4.02845703e-02 6.87487125e-01
7.21942663e-01 -2.27539361e-01 -6.02635264e-01 -4.02650684e-01
4.34531532e-02 1.55284882e-01 5.93226850e-01 4.90020066e-01
-1.36283720e+00 5.92365205e-01 5.81903601e+00 1.19226766e+00
-1.17552519e+00 -4.27062064e-01 8.63172948e-01 6.97780430e-01
-5.31283841e-02 7.36133987e-03 -8.34013224e-01 3.20057333e-01
9.08129871e-01 1.04089126e-01 4.61771369e-01 8.95090878e-01
2.38815770e-01 -1.75301477e-01 -1.01347971e+00 9.61832106e-01
2.45729275e-02 -1.16313314e+00 -1.82880774e-01 2.20755532e-01
3.97603333e-01 -4.31375921e-01 1.37475088e-01 2.11151123e-01
3.35597217e-01 -1.20942795e+00 4.16086078e-01 7.12314308e-01
4.05821353e-01 -8.05760324e-01 8.53236616e-01 9.03881848e-01
-9.25314128e-01 -4.10622954e-01 -1.30244359e-01 -1.30772024e-01
-1.49735108e-01 9.87989068e-01 -8.25085282e-01 6.38045549e-01
4.04879421e-01 3.65743577e-01 -3.41717511e-01 1.07833493e+00
-2.06913576e-01 9.47815835e-01 -2.09863886e-01 -4.54116255e-01
8.32052380e-02 -4.29207891e-01 8.60456705e-01 8.46951902e-01
7.93983638e-02 1.30199432e-01 5.42598069e-01 5.41372836e-01
3.25829804e-01 3.25782187e-02 -5.35332143e-01 -1.47198111e-01
5.46489894e-01 1.10627627e+00 -8.30571115e-01 -1.06848180e-01
-1.98103413e-02 2.20222816e-01 -2.85249472e-01 1.48553818e-01
-5.26757121e-01 -2.95400769e-01 6.26848996e-01 3.13131399e-02
-1.50575817e-01 -3.91298294e-01 -4.78596091e-01 -7.45962977e-01
-2.44657025e-01 -1.10965765e+00 6.09447300e-01 -3.18307072e-01
-1.48896706e+00 5.74206531e-01 1.02420352e-01 -1.32556856e+00
-3.73502642e-01 -1.08690894e+00 -6.64820671e-01 6.92723751e-01
-1.09459388e+00 -7.37613261e-01 1.20801523e-01 5.72418690e-01
2.92521656e-01 -7.72526324e-01 7.56944239e-01 -1.01474456e-01
-7.01074839e-01 2.31015235e-01 1.57563150e-01 1.81124315e-01
3.28419358e-01 -1.15765882e+00 -1.27731130e-01 1.03171003e+00
2.47042194e-01 3.61284584e-01 9.68421817e-01 -6.60006464e-01
-1.58639228e+00 -9.90149379e-01 6.00855291e-01 -4.88436371e-01
9.42302048e-01 -3.18212628e-01 -7.57267296e-01 3.55630487e-01
-3.51642638e-01 -4.45568234e-01 5.53952873e-01 3.78023796e-02
-4.09119964e-01 -5.23324788e-01 -1.21676326e+00 4.58630383e-01
1.60826162e-01 -5.38790643e-01 -6.01932108e-01 6.60364568e-01
4.46548015e-01 1.74371209e-02 -8.01946700e-01 4.62268621e-01
3.70179564e-01 -8.10867727e-01 9.69935119e-01 -8.36799502e-01
-1.54534191e-01 -4.70569283e-01 -1.30689695e-01 -9.70718980e-01
-1.26792118e-01 -7.29020059e-01 -1.00020915e-01 9.26022768e-01
4.33080107e-01 -9.34493601e-01 7.00279355e-01 3.64302009e-01
1.64890513e-02 -7.87062645e-01 -9.02813613e-01 -1.10267437e+00
-3.76421735e-02 -8.38893175e-01 2.57069707e-01 8.54574323e-01
-5.48270270e-02 4.96441983e-02 -4.34206992e-01 4.26411599e-01
9.03935432e-01 5.86022716e-03 6.88601255e-01 -1.64189684e+00
-4.81385946e-01 -6.10294759e-01 -7.04106033e-01 -5.75240135e-01
8.58174860e-02 -5.79372108e-01 5.62229119e-02 -1.00028491e+00
-9.17321220e-02 -5.40279627e-01 -3.85475516e-01 3.58950257e-01
4.35927100e-02 -9.42265838e-02 -4.80799973e-02 3.65485758e-01
-1.79339603e-01 1.78195894e-01 6.67170942e-01 -2.01859355e-01
-7.97107890e-02 7.03273118e-01 -6.13721550e-01 8.42980802e-01
9.17294919e-01 -5.63987076e-01 -1.74662575e-01 2.99350470e-01
2.53247201e-01 2.35867620e-01 5.20691156e-01 -9.27879453e-01
4.48694885e-01 -5.08054972e-01 4.82554317e-01 -3.04274589e-01
2.52470255e-01 -8.50242138e-01 -1.84311755e-02 9.79687870e-01
-2.11728022e-01 1.78816885e-01 -1.03423461e-01 4.87648845e-01
-2.77110934e-01 -6.08352840e-01 7.42583513e-01 7.04793110e-02
-7.64977694e-01 3.14237505e-01 -8.34759176e-01 -1.72071710e-01
1.11105168e+00 -3.20898890e-02 -1.86451063e-01 -4.58714426e-01
-5.67840815e-01 1.72194690e-01 -1.73233449e-01 3.54835689e-01
7.18821943e-01 -8.25155199e-01 -6.46062255e-01 2.65020221e-01
-1.35870531e-01 -3.32897305e-01 -1.44413218e-01 7.94199228e-01
-1.68709680e-01 5.70403337e-01 1.23357654e-01 -4.14199173e-01
-1.06307256e+00 6.20310962e-01 1.94213912e-01 -4.07141805e-01
-5.65519463e-03 6.19253337e-01 -4.07576859e-02 -2.61461705e-01
3.43653530e-01 4.81469817e-02 -3.01537126e-01 -3.14065278e-01
7.46684372e-01 3.78743976e-01 -1.80959273e-02 -4.12442535e-01
-3.59812617e-01 4.50029612e-01 2.35321432e-01 -1.65718351e-03
1.32023931e+00 4.44802940e-01 -2.45635182e-01 4.18450326e-01
7.69932747e-01 -1.25317737e-01 -7.35973954e-01 -2.16820329e-01
6.85492992e-01 -1.08996965e-01 -4.56584692e-02 -8.28549385e-01
-6.32234573e-01 7.54886210e-01 5.69922984e-01 6.57990277e-01
1.27493298e+00 -5.24607897e-02 4.79767203e-01 8.02283585e-01
7.28063107e-01 -9.79870498e-01 2.26765305e-01 5.76412976e-01
8.10060441e-01 -1.03738129e+00 3.86483185e-02 -3.25984299e-01
-4.90948439e-01 1.41914606e+00 2.61425346e-01 -1.61861837e-01
1.15768027e+00 4.59368557e-01 -1.53790653e-01 -1.04463622e-01
-4.73293453e-01 -5.24766035e-02 4.18539703e-01 5.92577159e-01
8.14736634e-02 4.52171043e-02 -2.85315514e-02 5.62226474e-01
-1.59231767e-01 -3.56310040e-01 1.83878958e-01 9.49891984e-01
-9.79107678e-01 -1.31900918e+00 -8.13673079e-01 7.09825158e-01
-3.83169115e-01 4.56567435e-03 -4.57706481e-01 6.71846211e-01
4.50179800e-02 1.12046611e+00 -3.19349617e-01 -6.79217994e-01
2.69466698e-01 -6.22252598e-02 1.59372732e-01 -3.70518267e-01
-3.56993139e-01 -1.94135919e-01 -8.32068399e-02 -3.22678626e-01
-1.56639650e-01 -8.30968380e-01 -8.62946749e-01 -1.64458960e-01
-3.78840059e-01 5.51340878e-01 8.48859608e-01 1.33258760e+00
-2.37106785e-01 1.98872045e-01 1.09677005e+00 -4.26657975e-01
-9.82245028e-01 -8.77772748e-01 -8.35284233e-01 -7.01066405e-02
2.73122657e-02 -6.41525209e-01 -4.95244086e-01 -1.54194817e-01] | [7.547479629516602, 2.6940598487854004] |
e3c4cdde-b241-4691-b692-5e10a0c5ca0f | estimating-uncertainty-of-earthquake-rupture | 1911.09660 | null | https://arxiv.org/abs/1911.09660v2 | https://arxiv.org/pdf/1911.09660v2.pdf | Estimating uncertainty of earthquake rupture using Bayesian neural network | Bayesian neural networks (BNN) are the probabilistic model that combines the strengths of both neural network (NN) and stochastic processes. As a result, BNN can combat overfitting and perform well in applications where data is limited. Earthquake rupture study is such a problem where data is insufficient, and scientists have to rely on many trial and error numerical or physical models. Lack of resources and computational expenses, often, it becomes hard to determine the reasons behind the earthquake rupture. In this work, a BNN has been used (1) to combat the small data problem and (2) to find out the parameter combinations responsible for earthquake rupture and (3) to estimate the uncertainty associated with earthquake rupture. Two thousand rupture simulations are used to train and test the model. A simple 2D rupture geometry is considered where the fault has a Gaussian geometric heterogeneity at the center, and eight parameters vary in each simulation. The test F1-score of BNN (0.8334), which is 2.34% higher than plain NN score. Results show that the parameters of rupture propagation have higher uncertainty than the rupture arrest. Normal stresses play a vital role in determining rupture propagation and are also the highest source of uncertainty, followed by the dynamic friction coefficient. Shear stress has a moderate role, whereas the geometric features such as the width and height of the fault are least significant and uncertain. | ['Md Mesbah Uddin', 'Sabber Ahamed'] | 2019-11-21 | null | null | null | null | ['small-data'] | ['computer-vision'] | [-4.36959594e-01 -3.56053561e-01 2.07661673e-01 2.41998062e-02
-7.16701329e-01 -2.64229387e-01 2.63747960e-01 1.05633098e-03
-3.43301594e-01 1.07432926e+00 1.95330605e-01 -3.44295114e-01
-5.87243915e-01 -9.89436328e-01 -7.65199423e-01 -1.22294068e+00
-2.59614617e-01 3.66745323e-01 6.17942393e-01 -2.91175038e-01
5.07119894e-01 5.54209769e-01 -1.20950556e+00 -2.08136752e-01
6.74874127e-01 8.96857023e-01 2.57797062e-01 3.81334722e-01
1.09414659e-01 3.89189571e-01 -7.13180602e-01 8.72143432e-02
-2.21267447e-01 1.07034408e-01 -4.10900950e-01 -7.15146244e-01
-6.17148578e-01 -4.51987207e-01 -2.67446220e-01 1.03449929e+00
8.21639657e-01 1.52725130e-01 1.03782403e+00 -7.34175146e-01
-4.11432475e-01 1.01318538e+00 -7.52869725e-01 2.67991066e-01
-1.20256856e-01 1.08055167e-01 2.16904268e-01 -9.09901142e-01
-1.16388276e-02 1.20405185e+00 9.91884947e-01 1.40442669e-01
-9.69635248e-01 -5.48936725e-01 -5.06081402e-01 1.32566631e-01
-1.63672245e+00 -1.00750983e-01 1.01338446e+00 -4.65110958e-01
5.74891329e-01 -9.23257843e-02 3.26246858e-01 1.14993453e+00
7.99466789e-01 1.01713531e-01 8.29452991e-01 -4.96524721e-01
4.65083838e-01 -3.31968695e-01 1.58417106e-01 1.63239315e-02
7.45228648e-01 3.10610980e-01 -2.84519583e-01 -3.24523389e-01
8.95625889e-01 1.93910077e-02 -2.02831045e-01 4.98119712e-01
-7.41229296e-01 9.74678516e-01 3.05542171e-01 3.66010547e-01
-5.04187226e-01 3.04776251e-01 2.80771822e-01 -2.24316552e-01
-2.20124959e-03 3.55018377e-01 -4.81412500e-01 -3.56439173e-01
-8.28460693e-01 5.15233934e-01 8.21685731e-01 1.57021165e-01
2.41744995e-01 5.95956504e-01 1.90113336e-01 8.31909537e-01
5.89163661e-01 9.75473464e-01 2.89629430e-01 -7.45652735e-01
3.03476661e-01 7.53615126e-02 2.13737696e-01 -1.33511603e+00
-4.15277213e-01 -2.19415113e-01 -1.27110863e+00 2.09501997e-01
6.40344858e-01 -6.56655192e-01 -8.62510026e-01 1.49104595e+00
-1.36208564e-01 1.54464975e-01 8.31146538e-02 8.63040805e-01
7.72553563e-01 9.22696471e-01 1.57842383e-01 -6.56825900e-02
1.30306256e+00 -3.33093368e-02 -6.19063318e-01 -1.56007990e-01
1.74161091e-01 -7.25594163e-01 4.78738874e-01 4.22519982e-01
-1.04768908e+00 -3.45813423e-01 -1.10471761e+00 8.24855983e-01
-2.26705685e-01 -2.52711564e-01 5.34682751e-01 5.18565118e-01
-5.65175533e-01 7.50004172e-01 -9.10966992e-01 1.56348601e-01
1.15543529e-01 6.03550747e-02 -1.01531424e-01 -1.35915011e-01
-2.00313187e+00 1.24110270e+00 7.09235847e-01 7.59060442e-01
-7.53420949e-01 -4.76671666e-01 -6.28420949e-01 1.67357653e-01
1.40530854e-01 -2.87029207e-01 8.56045783e-01 4.90497872e-02
-1.16919374e+00 -1.96343422e-01 2.26455107e-01 -2.00231418e-01
3.59179229e-01 -2.29337737e-01 -3.68377179e-01 6.26347736e-02
-1.73724264e-01 -2.94217244e-02 2.36762449e-01 -1.50474167e+00
-1.34075910e-01 -2.63152923e-03 -4.50248152e-01 -2.40354910e-01
-3.04072797e-02 1.27029642e-01 1.15539664e-02 -7.91823804e-01
5.17044544e-01 -7.36275375e-01 -3.99551719e-01 -7.76031852e-01
-5.44305921e-01 -2.42950931e-01 8.44224155e-01 -1.02088916e+00
1.43006396e+00 -1.98729730e+00 -3.00017655e-01 7.37716317e-01
-1.66028827e-01 9.76967588e-02 6.75911725e-01 6.34165227e-01
-7.41325989e-02 4.10698116e-01 -6.59095645e-01 4.70695019e-01
-1.15684457e-01 2.62633622e-01 -1.22833446e-01 1.89364806e-01
2.30054185e-01 3.12292755e-01 -5.18279493e-01 -4.39634770e-01
-2.51037240e-01 4.84437883e-01 -3.20449881e-02 1.31241888e-01
2.62649029e-01 -2.55123489e-02 -8.06047559e-01 7.96560228e-01
8.82525265e-01 1.36694446e-01 -2.92985439e-01 -3.04135740e-01
-1.50667489e-01 -4.84443903e-01 -1.65669990e+00 7.51136780e-01
1.50151094e-02 2.83938497e-01 -2.82011777e-01 -1.05752206e+00
1.52091634e+00 6.33463204e-01 1.71903834e-01 -1.55329958e-01
2.61905968e-01 4.32758003e-01 2.70582616e-01 -1.09522378e+00
2.63526708e-01 -3.02310020e-01 -2.81142920e-01 3.05643827e-01
-3.52978647e-01 -3.49097311e-01 5.12437113e-02 -1.20868191e-01
8.96890402e-01 -8.29145014e-02 -1.37380511e-01 -2.60956317e-01
-1.07974345e-02 -2.23758981e-01 9.74484026e-01 9.45678294e-01
9.65785310e-02 8.07608783e-01 8.60219181e-01 -4.35985565e-01
-1.12498868e+00 -1.20125294e+00 -2.62982070e-01 1.83282360e-01
9.16766524e-02 4.75074291e-01 -5.84842384e-01 1.84019879e-01
5.51362447e-02 8.35166931e-01 -4.76012349e-01 -3.37237328e-01
-6.14423335e-01 -1.37294543e+00 7.21457303e-01 9.70716238e-01
5.66580951e-01 -1.01042688e+00 -4.91501510e-01 3.98289353e-01
-2.75088400e-01 -6.22465611e-01 2.19637558e-01 2.57761180e-01
-1.02688360e+00 -8.59282911e-01 -5.06645441e-01 -5.16817570e-01
3.84313434e-01 -2.96583027e-01 8.27675462e-01 1.42810747e-01
-5.80251701e-02 -3.53136003e-01 -1.82854161e-01 -6.90318048e-01
-3.66731972e-01 -1.48735240e-01 7.31339529e-02 -4.41864520e-01
2.00015958e-02 -5.62230289e-01 -4.32199240e-01 4.35716122e-01
-1.01866102e+00 -5.58556616e-01 6.72907472e-01 9.48969126e-01
3.37863743e-01 8.44206393e-01 1.01569307e+00 -4.13229048e-01
8.29742789e-01 -8.34040821e-01 -4.50039238e-01 1.05748646e-01
-3.60854983e-01 -1.80523992e-01 2.69748271e-01 -4.65313107e-01
-1.31464601e+00 -3.75517607e-01 -1.72337115e-01 -3.04023534e-01
-2.75111914e-01 1.23669767e+00 -2.48975068e-01 3.04351658e-01
5.99752426e-01 -1.29713118e-01 6.65643886e-02 -4.59835470e-01
-5.11431515e-01 5.16858816e-01 4.96908039e-01 -8.87550771e-01
6.65735304e-01 -7.28325695e-02 3.24128956e-01 -8.58502805e-01
-3.93050104e-01 1.12015307e-01 -3.69674027e-01 -4.95123804e-01
7.73357749e-01 -5.00758529e-01 -7.88284123e-01 8.50997031e-01
-1.17402518e+00 -3.50641012e-02 3.03762853e-01 1.13175082e+00
-1.47870436e-01 1.58955559e-01 -7.96258271e-01 -1.21533608e+00
-1.37393996e-01 -1.17505872e+00 2.48324841e-01 7.50168324e-01
-1.24625422e-01 -9.16279137e-01 -2.21727386e-01 6.23353533e-02
7.20313311e-01 6.92613006e-01 1.08542442e+00 -6.90902650e-01
-1.50193572e-01 -5.07197022e-01 -1.94295466e-01 3.10542792e-01
-1.67652294e-01 5.98641098e-01 -7.41313994e-01 7.41102919e-02
2.52314031e-01 -3.31143104e-02 7.27202713e-01 1.04018760e+00
9.87929940e-01 3.74579504e-02 -2.47699946e-01 9.66242477e-02
1.64889860e+00 6.03059888e-01 9.98543262e-01 5.47293663e-01
5.49985111e-01 6.13720119e-01 4.12119567e-01 5.03516674e-01
1.22568190e-01 1.56350672e-01 4.85627204e-01 1.09680571e-01
4.31370616e-01 3.25799286e-02 1.43654332e-01 9.36331570e-01
-5.71304798e-01 -2.54022181e-01 -1.52275538e+00 5.38389027e-01
-1.66795051e+00 -1.17771423e+00 -5.81167996e-01 2.14275312e+00
9.63236928e-01 6.88082039e-01 -2.78133750e-01 4.59787518e-01
9.20308888e-01 -2.64261842e-01 -1.87035531e-01 -4.93912131e-01
-4.46824193e-01 -1.49100661e-01 6.35822356e-01 3.12229365e-01
-9.36767697e-01 2.42929488e-01 6.36464310e+00 9.98063624e-01
-1.11383545e+00 -4.02261704e-01 8.82368505e-01 3.89712363e-01
-2.50916719e-01 1.16812564e-01 -8.40686321e-01 1.00877583e+00
1.05085051e+00 1.55502379e-01 -1.26918584e-01 5.33674240e-01
7.13418603e-01 -7.52930701e-01 -4.24197614e-01 5.27812719e-01
-2.60050297e-01 -1.19206786e+00 -2.83278972e-01 -1.77428707e-01
6.22562945e-01 -2.74454534e-01 -3.23016852e-01 9.59940255e-02
3.05363357e-01 -1.06761992e+00 6.57020569e-01 1.31335938e+00
3.13909620e-01 -1.10680413e+00 1.77862191e+00 2.88338333e-01
-7.25199997e-01 -1.20945238e-01 -3.23074460e-01 -6.82348683e-02
5.37705064e-01 1.06870723e+00 -6.06953621e-01 6.11597836e-01
1.18208444e+00 -8.05458277e-02 -1.26701549e-01 1.29972446e+00
-2.34129325e-01 1.36015058e+00 -7.11018443e-01 -1.92672178e-01
2.16747895e-01 -2.24530220e-01 6.32022560e-01 9.83616531e-01
6.71484590e-01 2.14274257e-01 -2.03110445e-02 6.86406851e-01
4.06717628e-01 -4.04854953e-01 -4.79735672e-01 4.17292565e-01
1.14530826e+00 6.85613930e-01 -7.65209794e-01 1.70981679e-02
-2.00516619e-02 -2.52258480e-01 -4.11304176e-01 5.62996924e-01
-8.45235527e-01 -6.95120454e-01 6.50727227e-02 1.00152157e-01
3.80660683e-01 -2.43218869e-01 -7.34015703e-01 -2.21549153e-01
-2.77913045e-02 -4.02450562e-01 2.08383113e-01 -9.84588206e-01
-1.37052405e+00 2.41787657e-01 5.07599950e-01 -9.03893590e-01
-2.38412425e-01 -5.47702849e-01 -1.17682636e+00 1.21564031e+00
-1.04596376e+00 -8.27567220e-01 -5.54728843e-02 3.16176452e-02
1.39708236e-01 -1.65301442e-01 6.57244444e-01 1.77462623e-01
-8.04034054e-01 2.33586773e-01 4.77017790e-01 4.27527457e-01
4.92918879e-01 -9.25227106e-01 -1.74208358e-01 7.33379900e-01
-1.03671622e+00 6.50258362e-01 1.14929616e+00 -1.03741109e+00
-1.12756467e+00 -6.40861392e-01 7.26329088e-01 1.54032242e-02
7.39525795e-01 1.99986190e-01 -1.44446826e+00 -7.96409845e-02
-5.50938323e-02 -2.31875896e-01 5.05135238e-01 -7.24034086e-02
2.17793882e-01 -1.10265262e-01 -1.17741537e+00 5.59922338e-01
-1.58108864e-02 1.09050080e-01 -8.34473431e-01 -1.92596376e-01
4.62939620e-01 -1.68333948e-01 -1.33875322e+00 7.17280865e-01
6.15431309e-01 -6.72651291e-01 9.93008316e-01 -3.68197918e-01
6.74320221e-01 -2.41120592e-01 -2.57948071e-01 -1.17710447e+00
-3.01745951e-01 -3.09992999e-01 5.19726798e-02 1.61686623e+00
6.00917339e-01 -6.56897724e-01 5.33792913e-01 8.80769789e-01
-1.66138798e-01 -1.15194964e+00 -9.79273200e-01 -7.61401236e-01
3.53837371e-01 -4.09586161e-01 5.97909391e-01 7.80354857e-01
-3.63188058e-01 -1.17678404e-01 -2.83255607e-01 3.19987029e-01
8.22703362e-01 -4.57502693e-01 1.21038787e-01 -1.41149175e+00
9.60129797e-02 -4.74586308e-01 -1.10093534e-01 -1.80809721e-01
-4.13805783e-01 9.37712379e-03 3.89625460e-01 -1.43626523e+00
8.45169201e-02 -7.59132087e-01 -3.77075911e-01 3.20997506e-01
-3.67784828e-01 -1.70645624e-01 -3.14622462e-01 2.81831145e-01
4.58903879e-01 4.90232289e-01 9.12841022e-01 6.00119159e-02
-9.18362476e-03 1.90067440e-01 -4.84406620e-01 1.06660140e+00
1.18479764e+00 -7.19369113e-01 -9.78193283e-02 -4.21623498e-01
3.62820327e-01 6.15840554e-01 3.55655640e-01 -8.05849671e-01
4.16796625e-01 -4.35247213e-01 6.33603871e-01 -9.43694532e-01
3.79950702e-02 -5.62871873e-01 3.73069853e-01 4.47600126e-01
2.00504020e-01 3.22190821e-01 2.08441600e-01 5.67704320e-01
-3.59833598e-01 -9.28178906e-01 7.26474583e-01 -5.48864752e-02
-2.53587246e-01 -4.10478301e-02 -6.81191206e-01 -5.81752732e-02
7.18365967e-01 -3.86314690e-01 -1.93405412e-02 -2.31553704e-01
-6.21392787e-01 1.88957334e-01 -8.53067711e-02 -8.57450999e-03
6.99534237e-01 -1.23971450e+00 -1.06580412e+00 -9.04550999e-02
-5.42600751e-01 3.50395381e-01 6.76516652e-01 7.20508754e-01
-8.42596948e-01 -4.89710905e-02 -2.05384362e-02 -4.93069887e-01
-7.07148731e-01 -2.16378327e-02 2.15518281e-01 -9.37238559e-02
-3.23571980e-01 8.73260856e-01 -4.13842469e-01 3.90783548e-02
5.26979156e-02 4.99405041e-02 -6.12509549e-01 1.87587202e-01
3.83562654e-01 7.82494247e-01 -7.61876395e-03 -3.68787050e-01
-2.11080015e-01 5.76126456e-01 1.64001822e-01 -7.81744942e-02
1.66485441e+00 6.79939687e-02 -2.57382363e-01 6.01504564e-01
5.83536625e-01 -2.33916387e-01 -1.17295659e+00 3.87827754e-02
