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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
4cc79f8c-96e6-4a1d-aa2c-581d03d5c796 | dialects-identification-of-armenian-language | null | null | https://aclanthology.org/2022.digitam-1.2 | https://aclanthology.org/2022.digitam-1.2.pdf | Dialects Identification of Armenian Language | The Armenian language has many dialects that differ from each other syntactically, morphologically, and phonetically. In this work, we implement and evaluate models that determine the dialect of a given passage of text. The proposed models are evaluated for the three major variations of the Armenian language: Eastern, Western, and Classical. Previously, there were no instruments of dialect identification in the Armenian language. The paper presents three approaches: a statistical which relies on a stop words dictionary, a modified statistical one with a dictionary of most frequently encountered words, and the third one that is based on Facebook’s fastText language identification neural network model. Two types of neural network models were trained, one with the usage of pre-trained word embeddings and the other without. Approaches were tested on sentence-level and document-level data. The results show that the neural network-based method works sufficiently better than the statistical ones, achieving almost 98% accuracy at the sentence level and nearly 100% at the document level. | ['Karen Avetisyan'] | null | null | null | null | digitam-lrec-2022-6 | ['dialect-identification'] | ['natural-language-processing'] | [-3.47661436e-01 -3.88686657e-01 -1.97557554e-01 -2.45933115e-01
-2.30631858e-01 -5.97270608e-01 9.26551700e-01 6.04402363e-01
-1.01770031e+00 4.41568524e-01 6.33350492e-01 -7.41921008e-01
-1.31727427e-01 -8.68042350e-01 -3.36399376e-02 -5.58251321e-01
2.47867316e-01 5.34033418e-01 1.01979733e-01 -6.91811919e-01
6.21257663e-01 4.28738385e-01 -1.63896954e+00 2.46927887e-01
7.66200185e-01 5.00064313e-01 2.24522445e-02 5.84557891e-01
-7.48992205e-01 8.23784173e-01 -5.15392661e-01 -2.88072169e-01
4.62982245e-02 -3.07195187e-01 -8.98384333e-01 -4.02317405e-01
2.22329050e-01 -4.41771001e-02 -2.83069104e-01 9.81295049e-01
5.61513484e-01 6.80636242e-02 8.94564331e-01 -5.34942567e-01
-6.97412610e-01 9.64110613e-01 9.30657727e-04 2.36358330e-01
4.58612949e-01 -2.91425973e-01 8.82888377e-01 -7.97461867e-01
6.31134391e-01 1.22441864e+00 7.78652787e-01 2.61425048e-01
-9.00111139e-01 -4.55070853e-01 -1.64261892e-01 3.21302772e-01
-1.61666727e+00 -3.16797763e-01 8.60597908e-01 -8.34619761e-01
1.00992322e+00 1.52466953e-01 4.97972429e-01 8.66705298e-01
1.46853939e-01 3.40026706e-01 1.31695449e+00 -1.13725197e+00
1.53429881e-01 6.69280648e-01 8.30446243e-01 5.23610532e-01
3.48176420e-01 -1.03105225e-01 -3.38083446e-01 -2.64591098e-01
2.56318033e-01 -3.29363018e-01 -8.01489204e-02 -4.24308032e-02
-9.32292700e-01 1.22139084e+00 -2.97331631e-01 1.32052934e+00
-3.86193871e-01 -5.49256027e-01 7.59074450e-01 3.75527054e-01
5.33347785e-01 2.99752355e-01 -4.64297950e-01 -1.77208096e-01
-1.06128192e+00 -1.30358431e-02 1.02818346e+00 4.85301167e-01
4.60714221e-01 1.86099976e-01 8.85150731e-02 1.14061427e+00
2.88400799e-01 3.11069191e-01 1.22708416e+00 1.24611277e-02
5.02216041e-01 7.29552865e-01 -2.18279034e-01 -1.20266140e+00
-4.21715319e-01 -1.65541023e-01 -6.02789819e-01 1.19991116e-01
6.53785467e-01 -3.49021971e-01 -6.31758749e-01 1.52350855e+00
9.67633352e-02 -4.30624157e-01 1.61064044e-01 4.32685256e-01
1.16019595e+00 7.90547431e-01 1.13306202e-01 -1.26044024e-02
1.34975517e+00 -7.04693258e-01 -9.00447488e-01 1.05595939e-01
8.47797811e-01 -1.03731906e+00 1.08004808e+00 3.37754309e-01
-8.91250908e-01 -6.30314767e-01 -1.22688603e+00 -3.23817916e-02
-1.10016644e+00 2.85746992e-01 1.77224085e-01 1.27504611e+00
-1.08408725e+00 3.28885972e-01 -1.32397756e-01 -7.54268229e-01
-3.46342742e-01 1.15382083e-01 -4.29581493e-01 2.11604744e-01
-1.47027004e+00 1.10504544e+00 7.36592352e-01 -1.45876020e-01
-2.19396815e-01 -2.95506656e-01 -9.88031626e-01 -8.95662159e-02
-4.21833247e-01 -1.16099849e-01 7.54449427e-01 -1.29144907e+00
-1.66153157e+00 1.18770206e+00 6.06326424e-02 -2.32799158e-01
4.11344856e-01 6.33137822e-02 -8.18908095e-01 1.06246583e-02
-6.99814558e-02 9.88482311e-03 2.95674384e-01 -9.20352280e-01
-5.41505516e-01 -4.03090268e-01 -1.17326885e-01 1.32996395e-01
-9.30373728e-01 2.78375685e-01 -2.10042804e-01 -6.53240740e-01
-1.21501006e-01 -5.64455986e-01 1.26810864e-01 -6.38781190e-01
-1.99627653e-01 -3.89518529e-01 3.09899062e-01 -1.17429423e+00
1.63891304e+00 -2.11373854e+00 -1.73386067e-01 5.46540380e-01
-1.26750395e-01 4.56494659e-01 -3.59225124e-02 8.69377553e-01
-1.75630182e-01 1.47974893e-01 2.50911247e-02 -1.58067331e-01
6.66961744e-02 1.02822341e-01 7.46302083e-02 4.95968968e-01
-3.67931336e-01 3.17932904e-01 -6.38142467e-01 -4.72540796e-01
2.91775078e-01 5.73550820e-01 -2.91783929e-01 -1.21115558e-01
3.22285295e-01 -2.18243413e-02 -2.00077608e-01 3.59582573e-01
7.13292837e-01 5.58902204e-01 4.24264610e-01 -7.67327845e-02
-6.35761619e-01 5.13490796e-01 -1.30181766e+00 1.30806899e+00
-6.75011396e-01 6.14704967e-01 -4.79334146e-02 -1.01508117e+00
1.28284764e+00 3.51715893e-01 2.83111244e-01 -5.39047778e-01
5.44368684e-01 4.36257511e-01 9.20879468e-02 -6.64584398e-01
8.99382293e-01 -1.31726608e-01 -1.82975858e-01 3.17036808e-01
2.51511514e-01 2.65668571e-01 3.98337811e-01 -2.09355205e-01
7.32391059e-01 -2.29757547e-01 5.72228849e-01 -8.31711650e-01
1.13263631e+00 3.61002088e-02 2.34548956e-01 6.90073907e-01
-7.57363290e-02 4.30546761e-01 3.40769827e-01 -4.90530014e-01
-7.95582831e-01 -9.00062978e-01 -2.94131219e-01 1.28584754e+00
-1.20789163e-01 -3.64234775e-01 -1.06649721e+00 -4.55128044e-01
-4.63534240e-03 9.70118523e-01 -4.99594867e-01 -7.00569078e-02
-5.60338020e-01 -5.41922331e-01 7.92515099e-01 1.22552793e-02
3.70822072e-01 -1.20402896e+00 -1.01450995e-01 1.82397351e-01
-1.93200737e-01 -8.31421316e-01 -3.18336487e-01 7.73713458e-03
-5.61861753e-01 -9.33422208e-01 -6.64254546e-01 -1.27402997e+00
3.43750745e-01 -2.24689335e-01 9.71897483e-01 1.99872050e-02
4.69898172e-02 2.92115152e-01 -6.43799484e-01 -4.15804327e-01
-8.68879020e-01 4.97765034e-01 2.19683275e-01 2.32601315e-01
1.02924526e+00 -4.07599002e-01 -2.17179693e-02 -8.47244188e-02
-7.91767120e-01 -4.94850099e-01 5.47427118e-01 4.27662849e-01
4.03677896e-02 -8.69238675e-02 5.13776720e-01 -1.03451490e+00
9.30695295e-01 -5.78100085e-01 -3.46608371e-01 1.88432664e-01
-6.61502421e-01 5.80081604e-02 8.95201802e-01 -4.00195569e-01
-9.09292638e-01 3.11296266e-02 -8.78286779e-01 3.31554204e-01
-3.76727700e-01 6.16506875e-01 -1.60114586e-01 3.35071348e-02
6.82074487e-01 5.18108130e-01 -8.09204429e-02 -8.13766360e-01
2.72036821e-01 1.34145784e+00 3.51951659e-01 -4.30110782e-01
6.51101351e-01 1.52743965e-01 -7.21108198e-01 -1.38296390e+00
-1.96462736e-01 -7.44754434e-01 -9.75477099e-01 -2.88175613e-01
9.62970674e-01 -6.48009956e-01 -3.77765030e-01 7.95512259e-01
-1.16997409e+00 5.11314049e-02 -1.35354251e-01 7.59341478e-01
-2.17593089e-01 5.42867661e-01 -6.51133776e-01 -6.66726470e-01
-4.04647052e-01 -7.16310918e-01 4.08692360e-01 5.71428761e-02
-3.73932660e-01 -1.44575584e+00 5.34075975e-01 3.55295464e-02
4.22753006e-01 4.99480888e-02 1.36791861e+00 -1.12328839e+00
4.78685230e-01 -3.39410245e-01 1.63872376e-01 5.30399203e-01
3.03700626e-01 9.21218246e-02 -8.42761099e-01 -1.85773551e-01
1.47902742e-01 1.35395080e-01 6.89225137e-01 2.27493316e-01
4.82157588e-01 -1.80365235e-01 1.41644582e-01 2.75362045e-01
1.61815321e+00 4.49550807e-01 4.91804898e-01 7.38566458e-01
4.99074489e-01 6.69224143e-01 1.88229725e-01 2.94316262e-01
4.68682319e-01 3.43560278e-01 -1.46998718e-01 -9.45326984e-02
-5.16645350e-02 -2.50167012e-01 6.48782611e-01 1.61118722e+00
-3.53217497e-02 -1.76837862e-01 -1.14731479e+00 8.61714602e-01
-1.46142519e+00 -9.94390488e-01 -2.73760617e-01 2.36607814e+00
7.79941142e-01 2.40919217e-01 3.81765574e-01 5.09212255e-01
9.15088713e-01 2.20441669e-01 1.62804201e-01 -1.23065436e+00
-3.46607864e-01 3.94549578e-01 4.72196937e-01 8.36238980e-01
-1.14277411e+00 9.56108510e-01 6.57199144e+00 8.34195495e-01
-1.23774743e+00 5.42940274e-02 2.64910430e-01 3.85322124e-01
-2.75010109e-01 -2.17647329e-01 -9.94246125e-01 5.57800174e-01
1.21147990e+00 -2.51637101e-01 1.65813938e-01 6.98945701e-01
2.80203521e-01 -5.70668988e-02 -6.54322505e-01 9.46649313e-01
4.73223209e-01 -1.19687533e+00 2.88103580e-01 -1.18833408e-01
3.16457003e-01 1.49820238e-01 -2.19212279e-01 4.29252982e-01
1.34333149e-01 -8.16745639e-01 9.08418059e-01 4.09801185e-01
4.67948586e-01 -8.56093049e-01 1.10836709e+00 3.47055644e-01
-1.06280077e+00 1.06256321e-01 -3.86461914e-01 -2.98922837e-01
-1.44349203e-01 3.82620066e-01 -4.31485921e-01 6.57012939e-01
4.43953454e-01 5.04101217e-01 -7.80371726e-01 8.05140078e-01
-3.06334533e-02 8.88568819e-01 -2.45507613e-01 -6.92216635e-01
4.71996993e-01 -6.02898479e-01 4.63385433e-01 1.75719965e+00
3.53675246e-01 -4.94005412e-01 6.44409135e-02 3.65373403e-01
3.54692400e-01 1.09156060e+00 -8.66405189e-01 1.14972703e-01
4.50128227e-01 1.14884245e+00 -5.44482589e-01 -3.73444051e-01
-5.77822149e-01 7.72472262e-01 1.45380467e-01 1.41846284e-01
-4.65582788e-01 -8.42131674e-01 4.63420331e-01 1.23494096e-01
1.79974854e-01 -4.05138463e-01 -4.01409209e-01 -9.59186792e-01
-5.15836291e-02 -8.97917807e-01 5.26748240e-01 -2.15216056e-01
-1.38947940e+00 9.34298217e-01 -1.00742504e-01 -9.67313766e-01
-4.54206616e-01 -1.04743934e+00 -5.28095484e-01 1.03928173e+00
-1.26294696e+00 -9.52936232e-01 1.22704968e-01 5.83757341e-01
3.56985033e-01 -7.27914214e-01 1.08438551e+00 6.71158433e-01
-4.43180144e-01 6.30339563e-01 6.77034795e-01 6.46397948e-01
6.09444916e-01 -1.26965988e+00 8.09774399e-02 6.87138021e-01
1.61856905e-01 6.59434617e-01 6.48380339e-01 -3.33570927e-01
-9.36616778e-01 -7.94638395e-01 1.88108110e+00 -2.20209911e-01
7.55700707e-01 -5.66511571e-01 -6.99470520e-01 3.96656483e-01
5.66394031e-01 -7.17991292e-01 9.89842296e-01 3.05101275e-01
-3.49103719e-01 -8.85325968e-02 -1.17981851e+00 6.05344892e-01
5.25984764e-01 -7.22523630e-01 -1.08176863e+00 2.75637895e-01
2.81179965e-01 1.95058852e-01 -8.34487796e-01 -2.10387409e-01
6.26736522e-01 -1.10274768e+00 7.82603562e-01 -5.02253890e-01
2.26658493e-01 -5.09800911e-02 -3.04753035e-01 -1.29089665e+00
-3.25207740e-01 -2.33909383e-01 5.53346872e-01 1.58386588e+00
5.25650382e-01 -9.49433565e-01 3.95362824e-01 1.12007983e-01
1.06624827e-01 -2.13478953e-01 -9.22642469e-01 -1.00159240e+00
7.76166022e-01 -3.17398697e-01 4.19936240e-01 1.21530354e+00
1.64631575e-01 5.74430287e-01 8.46109446e-03 -2.29342490e-01
1.06311560e-01 -2.24413544e-01 4.93587315e-01 -1.34358346e+00
-1.28357708e-01 -8.97491813e-01 -6.55020356e-01 -5.99595010e-01
3.74015749e-01 -1.04705381e+00 -2.30167761e-01 -1.40994084e+00
-1.92524835e-01 -1.62364513e-01 -3.25960696e-01 1.57824308e-01
1.34687647e-01 1.12190560e-01 5.51247559e-02 -1.06077120e-01
1.41391203e-01 3.24421585e-01 3.81079316e-01 -2.03005582e-01
-3.42453510e-01 -1.74325764e-01 -5.22850871e-01 9.26013887e-01
1.00832045e+00 -4.33676541e-01 -1.76197231e-01 -3.81321073e-01
-2.74146162e-02 -6.05686843e-01 -2.71440148e-01 -1.23777437e+00
8.52257460e-02 1.48645341e-01 -2.21613273e-02 -5.59206247e-01
6.46817237e-02 -6.23335600e-01 -2.26934344e-01 6.42946661e-01
-2.85323679e-01 5.42027056e-01 2.13155493e-01 -1.88771337e-02
-4.43547040e-01 -8.09425592e-01 9.69175041e-01 1.83962099e-02
-8.25882494e-01 -1.93006143e-01 -1.06771803e+00 5.19651175e-02
7.59817004e-01 -3.64038467e-01 3.36184576e-02 -3.16834509e-01
-5.36440790e-01 -4.78248566e-01 2.30563030e-01 7.04547882e-01
2.04204634e-01 -1.39915872e+00 -8.91754985e-01 3.23752701e-01
1.85309500e-01 -9.99458015e-01 6.37046993e-02 5.64004958e-01
-9.37470973e-01 5.14477491e-01 -4.17738587e-01 -1.31061137e-01
-1.36678898e+00 5.50840139e-01 3.64567280e-01 -4.27085698e-01
-2.74174571e-01 3.02947015e-01 -2.77628779e-01 -1.07886422e+00
1.27770692e-01 -1.94223225e-01 -1.08019638e+00 4.37244356e-01
5.34153402e-01 3.26389015e-01 2.96616614e-01 -1.32146561e+00
-4.40009266e-01 9.11831498e-01 -7.63639808e-03 -3.33028644e-01
1.05777478e+00 -6.35318831e-02 -4.64572281e-01 9.62581515e-01
1.36761355e+00 5.07799506e-01 9.22796652e-02 -2.93293655e-01
3.41102451e-01 -5.36582544e-02 2.15723783e-01 -6.91897154e-01
-5.56151271e-01 8.93085182e-01 7.45811999e-01 4.69376296e-01
8.48991275e-01 -5.84006250e-01 6.01219952e-01 3.68743271e-01
1.91809312e-01 -1.46169794e+00 -6.76403105e-01 1.07918680e+00
5.18768430e-01 -7.81669319e-01 -3.95090610e-01 -1.21508487e-01
-4.02791798e-01 1.43037999e+00 9.50389728e-02 -2.21684903e-01
1.11056030e+00 3.05084512e-02 5.23818314e-01 1.11551151e-01
-1.40769929e-01 -3.66386563e-01 2.99751639e-01 5.63710093e-01
8.09911370e-01 7.14112595e-02 -1.13267899e+00 7.62187362e-01
-5.34516096e-01 -5.30420914e-02 5.19705832e-01 7.44185686e-01
-5.38645566e-01 -1.30102479e+00 -5.22529602e-01 2.61263013e-01
-5.67722440e-01 -3.35702002e-01 -5.84114611e-01 1.16835511e+00
2.96218902e-01 1.24058676e+00 1.47090301e-01 -6.13872170e-01
3.93690854e-01 4.50654656e-01 2.53213376e-01 -4.21347797e-01
-1.05660391e+00 -3.12732697e-01 3.94972056e-01 9.58957821e-02
-3.27701420e-01 -7.25995362e-01 -9.68825459e-01 -6.79882944e-01
1.19292755e-02 4.95197296e-01 9.19703126e-01 1.09908569e+00
-8.43587667e-02 9.47704837e-02 7.31754899e-01 -4.51338500e-01
-4.29607451e-01 -1.19580448e+00 -7.77728915e-01 4.80466843e-01
1.13801561e-01 -2.56336361e-01 -4.34757859e-01 -2.97300722e-02] | [10.262991905212402, 10.354966163635254] |
2723adbf-62ae-43c0-9506-8a4e75df245f | learning-prompt-enhanced-context-features-for | 2306.14451 | null | https://arxiv.org/abs/2306.14451v1 | https://arxiv.org/pdf/2306.14451v1.pdf | Learning Prompt-Enhanced Context Features for Weakly-Supervised Video Anomaly Detection | Video anomaly detection under weak supervision is challenging due to the absence of frame-level annotations during the training phase. Previous work has employed graph convolution networks or self-attention mechanisms to model temporal relations, along with multiple instance learning (MIL)-based classification loss to learn discriminative features. However, most of them utilize multi-branches to capture local and global dependencies separately, leading to increased parameters and computational cost. Furthermore, the binarized constraint of the MIL-based loss only ensures coarse-grained interclass separability, ignoring fine-grained discriminability within anomalous classes. In this paper, we propose a weakly supervised anomaly detection framework that emphasizes efficient context modeling and enhanced semantic discriminability. To this end, we first construct a temporal context aggregation (TCA) module that captures complete contextual information by reusing similarity matrix and adaptive fusion. Additionally, we propose a prompt-enhanced learning (PEL) module that incorporates semantic priors into the model by utilizing knowledge-based prompts, aiming at enhancing the discriminative capacity of context features while ensuring separability between anomaly sub-classes. Furthermore, we introduce a score smoothing (SS) module in the testing phase to suppress individual bias and reduce false alarms. Extensive experiments demonstrate the effectiveness of various components of our method, which achieves competitive performance with fewer parameters and computational effort on three challenging benchmarks: the UCF-crime, XD-violence, and ShanghaiTech datasets. The detection accuracy of some anomaly sub-classes is also improved with a great margin. | ['Shengjin Wang', 'Xiaoyu Wu', 'Yujiang Pu'] | 2023-06-26 | null | null | null | null | ['video-anomaly-detection', 'anomaly-detection-in-surveillance-videos', 'anomaly-detection', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'computer-vision', 'methodology', 'methodology'] | [ 1.84840262e-01 -3.06810707e-01 -1.97027057e-01 -6.12369716e-01
-4.61323947e-01 -1.45161003e-01 5.34508884e-01 4.91324216e-01
-5.34347415e-01 3.32439929e-01 -4.87976000e-02 -9.72090960e-02
-2.26626173e-01 -7.08821416e-01 -4.55078632e-01 -8.03612888e-01
-2.51299232e-01 -1.03155568e-01 5.85135281e-01 -6.27754405e-02
8.43468904e-02 4.34737742e-01 -1.54011083e+00 4.11157012e-01
1.39275718e+00 1.42828071e+00 -1.20940439e-01 3.16763282e-01
-4.71103787e-02 9.67348456e-01 -4.14885968e-01 -3.31027776e-01
1.71266764e-01 -3.74994755e-01 -4.45125937e-01 3.36715966e-01
5.10352492e-01 -5.06858706e-01 -6.12268269e-01 1.08182621e+00
2.00043276e-01 2.97272384e-01 3.80291760e-01 -1.44414687e+00
-3.62086177e-01 1.30783662e-01 -7.12420344e-01 6.74082875e-01
1.24315977e-01 4.26113278e-01 1.15396929e+00 -8.60418916e-01
2.63213124e-02 1.02263677e+00 4.99715030e-01 2.35687003e-01
-1.08132613e+00 -6.05158091e-01 8.56514096e-01 8.03048134e-01
-1.37529218e+00 -2.32512161e-01 1.00009751e+00 -2.18451798e-01
8.86258364e-01 1.87889665e-01 6.47234678e-01 1.24910522e+00
3.13472860e-02 1.01796722e+00 6.48515403e-01 -1.66891068e-01
3.00116211e-01 -2.07096443e-01 3.61990154e-01 1.03288639e+00
1.51980668e-01 -9.70507935e-02 -5.05518019e-01 -2.10733980e-01
6.57058597e-01 4.97135878e-01 -2.49999389e-01 -3.19560051e-01
-8.69100332e-01 7.16027319e-01 3.91012371e-01 2.48672694e-01
-3.98304433e-01 -1.21177010e-01 6.73746228e-01 2.71721840e-01
6.73071265e-01 -1.01281457e-01 -3.38096857e-01 -2.86494335e-03
-7.63810933e-01 9.34420750e-02 3.07179570e-01 9.16525960e-01
7.57124960e-01 2.09929958e-01 -5.93735933e-01 9.10384953e-01
1.69322580e-01 4.47298251e-02 4.80289519e-01 -6.19344711e-01
6.39487624e-01 9.29495215e-01 -2.65484273e-01 -1.30236304e+00
-4.61382657e-01 -7.02174366e-01 -8.92736197e-01 -2.30347328e-02
3.26784730e-01 6.77717701e-02 -1.07387245e+00 1.86871171e+00
3.63577217e-01 8.93328428e-01 -2.93351173e-01 9.68095064e-01
3.31988037e-01 4.40491736e-01 3.85638237e-01 -2.43957266e-01
1.25540924e+00 -9.65936482e-01 -6.71313167e-01 -2.65006006e-01
8.37475359e-01 -2.68983722e-01 1.14653361e+00 1.79019362e-01
-7.54609644e-01 -4.97439563e-01 -1.00797272e+00 1.11137874e-01
-1.68537170e-01 7.62858316e-02 5.48377275e-01 3.47134501e-01
-5.79816580e-01 5.59713542e-01 -1.19025278e+00 -9.77413431e-02
7.64802814e-01 1.56883404e-01 -2.07481921e-01 -1.68806896e-01
-1.17955589e+00 4.65255529e-01 4.72558886e-01 9.43228006e-02
-8.16993713e-01 -7.04118669e-01 -1.12957799e+00 3.14552218e-01
8.05057466e-01 -4.68230784e-01 7.70226717e-01 -1.19117069e+00
-1.18448603e+00 4.10098970e-01 -2.23174587e-01 -5.87794363e-01
4.38263655e-01 -3.93961221e-01 -6.98114574e-01 4.26783055e-01
3.63460667e-02 2.47659549e-01 9.92480755e-01 -9.17226553e-01
-9.61914897e-01 -4.93369043e-01 2.97446609e-01 1.98283419e-01
-6.96290791e-01 -1.07807390e-01 -6.96452856e-01 -1.23992670e+00
2.45450616e-01 -6.75878882e-01 -2.62224317e-01 -5.28481565e-02
-2.53262281e-01 -3.54879498e-01 1.09452999e+00 -7.41187215e-01
1.64661658e+00 -2.39756441e+00 -1.15493382e-03 3.96217406e-01
2.88768560e-01 4.29805309e-01 -1.97860613e-01 1.92540064e-02
-1.02090709e-01 -2.21996203e-01 -5.75815022e-01 -4.74114686e-01
-1.69690907e-01 4.72022593e-01 -1.78944141e-01 4.39212948e-01
7.60544002e-01 6.38677299e-01 -9.17037964e-01 -4.83256131e-01
2.73221433e-01 3.38947088e-01 -8.27112436e-01 2.09819004e-01
-2.12283641e-01 5.20103991e-01 -6.11910820e-01 8.42025161e-01
5.30536473e-01 -2.51023412e-01 -8.71599093e-02 -1.77534297e-01
2.08339736e-01 1.72921747e-01 -1.25972366e+00 1.81368530e+00
-3.78857255e-01 1.30067766e-01 -5.98930977e-02 -1.38234699e+00
5.73226810e-01 1.77043095e-01 5.13235092e-01 -8.07905078e-01
-6.85132444e-02 1.30404934e-01 -9.26894546e-02 -5.91269016e-01
1.97039112e-01 3.52542669e-01 1.33151069e-01 4.58382331e-02
9.37828422e-02 7.30066836e-01 2.52531976e-01 2.79477239e-01
1.30503821e+00 1.72989607e-01 8.88155252e-02 -1.04110219e-01
9.62906897e-01 -3.53709042e-01 1.09435821e+00 6.35497391e-01
-4.55068141e-01 5.43929338e-01 5.90203047e-01 -5.37709534e-01
-5.38797200e-01 -9.10564184e-01 5.39327879e-03 1.09485745e+00
4.55644846e-01 -6.18690610e-01 -5.62963784e-01 -1.31564450e+00
-9.11194012e-02 7.22131133e-01 -5.91484249e-01 -4.77038056e-01
-8.10658455e-01 -1.04826725e+00 2.80568838e-01 7.29374707e-01
6.33512139e-01 -8.29552114e-01 -3.68859708e-01 1.65713891e-01
-2.74858534e-01 -1.27602422e+00 -6.13232553e-01 9.70489904e-02
-8.12613904e-01 -1.14373791e+00 -3.41406435e-01 -5.13400078e-01
7.63593674e-01 2.32616067e-01 7.55005479e-01 2.75353789e-01
-2.73293108e-01 3.99533093e-01 -5.06809354e-01 -3.84898074e-02
1.41654238e-01 -1.00167707e-01 -3.19002718e-02 5.76429367e-01
5.03088474e-01 -7.48299420e-01 -7.41412938e-01 4.27852988e-01
-1.03496373e+00 -2.33012214e-01 6.45184994e-01 1.00638103e+00
6.55972540e-01 3.75852063e-02 6.28639162e-01 -6.79827690e-01
1.62211642e-01 -7.00414181e-01 -5.56979358e-01 1.37943208e-01
-6.47968829e-01 -5.27563617e-02 8.30868959e-01 -3.59158993e-01
-1.19251239e+00 -1.60819069e-01 1.94954760e-02 -7.27586925e-01
-2.96932340e-01 3.62584442e-01 -4.47679400e-01 9.85986739e-02
3.64290774e-01 4.14413899e-01 -1.07671589e-01 -3.14816833e-01
-1.14170266e-02 2.29370907e-01 5.16747475e-01 -6.33729458e-01
7.68132985e-01 6.08336151e-01 2.56917309e-02 -6.23133063e-01
-1.08757865e+00 -6.02097750e-01 -5.05475163e-01 -3.35535668e-02
8.46258879e-01 -1.05870664e+00 -4.02372837e-01 5.76957047e-01
-7.64139533e-01 -4.89975475e-02 -3.39897573e-01 4.70095813e-01
-2.94782788e-01 7.67112017e-01 -6.06040061e-01 -6.03286445e-01
-1.40908912e-01 -1.15462065e+00 1.12137532e+00 2.67084092e-01
7.58851245e-02 -8.75443816e-01 -3.25151384e-01 2.99286664e-01
2.53850579e-01 3.48086894e-01 8.10695708e-01 -9.33848441e-01
-6.34487092e-01 -2.59817719e-01 -3.91090363e-01 6.32242441e-01
1.97609633e-01 -2.63158232e-01 -9.52883601e-01 -3.00713509e-01
-1.21776782e-01 -5.76272346e-02 1.21115088e+00 2.39480793e-01
1.78349853e+00 -2.75313467e-01 -3.06464344e-01 6.49889112e-01
1.12808263e+00 2.61105031e-01 4.55970347e-01 3.35322499e-01
1.07256448e+00 4.73319709e-01 7.75917649e-01 6.99701250e-01
5.08612156e-01 6.93162978e-01 5.06570756e-01 3.93644953e-03
3.41291577e-02 4.47193161e-02 4.05808419e-01 4.33816731e-01
-4.08231318e-02 -2.02721953e-01 -7.60743499e-01 6.03048086e-01
-2.25384521e+00 -1.01014507e+00 -6.80456171e-04 2.13775611e+00
5.62703371e-01 3.49052370e-01 1.95058450e-01 2.96532452e-01
6.24748290e-01 3.06155384e-01 -6.00754917e-01 2.80980971e-02
-1.54270515e-01 6.66327104e-02 1.68004915e-01 3.00273031e-01
-1.45044410e+00 7.86376536e-01 3.97387528e+00 1.12127090e+00
-8.60330224e-01 2.32986640e-02 7.63426363e-01 -2.49340847e-01
1.53529262e-02 5.97456144e-03 -5.66677451e-01 6.89453125e-01
6.84710085e-01 1.62642494e-01 1.11126579e-01 7.96576738e-01
3.17757726e-01 -7.51915202e-03 -1.09981441e+00 7.98275471e-01
1.11936048e-01 -9.84844565e-01 2.10141838e-01 -2.49979347e-01
4.67350096e-01 -1.26014575e-01 -7.48948231e-02 4.23733532e-01
-1.51644900e-01 -5.35145164e-01 6.46316409e-01 4.55656320e-01
3.48462880e-01 -1.01824689e+00 7.45729268e-01 1.37787521e-01
-1.37459922e+00 -4.82938588e-01 -1.29399389e-01 1.65257454e-02
4.05627452e-02 7.96477973e-01 -3.42701405e-01 9.04177666e-01
8.18078220e-01 9.63254392e-01 -6.63280070e-01 1.00686944e+00
-2.32551754e-01 7.43000805e-01 -3.14227968e-01 4.41944391e-01
4.81313616e-01 -2.30887219e-01 8.36392641e-01 1.21126401e+00
8.48404318e-02 2.20345512e-01 6.60189569e-01 4.72835481e-01
8.75551626e-02 1.05744295e-01 -1.94583371e-01 3.14108670e-01
3.39701295e-01 1.25399888e+00 -5.44918478e-01 -3.74751836e-01
-8.34313273e-01 1.03128576e+00 2.56907284e-01 5.14922917e-01
-1.09994483e+00 -4.75215197e-01 9.08184409e-01 6.63634613e-02
3.85896564e-01 -1.60851717e-01 -2.70126224e-01 -1.47923315e+00
3.84635001e-01 -7.93757498e-01 9.65386331e-01 -5.18967248e-02
-1.30507755e+00 4.86018270e-01 2.14813128e-02 -1.39376664e+00
-3.16134212e-03 -3.40151727e-01 -8.25455546e-01 6.23602808e-01
-1.76360679e+00 -1.26750624e+00 -5.08732200e-01 8.88626933e-01
7.03946650e-01 -3.04067612e-01 3.83296341e-01 6.69554174e-01
-1.18489528e+00 8.81960034e-01 -3.68439525e-01 2.91053921e-01
6.87315583e-01 -1.13894045e+00 1.26842670e-02 1.30659294e+00
-2.75023077e-02 3.39919597e-01 3.55704904e-01 -5.99136949e-01
-8.42197180e-01 -1.62637985e+00 5.67419827e-01 -2.04609022e-01
7.48368442e-01 -1.35777056e-01 -1.46675110e+00 7.09620535e-01
-2.15096280e-01 5.89935005e-01 5.84662497e-01 1.23177633e-01
-5.23198128e-01 -3.04701060e-01 -9.97846067e-01 6.21031106e-01
1.31952047e+00 -4.91257846e-01 -4.37149853e-01 2.20317438e-01
7.12694407e-01 -2.94157714e-01 -6.44313753e-01 8.15394938e-01
1.63775533e-01 -1.07113338e+00 8.05275381e-01 -6.22415423e-01
3.44576925e-01 -5.73033273e-01 -1.12762474e-01 -1.05470753e+00
-4.10964191e-01 -2.70203412e-01 -6.52793467e-01 1.38186121e+00
1.64024144e-01 -7.01335907e-01 7.31588721e-01 5.30201077e-01
-4.97874141e-01 -9.98456419e-01 -1.00596821e+00 -9.30212796e-01
-5.49531221e-01 -7.05319285e-01 5.42857230e-01 1.09201396e+00
-2.21886218e-01 5.06306551e-02 -3.43508035e-01 6.93091035e-01
6.04734421e-01 3.38869654e-02 3.57965261e-01 -1.08406329e+00
-3.71190637e-01 -4.39966440e-01 -7.46584475e-01 -7.40071595e-01
2.68107653e-01 -8.06909919e-01 -3.48987073e-01 -9.19206619e-01
5.98918200e-02 -3.71081054e-01 -8.77720773e-01 7.97576368e-01
-6.79917812e-01 2.15029269e-02 -1.48140937e-01 1.16616212e-01
-1.03299236e+00 8.76898825e-01 8.00171018e-01 -7.34234080e-02
-2.37502038e-01 4.08170931e-03 -4.25470740e-01 9.55065131e-01
7.04519272e-01 -4.12873536e-01 -5.90839207e-01 -4.55911458e-01
-2.83546448e-01 -3.16892534e-01 6.59926951e-01 -1.12591422e+00
1.65791839e-01 -1.18954420e-01 4.50645506e-01 -4.36634511e-01
1.07905224e-01 -8.34661961e-01 -4.77977306e-01 2.74117380e-01
-2.80303419e-01 1.78327563e-03 2.20026150e-01 1.12724221e+00
-4.93010253e-01 1.25229642e-01 8.70411873e-01 2.81675428e-01
-1.03111815e+00 8.08590353e-01 -2.16146391e-02 9.10105929e-02
1.17070007e+00 -9.52536613e-02 -9.28946435e-02 -2.12826431e-02
-7.75872290e-01 6.25492871e-01 2.84337997e-01 5.72811604e-01
6.42066061e-01 -1.49315178e+00 -5.99827588e-01 4.35881644e-01
5.23612380e-01 1.80024356e-01 6.58821523e-01 1.13843429e+00
-1.46634459e-01 7.67754018e-02 -2.06522061e-03 -7.99731195e-01
-1.13107479e+00 5.05850971e-01 2.55598456e-01 -4.49019969e-01
-8.55929792e-01 7.89902151e-01 4.25470799e-01 7.19560236e-02
3.99007320e-01 -2.68117785e-01 -2.78093606e-01 -2.65519624e-03
6.40046299e-01 4.67583627e-01 1.93945646e-01 -5.46409905e-01
-5.40053308e-01 2.79871821e-01 -3.97337347e-01 4.13743138e-01
1.20230651e+00 -2.09083915e-01 1.37864381e-01 1.43902823e-01
9.89217937e-01 -1.69246390e-01 -1.59957922e+00 -6.78121746e-01
2.94997215e-01 -5.83915114e-01 1.38040408e-01 -5.26900411e-01
-1.28387308e+00 7.27701664e-01 6.41874611e-01 1.03529595e-01
1.63465273e+00 -2.12184146e-01 9.35498238e-01 2.19259188e-01
-1.76597219e-02 -1.14552128e+00 3.63054484e-01 4.17442828e-01
5.83411634e-01 -1.40131283e+00 -3.15707564e-01 -5.04593909e-01
-6.39051795e-01 9.36972022e-01 1.09527349e+00 -1.67891413e-01
5.06852329e-01 1.51707456e-02 -2.94066668e-01 -1.17937595e-01
-7.52659082e-01 -1.95930257e-01 5.39549887e-01 2.50545859e-01
1.74533129e-01 -2.21932575e-01 -2.56819963e-01 8.85283232e-01
5.74464202e-01 -2.95104593e-01 1.23797685e-01 1.05280864e+00
-3.99670601e-01 -8.61249804e-01 -8.61253496e-03 5.51077962e-01
-6.72952175e-01 -1.52047798e-01 1.36220887e-01 4.73896444e-01
3.64003301e-01 9.49006736e-01 2.04105735e-01 -4.13809925e-01
4.27992612e-01 1.71814963e-01 1.29617736e-01 -4.19085264e-01
-3.67136270e-01 1.26718089e-01 4.23849821e-02 -1.05914235e+00
-2.90012240e-01 -6.66323960e-01 -1.31323016e+00 4.35573943e-02
-2.62216687e-01 5.96929491e-02 3.70968170e-02 1.18939912e+00
6.36133134e-01 7.53415108e-01 7.60053813e-01 -5.02287149e-01
-4.47010487e-01 -6.72032893e-01 -4.66614336e-01 5.98228037e-01
5.47403455e-01 -8.30017567e-01 -4.39046741e-01 -9.22058076e-02] | [7.84399938583374, 1.6116762161254883] |
2ee55854-3c03-4396-a6bf-c42f95dd02e5 | scene-text-recognition-with-image-text | 2305.04524 | null | https://arxiv.org/abs/2305.04524v1 | https://arxiv.org/pdf/2305.04524v1.pdf | Scene Text Recognition with Image-Text Matching-guided Dictionary | Employing a dictionary can efficiently rectify the deviation between the visual prediction and the ground truth in scene text recognition methods. However, the independence of the dictionary on the visual features may lead to incorrect rectification of accurate visual predictions. In this paper, we propose a new dictionary language model leveraging the Scene Image-Text Matching(SITM) network, which avoids the drawbacks of the explicit dictionary language model: 1) the independence of the visual features; 2) noisy choice in candidates etc. The SITM network accomplishes this by using Image-Text Contrastive (ITC) Learning to match an image with its corresponding text among candidates in the inference stage. ITC is widely used in vision-language learning to pull the positive image-text pair closer in feature space. Inspired by ITC, the SITM network combines the visual features and the text features of all candidates to identify the candidate with the minimum distance in the feature space. Our lexicon method achieves better results(93.8\% accuracy) than the ordinary method results(92.1\% accuracy) on six mainstream benchmarks. Additionally, we integrate our method with ABINet and establish new state-of-the-art results on several benchmarks. | ['Umapada Pal', 'Yue Lu', 'Xiao Tu', 'Hongjian Zhan', 'Jiajun Wei'] | 2023-05-08 | null | null | null | null | ['scene-text-recognition', 'text-matching'] | ['computer-vision', 'natural-language-processing'] | [ 4.85527635e-01 -4.56903219e-01 -2.90935844e-01 -4.28072214e-01
-3.43310416e-01 -1.25492245e-01 7.98946619e-01 4.20165733e-02
-5.99245965e-01 9.34013128e-02 3.16512913e-01 -8.00823271e-02
1.68127269e-01 -5.79279602e-01 -6.41602159e-01 -5.23691177e-01
7.95981169e-01 4.26329553e-01 3.56853008e-01 -2.42278770e-01
6.62769735e-01 2.19736055e-01 -1.64075589e+00 7.25235701e-01
6.23262107e-01 1.11217940e+00 5.79830885e-01 3.28510076e-01
-6.52662575e-01 9.48544860e-01 -3.36770475e-01 -4.70055819e-01
4.41967309e-01 -4.79955286e-01 -4.86432463e-01 1.91382393e-01
8.79495442e-01 -1.84844524e-01 -5.21588445e-01 1.38362515e+00
3.91637176e-01 4.58884425e-02 8.73597443e-01 -1.21331823e+00
-1.07148707e+00 4.54453439e-01 -8.65785718e-01 1.21170737e-01
3.29119205e-01 -9.42618996e-02 1.05002332e+00 -1.56158674e+00
6.25373244e-01 1.27938223e+00 5.56016684e-01 3.72485459e-01
-9.09794748e-01 -7.03531682e-01 2.57549435e-01 4.48245913e-01
-1.89322233e+00 -4.72413599e-01 7.34358013e-01 -5.31720936e-01
1.15170312e+00 2.22869068e-01 5.75168371e-01 7.64703989e-01
4.00897235e-01 9.12075400e-01 9.17314827e-01 -6.73043549e-01
-2.86726058e-01 3.27622682e-01 3.96376662e-02 9.50017333e-01
1.85256958e-01 1.16409101e-01 -7.81314254e-01 3.58563066e-01
5.88666618e-01 2.94305891e-01 -2.11530820e-01 -2.15235606e-01
-1.16459548e+00 7.83866823e-01 5.27861416e-01 3.24377775e-01
-5.90368807e-02 1.20239690e-01 2.40441859e-01 3.28856051e-01
2.02924460e-02 2.14983389e-01 -4.69946899e-02 2.83833981e-01
-1.04256082e+00 -1.69584751e-01 4.01657522e-01 9.38440025e-01
7.29060292e-01 2.71273404e-01 -3.90239745e-01 9.35265899e-01
6.44669771e-01 8.37930799e-01 7.12157190e-01 -3.27663720e-01
6.88446522e-01 8.84863734e-01 -3.14890236e-01 -1.60293305e+00
-3.04282486e-01 -2.47249231e-01 -1.09726965e+00 4.62354720e-02
6.33160695e-02 4.24816072e-01 -1.04193413e+00 1.24815178e+00
-1.58056617e-01 4.53706570e-02 1.39132226e-02 1.06134641e+00
1.23379731e+00 6.87348545e-01 -8.00819770e-02 -1.56877413e-01
1.10299921e+00 -1.09619427e+00 -6.80876672e-01 -6.24305844e-01
5.30982852e-01 -1.14540672e+00 1.20854545e+00 2.60412931e-01
-7.10021913e-01 -7.75252879e-01 -1.14571345e+00 -2.23393008e-01
-5.22488832e-01 4.43566412e-01 1.73875853e-01 4.21155304e-01
-9.97295320e-01 2.55063236e-01 -2.49760479e-01 -5.39388955e-01
2.39894718e-01 3.19379508e-01 -3.63596618e-01 -2.01563612e-01
-9.05877948e-01 1.03336048e+00 4.40981984e-01 6.32999912e-02
-6.18058860e-01 -1.93259552e-01 -8.51624131e-01 -6.49138093e-02
2.80648470e-01 -3.68726969e-01 6.12047374e-01 -1.34494615e+00
-1.16592360e+00 1.15778220e+00 -2.93940008e-01 -3.97864133e-01
4.14817870e-01 8.83547496e-03 -5.12956500e-01 -8.61878097e-02
1.90323636e-01 1.15047622e+00 1.24381828e+00 -1.11891007e+00
-6.96660101e-01 -1.47926182e-01 -4.09075111e-01 4.75213468e-01
-2.14915097e-01 -1.04421996e-01 -7.83357441e-01 -8.77572238e-01
4.99523699e-01 -6.76701963e-01 1.57035604e-01 3.03692132e-01
-2.94619352e-01 -3.05929989e-01 8.91134202e-01 -4.20332968e-01
1.17966557e+00 -2.34767318e+00 -5.32596605e-03 1.45524099e-01
1.99393570e-01 3.38429898e-01 -3.68625849e-01 3.24856758e-01
5.70558906e-02 -6.06300980e-02 3.30916084e-02 -4.94447112e-01
-1.20500242e-02 2.61064231e-01 -6.00234985e-01 5.37586629e-01
9.03510116e-03 1.07519722e+00 -5.14207900e-01 -9.36237872e-01
6.26596451e-01 1.83596894e-01 -5.23860395e-01 -8.24190304e-02
-1.66348726e-01 -5.45888171e-02 -2.26560965e-01 6.56608999e-01
5.89116275e-01 -4.39432174e-01 -1.77736171e-02 -4.34833586e-01
-1.23256609e-01 3.06936670e-02 -1.31302941e+00 1.50176799e+00
-8.95868614e-02 9.65125918e-01 -5.56144059e-01 -1.08995807e+00
1.24278247e+00 -4.06594798e-02 2.72870451e-01 -1.22225761e+00
2.83334941e-01 9.79200304e-02 -3.31327580e-02 -4.58114266e-01
7.01640069e-01 4.93144654e-02 6.45768717e-02 6.50079548e-02
-5.70661500e-02 -1.74059346e-01 -9.25036147e-03 1.37786776e-01
3.68719161e-01 -1.33582532e-01 3.64518583e-01 -2.51121372e-01
8.18664312e-01 1.51181445e-01 4.17492479e-01 1.01405919e+00
-1.96661979e-01 6.15508914e-01 -3.01350262e-02 -7.29228556e-01
-8.27800333e-01 -8.61099958e-01 -1.41965002e-01 9.77715850e-01
6.59173369e-01 -6.59240603e-01 -2.30770260e-01 -5.91935635e-01
-8.69021192e-02 5.24563074e-01 -5.65885246e-01 -2.65377671e-01
-2.59953260e-01 -4.27659571e-01 6.03386998e-01 5.37289500e-01
8.78623903e-01 -9.25695777e-01 -2.46471778e-01 -1.47325456e-01
-3.29043299e-01 -1.22348261e+00 -8.89041185e-01 9.29119736e-02
-4.19882983e-01 -9.13494408e-01 -3.00189942e-01 -1.35639966e+00
8.32540631e-01 6.13073289e-01 9.04640019e-01 4.40131217e-01
-1.37597397e-01 2.83044577e-01 -3.24248344e-01 -2.71312982e-01
-3.95906210e-01 -3.18857759e-01 -4.41319421e-02 1.80662632e-01
6.28397048e-01 5.77958040e-02 -4.12180871e-01 4.29086506e-01
-7.26960480e-01 6.90432549e-01 5.38610816e-01 1.04294848e+00
9.59248960e-01 9.09750089e-02 2.05379695e-01 -4.02960122e-01
3.05674285e-01 2.39448976e-02 -7.50557363e-01 4.38030899e-01
-9.85077500e-01 4.06191759e-02 6.44662261e-01 -6.44052923e-01
-6.59625471e-01 3.74431282e-01 1.21166937e-01 -4.15662497e-01
1.24549434e-01 4.11401957e-01 -2.96496693e-02 -3.51880133e-01
5.34977198e-01 7.44160354e-01 -2.28243396e-01 -3.06885809e-01
2.75679976e-01 7.43386328e-01 6.13510668e-01 -1.42660946e-01
7.16764390e-01 4.90584850e-01 -9.77226719e-03 -8.07140708e-01
-8.78791630e-01 -6.51034594e-01 -6.30816042e-01 -2.40099251e-01
9.40960765e-01 -1.03750241e+00 -5.19351184e-01 4.87501174e-01
-1.21075583e+00 -3.83531749e-02 -1.03253603e-01 5.41745484e-01
-2.92433381e-01 7.24721313e-01 -3.45437407e-01 -4.59094137e-01
-4.75455314e-01 -1.25896692e+00 1.17911160e+00 2.18130380e-01
-5.61966561e-03 -7.59352922e-01 -3.61100852e-01 2.69726664e-01
1.50299728e-01 -4.57639009e-01 8.76359403e-01 -6.21744215e-01
-5.54087818e-01 -2.06521377e-01 -6.24247968e-01 3.67059559e-01
-8.23304281e-02 -1.36890233e-01 -6.98206425e-01 -2.40784526e-01
-3.76041085e-02 -1.24993533e-01 1.05122387e+00 3.48905355e-01
9.51335430e-01 -1.62144244e-01 -2.63222843e-01 7.38470912e-01
1.66313612e+00 2.25650758e-01 7.68809855e-01 4.65073496e-01
9.60915685e-01 2.36565500e-01 6.26855433e-01 3.15153420e-01
5.17292619e-01 7.56711662e-01 4.29753959e-01 -1.38541341e-01
-4.25621688e-01 -5.99425793e-01 4.89324600e-01 9.63200808e-01
4.92112428e-01 -3.90376538e-01 -1.01514614e+00 5.10184288e-01
-1.91970420e+00 -8.66310418e-01 -3.95975471e-01 2.04774666e+00
7.33457565e-01 2.19095170e-01 -3.81671399e-01 3.42713892e-02
7.82424390e-01 1.23099595e-01 -4.00080055e-01 -2.84411460e-01
-6.09007835e-01 -6.34661987e-02 6.23775959e-01 5.88120997e-01
-1.19910181e+00 1.35331261e+00 6.13405943e+00 9.64645445e-01
-1.38353074e+00 -2.49419548e-02 3.52591544e-01 2.56185621e-01
6.50033727e-03 7.57551268e-02 -1.08889079e+00 3.44207019e-01
1.77461311e-01 -2.66393032e-02 6.01464510e-01 6.36925936e-01
5.22836251e-03 -1.58633590e-01 -1.04531121e+00 1.63066494e+00
7.24964917e-01 -1.61986780e+00 6.56274438e-01 -2.80621469e-01
8.06600451e-01 1.93772480e-01 3.50559205e-02 2.25037515e-01
-9.59376991e-02 -1.25093341e+00 9.82753098e-01 6.87472343e-01
9.17039156e-01 -2.92903364e-01 7.28469610e-01 5.68526328e-01
-1.37249517e+00 7.47137591e-02 -6.62286878e-01 -9.39877424e-03
-2.93111950e-01 2.57698327e-01 -7.90492237e-01 2.05589175e-01
7.76259542e-01 1.18628764e+00 -9.89883661e-01 8.00307989e-01
-9.55217257e-02 2.54874676e-01 -3.10509890e-01 -1.66471481e-01
2.42048889e-01 -2.20022529e-01 4.43342745e-01 1.24784875e+00
1.81062501e-02 -2.29209691e-01 4.90640789e-01 7.76927888e-01
-1.11683816e-01 5.48237741e-01 -6.95934832e-01 1.76182970e-01
3.73874098e-01 8.81781340e-01 -7.07197011e-01 -5.15480518e-01
-6.61801457e-01 1.19081652e+00 4.78147268e-02 1.28518716e-01
-7.02418029e-01 -2.10972250e-01 5.55848926e-02 -1.22264557e-01
3.38511556e-01 -2.55136508e-02 -5.42149246e-01 -1.11275125e+00
1.03366762e-01 -9.94392991e-01 4.00263250e-01 -1.13004935e+00
-1.26174760e+00 4.99541193e-01 -3.25047761e-01 -1.49673092e+00
2.35609382e-01 -7.03506172e-01 -3.70252818e-01 7.93727994e-01
-1.50719643e+00 -1.25177169e+00 -4.19549614e-01 9.37812090e-01
8.79810810e-01 -4.96188968e-01 5.32182276e-01 3.39270920e-01
-3.95152539e-01 7.16756582e-01 -1.21417288e-02 3.61647695e-01
8.27942967e-01 -8.86396348e-01 3.56031537e-01 9.43013489e-01
6.01744950e-01 2.86522269e-01 3.82276863e-01 -7.58598149e-01
-1.40714669e+00 -1.21886039e+00 1.17469120e+00 -4.18762714e-01
5.21481037e-01 -1.97285637e-01 -8.56453359e-01 5.07585227e-01
9.86832231e-02 -1.74987148e-02 3.23312998e-01 -4.60345387e-01
-6.00576997e-01 -2.08645642e-01 -8.05303156e-01 8.93677831e-01
9.47507143e-01 -6.79186821e-01 -7.27091253e-01 3.27456206e-01
5.58038354e-01 -3.90460223e-01 -3.88198376e-01 3.29584152e-01
5.73804200e-01 -6.47088289e-01 8.77045929e-01 -2.72158235e-01
2.76612371e-01 -6.14259183e-01 -7.83726752e-01 -5.50155818e-01
-4.08461779e-01 -1.42693177e-01 3.10677677e-01 1.05900621e+00
2.83730417e-01 -5.29819369e-01 4.64874715e-01 -2.95136757e-02
-4.39326838e-03 -5.28648913e-01 -1.02937877e+00 -5.89653194e-01
1.95346791e-02 -5.37000835e-01 2.67861277e-01 9.88016605e-01
-1.79650620e-01 6.12190664e-01 -6.52259886e-01 2.27222666e-01
4.68744814e-01 1.94565967e-01 5.94219148e-01 -1.17697001e+00
-2.81537417e-02 -5.99519014e-01 -6.54915571e-01 -1.51687002e+00
1.55840501e-01 -1.33631325e+00 2.01603755e-01 -1.55710530e+00
4.29473639e-01 -1.79484606e-01 -1.32307127e-01 6.07587993e-01
4.14356627e-02 4.15012181e-01 5.56041121e-01 5.25923133e-01
-8.60912085e-01 5.47022343e-01 1.33168769e+00 -6.69718802e-01
-6.25386164e-02 -5.41980684e-01 -5.72675407e-01 8.69942129e-01
6.86708033e-01 -5.72355807e-01 -2.60863006e-01 -7.00446784e-01
3.18121374e-01 -1.41633049e-01 5.23917437e-01 -8.74503016e-01
6.27069473e-01 -3.04571122e-01 5.88276386e-01 -1.03857756e+00
2.28588328e-01 -1.05875254e+00 -4.98140715e-02 5.35494566e-01
-4.25429851e-01 1.99726507e-01 1.30281836e-01 5.57860732e-01
-2.11954713e-01 -3.63482833e-01 8.37650359e-01 1.04797937e-01
-1.18577981e+00 1.33160874e-01 -3.17506820e-01 5.93018644e-02
5.81300318e-01 -5.12893558e-01 -5.30827463e-01 -3.22747767e-01
-2.91553080e-01 2.91433066e-01 5.24303794e-01 8.42165053e-01
1.02917922e+00 -1.35647261e+00 -6.54281199e-01 5.37371278e-01
4.11769509e-01 -3.25423956e-01 -1.08229630e-01 9.11121011e-01
-3.84878576e-01 4.28848803e-01 4.25046831e-02 -1.08306062e+00
-1.44797313e+00 6.83825970e-01 6.27143443e-01 -3.30500454e-02
-7.69349098e-01 5.63922644e-01 4.03282911e-01 -1.44341946e-01
4.99359310e-01 -2.89979845e-01 -1.96601585e-01 -2.43019536e-01
3.43489677e-01 -8.54886174e-02 -2.41329893e-02 -1.01610577e+00
-5.31818271e-01 1.08928084e+00 -1.81548357e-01 1.69435546e-01
8.68635952e-01 -3.32178056e-01 -1.60581976e-01 4.05614972e-01
1.11716425e+00 -5.70241064e-02 -8.19457531e-01 -6.69435620e-01
-5.24728671e-02 -5.77009320e-01 2.92960882e-01 -8.35015893e-01
-1.10614836e+00 8.58392835e-01 9.33423102e-01 -2.36615434e-01
1.20955312e+00 -1.31506413e-01 6.03288114e-01 6.54257178e-01
1.99216511e-02 -1.28500307e+00 3.98267835e-01 7.19187260e-01
9.55859125e-01 -1.60664093e+00 1.22033350e-01 -2.14562774e-01
-8.78112078e-01 1.15005839e+00 6.98637366e-01 -1.17231801e-01
7.34115958e-01 9.37961042e-02 2.52461225e-01 -3.51834178e-01
-6.74611747e-01 -3.97774190e-01 6.92954302e-01 5.34397304e-01
1.72240496e-01 -4.98854220e-02 -2.86157787e-01 4.60449845e-01
2.07626056e-02 -2.76215166e-01 2.59200960e-01 5.55062592e-01
-5.77613890e-01 -8.66708338e-01 -3.91651601e-01 4.42677617e-01
-1.35662466e-01 -5.55049360e-01 -6.17172599e-01 7.51964271e-01
3.22399735e-01 9.91814733e-01 2.75592893e-01 -5.86022615e-01
3.94165128e-01 -2.34336313e-02 3.39203984e-01 -4.43105668e-01
-3.85181785e-01 1.91585854e-01 -4.27712739e-01 -1.67421177e-01
-5.31486452e-01 -3.97202730e-01 -1.42846656e+00 -2.80913204e-01
-5.73444247e-01 -3.37711602e-01 5.25253296e-01 1.07015944e+00
2.15540722e-01 1.91417664e-01 5.51778793e-01 -4.02365983e-01
-2.67951906e-01 -6.61484420e-01 -3.46481085e-01 5.76875746e-01
1.65740266e-01 -5.90935707e-01 -1.60297334e-01 3.84106725e-01] | [11.735645294189453, 2.0475828647613525] |
aa06f9be-671e-4e05-8245-6587c2b58e6f | uniform-hypergraph-partitioning-provable | 1602.06516 | null | http://arxiv.org/abs/1602.06516v4 | http://arxiv.org/pdf/1602.06516v4.pdf | Uniform Hypergraph Partitioning: Provable Tensor Methods and Sampling Techniques | In a series of recent works, we have generalised the consistency results in
the stochastic block model literature to the case of uniform and non-uniform
hypergraphs. The present paper continues the same line of study, where we focus
on partitioning weighted uniform hypergraphs---a problem often encountered in
computer vision. This work is motivated by two issues that arise when a
hypergraph partitioning approach is used to tackle computer vision problems:
(i) The uniform hypergraphs constructed for higher-order learning contain all
edges, but most have negligible weights. Thus, the adjacency tensor is nearly
sparse, and yet, not binary. (ii) A more serious concern is that standard
partitioning algorithms need to compute all edge weights, which is
computationally expensive for hypergraphs. This is usually resolved in practice
by merging the clustering algorithm with a tensor sampling strategy---an
approach that is yet to be analysed rigorously. We build on our earlier work on
partitioning dense unweighted uniform hypergraphs (Ghoshdastidar and Dukkipati,
ICML, 2015), and address the aforementioned issues by proposing provable and
efficient partitioning algorithms. Our analysis justifies the empirical success
of practical sampling techniques. We also complement our theoretical findings
by elaborate empirical comparison of various hypergraph partitioning schemes. | ['Ambedkar Dukkipati', 'Debarghya Ghoshdastidar'] | 2016-02-21 | null | null | null | null | ['hypergraph-partitioning'] | ['graphs'] | [ 3.46468449e-01 3.98692727e-01 -2.91371852e-01 2.15575427e-01
-4.19227451e-01 -6.96076572e-01 2.74918526e-01 2.28131384e-01
-1.17807686e-02 5.44065237e-01 9.41454843e-02 -3.56824279e-01
-6.46799147e-01 -9.63593960e-01 -3.70277107e-01 -1.10729694e+00
-3.05052161e-01 9.50233161e-01 4.36521590e-01 1.76497355e-01
3.15838486e-01 3.77822757e-01 -1.49761105e+00 -7.51298741e-02
6.55167103e-01 5.10253906e-01 -8.66972208e-02 8.70406926e-01
-2.03540064e-02 5.63008547e-01 -6.35107756e-02 -7.29433775e-01
4.87595379e-01 -3.54883373e-01 -1.33147395e+00 7.63656020e-01
5.22596598e-01 1.16689444e-01 -4.58498359e-01 1.34996843e+00
2.06677765e-01 6.24215230e-02 9.87458229e-01 -1.50970387e+00
-5.83520412e-01 1.12364650e+00 -1.30780149e+00 7.00326562e-02
4.84584942e-02 -2.75072336e-01 1.44379473e+00 -4.93591934e-01
5.70071518e-01 1.07192266e+00 6.71779633e-01 1.04124799e-01
-1.65943027e+00 -2.36991614e-01 2.10540950e-01 4.49996322e-01
-1.71845150e+00 -2.60778312e-02 7.24619985e-01 -6.72604203e-01
3.70573640e-01 6.64242506e-01 7.22993374e-01 5.46642303e-01
-2.60184526e-01 7.98513710e-01 1.26029873e+00 -5.67722559e-01
3.07664365e-01 -5.72457016e-02 5.88195920e-01 6.06311262e-01
9.79113042e-01 -2.36850828e-01 2.58350968e-01 -3.86897355e-01
5.95000029e-01 -2.09067076e-01 -5.31959891e-01 -1.00102377e+00
-1.10472929e+00 1.05533791e+00 2.29597971e-01 5.02870560e-01
-3.41958284e-01 8.67979825e-02 4.44633424e-01 2.43740141e-01
4.71311301e-01 7.02936947e-02 -1.50958925e-01 1.24302447e-01
-1.21997571e+00 1.69127852e-01 1.26128578e+00 1.22026110e+00
8.65494370e-01 -2.70846069e-01 7.23582059e-02 6.95480764e-01
1.21639356e-01 1.85036004e-01 -1.80100933e-01 -9.13159072e-01
2.65356362e-01 3.34054947e-01 -1.76652819e-01 -1.50975466e+00
-3.00418198e-01 -2.98621804e-01 -1.25950098e+00 -2.54073054e-01
4.87107545e-01 1.65370449e-01 -6.29130125e-01 1.44975257e+00
5.11170149e-01 1.87645718e-01 -2.65134990e-01 8.31687033e-01
4.97529119e-01 3.75971138e-01 -1.67921543e-01 -5.31129360e-01
1.32825541e+00 -8.58135939e-01 -4.72319782e-01 3.89571965e-01
4.09663856e-01 -7.09986031e-01 7.20721602e-01 5.10931492e-01
-1.18493319e+00 -1.05595335e-01 -8.22567523e-01 1.97452202e-01
-5.05629331e-02 -1.58920288e-01 8.37654293e-01 1.01025867e+00
-1.35816002e+00 3.73337805e-01 -5.59719861e-01 -8.23573947e-01
1.57726854e-01 2.78526783e-01 -1.73336655e-01 -3.43536794e-01
-8.23672533e-01 6.15326583e-01 5.20938754e-01 -1.05253659e-01
-3.43743771e-01 -4.96728539e-01 -6.50233507e-01 1.98890731e-01
6.56639814e-01 -7.59468555e-01 9.34008121e-01 -7.22512960e-01
-1.07823884e+00 9.03728247e-01 1.04575390e-02 -4.05702382e-01
4.81747568e-01 5.52037537e-01 2.17899173e-01 5.18265367e-01
-3.26136425e-02 2.68807411e-01 5.82726479e-01 -1.61425698e+00
-4.38953906e-01 -5.14533639e-01 4.57169443e-01 1.08046509e-01
-2.59215474e-01 -2.64228284e-01 -6.78669214e-01 -5.59047520e-01
2.81471401e-01 -1.33848643e+00 -6.02837920e-01 -6.80122614e-01
-6.91285968e-01 -1.88400835e-01 4.12529349e-01 -9.11595225e-02
1.45262086e+00 -1.83807409e+00 6.12434864e-01 6.53735399e-01
7.17643261e-01 -1.50665805e-01 1.56517923e-01 6.78548574e-01
-2.46253803e-01 1.58633724e-01 -5.61618984e-01 1.92580763e-02
1.86165124e-01 2.28005856e-01 -2.12959483e-01 9.46270704e-01
-4.76543307e-01 7.26384819e-01 -8.80933821e-01 -7.58777976e-01
2.24786997e-01 1.83732465e-01 -6.44367695e-01 -4.03091460e-01
5.29745184e-02 -1.75932664e-02 -1.81077242e-01 4.67841566e-01
1.05933714e+00 -4.57821429e-01 6.03892684e-01 -3.50526184e-01
5.01822829e-02 -1.29327506e-01 -1.68480802e+00 1.11851847e+00
5.63707948e-02 4.42375004e-01 2.64728397e-01 -1.44450593e+00
2.57276654e-01 2.65225589e-01 9.36904073e-01 1.37464017e-01
3.43973815e-01 -1.00415513e-01 1.44509049e-02 -3.30124974e-01
5.91507912e-01 -1.99326128e-01 -3.17864791e-02 6.87492609e-01
-1.93196923e-01 -2.40752265e-01 7.34977007e-01 6.38255894e-01
1.11684179e+00 -4.12262678e-01 3.16414654e-01 -9.46285248e-01
4.42495704e-01 2.65120059e-01 3.09251070e-01 7.77757108e-01
-1.26301795e-01 7.59389818e-01 9.66179788e-01 1.16543770e-02
-1.18717813e+00 -7.93290973e-01 -4.65259910e-01 7.60371685e-01
2.88225949e-01 -5.45503914e-01 -1.10269785e+00 -5.32962322e-01
-3.10223233e-02 1.80607796e-01 -9.89526153e-01 2.30714142e-01
-1.45065129e-01 -1.05045664e+00 2.61575639e-01 3.56598079e-01
-4.67033051e-02 -4.32899594e-01 -4.05700684e-01 1.84285700e-01
-3.25882614e-01 -1.10482740e+00 -4.13056612e-01 1.14109322e-01
-9.54174519e-01 -1.44686580e+00 -9.33065295e-01 -6.79322302e-01
8.78530383e-01 7.63173342e-01 1.31977940e+00 3.25263023e-01
-3.68851870e-01 8.68361950e-01 -5.59978604e-01 -2.26520747e-03
-3.46982032e-02 2.51237303e-01 -8.88175964e-02 1.39555067e-01
4.08313215e-01 -5.88122368e-01 -5.23441553e-01 4.88933563e-01
-1.13183463e+00 7.99363945e-04 4.52254236e-01 7.12929785e-01
5.23603797e-01 6.72850728e-01 -1.88477002e-02 -1.39137006e+00
3.31423551e-01 -6.66458607e-01 -6.51360393e-01 1.74632698e-01
-5.51033974e-01 -1.13143846e-01 3.97788346e-01 -3.18122208e-01
-6.03875458e-01 -1.24906674e-01 2.25722075e-01 -2.77338743e-01
9.58723649e-02 6.17271900e-01 -1.70120686e-01 -2.50169694e-01
2.93581247e-01 -3.73445489e-02 -2.59134352e-01 -8.60271230e-03
6.26390338e-01 5.91276169e-01 1.95264935e-01 -7.24210739e-01
9.36263204e-01 9.48121190e-01 4.29520249e-01 -1.07676339e+00
-5.67848682e-01 -8.71674836e-01 -8.18685532e-01 -1.58097401e-01
6.55605793e-01 -5.13821602e-01 -7.10248172e-01 1.49238303e-01
-7.86653817e-01 -1.11345664e-01 -1.24674812e-01 4.35758561e-01
-6.89699769e-01 1.07913542e+00 -7.27047920e-01 -9.09587204e-01
2.07294673e-01 -1.30868745e+00 7.93497920e-01 -2.32425794e-01
-2.17416584e-01 -1.05094886e+00 2.94098556e-01 3.22953194e-01
1.34964392e-01 4.69367430e-02 1.14035082e+00 -3.82059038e-01
-6.71459079e-01 -1.50508583e-02 -5.50934672e-01 -1.94616348e-01
-2.27625698e-01 2.45144069e-01 -6.48456037e-01 -4.20036942e-01
3.08484659e-02 -1.48583502e-01 9.72927213e-01 7.65737534e-01
1.08039069e+00 -2.23342717e-01 -4.64320600e-01 3.12153637e-01
1.72292912e+00 -3.09840471e-01 6.09411836e-01 1.47290051e-01
8.74157667e-01 9.32740688e-01 3.47290576e-01 4.07341033e-01
7.06164360e-01 4.49984908e-01 3.73492986e-01 -4.48381202e-03
-8.81930813e-04 2.31471360e-01 -3.51221383e-01 1.01675606e+00
-2.77896404e-01 -3.18329155e-01 -7.06982791e-01 1.03045213e+00
-2.07103348e+00 -1.03210211e+00 -9.00815368e-01 2.29073763e+00
8.11395884e-01 -1.36936128e-01 4.56199050e-01 5.83218873e-01
1.03860378e+00 8.15518051e-02 4.52152863e-02 -4.24076498e-01
5.02025038e-02 8.80425498e-02 9.85579133e-01 6.93748832e-01
-1.12756634e+00 6.18422329e-01 6.33396721e+00 9.24045444e-01
-5.07217407e-01 7.15243369e-02 6.58979535e-01 -5.51534295e-02
-6.14730418e-01 3.31346244e-01 -4.89884317e-01 4.16433901e-01
7.70087540e-01 -1.95955932e-01 2.91884154e-01 7.86204815e-01
-2.09273756e-01 -3.64988208e-01 -1.02038300e+00 9.39209104e-01
1.66542828e-01 -9.35249507e-01 -1.48199856e-01 5.77407002e-01
1.00341773e+00 -2.51612246e-01 9.09451395e-02 -1.82372518e-04
7.54180551e-01 -8.31166804e-01 3.35767716e-01 -6.20900691e-02
4.89097357e-01 -9.21716452e-01 6.25451028e-01 7.88439810e-02
-1.43410420e+00 1.73979372e-01 -7.88798511e-01 1.38550937e-01
1.11339919e-01 1.13686788e+00 -3.48384082e-01 8.45743299e-01
6.95423186e-01 1.70153290e-01 -1.83919296e-01 1.36498904e+00
1.53425887e-01 7.98208117e-01 -4.16488796e-01 1.50188118e-01
3.93454194e-01 -6.00460768e-01 5.54841161e-01 1.29956853e+00
-3.53428535e-02 2.16520041e-01 2.33224273e-01 5.27269363e-01
1.90115035e-01 2.38887727e-01 -6.85718179e-01 1.19288392e-01
1.88284472e-01 1.44503713e+00 -1.51858425e+00 -3.33534062e-01
-5.65839767e-01 5.71830630e-01 2.49243081e-01 3.01908255e-01
-8.05835962e-01 -3.48068655e-01 2.13455901e-01 2.18644589e-01
6.93953693e-01 -4.03856754e-01 -4.58400339e-01 -1.23073936e+00
-1.70990914e-01 -7.81894565e-01 6.16318822e-01 -1.23011954e-01
-1.42311323e+00 4.32907760e-01 4.28643048e-01 -9.32982802e-01
-2.47531924e-02 -3.83908063e-01 -1.72907352e-01 4.82804179e-01
-1.24986696e+00 -1.07266510e+00 -9.64579545e-03 4.31418747e-01
2.89784193e-01 4.39715177e-01 3.17767382e-01 4.65253055e-01
-5.30189633e-01 3.37140650e-01 8.24234262e-02 -2.89359912e-02
2.74546415e-01 -1.56754172e+00 2.38932699e-01 9.10388589e-01
3.29286397e-01 6.05662704e-01 9.25349534e-01 -6.31677449e-01
-1.59400809e+00 -9.16537642e-01 8.39308143e-01 -4.52870250e-01
9.53836322e-01 -2.89298356e-01 -9.48817492e-01 7.62484372e-01
4.54208225e-01 -6.99112564e-02 8.70941162e-01 3.23256671e-01
-2.48177439e-01 3.30717355e-01 -1.04197478e+00 6.37425125e-01
1.06885815e+00 -1.51773036e-01 -4.11677301e-01 3.49886119e-01
3.06086004e-01 5.75820580e-02 -8.57254267e-01 3.33488911e-01
4.80333596e-01 -1.06230235e+00 8.96656036e-01 -3.57053101e-01
1.87170163e-01 -3.02933395e-01 -2.68148839e-01 -1.05681515e+00
-6.68204546e-01 -5.35923541e-01 -7.50005618e-02 1.15318108e+00
1.12791777e-01 -4.64377165e-01 9.06676888e-01 5.09954095e-01
3.34332258e-01 -6.67590261e-01 -7.13332891e-01 -7.53701687e-01
9.59125757e-02 -5.01576483e-01 3.52213025e-01 1.12033582e+00
2.48633102e-01 2.27215797e-01 -1.43621981e-01 1.03913575e-01
1.12727356e+00 4.51558828e-01 8.04087877e-01 -1.43392920e+00
-5.37598073e-01 -8.47465873e-01 -6.40569329e-01 -9.57851768e-01
1.61618173e-01 -9.20607984e-01 -9.67894942e-02 -1.58401752e+00
7.74419963e-01 -5.53564131e-01 4.89490628e-02 3.14892828e-02
-8.30374286e-02 5.93123019e-01 8.77598971e-02 3.73087704e-01
-8.28843534e-01 2.02469155e-01 1.05195522e+00 -1.48344070e-01
1.76028669e-01 5.80987036e-02 -9.47714686e-01 6.71295166e-01
3.81215632e-01 -2.36791506e-01 -6.59234464e-01 -2.13385642e-01
3.65569741e-01 -5.41919507e-02 2.06383675e-01 -6.54544055e-01
3.19318712e-01 -1.75911874e-01 -2.59284854e-01 -7.64383674e-01
-1.12845995e-01 -1.15818584e+00 2.25233048e-01 3.74090850e-01
-9.15529728e-02 8.88988525e-02 -1.81585908e-01 7.96389222e-01
2.03117579e-02 -5.57684124e-01 8.98908198e-01 -2.40712211e-01
-2.72073030e-01 5.54623425e-01 -6.08179688e-01 1.22191563e-01
1.37526429e+00 -4.69902277e-01 -9.59967226e-02 -3.63927484e-01
-7.18750358e-01 2.00956032e-01 6.90231562e-01 -1.63641244e-01
2.42128044e-01 -1.19686198e+00 -6.76269591e-01 -9.32309628e-02
5.99931479e-02 2.20816173e-02 4.36487526e-01 1.28572869e+00
-6.58611178e-01 5.15501142e-01 1.33208677e-01 -8.34068894e-01
-1.22872126e+00 1.07955194e+00 -1.43726513e-01 -3.05618376e-01
-7.71780252e-01 7.16543198e-01 4.94897604e-01 -3.77122611e-02
9.50081572e-02 -1.00378864e-01 -5.71820289e-02 3.86492968e-01
2.17471510e-01 7.82667875e-01 1.11933574e-01 -7.41658807e-01
-2.11511180e-01 6.81413174e-01 -1.41691521e-01 -1.40712112e-01
1.19956696e+00 -4.62552071e-01 -6.13949597e-01 3.79414022e-01
1.13599133e+00 2.11034670e-01 -6.34957612e-01 -2.90838122e-01
1.46298721e-01 -6.08418643e-01 5.87656312e-02 1.35515817e-03
-1.30333316e+00 5.32499969e-01 -3.48210987e-03 1.27895463e+00
1.05886018e+00 2.87079811e-01 7.08224177e-01 1.28565161e-02
6.58761621e-01 -9.16215301e-01 -3.45263600e-01 3.37913454e-01
5.17086148e-01 -8.54639530e-01 3.05040598e-01 -8.13926637e-01
-5.65133214e-01 8.61489654e-01 3.27771544e-01 -2.67754108e-01
8.61775398e-01 3.60502601e-01 -4.55311090e-01 -4.41521227e-01
-5.70514202e-01 -7.11511970e-01 2.38188788e-01 5.01584828e-01
2.43560269e-01 2.32214183e-01 -7.67657042e-01 1.17848620e-01
-2.63363570e-01 -3.78307402e-01 8.59207451e-01 7.81630576e-01
-2.88650155e-01 -1.14506400e+00 -5.55521429e-01 6.16978943e-01
-5.15340447e-01 -1.42893255e-01 -4.98906374e-01 7.18993604e-01
-1.96733892e-01 9.73756313e-01 -1.59092844e-02 -2.88145185e-01
-9.49007571e-02 -3.07037115e-01 8.39922488e-01 -6.17599368e-01
-1.85107231e-01 3.04624856e-01 -1.24375641e-01 -1.15404531e-01
-7.47012854e-01 -8.69956970e-01 -6.98939621e-01 -7.10390568e-01
-7.39766657e-01 4.80478466e-01 3.42081428e-01 6.84664130e-01
-7.57792890e-02 3.38140875e-01 6.41894460e-01 -8.42596889e-01
-3.93200815e-01 -7.42016196e-01 -1.11873293e+00 2.27817461e-01
5.69098331e-02 -8.85357022e-01 -6.35502756e-01 9.89836827e-03] | [7.036774158477783, 5.223842144012451] |
8e18aefb-18f2-4827-b85b-b1f2a4521f3b | pangu-coder-program-synthesis-with-function | 2207.11280 | null | https://arxiv.org/abs/2207.11280v1 | https://arxiv.org/pdf/2207.11280v1.pdf | PanGu-Coder: Program Synthesis with Function-Level Language Modeling | We present PanGu-Coder, a pretrained decoder-only language model adopting the PanGu-Alpha architecture for text-to-code generation, i.e. the synthesis of programming language solutions given a natural language problem description. We train PanGu-Coder using a two-stage strategy: the first stage employs Causal Language Modelling (CLM) to pre-train on raw programming language data, while the second stage uses a combination of Causal Language Modelling and Masked Language Modelling (MLM) training objectives that focus on the downstream task of text-to-code generation and train on loosely curated pairs of natural language program definitions and code functions. Finally, we discuss PanGu-Coder-FT, which is fine-tuned on a combination of competitive programming problems and code with continuous integration tests. We evaluate PanGu-Coder with a focus on whether it generates functionally correct programs and demonstrate that it achieves equivalent or better performance than similarly sized models, such as CodeX, while attending a smaller context window and training on less data. | ['Qun Liu', 'Qianxiang Wang', 'Xin Jiang', 'Jiansheng Wei', 'Guangtai Liang', 'Yasheng Wang', 'Ignacio Iacobacci', 'Yuchi Ma', 'Xin Wang', 'Pingyi Zhou', 'Li Yan', 'Hao Yu', 'Lin Li', 'Bo Shen', 'Meng Xiao', 'Qi Zhang', 'Zhongqi Li', 'Yinpeng Guo', 'Guchun Zhang', 'Milan Gritta', 'Gerasimos Lampouras', 'Fenia Christopoulou'] | 2022-07-22 | null | null | null | null | ['program-synthesis', 'text-to-code-generation'] | ['computer-code', 'computer-code'] | [ 3.03948671e-01 4.73602653e-01 -1.11270271e-01 -3.28685939e-01
-9.85682905e-01 -5.50211668e-01 6.71023309e-01 4.86346662e-01
1.19424768e-01 3.19937319e-01 3.84984583e-01 -9.96361554e-01
5.04411638e-01 -9.04174685e-01 -1.12100303e+00 4.97755036e-02
-1.01352252e-01 3.52786869e-01 9.83146206e-02 -2.79229373e-01
4.77262884e-01 -3.82910252e-01 -1.44791186e+00 7.70679355e-01
1.02743173e+00 1.20463684e-01 4.63605136e-01 1.34927487e+00
-2.97719270e-01 1.47795343e+00 -4.53233808e-01 -3.10213178e-01
-2.45944768e-01 -7.45038033e-01 -8.92891169e-01 -4.12255377e-01
1.07026495e-01 1.29346149e-02 6.93302751e-02 8.06865036e-01
1.58055991e-01 -2.32221186e-01 4.76171941e-01 -1.13853157e+00
-1.10477936e+00 1.17684674e+00 -3.37676078e-01 -1.56850532e-01
6.42584980e-01 4.83846635e-01 1.29206145e+00 -9.86130059e-01
5.33483565e-01 1.51485121e+00 6.04080141e-01 1.04487669e+00
-1.99395466e+00 -3.93494368e-01 5.25283180e-02 -5.50296724e-01
-1.15971148e+00 -3.40083718e-01 3.16434056e-01 -9.60575044e-01
1.84220183e+00 -5.41623384e-02 8.57531130e-02 1.04527652e+00
6.22084737e-01 6.52084470e-01 6.95532739e-01 -7.59246051e-01
8.65208358e-02 5.26241995e-02 3.64536703e-01 1.09628868e+00
-1.55948430e-01 4.60418761e-01 -2.96280295e-01 -4.89030987e-01
4.90526646e-01 -5.20522952e-01 -1.11883700e-01 -1.05528452e-01
-1.29161334e+00 1.02297628e+00 3.14114451e-01 2.00481758e-01
7.05953091e-02 1.03263879e+00 4.87408012e-01 5.48321068e-01
2.38933302e-02 8.53774190e-01 -5.67986012e-01 -2.86153167e-01
-8.44786882e-01 6.29362106e-01 8.61905098e-01 1.52794945e+00
5.88600338e-01 3.82490277e-01 -2.48188913e-01 5.47525823e-01
7.68339634e-01 3.42715979e-01 4.66818541e-01 -5.71506679e-01
8.25404227e-01 7.84262121e-01 -3.49187046e-01 -5.63365877e-01
-1.15610845e-01 -1.95771590e-01 -7.13687912e-02 2.81083107e-01
1.37563229e-01 -3.57547522e-01 -8.06667149e-01 2.00233531e+00
-4.78397161e-01 -2.96930611e-01 3.55690986e-01 5.05962849e-01
7.94866145e-01 9.73015308e-01 2.64197648e-01 2.16859877e-01
9.45338428e-01 -1.14812648e+00 -5.56882471e-02 -6.67001367e-01
1.18848681e+00 -6.32522464e-01 1.40914798e+00 2.36332580e-01
-1.30585492e+00 -6.62833095e-01 -1.06124711e+00 -4.02844220e-01
-1.75460754e-03 2.28295460e-01 6.27501667e-01 3.54815483e-01
-1.40183449e+00 3.94051939e-01 -7.09884405e-01 -2.14614168e-01
-3.05930786e-02 5.37346117e-02 -6.04942394e-03 -5.00529222e-02
-8.02532852e-01 6.23073518e-01 5.60629249e-01 -3.13686490e-01
-1.49190617e+00 -9.43724215e-01 -1.37496507e+00 7.67091215e-02
-1.61222573e-02 -7.52552629e-01 1.63960099e+00 -1.03033078e+00
-1.28288150e+00 1.03481066e+00 -2.64550626e-01 -4.75618482e-01
2.79099673e-01 -7.96886533e-02 -1.95644200e-01 -5.39983094e-01
2.82018393e-01 8.24246287e-01 5.97785175e-01 -1.29145432e+00
-4.13799822e-01 2.95475066e-01 1.47837207e-01 -4.19601589e-01
3.50234330e-01 4.19328392e-01 -1.75943717e-01 -7.60665178e-01
-4.57812667e-01 -8.29062760e-01 -3.72690499e-01 -2.73672938e-01
-4.01557148e-01 -3.07047546e-01 3.33440691e-01 -7.25998759e-01
1.56848371e+00 -2.15217209e+00 5.81320345e-01 1.17813326e-01
1.62596330e-01 -1.86593920e-01 -5.07005751e-01 6.78336322e-01
-6.40238702e-01 4.90704358e-01 -7.08771467e-01 -1.71731368e-01
1.39561325e-01 5.43489866e-02 -8.51320267e-01 -5.27807027e-02
8.64450872e-01 1.21844256e+00 -1.16590130e+00 -3.40151638e-01
-2.22673312e-01 -1.48082614e-01 -1.43087900e+00 5.18574357e-01
-1.38587570e+00 2.19477758e-01 -1.61924377e-01 4.67369974e-01
7.22214505e-02 -1.77344561e-01 3.40533644e-01 6.46801889e-01
-3.90422285e-01 6.59178913e-01 -5.46068847e-01 2.14223194e+00
-1.06433511e+00 8.50097656e-01 -3.43503147e-01 -6.54786706e-01
9.82087135e-01 4.17567432e-01 -3.87855858e-01 -5.90192437e-01
-1.18663453e-01 4.34616029e-01 2.62096256e-01 -6.65681899e-01
4.41209584e-01 2.85959821e-02 -6.82127476e-01 8.45892847e-01
2.89406180e-01 -5.52873433e-01 2.06434175e-01 5.49200654e-01
1.48456621e+00 5.10528982e-01 2.36594975e-01 -5.81915259e-01
3.90662581e-01 3.32357317e-01 1.97257638e-01 8.93578649e-01
5.13521016e-01 6.10056639e-01 1.19866633e+00 -2.42993459e-01
-1.26221693e+00 -9.27948773e-01 1.86823592e-01 1.31581700e+00
-3.55412483e-01 -1.14668727e+00 -6.91353142e-01 -6.30688250e-01
-1.71384230e-01 1.62916183e+00 -7.82007396e-01 -3.98826391e-01
-9.19022441e-01 -7.48765051e-01 8.09009612e-01 3.94206107e-01
-1.63913518e-01 -1.30133259e+00 -5.84529817e-01 3.72331709e-01
4.06874865e-02 -4.29745287e-01 -5.30279875e-01 5.29223919e-01
-6.31943583e-01 -9.46385741e-01 -2.40759522e-01 -1.11506128e+00
8.91751647e-01 -2.82327175e-01 1.90926743e+00 6.80215657e-01
-2.61399001e-01 9.93562490e-02 -3.32374632e-01 -2.54859060e-01
-1.46211028e+00 1.93897173e-01 -6.62649512e-01 -6.49660885e-01
1.46845341e-01 -6.22351825e-01 8.25532377e-02 -1.20488264e-01
-8.53736579e-01 6.29626691e-01 4.24795002e-01 9.41791415e-01
-2.36556605e-02 -3.72811586e-01 3.17085683e-01 -9.89801764e-01
9.49804246e-01 -8.39382946e-01 -9.37957346e-01 3.63175720e-01
-5.20105422e-01 6.21054292e-01 7.45232105e-01 -2.73539990e-01
-1.01398945e+00 7.88825229e-02 -2.20359474e-01 6.08390234e-02
2.02903003e-01 9.03610110e-01 2.39505339e-03 3.88885260e-01
1.41339493e+00 4.31503445e-01 -3.73467416e-01 -1.08590961e-01
7.73704231e-01 3.76883179e-01 7.03491211e-01 -1.36017728e+00
8.48707616e-01 -4.59709495e-01 -4.71226692e-01 -2.43224904e-01
-2.22856402e-01 2.01477140e-01 -3.44986171e-01 1.81328490e-01
9.76913810e-01 -1.03818417e+00 -2.69398630e-01 1.93845302e-01
-1.65941203e+00 -1.12313783e+00 -1.02017589e-01 -2.08780184e-01
-6.97949946e-01 -1.93076417e-01 -6.26846194e-01 -6.56678081e-01
-1.42948821e-01 -1.39690387e+00 1.22042310e+00 -1.63973600e-01
-5.15176415e-01 -1.05622053e+00 4.23248798e-01 -1.88000798e-01
4.41480488e-01 2.06768826e-01 1.90174091e+00 -3.71068388e-01
-8.23277771e-01 1.74861833e-01 -2.84424961e-01 1.06044732e-01
-2.79622734e-01 3.34986508e-01 -6.88927531e-01 -8.11563805e-02
-3.19039881e-01 -8.10523629e-01 7.14319408e-01 -6.62834495e-02
9.68387663e-01 -4.27781999e-01 -2.46568352e-01 5.14012694e-01
1.60120571e+00 1.64391890e-01 5.64241409e-01 -1.09380119e-01
6.41460121e-01 4.72335339e-01 8.49414393e-02 1.67182282e-01
5.41917801e-01 5.02507627e-01 5.02719700e-01 1.66709363e-01
-1.51364371e-01 -8.11343610e-01 7.59445846e-01 7.76284099e-01
6.70064926e-01 -1.20025359e-01 -1.47568429e+00 7.97466576e-01
-2.01854467e+00 -6.16732359e-01 -4.05821055e-01 1.79795611e+00
1.33427417e+00 2.18941569e-01 -1.01921633e-01 -3.22491020e-01
3.10127854e-01 -1.26982912e-01 -2.00510949e-01 -9.13514733e-01
2.83630431e-01 4.51700836e-01 1.01881444e-01 7.98817277e-01
-5.61876535e-01 1.21927071e+00 6.62626982e+00 5.24197161e-01
-1.08746433e+00 1.97071865e-01 3.59415293e-01 1.32381320e-01
-9.12496269e-01 4.67906833e-01 -5.24757445e-01 3.51293504e-01
1.42447507e+00 -6.28217459e-01 7.45964527e-01 1.05322862e+00
5.94773889e-02 -1.57943264e-01 -1.81387973e+00 5.90590000e-01
-6.48649633e-02 -1.60479546e+00 1.30986288e-01 -4.04692948e-01
8.54807198e-01 2.45891571e-01 -3.93036216e-01 8.76896501e-01
1.14997804e+00 -1.29696083e+00 1.42175663e+00 4.04851526e-01
7.99141049e-01 -5.13955653e-01 2.05466136e-01 5.37312925e-01
-1.09748971e+00 -2.30916634e-01 -9.57097262e-02 -2.81550318e-01
-4.46051136e-02 4.32887137e-01 -7.77778983e-01 2.74137110e-01
1.72081202e-01 6.90027773e-01 -8.49787056e-01 6.47662818e-01
-6.87069178e-01 6.30701780e-01 3.68112415e-01 -1.32022619e-01
1.46177381e-01 4.14367676e-01 3.60034019e-01 1.69196332e+00
3.09476018e-01 -1.04625687e-01 1.93943903e-01 2.09145665e+00
-6.71411529e-02 -1.72802538e-01 -7.31179476e-01 -3.42523307e-01
1.96292520e-01 8.51592302e-01 -2.48282492e-01 -3.89544696e-01
-6.17294729e-01 7.67481267e-01 4.37951922e-01 1.21238172e-01
-9.47976351e-01 -4.92716759e-01 3.72542053e-01 -7.17102364e-03
-9.05766338e-02 -4.78513032e-01 -4.10324514e-01 -1.30142653e+00
-2.22151145e-01 -1.30921257e+00 1.25632674e-01 -1.10643053e+00
-7.80822396e-01 7.38650858e-01 4.20256183e-02 -7.41189897e-01
-8.41424823e-01 -5.99440813e-01 -9.59337950e-01 1.33499575e+00
-1.16076243e+00 -1.19096684e+00 4.16113995e-02 1.22918494e-01
7.61250436e-01 -3.14940989e-01 1.01345921e+00 1.17443420e-01
-4.39979911e-01 5.66162586e-01 -4.91740346e-01 2.09977582e-01
2.58522600e-01 -1.54311919e+00 1.24231553e+00 1.26456714e+00
-1.72801614e-01 1.07645142e+00 7.34051168e-01 -7.45095670e-01
-1.47778118e+00 -1.65530610e+00 1.17797792e+00 -8.31886053e-01
8.53491247e-01 -9.64773118e-01 -7.37884521e-01 9.80490685e-01
3.16213280e-01 -2.63603568e-01 3.80806178e-01 -9.41907763e-02
-7.32789099e-01 5.65866888e-01 -5.35430491e-01 8.46310318e-01
8.93429637e-01 -8.82014394e-01 -8.11733902e-01 2.84366578e-01
1.33254063e+00 -5.78825235e-01 -5.27408421e-01 -7.82256275e-02
1.98483780e-01 -6.57410085e-01 5.12734950e-01 -8.20081353e-01
1.54214358e+00 -6.64019644e-01 -3.86201471e-01 -1.40449739e+00
-1.01058431e-01 -8.27129841e-01 1.72044575e-01 1.27615058e+00
1.01701272e+00 -2.23198116e-01 1.27328902e-01 6.45254374e-01
-4.12695676e-01 -3.32206249e-01 -2.40852207e-01 -4.99938071e-01
5.25692940e-01 -8.62196743e-01 3.50128293e-01 7.37460315e-01
4.12983537e-01 6.01737022e-01 -3.33444774e-02 -1.34386169e-02
2.62199461e-01 3.13574255e-01 9.63596642e-01 -6.84405804e-01
-9.29411054e-01 -5.71239233e-01 2.73425698e-01 -9.86195147e-01
5.78932285e-01 -1.58660865e+00 5.18325567e-01 -1.20018649e+00
2.74908870e-01 -3.68564487e-01 4.82785195e-01 6.89968586e-01
-1.43727347e-01 -3.75579089e-01 1.30982354e-01 1.95911378e-01
-2.95157909e-01 4.44445491e-01 6.36650443e-01 -2.70459116e-01
-3.27803284e-01 -3.57577384e-01 -9.91329074e-01 3.56850654e-01
4.23963666e-01 -7.89628983e-01 -5.70202827e-01 -1.03454053e+00
9.16293621e-01 5.25040865e-01 5.42092383e-01 -8.87215137e-01
3.35322730e-02 -2.86228150e-01 -2.78631866e-01 6.41847774e-03
-6.10033929e-01 -2.90360034e-01 1.78372398e-01 6.99070454e-01
-1.00980079e+00 5.83501279e-01 5.79981327e-01 3.19499612e-01
-1.64895002e-02 -5.52922964e-01 6.45733297e-01 -5.14636457e-01
-6.35013103e-01 -2.52308011e-01 -7.55010366e-01 2.83051789e-01
7.10837245e-01 3.22481900e-01 -5.08655310e-01 -5.02143651e-02
-4.05890733e-01 4.02733266e-01 5.72875142e-01 8.34960938e-01
4.66001213e-01 -1.18008804e+00 -8.66951644e-01 2.66553223e-01
6.70623362e-01 1.98307764e-02 -3.78943652e-01 4.27293569e-01
-8.21316719e-01 5.45794547e-01 1.12254679e-01 -6.00179791e-01
-8.65749061e-01 6.35185361e-01 4.82463807e-01 -3.78818482e-01
-3.67750585e-01 9.30772185e-01 5.89474499e-01 -9.11403358e-01
-1.31707177e-01 -8.32113266e-01 2.91609734e-01 -7.24381328e-01
5.49314916e-01 -4.03871059e-01 -7.51727894e-02 -1.02538213e-01
-3.53870183e-01 1.65936813e-01 8.51210579e-02 -2.72086024e-01
1.27216172e+00 4.08087552e-01 -6.86816335e-01 5.00304461e-01
1.05535507e+00 -5.10598086e-02 -1.05101454e+00 -2.38228232e-01
6.13550544e-01 -5.76600805e-02 -8.85075256e-02 -9.98458862e-01
-6.21500552e-01 9.69983280e-01 -1.20713725e-03 -7.76970088e-02
6.90625608e-01 2.91579645e-02 3.41695487e-01 3.03134024e-01
4.37001735e-01 -5.49311221e-01 4.67665672e-01 9.68054235e-01
1.30607939e+00 -1.00816059e+00 -6.30800903e-01 -1.20290861e-01
-2.29405344e-01 1.44067752e+00 8.63965034e-01 -3.00510079e-01
3.83961797e-01 6.42099619e-01 -1.94439381e-01 -1.99018002e-01
-1.69964254e+00 1.04135565e-01 -1.49799674e-03 4.29534942e-01
1.03398049e+00 -6.41908795e-02 -9.78077650e-02 8.09653103e-01
-3.56049240e-01 2.23613754e-01 7.48229265e-01 1.11335385e+00
-9.30324718e-02 -1.28286517e+00 -2.07797080e-01 2.55584866e-01
-1.02510169e-01 -8.68119240e-01 -2.92316049e-01 5.54743588e-01
1.15145139e-01 8.40457618e-01 3.12904008e-02 -4.38841283e-01
1.65804148e-01 1.74619511e-01 4.30248320e-01 -1.47178721e+00
-9.57796037e-01 -3.26334238e-01 2.28031442e-01 -4.45085764e-01
2.77653396e-01 -5.17696261e-01 -1.39920104e+00 -3.27260166e-01
-2.29519196e-02 1.39056370e-01 4.65872943e-01 7.56618738e-01
4.14496511e-01 9.90988851e-01 2.67786533e-01 -7.56891668e-01
-2.50955433e-01 -5.98357201e-01 1.87119439e-01 4.78434153e-02
4.92886245e-01 1.13964401e-01 -6.12213314e-02 5.54849565e-01] | [7.803714275360107, 7.821127891540527] |
29480276-6728-488b-ac49-e87c901ed057 | diverse-text-generation-via-variational | 2204.01227 | null | https://arxiv.org/abs/2204.01227v1 | https://arxiv.org/pdf/2204.01227v1.pdf | Diverse Text Generation via Variational Encoder-Decoder Models with Gaussian Process Priors | Generating high quality texts with high diversity is important for many NLG applications, but current methods mostly focus on building deterministic models to generate higher quality texts and do not provide many options for promoting diversity. In this work, we present a novel latent structured variable model to generate high quality texts by enriching contextual representation learning of encoder-decoder models. Specifically, we introduce a stochastic function to map deterministic encoder hidden states into random context variables. The proposed stochastic function is sampled from a Gaussian process prior to (1) provide infinite number of joint Gaussian distributions of random context variables (diversity-promoting) and (2) explicitly model dependency between context variables (accurate-encoding). To address the learning challenge of Gaussian processes, we propose an efficient variational inference approach to approximate the posterior distribution of random context variables. We evaluate our method in two typical text generation tasks: paraphrase generation and text style transfer. Experimental results on benchmark datasets demonstrate that our method improves the generation quality and diversity compared with other baselines. | ['Yangfeng Ji', 'LiWei Wang', 'Jianqiao Zhao', 'Wanyu Du'] | 2022-04-04 | null | null | null | null | ['paraphrase-generation', 'text-style-transfoer', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing', 'natural-language-processing'] | [ 4.59577084e-01 3.12452435e-01 -1.78380489e-01 -3.24077189e-01
-1.43670046e+00 -3.98856938e-01 1.03860569e+00 -3.16690773e-01
-3.24090215e-04 1.26484168e+00 8.39415133e-01 -3.13984901e-01
2.80938566e-01 -9.53442097e-01 -7.65085340e-01 -6.88399374e-01
6.96311116e-01 9.22526300e-01 -2.31627092e-01 -2.05185726e-01
2.93404102e-01 -1.87932968e-01 -1.30726433e+00 4.06678349e-01
1.47401416e+00 3.13304454e-01 5.73486984e-01 9.17884767e-01
-5.55625498e-01 6.41501009e-01 -1.03646386e+00 -6.83111608e-01
-9.77994427e-02 -1.04068255e+00 -7.73746431e-01 -5.45218550e-02
7.19135031e-02 -3.37197334e-01 -1.29559562e-02 1.03576791e+00
6.80296481e-01 2.53910363e-01 1.20671940e+00 -1.05277646e+00
-1.35361385e+00 1.13408101e+00 -2.98060775e-01 -5.38465194e-02
4.15058613e-01 1.24056555e-01 1.26259482e+00 -8.81080806e-01
6.71371639e-01 1.62977362e+00 2.42263749e-01 8.73559952e-01
-1.50220644e+00 -4.79725838e-01 2.55159974e-01 -2.28271529e-01
-1.18214536e+00 -4.29180115e-01 6.73852205e-01 -3.12468827e-01
8.34459305e-01 7.56200925e-02 4.36465383e-01 1.88209248e+00
1.94792643e-01 1.20630348e+00 7.92338133e-01 -5.07135570e-01
3.94559532e-01 8.29258263e-02 -1.94228768e-01 2.49782681e-01
9.59851071e-02 -1.08074158e-01 -6.72543228e-01 -3.53035212e-01
7.41789699e-01 -1.96498439e-01 -1.69968098e-01 -9.35284346e-02
-1.33523452e+00 1.07572138e+00 -1.94779560e-01 4.86929715e-02
-3.44151258e-01 2.88484097e-01 1.27745405e-01 6.58723637e-02
5.72449923e-01 4.65547204e-01 -1.43802375e-01 -5.15918612e-01
-1.05367482e+00 7.51128972e-01 8.66006732e-01 1.45634747e+00
4.88589853e-01 2.63579726e-01 -1.17349255e+00 1.01413655e+00
3.79621059e-01 1.02648664e+00 4.91268992e-01 -7.72135794e-01
9.15381730e-01 5.21262996e-02 4.67718661e-01 -5.11405528e-01
3.06507498e-01 -1.84545875e-01 -9.87080932e-01 -3.58177185e-01
2.56633788e-01 -4.65555727e-01 -8.17513406e-01 1.77616882e+00
3.78165394e-02 1.11153737e-01 2.25303948e-01 5.67047358e-01
6.36967957e-01 1.21707463e+00 8.30101371e-02 -1.02745280e-01
1.03413522e+00 -1.00393975e+00 -1.08712280e+00 1.38655799e-02
2.71394759e-01 -8.63337576e-01 1.55138206e+00 2.48445526e-01
-1.47604036e+00 -5.34032583e-01 -7.75266647e-01 -2.65246451e-01
9.73546058e-02 1.85555354e-01 2.73056895e-01 6.64729893e-01
-1.09401333e+00 7.33904779e-01 -6.44435167e-01 2.79774249e-01
4.43323255e-01 -9.77211073e-02 2.86612272e-01 1.13420263e-01
-1.46369243e+00 5.90395689e-01 3.21662635e-01 -2.36051515e-01
-9.96620417e-01 -6.53327286e-01 -8.47340226e-01 1.98497280e-01
6.55811056e-02 -1.34714711e+00 1.49592185e+00 -7.20930815e-01
-2.19692612e+00 3.38353008e-01 -6.38326466e-01 -4.03422922e-01
5.41866422e-01 -5.37873626e-01 -4.43216637e-02 -9.55927372e-02
2.29057953e-01 7.20919609e-01 1.09767997e+00 -1.35667062e+00
-4.88720953e-01 2.98047155e-01 -2.77386755e-01 4.28950459e-01
-2.95394003e-01 -1.39815450e-01 -3.82653713e-01 -1.00907230e+00
-4.58167702e-01 -6.91045702e-01 -3.28077614e-01 -3.65203947e-01
-6.18090808e-01 -3.49446028e-01 2.31632918e-01 -7.96891749e-01
1.43265402e+00 -1.70912671e+00 5.74652791e-01 -8.05561244e-03
-1.10181324e-01 8.60605463e-02 -1.91652492e-01 5.09037375e-01
5.34368217e-01 3.41963351e-01 -4.15018439e-01 -8.41104031e-01
4.62950915e-01 3.11496973e-01 -9.03025091e-01 -2.01415792e-01
3.20159227e-01 1.21845317e+00 -1.23008740e+00 -5.59974968e-01
4.89930343e-03 6.83341742e-01 -7.78773010e-01 4.56781894e-01
-7.68613279e-01 4.20745939e-01 -4.50822026e-01 2.21519023e-01
4.53710765e-01 -3.67469251e-01 3.41868885e-02 4.31153953e-01
3.27874601e-01 5.91328263e-01 -1.03249431e+00 1.88365126e+00
-6.60603464e-01 4.35984999e-01 -4.98851866e-01 -2.31580526e-01
1.08365893e+00 4.47700530e-01 -1.81679532e-01 -2.12089583e-01
-9.60135181e-03 -4.24984246e-02 -4.44468915e-01 -3.35832268e-01
1.07093453e+00 -2.99568057e-01 -2.55580425e-01 6.17961168e-01
8.77659693e-02 -6.09357417e-01 1.31634593e-01 3.78040224e-01
5.80255628e-01 4.13272738e-01 5.61310090e-02 -8.70168135e-02
3.08991104e-01 -5.03675222e-01 4.49355841e-01 1.02125204e+00
1.56680822e-01 8.74841392e-01 8.36326838e-01 1.77730858e-01
-1.40833759e+00 -1.43584967e+00 6.59331307e-02 9.94053125e-01
-8.53172913e-02 -4.75263923e-01 -9.09652233e-01 -5.73459268e-01
-2.27049932e-01 1.48935091e+00 -5.57095647e-01 -4.39941660e-02
-4.76141989e-01 -6.52082860e-01 5.19734204e-01 5.38712442e-01
1.84574813e-01 -1.17140520e+00 -5.48130348e-02 3.05535972e-01
-6.54191494e-01 -7.91541517e-01 -8.91416132e-01 -1.97516784e-01
-7.63353109e-01 -2.75397688e-01 -1.07712650e+00 -6.46547079e-01
4.24982458e-01 -4.00402509e-02 1.44470668e+00 -3.97236198e-01
2.34771594e-01 2.06643809e-02 -3.28923076e-01 -4.81775343e-01
-9.32089150e-01 3.58004063e-01 -2.27815986e-01 -1.08097859e-01
1.12782620e-01 -5.06763399e-01 -3.86442155e-01 -1.41324267e-01
-9.57965195e-01 3.68889004e-01 5.86703539e-01 1.14651370e+00
6.97419107e-01 -4.61552620e-01 8.59241664e-01 -1.05602324e+00
1.25635064e+00 -6.50979996e-01 -4.19488162e-01 3.65717143e-01
-4.96827453e-01 6.78542256e-01 7.94862032e-01 -5.81218779e-01
-1.61792743e+00 -3.96278828e-01 -1.72783792e-01 -2.13642493e-01
-7.05777407e-02 3.39750499e-01 -3.78464907e-01 9.28810298e-01
5.69719613e-01 6.17089510e-01 -1.46948799e-01 -3.08984846e-01
8.44158828e-01 7.05997705e-01 3.47040385e-01 -1.02535367e+00
6.91708207e-01 2.54588604e-01 -1.45711482e-01 -5.47116876e-01
-8.49790156e-01 -2.35586222e-02 -4.54190701e-01 1.35727420e-01
8.69530857e-01 -1.02308011e+00 -2.19742041e-02 2.04246789e-01
-1.33870423e+00 -5.51495910e-01 -4.72989678e-01 2.46671051e-01
-9.14798975e-01 4.09720153e-01 -6.76148951e-01 -9.53044951e-01
-7.00375557e-01 -1.12588048e+00 1.65510094e+00 2.47220486e-01
-4.42320496e-01 -1.20420146e+00 4.11945313e-01 1.54528856e-01
4.70532268e-01 1.44476533e-01 1.05490518e+00 -3.25108409e-01
-6.55354559e-01 2.40700677e-01 3.01380083e-02 2.20664635e-01
7.81483054e-02 3.11822426e-02 -8.78812253e-01 -2.31453739e-02
-2.88235575e-01 -2.51951844e-01 7.65000999e-01 4.38011259e-01
1.01085627e+00 -5.41157007e-01 -2.49663293e-01 5.49919724e-01
1.05310011e+00 -6.56346083e-02 8.42962265e-01 -9.34516042e-02
7.36893833e-01 4.70330417e-01 4.22144949e-01 6.88090384e-01
3.96544874e-01 4.67724264e-01 -1.32452056e-01 2.87130326e-01
-7.21494928e-02 -8.81053686e-01 6.37551486e-01 1.00398386e+00
1.41166985e-01 -6.25961542e-01 -6.24667168e-01 6.79782093e-01
-1.85643864e+00 -1.26805866e+00 -1.07989706e-01 1.90073526e+00
1.37436521e+00 2.99176611e-02 8.79825056e-02 -3.32834125e-01
9.28867102e-01 1.97555885e-01 -4.03854668e-01 -3.96648973e-01
-2.26955652e-01 3.55157346e-01 1.14517258e-02 8.35529923e-01
-6.88987970e-01 1.25099456e+00 6.45907736e+00 1.06453061e+00
-5.25563598e-01 4.75463010e-02 5.94336152e-01 -4.09601182e-01
-1.18750000e+00 1.39843794e-02 -1.17717242e+00 8.10030162e-01
9.93261635e-01 -5.79981267e-01 3.68153781e-01 7.83618212e-01
3.50799412e-01 3.36728245e-01 -1.17361701e+00 9.67634201e-01
1.58429772e-01 -1.40116096e+00 7.13923216e-01 1.54120708e-02
1.38493299e+00 -5.71653545e-01 3.05462748e-01 6.43291235e-01
8.89965117e-01 -1.26412821e+00 9.17789102e-01 9.24734056e-01
7.75648654e-01 -9.40005481e-01 4.12499070e-01 4.95753974e-01
-6.96066201e-01 2.14845300e-01 -6.36370957e-01 1.77525431e-01
6.61570847e-01 7.05371141e-01 -8.01430643e-01 3.88300419e-01
8.72127861e-02 5.51677644e-01 -2.08333030e-01 5.02930462e-01
-6.99749112e-01 8.36642623e-01 1.39539137e-01 -5.46130896e-01
1.34637281e-01 -3.47356558e-01 6.56740189e-01 1.33250701e+00
5.32277942e-01 -1.01142772e-01 -1.12344153e-01 1.45633316e+00
-2.32667476e-01 2.28677422e-01 -6.05070829e-01 -2.06090938e-02
6.81650460e-01 8.22345793e-01 -2.11739972e-01 -5.45118988e-01
-1.06468096e-01 1.32079804e+00 4.32856947e-01 6.69025242e-01
-9.25066590e-01 -4.52901781e-01 6.21240735e-01 -1.86510369e-01
4.65218037e-01 -2.40396738e-01 -5.76373637e-01 -1.34192073e+00
3.61488238e-02 -9.77412701e-01 6.69322610e-02 -8.19240332e-01
-1.43162489e+00 5.49403369e-01 5.36514223e-02 -9.17798162e-01
-7.43705511e-01 -6.02795929e-02 -6.66680992e-01 1.53082764e+00
-1.45156312e+00 -1.01542234e+00 1.06446974e-01 3.81474257e-01
8.63175333e-01 -3.31294984e-01 7.80136406e-01 -1.84327796e-01
-2.74940282e-01 8.07195544e-01 4.91007388e-01 -1.14175148e-01
8.06843996e-01 -1.59051967e+00 9.72453058e-01 8.71854961e-01
1.72235921e-01 8.74159336e-01 8.18984926e-01 -9.81830418e-01
-9.79227662e-01 -1.22792149e+00 1.33043647e+00 -8.48218858e-01
3.63560975e-01 -6.41378045e-01 -7.77231932e-01 5.58691859e-01
5.28807402e-01 -8.72902751e-01 8.40842009e-01 5.37895970e-03
-1.47191927e-01 5.65676868e-01 -8.66899550e-01 1.04844856e+00
9.90313232e-01 -5.10178268e-01 -6.92872047e-01 2.77846634e-01
1.20685959e+00 -4.32486355e-01 -5.48018515e-01 -7.82503858e-02
2.67140716e-01 -6.70983493e-01 8.17008317e-01 -7.07910657e-01
1.06545496e+00 -1.42282188e-01 -1.21184602e-01 -1.65567625e+00
-4.03381586e-01 -1.15689850e+00 -5.77175081e-01 1.45834851e+00
5.97536385e-01 -3.01071823e-01 7.22516179e-01 5.73497236e-01
-1.67419940e-01 -8.10493290e-01 -6.21330917e-01 -5.62614262e-01
6.07211351e-01 -2.09623620e-01 9.90175843e-01 5.59837282e-01
-2.17605963e-01 6.41011238e-01 -6.98998928e-01 -8.43689516e-02
5.80192745e-01 2.64244169e-01 9.47082281e-01 -7.02359140e-01
-7.81128824e-01 -5.33521295e-01 3.76494646e-01 -1.71587777e+00
3.53194594e-01 -1.00976801e+00 3.24519783e-01 -1.86851108e+00
4.07982856e-01 -1.98879153e-01 3.47276330e-01 -1.47810102e-01
-9.80118096e-01 -4.14844811e-01 4.95854132e-02 1.65833712e-01
-6.35055780e-01 1.26496708e+00 1.46380353e+00 3.51437703e-02
-2.86300182e-01 6.89660609e-02 -9.68800783e-01 1.94180310e-01
7.98417926e-01 -4.37191993e-01 -7.99563885e-01 -6.22357070e-01
4.42349911e-01 3.00678521e-01 1.08891986e-01 -5.64472377e-01
-1.01042010e-01 -4.18144822e-01 2.89464504e-01 -7.23714828e-01
1.82579383e-01 -7.32012615e-02 2.40772925e-02 -4.35840711e-02
-8.98016512e-01 -8.47232565e-02 -1.52856216e-01 8.33925664e-01
-3.39073651e-02 -4.65957731e-01 5.54733515e-01 -5.81665598e-02
4.50976901e-02 2.33852476e-01 -5.21104634e-01 6.53330386e-01
5.94726384e-01 1.95423245e-01 -1.06079891e-01 -7.35298514e-01
-4.77483213e-01 3.40937138e-01 3.46543163e-01 5.30869722e-01
6.24512851e-01 -1.63799810e+00 -1.09974599e+00 -2.68202480e-02
-4.76948209e-02 2.91310817e-01 1.80199385e-01 -1.00763910e-01
-3.07149976e-01 3.74865115e-01 1.98696494e-01 -5.45276463e-01
-7.99729466e-01 4.77190048e-01 7.19724670e-02 -7.00700164e-01
-4.55315530e-01 1.10889387e+00 1.93513244e-01 -4.53633517e-01
1.67771012e-01 -3.53915155e-01 -1.30529165e-01 -1.37507409e-01
7.55494952e-01 2.65541881e-01 -4.89597440e-01 -2.41301641e-01
4.11495894e-01 2.18091346e-02 -1.62206516e-01 -6.83102608e-01
1.03762710e+00 -3.27597916e-01 1.55261904e-01 6.10903084e-01
9.97629702e-01 1.96101163e-02 -1.63870907e+00 -2.60570735e-01
-1.17502801e-01 -5.06306469e-01 -2.52911001e-01 -6.36251032e-01
-3.31765383e-01 1.10366356e+00 4.02326994e-02 -4.55239825e-02
6.42164469e-01 -9.84884202e-02 1.10412216e+00 2.48983160e-01
2.59010911e-01 -1.21556818e+00 3.90041262e-01 8.15523863e-01
9.50507224e-01 -8.36866915e-01 -4.25943345e-01 -1.60816833e-01
-1.21277773e+00 8.89556587e-01 4.45676059e-01 -1.42441802e-02
2.51646370e-01 1.54720366e-01 -3.12627032e-02 2.63659239e-01
-1.07995641e+00 -2.08235644e-02 2.68916637e-01 8.14495683e-01
6.66447341e-01 1.00540012e-01 -2.17004120e-01 8.61903906e-01
-6.50080264e-01 -3.82669680e-02 4.85957086e-01 5.64611673e-01
-4.68980849e-01 -1.43151188e+00 -2.66061246e-01 3.51540565e-01
-4.15047288e-01 -7.38413751e-01 -2.46666312e-01 8.85929242e-02
-2.55267501e-01 8.31971288e-01 4.23625335e-02 6.14675470e-02
4.89811301e-02 4.00344193e-01 4.30742174e-01 -8.30846488e-01
-4.71294671e-01 1.64065897e-01 5.93449920e-02 -2.58729979e-02
1.24799386e-01 -9.38641906e-01 -1.01651156e+00 -2.37452775e-01
-1.60864949e-01 3.03597242e-01 4.47955906e-01 7.56602526e-01
4.72476542e-01 7.59260297e-01 5.09367287e-01 -6.91528320e-01
-1.04155302e+00 -1.12661636e+00 -4.64734107e-01 5.50914943e-01
4.65379767e-02 -3.59779119e-01 -3.08524489e-01 3.92721117e-01] | [11.877233505249023, 9.126742362976074] |
2d349f62-b173-461c-beb2-380f148eb9a0 | portrait-eyeglasses-and-shadow-removal-by | 2203.10474 | null | https://arxiv.org/abs/2203.10474v1 | https://arxiv.org/pdf/2203.10474v1.pdf | Portrait Eyeglasses and Shadow Removal by Leveraging 3D Synthetic Data | In portraits, eyeglasses may occlude facial regions and generate cast shadows on faces, which degrades the performance of many techniques like face verification and expression recognition. Portrait eyeglasses removal is critical in handling these problems. However, completely removing the eyeglasses is challenging because the lighting effects (e.g., cast shadows) caused by them are often complex. In this paper, we propose a novel framework to remove eyeglasses as well as their cast shadows from face images. The method works in a detect-then-remove manner, in which eyeglasses and cast shadows are both detected and then removed from images. Due to the lack of paired data for supervised training, we present a new synthetic portrait dataset with both intermediate and final supervisions for both the detection and removal tasks. Furthermore, we apply a cross-domain technique to fill the gap between the synthetic and real data. To the best of our knowledge, the proposed technique is the first to remove eyeglasses and their cast shadows simultaneously. The code and synthetic dataset are available at https://github.com/StoryMY/take-off-eyeglasses. | ['Feng Xu', 'Zhibo Wang', 'Junfeng Lyu'] | 2022-03-20 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Lyu_Portrait_Eyeglasses_and_Shadow_Removal_by_Leveraging_3D_Synthetic_Data_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Lyu_Portrait_Eyeglasses_and_Shadow_Removal_by_Leveraging_3D_Synthetic_Data_CVPR_2022_paper.pdf | cvpr-2022-1 | ['shadow-removal'] | ['computer-vision'] | [ 3.69540393e-01 -1.53635927e-02 3.19813550e-01 -3.96672130e-01
-1.35687426e-01 -4.95861888e-01 4.30045038e-01 -4.77078676e-01
3.03432584e-01 7.13820994e-01 -1.47865981e-01 -2.46071228e-04
3.70724946e-01 -4.79137748e-01 -6.84575856e-01 -8.45040023e-01
3.73776376e-01 -1.54440001e-01 1.56560525e-01 -1.91633061e-01
-9.20911971e-03 5.54156303e-01 -1.89885724e+00 2.30434135e-01
8.70308995e-01 8.93918812e-01 2.37629861e-01 2.30485305e-01
9.82308239e-02 4.95793551e-01 -8.73122275e-01 -2.83470005e-01
3.82350147e-01 -4.82307255e-01 -3.47742112e-03 7.84595191e-01
8.52922499e-01 -6.10586405e-01 -2.93093175e-01 9.36854064e-01
3.38483930e-01 1.96909923e-02 4.66625124e-01 -1.55948746e+00
-6.56749010e-01 -1.79891571e-01 -1.18677402e+00 -1.81354433e-01
4.71531481e-01 1.63784832e-01 4.25838381e-01 -1.31368780e+00
5.91260970e-01 1.37195504e+00 4.38234836e-01 7.64688253e-01
-1.15991175e+00 -1.31948745e+00 2.43680999e-01 -6.09176308e-02
-1.58568072e+00 -1.05003226e+00 9.97287512e-01 -3.25246096e-01
2.76365340e-01 4.02626961e-01 6.45163298e-01 1.24245703e+00
-6.41475096e-02 9.43668544e-01 1.48393035e+00 -4.54495192e-01
-6.37485012e-02 4.94578063e-01 5.28183989e-02 8.39208186e-01
7.33551234e-02 1.62187621e-01 -6.04690313e-01 -1.09726012e-01
3.66442859e-01 2.75922120e-01 -5.57729304e-01 8.66217911e-02
-5.16804636e-01 1.56476736e-01 1.57891244e-01 -2.03277230e-01
-2.28576437e-02 -2.36718118e-01 -1.71817511e-01 8.57141092e-02
7.74950027e-01 -1.81978986e-01 -1.29641116e-01 4.92745966e-01
-1.12829804e+00 1.09474614e-01 5.56613684e-01 1.13042557e+00
8.03953469e-01 2.61772666e-02 -5.49815260e-02 1.06208539e+00
3.92625451e-01 8.24017048e-01 -5.48807941e-02 -4.24983084e-01
2.68107653e-01 7.42522955e-01 8.89205486e-02 -1.14648712e+00
-5.53634465e-02 2.21340246e-02 -7.10966766e-01 4.86330450e-01
3.42767864e-01 -1.44552127e-01 -1.21475315e+00 1.58374202e+00
8.21668804e-01 5.87740898e-01 -1.64360270e-01 1.10644925e+00
1.05821335e+00 5.91162145e-01 -3.20367277e-01 -3.39531064e-01
1.53740442e+00 -9.32336390e-01 -1.19197488e+00 -3.49888623e-01
3.51612009e-02 -1.25348592e+00 1.09113693e+00 4.14849997e-01
-8.23700845e-01 -3.96533132e-01 -1.24970996e+00 -1.60920963e-01
-7.71745220e-02 4.66490299e-01 2.97370404e-01 7.44773924e-01
-7.30354846e-01 2.86978692e-01 -6.32122636e-01 -3.24164718e-01
5.85624874e-01 9.73017588e-02 -2.63256788e-01 -1.88529074e-01
-8.82469356e-01 6.55216873e-01 -3.29987586e-01 4.88331079e-01
-8.89833987e-01 -6.06380403e-01 -8.60131264e-01 -3.39018285e-01
7.84877241e-01 -1.21790148e-01 1.10935974e+00 -1.11384916e+00
-1.54139495e+00 1.04834354e+00 -4.83267814e-01 2.57238150e-01
5.92627645e-01 -2.58314312e-01 -5.20397723e-01 -8.23963583e-02
-8.31318498e-02 2.25098282e-01 1.47818196e+00 -1.61186898e+00
-2.19490781e-01 -6.38749659e-01 -1.29409626e-01 1.06998257e-01
-4.78924334e-01 2.46001869e-01 -7.28287816e-01 -6.89532876e-01
1.12052392e-02 -1.20789766e+00 4.48327780e-01 4.01696235e-01
-6.83555484e-01 1.55319467e-01 1.49215448e+00 -7.24642515e-01
1.03449726e+00 -2.34589958e+00 -3.83010775e-01 -9.53081846e-02
3.23992908e-01 3.78124535e-01 -2.48008706e-02 3.34943831e-01
-8.63618478e-02 -9.52826627e-03 -2.08025917e-01 -8.45987618e-01
-2.14632452e-01 2.03644469e-01 -3.83109301e-01 7.06343293e-01
3.14068288e-01 3.37400228e-01 -5.88020086e-01 -4.78501350e-01
4.85176519e-02 7.19893932e-01 3.60095464e-02 1.41070664e-01
-6.13916591e-02 5.13716042e-01 -3.07728618e-01 1.22606957e+00
1.32492959e+00 1.68441087e-01 1.11348376e-01 -1.66215122e-01
5.58908060e-02 -8.03021342e-02 -1.24720788e+00 1.06989765e+00
-3.03854346e-01 7.65145540e-01 4.61263746e-01 -2.28186566e-02
1.06925237e+00 3.09078366e-01 1.88718691e-01 -6.61471963e-01
1.30699962e-01 1.90214545e-01 -3.57073247e-01 -5.57728410e-01
4.41265553e-01 -2.59342998e-01 2.79732555e-01 4.08114970e-01
-3.77927095e-01 -1.27704903e-01 1.76230296e-01 7.54835010e-02
7.96978593e-01 2.32608795e-01 1.02239035e-01 -1.38098374e-01
5.22320390e-01 -3.72704089e-01 8.28988314e-01 1.19708203e-01
-1.38459921e-01 7.45596647e-01 3.95607859e-01 -3.50500405e-01
-7.70792961e-01 -8.04906189e-01 -2.83080131e-01 9.25728381e-01
3.19776386e-01 -4.43169147e-01 -1.06239390e+00 -5.74462831e-01
1.51508972e-01 4.32245910e-01 -4.92846042e-01 -6.32735640e-02
-4.15673286e-01 -6.26647830e-01 4.02535558e-01 7.35174045e-02
5.22993147e-01 -9.67682242e-01 -3.74153703e-01 -3.63381118e-01
-8.94897431e-02 -1.32196891e+00 -8.08027625e-01 -4.18823659e-01
-4.69340116e-01 -1.29024279e+00 -5.65581143e-01 -4.98531818e-01
1.12009954e+00 7.82554448e-01 8.45765352e-01 5.54104984e-01
-6.12817466e-01 -6.83275685e-02 -2.53822684e-01 -8.13676178e-01
-3.03578407e-01 -4.08114821e-01 9.49521586e-02 6.22686207e-01
1.26016960e-01 -4.18309867e-01 -7.13052869e-01 7.69573927e-01
-8.56751621e-01 2.88239777e-01 3.19562197e-01 6.83193326e-01
4.30710942e-01 2.22611308e-01 1.30087942e-01 -1.02617753e+00
5.08680165e-01 -2.05927312e-01 -9.17206168e-01 -2.30921395e-02
-3.26244086e-01 -5.31585395e-01 5.29762387e-01 -4.18939471e-01
-1.25636268e+00 1.27314642e-01 2.86899716e-01 -7.45893180e-01
-2.81608641e-01 -2.90026724e-01 -4.76765454e-01 -1.99224800e-01
4.38731968e-01 -2.31496487e-02 -2.80276523e-03 -6.45365059e-01
-1.18261188e-01 8.09907973e-01 2.28360772e-01 -2.61975169e-01
1.06619394e+00 8.07952523e-01 2.86680106e-02 -1.22052300e+00
-1.06588209e+00 -1.55445874e-01 -3.80095333e-01 -5.23769498e-01
3.78449887e-01 -8.43457282e-01 -5.84244847e-01 8.33818913e-01
-1.08227396e+00 -2.28832424e-01 1.40621573e-01 -1.03989651e-03
7.65948072e-02 2.98499882e-01 -3.31460476e-01 -1.21667361e+00
-1.98297516e-01 -1.08707404e+00 1.38772905e+00 5.32947958e-01
1.03024460e-01 -4.20006782e-01 -3.35795283e-01 4.72956598e-01
8.00886974e-02 4.82085228e-01 2.37626269e-01 1.51712060e-01
-6.96908176e-01 -1.95637301e-01 -1.76939338e-01 3.92647624e-01
6.18893027e-01 3.67157549e-01 -1.59203517e+00 -3.22557032e-01
2.38116473e-01 -1.34411648e-01 7.17630148e-01 -3.77803743e-02
8.75541866e-01 -2.67092288e-01 -4.48979586e-01 7.48386919e-01
1.00248003e+00 2.03683048e-01 6.38635397e-01 5.99934794e-02
7.91931868e-01 7.72857547e-01 9.37872529e-01 5.17530560e-01
5.63215055e-02 8.50453436e-01 3.64636153e-01 -5.71580589e-01
-5.57050705e-01 -1.43584013e-01 4.70785111e-01 2.43624955e-01
-6.71350304e-03 -5.13520479e-01 -8.00198555e-01 2.37166181e-01
-1.55403471e+00 -6.17564976e-01 -4.60750461e-01 2.04565835e+00
8.62793386e-01 -3.11472952e-01 -1.95887864e-01 1.15477517e-01
9.67749357e-01 3.14836144e-01 -5.10329783e-01 -1.73673555e-02
-1.61555707e-01 -1.00363441e-01 2.50644863e-01 5.42701900e-01
-1.06161618e+00 1.02608860e+00 5.15679646e+00 7.06580997e-01
-1.28724742e+00 8.40633065e-02 5.30713201e-01 -3.91702354e-01
-7.07676113e-02 -4.31201980e-02 -1.01398230e+00 7.55412817e-01
2.00496897e-01 3.40424210e-01 4.41032708e-01 5.95595062e-01
3.69648457e-01 -3.51807743e-01 -8.89430702e-01 1.10245657e+00
5.28758645e-01 -6.05718970e-01 -4.05710459e-01 1.73598126e-01
7.29423523e-01 -4.92460787e-01 1.34416580e-01 -2.89878517e-01
-1.07036173e-01 -8.90865266e-01 7.63782084e-01 5.73799014e-01
9.10273731e-01 -4.71541077e-01 3.15047383e-01 1.48458838e-01
-1.04113901e+00 1.89611286e-01 -1.40311360e-01 2.34572794e-02
6.82490468e-02 8.83794129e-01 -9.78594542e-01 2.97949433e-01
4.73500997e-01 6.33319438e-01 -5.39365590e-01 7.81332791e-01
-7.59131968e-01 2.99070209e-01 -4.37346905e-01 2.01530814e-01
-3.02222669e-01 -4.03402388e-01 7.42841721e-01 9.35611248e-01
3.57552499e-01 3.47565114e-02 2.00165793e-01 9.81469870e-01
-3.03115368e-01 -1.39996111e-01 -7.28723526e-01 8.37516040e-02
5.01678467e-01 1.48979795e+00 -7.21114874e-01 -1.96365848e-01
-3.99991959e-01 1.10816741e+00 -1.46193594e-01 5.89259326e-01
-1.00321865e+00 -1.59255013e-01 8.12317252e-01 7.48997867e-01
3.93856131e-02 -1.17984302e-01 -1.51506096e-01 -1.24420595e+00
4.98666644e-01 -9.23434496e-01 5.47886454e-03 -7.71332264e-01
-1.03773224e+00 6.54834092e-01 -1.30113110e-01 -1.32982159e+00
3.04280221e-01 -4.69602466e-01 -6.40944719e-01 7.87115574e-01
-1.77087092e+00 -1.42026818e+00 -8.98975790e-01 6.70455456e-01
6.43485188e-01 1.72426254e-02 6.26046300e-01 5.31766295e-01
-9.54736471e-01 6.29980564e-01 2.83502527e-02 -3.75185944e-02
1.06016541e+00 -7.38101780e-01 2.42101207e-01 9.82334435e-01
-2.50013471e-02 5.31517208e-01 7.94435322e-01 -7.09398985e-01
-1.51574636e+00 -1.02096260e+00 4.38915074e-01 -3.63736659e-01
4.14649427e-01 -9.79270697e-01 -9.61735845e-01 6.23237371e-01
2.06239656e-01 1.93777457e-01 4.56792146e-01 -2.50550434e-02
-3.29044938e-01 -4.85885859e-01 -1.16504872e+00 6.86484575e-01
9.68493104e-01 -3.55695009e-01 -1.37136132e-01 5.56822121e-01
6.14542998e-02 -5.70162892e-01 -3.73874635e-01 3.00096750e-01
8.05009425e-01 -1.09235656e+00 7.24785447e-01 2.37690628e-01
2.10134894e-01 -6.02739632e-01 2.42242828e-01 -1.24148417e+00
3.65905792e-01 -1.03731740e+00 -1.30050540e-01 1.66810727e+00
1.10282734e-01 -7.95165300e-01 6.42158031e-01 5.93714297e-01
-1.05313145e-01 -5.41933239e-01 -5.75960994e-01 -6.88055575e-01
-8.16685975e-01 5.42470179e-02 7.70679891e-01 8.38245034e-01
-5.29039979e-01 2.83966064e-01 -8.04587722e-01 3.54944736e-01
8.41264963e-01 4.49346006e-01 1.16531086e+00 -1.06019652e+00
3.05686966e-02 5.76604083e-02 -9.43832844e-02 -6.21742725e-01
8.18894431e-02 -3.51786941e-01 2.17988372e-01 -9.86685872e-01
2.30364382e-01 -3.08650941e-01 1.20190576e-01 6.64377332e-01
-2.47720391e-01 5.59287846e-01 3.48871469e-01 2.71505415e-01
-1.64039642e-01 5.82149148e-01 1.43694890e+00 -5.97255072e-03
-2.34318584e-01 3.91998366e-02 -6.56934500e-01 1.05277979e+00
9.07303691e-01 -6.73215926e-01 -5.82865357e-01 -4.61786956e-01
-2.01482877e-01 -2.44416177e-01 5.03868818e-01 -7.68167257e-01
4.85383384e-02 -2.75373071e-01 5.18760383e-01 -5.79291105e-01
7.35630333e-01 -8.01515996e-01 1.84310332e-01 4.09295708e-02
2.78971672e-01 -2.69197404e-01 4.67321664e-01 5.41540265e-01
-7.46091530e-02 1.73014589e-02 1.03468812e+00 -1.41214892e-01
-2.97050983e-01 4.17835444e-01 -2.36360133e-01 -3.37104946e-01
1.21896183e+00 -3.26567769e-01 -5.07340074e-01 -3.31725121e-01
-4.78247434e-01 2.01555416e-01 8.89855504e-01 5.03126979e-01
8.21699739e-01 -1.06908143e+00 -6.83689535e-01 6.87811673e-01
1.05377935e-01 8.02090690e-02 3.84990841e-01 9.04042482e-01
-2.96578914e-01 -1.28279284e-01 -8.50318074e-02 -3.68102014e-01
-2.01134849e+00 4.36223567e-01 2.45473266e-01 3.35985452e-01
-7.16507316e-01 7.28661537e-01 8.59110892e-01 -2.42397767e-02
1.96575835e-01 -7.38772899e-02 -1.58983842e-02 1.81784153e-01
8.80409479e-01 2.79986173e-01 1.35092631e-01 -6.94049358e-01
-3.71390939e-01 6.18068933e-01 -2.20381394e-01 2.40370855e-01
1.20542312e+00 -3.69061112e-01 -1.93567947e-01 2.98930347e-01
9.56929982e-01 3.67813617e-01 -1.30800891e+00 -1.41497746e-01
-3.73582363e-01 -1.10253680e+00 -3.75854075e-02 -4.84111160e-01
-1.22645545e+00 8.62849891e-01 4.07603741e-01 7.39159063e-02
1.42497361e+00 -1.59167379e-01 1.06573212e+00 -2.23262049e-02
1.65810093e-01 -1.18461871e+00 1.58979625e-01 6.75880238e-02
1.20861816e+00 -1.13901746e+00 3.56305599e-01 -9.84501481e-01
-5.29975474e-01 8.13443601e-01 8.70790958e-01 1.54392533e-02
4.93192285e-01 6.05815828e-01 1.90254807e-01 -2.29351580e-01
-6.79778397e-01 -2.64571309e-01 1.74061045e-01 4.28769797e-01
2.65715718e-01 -5.48991524e-02 -7.13424534e-02 2.63307333e-01
-8.40733573e-02 -6.24148063e-02 5.47507823e-01 1.04576921e+00
-8.75435695e-02 -9.89699721e-01 -8.51280510e-01 1.59860909e-01
-3.90832454e-01 -8.29920098e-02 -7.97561109e-01 8.65691781e-01
3.46816123e-01 1.06940556e+00 -1.89563036e-01 -2.65716702e-01
4.27424192e-01 -2.68262029e-02 5.00194907e-01 -6.78666294e-01
-1.65852189e-01 5.22157609e-01 1.16651095e-01 -4.13885206e-01
-2.83510119e-01 -4.91928250e-01 -9.89444792e-01 -3.95789355e-01
-6.12679601e-01 -3.74736190e-01 6.76308513e-01 6.12544954e-01
3.22051167e-01 4.94834274e-01 8.41628134e-01 -8.86180341e-01
-1.11223847e-01 -1.10081756e+00 -9.57599044e-01 5.78930140e-01
6.43580317e-01 -1.04385197e+00 -4.26903009e-01 2.23336175e-01] | [12.902070999145508, -0.00472797779366374] |
3cce8e7e-4fa1-4277-a479-edd14bae6ca2 | squeeze-flow-of-micro-droplets-convolutional | 2211.09061 | null | https://arxiv.org/abs/2211.09061v1 | https://arxiv.org/pdf/2211.09061v1.pdf | Squeeze flow of micro-droplets: convolutional neural network with trainable and tunable refinement | We propose a platform based on neural networks to solve the image-to-image translation problem in the context of squeeze flow of micro-droplets. In the first part of this paper, we present the governing partial differential equations to lay out the underlying physics of the problem. We also discuss our developed Python package, sqflow, which can potentially serve as free, flexible, and scalable standardized benchmarks in the fields of machine learning and computer vision. In the second part of this paper, we introduce a residual convolutional neural network to solve the corresponding inverse problem: to translate a high-resolution (HR) imprint image with a specific liquid film thickness to a low-resolution (LR) droplet pattern image capable of producing the given imprint image for an appropriate spread time of droplets. We propose a neural network architecture that learns to systematically tune the refinement level of its residual convolutional blocks by using the function approximators that are trained to map a given input parameter (film thickness) to an appropriate refinement level indicator. We use multiple stacks of convolutional layers the output of which is translated according to the refinement level indicators provided by the directly-connected function approximators. Together with a non-linear activation function, such a translation mechanism enables the HR imprint image to be refined sequentially in multiple steps until the target LR droplet pattern image is revealed. The proposed platform can be potentially applied to data compression and data encryption. The developed package and datasets are publicly available on GitHub at https://github.com/sqflow/sqflow. | ['S. V. Sreenivasan', 'Shrawan Singhal', 'Aryan Mehboudi'] | 2022-11-16 | null | null | null | null | ['data-compression'] | ['time-series'] | [ 7.96081662e-01 -9.10519511e-02 3.25903952e-01 -2.37302721e-01
-3.86528432e-01 -6.18567050e-01 5.02702594e-01 -2.86408365e-02
-4.75665361e-01 5.04801571e-01 -4.98991311e-01 -2.49083519e-01
-7.17460439e-02 -1.15911174e+00 -1.19366312e+00 -9.93867636e-01
1.58144906e-01 2.34187528e-01 9.80761647e-02 -5.76137118e-02
4.04468656e-01 8.99743557e-01 -1.55659413e+00 5.37445486e-01
7.99194753e-01 1.32833207e+00 3.04994673e-01 1.03916657e+00
8.37599039e-02 4.06643927e-01 -3.22240174e-01 -1.73743919e-01
5.84419072e-01 -5.36031604e-01 -4.83878195e-01 -1.01804197e-01
6.10960543e-01 -2.34614879e-01 -1.60797182e-02 7.96979725e-01
4.74983156e-01 4.22005765e-02 6.95756793e-01 -7.17623115e-01
-6.99835598e-01 3.31391305e-01 -2.65709609e-01 1.70371935e-01
5.58776073e-02 4.29176748e-01 6.32176697e-01 -8.87687683e-01
7.27751851e-01 9.56333160e-01 5.45286477e-01 7.12336600e-01
-1.32667780e+00 -6.44342184e-01 -3.38558465e-01 -3.42109025e-01
-1.22980034e+00 -4.31519926e-01 5.46526253e-01 -4.82215017e-01
9.33153629e-01 4.08128411e-01 7.95784414e-01 6.34187639e-01
5.33511221e-01 -8.61528143e-02 9.98584032e-01 -2.46636838e-01
5.64072654e-02 3.31658721e-01 -3.25135797e-01 6.82101071e-01
2.52163351e-01 1.68253407e-01 -3.56959075e-01 2.59202182e-01
1.14717376e+00 1.72712225e-02 -3.84874254e-01 -3.39310504e-02
-8.45714092e-01 6.96465075e-01 7.36736894e-01 5.39942384e-01
-1.90907434e-01 1.44361779e-01 7.67546892e-02 3.37680638e-01
5.07247627e-01 9.44572449e-01 -2.93438822e-01 3.78713518e-01
-9.61294889e-01 3.00968587e-01 8.91933024e-01 6.84536636e-01
1.05897975e+00 5.91122545e-02 -2.39820614e-01 4.36968237e-01
1.58427760e-01 4.79773074e-01 2.54451156e-01 -9.31048334e-01
3.86500835e-01 6.38582885e-01 1.50490329e-01 -8.21318686e-01
-3.11371326e-01 2.51754951e-02 -1.06045306e+00 6.34784937e-01
2.68383563e-01 -2.61250824e-01 -8.06597233e-01 1.20117497e+00
3.59825253e-01 7.46660754e-02 2.88310140e-01 1.03536689e+00
1.12688136e+00 9.93159354e-01 -4.23359424e-01 -2.11482823e-01
1.27704847e+00 -7.18309820e-01 -5.37538111e-01 2.78442085e-01
7.27674961e-02 -7.52111793e-01 9.71218288e-01 2.36187607e-01
-1.38385618e+00 -7.74787605e-01 -1.22047448e+00 -3.57771099e-01
-6.60568595e-01 2.82329679e-01 9.34209824e-02 2.69788504e-01
-1.22737002e+00 1.14696777e+00 -7.81089067e-01 -9.31392163e-02
5.05554318e-01 7.20886230e-01 -2.40139037e-01 3.72365832e-01
-9.66180265e-01 5.08741260e-01 1.67671099e-01 3.54977965e-01
-6.91124618e-01 -1.12528133e+00 -8.88782859e-01 1.85765550e-02
-1.90812379e-01 -6.56901002e-01 7.75920331e-01 -1.09530103e+00
-1.79858673e+00 1.04379296e+00 -7.00737461e-02 -4.45131540e-01
5.78350902e-01 -8.13226774e-02 7.37450868e-02 2.06832930e-01
-3.72917414e-01 7.35389233e-01 1.29641593e+00 -1.23281872e+00
-4.81680661e-01 -1.47328496e-01 -1.23661056e-01 8.06337371e-02
-2.24427894e-01 -6.73474893e-02 -4.03570712e-01 -3.48259151e-01
-4.09895569e-01 -5.60341656e-01 -1.47016302e-01 1.27766952e-01
-3.07894111e-01 2.30730459e-01 1.08246100e+00 -3.93914759e-01
1.05189383e+00 -1.95695841e+00 4.08307672e-01 -5.80884889e-03
2.50968546e-01 5.31686544e-01 -1.62982732e-01 3.49473566e-01
-2.15119541e-01 2.31047109e-01 -4.65450138e-01 -3.30059469e-01
-3.90583545e-01 -2.67168641e-01 -3.24469745e-01 5.68262339e-01
7.13369489e-01 1.05556107e+00 -4.92598116e-01 -6.44648597e-02
4.72999603e-01 9.60061312e-01 -4.34017181e-01 6.62318170e-01
-4.16785568e-01 8.39470267e-01 -2.03540877e-01 2.86394536e-01
1.06350565e+00 -1.83623731e-01 -7.89183453e-02 -4.92909223e-01
-8.24923694e-01 -1.69371143e-02 -8.18333745e-01 1.17587304e+00
-3.18773478e-01 7.40293741e-01 4.16554093e-01 -6.49907529e-01
1.24654233e+00 1.71927676e-01 5.71490884e-01 -4.47682947e-01
3.63940150e-01 1.60095870e-01 -1.67144015e-01 -5.82940340e-01
4.76001471e-01 -2.61879474e-01 2.40348056e-01 3.81625295e-01
-1.80269882e-01 -4.92797852e-01 2.02333510e-01 -4.20161366e-01
8.22841167e-01 1.18002519e-01 -9.49539691e-02 -1.93477318e-01
8.55585515e-01 -1.00791022e-01 9.77175832e-02 4.79986161e-01
4.17869925e-01 9.20708656e-01 5.22521555e-01 -9.31883514e-01
-1.60316694e+00 -4.66664881e-01 -4.54357564e-01 6.43120944e-01
6.33118972e-02 -5.83521426e-02 -1.04310787e+00 -6.57815346e-03
1.21534541e-01 1.43653721e-01 -7.26578355e-01 6.51746020e-02
-8.61638606e-01 -6.72340572e-01 4.65680778e-01 3.14832069e-02
6.08845770e-01 -1.69005895e+00 -9.25376773e-01 3.83247547e-02
4.56921637e-01 -1.04441845e+00 -2.90110976e-01 2.24612415e-01
-7.17880189e-01 -1.01304424e+00 -6.27987802e-01 -7.50570118e-01
9.38472450e-01 -3.20698619e-02 8.99830103e-01 2.40190119e-01
-4.24390793e-01 5.33526763e-02 -1.04939841e-01 -1.73908547e-01
-6.81545317e-01 2.16876462e-01 -1.18274674e-01 3.42026293e-01
-4.33453545e-02 -5.05748093e-01 -8.52187991e-01 4.42533158e-02
-1.50939763e+00 2.75871396e-01 3.31735730e-01 4.68275934e-01
8.25730920e-01 -6.64113387e-02 8.34640935e-02 -8.67819905e-01
6.09140337e-01 -7.31707439e-02 -1.19908500e+00 -2.77598519e-02
-3.88529837e-01 1.05653167e-01 1.02601314e+00 -4.09318119e-01
-6.18978858e-01 3.13921005e-01 -2.74681807e-01 -7.45822191e-01
-1.22669794e-01 1.86390188e-02 4.76467423e-03 -5.39786398e-01
4.70937401e-01 1.86141446e-01 2.01936081e-01 -2.81134754e-01
7.40880966e-02 5.41702986e-01 2.93663919e-01 -3.16059381e-01
8.78502369e-01 5.29368997e-01 2.32727021e-01 -7.91394114e-01
-3.70678037e-01 -5.27679846e-02 -7.95639455e-01 -2.02740863e-01
1.06118703e+00 -5.05977094e-01 -1.09355438e+00 8.02114010e-01
-1.28492451e+00 -8.72916162e-01 -3.48336667e-01 -1.07537888e-01
-5.72859287e-01 -8.15452114e-02 -7.07197487e-01 -5.48916936e-01
-9.14608002e-01 -1.33492386e+00 1.20550025e+00 5.51213443e-01
1.95741147e-01 -9.95831370e-01 4.06250432e-02 -6.41283616e-02
6.38682961e-01 5.36628783e-01 7.80494690e-01 -2.37608686e-01
-9.45408881e-01 -1.17461458e-01 -2.64845520e-01 4.91659194e-01
3.32649291e-01 5.26155233e-01 -8.99647415e-01 -4.97687817e-01
3.22780460e-02 -2.80565321e-01 9.83744442e-01 5.59364319e-01
1.25324392e+00 -5.49112082e-01 -2.07704112e-01 1.26440215e+00
1.76839352e+00 1.75530121e-01 7.63823211e-01 1.26792505e-01
6.46240711e-01 3.94060194e-01 2.52166361e-01 2.93387055e-01
-1.13958031e-01 6.62977576e-01 4.52980369e-01 -3.92969012e-01
-1.34806350e-01 -6.84325723e-03 3.11266690e-01 3.09802294e-01
-3.26223910e-01 -3.86738122e-01 -3.75937939e-01 6.77429363e-02
-1.38353050e+00 -7.75687754e-01 -6.83988705e-02 2.33467436e+00
8.18755805e-01 -1.16104811e-01 -1.67177349e-01 -1.44781202e-01
7.92698920e-01 1.43192589e-01 -6.35106444e-01 -8.21799338e-01
-3.91367450e-02 6.30489826e-01 5.91850400e-01 8.93781185e-01
-1.18993616e+00 1.09503865e+00 5.60571289e+00 4.89885539e-01
-1.87852645e+00 -3.40130299e-01 6.43027246e-01 -4.37596329e-02
-3.41807753e-01 -4.11140829e-01 -1.05344605e+00 4.75064397e-01
9.63502765e-01 1.11044712e-01 8.12962770e-01 8.43311325e-02
3.63568872e-01 1.90215468e-01 -1.18312585e+00 6.72430575e-01
-1.62317500e-01 -1.78610849e+00 2.78465003e-01 -4.29250039e-02
7.03354180e-01 -7.63974488e-02 4.69278902e-01 -3.19700450e-01
-2.61808425e-01 -1.45410049e+00 4.97521788e-01 6.73590183e-01
1.47203767e+00 -5.71869612e-01 3.71731073e-01 8.80689099e-02
-1.18279552e+00 -4.46253978e-02 -5.21975219e-01 4.14925665e-02
-2.32338250e-01 3.65697205e-01 -9.80865180e-01 3.96963149e-01
7.29163468e-01 8.03205848e-01 -2.60697365e-01 5.94053984e-01
6.32531419e-02 2.44831264e-01 -4.35896903e-01 -1.09471850e-01
1.44184202e-01 -6.70742095e-01 5.43669879e-01 1.31215239e+00
6.66812301e-01 9.46048722e-02 -4.78057802e-01 1.53645861e+00
-3.03460360e-01 1.84272265e-03 -8.93024087e-01 -9.02631730e-02
1.37650624e-01 1.51709783e+00 -7.02850223e-01 -3.82352136e-02
-3.81754227e-02 9.78548586e-01 3.56443971e-01 4.14979100e-01
-6.05142653e-01 -6.30278528e-01 6.86363876e-01 3.97805005e-01
6.51901662e-01 7.45272413e-02 -2.57824808e-01 -8.70471537e-01
1.21516036e-02 -6.69089437e-01 -4.03077938e-02 -7.81395078e-01
-9.76458311e-01 8.39284599e-01 -2.86854774e-01 -1.16209567e+00
-7.24947155e-02 -8.54855239e-01 -7.20923007e-01 1.10645890e+00
-1.76126850e+00 -9.44814801e-01 -4.34522778e-01 3.34059894e-01
2.97975898e-01 -7.68529698e-02 9.46468651e-01 1.32445037e-01
-5.64211905e-01 3.38009149e-01 3.21376845e-02 1.30301431e-01
5.61048090e-01 -1.08014727e+00 4.40476656e-01 6.39745653e-01
-3.18689585e-01 5.21890759e-01 7.53916562e-01 -4.92407709e-01
-1.57661748e+00 -1.33305633e+00 5.43810785e-01 -3.38100016e-01
3.79010528e-01 -4.77981120e-01 -9.65062201e-01 4.51084793e-01
3.31924915e-01 3.49558294e-01 2.35058844e-01 -9.48102832e-01
-5.95290326e-02 -5.41852951e-01 -1.47834647e+00 2.03402698e-01
5.02097666e-01 -1.61361471e-01 1.75710656e-02 2.03008294e-01
8.76654625e-01 -8.98634315e-01 -9.82069492e-01 4.99745160e-01
5.37906229e-01 -1.08731902e+00 8.20139408e-01 -4.59793001e-01
1.01654017e+00 -4.11869854e-01 2.90569484e-01 -9.76893902e-01
-1.92536503e-01 -8.75019252e-01 7.94459954e-02 9.48236525e-01
5.86150885e-01 -6.90211654e-01 7.12052464e-01 5.16003072e-01
1.57366097e-02 -1.08698702e+00 -9.96750593e-01 -2.72663701e-02
2.85654932e-01 1.52611844e-02 6.76259995e-01 5.00133455e-01
-4.23919588e-01 1.07492626e-01 -2.80852884e-01 4.56328660e-01
3.90885085e-01 3.48105431e-01 6.93569779e-01 -9.74806130e-01
5.45565970e-02 -4.56271946e-01 -1.77862242e-01 -8.24533343e-01
-6.50023520e-02 -9.43837941e-01 1.67435139e-01 -1.24026537e+00
-1.02311552e-01 -5.26549518e-01 -9.97288972e-02 2.59042382e-01
4.62678783e-02 5.41271448e-01 2.94763297e-01 -3.67963389e-02
-1.77389160e-01 2.39372209e-01 1.66179335e+00 -5.81729114e-02
-7.02567995e-01 1.44664198e-01 -5.26105225e-01 9.47204307e-02
9.04061735e-01 -3.64707172e-01 3.62437405e-02 -4.28134233e-01
3.79334122e-01 -3.07396334e-02 4.58111137e-01 -9.98968959e-01
2.04378426e-01 6.47613779e-02 6.39956892e-01 -1.89053580e-01
3.53475481e-01 -1.00213015e+00 3.95152748e-01 6.34993851e-01
-4.71528172e-01 -1.41801029e-01 4.47315127e-01 7.57983699e-03
5.25350124e-02 -2.62650639e-01 1.07797205e+00 -1.54797301e-01
1.07748188e-01 6.50440395e-01 -2.33247980e-01 -4.54132080e-01
9.75170851e-01 -1.68220907e-01 -3.07681173e-01 -8.02702084e-02
-4.04046565e-01 -1.79166179e-02 7.74608970e-01 1.94550440e-01
8.95419896e-01 -1.15562570e+00 -7.77120888e-01 7.98172116e-01
-3.32827002e-01 4.30316985e-01 1.01854928e-01 4.74989921e-01
-1.01478612e+00 3.95744711e-01 -4.22282726e-01 -5.96507072e-01
-1.12042260e+00 5.05329072e-01 8.67126167e-01 -2.23785594e-01
-7.43442714e-01 8.14928770e-01 1.67637736e-01 -4.50493127e-01
-2.22234521e-02 -6.14418268e-01 -3.84528667e-01 -2.18570679e-01
7.12177753e-01 8.39829743e-02 5.66414110e-02 -5.41048706e-01
-5.70433103e-02 1.09455419e+00 1.57890633e-01 1.70547292e-01
1.62800372e+00 2.27368891e-01 -5.56676328e-01 1.22898467e-01
1.33069873e+00 2.99164038e-02 -1.65063906e+00 3.10730726e-01
-6.33522272e-01 -1.59703448e-01 -1.61586583e-01 -4.31586564e-01
-1.42046535e+00 8.61836612e-01 5.46746135e-01 3.77554148e-01
1.23772562e+00 -1.59153298e-01 7.03728855e-01 1.96512848e-01
-2.13383108e-01 -4.63171810e-01 -1.53211862e-01 4.57658678e-01
1.03360498e+00 -9.16892588e-01 2.49937531e-02 -3.67446601e-01
-3.04150879e-01 1.53938186e+00 4.58580136e-01 -4.54446673e-01
4.84267741e-01 8.04895043e-01 1.66235656e-01 -2.06823781e-01
-7.11160243e-01 8.76241997e-02 3.49218130e-01 4.21496034e-01
4.23216879e-01 -1.51483357e-01 4.12091352e-02 2.57104844e-01
-2.73084819e-01 2.81865239e-01 4.72898424e-01 4.55652624e-01
-3.62948984e-01 -1.01038849e+00 -3.76319706e-01 3.26943189e-01
-5.23545146e-01 -8.25416967e-02 -2.52052635e-01 4.39287186e-01
4.43063557e-01 5.07162988e-01 5.16554534e-01 -2.78067052e-01
3.49958986e-01 -1.78759053e-01 4.98795748e-01 -4.65824485e-01
-8.83917212e-01 -6.20144308e-02 -5.55494905e-01 -4.10793990e-01
-5.31390190e-01 -3.92929167e-01 -1.36010885e+00 -3.18651259e-01
1.73846841e-01 2.66803727e-02 5.57257652e-01 4.85286504e-01
3.80779892e-01 4.72860783e-01 7.89785445e-01 -1.39651346e+00
-2.07456369e-02 -8.45022321e-01 -2.80576169e-01 1.32052094e-01
9.53261614e-01 2.08973400e-02 -5.02808928e-01 2.42410377e-01] | [11.031299591064453, -1.1332733631134033] |
aac4cb9f-11c9-4d84-8f7a-f74f3580b09d | frequency-domain-learning-for-volumetric | 2302.08595 | null | https://arxiv.org/abs/2302.08595v2 | https://arxiv.org/pdf/2302.08595v2.pdf | Frequency-domain Learning for Volumetric-based 3D Data Perception | Frequency-domain learning draws attention due to its superior tradeoff between inference accuracy and input data size. Frequency-domain learning in 2D computer vision tasks has shown that 2D convolutional neural networks (CNN) have a stationary spectral bias towards low-frequency channels so that high-frequency channels can be pruned with no or little accuracy degradation. However, frequency-domain learning has not been studied in the context of 3D CNNs with 3D volumetric data. In this paper, we study frequency-domain learning for volumetric-based 3D data perception to reveal the spectral bias and the accuracy-input-data-size tradeoff of 3D CNNs. Our study finds that 3D CNNs are sensitive to a limited number of critical frequency channels, especially low-frequency channels. Experiment results show that frequency-domain learning can significantly reduce the size of volumetric-based 3D inputs (based on spectral bias) while achieving comparable accuracy with conventional spatial-domain learning approaches. Specifically, frequency-domain learning is able to reduce the input data size by 98% in 3D shape classification while limiting the average accuracy drop within 2%, and by 98% in the 3D point cloud semantic segmentation with a 1.48% mean-class accuracy improvement while limiting the mean-class IoU loss within 1.55%. Moreover, by learning from higher-resolution 3D data (i.e., 2x of the original image in the spatial domain), frequency-domain learning improves the mean-class accuracy and mean-class IoU by 3.04% and 0.63%, respectively, while achieving an 87.5% input data size reduction in 3D point cloud semantic segmentation. | ['Fengbo Ren', 'Suya You', 'Zifan Yu'] | 2023-02-16 | null | null | null | null | ['3d-shape-retrieval'] | ['computer-vision'] | [ 3.57743204e-01 2.98057586e-01 -1.50620684e-01 -1.38410330e-01
-7.04663277e-01 -5.34097850e-01 2.97703505e-01 1.25975952e-01
-4.29104775e-01 1.85047820e-01 -3.64680648e-01 -4.75100696e-01
-2.74447650e-01 -1.02969587e+00 -1.03868020e+00 -6.21058583e-01
-1.98086515e-01 1.58366144e-01 3.71424347e-01 9.56656262e-02
2.35689133e-01 9.39436197e-01 -1.82782161e+00 3.54587078e-01
7.36806154e-01 1.58828652e+00 1.27525434e-01 4.51195419e-01
-2.84397602e-01 8.92893970e-02 -8.71566355e-01 5.51302612e-01
5.06633937e-01 1.00727573e-01 -5.76299012e-01 -2.76189715e-01
9.77247357e-01 -4.42693532e-01 -3.33614796e-01 9.73935425e-01
6.41006887e-01 -5.54663837e-02 7.86628067e-01 -1.11113131e+00
-5.83088279e-01 1.31603003e-01 -7.30161190e-01 3.73586714e-01
-3.31539847e-03 2.78713822e-01 5.88220417e-01 -9.12937105e-01
3.70975703e-01 1.17871785e+00 9.40102935e-01 5.22457778e-01
-1.45895886e+00 -9.18623388e-01 9.79462862e-02 -1.72666103e-01
-1.33061397e+00 1.89655676e-01 1.11296523e+00 -4.61340755e-01
1.55903852e+00 -4.51677525e-03 7.49274790e-01 8.03418458e-01
2.30011523e-01 3.88456821e-01 1.04235446e+00 -2.10202128e-01
1.54181838e-01 -1.83866188e-01 -2.04696860e-02 3.78866643e-01
1.88399553e-01 3.63467425e-01 -5.34724712e-01 2.29764298e-01
1.26334274e+00 -3.51457715e-01 -1.11787524e-02 -2.03218922e-01
-7.52340555e-01 6.83227718e-01 7.79182494e-01 2.14460179e-01
7.43471012e-02 3.06277812e-01 2.43206248e-01 3.13785613e-01
7.42938519e-01 6.37095511e-01 -6.79698050e-01 -7.39400685e-02
-7.16591895e-01 3.61798376e-01 4.27667439e-01 9.09044862e-01
6.60120368e-01 3.15924972e-01 2.15149410e-02 9.64861572e-01
-4.17178264e-03 1.07599270e+00 2.22379696e-02 -1.13971543e+00
3.52289647e-01 7.10748315e-01 -1.98933363e-01 -8.43236804e-01
-8.79804492e-01 -6.33137465e-01 -9.14108992e-01 4.52653766e-01
7.22731054e-01 5.77826314e-02 -1.26064169e+00 1.64414287e+00
1.36987105e-01 -7.07167620e-03 -3.04855496e-01 1.04759479e+00
8.44250143e-01 5.02465844e-01 -1.05183698e-01 -4.87050302e-02
1.20639026e+00 4.54075634e-02 -1.37341365e-01 -2.21354529e-01
4.24909443e-01 -7.65684366e-01 1.40001178e+00 3.72149289e-01
-8.83592188e-01 -1.01739204e+00 -1.28179431e+00 -1.32149205e-01
-1.58962935e-01 -1.56449318e-01 4.02989626e-01 7.42732823e-01
-8.17801654e-01 6.95432007e-01 -7.44384646e-01 -1.97536096e-01
9.77218568e-01 6.51161075e-01 5.79988360e-02 7.48583004e-02
-1.03357685e+00 6.01740420e-01 2.79694706e-01 -5.22554994e-01
-4.93340015e-01 -1.69913125e+00 -5.29162824e-01 3.89501713e-02
1.23402670e-01 -3.52848172e-01 1.26648831e+00 -5.25361419e-01
-1.29031515e+00 1.09477413e+00 2.22344115e-01 -6.68543160e-01
2.27818444e-01 -1.55419216e-01 -3.10730278e-01 4.81751740e-01
3.00628003e-02 1.11875367e+00 9.80850220e-01 -1.18048465e+00
-6.94696128e-01 -5.49520254e-01 -9.28941462e-03 2.96936929e-02
-2.96784937e-01 -8.66644800e-01 -1.35515928e-01 -6.22894049e-01
5.81125021e-01 -9.49848354e-01 1.52114496e-01 2.24333242e-01
9.94548500e-02 -3.10184568e-01 1.16411448e+00 -1.30735636e-01
6.38596296e-01 -2.29556704e+00 -5.86913466e-01 1.24347687e-01
3.68243605e-01 2.68998563e-01 -1.07325748e-01 -3.55840594e-01
-1.70844406e-01 3.51004899e-01 -8.45528394e-02 2.66844511e-01
-3.27028632e-01 -1.24984039e-02 -2.85592675e-01 4.81299877e-01
7.07124114e-01 7.68916607e-01 -7.45312870e-01 -1.83296017e-02
6.23027682e-01 6.43919289e-01 -7.97656476e-01 -2.71277815e-01
-2.35768899e-01 4.91296411e-01 -3.11771005e-01 7.65219510e-01
1.11337399e+00 -2.67468363e-01 -2.67241985e-01 -6.07351124e-01
-1.89800411e-01 2.59485334e-01 -8.05221379e-01 1.77081835e+00
-7.30062246e-01 9.34941888e-01 -6.69517517e-02 -1.01607347e+00
1.24816740e+00 -1.20575868e-01 9.01278138e-01 -1.42679846e+00
2.03449473e-01 4.88113582e-01 1.85760453e-01 -2.79670715e-01
1.63628802e-01 -2.51497090e-01 -8.44682753e-02 2.19834223e-01
2.34001666e-01 -6.35968029e-01 -3.87258172e-01 -3.32446635e-01
8.73273909e-01 -1.29119650e-01 -2.31187567e-01 -4.99589801e-01
1.51871210e-02 3.10207069e-01 1.88031554e-01 8.90131176e-01
-2.18364567e-01 6.01355016e-01 4.55186009e-01 -5.29654860e-01
-1.33024848e+00 -1.25348949e+00 -7.23318636e-01 7.35795081e-01
4.94617671e-01 1.15568653e-01 -6.64462984e-01 -2.75260299e-01
6.09023750e-01 6.85932279e-01 -5.07601082e-01 -5.95013380e-01
-7.85601318e-01 -6.35477483e-01 7.47245193e-01 5.91822147e-01
8.40504527e-01 -5.58027565e-01 -1.03799355e+00 4.61675115e-02
1.75768584e-01 -1.27915370e+00 -1.11254573e-01 5.85890412e-01
-1.39595366e+00 -9.14365351e-01 -5.51828504e-01 -6.51588857e-01
4.51837331e-01 2.41932094e-01 1.00916076e+00 -3.28852057e-01
-4.19874489e-01 2.53287882e-01 -4.99848053e-02 -7.50410676e-01
-2.06847936e-02 3.65716219e-01 -2.48858072e-02 -4.94167179e-01
6.50340497e-01 -8.19321632e-01 -5.47207832e-01 3.18795532e-01
-5.53005040e-01 -1.99867889e-01 5.80284357e-01 8.09762955e-01
7.79068589e-01 3.29636899e-03 6.61571681e-01 -4.86201406e-01
2.98090577e-01 -8.53243694e-02 -6.27430499e-01 -6.49877608e-01
-7.32722461e-01 -9.05983672e-02 5.60535729e-01 -7.09217787e-01
-7.05261230e-01 6.98977476e-03 -1.47138506e-01 -1.07064128e+00
-3.14099371e-01 1.04291253e-02 1.05371907e-01 -2.56194711e-01
1.23089862e+00 6.37176335e-02 3.26146573e-01 -4.00532603e-01
9.12016034e-02 3.64998877e-01 5.54273307e-01 -6.09196544e-01
7.96868205e-01 6.89867318e-01 3.32092404e-01 -1.15826154e+00
-8.03188026e-01 -4.35882837e-01 -6.12437308e-01 -3.51281673e-01
7.96516299e-01 -9.86203909e-01 -9.40908134e-01 5.74282706e-01
-9.32301104e-01 -4.56715107e-01 -6.09357357e-01 4.30612713e-01
-4.91894811e-01 -2.97761410e-01 -3.66788805e-01 -7.01417506e-01
-1.37536243e-01 -9.81199026e-01 1.23481941e+00 1.14574298e-01
-1.72333986e-01 -5.95931828e-01 -7.45914400e-01 1.26351113e-03
3.74312788e-01 3.50824505e-01 1.34832680e+00 -3.37847799e-01
-6.58579171e-01 1.03070267e-01 -4.94794607e-01 3.19861382e-01
8.67464542e-02 -2.85698742e-01 -1.33776641e+00 -9.41588432e-02
1.24211162e-01 -2.65952125e-02 9.04415369e-01 9.83219087e-01
1.41806495e+00 2.08919153e-01 -2.64739990e-01 7.99780667e-01
1.43265569e+00 3.28887194e-01 2.91920274e-01 2.00853154e-01
5.63110173e-01 2.97593355e-01 4.95901823e-01 2.66962320e-01
-2.70932972e-01 5.31207860e-01 6.58311069e-01 -1.94066837e-01
-6.68244004e-01 -3.18380535e-01 -4.01062369e-01 2.12117776e-01
3.51111032e-02 -2.74742469e-02 -1.17683077e+00 5.99770546e-01
-1.32555878e+00 -6.05802000e-01 -1.52646318e-01 2.00305939e+00
5.58455646e-01 6.43282235e-01 7.89798275e-02 3.97999048e-01
5.59417903e-01 5.04034273e-02 -1.01734769e+00 -3.80869716e-01
-1.31789520e-01 5.33947885e-01 9.55185711e-01 3.33560169e-01
-1.23788679e+00 6.73942506e-01 5.59359074e+00 1.04946291e+00
-1.56100619e+00 -1.13056228e-01 8.27742577e-01 -4.83554423e-01
-2.14221537e-01 -5.99408746e-01 -8.23273599e-01 3.13209563e-01
6.43063128e-01 9.14392844e-02 1.90425009e-01 8.85912538e-01
1.48091003e-01 -1.95841879e-01 -1.05804920e+00 1.37066281e+00
-3.16391617e-01 -1.58286107e+00 8.38780552e-02 2.85396248e-01
7.74644434e-01 6.38730675e-02 3.56955916e-01 1.80204183e-01
-4.57680672e-01 -1.36426830e+00 8.50739598e-01 1.69153690e-01
1.38016713e+00 -1.14875627e+00 5.59862494e-01 3.95757228e-01
-1.12301028e+00 -1.47551581e-01 -3.75605255e-01 -1.91981092e-01
-2.12726146e-01 1.08082426e+00 -9.12834644e-01 1.76767334e-01
1.39363003e+00 6.21943891e-01 9.53186210e-03 6.25874996e-01
1.93564996e-01 5.84691465e-01 -8.43183815e-01 -1.58191532e-01
2.61426568e-01 8.41210932e-02 6.52977228e-01 9.83273149e-01
3.18595350e-01 2.85975158e-01 7.63635337e-03 1.14451528e+00
-2.03418881e-01 -4.32613790e-01 -6.33446753e-01 4.03981894e-01
7.77072191e-01 6.17483318e-01 -7.58990288e-01 1.82837024e-02
-2.11743489e-01 4.02228385e-01 -1.17004171e-01 3.03635091e-01
-6.72360718e-01 -4.97544706e-01 8.65174413e-01 5.44302583e-01
3.89385253e-01 -3.77773345e-01 -1.22622502e+00 -2.46184751e-01
9.12961438e-02 -3.21003079e-01 -6.77282289e-02 -5.44367969e-01
-1.18052340e+00 2.15798765e-01 9.40897390e-02 -1.34795260e+00
1.35604098e-01 -9.86705601e-01 -8.88907760e-02 1.07507372e+00
-1.52351046e+00 -9.07904208e-01 -3.56846720e-01 4.14710611e-01
3.83962393e-01 -2.51525529e-02 5.09596288e-01 2.84308463e-01
1.54750019e-01 6.48630321e-01 -2.02413857e-01 1.49618119e-01
4.32706863e-01 -1.12918842e+00 3.93097043e-01 3.46453935e-01
-3.21393639e-01 3.18620414e-01 4.65019196e-01 -4.92693454e-01
-1.37036574e+00 -1.29359305e+00 5.29935420e-01 -3.06950748e-01
2.33928174e-01 -5.66778958e-01 -9.74161506e-01 -4.96258549e-02
-6.38826370e-01 2.69061953e-01 2.32588693e-01 1.83902845e-01
-4.97915179e-01 -4.03440505e-01 -1.44913864e+00 3.78162146e-01
1.46554577e+00 -7.65264034e-01 -4.78922039e-01 1.27794653e-01
9.70711887e-01 -7.80893624e-01 -1.07725072e+00 1.00989318e+00
5.61278224e-01 -9.09873903e-01 1.38213944e+00 -3.01496685e-02
4.71571624e-01 -2.59177774e-01 -2.22380385e-01 -1.05509531e+00
-3.16164792e-01 9.62702781e-02 -1.23910584e-01 6.92495108e-01
3.58233929e-01 -5.68530738e-01 1.04626131e+00 -6.63439780e-02
-4.96885836e-01 -8.41178477e-01 -1.40641916e+00 -1.01395142e+00
4.93675262e-01 -1.01804376e+00 3.97360921e-01 7.00255215e-01
-5.29379129e-01 1.86998680e-01 4.38177288e-01 4.32328075e-01
5.93512118e-01 1.97774470e-01 3.40317786e-01 -1.40191627e+00
1.69273436e-01 -7.68968999e-01 -4.36719745e-01 -1.36347640e+00
-1.62501186e-01 -9.38658237e-01 -2.06092343e-01 -1.19246912e+00
-4.16636139e-01 -6.68643892e-01 -1.61583006e-01 3.77251059e-01
3.35084081e-01 5.02834678e-01 1.80171326e-01 1.87248558e-01
3.26150596e-01 2.13844165e-01 1.61316204e+00 -3.77657920e-01
-4.65969741e-01 -3.40828262e-02 -4.27164674e-01 6.05286062e-01
7.48301685e-01 -3.13679844e-01 -7.09980607e-01 -6.05438292e-01
3.49755548e-02 -2.07149357e-01 7.39293993e-01 -1.26624954e+00
3.44716161e-02 3.21640931e-02 1.00891650e+00 -9.57955599e-01
4.22667652e-01 -7.22633064e-01 -2.23877177e-01 6.35646403e-01
-1.40319848e-02 -6.78530872e-01 1.03335011e+00 3.42411220e-01
-7.47071728e-02 2.48707175e-01 1.18820632e+00 -6.80910721e-02
-7.94764638e-01 2.04151094e-01 -3.25319707e-01 2.48924673e-01
7.58993387e-01 -9.17073190e-01 -2.51294553e-01 7.16210008e-02
-6.70836866e-01 -1.40122667e-01 2.89651781e-01 4.12022561e-01
4.99412328e-01 -1.33545291e+00 -3.99978101e-01 5.58117807e-01
6.62093703e-03 5.18312752e-01 5.38776159e-01 4.03248042e-01
-3.61435056e-01 7.11120248e-01 -3.86925310e-01 -1.50808394e+00
-7.80370712e-01 1.20515190e-01 7.28252828e-01 2.86846578e-01
-7.15531707e-01 1.26732075e+00 2.04491153e-01 -4.87293452e-01
1.69678956e-01 -9.41104650e-01 1.09924495e-01 6.36673048e-02
3.89601104e-02 2.83270240e-01 5.32299936e-01 1.25900973e-02
-5.13396800e-01 1.09131444e+00 2.64144361e-01 2.69535005e-01
1.33177459e+00 3.98339152e-01 4.46794659e-01 4.97598916e-01
1.52307284e+00 -3.35821718e-01 -1.73943543e+00 -5.95762916e-02
-3.10686469e-01 -3.33130181e-01 3.89177114e-01 -9.13161099e-01
-1.08483469e+00 1.00716162e+00 1.05781150e+00 3.10737908e-01
1.40192413e+00 2.44906202e-01 7.03201830e-01 2.26075575e-02
2.33678579e-01 -1.26404786e+00 1.92395255e-01 8.45800042e-01
7.95368016e-01 -1.01361203e+00 -9.03024748e-02 -7.37578273e-01
1.42846435e-01 1.16276860e+00 7.86318481e-01 -4.56282556e-01
1.00906348e+00 5.03448546e-01 -6.07732907e-02 -3.70848954e-01
-3.63924026e-01 2.68842634e-02 4.64379370e-01 1.00521445e+00
3.77030820e-02 2.17809621e-02 3.36458564e-01 5.62873244e-01
-5.23750901e-01 -1.91057727e-01 5.77625632e-02 6.72406852e-01
-5.57020426e-01 -3.74832332e-01 -3.92926574e-01 7.57524073e-01
-4.72154245e-02 1.04604267e-01 -7.87329301e-02 1.06671619e+00
5.68713188e-01 7.16589749e-01 8.52861166e-01 -4.48344260e-01
8.33918691e-01 3.27582657e-02 7.69589901e-01 -5.42652726e-01
-3.03169221e-01 2.64078408e-01 -3.09287488e-01 -6.22124493e-01
-2.98075855e-01 -3.73026729e-01 -1.55449569e+00 -3.02216202e-01
-1.51945755e-01 -5.10115981e-01 6.56553686e-01 6.66756690e-01
4.05070513e-01 9.42303598e-01 5.19521773e-01 -9.17166889e-01
-3.02494019e-01 -7.17223763e-01 -5.52880347e-01 3.82656120e-02
5.43115973e-01 -8.46413314e-01 -5.66700995e-01 -1.51219189e-01] | [8.011353492736816, -3.426616668701172] |
879c21fb-352b-4b63-ae47-5300e1a1e4f8 | post-processing-recommender-systems-with | 2204.11241 | null | https://arxiv.org/abs/2204.11241v1 | https://arxiv.org/pdf/2204.11241v1.pdf | Post Processing Recommender Systems with Knowledge Graphs for Recency, Popularity, and Diversity of Explanations | Existing explainable recommender systems have mainly modeled relationships between recommended and already experienced products, and shaped explanation types accordingly (e.g., movie "x" starred by actress "y" recommended to a user because that user watched other movies with "y" as an actress). However, none of these systems has investigated the extent to which properties of a single explanation (e.g., the recency of interaction with that actress) and of a group of explanations for a recommended list (e.g., the diversity of the explanation types) can influence the perceived explaination quality. In this paper, we conceptualized three novel properties that model the quality of the explanations (linking interaction recency, shared entity popularity, and explanation type diversity) and proposed re-ranking approaches able to optimize for these properties. Experiments on two public data sets showed that our approaches can increase explanation quality according to the proposed properties, fairly across demographic groups, while preserving recommendation utility. The source code and data are available at https://github.com/giacoballoccu/explanation-quality-recsys. | ['Mirko Marras', 'Gianni Fenu', 'Ludovico Boratto', 'Giacomo Balloccu'] | 2022-04-24 | null | null | null | null | ['explainable-models', 'movie-recommendation'] | ['computer-vision', 'miscellaneous'] | [-3.69898707e-01 5.12918651e-01 -7.72878885e-01 -6.79721057e-01
2.15610087e-01 -4.26345319e-01 6.78420722e-01 3.10055673e-01
9.05559808e-02 4.06421334e-01 5.91413200e-01 -3.84135276e-01
-7.25864470e-01 -8.09281886e-01 -7.33767271e-01 -2.50301093e-01
-1.74874678e-01 5.28312147e-01 9.16789025e-02 -4.37680244e-01
5.25535405e-01 1.13849923e-01 -1.86418176e+00 4.01589632e-01
1.31931901e+00 5.22542238e-01 2.72886455e-01 4.17937517e-01
-1.01905398e-01 3.94631565e-01 -2.36839861e-01 -9.07117486e-01
1.90085873e-01 -5.14041603e-01 -5.73411524e-01 -3.64460088e-02
4.56423730e-01 -3.72162521e-01 -4.14347023e-01 7.66449928e-01
-1.08974993e-01 4.72666383e-01 7.50709713e-01 -1.43669248e+00
-1.37249351e+00 1.45521140e+00 -5.86848632e-02 7.63207972e-02
4.56328779e-01 -3.59498113e-01 1.59663820e+00 -7.77835071e-01
7.51198649e-01 1.00584829e+00 3.85041773e-01 5.00432491e-01
-1.12814867e+00 -3.47319990e-01 5.87057412e-01 3.51506710e-01
-9.76983011e-01 -1.43057052e-02 3.41626585e-01 -3.95027071e-01
6.48691714e-01 6.83493495e-01 8.29332530e-01 8.25711906e-01
2.53712207e-01 4.02038783e-01 6.32265508e-01 -1.66466951e-01
1.91672117e-01 7.85170674e-01 7.00179219e-01 4.05144423e-01
5.86308181e-01 7.30061829e-02 -6.17023170e-01 -3.12030017e-01
2.50133514e-01 4.62578893e-01 -3.01430881e-01 -3.67490947e-01
-5.69500327e-01 1.06282294e+00 5.55817962e-01 4.47827369e-01
-3.64200979e-01 -2.63406873e-01 -1.91139013e-01 4.36639220e-01
6.66445017e-01 9.91180003e-01 -7.32862175e-01 -1.05968326e-01
-4.99172062e-01 2.84587055e-01 1.03493667e+00 8.97339761e-01
8.72455359e-01 -3.26117843e-01 -1.12455107e-01 8.14734817e-01
4.42313015e-01 2.11596414e-01 4.34528887e-01 -8.36174309e-01
-1.14202991e-01 7.60061085e-01 1.75071970e-01 -1.16676366e+00
-4.07340974e-01 -5.40528238e-01 -2.64269710e-01 -2.12612167e-01
5.32963693e-01 -1.53871700e-01 -3.86790991e-01 1.65989327e+00
1.27006173e-01 -9.48344078e-03 -2.86958337e-01 1.07906079e+00
9.38992679e-01 4.25329357e-01 4.10826541e-02 -2.25938588e-01
1.27760661e+00 -1.03392863e+00 -5.11495948e-01 -1.44331321e-01
5.12437940e-01 -8.50934207e-01 1.26146472e+00 4.69620585e-01
-1.11296535e+00 -7.82112539e-01 -7.26709783e-01 1.76568761e-01
-4.60253060e-01 1.91348553e-01 9.91882980e-01 6.68985844e-01
-8.24812472e-01 1.13058531e+00 -2.91607857e-01 -7.08332181e-01
-1.69201270e-01 5.39098084e-01 -9.66281164e-03 6.16191775e-02
-1.31772828e+00 6.63312435e-01 -2.63719827e-01 -4.49547023e-01
-3.76096606e-01 -7.64452279e-01 -2.56048739e-01 5.80762684e-01
3.21862280e-01 -9.26701069e-01 8.46219182e-01 -1.36467755e+00
-1.26388574e+00 1.13012917e-01 -3.56459506e-02 -3.80141228e-01
8.86745285e-03 -5.11375070e-01 -8.33419204e-01 -3.69452626e-01
-5.65887429e-02 4.41622198e-01 6.17581308e-01 -1.18563783e+00
-6.02945328e-01 -3.97691131e-01 7.91296542e-01 2.92056441e-01
-7.17400432e-01 -2.19648421e-01 -2.81124085e-01 -3.17155838e-01
4.54537384e-02 -1.18625665e+00 -2.01295823e-01 -4.68836814e-01
-2.40909532e-01 -4.54553783e-01 8.54869112e-02 -2.65417904e-01
1.54822528e+00 -2.13212466e+00 2.46832728e-01 4.52338457e-01
3.38412970e-01 8.55111331e-03 -8.12104493e-02 5.95917761e-01
8.46098214e-02 4.98825103e-01 7.97255337e-01 -1.03333481e-01
2.02334836e-01 1.54444456e-01 3.12876794e-03 1.71304256e-01
-6.23323500e-01 5.37600398e-01 -8.58023405e-01 1.19616956e-01
-8.52244124e-02 6.54621899e-01 -1.23717976e+00 2.97822893e-01
-2.73847669e-01 1.83181152e-01 -5.68913817e-01 1.83409244e-01
3.71766239e-01 -6.11633241e-01 3.91982675e-01 -1.88520998e-01
-2.16045424e-01 7.08523095e-01 -1.05087221e+00 9.46592808e-01
-3.70237738e-01 2.37024948e-01 -4.89318699e-01 -1.46616906e-01
9.72160459e-01 -1.26596084e-02 5.12378693e-01 -4.17356282e-01
6.96489513e-02 1.34590715e-01 3.92324090e-01 -3.84055138e-01
9.46131170e-01 3.90497983e-01 2.29648635e-01 4.17099774e-01
-1.41790673e-01 4.04623210e-01 2.85863191e-01 7.89353788e-01
9.32705343e-01 -1.22143053e-01 3.75627309e-01 -4.37067091e-01
4.15195078e-01 -1.40146300e-01 3.99216861e-01 9.59821224e-01
2.48963058e-01 4.36370641e-01 3.64632457e-01 -4.45301324e-01
-8.84990036e-01 -7.44246840e-01 7.01703206e-02 1.43099809e+00
5.19580364e-01 -1.02897167e+00 -5.18624723e-01 -7.45688260e-01
3.84966910e-01 1.29317343e+00 -8.54054987e-01 -4.25274402e-01
-3.52988727e-02 -1.68547168e-01 -2.27722719e-01 5.61999567e-02
4.34035026e-02 -7.61298239e-01 -3.50610279e-02 -7.38833565e-03
-1.78321376e-01 -3.61072809e-01 -7.99937487e-01 -1.05857477e-01
-7.86794841e-01 -8.39298964e-01 -1.58304676e-01 8.58938694e-02
7.74259269e-01 7.18561172e-01 1.46407938e+00 9.33679760e-01
4.18260366e-01 5.35759151e-01 -8.80443692e-01 -1.04955330e-01
-1.82918403e-02 2.85757929e-01 3.83269995e-01 1.79815799e-01
4.41097885e-01 -6.90171897e-01 -9.42139447e-01 8.20662498e-01
-6.84118569e-01 -1.59106135e-01 4.66279656e-01 4.77747649e-01
2.38787904e-01 -8.54136869e-02 4.50680017e-01 -1.58903885e+00
6.83451533e-01 -1.06067741e+00 -5.08753657e-02 1.75426990e-01
-1.50521326e+00 -2.11365428e-02 6.75639093e-01 -5.64787626e-01
-1.19538927e+00 -2.01545879e-01 -2.07747400e-01 7.96627477e-02
-1.94053337e-01 5.26411593e-01 -5.14393635e-02 2.57292598e-01
7.75495410e-01 -2.51252353e-01 -3.84302676e-01 -5.64476907e-01
6.68357611e-01 3.68210286e-01 -1.87339902e-01 -2.94882208e-01
6.24347687e-01 2.86261737e-01 -5.13914585e-01 -3.37059170e-01
-8.70965242e-01 -5.10122418e-01 -2.66370267e-01 -3.84300202e-01
5.98724365e-01 -4.74870265e-01 -7.55624712e-01 -3.99211287e-01
-8.53171110e-01 7.72752836e-02 -3.05489302e-01 8.64577234e-01
-1.66199684e-01 9.61226150e-02 -2.48123243e-01 -5.01060069e-01
-3.94055583e-02 -7.93580949e-01 2.92004198e-01 5.57521999e-01
-7.06909120e-01 -9.87759233e-01 5.13481535e-02 5.09115517e-01
6.61811471e-01 -4.60764647e-01 8.77372324e-01 -1.12788343e+00
-6.51837289e-01 8.87840986e-03 6.53682873e-02 -1.44128025e-01
1.82093512e-02 2.26152465e-01 -4.07880932e-01 6.58582151e-02
-3.79170060e-01 4.76481080e-01 6.19177401e-01 7.64786720e-01
8.58029783e-01 -7.99479783e-01 -3.74025404e-01 3.45497817e-01
1.14097071e+00 5.92993014e-02 6.22426450e-01 1.73110828e-01
4.39040512e-01 5.76550543e-01 9.00169194e-01 7.78761566e-01
5.01946986e-01 1.05737054e+00 5.10514081e-01 -5.68129234e-02
-1.41072348e-01 -4.68358606e-01 3.83483678e-01 5.85093498e-01
-5.21629930e-01 -5.37031353e-01 -1.33420959e-01 2.98008442e-01
-1.97598815e+00 -9.67565656e-01 -6.67434752e-01 2.15323710e+00
1.60029843e-01 1.56316951e-01 3.77620280e-01 -4.65861946e-01
4.60715383e-01 -1.34968191e-01 -6.31060004e-01 -7.38865018e-01
3.63186374e-02 -4.48208928e-01 4.79378074e-01 8.57395053e-01
-1.53709084e-01 7.64544666e-01 5.18066597e+00 4.95208830e-01
-4.76002276e-01 5.95039763e-02 5.51441967e-01 -1.89006925e-01
-1.20447767e+00 3.80434543e-01 -9.95635331e-01 5.77166140e-01
1.00807869e+00 -5.71852803e-01 7.25805283e-01 1.10869956e+00
3.67879808e-01 9.49758962e-02 -1.09740186e+00 2.30835676e-01
1.51633933e-01 -1.17778611e+00 1.65748417e-01 3.84532064e-01
8.14653099e-01 -2.66689241e-01 3.44845831e-01 4.20239449e-01
5.10169506e-01 -5.43908536e-01 5.18358886e-01 6.59027278e-01
-9.24369991e-02 -5.27293801e-01 5.63937187e-01 2.54944772e-01
-9.45946932e-01 -2.92176902e-01 -5.56458235e-01 -2.39344269e-01
1.12477809e-01 7.51176536e-01 -6.64267421e-01 4.88510579e-01
6.68022633e-01 8.23807299e-01 -6.78708613e-01 7.25184798e-01
-3.71693045e-01 7.52154887e-01 1.83718890e-01 -2.81294346e-01
-6.94435462e-02 -3.84978026e-01 6.89446807e-01 6.78156257e-01
6.68201089e-01 5.45866907e-01 -1.38254672e-01 7.78435886e-01
-5.68174906e-02 5.89829326e-01 -4.99712557e-01 1.38985306e-01
7.19633043e-01 1.36488354e+00 -5.82633853e-01 -2.42196292e-01
-5.01112580e-01 7.50056386e-01 2.65818059e-01 1.55938983e-01
-1.07744527e+00 2.26898968e-01 9.93504822e-01 8.35892856e-01
2.22998857e-01 2.56365865e-01 -3.43481630e-01 -1.25236273e+00
-5.63234627e-01 -8.78966093e-01 5.41735411e-01 -9.05445635e-01
-1.29388463e+00 8.71624231e-01 -1.31406322e-01 -1.04849219e+00
-1.17369136e-02 -5.60022965e-02 -6.15697980e-01 4.80285853e-01
-1.04176152e+00 -8.86323869e-01 -2.22798049e-01 4.73701030e-01
5.62882960e-01 -2.27749467e-01 5.23306906e-01 4.05538499e-01
-1.87621906e-01 5.88636220e-01 5.99967316e-02 -8.80765021e-01
8.26596498e-01 -1.31834149e+00 1.91547945e-01 5.71380496e-01
5.75986266e-01 1.18182898e+00 1.24548542e+00 -7.31245220e-01
-1.20486009e+00 -6.49172902e-01 1.43686831e+00 -5.24010777e-01
4.15913135e-01 -1.76057667e-02 -8.26062858e-01 6.41888380e-01
3.55367869e-01 -7.54436612e-01 1.06288755e+00 1.09300411e+00
-4.04643178e-01 -9.64282230e-02 -1.00023568e+00 6.83768988e-01
1.30129564e+00 1.11410394e-01 -1.19314611e-01 4.67138410e-01
6.88132405e-01 9.71211195e-02 -1.05475760e+00 4.97422591e-02
9.05092359e-01 -1.44469154e+00 9.56067860e-01 -6.89160943e-01
7.65257239e-01 -6.14168905e-02 -6.27728403e-02 -1.61082292e+00
-1.02056670e+00 -3.72880071e-01 -2.39170313e-01 1.43663204e+00
9.02886093e-01 -6.69379115e-01 7.85788834e-01 1.15159786e+00
-2.07810417e-01 -8.20672333e-01 -1.94597751e-01 -5.37284374e-01
-2.53537983e-01 2.30305195e-01 1.04713309e+00 8.14245522e-01
4.36721772e-01 4.06771541e-01 -8.36359799e-01 3.91511559e-01
1.45706952e-01 3.16491127e-01 8.25263381e-01 -1.42559063e+00
-6.94568634e-01 -3.49365145e-01 1.55293986e-01 -1.20238733e+00
-1.15698159e-01 -9.95417416e-01 -5.18920481e-01 -1.41781425e+00
3.16578716e-01 -5.26494026e-01 -4.42950577e-01 1.64534986e-01
-3.94012868e-01 -7.24001601e-02 2.59097129e-01 5.71147382e-01
-8.85085344e-01 1.88338459e-01 9.81244802e-01 2.90022701e-01
-5.03137410e-01 3.91486049e-01 -1.26929724e+00 6.90412939e-01
9.46080744e-01 -6.55420303e-01 -7.25339949e-01 -3.26843649e-01
7.99898326e-01 1.42529756e-01 -1.09143004e-01 -5.15054584e-01
1.67585939e-01 -3.49072129e-01 -5.85670471e-02 -2.79583931e-01
2.25072712e-01 -1.00944936e+00 8.13246012e-01 2.82931060e-01
-8.31569552e-01 1.73115805e-01 -3.74840915e-01 5.68527520e-01
1.62617356e-01 -3.01787525e-01 2.16601312e-01 2.45524123e-01
-4.11513120e-01 3.43696743e-01 -2.94066936e-01 -4.40913707e-01
6.66377783e-01 -2.70932466e-01 -4.80257750e-01 -8.32427680e-01
-8.50503385e-01 -5.65860756e-02 5.72431624e-01 8.51534784e-01
4.27315563e-01 -1.24205053e+00 -6.76257193e-01 2.92398948e-02
2.18799248e-01 -1.22164619e+00 4.26058173e-01 5.85890114e-01
3.11059743e-01 3.88101339e-01 -3.18090290e-01 3.14077586e-02
-1.64610541e+00 5.05049765e-01 1.26340976e-02 -2.88472623e-01
-1.85788110e-01 8.13984036e-01 1.85305893e-01 -4.79481071e-01
-7.53533095e-02 -1.03157781e-01 -6.41363859e-01 -2.65783276e-02
2.71797210e-01 6.57208145e-01 -2.02373520e-01 -4.30221558e-01
-1.42165452e-01 1.96744844e-01 -2.69079268e-01 3.33376497e-01
1.35619473e+00 -6.15349293e-01 2.45359510e-01 1.70654178e-01
7.49826908e-01 4.87468749e-01 -8.48408222e-01 -2.33648755e-02
-2.10604221e-01 -1.03683400e+00 -2.30979115e-01 -9.49274242e-01
-1.32726288e+00 2.06299692e-01 1.72389701e-01 9.41780269e-01
8.11935127e-01 3.37113619e-01 5.26362836e-01 6.08395450e-02
2.82159030e-01 -7.13359177e-01 2.36831922e-02 1.79343864e-01
6.55392528e-01 -8.72538567e-01 2.99392104e-01 -7.29241967e-01
-1.01797616e+00 7.80258179e-01 7.61057734e-01 -1.06213233e-02
1.02713966e+00 -4.42046523e-01 -2.75019318e-01 -4.20101196e-01
-1.08319843e+00 -6.54410124e-02 6.54323757e-01 2.14214012e-01
9.37470257e-01 4.48005199e-01 -1.28685498e+00 1.14730251e+00
-3.62148762e-01 -2.20331788e-01 6.81363165e-01 -1.40564471e-01
-5.81167400e-01 -1.32781351e+00 7.55478293e-02 7.73469985e-01
-3.84725183e-01 -2.45055854e-01 -7.27599680e-01 6.03641748e-01
5.39569557e-01 1.22334182e+00 -1.76837355e-01 -7.62597740e-01
3.99508089e-01 -3.61802757e-01 -3.76947187e-02 -7.11329639e-01
-1.04581726e+00 -1.29411519e-01 4.09421504e-01 -5.03785849e-01
-2.14909032e-01 -8.49948704e-01 -1.02252030e+00 -8.56316090e-01
-7.76608229e-01 6.50852263e-01 5.14928579e-01 6.53468847e-01
7.60432184e-01 2.91201651e-01 7.42177010e-01 -2.29210287e-01
-3.35631490e-01 -7.38929212e-01 -9.84900475e-01 9.20512021e-01
-2.74823248e-01 -8.35512996e-01 -6.65795982e-01 5.40223978e-02] | [9.79137134552002, 5.750641822814941] |
a7f934f0-2d8e-40c4-b02d-824aed47d45d | 190408494 | 1904.08494 | null | https://arxiv.org/abs/1904.08494v2 | https://arxiv.org/pdf/1904.08494v2.pdf | Learning 2D to 3D Lifting for Object Detection in 3D for Autonomous Vehicles | We address the problem of 3D object detection from 2D monocular images in autonomous driving scenarios. We propose to lift the 2D images to 3D representations using learned neural networks and leverage existing networks working directly on 3D data to perform 3D object detection and localization. We show that, with carefully designed training mechanism and automatically selected minimally noisy data, such a method is not only feasible, but gives higher results than many methods working on actual 3D inputs acquired from physical sensors. On the challenging KITTI benchmark, we show that our 2D to 3D lifted method outperforms many recent competitive 3D networks while significantly outperforming previous state-of-the-art for 3D detection from monocular images. We also show that a late fusion of the output of the network trained on generated 3D images, with that trained on real 3D images, improves performance. We find the results very interesting and argue that such a method could serve as a highly reliable backup in case of malfunction of expensive 3D sensors, if not potentially making them redundant, at least in the case of low human injury risk autonomous navigation scenarios like warehouse automation. | ['Gaurav Sharma', 'Frederic Jurie', 'Siddharth Srivastava'] | 2019-03-27 | null | null | null | null | ['monocular-3d-object-localization', '3d-object-detection-from-monocular-images'] | ['computer-vision', 'computer-vision'] | [ 2.43812293e-01 2.79535472e-01 2.23832294e-01 -2.84903854e-01
-4.43274587e-01 -6.30098403e-01 4.95900661e-01 -2.05399513e-01
-6.40037358e-01 4.82801646e-01 -2.41949752e-01 -6.34323120e-01
-9.64496098e-03 -5.10454953e-01 -1.18371713e+00 -4.74209547e-01
-2.45721024e-02 5.81029713e-01 5.13922095e-01 -3.65310133e-01
2.56804854e-01 1.03523731e+00 -1.95027828e+00 -2.51984969e-02
2.91144341e-01 1.39355707e+00 4.60150212e-01 7.53183484e-01
2.24557683e-01 3.09962839e-01 -4.59069252e-01 1.21791020e-01
1.08033800e+00 4.88377437e-02 -2.59200513e-01 3.21909338e-01
6.53807521e-01 -5.90756893e-01 -1.78616822e-01 7.04683781e-01
5.71894705e-01 -2.34852329e-01 4.99780059e-01 -1.18914938e+00
-2.97405064e-01 -3.14241741e-03 -5.52750230e-01 7.32494816e-02
4.77228731e-01 5.17163694e-01 4.22298133e-01 -8.80801499e-01
6.48105204e-01 1.18138433e+00 7.60221004e-01 3.19769770e-01
-1.12165499e+00 -5.24539411e-01 9.66515690e-02 -5.39061725e-02
-1.01987219e+00 -2.26183072e-01 6.16795659e-01 -3.89993668e-01
1.25200367e+00 -1.11938789e-01 6.43858433e-01 1.05538321e+00
2.78835356e-01 9.00075555e-01 1.04120028e+00 -5.47185123e-01
7.89076984e-02 1.70074239e-01 -4.66827676e-02 8.50236356e-01
7.08831549e-01 5.43350458e-01 -3.41733158e-01 2.52687007e-01
6.99725986e-01 9.43428427e-02 -1.08729310e-01 -1.15067899e+00
-1.28622091e+00 6.35394275e-01 6.34063065e-01 -6.21516630e-02
-6.92577541e-01 3.89702618e-02 2.12202072e-01 3.55607539e-01
3.06769013e-01 5.97003043e-01 -5.85381091e-01 5.23333549e-02
-5.11189759e-01 3.28800946e-01 6.16264403e-01 1.25667036e+00
6.65768325e-01 1.88368075e-02 2.72296458e-01 4.44547683e-01
9.81262475e-02 7.20758796e-01 3.12097430e-01 -1.02166247e+00
5.54557562e-01 8.37081254e-01 5.52811444e-01 -5.92858672e-01
-8.72547925e-01 -4.61996019e-01 -3.34265888e-01 9.29648161e-01
4.05086398e-01 -6.02848493e-02 -1.24271393e+00 1.20253420e+00
3.31245363e-01 -4.69761193e-01 1.49124250e-01 1.13375652e+00
4.99752015e-01 2.73962796e-01 -4.98455733e-01 3.61108392e-01
1.05883777e+00 -5.40260851e-01 -2.18731716e-01 -5.89349985e-01
6.30980551e-01 -5.32068491e-01 7.74457812e-01 7.15124726e-01
-9.31247532e-01 -8.21403623e-01 -1.48702443e+00 2.66949385e-02
-8.15198958e-01 3.43507439e-01 4.28704560e-01 6.80600882e-01
-9.09916043e-01 4.22524482e-01 -6.70336902e-01 -6.76039100e-01
3.28328699e-01 5.19747376e-01 -7.23576307e-01 -1.63875684e-01
-7.97106266e-01 1.55798388e+00 4.97513950e-01 1.93714768e-01
-9.24613595e-01 -2.92657614e-01 -1.05279803e+00 -3.66003454e-01
5.76886237e-01 -6.30555272e-01 1.23000884e+00 -1.56633705e-01
-1.13609397e+00 1.07696903e+00 1.47464752e-01 -9.04344738e-01
7.15255976e-01 -7.06206858e-01 1.31100476e-01 2.48371400e-02
7.35325143e-02 1.10666037e+00 7.95415521e-01 -1.51472580e+00
-9.81976390e-01 -6.60943568e-01 1.74297214e-01 2.48741820e-01
2.32169151e-01 -6.22557044e-01 -4.26986217e-02 1.22558594e-01
4.81028318e-01 -1.03273308e+00 -3.11811060e-01 3.25165659e-01
-3.22512001e-01 -9.53100026e-02 9.69526350e-01 -4.04647380e-01
-3.39287557e-02 -1.98119473e+00 -1.15903996e-01 -7.83820227e-02
-5.38733564e-02 3.82315069e-01 2.11790297e-03 1.81181312e-01
-2.01966763e-02 -3.75361353e-01 -1.14168532e-01 -3.04498792e-01
5.88297211e-02 3.06267619e-01 -2.80357808e-01 6.58277452e-01
6.39789343e-01 9.21197832e-01 -8.01981986e-01 3.01605035e-02
6.31927490e-01 3.46134484e-01 -1.07148893e-01 2.86063820e-01
-4.27048728e-02 3.14261258e-01 -1.62496075e-01 7.67650366e-01
7.44552076e-01 5.53547144e-02 -3.32121342e-01 -1.70192122e-01
-2.20065117e-01 2.94589728e-01 -1.28862107e+00 1.74823141e+00
-4.17019069e-01 6.24099314e-01 1.81571439e-01 -1.12017524e+00
1.31431270e+00 6.22213855e-02 3.12755793e-01 -8.69787574e-01
1.92400113e-01 4.53126311e-01 -1.50943235e-01 -6.09610617e-01
6.41708255e-01 -7.88016096e-02 -2.98805684e-01 3.04823428e-01
9.15950388e-02 -7.55222380e-01 -4.39188071e-02 -1.19623102e-01
1.22158444e+00 7.85281718e-01 2.33923897e-01 6.78830594e-02
3.13119322e-01 3.50923419e-01 5.63083105e-02 1.17087638e+00
-2.80847490e-01 7.18261302e-01 1.91211373e-01 -7.86045611e-01
-1.28948450e+00 -1.07163072e+00 5.19986264e-02 5.71097016e-01
5.50228357e-01 3.07075918e-01 -3.05821508e-01 -8.35672319e-01
4.78652120e-01 7.61305869e-01 -5.58852255e-01 -2.83873260e-01
-6.68399751e-01 -4.97771680e-01 3.29197049e-01 7.28019714e-01
3.98525357e-01 -8.46598029e-01 -1.64776194e+00 2.63708562e-01
3.92981827e-01 -1.41158009e+00 2.75875032e-01 1.04327452e+00
-9.71743762e-01 -1.08924150e+00 -7.92849779e-01 -5.47139108e-01
5.31894147e-01 8.35893452e-01 8.45657587e-01 -2.73279279e-01
-3.43342543e-01 4.18133229e-01 -4.33850855e-01 -9.36908901e-01
-3.41260254e-01 2.20078826e-02 3.26789171e-01 -5.30725658e-01
7.12185979e-01 -2.63831198e-01 -4.90622789e-01 5.53333938e-01
-4.97278422e-01 8.44014436e-02 8.54811132e-01 8.04044664e-01
2.86948442e-01 -1.90305054e-01 4.11034167e-01 -4.14945602e-01
2.92193919e-01 -1.29548267e-01 -1.00967360e+00 -2.87575603e-01
-6.23321533e-01 1.78657845e-01 4.04735357e-01 -2.28391603e-01
-6.73164606e-01 7.53048539e-01 8.92684143e-03 -6.08172834e-01
-6.70953095e-01 -2.03262895e-01 1.66544408e-01 -1.96291491e-01
1.06199181e+00 1.37756526e-01 3.56045276e-01 -5.82005680e-01
2.92805254e-01 6.92358911e-01 6.49001300e-01 -6.70442358e-02
9.92601931e-01 8.74877453e-01 3.14220607e-01 -4.32053834e-01
-7.13224649e-01 -6.12561345e-01 -1.01287234e+00 -1.75222591e-01
7.15277195e-01 -1.05219340e+00 -5.05772173e-01 2.39290670e-01
-1.19868207e+00 -2.72172272e-01 -4.88025397e-01 6.37375295e-01
-8.52652431e-01 5.93376309e-02 -1.71712086e-01 -1.13924563e+00
4.59329709e-02 -1.24935174e+00 1.55244112e+00 2.52987221e-02
-1.23562422e-02 -3.58730853e-01 -3.54838997e-01 2.64919817e-01
2.71094590e-01 4.22143221e-01 4.30774480e-01 -6.06565654e-01
-6.69597268e-01 -8.61895084e-01 -4.22383636e-01 4.72854406e-01
1.43407941e-01 -6.24761522e-01 -1.11865973e+00 -6.59621358e-02
3.62262898e-03 -3.88121635e-01 1.02849579e+00 2.72616804e-01
5.83971500e-01 3.19600284e-01 -5.37161052e-01 3.74619156e-01
1.14706755e+00 2.76894569e-01 1.49012774e-01 5.69466412e-01
5.77811301e-01 7.20120251e-01 8.08255553e-01 1.94742888e-01
2.99973726e-01 6.07031703e-01 9.77200627e-01 -2.41727546e-01
-9.82575119e-02 -1.40254289e-01 3.00290644e-01 -3.45510207e-02
1.14463918e-01 -4.78642657e-02 -8.76005769e-01 6.66288972e-01
-1.82210541e+00 -5.03708959e-01 -1.52066676e-02 2.23283863e+00
1.62224337e-01 8.28285873e-01 3.21157694e-01 3.20029676e-01
5.09220481e-01 -2.61262208e-01 -8.33276272e-01 -3.40243042e-01
-1.16538592e-01 1.05848089e-02 1.16123271e+00 2.40285978e-01
-1.04070663e+00 4.72855747e-01 6.39671469e+00 -4.21546847e-02
-9.17260349e-01 8.85945186e-02 2.19525307e-01 -3.67094725e-01
3.07076842e-01 -2.28155553e-01 -1.04240131e+00 1.39214665e-01
7.36328840e-01 5.18656790e-01 -2.37172432e-02 1.11595452e+00
1.28413960e-01 -5.15133381e-01 -1.40670598e+00 9.96086478e-01
1.52682200e-01 -1.11043656e+00 -2.89059281e-01 2.55675137e-01
3.57268751e-01 3.55073959e-01 -2.14569300e-01 2.64704466e-01
3.44850421e-01 -8.76405597e-01 9.90251124e-01 2.47498527e-01
3.19705576e-01 -5.24455786e-01 9.83204901e-01 9.01336432e-01
-7.94283926e-01 -3.21976483e-01 -4.32817131e-01 -3.54400098e-01
3.74848455e-01 5.34536779e-01 -1.39369655e+00 4.88677531e-01
8.10451984e-01 4.21113580e-01 -4.66593504e-01 1.17460155e+00
-2.71571726e-01 -7.39215016e-02 -7.85312414e-01 -2.08679259e-01
4.69092488e-01 2.73160815e-01 6.89045846e-01 1.01232696e+00
4.43334043e-01 -2.20706344e-01 2.35372297e-02 8.10949564e-01
2.68904179e-01 -4.84777987e-01 -1.50936759e+00 4.82291549e-01
8.38721395e-02 1.11608911e+00 -6.43796980e-01 -2.20644757e-01
-3.73070955e-01 7.42866039e-01 3.51687036e-02 1.02297440e-01
-4.61075068e-01 -5.44287503e-01 3.18063051e-01 2.15623915e-01
7.60657609e-01 -5.71599901e-01 -3.76591474e-01 -7.02771544e-01
4.21239674e-01 -3.14775079e-01 3.94908749e-02 -1.25035048e+00
-1.08364332e+00 5.19960344e-01 2.16428131e-01 -1.44952595e+00
-6.16299987e-01 -1.44707739e+00 -1.12546377e-01 7.68762589e-01
-1.67036414e+00 -8.79518986e-01 -4.09847379e-01 3.44539918e-02
5.00005782e-01 -1.12357549e-01 4.82275456e-01 -3.57671417e-02
9.88557562e-02 1.04919061e-01 -3.03918898e-01 -1.98683858e-01
4.58438158e-01 -1.11840534e+00 7.48478055e-01 7.87275255e-01
8.11284482e-02 1.19895473e-01 7.63388693e-01 -6.09820843e-01
-1.67886591e+00 -7.98938394e-01 4.07585025e-01 -8.95072758e-01
1.22043200e-01 -7.10244298e-01 -5.19106805e-01 5.02886951e-01
7.02991784e-02 -3.06615513e-02 -7.18170702e-02 -2.81554520e-01
-1.13936260e-01 -2.00748265e-01 -1.30286908e+00 2.87455261e-01
1.42625880e+00 -6.94904178e-02 -8.45546722e-01 5.37532628e-01
7.06242204e-01 -6.93498135e-01 -3.99154246e-01 6.86263144e-01
6.45837843e-01 -1.12652385e+00 1.08133483e+00 -3.40195626e-01
5.91861121e-02 -4.51369643e-01 -4.14796650e-01 -1.30945694e+00
1.38741359e-02 -3.77906591e-01 -1.39171779e-01 2.01598853e-01
5.00213981e-01 -7.00182438e-01 1.02142060e+00 2.31012598e-01
-3.25112373e-01 -4.77356672e-01 -1.16938519e+00 -1.14522588e+00
-2.10910529e-01 -5.99209845e-01 4.63680476e-01 1.53559089e-01
-3.53532165e-01 2.96137899e-01 -2.15260476e-01 2.75600523e-01
6.17442012e-01 1.29948914e-01 1.20518100e+00 -1.43806982e+00
2.47472320e-02 -1.82839170e-01 -8.13836873e-01 -1.35165536e+00
-2.22874023e-02 -4.66831207e-01 5.33239901e-01 -1.55404198e+00
-4.22033608e-01 -3.00141364e-01 -6.27027005e-02 4.35157895e-01
2.41929531e-01 4.42061186e-01 2.29724824e-01 -4.58817035e-02
-3.62903535e-01 3.27058077e-01 1.30098462e+00 -4.15632501e-02
-9.05514732e-02 6.46114945e-02 -6.31633341e-01 6.37035787e-01
6.52489483e-01 -3.57369006e-01 -1.96901649e-01 -3.94848138e-01
3.00743338e-02 -2.50875890e-01 7.42637873e-01 -1.34439814e+00
2.24896759e-01 1.87849224e-01 8.40792954e-01 -1.12984967e+00
8.04508984e-01 -1.17137170e+00 -3.61366570e-01 7.22537994e-01
-2.91438750e-03 1.17956392e-01 3.56413662e-01 5.03736377e-01
2.10908934e-01 -2.68530190e-01 4.78835732e-01 -6.09334350e-01
-1.03144574e+00 -1.63160920e-01 -4.55964774e-01 -5.25893033e-01
1.07017291e+00 -7.89201498e-01 -1.53269768e-01 -4.02659029e-01
-5.67058504e-01 2.69331664e-01 4.64231759e-01 7.14455903e-01
6.84841752e-01 -1.23200619e+00 -5.57825267e-01 6.45841956e-01
3.22224051e-01 4.15053129e-01 -4.08281207e-01 5.89135528e-01
-4.16085541e-01 7.28326201e-01 -3.52224529e-01 -1.12733436e+00
-7.89584875e-01 6.76860154e-01 2.69182682e-01 1.71592429e-01
-8.88896704e-01 5.92286587e-01 1.20622881e-01 -8.09555113e-01
4.59059060e-01 -5.96314967e-01 1.91764504e-01 -1.43727556e-01
2.19088435e-01 4.05551977e-02 4.98414546e-01 -4.70730007e-01
-4.58190590e-01 5.87996602e-01 -4.62185927e-02 -1.23670429e-01
1.49653018e+00 -2.23401845e-01 6.91355467e-01 4.42741454e-01
9.49630618e-01 -5.37701786e-01 -1.84558475e+00 -1.42199486e-01
-7.30744377e-02 -3.93046737e-01 -7.93326572e-02 -7.98079848e-01
-7.11094379e-01 9.81513560e-01 1.01653147e+00 3.02834928e-01
8.15665781e-01 1.08638875e-01 5.43698370e-01 9.41884756e-01
7.53784597e-01 -1.00482392e+00 3.75920981e-02 5.73908985e-01
7.99002349e-01 -1.60919905e+00 1.29390821e-01 -8.72483999e-02
-3.87044072e-01 1.13599479e+00 6.78323388e-01 -5.25753677e-01
3.42752725e-01 3.66874278e-01 2.71281987e-01 -2.16356084e-01
-5.12838840e-01 -4.34640676e-01 1.08458437e-01 9.46400762e-01
-1.12303063e-01 -1.07198194e-01 3.62363160e-01 5.00584617e-02
-1.15248621e-01 -2.85144104e-03 5.93418002e-01 1.46420455e+00
-7.38045096e-01 -7.76342869e-01 -6.86450481e-01 4.57844198e-01
1.78768545e-01 3.74624908e-01 -4.50121105e-01 1.25228906e+00
3.88671875e-01 7.42312610e-01 1.34351060e-01 -3.52529913e-01
9.54109430e-01 2.25610822e-01 7.14403510e-01 -5.81905305e-01
-2.79989123e-01 -1.27260402e-01 9.06971321e-02 -6.69606686e-01
-3.93400401e-01 -6.97423577e-01 -9.09566224e-01 3.33179355e-01
-4.74047303e-01 -4.01894152e-01 1.21800411e+00 9.52320755e-01
5.75528443e-01 2.39085674e-01 3.63119185e-01 -1.72785902e+00
-7.92184651e-01 -9.35994506e-01 -3.43509465e-01 2.29112625e-01
7.77754307e-01 -1.21736777e+00 -2.79787600e-01 -1.18721500e-01] | [7.795698642730713, -2.5394504070281982] |
8dd35154-74d6-4da6-a658-3a2f65077516 | perception-framework-through-real-time | 2103.04136 | null | https://arxiv.org/abs/2103.04136v1 | https://arxiv.org/pdf/2103.04136v1.pdf | Perception Framework through Real-Time Semantic Segmentation and Scene Recognition on a Wearable System for the Visually Impaired | As the scene information, including objectness and scene type, are important for people with visual impairment, in this work we present a multi-task efficient perception system for the scene parsing and recognition tasks. Building on the compact ResNet backbone, our designed network architecture has two paths with shared parameters. In the structure, the semantic segmentation path integrates fast attention, with the aim of harvesting long-range contextual information in an efficient manner. Simultaneously, the scene recognition path attains the scene type inference by passing the semantic features into semantic-driven attention networks and combining the semantic extracted representations with the RGB extracted representations through a gated attention module. In the experiments, we have verified the systems' accuracy and efficiency on both public datasets and real-world scenes. This system runs on a wearable belt with an Intel RealSense LiDAR camera and an Nvidia Jetson AGX Xavier processor, which can accompany visually impaired people and provide assistive scene information in their navigation tasks. | ['Rainer Stiefelhagen', 'Jiaming Zhang', 'Kailun Yang', 'Haoye Chen', 'Yingzhi Zhang'] | 2021-03-06 | null | null | null | null | ['scene-parsing', 'scene-recognition'] | ['computer-vision', 'computer-vision'] | [ 4.73464638e-01 -5.72076887e-02 1.24707930e-01 -6.59792542e-01
-3.23905736e-01 -2.61329804e-02 -2.16788296e-02 -6.72783554e-02
-8.92688751e-01 5.05154967e-01 4.64963049e-01 -2.15948254e-01
-1.56757221e-01 -9.34757411e-01 -4.83122379e-01 -4.36083525e-01
1.98283106e-01 7.33692646e-02 4.29575771e-01 -2.32249632e-01
3.01792949e-01 3.51307362e-01 -2.23004532e+00 4.27574039e-01
1.02753234e+00 1.32168746e+00 9.19756055e-01 9.10845637e-01
-2.21586809e-01 1.07117903e+00 -2.87034363e-01 2.00196188e-02
2.13829547e-01 1.43727079e-01 -8.19303989e-01 2.24570129e-02
5.23911834e-01 -5.92347026e-01 -3.57145756e-01 1.14323413e+00
7.99213588e-01 2.11795256e-01 3.84121425e-02 -1.16930306e+00
-4.56615984e-01 2.30098069e-02 -2.66541928e-01 2.95136124e-01
4.19163793e-01 3.34728420e-01 9.58520353e-01 -8.11126530e-01
1.67972982e-01 1.36563599e+00 6.14503264e-01 2.48111531e-01
-4.61232066e-01 -2.75949329e-01 3.93389344e-01 8.65537584e-01
-1.14698935e+00 -5.60703039e-01 7.19462693e-01 -3.60477298e-01
1.30400896e+00 2.70462751e-01 1.12672269e+00 1.06395304e+00
-1.24042504e-01 7.95482635e-01 1.12101698e+00 -2.32183009e-01
3.31322759e-01 -1.95014626e-02 5.84992707e-01 7.17538357e-01
3.54082465e-01 -3.48374128e-01 -4.29584146e-01 2.88351566e-01
7.80455768e-01 2.77501583e-01 -3.29654127e-01 -1.11113805e-02
-7.49121964e-01 5.44260502e-01 9.69234824e-01 6.33984804e-02
-5.18251300e-01 3.03792715e-01 3.55363578e-01 -1.82228714e-01
2.81034231e-01 -2.67235398e-01 -6.07852697e-01 -5.12778461e-02
-2.13255778e-01 -1.04398914e-01 2.55624205e-01 1.06500471e+00
6.32688403e-01 -7.74866864e-02 -2.35926583e-01 1.02805555e+00
6.75799727e-01 6.08883023e-01 5.06490231e-01 -7.95685887e-01
4.34973300e-01 9.15578961e-01 -6.19403385e-02 -7.66804516e-01
-7.56037056e-01 -3.91442001e-01 -6.85537159e-01 3.26576114e-01
1.06963709e-01 -5.81196025e-02 -1.33364260e+00 1.51794279e+00
2.22390234e-01 1.69173330e-01 -8.93206894e-02 1.19905353e+00
1.25842667e+00 1.31958365e-01 7.32143879e-01 2.94284761e-01
1.93652213e+00 -1.01207602e+00 -4.19142187e-01 -5.76912522e-01
3.50729555e-01 -3.58871847e-01 1.53086233e+00 2.75105238e-01
-7.55792379e-01 -1.01030052e+00 -9.32362258e-01 -7.97166586e-01
-6.53312564e-01 4.11559343e-01 7.67004848e-01 6.14646018e-01
-1.33774865e+00 -5.51595353e-02 -7.56477654e-01 -6.86408699e-01
7.34253109e-01 3.78440589e-01 -1.85368031e-01 -2.43221328e-01
-8.54206979e-01 7.16670573e-01 5.49337447e-01 4.77445602e-01
-5.37384033e-01 -5.18843889e-01 -9.86293137e-01 1.77456915e-01
1.40700966e-01 -1.25757372e+00 1.02154732e+00 -8.43327045e-01
-1.51350617e+00 9.06352520e-01 -2.54776984e-01 -2.47954622e-01
4.12181795e-01 -7.91398168e-01 -2.50750154e-01 2.31198117e-01
2.68365979e-01 6.27520502e-01 7.25179136e-01 -9.25142467e-01
-9.60656285e-01 -8.31085026e-01 2.46107444e-01 7.20816910e-01
-1.84756249e-01 -6.23132847e-02 -4.62695628e-01 -5.40092066e-02
3.34662735e-01 -4.62395936e-01 -2.86627620e-01 2.07593083e-01
-7.51036167e-01 -3.32748026e-01 8.75001013e-01 -8.20784509e-01
4.32551652e-01 -2.26928115e+00 -9.06615406e-02 5.35853915e-02
2.74675727e-01 1.65767223e-01 4.63382676e-02 -1.70281947e-01
2.14324206e-01 -3.14535439e-01 -1.07643753e-01 -6.25407517e-01
-4.05153781e-02 4.49961334e-01 -2.59831756e-01 3.25637102e-01
-2.11721316e-01 8.07700694e-01 -8.40511322e-01 -5.69055498e-01
7.55573750e-01 6.63826168e-01 -7.14310706e-01 4.18757230e-01
-2.18634054e-01 4.13322449e-01 -6.31907403e-01 5.95687687e-01
6.24377012e-01 -1.22228526e-01 -5.46290129e-02 -3.03417385e-01
-1.28918380e-01 2.71513999e-01 -1.06853104e+00 2.07820415e+00
-5.05559444e-01 4.97199297e-01 2.98539549e-01 -8.46306205e-01
5.83299935e-01 -1.91171125e-01 3.02480638e-01 -1.05276597e+00
2.75077552e-01 2.14371663e-02 -5.54264843e-01 -1.05211401e+00
7.21114576e-01 4.00519997e-01 1.89572126e-01 -2.33951941e-01
1.46847228e-02 1.66607350e-01 -3.64734560e-01 -1.23271942e-01
1.06338263e+00 3.63823593e-01 2.20664859e-01 -8.57982337e-02
6.04521573e-01 3.82243879e-02 4.35174435e-01 6.67231083e-01
-2.43556216e-01 2.86713183e-01 -1.37257129e-01 -4.77936059e-01
-8.83339882e-01 -1.42062080e+00 1.19805394e-03 1.43215239e+00
4.92107868e-01 -2.93242812e-01 -5.13822317e-01 -2.78217226e-01
-2.20309600e-01 5.00680327e-01 -2.25553498e-01 9.80978757e-02
-3.79464358e-01 -4.56464410e-01 2.24139795e-01 9.85101700e-01
1.26523137e+00 -1.42030120e+00 -1.23587751e+00 6.82441890e-02
-1.94788113e-01 -1.38810098e+00 2.53907084e-01 -3.00384369e-02
-4.79513973e-01 -1.37363422e+00 -2.60175139e-01 -1.18213093e+00
5.05419552e-01 4.94444013e-01 7.57019341e-01 6.02003150e-02
-7.07079291e-01 5.41258454e-01 -3.15046608e-01 -5.35838485e-01
5.57254851e-01 -6.44923076e-02 -1.64690226e-01 -7.16450624e-03
4.04337287e-01 -7.75492489e-01 -1.15902674e+00 -3.05346549e-01
-3.91075253e-01 4.07930344e-01 4.65526164e-01 4.31185991e-01
6.59053564e-01 3.06648444e-02 2.20243424e-01 -4.62287307e-01
4.07724559e-01 -3.44204932e-01 -5.65767944e-01 6.03084313e-03
-1.19998835e-01 -3.54473025e-01 4.84569728e-01 2.88938284e-01
-1.30867171e+00 2.99849689e-01 -6.06363356e-01 -7.13708177e-02
-6.77236855e-01 5.33509515e-02 -5.43663085e-01 -4.13083425e-03
3.62163514e-01 2.54817396e-01 -2.70760834e-01 -6.44892395e-01
6.80000007e-01 1.16251743e+00 9.88731265e-01 -3.21794450e-01
2.00227469e-01 6.22664630e-01 -2.54978180e-01 -1.20799041e+00
-9.85866070e-01 -6.98756754e-01 -6.54262125e-01 -3.65875453e-01
1.42425096e+00 -1.16332030e+00 -1.19948626e+00 9.65559483e-01
-1.22912490e+00 -4.17352259e-01 -2.98438132e-01 5.38323879e-01
-6.88538551e-01 1.48115888e-01 -4.38776881e-01 -9.53423142e-01
-5.31339884e-01 -1.02870762e+00 1.37066793e+00 5.86313367e-01
2.22558215e-01 -5.95930457e-01 -2.60101348e-01 6.82289779e-01
3.72746706e-01 4.06869464e-02 1.09586728e+00 -2.18667716e-01
-8.29841018e-01 4.47221994e-01 -9.97921646e-01 3.40155423e-01
8.76993779e-03 -5.10718226e-01 -1.45193541e+00 -7.08532557e-02
5.82340285e-02 -3.02571505e-01 9.65133131e-01 6.35032952e-01
1.67737675e+00 -3.03511824e-02 -2.10231841e-01 8.45625639e-01
1.77121162e+00 1.20813437e-01 9.01508272e-01 5.13978839e-01
1.22513509e+00 5.15096605e-01 2.86643147e-01 3.03957433e-01
1.09243608e+00 4.06994045e-01 7.34718680e-01 -2.93215066e-01
-4.89419073e-01 -9.72682461e-02 9.10123512e-02 4.82419252e-01
-2.56105244e-01 -1.16608426e-01 -8.03379297e-01 4.56274420e-01
-1.85208583e+00 -8.07916343e-01 -3.79633695e-01 1.86154807e+00
1.88091889e-01 -3.66361067e-02 -9.20737758e-02 1.08045675e-01
5.95548987e-01 1.22112557e-01 -6.56315088e-01 -2.48590395e-01
-6.12800494e-02 3.90767545e-01 4.97522056e-01 4.07575548e-01
-1.07011688e+00 1.14145255e+00 5.58096838e+00 3.06409657e-01
-9.03965235e-01 2.78553486e-01 1.92759514e-01 -5.36108352e-02
4.87022847e-03 -4.01593328e-01 -5.12054980e-01 3.64904344e-01
6.90237343e-01 5.43381929e-01 4.44069147e-01 1.13834250e+00
2.47048631e-01 -2.67711580e-01 -6.79567337e-01 1.29849708e+00
-1.05081961e-01 -8.09105515e-01 -1.69353589e-01 -1.76082224e-01
-9.75830480e-02 6.92505062e-01 -2.10496798e-01 2.84552753e-01
2.12471649e-01 -9.92343247e-01 7.99404800e-01 6.95264637e-01
7.95992076e-01 -6.27879679e-01 6.65015757e-01 1.72715545e-01
-1.48519850e+00 -5.50248921e-01 -3.68310243e-01 -3.44614774e-01
9.62857679e-02 3.50450099e-01 -5.95819712e-01 4.86241966e-01
1.16349101e+00 7.30029404e-01 -6.73635602e-01 1.19840634e+00
-5.83339989e-01 2.43017808e-01 -2.86696106e-01 -3.25549059e-02
2.11585760e-02 -2.87148178e-01 3.82195175e-01 1.07953513e+00
2.21263558e-01 2.11062878e-01 2.77500749e-01 7.06551433e-01
2.25892976e-01 1.34078607e-01 -4.66262490e-01 6.52289033e-01
2.97230333e-01 1.15551209e+00 -6.04591191e-01 -4.19452250e-01
-4.85962749e-01 1.14987659e+00 4.13037270e-01 5.28525233e-01
-6.68123305e-01 -6.36428118e-01 9.47431028e-01 -3.68117131e-02
2.37653911e-01 -4.54124182e-01 -6.96517944e-01 -9.10730839e-01
4.07466233e-01 -3.08190525e-01 4.12891924e-01 -1.53479147e+00
-1.06378984e+00 8.00021231e-01 -4.19940472e-01 -7.36245871e-01
3.73345196e-01 -1.04782903e+00 -5.06505728e-01 9.39088821e-01
-1.79125059e+00 -1.46795893e+00 -1.36776030e+00 1.20075488e+00
7.22316802e-01 4.91590090e-02 8.51028979e-01 3.66216898e-01
-7.11078823e-01 9.75216180e-02 -5.33268392e-01 1.63610741e-01
-3.08310706e-02 -1.25086200e+00 2.95786351e-01 8.50006402e-01
-2.52478302e-01 3.24972779e-01 3.13683420e-01 -5.91483414e-01
-1.25867140e+00 -1.29864848e+00 4.71816063e-01 -8.25912952e-02
3.96100491e-01 -3.74813676e-01 -6.91746354e-01 6.99533343e-01
4.64656837e-02 6.61234260e-02 5.30290782e-01 9.64294598e-02
-2.38527998e-01 -1.19911946e-01 -1.32880676e+00 6.44031942e-01
1.95662117e+00 -8.25501680e-01 -7.18289912e-01 6.01492584e-01
9.19185221e-01 -4.77989852e-01 -4.96203810e-01 3.38632703e-01
4.50560540e-01 -9.84612405e-01 1.17640901e+00 -5.34136295e-01
6.31161109e-02 -4.25038040e-01 -6.79037213e-01 -8.65228176e-01
-3.74679089e-01 1.30717561e-01 1.49217233e-01 7.46860147e-01
-1.74438238e-01 -9.34186876e-01 8.30263257e-01 5.13384163e-01
-5.94846904e-01 -5.46256125e-01 -1.08744276e+00 -3.25195491e-01
-8.03792894e-01 -1.04267097e+00 6.14004493e-01 1.28533989e-01
-3.36699784e-01 4.38386470e-01 1.94184966e-02 6.61814034e-01
7.46371984e-01 -4.04940471e-02 5.46042681e-01 -1.36340523e+00
-5.37899211e-02 -2.24203587e-01 -9.83123243e-01 -1.11346519e+00
1.96674451e-01 -7.41247475e-01 8.67803693e-02 -2.20320392e+00
1.20527901e-01 -4.41595465e-01 -3.33203703e-01 5.61464608e-01
-9.58473012e-02 9.17445868e-02 1.13544486e-01 1.47903506e-02
-7.44126022e-01 6.12257123e-01 9.97050345e-01 -1.34757131e-01
-3.03500593e-01 -3.54258530e-02 -7.77874053e-01 1.12411928e+00
8.39886785e-01 5.72905988e-02 -7.81925917e-01 -8.16464484e-01
-7.06393197e-02 -3.66017818e-01 1.12708569e+00 -1.38316405e+00
3.43138486e-01 1.17741533e-01 5.59742212e-01 -7.33232379e-01
6.63245201e-01 -9.06871319e-01 -3.20970148e-01 5.71425736e-01
1.38916850e-01 -1.94609612e-01 1.02039598e-01 4.74000514e-01
2.65115917e-01 2.20471859e-01 6.48592710e-01 -1.54636219e-01
-1.51650143e+00 3.22619706e-01 -1.39343500e-01 -2.09261850e-01
9.71261501e-01 -4.78821665e-01 -6.82760537e-01 -3.94191593e-02
-7.99754322e-01 3.16800147e-01 8.17894191e-02 6.79356217e-01
9.85050142e-01 -1.02455580e+00 -3.21244866e-01 6.52648687e-01
1.14128485e-01 4.07578737e-01 5.63421249e-01 4.92896140e-01
-5.51361144e-01 3.26317817e-01 -5.37393689e-01 -8.88614416e-01
-1.29745948e+00 3.27185571e-01 3.94609123e-01 3.63093138e-01
-1.16379881e+00 8.38749647e-01 4.90850862e-03 -2.52837151e-01
3.76121640e-01 -6.86453164e-01 -4.94096756e-01 -1.27447709e-01
5.86160541e-01 4.72052693e-01 7.88174197e-02 -6.03669584e-01
-4.99035835e-01 7.93140173e-01 3.76482129e-01 7.97897428e-02
1.41819274e+00 -4.48741198e-01 -1.74775541e-01 1.34817958e-01
9.33224738e-01 -3.06339294e-01 -1.09968102e+00 -7.01898560e-02
-2.44522795e-01 -3.67361516e-01 1.53559268e-01 -8.93766880e-01
-9.76326346e-01 8.14328790e-01 1.09029758e+00 1.07934304e-01
1.56862319e+00 5.46191595e-02 7.84801006e-01 4.08088535e-01
8.06975245e-01 -1.05497253e+00 -1.13072321e-01 5.03856003e-01
8.00046504e-01 -1.16434419e+00 -2.55632609e-01 -6.42278373e-01
-3.34377974e-01 8.56845975e-01 9.90732610e-01 -5.27279079e-02
5.18740833e-01 2.59256214e-02 1.01193516e-02 -4.20141011e-01
-6.57555535e-02 -8.79878879e-01 1.16946176e-01 1.06310546e+00
-1.18596658e-01 3.20638776e-01 3.06176811e-01 6.24513805e-01
-5.33129990e-01 3.78233977e-02 7.95134529e-02 7.11262882e-01
-9.61980104e-01 -4.69477952e-01 -1.04863182e-01 4.60465640e-01
4.96061258e-02 -1.72398180e-01 -1.32102601e-03 5.16506553e-01
6.23686790e-01 1.04038310e+00 4.14141297e-01 -4.19101566e-01
6.49719298e-01 -1.40740082e-01 4.42775667e-01 -7.11307406e-01
-2.98379153e-01 -5.31779766e-01 3.51431906e-01 -1.10546923e+00
-2.23356143e-01 -2.34926045e-01 -1.50593400e+00 9.91376638e-02
3.29320520e-01 -5.36223114e-01 1.00352037e+00 1.05586159e+00
5.32463610e-01 9.61582303e-01 3.04953396e-01 -9.08705354e-01
-1.61154531e-02 -1.00442815e+00 -4.23130393e-01 2.83785790e-01
3.26980829e-01 -6.97259068e-01 6.22707568e-02 -1.92840368e-01] | [8.27701473236084, -1.5362908840179443] |
866df3da-4a70-40dd-a0ad-9a286f87a901 | explainable-slot-type-attentions-to-improve | 2210.10227 | null | https://arxiv.org/abs/2210.10227v1 | https://arxiv.org/pdf/2210.10227v1.pdf | Explainable Slot Type Attentions to Improve Joint Intent Detection and Slot Filling | Joint intent detection and slot filling is a key research topic in natural language understanding (NLU). Existing joint intent and slot filling systems analyze and compute features collectively for all slot types, and importantly, have no way to explain the slot filling model decisions. In this work, we propose a novel approach that: (i) learns to generate additional slot type specific features in order to improve accuracy and (ii) provides explanations for slot filling decisions for the first time in a joint NLU model. We perform an additional constrained supervision using a set of binary classifiers for the slot type specific feature learning, thus ensuring appropriate attention weights are learned in the process to explain slot filling decisions for utterances. Our model is inherently explainable and does not need any post-hoc processing. We evaluate our approach on two widely used datasets and show accuracy improvements. Moreover, a detailed analysis is also provided for the exclusive slot explainability. | ['Hongxia Jin', 'Akhila Yerukola', 'Vijay Srinivasan', 'Kalpa Gunaratna'] | 2022-10-19 | null | null | null | null | ['intent-detection', 'slot-filling'] | ['natural-language-processing', 'natural-language-processing'] | [ 6.23630702e-01 9.99037385e-01 -7.09040821e-01 -7.11121440e-01
-5.73843300e-01 -1.73290148e-01 5.84771752e-01 5.48735857e-01
-3.47693145e-01 9.26115930e-01 3.42124760e-01 -6.61049426e-01
-1.18459165e-01 -8.31201613e-01 -6.75099850e-01 -7.29314163e-02
1.65924668e-01 9.31949198e-01 3.43246683e-02 -2.46227890e-01
1.50944591e-01 -1.30836636e-01 -1.68141901e+00 3.88858795e-01
1.30909526e+00 8.95888925e-01 4.33348686e-01 5.61609328e-01
-7.58260965e-01 8.28799725e-01 -3.82332385e-01 -1.50484979e-01
-2.87685785e-02 -3.67215037e-01 -1.26200056e+00 3.63287807e-01
-4.78161201e-02 -2.38903373e-01 8.39690864e-02 5.95153868e-01
-5.42570502e-02 2.69192398e-01 4.40441787e-01 -1.31956267e+00
-4.43217903e-01 1.10001028e+00 -1.28256753e-01 -6.57075718e-02
5.23340285e-01 -2.95423716e-01 1.52573979e+00 -8.37230504e-01
5.54378748e-01 1.35369265e+00 3.65522683e-01 7.62242198e-01
-1.11331570e+00 -5.16604304e-01 5.25676787e-01 1.69939429e-01
-1.05394614e+00 -2.58868963e-01 6.96542501e-01 -1.53695434e-01
1.27677250e+00 3.78619075e-01 3.93428266e-01 6.36582375e-01
-2.01474771e-01 1.25266993e+00 8.02814484e-01 -7.75445938e-01
3.70452464e-01 3.82915676e-01 6.83941364e-01 6.14696145e-01
1.14295505e-01 -1.54476762e-01 -5.87873459e-01 -2.17430577e-01
4.26806539e-01 1.00874260e-01 -3.07556968e-02 -3.20488751e-01
-9.61297214e-01 1.21839952e+00 2.85024852e-01 2.56440043e-01
-3.53974044e-01 3.79312858e-02 2.99793422e-01 2.44009539e-01
7.39169836e-01 5.94500661e-01 -9.33919966e-01 -3.97672355e-02
-7.66312778e-01 3.67035687e-01 8.11657131e-01 9.79546249e-01
1.41119730e+00 -3.54777843e-01 -3.36152852e-01 9.62360322e-01
4.14159656e-01 -2.30571609e-02 7.23544896e-01 -4.88369644e-01
5.22608042e-01 8.32871556e-01 2.88561106e-01 -6.13865376e-01
-7.65080810e-01 -3.19079310e-01 -3.79722983e-01 -3.99771482e-01
1.64244790e-02 -3.59006152e-02 -1.19721222e+00 1.86549878e+00
4.55849141e-01 3.19018662e-02 3.45341891e-01 8.07455122e-01
8.22165787e-01 4.42185611e-01 7.13540196e-01 -7.41482079e-02
1.72285533e+00 -1.10080576e+00 -1.15423286e+00 -7.67589927e-01
1.03127301e+00 -6.01125360e-01 1.15123761e+00 -1.07904620e-01
-5.69678485e-01 -5.12594998e-01 -8.58097672e-01 -4.47023898e-01
-3.73163313e-01 4.13739026e-01 1.42048943e+00 4.78235573e-01
-7.99984813e-01 3.33760113e-01 -7.28985250e-01 -4.15257424e-01
1.59335971e-01 4.98826981e-01 -3.96781385e-01 5.65402731e-02
-1.51822495e+00 6.75104737e-01 6.19426489e-01 -3.15288931e-01
-1.90477908e-01 -6.30486012e-01 -1.42066669e+00 2.20712706e-01
6.31222665e-01 -6.47144556e-01 1.46595812e+00 -5.64957976e-01
-1.23948741e+00 7.96643615e-01 -6.98808908e-01 -8.91130865e-01
2.15244457e-01 -3.99587601e-01 -1.12545915e-01 -2.26615041e-01
5.50621510e-01 1.13255084e+00 6.27829194e-01 -1.06851149e+00
-9.52678442e-01 -2.82241851e-01 4.77036506e-01 3.79297346e-01
-2.45773464e-01 -3.79425704e-01 -3.76847714e-01 -4.94829506e-01
7.74460196e-01 -7.27444828e-01 -3.82992744e-01 -3.48689258e-01
-6.29744828e-01 -7.49983966e-01 6.35424256e-01 -6.36723816e-01
1.38384378e+00 -1.87855399e+00 -5.25907055e-02 -5.79577945e-02
1.48356244e-01 2.18907353e-02 2.39877943e-02 1.55602306e-01
-6.07252344e-02 2.47018546e-01 -4.59702253e-01 -9.66087282e-01
3.21802050e-01 6.09938741e-01 -7.01991439e-01 -1.25236422e-01
4.44964617e-01 9.98492420e-01 -6.04027689e-01 -3.51016879e-01
3.19115013e-01 5.75260818e-02 -1.07112050e+00 3.60330820e-01
-7.17912674e-01 1.23496540e-01 -6.42536283e-01 5.15943170e-01
5.79363048e-01 -2.95073062e-01 3.76090348e-01 -1.56866297e-01
-2.19126604e-02 8.89308572e-01 -1.11769831e+00 1.81171417e+00
-7.63930500e-01 3.20071489e-01 -2.61791676e-01 -1.18926823e+00
9.77395236e-01 3.96298587e-01 3.38188857e-01 -7.33117878e-01
2.79043801e-02 3.53909880e-01 -2.56645650e-01 -1.40594110e-01
9.67926383e-01 -5.57676442e-02 -4.12510157e-01 6.62001252e-01
2.14343071e-01 6.77874386e-02 1.80979162e-01 2.12640956e-01
8.58281732e-01 8.46555904e-02 6.91799521e-01 -2.48736635e-01
5.14338672e-01 2.68627822e-01 7.75092304e-01 1.03441536e+00
-1.75949261e-01 3.82938147e-01 3.44378918e-01 -6.88440681e-01
-9.34651434e-01 -5.55625856e-01 -3.67162377e-01 1.29766095e+00
2.97804654e-01 -3.62049550e-01 -6.56977177e-01 -8.66443157e-01
1.08036496e-01 9.55026031e-01 -7.73745477e-01 5.88650815e-03
-2.42668405e-01 -5.39927185e-01 4.54944484e-02 5.25798857e-01
4.28513080e-01 -1.26798749e+00 -5.60438335e-01 3.84429932e-01
-3.55831265e-01 -1.29379642e+00 -1.21731468e-01 7.53592312e-01
-7.42689967e-01 -9.27525222e-01 -2.41761699e-01 -6.13193870e-01
7.77008474e-01 2.28439078e-01 1.09895742e+00 5.79402030e-01
-1.92646891e-01 2.62119949e-01 -6.70265555e-01 -3.75247121e-01
-1.92922905e-01 4.81824964e-01 6.84713796e-02 -2.46710405e-01
7.90185988e-01 -3.05914700e-01 -2.01664597e-01 2.56272554e-01
-9.24638152e-01 5.32948554e-01 3.92899394e-01 1.11454940e+00
5.55281520e-01 -1.00963973e-01 6.05993986e-01 -1.31889188e+00
4.42844003e-01 -6.00574553e-01 -2.71263301e-01 2.47924969e-01
-6.82682276e-01 4.63266045e-01 2.53915012e-01 -8.88542272e-03
-1.09236145e+00 2.95524836e-01 -3.24018031e-01 1.68003708e-01
-4.59185302e-01 4.85588402e-01 -1.45571485e-01 3.15102667e-01
4.31666464e-01 6.68425635e-02 -3.23304981e-01 -6.59907699e-01
4.71869469e-01 6.87701523e-01 2.28709593e-01 -6.55455410e-01
6.82776988e-01 3.01078558e-01 -4.75547701e-01 -6.98138297e-01
-1.32952428e+00 -6.50741339e-01 -5.91067672e-01 4.11659569e-01
7.21267521e-01 -9.50576961e-01 -4.41599101e-01 -1.30527124e-01
-1.19144452e+00 -2.73227870e-01 -2.23017648e-01 4.38604742e-01
-6.49475098e-01 2.25994796e-01 -4.43880856e-01 -1.11179256e+00
-3.07446659e-01 -9.41677451e-01 1.31607664e+00 4.47087377e-01
-6.26560152e-01 -9.68380868e-01 -2.30404019e-01 6.14117920e-01
3.56800497e-01 -4.20652956e-01 9.61346686e-01 -1.10981679e+00
-5.53255498e-01 3.72686759e-02 -4.15069401e-01 -4.04932469e-01
3.09125721e-01 -6.19395971e-01 -1.10638857e+00 1.63926661e-01
-2.80405194e-01 -3.93201411e-01 1.11061835e+00 4.39140886e-01
1.04485905e+00 -2.81104326e-01 -4.45859045e-01 1.78629339e-01
9.91906643e-01 1.65213585e-01 3.72174054e-01 5.21570027e-01
3.51354718e-01 7.93368042e-01 1.23131311e+00 6.07539535e-01
7.87777126e-01 7.25261688e-01 2.71040380e-01 -3.58064830e-01
1.39218107e-01 -5.45911014e-01 -1.73154399e-01 1.79584220e-01
5.71264207e-01 -2.25235075e-01 -7.08062887e-01 7.46926129e-01
-2.23931265e+00 -3.52410614e-01 1.33281231e-01 1.79106677e+00
8.46902788e-01 9.96702239e-02 -2.57462412e-01 3.89927298e-01
6.06622279e-01 7.18292296e-02 -4.92638826e-01 -4.51866478e-01
4.74531762e-03 3.98859769e-01 2.00735286e-01 9.47630167e-01
-1.34483504e+00 1.59168255e+00 5.75347614e+00 5.71800470e-01
-6.87309146e-01 -3.63281295e-02 7.28428900e-01 4.45330501e-01
-8.34443212e-01 4.23868269e-01 -8.45362604e-01 -6.01762421e-02
3.69402260e-01 -3.53080221e-02 1.21498339e-01 1.10835350e+00
-6.21213913e-02 -3.32516491e-01 -9.26831543e-01 5.69029987e-01
-3.79884481e-01 -1.32572782e+00 1.41471103e-01 -1.73208252e-01
4.22869116e-01 -4.69176620e-01 -1.75087810e-01 6.28311515e-01
2.76114643e-01 -8.66625607e-01 6.32811368e-01 1.60852551e-01
4.31474656e-01 -6.44588411e-01 8.02418649e-01 4.45726573e-01
-1.18532157e+00 -2.08926946e-01 -5.86074173e-01 -5.58786452e-01
3.51497769e-01 7.08188355e-01 -1.05224490e+00 4.79877353e-01
3.60827148e-01 5.73240638e-01 -2.35367790e-01 6.31973624e-01
-5.18370986e-01 5.80313981e-01 -4.08677429e-01 -1.58729777e-02
3.08814973e-01 3.31577331e-01 2.71062970e-01 9.32788670e-01
1.78279132e-01 2.16530010e-01 4.54986542e-01 9.81655121e-01
2.04501778e-01 3.16275477e-01 -2.07648650e-01 -2.15478584e-01
5.90250254e-01 1.14554036e+00 -9.24034119e-01 -4.00937289e-01
-2.33594403e-01 1.02329862e+00 4.22808886e-01 1.30872950e-01
-5.30246258e-01 -6.49390668e-02 1.00860560e+00 -2.87487745e-01
1.81924596e-01 -1.88816767e-02 -6.43111348e-01 -1.21391094e+00
-1.11074530e-01 -2.62468398e-01 5.41604519e-01 -4.34552819e-01
-7.32511759e-01 5.66571176e-01 -7.21762255e-02 -7.60186553e-01
-6.53720915e-01 -4.59621757e-01 -4.66525108e-01 8.08591247e-01
-2.02077723e+00 -1.02724195e+00 -1.40583396e-01 1.87801778e-01
8.40403736e-01 -4.45664972e-02 1.11341894e+00 2.40605682e-01
-4.93938893e-01 5.41162312e-01 -6.89416945e-01 -1.56445697e-01
2.24850222e-01 -1.28970838e+00 6.81211710e-01 4.90033686e-01
3.56488436e-01 7.95276284e-01 1.14134932e+00 -6.92023158e-01
-9.19286370e-01 -7.51349151e-01 1.59926069e+00 -1.15137056e-01
3.11875910e-01 -4.33691531e-01 -1.00843167e+00 8.12166035e-01
-1.20397836e-01 -1.88914672e-01 8.75745595e-01 6.21874869e-01
-1.11589409e-01 4.85544294e-01 -1.08240545e+00 3.32228690e-01
1.06012547e+00 -4.87416804e-01 -8.14810574e-01 3.97730708e-01
1.30447078e+00 -5.49283326e-01 -3.49488497e-01 4.34197038e-01
3.78940284e-01 -7.66362727e-01 7.01261044e-01 -7.80817807e-01
3.00744295e-01 -3.62454116e-01 -2.35545367e-01 -8.96933854e-01
-4.89216521e-02 -3.36025089e-01 -1.25856772e-01 1.16426837e+00
7.84506738e-01 -6.16856396e-01 1.02307498e+00 1.02266788e+00
-1.94831967e-01 -8.06330144e-01 -8.58832061e-01 -1.14266664e-01
-4.79614913e-01 -8.02911520e-01 8.85817647e-01 8.38685334e-01
2.36485228e-01 3.17310214e-01 -5.27015805e-01 2.18815058e-01
3.99135977e-01 5.09579182e-01 7.80026138e-01 -1.27798164e+00
-2.17999578e-01 6.70284480e-02 -2.09336236e-01 -1.24099934e+00
5.12050867e-01 -7.76201308e-01 1.67391732e-01 -1.64653516e+00
1.43282980e-01 -9.48754013e-01 -2.55268335e-01 1.12287462e+00
-4.90154356e-01 -2.84817945e-02 6.24391921e-02 1.49958491e-01
-6.30622923e-01 7.85247624e-01 7.67824173e-01 -1.33577570e-01
-3.31963778e-01 1.38363689e-01 -1.03738534e+00 4.24044609e-01
8.39279532e-01 -4.23251390e-01 -5.89461625e-01 -1.71071574e-01
-1.13498792e-02 3.52288820e-02 9.78193358e-02 -7.88827300e-01
1.97235700e-02 -1.96549252e-01 5.26374988e-02 -5.97176075e-01
3.43566179e-01 -7.68528759e-01 -3.59918922e-01 3.86116445e-01
-5.50065637e-01 -5.58917761e-01 2.79502332e-01 3.20689976e-01
-3.17978829e-01 -3.73389274e-01 2.10155770e-01 5.70780830e-03
-1.11148846e+00 2.27291673e-01 -2.69836664e-01 -2.77480543e-01
7.10264981e-01 -1.14472911e-01 -3.53694595e-02 -4.47497070e-01
-9.01752949e-01 5.13993144e-01 1.89693078e-01 7.22998321e-01
2.79062718e-01 -1.22657931e+00 -2.04285949e-01 4.21882957e-01
3.51117015e-01 3.61820348e-02 4.11768347e-01 2.58357793e-01
2.30467208e-02 1.00693166e+00 1.89404078e-02 -5.04006565e-01
-9.62681472e-01 2.21379295e-01 1.54913291e-01 -6.46804869e-01
-4.53324854e-01 8.75928283e-01 3.29913259e-01 -1.00772345e+00
7.11514130e-02 -4.48826075e-01 -5.66554308e-01 7.95454681e-02
5.10408342e-01 -4.32061225e-01 7.72889936e-03 -5.56678593e-01
-3.24482501e-01 2.13365808e-01 -2.65178561e-01 -5.70280552e-02
1.20316482e+00 -4.62615132e-01 1.34853363e-01 2.48411343e-01
7.99177587e-01 -4.00292844e-01 -1.10877478e+00 -6.16795063e-01
4.25523430e-01 -3.33687872e-01 -6.55891374e-02 -6.77541256e-01
-6.01197839e-01 7.54340708e-01 2.77841985e-01 3.79600734e-01
8.45256329e-01 1.57927871e-01 8.69993389e-01 3.41527730e-01
3.93645823e-01 -1.07489657e+00 -3.58109266e-01 7.29164600e-01
5.29807389e-01 -1.44864953e+00 -2.32412502e-01 -7.42049694e-01
-6.27642989e-01 9.72810924e-01 8.56405914e-01 3.28111440e-01
3.98781896e-01 9.02381018e-02 2.01161191e-01 -1.73595384e-01
-9.14171696e-01 -5.17145097e-01 2.82639474e-01 4.54875976e-01
7.94830143e-01 2.44141862e-01 -8.53101969e-01 1.02887988e+00
-4.79917705e-01 -1.95100695e-01 1.19909391e-01 8.82699072e-01
-7.82869399e-01 -1.47042513e+00 -9.33486074e-02 4.79247868e-01
-1.34458870e-01 -1.06488273e-01 -2.20657393e-01 5.52538574e-01
-1.05108256e-05 9.66436982e-01 1.86887830e-01 -3.15351188e-01
9.01483223e-02 4.14884329e-01 -1.77185386e-01 -9.13335204e-01
-2.09469318e-01 -2.74284244e-01 2.10543260e-01 -6.21034145e-01
-2.28791654e-01 -4.96038854e-01 -1.83167648e+00 2.18937621e-01
-5.61974466e-01 5.90434194e-01 7.62246013e-01 1.59145892e+00
4.17011440e-01 6.90115869e-01 4.63754654e-01 -7.77151525e-01
-2.43238121e-01 -9.23109114e-01 -5.17420232e-01 3.39380771e-01
2.86408067e-01 -1.12363875e+00 -3.59992981e-01 -2.77296185e-01] | [12.520949363708496, 7.340747833251953] |
d57cb832-9bfb-43c3-8e7e-55ce9acb8b2f | dynamic-pose-robust-facial-expression | 1607.06250 | null | http://arxiv.org/abs/1607.06250v1 | http://arxiv.org/pdf/1607.06250v1.pdf | Dynamic Pose-Robust Facial Expression Recognition by Multi-View Pairwise Conditional Random Forests | Automatic facial expression classification (FER) from videos is a critical
problem for the development of intelligent human-computer interaction systems.