1.19807132e-01 -1.54619321e-01 2.75303721e-01 -7.03828454e-01
-9.76045370e-01 6.72431886e-01 4.22705323e-01 3.72208148e-01
8.03666115e-01 -1.55690789e-01 6.57726288e-01 2.56988466e-01
8.57850611e-02 -1.32192016e+00 -3.00296903e-01 7.21449792e-01
9.63871539e-01 -1.10334742e+00 8.06975588e-02 -5.27935810e-02
-4.86445099e-01 1.46289992e+00 4.44009483e-01 -1.96648523e-01
1.09204447e+00 8.40374708e-01 -1.20421842e-01 -3.26062888e-02
-4.33039308e-01 6.56105280e-01 -9.76509154e-02 3.04126203e-01
2.90104389e-01 6.69649094e-02 -4.57166433e-01 1.10037887e+00
-1.23284236e-01 -1.21516988e-01 5.61747015e-01 1.04979157e+00
-6.60258532e-01 -6.20508611e-01 -1.16916800e+00 5.36741674e-01
-7.47272134e-01 1.42765436e-02 3.81165087e-01 7.83983290e-01
4.01424523e-03 9.67432797e-01 4.85538915e-02 -2.99630195e-01
2.57665396e-01 -1.22781046e-01 -1.31525621e-01 5.46295755e-02
-2.28811547e-01 1.67097971e-01 1.34443283e-01 1.14328451e-01
-1.04339570e-02 -8.01276028e-01 -1.39642799e+00 -5.32383740e-01
-7.36721635e-01 5.52461445e-01 9.18930233e-01 1.01880217e+00
-1.65746838e-01 5.95239222e-01 6.99469686e-01 -8.73214364e-01
-7.28510141e-01 -1.27214766e+00 -7.92021632e-01 1.14697758e-02
-2.21548438e-01 -1.09479368e+00 -7.40406632e-01 1.02653995e-01] | [6.386481761932373, 3.0308454036712646] |
6e77fa82-ddd4-45ea-b9e2-a13385aaffb7 | ynu-hpcc-at-semeval-2019-task-9-using-a-bert | null | null | https://aclanthology.org/S19-2224 | https://aclanthology.org/S19-2224.pdf | YNU-HPCC at SemEval-2019 Task 9: Using a BERT and CNN-BiLSTM-GRU Model for Suggestion Mining | Consumer opinions towards commercial entities are generally expressed through online reviews, blogs, and discussion forums. These opinions largely express positive and negative sentiments towards a given entity,but also tend to contain suggestions for improving the entity. In this task, we extract suggestions from given the unstructured text, compared to the traditional opinion mining systems. Such suggestion mining is more applicability and extends capabilities. | ['Xue-jie Zhang', 'Jin Wang', 'Ping Yue'] | 2019-06-01 | null | null | null | semeval-2019-6 | ['suggestion-mining'] | ['natural-language-processing'] | [-6.46372318e-01 5.34763932e-01 -6.84793591e-01 -6.67266667e-01
-3.41345556e-03 -7.35690176e-01 5.51790118e-01 7.47574747e-01
-4.92547065e-01 1.01646113e+00 5.10525167e-01 -3.05410981e-01
3.42932135e-01 -9.22102511e-01 -1.04616508e-01 -3.08162093e-01
1.29597023e-01 -1.52753014e-02 3.03074300e-01 -6.34513974e-01
1.05810797e+00 -9.72296521e-02 -1.13003910e+00 4.56912398e-01
9.69609797e-01 1.06396103e+00 -2.03794166e-01 9.90240648e-02
-7.72798657e-01 9.18320477e-01 -7.64812648e-01 -1.25363827e+00
-1.35342136e-01 6.53958470e-02 -3.50083619e-01 3.53467315e-01
-8.18129349e-03 -6.64012507e-02 4.34021056e-01 1.32991791e+00
9.82573777e-02 7.19416663e-02 5.88102758e-01 -8.85677218e-01
-1.06307662e+00 7.92986095e-01 -7.65299916e-01 1.71468422e-01
4.96877909e-01 -4.78971332e-01 1.43390501e+00 -1.43064082e+00
6.59467459e-01 1.07058656e+00 2.41463155e-01 1.39795065e-01
-2.24763259e-01 -5.40400624e-01 8.65510404e-01 -1.40714079e-01
-9.84753728e-01 -1.23115294e-01 6.10787332e-01 -2.63717473e-01
9.38713491e-01 1.44036666e-01 9.15261269e-01 3.31370562e-01
5.10271966e-01 9.99483824e-01 7.28042006e-01 -2.82712609e-01
5.81060112e-01 1.14834404e+00 6.03051305e-01 2.38709152e-01
5.57777882e-01 -8.73101354e-01 -6.52016580e-01 -5.37551701e-01
-1.13069534e-01 4.98996824e-01 -7.50108361e-02 8.87021944e-02
-5.24447024e-01 1.05971467e+00 6.35388941e-02 2.37898946e-01
-7.35828698e-01 -5.86989403e-01 5.90738714e-01 5.73828459e-01
1.23785913e+00 5.82896829e-01 -1.06059635e+00 1.81813270e-03
-2.36224934e-01 1.75343931e-01 1.30205631e+00 6.93060696e-01
1.22313118e+00 -2.76683182e-01 3.59374374e-01 8.61523092e-01
9.00226891e-01 5.23961067e-01 8.64987493e-01 -3.96232933e-01
3.55234444e-01 1.21808839e+00 3.63709867e-01 -1.50616920e+00
-9.44961142e-03 -4.77367342e-01 -4.17055339e-01 -1.42328426e-01
-7.40039885e-01 -9.17138040e-01 -3.61168891e-01 8.46214294e-01
5.04095912e-01 -7.82859921e-01 3.14235032e-01 7.34068513e-01
1.23791909e+00 7.30762780e-01 1.43523008e-01 -6.97160304e-01
1.09691358e+00 -1.17498720e+00 -1.21542299e+00 -2.93692559e-01
8.66056800e-01 -1.01257277e+00 5.25837302e-01 6.69172704e-01
-8.73473823e-01 -3.91240150e-01 -9.13288236e-01 2.36634135e-01
-6.48798227e-01 2.93543339e-01 9.89834905e-01 6.47466540e-01
-7.56288111e-01 4.12265629e-01 -1.76970556e-01 -1.47624806e-01
2.35762104e-01 4.11829680e-01 -3.38026792e-01 3.61206144e-01
-1.25404382e+00 7.14605391e-01 -9.56747234e-02 2.79572815e-01
5.04790783e-01 -8.98569897e-02 -9.58212137e-01 -1.26757279e-01
3.11424345e-01 -3.42207432e-01 1.36723304e+00 -1.39287853e+00
-1.44047523e+00 4.13675845e-01 -4.60701972e-01 -2.87050933e-01
-2.66167998e-01 -5.23695409e-01 -1.09447932e+00 -2.63179094e-01
3.46680433e-01 2.18278736e-01 6.19644582e-01 -8.26074719e-01
-1.17099297e+00 -5.22227466e-01 2.55408376e-01 3.49457473e-01
-8.12014818e-01 4.47719187e-01 -3.30089658e-01 -5.52650988e-01
6.41902685e-02 -6.65474474e-01 -7.31213629e-01 -3.88368875e-01
-4.61691648e-01 -9.11774635e-01 8.13501894e-01 -8.70223865e-02
1.50002074e+00 -1.73095953e+00 -5.08896053e-01 4.99485999e-01
1.98467374e-01 3.11408848e-01 3.18455577e-01 6.72773421e-01
2.48465121e-01 6.76273823e-01 5.48176765e-01 1.21165469e-01
3.11661661e-02 -9.24548954e-02 -4.46633965e-01 1.99687183e-01
2.19876781e-01 5.99864185e-01 -9.59108770e-01 -4.58931208e-01
-1.85395226e-01 -8.46509933e-02 -3.67433786e-01 -1.16614075e-02
-1.62982211e-01 -1.05480112e-01 -1.46309531e+00 8.50880086e-01
3.66667062e-01 -6.25550210e-01 3.68503928e-01 -1.40942574e-01
-2.84941703e-01 5.78802347e-01 -8.22892189e-01 8.68390620e-01
-2.18615979e-01 4.95100945e-01 -2.11623777e-02 -8.43742192e-01
1.38212895e+00 2.69061863e-01 4.11313415e-01 -5.83617687e-01
3.37843090e-01 5.31748176e-01 -3.18217650e-02 -6.89958155e-01
9.64701414e-01 -9.11911055e-02 6.52012080e-02 7.06057906e-01
-2.83094078e-01 -1.58348382e-02 7.05511630e-01 6.15241528e-01
6.24332547e-01 -3.21883202e-01 8.87261391e-01 -1.72036052e-01
7.96325386e-01 2.90414363e-01 5.89124441e-01 8.05631056e-02
9.50413942e-02 2.48699654e-02 4.45014805e-01 -3.26149613e-01
-3.75935346e-01 -1.45830154e-01 5.26647363e-03 9.27938521e-01
2.23715022e-01 -9.84043598e-01 -1.28867537e-01 -1.19205701e+00
2.13345155e-01 4.93180007e-01 -6.98963284e-01 4.30010587e-01
-8.74645561e-02 -6.65167511e-01 -6.05075359e-01 4.48667586e-01
2.55203843e-01 -9.40524399e-01 2.82750666e-01 2.48415381e-01
2.92130888e-01 -5.76627851e-01 -4.85360712e-01 3.32061462e-02
-9.56923842e-01 -1.03781021e+00 -7.48022914e-01 -7.55846977e-01
1.11643040e+00 3.83304000e-01 1.25817513e+00 3.39676380e-01
6.12892032e-01 1.88012272e-01 -1.00765836e+00 -9.12884414e-01
8.35910067e-02 1.37374744e-01 3.67574096e-01 -5.18326499e-02
1.06115973e+00 -2.97307491e-01 -7.32650220e-01 6.00276172e-01
-6.24216974e-01 -5.45994818e-01 7.38954127e-01 4.14644629e-01
3.91201466e-01 2.40010038e-01 1.13610864e+00 -1.66755891e+00
1.39022994e+00 -9.37921762e-01 -1.32738322e-01 2.18774915e-01
-1.36536050e+00 -2.22134516e-01 6.24777138e-01 -2.28046641e-01
-1.22323871e+00 -5.31592250e-01 -1.47963598e-01 5.16896844e-01
1.66272923e-01 1.23114693e+00 2.46455684e-01 1.80560804e-03
7.65786409e-01 -3.61915737e-01 -3.16044539e-01 -2.00198427e-01
2.71152258e-01 9.11458910e-01 -4.30843145e-01 -1.16388492e-01
4.42966163e-01 2.23196819e-01 -6.66159928e-01 -5.56465328e-01
-1.05655730e+00 -9.06506956e-01 -2.42074043e-01 -2.83300340e-01
1.24550648e-01 -7.92544067e-01 -5.79842508e-01 -1.02444150e-01
-1.09880257e+00 6.33210361e-01 -1.99764937e-01 6.80359662e-01
4.05755699e-01 3.60651076e-01 -5.82326233e-01 -1.13371062e+00
-7.11384714e-01 -6.29231393e-01 4.34934288e-01 8.17237496e-01
-4.55576837e-01 -1.34019506e+00 3.01012337e-01 3.34180772e-01
5.52102387e-01 -3.91688764e-01 2.49419928e-01 -1.18667316e+00
1.46772623e-01 -6.34676814e-01 -3.10814325e-02 5.24675667e-01
6.18132949e-01 4.22351092e-01 -4.00760025e-01 7.31748417e-02
1.16694048e-01 -4.40716326e-01 6.67048037e-01 1.78553894e-01
4.97353286e-01 -7.84080207e-01 -5.38838089e-01 -5.72255671e-01
8.48454416e-01 3.58656913e-01 3.54585856e-01 5.44569671e-01
2.40762249e-01 1.06831431e+00 1.27860284e+00 7.21093774e-01
3.96561444e-01 -3.85197811e-02 4.42969948e-02 -9.53605846e-02
8.73240530e-01 2.86463965e-02 7.33788013e-01 1.41592467e+00
7.76391998e-02 -6.41624570e-01 -4.53075886e-01 6.27172649e-01
-2.05991888e+00 -6.41142488e-01 -4.00996000e-01 1.29033744e+00
6.69051349e-01 4.25624251e-01 -4.71084900e-02 -1.84619382e-01
6.43742561e-01 2.35699847e-01 -1.00969613e+00 -7.05286741e-01
-3.07564169e-01 -5.42627424e-02 4.78073657e-01 1.67986721e-01
-7.61568964e-01 8.28859389e-01 6.92974997e+00 5.75696230e-01
-1.07500982e+00 -3.18563312e-01 1.04270458e+00 2.73072720e-01
-1.00607395e+00 3.81359272e-02 -1.09127975e+00 1.98782802e-01
6.44019961e-01 -7.87175775e-01 -5.72023809e-01 1.31690407e+00
3.35952222e-01 -4.41428155e-01 -5.77539623e-01 5.06716251e-01
3.10887754e-01 -1.30117035e+00 2.93679181e-02 6.09134175e-02
1.14450669e+00 -1.73242763e-01 1.29698649e-01 3.11291367e-01
5.72133839e-01 -4.86360490e-01 3.46808225e-01 3.73332530e-01
-1.78423926e-01 -6.38195455e-01 1.42275488e+00 3.10980618e-01
-1.11934912e+00 8.68192315e-02 -6.09266222e-01 -5.00201106e-01
6.01162195e-01 1.15987611e+00 -8.74825001e-01 3.08935732e-01
7.92709172e-01 1.37181616e+00 -4.71617579e-01 6.32965684e-01
-7.23659694e-01 6.14698648e-01 7.41984323e-02 -9.59621131e-01
3.36782664e-01 -6.63850546e-01 6.53368756e-02 9.96365249e-01
2.54360974e-01 3.01840812e-01 1.53753683e-01 2.18097687e-01
-2.77466834e-01 1.13796723e+00 -4.88439500e-01 -5.35578728e-01
2.13292196e-01 1.62961066e+00 -5.83536923e-01 -6.26247466e-01
-8.51110339e-01 5.66656709e-01 8.02547410e-02 3.67202818e-01
-1.18691221e-01 -3.28407496e-01 2.91378796e-01 1.66670829e-01
2.90868074e-01 2.45803937e-01 -1.82632938e-01 -1.24717402e+00
3.65502357e-01 -8.16755831e-01 2.70724505e-01 -6.93758249e-01
-1.69691432e+00 6.38677657e-01 -5.37919998e-01 -1.43315136e+00
-2.26253957e-01 -8.17501724e-01 -1.13924813e+00 6.63593113e-01
-1.55748475e+00 -4.31605995e-01 3.65329534e-01 1.84836924e-01
6.55358553e-01 -4.37972039e-01 5.35870433e-01 1.83379024e-01
-4.43643898e-01 -6.74267486e-02 -2.34342232e-01 -1.99410692e-01
9.56482708e-01 -1.28703856e+00 2.92928249e-01 3.94004345e-01
-1.13080844e-01 1.32781959e+00 7.75509775e-01 -9.59140658e-01
-1.15105844e+00 -4.54368293e-01 1.17030430e+00 -9.96456891e-02
1.17640710e+00 5.05763233e-01 -6.65112317e-01 4.09783721e-01
8.56364131e-01 -3.01199198e-01 1.61671519e+00 7.42814779e-01
8.15791916e-03 -2.83895265e-02 -9.83411074e-01 5.81902504e-01
3.63540620e-01 -3.53715420e-01 -7.56748319e-01 6.83509052e-01
6.75184667e-01 2.05392227e-01 -9.38369930e-01 3.92403781e-01
7.02352285e-01 -9.27339315e-01 3.46577972e-01 -6.76112652e-01
4.31895733e-01 -4.80296016e-01 1.00042075e-01 -1.50975573e+00
1.01848908e-01 -4.08782333e-01 -1.55980349e-01 1.19608855e+00
1.29078913e+00 -6.84086978e-01 1.20707047e+00 1.35431516e+00
-1.05287194e-01 -1.12251580e+00 5.38698398e-02 8.48857313e-02
-3.41033041e-01 -6.88271672e-02 4.15171593e-01 1.11156619e+00
8.59216154e-01 9.39172983e-01 -1.51952207e-01 -2.57842630e-01
-9.75304991e-02 3.55195969e-01 7.23250926e-01 -1.35151041e+00
4.49972153e-02 -4.10098076e-01 -8.83263573e-02 -1.56206810e+00
-4.53406163e-02 -3.11715394e-01 -2.91863471e-01 -1.72263038e+00
-4.77850512e-02 -3.62215489e-01 -3.14464241e-01 1.11775532e-01
-3.09813231e-01 -1.18766226e-01 -5.53610802e-01 4.84972328e-01
-1.23233163e+00 5.33970058e-01 1.32046056e+00 -2.76867241e-01
-2.36253619e-01 4.24350947e-01 -1.50572026e+00 1.15004647e+00
9.20297086e-01 -4.07806545e-01 -4.85779107e-01 3.56330983e-02
1.61175513e+00 -3.12467545e-01 -7.80172110e-01 2.02958137e-01
6.51292682e-01 -3.91394585e-01 6.38668761e-02 -1.02922308e+00
-1.88504785e-01 -6.08145535e-01 -2.99688369e-01 1.05723910e-01
-3.85935277e-01 4.74142700e-01 -3.27955037e-01 7.22627640e-01
-9.16407347e-01 -4.30832207e-01 -2.24553794e-01 -2.46878847e-01
-7.80382395e-01 4.09262866e-01 -7.03309596e-01 -4.91910398e-01
5.91577649e-01 -1.30839795e-01 -4.26588595e-01 -7.78430343e-01
-6.96212351e-01 6.78492010e-01 3.72177750e-01 4.42439735e-01
7.10979819e-01 -1.15851319e+00 -5.83574593e-01 -3.44873726e-01
3.13133866e-01 -1.01861879e-01 -1.09342761e-01 7.34474838e-01
2.66360380e-02 2.16102898e-01 3.02475691e-01 6.47291467e-02
-1.14742863e+00 3.79285455e-01 -2.11384013e-01 -2.31573567e-01
8.37326646e-02 9.65662837e-01 1.30976915e-01 -4.87541646e-01
-6.68177828e-02 -6.14425205e-02 -1.41955698e+00 4.86071765e-01
9.10059512e-01 1.52578548e-01 6.82423934e-02 -5.96155703e-01
-1.93816870e-01 3.12048107e-01 -6.98915899e-01 3.70528772e-02
1.35389543e+00 -3.61892104e-01 -6.44114375e-01 6.83120012e-01
8.43449533e-01 8.56864452e-01 -2.50219107e-01 -3.43378365e-01
3.67580116e-01 -3.51057231e-01 -6.66064397e-02 -5.97218812e-01
-1.32635951e+00 1.87736467e-01 -2.42186815e-01 1.08099723e+00
5.46876669e-01 2.72756387e-02 6.43625140e-01 9.69716191e-01
2.56297916e-01 -1.70733094e+00 -3.52573059e-02 3.63252163e-01
7.62590349e-01 -1.52507770e+00 5.31102955e-01 -6.96020246e-01
-1.14065647e+00 1.36560297e+00 7.95170128e-01 -6.56813458e-02
1.36664796e+00 -1.12943359e-01 3.73899758e-01 -6.54022455e-01
-1.18167329e+00 6.54551685e-02 4.01402682e-01 1.44733921e-01
1.20882773e+00 -9.29925144e-02 -1.08370340e+00 9.80552077e-01
-1.32987216e-01 -2.89995193e-01 4.34598625e-01 9.91740227e-01
-9.47895527e-01 -1.40288568e+00 6.38924632e-03 9.01295602e-01
-8.30399930e-01 -3.62206280e-01 -9.49946702e-01 6.72710612e-02
-4.02142070e-02 1.67471230e+00 -2.41488382e-01 -3.25358331e-01
3.70707572e-01 -4.80562419e-01 -6.20558321e-01 -1.01305294e+00
-1.17500925e+00 1.21867046e-01 7.01939225e-01 -1.90671042e-01
-1.17079163e+00 -5.80389380e-01 -1.09578753e+00 -2.19157144e-01
-1.24713182e+00 1.09314024e+00 8.62578154e-01 9.95238364e-01
5.05473793e-01 2.30338965e-02 9.52605784e-01 -7.83428401e-02
-1.82111278e-01 -1.18165076e+00 -7.98804700e-01 2.78860301e-01
-6.30445480e-02 -3.71674955e-01 -6.57233417e-01 -1.86070278e-01] | [11.183854103088379, 6.775284290313721] |
6d910563-2a03-48c2-90f1-eabd0dfb20b8 | a-multi-task-network-for-joint-specular | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Fu_A_Multi-Task_Network_for_Joint_Specular_Highlight_Detection_and_Removal_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Fu_A_Multi-Task_Network_for_Joint_Specular_Highlight_Detection_and_Removal_CVPR_2021_paper.pdf | A Multi-Task Network for Joint Specular Highlight Detection and Removal | Specular highlight detection and removal are fundamental and challenging tasks. Although recent methods achieve promising results on the two tasks by supervised training on synthetic training data, they are typically solely designed for highlight detection or removal, and their performance usually deteriorates significantly on real-world images. In this paper, we present a novel network that aims to detect and remove highlights from natural images. To remove the domain gap between synthetic training samples and real test images, and support the investigation of learning-based approaches, we first introduce a dataset of 16K real images, each of which has the corresponding highlight detection and removal images. Using the presented dataset, we develop a multi-task network for joint highlight detection and removal, based on a new specular highlight image formation model. Experiments on the benchmark datasets and our new dataset show that our approach clearly outperforms the state-of-the-art methods for both highlight detection and removal. | ['Chunxia Xiao', 'Ping Li', 'Lei Zhu', 'Qing Zhang', 'Gang Fu'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['highlight-detection'] | ['computer-vision'] | [ 6.69789732e-01 -4.51613218e-01 3.00569564e-01 4.87524271e-03
-7.04291284e-01 -5.33615887e-01 6.00894690e-01 -2.31018085e-02
-3.54349375e-01 7.92724788e-01 -1.99677840e-01 9.45517328e-03
1.76512435e-01 -3.98818463e-01 -7.68347681e-01 -9.13816869e-01
-3.82583104e-02 -1.17819972e-01 6.50023937e-01 -1.01351924e-01
4.28307742e-01 4.94966000e-01 -1.80985188e+00 6.48656189e-01
1.00668192e+00 9.63541031e-01 2.50644177e-01 5.31501591e-01
-5.06605543e-02 9.39438522e-01 -9.31529641e-01 -8.62276480e-02
3.90778989e-01 -3.39084655e-01 -1.96125463e-01 2.14643270e-01
8.64256263e-01 -2.49430537e-01 -1.02588512e-01 9.84229565e-01
8.01570415e-01 1.85734525e-01 5.45260906e-01 -1.25859606e+00
-5.83311319e-01 4.23020750e-01 -9.14511800e-01 1.93470329e-01
7.56122544e-03 9.37722065e-03 6.72673047e-01 -1.23628044e+00
6.80770814e-01 6.81854725e-01 7.65111029e-01 1.72666550e-01
-1.32454264e+00 -8.06682110e-01 1.29124343e-01 2.97786057e-01
-1.25852585e+00 -4.49351698e-01 1.34684873e+00 -2.59027749e-01
4.35503662e-01 4.32545960e-01 5.55000722e-01 1.18268168e+00
4.79307733e-02 1.14029121e+00 1.69096982e+00 -6.15226865e-01
-1.64576881e-02 2.50564009e-01 -1.09544151e-01 4.96215492e-01
4.70466822e-01 2.62193859e-01 -9.03849423e-01 7.22697526e-02
5.14194250e-01 -1.50320828e-01 -6.75840318e-01 -6.23434067e-01
-1.17955756e+00 4.59797770e-01 3.54694605e-01 1.50632471e-01
-1.87248871e-01 -4.76053022e-02 5.65053999e-01 3.83605301e-01
7.41833150e-01 5.57487965e-01 -1.79507419e-01 4.16890353e-01
-1.46276724e+00 4.54146266e-01 6.52257502e-01 7.60453999e-01
6.40696049e-01 2.05332637e-01 -4.80691493e-01 1.02810168e+00
-4.21466500e-01 6.81375563e-01 2.58761197e-01 -3.74487698e-01
9.93117169e-02 2.84008950e-01 4.43367571e-01 -1.23801041e+00
-5.68649471e-01 -8.64923894e-01 -6.92830265e-01 5.01586914e-01
3.18101794e-01 -1.72300696e-01 -8.51368189e-01 1.38340414e+00
1.27239719e-01 4.22641963e-01 -2.75586993e-02 9.88870621e-01
9.69824731e-01 5.15139401e-01 -2.78908759e-01 -1.21024050e-01
8.93034399e-01 -1.50267625e+00 -7.96411991e-01 -1.24327995e-01
2.33300969e-01 -1.18747556e+00 1.03347254e+00 7.10871875e-01
-1.07005262e+00 -7.67805040e-01 -1.16387248e+00 -2.61701234e-02
-6.01209700e-01 6.76194489e-01 5.18287480e-01 5.97012639e-01
-9.73573267e-01 2.69020021e-01 -1.94653630e-01 -1.42435700e-01
2.41539061e-01 -1.44748554e-01 -1.41158998e-02 -5.61437830e-02
-7.81053722e-01 7.66752899e-01 4.38479036e-01 3.17565948e-01
-8.17872345e-01 -9.25052047e-01 -5.45171261e-01 1.37978464e-01
8.34657669e-01 -3.23766142e-01 9.40959096e-01 -1.08191836e+00
-1.36279976e+00 1.07183528e+00 1.90242514e-01 -4.86783445e-01
8.81847739e-01 -6.44096434e-01 -6.88260019e-01 1.97371230e-01
-6.52917251e-02 4.92678940e-01 1.42498243e+00 -1.88288581e+00
-6.98151469e-01 2.90798038e-01 -1.83511823e-01 -1.61785886e-01
-4.35123444e-01 2.41843760e-01 -4.89781410e-01 -9.82389092e-01
-3.90328705e-01 -7.13949442e-01 2.06502661e-01 9.92963314e-02
-8.64711404e-01 4.33670610e-01 1.24189734e+00 -5.35989344e-01
9.24875438e-01 -2.21311545e+00 -2.12344289e-01 1.26763552e-01
1.59589350e-01 8.16511571e-01 -5.44580221e-01 5.35569131e-01
-2.24055022e-01 -4.32077616e-01 -2.87313551e-01 -5.55745542e-01
-1.66885599e-01 -4.44464862e-01 -7.27029443e-01 3.49607050e-01
5.09100318e-01 7.08619475e-01 -7.92126715e-01 -3.86763960e-01
3.53180856e-01 6.09599173e-01 -1.70904845e-01 2.07524151e-01
-1.49596304e-01 1.79313079e-01 -2.61639468e-02 7.70063043e-01
1.19976115e+00 -1.44917533e-01 2.59703249e-02 -2.01978847e-01
-4.73049581e-01 -4.27748382e-01 -1.19415355e+00 1.26895320e+00
-4.21925336e-01 1.03661537e+00 5.91343082e-02 -7.90718257e-01
1.24486756e+00 -1.98654294e-01 3.33811820e-01 -9.10924852e-01
-2.11197394e-03 3.23003203e-01 -1.51141047e-01 -6.39134586e-01
7.25249887e-01 1.34580016e-01 2.46901840e-01 4.46923286e-01
-4.38918948e-01 -2.05190927e-01 4.05365914e-01 -1.49090085e-02
1.04230130e+00 2.78115898e-01 2.11118199e-02 -1.81969792e-01
8.65011573e-01 -7.20611736e-02 4.04815167e-01 8.50463033e-01
-1.08294137e-01 8.87082279e-01 4.06488091e-01 -5.04536808e-01
-8.80834818e-01 -1.13800585e+00 3.65607115e-03 1.23135173e+00
2.83215493e-01 -9.03470218e-02 -6.91647410e-01 -4.79969949e-01
2.46108249e-01 6.45681143e-01 -7.14614868e-01 -1.20758615e-01
-5.86922526e-01 -7.06985056e-01 3.93888563e-01 2.07821488e-01
7.40980625e-01 -1.37580836e+00 -1.08815730e+00 2.00460494e-01
-2.03197799e-03 -1.33936286e+00 -4.57892030e-01 2.59351879e-01
-3.06410044e-01 -1.37602282e+00 -1.10554218e+00 -8.05706263e-01
6.00812554e-01 1.02008152e+00 1.06677938e+00 2.88381755e-01
-7.35199153e-01 1.93288550e-01 -4.49473411e-01 -8.27311337e-01
-2.60252446e-01 -1.35418549e-01 -4.81699526e-01 3.98953766e-01
3.24206240e-02 -4.38783407e-01 -7.22377121e-01 3.26389760e-01
-1.25111425e+00 2.62919724e-01 9.34927344e-01 1.09220326e+00
4.60499674e-01 1.49642318e-01 3.60894233e-01 -1.18132365e+00
7.67998993e-01 2.19001725e-01 -6.59629107e-01 3.93048406e-01
-4.38771993e-01 -2.15385064e-01 7.72400081e-01 -5.72852910e-01
-1.34245992e+00 -4.02544066e-02 5.78277826e-01 -6.20441437e-01
-8.39723200e-02 3.65102053e-01 1.13393836e-01 -6.38665020e-01
7.45721042e-01 2.56951571e-01 -2.46385410e-01 -4.38385516e-01
3.14117819e-01 2.36378253e-01 8.58818114e-01 -2.60063529e-01
1.05333400e+00 9.11785781e-01 1.78980716e-02 -8.97245169e-01
-1.15262306e+00 -5.87482929e-01 -7.35700667e-01 -5.08278251e-01
2.68590987e-01 -8.79006088e-01 -2.40380615e-01 9.74097729e-01
-1.06753087e+00 -6.20911241e-01 -1.24802269e-01 -2.73621664e-03
-4.72258359e-01 4.63243932e-01 -2.30375290e-01 -7.72744536e-01
-4.46192443e-01 -7.80147195e-01 1.20541799e+00 2.28071958e-01
2.85475641e-01 -7.98526585e-01 5.45435436e-02 -1.04583263e-01
7.17252791e-01 7.69609272e-01 8.52354228e-01 -1.69190437e-01
-7.43798614e-01 -2.82394141e-01 -9.16474879e-01 4.86701459e-01
1.88340619e-01 2.66198456e-01 -1.26633823e+00 -3.33351552e-01
-1.13634370e-01 -2.15894029e-01 1.57091570e+00 3.42980802e-01
1.27757859e+00 3.36208314e-01 -4.30475205e-01 7.08733916e-01
1.70265162e+00 -1.49450200e-02 6.01943076e-01 6.95546210e-01
5.88149190e-01 8.46857667e-01 8.17016125e-01 4.78440583e-01
-1.27858773e-01 5.67273498e-01 4.98117685e-01 -7.84431934e-01
-6.36665821e-01 7.41777346e-02 1.70572594e-01 3.73001516e-01
-3.16220932e-02 -3.78536403e-01 -6.49238229e-01 4.76946443e-01
-1.86389923e+00 -9.78753626e-01 -3.31364661e-01 2.09431720e+00
4.57137883e-01 3.41628045e-01 1.83263034e-01 2.28058651e-01
8.76902699e-01 7.20739782e-01 -5.29244423e-01 -1.05228424e-01
-8.44721615e-01 2.82020241e-01 5.01943827e-01 2.60073990e-02
-1.42393291e+00 1.02062643e+00 6.35429430e+00 6.31049514e-01
-1.63492119e+00 -4.43744697e-02 4.74388361e-01 -2.62635291e-01
-4.06921953e-02 -4.77020472e-01 -4.36310589e-01 4.66213882e-01
4.48815644e-01 -6.40196577e-02 2.13077560e-01 5.64195752e-01
4.35323477e-01 -4.02045965e-01 -7.94189453e-01 1.11232960e+00
4.81020391e-01 -1.25946593e+00 -8.59298408e-02 -3.52220118e-01
1.25695956e+00 -1.45841211e-01 5.56606054e-01 1.60947546e-01
-2.33012155e-01 -6.78662658e-01 8.25847208e-01 7.92304397e-01
5.72055817e-01 -6.66029215e-01 6.06179178e-01 1.49323344e-02
-9.82167423e-01 -2.40265280e-01 -3.69671524e-01 2.57207632e-01
4.70197089e-02 9.57932830e-01 -4.95933712e-01 5.09965837e-01
5.67155361e-01 6.62351668e-01 -8.40174556e-01 1.71924686e+00
-3.97277087e-01 3.23949307e-01 -1.79949384e-02 -5.50786480e-02
2.52428442e-01 -1.48926511e-01 4.97826248e-01 1.77776802e+00
1.82048813e-01 -4.98504847e-01 2.33264670e-01 9.98371959e-01
-1.32987529e-01 2.55027533e-01 -4.64727789e-01 7.20021278e-02
1.56095043e-01 1.55556667e+00 -1.05195248e+00 -3.60661060e-01
-5.10630012e-01 1.02775347e+00 3.50678325e-01 5.83621621e-01
-9.85659659e-01 -6.66846871e-01 2.64047951e-01 -1.71252843e-02
5.83219588e-01 -1.05677359e-01 -2.84313232e-01 -9.06156123e-01
3.34599882e-01 -9.54671264e-01 2.38759406e-02 -9.81362402e-01
-1.05600476e+00 4.30973083e-01 -2.16719091e-01 -1.44580579e+00
2.42020711e-01 -8.52323771e-01 -1.11983120e+00 7.08157837e-01
-2.40340996e+00 -1.26603818e+00 -7.45280862e-01 3.83987844e-01
4.60077286e-01 -8.45783353e-02 4.32933629e-01 3.29825550e-01
-4.33393806e-01 4.30521548e-01 7.11325526e-01 -1.57270372e-01
1.37751651e+00 -1.11172950e+00 4.69837755e-01 1.03079987e+00
-1.27321817e-02 1.82528242e-01 9.11756516e-01 -5.22135317e-01
-1.14571679e+00 -1.18088961e+00 3.12778652e-01 -3.87848318e-02
5.78195333e-01 -5.48846185e-01 -1.04629230e+00 1.17355242e-01
2.66144514e-01 1.41396523e-01 4.56995994e-01 1.48713617e-02
-5.25187731e-01 -2.15389982e-01 -8.29991162e-01 4.97006178e-01
9.72227037e-01 -4.31385279e-01 -1.59813091e-01 5.88365734e-01
2.88951755e-01 -4.48233604e-01 -1.00020587e-01 4.68374819e-01
4.88802999e-01 -1.56813765e+00 9.70737278e-01 -2.13759750e-01
5.98989069e-01 -4.26320851e-01 3.24687719e-01 -1.47305930e+00
1.90784130e-02 -6.70961738e-01 -1.52202561e-01 9.79165852e-01
2.49554470e-01 -3.83597255e-01 8.04000258e-01 -1.53992862e-01
-2.61765689e-01 -6.52372956e-01 -4.30553645e-01 -9.55408037e-01
-3.95113975e-01 -1.77282870e-01 4.58372980e-01 7.99401581e-01
-6.76171124e-01 -6.04267418e-02 -7.32405543e-01 1.86965987e-01
8.18826854e-01 6.80642724e-01 1.10256422e+00 -1.24691463e+00
-6.60796911e-02 -6.09590888e-01 -8.29517767e-02 -5.83252370e-01
1.60527587e-01 -3.85222554e-01 3.10900688e-01 -1.30378950e+00
3.35838437e-01 -1.41025886e-01 -4.94572639e-01 2.86322415e-01
-3.82091343e-01 5.34629583e-01 2.89742172e-01 3.12803984e-02
-7.00281203e-01 5.74688435e-01 1.30046618e+00 -2.13102549e-01
-3.53320330e-01 -1.74352184e-01 -5.00285745e-01 6.97435796e-01
6.38997197e-01 -1.47786260e-01 -2.74153292e-01 -2.05963284e-01
9.98196006e-02 -4.31804240e-01 7.20188498e-01 -1.25943005e+00
1.78821981e-01 5.24349883e-02 5.83067060e-01 -1.06133306e+00
4.64597791e-01 -6.00823998e-01 -2.39958391e-01 3.66946965e-01
-3.44077408e-01 -2.07235858e-01 4.64771986e-01 5.20647228e-01
-3.75242800e-01 -6.55952170e-02 1.00998640e+00 7.50675201e-02
-9.23601508e-01 -7.93401450e-02 -1.65182158e-01 -4.29249667e-02
1.00075722e+00 -2.59301096e-01 -7.33743727e-01 -2.79805273e-01
-3.52497667e-01 1.87468342e-02 4.70489681e-01 4.32898939e-01
8.00914764e-01 -9.17500854e-01 -7.87391603e-01 2.24493071e-01
3.17125678e-01 -2.45535105e-01 4.83173043e-01 8.98050666e-01
-5.14833808e-01 2.21826136e-01 -5.48224032e-01 -1.87425882e-01
-1.50297415e+00 8.79534423e-01 2.74803370e-01 -1.84946001e-01
-7.34292746e-01 8.46378565e-01 5.34698904e-01 1.73982661e-02
2.63051271e-01 -3.38288397e-01 -1.04549058e-01 1.43189803e-01
6.90850139e-01 5.44986725e-01 7.69396424e-02 -4.08693969e-01
1.10544264e-01 5.91128826e-01 -4.05624300e-01 3.12183589e-01
1.38524401e+00 8.85753185e-02 -7.35243456e-03 2.81089664e-01
1.04564631e+00 3.08947444e-01 -1.36394465e+00 -4.37939018e-01
7.37631321e-02 -8.51067424e-01 -2.23089606e-02 -7.99302459e-01
-1.12114322e+00 8.05790305e-01 6.03271008e-01 4.73272324e-01
1.41867602e+00 -5.71143687e-01 6.38799787e-01 4.40631092e-01
1.74783003e-02 -1.43015218e+00 5.35933554e-01 2.68591672e-01
1.04049838e+00 -1.43420100e+00 2.08649963e-01 -8.87815475e-01
-3.87695968e-01 1.27658737e+00 5.74376702e-01 -4.51236308e-01
3.05212706e-01 3.43646705e-01 3.80688637e-01 -1.80734783e-01
-5.53252161e-01 -3.69359583e-01 4.35523659e-01 7.14065194e-01
3.51347744e-01 -1.33150011e-01 -1.63941428e-01 -2.90763900e-02
-2.59218216e-02 -1.84319243e-01 7.39367306e-01 8.09589744e-01
-6.54053748e-01 -8.86950135e-01 -6.52550519e-01 3.15258235e-01
-2.61001021e-01 -1.98723093e-01 -8.27240169e-01 1.18410218e+00
1.17708027e-01 7.16043413e-01 -1.06578462e-01 9.20525044e-02
4.27535921e-01 -1.61432475e-01 3.73957604e-01 -2.69901454e-01
-6.19589329e-01 1.64176822e-01 1.01585619e-01 -3.71441066e-01
-6.57721043e-01 -7.78917372e-01 -6.93153262e-01 5.28832786e-02
-4.21053797e-01 -3.53680812e-02 6.53679073e-01 4.25417930e-01
1.28750309e-01 8.60028565e-01 8.33543181e-01 -1.15592086e+00
-3.12598199e-01 -7.61757672e-01 -9.54198420e-01 6.29400432e-01
5.71041107e-01 -8.04439366e-01 -3.95175278e-01 7.86748156e-02] | [10.865083694458008, -2.7282960414886475] |
1c5df531-2fdd-4205-b305-117f39d7c398 | xjnlp-at-semeval-2017-task-12-clinical | null | null | https://aclanthology.org/s17-2178 | https://aclanthology.org/s17-2178.pdf | XJNLP at SemEval-2017 Task 12: Clinical temporal information ex-traction with a Hybrid Model | null | ['Chen Li', 'Yu Long', 'Xuan Wang', 'Zhijing Li'] | 2017-08-01 | xjnlp-at-semeval-2017-task-12-clinical-1 | https://aclanthology.org/S17-2178 | https://aclanthology.org/S17-2178.pdf | semeval-2017-8 | ['temporal-information-extraction'] | ['natural-language-processing'] | [-2.44508207e-01 3.89024585e-01 -2.65282035e-01 -2.15905145e-01
-8.60921741e-02 -7.76765764e-01 4.48510379e-01 -7.23253429e-01
-5.48377395e-01 1.31954515e+00 3.66348401e-02 -9.49533224e-01
-2.40340635e-01 -1.05564880e+00 -8.44053447e-01 -8.75781775e-01
-7.42435038e-01 6.86515033e-01 1.44298598e-01 -6.52004302e-01
8.47113907e-01 6.29996777e-01 -1.62287033e+00 5.77558100e-01
6.81926727e-01 5.49597681e-01 5.66466339e-02 1.04480565e+00
6.13576770e-02 1.59847176e+00 -6.05450153e-01 -4.56729174e-01
1.43710867e-01 -1.88022718e-01 -5.35770595e-01 -3.36825520e-01
7.91536197e-02 -5.88986814e-01 -3.10633808e-01 6.07356608e-01
1.14270830e+00 -5.53557873e-02 1.07025254e+00 -1.51067197e+00
-7.97060013e-01 5.44234335e-01 2.30399013e-01 1.20632313e-01
6.10446513e-01 -1.79472014e-01 3.51212233e-01 -1.36621380e+00
6.32680655e-01 8.64870071e-01 9.47046995e-01 5.99273622e-01
-1.20647264e+00 -5.60552299e-01 -7.17137158e-01 -2.57835120e-01
-1.59567785e+00 -6.30042374e-01 6.23617955e-02 -3.34735900e-01
1.69142139e+00 6.72587395e-01 1.35647905e+00 1.47341585e+00
1.36649036e+00 4.72517729e-01 1.13904178e+00 -3.16340059e-01
3.65351558e-01 6.43754780e-01 -2.29148138e-02 5.51889002e-01
1.19686353e+00 6.76312149e-01 -5.60917914e-01 -7.92779744e-01
1.11365557e+00 -3.12949389e-01 2.60771513e-01 -7.06989706e-01
-1.06566191e+00 7.88466871e-01 -1.33808568e-01 2.20882341e-01
-3.46203089e-01 -9.92989019e-02 2.24548295e-01 4.12961632e-01
-2.94082850e-01 4.77862865e-01 -8.89795244e-01 -1.87686309e-01
-1.01472771e+00 1.91863477e-01 1.30615628e+00 1.45811689e+00
4.72722203e-03 3.70100170e-01 1.66493103e-01 2.89523482e-01
5.29429734e-01 1.05927479e+00 4.36951250e-01 -1.33365822e+00
-9.55119133e-02 4.20878716e-02 6.12505674e-01 -1.13473701e+00
-7.12930143e-01 -2.41300568e-01 -9.54724610e-01 5.18771529e-01
-1.59270123e-01 -4.59180593e-01 -7.59686172e-01 4.79171664e-01
-1.81984559e-01 -3.19489211e-01 5.45619786e-01 2.69257158e-01
7.28577793e-01 1.45816103e-01 9.25893057e-03 -4.73371416e-01
8.38296890e-01 -1.26899445e+00 -1.11426139e+00 2.33304992e-01
6.82289064e-01 -1.11704338e+00 4.06794518e-01 4.29345578e-01
-1.87200487e+00 -9.69799384e-02 -1.08317137e+00 1.19022481e-01
-7.97166348e-01 -2.57194549e-01 6.66624963e-01 1.49922979e+00
-1.63972652e+00 6.47176981e-01 -6.05021298e-01 5.40335439e-02
-4.73218322e-01 9.07614470e-01 -3.47142309e-01 2.90655226e-01
-1.28751957e+00 9.52449858e-01 -1.04219764e-01 3.03568155e-01
-3.30324143e-01 -1.84501112e-01 -1.03385210e+00 -6.24127567e-01
-2.59417385e-01 -1.07312346e+00 1.36302292e+00 1.09815992e-01
-1.30379498e+00 1.00962746e+00 -4.14315462e-01 -1.14109404e-01
6.61359608e-01 -1.52096570e-01 -8.18174660e-01 1.58321753e-01
2.87593510e-02 3.47163439e-01 5.30214727e-01 -9.69232559e-01
-6.31655395e-01 -1.00153096e-01 -2.54188716e-01 2.48816490e-01
1.43904146e-03 2.05205917e-01 6.38460100e-01 -9.53604504e-02
1.91062525e-01 -8.61961663e-01 -3.21419090e-01 -7.61404037e-01
-1.77353188e-01 -3.84736449e-01 5.59762537e-01 -2.92163283e-01
1.33848476e+00 -1.84365213e+00 -4.54683900e-01 2.94456817e-03
2.53527164e-01 3.89016122e-02 2.33101144e-01 1.23831403e+00
-2.43849114e-01 1.06667292e+00 3.98064762e-01 9.52541176e-03
3.13382268e-01 5.07273376e-01 2.86674555e-02 2.51206756e-01
9.43330750e-02 1.15280068e+00 -1.09214199e+00 -5.20940363e-01
5.57267785e-01 4.60968643e-01 -3.84588838e-01 6.24593735e-01
9.30012882e-01 2.99172718e-02 -8.94021094e-02 1.33241999e+00
1.05377018e+00 -2.90812284e-01 -2.46628653e-02 3.44544262e-01
-6.44541264e-01 6.00222833e-02 -7.41804183e-01 9.49429095e-01
1.31349964e-02 1.99757457e-01 1.74998537e-01 -6.80728197e-01
3.94654661e-01 1.10545933e+00 4.45659161e-01 -9.24297392e-01
-1.69569388e-01 9.00310338e-01 1.03920102e-01 -9.62983370e-01
6.40755475e-01 -1.72503099e-01 7.69825354e-02 1.40852645e-01
-3.61913115e-01 -3.84120375e-01 5.52458428e-02 4.19574350e-01
1.00224769e+00 -1.33707494e-01 5.48685491e-01 -8.72328460e-01
6.88946426e-01 -2.01589778e-01 3.00597936e-01 1.23978651e+00
-2.90862411e-01 9.21478033e-01 3.19022417e-01 -5.37574828e-01
-5.72780192e-01 -1.20098352e+00 -4.80556756e-01 9.86326516e-01
-1.58388391e-01 -3.87671381e-01 -7.10155964e-01 -2.63155133e-01
-1.85071185e-01 4.93483543e-01 -5.93591571e-01 5.23159266e-01
-5.14705896e-01 -6.85349941e-01 7.65277922e-01 3.17972392e-01
6.49624616e-02 -1.49228525e+00 -5.70813596e-01 3.34754556e-01
-1.07132711e-01 -5.39811492e-01 -1.57590687e-01 6.12205923e-01
-1.19366693e+00 -1.01526415e+00 -8.75530541e-01 -1.03603637e+00
7.18889177e-01 2.44572356e-01 1.18463767e+00 2.84042627e-01
-2.87709564e-01 6.49904788e-01 -1.31030381e-02 -2.67939389e-01
-3.16364132e-02 -5.78750819e-02 5.01899779e-01 -1.07851994e+00
6.46393359e-01 -3.92214775e-01 -1.00622165e+00 5.35368681e-01
-3.85821253e-01 -2.04611585e-01 7.85597622e-01 1.04281425e+00
1.42742008e-01 -1.33182362e-01 2.41971418e-01 -1.16735983e+00
1.11498392e+00 -4.85333532e-01 1.30341025e-02 2.62960047e-01
-8.53362501e-01 -4.53584343e-01 2.27881029e-01 -1.33037493e-02
-6.33760452e-01 -2.51423597e-01 -4.14545417e-01 3.22794706e-01
-4.06866640e-01 3.34393084e-02 2.24873364e-01 -5.26000679e-01
5.21005154e-01 1.50258854e-01 -4.07556817e-02 1.21735156e-01
6.30975887e-02 3.75316888e-01 3.67786407e-01 -7.01224983e-01
5.19809723e-01 1.93325520e-01 -3.00995037e-02 -1.28791058e+00
2.66786069e-01 -1.64763972e-01 -6.60750508e-01 -2.69126564e-01
7.75121748e-01 -9.50198710e-01 -6.35705650e-01 5.21708906e-01
-9.54387486e-01 -1.36340424e-01 -5.25335252e-01 6.30965471e-01
-7.02474535e-01 2.94757709e-02 -4.17852998e-01 -1.17976546e+00
-4.78847146e-01 -1.38869667e+00 9.59160328e-01 3.85633826e-01
-2.05811903e-01 -1.19790471e+00 3.38975668e-01 8.99848714e-02
4.29175526e-01 7.39965662e-02 5.03531873e-01 -2.80004859e-01
-2.98304528e-01 -2.85470933e-01 1.59573574e-02 -1.65876746e-01
-3.09365243e-02 4.80770379e-01 -5.48391879e-01 -4.37962860e-01
2.44738594e-01 -4.86500829e-01 -7.51581416e-02 7.11018145e-01
4.09077644e-01 -6.02920949e-01 -7.12544680e-01 5.56552589e-01
1.27709007e+00 4.72497940e-01 5.88340998e-01 6.77552521e-01
9.28869769e-02 5.17871320e-01 4.48484391e-01 2.08113909e-01
1.07506551e-01 1.08483166e-01 1.14217520e-01 -2.17137560e-01
9.45900455e-02 -1.75644696e-01 4.44304377e-01 1.09980536e+00
-8.23837519e-01 -2.15223029e-01 -4.88283783e-01 6.68056428e-01
-1.69156313e+00 -1.31570101e+00 -6.70152605e-01 9.76623476e-01
3.78074080e-01 3.38884830e-01 1.58041969e-01 1.21232189e-01
5.47049284e-01 -4.76693302e-01 -1.27305150e-01 -9.16352808e-01
-3.82233024e-01 4.21136379e-01 7.14838982e-01 6.51479006e-01
-6.10602558e-01 5.08361042e-01 1.25505590e+01 5.49717486e-01
-5.10860123e-02 1.27314463e-01 3.14815581e-01 1.99914098e-01
-2.24020526e-01 1.62632689e-01 -1.04165065e+00 9.26906094e-02
1.64754760e+00 -4.36735392e-01 3.93212438e-01 6.44444764e-01
3.94769877e-01 -1.96429595e-01 -9.78620768e-01 5.21999180e-01
1.12709433e-01 -1.50397491e+00 -3.76861334e-01 9.64888573e-01
5.80093563e-01 -1.81895152e-01 6.81941509e-01 4.33209062e-01
7.39425182e-01 -1.14519978e+00 7.88427055e-01 3.18626672e-01
8.87266219e-01 -8.18826377e-01 1.07315242e+00 2.76328743e-01
-1.04355454e+00 1.15487054e-02 -5.90443075e-01 -9.22581375e-01
1.62244976e-01 2.02188566e-01 -5.61583102e-01 4.06869203e-01
6.19966745e-01 2.15250537e-01 -8.53338003e-01 1.45581341e+00
-4.29231405e-01 4.22687940e-02 -3.00976187e-01 -4.39922899e-01
4.83290881e-01 5.28384298e-02 3.26606274e-01 1.02295005e+00
2.89785862e-01 5.45581102e-01 -2.62208551e-01 2.92286038e-01
6.63503349e-01 5.84383309e-03 -1.38496268e+00 -8.68683681e-02
3.54620427e-01 9.69217539e-01 -3.99494410e-01 -3.56748998e-02
-6.11456513e-01 6.62115753e-01 -1.66942239e-01 5.75146377e-01
-2.92761147e-01 -8.89406264e-01 4.33969021e-01 -2.30117049e-02
-2.88469017e-01 -1.35338321e-01 -3.23372036e-01 -1.02459288e+00
-6.06416404e-01 -3.90122980e-01 7.32179061e-02 -8.63867998e-01
-1.79313409e+00 6.85844719e-01 -3.50327820e-01 -1.03178740e+00
-9.29034173e-01 -9.72984850e-01 -2.90242791e-01 6.67755604e-01
-6.59234166e-01 -1.30476952e+00 1.77773401e-01 4.65253919e-01