Still, it is a challenging problem that involves capturing high-dimensional
spatio-temporal patterns describing the variation of one's appearance over
time. Such representation undergoes great variability of the facial morphology
and environmental factors as well as head pose variations. In this paper, we
use Conditional Random Forests to capture low-level expression transition
patterns. More specifically, heterogeneous derivative features (e.g. feature
point movements or texture variations) are evaluated upon pairs of images. When
testing on a video frame, pairs are created between this current frame and
previous ones and predictions for each previous frame are used to draw trees
from Pairwise Conditional Random Forests (PCRF) whose pairwise outputs are
averaged over time to produce robust estimates. Moreover, PCRF collections can
also be conditioned on head pose estimation for multi-view dynamic FER. As
such, our approach appears as a natural extension of Random Forests for
learning spatio-temporal patterns, potentially from multiple viewpoints.
Experiments on popular datasets show that our method leads to significant
improvements over standard Random Forests as well as state-of-the-art
approaches on several scenarios, including a novel multi-view video corpus
generated from a publicly available database. | ['Séverine Dubuisson', 'Kévin Bailly', 'Arnaud Dapogny'] | 2016-07-21 | null | null | null | null | ['head-pose-estimation'] | ['computer-vision'] | [ 1.89762533e-01 -5.34071028e-01 1.44995302e-02 -8.26129854e-01
-7.30322123e-01 -4.59463596e-01 7.32166946e-01 -2.56432742e-01
-2.44526431e-01 7.06188679e-01 7.81927854e-02 6.50491714e-01
6.27477095e-02 -5.10102034e-01 -7.49543905e-01 -1.00762439e+00
-2.96407580e-01 3.45838100e-01 3.04035872e-01 -1.51303872e-01
1.92211531e-02 9.15561557e-01 -1.98001516e+00 4.90146786e-01
2.13925332e-01 1.27862906e+00 -6.72554504e-03 5.16688466e-01
1.25778660e-01 6.74925148e-01 -3.53679836e-01 -5.63662648e-01
-1.23280510e-01 -3.76767218e-02 -5.39808869e-01 5.16377151e-01
6.57167435e-01 -1.02502525e-01 2.36387447e-01 8.68214190e-01
3.57990384e-01 1.66843772e-01 7.35972822e-01 -1.42599237e+00
1.81537628e-01 -2.12895215e-01 -7.72233367e-01 -2.73028947e-02
7.76368022e-01 6.97803795e-02 6.04435444e-01 -1.17584741e+00
1.11383259e+00 1.39978349e+00 5.47960520e-01 4.45804507e-01
-1.37452888e+00 -7.83706963e-01 4.81857628e-01 4.27682877e-01
-1.36497951e+00 -8.65615249e-01 1.01518416e+00 -6.95215940e-01
4.61085528e-01 2.41231263e-01 9.74338710e-01 1.33940613e+00
2.78196543e-01 9.19091523e-01 1.38054037e+00 -3.58921647e-01
2.68134207e-01 -8.66964012e-02 -2.14624316e-01 8.15256953e-01
-4.29994673e-01 -2.28182271e-01 -9.64081168e-01 -3.55965555e-01
4.26586866e-01 -1.70693800e-01 -3.45400810e-01 -6.84768677e-01
-8.34064782e-01 6.06231034e-01 1.41643193e-02 1.09962806e-01
-5.54961681e-01 -2.95443118e-01 1.77524030e-01 3.74906845e-02
6.39072597e-01 -3.69484425e-01 -4.95202005e-01 -2.30877191e-01
-1.03021336e+00 5.00757575e-01 4.64006066e-01 6.24751210e-01
9.51297939e-01 -1.46087438e-01 -6.88190758e-03 1.01489639e+00
2.78472155e-01 5.27239561e-01 3.49174649e-01 -9.41887021e-01
2.10937709e-01 3.49093556e-01 -1.36162573e-02 -1.31915855e+00
-4.35044289e-01 1.73530668e-01 -8.25959563e-01 2.80152500e-01
6.34403348e-01 -1.69431213e-02 -8.24502289e-01 1.87518263e+00
7.93020129e-01 1.41233161e-01 -4.06652093e-01 6.23563528e-01
6.18728101e-01 3.44489545e-01 5.50063178e-02 -7.28046715e-01
1.24116981e+00 -4.88170683e-01 -6.37109995e-01 -3.41146924e-02
1.52645946e-01 -6.60307050e-01 5.09676754e-01 7.72444785e-01
-8.92568648e-01 -5.10615349e-01 -3.87277454e-01 4.21558261e-01
-1.89085841e-01 1.85810216e-02 5.22016168e-01 4.21889544e-01
-9.95074868e-01 4.48913872e-01 -9.71022546e-01 -3.19249511e-01
3.96289438e-01 5.00174046e-01 -8.62195373e-01 -3.87947038e-02
-8.51465583e-01 5.83415151e-01 -3.56581062e-01 4.77224976e-01
-5.63977718e-01 -2.93224931e-01 -8.98990512e-01 -5.07355034e-01
2.43540585e-01 -4.81886387e-01 1.08809865e+00 -1.40807152e+00
-1.63139427e+00 1.04603767e+00 -6.73930168e-01 -5.72453719e-04
6.38296247e-01 -1.94291174e-01 -3.99695694e-01 2.10471898e-01
8.50012004e-02 6.13989174e-01 1.27130556e+00 -1.31959307e+00
-4.40760225e-01 -9.76181388e-01 -2.80626744e-01 1.32794350e-01
3.79619449e-02 3.57587934e-01 -6.16669893e-01 -4.18267727e-01
1.63416192e-01 -1.24316454e+00 -1.50296614e-01 3.89341772e-01
-2.10062757e-01 -2.15547189e-01 7.46976614e-01 -5.84030628e-01
8.97999406e-01 -2.03881073e+00 4.29073393e-01 2.70546436e-01
-3.35213505e-02 -1.24553621e-01 2.20608935e-01 -1.16754845e-01
-9.34241712e-02 -2.86934644e-01 -9.13441926e-02 -4.33403045e-01
-3.21850181e-01 1.06096461e-01 -5.68302628e-03 7.65191674e-01
2.93361664e-01 4.12818074e-01 -5.99129915e-01 -8.08669686e-01
1.68681249e-01 5.81717074e-01 -4.72665876e-01 3.19348097e-01
-2.03241393e-01 8.26000512e-01 -3.79265398e-01 8.29798639e-01
6.99888706e-01 1.69670418e-01 2.21938819e-01 -2.50944018e-01
-4.66902703e-02 -3.35727483e-01 -1.18669331e+00 1.63914800e+00
-4.56511706e-01 5.22753596e-01 1.42613381e-01 -9.08953667e-01
8.96958590e-01 3.23684186e-01 7.81060398e-01 -3.58816952e-01
2.40432590e-01 -1.04020372e-01 -3.43941569e-01 -6.53672636e-01
1.73210949e-01 -2.53943294e-01 -1.80515852e-02 -2.21464373e-02
2.51594156e-01 7.93433264e-02 8.66813958e-02 -3.71811688e-01
7.39377797e-01 4.24956232e-01 4.64900881e-01 1.24513730e-01
6.02706194e-01 -4.93009865e-01 7.43696332e-01 1.95527747e-02
-2.47241378e-01 6.92865491e-01 5.13080895e-01 -6.26327455e-01
-4.32855994e-01 -9.97970045e-01 -2.96049207e-01 1.01294279e+00
-1.78482801e-01 -2.49714226e-01 -7.34675169e-01 -6.52332902e-01
-1.14266858e-01 1.87594175e-01 -9.06314850e-01 2.91957021e-01
-4.73289549e-01 -7.96598792e-01 1.14818856e-01 2.48323709e-01
3.11987162e-01 -1.08036256e+00 -7.43925333e-01 8.41274112e-02
-3.08860719e-01 -1.35853004e+00 -2.34670803e-01 -8.64917934e-02
-5.39305091e-01 -1.02027023e+00 -7.03063726e-01 -3.86086881e-01
4.04222399e-01 -3.84939574e-02 1.08974636e+00 -4.61638629e-01
-4.89833057e-01 4.21315193e-01 -2.69494265e-01 -1.56486124e-01
-6.27827942e-02 -4.60672498e-01 2.29141563e-01 8.29365015e-01
2.50395477e-01 -8.43181610e-01 -5.24148583e-01 5.71092546e-01
-4.76262748e-01 -5.09156436e-02 1.87138438e-01 8.40947509e-01
9.04592931e-01 -5.87847605e-02 3.90390418e-02 -6.99151337e-01
6.08978607e-02 -5.48872054e-01 -5.20586371e-01 3.24924529e-01
-5.05355783e-02 -1.29876375e-01 4.16604578e-01 -6.46901608e-01
-1.30769300e+00 5.97532451e-01 -1.31203830e-02 -6.85289025e-01
-5.35941124e-01 2.31825292e-01 -4.38023686e-01 6.95519000e-02
3.85313094e-01 1.78574175e-01 1.25166625e-01 -2.33851507e-01
9.20756534e-02 5.42175293e-01 3.63836408e-01 -6.53044224e-01
4.65979874e-01 6.55429125e-01 2.31061503e-01 -1.13182068e+00
-6.53122246e-01 -2.90514141e-01 -1.14619088e+00 -8.18549037e-01
1.02284801e+00 -9.16067362e-01 -5.79247594e-01 8.02794099e-01
-1.08141005e+00 -1.84558392e-01 2.82190502e-01 2.69762605e-01
-7.25509048e-01 8.49807560e-02 -1.72017872e-01 -1.11301255e+00
1.13146670e-01 -1.20711064e+00 1.53259468e+00 2.15775907e-01
-4.11756188e-01 -8.75735760e-01 2.44451821e-01 2.95546651e-01
-1.80658698e-01 7.30494976e-01 5.30430436e-01 -7.43779466e-02
-2.91305542e-01 -1.92549527e-01 3.87218446e-01 1.48501083e-01
1.55800030e-01 6.70520902e-01 -1.19873762e+00 -5.71135953e-02
-1.62157312e-01 -4.06731248e-01 4.78478253e-01 5.96111476e-01
1.21719456e+00 -1.89901322e-01 -3.69109809e-01 5.76456904e-01
1.03709531e+00 2.00146064e-01 4.53201145e-01 1.26664877e-01
5.76271176e-01 8.98639441e-01 8.56443644e-01 7.73591936e-01
3.74459505e-01 1.36596847e+00 2.88607299e-01 1.81910396e-01
2.20345199e-01 -2.16812521e-01 5.74663281e-01 3.27825248e-01
-4.73180771e-01 1.13193586e-01 -7.54232645e-01 2.34768897e-01
-1.61745048e+00 -1.27017629e+00 3.06178369e-02 2.32751322e+00
5.80628991e-01 -1.97446451e-01 4.77538377e-01 1.44445002e-01
6.74246728e-01 3.07650983e-01 -4.77127075e-01 -1.60098359e-01
-2.29281709e-01 3.15823197e-01 -1.41074106e-01 2.31321707e-01
-1.22262347e+00 7.38586545e-01 5.39056778e+00 6.28096223e-01
-1.47072494e+00 -8.69765282e-02 8.65981817e-01 -1.90004110e-01
2.41777673e-02 -4.13809001e-01 -8.09796095e-01 4.20651466e-01
6.19691133e-01 1.99336275e-01 1.56877607e-01 8.52173686e-01
1.81859434e-01 -4.02969301e-01 -8.36589515e-01 1.21678329e+00
1.84815764e-01 -8.96860480e-01 -5.39486706e-02 6.46667257e-02
5.42424202e-01 -2.79922068e-01 8.71131346e-02 -9.15268660e-02
1.39348954e-01 -9.03767765e-01 9.22981203e-01 7.77134418e-01
8.66010368e-01 -7.63751149e-01 4.08865094e-01 2.80785203e-01
-1.25726151e+00 7.72194713e-02 5.61485514e-02 1.31210595e-01
2.54409611e-01 5.19272149e-01 -6.52398229e-01 3.72061342e-01
9.39976871e-01 8.53527546e-01 -6.69591188e-01 6.54356182e-01
-3.61825451e-02 3.38724434e-01 -3.93566608e-01 1.28660779e-02
-3.20029914e-01 -1.56709999e-01 5.55217981e-01 1.09616578e+00
3.42356533e-01 1.43198013e-01 8.97156969e-02 2.15497941e-01
2.71749705e-01 2.95696557e-01 -8.31405878e-01 4.17757332e-01
1.54734969e-01 1.42757607e+00 -6.10371172e-01 2.22907159e-02
-3.45108211e-01 1.00656533e+00 5.08376837e-01 2.26096943e-01
-8.51093769e-01 4.37420934e-01 8.88721228e-01 3.48946989e-01
3.94139230e-01 -1.77037328e-01 3.99847388e-01 -1.32879424e+00
3.64602625e-01 -8.13743949e-01 3.52437347e-01 -8.83944571e-01
-1.19375229e+00 1.05352151e+00 1.62252724e-01 -1.45093763e+00
-5.98315358e-01 -4.66643155e-01 -3.99272889e-01 3.78808767e-01
-1.19281983e+00 -1.31827450e+00 -3.80064279e-01 1.12518668e+00
7.21740961e-01 -9.29309875e-02 1.08168828e+00 1.46350674e-02
-5.71857691e-01 6.11776412e-01 -2.44140431e-01 -2.86427122e-02
6.75027013e-01 -9.46105719e-01 -1.63779557e-01 5.65558434e-01
4.26356196e-01 2.38102466e-01 6.99075937e-01 -3.85919690e-01
-1.33031881e+00 -9.48407531e-01 6.85205996e-01 -4.35315728e-01
5.33945799e-01 -4.18865323e-01 -6.68791890e-01 5.24353027e-01
-2.21726060e-01 5.30034423e-01 8.43115389e-01 1.88629150e-01
-3.74892175e-01 -3.09580833e-01 -1.11662126e+00 4.33072567e-01
1.05056238e+00 -3.83067787e-01 -4.19033468e-02 6.40258938e-02
-3.34401608e-01 -4.44697201e-01 -8.77550483e-01 5.59667766e-01
1.16775692e+00 -1.46123290e+00 6.72496319e-01 -4.16650444e-01
4.45074350e-01 -1.76924974e-01 -4.71789241e-01 -1.20717585e+00
-8.50327834e-02 -7.68257678e-01 1.03569172e-01 1.29712951e+00
7.52337500e-02 -1.87796205e-01 8.75368774e-01 7.24874496e-01
4.95131552e-01 -1.02416754e+00 -1.17307663e+00 -4.28342342e-01
-2.10989609e-01 -5.92621744e-01 4.07941282e-01 7.62323260e-01
-1.20106041e-01 1.85410917e-01 -5.11685431e-01 1.86619446e-01
6.68643713e-01 2.90267617e-01 1.12948346e+00 -1.31315100e+00
-2.28209555e-01 -1.49151772e-01 -9.16726589e-01 -7.73515522e-01
7.11853981e-01 -3.98228765e-01 1.76590696e-01 -8.00878406e-01
3.08543980e-01 -2.34902889e-01 1.16661668e-01 3.04501951e-01
-7.31724650e-02 3.89188290e-01 1.00624561e-01 1.01489447e-01
-4.91970420e-01 6.55827522e-01 1.12139988e+00 2.89757587e-02
-1.26421094e-01 4.62634385e-01 5.97110018e-04 1.09816456e+00
3.95095348e-01 -3.09786707e-01 -1.87676400e-01 8.47284421e-02
-1.51288539e-01 6.79270208e-01 4.06296045e-01 -8.92844379e-01
-7.32101947e-02 -4.23099548e-01 7.30968118e-01 -4.35039997e-01
8.60274434e-01 -8.14182103e-01 5.21319628e-01 -7.78173730e-02
-2.38511227e-02 2.17930168e-01 -7.21547008e-02 6.47143185e-01
-5.14103293e-01 3.13947201e-01 8.82133007e-01 -1.28488228e-01
-7.10264444e-01 6.46978259e-01 -5.16609550e-01 -1.82601959e-01
1.13256645e+00 -4.25222397e-01 3.29629362e-01 -5.88739157e-01
-1.13446593e+00 -1.98851824e-01 4.81637836e-01 4.81580794e-01
5.66109896e-01 -1.38917959e+00 -5.29095113e-01 3.81524086e-01
2.30937630e-01 6.10286295e-02 5.12288809e-01 8.62323225e-01
3.78481075e-02 -5.57600781e-02 -5.21928489e-01 -1.09775925e+00
-2.06030178e+00 9.91783217e-02 3.06513488e-01 -2.37896621e-01
-2.84930885e-01 9.23702717e-01 2.89016515e-01 -1.89814940e-01
6.09084181e-02 -1.61862466e-02 -5.59065998e-01 5.38056970e-01
4.46860194e-01 1.23735413e-01 2.61137486e-02 -1.39455414e+00
-4.55144048e-01 1.04542232e+00 1.27903372e-01 -2.14567155e-01
1.54710102e+00 -1.93234369e-01 -1.24235384e-01 8.68047774e-01
1.39706242e+00 8.05037469e-02 -1.43372107e+00 -1.50422156e-01
-1.29074827e-01 -7.12177455e-01 -2.84102112e-01 -4.03325021e-01
-1.34524810e+00 8.04137826e-01 6.51815951e-01 -2.70660967e-01
1.49849963e+00 4.98910360e-02 2.13054270e-01 -1.23781618e-02
8.22167218e-01 -7.77540863e-01 5.07234968e-02 1.19494610e-01
1.10704517e+00 -1.15944672e+00 1.23629220e-01 -5.86718261e-01
-8.96758556e-01 1.29496229e+00 5.13981938e-01 1.31298929e-01
1.00733924e+00 2.61017472e-01 1.09532788e-01 -1.85474634e-01
-9.64759290e-01 -8.30523893e-02 2.86199510e-01 6.13601387e-01
4.27094877e-01 1.06590375e-01 1.69447154e-01 3.19791615e-01
-4.40336615e-01 2.37760291e-01 2.44073182e-01 8.12021494e-01
4.35578637e-04 -1.08738577e+00 -3.43998492e-01 2.61372387e-01
-4.38883573e-01 5.09679317e-01 -2.83826709e-01 7.63416171e-01
4.07218456e-01 6.60764813e-01 5.70653304e-02 -5.48864126e-01
3.66239667e-01 3.23513329e-01 8.67715538e-01 -3.66265059e-01
-2.02761099e-01 2.45901316e-01 1.20271318e-01 -8.57933521e-01
-9.53153849e-01 -1.51822793e+00 -7.69248545e-01 8.58401358e-02
-2.04180419e-01 -2.21227467e-01 6.40004277e-01 9.47511613e-01
1.09523669e-01 -1.26640201e-01 9.40565348e-01 -1.30019259e+00
-1.07682042e-01 -7.62782097e-01 -6.40670180e-01 6.27845526e-01
4.79617864e-01 -1.19846559e+00 -2.96711504e-01 5.17900825e-01] | [13.579751968383789, 1.565171241760254] |
d90f1792-f194-4ac3-b5f2-8b0c4062916e | exclusive-topic-modeling | 2102.03525 | null | https://arxiv.org/abs/2102.03525v1 | https://arxiv.org/pdf/2102.03525v1.pdf | Exclusive Topic Modeling | We propose an Exclusive Topic Modeling (ETM) for unsupervised text classification, which is able to 1) identify the field-specific keywords though less frequently appeared and 2) deliver well-structured topics with exclusive words. In particular, a weighted Lasso penalty is imposed to reduce the dominance of the frequently appearing yet less relevant words automatically, and a pairwise Kullback-Leibler divergence penalty is used to implement topics separation. Simulation studies demonstrate that the ETM detects the field-specific keywords, while LDA fails. When applying to the benchmark NIPS dataset, the topic coherence score on average improves by 22% and 10% for the model with weighted Lasso penalty and pairwise Kullback-Leibler divergence penalty, respectively. | ['Ying Chen', 'Hao Lei'] | 2021-02-06 | null | null | null | null | ['unsupervised-text-classification'] | ['natural-language-processing'] | [ 5.80671057e-03 1.93574101e-01 -5.01001954e-01 -3.34261179e-01
-9.70997453e-01 -3.43606502e-01 5.69393933e-01 6.96197748e-01
-4.10584867e-01 6.27523839e-01 1.89020529e-01 -5.29401936e-02
-4.67913687e-01 -6.04691625e-01 -3.79901767e-01 -8.21227491e-01
-4.44296330e-01 6.86081350e-01 2.53358763e-02 3.23076636e-01
2.20382899e-01 -3.77934501e-02 -1.33343458e+00 -2.15926781e-01
1.35866630e+00 9.92396951e-01 4.40414846e-01 1.32889092e-01
-1.73433378e-01 2.79685080e-01 -4.50809956e-01 -9.52044129e-02
6.38177916e-02 -9.67635140e-02 -4.65515822e-01 1.43345505e-01
4.54717129e-01 -2.01660141e-01 -2.11047664e-01 1.21491873e+00
-8.54355021e-05 3.04073483e-01 9.60907936e-01 -1.30873823e+00
-8.20955932e-02 8.27567518e-01 -9.61792409e-01 -9.61274654e-02
1.23470344e-01 -4.74344224e-01 1.44581878e+00 -1.04773748e+00
5.59221327e-01 1.31326926e+00 4.71075028e-01 -5.82823902e-02
-1.39349782e+00 -8.61455858e-01 4.70685035e-01 -1.04053803e-01
-1.62847400e+00 1.60468370e-01 8.11445117e-01 -5.26818335e-01
4.60438550e-01 2.25378171e-01 1.58042625e-01 8.55719864e-01
5.52777827e-01 8.70569229e-01 7.11972177e-01 -3.55100751e-01
3.94350797e-01 4.28076804e-01 7.20377386e-01 3.54224414e-01
5.68344235e-01 -5.39220929e-01 -8.06438863e-01 -6.84395194e-01
2.10651942e-02 2.55303383e-01 -1.96044371e-01 -3.38611722e-01
-1.15836716e+00 1.11533821e+00 -1.26435399e-01 7.02399909e-02
-6.19131446e-01 -2.47409299e-01 3.70703816e-01 -7.04653114e-02
1.07800162e+00 2.84570098e-01 -2.79635340e-01 -9.52132121e-02
-1.27413511e+00 5.01261234e-01 7.96154737e-01 1.11138690e+00
9.00931656e-01 -2.61922538e-01 -5.15699804e-01 1.05853605e+00
5.69763780e-01 7.01254129e-01 3.60494554e-01 -7.88830698e-01
5.99289477e-01 6.64592266e-01 1.84673741e-01 -1.21575534e+00
-2.17572659e-01 -4.33472693e-01 -8.76803994e-01 -4.50381875e-01
-1.72573123e-02 -1.73179552e-01 -7.35604882e-01 1.69163048e+00
2.36372054e-01 4.16775607e-02 -9.22579691e-02 7.34101295e-01
4.75518674e-01 9.05662000e-01 1.99169725e-01 -6.33617222e-01
1.47871959e+00 -5.47532380e-01 -1.07217276e+00 -1.95527092e-01
5.80239058e-01 -5.71822524e-01 1.02011526e+00 3.95841688e-01
-8.40522110e-01 -1.88114822e-01 -9.34039831e-01 2.14676380e-01
-2.22669497e-01 1.98661268e-01 4.46698070e-01 4.00540769e-01
-6.00648344e-01 3.02507520e-01 -7.98684597e-01 -3.12167972e-01
3.24497581e-01 1.99546695e-01 -2.46275708e-01 -3.13562229e-02
-1.28034294e+00 2.18851298e-01 5.73688269e-01 -3.95271271e-01
-8.47272456e-01 -7.43659437e-01 -7.45430410e-01 5.31500459e-01
4.08947319e-01 -3.28747690e-01 7.19686925e-01 -2.68747419e-01
-1.07645762e+00 5.49914539e-01 -4.28109109e-01 -6.97711587e-01
3.22719991e-01 -5.21960557e-01 -1.53848544e-01 4.03432816e-01
5.59172750e-01 5.00157714e-01 7.46053934e-01 -1.10504222e+00
-7.84751356e-01 -5.20414591e-01 -4.27032739e-01 2.81065851e-01
-1.14385927e+00 -7.49165267e-02 -5.68885267e-01 -1.06890702e+00
5.52439332e-01 -8.66270542e-01 -8.84435102e-02 -3.15502286e-01
-5.93096197e-01 -5.73694706e-01 9.96516824e-01 -5.15651822e-01
1.51908243e+00 -2.23437166e+00 1.66721605e-02 3.19454938e-01
5.47647774e-01 -2.34817877e-01 1.91525191e-01 4.23310399e-01
3.18131238e-01 -7.82304630e-02 -3.44230354e-01 -4.81408626e-01
-1.22918479e-01 -3.21699865e-02 -7.46493518e-01 4.15443718e-01
-1.31469697e-01 1.13435589e-01 -8.91200423e-01 -6.10381842e-01
-3.59363444e-02 1.17416382e-01 -5.88803649e-01 3.23159844e-02
-1.96813658e-01 -1.64148614e-01 -6.62430942e-01 4.03076023e-01
7.89260983e-01 -4.31990832e-01 3.52055341e-01 1.00241281e-01
-1.09359801e-01 4.01752889e-01 -1.28635788e+00 1.31087816e+00
-2.10198253e-01 7.31218636e-01 2.20256969e-01 -8.90418887e-01
1.16589272e+00 3.35067183e-01 7.05196559e-01 -1.35798320e-01
-7.06061348e-02 1.17932059e-01 -4.94347453e-01 -1.31178111e-01
8.24524164e-01 9.32215005e-02 -8.12790990e-02 4.42301750e-01
8.43499787e-03 1.56428739e-01 3.25326532e-01 7.14291573e-01
7.21064925e-01 -4.77261901e-01 7.26903602e-02 -8.80871892e-01
3.48756254e-01 4.08134796e-02 8.05120230e-01 9.32486773e-01
6.32718131e-02 4.38372165e-01 6.16327763e-01 -2.27010548e-02
-6.64764404e-01 -1.12915039e+00 -4.95863497e-01 1.25081766e+00
3.45977157e-01 -6.97957277e-01 -5.62731206e-01 -7.62540519e-01
3.49377513e-01 1.05812097e+00 -4.03870940e-01 -2.05238104e-01
-7.61322305e-02 -7.12058485e-01 2.72687525e-01 2.25945741e-01
1.44257531e-01 -4.22172755e-01 -3.17447186e-01 1.93041801e-01
-2.51269311e-01 -9.20120656e-01 -5.29450774e-01 4.11932021e-01
-9.49333608e-01 -7.43724167e-01 -9.72683966e-01 -6.39359832e-01
6.86045289e-01 3.81532639e-01 5.79069078e-01 -6.17016912e-01
5.68287484e-02 1.48545772e-01 -2.59536892e-01 -3.32527250e-01
9.83265694e-03 2.71998942e-01 2.96723843e-01 -9.31830052e-03
5.26542902e-01 -4.54673767e-01 -4.43793297e-01 4.30459499e-01
-6.96239710e-01 -1.45225033e-01 2.11285710e-01 1.04760981e+00
4.75208431e-01 3.67057920e-01 5.90574324e-01 -1.01749063e+00
6.16076231e-01 -5.73336124e-01 -6.28416061e-01 6.29804060e-02
-1.11931419e+00 -1.40507817e-01 3.22570920e-01 -5.13359070e-01
-1.00605440e+00 -3.48960668e-01 5.45224071e-01 -4.95626956e-01
2.98596341e-02 6.26268685e-01 3.82770188e-02 5.71465492e-01
3.27148020e-01 4.50472325e-01 -3.22326869e-01 -6.37823999e-01
1.10963687e-01 1.04363561e+00 2.40964100e-01 -6.85909808e-01
5.44036150e-01 6.59920812e-01 -2.10983202e-01 -1.01587296e+00
-9.21188653e-01 -9.52134550e-01 -3.80512446e-01 6.58772141e-02
6.54084504e-01 -1.11671865e+00 -4.36411560e-01 2.62389928e-01
-1.03200233e+00 1.12346649e-01 -7.44134113e-02 9.54832435e-01
-3.46630931e-01 5.44867635e-01 -5.53600550e-01 -1.01234984e+00
-5.96880078e-01 -8.52940261e-01 1.19755828e+00 2.75558047e-02
-5.36943495e-01 -8.70675385e-01 -1.00006536e-01 1.13396160e-01
-3.83303594e-03 -2.37879246e-01 1.14019370e+00 -1.09331250e+00
-7.50219729e-03 -4.15807843e-01 -2.58555263e-01 2.78495371e-01
2.60531962e-01 -1.99490398e-01 -7.15732217e-01 -4.50636148e-01
-5.51971793e-02 1.54035658e-01 1.11949134e+00 6.68711424e-01
9.48121905e-01 -5.22662938e-01 -8.68112445e-01 2.96165884e-01
9.75282848e-01 2.17850953e-01 4.30253036e-02 1.93329155e-01
5.01114070e-01 7.56183147e-01 1.00357676e+00 8.57124090e-01
3.61167490e-01 4.92287517e-01 8.10101181e-02 3.21234971e-01
4.18149352e-01 -3.56186539e-01 2.19610661e-01 8.69323492e-01
5.99179745e-01 -4.43642348e-01 -1.05276704e+00 7.29100287e-01
-1.91494775e+00 -4.52133030e-01 -2.29941174e-01 2.40381479e+00
9.43675816e-01 3.43276173e-01 2.34858952e-02 1.20142013e-01
9.69111502e-01 1.81616828e-01 -3.59383672e-01 1.55190423e-01
-2.75122244e-02 -3.55644166e-01 3.80009741e-01 5.39138377e-01
-1.29255700e+00 8.94831836e-01 6.06680870e+00 1.06846833e+00
-8.13738406e-01 -7.00356113e-03 5.78821599e-01 -5.30378520e-02
-2.47516260e-01 1.19093649e-01 -1.07426393e+00 6.99914694e-01
6.29173994e-01 -6.31959975e-01 -3.21972221e-01 1.26322985e+00
3.85756165e-01 -3.39244634e-01 -8.54403436e-01 7.95181513e-01
-6.72243396e-03 -8.66421402e-01 2.17836380e-01 2.41915271e-01
8.77345264e-01 -1.97207928e-01 8.67742822e-02 1.05269417e-01
2.61413991e-01 -4.17054266e-01 4.86325204e-01 3.71380419e-01
5.53552806e-01 -9.01501179e-01 6.41697943e-01 4.90057021e-01
-9.70260620e-01 -8.71639475e-02 -4.39342439e-01 1.95532754e-01
2.01748088e-02 1.36156583e+00 -8.47786665e-01 4.66316104e-01
7.79034793e-01 7.90862858e-01 -1.15410112e-01 9.12570298e-01
-4.83447313e-02 1.01505661e+00 -6.29559219e-01 -1.81504905e-01
2.56325483e-01 -4.94198948e-01 1.15847504e+00 1.26063693e+00
2.81189084e-01 -1.15891732e-01 3.96699458e-01 8.37606668e-01
-1.65658787e-01 3.50437731e-01 -2.27076575e-01 -7.09697530e-02
7.82187700e-01 1.15044475e+00 -8.07555318e-01 -5.72338998e-01
1.28772363e-01 7.70020068e-01 -1.21125586e-01 6.13858879e-01
-3.66900235e-01 -6.98321640e-01 5.42617142e-01 1.56624645e-01
2.80320793e-01 -1.70292452e-01 -4.43404734e-01 -1.09814775e+00
5.44224977e-02 -4.59263474e-01 6.12725019e-01 -4.30394888e-01
-1.28317297e+00 2.66123265e-01 2.92195439e-01 -1.00215948e+00
-8.60727727e-02 -1.97070107e-01 -5.65865040e-01 8.78778100e-01
-1.14562869e+00 -6.47617221e-01 -1.20015129e-01 8.98315012e-02
7.51203418e-01 -3.15412968e-01 5.65623939e-01 1.56335443e-01
-6.72691107e-01 5.26712179e-01 7.33552635e-01 -1.83622122e-01
9.25452471e-01 -1.29674184e+00 -7.23010674e-02 4.11032140e-01
-9.61451605e-02 9.63360727e-01 9.92798924e-01 -8.77223194e-01
-1.11214125e+00 -1.20314527e+00 1.16260958e+00 4.48196242e-03
6.12610757e-01 -5.82539082e-01 -1.13763011e+00 5.15750110e-01
-5.50648849e-03 -6.14419281e-01 6.47213876e-01 5.54552913e-01
-4.50210035e-01 -2.86673009e-01 -9.82912242e-01 5.01918077e-01
2.94092000e-01 -2.97150999e-01 -6.54353738e-01 7.41246939e-01
8.83131146e-01 1.66397259e-01 -8.74013960e-01 5.06639957e-01
5.37071526e-01 -4.19848442e-01 7.18689501e-01 -4.44860131e-01
3.20538938e-01 -4.07791808e-02 -2.76402473e-01 -1.02200937e+00
-2.96607435e-01 -7.31391668e-01 -3.39625895e-01 1.63903213e+00
3.68619055e-01 -6.74243510e-01 8.38777244e-01 2.67461628e-01
3.36095273e-01 -8.48711908e-01 -1.00649130e+00 -8.48803580e-01
2.31691264e-02 -4.53584015e-01 3.21536697e-02 9.21161056e-01
3.39194059e-01 3.41032624e-01 -2.97140002e-01 1.33513123e-01
8.51083398e-01 1.57201946e-01 4.72012341e-01 -1.59714377e+00
7.18457326e-02 -4.57349688e-01 -1.14711985e-01 -1.28007483e+00
2.70576447e-01 -6.32798254e-01 2.57524371e-01 -1.35305655e+00
5.78877568e-01 -3.75893384e-01 -4.49548483e-01 2.78573334e-01
-3.93139392e-01 -3.70365620e-01 1.78506761e-03 5.44115186e-01
-7.42900252e-01 8.79006982e-01 4.62189972e-01 -2.95708895e-01
-3.99148673e-01 2.51930296e-01 -5.27546525e-01 9.34512436e-01
5.07264793e-01 -8.64120126e-01 -3.29931438e-01 -5.05000800e-02
1.35944635e-01 -1.78936601e-01 1.72928870e-02 -7.11519301e-01
3.22102129e-01 -1.03234217e-01 1.70696136e-02 -1.28821993e+00
2.11244524e-01 -5.65069437e-01 -4.75947022e-01 3.12793523e-01
-7.12785721e-01 -5.39070785e-01 1.25604063e-01 9.69231784e-01
-2.07738161e-01 -2.42355511e-01 5.17357826e-01 3.84831637e-01
-1.02148570e-01 9.09990817e-02 -6.30854070e-01 2.32506990e-02
9.39499319e-01 1.48855478e-01 1.84471250e-01 -4.59674716e-01
-4.45353687e-01 6.85007036e-01 2.99683101e-02 3.56946498e-01
4.31776524e-01 -1.16250908e+00 -8.76447380e-01 -1.60293709e-02
2.39724725e-01 4.05600620e-03 2.14938998e-01 6.69051647e-01
1.15942299e-01 7.28953838e-01 5.27192891e-01 -9.23426390e-01
-1.17745924e+00 3.43903750e-01 -3.11843961e-01 -3.89001727e-01
-6.03131652e-01 8.56619716e-01 5.50562322e-01 -2.04096779e-01
6.17884099e-01 -5.29432148e-02 -2.03195304e-01 4.50607240e-01
3.57134581e-01 7.59824574e-01 -1.03089727e-01 -3.11107844e-01
-4.48630333e-01 4.61488396e-01 -5.66062689e-01 -2.12719455e-01
1.23822331e+00 -3.16532880e-01 -2.23729014e-01 6.69907272e-01
1.21141684e+00 -7.71172047e-02 -9.99709010e-01 -4.94230807e-01
3.50524157e-01 -2.56500512e-01 3.47260535e-01 -4.59915042e-01
-5.34713745e-01 8.75748396e-01 3.46431911e-01 4.82212961e-01
1.05577612e+00 1.40097052e-01 7.09613085e-01 4.26932037e-01
6.00156523e-02 -1.25909972e+00 -7.70721287e-02 5.17661393e-01
6.13927841e-01 -1.01018906e+00 2.75431633e-01 -7.46500194e-01
-5.77962399e-01 8.78588438e-01 3.18946511e-01 -1.31702721e-02
8.44631493e-01 6.24082722e-02 -3.47580999e-01 -2.74648756e-01
-7.92300403e-01 2.15935469e-01 4.99734074e-01 1.31976098e-01
2.79002964e-01 2.54685879e-02 -8.29606235e-01 8.56014192e-01
-8.60989094e-02 -4.32024628e-01 4.39929552e-02 5.45096695e-01
-7.35319972e-01 -5.65402150e-01 -4.41887528e-01 7.40610361e-01
-4.84710485e-01 -1.28225863e-01 -1.69738695e-01 3.80248070e-01
-3.19485754e-01 1.06224453e+00 2.99676031e-01 -1.05830327e-01
-8.88252724e-03 2.38590866e-01 -4.28685009e-01 -7.03358889e-01
-9.43281651e-02 6.78481519e-01 -1.77370444e-01 -1.87818006e-01
5.30226640e-02 -6.69471622e-01 -1.15530944e+00 8.54898915e-02
-6.77860081e-01 7.07750320e-01 6.66479707e-01 8.10938895e-01
6.82994545e-01 3.32507044e-01 9.40929234e-01 -2.74862260e-01
-7.91251719e-01 -1.26336670e+00 -8.98196399e-01 3.25648516e-01
1.23991340e-01 -7.40646124e-01 -6.30292475e-01 6.64885342e-02] | [10.382752418518066, 6.917177677154541] |
822d611c-a7a7-42c2-b94e-ad0851987616 | dynamic-community-detection-into-analyzing-of | 2011.01140 | null | https://arxiv.org/abs/2011.01140v1 | https://arxiv.org/pdf/2011.01140v1.pdf | Dynamic Community Detection into Analyzing of Wildfires Events | The study and comprehension of complex systems are crucial intellectual and scientific challenges of the 21st century. In this scenario, network science has emerged as a mathematical tool to support the study of such systems. Examples include environmental processes such as wildfires, which are known for their considerable impact on human life. However, there is a considerable lack of studies of wildfire from a network science perspective. Here, employing the chronological network concept -- a temporal network where nodes are linked if two consecutive events occur between them -- we investigate the information that dynamic community structures reveal about the wildfires' dynamics. Particularly, we explore a two-phase dynamic community detection approach, i.e., we applied the Louvain algorithm on a series of snapshots. Then we used the Jaccard similarity coefficient to match communities across adjacent snapshots. Experiments with the MODIS dataset of fire events in the Amazon basing were conducted. Our results show that the dynamic communities can reveal wildfire patterns observed throughout the year. | ['Marcos G Quiles', 'Elbert EN Macau', 'Leonardo N Ferreira', 'Moshé Cotacallapa', 'Didier A Vega-Oliveros', 'Alessandra Marli'] | 2020-11-02 | null | null | null | null | ['dynamic-community-detection'] | ['graphs'] | [ 1.92884013e-01 -3.77453417e-01 2.95798667e-02 1.57058284e-01
6.31832421e-01 -8.80859435e-01 7.65830338e-01 6.07316256e-01
-3.07482153e-01 7.20669448e-01 2.60048807e-01 -6.11702144e-01
-5.46091855e-01 -1.27875280e+00 -1.75143719e-01 -7.23330140e-01
-1.03692496e+00 3.42505351e-02 3.76157165e-01 -4.48684096e-01
3.62588644e-01 4.23774630e-01 -1.40170455e+00 -5.88913960e-03
7.51686275e-01 2.99393386e-01 3.14233184e-01 5.61588466e-01
-1.84934571e-01 3.73024881e-01 -4.20623630e-01 1.09878466e-01
1.68631211e-01 -4.09912944e-01 -6.25667810e-01 -1.60053104e-01
-1.85941294e-01 5.55813946e-02 -3.45496386e-01 1.07922530e+00
1.34213254e-01 7.06868768e-02 2.68059820e-01 -1.44027364e+00
-1.53632209e-01 7.12435424e-01 -8.86051834e-01 6.14669383e-01
3.55854452e-01 7.62966424e-02 1.12071967e+00 -6.31725669e-01
9.38367009e-01 1.35594785e+00 7.96773851e-01 -2.99725980e-01
-1.40947938e+00 -7.67699897e-01 1.43772677e-01 4.33329254e-01
-1.40019464e+00 -1.52831703e-01 5.90968132e-01 -9.25791204e-01
5.98872185e-01 3.41486990e-01 1.23023629e+00 6.47435725e-01
2.70227671e-01 3.81744504e-02 1.26702476e+00 -2.22065106e-01
1.82145521e-01 -6.47015095e-01 9.88420323e-02 3.58877122e-01
7.87134051e-01 3.20298374e-01 -3.62570107e-01 -4.09820467e-01
7.67375171e-01 4.87353861e-01 -2.84620792e-01 -6.97841644e-02
-1.44431210e+00 8.34898114e-01 7.68496394e-01 6.67143643e-01
-3.87857646e-01 -5.55510521e-02 4.09116417e-01 4.14611965e-01
5.85819781e-01 2.81077027e-01 4.96703126e-02 1.93515822e-01
-9.85080481e-01 1.95046902e-01 8.97948444e-01 3.03202331e-01
6.86458647e-01 -5.11152387e-01 4.87554193e-01 4.06587422e-01
4.80915189e-01 6.19431436e-01 -1.87628835e-01 -7.52572715e-01
2.35801190e-01 8.44851851e-01 -2.57860646e-02 -1.80877686e+00
-4.38714117e-01 -3.31258804e-01 -1.51472580e+00 2.30130389e-01
3.88245612e-01 1.03908349e-02 -3.22495788e-01 1.52808368e+00
4.83988464e-01 4.79012698e-01 -1.21157281e-01 7.85184026e-01
2.98082411e-01 6.14493906e-01 -4.83799279e-02 -6.97409272e-01
1.21270847e+00 -3.23188961e-01 -7.57712364e-01 2.28104472e-01
-1.10726424e-01 -6.13369465e-01 5.62625766e-01 2.73865927e-02
-5.62856257e-01 -1.42871007e-01 -8.13662231e-01 9.71683383e-01
-5.63118041e-01 -6.54329479e-01 5.79333603e-01 1.84521765e-01
-1.11626053e+00 8.25981677e-01 -8.29427540e-01 -1.13481402e+00
2.97910005e-01 -3.16831082e-01 -2.36643568e-01 -4.21685465e-02
-1.16755128e+00 6.84861004e-01 4.60628033e-01 5.55202305e-01
-7.57645130e-01 -4.90413755e-01 -2.28284627e-01 1.03629760e-01
5.33819795e-01 -5.97211480e-01 6.21471703e-01 -7.94629693e-01
-5.29380023e-01 6.60361588e-01 -5.62628172e-02 -2.67305136e-01
4.55318063e-01 2.13018820e-01 -6.21891618e-01 1.75548404e-01
2.72894830e-01 -1.47904336e-01 4.61213619e-01 -1.30719209e+00
-5.11248946e-01 -3.48011702e-01 1.29596055e-01 1.00668091e-02
-2.30971366e-01 1.40976682e-01 4.07557547e-01 -7.52815306e-01
4.76879597e-01 -9.42780674e-01 -6.13183379e-01 2.32937649e-01
-2.82536775e-01 9.99048725e-02 6.83768392e-01 -4.16814625e-01
1.68076372e+00 -2.08661151e+00 1.54166251e-01 6.44293070e-01
7.89115131e-01 1.05214022e-01 1.58755537e-02 1.27841544e+00
2.63817851e-02 5.84085822e-01 -6.54863477e-01 4.42854524e-01
-4.43475187e-01 4.48402643e-01 -2.93010414e-01 4.17833149e-01
4.22583893e-02 2.51841336e-01 -1.36039090e+00 -3.09881151e-01
-1.07907295e-01 1.21826597e-01 -2.41349563e-01 5.52583933e-02
-3.64195742e-02 4.36883330e-01 -3.59356493e-01 4.72294956e-01
6.66622937e-01 -3.35813314e-01 8.43520820e-01 1.20948300e-01
-7.25149930e-01 -1.89819261e-01 -1.16061139e+00 1.22483957e+00
1.50027677e-01 7.98855305e-01 3.88466150e-01 -9.20058966e-01
7.90020168e-01 3.08514237e-01 5.62169909e-01 -1.90283895e-01
-2.89326578e-01 1.45569459e-01 4.89978641e-01 -7.61816502e-01
1.65259138e-01 -1.81355998e-01 2.72161543e-01 8.19697380e-01
-6.79071188e-01 3.30740660e-01 6.03632271e-01 3.70749593e-01
1.46330500e+00 -2.89468795e-01 6.74044907e-01 -5.38648367e-01
3.66363943e-01 4.13126200e-01 4.42408353e-01 5.73838592e-01
-3.86690348e-01 1.30107507e-01 5.80320179e-01 -4.81202573e-01
-1.15289998e+00 -1.24132490e+00 8.73131976e-02 6.51789606e-01
1.86366022e-01 -5.05581141e-01 -1.62296787e-01 -1.91593766e-02
3.32020223e-01 1.05273342e-02 -5.32007217e-01 2.01284081e-01
-3.90920132e-01 -1.02888680e+00 3.68894815e-01 2.82102395e-02
6.17350817e-01 -1.08533943e+00 -7.34590888e-01 4.19898808e-01
-4.66354668e-01 -5.83257258e-01 -1.59127295e-01 -2.19973043e-01
-1.22921932e+00 -1.65219343e+00 -3.87788534e-01 -5.83588004e-01
6.12060905e-01 7.29479790e-01 1.15122020e+00 4.59210545e-01
-5.59820473e-01 3.29980731e-01 -6.55554771e-01 1.16741464e-01
-3.35746020e-01 -3.49016003e-02 2.45337009e-01 -4.71283309e-02
1.52153492e-01 -1.42449379e+00 -9.05048132e-01 4.52857316e-01
-9.40514445e-01 7.16340542e-02 5.28206646e-01 4.55622882e-01
8.50119740e-02 4.50056553e-01 6.23953700e-01 -5.66599071e-01
8.19398105e-01 -1.08935964e+00 -3.59552056e-01 3.46657783e-01
-8.14624190e-01 -3.17306191e-01 2.30327129e-01 -3.75800401e-01
-6.12280726e-01 -2.92612582e-01 6.52244031e-01 6.43068701e-02
-5.81348576e-02 1.30974960e+00 2.50507683e-01 1.88960090e-01
5.61355948e-01 -1.70876049e-02 2.74457365e-01 -3.06475163e-01
3.78213942e-01 6.65658176e-01 4.37374830e-01 -4.55289096e-01
9.99931097e-01 7.71277547e-01 1.20623931e-01 -1.13967931e+00
-1.62792504e-01 -7.73278654e-01 -7.98582494e-01 -8.24050367e-01
4.81014729e-01 -7.79018581e-01 -9.26717639e-01 3.46908480e-01
-1.29198158e+00 -4.82868627e-02 9.68012214e-02 3.87677938e-01
1.32137507e-01 4.43744957e-01 -3.16179514e-01 -9.22422945e-01
-2.10065559e-01 -2.00117484e-01 1.65724903e-01 1.19020082e-01
-3.81507158e-01 -1.15703607e+00 7.82274008e-01 -3.12774628e-01
6.13179266e-01 9.62609649e-01 8.63023341e-01 -2.04109907e-01
-6.72593057e-01 2.29441881e-01 -4.25656497e-01 -2.81992078e-01
4.59449172e-01 5.52087426e-01 -5.77093184e-01 -3.68395507e-01
-1.58109322e-01 7.88682818e-01 5.81605732e-01 1.73771873e-01
3.83518398e-01 -5.65744400e-01 -6.67840481e-01 7.33794868e-02
1.67051864e+00 2.42387339e-01 5.30777216e-01 5.04186332e-01
4.96700019e-01 9.70605433e-01 3.44163597e-01 6.61057293e-01
2.31486827e-01 1.60216182e-01 6.08932316e-01 1.33531420e-02
7.94965178e-02 -6.80519938e-02 1.92493945e-01 1.20408177e+00
-5.82967341e-01 -7.47364312e-02 -1.46713352e+00 7.57974207e-01
-2.10083818e+00 -1.59046555e+00 -6.96788311e-01 2.18187666e+00
6.02915943e-01 -5.83227463e-02 2.96936423e-01 1.57904774e-01
1.21086192e+00 5.57536960e-01 -2.96323568e-01 2.07921043e-01
-2.58516520e-01 3.85018364e-02 2.59496659e-01 4.31346029e-01
-8.60984206e-01 5.65725207e-01 6.42167568e+00 1.53280511e-01
-7.81465709e-01 4.34090495e-02 2.36332347e-03 1.94342062e-01
-4.36909884e-01 4.54334438e-01 -1.38629684e-02 2.93517083e-01
7.81471789e-01 -6.75007582e-01 5.95582128e-01 1.87237963e-01
9.57823336e-01 -4.45697814e-01 -4.95043844e-01 4.68838602e-01
-4.91784155e-01 -1.20450282e+00 -1.83319166e-01 3.58905941e-01
5.71601808e-01 7.76346028e-02 -5.10783017e-01 -2.80671746e-01
5.87814093e-01 -7.45111763e-01 3.37985516e-01 8.48035276e-01
5.28224826e-01 -3.36659044e-01 2.27120236e-01 3.91354173e-01
-1.77199459e+00 -1.44634262e-01 -1.67499557e-01 -5.86441040e-01
3.94381523e-01 1.19625723e+00 -6.53589606e-01 6.55463338e-01
9.41247523e-01 1.29088402e+00 -3.59075248e-01 1.46294141e+00
-1.77786231e-01 8.11110854e-01 -3.52804363e-01 1.62841585e-02
-4.64306120e-03 -9.33111966e-01 9.36809301e-01 7.79702663e-01
4.55282718e-01 2.63322175e-01 1.96983635e-01 9.11141813e-01
3.27582002e-01 -1.69469357e-01 -1.16241586e+00 -3.91116858e-01
8.33030164e-01 1.31073356e+00 -1.07180595e+00 -7.62259066e-02
-1.11205205e-01 4.39661086e-01 -7.95786232e-02 3.89382094e-01
-3.80416691e-01 -3.51956278e-01 9.37646985e-01 4.64202911e-01
1.13790771e-02 -8.39484811e-01 -6.04676902e-02 -1.01637852e+00
-1.57195032e-01 -6.33653402e-01 3.17240506e-01 -7.02763319e-01
-1.56887925e+00 3.27479362e-01 1.72512919e-01 -1.36676216e+00
1.56537384e-01 -4.40753251e-02 -1.11487508e+00 7.29284406e-01
-1.39046037e+00 -7.49420166e-01 -5.69674790e-01 5.17983675e-01
-6.91018626e-02 2.06157207e-01 7.63028145e-01 2.85135359e-01
-5.18915296e-01 -6.18837237e-01 6.13252044e-01 2.22129021e-02
3.47888857e-01 -1.04065037e+00 5.16776145e-01 1.07217133e+00
-1.80620775e-01 8.09111416e-01 7.24500000e-01 -1.11946857e+00
-1.08027267e+00 -8.99991632e-01 1.09155619e+00 3.66648622e-02
1.45519078e+00 -3.11537415e-01 -9.08844233e-01 4.11165446e-01
3.24794203e-01 -1.41668588e-01 6.57984912e-01 2.16063961e-01
-3.29199702e-01 -1.22561842e-01 -9.42949653e-01 8.09155703e-01
1.47251892e+00 -5.01405716e-01 -4.99538690e-01 2.10827440e-01
4.67266202e-01 6.63625181e-01 -1.09666085e+00 1.51851147e-01
8.32714975e-01 -8.46422255e-01 8.92914116e-01 -4.36646730e-01
3.51338238e-01 -7.39032388e-01 -1.18674591e-01 -1.45726860e+00
-5.03393352e-01 -5.64833999e-01 2.43197262e-01 1.20582700e+00
-7.14425743e-03 -9.53093886e-01 2.83710152e-01 -2.92133570e-01
4.19852525e-01 -5.61208427e-02 -7.73874104e-01 -8.97659779e-01
-2.10902825e-01 1.39385790e-01 5.16386747e-01 1.35496294e+00
3.31320614e-01 1.61752611e-01 2.29043085e-02 2.57800192e-01
8.14095914e-01 2.67525524e-01 5.44493854e-01 -1.88795757e+00
1.02496959e-01 -6.90405011e-01 -3.71817261e-01 -1.84246510e-01
-2.69427538e-01 -8.62062812e-01 -6.07890896e-02 -1.43540275e+00
3.50590706e-01 -5.70008755e-01 -8.00722018e-02 3.38092089e-01
-2.11143568e-01 -1.59771577e-01 1.96905151e-01 8.11953902e-01
-4.12360355e-02 3.23722750e-01 9.72660184e-01 -1.37919039e-01
-7.58669004e-02 -1.01279162e-01 -2.11009935e-01 4.51866388e-01
8.54428232e-01 -8.13020170e-01 -2.38527238e-01 6.60861582e-02
7.75282621e-01 1.45013466e-01 5.29027760e-01 -9.23728943e-01
3.05308670e-01 -5.45259535e-01 -4.38809752e-01 -4.48546112e-01
-4.31490451e-01 -1.10208070e+00 9.55692232e-01 9.06856060e-01
-2.17976123e-02 5.74109733e-01 -2.22223159e-02 1.03102112e+00
-2.97138840e-01 8.33944753e-02 2.63323456e-01 -2.68202275e-01
-5.49826324e-01 3.26366812e-01 -8.16772461e-01 -6.27806336e-02
1.03294408e+00 6.24691620e-02 -6.90066338e-01 -1.77228436e-01
-8.46760690e-01 5.08311749e-01 6.68093026e-01 2.71738678e-01
4.92239505e-01 -1.29746282e+00 -9.32271838e-01 -9.79383886e-02
6.41653091e-02 -2.39945069e-01 -4.80969576e-03 1.15763402e+00
-8.82368624e-01 -1.46790311e-01 -3.95756632e-01 -6.69778109e-01
-1.45585907e+00 4.46207523e-01 2.59389281e-01 -3.02788496e-01
-6.08058155e-01 1.35112703e-01 -3.09047401e-01 -3.01003397e-01
-1.69871345e-01 -8.86955485e-02 -3.96140784e-01 7.23488152e-01
4.22800839e-01 5.12860239e-01 -4.84644413e-01 -4.82951403e-01
-5.63578069e-01 5.70017219e-01 5.85799217e-01 -2.96512604e-01
1.70488381e+00 -4.67984051e-01 -9.87033665e-01 9.15214777e-01
6.23741627e-01 -2.94818640e-01 -7.98321247e-01 -2.27190375e-01
6.68815672e-01 -6.36618793e-01 -2.99703985e-01 -4.73862380e-01
-1.00096059e+00 6.23169005e-01 5.96303582e-01 1.02139747e+00
1.00494492e+00 -2.25811779e-01 2.78725952e-01 3.79992992e-01
4.33941692e-01 -6.01236403e-01 -1.90822870e-01 4.01577264e-01
8.39872599e-01 -9.72469687e-01 1.26428977e-01 -3.66886437e-01
-3.65199186e-02 1.20380199e+00 -2.30217092e-02 -1.75403163e-01
1.15035427e+00 1.15963351e-02 -2.35041186e-01 -4.83119518e-01
-6.41772270e-01 -4.96507674e-01 -2.72086352e-01 4.73156869e-01
3.17294985e-01 4.79131937e-01 -6.00046039e-01 -1.38454452e-01
5.40546328e-02 -4.03141156e-02 7.00767756e-01 1.16756272e+00
-7.26736486e-01 -8.70491803e-01 -6.26329780e-01 4.08180147e-01
1.83578208e-01 -2.75044084e-01 -5.95472634e-01 8.06322455e-01
-1.83486082e-02 1.14199328e+00 1.03873231e-01 -4.94194925e-01
2.31696755e-01 -1.00646786e-01 1.24185212e-01 -5.79565287e-01
-7.30242312e-01 -1.37583360e-01 -3.60682160e-02 -4.01744753e-01
-1.12150884e+00 -1.06867194e+00 -8.78717840e-01 -1.08203590e+00
-3.08535010e-01 2.67877162e-01 5.85891783e-01 7.00090289e-01
2.32533485e-01 4.42795545e-01 9.09437895e-01 -7.63079941e-01
3.10418103e-02 -1.15629303e+00 -9.65949476e-01 2.18770698e-01
3.11520755e-01 -5.42545974e-01 -6.86340809e-01 1.16603851e-01] | [7.308119297027588, 5.193799018859863] |
05041565-b490-4f33-81ee-b24e98a9318c | cogmen-contextualized-gnn-based-multimodal | 2205.02455 | null | https://arxiv.org/abs/2205.02455v1 | https://arxiv.org/pdf/2205.02455v1.pdf | COGMEN: COntextualized GNN based Multimodal Emotion recognitioN | Emotions are an inherent part of human interactions, and consequently, it is imperative to develop AI systems that understand and recognize human emotions. During a conversation involving various people, a person's emotions are influenced by the other speaker's utterances and their own emotional state over the utterances. In this paper, we propose COntextualized Graph Neural Network based Multimodal Emotion recognitioN (COGMEN) system that leverages local information (i.e., inter/intra dependency between speakers) and global information (context). The proposed model uses Graph Neural Network (GNN) based architecture to model the complex dependencies (local and global information) in a conversation. Our model gives state-of-the-art (SOTA) results on IEMOCAP and MOSEI datasets, and detailed ablation experiments show the importance of modeling information at both levels. | ['Ashutosh Modi', 'Atin Vikram Singh', 'Ayush Jain', 'Ashwani Bhat', 'Abhinav Joshi'] | 2022-05-05 | null | https://aclanthology.org/2022.naacl-main.306 | https://aclanthology.org/2022.naacl-main.306.pdf | naacl-2022-7 | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [-1.37507934e-02 1.41959310e-01 1.44695584e-02 -8.91119957e-01
5.71874417e-02 -3.84765267e-01 6.76269054e-01 1.98243320e-01
-1.92476347e-01 3.61963362e-01 7.16596067e-01 1.16427109e-01
2.50876695e-01 -5.64223945e-01 -1.08693548e-01 -2.74144977e-01
-2.80705959e-01 2.88541347e-01 -4.70614642e-01 -8.14199984e-01
-1.32733464e-01 3.79866213e-01 -1.28748369e+00 5.47460496e-01
4.53091592e-01 1.17795932e+00 -5.26559114e-01 1.08508992e+00
-4.60321546e-01 1.62550092e+00 -6.75661504e-01 -5.78162134e-01
-5.24340630e-01 -7.58577764e-01 -1.06261075e+00 3.20198797e-02
-9.12378281e-02 1.98421881e-01 -1.58604592e-01 9.05219615e-01
3.31627309e-01 5.07717729e-01 6.97781444e-01 -1.50809216e+00
-6.23710692e-01 9.08667922e-01 -2.15555444e-01 -1.14401892e-01
4.61543798e-01 -3.51465829e-02 1.03834772e+00 -4.83477443e-01
5.56140423e-01 1.46231878e+00 4.51637775e-01 7.31963098e-01
-6.43882275e-01 -3.76510501e-01 4.15776730e-01 3.89139742e-01
-9.49886262e-01 -5.19184649e-01 1.38283789e+00 -1.97419792e-01
1.36091888e+00 3.97361904e-01 7.11393476e-01 1.51994514e+00
2.92530745e-01 8.15750957e-01 8.61730099e-01 -5.00645995e-01
1.69717953e-01 1.92888424e-01 6.00304067e-01 5.65817177e-01
-6.37946069e-01 -1.94720611e-01 -1.05734360e+00 -1.28919289e-01
1.70549273e-01 -2.75156498e-01 2.17216481e-02 1.22681081e-01
-8.00485194e-01 7.89928377e-01 5.58975518e-01 5.46974301e-01
-4.88628775e-01 2.39655897e-01 7.66555965e-01 5.28165221e-01
2.96593696e-01 1.44209936e-01 -3.11667711e-01 -6.56785846e-01
-2.09773883e-01 -5.07064700e-01 1.19237518e+00 7.50510335e-01
6.72011018e-01 2.09233254e-01 6.72487691e-02 1.16109383e+00
5.16333282e-01 1.01391762e-01 2.55422115e-01 -5.57682157e-01
2.77459323e-01 1.05866230e+00 -4.42177236e-01 -1.46258473e+00
-7.62748778e-01 -1.19621888e-01 -1.18112683e+00 -1.95576414e-01
-1.27575129e-01 -6.91273451e-01 -8.39414001e-01 2.12916756e+00
2.31630445e-01 -4.73695733e-02 4.25591111e-01 7.93909490e-01
1.27922630e+00 7.85217285e-01 5.97749412e-01 -2.54494518e-01
1.28392720e+00 -9.50235784e-01 -1.18276668e+00 -6.22823954e-01
6.65598154e-01 -3.55061740e-01 7.05786705e-01 3.09962302e-01
-7.73744106e-01 -3.87141198e-01 -8.16904664e-01 -1.36151627e-01
-8.93640101e-01 -1.76879048e-01 9.68606830e-01 6.82656705e-01
-1.19910502e+00 1.69730410e-02 -4.34481055e-01 -6.33905470e-01
-6.56883838e-03 6.12458289e-01 -4.45237547e-01 2.67932296e-01
-1.70954955e+00 8.59688938e-01 2.07060754e-01 6.46421432e-01
-4.89960432e-01 2.11079568e-01 -1.09146452e+00 8.13950673e-02
1.81038633e-01 -3.75277311e-01 9.65683520e-01 -1.52340484e+00
-1.83877254e+00 5.24637640e-01 -3.37844223e-01 1.07606158e-01
-2.01809630e-01 1.37202084e-01 -9.22494113e-01 8.18614569e-03
-7.67597020e-01 7.05590367e-01 6.12289429e-01 -1.21320903e+00
-2.32705563e-01 -5.53729832e-01 1.83971561e-02 5.08456469e-01
-4.11289006e-01 3.50410342e-01 -5.68694055e-01 -7.96494037e-02
-1.94187909e-01 -9.38826978e-01 -3.91894169e-02 -7.74728477e-01
-5.85891843e-01 -5.43644965e-01 9.66651678e-01 -5.28829873e-01
1.28982854e+00 -2.24749804e+00 4.42636520e-01 5.48091471e-01
1.73084393e-01 -1.33488616e-02 -4.40098315e-01 7.59509921e-01
-1.78282246e-01 2.79414296e-01 4.61465977e-02 -5.79544127e-01
2.90492445e-01 4.86615598e-01 1.65798530e-01 -8.62644613e-02
2.28812635e-01 1.15362751e+00 -5.79641104e-01 -5.27247846e-01
1.74554676e-01 9.11273599e-01 -1.31609336e-01 5.14250457e-01
-1.04702763e-01 4.01192397e-01 -5.04147530e-01 6.68029904e-01
2.22142622e-01 -7.79373273e-02 6.53905630e-01 -4.18209970e-01
2.72305697e-01 1.95650803e-03 -8.39275539e-01 1.58953857e+00
-5.08396626e-01 7.47952342e-01 2.50704169e-01 -7.52455235e-01
1.04782605e+00 6.69303775e-01 9.78747383e-02 -6.52430892e-01
7.48699605e-01 -5.32378733e-01 1.18648842e-01 -8.22988927e-01
4.84111577e-01 -1.19651996e-01 -5.36498904e-01 4.96015817e-01
2.83376426e-01 2.58145064e-01 4.44322862e-02 5.58223128e-01
9.12037313e-01 -2.93064982e-01 4.50673074e-01 2.28642330e-01
6.19271100e-01 -3.69406641e-01 4.11674649e-01 4.49691325e-01
-6.25509977e-01 1.32942693e-02 9.67150450e-01 -5.51008284e-01
-1.90232649e-01 -4.22685713e-01 6.13751471e-01 1.60759556e+00
-2.55266819e-02 -4.26410615e-01 -6.92861438e-01 -7.64060497e-01
-4.66438144e-01 7.48424649e-01 -1.13032305e+00 -3.17810416e-01
-3.04679185e-01 -4.65429902e-01 7.31949508e-01 4.84392256e-01
4.92949456e-01 -1.72590339e+00 -3.84608328e-01 2.04125986e-01
-4.87229884e-01 -1.29355872e+00 -3.16042244e-01 3.31916243e-01
-3.58651638e-01 -6.40325129e-01 1.71901003e-01 -6.50060773e-01
3.80121708e-01 -3.65906715e-01 1.40881443e+00 -3.43239866e-02
-1.32721201e-01 6.51091754e-01 -5.65409482e-01 -3.15741688e-01
-3.52381110e-01 9.22907330e-03 -2.39686936e-01 5.24491131e-01
6.57180667e-01 -5.56988418e-01 -1.87420025e-01 6.85855970e-02
-8.86390626e-01 -1.63513720e-02 3.83127391e-01 5.64393938e-01
-2.24014260e-02 4.55603004e-02 7.41694868e-01 -8.54613543e-01
1.22086978e+00 -5.85982442e-01 1.52548328e-01 5.47357798e-01
6.40606955e-02 -1.50668487e-01 5.28078675e-01 -4.60056245e-01
-1.42358541e+00 6.51862193e-03 2.23141368e-02 1.04782684e-02
-6.79469705e-01 9.65048432e-01 -3.43942463e-01 4.39311415e-02
3.16554606e-01 -1.59771398e-01 -2.41802141e-01 1.27874583e-01
6.22170866e-01 9.07728493e-01 4.20409560e-01 -5.55109441e-01
-2.00888723e-01 3.38919647e-02 -1.17436841e-01 -1.06110764e+00
-4.89420593e-01 -3.19779783e-01 -5.43766618e-01 -9.44380462e-01
1.15373766e+00 -6.59293950e-01 -1.28117967e+00 4.76438254e-01
-1.23427153e+00 -3.54759216e-01 1.92180067e-01 2.98834234e-01
-2.28806674e-01 -2.26757191e-02 -9.03225958e-01 -1.27453673e+00
-4.74425405e-01 -9.28850710e-01 7.39861071e-01 5.85159183e-01
-7.55301058e-01 -1.45166790e+00 4.90783378e-02 5.03536940e-01
6.11138046e-01 4.50569481e-01 8.93683732e-01 -7.96065450e-01
1.71134458e-03 -1.63522050e-01 -1.50943324e-01 2.73261696e-01
2.59131491e-01 1.26983434e-01 -1.10623384e+00 2.77018160e-01
-1.52769208e-01 -8.16078126e-01 3.43312651e-01 1.13937646e-01
5.65743506e-01 -2.42548406e-01 -5.51289096e-02 1.14360698e-01
1.05848253e+00 3.85706514e-01 6.31005883e-01 -3.56386006e-01
8.97318482e-01 1.00897753e+00 2.07034964e-02 4.11273122e-01
7.63230026e-01 2.95807391e-01 4.20780480e-01 -6.05192743e-02
1.51842430e-01 1.54691711e-01 5.87104321e-01 1.25766563e+00
-9.68919024e-02 -7.63703287e-01 -9.36280549e-01 3.43066931e-01
-2.10634279e+00 -8.16315532e-01 -9.56036299e-02 1.35825813e+00
7.30947733e-01 -2.50117093e-01 -6.77587539e-02 -3.87963653e-01
5.97085655e-01 4.46398675e-01 -5.74953854e-01 -1.28654850e+00
-2.92725593e-01 -4.27867770e-02 -3.14512908e-01 8.60856652e-01
-8.36465299e-01 1.32592106e+00 6.31230259e+00 2.14402914e-01
-1.12119198e+00 -8.68813246e-02 8.94910097e-01 6.58227876e-02
-2.27409050e-01 -3.84600908e-01 -2.11575955e-01 -2.11869553e-01
1.30957162e+00 6.43114373e-02 8.21637392e-01 5.27333438e-01
-1.32544590e-02 -6.62739649e-02 -1.05764389e+00 1.09915435e+00
4.85169411e-01 -7.15878606e-01 -3.39880073e-03 -2.47575611e-01
2.76704609e-01 8.03368166e-02 -1.28399193e-01 5.82352042e-01
4.54813391e-01 -1.18346918e+00 1.42123908e-01 5.42738259e-01
3.05130154e-01 -1.11002088e+00 8.90542448e-01 1.80656061e-01
-1.24303854e+00 2.13577300e-01 2.67036527e-01 -2.54949987e-01
1.51307568e-01 1.51969180e-01 -6.87367618e-01 5.19828022e-01
6.34039044e-01 6.59194767e-01 -4.49165672e-01 -1.55930281e-01
-4.02855992e-01 6.33413732e-01 -2.35351697e-01 -5.94811618e-01
2.55385190e-01 -3.65867466e-01 2.27611661e-01 1.69166434e+00
-2.30199873e-01 4.97835159e-01 -1.10505551e-01 6.60325706e-01
-4.75567758e-01 3.55083227e-01 -6.69307292e-01 -5.33428431e-01
-1.45740611e-02 1.68665850e+00 -4.41997021e-01 -3.78735155e-01
-4.35887277e-01 1.14717603e+00 6.89752460e-01 7.01656699e-01
-6.91558063e-01 -3.96297753e-01 7.20263898e-01 -1.02849829e+00
-1.70582622e-01 -1.16751499e-01 -9.12209824e-02 -1.07447600e+00
-2.34316453e-01 -9.84279215e-01 7.04982221e-01 -9.99058902e-01
-1.43500423e+00 1.11365867e+00 -4.20484096e-01 -4.60862756e-01
-5.34335136e-01 -5.43298125e-01 -8.44089329e-01 6.39060318e-01
-1.06497955e+00 -1.36760724e+00 -3.92365307e-01 8.95777881e-01
2.76400328e-01 3.62948421e-03 1.17156756e+00 -9.50442404e-02
-7.88752198e-01 5.32615781e-01 -9.22526538e-01 5.42418420e-01
5.81039369e-01 -1.06643629e+00 -8.91683474e-02 4.74564373e-01
1.05231717e-01 6.01954877e-01 4.84009236e-01 -6.36419833e-01
-1.61972535e+00 -7.09700108e-01 1.16292739e+00 -3.00380945e-01
6.85834706e-01 -4.16731924e-01 -8.07470858e-01 9.58340704e-01
1.09188533e+00 -2.81121343e-01 1.16707575e+00 7.42987335e-01
-4.79094654e-01 4.48867679e-02 -1.01279974e+00 7.95621455e-01
7.56383836e-01 -8.31569493e-01 -4.88576055e-01 -7.18383044e-02
7.21279860e-01 -8.01304206e-02 -9.46006656e-01 3.58839065e-01
5.79359591e-01 -1.06550562e+00 5.20742834e-01 -8.31368506e-01
3.24438959e-01 7.24021494e-02 -3.10810059e-01 -1.49321461e+00
-1.25761613e-01 -7.62826502e-01 -4.61392477e-02 1.29296017e+00
6.39074028e-01 -5.32025337e-01 4.85337555e-01 1.31587410e+00
-2.98067853e-02 -5.13153791e-01 -8.93388927e-01 1.04958238e-02
-4.64073181e-01 -8.34932506e-01 5.19642234e-01 1.44773459e+00
9.41962600e-01 1.05184686e+00 -6.71172917e-01 2.80943662e-01
1.98127970e-01 -2.19644517e-01 6.59931242e-01 -1.01792145e+00
5.13063222e-02 -5.24596632e-01 -4.67408031e-01 -4.21917468e-01
5.16308784e-01 -5.93386889e-01 6.79683238e-02 -1.61241162e+00
8.24124366e-02 1.64775819e-01 -6.57267332e-01 7.59504139e-01
-6.67418092e-02 -1.26806885e-01 1.57081634e-01 -4.03304428e-01
-1.02205849e+00 7.78884351e-01 7.44277716e-01 -1.99985579e-01
-4.82919812e-01 -6.84819579e-01 -6.63087964e-01 7.22143769e-01
9.25769508e-01 -1.71007544e-01 -3.93021941e-01 -3.14183980e-01
5.96435368e-01 2.99094796e-01 -6.15443569e-03 -7.82793045e-01
6.64166868e-01 -2.47411922e-01 2.69953787e-01 -4.20100927e-01
7.66381979e-01 -8.86506557e-01 2.36490741e-02 -1.25109494e-01
-6.53358579e-01 1.14127547e-01 4.19507354e-01 3.60989094e-01
-6.22913361e-01 3.65191907e-01 3.31477821e-01 1.35901228e-01
-9.16443586e-01 2.01103047e-01 -6.71562135e-01 -1.94655716e-01
6.12250268e-01 2.45026588e-01 -4.86531377e-01 -1.19610655e+00
-1.07307243e+00 3.86894137e-01 -1.28372729e-01 8.32998216e-01
8.33557606e-01 -1.29719877e+00 -4.78421718e-01 6.65220257e-04
4.08246994e-01 -5.92600822e-01 5.83788097e-01 5.91632545e-01
-2.42298450e-02 1.31975994e-01 -1.91094309e-01 -3.22088808e-01
-1.64001822e+00 2.49862179e-01 5.46021760e-01 -2.59096175e-01
8.92281756e-02 9.84044731e-01 4.53780480e-02 -7.54454732e-01
4.17142689e-01 -1.35357976e-01 -7.62873352e-01 2.90607333e-01
4.25291806e-01 -3.93673219e-02 -2.81412363e-01 -1.06919563e+00
-4.86885577e-01 4.12178546e-01 2.98857018e-02 -3.82971406e-01
1.24137521e+00 -2.81740189e-01 -5.61484516e-01 9.18288708e-01
1.35375738e+00 -1.59575045e-01 -5.08628547e-01 -2.98881441e-01
-1.00445502e-01 1.92254782e-01 1.34831965e-01 -1.27592933e+00
-1.15946221e+00 1.01258385e+00 3.88440341e-01 4.10291314e-01
1.32351077e+00 2.19814211e-01 6.43893242e-01 6.01043344e-01
-5.64099103e-02 -1.57596695e+00 2.20843658e-01 8.52421522e-01
1.03356171e+00 -1.25471723e+00 -3.89051616e-01 -3.31181079e-01
-1.37240231e+00 1.21515620e+00 7.18314648e-01 4.95302379e-01
6.69526339e-01 1.07429661e-01 7.55387068e-01 -6.59836829e-01
-1.27663469e+00 -1.78441405e-01 4.06018466e-01 3.67824316e-01
8.18686306e-01 4.42439616e-01 2.82552004e-01 7.86634982e-01
3.98584455e-02 -2.26807117e-01 2.37814769e-01 9.34064507e-01
-9.18464214e-02 -1.08993697e+00 -5.81597202e-02 -3.25968973e-02
-1.75173268e-01 -1.39373466e-01 -1.33566737e+00 6.53070331e-01
3.92178483e-02 1.75775528e+00 -6.74550980e-02 -7.70023406e-01
4.30467337e-01 6.33913636e-01 4.46425788e-02 -1.50590032e-01
-1.01885295e+00 2.12331042e-02 6.58229828e-01 -7.53787637e-01
-8.35094988e-01 -3.48962307e-01 -1.55337214e+00 -2.22622827e-01
1.41585562e-02 1.14247903e-01 9.26012754e-01 1.06588852e+00
6.07550085e-01 7.77078331e-01 6.20224535e-01 -4.51259017e-01
5.52814782e-01 -1.17298460e+00 -5.27558982e-01 7.09418952e-01
1.18319064e-01 -1.32148772e-01 -4.34960634e-01 -1.14711775e-02] | [12.945305824279785, 6.1586174964904785] |
a1d0e170-36c0-4e57-9535-8c303d48ac9f | physics-based-deep-learning | 2109.05237 | null | https://arxiv.org/abs/2109.05237v3 | https://arxiv.org/pdf/2109.05237v3.pdf | Physics-based Deep Learning | This digital book contains a practical and comprehensive introduction of everything related to deep learning in the context of physical simulations. As much as possible, all topics come with hands-on code examples in the form of Jupyter notebooks to quickly get started. Beyond standard supervised learning from data, we'll look at physical loss constraints, more tightly coupled learning algorithms with differentiable simulations, as well as reinforcement learning and uncertainty modeling. We live in exciting times: these methods have a huge potential to fundamentally change what computer simulations can achieve. | ['Kiwon Um', 'Felix Trost', 'Patrick Schnell', 'Maximilian Mueller', 'Philipp Holl', 'Nils Thuerey'] | 2021-09-11 | null | null | null | null | ['physical-simulations'] | ['miscellaneous'] | [-7.84049273e-01 3.52576897e-02 -2.40193829e-01 -3.90919447e-01
-7.14231968e-01 -2.20663443e-01 3.79446864e-01 2.79269740e-02
-3.06733996e-01 1.25162601e+00 -2.17287436e-01 -4.01480973e-01
-3.40677917e-01 -6.77075922e-01 -7.15738297e-01 -7.29377806e-01
-5.69723964e-01 4.95076060e-01 -1.36443749e-01 -3.42990279e-01
4.18299943e-01 3.80181044e-01 -1.09141016e+00 -3.48120719e-01
6.00778580e-01 8.20047021e-01 -2.10542887e-01 7.00756550e-01
3.13646086e-02 9.05376136e-01 -3.07783455e-01 -4.26733583e-01
9.70283747e-02 -3.88403922e-01 -4.71445233e-01 -4.92175877e-01
-2.21576206e-02 -4.13182408e-01 -8.50731254e-01 8.67999971e-01
5.73914289e-01 4.31629509e-01 5.70854902e-01 -1.15186882e+00
-4.89849925e-01 6.58730626e-01 -4.41130608e-01 2.08941668e-01
1.82153791e-01 6.27904594e-01 7.59739935e-01 -4.16956812e-01
2.38470957e-01 1.15473890e+00 1.00088799e+00 3.51112723e-01
-1.00997925e+00 -5.54021537e-01 8.75483900e-02 1.03714362e-01
-8.48367453e-01 -1.46765649e-01 7.22977996e-01 -4.07415658e-01
1.07929480e+00 -4.89644371e-02 5.04033029e-01 1.23194981e+00
6.39176130e-01 8.34780633e-01 8.61287773e-01 -4.41332847e-01
5.51203728e-01 1.16256580e-01 8.24876428e-02 6.25624061e-01
1.22131310e-01 8.57690334e-01 -4.83905226e-01 -1.58459723e-01
8.87038350e-01 -3.75682712e-02 -6.39086962e-02 -6.22836113e-01
-1.09898508e+00 8.18352461e-01 2.19328284e-01 -7.49435350e-02
1.14803441e-01 8.14473569e-01 6.46240473e-01 3.37441623e-01
1.22733526e-01 5.55918157e-01 -6.57501459e-01 -4.78995025e-01
-9.38295662e-01 7.20737159e-01 1.00296295e+00 6.88333511e-01
2.29733974e-01 4.76853013e-01 2.66058981e-01 4.48720187e-01
4.95638758e-01 5.82691550e-01 2.05510214e-01 -1.41204476e+00
1.84672564e-01 -4.33626473e-01 4.89756584e-01 -3.79984707e-01
-6.08344674e-01 -2.93949723e-01 -7.01273859e-01 1.00485814e+00
3.82528186e-01 -8.64169240e-01 -5.02282858e-01 1.45301831e+00
9.92484763e-02 1.89953849e-01 -8.46899971e-02 8.63894284e-01
6.96305096e-01 6.97265744e-01 7.54662007e-02 -5.29259518e-02
6.73969865e-01 -8.17815959e-01 -6.19393468e-01 2.27670930e-02
4.64411765e-01 -6.13672137e-01 8.60843718e-01 7.29763567e-01
-1.40251720e+00 -3.69531184e-01 -1.32121003e+00 6.25243923e-03
-4.08241779e-01 -3.86123866e-01 1.27720404e+00 6.71164095e-01
-6.86075807e-01 1.65874374e+00 -1.33240080e+00 -1.89777538e-02
2.46909946e-01 2.80203789e-01 2.36805588e-01 2.31750131e-01
-1.47405589e+00 1.23578227e+00 2.04931200e-02 -2.46559501e-01
-9.21691597e-01 -9.49778199e-01 -7.90540040e-01 -1.26740197e-02
2.01925933e-01 -7.33315706e-01 1.46648598e+00 -4.77816850e-01
-1.81530225e+00 3.92050922e-01 2.93794662e-01 -6.77505195e-01
6.62965775e-01 -4.42809820e-01 -2.74711281e-01 -9.84455422e-02
-4.03076500e-01 2.69824564e-01 4.93963033e-01 -9.83547628e-01
-1.62245676e-01 2.85705049e-02 -3.81086841e-02 4.39219177e-02
2.56588548e-01 9.06103849e-02 4.58527505e-02 -5.94998598e-01
-4.38150346e-01 -5.48453033e-01 -5.13257980e-01 2.90463716e-01
-8.74874815e-02 -2.68536568e-01 6.50761425e-01 -3.66323143e-01
7.32584357e-01 -1.91311038e+00 5.48494644e-02 5.73684787e-03
1.01212211e-01 1.86271489e-01 3.36259246e-01 6.37674749e-01
-1.10134641e-02 -2.96043754e-02 -1.83758989e-01 -2.80276030e-01
6.35420620e-01 3.99218053e-02 -3.29042047e-01 5.07856548e-01
-1.76836133e-01 1.04999208e+00 -1.12976885e+00 -1.96674228e-01
6.78651214e-01 4.99376565e-01 -3.89776617e-01 1.29559198e-02
-5.28201461e-01 4.75880533e-01 -4.35378104e-01 2.75075406e-01
5.20630956e-01 -3.65921587e-01 5.86567447e-02 4.56255436e-01
-5.28165773e-02 4.32638586e-01 -1.12604403e+00 1.63184988e+00
-5.37460268e-01 6.42994761e-01 2.41035625e-01 -1.23638070e+00
4.75393206e-01 1.49560869e-01 6.77178562e-01 -6.58399343e-01
1.24069728e-01 6.51757978e-03 -9.75745916e-02 -3.72623980e-01
2.47110739e-01 -4.16872293e-01 -1.65404707e-01 5.11318147e-01
2.26178244e-01 -6.16393209e-01 -3.71323824e-02 3.41411918e-01
8.82473826e-01 6.76161408e-01 5.65027967e-02 -4.00969476e-01
2.86343996e-03 -1.52830884e-01 3.53483796e-01 9.39593852e-01
-3.55939448e-01 2.49815434e-01 3.88810307e-01 -6.42338991e-01
-1.16132784e+00 -1.42694354e+00 -3.58081698e-01 9.26655054e-01
4.02641222e-02 -3.72687697e-01 -5.76041937e-01 -1.71209574e-01
4.57928360e-01 6.77668571e-01 -3.95822763e-01 -1.21770397e-01
-2.93000847e-01 -6.32264614e-01 3.05044472e-01 7.03997970e-01
3.71383816e-01 -1.21199548e+00 -3.09749991e-01 1.42033681e-01
6.31287456e-01 -6.48192286e-01 -9.78366584e-02 6.41735494e-01
-1.04144514e+00 -6.51039541e-01 -3.90854180e-01 -4.98268902e-01
-7.44422823e-02 -4.14440930e-01 1.24655271e+00 5.71110286e-02
-6.64946318e-01 4.70676184e-01 4.01163623e-02 -3.82615805e-01
-2.35805571e-01 -3.04441392e-01 3.78884226e-01 -1.01069713e+00
2.46922180e-01 -9.53168273e-01 -4.04591084e-01 -2.33175546e-01
-3.80104661e-01 -3.11106354e-01 2.79518157e-01 8.09529245e-01
2.89640605e-01 1.01377755e-01 5.89352548e-01 -7.21228600e-01
5.36666214e-01 -4.68016624e-01 -1.03646588e+00 1.17689148e-01
-7.76444972e-01 3.00191373e-01 7.07971156e-01 -1.47361144e-01
-8.67104948e-01 -1.66416377e-01 -4.45518732e-01 -2.37786695e-01
-1.46227255e-01 4.13518697e-01 1.59735456e-01 -3.20212454e-01
7.42576122e-01 -1.04808733e-01 -1.50864143e-02 -4.91912454e-01
4.16024238e-01 4.09440130e-01 5.19181430e-01 -9.84667003e-01
6.80630684e-01 1.77780047e-01 1.75722852e-01 -6.84882760e-01
-6.36250019e-01 1.93054914e-01 -4.87246037e-01 -6.84503987e-02
5.81136167e-01 -7.33305097e-01 -1.14608550e+00 4.62641209e-01
-6.14955783e-01 -1.02906144e+00 -5.02348125e-01 8.68984640e-01
-8.74169350e-01 4.63337511e-01 -9.48781550e-01 -7.80178726e-01
-8.21771920e-02 -9.91854131e-01 6.78975105e-01 6.64615154e-01
-1.78317562e-01 -1.62225485e+00 3.53095144e-01 -7.06309825e-02
4.74063039e-01 3.00610155e-01 7.37541318e-01 -2.51154721e-01
-6.37404084e-01 -2.75662839e-01 2.20852524e-01 3.05101246e-01
-2.61640400e-01 3.62901628e-01 -8.65505457e-01 -3.15070957e-01
-9.09612775e-02 -8.61063063e-01 8.63012135e-01 6.86069071e-01
1.57248950e+00 -8.89699757e-02 -4.38148350e-01 8.04967880e-01
1.34091771e+00 1.53097168e-01 4.01562095e-01 2.15671211e-01
3.37821275e-01 8.63969252e-02 4.16988432e-01 8.36510181e-01
2.69399043e-02 1.39724523e-01 3.14361215e-01 1.39164582e-01
1.38515890e-01 -2.98949361e-01 2.39727661e-01 6.95793092e-01
9.53771733e-03 6.21289276e-02 -8.34040046e-01 6.63722083e-02
-1.75556934e+00 -1.21010113e+00 1.36165889e-02 2.39495444e+00
8.59023988e-01 5.49455881e-01 8.11952632e-03 -1.93930402e-01
4.46474016e-01 2.65563093e-02 -1.00875199e+00 -8.41608942e-01
3.70406032e-01 4.16108698e-01 5.64019680e-01 8.15320611e-01
-1.09970951e+00 5.65885723e-01 8.85402584e+00 8.97565782e-01
-1.15310252e+00 1.02114854e-02 5.94949126e-01 -5.71397483e-01
-3.32853943e-01 -1.58318430e-02 -6.34336472e-01 7.76414275e-01
1.29638791e+00 -3.28598022e-01 7.45729029e-01 1.12385333e+00
3.93875778e-01 -2.96342820e-01 -1.15911651e+00 1.00360322e+00
-4.16164249e-01 -1.52067864e+00 -6.38476372e-01 -3.09482105e-02
7.80624688e-01 4.32464451e-01 1.49679855e-01 7.11881280e-01
9.95281100e-01 -1.40404320e+00 3.67029995e-01 7.70994782e-01
6.30428910e-01 -9.24235404e-01 4.80486572e-01 3.55357707e-01
-4.88969594e-01 9.19906050e-02 -7.37368703e-01 -6.05728984e-01
3.76899749e-01 8.79944444e-01 -2.46053766e-02 2.01131731e-01
7.22701192e-01 8.69280875e-01 7.69228339e-02 1.50348103e+00
-1.39812469e-01 6.12613916e-01 -6.83651388e-01 -5.02199769e-01
1.79243505e-01 -7.36726820e-02 3.38670284e-01 1.21444714e+00
-1.14891477e-01 1.50774896e-01 1.76167622e-01 1.01430583e+00
-6.33227006e-02 -5.81393719e-01 -5.47469616e-01 -8.11258033e-02
2.45915338e-01 1.19051802e+00 -5.26638389e-01 -2.04657152e-01
-6.31621838e-01 5.30408680e-01 1.86419025e-01 2.84571558e-01
-1.18806589e+00 -6.42064631e-01 9.41504478e-01 -9.87838954e-02
2.23324746e-01 -5.78298390e-01 -4.35996950e-01 -1.07264030e+00
-3.86348069e-01 -5.64681292e-01 5.42122722e-02 -9.49123204e-01
-1.61682296e+00 -1.25225320e-01 -4.95254179e-04 -8.33515465e-01
-3.11348736e-01 -1.05837631e+00 -1.06352079e+00 6.91788137e-01
-1.14209199e+00 -3.67261142e-01 5.17405383e-02 2.92079508e-01
2.31013551e-01 -7.40227289e-03 7.39963293e-01 2.13977784e-01
-3.33901435e-01 6.64255142e-01 1.01460993e+00 7.62582421e-02
6.66879654e-01 -1.48498106e+00 5.75205922e-01 3.30480844e-01
5.61246388e-02 5.61768651e-01 1.05410731e+00 -4.54290509e-01
-1.41694880e+00 -4.84247804e-01 2.27123775e-04 -5.14631331e-01
1.10837281e+00 -2.82047570e-01 -8.27729404e-01 7.94553220e-01
3.31047893e-01 2.03764752e-01 3.29641044e-01 3.45972657e-01
-7.49781430e-02 4.02585045e-03 -1.03909147e+00 2.80070215e-01
7.26449788e-01 -4.97336060e-01 -4.08103973e-01 7.95731425e-01
6.57605350e-01 -7.99518824e-01 -9.59197700e-01 3.07129472e-01
5.92959344e-01 -8.35584879e-01 1.06271315e+00 -8.37274551e-01
6.86746359e-01 2.18435407e-01 2.97459155e-01 -1.41716516e+00
-1.62587419e-01 -1.11245859e+00 -4.82545972e-01 7.32379675e-01
3.57852370e-01 -5.33017039e-01 9.33503747e-01 7.90382147e-01
-8.61187503e-02 -1.00358546e+00 -7.11423934e-01 -9.54854965e-01
9.70431864e-01 -4.26139534e-01 3.08321387e-01 7.24870920e-01
6.92419410e-01 -6.43471405e-02 -3.33400488e-01 -1.38057515e-01
1.07836390e+00 9.29306298e-02 3.55069250e-01 -8.89718533e-01
-7.15559065e-01 -6.22982144e-01 -2.79465795e-01 -9.95293319e-01
4.33899313e-02 -6.68430746e-01 1.13831326e-01 -1.31144452e+00
2.33425349e-01 -4.59436297e-01 -4.27946538e-01 -2.43127998e-02
6.23461567e-02 -3.82034248e-03 -1.01210907e-01 -1.91036299e-01
-9.20358181e-01 6.87655926e-01 1.15433908e+00 1.61448970e-01
1.22204069e-02 2.42409408e-01 -6.53071940e-01 8.95930231e-01
1.02722502e+00 -2.61756003e-01 -3.78745109e-01 -2.26040170e-01
3.57396573e-01 4.50724602e-01 4.00286019e-01 -1.19978762e+00
1.01260304e-01 -5.41125596e-01 8.09632659e-01 -3.97555202e-01
3.38364720e-01 -3.78215671e-01 -2.74766684e-01 3.42382193e-01
-5.16774833e-01 -2.24784955e-01 5.79604745e-01 4.00145203e-01
1.28620580e-01 -4.00074303e-01 1.14393353e+00 -3.88557494e-01
-4.94232327e-01 2.24227712e-01 -4.15174097e-01 3.31094980e-01
1.05675006e+00 3.60394239e-01 -3.77613574e-01 -8.13772738e-01
-1.14847589e+00 6.51331127e-01 5.33770680e-01 1.33656487e-01
1.70402527e-01 -1.09565580e+00 -2.88071036e-01 -2.43107781e-01
-5.39767802e-01 -1.11438587e-01 2.14677110e-01 4.80194509e-01
-5.11563241e-01 5.47527909e-01 -2.87462264e-01 -2.38579974e-01
-5.33483744e-01 6.64236486e-01 8.19000542e-01 -1.35828584e-01
-8.44864607e-01 9.24884915e-01 -2.84641683e-01 -6.84288383e-01
5.35237491e-01 -3.01534712e-01 4.09030080e-01 -7.17602611e-01
6.23138547e-01 5.20593464e-01 -1.51908308e-01 2.82378316e-01
-1.94091946e-01 3.77592891e-01 1.20673552e-01 -2.41006017e-01
1.54778659e+00 -1.27593100e-01 1.24090016e-01 8.01054895e-01
1.03732026e+00 -6.27248362e-02 -1.92570305e+00 1.27429321e-01
-1.81434482e-01 -1.22002810e-01 6.42303228e-02 -1.22684968e+00
-8.71618509e-01 1.16139424e+00 6.28882766e-01 -4.49902713e-02
3.74612749e-01 -5.65615930e-02 9.26741958e-01 7.78580308e-01
5.72835445e-01 -1.45624125e+00 2.15762734e-01 6.23260379e-01
5.54728627e-01 -1.45614934e+00 5.41320205e-01 3.75724375e-01
-4.59699035e-01 1.20816016e+00 5.82566381e-01 -7.34169662e-01
1.19811821e+00 9.03634846e-01 -2.13188723e-01 2.97461990e-02
-8.73445153e-01 1.13370076e-01 -2.88222849e-01 8.12698245e-01
6.28784060e-01 -8.00947025e-02 2.88681090e-02 7.04519093e-01
-2.32281730e-01 1.49528027e-01 6.09242022e-01 6.90095544e-01
-7.43418634e-01 -1.25111723e+00 -2.57491648e-01 4.50849950e-01
-5.66590488e-01 6.06950969e-02 2.43163422e-01 8.60287130e-01
-2.47752860e-01 4.81183797e-01 -6.82700947e-02 -6.79423809e-02
-2.16688693e-01 2.28020892e-01 8.49204004e-01 -4.15131032e-01
-3.12118232e-01 -3.27652693e-01 -1.61833361e-01 -7.26813138e-01
4.39780951e-02 -5.36615252e-01 -1.68152225e+00 -1.13806152e+00
-8.75911713e-02 1.70020133e-01 8.39669466e-01 1.07739282e+00
-1.57082267e-02 5.61628878e-01 3.21217090e-01 -1.28136718e+00
-8.72277379e-01 -7.20224380e-01 -8.63603532e-01 -9.03232396e-02
3.99245262e-01 -7.56971717e-01 -5.05767643e-01 -4.21779245e-01] | [6.426987648010254, 3.5043442249298096] |
86b40853-42ef-47d3-8f65-5e7a78d5e5dc | appearance-consensus-driven-self-supervised | 2008.01341 | null | https://arxiv.org/abs/2008.01341v1 | https://arxiv.org/pdf/2008.01341v1.pdf | Appearance Consensus Driven Self-Supervised Human Mesh Recovery | We present a self-supervised human mesh recovery framework to infer human pose and shape from monocular images in the absence of any paired supervision. Recent advances have shifted the interest towards directly regressing parameters of a parametric human model by supervising them on large-scale datasets with 2D landmark annotations. This limits the generalizability of such approaches to operate on images from unlabeled wild environments. Acknowledging this we propose a novel appearance consensus driven self-supervised objective. To effectively disentangle the foreground (FG) human we rely on image pairs depicting the same person (consistent FG) in varied pose and background (BG) which are obtained from unlabeled wild videos. The proposed FG appearance consistency objective makes use of a novel, differentiable Color-recovery module to obtain vertex colors without the need for any appearance network; via efficient realization of color-picking and reflectional symmetry. We achieve state-of-the-art results on the standard model-based 3D pose estimation benchmarks at comparable supervision levels. Furthermore, the resulting colored mesh prediction opens up the usage of our framework for a variety of appearance-related tasks beyond the pose and shape estimation, thus establishing our superior generalizability. | ['R. Venkatesh Babu', 'Rahul Mysore Venkatesh', 'Mugalodi Rakesh', 'Jogendra Nath Kundu', 'Varun Jampani'] | 2020-08-04 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2788_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123460766.pdf | eccv-2020-8 | ['human-mesh-recovery'] | ['computer-vision'] | [ 3.00280869e-01 1.78698048e-01 1.47067443e-01 -4.09939021e-01
-5.32339931e-01 -4.27006871e-01 5.58249116e-01 -2.85721570e-01
-2.37110764e-01 4.43824857e-01 -2.14334846e-01 2.31762320e-01
2.57039756e-01 -3.07169914e-01 -1.05796874e+00 -6.31407261e-01
6.08379804e-02 1.05974662e+00 1.29295409e-01 -4.71717156e-02
-1.01711936e-01 6.38873100e-01 -1.76298070e+00 -6.33480325e-02
4.99241382e-01 9.32836115e-01 -1.25161201e-01 5.96070945e-01
4.29390371e-01 5.31228662e-01 -3.30610424e-01 -5.64724386e-01
7.61152327e-01 -2.29353532e-01 -7.40753949e-01 6.95446432e-01
1.42509270e+00 -4.72025245e-01 -2.53123760e-01 8.98449063e-01
4.65899318e-01 -3.43321264e-02 7.14313567e-01 -1.48529398e+00
-5.60652673e-01 -1.44942701e-01 -8.32000732e-01 -2.85001755e-01
6.13114953e-01 2.33830675e-01 9.20435250e-01 -8.83190155e-01
1.17755508e+00 1.37134922e+00 7.51667738e-01 5.85767090e-01
-1.63250625e+00 -4.86269563e-01 4.46115643e-01 4.40037027e-02
-1.34195554e+00 -5.03489137e-01 9.76886809e-01 -6.33571982e-01
4.50934052e-01 2.85081744e-01 1.00706112e+00 1.40388691e+00
-3.48167598e-01 7.78543234e-01 1.41427004e+00 -5.21016717e-01
7.44979158e-02 3.14370357e-02 -3.01627755e-01 1.12801051e+00
2.76208282e-01 7.12509640e-03 -6.84248149e-01 -1.46276444e-01
1.21326149e+00 -7.26241246e-02 -2.36674741e-01 -1.29160047e+00
-1.39454770e+00 3.26943994e-01 3.92568350e-01 -2.76827127e-01
-1.43812448e-01 2.42324606e-01 -2.42320285e-03 2.43605793e-01
7.40629971e-01 1.84433833e-01 -4.50234324e-01 4.21442240e-01
-7.31281459e-01 3.31825674e-01 7.90462315e-01 1.15397894e+00
9.33288932e-01 -1.16727993e-01 4.30848375e-02 5.75283349e-01
2.48837769e-01 7.87004471e-01 -3.83620411e-01 -1.24257207e+00
2.86568612e-01 6.58514082e-01 2.91801035e-01 -1.06656647e+00
-3.75268996e-01 -2.50148267e-01 -7.42187262e-01 4.64160174e-01
8.18175733e-01 1.26108423e-01 -1.03249943e+00 1.77275658e+00
9.38956976e-01 1.74415305e-01 -5.20390749e-01 1.12226987e+00
6.62930489e-01 -5.61848022e-02 -1.70765013e-01 1.93549946e-01
1.28770947e+00 -1.11268866e+00 -2.54086882e-01 -3.20867509e-01
2.42154319e-02 -6.55403197e-01 1.04868817e+00 6.60324275e-01
-1.09335697e+00 -5.94634473e-01 -9.84039307e-01 -3.35944146e-01
-2.02398211e-01 3.14496785e-01 7.22975135e-01 6.04631603e-01
-1.02813959e+00 5.80097437e-01 -8.50854635e-01 -4.72170889e-01
3.57662529e-01 4.41918135e-01 -9.17735279e-01 -1.39911085e-01
-5.77265739e-01 8.58194351e-01 -9.32520479e-02 4.49519038e-01
-7.86385596e-01 -6.95188701e-01 -1.02289331e+00 -6.74238205e-01
6.29091620e-01 -1.22625756e+00 8.16105843e-01 -1.32548451e+00
-1.51587224e+00 1.62068951e+00 -5.79833761e-02 -5.10752797e-02
1.18698335e+00 -3.95089477e-01 -3.36179547e-02 3.55502158e-01
2.15210263e-02 7.67261386e-01 1.23858619e+00 -1.72112870e+00
-2.66876966e-01 -6.75175011e-01 5.73558658e-02 2.96841830e-01
-3.73785943e-02 -1.60469174e-01 -9.48569894e-01 -5.79805136e-01
1.85274303e-01 -1.37225938e+00 -9.43651125e-02 8.25671077e-01
-5.92830241e-01 6.59219101e-02 3.88982207e-01 -9.09675241e-01
4.57167715e-01 -1.74898553e+00 7.32735157e-01 3.98592383e-01
3.79130483e-01 -2.08187640e-01 -1.28836155e-01 1.13670461e-01
7.79253021e-02 -4.35463369e-01 -1.99278325e-01 -9.69023883e-01
1.53695181e-01 2.01120749e-01 1.93728209e-01 1.06249416e+00
3.69938403e-01 9.01484907e-01 -9.09091473e-01 -8.13544035e-01
4.55477804e-01 5.97838402e-01 -6.61777258e-01 5.54344535e-01
-3.22292209e-01 1.00131273e+00 -1.56141028e-01 8.36246550e-01
7.50304461e-01 -5.38107574e-01 3.39820117e-01 -4.68792856e-01
2.83117175e-01 -1.99425340e-01 -1.48751462e+00 2.22621369e+00
-2.55249947e-01 3.03119540e-01 3.85247439e-01 -7.44547009e-01
5.32627165e-01 -2.64584552e-02 6.18453383e-01 -3.52946728e-01
1.36420920e-01 5.26509881e-02 -3.65790129e-01 -2.81614751e-01
5.56214340e-02 1.02264725e-01 1.91101536e-01 1.77032024e-01
1.87651962e-01 -5.37561439e-02 -1.06361821e-01 1.57436416e-01
8.00162017e-01 8.83701444e-01 -9.00122747e-02 -2.31829971e-01
4.16918963e-01 -2.22730771e-01 3.61663043e-01 3.65995407e-01
-7.34505281e-02 9.46930528e-01 1.42302498e-01 -5.26975214e-01
-1.26779497e+00 -1.41611838e+00 -1.28932014e-01 1.18866825e+00
3.38129669e-01 -2.07160518e-01 -6.87029123e-01 -7.21776009e-01
3.79039586e-01 -8.10002014e-02 -9.89277661e-01 3.67280215e-01
-6.98686600e-01 -3.40555847e-01 2.20725730e-01 3.88438046e-01
1.95646420e-01 -6.96865141e-01 -5.11291444e-01 -2.59492815e-01
-3.31500679e-01 -1.39319491e+00 -4.99743998e-01 -7.89991543e-02
-4.50410455e-01 -1.40792823e+00 -9.46915388e-01 -7.34843552e-01
1.06125188e+00 -2.13929247e-02 1.27831209e+00 1.57718509e-01
-6.60130858e-01 9.49582100e-01 2.15261634e-02 -2.12293537e-03
-1.67665318e-01 -2.22867906e-01 1.69831708e-01 3.72318864e-01
-1.59232810e-01 -6.86543047e-01 -9.83803213e-01 3.93160135e-01
-3.44012558e-01 4.37264830e-01 4.43975776e-01 5.80794990e-01
7.54581332e-01 -6.85656965e-01 -8.32420890e-04 -1.00019455e+00
-3.08434904e-01 -5.32877930e-02 -6.05977595e-01 5.11977673e-01
-4.25792068e-01 -4.52835951e-03 3.52430135e-01 -3.81175369e-01
-9.89846766e-01 4.69639927e-01 2.83192366e-01 -7.47624278e-01
-2.39887297e-01 -3.93597245e-01 -2.61931211e-01 -4.67795938e-01
4.31557268e-01 -2.94026602e-02 3.02670836e-01 -5.61573207e-01
6.22578681e-01 1.68519795e-01 8.99068058e-01 -8.50561976e-01
1.08709502e+00 8.31798017e-01 3.79010946e-01 -7.35513031e-01
-8.03610623e-01 -5.84754705e-01 -1.11746871e+00 -4.25522298e-01
9.86046493e-01 -1.20468485e+00 -8.54049206e-01 5.04110038e-01
-1.01685321e+00 -4.16606218e-01 1.25667332e-02 1.27009759e-02
-7.43494332e-01 7.70087600e-01 -6.61229134e-01 -8.14991355e-01
-2.06346899e-01 -9.22651052e-01 1.75366855e+00 -3.60265881e-01
-1.67216986e-01 -8.12222123e-01 7.28887506e-03 6.74627066e-01
-1.03939168e-01 7.51531661e-01 4.98523235e-01 -1.25905216e-01
-6.75500154e-01 -1.80837270e-02 -1.45829380e-01 2.91364253e-01
1.75385885e-02 1.29952710e-02 -1.13042986e+00 -3.90619874e-01
-4.17726457e-01 -6.15847886e-01 5.52378654e-01 1.23360626e-01
1.02744555e+00 -7.93052837e-02 -2.35772297e-01 8.90071809e-01
1.21885383e+00 -7.93420792e-01 1.92303285e-01 1.55665547e-01
1.32582629e+00 9.20462430e-01 3.42402369e-01 3.38010728e-01
5.21565139e-01 1.08166373e+00 4.47961628e-01 -4.84141022e-01
-4.09415424e-01 -4.26219761e-01 2.12521836e-01 4.14587289e-01
-5.46992838e-01 2.28775278e-01 -7.52927482e-01 1.95747495e-01
-1.86299467e+00 -5.43194890e-01 -1.54087111e-01 2.47009993e+00
8.41392398e-01 -1.82818517e-01 4.08541232e-01 -2.00120956e-01
6.00710452e-01 -1.03602141e-01 -5.63580751e-01 5.03828943e-01
-1.16392270e-01 2.41106540e-01 5.24875522e-01 4.42841113e-01
-1.36484444e+00 7.81507730e-01 5.75983477e+00 4.89708841e-01
-7.58495331e-01 3.43365013e-03 6.59004509e-01 -1.52850613e-01
-1.95814222e-01 -1.88866138e-01 -5.15739918e-01 1.30803823e-01
3.05068076e-01 4.56806630e-01 6.28193259e-01 7.76150107e-01
-1.17311433e-01 -8.13324153e-02 -1.58935285e+00 1.38649857e+00
4.38459426e-01 -9.81422842e-01 7.76564851e-02 1.03836551e-01
8.41068268e-01 -3.08840394e-01 5.95399588e-02 -3.03295225e-01
2.94543952e-01 -9.47605193e-01 1.10968614e+00 6.60003126e-01
9.90304112e-01 -3.89508814e-01 2.57299066e-01 -1.08195958e-03
-1.19133830e+00 3.26608062e-01 -1.71092331e-01 1.87020004e-02
2.07050785e-01 4.75360632e-01 -6.27833188e-01 7.87941992e-01
6.78912222e-01 7.82509148e-01 -9.18472350e-01 6.46509886e-01
-2.25745663e-01 -2.96393167e-02 -4.75279182e-01 4.83135879e-01
-3.83451819e-01 -4.55905825e-01 3.79438877e-01 8.98415089e-01
-8.25971588e-02 -2.41390079e-01 5.24887204e-01 9.69001234e-01
5.94184920e-03 1.21229604e-01 -2.98651576e-01 3.38452518e-01
1.48231462e-01 1.34207857e+00 -8.22283387e-01 -1.91548958e-01
-3.85419101e-01 1.56251240e+00 5.82553923e-01 2.75032997e-01
-8.57739031e-01 4.75018322e-01 5.03198504e-01 4.59055513e-01
3.11937600e-01 -2.64760792e-01 -6.64081201e-02 -1.42803848e+00
3.56235772e-01 -9.59444344e-01 2.33168557e-01 -7.64107108e-01
-1.63965285e+00 3.90814215e-01 1.18656352e-01 -1.21550667e+00
-1.41189098e-01 -7.86567569e-01 -2.93552764e-02 6.59120858e-01
-1.31890917e+00 -1.89938319e+00 -6.19567692e-01 8.32660079e-01
1.77375510e-01 2.72318512e-01 7.58606315e-01 3.20490748e-01
-4.84468281e-01 5.83459139e-01 -4.87007856e-01 1.62412331e-01
1.13525367e+00 -1.52341497e+00 4.13055152e-01 7.86800325e-01
2.90473461e-01 5.91963649e-01 6.83584452e-01 -6.86935246e-01
-1.84592640e+00 -9.88618016e-01 3.62297922e-01 -1.08913410e+00
3.81674349e-01 -8.76509666e-01 -4.80379850e-01 7.87542045e-01
-1.02325968e-01 3.81521672e-01 2.94978797e-01 1.88339502e-01
-6.37959301e-01 -6.65286602e-03 -9.73806918e-01 6.42505109e-01
1.66854608e+00 -4.92234409e-01 -2.80536175e-01 4.65084344e-01
3.06744993e-01 -6.60039067e-01 -8.80236208e-01 4.09975857e-01
7.72243798e-01 -9.65084791e-01 1.35064542e+00 -5.22806883e-01
3.26922297e-01 -4.34133828e-01 -9.36327428e-02 -8.35099518e-01
-6.08236119e-02 -6.58516407e-01 -2.26717770e-01 9.78321731e-01
-7.99059570e-02 -3.89243394e-01 9.35561538e-01 8.55242908e-01
1.60247818e-01 -6.28634274e-01 -8.60549688e-01 -6.64904118e-01
-2.16694146e-01 -9.41668898e-02 1.55540943e-01 7.70499825e-01
-4.62962151e-01 1.20572582e-01 -7.54364610e-01 3.22232693e-01
1.22102296e+00 2.52103031e-01 1.46657515e+00 -1.46029496e+00
-6.14002347e-01 -4.53363210e-02 -5.50651073e-01 -1.09603822e+00
3.60607773e-01 -8.05180073e-01 8.34686607e-02 -1.12480974e+00
4.01291400e-01 -3.51064742e-01 4.74245511e-02 4.76875216e-01
-1.88019559e-01 7.76662707e-01 1.81005821e-01 1.68024153e-01
-8.29423428e-01 5.47977865e-01 1.27699840e+00 -1.15138084e-01
3.03346604e-01 -1.49399027e-01 -1.83458716e-01 9.18827593e-01
1.25225618e-01 -1.10403858e-01 -3.49141985e-01 -3.80307198e-01
2.74232030e-01 -2.59586513e-01 1.03028798e+00 -8.29253495e-01
4.57830355e-02 -2.42250171e-02 8.52951884e-01 -2.55606145e-01
6.20919287e-01 -9.71095324e-01 5.48453569e-01 2.14382902e-01
-8.63278359e-02 1.29585102e-01 -1.20636664e-01 8.92574310e-01
3.78563643e-01 3.63190532e-01 6.30678594e-01 -2.92584568e-01
-7.58239388e-01 6.27746701e-01 4.70147818e-01 2.26820230e-01
8.91327620e-01 -3.89043272e-01 1.36110023e-01 -2.02246666e-01
-9.00504649e-01 -1.37017900e-02 1.28662217e+00 3.97946566e-01
4.97877985e-01 -1.27188885e+00 -6.93686187e-01 3.66266102e-01
3.33425969e-01 1.29187435e-01 2.89738744e-01 9.45860267e-01
-6.64543211e-01 -1.42226771e-01 -3.24966401e-01 -9.59936261e-01
-1.38736081e+00 5.31063616e-01 5.02586842e-01 -8.26697052e-02
-9.56358314e-01 8.11199665e-01 4.73900378e-01 -6.74628019e-01
3.72682005e-01 -1.01302102e-01 4.15017545e-01 -2.51026243e-01
2.28008494e-01 4.52384144e-01 1.21965706e-01 -7.96255887e-01
-5.44170260e-01 8.49191189e-01 2.88470387e-01 -1.41768664e-01
1.21391988e+00 -2.31959537e-01 -1.82455227e-01 4.96160656e-01
1.08761477e+00 1.79600835e-01 -1.67576933e+00 -5.79363741e-02
-3.64862114e-01 -6.54288769e-01 -3.05865675e-01 -7.87601650e-01
-1.14193010e+00 6.26923740e-01 5.74594796e-01 -4.72062111e-01
7.08220184e-01 1.50742665e-01 3.97088736e-01 2.33926311e-01
8.49278748e-01 -9.95444179e-01 1.99685141e-01 2.98989136e-02
9.63789940e-01 -1.44244874e+00 2.31793091e-01 -8.45799387e-01
-4.19155598e-01 9.95225847e-01 6.36713982e-01 -3.72786492e-01
4.52802837e-01 -1.90722607e-02 7.78588206e-02 -3.13977242e-01
-2.96528876e-01 -2.63080239e-01 8.14974785e-01 9.04974759e-01
3.14178467e-01 3.16619650e-02 1.69107944e-01 8.31521526e-02
-1.57573640e-01 -3.47969204e-01 9.36164483e-02 6.68427527e-01
1.65850863e-01 -1.13126266e+00 -3.79365683e-01 1.56901881e-01
-1.41987085e-01 1.61940157e-01 -6.91994548e-01 8.87998044e-01
2.09989578e-01 6.72291160e-01 -8.52300078e-02 -6.17075525e-03
2.99181581e-01 9.08125937e-03 1.29388094e+00 -6.24343753e-01
-3.29203278e-01 1.80022672e-01 1.41650692e-01 -8.80435109e-01
-8.37626636e-01 -7.86186159e-01 -7.52506196e-01 -2.58345336e-01
-1.22702859e-01 -4.05470699e-01 3.44554156e-01 8.64410639e-01
2.97160178e-01 1.90025195e-01 3.51361215e-01 -1.52107286e+00
-4.15898830e-01 -6.17580891e-01 -6.02102995e-01 1.14219391e+00
3.46525997e-01 -1.12902379e+00 -3.74958277e-01 3.20449889e-01] | [7.122588634490967, -1.1755578517913818] |
684bc794-6358-4da2-984f-60970a72afe0 | spa-vae-similar-parts-assignment-for | 2203.07825 | null | https://arxiv.org/abs/2203.07825v2 | https://arxiv.org/pdf/2203.07825v2.pdf | SPA-VAE: Similar-Parts-Assignment for Unsupervised 3D Point Cloud Generation | This paper addresses the problem of unsupervised parts-aware point cloud generation with learned parts-based self-similarity. Our SPA-VAE infers a set of latent canonical candidate shapes for any given object, along with a set of rigid body transformations for each such candidate shape to one or more locations within the assembled object. In this way, noisy samples on the surface of, say, each leg of a table, are effectively combined to estimate a single leg prototype. When parts-based self-similarity exists in the raw data, sharing data among parts in this way confers numerous advantages: modeling accuracy, appropriately self-similar generative outputs, precise in-filling of occlusions, and model parsimony. SPA-VAE is trained end-to-end using a variational Bayesian approach which uses the Gumbel-softmax trick for the shared part assignments, along with various novel losses to provide appropriate inductive biases. Quantitative and qualitative analyses on ShapeNet demonstrate the advantage of SPA-VAE. | ['Miaomiao Liu', 'Christian Walder', 'Shidi Li'] | 2022-03-15 | null | null | null | null | ['point-cloud-generation'] | ['computer-vision'] | [ 2.38687322e-01 5.60765386e-01 -3.45667712e-02 -4.91408706e-01
-9.46899116e-01 -5.56607485e-01 5.57382882e-01 -1.77338481e-01
3.15666616e-01 4.80104059e-01 9.45191756e-02 5.06911516e-01
-2.01720491e-01 -9.25077677e-01 -1.24158645e+00 -7.48483837e-01
2.34950900e-01 1.29141414e+00 6.95696846e-02 -1.27617754e-02
-9.09070373e-02 8.81196141e-01 -1.55149198e+00 1.05644606e-01
8.74209821e-01 9.92878735e-01 2.95352072e-01 1.50662065e-01
-1.52437732e-01 5.18550863e-04 -1.76568225e-01 -7.55192995e-01
5.98789513e-01 9.46469307e-02 -4.26167756e-01 6.29634798e-01
6.55913532e-01 -4.03941050e-03 -1.58387169e-01 7.80150473e-01
2.59935290e-01 1.96706012e-01 8.78529668e-01 -1.26692104e+00
-3.53435099e-01 7.05470562e-01 -7.47748613e-01 -8.41125011e-01
6.26277924e-03 2.55023032e-01 1.11858857e+00 -1.10805178e+00
9.46846306e-01 1.60885715e+00 7.04394221e-01 3.63061100e-01
-1.77420354e+00 -5.89214146e-01 -3.63734886e-02 -4.92519468e-01
-1.53509319e+00 -5.01759589e-01 1.26198745e+00 -5.44864178e-01
5.20871937e-01 2.42774427e-01 8.96594226e-01 1.05295658e+00
1.22252390e-01 8.09776247e-01 5.73419631e-01 -7.68920034e-02
4.80635464e-01 5.95252663e-02 -3.81711036e-01 6.74943089e-01
2.72272676e-01 -2.30860442e-01 -6.53316975e-01 -6.09848320e-01
1.00489461e+00 9.51575935e-02 3.09693813e-01 -1.24991846e+00
-1.05122745e+00 7.23264158e-01 6.17688239e-01 -2.15802625e-01
-4.47792441e-01 4.11454976e-01 5.92988497e-03 -3.16309541e-01
6.56918347e-01 5.23925781e-01 -4.86695796e-01 4.46647108e-01
-1.08868289e+00 6.49275303e-01 5.59476078e-01 1.41435850e+00
1.15305042e+00 1.81804925e-01 -1.52005583e-01 7.79587865e-01
6.26769304e-01 6.86325073e-01 -1.87945977e-01 -1.26026547e+00
3.76489401e-01 6.01653159e-01 -6.36667665e-03 -9.41793680e-01
-1.10548370e-01 -6.51319563e-01 -8.33717585e-01 3.28586817e-01
-1.25431661e-02 7.14389905e-02 -1.10554492e+00 1.95297289e+00
6.64124012e-01 6.25451654e-02 -2.41767898e-01 6.29880369e-01
3.69348228e-01 4.15133506e-01 -4.20998968e-02 1.35760603e-03
1.14911366e+00 -6.34546816e-01 -2.87529439e-01 -3.77216578e-01
-8.55089575e-02 -7.39765167e-01 7.76920080e-01 2.42615104e-01
-1.27666616e+00 -4.81024384e-01 -7.46496260e-01 -1.21630400e-01
1.01986296e-01 1.27346322e-01 7.37127602e-01 3.33514094e-01
-7.29546964e-01 7.62717724e-01 -1.03548515e+00 1.13515975e-02
9.44139242e-01 3.94179910e-01 -4.71718609e-01 -7.46975243e-02
-3.52587819e-01 4.43938404e-01 2.00836156e-02 1.01418709e-02
-8.77713144e-01 -8.14366758e-01 -1.07142079e+00 -2.04528905e-02
4.20017838e-01 -1.29147220e+00 8.33016217e-01 -8.70147228e-01
-1.30392110e+00 8.30524623e-01 -4.34166938e-01 -2.14474827e-01
6.05083287e-01 -1.10745087e-01 2.58976430e-01 -2.48059884e-01
4.76740837e-01 1.17929828e+00 1.17211914e+00 -1.84477222e+00
9.85843837e-02 -5.63238978e-01 -4.30975676e-01 1.13129102e-01
1.60973072e-01 -4.36991930e-01 -6.70299470e-01 -8.13264132e-01
7.81224966e-01 -1.05366910e+00 -3.31463963e-01 5.90540409e-01
-6.42564535e-01 -2.72538573e-01 5.36223710e-01 -5.50453305e-01
6.09481514e-01 -2.18317747e+00 4.07968491e-01 6.29986823e-01
1.40990987e-01 -5.12298107e-01 -7.77066052e-02 3.48660886e-01
1.78091884e-01 -3.66668999e-02 -8.27687860e-01 -8.50158393e-01
1.72726780e-01 5.20593703e-01 -4.04560506e-01 4.97542411e-01
5.54990828e-01 9.68850315e-01 -7.74486244e-01 -5.83626270e-01
2.40965039e-01 4.83729482e-01 -8.07931960e-01 4.79559004e-02
-7.39849329e-01 3.47189188e-01 -4.68545049e-01 7.68665493e-01
9.03753519e-01 -1.31900534e-01 1.93891525e-02 -3.71253580e-01
3.19279194e-01 9.25328508e-02 -1.37271082e+00 2.12457561e+00
-3.02985877e-01 9.45197269e-02 5.30127771e-02 -6.00265980e-01
1.12033510e+00 6.98380023e-02 6.13929749e-01 -5.45149855e-02
5.51347584e-02 2.04450756e-01 -1.46553099e-01 -1.07224293e-01
3.04658979e-01 -2.58023500e-01 -6.31642044e-02 2.46126935e-01
2.16489330e-01 -7.83068061e-01 -2.13168219e-01 1.43894136e-01
7.16554940e-01 8.54868770e-01 -1.24507211e-02 -3.31651866e-01
-2.02868536e-01 -1.77013706e-02 6.82017386e-01 4.35113460e-01
4.27959353e-01 1.15435255e+00 4.78766710e-02 -1.29026324e-01
-1.40230072e+00 -1.39507687e+00 -3.44966352e-01 6.39738142e-01
4.69985865e-02 -1.74615979e-01 -6.62951946e-01 -3.05235296e-01
5.66442370e-01 8.55765998e-01 -6.11592889e-01 -2.03878626e-01
-4.40513521e-01 -2.99818426e-01 7.69923031e-02 5.41806161e-01
-1.15423039e-01 -1.06884646e+00 -4.31446850e-01 3.93104017e-01
-1.13443375e-01 -7.75224686e-01 -4.82428968e-01 2.13259086e-01
-1.16276133e+00 -7.56820381e-01 -4.91990209e-01 -6.36768103e-01
1.11890018e+00 1.03578940e-01 1.20341718e+00 -2.85620332e-01
-4.42760229e-01 1.63818568e-01 1.32943109e-01 -2.62408823e-01
-3.24916154e-01 -2.05040097e-01 1.75577533e-02 2.51010448e-01
-2.42857710e-02 -9.21381831e-01 -2.26792946e-01 1.73227772e-01
-6.73063338e-01 2.99050838e-01 5.47087848e-01 7.07011700e-01
1.25082004e+00 -1.20429173e-01 1.74862534e-01 -6.63214564e-01
-6.90016383e-03 -4.93854880e-01 -5.65440178e-01 2.61308163e-01
-3.56118768e-01 3.92206550e-01 1.44810095e-01 -5.13438106e-01
-1.02150285e+00 5.49078703e-01 6.12297803e-02 -1.08922100e+00
3.07442658e-02 5.54572791e-02 -5.75003684e-01 1.25175431e-01
5.56557178e-01 1.92272723e-01 1.62623405e-01 -6.47250056e-01
8.25711489e-01 2.12606430e-01 6.74793780e-01 -1.05087650e+00
1.02475786e+00 6.86245978e-01 2.38445222e-01 -7.02733278e-01
-6.28363848e-01 -3.30637634e-01 -8.19678724e-01 -1.04167655e-01
7.52126932e-01 -1.00480604e+00 -5.03909290e-01 4.33171421e-01
-1.14889622e+00 -2.06494834e-02 -7.69624233e-01 1.71653926e-02
-7.85782158e-01 2.15049475e-01 -1.95872992e-01 -1.02319682e+00
-2.80236363e-01 -1.08086050e+00 1.61573756e+00 -5.70036024e-02
-5.05557179e-01 -4.88084674e-01 -1.53039694e-01 4.20553774e-01
-7.32294470e-02 5.77579498e-01 7.69463062e-01 -3.27510417e-01
-9.20214415e-01 -2.61355132e-01 1.42052233e-01 2.28480875e-01
2.18609318e-01 3.43832552e-01 -1.03237271e+00 -2.56384790e-01
-1.20590098e-01 -2.95552611e-01 6.23141408e-01 5.74375868e-01
1.19628870e+00 -3.71945739e-01 -6.25123441e-01 7.13127673e-01
1.27785337e+00 -7.05471113e-02 3.93592238e-01 -2.12676182e-01
9.22442615e-01 7.44876087e-01 4.60116565e-01 6.24606252e-01
3.40311348e-01 6.52302861e-01 5.78867257e-01 9.46013853e-02
-1.51344106e-01 -5.88901877e-01 -1.93553753e-02 6.23767614e-01
1.65315449e-01 -1.66409567e-01 -6.25108778e-01 7.29138434e-01
-1.74402559e+00 -6.45767212e-01 -8.22616071e-02 2.21648264e+00
8.46919417e-01 9.42156017e-02 -4.91758920e-02 -2.54310727e-01
7.73619473e-01 -1.99645036e-03 -7.88743973e-01 2.67099679e-01
-3.78591627e-01 2.91855782e-01 4.01690125e-01 4.03013319e-01
-8.86926174e-01 9.44595158e-01 5.70440674e+00 8.97911131e-01
-7.90456712e-01 -1.40116573e-03 6.13665044e-01 -1.93790361e-01
-9.40788984e-01 3.42221767e-01 -5.77064991e-01 2.79479116e-01
2.20829964e-01 1.40029505e-01 5.60584843e-01 1.08619618e+00
-1.33649167e-02 1.45151436e-01 -1.13569033e+00 8.23938668e-01
-4.13857214e-03 -1.48507297e+00 1.36503160e-01 2.20187306e-01
1.05148733e+00 2.50075571e-02 -1.78562209e-01 -2.73169607e-01
6.13911569e-01 -6.46470487e-01 1.40402281e+00 8.09130847e-01
7.65026867e-01 -7.06618965e-01 4.62728776e-02 3.21689487e-01
-1.11550522e+00 3.11970860e-01 -6.36674285e-01 3.70266110e-01
2.89754957e-01 7.91865587e-01 -7.82841027e-01 5.65406084e-01
4.38206643e-01 4.24678415e-01 -2.08429798e-01 9.45101023e-01
-2.20215410e-01 4.26917732e-01 -8.94997656e-01 3.83283615e-01
-1.74271092e-01 -4.80184287e-01 8.48210812e-01 5.94768465e-01
4.21950847e-01 -2.02514231e-01 1.77054107e-01 1.67653716e+00
-1.89216554e-01 -2.75365442e-01 -5.87816000e-01 1.63156971e-01
8.14031422e-01 1.28775263e+00 -7.82123506e-01 -2.75153458e-01
1.84199870e-01 8.94176006e-01 2.88364023e-01 6.59700781e-02
-4.80272889e-01 1.60810456e-01 7.65259862e-01 4.77721632e-01
4.88619298e-01 -1.52453825e-01 -8.63103688e-01 -8.57489705e-01
3.19681138e-01 -5.64611912e-01 -8.23546126e-02 -1.14877379e+00
-1.47785830e+00 3.43901992e-01 2.32435897e-01 -1.28017116e+00
-1.60477579e-01 -2.87182420e-01 -6.78039312e-01 9.47539210e-01
-6.50344849e-01 -1.66704619e+00 -8.13698992e-02 1.92525446e-01
6.03981674e-01 -4.69190218e-02 6.25063241e-01 -9.38488320e-02
-2.56851047e-01 4.25057590e-01 1.02755755e-01 -2.73725271e-01
4.78601873e-01 -1.10814810e+00 8.86254549e-01 8.28515172e-01
3.05384040e-01 7.83737957e-01 7.38161266e-01 -1.04033267e+00
-1.45113826e+00 -1.24734867e+00 5.85664093e-01 -6.20358825e-01
2.12008223e-01 -8.21647167e-01 -9.22599614e-01 7.14486361e-01
-2.29386851e-01 4.91971597e-02 1.98807418e-01 -6.42707711e-03
-4.10087407e-01 -1.11597821e-01 -1.35252237e+00 6.90480173e-01
1.32374144e+00 -4.33153152e-01 -4.89735991e-01 4.37573165e-01
8.46854568e-01 -4.69396889e-01 -8.38873744e-01 4.10037160e-01
6.46853805e-01 -5.57497501e-01 1.20668674e+00 -5.46028912e-01
5.24621785e-01 -4.67881382e-01 -3.27156603e-01 -1.21574330e+00
-5.09656966e-01 -6.88634813e-01 2.15188116e-02 1.39033568e+00
2.81967312e-01 -2.79214859e-01 9.89567280e-01 7.70955324e-01
-3.97104502e-01 -8.92731309e-01 -1.02723348e+00 -6.32367611e-01
-9.45457909e-03 -5.81563413e-01 9.18420553e-01 7.33859479e-01
-6.53820217e-01 1.78566277e-01 -2.83446819e-01 2.63885349e-01
1.14696646e+00 3.44272465e-01 1.04934573e+00 -1.49551070e+00
-5.70938349e-01 -3.02246869e-01 -2.47779384e-01 -9.77031171e-01
1.98453009e-01 -9.80783761e-01 3.22595298e-01 -1.59259915e+00
2.01492012e-01 -8.44214320e-01 3.03539813e-01 5.95875323e-01
4.04001996e-02 7.71730393e-02 9.95174050e-02 2.86275208e-01
-7.96717852e-02 9.01877999e-01 1.38544893e+00 -3.00343215e-01
-1.00783557e-01 1.74968928e-01 -6.81828380e-01 6.63258672e-01
3.01021069e-01 -7.65473783e-01 -2.80814767e-01 -3.65874708e-01
1.48894459e-01 8.50588307e-02 5.56283772e-01 -8.59593332e-01
-1.64847858e-02 -2.72459894e-01 5.39706051e-01 -7.90434897e-01
8.84646475e-01 -9.44753826e-01 9.20470417e-01 1.77396238e-01
-9.69434157e-02 -4.05251473e-01 3.06653138e-02 7.25888073e-01
1.75637394e-01 -1.89636230e-01 6.98992908e-01 -1.89068645e-01
-1.01515897e-01 4.42351252e-01 1.72060132e-01 -1.11234218e-01
8.38643014e-01 -6.51405275e-01 2.38946259e-01 -1.63995922e-01
-7.49289274e-01 1.66897237e-01 8.21319103e-01 4.48757350e-01
4.62927848e-01 -1.52876902e+00 -7.96843290e-01 5.20178437e-01
2.25541741e-01 9.48738396e-01 3.04066837e-01 3.71306658e-01
-1.35591015e-01 -1.97662011e-01 -2.25019567e-02 -9.46439207e-01
-9.55886543e-01 2.22695157e-01 1.44457728e-01 1.73034340e-01
-8.03627789e-01 1.15875459e+00 4.87658322e-01 -8.32094550e-01
-4.46467698e-02 -2.87409753e-01 5.35488069e-01 -1.27683938e-01
-2.99283564e-01 1.72516748e-01 -1.99425649e-02 -8.52107823e-01
-3.02915126e-01 6.87195897e-01 4.29318964e-01 -3.20646137e-01
1.53085113e+00 1.37763485e-01 -2.25675404e-01 5.30399680e-01
9.06788349e-01 2.28947937e-01 -1.63848329e+00 -3.64434630e-01
-3.56928557e-01 -5.13443768e-01 -1.38379276e-01 -6.38608694e-01
-9.56068754e-01 5.61057329e-01 1.72099724e-01 -2.55611241e-01
6.05187178e-01 4.61800814e-01 5.77386141e-01 2.60916799e-01
6.76849365e-01 -7.40503073e-01 -9.12942961e-02 1.54565275e-01
1.32798243e+00 -8.46987367e-01 9.62236300e-02 -7.89881825e-01
-5.72391987e-01 8.04579079e-01 4.52430308e-01 -3.60557646e-01
6.53965354e-01 2.10931972e-01 -3.65659326e-01 -3.99069220e-01
-5.96879721e-01 1.06493667e-01 4.77832288e-01 6.37798727e-01
-1.04421824e-01 3.74659121e-01 2.59084970e-01 5.49072802e-01
-5.98689258e-01 -3.82624209e-01 1.80251605e-04 8.83507550e-01
-2.22356409e-01 -9.61883068e-01 -4.76821452e-01 6.62563503e-01
1.47122473e-01 9.27830115e-02 -4.41250831e-01 4.53449547e-01
2.98974335e-01 4.40398186e-01 4.36973035e-01 -1.80430368e-01
3.18979234e-01 1.54681027e-01 7.52050638e-01 -7.24211454e-01
-3.13170731e-01 5.44204175e-01 -1.69177711e-01 -4.28946376e-01
-1.38245091e-01 -1.06731367e+00 -1.30706024e+00 -6.02590898e-03
-4.16287690e-01 -2.41930395e-01 7.20114768e-01 8.24767590e-01
6.55128658e-01 4.01816458e-01 6.27174675e-01 -1.38901699e+00
-6.82503939e-01 -9.36749697e-01 -5.00184357e-01 6.62482917e-01
-4.85682301e-02 -9.56452191e-01 -3.26557308e-01 2.16871202e-01] | [8.722339630126953, -3.5489590167999268] |
7afb77ea-7d4e-41ca-8d4e-ff20a51544d5 | deep-active-ensemble-sampling-for-image | 2210.05770 | null | https://arxiv.org/abs/2210.05770v1 | https://arxiv.org/pdf/2210.05770v1.pdf | Deep Active Ensemble Sampling For Image Classification | Conventional active learning (AL) frameworks aim to reduce the cost of data annotation by actively requesting the labeling for the most informative data points. However, introducing AL to data hungry deep learning algorithms has been a challenge. Some proposed approaches include uncertainty-based techniques, geometric methods, implicit combination of uncertainty-based and geometric approaches, and more recently, frameworks based on semi/self supervised techniques. In this paper, we address two specific problems in this area. The first is the need for efficient exploitation/exploration trade-off in sample selection in AL. For this, we present an innovative integration of recent progress in both uncertainty-based and geometric frameworks to enable an efficient exploration/exploitation trade-off in sample selection strategy. To this end, we build on a computationally efficient approximate of Thompson sampling with key changes as a posterior estimator for uncertainty representation. Our framework provides two advantages: (1) accurate posterior estimation, and (2) tune-able trade-off between computational overhead and higher accuracy. The second problem is the need for improved training protocols in deep AL. For this, we use ideas from semi/self supervised learning to propose a general approach that is independent of the specific AL technique being used. Taken these together, our framework shows a significant improvement over the state-of-the-art, with results that are comparable to the performance of supervised-learning under the same setting. We show empirical results of our framework, and comparative performance with the state-of-the-art on four datasets, namely, MNIST, CIFAR10, CIFAR100 and ImageNet to establish a new baseline in two different settings. | ['Donald A. Adjeroh', 'Gianfranco Doretto', 'Salman Mohamadi'] | 2022-10-11 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [-5.61837368e-02 2.91631997e-01 -3.26915205e-01 -6.31129563e-01
-1.46225023e+00 -3.36412370e-01 5.55170536e-01 3.59367371e-01
-8.03827345e-01 1.01131999e+00 -1.02433354e-01 -3.66790518e-02
-4.25677538e-01 -7.57868052e-01 -8.98605585e-01 -8.89917612e-01
-2.20378518e-01 8.79907131e-01 4.52515543e-01 3.36277783e-01
3.48616153e-01 6.27146363e-01 -1.64806986e+00 7.33644813e-02
7.88647890e-01 1.43070996e+00 1.53266434e-02 1.88348740e-01
-4.32614446e-01 7.31633067e-01 -3.64182115e-01 -1.97804093e-01
2.12105751e-01 -1.06557444e-01 -7.54448295e-01 -1.32503375e-01
3.88310105e-01 -3.26014429e-01 3.38721961e-01 7.50797510e-01
7.88705170e-01 3.20280701e-01 8.38485479e-01 -1.28977549e+00
-9.95271578e-02 1.03585243e+00 -7.03747451e-01 4.33946624e-02
-1.34703949e-01 -1.81022570e-01 9.59016442e-01 -1.02508318e+00
4.17691499e-01 1.27865815e+00 6.86850429e-01 4.97549742e-01
-1.31154060e+00 -6.74384594e-01 3.74692380e-01 2.52671272e-01
-1.44704509e+00 -8.11111987e-01 7.69714594e-01 -2.33765557e-01
7.64763296e-01 3.60110365e-02 5.90941548e-01 1.15098739e+00
-2.34631389e-01 1.42255950e+00 1.14222503e+00 -7.98913002e-01
1.06014895e+00 3.25460732e-01 2.54731834e-01 6.24788344e-01
3.15402210e-01 1.20264292e-01 -8.20358574e-01 -4.73056465e-01
5.10575593e-01 -1.86205447e-01 -6.02798015e-02 -8.97169113e-01
-6.17471337e-01 1.13871682e+00 1.07520387e-01 5.38193025e-02
-1.24802120e-01 3.90332401e-01 3.77432972e-01 1.01457566e-01
9.15590763e-01 2.25230679e-01 -7.33740866e-01 -2.05719233e-01
-1.44634187e+00 3.62946361e-01 9.77117777e-01 7.34726250e-01
8.30749154e-01 -1.99795991e-01 -1.32388726e-01 7.54345000e-01
1.01573873e+00 3.40095460e-02 2.81796575e-01 -8.68815243e-01
3.18080366e-01 4.57319170e-01 -3.34070288e-02 -4.03504342e-01
-2.51861632e-01 -4.58629936e-01 -4.03498083e-01 5.27510166e-01
3.93625379e-01 -2.13106662e-01 -7.89427221e-01 1.64285421e+00
4.32219446e-01 2.29025990e-01 -5.90124205e-02 3.70740235e-01
4.93169725e-01 2.31963828e-01 -1.28928246e-02 -4.52563763e-01
8.85723472e-01 -9.32706773e-01 -6.84567988e-01 -8.83880258e-02
6.62781298e-01 -4.61792082e-01 8.78683984e-01 7.10893929e-01
-1.04024327e+00 -2.29371130e-01 -1.11337852e+00 1.32100582e-01
-4.62340564e-01 1.80548906e-01 7.76708186e-01 8.30233753e-01
-1.10153461e+00 6.87057614e-01 -1.20435631e+00 -1.29024282e-01
8.96728218e-01 3.74793500e-01 2.59644818e-02 2.04172239e-01
-9.72622275e-01 6.27595961e-01 7.92921364e-01 -2.38367051e-01
-9.64398682e-01 -9.50765967e-01 -8.82920265e-01 2.15034001e-02
7.40854263e-01 -3.27357471e-01 1.42694318e+00 -8.70055139e-01
-1.77156150e+00 6.06174827e-01 1.04459904e-01 -1.06540132e+00
9.97983873e-01 -6.78570211e-01 1.97285965e-01 4.36941311e-02
-1.01900503e-01 9.54392910e-01 7.40403652e-01 -1.29042113e+00
-6.03717744e-01 -3.11039329e-01 -7.51368478e-02 1.60319954e-01
-2.48773634e-01 -9.86671671e-02 -3.49175632e-01 -5.21654129e-01
1.92101654e-02 -7.89442539e-01 -2.42981970e-01 3.77214134e-01
-2.59858906e-01 -6.50038958e-01 9.19797301e-01 1.06203116e-01
1.05999982e+00 -2.15152621e+00 -7.79847205e-02 1.17347926e-01
6.69636503e-02 2.18660861e-01 2.75497586e-01 3.40037972e-01
1.43331781e-01 9.58006233e-02 -5.03816485e-01 -9.69923973e-01
1.82328701e-01 2.65966326e-01 -4.58070457e-01 4.52107161e-01
2.79602081e-01 8.45784724e-01 -9.10205424e-01 -9.15144622e-01
2.21869841e-01 4.23932999e-01 -4.86264110e-01 1.60306379e-01
-5.79265833e-01 2.46928021e-01 -4.40594882e-01 5.88450670e-01
5.58707952e-01 -1.11728892e-01 -1.05618136e-02 -5.36157712e-02
-1.51890650e-01 4.09218550e-01 -1.54893386e+00 2.01055670e+00
-3.45406264e-01 3.77922595e-01 -8.27583671e-03 -1.10702050e+00
8.74681950e-01 2.98795611e-01 4.69232529e-01 -2.32076168e-01
9.99995619e-02 3.95065457e-01 -2.84580261e-01 -3.47621404e-02
2.70379961e-01 3.15624803e-01 2.48135179e-01 7.23446310e-01
3.69882047e-01 -2.59702448e-02 1.94703996e-01 1.61797643e-01
8.15645456e-01 7.49228418e-01 4.86325979e-01 -6.23105764e-01
3.17610145e-01 -2.63401657e-01 4.09737766e-01 1.00566292e+00
-3.20965976e-01 5.92751443e-01 5.38908184e-01 -2.54725784e-01
-5.18070281e-01 -8.94637346e-01 -5.19639790e-01 1.17113173e+00
-2.99832374e-01 -3.41843665e-01 -9.22633946e-01 -1.12821686e+00
-1.73367813e-01 8.79310668e-01 -7.41856515e-01 1.40109554e-01
-4.29804593e-01 -9.81555223e-01 5.60553908e-01 5.35204649e-01
6.97930574e-01 -8.24451387e-01 -9.25495565e-01 1.96932510e-01
1.54922724e-01 -7.72772074e-01 1.08262487e-01 6.55794203e-01
-1.11651766e+00 -8.85334074e-01 -5.51120937e-01 -1.27331853e-01
4.85679924e-01 -3.63146424e-01 1.14598525e+00 -4.21019316e-01
-1.36262015e-01 4.62261170e-01 -4.39570099e-01 -9.37287748e-01
1.40520942e-03 4.07055944e-01 6.92347363e-02 -8.36255476e-02
2.92132407e-01 -4.28814828e-01 -5.69583535e-01 1.64815351e-01
-8.97877216e-01 -2.23372877e-01 6.58706248e-01 6.98515534e-01
8.87356758e-01 -3.31512168e-02 6.36847198e-01 -1.09028578e+00
3.28853101e-01 -4.67159271e-01 -7.38601983e-01 2.78946400e-01
-8.98087084e-01 5.14639258e-01 2.64003992e-01 -3.72469991e-01
-1.14939046e+00 3.79413277e-01 -2.96348214e-01 -2.69670457e-01
-5.38808592e-02 3.44707489e-01 -2.85585821e-01 -6.24912754e-02
8.64660561e-01 -8.44886601e-02 -1.46968082e-01 -6.90239549e-01
4.17658478e-01 6.03072107e-01 -6.95762644e-03 -8.97786081e-01
2.83271074e-01 6.46735072e-01 -1.02785408e-01 -4.81801331e-01
-1.17778790e+00 -3.49661350e-01 -7.79701471e-01 7.44926333e-02
3.20188403e-01 -7.55878687e-01 -2.89121568e-01 3.44809264e-01
-8.97626817e-01 -5.45673430e-01 -7.82654941e-01 5.15202880e-01
-8.57142866e-01 4.04534340e-01 -3.24973166e-01 -1.20422804e+00
-5.24795830e-01 -1.28338671e+00 1.13436925e+00 1.53498590e-01
-5.13281189e-02 -1.08296442e+00 2.75363296e-01 1.14855655e-01
4.60910559e-01 1.39195055e-01 4.68392134e-01 -1.23421681e+00
-6.23870969e-01 7.22053573e-02 -1.96693167e-01 2.64628321e-01
-2.72183716e-01 -1.78730232e-03 -1.34222734e+00 -2.87696064e-01
1.63339213e-01 -8.24831784e-01 1.12591207e+00 5.33562541e-01
1.43393505e+00 -1.69894353e-01 -3.35779905e-01 4.52834487e-01
1.44689274e+00 -4.41219583e-02 5.80517232e-01 3.52427661e-01
2.06521615e-01 5.62157512e-01 8.94173503e-01 5.41716218e-01
3.01183313e-01 6.42388225e-01 5.56292295e-01 1.88857079e-01
2.41661761e-02 -8.83080289e-02 2.49644324e-01 3.43998998e-01
1.49307489e-01 -1.57281980e-01 -9.44592535e-01 4.67650115e-01
-2.30883479e+00 -6.75974131e-01 2.27889046e-01 2.44829154e+00
1.15036786e+00 4.89632100e-01 1.24637753e-01 4.42799658e-01
1.30436391e-01 1.31494060e-01 -6.40647888e-01 -1.79240704e-01
2.68335044e-01 3.41715962e-01 5.21684289e-01 4.38507080e-01
-1.31843841e+00 7.23814070e-01 6.01055908e+00 1.49548841e+00
-1.00543082e+00 3.96479696e-01 8.43508542e-01 -1.90776885e-01
-7.75443688e-02 -8.68813097e-02 -1.25070012e+00 3.62129897e-01
1.02808726e+00 2.99412221e-01 1.09827761e-02 1.24258280e+00
-1.11075349e-01 -3.96486223e-01 -1.42638922e+00 9.22089636e-01
1.06352024e-01 -1.45622051e+00 -2.87643701e-01 8.08676258e-02
5.59333801e-01 3.41281444e-01 -1.46787181e-01 3.81014109e-01
5.43326378e-01 -9.08907890e-01 8.42904389e-01 5.64517796e-01
5.91138303e-01 -7.53409624e-01 7.24961221e-01 5.64300656e-01
-9.29839909e-01 1.65780976e-01 -3.60797107e-01 1.42921239e-01
4.03399132e-02 1.03861523e+00 -7.69663870e-01 6.05223775e-01
8.37109387e-01 5.37955701e-01 -4.99996394e-01 1.12796187e+00
-2.42192045e-01 7.78991699e-01 -6.40673339e-01 -5.41108735e-02
2.32836455e-01 8.03223327e-02 4.10465240e-01 1.25757265e+00
2.13884413e-02 -3.40937883e-01 9.06402543e-02 1.07588387e+00
-1.30159343e-02 4.68428694e-02 -3.78093392e-01 1.33393973e-01
6.61343396e-01 1.24561942e+00 -9.72840488e-01 -2.54922062e-01
-5.47076464e-02 4.69656914e-01 7.12760031e-01 5.39111234e-02
-6.98844850e-01 -1.44400671e-01 1.53722418e-02 8.24678130e-03
3.51368517e-01 -1.41262576e-01 -2.91060805e-01 -9.56772447e-01
-1.34695426e-01 -5.84433436e-01 4.76862848e-01 -2.64207035e-01
-1.27506638e+00 4.40170228e-01 5.94593048e-01 -8.83007288e-01
-4.84371424e-01 -5.85062861e-01 -3.31287056e-01 6.02669597e-01
-1.59105814e+00 -9.40748155e-01 3.87053974e-02 2.95089453e-01
6.91600263e-01 -1.92896962e-01 8.38572204e-01 2.70210832e-01
-4.53950495e-01 7.19627082e-01 2.53822982e-01 -4.89056520e-02
6.48194015e-01 -1.49530220e+00 2.55587727e-01 7.02080846e-01
4.68157589e-01 5.06301582e-01 4.83643055e-01 -3.94910961e-01
-1.05408585e+00 -9.78252470e-01 5.02287149e-01 -4.95559126e-01
4.60682034e-01 -5.25927305e-01 -7.07737148e-01 5.97300529e-01
-9.03537273e-02 3.15880001e-01 7.16539025e-01 2.57072866e-01
-2.47949541e-01 -1.02063105e-01 -1.42019677e+00 2.62005121e-01
8.84631515e-01 -2.05393359e-01 -5.50575018e-01 1.80197686e-01
6.12012804e-01 -4.50663924e-01 -5.33141613e-01 6.80010676e-01
6.06269002e-01 -1.26292586e+00 9.43921328e-01 -1.21734872e-01
5.63022122e-02 -1.60675608e-02 -3.61661315e-01 -1.06222880e+00
2.37307727e-01 -5.55233657e-01 -6.63190305e-01 1.41286254e+00
3.93696159e-01 -6.81944847e-01 1.03425908e+00 5.44496775e-01
-5.23803905e-02 -1.33441627e+00 -1.11550891e+00 -5.99958420e-01
1.16031937e-01 -7.85170138e-01 3.17175359e-01 6.27904952e-01
-4.63713586e-01 1.07880540e-01 -2.76319161e-02 -3.05297464e-01
9.71212089e-01 -2.44179472e-01 5.22134960e-01 -1.50922632e+00
-4.07956660e-01 -2.93664604e-01 -2.05401346e-01 -1.05210769e+00
1.41160071e-01 -4.35874492e-01 9.06048492e-02 -1.48879409e+00
1.83248296e-01 -9.01204646e-01 -3.19426179e-01 6.42450929e-01
1.99937344e-01 4.55160737e-02 1.09635340e-02 3.43097329e-01
-1.00999665e+00 5.70119321e-01 4.65838164e-01 1.86231032e-01
-3.78993779e-01 1.13826998e-01 -5.90831339e-01 9.74884689e-01
7.44202793e-01 -6.68101490e-01 -6.98443353e-01 -2.63884574e-01
5.14462709e-01 -3.67146164e-01 2.77837306e-01 -1.05088174e+00
3.50287348e-01 7.64540657e-02 2.66564816e-01 -9.40175712e-01
3.48295510e-01 -8.62724423e-01 -1.33183435e-01 3.50613922e-01
-6.31643355e-01 -4.59619164e-01 2.29444131e-01 8.66183281e-01
-3.40749137e-02 -6.65088594e-01 8.57684195e-01 -2.40609348e-01
-5.86485505e-01 3.33513796e-01 2.36024503e-02 2.87935853e-01
1.01282632e+00 -1.10437430e-01 -8.56959820e-02 -5.22320569e-02
-5.49488127e-01 1.32879630e-01 2.11939991e-01 7.98762739e-02
3.63362700e-01 -9.93340194e-01 -5.96301019e-01 6.73820972e-02
5.05939052e-02 5.90867519e-01 -1.43927664e-01 9.58790481e-01
-4.45676535e-01 4.06115770e-01 2.18124479e-01 -8.54453504e-01
-8.83736372e-01 1.66811153e-01 3.18518311e-01 -5.69406509e-01
-3.42060089e-01 1.04858589e+00 1.33679835e-02 -5.77014744e-01
8.74441862e-01 -2.66433477e-01 -2.34234214e-01 4.31061804e-01
5.13140976e-01 4.85747546e-01 3.89247566e-01 -4.14077528e-02
-3.90803099e-01 1.96773052e-01 -2.68054724e-01 -4.27874804e-01
1.29561388e+00 -1.66835770e-01 9.35867354e-02 9.27131832e-01
9.92826402e-01 -9.05635431e-02 -1.56045008e+00 -6.14515185e-01
2.96156675e-01 -3.10906500e-01 4.95897293e-01 -9.56441283e-01
-8.84208918e-01 9.14847314e-01 9.83755529e-01 2.06896350e-01
8.34489346e-01 9.36154947e-02 2.75427878e-01 4.38208491e-01
6.88485682e-01 -1.21457386e+00 -1.64557323e-02 2.67955720e-01
6.48028791e-01 -1.43944240e+00 3.14343244e-01 -2.11680084e-01
-4.95668083e-01 1.07662177e+00 4.02155578e-01 -3.92131582e-02
9.32582855e-01 4.99939293e-01 -1.05991170e-01 -2.02003062e-01
-8.44045699e-01 -1.68997765e-01 4.28235739e-01 3.75122249e-01
5.71436524e-01 -2.55373657e-01 -3.38956803e-01 4.51426685e-01
1.59100428e-01 3.01911682e-01 -6.30794466e-03 1.12274730e+00
-5.97490609e-01 -1.16745019e+00 -2.72221565e-01 4.89221454e-01
-6.28237069e-01 -9.80602801e-02 -2.70824075e-01 8.62572491e-01
1.27254829e-01 7.96836853e-01 1.09034531e-01 2.82645643e-01
-2.11004913e-01 2.12340519e-01 6.65028334e-01 -7.52605677e-01
-2.30204761e-01 1.68357894e-01 9.80744138e-02 -5.82490206e-01
-8.21635962e-01 -7.38676190e-01 -1.22998405e+00 2.02335745e-01
-6.35749578e-01 2.38640875e-01 9.21000659e-01 1.20662463e+00
3.87576073e-01 2.45284945e-01 2.06555396e-01 -9.12758052e-01
-9.87324178e-01 -9.85937059e-01 -5.90290427e-01 -1.42672509e-01
1.08160838e-01 -9.51752901e-01 -4.87839431e-01 -3.23582888e-01] | [8.982686042785645, 3.87395977973938] |
c0c302bb-1503-48f3-9c22-a6924404476e | sentinel-2-time-series-analysis-with-3d | null | null | https://www.mdpi.com/2220-9964/10/7/483/htm | https://www.mdpi.com/2220-9964/10/7/483/pdf | Sentinel 2 Time Series Analysis with 3D Feature Pyramid Network and Time Domain Class Activation Intervals for Crop Mapping | In this paper, we provide an innovative contribution in the research domain dedicated to crop mapping by exploiting the of Sentinel-2 satellite images time series, with the specific aim to extract information on “where and when” crops are grown. The final goal is to set up a workflow able to reliably identify (classify) the different crops that are grown in a given area by exploiting an end-to-end (3+2)D convolutional neural network (CNN) for semantic segmentation. The method also has the ambition to provide information, at pixel level, regarding the period in which a given crop is cultivated during the season. To this end, we propose a solution called Class Activation Interval (CAI) which allows us to interpret, for each pixel, the reasoning made by CNN in the classification determining in which time interval, of the input time series, the class is likely to be present or not. Our experiments, using a public domain dataset, show that the approach is able to accurately detect crop classes with an overall accuracy of about 93% and that the network can detect discriminatory time intervals in which crop is cultivated. These results have twofold importance: (i) demonstrate the ability of the network to correctly interpret the investigated physical process (i.e., bare soil condition, plant growth, senescence and harvesting according to specific cultivated variety) and (ii) provide further information to the end-user (e.g., the presence of crops and its temporal dynamics). | ['Mirco Boschetti', 'Nicola Landro', 'Riccardo La Grassa', 'Ignazio Gallo'] | 2021-10-07 | null | null | null | isprs-international-journal-of-geo-1 | ['unet-segmentation'] | ['computer-vision'] | [ 4.08266366e-01 5.59423566e-02 8.23478475e-02 -1.29786059e-01
-7.69230351e-02 -1.15169036e+00 5.32861471e-01 8.83712471e-01
-3.92725356e-02 3.99736375e-01 -5.44372439e-01 -7.81399369e-01
-3.47558707e-01 -1.30231082e+00 -6.65061414e-01 -9.51286972e-01
-4.83232737e-01 2.55996615e-01 -6.83222637e-02 -3.50145012e-01
-2.69435644e-01 1.10950339e+00 -1.89854729e+00 2.27222338e-01
6.25591040e-01 1.26127124e+00 5.45751393e-01 7.20117688e-01
-9.45914071e-03 3.32934350e-01 -5.93161285e-01 4.80860293e-01
7.57520646e-02 -2.42503375e-01 -6.17345333e-01 3.04653674e-01
-8.41117650e-03 -2.65662950e-02 4.28981990e-01 1.07189739e+00
-9.14752111e-02 -3.78362775e-01 7.14013934e-01 -8.22793663e-01
8.58540833e-02 6.67694688e-01 -2.98648715e-01 1.10456198e-01
-1.17489398e-01 3.73310924e-01 5.37846923e-01 -2.19430223e-01
6.56028271e-01 8.96796286e-01 6.18564367e-01 -2.58455396e-01
-1.18472540e+00 -2.73003131e-01 3.08342636e-01 -9.95347202e-02
-1.17441666e+00 3.65256481e-02 4.94822055e-01 -8.38597417e-01
5.45419633e-01 3.59315157e-01 9.87385988e-01 6.49697900e-01
1.56654403e-01 6.25464499e-01 1.26583362e+00 -2.75997967e-01
5.62327266e-01 -2.91262746e-01 1.00989379e-01 1.30104631e-01
1.91818103e-01 2.28097335e-01 1.72883779e-01 2.40425363e-01
4.77210432e-01 -6.01645038e-02 -1.94828715e-02 -1.26439467e-01
-1.00466275e+00 6.14837468e-01 9.53202605e-01 6.48567677e-01
-1.10361123e+00 7.11144209e-02 4.76488769e-01 1.56385466e-01
7.41212964e-01 2.71789730e-01 -8.69123816e-01 4.64652359e-01
-1.08651555e+00 2.28637472e-01 6.04639232e-01 4.03670073e-01
8.79380167e-01 -1.70707747e-01 -1.03999130e-01 8.18137079e-02
2.05646232e-01 8.42825532e-01 -6.56236336e-02 -4.59010482e-01
-7.04914406e-02 5.64327359e-01 2.83651054e-01 -9.86832559e-01
-6.77724838e-01 -3.90711337e-01 -9.22071218e-01 4.60376292e-01
5.43008804e-01 -3.45186949e-01 -1.13160312e+00 1.49942327e+00
1.08908400e-01 4.11515050e-02 4.29578900e-01 8.02541792e-01
4.60505486e-01 7.65712738e-01 4.59318697e-01 -9.99960527e-02
1.79387486e+00 1.28350630e-02 -6.31627798e-01 -2.97611058e-01
5.47518313e-01 -4.93386179e-01 5.22550642e-01 2.65222698e-01
-4.82307136e-01 -5.74322224e-01 -1.12427950e+00 8.39334249e-01
-1.14759052e+00 4.22810376e-01 9.12847936e-01 2.48819754e-01
-9.23118353e-01 7.07678854e-01 -9.39995289e-01 -8.75743687e-01
3.58078778e-01 1.28789712e-03 -2.29014859e-01 3.30790728e-01
-1.26086056e+00 8.58098507e-01 9.51443970e-01 7.43201554e-01
-9.06331599e-01 -6.99067473e-01 -6.35674179e-01 2.97442853e-01
1.54078037e-01 -6.42445087e-02 8.10113370e-01 -1.48347855e+00
-1.05986118e+00 1.28127301e+00 6.21420071e-02 -8.75505984e-01
4.37073529e-01 1.73520803e-01 -4.26613003e-01 2.49193594e-01
1.70783564e-01 6.21447623e-01 5.02911568e-01 -1.39572942e+00
-1.04114556e+00 -6.72277987e-01 1.88201725e-01 -1.58988699e-01
-1.14658557e-01 -1.22050367e-01 5.45201004e-02 -4.80392188e-01
3.27232748e-01 -9.25148964e-01 -2.05523819e-01 1.09933326e-02
-2.32191309e-01 9.72461030e-02 9.93307650e-01 -9.22346592e-01
8.14992189e-01 -2.14830303e+00 -1.46395853e-02 3.96856964e-01
-8.01000372e-02 5.41384816e-01 1.87197417e-01 4.24627811e-01
-1.56386927e-01 2.57643163e-01 -7.09412158e-01 4.16849285e-01
-3.36238176e-01 3.01510602e-01 -3.14895272e-01 3.72520596e-01
6.13416195e-01 5.36576152e-01 -7.89897561e-01 -6.64457083e-02
5.27031124e-01 4.12334293e-01 4.93458986e-01 2.29790777e-01
-7.24450111e-01 6.50498331e-01 -3.74969453e-01 6.15897477e-01
1.10224903e+00 2.01216713e-01 5.06537318e-01 -1.39296770e-01
-7.89931536e-01 -3.67545903e-01 -9.70426083e-01 1.16411495e+00
-5.56999981e-01 7.75813818e-01 2.66319215e-01 -1.12248409e+00
1.19256592e+00 4.61531043e-01 4.38978255e-01 -5.28718650e-01
3.71041931e-02 4.52526838e-01 -1.84902623e-01 -4.73414123e-01
2.40216836e-01 3.17212999e-01 8.17304105e-03 4.68745641e-02
-1.74280077e-01 -6.83865398e-02 3.52468848e-01 -5.40604651e-01
6.50118947e-01 3.71168673e-01 2.64236987e-01 -6.99679554e-01
5.11340320e-01 5.01338482e-01 1.89261004e-01 4.45878625e-01
-2.40401432e-01 2.75697857e-01 9.22540426e-01 -7.58113384e-01
-8.98906350e-01 -7.85678864e-01 -2.42701441e-01 7.48663008e-01
2.09479555e-01 5.85567713e-01 -7.03058362e-01 -4.49295461e-01
1.58825636e-01 7.96594143e-01 -1.01018381e+00 1.18283764e-01
-2.90292680e-01 -9.56481099e-01 5.60826778e-01 3.90227169e-01
6.87349617e-01 -1.45197117e+00 -1.26191247e+00 3.99519682e-01
-2.08936095e-01 -1.08720815e+00 6.26002967e-01 8.21998060e-01
-9.82181847e-01 -1.15982902e+00 -5.65934241e-01 -5.39313614e-01
5.47503173e-01 1.82343528e-01 1.13754070e+00 -1.29497916e-01
-3.50985706e-01 -1.96703613e-01 -5.29287040e-01 -6.48096025e-01
-5.98996043e-01 2.85579413e-01 -8.24230433e-01 5.18223792e-02
4.50007439e-01 -3.78541827e-01 -6.18322432e-01 2.57227600e-01
-1.23941839e+00 -7.05667015e-04 4.91690099e-01 1.32459059e-01
5.58529854e-01 3.62406403e-01 2.53264517e-01 -6.55222654e-01
1.34148031e-01 -9.02021468e-01 -1.09198618e+00 4.88021642e-01
-4.57550138e-01 -1.50419891e-01 4.84320015e-01 -8.02621543e-02
-8.01260531e-01 5.44515252e-01 -9.33272615e-02 1.24366373e-01
-9.92077291e-01 1.04138982e+00 -1.84104517e-01 3.47975999e-01
6.22528732e-01 1.74766153e-01 4.61159796e-02 -3.24859083e-01
9.79638845e-02 5.92445970e-01 6.87146425e-01 -1.31671682e-01
5.38608432e-01 6.74565196e-01 2.69358754e-01 -8.77864778e-01
-6.74129128e-01 -6.34638727e-01 -9.54230368e-01 -6.42293453e-01
1.09264660e+00 -8.95067692e-01 -6.86109245e-01 7.79016197e-01
-1.14973783e+00 -6.67789757e-01 1.89923607e-02 2.30475053e-01
-3.73553485e-01 -1.77266598e-01 -3.15835164e-03 -1.01187062e+00
-4.50642169e-01 -8.21599782e-01 1.04635727e+00 4.73474294e-01
4.72689755e-02 -1.11803901e+00 -3.56248796e-01 -2.97891438e-01
5.21043956e-01 1.05160475e+00 9.49813247e-01 -4.15509731e-01
-1.60412520e-01 -2.70298272e-01 -4.44150269e-01 2.83878863e-01
2.12408021e-01 6.27584398e-01 -1.15608323e+00 7.97092170e-02
-2.89090455e-01 3.12782675e-01 8.18366468e-01 8.34892809e-01
9.48715091e-01 3.16077471e-01 -2.82312423e-01 5.22216976e-01
1.78606915e+00 3.41033131e-01 7.06528842e-01 3.91878933e-01
1.46446541e-01 9.83554602e-01 1.18068659e+00 5.15485942e-01
6.16586916e-02 5.63539326e-01 1.30061233e+00 -7.37607419e-01
9.79293138e-02 2.68722087e-01 3.35216880e-01 -2.59024233e-01
-1.07387155e-01 -4.46874231e-01 -1.18290555e+00 8.45082402e-01
-1.96668708e+00 -8.97448540e-01 -5.24059117e-01 2.24452949e+00
3.15833569e-01 1.50807336e-01 7.48618692e-02 3.71209055e-01
8.67165089e-01 2.68536240e-01 -5.39807081e-01 -5.24418235e-01
-5.94709992e-01 2.48901695e-01 1.11597109e+00 2.51235098e-01
-1.54870057e+00 1.13647664e+00 5.05424738e+00 3.47992510e-01
-1.64109588e+00 -4.03689891e-01 7.45737553e-01 8.32500398e-01
2.51714349e-01 9.88779590e-02 -5.57751179e-01 1.75344244e-01
1.00200438e+00 3.01079869e-01 2.79052287e-01 5.09279013e-01
6.28909707e-01 -5.60503960e-01 -7.23156393e-01 7.99201205e-02
-4.30580437e-01 -1.20075333e+00 -3.53385210e-01 -1.67820673e-03
4.50300276e-01 -4.59253564e-02 -2.41114363e-01 -9.35280919e-02
1.40311942e-01 -9.55392718e-01 8.42545927e-01 6.88474238e-01
5.62595248e-01 -6.27655923e-01 1.01104760e+00 3.98075461e-01
-1.48964071e+00 -6.72378093e-02 -8.44351854e-03 -8.65443945e-02
-2.25263119e-01 7.93424904e-01 -8.87559950e-01 9.79811013e-01
6.80932701e-01 6.80520296e-01 -4.87563580e-01 8.48692417e-01
-4.14126873e-01 8.06354821e-01 -4.76250499e-01 2.01130599e-01
7.14931309e-01 -1.17054269e-01 3.39833915e-01 1.22054911e+00
5.51793456e-01 -1.29494935e-01 6.54973537e-02 9.99902725e-01
5.86293340e-01 -1.44168243e-01 -6.47420585e-01 -1.95516273e-01
2.58835286e-01 1.30133879e+00 -1.37456071e+00 -1.14844672e-01
2.99443185e-01 8.22082520e-01 -4.15122420e-01 3.97918493e-01
-3.97934228e-01 -4.26560521e-01 7.48558581e-01 -1.87731721e-02
6.11244380e-01 -2.99868524e-01 -4.20846671e-01 -4.23672885e-01
6.91256523e-02 -3.58503342e-01 2.54707605e-01 -1.01492476e+00
-6.66876137e-01 5.54173350e-01 -4.79161292e-02 -9.90970790e-01
-3.43950093e-01 -9.54997838e-01 -6.54797018e-01 1.33428240e+00
-1.74736357e+00 -1.44436252e+00 -7.93739796e-01 2.05435976e-02
2.17528373e-01 2.75083184e-01 1.16404915e+00 -1.18528701e-01
-2.57007688e-01 -3.06896478e-01 2.44234428e-01 3.23740803e-02
7.72309527e-02 -1.37032223e+00 3.67804289e-01 9.99726593e-01
-3.49788725e-01 -2.14491427e-01 8.73161137e-01 -5.72691441e-01
-1.01411033e+00 -1.45726478e+00 9.40823734e-01 1.14269525e-01
6.49298429e-01 -6.26348704e-02 -7.38800943e-01 3.34372461e-01
-1.47241652e-01 1.44274101e-01 2.18576416e-01 -1.01720154e-01
-5.25903068e-02 -3.96540970e-01 -1.36881876e+00 1.32501364e-01
4.36366707e-01 -3.04069787e-01 6.61290959e-02 4.72215891e-01
4.83795881e-01 -1.57317325e-01 -9.31123078e-01 7.08231091e-01
5.50737739e-01 -9.29392278e-01 7.05227077e-01 -3.14721465e-01
6.13374949e-01 -3.58747870e-01 -6.54351041e-02 -1.36806750e+00
-1.20087951e-01 -1.41183138e-01 2.29393214e-01 1.08587384e+00
3.94413829e-01 -5.67186415e-01 4.44780022e-01 -1.08040102e-01
-5.15830284e-03 -4.54456925e-01 -5.04580379e-01 -7.03148782e-01
9.88426954e-02 -4.56049770e-01 1.02876353e+00 8.10840487e-01
-8.29893291e-01 -4.83746409e-01 3.03820014e-01 9.96585608e-01
1.38325199e-01 3.71900797e-01 3.66963744e-01 -1.67509949e+00
3.65419388e-01 -7.09370553e-01 -5.08476079e-01 -2.42126688e-01
-5.96211245e-03 -6.40003264e-01 1.43055156e-01 -1.42147112e+00
-5.03041148e-01 -6.75828278e-01 -3.16007584e-01 6.86358154e-01
-5.07296175e-02 5.54879718e-02 1.17688999e-02 1.49544224e-01
2.76406795e-01 -1.77974433e-01 7.15157807e-01 -2.19195932e-01
-2.91162878e-01 4.84959662e-01 -3.43422592e-01 4.09501076e-01
9.72252250e-01 -1.80723101e-01 1.22280143e-01 -3.95740360e-01
3.45496058e-01 2.64172226e-01 8.88719916e-01 -1.19026446e+00
-2.95722485e-01 -1.05238892e-01 2.05883861e-01 -9.32757497e-01
-2.88451135e-01 -1.17542851e+00 4.28719252e-01 7.91264892e-01
-3.29921216e-01 -2.70170480e-01 6.54307723e-01 5.23281768e-02
-2.93847799e-01 -3.25413913e-01 5.99329174e-01 5.17908530e-03
-1.07719290e+00 2.53405213e-01 -5.90639770e-01 -5.87556779e-01
1.13747084e+00 -3.22965160e-03 6.24115318e-02 -4.62076478e-02
-9.88019109e-01 3.90097141e-01 3.07125419e-01 4.21142995e-01
-1.93954155e-01 -7.90645063e-01 -9.84382153e-01 2.21990854e-01
6.09981477e-01 -3.93949822e-02 2.18436375e-01 4.92238283e-01
-1.07016599e+00 4.62448418e-01 -4.59008753e-01 -9.48087990e-01
-1.16112328e+00 3.53402197e-01 5.75348914e-01 -3.82273853e-01
-3.08099836e-01 4.27046001e-01 -2.48872310e-01 -3.46405745e-01
-7.86156654e-02 -6.35263324e-01 -7.04764187e-01 7.21640766e-01
1.70461640e-01 -1.83106273e-01 4.89085972e-01 -5.47389865e-01
-2.71151662e-01 3.58017921e-01 5.97666979e-01 7.88062215e-02
1.40003788e+00 1.83788911e-01 -3.30024928e-01 7.54661739e-01
7.43056595e-01 -6.35083377e-01 -1.56445909e+00 -3.52858528e-02
3.53739172e-01 -2.41572261e-02 2.06106201e-01 -1.16625059e+00
-1.29145193e+00 8.72375786e-01 1.30743814e+00 9.43248987e-01
1.50498509e+00 -1.62241027e-01 2.11682424e-01 1.91711277e-01
3.87947112e-02 -9.23277557e-01 -1.02834737e+00 5.50785542e-01
8.14721406e-01 -1.40492094e+00 -2.72900641e-01 -4.73117381e-01
-1.80221289e-01 1.57737005e+00 6.70948401e-02 7.23499358e-02
7.44154572e-01 2.22843200e-01 3.80374312e-01 -4.26981330e-01
-2.11916655e-01 -8.73786569e-01 -9.16780457e-02 8.46687436e-01
3.45490992e-01 8.47050011e-01 -1.92296460e-01 1.41030923e-01
7.25275949e-02 1.34507015e-01 1.06625959e-01 8.20249081e-01
-4.19439077e-01 -5.99641144e-01 -5.94340503e-01 2.29253083e-01
-1.18661471e-01 9.70407277e-02 -5.32809436e-01 7.80003071e-01
5.58784842e-01 9.48068738e-01 3.48003358e-01 -1.43600916e-02
3.02128524e-01 -8.23399872e-02 9.63400900e-02 -3.40971053e-01
-9.31692660e-01 -1.88645765e-01 -1.13597244e-01 -2.67851740e-01
-6.34549797e-01 -9.01675761e-01 -1.06855261e+00 -9.81166512e-02
-2.66843010e-02 -1.69954803e-02 1.41003418e+00 9.93342996e-01
1.82189733e-01 8.26067746e-01 8.49945068e-01 -1.13414562e+00
-8.62625167e-02 -9.29471314e-01 -7.66133189e-01 1.39482155e-01
3.25436652e-01 -3.68630379e-01 -2.80881643e-01 1.49534464e-01] | [9.34663200378418, -1.573038101196289] |
5513cce4-20b8-4a35-9d7f-c44998d9af3b | an-efficient-encoder-decoder-architecture | 2209.15200 | null | https://arxiv.org/abs/2209.15200v5 | https://arxiv.org/pdf/2209.15200v5.pdf | An efficient encoder-decoder architecture with top-down attention for speech separation | Deep neural networks have shown excellent prospects in speech separation tasks. However, obtaining good results while keeping a low model complexity remains challenging in real-world applications. In this paper, we provide a bio-inspired efficient encoder-decoder architecture by mimicking the brain's top-down attention, called TDANet, with decreased model complexity without sacrificing performance. The top-down attention in TDANet is extracted by the global attention (GA) module and the cascaded local attention (LA) layers. The GA module takes multi-scale acoustic features as input to extract global attention signal, which then modulates features of different scales by direct top-down connections. The LA layers use features of adjacent layers as input to extract the local attention signal, which is used to modulate the lateral input in a top-down manner. On three benchmark datasets, TDANet consistently achieved competitive separation performance to previous state-of-the-art (SOTA) methods with higher efficiency. Specifically, TDANet's multiply-accumulate operations (MACs) are only 5\% of Sepformer, one of the previous SOTA models, and CPU inference time is only 10\% of Sepformer. In addition, a large-size version of TDANet obtained SOTA results on three datasets, with MACs still only 10\% of Sepformer and the CPU inference time only 24\% of Sepformer. | ['Xiaolin Hu', 'Runxuan Yang', 'Kai Li'] | 2022-09-30 | null | null | null | null | ['speech-separation'] | ['speech'] | [ 2.21152455e-01 -4.01821919e-02 2.70318776e-01 -2.98686530e-02
-8.58170807e-01 3.44371684e-02 2.91783929e-01 -2.64051259e-01
-6.04966581e-01 4.89938796e-01 6.38962444e-03 -2.18343779e-01
3.67547455e-03 -5.11125565e-01 -7.00925469e-01 -9.00293648e-01
4.16363850e-02 1.55965984e-01 3.86428714e-01 -8.04817528e-02
1.66969728e-02 3.56272519e-01 -1.47309005e+00 3.91923815e-01
8.26705158e-01 1.23803818e+00 5.45278668e-01 8.53264630e-01
-2.73794699e-02 5.21643460e-01 -7.37640083e-01 -2.92525649e-01
7.37414286e-02 -5.72832882e-01 -4.43094850e-01 -3.50519717e-01
3.55450571e-01 -4.55770604e-02 -5.39400041e-01 1.02886093e+00
8.99834037e-01 1.04507044e-01 5.62035620e-01 -9.80444014e-01
-7.67597556e-01 8.12761664e-01 -6.13683760e-01 5.97316563e-01
-2.23403528e-01 1.69782303e-02 1.05673540e+00 -1.31365836e+00
-5.18419407e-02 1.24792230e+00 4.87082303e-01 4.66481775e-01
-1.21860611e+00 -9.20576394e-01 2.57335037e-01 3.38719994e-01
-1.43690002e+00 -8.40487897e-01 5.86464226e-01 -8.69954154e-02
1.46016228e+00 3.22495401e-01 5.23254454e-01 9.15540338e-01
4.35663462e-01 8.99494290e-01 8.29327881e-01 -2.65720844e-01
2.38951653e-01 -8.55670497e-02 1.19811103e-01 5.69190383e-01
-3.68341058e-02 -1.23284079e-01 -9.38212812e-01 1.88344836e-01
6.75402343e-01 -1.67100042e-01 -3.27285498e-01 3.35282654e-01
-1.18621981e+00 5.71197808e-01 6.14202380e-01 5.53443551e-01
-5.92093766e-01 2.03303114e-01 2.59838820e-01 2.95650780e-01
3.86990786e-01 3.55294168e-01 -5.95481098e-01 -5.13470843e-02
-1.20722437e+00 -2.97309667e-01 5.10657251e-01 8.92031908e-01
5.26873350e-01 5.39945483e-01 -3.49278420e-01 1.17585087e+00
2.24457487e-01 6.24781013e-01 8.03552032e-01 -6.50406480e-01
6.38345301e-01 1.14837997e-01 -5.04513383e-01 -9.06023562e-01
-3.73641640e-01 -1.22540200e+00 -1.15889180e+00 -4.26696911e-02
1.53119013e-01 -2.56640941e-01 -9.40144360e-01 1.76374233e+00
-1.66686773e-01 1.83186755e-01 1.26808807e-01 9.41222727e-01
7.04321921e-01 9.13718581e-01 -1.46172240e-01 -3.52492064e-01
1.52104545e+00 -1.29905200e+00 -8.46971273e-01 -7.84900963e-01
2.05243319e-01 -8.68466973e-01 6.52286887e-01 4.94959742e-01
-1.56866312e+00 -8.75340283e-01 -1.25839281e+00 -1.22193985e-01
-2.84245461e-01 6.14864945e-01 3.76428902e-01 4.18769866e-01
-1.26626480e+00 3.80745202e-01 -7.50679910e-01 6.92356303e-02
5.35970807e-01 6.14744604e-01 -1.78989828e-01 1.63545415e-01
-1.30853474e+00 7.07057714e-01 8.81284550e-02 3.32433075e-01
-1.00441825e+00 -6.73727274e-01 -8.54826510e-01 6.91851258e-01
3.11137974e-01 -6.86547101e-01 1.04713356e+00 -7.34555304e-01
-1.91118193e+00 2.38520801e-01 -4.55280811e-01 -6.86277211e-01
-2.25639477e-01 -2.92664200e-01 -5.57721972e-01 6.90531433e-02
3.31861991e-03 7.65689313e-01 1.16445875e+00 -7.59044170e-01
-6.09963655e-01 -2.75463730e-01 -4.13668871e-01 1.55518707e-02
-5.50327539e-01 7.10792392e-02 -6.85469925e-01 -9.10927713e-01
3.26063842e-01 -8.93935680e-01 4.27465774e-02 -3.10482144e-01
-2.66767502e-01 -6.11681230e-02 6.04192555e-01 -8.93050313e-01
1.53426254e+00 -2.47869396e+00 4.98786718e-01 -2.36631587e-01
2.88370013e-01 6.46050096e-01 -2.87573397e-01 1.58488527e-01
-2.08039165e-01 -2.49266624e-02 -5.56899130e-01 -5.30799806e-01
1.04501896e-01 -1.25557140e-01 -2.94810623e-01 1.98518634e-01
3.98773462e-01 7.81034350e-01 -6.90767944e-01 -2.42518067e-01
3.38966846e-02 4.98978078e-01 -6.99260533e-01 1.83729708e-01
3.51367831e-01 1.15953973e-02 -4.70910929e-02 3.07637513e-01
6.59621298e-01 2.80837119e-02 -4.38185930e-02 -1.67189851e-01
-1.20142490e-01 7.88346887e-01 -8.18513989e-01 1.61110044e+00
-7.32653081e-01 1.07376683e+00 4.93582070e-01 -9.68889654e-01
9.33656275e-01 5.42785883e-01 6.07846938e-02 -8.55704546e-01
3.11563581e-01 4.49929625e-01 5.93746185e-01 -2.07714707e-01
9.95831937e-02 -2.05911383e-01 1.55962780e-02 2.56530553e-01
3.07857603e-01 -4.27960642e-02 -1.24387220e-01 7.66374692e-02
1.08294642e+00 -3.47891122e-01 1.30871296e-01 -3.42315674e-01
7.46020138e-01 -6.05815291e-01 6.97739899e-01 4.54382837e-01
-2.02251136e-01 5.42693794e-01 2.72865027e-01 -6.03580624e-02
-7.02974677e-01 -9.92665768e-01 -3.53267193e-02 1.28236401e+00
-9.41177607e-02 -4.43806082e-01 -1.07894707e+00 -3.39580417e-01
-2.65827596e-01 5.82710862e-01 -4.06664908e-01 -3.47260386e-01
-7.13842452e-01 -7.62832403e-01 7.75340199e-01 7.70871699e-01
7.04353273e-01 -1.13848531e+00 -6.17131293e-01 3.60606641e-01
2.91243196e-02 -1.10920811e+00 -7.26749241e-01 6.79140985e-01
-5.58467507e-01 -3.80721718e-01 -9.70317721e-01 -7.88829327e-01
5.29427230e-01 3.22568625e-01 8.04585159e-01 -2.73102522e-01
-1.05199963e-01 -4.25425917e-01 -2.38134079e-02 -5.13742447e-01
6.08727522e-02 2.94795752e-01 1.65333271e-01 1.79694697e-01
3.79352838e-01 -8.22038531e-01 -4.99564856e-01 1.89586833e-01
-6.06826842e-01 1.82850525e-01 9.34266746e-01 1.10829258e+00
3.51947367e-01 1.58950105e-01 9.69684303e-01 -3.32064033e-01
4.76822376e-01 -3.26136410e-01 -4.05908763e-01 -1.77823082e-01
-2.03419015e-01 2.01176122e-01 8.94189060e-01 -3.93353403e-01
-9.41409647e-01 -2.87898034e-01 -4.23386425e-01 -5.69264650e-01
1.57555059e-01 2.99130052e-01 -3.37077975e-01 1.99603841e-01
2.05170661e-01 5.37084639e-01 -2.29404077e-01 -6.30125105e-01
-1.00868568e-01 9.36323822e-01 4.55977112e-01 -1.76931679e-01