1.82401046e-01 -5.88519633e-01 1.09945750e+00 -1.42670095e-01
-3.18711132e-01 1.08882773e+00 3.51253241e-01 -3.05974871e-01
8.32965672e-01 -1.16271269e+00 4.23785180e-01 1.09189451e-02
-6.59259140e-01 1.28460062e+00 6.81860268e-01 -6.81785107e-01
-1.16639841e+00 -1.52893141e-01 1.38110602e+00 -9.15031374e-01
5.27874053e-01 -1.02810413e-01 1.46300867e-01 6.57190859e-01
9.07351732e-01 -5.09461224e-01 1.08935249e+00 -1.13616258e-01
4.93123859e-01 8.23448122e-01 -1.38805819e+00 5.99272728e-01
1.21222687e+00 -4.81854081e-01 -8.27020168e-01 3.28937650e-01
9.94227231e-01 -5.60786903e-01 -1.34066856e+00 6.99696779e-01
1.14132881e+00 -1.04883564e+00 1.54237401e+00 -1.16912270e+00
4.08816934e-01 1.70770600e-01 -2.18455836e-01 -6.52935803e-01
-5.24122715e-01 -8.07741940e-01 -3.97186369e-01 3.40213925e-01
8.64144087e-01 -1.16800272e+00 5.53630650e-01 1.39166510e+00
-3.58625412e-01 -4.51715857e-01 -8.63314152e-01 -1.06149745e+00
-4.66870656e-03 1.81816339e-01 3.85523647e-01 8.18638563e-01
7.09820271e-01 2.02331603e-01 -2.05201268e-01 -4.37674336e-02
6.59316301e-01 -4.87512946e-01 5.82342803e-01 -8.74011815e-01
2.49149442e-01 -3.85976404e-01 -2.94489831e-01 -2.90531307e-01
-3.26795012e-01 -4.87107635e-01 -8.59293997e-01 -1.53084815e+00
8.49664882e-02 1.97367743e-01 -2.33256146e-01 3.44385266e-01
6.77823246e-01 2.45789096e-01 -2.27382466e-01 1.17054820e-01
-1.86927155e-01 -9.06788632e-02 9.18564737e-01 1.10557920e-03
-2.17724726e-01 3.92412305e-01 -1.01716614e+00 9.60534811e-01
7.85404518e-02 -6.09580100e-01 -5.51487744e-01 1.64939865e-01
5.71885824e-01 3.57775450e-01 4.49405432e-01 -9.70509052e-01
9.87445056e-01 -4.32846308e-01 6.12257063e-01 -1.43385470e+00
-2.65485227e-01 -9.00021493e-01 3.61826897e-01 6.29400074e-01
4.08042461e-01 3.53796333e-01 2.26811334e-01 -9.13986787e-02
-2.01378226e-01 -6.44658387e-01 4.43001747e-01 -7.45609999e-01
-4.55086857e-01 -3.90635371e-01 -1.02313221e+00 -9.67486948e-03
7.88939476e-01 -4.61360693e-01 -6.47767961e-01 -1.95199717e-02
-1.44608676e+00 7.30553791e-02 8.33678424e-01 1.34732248e-02
6.98022604e-01 -1.44444084e+00 -1.11748077e-01 6.18802965e-01
-5.42432606e-01 -4.98194307e-01 2.20674574e-01 1.06156552e+00
-1.31528914e+00 1.26915324e+00 -5.21896303e-01 -2.05424115e-01
-8.96585643e-01 7.19095051e-01 3.15837234e-01 1.75098896e-01
-3.93049836e-01 7.00151443e-01 1.12239346e-02 -6.58594191e-01
1.24011397e-01 2.97559589e-01 -4.87763733e-01 -2.52128989e-01
9.22277749e-01 1.17918742e+00 1.32526740e-01 -3.91715407e-01
-5.58191121e-01 4.39446121e-01 2.08604142e-01 -6.12734079e-01
1.38662541e+00 -2.20989168e-01 -7.44191706e-01 5.00318348e-01
7.50003338e-01 -7.33495504e-02 -3.28966863e-02 1.27011871e+00
4.80256975e-03 -5.47239244e-01 -3.11336160e-01 -7.82959402e-01
-3.42039734e-01 5.93214929e-01 6.17733955e-01 4.59080935e-01
9.69997942e-01 -6.47576571e-01 7.70057857e-01 7.64313757e-01
7.83133745e-01 -1.22997761e+00 -6.36951804e-01 3.69483382e-01
9.39836800e-01 -9.30032432e-01 1.30400375e-01 -3.14221263e-01
-5.55534840e-01 1.08689797e+00 5.38173854e-01 -1.62142932e-01
1.17091513e+00 6.96656525e-01 -1.75611794e-01 -4.53250527e-01
-1.04549897e+00 9.83817875e-02 2.10727617e-01 1.32608259e+00
8.26061428e-01 1.41275764e-01 -1.40426433e+00 8.46834600e-01
-6.34404242e-01 4.76996809e-01 9.47268784e-01 1.55018985e+00
-5.12773991e-01 -1.41692734e+00 -4.41162825e-01 2.31142953e-01
-2.82854587e-01 -3.49930972e-01 -8.62966061e-01 1.09929097e+00
1.97083559e-02 1.37330377e+00 -4.48758155e-01 -4.63731140e-01
8.27405751e-01 3.19600284e-01 4.41484600e-01 -1.25142381e-01
-8.94688129e-01 6.48750722e-01 3.84649158e-01 -1.05672014e+00
-7.94483483e-01 -1.27049911e+00 -1.08188879e+00 -1.16740942e+00
-3.07217389e-01 4.65777248e-01 2.31674016e-01 3.19533348e-01
-2.09303293e-02 1.56313553e-03 6.67447627e-01 -1.00111675e+00
-1.97397545e-01 -6.07395887e-01 -1.20775104e+00 -5.56946039e-01
3.94425273e-01 -5.85869551e-01 -1.15223050e+00 -6.24734610e-02] | [-1.5392025709152222, 15.86921501159668] |
a898b9e1-06ab-4ba2-ad09-b4eb0bd6667c | alzheimer-s-disease-detection-from | 2106.08689 | null | https://arxiv.org/abs/2106.08689v1 | https://arxiv.org/pdf/2106.08689v1.pdf | Alzheimer's Disease Detection from Spontaneous Speech through Combining Linguistic Complexity and (Dis)Fluency Features with Pretrained Language Models | In this paper, we combined linguistic complexity and (dis)fluency features with pretrained language models for the task of Alzheimer's disease detection of the 2021 ADReSSo (Alzheimer's Dementia Recognition through Spontaneous Speech) challenge. An accuracy of 83.1% was achieved on the test set, which amounts to an improvement of 4.23% over the baseline model. Our best-performing model that integrated component models using a stacking ensemble technique performed equally well on cross-validation and test data, indicating that it is robust against overfitting. | ['Elma Kerz', 'Daniel Wiechmann', 'Xuefeng Yin', 'Yu Qiao'] | 2021-06-16 | null | null | null | null | ['alzheimer-s-disease-detection'] | ['medical'] | [-2.70996034e-01 -1.31103426e-01 9.18161198e-02 -6.38426781e-01
-1.33298028e+00 -1.94708318e-01 5.21900058e-01 -2.11265936e-01
-6.90149844e-01 9.53010559e-01 6.16288006e-01 -3.29184592e-01
1.72355827e-02 -3.26843441e-01 -9.84059796e-02 -1.90815166e-01
-5.69344521e-01 4.91222620e-01 6.79079220e-02 -2.18787804e-01
-3.73587281e-01 3.53528559e-01 -1.13684583e+00 9.12325740e-01
9.32683110e-01 1.11411679e+00 2.72302888e-02 7.37960696e-01
1.01571091e-01 1.29958856e+00 -6.82930470e-01 -3.12093586e-01
-2.53951252e-01 -3.57975140e-02 -8.38760912e-01 5.52437343e-02
6.39183879e-01 -4.47173178e-01 -4.76096332e-01 6.74272120e-01
8.80027294e-01 -5.10952592e-01 5.95412314e-01 -6.56037271e-01
-5.72473347e-01 4.96668667e-01 2.65223294e-01 7.12494373e-01
3.63648981e-01 7.22027421e-02 8.13432217e-01 -1.12186098e+00
5.42428076e-01 1.41112983e+00 1.12787855e+00 7.16048717e-01
-1.25168741e+00 -4.50613111e-01 1.53351381e-01 2.56017208e-01
-1.19028592e+00 -8.94128442e-01 2.96343267e-01 -6.88918412e-01
1.31305778e+00 1.18247382e-01 7.38391817e-01 1.49929118e+00
5.29400766e-01 7.20593452e-01 1.40769470e+00 -3.60372752e-01
2.13783264e-01 1.95424169e-01 7.59666562e-01 8.97902787e-01
3.75728100e-03 2.23477632e-01 -1.17444448e-01 -6.57120824e-01
2.10706592e-01 -2.88685739e-01 -4.82166894e-02 3.26929659e-01
-1.18868375e+00 8.84338498e-01 3.42099428e-01 4.70329195e-01
-4.13529724e-01 -2.49474958e-01 6.81589663e-01 5.93412817e-01
8.08220446e-01 1.85721233e-01 -5.06004810e-01 1.31515309e-01
-1.01774502e+00 1.74400836e-01 9.46881115e-01 3.98132056e-01
-5.65316260e-01 2.04588816e-01 -1.53259158e-01 1.29605377e+00
5.83422124e-01 7.32721329e-01 8.01664054e-01 -9.06668484e-01
5.80879152e-01 6.96086764e-01 -1.18631169e-01 -1.90084100e-01
-9.58606303e-01 -8.13282311e-01 -7.81134725e-01 1.95577517e-01
3.32114428e-01 -3.93090934e-01 -1.05977345e+00 1.58618128e+00
-5.93953013e-01 -3.48075956e-01 4.09509927e-01 4.28649098e-01
8.26446712e-01 2.04033777e-01 5.26137888e-01 -2.51261413e-01
1.49837661e+00 -9.97311294e-01 -6.47238195e-01 -6.69799626e-01
9.19180751e-01 -6.39449239e-01 7.97292769e-01 4.67732757e-01
-9.27850068e-01 -5.37395537e-01 -1.07344580e+00 1.59320459e-01
-2.38684222e-01 5.87494493e-01 5.39658606e-01 7.97748864e-01
-1.42507279e+00 2.41308168e-01 -8.33734632e-01 -4.06246334e-01
6.04840577e-01 4.01579529e-01 -6.19508445e-01 -2.94393361e-01
-1.41867077e+00 1.16444016e+00 1.36855334e-01 -1.30253702e-01
-7.69455373e-01 -5.76802135e-01 -5.76074004e-01 -2.75016844e-01
-2.46648818e-01 -8.62768233e-01 1.28445435e+00 -8.43902051e-01
-1.09936774e+00 1.09999859e+00 -8.88592079e-02 -1.03067613e+00
6.84092939e-01 -6.02346122e-01 -1.12003100e+00 2.35248342e-01
1.09258331e-01 5.28607547e-01 1.66904822e-01 -6.73052967e-01
-4.88036096e-01 -7.40969658e-01 -4.12154347e-01 -1.69412747e-01
-2.15680107e-01 4.57315654e-01 2.48635665e-01 -6.78700984e-01
-2.09198505e-01 -9.57549214e-01 -2.38835022e-01 3.18515860e-02
-2.80074358e-01 -2.84930170e-01 4.64702666e-01 -1.34167027e+00
1.48260140e+00 -1.92940164e+00 -2.55035967e-01 -1.09507263e-01
5.19103646e-01 5.85756183e-01 -1.81955144e-01 -9.44011062e-02
-3.34137887e-01 2.25323349e-01 -2.52502650e-01 -4.42110538e-01
-3.41444492e-01 -4.99617718e-02 1.63710698e-01 1.96169108e-01
3.37004513e-01 8.44255090e-01 -5.79891741e-01 -2.95039505e-01
8.42031911e-02 6.98236048e-01 -6.62755132e-01 8.34519267e-02
7.77877569e-02 1.68983594e-01 -2.42107421e-01 7.43282735e-01
2.56402940e-01 -1.30691558e-01 2.04057202e-01 -1.86696842e-01
-3.90217127e-03 4.78670537e-01 -2.67547727e-01 1.22985268e+00
-3.35815012e-01 4.89111900e-01 -1.93700537e-01 -8.70467246e-01
8.14288974e-01 4.68334764e-01 4.53203201e-01 -6.96227729e-01
4.81012091e-02 5.80856025e-01 5.48636615e-01 -8.60754192e-01
-7.04724729e-01 -2.22320586e-01 -4.91101444e-02 -1.00636072e-01
7.03391805e-02 7.11517453e-01 1.05418101e-01 -4.60844710e-02
1.40371490e+00 -4.04486388e-01 5.11563659e-01 -4.30222034e-01
9.34622109e-01 -1.70780018e-01 4.41374451e-01 6.09387577e-01
-5.68684280e-01 5.07255733e-01 2.78010815e-01 -5.76975703e-01
-8.90628934e-01 -1.12575316e+00 -4.15547162e-01 5.52465796e-01
-1.18879592e+00 -6.57180846e-01 -8.07978988e-01 -9.37066197e-01
-9.92006809e-02 8.37378085e-01 -4.17895466e-01 -1.87491372e-01
-6.55973971e-01 -1.20178974e+00 9.14243758e-01 8.69714320e-01
6.75983012e-01 -8.12855542e-01 -1.08639799e-01 5.29851913e-01
-2.30056122e-01 -1.13694656e+00 -3.98567051e-01 2.71468729e-01
-1.02891123e+00 -1.23604333e+00 -1.06871498e+00 -1.00921488e+00
1.06872864e-01 -4.57821876e-01 1.17062664e+00 -4.28760827e-01
-2.95141637e-01 4.61286277e-01 -4.69377227e-02 -1.09030247e-01
-7.85031021e-01 2.66737975e-02 3.27777684e-01 -2.83761084e-01
5.87812424e-01 -3.13273102e-01 -3.46812755e-01 1.13011226e-01
-2.01566368e-01 -4.00497437e-01 7.74483562e-01 9.57977831e-01
3.85357171e-01 -2.78985411e-01 1.07065547e+00 -4.40041512e-01
6.77570999e-01 -4.83252585e-01 4.37009670e-02 4.71891105e-01
-7.15203822e-01 6.94546942e-03 4.17400509e-01 -3.42167526e-01
-5.34216583e-01 1.54966876e-01 -5.66133559e-01 -4.84570935e-02
-2.44777396e-01 5.20334899e-01 -3.46513808e-01 1.93324715e-01
5.89999020e-01 7.22699836e-02 4.81751740e-01 -8.61613274e-01
-2.06277817e-01 1.14337850e+00 4.01088804e-01 -9.74127725e-02
-1.97415620e-01 1.25382394e-02 -4.48770672e-01 -9.52116847e-01
-9.43206131e-01 -1.24336168e-01 -6.33701384e-01 -3.43585610e-02
9.44987655e-01 -1.31453681e+00 -1.18002653e-01 9.66309011e-01
-1.09734654e+00 -2.29789048e-01 1.41427755e-01 7.14644611e-01
-3.88038397e-01 2.77614325e-01 -5.68819404e-01 -7.13478446e-01
-7.05218971e-01 -1.24400842e+00 8.23779225e-01 -5.62544525e-01
-5.96422315e-01 -1.05906618e+00 8.07300210e-02 6.63108349e-01
5.78455389e-01 2.72059027e-04 1.05428350e+00 -1.14957881e+00
3.31282616e-01 -4.44101423e-01 -2.14034751e-01 9.29494023e-01
2.66642243e-01 -5.10276616e-01 -7.03843057e-01 -2.04886436e-01
2.70250201e-01 -2.80435294e-01 1.01464856e+00 3.84663820e-01
7.00626373e-01 -2.93768615e-01 -2.92718619e-01 -5.49368858e-02
9.38713610e-01 3.00033510e-01 6.59653366e-01 5.20129263e-01
2.91953266e-01 2.19429821e-01 -2.61329561e-01 4.05975357e-02
4.76567775e-01 8.58687580e-01 -1.46850690e-01 5.63848652e-02
-7.07965672e-01 2.58605510e-01 8.68995249e-01 7.15395153e-01
1.60867453e-01 -2.18037497e-02 -1.37737286e+00 1.06110573e-01
-1.43535602e+00 -1.00071716e+00 -2.61762589e-01 1.91082668e+00
6.65774465e-01 5.64345300e-01 4.77785647e-01 2.47653693e-01
5.64642251e-01 -7.73062035e-02 -3.45304430e-01 -1.99867144e-01
-4.71835762e-01 9.26275626e-02 2.14745358e-01 5.65590441e-01
-1.34497046e+00 5.47603428e-01 7.71564817e+00 3.16097260e-01
-1.05242467e+00 5.31625211e-01 8.99113655e-01 -3.46901000e-01
2.72969037e-01 -7.96925247e-01 -8.99694145e-01 7.27143824e-01
1.83186853e+00 9.99778360e-02 1.96088001e-01 6.88538194e-01
4.96206820e-01 2.00752005e-01 -8.38895977e-01 6.98418200e-01
3.37349266e-01 -8.29995036e-01 9.81295109e-02 -5.19668544e-03
2.54111290e-01 5.92189133e-01 -1.19611911e-01 5.72826505e-01
-3.49450298e-02 -1.15957928e+00 6.26207054e-01 8.61013532e-01
9.01975930e-01 -4.60484326e-01 1.05722821e+00 7.85405040e-02
-8.15935075e-01 -3.26811165e-01 2.69640356e-01 1.70282722e-01
2.92506725e-01 6.43508792e-01 -7.21823156e-01 7.19956160e-02
6.14269674e-01 8.00110817e-01 -9.22839165e-01 1.21410823e+00
2.47725725e-01 9.05724645e-01 -4.07506138e-01 1.70311466e-01
1.39781222e-01 5.52415907e-01 5.35652995e-01 1.47031891e+00
2.21292391e-01 -1.81670934e-01 3.61383200e-01 4.27615911e-01
-2.06612311e-02 1.09370708e-01 -3.74810696e-01 -1.48239866e-01
1.31290063e-01 5.64077377e-01 -9.43227634e-02 -4.57141221e-01
-7.48793721e-01 7.65747070e-01 2.93183357e-01 6.04187548e-02
-6.07845187e-01 -9.02288780e-02 3.25570047e-01 2.98070848e-01
7.25248009e-02 -1.58910826e-01 -2.29852587e-01 -1.16113162e+00
2.11415082e-01 -1.16965318e+00 4.81182992e-01 -5.49655139e-01
-1.59464836e+00 1.11018527e+00 -2.63486236e-01 -9.47053254e-01
-1.53562471e-01 -9.87499595e-01 -3.61563891e-01 8.47915947e-01
-1.17518175e+00 -1.06770873e+00 1.11006480e-02 3.21976662e-01
6.89292789e-01 -1.00158095e+00 1.33618104e+00 6.65206730e-01
-4.72605854e-01 6.45566761e-01 7.50973746e-02 1.85369253e-01
7.47050703e-01 -1.07938635e+00 2.04673439e-01 3.67147118e-01
-3.41453552e-01 5.09137332e-01 2.76420504e-01 -7.06077576e-01
-6.03962064e-01 -1.32351995e+00 1.48904347e+00 -4.05568212e-01
8.96058917e-01 -1.34905055e-01 -8.29245508e-01 5.88212013e-01
-1.85773954e-01 -1.07561700e-01 9.06747937e-01 -5.75614572e-02
-3.81491870e-01 -4.06555533e-02 -1.23083651e+00 1.64863780e-01
1.06292427e+00 -5.78717113e-01 -9.08509791e-01 6.10994756e-01
3.75189155e-01 1.40791371e-01 -1.23756540e+00 6.48082852e-01
6.65905297e-01 -7.86095321e-01 1.17960668e+00 -1.10141563e+00
2.71987200e-01 2.47998789e-01 -6.44116700e-01 -9.93818343e-01
-4.72396821e-01 -3.74740548e-02 -3.22189152e-01 9.65059161e-01
6.25793159e-01 -6.20228827e-01 2.40246192e-01 4.34160441e-01
-1.40229136e-01 -9.72737432e-01 -1.15210724e+00 -1.04335189e+00
4.69409794e-01 -6.38044655e-01 -1.04235254e-01 3.68724376e-01
-1.73345327e-01 6.66203022e-01 -8.61185342e-02 2.67909677e-03
5.15626907e-01 -9.09487188e-01 -3.81126702e-01 -1.63059771e+00
5.13041578e-02 -6.15628362e-01 -8.46273482e-01 -2.09924966e-01
5.52124083e-01 -1.34355009e+00 -5.71289003e-01 -1.61467338e+00
2.83193409e-01 -1.77895412e-01 -6.34148061e-01 5.64254463e-01
-1.13215528e-01 2.05186293e-01 7.85620660e-02 3.31048459e-01
-4.51198488e-01 4.68655884e-01 5.33723831e-01 -4.53916967e-01
-1.37352020e-01 3.69957775e-01 -9.53381836e-01 6.75584376e-01
1.12611485e+00 -1.55926093e-01 -1.79171324e-01 -3.30060244e-01
-5.50846159e-01 -8.81647393e-02 8.01365197e-01 -1.53099275e+00
-2.60413736e-01 7.26764143e-01 6.28983378e-01 -2.52815098e-01
3.76643807e-01 -4.41925019e-01 -3.51748951e-02 9.22537148e-01
-6.58255160e-01 1.34714946e-01 3.58963162e-01 3.10004771e-01
3.53326090e-02 -4.91034985e-03 8.78440559e-01 -1.66484028e-01
-4.13855463e-01 -1.33811697e-01 -8.72797072e-01 1.40455276e-01
6.52076066e-01 2.88412035e-01 -3.20969641e-01 4.42756228e-02
-1.58740354e+00 2.89579965e-02 -1.59206197e-01 6.17344320e-01
4.72362787e-01 -1.33951020e+00 -1.06360805e+00 5.38731627e-02
1.66923124e-02 -1.00860763e+00 7.44057745e-02 9.66897130e-01
-1.37892008e-01 9.30824697e-01 -3.35363187e-02 -3.59702259e-01
-1.45603955e+00 1.95981264e-01 1.02998805e+00 -5.26585579e-01
-8.45934808e-01 4.83542860e-01 -4.44312841e-01 -3.21486920e-01
4.86986369e-01 -2.10368976e-01 -3.33746463e-01 1.80313975e-01
1.07219708e+00 6.14039958e-01 3.84193569e-01 -6.29423738e-01
-7.35795915e-01 1.77637830e-01 -6.31210327e-01 -3.66693251e-02
1.50378358e+00 1.56836972e-01 -3.81309655e-03 6.29741073e-01
1.49795890e+00 -3.81446540e-01 -5.68731487e-01 -1.28018215e-01
3.83816898e-01 6.23711705e-01 5.57537496e-01 -1.62928033e+00
-8.99951816e-01 6.15087628e-01 1.60582030e+00 -4.45241593e-02
5.60795724e-01 1.54068485e-01 7.27954566e-01 6.82963073e-01
1.47385508e-01 -7.78803706e-01 -2.57564604e-01 5.71866572e-01
1.26768172e+00 -1.22880733e+00 -2.02135831e-01 -8.93002748e-02
-4.98996466e-01 1.39349473e+00 2.19225913e-01 -2.62228787e-01
1.04998195e+00 4.33568895e-01 1.85751647e-01 -1.35013118e-01
-8.33144188e-01 2.61156615e-02 6.52785599e-01 6.50862217e-01
8.02652359e-01 2.73492664e-01 -5.43835282e-01 1.28106880e+00
-2.27446124e-01 5.32164216e-01 -1.34998903e-01 6.31302416e-01
-6.62845612e-01 -9.10052598e-01 -1.54258177e-01 9.65325892e-01
-5.93836963e-01 -2.86579221e-01 -5.70236146e-01 6.61029637e-01
9.34502408e-02 9.66813862e-01 -1.30261868e-01 -3.00687045e-01
5.90398669e-01 8.90805781e-01 2.34235629e-01 -4.21895087e-01
-5.89130998e-01 -5.56322299e-02 8.65902126e-01 -5.42918086e-01
-5.25226235e-01 -9.20850456e-01 -1.14467978e+00 9.50600579e-02
2.15658054e-01 -2.15524584e-01 3.59829098e-01 1.06794965e+00
4.95771945e-01 8.25073242e-01 3.00565302e-01 -3.31012458e-01
-8.70652497e-01 -1.29935396e+00 -3.20765167e-01 -1.02028117e-01
4.83419508e-01 -6.83890998e-01 -5.64387143e-01 5.26923165e-02] | [13.910907745361328, 5.38214111328125] |
23e4180d-302e-4041-8e30-6ea5e452adad | beyond-part-models-person-retrieval-with | 1711.09349 | null | http://arxiv.org/abs/1711.09349v3 | http://arxiv.org/pdf/1711.09349v3.pdf | Beyond Part Models: Person Retrieval with Refined Part Pooling (and a Strong Convolutional Baseline) | Employing part-level features for pedestrian image description offers
fine-grained information and has been verified as beneficial for person
retrieval in very recent literature. A prerequisite of part discovery is that
each part should be well located. Instead of using external cues, e.g., pose
estimation, to directly locate parts, this paper lays emphasis on the content
consistency within each part.
Specifically, we target at learning discriminative part-informed features for
person retrieval and make two contributions. (i) A network named Part-based
Convolutional Baseline (PCB). Given an image input, it outputs a convolutional
descriptor consisting of several part-level features. With a uniform partition
strategy, PCB achieves competitive results with the state-of-the-art methods,
proving itself as a strong convolutional baseline for person retrieval.