4.63477403e-01 8.82446617e-02 -3.29934597e-01 -9.58269656e-01
-7.16370642e-01 -1.03048734e-01 -5.91406286e-01 1.63601354e-01
9.75927591e-01 -9.77701068e-01 -5.22058129e-01 7.17002749e-01
-1.16733646e+00 -3.24071378e-01 -6.21528625e-02 8.75306845e-01
-2.15508789e-01 -8.23195502e-02 -7.56352365e-01 -8.43052626e-01
-6.16456747e-01 -1.39454627e+00 1.02603662e+00 2.82918394e-01
-1.11698151e-01 -5.89870811e-01 -4.59722251e-01 3.97237509e-01
7.61439860e-01 -5.37508368e-01 8.98239017e-01 -5.53374290e-01
-4.76889670e-01 1.48321509e-01 -3.21628809e-01 6.41013861e-01
1.53846974e-02 -3.18451554e-01 -1.39525926e+00 -3.49893212e-01
3.35536212e-01 1.79326996e-01 1.19416749e+00 5.66501379e-01
1.31448078e+00 -7.80954435e-02 -2.66311556e-01 7.90873945e-01
9.09706473e-01 6.77562177e-01 5.51686108e-01 -7.98181444e-02
5.15169382e-01 2.48165295e-01 1.98478103e-01 9.08024758e-02
1.29003093e-01 8.57031584e-01 2.75420129e-01 -2.09100619e-01
-6.33965671e-01 1.92279205e-01 6.60644770e-01 1.74473727e+00
8.29828009e-02 -3.21836621e-01 -7.42417216e-01 7.44937062e-01
-1.57087672e+00 -7.99513459e-01 5.76338805e-02 2.06507111e+00
8.69886696e-01 3.59892219e-01 -2.81589836e-01 4.50002670e-01
6.13984406e-01 1.92452893e-01 -5.38665473e-01 -8.33723366e-01
-4.45786342e-02 7.14152575e-01 2.20300302e-01 4.63524342e-01
-9.56273437e-01 8.53858292e-01 5.80003548e+00 1.08991349e+00
-1.26276422e+00 3.33196670e-01 5.14090240e-01 -4.22344774e-01
1.47797406e-01 -3.97878736e-01 -8.43566716e-01 5.64340293e-01
1.28363907e+00 1.12503529e-01 6.22040927e-01 5.60048163e-01
-9.11304504e-02 9.05559883e-02 -9.97200787e-01 1.00462997e+00
1.27348125e-01 -8.60654533e-01 -1.55628562e-01 1.20090671e-01
4.54082608e-01 1.21241778e-01 1.42533913e-01 4.79088217e-01
-1.12419784e-01 -9.92558539e-01 6.66130304e-01 4.61704373e-01
7.85905361e-01 -9.14235175e-01 8.63785148e-01 4.15095091e-01
-1.40702796e+00 -3.00641418e-01 -5.99804521e-01 -1.83772027e-01
5.85975908e-02 7.98298538e-01 -6.52583361e-01 5.40011108e-01
7.47636497e-01 5.73356390e-01 -2.63672084e-01 8.27943325e-01
-3.33986580e-01 8.00697386e-01 -3.04002762e-01 8.76183156e-03
2.81748265e-01 2.05373839e-02 6.89346611e-01 1.53949046e+00
5.10906458e-01 9.90490466e-02 -3.70417625e-01 7.15496242e-01
-3.70412767e-01 -2.11736456e-01 -9.79275554e-02 1.63908392e-01
6.55549824e-01 1.07155013e+00 -3.85310888e-01 -4.73367959e-01
-2.42037237e-01 1.03631353e+00 4.09836978e-01 3.63858789e-01
-1.12204051e+00 -9.16623652e-01 8.43893945e-01 -1.47510558e-01
6.96162522e-01 -1.62133977e-01 -3.46927762e-01 -8.89150262e-01
-3.23935449e-02 -9.85817850e-01 1.27938995e-02 -8.08783591e-01
-9.43814158e-01 1.09576571e+00 -5.20930707e-01 -1.02403462e+00
1.54289410e-01 -6.28277719e-01 -4.97313589e-01 1.23856604e+00
-1.42861354e+00 -6.69284940e-01 -1.23012073e-01 3.61829311e-01
9.43563282e-01 -4.14482296e-01 7.47484446e-01 6.26767337e-01
-1.13196671e+00 9.43553448e-01 -9.41907540e-02 1.99582666e-01
6.22725010e-01 -1.11144555e+00 5.06331146e-01 9.71015215e-01
3.11329722e-01 7.73473561e-01 2.94763833e-01 -2.42428601e-01
-1.34173989e+00 -1.00974381e+00 9.74399388e-01 9.85702425e-02
4.88781363e-01 -5.49272537e-01 -1.10250258e+00 3.88610959e-01
5.72713137e-01 1.09478824e-01 5.72707295e-01 -4.28118221e-02
-2.71637201e-01 -5.06834865e-01 -6.12535596e-01 5.32408297e-01
1.02982199e+00 -6.43108130e-01 -7.67870307e-01 -2.41661981e-01
8.17969024e-01 -2.89752066e-01 -7.34608054e-01 2.61746347e-01
6.55357718e-01 -8.81332636e-01 1.05680192e+00 -8.00288245e-02
2.55473912e-01 -3.75650495e-01 -1.68425798e-01 -1.46585298e+00
-6.94046676e-01 -7.27584898e-01 -2.77660817e-01 1.20654857e+00
5.84711552e-01 -8.46298456e-01 8.70848820e-02 5.50017059e-02
-7.74711847e-01 -9.69784379e-01 -1.26539743e+00 -7.71910667e-01
1.81963500e-02 -5.11472166e-01 5.33208251e-01 4.86693710e-01
-7.60635315e-03 7.49035895e-01 -3.16186339e-01 3.67439061e-01
3.14297855e-01 -9.98482928e-02 3.28207612e-01 -9.26924169e-01
-5.24364412e-01 -7.96001971e-01 -2.03574568e-01 -1.34418726e+00
-4.48168926e-02 -9.41977739e-01 3.33569854e-01 -1.41685843e+00
1.99341611e-03 -1.52580708e-01 -6.90009296e-01 5.31180084e-01
-3.25328737e-01 2.38414153e-01 3.55690032e-01 1.40013816e-02
-2.09590152e-01 7.36928582e-01 1.21698248e+00 -3.15608859e-01
-1.44969374e-01 -1.93632275e-01 -6.12911224e-01 7.63248861e-01
7.70210862e-01 -5.95573008e-01 -3.73061121e-01 -8.59465063e-01
-2.39946380e-01 3.44705023e-02 1.29795820e-01 -1.54207528e+00
5.79763412e-01 3.39430869e-01 4.64267343e-01 -5.90777040e-01
7.78695941e-01 -5.81039548e-01 -1.12045579e-01 5.54718792e-01
-2.73712844e-01 -9.92060229e-02 4.95557010e-01 2.23663211e-01
-4.85050648e-01 4.82005671e-05 9.64426756e-01 3.39493155e-01
-2.81960487e-01 1.32463932e-01 -5.98351419e-01 -1.36123538e-01
7.25202858e-01 -6.61420589e-03 -3.01990390e-01 -7.13045448e-02
-7.83520699e-01 -5.24465516e-02 -3.48835677e-01 3.89297843e-01
6.62697077e-01 -1.18746841e+00 -8.29095185e-01 7.31454372e-01
-5.85094035e-01 -7.39463493e-02 4.11571473e-01 1.16152954e+00
-1.19606324e-01 9.04325366e-01 -3.98595966e-02 -5.43391824e-01
-1.15197754e+00 6.14949822e-01 3.80187362e-01 -1.17452726e-01
-5.22003889e-01 1.46008313e+00 7.46129096e-01 5.88427521e-02
1.54308081e-01 -4.93639529e-01 -1.38726965e-01 2.26850286e-01
7.09553301e-01 3.99390489e-01 2.33408570e-01 -5.08493900e-01
-7.11083889e-01 5.73951840e-01 -2.92652342e-02 -2.11559266e-01
1.37376857e+00 -2.50280630e-02 -1.02895193e-01 3.65609348e-01
1.11758184e+00 2.30694905e-01 -1.21434748e+00 -2.85864502e-01
-2.57133007e-01 -1.35657325e-01 5.18760204e-01 -8.24329913e-01
-1.38894701e+00 1.60438609e+00 3.83881688e-01 2.30543077e-01
1.61742318e+00 -1.06302232e-01 1.16065216e+00 7.34051168e-02
2.13031679e-01 -9.49170291e-01 2.55955346e-02 6.91445172e-01
1.12332726e+00 -7.36449540e-01 -2.58972019e-01 -2.74374515e-01
-4.62264389e-01 8.54097784e-01 6.76375568e-01 -6.01515211e-02
6.75168633e-01 7.10694969e-01 -2.36474425e-01 1.01151213e-01
-1.08183467e+00 -1.88535035e-01 4.20396596e-01 2.47964919e-01
4.63580459e-01 4.56280224e-02 -3.83842476e-02 1.23505390e+00
-2.85277247e-01 -3.55192065e-01 -1.07587613e-02 5.58930695e-01
-5.65920532e-01 -8.08073044e-01 -3.21333557e-01 3.09487969e-01
-5.58761775e-01 -6.07715428e-01 -1.87009618e-01 4.09195334e-01
3.15645009e-01 1.03733969e+00 1.59621745e-01 -5.32661498e-01
2.21637189e-01 2.02524468e-01 1.93094358e-01 -5.84851205e-01
-8.50265741e-01 5.50866961e-01 3.87270632e-03 -3.54586154e-01
-4.55918349e-02 -5.43679237e-01 -1.28674567e+00 -5.41192330e-02
-3.74232441e-01 2.21906140e-01 5.35194874e-01 1.01188970e+00
7.46587873e-01 1.33034003e+00 5.67952931e-01 -9.68486428e-01
-1.79925770e-01 -1.29254603e+00 -3.10586661e-01 -2.17924386e-01
5.61535656e-01 -6.02934062e-01 -4.90644664e-01 -7.14819282e-02] | [14.78361701965332, 5.8300933837890625] |
6a7245c1-63a6-431e-bc1d-825d4542d241 | proxy-graph-matching-with-proximal-matching | null | null | https://ojs.aaai.org/index.php/AAAI/article/view/17179 | https://ojs.aaai.org/index.php/AAAI/article/view/17179/16986 | Proxy Graph Matching with Proximal Matching Networks | Estimating feature point correspondence is a common technique in computer vision. A line of recent data-driven approaches utilizing the graph neural networks improved the matching accuracy by a large margin. However, these learning-based methods require a lot of labeled training data, which are expensive to collect. Moreover, we find most methods are sensitive to global transforms, for example, a random rotation. On the contrary, classical geometric approaches are immune to rotational transformation though their performance is generally inferior. To tackle these issues, we propose a new learning-based matching framework, which is designed to be rotationally invariant. The model only takes geometric information as input. It consists of three parts: a graph neural network to generate a high-level local feature, an attention-based module to normalize the rotational transform, and a global feature matching module based on proximal optimization. To justify our approach, we provide a convergence guarantee for the proximal method for graph matching. The overall performance is validated by numerical experiments. In particular, our approach is trained on the synthetic random graphs and then applied to several real-world datasets. The experimental results demonstrate that our method is robust to rotational transform and highlights its strong performance of matching accuracy. | ['Cheng-Lin Liu', 'Xu-Yao Zhang', 'Tie-Qiang Wang', 'Sitong Wu', 'Chuang Wang', 'Haoru Tan'] | 2021-10-16 | null | null | null | aaai-2021-10 | ['graph-matching'] | ['graphs'] | [ 7.87354410e-02 -7.58005232e-02 -2.67757118e-01 -3.28662306e-01
-7.35526979e-01 -1.99118242e-01 6.46686256e-01 -6.40216842e-03
-1.10163145e-01 9.19723287e-02 -2.00011492e-01 -1.47852734e-01
-2.72363573e-01 -9.42433953e-01 -9.07537699e-01 -6.95986986e-01
2.35797450e-01 2.43578121e-01 1.80065379e-01 -3.27661991e-01
4.81158853e-01 6.76648915e-01 -1.47565079e+00 -3.37451369e-01
8.76449347e-01 1.03477836e+00 -8.66464451e-02 2.33591944e-01
7.86693692e-02 6.57847285e-01 -5.94203100e-02 -2.61749685e-01
4.90373611e-01 -2.79541165e-01 -6.11402512e-01 1.18921906e-01
6.47748947e-01 -1.74901798e-01 -3.36573899e-01 1.30351651e+00
6.34444535e-01 1.04352027e-01 5.87524533e-01 -1.31772506e+00
-8.40522170e-01 3.44051093e-01 -7.75459826e-01 -3.35708112e-01
3.03748339e-01 -5.12387343e-02 1.03832304e+00 -1.19245565e+00
6.28680885e-01 1.14880061e+00 8.77787709e-01 2.20012486e-01
-7.91285932e-01 -4.48948562e-01 9.65795964e-02 2.23440453e-01
-1.47627783e+00 -2.94211477e-01 1.44793916e+00 -4.23441440e-01
5.16512632e-01 4.03373279e-02 6.31839156e-01 6.58345997e-01
1.51910886e-01 5.68423986e-01 8.13738346e-01 -6.47452652e-01
1.42155245e-01 -2.49604568e-01 2.98095793e-02 9.15060401e-01
2.94193089e-01 3.10901791e-01 -1.57224149e-01 -1.00649901e-01
9.25989449e-01 1.83906868e-01 -3.57620955e-01 -1.05707800e+00
-1.12468791e+00 9.67283130e-01 9.56996143e-01 1.81285203e-01
-2.61000037e-01 1.99224874e-01 2.45615616e-01 2.51655668e-01
4.74662185e-01 2.03925282e-01 -1.46828115e-01 3.74551743e-01
-6.91049874e-01 1.37974784e-01 6.01822495e-01 1.10025609e+00
8.27661633e-01 -9.54371840e-02 1.14469059e-01 8.51640642e-01
5.34434855e-01 5.88806033e-01 5.25570095e-01 -5.83501577e-01
5.95510662e-01 8.09187114e-01 -1.38429597e-01 -1.71746981e+00
-6.14623010e-01 -5.22794247e-01 -1.09215260e+00 1.26603529e-01
3.88142288e-01 2.32606038e-01 -7.03065932e-01 1.82520556e+00
5.18746734e-01 6.87175170e-02 -2.82726258e-01 1.12991524e+00
7.40061462e-01 3.28598350e-01 -2.67905265e-01 -1.39079513e-02
1.00901484e+00 -1.17587698e+00 -4.94913012e-01 1.02228142e-01
8.02627385e-01 -9.37184036e-01 9.77159083e-01 5.49353510e-02
-9.57353294e-01 -7.70010293e-01 -1.19031429e+00 -6.08479232e-02
-3.31672251e-01 1.87344477e-01 5.19158065e-01 5.14330924e-01
-9.70847011e-01 7.13797212e-01 -6.73774481e-01 -5.30792832e-01
2.73077875e-01 3.05995584e-01 -3.67976606e-01 -3.10242027e-02
-9.58076358e-01 6.96346462e-01 4.19024915e-01 3.41658801e-01
-2.87398398e-01 -3.69851589e-01 -1.18293786e+00 1.91180613e-02
2.90224880e-01 -1.04806817e+00 9.53128040e-01 -9.27636147e-01
-1.57555604e+00 8.64450932e-01 1.47062108e-01 -2.38369450e-01
7.23611236e-01 -7.48196319e-02 -1.12486288e-01 1.09036379e-01
3.38550918e-02 3.36315304e-01 1.10027254e+00 -1.18568087e+00
-3.59362036e-01 -4.19402927e-01 -7.75202513e-02 2.23737359e-01
-3.71153891e-01 -2.21467614e-01 -5.01030028e-01 -7.77451158e-01
4.98733222e-01 -9.62612033e-01 -3.02975297e-01 8.62976909e-02
-4.13814604e-01 -2.49890208e-01 6.52988195e-01 -2.93402463e-01
9.34777141e-01 -2.06710243e+00 -3.41266878e-02 4.90688890e-01
1.68572366e-01 1.43063411e-01 -1.86825424e-01 4.27455366e-01
-3.45078468e-01 -1.83087334e-01 -1.30883858e-01 -1.25023112e-01
7.41009489e-02 -3.22686225e-01 -2.20726594e-01 8.83842945e-01
2.51907796e-01 1.15834212e+00 -8.25062871e-01 -6.39324367e-01
6.60454035e-01 5.40242434e-01 -5.33872128e-01 2.72310019e-01
2.06799526e-02 1.61551178e-01 -5.63705623e-01 6.01315320e-01
8.66526067e-01 -4.47980285e-01 -6.84380531e-02 -6.32674754e-01
5.45992330e-03 -3.02568585e-01 -1.30518913e+00 1.81441319e+00
-2.98258036e-01 1.38730317e-01 -2.80148178e-01 -1.39281261e+00
1.32521605e+00 -9.57909375e-02 6.33296251e-01 -6.44588590e-01
5.39103627e-01 4.02632385e-01 -1.55603573e-01 -3.48188341e-01
3.14903855e-01 8.33579376e-02 6.47881031e-02 3.12655210e-01
-1.26113802e-01 -2.16283321e-01 -3.80001850e-02 1.43247331e-02
8.97661448e-01 3.24397951e-01 4.04168159e-01 -2.63222307e-01
9.50008273e-01 -3.52336243e-02 4.83254880e-01 4.94570374e-01
-1.72610328e-01 7.26239324e-01 7.76389390e-02 -7.44761348e-01
-1.06474948e+00 -7.21262693e-01 3.34295854e-02 6.23951375e-01
7.00940311e-01 -2.44010121e-01 -9.35125291e-01 -5.76635063e-01
-1.08585261e-01 1.75843716e-01 -4.77827251e-01 -5.77567279e-01
-7.92697906e-01 -6.10114396e-01 1.37616768e-01 5.22083759e-01
6.78820789e-01 -8.58760178e-01 -3.47284317e-01 5.27036935e-02
5.74845597e-02 -1.09544289e+00 -6.80574834e-01 -3.49613547e-01
-9.11599040e-01 -1.27632892e+00 -8.72752547e-01 -1.10498559e+00
1.01699066e+00 6.49349451e-01 8.93802702e-01 2.93161184e-01
-2.68412948e-01 3.76593828e-01 -3.43559682e-01 -1.87686592e-01
-1.07226335e-01 3.29886734e-01 -8.36441591e-02 3.73317838e-01
2.32695803e-01 -5.94651341e-01 -7.26264119e-01 5.29836655e-01
-7.71568835e-01 1.37783542e-01 8.70345950e-01 9.96019840e-01
6.38046682e-01 -1.80716619e-01 4.32794750e-01 -7.84021974e-01
5.14888644e-01 -2.01600656e-01 -8.47493291e-01 2.41007119e-01
-6.45997047e-01 2.11841211e-01 9.18761849e-01 -3.16919953e-01
-7.37186432e-01 4.47201431e-01 1.94394239e-03 -5.61633110e-01
8.05509537e-02 5.81291437e-01 -2.47640595e-01 -6.35398448e-01
6.02855980e-01 2.26170078e-01 9.30647850e-02 -2.54112273e-01
5.98526776e-01 3.88918519e-01 5.46555340e-01 -6.02144122e-01
1.42061603e+00 5.81185699e-01 2.56814361e-01 -8.10746074e-01
-7.25290358e-01 -4.54213113e-01 -7.38417864e-01 -2.98787296e-01
5.09385228e-01 -7.14076042e-01 -7.75893033e-01 6.48015976e-01
-9.91142213e-01 -7.82095790e-02 -1.20572545e-01 6.28341615e-01
-9.24576640e-01 6.55378699e-01 -4.43551332e-01 -4.16555583e-01
-5.41838467e-01 -1.15636063e+00 1.10906017e+00 3.06393474e-01
2.35334799e-01 -1.06789231e+00 2.20720530e-01 1.45191744e-01
3.62429112e-01 5.41705251e-01 8.47140253e-01 -5.75594842e-01
-7.38817751e-01 -4.29300994e-01 -5.18084466e-01 1.00420460e-01
1.78934142e-01 9.43906009e-02 -6.76605642e-01 -5.47927737e-01
2.40905434e-02 -2.72286922e-01 5.57739973e-01 3.26396704e-01
1.05421817e+00 -4.44837511e-02 -3.27801585e-01 8.47850621e-01
1.66604304e+00 -4.84449081e-02 3.37469816e-01 5.05108595e-01
1.06773829e+00 5.57077050e-01 7.70844042e-01 1.07103772e-01
4.31691885e-01 8.21194172e-01 7.04468966e-01 -3.66990566e-01
-1.05643474e-01 -4.80080694e-01 9.97234955e-02 1.21042180e+00
-1.38854146e-01 1.16809204e-01 -7.74852276e-01 3.79655451e-01
-2.04616833e+00 -7.63619125e-01 -8.80409479e-02 2.34458876e+00
3.59405100e-01 8.30909014e-02 4.47434708e-02 2.26389483e-01
1.01863241e+00 1.99537188e-01 -5.12277782e-01 -1.59667939e-01
4.87284958e-02 5.48260054e-03 5.28012097e-01 3.69271845e-01
-1.20280123e+00 1.08930802e+00 5.44933128e+00 7.57932246e-01
-1.32216835e+00 -3.05790752e-01 3.70196015e-01 4.80875552e-01
-1.52930498e-01 -8.34769011e-03 -4.08700854e-01 2.82197386e-01
2.65720278e-01 -2.30751708e-01 2.87949115e-01 1.12819219e+00
1.56429216e-01 4.48804528e-01 -1.04749870e+00 1.29427397e+00
2.83669233e-01 -1.29542625e+00 2.47041821e-01 -7.71102607e-02
6.76859379e-01 -7.29308650e-02 -5.84788471e-02 7.88170472e-02
-2.52001453e-04 -8.72834325e-01 5.65216422e-01 5.62197387e-01
6.32952213e-01 -7.96099901e-01 7.22179234e-01 3.12313199e-01
-1.44102848e+00 1.75345987e-01 -5.22601902e-01 1.10123686e-01
-2.87545342e-02 5.60512245e-01 -4.06234562e-01 1.04343498e+00
3.80816847e-01 8.87227058e-01 -7.25214899e-01 1.22278070e+00
-2.29695320e-01 1.12661995e-01 -2.91064918e-01 -1.40697271e-01
1.96840391e-01 -4.85239565e-01 2.72706151e-01 8.09367061e-01
3.41537297e-01 -1.70158714e-01 3.36064309e-01 7.08009601e-01
-9.82988849e-02 6.29675686e-01 -9.81918156e-01 3.83278310e-01
3.58736098e-01 1.41235614e+00 -7.17877448e-01 -5.93894497e-02
-4.77357864e-01 7.71512449e-01 6.18148506e-01 2.59737432e-01
-9.05280411e-01 -7.09529877e-01 3.86193469e-02 -5.33869192e-02
1.98397025e-01 -2.48140246e-01 -8.49254727e-02 -1.24965858e+00
3.64619672e-01 -7.50240684e-01 2.36389920e-01 -6.94950223e-01
-1.30433810e+00 4.83926505e-01 -2.23221138e-01 -1.60193384e+00
-2.99396992e-01 -7.31279314e-01 -6.01985574e-01 6.38861477e-01
-1.58582699e+00 -1.43484080e+00 -7.09907055e-01 6.94603384e-01
2.58195311e-01 -1.66759878e-01 4.82521445e-01 5.75267732e-01
-5.62127233e-01 8.42006207e-01 9.99398995e-03 5.41666389e-01
6.95918500e-01 -1.10255945e+00 4.73189950e-01 9.12337065e-01
2.68217266e-01 6.85013056e-01 4.77864683e-01 -3.46939415e-01
-1.76310039e+00 -1.12717569e+00 7.20218241e-01 -1.20598219e-01
6.42675877e-01 -1.20813437e-01 -8.34497809e-01 6.57922924e-01
-2.42391557e-01 4.65924770e-01 1.42589569e-01 -1.89222038e-01
-3.94996345e-01 -2.36164168e-01 -1.01573122e+00 6.94723845e-01
1.23085892e+00 -4.20538306e-01 -6.01970077e-01 3.93500358e-01
5.51877558e-01 -5.15931606e-01 -8.97120178e-01 6.25458002e-01
5.53881288e-01 -1.03843200e+00 9.78867412e-01 -4.09234971e-01
3.41558605e-01 -4.81618792e-01 1.98714249e-02 -1.27139544e+00
-3.70453447e-01 -6.45239770e-01 9.20647681e-02 1.18035150e+00
-1.41548598e-02 -1.01013923e+00 8.56338680e-01 3.21551174e-01
-8.96299407e-02 -8.36746514e-01 -7.51158655e-01 -9.68446791e-01
7.39778718e-03 -2.51366556e-01 6.69810653e-01 1.16841412e+00
-2.11958021e-01 3.97260070e-01 -3.43073636e-01 1.94873765e-01
8.53588998e-01 5.11049151e-01 1.17185104e+00 -1.19123340e+00
6.40849769e-02 -6.21122122e-01 -8.18780005e-01 -1.06728458e+00
1.84491307e-01 -9.81746674e-01 -6.23196550e-02 -1.39233518e+00
4.17884849e-02 -5.98773181e-01 -1.80615842e-01 1.35453254e-01
-2.26217359e-01 2.22764716e-01 7.66392611e-03 3.31172734e-01
-3.64222616e-01 8.12871873e-01 1.34442854e+00 -1.61043331e-01
-1.27223909e-01 -3.98086384e-03 -5.50417006e-01 9.38024938e-01
8.60682905e-01 -3.08266938e-01 -4.41575527e-01 -3.21569979e-01
2.63982594e-01 -2.25767761e-01 3.91720563e-01 -1.02038240e+00
4.97393429e-01 -1.18923455e-01 1.88607022e-01 -5.75422585e-01
8.44516754e-02 -1.01294649e+00 6.01180680e-02 5.95867693e-01
-1.62696481e-01 1.38740152e-01 -3.49200279e-01 5.10836720e-01
-3.27827126e-01 -5.43770492e-02 8.81303489e-01 1.60133615e-01
-5.14283180e-01 7.61590183e-01 4.46543127e-01 6.58763275e-02
9.90948021e-01 -3.21711212e-01 -1.73888221e-01 -3.62854630e-01
-1.94800809e-01 9.78879556e-02 6.47286236e-01 4.30304080e-01
5.61871767e-01 -1.81248975e+00 -6.25507653e-01 3.99271756e-01
4.72507358e-01 1.87856957e-01 5.41226333e-03 7.76532054e-01
-8.48796308e-01 3.79988074e-01 -9.54292044e-02 -8.49686623e-01
-9.96545136e-01 9.27916706e-01 4.75255609e-01 -7.17358366e-02
-7.61923075e-01 4.48222578e-01 9.44572762e-02 -6.93158388e-01
1.42534748e-01 -3.69287878e-01 -1.20932981e-01 -1.75745100e-01
1.35212913e-01 2.97499716e-01 1.06727071e-01 -8.28998744e-01
-3.60940039e-01 1.45698047e+00 1.05199806e-01 1.63200185e-01
1.12241459e+00 3.07279360e-02 -8.16026926e-02 1.32622853e-01
1.49838245e+00 -8.85735676e-02 -1.05399251e+00 -5.59589982e-01
1.61898598e-01 -5.34807265e-01 -1.90232709e-01 1.37625318e-02
-1.36519837e+00 7.85586059e-01 5.39098263e-01 3.93862754e-01
1.06143916e+00 -2.28938401e-01 8.58144522e-01 4.68178779e-01
4.94656414e-01 -1.02940011e+00 7.34616444e-02 3.07082295e-01
8.71740878e-01 -1.42403209e+00 1.56041935e-01 -6.35141313e-01
5.16076908e-02 1.41055036e+00 5.22039831e-01 -6.09409273e-01
7.39215076e-01 -1.92945957e-01 2.22656637e-01 -3.16765189e-01
-1.54896498e-01 -1.28195763e-01 5.33916116e-01 4.60427195e-01
2.65563488e-01 -1.99518055e-01 -5.31906486e-01 9.88519043e-02
-3.60218346e-01 -5.66297807e-02 1.18148260e-01 8.06440651e-01
-2.98147649e-01 -1.04658496e+00 -3.27337831e-01 8.91751796e-02
-1.58045858e-01 1.73421845e-01 -2.92645484e-01 1.08530319e+00
-2.07864836e-01 6.36258960e-01 -2.57132888e-01 -4.02504891e-01
6.01440668e-01 -2.92047232e-01 5.36146641e-01 -2.14255810e-01
-2.69772977e-01 -6.24303408e-02 -4.20716554e-01 -7.66973376e-01
-7.37273872e-01 -3.55724573e-01 -9.81139839e-01 -1.96134299e-01
-4.89234805e-01 1.04248952e-02 5.92030287e-01 8.13240826e-01
2.27280527e-01 2.63128519e-01 1.08009756e+00 -8.81205320e-01
-8.10883343e-01 -6.15245640e-01 -3.86181295e-01 8.11523378e-01
1.75404668e-01 -7.70745575e-01 -4.78236526e-01 -8.67641270e-02] | [8.435141563415527, -2.199864625930786] |
d04df750-6725-4c20-aedb-51b7230e4211 | proto-clip-vision-language-prototypical | 2307.03073 | null | https://arxiv.org/abs/2307.03073v2 | https://arxiv.org/pdf/2307.03073v2.pdf | Proto-CLIP: Vision-Language Prototypical Network for Few-Shot Learning | We propose a novel framework for few-shot learning by leveraging large-scale vision-language models such as CLIP. Motivated by the unimodal prototypical networks for few-shot learning, we introduce PROTO-CLIP that utilizes image prototypes and text prototypes for few-shot learning. Specifically, PROTO-CLIP adapts the image encoder and text encoder in CLIP in a joint fashion using few-shot examples. The two encoders are used to compute prototypes of image classes for classification. During adaptation, we propose aligning the image and text prototypes of corresponding classes. Such a proposed alignment is beneficial for few-shot classification due to the contributions from both types of prototypes. We demonstrate the effectiveness of our method by conducting experiments on benchmark datasets for few-shot learning as well as in the real world for robot perception. | ['Yu Xiang', 'Xinya Du', 'Yu-Wei Chao', 'Kamalesh Palanisamy', 'Jishnu Jaykumar P'] | 2023-07-06 | null | null | null | null | ['few-shot-image-classification', 'few-shot-learning'] | ['computer-vision', 'methodology'] | [ 1.53283164e-01 3.06024705e-03 -3.68712813e-01 -5.05467713e-01
-7.02911854e-01 -7.45538920e-02 8.30375135e-01 -7.92703778e-02
-4.18964505e-01 2.49223009e-01 2.86740303e-01 3.00226122e-01
1.90876350e-01 -6.78025603e-01 -9.88667548e-01 -4.30939376e-01
1.13332324e-01 1.58428892e-01 3.77029240e-01 -6.85180910e-03
1.53881386e-01 3.27654667e-02 -1.84849203e+00 4.04863864e-01
3.41160774e-01 9.35658395e-01 5.97229838e-01 6.51608348e-01
-2.00211003e-01 1.05598593e+00 -1.62352681e-01 -3.00042719e-01
1.88126922e-01 -4.88258183e-01 -3.51624310e-01 4.33341503e-01
6.45599306e-01 -5.56744039e-01 -6.33696973e-01 1.19597816e+00
3.10565501e-01 7.45728850e-01 8.99363279e-01 -1.27289009e+00
-8.86356771e-01 7.15071082e-01 -5.16188204e-01 8.09888244e-02
1.47208124e-01 5.39211929e-01 8.80243301e-01 -1.52507508e+00
8.90823245e-01 1.28093517e+00 4.13672477e-01 8.33051622e-01
-8.57157588e-01 -4.31632519e-01 9.87005085e-02 7.00873733e-01
-1.39472902e+00 -8.55793178e-01 8.14621449e-01 -4.95775521e-01
1.06098557e+00 -4.68434423e-01 7.17233717e-01 1.28455150e+00
6.19687289e-02 1.00378311e+00 4.66544360e-01 -6.11018658e-01
7.60892391e-01 1.74054056e-01 2.87343264e-01 8.32908690e-01
-4.05496806e-02 6.64857775e-02 -6.34408474e-01 6.46359101e-02
3.66437495e-01 6.02231681e-01 -5.98251037e-02 -8.06146324e-01
-9.23155010e-01 1.03755951e+00 5.07259548e-01 1.56170532e-01
-4.65115726e-01 2.91802049e-01 6.16866767e-01 2.02596024e-01
1.37548313e-01 9.06656012e-02 7.70835057e-02 -6.91093057e-02
-6.89282477e-01 -8.63175765e-02 6.98950589e-01 1.46899939e+00
1.02947092e+00 2.07511157e-01 -5.63420415e-01 1.24676776e+00
1.49933055e-01 4.16064441e-01 8.37066472e-01 -1.03117347e+00
1.02302678e-01 1.78773582e-01 -6.51951209e-02 -6.34950280e-01
9.00508165e-02 6.36681840e-02 -6.33123517e-01 -1.76809132e-02
-3.52703065e-01 2.83015799e-02 -1.22039068e+00 1.60579169e+00
-6.96994970e-03 7.74411619e-01 2.47842431e-01 7.24890351e-01
9.00638819e-01 7.48260975e-01 1.57450661e-01 -2.82883257e-01
1.08012807e+00 -1.46109378e+00 -4.58128810e-01 -4.99748051e-01
5.54563165e-01 -3.21603149e-01 1.22411847e+00 -8.79993662e-02
-8.47914457e-01 -8.42269242e-01 -1.17913842e+00 5.55382930e-02
-4.94193047e-01 -1.30540237e-01 1.96977928e-01 1.50157332e-01
-7.22159684e-01 4.04680878e-01 -7.57274806e-01 -8.17847431e-01
6.62156641e-01 -2.21856415e-01 -2.15440050e-01 -5.29005110e-01
-8.51593673e-01 6.88452005e-01 6.73980236e-01 -4.56308484e-01
-1.41195440e+00 -4.62831914e-01 -1.45742154e+00 2.83811629e-01
4.24834281e-01 -7.71211982e-01 1.47431564e+00 -6.37659252e-01
-1.41926098e+00 6.05077028e-01 -1.75479248e-01 -8.23358536e-01
1.07421063e-01 -1.33458242e-01 -3.48837554e-01 3.53502125e-01
2.11428925e-01 1.06219625e+00 1.22819602e+00 -1.07205474e+00
-6.15170062e-01 -9.13261920e-02 5.71414717e-02 2.30237052e-01
-6.75595701e-01 -2.18525946e-01 -7.67147183e-01 -3.69605780e-01
-3.10525686e-01 -6.21346474e-01 -3.67736876e-01 3.73716980e-01
-1.48418352e-01 -1.88128993e-01 8.61045718e-01 2.60006413e-02
7.83687711e-01 -2.34391499e+00 -6.98513165e-02 -3.38880032e-01
1.09402686e-01 3.55447263e-01 -5.37241995e-01 4.96120393e-01
1.29374146e-01 -4.30717319e-01 -2.27801189e-01 -5.73351860e-01
9.73519757e-02 4.57828373e-01 -5.56102753e-01 3.68943453e-01
2.25732028e-01 1.08884215e+00 -1.19654667e+00 -5.69257915e-01
7.60789692e-01 3.43580753e-01 -3.66598606e-01 5.53236365e-01
-2.58022130e-01 -4.40475941e-01 -2.38539875e-02 6.53456748e-01
2.33664289e-01 -2.37983584e-01 -1.83823735e-01 -1.97158679e-01
1.19528130e-01 -3.31025183e-01 -7.38686383e-01 2.09051895e+00
-4.55297351e-01 5.65550625e-01 -3.82023036e-01 -1.16377985e+00
1.01314795e+00 1.58563450e-01 1.24351516e-01 -6.08682215e-01
1.02575496e-01 -1.51641861e-01 -2.24629328e-01 -7.13712454e-01
5.06519139e-01 -3.72907519e-01 -6.66058138e-02 4.23564523e-01
9.03919756e-01 -2.22414002e-01 4.85958725e-01 5.42765141e-01
1.00601161e+00 4.04222570e-02 7.87185371e-01 2.93031275e-01
1.14203252e-01 -5.92069440e-02 4.59403396e-01 1.15660238e+00
-5.90620279e-01 6.53401673e-01 -9.90658626e-02 -4.89750952e-01
-1.14887214e+00 -1.10915983e+00 2.26224080e-01 1.36053741e+00
1.65756822e-01 -6.31885588e-01 -3.45976710e-01 -6.23807609e-01
-8.80471542e-02 9.96568918e-01 -7.63502896e-01 -6.11445785e-01
-1.43893108e-01 -1.01455212e-01 1.87261716e-01 6.76061869e-01
1.24299668e-01 -1.23862994e+00 -1.00829101e+00 2.43337497e-01
3.19563478e-01 -1.30014217e+00 -4.86766011e-01 3.96692634e-01
-5.53194106e-01 -9.60392594e-01 -9.23045576e-01 -1.13910687e+00
6.37980282e-01 8.81247222e-01 7.02605546e-01 -3.24528873e-01
-6.04504645e-01 7.65730619e-01 -6.49314582e-01 -4.88475084e-01
-2.88512409e-01 -4.01884705e-01 2.44378015e-01 2.15781834e-02
6.59578145e-01 -5.42494416e-01 -4.12621975e-01 -7.47883394e-02
-8.73033345e-01 -2.13410836e-02 7.02740729e-01 1.04080462e+00
6.09627306e-01 -4.29166108e-01 5.46936274e-01 -7.82495499e-01
5.13546467e-01 -7.56433547e-01 -3.17995459e-01 3.86598378e-01
-3.13106269e-01 -1.24134645e-01 7.97306478e-01 -7.10552871e-01
-8.94918919e-01 2.52622277e-01 2.10059449e-01 -1.27034020e+00
-4.23072636e-01 4.11474735e-01 2.02678606e-01 -3.26286145e-02
7.90780008e-01 5.15215576e-01 -1.63806587e-01 -2.27571428e-01
1.00010467e+00 8.44923794e-01 7.12370276e-01 -3.15749377e-01
6.59508526e-01 3.92701477e-01 -5.18488765e-01 -1.22219777e+00
-8.98054659e-01 -9.13326919e-01 -7.49503136e-01 -3.24529260e-01
7.07303882e-01 -1.08496332e+00 -2.27337986e-01 2.36478657e-01
-1.22227418e+00 -2.00757667e-01 -6.86051130e-01 5.23899674e-01
-9.87704396e-01 3.45384270e-01 -5.15612960e-01 -7.16263473e-01
-3.50164115e-01 -9.79811251e-01 9.79374170e-01 4.06959444e-01
5.79660349e-02 -8.13360810e-01 2.98503846e-01 -2.96396166e-01
1.93419725e-01 -2.72349894e-01 7.66648054e-01 -7.58087575e-01
-2.16059789e-01 -2.86235988e-01 -2.58369088e-01 2.06326112e-01
5.93557656e-02 -1.97951883e-01 -1.08852839e+00 -3.71380687e-01
1.09261917e-02 -9.84962404e-01 1.27527666e+00 3.00146490e-01
9.94227171e-01 -1.98114753e-01 -3.79484653e-01 6.53195441e-01
1.69023108e+00 2.44990557e-01 3.42527777e-01 -1.14514418e-01
7.60180771e-01 2.74708837e-01 6.86261892e-01 8.34532738e-01
2.62008101e-01 5.77204108e-01 2.57328421e-01 5.68274558e-01
-2.25713238e-01 -4.24192131e-01 4.32069361e-01 1.09879446e+00
4.41099316e-01 5.83193637e-02 -8.67080212e-01 8.65877092e-01
-2.14723825e+00 -1.30756009e+00 8.84754717e-01 1.91975796e+00
6.09417975e-01 1.55277222e-01 -6.79566413e-02 -5.99877238e-01
7.82022715e-01 4.95850444e-01 -9.13530171e-01 -3.21833611e-01
1.61566705e-01 7.46310130e-02 8.58434066e-02 1.24855004e-01
-1.10803068e+00 1.12909102e+00 6.31792450e+00 7.43975818e-01
-8.56476486e-01 2.87170976e-01 1.92111045e-01 -4.13990796e-01
1.30649298e-01 9.11862031e-02 -9.01504815e-01 3.04787874e-01
8.91295850e-01 -5.73484957e-01 3.86309385e-01 1.38641810e+00
-2.42489815e-01 1.62171364e-01 -1.26475072e+00 1.26479793e+00
7.31775522e-01 -1.68106484e+00 4.29824769e-01 -4.99275535e-01
9.52476799e-01 3.69629979e-01 -9.69697684e-02 7.78042316e-01
4.66961354e-01 -5.64504385e-01 6.97423637e-01 5.87580144e-01
7.30478108e-01 -5.19496500e-01 4.34240162e-01 4.74649787e-01
-1.26614726e+00 -4.72616494e-01 -1.04174602e+00 -7.28799477e-02
2.88912296e-01 3.10704559e-01 -1.04573083e+00 2.04107001e-01
4.26027566e-01 1.36198628e+00 -4.10761744e-01 1.26379192e+00
-1.91214055e-01 3.21012855e-01 9.77125466e-02 -9.92374271e-02
4.99532163e-01 -9.47926659e-03 5.02150178e-01 1.15496135e+00
2.72917956e-01 1.37247354e-01 4.38554347e-01 9.89013851e-01
-2.49679148e-01 1.82165280e-02 -1.21265948e+00 -1.81787521e-01
8.83614838e-01 1.27082884e+00 -5.93273938e-01 -9.16736603e-01
-8.72525930e-01 1.08213258e+00 7.22860932e-01 4.69515979e-01
-6.15212202e-01 -7.26080477e-01 4.79946613e-01 -5.20880103e-01
7.52560794e-01 -1.05408736e-01 2.15705842e-01 -1.22856295e+00
-2.79793173e-01 -4.68341708e-01 3.83837372e-01 -1.00076985e+00
-1.38599145e+00 4.40973461e-01 8.24290961e-02 -1.46110392e+00
-5.05173743e-01 -4.72805738e-01 -9.06355798e-01 2.78568268e-01
-1.38968205e+00 -1.23039258e+00 -4.57788825e-01 7.64110088e-01
1.29858363e+00 -4.89197701e-01 8.28421116e-01 -1.42321870e-01
-5.11570215e-01 4.41982388e-01 2.12205812e-01 1.23514086e-01
8.53819072e-01 -9.62334275e-01 6.49088740e-01 8.47458363e-01
5.81052303e-01 7.41630673e-01 5.91768324e-01 -5.15908837e-01
-1.43509340e+00 -1.35083282e+00 3.15283239e-01 3.44431661e-02
8.45749557e-01 -4.26859826e-01 -7.87668467e-01 7.76938081e-01
3.05321813e-01 4.37701881e-01 6.24452531e-01 -1.17332943e-01
-6.94332719e-01 3.64252813e-02 -7.55913258e-01 7.37563789e-01
9.89413083e-01 -8.45364094e-01 -1.09209526e+00 4.24098194e-01
1.03883374e+00 8.45720023e-02 -4.50818330e-01 1.19059030e-02
5.72534382e-01 -7.78864086e-01 9.36369181e-01 -6.32583439e-01
5.11966586e-01 6.87713772e-02 -6.61809385e-01 -1.57363057e+00
-4.76158559e-01 -1.71454847e-01 -5.22085965e-01 9.11443889e-01
-1.51530549e-01 -8.06441605e-02 5.29076576e-01 1.62601143e-01
-5.11564255e-01 -5.48863173e-01 -6.57728732e-01 -1.05072749e+00
-2.87678152e-01 -4.50321555e-01 2.23871917e-01 7.36418664e-01
3.19922805e-01 5.51561773e-01 -6.40805542e-01 -2.20856339e-01
9.70905006e-01 2.02641949e-01 8.48806381e-01 -9.81775463e-01
-4.56401259e-01 -2.79192626e-01 -7.28721559e-01 -8.55890095e-01
4.12159234e-01 -8.48930776e-01 6.51303172e-01 -1.40591037e+00
7.66140580e-01 2.11614639e-01 -4.48913097e-01 6.18020892e-01
-8.47939309e-03 3.14874858e-01 4.77974266e-01 3.41328710e-01
-1.16992402e+00 9.68999386e-01 7.13683248e-01 -5.82919776e-01
-9.29919630e-02 -4.45273012e-01 -4.27634120e-01 8.57614398e-01
4.52362984e-01 -3.55548382e-01 -6.81361318e-01 -2.36955747e-01
-3.70868325e-01 -1.44888043e-01 3.18772405e-01 -1.30706596e+00
6.26970828e-01 -2.59014219e-01 2.75993526e-01 -4.77160096e-01
6.30424380e-01 -5.78131855e-01 -4.25570130e-01 4.04213727e-01
-5.40340722e-01 -5.53419828e-01 -1.08858235e-01 9.82611537e-01
-2.04834521e-01 -5.44658661e-01 1.05422485e+00 -4.79092181e-01
-1.68980598e+00 6.01523519e-01 -1.61916718e-01 2.24956218e-02
1.30621147e+00 -2.45652676e-01 -4.88489777e-01 -3.88087422e-01
-7.65385926e-01 2.08115801e-01 5.97789466e-01 6.25440001e-01
1.15292501e+00 -1.48372304e+00 -2.76889533e-01 3.47278357e-01
1.06687438e+00 -4.04773861e-01 3.22445154e-01 6.20640337e-01
9.81385186e-02 3.14525306e-01 -5.54505706e-01 -5.37387669e-01
-1.02573466e+00 1.10816777e+00 1.13957882e-01 4.87064570e-01
-1.01172698e+00 7.85476506e-01 4.27624792e-01 -1.96931407e-01
5.61987042e-01 -8.13144725e-03 -1.40419111e-01 -3.79356854e-02
1.06457591e+00 1.63660303e-01 -4.12260860e-01 -5.32331645e-01
-2.00301215e-01 4.22392130e-01 -5.70015073e-01 -1.55145541e-01
1.40886271e+00 -3.51848036e-01 3.42017472e-01 1.17411125e+00
1.31849217e+00 -8.40945601e-01 -1.50492179e+00 -6.66411221e-01
-1.30346552e-01 -3.86046201e-01 -2.46716857e-01 -3.07302833e-01
-5.57422221e-01 1.21436977e+00 6.15234196e-01 -3.50085557e-01
6.24872446e-01 1.88435450e-01 7.85441518e-01 9.61921394e-01
5.87178290e-01 -1.30773044e+00 5.67327738e-01 7.23054111e-01
5.64892828e-01 -1.42565095e+00 -2.81802356e-01 1.64005324e-01
-7.85780311e-01 1.15942907e+00 7.61141717e-01 -3.13291818e-01
7.24342167e-01 -4.05135751e-03 -7.36303180e-02 -2.61337936e-01
-1.22501612e+00 -4.49750602e-01 1.40777245e-01 8.35588753e-01
-3.65196615e-02 -3.22887786e-02 1.10242672e-01 5.43825924e-01
3.73060226e-01 2.83716887e-01 6.78138793e-01 1.20780361e+00
-1.15142286e+00 -4.88704622e-01 1.21342190e-01 5.40532708e-01
4.06873673e-01 -8.69193748e-02 -2.04211384e-01 2.62571931e-01
-1.05597615e-01 7.38760352e-01 2.41296157e-01 -5.29063523e-01
6.61909282e-02 2.58114696e-01 4.89231765e-01 -1.22082663e+00
2.89079994e-01 -1.32384047e-01 -2.87823975e-01 -6.10774100e-01
-3.52065772e-01 -3.14709276e-01 -1.22923553e+00 2.54684746e-01
-6.64845109e-02 -1.59831978e-02 5.22462249e-01 1.16609824e+00
3.60369414e-01 2.96920717e-01 7.30774045e-01 -1.08959532e+00
-8.39958608e-01 -1.18847609e+00 -7.89470315e-01 5.08028269e-01
2.45101810e-01 -8.44554365e-01 -2.76685178e-01 3.03630948e-01] | [10.064921379089355, 2.4931466579437256] |
5e4cf8f6-4aa3-4115-8753-dfa975337b5c | cross-domain-video-anomaly-detection-without | 2212.07010 | null | https://arxiv.org/abs/2212.07010v1 | https://arxiv.org/pdf/2212.07010v1.pdf | Cross-Domain Video Anomaly Detection without Target Domain Adaptation | Most cross-domain unsupervised Video Anomaly Detection (VAD) works assume that at least few task-relevant target domain training data are available for adaptation from the source to the target domain. However, this requires laborious model-tuning by the end-user who may prefer to have a system that works ``out-of-the-box." To address such practical scenarios, we identify a novel target domain (inference-time) VAD task where no target domain training data are available. To this end, we propose a new `Zero-shot Cross-domain Video Anomaly Detection (zxvad)' framework that includes a future-frame prediction generative model setup. Different from prior future-frame prediction models, our model uses a novel Normalcy Classifier module to learn the features of normal event videos by learning how such features are different ``relatively" to features in pseudo-abnormal examples. A novel Untrained Convolutional Neural Network based Anomaly Synthesis module crafts these pseudo-abnormal examples by adding foreign objects in normal video frames with no extra training cost. With our novel relative normalcy feature learning strategy, zxvad generalizes and learns to distinguish between normal and abnormal frames in a new target domain without adaptation during inference. Through evaluations on common datasets, we show that zxvad outperforms the state-of-the-art (SOTA), regardless of whether task-relevant (i.e., VAD) source training data are available or not. Lastly, zxvad also beats the SOTA methods in inference-time efficiency metrics including the model size, total parameters, GPU energy consumption, and GMACs. | ['Amit K. Roy-Chowdhury', 'Kuan-Chuan Peng', 'Abhishek Aich'] | 2022-12-14 | null | null | null | null | ['video-anomaly-detection'] | ['computer-vision'] | [ 2.38620758e-01 -2.67526031e-01 -6.51755705e-02 -4.41172540e-01
-7.14003682e-01 -3.70676368e-01 5.24920046e-01 -1.37237579e-01
-1.84321523e-01 4.19250667e-01 -2.14707062e-01 -2.51739323e-01
2.94403732e-01 -7.53752410e-01 -1.10368860e+00 -6.15796626e-01
-2.32042134e-01 4.44402158e-01 4.40668195e-01 -7.63087645e-02
6.62022308e-02 3.19623470e-01 -1.68429458e+00 5.05448163e-01
8.42585623e-01 1.25580585e+00 -1.53549299e-01 8.79521489e-01
-9.23212841e-02 9.05668616e-01 -6.67082667e-01 -2.89172083e-01
5.21669030e-01 -6.22219205e-01 -4.67611164e-01 4.38366383e-01
7.50112832e-01 -7.54576921e-01 -6.04691207e-01 9.11242366e-01
4.35778201e-01 1.84170276e-01 5.65636694e-01 -1.68200552e+00
-4.84098703e-01 -6.69992268e-02 -5.71115792e-01 8.73804212e-01
2.71987170e-01 7.05774307e-01 5.64888775e-01 -9.72215176e-01
6.88439071e-01 1.03419542e+00 6.54319048e-01 9.20780897e-01
-1.10433519e+00 -6.87407613e-01 5.31808496e-01 4.71122652e-01
-1.10602498e+00 -4.43935156e-01 8.98188710e-01 -5.07062256e-01
1.20965910e+00 2.03552544e-01 6.17169619e-01 1.72617722e+00
1.94712326e-01 1.06422174e+00 4.23339069e-01 -1.14721045e-01
5.63844562e-01 -4.36128616e-01 -1.42612413e-01 8.41125786e-01
1.95363298e-01 3.58641185e-02 -7.84567356e-01 -3.04934829e-01
6.32477105e-01 3.31557803e-02 -2.04312965e-01 -6.40952706e-01
-1.13349473e+00 6.23396993e-01 -2.68438607e-01 5.41652367e-02
-4.08183187e-01 2.12838784e-01 1.04127407e+00 6.51123703e-01
7.10723042e-01 3.00607055e-01 -7.40881443e-01 -5.32160401e-01
-8.48045528e-01 3.09708267e-01 4.96160746e-01 9.92785871e-01
5.81465483e-01 6.63820446e-01 -3.40705454e-01 5.08909941e-01
-2.57499159e-05 4.29853171e-01 7.72161782e-01 -1.11958039e+00
4.49220210e-01 4.82385993e-01 -1.48014814e-01 -7.81333029e-01
-1.87437326e-01 -3.42599005e-01 -8.14156651e-01 4.11785483e-01
5.73350608e-01 -2.74116337e-01 -1.12867582e+00 1.80273020e+00
4.28136200e-01 9.43498313e-01 2.81387269e-01 7.82370448e-01
3.79234642e-01 7.33127713e-01 -1.08070999e-01 -3.10668916e-01
1.03661239e+00 -9.46666002e-01 -4.99327153e-01 -3.84897679e-01
9.53576863e-01 -5.37319243e-01 1.25100529e+00 6.18870616e-01
-1.03004050e+00 -6.13661468e-01 -9.74357426e-01 1.24053083e-01
-2.17162177e-01 4.69808504e-02 4.15740877e-01 5.44413090e-01
-8.80429208e-01 3.89265865e-01 -1.23277414e+00 -4.41791087e-01
7.13385046e-01 1.42391726e-01 -2.94213653e-01 -1.51954904e-01
-8.69355321e-01 5.41920424e-01 3.99986953e-01 -2.70653486e-01
-1.42895532e+00 -1.06829977e+00 -1.00990784e+00 -9.97112840e-02
6.21753573e-01 -8.73023629e-01 1.34563518e+00 -1.51300454e+00
-1.30486751e+00 8.19194794e-01 -2.37016171e-01 -7.45161414e-01
7.22764552e-01 -4.98882115e-01 -7.18051016e-01 2.44270965e-01
1.17251262e-01 4.49225932e-01 1.37528634e+00 -8.57910514e-01
-9.34849441e-01 -2.08642185e-01 -8.47033188e-02 1.52752697e-01
-1.02446340e-01 -1.89611852e-01 -6.02096915e-01 -1.11363459e+00
-1.71314999e-01 -7.43431091e-01 8.31861719e-02 4.36460495e-01
-2.27634832e-01 -2.70996034e-01 1.52379191e+00 -6.23106897e-01
1.25627494e+00 -2.51184487e+00 -9.69540700e-02 2.69841254e-01
2.30932325e-01 4.21170920e-01 -5.61942607e-02 -2.49772593e-01
-2.94166625e-01 -3.09079319e-01 -2.57803142e-01 -2.51495868e-01
-1.21767670e-01 5.54824829e-01 -4.42268461e-01 4.00051028e-01
4.87021118e-01 4.66074228e-01 -1.13767147e+00 -3.87817472e-01
2.15525493e-01 7.88836032e-02 -9.37293828e-01 5.33531010e-01
-4.62282002e-01 4.94660527e-01 -3.85762155e-01 8.89588654e-01
5.19059420e-01 -1.56786889e-01 -9.26757082e-02 -2.69329511e-02
3.65659297e-01 -7.64480457e-02 -9.71355438e-01 2.05324769e+00
-2.47588426e-01 6.19384050e-01 -3.19369316e-01 -1.33368623e+00
6.79931641e-01 3.35240632e-01 5.34927666e-01 -5.68162024e-01
-1.23054482e-01 2.46921092e-01 -2.99543235e-02 -6.87882721e-01
1.38149306e-01 3.04516703e-01 2.80282777e-02 3.14918697e-01
3.91904324e-01 5.15529454e-01 1.88862875e-01 2.71282434e-01
1.57014954e+00 4.42727327e-01 3.42733443e-01 2.52659321e-02
5.75157285e-01 7.64782876e-02 9.44006681e-01 8.68403375e-01
-3.98253202e-01 6.43552482e-01 7.97811627e-01 -7.54297376e-01
-1.24336290e+00 -1.38047886e+00 1.61342248e-01 1.27409875e+00
-2.75400449e-02 -3.80380690e-01 -7.64323533e-01 -1.22470355e+00
-1.72585696e-01 8.84804428e-01 -6.94153011e-01 -5.50987899e-01
-7.55426466e-01 -5.75583935e-01 7.06899822e-01 6.47628963e-01
5.90686321e-01 -8.63806486e-01 -7.17081726e-01 1.93057448e-01
-2.55666316e-01 -1.17597270e+00 -5.75002372e-01 -5.99317439e-02
-9.07203496e-01 -1.08863664e+00 -4.06663805e-01 -4.42404151e-01
6.56215847e-01 -2.99393106e-02 1.31125653e+00 -5.57286292e-02
-2.94589847e-01 7.54356861e-01 -4.58143145e-01 -3.03083956e-01
-4.94711161e-01 -2.15221226e-01 3.77605200e-01 3.16869944e-01
7.49056518e-01 -6.70269847e-01 -5.75219929e-01 1.68755844e-01
-9.37030077e-01 -1.17541812e-01 4.58366036e-01 9.01286185e-01
8.11200440e-01 4.38448915e-05 4.53648895e-01 -7.30105460e-01
1.42147571e-01 -8.54336798e-01 -6.04763269e-01 -5.73751843e-03
-3.98489445e-01 1.86650343e-02 8.49522650e-01 -6.69726193e-01
-1.10047090e+00 -1.31680584e-02 -1.34300753e-01 -1.30017257e+00
-4.97996569e-01 -3.39292884e-02 -3.33768696e-01 3.54791820e-01
8.33774924e-01 4.98876065e-01 1.00784026e-01 -3.10588658e-01
-7.55055547e-02 1.90808326e-01 9.61691737e-01 -7.35120296e-01
8.66368592e-01 5.34242809e-01 -2.98545305e-02 -7.03874946e-01
-7.17602611e-01 -3.30443531e-01 -3.77713263e-01 -1.83886021e-01
9.42678332e-01 -1.10965943e+00 -2.64263153e-01 7.55034566e-01
-9.53676403e-01 -4.84882444e-01 -6.88372493e-01 3.75873864e-01
-8.38192344e-01 5.33599675e-01 -2.59005785e-01 -5.24711490e-01
-2.69745976e-01 -1.17797470e+00 1.01661921e+00 -2.06995517e-01
-3.10211986e-01 -9.54928100e-01 -2.07978338e-02 -2.46316530e-02
1.81981966e-01 3.05299520e-01 9.24746215e-01 -1.21968007e+00
-5.24551213e-01 -1.42181844e-01 4.02092263e-02 6.02890909e-01
8.31087679e-02 -7.73081407e-02 -9.00481641e-01 -3.50812107e-01
6.05038814e-02 -8.26686919e-02 6.15823925e-01 3.75908524e-01
1.86917257e+00 -4.99931246e-01 -1.89779729e-01 9.45885777e-01
1.08670425e+00 2.93316662e-01 7.48575091e-01 3.18345666e-01
6.95716083e-01 -1.54811725e-01 8.61690521e-01 6.91723168e-01
2.55373865e-01 6.72126889e-01 4.89525855e-01 3.91034037e-02
-7.47152120e-02 -6.58766553e-02 9.08605278e-01 2.38315061e-01
1.84896272e-02 -4.34101254e-01 -7.79633462e-01 6.62906349e-01
-1.89333034e+00 -1.20626056e+00 2.33889282e-01 2.15179062e+00
5.38664043e-01 2.03741819e-01 2.94029564e-01 1.67275205e-01
5.56579411e-01 -2.64787637e-02 -9.53004301e-01 -4.60240453e-01
-2.15666443e-02 2.36695841e-01 4.33733650e-02 8.44084844e-02
-1.41518998e+00 7.37462282e-01 4.93738031e+00 9.53164756e-01
-1.06751513e+00 2.16547281e-01 7.65139461e-01 -4.19389099e-01
1.60551637e-01 -2.53786325e-01 -5.41464925e-01 8.74656498e-01
1.00757575e+00 2.75957976e-02 2.70356089e-01 1.13179183e+00
5.67285791e-02 6.22914620e-02 -1.49045312e+00 1.12654114e+00
2.65968233e-01 -1.16866434e+00 2.67485470e-01 -2.63624549e-01
4.75038618e-01 2.17193700e-02 2.01193634e-02 6.05163634e-01
7.07640871e-02 -4.38336730e-01 4.22547102e-01 5.03630400e-01
8.62140536e-01 -6.71473980e-01 5.85656583e-01 1.41965836e-01
-9.59185123e-01 -2.28402823e-01 -2.47353390e-01 1.54913351e-01
-4.51997016e-03 5.33533633e-01 -8.35471094e-01 4.09040093e-01
1.05294931e+00 8.18495810e-01 -6.57210052e-01 7.81972349e-01
1.88212305e-01 8.72104108e-01 -2.72820294e-01 7.94429243e-01
1.96537048e-01 1.28194332e-01 1.00827539e+00 1.10532343e+00
8.13395798e-01 4.19755792e-03 2.40108445e-01 5.67138076e-01
1.08914658e-01 -2.67030120e-01 -7.87076771e-01 3.21772695e-01
2.28504822e-01 7.80455649e-01 -3.28099221e-01 -6.79800570e-01
-6.74480379e-01 1.57420862e+00 -1.08266987e-01 4.28120434e-01
-1.15617502e+00 -4.05763723e-02 1.27207983e+00 1.33756429e-01
4.34496880e-01 1.59715816e-01 -4.80946004e-02 -1.54468775e+00
1.74084932e-01 -1.05882466e+00 8.98122132e-01 -7.10887194e-01
-1.64036846e+00 4.29427832e-01 -2.45067962e-02 -1.76899505e+00
-6.90908670e-01 -5.81337214e-01 -6.88629687e-01 3.69625241e-01
-1.27471328e+00 -1.03709888e+00 -4.61588711e-01 1.16075206e+00
9.87225771e-01 -6.82418704e-01 7.05804825e-01 3.49863738e-01
-6.67142928e-01 1.00655532e+00 -5.26655838e-02 4.11056966e-01
8.23261082e-01 -1.24849081e+00 7.38299966e-01 1.37030625e+00
-2.31056377e-01 1.67602077e-02 6.55307412e-01 -6.55648828e-01
-1.27994525e+00 -1.66538346e+00 3.25146735e-01 -5.34040749e-01
5.84801376e-01 -3.25996786e-01 -1.31943440e+00 1.05693269e+00
-8.60504061e-02 7.69196749e-01 6.18317246e-01 -2.58462101e-01
-4.28510845e-01 -2.52379447e-01 -1.19162226e+00 6.53023124e-01
1.43976462e+00 -3.50749433e-01 -5.99532604e-01 3.50407541e-01
6.59545481e-01 -7.68023789e-01 -5.50423861e-01 5.15362442e-01
2.36212134e-01 -1.02347207e+00 1.16815960e+00 -1.00482905e+00
3.94203007e-01 -4.53249037e-01 -3.55060309e-01 -1.18654668e+00
-4.81826514e-02 -5.79015911e-01 -1.11927927e+00 9.27396119e-01
8.37647468e-02 -6.17471874e-01 8.63274813e-01 4.62520152e-01
-6.60245061e-01 -7.31500506e-01 -1.20329034e+00 -1.06144953e+00
-2.42597580e-01 -8.21424186e-01 5.61333001e-01 1.05262673e+00
-3.42597872e-01 2.68808976e-02 -4.67988551e-01 5.35445511e-01
4.61702526e-01 -1.74206257e-01 9.87350583e-01 -9.73228991e-01
-4.32161748e-01 -2.52222389e-01 -9.62673724e-01 -9.68313098e-01
1.68917045e-01 -6.63820446e-01 -2.26567343e-01 -7.25124896e-01
-4.57623303e-02 -1.76859930e-01 -3.55081111e-01 5.92333198e-01
-2.41558552e-01 1.63442194e-01 -1.38017580e-01 -6.36229962e-02
-8.34642768e-01 4.65536296e-01 9.03953373e-01 8.72371439e-03
-2.96898335e-01 1.21284025e-02 -2.81896859e-01 9.43859279e-01
7.58916259e-01 -2.47127265e-01 -5.96667349e-01 -5.39893866e-01
-8.31397846e-02 -1.31674811e-01 6.20122015e-01 -1.39868569e+00
-3.95753272e-02 -1.61295205e-01 7.08836854e-01 -5.24256587e-01
3.18807453e-01 -9.05725002e-01 -1.23464189e-01 2.67664075e-01
1.03225140e-02 3.84594321e-01 3.43053669e-01 8.54427159e-01
-9.57007613e-03 1.59285124e-02 7.42687166e-01 1.32735372e-02
-1.26017916e+00 6.05635762e-01 -6.28212690e-01 3.89419496e-01
1.42752373e+00 -5.30613840e-01 -3.40562224e-01 -4.36161667e-01
-8.52584302e-01 1.11424051e-01 5.62332273e-01 5.26485860e-01
7.69630969e-01 -1.36293435e+00 -5.91288507e-01 6.32725954e-01
4.06216770e-01 2.98696369e-01 6.38202548e-01 8.71049702e-01
-4.53519762e-01 -4.01138254e-02 -2.10163087e-01 -9.89430189e-01
-1.11307693e+00 8.42048466e-01 4.19454843e-01 -2.21883133e-01
-9.36625898e-01 6.26057804e-01 4.81478930e-01 -1.53106093e-01
1.51288956e-01 -2.44318143e-01 3.41150552e-01 -1.95920065e-01
5.52133560e-01 5.05018592e-01 1.64030463e-01 -4.44398344e-01
-1.75960451e-01 3.94293703e-02 -2.37414673e-01 3.10373157e-01
1.02607119e+00 -5.78085706e-02 2.93621749e-01 4.11431819e-01
1.04131365e+00 -2.87303358e-01 -1.67637229e+00 -3.58508110e-01
-1.78350449e-01 -7.80569911e-01 -1.64185598e-01 -6.37944520e-01
-1.26910079e+00 7.93780446e-01 1.07077765e+00 -2.84890920e-01
1.66420305e+00 -7.62861520e-02 9.61954951e-01 4.79375571e-01
-3.57662491e-03 -1.13929260e+00 5.73328197e-01 4.90350634e-01
5.39941728e-01 -1.28374350e+00 -2.27166086e-01 -4.65477481e-02
-8.09466600e-01 9.67334747e-01 1.33804524e+00 -1.28630042e-01
4.92360741e-01 -6.49301410e-02 -9.51886922e-03 -1.43347427e-01
-9.29912388e-01 1.37819335e-01 2.20771924e-01 8.69231820e-01
-9.46650282e-02 -3.31739932e-01 3.70568693e-01 5.99174976e-01
7.53877535e-02 -2.72553302e-02 5.55908024e-01 9.41687703e-01
-1.63107932e-01 -8.82822692e-01 -3.30178112e-01 8.59479427e-01
-5.84592640e-01 3.02006938e-02 1.90754235e-01 7.93090165e-01
4.36033100e-01 4.54839349e-01 7.14178801e-01 -2.83990324e-01
2.03348473e-01 4.79668707e-01 2.45631725e-01 -4.79935408e-01
-9.57671702e-02 -6.83769584e-02 -8.35995451e-02 -9.40859497e-01
-1.35214061e-01 -8.11331987e-01 -1.12708449e+00 -3.84142548e-01
1.07020527e-01 -2.65118092e-01 5.36635108e-02 1.11264014e+00
7.56732404e-01 6.70805752e-01 4.79122698e-01 -6.94937110e-01
-2.01469645e-01 -6.45433426e-01 -3.28090757e-01 6.85273111e-01
6.46114051e-01 -7.54953563e-01 -2.15152383e-01 2.86505789e-01] | [7.848208904266357, 1.601791501045227] |
050e97e9-c5f2-4e38-8e1c-fd95f662e9c2 | ghost-in-the-minecraft-generally-capable | 2305.17144 | null | https://arxiv.org/abs/2305.17144v2 | https://arxiv.org/pdf/2305.17144v2.pdf | Ghost in the Minecraft: Generally Capable Agents for Open-World Environments via Large Language Models with Text-based Knowledge and Memory | The captivating realm of Minecraft has attracted substantial research interest in recent years, serving as a rich platform for developing intelligent agents capable of functioning in open-world environments. However, the current research landscape predominantly focuses on specific objectives, such as the popular "ObtainDiamond" task, and has not yet shown effective generalization to a broader spectrum of tasks. Furthermore, the current leading success rate for the "ObtainDiamond" task stands at around 20%, highlighting the limitations of Reinforcement Learning (RL) based controllers used in existing methods. To tackle these challenges, we introduce Ghost in the Minecraft (GITM), a novel framework integrates Large Language Models (LLMs) with text-based knowledge and memory, aiming to create Generally Capable Agents (GCAs) in Minecraft. These agents, equipped with the logic and common sense capabilities of LLMs, can skillfully navigate complex, sparse-reward environments with text-based interactions. We develop a set of structured actions and leverage LLMs to generate action plans for the agents to execute. The resulting LLM-based agent markedly surpasses previous methods, achieving a remarkable improvement of +47.5% in success rate on the "ObtainDiamond" task, demonstrating superior robustness compared to traditional RL-based controllers. Notably, our agent is the first to procure all items in the Minecraft Overworld technology tree, demonstrating its extensive capabilities. GITM does not need any GPU for training, but a single CPU node with 32 CPU cores is enough. This research shows the potential of LLMs in developing capable agents for handling long-horizon, complex tasks and adapting to uncertainties in open-world environments. See the project website at https://github.com/OpenGVLab/GITM. | ['Jifeng Dai', 'Zhaoxiang Zhang', 'Yu Qiao', 'Xiaogang Wang', 'Lewei Lu', 'Bin Li', 'Gao Huang', 'Chenyu Yang', 'Weijie Su', 'Chenxin Tao', 'Hao Tian', 'Yuntao Chen', 'Xizhou Zhu'] | 2023-05-25 | null | null | null | null | ['navigate', 'common-sense-reasoning'] | ['reasoning', 'reasoning'] | [-1.41815901e-01 -7.26986825e-02 -1.52860463e-01 1.87456533e-01
-4.67262834e-01 -6.92139328e-01 6.23436689e-01 -3.51382822e-01
-5.97224593e-01 9.35038745e-01 -9.19347107e-02 -2.06776410e-01
-1.54309005e-01 -7.79084623e-01 -7.60583162e-01 -6.66681588e-01
-3.87611777e-01 6.96993053e-01 1.18352979e-01 -8.13095987e-01
1.62644207e-01 1.41232461e-01 -1.82712817e+00 1.88791212e-02
8.95059288e-01 7.61712134e-01 7.78682649e-01 5.52583337e-01
3.43624562e-01 8.16304445e-01 -6.51990891e-01 5.49188405e-02
3.82344812e-01 7.75932223e-02 -7.48744130e-01 -2.63269991e-01
-7.52017498e-02 -3.94000947e-01 -2.35530026e-02 7.47453690e-01
6.18164599e-01 4.06876326e-01 7.38860890e-02 -1.46903610e+00
-2.38458097e-01 7.97695100e-01 -3.09851617e-01 -1.45186067e-01
3.90794158e-01 8.41233552e-01 1.02965546e+00 -3.76534939e-01
6.53306365e-01 1.37687361e+00 5.35314798e-01 7.39694357e-01
-9.76649582e-01 -6.64209664e-01 3.57554227e-01 4.53369580e-02
-1.10931933e+00 -2.19894096e-01 1.41300142e-01 -2.80101091e-01
1.56566846e+00 -3.91344214e-03 9.95646060e-01 1.56452417e+00
5.15899420e-01 9.57057476e-01 1.25775254e+00 -3.62733789e-02
5.93028963e-01 -2.08353475e-01 -5.15014768e-01 7.53702164e-01
2.39624321e-01 6.73668683e-01 -5.99544227e-01 -5.43196090e-02
8.74151409e-01 -1.36594564e-01 -2.42300965e-02 -3.94953072e-01
-1.54764950e+00 8.10935318e-01 4.45931375e-01 2.21186608e-01
-5.37881553e-01 5.42674899e-01 3.79125863e-01 3.77575070e-01
1.69354118e-02 1.14012468e+00 -6.66996300e-01 -5.63929558e-01
-3.50543201e-01 7.45036006e-01 9.96540010e-01 8.45607579e-01
6.02632999e-01 4.98450667e-01 -5.72435074e-02 4.80896086e-01
2.49968514e-01 8.18452597e-01 5.13608396e-01 -1.41153729e+00
4.66210753e-01 5.38302541e-01 4.23541218e-01 -8.16306472e-01
-8.93339038e-01 -6.89128876e-01 -5.39826632e-01 6.58087373e-01
2.50734210e-01 -6.66888416e-01 -5.95419168e-01 1.91282856e+00
4.59635437e-01 -3.31086032e-02 5.06029427e-01 1.05681896e+00
2.21345648e-01 7.61259019e-01 -1.64420214e-02 4.64000972e-03
1.19151092e+00 -1.14242494e+00 -4.46324319e-01 -6.33325815e-01
7.09532022e-01 -3.97101164e-01 1.27657545e+00 6.59722328e-01
-1.00696123e+00 -2.01549411e-01 -1.00875831e+00 5.20086586e-01
-4.65350717e-01 -1.77883819e-01 1.14544213e+00 4.22304958e-01
-1.17781591e+00 5.77327549e-01 -8.72020960e-01 -3.77002835e-01
1.80775926e-01 6.50368094e-01 -1.65222526e-01 1.70879200e-01
-1.33828557e+00 1.14239049e+00 6.20692432e-01 -3.45546044e-02
-1.56821072e+00 -6.09398603e-01 -7.77838290e-01 -1.36623755e-01
1.02907586e+00 -9.12733614e-01 1.60610282e+00 -8.32299590e-01
-2.10477042e+00 2.91719377e-01 4.98418003e-01 -6.33093178e-01
4.79844332e-01 -4.22668457e-01 -1.00831494e-01 -1.64099649e-01
1.06504761e-01 8.11764538e-01 9.64858770e-01 -1.12706351e+00
-7.51696467e-01 -8.30841511e-02 5.50718606e-01 6.09957218e-01
-1.28419295e-01 -1.91206262e-01 -7.94109404e-02 -5.72605669e-01
-5.41820943e-01 -1.28746009e+00 -5.86869597e-01 -5.64809918e-01
-9.34269950e-02 -2.34626025e-01 6.77550077e-01 -3.17977667e-02
8.23221207e-01 -1.74402976e+00 4.37766701e-01 -2.14241929e-02
1.88043818e-01 4.27432656e-01 -5.16242921e-01 7.14840293e-01
5.33085644e-01 -2.34996341e-02 -7.12881312e-02 -1.13928333e-01
4.39339757e-01 5.74004054e-01 -4.28857267e-01 -6.01329282e-03
1.06966823e-01 9.58531082e-01 -1.27297485e+00 -1.77660249e-02
2.32683957e-01 1.03929982e-01 -7.98317075e-01 1.93008780e-01
-1.03020847e+00 5.03942847e-01 -7.92749822e-01 5.17459869e-01
1.97774634e-01 -1.43316597e-01 2.92112321e-01 4.78333801e-01
-4.35539603e-01 1.02118310e-02 -1.11083305e+00 1.97391653e+00
-7.22211361e-01 2.15972543e-01 4.42650795e-01 -5.68833172e-01
7.85578489e-01 1.41523033e-01 4.26155835e-01 -9.11072314e-01
3.03583294e-01 3.35451901e-01 1.45497948e-01 -3.88098031e-01
6.45137906e-01 8.10612515e-02 -3.67318988e-01 5.05861104e-01
-1.24622628e-01 -6.27534986e-01 5.76221049e-01 4.38661501e-02
1.35316408e+00 5.79016328e-01 2.94613484e-02 -2.17654735e-01
1.71101868e-01 3.80648553e-01 6.13247514e-01 9.87572312e-01
-1.29973456e-01 -8.76546651e-02 4.27473843e-01 -4.94901627e-01
-6.89869702e-01 -8.74714971e-01 3.27650309e-01 1.46198511e+00
1.74157739e-01 -4.91600394e-01 -7.80610979e-01 -4.01909828e-01
1.05054490e-01 8.57733369e-01 -4.41288859e-01 -1.71194732e-01
-5.55221319e-01 -7.49972939e-01 7.49130785e-01 3.40808362e-01
7.50310838e-01 -1.57329226e+00 -1.38799310e+00 3.78884792e-01
-1.24053478e-01 -1.01404107e+00 -1.92748427e-01 2.60784715e-01
-5.98229349e-01 -1.03011966e+00 -3.10376495e-01 -6.07252479e-01
1.40833452e-01 1.32716507e-01 1.06501734e+00 1.56651344e-02
-2.14059129e-01 5.66132605e-01 -4.41076487e-01 -3.65980268e-01
-3.54662150e-01 2.36310154e-01 4.45413351e-01 -4.65737969e-01
-2.78953522e-01 -5.02785742e-01 -3.59186321e-01 4.02505815e-01
-8.02761793e-01 2.24792928e-01 8.22399199e-01 9.67534006e-01
1.74720362e-01 1.08428806e-01 7.22964883e-01 -3.83601338e-01
8.28522682e-01 -4.94579196e-01 -1.00826561e+00 -6.12088926e-02
-5.95585585e-01 1.07221335e-01 8.83141994e-01 -4.83347684e-01
-1.09616470e+00 -9.57800969e-02 4.02282849e-02 -2.25590035e-01
-1.18478112e-01 4.69993949e-01 1.65646836e-01 -1.99800357e-02
6.43973351e-01 1.91376284e-01 3.35513771e-01 -5.27472682e-02
3.96355718e-01 2.96267718e-01 2.33860001e-01 -1.00658214e+00
6.57902658e-01 2.55477339e-01 -9.08086449e-02 -5.03582537e-01
-5.67177117e-01 -5.73044550e-03 -7.28321355e-03 -7.47264400e-02
6.56530559e-01 -9.36550975e-01 -1.41677868e+00 6.16099179e-01
-8.34227324e-01 -1.16557229e+00 -4.11133587e-01 3.83213788e-01
-1.08670628e+00 -1.82203978e-01 -5.95582843e-01 -8.32796276e-01
-3.49766791e-01 -1.40965152e+00 9.09977019e-01 3.96705151e-01
-2.33627781e-01 -8.83399546e-01 2.43912160e-01 4.75085258e-01
8.69989634e-01 2.72395015e-01 7.40571856e-01 -4.53165233e-01
-7.27643013e-01 2.09446400e-01 2.84265071e-01 -4.22462970e-02
-7.60349333e-02 -2.10160017e-01 -5.49834788e-01 -5.49309194e-01
-3.08906198e-01 -8.89346361e-01 4.62291181e-01 1.39379770e-01
5.71275532e-01 -3.52394968e-01 -3.09149712e-01 4.00729179e-01
1.24245906e+00 2.60991156e-01 2.73683131e-01 8.81467164e-01
3.37729782e-01 4.58018124e-01 1.02041960e+00 9.08585548e-01
6.81595802e-01 7.00217545e-01 1.02296162e+00 2.20189735e-01
1.69453338e-01 -2.69455492e-01 9.90969598e-01 4.44780260e-01
-1.10422164e-01 -3.89486134e-01 -9.55740213e-01 4.19451565e-01
-2.23718500e+00 -9.09313023e-01 3.04077923e-01 1.97030663e+00
8.22909772e-01 1.03003956e-01 9.71412733e-02 -4.12179798e-01
3.20797026e-01 2.49889955e-01 -9.74151254e-01 -4.95116770e-01
1.16795115e-02 1.07744865e-01 3.42493504e-01 3.57048154e-01
-9.63623703e-01 1.53002727e+00 6.17436504e+00 9.13582563e-01
-1.04782462e+00 -1.63315743e-01 1.01748437e-01 -2.89169282e-01
3.88054177e-03 4.09763902e-02 -7.01120377e-01 2.06300393e-01
9.25426960e-01 -1.57904059e-01 9.94390428e-01 9.90183830e-01
5.42587042e-01 -3.62870663e-01 -8.62153053e-01 6.72750115e-01
-2.04555899e-01 -1.22700310e+00 -1.32724280e-02 -2.32661087e-02
8.54650617e-01 4.34481621e-01 3.74889761e-01 8.04980516e-01
1.13025475e+00 -1.22786295e+00 8.47731888e-01 3.15807432e-01
3.74778599e-01 -7.52875268e-01 6.94793642e-01 6.39048576e-01
-1.05883384e+00 -4.65592474e-01 -2.54613608e-01 -5.86284697e-01
7.49741197e-02 -1.10032015e-01 -9.72660363e-01 4.94545668e-01
8.59883308e-01 6.83478177e-01 -1.45507962e-01 7.32104540e-01
-2.53490597e-01 3.16467077e-01 -3.80124658e-01 -3.89398545e-01
6.39005482e-01 -2.32903525e-01 7.91468501e-01 7.38130331e-01
2.42725551e-01 -4.53376286e-02 6.19996905e-01 8.73608887e-01
1.83809802e-01 -1.15095675e-01 -6.76654637e-01 -3.38643938e-01
3.22237164e-01 1.33602118e+00 -4.55846399e-01 -1.35984197e-01
-6.05143085e-02 6.72114849e-01 4.00788844e-01 2.50339866e-01
-9.84257340e-01 -1.67003423e-01 9.63379741e-01 -3.23409438e-01
1.51983842e-01 -4.92558867e-01 -7.21596405e-02 -1.00652480e+00
-2.33321130e-01 -1.64460528e+00 2.63592869e-01 -8.27335417e-01
-9.49366033e-01 5.17375946e-01 -1.29357055e-01 -1.02656317e+00
-4.49502110e-01 -6.19499803e-01 -3.79343420e-01 3.84695917e-01
-1.40899992e+00 -1.06891274e+00 -2.63131648e-01 7.51767039e-01
7.15442777e-01 -4.68075395e-01 1.01535892e+00 -2.42768049e-01
-7.41124094e-01 1.90010339e-01 1.37516588e-01 -4.38589334e-01
6.02833092e-01 -1.25816751e+00 4.04061943e-01 4.73422706e-01
-3.17024708e-01 6.28039360e-01 8.69701505e-01 -7.00363696e-01
-2.18561220e+00 -1.01361644e+00 8.04781541e-02 -5.05490124e-01
9.71673489e-01 -2.68561035e-01 -2.68621504e-01 6.69219315e-01
3.33571434e-01 -4.57362890e-01 2.23330721e-01 2.40610559e-02
-1.87108055e-01 -2.91410536e-02 -1.00661850e+00 1.05422902e+00
1.18152189e+00 1.16906375e-01 -4.87757534e-01 5.09770811e-01
9.58118916e-01 -6.74482048e-01 -8.10932696e-01 2.96559691e-01
4.50391263e-01 -9.30419385e-01 8.34725022e-01 -4.76679444e-01
2.77057260e-01 -3.98356348e-01 -1.74084604e-01 -1.74031794e+00
-2.12285310e-01 -1.08750296e+00 -3.19624245e-02 7.77073085e-01
4.00963783e-01 -1.12700236e+00 5.89017093e-01 4.19388533e-01
-3.13975275e-01 -7.40237832e-01 -7.16155946e-01 -9.69698846e-01
8.53984430e-02 -4.00406182e-01 5.91068089e-01 6.58861220e-01
2.42695838e-01 2.00866997e-01 -5.11058748e-01 1.08158462e-01
3.91023070e-01 1.50626004e-01 9.94565785e-01 -8.32998872e-01
-6.97354734e-01 -5.36475956e-01 1.91755578e-01 -9.04918611e-01
2.49738872e-01 -8.37785602e-01 2.88284779e-01 -1.47002363e+00
-1.23946562e-01 -8.01808715e-01 -3.84231471e-02 8.69985044e-01
2.38448381e-01 -1.61923960e-01 6.17859423e-01 8.14207867e-02
-9.96771216e-01 8.75717819e-01 1.53883326e+00 -2.00781316e-01
-3.45964700e-01 -1.54721454e-01 -6.85401440e-01 8.46009374e-01
1.12814462e+00 -1.41221255e-01 -5.51728070e-01 -5.70751011e-01
5.10809004e-01 1.67021349e-01 3.04191321e-01 -1.21115291e+00
1.45138532e-01 -6.32396102e-01 -8.35569724e-02 -3.31098512e-02
4.67534542e-01 -7.01099217e-01 1.17488347e-01 8.72492492e-01
-1.95375398e-01 3.54712754e-01 5.48170447e-01 4.55880791e-01
1.47004262e-01 4.01910171e-02 3.87625158e-01 -4.08916980e-01
-1.02749872e+00 1.32585123e-01 -8.73324990e-01 2.86896884e-01
1.30229568e+00 6.21480793e-02 -6.55112147e-01 -2.57468909e-01
-4.69866097e-01 8.82465720e-01 5.54903090e-01 6.83386981e-01
5.29115617e-01 -7.84934580e-01 -5.06180346e-01 3.29284333e-02
-1.71320856e-01 3.75625715e-02 1.84175745e-01 7.97256112e-01
-4.91038352e-01 3.89409423e-01 -5.87477744e-01 -3.88575494e-01
-7.64692485e-01 5.91631711e-01 4.01131153e-01 -6.07054174e-01
-6.22597098e-01 6.14208579e-01 1.61940172e-01 -7.91135490e-01
1.39343560e-01 -3.34460407e-01 -1.10463530e-01 -1.77793637e-01
3.26810449e-01 4.91412908e-01 -2.38113314e-01 -1.55951902e-01
-4.39591944e-01 3.68595719e-01 1.94700304e-02 -2.02975497e-01
1.44551456e+00 1.79805219e-01 -1.87391713e-02 1.96614698e-01
3.06461990e-01 -2.29839042e-01 -1.58549690e+00 1.66829497e-01
-5.43951541e-02 -1.53039292e-01 -5.50797433e-02 -1.14489973e+00
-8.69385004e-01 4.46357459e-01 3.23335081e-01 -4.21025828e-02
8.35348129e-01 -3.90348226e-01 8.13793063e-01 8.71053874e-01
1.15166235e+00 -1.38578916e+00 5.01130104e-01 1.13607478e+00
1.03216076e+00 -1.07855153e+00 -2.91833192e-01 9.95564163e-02
-1.08808994e+00 9.20147836e-01 9.35774505e-01 -2.64859855e-01
-3.38421986e-02 5.30271888e-01 1.79174226e-02 -1.26916230e-01
-1.32922602e+00 -4.73161787e-01 -4.26846743e-01 8.96630764e-01
-1.19894877e-01 2.75316834e-01 9.83560383e-02 3.99236381e-01
-4.55702245e-01 -1.09766193e-01 6.81907773e-01 1.12823069e+00
-6.65190160e-01 -1.13141072e+00 -3.95114183e-01 1.78054154e-01
-3.23976018e-02 -2.59350911e-02 -1.72316447e-01 8.15658808e-01
6.23505078e-02 1.18064940e+00 -2.64293224e-01 -4.54869986e-01
2.91863859e-01 -2.61873484e-01 3.71183604e-01 -5.35873175e-01
-1.01374257e+00 -2.54683197e-01 3.14293236e-01 -1.04185545e+00
-1.67116120e-01 -5.67556679e-01 -1.56471276e+00 -4.91219550e-01
5.18630296e-02 8.08094293e-02 4.48331296e-01 8.92325938e-01
6.05467260e-01 6.29112184e-01 3.42841327e-01 -1.25830519e+00
-8.35945547e-01 -5.39334059e-01 -3.32930923e-01 4.30467092e-02
1.49403781e-01 -1.02768099e+00 -7.53109679e-02 -5.88254213e-01] | [4.091130256652832, 1.5072613954544067] |
c253efa8-f8fe-45da-8e40-215d5a0b1b10 | otw-optimal-transport-warping-for-time-series | 2306.00620 | null | https://arxiv.org/abs/2306.00620v1 | https://arxiv.org/pdf/2306.00620v1.pdf | OTW: Optimal Transport Warping for Time Series | Dynamic Time Warping (DTW) has become the pragmatic choice for measuring distance between time series. However, it suffers from unavoidable quadratic time complexity when the optimal alignment matrix needs to be computed exactly. This hinders its use in deep learning architectures, where layers involving DTW computations cause severe bottlenecks. To alleviate these issues, we introduce a new metric for time series data based on the Optimal Transport (OT) framework, called Optimal Transport Warping (OTW). OTW enjoys linear time/space complexity, is differentiable and can be parallelized. OTW enjoys a moderate sensitivity to time and shape distortions, making it ideal for time series. We show the efficacy and efficiency of OTW on 1-Nearest Neighbor Classification and Hierarchical Clustering, as well as in the case of using OTW instead of DTW in Deep Learning architectures. | ['Steven C. H. Hoi', 'Doyen Sahoo', 'Chenghao Liu', 'Fabian Latorre'] | 2023-06-01 | null | null | null | null | ['dynamic-time-warping'] | ['time-series'] | [-1.08116269e-01 -4.97928411e-01 -4.29913476e-02 -2.61200517e-01
-7.53202856e-01 -7.17746973e-01 6.47149563e-01 3.23986501e-01
-7.41001725e-01 2.20495045e-01 1.89351700e-02 -5.25859594e-01
-6.44045234e-01 -6.73709035e-01 -3.93337816e-01 -8.91921282e-01
-8.82473648e-01 1.57462180e-01 2.02797636e-01 -2.39346206e-01
1.65204480e-01 8.36835563e-01 -1.11235905e+00 -5.36572672e-02
4.21580344e-01 1.12074292e+00 -2.19890792e-02 4.25730824e-01
-8.52968693e-02 3.50060463e-01 -4.39435065e-01 -1.79411411e-01
6.12031162e-01 -2.29032934e-01 -7.17242956e-01 -4.27005291e-01
2.22770378e-01 -2.93767065e-01 -5.36164582e-01 6.74932957e-01
6.39772594e-01 3.58391523e-01 6.19936168e-01 -1.63543415e+00
-5.73625088e-01 3.77263486e-01 -4.60764557e-01 5.01220465e-01
-1.60926759e-01 -1.40933558e-01 1.20620739e+00 -5.43393075e-01
4.58254069e-01 9.58292842e-01 1.11436462e+00 2.06016898e-01
-1.53301823e+00 -3.23040038e-01 -5.37483767e-02 5.26385844e-01
-1.40694368e+00 -7.41176456e-02 8.21511745e-01 -3.44288081e-01
1.04277885e+00 2.01959193e-01 7.30663419e-01 9.25953925e-01
4.73612547e-01 7.74347067e-01 9.56379712e-01 -6.03641681e-02
4.57217604e-01 -8.08412731e-01 6.07846817e-03 3.44774485e-01
-3.28677356e-01 -2.56040674e-02 -3.75028908e-01 -6.97671100e-02
6.90698564e-01 1.66911393e-01 -1.53474435e-01 -4.04791832e-01
-1.51343191e+00 8.30804825e-01 4.72648352e-01 6.20287955e-01
-2.10999638e-01 1.83714524e-01 7.39348292e-01 9.30109739e-01
5.48332274e-01 4.71970439e-01 -6.80958867e-01 -4.28943843e-01
-1.00794375e+00 3.24524999e-01 4.16021854e-01 5.08289218e-01
3.65244567e-01 -3.19658108e-02 -1.60607565e-02 7.02249825e-01
-2.21845523e-01 1.72732100e-01 6.24683976e-01 -1.11331701e+00
3.06493223e-01 3.74634653e-01 -1.49983108e-01 -1.13424504e+00
-8.12704086e-01 -1.88234970e-01 -1.05398846e+00 2.46178895e-01
7.54888356e-01 1.26691133e-01 -6.48928523e-01 1.83636403e+00
2.40427479e-01 3.52765888e-01 -7.36688450e-02 8.41694653e-01
2.21886843e-01 7.32428610e-01 -2.41233021e-01 -4.47729915e-01
9.75502074e-01 -6.80707633e-01 -6.28717482e-01 1.33612007e-01
8.84948134e-01 -7.07036436e-01 1.06363964e+00 3.07200283e-01
-9.57011759e-01 -1.63414091e-01 -1.14204311e+00 -4.36167926e-01
-5.59940100e-01 -4.27543730e-01 5.05758524e-01 3.53009224e-01
-1.33306682e+00 1.29674041e+00 -1.48380899e+00 -3.37955207e-01
3.98797363e-01 5.09514630e-01 -4.20443118e-01 1.47773877e-01
-1.05155993e+00 8.79978776e-01 1.03131451e-01 2.16683932e-02
-5.50031722e-01 -9.09339130e-01 -6.35635912e-01 -6.36112764e-02
-8.67625177e-02 -2.20767125e-01 9.28731799e-01 -4.11433607e-01
-1.60797822e+00 5.24186015e-01 1.89146604e-02 -6.01590991e-01
7.86972404e-01 -1.86158586e-02 -3.81872058e-01 6.91874977e-03
-1.69296414e-01 4.83305126e-01 9.67241108e-01 -2.03367978e-01
-2.33587921e-01 -4.35657620e-01 6.02028854e-02 -1.21011995e-01
-6.24524236e-01 -4.91884425e-02 -1.00068323e-01 -8.72605681e-01
4.71733630e-01 -1.14722311e+00 -2.39245996e-01 4.46698755e-01
-3.04193608e-02 -4.85574037e-01 1.10780811e+00 -5.70311248e-01
1.11311758e+00 -2.28285980e+00 1.85816750e-01 2.11659655e-01
3.53561670e-01 1.55334324e-01 -4.49104190e-01 7.50465631e-01
-3.16802442e-01 -1.15731493e-01 -3.60460401e-01 -2.92139024e-01
9.44596305e-02 3.84255171e-01 -3.76564384e-01 8.53186071e-01
1.21188484e-01 7.24643588e-01 -8.94825637e-01 -2.85086960e-01
2.13680446e-01 7.19236732e-01 -5.10377824e-01 -1.28109649e-01
1.96840614e-01 3.42888325e-01 -1.73398003e-01 2.32008219e-01
4.90474820e-01 -1.74399525e-01 3.63073386e-02 -2.56809920e-01
-3.72231454e-01 3.54560465e-01 -8.22010100e-01 1.69392145e+00
-4.81243402e-01 1.06405425e+00 -3.87177974e-01 -1.20630157e+00
9.71499920e-01 3.45047295e-01 1.12130260e+00 -8.82057428e-01
1.30482480e-01 3.62257391e-01 7.57878646e-02 -3.64124149e-01
1.44303575e-01 3.80704775e-02 1.47770762e-01 8.20018470e-01
-1.06280044e-01 6.57767355e-02 6.20978512e-02 -1.89549327e-01
1.38793051e+00 3.55321765e-02 1.84477374e-01 -6.29776895e-01
1.04863800e-01 -1.15618967e-01 5.22652090e-01 3.25486332e-01
-3.29352558e-01 6.10630274e-01 4.92883980e-01 -8.95579040e-01
-1.43792760e+00 -1.18192863e+00 -2.34411851e-01 1.03795743e+00
-1.49490133e-01 -4.47745413e-01 -5.59557140e-01 -3.83832991e-01
-1.28488347e-01 2.59341776e-01 -5.16812146e-01 -1.88911438e-01
-9.70088720e-01 -8.83146644e-01 5.63571334e-01 6.42804801e-01
4.68297154e-01 -5.06919205e-01 -8.16562355e-01 4.54529256e-01
-3.96169990e-01 -9.35440600e-01 -8.03845465e-01 2.24887148e-01
-1.29072523e+00 -8.89586806e-01 -8.06839705e-01 -5.37489772e-01
3.51579040e-01 4.92646039e-01 6.48624361e-01 -2.61036158e-01
-1.73506439e-01 9.82137546e-02 -2.87486941e-01 -1.36726201e-01
-7.86263868e-02 2.10410923e-01 2.34748229e-01 4.43056934e-02
3.66762221e-01 -1.05984271e+00 -8.84194016e-01 5.84399045e-01
-1.01711786e+00 -1.59232721e-01 4.51613665e-02 7.81108916e-01
4.28537518e-01 2.78955668e-01 2.96995044e-01 -2.67972294e-02
7.11131573e-01 -1.45788044e-01 -7.02343822e-01 7.38105997e-02
-5.50199389e-01 1.47508010e-01 7.78949082e-01 -6.34753585e-01
-2.25998059e-01 -1.58547074e-01 -1.95388854e-01 -5.32247484e-01
3.31525266e-01 5.39802372e-01 3.88929486e-01 -2.21178785e-01
6.42101943e-01 9.98680145e-02 1.75280362e-01 -5.73622346e-01
1.65099084e-01 3.67347300e-01 3.42697382e-01 -3.75827581e-01
7.34089732e-01 9.11973953e-01 2.66670078e-01 -9.43406880e-01
-6.02046371e-01 -2.08511963e-01 -1.03864741e+00 -7.47754797e-02
7.50471294e-01 -4.38278913e-01 -8.10351670e-01 5.60105622e-01
-1.04599977e+00 -3.85913372e-01 -2.02026501e-01 7.26414144e-01
-5.78882575e-01 3.89397591e-01 -6.94952190e-01 -4.65765446e-01
-2.68950671e-01 -1.03559136e+00 9.13343430e-01 -5.33554673e-01
-4.56586331e-01 -1.29930568e+00 -3.15151475e-02 -1.50908813e-01
5.28602242e-01 7.17097402e-01 1.15559435e+00 -6.51852548e-01
-3.02915663e-01 -4.53573555e-01 -1.04367495e-01 1.40786156e-01
4.78744619e-02 1.11591078e-01 -8.36293459e-01 -5.95901728e-01
3.52158755e-01 7.52776936e-02 5.82180738e-01 4.71579760e-01
1.37690485e+00 -3.09080452e-01 1.90256648e-02 7.46565461e-01
1.22846270e+00 2.94211328e-01 3.43948126e-01 5.04984081e-01
5.91297746e-01 7.69616842e-01 3.04913133e-01 3.94012183e-01
4.53074962e-01 8.99772227e-01 2.61371911e-01 4.06625792e-02
1.43419206e-01 1.33602560e-01 3.48328620e-01 1.41861784e+00
-2.96234012e-01 1.75108224e-01 -1.17699385e+00 5.98749876e-01
-1.89494622e+00 -9.54867661e-01 -2.67679989e-01 2.26067519e+00
6.46486342e-01 2.64997538e-02 1.93107218e-01 7.18519807e-01
3.84557575e-01 4.51745301e-01 -6.42950118e-01 -4.38135475e-01
7.32327625e-03 2.15650573e-01 5.70735991e-01 2.82357544e-01
-1.19096720e+00 4.57954049e-01 6.68761873e+00 9.16633844e-01
-1.43142593e+00 3.32474291e-01 3.60103458e-01 -2.78806537e-01
-3.84586230e-02 -3.08818281e-01 -1.83950856e-01 4.20707405e-01
1.04553807e+00 -3.60566765e-01 5.99938512e-01 5.17166495e-01
5.48933029e-01 5.39995193e-01 -1.31133318e+00 1.24505806e+00
-4.73299325e-01 -1.21795309e+00 -3.22871029e-01 1.72639281e-01
4.37111735e-01 4.81037110e-01 2.25999281e-01 5.12122773e-02
1.76201135e-01 -8.75535131e-01 6.10669017e-01 -9.80549753e-02
5.78920960e-01 -7.70817280e-01 5.23069263e-01 2.07747534e-01
-1.37725532e+00 -6.93155453e-02 -3.50801051e-01 -7.35064894e-02
1.50229439e-01 6.15518868e-01 -4.68800157e-01 2.08161414e-01
7.53456831e-01 8.64205480e-01 -1.65888891e-01 8.80488157e-01
2.85067469e-01 2.75651693e-01 -7.26009250e-01 2.53634751e-01
6.46406591e-01 -3.37203234e-01 4.53211367e-01 8.54238749e-01
7.10089684e-01 -1.10626251e-01 5.83775714e-02 3.81412894e-01
1.53126568e-01 4.34242934e-02 -6.70756638e-01 -1.02643974e-01
5.54000080e-01 9.52410102e-01 -1.00669253e+00 1.08124599e-01
-3.99126679e-01 9.14068401e-01 3.09411347e-01 2.73395598e-01
-8.59789252e-01 -4.93254870e-01 1.02426732e+00 -2.19911616e-02
1.59903973e-01 -9.92724717e-01 -2.57082522e-01 -1.02743316e+00
1.38237685e-01 -5.15705943e-01 5.81515849e-01 -3.63554925e-01
-1.37563038e+00 6.85737312e-01 1.20908648e-01 -1.64215314e+00
-1.37988567e-01 -7.57627189e-01 -4.23570395e-01 5.24647415e-01
-1.36171436e+00 -7.66401291e-01 -4.15396728e-02 9.00931060e-01
4.57995266e-01 1.69570729e-01 7.79627025e-01 5.79970181e-01
-4.15600479e-01 5.54050267e-01 3.92670333e-01 1.10609546e-01
4.05984819e-01 -1.19406950e+00 9.56429482e-01 6.78290725e-01
1.65621668e-01 4.94581640e-01 7.60937572e-01 -9.47174206e-02
-1.57266414e+00 -1.29220819e+00 1.08497202e+00 -2.93688238e-01
1.09903550e+00 -3.16552401e-01 -1.10374391e+00 5.85554361e-01
-4.79444489e-02 2.29215696e-02 8.00109148e-01 -3.47933173e-02
-6.46252513e-01 -2.58163720e-01 -1.01757205e+00 9.48431253e-01
1.25512147e+00 -8.68869007e-01 -4.16083843e-01 6.00903988e-01
8.04199219e-01 -2.70212173e-01 -1.42199564e+00 3.34074825e-01
7.01016366e-01 -1.01958704e+00 1.03195834e+00 -4.24293488e-01
5.47603294e-02 -2.69474208e-01 -9.85009447e-02 -1.33288360e+00
-4.32744682e-01 -1.04454112e+00 9.36904028e-02 6.87379479e-01
2.40655482e-01 -9.16258097e-01 3.97160023e-01 5.57076454e-01
-8.88671279e-02 -8.19615424e-01 -1.33186722e+00 -1.30637836e+00
3.06208968e-01 -7.71918356e-01 7.89841771e-01 1.52860844e+00
1.00774160e-02 -1.54326916e-01 -1.74959615e-01 -3.90562229e-02
4.68071699e-01 1.88195214e-01 4.09853518e-01 -1.25666320e+00
1.52954813e-02 -6.61376357e-01 -6.97559893e-01 -9.37607586e-01
-7.41927791e-03 -8.57768118e-01 -2.58239090e-01 -1.08759630e+00
-4.69159096e-01 -2.17831314e-01 -4.58975166e-01 5.03794193e-01
4.48717922e-01 3.90178114e-01 1.16813131e-01 5.80031216e-01
-1.30148128e-01 6.04685485e-01 1.21866810e+00 -3.75478854e-03
-2.14454919e-01 -1.29862532e-01 3.69699970e-02 7.56919265e-01
9.80952442e-01 -4.92411792e-01 -4.80547488e-01 -8.60950708e-01
1.11142039e-01 -1.85330082e-02 2.08281860e-01 -9.46978092e-01
2.57789582e-01 -4.17223610e-02 3.15912557e-03 -5.84049702e-01
4.77282107e-01 -9.43467438e-01 1.59415558e-01 4.79594797e-01
-4.07910973e-01 9.05142009e-01 2.14258977e-03 3.42045009e-01
-2.78073072e-01 1.20722055e-01 8.53853583e-01 2.10688397e-01
-4.20041233e-01 6.75838947e-01 -3.29174668e-01 -1.11641958e-01
9.62844849e-01 -2.79382944e-01 -2.20050588e-01 -3.38385403e-01
-6.22673094e-01 1.66995570e-01 3.28861505e-01 5.56484640e-01
5.88155448e-01 -1.64048326e+00 -4.85676944e-01 2.02368796e-01
-6.41154498e-02 -2.79512763e-01 2.74184257e-01 1.38748622e+00
-8.36134374e-01 3.78161728e-01 -2.57933199e-01 -7.65980601e-01
-1.15044415e+00 5.96410990e-01 2.73585081e-01 -1.49190679e-01
-1.12071645e+00 4.94201958e-01 -8.13409910e-02 -3.71242046e-01
2.44421363e-01 -5.60276210e-01 -5.86859770e-02 2.23147362e-01
5.56241632e-01 5.79887450e-01 3.60475540e-01 -4.69534427e-01
-2.88387209e-01 7.59094596e-01 6.10803440e-03 -1.82976931e-01
1.55093944e+00 -1.79763362e-01 -3.23159337e-01 6.74894810e-01
1.78655457e+00 -7.17182040e-01 -1.36741352e+00 -2.48856202e-01
5.14949024e-01 -4.40090060e-01 2.35216469e-01 1.18409581e-01
-1.09338462e+00 1.15413153e+00 6.04744256e-01 6.80918872e-01
1.18525517e+00 -4.33690131e-01 1.29387200e+00 5.77318132e-01
4.22783077e-01 -9.78061974e-01 -7.27668554e-02 6.24732494e-01
7.79750943e-01 -8.63372743e-01 -1.47483423e-01 1.55564502e-01
-1.47537038e-01 1.42347372e+00 -2.35539135e-02 -2.69767672e-01
1.03215587e+00 2.47137949e-01 1.42157257e-01 -2.59607360e-02
-7.56381512e-01 -2.24702340e-02 1.98311076e-01 4.84893143e-01
3.83527309e-01 -1.04380816e-01 -4.32532072e-01 -2.65575588e-01
-4.57476079e-01 -2.52464473e-01 2.12382615e-01 8.73858750e-01
-3.33689041e-02 -1.08423364e+00 -4.86838445e-02 2.70046681e-01
-5.51670313e-01 7.25688860e-02 -3.08655039e-03 8.14311564e-01
-1.36345133e-01 7.34078467e-01 4.41317588e-01 -4.88370776e-01
2.65049517e-01 -1.13709513e-02 4.33927983e-01 9.23276246e-02
-4.53935742e-01 8.31068531e-02 -1.97257906e-01 -1.01121604e+00
-5.85372686e-01 -6.17195368e-01 -1.06538403e+00 -7.62856841e-01
1.16855413e-01 -6.53960556e-02 8.02440703e-01 1.10559022e+00
4.41712648e-01 1.85653448e-01 7.75247335e-01 -8.84806395e-01
-5.88475525e-01 -6.19493186e-01 -5.32653868e-01 4.23779011e-01
7.54837573e-01 -5.10871828e-01 -3.21124494e-01 -3.35862823e-02] | [7.32248592376709, 3.328270673751831] |
8432a602-f578-4090-bc80-e4726bb33d54 | enquire-one-s-parent-and-child-before | 2101.11268 | null | https://arxiv.org/abs/2101.11268v1 | https://arxiv.org/pdf/2101.11268v1.pdf | Enquire One's Parent and Child Before Decision: Fully Exploit Hierarchical Structure for Self-Supervised Taxonomy Expansion | Taxonomy is a hierarchically structured knowledge graph that plays a crucial role in machine intelligence. The taxonomy expansion task aims to find a position for a new term in an existing taxonomy to capture the emerging knowledge in the world and keep the taxonomy dynamically updated. Previous taxonomy expansion solutions neglect valuable information brought by the hierarchical structure and evaluate the correctness of merely an added edge, which downgrade the problem to node-pair scoring or mini-path classification. In this paper, we propose the Hierarchy Expansion Framework (HEF), which fully exploits the hierarchical structure's properties to maximize the coherence of expanded taxonomy. HEF makes use of taxonomy's hierarchical structure in multiple aspects: i) HEF utilizes subtrees containing most relevant nodes as self-supervision data for a complete comparison of parental and sibling relations; ii) HEF adopts a coherence modeling module to evaluate the coherence of a taxonomy's subtree by integrating hypernymy relation detection and several tree-exclusive features; iii) HEF introduces the Fitting Score for position selection, which explicitly evaluates both path and level selections and takes full advantage of parental relations to interchange information for disambiguation and self-correction. Extensive experiments show that by better exploiting the hierarchical structure and optimizing taxonomy's coherence, HEF vastly surpasses the prior state-of-the-art on three benchmark datasets by an average improvement of 46.7% in accuracy and 32.3% in mean reciprocal rank. | ['Bang Liu', 'Yefeng Zheng', 'Xi Chen', 'Ruihui Zhao', 'Suyuchen Wang'] | 2021-01-27 | null | null | null | null | ['taxonomy-expansion'] | ['natural-language-processing'] | [ 3.41458917e-02 2.66591012e-01 -5.85503638e-01 -2.57083595e-01
1.16081394e-01 -5.40389299e-01 3.32881063e-01 6.59332395e-01
-2.13296384e-01 5.72128534e-01 3.04474056e-01 -3.22225571e-01
-5.55021703e-01 -1.03551853e+00 -4.83239740e-02 -4.55048651e-01
-2.03458995e-01 6.73268676e-01 6.33276880e-01 -2.96969563e-01
2.43608564e-01 2.81972438e-01 -1.82556379e+00 4.62425873e-02
1.25184071e+00 1.05656755e+00 2.58673310e-01 -7.68798217e-02
-4.22825634e-01 5.67902744e-01 -4.97779578e-01 -6.87673748e-01
1.12623096e-01 -6.32292554e-02 -1.14006424e+00 -1.07488841e-01
1.99281141e-01 1.87838357e-02 -9.58988219e-02 1.34327686e+00
2.55920570e-02 -6.26742691e-02 3.99494529e-01 -1.33502090e+00
-2.49782696e-01 7.83790946e-01 -6.10119998e-01 2.51640677e-01
5.18390834e-01 -4.13780212e-01 1.67521691e+00 -6.80850089e-01
8.96612346e-01 1.13106132e+00 6.26102090e-01 -5.47598414e-02
-1.00833356e+00 -7.55105495e-01 4.05525744e-01 7.05170989e-01
-1.72322130e+00 1.19925030e-01 5.96375227e-01 -5.22709310e-01
1.26949191e+00 4.16972101e-01 7.18878984e-01 4.54557002e-01
-6.94456929e-03 2.96388209e-01 7.70693541e-01 -3.47741395e-01
1.53349325e-01 -2.03042682e-02 6.47840500e-01 8.31176996e-01
4.57063138e-01 1.67090483e-02 -6.21934772e-01 -2.21745461e-01
4.05222088e-01 -3.93641591e-01 -3.03428411e-01 -5.51647544e-01
-9.54094172e-01 6.14773452e-01 8.43266666e-01 4.09286171e-01
-2.82997996e-01 -7.75572479e-01 3.11596930e-01 2.34104082e-01
8.18843767e-02 8.65400195e-01 -4.62936729e-01 1.01963282e-01
-6.86644256e-01 1.09939342e-02 8.38485003e-01 1.02337921e+00
9.79017377e-01 -4.38831002e-01 5.89140207e-02 9.94786024e-01
1.78727377e-02 -2.03256905e-01 6.76421046e-01 -8.91668379e-01
1.86489940e-01 1.67565978e+00 -4.19726819e-01 -9.46536779e-01
-7.45542288e-01 -9.05366659e-01 -7.87146568e-01 -2.12709263e-01
-7.11353868e-02 6.01126492e-01 -6.82087004e-01 1.72206795e+00
7.70252168e-01 5.79281896e-02 -1.86779737e-01 6.11970127e-01
1.14718556e+00 1.60576254e-01 -8.52076858e-02 -4.95779097e-01
1.57408857e+00 -8.00150573e-01 -5.09614229e-01 7.62254223e-02
6.57110989e-01 -4.24202293e-01 9.56237912e-01 3.48582655e-01
-4.96028513e-01 -3.29361647e-01 -1.20401478e+00 -8.34264755e-02
-5.99705696e-01 -2.76889920e-01 7.82554746e-01 5.32864392e-01
-7.94617414e-01 5.79293847e-01 -5.04964173e-01 -4.98888433e-01
5.11626294e-03 4.88654077e-01 -5.18349469e-01 -4.95601594e-02
-1.55974805e+00 9.16203260e-01 1.07077742e+00 -3.59925210e-01
-3.05973947e-01 -8.35380495e-01 -8.85528803e-01 4.26813841e-01
8.96623254e-01 -8.56727540e-01 1.00019002e+00 -2.15989068e-01
-1.08865571e+00 7.35867441e-01 -2.32961357e-01 -3.21641862e-01
3.75752598e-02 1.55069083e-01 -5.48614204e-01 1.22104056e-01
3.94792765e-01 4.54155535e-01 8.15955475e-02 -1.06093347e+00
-1.12971628e+00 -6.62704587e-01 2.89669484e-01 4.43168849e-01
-6.94093823e-01 -1.84610412e-01 -7.68874824e-01 -4.81469244e-01
6.99241400e-01 -7.19527304e-01 2.67289318e-02 -5.17259717e-01
-5.01943648e-01 -7.11065531e-01 5.40180027e-01 -4.88612831e-01
1.86399162e+00 -1.58566833e+00 3.39117289e-01 4.24547344e-01
7.00529456e-01 1.41875833e-01 1.59931704e-01 5.77043414e-01
-3.40192169e-02 1.32520720e-02 -2.16511726e-01 -7.14923441e-02
-1.36057645e-01 4.74826157e-01 -6.68271407e-02 -2.15571225e-02
-3.36283863e-01 7.62069583e-01 -1.10313654e+00 -7.29022920e-01
7.56705180e-02 1.79752484e-01 -6.85956538e-01 9.42919683e-03
-1.05169900e-02 8.62054229e-02 -1.64909095e-01 7.49646187e-01
4.04464155e-01 -4.45135295e-01 7.04692900e-01 -3.52686644e-01
-2.88942922e-02 7.76062667e-01 -1.28150415e+00 1.56527412e+00
-1.76539794e-01 -3.82711925e-02 -1.50575474e-01 -7.02548385e-01
1.04106426e+00 5.04668728e-02 6.02994382e-01 -6.08244598e-01
-5.52677624e-02 2.88517922e-01 2.81637490e-01 -2.30156809e-01
5.20212650e-01 2.03954846e-01 1.91864297e-01 7.36899152e-02
1.11208737e-01 6.88380003e-02 5.57963431e-01 5.93551934e-01
1.36070418e+00 -4.04300056e-02 9.07122612e-01 -3.12470257e-01
7.48173237e-01 -1.03961930e-01 1.02019823e+00 4.11722541e-01
6.83745965e-02 -1.99457817e-02 4.84292388e-01 -2.87078738e-01
-5.70734262e-01 -8.74358714e-01 -3.70310038e-01 1.25225163e+00
5.09349406e-01 -1.12027204e+00 -4.04238373e-01 -8.11084867e-01
2.44078755e-01 7.85076797e-01 -5.42325795e-01 -3.15596372e-01
-2.33015969e-01 -5.99235773e-01 2.44698197e-01 4.23547149e-01
5.17714977e-01 -8.76257539e-01 -5.06716907e-01 1.22282162e-01
-4.13771689e-01 -1.18834412e+00 -2.02475101e-01 3.16084802e-01
-7.91420341e-01 -1.29186380e+00 1.16661154e-01 -4.71966714e-01
3.76315892e-01 3.83739084e-01 1.18681371e+00 2.77754694e-01
-1.60441741e-01 -1.02936395e-01 -7.36749470e-01 9.10195857e-02
-1.42745346e-01 6.77405357e-01 1.85977250e-01 -4.00829494e-01
5.55361092e-01 -1.04022145e+00 -2.95801103e-01 6.44534171e-01
-5.94776273e-01 1.08916517e-02 5.02476096e-01 9.22794521e-01
4.41908062e-01 6.18464947e-01 4.46317852e-01 -1.05301976e+00
4.78216618e-01 -4.07481104e-01 -6.62789822e-01 5.09970844e-01
-1.47391593e+00 1.64911658e-01 3.35614324e-01 -6.15047403e-02
-9.96690750e-01 -1.50305644e-01 -2.05648299e-02 -2.07232475e-01
2.70997554e-01 9.33253646e-01 -5.01819909e-01 1.00778826e-01
4.13639873e-01 -1.72814005e-03 -5.00621378e-01 -6.09337330e-01
2.62733042e-01 6.30192637e-01 9.13195491e-01 -7.23013818e-01
6.51840448e-01 1.87790170e-01 2.18945742e-01 -4.14358616e-01
-1.07288527e+00 -1.00543082e+00 -1.03797960e+00 5.43616191e-02
3.51122528e-01 -6.95901215e-01 -1.00329208e+00 -1.14433393e-01
-8.78260434e-01 4.25496936e-01 -4.46610749e-02 2.46204004e-01
1.64453074e-01 6.99029565e-01 -4.53073025e-01 -6.40897036e-01
-4.15676683e-01 -9.40961659e-01 9.31212127e-01 1.60590142e-01
-3.77324373e-01 -7.03220308e-01 2.49910844e-03 4.12548661e-01
-1.05130374e-01 3.77008761e-03 1.36769891e+00 -8.55482817e-01
-4.89931196e-01 -1.86301440e-01 -3.96677792e-01 -8.00670981e-02
2.54408896e-01 -9.82991084e-02 -7.25600541e-01 -1.81597158e-01
-2.79912621e-01 7.83827007e-02 1.02184212e+00 -1.80090442e-01
8.82865131e-01 -1.85143143e-01 -6.32492304e-01 7.70546973e-01
1.16163468e+00 1.41627088e-01 3.93360823e-01 6.48226678e-01
7.80249357e-01 9.65032101e-01 6.70604527e-01 3.92237365e-01
8.09218526e-01 1.00905192e+00 6.76991880e-01 4.15913016e-01
-1.06140830e-01 -4.88726079e-01 -1.21486321e-01 1.02134693e+00
-7.39675900e-03 1.84466653e-02 -1.03695011e+00 4.86348003e-01
-1.82903588e+00 -7.54460812e-01 -2.33497798e-01 2.48430514e+00
1.14170313e+00 4.12803829e-01 1.95146918e-01 4.53748703e-01
7.41040170e-01 -1.86135069e-01 -5.75693488e-01 9.81682073e-03
-1.48176000e-01 3.41998003e-02 2.74559855e-01 6.31202877e-01
-8.75436485e-01 1.11446702e+00 5.50767183e+00 8.26186478e-01
-6.58890903e-01 -5.20708486e-02 -9.92308930e-02 2.34824508e-01
-2.47739494e-01 2.91063994e-01 -1.12850463e+00 2.54684567e-01
4.25055534e-01 -4.87067819e-01 1.87110841e-01 8.79016519e-01
-5.04067600e-01 -1.09746903e-01 -1.04022086e+00 6.32376611e-01
-2.38101453e-01 -1.01961780e+00 1.72708347e-01 3.45206112e-01
4.69837159e-01 -9.72326100e-02 -3.23690593e-01 4.61770535e-01
6.45139277e-01 -7.66693771e-01 3.09841275e-01 2.14999914e-02
7.13105083e-01 -7.22577155e-01 9.60084915e-01 4.03008074e-01
-1.70578969e+00 -2.81502098e-01 -3.81255984e-01 -2.09760413e-01
-2.22233823e-03 8.53172779e-01 -1.09011340e+00 1.19205308e+00
8.61219764e-01 7.32138455e-01 -8.58211100e-01 1.18342578e+00
-6.34572864e-01 4.34877336e-01 -4.49941278e-01 2.04260334e-01
7.74513558e-02 -1.60114676e-01 7.81212747e-01 1.13297796e+00
4.05887663e-02 7.27465004e-02 5.00992954e-01 5.47908187e-01
9.11486056e-03 2.94740617e-01 -1.31181329e-01 2.61931956e-01
1.17985797e+00 1.36730647e+00 -7.85475612e-01 -3.84311557e-01
-1.11010902e-01 6.86905563e-01 6.19193673e-01 -1.17037762e-02
-5.07043958e-01 -4.95430201e-01 6.96862698e-01 3.73849347e-02
2.12705940e-01 1.48577020e-01 -4.22697455e-01 -1.12413895e+00
8.60290527e-02 -6.34858549e-01 9.68458951e-01 -7.00240672e-01
-1.16613114e+00 8.19209039e-01 1.09783448e-01 -1.06593382e+00
-3.70737284e-01 -4.29950982e-01 -1.59373283e-01 6.14991546e-01
-1.34273303e+00 -1.15253210e+00 -5.56728065e-01 2.98844546e-01
1.77892670e-01 -1.64815009e-01 9.42250013e-01 2.68467546e-01
-8.67838383e-01 8.58856618e-01 -3.39515001e-01 -1.58270791e-01
5.86457849e-01 -1.58065331e+00 3.25618863e-01 8.16213548e-01
2.23629445e-01 9.65266347e-01 6.42284632e-01 -9.68773484e-01
-6.71972752e-01 -7.87718236e-01 1.22422945e+00 -3.04322064e-01
8.78238022e-01 -1.16437450e-01 -1.35223532e+00 4.68407094e-01
-1.18774675e-01 -3.37465107e-01 7.02959180e-01 7.87604272e-01
-1.07927418e+00 -1.72748491e-01 -9.74277258e-01 3.74706954e-01
1.71912920e+00 -2.58906126e-01 -9.26900923e-01 4.11123298e-02
9.84451532e-01 -2.11881578e-01 -1.20796096e+00 9.12998796e-01
6.74568772e-01 -9.97639537e-01 9.66060042e-01 -4.09322917e-01
3.15915167e-01 -4.85503107e-01 -1.51977524e-01 -1.17043757e+00
-7.55227268e-01 -4.96500164e-01 -3.66586566e-01 1.42478037e+00
2.49529958e-01 -6.87778950e-01 7.16993034e-01 1.92570791e-01
-2.01076850e-01 -9.67330933e-01 -9.80156600e-01 -1.01743948e+00
-2.75026411e-01 -4.68091816e-02 1.01663959e+00 1.27204156e+00
4.94002253e-01 8.51750195e-01 9.92909260e-03 2.72413224e-01
5.90052128e-01 4.78906870e-01 4.73477900e-01 -2.00275087e+00
-1.10654846e-01 -8.43385816e-01 -6.94108427e-01 -8.53976071e-01
2.56682336e-01 -1.29054546e+00 -3.58711988e-01 -1.62531126e+00
3.87208432e-01 -4.07435864e-01 -5.07656753e-01 7.93101609e-01
-5.65755308e-01 -6.95544556e-02 -5.56461923e-02 5.14678657e-01
-7.94791639e-01 4.68172163e-01 9.18712020e-01 3.62532139e-02
-2.15762377e-01 -7.37352148e-02 -8.81307542e-01 7.89969981e-01
4.09959525e-01 -4.00271714e-01 -4.76872265e-01 4.39954631e-04
3.15815508e-01 -2.37153888e-01 -1.47364950e-02 -8.28823090e-01
5.86796403e-01 -3.63414846e-02 3.11149657e-02 -6.47616744e-01
1.28808871e-01 -7.90540576e-01 2.09630504e-01 4.52926964e-01
-2.33986422e-01 8.74360837e-03 -8.37385282e-02 5.17756104e-01
-2.83615679e-01 -3.20383072e-01 5.28201520e-01 1.21604167e-01
-1.09036410e+00 9.46514159e-02 2.89046735e-01 -1.06588729e-01
7.06960559e-01 -3.22357804e-01 -4.25304443e-01 1.46836760e-02
-6.10457838e-01 5.51531136e-01 6.10159814e-01 5.56716263e-01
3.62596542e-01 -1.19327188e+00 -4.28155363e-01 3.98263782e-02
6.74326479e-01 -2.29876935e-02 -7.80210495e-02 6.39545202e-01
3.12515199e-02 5.14823556e-01 -3.87304388e-02 -5.78839958e-01
-1.60383594e+00 7.01858521e-01 1.23066105e-01 -8.58554721e-01
-6.08151436e-01 1.07669055e+00 2.86206007e-01 -6.40717089e-01
3.89196515e-01 -1.71913460e-01 -5.77559173e-01 2.31260628e-01
3.99821877e-01 3.34792942e-01 4.34814572e-01 -7.08820105e-01
-6.23998702e-01 4.79871929e-01 -3.59497249e-01 1.97668329e-01
1.00647831e+00 -1.08879432e-01 -6.03796363e-01 2.89589077e-01
8.42656076e-01 3.81885655e-03 -5.71162999e-01 -7.56086588e-01
6.94457352e-01 -5.08114815e-01 2.85137743e-02 -1.20859814e+00
-8.50165725e-01 4.65312928e-01 1.00526951e-01 2.35625669e-01
1.25824189e+00 5.64648658e-02 6.15498960e-01 3.77911001e-01
6.66843712e-01 -7.13078558e-01 -1.30023643e-01 5.91260672e-01
7.82327354e-01 -8.27768147e-01 5.06586015e-01 -1.10028028e+00
-5.00829339e-01 8.47739518e-01 8.79653692e-01 5.96870840e-01
5.80177903e-01 1.75559551e-01 -3.55693549e-01 -3.76066834e-01
-8.78668427e-01 -5.95771611e-01 5.97383380e-01 5.37511766e-01
3.88019234e-01 1.60691544e-01 -6.16172016e-01 7.24831283e-01
-8.18559110e-01 -2.60996431e-01 -2.36402731e-02 4.75216150e-01
-6.47251964e-01 -1.26073635e+00 -5.75179122e-02 5.23619533e-01
-2.78140306e-02 -1.84108570e-01 -8.40528309e-01 6.49768591e-01
4.88993794e-01 9.42957163e-01 -1.71412468e-01 -8.02844882e-01
3.10893148e-01 1.46348909e-01 2.80460715e-01 -8.87378991e-01
-6.92354739e-01 -6.83993399e-02 1.74355730e-01 -4.93182540e-01
-1.71409965e-01 -4.48599756e-01 -1.42080605e+00 -2.79549211e-01
-6.24056160e-01 4.46804553e-01 2.88260639e-01 9.00844932e-01
5.87898135e-01 7.22315907e-01 4.03263271e-01 -9.00531486e-02
-4.73259449e-01 -1.09644556e+00 -6.06986940e-01 4.14753824e-01
-2.15907484e-01 -1.34439826e+00 -2.66949475e-01 -4.06903446e-01] | [9.19924545288086, 7.981405735015869] |
2d017a8d-7ac0-4c83-8baf-6679d034487c | designing-for-recommending-intermediate | 2010.04880 | null | https://arxiv.org/abs/2010.04880v1 | https://arxiv.org/pdf/2010.04880v1.pdf | Designing for Recommending Intermediate States in A Scientific Workflow Management System | To process a large amount of data sequentially and systematically, proper management of workflow components (i.e., modules, data, configurations, associations among ports and links) in a Scientific Workflow Management System (SWfMS) is inevitable. Managing data with provenance in a SWfMS to support reusability of workflows, modules, and data is not a simple task. Handling such components is even more burdensome for frequently assembled and executed complex workflows for investigating large datasets with different technologies (i.e., various learning algorithms or models). However, a great many studies propose various techniques and technologies for managing and recommending services in a SWfMS, but only a very few studies consider the management of data in a SWfMS for efficient storing and facilitating workflow executions. Furthermore, there is no study to inquire about the effectiveness and efficiency of such data management in a SWfMS from a user perspective. In this paper, we present and evaluate a GUI version of such a novel approach of intermediate data management with two use cases (Plant Phenotyping and Bioinformatics). The technique we call GUI-RISPTS (Recommending Intermediate States from Pipelines Considering Tool-States) can facilitate executions of workflows with processed data (i.e., intermediate outcomes of modules in a workflow) and can thus reduce the computational time of some modules in a SWfMS. We integrated GUI-RISPTS with an existing workflow management system called SciWorCS. In SciWorCS, we present an interface that users use for selecting the recommendation of intermediate states (i.e., modules' outcomes). We investigated GUI-RISP's effectiveness from users' perspectives along with measuring its overhead in terms of storage and efficiency in workflow execution. | ['Sristy Sumana Nath', 'Banani Roy', 'Debasish Chakroborti'] | 2020-10-10 | null | null | null | null | ['plant-phenotyping'] | ['computer-vision'] | [-2.08509602e-02 -3.04385751e-01 3.54751021e-01 -4.48768348e-01
5.61163984e-02 -9.65269804e-01 4.44327354e-01 8.63917291e-01
-2.37761199e-01 4.90877867e-01 -3.42980176e-01 -7.10744917e-01
-6.11005723e-01 -1.06522489e+00 -4.83152837e-01 -6.13285005e-01
-9.18096006e-02 3.68429869e-01 4.49798942e-01 3.93253952e-01
3.45007181e-01 1.06675982e+00 -1.74449444e+00 3.84931892e-01
9.85210598e-01 4.82392609e-01 8.23487043e-01 7.51631975e-01
-7.40617633e-01 1.81871742e-01 -1.03184772e+00 -2.87707373e-02
2.81059984e-02 -3.12008560e-01 -8.52605402e-01 -2.49599576e-01
-1.32852316e-01 9.51803848e-02 6.03543878e-01 6.85521185e-01
6.37777567e-01 -2.02969268e-01 1.50545493e-01 -1.66282773e+00
9.16654989e-02 4.38652664e-01 -1.83697939e-01 -1.54919520e-01
2.35521495e-01 6.92477942e-01 1.37301221e-01 -5.59772551e-01
1.02868176e+00 6.45324171e-01 3.17549616e-01 1.27730846e-01
-1.16134584e+00 -4.52271849e-01 -6.87856078e-02 2.50703096e-01
-1.46453810e+00 -5.50282061e-01 -1.55541062e-01 -6.86078310e-01
7.29480326e-01 9.30336773e-01 7.13235438e-01 3.60058665e-01
2.45563477e-01 -1.66900918e-01 8.87115300e-01 -3.66639644e-01
6.78523540e-01 3.36051106e-01 3.79750609e-01 6.81614399e-01
7.03196764e-01 -5.36507189e-01 -7.13182330e-01 -3.29461485e-01
3.04148376e-01 9.26121771e-02 -1.71439007e-01 -2.05638021e-01
-1.18540680e+00 -1.24101900e-01 -1.82085738e-01 4.83303398e-01
-3.24050874e-01 -3.32212150e-01 4.47273791e-01 3.38305056e-01
7.45191658e-03 4.45030153e-01 -7.41494536e-01 -2.76386976e-01
-6.98952019e-01 -1.39251292e-01 1.25182140e+00 1.38254416e+00
7.06069708e-01 -4.45298731e-01 -5.23819029e-01 1.93413883e-01
1.36350319e-01 3.34992222e-02 1.55555651e-01 -5.01650810e-01
-6.06604181e-02 1.22670078e+00 3.78149807e-01 -3.21531862e-01
-8.41233134e-01 -7.18930811e-02 -7.14247227e-01 4.00810987e-01
5.60132682e-01 2.77924221e-02 -6.44150913e-01 1.22297049e+00
7.54307210e-01 -1.67927071e-01 -1.68573439e-01 7.06649125e-01
9.11062658e-01 4.84839112e-01 5.83250523e-01 -5.33530414e-01
1.61793780e+00 -4.30571407e-01 -7.41316497e-01 4.25020635e-01
7.33787358e-01 -1.11030805e+00 1.20891941e+00 3.44339669e-01
-1.04140174e+00 -6.29188240e-01 -5.83513260e-01 1.18686385e-01
-9.31619585e-01 2.13372290e-01 6.66151822e-01 5.02051950e-01
-9.52666461e-01 1.13782310e+00 -9.78281617e-01 -9.96829927e-01
9.03865844e-02 1.60871670e-01 -3.84576410e-01 1.81323394e-01
-5.10984063e-01 9.07760143e-01 5.16742706e-01 1.88264504e-01
-7.36535072e-01 -1.08385718e+00 -3.23660016e-01 4.76682425e-01
4.11697745e-01 -8.11317444e-01 9.74834263e-01 -2.91015804e-01
-1.24952531e+00 5.31074762e-01 -2.09454834e-01 2.61147052e-01
5.22441089e-01 1.98549077e-01 -6.33062124e-01 -2.83582330e-01
-8.42618421e-02 -1.31946549e-01 1.93250507e-01 -1.06088388e+00
-7.93840587e-01 -5.11946023e-01 -5.20368554e-02 -2.18692526e-01
-1.89673707e-01 2.85644352e-01 -6.13220632e-01 1.62617832e-01
-9.82169062e-02 -5.47882676e-01 -2.08024368e-01 1.48519576e-01
-4.14939940e-01 -1.38469636e-01 6.39631271e-01 -5.54801345e-01
1.38786912e+00 -1.95371783e+00 -2.45804638e-02 2.77258545e-01
-1.24627553e-01 5.61530411e-01 7.83122052e-03 8.23777556e-01
-1.49244741e-02 5.84463716e-01 -1.11706220e-01 5.01470976e-02
1.46645963e-01 1.94221124e-01 3.58881801e-01 1.91132277e-01
-8.87212977e-02 1.95446849e-01 -7.64108479e-01 -4.05912042e-01
3.17657411e-01 3.88569653e-01 -8.00388008e-02 7.39633501e-01
-4.39397156e-01 7.02936590e-01 -4.01978791e-01 8.48368704e-01
9.91279542e-01 -3.25991392e-01 7.26953983e-01 -4.37799335e-01
-1.25978494e+00 6.35492802e-02 -1.48661458e+00 1.73393142e+00
-5.53334892e-01 -1.85403824e-02 1.95659548e-01 -4.50443745e-01
8.27445805e-01 6.11105919e-01 3.93855035e-01 -1.17525287e-01
-4.87452984e-01 8.62694755e-02 -1.06903896e-01 -1.03730190e+00
3.24367613e-01 6.28602743e-01 1.92907840e-01 7.49642909e-01
9.19766054e-02 2.41994038e-01 8.35327864e-01 1.67465165e-01
1.49349010e+00 8.35225046e-01 6.19338334e-01 -4.64126378e-01
5.85686147e-01 3.18638563e-01 6.20738864e-01 6.70871019e-01
1.89923540e-01 -3.73119265e-02 7.73468614e-01 -5.90893984e-01
-8.10421050e-01 -7.89284825e-01 -7.00887665e-02 1.09662676e+00
-2.02327549e-01 -1.08863020e+00 -3.99101228e-01 -5.15869737e-01
-6.09405413e-02 7.92107403e-01 -2.56237574e-02 2.76449084e-01
-2.44244933e-03 -6.20572388e-01 3.56425852e-01 5.21541759e-02
1.75230682e-01 -1.12949884e+00 -1.26244271e+00 1.58740968e-01
1.93724558e-01 -5.54431319e-01 -1.20588489e-01 2.14978456e-01
-7.60663688e-01 -1.47109580e+00 1.74967088e-02 -1.57041356e-01
7.90542662e-01 2.56642997e-01 1.23736191e+00 3.04308146e-01
-6.68267846e-01 1.11416519e-01 -5.18717408e-01 -7.65886009e-01
-5.13633072e-01 6.92283586e-02 -4.49268639e-01 -4.29106981e-01
4.10047285e-02 -4.77077514e-01 -5.38479626e-01 4.78677243e-01
-1.38253582e+00 4.14005518e-01 3.71016294e-01 1.82329610e-01
6.17289603e-01 5.38648963e-02 3.67210448e-01 -1.10922968e+00
4.84934717e-01 -5.48808157e-01 -1.23515332e+00 8.86421323e-01
-8.56953859e-01 7.63307065e-02 9.90935683e-01 7.03000575e-02
-1.17977881e+00 1.00085050e-01 -2.06421576e-02 -1.21417930e-02
-6.95919514e-01 8.64705145e-01 -5.80957353e-01 2.22162560e-01
2.27416649e-01 -1.83614884e-02 8.48456174e-02 -9.93506789e-01
1.54636592e-01 6.88829303e-01 1.07420154e-01 -5.35490930e-01
4.97779429e-01 2.23990262e-01 4.22238290e-01 -5.06453454e-01
8.40977430e-02 -5.89125156e-01 -6.59837127e-01 -3.88571292e-01
2.69955635e-01 -1.59936219e-01 -1.51521027e+00 2.31038094e-01
-1.28520560e+00 -1.17495328e-01 -2.71132231e-01 2.04135910e-01
-9.25013125e-02 1.75697699e-01 -6.79225847e-02 -8.57146263e-01
-5.18219173e-01 -1.02485335e+00 7.48842716e-01 3.20439935e-01
-1.48236901e-01 -5.42842388e-01 6.99971616e-03 -2.45543912e-01
7.42430806e-01 5.14716268e-01 9.13843274e-01 -4.35968280e-01
-7.88675487e-01 -4.07892950e-02 -1.10410362e-01 -3.23683619e-01
1.98176727e-01 8.15798998e-01 -9.51782584e-01 -2.25248203e-01
-3.60619038e-01 5.50498843e-01 5.21193668e-02 -2.00688899e-01
1.57272983e+00 -1.10843070e-01 -5.03144622e-01 6.61525607e-01
1.63019097e+00 3.12969357e-01 6.12956583e-01 3.84785473e-01
3.91852975e-01 9.21467245e-01 8.65664005e-01 8.21125329e-01
-1.72576115e-01 5.85003555e-01 4.71647799e-01 -1.47912383e-01
2.25577086e-01 2.48570785e-01 3.33778888e-01 6.29997432e-01
-3.74806613e-01 -2.91062295e-01 -9.97956634e-01 8.38247221e-03
-1.91946149e+00 -7.42408514e-01 -7.60863900e-01 2.64540887e+00
6.90068781e-01 -4.62539017e-01 -1.43176258e-01 -1.31196320e-01
7.08645999e-01 -4.73399937e-01 -5.30587137e-01 -5.34671366e-01
3.77393484e-01 2.22431272e-01 2.88216978e-01 4.81597669e-02
-4.50294256e-01 4.29476023e-01 4.64826775e+00 2.30486289e-01
-8.86851728e-01 1.95572764e-01 -5.36893532e-02 -6.46873936e-02
-9.52549875e-02 4.29415971e-01 -8.18934977e-01 5.60662806e-01
1.38747334e+00 -3.97424996e-01 2.54512370e-01 6.35595977e-01
9.30660725e-01 -4.88411307e-01 -1.30392647e+00 3.33903342e-01
-5.75276554e-01 -1.51813471e+00 -8.33559260e-02 2.72991341e-02
3.39917988e-02 -2.07007572e-01 -8.85582626e-01 -2.61790931e-01
3.13378960e-01 -5.57004690e-01 6.89819157e-01 9.97103035e-01
8.82569671e-01 -5.48549712e-01 8.23476195e-01 2.77679056e-01
-1.20351100e+00 3.72407645e-01 -2.55293250e-01 1.68677151e-01
1.11683600e-01 1.10164058e+00 -1.04312515e+00 1.37226892e+00
1.11560643e+00 8.03829804e-02 -8.70602787e-01 1.27988577e+00
6.38562888e-02 3.75850111e-01 -2.42211699e-01 1.61374152e-01
-5.96358955e-01 -4.92816925e-01 2.13911250e-01 1.53137016e+00
7.34814107e-01 -2.76724339e-01 -1.11255655e-02 7.76876986e-01
4.34821576e-01 3.95447642e-01 -4.24046010e-01 -1.81072056e-01
7.02612519e-01 1.88233757e+00 -9.50718045e-01 -4.76432353e-01
-1.67100683e-01 5.47528148e-01 -3.33112366e-02 2.95965493e-01
-6.43971086e-01 -5.66243291e-01 8.34618211e-01 2.70295769e-01
-1.65483952e-01 -1.93376049e-01 -2.17295125e-01 -8.10038984e-01
4.98728193e-02 -5.68290770e-01 5.82143009e-01 -9.57102716e-01
-1.06624293e+00 4.67831969e-01 -6.43874854e-02 -1.08008659e+00
2.39369646e-01 -2.08044693e-01 -6.46570206e-01 1.30446088e+00
-1.05277085e+00 -1.05061603e+00 -1.02690184e+00 6.33777142e-01
-9.23514888e-02 2.87657112e-01 1.23728383e+00 1.97793841e-01
-8.91856134e-01 -9.53195468e-02 1.53142154e-01 -7.17116058e-01
1.03986764e+00 -1.27244854e+00 2.09170654e-01 1.07145119e+00
-3.74739826e-01 1.11514652e+00 6.10635936e-01 -7.35010207e-01
-1.93462706e+00 -1.20361793e+00 8.45476151e-01 -1.42996773e-01
3.82605404e-01 -5.88102460e-01 -1.02098060e+00 4.55184639e-01
3.51021647e-01 -9.24769342e-02 1.08087420e+00 1.81627795e-01
4.64299656e-02 -4.45562601e-01 -1.33801222e+00 6.41466320e-01
1.03170133e+00 7.85971209e-02 2.99751103e-01 4.78731751e-01
3.09532285e-01 -1.32694542e-01 -1.35278869e+00 1.56220049e-01
4.79490459e-01 -7.06261992e-01 5.50499260e-01 -6.90683544e-01
-1.10948700e-02 -1.12829435e+00 3.83908719e-01 -1.27892244e+00
-2.14191362e-01 -7.63731301e-01 1.31630033e-01 1.64739823e+00
1.22203678e-01 -5.87392688e-01 2.17881193e-03 8.35743368e-01
-3.51886988e-01 -1.69775277e-01 -3.99960637e-01 -7.80142188e-01
-8.40940773e-01 -1.79235458e-01 1.10313559e+00 9.54267204e-01
1.44322351e-01 -2.96209045e-02 2.84763336e-01 5.41830063e-01
4.10091937e-01 3.19032341e-01 1.19036758e+00 -1.43148243e+00
-1.50806487e-01 -1.13608904e-01 -1.33471191e-01 2.73900390e-01
-6.07704639e-01 -1.09958804e+00 -1.66077837e-01 -2.06636977e+00
9.93283316e-02 -6.39284432e-01 -6.49768561e-02 7.43714988e-01
1.96741857e-02 -4.35684234e-01 3.87692690e-01 5.87756932e-01
-4.25724864e-01 -5.36609590e-01 8.71026874e-01 1.93498954e-01
-3.21236998e-01 -1.51255652e-02 -3.91512662e-01 4.60736513e-01
9.28592980e-01 -7.13547051e-01 -1.29382819e-01 -2.92266279e-01
2.57340282e-01 -2.59942282e-02 3.45255554e-01 -9.60468471e-01
7.35376477e-01 -5.77115953e-01 4.19694424e-01 -6.93102956e-01
-3.88341308e-01 -1.12992489e+00 1.55751061e+00 6.05275214e-01
-9.15752053e-02 2.08048150e-01 2.11420864e-01 1.34083152e-01
3.06455165e-01 -5.96753538e-01 2.73061872e-01 -3.85696888e-01
-5.35736263e-01 1.88690394e-01 -6.01949155e-01 -7.41272509e-01
1.33137429e+00 -4.08248371e-03 -8.34316194e-01 7.39623845e-01
-7.94444263e-01 3.22960466e-01 8.01790655e-01 1.54189959e-01
6.74845055e-02 -4.58944947e-01 -3.05910558e-01 4.17515337e-01
2.96600163e-01 2.03314140e-01 2.82923222e-01 8.99757743e-01
-9.57409143e-01 1.54302463e-01 -6.47759736e-01 -5.42943418e-01
-1.67902148e+00 6.52887285e-01 -9.26303938e-02 -2.71216452e-01
-3.62796575e-01 1.48206145e-01 -4.13142860e-01 -4.63841230e-01
-5.40117128e-03 -2.69020975e-01 -1.17990889e-01 2.26255774e-01
7.86570907e-01 6.07989728e-01 7.69491792e-01 2.10246563e-01
-6.19429708e-01 1.14822239e-01 3.41773838e-01 1.41672373e-01
1.41435564e+00 1.12501159e-01 -7.24733829e-01 6.08317792e-01
2.36306652e-01 1.48156524e-01 -9.76362824e-01 3.23774666e-01
3.88877749e-01 -5.54376960e-01 -5.10881424e-01 -1.28654718e+00
-9.78492260e-01 6.43984318e-01 5.33156753e-01 3.65638703e-01
1.22835433e+00 -8.64810124e-02 -2.88308203e-01 2.39880890e-01
5.79618752e-01 -9.13414598e-01 -8.09991360e-01 -8.65428708e-03
9.04614627e-01 -6.93772197e-01 2.21271798e-01 -6.11031473e-01
-2.86520302e-01 1.57214046e+00 4.44304228e-01 9.25805390e-01
6.99418604e-01 6.02541387e-01 1.65032931e-02 -3.10113698e-01
-1.05706191e+00 -2.41861463e-01 -4.94310915e-01 6.62796080e-01
7.59892881e-01 3.04590017e-01 -9.11207795e-01 6.30838037e-01
4.12291616e-01 6.43103957e-01 7.17582643e-01 1.51536500e+00
-3.08399647e-01 -1.42626631e+00 -6.43869817e-01 6.78721070e-01
-2.26025566e-01 1.22254029e-01 -1.56727865e-01 5.55129886e-01
6.02059305e-01 9.12211657e-01 1.89173818e-01 -4.68942482e-04
6.96262240e-01 3.36506814e-01 4.92070094e-02 -7.48672366e-01
-1.32186913e+00 -8.07812735e-02 2.51173466e-01 -7.46209741e-01
-2.67885983e-01 -6.52896583e-01 -1.23751676e+00 -6.85022652e-01
-5.21920323e-02 3.84764373e-01 1.42731214e+00 5.73654413e-01
1.05879366e+00 9.26320791e-01 2.12713048e-01 -4.58946735e-01
-1.66790456e-01 -8.91661227e-01 -7.16225028e-01 2.85200030e-01
-6.64105833e-01 -1.75413013e-01 1.43094510e-01 5.65909505e-01] | [9.056949615478516, 7.744824409484863] |
5d19adb2-47b0-4d17-b279-3b58df93bbbd | surpassing-the-human-accuracy-detecting | 2204.11433 | null | https://arxiv.org/abs/2204.11433v1 | https://arxiv.org/pdf/2204.11433v1.pdf | Surpassing the Human Accuracy: Detecting Gallbladder Cancer from USG Images with Curriculum Learning | We explore the potential of CNN-based models for gallbladder cancer (GBC) detection from ultrasound (USG) images as no prior study is known. USG is the most common diagnostic modality for GB diseases due to its low cost and accessibility. However, USG images are challenging to analyze due to low image quality, noise, and varying viewpoints due to the handheld nature of the sensor. Our exhaustive study of state-of-the-art (SOTA) image classification techniques for the problem reveals that they often fail to learn the salient GB region due to the presence of shadows in the USG images. SOTA object detection techniques also achieve low accuracy because of spurious textures due to noise or adjacent organs. We propose GBCNet to tackle the challenges in our problem. GBCNet first extracts the regions of interest (ROIs) by detecting the GB (and not the cancer), and then uses a new multi-scale, second-order pooling architecture specializing in classifying GBC. To effectively handle spurious textures, we propose a curriculum inspired by human visual acuity, which reduces the texture biases in GBCNet. Experimental results demonstrate that GBCNet significantly outperforms SOTA CNN models, as well as the expert radiologists. Our technical innovations are generic to other USG image analysis tasks as well. Hence, as a validation, we also show the efficacy of GBCNet in detecting breast cancer from USG images. Project page with source code, trained models, and data is available at https://gbc-iitd.github.io/gbcnet | ['Chetan Arora', 'Pankaj Gupta', 'Pratyaksha Rana', 'Mayank Gupta', 'Soumen Basu'] | 2022-04-25 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Basu_Surpassing_the_Human_Accuracy_Detecting_Gallbladder_Cancer_From_USG_Images_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Basu_Surpassing_the_Human_Accuracy_Detecting_Gallbladder_Cancer_From_USG_Images_CVPR_2022_paper.pdf | cvpr-2022-1 | ['gallbladder-cancer-detection'] | ['computer-vision'] | [ 4.16057296e-02 2.49527097e-01 -4.77463864e-02 1.54626325e-01
-8.80824387e-01 -1.64850265e-01 1.58994570e-01 6.90338686e-02
-1.98771372e-01 1.27124965e-01 1.38605744e-01 -6.11662805e-01
1.55077085e-01 -7.20221579e-01 -8.29934537e-01 -1.02655387e+00
-3.60967070e-01 -4.99339662e-02 4.37430561e-01 -2.58664228e-03
-7.57614076e-02 3.33068192e-01 -1.02625835e+00 5.49758673e-01
7.75842607e-01 1.35377300e+00 7.03923762e-01 9.01317477e-01
3.28715257e-02 7.54415512e-01 -3.66061509e-01 1.15000270e-01
3.36640805e-01 -5.82444727e-01 -5.88519394e-01 -1.38908267e-01
5.19662559e-01 -5.34138799e-01 -3.77742320e-01 1.18307281e+00
7.06229866e-01 -2.57191002e-01 4.12610561e-01 -6.71829343e-01
-6.41643524e-01 2.95338452e-01 -4.20145422e-01 5.70941985e-01
-3.48600443e-04 1.23787589e-01 4.09027278e-01 -7.28226542e-01
5.76594353e-01 9.87753689e-01 8.15271497e-01 8.86324406e-01
-8.04387629e-01 -6.89732254e-01 9.90942493e-03 4.80079427e-02
-1.00831330e+00 -1.43693537e-01 3.67938042e-01 -2.53544718e-01
6.42096341e-01 5.99629104e-01 1.09388959e+00 1.17950320e+00
7.55261898e-01 1.09399712e+00 1.22482026e+00 -5.71197927e-01
1.34393349e-01 -1.29476935e-01 1.64185688e-02 1.28873408e+00
5.31149447e-01 1.81278035e-01 -1.13678992e-01 -8.41522589e-02
1.32577443e+00 2.76298225e-01 -7.08732784e-01 -5.25651574e-01
-1.18601620e+00 7.55361617e-01 8.32854390e-01 5.00178397e-01
-3.14816654e-01 3.85135800e-01 2.01439351e-01 1.31707340e-01
3.18048835e-01 3.77614528e-01 -3.71503741e-01 3.98390025e-01
-3.62681270e-01 -1.73179910e-01 8.29308271e-01 1.05141234e+00
2.92754054e-01 -2.82276839e-01 -1.88032061e-01 6.29698873e-01
3.19146305e-01 4.96410459e-01 8.19361746e-01 -5.06511211e-01
-4.66646627e-02 5.50654054e-01 -2.81617820e-01 -9.99761462e-01
-6.85228586e-01 -6.95396185e-01 -1.07400334e+00 3.82213816e-02
6.76345587e-01 1.21126205e-01 -1.38139617e+00 1.19736564e+00
1.14714587e-02 3.81169409e-01 -1.83421031e-01 1.30429840e+00
1.58142161e+00 1.04927957e-01 2.92905688e-01 3.10101449e-01
1.88509595e+00 -9.27724183e-01 -4.43998635e-01 -3.40566635e-01
8.21167469e-01 -5.51093638e-01 9.68675554e-01 1.90526411e-01
-7.41367579e-01 -2.53807664e-01 -9.26738560e-01 -7.36736730e-02
-2.32985452e-01 4.04836446e-01 7.88420200e-01 8.30713511e-01
-1.10586786e+00 1.90058663e-01 -1.53201187e+00 -6.83108568e-01
6.41129673e-01 1.87915593e-01 -2.11218625e-01 -3.81298393e-01
-8.69007587e-01 9.15377855e-01 1.65599853e-01 2.59555042e-01
-7.98356116e-01 -5.95970154e-01 -1.09401798e+00 2.12021526e-02
5.48175752e-01 -6.32042587e-01 1.17706203e+00 -8.20751250e-01
-1.31288517e+00 9.48794723e-01 1.61722869e-01 -4.71270055e-01
5.35547614e-01 1.68983549e-01 -1.37419775e-01 5.41212738e-01
2.98142675e-02 5.59332788e-01 7.17086017e-01 -1.06958616e+00
-6.51688993e-01 -2.94259757e-01 -6.14955649e-02 -1.45263284e-01
4.81437966e-02 -2.83720016e-01 -7.79678404e-01 -8.25050950e-01
6.15036607e-01 -9.87904847e-01 -5.29314280e-01 2.02493578e-01
-2.48497784e-01 1.90231483e-02 5.90390742e-01 -7.58197427e-01
9.16342735e-01 -2.17013454e+00 -3.49738479e-01 1.10376768e-01
4.74956632e-01 2.23247960e-01 -1.93041027e-01 -4.36402112e-01
2.22086117e-01 2.21820772e-01 1.13957986e-01 -4.29145657e-02
-3.61223608e-01 1.87973693e-01 3.14839005e-01 5.96268475e-01
1.28103226e-01 1.22401977e+00 -1.15858579e+00 -6.74949884e-01
5.12051225e-01 3.11725348e-01 -4.13332254e-01 3.22234742e-02
5.44768870e-02 5.94331980e-01 -7.94417083e-01 1.18469810e+00
7.25458145e-01 -7.05786586e-01 3.28390837e-01 -4.69558865e-01
8.15325007e-02 1.23349026e-01 -8.13913822e-01 1.77195191e+00
-4.99389589e-01 5.36876261e-01 2.89564997e-01 -1.15012419e+00
5.65382242e-01 4.14263695e-01 2.39593655e-01 -1.07204783e+00
3.19586575e-01 6.93801343e-01 2.18189135e-01 -9.08331633e-01
-9.03434381e-02 -7.90228974e-03 1.63069963e-01 1.08953908e-01
3.09132367e-01 -1.12471424e-01 -6.97862240e-04 -7.28600472e-02
1.03324628e+00 -9.48440358e-02 6.76054716e-01 -7.48561442e-01
1.64167285e-01 8.41638297e-02 5.87550998e-01 1.36757278e+00
-5.31540811e-01 7.95814157e-01 3.72049719e-01 -9.01815236e-01
-5.90840816e-01 -8.02652001e-01 -4.06390399e-01 5.82952797e-01
5.28357327e-01 1.78563565e-01 -6.51696742e-01 -8.71179342e-01
7.36052319e-02 -8.58498737e-02 -1.15323842e+00 1.85834661e-01
-8.73957515e-01 -9.81845558e-01 3.57374728e-01 4.79475081e-01
6.07756734e-01 -1.05310035e+00 -1.18912423e+00 3.50694686e-01
-3.32852721e-01 -1.01898491e+00 -2.72707373e-01 2.10789829e-01
-9.69019473e-01 -1.61398113e+00 -1.26333094e+00 -1.19879794e+00
7.84381568e-01 4.40198570e-01 1.24601912e+00 6.01057768e-01
-8.81389976e-01 3.91800433e-01 -2.82396823e-01 -6.76280856e-01
-4.00036871e-01 8.20966531e-03 -4.04133946e-01 -3.74998361e-01
4.83043104e-01 1.67505547e-01 -1.18967307e+00 2.18210906e-01
-7.49304414e-01 2.23289981e-01 9.48843062e-01 1.05800891e+00
4.84364808e-01 -4.34309363e-01 2.65178800e-01 -8.76039684e-01
1.05118886e-01 -3.01028311e-01 -4.87878323e-01 2.59916753e-01
-8.55757520e-02 -4.56946969e-01 1.64728269e-01 -4.65855837e-01
-7.10572660e-01 -2.46793345e-01 5.05825803e-02 -3.92815411e-01
-3.19994569e-01 4.53945279e-01 5.51814675e-01 -5.33261895e-01
6.03655279e-01 3.25863570e-01 2.22086191e-01 -3.02068412e-01
-3.56258363e-01 4.54247147e-01 3.63561153e-01 -2.47156709e-01
6.15869574e-02 6.80714011e-01 -1.02876328e-01 -8.34788978e-01
-6.70713425e-01 -6.31167233e-01 -1.00818329e-01 -1.79604784e-01
8.41796994e-01 -1.10373354e+00 -5.81113994e-01 4.23639923e-01
-8.40507030e-01 -6.14977360e-01 9.57345068e-02 8.65344942e-01
-1.94879726e-01 2.27320001e-01 -1.25799084e+00 -5.29107511e-01
-5.84567308e-01 -1.49747801e+00 1.00531328e+00 4.10429269e-01
2.59850293e-01 -1.02338445e+00 -5.60438335e-01 4.80312705e-02
7.64039099e-01 4.70956475e-01 8.56025457e-01 -3.75454366e-01
-1.01999354e+00 -1.92125753e-01 -5.70198774e-01 8.90697315e-02
3.16449642e-01 -5.60743392e-01 -8.58959079e-01 -4.83013421e-01
9.80039537e-02 -5.49128503e-02 1.23746467e+00 1.19798684e+00
1.35161376e+00 -6.86593577e-02 -5.96333623e-01 9.96902525e-01
1.49203467e+00 3.74259323e-01 5.10229707e-01 5.97781122e-01
4.61413682e-01 3.18286628e-01 3.65550190e-01 -5.01563735e-02
6.14500120e-02 2.54920274e-01 6.35653913e-01 -8.10204685e-01
-5.06688714e-01 1.79825425e-01 -3.26342732e-01 5.61452448e-01
-2.99567729e-01 -1.05344683e-01 -9.43142831e-01 7.28111923e-01
-1.50750053e+00 -3.52638274e-01 -1.21800795e-01 1.75162697e+00
5.53028345e-01 -2.68406034e-01 -3.13087225e-01 -4.75183338e-01
3.68281484e-01 -1.87907636e-01 -1.42961293e-01 1.70212969e-01
9.30001494e-03 3.15554082e-01 9.71773148e-01 8.29155073e-02
-1.53274739e+00 5.14670670e-01 5.76801634e+00 7.51150131e-01
-1.49057698e+00 2.13271052e-01 1.07877409e+00 1.86915398e-01
1.12413816e-01 -5.79372644e-01 -3.90202373e-01 4.30942655e-01
-2.43751686e-02 5.60812950e-01 5.52845076e-02 7.43834138e-01
-2.66148001e-02 -3.95802230e-01 -8.06063175e-01 9.52621818e-01
2.58148491e-01 -1.45522296e+00 -1.59448057e-01 -3.08855809e-02
6.85190856e-01 2.17883646e-01 4.08324823e-02 1.84746683e-01
-2.03972273e-02 -9.96229827e-01 4.76570338e-01 2.18565881e-01
9.63744164e-01 -1.22138232e-01 1.29060650e+00 2.47682389e-02
-9.72757936e-01 -7.66405985e-02 -3.92990887e-01 2.57658869e-01
-5.12442172e-01 4.32248086e-01 -1.12143540e+00 3.32201928e-01
1.24401498e+00 4.58436906e-01 -6.92505777e-01 1.32595456e+00
-7.78780803e-02 5.75597346e-01 -2.99100131e-01 -2.99769789e-01
4.58148926e-01 1.30613580e-01 4.47845072e-01 1.57205868e+00
7.29391396e-01 3.91858131e-01 2.24361792e-01 8.51505637e-01
1.00802384e-01 1.72108322e-01 -3.74034494e-01 3.77657861e-01
4.41383682e-02 1.28830361e+00 -1.24399269e+00 -4.10788804e-01
-6.02184892e-01 6.87662959e-01 -1.86886162e-01 5.02805114e-01
-2.85267025e-01 -1.29711747e-01 9.76112634e-02 1.20115377e-01
3.13566923e-01 2.18675837e-01 -3.78668457e-01 -1.33612537e+00
-1.93295360e-01 -8.27753484e-01 6.10915244e-01 -4.01238024e-01
-1.03756893e+00 5.79595983e-01 -3.77484620e-01 -1.56390965e+00
2.33297348e-01 -1.04090929e+00 -3.55665594e-01 6.32235229e-01
-1.84388530e+00 -1.30175066e+00 -9.37410831e-01 3.88171792e-01
5.44163048e-01 1.91548597e-02 9.86764848e-01 2.30151832e-01
-2.63837874e-01 4.45003599e-01 -4.80901003e-02 5.60734987e-01
5.74258924e-01 -1.42355871e+00 1.59197479e-01 6.72972918e-01
-4.60797369e-01 6.75442636e-01 2.80465662e-01 -6.33040667e-01
-1.50860119e+00 -9.51006174e-01 4.18054670e-01 -1.93056583e-01
4.15160239e-01 -3.57641548e-01 -7.38727272e-01 7.22256899e-01
-1.40378267e-01 7.76216745e-01 5.27046502e-01 -3.18821520e-01
1.38868447e-02 2.62490392e-01 -1.19279730e+00 4.42380548e-01
6.76917791e-01 -5.04779592e-02 -2.54270732e-01 1.47831351e-01
3.16567183e-01 -1.21988630e+00 -6.51615858e-01 5.52116156e-01
6.89590096e-01 -9.47017133e-01 1.03063667e+00 -2.31445551e-01
4.46977258e-01 -7.10275397e-02 1.05084695e-01 -1.18411922e+00
-3.51816833e-01 -1.50633737e-01 6.24568202e-02 6.03440702e-02
1.10858187e-01 -8.20182800e-01 9.75653768e-01 1.42478555e-01
-4.37412381e-01 -7.56465852e-01 -9.41751778e-01 -4.42382812e-01
-5.42863645e-02 -2.39809349e-01 3.10622364e-01 1.04219162e+00
-7.46022835e-02 -5.56342661e-01 -3.46748494e-02 3.53667140e-01
8.39314103e-01 3.70339155e-01 4.36836541e-01 -9.14582312e-01
-3.94961655e-01 -5.02820492e-01 -3.93149197e-01 -9.42749202e-01
-7.72223055e-01 -7.59346962e-01 2.54966795e-01 -1.66626012e+00
3.44777733e-01 -7.02477396e-01 -4.89223927e-01 4.55063850e-01
-4.08207566e-01 8.18424106e-01 1.92283779e-01 1.74853355e-01
-4.63702589e-01 -1.61060989e-01 1.90909195e+00 -2.84899026e-01
-1.11058004e-01 4.93869670e-02 -4.88062710e-01 7.51546800e-01
6.24217570e-01 -2.76193082e-01 2.70197749e-01 -4.31400597e-01
-2.85978228e-01 3.46235871e-01 7.88932145e-01 -8.59125078e-01
2.03593254e-01 1.38385803e-01 8.21995676e-01 -3.69404584e-01
-7.50845894e-02 -6.83408201e-01 -1.31242782e-01 1.11801004e+00
1.58777803e-01 -6.94357395e-01 3.47766161e-01 4.78842825e-01
-1.89962283e-01 -6.48161843e-02 9.21728730e-01 -8.71200323e-01
-7.63380289e-01 1.63564846e-01 -5.69033384e-01 -3.58904064e-01
6.14446282e-01 -2.51415044e-01 -2.92115450e-01 -1.94870651e-01
-7.97833622e-01 1.68694630e-01 1.65490732e-01 1.98115304e-01
7.47708976e-01 -9.52908814e-01 -7.53650308e-01 6.59222186e-01
2.80474603e-01 4.26314890e-01 5.47495663e-01 1.49232733e+00
-1.05498111e+00 5.64518034e-01 -8.79611820e-02 -9.35052752e-01
-1.09150648e+00 2.36870691e-01 8.95343065e-01 -3.91429186e-01
-1.14047873e+00 9.53381300e-01 7.75769472e-01 -3.97269428e-01
3.27363908e-01 -1.08894420e+00 -1.05481647e-01 -4.22098696e-01
6.17147923e-01 1.05066567e-04 4.76113051e-01 6.05374165e-02
-2.83256978e-01 4.84479010e-01 -1.94257230e-01 7.54630804e-01
1.03618443e+00 -2.13527530e-02 2.13701800e-01 -2.59540439e-01
1.03835618e+00 -3.76818627e-01 -1.00936973e+00 -3.34004939e-01
-1.30891770e-01 -6.01975441e-01 3.01774353e-01 -8.68623734e-01
-1.36112547e+00 6.49137795e-01 1.06841338e+00 1.91450402e-01
1.17379355e+00 1.54078126e-01 7.79315174e-01 6.78324923e-02
5.24035931e-01 -5.32667041e-01 1.36825472e-01 3.37284625e-01
7.87159383e-01 -1.55088210e+00 -2.13254854e-01 -6.11994267e-01
-2.57051736e-01 1.37702870e+00 5.68163693e-01 -4.06685889e-01
7.43512332e-01 4.16237980e-01 6.50791407e-01 -4.19452220e-01
-1.55231789e-01 -1.82518542e-01 2.90501684e-01 6.34281397e-01
4.16172355e-01 2.82728553e-01 -3.14115196e-01 6.14498973e-01
1.37622446e-01 1.38778776e-01 5.18734753e-01 9.71879303e-01
-2.77885497e-01 -4.36124355e-01 -5.22073448e-01 7.97693849e-01
-9.70584929e-01 -1.62753537e-01 3.10139567e-01 1.13679874e+00
1.20123446e-01 6.12972260e-01 2.17328280e-01 2.05054596e-01
2.35076860e-01 -5.98171651e-01 5.66643834e-01 -5.64213693e-01
-6.24838769e-01 7.95061529e-01 -2.19019696e-01 -7.67283738e-01
-3.52567524e-01 -3.81606668e-01 -1.02928972e+00 2.23600999e-01
-3.86267722e-01 -2.01751664e-01 5.81242025e-01 6.56331360e-01
1.42655345e-02 7.22202837e-01 1.14568658e-01 -9.66265321e-01
-2.43100718e-01 -1.25908792e+00 -7.45305419e-01 4.98270035e-01
6.90864086e-01 -5.26841938e-01 -3.44747484e-01 -1.48295447e-01] | [15.026774406433105, -2.455826759338379] |
df9c6ef9-a34d-4906-857f-0349ff9a4cf9 | em-fusion-dynamic-object-level-slam-with | 1904.11781 | null | https://arxiv.org/abs/1904.11781v2 | https://arxiv.org/pdf/1904.11781v2.pdf | EM-Fusion: Dynamic Object-Level SLAM with Probabilistic Data Association | The majority of approaches for acquiring dense 3D environment maps with RGB-D cameras assumes static environments or rejects moving objects as outliers. The representation and tracking of moving objects, however, has significant potential for applications in robotics or augmented reality. In this paper, we propose a novel approach to dynamic SLAM with dense object-level representations. We represent rigid objects in local volumetric signed distance function (SDF) maps, and formulate multi-object tracking as direct alignment of RGB-D images with the SDF representations. Our main novelty is a probabilistic formulation which naturally leads to strategies for data association and occlusion handling. We analyze our approach in experiments and demonstrate that our approach compares favorably with the state-of-the-art methods in terms of robustness and accuracy. | ['Jörg Stückler', 'Michael Strecke'] | 2019-04-26 | em-fusion-dynamic-object-level-slam-with-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Strecke_EM-Fusion_Dynamic_Object-Level_SLAM_With_Probabilistic_Data_Association_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Strecke_EM-Fusion_Dynamic_Object-Level_SLAM_With_Probabilistic_Data_Association_ICCV_2019_paper.pdf | iccv-2019-10 | ['occlusion-handling'] | ['computer-vision'] | [-4.30481136e-03 -2.97868162e-01 -1.56065179e-02 -4.96194661e-01
-6.25415623e-01 -6.85096145e-01 6.58546627e-01 -1.35735320e-02
-4.57068413e-01 5.69713950e-01 -2.21438661e-01 -4.86855209e-03
-3.49821597e-01 -4.38554555e-01 -8.99577975e-01 -4.61082667e-01
-1.72157317e-01 1.15760481e+00 7.16445029e-01 4.76359278e-02
2.60606408e-01 1.19312894e+00 -1.61568856e+00 -3.01935524e-01
5.52172601e-01 9.14966822e-01 1.47199452e-01 6.84565783e-01
-5.18555567e-02 6.00916207e-01 -2.68043160e-01 -7.08336160e-02
5.93207836e-01 1.30789587e-02 -4.98185635e-01 2.86129653e-01
8.62792552e-01 -2.80240208e-01 -4.96334612e-01 9.79029298e-01
4.40097868e-01 2.06123024e-01 4.45809215e-01 -1.46209109e+00
-3.29347312e-01 -2.07497120e-01 -7.31406748e-01 3.40911895e-02
9.25569236e-01 -1.80437312e-01 3.23777169e-01 -1.10805047e+00
9.65019524e-01 1.43738663e+00 1.00660801e+00 3.75141352e-01
-1.34966636e+00 -2.07826614e-01 1.81061342e-01 -7.42607284e-03
-1.64194524e+00 -5.24950683e-01 7.23092318e-01 -4.87555802e-01
8.18323314e-01 3.64833385e-01 7.49264181e-01 7.28192806e-01
4.81499612e-01 7.71013796e-01 1.17359722e+00 -4.72393930e-01
3.06499898e-01 -1.20056547e-01 8.09810609e-02 6.01324737e-01
7.77356923e-01 3.39745015e-01 -8.01530302e-01 -3.88095766e-01
9.76641774e-01 1.60092965e-01 -7.96861574e-02 -1.60784996e+00
-1.46041310e+00 5.44405937e-01 4.69581664e-01 -9.66465920e-02
-3.74587566e-01 7.10308790e-01 1.47138676e-02 -1.21407792e-01
4.66581076e-01 -1.43803880e-01 -3.31395656e-01 3.48090641e-02
-8.86156261e-01 5.02932012e-01 6.11281872e-01 1.51205146e+00
8.06942701e-01 -7.73487985e-02 5.06394804e-01 1.52108774e-01
9.29481983e-01 9.46324944e-01 1.22245438e-01 -1.13657773e+00
1.06225692e-01 2.70240664e-01 4.93867069e-01 -9.97677684e-01
-4.58559841e-01 9.17703882e-02 -4.05352980e-01 6.50321186e-01
2.20591292e-01 4.47695643e-01 -9.80166376e-01 1.10928440e+00
7.84448743e-01 3.67399544e-01 1.25901595e-01 9.45814133e-01
4.88386959e-01 6.35139942e-02 -4.15700585e-01 -1.51856869e-01
7.79721260e-01 -4.12055075e-01 -8.13878357e-01 -4.20046717e-01
1.41320243e-01 -8.08284640e-01 2.59658307e-01 2.36294836e-01
-9.60200012e-01 -4.08333957e-01 -8.37967634e-01 -1.87479943e-01
-9.05612484e-02 -2.48335227e-01 7.57078767e-01 5.42894065e-01
-1.13467717e+00 5.37533641e-01 -1.34646583e+00 -6.51100576e-01
2.25610957e-01 8.10924351e-01 -5.59997082e-01 -1.98259696e-01
-3.66548210e-01 1.32452369e+00 3.56428862e-01 1.63793713e-01
-6.60529613e-01 -3.30722332e-01 -1.21707416e+00 -8.41196418e-01
1.39426157e-01 -8.12365174e-01 1.11532319e+00 -3.60399514e-01
-1.47844231e+00 1.09937656e+00 -4.76423144e-01 -3.32264453e-01
9.13563490e-01 -5.67924380e-01 -7.09596574e-02 -7.16348970e-03
-1.49741918e-02 5.41819751e-01 6.56695724e-01 -1.84330964e+00
-5.69964528e-01 -7.75369883e-01 -1.53325528e-01 3.25043052e-01
5.37110806e-01 -8.86383578e-02 -4.10945445e-01 -2.65568227e-01
1.23209059e+00 -1.29139256e+00 -5.00555336e-01 5.45145333e-01
-1.18072435e-01 2.11470887e-01 9.22401547e-01 -1.86195999e-01
3.82567257e-01 -1.98595917e+00 4.25395221e-01 4.58707452e-01
-1.59513019e-02 -4.57073420e-01 2.91758835e-01 -3.92169021e-02
3.47698629e-01 -3.70799184e-01 -2.14996338e-02 -8.63123894e-01
9.36983526e-02 6.79577351e-01 -4.25292462e-01 1.47369599e+00
-1.51037857e-01 7.80468941e-01 -1.07097363e+00 -5.34485042e-01
7.39552379e-01 7.10784972e-01 -1.05766200e-01 4.67195511e-02
-9.99658369e-03 5.55222571e-01 -4.29279476e-01 1.05473650e+00
1.00028813e+00 2.44010031e-01 -1.85507573e-02 -1.32368147e-01
-3.28384876e-01 1.68570772e-01 -1.70476270e+00 2.27238464e+00
8.38169456e-03 4.74713326e-01 5.88361323e-02 -2.66492814e-01
9.58148122e-01 2.37261951e-02 8.36056113e-01 -3.28443199e-01
9.00179520e-02 4.52276409e-01 -4.83714014e-01 8.98307469e-03
9.46227729e-01 -8.63395780e-02 -2.15062872e-01 1.01593785e-01
-1.30347773e-01 -7.68688440e-01 -4.21146631e-01 1.19974211e-01
1.00965250e+00 8.81471217e-01 4.87832069e-01 -1.61344811e-01
2.18616754e-01 3.84091347e-01 5.63150704e-01 7.10009396e-01
-4.32432115e-01 7.69612074e-01 -4.36935723e-01 -5.26547909e-01
-8.98753166e-01 -1.53112328e+00 -3.84824306e-01 3.44671726e-01
8.78663480e-01 -4.58861664e-02 6.06996678e-02 -4.10836786e-01
7.48191178e-01 3.71025920e-01 -4.68941182e-01 1.56717181e-01
-8.45591068e-01 -2.68566340e-01 5.18606901e-02 6.30090415e-01
-5.67414500e-02 -5.97997785e-01 -1.10992682e+00 1.80481866e-01
2.16252759e-01 -1.20025706e+00 1.69024933e-02 3.74619782e-01
-1.28915226e+00 -9.68141019e-01 -4.64850694e-01 -3.85706663e-01
8.31992507e-01 8.06427598e-01 9.38599944e-01 -2.57035196e-01
-3.00413489e-01 9.80147898e-01 -3.48009944e-01 -4.73812699e-01
-1.90326348e-01 -4.82223511e-01 7.63546228e-01 -3.46925527e-01
3.76880586e-01 -4.07357544e-01 -2.95607328e-01 4.96244460e-01
-5.50370574e-01 -2.24052802e-01 3.01184952e-01 1.77790090e-01
1.19170451e+00 -4.36743051e-01 -4.02702570e-01 -4.40470099e-01
-2.23357603e-01 -2.57681698e-01 -1.07024527e+00 6.84775561e-02
-3.20631683e-01 -3.45371850e-02 -3.78195018e-01 -4.34961736e-01
-6.55148506e-01 9.14196968e-01 4.21872914e-01 -1.08913398e+00
-1.04821712e-01 -4.78349719e-03 -8.14082325e-02 -8.03703547e-01
4.99985129e-01 -1.06411306e-02 -2.44877953e-03 -4.74688232e-01
5.95100284e-01 2.26183504e-01 7.46281862e-01 -6.44801676e-01
1.12321448e+00 1.08031940e+00 3.64391387e-01 -5.66020966e-01
-5.24902761e-01 -9.90832031e-01 -1.28329670e+00 -4.13670689e-01
5.76738000e-01 -1.00764251e+00 -6.83373988e-01 3.28670979e-01
-1.38473558e+00 -6.17519356e-02 -6.43808663e-01 7.73185492e-01
-1.11830497e+00 3.60568196e-01 -2.14253753e-01 -1.17061579e+00
2.64226884e-01 -1.10322368e+00 1.53835154e+00 -3.20082568e-02
-1.75015166e-01 -7.85279095e-01 5.02466381e-01 -1.21961355e-01
-1.69520881e-02 7.69049227e-01 -2.23920122e-01 -2.27961853e-01
-1.11682904e+00 -3.63202780e-01 1.13580182e-01 -3.00895452e-01
1.14413694e-01 -3.91999818e-02 -1.00497448e+00 -4.69327807e-01
6.64901361e-02 -1.65263638e-02 5.16726315e-01 4.77826416e-01
5.78854144e-01 2.51176983e-01 -7.45687544e-01 7.27675200e-01
1.70866013e+00 -4.61339951e-02 5.99638343e-01 7.49009013e-01
8.34638536e-01 5.32966137e-01 1.07089686e+00 4.34531271e-01
4.92098421e-01 8.86740327e-01 8.43296170e-01 2.93332875e-01
-4.89694707e-04 5.96947223e-02 3.19281578e-01 6.43292308e-01
-1.80824652e-01 8.10161158e-02 -1.18090248e+00 5.41815341e-01
-2.24891782e+00 -6.27800167e-01 -5.94484925e-01 2.38057804e+00
3.04959208e-01 5.88496067e-02 -4.37843315e-02 1.44811664e-02
6.35457158e-01 5.68500254e-03 -5.92322528e-01 1.89454660e-01
-2.34685272e-01 1.67583004e-02 1.02015984e+00 7.22444057e-01
-1.09348917e+00 8.86066973e-01 6.86665630e+00 -7.54526928e-02
-6.12611830e-01 2.90120482e-01 -4.32768285e-01 -1.91168442e-01
-7.10191205e-02 2.45096907e-01 -1.11895883e+00 2.30642986e-02
7.63331234e-01 -9.16585028e-02 3.91461775e-02 9.43661153e-01
-1.94290146e-01 -4.43717062e-01 -1.34080923e+00 1.11489081e+00
3.47292900e-01 -1.08979833e+00 -2.82600760e-01 3.89667511e-01
7.18054593e-01 4.50969696e-01 -2.43083909e-01 -2.93355167e-01
6.64323688e-01 -5.30451238e-01 1.42494106e+00 7.83205926e-01
4.49431270e-01 -4.86090720e-01 6.60381734e-01 2.28082180e-01
-1.34895945e+00 3.69633377e-01 -5.29228687e-01 3.30988467e-02
5.17394841e-01 5.51357985e-01 -6.91273808e-01 9.49467123e-01
8.24523211e-01 7.59458542e-01 -4.83636975e-01 1.46540320e+00
1.58974510e-02 -3.93299788e-01 -8.06768715e-01 2.31264785e-01
-1.64219961e-01 -1.80586845e-01 7.25278556e-01 8.65350127e-01
4.94021982e-01 5.83340488e-02 5.62846720e-01 4.71771687e-01
3.14446777e-01 -2.91514933e-01 -9.60457921e-01 7.14106321e-01
4.80685830e-01 8.12750280e-01 -1.05861270e+00 -2.19808400e-01
-2.90454447e-01 1.12553465e+00 2.05257669e-01 -1.01593193e-02
-6.77174568e-01 2.00375378e-01 7.31442511e-01 1.31891653e-01
3.55924398e-01 -9.20782328e-01 -3.39900404e-01 -1.29340756e+00
1.79076657e-01 -2.72982210e-01 9.60254520e-02 -9.60954547e-01
-1.11564088e+00 4.56176460e-01 2.23367542e-01 -1.67978287e+00
-1.72500238e-01 -7.53301084e-01 1.99276850e-01 6.36454999e-01
-1.51658654e+00 -1.37781513e+00 -6.70727849e-01 7.43980467e-01
2.79153168e-01 3.94375831e-01 6.90086365e-01 1.19752496e-01
1.72178656e-01 -1.72051519e-01 3.56880277e-01 -2.87971556e-01
6.27974868e-01 -1.31310320e+00 5.20333469e-01 9.85511780e-01
4.15727139e-01 6.13601625e-01 1.04727852e+00 -1.02182615e+00
-2.01600814e+00 -9.28630173e-01 4.78190243e-01 -1.08721030e+00
5.78570902e-01 -5.31572223e-01 -6.51659489e-01 1.20636821e+00
-4.31686848e-01 7.13031769e-01 2.76210725e-01 -1.40855789e-01
-1.40928090e-01 9.78684872e-02 -1.32895851e+00 5.85331991e-02
1.31521928e+00 -4.81981307e-01 -8.17605674e-01 5.38901508e-01
5.70455611e-01 -1.26103592e+00 -7.71477520e-01 6.13045454e-01
6.69156730e-01 -8.44206274e-01 1.36448693e+00 -1.68165550e-01
-6.68468595e-01 -8.76484454e-01 -7.78419316e-01 -9.17382002e-01
-1.56479523e-01 -5.78548193e-01 -6.55702651e-01 8.82781088e-01
-4.68494475e-01 -4.62234855e-01 9.62338448e-01 7.22339928e-01
-2.42552355e-01 5.07637113e-03 -1.38154221e+00 -1.13727283e+00
-4.59274441e-01 -6.58575356e-01 2.34548897e-01 8.28400016e-01
-5.91454744e-01 -4.54474956e-01 -2.56305218e-01 7.47104585e-01
1.29245830e+00 3.12767297e-01 1.23293221e+00 -1.66236734e+00
1.73856243e-01 1.21777076e-02 -1.13219190e+00 -1.03173542e+00
2.80826509e-01 -5.49097657e-01 5.36403894e-01 -1.62748992e+00
-2.01428711e-01 -8.31177950e-01 -7.09856860e-03 1.08654268e-01
3.11327219e-01 3.65534246e-01 2.36372396e-01 5.31062067e-01
-9.10153985e-01 5.15220821e-01 5.94779849e-01 7.38738328e-02
-1.96588099e-01 -1.02011506e-02 1.57354310e-01 9.62292850e-01
3.29164416e-01 -7.25709438e-01 8.50018635e-02 -7.10859537e-01
2.99357306e-02 -1.70589462e-01 6.89292789e-01 -1.21538198e+00
5.27041674e-01 -2.69405067e-01 5.59603810e-01 -1.29997694e+00
7.49900341e-01 -1.42734444e+00 7.60216177e-01 6.03174269e-01
3.69129241e-01 4.64470416e-01 2.63765991e-01 7.64641941e-01
1.24879129e-01 -1.50545329e-01 6.11765265e-01 -2.29302153e-01
-9.64237094e-01 3.69896203e-01 -4.27863523e-02 -4.49900836e-01
1.36958480e+00 -7.77464926e-01 3.96838821e-02 8.39506462e-03
-7.78850436e-01 7.55908489e-02 1.24022353e+00 4.86160189e-01
1.05403757e+00 -1.46311128e+00 -4.49617863e-01 1.83029905e-01
1.59427240e-01 5.89197934e-01 -2.12421417e-01 8.06636751e-01
-8.97730231e-01 3.00119221e-01 -1.67089447e-01 -1.38944662e+00
-1.29185820e+00 5.38044691e-01 2.13360325e-01 2.87190646e-01
-7.65053511e-01 6.65427804e-01 -2.59443939e-01 -6.54176831e-01
3.03065628e-01 -5.10835886e-01 5.15334666e-01 -2.09844887e-01
2.86224514e-01 5.91153145e-01 1.48960620e-01 -1.21140957e+00
-1.01750481e+00 1.04843819e+00 2.15931118e-01 -3.32342803e-01
1.42980027e+00 -4.50233072e-01 -2.17767939e-01 9.15716469e-01
7.83195257e-01 1.12576000e-01 -1.40399253e+00 -3.36970180e-01
3.94795656e-01 -1.11595774e+00 -7.28007406e-02 -1.83340415e-01
-5.33723474e-01 4.46157932e-01 9.58410561e-01 -1.94637060e-01
4.63831067e-01 3.42950314e-01 1.59012273e-01 4.68505025e-01
1.38178086e+00 -5.14957011e-01 -3.43679376e-02 6.04717374e-01
7.68584788e-01 -1.30267596e+00 7.00556695e-01 -5.22394896e-01
-1.98596567e-01 9.45409477e-01 4.28391695e-01 -5.57734668e-01
5.54476261e-01 6.11653090e-01 1.04548663e-01 -2.99455345e-01
-1.58658922e-02 -2.22239524e-01 2.41848290e-01 9.05431569e-01
-4.91940826e-02 -1.58251449e-01 3.27703804e-01 -4.02331531e-01
-2.79733520e-02 4.05598152e-03 4.21320319e-01 1.70585787e+00
-5.75941801e-01 -9.77443159e-01 -9.93109345e-01 -6.10460192e-02
2.49857120e-02 3.92216563e-01 -2.78260529e-01 1.06071270e+00
6.65190304e-03 5.26999235e-01 2.38614410e-01 -6.47985563e-02
5.82899392e-01 -3.05471830e-02 1.06302214e+00 -5.82353175e-01
-2.62723714e-01 2.55535573e-01 -2.46452257e-01 -1.00487351e+00
-1.11851466e+00 -1.22531807e+00 -1.42394114e+00 -1.51767433e-01
-6.74883306e-01 -2.20549047e-01 1.16170478e+00 6.95212662e-01
2.67117739e-01 2.41695523e-01 3.15751404e-01 -1.64320195e+00
-2.71172613e-01 -4.92729336e-01 -6.21464789e-01 2.03462303e-01
6.70029700e-01 -1.16635668e+00 -2.58316636e-01 5.11067500e-03] | [7.365808486938477, -2.3350038528442383] |
671bd0ed-4304-46bc-850d-299dfbdeb467 | discourse-planning-with-an-n-gram-model-of | null | null | https://aclanthology.org/D15-1230 | https://aclanthology.org/D15-1230.pdf | Discourse Planning with an N-gram Model of Relations | null | ['Kathleen McKeown', 'Or Biran'] | 2015-09-01 | null | null | null | emnlp-2015-9 | ['concept-to-text-generation'] | ['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] | [-7.38763952255249, 3.6979904174804688] |
91ac1634-8444-4580-955d-e27060f04f6b | video-object-segmentation-using-space-time | 1904.00607 | null | https://arxiv.org/abs/1904.00607v2 | https://arxiv.org/pdf/1904.00607v2.pdf | Video Object Segmentation using Space-Time Memory Networks | We propose a novel solution for semi-supervised video object segmentation. By the nature of the problem, available cues (e.g. video frame(s) with object masks) become richer with the intermediate predictions. However, the existing methods are unable to fully exploit this rich source of information. We resolve the issue by leveraging memory networks and learn to read relevant information from all available sources. In our framework, the past frames with object masks form an external memory, and the current frame as the query is segmented using the mask information in the memory. Specifically, the query and the memory are densely matched in the feature space, covering all the space-time pixel locations in a feed-forward fashion. Contrast to the previous approaches, the abundant use of the guidance information allows us to better handle the challenges such as appearance changes and occlussions. We validate our method on the latest benchmark sets and achieved the state-of-the-art performance (overall score of 79.4 on Youtube-VOS val set, J of 88.7 and 79.2 on DAVIS 2016/2017 val set respectively) while having a fast runtime (0.16 second/frame on DAVIS 2016 val set). | ['Joon-Young Lee', 'Seoung Wug Oh', 'Seon Joo Kim', 'Ning Xu'] | 2019-04-01 | video-object-segmentation-using-space-time-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Oh_Video_Object_Segmentation_Using_Space-Time_Memory_Networks_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Oh_Video_Object_Segmentation_Using_Space-Time_Memory_Networks_ICCV_2019_paper.pdf | iccv-2019-10 | ['interactive-video-object-segmentation', 'one-shot-visual-object-segmentation'] | ['computer-vision', 'computer-vision'] | [ 1.95259318e-01 -1.32027224e-01 -5.08890986e-01 -2.66846001e-01
-5.94535708e-01 -3.85957003e-01 1.86311409e-01 1.00076392e-01
-5.91406107e-01 5.95035136e-01 -3.16126598e-03 1.87452048e-01
2.33104259e-01 -5.68298817e-01 -1.10967898e+00 -4.38661546e-01
-1.22154467e-01 1.56671837e-01 8.86872649e-01 9.91707817e-02
2.12456182e-01 1.38596147e-01 -1.79321504e+00 5.11790752e-01
7.93398023e-01 1.47905624e+00 6.11773372e-01 5.97102940e-01
-8.74610320e-02 7.98864961e-01 -3.52748066e-01 -9.50206965e-02
4.33231086e-01 -1.19657747e-01 -8.17756116e-01 3.30287874e-01
7.79312968e-01 -6.71792865e-01 -7.55278945e-01 9.84502554e-01
3.35607350e-01 2.58853763e-01 2.40151763e-01 -1.03544450e+00
-3.35704893e-01 3.26966882e-01 -6.93049550e-01 4.86321032e-01
3.16982001e-01 2.64977187e-01 6.79389298e-01 -1.25063133e+00
1.08043885e+00 1.01570511e+00 2.87124962e-01 5.50647438e-01
-8.14389944e-01 -6.16855919e-01 6.78066552e-01 5.07885456e-01
-1.33629203e+00 -7.13800192e-01 6.57757342e-01 -4.66303736e-01
8.56954694e-01 1.00830853e-01 7.71123648e-01 8.66974235e-01
-8.90763253e-02 1.28445721e+00 8.00689578e-01 -2.96971481e-02
8.37839916e-02 1.07876241e-01 2.47718349e-01 7.65755713e-01
-1.51830509e-01 3.03727749e-04 -8.95723581e-01 2.73618251e-01
7.80150831e-01 1.55317560e-01 -4.31179553e-01 -2.70507127e-01
-1.23741078e+00 3.75099957e-01 4.29614127e-01 -1.33273661e-01
-4.54173625e-01 5.64935319e-02 4.00516212e-01 1.95893958e-01
4.53185260e-01 7.65140951e-02 -6.31957233e-01 -2.07576439e-01
-1.27124476e+00 7.45111033e-02 6.63257658e-01 1.13427484e+00
8.13543737e-01 2.37462367e-03 -4.45129335e-01 6.33485615e-01
3.82046908e-01 4.35114384e-01 2.24484891e-01 -1.20421171e+00
7.74255872e-01 4.40837890e-01 2.59107858e-01 -1.05233943e+00
-9.09754038e-02 -4.51538652e-01 -6.35286748e-01 -1.90835409e-02
3.71935099e-01 -1.33537412e-01 -1.17782331e+00 1.63392329e+00
4.69271898e-01 7.54539013e-01 -1.09124258e-01 1.16395605e+00
1.11864018e+00 8.71216714e-01 3.08019929e-02 -5.05140543e-01
1.10228145e+00 -1.47619033e+00 -6.33969486e-01 -5.28278410e-01
2.66333610e-01 -7.69698799e-01 7.55343139e-01 4.21866775e-01
-1.39323080e+00 -9.62668538e-01 -1.12374747e+00 2.48691905e-02
-5.27525544e-02 1.52101085e-01 4.13647264e-01 1.29412696e-01
-1.11674190e+00 7.22695827e-01 -9.34797585e-01 -2.20339209e-01
8.21822703e-01 4.46661443e-01 -1.77541927e-01 -2.43813023e-01
-1.00614166e+00 3.36464584e-01 3.52113634e-01 1.22861415e-01
-1.19768572e+00 -6.51322603e-01 -7.25325644e-01 -1.40332595e-01
9.07311916e-01 -4.92052644e-01 1.04299390e+00 -1.16470575e+00
-1.28033543e+00 8.01365674e-01 -3.73702884e-01 -6.53221905e-01
7.62794197e-01 -6.36280417e-01 -2.39362657e-01 5.54790020e-01
1.61786973e-01 1.18687987e+00 1.00067854e+00 -1.17053425e+00
-9.39348340e-01 -3.23639363e-01 1.67847708e-01 2.12295040e-01
-4.00041342e-01 1.99072435e-02 -1.32501841e+00 -6.47758067e-01
1.91501111e-01 -9.35358882e-01 -1.48334295e-01 4.50988680e-01
-2.40438491e-01 -3.84679586e-02 1.00420916e+00 -7.52965331e-01
1.43647313e+00 -2.18248224e+00 1.83879063e-01 -8.47998410e-02
2.87575781e-01 6.38071418e-01 -1.42012149e-01 -5.92360124e-02
1.47923544e-01 -8.52980185e-03 -2.50938684e-02 -6.23737216e-01
-3.22371721e-01 2.07582057e-01 -3.16100448e-01 3.34109426e-01
2.13740408e-01 9.28224146e-01 -9.34295237e-01 -7.54714489e-01
2.51156807e-01 4.31998670e-01 -6.31271183e-01 4.71360028e-01
-4.09747630e-01 4.13033098e-01 -5.19065917e-01 6.48498654e-01
6.22527003e-01 -4.67910081e-01 -1.88100189e-01 -3.19465041e-01
-1.10229095e-02 1.11314856e-01 -1.37080204e+00 2.04806495e+00
3.11404257e-03 6.55364633e-01 9.30979550e-02 -7.02017903e-01
6.98761225e-01 1.76968172e-01 5.92216372e-01 -7.33247459e-01
-1.19901381e-01 1.42168999e-01 -2.77292043e-01 -6.46120310e-01
5.33608615e-01 7.37328351e-01 3.64160419e-01 2.46188819e-01
1.82874382e-01 4.13734972e-01 3.90041381e-01 3.22359174e-01
8.48447323e-01 5.20134687e-01 -4.99395542e-02 -4.50543091e-02
4.21815306e-01 9.80095938e-04 8.72383714e-01 7.40634084e-01
-4.17404681e-01 7.97573626e-01 3.36144567e-01 -4.24452931e-01
-7.73984551e-01 -8.54293406e-01 8.55343565e-02 1.03094554e+00
6.03916705e-01 -7.48100340e-01 -7.71659493e-01 -6.66148841e-01
-2.56362021e-01 2.03395814e-01 -5.69013119e-01 -8.45578685e-02
-8.34778666e-01 -3.84341419e-01 7.27114305e-02 7.73280680e-01
7.27537811e-01 -1.15156353e+00 -8.63685906e-01 2.85666257e-01
-1.71660230e-01 -1.36355877e+00 -7.06147790e-01 -1.23741858e-01
-1.04705012e+00 -1.10013545e+00 -6.75025642e-01 -6.94414020e-01
5.99962473e-01 4.60601360e-01 1.18801332e+00 2.88073987e-01
-2.82162338e-01 3.54368031e-01 -2.38453269e-01 6.20654374e-02
9.04773921e-03 -1.30891860e-01 -7.02515841e-02 1.94434062e-01
1.03095680e-01 -3.43311429e-01 -1.07231343e+00 4.86742675e-01
-7.54691839e-01 4.47362900e-01 5.29994428e-01 7.47537017e-01
1.00513577e+00 -1.95448741e-01 4.54013377e-01 -8.44371438e-01
-2.38715187e-01 -6.59113050e-01 -5.55131614e-01 4.45280895e-02
-4.67053443e-01 -2.27957189e-01 2.66171217e-01 -5.47557235e-01
-9.86128807e-01 2.95789063e-01 6.76091984e-02 -8.35755527e-01
-1.29987121e-01 2.40350455e-01 -1.61882579e-01 6.59415051e-02
2.17811242e-01 3.29316854e-01 -2.23735362e-01 -5.94361961e-01
6.96549714e-02 4.15547013e-01 7.14504182e-01 -5.31948149e-01
4.39833015e-01 5.97547114e-01 -2.27148429e-01 -4.81858164e-01
-9.45302248e-01 -5.46607316e-01 -7.20721662e-01 -3.71816307e-01
9.41173792e-01 -1.15104306e+00 -4.52116758e-01 4.93688524e-01
-1.14223576e+00 -3.17694515e-01 -2.26073489e-01 3.37098837e-01
-5.26389003e-01 3.68874550e-01 -7.35373020e-01 -6.08125925e-01
-3.33045125e-01 -1.40447819e+00 9.05866086e-01 4.09430712e-01
-8.47457871e-02 -5.22934973e-01 -3.65137398e-01 3.77258986e-01
1.90797016e-01 1.96940020e-01 4.65942562e-01 -4.73049879e-01
-1.19082558e+00 1.13778815e-01 -3.78255099e-01 2.77982503e-01
-7.97791779e-02 5.45858080e-03 -9.58263934e-01 -3.15582275e-01
-2.84649455e-03 -1.79250285e-01 1.25576198e+00 4.21384960e-01
1.37417364e+00 -1.50563762e-01 -4.83688146e-01 7.27399945e-01
1.36519313e+00 2.87912607e-01 4.23793197e-01 1.67450368e-01
8.37793589e-01 4.84355062e-01 1.04059541e+00 5.14288664e-01
3.83963197e-01 8.24357271e-01 6.21164441e-01 -1.03946626e-01
-4.71766591e-01 -3.04156780e-01 4.00325209e-01 6.26247704e-01
2.35614576e-03 -2.91731209e-01 -7.80577719e-01 6.68483138e-01
-2.25727081e+00 -7.21145689e-01 -3.67807597e-02 2.20399904e+00
8.51125956e-01 4.76813853e-01 1.40147135e-02 -2.18612656e-01
7.27805257e-01 4.56132919e-01 -8.57063830e-01 2.52970457e-01
-2.56361276e-01 -7.12943375e-02 3.85641217e-01 3.34520996e-01
-1.31608915e+00 1.00936580e+00 6.02022171e+00 7.59600818e-01
-1.02666557e+00 1.07177123e-01 9.64478433e-01 -5.23445487e-01
1.77404478e-01 -1.49034753e-01 -9.58880365e-01 6.21429026e-01
7.01156735e-01 2.25926384e-01 3.50635827e-01 7.32559085e-01
1.86450422e-01 -4.49040562e-01 -1.34797740e+00 1.24856460e+00
2.48723030e-01 -1.59966326e+00 4.32873555e-02 -2.62539774e-01
9.19733047e-01 2.03939855e-01 1.54404461e-01 1.39856830e-01
-2.87916511e-01 -9.28621948e-01 1.00081384e+00 6.99193895e-01
8.02578211e-01 -4.77900416e-01 5.50088823e-01 2.52932787e-01
-1.36273837e+00 -5.47851101e-02 -2.90217131e-01 -9.80749205e-02
2.25975290e-01 3.86246622e-01 -3.60252678e-01 4.09750462e-01
1.00763154e+00 1.16402304e+00 -7.20457256e-01 1.09636998e+00
3.45180295e-02 6.61078155e-01 -3.85360271e-01 3.91675562e-01
2.92425483e-01 1.27522796e-01 6.64965570e-01 1.08182478e+00
1.72930345e-01 1.95565909e-01 5.87251782e-01 6.23409748e-01
-1.29620299e-01 1.94572043e-02 -1.14893764e-01 1.40380025e-01
3.77515703e-01 9.62215900e-01 -8.54792655e-01 -6.94724143e-01
-6.51861429e-01 9.48788583e-01 2.73868740e-01 5.55484951e-01
-9.05602515e-01 7.25865588e-02 6.61216795e-01 2.83331335e-01
6.51387870e-01 -1.05163030e-01 -1.03653267e-01 -1.11677492e+00
3.15072805e-01 -8.20739806e-01 4.67624903e-01 -7.87812233e-01
-9.41049337e-01 6.29896462e-01 -5.92255108e-02 -1.35684240e+00
-1.48128495e-01 -4.21097368e-01 -3.52846801e-01 5.58311582e-01
-1.58884251e+00 -8.50290120e-01 -3.61015886e-01 6.29240930e-01
1.02305794e+00 -1.33375302e-01 4.65684295e-01 5.30866325e-01
-6.35082781e-01 4.90909874e-01 -1.87278658e-01 1.73068330e-01
8.78389478e-01 -8.97188246e-01 2.33885556e-01 1.00101495e+00
3.10014367e-01 3.70604485e-01 4.55046654e-01 -8.00454080e-01
-1.52460599e+00 -1.16078758e+00 3.66640985e-01 -3.13559473e-01
4.18223441e-01 -2.75786012e-01 -1.03653550e+00 4.95877296e-01
1.63394362e-01 5.15682340e-01 2.97426283e-01 -3.90271366e-01
-1.92981750e-01 -1.33044988e-01 -7.84212172e-01 5.68108380e-01
1.44366646e+00 -3.35222900e-01 -3.70844275e-01 1.93392918e-01
9.78704214e-01 -8.28500450e-01 -7.07319021e-01 4.64309067e-01
4.13932592e-01 -1.04369628e+00 1.01982975e+00 -5.82338035e-01
5.65533221e-01 -5.93746781e-01 -1.98986843e-01 -7.13676572e-01
1.78521313e-02 -8.97942126e-01 -7.54835248e-01 1.17907512e+00
2.20824853e-01 -1.12860449e-01 8.59620214e-01 7.34163165e-01
-1.82670936e-01 -1.26046479e+00 -8.25231791e-01 -3.79409939e-01
-5.51882565e-01 -5.76398551e-01 4.25216436e-01 4.56037343e-01
-3.22790027e-01 7.46897310e-02 -5.81957638e-01 1.37505621e-01
5.28850675e-01 4.42713708e-01 7.17376292e-01 -9.18543339e-01
-3.68226618e-01 -1.80726781e-01 -3.76768470e-01 -1.79232597e+00
1.13227367e-01 -6.36583447e-01 -7.91419521e-02 -1.49552572e+00
3.61464709e-01 -4.75593835e-01 -5.75200737e-01 4.08988804e-01
-4.27220136e-01 4.15920466e-01 5.64587235e-01 4.20874327e-01
-1.22384715e+00 3.04051250e-01 1.22625268e+00 -1.71230078e-01
-2.94563562e-01 3.51191983e-02 -4.04738784e-01 9.36219275e-01
5.20123243e-01 -4.63026017e-01 -4.45171773e-01 -6.83645427e-01
-3.96560552e-03 2.18324408e-01 5.00157237e-01 -1.06821942e+00
5.68427861e-01 -1.28533721e-01 6.80123091e-01 -1.00584590e+00
5.84731877e-01 -7.53129005e-01 7.38321766e-02 2.87569433e-01
-2.33461216e-01 -4.24748845e-02 2.58236796e-01 7.84287453e-01
-4.01058584e-01 -1.52363935e-02 6.08530879e-01 -1.60899218e-02
-1.27039158e+00 8.61513078e-01 5.48559241e-02 2.48850822e-01
1.05899084e+00 -4.82114941e-01 -1.87628955e-01 -2.71903306e-01
-8.94694686e-01 5.76441526e-01 5.59061646e-01 6.78437531e-01
8.45701993e-01 -1.17896092e+00 -4.91889745e-01 3.49054366e-01
-5.30775189e-02 2.96415955e-01 5.15388727e-01 9.03557360e-01
-1.51284307e-01 2.28131041e-01 -2.73260236e-01 -9.46819842e-01
-1.22618556e+00 6.73927903e-01 1.11587770e-01 -7.99246281e-02
-7.31502473e-01 9.46146250e-01 4.82353330e-01 4.04571086e-01
4.94817942e-01 -2.01324224e-01 -3.69671226e-01 2.27269515e-01
6.97391629e-01 3.71478975e-01 -2.22505152e-01 -7.98375607e-01
-2.67194182e-01 7.59680569e-01 -1.63024575e-01 3.32593620e-02
1.36876214e+00 -4.14461046e-01 7.26290047e-02 4.67017412e-01
1.22560751e+00 -1.97783127e-01 -1.95130348e+00 -6.21157765e-01
-2.11328223e-01 -8.41174662e-01 -2.31033526e-02 -6.10647559e-01
-1.47037947e+00 8.93629551e-01 7.31480420e-01 -2.56485373e-01
1.22853279e+00 7.80562237e-02 1.04229367e+00 1.83415860e-01
2.84222424e-01 -1.25539601e+00 1.96996257e-01 3.70814472e-01
6.65196955e-01 -1.43956292e+00 2.00100224e-02 -6.27059877e-01
-6.46362543e-01 9.74052846e-01 9.79880452e-01 -1.47904485e-01
6.59874499e-01 2.70408183e-01 -4.14713770e-02 -3.97147089e-02
-9.36312556e-01 -8.61577541e-02 4.57815856e-01 4.13009316e-01
8.51334110e-02 -3.11968148e-01 7.48288706e-02 7.07025051e-01
2.89683700e-01 4.76914309e-02 2.62875736e-01 8.84848475e-01
-4.66020674e-01 -7.38924623e-01 -2.22381249e-01 6.32550895e-01
-4.96561378e-01 -8.98453221e-02 4.74563427e-02 4.88736004e-01
2.35357538e-01 8.98277700e-01 1.65137529e-01 -2.76515394e-01
1.75382718e-01 -1.17636144e-01 2.18439773e-01 -6.58739865e-01
-2.86212981e-01 3.45489204e-01 -4.68376055e-02 -1.14542329e+00
-6.40311360e-01 -6.51622295e-01 -1.45760977e+00 3.48746851e-02
-1.24285266e-01 -1.71193480e-01 3.25646371e-01 8.76177073e-01
5.68935454e-01 6.29758596e-01 4.10533547e-01 -1.01474547e+00
-2.00216398e-01 -6.28220141e-01 -1.86637282e-01 4.87498075e-01
4.35571104e-01 -8.49569380e-01 -1.27367582e-02 4.41169769e-01] | [9.20274829864502, -0.03192989155650139] |
13ec277e-c787-4e67-b097-8a57e6e828a1 | audio-video-emotion-recognition-in-the-wild | 2002.09023 | null | https://arxiv.org/abs/2002.09023v1 | https://arxiv.org/pdf/2002.09023v1.pdf | Audio-video Emotion Recognition in the Wild using Deep Hybrid Networks | This paper presents an audiovisual-based emotion recognition hybrid network. While most of the previous work focuses either on using deep models or hand-engineered features extracted from images, we explore multiple deep models built on both images and audio signals. Specifically, in addition to convolutional neural networks (CNN) and recurrent neutral networks (RNN) trained on facial images, the hybrid network also contains one SVM classifier trained on holistic acoustic feature vectors, one long short-term memory network (LSTM) trained on short-term feature sequences extracted from segmented audio clips, and one Inception(v2)-LSTM network trained on image-like maps, which are built based on short-term acoustic feature sequences. Experimental results show that the proposed hybrid network outperforms the baseline method by a large margin. | ['Luisa F. Polanía', 'Xin Guo', 'Kenneth E. Barner'] | 2020-02-20 | null | null | null | null | ['video-emotion-recognition'] | ['computer-vision'] | [ 3.20041597e-01 -1.35835679e-02 -3.04045584e-02 -4.59734768e-01
-7.79732764e-01 -5.22839986e-02 5.33449709e-01 -3.14994484e-01
-5.69692016e-01 3.22272778e-01 2.36581951e-01 1.86719477e-01
4.06873912e-01 -4.22586083e-01 -7.51806915e-01 -6.92094386e-01
-1.78596213e-01 -2.65428871e-01 4.45112847e-02 -1.59188852e-01
1.13423303e-01 3.14686835e-01 -1.60052037e+00 5.22678554e-01
3.76219153e-01 1.98802948e+00 -1.36454403e-01 7.58084178e-01
-2.04737499e-01 1.19600356e+00 -4.74976152e-01 -3.59833658e-01
-5.42974547e-02 -4.62959200e-01 -4.31150913e-01 -2.25182578e-01
2.89090037e-01 -1.40925705e-01 -5.65014482e-01 1.05894578e+00
6.96178496e-01 1.58432961e-01 5.25158703e-01 -1.55724192e+00
-8.16356659e-01 3.42848510e-01 -1.90629810e-01 7.77436467e-03
4.01432306e-01 6.30880296e-02 7.36081719e-01 -1.32867050e+00
4.04734224e-01 1.04469764e+00 1.03638124e+00 4.10020620e-01
-8.99457932e-01 -9.44488645e-01 -4.98141013e-02 5.98655820e-01
-1.32808626e+00 -5.84643185e-01 1.22678351e+00 -3.08685690e-01
1.28389513e+00 -1.10173877e-02 8.17322135e-01 1.75570607e+00
5.18299460e-01 9.41745937e-01 8.25433791e-01 -4.00129527e-01
1.38023913e-01 3.42519432e-01 -8.55328813e-02 5.00503361e-01
-8.96846056e-01 3.45380664e-01 -6.31230235e-01 -2.04824865e-01
5.29282928e-01 -1.42688528e-01 -1.90746218e-01 -2.09056005e-01
-6.73578262e-01 9.80908215e-01 5.52412510e-01 3.52668583e-01
-6.78787589e-01 1.46326870e-01 1.09649563e+00 5.13346136e-01
4.97803867e-01 8.28678757e-02 -4.09843296e-01 -2.65023023e-01
-1.17055738e+00 -1.31140739e-01 4.77904320e-01 5.27573228e-01
6.68892801e-01 7.92696059e-01 -1.12987161e-01 1.15299726e+00
2.06390098e-01 3.75991464e-01 1.03853953e+00 -8.06237161e-01
-1.08429782e-01 2.79287398e-01 -4.12360162e-01 -1.33634591e+00
-3.31050843e-01 -2.57255137e-01 -8.00592482e-01 2.20407471e-01
-2.48888269e-01 -2.70306289e-01 -9.41342354e-01 1.75825286e+00
-1.35434762e-01 6.40872657e-01 2.09885687e-01 8.67848814e-01
1.19992638e+00 9.40258503e-01 1.30449012e-01 -1.40662938e-01
9.17369962e-01 -1.34837008e+00 -1.01932764e+00 -4.46738973e-02
2.33093783e-01 -5.44192433e-01 7.99937248e-01 5.87740362e-01
-1.05421495e+00 -9.94101107e-01 -1.02890873e+00 1.35875698e-02
-8.37242961e-01 2.91299671e-02 1.50649428e-01 5.29983521e-01
-1.47227395e+00 3.22109312e-01 -5.62965989e-01 -3.15427542e-01
2.84514099e-01 3.25707495e-01 -5.81861913e-01 3.16780955e-01
-1.45776856e+00 7.89578259e-01 1.45758465e-01 5.03929734e-01
-1.14920950e+00 -2.36907482e-01 -1.35815251e+00 1.63381979e-01
1.04642726e-01 -1.59319267e-01 1.01453304e+00 -2.10007954e+00
-2.04721856e+00 5.66370904e-01 -2.46016402e-02 -6.74209714e-01
1.02127818e-02 -1.55763730e-01 -7.21810281e-01 6.23433828e-01
-3.29091549e-01 1.00752556e+00 1.27212453e+00 -1.01200449e+00
-2.08243564e-01 -1.85876712e-01 -4.05837715e-01 -6.20437302e-02
-8.95838559e-01 5.40520430e-01 -2.60957330e-01 -9.06779230e-01
-3.85630161e-01 -9.27262187e-01 -1.81926683e-01 -1.22935968e-02
-3.31303328e-01 -8.73471051e-02 1.08756709e+00 -8.32773507e-01
1.08905792e+00 -2.35103393e+00 3.22142877e-02 1.43513441e-01
-3.09394836e-01 4.33021873e-01 -7.50510395e-01 2.22909778e-01
-5.15868902e-01 -5.15635237e-02 -1.17546255e-02 -2.07590237e-01
7.31910542e-02 -2.95107029e-02 -5.11363924e-01 2.57318407e-01
5.85963905e-01 1.15338898e+00 -5.90231180e-01 -5.15232801e-01
3.47465336e-01 9.34507549e-01 -2.92834252e-01 2.94100821e-01
1.00654557e-01 1.17830254e-01 -3.25882733e-02 7.41925418e-01
5.97221494e-01 2.80383706e-01 -2.17007443e-01 -1.40598893e-01
1.69183258e-02 -7.81283528e-02 -7.50820518e-01 1.92275703e+00
-5.77446043e-01 8.92717361e-01 1.17150836e-01 -1.32138646e+00
1.08705640e+00 8.69362414e-01 5.15491486e-01 -8.76564562e-01
3.98141146e-01 2.58862842e-02 -4.01854753e-01 -6.73210859e-01
4.32190359e-01 -2.55749255e-01 -2.17983574e-01 2.69623306e-02
8.62070382e-01 1.48793995e-01 -4.22451794e-01 -1.08200632e-01
1.02347469e+00 1.62215233e-01 -6.47179261e-02 1.76915288e-01
6.68822944e-01 -4.84553039e-01 6.47159815e-01 4.05504107e-01
-4.97141778e-01 7.65492499e-01 3.47111732e-01 -4.19891030e-01
-7.57162333e-01 -8.92160177e-01 -1.82333648e-01 1.20261240e+00
-1.91323817e-01 -5.81450641e-01 -7.05980837e-01 -5.66758633e-01
-2.08106667e-01 4.10840809e-01 -1.11473739e+00 -4.28656340e-01
-1.86673149e-01 -2.13026837e-01 9.47829247e-01 8.64337981e-01
3.70948195e-01 -1.70563197e+00 -6.06464863e-01 2.61708468e-01
-1.07163213e-01 -1.09486556e+00 -2.50034630e-01 3.76711845e-01
-4.21326876e-01 -4.83836949e-01 -8.45784426e-01 -9.09755707e-01
-1.14302769e-01 -3.36028337e-01 6.84981704e-01 -3.69292021e-01
-2.94487625e-01 5.62876821e-01 -5.49326062e-01 -5.51382601e-01
9.02116671e-03 -2.04931274e-01 1.54390186e-01 6.30911827e-01
6.01893544e-01 -7.56455839e-01 -1.32543996e-01 2.51879785e-02
-7.52408326e-01 -4.52926576e-01 4.73100662e-01 1.09618497e+00
5.54031312e-01 -2.40080133e-01 9.14241552e-01 -2.66970664e-01
4.47829247e-01 -4.92335916e-01 -1.00907937e-01 9.98700410e-02
-1.41417235e-01 -4.56867665e-01 5.97032726e-01 -7.85296798e-01
-1.03136909e+00 2.65336305e-01 -6.23864472e-01 -1.24777937e+00
-3.13576400e-01 7.56433010e-01 -5.76528944e-02 -1.64144620e-01
4.60988402e-01 5.23502588e-01 1.32303700e-01 -7.23879635e-02
2.50429392e-01 8.68764758e-01 7.66287744e-01 -6.76244274e-02
2.70137042e-01 2.08012626e-01 -5.79173684e-01 -1.24270391e+00
-5.93627989e-01 -1.65171102e-01 -5.90155303e-01 -6.26082718e-01
1.19720781e+00 -1.09644163e+00 -4.94598001e-01 7.32763529e-01
-1.00978816e+00 -3.17929238e-01 -1.90390006e-01 6.26595080e-01
-6.73118234e-01 -6.54099807e-02 -1.10084677e+00 -1.03945637e+00
-1.84909776e-01 -1.09099567e+00 9.77958798e-01 1.24317050e-01
-3.25990438e-01 -9.49367166e-01 1.98759228e-01 -1.85643658e-02
6.13254309e-01 5.69813788e-01 5.47770500e-01 -7.92473137e-01
1.00615472e-01 -5.23638844e-01 -6.57734349e-02 8.51300001e-01
-2.17190862e-01 2.20300525e-01 -1.47777867e+00 -1.54872295e-02
1.84853509e-01 -1.16671586e+00 1.26550591e+00 7.09361136e-01
1.45835626e+00 -2.01101482e-01 -4.89460677e-02 6.67551816e-01
1.19326437e+00 4.19923007e-01 9.11380947e-01 1.65018722e-01
4.75893945e-01 5.09130239e-01 3.47551465e-01 2.59281665e-01
2.41225764e-01 6.69576347e-01 6.87349856e-01 -3.20852339e-01
2.17662722e-01 -1.47996500e-01 7.95998991e-01 1.04340255e+00
6.71370775e-02 4.80530635e-02 -5.92515409e-01 5.92424810e-01
-1.73084486e+00 -1.18424881e+00 3.50547045e-01 1.64968741e+00
5.63691974e-01 7.20112920e-02 1.12515181e-01 2.01016873e-01
5.92273235e-01 3.07400614e-01 -5.37160814e-01 -9.30038750e-01
-3.69860291e-01 6.87670350e-01 -3.66959065e-01 1.49956480e-01
-1.34618270e+00 9.91743982e-01 6.58697796e+00 8.87754560e-01
-1.78553605e+00 1.35757476e-01 4.51528132e-01 -3.19399357e-01
1.30536884e-01 -4.48681295e-01 -1.71468481e-01 3.28218460e-01
1.58147538e+00 2.45476365e-01 6.59969077e-02 1.12502694e+00
-1.47041932e-01 3.03500146e-01 -7.13091791e-01 1.07640159e+00
6.17102504e-01 -1.11781383e+00 1.23924091e-01 -1.47861481e-01
5.14878213e-01 2.03721657e-01 5.86137593e-01 7.72109866e-01
-1.45222470e-01 -1.25546896e+00 8.87072265e-01 7.52682030e-01
9.04596150e-01 -1.12969625e+00 8.89508724e-01 -8.40112790e-02
-1.31027758e+00 -3.97226065e-01 -2.83121139e-01 -1.80707555e-02
-8.58955178e-03 2.21609652e-01 -1.28673509e-01 4.74794567e-01
1.07101440e+00 1.03750169e+00 -3.86588186e-01 6.64052665e-01
-4.44119535e-02 7.40011752e-01 5.00354059e-02 9.53989103e-02
5.55420935e-01 8.69451836e-02 2.79947281e-01 1.58859289e+00
2.89620250e-01 -9.63328481e-02 1.32687569e-01 5.90395033e-01
-1.18586592e-01 2.71447688e-01 -8.53001356e-01 -2.69059896e-01
-1.88135989e-02 1.64875329e+00 -3.60011458e-01 -3.60931218e-01
-6.66264117e-01 1.24185014e+00 1.25829250e-01 5.57884276e-01
-9.63986993e-01 -1.01638353e+00 5.89438140e-01 -4.02336061e-01
5.38207710e-01 2.14368880e-01 1.71731398e-01 -1.09222257e+00
-1.83962122e-01 -7.30439067e-01 2.02390730e-01 -1.18874228e+00
-1.35294199e+00 1.35024357e+00 -6.22245193e-01 -1.18090534e+00
-5.37422121e-01 -5.75231016e-01 -8.52622092e-01 6.37517929e-01
-1.60338330e+00 -1.31179368e+00 -3.20532247e-02 8.98474753e-01
5.33128858e-01 -3.99363637e-01 1.16324890e+00 2.87438095e-01
-5.78146994e-01 8.28191340e-01 -8.33989233e-02 5.09304821e-01
8.32230449e-01 -6.99601829e-01 -1.31486878e-01 2.50111133e-01
2.15019554e-01 4.21020895e-01 3.24292123e-01 -2.35325977e-01
-9.59682882e-01 -1.26462114e+00 8.31486881e-01 -9.84353423e-02
5.95452547e-01 -4.85329956e-01 -1.08017862e+00 7.51076460e-01
7.36872852e-01 3.61571908e-01 1.05328298e+00 -1.27862528e-01
-5.32293499e-01 -8.14057514e-02 -1.00081086e+00 7.24764392e-02
3.60658169e-01 -1.11019135e+00 -5.04207015e-01 -1.65014803e-01
6.68067515e-01 -1.41919374e-01 -7.96686411e-01 6.86309814e-01
9.61381733e-01 -9.92250025e-01 8.41156363e-01 -6.76359296e-01
7.26903498e-01 1.61790311e-01 -4.83249635e-01 -1.34187686e+00
-3.23649868e-02 -2.14187995e-01 9.04708132e-02 1.04777145e+00
3.56004000e-01 -3.49974155e-01 5.36566973e-01 6.15937961e-03
-3.18004936e-01 -8.48823488e-01 -1.08848250e+00 -7.15676010e-01
-4.35733683e-02 -8.47651720e-01 2.05695182e-01 1.14145446e+00
1.37867615e-01 3.95076603e-01 -6.17381871e-01 -9.07226726e-02
8.80783945e-02 -6.33337125e-02 4.86132801e-01 -1.04748631e+00
1.20252423e-01 -5.15486062e-01 -8.16343725e-01 -4.91706431e-01
9.05989945e-01 -8.01128030e-01 2.51066238e-01 -9.43815410e-01
1.42642230e-01 3.17238539e-01 -1.00201261e+00 6.46102130e-01
2.78704047e-01 8.27156007e-01 5.43834008e-02 -7.02157974e-01
-7.19064295e-01 1.23913765e+00 5.61488152e-01 -2.30807886e-01
-2.64380455e-01 -2.95754462e-01 -2.72915751e-01 1.03574014e+00
5.21497548e-01 -3.15269738e-01 -1.40635744e-01 -1.09950695e-02
-1.36899594e-02 3.55952621e-01 5.64421952e-01 -1.18154931e+00
4.34892088e-01 1.72320649e-01 7.67773449e-01 -4.95648414e-01
9.84180808e-01 -7.16976643e-01 -1.35847270e-01 2.84216732e-01
-6.69469297e-01 -1.29600521e-02 4.72814918e-01 5.05268037e-01
-9.04549360e-01 1.51117966e-01 7.80615866e-01 1.59929618e-02
-6.41902030e-01 3.93026114e-01 -8.57183814e-01 -4.82923031e-01
9.32693779e-01 -3.94606650e-01 1.83817402e-01 -7.76709497e-01
-1.13359141e+00 -4.12300900e-02 3.90047133e-02 8.30341101e-01
1.10978031e+00 -1.89013064e+00 -4.34246093e-01 5.58163464e-01
2.04632282e-01 -8.08099985e-01 4.92726952e-01 7.11338639e-01
1.50977254e-01 5.38341641e-01 -7.16271043e-01 -6.58552229e-01
-1.20987844e+00 6.57517910e-01 3.41185778e-01 2.16210499e-01
-1.80347338e-01 1.21321595e+00 1.82349384e-01 -6.47311091e-01
5.71187496e-01 -1.30124182e-01 -5.26024163e-01 4.80762035e-01
7.58535922e-01 -3.71174440e-02 -1.94720607e-02 -1.19411874e+00
-3.76838267e-01 5.73555589e-01 1.26218110e-01 -4.06111062e-01
1.54392564e+00 9.89176482e-02 -1.22807592e-01 8.85678709e-01
1.66386235e+00 -4.51168001e-01 -1.10201323e+00 -3.98553967e-01
-1.31384686e-01 -2.20981687e-02 2.92353511e-01 -6.24615371e-01
-1.40095174e+00 1.32000077e+00 8.11065376e-01 -7.65743032e-02
1.54321778e+00 -4.74349916e-01 9.31582332e-01 2.68553555e-01
-1.23236850e-02 -1.27836204e+00 7.87049234e-01 6.02200508e-01
1.06687391e+00 -9.79175627e-01 -9.09722209e-01 2.87980705e-01
-9.69309926e-01 1.27503216e+00 6.74287498e-01 -3.80551904e-01
1.17665696e+00 2.95881569e-01 4.12035882e-01 8.21624696e-02
-1.11157787e+00 -1.30656421e-01 3.21097642e-01 6.09086931e-01
2.84855962e-01 -4.17934626e-01 3.74369979e-01 1.06490052e+00
-2.03150585e-01 1.63008779e-01 3.26127708e-01 6.96879566e-01
-2.65948087e-01 -7.64748812e-01 -5.52440658e-02 4.02023703e-01
-6.27647281e-01 -1.22260340e-01 -5.10388732e-01 5.06233573e-01
1.52561232e-01 8.69224191e-01 3.60030711e-01 -1.02368653e+00
9.28849280e-02 5.31972289e-01 1.91841841e-01 -3.17728281e-01
-8.86649728e-01 2.83335179e-01 -1.94469914e-01 -8.46335828e-01
-5.73749244e-01 -4.28389549e-01 -1.06079519e+00 2.33384028e-01
-4.21820790e-01 8.01398084e-02 8.04636717e-01 8.75392735e-01
2.06193894e-01 5.29974103e-01 8.23733985e-01 -1.18675828e+00
-4.43336032e-02 -1.16961026e+00 -8.05136919e-01 1.36777782e-03
7.16185510e-01 -5.44250607e-01 -2.62687892e-01 9.15439948e-02] | [13.35219955444336, 5.138749599456787] |
83fab808-e252-4f4f-92ce-b148ecfe9630 | using-eeg-signals-to-assess-workload-during | 2305.08044 | null | https://arxiv.org/abs/2305.08044v1 | https://arxiv.org/pdf/2305.08044v1.pdf | Using EEG Signals to Assess Workload during Memory Retrieval in a Real-world Scenario | Objective: The Electroencephalogram (EEG) is gaining popularity as a physiological measure for neuroergonomics in human factor studies because it is objective, less prone to bias, and capable of assessing the dynamics of cognitive states. This study investigated the associations between memory workload and EEG during participants' typical office tasks on a single-monitor and dual-monitor arrangement. We expect a higher memory workload for the single-monitor arrangement. Approach: We designed an experiment that mimics the scenario of a subject performing some office work and examined whether the subjects experienced various levels of memory workload in two different office setups: 1) a single-monitor setup and 2) a dual-monitor setup. We used EEG band power, mutual information, and coherence as features to train machine learning models to classify high versus low memory workload states. Main results: The study results showed that these characteristics exhibited significant differences that were consistent across all participants. We also verified the robustness and consistency of these EEG signatures in a different data set collected during a Sternberg task in a prior study. Significance: The study found the EEG correlates of memory workload across individuals, demonstrating the effectiveness of using EEG analysis in conducting real-world neuroergonomic studies. | ['Tzyy-Ping Jung', 'Chung-Kuan Cheng', 'Steven Dong', 'Kuan-Jung Chiang'] | 2023-05-14 | null | null | null | null | ['eeg', 'eeg'] | ['methodology', 'time-series'] | [ 7.49882236e-02 -3.59821171e-01 1.20773226e-01 -2.84785658e-01
6.29138201e-02 -3.05058688e-01 2.74848372e-01 2.71483064e-01
-6.39417052e-01 7.82676101e-01 1.03420354e-01 -2.68748879e-01
-5.26803672e-01 -4.63552773e-01 -3.91516149e-01 -4.71337318e-01
-3.08474153e-01 5.76882577e-03 -8.75819325e-02 7.01961368e-02
5.83418310e-01 3.96831065e-01 -1.57384145e+00 8.27695653e-02
5.35391331e-01 9.95554626e-01 5.98750234e-01 4.19059992e-01
6.12815917e-01 1.36152387e-01 -1.05680561e+00 7.29082152e-02
2.77346283e-01 -5.11430800e-01 -5.01596749e-01 7.14442208e-02
-4.39191610e-02 1.49212763e-01 -2.21162692e-01 7.33790815e-01
1.03631973e+00 3.77763063e-01 3.43209416e-01 -1.17278481e+00
-3.09743524e-01 1.51311576e-01 -5.03311157e-01 9.35280561e-01
7.41628706e-01 3.50596905e-01 2.00205669e-01 -1.77121788e-01
3.43820415e-02 6.66287482e-01 4.58308518e-01 2.01695114e-02
-1.46263373e+00 -1.08976603e+00 -6.83057159e-02 5.17894626e-01
-1.51971459e+00 -5.76555967e-01 6.20733738e-01 -6.55959666e-01
1.05517244e+00 2.73967981e-01 1.11040509e+00 1.24767005e+00
1.01043022e+00 -4.36587274e-01 1.87839580e+00 -2.84649134e-01
3.27254146e-01 6.12697542e-01 6.16113484e-01 6.70408681e-02
6.25735462e-01 2.66040891e-01 -9.87365246e-01 -2.32654139e-01
5.72351098e-01 -3.28564614e-01 -6.63156927e-01 2.88906872e-01
-1.15770543e+00 2.40680367e-01 -2.26335257e-01 7.18749762e-01
-7.18289673e-01 -2.83655614e-01 4.16662037e-01 3.69756639e-01
3.24632943e-01 8.01370442e-01 -5.70643507e-02 -6.18578672e-01
-9.58884358e-01 2.20816545e-02 6.85875297e-01 5.07940888e-01
1.04171991e-01 -9.36947986e-02 -4.20381278e-01 5.69918394e-01
-1.30863935e-01 3.71468246e-01 8.73210311e-01 -5.85093617e-01
3.67337853e-01 2.03136891e-01 2.05126137e-01 -1.25860977e+00
-1.06229556e+00 -6.12845540e-01 -4.19926137e-01 8.15650001e-02
1.60307959e-01 -1.11844048e-01 -2.29383126e-01 1.65675497e+00
-4.17857230e-01 3.84022705e-02 -6.18024886e-01 9.99759138e-01
1.60408258e-01 1.99303836e-01 4.71691713e-02 -7.41870821e-01
1.58688235e+00 -1.67563595e-02 -1.01052785e+00 -2.21823856e-01
1.13114499e-01 -5.19609451e-01 1.34515345e+00 5.85605979e-01
-1.13535607e+00 -9.90074813e-01 -1.40327883e+00 7.17019439e-01
-3.23286712e-01 -6.60781190e-02 2.44478077e-01 1.32217395e+00
-9.87944782e-01 6.87027454e-01 -9.87667680e-01 -3.72263581e-01
-2.16754466e-01 3.45480800e-01 -2.87114382e-01 5.77253819e-01
-1.05820918e+00 1.38551831e+00 3.09620768e-01 1.68248504e-01
-5.50497293e-01 -4.55523580e-01 -3.98227721e-01 1.32935867e-01
-7.96265006e-02 -4.30374563e-01 7.18544126e-01 -4.77411836e-01
-1.34376800e+00 7.05756426e-01 -3.32796514e-01 -3.98326293e-02
-3.94920334e-02 -1.00130379e-01 -8.77623677e-01 -7.75025561e-02
6.63104653e-02 -7.98435509e-02 5.11786640e-01 -7.42083728e-01
1.67070076e-01 -8.88659477e-01 -3.39748442e-01 1.03686109e-01
-1.45063281e-01 1.62568659e-01 5.86858809e-01 -3.92865509e-01
1.91039860e-01 -8.98153961e-01 5.45667291e-01 -9.44547057e-01
-1.29847586e-01 8.50632936e-02 8.94277319e-02 -7.38743305e-01
1.47373617e+00 -2.08513999e+00 -3.05905491e-01 6.62025332e-01
1.62481159e-01 -2.70640552e-01 1.46526054e-01 3.49595189e-01
-3.19495410e-01 3.52043062e-02 1.43066972e-01 1.23947538e-01
9.87304822e-02 -3.22077096e-01 1.12496413e-01 9.54672039e-01
-2.21634910e-01 5.44937909e-01 -5.92357934e-01 -7.36911744e-02
3.29821855e-02 1.09037042e-01 -3.10337767e-02 3.32653135e-01
8.44462097e-01 5.70675075e-01 4.37759906e-01 3.06311250e-01
5.06491125e-01 1.27612844e-01 1.73149034e-01 -2.14064464e-01
-4.72623467e-01 4.11354750e-01 -9.80608344e-01 1.31136787e+00
-3.02499175e-01 1.18168652e+00 -1.63835362e-01 -6.99079514e-01
9.62545812e-01 5.16966224e-01 1.22218929e-01 -1.40174639e+00
4.47523713e-01 -3.18002887e-02 6.63946986e-01 -8.33782613e-01
3.47727090e-01 -3.07772160e-01 1.44479081e-01 8.65497172e-01
-9.70561802e-02 -1.32865027e-01 7.63891637e-02 -3.54744196e-01
1.00359762e+00 -2.90891796e-01 2.78969437e-01 -7.45191813e-01
-1.95670724e-02 -7.34345615e-01 2.19783753e-01 7.81222522e-01
-4.08938140e-01 2.00037777e-01 4.53931630e-01 -4.86137113e-03
-3.75469834e-01 -1.00779593e+00 -3.96885902e-01 6.90426350e-01
2.00408548e-01 -3.10633093e-01 -6.75382674e-01 2.73073673e-01
-2.62188822e-01 1.11348701e+00 -5.94500721e-01 -8.02392006e-01
-7.53184259e-02 -9.14179504e-01 3.61780852e-01 5.20345330e-01
5.12471914e-01 -8.39485884e-01 -1.38564205e+00 -1.26032352e-01
-3.09784830e-01 -8.87516320e-01 -3.53298098e-01 3.72248620e-01
-7.88024843e-01 -1.00499237e+00 -2.73687512e-01 -2.43110612e-01
3.99522364e-01 2.29190066e-01 1.02256560e+00 -4.47828561e-01
-3.46805125e-01 7.39649117e-01 2.66292095e-02 -8.82106721e-01
2.35195160e-01 -1.72823220e-01 5.59938073e-01 -8.53867978e-02
5.56767523e-01 -9.26286459e-01 -7.38629162e-01 4.95797455e-01
-4.47373748e-01 -4.78906155e-01 6.67162657e-01 2.76506335e-01
1.44416451e-01 2.60203481e-01 6.25519156e-01 -7.33961351e-03
1.55375063e+00 -5.35854101e-01 -1.70541480e-01 -3.02298442e-02
-1.04230142e+00 -4.29713070e-01 2.04919726e-02 -4.95712161e-01
-8.40710402e-01 -7.04294801e-01 6.20977938e-01 2.29041293e-01
-4.32027906e-01 1.75643355e-01 -1.94351241e-01 3.77641320e-02
7.40540683e-01 1.01295218e-01 3.60855944e-02 -5.28947562e-02
-6.72493577e-01 8.57173979e-01 5.17392635e-01 -4.56065059e-01
4.35395492e-03 3.11860815e-02 -3.67789157e-02 -1.09971261e+00
-1.26728892e-01 -2.87566006e-01 -6.48541391e-01 -5.56804478e-01
9.43020165e-01 -7.16071844e-01 -1.02650023e+00 2.93817163e-01
-9.22779918e-01 -3.64092022e-01 3.24465305e-01 1.21085298e+00
-3.90636742e-01 1.32026717e-01 -1.35572672e-01 -1.05174208e+00
-1.74238026e-01 -9.23535049e-01 5.70572257e-01 1.44262984e-01
-1.05740225e+00 -7.93971956e-01 2.05092087e-01 3.33299458e-01
4.92540926e-01 2.79803216e-01 7.41954982e-01 -4.70601290e-01
2.38529742e-01 -2.97122777e-01 2.41832882e-01 9.76672396e-02
3.70808691e-01 -7.18971848e-01 -1.19771504e+00 -4.02017653e-01
8.98329675e-01 -5.09123839e-02 -2.61366665e-02 3.94241363e-01
1.03914213e+00 1.05081826e-01 3.45286392e-02 1.35583997e-01
1.04525626e+00 6.77263558e-01 8.34698617e-01 3.88894945e-01
-9.20679271e-02 6.98703110e-01 1.25917643e-01 2.24066079e-01
-4.35240008e-02 6.58837199e-01 -1.15376100e-01 1.84955165e-01
4.44136143e-01 1.88615456e-01 4.48318660e-01 6.10818088e-01
-5.38763046e-01 3.73122618e-02 -6.59192920e-01 1.26580179e-01
-1.17769599e+00 -1.05559349e+00 -2.22656876e-01 2.65773511e+00
3.73214960e-01 2.63100266e-01 3.85086685e-01 3.59235287e-01
6.57452822e-01 -2.14356959e-01 -2.90646762e-01 -6.76137507e-01
1.02435119e-01 4.54317838e-01 4.37857300e-01 -2.28807941e-01
-4.06817198e-01 -4.06070650e-01 7.74544907e+00 -6.19888753e-02
-1.11477530e+00 3.48938406e-01 4.59102333e-01 -8.27219307e-01
3.79830629e-01 -4.27686125e-01 -2.35565096e-01 9.63088334e-01
1.71627343e+00 -4.68310207e-01 3.99784058e-01 1.27573431e-01
9.31911170e-01 -8.77197266e-01 -1.28114855e+00 1.13385510e+00
2.14132056e-01 -4.96499777e-01 -6.40816927e-01 2.74104029e-01
1.27087295e-01 -4.00760800e-01 6.19539469e-02 3.86366062e-02
-1.10104620e+00 -1.04476297e+00 8.10716867e-01 8.71391892e-01
7.09310591e-01 -5.56536198e-01 8.71053159e-01 3.84773463e-01
-6.61723733e-01 -1.44776911e-01 -1.43776730e-01 -7.77356148e-01
2.14079052e-01 4.83282268e-01 -5.43900788e-01 1.81514546e-01
7.61750162e-01 -7.56789446e-02 -5.41803002e-01 9.50399935e-01
2.02992052e-01 6.80909753e-01 4.42185393e-03 -1.07803062e-01
-3.92451704e-01 -1.79776117e-01 3.55804116e-01 1.06407642e+00
4.23685938e-01 1.13183807e-03 -3.75603676e-01 1.24164951e+00
5.43295979e-01 -1.22113973e-01 -6.10619366e-01 -2.68271659e-02
4.94571775e-01 9.57549036e-01 -9.98265445e-01 -2.86025405e-02
-4.34505939e-01 7.85298526e-01 -2.49451160e-01 2.97002316e-01
-7.97467172e-01 -6.89641893e-01 3.51580650e-01 3.60080063e-01
-5.45510352e-01 -3.42181027e-01 -6.96261942e-01 -6.82370126e-01
3.86936218e-01 -5.93285501e-01 -5.56682050e-02 -9.19027030e-01
-9.02962148e-01 4.70541120e-01 7.01722443e-01 -8.25279713e-01
1.19548246e-01 -4.91662472e-01 -7.68578291e-01 1.41305363e+00
-6.61250472e-01 -5.95696568e-02 -7.31621206e-01 5.34694612e-01
2.55299836e-01 3.62722240e-02 8.16126168e-01 2.74771392e-01
-7.66050279e-01 4.52976078e-01 -4.56519634e-01 -5.16084433e-01
8.08279991e-01 -9.32661176e-01 -2.30910212e-01 6.52982473e-01
-3.14257085e-01 1.24945140e+00 7.55327523e-01 -7.49339342e-01
-1.09984231e+00 -2.90640801e-01 6.00746274e-01 -2.84866184e-01
3.31054360e-01 -5.47665417e-01 -6.36619389e-01 3.36546659e-01
5.92054546e-01 -7.93547869e-01 1.11004162e+00 1.94630116e-01
3.24993432e-01 -2.04197526e-01 -1.09349370e+00 3.25227827e-01
9.95789826e-01 -7.65948355e-01 -9.11371648e-01 9.30597559e-02
-4.61940421e-03 -3.10756322e-02 -1.12089980e+00 5.93659244e-02
1.00774288e+00 -1.25186288e+00 4.20046300e-01 -1.74868882e-01
1.68871116e-02 1.32071033e-01 2.10943297e-01 -1.62447059e+00
-6.61865175e-01 -7.23900557e-01 2.73317844e-01 8.12189221e-01
2.83888817e-01 -1.15576422e+00 1.19551867e-02 9.97101128e-01
-2.59442329e-01 -5.47821879e-01 -6.89915061e-01 -1.00552475e+00
-4.09755349e-01 -3.64125907e-01 4.44362879e-01 5.55313885e-01
8.65083337e-01 2.44115159e-01 1.68304285e-03 3.26132737e-02
1.80428714e-01 -1.80508479e-01 2.94266939e-01 -1.17126596e+00
-5.18605232e-01 -5.31772971e-01 -8.57650101e-01 -1.37848094e-01
9.73597765e-02 -7.26025164e-01 -1.80249631e-01 -1.31677973e+00
4.59380656e-01 2.00475916e-01 -4.60271239e-01 3.83060873e-02
-1.18553065e-01 5.61095029e-03 -1.27680942e-01 -9.53220576e-02
1.70636356e-01 1.02265298e-01 7.39089966e-01 2.71644056e-01
-5.50983250e-01 1.98115185e-01 -8.99482787e-01 2.15412959e-01
1.08068657e+00 -2.89684892e-01 -9.25197661e-01 -5.94224893e-02
8.51764381e-02 -7.17956759e-03 3.58624220e-01 -1.37848580e+00
6.02565259e-02 1.25131667e-01 8.18963647e-01 -1.60189539e-01
3.94526154e-01 -6.69929981e-01 4.47670281e-01 4.66616541e-01
-1.94363296e-01 7.76955724e-01 6.78306520e-01 2.35314265e-01
2.05081180e-01 -1.29621223e-01 5.53418100e-01 -2.30585709e-01
-1.64690629e-01 -4.90154684e-01 -7.39673972e-01 -2.21003607e-01
1.08668244e+00 -8.66533816e-01 -3.88883442e-01 -3.28125000e-01
-6.45131648e-01 -2.42716014e-01 2.15396866e-01 3.15048844e-01
3.54234517e-01 -1.07603598e+00 -1.83264136e-01 4.06755596e-01
-1.18100308e-01 -9.16967988e-01 3.14253837e-01 1.76326811e+00
-1.27537489e-01 6.72472119e-01 -8.70117843e-01 -4.50556993e-01
-1.11496115e+00 4.00257677e-01 4.37002093e-01 2.88782448e-01
-4.87683892e-01 3.44920188e-01 1.75107475e-02 5.54929852e-01
2.72074372e-01 -2.92402297e-01 -8.51700678e-02 2.79196233e-01
7.55393028e-01 1.14204919e+00 5.92965603e-01 -3.00692797e-01
-2.70324141e-01 1.00921519e-01 5.22223532e-01 -2.61009693e-01
9.75077987e-01 -4.24352229e-01 -4.10987407e-01 1.21795893e+00
8.12777579e-01 1.32247992e-02 -5.63793421e-01 7.07468212e-01
-3.06033976e-02 -5.05136490e-01 1.80447549e-02 -8.08912396e-01
-4.91964519e-01 7.01088190e-01 1.23527396e+00 3.65813017e-01
1.32398522e+00 -4.09992129e-01 1.58710167e-01 2.35828891e-01
6.51643753e-01 -1.47662055e+00 4.28417213e-02 -1.98391333e-01
1.05317616e+00 -3.69334400e-01 1.41174600e-01 1.91177011e-01
-4.92422491e-01 8.68170857e-01 6.53320730e-01 1.02694534e-01
8.35228562e-01 2.31316909e-01 -3.77796710e-01 -5.20274341e-01
-6.25462890e-01 1.28417939e-01 3.22306305e-01 6.04760289e-01
7.38474071e-01 2.91195214e-01 -8.19268107e-01 9.13774014e-01
-6.62655890e-01 -5.20664407e-03 7.15855479e-01 1.01470876e+00
-2.89700478e-01 -5.33341467e-01 -5.88346243e-01 9.28571105e-01
-2.99545258e-01 2.47780830e-01 -4.47621673e-01 9.06829834e-01
3.22575033e-01 1.38570237e+00 3.97881150e-01 -6.37868345e-01
8.99133444e-01 6.89952135e-01 7.93801665e-01 -6.82195246e-01
-6.90085888e-01 -1.77112266e-01 5.15488861e-03 -7.00646520e-01
-3.50812167e-01 -8.10375631e-01 -7.61093915e-01 -1.52817279e-01
-3.60145777e-01 2.11194828e-01 8.41888189e-01 9.16213393e-01
6.09180570e-01 8.69749188e-01 2.80959368e-01 -8.91407371e-01
-2.29111373e-01 -1.52657783e+00 -1.21978784e+00 1.60236359e-01
-2.50644404e-02 -9.45489824e-01 -4.62247729e-01 -8.42722952e-02] | [13.34228515625, 3.288900852203369] |
6212a61d-4d91-4dc0-af3c-8e7c263ae70a | tea-pse-3-0-tencent-ethereal-audio-lab | 2303.07704 | null | https://arxiv.org/abs/2303.07704v1 | https://arxiv.org/pdf/2303.07704v1.pdf | TEA-PSE 3.0: Tencent-Ethereal-Audio-Lab Personalized Speech Enhancement System For ICASSP 2023 DNS Challenge | This paper introduces the Unbeatable Team's submission to the ICASSP 2023 Deep Noise Suppression (DNS) Challenge. We expand our previous work, TEA-PSE, to its upgraded version -- TEA-PSE 3.0. Specifically, TEA-PSE 3.0 incorporates a residual LSTM after squeezed temporal convolution network (S-TCN) to enhance sequence modeling capabilities. Additionally, the local-global representation (LGR) structure is introduced to boost speaker information extraction, and multi-STFT resolution loss is used to effectively capture the time-frequency characteristics of the speech signals. Moreover, retraining methods are employed based on the freeze training strategy to fine-tune the system. According to the official results, TEA-PSE 3.0 ranks 1st in both ICASSP 2023 DNS-Challenge track 1 and track 2. | ['Shidong Shang', 'Tao Yu', 'Yannan Wang', 'Weixin Zhu', 'Wei Rao', 'Shulin He', 'Shimin Zhang', 'Jun Chen', 'Yukai Ju'] | 2023-03-14 | null | null | null | null | ['speech-enhancement'] | ['speech'] | [ 2.40988489e-02 -4.05704290e-01 6.57699183e-02 -2.54081845e-01
-1.02810991e+00 -5.74490368e-01 5.07574081e-01 -7.28501976e-01
-6.51258588e-01 5.13927996e-01 5.69630623e-01 -4.11411464e-01
-2.60440968e-02 2.48576645e-02 -4.93203998e-01 -5.91135383e-01
-1.05222724e-01 -3.16731423e-01 -1.06119793e-02 -4.52447563e-01
-2.54328251e-01 4.87693787e-01 -1.03571725e+00 3.96746606e-01
7.23644018e-01 1.19514596e+00 1.41509473e-01 7.95127332e-01
9.13661197e-02 9.22313750e-01 -7.80083656e-01 -1.49255902e-01
4.33697075e-01 -4.18177068e-01 -6.51872873e-01 -6.03897035e-01
4.10611570e-01 -3.19450200e-01 -9.79140401e-01 1.13687897e+00
1.00774086e+00 6.34672761e-01 -6.04291596e-02 -9.52842653e-01
-3.78647953e-01 1.00259507e+00 -3.39655757e-01 7.61074483e-01
-1.11537412e-01 2.68593609e-01 7.40975082e-01 -8.86161387e-01
2.86644489e-01 1.30578923e+00 1.07992506e+00 6.86339378e-01
-1.12377048e+00 -1.07293546e+00 2.04401985e-01 6.77988231e-01
-1.37043655e+00 -9.33729649e-01 8.28168571e-01 -8.25733319e-02
1.18860686e+00 4.56699282e-01 2.94940829e-01 1.82874584e+00
-2.89683372e-01 8.51791203e-01 8.28507483e-01 -1.76621303e-01
1.14584677e-01 -4.24952686e-01 1.91306308e-01 9.57162082e-02
-4.35264260e-01 3.01526219e-01 -9.02627230e-01 -1.15979724e-01
6.74337804e-01 -4.92989361e-01 -4.01171207e-01 5.17875671e-01
-1.16771221e+00 5.37086487e-01 3.24377030e-01 4.67287362e-01
-6.71577215e-01 1.10866569e-01 8.74009848e-01 4.36606437e-01
6.09161735e-01 2.38504797e-01 -8.57580662e-01 -6.95442438e-01
-1.28093946e+00 2.00307205e-01 5.30644417e-01 5.84196448e-01
-1.56526379e-02 8.46144378e-01 -3.65618825e-01 1.20251024e+00
-8.74593183e-02 2.99702972e-01 8.17763269e-01 -1.26488447e+00
5.20268857e-01 -3.84962112e-01 -5.82223982e-02 -5.96814096e-01
-5.54510355e-01 -9.89615977e-01 -1.09235775e+00 -3.35843712e-01
-3.30671147e-02 -5.71260750e-01 -9.35089290e-01 2.10648394e+00
4.62561324e-02 7.07015634e-01 -3.83049287e-02 9.26687539e-01
8.13925028e-01 8.36803615e-01 -5.78079708e-02 -2.49016702e-01
1.03735232e+00 -1.25385523e+00 -1.20887995e+00 -2.83184767e-01
1.66586071e-01 -8.10997665e-01 7.88917303e-01 5.94792724e-01
-1.05086184e+00 -8.16180110e-01 -9.02272522e-01 1.22860231e-01
-1.06747851e-01 3.96103531e-01 -3.83816622e-02 6.13865674e-01
-1.22115719e+00 6.77680612e-01 -8.59688461e-01 -1.75953105e-01
5.50347753e-02 2.35342920e-01 -1.55618846e-01 2.72243351e-01
-1.74371982e+00 8.53902876e-01 2.77356476e-01 4.68472302e-01
-8.02463949e-01 -8.96090925e-01 -5.18167019e-01 6.59697875e-02
2.46685490e-01 -3.56270850e-01 1.61687148e+00 -5.51022649e-01
-1.92590880e+00 4.68260616e-01 -1.70536742e-01 -9.77177918e-01
4.58098680e-01 -5.47630727e-01 -1.15007675e+00 1.01228185e-01
-3.78069282e-02 3.80361855e-01 1.05460083e+00 -6.51921749e-01
-4.43603009e-01 4.00820114e-02 -5.18469274e-01 3.14269438e-02
-3.02500039e-01 4.38326001e-01 -2.99798280e-01 -1.41549850e+00
1.50973424e-01 -7.88669646e-01 -2.91782290e-01 -8.00058067e-01
-5.36019325e-01 -8.66929293e-02 9.41915870e-01 -1.38642192e+00
1.59055281e+00 -2.52535605e+00 -8.91677141e-02 5.86615205e-02
1.24444962e-01 8.95173490e-01 -5.69797456e-01 2.73602128e-01
-4.83031631e-01 9.05994885e-03 6.98976070e-02 -6.40042365e-01
1.79688916e-01 -1.01794876e-01 -6.36977911e-01 3.35734427e-01
4.47518863e-02 6.07472241e-01 -5.15752196e-01 3.13430458e-01
3.03783953e-01 7.46010661e-01 -2.80388296e-01 -2.10355520e-01
2.69906402e-01 5.59671700e-01 -1.74092427e-02 1.96337759e-01
8.90328228e-01 1.59786433e-01 1.39246685e-02 -4.40607697e-01
-3.87092799e-01 1.06151056e+00 -9.36479330e-01 1.78199053e+00
-3.42142075e-01 8.47836971e-01 6.06768548e-01 -8.31663430e-01
9.19581950e-01 7.84125090e-01 5.52316725e-01 -8.50672424e-01
1.23440064e-01 3.00961226e-01 8.75249431e-02 -2.78940111e-01
4.61955994e-01 -9.13447291e-02 1.57293633e-01 7.96256866e-03
3.31509382e-01 2.31610492e-01 -1.07458541e-02 7.10350573e-02
1.25418830e+00 -5.07101379e-02 4.68261838e-02 -2.13026911e-01
6.50608540e-01 -5.93544364e-01 9.89737689e-01 1.00925577e+00
-6.70349479e-01 7.85111666e-01 1.04361057e-01 -2.32836589e-01
-1.07046437e+00 -1.07286906e+00 -4.49025780e-02 1.22078073e+00
-4.29824919e-01 -5.01639724e-01 -9.35826182e-01 -5.32951117e-01
-4.10710216e-01 8.90215218e-01 -3.50433677e-01 -1.53304294e-01
-1.02063596e+00 -6.85841143e-01 1.25486743e+00 3.94185632e-01
6.27337396e-01 -1.05939603e+00 9.15536657e-02 5.90379655e-01
-7.15845942e-01 -1.40364718e+00 -1.01751208e+00 3.46436977e-01
-5.54719388e-01 -5.67189097e-01 -6.70922339e-01 -6.19648933e-01
-3.31790745e-01 2.52629429e-01 6.24538124e-01 -5.57131171e-01
1.96168181e-02 1.67971030e-01 -5.18679500e-01 -1.10564485e-01
-1.89617872e-01 3.66103232e-01 4.36619639e-01 5.03828488e-02
4.91861939e-01 -8.63587022e-01 -4.06144828e-01 1.61811322e-01
-6.44476414e-01 -3.05918425e-01 3.72884363e-01 9.05511916e-01
7.25162625e-02 1.57789052e-01 9.33300257e-01 -1.48471951e-01
7.50918925e-01 -2.60047615e-01 -5.28726101e-01 -2.28437576e-02
-2.11128443e-01 -1.90444306e-01 7.17850327e-01 -7.75913060e-01
-1.23351109e+00 -2.66959995e-01 -7.57106006e-01 -7.05942512e-01
1.53298765e-01 3.98028642e-01 -1.54098526e-01 4.83118147e-02
6.14503145e-01 5.20246506e-01 -2.92943984e-01 -1.00202990e+00
2.17815191e-01 7.35002160e-01 9.69315112e-01 -2.11396366e-01
7.90695608e-01 8.84487182e-02 -5.00497580e-01 -1.06153524e+00
-1.00558686e+00 -5.10832310e-01 -2.70302355e-01 1.24434009e-02
5.27847946e-01 -1.11756587e+00 -8.58311236e-01 8.05752158e-01
-1.42554688e+00 -2.65013844e-01 -2.47205839e-01 9.94160116e-01
-1.51472256e-01 3.41644198e-01 -1.19075835e+00 -9.49943304e-01
-7.93834090e-01 -9.82782662e-01 5.67846835e-01 1.04414327e-02
-2.31848329e-01 -5.74066222e-01 1.25347883e-01 3.43006819e-01
9.33540523e-01 -3.25248897e-01 2.84891188e-01 -8.48147571e-01
-1.36205763e-01 4.25720625e-02 -2.54718736e-02 1.13608241e+00
-1.35695592e-01 -2.89711565e-01 -1.30873930e+00 -3.01797122e-01
5.37961841e-01 1.06509000e-01 1.11941850e+00 8.33918750e-01
1.14092422e+00 -4.82221723e-01 1.49929464e-01 8.26732993e-01
5.72762191e-01 4.71681297e-01 7.34103560e-01 5.14321208e-01
5.28538883e-01 2.02022448e-01 6.97369128e-02 4.16003555e-01
1.56057715e-01 7.67438531e-01 -3.65628675e-02 4.71593067e-02
-4.90134269e-01 -9.15462375e-02 8.08514297e-01 1.56202328e+00
2.39033401e-01 -1.69382289e-01 -6.17054343e-01 4.45363104e-01
-1.53828001e+00 -1.16714907e+00 -1.38927147e-01 2.01103759e+00
9.46087956e-01 1.08678669e-01 2.20136225e-01 1.59528822e-01
1.03201008e+00 6.21886253e-01 -6.17787063e-01 -1.26023382e-01
-5.01542389e-01 4.02793348e-01 4.70483154e-01 4.63783890e-01
-1.33874464e+00 9.21306551e-01 6.34812355e+00 1.33900785e+00
-1.37027752e+00 4.10339683e-01 2.48495564e-01 -4.59863365e-01
2.40162805e-01 -3.99586409e-01 -7.70905912e-01 5.42589128e-01
1.45274925e+00 -2.30895057e-01 9.16281283e-01 5.71005285e-01
5.70139825e-01 6.51785135e-01 -4.01583612e-01 1.20686376e+00
-2.09753379e-01 -1.17362964e+00 -3.90218645e-01 -7.40798861e-02
4.54001963e-01 6.98018312e-01 3.80129933e-01 7.00401127e-01
2.11346984e-01 -8.04802775e-01 8.83634686e-01 3.44060749e-01
6.83711588e-01 -8.20426583e-01 6.98645592e-01 1.46058634e-01
-1.35075569e+00 -3.35435778e-01 -1.02139495e-01 2.56328285e-01
5.33404529e-01 7.37244308e-01 -7.26005793e-01 7.05189109e-01
7.49322653e-01 5.00703454e-01 -8.10555071e-02 1.10699272e+00
-2.24456519e-01 1.23812497e+00 -3.54485512e-01 5.56326210e-01
3.46340954e-01 4.66480590e-02 1.10124481e+00 1.38139069e+00
4.54516977e-01 5.72041348e-02 -1.01437375e-01 5.54893911e-01
-3.10632795e-01 -4.50756788e-01 1.27828360e-01 -1.30019143e-01
8.32247615e-01 1.04946566e+00 -1.62189052e-01 -3.61427635e-01
-1.28023446e-01 9.29212451e-01 -3.30430977e-02 7.30947733e-01
-8.94219398e-01 -5.80462933e-01 1.10274959e+00 -3.39019388e-01
5.65780461e-01 -3.46637219e-01 -1.44660845e-01 -1.16993904e+00
-1.97710544e-01 -1.12241435e+00 1.26773328e-01 -7.36359417e-01
-1.40881538e+00 1.16008997e+00 -5.07068694e-01 -9.85764980e-01
-2.21281856e-01 -3.24958980e-01 -4.88463551e-01 1.11694992e+00
-1.32511604e+00 -9.69796240e-01 2.12124377e-01 4.38380182e-01
6.83737636e-01 -3.07304919e-01 6.20244086e-01 8.75086129e-01
-8.62619638e-01 7.68276632e-01 2.74336487e-01 2.42416888e-01
7.34172761e-01 -8.24796557e-01 1.18494570e+00 1.25317860e+00
-1.16616122e-01 8.12104523e-01 7.28941262e-01 -5.72640538e-01
-8.26928258e-01 -1.16330934e+00 9.14253652e-01 -1.56354085e-02
8.90684783e-01 -4.46274400e-01 -1.05380154e+00 7.47047782e-01
2.59541690e-01 -7.77266026e-02 4.39335257e-01 8.88044313e-02
-5.76803148e-01 -2.29148418e-01 -7.31668234e-01 5.39460242e-01
1.01352143e+00 -1.04393089e+00 -7.02257276e-01 -8.91043097e-02
1.28906965e+00 -4.65911359e-01 -5.98799109e-01 5.56182623e-01
4.40226585e-01 -7.28487134e-01 1.22087657e+00 -4.88701701e-01
-3.30427527e-01 -2.51412272e-01 -4.13982302e-01 -1.45522761e+00
-4.17993426e-01 -1.30061400e+00 -1.47904232e-01 1.29361951e+00
1.62611648e-01 -5.76264024e-01 5.14519870e-01 -3.94467004e-02
-6.56552196e-01 -2.72228830e-02 -1.30455184e+00 -1.19471872e+00
-2.55390555e-01 -8.09391975e-01 4.64565367e-01 1.06457663e+00
-3.28286737e-01 2.11114690e-01 -8.53636026e-01 2.59855002e-01
5.44378519e-01 -7.20276237e-01 3.36554646e-01 -8.55340064e-01
-2.83454269e-01 -5.96888065e-01 1.91155300e-02 -1.28121495e+00
1.47392139e-01 -5.98441541e-01 1.13312539e-03 -1.01903570e+00
-4.53209519e-01 6.90168291e-02 -7.71453202e-01 3.79105747e-01
-1.12319626e-01 2.02934131e-01 4.09296304e-01 -8.40920582e-02
-4.59203929e-01 8.92807662e-01 6.97413564e-01 4.51141633e-02
-3.19499642e-01 2.02823684e-01 -4.45313513e-01 6.41760349e-01
9.84719515e-01 -4.33838516e-01 -1.39362976e-01 -2.53385186e-01
-3.84871513e-01 4.66723405e-02 3.89733285e-01 -1.03992093e+00
2.86646426e-01 2.03407675e-01 1.82617128e-01 -9.83974457e-01
7.54170001e-01 -3.10855836e-01 1.97177976e-01 4.24167365e-01
-4.60477203e-01 -2.11521357e-01 5.01395583e-01 2.48560250e-01
-4.19824094e-01 -6.98610395e-03 8.81723642e-01 1.91247225e-01
-5.10968506e-01 1.93037033e-01 -6.51373923e-01 2.18919247e-01
1.47213534e-01 2.89421976e-01 -4.91525680e-01 -4.70963627e-01
-1.00997174e+00 -6.54986948e-02 -4.11966294e-01 4.64054912e-01
3.77550721e-01 -1.53616703e+00 -8.82099748e-01 3.16374362e-01
-4.56204653e-01 -6.38431311e-01 1.00218403e+00 1.02062321e+00
-1.03915110e-01 6.37283146e-01 7.53945410e-02 -1.20287791e-01
-1.02638721e+00 2.23823503e-01 4.56341118e-01 -3.04556668e-01
-7.51670241e-01 1.19547129e+00 1.25340730e-01 -8.28254342e-01
6.46190464e-01 -1.36276707e-01 2.87164003e-02 -1.24193586e-01
9.89951551e-01 8.56671154e-01 4.08542305e-01 -5.68206370e-01
-5.49653471e-01 -5.60749089e-03 -1.09004974e-01 -5.51113904e-01
1.51756763e+00 -4.39796925e-01 -9.69486460e-02 4.33731079e-01
1.26823878e+00 1.82526648e-01 -1.17658913e+00 -5.95477462e-01
2.19586939e-01 -3.15701365e-02 4.40716803e-01 -1.10729492e+00
-9.84924138e-01 8.26648891e-01 6.20061815e-01 1.65992379e-01
1.39290571e+00 -6.37198150e-01 1.31823969e+00 1.64402068e-01
-9.53182429e-02 -1.18413460e+00 -9.62079763e-02 1.19474030e+00
1.01622927e+00 -6.29314244e-01 -5.55131972e-01 -2.57936735e-02
-4.62672859e-01 1.05982339e+00 3.67183864e-01 1.30929694e-01
5.66119969e-01 4.37821269e-01 1.73072785e-01 3.38109821e-01
-7.39285111e-01 -4.77260053e-02 2.94173211e-01 5.45308232e-01
2.33424351e-01 -9.37486440e-02 5.36571592e-02 1.08798099e+00
-5.18823326e-01 -1.07657589e-01 2.60495454e-01 3.70777816e-01
-3.03817958e-01 -1.08165431e+00 -6.25756681e-01 9.90472063e-02
-6.61405623e-01 -5.92154384e-01 -1.37089744e-01 3.73407543e-01
-6.01071455e-02 1.08585691e+00 -2.22728923e-01 -8.36108506e-01
4.14573759e-01 3.08405310e-01 3.62871028e-02 -1.06500432e-01
-9.67359245e-01 6.92736506e-01 2.20900893e-01 -6.50364995e-01
-2.82199960e-02 -6.33861005e-01 -8.91199589e-01 -6.21737003e-01
-6.26939163e-02 2.67277390e-01 9.14477289e-01 7.57717550e-01
5.83946884e-01 1.04080045e+00 6.60103798e-01 -8.64938498e-01
-8.71672928e-01 -1.33823025e+00 -7.12295175e-01 -5.91488779e-02
9.15550411e-01 -2.07210004e-01 -4.52391893e-01 -6.50147349e-02] | [14.88580322265625, 5.991230487823486] |
4994af7e-f532-4c6f-b385-019fe591e898 | anchorformer-point-cloud-completion-from | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chen_AnchorFormer_Point_Cloud_Completion_From_Discriminative_Nodes_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_AnchorFormer_Point_Cloud_Completion_From_Discriminative_Nodes_CVPR_2023_paper.pdf | AnchorFormer: Point Cloud Completion From Discriminative Nodes | Point cloud completion aims to recover the completed 3D shape of an object from its partial observation. A common strategy is to encode the observed points to a global feature vector and then predict the complete points through a generative process on this vector. Nevertheless, the results may suffer from the high-quality shape generation problem due to the fact that a global feature vector cannot sufficiently characterize diverse patterns in one object. In this paper, we present a new shape completion architecture, namely AnchorFormer, that innovatively leverages pattern-aware discriminative nodes, i.e., anchors, to dynamically capture regional information of objects. Technically, AnchorFormer models the regional discrimination by learning a set of anchors based on the point features of the input partial observation. Such anchors are scattered to both observed and unobserved locations through estimating particular offsets, and form sparse points together with the down-sampled points of the input observation. To reconstruct the fine-grained object patterns, AnchorFormer further employs a modulation scheme to morph a canonical 2D grid at individual locations of the sparse points into a detailed 3D structure. Extensive experiments on the PCN, ShapeNet-55/34 and KITTI datasets quantitatively and qualitatively demonstrate the efficacy of AnchorFormer over the state-of-the-art point cloud completion approaches. Source code is available at https://github.com/chenzhik/AnchorFormer. | ['Tao Mei', 'Jiebo Luo', 'Wengang Zhou', 'Ting Yao', 'Zhaofan Qiu', 'Fuchen Long', 'Zhikai Chen'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['point-cloud-completion'] | ['computer-vision'] | [ 2.82866904e-03 -2.21839041e-01 7.52739375e-03 -2.85864532e-01
-8.47435832e-01 -5.97378969e-01 4.05822307e-01 -1.16305538e-01
4.13178474e-01 1.87781364e-01 8.31690505e-02 3.90270561e-01
-2.13677332e-01 -8.36760700e-01 -8.05915892e-01 -8.35674882e-01
1.50197238e-01 9.12328184e-01 8.11802745e-02 3.66870873e-02
2.67714947e-01 1.04777384e+00 -1.45114315e+00 1.80534512e-01
7.35328734e-01 1.07267535e+00 5.12635469e-01 1.09039076e-01
-2.65734613e-01 -1.25855818e-01 -3.24840724e-01 1.31917357e-01
4.43620890e-01 9.32894349e-02 -4.24415357e-02 4.52165931e-01
4.81570035e-01 -1.65476188e-01 -4.91953641e-01 8.97442758e-01
1.99827865e-01 2.47133467e-02 8.07871640e-01 -1.16290855e+00
-7.82402635e-01 4.16259794e-03 -8.70657384e-01 -2.24440426e-01
1.34486094e-01 2.36347124e-01 8.76599789e-01 -1.55518770e+00
6.44531012e-01 1.08705950e+00 6.28083229e-01 2.05651447e-01
-1.32640207e+00 -8.27958643e-01 1.91244647e-01 -7.44332224e-02
-1.74207318e+00 -3.69865626e-01 1.35538125e+00 -5.49431682e-01
4.02263492e-01 1.05725810e-01 6.96202993e-01 7.68993318e-01
9.47638825e-02 6.26922846e-01 7.15904117e-01 4.62094620e-02
2.57836521e-01 -2.96320111e-01 -2.07964405e-01 6.38903022e-01
2.72563875e-01 5.74001968e-02 -7.07623720e-01 -5.19654632e-01
1.33922338e+00 7.21506000e-01 -4.21575606e-01 -9.26034689e-01
-1.31751132e+00 6.29617572e-01 7.25581825e-01 1.01179555e-01
-7.65344381e-01 5.42421341e-02 -3.62835020e-01 -1.83114067e-01
3.94294590e-01 7.90850967e-02 -5.00451267e-01 2.17209369e-01
-8.69704306e-01 4.18477982e-01 4.03584927e-01 1.41357028e+00
1.44247830e+00 8.91577303e-02 -1.64622799e-01 8.04584503e-01
6.76194310e-01 9.04708564e-01 1.18016340e-02 -1.12322581e+00
6.18020594e-01 9.76368129e-01 9.45453867e-02 -1.25567222e+00
-1.44017249e-01 -7.41481304e-01 -1.05483735e+00 2.35102251e-01
1.06989570e-01 2.58285135e-01 -1.03288472e+00 1.44084680e+00
6.09319806e-01 5.17909527e-01 -3.43331695e-01 9.47648585e-01
7.36463308e-01 7.15127647e-01 -5.12065470e-01 3.69903147e-02
1.08212566e+00 -5.96073806e-01 -2.25972205e-01 -1.08918317e-01
-8.48899409e-02 -6.54995263e-01 8.67706716e-01 2.05912456e-01
-1.07764208e+00 -6.04180157e-01 -8.89474988e-01 8.51977989e-02
9.36979130e-02 3.00102562e-01 4.65441972e-01 2.89121512e-02
-7.02411473e-01 4.40304220e-01 -9.91339087e-01 2.54115105e-01
7.44741559e-01 1.79684445e-01 -4.02849644e-01 -4.06770319e-01
-3.68810207e-01 1.21500716e-01 -1.66054100e-01 1.72154963e-01
-9.85510647e-01 -1.21689725e+00 -7.58596480e-01 1.03777409e-01
2.58131683e-01 -8.61421347e-01 7.81996727e-01 -3.85525614e-01
-9.62608814e-01 5.60740829e-01 -4.81974274e-01 1.89649805e-01
6.56706467e-02 1.20933935e-01 -4.27329987e-02 1.61419258e-01
4.23422754e-01 5.28378010e-01 1.05335927e+00 -1.82478559e+00
-4.89844203e-01 -7.95050681e-01 -3.98238212e-01 1.27219856e-01
1.37236893e-01 -6.32904112e-01 -7.60062873e-01 -8.34198892e-01
9.54872191e-01 -9.55805779e-01 -3.43388051e-01 3.47429365e-01
-2.62491584e-01 -1.57672957e-01 9.67762411e-01 -3.45897168e-01
7.30597317e-01 -2.38896966e+00 2.12996349e-01 5.22698164e-01
4.48837161e-01 -2.23315820e-01 -3.01735938e-01 4.97736067e-01
6.46184310e-02 -1.80849582e-01 -5.16682327e-01 -7.29327679e-01
-3.51752229e-02 4.03024852e-01 -3.96758884e-01 6.58725917e-01
1.87953547e-01 1.02631938e+00 -9.64828610e-01 -1.39193073e-01
3.56751889e-01 7.59402871e-01 -6.47237301e-01 1.23267531e-01
-2.81558394e-01 7.22348571e-01 -8.59763205e-01 1.15511417e+00
1.09073603e+00 -3.06710750e-01 -3.59901905e-01 -4.04322296e-01
-3.30795608e-02 -2.21797436e-01 -1.32439733e+00 2.33294773e+00
-1.57738879e-01 1.25506580e-01 2.68942714e-01 -6.70955002e-01
1.20107114e+00 1.55231446e-01 8.96152377e-01 -2.13223934e-01
-1.84772357e-01 2.89768845e-01 -3.70748490e-01 -1.28951669e-01
4.13187027e-01 -1.75072812e-02 -3.27155888e-02 9.52748209e-02
1.73373930e-02 -2.44005457e-01 -4.86258000e-01 5.93937412e-02
1.05917466e+00 1.60524383e-01 -1.56722125e-02 -1.26612908e-03
3.37374598e-01 4.55594622e-02 8.29407632e-01 5.10559618e-01
1.06417760e-01 1.27105761e+00 -5.35306195e-03 -3.32601219e-01
-1.06107199e+00 -1.43407965e+00 -2.31681988e-01 4.25959706e-01
2.79022187e-01 -3.43056530e-01 -1.97995260e-01 -5.01143754e-01
4.23422247e-01 3.56148720e-01 -5.95827341e-01 -1.23936813e-02
-6.11275971e-01 -9.26206559e-02 1.43789651e-03 4.85218614e-01
2.27423668e-01 -9.48274314e-01 -3.08926791e-01 2.54457235e-01
-1.42748982e-01 -8.54104459e-01 -5.94332516e-01 -4.81151342e-02
-1.10842848e+00 -9.07514393e-01 -6.37340248e-01 -6.88812494e-01
1.05760193e+00 7.43654072e-01 8.97426605e-01 1.19665027e-01
-1.07180990e-01 4.99897629e-01 -4.22542691e-01 -1.17391452e-01
7.07878694e-02 -5.03812209e-02 -1.45653868e-02 3.38683605e-01
2.06385940e-01 -1.17996979e+00 -7.97936499e-01 4.24781829e-01
-8.29951406e-01 4.66019921e-02 8.78450513e-01 7.44586468e-01
1.36828613e+00 -2.17233144e-04 3.02203625e-01 -5.95606446e-01
-7.38026500e-02 -7.64094114e-01 -4.36036617e-01 -6.88145012e-02
-1.50963336e-01 -1.30059689e-01 3.23531628e-01 -3.33812237e-01
-7.24022269e-01 4.45292383e-01 6.17743023e-02 -1.33519816e+00
-2.89301068e-01 4.08298880e-01 -5.03931344e-01 -1.22907586e-01
3.50944728e-01 5.15993476e-01 4.39436473e-02 -9.10845518e-01
2.76056588e-01 2.22380802e-01 5.46966672e-01 -8.37218940e-01
1.26933599e+00 8.16214025e-01 2.61328042e-01 -7.35628366e-01
-5.46950340e-01 -7.76270032e-01 -7.06318617e-01 -1.70409784e-01
4.14813817e-01 -1.10504675e+00 -3.87392104e-01 3.61679941e-01
-1.09522748e+00 -1.49245620e-01 -5.86134732e-01 1.89956710e-01
-5.22125602e-01 3.34375709e-01 -2.04434842e-01 -4.46514815e-01
-8.88427943e-02 -8.85702670e-01 1.56825793e+00 1.10870846e-01
1.20576672e-01 -6.93374932e-01 6.30965233e-02 2.34476775e-01
1.40255615e-01 4.77486253e-01 7.10195780e-01 -1.68828949e-01
-1.22315216e+00 -3.67885321e-01 -3.84102352e-02 7.34068900e-02
3.74040276e-01 -3.01813781e-02 -7.15361834e-01 -4.70260471e-01
1.83914199e-01 1.06388718e-01 5.34500241e-01 5.35530508e-01
1.29363000e+00 -1.78292319e-01 -5.24406910e-01 8.75349343e-01
1.63298893e+00 -1.17398590e-01 4.06788319e-01 -1.36753455e-01
1.09436607e+00 2.99591869e-01 6.29759014e-01 7.25035191e-01
4.59119976e-01 7.74723649e-01 7.27963686e-01 1.73576549e-01
-2.73634136e-01 -6.98569775e-01 -8.95558521e-02 8.37657273e-01
-2.22099230e-01 8.10726359e-02 -8.76722395e-01 5.74109793e-01
-1.78454888e+00 -6.95126951e-01 -2.04211473e-01 2.08833361e+00
3.78862411e-01 -2.24105328e-01 -2.52160370e-01 -1.05138794e-01
6.79791331e-01 1.65993929e-01 -7.63652563e-01 6.50993168e-01
-1.47241592e-01 2.34864414e-01 3.49226713e-01 3.48082036e-01
-5.76244712e-01 6.91051424e-01 4.81178188e+00 9.46664929e-01
-9.03372824e-01 1.26080289e-01 2.45163664e-01 -3.06863617e-02
-6.59408033e-01 1.96919352e-01 -7.69004643e-01 4.94582921e-01
1.83983505e-01 1.02116995e-01 3.99776906e-01 8.51144373e-01
1.45972982e-01 2.34137133e-01 -9.15536106e-01 1.20486319e+00
8.10461864e-03 -1.52073514e+00 3.43486786e-01 4.19661224e-01
1.00007558e+00 2.64738977e-01 1.09017856e-01 4.91349772e-02
-1.31906196e-02 -7.58845150e-01 8.80057454e-01 8.75973642e-01
8.19704354e-01 -5.91662884e-01 2.76525080e-01 6.36209071e-01
-1.49467206e+00 9.67619345e-02 -7.35328197e-01 6.79875761e-02
1.09533466e-01 6.84807837e-01 -6.31024003e-01 7.65720010e-01
7.32465327e-01 1.03931856e+00 -3.96690458e-01 1.18091011e+00
-1.29202545e-01 4.41371232e-01 -5.65411806e-01 4.46685702e-01
8.81020129e-02 -4.75721955e-01 9.57848966e-01 6.26392961e-01
7.04877079e-01 4.08133656e-01 3.25184047e-01 1.23001862e+00
2.22423556e-03 -1.36704460e-01 -5.75436711e-01 4.21097189e-01
9.44712758e-01 1.48891068e+00 -6.41618431e-01 -1.27970725e-01
-2.55901903e-01 6.24709129e-01 3.19979846e-01 5.52045345e-01
-4.14601386e-01 1.68327451e-01 9.04138625e-01 5.08243620e-01
7.51333594e-01 -6.01937830e-01 -6.82157278e-01 -1.12871063e+00
3.70167226e-01 -5.66060960e-01 -2.93127727e-03 -9.62092638e-01
-1.42456698e+00 3.62762481e-01 -2.02468410e-02 -1.73552060e+00
1.12379856e-01 -2.92843044e-01 -7.13216960e-01 1.05320144e+00
-1.26272559e+00 -1.33137894e+00 -7.56560385e-01 8.01383913e-01
5.91567099e-01 -2.00554386e-01 6.04083717e-01 1.51600391e-01
-6.81876093e-02 1.99610263e-01 1.87396094e-01 -1.42514855e-02
3.37605894e-01 -8.76505315e-01 3.78871828e-01 6.03776753e-01
3.39899153e-01 6.33667707e-01 3.28324676e-01 -8.62944245e-01
-1.72916770e+00 -1.22448146e+00 3.95367503e-01 -5.96670389e-01
3.04389119e-01 -4.02535796e-01 -1.12674880e+00 6.44306719e-01
-3.37266624e-01 4.54370111e-01 3.66641611e-01 -2.01446310e-01
-2.43638173e-01 -2.36277431e-01 -1.15768135e+00 3.68620872e-01
1.31297243e+00 -3.07153434e-01 -5.27304113e-01 6.42446727e-02
6.65897727e-01 -5.99742353e-01 -9.32599306e-01 5.32787323e-01
3.89235884e-01 -8.34926963e-01 1.26109195e+00 -8.46719295e-02
4.66010123e-01 -7.90664017e-01 -4.74960864e-01 -1.34920871e+00
-8.68422985e-01 -2.94175446e-01 -3.31860662e-01 1.35655034e+00
6.78495616e-02 -7.02037454e-01 1.16747642e+00 3.95318061e-01
-5.89907408e-01 -9.83456254e-01 -1.09939170e+00 -5.25404274e-01
-7.42942840e-02 -4.48333621e-01 1.14598632e+00 8.15592110e-01
-6.62911236e-01 1.04130737e-01 -6.59866631e-02 6.62153959e-01
1.03277063e+00 5.76827347e-01 9.96193826e-01 -1.36419952e+00
-1.34202987e-01 -2.38131642e-01 -6.22632146e-01 -1.47680414e+00
1.39771104e-02 -1.02640057e+00 3.58828865e-02 -1.50984013e+00
1.22234540e-03 -1.04244626e+00 -1.21005900e-01 4.83770370e-01
-8.72221366e-02 3.56039762e-01 2.11934373e-01 7.78294623e-01
-1.79582670e-01 9.77715254e-01 1.60189867e+00 -1.99252248e-01
-1.71677768e-01 1.75746620e-01 -6.78254008e-01 6.04782999e-01
5.68052888e-01 -5.92431605e-01 -3.02501708e-01 -6.00223482e-01
-1.45293605e-02 2.72381186e-01 6.54719889e-01 -9.98049021e-01
3.87235820e-01 -2.93163478e-01 7.75851369e-01 -1.34378481e+00
7.63562679e-01 -1.24457252e+00 7.10713625e-01 -1.78406946e-02
2.11681738e-01 -7.87105188e-02 -1.00448333e-01 9.07610416e-01
-1.48014531e-01 4.00019027e-02 5.06376863e-01 -1.54552758e-01
-3.93180102e-01 1.07912397e+00 2.87781477e-01 -2.20772788e-01
9.03012693e-01 -5.19499838e-01 -9.13639516e-02 -4.81860302e-02
-6.00279570e-01 2.64918417e-01 9.48317170e-01 3.83253545e-01
1.07575810e+00 -1.70459700e+00 -8.15472424e-01 7.94823110e-01
3.31836045e-01 8.96710813e-01 5.67421436e-01 6.60483837e-01
-2.40119010e-01 9.99110341e-02 -6.86093047e-02 -1.11695516e+00
-7.91387796e-01 3.81402850e-01 5.73264435e-02 1.44164145e-01
-1.01427698e+00 6.46085620e-01 4.96065944e-01 -4.75707889e-01
-3.72389182e-02 -4.08667862e-01 1.62730694e-01 -2.22398877e-01
2.57628649e-01 2.12339833e-01 7.22640529e-02 -9.01695907e-01
-2.90761530e-01 1.09511590e+00 2.27400005e-01 9.91089046e-02
1.56299174e+00 -3.58361937e-02 -1.15773775e-01 4.12933499e-01
1.16339195e+00 3.14226508e-01 -1.76789451e+00 -6.01102173e-01
-4.49792475e-01 -1.01601326e+00 -8.16061497e-02 -3.20492476e-01
-1.40404689e+00 5.40423334e-01 2.48598009e-01 -1.50787249e-01
9.18856919e-01 3.89363557e-01 6.37825906e-01 4.63293344e-02
7.68829882e-01 -4.13830847e-01 7.58490041e-02 2.30823889e-01
1.26188922e+00 -9.07969832e-01 1.43398615e-02 -5.82287431e-01
-3.25145602e-01 9.23499346e-01 4.51812863e-01 -6.07472181e-01
9.45423543e-01 4.07592915e-02 -1.99076474e-01 -6.40454888e-01
-5.57252467e-01 6.98125213e-02 4.31275129e-01 7.75674760e-01
-3.32976818e-01 3.39741200e-01 3.58140111e-01 7.23121464e-01
-1.56859025e-01 -2.18975887e-01 1.28140122e-01 7.82965481e-01
-4.19900954e-01 -9.32390153e-01 -7.88497984e-01 5.92315137e-01
1.40160158e-01 2.01551557e-01 -1.14969403e-01 4.88095343e-01
2.01773062e-01 5.11466146e-01 1.94361612e-01 -3.87183756e-01
5.43702066e-01 -2.87620425e-01 3.35053772e-01 -8.92266989e-01
-8.46778676e-02 3.38749945e-01 -6.27648234e-01 -5.82386911e-01
-2.94932693e-01 -1.04264903e+00 -1.29382563e+00 -1.89048201e-01
-2.27188319e-02 -8.88388418e-03 6.43800437e-01 5.18404901e-01
6.98871374e-01 2.44135812e-01 9.83527005e-01 -1.56395209e+00
-4.36931431e-01 -8.27151060e-01 -8.30495656e-01 3.42374325e-01
3.63118529e-01 -9.86322045e-01 -5.25275350e-01 -1.13380648e-01] | [8.393777847290039, -3.5532195568084717] |
15c3920c-7c53-4d03-8c90-f524ded13b1e | pathways-asynchronous-distributed-dataflow | 2203.12533 | null | https://arxiv.org/abs/2203.12533v1 | https://arxiv.org/pdf/2203.12533v1.pdf | Pathways: Asynchronous Distributed Dataflow for ML | We present the design of a new large scale orchestration layer for accelerators. Our system, Pathways, is explicitly designed to enable exploration of new systems and ML research ideas, while retaining state of the art performance for current models. Pathways uses a sharded dataflow graph of asynchronous operators that consume and produce futures, and efficiently gang-schedules heterogeneous parallel computations on thousands of accelerators while coordinating data transfers over their dedicated interconnects. Pathways makes use of a novel asynchronous distributed dataflow design that lets the control plane execute in parallel despite dependencies in the data plane. This design, with careful engineering, allows Pathways to adopt a single-controller model that makes it easier to express complex new parallelism patterns. We demonstrate that Pathways can achieve performance parity (~100% accelerator utilization) with state-of-the-art systems when running SPMD computations over 2048 TPUs, while also delivering throughput comparable to the SPMD case for Transformer models that are pipelined across 16 stages, or sharded across two islands of accelerators connected over a data center network. | ['Yonghui Wu', 'Chandramohan A. Thekkath', 'Laurent El Shafey', 'Ryan Sepassi', 'Parker Schuh', 'Brennan Saeta', 'Sudip Roy', 'Ruoming Pang', 'Hyeontaek Lim', 'Michael Isard', 'Dan Hurt', 'Steven Hand', 'Sanjay Ghemawat', 'Jeff Dean', 'Aakanksha Chowdhery', 'Paul Barham'] | 2022-03-23 | null | null | null | null | ['2048'] | ['playing-games'] | [-6.36546791e-01 8.51021856e-02 -5.12518525e-01 -3.34742576e-01
1.33323833e-01 -5.78951240e-01 7.98960984e-01 3.55585843e-01
-1.74519420e-01 2.69792050e-01 5.75515807e-01 -9.42604423e-01
1.56449944e-01 -9.55372036e-01 -3.72787803e-01 -4.30148691e-01
-6.65989876e-01 4.98104990e-01 6.62073553e-01 -3.60574603e-01
8.51353556e-02 9.56874907e-01 -1.41383696e+00 8.09864223e-01
1.12621114e-01 1.01751137e+00 -2.70283222e-01 7.48574257e-01
-2.53922194e-01 9.42455709e-01 -7.08754897e-01 3.60263109e-01
3.51875722e-01 -5.35295121e-02 -7.71451294e-01 -4.22029436e-01
2.15069965e-01 -3.29069823e-01 -7.70796299e-01 2.16473654e-01
3.18597943e-01 -3.64117801e-01 -1.45461902e-01 -1.38546216e+00
4.19997312e-02 8.02287221e-01 -6.18798077e-01 7.43256450e-01
-3.42932269e-02 8.62755775e-01 8.04190040e-01 -5.41340351e-01
8.85890365e-01 9.70447779e-01 4.99515057e-01 6.62988350e-02
-1.73348486e+00 -5.52151978e-01 -1.14203714e-01 -5.99123597e-01
-1.03401065e+00 -8.46452773e-01 -2.56783128e-01 -3.14247817e-01
1.60464120e+00 4.00639802e-01 9.19340074e-01 5.30270934e-01
1.24745739e+00 1.84938088e-01 8.34571719e-01 -1.29815772e-01
5.97521186e-01 -2.68646002e-01 5.96399844e-01 5.67579865e-01
2.72111446e-01 2.32557237e-01 -1.17143905e+00 -3.89674664e-01
9.97507691e-01 -2.25335419e-01 -3.34398784e-02 -1.10264361e-01
-1.85900617e+00 5.09523153e-01 7.20709443e-01 4.39213395e-01
-2.34546930e-01 4.88644212e-01 1.13014758e+00 5.65179050e-01
2.76442260e-01 7.75315166e-01 -4.09849614e-01 -1.00345738e-01
-1.36209905e+00 5.49737751e-01 1.16717255e+00 1.40796471e+00
2.98651665e-01 1.69964418e-01 -7.00913846e-01 -4.28794563e-01
2.52586186e-01 2.72451013e-01 2.02181101e-01 -7.18766332e-01
2.28326097e-01 4.70858067e-01 -1.12191424e-01 -3.12009513e-01
-1.09420085e+00 -6.50308847e-01 -8.38243663e-01 6.31397724e-01
2.94023544e-01 -2.30041578e-01 -4.03666347e-01 1.11221635e+00
4.25043017e-01 -3.36523466e-02 -4.33424339e-02 9.88654315e-01
4.62636471e-01 1.07788587e+00 1.94712266e-01 1.50998861e-01
1.66915679e+00 -1.45083189e+00 -7.15847433e-01 1.87212005e-01
9.86866534e-01 -1.06588697e+00 6.93555534e-01 3.42758745e-01
-1.43481648e+00 -6.83220029e-01 -1.33782363e+00 -4.09841508e-01
-1.50810912e-01 -4.00058955e-01 1.22393560e+00 5.00182390e-01
-1.67747200e+00 4.65176493e-01 -1.57710171e+00 -2.38026768e-01
1.10205647e-03 6.11532927e-01 -1.17557757e-01 4.73122716e-01
-4.65231508e-01 3.06318194e-01 3.80122453e-01 -1.86742440e-01
-1.06310809e+00 -1.66799784e+00 -2.60177284e-01 3.71914059e-01
-2.18529224e-01 -1.17725992e+00 1.43373859e+00 -4.10466194e-01
-1.62928581e+00 4.13329601e-01 -2.02649981e-02 -7.83800066e-01
-6.71549886e-02 3.01277548e-01 -8.18643391e-01 -2.06129178e-01
-5.50410040e-02 5.84031641e-01 3.39668065e-01 -2.28333160e-01
-6.25845909e-01 -1.48895681e-01 -1.58122137e-01 -1.71941638e-01
-1.88598216e-01 4.10498679e-01 -8.37504864e-02 -1.41892627e-01
-1.93343878e-01 -8.30088556e-01 -5.47179401e-01 8.78957063e-02
-3.71163130e-01 1.14730299e-01 8.84180665e-01 2.51009375e-01
1.27838564e+00 -2.21025443e+00 4.68575917e-02 2.59159118e-01
6.63009346e-01 -1.42458037e-01 3.43043387e-01 7.39262104e-01
1.35711133e-01 -1.88551113e-01 4.27199185e-01 -1.89106703e-01
1.97134271e-01 1.31257698e-01 -6.86917603e-01 4.42014635e-01
-1.05160894e-02 4.15738434e-01 -7.33836532e-01 2.51043327e-02
-1.74919412e-01 5.74997701e-02 -9.28429365e-01 2.30135843e-01
-3.13794494e-01 5.96328974e-01 -6.14099920e-01 4.74057764e-01
9.70094323e-01 -6.85888529e-01 4.32478905e-01 -1.90603286e-01
-1.38700271e+00 7.08488822e-01 -9.57340419e-01 2.21042538e+00
-6.23550475e-01 6.43244147e-01 6.95723176e-01 -1.64390951e-01
5.24632454e-01 2.75988072e-01 5.04592657e-01 -7.43515432e-01
-1.06979564e-01 2.72060871e-01 1.52410954e-01 3.43815327e-01
1.13064277e+00 5.72736673e-02 -4.29716170e-01 8.68186116e-01
-7.85040669e-03 -3.00520211e-01 3.75126451e-01 6.29146397e-01
1.94501865e+00 -2.57179946e-01 -4.27410364e-01 -1.37564218e+00
2.20303923e-01 7.69886613e-01 4.16584253e-01 7.80533195e-01
-9.83336791e-02 -3.55205417e-01 9.96131659e-01 -9.99341011e-01
-1.38400185e+00 -1.44536757e+00 -5.43800473e-01 1.09611893e+00
-2.11407885e-01 -1.57664466e+00 -3.12075436e-01 -4.01388630e-02
1.69540733e-01 6.77363157e-01 2.10101828e-01 2.74229378e-01
-6.97134137e-01 -9.27368879e-01 5.57189167e-01 4.32820469e-01
3.06638658e-01 -3.87506366e-01 -9.76345599e-01 6.67143106e-01
8.91479433e-01 -8.78473878e-01 -4.99910653e-01 5.35535395e-01
-9.84186172e-01 -4.34500337e-01 2.50249535e-01 -3.57611626e-01
5.83980978e-01 1.88025579e-01 1.51995850e+00 -6.59949183e-02
-5.54454327e-01 -2.89545178e-01 1.75151914e-01 -1.54226786e-02
-5.48924088e-01 3.54973078e-01 8.99576545e-02 -4.80050594e-01
-1.56870149e-02 -7.46212065e-01 -7.48934865e-01 4.04445976e-01
-8.11482430e-01 6.11839652e-01 1.80971622e-01 6.95029855e-01
6.44154251e-01 -4.08384725e-02 -9.39091817e-02 -9.26588595e-01
3.45952511e-01 -6.51920974e-01 -1.21828115e+00 -6.17896318e-01
-5.31016469e-01 4.76190865e-01 1.22583902e+00 2.23987907e-01
-1.11811733e+00 -9.91680566e-03 -1.14860818e-01 -1.51435748e-01
2.57378936e-01 1.89734623e-03 2.99892455e-01 -3.43896270e-01
7.88237095e-01 -3.77712011e-01 4.06015635e-01 -1.60417497e-01
4.90309387e-01 1.60505101e-01 2.27238372e-01 -9.35913444e-01
2.52093613e-01 7.80702591e-01 5.61317146e-01 -5.88930488e-01
3.43999006e-02 -4.94264103e-02 -3.57171804e-01 4.72880015e-03
7.59670496e-01 -1.25440168e+00 -1.23268986e+00 2.26883471e-01
-1.07340944e+00 -7.49964118e-01 -4.31400716e-01 2.18953922e-01
-2.60136038e-01 -5.50030291e-01 -1.45020413e+00 6.62764907e-02
-3.89667451e-01 -1.55376136e+00 1.22036648e+00 3.18462402e-01
-6.16563439e-01 -6.95787311e-01 -1.63714901e-01 -5.15057325e-01
1.42982185e+00 1.56964839e-01 9.09016430e-01 -1.81915134e-01
-1.35278833e+00 4.43092018e-01 -3.86484087e-01 -4.98324424e-01
-3.84903729e-01 3.56726736e-01 -5.94114125e-01 -5.33525467e-01
-5.10412455e-02 4.35116850e-02 4.27071184e-01 1.91620454e-01
1.14674079e+00 3.73279606e-03 -9.32328582e-01 9.39640641e-01
1.38195491e+00 -2.22141653e-01 3.28806788e-01 8.10528845e-02
2.62752652e-01 4.34338190e-02 -2.97223907e-02 7.39455581e-01
-4.14550900e-02 7.96439290e-01 2.98434138e-01 -2.91075140e-01
-1.23663396e-01 2.11959124e-01 4.91377652e-01 1.22053587e+00
4.45160985e-01 -1.01537243e-01 -1.14670503e+00 -1.64806396e-01
-1.57534361e+00 -7.31747568e-01 -7.86874533e-01 2.04661441e+00
5.36318004e-01 5.99100411e-01 4.78381626e-02 -4.20869231e-01
-1.04451351e-01 2.20291391e-01 -3.87119085e-01 -1.18336785e+00
2.20142394e-01 8.81999075e-01 1.14900303e+00 2.88281143e-01
-5.34680724e-01 7.53580093e-01 7.51379585e+00 5.75117648e-01
-1.27847028e+00 5.61434150e-01 6.86834514e-01 -5.90444386e-01
-2.81520873e-01 3.36884111e-01 -1.28537130e+00 6.07299089e-01
1.96899629e+00 -5.69865704e-01 1.62062705e-01 6.21830225e-01
5.61450481e-01 -2.44118273e-01 -1.44605660e+00 3.51901710e-01
-7.30826616e-01 -2.05934644e+00 -2.83586532e-01 3.41397166e-01
6.43160522e-01 5.07524490e-01 -3.92035730e-02 1.30860478e-01
7.88219154e-01 -7.45414615e-01 9.28555012e-01 5.33143759e-01
4.94871199e-01 -8.08704138e-01 1.96386889e-01 4.07599052e-03
-1.27048230e+00 -4.00154404e-02 5.93154225e-04 -4.65039760e-01
4.52525049e-01 9.33352530e-01 -5.82352936e-01 6.52386546e-01
5.43515146e-01 3.31705868e-01 -2.05155671e-01 8.19295287e-01
3.26744646e-01 4.78822798e-01 -4.80733484e-01 5.16672805e-02
4.10932451e-01 6.22320026e-02 3.01314384e-01 1.26581919e+00
2.03093052e-01 -1.80873409e-01 2.07566366e-01 1.06161940e+00
1.77384779e-01 -3.81317556e-01 -4.20631975e-01 1.81981027e-01
5.18796921e-01 1.54833126e+00 -1.00976169e+00 -5.34175038e-01
-7.27927685e-01 6.91449583e-01 1.34934530e-01 -8.25086311e-02
-8.35829616e-01 -2.77651101e-01 1.34541404e+00 4.27744776e-01
-1.12999566e-01 -9.76036668e-01 -6.34893239e-01 -9.29537356e-01
-5.05706847e-01 -6.98386073e-01 1.97489440e-01 -3.76651615e-01
-1.20111108e+00 9.51942742e-01 -3.61400634e-01 -6.56440437e-01
1.55368641e-01 -4.65725064e-01 -7.98913002e-01 1.10049093e+00
-8.83020222e-01 -7.07750976e-01 -9.79842693e-02 4.58568335e-01
1.37886390e-01 -5.69087937e-02 8.46884489e-01 4.41938013e-01
-5.11332333e-01 1.80007264e-01 2.38386601e-01 -5.43067575e-01
6.33894384e-01 -9.30921674e-01 1.22298181e+00 7.78958321e-01
-5.70761383e-01 8.01724613e-01 7.69727647e-01 -5.41505635e-01
-2.64347315e+00 -7.97622919e-01 5.93398929e-01 -3.42152178e-01
1.14685261e+00 -8.80341887e-01 -3.63641828e-01 8.73212755e-01
8.86387885e-01 5.78086853e-01 7.96091735e-01 3.53254288e-01
9.88550112e-02 -4.84115630e-01 -7.18332350e-01 5.55732369e-01
9.94178534e-01 -2.45772943e-01 1.76575437e-01 8.14301550e-01
1.03367901e+00 -1.05075264e+00 -1.26838386e+00 -3.53050351e-01
1.15071535e-01 -9.43164647e-01 6.63310349e-01 -6.70328856e-01
3.13091666e-01 -5.89678884e-01 1.74193755e-01 -1.14456010e+00
-5.95579028e-01 -1.12259543e+00 1.90708041e-02 6.48563564e-01
3.03431422e-01 -6.96318805e-01 6.30961716e-01 6.11747384e-01
-1.10400629e+00 -3.45717847e-01 -1.00538146e+00 -7.24288642e-01
-1.41081125e-01 -4.02093865e-02 1.02850676e+00 6.57795489e-01
7.20354021e-01 5.59160531e-01 3.59919459e-01 4.10821050e-01
4.73097473e-01 1.89885899e-01 8.35696638e-01 -6.30876303e-01
-5.99670827e-01 -5.92734694e-01 -3.56953263e-01 -1.27982998e+00
-2.11101532e-01 -1.47669518e+00 -5.05748689e-01 -9.25902724e-01
-4.87372428e-01 -9.58726108e-01 -1.10165244e-02 4.83250648e-01
8.96670282e-01 -2.50549108e-01 2.94502825e-01 3.07311803e-01
-4.77269799e-01 6.00581393e-02 1.40898836e+00 1.60588309e-01
-2.29438454e-01 -3.89756590e-01 -4.58600789e-01 -3.35988291e-02
5.76912582e-01 -2.27733687e-01 -3.07218581e-01 -8.66547227e-01
2.36310035e-01 4.23328161e-01 1.62015110e-01 -1.49005783e+00
9.80117083e-01 1.62436172e-01 4.42887962e-01 -3.43746781e-01
2.10397780e-01 -6.94484532e-01 8.46442223e-01 7.52868533e-01
-7.25284144e-02 1.09961641e+00 7.61686683e-01 9.89291146e-02
-2.45455325e-01 7.09553421e-01 5.22248745e-01 -2.33711246e-02
-5.36822319e-01 3.87166291e-01 -9.80498791e-01 -2.36201823e-01
1.09443200e+00 7.06402481e-01 -8.05770516e-01 6.93815053e-01
-7.03202307e-01 4.77557957e-01 6.39950931e-01 1.24714576e-01
-1.75919190e-01 -1.34188163e+00 -4.17517334e-01 8.86304259e-01
-2.51335740e-01 -3.29046249e-02 4.29537833e-01 8.88453543e-01
-1.41441917e+00 8.62466455e-01 -6.45142198e-01 -8.11120450e-01
-6.56080782e-01 6.80361927e-01 1.03644833e-01 -6.43129110e-01
-1.02089298e+00 6.19862556e-01 8.19370337e-03 1.40403226e-01
-4.60142612e-01 -7.81390607e-01 9.40033913e-01 -3.68730605e-01
8.41207564e-01 1.31339177e-01 6.61751747e-01 1.82938710e-01
-3.62447351e-01 -4.96897399e-01 -2.99890906e-01 -8.71892348e-02
9.16235328e-01 3.00793737e-01 -8.12083662e-01 8.93293023e-02
8.71161222e-01 8.92699659e-02 -9.35793221e-01 7.48485699e-02
-1.24437042e-01 -3.50363791e-01 6.37416065e-01 -5.18193245e-01
-1.03125978e+00 5.45822382e-01 7.99332559e-02 3.92154336e-01
9.84602153e-01 -2.78918922e-01 9.74651456e-01 -9.88886133e-02
6.90405309e-01 -9.40983057e-01 -2.51675397e-01 7.54095972e-01
1.48489177e-01 -9.00897011e-02 6.74193352e-03 -5.39357483e-01
4.55497950e-02 1.46800733e+00 7.74453342e-01 -6.89872354e-02
1.02283633e+00 1.49514067e+00 -5.67959368e-01 -4.90088701e-01
-1.53268170e+00 5.41701496e-01 -6.81077361e-01 -6.05656914e-02
7.08459258e-01 3.96392941e-01 -5.43826938e-01 1.57394379e-01
-2.57240802e-01 2.29280531e-01 7.59647548e-01 1.38857806e+00
-2.49709293e-01 -1.64745212e+00 -1.71718836e-01 4.84927654e-01
1.25557685e-03 -3.38276744e-01 7.15201735e-01 8.95366967e-01
-8.03058669e-02 4.79803503e-01 1.11860311e+00 -6.49492964e-02
1.55971348e-01 -2.27450997e-01 2.70106584e-01 -6.40306413e-01
-1.52474046e+00 -2.81655759e-01 3.17104816e-01 -1.13016081e+00
4.16950375e-01 -4.22780216e-01 -1.40596306e+00 -1.35999966e+00
4.66801196e-01 2.54971296e-01 8.45562100e-01 1.16979338e-01
1.32758224e+00 1.23223424e+00 6.42310381e-01 -9.33413506e-01
-3.30208898e-01 -7.64968172e-02 -8.60204697e-01 -3.61974299e-01
1.43840581e-01 -2.08487898e-01 1.57550380e-01 -3.28261465e-01] | [8.454631805419922, 3.3543128967285156] |
5eef94c0-c661-41de-9fb0-d4214762457c | em-network-oracle-guided-self-distillation | 2306.10058 | null | https://arxiv.org/abs/2306.10058v1 | https://arxiv.org/pdf/2306.10058v1.pdf | EM-Network: Oracle Guided Self-distillation for Sequence Learning | We introduce EM-Network, a novel self-distillation approach that effectively leverages target information for supervised sequence-to-sequence (seq2seq) learning. In contrast to conventional methods, it is trained with oracle guidance, which is derived from the target sequence. Since the oracle guidance compactly represents the target-side context that can assist the sequence model in solving the task, the EM-Network achieves a better prediction compared to using only the source input. To allow the sequence model to inherit the promising capability of the EM-Network, we propose a new self-distillation strategy, where the original sequence model can benefit from the knowledge of the EM-Network in a one-stage manner. We conduct comprehensive experiments on two types of seq2seq models: connectionist temporal classification (CTC) for speech recognition and attention-based encoder-decoder (AED) for machine translation. Experimental results demonstrate that the EM-Network significantly advances the current state-of-the-art approaches, improving over the best prior work on speech recognition and establishing state-of-the-art performance on WMT'14 and IWSLT'14. | ['Nam Soo Kim', 'Seok Min Kim', 'Minchan Kim', 'Hyeonseung Lee', 'Sunghwan Ahn', 'Ji Won Yoon'] | 2023-06-14 | null | null | null | null | ['machine-translation'] | ['natural-language-processing'] | [ 8.20107758e-01 3.81112367e-01 -3.80373597e-01 -4.44959104e-01
-1.11946845e+00 -4.72223431e-01 5.86940706e-01 -6.88976347e-01
-4.16070908e-01 6.73203707e-01 2.68156111e-01 -1.12802410e+00
5.33059359e-01 -1.11488953e-01 -8.49685967e-01 -7.75312364e-01
1.28022835e-01 6.25426829e-01 9.05717760e-02 -1.83027327e-01
-2.05019079e-02 -2.76252538e-01 -9.26244080e-01 7.50438273e-01
9.57890630e-01 8.17945957e-01 8.63324285e-01 8.10157895e-01
-1.77279577e-01 1.11464036e+00 -5.00487626e-01 -2.82305181e-01
-8.27462897e-02 -9.44249153e-01 -9.32223558e-01 -3.06678444e-01
8.22182819e-02 -3.81357849e-01 -6.96938872e-01 8.27727437e-01
6.91947699e-01 6.55016974e-02 5.83548129e-01 -7.83719182e-01
-7.57032871e-01 8.18669617e-01 -1.09124862e-01 3.86669993e-01
1.96507528e-01 3.98585618e-01 1.13691092e+00 -1.10699594e+00
6.79566979e-01 1.29193032e+00 3.74071479e-01 9.19450879e-01
-1.15249574e+00 -5.48248589e-01 4.19570208e-01 6.71449184e-01
-9.94673491e-01 -9.56120491e-01 3.87178808e-01 -3.07852551e-02
1.81868291e+00 3.63519758e-01 3.54234666e-01 1.73279166e+00
1.26960397e-01 1.40735173e+00 9.41734552e-01 -5.47596753e-01
1.85629800e-01 -1.83877453e-01 -1.81395575e-01 5.41779160e-01
-7.22547829e-01 5.55966139e-01 -7.38228500e-01 1.75182015e-01
4.84310925e-01 -3.46579731e-01 -3.02522242e-01 7.15254946e-03
-1.57934487e+00 5.16324639e-01 3.86797078e-02 2.67025739e-01
-2.57570267e-01 7.09410235e-02 5.23842931e-01 6.91334605e-01
6.10673249e-01 1.34405911e-01 -9.02344823e-01 -7.17830181e-01
-1.05143571e+00 -4.69058722e-01 9.06675875e-01 1.04731512e+00
2.74774432e-01 5.48048615e-01 -5.09188890e-01 9.02606368e-01
1.14931121e-01 6.30569339e-01 8.09931338e-01 -5.84037840e-01
9.45206285e-01 7.92242959e-02 -3.76360565e-01 -1.11145481e-01
1.19124942e-01 -6.66916668e-01 -9.16642070e-01 -5.80559969e-01
-7.33449906e-02 -2.49859050e-01 -1.04815269e+00 1.95982862e+00
-2.73976196e-03 8.38866770e-01 4.45564359e-01 5.71228266e-01
5.95618367e-01 1.10038400e+00 -2.50960529e-01 -3.57360721e-01
9.02500093e-01 -1.63710105e+00 -7.06565857e-01 -5.95986307e-01
9.90712523e-01 -5.39065540e-01 7.73761094e-01 2.19475366e-02
-1.16811848e+00 -5.78703105e-01 -8.60858440e-01 1.56868890e-01
-8.76728445e-02 2.53381938e-01 9.41080973e-02 1.75575092e-01
-1.36330402e+00 5.50561309e-01 -9.37168956e-01 -2.63908029e-01
1.64620399e-01 2.56169409e-01 -1.84181139e-01 -1.25763640e-01
-1.50409663e+00 1.23159397e+00 5.89564443e-01 2.75877148e-01
-1.39129531e+00 -5.77698290e-01 -9.07278776e-01 5.28784767e-02
4.01297539e-01 -6.79062903e-01 1.86733341e+00 -1.12504661e+00
-1.97465253e+00 6.93755269e-01 -7.78031826e-01 -7.24120557e-01
2.47703761e-01 -1.05653763e-01 -5.71075201e-01 1.34823725e-01
-6.30845428e-02 7.17538357e-01 8.31727266e-01 -6.50323510e-01
-5.71852982e-01 9.61272046e-02 -3.92192841e-01 2.45044708e-01
-2.30137065e-01 1.37265176e-01 -4.88063306e-01 -7.99554110e-01
-2.40997255e-01 -1.08162665e+00 -1.49621144e-01 -5.30250132e-01
-6.78673446e-01 -2.15316564e-01 7.73178875e-01 -1.01245904e+00
1.41775072e+00 -2.07586575e+00 5.30012548e-01 -2.29593143e-01
-1.89519808e-01 8.99239361e-01 -7.35138357e-01 7.79878020e-01
-2.95492947e-01 -1.14751875e-01 -4.59434181e-01 -5.49924433e-01
-2.94070039e-02 4.38585043e-01 -5.23242891e-01 5.52333742e-02
5.81408679e-01 1.41109896e+00 -1.04347467e+00 -2.57387370e-01
-2.29159906e-03 3.55708629e-01 -4.32090640e-01 6.32850826e-01
-5.87183177e-01 7.83770561e-01 -3.17384273e-01 2.95825571e-01
3.24252814e-01 -3.74958336e-01 6.47447765e-01 3.39289516e-01
1.16188474e-01 1.23107862e+00 -3.44589025e-01 2.09240603e+00
-5.85149527e-01 7.54272759e-01 -3.60510536e-02 -1.29620945e+00
7.31906831e-01 8.12184453e-01 3.46944444e-02 -1.06027102e+00
-2.12242573e-01 4.73821491e-01 3.07511866e-01 -4.97545540e-01
1.38493136e-01 2.27361359e-02 3.50184619e-01 6.29097939e-01
1.59175768e-01 4.17971879e-01 -2.40771934e-01 2.09096462e-01
1.25599253e+00 4.27353740e-01 6.36447594e-02 4.33347784e-02
5.14793396e-01 -1.87820733e-01 5.55022895e-01 7.51078963e-01
9.83487666e-02 2.48734936e-01 2.99514920e-01 3.37738506e-02
-1.18814731e+00 -9.19211805e-01 2.67023742e-01 1.23588848e+00
-3.36870104e-01 -3.92016768e-01 -1.03352451e+00 -8.85766745e-01
-4.22648221e-01 1.03960562e+00 -4.63952333e-01 -5.11927128e-01
-8.84086609e-01 -3.97281140e-01 7.25989878e-01 7.05994546e-01
4.60380316e-01 -1.08533466e+00 -1.01326644e-01 4.64152247e-01
-7.00458825e-01 -1.24295604e+00 -1.06853986e+00 5.23320138e-01
-1.10606408e+00 -5.28243423e-01 -6.93317294e-01 -1.10754120e+00
5.23765743e-01 2.09575504e-01 1.23965323e+00 -5.41313877e-03
2.74074227e-01 -6.14289343e-02 -4.55345094e-01 1.10984802e-01
-9.58131313e-01 5.19292235e-01 3.01716086e-02 -1.09655820e-01
3.22299182e-01 -6.68779552e-01 -3.90058279e-01 3.47670525e-01
-5.83537102e-01 4.86913651e-01 1.03369868e+00 1.36929834e+00
2.73593903e-01 -4.41463917e-01 8.45535278e-01 -6.89125061e-01
2.20705166e-01 -5.63775361e-01 -1.63061693e-01 4.33057934e-01
-5.63781857e-01 3.84089321e-01 6.90598130e-01 -6.62201703e-01
-1.08572197e+00 -2.47508958e-01 -4.97434944e-01 -5.38063705e-01
6.44043833e-02 6.70801222e-01 -1.87030181e-01 3.51419121e-01
6.44086897e-02 1.10181940e+00 1.06051996e-01 -6.70839429e-01
4.98549432e-01 9.40324962e-01 5.99846661e-01 -4.98360842e-01
5.12585521e-01 -1.38423428e-01 -5.67380488e-01 -6.02365077e-01
-9.18446898e-01 -3.10546160e-01 -6.80149078e-01 1.24124713e-01
6.51420176e-01 -9.93723094e-01 -6.08638167e-01 4.24020410e-01
-1.58420122e+00 -7.35869348e-01 -1.12745054e-02 4.11067992e-01
-7.03123152e-01 3.10563892e-01 -8.98519337e-01 -8.73423815e-01
-5.33813298e-01 -1.17906916e+00 1.10961902e+00 -4.73990202e-01
-1.89027101e-01 -1.16338909e+00 1.56226978e-01 2.39689559e-01
6.05483234e-01 -5.38960934e-01 1.06407678e+00 -1.07896650e+00
-7.74289191e-01 2.93444157e-01 8.66872966e-02 6.73130989e-01
-1.57648280e-01 -5.54155111e-01 -1.14855242e+00 -4.33665067e-01
2.42224231e-01 -2.79568851e-01 1.03901684e+00 1.19403288e-01
9.54772830e-01 -7.01602876e-01 -4.83082622e-01 6.41959131e-01
9.28482115e-01 5.09566486e-01 8.20929050e-01 1.73166528e-01
5.31146884e-01 3.92414421e-01 4.47603971e-01 -3.55860777e-02
4.70902711e-01 8.41841996e-01 2.87266046e-01 -8.68619308e-02
-2.95688629e-01 -7.22011447e-01 1.14331162e+00 1.79796565e+00
4.33618069e-01 -4.03873235e-01 -6.82672799e-01 4.47915345e-01
-2.07043457e+00 -1.02897727e+00 3.24320048e-01 1.87531853e+00
1.22257423e+00 4.21273336e-02 -1.07488766e-01 -1.63880318e-01
6.47745550e-01 2.97847301e-01 -8.19975078e-01 -4.52084631e-01
-2.14019716e-01 3.18803430e-01 1.51021659e-01 5.19239545e-01
-7.12210715e-01 1.20018411e+00 6.86125040e+00 1.29900706e+00
-1.15318060e+00 4.26257104e-01 6.09542906e-01 -4.00946029e-02
-3.39395165e-01 1.06852703e-01 -8.37175667e-01 6.77826345e-01
1.73479831e+00 1.25882328e-01 6.19207740e-01 5.53092062e-01
2.40546048e-01 3.70203614e-01 -1.50607145e+00 7.76268959e-01
8.20161253e-02 -1.33899593e+00 1.01423739e-02 8.03454444e-02
6.40090406e-01 5.74727893e-01 2.76400447e-01 6.30043149e-01
3.58159930e-01 -1.05612814e+00 7.37845480e-01 4.53612298e-01
1.13064778e+00 -4.76310015e-01 5.01206040e-01 8.60638559e-01
-1.07287729e+00 3.06735076e-02 -1.34845436e-01 1.15611451e-02
3.27134758e-01 3.59305054e-01 -1.16515708e+00 6.45234048e-01
1.09189861e-01 9.95208740e-01 5.94022162e-02 4.95804042e-01
-4.11945075e-01 1.22888207e+00 5.81324473e-02 -1.40453592e-01
5.82278728e-01 -8.73477757e-02 7.12797523e-01 1.58607292e+00
3.75111252e-01 -1.14237539e-01 6.66564852e-02 7.70025432e-01
-1.69839084e-01 -4.28939126e-02 -5.75304091e-01 -4.33866352e-01
6.34473681e-01 6.69844806e-01 2.59934030e-02 -6.35069430e-01
-4.84965622e-01 1.41096365e+00 5.41476369e-01 6.19875371e-01
-7.94876516e-01 -1.33965895e-01 7.36049056e-01 -3.00379425e-01
9.11517441e-01 -3.79975110e-01 2.01342583e-01 -1.35734546e+00
1.43491685e-01 -1.40630460e+00 5.74438311e-02 -7.56496251e-01
-1.00548732e+00 8.97689283e-01 -4.91646677e-01 -1.17935824e+00
-8.86983991e-01 -6.66717827e-01 -6.87480450e-01 1.20491135e+00
-1.73544478e+00 -1.09818792e+00 6.38464093e-01 3.61470729e-01
9.74282980e-01 -5.25611043e-01 1.04099464e+00 2.29373470e-01
-6.91739976e-01 9.68057930e-01 6.20574355e-01 2.31058717e-01
5.12241840e-01 -8.81109476e-01 1.08902311e+00 9.21555161e-01
2.06389740e-01 5.66921234e-01 1.65444002e-01 -6.27638936e-01
-1.67614067e+00 -1.17558527e+00 1.22837746e+00 -5.81637621e-01
7.84949243e-01 -5.66857398e-01 -8.04169953e-01 9.21439648e-01
5.98359942e-01 -4.60836828e-01 6.05086267e-01 -1.06255330e-01
-4.71964955e-01 2.28467807e-01 -4.00034934e-01 7.59141743e-01
1.40249717e+00 -1.11901248e+00 -7.43056357e-01 1.86689630e-01
1.27087665e+00 -5.47450483e-01 -5.98862886e-01 2.93249667e-01
3.07575017e-01 -5.77793241e-01 8.79017711e-01 -1.05215061e+00
4.78818923e-01 -5.46818264e-02 -2.35354677e-01 -1.72856116e+00
-2.17580602e-01 -1.17401767e+00 -6.77387774e-01 9.50133622e-01
8.52677166e-01 -7.15650380e-01 5.26776314e-01 -5.81910312e-02
-7.68794000e-01 -1.01228905e+00 -1.18006122e+00 -1.27323258e+00
1.67311460e-01 -5.89453399e-01 5.82963049e-01 8.68365705e-01
2.70600140e-01 7.20861018e-01 -7.25631595e-01 -1.10984124e-01
2.48747230e-01 -6.95050210e-02 4.27805334e-01 -5.05791783e-01
-8.00593674e-01 -3.60983014e-01 1.88446194e-02 -2.14546704e+00
5.95455229e-01 -1.44905889e+00 3.75013411e-01 -1.28826642e+00
3.03659886e-01 5.77238156e-03 -3.73694628e-01 5.44093013e-01
-2.86207110e-01 -2.95212626e-01 1.12093389e-01 2.44938195e-01
-6.75896883e-01 9.26263213e-01 1.38584948e+00 -2.26553261e-01
1.85869023e-01 -7.63222873e-02 -4.43945736e-01 7.78462514e-02
5.96004367e-01 -4.61469918e-01 -5.58112085e-01 -8.59197199e-01
-1.21486478e-01 4.53612000e-01 9.10793245e-02 -4.48696136e-01
5.11634290e-01 8.02255571e-02 -1.57695383e-01 -6.29717946e-01
3.77868950e-01 -4.43421274e-01 -8.83640572e-02 5.41027665e-01
-7.78728426e-01 7.77685195e-02 1.95996612e-01 6.10042632e-01
-3.64146531e-01 -6.73478842e-03 4.91476178e-01 -6.04840554e-02
-5.88138878e-01 3.24837059e-01 -5.73894203e-01 2.89129227e-01
4.47150826e-01 -1.83181632e-02 -3.42035323e-01 -6.33772254e-01
-7.21652150e-01 3.76596689e-01 -1.19357491e-02 5.48877537e-01
7.29540825e-01 -1.49025941e+00 -9.84167039e-01 2.90665776e-01
9.99762788e-02 -3.63688141e-01 6.08371943e-02 1.20777595e+00
4.03456360e-01 9.70977008e-01 3.15983266e-01 -7.21452117e-01
-9.97131407e-01 5.76898754e-01 3.82863343e-01 -4.25240755e-01
-6.13112628e-01 8.84045601e-01 4.46220726e-01 -7.30693221e-01
3.75684738e-01 -2.98791230e-01 3.84560019e-01 -4.92617339e-01
5.23086071e-01 8.14099237e-02 5.27882203e-02 -4.82223511e-01
-3.41920823e-01 4.23230380e-02 -4.23430622e-01 -4.65691864e-01
1.14768648e+00 -2.86547184e-01 -5.30407429e-02 5.20640373e-01
1.41170752e+00 -6.19705915e-01 -1.33250034e+00 -7.24404395e-01
2.46050194e-01 -4.53830399e-02 -1.37845576e-01 -1.08356166e+00
-7.82360315e-01 1.34497368e+00 1.45631924e-01 -2.69324005e-01
9.81680274e-01 -8.76814798e-02 1.27568114e+00 6.47410750e-01
2.45913357e-01 -1.10426617e+00 2.43241489e-01 1.22150159e+00
6.91473901e-01 -9.50494826e-01 -1.06940949e+00 -1.03347145e-01
-6.94501758e-01 9.52455938e-01 4.09116060e-01 4.86097753e-01
4.25251901e-01 5.21996558e-01 8.21436718e-02 3.76481384e-01
-1.54986942e+00 -8.60764086e-02 2.94685096e-01 6.25194609e-01
5.74849606e-01 4.59536053e-02 3.27376544e-01 5.62323213e-01
2.86714584e-02 -5.05888276e-02 -1.16561353e-01 8.49353135e-01
-4.59828317e-01 -1.38211203e+00 1.04706444e-01 3.13932300e-01
-1.52406707e-01 -9.77131307e-01 -3.75999719e-01 1.95944801e-01
-4.24977213e-01 9.60068941e-01 -1.89279933e-02 -7.36654401e-01
1.33562356e-01 5.88812113e-01 3.45242262e-01 -7.01732039e-01
-4.92418528e-01 2.56213695e-01 5.30941546e-01 -6.93853319e-01
-1.20726079e-01 -6.00639939e-01 -1.04879105e+00 -1.95498094e-01
-3.00638795e-01 2.86286175e-01 5.41132450e-01 1.26022136e+00
7.71015286e-01 6.63340628e-01 8.22401762e-01 -6.06325805e-01
-9.62950110e-01 -1.31942391e+00 -6.40608668e-02 -1.22399814e-01
5.93213379e-01 -2.53965795e-01 -1.90017506e-01 1.23202048e-01] | [14.472229957580566, 7.1801347732543945] |
5410ed75-5185-4636-9c56-86bdfd29a19e | lip-flow-learning-inference-time-priors-for | 2203.07881 | null | https://arxiv.org/abs/2203.07881v1 | https://arxiv.org/pdf/2203.07881v1.pdf | LiP-Flow: Learning Inference-time Priors for Codec Avatars via Normalizing Flows in Latent Space | Neural face avatars that are trained from multi-view data captured in camera domes can produce photo-realistic 3D reconstructions. However, at inference time, they must be driven by limited inputs such as partial views recorded by headset-mounted cameras or a front-facing camera, and sparse facial landmarks. To mitigate this asymmetry, we introduce a prior model that is conditioned on the runtime inputs and tie this prior space to the 3D face model via a normalizing flow in the latent space. Our proposed model, LiP-Flow, consists of two encoders that learn representations from the rich training-time and impoverished inference-time observations. A normalizing flow bridges the two representation spaces and transforms latent samples from one domain to another, allowing us to define a latent likelihood objective. We trained our model end-to-end to maximize the similarity of both representation spaces and the reconstruction quality, making the 3D face model aware of the limited driving signals. We conduct extensive evaluations where the latent codes are optimized to reconstruct 3D avatars from partial or sparse observations. We show that our approach leads to an expressive and effective prior, capturing facial dynamics and subtle expressions better. | ['Otmar Hilliges', 'Jason Saragih', 'Shih-En Wei', 'Alexander Richard', 'Stanislav Pidhorskyi', 'Akin Caliskan', 'Shugao Ma', 'Emre Aksan'] | 2022-03-15 | null | null | null | null | ['face-model'] | ['computer-vision'] | [ 9.46783870e-02 4.01369929e-01 -3.20048213e-01 -6.68452799e-01
-6.81848347e-01 -4.25635785e-01 6.88538194e-01 -1.12842405e+00
6.18370473e-02 1.92626134e-01 6.62732244e-01 2.14657769e-01
2.47273907e-01 -4.29595053e-01 -9.41614807e-01 -4.12163854e-01
2.82973617e-01 4.33613986e-01 -5.44552207e-01 2.27822527e-01
-1.37289256e-01 6.06258988e-01 -1.60753679e+00 2.58716941e-01
3.40479553e-01 1.05439615e+00 3.36612649e-02 6.28255427e-01
3.52840871e-02 8.36498857e-01 -2.18287498e-01 -5.30004203e-01
5.18212378e-01 -3.44796717e-01 -3.39790314e-01 5.88469684e-01
8.23032260e-01 -7.89662898e-01 -5.72506547e-01 8.51248920e-01
4.47560132e-01 8.43239352e-02 5.61236918e-01 -1.47311938e+00
-7.14464128e-01 1.81026608e-01 -6.63256407e-01 -4.34694141e-01
5.41943312e-01 2.73433536e-01 7.58468390e-01 -1.12439466e+00
9.83600736e-01 1.66119254e+00 4.83060300e-01 1.05164063e+00
-1.35005617e+00 -7.08037734e-01 3.52357149e-01 -2.46472955e-01
-1.35272956e+00 -1.31767821e+00 1.13162374e+00 -4.27490085e-01
5.57855368e-01 -4.01268937e-02 6.56075656e-01 1.76346397e+00
-2.97767609e-01 8.25777054e-01 7.61427999e-01 -8.06261450e-02
1.73646867e-01 1.73818976e-01 -5.17460406e-01 7.69500971e-01
-3.54667991e-01 1.33698583e-01 -1.04330122e+00 -1.46946862e-01
1.10260236e+00 2.43243128e-01 -4.90609437e-01 -7.44041920e-01
-8.61201704e-01 7.67802417e-01 3.10337497e-03 -1.96982205e-01
-4.20125961e-01 2.21270338e-01 1.17117740e-01 6.91637397e-02
5.96781075e-01 -1.43672258e-01 -2.19179511e-01 -2.60176480e-01
-1.06567013e+00 1.08770497e-01 7.54229188e-01 1.05367839e+00
6.28697991e-01 2.35895619e-01 4.88905460e-02 8.09091866e-01
7.77173519e-01 7.07336307e-01 2.44424775e-01 -1.54443240e+00
4.75479066e-01 3.65982950e-01 -9.01147053e-02 -8.20747793e-01
-1.56382471e-02 -1.63781300e-01 -5.57951331e-01 3.14265221e-01
3.04464400e-01 -3.95700634e-02 -9.04228926e-01 2.33709884e+00
3.92605394e-01 4.22825813e-01 -1.84746590e-02 1.16935039e+00
5.55298865e-01 6.20382190e-01 -3.67878377e-01 -2.46142372e-01
1.02937090e+00 -9.58512187e-01 -1.00056636e+00 -3.06599945e-01
1.14789963e-01 -5.27165234e-01 1.24537599e+00 1.83824643e-01
-1.46275210e+00 -4.73820835e-01 -9.04372811e-01 -3.58752251e-01
3.58983517e-01 2.67703682e-01 2.52439976e-01 4.70039696e-01
-1.09097898e+00 4.84910131e-01 -1.16279864e+00 -2.28798911e-01
4.84475702e-01 2.22444713e-01 -7.19246447e-01 -1.67199463e-01
-5.78637540e-01 5.98809302e-01 -4.81779844e-01 1.75972879e-01
-1.47936606e+00 -6.94714725e-01 -1.14440000e+00 -5.15591167e-03
3.10334653e-01 -1.01902330e+00 1.12948406e+00 -1.17750812e+00
-2.13965869e+00 1.05182791e+00 -4.60377723e-01 3.65357660e-02
7.48724639e-01 -2.97884256e-01 -3.23210210e-02 3.36392820e-01
4.80062999e-02 6.92472816e-01 1.32875550e+00 -1.46599472e+00
2.31990535e-02 -6.63304806e-01 6.12522699e-02 3.72350216e-01
-3.56579840e-01 -5.99402934e-02 -8.26120138e-01 -4.84339029e-01
-3.37597355e-02 -1.11490464e+00 1.18804187e-01 7.40510583e-01
-2.75996625e-01 3.64082396e-01 1.12176323e+00 -6.52136385e-01
5.97374439e-01 -2.35827136e+00 6.03689492e-01 1.65278643e-01
3.23295891e-01 -1.85728699e-01 -3.13887328e-01 2.99454965e-02
-6.33605346e-02 -1.63461566e-01 -9.36635286e-02 -1.23840618e+00
2.78113186e-02 4.06106859e-01 -4.53460574e-01 6.82635963e-01
1.95474893e-01 7.49236763e-01 -7.03891158e-01 -3.09640199e-01
2.02320561e-01 1.09825480e+00 -8.62906098e-01 7.28716671e-01
-1.62726849e-01 7.97468007e-01 -2.02904671e-01 6.60345495e-01
7.61546731e-01 -2.34988749e-01 1.66337878e-01 -1.66096389e-01
2.25039393e-01 1.13281041e-01 -1.01998222e+00 2.35849047e+00
-7.73949027e-01 5.44664681e-01 5.64157128e-01 -5.40250003e-01
9.60210323e-01 4.84222591e-01 5.69043636e-01 -4.29739237e-01
1.69811949e-01 -4.09065783e-02 -5.81008852e-01 -6.02848947e-01
1.22992918e-01 -3.66527677e-01 3.21991980e-01 4.71538693e-01
4.49685454e-01 -1.27733633e-01 -3.46872509e-01 1.42757654e-01
7.15457201e-01 7.12868273e-01 -2.92340338e-01 2.65141308e-01
2.91880250e-01 -8.59150529e-01 6.64135396e-01 3.76335233e-02
-4.33032401e-02 1.03598356e+00 6.78470910e-01 -3.80967349e-01
-1.01760161e+00 -1.16010439e+00 1.26587749e-01 9.12464678e-01
-1.21324696e-01 -6.14430726e-01 -7.03314126e-01 -7.26352572e-01
-9.38088670e-02 5.13996065e-01 -7.21852183e-01 -3.28622460e-01
-4.60629761e-01 4.26309407e-02 5.22294700e-01 4.16715801e-01
1.39119536e-01 -5.95093310e-01 -6.37694895e-01 -1.71184823e-01
-3.49273801e-01 -1.34749413e+00 -8.63262951e-01 -2.03242764e-01
-8.51113796e-01 -6.52072132e-01 -9.48721290e-01 -3.63052458e-01
8.71662557e-01 9.55064446e-02 1.02544808e+00 -3.31712097e-01
9.99206528e-02 6.15128875e-01 4.12961934e-03 1.35281039e-02
-3.55336726e-01 -4.35133576e-01 3.82154584e-01 7.05334544e-01
-4.83704498e-03 -1.15145791e+00 -4.75374371e-01 2.95309454e-01
-6.86134517e-01 4.11141872e-01 3.25392485e-01 7.14362562e-01
6.09599888e-01 -8.75730038e-01 1.43461358e-02 -5.64461946e-01
1.51753783e-01 -4.33386564e-01 -5.78119278e-01 8.27246904e-03
-1.84976757e-01 2.86524802e-01 6.69532239e-01 -6.95524216e-01
-1.14973402e+00 4.70180035e-01 -2.35043034e-01 -1.57943022e+00
-5.73065355e-02 7.17753097e-02 -7.82316327e-01 8.65006074e-02
4.49679911e-01 1.19530827e-01 4.01897103e-01 -5.99417686e-01
7.38321543e-01 5.55511177e-01 7.75327563e-01 -8.31485152e-01
6.88385665e-01 7.63949931e-01 -9.84568894e-02 -7.78192461e-01
-6.22604311e-01 -1.84275303e-02 -5.63533127e-01 -4.34715688e-01
7.29968309e-01 -1.14412296e+00 -8.61226201e-01 3.00941080e-01
-1.22066820e+00 -2.80519992e-01 -5.37573695e-01 3.84427667e-01
-9.26977456e-01 1.78738907e-01 -4.40228552e-01 -9.44790244e-01
-5.90494126e-02 -1.31510842e+00 1.62865007e+00 1.10428333e-01
-2.19081953e-01 -8.63218069e-01 2.33986810e-01 4.58157748e-01
1.29139960e-01 3.62193704e-01 5.33760309e-01 -1.12637440e-02
-5.85050046e-01 -5.60276136e-02 2.26249751e-02 4.47249532e-01
7.29179680e-02 1.27829537e-01 -1.56786799e+00 -2.34719008e-01
2.46410757e-01 -5.72307229e-01 6.86200798e-01 4.24991012e-01
1.09675395e+00 -6.30529404e-01 -1.21174313e-01 1.33698237e+00
1.06111622e+00 -1.26317009e-01 4.60318863e-01 -4.41342443e-01
8.48253846e-01 6.60684288e-01 -4.65367325e-02 6.57788932e-01
5.66005170e-01 8.57764482e-01 6.65542305e-01 9.87275913e-02
-2.11993217e-01 -9.63059723e-01 8.04176509e-01 6.73243284e-01
-2.22114459e-01 -7.93183967e-02 -5.36676943e-01 4.10514325e-01
-1.65700674e+00 -9.56693411e-01 6.44966185e-01 2.05614877e+00
7.13468671e-01 -2.22636566e-01 1.38169304e-01 -2.13212430e-01
4.14730132e-01 3.52734745e-01 -8.54029834e-01 -1.72809541e-01
9.61843506e-03 -6.51115775e-02 -1.01600304e-01 7.39511192e-01
-6.04290962e-01 7.44472384e-01 6.03615904e+00 3.26221138e-01
-1.38867986e+00 1.09325029e-01 5.22261143e-01 -6.40718281e-01
-7.27010906e-01 6.60748258e-02 -5.38878322e-01 3.85579884e-01
8.66146445e-01 2.34312546e-02 6.10276401e-01 9.51139808e-01
3.41040641e-01 4.86644536e-01 -1.53761578e+00 1.37803495e+00
3.94197226e-01 -1.31418705e+00 2.96900213e-01 4.39411879e-01
6.11462831e-01 -2.65857041e-01 4.04242665e-01 9.86117497e-02
2.40851089e-01 -1.09760416e+00 1.14264452e+00 7.97095299e-01
1.30164981e+00 -4.95455891e-01 4.60787937e-02 4.28645998e-01
-8.99314582e-01 8.81810933e-02 -1.39720514e-01 7.97945261e-02
4.07730281e-01 2.21996799e-01 -5.23501217e-01 2.24578455e-01
4.53790247e-01 9.14523959e-01 6.94987848e-02 2.44005889e-01
-1.60783067e-01 3.18784386e-01 -3.71949404e-01 6.01418257e-01
-1.08046055e-01 -4.10496175e-01 6.29678011e-01 6.46976888e-01
4.17899579e-01 5.34320660e-02 7.07855076e-02 1.24028039e+00
-4.00972605e-01 -3.36392462e-01 -8.99926841e-01 -1.76237505e-02
2.35925719e-01 1.18187296e+00 -1.41795397e-01 -1.15446665e-03
-5.21965265e-01 1.18396902e+00 4.77813572e-01 5.90978265e-01
-8.82568896e-01 2.60463059e-01 1.15208125e+00 3.05816591e-01
5.74875027e-02 -2.81908870e-01 -4.37379070e-02 -1.60767281e+00
1.15061119e-01 -8.24095190e-01 -2.83708870e-02 -1.00244761e+00
-1.08851218e+00 7.10399806e-01 -1.23518303e-01 -1.15887117e+00
-6.22276545e-01 -4.74271894e-01 -4.38003123e-01 8.09197009e-01
-1.42562854e+00 -1.46854496e+00 -3.25001955e-01 8.92955601e-01
5.04021227e-01 -1.74857125e-01 8.92466307e-01 2.98017740e-01
-6.20851159e-01 8.85485411e-01 -3.08440864e-01 2.57892702e-02
7.24187732e-01 -8.35506260e-01 2.38180354e-01 6.35178447e-01
4.03780848e-01 5.87115228e-01 5.57822883e-01 -2.25888804e-01
-1.66034937e+00 -7.76959419e-01 5.79560876e-01 -6.33852601e-01
2.09735721e-01 -7.26827323e-01 -6.89700305e-01 9.97762144e-01
6.29358515e-02 5.40558755e-01 7.97429085e-01 -2.61020493e-02
-8.00032496e-01 -1.94734931e-01 -1.06393611e+00 5.53362072e-01
1.36133373e+00 -1.03180981e+00 -2.67649055e-01 -6.50398955e-02
5.31667113e-01 -6.27209187e-01 -6.49845183e-01 1.73204005e-01
9.96065021e-01 -1.02377439e+00 8.34592283e-01 -6.14377379e-01
6.97840273e-01 -8.20613131e-02 -3.39031011e-01 -1.24657142e+00
1.97682995e-02 -1.04210687e+00 -5.31671345e-01 1.16939986e+00
1.49081409e-01 -1.01466447e-01 1.04405010e+00 8.76549602e-01
-1.12777442e-01 -7.72795081e-01 -1.04266071e+00 -4.87816811e-01
-1.22121021e-01 -4.70560253e-01 6.54715300e-01 8.70673716e-01
-1.59696862e-01 5.10926604e-01 -8.43145251e-01 1.41066611e-01
7.12808132e-01 7.75988325e-02 1.12765658e+00 -1.01980221e+00
-3.40567946e-01 -1.64390206e-01 -2.26267278e-01 -1.55022883e+00
6.36481643e-01 -8.67001474e-01 -9.52350572e-02 -1.01664257e+00
1.88190386e-01 -7.45375007e-02 2.56788045e-01 4.30643737e-01
2.56290138e-01 9.95399132e-02 4.02582198e-01 2.46911705e-01
-3.16368908e-01 1.13938129e+00 1.39799249e+00 5.58975376e-02
-1.26126707e-01 -1.37277648e-01 -5.65659285e-01 9.90202188e-01
7.04859942e-02 -3.54353130e-01 -7.34129786e-01 -8.93702805e-01
2.98607528e-01 5.71180165e-01 4.39986587e-01 -5.66698790e-01
1.94034547e-01 -2.17381895e-01 3.94228429e-01 -1.56844676e-01
1.00520229e+00 -1.01769304e+00 3.78370464e-01 -1.84861377e-01
-4.40718681e-01 -1.93870500e-01 -6.55938610e-02 4.69753057e-01
-1.30006164e-01 2.60190874e-01 9.19578254e-01 2.70894379e-04
-1.00959972e-01 7.37306833e-01 4.25289832e-02 1.39807075e-01
7.90222466e-01 -3.32334697e-01 1.72571152e-01 -9.75520194e-01
-8.48628700e-01 6.16634004e-02 8.13108325e-01 5.54605782e-01
8.60834599e-01 -1.63044107e+00 -7.32825220e-01 8.09507191e-01
2.29456788e-03 1.06136911e-01 3.80852103e-01 5.18647790e-01
-2.73151547e-01 -4.39974293e-02 -4.04666096e-01 -8.47394884e-01
-1.14165139e+00 4.27598208e-01 4.64461267e-01 1.06627792e-01
-7.84245729e-01 8.93261254e-01 4.38905597e-01 -5.79912484e-01
5.29739082e-01 -1.37874559e-01 1.24888100e-01 -7.41198659e-02
4.83080149e-01 1.25621215e-01 -3.29234093e-01 -9.85543430e-01
-1.94705859e-01 8.94295692e-01 1.81998596e-01 -5.59976637e-01
1.40407562e+00 -2.41189450e-01 2.94316173e-01 3.37931097e-01
1.51700258e+00 2.51384646e-01 -2.12829971e+00 -2.04161495e-01
-6.58523619e-01 -8.76361668e-01 -8.36017579e-02 -3.79987627e-01
-1.50086164e+00 1.11827970e+00 3.74993414e-01 -5.17075896e-01
1.13152933e+00 8.45174193e-02 7.71722555e-01 1.64392740e-02
2.47491673e-01 -7.22135603e-01 4.30990607e-01 3.71552527e-01
1.12112415e+00 -9.35087204e-01 -3.03277254e-01 -2.88176656e-01
-7.44711220e-01 9.92033839e-01 5.38969338e-01 -1.11742951e-01
6.62529707e-01 5.43136418e-01 1.67585954e-01 -1.54602006e-01
-1.07097733e+00 1.54188976e-01 1.64709151e-01 7.47717559e-01
7.19023198e-02 -2.16745242e-01 5.39610624e-01 8.26031685e-01
-3.06908399e-01 1.39341995e-01 3.53493184e-01 4.68629211e-01
2.63390064e-01 -8.39477301e-01 -3.27228963e-01 -1.37302607e-01
-2.60230750e-01 3.73743683e-01 -4.29652423e-01 3.90795588e-01
8.96469951e-02 7.28083491e-01 2.69550472e-01 -4.15006816e-01
4.85832483e-01 2.31205672e-01 6.73363209e-01 -4.48780596e-01
7.77926371e-02 2.89489239e-01 -1.09146692e-01 -1.01073158e+00
-3.31712753e-01 -6.22354686e-01 -9.64568198e-01 -4.10148263e-01
-1.27735615e-01 -1.62374526e-01 6.34810209e-01 6.88748300e-01
5.64875305e-01 1.68224633e-01 1.04371655e+00 -1.32847381e+00
-6.19212270e-01 -7.20776916e-01 -4.80986089e-01 5.66779375e-01
6.87933862e-01 -7.00922608e-01 -4.67871666e-01 3.65757704e-01] | [12.816967964172363, -0.3553785979747772] |
dad352ae-7bae-42c5-9b83-1684ecff61d8 | a-deep-forgetful-novelty-seeking-movie | 1909.01811 | null | https://arxiv.org/abs/1909.01811v1 | https://arxiv.org/pdf/1909.01811v1.pdf | A Deep, Forgetful Novelty-Seeking Movie Recommender Model | As more and more people shift their movie watching online, competition between movie viewing websites are getting more and more intense. Therefore, it has become incredibly important to accurately predict a given user's watching list to maximize the chances of keeping the user on the platform. Recent studies have suggested that the novelty-seeking propensity of users can impact their viewing behavior. In this paper, we aim to accurately model and describe this novelty-seeking trait across many users and timestamps driven by data, taking into consideration user forgetfulness. Compared to previous studies, we propose a more robust measure for novelty. Our model, termed Deep Forgetful Novelty-Seeking Model (DFNSM), leverages demographic information about users, genre information about movies, and novelty-seeking traits to predict the most likely next actions of a user. To evaluate the performance of our model, we conducted extensive experiments on a large movie rating dataset. The results reveal that DFNSM is very effective for movie recommendation. | ['Ruomu Zou'] | 2019-09-02 | null | null | null | null | ['movie-recommendation'] | ['miscellaneous'] | [-4.70103115e-01 -6.84620738e-01 -4.14757222e-01 -6.68442011e-01
-7.63733909e-02 -3.92322510e-01 3.30584347e-01 3.14556867e-01
-3.57490510e-01 1.13115571e-01 4.85757798e-01 -1.67677283e-01
-8.04862604e-02 -7.02355802e-01 -3.78255427e-01 -9.32403281e-02
-5.82009852e-02 -2.98145622e-01 2.11681008e-01 -2.43511915e-01
7.28913963e-01 5.22867106e-02 -1.45157135e+00 1.85761765e-01
9.73636925e-01 1.14554536e+00 3.23626846e-01 2.58366644e-01
2.56310374e-01 7.43049026e-01 -1.71782851e-01 -8.64512563e-01
4.30885047e-01 -3.63481939e-01 -3.30904782e-01 1.40539050e-01
5.38914382e-01 -6.36604488e-01 -6.00482523e-01 1.05320275e+00
2.22178355e-01 9.45416391e-01 3.84629637e-01 -8.76123190e-01
-1.21706951e+00 7.11168706e-01 -2.36093223e-01 7.97087073e-01
3.39598149e-01 3.30072492e-02 1.23171926e+00 -8.75132501e-01
4.54481184e-01 7.95021892e-01 5.80131173e-01 3.99060071e-01
-8.51422668e-01 -3.71612519e-01 6.04598701e-01 4.50871408e-01
-1.11041307e+00 -4.14533198e-01 8.09088528e-01 -4.04074520e-01
5.65366268e-01 3.29077840e-01 9.40062642e-01 1.10874212e+00
5.45546830e-01 8.42815459e-01 6.36553288e-01 -5.00909202e-02
4.17505473e-01 3.07978421e-01 2.14020595e-01 2.93979943e-01
1.63365766e-01 -4.23759013e-01 -8.25414002e-01 -2.24245727e-01
6.01111233e-01 9.71728742e-01 7.59229064e-02 -1.68530524e-01
-6.07066453e-01 7.31871665e-01 1.74021512e-01 3.47360492e-01
-7.13540018e-01 -3.76032770e-01 1.04703363e-02 5.45687556e-01
6.90706670e-01 6.95159376e-01 -3.32480073e-01 -5.84126830e-01
-9.48899746e-01 3.66136342e-01 5.15053272e-01 8.30579460e-01
5.09217799e-01 -1.28290057e-01 -3.26757014e-01 9.33439791e-01
8.74296948e-02 4.21686135e-02 9.04703975e-01 -1.15146995e+00
-1.16539173e-01 7.06055939e-01 2.23792046e-01 -1.37453830e+00
-5.72181195e-02 -6.14635110e-01 -4.61181104e-01 -4.27655488e-01
2.31825620e-01 2.56637321e-03 -2.75877655e-01 1.71923089e+00
-2.18706280e-02 6.23217747e-02 -5.66139162e-01 8.08062851e-01
2.27005064e-01 5.34037709e-01 -7.29427561e-02 -7.06074417e-01
1.02338219e+00 -9.59543824e-01 -6.59438610e-01 -2.31981754e-01
2.60548919e-01 -4.89010155e-01 1.47740626e+00 8.52221489e-01
-1.22809684e+00 -7.20782459e-01 -6.51784897e-01 -2.52253227e-02
9.17112678e-02 -9.42865908e-02 8.07754338e-01 4.87478703e-01
-8.39931309e-01 1.09215486e+00 -6.87074125e-01 -5.57434440e-01
2.49751836e-01 -8.64603147e-02 1.43440649e-01 -1.72994778e-01
-1.24696243e+00 4.72635686e-01 -2.90544569e-01 -3.11315447e-01
-9.13331449e-01 -5.55847406e-01 -2.46209398e-01 3.93961221e-01
5.01423120e-01 -5.64197838e-01 1.48562825e+00 -1.05313075e+00
-1.35350800e+00 2.47337952e-01 -1.89938605e-01 -2.92287320e-01
4.64756116e-02 -4.03737545e-01 -8.45039546e-01 -2.64866441e-01
-4.54639969e-03 8.00632238e-02 1.03981888e+00 -5.20282447e-01
-7.42340982e-01 -6.04056358e-01 2.70500213e-01 2.14512318e-01
-1.11868656e+00 -1.06367007e-01 -4.02609199e-01 -7.95554757e-01
3.82905938e-02 -7.24062383e-01 -2.26041242e-01 -2.18181595e-01
1.38555869e-01 -4.80952889e-01 1.80526212e-01 -6.42845333e-01
2.04856801e+00 -2.45877099e+00 7.70388637e-03 -4.15079594e-02
3.48643631e-01 1.67576313e-01 -8.85876417e-02 5.93731284e-01
5.30828297e-01 1.24437533e-01 5.08695364e-01 -4.12970722e-01
-1.98078141e-01 -1.22994870e-01 -3.15009385e-01 1.55701488e-01
-4.56513375e-01 6.27606809e-01 -8.11306596e-01 6.88435659e-02
-3.07006925e-01 1.10665962e-01 -1.01612306e+00 3.56470466e-01
-2.28785798e-01 2.22489327e-01 -5.29735029e-01 7.48171329e-01
4.27279592e-01 -6.03542149e-01 2.34624390e-02 2.23088741e-01
-2.15775222e-01 3.60746652e-01 -6.26184344e-01 1.70365703e+00
-3.59636158e-01 2.94720203e-01 -3.78767312e-01 -3.29037130e-01
8.70979190e-01 -5.16245179e-02 4.06912059e-01 -9.59885478e-01
1.17386699e-01 -1.85082436e-01 2.73409300e-02 -6.73435867e-01
9.47980702e-01 -4.21138760e-03 -3.12757306e-03 4.35677618e-01
-1.79172516e-01 7.15658665e-01 2.16881111e-01 6.95596278e-01
1.07804728e+00 -4.25744861e-01 3.76800835e-01 -9.51899514e-02
2.03699037e-01 -4.64841902e-01 8.68992329e-01 9.74726260e-01
-5.60958683e-01 3.08908194e-01 2.85587490e-01 -5.15576065e-01
-7.63420820e-01 -1.00430214e+00 2.65328944e-01 1.44764245e+00
1.43623278e-01 -8.28193843e-01 -4.15757567e-01 -5.25178432e-01
1.40348181e-01 1.07242274e+00 -6.06975198e-01 -7.42047191e-01
-6.39767796e-02 -3.27911466e-01 -3.14716309e-01 2.36394972e-01
2.26719350e-01 -7.82200933e-01 1.09662078e-02 1.97401658e-01
-2.62289792e-01 -5.94769835e-01 -1.04135323e+00 -7.39008307e-01
-1.26494825e+00 -4.54839468e-01 -5.51918745e-01 -2.69111782e-01
3.92576128e-01 9.45916116e-01 9.62845027e-01 6.57271072e-02
1.66807652e-01 4.44968134e-01 -6.46585882e-01 -3.22558731e-02
1.34459645e-01 1.32594794e-01 6.89237893e-01 4.21200693e-01
7.29215384e-01 -8.99443328e-01 -1.15281129e+00 3.19736183e-01
-9.66380239e-01 -4.30610538e-01 4.46269572e-01 1.36815146e-01
3.10503036e-01 4.18699533e-01 8.21070910e-01 -7.95291305e-01
9.55379546e-01 -1.06469524e+00 -2.85622962e-02 1.51081681e-01
-1.31120074e+00 -6.22359335e-01 8.68754148e-01 -7.60239661e-01
-1.15699112e+00 -3.03410888e-01 -9.77616459e-02 -2.26663142e-01
1.74283966e-01 5.31550527e-01 2.97417864e-03 3.37540358e-01
5.99747181e-01 2.11664364e-01 -1.65964097e-01 -9.81513381e-01
2.50175655e-01 7.46855140e-01 1.71294689e-01 9.18335468e-02
3.20033997e-01 4.38132972e-01 -5.74036896e-01 -7.64291286e-01
-1.41014540e+00 -7.54839540e-01 -3.18606645e-01 -5.38501501e-01
3.55748445e-01 -9.23094153e-01 -7.31793523e-01 3.81803423e-01
-6.27364516e-01 1.93677023e-01 -1.98449180e-01 6.15831554e-01
-2.90247463e-02 6.73928857e-01 -7.80546844e-01 -8.28158975e-01
-6.49090186e-02 -6.32959723e-01 2.24441141e-01 5.33377528e-01
-3.26068282e-01 -1.02577174e+00 2.57844418e-01 5.99045575e-01
7.43288457e-01 -4.87181664e-01 6.12746954e-01 -8.11285555e-01
-6.22307360e-01 -3.61929148e-01 2.35194102e-01 5.43446124e-01
3.30323339e-01 -2.55435079e-01 -4.38730061e-01 -4.64513570e-01
4.56785202e-01 4.76643033e-02 7.38380492e-01 4.80562359e-01
1.21583486e+00 -6.37403071e-01 4.47139740e-02 2.86512911e-01
1.20537221e+00 2.18706042e-01 4.43598151e-01 2.09293619e-01
4.63332057e-01 2.90288568e-01 8.84272754e-01 1.08953202e+00
6.31726265e-01 5.80409586e-01 3.50227535e-01 7.22445011e-01
4.26014304e-01 -6.55219734e-01 7.80694306e-01 1.25706410e+00
-4.94800843e-02 -5.83114088e-01 -2.00310901e-01 3.50420624e-01
-1.76919913e+00 -1.11416423e+00 8.75335783e-02 2.41145658e+00
5.43294370e-01 2.31216356e-01 5.90893447e-01 -2.93969959e-01
4.71325219e-01 2.57081389e-01 -9.54955578e-01 -2.44129404e-01
1.99977413e-01 -4.06057835e-01 8.27601701e-02 2.29485780e-01
-5.53475916e-01 7.76361346e-01 6.09171295e+00 4.84104663e-01
-8.34492743e-01 2.67509669e-01 6.99055731e-01 -8.50041449e-01
-4.85811204e-01 3.40371244e-02 -8.62884879e-01 8.39546382e-01
1.07673621e+00 -4.84107554e-01 8.23506057e-01 9.43270326e-01
5.35047412e-01 -1.74876362e-01 -9.82792675e-01 8.14508021e-01
5.65507114e-01 -9.38397944e-01 6.77989498e-02 5.89043684e-02
7.60417700e-01 -2.14338943e-01 4.13866520e-01 6.42630219e-01
-5.24290167e-02 -2.25561962e-01 4.59155589e-01 1.05149984e+00
-1.22815594e-02 -5.67056537e-01 1.38643920e-01 7.48780549e-01
-5.09071648e-01 -6.11389220e-01 -5.71843803e-01 -7.25710392e-01
4.91749525e-01 8.96723449e-01 -2.50328541e-01 -4.89987619e-02
8.93346250e-01 1.16038001e+00 -9.41249073e-01 1.12897813e+00
2.46936053e-01 8.08691680e-01 3.20136458e-01 -9.30448100e-02
-1.70003876e-01 -3.39154869e-01 5.38423300e-01 4.65174854e-01
7.30598807e-01 2.50136912e-01 -1.21070996e-01 6.27040982e-01
-2.86279529e-01 4.49191362e-01 -4.26076621e-01 -5.50123632e-01
4.41435546e-01 1.38058543e+00 -4.36213046e-01 -2.65062988e-01
-5.86114585e-01 1.25151622e+00 3.81615937e-01 5.93962744e-02
-6.73768997e-01 -5.16218729e-02 9.09091651e-01 6.24012172e-01
2.19606176e-01 -3.87447327e-01 -2.64547821e-02 -1.64922953e+00
-4.66117775e-03 -7.41079569e-01 4.62346375e-01 -6.35542691e-01
-1.72929990e+00 3.67033005e-01 -5.81576407e-01 -1.09824753e+00
1.01876989e-01 -1.34069681e-01 -7.97231853e-01 4.42327142e-01
-1.12355423e+00 -4.34092641e-01 -1.79234281e-01 2.90508628e-01
7.98163414e-01 -2.12282211e-01 1.72931001e-01 5.35096347e-01
-6.84458971e-01 5.82562447e-01 3.48244160e-01 -5.89235902e-01
1.00469255e+00 -7.74053752e-01 5.80284238e-01 9.17779207e-01
1.64851874e-01 1.14181793e+00 5.60029387e-01 -8.91185820e-01
-1.46712244e+00 -8.68951321e-01 1.18754268e+00 -8.04940641e-01
5.65579116e-01 -2.80560046e-01 -9.55681324e-01 5.85651278e-01
-4.76631038e-02 -6.86904430e-01 1.12497520e+00 5.43140531e-01
-2.10078403e-01 -3.64912689e-01 -8.55781734e-01 6.71994328e-01
1.25960672e+00 -3.92632723e-01 -4.57524866e-01 3.11974496e-01
8.31199527e-01 3.95363897e-01 -9.21027303e-01 -1.05224494e-02
5.83167076e-01 -1.26130044e+00 6.65470600e-01 -8.19217265e-01
5.09644926e-01 1.71048686e-01 -1.40362680e-01 -1.26282275e+00
-1.19660854e+00 -6.30058467e-01 -4.68019485e-01 1.22402394e+00
2.15985358e-01 -2.13807300e-01 8.19304943e-01 1.12763512e+00
2.99220923e-02 -6.21793151e-01 -4.86056775e-01 -7.91379869e-01
-3.61774266e-01 -3.47700380e-02 4.68449146e-01 9.57501769e-01
3.62136453e-01 5.70751846e-01 -1.07983947e+00 -7.40675330e-02
1.91919208e-01 1.47165745e-01 5.14653742e-01 -1.53638804e+00
-6.46438777e-01 -3.36471289e-01 1.26104532e-02 -1.48323798e+00
-2.83344030e-01 -6.40051007e-01 -4.42023516e-01 -1.22823191e+00
6.05016589e-01 1.55956522e-02 -8.53333294e-01 8.90470203e-03
-3.31897259e-01 7.89615065e-02 2.04713214e-02 5.04244804e-01
-1.41457653e+00 7.17329800e-01 1.31897068e+00 3.93164545e-01
-3.47683817e-01 7.19239414e-01 -1.23425484e+00 5.81270993e-01
8.12223673e-01 -2.00753346e-01 -6.48468316e-01 -3.37364733e-01
1.02801514e+00 2.43888125e-01 -1.44791499e-01 -8.02171588e-01
3.10458690e-01 -3.61413211e-01 1.68208823e-01 -3.84298325e-01
3.94875377e-01 -7.03992069e-01 1.95515063e-02 1.32488415e-01
-6.30056322e-01 5.04148483e-01 -3.98028821e-01 9.51781392e-01
5.34121133e-02 -3.35564852e-01 4.15618122e-01 -1.00871444e-01
-6.10033274e-01 8.78567815e-01 -6.81233525e-01 -6.58202916e-02
8.20564866e-01 5.91396075e-03 1.26008792e-02 -9.47035372e-01
-7.27529466e-01 -3.08922604e-02 6.43065393e-01 9.41651285e-01
5.98580241e-01 -1.37167323e+00 -1.47103041e-01 1.56040519e-01
1.98977157e-01 -8.60128522e-01 8.89932275e-01 6.95892334e-01
2.85243034e-01 2.39089310e-01 -2.07664609e-01 1.01916768e-01
-1.06093645e+00 9.68683600e-01 -2.25818709e-01 2.47416869e-01
-1.25393033e-01 8.00352573e-01 -1.14076592e-01 5.67940287e-02
3.00268918e-01 -2.77198762e-01 -4.45578754e-01 1.18030451e-01
8.13477397e-01 9.53452706e-01 -2.02879518e-01 -2.76124686e-01
4.74104583e-02 -1.99510321e-01 -7.46993721e-01 1.81658834e-01
1.41763830e+00 -7.00980008e-01 -4.32246067e-02 8.80239487e-01
9.21363413e-01 3.36238354e-01 -1.28958821e+00 -5.57841122e-01
-6.06057718e-02 -1.06401730e+00 9.93170142e-02 -6.71212971e-01
-9.47240472e-01 6.23216987e-01 3.87028605e-01 6.46966696e-01
1.12206244e+00 -1.51733845e-01 1.37581158e+00 2.19230652e-01
3.63742322e-01 -1.31849325e+00 5.51910579e-01 3.96773130e-01
3.68305624e-01 -1.36326873e+00 4.60819416e-02 -1.13236951e-02
-9.56078172e-01 5.34597337e-01 9.13509607e-01 -3.16172093e-02
9.73531604e-01 -8.09909701e-01 -1.08096138e-01 -3.33502777e-02
-1.16927421e+00 3.80167514e-01 3.55374962e-01 -3.92922647e-02
5.17798424e-01 5.26021942e-02 -6.66155934e-01 1.28473473e+00
2.88907252e-02 2.08850816e-01 6.84237063e-01 5.88853836e-01
-9.04337823e-01 -1.29123795e+00 4.48284447e-01 9.03235197e-01
-8.15416813e-01 -1.15419246e-01 -3.51286650e-01 -2.18827561e-01
-9.34447497e-02 1.01744390e+00 1.48228267e-02 -8.66387308e-01
3.76470506e-01 1.03320003e-01 7.84301385e-02 -8.36110950e-01
-5.67791581e-01 -2.41229564e-01 -2.29043856e-01 -7.45724797e-01
-3.26406956e-02 -1.03076565e+00 -5.17948806e-01 -6.51699305e-01
-2.46273026e-01 1.84990004e-01 4.16308880e-01 6.66834474e-01
9.28075254e-01 -1.26599716e-02 1.13234472e+00 -2.92606920e-01
-9.65757608e-01 -8.79275203e-01 -1.09204960e+00 6.22879028e-01
-4.00856100e-02 -4.23100829e-01 -5.56328833e-01 -1.97404847e-01] | [10.15149211883545, 5.644640922546387] |
926a90f3-f9f6-47cd-b022-d88cafe30d54 | fedcbo-reaching-group-consensus-in-clustered | 2305.02894 | null | https://arxiv.org/abs/2305.02894v1 | https://arxiv.org/pdf/2305.02894v1.pdf | FedCBO: Reaching Group Consensus in Clustered Federated Learning through Consensus-based Optimization | Federated learning is an important framework in modern machine learning that seeks to integrate the training of learning models from multiple users, each user having their own local data set, in a way that is sensitive to data privacy and to communication loss constraints. In clustered federated learning, one assumes an additional unknown group structure among users, and the goal is to train models that are useful for each group, rather than simply training a single global model for all users. In this paper, we propose a novel solution to the problem of clustered federated learning that is inspired by ideas in consensus-based optimization (CBO). Our new CBO-type method is based on a system of interacting particles that is oblivious to group memberships. Our model is motivated by rigorous mathematical reasoning, including a mean field analysis describing the large number of particles limit of our particle system, as well as convergence guarantees for the simultaneous global optimization of general non-convex objective functions (corresponding to the loss functions of each cluster of users) in the mean-field regime. Experimental results demonstrate the efficacy of our FedCBO algorithm compared to other state-of-the-art methods and help validate our methodological and theoretical work. | ['Yuhua Zhu', 'Sixu Li', 'Nicolas Garcia Trillos', 'Jose A. Carrillo'] | 2023-05-04 | null | null | null | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [-2.63842463e-01 4.76843752e-02 7.78860897e-02 -1.81550145e-01
-6.75431788e-01 -4.69734609e-01 5.11266589e-01 6.27988815e-01
-4.37258750e-01 7.84456968e-01 -1.59017101e-01 -3.67690213e-02
-7.25313008e-01 -8.92074764e-01 -9.52208996e-01 -1.41826642e+00
-4.75563496e-01 8.45346212e-01 -3.08289528e-01 -9.19195116e-02
-4.08446826e-02 4.73015279e-01 -1.35853553e+00 5.00735603e-02
1.01989615e+00 1.08160734e+00 -2.36543920e-03 5.84066093e-01
-1.42707825e-02 9.29710686e-01 -5.20262301e-01 -4.41399306e-01
4.62767810e-01 -2.77666271e-01 -7.73948848e-01 -1.47017874e-02
3.93484801e-01 3.35162461e-01 -1.24130361e-01 1.35822141e+00
6.24605298e-01 2.85275251e-01 3.72001708e-01 -1.48177958e+00
-6.98493063e-01 7.40838945e-01 -8.99512395e-02 -1.83660045e-01
-2.31425270e-01 -1.88655108e-01 9.59647834e-01 -4.43306029e-01
3.57704222e-01 1.00442946e+00 7.16854632e-01 7.38621891e-01
-1.17823148e+00 -5.87609529e-01 1.03030525e-01 2.04739809e-01
-1.25674736e+00 -3.81634891e-01 5.47972679e-01 -3.82855386e-01