(ii) A refined part pooling (RPP) method. Uniform partition inevitably incurs
outliers in each part, which are in fact more similar to other parts. RPP
re-assigns these outliers to the parts they are closest to, resulting in
refined parts with enhanced within-part consistency. Experiment confirms that
RPP allows PCB to gain another round of performance boost. For instance, on the
Market-1501 dataset, we achieve (77.4+4.2)% mAP and (92.3+1.5)% rank-1
accuracy, surpassing the state of the art by a large margin. | ['Shengjin Wang', 'Qi Tian', 'Yi Yang', 'Yifan Sun', 'Liang Zheng'] | 2017-11-26 | beyond-part-models-person-retrieval-with-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Yifan_Sun_Beyond_Part_Models_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Yifan_Sun_Beyond_Part_Models_ECCV_2018_paper.pdf | eccv-2018-9 | ['person-retrieval'] | ['computer-vision'] | [-2.06576034e-01 -1.00277968e-01 2.19725184e-02 -3.00537467e-01
-9.91610229e-01 -4.32857960e-01 6.65856481e-01 2.88917422e-01
-4.04866338e-01 6.26345277e-01 4.61919725e-01 5.21394312e-01
2.71408353e-02 -7.51906931e-01 -9.98991251e-01 -5.76855898e-01
-6.42164573e-02 5.21462679e-01 2.88142204e-01 -1.75629407e-01
-2.68511102e-02 7.09739983e-01 -1.71864748e+00 4.10962939e-01
5.55915475e-01 1.08232749e+00 -2.48734653e-01 2.65663594e-01
3.11622280e-03 8.34882632e-02 -7.55480468e-01 -8.97829890e-01
4.23700392e-01 1.37220621e-01 -8.19229126e-01 2.10394204e-01
8.60060811e-01 -4.42195982e-01 -5.69434226e-01 7.41792560e-01
7.18935490e-01 4.39925306e-02 8.09530437e-01 -1.28084540e+00
-5.47312081e-01 3.32578808e-01 -7.27361262e-01 -1.17940716e-01
5.31437814e-01 6.44455254e-02 9.41713035e-01 -1.05553305e+00
4.10463363e-01 1.38555193e+00 8.58421445e-01 4.56423640e-01
-1.12106621e+00 -5.49050987e-01 3.86284024e-01 5.99824004e-02
-1.89787245e+00 -3.90952855e-01 4.98916835e-01 -3.69729370e-01
8.79030645e-01 3.55336994e-01 6.95106208e-01 1.10384619e+00
9.36326832e-02 1.07438385e+00 8.52819979e-01 -9.83503163e-02
-1.24944355e-02 1.65532798e-01 2.94750065e-01 5.20663023e-01
7.23859251e-01 -8.18590000e-02 -6.10306323e-01 -1.58551350e-01
4.49478477e-01 1.81092709e-01 -2.56591111e-01 -5.38125753e-01
-1.00864542e+00 4.02127266e-01 9.42763627e-01 7.93847144e-02
-4.18590128e-01 2.66189784e-01 4.94040281e-01 -3.97874303e-02
3.36334705e-01 1.68521941e-01 -3.70426148e-01 2.01379880e-01
-1.04795277e+00 7.16227949e-01 5.13427436e-01 1.16909492e+00
6.68812871e-01 -5.28108954e-01 -7.19262540e-01 8.43131065e-01
3.17061067e-01 6.67176604e-01 1.46727664e-02 -4.99008507e-01
4.89947677e-01 7.80139327e-01 3.26183438e-01 -9.15523767e-01
-4.00337934e-01 -7.43713379e-01 -1.00407326e+00 -5.23572378e-02
5.98176241e-01 3.64946246e-01 -9.95642543e-01 1.64151657e+00
2.07492068e-01 -3.55825573e-01 -3.78721088e-01 1.09070253e+00
1.00299573e+00 2.30962813e-01 1.61965370e-01 3.66875678e-01
1.72645330e+00 -9.30459738e-01 -1.71313673e-01 -1.29394487e-01
1.86624095e-01 -7.66392171e-01 7.38286436e-01 2.86994666e-01
-9.32725668e-01 -9.10048604e-01 -9.77500618e-01 6.28218502e-02
-5.39952278e-01 3.05884957e-01 4.74479347e-01 8.35937202e-01
-1.19126701e+00 6.02704763e-01 -6.36843979e-01 -5.96964538e-01
6.74432218e-01 4.62403893e-01 -8.28405619e-01 -4.17038977e-01
-9.12253797e-01 6.33270264e-01 7.67806098e-02 1.18425593e-01
-6.29931867e-01 -8.36455464e-01 -8.20586443e-01 1.29425526e-01
1.75405502e-01 -1.03442442e+00 8.56734097e-01 -6.08288884e-01
-9.89113092e-01 9.68504369e-01 -3.24525356e-01 -4.17637885e-01
8.94788325e-01 -5.60955465e-01 -2.53715098e-01 3.44182163e-01
4.19891655e-01 9.46674109e-01 9.75502670e-01 -1.40587080e+00
-5.25696099e-01 -5.04555821e-01 -4.59889434e-02 -3.48403193e-02
-2.54468977e-01 1.02274656e-01 -1.03688216e+00 -5.68923414e-01
1.53951883e-01 -8.65936041e-01 -4.78567928e-02 3.42772216e-01
-6.90474510e-01 -4.12763953e-01 3.11038435e-01 -7.70065606e-01
9.02590215e-01 -1.87600279e+00 -1.87563226e-01 3.29898298e-01
4.70915943e-01 3.13072115e-01 -7.01517761e-02 5.93559206e-01
-1.10526688e-01 -1.01804473e-02 7.33925775e-02 -7.71757364e-01
3.55450302e-01 -2.73378521e-01 2.04298310e-02 7.67311335e-01
5.52939177e-01 1.19841588e+00 -5.21723032e-01 -4.84889656e-01
2.83562273e-01 6.93079650e-01 -5.16265631e-01 -1.22859083e-01
4.09043789e-01 1.69014394e-01 -4.56381738e-01 1.02390611e+00
1.08183444e+00 -2.08915874e-01 -3.54486853e-01 -4.70922053e-01
4.21256498e-02 -1.78115159e-01 -1.29470313e+00 1.48221755e+00
5.58842421e-02 4.36942130e-01 -1.54186308e-01 -6.79707289e-01
8.37045908e-01 1.31641895e-01 4.91763115e-01 -7.78023958e-01
-7.72931054e-02 1.06747121e-01 -3.28744739e-01 -1.32655486e-01
6.79913700e-01 3.34324032e-01 -2.89261490e-01 -9.49706212e-02
1.37597442e-01 4.32678461e-01 1.77248538e-01 2.99762517e-01
1.19423819e+00 2.43994609e-01 2.43229285e-01 -3.82246852e-01
6.30070567e-01 -2.72016376e-01 4.42029387e-01 9.60977972e-01
-5.58694839e-01 1.16517186e+00 2.77914643e-01 -5.75877070e-01
-1.16812623e+00 -1.23587441e+00 -3.21249247e-01 7.36309886e-01
2.88553864e-01 -5.72353065e-01 -7.53530741e-01 -6.27243042e-01
4.54184383e-01 -1.41383067e-01 -8.53879094e-01 -1.16406575e-01
-6.40289903e-01 -5.36769986e-01 6.59481525e-01 8.42463255e-01
8.88005137e-01 -6.60751343e-01 -4.03045297e-01 -1.59850391e-03
-2.57881910e-01 -1.01766801e+00 -5.80470920e-01 -9.82806981e-02
-4.35354114e-01 -1.22607374e+00 -1.31338155e+00 -5.37797332e-01
7.93324709e-01 5.18170238e-01 1.35588372e+00 1.77670404e-01
-3.83106142e-01 5.14403045e-01 -4.46078181e-01 -2.27278978e-01
3.42973560e-01 5.47153093e-02 1.79722100e-01 5.87925315e-02
4.74682063e-01 -1.47555009e-01 -1.09460306e+00 4.78023857e-01
-4.59854275e-01 -3.46835971e-01 9.42030907e-01 7.52006531e-01
6.20967209e-01 1.25045395e-02 2.18062192e-01 -2.85139918e-01
1.82331294e-01 -8.45496505e-02 -2.12979868e-01 2.65028745e-01
-2.69161642e-01 -1.34952620e-01 4.09617424e-01 -1.71435535e-01
-7.55975604e-01 2.23582193e-01 -3.78398895e-02 -3.70108724e-01
-5.40486574e-01 -1.22426048e-01 -5.90774953e-01 -1.47125989e-01
5.36468863e-01 2.74411798e-01 -1.55099258e-01 -6.58465743e-01
2.40010664e-01 6.21751666e-01 5.66256762e-01 -8.44913304e-01
9.95345533e-01 6.48331642e-01 -5.32182194e-02 -7.45278537e-01
-6.73121333e-01 -9.16125178e-01 -9.40089941e-01 -2.24080056e-01
7.39367247e-01 -1.29412079e+00 -8.71390045e-01 5.93868911e-01
-9.91548240e-01 1.24842614e-01 -1.89424545e-01 1.10033952e-01
-1.28145382e-01 5.03822327e-01 -5.54163575e-01 -7.23642588e-01
-4.35066044e-01 -9.04390037e-01 1.58777201e+00 3.11433911e-01
-2.20542341e-01 -2.53938586e-01 -2.90660888e-01 4.31581169e-01
3.09296966e-01 3.22692335e-01 3.38291705e-01 -7.62617648e-01
-6.42802119e-01 -7.79677391e-01 -6.39222205e-01 1.58188060e-01
-1.53207585e-01 -8.63883793e-02 -1.21446288e+00 -7.53169835e-01
-7.35414565e-01 -9.31595266e-02 1.18846178e+00 3.68399918e-01
9.55199361e-01 -9.61492956e-02 -7.59359062e-01 3.59554648e-01
1.26578653e+00 -4.00723279e-01 7.56342947e-01 4.34505939e-01
5.53533673e-01 6.07719302e-01 4.76490706e-01 5.49857974e-01
3.48348409e-01 1.05808449e+00 2.54290521e-01 -1.00139633e-01
-5.35923719e-01 -2.82991201e-01 1.50700793e-01 5.00205159e-02
-2.23373145e-01 -1.77659899e-01 -6.96536720e-01 6.78543508e-01
-1.84767282e+00 -9.79114950e-01 -1.73908636e-01 2.44812989e+00
3.82892191e-01 2.17955261e-01 5.38091242e-01 8.43254328e-02
7.97322035e-01 -4.16736118e-02 -1.94977254e-01 2.19344690e-01
-1.79392412e-01 2.78570533e-01 5.76203287e-01 1.15390187e-02
-1.57280731e+00 8.14878702e-01 5.27619839e+00 7.92216122e-01
-5.62519610e-01 -1.46901742e-01 5.07601857e-01 -1.73825398e-02
2.70467103e-01 -3.24660033e-01 -1.25398338e+00 5.47012091e-01
4.04982150e-01 2.15298444e-01 -6.63634539e-02 8.08618128e-01
-1.04798593e-01 -1.44919798e-01 -1.20842278e+00 1.11490583e+00
3.01139057e-01 -9.04553950e-01 1.93159640e-01 2.63784736e-01
6.82105541e-01 -2.75533855e-01 -4.74806540e-02 4.67969894e-01
-1.91655025e-01 -1.04425764e+00 1.03530741e+00 8.15217972e-01
6.20238781e-01 -1.04744470e+00 1.09021890e+00 5.04199183e-03
-1.63105798e+00 8.92760158e-02 -6.36457086e-01 2.10851356e-01
-5.10005951e-02 5.74707150e-01 -4.38903958e-01 9.00252461e-01
1.22228956e+00 5.24442554e-01 -1.02863264e+00 1.71024311e+00
-2.60216128e-02 1.96329623e-01 -3.84324491e-01 9.16635171e-02
5.45343943e-02 1.92739561e-01 4.68195111e-01 1.43788040e+00
6.94763958e-02 -1.05476633e-01 3.68546546e-01 8.05622816e-01
-1.89724743e-01 2.36562770e-02 -2.66895741e-01 3.57891232e-01
4.06147927e-01 1.17196226e+00 -6.95413172e-01 -3.22793633e-01
-4.27560002e-01 1.36787760e+00 3.03340197e-01 3.29624176e-01
-7.45023012e-01 -4.81659561e-01 9.86168444e-01 3.47402900e-01
6.31121516e-01 3.18093761e-03 -7.22888783e-02 -9.45699275e-01
3.97828817e-01 -6.34471059e-01 2.97768414e-01 -4.88797933e-01
-1.68184876e+00 5.38554728e-01 1.21963300e-01 -1.35663271e+00
1.39061347e-01 -8.02788675e-01 -3.95319611e-01 1.05803990e+00
-1.41704512e+00 -1.68131208e+00 -6.40991151e-01 5.70893586e-01
3.23834956e-01 -4.98375408e-02 7.28245020e-01 5.46452522e-01
-5.70222318e-01 1.06129670e+00 -1.22405916e-01 6.14511490e-01
9.54977274e-01 -1.23426306e+00 5.31967103e-01 9.00635183e-01
-3.85753326e-02 9.14448321e-01 6.09564662e-01 -7.29525506e-01
-1.23351336e+00 -1.24566138e+00 9.81006324e-01 -7.36729562e-01
1.69815615e-01 -4.96518970e-01 -6.71138287e-01 3.37583929e-01
-1.66399479e-01 3.28246206e-01 4.50886190e-01 2.24875346e-01
-7.53895879e-01 -3.12374711e-01 -1.28303826e+00 5.12636483e-01
1.17503023e+00 -4.52263296e-01 -3.92371804e-01 2.40995675e-01
5.15191615e-01 -3.67276043e-01 -7.68557668e-01 4.73049492e-01
7.84473538e-01 -1.07243598e+00 1.52357972e+00 -5.40137708e-01
2.10737035e-01 -4.84739423e-01 -3.44038248e-01 -8.12045813e-01
-7.37957418e-01 -2.48518154e-01 -6.27347007e-02 1.39268506e+00
5.20036258e-02 -4.46082324e-01 9.05238688e-01 8.09406996e-01
-1.80829018e-01 -7.80034304e-01 -8.20069015e-01 -1.04116058e+00
1.62493065e-01 -3.32031012e-01 7.54388273e-01 3.46030504e-01
-3.93137157e-01 -4.58546430e-02 -2.70631522e-01 4.76633817e-01
8.89746785e-01 6.50490150e-02 1.07310426e+00 -1.37000263e+00
-1.24092974e-01 -6.63737833e-01 -8.92400086e-01 -1.12946713e+00
-1.83965098e-02 -6.69606626e-01 1.05049245e-01 -1.67234993e+00
6.96443796e-01 -3.60876024e-01 -4.25634265e-01 5.54270446e-01
-3.28238755e-01 7.51268506e-01 2.38378733e-01 2.12529704e-01
-8.92579377e-01 4.60319996e-01 8.95409584e-01 -4.24073666e-01
8.84755105e-02 2.36778677e-01 -7.35688627e-01 3.86689693e-01
6.30551517e-01 -1.60854027e-01 2.32870162e-01 -3.40151787e-02
-1.46179274e-01 -4.97112215e-01 9.53850508e-01 -1.42873108e+00
1.73895657e-01 5.62485993e-01 1.16945660e+00 -9.97765005e-01
6.27232015e-01 -7.33702481e-01 8.22947249e-02 5.05703926e-01
6.16333820e-02 -1.22206844e-01 2.12018669e-01 6.69960022e-01
-1.54215902e-01 9.87042040e-02 5.81572235e-01 -9.87713858e-02
-8.15468431e-01 5.08979023e-01 3.95537615e-02 -1.29813641e-01
7.75782049e-01 -4.99340653e-01 -3.55854720e-01 -1.60348251e-01
-5.28285801e-01 1.82862207e-01 5.18037081e-01 4.81729656e-01
5.69603264e-01 -1.42128778e+00 -8.38615179e-01 1.00195900e-01
5.76504707e-01 -2.80850977e-01 3.50486070e-01 7.46376455e-01
-2.66169399e-01 7.32294023e-01 -1.33352444e-01 -8.08357656e-01
-1.33512437e+00 4.42646086e-01 3.23936105e-01 -9.43268389e-02
-8.33265185e-01 1.06291699e+00 2.11526528e-01 -3.25831652e-01
3.70804995e-01 -1.12601228e-01 -1.93325593e-03 8.60953331e-02
6.58811986e-01 3.57671738e-01 3.29835296e-01 -9.13749933e-01
-8.34723651e-01 6.94403291e-01 -2.40235224e-01 1.92882210e-01
1.13311911e+00 5.63008785e-02 7.59831667e-02 -1.72785789e-01
1.29598522e+00 8.57544392e-02 -1.35180795e+00 -2.92616695e-01
-8.54589790e-02 -7.37658858e-01 -3.63650680e-01 -9.06474948e-01
-8.82277787e-01 6.21186137e-01 7.07289517e-01 -1.62903890e-01
7.99359024e-01 3.07748944e-01 8.05673361e-01 3.17214340e-01
8.06580245e-01 -9.30416942e-01 1.38576820e-01 2.99137920e-01
9.90986109e-01 -1.36176991e+00 2.98845679e-01 -4.55437869e-01
-3.87464702e-01 9.03134644e-01 5.94014466e-01 -2.74483621e-01
5.03835559e-01 -6.90248162e-02 -2.71036506e-01 -2.30811059e-01
-2.47121707e-01 -6.03908896e-01 8.78517687e-01 7.40735590e-01
3.99342209e-01 3.02752584e-01 -9.39126406e-03 8.03892314e-01
-1.43697068e-01 -3.94864768e-01 -2.32142389e-01 6.51114345e-01
-4.86262441e-01 -1.08066869e+00 -6.45107567e-01 5.34272671e-01
-2.89533913e-01 8.53732973e-02 -5.05774736e-01 1.05042517e+00
2.58811623e-01 9.31196451e-01 1.07904077e-01 -2.03273952e-01
7.10386157e-01 2.06238199e-02 6.29718959e-01 -3.07324260e-01
-6.97247982e-01 -6.02179952e-02 1.16668809e-02 -1.00964808e+00
-8.48304927e-02 -7.90483594e-01 -9.11616206e-01 -4.00957525e-01
-1.32551402e-01 -1.51894772e-02 3.56158465e-01 8.16175461e-01
3.86577487e-01 4.10413623e-01 2.83419639e-01 -1.07730126e+00
-5.08824706e-01 -7.84188032e-01 -5.17877281e-01 7.31048942e-01
1.83027744e-01 -8.42670441e-01 -2.75172610e-02 -2.76501179e-01] | [14.684309005737305, 0.8924669027328491] |
7dbf4046-bc7b-40d9-b500-f04620100e26 | participatory-research-as-a-path-to-community | 2306.08906 | null | https://arxiv.org/abs/2306.08906v1 | https://arxiv.org/pdf/2306.08906v1.pdf | Participatory Research as a Path to Community-Informed, Gender-Fair Machine Translation | Recent years have seen a strongly increased visibility of non-binary people in public discourse. Accordingly, considerations of gender-fair language go beyond a binary conception of male/female. However, language technology, especially machine translation (MT), still suffers from binary gender bias. Proposing a solution for gender-fair MT beyond the binary from a purely technological perspective might fall short to accommodate different target user groups and in the worst case might lead to misgendering. To address this challenge, we propose a method and case study building on participatory action research to include experiential experts, i.e., queer and non-binary people, translators, and MT experts, in the MT design process. The case study focuses on German, where central findings are the importance of context dependency to avoid identity invalidation and a desire for customizable MT solutions. | ['Katharina Bühn', 'Daniela Duh', 'Sigrid Schefer-Wenzl', 'Igor Miladinovic', 'Arthur Mettinger', 'Lukas Daniel Klausner', 'Sabrina Burtscher', 'Katta Spiel', 'Manuel Lardelli', 'Dagmar Gromann'] | 2023-06-15 | null | null | null | null | ['machine-translation'] | ['natural-language-processing'] | [ 3.83297890e-01 7.23648906e-01 -3.84312391e-01 -2.67107934e-01
-5.81676126e-01 -8.79765689e-01 8.03581834e-01 2.68114656e-01
-5.53797305e-01 8.38224649e-01 7.29771733e-01 -1.03529894e+00
1.74228251e-01 -3.19345295e-01 -1.75419480e-01 -3.02990586e-01
1.10361052e+00 3.45060378e-01 -5.20213485e-01 -3.05003136e-01
5.99763095e-01 2.49061808e-02 -1.09224415e+00 1.50655016e-01
1.24106014e+00 1.06484942e-01 -3.07799488e-01 1.30798355e-01
-3.83925825e-01 5.56529999e-01 -6.96429968e-01 -9.18181658e-01
7.92947561e-02 -5.18434227e-01 -8.39225173e-01 -1.18983183e-02
4.54559863e-01 -3.10621262e-01 4.74637270e-01 1.04603004e+00
8.01916599e-01 -3.04288208e-01 2.26756334e-01 -1.02011943e+00
-1.05625117e+00 1.05567253e+00 -6.13064945e-01 -1.90603480e-01
4.66109693e-01 2.46623859e-01 7.59399652e-01 -7.86089718e-01
9.72917259e-01 1.54058433e+00 7.66308963e-01 4.74473029e-01
-1.51886249e+00 -7.23432481e-01 -4.68195640e-02 -2.56933391e-01
-1.43310761e+00 -9.35738862e-01 2.91101664e-01 -8.09935033e-01
5.13342679e-01 7.33539522e-01 9.33817923e-01 1.19983196e+00
1.29248559e-01 -7.78743550e-02 1.45756197e+00 -7.18228340e-01
-9.36785713e-02 5.85419834e-01 -3.85516673e-01 1.09371364e-01
8.20551336e-01 -3.97635251e-01 -4.65738267e-01 -2.53695101e-01
2.96364546e-01 -4.19882834e-01 1.05977260e-01 1.06228605e-01
-1.48043346e+00 6.42005444e-01 -3.31962615e-01 7.80340493e-01
-2.11602882e-01 -8.86814017e-03 4.48727012e-01 2.37504601e-01
6.11473858e-01 5.42952836e-01 -4.17365767e-02 -6.91769898e-01
-7.90282786e-01 3.64480585e-01 7.98975050e-01 3.51434797e-01
3.21112067e-01 -2.08194494e-01 -4.43874836e-01 5.82242906e-01
5.57351947e-01 6.15924239e-01 2.10153535e-02 -9.85435307e-01
6.52130365e-01 7.36153483e-01 5.10058761e-01 -1.24716294e+00
4.59612124e-02 -2.69727916e-01 -1.64870769e-01 1.31859809e-01
7.41923869e-01 -3.55403692e-01 -6.87206209e-01 1.60489774e+00
3.78480911e-01 -9.81219292e-01 -3.34433436e-01 1.03479087e+00
2.17526123e-01 3.02942544e-01 6.74192250e-01 -3.64515811e-01
1.22009742e+00 -3.68328951e-02 -8.55084777e-01 -5.33001542e-01
8.29915762e-01 -1.12449598e+00 1.23602140e+00 -1.07092172e-01
-1.22457027e+00 1.85154378e-02 -8.06492448e-01 -3.26699555e-01
-2.28112370e-01 1.35149360e-01 4.41650569e-01 1.61398065e+00
-9.72910762e-01 3.37051272e-01 -5.04950523e-01 -1.04021025e+00
2.40616858e-01 2.74609119e-01 -3.67983282e-02 2.67726511e-01
-6.93560600e-01 1.08584332e+00 -8.78873840e-02 4.74548906e-01
2.13351578e-01 -5.91169477e-01 -3.91494185e-01 -3.52735788e-01
1.61254376e-01 -6.87240362e-01 1.26511991e+00 -1.69936991e+00
-1.15900707e+00 1.14629507e+00 6.59154952e-02 2.40377471e-01
9.78707314e-01 -8.17011371e-02 -4.02312487e-01 -2.74814576e-01
5.44912040e-01 5.87632716e-01 3.03183347e-01 -1.32013810e+00
-6.20821774e-01 -4.56931621e-01 1.59146264e-02 2.08424747e-01
-3.67752165e-01 7.17547238e-01 6.56347573e-01 -2.86213636e-01
3.02775264e-01 -7.11583614e-01 4.15021880e-03 -7.50783235e-02
-3.85176271e-01 3.35353576e-02 6.14666283e-01 -8.90978992e-01
1.34134638e+00 -1.98140037e+00 -3.84494960e-01 1.02391064e-01
5.09703636e-01 -8.82565156e-02 4.66366708e-01 9.43513215e-01
2.28730142e-01 7.00421691e-01 2.14587361e-01 -8.73154253e-02
4.63135600e-01 -1.65228412e-01 -1.70560330e-01 6.62692785e-01
-1.93007141e-01 6.82503164e-01 -1.05609143e+00 -8.13349903e-01
-1.88611180e-01 3.25358897e-01 -2.48688892e-01 -7.90303051e-01
1.64776862e-01 6.21806383e-01 -1.31758109e-01 7.34616995e-01
7.70991504e-01 1.09050594e-01 8.36561263e-01 1.91693082e-01
-1.05716157e+00 2.33839303e-01 -7.27301776e-01 1.05966151e+00
-4.27840382e-01 8.44539583e-01 5.85328996e-01 -2.70936996e-01
9.03643072e-01 3.26879293e-01 2.83995420e-01 -9.31125581e-01
3.92288834e-01 9.05222714e-01 7.02149689e-01 -3.59258235e-01
1.07070553e+00 -1.85999885e-01 -1.30066067e-01 6.26017451e-01
-6.07386112e-01 1.37376204e-01 3.21588516e-02 -9.15411711e-02
4.43667203e-01 4.31153685e-01 7.56780356e-02 -7.38532066e-01
-1.57448262e-01 2.03755930e-01 7.93550968e-01 4.01854843e-01
-4.55686569e-01 2.44240910e-01 4.54522341e-01 -8.88875201e-02
-1.24693167e+00 -6.56745136e-01 4.69261669e-02 1.14280951e+00
3.14157337e-01 -4.08073694e-01 -6.89546347e-01 -3.01987052e-01
-2.13857189e-01 1.07848692e+00 -9.62627009e-02 -1.75966434e-02
-4.19062942e-01 -5.32938123e-01 5.59850395e-01 -2.01349214e-01
5.80455005e-01 -2.56070614e-01 -1.21105206e+00 2.47185931e-01
-4.67158705e-01 -8.77454102e-01 -3.77161235e-01 -6.60772264e-01
-5.71747661e-01 -5.83503783e-01 -7.76781201e-01 -4.95293111e-01
6.83608949e-01 1.72091156e-01 7.44719207e-01 -4.60152850e-02
1.51730627e-01 1.80571958e-01 -2.74193704e-01 -5.90613008e-01
-8.78612280e-01 9.99135301e-02 -2.98301518e-01 2.59306245e-02
4.74221349e-01 -5.77683449e-01 -6.88375413e-01 2.64259636e-01
-4.93814260e-01 5.93929887e-01 2.89318651e-01 2.13297293e-01
-2.76273727e-01 -7.56129444e-01 5.97440660e-01 -7.67019749e-01
7.96757877e-01 -4.13995653e-01 1.06844239e-01 3.84158641e-01
-8.31909180e-01 -7.31988251e-01 1.22764222e-01 -5.91281295e-01
-1.26585674e+00 -6.10871851e-01 2.18115792e-01 5.34694076e-01
3.57866079e-01 5.25796771e-01 -2.15392798e-01 -2.62405798e-02
9.59045351e-01 -5.38808346e-01 2.79532075e-01 -1.87119529e-01
1.20762892e-01 8.93906772e-01 4.46287878e-02 -6.34705961e-01
5.77830434e-01 2.86841482e-01 -2.86416471e-01 -7.20411897e-01
2.71008015e-02 2.35309735e-01 -3.28149796e-01 -6.79693401e-01
6.29603624e-01 -8.12102377e-01 -7.27939725e-01 1.07015766e-01
-9.63977754e-01 -6.52154148e-01 -2.74623573e-01 6.16219521e-01
-2.82254040e-01 5.31115271e-02 -3.21171492e-01 -1.06969559e+00
-1.96346819e-01 -8.33794892e-01 7.80295253e-01 4.02716607e-01
-1.10347903e+00 -8.87238085e-01 -1.89971253e-01 5.93350768e-01
7.44614363e-01 6.70131326e-01 9.78328288e-01 -4.82985288e-01
-2.92356223e-01 -1.24496184e-02 -3.18306416e-01 -5.74287772e-01
5.40602744e-01 3.56779069e-01 -6.59020782e-01 -5.21155912e-03
-9.28920060e-02 -1.60056486e-04 -1.45478606e-01 -1.96617663e-01
6.78642839e-02 -7.43152440e-01 -3.20541263e-01 -1.83737427e-01