2.62703121e-01 5.94619989e-01 4.45390642e-01 5.94346762e-01
3.06032777e-01 7.34774530e-01 8.56685698e-01 -6.49166644e-01
8.28779101e-01 4.13860083e-01 2.61515856e-01 7.66900182e-01
6.39245570e-01 8.33884776e-02 -5.88844895e-01 -8.11840475e-01
5.44808358e-02 4.00367320e-01 -4.63579386e-01 -9.40714180e-01
-8.36736858e-01 1.15856981e+00 4.08699036e-01 2.55397201e-01
-4.71366763e-01 -2.68448377e-03 1.27664715e-01 3.59504700e-01
8.49303544e-01 3.37654464e-02 -4.01288867e-01 5.60156047e-01
-6.73912644e-01 2.59509981e-01 1.39538002e+00 7.35349834e-01
9.59106565e-01 -3.85125726e-01 -9.01653990e-02 3.58691961e-01
5.05879164e-01 6.54128730e-01 2.10385084e-01 -9.61799979e-01
2.89505064e-01 4.34903234e-01 5.09287119e-01 -7.80527294e-01
-5.48103824e-02 -6.75060928e-01 -1.00426245e+00 2.55128503e-01
4.25401002e-01 -4.98842984e-01 -1.14329141e-02 1.86840665e+00
7.69826949e-01 1.37168691e-01 1.26994327e-01 9.04923379e-01
3.03928889e-02 5.41775405e-01 -3.12623292e-01 -4.90082413e-01
6.85493648e-01 -1.04136372e+00 -4.87859458e-01 3.31171900e-01
7.77256966e-01 -2.00525552e-01 3.26561511e-01 5.12694538e-01
-9.89858150e-01 4.24365625e-02 -7.13976264e-01 4.61449593e-01
-2.85344929e-01 -5.45567274e-01 5.66273868e-01 9.02258396e-01
-1.42346859e+00 8.33724499e-01 -8.10274363e-01 -5.50369740e-01
6.58852279e-01 6.99458480e-01 -3.59057784e-01 -1.24043673e-01
-6.94959164e-01 4.29902881e-01 6.58271909e-02 3.57829630e-02
-8.41416240e-01 -8.17198992e-01 -3.77227992e-01 3.79184753e-01
2.69493073e-01 -1.05020022e+00 9.69445169e-01 -9.26005960e-01
-1.08840692e+00 4.11710769e-01 -8.78523514e-02 -7.06056535e-01
7.18308389e-01 -3.15889083e-02 -2.30622590e-02 7.42923766e-02
6.24762848e-03 -2.41331324e-01 6.96848214e-01 -1.54039156e+00
-9.08868313e-01 -8.43448281e-01 -1.29780650e-01 6.94718063e-02
-7.67559350e-01 -1.86302185e-01 -1.03051633e-01 -1.09481148e-01
-3.34889054e-01 -7.92766869e-01 -5.18280447e-01 1.03640527e-01
-1.11712545e-01 -4.26627278e-01 1.10173202e+00 -5.59642613e-02
7.42235065e-01 -1.84987736e+00 1.66914910e-01 3.61922592e-01
4.78635937e-01 1.54809847e-01 -1.07233055e-01 5.98042786e-01
4.54434961e-01 1.30088240e-01 -7.86320865e-02 -8.98441672e-01
3.46713424e-01 2.34248355e-01 -1.62147522e-01 1.06250679e+00
-6.47252321e-01 5.40672898e-01 -1.03876746e+00 -5.50483525e-01
9.40888226e-02 5.25928974e-01 -5.50758064e-01 2.87544399e-01
-1.89025402e-01 5.41228771e-01 -8.55391026e-01 5.08483171e-01
8.57782543e-01 -5.68046749e-01 4.02240455e-01 1.12902403e-01
-1.21479928e-01 -4.60998833e-01 -1.17736912e+00 1.50200510e+00
-5.65460384e-01 2.90255211e-02 9.93741095e-01 -1.30454707e+00
4.38298821e-01 3.80206019e-01 1.03842795e+00 -2.10328698e-01
2.40397424e-01 1.31414264e-01 -5.15217125e-01 -2.74395555e-01
-6.93785176e-02 -2.06265256e-01 1.94858551e-01 9.03131127e-01
2.92026877e-01 5.72429955e-01 -7.26973340e-02 3.19160402e-01
1.14615881e+00 -6.71023726e-01 -1.32651418e-01 -7.94248402e-01
6.24921501e-01 -2.13717714e-01 5.52107751e-01 1.28616130e+00
-1.98557466e-01 1.27691701e-01 -9.85196084e-02 -6.69780374e-01
-8.28878105e-01 -8.69734347e-01 -1.16406813e-01 1.06428742e+00
4.21667278e-01 -3.34818184e-01 -9.16611969e-01 -9.45004046e-01
3.23341995e-01 5.03206789e-01 -6.92379653e-01 8.43628030e-03
-2.61893988e-01 -1.05036986e+00 1.79176897e-01 -3.32217574e-01
1.54114902e-01 -7.05102205e-01 -3.91382784e-01 2.56422102e-01
3.38242613e-02 -5.62017024e-01 -5.30738473e-01 2.15321660e-01
-6.93138540e-01 -1.50458169e+00 -3.89409393e-01 -6.25388741e-01
7.40672350e-01 2.63017148e-01 9.17364538e-01 2.89560914e-01
-2.12055624e-01 1.00331903e+00 -3.12911361e-01 -5.87436855e-01
-3.45461249e-01 -1.13009162e-01 4.28573847e-01 8.66364300e-01
1.88593373e-01 -5.19566000e-01 -6.76263094e-01 2.33796299e-01
-8.93154442e-01 -4.38315153e-01 1.86504081e-01 7.86314964e-01
3.99643123e-01 2.22624734e-01 4.89276916e-01 -1.03598976e+00
5.46484590e-01 -8.51396024e-01 -6.68923616e-01 6.81050599e-01
-9.64458764e-01 3.55692655e-02 1.08248627e+00 -1.22148909e-01
-8.80985916e-01 9.90600213e-02 5.16069591e-01 -5.03617108e-01
6.61653057e-02 1.00917138e-01 -3.01879317e-01 -6.81048870e-01
5.42139173e-01 1.77064002e-01 3.90068054e-01 -6.87230527e-01
4.89210844e-01 8.99373651e-01 3.44767362e-01 -1.05989528e+00
7.66506255e-01 9.62288618e-01 1.83656439e-01 -8.37332308e-01
-8.60926449e-01 -5.43108106e-01 -4.80140001e-01 -2.88259268e-01
4.63389426e-01 -6.82123899e-01 -1.40160859e+00 3.10679793e-01
-1.05720627e+00 -2.32448325e-01 -7.40758717e-01 5.33166409e-01
-4.92662609e-01 3.93192738e-01 -3.81442726e-01 -1.24160969e+00
-9.13501263e-01 -6.85675502e-01 6.40085518e-01 2.47163832e-01
5.78990281e-01 -1.58230996e+00 3.95263016e-01 5.36677003e-01
6.41101778e-01 2.01926827e-01 6.41000032e-01 -1.03985810e+00
-6.04275107e-01 -3.47668171e-01 2.39513993e-01 2.97782153e-01
3.09648886e-02 -4.98277336e-01 -7.10716546e-01 -1.00023222e+00
6.40989959e-01 -3.52386087e-01 6.19588435e-01 2.48494431e-01
1.09755540e+00 -1.06482613e+00 -6.39335990e-01 6.48733377e-01
1.76890755e+00 -3.19811791e-01 -2.48458549e-01 -5.57787856e-03
6.62937701e-01 6.50062323e-01 1.72127083e-01 8.79196763e-01
4.40295011e-01 3.87710005e-01 8.41745019e-01 8.42604488e-02
4.84801859e-01 8.37023258e-02 1.66348800e-01 6.25804842e-01
1.62288979e-01 -5.52287221e-01 -4.98630106e-01 5.38953125e-01
-2.31231451e+00 -9.98069823e-01 -1.62570894e-01 2.62248015e+00
4.62577343e-01 -7.30502546e-01 4.51226272e-02 -3.23423475e-01
1.06251860e+00 -1.23574175e-01 -5.52079260e-01 -7.88432509e-02
-1.95188805e-01 1.64993703e-01 7.92724967e-01 5.54804862e-01
-1.02036059e+00 3.00334871e-01 5.79909897e+00 6.55582070e-01
-1.04940486e+00 6.89957678e-01 5.15113413e-01 -4.13448364e-01
-1.91699713e-01 4.46411707e-02 -6.73723340e-01 6.61876261e-01
8.92818630e-01 -5.25180757e-01 7.95330763e-01 7.22155392e-01
2.22568542e-01 2.94559181e-01 -1.16827834e+00 8.76364946e-01
1.08330488e-01 -1.56788540e+00 -1.72887459e-01 5.88618636e-01
1.08220148e+00 5.32399833e-01 -3.72054242e-02 -1.15311503e-01
7.95141637e-01 -7.56627798e-01 5.26782930e-01 7.77280331e-01
9.68065783e-02 -7.95278907e-01 5.86780727e-01 8.14722002e-01
-8.70523870e-01 -6.19557559e-01 -4.99375343e-01 6.85425252e-02
-5.09795174e-02 7.25317955e-01 -1.09127589e-01 8.64645898e-01
8.40662062e-01 5.14865100e-01 -3.03934216e-01 1.35850966e+00
5.49294949e-01 6.25335336e-01 -6.68100893e-01 -7.97707662e-02
8.26169327e-02 -4.75670248e-01 6.56200051e-01 9.02475834e-01
3.04876536e-01 -1.47952922e-02 3.21486771e-01 7.64178813e-01
-3.22044492e-01 5.02230823e-01 -5.05456984e-01 3.70281965e-01
6.07491672e-01 1.51359558e+00 -1.98964849e-01 -1.32532209e-01
-4.34493065e-01 7.75692523e-01 7.11697936e-01 1.71985626e-01
-3.63914192e-01 -6.70534447e-02 6.68667555e-01 -3.81615385e-02
4.50473994e-01 9.70807970e-02 -2.73443665e-02 -1.36967635e+00
3.52053382e-02 -5.73296428e-01 7.87212789e-01 1.58587366e-01
-2.03877258e+00 3.14646095e-01 -4.19883937e-01 -9.23862815e-01
1.26135036e-01 -2.52335995e-01 -1.03040671e+00 7.31274188e-01
-1.32845855e+00 -1.33467448e+00 -3.39055061e-02 1.04679537e+00
-2.43671194e-01 -3.49288762e-01 9.28529024e-01 4.31637526e-01
-4.69376445e-01 6.05979025e-01 1.03503120e+00 -1.44992664e-01
5.97589493e-01 -1.27036405e+00 -3.55357587e-01 7.46393144e-01
2.09409758e-01 6.43081784e-01 6.15271926e-01 -3.43968660e-01
-1.97218740e+00 -1.48460758e+00 9.41304505e-01 -6.11568213e-01
6.71877027e-01 -5.16273797e-01 -6.68928623e-01 5.85700810e-01
8.38622582e-05 8.43875110e-01 9.20106351e-01 1.92345321e-01
1.14761265e-02 -5.17497241e-01 -1.56076908e+00 -2.92831454e-02
7.27599323e-01 -1.11509487e-01 -3.07226442e-02 8.92708898e-01
4.93930787e-01 8.76445621e-02 -8.89454842e-01 -1.74662516e-01
2.04395324e-01 -8.45261097e-01 7.63349712e-01 -9.31930244e-01
-5.77146292e-01 -2.85443217e-01 -3.64482820e-01 -1.33607519e+00
-2.56342888e-01 -1.01968622e+00 -5.71098089e-01 1.14427245e+00
4.65916544e-02 -8.94055307e-01 1.16684413e+00 8.09712887e-01
1.08620882e-01 -6.48736775e-01 -1.25779688e+00 -8.53549540e-01
3.63574445e-01 -2.97065489e-02 5.75950503e-01 1.12277699e+00
1.67378768e-01 6.90687224e-02 -4.35301840e-01 5.80972672e-01
1.61613917e+00 2.15866849e-01 6.94273412e-01 -1.69510627e+00
-4.16195214e-01 -9.06278938e-02 -1.56787112e-01 -4.90201980e-01
4.67189163e-01 -1.16566825e+00 -8.06532800e-02 -1.25430024e+00
2.89676875e-01 -1.03570521e+00 -5.45951545e-01 3.48536313e-01
1.62042707e-01 -6.68544173e-02 5.07465303e-01 4.48419422e-01
-1.30189753e+00 6.43730879e-01 7.16584682e-01 -3.50384116e-01
-6.32082596e-02 4.08749044e-01 -8.07806849e-01 3.57837021e-01
5.44836164e-01 -7.57911265e-01 -1.73287183e-01 -4.62681115e-01
-2.87189856e-02 -5.11655882e-02 6.46793723e-01 -1.02932096e+00
1.12559366e+00 -1.52903110e-01 6.17531314e-02 -1.30027995e-01
1.31989777e-01 -1.19788468e+00 2.30398074e-01 6.82644367e-01
-2.34638810e-01 -4.86428559e-01 -5.28373837e-01 9.77819204e-01
9.40512270e-02 -3.34255427e-01 7.85943806e-01 -2.91996747e-01
7.83775523e-02 8.08633208e-01 3.90652195e-03 8.78710300e-02
1.19318700e+00 3.84606600e-01 -4.36597973e-01 -4.76681054e-01
-7.14904904e-01 6.66602314e-01 6.12804472e-01 -4.22499962e-02
5.02774790e-02 -1.09410179e+00 -8.70462298e-01 2.06733271e-02
-1.19294718e-01 4.76324782e-02 3.30505490e-01 7.98159003e-01
-3.44013013e-02 3.76667887e-01 2.78052717e-01 -5.39532602e-01
-1.03005254e+00 8.16546381e-01 6.28265142e-01 -1.74641028e-01
-5.14963508e-01 6.14058852e-01 5.14441170e-02 -7.76990712e-01
4.39724833e-01 5.18702209e-01 2.51105905e-01 -2.15564251e-01
5.42766213e-01 7.00562179e-01 3.28478485e-01 -5.11784196e-01
-2.27150187e-01 1.31335020e-01 -1.53910089e-02 1.89040601e-01
1.53509569e+00 -2.58275896e-01 -5.37441432e-01 8.44881535e-02
1.33206666e+00 5.20255715e-02 -1.22443509e+00 -5.28046191e-01
-2.11264968e-01 -4.30323005e-01 2.88975574e-02 -6.53015375e-01
-1.22578681e+00 4.31305826e-01 7.48055279e-01 5.50193667e-01
6.99592650e-01 1.55894786e-01 6.82018578e-01 5.54614186e-01
9.34197068e-01 -8.67879212e-01 -4.59130466e-01 3.32140565e-01
2.80982763e-01 -1.22693002e+00 -2.20894650e-01 8.30596611e-02
1.42575866e-02 9.81128812e-01 4.06280637e-01 -1.04961284e-01
1.03335154e+00 1.67532295e-01 -1.56861126e-01 -2.11111248e-01
-1.12078428e+00 1.66303143e-01 -3.80086899e-02 6.31398201e-01
-7.77350664e-02 4.53151539e-02 -2.30321392e-01 7.47641444e-01
1.61720902e-01 -3.49899493e-02 1.97110787e-01 1.01227856e+00
-5.37487090e-01 -1.32737219e+00 -5.07956326e-01 4.58848745e-01
-6.08723879e-01 2.65951931e-01 -2.86809742e-01 2.90313631e-01
3.14292550e-01 1.17666185e+00 1.87798496e-02 2.60216780e-02
-1.26482457e-01 -4.83028553e-02 3.22442055e-01 -3.84421468e-01
-7.30974793e-01 -2.24693567e-01 -6.51140571e-01 -5.93416095e-01
-4.74770039e-01 -7.08955526e-01 -8.95812094e-01 -5.43364882e-01
-5.12391448e-01 1.07116008e+00 9.28892255e-01 1.04509008e+00
6.20139062e-01 -2.07541689e-01 1.29247785e+00 -7.69109845e-01
-1.03222454e+00 -5.42222202e-01 -1.02851093e+00 3.43858451e-01
6.12092078e-01 -1.29945174e-01 -8.89165998e-01 -1.77926406e-01] | [5.872522354125977, 6.174613952636719] |
c63a7d90-fe9c-4722-a49b-5b129b4d96f0 | learning-to-track-for-spatio-temporal-action | 1506.01929 | null | http://arxiv.org/abs/1506.01929v2 | http://arxiv.org/pdf/1506.01929v2.pdf | Learning to track for spatio-temporal action localization | We propose an effective approach for spatio-temporal action localization in
realistic videos. The approach first detects proposals at the frame-level and
scores them with a combination of static and motion CNN features. It then
tracks high-scoring proposals throughout the video using a
tracking-by-detection approach. Our tracker relies simultaneously on
instance-level and class-level detectors. The tracks are scored using a
spatio-temporal motion histogram, a descriptor at the track level, in
combination with the CNN features. Finally, we perform temporal localization of
the action using a sliding-window approach at the track level. We present
experimental results for spatio-temporal localization on the UCF-Sports, J-HMDB
and UCF-101 action localization datasets, where our approach outperforms the
state of the art with a margin of 15%, 7% and 12% respectively in mAP. | ['Zaid Harchaoui', 'Cordelia Schmid', 'Philippe Weinzaepfel'] | 2015-06-05 | learning-to-track-for-spatio-temporal-action-1 | http://openaccess.thecvf.com/content_iccv_2015/html/Weinzaepfel_Learning_to_Track_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Weinzaepfel_Learning_to_Track_ICCV_2015_paper.pdf | iccv-2015-12 | ['spatio-temporal-action-localization'] | ['computer-vision'] | [-1.04478680e-01 -5.94994485e-01 -4.93233353e-01 -8.43813792e-02
-1.24456549e+00 -6.92593634e-01 6.06229722e-01 2.80780375e-01
-8.47027481e-01 5.15444994e-01 3.12413514e-01 5.20603716e-01
1.05179362e-01 -5.17748594e-01 -7.62286484e-01 -6.36408925e-01
-5.07912517e-01 1.46470964e-01 1.39403164e+00 1.83261082e-01
2.00235814e-01 4.41475272e-01 -1.60695636e+00 8.32989275e-01
-8.50137621e-02 1.34104705e+00 7.52168149e-02 1.03103495e+00
3.77210021e-01 1.22667825e+00 -5.02757490e-01 -1.28361821e-01
3.17528754e-01 -2.41133168e-01 -5.72226524e-01 4.17087078e-02
1.05808663e+00 -3.75058651e-01 -5.96377313e-01 8.01318049e-01
3.54369640e-01 2.15488419e-01 2.79114038e-01 -1.20530105e+00
2.44171008e-01 4.96029109e-02 -6.00786865e-01 6.62577391e-01
6.88762069e-01 2.25893646e-01 8.03986490e-01 -9.89279330e-01
1.08294189e+00 1.28175890e+00 9.14115667e-01 2.78436840e-01
-8.27305436e-01 -4.95327383e-01 1.94921806e-01 5.05804479e-01
-1.37010515e+00 -2.50314236e-01 1.58361837e-01 -6.73456907e-01
1.05883944e+00 -1.01623252e-01 8.01711977e-01 7.98696876e-01
5.07585466e-01 1.13321769e+00 8.21267188e-01 -2.23068967e-01
2.71062791e-01 -5.57060540e-01 -1.92897230e-01 9.62423503e-01
-2.55518168e-01 3.65222752e-01 -1.09316897e+00 -7.08344355e-02
8.46888006e-01 1.74368881e-02 2.25849196e-01 -4.90940988e-01
-1.55187035e+00 5.31002522e-01 4.86374795e-01 3.63115489e-01
-7.32474864e-01 6.27077758e-01 7.30859458e-01 -1.16859548e-01
4.14620996e-01 -1.80843413e-01 -5.24928153e-01 -5.17176926e-01
-1.60772777e+00 4.98782635e-01 2.29354307e-01 8.29614699e-01
3.37333292e-01 -1.56050891e-01 -8.95455062e-01 3.28116179e-01
2.67187476e-01 4.38843399e-01 1.94303393e-01 -1.24452651e+00
4.34081733e-01 4.30274755e-01 5.26868045e-01 -9.84292030e-01
-3.46924186e-01 -2.24290967e-01 -1.04526401e-01 4.68047023e-01
6.66106105e-01 1.65666640e-01 -9.35141683e-01 1.57390261e+00
6.29218221e-01 7.60006487e-01 -4.47659165e-01 9.65294480e-01
5.69230795e-01 4.76085961e-01 4.39759195e-01 -9.53911394e-02
1.50673604e+00 -1.24029326e+00 -7.01504827e-01 1.82779998e-01
6.40160501e-01 -6.59542203e-01 5.01201808e-01 2.83460945e-01
-1.29862893e+00 -9.49883163e-01 -7.79320002e-01 1.21605538e-01
-3.72013777e-01 8.43329966e-01 1.51563764e-01 2.73540378e-01
-1.13899016e+00 6.11305237e-01 -1.29024351e+00 -6.81970656e-01
5.80107868e-01 2.89798051e-01 -5.94668329e-01 -7.99944974e-04
-8.67893219e-01 7.82423794e-01 5.20837665e-01 -2.19371781e-01
-1.40171969e+00 -4.90815282e-01 -7.86776483e-01 -8.21487457e-02
2.12578788e-01 -4.11639094e-01 1.24193335e+00 -7.01156259e-01
-1.21835780e+00 6.71416998e-01 -3.05000752e-01 -8.13040495e-01
9.07957375e-01 -3.60626578e-01 -4.55435634e-01 5.30780971e-01
4.08870280e-01 1.03600371e+00 5.98904908e-01 -7.51530647e-01
-1.58382702e+00 -1.22312933e-01 1.73917964e-01 -7.06497114e-03
3.40267606e-02 3.43921304e-01 -8.73823047e-01 -5.25119603e-01
-8.19126290e-05 -1.07110381e+00 -1.00489751e-01 4.16142315e-01
-2.01235898e-02 -3.80412996e-01 1.08082056e+00 -5.97481012e-01
1.28509414e+00 -2.20022917e+00 -3.96823399e-02 -1.63662389e-01
-1.62609205e-01 2.88770318e-01 -1.22205481e-01 2.97470719e-01
1.53376386e-01 -4.76379126e-01 4.40925628e-01 -4.25195247e-01
-1.56271383e-01 -1.73933432e-01 9.58073810e-02 7.44177163e-01
1.48409307e-01 9.01331007e-01 -1.17318416e+00 -8.84793341e-01
6.17785573e-01 4.98208016e-01 -4.19079512e-01 -8.08563754e-02
-1.61985084e-01 2.48108730e-01 -4.10977811e-01 9.75492954e-01
2.47177035e-01 -1.41257094e-02 -1.06245913e-01 -3.02684814e-01
-5.76459527e-01 8.67917687e-02 -1.42986393e+00 2.04213095e+00
1.73935249e-01 7.73517787e-01 -2.11799264e-01 -5.47055840e-01
4.30951238e-01 4.02054846e-01 8.32226872e-01 -5.85348666e-01
-2.27176055e-01 5.85605241e-02 -3.47370058e-01 -3.15593183e-01
7.43995428e-01 2.56739676e-01 -2.62054145e-01 -4.40027602e-02
4.02352273e-01 8.88593912e-01 6.05731547e-01 2.89454788e-01
1.70961893e+00 7.52733171e-01 2.17499942e-01 3.04259062e-02
5.21334767e-01 4.00304466e-01 5.75003743e-01 1.02022982e+00
-7.33884096e-01 6.29568219e-01 2.74890214e-01 -7.12236166e-01
-7.95039892e-01 -1.09208548e+00 1.34858325e-01 1.34220290e+00
2.23844647e-02 -5.51175594e-01 -7.38601625e-01 -9.91645217e-01
1.29002392e-01 1.77557394e-01 -9.75357831e-01 1.10982500e-01
-7.42762566e-01 -1.25717357e-01 4.64174300e-01 9.29470003e-01
5.57089806e-01 -1.19298148e+00 -1.02673626e+00 4.52239871e-01
-2.04030737e-01 -1.28998029e+00 -5.10711133e-01 1.37813672e-01
-8.87981713e-01 -1.04468989e+00 -7.92172551e-01 -4.87216830e-01
2.16017202e-01 2.87022412e-01 1.01976883e+00 -1.57937109e-01
-4.36470032e-01 4.40523714e-01 -3.96143049e-01 -4.57776673e-02
-9.16410461e-02 -1.02309242e-01 -1.50575321e-02 1.33121639e-01
3.71291667e-01 6.41278028e-02 -9.23826098e-01 5.56861937e-01
-5.12229085e-01 -4.63657528e-01 5.44540763e-01 4.80670899e-01
8.76145661e-01 -2.19104365e-01 8.71521309e-02 -1.03375629e-01
-2.51429677e-01 -1.70400232e-01 -7.80136049e-01 2.93467253e-01
2.10652217e-01 -4.82967764e-01 -1.74290165e-01 -5.91787159e-01
-6.91312075e-01 9.53757405e-01 3.08538407e-01 -9.06490684e-01
-3.31132591e-01 1.02063697e-02 1.97400928e-01 -1.82280168e-01
7.30453253e-01 2.98202217e-01 -4.29447412e-01 -3.22707415e-01
4.84785318e-01 1.13403097e-01 8.55266154e-01 -8.01665634e-02
6.03840649e-01 8.91718805e-01 -1.16826650e-02 -3.28905851e-01
-9.63825762e-01 -1.12597275e+00 -1.22078860e+00 -9.54040170e-01
1.50662243e+00 -1.31317902e+00 -7.05177784e-01 3.15661669e-01
-1.11554408e+00 -2.67255664e-01 -3.17322791e-01 7.85716653e-01
-7.92453825e-01 1.78945646e-01 -8.22978973e-01 -7.65849233e-01
1.09195866e-01 -9.90782261e-01 1.76868260e+00 1.02200747e-01
-2.87433952e-01 -6.82422757e-01 3.75744492e-01 1.82615131e-01
1.27677083e-01 6.16471529e-01 -3.84370722e-02 -2.80107617e-01
-1.05241072e+00 -6.50676906e-01 -6.93547130e-02 -3.16962227e-02
-2.92920470e-01 1.29471738e-02 -7.85820425e-01 -2.25473508e-01
-8.04184735e-01 -3.65463346e-01 1.21628678e+00 8.43815446e-01
8.42307925e-01 2.49796808e-01 -7.13501930e-01 2.10301906e-01
1.47049820e+00 1.66084245e-01 7.56035089e-01 5.64609826e-01
5.08126915e-01 1.60799012e-01 1.31839395e+00 5.60710371e-01
1.79267526e-01 1.27638197e+00 4.03552055e-01 -8.80771503e-02
-3.93209696e-01 -1.64860010e-01 6.88775837e-01 -2.94772536e-03
-3.85454118e-01 -9.99043882e-02 -6.63225174e-01 8.59909117e-01
-2.30997372e+00 -1.57136524e+00 -3.21087450e-01 2.13671088e+00
4.14705098e-01 3.04133654e-01 7.81323135e-01 7.29703158e-02
7.65220761e-01 1.62570775e-01 -5.97618185e-02 1.65192768e-01
1.09919673e-02 -1.07934259e-01 7.87483335e-01 9.34072733e-02
-1.94669676e+00 1.23009801e+00 6.69702578e+00 9.74828124e-01
-8.70848179e-01 5.06637633e-01 2.50992298e-01 -3.82256180e-01
9.90364194e-01 -2.18732119e-01 -1.10043943e+00 4.98923600e-01
7.87289679e-01 4.11352694e-01 -3.21354628e-01 1.07085967e+00
4.16235059e-01 -6.03327155e-01 -9.85987484e-01 7.90998340e-01
5.05774170e-02 -1.59768713e+00 -4.80745614e-01 2.30499674e-02
8.06600332e-01 2.84337848e-01 -1.56166732e-01 3.74901384e-01
1.52715370e-01 -5.80034971e-01 1.26569200e+00 7.57315755e-01
4.08465087e-01 -7.50529587e-01 6.96419895e-01 1.84486926e-01
-1.74523568e+00 -2.05879807e-01 -2.69115090e-01 1.06968179e-01
4.62791890e-01 1.59302324e-01 -4.73906368e-01 3.77612770e-01
1.04289818e+00 9.93656218e-01 -6.37537956e-01 1.52753389e+00
-5.83263785e-02 4.87656355e-01 -1.33704022e-01 2.13655323e-01
5.03963590e-01 4.06751961e-01 3.85713339e-01 1.72411823e+00
4.65910852e-01 3.89673188e-02 6.89143777e-01 2.91444778e-01
3.89656723e-01 -1.72570929e-01 -2.83330172e-01 2.82542109e-01
1.76270500e-01 1.44015408e+00 -9.61034715e-01 -6.76997960e-01
-3.56452674e-01 9.83615875e-01 1.27312705e-01 7.24469870e-02
-1.24246144e+00 6.00014627e-02 4.78978515e-01 3.52233112e-01
9.18487132e-01 -3.55154365e-01 4.10475016e-01 -8.96303892e-01
-1.06165763e-02 -4.99987930e-01 6.65909231e-01 -8.65520477e-01
-6.94789529e-01 4.49568719e-01 4.50333133e-02 -1.78866029e+00
-1.74051762e-01 -5.42977214e-01 -4.36979711e-01 3.62001956e-01
-1.12289703e+00 -1.30604041e+00 -3.82982433e-01 7.15887070e-01
6.69495881e-01 -1.74322367e-01 5.62886477e-01 7.03642368e-01
-3.15956384e-01 3.57226282e-01 -1.51124150e-01 4.83304501e-01
6.55171514e-01 -1.11261833e+00 2.60002702e-01 7.81940460e-01
5.12440741e-01 -2.91388165e-02 5.18931508e-01 -7.86266148e-01
-1.02111864e+00 -1.45606494e+00 1.10893857e+00 -9.08123195e-01
7.71592855e-01 -3.92148584e-01 -5.44534922e-01 5.15392184e-01
-3.32813174e-03 6.20236933e-01 1.76463857e-01 -2.93371707e-01
-6.77122176e-02 -4.26730551e-02 -9.76628661e-01 1.12756543e-01
1.14767301e+00 -4.32684213e-01 -3.10897022e-01 6.82717800e-01
2.59126037e-01 -7.00258136e-01 -8.91210377e-01 2.95097560e-01
8.46937299e-01 -9.85989988e-01 1.06761050e+00 -5.67033052e-01
2.84026533e-01 -8.17002416e-01 -2.65597403e-01 -6.76214397e-01
-6.85861111e-01 -3.04017574e-01 -1.60653546e-01 8.81308317e-01
1.66597724e-01 4.35500711e-01 9.32446599e-01 -2.21458986e-01
-7.83553049e-02 -5.70126534e-01 -1.36499453e+00 -1.06991386e+00
-5.76860785e-01 -5.71135998e-01 -1.37885600e-01 4.13115740e-01
-2.49810085e-01 -3.02021414e-01 -5.37508488e-01 1.12666249e-01
5.33847690e-01 -1.22030504e-01 7.81259537e-01 -8.39063525e-01
-2.10605443e-01 -2.48225600e-01 -1.18691576e+00 -9.03596282e-01
-5.64662628e-02 -4.58908200e-01 3.85331690e-01 -1.43458617e+00
3.99988443e-01 2.29824305e-01 -6.62382126e-01 5.39499283e-01
8.91965404e-02 7.06071973e-01 3.68568689e-01 2.89852619e-01
-1.62684357e+00 1.06393270e-01 7.61226475e-01 6.21822886e-02
2.13511214e-01 -1.29002497e-01 6.24679387e-01 7.02066600e-01
2.97513634e-01 -9.29506004e-01 1.59197569e-01 -2.62512472e-02
-2.51524389e-01 2.56420463e-01 8.13074887e-01 -1.77436256e+00
4.32609111e-01 4.58727032e-02 7.81416357e-01 -1.24252331e+00
6.43306375e-01 -5.69344401e-01 -6.18048087e-02 7.99055517e-01
-4.65797454e-01 3.55866939e-01 1.53375193e-01 9.32247281e-01
-4.36898470e-01 3.14802468e-01 9.69444931e-01 -1.92591235e-01
-1.19595051e+00 3.97341430e-01 -5.23036897e-01 -2.05516875e-01
1.42811227e+00 -3.73077929e-01 -1.47003323e-01 -2.42276490e-01
-1.15596318e+00 1.41802832e-01 2.86989659e-01 5.19410968e-01
4.46317643e-01 -1.69517314e+00 -5.81711411e-01 -2.33619869e-01
3.19782287e-01 -7.61223793e-01 3.05868655e-01 1.36960530e+00
-4.82524902e-01 6.41686857e-01 -3.66839290e-01 -1.14103007e+00
-1.70170319e+00 5.18030226e-01 3.34121227e-01 -3.92657250e-01
-4.10973072e-01 9.41576481e-01 1.40840318e-02 2.44845152e-01
6.49751127e-01 -4.14150268e-01 -2.20240116e-01 1.89354867e-01
8.30966294e-01 4.35968906e-01 -1.27351508e-01 -1.07205176e+00
-8.34503770e-01 6.76336825e-01 3.61030877e-01 -3.61223638e-01
9.51158762e-01 9.21788812e-02 4.11206007e-01 4.37590569e-01
1.04700947e+00 -3.29429716e-01 -1.69629169e+00 -2.22690076e-01
3.12326550e-01 -7.32544839e-01 2.51847804e-02 -8.58644605e-01
-1.01181614e+00 7.02821314e-01 1.11339152e+00 -1.27280593e-01
8.93696070e-01 2.56944925e-01 6.10711515e-01 2.14277804e-02
5.58278322e-01 -1.25919175e+00 5.06508827e-01 4.71064895e-01
6.15395665e-01 -1.15161967e+00 6.03765249e-02 -1.73890248e-01
-6.09414756e-01 9.84996140e-01 5.25013804e-01 -3.80782157e-01
5.02296865e-01 3.43124002e-01 3.83295259e-03 -2.15876326e-01
-7.96124399e-01 -6.36312127e-01 5.14651358e-01 4.48910385e-01
4.27181095e-01 -6.72579780e-02 -1.55203566e-01 -8.36924762e-02
5.27740896e-01 3.58078927e-01 1.83888264e-02 1.04686940e+00
-8.36414516e-01 -8.53353262e-01 -3.72136295e-01 -4.18102555e-02
-5.40038288e-01 2.85502136e-01 -3.62252653e-01 1.12200415e+00
6.36382639e-01 8.74479413e-01 1.76979125e-01 -5.18810570e-01
4.92050260e-01 1.03528239e-01 5.50916433e-01 -4.69059050e-01
-8.30979347e-01 5.31977892e-01 2.84730822e-01 -1.38326573e+00
-8.99805486e-01 -1.33731508e+00 -1.21572423e+00 1.14214011e-01
-1.75239742e-01 -2.23995626e-01 3.98682535e-01 8.63521695e-01
4.24749792e-01 4.90089327e-01 1.47033140e-01 -1.09987414e+00
-2.60578215e-01 -9.22419190e-01 -4.51411605e-01 3.50282222e-01
1.29461691e-01 -9.80920970e-01 1.82941675e-01 2.74734467e-01] | [8.290385246276855, 0.41444993019104004] |
2d432c70-6abd-4c87-9dfd-2270ce36156f | nlm_nih-at-semeval-2017-task-3-from-question | null | null | https://aclanthology.org/S17-2057 | https://aclanthology.org/S17-2057.pdf | NLM\_NIH at SemEval-2017 Task 3: from Question Entailment to Question Similarity for Community Question Answering | This paper describes our participation in SemEval-2017 Task 3 on Community Question Answering (cQA). The Question Similarity subtask (B) aims to rank a set of related questions retrieved by a search engine according to their similarity to the original question. We adapted our feature-based system for Recognizing Question Entailment (RQE) to the question similarity task. Tested on cQA-B-2016 test data, our RQE system outperformed the best system of the 2016 challenge in all measures with 77.47 MAP and 80.57 Accuracy. On cQA-B-2017 test data, performances of all systems dropped by around 30 points. Our primary system obtained 44.62 MAP, 67.27 Accuracy and 47.25 F1 score. The cQA-B-2017 best system achieved 47.22 MAP and 42.37 F1 score. Our system is ranked sixth in terms of MAP and third in terms of F1 out of 13 participating teams. | ['Dina Demner-Fushman', 'Asma Ben Abacha'] | 2017-08-01 | null | null | null | semeval-2017-8 | ['question-similarity'] | ['natural-language-processing'] | [-1.12847254e-01 -8.56713355e-02 6.67288065e-01 -1.24786265e-01
-1.70959651e+00 -8.32773805e-01 7.92900801e-01 5.79931796e-01
-7.76885986e-01 5.23328125e-01 4.12656665e-01 -4.71500486e-01
-4.75080490e-01 -5.84693968e-01 -5.90037227e-01 -4.65874486e-02
3.44322890e-01 5.18414319e-01 6.26836896e-01 -6.71380103e-01
7.74456441e-01 -3.34417641e-01 -1.55235469e+00 9.65622008e-01
1.17990804e+00 1.16618598e+00 -1.73498914e-02 1.07267559e+00
-1.01422764e-01 1.03958929e+00 -9.36746359e-01 -6.74683511e-01
-2.49624863e-01 -4.77188081e-01 -1.70344615e+00 -5.50145447e-01
1.25677657e+00 1.47887826e-01 -1.87686518e-01 4.92016822e-01
5.44041336e-01 7.51379505e-02 6.03107214e-01 -1.16053092e+00
-7.01083839e-01 1.15922708e-02 1.21055372e-01 7.31721222e-01
1.16434002e+00 -6.54526576e-02 1.56626773e+00 -1.28219950e+00
6.61530912e-01 9.08220053e-01 5.71146667e-01 4.29610044e-01
-9.38521206e-01 -3.85003090e-01 -3.25095326e-01 6.89486265e-01
-1.24187100e+00 -2.65754789e-01 -1.11530878e-01 -5.26254177e-01
1.39344454e+00 7.01936424e-01 1.81351975e-01 5.42069972e-01
-5.67300208e-02 6.90996051e-01 1.29630637e+00 -2.71099657e-01
1.85798988e-01 2.66092438e-02 5.03041804e-01 5.00332654e-01
2.26307005e-01 -8.58656704e-01 -7.71419883e-01 -5.14152706e-01
-3.60563159e-01 -6.42908812e-01 -4.01953846e-01 4.37507212e-01
-1.13965559e+00 8.36709738e-01 4.92765218e-01 4.38601434e-01
-3.47470075e-01 -1.53402671e-01 1.89115688e-01 8.07423055e-01
1.79041550e-01 1.35733390e+00 -5.62156618e-01 -2.05711052e-01
-6.93774343e-01 1.13866699e+00 1.19017339e+00 7.95918882e-01
4.76199210e-01 -7.65776277e-01 -8.66878510e-01 9.38339710e-01
1.75249115e-01 7.43081689e-01 3.13105553e-01 -1.34100378e+00
6.66425705e-01 8.28376353e-01 2.85486102e-01 -1.03720975e+00
-4.51784223e-01 -5.91956317e-01 -3.74067724e-01 -6.39497519e-01
7.54063189e-01 1.58251867e-01 -2.34052882e-01 1.46534920e+00
3.24085131e-02 -4.66618150e-01 6.03512377e-02 7.46038914e-01
1.44842899e+00 6.85892999e-01 -5.30948602e-02 1.45120218e-01
1.91744637e+00 -9.32807326e-01 -5.97623765e-01 -1.96702048e-01
8.80690694e-01 -1.23663211e+00 1.22590578e+00 3.65354151e-01
-1.29176497e+00 -5.26611030e-01 -7.00245321e-01 -1.81214437e-01
-3.24365735e-01 -1.76277429e-01 -1.85055673e-01 5.50745845e-01
-1.48482680e+00 1.71792150e-01 2.88458616e-01 -6.63110614e-01
1.55320078e-01 -1.00197621e-01 -1.44693270e-01 -5.88883281e-01
-1.52575469e+00 1.01145577e+00 -3.14516604e-01 -5.20269811e-01
-6.91991150e-01 -9.96431768e-01 -4.11614269e-01 1.99869245e-01
2.88727552e-01 -8.93281102e-01 1.74663055e+00 -9.61451605e-02
-7.45784581e-01 1.12455440e+00 -3.54345530e-01 -4.28443819e-01
2.03099832e-01 -3.28882843e-01 -5.59538484e-01 6.12328410e-01
7.12022841e-01 4.43463862e-01 1.91613942e-01 -6.78777516e-01
-8.11219692e-01 -3.15826446e-01 3.31747919e-01 7.14500397e-02
-2.95437187e-01 4.16550845e-01 -9.85618383e-02 -3.01019531e-02
1.01002581e-01 -7.57812977e-01 2.21469522e-01 -5.07128477e-01
-5.45140030e-03 -1.10592735e+00 4.98762310e-01 -1.00705659e+00
1.49633634e+00 -1.46496463e+00 -4.69145358e-01 -5.51848114e-02
5.36355853e-01 3.85924667e-01 -3.95904779e-01 9.83943641e-01
2.19455093e-01 3.05772036e-01 -9.74717513e-02 -1.49574846e-01
1.07104175e-01 -2.71746665e-01 -2.98335791e-01 7.40549862e-02
3.59586716e-01 9.36837792e-01 -1.19411659e+00 -5.09300828e-01
-6.13383949e-01 -7.56399482e-02 -6.68782771e-01 5.40237486e-01
-2.85387129e-01 1.79335564e-01 -4.35942709e-01 5.27372181e-01
5.04963577e-01 -5.83549976e-01 -3.03372085e-01 2.55883634e-01
-6.39144406e-02 9.73375320e-01 -6.42316580e-01 1.45837653e+00
-4.68692183e-01 9.18839872e-01 -1.22653246e-01 -6.01000190e-01
1.01073265e+00 5.79348445e-01 2.42770165e-01 -1.22501719e+00
-3.05923700e-01 4.72714990e-01 2.84266829e-01 -9.14029300e-01
7.05975950e-01 3.33423257e-01 -3.35155159e-01 7.07354367e-01
1.25251822e-02 -3.56114209e-01 4.53312665e-01 7.41287231e-01
1.81723773e+00 -4.83579308e-01 6.33071177e-03 -7.82594383e-01
1.12396383e+00 8.14708173e-02 -8.79245922e-02 1.17695796e+00
-5.01494229e-01 7.78906286e-01 5.91407478e-01 -3.23045403e-01
-7.56296456e-01 -9.46772039e-01 -1.44060642e-01 1.42066777e+00
-2.73137033e-01 -8.10355663e-01 -7.40202308e-01 -6.04971588e-01
-8.83157924e-02 8.77918482e-01 -5.48537016e-01 -2.52301246e-02
-6.36691391e-01 -1.54451951e-01 7.50314355e-01 7.44224340e-02
6.02096498e-01 -1.04702497e+00 -3.11038852e-01 6.44981638e-02
-1.04090917e+00 -1.18957794e+00 -6.21953666e-01 -6.29906952e-01
-3.89614522e-01 -1.53060973e+00 -8.79497290e-01 -9.73173797e-01
4.44210172e-02 4.77693170e-01 1.77771330e+00 3.83378148e-01
-7.18954206e-02 9.06804204e-01 -7.10013747e-01 -2.66349167e-01
-2.24689394e-01 4.95307356e-01 -5.49221098e-01 -2.76473314e-01
5.82644641e-01 -4.12878767e-02 -9.84678209e-01 5.43706834e-01
-6.10065877e-01 -5.74992597e-01 1.95012957e-01 6.39224410e-01
1.28919616e-01 -8.01049471e-01 1.13434494e+00 -3.62659931e-01
1.18700838e+00 -7.32634723e-01 -6.17473014e-02 5.02195060e-01
-9.71162319e-01 -2.62341231e-01 4.10211116e-01 2.08186164e-01
-8.29195440e-01 -7.50926554e-01 -3.01150471e-01 4.95129526e-01
3.25905755e-02 3.16058934e-01 2.39065811e-01 1.33682698e-01
1.15308869e+00 1.51436061e-01 -3.40586990e-01 -5.07641554e-01
8.86274651e-02 9.69753861e-01 5.72592437e-01 -5.48944235e-01
3.57507020e-01 2.85683870e-02 -2.95052022e-01 -7.07770109e-01
-1.33827710e+00 -1.21783733e+00 -1.52323753e-01 -5.16224444e-01
1.03357291e+00 -7.55787373e-01 -1.25116253e+00 2.93806672e-01
-1.44085538e+00 1.44399717e-01 8.93940553e-02 5.22714965e-02
-2.52546042e-01 3.19838285e-01 -4.25663173e-01 -6.33126318e-01
-8.24456930e-01 -7.53566265e-01 9.64642882e-01 2.10545912e-01
-5.62949777e-01 -5.50964594e-01 4.77029860e-01 1.29262185e+00
7.71886528e-01 -1.59699738e-01 6.85177088e-01 -1.06099701e+00
-1.63397521e-01 -2.33076885e-01 -2.44717047e-01 2.55225748e-01
-2.40343317e-01 -4.71656233e-01 -9.79340672e-01 -3.29414606e-01
-7.82295316e-02 -8.32363129e-01 8.03869784e-01 -2.53611177e-01
9.35168624e-01 -2.14228913e-01 2.16488943e-01 -3.14264029e-01
1.14097393e+00 -2.37028763e-01 5.05629420e-01 7.11979449e-01
1.57054454e-01 9.09100533e-01 5.41528285e-01 1.58877134e-01
1.00237226e+00 6.32294059e-01 4.09523517e-01 6.90891802e-01
-1.98599771e-01 -1.04050949e-01 2.39495859e-01 1.03821218e+00
2.02538610e-01 -2.89211214e-01 -1.53579760e+00 1.06824589e+00
-1.69931996e+00 -8.99963439e-01 -1.02524400e+00 1.96210027e+00
7.53086090e-01 -1.13865644e-01 2.82861024e-01 -2.13911086e-02
5.30530691e-01 -6.42247722e-02 9.31807887e-03 -5.53814769e-01
-1.56661302e-01 4.96722668e-01 -1.94160655e-01 4.19144839e-01
-7.77162313e-01 4.11646634e-01 5.73309612e+00 6.80203259e-01
-9.33732241e-02 3.82151842e-01 4.35350955e-01 1.09325476e-01
-4.49484289e-01 -1.14500355e-02 -6.05878770e-01 3.45266014e-01
1.43725431e+00 -3.24897468e-01 1.05816051e-01 4.59583312e-01
-2.71577924e-01 -3.29065412e-01 -1.00286078e+00 7.73614109e-01
5.06429672e-01 -1.18349576e+00 3.82249169e-02 -2.85878271e-01
1.02047181e+00 2.00566962e-01 1.85234770e-02 8.12325954e-01
-4.93117096e-03 -1.32115841e+00 4.65123713e-01 7.75440514e-01
3.00468624e-01 -5.17373681e-01 1.21309996e+00 4.90877002e-01
-8.17382991e-01 3.67820561e-02 -2.89406210e-01 -3.85816425e-01
-5.88789135e-02 2.66927510e-01 -9.99092042e-01 5.10768890e-01
1.25821579e+00 1.32014170e-01 -1.16334093e+00 1.21499479e+00
-3.22207570e-01 9.37398076e-01 -1.38909936e-01 -6.29771113e-01
4.57516313e-01 2.55670488e-01 5.34949839e-01 1.35400975e+00
5.43429665e-02 1.41231015e-01 -2.54354417e-01 4.93122071e-01
-3.94756913e-01 3.34573448e-01 -1.72638521e-01 3.14241201e-01
6.15718722e-01 1.12385094e+00 2.08391785e-03 -4.23702210e-01
-2.44429544e-01 6.64921105e-01 4.01762426e-01 1.33449033e-01
-5.07793725e-01 -7.09373534e-01 2.45051190e-01 2.41903007e-01
3.01578850e-01 2.22208127e-01 -2.23357230e-02 -8.89704466e-01
4.42555994e-01 -1.21196377e+00 7.21482396e-01 -9.16573822e-01
-1.58428001e+00 8.47266853e-01 -2.71057069e-01 -8.47042918e-01
-2.86225349e-01 -3.91823471e-01 -6.28522336e-01 1.00247228e+00
-1.49672163e+00 -8.11111212e-01 -5.83865166e-01 5.54030895e-01
3.97681087e-01 -1.51911259e-01 9.22915816e-01 3.84429634e-01
-4.80256267e-02 6.65192962e-01 1.25055924e-01 1.14921276e-02
1.19324636e+00 -1.47654271e+00 4.47418302e-01 4.55699742e-01
2.66111214e-02 6.59769416e-01 6.80364549e-01 -2.50680268e-01
-1.09941554e+00 -8.58624637e-01 1.99265337e+00 -1.06617129e+00
8.25611234e-01 -1.49569646e-01 -1.22075891e+00 1.79619148e-01
6.67068243e-01 -3.69945168e-01 8.88889194e-01 2.44410455e-01
-6.63704574e-01 -1.62753239e-01 -1.15080488e+00 2.50199407e-01
6.99858546e-01 -1.10324240e+00 -1.10976875e+00 6.39845848e-01
8.04378867e-01 -1.79051328e-02 -1.40218401e+00 2.66806334e-01
3.14698786e-01 -1.01034069e+00 8.41699362e-01 -8.53824377e-01
5.65365672e-01 -3.19786340e-01 -3.59910846e-01 -7.80568361e-01
-3.79900366e-01 -3.91641498e-01 1.31324306e-01 1.11963999e+00
7.18723476e-01 -6.29887700e-01 4.02705163e-01 4.93219256e-01
-2.43334100e-01 -9.13893938e-01 -1.19112003e+00 -5.80768704e-01
6.07976079e-01 -2.06487834e-01 4.51404393e-01 8.80417109e-01
8.45223013e-03 7.55428910e-01 3.95655185e-01 -1.72493055e-01
2.82793313e-01 2.60771245e-01 7.35665917e-01 -1.13635421e+00
-9.51020494e-02 -6.57720804e-01 2.80155074e-02 -9.59627688e-01
-1.67278498e-01 -1.02809393e+00 2.87784506e-02 -1.91372514e+00
5.26808143e-01 1.45167828e-01 -1.43948883e-01 -2.80855838e-02
-6.08524323e-01 3.63380581e-01 3.07590812e-01 3.34764451e-01
-1.40290534e+00 2.87283927e-01 1.16608036e+00 -7.81627893e-02
5.35949767e-01 8.90862569e-02 -9.27725196e-01 1.79701954e-01
1.08321440e+00 -4.92032796e-01 -9.57422107e-02 -4.53284264e-01
7.32812762e-01 -1.93825606e-02 4.87999767e-01 -9.99617100e-01
5.01838207e-01 3.63838315e-01 -1.23469494e-01 -6.59555435e-01
1.38214771e-02 -9.68857035e-02 -5.63420355e-01 6.25859261e-01
-6.64654136e-01 5.12034416e-01 7.35445842e-02 3.00261647e-01
-3.91199619e-01 -4.97094035e-01 2.82108873e-01 -1.55617911e-02
-2.18321487e-01 -1.59963503e-01 -6.53192580e-01 9.81509864e-01
5.44052839e-01 1.94220245e-01 -8.96537662e-01 -7.72803366e-01
-3.33443671e-01 6.19747996e-01 -4.81611267e-02 5.69782853e-01
4.98862863e-01 -1.17246497e+00 -1.51481795e+00 -4.68394667e-01
6.81611776e-01 -1.81646109e-01 4.72023964e-01 8.93068552e-01
-4.85557735e-01 1.06690776e+00 1.84594855e-01 -5.77375591e-01
-1.30214369e+00 -3.02323066e-02 2.82888889e-01 -7.21165538e-01
2.01823339e-01 9.62511897e-01 -3.56340230e-01 -6.86598182e-01
-1.41572878e-01 -9.87905413e-02 -5.67494810e-01 4.85680178e-02
9.06814933e-01 9.31929469e-01 6.51292980e-01 -5.01651227e-01
-5.47933817e-01 4.44892347e-01 -7.92887062e-02 -2.69528866e-01
9.80877578e-01 7.05764890e-02 -4.91650373e-01 1.63475528e-01
1.50482082e+00 7.76308822e-03 -1.19608760e-01 -3.75808090e-01
5.55879474e-01 -3.13785464e-01 -2.43903384e-01 -1.34479189e+00
-1.48351923e-01 8.19415152e-01 2.78792441e-01 6.09796584e-01
8.25540543e-01 3.96805644e-01 9.01795745e-01 8.21588039e-01
-1.36255973e-03 -1.01218367e+00 5.23967206e-01 1.07560444e+00
1.41530192e+00 -1.25802445e+00 -1.65645778e-01 5.43712731e-03
-4.80647683e-01 7.99198687e-01 7.96325147e-01 -5.62444702e-02
3.10959369e-01 -7.13580251e-01 5.19145541e-02 -7.86586106e-01
-1.09351909e+00 -2.56233633e-01 8.90957177e-01 3.51485938e-01
6.65590525e-01 -1.59615487e-01 -6.91645265e-01 7.30890095e-01
-5.73805213e-01 -4.90646333e-01 4.02470201e-01 7.32015312e-01
-6.93463802e-01 -6.01952791e-01 -4.32734340e-01 5.44125974e-01
-7.65815020e-01 -2.08112508e-01 -6.25299335e-01 5.81411421e-01
-2.44342595e-01 1.89294636e+00 2.05869243e-01 -4.37855333e-01
6.39932692e-01 2.45526299e-01 2.01622218e-01 -5.03287673e-01
-1.39554155e+00 -6.32976413e-01 6.59235656e-01 -5.86861134e-01
-2.49280095e-01 -9.36380446e-01 -8.81872177e-01 -4.74437684e-01
-2.27889076e-01 6.97314978e-01 4.90728021e-01 8.68716300e-01
5.81173062e-01 2.53768891e-01 6.42029226e-01 5.01969099e-01
-7.17957139e-01 -1.45534921e+00 -2.64028274e-02 5.95295727e-01
2.65943646e-01 9.18924343e-03 -3.42693508e-01 -3.57949257e-01] | [11.377742767333984, 8.037906646728516] |
9250e1f7-5c0f-48dd-afd1-2b14cfb74c27 | entity-aware-negative-sampling-with-auxiliary | 2210.06242 | null | https://arxiv.org/abs/2210.06242v1 | https://arxiv.org/pdf/2210.06242v1.pdf | Entity Aware Negative Sampling with Auxiliary Loss of False Negative Prediction for Knowledge Graph Embedding | Knowledge graph (KG) embedding is widely used in many downstream applications using KGs. Generally, since KGs contain only ground truth triples, it is necessary to construct arbitrary negative samples for representation learning of KGs. Recently, various methods for sampling high-quality negatives have been studied because the quality of negative triples has great effect on KG embedding. In this paper, we propose a novel method called Entity Aware Negative Sampling (EANS), which is able to sample negative entities resemble to positive one by adopting Gaussian distribution to the aligned entity index space. Additionally, we introduce auxiliary loss for false negative prediction that can alleviate the impact of the sampled false negative triples. The proposed method can generate high-quality negative samples regardless of negative sample size and effectively mitigate the influence of false negative samples. The experimental results on standard benchmarks show that our EANS outperforms existing the state-of-the-art methods of negative sampling on several knowledge graph embedding models. Moreover, the proposed method achieves competitive performance even when the number of negative samples is limited to only one. | ['Sang-hyun Je'] | 2022-10-12 | null | null | null | null | ['knowledge-graph-embedding'] | ['graphs'] | [-2.09480464e-01 3.83448213e-01 -6.41305029e-01 -1.79346338e-01
-5.85120916e-01 -3.71581465e-01 4.66084540e-01 2.17422143e-01
-3.87330681e-01 1.06052423e+00 -1.03167355e-01 1.78153701e-02
-4.59200256e-02 -1.57321930e+00 -8.64427328e-01 -6.18413985e-01
-1.41243180e-02 6.54691100e-01 5.75499594e-01 -1.57117993e-01
-5.78862103e-03 4.49455418e-02 -1.21497118e+00 -8.18211734e-02
1.14412987e+00 7.57785022e-01 5.80623932e-02 1.20482191e-01
-2.57560492e-01 6.78426206e-01 -6.10058546e-01 -9.56507146e-01
3.42205435e-01 -1.52338237e-01 -5.05655229e-01 -4.83937532e-01
4.17565927e-02 -1.76068798e-01 -2.48880357e-01 1.38354433e+00
6.70008957e-01 5.78990839e-02 9.34819937e-01 -1.73741698e+00
-9.82005000e-01 9.26736891e-01 -8.59666526e-01 1.65768892e-01
7.38771781e-02 -6.63833469e-02 1.24545372e+00 -1.18619108e+00
7.90160358e-01 1.34470069e+00 6.07626438e-01 3.23303342e-01
-1.05863440e+00 -9.07763779e-01 7.86821470e-02 4.97246861e-01
-1.62424970e+00 -1.25193313e-01 8.95000517e-01 -3.60728651e-01
5.15499830e-01 1.64755493e-01 9.74348247e-01 9.29491520e-01
-2.28628740e-01 1.07187390e+00 6.50070310e-01 -4.16499078e-01
2.32302293e-01 5.26386321e-01 3.58006686e-01 6.70455515e-01
1.16471601e+00 -7.16786012e-02 -4.60411996e-01 -5.17223477e-01
4.46837693e-01 -1.92031711e-01 -5.20374060e-01 -9.85756338e-01
-1.14772737e+00 8.62658441e-01 5.38459539e-01 3.12407650e-02
-2.60958135e-01 2.21411452e-01 5.97236335e-01 4.15939763e-02
4.23562139e-01 1.03194073e-01 -6.41239405e-01 1.81610540e-01
-3.99137795e-01 1.88070223e-01 7.76851416e-01 1.12952507e+00
8.83594871e-01 7.44468346e-02 -3.75222862e-01 7.16030359e-01
3.63529325e-01 5.71893275e-01 3.32507789e-01 -1.81384549e-01
8.57700646e-01 8.95619929e-01 2.73571342e-01 -1.30775678e+00
-5.88808879e-02 -4.45681959e-01 -9.48365271e-01 -3.56956542e-01
1.14056081e-01 -3.54172178e-02 -7.91826844e-01 1.58636737e+00
6.72979176e-01 3.21879625e-01 2.59654015e-01 7.86001086e-01
9.75009441e-01 4.64583427e-01 2.95314312e-01 -2.44267881e-01
1.03106701e+00 -8.29573750e-01 -8.36791992e-01 2.19425354e-02
9.63549733e-01 -2.96117783e-01 1.08315647e+00 1.30585074e-01
-5.72168410e-01 1.01232957e-02 -1.02560723e+00 1.75109714e-01
-5.30904114e-01 5.26945889e-01 9.26293194e-01 8.41937542e-01
-6.59022510e-01 4.08709437e-01 -5.69187224e-01 -1.63919274e-02
5.95671535e-01 3.61028343e-01 -4.92766768e-01 -1.68175206e-01
-1.67693448e+00 5.28222859e-01 8.68906558e-01 2.45209455e-01
-5.11328518e-01 -9.50227797e-01 -9.65013444e-01 2.41082072e-01
6.07514262e-01 -7.47848570e-01 5.23712814e-01 -6.70584202e-01
-8.63362014e-01 4.61103052e-01 -6.23814240e-02 -2.46912584e-01
6.56910181e-01 -6.93682358e-02 -5.82253814e-01 -2.75620651e-02
2.79675156e-01 1.69738576e-01 6.03693843e-01 -1.35387051e+00
-2.81604111e-01 -5.12865901e-01 6.43102303e-02 7.75143579e-02
-7.82061458e-01 -6.04319036e-01 -5.28157115e-01 -4.78289872e-01
8.95528644e-02 -6.40546143e-01 -1.72194496e-01 -2.44052678e-01
-7.11096466e-01 -5.92320502e-01 5.68127453e-01 -3.27236503e-01
1.33820903e+00 -1.91797006e+00 -7.42121460e-03 6.22915030e-01
2.57994413e-01 4.79285508e-01 1.54807255e-01 5.43938994e-01
-3.84231396e-02 4.69891965e-01 1.43957600e-01 7.40140676e-02
1.41314834e-01 -4.84293625e-02 -2.21941605e-01 2.85117954e-01
2.35786378e-01 8.49892259e-01 -1.35672021e+00 -7.38372922e-01
1.10905565e-01 3.43374938e-01 -4.47056919e-01 -1.29514396e-01
-6.95213825e-02 -2.99424559e-01 -7.03411698e-01 7.54796028e-01
9.96655583e-01 -4.29538935e-01 4.67656910e-01 -7.58306682e-01
6.94508255e-01 -7.00561777e-02 -1.47691178e+00 8.32518280e-01
1.38176501e-01 -4.67050970e-02 -3.31938088e-01 -1.04739881e+00
8.91430318e-01 7.33181369e-03 -9.93067473e-02 -2.45474041e-01
9.70255509e-02 2.65381664e-01 -1.97056219e-01 -4.18985128e-01
7.81240284e-01 -2.56170779e-01 7.13227242e-02 1.47204092e-02
3.34725007e-02 3.19966763e-01 3.89988452e-01 6.33152843e-01
9.16093409e-01 -1.34785220e-01 4.32056755e-01 -8.23356137e-02
6.73320651e-01 -1.77375719e-01 1.10174513e+00 5.82001388e-01
-4.07066494e-01 1.28563285e-01 8.44539404e-01 -1.31707519e-01
-7.05571711e-01 -1.11015570e+00 1.27452776e-01 6.15887523e-01
5.40970445e-01 -6.63598418e-01 -3.61088663e-01 -1.20125413e+00
4.20490503e-01 5.70007026e-01 -8.11236262e-01 -5.19143641e-01
-1.96820021e-01 -1.07167840e+00 5.23441672e-01 7.15782464e-01
4.31009978e-01 -9.53997493e-01 5.68225622e-01 -5.94379380e-02
-1.66282848e-01 -7.16892302e-01 -1.99303374e-01 -2.34907731e-01
-6.40442491e-01 -1.51783252e+00 -4.46166992e-01 -8.80226851e-01
9.68335271e-01 2.63841391e-01 1.17436194e+00 -2.59314366e-02
8.52216333e-02 1.41708657e-01 -4.79080200e-01 -3.60423058e-01
-3.41960192e-02 2.68093586e-01 1.38128921e-01 -7.93226734e-02
7.32232153e-01 -4.40821588e-01 -4.84058857e-01 -8.69093388e-02
-8.56716931e-01 -2.09716037e-01 7.62528121e-01 1.04921567e+00
7.67220557e-01 2.20931426e-01 1.30762267e+00 -1.50273049e+00
6.99793756e-01 -6.68415666e-01 -4.48176950e-01 6.41106069e-01
-8.04549694e-01 1.98304787e-01 8.45836580e-01 -4.31109279e-01
-8.37820351e-01 -3.92022252e-01 2.37673432e-01 -4.82474923e-01
6.26452386e-01 5.58263063e-01 -5.07125497e-01 -1.75183490e-01
5.27488410e-01 2.30620444e-01 -3.43172133e-01 -2.12821811e-01
4.30937707e-01 4.42385882e-01 2.62233801e-02 -5.95544815e-01
1.07523263e+00 3.17856640e-01 1.39196858e-01 -5.37214041e-01
-6.83848917e-01 -3.50890785e-01 -1.74801141e-01 -2.20339522e-02
2.52083004e-01 -9.63784695e-01 -8.27628791e-01 3.28626484e-01
-6.80575311e-01 1.91990361e-01 -2.34590724e-01 5.86422980e-01
-1.74020484e-01 6.22811615e-01 -4.06232655e-01 -1.06353462e+00
-5.71072102e-01 -6.72066092e-01 8.09738517e-01 1.63993686e-01
2.48701498e-01 -8.97176504e-01 -1.17858887e-01 2.02279717e-01
2.60935724e-02 2.47958273e-01 1.04562807e+00 -7.53552139e-01
-9.06593084e-01 -3.37103218e-01 -2.51159191e-01 3.36203426e-01
3.82238217e-02 -1.03176638e-01 -6.31275892e-01 -3.59602779e-01
-6.92770600e-01 -4.53718483e-01 1.02407634e+00 -6.98102489e-02
1.09580731e+00 -2.51134336e-01 -7.99454212e-01 3.12351286e-01
1.58062828e+00 -2.81225652e-01 7.59199679e-01 3.10214818e-01
9.49004531e-01 2.51539469e-01 1.23189533e+00 4.64984596e-01
5.50947487e-01 2.01809764e-01 4.55043942e-01 1.46218657e-01
3.99718136e-01 -5.86504638e-01 5.62775172e-02 7.08820581e-01
1.79679282e-02 -2.97362924e-01 -5.05110860e-01 1.01523423e+00
-1.84374607e+00 -1.04295087e+00 -3.51793170e-01 2.36166739e+00
1.06703794e+00 2.10864484e-01 -5.93663566e-02 3.22458655e-01
9.86162841e-01 1.34200215e-01 -5.22038579e-01 3.33298564e-01
-7.25095868e-02 -2.88622286e-02 7.45050669e-01 1.39220253e-01
-8.45603943e-01 8.20203841e-01 5.22004747e+00 1.21257424e+00
-7.50323832e-01 -8.40340853e-02 4.22319829e-01 7.82648772e-02
-8.73900294e-01 -3.07472907e-02 -1.19942701e+00 6.52992725e-01
3.91539723e-01 -5.57419479e-01 3.09986807e-02 1.16607606e+00
-1.20313749e-01 7.18139037e-02 -7.06648469e-01 8.57214451e-01
4.00167741e-02 -1.28962648e+00 3.54372054e-01 -1.87316298e-01
7.19346046e-01 -3.30366760e-01 -1.55694902e-01 7.19751894e-01
4.11488503e-01 -6.45709634e-01 4.29246932e-01 4.91977334e-01
7.29818404e-01 -1.10714054e+00 1.03687239e+00 3.60240817e-01
-1.20939517e+00 1.10022165e-01 -7.26240456e-01 4.79141206e-01
2.99893301e-02 1.08549750e+00 -1.10242295e+00 9.17563975e-01
7.06360638e-01 7.29872346e-01 -6.61405027e-01 8.75637829e-01
-2.80392587e-01 5.70101202e-01 -3.75859469e-01 -3.03368419e-01
-2.88249671e-01 -3.42108011e-01 4.01768804e-01 8.14708710e-01
2.45832413e-01 -2.88056936e-02 8.91275927e-02 7.08021045e-01
-4.99177456e-01 4.55771685e-01 -7.28886604e-01 -3.37144464e-01
9.27815855e-01 1.24063802e+00 -3.88952136e-01 -4.85199958e-01
-2.75143921e-01 6.45566285e-01 8.27307999e-01 3.80888283e-01
-7.09920347e-01 -6.53123736e-01 2.38851011e-01 1.46440133e-01
5.72799206e-01 4.47835773e-01 3.00030317e-02 -1.33105934e+00
4.33756918e-01 -3.73790443e-01 5.53255737e-01 -6.01046085e-01
-1.59578204e+00 3.67702581e-02 -2.07266897e-01 -1.20508826e+00
2.97407918e-02 -3.47217530e-01 -3.36552858e-01 8.02694380e-01
-1.57655835e+00 -1.19302821e+00 -2.35795796e-01 3.17318708e-01
-4.12441313e-01 -2.34510265e-02 6.59268498e-01 4.70041782e-01
-5.42027771e-01 1.05755556e+00 1.90416798e-01 3.71624500e-01
7.26665378e-01 -1.30972898e+00 -5.70207164e-02 5.48370183e-01
-3.29569101e-01 9.46457207e-01 5.20281672e-01 -9.79269624e-01
-1.42972350e+00 -1.41236806e+00 8.23170781e-01 -2.82677323e-01
4.67061520e-01 -1.47718444e-01 -8.96126390e-01 8.86120915e-01
-3.71402681e-01 5.91426730e-01 8.87599349e-01 2.50723302e-01
-3.95833492e-01 -5.08320034e-01 -1.64970398e+00 5.46045661e-01
1.03256154e+00 -2.15661570e-01 -3.86998892e-01 3.25840473e-01
8.36740911e-01 -1.88420102e-01 -9.51323628e-01 7.62278855e-01
4.27262455e-01 -7.95256495e-01 1.04605520e+00 -6.27514362e-01
1.00711718e-01 -6.41757011e-01 1.31403968e-01 -1.30987918e+00
-2.79774010e-01 6.49014562e-02 -6.07561707e-01 1.49009299e+00
3.75682324e-01 -9.65901375e-01 1.18704998e+00 4.19160873e-01
1.52909309e-01 -9.36973572e-01 -5.73544979e-01 -8.16071630e-01
-2.47954488e-01 1.13976322e-01 8.42204571e-01 1.29401994e+00
-5.64874820e-02 4.74945068e-01 -3.60439062e-01 2.52061397e-01
7.33335555e-01 1.94446612e-02 9.76271868e-01 -1.29499555e+00
-3.35191667e-01 2.02363506e-01 -8.88435245e-01 -6.82173371e-01
1.89277992e-01 -1.01684213e+00 -3.91959578e-01 -1.64646578e+00
4.82952416e-01 -6.51690662e-01 -3.47488075e-01 2.55901903e-01
-6.47492290e-01 1.80044636e-01 -1.85960874e-01 -1.69434279e-01
-6.97782993e-01 9.66406941e-01 1.20025480e+00 -1.19397730e-01
-7.84743503e-02 -2.48758495e-01 -8.55669677e-01 5.27495742e-01
6.27813756e-01 -3.27277273e-01 -7.83585370e-01 1.41028404e-01
7.42123008e-01 -1.48227230e-01 1.88790858e-01 -6.11897588e-01
-7.01928958e-02 -2.18039483e-01 5.00213206e-01 -9.60002542e-01
1.93855971e-01 -6.86047614e-01 2.24385604e-01 1.69763476e-01
-1.03056272e-02 -2.73876309e-01 -1.59451842e-01 1.25931764e+00
-1.22844100e-01 -4.03399885e-01 4.18328375e-01 -4.53608260e-02
-6.94305301e-01 4.14888203e-01 4.00663793e-01 4.59170014e-01
1.24438310e+00 1.72129627e-02 -6.48550808e-01 -3.61985005e-02
-4.26120132e-01 5.53433836e-01 4.55631912e-01 2.21989393e-01
5.90395808e-01 -1.79870307e+00 -3.74548882e-01 7.95973092e-02
5.23182034e-01 1.87484100e-01 1.47785336e-01 8.24210346e-01
-3.38624805e-01 1.47022069e-01 4.77906287e-01 -1.83545887e-01
-1.12855375e+00 6.09163105e-01 7.71794021e-02 -5.04439890e-01
-2.23990455e-01 8.54474127e-01 1.57457501e-01 -7.46551633e-01
3.81662436e-02 -2.95399874e-02 -4.16560203e-01 6.21340759e-02
4.02239591e-01 5.59946358e-01 -1.42189802e-03 -4.04942840e-01
-3.86451244e-01 3.31631936e-02 -2.85486609e-01 3.94409835e-01
1.22567284e+00 2.61391938e-01 -2.47731194e-01 5.22775233e-01
1.02947223e+00 3.50825250e-01 -4.89701122e-01 -2.00792328e-01
-1.29097492e-01 -7.63206899e-01 -3.36794227e-01 -4.60048616e-01
-1.22237909e+00 5.03554344e-01 2.01195374e-01 2.47524902e-01
7.50331104e-01 -1.24938168e-01 7.15305686e-01 5.04302919e-01
6.21560156e-01 -1.17896104e+00 -1.14850678e-01 7.72163942e-02
4.52157825e-01 -1.34437835e+00 1.57991216e-01 -1.13426721e+00
-5.44713974e-01 7.11187840e-01 1.02147496e+00 -1.71730995e-01
6.59807503e-01 -2.44204447e-01 -4.64849353e-01 -2.04931974e-01
-5.40987790e-01 2.09328625e-02 -2.19410146e-03 8.45560253e-01
2.57363617e-01 3.46247375e-01 -8.38013291e-01 9.99986470e-01
-9.80694443e-02 1.66376978e-01 5.33947825e-01 5.65073073e-01
-2.94613093e-01 -1.20618010e+00 -7.80561268e-02 1.12064934e+00
-3.25139761e-01 -1.31108776e-01 -3.80612135e-01 1.03113067e+00
-2.33274717e-02 4.88978058e-01 -4.85573977e-01 -4.40746814e-01
3.11590761e-01 1.01015724e-01 5.25691926e-01 -5.21100938e-01
-1.27581879e-01 -6.66468740e-01 3.57186377e-01 -1.84144527e-01
-2.74663866e-01 -4.45342302e-01 -1.37026083e+00 -3.87140572e-01
-1.02559257e+00 4.78129178e-01 1.21740680e-02 5.05774856e-01
4.98486310e-01 2.73056775e-01 4.39024210e-01 -2.54766136e-01
-8.68698835e-01 -1.04193592e+00 -1.11572289e+00 5.24817050e-01
-1.02036297e-01 -1.02027380e+00 -5.95645666e-01 -3.66403967e-01] | [8.761134147644043, 7.878274917602539] |
562088cf-4b3f-4de8-9699-95b99067af07 | learning-trajectory-aware-transformer-for | 2204.04216 | null | https://arxiv.org/abs/2204.04216v3 | https://arxiv.org/pdf/2204.04216v3.pdf | Learning Trajectory-Aware Transformer for Video Super-Resolution | Video super-resolution (VSR) aims to restore a sequence of high-resolution (HR) frames from their low-resolution (LR) counterparts. Although some progress has been made, there are grand challenges to effectively utilize temporal dependency in entire video sequences. Existing approaches usually align and aggregate video frames from limited adjacent frames (e.g., 5 or 7 frames), which prevents these approaches from satisfactory results. In this paper, we take one step further to enable effective spatio-temporal learning in videos. We propose a novel Trajectory-aware Transformer for Video Super-Resolution (TTVSR). In particular, we formulate video frames into several pre-aligned trajectories which consist of continuous visual tokens. For a query token, self-attention is only learned on relevant visual tokens along spatio-temporal trajectories. Compared with vanilla vision Transformers, such a design significantly reduces the computational cost and enables Transformers to model long-range features. We further propose a cross-scale feature tokenization module to overcome scale-changing problems that often occur in long-range videos. Experimental results demonstrate the superiority of the proposed TTVSR over state-of-the-art models, by extensive quantitative and qualitative evaluations in four widely-used video super-resolution benchmarks. Both code and pre-trained models can be downloaded at https://github.com/researchmm/TTVSR. | ['Xueming Qian', 'Jianlong Fu', 'Huan Yang', 'Chengxu Liu'] | 2022-04-08 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Liu_Learning_Trajectory-Aware_Transformer_for_Video_Super-Resolution_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Liu_Learning_Trajectory-Aware_Transformer_for_Video_Super-Resolution_CVPR_2022_paper.pdf | cvpr-2022-1 | ['video-super-resolution'] | ['computer-vision'] | [ 2.93211281e-01 -4.47916031e-01 -3.89901936e-01 -2.22595289e-01
-1.10358834e+00 -3.40880424e-01 5.01929998e-01 -4.84990209e-01
-2.15162918e-01 7.11929202e-01 5.46824455e-01 6.37286678e-02
6.27000292e-04 -5.77370048e-01 -9.06083941e-01 -5.32006621e-01
3.33215483e-02 -2.91842192e-01 6.01616919e-01 -2.44609207e-01
1.25275448e-01 4.01128769e-01 -1.60315478e+00 6.16853595e-01
7.95109808e-01 7.12328017e-01 6.91260397e-01 5.98925889e-01
-9.89846699e-03 1.02852392e+00 -1.96382433e-01 -5.19066118e-02
2.97830164e-01 -3.83131593e-01 -8.29949915e-01 2.27140248e-01
6.93367898e-01 -7.69290447e-01 -7.70191610e-01 1.00918984e+00
3.31837207e-01 4.05361533e-01 1.27328560e-01 -7.99503207e-01
-9.43643987e-01 5.67889512e-01 -8.88045371e-01 6.82901919e-01
6.39001608e-01 1.60709009e-01 9.07252014e-01 -1.15704429e+00
7.74439096e-01 1.34557736e+00 3.87557447e-01 6.53221250e-01
-1.05258381e+00 -7.19664991e-01 4.28545952e-01 5.31079888e-01
-1.49360645e+00 -5.81825197e-01 6.17370963e-01 -2.80259639e-01
8.36662471e-01 1.57379687e-01 5.08407056e-01 1.22099209e+00
8.54835734e-02 7.18734443e-01 9.33693826e-01 2.14409083e-02
-7.77051672e-02 -3.38016272e-01 -1.88254938e-01 4.83308613e-01
-3.59392986e-02 6.15058243e-02 -7.94878244e-01 2.34799325e-01
1.40868390e+00 3.30492616e-01 -5.74676991e-01 -2.46723577e-01
-1.37399876e+00 4.69346166e-01 4.54712778e-01 4.39262241e-01
-5.04303873e-01 5.14784874e-03 3.93341660e-01 6.72613606e-02
4.72140580e-01 1.77758001e-02 -2.79388487e-01 -1.66212276e-01
-1.07501280e+00 1.28326982e-01 -1.41447723e-01 1.10963476e+00
6.96289897e-01 3.27235401e-01 -3.88952285e-01 8.82595062e-01
-1.31718904e-01 2.91886568e-01 4.37680066e-01 -1.27254725e+00
5.37809789e-01 1.76905632e-01 3.97551119e-01 -8.52893591e-01
3.25295180e-02 -2.14461803e-01 -1.01571500e+00 -8.54061171e-02
2.31740370e-01 2.24480942e-01 -9.72656906e-01 1.44055307e+00
3.10450405e-01 7.05781639e-01 9.63667184e-02 1.31823301e+00
9.33216274e-01 9.22926903e-01 -5.43862097e-02 -5.27830899e-01
1.37068880e+00 -1.06159711e+00 -8.29029799e-01 1.42404810e-02
-4.30798009e-02 -7.83542454e-01 1.11935186e+00 1.71491995e-01
-1.36504245e+00 -8.33307445e-01 -8.12275112e-01 -3.91503513e-01
1.99935529e-02 -3.43035907e-02 2.45377779e-01 -1.30953565e-01
-1.10588169e+00 6.30652726e-01 -8.87355566e-01 -1.31056070e-01
5.68751037e-01 3.70583534e-02 -4.23337549e-01 -5.30941725e-01
-1.25765944e+00 4.74422038e-01 2.25305662e-01 1.68613702e-01
-1.01815689e+00 -8.87929559e-01 -8.81765842e-01 -8.10516626e-02
5.89455545e-01 -5.36325991e-01 1.13330841e+00 -8.65741968e-01
-1.31074035e+00 4.94945884e-01 -6.39300764e-01 -3.90911967e-01
4.97236848e-01 -4.24692035e-01 -4.91367638e-01 4.77316409e-01
1.73504010e-01 5.02580941e-01 1.09164512e+00 -1.27961802e+00
-1.02806771e+00 -6.42771050e-02 2.94906706e-01 3.67049932e-01
-2.89667726e-01 4.10142004e-01 -7.98551857e-01 -8.91929984e-01
-3.23862918e-02 -6.35958791e-01 -1.75542444e-01 -2.48044744e-01
-1.16404191e-01 -9.16651115e-02 8.46326351e-01 -9.68043089e-01
1.24845576e+00 -2.03279305e+00 3.40398431e-01 -4.57130641e-01
2.62675494e-01 4.45720822e-01 -2.80885071e-01 9.64154080e-02
-1.24209762e-01 5.54775298e-02 -1.70866959e-02 -2.50852823e-01
-5.16461313e-01 7.90304542e-02 -5.45054138e-01 3.17398518e-01
2.64004171e-01 8.34916353e-01 -1.06605220e+00 -5.73931038e-01
4.76591676e-01 1.03877199e+00 -4.42979634e-01 2.16869131e-01
-1.93283539e-02 7.52488494e-01 -5.73311031e-01 6.58048570e-01
6.63607776e-01 -3.33168298e-01 1.15588699e-02 -5.73480904e-01
-3.96680236e-01 1.58882797e-01 -9.34015274e-01 1.80348265e+00
-4.66667712e-01 6.28912389e-01 -1.39734954e-01 -7.25731850e-01
6.55849099e-01 2.86238164e-01 6.57424510e-01 -9.34430301e-01
-1.88082665e-01 -4.50723544e-02 -4.89765465e-01 -4.76277769e-01
8.61838520e-01 1.36225730e-01 3.01790833e-01 -8.53381157e-02
-1.20652676e-01 5.31420529e-01 2.54774153e-01 2.35479102e-01
9.73488569e-01 5.20220697e-01 3.44431400e-01 1.91528231e-01
6.62196457e-01 -2.32353136e-01 1.01357400e+00 4.28802758e-01
-2.53786176e-01 1.00925958e+00 3.68586108e-02 -3.71134281e-01
-1.24146938e+00 -1.22227848e+00 6.01800606e-02 1.15420830e+00
4.30490851e-01 -4.91584390e-01 -5.75452864e-01 -4.41098362e-01
-3.62033367e-01 4.02007192e-01 -5.41282177e-01 2.20688641e-01
-1.01092112e+00 -3.57857823e-01 6.89831674e-02 7.02261746e-01
5.56026101e-01 -9.30927277e-01 -5.27303576e-01 1.87549993e-01
-6.97024286e-01 -1.65233767e+00 -8.68996263e-01 -5.61074018e-01
-8.45503867e-01 -7.93617666e-01 -9.81024563e-01 -7.21046746e-01
4.37451571e-01 1.09110916e+00 9.93465364e-01 -1.19016565e-01
-1.99600503e-01 2.42615879e-01 -7.43633151e-01 4.13546890e-01
-8.44022632e-02 -1.36850327e-01 1.38857597e-02 1.16356775e-01
1.76766515e-01 -6.16982639e-01 -8.48222315e-01 4.59824055e-01
-9.32377398e-01 3.56271595e-01 6.62833273e-01 8.11064243e-01
1.07243264e+00 1.12747811e-01 5.67507803e-01 -4.99208599e-01
4.58826236e-02 -3.72463167e-01 -5.08592844e-01 2.54794449e-01
-2.31101230e-01 -1.83794469e-01 8.32587779e-01 -5.06982565e-01
-1.30637932e+00 4.41302225e-04 1.41026497e-01 -1.08425212e+00
-9.50369164e-02 1.30893692e-01 -1.24476776e-01 9.44338888e-02
2.26326048e-01 5.72594523e-01 -3.15395296e-01 -5.38906872e-01
4.01931912e-01 4.99507487e-01 7.32649148e-01 -4.38563526e-01
9.95696127e-01 6.42851412e-01 -2.47473583e-01 -8.76023650e-01
-9.62413371e-01 -5.56785583e-01 -6.29719555e-01 -2.94511944e-01
1.01815498e+00 -1.39014018e+00 -2.99038798e-01 2.24906489e-01
-8.74425650e-01 -3.72704268e-01 -1.48707643e-01 5.46930850e-01
-6.11624181e-01 5.04499435e-01 -7.62925923e-01 -4.85891044e-01
-2.36710623e-01 -1.09836280e+00 1.16485357e+00 4.24136966e-01
2.60305345e-01 -6.78481519e-01 -2.67985225e-01 6.03681564e-01
4.48201627e-01 1.82360411e-01 2.80449122e-01 5.85608296e-02
-1.10087335e+00 3.19328099e-01 -5.58338761e-01 3.45683217e-01
3.53099912e-01 5.66618191e-03 -6.43715203e-01 -5.35779595e-01
-2.16027260e-01 -1.40367657e-01 1.01410174e+00 5.34802079e-01
1.29726589e+00 -2.64140308e-01 -1.48349106e-01 8.09376717e-01
1.39615643e+00 6.54594377e-02 8.43592227e-01 3.73560548e-01
1.18292236e+00 3.99508923e-01 1.25038338e+00 4.06527042e-01
6.22579634e-01 9.69319224e-01 3.08412284e-01 -5.24507090e-02
-4.82942849e-01 -3.79418939e-01 6.90847218e-01 7.08571255e-01
-7.48679876e-01 1.38501346e-01 -4.86830741e-01 6.54597461e-01
-1.95504093e+00 -1.39486134e+00 -6.69015497e-02 2.19774127e+00
8.76787186e-01 -1.66809082e-01 1.54205635e-01 -1.71056733e-01
9.27488804e-01 5.57198763e-01 -6.34831071e-01 1.95354834e-01
-2.03459561e-01 -2.60735024e-02 4.65283453e-01 4.95850861e-01
-1.14415991e+00 1.16951072e+00 4.97448492e+00 1.09112954e+00
-1.09243691e+00 3.09712499e-01 4.92891669e-01 -5.25032103e-01
-2.37172469e-01 -1.21251367e-01 -8.20998192e-01 4.40311879e-01
7.22879529e-01 -2.31228545e-01 6.99639857e-01 5.40699303e-01
6.87807620e-01 2.12029651e-01 -9.26252663e-01 1.20401561e+00
-9.13538504e-03 -1.52881622e+00 2.21328661e-01 -6.59541562e-02
9.19500053e-01 2.44254395e-02 1.91602200e-01 2.45073572e-01
8.86448845e-02 -9.00280058e-01 6.95892513e-01 5.75174510e-01
1.14328408e+00 -8.76225471e-01 3.85399133e-01 -4.71445099e-02
-1.85452676e+00 -1.71528026e-01 -6.00492239e-01 2.98039675e-01
3.72090936e-01 3.51232767e-01 -3.62086684e-01 9.47318017e-01
1.19909298e+00 1.27429354e+00 -5.31733990e-01 6.89044595e-01
-2.23643500e-02 3.98551911e-01 1.40661597e-01 7.12434471e-01
1.06333710e-01 -2.02505037e-01 6.80995405e-01 1.21041763e+00
4.55586433e-01 4.91126537e-01 1.97647721e-01 6.98349595e-01
-1.29781980e-02 -6.82307407e-02 -3.31745595e-01 1.53241336e-01
6.14434838e-01 1.33893323e+00 -5.17996192e-01 -4.08138782e-01
-8.22017133e-01 1.10302925e+00 2.37602666e-01 6.12213254e-01
-1.11457050e+00 -1.50578748e-02 8.29029143e-01 3.76634210e-01
7.31328905e-01 -2.02120945e-01 2.03673512e-01 -1.56556845e+00
2.18836173e-01 -9.58591878e-01 3.40627462e-01 -8.86139333e-01
-1.12243390e+00 6.98158145e-01 4.62431274e-02 -1.68125737e+00
-2.46151268e-01 1.57631245e-02 -1.98109806e-01 6.19546533e-01
-1.90172112e+00 -1.25619507e+00 -5.51515162e-01 9.28157210e-01
1.16857588e+00 9.06570163e-03 2.26569727e-01 5.34492612e-01
-6.83141470e-01 5.24264395e-01 9.25884917e-02 1.76303193e-01
8.33340287e-01 -9.26643312e-01 4.57242429e-01 1.25636470e+00
-7.19681457e-02 5.76396286e-01 7.18072653e-01 -6.44397616e-01
-1.47754836e+00 -1.51411855e+00 4.68404889e-01 -2.38834575e-01
5.49548209e-01 -9.00839940e-02 -1.28769207e+00 8.07638764e-01
1.00426584e-01 4.24802244e-01 1.67178735e-01 -2.89046973e-01
-5.01311064e-01 -1.81445256e-01 -8.84608567e-01 6.49199307e-01
1.46414483e+00 -5.90130329e-01 -4.90430474e-01 9.16918833e-03
9.35792327e-01 -4.56566453e-01 -1.14152527e+00 4.89064038e-01
5.29565513e-01 -1.16538107e+00 1.41238582e+00 -2.87038177e-01
7.41942167e-01 -6.57831371e-01 -3.15755010e-01 -1.03964162e+00
-4.97426897e-01 -6.70854449e-01 -3.43739450e-01 1.37692654e+00
-2.39200458e-01 -3.90896469e-01 4.54805404e-01 3.72815877e-01
-1.61528736e-01 -7.59620190e-01 -7.36945271e-01 -7.49837697e-01
-2.10895240e-01 -1.29765570e-01 5.04995465e-01 1.03864634e+00
-3.51448715e-01 1.87518284e-01 -7.69033432e-01 4.11517322e-01
9.27365363e-01 2.76693851e-01 5.49956024e-01 -7.55748868e-01
-1.58084929e-01 -2.47098103e-01 -2.73532838e-01 -1.26530468e+00
7.43511468e-02 -5.04257441e-01 -2.24291477e-02 -1.65415561e+00
4.32989448e-01 -1.76655486e-01 -4.54434454e-01 3.65015566e-01
-3.44998986e-01 4.22870874e-01 3.34443212e-01 4.03622419e-01
-9.28452730e-01 6.19284451e-01 1.52485240e+00 2.16192648e-01
-2.68581539e-01 -3.46983403e-01 -5.55208206e-01 6.68021739e-01
7.47929037e-01 -9.84608531e-02 -4.64301854e-01 -4.70822603e-01
-2.18753666e-01 5.20215273e-01 4.84668046e-01 -7.51442015e-01
5.75912818e-02 -3.55853021e-01 4.61132467e-01 -7.18115628e-01
4.34029967e-01 -5.52445710e-01 2.04550385e-01 -5.93734346e-02
-3.74057919e-01 2.61759646e-02 -3.05461977e-02 7.35821784e-01
-3.90271246e-01 3.43435526e-01 9.37196970e-01 -9.46460068e-02
-1.14938760e+00 6.85975313e-01 -1.35810554e-01 1.34217711e-02
9.34818685e-01 -2.95604557e-01 -4.64981139e-01 -3.55594069e-01
-5.98266482e-01 2.28631034e-01 7.69364715e-01 8.24943841e-01
1.15060019e+00 -1.37556553e+00 -9.26124275e-01 -1.33475631e-01
-3.35642844e-02 1.98025450e-01 9.92755532e-01 8.52718949e-01
-1.51055977e-01 3.62789184e-01 -4.02407944e-01 -5.93936324e-01
-1.56257379e+00 8.07474434e-01 1.98363796e-01 -5.64068407e-02
-1.22840679e+00 5.18419683e-01 6.02328062e-01 2.82939076e-01
2.28162855e-02 -3.63941222e-01 -4.85691816e-01 -1.38041422e-01
1.04406762e+00 4.28547263e-01 -3.25557441e-01 -1.06398535e+00
-1.99924737e-01 9.00982797e-01 -3.93332660e-01 1.14805788e-01
1.40392661e+00 -6.69217825e-01 1.30556121e-01 3.53450537e-01
1.12157786e+00 -3.62587385e-02 -1.87076128e+00 -5.78651428e-01
-4.32541430e-01 -1.03120041e+00 1.02278419e-01 -2.69990444e-01
-1.26228619e+00 6.24911904e-01 5.26306570e-01 -6.25582486e-02
1.40065444e+00 1.77155472e-02 1.08676386e+00 -9.68418941e-02
5.63167393e-01 -8.86900067e-01 2.47747004e-01 3.54247034e-01
9.51309085e-01 -1.27632320e+00 3.04248687e-02 -5.35385311e-01
-8.18459034e-01 9.82199073e-01 6.65009975e-01 -2.60291308e-01
9.89461765e-02 5.26968837e-02 -2.00794399e-01 2.40284592e-01
-9.32770669e-01 -3.82785559e-01 2.78680682e-01 6.60480678e-01
4.19585705e-01 -8.87645632e-02 -5.60436472e-02 3.83484215e-01
1.20669089e-01 2.28607967e-01 7.74848044e-01 6.44370317e-01
-4.69214469e-01 -7.89120615e-01 -3.33173931e-01 1.96357235e-01
-6.15700781e-01 -1.02497116e-01 2.93646365e-01 6.25591874e-01
-7.89139345e-02 9.24563348e-01 1.05763804e-02 -4.81022388e-01
3.02372277e-01 -4.78993952e-01 5.06880283e-01 -3.31531554e-01
-9.80911478e-02 3.49711508e-01 -4.66797352e-02 -1.10438943e+00
-7.84689963e-01 -8.60443652e-01 -1.33876896e+00 -3.30600768e-01
7.50949234e-02 -2.08664596e-01 1.43929869e-01 6.41947865e-01
4.75816935e-01 8.64379585e-01 7.01548874e-01 -1.25219429e+00
-2.22555295e-01 -7.09797144e-01 -4.18692052e-01 4.25867826e-01
6.13594830e-01 -6.63548529e-01 -2.01022685e-01 4.48768646e-01] | [11.034467697143555, -1.8888919353485107] |
60098c21-8244-4b67-9a4c-621a1da2c072 | exploring-vanilla-u-net-for-lesion | 2210.07490 | null | https://arxiv.org/abs/2210.07490v1 | https://arxiv.org/pdf/2210.07490v1.pdf | Exploring Vanilla U-Net for Lesion Segmentation from Whole-body FDG-PET/CT Scans | Tumor lesion segmentation is one of the most important tasks in medical image analysis. In clinical practice, Fluorodeoxyglucose Positron-Emission Tomography~(FDG-PET) is a widely used technique to identify and quantify metabolically active tumors. However, since FDG-PET scans only provide metabolic information, healthy tissue or benign disease with irregular glucose consumption may be mistaken for cancer. To handle this challenge, PET is commonly combined with Computed Tomography~(CT), with the CT used to obtain the anatomic structure of the patient. The combination of PET-based metabolic and CT-based anatomic information can contribute to better tumor segmentation results. %Computed tomography~(CT) is a popular modality to illustrate the anatomic structure of the patient. The combination of PET and CT is promising to handle this challenge by utilizing metabolic and anatomic information. In this paper, we explore the potential of U-Net for lesion segmentation in whole-body FDG-PET/CT scans from three aspects, including network architecture, data preprocessing, and data augmentation. The experimental results demonstrate that the vanilla U-Net with proper input shape can achieve satisfactory performance. Specifically, our method achieves first place in both preliminary and final leaderboards of the autoPET 2022 challenge. Our code is available at https://github.com/Yejin0111/autoPET2022_Blackbean. | ['Junjun He', 'Jingqi Niu', 'Meng Wei', 'Yuncheng Yang', 'Qian Wu', 'Can Tu', 'Yanzhou Su', 'Zhongying Deng', 'Ziyan Huang', 'Haoyu Wang', 'Jin Ye'] | 2022-10-14 | null | null | null | null | ['tumor-segmentation'] | ['computer-vision'] | [ 8.97276253e-02 -2.50119120e-01 -9.17102277e-01 -3.13748956e-01
-5.90608656e-01 -3.53731275e-01 7.89981261e-02 1.07947655e-01
-5.63380420e-01 8.41449440e-01 -7.15104640e-02 -7.57379353e-01
2.80554503e-01 -9.10584390e-01 -1.80722430e-01 -7.87666917e-01
2.39083227e-02 6.90216839e-01 1.44541904e-01 2.19936028e-01
-4.77778971e-01 7.18683362e-01 -6.25748515e-01 1.44062594e-01
8.22754323e-01 1.00907004e+00 4.94059712e-01 5.14304161e-01
-2.93784887e-01 4.35505539e-01 -1.52721405e-01 1.84843078e-01
2.63259113e-01 -6.55856133e-01 -8.80024254e-01 3.65268253e-02
-3.70017141e-01 -3.88307810e-01 -3.38077009e-01 1.24378622e+00
4.63007659e-01 -8.37727413e-02 4.19825256e-01 -1.34250188e+00
8.73993337e-02 6.94563091e-01 -8.00328553e-01 7.01185286e-01
-9.17277411e-02 3.99485409e-01 5.02872348e-01 -7.15523124e-01
5.66409230e-01 6.80110812e-01 4.94320422e-01 6.22880042e-01
-8.41478705e-01 -7.92715073e-01 -6.66894997e-03 2.01096267e-01
-1.39927626e+00 -1.16646234e-02 4.59100395e-01 -1.43569201e-01
5.03663003e-01 4.91541296e-01 1.09518909e+00 1.03539693e+00
3.05386126e-01 9.78400826e-01 1.15642726e+00 -7.54784932e-03
8.01596865e-02 -1.04482867e-01 1.26366302e-01 9.40312684e-01
1.46019369e-01 2.79126409e-02 1.46574110e-01 3.46760228e-02
1.13257802e+00 4.27638441e-01 -7.35761642e-01 -9.42756757e-02
-1.40533292e+00 6.48697674e-01 8.92400801e-01 5.37419379e-01
-3.51920426e-01 4.85308319e-01 5.64265966e-01 -1.94452256e-01
4.49070424e-01 1.11829549e-01 -1.65197104e-01 -2.18804702e-01
-8.22686493e-01 -2.54141510e-01 3.76098186e-01 5.87510169e-01
1.56055793e-01 -7.66803324e-02 -1.10744424e-01 5.89166522e-01
3.66043717e-01 1.64032012e-01 1.06419694e+00 -5.87079644e-01
1.75168455e-01 5.46053886e-01 -8.37535039e-02 -3.43669713e-01
-8.20159554e-01 -6.40029550e-01 -1.19286454e+00 -6.01209477e-02
7.28968859e-01 -3.23802441e-01 -1.39068592e+00 1.59121370e+00
3.86809081e-01 3.47490996e-01 -3.14167023e-01 1.28457582e+00
8.86934221e-01 3.36612612e-01 3.09313983e-01 -4.87422198e-01
1.47768474e+00 -1.05466568e+00 -6.66461527e-01 -3.88761833e-02
8.27328026e-01 -4.38319474e-01 8.08262825e-01 2.63251867e-02
-1.00500679e+00 -1.10100932e-01 -8.88399124e-01 2.13156804e-01
-2.75835812e-01 -2.73585878e-02 9.26405668e-01 6.98711574e-01
-8.45023930e-01 4.88139957e-01 -1.42148554e+00 -3.98119718e-01
6.87439203e-01 4.90800500e-01 -2.48214036e-01 -3.33822578e-01
-1.27736831e+00 8.11924696e-01 6.51823223e-01 -3.23910899e-02
-9.78739798e-01 -7.99850464e-01 -5.43785214e-01 -1.53394297e-01
6.20217741e-01 -9.26271737e-01 1.43634057e+00 -6.88629568e-01
-1.29962289e+00 5.20153463e-01 -5.96729815e-02 -3.30972612e-01
9.36942995e-01 6.30751908e-01 -1.18284814e-01 4.61028069e-01
-7.17386156e-02 9.06837285e-01 1.91737354e-01 -6.83527231e-01
-6.62135601e-01 -4.24718052e-01 -3.07291925e-01 3.25049251e-01
1.23440064e-01 -1.40179068e-01 -6.88695669e-01 -4.99903947e-01
4.40330058e-01 -7.79957592e-01 -5.40063202e-01 9.97491851e-02
-5.65553606e-01 -4.43737954e-03 7.92127073e-01 -5.61609387e-01
9.29798126e-01 -1.87226045e+00 -2.15915307e-01 2.66758651e-01
4.01269019e-01 1.56318411e-01 2.69673347e-01 -2.95058697e-01
-3.60217184e-01 5.37532926e-01 -2.03807741e-01 1.01604365e-01
-3.53046924e-01 1.53296113e-01 5.24810433e-01 6.23128653e-01
-1.99421257e-01 1.35125566e+00 -1.08335435e+00 -7.26607919e-01
1.78023338e-01 2.38756105e-01 3.08883302e-02 -2.55069494e-01
-1.53709441e-01 8.56178224e-01 -7.70976901e-01 1.04834092e+00
3.73484015e-01 -4.28476870e-01 2.82664061e-01 -2.00247064e-01
8.62872303e-02 -1.40536517e-01 -6.03881776e-01 1.70781565e+00
-4.49562892e-02 3.19291621e-01 1.62304893e-01 -7.60224283e-01
3.58638376e-01 5.42847097e-01 1.08779359e+00 -7.31566727e-01
7.04677761e-01 2.51619428e-01 4.98631477e-01 -3.60000640e-01
-2.47226357e-02 -4.62423146e-01 2.27878392e-01 4.12402540e-01
-5.26881754e-01 -1.05347700e-01 3.11914057e-01 -1.01702094e-01
1.22229660e+00 -1.77282348e-01 3.92430335e-01 -1.11531261e-02
1.46622822e-01 2.57216245e-01 6.91768646e-01 2.83651054e-01
-6.51446342e-01 4.71624792e-01 5.68128824e-01 -2.78821021e-01
-8.60797882e-01 -9.90600586e-01 -2.89191455e-01 4.81480658e-01
1.64621830e-01 9.26003754e-02 -6.25475287e-01 -9.47939754e-01
-6.04085103e-02 5.29099464e-01 -7.95201063e-01 4.44669798e-02
-4.64346081e-01 -1.41874647e+00 6.27528608e-01 8.30532074e-01
8.90340984e-01 -6.87982917e-01 -6.41450822e-01 4.70250845e-01
-4.19515580e-01 -1.00635576e+00 -3.56034577e-01 5.26286840e-01
-1.23099101e+00 -1.45206201e+00 -9.59206760e-01 -4.10145223e-01
9.84617174e-01 3.00202966e-01 7.52310276e-01 3.00477177e-01
-5.56994021e-01 9.39731747e-02 -2.03102961e-01 -4.39408898e-01
-3.67017031e-01 7.62348697e-02 -2.27438807e-01 -3.16084951e-01
1.22970060e-01 -4.23554420e-01 -7.78565645e-01 5.42445302e-01
-9.53225434e-01 5.37313640e-01 7.55912840e-01 7.11594045e-01
1.18816376e+00 2.05767751e-01 2.32211515e-01 -9.34634626e-01
5.26323557e-01 -6.69293284e-01 -3.16823959e-01 8.29906985e-02
-2.83829629e-01 -4.90761399e-01 5.20940423e-01 -2.97334611e-01
-6.86524749e-01 3.02258402e-01 -2.78913200e-01 -4.80163693e-01
-1.35227472e-01 7.81897366e-01 3.13858327e-04 -1.66358463e-02
5.05865097e-01 1.35540769e-01 1.41274184e-01 -1.01817265e-01
-3.94119471e-02 2.23333836e-01 5.01069188e-01 -2.67746150e-01
3.63410264e-01 6.17776990e-01 3.29201967e-01 -5.86410999e-01
-4.62074101e-01 -5.33454716e-01 -4.65699106e-01 -2.81137258e-01
1.03375840e+00 -6.56441867e-01 -5.84221601e-01 3.94276768e-01
-6.99620247e-01 -5.00755668e-01 -2.97765553e-01 7.93868542e-01
-3.22261810e-01 1.31077230e-01 -7.90409565e-01 -1.58635184e-01
-3.63762826e-01 -1.50225508e+00 3.79046649e-01 4.64630514e-01
6.65482059e-02 -1.06116164e+00 -2.61154741e-01 1.29450127e-01
3.50258797e-01 6.84521914e-01 9.41388488e-01 -6.87163532e-01
-5.11047900e-01 -2.55695581e-01 -4.64453340e-01 1.08831204e-01
2.87131667e-01 -2.19171904e-02 -6.49908245e-01 -4.34106728e-03
-1.07210137e-01 -4.60627303e-02 7.29433239e-01 7.58802235e-01
1.61130726e+00 3.43986541e-01 -7.88955271e-01 8.23086500e-01
1.32515442e+00 6.77577972e-01 3.13903272e-01 1.31441444e-01
7.38148987e-01 1.52490899e-01 5.25390446e-01 2.92502105e-01
2.17328131e-01 3.29809666e-01 9.32615101e-01 -3.71764272e-01
-2.14899600e-01 6.06071018e-02 -1.42392114e-01 5.99907696e-01
-1.50953963e-01 -4.69679683e-01 -1.25181568e+00 4.78021741e-01
-1.50851107e+00 -6.59274638e-01 -3.68146569e-01 1.79491842e+00
7.17781544e-01 -9.85086779e-04 1.53200194e-01 7.86331296e-02
8.07776392e-01 -2.44201958e-01 -9.01632905e-01 5.92216924e-02
3.29830021e-01 2.17370749e-01 8.24422061e-01 8.49758908e-02
-8.64097238e-01 4.79642689e-01 5.98798370e+00 7.81370342e-01
-1.36237073e+00 4.69232798e-01 1.15076292e+00 -1.15922712e-01
-1.29346609e-01 -4.93419021e-01 -4.55102026e-01 6.08660579e-01
8.00527990e-01 -3.41062605e-01 1.03436351e-01 6.82207644e-01
4.81891125e-01 -5.41391373e-01 -1.17139661e+00 9.19872701e-01
-5.14351010e-01 -1.30838227e+00 -2.77557671e-01 2.64838845e-01
4.23651278e-01 3.49811196e-01 -3.81493047e-02 1.73406154e-01
3.01574826e-01 -1.05972087e+00 1.79726660e-01 2.26691410e-01
1.06207383e+00 -7.06899643e-01 9.97260869e-01 4.73852247e-01
-1.24078095e+00 1.52156115e-01 -1.17440782e-01 3.32398951e-01
2.14907572e-01 4.89341080e-01 -1.47377288e+00 6.42569005e-01
4.40790147e-01 5.08840322e-01 -2.96137750e-01 1.35093892e+00
-2.78098494e-01 5.32121658e-01 -5.73237240e-01 -1.32250071e-01
4.36056465e-01 -4.50627692e-02 2.05359116e-01 9.73504841e-01
4.27607298e-01 5.20189106e-01 5.31459808e-01 8.09134722e-01
-3.10318589e-01 1.32201493e-01 -1.99319273e-01 -1.65413216e-01
1.56661034e-01 1.42864645e+00 -1.49040842e+00 -4.87135202e-01
-2.05649152e-01 6.84930384e-01 -1.85575604e-01 1.54655218e-01
-1.22871757e+00 1.22225277e-01 3.03668499e-01 1.55653089e-01
-2.99421638e-01 2.61125490e-02 -2.61947572e-01 -9.70403850e-01
-5.06152928e-01 -3.62925738e-01 3.67837131e-01 -6.89811766e-01
-8.32075059e-01 5.15048981e-01 1.56350479e-01 -1.16652906e+00
3.12144216e-02 -4.31015819e-01 -8.39327216e-01 8.74167562e-01
-1.60054636e+00 -8.30198050e-01 -8.45926106e-01 6.04454100e-01
5.48106432e-01 3.99272054e-01 6.13765597e-01 3.92317355e-01
-9.68034029e-01 4.09899771e-01 -1.76462591e-01 2.19574064e-01
3.79775465e-01 -1.43321180e+00 1.17207862e-01 6.81757510e-01
-4.13877904e-01 1.55347183e-01 2.06423417e-01 -8.82631004e-01
-1.10437346e+00 -1.30750787e+00 9.28127617e-02 4.50206309e-04
5.64646542e-01 2.64672250e-01 -6.20873868e-01 6.29418015e-01
-2.36441381e-02 6.06401205e-01 7.31973290e-01 -7.67331123e-01
4.41674411e-01 1.58317864e-01 -1.44417119e+00 4.14182961e-01
6.80704355e-01 -1.93405598e-02 1.62758511e-02 5.70836365e-01
5.14740229e-01 -9.91241693e-01 -1.06626618e+00 3.99009645e-01
3.54833454e-01 -6.05344236e-01 7.98202276e-01 -3.71359915e-01
3.12525243e-01 -2.70217121e-01 1.88423574e-01 -1.52293086e+00
-2.25244507e-01 -1.64925796e-03 2.77901024e-01 5.31687260e-01
4.45497364e-01 -6.48153186e-01 1.21327972e+00 4.84284878e-01
-3.80419791e-01 -1.17137933e+00 -8.12370479e-01 -6.01614296e-01
-2.65634328e-04 -5.75616360e-01 4.21571374e-01 1.06545079e+00
-1.11611404e-01 -1.41050324e-01 1.11662947e-01 -7.81373903e-02
3.84763360e-01 -2.67353565e-01 2.81612873e-01 -1.01867223e+00
-1.34106994e-01 -6.41304851e-01 -2.27797076e-01 -7.59957433e-01
-1.82049289e-01 -1.32006788e+00 -1.96240246e-01 -1.71264660e+00
4.07708704e-01 -4.95402724e-01 -5.13588667e-01 6.02576733e-01
-6.99709579e-02 4.39795911e-01 8.09843317e-02 1.72656566e-01
-1.66969597e-01 1.15988918e-01 1.93994999e+00 -1.60434306e-01
-9.43022966e-02 1.83149084e-01 -4.00413603e-01 9.98268843e-01
1.09777546e+00 -4.13427383e-01 -4.78392452e-01 -8.01834017e-02
-3.14947546e-01 6.10796511e-01 3.34019125e-01 -8.11496735e-01
2.01522037e-01 -2.81193078e-01 8.86943877e-01 -7.14603662e-01
2.62864023e-01 -1.01453209e+00 3.39050353e-01 1.01419282e+00
1.32650554e-01 2.23435611e-01 2.59657502e-01 3.50886226e-01
-2.99608737e-01 -2.56612569e-01 1.01905406e+00 -8.15733373e-01
-6.23286068e-01 7.75834560e-01 -5.13285041e-01 -1.78691804e-01
1.43490350e+00 -2.90021092e-01 -4.41913247e-01 -1.61382571e-01
-1.02578866e+00 5.67042708e-01 2.34621450e-01 -1.53119400e-01
6.29597664e-01 -1.27339745e+00 -4.15076226e-01 -2.85081323e-02
-1.67720601e-01 2.66968310e-01 3.01340014e-01 1.61816907e+00
-1.01667750e+00 3.60028982e-01 -3.80444854e-01 -6.99080527e-01
-1.09241462e+00 3.45032394e-01 8.39803040e-01 -6.86113656e-01
-5.72845995e-01 1.00766349e+00 2.15817422e-01 -5.17454334e-02
7.22607896e-02 -8.01544309e-01 -3.38278450e-02 -2.44344268e-02
3.81026685e-01 3.06379557e-01 2.24460568e-02 -2.32237294e-01
-3.86685103e-01 -7.07512349e-03 -8.38047341e-02 -5.74740283e-02
1.07545125e+00 2.04221178e-02 -1.17903046e-01 1.00256540e-01
1.08322394e+00 -5.39961040e-01 -9.61183608e-01 -7.49388337e-02
-1.83287337e-01 -2.93186247e-01 4.67724591e-01 -1.23007357e+00
-1.62401426e+00 6.92547560e-01 7.35583603e-01 -5.27958684e-02
1.19354916e+00 -2.40385816e-01 8.49826574e-01 -3.61096151e-02
3.42919648e-01 -6.69717848e-01 -4.12238948e-02 1.87178347e-02
4.76289868e-01 -1.26811767e+00 4.24492992e-02 -7.98457980e-01
-5.08436978e-01 1.22636139e+00 8.71750534e-01 2.50155002e-01
5.24776876e-01 3.58338863e-01 6.12505004e-02 -2.29807541e-01
-5.52590847e-01 -1.00147411e-01 -1.49783611e-01 1.79364130e-01
4.05293405e-01 3.76658112e-01 -2.46901751e-01 6.32420897e-01
-1.02808110e-01 1.28170595e-01 6.15968525e-01 1.13015532e+00
-2.98465699e-01 -1.17825472e+00 -3.93584043e-01 7.39286602e-01
-7.51898706e-01 8.54384899e-03 -3.77851456e-01 1.28406978e+00
1.87328234e-01 4.65232402e-01 6.36441587e-03 -8.53326842e-02
1.86487570e-01 1.60216838e-01 5.15407860e-01 -6.40028477e-01
-5.81008196e-01 4.65764999e-01 -3.42983343e-02 -5.69391489e-01
-3.75787377e-01 -5.54296374e-01 -1.68988895e+00 -2.43798539e-01
-2.95435071e-01 6.21418320e-02 1.05468237e+00 7.49727786e-01
-2.48311713e-01 1.14165902e+00 4.66784090e-01 -5.08948922e-01
-1.33745626e-01 -8.82995129e-01 -7.19733059e-01 -1.77572533e-01
5.42170629e-02 -6.17854118e-01 -6.01029918e-02 -2.32590541e-01] | [14.679407119750977, -2.479750394821167] |
3a2d5628-acf7-47a9-8a45-2209eb6a7e22 | referee-towards-reference-free-cross-speaker | 2109.03439 | null | https://arxiv.org/abs/2109.03439v1 | https://arxiv.org/pdf/2109.03439v1.pdf | Referee: Towards reference-free cross-speaker style transfer with low-quality data for expressive speech synthesis | Cross-speaker style transfer (CSST) in text-to-speech (TTS) synthesis aims at transferring a speaking style to the synthesised speech in a target speaker's voice. Most previous CSST approaches rely on expensive high-quality data carrying desired speaking style during training and require a reference utterance to obtain speaking style descriptors as conditioning on the generation of a new sentence. This work presents Referee, a robust reference-free CSST approach for expressive TTS, which fully leverages low-quality data to learn speaking styles from text. Referee is built by cascading a text-to-style (T2S) model with a style-to-wave (S2W) model. Phonetic PosteriorGram (PPG), phoneme-level pitch and energy contours are adopted as fine-grained speaking style descriptors, which are predicted from text using the T2S model. A novel pretrain-refinement method is adopted to learn a robust T2S model by only using readily accessible low-quality data. The S2W model is trained with high-quality target data, which is adopted to effectively aggregate style descriptors and generate high-fidelity speech in the target speaker's voice. Experimental results are presented, showing that Referee outperforms a global-style-token (GST)-based baseline approach in CSST. | ['Dong Yu', 'Dan Su', 'Shan Yang', 'Songxiang Liu'] | 2021-09-08 | null | null | null | null | ['expressive-speech-synthesis'] | ['speech'] | [ 5.67763269e-01 2.91038379e-02 4.18840572e-02 -6.83455169e-01
-1.62209713e+00 -6.70817733e-01 6.27599716e-01 -4.10414606e-01
-1.65649131e-02 3.91745180e-01 4.59202498e-01 -1.85207874e-01
4.43557024e-01 -3.98594528e-01 -6.70746267e-01 -7.25528240e-01
6.14325643e-01 4.87413436e-01 -4.07905318e-02 -5.36310792e-01
-3.10448911e-02 3.11560184e-01 -1.31879687e+00 4.53868061e-01
9.12503064e-01 1.13309693e+00 4.58090723e-01 1.10239494e+00
-3.73643130e-01 2.98859149e-01 -9.44683611e-01 -2.18896255e-01
1.17938578e-01 -8.76587689e-01 -6.57279253e-01 -2.05246322e-02
5.18416524e-01 -1.73151530e-02 2.45186370e-02 7.44957328e-01
7.66128361e-01 3.23270828e-01 6.65356874e-01 -8.29348683e-01
-4.10612792e-01 5.65223932e-01 4.08517802e-03 -1.92095682e-01
4.58947271e-01 3.23856741e-01 9.88318324e-01 -1.09060192e+00
6.31087959e-01 1.69017470e+00 4.59634006e-01 1.03453612e+00
-1.42529893e+00 -7.66263187e-01 -4.10987675e-04 -2.78445423e-01
-1.17559230e+00 -9.98999000e-01 1.10611260e+00 -1.48138955e-01
9.01880860e-01 6.60977781e-01 6.64682925e-01 1.49740231e+00
1.59909911e-02 9.20603275e-01 1.27548695e+00 -6.26216471e-01
2.84687370e-01 2.46684238e-01 -4.23897028e-01 2.64432758e-01
-9.58619356e-01 3.64621192e-01 -9.74027216e-01 4.81329635e-02
5.72306514e-01 -7.72273302e-01 -1.52050033e-01 3.48536044e-01
-1.16508567e+00 6.17264926e-01 9.65819806e-02 3.50609571e-01
-2.22801253e-01 -2.34426767e-01 3.68818730e-01 4.92183954e-01
8.43506575e-01 3.46587837e-01 -6.05069041e-01 -5.39019525e-01
-1.39293861e+00 4.28196400e-01 8.23763132e-01 1.29298544e+00
5.68731844e-01 7.78278232e-01 -5.87797523e-01 1.18524718e+00
2.65250862e-01 1.08510125e+00 5.78799129e-01 -9.09791589e-01
6.93610847e-01 1.03009527e-03 -1.31889820e-01 -2.90901542e-01
2.02787057e-01 -3.49310249e-01 -6.75728858e-01 9.22241658e-02
1.68330163e-01 -2.09864974e-01 -9.08329964e-01 1.86068738e+00
3.30722809e-01 6.06922694e-02 2.22540811e-01 7.07651258e-01
8.03396940e-01 1.04365015e+00 -2.00007468e-01 -3.67310166e-01
1.05236816e+00 -9.96565163e-01 -8.32595527e-01 -1.71998516e-01
2.21492723e-01 -1.06135762e+00 1.59597194e+00 3.66112292e-01
-1.43881309e+00 -1.04271734e+00 -7.16528177e-01 -1.31224275e-01
-2.29839623e-01 1.64221615e-01 -1.82590187e-01 7.08610952e-01
-1.16827226e+00 6.32758439e-01 -5.83722949e-01 -1.15338765e-01
1.03145361e-01 2.43107662e-01 -5.56487292e-02 3.27321529e-01
-1.24603033e+00 6.50774717e-01 9.58716720e-02 -3.61187458e-02
-9.55361426e-01 -1.04088199e+00 -9.26632822e-01 -1.45780146e-01
3.27271707e-02 -5.78013778e-01 1.59170675e+00 -1.10860240e+00
-2.62711573e+00 6.27049208e-01 -7.20083356e-01 -8.49923119e-02
4.09339905e-01 -1.30992353e-01 -6.00719810e-01 1.72415003e-01
-1.02680333e-01 8.56843174e-01 1.34786153e+00 -1.23616147e+00
-5.52648604e-01 7.38021731e-02 -6.48106813e-01 3.76888812e-01
-1.59563765e-01 3.65187615e-01 -3.07889581e-01 -1.10115528e+00
-1.25314623e-01 -9.45532203e-01 2.72886127e-01 -3.97068590e-01
-5.75044334e-01 -4.93367136e-01 9.47466373e-01 -8.45997453e-01
1.19536710e+00 -2.06835532e+00 4.57902282e-01 -7.67747611e-02
-2.37621456e-01 3.55081558e-01 -4.15997267e-01 4.04129088e-01
8.16712603e-02 7.66044774e-04 -2.62698859e-01 -9.48934436e-01
2.62433678e-01 3.28806579e-01 -5.76647162e-01 -1.12292752e-01
5.75682640e-01 8.30701768e-01 -7.86139607e-01 -5.86359918e-01
2.69594789e-01 5.75018287e-01 -7.23742008e-01 7.93893516e-01
-3.34520966e-01 1.04001331e+00 -3.06993902e-01 3.47364664e-01
4.11351204e-01 6.22263432e-01 -1.69999748e-01 -2.29418367e-01
-3.24472159e-01 8.41004550e-01 -7.87594736e-01 1.95260990e+00
-1.00882161e+00 3.81925553e-01 2.15149716e-01 -4.93311644e-01
1.24776363e+00 6.73494339e-01 1.70209095e-01 -4.87559199e-01
1.54517610e-02 4.14555103e-01 -1.73218027e-01 -3.06447029e-01
3.01174581e-01 -6.10574841e-01 -3.08678269e-01 7.51153529e-02
6.12400472e-01 -1.34393871e+00 -1.36223540e-01 -2.07737923e-01
6.87274098e-01 4.43207234e-01 -1.51337609e-01 -3.74517560e-01
7.06132174e-01 -4.87600386e-01 7.79489219e-01 4.33756977e-01
5.13999909e-02 9.40959454e-01 2.26137079e-02 1.05466716e-01
-1.02484035e+00 -1.45104778e+00 9.32919830e-02 1.24534488e+00
-5.00273287e-01 -3.33791256e-01 -1.30082834e+00 -6.07570946e-01
-4.09707934e-01 1.13089848e+00 -4.31998968e-01 -5.29237241e-02
-8.26399982e-01 3.67344096e-02 7.45544314e-01 3.15211296e-01
3.02965045e-01 -1.37577903e+00 1.20800890e-01 3.65211695e-01
-3.38992208e-01 -1.07046211e+00 -1.33266902e+00 1.24586262e-01
-6.96612358e-01 -2.65621573e-01 -1.01857603e+00 -7.78932154e-01
2.19765648e-01 -3.99921030e-01 9.16048169e-01 -6.16678357e-01
1.41158342e-01 1.86524048e-01 -2.94182956e-01 -5.63015223e-01
-1.17397022e+00 1.86837509e-01 3.96491200e-01 4.08579260e-01
-7.18299001e-02 -6.18895173e-01 -2.39316568e-01 1.73482329e-01
-6.78363264e-01 2.16651946e-01 3.98492157e-01 9.10253465e-01
7.43061543e-01 -2.10436761e-01 8.66755843e-01 -5.57816684e-01
4.72936064e-01 2.93323454e-02 -5.19208968e-01 2.80803647e-02
-3.79281282e-01 1.05135135e-01 9.82445300e-01 -6.29086971e-01
-1.59003639e+00 -1.66220590e-01 -7.93101490e-01 -7.06582546e-01
-4.67473090e-01 9.63749141e-02 -7.49505818e-01 2.80725181e-01
4.22558218e-01 6.33317232e-01 -9.19835344e-02 -6.76569402e-01
4.87220556e-01 1.04386365e+00 8.23107183e-01 -8.01645398e-01
1.05236304e+00 -1.07226588e-01 -4.07555312e-01 -1.23581314e+00
-9.95566845e-01 -2.98526555e-01 -7.15540111e-01 -1.25140876e-01
8.68493140e-01 -7.49484241e-01 -3.68138611e-01 8.74820232e-01
-1.20085692e+00 -7.86083996e-01 -5.94133079e-01 4.95241582e-01
-1.00804758e+00 -5.37366457e-02 -6.07218444e-01 -9.25137699e-01
-5.70207059e-01 -1.17390358e+00 1.61019158e+00 1.15628541e-01
-5.14914215e-01 -1.18338740e+00 1.32606313e-01 4.39775616e-01
7.41913080e-01 -8.89289193e-03 7.63955951e-01 -5.65174997e-01
-8.20703954e-02 1.73767060e-01 3.51395249e-01 9.12311673e-01
5.68044901e-01 -1.87140629e-02 -1.42576182e+00 -3.19579780e-01
3.18099439e-01 -3.09038788e-01 3.20222169e-01 3.55313718e-01
8.89566123e-01 -5.80838144e-01 2.46555403e-01 7.87897766e-01
6.98325515e-01 3.03097904e-01 3.57721120e-01 -3.96106869e-01
8.86905849e-01 6.95614338e-01 5.40492237e-01 1.59859598e-01
2.78742075e-01 9.41372931e-01 -4.38110560e-01 -3.28618437e-01
-7.53899455e-01 -5.09901464e-01 9.11489785e-01 1.79848850e+00
1.29476339e-01 -1.88761830e-01 -4.42999601e-01 3.58289838e-01
-1.08986473e+00 -8.71430874e-01 1.81658283e-01 2.03122735e+00
1.53566563e+00 2.70989776e-01 1.51261479e-01 2.34773859e-01
6.15717590e-01 1.98135123e-01 -6.32140279e-01 -7.79158950e-01
-8.44766945e-02 6.56285048e-01 -1.73470289e-01 7.60237157e-01
-6.42183244e-01 1.29600167e+00 5.80620909e+00 1.34508371e+00
-1.55823612e+00 1.43235996e-01 4.88586873e-01 -1.91562966e-01
-6.06218517e-01 -3.23807508e-01 -1.05410683e+00 5.18403471e-01
1.43689919e+00 -2.69121438e-01 7.61058033e-01 5.05125761e-01
6.84457242e-01 4.99692261e-01 -1.20601165e+00 8.77792060e-01
-2.05971785e-02 -1.01308668e+00 7.05468506e-02 -2.49251261e-01
5.91277242e-01 3.72425769e-03 3.69857460e-01 5.42021394e-01
1.39550447e-01 -7.60502756e-01 1.29624784e+00 5.08523405e-01
1.46477294e+00 -6.93862498e-01 1.84986204e-01 1.77285239e-01
-1.45472074e+00 4.17996436e-01 1.23196788e-01 4.27502066e-01
2.56007403e-01 2.89155155e-01 -9.81133461e-01 7.38307416e-01
4.33483154e-01 6.37021303e-01 -1.90745741e-01 1.11437194e-01
-7.19419420e-02 1.27782071e+00 -3.16783935e-01 1.25233263e-01
1.32323638e-01 -1.75305530e-01 9.08897698e-01 1.54089653e+00
4.68451828e-01 -1.40049532e-01 1.72181368e-01 1.16828549e+00
-3.81425861e-03 2.85187095e-01 -2.82735050e-01 -1.99499771e-01
5.62929690e-01 1.13206267e+00 -1.32857680e-01 -5.15761256e-01
-2.14047104e-01 1.19135475e+00 -1.44848794e-01 4.11616534e-01
-5.13849676e-01 -4.17699248e-01 8.04298878e-01 -7.25802928e-02
3.94309908e-01 -4.68416959e-02 -7.67926946e-02 -9.72031057e-01
-8.30230564e-02 -1.10799778e+00 -1.88926280e-01 -8.17161143e-01
-1.36088574e+00 1.00592947e+00 -4.37224023e-02 -1.23044169e+00
-6.49134755e-01 -3.81393641e-01 -9.49656188e-01 1.51900470e+00
-1.47647834e+00 -1.50548267e+00 1.04688093e-01 6.06329024e-01
1.29025590e+00 -4.46899354e-01 1.04947960e+00 -4.64883111e-02
-3.70746374e-01 9.72205698e-01 2.85083316e-02 -8.40281919e-02
8.66329074e-01 -1.58575344e+00 8.89724851e-01 5.64034402e-01
3.79105508e-02 3.87828529e-01 8.33202362e-01 -4.58812892e-01
-1.36522985e+00 -1.41195405e+00 1.06121445e+00 -4.17661577e-01
4.77394462e-01 -8.11375737e-01 -9.59474862e-01 3.11946124e-01
3.54931593e-01 1.49202086e-02 6.07516766e-01 -1.07908905e-01
-1.40513495e-01 -3.60065073e-01 -1.02195585e+00 5.62264025e-01
8.85871887e-01 -1.01615798e+00 -7.41244435e-01 -1.37147516e-01
1.05886137e+00 -5.36967218e-01 -1.05984068e+00 1.83116466e-01
4.40381110e-01 -6.82875991e-01 6.76030397e-01 -2.96116620e-01
4.28809077e-02 -1.01744518e-01 -3.02935570e-01 -1.82478893e+00
1.08655885e-01 -1.41033053e+00 2.37039015e-01 1.89914680e+00
4.33546752e-01 -4.37780768e-01 3.16485584e-01 1.52591884e-01
-8.10321271e-01 -4.15774137e-01 -1.20548785e+00 -1.05001462e+00
5.19556165e-01 -5.68449259e-01 9.13671553e-01 6.46334410e-01
-2.03384101e-01 5.78899503e-01 -2.40150779e-01 3.78360674e-02
4.43271816e-01 -6.48370609e-02 6.82558656e-01 -9.05680716e-01
-4.46004927e-01 -4.74528134e-01 2.14040115e-01 -1.06640661e+00
4.12607759e-01 -9.51469719e-01 5.44921637e-01 -9.95338678e-01
-5.01233637e-01 -5.12170076e-01 -2.51640230e-01 1.39191493e-01
-3.70081455e-01 -7.34348455e-03 3.18279475e-01 -1.08164258e-01
-1.25185192e-01 1.09288359e+00 1.65834630e+00 -1.11954592e-01
-4.93366808e-01 3.86463434e-01 -2.16577321e-01 6.92816019e-01
7.63988197e-01 -3.81522030e-01 -4.46326196e-01 -8.42440128e-02
-5.15937328e-01 5.61308861e-01 1.23823583e-02 -9.85091686e-01
-1.55725017e-01 -2.69871503e-01 8.37472975e-02 -5.26826262e-01
7.84129083e-01 -3.45825344e-01 -1.56783864e-01 9.97843873e-03
-6.87198460e-01 -2.39570975e-01 3.05942774e-01 1.75761446e-01
-4.14751887e-01 -1.56217441e-02 9.99040961e-01 1.80805385e-01
-1.15815207e-01 2.89274126e-01 -4.08162951e-01 2.83636600e-01
2.68721759e-01 -1.72363259e-02 3.11814666e-01 -4.69663918e-01
-7.11323321e-01 -2.32551992e-01 1.89270377e-01 6.85507894e-01
6.14963591e-01 -1.68079364e+00 -1.01833367e+00 5.49691260e-01
6.52963519e-02 7.99819380e-02 2.35263899e-01 4.39906061e-01
2.64866501e-01 4.03865963e-01 2.09081396e-01 -7.04615116e-01
-1.22562838e+00 2.26272106e-01 3.72469485e-01 -7.23372996e-02
-5.38045168e-01 1.11664987e+00 4.24128443e-01 -7.93328285e-01
6.70201033e-02 -6.09246790e-01 2.47584805e-01 -1.35418236e-01
4.97124195e-01 1.03227720e-01 1.63602471e-01 -8.26635182e-01
-6.49340898e-02 5.81827939e-01 7.74802193e-02 -7.74687946e-01
1.10412800e+00 -3.62376928e-01 3.37908924e-01 1.00798690e+00
1.34263623e+00 4.72333550e-01 -1.60629308e+00 -2.70381123e-01
-2.12273344e-01 -1.79526761e-01 2.24810064e-01 -8.35329592e-01
-7.34324396e-01 1.09320021e+00 4.34659898e-01 -3.99566777e-02
1.22727919e+00 -5.41241579e-02 1.32660842e+00 -9.07935575e-02
1.07739769e-01 -1.29397762e+00 1.84637621e-01 7.18168974e-01
1.35526228e+00 -1.00957799e+00 -9.05407190e-01 -2.65619606e-01
-8.69263232e-01 1.16816986e+00 3.45553964e-01 2.28752613e-01
5.52919149e-01 3.80307078e-01 5.74289560e-01 3.76991749e-01
-8.56866002e-01 -2.41048336e-02 7.44089603e-01 6.62471771e-01
4.58196044e-01 8.22290182e-02 5.53069592e-01 6.60188556e-01
-9.17211056e-01 -2.44651347e-01 -7.83794746e-02 4.49148238e-01
-2.49692127e-01 -1.44352388e+00 -4.57377374e-01 1.25622794e-01
-2.33811751e-01 -4.05427516e-01 -3.22159201e-01 1.55146763e-01
2.02082805e-02 1.12350094e+00 4.62289490e-02 -4.37205821e-01
5.56336761e-01 5.21949291e-01 5.04266381e-01 -7.83260167e-01
-7.95089543e-01 7.50978589e-01 1.80311084e-01 -4.86775726e-01
-8.46040472e-02 -8.80977690e-01 -1.04599047e+00 -6.51982799e-02
-1.48990408e-01 2.62937248e-01 7.45106101e-01 1.14782035e+00
1.20338954e-01 7.88695157e-01 1.25828147e+00 -1.13128400e+00
-4.73519176e-01 -1.32171416e+00 -5.65795481e-01 3.83485317e-01
6.48868382e-01 -2.57646739e-01 -4.37963754e-01 4.90216464e-01] | [14.971290588378906, 6.550217628479004] |
4537eb4b-16fc-4271-a78d-94e975ae1b31 | comparative-layer-wise-analysis-of-self | 2211.03929 | null | https://arxiv.org/abs/2211.03929v3 | https://arxiv.org/pdf/2211.03929v3.pdf | Comparative layer-wise analysis of self-supervised speech models | Many self-supervised speech models, varying in their pre-training objective, input modality, and pre-training data, have been proposed in the last few years. Despite impressive successes on downstream tasks, we still have a limited understanding of the properties encoded by the models and the differences across models. In this work, we examine the intermediate representations for a variety of recent models. Specifically, we measure acoustic, phonetic, and word-level properties encoded in individual layers, using a lightweight analysis tool based on canonical correlation analysis (CCA). We find that these properties evolve across layers differently depending on the model, and the variations relate to the choice of pre-training objective. We further investigate the utility of our analyses for downstream tasks by comparing the property trends with performance on speech recognition and spoken language understanding tasks. We discover that CCA trends provide reliable guidance to choose layers of interest for downstream tasks and that single-layer performance often matches or improves upon using all layers, suggesting implications for more efficient use of pre-trained models. | ['Karen Livescu', 'Bowen Shi', 'Ankita Pasad'] | 2022-11-08 | null | null | null | null | ['spoken-language-understanding', 'spoken-language-understanding'] | ['natural-language-processing', 'speech'] | [ 3.07473183e-01 -1.23534583e-01 -3.63900214e-02 -6.78346217e-01
-8.47991109e-01 -7.87588537e-01 9.37227190e-01 2.09731251e-01
-6.11616135e-01 2.21401066e-01 8.32418203e-01 -5.78156948e-01
-3.64466339e-01 -1.60778210e-01 -4.42615688e-01 -5.72314978e-01
-2.33576685e-01 2.16873616e-01 4.72220741e-02 -1.17772602e-01
2.03059539e-01 2.05415070e-01 -1.55448067e+00 6.22065961e-01
6.05694652e-01 8.71517479e-01 2.92994857e-01 8.02806735e-01
-1.62637144e-01 4.13883150e-01 -3.88062626e-01 -2.07821459e-01
-1.04323467e-02 -6.48190022e-01 -7.86438406e-01 -1.47012129e-01
3.60814482e-01 1.83673874e-01 -1.84731767e-01 6.46119773e-01
5.50531209e-01 9.12174433e-02 8.18554103e-01 -7.94106841e-01
-5.54353476e-01 1.03902662e+00 1.89036950e-02 5.15426397e-01
1.03812283e-02 1.65148288e-01 1.31676316e+00 -7.90633738e-01
3.82364541e-01 1.24615550e+00 8.70634973e-01 6.63449645e-01
-1.57640421e+00 -5.43232560e-01 4.12659287e-01 1.92260966e-01
-1.22015405e+00 -1.24264753e+00 5.73674381e-01 -5.51636934e-01
1.10318696e+00 4.04679216e-02 3.64315063e-01 1.15691781e+00
-4.21207584e-02 6.56759501e-01 1.30366671e+00 -5.07945240e-01
1.99811637e-01 2.57575661e-01 1.75702155e-01 3.20550829e-01
-1.53048143e-01 1.83230162e-01 -1.07527697e+00 -1.24152869e-01
2.71304876e-01 -6.44411683e-01 -3.18625182e-01 -2.00583383e-01
-1.10482728e+00 5.72502196e-01 1.80297986e-01 6.33184314e-01
-5.19619361e-02 9.31111425e-02 5.16742826e-01 5.48967421e-01
5.97502470e-01 5.53966880e-01 -9.77051497e-01 -5.03478587e-01
-1.05907464e+00 -1.47446290e-01 7.84165800e-01 5.06416917e-01
6.93911135e-01 1.49416134e-01 5.31016737e-02 1.28301561e+00
5.17404616e-01 3.19074877e-02 8.31037045e-01 -8.68712664e-01
5.48066735e-01 9.75847542e-02 -6.03667855e-01 -5.46100974e-01
-3.75428587e-01 -5.90958178e-01 -3.67356092e-01 -6.04188889e-02
6.08727574e-01 -1.64216191e-01 -7.82625079e-01 2.11068392e+00
-2.66646594e-01 -7.25930929e-02 3.22315425e-01 4.37586546e-01
5.57769477e-01 4.59505945e-01 4.15456384e-01 -1.68547705e-01
1.05544317e+00 -7.14022696e-01 -4.39864963e-01 -5.42484641e-01
1.01729918e+00 -8.16021740e-01 1.17066157e+00 3.42777282e-01
-1.04952550e+00 -6.74938560e-01 -8.90045047e-01 2.53182594e-02
-3.99408907e-01 1.09513015e-01 5.00433683e-01 8.06743681e-01
-1.32253778e+00 9.48399365e-01 -6.75788701e-01 -4.29950655e-01
1.92438617e-01 3.94405693e-01 -3.84441733e-01 1.44131303e-01
-9.08412755e-01 9.17724848e-01 1.91085741e-01 -8.66895691e-02
-6.84881091e-01 -9.03026223e-01 -6.07487798e-01 1.27585247e-01
-2.63459682e-01 -4.34019387e-01 1.39615190e+00 -9.16856706e-01
-1.57208514e+00 8.55801046e-01 -3.76170307e-01 -2.35321850e-01
1.30968532e-02 -1.32572711e-01 -3.19439173e-01 -3.12602818e-01
-2.21765801e-01 6.82057977e-01 6.31663740e-01 -1.24471498e+00
-5.55726886e-01 -5.04055858e-01 -3.44353557e-01 2.90061593e-01
-4.28716511e-01 6.41299263e-02 -2.60966599e-01 -5.52803993e-01
4.03629333e-01 -8.97367179e-01 3.73183005e-02 -2.19163418e-01
-4.91162278e-02 -1.99805737e-01 3.37783962e-01 -6.48500025e-01
1.26326120e+00 -2.39581609e+00 3.18553239e-01 1.48693889e-01
-1.52832285e-01 6.78460225e-02 -4.08282876e-01 5.96120358e-01
-3.69996801e-02 4.19305921e-01 -4.02368098e-01 -7.77358949e-01
-2.67134588e-02 2.05228284e-01 -4.28277671e-01 3.87125939e-01
3.69048268e-01 6.38192415e-01 -6.92876041e-01 -2.16732860e-01
2.43637478e-03 4.86025006e-01 -6.41907156e-01 2.11685330e-01
-1.17445149e-01 4.36776817e-01 7.10735023e-02 2.66427755e-01
2.74454653e-01 7.44484067e-02 3.61855179e-01 -7.60218278e-02
-3.35198879e-01 1.14136851e+00 -6.52829170e-01 1.74289989e+00
-7.48442888e-01 1.07792890e+00 2.85655767e-01 -9.52992320e-01
8.28513324e-01 3.62399399e-01 1.53657109e-01 -6.01413429e-01
8.81971698e-03 3.30170602e-01 8.36090147e-01 -2.65864640e-01
1.64462760e-01 -3.10692012e-01 2.51414299e-01 6.16926849e-01
2.92743474e-01 -1.77762568e-01 -4.18096669e-02 -1.12255245e-01
1.11830306e+00 6.71818033e-02 3.32714580e-02 -4.09275144e-01
2.71097124e-01 -1.56878188e-01 2.99769580e-01 6.54166400e-01
-1.48947105e-01 7.18481243e-01 4.62591410e-01 1.26540840e-01
-8.97620916e-01 -1.05793345e+00 -6.06804907e-01 1.45058477e+00
-5.85289240e-01 -5.92155874e-01 -4.72934216e-01 -1.55466869e-01
-3.10939108e-03 6.94310784e-01 -5.05553842e-01 -3.64286989e-01
-4.74067122e-01 -6.90923572e-01 7.59074211e-01 7.65256047e-01
8.29634741e-02 -1.03635669e+00 -8.83560330e-02 1.53310478e-01
2.00498477e-01 -1.20416009e+00 -5.18324785e-02 5.82982540e-01
-1.16398001e+00 -5.34272909e-01 -3.77359569e-01 -7.12670267e-01
2.91108221e-01 1.73335508e-01 1.07103693e+00 -7.35320663e-03
2.09859058e-01 5.95938265e-01 -3.97198141e-01 -2.22473174e-01
-7.00749576e-01 4.91673946e-01 1.93923622e-01 -7.83425122e-02
2.98627555e-01 -7.12891877e-01 -2.92885035e-01 2.08943456e-01
-5.04952312e-01 -7.66795129e-02 6.82189643e-01 5.76124430e-01
2.15040252e-01 -3.09051961e-01 5.37546754e-01 -7.75973678e-01
8.65411282e-01 -5.32880187e-01 -5.10945246e-02 9.57701430e-02
-7.56339490e-01 4.45226490e-01 4.86446351e-01 -2.91460156e-01
-1.03950059e+00 -2.58562505e-01 -3.36510748e-01 -1.10322960e-01
-3.96204680e-01 6.87369108e-01 -4.50098701e-02 2.22322479e-01
6.32379234e-01 1.05932668e-01 1.07500218e-01 -8.42564821e-01
4.54430401e-01 8.14406216e-01 3.23501885e-01 -7.82046497e-01
5.12492776e-01 2.29637042e-01 -4.97982889e-01 -1.03069031e+00
-7.94346035e-01 -3.45313191e-01 -9.23345089e-01 3.68270953e-03
7.85174549e-01 -8.75448048e-01 -3.65565568e-01 4.31788534e-01
-9.94034231e-01 -7.74599254e-01 -5.06289005e-02 6.33723319e-01
-4.03427482e-01 1.97414160e-01 -5.32587528e-01 -9.55621421e-01
-1.44732110e-02 -1.12981749e+00 8.58041644e-01 -1.75753549e-01
-5.80042243e-01 -1.21054995e+00 1.91807285e-01 4.70328212e-01
6.41578138e-01 -6.03586435e-01 1.25995302e+00 -7.98215687e-01
-2.15636164e-01 2.92834610e-01 -8.64736214e-02 4.93752956e-01
3.12666088e-01 1.58320114e-01 -1.51581788e+00 -1.58235446e-01
-1.32573277e-01 -2.44601831e-01 1.22430944e+00 5.39755046e-01
1.09100449e+00 -3.70876491e-02 -2.24775195e-01 6.20295048e-01
8.40071142e-01 7.53404424e-02 3.23873192e-01 2.18134850e-01
5.09282768e-01 1.04614568e+00 6.68701250e-03 -9.41637456e-02
3.10529858e-01 5.57940781e-01 -1.73121914e-01 2.32449651e-01
-3.06942016e-01 -2.80323595e-01 6.48852706e-01 1.42193234e+00
1.27317131e-01 -6.31204844e-02 -1.09657359e+00 7.32096791e-01
-1.60226810e+00 -7.03638911e-01 2.12407097e-01 2.17894053e+00
1.07348537e+00 3.35660458e-01 1.08392514e-01 1.38006464e-01
2.63222486e-01 4.17022586e-01 -4.39962924e-01 -6.14375293e-01
-2.39451274e-01 3.08096379e-01 2.07750186e-01 5.78853369e-01
-7.42283344e-01 9.33128238e-01 7.63011456e+00 6.15419328e-01
-1.24596024e+00 5.10228202e-02 6.99455202e-01 -9.48293954e-02
-5.01182020e-01 1.76881626e-01 -6.99241698e-01 2.27873832e-01
1.44396055e+00 1.65132503e-03 5.55912614e-01 5.24832189e-01
3.15833092e-01 7.59657919e-02 -1.42206454e+00 5.13446569e-01
-5.69595583e-02 -1.11297095e+00 -1.71991929e-01 5.64395897e-02
4.93829638e-01 4.92880315e-01 3.57356995e-01 3.34629148e-01
1.77156806e-01 -1.04248345e+00 7.18035221e-01 3.48094672e-01
6.70013011e-01 -4.31476355e-01 3.12899977e-01 1.42815366e-01
-1.08647847e+00 -1.93850860e-01 -1.40460402e-01 -2.73095995e-01
-3.24704945e-02 1.85023218e-01 -8.81802499e-01 -2.25052610e-02
6.79284751e-01 6.01873755e-01 -6.18389845e-01 6.64904177e-01
-1.64200634e-01 1.21320057e+00 -2.73302943e-01 6.15317263e-02
1.35164276e-01 -1.27244115e-01 3.25778604e-01 1.48182952e+00
1.58625677e-01 -1.73750654e-01 -2.49122933e-01 6.60649300e-01
1.04848474e-01 2.00005919e-01 -5.77911615e-01 -4.09187824e-01
6.45687282e-01 8.14290464e-01 -4.47985888e-01 -7.82244578e-02
-6.27658129e-01 5.23322403e-01 5.50410628e-01 5.10280728e-01
-2.41984129e-01 5.92818521e-02 1.34972286e+00 1.36153311e-01
2.08809301e-01 -7.69529760e-01 -6.67892337e-01 -9.76976931e-01
-5.61301038e-02 -9.29390848e-01 1.78952202e-01 -5.98038435e-01
-1.41857791e+00 6.25972211e-01 6.47043437e-02 -6.93655491e-01
-4.73747611e-01 -8.59257102e-01 -7.43970215e-01 9.18263733e-01
-1.32274401e+00 -8.54370713e-01 2.54867673e-01 1.92982569e-01
6.74122393e-01 -4.01480407e-01 8.12889636e-01 5.05768955e-02
-5.85521162e-01 6.87862456e-01 3.27297419e-01 1.68384463e-01
7.07101882e-01 -1.02472568e+00 6.27235353e-01 6.80131018e-01
6.60232544e-01 9.04943347e-01 5.73438525e-01 -2.84198791e-01
-1.12212741e+00 -5.47780156e-01 8.99296582e-01 -5.94597399e-01
9.82125461e-01 -4.87525463e-01 -1.07527530e+00 6.08888507e-01
2.97574848e-01 -4.10842329e-01 1.07903481e+00 9.04207826e-01
-6.55978858e-01 -1.84999257e-01 -4.77073491e-01 6.32914662e-01
1.32427216e+00 -1.20930171e+00 -5.79554856e-01 -1.65028498e-01
7.66763210e-01 4.65851836e-02 -7.31164336e-01 3.21799457e-01
6.35023892e-01 -1.10526049e+00 7.71765292e-01 -9.07106876e-01
2.50107348e-01 8.94990563e-02 -5.42452931e-01 -1.63859630e+00
-3.98523599e-01 -3.38889003e-01 2.93890089e-01 1.45663595e+00
1.00345397e+00 -5.94242454e-01 5.10587513e-01 7.58636415e-01
-4.47805166e-01 -7.22642958e-01 -1.06305909e+00 -5.35082400e-01
5.51471412e-01 -9.62383330e-01 2.89813787e-01 7.52689540e-01
1.58823133e-01 4.40692186e-01 1.57734126e-01 5.31746298e-02
2.15186268e-01 -1.80187389e-01 4.99757886e-01 -1.16692805e+00
-4.54571337e-01 -7.76335359e-01 -2.51410395e-01 -1.14527297e+00
3.62688959e-01 -9.98914361e-01 1.30671576e-01 -1.15060532e+00
-1.15600847e-01 -6.74273431e-01 -4.38407481e-01 5.28791487e-01
1.83403492e-03 -3.03898938e-02 2.69631773e-01 2.84212083e-01
-1.05817251e-01 4.18596268e-01 6.32500052e-01 5.34630865e-02
-4.79714513e-01 1.16616860e-03 -8.73266816e-01 5.90912819e-01
9.15451467e-01 -2.64146715e-01 -5.17394781e-01 -8.53331506e-01
1.19499594e-01 -3.70242566e-01 7.87587762e-02 -1.11344075e+00
1.53523952e-01 1.04156069e-01 2.35940591e-01 -1.82704240e-01
5.25521815e-01 -4.13865358e-01 -2.12119728e-01 9.55845341e-02
-8.41687918e-01 2.91137211e-02 5.29386938e-01 4.49308395e-01
-2.11038008e-01 -3.18192989e-01 6.80724561e-01 -2.28102263e-02
-6.49636209e-01 -1.10761210e-01 -6.72222197e-01 1.68777242e-01
2.42194459e-01 -1.85040161e-01 -1.20000765e-01 -3.53435963e-01
-9.46535766e-01 -1.60307541e-01 1.76655233e-01 7.17688501e-01
2.47717783e-01 -1.09920120e+00 -7.43413150e-01 3.94106477e-01
1.73576459e-01 -4.48626101e-01 -2.89669354e-02 7.74094880e-01
1.01896472e-01 5.17009854e-01 9.50978398e-02 -6.30482078e-01
-1.22094703e+00 -4.93081771e-02 3.32519263e-01 1.44810513e-01
-1.62540063e-01 9.83064234e-01 2.04484433e-01 -4.55455989e-01
2.87249982e-01 -4.81954277e-01 -1.00191809e-01 4.79799062e-01
1.71137080e-01 1.86897188e-01 1.79910332e-01 -7.43423700e-01
-4.15924966e-01 4.99328554e-01 -1.55448526e-01 -5.20424545e-01
1.45578754e+00 -2.89416105e-01 1.21561311e-01 1.06696332e+00
1.34065843e+00 -6.91079050e-02 -1.17786992e+00 -4.43734646e-01
2.23762617e-01 -1.44349456e-01 2.38331184e-01 -5.25207102e-01
-8.89237404e-01 1.20915198e+00 5.36359251e-01 1.53342023e-01
7.78644085e-01 3.26543093e-01 3.44868302e-01 3.63539457e-01
6.10064268e-02 -1.10260689e+00 -8.18361789e-02 7.27257848e-01
8.57368529e-01 -1.03439093e+00 -2.15314761e-01 5.86766340e-02
-7.04165637e-01 9.79673386e-01 4.03435528e-01 1.02065675e-01
9.45381045e-01 2.93898702e-01 4.28372502e-01 -2.85452753e-02
-1.14971304e+00 -3.96579623e-01 3.66013229e-01 6.18638635e-01
1.01470566e+00 -2.85637360e-02 -3.90443653e-02 4.12763149e-01
-6.31771445e-01 -6.22509778e-01 7.45205507e-02 5.45166373e-01
-4.70066786e-01 -1.29969144e+00 -1.32990897e-01 5.39674103e-01
-2.58815706e-01 -4.56239074e-01 -7.48651206e-01 5.75902164e-01
4.75624669e-03 1.12050247e+00 3.16283584e-01 -6.42179728e-01
1.88998535e-01 7.72032678e-01 3.65488827e-01 -9.37097132e-01
-6.94112241e-01 -5.28559126e-02 4.23500627e-01 -2.44891912e-01
-3.51272941e-01 -1.08785594e+00 -1.04744720e+00 -5.14457449e-02
-4.52865601e-01 1.25510558e-01 8.53496611e-01 1.24179077e+00
4.43206728e-01 3.68251085e-01 4.09694344e-01 -9.66633677e-01
-5.98812461e-01 -1.21345234e+00 -3.28425318e-01 2.52087384e-01
5.04971921e-01 -5.45096815e-01 -6.90807462e-01 7.23089948e-02] | [14.309172630310059, 6.880849838256836] |
2af0a55e-f3c0-4930-a7fc-1354dfb6347e | audit-audio-editing-by-following-instructions | 2304.00830 | null | https://arxiv.org/abs/2304.00830v2 | https://arxiv.org/pdf/2304.00830v2.pdf | AUDIT: Audio Editing by Following Instructions with Latent Diffusion Models | Audio editing is applicable for various purposes, such as adding background sound effects, replacing a musical instrument, and repairing damaged audio. Recently, some diffusion-based methods achieved zero-shot audio editing by using a diffusion and denoising process conditioned on the text description of the output audio. However, these methods still have some problems: 1) they have not been trained on editing tasks and cannot ensure good editing effects; 2) they can erroneously modify audio segments that do not require editing; 3) they need a complete description of the output audio, which is not always available or necessary in practical scenarios. In this work, we propose AUDIT, an instruction-guided audio editing model based on latent diffusion models. Specifically, AUDIT has three main design features: 1) we construct triplet training data (instruction, input audio, output audio) for different audio editing tasks and train a diffusion model using instruction and input (to be edited) audio as conditions and generating output (edited) audio; 2) it can automatically learn to only modify segments that need to be edited by comparing the difference between the input and output audio; 3) it only needs edit instructions instead of full target audio descriptions as text input. AUDIT achieves state-of-the-art results in both objective and subjective metrics for several audio editing tasks (e.g., adding, dropping, replacement, inpainting, super-resolution). Demo samples are available at https://audit-demo.github.io/. | ['Sheng Zhao', 'Jiang Bian', 'Zhizheng Wu', 'Lei He', 'Xu Tan', 'Zeqian Ju', 'Yuancheng Wang'] | 2023-04-03 | null | null | null | null | ['audio-generation'] | ['audio'] | [ 4.24031734e-01 -1.97536603e-01 2.75791794e-01 -3.59805375e-02
-9.82030034e-01 -4.60709572e-01 1.73216417e-01 1.67276021e-02
-3.83924007e-01 4.59840178e-01 5.27121842e-01 1.83447003e-01
-8.49967450e-02 -6.29634738e-01 -6.11443579e-01 -4.96901125e-01
8.82634521e-02 2.07119003e-01 3.49281490e-01 -2.62861371e-01
2.47954398e-01 7.43248016e-02 -1.91872358e+00 5.63572705e-01
1.18262625e+00 8.67244542e-01 4.45471555e-01 1.07705200e+00
-5.53868078e-02 6.35843813e-01 -1.01718366e+00 -6.46307543e-02
4.31282818e-02 -7.67006278e-01 -5.43877900e-01 -8.64254218e-03
4.11970705e-01 -7.00433135e-01 -3.96074057e-01 1.15980649e+00
1.00671506e+00 4.44104105e-01 5.98418117e-01 -1.25786531e+00
-1.03825021e+00 8.32200885e-01 -3.05656582e-01 6.16491102e-02
6.89027429e-01 2.36850381e-01 7.63478935e-01 -9.64686751e-01
5.39316058e-01 1.23153591e+00 5.19034445e-01 6.66198850e-01
-1.19221568e+00 -1.05712569e+00 3.55671868e-02 3.87453258e-01
-1.36375797e+00 -6.81513429e-01 8.69358540e-01 -5.49017251e-01
6.64667606e-01 4.52182353e-01 6.51343644e-01 1.28743398e+00
1.99172303e-01 6.93703771e-01 8.31021786e-01 -3.81954849e-01
2.85726488e-01 -1.35270983e-01 -2.76910454e-01 2.28076369e-01
-7.07743883e-01 2.57461935e-01 -9.98682022e-01 -7.28049278e-02
7.32798278e-01 -3.00627023e-01 -5.21864116e-01 3.29577267e-01
-1.21879160e+00 6.86197817e-01 -1.70204759e-01 2.50767648e-01
-5.46018839e-01 1.75723303e-02 6.04123950e-01 8.24409783e-01
5.65989971e-01 2.05447018e-01 -1.81222185e-01 -4.96159375e-01
-1.21916401e+00 1.29904479e-01 4.14242566e-01 9.64086413e-01
4.05216962e-01 4.80877072e-01 -3.14412266e-01 1.17372251e+00
-1.12381540e-02 4.02389199e-01 7.45048165e-01 -1.20225394e+00
5.96013188e-01 -5.54822385e-02 6.16207570e-02 -7.66598344e-01
-6.38168305e-02 -1.64149374e-01 -9.94886279e-01 3.95377189e-01
5.26302271e-02 -3.07247013e-01 -7.26102412e-01 1.72993803e+00
1.93879306e-01 4.13341999e-01 -7.19367638e-02 8.46109748e-01
1.00186753e+00 9.16426480e-01 -2.88832515e-01 -4.69097108e-01
1.04273474e+00 -1.05287087e+00 -1.33591795e+00 -2.05043331e-02
4.46080446e-01 -1.34159088e+00 1.66151881e+00 9.00007129e-01
-1.24894834e+00 -9.25881505e-01 -9.69083369e-01 -1.70931354e-01
-4.07610759e-02 2.57005423e-01 5.42734377e-02 3.84656310e-01
-1.00279260e+00 7.49217749e-01 -5.46025991e-01 4.41484749e-02
-9.74964276e-02 1.41135260e-01 -2.92620540e-01 1.45801738e-01
-1.39059520e+00 3.15362185e-01 2.09357083e-01 -1.96089074e-02
-1.31675589e+00 -9.82876658e-01 -7.10694075e-01 1.73745856e-01
5.22851825e-01 -4.59672481e-01 1.41579008e+00 -1.10632801e+00
-1.95209515e+00 2.99999028e-01 -4.79646400e-02 -1.14519164e-01
7.46941447e-01 -7.36052215e-01 -6.61563039e-01 1.15860663e-01
2.27329303e-02 6.30463421e-01 1.27669096e+00 -9.72457051e-01
-6.41729832e-01 -4.88846526e-02 -1.40768439e-01 1.44726157e-01
-4.99820292e-01 5.73412776e-02 -5.28361261e-01 -1.33956003e+00
-9.10516530e-02 -8.15393686e-01 2.69802511e-01 1.55189410e-01
-5.11681139e-01 8.48781019e-02 1.06559837e+00 -1.28000033e+00
1.60161018e+00 -2.74625516e+00 3.58855307e-01 -1.09410234e-01
-3.13897617e-02 3.00178796e-01 -4.21929896e-01 5.97621024e-01
-4.11362201e-02 1.60736084e-01 -2.59026349e-01 -5.81332266e-01
3.50541733e-02 -9.47718248e-02 -5.88002563e-01 8.02007169e-02
-1.29661143e-01 2.31819555e-01 -9.21123385e-01 -4.84334469e-01
1.53269798e-01 7.30003655e-01 -9.08798993e-01 3.10973316e-01
-1.09473154e-01 7.00051606e-01 1.22095115e-01 4.61944669e-01
4.63407129e-01 6.02288306e-01 -3.66403937e-01 -2.10073233e-01
-2.57496893e-01 2.79897332e-01 -1.77749741e+00 1.89106929e+00
-6.94114923e-01 8.70704353e-01 3.16571385e-01 -3.98684084e-01
8.69591773e-01 7.25809276e-01 3.63072842e-01 -5.06232917e-01
-1.72015369e-01 1.27571478e-01 -1.56844720e-01 -6.62683189e-01
5.79310298e-01 -8.95371512e-02 6.49990737e-02 6.66966379e-01
2.01788381e-01 -4.64768052e-01 2.63900369e-01 1.87529832e-01
1.06258595e+00 -1.48865860e-02 -1.64345041e-01 3.15043002e-01
3.52189362e-01 -5.83722651e-01 6.01104140e-01 5.81884086e-01
1.31884199e-02 9.47518170e-01 3.42918962e-01 3.03800553e-01
-7.95424938e-01 -1.10420167e+00 2.08954781e-01 1.39995992e+00
-1.20315023e-01 -6.35718942e-01 -9.35130060e-01 -2.80171126e-01
-2.94183046e-01 9.99026954e-01 -4.14729625e-01 -4.76507962e-01
-2.97940522e-01 2.02474594e-02 8.38942826e-01 4.07253027e-01
4.58450466e-01 -1.30780542e+00 -2.29698002e-01 4.61320966e-01
-6.10536993e-01 -6.62851393e-01 -1.27971387e+00 -1.26995649e-02
-7.90357232e-01 -6.51902080e-01 -7.04642236e-01 -8.02641511e-01
3.96375924e-01 6.38679266e-02 6.12283349e-01 1.37733632e-05
-1.19682811e-01 4.14323598e-01 -3.99679810e-01 -3.94880444e-01
-6.55693054e-01 -1.59806952e-01 3.86178493e-01 1.71449780e-01
-2.13120207e-01 -9.15480494e-01 -4.74358827e-01 4.36770767e-01
-1.22226119e+00 -3.14758122e-02 4.28319216e-01 7.37316191e-01
7.26765871e-01 3.07522744e-01 8.09994400e-01 -7.03552425e-01
1.21764195e+00 -1.09389760e-01 -2.44000569e-01 -4.35528234e-02
-2.84113944e-01 -2.10394531e-01 7.22937286e-01 -1.04052913e+00
-1.17560041e+00 -1.49132401e-01 -5.39441347e-01 -7.01928020e-01
-2.45651543e-01 5.47604144e-01 -1.94268405e-01 4.56044793e-01
7.83234060e-01 1.97321177e-01 -1.33153245e-01 -8.28713715e-01
3.72196823e-01 1.10421538e+00 6.55976534e-01 -3.91992033e-01
7.03257203e-01 1.44470647e-01 -5.64856410e-01 -9.21720684e-01
-5.04163325e-01 -2.13972241e-01 -6.76177680e-01 -4.47166026e-01
6.54837251e-01 -9.07678425e-01 -2.93543667e-01 6.75355434e-01
-1.13405383e+00 -5.70928335e-01 -5.06234884e-01 7.27028549e-01
-6.97329581e-01 4.83666182e-01 -8.11206579e-01 -6.06501043e-01
-2.72019982e-01 -1.21269202e+00 9.28357959e-01 -5.37072234e-02
-6.05000556e-01 -6.26202106e-01 -3.95593345e-02 2.00141132e-01
3.86475086e-01 4.09736633e-02 9.67117369e-01 -1.96475089e-01
-2.08623067e-01 -8.08186680e-02 3.85057449e-01 7.72013605e-01
4.19215977e-01 4.17600006e-01 -9.63514209e-01 -2.84545243e-01
2.13869233e-02 -2.42923856e-01 6.93571806e-01 3.08770895e-01
1.23362553e+00 -4.40155774e-01 1.78345546e-01 5.69376826e-01
7.07285225e-01 3.18556547e-01 7.27596283e-01 -1.51525930e-01
3.70662481e-01 4.98610854e-01 9.76745784e-01 6.53352439e-01
-1.69168621e-01 7.67548978e-01 1.62403047e-01 1.31871089e-01
-5.70703089e-01 -5.44888556e-01 1.01044297e+00 1.52075613e+00
1.44186523e-02 -1.72641426e-01 -5.16164005e-01 6.11222029e-01
-1.56484330e+00 -1.04528117e+00 -1.32269710e-01 2.38880801e+00
1.28192937e+00 7.68743530e-02 2.21417472e-02 5.26344419e-01
8.58049631e-01 7.48346448e-02 -6.52512133e-01 -4.25728709e-01
1.78342029e-01 2.83118725e-01 -2.53654718e-01 6.92043722e-01
-7.26228952e-01 7.91872263e-01 5.64118052e+00 1.06221139e+00
-1.27401459e+00 4.18154359e-01 1.00367423e-02 -5.40728331e-01
-3.33191931e-01 -2.37878725e-01 -5.15580058e-01 6.64716065e-01
8.21535587e-01 -2.91118771e-01 7.40428329e-01 4.63680178e-01
6.34412646e-01 -2.11541746e-02 -1.18073857e+00 1.02529812e+00
-9.18392017e-02 -9.56873178e-01 3.11817795e-01 -3.42456669e-01
6.34210467e-01 -5.11038959e-01 3.58385980e-01 5.21363258e-01
-7.71119222e-02 -7.60143697e-01 1.08761346e+00 6.89996064e-01
1.21881449e+00 -8.08564484e-01 4.30834264e-01 4.85756487e-01
-1.13236558e+00 -4.47008610e-02 -1.50650457e-01 -3.32029685e-02
2.64091730e-01 6.75487876e-01 -3.56213927e-01 3.38215142e-01
8.36218476e-01 5.37660480e-01 -2.78362781e-01 1.12529087e+00
-5.79076290e-01 8.62529099e-01 -2.43364751e-01 4.63710606e-01
-2.53453285e-01 -1.19288459e-01 8.65024447e-01 1.20451713e+00
8.42442453e-01 1.77189130e-02 2.69981138e-02 8.54571104e-01
-1.25498071e-01 2.79339731e-01 -4.00043637e-01 -1.68451726e-01
7.51335800e-01 8.06604326e-01 -3.22849095e-01 -2.00457618e-01
-9.31536183e-02 1.28246272e+00 -2.47568697e-01 6.49118543e-01
-9.87888813e-01 -9.63243306e-01 6.34389579e-01 1.17563330e-01
1.40432656e-01 -1.51477158e-01 1.98408589e-02 -8.60728204e-01
3.21349464e-02 -1.07296538e+00 1.11093402e-01 -1.17890882e+00
-9.74608660e-01 4.74875778e-01 -2.03087896e-01 -1.54331124e+00
-2.61432528e-01 8.32586661e-02 -7.12214410e-01 8.39138806e-01
-1.03801429e+00 -6.72949016e-01 -3.90733689e-01 7.71051168e-01
9.36382771e-01 -9.68001038e-02 7.82169819e-01 9.20724511e-01
-3.76165807e-01 6.81586146e-01 1.59778558e-02 -1.74413528e-02
1.36443412e+00 -1.02857876e+00 8.14409703e-02 6.28122389e-01
2.10807800e-01 3.86390120e-01 9.47128475e-01 -7.64941633e-01
-1.01677680e+00 -1.10183620e+00 7.79108942e-01 -4.79436256e-02
5.29444635e-01 -3.06517839e-01 -1.12692356e+00 6.42751038e-01
2.71545112e-01 -4.76982534e-01 7.82680392e-01 -7.17458799e-02
-4.86265607e-02 -1.42449006e-01 -9.35092390e-01 6.92113042e-01
1.10010719e+00 -9.27215457e-01 -5.00572026e-01 1.74743369e-01
9.92443442e-01 -6.29156768e-01 -8.71092141e-01 2.53293961e-02
3.53789747e-01 -9.57391679e-01 7.36490309e-01 -1.68414026e-01
5.19619823e-01 -4.96086746e-01 -7.49369292e-03 -1.73122716e+00
-1.81180134e-01 -1.02151740e+00 -1.29582539e-01 1.76181805e+00
2.35409200e-01 -3.21972013e-01 1.27981141e-01 1.24017060e-01
-4.46770251e-01 -4.32539105e-01 -9.88488793e-01 -7.04202890e-01
-2.54832119e-01 -6.85684443e-01 6.32395387e-01 9.80182409e-01
-1.29282862e-01 1.65828094e-01 -7.14791775e-01 1.77696124e-01
3.18014383e-01 -3.01076621e-01 7.70587027e-01 -1.10893500e+00
-4.10072237e-01 -2.86413848e-01 -3.33563313e-02 -1.02133358e+00
3.96707691e-02 -5.95610082e-01 1.22378930e-01 -1.46126235e+00
-3.17544043e-01 -1.65235668e-01 -5.83589785e-02 5.55849791e-01
-5.07824756e-02 8.87596756e-02 2.84759074e-01 1.70134470e-01
-2.46221870e-01 7.61026144e-01 1.35798645e+00 -2.96881437e-01
-6.73149407e-01 2.15359569e-01 -4.52419281e-01 5.99660695e-01
7.88835824e-01 -6.66534543e-01 -5.50341606e-01 -5.26088834e-01
2.29382545e-01 3.92912954e-01 2.82144129e-01 -1.20155215e+00
2.79916823e-01 -6.11622669e-02 1.11150309e-01 -4.40657556e-01
4.84730601e-01 -7.12345421e-01 4.19899076e-01 2.92626411e-01
-6.43328846e-01 7.26711079e-02 3.40584189e-01 5.50443828e-01
-6.23828948e-01 -5.55368483e-01 7.44253099e-01 2.36997053e-01
-5.11957407e-01 8.27358961e-02 -6.48847163e-01 -9.28919390e-03
7.59624064e-01 -1.06909171e-01 -1.33972836e-03 -8.75700116e-01
-1.18365037e+00 -4.66990285e-02 2.63023138e-01 6.38156950e-01
7.96562254e-01 -1.52449477e+00 -8.11887920e-01 2.49504939e-01
-2.11425886e-01 -4.46579866e-02 7.33526230e-01 7.83856034e-01
-2.35506803e-01 -2.43717134e-01 -1.48956984e-01 -4.24029529e-01
-1.70928514e+00 4.84652311e-01 3.24216932e-02 1.16143212e-01
-6.39776826e-01 7.93478668e-01 1.00239553e-01 -3.34890485e-01
6.51752412e-01 -4.95111138e-01 -3.20301726e-02 4.41316843e-01
6.86656892e-01 4.88248229e-01 6.55838102e-02 -2.80390143e-01
2.22166494e-01 5.04070818e-01 7.68986195e-02 -4.91466194e-01
1.18871224e+00 -2.07186833e-01 1.14539862e-01 8.26409936e-01
8.46528828e-01 4.60217685e-01 -1.29500997e+00 -2.51719981e-01
-5.04891336e-01 -5.50696135e-01 2.28203714e-01 -7.59025574e-01
-1.04601836e+00 1.23407137e+00 8.11644554e-01 9.13644135e-02
1.34171903e+00 -3.11953276e-01 9.58734155e-01 6.82235360e-02
1.08577669e-01 -1.56540143e+00 4.91770297e-01 5.77688754e-01
1.45748830e+00 -6.11115575e-01 -2.48123124e-01 -2.31321946e-01
-7.04953074e-01 1.03016782e+00 4.33075666e-01 2.97539592e-01
6.37558758e-01 3.69517952e-01 2.52204657e-01 1.74427301e-01
-8.82576346e-01 5.66914305e-02 2.97031730e-01 8.05770338e-01
4.85044986e-01 -1.48584679e-01 -4.84487079e-02 8.00816357e-01
-5.00221431e-01 1.09197751e-01 6.04009151e-01 7.83582568e-01
-3.96931410e-01 -9.28402960e-01 -6.79119945e-01 3.39377493e-01
-2.67158091e-01 -1.88413247e-01 -5.13923466e-01 3.36449742e-01
2.14392513e-01 1.13893688e+00 -1.96419343e-01 -5.92938006e-01
6.41610384e-01 2.26459861e-01 2.79607087e-01 -6.79173231e-01
-7.23842621e-01 3.10244858e-01 -1.74073726e-01 -3.26427937e-01
-1.24802273e-02 -5.59987307e-01 -1.35807121e+00 -3.06137979e-01
-5.00367105e-01 2.01618657e-01 4.77271020e-01 7.07039714e-01
4.92206722e-01 1.10072947e+00 4.19763744e-01 -1.07082927e+00
-5.38931489e-01 -1.28248203e+00 -7.96954572e-01 4.54034925e-01
4.55496669e-01 -5.56582689e-01 -5.15036047e-01 5.05142450e-01] | [15.391290664672852, 5.719313621520996] |
21880dda-ffab-47fa-9c32-bc30a4b8ad2e | ctbl-augmenting-large-language-models-for | 2303.12024 | null | https://arxiv.org/abs/2303.12024v3 | https://arxiv.org/pdf/2303.12024v3.pdf | cTBLS: Augmenting Large Language Models with Conversational Tables | Optimizing accuracy and performance while eliminating hallucinations of open-domain conversational large language models (LLMs) is an open research challenge. A particularly promising direction is to augment and ground LLMs with information from structured sources. This paper introduces Conversational Tables (cTBLS), a three-step architecture to retrieve and generate dialogue responses grounded on retrieved tabular information. cTBLS uses Transformer encoder embeddings for Dense Table Retrieval and obtains up to 125% relative improvement over the retriever in the previous state-of-the-art system on the HyrbiDialogue dataset. cTBLS then uses a shared process between encoder and decoder models to perform a coarse+fine tabular knowledge (e.g., cell) ranking combined with a GPT-3.5 LLM response generator to yield a 2x relative improvement in ROUGE scores. Finally, human evaluators prefer cTBLs +80% of the time (coherency, fluency) and judge informativeness to be 4x better than the previous state-of-the-art. | ['Larry Heck', 'Anirudh S Sundar'] | 2023-03-21 | null | null | null | null | ['response-generation', 'table-retrieval'] | ['natural-language-processing', 'natural-language-processing'] | [-1.05007507e-01 6.50325418e-01 -6.16684668e-02 -4.23003644e-01
-1.70155621e+00 -5.54911315e-01 7.86189139e-01 3.67920995e-01
-2.18280256e-01 1.09331799e+00 1.08536100e+00 -1.17981784e-01
1.29953548e-01 -8.19819152e-01 -6.93319798e-01 -3.24475430e-02
3.60198587e-01 1.49224234e+00 -5.32054007e-02 -7.54250765e-01
2.71780372e-01 -3.27782519e-02 -1.17345560e+00 1.24249804e+00
9.77941811e-01 9.90443170e-01 -7.30237439e-02 8.65031481e-01
-5.76978147e-01 1.43693388e+00 -8.21464837e-01 -9.35958028e-01
-1.89943597e-01 -4.58780885e-01 -1.34521437e+00 -2.59257078e-01
4.84255016e-01 -5.19359589e-01 -4.77637619e-01 4.32627320e-01
8.22206378e-01 2.04968929e-01 6.02299213e-01 -9.65193689e-01
-9.98452008e-01 1.05318856e+00 6.52048290e-02 -3.26278657e-02
6.47982121e-01 1.47615790e-01 1.36029613e+00 -1.06713796e+00
7.89503872e-01 1.78381670e+00 2.66864359e-01 6.59957767e-01
-1.30275059e+00 -3.61211509e-01 -2.21889108e-01 1.77475035e-01
-1.15176415e+00 -6.18070245e-01 2.15640411e-01 -1.77033082e-01
1.58402300e+00 4.82752919e-01 2.76866496e-01 1.48796189e+00
2.31263727e-01 1.03967893e+00 1.02692139e+00 -4.15002465e-01
1.96569085e-01 7.55770802e-01 1.52284309e-01 7.35649228e-01
5.14737777e-02 -4.73013639e-01 -1.14093244e+00 -2.67623365e-01
3.59049231e-01 -5.21466672e-01 -2.13851765e-01 -4.93462272e-02
-1.24767983e+00 1.04810572e+00 4.73606616e-01 -1.40216067e-01
-3.66320223e-01 -1.48687303e-01 4.98371333e-01 4.30320233e-01
6.12803161e-01 1.17168093e+00 -4.76605564e-01 -4.28023010e-01
-7.12025464e-01 5.24875104e-01 1.35360801e+00 1.09568095e+00
4.62486416e-01 -1.66063294e-01 -7.33865798e-01 1.33559275e+00
1.22877367e-01 5.70316792e-01 6.28292203e-01 -8.50003839e-01
9.30265307e-01 6.84774637e-01 6.60531372e-02 -6.69556201e-01
-2.83979237e-01 -4.09558058e-01 -5.47708213e-01 -5.23197412e-01
1.32795081e-01 -1.69859335e-01 -3.40602726e-01 1.40877056e+00
-7.34657198e-02 -6.26922309e-01 4.95587856e-01 8.26738596e-01
1.27146828e+00 1.02787292e+00 -1.02615178e-01 5.69739081e-02
1.31877232e+00 -1.25095367e+00 -7.94176638e-01 -5.15086293e-01
6.88370645e-01 -8.58943522e-01 1.45821047e+00 3.57533514e-01
-1.23878968e+00 -3.82514119e-01 -1.01279783e+00 -5.84132254e-01
-3.47808570e-01 1.33093283e-01 7.11929858e-01 3.85198742e-01
-1.33771896e+00 1.27903178e-01 -3.51504922e-01 -5.33811033e-01
3.11871637e-02 1.42791897e-01 -1.90289780e-01 -5.87066896e-02
-1.60308635e+00 1.37263489e+00 1.57130212e-01 -1.81461528e-01
-9.17326868e-01 -7.54477024e-01 -8.09001088e-01 2.47042596e-01
3.88262928e-01 -8.83867145e-01 1.63912284e+00 -1.90069064e-01
-1.73076940e+00 8.49172950e-01 -1.94262296e-01 -7.53937304e-01
3.53816599e-01 -3.36216688e-01 -3.96032661e-01 3.87778729e-02
2.10740939e-01 9.57575083e-01 1.32217005e-01 -9.12588537e-01
-1.79584384e-01 -7.78233334e-02 1.62260592e-01 7.75262773e-01
-4.58787560e-01 -8.46430287e-02 -6.37034535e-01 -2.90044844e-01
-2.84973830e-01 -8.48388135e-01 1.81969292e-02 -5.20230353e-01
-6.88086629e-01 -4.85625148e-01 3.59001607e-02 -9.24221396e-01
1.41997957e+00 -1.50300169e+00 2.22160071e-01 -2.36165106e-01
3.87933552e-01 1.69528052e-02 -3.95032197e-01 9.82189715e-01
3.36936563e-01 1.62613019e-01 2.28530943e-01 -3.25717926e-01
2.67660946e-01 -1.99771851e-01 -5.31635821e-01 -2.35965386e-01
3.68033171e-01 1.20251358e+00 -7.76745260e-01 -3.43028635e-01
-1.80432182e-02 2.37906858e-01 -7.78015435e-01 5.16551733e-01
-5.59912980e-01 -1.10164538e-01 -3.81823510e-01 4.80841190e-01
3.07405982e-02 -5.49106658e-01 2.30042815e-01 -7.51346946e-02
3.05382282e-01 1.18843150e+00 -6.60306573e-01 1.64818263e+00
-9.81657088e-01 8.57920945e-01 -3.74157727e-01 -1.35154903e-01
1.05509424e+00 2.48912692e-01 3.48961651e-02 -1.06832027e+00
-2.41045561e-02 1.92393199e-01 -3.01730543e-01 -4.60305631e-01
1.09765422e+00 1.25850901e-01 -4.53592151e-01 6.08084857e-01
2.67031223e-01 -2.05240041e-01 1.58147797e-01 7.70146966e-01
1.23314309e+00 -1.72141746e-01 2.09055483e-01 -2.26900876e-01
4.75837111e-01 2.70829052e-01 -1.80081233e-01 8.53697300e-01
3.04982454e-01 3.60961974e-01 6.64985299e-01 -2.88256705e-01
-1.00329781e+00 -1.03165376e+00 1.33128136e-01 1.36040556e+00
-2.37354666e-01 -6.74695730e-01 -8.92051458e-01 -4.94057775e-01
5.80159724e-02 1.24383211e+00 -4.71214712e-01 -2.15951532e-01
-3.22587997e-01 -6.17061198e-01 6.69560313e-01 5.41669130e-01
4.01709974e-01 -1.43087173e+00 -1.88718691e-01 3.89966726e-01
-9.04772878e-01 -1.19206572e+00 -3.81013513e-01 1.56770140e-01
-6.21871233e-01 -5.90635955e-01 -4.36051816e-01 -5.11350155e-01
2.22277343e-01 -9.21703428e-02 1.80514216e+00 -3.54605436e-01
-4.25361469e-02 8.89365748e-03 -3.97668570e-01 -2.21274689e-01
-7.39496648e-01 5.28500140e-01 -8.53479207e-02 -3.05426151e-01
5.49817562e-01 -1.41000524e-02 -3.01580429e-01 9.46194679e-02
-4.25840884e-01 3.05413365e-01 7.21285105e-01 1.10530388e+00
3.41670632e-01 -5.39125443e-01 7.70198762e-01 -1.08663487e+00
1.44267249e+00 -3.69183272e-01 -3.12563926e-01 4.81714249e-01
-7.45830059e-01 5.95999122e-01 5.80143631e-01 -1.14910461e-01
-1.09349716e+00 -5.24148047e-01 -2.28086352e-01 -7.74000734e-02
2.53192842e-01 5.23001075e-01 -1.20469607e-01 5.11583447e-01
1.02308524e+00 1.00869074e-01 -4.95318398e-02 -4.16165799e-01
7.03243971e-01 1.14755571e+00 3.29254299e-01 -5.39649606e-01
1.21915400e-01 -6.01804703e-02 -8.01012695e-01 -5.86239398e-01
-9.46309209e-01 -3.22324961e-01 -2.05408767e-01 -2.56696731e-01
8.13134491e-01 -1.07429266e+00 -9.11319673e-01 -1.00051500e-01
-1.25049675e+00 -4.90985721e-01 -1.52429193e-01 1.81489825e-01
-6.42957807e-01 -3.16731274e-01 -1.01664102e+00 -7.28011966e-01
-8.34185541e-01 -1.06466997e+00 1.20927858e+00 -2.82157902e-02
-8.60558927e-01 -7.78543234e-01 4.44223620e-02 9.81005609e-01
5.37864745e-01 -3.47525030e-01 1.02308547e+00 -1.10949230e+00
-6.98163569e-01 -1.03507943e-01 -2.64878243e-01 9.61794779e-02
-3.68530899e-01 -4.43576247e-01 -1.16424811e+00 -1.39580086e-01
-2.51048088e-01 -1.13881207e+00 8.73558044e-01 -2.73922067e-02
8.46510172e-01 -6.17087305e-01 -3.24225016e-02 1.44141942e-01
1.11083424e+00 2.77304262e-01 7.87756205e-01 2.59551674e-01
4.83295888e-01 7.80325413e-01 5.54528832e-01 5.30160189e-01
7.35171974e-01 6.74375057e-01 -1.27558544e-01 1.45318106e-01
-2.22603634e-01 -7.88217127e-01 6.67500556e-01 1.25150168e+00
8.14699531e-01 -6.82725012e-01 -8.43119025e-01 4.89632457e-01
-1.55001068e+00 -7.17189372e-01 1.55354887e-01 2.01371932e+00
1.16174519e+00 3.19483429e-01 -2.16431558e-01 -3.13717902e-01
3.04768354e-01 1.83756858e-01 -4.97338235e-01 -8.50363970e-01
-1.39681712e-01 3.10514778e-01 2.34069437e-01 8.67223799e-01
-4.74636018e-01 1.30405867e+00 6.08812618e+00 8.99386704e-01
-6.56997681e-01 4.72211419e-03 1.04613829e+00 -2.93939412e-01
-6.80099070e-01 -2.21606568e-01 -1.21236372e+00 1.43927515e-01
1.55865681e+00 -3.64186883e-01 7.41762400e-01 8.31616521e-01
-6.30607679e-02 2.77322643e-02 -1.24424732e+00 9.56543446e-01
5.93237698e-01 -1.64967895e+00 4.49449420e-01 -1.79025143e-01
5.09289622e-01 1.14911467e-01 -4.81800959e-02 1.07570505e+00
7.36632884e-01 -1.24142289e+00 4.89121825e-01 6.39726698e-01
9.44350481e-01 -7.42325842e-01 8.55380595e-01 2.38080844e-01
-5.80052495e-01 1.74190521e-01 -4.68617827e-01 1.64899349e-01
5.38613982e-02 1.85253978e-01 -1.62192059e+00 2.13720247e-01
5.74103296e-01 1.54942796e-01 -7.35811710e-01 3.76854956e-01
-1.61159530e-01 4.80394095e-01 -1.62365124e-01 -8.55895996e-01
2.49418050e-01 1.80896163e-01 2.09562331e-01 1.36369526e+00
1.29333860e-03 6.57350346e-02 3.40951718e-02 8.81599367e-01
-5.90328932e-01 3.41109544e-01 -5.21873236e-01 -3.44579816e-01
8.36911738e-01 1.15002048e+00 -1.16849914e-01 -6.88495755e-01
-1.03266433e-01 1.10024893e+00 6.87653661e-01 9.67621654e-02
-5.35130024e-01 -3.50254118e-01 4.80500013e-01 5.97018041e-02
1.40085921e-03 3.54405254e-01 -4.09018785e-01 -1.22968376e+00
-9.99685004e-03 -1.53106129e+00 3.80172670e-01 -1.15934706e+00
-1.42478788e+00 1.21485424e+00 -1.82521001e-01 -1.03132403e+00
-1.02606714e+00 -3.39788377e-01 7.28712305e-02 1.11915112e+00
-1.10830605e+00 -9.87693667e-01 -2.57885098e-01 3.52640003e-01
1.10700703e+00 -3.65086108e-01 1.25968957e+00 1.73956856e-01
-2.86326557e-01 7.78881431e-01 3.83025371e-02 1.29925951e-01
1.01009226e+00 -1.28914392e+00 6.60778284e-01 1.54588789e-01
1.92429677e-01 9.47329342e-01 6.05129540e-01 -6.99126840e-01
-1.50866330e+00 -1.05766535e+00 1.45049596e+00 -1.01634896e+00
7.34940708e-01 -8.32144856e-01 -7.25457251e-01 6.04922831e-01
7.87902057e-01 -6.28934920e-01 6.79268301e-01 4.95159924e-01
-5.15716493e-01 -2.79239655e-01 -8.61639857e-01 8.51472378e-01
6.32061839e-01 -8.98085058e-01 -7.13485599e-01 6.04674578e-01
1.01533687e+00 -7.50631094e-01 -8.23714197e-01 1.87942982e-02
4.71137226e-01 -8.35118711e-01 8.13380778e-01 -7.49066770e-01
8.83780241e-01 2.40778282e-01 -3.03030252e-01 -1.58292222e+00
-2.12513804e-01 -7.99440324e-01 -6.32716566e-02 1.09499383e+00
1.03406358e+00 -2.62933046e-01 5.11645436e-01 4.62562054e-01
-2.66739637e-01 -9.08245087e-01 -5.76286614e-01 -3.31390142e-01
1.28551498e-01 -2.41832957e-01 5.53236604e-01 6.34552419e-01
4.78159755e-01 1.24847162e+00 -3.80521089e-01 -2.94770479e-01
2.10091487e-01 4.04068381e-02 9.68814552e-01 -8.83447409e-01
-3.04417193e-01 -2.96348453e-01 1.01757728e-01 -1.33690584e+00
1.55618921e-01 -1.04313719e+00 6.31058589e-02 -2.01272631e+00
3.00733238e-01 -5.93138859e-02 -6.44787699e-02 3.48327130e-01
-4.09650542e-02 6.18795902e-02 1.85099974e-01 -1.26343891e-01
-9.04160559e-01 7.63861179e-01 1.07801616e+00 -2.25802630e-01
-2.05692545e-01 -4.60951000e-01 -1.22568500e+00 1.16860189e-01
5.91248214e-01 -3.12167555e-01 -6.46502435e-01 -4.36294436e-01
3.46096039e-01 8.51207614e-01 -1.23183234e-02 -1.00272465e+00
3.67482543e-01 9.05872434e-02 4.09571975e-01 -7.84881711e-01
8.11610401e-01 1.01577463e-02 -1.75232112e-01 3.18402052e-02
-1.26185381e+00 4.07209337e-01 4.66387987e-01 1.75981686e-01
-3.30913335e-01 6.17884099e-02 4.09093082e-01 -2.43927225e-01
-3.63362908e-01 -1.59178033e-01 -5.82865894e-01 5.31411827e-01
1.58322647e-01 4.29966688e-01 -7.81399012e-01 -8.94468963e-01
-3.96921098e-01 3.30986351e-01 -7.09354803e-02 8.82352591e-01
8.42413008e-01 -1.21450865e+00 -9.91714895e-01 1.03979431e-01
4.12585020e-01 -4.22610044e-01 2.19874740e-01 3.64335507e-01
-5.69589674e-01 1.08700132e+00 -1.44875750e-01 -2.13064224e-01
-9.94830966e-01 1.94664672e-01 2.21338436e-01 -7.15287745e-01
-4.02423024e-01 1.14011180e+00 -7.14475289e-02 -8.06295514e-01
2.85964191e-01 -2.48801678e-01 -2.48517901e-01 3.05248767e-01
6.43162012e-01 2.34950259e-01 5.63407958e-01 -4.36126776e-02
-2.06041366e-01 -3.16153675e-01 -5.43364584e-01 -6.90209985e-01
1.07291234e+00 1.14485929e-02 -8.00741389e-02 6.13042593e-01
1.31375229e+00 2.46653557e-01 -6.47020876e-01 -2.57272810e-01
-1.51541578e-02 -1.87313259e-01 -4.73471172e-02 -1.58460796e+00
-4.82491612e-01 8.01919639e-01 9.40233171e-02 -8.38874374e-03
5.93566477e-01 2.42174380e-02 9.36140001e-01 9.36898828e-01
3.27858239e-01 -1.06412959e+00 4.90669817e-01 8.56301904e-01
1.42140400e+00 -1.22240984e+00 -2.06041127e-01 -7.97744393e-02
-1.42603779e+00 9.22112048e-01 9.10006642e-01 1.65787905e-01
-7.01993704e-02 2.47849360e-01 2.60768682e-01 -2.39340246e-01
-1.68244076e+00 -1.68264378e-02 3.73968065e-01 2.92875051e-01
9.43700910e-01 1.85407251e-01 -1.24877840e-01 7.87142634e-01
-7.36058772e-01 -2.37128645e-01 4.89037365e-01 3.39818627e-01
-3.47556114e-01 -8.54619265e-01 -1.58405565e-02 7.54995883e-01
-4.49232981e-02 -6.61406457e-01 -8.33078921e-01 6.93401635e-01
-6.79214358e-01 1.30561936e+00 5.06013706e-02 -7.64966190e-01
4.22531813e-01 3.39213163e-01 8.54093209e-02 -7.67588496e-01
-9.51746047e-01 -8.68656021e-03 9.35750961e-01 -7.40046203e-01
3.92738104e-01 -4.58848745e-01 -1.10827756e+00 -5.40723145e-01
-1.89474419e-01 4.72570449e-01 4.08600628e-01 4.88070875e-01
5.65207303e-01 6.55479729e-01 4.80043918e-01 -8.70559067e-02
-9.14652407e-01 -1.30819476e+00 -9.92328152e-02 1.90378293e-01
-8.69356096e-02 -3.34235251e-01 -2.26884067e-01 -3.65919113e-01] | [11.76514720916748, 8.37051010131836] |
a81decd1-d9fa-42d3-b76d-008fda410f45 | research-note-on-uncertain-probabilities-and | 2208.10932 | null | https://arxiv.org/abs/2208.10932v1 | https://arxiv.org/pdf/2208.10932v1.pdf | Research Note on Uncertain Probabilities and Abstract Argumentation | The sixth assessment of the international panel on climate change (IPCC) states that "cumulative net CO2 emissions over the last decade (2010-2019) are about the same size as the 11 remaining carbon budget likely to limit warming to 1.5C (medium confidence)." Such reports directly feed the public discourse, but nuances such as the degree of belief and of confidence are often lost. In this paper, we propose a formal account for allowing such degrees of belief and the associated confidence to be used to label arguments in abstract argumentation settings. Differently from other proposals in probabilistic argumentation, we focus on the task of probabilistic inference over a chosen query building upon Sato's distribution semantics which has been already shown to encompass a variety of cases including the semantics of Bayesian networks. Borrowing from the vast literature on such semantics, we examine how such tasks can be dealt with in practice when considering uncertain probabilities, and discuss the connections with existing proposals for probabilistic argumentation. | ['Murat Sensoy', 'Lance M. Kaplan', 'Massimiliano Giacomin', 'Federico Cerutti', 'Pietro Baroni'] | 2022-08-23 | null | null | null | null | ['abstract-argumentation', 'abstract-argumentation'] | ['natural-language-processing', 'reasoning'] | [ 3.13354820e-01 8.85635376e-01 -4.88531351e-01 -4.44772691e-01
-6.16593540e-01 -9.21352446e-01 1.06597197e+00 6.25069499e-01
-4.59640235e-01 1.05574572e+00 4.19197738e-01 -1.11487281e+00
-4.59690571e-01 -1.09028506e+00 -7.12626219e-01 -6.55746043e-01
2.25998431e-01 3.88533086e-01 3.38494003e-01 -2.02728119e-02
4.48945552e-01 2.14960128e-01 -1.57384872e+00 3.24651934e-02
8.14756095e-01 7.04851687e-01 -1.75261274e-01 4.23582703e-01
-3.29275936e-01 5.95555544e-01 -6.94782197e-01 -6.17557406e-01
-3.19623172e-01 8.89931917e-02 -1.02274823e+00 -6.35623991e-01
-3.18106785e-02 -3.18928480e-01 2.00482175e-01 1.45693779e+00
1.29943922e-01 -4.10909764e-02 9.57624257e-01 -1.28752637e+00
-9.07223597e-02 1.43450260e+00 -3.38091403e-01 1.86461776e-01
4.16477650e-01 -3.66871536e-01 1.18023670e+00 -4.76546705e-01
3.91320646e-01 1.55070806e+00 5.86106300e-01 3.66676658e-01
-9.48169649e-01 -4.98574555e-01 6.54498756e-01 2.48628054e-02
-9.45434153e-01 -4.30783294e-02 4.88429695e-01 -5.35616040e-01
5.28205991e-01 5.21689296e-01 5.59913278e-01 9.36205566e-01
1.38791129e-02 4.30051953e-01 1.30071270e+00 -6.33467078e-01
6.98797464e-01 -7.31156720e-03 2.54024982e-01 -5.62817371e-03
8.28639388e-01 1.34352222e-01 -4.23457623e-01 -5.64148545e-01
4.39830646e-02 -6.16924644e-01 -2.41402566e-01 -1.64559081e-01
-1.16510773e+00 1.11317217e+00 2.40682811e-01 2.08010584e-01
-1.67632222e-01 6.16968334e-01 2.66574562e-01 -9.59852561e-02
5.17497361e-01 9.67792645e-02 -5.62006176e-01 7.33714402e-02
-6.69648170e-01 4.98754114e-01 1.10142374e+00 6.13789499e-01
2.50062644e-01 -3.51530105e-01 1.53359458e-01 2.44360819e-01
1.15689981e+00 9.62847412e-01 -2.89100617e-01 -1.50637865e+00
4.98827726e-01 3.41809355e-02 8.56808364e-01 -9.31693673e-01
-1.72248170e-01 -2.40786582e-01 -4.01270390e-01 3.79369557e-01
7.05734551e-01 -3.74215156e-01 -5.95803738e-01 2.09123898e+00
3.46217126e-01 -3.28939080e-01 2.70639539e-01 6.04165018e-01
4.69226480e-01 7.05739677e-01 6.33770585e-01 -2.53620744e-01
1.40286338e+00 1.89550161e-01 -8.08757365e-01 3.28015313e-02
4.62650001e-01 -4.21442360e-01 7.17933834e-01 4.20021474e-01
-8.43442321e-01 2.68063873e-01 -1.07925355e+00 3.41699779e-01
-5.47478795e-01 -5.53885460e-01 7.38374352e-01 1.06209815e+00
-7.37965763e-01 5.36683083e-01 -6.78567946e-01 -3.42139810e-01
-9.65296105e-03 -1.96643263e-01 2.85408169e-01 2.76662320e-01
-1.80571544e+00 1.51695275e+00 6.04605556e-01 4.91183698e-01
-6.87062263e-01 -4.66582865e-01 -6.66081190e-01 7.53325522e-02
4.98292625e-01 -6.67570829e-01 1.35883570e+00 -3.45937222e-01
-1.29893064e+00 7.07662344e-01 7.75693431e-02 -4.97116745e-01
7.79730320e-01 -3.19059521e-01 -4.03066009e-01 -2.11347546e-03
1.19681723e-01 7.13262916e-01 2.57432371e-01 -1.27365601e+00
-7.95522451e-01 -3.89287829e-01 7.24288881e-01 3.22438449e-01
1.97827950e-01 3.11530083e-01 3.47940892e-01 -2.78816849e-01
2.23591715e-01 -1.00479460e+00 -3.07349175e-01 4.42029461e-02
-5.60264647e-01 -5.49483657e-01 4.58478630e-01 -1.17490403e-01
1.00503731e+00 -1.62639141e+00 -1.09027155e-01 5.25254607e-01
-3.08585793e-01 -2.13006184e-01 6.08737886e-01 3.98869663e-01
1.04971565e-02 6.78409100e-01 -7.22944617e-01 3.70784909e-01
6.05219901e-01 4.85382318e-01 -9.31715488e-01 5.36574483e-01
-1.35145158e-01 4.33630317e-01 -1.17664337e+00 -3.38349730e-01
1.75521687e-01 2.53420264e-01 -9.11897793e-02 -2.48461217e-01
-5.59969723e-01 9.15907994e-02 -6.83142900e-01 3.20947319e-01
5.65995336e-01 -1.26612514e-01 7.03990817e-01 1.50611913e-02
-3.92697036e-01 5.20663977e-01 -1.26717854e+00 1.20119524e+00
-5.10132276e-02 2.90622175e-01 1.99352652e-01 -6.80638134e-01
5.45058668e-01 4.07468706e-01 -8.81218910e-02 1.23344334e-02
-1.14136800e-01 3.10592175e-01 3.67154144e-02 -2.09277377e-01
4.82237816e-01 -4.33893144e-01 -4.93586063e-01 6.32139683e-01
-4.84024137e-01 -1.00162458e+00 6.41021356e-02 3.87550563e-01
3.40758443e-01 4.32761878e-01 3.45914990e-01 -7.88127720e-01
5.80513000e-01 1.88814979e-02 4.01888400e-01 8.53682399e-01
-7.76257962e-02 1.55951709e-01 8.23984325e-01 -1.52363688e-01
-8.53387594e-01 -1.33632159e+00 -5.26208878e-01 9.28410590e-01
1.10507369e-01 -2.07712665e-01 -6.84576273e-01 -6.15942538e-01
3.87911908e-02 1.53453243e+00 -5.87596536e-01 2.51384228e-01
-1.08840369e-01 -1.02762544e+00 7.16736913e-01 3.41596872e-01
2.00715989e-01 -7.27096021e-01 -8.89777422e-01 7.87004530e-02
-3.17819566e-01 -6.96531773e-01 3.24255735e-01 2.86088139e-01
-7.54291058e-01 -1.27082062e+00 -5.84229648e-01 2.04702958e-01
3.11648846e-01 -2.44436026e-01 1.15364099e+00 -2.91897804e-01
5.43773413e-01 6.18113339e-01 -1.47717878e-01 -1.05759537e+00
-7.45086432e-01 -2.51005232e-01 -1.47300407e-01 -7.82735348e-01
2.10935667e-01 -5.41407764e-01 -5.49752474e-01 -2.36232020e-02
-1.02301490e+00 -1.95528880e-01 1.21469624e-01 3.56814623e-01
2.15500295e-01 1.59867331e-01 7.38157153e-01 -1.15184450e+00
7.36191511e-01 -7.43588030e-01 -8.19650292e-01 5.36159515e-01
-8.15999687e-01 1.95391223e-01 -2.18673706e-01 -1.29821580e-02
-1.40464294e+00 -5.38949609e-01 8.89847986e-03 5.24069309e-01
-2.91759312e-01 1.03076732e+00 -2.61357367e-01 4.05765235e-01
3.12597275e-01 -3.60744923e-01 -2.85303682e-01 -1.65365815e-01
9.73034441e-01 4.06689674e-01 5.30065477e-01 -1.30871630e+00
5.67492425e-01 6.42051756e-01 1.99296027e-01 -4.21806067e-01
-1.28284097e+00 -2.32882034e-02 -2.77685493e-01 -3.16745311e-01
6.53795063e-01 -7.57517040e-01 -7.55802572e-01 2.60012031e-01
-1.39414990e+00 -1.52440853e-02 -2.90109307e-01 6.91717207e-01
-6.41127169e-01 4.75374103e-01 -1.18065916e-01 -1.36605632e+00
-1.05261534e-01 -7.91532099e-01 5.78943431e-01 4.70706858e-02
-4.84368384e-01 -1.20507944e+00 1.78916350e-01 4.93331924e-02
4.35637593e-01 4.93708253e-01 1.20238400e+00 -6.04502022e-01
-1.10427283e-01 1.99580535e-01 -2.44229987e-01 2.08773613e-01
-1.11451723e-01 3.35347503e-01 -1.07791853e+00 2.30952486e-01
1.84860229e-01 -2.10191861e-01 9.28429246e-01 4.77254063e-01
7.58437872e-01 -3.80603075e-01 -3.04477423e-01 -4.11334693e-01
1.51690042e+00 4.85288948e-02 6.02081835e-01 4.17331874e-01
1.27352569e-02 1.21960878e+00 5.13837755e-01 3.60055238e-01
6.20734334e-01 2.64509767e-01 8.83026123e-01 5.42686164e-01
2.96176612e-01 -1.66968912e-01 1.52234077e-01 4.71068174e-01
-1.55717537e-01 -4.24037367e-01 -1.09104753e+00 6.17805421e-01
-1.90979290e+00 -1.21120536e+00 -2.52045661e-01 2.31900883e+00
1.07051158e+00 5.01275599e-01 -1.06016420e-01 1.90776959e-01
9.83572960e-01 3.84447157e-01 -2.95205891e-01 -5.04678428e-01
-2.91210145e-01 2.80900276e-03 7.01608539e-01 8.87217343e-01
-9.20205951e-01 3.65071505e-01 7.57749891e+00 5.77196240e-01
-5.14881074e-01 1.65374383e-01 6.84713960e-01 4.02421802e-01
-1.24607122e+00 6.31456971e-01 -6.23238385e-01 4.16085809e-01
1.32367492e+00 -2.21100867e-01 -1.64038777e-01 4.74325180e-01
7.90316090e-02 -6.05948985e-01 -9.25418019e-01 5.05290739e-02
-6.04993403e-01 -1.08522666e+00 1.12946533e-01 -4.13708426e-02
6.58290088e-01 2.77183168e-02 -1.16462082e-01 -6.12525530e-02
1.29719591e+00 -9.97231960e-01 1.23611963e+00 6.90060675e-01
6.21787131e-01 -7.08176970e-01 8.93786252e-01 3.70810956e-01
-7.56501019e-01 2.97259204e-02 -2.91972429e-01 -1.49892539e-01
4.55947161e-01 1.02001071e+00 -3.75871748e-01 7.03809857e-01
5.89888275e-01 1.98331907e-01 1.94352376e-03 7.79038310e-01
-9.10527170e-01 8.30657184e-01 -8.41664732e-01 -2.05181181e-01
4.22155410e-01 -1.57689363e-01 5.80296576e-01 1.20343959e+00
4.56120640e-01 1.05673619e-01 -3.46725374e-01 9.29251969e-01
2.14238852e-01 -3.18556607e-01 -5.39852262e-01 1.40603647e-01
8.32273841e-01 5.70453644e-01 -7.03161538e-01 -3.37506026e-01
-2.84168392e-01 -1.45745471e-01 -1.80498481e-01 3.84907871e-01
-8.44852626e-01 -2.39598006e-01 2.86239356e-01 -2.59616256e-01
9.03737824e-03 9.46930330e-03 -2.44145826e-01 -8.30051839e-01
-1.16583847e-01 -4.37224865e-01 5.79251647e-01 -8.84999990e-01
-1.18368042e+00 -5.27185276e-02 8.75281155e-01 -7.48492360e-01
-2.76117265e-01 -7.10823596e-01 -7.85361707e-01 1.00813568e+00
-1.79479134e+00 -8.80931139e-01 4.76499200e-01 -9.99749824e-02
-6.06350042e-02 7.46377468e-01 1.09476674e+00 -4.55173135e-01
-8.02587792e-02 -3.80477160e-01 5.00927605e-02 -2.80957282e-01
4.70194668e-01 -1.58998287e+00 2.38053858e-01 7.51374304e-01
-3.82892162e-01 7.81016409e-01 1.22788906e+00 -5.84969223e-01
-9.14891779e-01 -6.60376430e-01 1.01221406e+00 -5.62572718e-01
9.92459178e-01 7.47135878e-02 -6.21052742e-01 5.93540967e-01
3.81290704e-01 -2.76476920e-01 5.78347623e-01 3.77270192e-01
-7.90081859e-01 2.74507582e-01 -1.28683031e+00 4.65203822e-01
5.77310205e-01 -3.96285862e-01 -1.08484805e+00 2.98608780e-01
7.84681559e-01 -8.21982920e-02 -1.01074409e+00 6.51085854e-01
1.01037610e+00 -8.84544313e-01 9.54638660e-01 -6.62477136e-01
2.03260593e-02 -7.00401425e-01 -7.23208070e-01 -1.11377835e+00
6.00080602e-02 -4.68524456e-01 5.72748482e-02 1.20211649e+00
6.99020147e-01 -8.61153841e-01 4.85726386e-01 9.57334220e-01
1.04273975e-01 -2.49191910e-01 -1.39256716e+00 -5.05900383e-01
7.75887489e-01 -8.55334938e-01 6.16184652e-01 8.46170723e-01
1.25366107e-01 -2.16530830e-01 4.58054282e-02 4.78983939e-01
9.80899453e-01 1.06560841e-01 4.05154899e-02 -1.70500553e+00
1.78164318e-01 -5.89720368e-01 1.16919979e-01 -5.57643652e-01
2.41815522e-01 -5.97068131e-01 2.27779821e-01 -1.67547262e+00
1.69663295e-01 -6.18996024e-01 -4.16476756e-01 2.40465567e-01
-1.41887382e-01 -3.00878912e-01 1.65302604e-01 -3.00607886e-02
-2.87785947e-01 3.75332803e-01 7.10158527e-01 -3.86384785e-01
5.14131129e-01 1.07946016e-01 -1.07193303e+00 1.17316771e+00
8.14882040e-01 -7.37201095e-01 -4.37205672e-01 -2.13031277e-01
1.32592738e+00 1.07474402e-01 5.06576002e-01 -4.51823384e-01
1.38772696e-01 -7.85900414e-01 -6.96938531e-03 -6.88829005e-01
9.58092883e-02 -8.30398619e-01 3.91718835e-01 6.25295877e-01
-7.82081485e-01 -1.95171371e-01 1.65014211e-02 9.65600610e-01
9.28981900e-02 -7.80334294e-01 3.61540854e-01 -1.46982804e-01
-3.36200774e-01 -3.97548765e-01 -7.44184613e-01 7.19763637e-02
6.50963366e-01 2.05122575e-01 -8.23266447e-01 -4.48257297e-01
-9.71625149e-01 3.56444657e-01 3.08624119e-01 4.42228653e-02
1.44553870e-01 -1.02579689e+00 -8.26577365e-01 -9.27204847e-01
2.75163893e-02 -2.23302506e-02 -9.84761715e-02 5.56769848e-01
-3.19089442e-01 5.27466118e-01 2.37486407e-01 -2.67902821e-01
-4.63703424e-01 3.71880800e-01 2.55862296e-01 -2.03905210e-01
-2.43668735e-01 4.93630677e-01 7.37079605e-02 -5.96375763e-01
2.56453007e-01 -6.93243563e-01 -3.72835219e-01 2.86825687e-01
4.13558424e-01 5.09217262e-01 -4.34414983e-01 -3.29006523e-01
-6.86264455e-01 5.36058009e-01 4.26687777e-01 -7.61799335e-01
1.02032650e+00 -4.97378349e-01 -4.60585326e-01 9.60253537e-01
1.90339148e-01 4.16669399e-01 -9.48138118e-01 8.37716535e-02
4.64302212e-01 2.71603949e-02 -1.00441001e-01 -1.20994937e+00
-3.67190719e-01 9.06279564e-01 2.73740560e-01 5.44809222e-01
4.62529451e-01 1.45328507e-01 -5.77939041e-02 5.17514229e-01
4.63535100e-01 -9.47955012e-01 -7.64765978e-01 2.56035894e-01
1.09106910e+00 -9.55978334e-01 3.75401229e-01 -4.79146004e-01
-2.92343885e-01 1.29283285e+00 -1.01305977e-01 2.85663128e-01
8.46013129e-01 2.37812921e-01 -1.11814603e-01 -1.71670616e-01
-7.26614594e-01 -1.66420825e-02 -7.28352070e-02 6.56872869e-01
5.63489020e-01 6.13899291e-01 -5.79975009e-01 4.76557225e-01
-2.40236446e-01 -2.97035784e-01 9.14089859e-01 9.41388667e-01
-8.56552362e-01 -9.76742029e-01 -5.25573492e-01 1.25568375e-01
-8.42274964e-01 -2.33524933e-01 -3.08666587e-01 6.65184617e-01
-4.61034104e-02 1.38438058e+00 -1.64680719e-01 3.82973045e-01
3.12144347e-02 1.29889384e-01 4.64977950e-01 -4.75136012e-01
-1.75232187e-01 -2.42375240e-01 8.35238159e-01 -1.20630980e-01
-1.26788044e+00 -7.85776019e-01 -1.13842571e+00 -3.67806464e-01
-4.72680807e-01 7.54985452e-01 9.90217090e-01 1.05261862e+00
-1.79918334e-01 2.17543200e-01 5.17740147e-03 -4.86516595e-01
-9.35322583e-01 -1.08428466e+00 -6.07468426e-01 -1.99780703e-01
1.97316840e-01 -6.85854554e-01 -8.50463629e-01 -1.10513084e-01] | [8.285189628601074, 5.762246131896973] |
6b8e7c51-d368-498b-b346-6d84529b5094 | using-massive-multilingual-pre-trained | 2210.06068 | null | https://arxiv.org/abs/2210.06068v2 | https://arxiv.org/pdf/2210.06068v2.pdf | Investigating Massive Multilingual Pre-Trained Machine Translation Models for Clinical Domain via Transfer Learning | Massively multilingual pre-trained language models (MMPLMs) are developed in recent years demonstrating superpowers and the pre-knowledge they acquire for downstream tasks. This work investigates whether MMPLMs can be applied to clinical domain machine translation (MT) towards entirely unseen languages via transfer learning. We carry out an experimental investigation using Meta-AI's MMPLMs ``wmt21-dense-24-wide-en-X and X-en (WMT21fb)'' which were pre-trained on 7 language pairs and 14 translation directions including English to Czech, German, Hausa, Icelandic, Japanese, Russian, and Chinese, and the opposite direction. We fine-tune these MMPLMs towards English-\textit{Spanish} language pair which \textit{did not exist at all} in their original pre-trained corpora both implicitly and explicitly. We prepare carefully aligned \textit{clinical} domain data for this fine-tuning, which is different from their original mixed domain knowledge. Our experimental result shows that the fine-tuning is very successful using just 250k well-aligned in-domain EN-ES segments for three sub-task translation testings: clinical cases, clinical terms, and ontology concepts. It achieves very close evaluation scores to another MMPLM NLLB from Meta-AI, which included Spanish as a high-resource setting in the pre-training. To the best of our knowledge, this is the first work on using MMPLMs towards \textit{clinical domain transfer-learning NMT} successfully for totally unseen languages during pre-training. | ['Goran Nenadic', 'Serge Gladkoff', 'Irina Sorokina', 'Gleb Erofeev', 'Lifeng Han'] | 2022-10-12 | null | null | null | null | ['zero-shot-machine-translation'] | ['natural-language-processing'] | [ 3.39058191e-01 4.00308549e-01 -4.64497298e-01 -3.94136280e-01
-1.55682003e+00 -5.15752137e-01 4.17792827e-01 -3.63945439e-02
-8.38291049e-01 1.26145089e+00 2.42690563e-01 -9.67238367e-01
5.77004440e-02 -4.13231730e-01 -8.63583863e-01 -2.25674152e-01
1.29948214e-01 1.27571189e+00 -1.39442265e-01 -4.58985478e-01
-4.83270675e-01 -6.83002919e-02 -1.75235465e-01 9.92811441e-01
1.03531671e+00 1.57338262e-01 1.96124434e-01 6.15487814e-01
1.06101096e-01 6.01629734e-01 -2.78411001e-01 -7.43444383e-01
1.34086058e-01 -6.48332179e-01 -1.32761681e+00 -3.94249976e-01
1.65500522e-01 -1.51037589e-01 -8.72369558e-02 6.87690258e-01
7.36883163e-01 -4.60521281e-01 7.65194356e-01 -6.20264471e-01
-1.00406539e+00 7.25477457e-01 -2.83092499e-01 2.16003209e-01
2.66511858e-01 3.21532279e-01 7.09770560e-01 -9.12906170e-01
1.30482423e+00 1.10267973e+00 7.69029915e-01 9.91404533e-01
-1.06818140e+00 -8.32451046e-01 -2.55601645e-01 1.03386575e-02
-1.29901528e+00 -3.09957325e-01 1.24010257e-01 -4.25652206e-01
1.50807559e+00 4.46817949e-02 1.01515979e-01 1.66874254e+00
4.25106287e-01 7.01802790e-01 1.27428257e+00 -7.94503450e-01
-7.24413320e-02 4.09702063e-01 -3.88376892e-01 5.43788433e-01
-1.93407744e-01 1.47078022e-01 -3.18004608e-01 -2.80432552e-01
6.29497468e-01 -4.47504342e-01 -1.67742670e-01 2.26829246e-01
-1.82286417e+00 8.14065218e-01 8.92070308e-02 6.56765938e-01
-1.45528018e-01 -3.15007389e-01 6.58390343e-01 8.72368991e-01
6.82869971e-01 4.84981209e-01 -1.12660086e+00 -1.61939964e-01
-7.23926783e-01 -2.55046755e-01 7.54216731e-01 1.35628736e+00
2.10447833e-01 -2.32681453e-01 -1.73993215e-01 1.18129587e+00
-1.87306881e-01 7.17888713e-01 8.70083332e-01 -4.43157136e-01
1.18842852e+00 3.65174711e-01 -2.21210703e-01 -4.77471538e-02
-5.19915402e-01 -5.77245772e-01 -9.98687744e-01 -4.80663538e-01
4.07791406e-01 -6.84830546e-01 -1.04556918e+00 1.99595225e+00
-3.77825350e-02 -8.94166455e-02 6.93916678e-01 5.29118538e-01
8.15483570e-01 4.61319536e-01 4.46445227e-01 -2.01519355e-01
1.36714685e+00 -1.11316562e+00 -5.71691692e-01 -3.23050976e-01
1.35281980e+00 -1.10227001e+00 1.08013129e+00 1.69222355e-01
-1.11364365e+00 -4.60196614e-01 -5.70894361e-01 -1.16907373e-01
-4.10197616e-01 4.38060820e-01 2.84949005e-01 3.19377393e-01
-1.26623440e+00 3.61277014e-01 -9.37029541e-01 -7.92010546e-01
2.81901062e-01 4.14656997e-01 -7.00702786e-01 -3.90721977e-01
-1.72370708e+00 1.31581116e+00 5.90427160e-01 -1.53524533e-01
-9.11573827e-01 -7.86521196e-01 -6.78559244e-01 -5.51402569e-01
-4.68595214e-02 -1.11149013e+00 1.36654997e+00 -9.46091831e-01
-1.52285945e+00 1.47330892e+00 9.02499165e-03 -3.98025781e-01
7.72946656e-01 -9.07754526e-02 -9.25394237e-01 -4.21647578e-02
2.88654268e-01 9.25288022e-01 3.60872805e-01 -4.44971830e-01
-4.44914907e-01 -1.95266902e-01 -2.28999317e-01 2.55343884e-01
-3.05958062e-01 3.00517440e-01 -2.68407494e-01 -9.25907373e-01
-3.35846722e-01 -1.07206917e+00 -3.69646847e-01 -4.82147396e-01
-4.41927016e-01 -8.45540389e-02 -1.17754854e-01 -8.90823364e-01
1.08255541e+00 -1.83587134e+00 2.32654393e-01 -8.72861594e-02
-2.07480997e-01 4.93118644e-01 -5.00259399e-01 8.19983780e-01
-3.41388851e-01 2.28886664e-01 -3.22140753e-01 -1.65770099e-01
-2.77837515e-01 4.77881342e-01 -9.83349904e-02 2.79519409e-01
4.63757485e-01 1.28725350e+00 -1.07970178e+00 -7.08096087e-01
-1.69438839e-01 2.14957640e-01 -5.67428410e-01 1.08064167e-01
-1.69511825e-01 7.96225488e-01 -4.58822727e-01 5.88328302e-01
4.08244818e-01 -2.68720269e-01 3.69924039e-01 9.93318111e-02
2.88931370e-01 4.60765481e-01 -1.83694586e-01 2.19664359e+00
-6.45418882e-01 1.88669667e-01 -1.66299537e-01 -9.32571590e-01
6.12054408e-01 8.17235529e-01 5.11008263e-01 -8.66121054e-01
-6.53249538e-03 9.26963151e-01 3.38568717e-01 -6.31345391e-01
5.76235950e-02 -7.58213103e-01 -1.37870491e-01 2.84325123e-01
4.04824972e-01 -6.82562962e-02 2.38230173e-02 -8.72496143e-02
1.12898719e+00 2.15219036e-01 4.32493418e-01 -2.73763657e-01
7.10548818e-01 5.08151233e-01 4.96895194e-01 4.36296761e-01
-8.45161825e-02 5.41873574e-01 6.84734955e-02 -2.56767035e-01
-1.13169277e+00 -1.10926378e+00 -4.62516963e-01 9.86407518e-01
-6.11119509e-01 -2.39201486e-01 -8.14377725e-01 -1.00770223e+00
-4.05161023e-01 7.65388787e-01 -4.96968001e-01 -2.00111166e-01
-9.16460514e-01 -9.68971550e-01 1.33409905e+00 3.93402517e-01
2.86346942e-01 -1.15513074e+00 1.18917786e-01 5.53903341e-01
-5.75858593e-01 -1.46908736e+00 -6.54779971e-01 3.05210054e-01
-1.10888278e+00 -7.90972233e-01 -1.22543705e+00 -1.23888719e+00
6.18994594e-01 -5.72536349e-01 1.32243490e+00 -5.59790194e-01
-2.35567149e-02 -3.59270023e-03 -3.91972423e-01 -4.73815858e-01
-1.11406648e+00 6.71461642e-01 7.87850246e-02 -5.62682629e-01
8.55132878e-01 -3.74178290e-01 -3.01499575e-01 4.00756091e-01
-7.85718620e-01 4.06232446e-01 1.11338484e+00 1.12688828e+00
6.86001122e-01 -6.14388227e-01 8.83104622e-01 -1.18161881e+00
6.94563568e-01 -6.56338871e-01 5.62598035e-02 5.54696262e-01
-3.61239105e-01 1.11049816e-01 8.32123697e-01 -6.64868891e-01
-8.32833350e-01 -2.37278059e-01 -5.71449935e-01 -1.24216609e-01
-3.57696533e-01 5.93488038e-01 6.15942329e-02 4.22139376e-01
1.04092789e+00 2.84462869e-01 -1.01701878e-01 -6.52668595e-01
4.78100330e-01 1.00214827e+00 5.15263021e-01 -7.42609739e-01
6.37382329e-01 2.05377251e-01 -3.42923552e-01 -4.52328146e-01
-6.27462983e-01 -4.99123901e-01 -7.90300310e-01 5.89786470e-01
1.14738655e+00 -1.18520427e+00 2.43657064e-02 3.04700434e-01
-1.12620664e+00 -6.86511874e-01 -1.51662320e-01 9.86062288e-01
-6.61744654e-01 1.03903197e-01 -1.18817115e+00 9.60354134e-02
-7.73874044e-01 -1.16945958e+00 1.20788229e+00 -6.49633110e-01
-4.45497066e-01 -1.33526659e+00 4.97051388e-01 4.10821676e-01
3.33285183e-01 3.95157258e-05 1.32167256e+00 -8.91887784e-01
2.61922359e-01 3.57263535e-02 -1.39933512e-01 4.87534672e-01
3.85757715e-01 -7.64323294e-01 -5.07382691e-01 -6.03512287e-01
-4.15634066e-02 -6.08643413e-01 5.95369995e-01 7.55226985e-02
4.76869136e-01 -2.53813952e-01 -5.14678299e-01 6.74661279e-01
1.24881959e+00 1.11566119e-01 5.52510083e-01 4.10150290e-01
4.68623787e-01 3.59139532e-01 6.64950013e-01 -2.19230935e-01
4.66506273e-01 7.36574173e-01 -2.04880744e-01 -7.85285950e-01
-2.02320784e-01 -1.71008632e-01 6.96987331e-01 1.39948285e+00
-5.39185219e-02 -9.08260867e-02 -1.18174052e+00 7.74060249e-01
-1.55496478e+00 -5.22555530e-01 3.56282964e-02 1.85073233e+00
1.63313127e+00 4.76507284e-02 -1.84982881e-01 -6.71484590e-01
4.34196919e-01 -5.48061728e-01 -4.74347949e-01 -8.75051498e-01
-1.58058509e-01 7.08087504e-01 4.71115768e-01 4.65205580e-01
-1.01748049e+00 1.28253913e+00 5.54381609e+00 1.25199151e+00
-1.19291627e+00 6.38512194e-01 6.47173524e-01 -7.01536052e-03
-2.13849291e-01 -2.04204187e-01 -8.99104834e-01 3.18293601e-01
1.55457020e+00 4.99662384e-02 1.42365813e-01 5.84402263e-01
2.07265273e-01 5.88471591e-01 -1.48439920e+00 7.07959473e-01
-6.05017319e-02 -1.22181928e+00 1.59352526e-01 1.77298099e-01
8.60826075e-01 7.11654663e-01 4.87674363e-02 8.14136744e-01
5.70364773e-01 -1.23177528e+00 3.05640310e-01 1.30955681e-01
1.64114892e+00 -3.52918565e-01 1.03845251e+00 6.10305011e-01
-6.21434987e-01 3.23137313e-01 -3.82689834e-01 3.94741386e-01
3.08540016e-01 4.12448376e-01 -1.48556447e+00 9.94100571e-01
3.48939419e-01 5.48347771e-01 -3.50812525e-01 3.94176066e-01
-2.31253043e-01 8.51206958e-01 -2.35677287e-01 2.25525185e-01
6.19888783e-01 1.05607569e-01 3.96590322e-01 1.73094082e+00
3.77056777e-01 -7.90957659e-02 2.68714726e-01 5.42440414e-01
-2.73433030e-01 6.58596158e-01 -5.44536293e-01 -2.47276518e-02
-5.35792597e-02 8.78494799e-01 -2.52002805e-01 -4.75902408e-01
-5.17505825e-01 1.12751639e+00 4.15694684e-01 2.88214177e-01
-8.66747737e-01 -2.01834813e-02 4.53957528e-01 2.10422128e-01
-4.61449064e-02 -3.90606336e-02 4.58238795e-02 -1.41121900e+00
-1.14709117e-01 -1.48744774e+00 7.09568739e-01 -5.53514242e-01
-1.63660777e+00 1.15515411e+00 9.60321352e-03 -1.52494597e+00
-7.41080523e-01 -8.44622254e-01 -3.58724669e-02 1.18473792e+00
-1.61518800e+00 -1.71485364e+00 5.95387936e-01 8.81284714e-01
7.33959258e-01 -5.49385786e-01 1.29763031e+00 7.10807145e-01
-1.98087901e-01 1.03528523e+00 3.85604024e-01 4.03145105e-01
1.43557119e+00 -9.38106239e-01 4.83630091e-01 4.27093774e-01
-7.77083263e-02 7.85119295e-01 1.22037806e-01 -7.41269171e-01
-9.32367623e-01 -1.42985165e+00 1.47724032e+00 -7.26750195e-01
7.64485002e-01 -4.01559472e-01 -8.27927768e-01 9.35878813e-01
3.39825720e-01 -4.35211927e-01 1.01117325e+00 7.55443275e-02
-1.80932671e-01 3.90289456e-01 -1.07503557e+00 6.29647493e-01
1.08061624e+00 -7.09782541e-01 -9.78851974e-01 8.77399921e-01
9.27799642e-01 -6.28228605e-01 -1.36045289e+00 6.73841476e-01
2.36422524e-01 1.99492015e-02 9.95530188e-01 -1.32555294e+00
7.11411357e-01 7.76795372e-02 -1.33929923e-01 -1.41997457e+00
-1.41025260e-01 -6.15574658e-01 5.44951379e-01 9.13506031e-01
1.13077962e+00 -7.07425892e-01 3.45679671e-01 -5.37845381e-02
-5.00523150e-01 -9.03331995e-01 -1.14269376e+00 -8.96345317e-01
1.18297672e+00 -4.73803967e-01 3.20770115e-01 1.42763472e+00
3.00777584e-01 5.87241292e-01 -2.03534067e-01 -1.81322977e-01
7.10017085e-02 -2.42390051e-01 5.54122925e-01 -6.54217660e-01
-6.33007646e-01 -1.29238591e-01 -2.87527293e-02 -7.31927395e-01
1.20945297e-01 -1.63999367e+00 -2.91626245e-01 -1.53394663e+00
3.47852945e-01 -5.07319510e-01 -3.11858833e-01 9.74990487e-01
-2.05449581e-01 2.72409320e-01 -2.78514445e-01 3.72098953e-01
-1.92339748e-01 1.32406354e-01 1.38124573e+00 -2.33703151e-01
-2.06506804e-01 -9.75758061e-02 -6.18475795e-01 5.79818666e-01
7.32145727e-01 -7.27697611e-01 -2.89948344e-01 -8.91880989e-01
1.82913207e-02 1.42363027e-01 -6.66173249e-02 -5.41409254e-01
4.25336473e-02 -1.60747245e-01 1.60718679e-01 -3.22204649e-01
3.27079445e-02 -5.63515186e-01 2.60257930e-01 6.74105108e-01
-5.15017509e-01 2.18360305e-01 5.88037610e-01 8.86488259e-02
-2.59837449e-01 -1.13192983e-01 6.67822123e-01 -4.70154732e-01
-3.91471684e-01 2.18164966e-01 -3.04387301e-01 4.76592749e-01
6.82956040e-01 5.57233505e-02 -3.63451540e-01 -1.00557618e-01
-1.12563550e+00 3.06826949e-01 1.93950191e-01 5.33268929e-01
3.00761402e-01 -1.16524470e+00 -1.49334478e+00 8.89040828e-02
3.80929291e-01 -3.06781828e-01 1.34979367e-01 1.22451055e+00
-5.05504429e-01 8.64192724e-01 -1.27261981e-01 -7.65263081e-01
-9.90342379e-01 5.95259011e-01 3.15837979e-01 -1.13523042e+00
-5.72407007e-01 6.94066942e-01 3.23107243e-01 -1.21543300e+00
-2.47188985e-01 -4.30332243e-01 2.02379152e-01 -3.25275838e-01
1.06995292e-01 -9.69309285e-02 5.56062102e-01 -5.76727450e-01
-4.90077645e-01 6.32481456e-01 -4.45697844e-01 -3.72628421e-01
1.30670786e+00 5.37718683e-02 -9.24233198e-02 7.43664131e-02
1.43814635e+00 -7.64525011e-02 -2.77015090e-01 -4.68539238e-01
2.21806452e-01 3.21071446e-01 -5.12641609e-01 -1.38781679e+00
-6.06567562e-01 1.04818022e+00 5.68176866e-01 -8.99829149e-01
1.09704173e+00 1.57602102e-01 1.06596136e+00 4.70342427e-01
5.97813010e-01 -1.08279312e+00 -1.89754799e-01 7.01818883e-01
8.10137630e-01 -1.18178797e+00 -4.14951384e-01 -1.34469762e-01
-7.57896245e-01 9.30692911e-01 3.56314659e-01 2.97262073e-01
4.44831938e-01 2.16246501e-01 4.23210919e-01 5.41511662e-02
-8.66679549e-01 -6.12319820e-02 4.21894491e-01 6.82123780e-01
7.85022974e-01 2.48072729e-01 -6.48411691e-01 5.76579630e-01
-4.31297362e-01 4.09265369e-01 9.94348675e-02 7.61916280e-01
2.33302653e-01 -1.71228540e+00 -7.56309703e-02 3.13496411e-01
-1.01197922e+00 -6.93153501e-01 -4.81242776e-01 1.12056005e+00
4.03210849e-01 7.76070118e-01 -3.85107905e-01 -2.77838588e-01
4.34665173e-01 3.81417871e-01 4.51623350e-01 -9.89760816e-01
-8.55950415e-01 2.86414653e-01 4.30417985e-01 -2.12192312e-01
-1.81119993e-01 -5.60434639e-01 -1.27431059e+00 7.75606260e-02
-8.28855559e-02 2.95093387e-01 3.47355992e-01 9.36725497e-01
4.01204914e-01 6.33555889e-01 8.98484215e-02 2.60809511e-02
-6.76155269e-01 -1.40939713e+00 1.36509329e-01 4.37224984e-01
5.77228609e-03 -2.78955996e-02 8.50523636e-02 1.81271702e-01] | [11.422256469726562, 10.308087348937988] |
349e4d65-fa8f-4077-be62-2da7eb5c7a8a | video-quality-assessment-for-computer | null | null | https://www.researchgate.net/publication/220183765_Video_Quality_Assessment_for_Computer_Graphics_Applications | https://www.researchgate.net/publication/220183765_Video_Quality_Assessment_for_Computer_Graphics_Applications | Video Quality Assessment for Computer Graphics Applications | Numerous current Computer Graphics methods produce video sequences as their outcome. The merit of these methods is often judged by assessing the quality of a set of results through lengthy user studies. We present a full-reference video quality metric geared specifically towards the requirements of Computer Graphics applications as a faster computational alternative to subjective evaluation. Our metric can compare a video pair with arbitrary dynamic ranges, and comprises a human visual system model for a wide range of luminance levels, that predicts distortion visibility through models of luminance adaptation, spatiotemporal contrast sensitivity and visual masking. We present applications of the proposed metric to quality prediction of HDR video compression and temporal tone mapping, comparison of different rendering approaches and qualities, and assessing the impact of variable frame rate to perceived quality. | ['Hans-Peter Seidel', 'Karol Myszkowski', 'Martin Cadik', 'Tunc Ozan Aydin'] | 2010-12-01 | null | null | null | acm-transactions-on-graphics-2010-12 | ['video-quality-assessment', 'video-compression', 'tone-mapping', 'video-quality-assessment'] | ['computer-vision', 'computer-vision', 'computer-vision', 'time-series'] | [ 4.65775877e-01 -5.83313882e-01 1.57889664e-01 -3.66378248e-01
-4.80873168e-01 -3.46290946e-01 6.59133196e-01 -3.60156082e-05
-2.51932502e-01 4.71002162e-01 1.54109478e-01 -4.92853343e-01
3.42381448e-02 -6.37612879e-01 -2.92283952e-01 -1.82537600e-01
-3.73868644e-01 -3.19406718e-01 7.53580153e-01 -3.48900735e-01
6.16567492e-01 5.03533900e-01 -2.01928258e+00 4.27825123e-01
6.55671418e-01 1.33847594e+00 2.46079206e-01 1.33023500e+00
4.35176581e-01 9.31834579e-01 -7.73850620e-01 -3.65743518e-01
2.72392213e-01 -6.47192895e-01 -2.71873564e-01 2.59764314e-01
4.06869471e-01 -5.42266548e-01 -4.27008092e-01 9.85884190e-01
5.63125074e-01 -1.45860165e-02 5.60692012e-01 -1.14037395e+00
-6.80759847e-01 -1.80676803e-01 -2.72555768e-01 7.16387987e-01
1.04616535e+00 4.12978500e-01 6.56137645e-01 -6.67135954e-01
3.52421135e-01 1.00993717e+00 4.68036890e-01 2.03448579e-01
-1.24425304e+00 -4.06600952e-01 -3.25300395e-01 4.96063232e-01
-1.40401971e+00 -5.51081240e-01 4.15858209e-01 -5.57252824e-01
1.06501198e+00 7.08219051e-01 6.77619457e-01 4.92303014e-01
4.57779914e-01 3.08412388e-02 9.91270363e-01 -2.77816504e-01
3.31038982e-01 6.48968969e-04 -3.38861644e-01 3.41011226e-01
-1.46564946e-03 5.43091238e-01 -6.14668548e-01 -1.50052786e-01
1.21782184e+00 -7.45754778e-01 -6.66905522e-01 -2.22170681e-01
-8.18951547e-01 2.54876047e-01 -1.57052293e-01 1.53189152e-01
-1.53397828e-01 3.77486832e-02 4.02405858e-01 5.31825900e-01
2.15116441e-01 1.22200117e-01 -4.03497852e-02 -4.84241039e-01
-9.60779071e-01 3.17313552e-01 5.54745913e-01 1.03379679e+00
3.40094179e-01 4.57054764e-01 -1.67988330e-01 8.00265372e-01
4.14506286e-01 4.73932803e-01 3.81157041e-01 -1.37381673e+00
1.19994476e-01 -6.81936517e-02 3.35418999e-01 -1.07970989e+00
-9.79435146e-02 2.03252807e-01 -5.05210102e-01 1.03425181e+00
2.59264648e-01 3.48227829e-01 -4.27508771e-01 1.26329648e+00
-1.75547659e-01 9.38345566e-02 -2.05134168e-01 1.09491765e+00
6.31657600e-01 7.34470189e-01 6.17936403e-02 -6.19784594e-01
1.03948665e+00 -3.19199622e-01 -9.00157928e-01 2.66758591e-01
1.13872103e-01 -1.05362940e+00 1.16457784e+00 9.01511371e-01
-1.62732840e+00 -1.06264198e+00 -1.39401412e+00 -1.16807684e-01
1.32444441e-01 -4.64825511e-01 7.21378103e-02 1.29413664e+00
-1.64716208e+00 7.04947233e-01 -2.97303468e-01 8.15726966e-02
-1.48705274e-01 3.03535104e-01 1.16872966e-01 5.24385758e-02
-1.08787501e+00 1.00993025e+00 5.13456995e-03 -2.91022152e-01
-5.95239341e-01 -6.71129882e-01 -7.26190686e-01 1.33691013e-01
-7.26662576e-02 -4.61490810e-01 1.29000008e+00 -1.17113304e+00
-1.64368212e+00 7.40691900e-01 -4.95585576e-02 -3.70534301e-01
6.32315695e-01 -5.29100709e-02 -1.01244378e+00 2.69919574e-01
-4.18624431e-01 2.83029705e-01 8.86077881e-01 -1.33824170e+00
-7.76377559e-01 -2.73862015e-02 5.77993179e-03 3.65340889e-01
2.80663341e-01 3.66106063e-01 -6.45881534e-01 -6.46691442e-01
-1.18149467e-01 -2.70122617e-01 -2.60712150e-02 1.66014507e-01
3.55633736e-01 3.55591744e-01 7.50841498e-01 -7.68500328e-01
1.61557007e+00 -2.10961413e+00 -1.82036519e-01 2.41445407e-01
1.23592913e-01 2.86699057e-01 -3.93673815e-02 8.88371244e-02
-1.23664394e-01 1.42302901e-01 3.66786830e-02 7.26255029e-02
2.40801703e-02 -3.35725248e-01 -1.31610304e-01 3.66363108e-01
-8.08530077e-02 2.78752893e-01 -7.00351596e-01 -4.38536257e-01
6.95876241e-01 6.96439207e-01 -5.94592333e-01 4.64123577e-01
1.05813146e-01 1.08416818e-01 3.11885595e-01 4.96575266e-01
7.03957856e-01 1.80414811e-01 1.12856910e-01 -4.66282696e-01
-3.78765225e-01 1.29634529e-01 -1.20084476e+00 1.11958492e+00
-3.15585971e-01 1.26408148e+00 -2.34115198e-01 -1.48932174e-01
9.66120362e-01 5.25002420e-01 2.12791771e-01 -1.34041619e+00
1.91289276e-01 1.96066692e-01 1.97742030e-01 -6.39277697e-01
8.48252535e-01 1.43813659e-02 4.41495955e-01 2.09780782e-01
-2.17707813e-01 -3.88154685e-01 2.32270107e-01 1.61606874e-02
9.65760529e-01 3.29478890e-01 5.12416840e-01 -2.75939524e-01
5.78784704e-01 -6.00048423e-01 2.03180999e-01 5.81522226e-01
-5.29310644e-01 1.00486112e+00 4.73028094e-01 -3.49910051e-01
-1.66570330e+00 -1.21049368e+00 -2.44006976e-01 1.02058303e+00
4.62868005e-01 -4.18465644e-01 -6.78304970e-01 2.68267125e-01
-4.45003420e-01 7.26061583e-01 -2.91056037e-01 -2.00267434e-02
-5.16831994e-01 -4.34461504e-01 3.02412957e-01 2.12892652e-01
3.76619160e-01 -8.65962088e-01 -1.13370383e+00 1.38952047e-01
1.67909525e-02 -1.22040224e+00 -3.93889517e-01 -4.37318444e-01
-6.89708471e-01 -9.64598715e-01 -5.76181829e-01 -5.12260497e-01
-6.31078854e-02 2.56415904e-01 1.48931444e+00 3.64672720e-01
-4.27944362e-01 5.29997647e-01 -4.29003865e-01 1.88963979e-01
-7.34248042e-01 -9.58863318e-01 2.61141136e-02 -1.50313556e-01
7.35537857e-02 -4.97867674e-01 -9.47341800e-01 7.20952749e-01
-1.08609879e+00 1.58135459e-01 -8.36321786e-02 2.77464479e-01
3.07491750e-01 1.26043543e-01 -4.80268802e-03 -3.92010123e-01
7.89536357e-01 -1.20078556e-01 -6.92575574e-01 1.19642191e-01
-4.93801445e-01 -3.83909196e-01 4.87062842e-01 -2.59572595e-01
-1.22310686e+00 -5.92735529e-01 -1.20444946e-01 -1.21404111e-01
-9.31010395e-02 -3.83689031e-02 -1.69121146e-01 -2.99363226e-01
8.40008020e-01 2.35251173e-01 -2.50332803e-01 -1.12049080e-01
1.28465727e-01 5.44003546e-01 9.21631277e-01 -3.28665167e-01
4.40740764e-01 1.79240286e-01 -6.76034018e-02 -8.97998393e-01
2.47829214e-01 -1.80572152e-01 -3.48255843e-01 -8.73591065e-01
6.59812689e-01 -7.02119708e-01 -8.27822328e-01 6.11528039e-01
-1.02676427e+00 -3.33602548e-01 -3.15706208e-02 6.01274848e-01
-9.42462981e-01 6.23849750e-01 -6.30812764e-01 -8.98501396e-01
1.46431267e-01 -1.37271738e+00 8.06733489e-01 1.34279549e-01
-1.69483453e-01 -9.62506592e-01 1.65616944e-01 -1.49141336e-02
6.48916245e-01 1.69881776e-01 8.46820951e-01 3.23013812e-01
-6.86710179e-01 1.23478010e-01 -4.53017980e-01 3.77007663e-01
1.15249068e-01 4.78172451e-01 -1.03227162e+00 -2.23307073e-01
2.51891851e-01 1.54922502e-02 2.37151429e-01 6.20542884e-01
1.14994788e+00 1.83880813e-02 2.93134719e-01 7.35563934e-01
1.85996222e+00 9.11474049e-01 1.35325491e+00 3.14055622e-01
3.42796534e-01 4.25647646e-01 6.85363948e-01 7.26070642e-01
-7.04957098e-02 9.28443253e-01 3.00810397e-01 -1.97215259e-01
-2.95224279e-01 1.64947167e-01 1.83291599e-01 6.52547419e-01
-5.76014102e-01 -5.15361786e-01 -7.61404395e-01 -2.93213930e-02
-1.22392035e+00 -1.03311384e+00 -1.91717252e-01 2.41626549e+00
4.35943753e-01 4.07559335e-01 3.37396145e-01 5.08427739e-01
9.71136749e-01 6.14920557e-02 -1.77434832e-01 -9.15331423e-01
-1.87515184e-01 1.83518633e-01 4.59194720e-01 7.31335819e-01
-5.24591148e-01 3.75603557e-01 8.36368370e+00 7.67792761e-01
-9.74543095e-01 -2.23428175e-01 8.75816405e-01 -1.06895596e-01
-3.78388464e-01 -1.32647902e-01 -1.84383318e-01 5.24905920e-01
1.14328444e+00 -4.20886725e-01 4.94987547e-01 4.70756263e-01
7.63364851e-01 -4.54733938e-01 -1.24143517e+00 1.29754031e+00
2.69111037e-01 -1.09444880e+00 2.16811031e-01 -1.95773393e-02
4.29250032e-01 -5.86964369e-01 5.62966287e-01 -1.68051288e-01
-2.66632587e-01 -1.22165394e+00 1.04635572e+00 3.64385724e-01
1.11606050e+00 -6.64934635e-01 3.68615836e-01 -2.68991500e-01
-1.23817861e+00 6.78400919e-02 -2.71831185e-01 -2.71879882e-01
4.99003768e-01 -7.71318376e-02 -2.24675089e-01 5.05113713e-02
7.68705547e-01 1.53301388e-01 -5.96835434e-01 1.51803160e+00
3.65878254e-01 2.15729162e-01 5.51595502e-02 2.21850887e-01
-2.54378170e-01 -2.21545324e-01 4.79080617e-01 1.41539323e+00
5.45786381e-01 4.04948384e-01 -5.18446565e-01 4.88289952e-01
5.63250422e-01 2.87272722e-01 -4.50354904e-01 3.85463268e-01
3.75393152e-01 5.71387589e-01 -5.23375452e-01 -3.26966316e-01
-6.39815152e-01 1.03373849e+00 -3.54282916e-01 4.50603724e-01
-9.66490686e-01 -1.73354432e-01 6.23121142e-01 4.25508857e-01
1.88592032e-01 -2.84068018e-01 -5.08092046e-01 -7.62335300e-01
-1.10816024e-01 -1.13427079e+00 2.59873690e-03 -1.31915033e+00
-6.75233781e-01 7.81366527e-01 8.88672844e-02 -1.62367463e+00
-2.68353671e-01 -7.41757154e-01 -4.35801864e-01 9.43258166e-01
-1.13857818e+00 -2.95796454e-01 -4.42310780e-01 5.43624997e-01
6.47585034e-01 -3.41916263e-01 5.19609809e-01 5.88011682e-01
3.22085135e-02 5.54983974e-01 -1.83350116e-01 -5.35500228e-01
4.72557724e-01 -1.11113691e+00 5.32972574e-01 9.60813820e-01
-3.47223490e-01 9.15374979e-02 1.24431944e+00 -3.73426497e-01
-1.02193189e+00 -4.06147897e-01 4.88780528e-01 -1.78967133e-01
3.53541881e-01 5.68202995e-02 -1.00394952e+00 5.47322072e-02
2.88490206e-01 -2.75321513e-01 6.38024449e-01 -4.22925264e-01
-2.96991795e-01 7.16011003e-02 -1.15606964e+00 7.33934343e-01
9.15822268e-01 -5.72067797e-01 -1.05333045e-01 -3.09848338e-01
4.49231565e-01 -3.89211476e-01 -1.01691997e+00 4.04582351e-01
8.03830326e-01 -1.87408566e+00 8.96214724e-01 9.64629799e-02
4.89488333e-01 -5.13116479e-01 -5.25967598e-01 -8.77067626e-01
-5.62504053e-01 -7.01953173e-01 1.76179469e-01 7.95220315e-01
2.20951185e-01 -9.77312867e-03 4.04044092e-01 7.49175251e-01
2.87954789e-02 -1.54152751e-01 -8.43999028e-01 -6.15686476e-01
-5.84546328e-01 -7.37554848e-01 4.96447057e-01 4.44793642e-01
7.39601254e-02 -4.45957519e-02 -7.56212950e-01 4.05983105e-02
6.36856973e-01 -4.84731793e-01 5.60226560e-01 -7.62592316e-01
-5.54379761e-01 -7.11632490e-01 -1.03797150e+00 -8.98041487e-01
-6.51058555e-01 7.28796795e-02 -7.50061646e-02 -9.27108943e-01
3.57579440e-02 -1.18918531e-01 -1.12375595e-01 -6.71656251e-01
-3.10006458e-02 7.53195941e-01 3.33282620e-01 4.93474156e-02
-3.76631051e-01 1.54564053e-01 1.06766367e+00 3.36860687e-01
-2.01791167e-01 -9.47706774e-02 -2.32405975e-01 7.44734108e-01
5.37618518e-01 1.43724427e-01 -6.10053778e-01 -3.47392142e-01
3.84998381e-01 7.41753817e-01 3.75134051e-01 -1.43765008e+00
-1.90537553e-02 -1.17909603e-01 4.81169850e-01 -2.59712130e-01
3.42209160e-01 -8.36893320e-01 6.17082715e-01 3.90952498e-01
-3.93224955e-01 5.11109352e-01 1.59274101e-01 3.47122133e-01
-3.32574308e-01 -1.62525594e-01 1.08312809e+00 9.33390707e-02
-1.18820667e+00 4.75826524e-02 -5.70752978e-01 -1.26943499e-01
8.25523138e-01 -8.87133896e-01 -1.27315894e-01 -7.29091883e-01
-7.96618164e-01 -4.15361702e-01 8.59998643e-01 3.10410917e-01
9.74799156e-01 -1.37973309e+00 -6.61649644e-01 3.48443210e-01
-1.11220926e-01 -9.81144667e-01 4.05767083e-01 2.73353904e-01
-1.22944570e+00 7.99863189e-02 -7.17872739e-01 -6.49628520e-01
-1.62265754e+00 6.82824850e-01 5.56852162e-01 3.13556939e-01
-2.97758877e-01 4.28396046e-01 3.93435001e-01 7.76346624e-01
2.19403923e-01 -2.70067245e-01 -3.61743748e-01 -4.66861248e-01
8.74609232e-01 7.44135737e-01 9.19344053e-02 -8.98206890e-01
-1.36264777e-02 6.48741364e-01 4.00360823e-01 -5.51082730e-01
5.29876947e-01 -7.33548880e-01 3.14843655e-01 5.48804045e-01
1.13486671e+00 -1.55818179e-01 -1.49767804e+00 9.98842642e-02
-3.22278529e-01 -1.13468552e+00 4.46471125e-02 -6.20697141e-01
-7.38353550e-01 6.20565414e-01 1.35592592e+00 5.04713714e-01
1.66004443e+00 -5.15915096e-01 3.34928215e-01 -3.18381131e-01
3.35808992e-01 -1.11948872e+00 1.93969861e-01 1.62360743e-01
8.29961598e-01 -1.03422987e+00 1.30433366e-02 -6.64706349e-01
-6.36459529e-01 1.19095755e+00 4.81564403e-01 -6.57699928e-02
5.90765417e-01 5.38707256e-01 2.97922641e-01 1.62783086e-01
-8.29377651e-01 -6.85757445e-03 4.36034530e-01 9.85330701e-01
7.84967005e-01 -1.58947602e-01 -4.74587560e-01 -1.29763335e-01
-3.21366936e-01 1.64421931e-01 7.00033188e-01 6.56382501e-01
-7.18526244e-01 -7.76597559e-01 -4.27126467e-01 1.64753929e-01
-6.96167111e-01 -1.93220854e-01 2.32341588e-01 5.95794082e-01
1.53735578e-01 1.04276586e+00 3.96136701e-01 -4.91241455e-01
5.38220108e-01 -4.77290660e-01 7.59789526e-01 -7.80389383e-02
-4.38959926e-01 3.32888037e-01 8.33631828e-02 -7.49754131e-01
-5.80982566e-01 -4.88199234e-01 -6.47259057e-01 -6.12135947e-01
7.03867823e-02 -2.09644839e-01 5.92441022e-01 5.58659077e-01
-1.20012283e-01 5.22411346e-01 5.53861916e-01 -9.44449842e-01
4.74540144e-02 -6.02292061e-01 -8.02648008e-01 6.69562340e-01
4.14003670e-01 -4.08265203e-01 -3.76427144e-01 5.12778521e-01] | [11.6061372756958, -1.9652308225631714] |
d59df89b-242f-482d-bda1-19e4c251bbe6 | network-traffic-anomaly-detection-method | 2205.03907 | null | https://arxiv.org/abs/2205.03907v1 | https://arxiv.org/pdf/2205.03907v1.pdf | Network Traffic Anomaly Detection Method Based on Multi scale Residual Feature | To address the problem that traditional network traffic anomaly detection algorithms do not suffi-ciently mine potential features in long time domain, an anomaly detection method based on mul-ti-scale residual features of network traffic is proposed. The original traffic is divided into subse-quences of different time spans using sliding windows, and each subsequence is decomposed and reconstructed into data sequences of different levels using wavelet transform technique; the stacked autoencoder (SAE) constructs similar feature space using normal network traffic, and gen-erates reconstructed error vector using the difference between reconstructed samples and input samples in the similar feature space; the multi-path residual group is used to learn reconstructed error The traffic classification is completed by a lightweight classifier. The experimental results show that the detection performance of the proposed method for anomalous network traffic is sig-nificantly improved compared with traditional methods; it confirms that the longer time span and more S transformation scales have positive effects on discovering potential diversity information in the original network traffic. | ['Kun Wang', 'Yu Fu', 'Xueyuan Duan'] | 2022-05-08 | null | null | null | null | ['traffic-classification'] | ['miscellaneous'] | [ 4.57482412e-03 -7.01072276e-01 -3.56242694e-02 -1.16536215e-01
1.10670686e-01 -4.85118739e-02 3.09351802e-01 -7.95661062e-02
-2.16334499e-03 5.72929144e-01 1.02917729e-02 -5.46646595e-01
-2.86703736e-01 -9.96282458e-01 -2.36145422e-01 -7.22948611e-01
-4.97496814e-01 1.53153569e-01 6.66486800e-01 -3.73792440e-01
5.63113809e-01 8.55380833e-01 -2.03861761e+00 3.66103977e-01
7.44267881e-01 1.39836764e+00 -5.29322743e-01 7.11059868e-01
-6.56417072e-01 9.49558377e-01 -9.83800769e-01 -7.10538104e-02
6.43991649e-01 -4.54059422e-01 -4.96773988e-01 4.37542826e-01
2.84082085e-01 -2.64407039e-01 -8.76740098e-01 1.16262817e+00
-1.48476154e-01 4.30631340e-01 3.44978571e-01 -1.75104880e+00
-3.11693192e-01 1.70281172e-01 -6.39556825e-01 1.17069507e+00
6.40000179e-02 -1.28388712e-02 8.54518175e-01 -7.30994701e-01
2.92364091e-01 1.42935312e+00 6.97067797e-01 5.56893423e-02
-1.16522563e+00 -1.01577830e+00 2.38760039e-01 8.18634331e-01
-1.03581977e+00 -8.76475871e-02 1.02564669e+00 -2.33103618e-01
1.01739299e+00 4.45957154e-01 6.18594885e-01 1.08838284e+00
6.51711702e-01 5.25005996e-01 6.54636204e-01 -1.10396259e-01
-7.08358735e-02 -1.76464804e-02 5.70641339e-01 6.19988441e-01
4.64203715e-01 2.47954860e-01 -1.20670669e-01 -6.68295622e-01
2.83151895e-01 7.81515956e-01 7.11705387e-02 3.05071712e-01
-1.08160937e+00 8.82438540e-01 4.12389636e-03 7.65227377e-01
-8.27967703e-01 -2.81688452e-01 9.88471031e-01 1.15186477e+00
4.43195224e-01 3.03908084e-02 -4.53366816e-01 -1.91043153e-01
-8.05659115e-01 1.02120582e-02 6.64057791e-01 4.53167081e-01
8.01580906e-01 1.07163703e+00 2.45989561e-01 5.05401790e-01
1.74627490e-02 4.38668251e-01 9.67524827e-01 -6.24872863e-01
2.91544944e-01 8.95415187e-01 -5.22520363e-01 -1.51542139e+00
-9.97929648e-02 -1.70940042e-01 -7.57152796e-01 3.64079177e-01
2.72964388e-01 -2.02206954e-01 -5.20777106e-01 1.08979416e+00
3.17031354e-01 1.06627798e+00 1.28997535e-01 2.20322847e-01
2.89861470e-01 9.40695107e-01 8.87788832e-02 -4.46643800e-01
1.07244921e+00 -5.63603580e-01 -8.87549877e-01 2.55919129e-01
5.82408786e-01 -6.72583759e-01 5.43791533e-01 3.85537177e-01
-5.17727852e-01 -8.67492855e-01 -1.07981956e+00 7.73781002e-01
-5.63780546e-01 -3.41593683e-01 4.12888557e-01 6.02425098e-01
-5.40075302e-01 7.09817767e-01 -5.36214173e-01 -4.12117749e-01
3.32775354e-01 2.31949538e-01 -4.75308627e-01 3.02242488e-01
-1.09294868e+00 3.72148633e-01 6.73699498e-01 -1.32620484e-01
-3.36673200e-01 -5.72702408e-01 -6.97880208e-01 2.94757653e-02
3.12537462e-01 1.22567624e-01 5.87678909e-01 -1.21005845e+00
-9.14215565e-01 2.17980951e-01 -1.88151523e-01 -6.39694273e-01
8.97540078e-02 3.61055225e-01 -1.90656364e+00 -2.04164926e-02
1.21475197e-01 -4.04072434e-01 1.32390344e+00 -7.40728140e-01
-1.27804184e+00 -4.63415980e-01 -6.74743652e-01 -5.81410110e-01
-6.90341771e-01 1.47308797e-01 4.66825753e-01 -9.49833989e-01
4.61886913e-01 -5.45697212e-01 -2.14685090e-02 -5.17500639e-01
-2.75010735e-01 -4.43502992e-01 2.08003759e+00 -7.34562516e-01
1.71619654e+00 -2.29968238e+00 -5.65745771e-01 9.08818364e-01
1.80786416e-01 4.69877392e-01 -1.07397772e-01 2.76303202e-01
-6.13896906e-01 -5.15305251e-02 -9.41433385e-02 5.83637893e-01
-4.44366157e-01 4.53119397e-01 -7.08207786e-01 3.49349320e-01
1.62295282e-01 3.08723062e-01 -5.93169808e-01 -3.22008669e-01
3.13307613e-01 3.53406705e-02 -3.12361181e-01 2.01018050e-01
2.91892737e-01 2.97015429e-01 -6.97099745e-01 7.00528443e-01
6.42136157e-01 1.53063461e-01 -3.61860275e-01 -2.09107131e-01
-1.79962486e-01 -1.38884187e-01 -1.49498749e+00 5.86878359e-01
-9.20523256e-02 8.74341190e-01 -2.98658341e-01 -1.54960716e+00
1.43131876e+00 3.49213332e-01 8.74787748e-01 -7.78140545e-01
1.39513060e-01 2.71352559e-01 1.64567173e-01 -8.44471037e-01
2.55153298e-01 1.55709669e-01 1.92312628e-01 6.59011483e-01
5.17790392e-03 8.34118724e-01 2.82434344e-01 -1.63046028e-02
1.45272088e+00 -5.64289749e-01 1.78646371e-01 -1.16818376e-01
1.40811753e+00 -1.36933759e-01 8.68299961e-01 5.70533395e-01
-4.77779329e-01 -3.40682328e-01 5.62648714e-01 -1.07440674e+00
-1.12385964e+00 -1.05046999e+00 -1.49691507e-01 1.21641564e+00
-1.83611095e-01 -8.38313624e-02 -1.93775579e-01 -1.15889633e+00
3.76290083e-01 9.48095798e-01 -5.26148677e-01 -4.56949800e-01
-9.90942657e-01 -6.45795524e-01 8.18006754e-01 3.78416598e-01
7.32662380e-01 -1.15729821e+00 -1.02968417e-01 4.69625175e-01
1.60049975e-01 -9.74202394e-01 -2.41220042e-01 -1.34949416e-01
-1.14442718e+00 -1.23153806e+00 2.79721856e-01 -7.78748333e-01
5.63671589e-01 4.95121479e-01 7.34076858e-01 3.63041371e-01
-4.68238115e-01 1.36562556e-01 -3.39943022e-01 -3.39775652e-01
-6.64932609e-01 -4.15934741e-01 4.89674270e-01 7.39251435e-01
1.13843870e+00 -1.10386133e+00 -1.95389554e-01 3.72236252e-01
-8.05524707e-01 -9.24976408e-01 5.51244259e-01 8.07173252e-01
2.64999419e-01 7.63615251e-01 9.10520315e-01 -7.12099433e-01
8.24486554e-01 -9.72913742e-01 -3.22964787e-01 -2.97806263e-01
-7.76562870e-01 -2.59965789e-02 9.72586870e-01 -5.11702597e-01
-9.16928649e-01 -6.82354629e-01 -1.28911808e-01 -6.21516466e-01
-4.82045412e-01 -5.02456650e-02 1.06003501e-01 1.32258795e-02
6.06716394e-01 7.87898302e-01 4.21935290e-01 -4.65379179e-01
-1.79453403e-01 7.94659674e-01 4.51183110e-01 -4.23031777e-01
1.02384579e+00 4.47240353e-01 1.84283033e-01 -1.11630821e+00
-1.09490559e-01 -4.07709450e-01 -4.62556154e-01 -3.24853003e-01
5.06035745e-01 -3.39599282e-01 -8.11247408e-01 1.61754027e-01
-7.26880372e-01 7.58276939e-01 -6.36547208e-01 5.29768884e-01
-1.35809615e-01 8.92403126e-01 -6.36680186e-01 -8.57982874e-01
-2.31429517e-01 -8.28154325e-01 3.74323606e-01 -3.88526022e-02
-1.03909761e-01 -9.25261080e-01 1.63064413e-02 -8.17244723e-02
5.55538297e-01 4.06057328e-01 1.27592111e+00 -1.65053439e+00
-2.99446642e-01 -7.43158877e-01 -8.05726126e-02 8.19596469e-01
2.73815274e-01 2.76659042e-01 -6.11832678e-01 -2.00427949e-01
4.18250710e-01 4.15675372e-01 6.17617726e-01 -6.53842539e-02
1.50389993e+00 -6.14055336e-01 -1.76407009e-01 3.27223718e-01
1.05570066e+00 7.00247109e-01 5.98857939e-01 5.57003677e-01
5.36547959e-01 3.68129075e-01 4.76288378e-01 3.61962348e-01
-2.20740989e-01 1.37395367e-01 3.43082190e-01 2.99140841e-01
1.94184303e-01 8.48642215e-02 5.70358634e-01 1.20712948e+00
-1.56985670e-01 5.60421534e-02 -5.76071918e-01 2.82955199e-01
-1.65423131e+00 -1.80635786e+00 -4.32114154e-01 1.97064114e+00
-2.15345062e-02 5.64790606e-01 4.78583634e-01 1.00160670e+00
9.66901541e-01 2.97379345e-01 -5.59399307e-01 -7.87013710e-01
-6.16933890e-02 2.51513511e-01 2.67396718e-01 1.13892116e-01
-8.58661652e-01 5.06721079e-01 6.16945791e+00 9.10418153e-01
-1.31417620e+00 -6.27334341e-02 3.98502290e-01 2.10900232e-01
7.25910589e-02 -7.82829374e-02 -4.59795296e-01 9.40916538e-01
1.56283367e+00 -3.46475214e-01 1.50255531e-01 1.02777529e+00
2.47254059e-01 4.09531325e-01 -6.92852557e-01 8.00623357e-01
4.22925595e-03 -1.01893365e+00 4.80311900e-01 2.27452487e-01
3.99398416e-01 -1.30243450e-01 7.20488355e-02 7.40256310e-01
8.58097374e-02 -6.77269340e-01 -1.63088500e-01 6.80346429e-01
2.03100324e-01 -1.16402757e+00 9.40998077e-01 9.82495397e-02
-1.45920396e+00 -7.42442966e-01 -4.91389215e-01 -2.71870196e-02
-6.92407116e-02 7.18871057e-01 -6.91996098e-01 6.42823756e-01
5.45309961e-01 8.34611058e-01 -6.44487679e-01 6.84025049e-01
7.63219297e-01 1.09696448e+00 -1.34692401e-01 2.87458897e-01
2.83563942e-01 -4.88004893e-01 1.14869630e+00 1.07122529e+00
7.34275997e-01 -2.66558945e-01 3.75262558e-01 2.86860317e-01
3.28603029e-01 1.70621097e-01 -8.59992266e-01 -1.84613928e-01
7.11652160e-01 1.04241180e+00 -4.09126848e-01 -4.91618037e-01
-7.99890161e-01 8.80845547e-01 -3.70510042e-01 4.92450237e-01
-8.04345429e-01 -9.34570014e-01 9.26228106e-01 2.09910706e-01
3.96173656e-01 3.35685648e-02 -6.52388781e-02 -9.92028534e-01
1.34799033e-01 -9.48617339e-01 1.00389123e+00 -1.28524587e-01
-1.69780934e+00 7.37395942e-01 -3.53288263e-01 -1.69898391e+00
-4.17998821e-01 -5.95917106e-01 -1.15086985e+00 6.49010479e-01
-1.00925672e+00 -8.73736382e-01 -1.49374783e-01 1.07364297e+00
8.32571626e-01 -1.21236753e+00 7.39650428e-01 4.33913589e-01
-8.82114112e-01 5.63722551e-01 3.00746590e-01 2.80344337e-01
2.04482988e-01 -8.59189928e-01 5.09214997e-01 9.94489968e-01
-6.96776062e-02 5.47317743e-01 7.76723206e-01 -8.80331755e-01
-1.14566040e+00 -1.23417974e+00 6.86621070e-01 -3.03442419e-01
1.15848923e+00 3.49557132e-01 -1.34221685e+00 9.76268232e-01
-2.67898589e-01 5.14545679e-01 8.08516800e-01 -1.56749576e-01
-5.40386379e-01 -5.95302999e-01 -1.57841635e+00 3.74840587e-01
7.61524260e-01 -3.96338403e-01 -9.48172629e-01 -1.17946841e-01
6.04306519e-01 3.85362238e-01 -8.85959625e-01 3.43324512e-01
3.07947785e-01 -1.07633495e+00 1.03093255e+00 -1.08719420e+00
-3.14716548e-01 -4.92772102e-01 -3.48387152e-01 -8.14726472e-01
-6.31982803e-01 -2.90806264e-01 -6.23634994e-01 1.10217738e+00
-4.32890803e-02 -1.19935191e+00 8.03449631e-01 -1.65390968e-01
-1.07916676e-01 -5.73092461e-01 -1.27415729e+00 -9.26845670e-01
-5.30884564e-01 -6.03182375e-01 8.63846302e-01 1.18837988e+00
-2.84724802e-01 7.56214187e-02 -2.31485873e-01 2.07825050e-01
6.84151292e-01 -2.22267568e-01 6.98629975e-01 -1.84731340e+00
1.38286218e-01 -4.61293906e-01 -1.33607841e+00 -4.59023505e-01
4.32471156e-01 -7.17937887e-01 -7.35359907e-01 -4.08042192e-01
-4.36498642e-01 -1.20941728e-01 -8.05511773e-01 5.45268543e-02
1.52219031e-02 1.45510654e-04 -3.63671631e-01 4.29596692e-01
-3.80918860e-01 4.93976891e-01 5.69378138e-01 -2.02517249e-02
-1.20638967e-01 2.83165485e-01 -3.98147851e-01 7.53632486e-01
7.68489718e-01 -4.02896821e-01 -4.11910713e-01 4.22313929e-01
-4.99504507e-01 -5.05977161e-02 1.82198137e-02 -1.52811491e+00
7.03604594e-02 -3.64030063e-01 6.07136309e-01 -8.68090570e-01
-1.70188785e-01 -1.31266069e+00 -5.05670020e-03 5.77465057e-01
5.89215942e-02 7.87014306e-01 1.46940485e-01 1.05035794e+00
-3.40987206e-01 -5.71525916e-02 5.88675141e-01 1.93342641e-01
-1.07467711e+00 4.64741260e-01 -8.83663654e-01 -2.31790692e-01
1.29728222e+00 -8.74506831e-01 1.73679933e-01 -1.21156521e-01
-6.74947560e-01 -2.52984576e-02 -1.19189836e-01 6.84800923e-01
5.82434535e-01 -1.89060509e+00 -6.48291051e-01 8.62886727e-01
6.75012171e-02 -8.08418214e-01 4.36471879e-01 7.26909041e-01
-5.05010664e-01 3.07307571e-01 -6.32260084e-01 -6.09712303e-01
-1.38635635e+00 6.76321149e-01 1.43884197e-01 -1.57739252e-01
-1.01460087e+00 1.98885486e-01 -5.25105238e-01 -1.67589560e-01
3.98160033e-02 1.20965913e-01 -4.85222191e-01 -2.26450115e-02
9.05913234e-01 1.17069662e+00 2.01527312e-01 -1.04378366e+00
-8.42745602e-02 3.71917337e-01 -3.57335687e-01 3.28852594e-01
1.32964039e+00 -1.78494424e-01 -4.51002270e-01 5.94942868e-01
1.38019860e+00 1.19167201e-01 -6.71789289e-01 -4.90876585e-01
3.44667971e-01 -8.42631876e-01 -2.75246073e-02 -2.60957535e-02
-1.32600319e+00 5.01169145e-01 1.00665689e+00 8.13798010e-01
1.16585636e+00 -5.68245828e-01 1.22895932e+00 7.47764647e-01
-2.37449072e-02 -1.01259947e+00 3.79926085e-01 7.55508482e-01
3.79702717e-01 -1.03112006e+00 -3.34962010e-01 1.05930204e-02
-5.30292630e-01 1.69904900e+00 8.23251545e-01 -6.18890285e-01
8.48457754e-01 6.46855533e-02 -2.15430468e-01 -3.83605927e-01
-6.59645259e-01 -1.21011399e-02 6.82809651e-02 5.97488463e-01
1.45105243e-01 -2.83874780e-01 -1.41825959e-01 2.67363697e-01
-2.14293242e-01 -5.78161061e-01 3.96794498e-01 8.57358932e-01
-9.71015573e-01 -8.56743634e-01 -7.51839936e-01 9.42982256e-01
-6.13945186e-01 4.27324980e-01 9.12618712e-02 7.44826734e-01
3.01024139e-01 1.08376956e+00 8.05654287e-01 -9.52413678e-01
5.12235224e-01 6.03202641e-01 -3.19878608e-01 -6.43410385e-02
-2.93598086e-01 -4.41713631e-01 -4.16002907e-02 -8.33884597e-01
2.07753941e-01 -5.93461514e-01 -1.17476714e+00 -1.04100192e+00
-1.46383777e-01 4.98243213e-01 3.38536799e-01 1.00514698e+00
4.34675366e-01 8.02186549e-01 1.35273314e+00 -3.20836723e-01
-4.31069076e-01 -9.63307559e-01 -6.04320943e-01 9.39509273e-01
6.06121600e-01 -6.03594184e-01 -8.77772093e-01 -2.47149751e-01] | [7.475388526916504, 2.313858985900879] |
2ebd5587-d5bf-4007-8cd6-041cd2211b04 | generating-sequences-by-learning-to-self | 2211.00053 | null | https://arxiv.org/abs/2211.00053v1 | https://arxiv.org/pdf/2211.00053v1.pdf | Generating Sequences by Learning to Self-Correct | Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot demonstrations, frequently violate these constraints, and lack a mechanism to iteratively revise their outputs. Moreover, some powerful language models are of extreme scale or inaccessible, making it inefficient, if not infeasible, to update their parameters for task-specific adaptation. We present Self-Correction, an approach that decouples an imperfect base generator (an off-the-shelf language model or supervised sequence-to-sequence model) from a separate corrector that learns to iteratively correct imperfect generations. To train the corrector, we propose an online training procedure that can use either scalar or natural language feedback on intermediate imperfect generations. We show that Self-Correction improves upon the base generator in three diverse generation tasks - mathematical program synthesis, lexically-constrained generation, and toxicity control - even when the corrector is much smaller than the base generator. | ['Yejin Choi', 'Daniel Khashabi', 'Tianxiao Shen', 'Faeze Brahman', 'Peter West', 'Ximing Lu', 'Sean Welleck'] | 2022-10-31 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 5.41880965e-01 3.96159112e-01 -1.85349897e-01 -2.31006607e-01
-9.14150119e-01 -1.07981110e+00 6.48959219e-01 2.12511811e-02
-1.73041731e-01 1.04597485e+00 -5.36574125e-02 -6.39372766e-01
4.18866605e-01 -8.75949323e-01 -1.20650613e+00 -1.91554353e-01
1.83916196e-01 4.96192455e-01 2.27548152e-01 -4.99376476e-01
5.11796951e-01 1.73536614e-01 -1.64180434e+00 9.63433832e-02
1.30304718e+00 1.61497936e-01 5.51947713e-01 1.26535022e+00
-3.30294371e-01 9.46482301e-01 -7.68980324e-01 -3.21692944e-01
-5.52585199e-02 -8.61222386e-01 -6.53265417e-01 -1.41058922e-01
2.28133827e-01 -4.07045513e-01 6.46497980e-02 1.19459903e+00
2.93647438e-01 8.54321718e-02 4.44454223e-01 -1.38342428e+00
-9.81862187e-01 9.35404241e-01 -5.41466698e-02 -1.12659276e-01
6.23978257e-01 6.54406965e-01 9.53483045e-01 -5.41858315e-01
6.88529372e-01 1.19801044e+00 5.79944611e-01 1.06532073e+00
-1.45949423e+00 -4.70655233e-01 1.94222331e-01 -4.57673401e-01
-1.19042397e+00 -4.24667358e-01 5.27376115e-01 -5.45180857e-01
1.26624012e+00 1.80100694e-01 5.71128130e-01 1.12960970e+00
8.81850943e-02 5.55597901e-01 7.02057838e-01 -6.02081895e-01
4.72167671e-01 1.67319492e-01 -8.41234028e-02 1.22218561e+00
2.65248388e-01 4.17783588e-01 -4.44321960e-01 -5.26255071e-01
8.24405193e-01 -4.71829176e-01 -3.84923100e-01 -4.83993828e-01
-1.21440029e+00 8.43423605e-01 -7.59442002e-02 6.66706264e-02
-3.41397114e-02 7.45141149e-01 3.83191913e-01 5.58625638e-01
2.44125593e-02 1.23912120e+00 -6.18870080e-01 -5.06568551e-01
-9.11337554e-01 5.82365513e-01 1.15812433e+00 1.20470178e+00
7.86559284e-01 4.16725934e-01 -1.34557471e-01 3.76731545e-01
-3.30937952e-02 4.47172463e-01 7.59799778e-01 -1.11107683e+00
3.17028970e-01 4.68409210e-01 3.83541167e-01 -4.59661931e-01
-7.32125938e-02 -2.68496394e-01 -3.78912807e-01 4.76914853e-01
3.50225121e-01 -4.13111925e-01 -9.42332447e-01 2.26806474e+00
1.56576678e-01 -4.14420944e-03 1.37186527e-01 5.94039917e-01
3.36398304e-01 9.39920664e-01 4.07384150e-02 -2.69102752e-01
8.07000101e-01 -1.04850161e+00 -3.52459550e-01 -2.26780698e-01
9.53995407e-01 -5.05818725e-01 1.68056011e+00 3.94054353e-01
-1.54520190e+00 -5.21979809e-01 -1.27108920e+00 5.10648303e-02
-3.51484120e-01 -1.66107178e-01 6.21387124e-01 6.72201037e-01
-1.21899223e+00 9.44840252e-01 -8.08465242e-01 -1.09809473e-01
-8.22032765e-02 3.51109117e-01 3.54633182e-02 2.69264013e-01
-1.03713989e+00 1.00410545e+00 5.32549262e-01 -4.31672156e-01
-1.30804574e+00 -9.59188521e-01 -1.03165102e+00 8.43509957e-02
4.18747902e-01 -1.14103198e+00 1.87176919e+00 -1.28582549e+00
-1.97248352e+00 6.12448215e-01 -1.51179805e-02 -3.67421210e-01
6.02443516e-01 -2.63473149e-02 -9.64646712e-02 -1.02775171e-01
1.13890851e-02 7.03576982e-01 9.68238473e-01 -1.08751166e+00
-5.32532454e-01 2.01480225e-01 3.62837583e-01 1.51446596e-01
-7.61825293e-02 -1.19143307e-01 -8.12114924e-02 -5.93254209e-01
-4.27033156e-01 -9.69068050e-01 -4.61740524e-01 -3.26782018e-02
-4.87204671e-01 -1.90597460e-01 4.79890972e-01 -4.75513995e-01
1.23847044e+00 -1.75088429e+00 4.80703026e-01 4.38692456e-04
-6.55666366e-02 2.96311885e-01 -5.21830201e-01 4.77913737e-01
9.65422094e-02 4.63468939e-01 -4.86367047e-01 6.74051195e-02
3.26420337e-01 1.46392703e-01 -3.37771416e-01 2.52984501e-02
4.00088727e-01 1.08109081e+00 -1.35440516e+00 -3.06918263e-01
-3.57561968e-02 1.69815551e-02 -9.97485638e-01 4.51832891e-01
-1.09461796e+00 2.84165025e-01 -3.79988223e-01 3.42522621e-01
1.15841776e-01 -3.81009698e-01 1.38129801e-01 3.48916024e-01
-7.56962225e-02 2.94914275e-01 -1.05068898e+00 1.87768435e+00
-9.38623190e-01 3.84265512e-01 -1.20545693e-01 -6.45961225e-01
7.86622703e-01 2.30439767e-01 -1.73807278e-01 -4.72136736e-01
-2.47187972e-01 5.50089180e-01 1.53806359e-02 -6.01766467e-01
6.94493473e-01 -7.64714628e-02 -3.67884517e-01 8.64157379e-01
6.52895048e-02 -9.65247929e-01 5.90745330e-01 3.62904221e-01
1.12197626e+00 4.16026115e-01 4.31441396e-01 -9.17626098e-02
5.22092879e-01 1.85935780e-01 3.84419769e-01 1.13731647e+00
2.13082597e-01 5.78171015e-01 6.39502406e-01 -9.31405276e-02
-1.54543269e+00 -1.00072932e+00 7.02072382e-01 1.29164243e+00
-1.34717852e-01 -4.68501627e-01 -1.05530250e+00 -6.04928255e-01
-5.88624850e-02 1.20222032e+00 -2.95952708e-01 -5.98489046e-01
-8.59392941e-01 -2.78831631e-01 7.61071622e-01 5.35208523e-01
1.33431122e-01 -1.17643702e+00 -6.04077697e-01 4.76274490e-01
4.42086048e-02 -5.87972164e-01 -9.53801155e-01 1.60450742e-01
-8.03137839e-01 -9.12594140e-01 -6.01684272e-01 -9.83743012e-01
9.20104682e-01 -2.23804861e-01 1.36361754e+00 5.53772569e-01
-1.43690929e-01 3.68103683e-01 4.11081463e-02 -9.28245634e-02
-1.28749228e+00 3.10543239e-01 -2.18709335e-01 -7.60775089e-01
-2.55097151e-01 -7.17243969e-01 -1.89233333e-01 -1.14475057e-01
-8.95712674e-01 3.04051518e-01 3.97696167e-01 1.07070196e+00
2.98305541e-01 -3.15954268e-01 8.70609701e-01 -1.01829457e+00
1.06563270e+00 -9.60005149e-02 -1.04886723e+00 4.29057717e-01
-6.02022171e-01 5.45665324e-01 1.12476015e+00 -7.48829961e-01
-8.21572244e-01 1.21248320e-01 -9.83057171e-02 -3.45704526e-01
1.63427740e-01 5.08184314e-01 -1.32283911e-01 -4.68059555e-02
1.04388511e+00 5.36646664e-01 -4.15292494e-02 -8.72320309e-03
7.07063973e-01 2.55531013e-01 8.02069068e-01 -1.00479317e+00
1.09077036e+00 -3.01141560e-01 -2.82864571e-01 -4.06438351e-01
-5.26604593e-01 1.74909160e-01 -2.27354273e-01 2.95162499e-01
3.74811918e-01 -6.51551366e-01 -6.20671153e-01 4.16956037e-01
-1.39466059e+00 -9.92540717e-01 -5.52918375e-01 4.55134995e-02
-1.06261957e+00 1.96708098e-01 -6.83636725e-01 -7.01296747e-01
-4.67145205e-01 -1.18840313e+00 9.27892804e-01 2.51600653e-01
-6.90886438e-01 -9.23574328e-01 1.95406318e-01 -2.52929926e-01
6.57130718e-01 1.41084254e-01 1.44157124e+00 -6.08560443e-01
-8.11156213e-01 -8.26544985e-02 1.95496961e-01 3.03061783e-01
8.71009305e-02 3.92926753e-01 -4.93759871e-01 -1.76016629e-01
-3.30619633e-01 -6.98610902e-01 4.17248219e-01 3.46750841e-02
1.10074091e+00 -6.57609463e-01 -1.28427804e-01 5.05602956e-01
1.21599424e+00 1.74955264e-01 3.47768158e-01 1.83039948e-01
4.40111965e-01 1.57457665e-01 2.56132960e-01 3.70518774e-01
1.82033047e-01 4.16833967e-01 1.65554240e-01 2.56779075e-01
-1.01812556e-01 -9.03695524e-01 5.00969052e-01 7.11274445e-01
2.38595009e-01 -2.66553223e-01 -7.17870116e-01 6.42829537e-01
-1.74335003e+00 -1.06494057e+00 3.90436649e-01 2.21197033e+00
1.54305899e+00 2.68379420e-01 2.02740226e-02 -2.01167062e-01
6.79195106e-01 3.51743726e-03 -7.49168992e-01 -7.02345252e-01
1.38999209e-01 3.81326407e-01 1.61523595e-01 9.42543805e-01
-5.98181307e-01 1.25490701e+00 6.82166386e+00 6.15409911e-01
-1.01969016e+00 -6.69196174e-02 3.28830272e-01 -3.01125944e-02
-8.84424508e-01 1.98825046e-01 -5.09962618e-01 4.62704390e-01
1.00911915e+00 -7.59751022e-01 9.55855548e-01 1.10388339e+00
2.84705311e-01 7.20191896e-02 -1.52129924e+00 4.82776582e-01
-6.58745989e-02 -1.72379434e+00 1.51300877e-01 -4.93292332e-01
1.01039088e+00 -3.47753286e-01 6.37504831e-02 6.22325122e-01
9.18811142e-01 -1.15768826e+00 9.85151827e-01 3.77292931e-01
9.61498916e-01 -7.65193105e-01 -9.70592909e-03 7.21249282e-01
-8.46286893e-01 1.15667418e-01 -1.67193547e-01 -6.21291362e-02
2.38244221e-01 8.36870912e-03 -9.14971292e-01 -9.22176614e-02
-1.73830777e-01 2.24829867e-01 -3.84538800e-01 7.82613456e-01
-5.71895957e-01 3.66006225e-01 -1.11645065e-01 -4.92242992e-01
1.42218426e-01 1.01031199e-01 5.80110967e-01 1.20689499e+00
6.28909945e-01 2.64745444e-01 2.63100713e-01 1.46929991e+00
-1.70852512e-01 -2.30316311e-01 -6.12758934e-01 -6.11262918e-01
5.88775039e-01 9.21965599e-01 -2.54481554e-01 -6.22145951e-01
-1.30968153e-01 1.10266769e+00 3.40037882e-01 4.98727649e-01
-9.48376656e-01 -6.59616709e-01 6.63551092e-01 -1.30102947e-01
1.85788646e-01 -4.17687774e-01 -3.38485360e-01 -1.16657031e+00
-2.35344291e-01 -1.36401451e+00 -5.00605404e-02 -9.76766348e-01
-8.22111130e-01 4.79880154e-01 -2.10407913e-01 -7.87569463e-01
-9.99894202e-01 -3.04646224e-01 -7.23255813e-01 9.90835190e-01
-1.34836531e+00 -8.01285207e-01 -5.59829548e-02 3.31214100e-01
6.98385239e-01 -1.34032458e-01 9.50315416e-01 -1.38720214e-01
-3.48360479e-01 7.93391347e-01 -2.75270492e-01 -2.31568590e-01
2.72575855e-01 -1.41632020e+00 7.28284895e-01 9.13347006e-01
-2.19410703e-01 8.60680819e-01 1.07882416e+00 -7.52993345e-01
-1.72387946e+00 -1.10482311e+00 8.66028905e-01 -3.55768234e-01
9.38731670e-01 -3.93574983e-01 -9.21523273e-01 7.54858732e-01
4.22101747e-03 -3.09124440e-01 1.66224957e-01 -2.47471496e-01
-3.86739582e-01 4.03310269e-01 -1.04537129e+00 1.03734660e+00
1.27056968e+00 -6.80982172e-01 -6.14199579e-01 5.01044214e-01
1.22838855e+00 -7.37474799e-01 -4.19803917e-01 4.63575050e-02
3.39106113e-01 -4.97734338e-01 7.39560366e-01 -9.11295235e-01
8.14393997e-01 -4.24264848e-01 4.14605159e-03 -1.67682648e+00
-3.92765962e-02 -1.26532543e+00 -3.59617382e-01 9.47878361e-01
8.20990086e-01 -5.49836099e-01 8.25969994e-01 7.49059796e-01
-4.29212540e-01 -4.22376215e-01 -4.18566167e-01 -9.10058141e-01
3.35001320e-01 -1.87613949e-01 8.85785937e-01 6.80691779e-01
4.16354835e-01 4.42678958e-01 -2.95109332e-01 4.31651622e-02
1.41265541e-01 3.42577875e-01 9.72186863e-01 -6.77444518e-01
-7.84030020e-01 -6.45061672e-01 1.54409721e-01 -1.15008426e+00
3.79558384e-01 -9.35253978e-01 4.91458595e-01 -1.23313296e+00
1.03564207e-02 -4.27594125e-01 2.85914034e-01 4.32496727e-01
-3.52233440e-01 -3.95932138e-01 6.50097728e-02 -1.21213071e-01
-3.69408131e-01 5.24973333e-01 1.38978803e+00 -1.84139118e-01
-4.81159955e-01 2.41690092e-02 -8.61794770e-01 5.75324953e-01
7.11992919e-01 -3.05808246e-01 -7.02528715e-01 -4.55931515e-01
5.97674966e-01 4.36069548e-01 2.95765549e-01 -7.83839464e-01
3.64864230e-01 -5.70804775e-01 -7.78851137e-02 -6.56586364e-02
4.65603881e-02 -2.23716959e-01 2.14846998e-01 7.15584338e-01
-9.09496486e-01 4.39302951e-01 2.03823656e-01 4.09784228e-01
1.65721908e-01 -6.95807993e-01 7.19746888e-01 -5.70381403e-01
-6.98988974e-01 1.01410195e-01 -5.90090394e-01 2.94102967e-01
8.15464556e-01 -7.32844695e-02 -5.01765549e-01 -4.18511033e-01
-4.77647662e-01 1.88318536e-01 8.61322224e-01 2.57628173e-01
5.70306957e-01 -1.03636849e+00 -6.27101541e-01 1.60043746e-01
9.95959416e-02 -3.82282101e-02 -1.82279810e-01 2.34875590e-01
-5.70498943e-01 2.31711909e-01 7.31474832e-02 -2.56512880e-01
-9.16193485e-01 7.77129471e-01 5.91146231e-01 -3.82947832e-01
-3.92131567e-01 8.39955449e-01 1.71175867e-01 -7.55963326e-01
1.75869226e-01 -5.53154051e-01 4.31467891e-01 -5.23283184e-01
4.66736346e-01 -3.27828266e-02 -1.68418527e-01 1.50608391e-01
6.44661933e-02 2.77001411e-01 4.47301380e-02 -5.35724461e-01
1.05915141e+00 2.24204347e-01 -4.16587256e-02 4.02978837e-01
8.60887766e-01 -1.48463249e-01 -1.26234949e+00 -8.96714032e-02
7.50403898e-03 -2.16473997e-01 -3.96622002e-01 -9.44326818e-01
-4.78448451e-01 7.57401228e-01 -2.85686292e-02 2.01558977e-01
7.72692740e-01 -3.61631036e-01 8.37193847e-01 7.71216512e-01
5.47221124e-01 -1.10058260e+00 4.96585667e-01 6.43978596e-01
9.42033529e-01 -9.17942762e-01 -3.94523293e-01 -2.25853443e-01
-5.51152587e-01 1.07424486e+00 8.50579441e-01 -1.35684669e-01
2.20686346e-02 5.06891251e-01 -2.86067784e-01 1.97405547e-01
-1.25927234e+00 7.13515058e-02 -5.16490750e-02 7.75458097e-01
3.57582808e-01 -4.07522842e-02 -1.74427703e-01 4.03182626e-01
-6.61295593e-01 3.22644711e-01 8.89523864e-01 1.10521889e+00
-6.80595756e-01 -1.14387286e+00 -1.78058401e-01 3.26654673e-01
-1.24204770e-01 -3.29333544e-01 -3.06777656e-01 7.88534820e-01
-2.06121892e-01 7.59527445e-01 -6.51077181e-02 -1.28515989e-01
1.71986118e-01 2.50501305e-01 7.48724103e-01 -9.95858431e-01
-7.24588633e-01 -3.26324105e-01 3.17430735e-01 -3.94062966e-01
2.11048990e-01 -3.85907263e-01 -1.37797558e+00 -4.82041359e-01
-3.70639175e-01 2.26935789e-01 5.59234381e-01 8.38736773e-01
2.89167553e-01 4.95916307e-01 5.30062199e-01 -5.95688462e-01
-1.07368898e+00 -5.73513269e-01 -1.87191606e-01 1.19331822e-01
5.08190036e-01 -1.25904560e-01 -2.74107307e-01 3.44629824e-01] | [8.203746795654297, 7.501020431518555] |
8a780367-142e-48bc-b0b6-b121130a9973 | pmhld-patch-map-based-hybrid-learning | null | null | https://ieeexplore.ieee.org/document/9094006 | https://ieeexplore.ieee.org/document/9094006 | PMHLD: Patch Map Based Hybrid Learning DehazeNet for Single Image Haze Removal | Images captured in a hazy environment usually suffer from bad visibility and missing information. Over many years, learning-based and handcrafted prior-based dehazing algorithms have been rigorously developed. However, both algorithms exhibit some weaknesses in terms of haze removal performance. Therefore, in this work, we have proposed the patch-map-based hybrid learning DehazeNet, which integrates these two strategies by using a hybrid learning technique involving the patch map and a bi-attentive generative adversarial network. In this method, the reasons limiting the performance of the dark channel prior (DCP) have been analyzed. A new feature called the patch map has been defined for selecting the patch size adaptively. Using this map, the limitations of the DCP (e.g., color distortion and failure to recover images involving white scenes) can be addressed efficiently. In addition, to further enhance the performance of the method for haze removal, a patch-map-based DCP has been embedded into the network, and this module has been trained with the atmospheric light generator, patch map selection module, and refined module simultaneously. A combination of traditional and learning-based methods can efficiently improve the haze removal performance of the network. Experimental results show that the proposed method can achieve better reconstruction results compared to other state-of-the-art haze removal algorithms. | ['Sy-Yen Kuo', 'Jian-Jiun Ding', 'Hao-Yu Feng', 'Wei-Ting Chen'] | 2020-05-14 | null | null | null | ieee-transaction-on-image-processing-2020-5 | ['single-image-haze-removal', 'single-image-deraining', 'computational-phenotyping'] | ['computer-vision', 'computer-vision', 'medical'] | [ 2.62618631e-01 -2.67308682e-01 5.03923416e-01 7.14529902e-02
-4.29014117e-01 1.58975739e-02 4.11314040e-01 -4.12116826e-01
-1.87066704e-01 7.86327839e-01 -7.99508467e-02 -6.52382001e-02
1.92361511e-02 -1.09125662e+00 -5.38529575e-01 -1.49912488e+00
2.83191025e-01 -2.67632663e-01 6.15543008e-01 -4.21643049e-01
2.86845773e-01 3.54892015e-01 -1.74794269e+00 6.30687848e-02
1.59077680e+00 7.81368852e-01 4.78208691e-01 4.50685591e-01
-7.83469155e-02 9.34654474e-01 -7.73980200e-01 -4.50379364e-02
5.78843057e-01 -7.89864063e-01 1.24925502e-01 9.76121649e-02
5.05205810e-01 -5.73212564e-01 -4.41341698e-01 1.25503612e+00
8.37804019e-01 3.18890989e-01 5.27874589e-01 -1.17723858e+00
-6.06688619e-01 -7.25243147e-03 -5.82526445e-01 2.36189082e-01
-1.02685690e-01 3.17095160e-01 1.94391444e-01 -1.15806603e+00
4.46337759e-02 9.85494614e-01 5.98342597e-01 3.90414864e-01
-8.62174213e-01 -1.07354105e+00 -8.83486420e-02 4.23824191e-01
-1.69401479e+00 -3.58393461e-01 1.27320659e+00 -3.68057072e-01
5.07566392e-01 2.03712717e-01 8.04583073e-01 4.79453385e-01
6.09821022e-01 5.31482875e-01 1.46924818e+00 -7.11113214e-01
1.65502965e-01 3.12365443e-01 -3.73271823e-01 6.47138357e-01
4.25121963e-01 3.90532851e-01 -2.93367773e-01 2.19424590e-01
7.73070157e-01 1.69246972e-01 -4.70430672e-01 -1.58701986e-01
-4.53430474e-01 7.03866541e-01 6.09728456e-01 1.49910718e-01
-2.70544291e-01 -1.61621377e-01 -3.91318828e-01 1.50690094e-01
5.98424971e-01 4.08939481e-01 1.04672104e-01 6.30400419e-01
-1.08240485e+00 2.64481962e-01 4.86742556e-01 7.60663927e-01
1.23747087e+00 6.75056696e-01 -3.10564667e-01 9.12610829e-01
5.44319808e-01 8.82158577e-01 2.20581025e-01 -7.09082007e-01
2.66621143e-01 4.75699484e-01 2.54673004e-01 -1.31812227e+00
-1.73104333e-03 -4.93971884e-01 -1.09419167e+00 7.74127662e-01
-4.71126251e-02 -2.48674423e-01 -1.46154964e+00 1.30010569e+00
4.67480361e-01 5.48273861e-01 1.74288020e-01 1.08739495e+00
8.12604070e-01 1.18082690e+00 -2.48132125e-01 -3.92450541e-01
6.69102550e-01 -1.16384208e+00 -1.18011987e+00 -2.42759094e-01
-9.32729021e-02 -9.61849391e-01 7.10653543e-01 6.30007207e-01
-1.04688132e+00 -7.30273843e-01 -1.40205264e+00 2.21588627e-01
-5.28304577e-01 -1.74651518e-01 2.15622008e-01 7.89307475e-01
-1.09045053e+00 1.88906178e-01 -5.18679023e-01 1.20329112e-02
1.74624220e-01 1.48092568e-01 2.25691974e-01 -5.10117233e-01
-1.44861436e+00 1.13962233e+00 4.71255362e-01 4.93656933e-01
-1.15998554e+00 -5.78114927e-01 -6.83835208e-01 -1.04838960e-01
3.93299997e-01 -6.73087478e-01 6.03862584e-01 -1.10949063e+00
-1.65927303e+00 2.19769880e-01 1.18635699e-01 -1.47479877e-01
2.21853077e-01 -3.70617747e-01 -8.33223879e-01 3.09961498e-01
-9.56635773e-02 3.07958841e-01 1.48164856e+00 -1.97766995e+00
-7.58673072e-01 -6.84065074e-02 -1.75627191e-02 3.19457799e-01
-2.00800151e-01 -1.94074661e-01 -4.61685598e-01 -9.10102069e-01
3.40831652e-02 -6.93286955e-01 -2.58346379e-01 -3.66411544e-02
-2.05852136e-01 3.01546037e-01 1.12633967e+00 -8.90846312e-01
1.40204740e+00 -2.13371134e+00 -3.15247700e-02 3.03042740e-01
1.67996377e-01 7.59959280e-01 2.80093886e-02 4.64857668e-01
8.86370838e-02 -5.91818839e-02 -6.17813885e-01 -1.50179386e-01
-3.71839613e-01 3.32653999e-01 -1.02929242e-01 7.25772917e-01
2.43989721e-01 4.09919798e-01 -7.50752747e-01 -5.74134946e-01
6.58008039e-01 8.13255191e-01 -3.79083127e-01 6.69857502e-01
-3.90852168e-02 5.48220515e-01 -2.24624246e-01 6.32244468e-01
1.24185455e+00 4.09409136e-01 -4.80491847e-01 -1.96921989e-01
-4.66413021e-01 -3.74379188e-01 -1.39034224e+00 1.17490399e+00
-3.92233968e-01 3.60516310e-01 3.46125633e-01 -6.66116893e-01
1.00947726e+00 2.78087497e-01 1.88885212e-01 -6.45266414e-01
3.26503068e-02 2.25483030e-01 -1.25076115e-01 -8.07563663e-01
3.83080095e-01 -4.61609155e-01 6.17785871e-01 -6.96104541e-02
-2.41316840e-01 -4.88448650e-01 -2.11795151e-01 -8.62902850e-02
7.47875214e-01 1.64818704e-01 -4.07272950e-02 -1.90121531e-01
8.86306226e-01 5.02056442e-02 8.73302877e-01 7.23187208e-01
-1.71878785e-01 9.05607939e-01 -2.70060450e-01 -3.36045980e-01
-9.73255098e-01 -8.14739704e-01 -1.68308504e-02 6.46923840e-01
7.70840168e-01 -1.46305319e-02 -8.20674121e-01 -1.48308963e-01
-3.47133487e-01 8.42186630e-01 -5.77168882e-01 -5.85797846e-01
-7.29475200e-01 -1.09653640e+00 2.44340926e-01 8.14943947e-03
1.09951115e+00 -8.76770377e-01 -2.25454941e-01 1.81421652e-01
-8.46830159e-02 -8.64111066e-01 -1.85610905e-01 -7.26350546e-02
-6.84219003e-01 -9.47304130e-01 -6.68684363e-01 -8.33093286e-01
7.42216051e-01 9.04532850e-01 7.23257363e-01 3.99939775e-01
-5.95436506e-02 2.22166002e-01 -6.74662411e-01 -7.77698398e-01
-4.29660082e-01 -4.51245695e-01 -3.34710300e-01 5.78954399e-01
-4.47757281e-02 -7.48031080e-01 -1.06063652e+00 4.04821366e-01
-1.27356040e+00 1.22053757e-01 9.65179861e-01 9.20678318e-01
5.12075424e-01 8.33784759e-01 2.83963203e-01 -8.35597396e-01
2.74058372e-01 -5.78531325e-01 -7.34769464e-01 1.18501827e-01
-9.06815708e-01 -3.75110954e-01 6.47569180e-01 -3.47521693e-01
-1.64224374e+00 -1.14140078e-01 -3.79443988e-02 -6.29931808e-01
-1.82057798e-01 3.82619321e-01 -5.08437872e-01 -5.93482554e-01
4.40863609e-01 6.23181403e-01 1.44071475e-01 -3.12868804e-01
1.54460996e-01 5.32402277e-01 5.32154799e-01 -1.46843761e-01
1.65077901e+00 5.37501216e-01 -2.17191856e-02 -8.85353863e-01
-7.76522458e-01 -4.14643705e-01 -3.12838346e-01 -4.57442254e-01
9.79571164e-01 -1.24830365e+00 5.38765565e-02 9.86342192e-01
-1.01552248e+00 -2.79920876e-01 7.14907423e-02 4.88432080e-01
-1.57613158e-02 4.53353107e-01 -4.37052459e-01 -1.08935761e+00
-4.23429966e-01 -9.64861691e-01 6.12706602e-01 6.35335803e-01
8.73857975e-01 -9.26673949e-01 7.79491514e-02 4.66638327e-01
8.66603017e-01 4.51709539e-01 7.02627361e-01 -1.11167668e-03
-1.11907494e+00 -8.89935046e-02 -1.09692141e-01 8.29177499e-01
2.52747625e-01 -4.26258706e-02 -1.07932246e+00 -5.27129889e-01
3.64417255e-01 2.95244575e-01 1.03293931e+00 3.66553694e-01
9.76874292e-01 -3.67051899e-01 -1.73846856e-01 1.09113204e+00
1.96013117e+00 5.52787066e-01 1.12030458e+00 6.95556819e-01
8.49472404e-01 3.21608335e-01 7.05936253e-01 2.90569186e-01
1.56114588e-03 4.79391575e-01 9.44437087e-01 -4.58031207e-01
-5.04274130e-01 2.02567559e-02 3.10768843e-01 7.15379715e-01
-3.42302501e-01 -5.04115820e-01 -5.19188046e-01 4.30809647e-01
-1.62719882e+00 -8.72448146e-01 -7.72285163e-02 2.09820366e+00
7.62383044e-01 1.87441241e-02 -5.06300211e-01 1.27772987e-01
9.11594212e-01 4.71184164e-01 -2.86669642e-01 2.22744737e-02
-3.82815719e-01 1.67168722e-01 6.01239741e-01 5.49157500e-01
-1.10648918e+00 9.61601019e-01 5.90202332e+00 9.86240447e-01
-1.18536758e+00 2.12132722e-01 2.28393218e-03 2.71345079e-01
-4.31753367e-01 2.67178901e-02 -6.86507523e-01 7.95682251e-01
4.62650806e-01 3.19730461e-01 7.16511190e-01 3.46551329e-01
4.90883470e-01 -2.31572717e-01 -1.97486550e-01 9.43169832e-01
6.17855728e-01 -8.99772286e-01 1.09078161e-01 -4.88478914e-02
1.07604706e+00 -2.60616481e-01 1.01983353e-01 2.16577962e-01
7.37325177e-02 -9.25440609e-01 5.89824200e-01 7.01593816e-01
5.59504330e-01 -7.85966516e-01 1.11270475e+00 3.45946878e-01
-8.80706906e-01 -1.37228549e-01 -5.55817664e-01 -2.00092699e-02
2.32654829e-02 9.91875052e-01 -4.95855153e-01 9.15976703e-01
8.25576127e-01 4.70614552e-01 -5.97727180e-01 1.50160372e+00
-7.35806584e-01 8.07857573e-01 -1.46964446e-01 4.02260780e-01
1.38953626e-01 -4.88957673e-01 7.85814583e-01 1.09107554e+00
4.65359718e-01 4.74087477e-01 1.91992268e-01 9.61921155e-01
3.67177784e-01 -1.19165450e-01 -5.59639335e-01 4.63922024e-01
4.35173094e-01 1.15263963e+00 -3.05141330e-01 -1.74680620e-01
-5.54948032e-01 7.09126353e-01 -2.85816699e-01 6.95686102e-01
-9.74534214e-01 -5.49335539e-01 4.13066179e-01 2.04385057e-01
4.81332749e-01 -1.56675309e-01 -6.91879094e-02 -9.26861942e-01
-1.53733879e-01 -9.31308270e-01 2.91074533e-02 -1.04040813e+00
-1.31150234e+00 5.01907587e-01 3.56314853e-02 -1.36105883e+00
4.16029513e-01 -4.40820217e-01 -1.00594616e+00 1.07039547e+00
-2.19292545e+00 -1.25479376e+00 -8.68716061e-01 8.63333523e-01
5.81192315e-01 -2.76632726e-01 3.33521217e-01 5.49520314e-01
-9.39359248e-01 2.86647081e-01 3.11751604e-01 -1.33330002e-01
8.53085399e-01 -1.10761940e+00 -3.95886898e-01 1.61322248e+00
-3.94243926e-01 3.23159337e-01 9.35912192e-01 -7.16360390e-01
-1.30032945e+00 -1.49001706e+00 1.57337025e-01 -5.70587479e-02
1.64015144e-01 -6.70952257e-03 -1.15107000e+00 2.99280286e-01
5.50775588e-01 5.59035167e-02 4.38466102e-01 -6.82188392e-01
-8.64401683e-02 -6.25822425e-01 -1.15998507e+00 5.46213746e-01
4.36821073e-01 -1.74690962e-01 -6.79919124e-01 1.74661115e-01
7.53759146e-01 -4.40083116e-01 -6.22116864e-01 6.67165697e-01
1.76556319e-01 -1.22135556e+00 1.05530572e+00 1.04240566e-01
2.79731870e-01 -9.22772944e-01 -1.36315644e-01 -1.55953407e+00
-6.64457440e-01 -6.35427296e-01 -2.38210127e-01 1.30436361e+00
1.08686671e-01 -7.60715127e-01 4.71378714e-01 2.00502396e-01
-6.52102411e-01 -6.17375910e-01 -4.89228576e-01 -6.14595711e-01
1.10455696e-02 2.06265688e-01 6.43522203e-01 9.46116328e-01
-8.79345179e-01 7.16795996e-02 -8.52786064e-01 9.83011007e-01
8.69927347e-01 6.73673004e-02 8.57591391e-01 -9.89204943e-01
-5.63535541e-02 -1.60444990e-01 -6.50163665e-02 -7.28377938e-01
-1.91192493e-01 -3.28372717e-01 5.02215505e-01 -1.63978195e+00
-1.39486266e-03 -3.61634642e-01 -5.55454016e-01 2.52147973e-01
-7.47259438e-01 3.13523769e-01 1.00963645e-01 4.60528731e-01
-2.37572491e-01 8.04421604e-01 1.66060293e+00 -3.67407978e-01
-2.71786422e-01 -1.26764044e-01 -5.00558078e-01 5.34595430e-01
1.02646160e+00 -5.81448197e-01 -4.36600834e-01 -5.97367764e-01
2.14218199e-02 -2.16340661e-01 3.92910242e-01 -1.33130825e+00
5.33852518e-01 -3.69355530e-01 5.68448544e-01 -7.17574358e-01
4.35426801e-01 -1.00501955e+00 2.36208931e-01 3.48328739e-01
2.54019737e-01 -3.40871990e-01 1.14337519e-01 6.97326243e-01
-5.17933249e-01 -1.19787604e-01 1.11517763e+00 -2.84166485e-01
-8.91289175e-01 2.56599039e-01 -4.14759547e-01 -2.12107673e-01
1.04175639e+00 -6.33041918e-01 -5.13073146e-01 -4.29102182e-01
-1.64442718e-01 1.31406426e-01 4.62175995e-01 5.22300117e-02
1.02204025e+00 -1.03538287e+00 -7.92416871e-01 4.81318384e-01
-4.23134863e-02 2.68849462e-01 4.24384773e-01 8.92428815e-01
-8.73896778e-01 -2.25295082e-01 -3.21309209e-01 -2.40024656e-01
-8.72618616e-01 6.82740867e-01 4.77222562e-01 9.67784151e-02
-6.60931885e-01 6.40083492e-01 3.21004391e-01 -2.03212440e-01
6.72942325e-02 2.09025204e-01 -3.08707505e-01 -3.97059053e-01
5.69727004e-01 5.48476517e-01 -4.35613357e-02 -4.66319442e-01
1.17412321e-01 6.44665480e-01 2.87782431e-01 1.87874064e-01
1.34033120e+00 -5.48305154e-01 -5.56495249e-01 5.99525273e-02
6.71580791e-01 4.13475722e-01 -1.40743136e+00 -1.40854925e-01
-7.59548366e-01 -8.62050474e-01 5.20436287e-01 -9.67643380e-01
-1.47216618e+00 7.88248181e-01 9.13397014e-01 2.09066585e-01
1.59783340e+00 -5.93896389e-01 1.03439915e+00 -5.27181141e-02
1.67997062e-01 -9.53864932e-01 1.48450181e-01 1.57831803e-01
7.51962960e-01 -1.19414067e+00 1.99332505e-01 -7.40466416e-01
-1.72594205e-01 8.77987087e-01 9.94281590e-01 -2.54905432e-01
6.61235869e-01 1.16104081e-01 3.43079001e-01 -1.37743056e-01
-3.37349921e-01 -1.48551002e-01 2.42516115e-01 6.60313725e-01
-1.81413203e-01 -2.56077856e-01 -1.68998897e-01 2.07146734e-01
1.39913350e-01 -3.19640428e-01 6.69104457e-01 9.98471916e-01
-9.72964108e-01 -8.24203908e-01 -9.90788102e-01 9.68120545e-02
-3.88260216e-01 -2.45020628e-01 5.80305718e-02 7.37758279e-01
8.37859094e-01 1.30751526e+00 -3.90771806e-01 -5.41518152e-01
2.00135633e-01 -2.20731169e-01 2.59794295e-01 -4.21234339e-01
-3.18189263e-01 2.85642445e-01 -3.42166513e-01 -2.46213004e-01
-7.30839491e-01 -2.42924951e-02 -7.78382242e-01 -4.36736308e-02
-7.66489089e-01 2.02469841e-01 4.59055662e-01 7.91419983e-01
1.16520539e-01 6.06784642e-01 8.45374823e-01 -1.02457201e+00
-1.03266478e-01 -9.50957596e-01 -8.35781276e-01 1.25090569e-01
6.06896758e-01 -8.31061125e-01 -8.24894428e-01 9.29794610e-02] | [10.901021003723145, -3.1564838886260986] |
01ea9f89-65d6-49c8-8456-43157440eef0 | precise-affordance-annotation-for-egocentric | 2206.05424 | null | https://arxiv.org/abs/2206.05424v1 | https://arxiv.org/pdf/2206.05424v1.pdf | Precise Affordance Annotation for Egocentric Action Video Datasets | Object affordance is an important concept in human-object interaction, providing information on action possibilities based on human motor capacity and objects' physical property thus benefiting tasks such as action anticipation and robot imitation learning. However, existing datasets often: 1) mix up affordance with object functionality; 2) confuse affordance with goal-related action; and 3) ignore human motor capacity. This paper proposes an efficient annotation scheme to address these issues by combining goal-irrelevant motor actions and grasp types as affordance labels and introducing the concept of mechanical action to represent the action possibilities between two objects. We provide new annotations by applying this scheme to the EPIC-KITCHENS dataset and test our annotation with tasks such as affordance recognition. We qualitatively verify that models trained with our annotation can distinguish affordance and mechanical actions. | ['Yoichi Sato', 'Yusuke Goutsu', 'Takuma Yagi', 'Ryosuke Furuta', 'Yifei HUANG', 'Zecheng Yu'] | 2022-06-11 | null | null | null | null | ['affordance-recognition', 'action-anticipation'] | ['computer-vision', 'computer-vision'] | [ 2.25951225e-01 2.01011866e-01 -3.83662701e-01 -3.33375305e-01
-4.99885976e-02 -6.76512659e-01 7.20238388e-01 -5.44269867e-02
-3.69421482e-01 5.32052279e-01 4.67067450e-01 -1.17399298e-01
-1.25836059e-01 -3.61731559e-01 -5.55203438e-01 -3.02631438e-01
-1.34977162e-01 4.80365098e-01 3.93700659e-01 -1.46126002e-01
3.99166137e-01 8.02624702e-01 -1.69138408e+00 3.73016208e-01
7.62144446e-01 8.86633396e-01 5.78766882e-01 5.55041313e-01
1.47575304e-01 7.50044823e-01 -2.94101924e-01 1.36316538e-01
4.98853564e-01 -3.62723470e-01 -1.18295419e+00 2.15203881e-01
7.04849362e-02 -6.64193392e-01 -3.04801017e-01 8.21249843e-01
2.12076996e-02 2.21368134e-01 7.32179761e-01 -1.64009678e+00
-9.33930814e-01 7.30402648e-01 -9.49085578e-02 -1.02164045e-01
7.56092250e-01 5.48843265e-01 1.33627343e+00 -4.97408867e-01
8.01022410e-01 1.60070491e+00 2.48531371e-01 8.66688490e-01
-1.20513189e+00 -1.03333145e-01 2.11124241e-01 3.09716493e-01
-9.84581530e-01 -3.12103271e-01 7.01131821e-01 -6.98073506e-01
1.31971288e+00 5.09191155e-01 7.82732129e-01 1.03012812e+00
2.12426595e-02 1.32426298e+00 9.75177646e-01 -3.68295521e-01
-1.18874749e-02 -5.34955025e-01 5.10227941e-02 4.78099525e-01
-2.01481325e-03 1.71557382e-01 -5.35851903e-02 -1.12272277e-01
1.11974132e+00 3.93320084e-01 -2.29021609e-01 -4.50852960e-01
-1.87092984e+00 3.24166626e-01 7.99889207e-01 3.67395461e-01
-4.70593363e-01 5.51725626e-01 1.86188743e-01 2.89333224e-01
-8.22030604e-02 9.38271165e-01 -6.88302696e-01 -1.84577197e-01
-3.66661756e-04 5.76741517e-01 7.19669521e-01 1.31930077e+00
7.30056822e-01 -3.58015805e-01 -3.54443133e-01 6.48460865e-01
3.18437189e-01 2.70064026e-01 4.54914629e-01 -1.37383187e+00
3.08210701e-01 9.65773880e-01 5.74333608e-01 -4.62360919e-01
-6.19768381e-01 6.28934801e-01 -1.19934417e-01 3.07757735e-01
5.97009540e-01 4.38346028e-01 -9.79600728e-01 1.63939261e+00
3.85778338e-01 -3.81911635e-01 -2.19916888e-02 1.30243313e+00
7.37075269e-01 -1.81498304e-01 4.67033625e-01 2.34094799e-01
1.52107084e+00 -1.01006031e+00 -7.06502914e-01 -5.79656772e-02
1.19834173e+00 -5.67395627e-01 1.55462873e+00 1.86286792e-01
-5.44729412e-01 -4.82512683e-01 -8.70085955e-01 -5.10642409e-01
-3.68604720e-01 2.88712233e-01 1.24815750e+00 1.25881992e-02
-4.27449346e-01 1.05302751e+00 -1.13818085e+00 -4.59797561e-01
3.39546263e-01 4.33780879e-01 -4.98850822e-01 1.72159240e-01
-8.28962266e-01 1.20305467e+00 5.95765173e-01 -1.44823492e-02
-7.19385207e-01 -6.77882656e-02 -9.92842972e-01 -6.80481419e-02
6.35591090e-01 -4.52839434e-01 1.43531990e+00 -7.19293654e-01
-1.40403461e+00 7.56194532e-01 2.68181592e-01 -1.34878874e-01
5.24758160e-01 -6.57834351e-01 -1.66468322e-01 -1.04239890e-02
2.45922506e-01 9.83533263e-01 7.19283938e-01 -1.00386238e+00
-4.94711399e-01 -3.77869636e-01 5.58769166e-01 3.87722909e-01
1.37500212e-01 -4.17151414e-02 4.78764847e-02 -7.52287090e-01
7.38950908e-01 -1.25431001e+00 -7.94948190e-02 4.30240244e-01
-7.81051099e-01 -8.48488986e-01 8.16487432e-01 -3.17553997e-01
7.51921892e-01 -2.06760359e+00 5.36862850e-01 -4.14681733e-01
2.75399774e-01 -1.15696102e-01 -3.26381326e-01 4.91367131e-01
4.99880128e-02 1.06639564e-01 -1.60698265e-01 5.33031709e-02
3.90415847e-01 7.12071896e-01 -2.25095108e-01 3.24578106e-01
3.67223024e-01 1.26909578e+00 -1.26893842e+00 -5.33884883e-01
4.34674203e-01 5.93966357e-02 -6.03398323e-01 4.22018468e-01
-4.96359289e-01 7.67211556e-01 -8.55724096e-01 1.03355968e+00
4.86170053e-02 -1.30742360e-02 3.05712998e-01 -3.14249724e-01
-7.33655393e-02 7.06734180e-01 -1.11290908e+00 1.75546694e+00
-2.62854040e-01 3.90743643e-01 -2.61813045e-01 -5.11540115e-01
6.02783501e-01 3.67668211e-01 5.27983069e-01 -3.27571452e-01
2.37583574e-02 3.83584768e-01 5.01620233e-01 -1.15306568e+00
3.84175420e-01 3.94418128e-02 -1.46868452e-01 5.77837050e-01
6.69965893e-02 -5.91670513e-01 5.64089231e-02 -1.99284405e-01
1.12286174e+00 1.14183748e+00 6.28836513e-01 -3.09632540e-01
1.76355794e-01 -8.79557207e-02 3.13942015e-01 4.47581977e-01
-4.38684553e-01 3.19986761e-01 5.42424083e-01 -5.84937215e-01
-1.09103203e+00 -1.05784357e+00 3.48797664e-02 1.25747013e+00
2.81151891e-01 -2.58137912e-01 -4.26698267e-01 -7.56696343e-01
2.64058381e-01 7.29881406e-01 -6.84795678e-01 1.40732331e-02
-8.43972325e-01 -2.17742592e-01 1.92087874e-01 8.71532619e-01
1.42927855e-01 -1.74943626e+00 -1.25265443e+00 2.60961410e-02
-2.41719425e-01 -1.18444192e+00 -3.31015527e-01 1.90189064e-01
-8.58117044e-01 -1.35027361e+00 -3.90489250e-01 -5.77733636e-01
7.27422357e-01 1.55360073e-01 8.82434547e-01 4.93469596e-01
-3.65543276e-01 3.34874898e-01 -6.60983384e-01 -1.71639487e-01
-4.72680122e-01 2.16289107e-02 3.23431939e-01 -3.82935226e-01
3.81848246e-01 -5.52688360e-01 -6.84741378e-01 4.65864897e-01
-8.71452868e-01 3.98453414e-01 6.22584760e-01 6.90494359e-01
3.38110596e-01 -5.95023036e-01 3.00766289e-01 -2.75409400e-01
3.78693491e-01 -1.40031219e-01 -1.90573692e-01 1.92944944e-01
-1.41365856e-01 3.56807739e-01 1.17856942e-01 -9.22742903e-01
-7.93771923e-01 4.00797188e-01 -1.15280906e-02 -2.42125168e-01
-5.34804225e-01 3.73207740e-02 -4.30334568e-01 1.32807106e-01
5.20612240e-01 -2.19115198e-01 -2.08902121e-01 -7.10889935e-01
6.07465088e-01 7.43328273e-01 4.63535458e-01 -9.46509182e-01
4.31132913e-01 4.65514272e-01 1.13181166e-01 -6.45768523e-01
-6.63882911e-01 -4.41644371e-01 -1.35233128e+00 -9.86992940e-02
1.07238853e+00 -4.09751385e-01 -1.20782161e+00 1.53804600e-01
-1.33817983e+00 -6.24437571e-01 -4.56933141e-01 8.62105489e-01
-1.12243462e+00 4.37720090e-01 -7.62209952e-01 -8.99690926e-01
1.80017084e-01 -1.40916502e+00 1.47175801e+00 -1.17065683e-01
-8.52932632e-01 -3.55270863e-01 -4.67638105e-01 -1.30398059e-02
-1.05493948e-01 4.11470622e-01 8.91699970e-01 -5.82755446e-01
-7.34789550e-01 -1.92573473e-01 -2.97765017e-01 4.66220267e-02
5.15399873e-01 -2.92877406e-01 -6.49413586e-01 7.12790191e-02
-9.43103209e-02 -6.54517412e-01 5.59986234e-01 -2.37296354e-02
9.24855471e-01 -3.75891894e-01 -5.80387115e-01 7.12776333e-02
1.06201482e+00 4.97049056e-02 6.00953698e-01 2.94772953e-01
9.80182528e-01 1.09502864e+00 9.00519729e-01 1.58841312e-01
2.42658019e-01 8.84709001e-01 6.98562145e-01 3.38942796e-01
-8.57358053e-02 -4.71309096e-01 7.67032653e-02 3.62235874e-01
-4.82444584e-01 -6.37125373e-02 -7.47653246e-01 3.94144744e-01
-2.17266107e+00 -6.85859084e-01 -4.17324543e-01 1.87103343e+00
7.93855786e-01 2.34110840e-02 1.07492872e-01 2.22868696e-01
2.54925996e-01 -1.87660798e-01 -6.72377586e-01 -1.34294614e-01
2.36107841e-01 -2.28635803e-01 2.23375306e-01 2.64244556e-01
-1.03282952e+00 1.24905372e+00 6.23835468e+00 3.46353948e-01
-7.21123874e-01 4.62478064e-02 -2.27107406e-01 1.61188155e-01
-8.18920955e-02 2.18346938e-01 -3.32803220e-01 2.17988431e-01
2.06324253e-02 4.74498719e-01 6.01126492e-01 9.39942420e-01
-4.86130118e-02 -2.54393131e-01 -1.83071625e+00 6.44885659e-01
-3.53530228e-01 -5.96666038e-01 -6.99135661e-02 -1.52739529e-02
2.29381278e-01 -1.79547355e-01 -4.65728492e-01 2.04593286e-01
3.81932914e-01 -1.05781841e+00 9.57860053e-01 5.71083367e-01
5.95807195e-01 9.93252471e-02 1.09272063e-01 5.09086847e-01
-1.18435609e+00 -3.06332201e-01 -1.97346538e-01 -6.46460950e-01
2.80032456e-01 -1.52281746e-01 -6.74661040e-01 2.91870296e-01
5.81165969e-01 6.40371561e-01 -3.35688740e-01 6.62640572e-01
-7.98335195e-01 -1.68742761e-01 -4.46571410e-01 -2.91833878e-01
5.97407483e-02 -1.39011487e-01 7.65444279e-01 6.59261465e-01
1.34711444e-01 2.19168857e-01 3.16898406e-01 1.17781198e+00
3.19198042e-01 -1.19057253e-01 -7.35183001e-01 -4.08981234e-01
3.17936718e-01 9.78125215e-01 -9.50498104e-01 -3.12863380e-01
-3.91746163e-01 1.31223261e+00 3.99645537e-01 3.34612101e-01
-5.43838918e-01 -1.58587813e-01 9.41367805e-01 -1.03997663e-02
9.93441120e-02 -7.12244213e-01 -4.41749729e-02 -1.06446338e+00
2.55481869e-01 -4.75760549e-01 -1.08866449e-02 -1.08671534e+00
-9.89345670e-01 3.77905875e-01 2.29059026e-01 -1.40679872e+00
-3.24341297e-01 -1.10845947e+00 -5.88300414e-02 6.13981605e-01
-9.84707892e-01 -1.43775856e+00 -2.00070649e-01 1.12939641e-01
6.81846321e-01 3.64177346e-01 9.10954058e-01 -3.77167314e-02
9.33807269e-02 -3.56947035e-02 -7.80633330e-01 3.02471220e-02
3.79653335e-01 -1.38991523e+00 4.69785780e-01 2.92041481e-01
1.80412367e-01 6.78264618e-01 8.36854875e-01 -7.74581254e-01
-1.33252645e+00 -4.29706633e-01 9.63141263e-01 -1.19191456e+00
7.28195906e-01 -4.42406297e-01 -8.50582302e-01 1.03818083e+00
-2.32550472e-01 1.57223240e-01 2.97913790e-01 9.36134905e-02
-3.56261849e-01 7.34730601e-01 -1.04961812e+00 1.01734984e+00
1.84654355e+00 -5.65729618e-01 -1.11305499e+00 5.52850783e-01
1.01395702e+00 -5.84601760e-01 -9.12844598e-01 6.78860962e-01
8.84884536e-01 -6.43717051e-01 9.44246233e-01 -1.06312990e+00
3.76143545e-01 -3.66148233e-01 -3.80784750e-01 -9.02630985e-01
-3.55771661e-01 -4.72815812e-01 -3.24594826e-01 6.54417813e-01
6.98939804e-03 -3.15287411e-01 2.69499660e-01 7.54529297e-01
-2.41009519e-01 -6.03116512e-01 -8.22906315e-01 -1.04741514e+00
-3.11820686e-01 -5.28355479e-01 1.15159953e+00 9.89665449e-01
4.87413168e-01 1.35698095e-01 -2.16630235e-01 -2.05670401e-01
1.10559821e-01 2.54004091e-01 8.68424833e-01 -1.50613546e+00
-3.02947134e-01 -7.48986423e-01 -3.98614496e-01 -1.38872504e+00
3.85007918e-01 -8.09806049e-01 3.58439445e-01 -1.48174536e+00
5.38969040e-02 -5.23765028e-01 -1.14417821e-01 1.00662267e+00
-2.21243754e-01 8.30468163e-02 3.24890703e-01 7.20519722e-01
-5.03042698e-01 4.94650364e-01 1.85114896e+00 6.29167259e-02
-3.11003149e-01 -2.16339961e-01 -1.74584240e-01 8.23141754e-01
5.25988340e-01 -3.61131370e-01 -7.15651810e-02 -4.16642904e-01
1.89193010e-01 -2.71763578e-02 8.32083106e-01 -6.72935843e-01
-4.11374122e-01 -6.16557360e-01 8.15719143e-02 -1.69278398e-01
3.52399915e-01 -9.59410906e-01 -1.68021783e-01 7.87233591e-01
-5.09602606e-01 -2.14680210e-01 -1.22745447e-01 5.17598271e-01
2.71222621e-01 -1.82896107e-01 9.57390517e-02 -2.96349764e-01
-1.10476875e+00 5.26760146e-02 -2.49642640e-01 -4.04521316e-01
9.51929629e-01 -2.30612963e-01 -1.29280379e-02 7.72458538e-02
-1.07112467e+00 2.52355278e-01 7.12066531e-01 9.84049022e-01
3.57647389e-01 -1.50734401e+00 -3.39430153e-01 2.32023805e-01
4.81054187e-01 -1.27132595e-01 -9.32852179e-02 7.78332174e-01
-3.64586234e-01 2.93184817e-01 -4.41592693e-01 -7.38077879e-01
-9.64592040e-01 6.96798801e-01 1.72355741e-01 2.25951314e-01
-8.71475875e-01 6.76710486e-01 2.68616498e-01 -9.72740948e-01
5.74732549e-04 -9.23562527e-01 -2.68935382e-01 -4.60889697e-01
-1.38642006e-02 3.65611643e-01 -5.24517238e-01 -7.90397346e-01
-2.74785995e-01 2.92879611e-01 5.26018620e-01 -7.37310722e-02
9.87450421e-01 -9.68124345e-02 -2.99270302e-01 6.63804650e-01
8.71397614e-01 -6.97976768e-01 -1.55208302e+00 -8.46151635e-02
5.14023662e-01 -5.17233193e-01 -5.95082045e-01 -6.78372681e-01
-2.02164546e-01 8.79370391e-01 2.71289140e-01 3.52990031e-01
6.72341526e-01 5.55571914e-01 6.37172878e-01 8.16915810e-01
6.41277194e-01 -1.14519215e+00 3.47383678e-01 3.91200542e-01
1.42046463e+00 -1.46947575e+00 1.25721112e-01 -6.34896219e-01
-5.91355205e-01 1.01851487e+00 8.71298611e-01 -1.95504159e-01
3.84657115e-01 -2.77540758e-02 -1.04862049e-01 -2.65014827e-01
-4.69017118e-01 -5.35227060e-01 6.55958772e-01 6.05823457e-01
5.88263988e-01 5.75281084e-01 -6.04689896e-01 5.77155769e-01
-2.34544978e-01 1.60067052e-01 9.45529342e-02 1.25188696e+00
-3.77149671e-01 -1.03712428e+00 -2.14388575e-02 6.15941107e-01
-1.80263728e-01 2.13253394e-01 -4.73655015e-01 9.95712996e-01
1.96040571e-01 6.30832732e-01 1.19540706e-01 -3.04323792e-01
4.43836272e-01 1.61267459e-01 1.06524634e+00 -9.31548417e-01
-2.71967649e-01 3.30189317e-02 4.47547920e-02 -8.80804062e-01
-6.29750550e-01 -7.93534875e-01 -1.78587818e+00 2.67130196e-01
-4.33912694e-01 -3.24612200e-01 5.80518484e-01 1.34213853e+00
2.07415730e-01 4.54401135e-01 -1.35534316e-01 -1.36282825e+00
-8.37881207e-01 -1.31161642e+00 -5.96139610e-01 8.95449936e-01
4.16976929e-01 -1.04824638e+00 -1.39205664e-01 -2.00595940e-03] | [5.1043701171875, -0.002445972990244627] |
d0710a92-9624-4fc9-aa67-50510011690d | conmix-for-source-free-single-and-multi | 2211.03876 | null | https://arxiv.org/abs/2211.03876v1 | https://arxiv.org/pdf/2211.03876v1.pdf | CoNMix for Source-free Single and Multi-target Domain Adaptation | This work introduces the novel task of Source-free Multi-target Domain Adaptation and proposes adaptation framework comprising of \textbf{Co}nsistency with \textbf{N}uclear-Norm Maximization and \textbf{Mix}Up knowledge distillation (\textit{CoNMix}) as a solution to this problem. The main motive of this work is to solve for Single and Multi target Domain Adaptation (SMTDA) for the source-free paradigm, which enforces a constraint where the labeled source data is not available during target adaptation due to various privacy-related restrictions on data sharing. The source-free approach leverages target pseudo labels, which can be noisy, to improve the target adaptation. We introduce consistency between label preserving augmentations and utilize pseudo label refinement methods to reduce noisy pseudo labels. Further, we propose novel MixUp Knowledge Distillation (MKD) for better generalization on multiple target domains using various source-free STDA models. We also show that the Vision Transformer (VT) backbone gives better feature representation with improved domain transferability and class discriminability. Our proposed framework achieves the state-of-the-art (SOTA) results in various paradigms of source-free STDA and MTDA settings on popular domain adaptation datasets like Office-Home, Office-Caltech, and DomainNet. Project Page: https://sites.google.com/view/conmix-vcl | ['Anirban Chakraborty', 'Himanshu Patil', 'Rohit Lal', 'Vikash Kumar'] | 2022-11-07 | null | null | null | null | ['multi-target-domain-adaptation'] | ['computer-vision'] | [ 3.27853739e-01 -1.98117644e-01 -3.97501856e-01 -6.76842153e-01
-1.08547282e+00 -8.15504253e-01 6.34417772e-01 -2.23664105e-01
-5.83208859e-01 1.06787729e+00 -1.32739032e-02 -8.72724950e-02
-1.76288143e-01 -5.50305724e-01 -7.22525954e-01 -8.14831734e-01
4.37474847e-01 6.10390306e-01 -3.39077823e-02 -1.71952873e-01
-2.98347414e-01 2.47139603e-01 -1.20115519e+00 2.20122620e-01
9.67300236e-01 1.26048470e+00 8.74157920e-02 3.52593064e-01
1.23620823e-01 6.01899564e-01 -4.24408853e-01 -6.42324805e-01
8.52030158e-01 -2.58325636e-01 -9.14547801e-01 -2.86457501e-03
6.77331448e-01 -3.35692316e-01 -3.52684677e-01 1.17052078e+00
6.96629703e-01 4.55627531e-01 9.65063691e-01 -1.84382427e+00
-1.09952354e+00 1.31269231e-01 -7.82106459e-01 1.60924926e-01
-1.61569595e-01 -6.24398421e-03 8.26880395e-01 -8.59014988e-01
6.33843720e-01 1.23511553e+00 5.14527619e-01 7.00968802e-01
-1.27031064e+00 -1.19195104e+00 3.41317713e-01 2.19573453e-01
-1.45401585e+00 -5.26106238e-01 4.64068115e-01 -4.14669663e-01
5.88626683e-01 4.43236232e-02 -1.30419284e-01 1.64878035e+00
-2.66817868e-01 8.21997464e-01 1.48643482e+00 -4.21079129e-01
2.24828750e-01 5.42544603e-01 2.27096528e-01 4.65403259e-01
9.36645269e-02 2.17668146e-01 -5.37107289e-01 -3.36825043e-01
7.60587931e-01 1.29546538e-01 -2.96737880e-01 -6.58111095e-01
-1.27754712e+00 9.37472761e-01 2.43561983e-01 -1.53625250e-01
-1.80768892e-01 -2.43268654e-01 4.67750549e-01 6.29017293e-01
4.69926059e-01 2.23321065e-01 -8.84833694e-01 2.84130394e-01
-8.11111510e-01 2.95984000e-01 5.80742121e-01 1.56010234e+00
9.57764208e-01 7.31492043e-02 -4.00768638e-01 1.22260082e+00
3.16791058e-01 9.05354679e-01 7.17656136e-01 -1.12028110e+00
5.79784393e-01 2.62860060e-01 5.96426614e-02 -4.54709798e-01
-5.66493496e-02 -5.99623680e-01 -9.21797872e-01 3.33368093e-01
4.58722711e-01 -3.53375524e-01 -1.24383426e+00 2.08463526e+00
4.43683118e-01 4.68945175e-01 3.81072938e-01 7.69721031e-01
5.67672610e-01 3.09035450e-01 3.13098997e-01 -1.45907968e-01
1.32873809e+00 -1.14717829e+00 -6.10444844e-01 -2.17680603e-01
6.79059923e-01 -5.72582066e-01 9.75990713e-01 2.86698788e-01
-5.09226620e-01 -4.88858879e-01 -9.84071255e-01 -1.53025895e-01
-6.41770244e-01 1.34151667e-01 4.53348011e-01 7.33514965e-01
-7.53962874e-01 2.18983158e-01 -5.34475505e-01 -5.06274104e-01
9.08813298e-01 3.66700262e-01 -7.65309453e-01 -4.72037077e-01
-1.16465151e+00 8.06477547e-01 5.55229783e-01 -4.79198605e-01
-1.12859464e+00 -9.62080061e-01 -7.68712401e-01 -3.27592909e-01
6.26867115e-01 -9.36323225e-01 1.19857383e+00 -1.07070136e+00
-1.43679535e+00 1.10044360e+00 -1.94895610e-01 -6.18246734e-01
5.99041641e-01 -2.24553823e-01 -6.58178389e-01 -1.26595750e-01
4.14003938e-01 7.80983865e-01 9.96616483e-01 -1.20016825e+00
-8.35298896e-01 -5.02575338e-01 -1.93524748e-01 5.14147401e-01
-4.61914480e-01 -1.94377825e-01 -2.96528220e-01 -9.03534532e-01
-2.12393716e-01 -1.07607126e+00 -7.68454652e-03 4.86267433e-02
-3.94394338e-01 -1.90480519e-02 1.07806182e+00 -5.61920345e-01
8.71559560e-01 -2.23435688e+00 -5.74393980e-02 2.69443244e-01
1.38697430e-01 4.13973063e-01 -3.32695305e-01 -5.61954603e-02
-1.31243363e-01 -9.48720351e-02 -4.26893175e-01 -5.64691305e-01
1.31123573e-01 4.96107966e-01 -4.20553476e-01 4.66391593e-01
-1.42154455e-01 7.50494599e-01 -6.06411994e-01 -4.62348014e-01
1.92505985e-01 3.83377731e-01 -4.34260339e-01 1.29807472e-01
-2.04363018e-01 7.35683441e-01 -6.11670315e-01 8.70684266e-01
1.11007011e+00 -2.71977186e-01 -1.71869606e-01 -3.07624459e-01
3.12540501e-01 -2.74480224e-01 -1.47433209e+00 2.04904699e+00
-2.29453936e-01 1.49085373e-01 1.79387584e-01 -1.03235650e+00
7.80109107e-01 2.98933238e-01 5.16502380e-01 -5.59400916e-01
1.51881889e-01 2.11748302e-01 -4.17080760e-01 -1.63299412e-01
2.88739532e-01 -1.05665267e-01 -9.26691145e-02 2.37094387e-01
5.58642447e-01 3.41323256e-01 -1.41021371e-01 1.44559279e-01
9.50294316e-01 2.49304280e-01 4.08255875e-01 -1.67834148e-01
4.32111651e-01 3.38920541e-02 8.14532697e-01 8.58534813e-01
-6.39586866e-01 6.79614604e-01 -9.29523483e-02 1.59265734e-02
-8.49414289e-01 -1.16906416e+00 -1.80570215e-01 1.57199919e+00
1.17362693e-01 5.66856787e-02 -5.52573264e-01 -1.16208327e+00
1.79667145e-01 7.37029076e-01 -7.15876698e-01 -1.26680657e-01
-2.20057636e-01 -7.61146963e-01 9.15818095e-01 4.07200485e-01
9.18586195e-01 -6.25305593e-01 4.25731763e-02 -5.48675321e-02
-3.19191813e-01 -1.23597169e+00 -7.19926655e-01 4.09261525e-01
-5.45922339e-01 -9.94704664e-01 -1.12115288e+00 -6.46918952e-01
4.36175108e-01 1.94967985e-01 7.02798903e-01 -9.73362565e-01
2.70354748e-02 5.22504985e-01 -3.21535408e-01 -5.84794343e-01
-8.18802565e-02 1.11137904e-01 4.13529426e-01 2.22202763e-01
6.57849431e-01 -7.39714861e-01 -5.33425331e-01 4.48501170e-01
-7.33233571e-01 -2.95184225e-01 5.20065844e-01 8.73487532e-01
1.05667937e+00 -1.60381004e-01 7.52449334e-01 -1.10541010e+00
6.52686357e-01 -8.32818449e-01 -3.85241807e-01 4.73288506e-01
-1.00466275e+00 -7.59962993e-03 5.35131633e-01 -7.82881439e-01
-1.27675128e+00 1.08614996e-01 2.58862048e-01 -1.11544394e+00
-5.90770125e-01 -3.57926078e-02 -4.67435986e-01 -1.71219960e-01
9.10479546e-01 3.35526347e-01 -1.25440866e-01 -5.81336677e-01
6.14777565e-01 8.10365736e-01 5.85284889e-01 -7.93897808e-01
8.99626434e-01 5.20703197e-01 -1.55303985e-01 -5.19258082e-01
-9.90993202e-01 -7.08012938e-01 -6.67936146e-01 3.51727515e-01
9.03394222e-01 -1.44485581e+00 -3.24118435e-01 5.89710832e-01
-7.52290249e-01 -1.98212445e-01 -4.38064277e-01 5.97543240e-01
-6.73854232e-01 2.75162786e-01 -6.88336119e-02 -4.71254408e-01
-4.33767587e-01 -1.06259120e+00 8.94293785e-01 3.98230523e-01
1.48600623e-01 -1.00949895e+00 2.91306078e-02 5.66313565e-01
4.19206828e-01 4.11132872e-01 5.93562245e-01 -1.30286980e+00
-3.35780799e-01 1.96332768e-01 -4.45191473e-01 8.03277433e-01
2.77522206e-01 -9.14908051e-01 -1.33906531e+00 -6.18265092e-01
-7.94586688e-02 -5.47262967e-01 8.04050982e-01 4.58429277e-01
9.68401194e-01 -4.03391391e-01 -4.10042316e-01 1.00611198e+00
1.51002729e+00 1.21356927e-01 2.37248242e-01 3.75302166e-01
9.69658911e-01 2.46802300e-01 8.31315398e-01 5.29401779e-01
6.03085816e-01 5.93663394e-01 1.89762548e-01 -2.36722324e-02
-3.95949185e-01 -2.20023796e-01 3.32274914e-01 3.37865353e-01
1.11895755e-01 -2.43462279e-01 -7.24703670e-01 7.92221844e-01
-1.92643178e+00 -7.91684747e-01 9.67038274e-02 2.14534521e+00
9.83567536e-01 -3.87925833e-01 1.83584392e-01 -4.39491481e-01
6.72618866e-01 -6.72067888e-03 -1.05012083e+00 -1.24665292e-03
-4.43759978e-01 2.94630527e-01 1.10221231e+00 3.64570379e-01
-1.39797938e+00 9.64498341e-01 5.15983295e+00 1.27633917e+00
-8.74027967e-01 7.22332537e-01 4.91041541e-01 -1.94205046e-01
8.62257183e-02 -1.79089800e-01 -1.04911125e+00 3.91825974e-01
8.05486500e-01 -3.83845896e-01 5.23689985e-01 1.21925724e+00
-2.08349437e-01 1.10775277e-01 -1.12330735e+00 1.15461755e+00
9.23097879e-02 -9.90418494e-01 -1.32973790e-02 1.93755224e-01
7.89273739e-01 3.35162371e-01 3.90473962e-01 7.01494277e-01
7.65176594e-01 -7.73796678e-01 4.64452803e-01 1.61269858e-01
1.34817088e+00 -6.23569310e-01 4.55695897e-01 2.52230167e-01
-1.02894926e+00 -2.07113534e-01 -4.65206236e-01 4.30159986e-01
-1.07920624e-01 4.31091666e-01 -8.96788597e-01 8.54741156e-01
8.36264491e-01 7.41538882e-01 -4.78488833e-01 7.57093370e-01
8.26156363e-02 5.42720795e-01 -4.74558145e-01 6.50447130e-01
4.61613275e-02 -6.10125400e-02 6.47763312e-01 1.28843570e+00
3.44081551e-01 6.70531914e-02 4.80436057e-01 6.77859128e-01
-3.86827976e-01 7.74290860e-02 -6.78057313e-01 3.40650201e-01
8.57378542e-01 9.99744952e-01 -1.78697661e-01 -2.39303440e-01
-3.53963345e-01 1.38831401e+00 2.65619755e-01 7.72440553e-01
-8.07932138e-01 -2.12441161e-01 1.12916183e+00 -7.30014220e-02
3.88120770e-01 4.27854173e-02 -2.42279962e-01 -1.36435926e+00
-6.97312877e-02 -1.05276060e+00 8.87615502e-01 -5.47033310e-01
-1.72218132e+00 3.23936701e-01 2.95061618e-01 -1.24177217e+00
8.26921016e-02 -4.45438802e-01 -1.64629787e-01 1.24373507e+00
-1.97846735e+00 -1.69735134e+00 -1.28516153e-01 1.48679197e+00
5.12575805e-01 -6.39658391e-01 1.03570902e+00 6.05122983e-01
-4.95434105e-01 1.20418477e+00 5.45004070e-01 1.79391384e-01
1.46637273e+00 -1.20422292e+00 -5.46932481e-02 7.94098854e-01
3.05089150e-02 4.92658049e-01 3.62842649e-01 -6.48470402e-01
-9.00913358e-01 -1.65780294e+00 4.24505711e-01 -5.92408061e-01
4.05996263e-01 -2.75783807e-01 -7.69418895e-01 1.15159094e+00
1.01653993e-01 4.15312111e-01 8.86724114e-01 2.14356370e-02
-9.27792549e-01 -1.96844772e-01 -1.74202836e+00 1.85554296e-01
9.50356781e-01 -5.30716062e-01 -4.54066515e-01 3.73448402e-01
9.35360074e-01 -2.69217402e-01 -1.08061957e+00 3.83030742e-01
2.81671464e-01 -5.92423141e-01 1.07232213e+00 -4.98345971e-01
-1.26595035e-01 -4.73683596e-01 -6.70334220e-01 -1.31656075e+00
-4.11771178e-01 -5.09480298e-01 -1.86332360e-01 1.55102944e+00
3.89237463e-01 -8.61096382e-01 6.49712741e-01 5.99083424e-01
6.28095865e-02 -2.60698617e-01 -1.14918935e+00 -1.09270263e+00
3.27864528e-01 -2.65462399e-01 5.77602386e-01 1.62106359e+00
-6.15149856e-01 2.51105011e-01 -6.48039937e-01 4.95728165e-01
1.03971303e+00 -2.34093517e-01 6.57974839e-01 -1.21101642e+00
-2.72732764e-01 9.49929506e-02 -3.83459926e-02 -7.98805892e-01
2.58226097e-01 -1.18073356e+00 -3.81612867e-01 -1.14609337e+00
1.10874772e-01 -6.84435129e-01 -8.09815168e-01 1.01274836e+00
4.30495031e-02 1.28119379e-01 3.26720357e-01 4.58796799e-01
-7.25616336e-01 6.17281914e-01 1.13558543e+00 -1.10637404e-01
-1.23103738e-01 2.48331372e-02 -9.40059543e-01 5.13357401e-01
7.82560170e-01 -7.91971028e-01 -7.81246901e-01 -4.85075980e-01
-5.59572697e-01 -3.14382970e-01 3.59853268e-01 -9.32542980e-01
2.05783814e-01 -2.69623667e-01 4.39110160e-01 -2.79552966e-01
4.65958267e-01 -1.03397608e+00 2.28651129e-02 -1.62138313e-01
-4.82081532e-01 -2.70633042e-01 2.83578932e-01 9.58244324e-01
-5.08981571e-02 6.31126463e-02 1.19748974e+00 -1.51038557e-01
-9.50231194e-01 6.15298688e-01 3.27927321e-01 3.55286241e-01
1.27226377e+00 -2.66468644e-01 -4.86529380e-01 -6.41274750e-02
-8.48182082e-01 4.44147378e-01 3.28464419e-01 5.07331550e-01
3.55058163e-01 -1.58685791e+00 -8.49363327e-01 1.82556868e-01
4.33639020e-01 1.06346689e-03 4.03144687e-01 6.39370143e-01
1.94850773e-01 2.28383094e-01 -3.41764778e-01 -4.68042642e-01
-1.30024302e+00 5.41739225e-01 4.32403356e-01 -3.60410422e-01
-3.11009914e-01 1.22565722e+00 5.76935649e-01 -8.91368449e-01
3.14210624e-01 1.39939478e-02 -4.51076925e-02 -4.98487242e-03
4.49651361e-01 4.75382924e-01 -1.11957431e-01 -7.27795422e-01
-4.03726935e-01 3.80466312e-01 -5.67653477e-01 -1.21273838e-01
1.04592347e+00 -4.57250535e-01 3.97916853e-01 1.89022735e-01
1.35001373e+00 -3.51757765e-01 -1.38728023e+00 -7.83211768e-01
-2.49427870e-01 -4.36462194e-01 -7.56790675e-03 -1.45638394e+00
-9.18597877e-01 6.00687146e-01 1.21340775e+00 -5.40248334e-01
1.21711254e+00 -1.68121889e-01 8.46742809e-01 4.29451078e-01
3.44017088e-01 -1.30052447e+00 -1.84619516e-01 5.00973642e-01
6.25493944e-01 -1.72069120e+00 -1.57679528e-01 -1.85331628e-01
-1.10339069e+00 4.53947276e-01 7.55447388e-01 1.30507082e-01
8.32033813e-01 4.36940007e-02 1.29905075e-01 1.28186256e-01
-4.88304853e-01 -3.87220979e-01 2.03905284e-01 1.31577611e+00
-2.12892383e-01 2.26247683e-01 1.94613695e-01 8.49518955e-01
1.32163957e-01 1.73576340e-01 7.70156160e-02 8.64882588e-01
-1.75151363e-01 -1.35298145e+00 -5.01934886e-01 3.48343104e-01
-5.60996473e-01 -1.50931925e-01 -1.68333024e-01 7.23759353e-01
5.22695661e-01 9.49362874e-01 -4.48987603e-01 -3.26587826e-01
4.06243026e-01 4.35256302e-01 3.41045737e-01 -4.84073639e-01
-4.03027236e-01 1.56032238e-02 -4.75168303e-02 -4.67615187e-01
-4.36697990e-01 -7.25135982e-01 -1.07407069e+00 -1.61226511e-01
-1.65818274e-01 -1.26519695e-01 5.47789514e-01 8.11908126e-01
6.98478222e-01 3.35823774e-01 4.01532322e-01 -2.25455776e-01
-6.95990503e-01 -8.73689294e-01 -8.05527449e-01 6.52875006e-01
4.61103171e-01 -8.00633907e-01 -2.08387554e-01 3.37604314e-01] | [10.381721496582031, 3.138502359390259] |
5a70b2a9-1fe8-495c-9110-250fc55168e7 | adaptive-period-embedding-for-representing | 1906.09447 | null | https://arxiv.org/abs/1906.09447v1 | https://arxiv.org/pdf/1906.09447v1.pdf | Adaptive Period Embedding for Representing Oriented Objects in Aerial Images | We propose a novel method for representing oriented objects in aerial images named Adaptive Period Embedding (APE). While traditional object detection methods represent object with horizontal bounding boxes, the objects in aerial images are oritented. Calculating the angle of object is an yet challenging task. While almost all previous object detectors for aerial images directly regress the angle of objects, they use complex rules to calculate the angle, and their performance is limited by the rule design. In contrast, our method is based on the angular periodicity of oriented objects. The angle is represented by two two-dimensional periodic vectors whose periods are different, the vector is continuous as shape changes. The label generation rule is more simple and reasonable compared with previous methods. The proposed method is general and can be applied to other oriented detector. Besides, we propose a novel IoU calculation method for long objects named length independent IoU (LIIoU). We intercept part of the long side of the target box to get the maximum IoU between the proposed box and the intercepted target box. Thereby, some long boxes will have corresponding positive samples. Our method reaches the 1st place of DOAI2019 competition task1 (oriented object) held in workshop on Detecting Objects in Aerial Images in conjunction with IEEE CVPR 2019. | ['Jun Du', 'Yixing Zhu', 'Xueqing Wu'] | 2019-06-22 | null | null | null | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [ 3.62259865e-01 -2.17259064e-01 -2.60765463e-01 -1.11213282e-01
2.25921690e-01 -6.68635368e-01 3.62908751e-01 -1.50654810e-02
-4.29422468e-01 2.69480914e-01 -1.28644571e-01 1.89473480e-02
-1.80413052e-01 -9.72063661e-01 -5.13875663e-01 -7.75128901e-01
-2.05996871e-01 7.06282109e-02 8.24379265e-01 -2.16263324e-01
1.24316551e-01 4.98689890e-01 -1.44824433e+00 3.59105885e-01
3.18477809e-01 1.12455118e+00 -4.71913815e-02 6.81665242e-01
-7.87239987e-03 2.53231287e-01 -9.36838508e-01 8.06913003e-02
6.31797493e-01 -3.18115145e-01 -7.00472713e-01 3.65283668e-01
3.16970229e-01 -4.04481679e-01 -1.52606025e-01 1.21364856e+00
3.61628443e-01 -1.35095427e-02 1.12659991e+00 -1.44402599e+00
-7.62619853e-01 7.33497679e-01 -9.18372869e-01 3.60203654e-01
1.12237282e-01 -1.52149931e-01 9.20078576e-01 -8.08499873e-01
4.14108604e-01 1.15149307e+00 7.88803756e-01 3.66572112e-01
-8.89464378e-01 -4.91730571e-01 5.62860548e-01 1.88522458e-01
-1.52947819e+00 3.10705632e-01 7.37600863e-01 -5.64130366e-01
6.06363237e-01 6.18582904e-01 8.18450630e-01 7.05037117e-01
7.55478442e-02 9.06770468e-01 9.39417064e-01 -5.10389507e-01
-4.80128042e-02 2.77548969e-01 4.37242955e-01 5.62119186e-01
5.04690349e-01 -8.52120051e-04 2.32453465e-01 7.20214890e-03
5.39825678e-01 3.30256253e-01 -3.45531523e-01 -2.19712406e-01
-1.36817086e+00 7.12040424e-01 8.86676967e-01 3.34240675e-01
1.94268357e-02 -2.29627099e-02 3.33948165e-01 -8.62108395e-02
1.69655651e-01 3.60876620e-01 -2.66110897e-01 6.43007755e-01
-7.08621860e-01 3.19966614e-01 3.85363579e-01 1.17476130e+00
3.45773578e-01 -4.46410924e-02 -4.41592216e-01 5.63378513e-01
2.86654472e-01 4.98832494e-01 5.78779876e-01 -1.23148166e-01
3.27243090e-01 9.54537988e-01 1.43305063e-01 -1.41492081e+00
-5.90003014e-01 -1.89747974e-01 -7.79003441e-01 2.96137094e-01
3.83420646e-01 -1.11914299e-01 -8.85257125e-01 1.09234977e+00
5.07359445e-01 -7.03218952e-02 -4.08206470e-02 1.03114831e+00
1.01899064e+00 1.15852118e+00 -1.70971200e-01 -1.66425973e-01
1.70445538e+00 -1.00110912e+00 -6.06373787e-01 -1.38395265e-01
4.50539380e-01 -7.66691446e-01 8.14038396e-01 4.65594411e-01
-5.28980076e-01 -6.40374720e-01 -1.40342057e+00 2.65288174e-01
-7.36303866e-01 7.83563256e-01 5.20439267e-01 4.58655506e-01
-4.68455225e-01 4.31843907e-01 -4.80376303e-01 -2.92533129e-01
2.84077674e-01 4.87998247e-01 -5.25974557e-02 4.52139556e-01
-9.71723378e-01 4.77302432e-01 8.29068422e-01 2.85795033e-01
-3.55402142e-01 -1.74011037e-01 -7.09754348e-01 -1.11202106e-01
5.02216756e-01 -4.49762270e-02 8.92470539e-01 -1.00321877e+00
-1.12097394e+00 5.10658503e-01 2.97584176e-01 -4.70555633e-01
4.69395012e-01 -5.21140248e-02 -4.84257847e-01 -1.87655922e-03
-1.99927762e-02 8.00211608e-01 1.23800361e+00 -9.63492036e-01
-8.26584160e-01 -3.37297440e-01 1.65598661e-01 2.21148431e-01
-8.91417980e-01 3.07682194e-02 -1.86874956e-01 -6.86887205e-01
3.56790513e-01 -8.83973241e-01 -2.63995230e-02 4.04898196e-01
-5.83870113e-01 -5.41377544e-01 1.50107849e+00 -4.12209779e-01
1.65038466e+00 -2.25827622e+00 -1.49658680e-01 5.09957783e-02
7.49687254e-02 6.51415408e-01 -2.72175893e-02 8.00553933e-02
-1.39461607e-01 2.42961258e-01 -2.47984990e-01 5.21001756e-01
-9.09911394e-02 -3.23958218e-01 -4.45712984e-01 4.24694717e-01
4.47106987e-01 5.51609576e-01 -8.70427072e-01 -7.11331010e-01
1.24134660e-01 1.86601743e-01 -2.96894073e-01 -5.01123033e-02
-8.11068341e-03 -3.43297154e-01 -6.02440655e-01 7.09591210e-01
9.67050016e-01 -8.10268298e-02 -8.09952691e-02 -6.59989893e-01
-4.73049492e-01 -1.31663144e-01 -1.57550192e+00 5.83590269e-01
-2.31566709e-02 6.71947837e-01 -5.54568470e-01 -8.12613130e-01
1.28688288e+00 1.91421777e-01 3.79512846e-01 -1.08961808e-02
2.31764764e-01 -1.56786785e-01 6.58166781e-02 -5.20573616e-01
6.01710498e-01 2.23992333e-01 1.17176235e-01 -4.37088721e-02
-4.22348171e-01 -3.64656180e-01 3.52260888e-01 -2.81925410e-01
6.68726206e-01 2.02514216e-01 8.80257785e-01 -2.31216133e-01
6.22099996e-01 2.58646794e-02 5.15281320e-01 4.69485044e-01
-3.55495811e-01 5.94436347e-01 5.26464939e-01 -8.72814059e-01
-9.47967172e-01 -7.44376063e-01 -6.51802599e-01 6.26433015e-01
5.49163878e-01 -4.19894308e-01 -4.92824435e-01 -1.08392572e+00
-4.79214229e-02 2.59134203e-01 -8.45150828e-01 -1.71652753e-02
-5.15037775e-01 -9.91108596e-01 4.00721520e-01 6.87119067e-01
8.64152789e-01 -1.11321712e+00 -9.36240911e-01 1.16799437e-01
2.04985682e-02 -1.08274281e+00 -5.84913790e-01 9.02209133e-02
-8.16629350e-01 -1.16692090e+00 -9.37384009e-01 -9.03197885e-01
8.58477473e-01 5.67183793e-01 5.11023819e-01 -1.17085397e-01
-5.35391748e-01 -7.49833658e-02 -5.79511940e-01 -9.54136252e-01
1.50553271e-01 -1.11180417e-01 2.77991462e-02 2.33142320e-02
4.71227169e-01 7.82135725e-02 -6.59952462e-01 8.77604902e-01
-9.63052928e-01 -1.60564110e-01 5.74124455e-01 8.03657830e-01
6.45945907e-01 3.25279206e-01 1.62805319e-01 -4.16455805e-01
2.89330333e-01 -9.89402682e-02 -1.04213583e+00 3.09131175e-01
-3.56275469e-01 -9.62363407e-02 8.53650689e-01 -8.82963777e-01
-3.41282308e-01 2.09522232e-01 3.15583318e-01 -3.08756322e-01
-2.06719711e-01 2.69871429e-02 -7.48446435e-02 -1.16669074e-01
6.99454606e-01 -8.11259728e-03 -4.88709986e-01 9.08504706e-03
-1.05422005e-01 7.88939238e-01 1.29695401e-01 4.35216054e-02
8.98292065e-01 5.66092789e-01 2.39990890e-01 -1.16711879e+00
-6.62860751e-01 -5.73247552e-01 -6.99709952e-01 -2.90707082e-01
8.46529722e-01 -4.81088787e-01 -7.31106043e-01 3.05384666e-01
-1.28872609e+00 2.46392578e-01 -2.90418625e-01 6.55319571e-01
3.38842608e-02 4.95396852e-01 -8.02856833e-02 -1.03582859e+00
-3.88130188e-01 -1.07640815e+00 1.04855621e+00 3.22597563e-01
3.39293890e-02 -5.94863117e-01 -3.50563452e-02 -3.60342294e-01
-5.18646976e-03 4.50107396e-01 7.80016243e-01 -5.58079660e-01
-3.85344088e-01 -6.45171046e-01 -3.58640879e-01 5.20382285e-01
1.23911016e-01 4.64920223e-01 -7.11755395e-01 -1.52951777e-01
-1.83222935e-01 -2.72786133e-02 9.41576421e-01 5.12577057e-01
1.22713971e+00 -5.91491342e-01 -7.13237286e-01 3.66690457e-01
1.28378141e+00 4.96436924e-01 4.77064282e-01 4.53836411e-01
5.34289837e-01 5.25930047e-01 9.11045432e-01 4.69190031e-01
-1.61538661e-01 9.80929255e-01 5.06168365e-01 -2.55390331e-02
5.50065674e-02 7.44938850e-04 5.49683928e-01 4.80556041e-01
-3.01703453e-01 -5.61397493e-01 -6.99427962e-01 4.28786695e-01
-1.62290859e+00 -9.81467187e-01 -5.19671321e-01 2.22453856e+00
4.04462785e-01 3.92075658e-01 3.25741768e-01 5.37249863e-01
7.89862692e-01 3.12611371e-01 -3.63328278e-01 -3.08884919e-01
-1.23591118e-01 -3.12238336e-01 7.41118908e-01 -4.70378175e-02
-1.70783234e+00 7.49563634e-01 6.25943041e+00 8.80909920e-01
-1.16595960e+00 -3.72769296e-01 2.85140872e-01 2.20259339e-01
2.63740838e-01 -8.36498067e-02 -1.50677252e+00 4.85262811e-01
2.00445831e-01 -1.36502571e-02 -8.39763135e-02 1.23466909e+00
-6.82837814e-02 2.45528738e-03 -8.94208908e-01 8.02176595e-01
1.78249642e-01 -8.55286121e-01 3.06576699e-01 -4.54975516e-02
5.99245906e-01 -6.34725034e-01 2.62642771e-01 1.85834199e-01
-3.38715017e-01 -7.31416225e-01 6.46153569e-01 1.61937848e-01
6.46510184e-01 -5.52873194e-01 9.18434262e-01 2.60535955e-01
-1.75867009e+00 -3.82541478e-01 -9.51646984e-01 4.47039120e-02
-2.73864746e-01 3.52505684e-01 -1.05033195e+00 4.50329304e-01
8.31055284e-01 7.13261962e-01 -7.40166426e-01 1.31921840e+00
-1.68620840e-01 5.79091489e-01 -5.06273627e-01 -6.04822636e-01
5.39537609e-01 -2.66983151e-01 8.77835870e-01 1.06023407e+00
4.71839219e-01 1.41205341e-01 3.45748901e-01 7.23337710e-01
1.18210815e-01 3.22589755e-01 -9.71965432e-01 -9.48022082e-02
2.05732942e-01 1.30890107e+00 -1.19572151e+00 -3.88005644e-01
-2.84342170e-01 8.28590095e-01 -2.38772660e-01 4.13985588e-02
-1.09952974e+00 -8.65338624e-01 -8.58592466e-02 3.77665497e-02
5.95746756e-01 -8.42032060e-02 -3.11184302e-02 -7.14599431e-01
2.02721804e-01 -6.77516818e-01 4.80305731e-01 -5.64316332e-01
-9.76621568e-01 8.63528073e-01 3.00958812e-01 -1.90302777e+00
2.89217472e-01 -1.03905523e+00 -6.68945611e-01 2.65104324e-01
-1.15440321e+00 -1.16249716e+00 -4.93021876e-01 2.24262103e-01
8.13448906e-01 -4.64868844e-02 6.33985102e-01 7.82646611e-02
-6.48281693e-01 5.48460722e-01 3.06922197e-02 4.32349175e-01
4.85493332e-01 -1.06138778e+00 2.92436123e-01 8.10510755e-01
3.71114433e-01 3.39072108e-01 6.52636468e-01 -5.88846326e-01
-8.19808185e-01 -1.24826944e+00 6.50918484e-01 -3.15250278e-01
4.11658615e-01 -3.59175175e-01 -6.74044013e-01 5.93308449e-01
-2.90797185e-02 1.82154670e-01 3.05796504e-01 -3.76418293e-01
-1.98204249e-01 -9.74672362e-02 -8.63313258e-01 5.89682698e-01
7.83725142e-01 3.06551903e-01 -6.34704351e-01 7.11138666e-01
7.25437641e-01 -2.67889917e-01 -3.67041975e-01 8.47554803e-01
6.21974945e-01 -7.30670273e-01 1.10532618e+00 -6.33320630e-01
2.79358268e-01 -8.93409252e-01 2.72272676e-02 -8.59490275e-01
-5.36033750e-01 -2.63865978e-01 4.28783111e-02 8.89507949e-01
2.97135472e-01 -3.86041045e-01 5.80907464e-01 -2.41459802e-01
-1.71276566e-03 -9.80101109e-01 -5.27832150e-01 -1.29448581e+00
-4.00174141e-01 1.56892873e-02 6.30954862e-01 6.72312975e-01
-1.33756399e-01 2.58084863e-01 -3.89309078e-01 7.16920257e-01
3.37246090e-01 1.97876215e-01 6.14901602e-01 -1.18627691e+00
-1.60332352e-01 -4.54818845e-01 -9.80517089e-01 -1.22271574e+00
-3.39759767e-01 -8.21769714e-01 -7.58725181e-02 -1.38436973e+00
-2.92584281e-02 -2.79245704e-01 -6.48177937e-02 4.10414398e-01
-9.61962864e-02 4.13513869e-01 2.67841935e-01 6.24869019e-02
-4.56835449e-01 4.87918109e-01 1.15999985e+00 -5.24769843e-01
-5.11834860e-01 3.06450754e-01 6.91203251e-02 1.18861830e+00
6.82816446e-01 -5.53052187e-01 -1.58890471e-01 -2.77671784e-01
1.05105480e-03 -4.03309852e-01 4.02088493e-01 -1.33101547e+00
9.50948521e-02 -1.58008769e-01 7.44440556e-01 -1.13070631e+00
2.27702543e-01 -1.07506382e+00 -3.51343483e-01 6.62959695e-01
-1.33461311e-01 1.29297659e-01 4.26631123e-01 4.64720488e-01
-4.11626920e-02 -7.70347595e-01 8.73624265e-01 -2.03377213e-02
-8.17458689e-01 4.42486107e-01 -4.01549578e-01 -2.53244251e-01
1.61582971e+00 -6.33922994e-01 -2.12100357e-01 1.74944833e-01
-5.68798602e-01 1.39820531e-01 2.47068256e-02 3.42800468e-01
8.37469518e-01 -1.40102577e+00 -4.68080431e-01 2.29228571e-01
3.18721801e-01 3.24114114e-02 -3.87235358e-02 7.13386536e-01
-6.49921656e-01 4.49557453e-01 -2.94005811e-01 -7.54403949e-01
-1.58033860e+00 8.22193265e-01 2.77677238e-01 -1.55153260e-01
-6.32638752e-01 9.07348573e-01 4.26423222e-01 -2.92202774e-02
1.84319392e-01 -6.74911082e-01 -7.87915528e-01 3.60926300e-01
6.79154634e-01 4.34342116e-01 -2.70734489e-01 -4.96454775e-01
-3.48444313e-01 9.62417424e-01 -1.21111199e-01 1.93221182e-01
1.01665854e+00 3.89188588e-01 9.92372110e-02 5.55345178e-01
1.11550725e+00 -3.34220976e-02 -9.86189663e-01 5.30425347e-02
-1.05847776e-01 -4.74590242e-01 -1.46971881e-01 -3.32768172e-01
-6.55457556e-01 8.62525940e-01 9.89279389e-01 4.47630227e-01
1.02494586e+00 -3.52540493e-01 6.05421960e-01 6.62290514e-01
1.41043171e-01 -1.05599022e+00 3.61719280e-01 4.18698430e-01
1.01146591e+00 -1.16008949e+00 4.82254177e-01 -6.47046864e-01
-6.30410433e-01 1.50124264e+00 8.37173104e-01 -2.51940668e-01
6.85336709e-01 1.21144064e-01 -2.91260362e-01 -1.14986666e-01
-2.64010727e-01 -2.95804858e-01 6.79493725e-01 5.40558219e-01
1.96002975e-01 8.57605338e-02 -6.34363174e-01 5.50992072e-01
-1.16930746e-01 -2.64982790e-01 4.99946684e-01 9.20582294e-01
-6.60605371e-01 -6.93898261e-01 -6.98698044e-01 4.43053693e-01
-1.97300509e-01 3.89618874e-02 -4.47875500e-01 1.08977246e+00
4.82470393e-01 5.70096552e-01 2.19786242e-01 -5.21251380e-01
5.06447732e-01 -1.82778403e-01 1.81244269e-01 -5.53640068e-01
-1.81465060e-01 1.00279562e-02 -4.32660729e-01 -2.21456423e-01
-2.88867831e-01 -5.16341627e-01 -1.18550193e+00 3.21948647e-01
-9.48281109e-01 6.19403087e-02 5.29285848e-01 4.97865617e-01
-1.24921091e-01 4.81120497e-01 8.67288351e-01 -5.79026699e-01
-5.89017451e-01 -8.59176040e-01 -5.47496974e-01 7.52021298e-02
1.87350318e-01 -8.62320125e-01 -4.49294209e-01 1.70501798e-01] | [8.688843727111816, -0.7501952648162842] |
d67dfe86-3232-404c-abab-2d8f4ce17a69 | end-to-end-learning-of-geometry-and-context | 1703.04309 | null | http://arxiv.org/abs/1703.04309v1 | http://arxiv.org/pdf/1703.04309v1.pdf | End-to-End Learning of Geometry and Context for Deep Stereo Regression | We propose a novel deep learning architecture for regressing disparity from a
rectified pair of stereo images. We leverage knowledge of the problem's
geometry to form a cost volume using deep feature representations. We learn to
incorporate contextual information using 3-D convolutions over this volume.