1.23325586e+00 2.88299173e-01 5.76584220e-01 5.97524405e-01
7.56547272e-01 9.01514173e-01 6.49148941e-01 5.50827086e-01
8.67095590e-01 5.33198714e-01 -2.23984316e-01 -2.01607928e-01
3.77435088e-02 -3.91181856e-01 2.41107717e-01 6.29669905e-01
-1.71242297e-01 9.63884592e-02 -1.41124845e+00 6.51022673e-01
-1.67614567e+00 -8.19072664e-01 -4.77588922e-01 2.10222721e+00
8.06560278e-01 7.37223327e-02 4.46110576e-01 1.56604171e-01
9.54702139e-01 -6.25523087e-03 1.02319196e-01 -1.10265064e+00
5.66505678e-02 -1.38178244e-01 5.00184476e-01 3.55702013e-01
-4.32110816e-01 6.69420362e-01 5.58957052e+00 5.46201289e-01
-1.42779601e+00 3.92724693e-01 9.42583025e-01 1.06278531e-01
-9.74791408e-01 4.32667702e-01 -4.32607859e-01 2.82108784e-01
7.21299231e-01 -9.09594059e-01 1.06861368e-01 9.48143899e-02
6.20139778e-01 -3.37309152e-01 -8.42543662e-01 8.22118163e-01
-2.54491091e-01 -1.05065346e+00 -4.22925919e-01 4.66861516e-01
7.54097879e-01 -5.50691545e-01 2.14383885e-01 -6.35633394e-02
1.10603921e-01 -1.07639158e+00 1.55917645e+00 7.12713897e-02
1.06743062e+00 -3.91044170e-01 4.35320735e-01 1.81482852e-01
-4.86486524e-01 -1.40920833e-01 1.20074891e-01 -5.64727902e-01
2.89855927e-01 3.29085112e-01 -7.23919988e-01 3.44728768e-01
2.01464549e-01 1.52820209e-03 -3.99750739e-01 7.13622987e-01
2.29666695e-01 4.52305645e-01 -1.58481702e-01 -2.62082797e-02
3.53586376e-02 -4.45698529e-01 4.57276374e-01 1.10687411e+00
5.71176052e-01 -1.62252054e-01 -2.22180918e-01 7.74087787e-01
2.98242331e-01 5.14468551e-01 -4.74312723e-01 -6.28520191e-01
8.56953442e-01 1.12153566e+00 -1.18435550e+00 1.48104966e-01
-2.97268301e-01 5.23488224e-01 -4.22298461e-01 5.63295126e-01
-5.01967669e-01 -5.63952327e-02 5.11668563e-01 7.77580619e-01
-4.24005389e-01 -3.87913585e-01 -9.48738515e-01 -6.43512011e-01
7.67514855e-02 -1.32401872e+00 -2.67685771e-01 -1.02749959e-01
-5.83203495e-01 -3.07132721e-01 5.06935548e-03 -7.97316074e-01
-1.07149467e-01 -9.75870788e-02 -3.39819580e-01 1.10056651e+00
-3.30693722e-01 -1.67171848e+00 -1.29601099e-02 -4.22934324e-01
2.88934596e-02 3.82002354e-01 4.51051176e-01 5.27449787e-01
-3.84309798e-01 6.93596065e-01 8.09486136e-02 -4.47076261e-01
7.86823094e-01 -6.83847845e-01 5.39522111e-01 7.21999407e-01
-4.74568486e-01 1.10970175e+00 1.04237759e+00 -9.52984214e-01
-1.48788035e+00 -4.80413079e-01 1.71524823e+00 -4.00644124e-01
4.41932678e-01 -5.74301004e-01 -1.71869382e-01 6.61050856e-01
3.05544466e-01 -8.53532195e-01 7.20105171e-01 2.15268463e-01
-3.32142264e-02 7.39922449e-02 -1.55560982e+00 1.34113514e+00
1.38409853e+00 -8.33669603e-01 -7.59532303e-02 4.47539240e-02
4.91290599e-01 -3.98965716e-01 -7.21003652e-01 4.97996211e-02
1.04351616e+00 -7.72031009e-01 4.66338843e-01 8.01223963e-02
3.18781018e-01 -1.79806516e-01 -4.66497093e-02 -6.90038860e-01
-6.79341331e-02 -1.13093615e+00 7.10524023e-01 1.93732119e+00
3.97758007e-01 -8.12880397e-01 4.80850846e-01 1.35194874e+00
-7.40349665e-02 -4.32884693e-01 -1.27119541e+00 -3.08273733e-01
3.67374927e-01 -5.60953990e-02 8.59762490e-01 1.30670762e+00
3.18532079e-01 1.25714242e-01 -1.73384756e-01 -3.36497515e-01
8.55246112e-02 -1.21340841e-01 8.73541296e-01 -1.23515618e+00
8.09843838e-02 -6.38161480e-01 -2.28439599e-01 -2.57264704e-01
-2.95377165e-01 -6.15760446e-01 -1.07541509e-01 -1.78137457e+00
1.97742507e-01 -3.95382553e-01 6.95131838e-01 1.57090962e-01
1.00940950e-01 2.36845151e-01 4.10257727e-01 1.25194743e-01
-6.20778650e-03 1.37363166e-01 1.21157026e+00 1.11342832e-01
-5.04310846e-01 -3.10093462e-01 -1.49666858e+00 3.59477073e-01
9.42567527e-01 -3.90858799e-01 -4.82877463e-01 -3.81428391e-01
1.01574469e+00 -2.92410105e-01 4.91114229e-01 -4.88315791e-01
9.72058028e-02 -7.60240614e-01 2.01228246e-01 -5.55741899e-02
-1.23818628e-02 -6.62179530e-01 8.81085157e-01 4.23479050e-01
-1.88324735e-01 1.07410654e-01 7.76939318e-02 -3.81873548e-01
3.25446427e-01 -8.29579756e-02 2.16289431e-01 -3.17964777e-02
1.03023283e-01 -3.83368909e-01 -6.38891995e-01 1.18232621e-02
7.92563856e-01 -1.06221306e+00 -6.28450453e-01 -2.77671188e-01
-3.68477792e-01 -4.64453129e-03 1.12436199e+00 3.47294420e-01
-9.34869945e-02 -1.07828200e+00 -6.91614926e-01 -4.14123356e-01
2.16307584e-02 -5.24094760e-01 1.07742518e-01 1.12574148e+00
-5.51123679e-01 2.84482777e-01 -2.54587203e-01 1.44411074e-02
-1.19530475e+00 -1.54440388e-01 2.94367343e-01 3.34567666e-01
-2.96114624e-01 4.05592412e-01 4.36243750e-02 -4.38798875e-01
-3.66899401e-01 6.43079635e-03 6.56925738e-02 4.26213384e-01
-4.80561815e-02 8.47288966e-01 -1.61567777e-01 -1.04887211e+00
-3.99029762e-01 2.01115191e-01 3.71740431e-01 -1.07665682e+00
6.88173294e-01 -7.03124881e-01 -4.59516704e-01 7.08480716e-01
6.19138956e-01 6.47092760e-01 -7.38007486e-01 5.00780404e-01
3.59950185e-01 -1.18559492e+00 -5.68933547e-01 -9.53202903e-01
-2.51582652e-01 5.02059698e-01 4.66791689e-01 4.02587891e-01
6.89267516e-01 -2.65056551e-01 3.87393564e-01 -2.26716906e-01
4.41612422e-01 -1.58997738e+00 -5.50536156e-01 7.30021298e-02
8.27736974e-01 -6.31077766e-01 4.77280840e-02 -4.08355325e-01
-3.90519798e-01 9.06519711e-01 4.90473092e-01 7.45090485e-01
1.45597324e-01 -2.99664170e-01 3.11693519e-01 4.27793376e-02
-4.13363338e-01 2.24567088e-03 -3.88784051e-01 7.09823012e-01
1.09017968e+00 4.81331050e-01 -1.57221866e+00 4.20901895e-01
-5.84532440e-01 2.53687412e-01 6.46453023e-01 1.00556219e+00
-1.71070248e-01 -1.30763412e+00 -7.40412474e-01 1.83056578e-01
-9.50705588e-01 2.57983446e-01 -9.48386908e-01 9.57433939e-01
6.44804239e-01 1.33643556e+00 2.59426594e-01 -2.79297113e-01
1.45010009e-01 8.39682594e-02 6.26575768e-01 -3.53804320e-01
-1.28130913e+00 1.06289886e-01 9.12667751e-01 4.26725522e-02
-2.42604345e-01 -1.01760292e+00 -7.65720129e-01 -9.89912808e-01
7.53681064e-02 3.52279134e-02 8.70512247e-01 8.80022168e-01
5.84511936e-01 -2.59100407e-01 2.94266760e-01 -4.13493276e-01
-3.16739291e-01 -8.32862496e-01 -3.13654661e-01 3.01866420e-02
1.19882099e-01 -1.77037150e-01 1.20737411e-01 -1.81042880e-01] | [9.154427528381348, 9.898896217346191] |
5fd40019-951f-409f-9936-6d8634e17990 | motion-comfort-optimization-for-autonomous | 2306.09462 | null | https://arxiv.org/abs/2306.09462v1 | https://arxiv.org/pdf/2306.09462v1.pdf | Motion Comfort Optimization for Autonomous Vehicles: Concepts, Methods, and Techniques | This article outlines the architecture of autonomous driving and related complementary frameworks from the perspective of human comfort. The technical elements for measuring Autonomous Vehicle (AV) user comfort and psychoanalysis are listed here. At the same time, this article introduces the technology related to the structure of automatic driving and the reaction time of automatic driving. We also discuss the technical details related to the automatic driving comfort system, the response time of the AV driver, the comfort level of the AV, motion sickness, and related optimization technologies. The function of the sensor is affected by various factors. Since the sensor of automatic driving mainly senses the environment around a vehicle, including "the weather" which introduces the challenges and limitations of second-hand sensors in autonomous vehicles under different weather conditions. The comfort and safety of autonomous driving are also factors that affect the development of autonomous driving technologies. This article further analyzes the impact of autonomous driving on the user's physical and psychological states and how the comfort factors of autonomous vehicles affect the automotive market. Also, part of our focus is on the benefits and shortcomings of autonomous driving. The goal is to present an exhaustive overview of the most relevant technical matters to help researchers and application developers comprehend the different comfort factors and systems of autonomous driving. Finally, we provide detailed automated driving comfort use cases to illustrate the comfort-related issues of autonomous driving. Then, we provide implications and insights for the future of autonomous driving. | ['Ala Al-Fuqaha', 'Mohsen Guizani', 'Basheer Qolomany', 'Junaid Qadir', 'Mohamed Rahouti', 'Mohammed Aledhari'] | 2023-06-15 | null | null | null | null | ['autonomous-vehicles'] | ['computer-vision'] | [-7.95750558e-01 1.65549934e-01 -8.58460441e-02 -3.28974962e-01
-1.57852113e-01 -6.26139641e-01 6.06687106e-02 -2.49347478e-01
-2.73120552e-01 4.38572347e-01 -8.15992057e-02 -3.18129450e-01
2.07988396e-01 -3.50685477e-01 -2.68408179e-01 -9.35533524e-01
2.95945406e-01 -4.15743023e-01 -4.21601161e-02 -6.01306438e-01
1.95848122e-01 6.70730710e-01 -2.47232032e+00 -3.65571588e-01
1.01774275e+00 1.11231887e+00 3.76776367e-01 5.57792068e-01
4.27274019e-01 4.84086275e-01 -3.89445245e-01 -9.93724093e-02
-7.99628794e-02 -6.78710490e-02 1.13549002e-01 -8.64698887e-02
-1.62671298e-01 -5.47420204e-01 -5.33070147e-01 5.35281241e-01
5.01616955e-01 2.92671323e-01 5.95859766e-01 -1.96111369e+00
-1.73506081e-01 -5.30257344e-01 4.36341107e-01 -1.13830402e-01
2.02584103e-01 5.98329246e-01 -5.59640899e-02 -8.34201634e-01
2.00313423e-02 9.58890915e-01 4.56408769e-01 7.51670957e-01
-6.19116127e-01 -6.13931537e-01 -1.42677571e-04 6.36821926e-01
-1.43861556e+00 -9.24733579e-01 5.75921714e-01 -5.79872966e-01
1.03879273e+00 4.99667704e-01 8.04617345e-01 1.05201948e+00
1.09242845e+00 3.39110792e-01 8.51864457e-01 1.53894827e-01
6.01499081e-01 1.21577048e+00 6.32845521e-01 6.77055568e-02
3.74336541e-01 3.85359883e-01 -3.44225168e-02 2.39140471e-03
-2.69475251e-01 -4.70190495e-01 -3.98987764e-03 -2.23894313e-01
-1.61730766e-01 7.87707031e-01 -1.93615153e-01 1.08954959e-01
-2.94316143e-01 8.50659460e-02 5.36932826e-01 2.23310858e-01
-4.43122201e-02 2.47039735e-01 -2.15205789e-01 -3.73021990e-01
-7.21650496e-02 3.44943494e-01 7.64857113e-01 1.05576098e+00
5.80863118e-01 2.46369645e-01 -9.35084820e-02 5.14920890e-01
5.75500488e-01 1.00824070e+00 7.00613996e-03 -1.23450196e+00
-5.50498724e-01 4.18912852e-03 4.74536896e-01 -8.31412733e-01
-5.11912346e-01 -1.24529503e-01 1.13740548e-01 7.55248487e-01
-3.36172193e-01 -4.77837831e-01 -4.91077691e-01 1.24246728e+00
4.10356894e-02 -4.95302975e-01 4.71730024e-01 8.35785627e-01
1.19654024e+00 7.15186059e-01 3.08987349e-01 -5.68311691e-01
1.51081812e+00 -9.62028384e-01 -1.79528880e+00 -6.68711364e-01
5.94858944e-01 -6.07374489e-01 9.87355173e-01 3.50605071e-01
-9.67527151e-01 -8.35645676e-01 -1.48544693e+00 -1.02890469e-01
-7.78974950e-01 -1.02425165e-01 1.55516431e-01 1.32668221e+00
-1.14015305e+00 -2.94829458e-01 -4.63060200e-01 -5.33143163e-01
-3.78213853e-01 1.75903052e-01 5.58474474e-02 -5.51976683e-03
-1.29179108e+00 1.75786233e+00 -3.62910032e-02 1.18000768e-01
-5.82461059e-01 -5.06016612e-01 -1.12132537e+00 -3.49903405e-01
9.99954492e-02 -4.48428541e-01 1.33643758e+00 -4.69894350e-01
-1.44178641e+00 5.81981301e-01 -5.61528146e-01 -8.33837837e-02
2.23593295e-01 -5.40998578e-01 -1.24589860e+00 -1.60771776e-02
1.72000989e-01 6.08705282e-01 4.49153960e-01 -1.43025267e+00
-6.58325434e-01 -1.22819714e-01 -4.06834304e-01 2.02236250e-01
1.65621832e-01 4.27985709e-04 -2.28725627e-01 4.36045706e-01
-6.06056809e-01 -1.15196145e+00 -1.13458179e-01 -2.47406870e-01
1.80334225e-01 -4.28441226e-01 1.52430606e+00 -2.77028322e-01
1.24170887e+00 -2.96612191e+00 -6.22850537e-01 -1.56369116e-02
1.32826060e-01 7.33916014e-02 1.35842994e-01 5.48783123e-01
1.15367636e-01 -3.41347933e-01 3.76060963e-01 1.23599637e-03
1.66233867e-01 4.02497292e-01 -1.49951398e-01 5.48382103e-01
1.23468481e-01 7.59369016e-01 -4.60460931e-01 -2.13776514e-01
8.69833410e-01 7.89699852e-01 8.70838994e-04 4.24905926e-01
4.34073001e-01 2.62564123e-01 -1.12070315e-01 5.95437407e-01
1.04124045e+00 1.08543527e+00 -6.43145323e-01 -1.47497609e-01
-8.00445974e-01 5.05785979e-02 -5.01029015e-01 7.56686032e-01
-2.95574784e-01 9.34296072e-01 3.25157106e-01 -1.31529421e-01
1.07857621e+00 6.40178025e-01 3.45113963e-01 -1.34848177e+00
4.28391725e-01 4.87583637e-01 1.54001487e-03 -1.40039301e+00
8.35749745e-01 1.08749997e-02 -2.02358589e-01 -2.15055406e-01
-6.40144348e-01 -4.02276427e-01 -3.29169810e-01 -8.56417120e-02
4.75477993e-01 -2.36514077e-01 1.94494482e-02 -6.21152580e-01
5.69791734e-01 -2.33519319e-02 4.43588972e-01 9.87212807e-02
-1.20421767e+00 -2.62222558e-01 5.19574106e-01 -9.83031914e-02
-6.64901793e-01 -9.37212229e-01 -5.44017196e-01 8.53250742e-01
7.93130338e-01 -2.46420786e-01 -9.90726292e-01 1.01600692e-01
2.42684782e-01 1.68907273e+00 -5.12558579e-01 -7.69192576e-01
5.24130613e-02 -2.02823028e-01 2.18775854e-01 5.67945719e-01
1.99565411e-01 -6.63276017e-01 -1.25530457e+00 -1.05082043e-01
-3.29606116e-01 -1.37553251e+00 5.54757267e-02 6.27371967e-02
-2.27350131e-01 -4.38518196e-01 2.14625612e-01 -5.35033345e-01
2.12490156e-01 8.07557046e-01 5.29427826e-01 4.39642444e-02
6.77290261e-02 7.56286323e-01 2.92312559e-02 -1.01385272e+00
-3.24187249e-01 -4.29588079e-01 3.84408653e-01 -2.44285688e-01
1.03551388e+00 1.34456918e-01 -6.95918202e-01 7.48527229e-01
-4.06762153e-01 -2.96047300e-01 2.06473053e-01 -1.72854718e-02
1.50496230e-01 3.29202563e-01 8.05780411e-01 6.85222298e-02
8.02254260e-01 -5.76663017e-01 -2.05918029e-01 -3.81899655e-01
-8.58161569e-01 -4.37700212e-01 5.35909720e-02 -1.79227833e-02
-9.48864281e-01 1.48723731e-02 -2.91685134e-01 3.99390981e-02
-5.79440236e-01 -2.13149056e-01 -6.95579588e-01 -1.06917560e-01
5.46047330e-01 -2.87157089e-01 6.72305346e-01 3.39155376e-01
1.01771519e-01 1.11444890e+00 4.64860618e-01 1.13708749e-01
4.94603604e-01 3.58785391e-01 -1.52129382e-01 -1.15474582e+00
-4.07092646e-02 -5.17631948e-01 -2.21726164e-01 -1.07343435e+00
1.23603618e+00 -1.04642653e+00 -1.04440057e+00 4.70070958e-01
-8.83435965e-01 -3.30826849e-01 -9.91225243e-03 5.88274419e-01
-6.42249227e-01 -8.26094151e-02 -2.77824134e-01 -1.60745585e+00
-8.23500752e-02 -1.82789850e+00 9.44577634e-01 3.86001110e-01
-6.07806802e-01 -7.66023874e-01 -2.65759647e-01 7.16988087e-01
6.40338421e-01 4.81089622e-01 5.82007408e-01 1.72636315e-01
-1.84702188e-01 -3.42697114e-01 2.55968630e-01 3.21939230e-01
-1.02240831e-01 2.16439128e-01 -1.76741624e+00 -2.42642626e-01
4.96219039e-01 -8.63169692e-03 3.25289875e-01 6.36165977e-01
4.71812397e-01 2.26921946e-01 -5.02523184e-01 2.47479185e-01
1.33403349e+00 1.10306108e+00 1.06795895e+00 3.57172161e-01
1.84886366e-01 1.45689857e+00 1.48827910e+00 1.01958342e-01
3.56993496e-01 2.95168579e-01 7.18544543e-01 -2.19361559e-01
4.32533562e-01 1.85206890e-01 7.58980453e-01 6.07335806e-01
1.22400753e-01 3.97168612e-03 -3.31638068e-01 3.00753951e-01
-1.46798849e+00 -5.06174028e-01 -8.38464737e-01 2.06337094e+00
-1.60359442e-01 7.26519451e-02 2.06710085e-01 2.72451639e-01
3.81128460e-01 -3.49984765e-01 -6.67476475e-01 -1.41461205e+00
-9.55603551e-03 -4.04791147e-01 7.37992942e-01 1.01890624e+00
-7.99074650e-01 6.04683697e-01 7.51348114e+00 3.17745030e-01
-8.41360688e-01 7.62536079e-02 5.50949872e-01 -2.33995900e-01
-3.63859415e-01 -1.48900956e-01 -6.78405166e-01 4.98353720e-01
1.57104015e+00 -2.62968719e-01 2.78979260e-02 1.17945349e+00
1.01405585e+00 -9.13170636e-01 -1.03462744e+00 9.14133847e-01
1.73228711e-01 -2.84728348e-01 -7.21098602e-01 2.82363951e-01
6.35667294e-02 -1.41244784e-01 3.65650594e-01 4.45262074e-01
-9.02376413e-01 -1.04259002e+00 8.36331964e-01 4.82647210e-01
7.39804208e-01 -1.37416649e+00 1.06286871e+00 4.09233831e-02
-9.90065217e-01 -2.52432346e-01 -2.43559018e-01 -4.93205875e-01
3.46352577e-01 3.53378445e-01 -2.44209051e-01 -1.35609046e-01
8.06588352e-01 -4.63723950e-02 -4.53189731e-01 5.29320002e-01
2.81199127e-01 -7.40589350e-02 1.61443070e-01 -4.53819066e-01
2.05393776e-01 -4.18425769e-01 2.99220890e-01 1.01441896e+00
2.41597325e-01 2.43653685e-01 -4.69181925e-01 7.84613311e-01
1.22947395e+00 8.57700780e-03 -1.18892753e+00 1.08244285e-01
4.41059977e-01 1.41960859e+00 -1.25765875e-01 -1.41504794e-01
-6.26493037e-01 6.20852590e-01 -6.94216013e-01 6.52219713e-01
-1.15253842e+00 -6.43725395e-01 1.51347327e+00 4.18031335e-01
-4.28436548e-01 -2.45298892e-01 -8.19621384e-01 -5.74554503e-03
7.93067291e-02 -3.69042635e-01 -2.86669344e-01 -1.16538012e+00
-4.85616446e-01 5.36913574e-01 8.54973942e-02 -1.22574055e+00
2.80558225e-02 -7.62515426e-01 -4.08552885e-01 7.97335148e-01
-1.28305507e+00 -5.28355002e-01 -8.14476132e-01 2.32027829e-01
4.72489268e-01 -7.72009417e-02 7.54520178e-01 1.93530217e-01
-5.68785071e-01 5.42335510e-01 6.34212792e-02 -8.58140528e-01
9.67370570e-01 -5.76212466e-01 9.47910771e-02 6.05190575e-01
-1.70770609e+00 2.73095280e-01 1.34939575e+00 -5.88983893e-01
-1.81699443e+00 -8.19441795e-01 7.67095327e-01 -7.93940246e-01
1.68227330e-01 -5.88634670e-01 -7.23158777e-01 2.73924261e-01
7.44919956e-01 -8.48639667e-01 9.42881286e-01 -2.44550064e-01
3.80793869e-01 -4.38853800e-01 -1.37348640e+00 1.00293863e+00
4.19966727e-01 -5.41325867e-01 -2.04113894e-03 6.19902276e-02
8.15031648e-01 -1.68591112e-01 -5.44434428e-01 1.90026000e-01
9.35941160e-01 -1.19633961e+00 3.95168513e-01 2.36691117e-01
-8.59215260e-02 -2.65105933e-01 -2.81379391e-02 -8.98662865e-01
-6.04351640e-01 -5.29587448e-01 2.33135924e-01 6.75172091e-01
3.43751818e-01 -1.26129103e+00 5.60031943e-02 1.64714003e+00
-7.33473897e-01 -6.63432479e-01 -7.29244649e-01 -5.38735330e-01
-1.16805263e-01 -7.78582454e-01 2.93481439e-01 3.68568897e-01
5.50489008e-01 2.65390813e-01 -1.00177318e-01 4.21382159e-01
1.56755373e-01 -8.73254597e-01 8.22511673e-01 -1.10168326e+00
5.06452858e-01 -2.53562540e-01 -7.42099822e-01 -2.60027885e-01
1.37802929e-01 9.83232558e-02 3.47757280e-01 -1.52007246e+00
-1.70183927e-01 9.25671011e-02 1.86294228e-01 -3.18605334e-01
-1.88876182e-01 -2.54390299e-01 -3.73062789e-02 -3.98071498e-01
1.11189410e-01 3.64456832e-01 1.06626415e+00 2.94596732e-01
-3.48552436e-01 -5.95132597e-02 -8.46201301e-01 3.18537742e-01
1.10720921e+00 1.29441261e-01 -1.01848102e+00 3.24296355e-02
1.44513026e-02 -1.09922597e-02 2.74059355e-01 -1.02207816e+00
2.32947841e-01 -4.61976647e-01 1.08629046e-03 -9.32990074e-01
7.67415464e-01 -1.44803751e+00 6.07744992e-01 8.18064809e-01
2.47097924e-01 2.82161057e-01 7.50099480e-01 2.17119843e-01
3.94497737e-02 -1.63340956e-01 1.10979176e+00 5.27776778e-01
-8.80200148e-01 -3.07080626e-01 -1.65226579e+00 -7.63953745e-01
1.89838767e+00 -8.81607175e-01 -4.04923618e-01 -6.40544772e-01
-4.86040473e-01 6.64963901e-01 6.51390910e-01 8.37818980e-01
6.17013156e-01 -1.32590652e+00 -2.65297592e-01 4.95649159e-01
4.67444718e-01 -6.43233299e-01 6.30287528e-01 1.12705290e+00
-4.82760698e-01 7.90483773e-01 -5.16371608e-01 -6.06036842e-01
-1.37024403e+00 8.29450727e-01 6.55487359e-01 1.16236079e+00
-2.85371929e-01 2.61869252e-01 6.45379305e-01 1.96612477e-01
2.85598010e-01 -1.53243884e-01 -2.89409965e-01 -1.18037991e-01
4.56881255e-01 7.92532384e-01 2.94593304e-01 -8.49732518e-01
-6.20621264e-01 3.46134394e-01 2.47007564e-01 -3.28711987e-01
9.89026502e-02 -7.62619257e-01 -4.77160588e-02 9.81454849e-01
1.10686243e+00 -1.49457380e-01 -9.55087662e-01 8.50979149e-01
-5.64723074e-01 -2.95922846e-01 2.10994974e-01 -5.25202751e-01
-7.03070819e-01 7.02284038e-01 1.27822745e+00 1.17091380e-01
1.19169140e+00 -6.28914386e-02 8.75630319e-01 1.75889343e-01
7.51853436e-02 -1.94609547e+00 -2.70465493e-01 1.94635153e-01
9.45131123e-01 -1.19989944e+00 -5.17377496e-01 -7.60770440e-01
-1.12439239e+00 7.85637200e-01 9.26628649e-01 2.16709539e-01
8.96420062e-01 7.22757220e-01 7.92490542e-01 -1.38541430e-01
-8.16194475e-01 -3.50520849e-01 1.45301074e-01 1.04659808e+00
4.32436168e-01 4.30222631e-01 -6.21023476e-01 5.14987946e-01
-4.28543657e-01 -3.61968935e-01 5.60173571e-01 7.87859142e-01
-8.60492945e-01 -4.45520431e-01 -6.99112713e-01 1.90285876e-01
5.93575500e-02 6.38893247e-01 -5.12646019e-01 7.53668666e-01
4.09895927e-01 2.05386353e+00 1.75775409e-01 -9.06593323e-01
7.55498528e-01 3.18919361e-01 -2.81391203e-01 1.76510662e-01
-1.52619019e-01 1.92443520e-01 5.67831457e-01 -7.58808136e-01
-1.13224583e-02 -7.86772311e-01 -1.49176216e+00 -6.89943314e-01
-1.14775673e-01 1.79396942e-01 1.13660169e+00 8.37010145e-01
3.40192944e-01 7.11967766e-01 8.38491082e-01 -4.85597372e-01
2.01190308e-01 -5.56046188e-01 -8.31221044e-01 -1.82372749e-01
6.54021382e-01 -8.79764080e-01 -6.96037412e-01 -3.97878408e-01] | [5.712462425231934, 1.0746855735778809] |
54853f1c-58c6-4718-9808-36273b945e6d | bootstrapping-multilingual-amr-with | 2102.02189 | null | https://arxiv.org/abs/2102.02189v1 | https://arxiv.org/pdf/2102.02189v1.pdf | Bootstrapping Multilingual AMR with Contextual Word Alignments | We develop high performance multilingualAbstract Meaning Representation (AMR) sys-tems by projecting English AMR annotationsto other languages with weak supervision. Weachieve this goal by bootstrapping transformer-based multilingual word embeddings, in partic-ular those from cross-lingual RoBERTa (XLM-R large). We develop a novel technique forforeign-text-to-English AMR alignment, usingthe contextual word alignment between En-glish and foreign language tokens. This wordalignment is weakly supervised and relies onthe contextualized XLM-R word embeddings.We achieve a highly competitive performancethat surpasses the best published results forGerman, Italian, Spanish and Chinese. | ['Todd Ward', 'Salim Roukos', 'Radu Florian', 'Tahira Naseem', 'Ramon Fernandez Astudillo', 'Young-suk Lee', 'Janaki Sheth'] | 2021-02-03 | null | https://aclanthology.org/2021.eacl-main.30 | https://aclanthology.org/2021.eacl-main.30.pdf | eacl-2021-2 | ['multilingual-word-embeddings'] | ['methodology'] | [-6.95891753e-02 6.32967055e-02 -6.38116181e-01 -2.66855568e-01
-1.26233566e+00 -7.86046565e-01 8.96150351e-01 3.27730209e-01
-9.96402502e-01 8.04302037e-01 7.29410052e-01 -7.99690247e-01
4.84934151e-01 -6.90840304e-01 -4.86423016e-01 -3.68725538e-01
3.95559758e-01 6.56010985e-01 -3.24367851e-01 -5.13875484e-01
-2.38633618e-01 2.55400658e-01 -7.23266721e-01 3.89885932e-01
7.93896496e-01 2.44412825e-01 8.06081891e-02 6.24347925e-01
-4.83561158e-01 7.21038759e-01 -1.92361012e-01 -7.37727046e-01
-6.61517009e-02 -9.68842953e-02 -1.01499069e+00 -4.09462780e-01
3.10834348e-01 4.08855170e-01 -7.38297924e-02 8.24154496e-01
1.26941830e-01 -2.41067577e-02 6.42513573e-01 -8.59972835e-01
-1.32420063e+00 1.20565319e+00 -5.62317908e-01 2.94611484e-01
6.13109529e-01 -4.28964525e-01 1.70528209e+00 -1.42540348e+00
8.84358227e-01 1.29427099e+00 7.22732663e-01 5.55587769e-01
-1.53247023e+00 -6.12269938e-01 4.72770959e-01 1.67066261e-01
-1.61008906e+00 -4.54488665e-01 3.75746936e-01 -2.71027565e-01
1.74268675e+00 4.62677367e-02 3.85812402e-01 1.30007744e+00
3.77071887e-01 7.31362939e-01 1.13324630e+00 -9.42147672e-01
-2.82179087e-01 2.63704181e-01 4.34226245e-01 5.45499206e-01
4.20145690e-01 -3.16777080e-01 -4.68690991e-01 4.59476486e-02
3.52762043e-01 -2.60948747e-01 2.86369715e-02 1.79412320e-01
-1.37540305e+00 8.44290674e-01 4.35485244e-02 7.43993998e-01
-3.28842849e-01 3.31131369e-01 4.86936182e-01 6.39794886e-01
4.56039220e-01 4.54889685e-01 -8.33348215e-01 -1.08267076e-01
-3.48717839e-01 -4.01251875e-02 5.84454298e-01 1.08193886e+00
8.08992803e-01 4.94719088e-01 2.07504869e-01 1.19108331e+00
3.46015930e-01 7.86823869e-01 1.02112484e+00 -3.89685482e-01
5.32306373e-01 1.72564104e-01 -1.52526349e-01 -4.38737154e-01
-1.91577137e-01 3.17282230e-02 -4.52703357e-01 -1.52983129e-01
1.23598546e-01 -1.59545988e-01 -8.11340570e-01 1.59160399e+00
1.58072978e-01 -3.45510900e-01 6.46574557e-01 3.79078865e-01
8.30545723e-01 7.30508149e-01 3.70479435e-01 5.78007028e-02
1.74982655e+00 -1.01522338e+00 -8.54702234e-01 -4.58188415e-01
1.11045694e+00 -9.74397123e-01 1.42594862e+00 2.43941873e-01
-7.48195410e-01 -5.46832561e-01 -8.46580088e-01 -3.46629113e-01
-8.39081943e-01 7.48735368e-02 4.69203681e-01 4.37889934e-01
-8.63915026e-01 1.25484407e-01 -6.37641788e-01 -5.12055278e-01
5.33659644e-02 -7.85529837e-02 -1.07595181e+00 -2.93207318e-01
-1.32505524e+00 1.32715154e+00 6.14144087e-01 -2.91706294e-01
-6.90516233e-01 -6.56939447e-01 -1.45499539e+00 -4.69603956e-01
1.79469511e-02 -2.28114143e-01 1.18191862e+00 -8.34318042e-01
-1.41052425e+00 1.39166498e+00 -2.47066259e-01 -4.81771708e-01
-2.04776898e-01 -7.72417784e-01 -8.90763164e-01 -4.42619950e-01
1.79907501e-01 4.44698751e-01 3.82030606e-01 -9.64101076e-01
-5.20576119e-01 -2.77283520e-01 -2.00872198e-01 2.68114269e-01
-5.40606797e-01 5.54651260e-01 -4.69755419e-02 -1.09379017e+00
-1.53606325e-01 -8.31935823e-01 -3.32022101e-01 -5.32307982e-01
-1.99313432e-01 -6.27581894e-01 4.22220170e-01 -9.82435763e-01
1.22013330e+00 -1.93810070e+00 3.43370974e-01 -9.98862833e-02
-1.50291502e-01 1.96056023e-01 -6.30353630e-01 7.09776342e-01
-6.22462571e-01 9.01388824e-02 -3.50605063e-02 -4.85173732e-01
2.37399116e-01 8.89243007e-01 -5.21291316e-01 3.89274716e-01
5.23631394e-01 1.21248996e+00 -1.14103734e+00 -4.36460048e-01
2.77241647e-01 3.27215940e-01 -4.29875821e-01 4.23113167e-01
-3.24287228e-02 -8.09027776e-02 -6.62597194e-02 5.95459402e-01
1.11689150e-01 3.52911174e-01 6.13900840e-01 3.24835703e-02
-2.41503835e-01 6.53622806e-01 -8.50208759e-01 1.86451662e+00
-1.35585153e+00 4.70526397e-01 -1.68298140e-01 -9.15556610e-01
9.47896242e-01 6.19266987e-01 2.76579469e-01 -4.10998255e-01
8.91336128e-02 4.82037365e-01 -8.82262811e-02 -1.74036920e-01
9.84281361e-01 -6.23972356e-01 -6.45019948e-01 3.73751342e-01
6.30474746e-01 -4.05052751e-01 -6.20112754e-02 -8.22055638e-02
8.45204473e-01 5.30683994e-01 1.01408410e+00 -4.52942252e-01
5.81526577e-01 -4.31892648e-02 3.91862303e-01 3.31440449e-01
2.07988903e-01 3.10795188e-01 1.16601802e-01 -4.75804746e-01
-1.13472700e+00 -1.20910263e+00 -3.65948051e-01 1.60511339e+00
-3.84568453e-01 -1.05549169e+00 -3.62320095e-01 -8.92908394e-01
-2.50458717e-02 1.29157233e+00 -4.71091688e-01 1.17487662e-01
-7.98167825e-01 -5.86842120e-01 8.28130901e-01 8.42707694e-01
-3.97646219e-01 -1.08109474e+00 6.38788566e-02 5.62823474e-01
1.97758507e-02 -1.59063125e+00 -4.90545720e-01 4.52208638e-01
-4.14633304e-01 -6.90383255e-01 -4.73333687e-01 -1.10674524e+00
3.38824958e-01 -3.64681900e-01 1.42699420e+00 -4.36233729e-01
-1.52522072e-01 3.14416289e-01 -5.01591802e-01 -3.96674991e-01
-6.48939550e-01 2.91231394e-01 3.53086144e-01 -3.39684069e-01
9.67950702e-01 -3.77577424e-01 4.11887795e-01 -2.74941713e-01
-6.73387885e-01 -1.75185055e-01 4.22228664e-01 8.79645765e-01
9.84500647e-01 -6.21193767e-01 4.40285087e-01 -1.13336337e+00
5.18983424e-01 -5.05952120e-01 -4.83473420e-01 3.68616819e-01
-4.50512141e-01 3.07004243e-01 1.08630836e+00 -5.08251250e-01
-8.82926404e-01 -5.89439832e-02 -7.94055521e-01 -7.28491023e-02
-1.34467661e-01 4.08282906e-01 -4.18975949e-01 3.25871527e-01
6.18108511e-01 5.43840006e-02 -6.76777899e-01 -6.57412827e-01
1.33590233e+00 9.05873537e-01 6.02436244e-01 -1.02185631e+00
8.87699366e-01 -8.08281824e-02 -4.57912177e-01 -1.09004247e+00
-9.02521193e-01 -5.43791056e-01 -6.85102284e-01 4.11000043e-01
1.24636757e+00 -1.16092443e+00 -1.55895867e-03 5.92573313e-03
-1.40272462e+00 -1.40070453e-01 -7.57277966e-01 6.32120848e-01
-3.41813922e-01 2.65642732e-01 -9.53205705e-01 -5.82097828e-01
-4.92860883e-01 -6.80043876e-01 1.03141892e+00 -4.30378348e-01
-7.48486340e-01 -1.39519036e+00 5.89667797e-01 -5.18310815e-02
2.18863145e-01 4.96534305e-03 1.11562252e+00 -1.14519918e+00
1.01276010e-01 -3.09830874e-01 -1.78351685e-01 5.54791629e-01
2.96341747e-01 -3.33235443e-01 -9.08249974e-01 -2.69720167e-01
-5.11005342e-01 -4.64581102e-01 6.46665990e-01 -1.44044787e-01
6.32029891e-01 -3.52632701e-01 -1.50065437e-01 4.96892333e-01
1.53377485e+00 -3.43928337e-01 3.47496867e-01 4.27466601e-01
1.08772492e+00 2.83481479e-01 4.40897942e-01 1.20024029e-02
6.59571409e-01 7.26929545e-01 -1.15979858e-01 -2.00918958e-01
-3.78191292e-01 -4.92440313e-01 9.75169480e-01 1.87874818e+00
1.75626129e-02 2.06120256e-02 -1.02243590e+00 1.02751434e+00
-1.59244585e+00 -3.54880303e-01 -1.71711028e-01 2.09974527e+00
1.29911947e+00 -7.75051191e-02 -3.06139141e-01 -2.89779127e-01
1.47952378e-01 3.86052221e-01 1.64529771e-01 -9.97503281e-01
-3.51065367e-01 9.40880239e-01 5.52046359e-01 1.01196587e+00
-8.74915481e-01 1.67385149e+00 6.76975536e+00 7.34725535e-01
-5.80338836e-01 8.14014494e-01 -5.74683100e-02 4.00947511e-01
-8.92340660e-01 1.81117311e-01 -1.03161216e+00 -3.15934159e-02
1.28648245e+00 -2.96071947e-01 4.43089098e-01 9.17486846e-01
-4.17397738e-01 5.99971712e-01 -1.20899510e+00 9.38656867e-01
3.79716426e-01 -9.59943771e-01 3.84819597e-01 -2.63659716e-01
7.07364082e-01 6.31461084e-01 -3.27669740e-01 5.41166008e-01
1.21538019e+00 -1.22236907e+00 4.06303704e-01 2.07780153e-01
1.21350491e+00 -8.93022537e-01 9.68659937e-01 -3.17271024e-01
-1.57957244e+00 2.95588136e-01 -5.96267045e-01 2.71675229e-01
4.38866138e-01 4.75473732e-01 -5.82125485e-01 9.58980083e-01
3.75715375e-01 1.10922849e+00 -3.31677198e-01 1.01110619e-02
-8.35837364e-01 6.87415361e-01 -8.55571255e-02 1.42115787e-01
2.42689759e-01 -2.80616432e-01 6.27867579e-01 1.95810831e+00
3.13081264e-01 -3.87998402e-01 5.28645217e-01 3.68576407e-01
-3.74436855e-01 9.29497898e-01 -1.00598264e+00 -4.55247909e-01
3.52134526e-01 1.18148887e+00 -2.08894923e-01 -6.51253164e-01
-6.95401311e-01 1.07566082e+00 6.82483733e-01 2.03849763e-01
-5.35674274e-01 -5.40756881e-01 1.16352630e+00 -5.99687509e-02
2.17492014e-01 -4.92987335e-01 -1.77538320e-02 -1.64737606e+00
-6.59518242e-02 -9.35504317e-01 6.55540824e-01 -7.02053547e-01
-1.52756059e+00 1.04243791e+00 -5.30095324e-02 -6.80689871e-01
-3.22114497e-01 -1.03470564e+00 -4.10305589e-01 9.80208278e-01
-2.02235198e+00 -1.63708270e+00 5.25500238e-01 3.89267802e-01
7.79588580e-01 -6.67877853e-01 1.64812386e+00 1.93244413e-01
-3.51801574e-01 7.72619486e-01 -1.21507188e-02 3.59938771e-01
7.38230884e-01 -1.71388412e+00 8.26754451e-01 9.58731890e-01
8.90054166e-01 7.30627000e-01 5.20674944e-01 -4.62611586e-01
-1.33367169e+00 -1.29803741e+00 1.63105440e+00 -9.27601099e-01
1.49937296e+00 -7.38749802e-01 -9.84083295e-01 1.24539077e+00
8.03254545e-01 2.04650566e-01 1.01117611e+00 7.93933392e-01
-8.13280582e-01 1.73106324e-02 -5.07066429e-01 8.20768893e-01
8.96417260e-01 -1.02076614e+00 -1.34813762e+00 3.09488505e-01
1.07836485e+00 -3.44205238e-02 -1.24077845e+00 2.99252812e-02
5.44533789e-01 5.87867722e-02 8.34518731e-01 -1.25035548e+00
2.35905796e-01 -1.96496502e-01 -6.48301840e-01 -1.61911011e+00
-2.45065928e-01 -6.11194372e-01 1.03934094e-01 1.26003444e+00
5.51657736e-01 -7.54194796e-01 2.50453740e-01 -2.66239643e-01
-2.76336551e-01 -3.95646125e-01 -1.03751707e+00 -1.00639427e+00
8.45506430e-01 -8.48406672e-01 6.05434835e-01 1.38504100e+00
3.04133266e-01 8.46397638e-01 -2.10935593e-01 1.02039739e-01
3.78591418e-01 -1.24240115e-01 5.76237440e-01 -9.52577472e-01
-2.53648609e-01 -1.54886216e-01 -6.75782323e-01 -8.42569768e-01
8.99060607e-01 -1.55870211e+00 4.70957048e-02 -1.43059111e+00
-1.78293288e-01 -1.91808462e-01 -7.67954350e-01 1.00648463e+00
-2.77868658e-01 3.75197113e-01 -3.78792509e-02 -4.40490931e-01
-5.46218872e-01 6.63344383e-01 5.77741325e-01 1.19134281e-02
2.41261423e-01 -8.41670334e-01 -4.83052284e-01 6.57236397e-01
6.31593287e-01 -8.46544206e-01 -4.70972285e-02 -7.28669167e-01
3.43887359e-01 -4.82792586e-01 -2.57078171e-01 -4.61044848e-01
-3.13961387e-01 -2.94672757e-01 1.32702768e-01 -2.96500981e-01
1.28506735e-01 -7.56551504e-01 -1.25246167e-01 1.14551686e-01
-3.18315119e-01 8.13022554e-01 3.37443620e-01 8.91626105e-02
-2.63384402e-01 -3.85985315e-01 5.69868445e-01 -2.10553929e-01
-7.88394690e-01 5.34331538e-02 -7.79007375e-01 5.26827693e-01
7.12868989e-01 1.67291135e-01 1.26264542e-01 -3.14071239e-03
-5.96537471e-01 2.04231869e-02 4.34937119e-01 9.01253998e-01
4.52119291e-01 -1.82143414e+00 -1.12555790e+00 4.47630674e-01
7.81531572e-01 -5.35722613e-01 -4.29642141e-01 4.52082902e-01
-2.06478283e-01 4.18918431e-01 6.19004183e-02 -8.97432491e-02
-1.30145359e+00 5.32338381e-01 1.08707756e-01 -3.85536879e-01
-7.52605498e-01 7.34050870e-01 4.37907204e-02 -1.18322313e+00
-2.64770329e-01 -4.79476094e-01 -3.19882423e-01 -1.31508829e-02
5.72649181e-01 -7.66050965e-02 1.13083042e-01 -1.01842785e+00
-5.60081124e-01 5.36628127e-01 -1.51408106e-01 -1.28859311e-01
1.42278707e+00 -8.03824980e-03 -4.00278419e-01 8.64619255e-01
1.25099301e+00 6.65448844e-01 -3.01733524e-01 -4.39022213e-01
4.69029665e-01 -2.62613684e-01 -1.45518169e-01 -2.08038703e-01
-5.48032284e-01 8.46420228e-01 2.80613571e-01 -2.86242425e-01
5.08442402e-01 1.57587931e-01 8.61529171e-01 6.31921887e-01
2.49489978e-01 -1.37399232e+00 -8.23245868e-02 9.76278663e-01
7.26375759e-01 -1.00009298e+00 -2.79663384e-01 7.85321519e-02
-7.27976143e-01 1.28039026e+00 2.59865373e-01 -4.32355106e-01
6.30869508e-01 4.58787620e-01 4.45990682e-01 2.56200656e-02
-7.90188491e-01 -3.01875502e-01 2.95156807e-01 8.20088029e-01
9.73382413e-01 5.49664140e-01 -4.59249318e-01 7.85796463e-01
-4.65162933e-01 -4.11286473e-01 3.71589243e-01 7.30213046e-01
-1.07173316e-01 -1.87404394e+00 -5.89353815e-02 1.54988453e-01
-7.98275650e-01 -8.08114767e-01 -1.52503997e-01 8.85195673e-01
3.77153531e-02 6.47803009e-01 4.68153507e-02 -1.09767146e-01
4.41162884e-01 3.35620254e-01 3.96452844e-01 -1.19409811e+00
-3.48766714e-01 -6.04113229e-02 6.84113622e-01 -4.36593294e-01
-1.56841204e-01 -5.26373267e-01 -1.13761187e+00 -4.43703979e-02
-2.52683192e-01 3.37276220e-01 4.54204500e-01 1.09267461e+00
-1.57818422e-01 4.73171085e-01 3.82133126e-01 -3.45859230e-01
-2.09666967e-01 -1.00769925e+00 -3.44890326e-01 4.95386928e-01
6.21864758e-02 -1.23854078e-01 1.00971811e-01 2.14821640e-02] | [11.009774208068848, 9.964348793029785] |
3ca0e8d7-49c0-45fd-a7b0-68e7103d8e56 | inverse-rendering-of-translucent-objects | 2305.08336 | null | https://arxiv.org/abs/2305.08336v1 | https://arxiv.org/pdf/2305.08336v1.pdf | Inverse Rendering of Translucent Objects using Physical and Neural Renderers | In this work, we propose an inverse rendering model that estimates 3D shape, spatially-varying reflectance, homogeneous subsurface scattering parameters, and an environment illumination jointly from only a pair of captured images of a translucent object. In order to solve the ambiguity problem of inverse rendering, we use a physically-based renderer and a neural renderer for scene reconstruction and material editing. Because two renderers are differentiable, we can compute a reconstruction loss to assist parameter estimation. To enhance the supervision of the proposed neural renderer, we also propose an augmented loss. In addition, we use a flash and no-flash image pair as the input. To supervise the training, we constructed a large-scale synthetic dataset of translucent objects, which consists of 117K scenes. Qualitative and quantitative results on both synthetic and real-world datasets demonstrated the effectiveness of the proposed model. | ['Hajime Nagahara', 'Trung Thanh Ngo', 'Chenhao Li'] | 2023-05-15 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_Inverse_Rendering_of_Translucent_Objects_Using_Physical_and_Neural_Renderers_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_Inverse_Rendering_of_Translucent_Objects_Using_Physical_and_Neural_Renderers_CVPR_2023_paper.pdf | cvpr-2023-1 | ['inverse-rendering'] | ['computer-vision'] | [ 6.80268288e-01 -2.62233824e-01 8.54418457e-01 -5.05249262e-01
-4.48704720e-01 -3.18494469e-01 4.60600436e-01 -6.04594648e-01
-2.15394840e-01 4.71342146e-01 -1.46000758e-01 -2.67706484e-01
-9.99811590e-02 -7.86026180e-01 -8.65507424e-01 -8.50550473e-01
4.81514663e-01 2.32529461e-01 -1.01962574e-02 7.95896165e-03
2.55312920e-01 4.42291945e-01 -1.53632188e+00 1.20709427e-01
1.27690661e+00 1.13701248e+00 7.06622958e-01 6.64135993e-01
6.66795224e-02 7.72119045e-01 -3.26600373e-01 -1.42430574e-01
4.13129985e-01 -2.58295715e-01 -3.60024154e-01 3.58680844e-01
3.58054489e-01 -8.47400010e-01 -3.30525786e-01 9.03053641e-01
6.30040050e-01 4.35452104e-01 7.78841019e-01 -8.67564619e-01
-8.62738967e-01 -1.73072964e-01 -6.51578903e-01 -3.03939044e-01
2.59740323e-01 3.03871930e-01 3.98814052e-01 -9.39010024e-01
2.89176613e-01 1.25760126e+00 3.99576426e-01 3.57082039e-01
-9.79583025e-01 -6.99085355e-01 1.27580687e-01 -1.93450972e-01
-1.11891437e+00 -6.19052052e-01 1.10283530e+00 -2.58984864e-01
4.67984229e-01 4.27669913e-01 7.17784822e-01 7.20341861e-01
-3.05296890e-02 5.36022246e-01 1.51360023e+00 -5.13368309e-01
9.77671221e-02 1.73709854e-01 -1.33120358e-01 8.85436237e-01
-6.56096637e-02 4.36910391e-01 -2.30908766e-01 -3.06965441e-01
1.09007764e+00 1.18191250e-01 -5.20163059e-01 -1.85874075e-01
-1.10514092e+00 3.52195561e-01 4.46069092e-01 -5.81088364e-01
-3.97243142e-01 8.59764740e-02 -1.30163897e-02 2.65430361e-01
8.41064453e-01 2.57881522e-01 -7.24576116e-02 3.19936097e-01
-3.28673273e-01 -2.79093292e-02 6.79019690e-01 8.66407096e-01
8.18357468e-01 1.67034641e-01 -4.18270640e-02 1.08103955e+00
8.88589382e-01 1.09096432e+00 -8.15477297e-02 -1.33262527e+00
5.29601157e-01 3.05663496e-01 4.36859220e-01 -8.53724778e-01
4.64349501e-02 -3.75110596e-01 -9.61844623e-01 5.14314473e-01
-2.28910390e-02 4.96546924e-02 -1.14398050e+00 1.45250285e+00
6.70271635e-01 5.13392389e-01 1.22154407e-01 1.34014511e+00
6.86129510e-01 9.71004844e-01 -3.05898607e-01 -1.93863124e-01
9.81269598e-01 -1.06171858e+00 -5.95932722e-01 2.36669686e-02
1.53636634e-01 -9.98005033e-01 1.27140391e+00 5.14140606e-01
-1.11817074e+00 -3.95128995e-01 -9.15901601e-01 -3.24203491e-01
2.20636725e-01 4.49121535e-01 6.29979968e-01 3.12090784e-01
-6.34598017e-01 5.23009419e-01 -7.21934795e-01 9.98171121e-02
1.93295643e-01 3.78262885e-02 6.44991249e-02 -4.23950195e-01
-9.38588440e-01 4.64821041e-01 8.79433472e-03 7.24860728e-01
-1.19059336e+00 -7.58340716e-01 -5.75994194e-01 -1.71807662e-01
2.64605194e-01 -9.26134109e-01 8.13769639e-01 -1.01732695e+00
-1.97213781e+00 7.93442607e-01 5.84036633e-02 4.21056241e-01
6.06622994e-01 -3.51861835e-01 -4.36644644e-01 1.82887897e-01
-2.09835827e-01 4.27697375e-02 1.10590374e+00 -1.98108220e+00
-2.72896618e-01 -2.13097647e-01 2.05403879e-01 5.49681246e-01
4.89299633e-02 -1.72870591e-01 -4.13562059e-01 -4.30105567e-01
4.50275123e-01 -6.39890254e-01 -1.53567672e-01 6.40813231e-01
-5.83890676e-01 5.73836863e-01 6.30579054e-01 -1.01967251e+00
4.50079769e-01 -2.29839087e+00 -1.11431189e-01 1.37262493e-01
2.45383531e-01 2.80685902e-01 -5.87071598e-01 7.23090991e-02
7.31912777e-02 -2.91919202e-01 -5.18308997e-01 -5.99502504e-01
-1.72074378e-01 1.81576326e-01 -5.01479328e-01 6.06435955e-01
2.07603793e-03 6.20533764e-01 -7.92417884e-01 -2.49370754e-01
4.02198493e-01 9.41073120e-01 -4.85602826e-01 6.69619024e-01
-3.40884060e-01 7.19656944e-01 -7.10275531e-01 5.26517212e-01
1.24986243e+00 -1.43733054e-01 -2.94460684e-01 -3.60617936e-01
-2.24762738e-01 -2.97916331e-03 -1.09192121e+00 1.71112609e+00
-9.96538460e-01 3.00946146e-01 3.97773385e-01 -5.13495624e-01
1.18053102e+00 9.12996009e-02 1.96986973e-01 -9.34397519e-01
2.05943823e-01 2.34139472e-01 -2.93442965e-01 -7.94483125e-01
3.96788478e-01 -2.07324401e-01 8.68250489e-01 7.08088458e-01
-5.25757134e-01 -5.26904345e-01 -5.64435124e-01 -1.45374730e-01
8.00434589e-01 4.75116521e-01 -2.52888292e-01 -5.11543192e-02
5.98635018e-01 -3.81771266e-01 3.88909489e-01 5.01098931e-01
3.61973464e-01 1.00948191e+00 2.70265248e-02 -5.28753281e-01
-1.13551211e+00 -1.27380407e+00 4.33484726e-02 7.44837105e-01
5.25498509e-01 3.91348213e-01 -5.93111753e-01 -5.94382882e-02
-1.80259719e-01 6.60998702e-01 -3.19299161e-01 -9.06745791e-02
-6.81218326e-01 -7.41663873e-01 8.18260610e-02 -9.24078654e-03
9.34424400e-01 -9.50815678e-01 -5.49462140e-01 -1.00021623e-01
-2.40859464e-01 -1.00174093e+00 -5.01457989e-01 -1.97608963e-01
-7.59825468e-01 -1.07648420e+00 -7.26484239e-01 -5.90143144e-01
9.60690439e-01 5.95059454e-01 1.11982048e+00 2.25521684e-01
-2.80342966e-01 2.97145635e-01 -1.03321657e-01 -3.29040647e-01
-5.15733600e-01 -4.63398695e-01 -3.91414672e-01 4.62075353e-01
-6.71645284e-01 -7.24603295e-01 -7.89214551e-01 3.85794520e-01
-1.02381551e+00 5.43694854e-01 4.52456445e-01 7.67362773e-01
6.50420606e-01 -1.94898665e-01 1.32116035e-01 -7.98761427e-01
5.29953539e-01 -2.69924045e-01 -9.43747520e-01 4.46179539e-01
-5.61313987e-01 -1.01864576e-01 6.38310313e-01 -5.37358344e-01
-1.91180968e+00 -4.54370640e-02 9.30340681e-03 -6.55342460e-01
-3.19025479e-02 3.12031806e-01 -3.41828018e-01 -2.70301819e-01
4.58950728e-01 4.52653170e-01 3.88737000e-03 -5.39067328e-01
1.49214670e-01 7.27324665e-01 5.14594913e-01 -8.06386709e-01
9.44692314e-01 8.84422898e-01 1.03100970e-01 -9.05341744e-01
-7.30575562e-01 -1.21559463e-01 -2.63146579e-01 -3.19401681e-01
6.22693062e-01 -9.84880328e-01 -9.55342174e-01 9.55474615e-01
-1.37877810e+00 -6.81744874e-01 -2.50236243e-01 4.68174666e-01
-5.71163714e-01 3.84411365e-01 -5.70805669e-01 -9.37245727e-01