Disparity values are regressed from the cost volume using a proposed
differentiable soft argmin operation, which allows us to train our method
end-to-end to sub-pixel accuracy without any additional post-processing or
regularization. We evaluate our method on the Scene Flow and KITTI datasets and
on KITTI we set a new state-of-the-art benchmark, while being significantly
faster than competing approaches. | ['Abraham Bachrach', 'Hayk Martirosyan', 'Peter Henry', 'Saumitro Dasgupta', 'Alex Kendall', 'Adam Bry', 'Ryan Kennedy'] | 2017-03-13 | end-to-end-learning-of-geometry-and-context-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Kendall_End-To-End_Learning_of_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Kendall_End-To-End_Learning_of_ICCV_2017_paper.pdf | iccv-2017-10 | ['stereo-lidar-fusion'] | ['computer-vision'] | [ 3.49806398e-01 -5.48101589e-03 1.50405660e-01 -7.58760333e-01
-8.33380997e-01 -3.14289182e-01 5.36537886e-01 -2.25278467e-01
-7.40227938e-01 7.30051696e-01 1.57372773e-01 -3.35872740e-01
3.49545479e-01 -7.78387964e-01 -1.04410982e+00 -2.23789960e-01
-1.19027287e-01 1.45102277e-01 1.74631998e-01 -1.30473346e-01
6.29213512e-01 5.15345275e-01 -1.60922074e+00 2.89648205e-01
8.56873274e-01 1.22901213e+00 -1.10151649e-01 9.90056217e-01
1.79418221e-01 1.14788556e+00 -2.13126257e-01 -1.87602788e-01
8.58833969e-01 -1.03522219e-01 -9.32668924e-01 1.15721166e-01
1.43836761e+00 -8.95912707e-01 -8.67050946e-01 7.87798464e-01
3.86221975e-01 3.49004179e-01 6.48213685e-01 -6.78263724e-01
-3.70594084e-01 -1.75245374e-01 -8.08699250e-01 5.45745969e-01
2.42617264e-01 6.18248641e-01 9.45409119e-01 -1.04770970e+00
7.06693232e-01 1.42219543e+00 6.65997922e-01 2.30908841e-01
-1.44984174e+00 -3.82315367e-01 6.28990680e-02 1.08621819e-02
-1.06823862e+00 -5.78274488e-01 6.72435760e-01 -6.41032457e-01
1.51531386e+00 -2.29205698e-01 5.82846999e-01 5.54993212e-01
3.69187355e-01 4.65582132e-01 1.01940095e+00 -6.79938421e-02
-4.81895693e-02 -4.30737257e-01 -1.49114713e-01 1.06688499e+00
-1.44606322e-01 4.64717269e-01 -4.50832009e-01 3.59033316e-01
1.23528147e+00 -1.33400261e-01 -2.17484564e-01 -3.94269496e-01
-9.91893053e-01 9.10215259e-01 1.09419858e+00 -5.19913852e-01
-2.49662310e-01 6.34232581e-01 2.36975238e-01 1.69419944e-01
7.90473759e-01 9.40388516e-02 -4.29018289e-01 -1.93874106e-01
-7.16784298e-01 4.32206064e-01 4.82295156e-01 5.72324336e-01
1.38338506e+00 6.63771853e-02 8.07294436e-03 6.53626621e-01
1.58975855e-01 3.85538071e-01 9.55131501e-02 -1.51221848e+00
8.37541997e-01 4.96716261e-01 7.36757473e-04 -7.47543097e-01
-4.06446636e-01 -3.26340348e-01 -5.40099978e-01 8.92703891e-01
4.28628385e-01 -3.34580004e-01 -1.30594337e+00 1.44054830e+00
3.11073780e-01 6.39464855e-01 -1.10024743e-01 1.27943993e+00
7.21861005e-01 6.87810123e-01 -2.09398717e-01 3.03057909e-01
6.96807384e-01 -1.11612761e+00 -1.29185412e-02 -6.28242016e-01
7.20781624e-01 -7.58466661e-01 1.01151836e+00 2.97942132e-01
-1.30220234e+00 -6.08586848e-01 -1.14612365e+00 -8.74203980e-01
-1.26119569e-01 -1.37958944e-01 6.57099187e-01 1.37067840e-01
-1.32643282e+00 9.12959158e-01 -8.59501183e-01 9.38691422e-02
8.75192642e-01 5.24295866e-01 -3.84687275e-01 -2.11071119e-01
-5.46706915e-01 9.11844254e-01 -1.82249710e-01 2.27799818e-01
-8.82472515e-01 -1.18881190e+00 -1.47477984e+00 -3.39562930e-02
-4.52903695e-02 -1.10793054e+00 1.21702647e+00 -7.63788402e-01
-1.83077919e+00 1.15779805e+00 -3.61795187e-01 -5.12165546e-01
7.75224686e-01 -4.44272727e-01 1.88800588e-01 1.65536270e-01
2.72259325e-01 1.03073537e+00 8.58381093e-01 -8.71515334e-01
-8.88272643e-01 -2.96961725e-01 2.34972745e-01 3.00175667e-01
2.59363532e-01 -4.30893511e-01 -2.67608047e-01 -4.36073005e-01
1.38887137e-01 -8.16479445e-01 -5.45287907e-01 5.94440103e-01
-2.72947729e-01 2.19260529e-01 8.20784032e-01 -6.10302091e-01
5.06926715e-01 -2.01996112e+00 1.35005280e-01 -1.58216599e-02
4.97316986e-01 6.32746369e-02 -2.82045841e-01 -2.42892385e-01
3.83184142e-02 -3.34188074e-01 -6.73457325e-01 -7.23744512e-01
-3.67570698e-01 1.73353553e-01 -3.66862297e-01 8.87034714e-01
6.65329933e-01 8.77246261e-01 -8.83387625e-01 -2.95508832e-01
8.28651726e-01 7.65076876e-01 -1.08803582e+00 4.42463815e-01
-9.86064151e-02 7.33778894e-01 -5.99325150e-02 2.99222857e-01
8.51279914e-01 -1.10602036e-01 -5.66139877e-01 -5.26529290e-02
-3.18873525e-01 6.83257282e-01 -9.51238871e-01 2.16588020e+00
-8.20667446e-01 1.17526233e+00 -1.07034229e-01 -1.02684820e+00
9.01329935e-01 -4.53359187e-01 4.58352268e-01 -9.30747449e-01
1.49450168e-01 2.19239086e-01 -1.08292259e-01 -3.47620249e-01
4.71504003e-01 -7.66437948e-02 1.88293651e-01 1.78920731e-01
8.00906569e-02 -5.60961962e-01 -7.92258419e-03 -9.81184468e-03
1.15334916e+00 4.92699504e-01 3.57985846e-03 -2.95905977e-01
6.37350500e-01 -3.84297445e-02 5.27894497e-01 3.29509526e-01
-1.76963136e-01 9.27426457e-01 4.96587455e-01 -1.03259099e+00
-9.94980097e-01 -1.29861248e+00 -2.65324295e-01 7.22818911e-01
3.39268714e-01 7.59753138e-02 -4.15403664e-01 -7.13992596e-01
3.96567523e-01 2.35449970e-01 -8.60519826e-01 1.14969499e-01
-9.82481599e-01 -1.94793671e-01 1.65763721e-01 7.84046531e-01
7.04252481e-01 -6.77512050e-01 -1.02802241e+00 2.74917066e-01
1.93564910e-02 -1.31354618e+00 -6.38021290e-01 1.13181181e-01
-1.02473378e+00 -1.15634882e+00 -5.44979572e-01 -8.78991663e-01
6.54758871e-01 3.05744320e-01 1.37447810e+00 9.77294594e-02
-6.83413982e-01 5.23004942e-02 4.06445205e-01 -4.11091782e-02
2.23195814e-02 -6.58472329e-02 -3.42381924e-01 -8.75291526e-02
-3.99771472e-03 -7.69350350e-01 -1.19279146e+00 1.66625574e-01
-6.44899249e-01 1.92186192e-01 2.95582116e-01 7.51942337e-01
6.34015381e-01 -7.43899167e-01 3.84769924e-02 -6.35641336e-01
4.57382351e-02 -1.57185838e-01 -1.16195142e+00 -4.09739286e-01
-1.10167794e-01 3.56746852e-01 4.98605877e-01 4.36076224e-02
-1.21365666e+00 2.70170033e-01 -2.85141766e-01 -6.00530863e-01
-4.65711541e-02 5.46797775e-02 3.96647394e-01 -4.65845197e-01
6.94995940e-01 -2.02963009e-01 1.80902835e-02 -2.26539657e-01
5.41996598e-01 2.14999512e-01 1.06700158e+00 -4.47345138e-01
7.36660302e-01 1.03173339e+00 4.39214021e-01 -6.44798577e-01
-9.95616198e-01 -3.89489323e-01 -8.57271612e-01 -1.14221543e-01
9.04527426e-01 -1.11582398e+00 -8.15834820e-01 4.45175141e-01
-1.21026540e+00 -8.02105665e-01 -2.95769393e-01 6.14612520e-01
-8.84785533e-01 1.91557333e-01 -7.90938437e-01 -2.48315141e-01
-3.64184916e-01 -1.20274222e+00 1.25066316e+00 3.23513418e-01
7.21866190e-02 -1.20580983e+00 2.96525568e-01 2.73599863e-01
4.45489824e-01 4.58063096e-01 4.84450072e-01 2.69707948e-01
-1.02153778e+00 1.67320356e-01 -7.46522665e-01 5.23669362e-01
4.99705523e-02 7.46456608e-02 -1.37966573e+00 -3.27124238e-01
-1.26521528e-01 -5.33935189e-01 1.41821313e+00 7.20888436e-01
1.23213637e+00 -1.01158738e-01 -3.78920324e-03 1.63148570e+00
1.72292650e+00 -2.69157112e-01 7.63905644e-01 3.69280875e-01
8.76361370e-01 5.80498099e-01 4.29338813e-01 3.84966582e-01
7.08518565e-01 4.26667362e-01 7.49069214e-01 -3.45638096e-01
-3.49937260e-01 -1.93864912e-01 5.48777357e-02 2.14419127e-01
-1.23658292e-01 2.57111728e-01 -8.84473264e-01 4.81559753e-01
-1.77723205e+00 -7.42417693e-01 -5.74380159e-02 2.11685276e+00
5.65804958e-01 3.25122714e-01 -1.56539395e-01 -1.39481217e-01
1.83085009e-01 3.63443017e-01 -8.02189946e-01 -5.80987573e-01
-5.83236031e-02 5.59639752e-01 7.70470798e-01 1.09399033e+00
-1.31235898e+00 1.15579271e+00 6.66276932e+00 7.06658736e-02
-1.56621957e+00 -3.16564590e-01 1.01948726e+00 -3.38954329e-01
-1.89955056e-01 4.16921526e-02 -4.34480429e-01 8.21920261e-02
5.88843524e-01 -5.40009961e-02 4.92576718e-01 5.87292910e-01
2.40159988e-01 -4.28295702e-01 -1.31279945e+00 1.28402686e+00
7.29333311e-02 -1.76607585e+00 -2.19448879e-01 1.72867745e-01
1.04579008e+00 6.90466225e-01 2.02586427e-01 2.46552452e-01
5.20438731e-01 -1.25024378e+00 5.23246229e-01 3.45891416e-01
8.90190005e-01 -6.89070940e-01 5.56074500e-01 -1.74671680e-01
-1.11769652e+00 -1.19511962e-01 -4.73837525e-01 -7.16620743e-01
3.99733722e-01 5.03547549e-01 -5.44991374e-01 1.90353259e-01
7.44414151e-01 1.35103679e+00 -4.11152005e-01 1.12234783e+00
-2.12543815e-01 1.33719906e-01 -3.33016992e-01 5.73999822e-01
5.59122264e-01 -3.80932033e-01 3.42231303e-01 1.19297707e+00
1.60055041e-01 1.31481335e-01 3.13897990e-02 9.92830575e-01
-3.53422940e-01 -3.35413218e-01 -7.46892214e-01 8.09145749e-01
-4.59920913e-02 1.11292076e+00 -3.35137606e-01 -4.63145614e-01
-5.87042868e-01 1.09289682e+00 7.89407432e-01 4.82423216e-01
-6.04015648e-01 -3.95745754e-01 1.32763219e+00 2.64201313e-02
4.82701629e-01 -2.17958778e-01 -8.44039679e-01 -1.28230822e+00
2.40455404e-01 -3.01186562e-01 1.83769777e-01 -7.16986477e-01
-1.12356162e+00 5.03047824e-01 -3.08761925e-01 -1.23457491e+00
-4.74796087e-01 -6.94173515e-01 -8.84114385e-01 1.12647760e+00
-2.22451234e+00 -8.32850099e-01 -6.60998464e-01 6.47835553e-01
5.13322771e-01 1.62980393e-01 2.54979461e-01 2.83660054e-01
-4.02310610e-01 3.04603487e-01 -2.17404231e-01 1.88022152e-01
7.41836965e-01 -1.25116503e+00 9.47823346e-01 8.17255020e-01
-1.98943853e-01 8.43529627e-02 4.52719390e-01 -1.69825360e-01
-1.15886462e+00 -1.24297476e+00 6.20564699e-01 -5.60373724e-01
5.22693038e-01 -2.62191564e-01 -9.01971877e-01 8.99114430e-01
1.18946269e-01 9.42832410e-01 -5.79913817e-02 -2.39810169e-01
-6.21158898e-01 -2.46711016e-01 -1.31131983e+00 4.33681875e-01
1.53335249e+00 -6.06091380e-01 -4.80211437e-01 9.22843218e-02
6.86820865e-01 -9.20352578e-01 -5.76225281e-01 3.98312807e-01
4.41870600e-01 -1.30040002e+00 1.25108004e+00 -5.07503569e-01
1.02899265e+00 -1.90752998e-01 -8.45261589e-02 -1.48993003e+00
-7.98372999e-02 -7.36994922e-01 -1.59914196e-02 5.00983834e-01
2.62688905e-01 -7.41065323e-01 9.48489428e-01 7.41831183e-01
-4.46613908e-01 -8.08848560e-01 -1.19264066e+00 -4.01291877e-01
3.71618748e-01 -4.70707148e-01 3.79105419e-01 5.36716282e-01
-2.43008882e-01 2.15804890e-01 -2.69159496e-01 1.37987241e-01
9.96382535e-01 9.08119455e-02 1.01671660e+00 -9.32738245e-01
-1.20576955e-01 -5.43722034e-01 -8.33714008e-01 -1.59852672e+00
3.38329494e-01 -6.91797912e-01 2.32648492e-01 -1.50701070e+00
-1.93555772e-01 -2.78682977e-01 -6.27049580e-02 3.27597529e-01
-2.21328869e-01 6.07027888e-01 6.36064261e-02 -1.19905889e-01
-1.72070041e-01 6.69443071e-01 1.37231827e+00 -2.78744161e-01
-3.37235719e-01 -2.65689015e-01 -3.02156925e-01 7.62147307e-01
5.89949846e-01 -2.24901006e-01 -4.08136517e-01 -1.11831558e+00
-1.35680348e-01 9.42568630e-02 6.25000417e-01 -1.10481131e+00
1.55974925e-01 -2.01107129e-01 5.57899892e-01 -5.79532683e-01
5.85959196e-01 -4.07872796e-01 -7.04092264e-01 2.84317344e-01
-3.11261684e-01 3.11228056e-02 3.68194371e-01 2.78207004e-01
-1.67645961e-01 4.55177426e-01 1.04112756e+00 1.72146887e-01
-9.08187985e-01 6.08556211e-01 5.98841486e-03 4.65849996e-01
8.35987687e-01 -2.48557240e-01 -3.24045748e-01 -3.41959029e-01
-2.31888339e-01 2.79824823e-01 6.14899993e-01 3.64584118e-01
9.51128781e-01 -1.06428063e+00 -8.59637976e-01 6.04731619e-01
2.28285347e-03 6.43194497e-01 2.30794817e-01 5.66763282e-01
-1.26303709e+00 3.34945500e-01 -4.90672886e-01 -1.03649044e+00
-8.13928366e-01 1.57116562e-01 6.78092778e-01 9.96377692e-03
-1.05089402e+00 9.67226326e-01 4.73049492e-01 -4.51851666e-01
2.70216167e-01 -6.92434788e-01 1.12871386e-01 -5.30641317e-01
4.81296867e-01 3.19373876e-01 5.16194813e-02 -6.47960901e-01
-3.90211433e-01 9.94475186e-01 9.80140641e-02 -3.15323561e-01
1.41537559e+00 -1.80068299e-01 1.65708572e-01 2.39463691e-02
1.85675359e+00 -3.10847014e-01 -2.30852485e+00 -2.82572985e-01
-3.45030636e-01 -1.17836022e+00 4.54114228e-01 -4.24404681e-01
-1.52279198e+00 1.23947501e+00 5.80384314e-01 -6.84307694e-01
8.46632004e-01 -3.33686203e-01 9.67387855e-01 3.82601976e-01
1.19900659e-01 -8.72104466e-01 1.64137378e-01 7.20078528e-01
7.19698012e-01 -1.78384626e+00 9.68046784e-02 -3.18434238e-01
-2.85355002e-01 1.14291608e+00 8.22018266e-01 -8.51911426e-01
8.02156985e-01 4.56115723e-01 5.27780414e-01 -2.81255729e-02
-7.28991330e-01 -2.62575239e-01 3.83111954e-01 5.84925473e-01
3.65592062e-01 -2.53688306e-01 1.41728923e-01 -3.79841417e-01
-2.20489010e-01 2.55617619e-01 6.30777240e-01 1.05549800e+00
-3.29209745e-01 -6.66651607e-01 2.72427965e-02 3.74736995e-01
-1.06610492e-01 -1.45250499e-01 -2.04002019e-02 5.11468887e-01
-1.16799518e-01 7.80137300e-01 7.72202671e-01 -2.25844368e-01
4.95576501e-01 -5.99551439e-01 5.56346714e-01 -5.41174054e-01
-3.47802520e-01 -2.21823037e-01 -2.16042995e-03 -1.20432591e+00
-2.26923749e-01 -4.67886209e-01 -1.31792510e+00 -5.20141482e-01
4.35248822e-01 -5.27287841e-01 6.08717382e-01 9.20462489e-01
3.42420667e-01 3.51461828e-01 9.83865917e-01 -1.45178711e+00
-2.65477329e-01 -6.27565742e-01 5.27489781e-02 6.04329884e-01
9.84297574e-01 -5.46237469e-01 -5.09910405e-01 7.93847665e-02] | [8.649645805358887, -2.087611198425293] |
a0e7fff4-92a9-4191-be8a-34c327cea2cf | peer-to-peer-federated-continual-learning-for | 2304.07421 | null | https://arxiv.org/abs/2304.07421v1 | https://arxiv.org/pdf/2304.07421v1.pdf | Peer-to-Peer Federated Continual Learning for Naturalistic Driving Action Recognition | Naturalistic driving action recognition (NDAR) has proven to be an effective method for detecting driver distraction and reducing the risk of traffic accidents. However, the intrusive design of in-cabin cameras raises concerns about driver privacy. To address this issue, we propose a novel peer-to-peer (P2P) federated learning (FL) framework with continual learning, namely FedPC, which ensures privacy and enhances learning efficiency while reducing communication, computational, and storage overheads. Our framework focuses on addressing the clients' objectives within a serverless FL framework, with the goal of delivering personalized and accurate NDAR models. We demonstrate and evaluate the performance of FedPC on two real-world NDAR datasets, including the State Farm Distracted Driver Detection and Track 3 NDAR dataset in the 2023 AICity Challenge. The results of our experiments highlight the strong competitiveness of FedPC compared to the conventional client-to-server (C2S) FLs in terms of performance, knowledge dissemination rate, and compatibility with new clients. | ['Ziran Wang', 'Lu Su', 'Yunsheng Ma', 'Liangqi Yuan'] | 2023-04-14 | null | null | null | null | ['action-recognition-in-videos'] | ['computer-vision'] | [-4.14118946e-01 5.16723134e-02 -3.89463305e-01 -5.71510911e-01
-1.12899494e+00 -5.31367779e-01 5.59018791e-01 -2.05076426e-01
-4.10861582e-01 5.60184062e-01 2.57644176e-01 -3.67451817e-01
-3.27711135e-01 -3.90372574e-01 -7.92049706e-01 -6.79990768e-01
3.25132430e-01 4.75196578e-02 5.44321716e-01 5.96442372e-02
-9.11077335e-02 9.23250914e-01 -1.87712753e+00 1.36796683e-01
3.74642670e-01 1.10623670e+00 -1.17246002e-01 4.61921841e-01
5.65605760e-02 1.20018983e+00 -4.60436642e-01 -1.04436147e+00
5.41734576e-01 3.32238287e-01 -3.60953778e-01 -3.35506350e-01
5.40440559e-01 -8.78405750e-01 -9.70242262e-01 5.16602576e-01
6.12250447e-01 5.68812080e-02 -2.30833068e-02 -2.11053205e+00
-3.45120281e-01 1.07941508e-01 -5.16371787e-01 3.06248248e-01
-9.93106365e-02 3.24298412e-01 6.47363365e-01 -4.48600888e-01
3.32862854e-01 9.80790973e-01 5.47212183e-01 7.00918198e-01
-6.74965382e-01 -1.01679409e+00 1.88192055e-01 7.67712355e-01
-1.38405955e+00 -7.93018579e-01 4.58513379e-01 -2.98442960e-01
6.78018153e-01 5.00422299e-01 3.75816524e-01 1.02800763e+00
2.19234735e-01 1.12217116e+00 7.49632359e-01 5.14615029e-02
3.45018476e-01 4.25143123e-01 3.78778458e-01 2.77051508e-01
1.74947530e-01 3.24442714e-01 -1.17504370e+00 -4.67332065e-01
7.97738880e-02 3.21695626e-01 3.40492185e-03 -4.07865912e-01
-5.75238585e-01 5.94034970e-01 -8.14996585e-02 -3.84762377e-01
-6.52699888e-01 3.38863611e-01 5.22984982e-01 1.56305760e-01
5.86923778e-01 -3.86324883e-01 -4.73226964e-01 -4.90085840e-01
-4.76131856e-01 4.03118193e-01 8.24134290e-01 1.05741084e+00
6.84298694e-01 -2.21369997e-01 -5.13883531e-01 5.52349210e-01
1.34766966e-01 7.59234548e-01 6.00596927e-02 -1.52737546e+00
2.35426098e-01 2.80317485e-01 2.52347112e-01 -8.23732078e-01
1.02094062e-01 -1.80876419e-01 -3.76275554e-02 1.91648364e-01
-1.27698749e-01 -4.45550054e-01 -2.20335331e-02 1.42647326e+00
5.35584390e-01 2.80183464e-01 2.17629671e-01 9.58064735e-01
7.36807704e-01 5.36099553e-01 3.12016103e-02 -1.80086970e-01
9.73427236e-01 -9.38500285e-01 -9.56811845e-01 -7.14690611e-02
5.36074281e-01 -2.88142115e-01 6.03830159e-01 3.52762282e-01
-8.59516740e-01 -9.82913151e-02 -5.39110363e-01 9.97087136e-02
-9.87753496e-02 -6.60140142e-02 3.51205617e-01 8.43180954e-01
-9.05275941e-01 -2.72836685e-01 -7.31246710e-01 -2.11659491e-01
8.32416713e-01 2.67930269e-01 -5.44327974e-01 -3.16225022e-01
-7.86051452e-01 8.11046004e-01 -2.86318749e-01 -1.77870035e-01
-1.15366054e+00 -1.10342872e+00 -2.71055579e-01 1.99489951e-01
6.82668447e-01 -3.34360451e-01 1.57985890e+00 -4.89901006e-01
-1.42828774e+00 5.54578662e-01 -2.65402406e-01 -9.27820146e-01
7.45514274e-01 -4.52819824e-01 -6.00113332e-01 2.08784997e-01
-9.09282714e-02 6.36207402e-01 6.96741700e-01 -1.11037314e+00
-1.10949206e+00 -5.46443701e-01 -6.96749985e-02 1.91645652e-01
-6.55753493e-01 3.49018902e-01 -4.21041697e-01 2.47226283e-01
-7.61466801e-01 -8.19966376e-01 1.38244197e-01 3.46844494e-01
5.54073341e-02 -6.75322354e-01 1.54983771e+00 -6.72963321e-01
7.70280719e-01 -2.79818416e+00 -8.74503791e-01 1.35873288e-01
4.38279867e-01 5.03261566e-01 -5.27377501e-02 3.86506319e-01
4.51981068e-01 -4.54826534e-01 5.36228120e-01 -3.34052145e-01
1.89335883e-01 3.87179971e-01 -2.98289090e-01 5.14413774e-01
-1.78925037e-01 8.14465284e-01 -4.91782784e-01 -5.28685808e-01
1.86123669e-01 5.37052274e-01 -3.91047478e-01 2.75269896e-01
3.71002145e-02 2.62197912e-01 -4.84343946e-01 3.54644030e-01
5.66753924e-01 2.11007431e-01 -1.77126974e-01 1.90021500e-01
-3.30150008e-01 -6.15578555e-02 -7.80463338e-01 1.00507832e+00
-1.86706364e-01 9.70831990e-01 5.11092186e-01 -5.68691552e-01
1.00325203e+00 5.00016928e-01 8.17200840e-01 -9.65882480e-01
5.59136942e-02 -1.52167201e-01 -6.08684421e-01 -7.90582478e-01
3.53229761e-01 4.59486604e-01 1.03686810e-01 3.06045830e-01
-3.24236065e-01 5.38337409e-01 -3.80263746e-01 3.36197555e-01
1.51463830e+00 -3.71205628e-01 -3.35192919e-01 1.23064056e-01
2.80438155e-01 9.84168872e-02 6.97194278e-01 5.60423732e-01
-9.11738515e-01 3.54820937e-02 4.17565733e-01 -4.05020356e-01
-4.33802307e-01 -9.38956678e-01 2.01166213e-01 1.08286583e+00
8.04306865e-02 -2.95046151e-01 -7.12202489e-01 -7.46512234e-01
4.64298099e-01 1.12769699e+00 -1.03937062e-02 -2.79912978e-01
-1.75863475e-01 7.55755231e-02 5.92159450e-01 4.32569176e-01
7.49956727e-01 -5.69357514e-01 -7.47826934e-01 -1.80344716e-01
-3.82727236e-01 -1.41515851e+00 -5.74115872e-01 -4.42225307e-01
-1.73623577e-01 -1.08023465e+00 -2.38094702e-01 -2.74044424e-01
2.12744460e-01 9.73887682e-01 5.15969098e-01 -5.82467139e-01
-2.54833758e-01 1.24749207e+00 5.19794524e-02 -1.12029636e+00
-1.88898280e-01 -2.39877149e-01 2.01788321e-01 5.39695084e-01
1.13326287e+00 -2.03031421e-01 -7.82863379e-01 5.89794695e-01
-5.35610020e-01 -4.20876235e-01 5.01370966e-01 1.76372334e-01
6.10968709e-01 3.00059885e-01 8.42339993e-01 -5.74810386e-01
7.44861484e-01 -5.52289844e-01 -9.97203648e-01 1.86623782e-01
-9.15024042e-01 -6.75360084e-01 1.18387774e-01 -8.76672417e-02
-1.35221803e+00 2.85529703e-01 7.90253580e-02 -8.19828987e-01
-3.93991619e-01 -2.72778515e-02 -1.19611144e-01 -3.16180915e-01
3.65435004e-01 1.59405038e-01 6.31685972e-01 -4.01484638e-01
4.42470402e-01 1.27254260e+00 8.67799163e-01 1.83110852e-02
5.01768708e-01 7.02018559e-01 -2.68325537e-01 -7.10719824e-01
-4.98333901e-01 -8.86658609e-01 -1.62251629e-02 -5.66149354e-01
6.72042429e-01 -1.39129460e+00 -1.45498753e+00 3.94722670e-01
-9.84743536e-01 8.64810497e-03 -5.11539996e-01 7.66714215e-01
-6.08409584e-01 6.81275083e-03 -2.91532397e-01 -1.17148125e+00
-4.37744707e-01 -6.94908679e-01 6.46103621e-01 1.79760322e-01
1.75536498e-01 -1.19973682e-01 9.51399952e-02 1.10320175e+00
6.27063453e-01 -1.10766813e-01 5.11598766e-01 -6.92514002e-01
-9.59507704e-01 -7.74585426e-01 -3.20567459e-01 6.55222893e-01
-3.00334632e-01 -2.49096781e-01 -1.42893386e+00 -3.70841682e-01
-1.29743099e-01 -2.00556323e-01 2.16369465e-01 1.19636349e-01
1.28300190e+00 -5.89402497e-01 -4.34141666e-01 1.38053551e-01
1.17482114e+00 6.59512281e-02 5.79872787e-01 -7.93415867e-03
4.10380930e-01 8.64916623e-01 7.24859357e-01 7.14375436e-01
7.60835350e-01 4.27133024e-01 6.69874787e-01 1.41099885e-01
-1.56424537e-01 -4.50268775e-01 6.61897898e-01 3.33463103e-01
4.14196461e-01 -1.63927600e-01 -7.65327990e-01 9.24919844e-01
-2.10225177e+00 -9.73542273e-01 -6.89698532e-02 2.22443748e+00
2.60136604e-01 -3.06571484e-01 4.46435094e-01 -1.85740933e-01
5.15566289e-01 -1.60614178e-01 -9.45465505e-01 -3.16509962e-01
1.61653087e-01 -5.05643070e-01 1.08762610e+00 -1.19512737e-01
-5.86711764e-01 4.69175935e-01 6.21194649e+00 8.19520772e-01
-9.47383404e-01 7.31088936e-01 4.73150253e-01 -8.03794742e-01
-1.04255159e-03 -4.44585472e-01 -1.03848326e+00 3.74524295e-01
1.63066387e+00 -5.31486392e-01 5.07461905e-01 1.26869822e+00
4.38079953e-01 1.52657941e-01 -8.47367048e-01 1.21139884e+00
3.43854949e-02 -1.49990654e+00 -5.42721808e-01 2.25500032e-01
3.77677232e-01 7.69595325e-01 -1.14857610e-02 3.32582951e-01
3.25517744e-01 -5.39466262e-01 6.05613351e-01 5.45151055e-01
5.76021731e-01 -1.17116141e+00 6.41295791e-01 6.44704938e-01
-7.09392428e-01 -6.40814185e-01 -2.61722505e-01 3.01637322e-01
1.05031237e-01 2.85894990e-01 -5.11393547e-01 2.45494142e-01
1.15830290e+00 3.47901255e-01 -2.62288690e-01 1.10625529e+00
3.85532141e-01 8.23493421e-01 -3.49512458e-01 3.47789153e-02
5.05152531e-02 2.10345462e-01 5.67411304e-01 8.63428175e-01
7.52778649e-02 2.09284052e-01 -1.05507985e-01 5.96818447e-01
-2.29773968e-01 2.33387831e-03 -9.85410333e-01 2.36577868e-01
7.15516567e-01 1.30368435e+00 2.70722181e-01 1.76300734e-01
-8.98377061e-01 4.74154294e-01 5.21693192e-02 3.49841505e-01
-8.36612523e-01 -1.47127435e-01 1.25504160e+00 4.82382238e-01
4.44359362e-01 -1.09516911e-01 5.75263947e-02 -3.73632997e-01
2.45417669e-01 -6.70999944e-01 4.85237151e-01 -6.95708215e-01
-1.38346672e+00 1.69572368e-01 2.32843846e-01 -9.25214589e-01
6.60278425e-02 -2.22485542e-01 -3.76796275e-01 5.17768919e-01
-2.02767563e+00 -1.03997540e+00 -4.06284451e-01 1.14504409e+00
1.46932364e-01 -3.92468482e-01 4.86025184e-01 6.38548791e-01
-6.78784966e-01 9.14561391e-01 3.37232172e-01 -2.65500098e-01
8.12734902e-01 -3.25532407e-01 -1.90538093e-02 6.42726243e-01
-2.81416774e-01 -1.65841728e-01 2.17824817e-01 -3.55884492e-01
-2.12856817e+00 -1.58456910e+00 9.01826382e-01 -6.44083738e-01
3.45920473e-01 -3.04489076e-01 -6.71952009e-01 6.90724432e-01
2.52574801e-01 3.53704482e-01 7.85149932e-01 -1.37534320e-01
-3.39783221e-01 -9.91521299e-01 -1.43708682e+00 2.92571068e-01
8.68544757e-01 -7.23232985e-01 -8.58190842e-03 6.68901145e-01
8.46672177e-01 -4.07069325e-02 -6.86543703e-01 -1.37089357e-01
4.46887493e-01 -8.27268064e-01 7.47584045e-01 -4.94365960e-01
-2.75565952e-01 -7.49058947e-02 -1.52038768e-01 -7.41519213e-01
-2.93200552e-01 -1.19873726e+00 -2.71631598e-01 1.34780836e+00
3.81855927e-02 -1.08285069e+00 9.54953015e-01 1.36691415e+00
-8.61753151e-03 -3.67141217e-01 -1.41690946e+00 -1.01172101e+00
-3.29754531e-01 -7.05174029e-01 9.10485089e-01 3.48712295e-01
-1.66967958e-01 -7.54218698e-02 -4.58763272e-01 5.58057010e-01
7.55446017e-01 -4.18188870e-01 1.21091723e+00 -1.11903644e+00
-4.41792347e-02 1.72260627e-01 -5.69919050e-01 -7.75208950e-01
3.21413130e-01 -5.34321904e-01 1.28774911e-01 -1.07482004e+00
1.96955562e-01 -3.16473335e-01 -4.72629637e-01 6.82679474e-01
4.54677999e-01 -8.96438658e-02 2.72938430e-01 4.14734960e-01
-8.18512857e-01 5.27462482e-01 6.04675829e-01 -3.12314872e-02
9.47941095e-02 2.33721823e-01 -6.77715719e-01 2.82301396e-01
8.13443422e-01 -5.32536566e-01 -6.17669284e-01 -6.00450873e-01
-2.84785002e-01 1.93869889e-01 6.64493918e-01 -7.91749060e-01
7.25257277e-01 -3.53559762e-01 -3.32298845e-01 -9.04853702e-01
4.90653545e-01 -1.30901313e+00 2.66372740e-01 3.06393027e-01
-2.82450587e-01 -6.39222413e-02 2.92969555e-01 1.09415770e+00
-5.60531802e-02 3.48220766e-01 3.86533976e-01 4.34939742e-01
-7.97503710e-01 5.58659315e-01 -5.67501068e-01 -3.50772381e-01
1.68751788e+00 -5.67598715e-02 -9.18256700e-01 -6.11993194e-01
-2.00240761e-02 7.48640835e-01 6.04308210e-02 6.92387164e-01
7.36596584e-01 -1.02942383e+00 -8.13611209e-01 4.00706947e-01
5.01183450e-01 -3.60948533e-01 3.86154234e-01 8.89717996e-01
-3.00371032e-02 7.41252065e-01 -4.66015153e-02 -5.58215797e-01
-1.55470729e+00 6.42673790e-01 2.82128692e-01 4.49492067e-01
-9.42332029e-01 5.92499971e-01 -5.81620969e-02 -3.29383105e-01
8.19098651e-01 4.82902020e-01 7.62797147e-02 -2.96038359e-01
7.60993421e-01 9.86029744e-01 2.31157094e-01 -4.61384267e-01
-3.87526512e-01 -2.39821598e-01 -1.45001933e-01 1.16183385e-01
1.13159668e+00 -5.28589308e-01 9.07127336e-02 7.41940364e-02
1.10929060e+00 -1.41893461e-01 -1.64285004e+00 -5.14881253e-01
-2.32961074e-01 -7.72613406e-01 5.65332413e-01 -7.80220509e-01
-1.45286667e+00 4.67151016e-01 1.10873449e+00 -1.11829817e-01
8.74338865e-01 6.71407580e-02 1.42293775e+00 5.23089468e-01
4.90960300e-01 -1.16722977e+00 -3.22496504e-01 -1.88316658e-01
5.21053493e-01 -1.20202553e+00 -4.97350365e-01 -2.41878569e-01
-9.58665311e-01 6.35880888e-01 6.73860431e-01 3.83657455e-01
8.38157475e-01 2.73517340e-01 4.12174255e-01 -2.94745028e-01
-1.23079515e+00 1.87086284e-01 -1.67608932e-01 8.01703453e-01
-5.11293709e-01 2.17040032e-01 -3.32542397e-02 5.57600677e-01
1.51165366e-01 5.26628315e-01 2.64731050e-01 1.08268023e+00
-2.04684362e-01 -7.91019440e-01 -1.97969645e-01 5.29395103e-01
-3.17350715e-01 5.46286762e-01 -3.50518852e-01 2.68109322e-01
-5.14182970e-02 1.45157397e+00 1.41383126e-01 -3.37091744e-01
4.77300853e-01 1.82175085e-01 -1.46649867e-01 1.79139664e-03
-5.74048758e-01 -3.81698310e-01 2.31415078e-01 -1.02245522e+00
-1.97352037e-01 -7.52361417e-01 -8.63775432e-01 -8.48665237e-01
-1.67943925e-01 1.68238491e-01 1.11961937e+00 6.46061480e-01
1.17985594e+00 1.92959398e-01 1.02579975e+00 -1.66569203e-02
-9.97628987e-01 -1.97205096e-01 -6.14703596e-01 1.25553191e-01
4.00806218e-01 -3.00628960e-01 -1.96576461e-01 -1.73154876e-01] | [5.86002779006958, 6.283297538757324] |
17845561-4be9-4783-bd06-277649011e9e | how-to-learn-and-generalize-from-three | 2306.06335 | null | https://arxiv.org/abs/2306.06335v1 | https://arxiv.org/pdf/2306.06335v1.pdf | How to Learn and Generalize From Three Minutes of Data: Physics-Constrained and Uncertainty-Aware Neural Stochastic Differential Equations | We present a framework and algorithms to learn controlled dynamics models using neural stochastic differential equations (SDEs) -- SDEs whose drift and diffusion terms are both parametrized by neural networks. We construct the drift term to leverage a priori physics knowledge as inductive bias, and we design the diffusion term to represent a distance-aware estimate of the uncertainty in the learned model's predictions -- it matches the system's underlying stochasticity when evaluated on states near those from the training dataset, and it predicts highly stochastic dynamics when evaluated on states beyond the training regime. The proposed neural SDEs can be evaluated quickly enough for use in model predictive control algorithms, or they can be used as simulators for model-based reinforcement learning. Furthermore, they make accurate predictions over long time horizons, even when trained on small datasets that cover limited regions of the state space. We demonstrate these capabilities through experiments on simulated robotic systems, as well as by using them to model and control a hexacopter's flight dynamics: A neural SDE trained using only three minutes of manually collected flight data results in a model-based control policy that accurately tracks aggressive trajectories that push the hexacopter's velocity and Euler angles to nearly double the maximum values observed in the training dataset. | ['Ufuk Topcu', 'Cyrus Neary', 'Franck Djeumou'] | 2023-06-10 | null | null | null | null | ['model-based-reinforcement-learning'] | ['reasoning'] | [-2.26653188e-01 1.95561111e-01 -4.13068593e-01 -4.89528030e-02
-1.33332953e-01 -7.22527623e-01 4.32515711e-01 1.57985147e-02
-4.65318888e-01 1.07361293e+00 -3.61946613e-01 -4.08633411e-01
-5.02512574e-01 -6.74876273e-01 -9.49087918e-01 -9.32282329e-01
-8.56040657e-01 7.98561215e-01 2.20353186e-01 -6.46138251e-01
-1.64331898e-01 7.73033679e-01 -1.33178115e+00 -4.79266465e-01
8.99034381e-01 9.66609895e-01 -2.13493127e-02 8.89257967e-01
2.99797207e-01 1.05984676e+00 -5.62387884e-01 3.34994346e-01
5.83421946e-01 -2.29551584e-01 -3.03594381e-01 -3.92686129e-01
-1.50604308e-01 -3.87212783e-01 -7.48281598e-01 8.84255230e-01
2.11957335e-01 6.54320955e-01 6.04976594e-01 -1.10851538e+00
-1.47564203e-01 5.07745028e-01 1.27109140e-01 1.74244061e-01
-3.02680403e-01 8.35285068e-01 3.38860542e-01 -1.90383643e-01
5.59116483e-01 1.01838112e+00 9.53720033e-01 8.55258048e-01
-1.30899310e+00 -6.24371409e-01 4.02420878e-01 -3.91142488e-01
-1.35284090e+00 -9.56290737e-02 3.89033765e-01 -7.93015778e-01
9.91481900e-01 -2.00005025e-01 1.02061677e+00 1.03252625e+00
7.80813932e-01 3.27111512e-01 7.41163075e-01 3.16640556e-01
7.91440845e-01 -8.10728446e-02 -2.68831074e-01 9.87455785e-01
1.61264852e-01 1.19959593e+00 -2.60378301e-01 -1.36690289e-01
9.24977362e-01 -1.14618190e-01 -4.77319688e-01 -5.95207274e-01
-1.07886100e+00 7.38673210e-01 5.55146456e-01 -2.95704663e-01
-6.47513926e-01 5.48931062e-01 2.01442733e-01 4.52990115e-01
2.94490516e-01 9.17015016e-01 -7.23414302e-01 -3.59224766e-01
-6.88947976e-01 8.29978526e-01 1.02012384e+00 9.32344913e-01
4.69745338e-01 6.98933780e-01 -1.62500495e-04 2.22062439e-01
-1.54426545e-01 8.23073506e-01 4.66563106e-01 -1.14305508e+00
2.24841267e-01 3.88629079e-01 7.72640586e-01 -6.67358637e-01
-5.25513828e-01 -6.44117177e-01 -6.44209504e-01 6.21669948e-01
2.22758085e-01 -9.92064655e-01 -9.55693364e-01 1.92552149e+00
4.18489039e-01 3.03554833e-01 2.60545701e-01 1.14154887e+00
-2.15672672e-01 9.51352417e-01 -2.51130760e-01 -1.75635487e-01
4.33123112e-01 -5.38678467e-01 -4.60784674e-01 -1.63905144e-01
8.55558276e-01 1.30012840e-01 7.92353272e-01 3.86036605e-01
-1.14170110e+00 -5.62555015e-01 -1.29657900e+00 8.68090034e-01
-3.72395784e-01 -7.16008097e-02 2.64543176e-01 1.87484697e-01
-8.99221361e-01 1.36998725e+00 -1.34495175e+00 -1.19109727e-01
-1.12533152e-01 5.47768116e-01 2.18338490e-01 4.58420396e-01
-1.17516756e+00 1.20899272e+00 6.88412309e-01 1.47095978e-01
-1.77022541e+00 -1.08607709e+00 -5.54243624e-01 -1.93198800e-01
4.11905020e-01 -6.46027446e-01 1.55932820e+00 -6.65234625e-01
-1.91695678e+00 4.53984663e-02 2.16886595e-01 -1.08108532e+00
4.55794334e-01 -2.12274253e-01 -4.62517649e-01 -9.96442288e-02
-3.76636267e-01 5.88483632e-01 8.35321605e-01 -1.11588371e+00
-5.40496051e-01 9.01676621e-03 1.86779514e-01 1.82053775e-01
-2.19158959e-02 -7.29647696e-01 3.11532140e-01 -4.49482530e-01
-3.46198529e-02 -1.47027397e+00 -6.54846549e-01 1.15579583e-01
-1.41915813e-01 2.37898529e-01 8.82938504e-01 -3.35623294e-01
1.04949796e+00 -1.63152337e+00 4.96422201e-01 2.43228287e-01
-8.66120458e-02 4.72275734e-01 2.28030965e-01 6.10649586e-01
1.30451992e-01 -5.94628938e-02 -3.73095095e-01 8.94296169e-02
-1.18461594e-01 5.53273976e-01 -8.40757668e-01 3.71312559e-01
4.37448949e-01 6.84110045e-01 -1.03413677e+00 1.49961770e-01
1.67104319e-01 2.76512235e-01 -4.69082683e-01 4.49223131e-01
-8.41899514e-01 8.13041806e-01 -6.13942742e-01 7.04634413e-02
2.51654178e-01 1.11255705e-01 -4.12266701e-02 2.82503128e-01
-4.48814541e-01 1.24460328e-02 -9.09071505e-01 1.24674809e+00
-7.25366116e-01 5.02379179e-01 2.34156802e-01 -8.62203300e-01
1.11113739e+00 -5.61799072e-02 5.83625197e-01 -2.17401594e-01
2.93232799e-01 1.24070719e-01 1.57130495e-01 -3.87900144e-01
4.15219396e-01 -4.39374447e-01 4.57138866e-02 2.53077626e-01
1.96851656e-01 -8.70055199e-01 -4.32094373e-02 -5.82409799e-02
1.12490642e+00 2.52933294e-01 -1.83618918e-01 -5.15791655e-01
3.63159537e-01 5.37676632e-01 7.05696642e-01 6.92044377e-01
-9.56603289e-02 -1.08409375e-01 4.57955331e-01 -7.04791248e-01
-1.09981263e+00 -1.10386384e+00 5.23190387e-02 4.80253518e-01
3.09982479e-01 -1.97616726e-01 -7.15382755e-01 -4.04991925e-01
4.09289718e-01 1.01135707e+00 -7.33269751e-01 -8.01133454e-01
-5.36371469e-01 -4.57077265e-01 2.84251273e-01 5.50613225e-01
2.34728366e-01 -5.75338066e-01 -9.14892435e-01 3.65189433e-01
5.12489974e-01 -8.59892607e-01 -2.26289213e-01 5.14139235e-01
-8.99801314e-01 -1.12439132e+00 -3.88306022e-01 -2.99564004e-01
5.18161178e-01 -4.93763775e-01 7.71929622e-01 -1.04919523e-01
-1.68673173e-01 5.28478563e-01 1.91187590e-01 -6.63224339e-01
-6.32722497e-01 -9.49964300e-02 8.60026658e-01 -3.38600904e-01
-2.16346920e-01 -4.97911006e-01 -4.67620730e-01 4.37647313e-01
-6.29061759e-01 -1.48327246e-01 2.87150554e-02 9.38498080e-01
6.28468215e-01 2.99148172e-01 2.94996023e-01 -3.55645984e-01
7.64078021e-01 -5.19435346e-01 -1.41570854e+00 -3.80731635e-02
-7.85527587e-01 5.00292957e-01 1.09501135e+00 -8.90686035e-01
-8.70583117e-01 6.58527613e-02 1.58279076e-01 -1.01345801e+00
1.21762484e-01 4.33492154e-01 6.78790092e-01 -1.86398029e-01
7.41642475e-01 1.78441167e-01 3.09496254e-01 -1.39740616e-01
9.18807089e-02 1.85470074e-01 5.59017003e-01 -9.78207648e-01
9.25048470e-01 2.93664187e-01 4.65924472e-01 -6.60933197e-01
-8.01587641e-01 1.08454511e-01 -5.30476868e-01 -4.73669469e-01
6.09726727e-01 -9.03930068e-01 -1.01345432e+00 3.11889380e-01
-6.91493452e-01 -1.05820525e+00 -8.28934789e-01 7.73466706e-01
-8.67712200e-01 -5.50506711e-01 -4.76869434e-01 -1.21340930e+00
-2.00081691e-02 -8.66355658e-01 5.77266157e-01 4.40919518e-01
-9.43706110e-02 -1.35002029e+00 4.57218856e-01 -7.65779972e-01
4.51883823e-01 7.03187108e-01 8.76686633e-01 -3.96799237e-01
-4.30242896e-01 -3.16190064e-01 5.71710944e-01 5.84541559e-01
-2.86227614e-01 2.30184361e-01 -5.04999042e-01 -6.15834892e-01
6.34935051e-02 -4.80439186e-01 5.17682970e-01 5.65427184e-01
9.35221732e-01 -4.65125561e-01 -5.48811436e-01 6.77699804e-01
1.32443058e+00 3.90751034e-01 -1.91663593e-01 7.21201450e-02
3.19104195e-01 4.63732183e-01 7.07091689e-01 6.21636868e-01
7.92920440e-02 3.60067457e-01 6.47495866e-01 3.76792222e-01
4.29658920e-01 -7.86849797e-01 6.07161224e-01 8.09557855e-01
1.08163454e-01 -6.14399388e-02 -1.08787966e+00 4.17904973e-01
-1.79065609e+00 -8.14020991e-01 1.68004796e-01 2.60024023e+00
5.83548665e-01 4.01092738e-01 9.99377519e-02 -4.05168414e-01
3.82955432e-01 -5.34230396e-02 -1.29979479e+00 -5.71389973e-01
1.30448148e-01 1.27336802e-02 1.01622152e+00 5.74683070e-01
-8.40325356e-01 8.89970422e-01 7.19523811e+00 3.51696730e-01
-1.35991561e+00 -3.45420361e-01 3.92204106e-01 -5.26988328e-01
6.65562600e-02 5.24579957e-02 -9.14117157e-01 4.71210569e-01
1.56222606e+00 -6.24068737e-01 8.24795961e-01 1.09656298e+00
5.60858607e-01 -1.20592654e-01 -9.89128232e-01 4.99761492e-01
-4.96213973e-01 -1.42690384e+00 -1.90158501e-01 9.61275995e-02
1.13926148e+00 3.43377501e-01 2.79139251e-01 8.05088580e-01
1.06947982e+00 -9.42838073e-01 8.45602989e-01 8.11533153e-01
5.32064974e-01 -9.72500443e-01 6.22421682e-01 9.34427679e-01
-9.01918173e-01 -6.00573838e-01 -5.10613918e-01 -2.49047160e-01
3.29381347e-01 3.92261386e-01 -9.21216071e-01 4.86510694e-02
4.63295758e-01 6.74733281e-01 3.50939631e-02 9.00681317e-01
-3.71479541e-02 7.57451355e-01 -7.21986473e-01 -5.11216700e-01
6.24434054e-01 -3.87193978e-01 8.68127108e-01 5.62174976e-01
5.34231484e-01 1.58401877e-01 5.66886246e-01 1.05693841e+00
2.58826882e-01 -5.46530426e-01 -9.26130712e-01 -3.25401396e-01
2.69532919e-01 8.03207934e-01 -2.54288971e-01 -2.63715565e-01
4.12520289e-01 2.92010695e-01 1.92803115e-01 5.96948266e-01
-1.14155018e+00 -4.18891311e-01 1.25150108e+00 -5.70177147e-03
1.96136683e-01 -8.39882672e-01 1.72153398e-01 -1.01664519e+00
-3.18227559e-01 -4.79283333e-01 -1.19107865e-01 -6.35442138e-01
-9.18610156e-01 6.99829519e-01 2.52031982e-01 -1.35196567e+00
-1.03934157e+00 -8.51330101e-01 -6.49900377e-01 8.90896082e-01
-1.11528075e+00 -2.61493564e-01 1.48052061e-02 2.91662306e-01
2.40098879e-01 -2.98833936e-01 6.21116161e-01 -4.96869922e-01
-6.60038114e-01 -7.88030773e-02 5.36223292e-01 -2.48462006e-01
8.57772455e-02 -1.17849636e+00 4.90629822e-01 5.39220095e-01
-3.73153090e-01 4.70617324e-01 1.31010234e+00 -8.34993541e-01
-1.53015375e+00 -1.31107891e+00 -1.63981885e-01 -4.99449134e-01
1.02749479e+00 -1.28123298e-01 -1.03437006e+00 5.88650346e-01
-3.21217865e-01 2.12616414e-01 -8.59183744e-02 -2.32935533e-01
3.19547564e-01 -1.48397192e-01 -1.17403090e+00 6.13082945e-01
9.82959449e-01 -3.16598207e-01 -2.61226505e-01 4.07137573e-01
8.12201619e-01 -8.51653874e-01 -9.58757937e-01 5.75808287e-01
4.85908866e-01 -5.60238242e-01 6.19547427e-01 -1.13280427e+00
1.09263845e-01 -3.16226602e-01 8.97022113e-02 -2.09115767e+00
4.34230939e-02 -1.02801025e+00 -6.48390412e-01 5.84986389e-01
2.95095205e-01 -6.99587345e-01 7.59698510e-01 5.04127681e-01
-1.67626813e-01 -1.13761079e+00 -9.73429263e-01 -1.34064734e+00
5.22362649e-01 -2.76705474e-01 6.89615071e-01 6.39101744e-01
-1.03111938e-01 -2.77609408e-01 -2.54262477e-01 6.67956233e-01
4.82064337e-01 -1.18848354e-01 6.83535933e-01 -1.01794267e+00
-2.94117570e-01 -2.46970475e-01 -1.92192629e-01 -1.03984404e+00
5.75985968e-01 -4.99843180e-01 4.59057003e-01 -9.67050195e-01
-7.13704526e-01 -4.92531568e-01 -6.28154650e-02 1.24607841e-02
3.86448830e-01 -6.37852848e-01 -1.19105466e-02 1.31563395e-01
-7.63799325e-02 1.02547669e+00 1.06571686e+00 1.51439384e-02
-5.28396308e-01 3.75956535e-01 1.06995240e-01 7.22699940e-01
8.88045251e-01 -5.27136087e-01 -8.15759242e-01 -2.80785978e-01
-1.02480732e-01 4.03720498e-01 3.99444401e-01 -1.67322195e+00
3.10480207e-01 -6.48962855e-01 3.42340380e-01 -3.53102803e-01
4.43601131e-01 -6.55857921e-01 2.25168824e-01 9.50270355e-01
-7.05260158e-01 3.39871526e-01 5.50502837e-01 9.35255408e-01
-7.78857991e-02 2.86920145e-02 9.37061727e-01 -2.89062355e-02
-5.16937315e-01 7.05558360e-01 -8.09363246e-01 2.46840760e-01
1.30149591e+00 2.46935889e-01 -2.96347421e-02 -5.19308865e-01
-8.00266147e-01 6.75079644e-01 5.50205469e-01 4.36023057e-01
4.13905859e-01 -1.06297278e+00 -1.41854316e-01 3.30839366e-01
-1.40842184e-01 -9.01434664e-03 8.41490626e-02 4.24186260e-01
-6.06983244e-01 5.37171602e-01 -7.07173273e-02 -7.29564130e-01
-3.15933049e-01 7.19607711e-01 1.05000556e+00 -1.41813263e-01
-6.29492939e-01 6.27159417e-01 -1.39788330e-01 -7.20983386e-01
1.41861141e-01 -8.93218279e-01 4.65941757e-01 -6.00164592e-01
3.21396947e-01 1.59952983e-01 -1.52668998e-01 -3.86317521e-01
-1.62716992e-02 3.64553928e-01 4.31044579e-01 -2.38657624e-01
1.09248245e+00 9.96006206e-02 5.23404479e-01 8.66161525e-01
7.99982011e-01 -5.83232939e-01 -2.11281395e+00 -5.79096656e-03
-3.06241792e-02 3.20141832e-03 4.79712397e-01 -9.30379450e-01
-7.49898970e-01 8.62628460e-01 7.10841775e-01 1.09246291e-01
6.35988712e-01 -5.51089227e-01 6.62266433e-01 9.95564699e-01
7.47457206e-01 -1.31970656e+00 -2.25728944e-01 8.87402952e-01
8.49615693e-01 -8.70579839e-01 -2.93170184e-01 3.27778131e-01
-8.75636756e-01 1.24958467e+00 7.84444988e-01 -7.10976005e-01
9.07492936e-01 6.53042197e-01 -4.94854115e-02 1.29981443e-01
-1.39351773e+00 4.50349413e-02 1.22668378e-01 7.40728736e-01
-2.86654294e-01 2.31069364e-02 4.81777430e-01 3.43028784e-01
-1.47879362e-01 2.23302349e-01 5.21635354e-01 1.02268279e+00
-6.47769272e-01 -5.30981958e-01 -1.09642357e-01 3.53562593e-01
1.48345232e-01 3.56353104e-01 -7.65293390e-02 1.03972435e+00
-1.76582888e-01 4.83620703e-01 2.42235482e-01 -6.98608696e-01
5.04797459e-01 8.58743116e-02 2.60385513e-01 -3.74936163e-01
-4.16696459e-01 -6.86829090e-01 -4.80720401e-03 -7.87417829e-01
2.94578344e-01 -5.39303243e-01 -1.43845642e+00 -4.70386952e-01
-7.73839429e-02 5.06226540e-01 8.21027219e-01 8.88411760e-01
4.05911565e-01 5.79409957e-01 7.71434724e-01 -1.04445910e+00
-1.36552453e+00 -7.37542033e-01 -5.80221117e-01 -1.81293190e-01
7.24687994e-01 -1.00183237e+00 -5.92303097e-01 -2.18683705e-01] | [4.814412593841553, 2.16741943359375] |
57e73e12-3505-4fe6-822d-463a29bb5530 | supergf-unifying-local-and-global-features | 2212.13105 | null | https://arxiv.org/abs/2212.13105v1 | https://arxiv.org/pdf/2212.13105v1.pdf | SuperGF: Unifying Local and Global Features for Visual Localization | Advanced visual localization techniques encompass image retrieval challenges and 6 Degree-of-Freedom (DoF) camera pose estimation, such as hierarchical localization. Thus, they must extract global and local features from input images. Previous methods have achieved this through resource-intensive or accuracy-reducing means, such as combinatorial pipelines or multi-task distillation. In this study, we present a novel method called SuperGF, which effectively unifies local and global features for visual localization, leading to a higher trade-off between localization accuracy and computational efficiency. Specifically, SuperGF is a transformer-based aggregation model that operates directly on image-matching-specific local features and generates global features for retrieval. We conduct experimental evaluations of our method in terms of both accuracy and efficiency, demonstrating its advantages over other methods. We also provide implementations of SuperGF using various types of local features, including dense and sparse learning-based or hand-crafted descriptors. | ['Takayuki Okatani', 'Boshu Lei', 'Ran Yan', 'Wenzheng Song'] | 2022-12-23 | null | null | null | null | ['visual-localization', 'sparse-learning'] | ['computer-vision', 'methodology'] | [-2.71493375e-01 -7.78635740e-01 -3.78356546e-01 -4.68845814e-01
-1.33837509e+00 -7.80622303e-01 6.57784641e-01 1.30622774e-01
-3.22123080e-01 3.10741276e-01 2.96500444e-01 9.57979187e-02
-3.22436243e-01 -4.93091464e-01 -5.67008078e-01 -5.07243574e-01
-4.73502316e-02 2.74014175e-01 1.90563366e-01 2.01675370e-01
4.04914290e-01 7.59224117e-01 -1.48304296e+00 2.00755838e-02
5.64171135e-01 1.42471349e+00 3.95959467e-01 5.40522993e-01
1.86815768e-01 6.31437480e-01 -4.74323481e-01 -1.41561553e-01
4.21573848e-01 -4.57446389e-02 -5.62353194e-01 -4.55671996e-02
7.07940042e-01 -5.46730638e-01 -3.10037643e-01 8.95807505e-01
6.74053788e-01 -4.74618413e-02 4.81562972e-01 -1.33759964e+00
-5.82538307e-01 -1.47761151e-01 -5.49951613e-01 7.78191090e-02
6.92849874e-01 7.06338957e-02 1.08763778e+00 -1.22366524e+00
5.17094016e-01 1.20583260e+00 7.59526730e-01 -8.72754976e-02
-1.10083497e+00 -6.37492359e-01 -1.40231277e-03 2.07927659e-01
-2.02470231e+00 -5.72412729e-01 6.40529096e-01 -4.51261431e-01
8.64925086e-01 2.18317226e-01 4.64188308e-01 7.44544804e-01
2.46670008e-01 7.39866495e-01 9.81661439e-01 -3.69295627e-01
1.25518277e-01 -1.68669611e-01 -2.81271011e-01 9.03049648e-01
2.51683176e-01 1.17319576e-01 -9.44804788e-01 -4.15518224e-01
1.02138650e+00 3.51083100e-01 -7.24838227e-02 -7.46477962e-01
-1.57135487e+00 8.37052763e-01 6.53106153e-01 7.26041198e-02
-3.01327199e-01 4.65161204e-01 1.83412418e-01 -2.23199949e-02
2.81138062e-01 3.41509819e-01 -4.22649831e-01 5.25346119e-03
-1.24506855e+00 1.36827141e-01 5.17010450e-01 1.30715990e+00
1.32691622e+00 -4.91057068e-01 -3.91127974e-01 7.09954262e-01
4.31578189e-01 6.99126065e-01 3.93956482e-01 -9.61425006e-01
2.92512000e-01 6.37388229e-01 2.17685863e-01 -1.50520849e+00
-2.87988573e-01 -2.95940608e-01 -6.64852500e-01 -2.25781664e-01
-4.89070900e-02 2.23038509e-01 -8.93204391e-01 1.62492597e+00
4.05857116e-01 1.54681772e-01 -2.61772484e-01 1.25872850e+00
6.49888635e-01 4.19979095e-01 -1.63652927e-01 1.53894871e-01
1.24802327e+00 -1.34213102e+00 -4.79724526e-01 -1.41534582e-01
4.99521524e-01 -9.41076517e-01 8.59952450e-01 4.86734845e-02
-8.31426680e-01 -5.07649302e-01 -9.26173508e-01 -4.57727373e-01
-3.58960599e-01 4.90548790e-01 8.57931972e-01 2.21179202e-01
-1.29313397e+00 2.49166906e-01 -8.63680720e-01 -3.47323298e-01
1.43261179e-01 5.14863491e-01 -6.22915983e-01 -2.12343216e-01
-7.83634543e-01 6.42000377e-01 2.16189653e-01 8.21404625e-03
-9.70975220e-01 -5.31020939e-01 -1.08337140e+00 2.06284225e-01
2.88183510e-01 -9.06643808e-01 1.07486248e+00 -4.62175459e-01
-1.37230003e+00 6.16181374e-01 -3.89636159e-01 -1.48294091e-01
2.06085533e-01 -3.54417652e-01 5.88697270e-02 3.97608787e-01
4.38896984e-01 7.49125540e-01 9.96938467e-01 -8.87212813e-01
-5.77785194e-01 -4.59470898e-01 1.09186076e-01 2.77086109e-01
-4.72349375e-01 6.85802177e-02 -1.02671039e+00 -5.62316060e-01
3.64932895e-01 -8.84159863e-01 -3.33843768e-01 3.25335979e-01
-3.22040528e-01 -2.51283318e-01 7.69219518e-01 -3.96752059e-01
1.16471183e+00 -2.08160186e+00 3.99270505e-02 3.37220073e-01
5.36303043e-01 -6.52589574e-02 -2.99443662e-01 6.00477040e-01
4.88554627e-01 -1.26189560e-01 3.23808938e-01 -5.43090582e-01
1.10996649e-01 -6.24369979e-02 -7.20691532e-02 6.51069462e-01
1.58427536e-01 1.13554823e+00 -1.03381503e+00 -7.52664506e-01
4.57070231e-01 5.54293454e-01 -5.17385185e-01 2.47716516e-01
1.97207481e-01 3.57655466e-01 -7.68628240e-01 1.07366872e+00
7.67386198e-01 -4.65604931e-01 -4.95344549e-02 -5.98136187e-01
-1.08575307e-01 7.06209615e-02 -1.00929964e+00 2.26233959e+00
-7.55474031e-01 4.05538619e-01 2.05278341e-02 -4.87472534e-01
8.99410307e-01 -1.35770872e-01 5.05231678e-01 -7.23081708e-01
8.92532170e-02 2.80049980e-01 -8.79943907e-01 -1.52139962e-01
7.21269548e-01 6.55475318e-01 -3.93725514e-01 1.66115075e-01
2.61385888e-01 3.96263599e-02 -4.40202057e-02 2.48229131e-01
1.21026015e+00 2.44176358e-01 5.37103176e-01 -3.25750440e-01
4.93254095e-01 -9.29777771e-02 3.77102345e-01 8.52202058e-01
-3.85332182e-02 8.40514660e-01 9.16559696e-02 -4.73737806e-01
-9.15025234e-01 -1.01324177e+00 7.93391280e-03 8.84551704e-01
6.67036057e-01 -1.05380929e+00 -4.53534663e-01 -5.05958557e-01
1.30440801e-01 -2.13708773e-01 -2.90591598e-01 -2.94523090e-02
-3.03333312e-01 -2.10962564e-01 3.81912708e-01 5.08545816e-01
5.30203402e-01 -3.18716586e-01 -7.70140469e-01 -9.71531793e-02
-8.41203555e-02 -1.23224580e+00 -8.93323720e-01 4.65248711e-02
-6.16937637e-01 -9.81351852e-01 -7.88299084e-01 -6.89529657e-01
8.66170406e-01 8.91763628e-01 8.54452491e-01 3.45129723e-04
-4.04333293e-01 5.24252594e-01 -3.66140425e-01 1.82236359e-01
5.01980245e-01 3.29173952e-01 1.04202203e-01 8.22526519e-04
2.93615043e-01 -4.22260523e-01 -9.74988997e-01 4.93383378e-01
-7.83939898e-01 -1.58631280e-01 9.59584951e-01 1.01650023e+00
9.16778743e-01 -3.80784541e-01 4.39023077e-02 -2.62127429e-01
5.63231230e-01 -3.31123680e-01 -1.01646042e+00 4.74426180e-01
-4.86370921e-01 3.04923564e-01 4.14548129e-01 -3.85668010e-01
-5.11558771e-01 6.30910993e-01 1.67228311e-01 -1.00996339e+00
1.71421245e-01 5.69672048e-01 -1.55824319e-01 -8.37228954e-01
4.90972221e-01 3.64441395e-01 -6.24169894e-02 -4.87662792e-01
6.17249191e-01 6.92860961e-01 4.88274664e-01 -6.32406414e-01
9.65782940e-01 3.99625719e-01 2.74355561e-02 -4.74996716e-01
-8.36247027e-01 -8.08053434e-01 -6.69439077e-01 2.64970828e-02
5.94001949e-01 -1.30605066e+00 -8.15260231e-01 3.19334298e-01
-1.17537904e+00 1.13850921e-01 8.58655348e-02 5.24378538e-01
-4.75407571e-01 4.28764343e-01 -4.53207076e-01 -4.11256820e-01
-3.27952534e-01 -1.40160716e+00 1.95800543e+00 2.21667796e-01
1.25668600e-01 -7.04098642e-01 9.91966426e-02 1.35189682e-01
6.93493247e-01 1.81626454e-01 3.35738301e-01 -1.95847034e-01
-1.17229664e+00 -4.42429841e-01 -6.58386886e-01 3.32140960e-02
3.86815257e-02 -3.33161473e-01 -8.00085783e-01 -6.42075658e-01
-4.58457917e-01 -5.05065143e-01 7.35610485e-01 1.51511759e-01
1.16665864e+00 -3.57531965e-01 -5.78776300e-01 1.01172900e+00
1.58162236e+00 -2.90138572e-01 3.46150815e-01 3.46344203e-01
7.61300087e-01 6.96106926e-02 7.83775091e-01 5.63140392e-01
7.16097295e-01 1.11175394e+00 3.49370003e-01 -1.81832790e-01
-9.57015827e-02 -5.38039386e-01 1.73400551e-01 7.45777965e-01
3.12829703e-01 1.21041639e-02 -8.39354694e-01 5.32895088e-01
-2.11197138e+00 -4.41131920e-01 3.64492238e-01 2.25032496e+00
6.05535507e-01 -4.86530185e-01 -1.38608545e-01 -3.66064221e-01
6.29581630e-01 2.28301406e-01 -3.91046822e-01 1.88581273e-01
8.79653618e-02 3.74976397e-02 8.00900102e-01 5.18104851e-01
-1.35182881e+00 1.25859153e+00 6.55089808e+00 1.06464159e+00
-1.23008549e+00 1.97074160e-01 2.54359245e-01 -3.51584591e-02
-1.05060801e-01 1.83800906e-01 -9.90926623e-01 2.81112462e-01
4.61262673e-01 -9.18020681e-02 4.51495916e-01 1.15279508e+00
-2.22938254e-01 -2.05979019e-01 -1.07417274e+00 1.57621527e+00
3.15777540e-01 -1.42641509e+00 1.57660708e-01 1.34161785e-01
7.00274527e-01 7.65154660e-02 -2.97234431e-02 4.45144344e-03
7.70379305e-02 -8.31026793e-01 8.29620361e-01 4.52688426e-01
1.09481990e+00 -6.73660517e-01 6.31474614e-01 2.04062045e-01
-1.51310456e+00 1.33152883e-02 -4.07941490e-01 1.51461303e-01
2.59103775e-02 7.13284135e-01 -6.42445207e-01 6.65185571e-01
7.87488580e-01 8.12563598e-01 -8.77098143e-01 1.24699950e+00
-2.08441198e-01 -9.37933698e-02 -5.90907454e-01 1.57291684e-02
2.51312643e-01 7.13027790e-02 2.74584353e-01 1.16577113e+00
5.07488310e-01 -2.07407355e-01 5.66498995e-01 7.34781742e-01
2.26674154e-02 -6.73525855e-02 -6.89948320e-01 1.61299080e-01
8.74490261e-01 1.67314625e+00 -5.10151744e-01 -1.34466350e-01
-3.76420557e-01 1.11756492e+00 4.65025663e-01 2.65398353e-01
-8.11424673e-01 -6.89110518e-01 6.98545456e-01 -3.90275866e-02
3.46004605e-01 -6.81682289e-01 1.83024973e-01 -1.45965660e+00
1.34678736e-01 -5.01405537e-01 5.40222935e-02 -8.45102072e-01
-1.07864141e+00 6.68247759e-01 2.86544692e-02 -1.35349512e+00
-3.29360396e-01 -4.14667040e-01 -7.78072327e-02 7.48185873e-01
-1.60273695e+00 -1.58376789e+00 -8.28511119e-01 9.16731715e-01
2.78314739e-01 -1.51867117e-03 6.90836370e-01 4.30987954e-01
-1.84326619e-01 9.08071339e-01 9.01878551e-02 1.19441487e-01
9.07549083e-01 -8.22285473e-01 1.62610084e-01 7.90184438e-01
2.90519148e-01 7.95062482e-01 1.02643602e-01 -2.83674181e-01
-1.96417820e+00 -1.18425953e+00 9.57140803e-01 -6.01494193e-01
6.28598273e-01 -5.92808604e-01 -1.80665061e-01 5.55882394e-01
-2.02400088e-01 5.39974511e-01 3.76738667e-01 4.06327806e-02
-5.75676262e-01 -4.56928790e-01 -9.60357428e-01 4.07321244e-01
1.17253661e+00 -8.94639969e-01 -1.85855299e-01 5.05116582e-01
6.83666825e-01 -6.65563822e-01 -8.88834715e-01 3.46289307e-01
6.80315435e-01 -8.98828924e-01 1.24547589e+00 7.58415312e-02
4.40824265e-03 -7.08198309e-01 -5.33369303e-01 -8.65985334e-01
-5.62976599e-01 -5.60718358e-01 -2.03668728e-01 1.06861889e+00
4.69680391e-02 -4.39697951e-01 5.83536148e-01 2.21354634e-01
2.05444060e-02 -6.53019130e-01 -9.17612672e-01 -8.22044432e-01
-5.91147244e-01 -1.10102564e-01 7.28111267e-01 6.91170037e-01
-1.87542930e-01 3.31159413e-01 -4.91284817e-01 3.96486074e-01
5.76116920e-01 3.66819978e-01 1.04084957e+00 -9.43268001e-01
-1.90909170e-02 -6.22774921e-02 -9.59867537e-01 -1.54928231e+00
1.38142437e-01 -7.27953613e-01 1.03262447e-01 -1.43527126e+00
4.97957200e-01 -5.41053236e-01 -3.54785681e-01 7.37762749e-01
-1.25507619e-02 5.22526383e-01 2.27783367e-01 6.17075384e-01
-1.36189222e+00 4.85789061e-01 9.62512076e-01 -4.91104275e-02
-7.98530057e-02 -3.11975181e-01 -5.44363856e-01 2.53347516e-01
3.80702436e-01 -3.99877757e-01 -4.22550052e-01 -7.59007573e-01
1.38549194e-01 8.04898292e-02 6.50526106e-01 -1.01535690e+00
7.06717730e-01 -5.85601404e-02 5.53467274e-01 -6.51455998e-01
4.99003589e-01 -8.23392689e-01 1.27068371e-01 1.58746600e-01
-7.55951628e-02 4.24413025e-01 -2.16808602e-01 6.40697777e-01
-6.70779586e-01 2.65296847e-01 3.98666859e-01 -5.71011491e-02
-9.73778486e-01 5.30666411e-01 9.81579125e-02 -3.31086874e-01
1.08707643e+00 -1.79950088e-01 -2.79600918e-01 -5.24002790e-01
-2.35695571e-01 2.88029104e-01 6.73558295e-01 4.30374295e-01
8.40327919e-01 -1.64175928e+00 -2.85132855e-01 3.39796513e-01
5.03210664e-01 5.43672778e-02 -2.93126814e-02 1.09888208e+00
-6.98735118e-01 7.61137247e-01 -2.93382648e-02 -1.02608454e+00
-1.15722477e+00 6.14998519e-01 2.79671513e-02 -3.28993559e-01
-2.93628275e-01 7.73275256e-01 2.72628844e-01 -3.94479454e-01
2.56760240e-01 -1.87428042e-01 3.10579449e-01 -9.94700715e-02
5.61219394e-01 1.91453800e-01 1.07118234e-01 -6.11331761e-01
-7.62716889e-01 1.02275777e+00 -2.51370184e-02 7.29565471e-02
1.01223946e+00 -3.79052609e-01 -2.76970178e-01 -1.53080255e-01
1.53483212e+00 7.95754045e-02 -1.26769769e+00 -3.63951802e-01
9.05987099e-02 -8.71405184e-01 4.10303205e-01 -5.79148293e-01
-1.14337373e+00 6.99586153e-01 5.76046407e-01 -3.12966228e-01
1.28053069e+00 2.29868814e-01 7.56058872e-01 4.50667560e-01
1.02538264e+00 -7.50165939e-01 2.15988979e-01 4.05647933e-01
6.77561283e-01 -1.26751328e+00 2.27413803e-01 -3.82307768e-01
-2.42308989e-01 1.02210200e+00 4.74402577e-01 -2.78451085e-01
5.55505455e-01 1.92471400e-01 -2.68200729e-02 -1.39143035e-01
-4.82147992e-01 -2.96935141e-01 5.05589426e-01 4.25660133e-01
7.51662925e-02 -1.44978166e-01 -1.45548537e-01 4.09916490e-01
6.07271083e-02 3.87685420e-03 -2.15988159e-01 1.09016550e+00
-2.70801485e-01 -1.09141755e+00 -3.54238093e-01 1.68783054e-01
-8.06449875e-02 -2.45652616e-01 -3.43044281e-01 6.40845776e-01
2.17355229e-02 7.44273245e-01 -1.75832391e-01 -6.96729898e-01
-3.83055545e-02 -4.99328762e-01 5.02398789e-01 -4.31959331e-01
-2.87191093e-01 2.65377760e-01 -2.97555119e-01 -1.23654389e+00
-4.43223238e-01 -3.98025990e-01 -7.18717813e-01 -2.06672668e-01
-5.98387003e-01 1.12063862e-01 8.97562683e-01 7.01537430e-01
1.06233299e+00 7.96291679e-02 7.66049922e-01 -1.39376247e+00
-5.61147332e-01 -6.32849753e-01 -3.35437417e-01 8.52452666e-02
5.61828613e-01 -8.30238581e-01 -2.56771535e-01 -2.14746132e-01] | [7.831641674041748, -2.044353485107422] |
8ca170dc-671d-4e0f-833e-5f8cd35ea2db | 3d-self-supervised-methods-for-medical | 2006.03829 | null | https://arxiv.org/abs/2006.03829v3 | https://arxiv.org/pdf/2006.03829v3.pdf | 3D Self-Supervised Methods for Medical Imaging | Self-supervised learning methods have witnessed a recent surge of interest after proving successful in multiple application fields. In this work, we leverage these techniques, and we propose 3D versions for five different self-supervised methods, in the form of proxy tasks. Our methods facilitate neural network feature learning from unlabeled 3D images, aiming to reduce the required cost for expert annotation. The developed algorithms are 3D Contrastive Predictive Coding, 3D Rotation prediction, 3D Jigsaw puzzles, Relative 3D patch location, and 3D Exemplar networks. Our experiments show that pretraining models with our 3D tasks yields more powerful semantic representations, and enables solving downstream tasks more accurately and efficiently, compared to training the models from scratch and to pretraining them on 2D slices. We demonstrate the effectiveness of our methods on three downstream tasks from the medical imaging domain: i) Brain Tumor Segmentation from 3D MRI, ii) Pancreas Tumor Segmentation from 3D CT, and iii) Diabetic Retinopathy Detection from 2D Fundus images. In each task, we assess the gains in data-efficiency, performance, and speed of convergence. Interestingly, we also find gains when transferring the learned representations, by our methods, from a large unlabeled 3D corpus to a small downstream-specific dataset. We achieve results competitive to state-of-the-art solutions at a fraction of the computational expense. We publish our implementations for the developed algorithms (both 3D and 2D versions) as an open-source library, in an effort to allow other researchers to apply and extend our methods on their datasets. | ['Julius Severin', 'Aiham Taleb', 'Winfried Loetzsch', 'Noel Danz', 'Christoph Lippert', 'Benjamin Bergner', 'Thomas Gaertner'] | 2020-06-06 | null | http://proceedings.neurips.cc/paper/2020/hash/d2dc6368837861b42020ee72b0896182-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/d2dc6368837861b42020ee72b0896182-Paper.pdf | neurips-2020-12 | ['diabetic-retinopathy-detection'] | ['medical'] | [ 2.82457292e-01 6.11379862e-01 -3.91008973e-01 -4.62134033e-01
-9.01425242e-01 -4.28187728e-01 4.29443032e-01 8.41820911e-02
-3.64375830e-01 5.04528940e-01 2.31653914e-01 -3.08670700e-01
-2.61175901e-01 -3.36011797e-01 -8.38383615e-01 -6.87291205e-01
-2.57496327e-01 7.90910959e-01 7.94903114e-02 2.40165547e-01
1.09993771e-01 5.80611527e-01 -1.34613836e+00 1.89569294e-01
8.28221440e-01 1.03224552e+00 1.98659688e-01 3.88019979e-01
1.05120772e-02 6.57172382e-01 -8.84186551e-02 -5.87358512e-02
5.76634943e-01 -2.57510453e-01 -9.56012785e-01 6.64153576e-01
2.77171403e-01 -1.97671950e-01 -1.44695669e-01 8.22858572e-01
4.31720227e-01 -1.42668828e-01 8.23485255e-01 -8.69945347e-01
-5.83848119e-01 3.19112927e-01 -5.03786147e-01 1.98640488e-02
1.84732020e-01 1.53393283e-01 6.21193647e-01 -6.24352157e-01
7.38564432e-01 7.77865887e-01 8.38529766e-01 6.41363919e-01
-1.25113249e+00 -2.72362202e-01 -3.64341289e-02 5.63882701e-02
-1.14292419e+00 -3.76913995e-01 6.94727957e-01 -5.44636011e-01
1.06046093e+00 -5.47967106e-02 6.42192960e-01 1.23925078e+00
-4.41851914e-02 1.03509343e+00 1.24212587e+00 -6.32719040e-01
3.50563496e-01 7.68318847e-02 1.41071513e-01 1.06608546e+00
1.57278433e-01 2.25380450e-01 -1.92091733e-01 -1.98064566e-01
9.67975855e-01 2.26393044e-02 -3.16490114e-01 -8.55128944e-01
-1.24988985e+00 9.13079500e-01 4.28010553e-01 1.92569196e-01
-3.25743407e-01 -5.09057194e-02 3.36009383e-01 2.79286504e-01
8.15076470e-01 4.57841724e-01 -7.42906451e-01 9.92847905e-02
-7.95329154e-01 -3.74381170e-02 8.45782518e-01 1.14326084e+00
6.01525724e-01 -3.72778475e-01 9.00083110e-02 7.49119103e-01
2.18710303e-01 1.33036837e-01 8.39755952e-01 -9.29192126e-01
1.37465373e-01 7.71884263e-01 -1.20556012e-01 -4.62654144e-01
-9.71221805e-01 -6.26459777e-01 -8.06646347e-01 1.60676464e-01
5.00619948e-01 -1.55007206e-02 -1.16507256e+00 1.42620194e+00
4.22386199e-01 2.86622077e-01 2.86644608e-01 9.87887502e-01
8.23948503e-01 2.64196754e-01 -9.81427878e-02 -2.52610058e-01
1.23663449e+00 -1.22816646e+00 -1.70549020e-01 -1.74312577e-01
1.12573814e+00 -5.74439466e-01 8.05184782e-01 3.88649166e-01
-1.07938612e+00 -4.00477141e-01 -7.28547275e-01 -8.35673437e-02
-1.69624910e-01 3.62642229e-01 8.18275273e-01 4.81125981e-01
-1.20930278e+00 6.87694132e-01 -1.11374664e+00 -6.12145185e-01
7.78307796e-01 4.05385166e-01 -5.67403078e-01 -1.72034144e-01
-6.42503023e-01 1.11115003e+00 2.79346079e-01 -2.10845321e-01
-8.79611015e-01 -8.45475614e-01 -8.90880823e-01 -1.91180125e-01
3.54073733e-01 -9.72493887e-01 1.27152109e+00 -1.03027415e+00
-1.50118756e+00 1.58636534e+00 3.30072902e-02 -6.67839110e-01
5.97739458e-01 -1.80882573e-01 8.94509181e-02 3.20589989e-01
9.82163399e-02 8.49244475e-01 7.63253450e-01 -9.69681144e-01
-3.70679766e-01 -6.26326621e-01 -9.23302174e-02 2.17549145e-01
-8.21050331e-02 -3.80403876e-01 -5.56910157e-01 -6.73106432e-01
3.54324639e-01 -1.09668517e+00 -6.43162072e-01 2.96277106e-01
-4.83522862e-01 -1.84352025e-01 2.57538974e-01 -4.65986580e-01
3.83948416e-01 -2.21845889e+00 2.91235119e-01 1.06755726e-01
2.84752846e-01 2.37939909e-01 -7.05359280e-02 -1.21071100e-01
-2.99144119e-01 -1.78665876e-01 -4.94227678e-01 -3.65992486e-01
-2.18035296e-01 4.77584124e-01 1.09659269e-01 6.02665424e-01
4.64631170e-01 9.41072345e-01 -8.86207700e-01 -4.91781980e-01
1.70241132e-01 2.66772628e-01 -6.88799620e-01 1.62068188e-01
-2.44747370e-01 7.35952079e-01 -5.65056801e-01 6.93232656e-01
3.86813551e-01 -7.37401307e-01 1.95219055e-01 -2.15709910e-01
6.05050214e-02 1.90280810e-01 -7.36684024e-01 2.31734300e+00
-4.29378867e-01 3.16831142e-01 -2.68157929e-01 -1.66202772e+00
9.30339932e-01 3.88081998e-01 6.75646901e-01 -6.67960644e-01
8.12794343e-02 3.69476557e-01 -2.08451629e-01 -8.91453087e-01
-1.49632543e-01 2.59265769e-03 1.63574249e-01 5.25615275e-01
5.15778184e-01 -2.65055478e-01 5.79642057e-02 -1.38667107e-01
1.16263926e+00 6.13533437e-01 7.13410378e-01 -2.56574899e-01
3.24207902e-01 4.41418111e-01 1.98021874e-01 6.31561220e-01
-1.45381510e-01 6.53950155e-01 6.47889912e-01 -7.98044443e-01
-1.07485986e+00 -8.52580011e-01 -3.49262923e-01 6.44648075e-01
-4.34159860e-02 -7.56490976e-02 -7.06068397e-01 -9.90046203e-01
9.21585634e-02 5.05094707e-01 -8.80680263e-01 3.03146187e-02
-6.07829988e-01 -9.03871357e-01 3.08284283e-01 6.34417772e-01
2.87807167e-01 -7.15415657e-01 -8.84829998e-01 3.15809585e-02
2.91971058e-01 -1.13772273e+00 -3.69198769e-02 6.64120197e-01
-1.42765284e+00 -1.29018724e+00 -1.04555738e+00 -1.04695570e+00
9.37659681e-01 2.20991611e-01 1.05183959e+00 -2.96592209e-02
-5.32150090e-01 5.02237082e-01 -4.78254646e-01 -2.93334305e-01
-3.57650578e-01 1.39184594e-01 -8.81857574e-02 -3.32128435e-01
3.02159578e-01 -7.01693475e-01 -5.18311203e-01 3.11955273e-01
-7.38912523e-01 3.24456483e-01 8.78694057e-01 9.86391068e-01
9.35660839e-01 -4.96011823e-01 3.24217409e-01 -1.18933201e+00
1.20272562e-01 -4.84319031e-01 -5.75386763e-01 6.08999021e-02
-7.18551219e-01 2.31077433e-01 4.77720350e-01 -2.76092231e-01
-7.31755733e-01 5.63872039e-01 -3.06153335e-02 -7.16174304e-01
-5.99890471e-01 4.56762046e-01 3.56580436e-01 -1.55842409e-01
9.81907964e-01 3.77543956e-01 4.88207966e-01 -5.57555377e-01
4.92654353e-01 4.84957129e-01 4.29487735e-01 -4.25521463e-01
5.40857494e-01 5.74198186e-01 9.77706686e-02 -4.45136428e-01
-1.12084687e+00 -4.67037916e-01 -8.01058650e-01 2.01898843e-01
8.01307499e-01 -1.02480841e+00 -3.86209220e-01 1.31180599e-01
-9.14116144e-01 -4.78158444e-01 -4.82254654e-01 7.10453928e-01
-1.01326895e+00 3.23595583e-01 -5.37338257e-01 -3.27518225e-01
-2.59515077e-01 -1.39415514e+00 1.20485854e+00 9.75593030e-02
-8.46685767e-02 -1.07469618e+00 6.23478554e-02 4.33363140e-01
2.80126721e-01 2.17715874e-01 1.22813547e+00 -1.01402032e+00
-5.37282169e-01 -8.78377557e-02 -3.10267925e-01 2.92609692e-01
5.83635420e-02 -5.77146769e-01 -8.83623064e-01 -7.89804757e-02
1.74775328e-02 -5.29891372e-01 9.65982676e-01 6.67551458e-01
1.31104827e+00 -3.86252254e-02 -5.50994575e-01 9.63161826e-01
1.21408033e+00 1.37494728e-02 2.67563254e-01 3.75069439e-01
4.04521167e-01 6.73928201e-01 4.05318499e-01 3.09429109e-01
2.46009678e-01 5.84931850e-01 5.36498904e-01 -4.18252409e-01
-2.80063510e-01 5.28006367e-02 -9.23058987e-02 7.58491337e-01
-2.35975757e-01 1.50633648e-01 -1.01511562e+00 5.05694330e-01
-1.91510677e+00 -3.97000909e-01 1.13527037e-01 1.90577340e+00
7.08008468e-01 -2.12833583e-02 1.11279622e-01 -1.47106871e-01
3.63559723e-01 -2.13382989e-01 -7.47429371e-01 2.20715810e-04
1.51538834e-01 3.33195090e-01 4.29469824e-01 5.49194328e-02
-1.39103794e+00 8.35051835e-01 6.39300823e+00 4.30755943e-01
-1.03657627e+00 1.87037885e-01 9.43406641e-01 -1.19221084e-01
2.35563386e-02 -1.37457892e-01 -4.82642382e-01 5.88885322e-02
6.91652715e-01 2.57639736e-01 1.77355543e-01 1.17508447e+00
7.61729851e-02 4.09396291e-02 -1.42761064e+00 1.12734139e+00
3.19255710e-01 -1.64168096e+00 -1.70271501e-01 -6.07401580e-02
6.07683241e-01 2.64068305e-01 -1.30536005e-01 2.13692233e-01
1.50981098e-01 -1.04180491e+00 3.20057482e-01 4.22766328e-01
7.94723749e-01 -4.69194084e-01 7.30559647e-01 3.17939937e-01
-4.85485673e-01 -2.01934297e-02 -3.79101962e-01 1.58354998e-01
-2.60353573e-02 6.58065259e-01 -1.06014454e+00 5.87762654e-01
7.07579851e-01 9.47016537e-01 -4.14405763e-01 1.14034641e+00
-1.54880345e-01 5.12740433e-01 -1.78727418e-01 3.29998769e-02
3.90921861e-01 -1.39766231e-01 4.03503239e-01 1.01743841e+00
2.55326986e-01 9.74691659e-02 1.70358121e-01 8.45097959e-01
2.09931135e-02 1.68751106e-01 -6.90962851e-01 1.92546621e-01
-4.97947857e-02 1.19514453e+00 -8.57467055e-01 -2.44652376e-01
-5.20387828e-01 9.96696651e-01 4.52619314e-01 2.24791810e-01
-6.95258617e-01 -1.93212796e-02 3.54180902e-01 -1.05592780e-01
4.62648004e-01 -2.88678799e-02 -2.69597501e-01 -1.14053202e+00
-1.31360367e-01 -7.69720376e-01 4.15297568e-01 -8.82913649e-01
-1.59241426e+00 6.37315989e-01 -5.01175411e-02 -1.35084176e+00
-3.55202377e-01 -1.03708732e+00 -1.64758861e-01 4.70123857e-01
-1.75918162e+00 -1.09080160e+00 -3.11929911e-01 6.70783401e-01
4.74465638e-01 -3.94878238e-01 1.11721587e+00 8.26319978e-02
-3.02295417e-01 3.40630502e-01 3.89802679e-02 1.09728724e-01
7.09314287e-01 -1.21954632e+00 1.72868103e-01 2.58221656e-01
2.86570638e-01 1.44606948e-01 1.79976940e-01 -2.45646730e-01
-1.37737191e+00 -1.07210231e+00 6.09506488e-01 -3.70731205e-01
5.60523748e-01 -8.58882666e-02 -6.23503864e-01 8.11090350e-01
-1.26844883e-01 4.47290927e-01 9.48007405e-01 1.29883707e-01
-1.74018145e-01 1.27102494e-01 -1.27722061e+00 3.76623213e-01
1.42166793e+00 -1.34745732e-01 -7.23925292e-01 8.85691464e-01
8.53033662e-01 -6.77266359e-01 -9.82955992e-01 4.06576127e-01
3.21663052e-01 -9.63756800e-01 9.91779387e-01 -7.76000261e-01
6.98899209e-01 -2.94185504e-02 -1.53044835e-01 -1.30273366e+00
-2.64687777e-01 -4.54951257e-01 -8.11039358e-02 4.00128394e-01
5.40577352e-01 -5.92875540e-01 9.93701220e-01 3.14556807e-01
-6.70732379e-01 -1.06931674e+00 -9.56069410e-01 -4.84036088e-01
6.85944781e-02 -4.09218997e-01 1.41187757e-01 1.00836587e+00
1.44168744e-02 2.13131949e-01 -2.81068180e-02 8.73178914e-02
6.03231788e-01 5.08453488e-01 6.76318109e-01 -1.23920155e+00
-5.01562774e-01 -4.10499394e-01 -6.95387423e-01 -1.13017595e+00
2.09273860e-01 -1.36035776e+00 -1.88759699e-01 -1.33666134e+00
3.86630774e-01 -5.80991566e-01 -2.24715155e-02 8.40297580e-01
1.18580416e-01 2.21181482e-01 -5.58405258e-02 3.88550848e-01
-5.31455398e-01 2.65438914e-01 1.43060780e+00 -1.66751280e-01
-1.91954449e-01 1.29189849e-01 -7.09604561e-01 8.71860206e-01
6.19150758e-01 -4.69978034e-01 -3.49800855e-01 -4.96890664e-01
-1.68613061e-01 2.57299066e-01 5.50619185e-01 -9.05134439e-01
1.35774791e-01 2.33231649e-01 6.46228135e-01 -3.29832971e-01
2.34408051e-01 -7.86080599e-01 -2.43144885e-01 4.94503349e-01
-4.50027883e-01 -3.26744616e-01 1.74592331e-01 4.91850853e-01
-5.78921437e-02 -3.69115144e-01 7.51742482e-01 -5.09639859e-01
-6.40447915e-01 3.80122185e-01 -1.68951556e-01 -4.70711850e-02
1.15276539e+00 -2.65232205e-01 -1.78778023e-01 -9.65849459e-02
-1.20663583e+00 8.34885538e-02 4.29436952e-01 1.26277864e-01
5.72182894e-01 -9.07354772e-01 -6.75430417e-01 4.46406484e-01
1.84699461e-01 3.22660059e-01 1.61183566e-01 1.20192575e+00
-6.64273560e-01 5.84226906e-01 -3.10086757e-01 -1.13992596e+00
-7.74786592e-01 7.06437290e-01 4.45145190e-01 -4.52194840e-01
-1.02440560e+00 7.28118241e-01 2.88382083e-01 -6.68824315e-01
3.42588425e-01 -5.54661274e-01 -2.04764247e-01 -2.31377915e-01
3.44834715e-01 -5.12349308e-02 2.62695968e-01 -1.15251042e-01
-2.29097277e-01 7.44427085e-01 -8.76373872e-02 3.38675797e-01
1.71298480e+00 1.21384338e-01 2.12518886e-01 1.10544808e-01
1.20314991e+00 -5.30536354e-01 -1.34798419e+00 -4.79216278e-01
1.42047063e-01 -9.43156406e-02 2.44721770e-01 -8.38333189e-01
-1.25554109e+00 5.95241427e-01 8.05336893e-01 1.53965561e-03
1.18984222e+00 4.17331368e-01 5.23585856e-01 5.12630641e-01
4.42484260e-01 -6.44354641e-01 1.07226036e-01 2.83398271e-01
6.40850008e-01 -1.35358477e+00 8.00108463e-02 -4.65940952e-01
-7.00317264e-01 1.18195295e+00 3.17129374e-01 -4.43870902e-01
7.21040070e-01 1.86179727e-01 2.19763443e-01 -3.14302742e-01
-7.25434065e-01 -2.67365545e-01 3.41811478e-01 8.78790498e-01
4.25746202e-01 -1.06464162e-01 -4.13181596e-02 5.32495558e-01
1.40633821e-01 2.28598312e-01 3.32113326e-01 8.59474480e-01
-1.05519928e-01 -1.06046772e+00 -1.44638762e-01 5.53250194e-01
-1.83750302e-01 -6.90408098e-03 -1.32674605e-01 1.02535963e+00
9.46408138e-02 2.84461290e-01 1.72418088e-01 3.65452133e-02
2.38229156e-01 2.24589005e-01 8.55087042e-01 -6.93844378e-01
-3.51509362e-01 2.10740477e-01 5.52455820e-02 -7.22648859e-01
-8.99522364e-01 -6.42912745e-01 -1.23814344e+00 4.64093298e-01
-1.26752838e-01 -1.58601403e-01 7.29254961e-01 1.06763828e+00
6.24829590e-01 3.32635790e-01 4.60552931e-01 -1.12452197e+00
-7.05993295e-01 -9.58297312e-01 -4.44408596e-01 4.44824785e-01
8.83053839e-02 -8.65565777e-01 -1.82723284e-01 2.54647940e-01] | [14.707958221435547, -2.3294148445129395] |
60c9db0d-a473-4791-81fe-201ab016f6fe | qurg-question-rewriting-guided-context | 2305.06655 | null | https://arxiv.org/abs/2305.06655v2 | https://arxiv.org/pdf/2305.06655v2.pdf | QURG: Question Rewriting Guided Context-Dependent Text-to-SQL Semantic Parsing | Context-dependent Text-to-SQL aims to translate multi-turn natural language questions into SQL queries. Despite various methods have exploited context-dependence information implicitly for contextual SQL parsing, there are few attempts to explicitly address the dependencies between current question and question context. This paper presents QURG, a novel Question Rewriting Guided approach to help the models achieve adequate contextual understanding. Specifically, we first train a question rewriting model to complete the current question based on question context, and convert them into a rewriting edit matrix. We further design a two-stream matrix encoder to jointly model the rewriting relations between question and context, and the schema linking relations between natural language and structured schema. Experimental results show that QURG significantly improves the performances on two large-scale context-dependent datasets SParC and CoSQL, especially for hard and long-turn questions. | ['Zhao Yan', 'Zhoujun Li', 'Yunbo Cao', 'Qian-Wen Zhang', 'Liqun Yang', 'Jian Yang', 'Dongling Xiao', 'Linzheng Chai'] | 2023-05-11 | null | null | null | null | ['text-to-sql', 'semantic-parsing', 'question-rewriting'] | ['computer-code', 'natural-language-processing', 'natural-language-processing'] | [ 3.05979431e-01 2.90269196e-01 -9.29235891e-02 -1.09402180e+00
-1.22729051e+00 -9.39494014e-01 4.42154646e-01 4.23526555e-01
-3.68021727e-01 3.24667811e-01 4.76260900e-01 -8.54528010e-01
4.63646725e-02 -1.26887929e+00 -1.05665648e+00 5.45985818e-01
5.28467834e-01 3.94594222e-01 4.63954508e-01 -6.22077882e-01
3.76733281e-02 1.84373390e-02 -1.34720278e+00 9.67858136e-01
1.30735576e+00 7.10856616e-01 2.57358760e-01 8.73553097e-01
-1.03187966e+00 1.42233348e+00 -5.40539324e-01 -1.05930042e+00
-2.26516370e-02 -6.94648027e-01 -1.23366439e+00 -2.53662497e-01
6.49996340e-01 -1.50863037e-01 -1.42690822e-01 8.34767461e-01
4.67102863e-02 1.23803066e-02 -6.16771430e-02 -8.09014320e-01
-9.87617552e-01 9.29003835e-01 2.68953532e-01 7.96385482e-02
9.85583544e-01 -9.44797695e-03 1.59063578e+00 -8.69052947e-01
9.06314135e-01 1.22091985e+00 4.86522555e-01 4.87243384e-01
-9.11237538e-01 -6.39377534e-02 3.08245629e-01 4.64917690e-01
-7.82172561e-01 -1.85764596e-01 5.56974471e-01 8.16858001e-03
1.54507339e+00 6.94498479e-01 3.39436680e-01 7.84173012e-01
2.08880723e-01 5.15672147e-01 9.35984790e-01 -5.65181315e-01
-4.91022058e-02 2.47398913e-01 4.95147854e-01 8.46360803e-01
-8.02971199e-02 -1.69549838e-01 -3.23664635e-01 -1.20789938e-01
1.19647585e-01 -2.54342139e-01 1.43540725e-01 -6.22061640e-03
-8.52639556e-01 9.43153679e-01 2.00075358e-01 1.48272976e-01
-9.47674736e-02 5.58031537e-02 5.33593833e-01 7.34057248e-01
-2.01623857e-01 9.01396096e-01 -7.57258415e-01 -7.12532476e-02
-1.79620951e-01 5.19392312e-01 1.38761330e+00 1.61392105e+00
1.04511034e+00 -5.22395432e-01 -3.67732406e-01 7.62766480e-01
2.58407652e-01 7.10345209e-01 1.02870442e-01 -1.26166832e+00
9.92833793e-01 1.30820096e+00 -1.08783990e-01 -9.39157069e-01
-1.58986431e-02 1.24409974e-01 -1.14089608e-01 -6.75008833e-01
1.37634769e-01 -6.12401776e-02 -5.83361268e-01 1.66275930e+00
5.46699941e-01 -5.00400901e-01 3.94654274e-01 6.65292680e-01
1.32087648e+00 6.77244842e-01 2.23829150e-01 3.93640697e-02
1.86177087e+00 -8.90071154e-01 -8.56009603e-01 -6.06914461e-01
7.88007021e-01 -7.57980227e-01 1.78321409e+00 -1.89254284e-01
-1.07962000e+00 -5.66324472e-01 -8.90123427e-01 -9.74724948e-01
-5.64603746e-01 -4.61467877e-02 5.55894554e-01 4.01134968e-01
-4.54192936e-01 1.25493538e-02 -5.52416444e-01 -5.08906007e-01
-1.13103226e-01 -1.45736590e-01 -1.70757100e-01 -3.41808081e-01
-1.54222620e+00 8.41896176e-01 3.75914067e-01 -8.17194465e-04
-4.33693826e-01 -9.53293324e-01 -1.08734190e+00 1.46227196e-01
1.02331328e+00 -8.26679111e-01 1.53334320e+00 -3.15575331e-01
-1.27226484e+00 9.25961673e-01 -5.18094122e-01 -4.31217700e-01
4.48895469e-02 -5.13590038e-01 -6.11861825e-01 -2.46177372e-02
3.48709881e-01 2.77714282e-01 1.00200199e-01 -1.02828622e+00
-6.86328113e-01 -4.07220423e-01 7.16083229e-01 2.12917164e-01
1.72361076e-01 3.47823560e-01 -9.64448094e-01 -2.71151602e-01
5.05271219e-02 -5.82644403e-01 -2.25912377e-01 -3.24809104e-01
-5.27516842e-01 -4.48816597e-01 3.86451274e-01 -8.59451830e-01
1.64807189e+00 -1.54325294e+00 5.13341976e-04 8.03378876e-03
-1.29562780e-01 1.96410462e-01 -2.38538548e-01 1.02234030e+00
-3.49738449e-02 1.66406542e-01 -5.56772292e-01 1.78398922e-01
2.58499682e-01 7.82716274e-01 -7.68915653e-01 -6.72868431e-01
5.20093799e-01 1.45939100e+00 -6.29735112e-01 -7.02106297e-01
-2.02084646e-01 -1.56430706e-01 -8.55758190e-01 8.86574090e-01
-1.21658862e+00 -2.53868818e-01 -4.83777434e-01 5.77003837e-01
4.91836905e-01 -2.52646178e-01 5.84783137e-01 -3.62204015e-01
1.95662737e-01 8.73425186e-01 -1.04025531e+00 1.87630558e+00
-7.66976357e-01 2.71444201e-01 -1.19797535e-01 -5.31729817e-01
1.02572286e+00 2.33348808e-03 -2.81077567e-02 -1.07863176e+00
-4.63443905e-01 1.59504712e-01 -4.73906398e-01 -1.32180572e+00
7.10706413e-01 -1.78527862e-01 -5.67204595e-01 4.10465658e-01
-1.02805339e-01 -4.47644800e-01 4.78512198e-01 5.53607583e-01
1.10425758e+00 3.89594436e-01 3.01151603e-01 1.20428234e-01
9.61252630e-01 3.17854971e-01 6.76873624e-01 7.16065049e-01
3.13623756e-01 3.72854233e-01 9.20261919e-01 -4.22726244e-01
-6.98284507e-01 -1.16844940e+00 2.68341810e-01 1.45542777e+00
2.10988950e-02 -9.76922452e-01 -7.87385285e-01 -1.12944090e+00
1.41900986e-01 1.26228726e+00 -5.74996114e-01 -6.47912696e-02
-1.21872568e+00 -1.65816277e-01 7.09271431e-01 5.79390109e-01
3.81783396e-01 -1.18525577e+00 -5.81123829e-01 4.77267027e-01
-6.12639725e-01 -1.57379282e+00 -6.73718214e-01 1.18507139e-01
-6.30544066e-01 -1.53393948e+00 5.77086270e-01 -8.10592771e-01
2.29791477e-01 -1.26297072e-01 1.74836361e+00 2.57360458e-01
-9.55894068e-02 5.49503267e-01 -4.30567801e-01 -3.44556093e-01
-8.16001475e-01 3.43088061e-01 -9.82829273e-01 -2.53816128e-01
8.46887887e-01 -2.40094766e-01 -2.70538390e-01 2.83682436e-01
-1.38803840e+00 1.52870817e-02 3.02615970e-01 5.10440350e-01
8.68980289e-01 -7.50522435e-01 7.47849882e-01 -1.76980758e+00
9.06485736e-01 -3.20626110e-01 -5.80038428e-01 1.11722958e+00
-7.44337916e-01 4.74714428e-01 8.89068902e-01 2.34115377e-01
-1.53521907e+00 -1.45025477e-01 -5.03552020e-01 2.40153387e-01
-1.81716800e-01 9.78349507e-01 -6.32817626e-01 5.95669806e-01
8.17257106e-01 1.28115788e-01 -3.33839536e-01 -4.44077343e-01
1.02544045e+00 4.41914767e-01 8.14465940e-01 -1.11336029e+00
6.89817727e-01 8.34979787e-02 -3.14069808e-01 -3.41320574e-01
-1.25169134e+00 -3.97492349e-01 -4.14367229e-01 1.95534006e-01
1.08100414e+00 -6.37809694e-01 -5.71231008e-01 -1.90199941e-01
-1.40699756e+00 -1.13814697e-01 -6.74206078e-01 -1.45122483e-01
-5.89365304e-01 5.11140108e-01 -5.08376360e-01 -3.75468969e-01
-4.96679813e-01 -8.90413702e-01 8.78036022e-01 1.86779276e-01
-6.65212870e-02 -9.83889699e-01 4.20588791e-01 6.41960621e-01
3.32441777e-01 1.46899939e-01 1.68353868e+00 -6.24785841e-01
-9.73618090e-01 -4.48802970e-02 -3.19723040e-01 3.80647480e-01
1.90500468e-01 -5.38611002e-02 -5.10469854e-01 1.94696411e-01
9.99008492e-02 -5.18350780e-01 3.39990586e-01 -6.67662382e-01
1.02250910e+00 -5.00965059e-01 9.33005586e-02 6.72162473e-01
1.70378661e+00 2.61521786e-01 7.32448757e-01 4.77193184e-02
7.06004143e-01 7.02154100e-01 6.82928383e-01 -6.56500645e-03
1.22961771e+00 4.06872183e-01 3.39432389e-01 3.92087996e-01
-1.84254467e-01 -8.62366080e-01 5.99504635e-02 1.26439703e+00
7.60326505e-01 6.84113894e-03 -9.41976666e-01 5.15939176e-01
-1.65699828e+00 -6.31898522e-01 -4.82046008e-01 1.69705939e+00
1.32335413e+00 -1.83861315e-01 -5.24422348e-01 -5.53374946e-01
2.30059698e-01 2.38906547e-01 -7.13156819e-01 -7.94515908e-01
-3.02039862e-01 5.23283720e-01 6.29113242e-02 8.21798086e-01
-6.78493857e-01 1.18212795e+00 5.85056114e+00 2.47123897e-01
-6.27437055e-01 2.79573370e-02 -3.49227451e-02 6.11198228e-03
-1.18318331e+00 5.97778201e-01 -8.84723902e-01 6.03803545e-02
1.00204706e+00 -4.28656638e-01 4.48746800e-01 6.56120896e-01
-2.62665123e-01 -3.86454836e-02 -1.49296117e+00 4.35122371e-01
7.15666860e-02 -1.38067651e+00 4.00076330e-01 -5.92798293e-01
5.12342572e-01 -2.24806607e-01 -4.40504998e-01 8.80818903e-01
3.79849046e-01 -8.21118236e-01 3.42801034e-01 9.35517967e-01
7.60781109e-01 -6.30050421e-01 4.92228180e-01 2.78415620e-01
-1.27323508e+00 -1.35955140e-01 -4.92807418e-01 4.06341016e-01
1.20247759e-01 3.15125316e-01 -6.24275208e-01 9.86565650e-01
5.56016922e-01 2.72429943e-01 -1.10485625e+00 1.83335558e-01
-5.92263281e-01 5.57068348e-01 -1.22316360e-01 -2.78269291e-01
9.04116780e-02 -3.13426733e-01 1.80981353e-01 1.28660285e+00
-5.76359443e-02 4.48780477e-01 1.14612384e-02 1.01500106e+00
-3.41826946e-01 4.40926641e-01 -4.55283940e-01 -1.16374791e-01
4.55977350e-01 9.67018783e-01 2.57857472e-01 -5.89537024e-01
-9.10751462e-01 7.24138319e-01 7.46594846e-01 4.06528085e-01
-8.97780001e-01 -8.07276607e-01 5.84676564e-01 6.76484406e-02
3.37427974e-01 -1.19851582e-01 -4.83235747e-01 -1.53004849e+00
8.39948654e-01 -1.51383507e+00 9.06527340e-01 -8.93113375e-01
-9.84404683e-01 5.62679052e-01 -1.19998820e-01 -4.93896335e-01
-3.92818302e-01 -2.98551381e-01 -5.96760094e-01 8.73019814e-01
-1.61856019e+00 -1.14979362e+00 -2.62451977e-01 5.82509339e-01
5.06087303e-01 1.07276089e-01 9.59301293e-01 3.76088917e-01
-4.64548409e-01 6.94156170e-01 -4.17594463e-01 2.17445284e-01
7.62202621e-01 -1.47770214e+00 8.27792585e-01 9.84875739e-01
2.36523807e-01 1.21510148e+00 4.58884776e-01 -7.42227912e-01
-1.98358965e+00 -1.30329025e+00 1.49205637e+00 -9.00436699e-01
5.21024287e-01 -4.76676941e-01 -1.46873724e+00 9.15343225e-01
5.96084952e-01 -2.35865638e-01 8.35394859e-01 1.85150459e-01
-9.65800524e-01 -4.19148475e-01 -8.00802171e-01 6.02175653e-01
1.07361889e+00 -1.17400622e+00 -1.35949898e+00 2.21262187e-01
1.38967538e+00 -5.38849175e-01 -1.11077380e+00 4.70002860e-01
2.71493316e-01 -9.34996486e-01 6.58053041e-01 -1.37526822e+00
7.61727989e-01 -3.75380397e-01 -7.38193035e-01 -7.90913403e-01
1.35714129e-01 -4.66385871e-01 -3.19650233e-01 1.41537535e+00
7.82014728e-01 -4.17080998e-01 8.66239309e-01 9.37893867e-01
-1.54495403e-01 -9.87135649e-01 -6.81613386e-01 -3.63238394e-01
2.84229785e-01 -7.05116749e-01 1.17883682e+00 7.91761756e-01
5.87982498e-02 7.56967545e-01 1.20859832e-01 2.15907142e-01
1.85560733e-01 8.79037917e-01 9.01833236e-01 -6.25773251e-01
-4.80489254e-01 1.30787253e-01 4.26972121e-01 -1.38090980e+00
5.09486347e-02 -1.05982649e+00 6.03883378e-02 -1.96326375e+00
-6.52315915e-02 -2.89596140e-01 3.04707646e-01 2.74984509e-01
-7.90287197e-01 -8.34043086e-01 1.75830141e-01 -2.36881509e-01
-7.71396518e-01 3.98149580e-01 1.14965320e+00 -3.86006534e-02
1.02125818e-03 -2.44825602e-01 -1.08213186e+00 1.55039549e-01
5.18948197e-01 -5.53451836e-01 -7.39544749e-01 -1.08603287e+00
1.00209665e+00 6.84220374e-01 4.53414507e-02 -6.88153028e-01
5.70037961e-01 -6.60033107e-01 -3.28519881e-01 -7.58571267e-01
-1.06509440e-01 -7.57225275e-01 -7.52425566e-02 1.85148288e-02
-9.84461069e-01 5.71049333e-01 3.73538584e-02 4.54014331e-01
-4.69707429e-01 -4.77260441e-01 4.47400779e-01 -3.76046687e-01
-5.89017630e-01 1.89736411e-01 -1.29218802e-01 1.15909469e+00
7.14826882e-01 4.27541167e-01 -5.17342329e-01 -3.19718689e-01
-6.57544553e-01 6.56779528e-01 1.71004608e-01 7.32965231e-01
5.88513792e-01 -1.24117708e+00 -6.43946826e-01 2.96520412e-01
5.45999289e-01 2.32240707e-01 1.04512818e-01 5.07729985e-02
-7.21452057e-01 7.83632517e-01 2.09807813e-01 -2.47746125e-01
-9.73385394e-01 6.90363526e-01 5.55755913e-01 -6.22086048e-01
-1.83913976e-01 7.82411456e-01 -4.80775908e-02 -1.40830481e+00
6.48799390e-02 -8.43576670e-01 -5.94051741e-02 -1.40386000e-01
3.57303321e-01 -1.76290050e-02 3.41739237e-01 -2.70536300e-02
-3.11580509e-01 5.85312724e-01 -1.93129390e-01 -6.61351681e-02
9.37065005e-01 -1.97135523e-01 -4.82487053e-01 2.29222789e-01
1.43574429e+00 2.05857515e-01 -7.17622221e-01 -7.64720500e-01
6.25180125e-01 -2.90915906e-01 -8.44177961e-01 -1.07821167e+00
-5.76571941e-01 9.19008791e-01 -5.31752445e-02 3.74930620e-01
1.04825354e+00 8.41800719e-02 1.10581994e+00 1.23336852e+00
1.19722739e-01 -1.05566728e+00 1.11631520e-01 1.05619836e+00
1.22346365e+00 -1.06037021e+00 -1.63990229e-01 -7.52418637e-01
-6.29873574e-01 1.05619276e+00 1.10293257e+00 1.67541578e-01
2.64572293e-01 2.09226027e-01 4.87867922e-01 -4.25095767e-01
-1.32474494e+00 -2.59149134e-01 1.89664543e-01 4.57039297e-01
4.89990473e-01 -1.53745070e-01 -2.23656192e-01 8.46519709e-01
-4.93961275e-01 1.28654972e-01 5.41331530e-01 1.17531145e+00
-4.52181876e-01 -1.63848913e+00 5.03905443e-03 5.77236295e-01
-5.06388485e-01 -2.96374023e-01 -5.70455849e-01 6.54932439e-01
-7.82115161e-02 1.03948367e+00 -1.57131687e-01 -2.75566608e-01
1.27703059e+00 5.36300838e-01 3.21818501e-01 -1.03105748e+00
-1.24518657e+00 -7.44129717e-01 6.65323317e-01 -8.35174441e-01
1.87059298e-01 -5.01891434e-01 -1.55674767e+00 -4.10373211e-02
1.38068907e-02 4.90834087e-01 6.17663443e-01 8.00276160e-01
6.81554377e-01 4.44542974e-01 6.71039402e-01 1.07193136e+00
-8.58691990e-01 -5.10793924e-01 2.49998644e-01 7.04572797e-01
-2.48978641e-02 2.80718386e-01 8.60313699e-02 8.11105296e-02] | [9.958758354187012, 7.8564043045043945] |
313ca3b6-4e7d-4828-91f2-19c037e8d66a | why-deep-models-often-cannot-beat-non-deep | 2306.17702 | null | https://arxiv.org/abs/2306.17702v1 | https://arxiv.org/pdf/2306.17702v1.pdf | Why Deep Models Often cannot Beat Non-deep Counterparts on Molecular Property Prediction? | Molecular property prediction (MPP) is a crucial task in the drug discovery pipeline, which has recently gained considerable attention thanks to advances in deep neural networks. However, recent research has revealed that deep models struggle to beat traditional non-deep ones on MPP. In this study, we benchmark 12 representative models (3 non-deep models and 9 deep models) on 14 molecule datasets. Through the most comprehensive study to date, we make the following key observations: \textbf{(\romannumeral 1)} Deep models are generally unable to outperform non-deep ones; \textbf{(\romannumeral 2)} The failure of deep models on MPP cannot be solely attributed to the small size of molecular datasets. What matters is the irregular molecule data pattern; \textbf{(\romannumeral 3)} In particular, tree models using molecular fingerprints as inputs tend to perform better than other competitors. Furthermore, we conduct extensive empirical investigations into the unique patterns of molecule data and inductive biases of various models underlying these phenomena. | ['Stan Z. Li', 'Xiao Zhu', 'Lecheng Zhang', 'Jun Xia'] | 2023-06-30 | null | null | null | null | ['drug-discovery', 'property-prediction', 'molecular-property-prediction'] | ['medical', 'medical', 'miscellaneous'] | [ 2.58610517e-01 -1.86274841e-01 -5.38125515e-01 -2.42939517e-01
-2.99795508e-01 -5.08445084e-01 4.57178175e-01 4.99472827e-01
-1.56636268e-01 1.23284745e+00 -1.53902117e-02 -7.21176863e-01
-6.26383185e-01 -8.63011777e-01 -9.53323364e-01 -8.32785010e-01
-4.25389677e-01 3.88053209e-01 6.20391108e-02 -3.51672649e-01
5.15593171e-01 8.41327965e-01 -1.31437266e+00 3.23009968e-01
8.93101394e-01 9.46387351e-01 -1.52251467e-01 2.63015509e-01
3.28386948e-02 8.76152337e-01 -5.07792413e-01 -6.49526596e-01
6.74403757e-02 -2.46421218e-01 -8.31751287e-01 -6.40919089e-01
5.41264594e-01 -1.67214558e-01 -2.81169593e-01 1.15706396e+00
8.76586199e-01 1.36316642e-02 9.76947665e-01 -9.57588136e-01
-7.25283146e-01 5.85772753e-01 -5.69747448e-01 4.12317336e-01
3.93818200e-01 2.49265715e-01 1.22345889e+00 -9.38420773e-01
6.59683704e-01 1.13826656e+00 1.00034654e+00 4.31981355e-01
-1.43040788e+00 -8.62822115e-01 7.48700136e-03 2.22702324e-01
-1.48738515e+00 -3.03246409e-01 4.26228195e-01 -4.44163471e-01
1.13511562e+00 2.71395624e-01 1.83813855e-01 1.43222177e+00
5.95336437e-01 4.89854395e-01 1.12935436e+00 -4.14327644e-02
2.27150604e-01 9.41754729e-02 9.99137238e-02 6.38816595e-01
5.00458360e-01 3.98693830e-01 -6.20628953e-01 -4.65155959e-01
7.18510866e-01 1.19932573e-02 -3.17800641e-01 -2.09381610e-01
-7.96733141e-01 9.97086763e-01 6.49594426e-01 3.42601478e-01
-6.33715510e-01 -1.44004002e-02 2.41393209e-01 3.90031278e-01
2.82902181e-01 9.40727174e-01 -8.51662219e-01 -9.14670900e-02
-7.24550366e-01 4.20342028e-01 7.57010520e-01 7.29595900e-01
7.02517211e-01 -2.38756109e-02 -2.70406827e-02 8.36550951e-01
-1.28720015e-01 2.35751159e-02 3.20495576e-01 -5.10860980e-01
4.23829883e-01 4.95315492e-01 3.14322598e-02 -1.10016644e+00
-7.24340022e-01 -6.65295482e-01 -1.01710856e+00 2.99707265e-03
6.01790905e-01 -8.05944353e-02 -8.68591785e-01 1.50563467e+00
-6.02825498e-03 -3.04046571e-01 -7.45826438e-02 5.11244118e-01
1.22611415e+00 6.03850305e-01 5.95048964e-01 -4.43525054e-02
1.09710598e+00 -5.91774940e-01 -2.86797136e-01 2.21088320e-01
7.80164242e-01 -7.61462986e-01 8.44247282e-01 6.13245368e-01
-6.46422029e-01 -5.22000551e-01 -1.01100039e+00 6.38480932e-02
-8.54525208e-01 -2.95957029e-01 1.06813419e+00 7.39891827e-01
-6.03197396e-01 1.33808279e+00 -2.80443996e-01 -2.04073191e-01
8.22107017e-01 9.07797217e-01 -5.71938396e-01 -1.83486193e-01
-1.34364486e+00 1.00884604e+00 5.86733937e-01 -1.47193089e-01
-1.08574748e+00 -9.05019701e-01 -5.14797509e-01 -4.40806299e-02
2.05831289e-01 -5.28753221e-01 7.70534873e-01 -8.60465348e-01
-1.36451530e+00 5.98961115e-01 3.83290648e-02 -6.02591693e-01
3.70633274e-01 -3.33663411e-02 -4.92648631e-01 2.21096780e-02
-1.42456651e-01 7.29502439e-01 4.18733329e-01 -1.19756007e+00
-3.47102463e-01 -4.77038324e-01 1.54684052e-01 -1.95076808e-01
-3.85050356e-01 -5.07236086e-02 1.14496395e-01 -6.47136569e-01
-2.76432276e-01 -7.58302212e-01 -3.31999481e-01 -5.04874527e-01
-7.21089303e-01 -5.60098410e-01 2.47070253e-01 -1.65210888e-01
1.34469199e+00 -1.83582556e+00 1.73715025e-01 2.82173634e-01
3.24979335e-01 3.43440801e-01 -2.63943851e-01 7.18860149e-01
-6.05369985e-01 4.05749887e-01 -5.91487437e-02 2.44296759e-01
-1.32052884e-01 1.40589491e-01 -1.22830994e-01 4.32151198e-01
1.91069439e-01 8.58847439e-01 -8.74916673e-01 -3.44671488e-01
5.67001365e-02 4.20516610e-01 -4.25426781e-01 -2.71861017e-01
-5.01377940e-01 3.11155945e-01 -6.90698981e-01 1.11677849e+00
8.32694173e-01 -4.44161743e-01 1.15610017e-02 -2.01868832e-01
-2.82773376e-01 3.51299167e-01 -5.29763758e-01 1.39635992e+00
1.67274117e-01 4.18035269e-01 -5.06609797e-01 -9.55207765e-01
9.02796209e-01 2.85377651e-01 5.62978804e-01 -8.14056814e-01
1.79511771e-01 4.13225263e-01 5.95522881e-01 -2.80782640e-01
2.86522597e-01 -4.21189547e-01 4.35623109e-01 9.33982208e-02
1.15585022e-01 1.83180898e-01 9.76661295e-02 -2.92103291e-01
1.07052064e+00 4.68251668e-02 2.80085653e-01 -3.87703419e-01
5.10483384e-01 -3.60590406e-02 5.56854844e-01 8.92991304e-01
-1.22623451e-01 3.94895107e-01 8.61433268e-01 -9.30501223e-01
-9.08096254e-01 -6.35823309e-01 -7.30191112e-01 1.21026635e+00
-1.16649911e-01 -5.94602704e-01 -4.42572236e-01 -8.17641973e-01
1.98020250e-01 4.48922962e-01 -9.92377639e-01 -2.19677001e-01
-1.97588354e-01 -1.39340067e+00 8.11354816e-01 5.41714966e-01
2.63503939e-01 -1.19301033e+00 -1.66508764e-01 2.95179814e-01
3.10633481e-01 -7.63581097e-01 2.54038453e-01 7.23887622e-01
-9.35281694e-01 -1.10985076e+00 -7.45720088e-01 -4.44560021e-01
3.51683259e-01 -1.71656147e-01 1.22792137e+00 1.49389327e-01
-8.07455331e-02 -4.05581415e-01 -3.72867912e-01 -7.70309091e-01
-3.31703424e-01 4.56713259e-01 3.01045598e-03 -2.33391732e-01
8.61225605e-01 -5.93920708e-01 -8.87965620e-01 3.29185784e-01
-7.42628813e-01 -3.20669979e-01 1.00723469e+00 7.87313521e-01
5.73941886e-01 1.68328762e-01 8.27890515e-01 -1.11791360e+00
6.38977170e-01 -7.21480846e-01 -1.95486724e-01 -3.30716297e-02
-6.68357432e-01 4.57382426e-02 8.26245248e-01 -4.50436145e-01
-4.43447769e-01 -4.37363423e-02 -6.42293930e-01 -1.46021754e-01
-3.02830786e-01 7.72176325e-01 -7.64560103e-02 -2.65308857e-01
9.18572366e-01 9.53658521e-02 -3.87297928e-01 -8.13308775e-01
-1.07166298e-01 2.91394472e-01 7.92547595e-03 -7.42474973e-01
3.87834877e-01 2.39553884e-01 4.63715315e-01 -8.52696240e-01
-8.36412728e-01 7.56394044e-02 -5.72750390e-01 2.77608633e-01
6.92586124e-01 -6.90439880e-01 -1.03700721e+00 2.30637938e-01
-1.01827586e+00 -2.93668330e-01 4.40209918e-02 3.31204146e-01
-1.84813455e-01 5.56037605e-01 -6.50204062e-01 -4.83022839e-01
-3.56307685e-01 -1.16486526e+00 7.35595465e-01 8.18640292e-02
-5.21393776e-01 -1.07430899e+00 -1.37953311e-02 2.07821801e-01
2.53144950e-01 6.27529442e-01 1.52727795e+00 -9.94556963e-01
-2.85361916e-01 -2.58884341e-01 -3.49595368e-01 1.42631680e-01
2.12833926e-01 1.70185179e-01 -1.20386112e+00 -1.66388676e-01
-3.73874724e-01 -4.27610755e-01 9.92367387e-01 5.43952405e-01
1.49797297e+00 -9.62179676e-02 -3.73421192e-01 5.76861501e-01
1.32272696e+00 5.03688931e-01 7.67030358e-01 5.32549918e-01
7.45630622e-01 6.75059259e-01 2.35790953e-01 4.31670129e-01
7.36488961e-03 4.85118806e-01 7.31872797e-01 -2.92197764e-01
2.42148712e-01 -3.12757105e-01 1.83743075e-01 8.20269436e-02
-4.13629025e-01 -2.87054271e-01 -9.33767736e-01 2.40694880e-01
-1.49702656e+00 -8.71193230e-01 -4.26378936e-01 2.24669957e+00
1.06603765e+00 2.98647821e-01 1.70160234e-01 3.86124589e-02
2.90331215e-01 1.09449379e-01 -8.37480664e-01 -4.53190416e-01
-4.68016446e-01 6.70007169e-01 4.74772006e-01 1.05311058e-01
-1.18221712e+00 9.86875892e-01 6.65980625e+00 1.16359055e+00
-1.30879390e+00 -2.34115750e-01 8.89876246e-01 1.94685534e-01
-3.10054332e-01 -1.76397040e-01 -8.70428503e-01 5.29969573e-01
1.12776351e+00 1.08700320e-01 -1.24601983e-01 6.50028706e-01
-6.03501946e-02 1.32086463e-02 -1.71332264e+00 8.19370031e-01
-2.50909328e-01 -1.66360402e+00 6.45314395e-01 5.35880148e-01
7.29883313e-01 1.65393770e-01 3.39260131e-01 2.43748546e-01
2.48944923e-01 -1.80643547e+00 2.77186155e-01 1.78071290e-01
6.82312012e-01 -8.45995426e-01 7.68274248e-01 1.54662982e-01
-7.51114011e-01 -1.31312728e-01 -6.93785250e-01 -2.77289659e-01
-5.03833592e-01 6.10648870e-01 -6.19364679e-01 9.44057703e-01
7.18837082e-01 9.53775167e-01 -5.84554732e-01 9.67934430e-01
1.06450971e-02 4.43500727e-01 -5.74539974e-02 -1.29833385e-01
4.24666613e-01 -8.32153112e-02 2.39685386e-01 1.31170678e+00
4.12655398e-02 -1.42061889e-01 -7.96643719e-02 8.66558909e-01
-4.01116997e-01 2.69311011e-01 -8.21279049e-01 -2.07723200e-01
3.16468358e-01 8.66694689e-01 -5.81751406e-01 -1.75101489e-01
-3.91493887e-01 4.09194797e-01 1.44097716e-01 2.35929236e-01
-6.55439794e-01 -2.15405956e-01 8.25741053e-01 1.13769649e-02
3.07839900e-01 2.02759534e-01 -4.68684584e-01 -7.54124165e-01
-3.95037234e-01 -1.20951331e+00 4.99358833e-01 -3.90877694e-01
-1.60560215e+00 5.72056830e-01 -2.89996415e-01 -9.56036270e-01
4.14685518e-01 -1.05287182e+00 -3.06764483e-01 9.35590148e-01
-1.56408215e+00 -6.95833206e-01 5.52406944e-02 5.73495150e-01
2.90995419e-01 -2.26219356e-01 1.23157096e+00 5.08799613e-01
-8.59840333e-01 6.18357360e-01 2.75570333e-01 4.48766164e-02
6.41954601e-01 -1.15484643e+00 3.49088430e-01 1.84453800e-01
4.08422351e-02 9.14627135e-01 7.63381898e-01 -5.60664237e-01
-1.20354390e+00 -9.68016565e-01 8.51409674e-01 -7.26644874e-01
6.38724744e-01 -4.83537130e-02 -1.04622912e+00 3.92346531e-01
2.08575726e-01 -2.44914204e-01 1.31753159e+00 3.44356447e-01
-4.74602938e-01 4.88550067e-02 -1.00527191e+00 4.94180948e-01
8.97054315e-01 -4.03096437e-01 -3.75942379e-01 5.14022946e-01
4.19578910e-01 -2.41106287e-01 -1.30318773e+00 8.70380878e-01
7.42295682e-01 -1.12646616e+00 1.15088904e+00 -1.04286981e+00
8.23057532e-01 -9.69606042e-02 -6.13655746e-02 -9.31207955e-01
-4.58142638e-01 -4.62826818e-01 -1.52524691e-02 7.36807525e-01
8.08290601e-01 -7.28896618e-01 9.64781404e-01 3.50567460e-01
5.10010980e-02 -1.24498963e+00 -7.81659603e-01 -8.06636810e-01
6.49698436e-01 -1.88977599e-01 6.62187338e-01 1.18428266e+00
-2.46582597e-01 4.72953558e-01 -3.46254885e-01 -2.91303545e-02
3.04953039e-01 1.84154540e-01 5.50256789e-01 -1.58458328e+00
-2.49221250e-01 -7.51532018e-01 -3.30292791e-01 -1.03406072e+00
6.85134083e-02 -7.58942306e-01 -5.34182191e-01 -1.32514167e+00
4.94479269e-01 -5.46645045e-01 -8.20914567e-01 6.43014908e-01
-4.35625315e-02 2.31815264e-01 -2.01257989e-01 2.43067205e-01
-1.97049618e-01 4.60741639e-01 1.28738213e+00 -3.15433264e-01
-6.65154159e-02 7.99725875e-02 -9.40690935e-01 7.05554664e-01
7.30772376e-01 -5.29033184e-01 -2.30345353e-01 -1.59983799e-01
6.25947535e-01 -2.24623889e-01 2.79339284e-01 -6.74638987e-01
-5.42698167e-02 -1.06435075e-01 8.70344341e-01 -6.88765883e-01
2.91304111e-01 -6.31398797e-01 1.96163088e-01 5.74125767e-01
-1.37898847e-01 -8.95572752e-02 3.91894907e-01 4.84398991e-01
-1.87920675e-01 -2.57523805e-01 7.26469100e-01 -3.06459308e-01
-3.48719001e-01 6.49580121e-01 -3.90281707e-01 -2.91979283e-01
7.15592623e-01 -4.78258908e-01 -3.69314134e-01 2.82986872e-02
-6.73016071e-01 8.93416405e-02 2.84520566e-01 3.23584557e-01
3.86860847e-01 -1.08890963e+00 -4.73263830e-01 -1.45589918e-01
1.57041952e-01 3.07296365e-02 1.24094062e-01 9.48192477e-01
-7.11819232e-01 7.24242330e-01 -3.38945419e-01 -4.13075536e-01
-1.13514221e+00 6.72658980e-01 5.81335008e-01 -1.96946591e-01
-2.74455756e-01 1.04878128e+00 4.15393531e-01 -2.38176405e-01
4.21380490e-01 -2.47194633e-01 -2.80574471e-01 2.66555279e-01
3.50460142e-01 4.44909781e-01 2.51060635e-01 -3.36361319e-01
-5.67372918e-01 4.81361806e-01 -5.56035340e-01 7.67106473e-01
1.55395770e+00 6.78847551e-01 -1.29292831e-01 2.01601088e-01
1.14817536e+00 -2.56474942e-01 -9.50964510e-01 3.68364602e-02
2.49903411e-01 -2.75061756e-01 -2.62597531e-01 -1.06409991e+00
-7.09492981e-01 1.10700643e+00 4.72250074e-01 5.28042167e-02
7.52852023e-01 -2.94043988e-01 5.84097087e-01 5.96117258e-01
2.72305757e-01 -9.44258988e-01 -2.81902775e-03 4.86461610e-01
8.43719542e-01 -1.24690270e+00 4.60746378e-01 -3.15413564e-01
-3.33069265e-01 1.21820092e+00 4.30060118e-01 1.98425367e-01
5.55832565e-01 -3.05533171e-01 -3.45977128e-01 -5.80544293e-01
-6.49934888e-01 -1.03141323e-01 1.01364508e-01 4.36747938e-01
8.77905488e-01 5.19989617e-02 -3.00203711e-01 4.44305450e-01
-1.30996421e-01 4.22164761e-02 9.99068320e-02 8.07087481e-01
-3.00378799e-01 -1.41869700e+00 -4.38920930e-02 4.46675271e-01
-8.94509196e-01 -4.96693075e-01 -9.72160220e-01 9.15566802e-01
3.98084998e-01 8.36139977e-01 -3.90752256e-01 -4.48474556e-01
3.68266642e-01 -1.15695603e-01 6.17741883e-01 -5.02835453e-01
-9.58056688e-01 -1.04602411e-01 1.95586011e-01 -3.65416110e-01
-4.54900056e-01 -4.17726547e-01 -9.94955719e-01 -6.89901531e-01
-1.89137533e-01 3.45658064e-01 2.54033476e-01 7.54448116e-01
4.40177947e-01 4.08038348e-01 4.63746041e-01 -5.62329292e-01
-4.91166800e-01 -1.04563332e+00 -5.48425138e-01 2.89457202e-01
2.11332530e-01 -8.20370376e-01 -1.84156075e-01 -3.07769209e-01] | [5.2082905769348145, 5.756438732147217] |
0cec79d0-ac1d-49c2-89e0-f0ea1a2a5687 | image-augmentation-based-momentum-memory | 2205.09448 | null | https://arxiv.org/abs/2205.09448v1 | https://arxiv.org/pdf/2205.09448v1.pdf | Image Augmentation Based Momentum Memory Intrinsic Reward for Sparse Reward Visual Scenes | Many scenes in real life can be abstracted to the sparse reward visual scenes, where it is difficult for an agent to tackle the task under the condition of only accepting images and sparse rewards. We propose to decompose this problem into two sub-problems: the visual representation and the sparse reward. To address them, a novel framework IAMMIR combining the self-supervised representation learning with the intrinsic motivation is presented. For visual representation, a representation driven by a combination of the imageaugmented forward dynamics and the reward is acquired. For sparse rewards, a new type of intrinsic reward is designed, the Momentum Memory Intrinsic Reward (MMIR). It utilizes the difference of the outputs from the current model (online network) and the historical model (target network) to present the agent's state familiarity. Our method is evaluated on the visual navigation task with sparse rewards in Vizdoom. Experiments demonstrate that our method achieves the state of the art performance in sample efficiency, at least 2 times faster than the existing methods reaching 100% success rate. | ['Guizhong Liu', 'Biao Zhao', 'Zheng Fang'] | 2022-05-19 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 4.25871909e-02 1.76722750e-01 -4.31795597e-01 -1.47836491e-01
-2.51650810e-01 -1.21291585e-01 8.44015658e-01 -2.57389843e-01
-6.31119788e-01 7.95369804e-01 -5.61696012e-04 1.76952288e-01
-1.40748974e-02 -3.96132320e-01 -7.08228469e-01 -9.09579933e-01
-1.55137092e-01 3.02055568e-01 6.73022345e-02 -3.41114908e-01
4.79659021e-01 1.24014437e-01 -1.74458504e+00 -2.83993711e-03
9.23677742e-01 1.16421521e+00 8.14455092e-01 3.75194073e-01
-1.34966493e-01 1.52361155e+00 -1.25894070e-01 2.19081044e-01
3.58327955e-01 -5.32686293e-01 -5.69559574e-01 8.32102299e-02
6.15070648e-02 -4.68405515e-01 -4.21273500e-01 1.16999030e+00
2.66397357e-01 5.80030262e-01 5.34898520e-01 -1.21257222e+00
-6.39696121e-01 5.98355770e-01 -6.49672806e-01 2.06264809e-01
1.27618119e-01 3.69698584e-01 8.31216693e-01 -9.49457526e-01
9.54778016e-01 1.34148252e+00 -1.52849853e-01 8.34824502e-01
-1.15216148e+00 -3.83774251e-01 9.79859710e-01 3.53251398e-01
-7.28963494e-01 -3.60851973e-01 6.50458515e-01 -5.60041904e-01
9.50049102e-01 -1.10953033e-01 8.68716419e-01 1.26807809e+00
1.40506476e-01 1.08201659e+00 1.38234282e+00 -3.45092118e-02
6.83702767e-01 2.80596733e-01 -7.57038686e-03 8.11308503e-01
-2.10617427e-02 1.01444376e+00 -8.28923404e-01 -1.04711294e-01
1.08544862e+00 3.46726000e-01 -1.29799828e-01 -8.80640984e-01
-1.33268809e+00 1.00306082e+00 1.04020333e+00 1.10413849e-01
-7.10041285e-01 4.52414155e-01 1.37550056e-01 3.84608567e-01
1.08488537e-02 3.83362621e-01 -4.52044755e-02 -1.09366991e-01
-4.94050533e-01 2.04266593e-01 2.83197612e-01 7.95561492e-01
7.30657041e-01 8.03164363e-01 -2.82092720e-01 6.56610966e-01
5.56932330e-01 6.06786728e-01 8.15713346e-01 -1.09041035e+00
1.85790494e-01 3.90628427e-01 3.64624977e-01 -9.20581520e-01
-1.47025168e-01 -6.11767113e-01 -9.59427893e-01 9.38892126e-01
6.37963936e-02 1.03724161e-02 -1.09700060e+00 2.00615144e+00
1.74859092e-01 3.12802494e-01 4.24472034e-01 1.28355920e+00
8.84046912e-01 8.56696725e-01 6.59474954e-02 -3.75992298e-01
5.94260395e-01 -1.38998616e+00 -6.04475319e-01 -6.38062239e-01
3.54929566e-01 -9.10485014e-02 8.39132249e-01 2.54069090e-01
-1.20753825e+00 -6.15048230e-01 -9.23159838e-01 4.22265917e-01
-2.47467086e-02 1.52745679e-01 8.21007848e-01 -6.60122335e-02
-1.08327198e+00 7.66442299e-01 -7.15049386e-01 -2.45998189e-01
4.05098021e-01 3.19912106e-01 -2.72465020e-01 1.76780391e-02
-7.04781473e-01 8.89784575e-01 2.21478283e-01 2.00869635e-01
-1.69819641e+00 -7.25438595e-02 -8.78837824e-01 -2.01881099e-02
5.26543260e-01 -7.90392935e-01 1.01543474e+00 -1.49845219e+00
-1.68297517e+00 6.20180130e-01 -4.25011441e-02 -5.86069107e-01
5.55999696e-01 -2.16519386e-01 3.09022546e-01 -8.63053948e-02
1.94747940e-01 9.13188815e-01 1.36479199e+00 -1.39168227e+00
-5.82650483e-01 -2.03157231e-01 1.00681029e-01 6.50042117e-01
-1.59500539e-01 -4.47296500e-01 -2.27183655e-01 -5.24389386e-01
2.29383055e-02 -9.68318701e-01 -8.90287042e-01 8.96729454e-02
-8.46596062e-02 4.48536174e-03 3.53770882e-01 -2.16649011e-01
5.51573634e-01 -2.23740506e+00 7.96297193e-01 1.89949170e-01
2.76655585e-01 5.98433092e-02 -5.06613016e-01 5.91613501e-02
-3.61613594e-02 -4.56535876e-01 -9.65889692e-02 -3.80045086e-01
-1.94190472e-01 4.09038991e-01 -5.90628207e-01 3.74000341e-01
1.01797782e-01 9.48066294e-01 -1.31539428e+00 -2.98576087e-01
9.10083354e-02 1.63347647e-01 -5.23835838e-01 5.85519135e-01
-2.19846532e-01 8.02589059e-01 -7.12524533e-01 4.41768855e-01
3.77338737e-01 -4.49297667e-01 1.65892884e-01 3.32467288e-01
-1.35288849e-01 -1.48083672e-01 -1.21826100e+00 1.93250418e+00
-1.24230266e-01 4.63260114e-02 1.83889836e-01 -8.97460878e-01
1.03903019e+00 -9.32066068e-02 4.08452362e-01 -1.01507652e+00
1.84295326e-01 2.16327101e-01 -5.89359999e-02 -2.56461412e-01
6.76477730e-01 1.85871392e-01 2.09445432e-01 3.46100360e-01
2.11371467e-01 8.45096782e-02 4.80951443e-02 5.92876494e-01
9.17211354e-01 5.80066919e-01 1.26388818e-01 -2.18293276e-02
3.42239588e-01 1.80982783e-01 6.14211798e-01 1.02799249e+00
-2.91972399e-01 1.24735095e-01 3.01136672e-01 -4.07680035e-01
-8.10666502e-01 -1.05281091e+00 5.34795105e-01 1.05174124e+00
4.67608303e-01 -7.83792958e-02 -1.80831388e-01 -6.07724488e-01
3.29597145e-02 3.64573598e-01 -8.62402439e-01 -3.33177209e-01
-3.95659536e-01 -4.39712733e-01 -3.51164401e-01 5.74407101e-01
4.12374765e-01 -1.59050012e+00 -1.03727233e+00 1.90135613e-01
2.96408199e-02 -6.53280854e-01 -3.28037649e-01 4.68806952e-01
-1.08594310e+00 -5.86156905e-01 -9.84397233e-01 -7.78997660e-01
8.06844652e-01 4.78069156e-01 9.48987782e-01 6.99494332e-02
-1.59986243e-01 6.03596807e-01 -2.87485391e-01 8.24861974e-02
-3.69909406e-02 -3.66485238e-01 1.43503428e-01 1.09478883e-01
-4.28724051e-01 -4.62969869e-01 -4.64924544e-01 6.95152953e-03
-5.80133557e-01 2.04916239e-01 9.39464748e-01 1.30160630e+00
8.82909954e-01 -5.91112316e-01 6.99867308e-01 -6.01080835e-01
5.21642268e-01 -6.13025606e-01 -8.20769846e-01 -2.25869969e-01
-7.84012437e-01 3.79291207e-01 4.10268754e-01 -7.28593528e-01
-1.15890563e+00 4.60166782e-01 3.96215230e-01 -7.61665285e-01
2.13130757e-01 6.25193000e-01 4.96650219e-01 -2.12434277e-01
6.18880212e-01 5.12757659e-01 2.87951887e-01 -2.16789931e-01
7.22241640e-01 -3.60209681e-02 4.72803056e-01 -3.96291852e-01
7.50004530e-01 4.40785170e-01 7.11891204e-02 -3.84925038e-01
-6.75773025e-01 -3.92165482e-01 -7.07176626e-02 -6.22299135e-01
6.09955490e-01 -1.07609534e+00 -8.09738457e-01 3.85094017e-01
-9.45765316e-01 -4.84898657e-01 -6.98763192e-01 7.09921539e-01
-9.74253654e-01 2.67547250e-01 -5.25758743e-01 -1.16754699e+00
-6.10397123e-02 -1.16111636e+00 5.28294325e-01 4.88858908e-01
4.21529502e-01 -5.48202932e-01 2.51613021e-01 -1.42504960e-01
6.15704954e-01 8.91650245e-02 5.81093371e-01 -8.56845304e-02
-1.11219263e+00 3.00817132e-01 -2.07617834e-01 1.28132239e-01
-2.75961429e-01 -5.43186426e-01 -8.59914720e-01 -5.67172587e-01
1.62197307e-01 -1.00312150e+00 1.36126137e+00 5.76781452e-01
7.18320966e-01 -2.98349857e-01 -3.05971771e-01 4.46617901e-01
1.47200871e+00 2.31983766e-01 5.32989323e-01 4.59836245e-01
5.36029398e-01 5.61508060e-01 9.97765958e-01 6.49529457e-01
3.32989961e-01 5.23883641e-01 1.24547338e+00 4.03681472e-02
1.05347455e-01 -2.73060232e-01 9.02260840e-01 7.14211524e-01
-4.73512352e-01 2.13316843e-01 -2.66138524e-01 3.81754667e-01
-2.53945303e+00 -1.23748171e+00 3.10253978e-01 2.37990427e+00
4.03709561e-01 9.86727923e-02 1.67723447e-01 -4.74752367e-01
4.98784602e-01 3.57716948e-01 -1.20184803e+00 -3.05893332e-01
-1.72021501e-02 -3.85081917e-01 2.01862484e-01 4.87096727e-01
-7.95614064e-01 1.09631693e+00 6.36603069e+00 5.25057852e-01
-1.07071531e+00 -8.96219816e-03 6.47811472e-01 -2.28530526e-01
-1.98628977e-01 3.29012536e-02 -6.08834624e-01 1.61725983e-01
5.46712399e-01 -1.09844916e-01 9.87867355e-01 1.37505615e+00
-4.52421196e-02 -3.19726199e-01 -9.22482014e-01 1.20636749e+00
1.42404571e-01 -1.36863434e+00 8.00013915e-02 1.26212478e-01
7.58453727e-01 3.43992978e-01 6.65727139e-01 5.93810678e-01
7.36980677e-01 -1.01623952e+00 9.43810880e-01 8.51735413e-01
4.65766221e-01 -3.50152820e-01 4.08387661e-01 6.10617399e-01
-1.23100352e+00 -5.44824958e-01 -5.60957372e-01 -2.20610589e-01
2.03724820e-02 3.78067791e-02 -3.00473481e-01 5.12305439e-01
6.56644046e-01 1.34470046e+00 -2.99152762e-01 1.01830673e+00
-1.96910441e-01 -1.39905170e-01 7.19220266e-02 -2.08069935e-01
6.19088769e-01 -4.71765965e-01 7.14280725e-01 4.39099342e-01
2.72896409e-01 -2.21837312e-01 6.56688273e-01 1.01791203e+00
2.89462566e-01 -3.13064866e-02 -8.53603125e-01 -5.55254482e-02
2.27316357e-02 1.35616815e+00 -4.97149199e-01 -4.02875215e-01
-6.08430542e-02 1.05137694e+00 9.06328380e-01 5.20352602e-01
-5.99995732e-01 1.88833892e-01 3.31968337e-01 -4.83525515e-01
3.57663691e-01 -1.93354130e-01 1.22447148e-01 -1.38963449e+00
-1.25322759e-01 -6.92570508e-01 2.57164985e-01 -8.61261845e-01
-1.30307007e+00 9.32579279e-01 -3.30841422e-01 -1.47286665e+00
-5.77551603e-01 -4.44076091e-01 -4.95889604e-01 8.07389975e-01
-1.97356737e+00 -8.22139323e-01 -4.40189481e-01 9.15286720e-01
5.57913721e-01 -5.30110240e-01 8.01742315e-01 -1.54868588e-01
-4.52487171e-01 1.03556708e-01 1.98967859e-01 -4.21807557e-01
3.97969186e-01 -1.24660122e+00 -1.11448303e-01 6.09555185e-01
1.25934109e-01 1.69062912e-01 4.98765886e-01 -6.61998689e-01
-1.62041306e+00 -9.09892738e-01 3.61353993e-01 -7.72721395e-02
4.69963253e-01 3.73794860e-03 -6.28452122e-01 5.94971478e-01
1.47024557e-01 2.58797735e-01 8.08913186e-02 -2.08547905e-01
-3.61903161e-01 4.07785513e-02 -9.96530831e-01 6.53739393e-01
1.14443660e+00 -4.71364334e-02 -5.91168642e-01 1.83319524e-01
5.47743678e-01 -4.12713706e-01 -1.51076794e-01 -1.56405102e-02
3.00933957e-01 -9.14417326e-01 1.00469589e+00 -7.01743007e-01
5.16486526e-01 -4.68320638e-01 -2.01018348e-01 -1.54283845e+00
-7.12591052e-01 -5.31912327e-01 -3.69856507e-01 6.07085586e-01
2.10656956e-01 -5.99544406e-01 8.40602040e-01 2.32105494e-01
-5.70833087e-02 -6.56068802e-01 -9.39090610e-01 -9.68172729e-01
-3.46944720e-01 8.62762630e-02 1.45700648e-02 6.62142575e-01
9.61727351e-02 4.84215945e-01 -8.76661658e-01 -1.38838530e-01
8.35994303e-01 4.69737887e-01 6.67117953e-01 -1.17934453e+00
-4.82607156e-01 -4.22878653e-01 -1.91893622e-01 -1.33071291e+00
2.03265548e-01 -1.10313880e+00 3.70489001e-01 -1.62024188e+00
5.64193189e-01 -5.44571161e-01 -7.18314409e-01 4.11043376e-01
-3.59221026e-02 -2.98104972e-01 6.98328495e-01 5.96072495e-01
-9.72927868e-01 1.20811379e+00 1.60607243e+00 -3.03730398e-01
-4.26670462e-01 -1.96725518e-01 -5.78630269e-01 6.52230322e-01
6.05600655e-01 -4.90167826e-01 -5.68192840e-01 -3.78261298e-01
1.43685460e-01 3.87097090e-01 3.59541982e-01 -1.06951332e+00
1.76693380e-01 -5.76463044e-01 3.23363155e-01 -5.37132621e-01
7.87002504e-01 -7.49368548e-01 -1.54716238e-01 9.56861496e-01
-6.25483871e-01 1.69950753e-01 -1.03898883e-01 1.21540260e+00
-1.33951694e-01 -2.63087720e-01 6.98784232e-01 -5.25534987e-01
-1.02233005e+00 5.17837644e-01 -5.17575681e-01 -9.61523354e-02
9.83427227e-01 -4.76281065e-03 -5.08860767e-01 -4.58654761e-01
-9.98872757e-01 5.31468689e-01 3.25348556e-01 3.81940067e-01
1.19493043e+00 -1.54285681e+00 -4.99573529e-01 1.74733117e-01
6.53522238e-02 -4.17078972e-01 4.17335302e-01 7.72374034e-01
2.65025258e-01 1.73019856e-01 -7.65271485e-01 -6.67212188e-01
-8.47851038e-01 9.06778097e-01 2.76001453e-01 -4.85815436e-01
-7.20853269e-01 7.52719581e-01 4.90374684e-01 -2.43434012e-01
5.66842735e-01 -1.81132779e-02 -7.42178082e-01 -1.58883154e-01
4.68818218e-01 1.43013909e-01 -7.55577862e-01 -6.36005938e-01
-2.02137321e-01 4.40509319e-01 -2.50851840e-01 -2.95145512e-01
1.47226059e+00 -1.48335874e-01 1.21477008e-01 6.42874718e-01
5.14941275e-01 -5.71315825e-01 -1.91974401e+00 -3.78857464e-01
-3.31160843e-01 -6.10880375e-01 1.75239593e-01 -8.01354110e-01
-1.29516327e+00 7.26535499e-01 9.83820736e-01 -9.32170302e-02
7.85262465e-01 -6.32647425e-02 3.19458127e-01 7.77149916e-01
7.01062977e-01 -1.08242548e+00 8.34680736e-01 8.23476732e-01
1.18830180e+00 -1.51876187e+00 -3.32575470e-01 1.14430174e-01
-1.22120130e+00 7.71597147e-01 9.04408038e-01 -6.35642469e-01
3.84779364e-01 -3.08753252e-01 -6.40761703e-02 -1.51702538e-01
-1.07864738e+00 -4.67220664e-01 -3.23782451e-02 7.90498853e-01
-2.04300433e-01 -8.03444088e-02 -7.41496077e-03 4.81453925e-01
4.09123302e-01 1.41218388e-02 5.64893186e-01 1.19857860e+00
-7.24413693e-01 -9.18213785e-01 -1.03191501e-02 4.63309944e-01
1.90621346e-01 2.51723472e-02 -1.54769868e-01 1.59925371e-01
-3.95163298e-01 7.81350553e-01 -2.26834610e-01 -3.00904870e-01
1.43890202e-01 -4.85369086e-01 4.77289408e-01 -6.52653456e-01
-6.29186094e-01 3.67257744e-02 -2.68421650e-01 -1.14092183e+00
-6.93688035e-01 -4.27749991e-01 -1.19503748e+00 7.58529678e-02
-1.40248373e-01 1.00239217e-01 5.32484710e-01 6.32071376e-01
4.16256547e-01 5.98830581e-01 7.76316762e-01 -1.31898904e+00
-7.88157463e-01 -5.38020313e-01 -6.72953427e-01 2.41964787e-01
3.61982793e-01 -1.09648585e+00 -3.41494620e-01 -9.95181277e-02] | [4.246084690093994, 1.3570404052734375] |
46e56878-3491-4818-8470-8f43f45de1e6 | a-novel-metric-for-evaluating-semantics-1 | null | null | https://openreview.net/forum?id=mVJ-hJVpq3r | https://openreview.net/pdf?id=mVJ-hJVpq3r | A Novel Metric for Evaluating Semantics Preservation | In this paper, we leverage pre-trained language models (PLMs) to precisely evaluate the semantics preservation of edition process on sentences. Our metric, Neighboring Distribution Divergence (NDD), evaluates the disturbance on predicted distribution of neighboring words from mask language model (MLM). NDD is capable of detecting precise changes in semantics which are easily ignored by text similarity. By exploiting the property of NDD, we implement a unsupervised and even training-free algorithm for extractive sentence compression. We show that NDD-based algorithm outperforms previous perplexity-based unsupervised algorithm by a large margin. For further exploration on interpretability, we evaluate NDD by pruning on syntactic dependency treebanks and apply NDD for predicate detection as well. | ['Anonymous'] | 2021-10-16 | null | null | null | acl-arr-october-2021-10 | ['predicate-detection', 'sentence-compression'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.65151531e-01 3.13303769e-01 -3.61041248e-01 -5.24957180e-01
-7.38300860e-01 -7.00539768e-01 7.20538497e-01 8.00054610e-01
-5.73963583e-01 7.58366644e-01 8.76081049e-01 -5.05824327e-01
4.18941677e-02 -5.47818542e-01 -6.45784855e-01 -3.23530734e-01
-1.91090822e-01 4.78839606e-01 4.62604403e-01 -8.06898773e-02
5.08449674e-01 4.03557271e-01 -1.16888034e+00 6.06417120e-01
6.81466937e-01 3.14687729e-01 4.71456617e-01 6.85910821e-01
-5.15067339e-01 8.19252133e-01 -7.48116314e-01 -3.84178966e-01
-5.24759181e-02 -5.52760899e-01 -9.55673695e-01 -2.62178779e-01
6.74356520e-02 -4.41680327e-02 -4.09212768e-01 1.32222259e+00
4.43989038e-01 2.35118836e-01 8.20746899e-01 -6.47222877e-01
-3.34839463e-01 1.25030315e+00 -3.37869585e-01 9.00140226e-01
4.85504657e-01 -1.19086586e-01 1.37643862e+00 -1.01485705e+00
9.81196225e-01 1.62382674e+00 4.17197376e-01 3.66369605e-01
-1.25587034e+00 -2.52683051e-02 1.61090821e-01 -6.25203848e-02
-1.17816889e+00 -4.83555883e-01 6.31608665e-01 -9.58474576e-02
1.49734366e+00 4.08569634e-01 1.09228775e-01 1.04999471e+00
5.27968466e-01 9.76648569e-01 8.87234926e-01 -8.16915572e-01
4.85625803e-01 -5.31480871e-02 4.55427170e-01 6.00125313e-01
4.19181466e-01 -1.18796036e-01 -8.13504338e-01 -2.19234094e-01
7.90758952e-02 -3.99009794e-01 -1.34848416e-01 -5.37236966e-03
-1.04572332e+00 6.86080933e-01 -3.47635329e-01 4.56361681e-01
-5.11109978e-02 4.32251729e-02 7.21615314e-01 6.67162776e-01
7.84012675e-01 3.74617845e-01 -8.69633734e-01 -3.63912791e-01
-1.10152113e+00 1.07704923e-01 9.64342356e-01 1.03108323e+00
5.12508571e-01 -2.66321063e-01 -4.68074471e-01 7.70656466e-01
2.96475440e-01 3.98797601e-01 1.04490209e+00 -5.95702529e-01
5.68895519e-01 5.25272369e-01 -3.44542354e-01 -7.09913731e-01
-2.81335175e-01 -2.89174825e-01 -5.75019300e-01 -3.58533025e-01
2.02132702e-01 1.62763208e-01 -7.79308200e-01 1.79054189e+00
1.41735300e-01 -5.79321906e-02 3.18169296e-01 2.92713195e-01
3.94398987e-01 5.36628544e-01 1.52029052e-01 -6.26736939e-01
1.05220795e+00 -6.07742071e-01 -7.27130413e-01 -3.02426785e-01
1.25379729e+00 -7.63452709e-01 1.37588501e+00 1.03230953e-01
-1.13455164e+00 -9.89785418e-02 -9.47896183e-01 5.38394961e-04
-1.46443114e-01 -1.70450255e-01 2.51097888e-01 4.55602646e-01
-9.09685135e-01 1.01134610e+00 -1.00986421e+00 -5.70408106e-01
2.09477514e-01 1.50004968e-01 -1.12325363e-01 3.53164911e-01
-1.30257022e+00 9.70391870e-01 6.03724003e-01 -4.63285446e-01
-6.35629475e-01 -6.23395085e-01 -9.44525301e-01 3.39278169e-02
2.08271325e-01 -3.03095043e-01 1.37616265e+00 -7.27070868e-01
-1.25492620e+00 1.09756243e+00 -7.36698687e-01 -9.57390189e-01
2.09750339e-01 -1.90661967e-01 -4.39901501e-01 1.70774713e-01
1.04210459e-01 1.73911259e-01 6.82487965e-01 -8.60386014e-01
-5.31204402e-01 -2.61874110e-01 -4.18027878e-01 2.54922211e-01
-3.97945821e-01 3.36130857e-01 -2.37939969e-01 -9.11950529e-01
2.34102666e-01 -4.90413040e-01 -8.73799250e-02 -5.61586618e-01
-7.39676654e-01 -2.90627629e-01 6.60971940e-01 -8.19276214e-01
1.85993278e+00 -2.28208780e+00 -5.90990260e-02 3.27162892e-01
5.37341461e-02 3.04327399e-01 -2.08680987e-01 5.80907226e-01
1.81247890e-01 2.64833808e-01 -5.80010831e-01 -6.45290613e-01
4.54222485e-02 5.21688282e-01 -7.11986005e-01 3.30835015e-01
1.33803859e-01 8.75342369e-01 -1.02864218e+00 -5.45131207e-01
-1.08379096e-01 -2.24023864e-01 -4.68690872e-01 -4.27383184e-02
-5.54688454e-01 1.34205610e-01 -1.59413099e-01 5.57083070e-01
4.99931991e-01 1.04532428e-01 6.73413277e-01 1.92691684e-01
7.28288619e-03 1.06832385e+00 -7.21913338e-01 1.68372810e+00
-3.39511275e-01 6.72729969e-01 -5.36539912e-01 -9.56981719e-01
9.36618209e-01 1.19811743e-02 -1.94948286e-01 -6.68519914e-01
-6.64936900e-02 2.32402802e-01 1.19518302e-01 -3.26610476e-01
6.17387295e-01 -1.90562308e-01 -2.67965049e-01 6.95372522e-01
1.36380091e-01 1.57681018e-01 2.85986245e-01 4.87294883e-01
1.48353612e+00 -2.43572161e-01 5.81891418e-01 -7.46590376e-01
5.06876111e-01 -2.28597447e-01 5.05366445e-01 9.57087040e-01
-2.05240771e-01 2.18525186e-01 8.14545214e-01 3.44333462e-02
-1.17852080e+00 -1.40200043e+00 -1.72427848e-01 1.14176869e+00
1.22321360e-01 -7.74264216e-01 -8.08821499e-01 -9.63208258e-01
9.58323553e-02 1.31133938e+00 -2.17902839e-01 -3.32877010e-01
-7.08011210e-01 -7.94459581e-01 8.28070104e-01 5.26324093e-01
9.40350741e-02 -1.00591409e+00 -2.80748963e-01 1.71728835e-01
-4.66483273e-02 -9.90399718e-01 -5.23035765e-01 4.67439145e-01
-1.14978313e+00 -7.59714782e-01 -6.07512221e-02 -9.39441621e-01
4.96366084e-01 -1.59962967e-01 1.12823629e+00 -1.62232861e-01
-3.74000706e-02 1.27164528e-01 -3.31269413e-01 -3.33157629e-01
-1.01159573e+00 3.21963876e-01 2.52180099e-01 -5.10134339e-01
7.94005871e-01 -8.14093053e-01 -1.92907065e-01 -1.65610090e-01
-8.55803967e-01 -3.56927216e-01 4.71915781e-01 5.73544323e-01
7.03208625e-01 1.74122587e-01 5.45574367e-01 -1.16644263e+00
1.05599082e+00 -4.02684659e-01 -3.45422804e-01 2.72237957e-01
-7.88524508e-01 7.39180028e-01 5.91178536e-01 -2.81241179e-01
-1.14542699e+00 -3.16147089e-01 -2.50223160e-01 -1.10820949e-01
-1.79641485e-01 7.29914188e-01 -2.32692748e-01 7.35798419e-01
4.74222749e-01 5.12239575e-01 -3.64477113e-02 -7.30787814e-01
4.12598103e-01 7.01125205e-01 5.90449929e-01 -2.94055760e-01
6.33314192e-01 3.77849102e-01 -1.97728485e-01 -9.91925180e-01
-1.01823378e+00 -6.08900607e-01 -8.59026194e-01 2.38079697e-01
6.87881947e-01 -7.47747958e-01 -1.21666156e-01 1.44659638e-01
-1.35291171e+00 -2.92593092e-01 -4.10935849e-01 4.40436065e-01
-3.05599183e-01 7.67753959e-01 -7.42674947e-01 -8.58501315e-01
-3.91200364e-01 -4.16090965e-01 9.11186397e-01 -3.50496531e-01
-8.01763773e-01 -1.50405121e+00 3.14834505e-01 -2.09038407e-01
-1.02399392e-02 -1.92066073e-01 1.35357893e+00 -1.50452971e+00
-1.71022177e-01 3.89164947e-02 2.50763506e-01 5.69353878e-01
1.53215542e-01 -1.44787207e-01 -8.13689888e-01 -2.01328188e-01
2.95291126e-01 -6.38373708e-03 1.19258893e+00 3.31269413e-01
8.76617432e-01 -4.83141959e-01 -4.14077997e-01 3.33693445e-01
1.23467267e+00 -1.50334790e-01 5.06852686e-01 4.42519218e-01
3.61492097e-01 5.23813069e-01 7.49041855e-01 5.28178513e-01
8.02802145e-02 1.20321162e-01 1.07481368e-01 3.67538095e-01
-1.37216374e-01 -6.49837494e-01 8.67461026e-01 1.36798191e+00
4.56166148e-01 -5.24596870e-01 -7.94308066e-01 5.98353148e-01
-1.68890607e+00 -9.33093846e-01 -1.99504465e-01 2.21846342e+00
1.13091922e+00 7.46920943e-01 -1.25951812e-01 1.59885213e-01
8.84182572e-01 3.11318278e-01 -2.66104132e-01 -8.06898892e-01
-4.64396805e-01 3.98540616e-01 6.35485709e-01 8.59203994e-01
-9.52469766e-01 1.25544083e+00 6.68102646e+00 1.07086480e+00
-6.79105639e-01 1.86801106e-01 4.12451506e-01 -2.25052789e-01
-7.87441611e-01 1.69694573e-01 -9.93703187e-01 6.40969157e-01
1.25518942e+00 -3.29082578e-01 -1.98942646e-02 6.98880494e-01
5.00842035e-01 -2.43551537e-01 -1.16589034e+00 5.26982188e-01
1.38556853e-01 -1.23397338e+00 2.91387737e-01 -1.63999021e-01
4.98012215e-01 1.77006811e-01 -2.39887521e-01 2.34940737e-01
2.84076065e-01 -6.75039947e-01 7.82448292e-01 3.62348229e-01
5.72647810e-01 -8.82882416e-01 6.94020271e-01 6.66609168e-01
-9.07302022e-01 1.83147669e-01 -6.59877121e-01 -2.51966536e-01
3.23875993e-01 9.83261824e-01 -9.86960292e-01 2.72835404e-01
2.53105849e-01 5.85293233e-01 -5.55122256e-01 4.85731423e-01
-5.22094250e-01 1.09139514e+00 -3.03858340e-01 -2.44528890e-01
1.47808477e-01 -7.23937228e-02 1.01211691e+00 1.75263011e+00
1.45468384e-01 -1.63637280e-01 -1.94687188e-01 7.17450500e-01
-2.73298651e-01 1.31010577e-01 -5.43061733e-01 -1.55110031e-01
8.59858215e-01 7.18780458e-01 -7.62043655e-01 -4.69590545e-01
-1.41943529e-01 1.40512395e+00 4.29509342e-01 8.28929767e-02
-6.32580757e-01 -2.37488255e-01 7.10959375e-01 -6.19014204e-02
4.46056277e-01 -2.52253950e-01 -3.08340102e-01 -1.19522142e+00
1.84418961e-01 -6.42432749e-01 3.90722781e-01 -2.89694995e-01
-1.38128638e+00 4.31434333e-01 1.81855354e-02 -8.25084865e-01
-3.18150133e-01 -3.62132549e-01 -7.36138940e-01 7.01292455e-01
-1.53043330e+00 -5.25319517e-01 5.64642608e-01 3.52158129e-01
8.01782131e-01 -2.64394581e-01 8.38215411e-01 1.97242796e-02
-5.94519198e-01 8.04557800e-01 3.76271635e-01 1.08503357e-01
6.45599604e-01 -1.37995541e+00 7.35442877e-01 1.13463259e+00
3.99373680e-01 7.95686543e-01 9.43444431e-01 -1.01573086e+00
-1.04909909e+00 -1.39411271e+00 1.75773811e+00 -5.65395355e-01
8.40068579e-01 -5.77376604e-01 -1.00610387e+00 7.88199246e-01
1.28242090e-01 -5.40135145e-01 7.70529687e-01 1.23609290e-01
-3.63148808e-01 2.21149221e-01 -9.45933878e-01 6.86286390e-01
1.21448648e+00 -8.51131201e-01 -1.34703851e+00 5.01102269e-01
1.14358151e+00 8.03945062e-04 -5.14436901e-01 2.99586058e-01
9.13593161e-04 -9.04580414e-01 5.33007443e-01 -8.03664207e-01
4.17287022e-01 -4.92167771e-02 -4.42427069e-01 -1.01210284e+00
-8.77824128e-02 -8.57113659e-01 -3.87387961e-01 1.46006322e+00
7.95803249e-01 -5.75993598e-01 5.39466798e-01 1.91245645e-01
-2.58188248e-01 -6.21658623e-01 -8.33355665e-01 -1.15437269e+00
1.85613811e-01 -7.29248643e-01 3.82441729e-01 8.05729449e-01
2.96465188e-01 4.74998713e-01 4.77115735e-02 2.71227479e-01
3.48898321e-01 -1.39623582e-01 1.19596265e-01 -7.93259919e-01
-4.67826903e-01 -3.50540578e-01 -3.99202943e-01 -1.28055441e+00
5.63003600e-01 -1.32179534e+00 4.47845012e-02 -1.16575325e+00
3.17537636e-01 1.40014857e-01 -3.65592688e-01 1.66034296e-01
-3.12463939e-02 -2.12252557e-01 -9.72381085e-02 3.73485595e-01
-7.05918968e-01 6.23370945e-01 5.82411587e-01 3.46458517e-02
-3.04089665e-01 -2.60162264e-01 -3.80382627e-01 1.09482312e+00
1.03839576e+00 -8.38880241e-01 -3.71793628e-01 -3.44151497e-01
1.28643990e-01 -3.09274673e-01 -1.24013191e-02 -6.05878413e-01
1.41268566e-01 4.74677049e-02 -3.70485149e-02 -7.43809104e-01
-2.96518534e-01 -1.23086981e-01 -6.94204569e-01 5.89043915e-01
-8.43270600e-01 1.25050977e-01 1.02283187e-01 6.32021189e-01
-4.06034142e-01 -6.58689260e-01 5.45812368e-01 -1.78666607e-01
-8.17830145e-01 -7.44505748e-02 -8.04940879e-01 4.80241984e-01
7.22293556e-01 3.71346511e-02 -1.90428123e-01 -1.47048831e-01
-7.24418581e-01 -1.09242253e-01 3.88522565e-01 1.60869271e-01
6.43585443e-01 -9.92564440e-01 -5.80487907e-01 2.26596072e-01
2.94324066e-02 -2.15847701e-01 -2.82218784e-01 8.14160883e-01
-5.56246877e-01 3.64021540e-01 5.92283428e-01 -3.18379700e-01
-1.41735351e+00 4.80131030e-01 -9.76064950e-02 -5.31350315e-01
-6.23195112e-01 8.99212956e-01 -1.91222191e-01 -5.11463821e-01
1.94373608e-01 -5.96584678e-01 2.38337308e-01 -4.81414348e-02
4.35706735e-01 2.53632128e-01 1.72026590e-01 -1.83273241e-01
-5.21791697e-01 -1.29841968e-01 -2.79302150e-01 -3.57332289e-01
1.18699265e+00 -5.52128255e-01 -4.77577060e-01 6.67457998e-01
1.23637557e+00 3.95640552e-01 -7.87182212e-01 -2.57660210e-01
9.05980170e-01 -1.40705571e-01 -7.98246637e-02 -5.95239878e-01
-2.32573956e-01 6.50390029e-01 2.14482695e-02 1.09393217e-01
9.82959688e-01 3.35714221e-01 8.76871288e-01 5.89125454e-01
-3.96332070e-02 -1.42070401e+00 1.22162122e-02 1.06443799e+00
5.18745542e-01 -7.35806346e-01 -9.22230408e-02 -3.61163616e-01
-5.80612004e-01 9.41518724e-01 1.44190550e-01 -3.68964672e-02
6.88749433e-01 4.92811441e-01 -1.75392091e-01 -3.41602936e-02
-1.01942575e+00 -8.92137364e-02 7.24027902e-02 3.41494620e-01
4.66520399e-01 -1.81961387e-01 -7.00391531e-01 6.13292575e-01
-4.47002679e-01 -5.17904341e-01 4.68731403e-01 1.02244866e+00
-9.37310815e-01 -1.30935133e+00 4.84453514e-02 4.17942584e-01
-5.22326171e-01 -7.57483721e-01 -5.01812994e-01 6.54230893e-01
-1.54009402e-01 8.18251371e-01 3.93675357e-01 -2.42587462e-01
8.78613293e-02 4.53868836e-01 4.72090751e-01 -8.56581330e-01
-1.82142347e-01 7.54645169e-02 5.01281083e-01 -4.35980648e-01
-2.07264990e-01 -8.10289621e-01 -1.66094184e+00 -3.65254343e-01
-2.46441215e-01 5.70502318e-02 3.97679955e-01 1.28152812e+00
4.02772039e-01 2.74696022e-01 5.00054657e-01 6.26017228e-02
-1.09658933e+00 -1.10114253e+00 -9.20038283e-01 5.51187754e-01
1.19627595e-01 -8.23532864e-02 -6.21522903e-01 9.79808420e-02] | [10.95804500579834, 8.873529434204102] |
595b67dd-1cbd-4dec-aaaa-2a6e19570eca | exploring-representation-learning-for-small | 2303.10912 | null | https://arxiv.org/abs/2303.10912v1 | https://arxiv.org/pdf/2303.10912v1.pdf | Exploring Representation Learning for Small-Footprint Keyword Spotting | In this paper, we investigate representation learning for low-resource keyword spotting (KWS). The main challenges of KWS are limited labeled data and limited available device resources. To address those challenges, we explore representation learning for KWS by self-supervised contrastive learning and self-training with pretrained model. First, local-global contrastive siamese networks (LGCSiam) are designed to learn similar utterance-level representations for similar audio samplers by proposed local-global contrastive loss without requiring ground-truth. Second, a self-supervised pretrained Wav2Vec 2.0 model is applied as a constraint module (WVC) to force the KWS model to learn frame-level acoustic representations. By the LGCSiam and WVC modules, the proposed small-footprint KWS model can be pretrained with unlabeled data. Experiments on speech commands dataset show that the self-training WVC module and the self-supervised LGCSiam module significantly improve accuracy, especially in the case of training on a small labeled dataset. | ['Yujun Wang', 'Peng Gao', 'Quandong Wang', 'Liyong Guo', 'Fan Cui'] | 2023-03-20 | null | null | null | null | ['small-footprint-keyword-spotting', 'keyword-spotting'] | ['speech', 'speech'] | [ 2.40174472e-01 -6.59794137e-02 -4.01797324e-01 -4.73809093e-01
-1.37479758e+00 -1.31045535e-01 3.43249232e-01 -2.69801885e-01
-4.02987421e-01 4.57019329e-01 3.63474071e-01 -2.99029976e-01
2.36315817e-01 -2.46714994e-01 -7.88340509e-01 -4.36284900e-01
2.16652095e-01 2.68734582e-02 1.04067527e-01 -1.36811463e-02
-1.55467719e-01 2.37758741e-01 -1.48234725e+00 3.40945333e-01
7.36000538e-01 1.24756706e+00 8.34694088e-01 8.81903887e-01
-1.34157166e-01 8.29127669e-01 -5.21055639e-01 2.03329161e-01
1.81821197e-01 -6.29398823e-01 -7.19737291e-01 -1.81109071e-01
6.36653900e-01 -3.04333687e-01 -4.31266189e-01 7.11646914e-01
8.80455494e-01 4.95558769e-01 7.34788418e-01 -1.03733563e+00
-5.74426830e-01 6.34530067e-01 -4.38714445e-01 1.94587573e-01
5.63767180e-02 8.86846483e-02 1.25720787e+00 -1.22071505e+00
2.80696899e-01 1.15236318e+00 4.43864226e-01 7.13898361e-01
-9.89854515e-01 -8.64774525e-01 2.50188023e-01 5.41107297e-01
-1.75759971e+00 -8.59377086e-01 9.50993001e-01 1.97929963e-02
1.29631305e+00 4.49601859e-01 3.27416837e-01 9.48282957e-01
-3.45358193e-01 1.31183898e+00 7.04367638e-01 -6.79823279e-01
3.83984685e-01 2.99314648e-01 4.01281789e-02 6.12042964e-01
-4.00393128e-01 -2.33509049e-01 -1.10631061e+00 -1.88656822e-01
4.90917563e-01 -2.29418606e-01 -3.67278397e-01 -3.60404521e-01
-8.99069905e-01 5.97736359e-01 1.42207056e-01 4.46371973e-01
3.67273428e-02 3.11737329e-01 5.79218745e-01 3.75549674e-01
6.25011683e-01 3.12774301e-01 -7.42426932e-01 -1.41884074e-01
-1.45965338e+00 -1.61767840e-01 3.05345386e-01 1.22264421e+00
7.13576138e-01 8.62536609e-01 -3.12434256e-01 1.24432039e+00
3.32055867e-01 8.01751196e-01 1.11186182e+00 -5.37059784e-01
6.25749588e-01 -1.47817492e-01 -4.24711496e-01 -4.85233933e-01
-5.53087778e-02 -4.56655920e-01 -7.28195846e-01 -3.44689637e-01
-1.86032474e-01 5.83149493e-02 -8.62081349e-01 1.81477737e+00
-1.09889340e-02 8.86885285e-01 3.60315830e-01 9.38924491e-01
9.70679343e-01 1.01739168e+00 5.19251153e-02 -4.92377728e-01
1.01738858e+00 -1.33658803e+00 -9.26453769e-01 -9.07064006e-02
7.88535893e-01 -6.99612379e-01 1.53850615e+00 2.33468086e-01
-1.15490949e+00 -8.99600685e-01 -1.18894351e+00 -6.18764050e-02
-3.52009088e-01 5.30066013e-01 2.43997529e-01 5.30446768e-01
-9.74119961e-01 4.64268446e-01 -6.63774490e-01 -1.40113279e-01
3.94594759e-01 2.91026533e-01 -6.69920221e-02 5.92742674e-02
-1.26588535e+00 4.71423239e-01 2.35112578e-01 -6.57911375e-02
-1.26125157e+00 -7.63757527e-01 -1.05442798e+00 2.97363907e-01
4.46307838e-01 -1.96285307e-01 1.38669574e+00 -8.54676306e-01
-2.12482142e+00 5.08224547e-01 -2.42754892e-01 -6.39976144e-01
-5.49525060e-02 -3.57640207e-01 -7.25020230e-01 1.56271741e-01
-1.49701163e-01 4.11846668e-01 1.34604287e+00 -1.08080065e+00
-4.08092588e-01 -4.84108701e-02 -5.47209263e-01 4.64848191e-01
-7.02458024e-01 -6.80729821e-02 -7.43775427e-01 -9.47026491e-01
-8.39353874e-02 -4.99900430e-01 -7.95985535e-02 -3.37472558e-01
-3.70014876e-01 -3.70383382e-01 9.75393116e-01 -7.66668916e-01
1.39546418e+00 -2.40595794e+00 6.22749776e-02 2.12803856e-01
-1.35520831e-01 6.57907426e-01 -4.88264024e-01 8.71908888e-02
6.11458570e-02 -2.27471143e-01 -1.11417309e-03 -8.13091278e-01
-7.18831643e-02 1.75523713e-01 -5.21439016e-01 3.44606608e-01
3.41811419e-01 8.88419986e-01 -8.90336990e-01 -5.63829601e-01
3.52128267e-01 4.13141847e-01 -5.27537882e-01 8.27071786e-01
-2.30569065e-01 2.10167065e-01 -9.40304920e-02 4.82656956e-01
5.10822475e-01 8.25004131e-02 1.04480557e-01 -3.25317532e-01
-1.19333330e-03 5.17856300e-01 -1.23083389e+00 2.32762718e+00
-9.68168914e-01 5.92997611e-01 -3.49986888e-02 -1.35725427e+00
9.54769373e-01 3.83491069e-01 2.10872680e-01 -7.98852801e-01
-1.50259947e-02 2.61589199e-01 -3.77502650e-01 -3.58365983e-01
5.36833584e-01 1.12044647e-01 -1.49728300e-03 3.91829073e-01
7.53171146e-01 -4.60555106e-01 -4.99977171e-01 3.18315864e-01
7.59673238e-01 1.79353073e-01 1.35193244e-01 -1.44015789e-01
5.90559721e-01 -5.12830436e-01 5.78054667e-01 7.02428162e-01
1.20150134e-01 8.78205359e-01 -3.00549477e-01 -4.30185534e-02
-8.38298023e-01 -1.00278759e+00 1.00270741e-01 1.36535585e+00
-2.53690947e-02 -5.89033484e-01 -5.98179877e-01 -5.12295902e-01
-2.80257165e-01 6.76999927e-01 -1.55199081e-01 -3.29094797e-01
-4.53041047e-01 -1.77197963e-01 6.64649367e-01 6.65092647e-01
3.88865471e-01 -8.21115851e-01 -5.49252927e-01 2.43478984e-01
1.92957163e-01 -1.14154291e+00 -9.02577400e-01 4.64813471e-01
-5.19780338e-01 -6.12483025e-01 -1.01482153e+00 -9.54369068e-01
3.96524668e-01 4.48444843e-01 6.04661047e-01 -3.16578418e-01
-3.23030055e-01 4.69218373e-01 -5.50044000e-01 -3.19540918e-01
-8.61946493e-02 1.38294622e-01 4.76829410e-01 7.89815336e-02
1.08758628e-01 -5.01234412e-01 -3.82067919e-01 5.90576418e-02
-5.63558340e-01 3.44038233e-02 3.49578470e-01 1.09815776e+00
8.59332442e-01 -1.54371738e-01 7.96091259e-01 -7.00395226e-01
5.55215299e-01 -3.04682165e-01 -4.89301115e-01 4.27942574e-01
-7.35668957e-01 1.40735313e-01 8.84834409e-01 -8.44668746e-01
-1.02637422e+00 5.64856902e-02 -1.95390761e-01 -9.41577375e-01
1.37737051e-01 4.67152506e-01 -4.48295623e-01 1.26063257e-01
3.46869290e-01 4.86266106e-01 -2.32979998e-01 -6.31920159e-01
6.99799478e-01 1.07682776e+00 6.65877283e-01 -4.56435591e-01
7.59645641e-01 2.41282489e-02 -5.68876803e-01 -1.07815623e+00
-6.87737226e-01 -6.80035770e-01 -2.86013305e-01 -2.34144647e-02
5.92703819e-01 -1.19450486e+00 -2.55754024e-01 2.32034355e-01
-9.30036604e-01 -6.05930507e-01 -8.38305652e-01 9.24703062e-01
-6.11508727e-01 3.54709178e-01 -5.17559648e-01 -1.16130292e+00
-5.69251716e-01 -1.00630295e+00 1.02722168e+00 1.11906990e-01
-1.90896839e-01 -7.76093781e-01 9.58963558e-02 2.50995576e-01
6.62815332e-01 -5.30556738e-01 6.85965240e-01 -9.91323233e-01
-1.82659954e-01 2.13308204e-02 6.45629615e-02 8.53393435e-01
1.53211042e-01 -4.97386396e-01 -1.47059488e+00 -3.03162187e-01
3.75923477e-02 -7.59292543e-01 8.40247929e-01 3.00120503e-01
1.58991015e+00 -2.15058714e-01 -3.32043618e-02 7.86719143e-01
1.09883451e+00 3.07049662e-01 3.06294531e-01 -2.09247082e-01
8.26955080e-01 1.00584291e-01 7.29511440e-01 4.03742909e-01
9.07835439e-02 8.68799865e-01 -1.25504211e-01 -2.78018534e-01
-4.95714277e-01 -5.71949601e-01 3.83201927e-01 1.73972511e+00
2.82274663e-01 -1.36444613e-01 -5.52249968e-01 5.43226421e-01
-1.87603259e+00 -7.04762399e-01 4.79986310e-01 2.23177838e+00
9.64144528e-01 -8.04386213e-02 -8.52514356e-02 2.64130980e-01
5.01431227e-01 4.49390590e-01 -7.11981058e-01 -4.31390166e-01
-2.17623979e-01 6.18142128e-01 3.82494539e-01 4.93553162e-01
-9.21390116e-01 1.27662253e+00 5.66678143e+00 1.65138113e+00
-1.16297507e+00 4.83433396e-01 5.75750589e-01 -1.87683538e-01
-3.56606245e-01 -2.84097761e-01 -7.10736513e-01 3.34791303e-01
1.23670554e+00 -3.54320370e-02 5.01029313e-01 9.87633049e-01
2.11552352e-01 1.66458786e-01 -9.95936096e-01 1.40983701e+00
4.10631210e-01 -1.31288815e+00 -5.42994812e-02 -4.46295738e-01
6.76320136e-01 -1.22684583e-01 4.51763600e-01 6.60258472e-01
1.13140285e-01 -8.68321300e-01 6.91958547e-01 3.02080750e-01
1.43295729e+00 -7.89406359e-01 5.67809284e-01 1.90759435e-01
-1.61298621e+00 1.38663799e-01 -6.24269128e-01 1.44856900e-01
5.63687012e-02 3.26767743e-01 -8.86841357e-01 5.92276573e-01
5.46574771e-01 5.73449254e-01 -5.84126152e-02 5.70734262e-01
-5.95158301e-02 1.18853688e+00 -3.56687546e-01 -2.51389761e-02
1.97444856e-01 7.18517825e-02 2.43668571e-01 1.55633962e+00
2.17732042e-01 6.60015345e-02 1.28257066e-01 5.78743279e-01
-2.42819726e-01 4.97119009e-01 -3.90624404e-01 -1.62478358e-01
5.98387837e-01 1.08905745e+00 -2.10689485e-01 -6.06691837e-01
-4.27504331e-01 1.32715654e+00 4.30872202e-01 6.02946997e-01
-6.92807019e-01 -3.55171919e-01 5.29553056e-01 -1.03524752e-01
3.94281358e-01 -1.53127998e-01 1.03845790e-01 -1.22643280e+00
-2.18167648e-01 -7.25895822e-01 2.29838744e-01 -8.34204733e-01
-1.01421559e+00 6.42787933e-01 -1.35196209e-01 -1.35299814e+00
-2.98295259e-01 -3.08853179e-01 -6.63185239e-01 9.48607385e-01
-1.81043434e+00 -1.24620652e+00 1.10708266e-01 8.87504637e-01
1.05257559e+00 -8.53590429e-01 1.08355892e+00 2.18913436e-01
-4.84468430e-01 1.17203319e+00 2.93571085e-01 -3.86022627e-02
8.51774871e-01 -1.07555604e+00 2.78566271e-01 6.49237990e-01
6.86944187e-01 4.96876299e-01 3.50125074e-01 -4.01224792e-01
-1.53641427e+00 -1.20713067e+00 5.15473545e-01 1.68771908e-01
5.99776030e-01 -6.56465471e-01 -9.17821586e-01 2.68065631e-01
2.13105977e-01 4.11021858e-01 1.05305684e+00 7.64145106e-02
-5.94880879e-01 -2.59633034e-01 -8.28112245e-01 3.50113273e-01
1.01021719e+00 -1.30445564e+00 -4.35255915e-01 4.61465299e-01
1.08273530e+00 -3.90050203e-01 -4.91764694e-01 2.56776959e-01
4.33840901e-01 -2.03287408e-01 8.71418476e-01 -7.34531641e-01
-9.39259827e-02 -2.54488051e-01 -6.34633303e-01 -1.39810705e+00
6.44100308e-02 -7.10984349e-01 -5.00641108e-01 1.40572166e+00
4.62706715e-01 -1.34235218e-01 7.36545205e-01 2.09011987e-01
-4.66608822e-01 -7.16458082e-01 -1.27521706e+00 -1.09424937e+00
-2.23703206e-01 -9.92276788e-01 4.95469630e-01 9.38331306e-01
1.72880441e-01 5.64599097e-01 -9.29433167e-01 1.63443342e-01
2.95421749e-01 -1.30570112e-02 6.80119157e-01 -6.52846873e-01
-7.02980876e-01 8.02047104e-02 -3.65090281e-01 -1.39775801e+00
5.41159570e-01 -1.06987429e+00 2.79947698e-01 -1.04085028e+00
7.61163682e-02 -3.38530570e-01 -5.33388197e-01 3.42004120e-01
-7.20857680e-02 1.00151971e-01 2.61087298e-01 1.63095042e-01
-8.98217261e-01 9.94311213e-01 7.89221406e-01 -4.55045730e-01
-4.01405841e-01 -9.09256637e-02 -2.00510353e-01 4.31533307e-01
8.06136310e-01 -2.84657091e-01 -8.26219499e-01 -3.50701541e-01
-4.32440400e-01 -4.30972427e-02 -7.45503306e-02 -1.11352003e+00
4.42085803e-01 -2.80950572e-02 1.69932276e-01 -4.68411446e-01
7.51081407e-01 -7.78201461e-01 -3.07512403e-01 1.74604475e-01
-6.03811026e-01 -5.43811858e-01 2.58794755e-01 5.96099019e-01
-3.51573318e-01 -3.10289174e-01 7.12512016e-01 1.80691749e-01
-8.43257070e-01 2.63261199e-01 -5.02649963e-01 3.43954042e-02
6.53681755e-01 1.35303773e-02 3.18937391e-01 -7.16780901e-01
-6.57160878e-01 1.90483525e-01 -2.32412949e-01 3.56175393e-01
1.06011093e+00 -1.56968641e+00 -4.44981456e-01 5.16010463e-01
3.09086531e-01 3.67585756e-02 3.24358404e-01 2.49502659e-01
1.64640009e-01 4.08395231e-01 3.42554539e-01 -4.76584762e-01
-1.31533885e+00 4.86542970e-01 9.57599208e-02 -1.47710562e-01
-3.50689441e-01 1.26455402e+00 8.28661397e-02 -4.47326303e-01
6.37625813e-01 -5.07872522e-01 -1.06675230e-01 -1.59114182e-01
7.70817697e-01 1.24106444e-01 4.77386871e-03 -6.28547370e-01
-3.04019988e-01 4.48436499e-01 -1.66144118e-01 -3.58220249e-01
1.27780521e+00 -1.11583017e-01 5.97217262e-01 7.53937781e-01
1.52543736e+00 2.68306714e-02 -1.09466505e+00 -7.19315469e-01
-2.78709352e-01 -2.70670712e-01 3.76312613e-01 -3.45991552e-01
-1.02862024e+00 1.12015617e+00 8.69292974e-01 -4.61930364e-01
1.14524281e+00 8.31343681e-02 8.68323803e-01 5.10746241e-01
2.85909086e-01 -1.73511815e+00 4.16940570e-01 5.20132065e-01
7.06129253e-01 -9.89192426e-01 -2.25958735e-01 -1.32697314e-01
-8.68613243e-01 8.73158693e-01 6.43101990e-01 -2.12414302e-02
6.90547884e-01 3.94701868e-01 1.95401795e-02 1.68700203e-01
-6.03905618e-01 -3.27515095e-01 5.38795948e-01 6.16360366e-01
3.64945382e-01 3.82549018e-02 1.00284837e-01 9.42929447e-01
-4.46494408e-02 -2.75223255e-01 -6.00478686e-02 7.86349535e-01
-3.38762969e-01 -9.01373327e-01 -9.98402983e-02 5.13448000e-01
-1.24649204e-01 -4.14404660e-01 -5.49964383e-02 1.70421585e-01
-1.03791893e-01 8.93854082e-01 3.82043160e-02 -7.85446703e-01
3.82478833e-01 2.97457904e-01 2.00870872e-01 -8.18969429e-01
-3.40618789e-01 4.83113408e-01 -3.35214660e-02 -5.51447451e-01
-3.39566886e-01 -2.93915838e-01 -1.25363028e+00 2.88198113e-01
-8.38131130e-01 4.82220411e-01 6.53587103e-01 8.99266422e-01
2.58615613e-01 7.35348165e-01 9.25634265e-01 -7.87042737e-01
-6.38945520e-01 -1.13022256e+00 -8.90556931e-01 2.69044489e-01
2.98660278e-01 -3.26750219e-01 -2.72782952e-01 9.23362970e-02] | [14.570502281188965, 6.308271408081055] |
f55faf8e-26c6-4e57-ac94-88121a18d4d9 | quick-dense-retrievers-consume-kale-post | 2304.01016 | null | https://arxiv.org/abs/2304.01016v3 | https://arxiv.org/pdf/2304.01016v3.pdf | Quick Dense Retrievers Consume KALE: Post Training Kullback Leibler Alignment of Embeddings for Asymmetrical dual encoders | In this paper, we consider the problem of improving the inference latency of language model-based dense retrieval systems by introducing structural compression and model size asymmetry between the context and query encoders. First, we investigate the impact of pre and post-training compression on the MSMARCO, Natural Questions, TriviaQA, SQUAD, and SCIFACT, finding that asymmetry in the dual encoders in dense retrieval can lead to improved inference efficiency. Knowing this, we introduce Kullback Leibler Alignment of Embeddings (KALE), an efficient and accurate method for increasing the inference efficiency of dense retrieval methods by pruning and aligning the query encoder after training. Specifically, KALE extends traditional Knowledge Distillation after bi-encoder training, allowing for effective query encoder compression without full retraining or index generation. Using KALE and asymmetric training, we can generate models which exceed the performance of DistilBERT despite having 3x faster inference. | ['ChengXiang Zhai', 'Alessandro Magnani', 'Daniel Campos'] | 2023-03-31 | null | null | null | null | ['natural-questions', 'triviaqa'] | ['miscellaneous', 'miscellaneous'] | [-1.49989694e-01 5.93707114e-02 -3.06304544e-01 -4.79094684e-02
-1.16211355e+00 -8.16982985e-01 7.83383489e-01 2.26176262e-01
-7.79438317e-01 5.41244507e-01 3.96087557e-01 -6.22740388e-01
-4.64446604e-01 -8.74514341e-01 -8.66084754e-01 -1.55499339e-01
-1.20296955e-01 1.12296975e+00 2.13173240e-01 -2.22707182e-01
2.92273253e-01 4.15355086e-01 -1.52209067e+00 3.61217976e-01
6.07288539e-01 7.31082499e-01 2.87778646e-01 1.14177477e+00
-9.11214724e-02 9.60868835e-01 -5.13349891e-01 -7.79503047e-01
3.15069348e-01 1.50822354e-02 -1.20638263e+00 -8.46868813e-01
5.90657413e-01 -1.11003053e+00 -8.79087389e-01 6.94618225e-01
7.07698107e-01 -1.71671454e-02 7.23400533e-01 -8.41332138e-01
-5.45708120e-01 1.12021101e+00 -1.77718058e-01 4.47855085e-01
2.76392132e-01 1.55832022e-02 1.43036437e+00 -7.38921523e-01
8.37798417e-01 1.54359436e+00 3.96227986e-01 4.07129377e-01
-1.20964742e+00 -5.79761684e-01 -5.77513278e-01 3.87003541e-01
-1.75371552e+00 -6.91940129e-01 -7.22171925e-03 4.57897335e-02
1.63159311e+00 4.28670228e-01 3.25986445e-01 6.43058300e-01
-3.23890382e-03 8.45049262e-01 3.64346921e-01 -6.22174203e-01
1.50331333e-01 1.48644343e-01 1.28732920e-01 7.04544663e-01
5.49798846e-01 2.37889305e-01 -4.96354908e-01 -5.92831075e-01
3.30955923e-01 -1.87222406e-01 2.32206602e-02 6.51577860e-02
-8.95845294e-01 8.86064887e-01 3.16420197e-01 1.53769553e-01
-2.35474125e-01 5.00813127e-01 5.00003695e-01 5.51302373e-01
1.81855306e-01 1.11251104e+00 -4.53398526e-01 -4.82828438e-01
-1.30774999e+00 6.90791130e-01 1.19743872e+00 1.05930781e+00
7.70403087e-01 -4.45711732e-01 -5.62855363e-01 6.17686152e-01
1.79709926e-01 8.40574622e-01 5.58311045e-01 -1.30582964e+00
5.68110704e-01 2.96209574e-01 -1.72456190e-01 -7.88536370e-01
4.90461625e-02 -3.11884463e-01 -4.30425107e-01 -6.42269611e-01
6.00973591e-02 -1.22294165e-01 -7.22300112e-01 1.55439794e+00
5.22538386e-02 -2.44634338e-02 2.86007583e-01 4.42926258e-01
2.97680646e-01 9.49149132e-01 -1.40869349e-01 -3.55558135e-02
1.36830986e+00 -8.65104020e-01 -6.71100795e-01 -1.91874728e-01
1.38152456e+00 -8.54857445e-01 9.63462234e-01 1.93620846e-01
-1.45037711e+00 -2.97425777e-01 -1.12059414e+00 -8.48495185e-01
-4.23094243e-01 -2.40615115e-01 8.09905350e-01 2.30417520e-01
-1.28972363e+00 5.71984291e-01 -9.62924719e-01 -3.98613960e-02
1.21073034e-02 4.27451640e-01 -1.87288180e-01 -5.22662222e-01
-1.45526183e+00 1.04058015e+00 6.17022336e-01 -2.59627432e-01
-8.28242183e-01 -1.02922654e+00 -5.88736176e-01 5.91050446e-01
1.68232843e-01 -8.73261392e-01 1.47304904e+00 -5.37742563e-02
-1.19275677e+00 4.73917246e-01 -2.74409711e-01 -1.01283312e+00
-1.10834099e-01 -3.91960055e-01 -4.91829729e-03 4.69710886e-01
-3.05184394e-01 9.85171437e-01 5.43273509e-01 -7.08775878e-01
-1.93909198e-01 -3.39436829e-01 3.03468734e-01 2.09014177e-01
-4.48062330e-01 -1.10592484e-01 -7.65693843e-01 -3.13255876e-01
-3.89848351e-01 -7.92352378e-01 7.12157190e-02 -3.03617805e-01
-3.38995308e-01 -3.78090471e-01 6.69113934e-01 -8.95571828e-01
1.85339749e+00 -2.22102642e+00 2.68238366e-01 3.71786147e-01
4.12626565e-01 5.91365993e-01 -4.11458045e-01 7.36439347e-01
6.47375822e-01 2.02406764e-01 -2.02636030e-02 -4.02935386e-01
3.24012965e-01 6.91090524e-01 -6.01763725e-01 -9.53729078e-02
3.37458476e-02 1.31790626e+00 -8.23204756e-01 -9.12244976e-01
-1.99001759e-01 5.52619219e-01 -1.06675756e+00 3.63643557e-01
-5.60719848e-01 -6.02904201e-01 -2.57175297e-01 4.82213169e-01
4.58878696e-01 -3.47192019e-01 3.41312051e-01 -7.94513896e-02
3.88065010e-01 7.91920602e-01 -7.25700319e-01 1.79247677e+00
-7.15326369e-01 7.82867014e-01 -1.50432140e-01 -5.15343010e-01
4.70983058e-01 1.42407492e-01 1.47048146e-01 -1.13555622e+00
-2.32494026e-01 4.59812194e-01 -7.47626349e-02 -3.21240336e-01
1.10991263e+00 4.23933864e-02 6.32955693e-04 7.43500471e-01
2.60826081e-01 -3.42990130e-01 4.85065699e-01 8.12011659e-01
1.24583375e+00 -4.76973921e-01 -3.67727965e-01 1.56671822e-03
1.70406833e-01 -2.73132890e-01 -1.04389705e-01 1.11028910e+00
5.73716700e-01 2.30012044e-01 4.03664887e-01 7.10800849e-03
-1.35668910e+00 -1.03172088e+00 -2.57462531e-01 1.20287454e+00
-1.74373627e-01 -1.10041153e+00 -7.15776861e-01 -3.38027984e-01
4.07749176e-01 9.34589326e-01 -1.90478504e-01 -6.92921579e-01
-7.49695897e-01 -3.56217086e-01 1.31966698e+00 4.69994724e-01
4.05364007e-01 -5.44947386e-01 -5.37593007e-01 4.93995585e-02
-2.01993704e-01 -1.08437097e+00 -4.68950748e-01 1.11226954e-01
-1.07979655e+00 -7.29694307e-01 -5.15403569e-01 -3.24334741e-01
2.42871508e-01 5.68284877e-02 1.59559822e+00 3.16583425e-01
-3.50659460e-01 4.56354499e-01 -2.37695128e-01 -1.98048696e-01
-4.75079060e-01 7.81918764e-01 -1.46835521e-01 -1.02240694e+00
5.59040487e-01 -5.76337874e-01 -6.56021535e-01 -1.11245908e-01
-1.35704172e+00 -1.57361880e-01 9.71924484e-01 7.95683682e-01
4.72590625e-01 -1.75309926e-01 1.95836574e-01 -8.08238745e-01
8.81809652e-01 -4.66725826e-01 -7.82236397e-01 6.17848754e-01
-1.14320517e+00 9.64726865e-01 2.11740613e-01 -1.09668978e-01
-7.07839966e-01 -6.22910500e-01 -2.48435035e-01 -5.50643384e-01
5.74112833e-01 6.73483133e-01 4.32982743e-01 5.94166741e-02
7.76565373e-01 2.04233289e-01 1.11582607e-01 -7.16181815e-01
8.37267756e-01 7.70312846e-01 4.47399557e-01 -8.89468908e-01
5.36461353e-01 1.51483025e-02 1.30861579e-02 -4.92240876e-01
-5.73300719e-01 -5.51917315e-01 -3.60678405e-01 5.49367070e-01
5.15045166e-01 -9.27305639e-01 -5.49415052e-01 -1.01357132e-01
-1.37642848e+00 -2.13253915e-01 -6.81166470e-01 3.90132248e-01
-3.53194207e-01 3.49900335e-01 -9.28210616e-01 -5.24596453e-01
-8.44587028e-01 -9.83877122e-01 1.49709177e+00 -1.55850932e-01
-1.19319573e-01 -9.98590648e-01 4.03615832e-01 2.62872815e-01
7.63905287e-01 -6.43660843e-01 1.38999939e+00 -7.54402399e-01
-7.89915502e-01 -3.86865735e-01 -4.09247607e-01 5.53915322e-01
-5.58732033e-01 -3.24523270e-01 -8.23637426e-01 -4.46088135e-01
-3.51607740e-01 -6.76029503e-01 9.34931517e-01 -1.50432214e-01
1.17609835e+00 -8.22765946e-01 -2.59315819e-01 7.97506630e-01
1.48386037e+00 -1.98277712e-01 8.62027287e-01 1.19522542e-01