-3.75097126e-01 -1.01761913e+00 1.33262670e+00 2.47366235e-01
6.61510706e-01 -9.43439603e-01 1.68108921e-02 4.61194664e-01
4.42219943e-01 2.44093046e-01 7.67224252e-01 4.87039596e-01
-1.14450908e+00 3.04303706e-01 -7.93782711e-01 6.64158940e-01
3.21306646e-01 9.37669277e-02 -1.32420158e+00 -2.08419397e-01
4.14818674e-01 -3.65020484e-01 8.46453547e-01 3.46486628e-01
1.36484742e+00 -2.32912555e-01 7.09439255e-03 1.34733474e+00
1.52814925e+00 2.29697317e-01 7.62347341e-01 7.99052417e-02
1.04794359e+00 7.15762734e-01 7.70744741e-01 5.44220030e-01
3.51871848e-01 3.13910246e-01 7.27537155e-01 -3.51277888e-01
-2.79749155e-01 -1.68246344e-01 8.99858475e-02 1.00165308e+00
-4.42048907e-01 -6.51557505e-01 -5.89286208e-01 3.46273854e-02
-1.42856622e+00 -4.67091888e-01 -2.61500657e-01 2.29277730e+00
6.41067684e-01 -3.37781340e-01 -8.03739130e-01 -2.43949458e-01
5.02769887e-01 3.06841403e-01 -7.17145920e-01 -2.55915821e-02
-1.51690468e-01 -1.09118044e-01 3.09907973e-01 7.47690439e-01
-6.14403784e-01 6.72249973e-01 5.90276814e+00 6.46402001e-01
-1.45371270e+00 -3.66943404e-02 5.95299363e-01 5.86482212e-02
-1.05617559e+00 -4.45364267e-02 -2.07992688e-01 4.45598215e-01
3.73232126e-01 3.07095736e-01 9.79556143e-01 2.33459473e-01
5.04821718e-01 -3.20266157e-01 -8.27070236e-01 1.04908848e+00
1.12802096e-01 -1.02838206e+00 4.23054285e-02 -1.11856461e-01
8.14623296e-01 1.83659550e-02 3.10603648e-01 -1.62817404e-01
2.83154398e-01 -8.17225099e-01 7.66127706e-01 1.09071314e+00
9.81935024e-01 -3.48825336e-01 1.65372178e-01 4.37160879e-01
-9.40517962e-01 2.43168607e-01 -4.17290539e-01 2.71447971e-02
2.60043830e-01 7.95479178e-01 -4.68501061e-01 7.15385854e-01
3.84788394e-01 5.73727489e-01 -1.61110401e-01 9.96336162e-01
-3.00812542e-01 4.50124294e-01 -3.83707911e-01 3.73025715e-01
-1.54329017e-01 -7.59478629e-01 4.42499995e-01 6.48412824e-01
3.92655373e-01 5.75156629e-01 1.12034507e-01 1.39179885e+00
-4.72235717e-02 -2.62791455e-01 -3.79436791e-01 3.25727999e-01
3.25142086e-01 1.26844203e+00 -4.06847119e-01 -1.81554198e-01
-2.10469261e-01 1.25151718e+00 1.92788258e-01 1.08239043e+00
-9.39864099e-01 -2.98351645e-01 5.28432429e-01 -4.40130606e-02
-2.75840253e-01 -2.63049811e-01 -8.10308531e-02 -1.21572304e+00
2.95210063e-01 -5.88847041e-01 -4.73702878e-01 -1.47148013e+00
-1.25915265e+00 4.77355152e-01 -1.92704231e-01 -1.21885657e+00
3.02104712e-01 -5.97101212e-01 -6.69593275e-01 1.48088813e+00
-2.09987020e+00 -1.31095541e+00 -8.07913423e-01 5.03409624e-01
1.85774043e-01 1.97781041e-01 5.29563367e-01 3.33818257e-01
-2.71860540e-01 -1.32332146e-01 3.64876688e-01 -1.87071174e-01
7.89428949e-01 -8.15918148e-01 3.13089669e-01 6.33316576e-01
-5.06012976e-01 4.74681050e-01 5.20559788e-01 -4.91285354e-01
-1.56412733e+00 -1.22405934e+00 -1.26601604e-03 4.68388293e-03
1.98212147e-01 -3.31622750e-01 -9.72748637e-01 4.72913921e-01
-4.57109921e-02 4.11680251e-01 8.77179727e-02 -4.84786898e-01
-2.05494165e-01 -2.32531145e-01 -1.24048638e+00 4.05952811e-01
1.14639127e+00 -6.83384776e-01 -2.99093430e-03 3.43229800e-01
9.67798889e-01 -8.31711233e-01 -6.79467916e-01 2.89431751e-01
6.69697464e-01 -1.02896762e+00 1.18576407e+00 -1.39626358e-02
7.40683734e-01 -3.49061728e-01 -1.66821018e-01 -1.54054654e+00
-1.98301226e-01 -4.55587655e-01 5.31457104e-02 1.22350442e+00
-3.19785252e-02 -1.03946817e+00 6.23806715e-01 7.85292804e-01
-4.59996253e-01 -6.51503682e-01 -4.70188439e-01 -5.69743693e-01
-2.77082026e-01 -1.90943196e-01 8.14027071e-01 7.93605924e-01
-1.17476714e+00 7.37483725e-02 -4.89787251e-01 6.46192789e-01
1.06016588e+00 3.69883448e-01 6.38489544e-01 -1.04686999e+00
-4.89657104e-01 1.52273014e-01 3.65423381e-01 -1.38382113e+00
3.00240129e-01 -6.29108667e-01 4.13573921e-01 -1.36896110e+00
2.69268036e-01 -1.09384120e+00 -1.47202209e-01 -4.08653077e-03
-1.66735500e-01 -2.63215750e-02 -9.51333344e-02 2.56207526e-01
-1.38411298e-01 1.05128038e+00 2.05144644e+00 -2.35256717e-01
-1.38625354e-01 1.45052582e-01 -3.47875267e-01 5.71268976e-01
6.52601779e-01 -3.83471876e-01 -7.74282753e-01 -1.23047662e+00
3.58606309e-01 3.50367248e-01 7.77217150e-01 -5.26795208e-01
-6.39029741e-02 -4.62842405e-01 2.49450937e-01 -4.00161773e-01
7.74130642e-01 -9.92261231e-01 2.40766138e-01 1.37349978e-01
-3.37604880e-01 -5.17128825e-01 -1.02273993e-01 5.42910814e-01
3.12542878e-02 -1.88572764e-01 8.23782563e-01 -1.38132691e-01
-3.21730822e-01 6.03685379e-01 3.11087996e-01 -1.95259348e-01
5.82176030e-01 -7.26479068e-02 -4.71134663e-01 -1.77211687e-01
-1.25644788e-01 1.36914268e-01 7.57626116e-01 2.36022919e-02
9.30582225e-01 -1.15970695e+00 -6.83751285e-01 4.76911724e-01
-3.82590070e-02 5.42089403e-01 5.32724261e-01 6.80049181e-01
-9.08826172e-01 -2.88451195e-01 -1.21975504e-01 -4.70876426e-01
-1.06516397e+00 2.86450118e-01 5.88741660e-01 1.58820346e-01
-6.03041172e-01 8.05158138e-01 6.80563152e-01 -9.66785192e-01
-1.42182916e-01 -5.56032062e-01 2.05890656e-01 -6.43980205e-01
4.18345153e-01 5.78395545e-01 -8.02704915e-02 -3.49061787e-01
9.90934372e-02 6.77310824e-01 3.83169234e-01 -2.28856161e-01
1.44279635e+00 -4.86330479e-01 -3.56335640e-01 4.14491415e-01
1.20294940e+00 -1.06710725e-01 -1.75641215e+00 -2.86685079e-01
-9.44115877e-01 -7.88183153e-01 2.55548120e-01 -7.62927949e-01
-1.33495426e+00 1.10856497e+00 4.96698916e-01 -2.16827527e-01
1.39627457e+00 -4.37138945e-01 7.43419528e-01 5.13825893e-01
2.99977720e-01 -7.76157737e-01 8.93108100e-02 5.10826349e-01
1.06639040e+00 -1.28454638e+00 2.05781236e-02 -7.82702684e-01
-2.63973027e-01 1.03753793e+00 5.18585920e-01 -1.44795448e-01
5.66224575e-01 3.22242916e-01 2.85555631e-01 -1.09158099e-01
-6.15961373e-01 4.06154633e-01 2.85090208e-01 6.15426302e-01
1.70112252e-01 -4.10229638e-02 3.42836022e-01 -4.46078256e-02
1.27658963e-01 2.26579517e-01 5.17932057e-01 6.89884543e-01
-2.43104756e-01 -6.96395099e-01 -5.60960770e-01 3.96831393e-01
2.01409712e-01 -1.50129274e-01 1.49597988e-01 1.95983291e-01
-5.23245744e-02 6.83607399e-01 -4.50646356e-02 -1.57265723e-01
2.44010791e-01 -5.61800897e-01 5.74221551e-01 -3.71826887e-01
-1.14137582e-01 8.30548555e-02 -1.15389511e-01 -5.77395380e-01
-4.20989573e-01 -4.17239487e-01 -1.13046277e+00 -2.52342373e-01
-4.93725300e-01 -2.25387380e-01 8.43995094e-01 7.16955364e-01
1.60240099e-01 7.09750891e-01 1.04963672e+00 -1.02224612e+00
-5.64433098e-01 -7.30879188e-01 -1.04341447e+00 2.54230767e-01
6.38855994e-01 -5.39490759e-01 -5.75409651e-01 2.06353933e-01] | [9.80324649810791, -3.0611770153045654] |
6e796397-d8d9-4fac-81cf-4c9a1690e392 | deep-learning-for-detecting-multiple-space | 1608.01529 | null | http://arxiv.org/abs/1608.01529v1 | http://arxiv.org/pdf/1608.01529v1.pdf | Deep Learning for Detecting Multiple Space-Time Action Tubes in Videos | In this work, we propose an approach to the spatiotemporal localisation
(detection) and classification of multiple concurrent actions within temporally
untrimmed videos. Our framework is composed of three stages. In stage 1,
appearance and motion detection networks are employed to localise and score
actions from colour images and optical flow. In stage 2, the appearance network
detections are boosted by combining them with the motion detection scores, in
proportion to their respective spatial overlap. In stage 3, sequences of
detection boxes most likely to be associated with a single action instance,
called action tubes, are constructed by solving two energy maximisation
problems via dynamic programming. While in the first pass, action paths
spanning the whole video are built by linking detection boxes over time using
their class-specific scores and their spatial overlap, in the second pass,
temporal trimming is performed by ensuring label consistency for all
constituting detection boxes. We demonstrate the performance of our algorithm
on the challenging UCF101, J-HMDB-21 and LIRIS-HARL datasets, achieving new
state-of-the-art results across the board and significantly increasing
detection speed at test time. We achieve a huge leap forward in action
detection performance and report a 20% and 11% gain in mAP (mean average
precision) on UCF-101 and J-HMDB-21 datasets respectively when compared to the
state-of-the-art. | ['Fabio Cuzzolin', 'Suman Saha', 'Philip H. S. Torr', 'Michael Sapienza', 'Gurkirt Singh'] | 2016-08-04 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 4.57623959e-01 -2.70643085e-01 -1.11553684e-01 -4.33044843e-02
-7.70752788e-01 -5.48218548e-01 6.15057588e-01 -6.42955974e-02
-7.60361731e-01 6.58845544e-01 6.84611648e-02 1.90357283e-01
-2.92841911e-01 -2.07041442e-01 -6.46203697e-01 -7.78196633e-01
-6.21659338e-01 2.01750517e-01 1.01054585e+00 1.40442818e-01
2.23656416e-01 4.18818891e-01 -1.60973489e+00 7.12972462e-01
3.83962303e-01 1.24605691e+00 2.39367813e-01 1.09615326e+00
3.13221157e-01 1.31061876e+00 -4.61304039e-01 -9.89765525e-02
3.54020089e-01 -4.53232288e-01 -8.68819714e-01 2.73376942e-01
7.17813194e-01 -3.30314517e-01 -4.18426514e-01 7.80712068e-01
3.68064135e-01 5.76122582e-01 5.52569270e-01 -1.36411023e+00
4.50739004e-02 7.66028985e-02 -8.66233945e-01 5.41778982e-01
4.53059345e-01 2.32558444e-01 9.60735738e-01 -8.47975254e-01
8.92510653e-01 1.21972287e+00 4.31883216e-01 5.84699452e-01
-1.18396235e+00 -4.40952629e-01 2.67998517e-01 4.84593421e-01
-1.23379016e+00 -5.35592675e-01 4.03506517e-01 -5.53046644e-01
1.20196295e+00 2.37562731e-01 6.30410492e-01 8.68226230e-01
-8.84760395e-02 1.03063643e+00 7.11163461e-01 -3.92640054e-01
2.90919721e-01 -1.38072222e-01 -2.94604331e-01 7.97252834e-01
-4.03258920e-01 9.52707157e-02 -7.47603714e-01 1.55212730e-01
8.91054332e-01 -1.89411625e-01 -7.69033581e-02 -4.88076061e-01
-1.56562483e+00 5.33856869e-01 4.75690186e-01 2.21932352e-01
-6.28369570e-01 4.32627171e-01 4.65018362e-01 3.83954123e-03
3.83005917e-01 1.58890322e-01 -3.79089624e-01 -4.77039129e-01
-1.28181422e+00 1.80048943e-02 4.71201539e-01 7.26497233e-01
4.33391958e-01 -3.12207282e-01 -5.72877586e-01 8.13060045e-01
1.60759419e-01 1.36448041e-01 3.37855875e-01 -1.28797042e+00
5.29226542e-01 6.70744240e-01 2.82548100e-01 -1.01842272e+00
-4.91389632e-01 -1.43289119e-01 -6.26888573e-01 3.80799264e-01
5.72034061e-01 -1.62957404e-02 -9.83434618e-01 1.72342157e+00
6.05965614e-01 5.61407387e-01 6.20978244e-05 9.92715478e-01
2.83390880e-01 6.42196000e-01 1.73231497e-01 -3.52301776e-01
1.25826395e+00 -1.31555462e+00 -3.66863728e-01 -3.30036730e-01
6.08246148e-01 -6.20825112e-01 5.02493382e-01 5.75861812e-01
-1.17208326e+00 -8.57555509e-01 -8.63337815e-01 2.32765630e-01
-2.20131189e-01 4.94577527e-01 4.03582215e-01 1.89922273e-01
-1.11466432e+00 6.45990968e-01 -1.01625693e+00 -4.77536559e-01
5.58310270e-01 2.89840400e-01 -5.49853384e-01 -4.31957170e-02
-9.08907771e-01 6.32044375e-01 4.71344382e-01 1.44658312e-01
-1.31064534e+00 -3.36612016e-01 -5.53483665e-01 -3.61177355e-01
5.64619780e-01 -4.03425664e-01 1.10798061e+00 -1.05487370e+00
-1.17831826e+00 7.89038956e-01 -2.42467955e-01 -7.45097339e-01
8.42130482e-01 -2.15546504e-01 -4.67514962e-01 5.69828928e-01
1.13248706e-01 1.20293188e+00 8.06702971e-01 -8.83920252e-01
-1.50881755e+00 -1.11617602e-01 2.75634050e-01 3.13589185e-01
-1.11310199e-01 2.95300335e-01 -9.96738434e-01 -4.65045720e-01
1.06175862e-01 -9.71352577e-01 -1.78129658e-01 2.83007473e-01
-2.19703272e-01 -1.94773450e-01 8.32676828e-01 -8.04712236e-01
1.45401359e+00 -2.12454629e+00 4.95168626e-01 5.43237552e-02
2.03044247e-02 2.06457123e-01 -1.12241887e-01 -7.96878152e-03
-3.91069092e-02 -2.75741756e-01 -1.71473235e-01 -4.59324419e-01
-1.84362665e-01 9.75123495e-02 8.12229738e-02 7.07392752e-01
3.08526725e-01 7.52548337e-01 -1.13810027e+00 -6.49524450e-01
6.51880920e-01 3.62335414e-01 -4.15703535e-01 2.96321157e-02
-1.76243186e-01 4.84851390e-01 -5.23341000e-02 6.55315399e-01
2.13526428e-01 -1.46636382e-01 1.58027578e-02 -1.82327583e-01
-6.79899976e-02 3.45640490e-03 -1.63103831e+00 1.82707345e+00
-1.70173869e-01 7.82143414e-01 4.34301607e-03 -9.22864437e-01
5.49151719e-01 3.30676645e-01 8.45526099e-01 -7.60820627e-01
-2.97338992e-01 -8.63757282e-02 1.09766431e-01 -5.76556206e-01
4.92010057e-01 4.68283892e-01 -5.19295223e-03 2.72794604e-01
1.13273896e-01 4.94806677e-01 7.26153791e-01 2.86964267e-01
1.32404125e+00 6.13742292e-01 2.22177297e-01 1.29320934e-01
7.93695092e-01 1.78931449e-02 4.70452994e-01 7.32300758e-01
-6.02466881e-01 6.38344109e-01 5.97680509e-01 -3.64351749e-01
-8.55938613e-01 -9.49501693e-01 1.46077216e-01 1.24455881e+00
1.25415802e-01 -2.80370891e-01 -7.34701455e-01 -9.08087730e-01
-1.28346682e-01 3.24769080e-01 -9.58497524e-01 -1.70754902e-02
-6.77383006e-01 -6.00771785e-01 4.63488758e-01 6.67625129e-01
7.01520085e-01 -1.37378848e+00 -1.14664340e+00 4.52103794e-01
-3.09982508e-01 -1.22585642e+00 -5.46991706e-01 2.35384911e-01
-7.18328953e-01 -1.19299018e+00 -7.75336981e-01 -5.81667542e-01
6.00046873e-01 2.49262586e-01 9.19788241e-01 -3.04676984e-02
-6.63986146e-01 2.54953265e-01 -3.92563671e-01 2.32048467e-01
-2.05458716e-01 -3.26836020e-01 -4.55916338e-02 4.20586109e-01
1.76990982e-02 -1.60356596e-01 -1.03870869e+00 7.09021986e-01
-6.57535017e-01 -3.75382118e-02 6.60857737e-01 5.16963661e-01
7.11952329e-01 8.99892207e-03 8.57631564e-02 -3.10952246e-01
-9.59967375e-02 -3.54823261e-01 -6.46632552e-01 2.64118463e-01
-1.15147784e-01 -1.92267969e-01 1.66331038e-01 -6.25801384e-01
-9.07132745e-01 6.11273646e-01 3.08290720e-01 -7.44852066e-01
-1.97440773e-01 -4.53767404e-02 2.27074608e-01 -3.82258967e-02
6.36112750e-01 2.61068225e-01 -2.74504364e-01 -2.21376151e-01
5.10159016e-01 4.22429383e-01 9.79782760e-01 -1.03018150e-01
3.05113465e-01 7.04869151e-01 -3.34919095e-02 -5.70343316e-01
-6.43906832e-01 -9.80200529e-01 -1.03241956e+00 -9.24412310e-01
1.34523261e+00 -8.03582489e-01 -4.66651589e-01 5.05203009e-01
-9.95429575e-01 -4.47505683e-01 -1.49610072e-01 4.14322764e-01
-7.02715755e-01 2.45269343e-01 -3.78190786e-01 -1.05255830e+00
-1.01048104e-01 -1.00690401e+00 1.25537634e+00 1.18507497e-01
-2.66335160e-01 -7.88292766e-01 7.79450387e-02 2.59015650e-01
8.32757354e-02 4.89529431e-01 3.05458933e-01 -3.98805141e-01
-6.98967397e-01 -3.24541777e-01 -2.31769562e-01 3.64047587e-01
6.66155131e-04 1.37579843e-01 -8.85829449e-01 -2.90808529e-01
-5.80828607e-01 -1.30137473e-01 9.08484101e-01 6.38038397e-01
9.13209736e-01 9.91190691e-03 -5.45242012e-01 3.83692801e-01
1.28032768e+00 4.69629079e-01 6.67592824e-01 5.18091023e-01
5.34854531e-01 6.81634486e-01 9.91531610e-01 4.37756658e-01
-4.76125255e-02 1.17087579e+00 5.21811128e-01 -1.20157897e-01
-4.22986776e-01 1.76981390e-02 6.11321807e-01 4.05077115e-02
-1.87748268e-01 -1.44968748e-01 -7.13602006e-01 7.91152179e-01
-2.20431614e+00 -1.25974274e+00 -2.99956769e-01 2.40430641e+00
3.32215697e-01 3.75505596e-01 6.97869420e-01 2.32363626e-01
7.59051681e-01 2.46858135e-01 -3.88739228e-01 8.84092599e-02
-1.24444351e-01 -1.90587118e-01 4.59120661e-01 3.39669883e-01
-1.63692045e+00 9.14365947e-01 5.71064997e+00 7.70953357e-01
-7.26328015e-01 2.68420398e-01 7.15895951e-01 -6.21857524e-01
7.67378986e-01 -4.94165309e-02 -7.12437928e-01 5.59325755e-01
7.95273185e-01 4.69584435e-01 4.42621917e-01 4.78843957e-01
4.69823390e-01 -6.84394181e-01 -1.14433205e+00 9.26966012e-01
1.63476225e-02 -1.17658734e+00 -4.91230339e-01 -1.15914918e-01
8.90548289e-01 7.73555264e-02 -2.64441729e-01 8.87738094e-02
2.17539325e-01 -7.23355830e-01 9.37119842e-01 6.65099144e-01
5.94271064e-01 -7.96283245e-01 5.33219218e-01 2.77644634e-01
-1.51348245e+00 -4.08941269e-01 -6.45785555e-02 1.29812770e-02
4.37259078e-01 2.38082036e-01 -5.97345531e-01 4.26085293e-01
9.02771533e-01 9.28601086e-01 -6.07172787e-01 1.30818653e+00
-3.46321493e-01 3.29014927e-01 -2.12402970e-01 1.30160168e-01
4.92429882e-01 7.25502521e-02 5.18130302e-01 1.40690613e+00
1.35589242e-01 -1.35376811e-01 2.79338866e-01 4.69803512e-01
2.95145750e-01 -2.11967915e-01 -1.91176951e-01 2.72628635e-01
3.38406563e-01 1.48974383e+00 -1.04680014e+00 -5.20336747e-01
-3.53248596e-01 1.34099007e+00 2.23996237e-01 2.81184822e-01
-1.21294487e+00 -1.67429164e-01 5.91312647e-01 -1.06864616e-01
6.05590880e-01 -4.34753783e-02 2.54599869e-01 -6.86102748e-01
8.63942727e-02 -6.56095862e-01 6.83936119e-01 -9.00546491e-01
-6.83312774e-01 5.07941544e-01 8.89322981e-02 -1.55159545e+00
-4.24748778e-01 -4.18302655e-01 -3.61401916e-01 5.73287964e-01
-1.05807519e+00 -8.78795564e-01 -3.40632081e-01 5.34821093e-01
7.52542734e-01 5.90152554e-02 4.69460130e-01 4.30888504e-01
-7.19105721e-01 2.79357016e-01 1.42001817e-02 1.06696658e-01
6.08864367e-01 -1.21582472e+00 3.94917250e-01 1.05981147e+00
6.25117719e-01 -5.89265302e-03 4.59422469e-01 -5.92122674e-01
-9.17127132e-01 -1.16440845e+00 6.18832946e-01 -5.76277494e-01
6.59919560e-01 -3.62556309e-01 -6.48131430e-01 3.44027013e-01
-1.07814565e-01 1.94366142e-01 7.37150311e-02 -4.95603234e-01
-1.34866731e-02 -5.22824563e-02 -1.04389226e+00 3.76067728e-01
1.34057951e+00 -3.54557484e-01 -2.84251779e-01 4.91756767e-01
2.85588413e-01 -4.17338133e-01 -6.93917811e-01 2.56956249e-01
7.57019639e-01 -1.10600388e+00 1.10361731e+00 -5.82191586e-01
4.30725694e-01 -6.57689750e-01 -1.29397571e-01 -8.55781436e-01
-3.93155724e-01 -6.80645764e-01 -3.17135692e-01 1.05131114e+00
3.85228515e-01 9.60352644e-02 8.93022597e-01 2.56155938e-01
1.08458005e-01 -8.16753328e-01 -1.25455654e+00 -6.92333281e-01
-6.33246720e-01 -6.62795901e-01 -1.89035282e-01 4.47074115e-01
-9.08212885e-02 -2.13469658e-03 -5.46077847e-01 -1.50968991e-02
5.08838713e-01 -1.88059047e-01 7.44737685e-01 -8.54567885e-01
-5.70222914e-01 -5.00627637e-01 -6.26696765e-01 -1.17129910e+00
-2.70945460e-01 -5.79366386e-01 2.59250313e-01 -1.50145149e+00
3.63056332e-01 5.39244153e-02 -5.45006812e-01 5.70545077e-01
-1.07495874e-01 5.73497355e-01 3.77173841e-01 3.24511707e-01
-1.26968360e+00 -1.29654016e-02 9.55207407e-01 -1.63337365e-02
-3.04915428e-01 -6.66413009e-02 7.96634629e-02 6.33523762e-01
2.68175423e-01 -4.74146038e-01 -2.72598296e-01 -1.83860540e-01
-1.21900506e-01 1.49998397e-01 6.39789701e-01 -1.47504139e+00
4.34641540e-01 -2.67776698e-02 6.32687032e-01 -8.03671598e-01
4.65512782e-01 -7.30906367e-01 1.91700310e-01 6.16829276e-01
-4.67979759e-01 1.85267359e-01 2.98593462e-01 8.67508054e-01
-9.71725583e-02 -1.36720669e-02 8.22862566e-01 -1.07200839e-01
-1.34471202e+00 8.63750353e-02 -3.59546810e-01 -1.01418778e-01
1.46134996e+00 -5.11334538e-01 -8.75446573e-02 -2.91974366e-01
-9.31193113e-01 4.02778745e-01 2.53783256e-01 6.29936218e-01
6.06371284e-01 -1.21743119e+00 -6.03634775e-01 5.34129404e-02
9.73978564e-02 -2.02092528e-01 4.85560000e-01 1.36616170e+00
-4.14021224e-01 4.37247902e-01 -1.30193412e-01 -8.66042316e-01
-1.56405485e+00 4.82131958e-01 4.70609337e-01 -3.66708189e-01
-5.41314900e-01 1.18941450e+00 2.43238181e-01 2.37787902e-01
6.60745919e-01 -1.40971571e-01 -2.75817484e-01 2.95650244e-01
6.21902287e-01 6.64938450e-01 -5.29670790e-02 -8.70413780e-01
-7.07270682e-01 6.28773630e-01 1.07348315e-01 -3.84440660e-01
1.06549263e+00 -1.47646502e-01 2.07509756e-01 1.81942180e-01
1.16189969e+00 -4.77482557e-01 -1.87860894e+00 -8.06568414e-02
1.93504959e-01 -5.81880391e-01 -7.57102016e-03 -1.04053605e+00
-1.17673671e+00 6.67149127e-01 9.42352057e-01 2.19438951e-02
1.41870034e+00 2.34274134e-01 3.28967839e-01 -2.19577458e-02
1.45453751e-01 -1.35025525e+00 4.39093173e-01 3.15355688e-01
7.07435131e-01 -1.12049651e+00 -1.49050534e-01 -1.11140840e-01
-7.33577192e-01 9.80927587e-01 6.21578336e-01 -4.94821928e-02
8.74564126e-02 1.95399448e-01 -2.44443998e-01 -2.02811286e-01
-8.47950041e-01 -4.11887527e-01 5.06579280e-01 4.53483164e-01
1.68234721e-01 -1.78612575e-01 -1.52807012e-01 -9.78994071e-02
4.74792033e-01 2.53196210e-02 9.76760909e-02 1.01293266e+00
-5.66651165e-01 -6.69393599e-01 -3.39650303e-01 3.78433645e-01
-4.16555882e-01 1.09334335e-01 -4.84717876e-01 8.30404222e-01
3.94241393e-01 1.06076491e+00 3.49765569e-01 -5.06509602e-01
4.32518721e-01 4.38780561e-02 4.60229665e-01 -2.99864411e-01
-3.47682476e-01 3.47011238e-01 3.46926510e-01 -1.03591394e+00
-7.87238598e-01 -9.34865236e-01 -1.27135015e+00 1.95010617e-01
-1.85963318e-01 -2.58668453e-01 5.21639705e-01 8.65130067e-01
3.06597680e-01 8.21998119e-01 3.91453385e-01 -1.10818183e+00
-1.98341802e-01 -9.46007669e-01 -3.62858266e-01 5.67027926e-01
2.47014567e-01 -8.85967851e-01 -3.61835122e-01 2.81323135e-01] | [8.320929527282715, 0.4316580891609192] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.