3.41568977e-01 1.30755201e-01 1.88261971e-01 3.97241473e-01
2.62912244e-01 6.83409989e-01 5.56113105e-03 3.97918582e-01
-2.87184119e-01 -6.03090346e-01 4.26283747e-01 1.18492663e+00
2.56158829e-01 -2.11630613e-01 -8.92855227e-01 5.09799480e-01
-1.33522213e+00 -7.69346535e-01 6.22403920e-01 2.30444527e+00
1.50090015e+00 -1.42920576e-02 -1.81830958e-01 -4.32808660e-02
-1.43893259e-02 4.63419966e-02 -3.62867951e-01 -7.97323823e-01
5.58850393e-02 9.85015869e-01 9.30443108e-01 9.36812520e-01
-3.08357924e-01 9.52910244e-01 6.92416763e+00 1.08223128e+00
-8.47240031e-01 1.70194551e-01 1.96173325e-01 -6.19346619e-01
-6.83226407e-01 1.35361195e-01 -1.20915139e+00 3.20517331e-01
1.75937235e+00 -5.34470677e-01 7.91880488e-01 7.62407899e-01
-6.06347978e-01 -1.85200348e-02 -1.46131074e+00 7.75463939e-01
-1.59483775e-02 -1.56573451e+00 4.82186347e-01 1.94646835e-01
4.06906724e-01 1.74370781e-01 1.05694324e-01 8.61477494e-01
2.14653179e-01 -1.13734937e+00 2.18802437e-01 6.17367983e-01
9.67658520e-01 -6.57307267e-01 8.63972545e-01 3.21929038e-01
-6.10464811e-01 -6.91156611e-02 -6.97913468e-01 1.13045074e-01
3.94631438e-02 4.24866140e-01 -1.15620995e+00 3.60603988e-01
1.16235837e-01 9.44583192e-02 -8.39865685e-01 7.42015600e-01
2.02499911e-01 5.91068506e-01 -6.97761297e-01 -1.16707020e-01
2.73724943e-01 2.58868128e-01 2.62115687e-01 1.41579485e+00
2.67881930e-01 -9.17363986e-02 -3.38875115e-01 6.82112277e-01
-4.41857994e-01 -2.85954326e-01 -7.01567471e-01 -5.04188836e-01
8.94025266e-01 6.53929353e-01 2.90341765e-01 -6.79902673e-01
-6.77449703e-02 9.66032445e-01 3.94050241e-01 3.11947137e-01
-4.92290020e-01 -6.30811155e-01 3.84234339e-01 3.08070153e-01
3.51956874e-01 -3.49731028e-01 8.33607614e-02 -1.11899042e+00
6.54842108e-02 -1.06057549e+00 4.54729289e-01 -5.79168141e-01
-7.90137351e-01 1.36905149e-01 6.17158592e-01 -3.63549381e-01
-1.01894021e+00 -2.63768941e-01 8.59131515e-02 9.86725807e-01
-1.68455827e+00 -7.40320146e-01 2.61178553e-01 3.76242220e-01
1.80647224e-01 1.77767351e-02 1.02237284e+00 7.90117860e-01
-4.25512753e-02 1.16902030e+00 5.20308375e-01 -3.15604694e-02
4.61270690e-01 -1.13147032e+00 3.58434021e-01 3.95606428e-01
4.02140796e-01 1.24065483e+00 4.03722644e-01 -3.31602663e-01
-2.13774490e+00 -8.25506508e-01 1.28869140e+00 -5.83753824e-01
6.10729635e-01 -3.53475004e-01 -8.66431117e-01 7.12699592e-01
3.43238711e-01 -2.57372171e-01 5.70957124e-01 4.21130985e-01
-7.26074934e-01 -2.71086961e-01 -9.33389783e-01 4.65397090e-01
7.82705247e-01 -1.21853089e+00 -7.96044230e-01 5.68558574e-01
1.15552819e+00 -3.98514539e-01 -1.20877981e+00 4.39304382e-01
7.97092378e-01 -4.77039754e-01 1.24436760e+00 -7.40729332e-01
5.33347547e-01 2.77665675e-01 -3.68862838e-01 -7.59209633e-01
-7.61043951e-02 -5.04009545e-01 -1.03858364e+00 8.59739184e-01
5.00959992e-01 -3.37888926e-01 7.17177212e-01 6.66332960e-01
1.11668199e-01 -9.41887140e-01 -9.82031286e-01 -7.40991235e-01
4.57453132e-01 -2.76891142e-01 7.42840707e-01 3.99890751e-01
3.24846692e-02 5.60873032e-01 -6.42684251e-02 -1.80675060e-01
2.89210916e-01 -1.63295701e-01 8.25582445e-01 -7.22310722e-01
-8.38110268e-01 -3.31503183e-01 -2.58397520e-01 -1.62235475e+00
1.71636537e-01 -1.07869983e+00 -2.11232767e-01 -1.25208604e+00
3.21367949e-01 -4.92107987e-01 -9.59366038e-02 2.41802424e-01
-7.35216439e-02 -1.77852064e-01 1.64964467e-01 3.50226909e-01
-8.66210639e-01 5.01336992e-01 9.58035827e-01 -1.99373528e-01
1.74466521e-01 -3.85320216e-01 -5.81742525e-01 -3.75050567e-02
3.03253472e-01 -3.88933867e-01 -8.09034765e-01 -1.13201046e+00
8.11917424e-01 2.80580640e-01 2.23380655e-01 -8.34294081e-01
5.82856596e-01 5.37824929e-01 -1.08738057e-01 -6.22995377e-01
3.83722186e-01 -7.11490214e-01 -1.58765391e-01 4.39337879e-01
-8.43227863e-01 6.23237193e-01 4.78900343e-01 4.37043279e-01
-5.03119946e-01 -5.10100007e-01 3.67161691e-01 -2.16621861e-01
-2.79952079e-01 2.85187304e-01 -5.95991090e-02 5.14438391e-01
2.30647519e-01 3.37080926e-01 -4.94236887e-01 -3.45091105e-01
-2.93515712e-01 3.21504980e-01 2.23500103e-01 4.58434150e-02
6.90021455e-01 -1.18617284e+00 -5.52712679e-01 3.29091132e-01
-3.08103133e-02 2.34353900e-01 1.32188033e-02 6.78703368e-01
-7.85478055e-01 1.18150675e+00 2.82009989e-01 -2.78760880e-01
-1.05709529e+00 4.05293256e-01 2.95507669e-01 -9.31604743e-01
-3.02211672e-01 1.03333819e+00 -3.92795414e-01 -3.24608743e-01
5.01659572e-01 -3.38003844e-01 4.47732627e-01 -2.05353558e-01
5.78268588e-01 4.38346416e-01 3.24781448e-01 1.56656697e-01
-1.14198707e-01 7.73093030e-02 -7.10746288e-01 -4.62945819e-01
9.64400113e-01 1.79586858e-01 -4.86525595e-01 5.48024289e-02
1.87931097e+00 -3.49603058e-03 -4.80257511e-01 -4.15586889e-01
1.39475077e-01 -4.37733680e-01 4.41843152e-01 -6.51868999e-01
-6.90580904e-01 1.07694435e+00 5.71397245e-01 5.00549264e-02
9.44629848e-01 -4.81557176e-02 1.43437517e+00 1.20448983e+00
3.04348320e-01 -9.95809495e-01 -9.90856662e-02 7.56184578e-01
5.42190135e-01 -6.79218352e-01 2.85437405e-01 2.86181033e-01
-6.41854405e-02 8.49868774e-01 2.35169962e-01 9.40541103e-02
6.96352720e-01 3.17813277e-01 -5.74140608e-01 -4.50226814e-01
-1.29928935e+00 1.21166699e-01 3.39901179e-01 1.16171755e-01
2.83048898e-01 -1.50070980e-01 -2.65850931e-01 1.05090782e-01
-4.96703535e-01 9.80809256e-02 -6.08301349e-02 1.05363071e+00
-4.10044163e-01 -1.23982465e+00 1.12665571e-01 7.74138570e-01
-6.13544166e-01 -7.86987782e-01 -2.27213115e-01 7.13932753e-01
-2.94581026e-01 5.52756310e-01 4.50673848e-01 -5.14625847e-01
4.55070548e-02 3.71292114e-01 7.20226526e-01 -5.46299040e-01
-5.62265396e-01 -3.81375819e-01 3.23160827e-01 -5.30203223e-01
2.02310279e-01 -2.21047863e-01 -9.21157897e-01 -7.81933069e-01
-5.89756191e-01 7.01634586e-01 7.00390995e-01 7.85846651e-01
1.10551536e+00 7.22023919e-02 3.78254533e-01 -1.89265594e-01
-1.26977706e+00 -1.15343821e+00 -3.53290975e-01 1.85133651e-01
2.78851360e-01 -3.22659403e-01 -5.46742976e-01 -2.43555292e-01] | [11.337923049926758, 7.691100597381592] |
09effb8a-083c-4190-9c78-a182dd00e937 | learning-to-incorporate-texture-saliency | 2208.01587 | null | https://arxiv.org/abs/2208.01587v2 | https://arxiv.org/pdf/2208.01587v2.pdf | Learning to Incorporate Texture Saliency Adaptive Attention to Image Cartoonization | Image cartoonization is recently dominated by generative adversarial networks (GANs) from the perspective of unsupervised image-to-image translation, in which an inherent challenge is to precisely capture and sufficiently transfer characteristic cartoon styles (e.g., clear edges, smooth color shading, abstract fine structures, etc.). Existing advanced models try to enhance cartoonization effect by learning to promote edges adversarially, introducing style transfer loss, or learning to align style from multiple representation space. This paper demonstrates that more distinct and vivid cartoonization effect could be easily achieved with only basic adversarial loss. Observing that cartoon style is more evident in cartoon-texture-salient local image regions, we build a region-level adversarial learning branch in parallel with the normal image-level one, which constrains adversarial learning on cartoon-texture-salient local patches for better perceiving and transferring cartoon texture features. To this end, a novel cartoon-texture-saliency-sampler (CTSS) module is proposed to dynamically sample cartoon-texture-salient patches from training data. With extensive experiments, we demonstrate that texture saliency adaptive attention in adversarial learning, as a missing ingredient of related methods in image cartoonization, is of significant importance in facilitating and enhancing image cartoon stylization, especially for high-resolution input pictures. | ['Yingjie Tian', 'Yuqi Zhang', 'Xiang Gao'] | 2022-08-02 | null | null | null | null | ['unsupervised-image-to-image-translation'] | ['computer-vision'] | [ 8.86813641e-01 2.46475250e-01 -2.58296747e-02 -1.09793998e-01
-5.92922091e-01 -6.93329990e-01 7.21951663e-01 -4.27912503e-01
6.02823123e-02 9.48500931e-01 1.31324798e-01 -2.92373598e-02
2.53165931e-01 -1.08719134e+00 -1.15427816e+00 -8.93741667e-01
3.36910129e-01 1.66866824e-01 2.08315402e-01 -6.25251889e-01
2.21800879e-01 6.51995778e-01 -1.19632185e+00 5.19233823e-01
1.08516419e+00 7.32048690e-01 3.56446445e-01 7.38752723e-01
-1.22303881e-01 9.30822849e-01 -3.63208294e-01 -5.56879044e-01
4.11456287e-01 -7.76414394e-01 -6.11964881e-01 1.59452766e-01
7.73997307e-01 -2.36326545e-01 -1.28581464e-01 1.31057942e+00
2.90596902e-01 -1.72376409e-01 8.10669601e-01 -1.32418191e+00
-1.19148493e+00 4.30967987e-01 -9.50380206e-01 -1.99253157e-01
3.05258334e-01 3.03301156e-01 7.39432156e-01 -8.05059671e-01
7.83686817e-01 1.62196040e+00 5.14725745e-01 6.69889450e-01
-1.56353581e+00 -8.08199227e-01 1.29473403e-01 -4.64943573e-02
-8.10876071e-01 -7.72733465e-02 1.33864677e+00 -2.07744926e-01
-9.68248956e-03 5.80649972e-01 6.46802545e-01 1.55316937e+00
5.98841846e-01 9.07788575e-01 1.61992908e+00 -2.75941432e-01
-1.00421757e-01 4.06315833e-01 -7.90133953e-01 5.86391032e-01
1.83052883e-01 3.21350366e-01 -2.42649242e-01 2.37448260e-01
1.40495038e+00 1.34725839e-01 -2.85039037e-01 -8.17446947e-01
-1.12741184e+00 7.95062542e-01 7.85735369e-01 1.17313825e-01
-2.80623436e-01 1.85480013e-01 2.62222201e-01 3.54222447e-01
4.22608584e-01 7.16172397e-01 -2.10233495e-01 2.29477540e-01
-8.14331412e-01 2.82262802e-01 2.86056042e-01 1.19270074e+00
9.84991074e-01 7.11778641e-01 -4.88647252e-01 8.86383176e-01
-1.90327168e-01 9.98720348e-01 4.08127040e-01 -6.94239140e-01
8.90131742e-02 5.09512901e-01 -3.90674956e-02 -1.06195450e+00
1.02516502e-01 -3.45632941e-01 -1.30723667e+00 7.88705885e-01
1.99047670e-01 -1.29594982e-01 -9.77219164e-01 1.71868742e+00
1.43500894e-01 4.79430221e-02 7.74990097e-02 9.96989846e-01
5.53135276e-01 4.28510219e-01 2.60345638e-01 1.75747424e-01
1.56631243e+00 -9.45360482e-01 -5.79357445e-01 -4.21734840e-01
-2.69038796e-01 -1.12593913e+00 1.49577463e+00 8.75986591e-02
-1.13710153e+00 -7.89105773e-01 -9.85538185e-01 -1.60834968e-01
-3.67662996e-01 -2.25284293e-01 6.07018352e-01 6.60189271e-01
-9.75474954e-01 5.56990862e-01 -2.24294618e-01 -2.01220900e-01
6.98141694e-01 -1.86191842e-01 -3.37481230e-01 -2.47471649e-02
-1.24906147e+00 6.70760214e-01 3.37125868e-01 -1.79919943e-01
-9.20416713e-01 -6.88191175e-01 -9.51022804e-01 -5.03251255e-02
2.34768853e-01 -1.00845754e+00 5.89070737e-01 -1.94158781e+00
-1.82161462e+00 9.31265116e-01 2.17054889e-01 -4.32324558e-01
7.73916960e-01 -1.86193883e-01 -3.74715388e-01 1.83821529e-01
1.45972103e-01 9.18264985e-01 1.88207316e+00 -1.80139863e+00
-5.18340468e-01 1.53723231e-03 1.66931510e-01 3.62692952e-01
-2.41097569e-01 -1.69984907e-01 -1.94359764e-01 -1.42003846e+00
-3.37002367e-01 -9.26477730e-01 -1.97012872e-01 1.63084269e-01
-6.45026386e-01 4.09610957e-01 1.04966521e+00 -6.59402847e-01
5.89985013e-01 -1.92998362e+00 3.81713748e-01 4.81393971e-02
2.43571743e-01 1.31023273e-01 -3.71609777e-01 3.07354808e-01
-4.54343170e-01 4.04593609e-02 -4.64048564e-01 -5.50935455e-02
-1.85553521e-01 1.03202753e-01 -6.02924526e-01 1.59776405e-01
8.81407797e-01 1.32469773e+00 -1.10680020e+00 -5.27770877e-01
4.11745787e-01 4.52795893e-01 -8.18681777e-01 3.50026935e-01
-2.62107581e-01 9.82270300e-01 -5.24796665e-01 7.01229692e-01
1.00989914e+00 9.21905637e-02 -2.46060714e-02 -3.37398022e-01
1.47747487e-01 -6.09954953e-01 -7.66647696e-01 1.48594880e+00
-7.70720661e-01 6.83101177e-01 6.62136003e-02 -9.24495995e-01
9.23940122e-01 -9.92982239e-02 2.86454082e-01 -8.56504023e-01
6.98894858e-02 -1.87966786e-03 -2.09474027e-01 -2.48993471e-01
6.36846244e-01 -4.36408430e-01 -2.51946211e-01 2.61885494e-01
-1.55469909e-01 -7.35800326e-01 -2.66915262e-01 1.87632158e-01
4.39691722e-01 3.08098555e-01 1.48526520e-01 -3.76015157e-01
6.45612180e-01 -2.11723298e-01 3.11676204e-01 5.39378524e-01
8.29978585e-02 1.06091189e+00 5.59517562e-01 -3.70016277e-01
-1.49438632e+00 -1.31136465e+00 -1.05125666e-01 1.06445348e+00
4.21338856e-01 3.17209274e-01 -9.79214847e-01 -7.26311922e-01
-1.31358132e-01 7.15264142e-01 -9.62337673e-01 -4.04682100e-01
-7.79268026e-01 -2.97158778e-01 3.00167978e-01 3.81636351e-01
6.85354829e-01 -1.66653562e+00 -3.61525118e-01 1.37331337e-01
-1.12931840e-01 -1.10523868e+00 -8.64329159e-01 6.08903095e-02
-6.48290873e-01 -8.89224291e-01 -1.36723626e+00 -1.04168546e+00
7.63836980e-01 2.03146651e-01 1.30452538e+00 -1.50087744e-01
-2.70142555e-01 2.59893596e-01 -4.80623871e-01 -4.90844727e-01
-7.22275138e-01 -1.56682983e-01 -1.82197437e-01 4.78113860e-01
-3.55561852e-01 -7.65420496e-01 -7.47453094e-01 2.13224873e-01
-1.37145805e+00 6.71700776e-01 9.66776311e-01 1.33289540e+00
4.24125552e-01 -1.46947250e-01 5.94819784e-01 -1.38545740e+00
4.88730282e-01 -2.41348043e-01 -3.57250810e-01 2.71452397e-01
-3.47161531e-01 5.42745851e-02 1.17052877e+00 -6.00241065e-01
-1.19326949e+00 -1.40387356e-01 -2.42994785e-01 -7.97451735e-01
-2.72080302e-01 -3.10813904e-01 -3.65244657e-01 -3.45032185e-01
6.27901852e-01 7.28023708e-01 8.41081589e-02 -4.69140848e-03
9.39496338e-01 5.00329249e-02 6.54353082e-01 -7.61983097e-01
1.55642080e+00 7.31785774e-01 -4.74672876e-02 -6.46227419e-01
-6.68212652e-01 2.57262588e-01 -6.16613626e-01 -1.20622374e-01
1.03940284e+00 -8.50044906e-01 -3.84712875e-01 6.67078555e-01
-8.86415720e-01 -6.57345653e-01 -6.16152406e-01 -2.78632283e-01
-9.50778246e-01 4.41854388e-01 -4.33159441e-01 -3.76618207e-01
-4.73586768e-01 -1.05635452e+00 1.41189718e+00 4.49742138e-01
7.51523301e-02 -1.04944146e+00 -2.20913708e-01 1.24200583e-01
6.51309013e-01 8.20680916e-01 1.01868069e+00 7.82932416e-02
-6.85494781e-01 1.38710469e-01 -6.52901113e-01 5.82489014e-01
2.36000940e-01 -2.24813178e-01 -9.11938071e-01 -4.88510460e-01
-6.88376203e-02 -3.18779737e-01 7.41656899e-01 2.86407858e-01
1.37047493e+00 -5.07196724e-01 5.64617738e-02 9.83737469e-01
1.58369076e+00 1.93168581e-01 9.39410329e-01 2.17622310e-01
1.11013258e+00 4.97040898e-01 5.36663771e-01 9.33045074e-02
-1.67771623e-01 6.57969475e-01 3.74668390e-01 -9.39397812e-01
-5.74943900e-01 -6.58241153e-01 3.58426392e-01 1.11199126e-01
-2.71829199e-02 -1.91583440e-01 -1.30382508e-01 3.83612812e-01
-1.39776540e+00 -9.88727748e-01 3.56511891e-01 1.70192456e+00
9.47095692e-01 1.99723974e-01 1.13414049e-01 -5.51876687e-02
8.79693985e-01 3.25102091e-01 -7.49421060e-01 -6.32982671e-01
-4.40975308e-01 4.54078496e-01 6.38174951e-01 3.87218863e-01
-1.03115392e+00 1.43365085e+00 4.90493107e+00 1.43137169e+00
-1.42535472e+00 1.57675873e-02 9.23182845e-01 4.05304909e-01
-6.73720777e-01 -1.06115066e-01 -1.51645511e-01 6.96550488e-01
2.66679656e-02 -1.71402059e-02 6.84507847e-01 8.99257004e-01
2.72619873e-02 2.47127891e-01 -7.87914395e-01 9.60960031e-01
7.95975253e-02 -1.30517137e+00 6.47943854e-01 -1.65251777e-01
1.15971553e+00 -6.12602293e-01 6.91677511e-01 2.40173236e-01
5.79046249e-01 -1.04503429e+00 9.69132125e-01 6.24195635e-01
1.51277018e+00 -8.33682716e-01 3.56155753e-01 -2.74260253e-01
-1.16500413e+00 1.26470141e-02 -3.69553149e-01 4.45367187e-01
1.58665508e-01 3.38711679e-01 -5.08595645e-01 8.19664538e-01
6.09298587e-01 8.14920306e-01 -6.83905363e-01 4.66976941e-01
-3.77842247e-01 2.82183945e-01 2.59179831e-01 3.85862701e-02
3.98820937e-01 -4.38446105e-01 7.83568621e-01 1.23159182e+00
1.21801794e-01 -1.96245134e-01 4.84442040e-02 1.17506361e+00
-9.56656262e-02 5.28364703e-02 -9.10546899e-01 2.06291974e-01
9.21332836e-02 1.34611487e+00 -7.79798746e-01 -4.82468545e-01
-1.66774079e-01 1.48985732e+00 2.07336888e-01 4.87111866e-01
-1.06402516e+00 -4.62894827e-01 5.83921015e-01 2.16824040e-01
2.69571453e-01 1.74756929e-01 -4.72883791e-01 -1.14616728e+00
-2.49154910e-01 -1.25433731e+00 -1.86052606e-01 -1.10938752e+00
-1.50530136e+00 8.57237458e-01 -4.31767732e-01 -1.50164258e+00
5.86995669e-02 -4.18283314e-01 -9.99361455e-01 9.20757055e-01
-1.64007151e+00 -1.66940904e+00 -4.32180882e-01 1.00931489e+00
7.99554706e-01 -3.14216793e-01 5.77912450e-01 4.58231047e-02
-2.86967665e-01 7.76757240e-01 1.25772595e-01 1.34483337e-01
9.36999977e-01 -1.61894464e+00 5.09117603e-01 8.34506273e-01
1.60644781e-02 3.37680072e-01 7.92836070e-01 -5.85626900e-01
-1.33309782e+00 -1.45331764e+00 1.45537350e-02 -6.50859773e-01
5.18877566e-01 -5.49591243e-01 -7.98170865e-01 3.68155003e-01
6.09512329e-01 -2.24750981e-01 -1.75838787e-02 -4.00758266e-01
-2.43052632e-01 -1.96559519e-01 -1.15672481e+00 1.12188160e+00
9.89309788e-01 -5.44749498e-01 -4.13666934e-01 2.70660669e-01
7.57917047e-01 -3.63629937e-01 -6.45195663e-01 3.82305503e-01
5.77247798e-01 -1.11300540e+00 1.38531017e+00 -4.69371736e-01
8.55582297e-01 -2.87458628e-01 1.99135765e-01 -1.41962874e+00
-4.74004567e-01 -8.50423872e-01 5.05120516e-01 1.19956553e+00
9.57519412e-02 -6.36393964e-01 6.64798856e-01 -5.05298339e-02
-1.27613515e-01 -5.68217278e-01 -4.60528433e-01 -5.24616838e-01
3.81546617e-01 8.81139114e-02 6.45128012e-01 1.07230377e+00
-6.79247439e-01 2.65314579e-01 -1.00514436e+00 -8.27269331e-02
7.04645753e-01 4.93520439e-01 1.09520423e+00 -6.86317861e-01
-2.11895898e-01 -6.00694060e-01 -2.29118302e-01 -8.21693778e-01
3.81665342e-02 -7.62736320e-01 -7.03341290e-02 -1.04577720e+00
1.13020204e-01 -4.94913459e-01 -3.34735274e-01 1.73439115e-01
-4.60321277e-01 7.88819909e-01 4.08013105e-01 -2.18821261e-02
-2.73574650e-01 8.39114249e-01 2.25696516e+00 -3.60314041e-01
2.09840000e-01 -1.30436823e-01 -9.65549827e-01 6.85277522e-01
5.88007271e-01 -2.36186147e-01 -3.38491201e-01 -4.62524593e-02
-2.25455001e-01 2.02726480e-02 6.02618456e-01 -9.15084481e-01
-3.14364165e-01 -2.98941165e-01 7.72026956e-01 -2.80972302e-01
2.93605864e-01 -8.26140583e-01 8.09512809e-02 4.97018248e-01
-2.05408812e-01 -5.66450097e-02 2.27069706e-01 7.60429800e-01
-2.95492351e-01 2.28206694e-01 1.33195722e+00 -1.99194357e-01
-9.33119416e-01 4.30418968e-01 -2.33931467e-02 2.93163836e-01
1.01457286e+00 -3.64969939e-01 -8.57678577e-02 -4.72998530e-01
-4.86956954e-01 -2.25403771e-01 8.69470656e-01 7.72349954e-01
7.29879975e-01 -1.50885928e+00 -8.94807279e-01 6.34468198e-01
1.27888814e-01 -9.52593610e-02 6.33432686e-01 4.59564388e-01
-7.14092314e-01 3.75466086e-02 -9.63350713e-01 -6.27844632e-01
-8.47691655e-01 9.22504783e-01 9.04279575e-02 -2.81222939e-01
-7.67615139e-01 6.87259495e-01 1.08951974e+00 -1.44801617e-01
-2.62406051e-01 -2.19704106e-01 -1.57328010e-01 -1.83521047e-01
1.83350772e-01 -1.44978821e-01 -4.57014829e-01 -4.80900079e-01
1.67409018e-01 1.02345240e+00 -1.07838430e-01 1.41198725e-01
1.06002152e+00 -1.58503607e-01 -1.50616290e-02 -3.91209126e-02
9.91089523e-01 4.51730847e-01 -1.63004053e+00 -1.82076693e-01
-5.44909596e-01 -6.05307460e-01 -3.29791039e-01 -7.29711890e-01
-1.24297202e+00 7.49720693e-01 7.08678067e-01 2.54298598e-01
1.40008140e+00 -2.31550187e-01 1.00499952e+00 -1.75735474e-01
4.44927871e-01 -6.93972290e-01 5.59536040e-01 2.11144656e-01
1.47676003e+00 -1.37786293e+00 -3.46934080e-01 -3.92504096e-01
-1.00960946e+00 9.14550245e-01 7.41532505e-01 -6.96348906e-01
2.88892776e-01 1.89532459e-01 9.72057283e-02 -4.32006046e-02
-2.21638918e-01 -1.02547616e-01 5.65485477e-01 8.78587008e-01
1.35696366e-01 2.33617172e-01 9.46272984e-02 2.90836513e-01
-3.08617443e-01 -5.01635611e-01 4.42451686e-01 3.66944075e-01
-1.09137952e-01 -1.09799421e+00 -5.12867689e-01 1.88485280e-01
-3.12188089e-01 -5.11209786e-01 -2.59855688e-01 9.10320580e-01
2.33963236e-01 4.86368269e-01 -1.17444815e-02 -1.94643661e-01
1.75972581e-01 -4.07478720e-01 5.11547089e-01 -4.79267895e-01
-3.94687325e-01 2.61411846e-01 -3.11198592e-01 -5.68385839e-01
-2.33272418e-01 -3.63205075e-01 -5.40825784e-01 -3.18592638e-01
4.85138521e-02 -1.55821785e-01 1.77385435e-01 4.71697718e-01
1.39728576e-01 1.02434051e+00 9.20809329e-01 -1.11743915e+00
2.92003471e-02 -9.27415252e-01 -6.98743939e-01 9.37254548e-01
2.75084376e-01 -6.97408676e-01 -1.40408322e-01 2.99649000e-01] | [11.636970520019531, -0.626628577709198] |
be450b03-b4db-4c36-a3af-c53c7c4420be | video-event-extraction-via-tracking-visual | 2211.01781 | null | https://arxiv.org/abs/2211.01781v2 | https://arxiv.org/pdf/2211.01781v2.pdf | Video Event Extraction via Tracking Visual States of Arguments | Video event extraction aims to detect salient events from a video and identify the arguments for each event as well as their semantic roles. Existing methods focus on capturing the overall visual scene of each frame, ignoring fine-grained argument-level information. Inspired by the definition of events as changes of states, we propose a novel framework to detect video events by tracking the changes in the visual states of all involved arguments, which are expected to provide the most informative evidence for the extraction of video events. In order to capture the visual state changes of arguments, we decompose them into changes in pixels within objects, displacements of objects, and interactions among multiple arguments. We further propose Object State Embedding, Object Motion-aware Embedding and Argument Interaction Embedding to encode and track these changes respectively. Experiments on various video event extraction tasks demonstrate significant improvements compared to state-of-the-art models. In particular, on verb classification, we achieve 3.49% absolute gains (19.53% relative gains) in F1@5 on Video Situation Recognition. | ['Heng Ji', 'Shih-Fu Chang', 'Jiajie Zhang', 'Xudong Lin', 'Manling Li', 'Guang Yang'] | 2022-11-03 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 5.47082782e-01 -7.37652257e-02 -3.52930099e-01 -2.70118624e-01
-5.13623059e-01 -5.57291269e-01 7.41722822e-01 4.80053157e-01
-4.98123318e-01 5.18315434e-01 7.58952200e-01 2.13772908e-01
2.59882748e-01 -5.24017215e-01 -1.01306653e+00 -7.66627967e-01
-2.83691287e-01 -4.44671303e-01 7.36038625e-01 1.27107222e-02
2.61651337e-01 4.18664783e-01 -1.61467946e+00 9.19845641e-01
1.44493669e-01 1.04197645e+00 1.93012580e-01 7.10926950e-01
-2.04154745e-01 1.62630546e+00 -7.09718585e-01 -3.17280442e-01
-4.75721002e-01 -3.53038490e-01 -7.81531751e-01 4.59822536e-01
3.07738960e-01 -4.81637031e-01 -7.04284370e-01 1.08180904e+00
1.74439028e-01 3.56767252e-02 4.70590442e-01 -1.35581851e+00
-3.86300772e-01 5.15591741e-01 -6.36076331e-01 1.02437520e+00
7.19700456e-01 1.79244149e-02 9.94287193e-01 -1.01900470e+00
1.09488606e+00 1.35382652e+00 1.97724476e-01 3.31671208e-01
-8.63528252e-01 -4.97935563e-01 6.67590201e-01 8.40179861e-01
-9.06732261e-01 -6.30250812e-01 1.17718124e+00 -4.80056405e-01
8.34844589e-01 3.64274025e-01 8.86629403e-01 1.30853641e+00
1.74758911e-01 1.46580589e+00 5.12435436e-01 -3.04523915e-01
2.49156728e-02 -2.43097380e-01 1.16213568e-01 7.76522100e-01
1.31988063e-01 -3.61791134e-01 -9.25189912e-01 1.78161841e-02
5.55649042e-01 2.28427991e-01 -2.97991782e-01 -3.45677167e-01
-1.57016373e+00 6.25572324e-01 3.66966039e-01 5.71959578e-02
-7.03146696e-01 4.36093837e-01 7.12127686e-01 -1.07951291e-01
3.88318568e-01 -2.49066025e-01 -3.65933537e-01 -3.05041492e-01
-4.18683410e-01 2.20201820e-01 2.52705067e-01 7.94853330e-01
3.92206132e-01 -6.35993034e-02 -6.48456514e-01 4.54641491e-01
3.28104675e-01 2.43985280e-01 2.64910251e-01 -7.74066508e-01
7.23507941e-01 9.93517816e-01 2.03007624e-01 -1.46475482e+00
-1.72370851e-01 1.52372057e-02 -3.57348621e-01 -2.11156875e-01
3.17765236e-01 -7.28169978e-02 -6.31505251e-01 1.67531288e+00
7.05224097e-01 6.02884054e-01 6.44007772e-02 7.56703436e-01
1.05563378e+00 1.00207496e+00 4.33468372e-01 -5.85754335e-01
1.84973133e+00 -7.88188219e-01 -1.06728947e+00 -3.79998982e-01
3.08305234e-01 -7.71766543e-01 7.51266420e-01 -2.30444949e-02
-1.27151573e+00 -6.28348589e-01 -8.58605027e-01 4.59507760e-03
-2.86361843e-01 1.07516930e-01 6.00617826e-01 9.98104289e-02
-3.87925446e-01 2.00059116e-01 -9.71950412e-01 -7.00343400e-02
5.96988916e-01 2.30247974e-02 -3.09382588e-01 2.89738141e-02
-1.26075804e+00 3.69541347e-01 5.40788829e-01 1.58992603e-01
-1.00778055e+00 -4.39622879e-01 -1.25056195e+00 3.79023626e-02
6.57216728e-01 -1.49937406e-01 1.11292982e+00 -1.08474672e+00
-8.46827745e-01 7.80335486e-01 -5.44166803e-01 -4.48743969e-01
2.34241351e-01 -2.93412000e-01 -5.06173849e-01 6.88398242e-01
1.02792889e-01 6.51276290e-01 9.88127172e-01 -1.01019728e+00
-1.07951868e+00 -2.30856270e-01 2.13077933e-01 3.11111659e-01
-5.44737220e-01 6.80922151e-01 -8.75443518e-01 -8.34737539e-01
4.64718752e-02 -6.86166763e-01 1.96092859e-01 1.49205744e-01
-1.80368349e-01 -4.05623287e-01 1.32201600e+00 -7.82051086e-01
1.40823770e+00 -2.48054862e+00 1.78332224e-01 -5.01642883e-01
3.12287033e-01 1.09074034e-01 1.68059971e-02 9.68160257e-02
-1.59679085e-01 -1.90146834e-01 1.69259757e-01 1.12403743e-02
-1.78191572e-01 1.04329519e-01 -4.28295642e-01 6.52316511e-01
5.12003958e-01 9.49631810e-01 -1.23282313e+00 -8.97350192e-01
3.87020767e-01 5.57737112e-01 -4.26905960e-01 1.19734183e-01
-2.40206689e-01 1.33275807e-01 -7.68257201e-01 7.51562119e-01
3.52435231e-01 -2.97982514e-01 2.16022417e-01 -6.69717669e-01
-4.07794900e-02 2.21508026e-01 -1.38427603e+00 1.32601023e+00
8.12091678e-02 9.75101888e-01 -1.39757186e-01 -1.00181174e+00
2.90931612e-01 5.26820540e-01 6.38085902e-01 -5.29550016e-01
1.70688480e-01 -3.10993850e-01 -1.96545824e-01 -8.24166059e-01
4.67003167e-01 4.01173592e-01 -3.76997501e-01 7.45786130e-02
-8.88912156e-02 7.08419919e-01 5.12430727e-01 3.77751946e-01
1.08868527e+00 1.32539421e-01 5.62898755e-01 8.59590322e-02
6.53108656e-01 -1.51329830e-01 7.14610934e-01 4.22411919e-01
-6.52476192e-01 3.57945442e-01 8.64506662e-01 -6.30993783e-01
-5.12594819e-01 -1.04861248e+00 3.57147515e-01 1.13582158e+00
7.03229308e-01 -5.81346571e-01 -6.92868769e-01 -9.62244928e-01
-2.10310251e-01 4.34359282e-01 -8.27650785e-01 -2.22731441e-01
-9.92267728e-01 -7.82920420e-01 1.54822424e-01 7.73389697e-01
6.30928457e-01 -1.47111213e+00 -1.15745306e+00 2.64724284e-01
-6.88680649e-01 -1.53708506e+00 -6.03525221e-01 -1.12236038e-01
-6.44421637e-01 -1.36509693e+00 -3.15289438e-01 -1.04513431e+00
5.55055737e-01 3.15087020e-01 9.71685529e-01 -4.37701773e-03
-4.11993057e-01 3.67714435e-01 -4.26362842e-01 -2.14776218e-01
-2.24369869e-01 -6.84208214e-01 -2.95063317e-01 4.93187159e-01
3.17190111e-01 -3.68138626e-02 -8.11383784e-01 1.49880618e-01
-8.82325828e-01 1.94503963e-01 3.90635073e-01 6.06286764e-01
7.24516988e-01 1.28319204e-01 1.90260425e-01 -6.00233138e-01
1.51461184e-01 -4.24019307e-01 -2.48829842e-01 3.14090282e-01
1.53544888e-01 -6.50914907e-02 3.19542050e-01 -6.52769685e-01
-1.31179881e+00 2.63365299e-01 3.83295000e-01 -4.04090732e-01
-3.46532196e-01 1.12179555e-01 -1.77293539e-01 4.68522817e-01
2.23836198e-01 3.89212459e-01 -6.26455963e-01 2.41057053e-02
2.98917711e-01 3.95475119e-01 5.41305304e-01 -3.82707775e-01
3.34939361e-01 9.33888376e-01 -1.69561684e-01 -8.58423233e-01
-9.23707724e-01 -6.68487608e-01 -5.37914991e-01 -6.43342078e-01
1.32157457e+00 -1.08644319e+00 -6.37214363e-01 3.26063663e-01
-1.39486492e+00 9.81396809e-02 -2.24158362e-01 3.69260311e-01
-5.49531877e-01 5.08705556e-01 -7.91597962e-01 -6.62618220e-01
-1.26057833e-01 -1.30948400e+00 1.35767448e+00 2.61810362e-01
-2.41413593e-01 -7.25528300e-01 -3.74994338e-01 2.62759864e-01
-2.76679575e-01 6.78384423e-01 7.43064165e-01 -2.27350935e-01
-6.65614069e-01 -1.96993738e-01 -3.43281120e-01 -6.29788265e-02
2.72991300e-01 1.77575558e-01 -7.29644775e-01 -5.85335754e-02
-1.27743194e-02 8.03999379e-02 7.80629694e-01 5.36446571e-01
1.10548127e+00 -3.29649121e-01 -7.70780444e-01 2.60465264e-01
1.07409620e+00 5.61049640e-01 6.59340024e-01 3.25224578e-01
7.69776344e-01 5.21723032e-01 9.43081498e-01 6.18492842e-01
2.73799628e-01 8.68274927e-01 6.78641260e-01 -5.34405336e-02
-4.11014050e-01 -8.50060433e-02 8.45386922e-01 3.12418193e-01
-4.39955443e-02 -3.04043770e-01 -5.46667755e-01 7.84483790e-01
-2.12339616e+00 -1.48029280e+00 -3.57358038e-01 1.71662760e+00
7.07550943e-01 3.93141061e-01 2.29511485e-01 2.33689964e-01
1.08239067e+00 6.20826006e-01 -4.82606292e-01 9.54402424e-03
-9.14075002e-02 -3.75748843e-01 2.25419313e-01 -1.67447515e-02
-1.57406795e+00 9.66891825e-01 5.71128130e+00 5.22109687e-01
-8.84914637e-01 1.19495228e-01 6.24043405e-01 -3.37738633e-01
1.55007213e-01 -1.06753230e-01 -8.01415205e-01 5.71790338e-01
6.08823121e-01 8.92730579e-02 -1.24938197e-01 6.09144628e-01
4.46249306e-01 -1.04504183e-01 -1.00433850e+00 9.75823462e-01
2.92102218e-01 -1.55493522e+00 1.82343140e-01 -3.26970756e-01
6.06091380e-01 -4.56460118e-01 -2.15025261e-01 7.69680291e-02
-1.36376172e-01 -3.22511047e-01 1.26503956e+00 3.15406114e-01
3.93573701e-01 -7.06635833e-01 4.36359107e-01 -2.88433526e-02
-1.68870163e+00 -2.24643618e-01 -2.28908937e-02 -1.35961980e-01
3.55087012e-01 2.73700178e-01 -5.54937065e-01 1.79879814e-01
9.26154256e-01 1.17350507e+00 -5.48126221e-01 5.40850639e-01
-4.76670295e-01 6.72791898e-01 2.49598064e-02 -1.88564330e-01
1.18380800e-01 3.15863669e-01 7.16534734e-01 1.47047603e+00
-1.57481059e-01 2.44472578e-01 1.25380367e-01 5.03862321e-01
-5.51032685e-02 -8.46967176e-02 -3.18098277e-01 1.36395711e-02
3.61582935e-01 1.22993422e+00 -1.11897588e+00 -6.78910613e-01
-7.19335139e-01 1.32612491e+00 1.64678246e-01 2.68235803e-01
-1.41042495e+00 -2.90253192e-01 9.19840693e-01 8.27906653e-02
7.06936240e-01 -5.80544434e-02 4.21664268e-01 -1.37577939e+00
3.51057261e-01 -7.17717290e-01 7.84499407e-01 -8.47450376e-01
-6.24366045e-01 3.66019189e-01 1.52347744e-01 -1.31130242e+00
-1.19886838e-01 -6.01694942e-01 -4.79050636e-01 1.20816797e-01
-1.58250451e+00 -1.01872540e+00 -5.41262209e-01 7.42273033e-01
1.19322133e+00 2.09459752e-01 3.20833057e-01 2.53102243e-01
-7.31373787e-01 2.51759917e-01 -2.88056016e-01 4.84211266e-01
3.94621670e-01 -1.01077604e+00 2.39017263e-01 1.31273949e+00
4.36102211e-01 3.33029956e-01 6.53864443e-01 -6.78944528e-01
-1.63760042e+00 -1.20871150e+00 8.35576594e-01 -5.17713666e-01
7.35245287e-01 -3.86814415e-01 -5.83563447e-01 7.25794375e-01
1.46631539e-01 3.58450264e-01 1.88137129e-01 -4.23300683e-01
-1.99644566e-01 8.78973957e-03 -8.15501451e-01 8.51035058e-01
1.19543684e+00 -5.28471828e-01 -8.79715323e-01 2.77274370e-01
7.56409228e-01 -4.83087510e-01 -4.90589082e-01 3.83984774e-01
5.16521990e-01 -7.99093187e-01 1.47267675e+00 -7.21041918e-01
5.93918264e-01 -4.81123030e-01 -2.61784524e-01 -5.19450963e-01
-3.32478493e-01 -4.05802071e-01 -7.20016599e-01 1.30780911e+00
-1.74471244e-01 6.86875284e-02 7.52536356e-01 2.22199887e-01
1.42722083e-02 -5.13478398e-01 -7.02058077e-01 -4.07996058e-01
-8.13472748e-01 -4.41786170e-01 2.53999859e-01 8.89564812e-01
1.08299732e-01 3.05876136e-01 -3.10064971e-01 4.21564400e-01
5.81315100e-01 3.03597599e-01 4.74968463e-01 -8.34418654e-01
-1.07188746e-01 -3.70848387e-01 -8.03108156e-01 -8.97651672e-01
2.64734387e-01 -4.81453866e-01 8.11024830e-02 -1.54463720e+00
6.81220353e-01 4.02529716e-01 -5.15225708e-01 3.82703036e-01
-6.25116765e-01 2.97953427e-01 2.56897211e-01 4.17195037e-02
-1.11824787e+00 3.57164443e-01 1.08643866e+00 -2.39559516e-01
-8.48430849e-04 -4.45586562e-01 -3.44807327e-01 1.09163368e+00
4.03463066e-01 -3.81985456e-01 -3.53689075e-01 -3.92972440e-01
6.45439476e-02 1.44521147e-01 6.33694947e-01 -9.31198716e-01
-1.92422401e-02 -4.16083485e-01 6.65789366e-01 -7.15726435e-01
4.95466560e-01 -8.42245340e-01 1.47050684e-02 5.78336179e-01
-4.10158962e-01 1.98649406e-01 3.12191755e-01 1.01680863e+00
-4.29217339e-01 -6.81861341e-02 5.19497156e-01 3.94477621e-02
-1.42126703e+00 1.50492653e-01 -5.14394641e-01 6.54101521e-02
1.50297821e+00 -1.62839308e-01 -3.95407319e-01 -2.67008573e-01
-8.34807515e-01 -5.20687476e-02 2.53110658e-02 6.74674332e-01
9.21079695e-01 -1.31031096e+00 -5.28146505e-01 -3.19487713e-02
2.75596112e-01 -3.10228795e-01 3.79936099e-01 7.25023329e-01
-3.52528602e-01 1.10028133e-01 -3.75975259e-02 -5.67899644e-01
-1.92619920e+00 7.12480664e-01 -2.33323909e-02 -9.88144875e-02
-7.73378253e-01 9.29379821e-01 5.12476742e-01 6.74322128e-01
2.75927722e-01 -6.38784707e-01 -5.77292979e-01 4.78460014e-01
8.91613364e-01 2.51895785e-01 -2.64350355e-01 -9.24210012e-01
-6.57402933e-01 4.71985489e-01 -2.61681490e-02 2.12021723e-01
1.28946805e+00 -1.72148198e-01 1.05572909e-01 4.31406438e-01
1.35194135e+00 9.05287415e-02 -1.77514184e+00 -1.99080318e-01
-5.45686930e-02 -4.36006755e-01 1.11549255e-02 -3.71193081e-01
-1.13140440e+00 7.47844696e-01 6.52223706e-01 1.54431120e-01
1.25332284e+00 4.57376748e-01 8.06768119e-01 4.00085784e-02
3.20533775e-02 -8.42960298e-01 5.26265681e-01 1.97934851e-01
7.19065011e-01 -1.24312484e+00 3.81916389e-02 -7.36249030e-01
-6.46400750e-01 1.13335299e+00 4.20686126e-01 -3.27806571e-03
3.86350632e-01 2.56786019e-01 -2.85041839e-01 -4.68208611e-01
-8.37434471e-01 -2.92416155e-01 4.45730329e-01 2.00340196e-01
3.81042987e-01 -1.05342299e-01 -1.59342423e-01 3.99336338e-01
5.67515492e-01 -1.87320367e-01 3.88819039e-01 1.14087951e+00
-5.14344275e-01 -5.11878014e-01 -3.36338192e-01 2.21815199e-01
-6.74366891e-01 4.66404334e-02 -2.41356149e-01 7.79615939e-01
2.28196800e-01 9.95208383e-01 4.04575676e-01 -1.32046133e-01
5.04754961e-01 1.15902107e-02 5.36707938e-01 -4.07898456e-01
-2.87213653e-01 1.71154514e-01 2.76012897e-01 -7.93072343e-01
-8.54123056e-01 -1.32680249e+00 -1.54278171e+00 1.76056311e-01
-1.29764482e-01 -2.11549908e-01 3.37400794e-01 9.22178924e-01
3.50860417e-01 9.29525495e-01 3.82187814e-01 -6.28455400e-01
6.95890188e-02 -5.14008462e-01 -1.11354150e-01 1.10241055e+00
4.38425392e-01 -8.51577342e-01 -3.24519008e-01 7.99851120e-01] | [8.469680786132812, 0.5917987823486328] |
6ff892f7-aa11-4ff3-9fea-98353c630810 | automatic-tracking-of-protein-vesicles | 1506.02083 | null | http://arxiv.org/abs/1506.02083v1 | http://arxiv.org/pdf/1506.02083v1.pdf | Automatic tracking of protein vesicles | With the advance of fluorescence imaging technologies, recently cell
biologists are able to record the movement of protein vesicles within a living
cell. Automatic tracking of the movements of these vesicles become key for
qualitative analysis of dynamics of theses vesicles. In this thesis, we
formulate such tracking problem as video object tracking problem, and design a
dynamic programming method for tracking single object. Our experiments on
simulation data show that the method can identify a track with high accuracy
which is robust to the choose of tracking parameters and presence of high level
noise. We then extend this method to the tracking multiple objects using the
track elimination strategy. In multiple object tracking, the above approach
often fails to correctly identify a track when two tracks cross. We solve this
problem by incorporating the Kalman filter into the dynamic programming
framework. Our experiments on simulated data show that the tracking accuracy is
significantly improved. | ['Min Xu'] | 2015-06-05 | null | null | null | null | ['video-object-tracking'] | ['computer-vision'] | [-9.49494988e-02 -7.66052425e-01 1.33117497e-01 1.79433361e-01
-6.31174862e-01 -8.91830444e-01 2.05095828e-01 1.43113390e-01
-7.18420565e-01 9.10310268e-01 -4.88612175e-01 7.05855340e-02
-9.97409150e-02 -4.09929931e-01 -7.77530730e-01 -1.16166615e+00
1.94096506e-01 5.53560972e-01 7.06110358e-01 3.19254160e-01
4.00840491e-01 7.51708210e-01 -1.29680324e+00 -2.04898342e-01
7.60020167e-02 6.05753541e-01 5.34194291e-01 1.33588099e+00
-4.19950694e-01 4.17254448e-01 -4.00293291e-01 1.99746504e-01
2.76768148e-01 -5.29052168e-02 -3.36367786e-01 1.77522123e-01
1.38750032e-01 -5.64827360e-02 -3.67244631e-02 1.25083482e+00
4.31842715e-01 -7.83062167e-03 2.55602568e-01 -1.31385028e+00
9.83409584e-03 3.71086895e-02 -7.23259628e-01 6.57616198e-01
4.02255923e-01 -6.90791905e-02 2.59575337e-01 -7.77930677e-01
9.20011997e-01 1.08159089e+00 9.23363030e-01 5.41357398e-01
-1.38137245e+00 -5.48104882e-01 1.65413156e-01 -2.85647810e-01
-1.50982141e+00 -4.08799380e-01 3.30560118e-01 -8.99352908e-01
4.11278188e-01 1.56003833e-01 8.26690555e-01 4.66754854e-01
4.87599373e-01 6.01292133e-01 1.12903976e+00 -3.15626085e-01
1.82849765e-02 -4.96592112e-02 5.27200758e-01 6.74352765e-01
3.74226689e-01 3.48741561e-01 -2.65701711e-01 -2.08066732e-01
9.54649985e-01 2.08848432e-01 -3.57691467e-01 -2.73172617e-01
-1.32878351e+00 4.52769876e-01 -4.24480617e-01 3.78058761e-01
-2.42429703e-01 4.75132316e-01 1.51636481e-01 7.84929320e-02
7.04696998e-02 -1.84342682e-01 -4.25731659e-01 -3.27158660e-01
-8.13768029e-01 5.66417217e-01 7.77665973e-01 9.18809175e-01
4.05164927e-01 -1.55560166e-01 9.18888226e-02 2.32294247e-01
5.75423837e-01 7.41661131e-01 2.87326306e-01 -9.63711560e-01
-4.73066373e-03 5.56638360e-01 7.08903790e-01 -7.36844659e-01
-4.39167917e-01 2.67723709e-01 -4.39348817e-01 6.87881291e-01
1.02996087e+00 -3.43201369e-01 -8.36709082e-01 1.46493864e+00
6.91574633e-01 4.09860015e-01 7.81652778e-02 4.69870299e-01
5.02638221e-01 8.24878037e-01 -6.62110820e-02 -9.22470391e-01
1.09383917e+00 -4.43887770e-01 -1.37073386e+00 5.16273499e-01
4.12947536e-01 -8.50600123e-01 1.99069664e-01 3.52148384e-01
-1.18862557e+00 -4.18557703e-01 -6.53162658e-01 4.74724948e-01
-3.12795877e-01 6.20094426e-02 2.54118919e-01 4.94078338e-01
-8.87841463e-01 6.49860322e-01 -1.36609626e+00 -4.84400004e-01
2.41183802e-01 7.72544205e-01 -1.95760682e-01 5.14764786e-01
-3.95734668e-01 8.06520164e-01 4.20680046e-01 2.13012606e-01
-7.33608842e-01 -6.89468861e-01 -3.54167521e-01 -2.62380004e-01
7.68303201e-02 -4.08612996e-01 1.19437790e+00 -6.35401666e-01
-1.47009957e+00 7.23459482e-01 -6.73401654e-01 -2.24495560e-01
7.10855186e-01 -3.61141898e-02 -3.39499623e-01 2.39963420e-02
-3.69057134e-02 2.90045559e-01 6.06065571e-01 -1.27855825e+00
-8.81966829e-01 -3.02365959e-01 -3.73616219e-01 -1.40283421e-01
1.48616880e-01 3.58192831e-01 -4.23579007e-01 -1.99938893e-01
9.30246487e-02 -1.11727667e+00 -3.40289295e-01 4.27114636e-01
1.50957322e-02 2.58122943e-02 1.37304127e+00 -1.08639032e-01
1.09577668e+00 -2.04259515e+00 1.99564099e-01 6.48167683e-03
1.72608942e-01 3.70965183e-01 3.75361383e-01 2.54685909e-01
1.70241907e-01 -8.80049318e-02 2.10344106e-01 -1.69504821e-01
-4.48890150e-01 2.89707780e-01 -1.75033882e-01 7.35151172e-01
-3.20826352e-01 6.31401360e-01 -8.52488518e-01 -5.97716153e-01
2.95679301e-01 4.50133502e-01 -5.25514362e-03 -3.44070420e-02
1.03758080e-02 5.89483321e-01 -6.89460516e-01 6.27180278e-01
7.74192631e-01 -4.99265820e-01 -4.81512621e-02 -1.56580374e-01
-6.98017359e-01 -6.34776831e-01 -1.83340585e+00 1.08858645e+00
3.23856264e-01 7.26792753e-01 4.96328026e-01 -6.16023064e-01
8.14301968e-01 5.46726465e-01 1.07031310e+00 -2.46575754e-02
2.46765971e-01 9.15621892e-02 -4.23835814e-02 -5.52750468e-01
2.25457177e-01 -5.13183534e-01 4.78747070e-01 3.46487135e-01
-3.59260976e-01 2.46380523e-01 3.28600526e-01 6.74729273e-02
1.07322872e+00 4.39774133e-02 1.11456849e-01 -4.81011391e-01
7.76931226e-01 2.83759892e-01 7.57912874e-01 6.74090564e-01
-4.94729608e-01 3.70123893e-01 -1.23197390e-02 -5.56686878e-01
-1.01129723e+00 -9.47078049e-01 -2.99244255e-01 3.61584127e-01
5.33911526e-01 -1.76105052e-01 -5.33404529e-01 1.33679714e-03
3.50337923e-01 -3.02474886e-01 -5.17163157e-01 5.36898434e-01
-6.21042430e-01 -1.09823298e+00 3.56607199e-01 5.43259859e-01
-3.64364795e-02 -7.67789602e-01 -7.47867286e-01 6.82283998e-01
1.91156790e-01 -1.28062832e+00 -1.94308206e-01 1.08497888e-01
-8.72881830e-01 -1.35881400e+00 -7.62138247e-01 -7.71145880e-01
7.43390858e-01 3.32033455e-01 3.26866090e-01 3.91047269e-01
-4.38875794e-01 5.96590996e-01 3.87643115e-03 -5.27294695e-01
-5.67660868e-01 -6.49363697e-01 3.59869003e-01 -1.62844181e-01
1.56226397e-01 -1.08085655e-01 -1.72362298e-01 5.47914326e-01
-6.68475091e-01 -2.76203871e-01 -1.86316326e-01 6.24613404e-01
1.10040700e+00 4.28005457e-01 1.04278497e-01 -4.21874225e-01
3.44155401e-01 -7.06099495e-02 -1.50211954e+00 1.23744942e-01
-2.62463331e-01 -2.34639004e-01 3.36914510e-01 -9.22683179e-01
-7.53201485e-01 6.14089847e-01 9.69216824e-02 -5.96147776e-01
5.90930767e-02 1.48314536e-01 8.27774704e-02 -7.92461872e-01
5.83662018e-02 3.79881293e-01 4.86081213e-01 -3.40704799e-01
-1.62877038e-01 1.67499185e-01 3.88502538e-01 -4.04162675e-01
7.70066321e-01 1.02293921e+00 5.04253566e-01 -8.76269162e-01
-3.78684938e-01 -9.91476595e-01 -7.00961232e-01 -6.09546483e-01
7.98071980e-01 -8.07249367e-01 -1.45965338e+00 7.29458272e-01
-1.19153726e+00 -3.09471220e-01 4.73151952e-02 5.71536660e-01
-6.30232036e-01 5.69283724e-01 -6.52917683e-01 -1.04894423e+00
-5.38699701e-03 -1.40502310e+00 9.07550216e-01 5.77761412e-01
5.62428981e-02 -1.31760812e+00 6.21116102e-01 -2.17806071e-01
1.34300277e-01 3.96998137e-01 1.68673351e-01 -4.17087913e-01
-9.78852928e-01 -3.91287684e-01 5.34923002e-02 -2.57521003e-01
1.13226160e-01 7.45241880e-01 -5.88392138e-01 -5.48316300e-01
-2.36452278e-02 4.29324150e-01 5.09278238e-01 9.76307392e-01
7.21715569e-01 1.76177770e-01 -1.18298197e+00 5.00825047e-01
1.85968447e+00 7.58800268e-01 3.70679349e-01 3.61235052e-01
5.94932735e-01 2.61392385e-01 7.15806067e-01 3.60549599e-01
-7.63267949e-02 7.42191792e-01 1.83788329e-01 2.02980399e-01
2.29987070e-01 4.12120968e-01 2.74618447e-01 6.80293798e-01
-4.21646126e-02 -4.29946125e-01 -7.99625158e-01 5.66134751e-01
-2.04805183e+00 -1.24255121e+00 -7.83870280e-01 2.15679908e+00
5.33686399e-01 8.57107416e-02 2.71987051e-01 6.45971447e-02
1.09132469e+00 -4.91109163e-01 -4.07453388e-01 4.32102792e-02
-3.82129033e-03 -3.13687503e-01 8.49188387e-01 8.89024973e-01
-1.04006100e+00 7.05890894e-01 7.85055017e+00 4.20296162e-01
-8.79650772e-01 -1.17495842e-01 -1.70503274e-01 6.34461418e-02
1.87664777e-01 6.50576949e-02 -1.84501302e+00 7.45722532e-01
7.83629298e-01 -4.25946444e-01 1.53514996e-01 2.15170652e-01
5.30446470e-01 -4.11763012e-01 -7.89453745e-01 1.04087901e+00
-3.57899725e-01 -1.57300186e+00 -2.86959559e-01 2.52600014e-01
5.44637680e-01 -2.03508943e-01 -2.62937605e-01 -2.93534487e-01
4.47955847e-01 -5.38915277e-01 3.60379845e-01 9.63741183e-01
-5.97244129e-04 -4.16133463e-01 5.08745432e-01 5.89817524e-01
-1.48289669e+00 -5.37373498e-02 -3.92263085e-01 1.15163755e-02
8.15717816e-01 3.72681171e-01 -6.39365494e-01 1.56152934e-01
5.58031678e-01 5.61306119e-01 -2.44083442e-02 1.75663519e+00
5.24217188e-01 1.12718754e-01 -8.95775378e-01 -1.43965513e-01
-5.95422424e-02 -3.62851769e-01 9.82387066e-01 1.02991247e+00
4.13936526e-01 4.80199754e-01 2.32275039e-01 6.37114108e-01
4.33282822e-01 -1.24810621e-01 -5.43853760e-01 -2.04679579e-01
3.47375721e-01 1.32375133e+00 -1.24762070e+00 -2.85894513e-01
-5.16474724e-01 4.88597184e-01 -1.99956018e-02 1.42275035e-01
-7.45959759e-01 -1.74672063e-02 7.18911946e-01 2.49107718e-01
5.80572605e-01 -3.45023304e-01 1.24530442e-01 -8.67210090e-01
-2.56738216e-01 -3.11316490e-01 3.00293807e-02 -8.13112676e-01
-8.65963042e-01 1.14502035e-01 3.42727676e-02 -1.15505588e+00
1.99686483e-01 -7.73091018e-01 -5.85712910e-01 5.90748429e-01
-1.21147692e+00 -8.70175004e-01 -1.05348215e-01 5.47925413e-01
3.29826802e-01 5.49135990e-02 4.43250418e-01 4.48154449e-01
-6.55108273e-01 -8.35155174e-02 6.69174552e-01 1.00229971e-01
4.18031305e-01 -1.16326439e+00 -4.44015227e-02 8.66498530e-01
-1.10768788e-01 6.08672738e-01 1.07966697e+00 -9.90644991e-01
-1.46246064e+00 -7.51307547e-01 6.27667844e-01 -6.46447659e-01
9.32647884e-01 -6.76044077e-02 -1.05621600e+00 8.34427774e-01
-1.76229268e-01 3.63822639e-01 5.11497021e-01 -4.19664085e-01
5.87501645e-01 8.26043934e-02 -1.03373492e+00 2.26845413e-01
5.98667383e-01 -1.01946473e-01 -2.48009041e-01 4.37227130e-01
1.78972378e-01 -7.51185536e-01 -1.09396088e+00 2.27634877e-01
8.33352685e-01 -5.13296247e-01 9.48894083e-01 -5.56142390e-01
-6.01313591e-01 -1.00561798e+00 -2.24311417e-03 -7.34234095e-01
-1.28564268e-01 -9.01838422e-01 -1.28023639e-01 1.22787118e+00
3.77793536e-02 -3.57732087e-01 1.04332185e+00 4.11503643e-01
1.85648099e-01 -2.77787447e-01 -1.00956249e+00 -1.04702115e+00
-9.05927345e-02 2.01670408e-01 9.11201611e-02 7.25605309e-01
-1.21202730e-01 -2.58570433e-01 -1.75200120e-01 5.32424271e-01
1.01230752e+00 1.34497344e-01 8.51180851e-01 -1.60743010e+00
-3.52144875e-02 -1.92883879e-01 -7.15365052e-01 -1.02746248e+00
8.20343196e-02 -2.82746762e-01 2.61772543e-01 -1.25785124e+00
4.80587691e-01 -2.54309744e-01 -1.31747156e-01 -1.48275085e-02
-1.68351263e-01 -5.32331020e-02 1.83280006e-01 2.70502061e-01
-6.21112168e-01 -8.94417390e-02 1.29287863e+00 1.55848965e-01
-1.67613640e-01 3.02751333e-01 7.20186904e-02 9.95302022e-01
4.93088454e-01 -8.69886339e-01 1.55797198e-01 -1.94517925e-01
1.18370481e-01 3.58221769e-01 3.72547776e-01 -1.01960897e+00
8.37400913e-01 -7.12235793e-02 3.75068575e-01 -1.11041021e+00
5.11538267e-01 -1.18814659e+00 7.99841225e-01 6.80310965e-01
1.05115473e-01 4.72143948e-01 5.86932600e-01 9.83311534e-01
-9.69931409e-02 -4.36865270e-01 1.07578659e+00 -2.46316612e-01
-6.64280355e-01 4.33651119e-01 -8.16136599e-01 -2.49358863e-01
1.42044663e+00 -7.01485515e-01 -1.88921183e-01 1.36052014e-03
-1.30064976e+00 3.77236843e-01 7.26900816e-01 -1.07471019e-01
3.72581422e-01 -1.12471354e+00 -9.62381214e-02 1.28056845e-02
-3.06525767e-01 -2.27308109e-01 -9.68813375e-02 1.10974300e+00
-7.00049758e-01 2.72253901e-01 -2.71335930e-01 -9.89739478e-01
-2.08268833e+00 6.13846362e-01 4.52759206e-01 -1.24123633e-01
-5.23571849e-01 6.73640430e-01 -1.29060447e-01 9.90108252e-02
1.25545830e-01 -3.83910418e-01 -3.39072108e-01 -1.56404115e-02
7.20970392e-01 5.88123202e-01 -4.68025833e-01 -8.21934402e-01
-3.28707606e-01 1.34423232e+00 -1.53897509e-01 1.11387866e-02
1.09487569e+00 -5.41384518e-01 -9.79987998e-03 7.10898817e-01
8.73818099e-01 -9.78625286e-03 -1.42359388e+00 1.85858086e-01
6.19071210e-03 -4.82676148e-01 -1.69602465e-02 -2.54866779e-01
-9.06833053e-01 6.33580625e-01 8.61879051e-01 3.31999600e-01
6.57742679e-01 -3.41538668e-01 4.94059533e-01 3.39303821e-01
3.38456690e-01 -7.07979739e-01 -2.02509448e-01 4.36335295e-01
7.58524463e-02 -1.15982723e+00 2.45749146e-01 -5.52599788e-01
-1.72646359e-01 1.31189883e+00 4.44695145e-01 -1.90077662e-01
9.52284873e-01 1.06043077e+00 2.12956369e-01 -2.87412345e-01
-7.78002024e-01 -1.83978438e-01 -4.19465423e-01 4.75127995e-01
2.26340201e-02 -2.65872627e-01 -2.53511727e-01 -6.68267235e-02
6.34329915e-01 4.23082858e-01 7.47021556e-01 1.17656481e+00
-9.12246525e-01 -1.13368046e+00 -6.47226930e-01 3.00709587e-02
-8.69060934e-01 4.61993903e-01 3.79909724e-02 1.08028221e+00
-1.26539290e-01 5.33374131e-01 2.08894059e-01 1.86918885e-01
3.28167379e-01 4.92143445e-02 7.66696513e-01 -1.03072964e-01
-4.64388818e-01 6.02000594e-01 -4.77213591e-01 -3.05093437e-01
-1.02987635e+00 -1.10473382e+00 -1.57249069e+00 -2.90407151e-01
-8.69429290e-01 2.65465230e-01 6.34312332e-01 9.37924445e-01
5.11413477e-02 3.46158028e-01 1.61222100e-01 -5.89702487e-01
-2.17090338e-01 -2.64068335e-01 -8.77520382e-01 1.70317218e-01
6.98618472e-01 -9.93882775e-01 -4.81888443e-01 4.93740767e-01] | [6.714853286743164, -1.9927723407745361] |
219e7334-db59-4078-ac11-5a3e58cc0e99 | caila-concept-aware-intra-layer-adapters-for | 2305.16681 | null | https://arxiv.org/abs/2305.16681v1 | https://arxiv.org/pdf/2305.16681v1.pdf | CAILA: Concept-Aware Intra-Layer Adapters for Compositional Zero-Shot Learning | Compositionality, the ability to combine existing concepts and generalize towards novel compositions, is a key functionality for intelligent entities. Here, we study the problem of Compositional Zero-Shot Learning (CZSL), which aims at recognizing novel attribute-object compositions. Recent approaches build their systems on top of large-scale Vision-Language Pre-trained (VLP) models, e.g. CLIP, and observe significant improvements. However, these methods treat CLIP as a black box and focus on pre- and post-CLIP operations. Here, we propose to dive deep into the architecture and insert adapters, a parameter-efficient technique proven to be effective among large language models, to each CLIP encoder layer. We further equip adapters with concept awareness so that concept-specific features of "object", "attribute" and "composition" can be extracted. We name our method CAILA, Concept-Aware Intra-Layer Adapters. Quantitative evaluations performed on three popular CZSL datasets, MIT-States, C-GQA, and UT-Zappos, reveal that CAILA achieves double-digit relative improvements against the current state-of-the-art on all benchmarks. | ['Ram Nevatia', 'Haidong Zhu', 'Zhaoheng Zheng'] | 2023-05-26 | null | null | null | null | ['compositional-zero-shot-learning'] | ['computer-vision'] | [ 2.70816863e-01 1.15083456e-01 -2.07394913e-01 -3.58934999e-01
-8.25258136e-01 -6.06182098e-01 8.60242248e-01 2.18253627e-01
-3.03075105e-01 4.61308599e-01 4.19736892e-01 2.05419436e-02
3.28621149e-01 -6.59073412e-01 -1.09137321e+00 -3.46199691e-01
1.27916172e-01 5.37358463e-01 4.79422092e-01 -2.17159450e-01
-9.99034122e-02 -1.14778979e-02 -2.05587196e+00 9.26608145e-01
8.99802268e-01 1.09565425e+00 1.62108287e-01 6.10396385e-01
-5.69916427e-01 1.10495758e+00 -6.72052428e-02 -9.51933682e-01
9.63511541e-02 -3.93236578e-02 -9.14406836e-01 -3.66424806e-02
8.93333316e-01 -1.76860243e-01 -1.76777229e-01 1.03508961e+00
3.32794636e-01 1.63704172e-01 6.85456216e-01 -1.36205137e+00
-1.16246629e+00 1.30995274e+00 -1.41843542e-01 -1.67369798e-01
3.63520592e-01 3.43675703e-01 1.64978278e+00 -1.38422346e+00
1.07495189e+00 1.26097631e+00 7.17302918e-01 8.42918158e-01
-1.32159603e+00 -8.62639248e-01 5.72732866e-01 6.84065282e-01
-1.45633209e+00 -6.96087420e-01 5.80469429e-01 -5.09878159e-01
1.39901054e+00 1.77141801e-01 5.40551543e-01 1.30297029e+00
-3.45744461e-01 1.36216044e+00 5.32000542e-01 -3.88806641e-01
3.28953356e-01 1.93106115e-01 1.04048863e-01 8.28190565e-01
2.35488087e-01 -2.33399928e-01 -7.59693623e-01 7.30770752e-02
-2.62308121e-02 5.16141728e-02 -1.56773999e-01 -6.28825903e-01
-1.15630949e+00 4.93334413e-01 5.45309007e-01 3.01057190e-01
-3.99043709e-01 2.40227982e-01 5.33658147e-01 1.54642195e-01
2.06619740e-01 7.59151995e-01 -6.36259317e-01 -1.93274468e-02
-9.20223832e-01 4.05121803e-01 7.47038007e-01 1.73685932e+00
7.14968503e-01 -2.34635144e-01 -4.69696075e-01 6.67092860e-01
2.50120442e-02 2.21268415e-01 5.96754253e-01 -7.71624446e-01
4.88405854e-01 9.68433499e-01 -2.33029902e-01 -2.34190419e-01
-3.21475565e-01 -5.31869769e-01 -5.60550034e-01 -3.93079221e-01
-3.85964543e-01 7.49725252e-02 -9.92492020e-01 1.75194311e+00
2.87232518e-01 5.76526999e-01 3.32251161e-01 5.86060524e-01
1.09282053e+00 6.08710408e-01 4.20454770e-01 6.86141476e-02
1.38511133e+00 -1.57782042e+00 -4.29285616e-01 -1.36101604e-01
6.24844193e-01 -2.85613149e-01 1.45764124e+00 2.71168441e-01
-1.28992224e+00 -6.81197107e-01 -1.11800921e+00 -5.64241409e-01
-9.42264736e-01 2.77300738e-02 8.32515299e-01 2.72107929e-01
-7.89409161e-01 4.76917565e-01 -6.11078501e-01 -3.79632413e-01
6.37372673e-01 7.81519413e-02 -4.72605973e-01 -1.66538402e-01
-1.14158356e+00 7.11911678e-01 6.39359117e-01 -5.42516768e-01
-1.11041212e+00 -1.39071178e+00 -1.05572081e+00 5.28653502e-01
7.82351911e-01 -1.00287354e+00 1.53996813e+00 -8.21013272e-01
-1.18128872e+00 7.19920278e-01 -2.87037134e-01 -8.34061682e-01
8.48638937e-02 -2.12296590e-01 -5.51694334e-01 2.81238794e-01
4.77369837e-02 1.08224308e+00 7.71684468e-01 -1.37958694e+00
-1.13784957e+00 -1.42833769e-01 3.14891458e-01 -6.61539799e-03
-5.47823250e-01 -2.84814741e-02 -6.14771724e-01 -3.71442229e-01
-1.94299683e-01 -7.30912864e-01 1.79088667e-01 2.33848304e-01
-2.43929699e-01 -5.10312438e-01 5.34809768e-01 -4.45341408e-01
1.21792591e+00 -2.06025720e+00 2.85227329e-01 -2.70280421e-01
4.17598695e-01 4.59660023e-01 -2.11633608e-01 4.97650921e-01
5.51923998e-02 -2.79893056e-02 -3.21910620e-01 -7.61554301e-01
4.93931651e-01 1.09892637e-01 -5.40493906e-01 -8.10915157e-02
4.87762719e-01 1.42486668e+00 -1.15013683e+00 -4.42204863e-01
8.68064016e-02 2.99784333e-01 -8.24884713e-01 1.22562610e-01
-8.17415178e-01 -2.45441794e-01 1.46012798e-01 1.00243056e+00
4.33789134e-01 -4.62159693e-01 1.32956982e-01 -4.63428676e-01
-2.41567753e-02 3.28208894e-01 -8.81772816e-01 2.31168580e+00
-5.96632838e-01 4.98863697e-01 -4.28806216e-01 -7.08502591e-01
6.23219669e-01 3.01489025e-01 1.77225456e-01 -6.56298876e-01
-2.70080250e-02 9.68664885e-02 -1.88762337e-01 -4.99589741e-01
8.56051087e-01 -8.38223659e-03 -1.29754081e-01 4.62722741e-02
8.69308233e-01 5.47809452e-02 4.98493314e-01 6.14203274e-01
1.15141690e+00 2.49130279e-01 4.51376289e-01 -9.39185247e-02
6.72775328e-01 -1.78755432e-01 5.01623929e-01 6.23885930e-01
-2.71882564e-02 4.89827782e-01 3.65803510e-01 -3.03537637e-01
-1.02350843e+00 -1.28047609e+00 4.09351259e-01 1.55872178e+00
1.73588336e-01 -9.34795439e-01 -4.55154032e-01 -8.55685651e-01
1.79220676e-01 1.12211192e+00 -7.09006548e-01 -4.01319414e-01
-3.42362404e-01 -2.15445057e-01 5.74609816e-01 8.75575483e-01
4.18534875e-01 -9.07051682e-01 -5.38542926e-01 3.45740587e-01
-4.51877825e-02 -1.44191754e+00 -4.76780504e-01 -3.25052515e-02
-4.42707509e-01 -8.92815828e-01 -4.91535664e-01 -9.35607791e-01
5.16427100e-01 2.53294766e-01 1.49723125e+00 -2.25747883e-01
-3.49730879e-01 4.58304882e-01 -6.58932865e-01 -5.60370326e-01
-9.96842235e-02 3.66021961e-01 -8.46388713e-02 2.53186196e-01
6.34970248e-01 -7.05158114e-01 -4.25371528e-01 -2.63184041e-01
-8.01265180e-01 4.68951613e-01 8.52448821e-01 8.36240470e-01
6.01978540e-01 -3.39227021e-01 5.05386829e-01 -1.14701617e+00
3.24515194e-01 -4.42380786e-01 -2.58468807e-01 7.52846658e-01
-6.22663677e-01 1.96647078e-01 6.90292597e-01 -5.76387584e-01
-1.08711517e+00 1.29505247e-01 7.17372149e-02 -8.59993517e-01
-1.74520854e-02 4.59981471e-01 -5.64912140e-01 1.99993402e-01
5.96401811e-01 5.00285149e-01 -5.50701976e-01 -5.16621888e-01
1.09389222e+00 7.11741507e-01 1.12044549e+00 -8.59340370e-01
5.72911859e-01 5.86043715e-01 -3.03530604e-01 -5.98368883e-01
-9.84461010e-01 -8.71600986e-01 -5.38408458e-01 1.21930979e-01
6.92441285e-01 -1.40252709e+00 -7.13994801e-01 1.94018438e-01
-1.10922384e+00 -1.19727023e-01 -6.53196692e-01 -5.09572960e-02
-4.94597614e-01 8.22800305e-03 -4.44332153e-01 -6.18758082e-01
-8.41324747e-01 -8.62190306e-01 1.23077106e+00 4.04673517e-01
-1.68529138e-01 -7.18591213e-01 -1.90777630e-01 2.12679669e-01
5.14839768e-01 -1.20927073e-01 1.10334611e+00 -8.04844677e-01
-7.67482996e-01 8.77602175e-02 -4.29010451e-01 3.00867051e-01
-2.41636574e-01 -1.50385529e-01 -1.15154862e+00 -1.93170369e-01
-5.31800032e-01 -5.46248019e-01 1.16801870e+00 -3.69223565e-01
1.34718311e+00 -4.62980688e-01 -4.04844075e-01 7.73832262e-01
1.52215958e+00 -1.43263415e-01 5.16422689e-01 -5.03837988e-02
8.29690039e-01 3.95759106e-01 3.58244419e-01 4.14011270e-01
1.08199966e+00 6.10305846e-01 3.14776868e-01 3.12262565e-01
-5.34736276e-01 -7.91522264e-01 4.08174187e-01 8.59487414e-01
7.25725070e-02 -1.60563588e-01 -7.61769891e-01 1.01627839e+00
-1.98502505e+00 -9.74407494e-01 3.03646803e-01 1.63188899e+00
1.01530123e+00 1.34926289e-01 -8.77022222e-02 -2.18583375e-01
2.44863778e-01 3.67130078e-02 -6.56785011e-01 -9.07602087e-02
-1.55543268e-01 5.10884941e-01 3.27810824e-01 1.96835503e-01
-1.13277984e+00 1.35168266e+00 5.02996445e+00 8.64461362e-01
-9.48800385e-01 3.55008096e-01 -5.73287308e-02 -4.11300480e-01
-6.06818914e-01 1.53667465e-01 -1.11978471e+00 4.39311624e-01
7.99061179e-01 -3.71315747e-01 4.32045341e-01 9.67155755e-01
-6.48113906e-01 2.31781319e-01 -1.62942648e+00 1.00978994e+00
6.25424504e-01 -1.68263817e+00 6.58722579e-01 -5.43843031e-01
9.81334269e-01 1.14595838e-01 1.62291810e-01 1.06346297e+00
4.30852056e-01 -6.98724687e-01 8.98854792e-01 6.71597481e-01
9.50442433e-01 -5.07559121e-01 5.49323380e-01 4.41794656e-02
-1.56870329e+00 -3.17150831e-01 -3.93188506e-01 1.41793355e-01
2.03058392e-01 1.94534078e-01 -8.03009510e-01 6.41356885e-01
6.64253592e-01 1.00834131e+00 -7.87699521e-01 9.18497860e-01
-4.16286647e-01 3.90369713e-01 6.51164800e-02 -1.00014679e-01
2.90893167e-01 4.70685005e-01 3.31911832e-01 1.48347604e+00
1.74735412e-01 -1.42992800e-03 3.01262178e-02 1.23317409e+00
-6.60206735e-01 -1.28289625e-01 -1.53347000e-01 -3.78177464e-01
6.10570729e-01 1.25837874e+00 -1.95499539e-01 -1.09831536e+00
-8.40377748e-01 1.03633130e+00 5.26959717e-01 1.72076210e-01
-8.31320524e-01 -4.76815253e-01 9.10649180e-01 -1.32419392e-01
8.63881409e-01 6.01702891e-02 -1.38683766e-01 -1.59820020e+00
6.85994402e-02 -9.39672410e-01 4.33285087e-01 -8.69327605e-01
-1.43130922e+00 3.22979987e-01 -7.40028918e-02 -1.10637403e+00
4.81730560e-03 -5.18261194e-01 -4.66450304e-01 4.86330211e-01
-1.56520808e+00 -1.90753531e+00 -4.17471766e-01 6.49747312e-01
1.04330945e+00 -4.50250506e-01 9.33797598e-01 4.93287116e-01
-1.81880176e-01 9.29801047e-01 -1.77957416e-01 -1.03887310e-02
6.31685019e-01 -1.49294376e+00 8.51969123e-01 9.60482597e-01
5.62238753e-01 8.07977498e-01 6.39966965e-01 -5.43877125e-01
-1.52130616e+00 -1.37475657e+00 1.27420294e+00 -5.48991919e-01
8.06956470e-01 -8.36792767e-01 -9.29699779e-01 9.02657807e-01
2.64743149e-01 1.98984042e-01 6.99537039e-01 2.49266922e-01
-1.11882794e+00 -1.45378307e-01 -9.07570004e-01 7.48411655e-01
1.56167495e+00 -8.63166928e-01 -1.02760482e+00 2.07914069e-01
1.52944446e+00 -2.48458549e-01 -7.33940303e-01 3.86855185e-01
8.05153728e-01 -6.64630413e-01 1.03085279e+00 -9.28614378e-01
6.12318993e-01 -4.26121414e-01 -3.84943396e-01 -1.30275714e+00
-4.33118403e-01 -5.58387399e-01 -8.89695942e-01 1.34220195e+00
5.96241474e-01 -8.14481825e-02 6.07050657e-01 6.06028378e-01
-5.19444227e-01 -6.71788335e-01 -5.23352742e-01 -9.37293947e-01
-1.11778371e-01 -5.07560015e-01 1.06740975e+00 9.33673978e-01
2.79658914e-01 6.47980392e-01 -7.52546638e-02 1.83952868e-01
5.09288311e-01 1.80607989e-01 8.16652954e-01 -1.07103014e+00
-4.54530448e-01 -5.25738955e-01 -5.34929574e-01 -9.98644471e-01
3.63834292e-01 -1.25904393e+00 -5.39039336e-02 -1.61098635e+00
4.75201756e-01 -1.34029716e-01 -4.27050501e-01 6.77967966e-01
-2.63363510e-01 -6.24540374e-02 5.41960239e-01 -2.64091603e-02
-1.23859680e+00 7.22085714e-01 8.36727738e-01 -8.29493642e-01
-1.51662841e-01 -4.86896932e-01 -9.30222392e-01 5.33133805e-01
3.36794585e-01 -1.73837334e-01 -5.40551066e-01 -6.20433450e-01
3.21711600e-01 -6.83868170e-01 2.96159327e-01 -1.28496289e+00
6.73615396e-01 1.44819468e-01 6.65043816e-02 -6.23437047e-01
5.35064638e-01 -8.97190273e-01 -8.41161013e-02 1.75625816e-01
-7.44272351e-01 -8.92402902e-02 7.25259930e-02 4.87919003e-01
-2.14500293e-01 -8.69284496e-02 4.42367077e-01 -2.47631103e-01
-1.32693422e+00 4.64862525e-01 2.14950323e-01 1.30317613e-01
1.09484184e+00 2.63213784e-01 -5.94962180e-01 5.02392314e-02
-5.69413304e-01 4.42017257e-01 4.35595483e-01 7.48786569e-01
5.86514115e-01 -1.32397473e+00 -6.29677117e-01 1.57514349e-01
1.05811048e+00 1.10344797e-01 4.20844376e-01 6.10138297e-01
-1.85968906e-01 5.21740258e-01 -1.01788096e-01 -2.99304903e-01
-1.21542633e+00 1.18420792e+00 1.35800943e-01 -1.26395658e-01
-8.22627544e-01 1.31752837e+00 2.59084702e-01 -3.96664917e-01
4.79946464e-01 -7.47569919e-01 -9.25637186e-02 9.05876011e-02
9.06307697e-01 9.93751809e-02 6.90711960e-02 -4.35553998e-01
-2.79639363e-01 2.79118627e-01 -4.68851924e-01 1.07803009e-01
1.31357729e+00 -1.16053492e-01 -8.66293311e-02 5.65899968e-01
1.18519640e+00 -2.82080114e-01 -1.15727448e+00 -8.21293116e-01
3.56801808e-01 -1.16801687e-01 -2.72289634e-01 -9.71664906e-01
-7.67464638e-01 9.87959743e-01 3.27338338e-01 -2.11234301e-01
9.37607408e-01 2.52008975e-01 1.00525737e+00 7.51172960e-01
4.76752758e-01 -1.16485298e+00 -3.45893204e-02 5.60798466e-01
8.56969774e-01 -1.10607266e+00 -1.17070980e-01 -3.64070624e-01
-8.42802525e-01 6.81874216e-01 8.60549867e-01 2.45040748e-02
3.69754076e-01 9.05632451e-02 -5.28935611e-01 3.53327319e-02
-1.48212004e+00 -6.75754189e-01 4.51879263e-01 5.76270223e-01
1.59475565e-01 3.29201788e-01 6.95341080e-02 1.05777454e+00
-2.10369512e-01 2.37825543e-01 2.28069723e-01 8.02568793e-01
-4.33171421e-01 -8.46720994e-01 4.81030047e-01 4.48667258e-01
8.74241516e-02 -3.43017131e-01 -4.32683915e-01 4.35003579e-01
6.75857067e-01 5.19609451e-01 1.66382030e-01 -4.62045372e-01
5.14951468e-01 5.42139828e-01 4.02103603e-01 -8.88402045e-01
-5.78329206e-01 -6.37981236e-01 1.90695778e-01 -7.51794338e-01
-3.95845234e-01 -6.35161817e-01 -1.32328308e+00 -3.25732827e-02
-8.51802751e-02 -1.32422313e-01 5.03274262e-01 1.00586855e+00
6.53740108e-01 7.30575323e-01 8.50091502e-03 -4.67320830e-01
-3.88415605e-01 -9.73487198e-01 -2.43400708e-01 5.72874665e-01
2.76856184e-01 -6.38205230e-01 -5.16355857e-02 3.12868178e-01] | [10.179327011108398, 2.014559268951416] |
4a0b6169-e525-4169-87a4-7cafa2ea1d54 | prompting-large-language-model-for-machine | 2301.07069 | null | https://arxiv.org/abs/2301.07069v2 | https://arxiv.org/pdf/2301.07069v2.pdf | Prompting Large Language Model for Machine Translation: A Case Study | Research on prompting has shown excellent performance with little or even no supervised training across many tasks. However, prompting for machine translation is still under-explored in the literature. We fill this gap by offering a systematic study on prompting strategies for translation, examining various factors for prompt template and demonstration example selection. We further explore the use of monolingual data and the feasibility of cross-lingual, cross-domain, and sentence-to-document transfer learning in prompting. Extensive experiments with GLM-130B (Zeng et al., 2022) as the testbed show that 1) the number and the quality of prompt examples matter, where using suboptimal examples degenerates translation; 2) several features of prompt examples, such as semantic similarity, show significant Spearman correlation with their prompting performance; yet, none of the correlations are strong enough; 3) using pseudo parallel prompt examples constructed from monolingual data via zero-shot prompting could improve translation; and 4) improved performance is achievable by transferring knowledge from prompt examples selected in other settings. We finally provide an analysis on the model outputs and discuss several problems that prompting still suffers from. | ['Alexandra Birch', 'Barry Haddow', 'Biao Zhang'] | 2023-01-17 | null | null | null | null | ['semantic-textual-similarity'] | ['natural-language-processing'] | [ 3.69506031e-01 7.97117036e-03 -4.67331797e-01 -4.62787420e-01
-1.43949664e+00 -9.69720960e-01 1.08336508e+00 2.08897799e-01
-5.17685473e-01 9.71377313e-01 7.09724903e-01 -8.38033497e-01
-1.08690172e-01 -2.74898577e-02 -7.41007388e-01 -2.79124588e-01
2.44127899e-01 6.91169202e-01 -1.87594533e-01 -5.80483496e-01
5.13270259e-01 1.53528154e-01 -9.42625761e-01 5.80755413e-01
1.11640358e+00 3.17502499e-01 5.08905530e-01 4.98678029e-01
-3.01024497e-01 5.94921589e-01 -7.96646714e-01 -2.73495317e-01
3.69736493e-01 -7.54479647e-01 -1.13073957e+00 1.91751905e-02
6.63725615e-01 -4.05669630e-01 4.36640568e-02 7.07326353e-01
6.30771875e-01 4.53919657e-02 5.74548185e-01 -1.16875350e+00
-1.07738853e+00 7.25292683e-01 6.27399608e-02 3.51441771e-01
8.15814257e-01 4.17157292e-01 9.91144300e-01 -1.13630474e+00
7.23167479e-01 1.10806656e+00 5.49560487e-01 6.25511646e-01
-1.24711263e+00 -4.87265587e-01 4.94286157e-02 8.91682878e-02
-8.44336927e-01 -6.30182624e-01 4.26438391e-01 -3.80900502e-01
1.08638954e+00 2.25568816e-01 3.47243696e-01 1.43148828e+00
6.42567351e-02 7.61635303e-01 1.67496097e+00 -8.74438286e-01
-2.07956791e-01 5.10943294e-01 7.63281956e-02 3.20998490e-01
-1.09251896e-02 4.40514743e-01 -9.36553180e-01 -2.68270373e-01
5.49421072e-01 -6.15236521e-01 -3.43476266e-01 1.57414377e-01
-1.68073893e+00 8.02380741e-01 1.37311621e-02 3.80663782e-01
-3.12867165e-01 -3.39128643e-01 4.73351926e-01 1.10146224e+00
4.70867664e-01 1.06859124e+00 -7.91779101e-01 -4.66642499e-01
-9.60009396e-01 4.54017282e-01 9.36931670e-01 1.20577943e+00
7.72708893e-01 -8.63250792e-02 -4.18383807e-01 9.61544275e-01
-7.40979761e-02 6.19414866e-01 5.97667575e-01 -7.72091448e-01
9.98872399e-01 3.75274003e-01 4.66698140e-01 -4.41268921e-01
-3.38458329e-01 -1.05595715e-01 -3.79637443e-02 -4.23516601e-01
8.01006317e-01 -5.54873407e-01 -5.06725013e-01 1.77288663e+00
-2.62056366e-02 -4.51658785e-01 1.60199434e-01 1.11434114e+00
7.33343959e-01 7.02687919e-01 2.92980045e-01 -4.72659260e-01
1.17258906e+00 -1.11348283e+00 -5.98172188e-01 -4.99989569e-01
1.07937694e+00 -1.48235142e+00 1.74964404e+00 4.89468910e-02
-9.96608019e-01 -5.87392926e-01 -6.25181139e-01 -1.61257520e-01
-1.20572545e-01 2.95491159e-01 5.87244630e-01 3.54693443e-01
-1.14679229e+00 5.26471972e-01 -4.18352991e-01 -9.43909943e-01
-2.65819728e-01 1.77955195e-01 -2.98436016e-01 -2.11083353e-01
-1.29653323e+00 1.36771011e+00 -3.22760828e-02 -3.49166870e-01
-5.02000988e-01 -7.67264307e-01 -7.68968284e-01 -1.46277875e-01
9.31480601e-02 -6.54175937e-01 1.72237337e+00 -1.28245389e+00
-1.57554734e+00 7.08824754e-01 -1.91713288e-01 -4.32237424e-02
2.81518161e-01 -4.49455976e-01 -3.20521593e-01 -7.05342274e-03
4.39275742e-01 9.55681086e-01 4.24416155e-01 -9.60687876e-01
-6.57732248e-01 -5.31478859e-02 8.28096122e-02 7.89562941e-01
-3.48816961e-01 7.28571534e-01 1.72913998e-01 -6.51036203e-01
-1.47169903e-01 -1.01433980e+00 3.68355215e-03 -4.49467927e-01
-1.04137786e-01 -5.26283681e-01 3.78399581e-01 -9.10090506e-01
8.34948599e-01 -1.98672390e+00 -1.12714514e-01 -3.24984401e-01
-4.64632869e-01 1.54468834e-01 -6.28294468e-01 9.64668751e-01
5.55772409e-02 1.01367064e-01 1.16765261e-01 1.07771143e-01
1.22683553e-03 -4.36296649e-02 -4.39739823e-01 2.47339979e-01
3.45284104e-01 1.10501599e+00 -1.09213662e+00 -4.23671305e-01
2.50177309e-02 1.09241769e-01 -3.47552896e-01 4.00009036e-01
-2.97308087e-01 7.24222839e-01 -1.76649362e-01 4.58081424e-01
1.75930671e-02 -3.51676911e-01 2.85130829e-01 3.47267956e-01
-1.72522545e-01 1.05014956e+00 -5.88169515e-01 1.78352129e+00
-4.79667246e-01 8.42405796e-01 3.13235410e-02 -6.34648561e-01
1.12557435e+00 5.86387038e-01 2.01235503e-01 -9.99163866e-01
-2.83215702e-01 6.42611206e-01 2.87066191e-01 -6.74888670e-01
8.40408802e-01 -3.22118342e-01 -6.14550197e-03 8.24567795e-01
-1.85974268e-03 -2.04432711e-01 1.79499611e-01 2.25105658e-01
8.80176842e-01 4.48704600e-01 2.16288909e-01 -5.18687785e-01
1.04054898e-01 7.30716825e-01 4.32424277e-01 6.21900201e-01
-1.80860490e-01 4.36545134e-01 7.71169215e-02 -6.15335703e-02
-9.95312274e-01 -7.00314105e-01 5.30052185e-02 1.37302017e+00
3.17680687e-02 -3.16284508e-01 -6.54622316e-01 -7.87446260e-01
-1.08299263e-01 1.09129810e+00 -2.00577438e-01 1.52071297e-01
-7.81479180e-01 -4.04291660e-01 3.50774169e-01 5.12322307e-01
-5.75205348e-02 -1.17967808e+00 -4.76353228e-01 2.99862176e-01
-7.57938921e-01 -1.12028253e+00 -8.10215712e-01 2.52338558e-01
-1.16910994e+00 -7.33298898e-01 -8.41870666e-01 -1.00832534e+00
6.22381270e-01 5.52170277e-01 1.28345597e+00 9.26680211e-03
5.13653100e-01 6.04767919e-01 -6.15247190e-01 -2.22966924e-01
-8.00795555e-01 3.26730222e-01 1.79934114e-01 -6.37348413e-01
6.18429661e-01 -2.45866597e-01 -2.67700791e-01 4.08145308e-01
-4.00430530e-01 1.43045202e-01 7.98136175e-01 1.09356999e+00
1.39775023e-01 -9.01596129e-01 8.96751702e-01 -8.88135791e-01
1.16078830e+00 -4.40259606e-01 -2.86223799e-01 4.07099128e-01
-8.54955077e-01 -8.77875984e-02 6.77774847e-01 -6.06600106e-01
-1.13101745e+00 -3.96447122e-01 1.20513648e-01 1.86553240e-01
-4.63964462e-01 5.82044661e-01 2.16128111e-01 1.09287230e-02
1.01611245e+00 2.10795164e-01 4.46942542e-03 -6.46911144e-01
4.78535324e-01 8.05842042e-01 1.76108867e-01 -1.08438134e+00
5.96104980e-01 -2.05205500e-01 -5.33767641e-01 -5.53101659e-01
-5.91553211e-01 -4.16924745e-01 -4.55138385e-01 -5.49532920e-02
3.61553133e-01 -9.48910892e-01 -4.21074145e-02 -9.12856981e-02
-1.42657995e+00 -9.13892090e-01 -1.46124303e-01 8.65873873e-01
-7.80975819e-01 1.82755470e-01 -8.89398038e-01 -3.88365030e-01
-3.46871972e-01 -1.31144559e+00 1.05126369e+00 8.90022293e-02
-8.56965840e-01 -1.02088308e+00 5.76545335e-02 4.97176200e-01
5.51922321e-01 -3.81874233e-01 1.13031650e+00 -1.14058900e+00
-4.94136214e-01 1.88718036e-01 -3.33862789e-02 1.95719242e-01
3.46535623e-01 -3.13108444e-01 -4.00118172e-01 -2.57703722e-01
-3.02353390e-02 -6.86157286e-01 6.79690391e-02 1.53052896e-01
1.86742529e-01 -6.20021820e-01 -1.25640333e-01 1.73805892e-01
1.02560842e+00 8.96644294e-02 1.72921628e-01 5.83815694e-01
3.77594560e-01 1.07449484e+00 1.18895411e+00 -1.08197205e-01
4.76150125e-01 8.74839544e-01 -3.88709754e-01 -1.18305221e-01
-4.04481262e-01 -4.79402661e-01 8.51622403e-01 1.18521512e+00
3.14020544e-01 7.41571262e-02 -9.64125693e-01 6.38788044e-01
-1.61966717e+00 -6.67779088e-01 -2.18911231e-01 1.99659252e+00
1.16625583e+00 -1.81369677e-01 2.03059047e-01 -3.28772247e-01
5.93881369e-01 -1.27816677e-01 -1.24666318e-01 -8.02741587e-01
-1.59293085e-01 4.53400649e-02 2.66402930e-01 7.13646233e-01
-4.69185352e-01 1.28804207e+00 6.85352755e+00 4.15418625e-01
-1.16936553e+00 2.02784792e-01 5.21833420e-01 9.35834944e-02
-5.01771212e-01 3.68949413e-01 -8.52980137e-01 4.42351639e-01
1.00238752e+00 -3.53929818e-01 5.75849891e-01 6.17889285e-01
3.62789571e-01 -6.10540807e-02 -1.52003562e+00 6.12718046e-01
1.58720061e-01 -9.93630469e-01 -1.91368595e-01 -1.67761207e-01
9.02025342e-01 1.17217742e-01 -1.94510981e-01 4.89199132e-01
3.58407289e-01 -7.30036020e-01 7.34151244e-01 -8.49103630e-02
8.52113068e-01 -2.51444310e-01 4.67290103e-01 6.16500556e-01
-4.55001056e-01 7.51204342e-02 -1.66039497e-01 -4.53501493e-01
2.69616932e-01 1.33480191e-01 -1.32091594e+00 6.39272690e-01
3.07019681e-01 3.16857785e-01 -6.59098685e-01 6.65544748e-01
-5.94791472e-01 1.02593327e+00 -1.33553490e-01 -2.92585492e-01
3.74801934e-01 -2.29180261e-01 4.29099947e-01 1.39629090e+00
4.37781423e-01 1.82570703e-02 3.41729939e-01 3.99770528e-01
8.21508989e-02 6.36650980e-01 -7.38779604e-01 -2.00855747e-01
7.44727135e-01 9.83936667e-01 -4.81328279e-01 -4.12215590e-01
-7.07944155e-01 8.54438663e-01 2.98369944e-01 6.58050120e-01
-3.35579515e-01 -8.83834735e-02 3.87845159e-01 1.61393330e-01
-3.30722332e-01 -3.95885617e-01 -7.21760571e-01 -1.03291523e+00
1.52442887e-01 -1.55945373e+00 2.35323146e-01 -8.27234566e-01
-1.33878314e+00 4.66467083e-01 8.27514231e-02 -1.13186741e+00
-5.81903934e-01 -5.32635510e-01 -4.81339544e-01 1.33390391e+00
-1.42820430e+00 -8.03282320e-01 1.25039801e-01 3.73419374e-01
9.23648953e-01 -1.48819402e-01 8.34272921e-01 2.76289642e-01
-2.58101020e-02 6.06477261e-01 -1.65936366e-01 -1.29615083e-01
1.56605494e+00 -1.14015853e+00 6.74391210e-01 8.37892175e-01
4.24908102e-01 1.00662315e+00 7.91604161e-01 -8.21237087e-01
-1.43849754e+00 -6.70206308e-01 1.76598489e+00 -8.08448076e-01
5.56328595e-01 -1.11754157e-01 -8.39426696e-01 8.48726630e-01
7.92514920e-01 -5.95349610e-01 7.48197913e-01 5.28500259e-01
-2.13680223e-01 1.83203250e-01 -7.74683714e-01 8.96814823e-01
1.04596126e+00 -6.21750176e-01 -8.36616635e-01 8.45662236e-01
8.96641195e-01 -5.03783703e-01 -7.49825776e-01 3.05462718e-01
2.39078179e-01 -5.80657721e-01 5.40156960e-01 -8.17449629e-01
4.56868768e-01 3.69184352e-02 -1.96896642e-01 -1.68339956e+00
-4.07819778e-01 -8.86308312e-01 3.14006537e-01 1.33203149e+00
7.62771666e-01 -4.79179651e-01 3.56111497e-01 6.62993491e-01
-3.96447837e-01 -8.39077950e-01 -5.10053456e-01 -1.07591546e+00
5.33826947e-01 -1.52081132e-01 6.61434770e-01 1.15572250e+00
5.13492346e-01 9.73941922e-01 -2.85501003e-01 -3.16003531e-01
1.53527502e-02 2.07150564e-01 9.23920453e-01 -7.84011781e-01
-3.84621024e-01 -3.38489741e-01 4.06522661e-01 -1.49525833e+00
1.65307134e-01 -1.20859087e+00 2.53033102e-01 -1.67862999e+00
1.70860484e-01 -6.12437189e-01 1.38762727e-01 5.98997116e-01
-6.71127856e-01 -2.11492956e-01 3.52664500e-01 5.61181605e-01
-2.56405294e-01 3.11180919e-01 1.37271070e+00 2.53049046e-01
-3.04974079e-01 -2.02091768e-01 -1.01348925e+00 8.22762251e-02
8.65179300e-01 -3.58901858e-01 -4.43748713e-01 -8.98168564e-01
-1.39784561e-02 3.81561905e-01 1.91678368e-02 -3.39868575e-01
1.67874873e-01 -4.65183705e-01 5.24826124e-02 -1.89735442e-01
1.45705221e-02 -5.21321416e-01 -2.49587566e-01 2.03633413e-01
-7.15478599e-01 6.72269881e-01 2.88608938e-01 7.48674721e-02
-9.51386690e-02 -2.15318143e-01 3.37748051e-01 -8.46452191e-02
-6.62107408e-01 -1.43776044e-01 -5.10485172e-01 3.33997339e-01
4.38926607e-01 1.67795941e-02 -5.45010328e-01 -4.70226020e-01
-8.77942294e-02 1.34344682e-01 7.07662046e-01 8.48594964e-01
1.06567621e-01 -1.47411966e+00 -1.05078351e+00 -1.81900516e-01
3.04903477e-01 -4.80571926e-01 -3.75391841e-01 1.01731658e+00
-4.04053181e-02 8.74603391e-01 -1.14626348e-01 -7.58565724e-01
-1.09630287e+00 4.02831495e-01 -1.79422528e-01 -2.16656953e-01
-5.60236394e-01 6.20671213e-01 -5.68208843e-02 -7.74162173e-01
2.64984407e-02 1.13604091e-01 1.85374081e-01 -5.55258803e-02
2.51231819e-01 2.56297708e-01 1.35452569e-01 -5.09846091e-01
-3.35646033e-01 1.87544405e-01 -2.71144420e-01 -7.00855672e-01
8.86241198e-01 -2.13726908e-01 1.16923742e-01 5.23360133e-01
9.61083531e-01 -1.27930209e-01 -1.06982374e+00 -4.87707824e-01
4.04488534e-01 -4.02636856e-01 -4.92405981e-01 -1.35598636e+00
-1.67095169e-01 8.52645397e-01 1.11889943e-01 -1.53716281e-01
8.45485151e-01 -8.94964337e-02 7.49807715e-01 5.02562284e-01
4.92908984e-01 -1.20978808e+00 1.11679554e-01 7.68964231e-01
9.28163171e-01 -1.33965278e+00 -2.93642491e-01 -1.71880528e-01
-9.16572988e-01 1.08307707e+00 7.41759717e-01 3.49859357e-01
-1.99444219e-01 2.95294039e-02 7.70409822e-01 6.52032718e-02
-9.53618288e-01 2.20944330e-01 1.65614516e-01 5.30924082e-01
1.01957047e+00 2.51610398e-01 -9.15264368e-01 3.92577350e-01
-4.78049248e-01 -4.24023978e-02 3.14401627e-01 8.95271897e-01
-3.52108091e-01 -1.47315788e+00 -5.12371063e-01 3.98420662e-01
-3.01946193e-01 -6.56743109e-01 -7.25903571e-01 7.26496696e-01
-3.86199683e-01 1.37998521e+00 -2.80279189e-01 -2.86906272e-01
2.21292391e-01 5.36079049e-01 7.30929792e-01 -1.09719098e+00
-1.05306256e+00 6.58840835e-02 5.56223869e-01 -2.60938138e-01
-8.33724961e-02 -1.04343450e+00 -8.16649079e-01 -2.50796765e-01
-4.53350335e-01 5.22910953e-01 7.92939365e-01 1.18215489e+00
4.72253740e-01 5.73011413e-02 5.76398194e-01 -5.02762139e-01
-1.09764326e+00 -1.42388284e+00 1.87521968e-02 4.65104431e-01
1.16915442e-01 -3.67969394e-01 -1.78197771e-01 2.17269465e-01] | [11.591409683227539, 10.252449035644531] |
c5d4acb4-65d2-454a-8492-2196e6f27580 | face-presentation-attack-detection-in-learned | 1810.13170 | null | http://arxiv.org/abs/1810.13170v2 | http://arxiv.org/pdf/1810.13170v2.pdf | Face Presentation Attack Detection in Learned Color-liked Space | Face presentation attack detection (PAD) has become a thorny problem for
biometric systems and numerous countermeasures have been proposed to address
it. However, majority of them directly extract feature descriptors and
distinguish fake faces from the real ones in existing color spaces (e.g. RGB,
HSV and YCbCr). Unfortunately, it is unknown for us which color space is the
best or how to combine different spaces together. To make matters worse, the
real and fake faces are overlapped in existing color spaces. So, in this paper,
a learned distinguishable color-liked space is generated to deal with the
problem of face PAD. More specifically, we present an end-to-end deep learning
network that can map existing color spaces to a new learned color-liked space.
Inspired by the generator of generative adversarial network (GAN), the proposed
network consists of a space generator and a feature extractor. When training
the color-liked space, a new triplet combination mechanism of points-to-center
is explored to maximize interclass distance and minimize intraclass distance,
and also keep a safe margin between the real and presented fake faces.
Extensive experiments on two standard face PAD databases, i.e., Relay-Attack
and OULU-NPU, indicate that our proposed color-liked space analysis based
countermeasure significantly outperforms the state-of-the-art methods and show
excellent generalization capability. | ['Xiaoyi Feng', 'Zhaoqiang Xia', 'Lei Li', 'Fabio Roli', 'Xiaoyue Jiang'] | 2018-10-31 | null | null | null | null | ['face-presentation-attack-detection'] | ['computer-vision'] | [ 3.55851976e-03 -4.64010566e-01 9.34849679e-02 -2.83571899e-01
-4.56003398e-01 -7.21151352e-01 2.94656634e-01 -7.11275160e-01
-1.08412988e-02 7.33720541e-01 -4.56588596e-01 -4.14511561e-01
1.29656553e-01 -8.26859176e-01 -4.43304807e-01 -9.55122292e-01
2.05934793e-01 6.63772374e-02 -9.59646180e-02 -4.03340399e-01
3.02297324e-01 8.05182397e-01 -1.24702287e+00 1.63452625e-01
1.00105345e+00 1.11609948e+00 -4.84763026e-01 1.65099218e-01
-1.97744623e-01 1.29377127e-01 -9.73615885e-01 -7.58010626e-01
5.37942886e-01 -8.44813883e-01 -2.27141961e-01 1.98605098e-02
5.26088774e-01 -3.47484231e-01 -6.49852753e-01 1.28912210e+00
5.12093127e-01 -3.07557434e-01 6.40173256e-01 -1.89001417e+00
-1.22439003e+00 -1.65370144e-02 -9.17027771e-01 -1.78891979e-02
2.85688996e-01 2.52465695e-01 1.43341154e-01 -8.67533922e-01
3.37078243e-01 1.37830675e+00 3.86042029e-01 9.45265770e-01
-9.48044240e-01 -1.44790792e+00 -1.41228616e-01 4.43391383e-01
-1.62846124e+00 -1.54796854e-01 1.14703536e+00 -9.40111503e-02
6.77705184e-02 3.39293838e-01 5.20865858e-01 1.16880655e+00
1.62097484e-01 5.46560884e-01 1.38728392e+00 -3.04411024e-01
-3.91665772e-02 2.69352704e-01 -3.64331573e-01 8.10378850e-01
5.41506290e-01 3.33644986e-01 -2.88629889e-01 -1.82285249e-01
7.97085881e-01 1.52533084e-01 -4.61967587e-01 -3.08505684e-01
-5.54321229e-01 7.22971082e-01 6.43428028e-01 2.47018814e-01
-3.26376930e-02 -1.25169978e-01 2.48019099e-02 2.42382362e-01
1.94291532e-01 2.24728882e-01 -3.29376012e-02 3.13775420e-01
-7.60320783e-01 1.00622512e-01 4.73999143e-01 7.92972445e-01
7.00980604e-01 3.24641168e-01 -5.03427573e-02 6.74518824e-01
3.48624349e-01 8.92376184e-01 3.50777894e-01 -3.28808337e-01
5.08126020e-01 5.16361713e-01 8.68283808e-02 -1.53846467e+00
-3.30637097e-02 -1.13932505e-01 -9.36144114e-01 4.17625546e-01
4.86757636e-01 -2.31697872e-01 -8.86610448e-01 1.61935127e+00
3.74433458e-01 6.42839253e-01 9.30381566e-02 1.12641990e+00
5.91947794e-01 5.76909781e-01 -1.45295724e-01 -1.58151448e-01
1.30401480e+00 -5.26805460e-01 -7.94738650e-01 -1.85004294e-01
1.19125396e-01 -8.44591439e-01 8.13865423e-01 4.44198221e-01
-6.88843489e-01 -6.30258322e-01 -1.36866975e+00 2.41574228e-01
-6.34775102e-01 2.98553586e-01 5.42383850e-01 1.41920555e+00
-8.20149004e-01 3.68881553e-01 -3.12581271e-01 -1.20486319e-03
8.29089880e-01 3.15303653e-01 -6.46242440e-01 -2.99399942e-01
-1.29567719e+00 6.25648737e-01 2.09516108e-01 5.34999609e-01
-6.47973478e-01 -3.13075215e-01 -5.90093970e-01 -9.92583707e-02
2.73850381e-01 9.21226367e-02 5.04798651e-01 -1.05749452e+00
-1.62583005e+00 7.24245191e-01 2.46289045e-01 -2.79490016e-02
4.58081871e-01 1.41518742e-01 -9.61189389e-01 1.13966413e-01
-2.15660155e-01 2.82979041e-01 1.22989345e+00 -1.51647508e+00
-3.08916301e-01 -6.76926732e-01 -2.59604663e-01 -1.50440395e-01
-6.22868001e-01 2.83307601e-02 -3.69394571e-01 -7.36290634e-01
2.82480359e-01 -9.94929016e-01 4.09592867e-01 2.89222091e-01
-6.08760536e-01 8.10862258e-02 1.30044770e+00 -7.88718224e-01
9.83900607e-01 -2.22831249e+00 -1.24414362e-01 5.63559771e-01
9.74065214e-02 7.56134987e-01 -3.73675704e-01 1.04027718e-01
-3.72559667e-01 1.43183365e-01 -2.08517373e-01 7.70351067e-02
-1.21768676e-01 -2.39653559e-03 -4.27706748e-01 8.99619818e-01
4.62081581e-01 7.15864241e-01 -8.65082860e-01 -4.06430304e-01
4.17101867e-02 8.06629896e-01 -2.95092225e-01 2.09366083e-01
2.42404938e-01 5.53483725e-01 -4.43256944e-01 9.84321177e-01
1.64902484e+00 3.06301087e-01 2.46997729e-01 -3.26069713e-01
2.48898983e-01 -3.49323928e-01 -1.26371408e+00 1.08420944e+00
-1.63326278e-01 6.02356255e-01 -1.47296250e-01 -8.72308433e-01
1.35043168e+00 1.71242326e-01 2.84539968e-01 -7.95319557e-01
2.91835189e-01 3.45320523e-01 1.30000591e-01 -2.38693655e-01
1.03678010e-01 -8.61549675e-02 1.68192968e-01 2.93201774e-01
-1.32806033e-01 1.52645305e-01 -3.62996459e-01 -2.24430397e-01
6.83518708e-01 2.39617340e-02 -1.97711647e-01 1.09108038e-01
9.05481756e-01 -4.85463351e-01 7.25077689e-01 2.79529214e-01
-4.36278701e-01 5.30599296e-01 6.18856132e-01 -3.19220901e-01
-5.98186195e-01 -1.12860441e+00 -1.16257086e-01 4.43025708e-01
5.83633840e-01 4.80707958e-02 -8.60630751e-01 -1.04662001e+00
2.25987017e-01 3.29984605e-01 -6.36434793e-01 -5.47317266e-01
-7.33129501e-01 -6.19330645e-01 9.75574076e-01 2.06584543e-01
9.24643755e-01 -8.23644102e-01 -7.24722147e-02 -5.48043735e-02
3.16318534e-02 -9.77566004e-01 -5.64832091e-01 -5.18066227e-01
-4.52963412e-01 -1.37723899e+00 -7.93956637e-01 -8.42426836e-01
8.38411033e-01 4.49022293e-01 5.61007440e-01 4.02279735e-01
-4.65322345e-01 -3.18868421e-02 -3.65687609e-01 -4.20324385e-01
-2.98612624e-01 -4.95762616e-01 1.05025023e-01 5.81171095e-01
4.73789275e-01 -1.27284229e-01 -7.95223057e-01 5.46463251e-01
-9.49669719e-01 -3.76285136e-01 5.39583206e-01 9.91955459e-01
1.87766135e-01 2.61847496e-01 4.82422829e-01 -6.38637125e-01
5.62622726e-01 -3.01050365e-01 -7.31667519e-01 5.88962376e-01
-4.10976768e-01 -8.21025670e-02 8.12530696e-01 -5.91933966e-01
-9.57683802e-01 -3.38888057e-02 4.58093360e-02 -6.20806217e-01
3.67911719e-02 -6.53667301e-02 -6.52112603e-01 -6.33421540e-01
4.53458905e-01 4.31058735e-01 2.20288575e-01 -1.84255391e-01
4.61122930e-01 8.21026027e-01 6.27484441e-01 -5.32483101e-01
1.47241211e+00 4.97444957e-01 1.72347620e-01 -5.91074526e-01
-9.42268129e-03 8.95816684e-02 -3.51578623e-01 -2.45287627e-01
5.45555294e-01 -5.50328970e-01 -9.10560310e-01 9.31583166e-01
-1.14606619e+00 3.15312475e-01 3.57524842e-01 2.16867164e-01
-1.60494164e-01 5.02001524e-01 -3.99970651e-01 -9.67375875e-01
-2.41728947e-01 -1.21755242e+00 1.03321731e+00 6.89038992e-01
6.40537620e-01 -6.39277935e-01 -2.86964476e-01 1.41999885e-01
4.62785393e-01 6.21858656e-01 7.85909116e-01 -2.82573819e-01
-5.73519468e-01 -5.28389573e-01 -4.66844708e-01 3.81143004e-01
5.83017468e-01 1.27829298e-01 -9.65496719e-01 -5.45124769e-01
1.73624360e-03 -6.42412379e-02 6.39163435e-01 -1.57726750e-01
1.42699671e+00 -4.76092011e-01 -4.42867458e-01 8.32953691e-01
1.38098145e+00 6.55815184e-01 9.97965574e-01 8.92320201e-02
6.74591362e-01 5.31200290e-01 5.18788993e-01 2.21403360e-01
7.26501495e-02 7.52480924e-01 5.37647009e-01 -1.55353546e-01
4.43415977e-02 -2.34696865e-01 2.23407373e-01 1.63169473e-01
1.39416799e-01 -2.35560730e-01 -5.44162452e-01 -4.55693863e-02
-1.29936826e+00 -8.75611126e-01 2.88463622e-01 2.28882766e+00
6.95295870e-01 -7.42521062e-02 7.33860433e-02 2.55866408e-01
1.00896811e+00 3.20945680e-02 -6.15971267e-01 -2.84861624e-01
-4.24126685e-01 4.69033033e-01 4.24342930e-01 1.58640712e-01
-1.07196569e+00 1.03374887e+00 4.45161438e+00 8.14148962e-01
-1.73167980e+00 -2.04323567e-02 7.73826122e-01 2.47162074e-01
3.08979154e-02 -1.94273666e-01 -4.62311059e-01 7.53851831e-01
3.80555034e-01 6.79383948e-02 7.03038871e-01 6.82664156e-01
-2.36371517e-01 3.87458742e-01 -7.20171690e-01 1.37798369e+00
3.54824811e-01 -1.09303260e+00 7.19101503e-02 2.39046544e-01
6.09062970e-01 -7.18415141e-01 6.11037254e-01 1.33087695e-01
-1.86914474e-01 -1.15755892e+00 3.95206571e-01 4.95806336e-01
1.17257833e+00 -9.68317330e-01 7.77425110e-01 -1.75356120e-01
-1.18546641e+00 -1.40659884e-01 -4.98546064e-01 5.52578926e-01
-3.38703960e-01 1.39722139e-01 -4.83877242e-01 7.14590073e-01
4.50235903e-01 3.76454026e-01 -7.32908487e-01 9.80678558e-01
-2.34361723e-01 3.14051479e-01 -9.77411866e-03 7.01284185e-02
1.31499529e-01 -2.99893588e-01 3.47799420e-01 8.34613502e-01
6.38735175e-01 1.95739254e-01 -1.05059519e-01 9.95129883e-01
-2.65925914e-01 -1.51646240e-02 -6.34334207e-01 -7.08061680e-02
7.58578956e-01 1.24290204e+00 -6.33198977e-01 5.12411520e-02
-2.39341632e-01 1.22521937e+00 6.41255453e-03 4.23107028e-01
-1.07081509e+00 -7.16576338e-01 9.06814575e-01 -8.75222385e-02
1.56289816e-01 -3.25202756e-02 -6.43810928e-02 -1.17425883e+00
1.17827550e-01 -1.05039954e+00 2.67933249e-01 -5.94627142e-01
-1.50954747e+00 8.23903322e-01 -3.51619452e-01 -1.56918633e+00
9.11187530e-02 -9.63721275e-01 -7.70714760e-01 1.13386595e+00
-1.51863480e+00 -1.10821688e+00 -5.72107494e-01 9.24191475e-01
-1.56165525e-01 -6.46783531e-01 6.68533564e-01 3.21900517e-01
-6.72412157e-01 1.31264555e+00 1.75275624e-01 5.75766802e-01
7.82336831e-01 -8.41185033e-01 2.61731982e-01 8.75011861e-01
1.36293516e-01 6.38794720e-01 2.44926289e-01 -5.83236039e-01
-1.82653213e+00 -1.02957129e+00 8.88138711e-02 -1.79089174e-01
2.32787073e-01 -3.08045477e-01 -1.02353036e+00 1.19755775e-01
1.23766176e-02 4.43259835e-01 4.90221083e-01 -6.16069078e-01
-8.12682807e-01 -6.04576766e-01 -1.62021887e+00 5.89408159e-01
7.77439117e-01 -6.52536273e-01 -8.86815134e-03 2.63928086e-01
1.21895112e-01 -4.01109219e-01 -3.73811603e-01 2.77737677e-01
7.53976405e-01 -9.98583138e-01 1.13313603e+00 -6.09126747e-01
3.19379754e-02 -5.65207362e-01 -4.45674136e-02 -1.41221130e+00
-6.92374408e-02 -6.25862598e-01 -7.20489770e-02 1.21810853e+00
-8.49224329e-02 -9.75057364e-01 9.08325732e-01 2.27969229e-01
2.15722889e-01 -6.64174914e-01 -1.03192163e+00 -9.60986257e-01
1.54009834e-01 9.09382179e-02 1.08110571e+00 9.97246623e-01
-2.72107542e-01 -3.37763041e-01 -5.16650736e-01 4.89573628e-01
9.08623815e-01 1.03667788e-01 7.44807959e-01 -1.16869879e+00
1.10278726e-01 -5.00335276e-01 -7.26763368e-01 -4.95835423e-01
3.02650303e-01 -6.16637588e-01 -1.08040601e-01 -7.87781715e-01
-2.29785740e-01 -5.79228163e-01 -4.47207451e-01 4.67329353e-01
-3.42874855e-01 7.26820648e-01 3.14787090e-01 -3.80719267e-02
1.52525380e-01 4.85725194e-01 1.40862608e+00 -4.20866311e-01
-1.49421915e-01 -2.24816695e-01 -7.80064285e-01 3.10333818e-01
8.97446513e-01 -2.64095157e-01 -3.07921737e-01 -2.31752604e-01
-4.15569723e-01 1.47645041e-01 5.69894373e-01 -1.16028357e+00
8.34278688e-02 -3.63071114e-01 7.71084547e-01 -3.46325874e-01
4.58626002e-01 -9.23620224e-01 2.63541732e-02 5.07592142e-01
1.32950813e-01 -4.92754802e-02 2.72864938e-01 4.40891534e-01
-1.77673563e-01 2.02927858e-01 1.06090212e+00 1.41956165e-01
-4.85303760e-01 5.66231608e-01 2.70519853e-01 -1.31261930e-01
1.14611602e+00 -5.04211128e-01 -8.05612385e-01 -2.99524039e-01
-1.98076293e-01 -2.03429341e-01 2.50130177e-01 7.26851165e-01
1.03952789e+00 -1.72354066e+00 -7.95155764e-01 8.06448936e-01
1.72743782e-01 -5.68035841e-01 2.93821633e-01 3.67807657e-01
-6.21560872e-01 2.47568339e-01 -7.08568871e-01 -2.81573266e-01
-1.24659669e+00 6.48395121e-01 4.93054122e-01 3.62750053e-01
-2.05114603e-01 7.93269157e-01 7.00375140e-02 -2.47008801e-01
-3.01319943e-03 2.19624937e-01 -5.51455393e-02 -2.03191489e-01
7.08571196e-01 2.07082957e-01 1.52490847e-02 -9.18277860e-01
-5.89164019e-01 6.91969156e-01 6.16305973e-03 1.53947666e-01
1.02620649e+00 2.05397680e-01 -3.39576364e-01 -3.98642451e-01
1.44073403e+00 4.06862311e-02 -1.16605723e+00 -1.00832187e-01
-1.81863949e-01 -1.18153298e+00 -3.02951217e-01 -7.21594870e-01
-1.61727631e+00 9.24398661e-01 1.24004817e+00 2.68888354e-01
1.31992114e+00 -4.94054258e-01 8.08234632e-01 1.84237421e-01
4.68505323e-01 -8.16800475e-01 3.54906678e-01 1.77106343e-03
8.32490683e-01 -1.23398507e+00 -2.05120325e-01 -5.70345104e-01
-4.34227467e-01 1.31802630e+00 1.02062845e+00 -1.64440155e-01
6.35843813e-01 -1.20361008e-01 1.95933476e-01 4.01279256e-02
4.74999696e-02 7.63271526e-02 3.70199561e-01 1.01172686e+00
1.09941036e-01 4.02409770e-02 -2.02381819e-01 3.90677780e-01
-1.89701855e-01 -1.09084100e-01 4.24485326e-01 6.75550640e-01
-4.09395546e-02 -1.31522548e+00 -8.69707406e-01 2.06721649e-01
-2.43598863e-01 2.29231775e-01 -4.22842622e-01 8.93215597e-01
4.48937416e-01 1.05139792e+00 4.52625006e-02 -6.89999998e-01
1.92399204e-01 -1.47160068e-01 6.67470932e-01 -9.93856862e-02
-1.60667256e-01 -1.73355669e-01 -6.03582144e-01 -4.45103526e-01
-1.03240147e-01 -2.18072191e-01 -9.96978343e-01 -5.46754181e-01
-4.87920880e-01 1.50287300e-01 6.71593487e-01 7.04103768e-01
2.50820279e-01 1.05993636e-01 1.39790130e+00 -6.40029430e-01
-5.27789533e-01 -6.16405368e-01 -7.13747263e-01 3.97573680e-01
3.37418854e-01 -8.23873639e-01 -1.30574390e-01 -3.22026849e-01] | [12.849611282348633, 1.0021353960037231] |
48c398d4-5993-4712-93ae-8c362efd0be3 | universal-semi-supervised-model-adaptation | 2307.03449 | null | https://arxiv.org/abs/2307.03449v1 | https://arxiv.org/pdf/2307.03449v1.pdf | Universal Semi-supervised Model Adaptation via Collaborative Consistency Training | In this paper, we introduce a realistic and challenging domain adaptation problem called Universal Semi-supervised Model Adaptation (USMA), which i) requires only a pre-trained source model, ii) allows the source and target domain to have different label sets, i.e., they share a common label set and hold their own private label set, and iii) requires only a few labeled samples in each class of the target domain. To address USMA, we propose a collaborative consistency training framework that regularizes the prediction consistency between two models, i.e., a pre-trained source model and its variant pre-trained with target data only, and combines their complementary strengths to learn a more powerful model. The rationale of our framework stems from the observation that the source model performs better on common categories than the target-only model, while on target-private categories, the target-only model performs better. We also propose a two-perspective, i.e., sample-wise and class-wise, consistency regularization to improve the training. Experimental results demonstrate the effectiveness of our method on several benchmark datasets. | ['Guanbin Li', 'Shuguang Cui', 'Xiaoguang Han', 'Yipeng Qin', 'Yushuang Wu', 'Zizheng Yan'] | 2023-07-07 | null | null | null | null | ['domain-adaptation'] | ['methodology'] | [ 2.39527270e-01 1.34068891e-01 -6.12909019e-01 -5.60488343e-01
-7.70671904e-01 -4.04989272e-01 4.27575380e-01 -1.08210206e-01
-2.28327766e-01 7.80979276e-01 -1.88112527e-01 1.49545342e-01
1.43093854e-01 -5.24708807e-01 -7.18318641e-01 -7.41294801e-01
5.66289485e-01 6.21132374e-01 4.03676182e-01 7.53411129e-02
-3.47900599e-01 -1.20855890e-01 -1.12007177e+00 6.10433929e-02
1.14361978e+00 1.29756582e+00 3.13485652e-01 -9.12635177e-02
3.25541236e-02 9.54372466e-01 -2.90265828e-01 -2.96149939e-01
3.28857422e-01 -6.52762055e-01 -7.35228181e-01 4.63631123e-01
3.32982779e-01 -1.41212791e-01 -5.46882749e-02 1.12988067e+00
7.96779245e-02 1.72813565e-01 6.76031649e-01 -1.45450675e+00
-7.94652462e-01 4.23814029e-01 -5.71911275e-01 -2.97319531e-01
-1.55688450e-01 -1.40202763e-02 8.85007679e-01 -7.48049319e-01
6.13522053e-01 9.24670041e-01 6.18427932e-01 7.80725002e-01
-1.46636546e+00 -9.81995761e-01 5.87548673e-01 7.48880149e-04
-1.51528394e+00 -3.99010032e-01 9.54586267e-01 -4.31679130e-01
2.30307013e-01 -1.32868648e-01 2.01171502e-01 1.08946264e+00
-2.02376246e-01 7.85065591e-01 1.47178125e+00 -3.55251312e-01
3.90841722e-01 6.71450973e-01 2.46159956e-01 4.45681691e-01
1.99101567e-01 2.27596704e-02 -2.63211936e-01 -3.23797435e-01
6.18898392e-01 2.60564953e-01 -2.61771083e-01 -8.70325148e-01
-1.11920345e+00 8.66257012e-01 2.69568026e-01 3.40811253e-01
-2.14127839e-01 -4.28714752e-01 4.32939768e-01 2.99883485e-01
5.44288933e-01 -6.97320560e-03 -8.30906868e-01 3.69120598e-01
-7.58560717e-01 -1.78314734e-03 6.30677462e-01 1.25669575e+00
1.12211621e+00 -2.19893724e-01 2.81231310e-02 1.24903667e+00
4.72346604e-01 5.78781605e-01 8.45607877e-01 -7.06040561e-01
4.68009472e-01 6.20861292e-01 2.93727845e-01 -5.82509160e-01
-1.55821443e-01 -5.29681325e-01 -9.15350497e-01 -8.26359242e-02
5.26982963e-01 -1.36080191e-01 -8.65901351e-01 2.31193304e+00
5.62821686e-01 5.36561966e-01 2.53726900e-01 8.04837227e-01
5.55525422e-01 5.58042347e-01 2.05478668e-01 -3.48242790e-01
9.01450098e-01 -1.41326416e+00 -6.78370357e-01 -4.63190317e-01
7.47731924e-01 -4.79340076e-01 1.08621645e+00 1.88766703e-01
-9.13448572e-01 -8.00197959e-01 -1.12171316e+00 2.07859144e-01
-1.77645147e-01 4.35595661e-01 2.74167180e-01 2.63365746e-01
-5.05577385e-01 3.47287804e-01 -7.04406321e-01 -3.08642298e-01
3.50261658e-01 5.87758496e-02 -4.54621464e-01 -2.66551226e-01
-1.11266422e+00 7.21169829e-01 6.03313565e-01 -4.13923889e-01
-9.36868429e-01 -6.04820728e-01 -7.93821931e-01 -3.00836265e-02
5.29271364e-01 -4.54088062e-01 1.43541098e+00 -1.44526839e+00
-1.61030161e+00 9.36950088e-01 -3.81906480e-01 -2.57742375e-01
4.23976451e-01 -1.63684770e-01 -5.98019242e-01 -1.25131249e-01
2.98169792e-01 3.68953377e-01 9.46496725e-01 -1.64711750e+00
-8.54665220e-01 -2.87030429e-01 -2.42032081e-01 1.10130616e-01
-6.50000393e-01 -2.20415950e-01 -7.13705420e-01 -7.47197926e-01
2.29874536e-01 -1.05900288e+00 -8.67962092e-02 1.88498590e-02
-3.65708917e-01 -2.27601543e-01 8.21185708e-01 -5.75668991e-01
1.35774755e+00 -2.22650337e+00 2.02865735e-01 3.96173120e-01
1.57399997e-01 4.43208605e-01 -2.41236657e-01 7.92059451e-02
-2.05821380e-01 -1.38004124e-01 -5.58722556e-01 -5.22884607e-01
-1.35125816e-01 4.13145751e-01 -4.39866811e-01 2.85729825e-01
1.06251590e-01 5.60192168e-01 -9.54072535e-01 -5.31835496e-01
-1.30304340e-02 1.68955848e-01 -4.75936413e-01 4.70654190e-01
-2.21175298e-01 7.40369976e-01 -6.27982140e-01 4.23276424e-01
9.15568650e-01 -6.61419094e-01 5.56588650e-01 -1.66806191e-01
2.83062696e-01 1.87590271e-01 -1.38777339e+00 1.62344849e+00
-6.47794604e-01 2.51923390e-02 6.69559166e-02 -1.14749014e+00
1.10789049e+00 3.17048728e-01 3.12551051e-01 -5.25537193e-01
-7.57806748e-02 4.98001099e-01 -1.80929571e-01 -1.53315738e-01
2.23108754e-02 -3.34021598e-01 4.79564704e-02 6.51809156e-01
3.06737572e-01 2.40260556e-01 -1.51308123e-02 7.15574473e-02
6.60761297e-01 2.95525700e-01 4.83950794e-01 -8.27073157e-02
7.40943730e-01 -2.63490975e-01 1.21456492e+00 6.00665808e-01
-3.83742869e-01 5.99977493e-01 2.62314975e-01 -1.54597431e-01
-8.41248870e-01 -9.82185304e-01 -1.98712870e-01 1.20835674e+00
4.78777230e-01 -1.16553567e-01 -6.21215165e-01 -1.27907240e+00
2.45601702e-02 6.39911234e-01 -7.36965358e-01 -3.59523326e-01
-3.63117158e-01 -4.35898006e-01 1.27268016e-01 5.90613961e-01
6.97503805e-01 -5.59536815e-01 5.25006056e-02 1.45057246e-01
-2.26460814e-01 -1.04084182e+00 -7.08855569e-01 1.71425521e-01
-9.04364944e-01 -9.55800295e-01 -7.66183376e-01 -1.00744534e+00
8.82794678e-01 4.15504515e-01 1.01614928e+00 2.51278915e-02
7.16534376e-01 2.25069299e-01 -4.89176661e-01 -2.14840278e-01
-4.92350549e-01 2.53122840e-02 1.90824538e-01 4.76620495e-01
4.39846992e-01 -5.58390677e-01 -1.82173371e-01 7.05346227e-01
-7.96814978e-01 1.24148868e-01 6.18996739e-01 1.11813796e+00
9.57761228e-01 2.80436464e-02 9.06073093e-01 -1.32034492e+00
2.78284073e-01 -8.70587528e-01 -4.44576442e-01 6.45419121e-01
-9.35854912e-01 -1.64993167e-01 9.65939999e-01 -8.45202804e-01
-1.21499622e+00 2.37919793e-01 2.58892626e-01 -7.66507626e-01
-2.46743500e-01 5.71264505e-01 -5.82209468e-01 8.10468048e-02
5.89184940e-01 3.37242335e-01 1.06581949e-01 -6.31760776e-01
1.96014330e-01 7.92385459e-01 5.03444970e-01 -6.21743917e-01
1.03613257e+00 5.32463014e-01 -4.72028792e-01 -3.24900210e-01
-1.21870959e+00 -5.90974033e-01 -8.00841868e-01 1.18247278e-01
5.81791282e-01 -1.19505596e+00 1.46878421e-01 7.64479756e-01
-9.07435715e-01 -6.04457498e-01 -3.74191135e-01 5.11419058e-01
-3.33060026e-01 3.39304417e-01 -3.14229578e-01 -5.76696634e-01
-4.01657373e-02 -8.74396443e-01 7.55681515e-01 2.85860091e-01
-2.08853614e-02 -1.23859107e+00 1.64132386e-01 2.49890253e-01
2.47888580e-01 1.28901424e-02 7.06929326e-01 -1.20567989e+00
-1.27601415e-01 -2.41151094e-01 -3.03766668e-01 8.87445748e-01
5.00894368e-01 -3.23141575e-01 -9.28431392e-01 -4.47671533e-01
1.00208431e-01 -6.38013244e-01 7.36966133e-01 8.75125453e-02
1.08623660e+00 -2.73584843e-01 -4.61919308e-01 4.99386072e-01
1.30125225e+00 1.50602669e-01 2.04052374e-01 1.18538551e-01
7.70355344e-01 6.19609058e-01 8.59972537e-01 4.49906170e-01
6.40464067e-01 8.26063693e-01 1.72008388e-02 -2.34501496e-01
-6.72426000e-02 -5.51912844e-01 2.93291152e-01 9.78356004e-01
3.15509379e-01 -1.93359666e-02 -7.51113057e-01 4.90440190e-01
-2.04074097e+00 -5.78890622e-01 1.48455396e-01 2.45416117e+00
1.13021231e+00 5.69468401e-02 2.24888280e-01 -1.19219467e-01
8.69726002e-01 -3.32220495e-02 -9.11129951e-01 1.37092024e-01
-1.08183570e-01 1.07411020e-01 2.91654795e-01 2.87920326e-01
-1.32460046e+00 8.14717114e-01 5.39441347e+00 9.35889840e-01
-1.21753395e+00 4.16753799e-01 5.67909777e-01 1.82581306e-01
-2.47790813e-01 1.44642904e-01 -7.70719409e-01 8.39561284e-01
8.32478464e-01 -2.19707355e-01 2.91047633e-01 1.26934016e+00
-8.02836344e-02 1.57159552e-01 -1.19397831e+00 7.37899482e-01
6.45150766e-02 -9.76563871e-01 -5.55843711e-02 -3.91171537e-02
9.16666090e-01 -1.59248650e-01 -6.42611831e-02 6.10770643e-01
5.30535161e-01 -4.97570813e-01 7.88322628e-01 3.38007927e-01
7.67690361e-01 -5.63688517e-01 6.93978667e-01 6.50421858e-01
-1.21051860e+00 -1.01788051e-01 -4.14873332e-01 1.84895977e-01
-1.57453045e-02 6.13840044e-01 -2.02200368e-01 7.39154756e-01
6.82857394e-01 1.09388769e+00 -4.66348082e-01 7.81626761e-01
-3.70185763e-01 8.41952384e-01 -2.02336639e-01 4.66377109e-01
-1.14313059e-01 -4.26936835e-01 2.55929470e-01 9.16419387e-01
7.70430788e-02 -2.48977984e-03 6.23435080e-01 7.84878731e-01
-1.81624964e-01 1.75986782e-01 -2.20213532e-01 2.13083953e-01
8.97780836e-01 1.13268685e+00 -2.63177842e-01 -6.32089674e-01
-7.68718779e-01 1.01479995e+00 6.65321231e-01 5.49204350e-01
-9.43100452e-01 -2.18596891e-01 4.95421708e-01 -7.88119808e-02
3.50547761e-01 2.21446350e-01 -2.84109056e-01 -1.45568740e+00
8.61612707e-02 -8.99724245e-01 6.09462619e-01 -4.85709727e-01
-1.75943017e+00 5.14905035e-01 -8.92551616e-03 -1.64797652e+00
-2.04021245e-01 -2.83213526e-01 -6.49277747e-01 1.07105243e+00
-1.70651793e+00 -1.40540898e+00 -2.84334630e-01 8.27050209e-01
2.00874284e-01 -2.02485621e-01 8.33251476e-01 3.82631123e-01
-8.20280552e-01 9.48928177e-01 5.59695482e-01 1.26770407e-01
1.10253596e+00 -9.51911211e-01 -1.10601872e-01 8.03164780e-01
-3.95004041e-02 6.53200567e-01 2.62619883e-01 -5.10607481e-01
-8.29884052e-01 -1.52231991e+00 8.65662992e-01 -2.77891219e-01
5.37086248e-01 -2.52437949e-01 -1.34032571e+00 1.01621771e+00
-2.73850173e-01 1.64465114e-01 1.03607082e+00 1.30632430e-01
-8.63744795e-01 -3.17870021e-01 -1.31788003e+00 2.72311628e-01
7.53738225e-01 -5.22786796e-01 -6.48374140e-01 4.00886744e-01
6.67053223e-01 -3.71312231e-01 -8.59281003e-01 4.38883960e-01
3.71365577e-01 -8.29938233e-01 7.17581332e-01 -5.73491275e-01
3.42085570e-01 -4.68158901e-01 -2.88979918e-01 -1.42719853e+00
-4.65677708e-01 -1.18972331e-01 -2.75863498e-01 1.49403095e+00
4.46730703e-01 -9.66256142e-01 5.05642891e-01 7.13130355e-01
-9.74974334e-02 -7.45170534e-01 -8.69424880e-01 -1.23409712e+00
3.23168933e-01 -2.51319557e-01 6.53124630e-01 1.42555821e+00
-2.24059924e-01 2.92513758e-01 -7.25608408e-01 2.74855673e-01
5.22776723e-01 4.28872764e-01 9.19261158e-01 -1.37523234e+00
-5.45409262e-01 -5.52713946e-02 1.28598809e-01 -1.24009764e+00
3.75344664e-01 -8.95765722e-01 7.21190572e-02 -1.09100258e+00
4.70825195e-01 -9.12793458e-01 -6.47650361e-01 8.05063248e-01
-5.02319515e-01 1.07960910e-01 1.00990430e-01 7.18997538e-01
-7.94719040e-01 6.43881202e-01 1.04957700e+00 1.60083640e-02
-4.01536226e-01 2.97385696e-02 -9.35135841e-01 7.34074950e-01
6.91506028e-01 -6.03227258e-01 -5.83538711e-01 -2.91296214e-01
-2.76094556e-01 -3.31900567e-01 1.59804866e-01 -9.94360983e-01
-5.94481044e-02 -4.05156493e-01 2.75625676e-01 -2.97610722e-02
4.71638292e-02 -9.21611309e-01 5.86640127e-02 3.32151234e-01
-5.33005536e-01 -7.00971007e-01 2.53877211e-02 7.68283546e-01
-3.18270355e-01 -2.62724519e-01 1.25211406e+00 1.54103771e-01
-7.54282057e-01 3.98641616e-01 2.73676336e-01 2.34734759e-01
1.04875791e+00 3.35101262e-02 -1.46617308e-01 -2.26020813e-01
-6.86731637e-01 3.62759829e-01 7.37898171e-01 4.59490538e-01
2.04249367e-01 -1.67133653e+00 -6.42042458e-01 1.57547757e-01
5.68522334e-01 9.61430296e-02 3.35943252e-01 7.02013254e-01
3.42108727e-01 1.37337267e-01 -1.64992791e-02 -6.61194384e-01
-1.02428234e+00 7.97637105e-01 2.48664454e-01 -5.82088530e-01
-2.71962941e-01 9.17152226e-01 6.10481203e-01 -9.16049480e-01
1.29301637e-01 -2.63408087e-02 -1.39555484e-01 -2.04764202e-01
5.72285354e-01 1.67867050e-01 -2.37753153e-01 -9.29424942e-01
-3.02738965e-01 6.06896996e-01 -3.89669746e-01 1.93902612e-01
9.96738791e-01 -3.10032755e-01 3.53288092e-02 6.63432419e-01
1.12829816e+00 -7.83266202e-02 -1.52446175e+00 -1.02224636e+00
-8.23718309e-02 -4.48165148e-01 -8.71348605e-02 -9.76478040e-01
-1.15615368e+00 5.87987185e-01 5.63726008e-01 -1.23448804e-01
1.28040850e+00 1.00392990e-01 7.97311187e-01 1.50019720e-01
3.87490928e-01 -1.37008739e+00 8.45774785e-02 4.28893119e-01
7.18470812e-01 -1.56339657e+00 -1.11265033e-01 -4.56110924e-01
-1.00478518e+00 7.01493800e-01 9.12774682e-01 3.56980339e-02
7.16753662e-01 -2.64993221e-01 1.13232315e-01 3.35514158e-01
-7.28335142e-01 -1.05416067e-01 3.43509823e-01 7.30334282e-01
2.69890785e-01 1.64220527e-01 -1.04116455e-01 1.22433448e+00
4.28065062e-01 3.09971780e-01 6.75257146e-02 9.33184683e-01
-2.00223669e-01 -1.25022256e+00 -2.43680283e-01 4.07928675e-01
-1.23672737e-02 1.31245241e-01 -3.58438194e-01 7.06868947e-01
3.68657082e-01 9.58349466e-01 -1.59834713e-01 -5.79045534e-01
3.02388757e-01 2.39673346e-01 7.85766691e-02 -8.40545952e-01
-2.64342070e-01 4.67253327e-02 -2.33848155e-01 -3.60081404e-01
-4.96868104e-01 -6.93073511e-01 -1.21026993e+00 6.46542311e-02
-4.72718239e-01 2.87304014e-01 2.97017395e-01 1.02112114e+00
5.09007215e-01 1.45519152e-01 1.00359988e+00 -5.43428719e-01
-8.51178348e-01 -1.08492446e+00 -7.18520284e-01 7.44458199e-01
2.39611328e-01 -8.99770498e-01 -4.82346863e-01 2.81507969e-01] | [10.329150199890137, 3.1033544540405273] |
e2c92e30-3779-489a-9447-55c7de02b903 | transductive-zero-shot-hashing-for-multi | 1911.07192 | null | https://arxiv.org/abs/1911.07192v2 | https://arxiv.org/pdf/1911.07192v2.pdf | Transductive Zero-Shot Hashing for Multilabel Image Retrieval | Hash coding has been widely used in approximate nearest neighbor search for large-scale image retrieval. Given semantic annotations such as class labels and pairwise similarities of the training data, hashing methods can learn and generate effective and compact binary codes. While some newly introduced images may contain undefined semantic labels, which we call unseen images, zeor-shot hashing techniques have been studied. However, existing zeor-shot hashing methods focus on the retrieval of single-label images, and cannot handle multi-label images. In this paper, for the first time, a novel transductive zero-shot hashing method is proposed for multi-label unseen image retrieval. In order to predict the labels of the unseen/target data, a visual-semantic bridge is built via instance-concept coherence ranking on the seen/source data. Then, pairwise similarity loss and focal quantization loss are constructed for training a hashing model using both the seen/source and unseen/target data. Extensive evaluations on three popular multi-label datasets demonstrate that, the proposed hashing method achieves significantly better results than the competing methods. | ['Song Wang', 'Qin Zou', 'Long Chen', 'Ling Cao', 'Zheng Zhang'] | 2019-11-17 | null | null | null | null | ['multi-label-image-retrieval'] | ['computer-vision'] | [ 1.12199455e-01 -3.05991352e-01 -6.15183353e-01 -6.38999879e-01
-1.32981479e+00 -3.30664575e-01 4.54269141e-01 5.19823253e-01
-2.84766495e-01 5.17106652e-01 1.53743222e-01 6.34403408e-01
-2.75427192e-01 -6.64953113e-01 -4.31551874e-01 -9.45536852e-01
3.11963111e-01 7.07203746e-01 3.04181069e-01 2.01950267e-01
4.24097419e-01 1.11036666e-01 -2.03117180e+00 3.78570378e-01
1.51171297e-01 1.11003196e+00 1.48507342e-01 2.12880701e-01
6.00473061e-02 7.54720747e-01 -2.33288676e-01 -2.62803525e-01
1.88679114e-01 -2.69877017e-01 -8.33951354e-01 -6.97413459e-02
7.45432615e-01 -5.86812079e-01 -5.71999192e-01 1.19170880e+00
7.20313787e-01 3.24859738e-01 8.76096964e-01 -1.61455441e+00
-1.14842463e+00 3.35708261e-01 -3.90105724e-01 -3.55081111e-02
4.93224710e-01 -3.41880679e-01 1.38055110e+00 -1.19156432e+00
7.46290803e-01 1.42383647e+00 5.72289705e-01 4.52635974e-01
-1.07909632e+00 -8.76000404e-01 -3.96398872e-01 5.34421265e-01
-2.14411569e+00 -3.21734995e-01 6.36955380e-01 -2.43418962e-01
4.55256492e-01 3.95708382e-02 3.37573022e-01 7.78321147e-01
-1.17363855e-01 8.93859029e-01 9.48062360e-01 -3.80617082e-01
3.35351616e-01 2.92012036e-01 1.00278303e-01 7.07885861e-01
2.04785634e-02 6.02393374e-02 -7.23535478e-01 -5.02732873e-01
2.36666486e-01 6.62568867e-01 -2.12390840e-01 -6.96887612e-01
-1.18609476e+00 1.22512972e+00 7.63296068e-01 2.59957850e-01
-9.82908830e-02 2.98567533e-01 5.91231644e-01 6.87083080e-02
2.55499721e-01 4.22439799e-02 1.77369803e-01 3.03189069e-01
-1.17250264e+00 2.30309457e-01 4.20191854e-01 1.13065958e+00
1.29476070e+00 -7.16463983e-01 -3.90632182e-01 1.04017055e+00
5.05326450e-01 7.67759144e-01 8.22038829e-01 -7.73309410e-01
1.00403540e-01 4.78206128e-01 5.78077473e-02 -1.36839199e+00
-1.71226971e-02 2.08485916e-01 -7.41317391e-01 -4.25620437e-01
-1.67687476e-01 8.43335032e-01 -1.05600417e+00 1.55878866e+00
5.31441152e-01 5.22171915e-01 8.48887861e-02 1.09392416e+00
9.31784689e-01 8.96460414e-01 1.71770185e-01 -1.85011283e-01
1.33566427e+00 -1.09723341e+00 -8.42977703e-01 8.39856453e-03
7.47961640e-01 -9.37326372e-01 7.47741818e-01 -4.05727793e-03
-7.75695503e-01 -5.65628707e-01 -1.03965867e+00 -4.82903004e-01
-6.65721178e-01 -1.79407820e-01 2.27722406e-01 4.25748110e-01
-9.75200236e-01 3.46869409e-01 -3.58643383e-01 -4.07796741e-01
3.54836613e-01 1.73524559e-01 -4.53605413e-01 -9.82438803e-01
-1.44119751e+00 6.34264231e-01 8.20274055e-01 -3.83802056e-01
-1.28503883e+00 -3.91580284e-01 -1.07470572e+00 1.21500880e-01
1.35254771e-01 -2.37273440e-01 1.02281010e+00 -4.39447433e-01
-5.50301552e-01 1.18592858e+00 -1.19399138e-01 -2.08973140e-01
-3.15358609e-01 1.69462368e-01 -3.52437675e-01 5.76760948e-01
5.28467238e-01 1.25358880e+00 8.59494030e-01 -1.29661834e+00
-6.57299161e-01 -4.00516808e-01 -1.61098853e-01 4.03023988e-01
-6.26014590e-01 -1.87006608e-01 -1.99662432e-01 -6.15153551e-01
4.83812869e-01 -9.99786258e-01 2.84464985e-01 3.53180915e-01
-1.17949367e-01 -4.17729348e-01 1.03223658e+00 -2.68200666e-01
1.07578552e+00 -2.55751348e+00 -1.13434149e-02 1.10905126e-01
1.22893997e-01 7.86153525e-02 -2.47632444e-01 5.00749171e-01
1.09041356e-01 -1.83721140e-01 -7.29207098e-02 -5.16049385e-01
2.52207130e-01 3.48884553e-01 -4.05839682e-01 6.12615466e-01
-2.76496857e-01 9.21731889e-01 -1.09387887e+00 -1.14773750e+00
1.99622825e-01 6.64838374e-01 -2.62017876e-01 2.47916564e-01
1.91964760e-01 -1.34911746e-01 -4.59668398e-01 9.50461209e-01
5.87630987e-01 -5.59468329e-01 -3.44337523e-01 -3.57226789e-01
3.55562240e-01 -2.29536042e-01 -1.11637950e+00 2.02770877e+00
-2.29636654e-01 3.10079157e-01 -4.54085469e-01 -7.91892111e-01
9.07899916e-01 5.63863158e-01 4.59168255e-01 -7.90944576e-01
3.22659165e-02 5.06778836e-01 -9.23016667e-01 -2.90873468e-01
7.98529088e-01 -5.40549457e-01 -2.90406555e-01 6.01590991e-01
2.76765317e-01 -1.29758164e-01 -1.23416968e-02 3.62310618e-01
7.14605689e-01 -4.05062407e-01 3.59082282e-01 4.19839099e-02
4.80068684e-01 5.69519512e-02 4.46190476e-01 5.72550178e-01
-4.61734861e-01 1.02326977e+00 -3.19565505e-01 -5.64468086e-01
-1.24993706e+00 -1.10429978e+00 -4.63834465e-01 1.16953731e+00
9.13390279e-01 -5.31747818e-01 -5.56209981e-01 -4.61310536e-01
2.00580075e-01 2.80062467e-01 -5.56392789e-01 -6.36489391e-01
-8.15286580e-03 -3.28949541e-01 6.73342466e-01 2.06600025e-01
4.77277368e-01 -1.04653955e+00 -2.91232467e-01 1.69482872e-01
-5.14020443e-01 -9.42298532e-01 -7.95326591e-01 -6.52025733e-03
-5.87886870e-01 -9.22551990e-01 -9.05332088e-01 -1.36984682e+00
5.50366282e-01 8.04287732e-01 7.22386837e-01 2.72944868e-01
-7.42677569e-01 4.68647659e-01 -6.81593776e-01 2.45894268e-01
-7.44898468e-02 -2.43619800e-01 7.80027211e-02 1.45491019e-01
7.25258887e-01 -1.63938060e-01 -9.11435843e-01 6.43735588e-01
-1.18561971e+00 -3.90850812e-01 5.59670687e-01 1.14028692e+00
1.03887546e+00 2.62210313e-02 4.80327100e-01 -4.62949663e-01
2.10176408e-01 -6.20458543e-01 -1.98109701e-01 5.77910841e-01
-8.36961985e-01 2.05874786e-01 2.67482102e-01 -5.28440893e-01
-4.77590114e-01 1.09856457e-01 2.91420668e-01 -8.22087109e-01
-2.95742676e-02 2.22677499e-01 1.68791190e-01 -3.18615973e-01
4.74552304e-01 4.01289701e-01 -2.06271231e-01 -2.17855096e-01
5.86863399e-01 1.08671355e+00 4.22311783e-01 -3.90944034e-01
6.48772657e-01 7.74143815e-01 -1.08292945e-01 -3.41289014e-01
-1.03945673e+00 -1.24478483e+00 -4.78256494e-01 4.89314608e-02
1.05813360e+00 -1.20538688e+00 -3.95573735e-01 3.36685956e-01
-9.72684562e-01 3.89725715e-01 -2.70084471e-01 5.15325546e-01
-7.13310897e-01 5.03232539e-01 -7.57341683e-01 -5.00582576e-01
-4.12320048e-01 -1.25500965e+00 1.80909753e+00 1.53250784e-01
3.72746252e-02 -8.40887606e-01 3.33312213e-01 4.72238123e-01
2.99155593e-01 -1.39812902e-01 8.85810614e-01 -7.05994844e-01
-7.24404573e-01 -5.18679023e-01 -5.71801782e-01 1.76774815e-01
7.70825744e-02 -5.24721742e-01 -1.05384278e+00 -7.48062909e-01
-6.90950528e-02 -1.16492355e+00 7.80786216e-01 2.53035687e-02
9.25340950e-01 -3.90057355e-01 -4.89048183e-01 2.64467627e-01
1.76831007e+00 1.81837939e-02 5.82842112e-01 4.74080369e-02
6.84939444e-01 4.05662358e-01 1.17234099e+00 6.32029235e-01
5.51688612e-01 6.74613655e-01 4.22838241e-01 2.33789250e-01
-1.24237530e-01 -5.64462960e-01 -5.34897037e-02 8.71053457e-01
8.03652942e-01 -2.90189207e-01 -7.06320643e-01 7.69582033e-01
-1.73756135e+00 -1.00639153e+00 3.57564211e-01 2.27165627e+00
9.30217028e-01 -2.53189683e-01 -3.09074372e-01 2.18310460e-01
1.20631051e+00 3.21421653e-01 -4.20634151e-01 4.00704741e-02
-3.02209817e-02 -8.20911527e-02 3.35416466e-01 3.49243164e-01
-1.27264261e+00 8.75735164e-01 5.70916176e+00 1.32418907e+00
-9.33802903e-01 3.87092322e-01 4.53413397e-01 -9.13830251e-02
-2.78190702e-01 1.89060032e-01 -1.07011819e+00 4.42460388e-01
7.89764702e-01 -1.69560879e-01 2.08757937e-01 1.05854297e+00
-4.28400755e-01 6.24028370e-02 -1.19004261e+00 1.51274431e+00
6.10201955e-01 -1.01823235e+00 4.90771770e-01 1.85458467e-03
7.78547466e-01 -2.03691885e-01 2.61566937e-01 3.56125116e-01
-1.11847250e-02 -9.97738719e-01 6.29702687e-01 4.24879909e-01
1.09154952e+00 -6.89774454e-01 8.18503201e-01 1.52322188e-01
-1.46981132e+00 -1.82014957e-01 -7.94065595e-01 3.10452968e-01
9.90639627e-02 4.38325137e-01 -6.32639289e-01 2.76848286e-01
7.10891604e-01 1.01808929e+00 -6.48807347e-01 1.09775448e+00
1.39969155e-01 5.28341755e-02 -1.15239158e-01 1.57568417e-02
4.52652782e-01 1.85017481e-01 4.09623096e-03 7.68851519e-01
4.35710669e-01 1.73642039e-01 4.52390641e-01 5.87593138e-01
-1.86531752e-01 3.94546360e-01 -7.82703280e-01 3.26951817e-02
9.00349498e-01 1.22983885e+00 -5.28412342e-01 -4.99237418e-01
-3.23048115e-01 1.12029886e+00 3.41736883e-01 1.72775820e-01
-7.28106022e-01 -4.42690581e-01 3.67238969e-01 -2.24336803e-01
3.26881796e-01 2.38905445e-01 3.91554803e-01 -1.02725160e+00
-1.12832725e-01 -5.41366994e-01 7.79961407e-01 -1.03315198e+00
-1.32316971e+00 2.85970777e-01 -9.64785088e-03 -1.41411722e+00
-1.54686779e-01 -1.37613848e-01 1.94194000e-02 4.50267494e-01
-1.59696949e+00 -1.27440464e+00 -5.39708555e-01 7.69588292e-01
4.50694263e-01 -1.17774529e-03 1.13195717e+00 7.25626647e-01
9.17929038e-02 6.88340843e-01 5.72200358e-01 9.27363038e-02
1.17278457e+00 -9.79181826e-01 -1.60042763e-01 2.31254339e-01
3.61485928e-01 5.09154797e-01 4.53451753e-01 -4.93081033e-01
-1.11225736e+00 -1.11704016e+00 1.04429638e+00 1.81140732e-02
4.53517854e-01 -2.77639508e-01 -9.87274408e-01 5.46194494e-01
-1.13030441e-01 6.08579159e-01 8.79849792e-01 -4.01630700e-01
-9.76185024e-01 -2.23072231e-01 -1.50771940e+00 1.84976026e-01
6.50196135e-01 -1.07594454e+00 -6.60411417e-01 5.96780300e-01
9.21542704e-01 -3.20564993e-02 -1.10821390e+00 3.07606936e-01
4.82831955e-01 -7.39152551e-01 1.19592381e+00 -1.18812352e-01
2.56304681e-01 -4.56169695e-01 -7.53217161e-01 -9.66719925e-01
-4.18136805e-01 2.25989074e-01 2.73786664e-01 1.15850270e+00
-7.27424920e-02 -2.92935103e-01 6.25271738e-01 3.46539825e-01
1.78208217e-01 -5.07843614e-01 -1.15110934e+00 -7.86874175e-01
-1.41319185e-01 3.09290320e-01 5.60477138e-01 1.04226661e+00
-8.10207203e-02 2.43852749e-01 -5.94921470e-01 1.29689604e-01
1.03298962e+00 3.65837961e-01 1.40246868e-01 -1.21270669e+00
6.21885620e-02 8.35643336e-02 -1.04086411e+00 -8.89617085e-01
2.47736722e-01 -1.23494625e+00 3.79064888e-01 -1.26960027e+00
8.25969040e-01 -7.13522255e-01 -6.31705999e-01 6.28592849e-01
-1.06908582e-01 8.81027102e-01 2.41223965e-02 8.01150262e-01
-1.31202173e+00 9.03300941e-01 9.61458921e-01 -5.12716472e-01
5.48260272e-01 -6.40121520e-01 -2.67864257e-01 7.11809099e-02
4.10104007e-01 -9.33968246e-01 -5.18794537e-01 -1.65615514e-01
1.67797819e-01 9.65641066e-02 5.45126855e-01 -1.07961452e+00
5.74042201e-01 8.40659961e-02 -1.71031822e-02 -6.44190967e-01
6.24361336e-01 -9.18024302e-01 1.31575316e-01 3.60373825e-01
-6.86537564e-01 -1.35499775e-01 -4.08185452e-01 9.03069615e-01
-7.45721936e-01 -6.42210305e-01 9.90701497e-01 -1.02109551e-01
-8.34450066e-01 5.68156421e-01 2.46909916e-01 1.46184415e-01
1.16786516e+00 -2.34336689e-01 -3.66609454e-01 -2.51675963e-01
-4.72631007e-01 2.48796895e-01 7.05148697e-01 4.83351558e-01
1.13971055e+00 -1.99458957e+00 -5.19644499e-01 6.82262853e-02
8.56736302e-01 -6.70349076e-02 3.84055972e-01 3.09455633e-01
-3.04863989e-01 5.13296664e-01 -1.34004846e-01 -6.66456342e-01
-1.33513260e+00 1.15768230e+00 -4.48916890e-02 8.89967978e-02
-3.56152683e-01 8.16928446e-01 2.08971590e-01 -3.85187119e-01
3.39742720e-01 2.32131824e-01 -3.45794261e-02 3.03429514e-01
7.24349499e-01 2.87726581e-01 -1.38729930e-01 -1.06939554e+00
-2.85050392e-01 9.55356061e-01 -3.18596929e-01 6.87005445e-02
8.72284353e-01 -5.57564974e-01 -3.93609792e-01 7.61157036e-01
2.11150694e+00 -7.20155776e-01 -7.12825954e-01 -6.74467862e-01
4.55633700e-02 -7.96978951e-01 2.14827303e-02 -1.40012711e-01
-9.26298380e-01 9.72514510e-01 1.08969760e+00 -9.99784246e-02
9.97045994e-01 2.00107634e-01 1.34079385e+00 4.65072364e-01
7.32988894e-01 -1.10510635e+00 5.85598052e-01 9.09085050e-02
3.99154902e-01 -1.64599669e+00 -6.16695620e-02 -2.54757315e-01
-3.58020395e-01 7.56654561e-01 2.67338604e-01 2.19971389e-02
6.65744185e-01 -3.67239177e-01 -5.61196320e-02 -4.00218397e-01
-5.86589158e-01 -3.89040969e-02 2.05988422e-01 3.93817008e-01
-2.32133642e-02 -4.02223729e-02 -1.71061501e-01 -4.86173853e-02
2.44939223e-01 -8.45981240e-02 1.41157582e-01 1.07534063e+00
-9.77681100e-01 -9.46740568e-01 -4.96460736e-01 3.24350357e-01
-1.23909064e-01 -2.52652943e-01 -4.64675948e-02 3.71252030e-01
5.23424633e-02 8.92620862e-01 8.25179294e-02 -3.78124535e-01
-1.93031147e-01 2.02115685e-01 1.86613426e-01 -7.52638221e-01
-1.11126997e-01 -1.80665731e-01 -6.41080439e-01 -3.84720743e-01
-5.38165033e-01 -4.51329559e-01 -1.42835939e+00 -1.01139262e-01
-6.49031520e-01 4.55945641e-01 6.23192787e-01 6.86219215e-01
2.15762973e-01 -3.49317998e-01 8.83045495e-01 -7.05131888e-01
-7.59677112e-01 -6.27191186e-01 -1.03453696e+00 9.97726381e-01
3.04616779e-01 -9.99852896e-01 -7.33103156e-01 2.79170144e-02] | [11.375194549560547, 1.0109940767288208] |
113a68b0-9d5f-4560-a96a-42d76d0596af | pavementscapes-a-large-scale-hierarchical | 2208.00775 | null | https://arxiv.org/abs/2208.00775v1 | https://arxiv.org/pdf/2208.00775v1.pdf | Pavementscapes: a large-scale hierarchical image dataset for asphalt pavement damage segmentation | Pavement damage segmentation has benefited enormously from deep learning. % and large-scale datasets. However, few current public datasets limit the potential exploration of deep learning in the application of pavement damage segmentation. To address this problem, this study has proposed Pavementscapes, a large-scale dataset to develop and evaluate methods for pavement damage segmentation. Pavementscapes is comprised of 4,000 images with a resolution of $1024 \times 2048$, which have been recorded in the real-world pavement inspection projects with 15 different pavements. A total of 8,680 damage instances are manually labeled with six damage classes at the pixel level. The statistical study gives a thorough investigation and analysis of the proposed dataset. The numeral experiments propose the top-performing deep neural networks capable of segmenting pavement damages, which provides the baselines of the open challenge for pavement inspection. The experiment results also indicate the existing problems for damage segmentation using deep learning, and this study provides potential solutions. | ['Weiguang Zhang', 'Ju Huyan', 'Tao Ma', 'Zheng Tong'] | 2022-07-24 | null | null | null | null | ['2048'] | ['playing-games'] | [-2.08379373e-01 1.55203685e-01 5.17988861e-01 -2.79313177e-01
-8.22931111e-01 -1.02316447e-01 -1.79300755e-01 2.49404937e-01
-1.62471831e-01 6.51739597e-01 5.45378029e-03 -1.49044201e-01
-1.47630140e-01 -1.61622298e+00 -7.31580079e-01 -1.03384209e+00
-6.47731602e-01 3.28665614e-01 3.13934684e-01 -4.10424262e-01
4.01961625e-01 7.43322670e-01 -1.59135044e+00 4.81569588e-01
9.24807370e-01 1.18342352e+00 -2.29509547e-03 8.25962961e-01
4.28490281e-01 4.73004580e-01 -1.09537232e+00 -3.45101923e-01
4.90470231e-01 3.50555420e-01 -1.27722597e+00 7.22211301e-02
6.76928103e-01 -1.18698430e+00 -6.65956199e-01 8.21235120e-01
7.99803972e-01 7.30423480e-02 7.59096622e-01 -1.18356621e+00
-7.77968943e-01 4.51109380e-01 -8.03931057e-01 5.02566278e-01
-4.90425766e-01 2.18751565e-01 9.69236910e-01 -6.30186498e-01
4.60884869e-02 1.61766934e+00 8.79699051e-01 3.79321091e-02
-4.06661391e-01 -4.06746045e-02 -2.31946498e-01 5.37794828e-01
-1.16320205e+00 3.11906010e-01 6.32728159e-01 -5.96199512e-01
9.11623836e-01 2.58005887e-01 4.94434118e-01 5.84179699e-01
2.86527723e-01 1.00856590e+00 7.06591785e-01 -2.05527432e-02
1.99159216e-02 -3.89255553e-01 4.47152555e-01 6.81588113e-01
6.06437802e-01 2.97262073e-01 5.01578689e-01 1.23165831e-01
7.00435400e-01 -1.96831226e-01 1.63978025e-01 3.73814136e-01
-5.26653707e-01 9.00055766e-01 8.13540936e-01 1.50322169e-01
-5.58142722e-01 3.38455379e-01 8.24652493e-01 3.24944973e-01
5.83678246e-01 2.98238069e-01 -5.26973367e-01 3.85986380e-02
-6.08711958e-01 4.54774708e-01 5.47374129e-01 4.57683831e-01
7.88737118e-01 3.49563867e-01 -1.31942943e-01 1.24678433e+00
3.09475362e-01 3.44170243e-01 -4.21333343e-01 -1.07135844e+00
6.74829304e-01 5.89516342e-01 1.77968800e-01 -1.26349080e+00
-6.89479232e-01 -9.02710669e-03 -1.02173972e+00 6.34645939e-01
4.33436662e-01 -5.79698205e-01 -1.10344481e+00 7.75655508e-01
2.20108554e-01 -9.73732695e-02 -1.23469345e-01 7.00625777e-01
8.86834502e-01 8.97255301e-01 1.35418087e-01 5.76319516e-01
1.25185740e+00 -9.35217619e-01 -2.55062044e-01 -4.75292243e-02
7.72878528e-01 -4.96543914e-01 9.65419888e-01 7.14332938e-01
-8.67662787e-01 -6.37342811e-01 -1.10397947e+00 1.47493169e-01
-7.18343914e-01 2.75388628e-01 4.19786066e-01 8.67040515e-01
-7.39481390e-01 6.99100137e-01 -6.81758106e-01 -2.18693271e-01
1.14907253e+00 -7.76719898e-02 -1.06587276e-01 -3.07537049e-01
-1.41664648e+00 1.20545447e+00 4.62162465e-01 7.19043255e-01
-1.84916151e+00 -5.91236711e-01 -7.12952375e-01 1.64179027e-01
-5.81868477e-02 9.97309834e-02 9.62752044e-01 -1.71017647e-01
-6.32871330e-01 9.42319870e-01 1.07762527e+00 -5.07406712e-01
3.89617473e-01 -7.18023479e-01 -4.38908070e-01 4.30011392e-01
-1.54608944e-02 6.67387724e-01 3.90075535e-01 -1.68277407e+00
-9.03460145e-01 -2.10565910e-01 4.34448242e-01 -3.39037538e-01
-4.37932253e-01 1.00029139e-02 1.75973058e-01 -6.25198126e-01
-2.87584543e-01 -7.19575226e-01 -4.80628520e-01 -2.05111921e-01
-6.67774022e-01 -1.44688696e-01 1.03760302e+00 -1.15637565e+00
1.21112680e+00 -1.93612099e+00 -5.56119323e-01 1.84892446e-01
2.08138637e-02 6.59443557e-01 -3.59892100e-01 7.57868230e-01
-3.62186521e-01 1.45995080e-01 -7.75331438e-01 -9.28605497e-02
1.19976357e-01 5.32809079e-01 -1.01507038e-01 6.13555968e-01
2.85968900e-01 4.93085861e-01 -5.56026101e-01 -5.01873970e-01
2.69933224e-01 -1.19678795e-01 -2.87864894e-01 2.17450649e-01
1.56010270e-01 5.10452874e-02 -3.20797920e-01 1.12658083e+00
1.33471596e+00 6.97337449e-01 -9.10257876e-01 -2.60528058e-01
-2.72652090e-01 -2.53191292e-01 -1.00623477e+00 1.41484737e+00
-8.87908787e-02 6.90132201e-01 1.74268499e-01 -1.58613098e+00
1.16073263e+00 1.83551624e-01 8.47632945e-01 -3.69098574e-01
4.11778152e-01 1.29764572e-01 -6.17740154e-02 -1.21949673e+00
6.93278611e-01 1.39542595e-01 -6.79721236e-01 2.43024111e-01
-1.84312567e-01 -2.35846624e-01 4.54708070e-01 5.75490296e-02
1.24182332e+00 -4.37841773e-01 -6.41018927e-01 -2.01874197e-01
1.78335056e-01 2.20498636e-01 4.45853055e-01 5.74522316e-01
-5.04411459e-01 6.96181536e-01 7.48575866e-01 -8.38813305e-01
-1.01932573e+00 -1.02168798e+00 -5.15100360e-01 8.86251032e-01
2.13853363e-02 6.18753195e-01 -1.25381458e+00 -6.26548946e-01
4.43874478e-01 3.39216083e-01 -9.46653187e-01 -3.50578099e-01
-7.36306429e-01 -1.49580252e+00 1.42981279e+00 8.66020441e-01
9.14115250e-01 -1.45115924e+00 -5.29487669e-01 4.05558705e-01
-1.46898285e-01 -6.65101349e-01 2.39200875e-01 -2.60908514e-01
-8.05778444e-01 -1.46270669e+00 -7.68084407e-01 -1.09427142e+00
4.80448633e-01 2.95267940e-01 1.12291121e+00 3.55484605e-01
-8.32329273e-01 2.02898845e-01 -4.43622977e-01 -5.57252765e-01
-5.51837564e-01 8.11127126e-02 -3.84953052e-01 -3.13328832e-01
3.95272434e-01 -5.76615453e-01 -8.36437702e-01 3.63109916e-01
-9.82729852e-01 -8.97631466e-01 8.97267342e-01 6.26927197e-01
1.88125089e-01 4.03255343e-01 1.46848178e+00 -5.11809587e-01
7.38459647e-01 -8.68310630e-01 -3.31895977e-01 9.39460173e-02
-5.30744493e-01 -6.51993036e-01 8.09939578e-02 4.18336391e-01
-1.35180283e+00 -5.15567899e-01 -8.67596805e-01 1.63435459e-01
-8.50689054e-01 8.93852711e-01 -3.28670323e-01 9.83691439e-02
7.19351530e-01 -6.12778127e-01 -1.57551527e-01 -9.96533573e-01
3.95601392e-01 7.93555439e-01 8.50638568e-01 -1.06316078e+00
4.64001596e-01 3.04515630e-01 -1.30375087e-01 -8.41686964e-01
-5.38520396e-01 -3.04804623e-01 -7.48791456e-01 -3.59979808e-01
9.05152917e-01 -6.88733220e-01 -4.77327436e-01 1.54069924e+00
-1.28378665e+00 -2.54347295e-01 -5.07367730e-01 3.21401432e-02
-1.78061202e-01 7.58797586e-01 -1.15858781e+00 -4.87357438e-01
-6.88280761e-01 -9.83932436e-01 8.11912656e-01 6.90688100e-03
5.11749029e-01 -6.21034503e-01 4.49550837e-01 6.78406477e-01
1.11403495e-01 9.50668275e-01 1.06875670e+00 -2.41415933e-01
-3.44870448e-01 -7.52470434e-01 -5.90671778e-01 1.23430717e+00
1.43471703e-01 4.59735751e-01 -1.06258726e+00 -2.27168694e-01
-5.37783980e-01 -5.95357835e-01 1.19584584e+00 7.27817059e-01
1.61585927e+00 -1.68571260e-03 -1.12284124e-01 5.88849485e-02
1.44033623e+00 3.30728859e-01 1.46491122e+00 5.37558138e-01
9.70294833e-01 9.58795011e-01 9.27346945e-01 7.99675524e-01
5.33375084e-01 -1.96433917e-01 1.50424421e+00 -6.13436341e-01
2.12457534e-02 2.01437160e-01 -7.04846606e-02 6.79744244e-01
-3.47273737e-01 -4.75391328e-01 -1.24197137e+00 1.48969829e+00
-1.27251768e+00 -1.12150776e+00 -5.67893744e-01 1.56755650e+00
5.72414517e-01 3.67616937e-02 1.66552335e-01 6.75005913e-01
7.20648408e-01 4.09229368e-01 -5.94731748e-01 -8.61193538e-01
-1.57781150e-02 3.55393797e-01 6.39431119e-01 1.57589617e-03
-1.52202332e+00 5.57731509e-01 6.36733246e+00 8.41759562e-01
-3.73084337e-01 -1.41083866e-01 7.68056750e-01 1.77892745e-01
1.43284246e-01 -4.84021842e-01 -5.70424616e-01 3.24811250e-01
9.02609408e-01 5.55594563e-01 -4.23461229e-01 9.61350441e-01
4.03296947e-01 -2.20797554e-01 -5.83031595e-01 1.86970368e-01
-2.59602606e-01 -1.13148022e+00 -2.17114657e-01 8.74036327e-02
1.04990625e+00 2.61030734e-01 -4.69733868e-03 4.21664149e-01
4.51384157e-01 -9.69512343e-01 3.53466928e-01 3.61451536e-01
4.59523618e-01 -1.10056019e+00 1.11763752e+00 4.50426564e-02
-1.05270827e+00 -5.24278700e-01 -6.72079206e-01 -1.33603830e-02
4.80614930e-01 6.98651135e-01 -5.94840825e-01 5.41568577e-01
1.27994287e+00 8.23548198e-01 -4.47378397e-01 1.51249349e+00
-9.36490670e-02 9.32633340e-01 -4.19326462e-02 6.77884221e-01
6.95006430e-01 -9.74469930e-02 2.32526988e-01 1.39926457e+00
3.48167449e-01 -2.10487857e-01 1.58160672e-01 4.69182909e-01
-9.90118608e-02 -3.55454445e-01 -3.45128685e-01 2.14841023e-01
3.79242897e-01 1.00326169e+00 -3.46633226e-01 -1.25335053e-01
-3.56427699e-01 1.98499039e-01 -3.27152282e-01 2.21817091e-01
-1.01850498e+00 -9.45064962e-01 1.03392756e+00 -1.05587795e-01
3.99181813e-01 -9.48682576e-02 -5.80228627e-01 -3.70306939e-01
-1.16913378e-01 -5.92071772e-01 4.95505512e-01 -6.05887949e-01
-1.67583132e+00 7.82012045e-02 3.71983171e-01 -1.30024946e+00
8.05574775e-01 -7.10368872e-01 -1.20390022e+00 5.17722130e-01
-1.72925353e+00 -1.41753733e+00 -6.58040941e-01 3.42128515e-01
7.39999056e-01 -5.95846213e-02 4.74615335e-01 1.01450980e+00
-1.06852388e+00 4.53924626e-01 3.01036000e-01 6.10976219e-01
4.75557864e-01 -1.31094718e+00 8.76635730e-01 9.73266363e-01
-6.39921904e-01 -3.39212745e-01 4.92138743e-01 -5.19886672e-01
-5.11795342e-01 -1.50303113e+00 7.23732859e-02 -1.30004108e-01
4.38200951e-01 4.35746163e-01 -1.34553802e+00 4.96261865e-01
3.85020941e-01 -9.64738950e-02 5.17917037e-01 -3.34979802e-01
1.67315871e-01 -2.89723635e-01 -1.64511085e+00 -1.12063326e-01
6.70662344e-01 1.48397190e-02 -6.92145288e-01 3.09021622e-01
5.23232281e-01 -4.19001251e-01 -1.51943016e+00 4.24742728e-01
3.96348506e-01 -7.35278130e-01 1.31539690e+00 -4.81934994e-01
9.19753373e-01 2.35688668e-02 -9.77178365e-02 -1.39748693e+00
-2.39855483e-01 2.37324446e-01 3.48381475e-02 1.42340696e+00
-3.18383425e-02 -4.79822159e-01 8.04271758e-01 3.62848729e-01
-1.30396092e+00 -9.08681989e-01 -8.91448379e-01 -5.36410928e-01
8.16996932e-01 -5.11110187e-01 1.11777222e+00 6.89577699e-01
-5.66896915e-01 -6.25275075e-01 -3.71191502e-01 7.05394626e-01
6.88037276e-01 -1.57761097e-01 4.31209594e-01 -1.40965760e+00
4.50221181e-01 -4.79523510e-01 -5.74278176e-01 -9.97141123e-01
-1.85211211e-01 -2.50004232e-01 4.46815521e-01 -2.04868174e+00
-5.08873641e-01 -7.84505606e-01 -5.51735759e-01 7.83893287e-01
1.63039133e-01 4.40570302e-02 -4.41001415e-01 -2.53882468e-01
1.45377830e-01 6.06523991e-01 1.54250312e+00 -8.16649675e-01
2.84512341e-01 2.39758804e-01 -5.28070211e-01 6.63466811e-01
1.03525972e+00 -5.45208871e-01 -2.23481029e-01 -8.48856986e-01
-6.59049451e-02 -2.05039725e-01 5.67980587e-01 -1.10141110e+00
-3.47886056e-01 -2.01580256e-01 -1.28330052e-01 -1.23143637e+00
-4.31408472e-02 -5.86764395e-01 -4.42230225e-01 6.06969833e-01
-3.90716083e-02 -2.25908294e-01 5.81525326e-01 4.82182533e-01
-5.51935017e-01 -5.04083633e-01 9.75263417e-01 -2.31753960e-01
-1.23940396e+00 6.38705432e-01 -7.27680564e-01 6.41842932e-02
1.10052121e+00 -3.03317040e-01 -8.20253074e-01 3.58201921e-01
-6.19856656e-01 3.59175175e-01 -1.28993586e-01 4.79680479e-01
8.77603233e-01 -1.27241623e+00 -1.10809386e+00 -3.79650481e-02
-5.05446941e-02 5.55268884e-01 8.95168066e-01 4.54809427e-01
-1.23256338e+00 4.14709002e-02 -8.63714814e-01 -3.48622859e-01
-8.61806452e-01 1.43423647e-01 4.66171056e-01 -2.64081478e-01
-5.60122192e-01 1.12125194e+00 -1.29721120e-01 -1.44888178e-01
-4.12070788e-02 -7.47898459e-01 -6.80163383e-01 4.25135523e-01
2.80136049e-01 1.35540974e+00 5.22468507e-01 -5.98381698e-01
-7.48101473e-02 3.10050994e-01 -9.74376034e-03 6.36598289e-01
1.69111848e+00 -1.11879088e-01 -3.31440002e-01 -3.16054702e-01
1.25665343e+00 -8.21565807e-01 -1.56105399e+00 1.85767129e-01
-6.16901964e-02 -6.21870279e-01 1.81557849e-01 -9.04990911e-01
-1.86801827e+00 1.40112078e+00 9.68723834e-01 5.98683774e-01
1.16080117e+00 -3.42624068e-01 1.89346814e+00 4.16897535e-01
3.00399661e-01 -1.40988052e+00 3.33061397e-01 5.40069580e-01
1.19537592e+00 -1.40439093e+00 -2.11017892e-01 1.61111802e-01
-5.61181784e-01 9.69872534e-01 9.93256032e-01 -5.61427891e-01
7.86697745e-01 1.94086894e-01 5.99058867e-02 -7.82621741e-01
2.01365203e-02 -5.16025983e-02 -1.92697421e-01 8.59786153e-01
-2.23420456e-01 2.29186296e-01 -6.64550662e-02 6.09265625e-01
1.90847889e-01 -3.01815957e-01 7.32391059e-01 1.20719767e+00
-1.14959216e+00 -8.78327012e-01 -5.91536283e-01 5.80273569e-01
-4.22784954e-01 2.76381344e-01 1.51318744e-01 8.43753994e-01
6.97534978e-01 1.16597760e+00 -1.93403855e-01 -3.29116702e-01
7.97500968e-01 -7.46170163e-01 1.23976037e-01 -4.52381402e-01
-7.28457332e-01 -6.27347052e-01 4.07407165e-01 -1.33028060e-01
-1.32319987e-01 -5.68228066e-01 -1.16223156e+00 -4.74485397e-01
-2.40757450e-01 -2.21051164e-02 4.31609213e-01 9.62283909e-01
3.76103595e-02 6.96247756e-01 9.41941977e-01 -1.00118101e+00
-5.27915597e-01 -1.26913261e+00 -1.11378407e+00 3.44400227e-01
3.67397875e-01 -8.99131000e-01 -2.95225978e-01 2.15837196e-01] | [7.412198066711426, 1.2404476404190063] |
6e4cfd0a-7cb9-463c-a4b6-68df83af8901 | camera-adversarial-transfer-for-unsupervised | 1904.01308 | null | https://arxiv.org/abs/1904.01308v2 | https://arxiv.org/pdf/1904.01308v2.pdf | CANU-ReID: A Conditional Adversarial Network for Unsupervised person Re-IDentification | Unsupervised person re-ID is the task of identifying people on a target data set for which the ID labels are unavailable during training. In this paper, we propose to unify two trends in unsupervised person re-ID: clustering & fine-tuning and adversarial learning. On one side, clustering groups training images into pseudo-ID labels, and uses them to fine-tune the feature extractor. On the other side, adversarial learning is used, inspired by domain adaptation, to match distributions from different domains. Since target data is distributed across different camera viewpoints, we propose to model each camera as an independent domain, and aim to learn domain-independent features. Straightforward adversarial learning yields negative transfer, we thus introduce a conditioning vector to mitigate this undesirable effect. In our framework, the centroid of the cluster to which the visual sample belongs is used as conditioning vector of our conditional adversarial network, where the vector is permutation invariant (clusters ordering does not matter) and its size is independent of the number of clusters. To our knowledge, we are the first to propose the use of conditional adversarial networks for unsupervised person re-ID. We evaluate the proposed architecture on top of two state-of-the-art clustering-based unsupervised person re-identification (re-ID) methods on four different experimental settings with three different data sets and set the new state-of-the-art performance on all four of them. Our code and model will be made publicly available at https://team.inria.fr/perception/canu-reid/. | ['Xavier Alameda-Pineda', 'Stephane Lathuilière', 'Guillaume Delorme', 'Yihong Xu', 'Radu Horaud'] | 2019-04-02 | null | null | null | null | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 1.32148191e-01 -1.19747803e-01 1.82578340e-01 -3.97100180e-01
-3.40064794e-01 -8.65534544e-01 9.96693730e-01 -1.06938146e-01
-7.12872565e-01 6.01678193e-01 2.71607548e-01 3.15033257e-01
3.24301198e-02 -7.32748330e-01 -6.68965340e-01 -6.64745390e-01
4.43291441e-02 9.66815233e-01 5.08331414e-03 -4.08754684e-02
-7.23806471e-02 4.71209317e-01 -1.59627616e+00 1.78949609e-01
7.62609243e-01 4.81652498e-01 -2.22293392e-01 7.09803760e-01
3.38675022e-01 5.13013899e-01 -7.99565196e-01 -6.81445539e-01
5.69417894e-01 -4.72211689e-01 -9.08209503e-01 1.17242098e-01
5.78178823e-01 -1.12145156e-01 -5.68174601e-01 1.18887627e+00
6.48661554e-01 4.80686933e-01 1.05528378e+00 -1.43075299e+00
-9.58873272e-01 4.53333288e-01 -4.20409590e-01 3.49400775e-03
4.34510708e-01 4.13995869e-02 4.98751134e-01 -6.53489530e-01
6.21419251e-01 1.18419123e+00 7.25595057e-01 1.04658926e+00
-1.50687516e+00 -7.76951194e-01 1.36104301e-01 3.38545799e-01
-1.72069716e+00 -4.18885052e-01 7.63464630e-01 -6.08582258e-01
5.02720058e-01 1.89531133e-01 4.23396051e-01 1.52731395e+00
-4.49758381e-01 4.63029355e-01 1.22299778e+00 -4.98540342e-01
2.15329438e-01 4.68665957e-01 1.14708886e-01 2.49396339e-01
9.00041163e-02 3.05955708e-01 -2.21383855e-01 1.84049970e-03
5.56485415e-01 2.26270899e-01 -3.27663347e-02 -6.80516303e-01
-1.20549560e+00 7.93160439e-01 5.44726908e-01 2.39409149e-01
-1.11741096e-01 8.11213069e-03 3.48020941e-01 3.60842288e-01
1.24143295e-01 3.29432845e-01 5.29129691e-02 2.46418536e-01
-7.34781384e-01 4.51558292e-01 7.23439336e-01 8.99490893e-01
7.43674219e-01 -3.80188286e-01 -2.62889832e-01 1.03348005e+00
5.62113747e-02 7.59793103e-01 6.70114338e-01 -8.47421169e-01
2.19500333e-01 3.90915841e-01 3.13114613e-01 -8.85393381e-01
-3.87638003e-01 -1.91102013e-01 -1.06505597e+00 3.63662183e-01
8.16134572e-01 -2.43480831e-01 -1.08901548e+00 1.96230316e+00
1.45602092e-01 2.93199122e-01 1.43347055e-01 9.43368614e-01
7.33142674e-01 2.37568304e-01 1.37562841e-01 2.33796567e-01
1.34718275e+00 -8.52349818e-01 -3.17685425e-01 -2.40730941e-01
8.28884467e-02 -4.90425766e-01 7.43334293e-01 3.12537968e-01
-6.90790117e-01 -9.93636549e-01 -9.60746884e-01 2.13467613e-01
-7.62415826e-01 3.02418262e-01 1.34785414e-01 8.82717788e-01
-1.05924284e+00 3.98578823e-01 -5.69890916e-01 -7.70387650e-01
2.89144933e-01 5.26389599e-01 -7.14466393e-01 -1.83812335e-01
-1.19098973e+00 7.92898834e-01 4.72702265e-01 -2.87204266e-01
-1.04676259e+00 -4.44922209e-01 -6.23973668e-01 -1.19653061e-01
2.06186533e-01 -8.71931612e-01 8.47703338e-01 -1.40790486e+00
-1.30152869e+00 1.35793090e+00 -1.41736791e-01 -4.99590218e-01
6.85631216e-01 -6.07122667e-02 -6.96422756e-01 4.71428260e-02
3.43747288e-01 6.88505828e-01 1.20328963e+00 -1.70480168e+00
-5.68341494e-01 -5.53277075e-01 2.98516992e-02 9.44663435e-02
-3.95725071e-01 -1.47294002e-02 -6.80021822e-01 -7.91023433e-01
-4.27732974e-01 -1.32003129e+00 -1.24692179e-01 -4.72307086e-01
-5.48283994e-01 -2.22715214e-01 4.97970730e-01 -5.99767148e-01
8.29421163e-01 -2.22641397e+00 3.97765458e-01 4.77693081e-01
2.93730140e-01 3.08809429e-01 -1.75605416e-01 3.21592480e-01
-5.38882554e-01 1.14939958e-02 -3.01335037e-01 -6.33379698e-01
1.45602182e-01 2.09858734e-03 -3.06594744e-03 6.64257944e-01
-7.15931877e-02 8.38065147e-01 -7.99644470e-01 -2.90755600e-01
4.43984598e-01 4.25417453e-01 -4.27283645e-01 4.86515164e-01
2.53372014e-01 7.19506800e-01 -1.37968689e-01 2.99519718e-01
9.13106322e-01 5.79426177e-02 2.24097148e-01 -2.28960052e-01
1.57678634e-01 -3.98469299e-01 -1.48495126e+00 1.44580436e+00
-1.30090520e-01 2.54137695e-01 -1.91085279e-01 -1.13412106e+00
7.69039392e-01 9.67780426e-02 4.26707268e-01 -4.77397263e-01
2.52617132e-02 -2.12613195e-01 -1.67121634e-01 -2.10347816e-01
2.85895556e-01 -3.40165086e-02 -3.54463249e-01 4.37897801e-01
2.53302693e-01 4.38672185e-01 2.10886627e-01 2.01673701e-01
8.67752433e-01 -2.79248022e-02 2.29674950e-01 -1.63165197e-01
8.27310145e-01 -6.66355118e-02 3.48831564e-01 1.01699901e+00
-4.27296072e-01 7.84166098e-01 1.55147299e-01 -3.35028678e-01
-1.19734108e+00 -1.49448514e+00 -7.52895558e-03 1.32614243e+00
2.85610557e-01 -1.03652716e-01 -9.70558405e-01 -9.77265179e-01
1.79800302e-01 5.51203012e-01 -1.02041221e+00 -1.36666402e-01
-5.90698540e-01 -6.53616965e-01 7.76676238e-01 4.80478287e-01
4.86615002e-01 -1.03042722e+00 1.02422886e-01 -5.55307269e-02
-2.87682563e-01 -1.01045012e+00 -4.97109294e-01 3.70562598e-02
-1.80493161e-01 -1.28672719e+00 -1.06164610e+00 -8.88816476e-01
8.98438275e-01 1.56181619e-01 1.17372489e+00 -1.80334121e-01
-1.53254062e-01 7.92788208e-01 -5.49234152e-01 -2.61373371e-01
-5.38608015e-01 -5.17556779e-02 6.32742703e-01 3.76405180e-01
7.94429362e-01 -6.02940679e-01 -6.11614227e-01 5.52560568e-01
-7.23979652e-01 -4.06257331e-01 2.09274009e-01 9.21135247e-01
4.83413696e-01 2.74812132e-01 3.73175263e-01 -1.04484272e+00
4.57800597e-01 -3.50791723e-01 -3.30388486e-01 1.61696732e-01
-2.27533355e-01 -4.82634045e-02 8.96259010e-01 -6.21783614e-01
-1.03856647e+00 2.81807870e-01 -1.52497077e-02 -4.87484425e-01
-8.13724935e-01 -9.76708904e-02 -3.73332977e-01 -7.89878983e-03
8.74950707e-01 1.35737658e-01 -1.45829961e-01 -3.96998048e-01
6.73895538e-01 6.80741251e-01 1.04676342e+00 -4.45892721e-01
1.29281235e+00 7.75520682e-01 -3.82345527e-01 -5.70710301e-01
-4.22330201e-01 -6.87662423e-01 -1.13979304e+00 -1.80213198e-01
1.17522478e+00 -1.06269884e+00 -7.24395096e-01 7.23227680e-01
-8.20644557e-01 -3.55369747e-01 -3.99796873e-01 3.73949766e-01
-5.28739452e-01 5.05381942e-01 -2.12529197e-01 -5.51106989e-01
-1.23418733e-01 -8.41936290e-01 6.96373999e-01 2.62013137e-01
-2.66610742e-01 -9.32094634e-01 3.04903954e-01 3.45375240e-01
8.68350342e-02 3.48131478e-01 3.79351109e-01 -9.62105811e-01
-1.10497631e-01 -3.45377356e-01 -1.89900190e-01 5.21907926e-01
2.42650732e-01 -4.45059210e-01 -1.24803865e+00 -6.56195045e-01
-3.29930276e-01 -2.06785738e-01 8.87651682e-01 1.83836818e-01
1.16892529e+00 -1.60936385e-01 -6.09365344e-01 7.93059766e-01
1.30017173e+00 -1.05448134e-01 6.59896731e-01 4.95946109e-01
1.06843615e+00 7.86621630e-01 3.42672080e-01 4.37938035e-01
5.65653682e-01 8.67568910e-01 1.66549236e-01 -1.72278464e-01
-2.44683638e-01 -3.34282726e-01 3.18308979e-01 1.04186371e-01
-4.51878101e-01 -2.72405863e-01 -8.67397249e-01 6.54097557e-01
-1.78030539e+00 -1.29893816e+00 2.66680509e-01 2.46712041e+00
4.81974214e-01 -1.86804637e-01 6.77735686e-01 -1.72608290e-02
1.05298829e+00 -1.57521412e-01 -4.84257132e-01 1.32328076e-02
-5.16548045e-02 6.92771971e-02 6.93835795e-01 5.55486917e-01
-1.53768277e+00 1.02019107e+00 5.43805695e+00 6.46809578e-01
-7.18825877e-01 2.34435678e-01 2.71364063e-01 5.03527299e-02
1.61257982e-01 -3.65729839e-01 -8.03839445e-01 6.88715279e-01
6.71392381e-01 -6.73624873e-02 7.51825690e-01 7.68803537e-01
-1.34648964e-01 2.62955576e-01 -1.46717119e+00 1.25184786e+00
3.66662264e-01 -7.94495523e-01 -2.46816427e-02 3.46943401e-02
8.56585562e-01 -7.80544132e-02 1.63802594e-01 3.40435207e-01
7.71854639e-01 -1.13177168e+00 5.60356140e-01 6.36277616e-01
9.77289319e-01 -1.06193447e+00 8.31612885e-01 1.08489163e-01
-1.04296947e+00 -2.48854876e-01 -3.45250279e-01 1.96078345e-01
-1.01719666e-02 1.50197729e-01 -6.66265666e-01 4.82957482e-01
1.14068532e+00 7.43602693e-01 -8.47345769e-01 6.31180048e-01
-1.98414505e-01 3.54030222e-01 -1.84301317e-01 4.29083586e-01
-2.38856554e-01 -5.70567101e-02 5.77116013e-01 1.40830398e+00
8.87184124e-03 -7.93977976e-02 3.39138627e-01 7.10655332e-01
-1.83697373e-01 -2.60454923e-01 -6.02755368e-01 3.58838528e-01
4.51342791e-01 1.07154632e+00 -4.90235686e-01 -4.41327751e-01
-3.86521190e-01 1.59379268e+00 4.83961940e-01 5.11817753e-01
-8.76580536e-01 -3.08975399e-01 9.20319974e-01 1.29487291e-01
4.18832868e-01 -7.21720383e-02 -1.85111109e-02 -1.23788917e+00
-2.07292810e-01 -9.64393973e-01 6.80428505e-01 -3.58798295e-01
-1.96343780e+00 4.55599606e-01 1.38599798e-01 -1.30428457e+00
-4.81547654e-01 -5.56984961e-01 -4.35884774e-01 9.33440268e-01
-1.27604914e+00 -1.39626753e+00 -6.00281537e-01 1.22532618e+00
6.67366087e-02 -6.40938461e-01 9.77394760e-01 3.79687190e-01
-4.16344285e-01 1.18598270e+00 3.55607450e-01 6.96736634e-01
1.30551994e+00 -1.48209691e+00 4.87685144e-01 1.01536465e+00
-9.06117447e-03 7.30854332e-01 6.06076777e-01 -5.47549009e-01
-9.59439754e-01 -1.33874869e+00 5.69638789e-01 -8.82553935e-01
4.50476468e-01 -6.08639896e-01 -6.66372597e-01 9.19745684e-01
2.25168336e-02 -9.41492692e-02 8.98023546e-01 1.11492172e-01
-5.61084986e-01 -2.38780946e-01 -1.43683910e+00 5.24340868e-01
1.12120366e+00 -5.75298786e-01 -5.79291880e-01 3.57473105e-01
3.03540528e-01 -1.34929433e-01 -1.01408744e+00 1.05437309e-01
4.15811449e-01 -1.08044410e+00 1.52077329e+00 -4.41629052e-01
3.65987904e-02 -6.21633291e-01 -4.58762422e-02 -1.54722381e+00
-8.04632366e-01 -2.34339193e-01 1.62726656e-01 1.61565137e+00
-4.88865897e-02 -8.19764316e-01 8.16804051e-01 4.86284912e-01
2.68622994e-01 2.47982487e-01 -7.52885938e-01 -8.19490373e-01
1.93454176e-01 -1.39256597e-01 6.99979544e-01 1.08123314e+00
-3.02984446e-01 1.72822267e-01 -6.42625630e-01 5.37285864e-01
1.05106449e+00 -9.66152176e-02 1.26207578e+00 -1.21264422e+00
-4.55555439e-01 -2.66194850e-01 -5.24527669e-01 -7.59369135e-01
4.01607394e-01 -1.00156617e+00 -5.34184203e-02 -1.06849694e+00
4.40313578e-01 -4.99128550e-01 -2.64462650e-01 4.84012157e-01
-2.01310366e-01 6.77676618e-01 4.61312383e-01 4.11828518e-01
-5.54525375e-01 2.84097701e-01 7.62234747e-01 -4.77249712e-01
-2.41007924e-01 5.74398339e-02 -8.20297956e-01 7.00921357e-01
8.54606569e-01 -4.44420993e-01 -1.52843565e-01 -2.12160751e-01
-3.92870456e-01 -6.23312414e-01 7.48896241e-01 -1.34422302e+00
3.91349912e-01 9.18106660e-02 1.01594019e+00 -2.32602477e-01
3.62366050e-01 -1.01669037e+00 2.84498870e-01 2.42490798e-01
-2.48590410e-01 -1.64995879e-01 4.26363992e-03 5.85629940e-01
-3.31857353e-02 -1.64312795e-01 1.07861209e+00 -2.85204232e-01
-1.01924753e+00 3.83793086e-01 -1.97624907e-01 1.38212308e-01
1.15777767e+00 -2.05992982e-01 -1.72931001e-01 -3.65821093e-01
-9.91771221e-01 1.35006636e-01 8.86843145e-01 5.55913508e-01
3.89473230e-01 -1.40101969e+00 -9.64099705e-01 3.91506940e-01
4.06717390e-01 -3.75870347e-01 4.38283443e-01 1.71950281e-01
-1.63233861e-01 1.42439470e-01 -5.31455755e-01 -6.04141295e-01
-1.37842178e+00 1.11355889e+00 4.84354794e-01 -3.18558306e-01
-3.84319752e-01 6.78481698e-01 4.18603808e-01 -8.33522081e-01
2.20125645e-01 5.17811894e-01 -4.57674950e-01 8.35388005e-02
6.76497340e-01 5.60842633e-01 -2.62777299e-01 -9.94966805e-01
-5.71477771e-01 7.45447755e-01 -1.54600710e-01 -9.30491909e-02
1.02102244e+00 -3.08111370e-01 -8.18872675e-02 5.81676848e-02
1.17005765e+00 2.54090242e-02 -1.11355805e+00 -2.66909242e-01
-3.58949274e-01 -3.86703312e-01 -5.81091166e-01 -8.83744419e-01
-7.80642271e-01 4.21201497e-01 1.12304342e+00 1.27581879e-01
1.14794517e+00 1.61837503e-01 4.17951047e-01 2.35480711e-01
2.63395786e-01 -1.18441033e+00 1.33580714e-01 4.44892675e-01
6.68825626e-01 -1.43787098e+00 -1.39099747e-01 -2.80824691e-01
-8.09878349e-01 6.36023641e-01 6.73829377e-01 -4.11792994e-01
4.84610856e-01 -1.91189330e-02 2.05589682e-01 1.76870689e-01
1.93855330e-01 -5.60507178e-01 3.10464233e-01 1.39573467e+00
1.50944097e-02 4.39035177e-01 1.46126106e-01 6.28332555e-01
-4.02518421e-01 -8.68366882e-02 2.92143971e-01 3.62988919e-01
7.43281394e-02 -1.33991349e+00 -8.22394073e-01 1.65150955e-01
-3.05706948e-01 1.03581615e-01 -5.43801129e-01 7.67903149e-01
5.43921351e-01 1.02550364e+00 1.90319851e-01 -5.99011838e-01
4.93549764e-01 -4.06692214e-02 4.99925911e-01 -4.53007370e-01
-6.44125819e-01 -4.79479879e-01 -1.58971116e-01 -3.95833313e-01
-6.01741076e-01 -8.08963358e-01 -8.44746590e-01 -5.27160525e-01
2.43884861e-01 1.23070888e-02 2.40039423e-01 5.90147853e-01
2.30263218e-01 3.63196403e-01 7.34383225e-01 -1.11024106e+00
-3.47644091e-01 -9.45393920e-01 -5.74452460e-01 1.16521275e+00
2.68222272e-01 -7.40470052e-01 -2.55720437e-01 4.47630405e-01] | [14.722857475280762, 1.0470106601715088] |
0f8a97e8-33df-47fe-8b97-9d27451f6f66 | iterative-optimization-of-pseudo-ground-truth | 2208.14683 | null | https://arxiv.org/abs/2208.14683v1 | https://arxiv.org/pdf/2208.14683v1.pdf | Iterative Optimization of Pseudo Ground-Truth Face Image Quality Labels | While recent face recognition (FR) systems achieve excellent results in many deployment scenarios, their performance in challenging real-world settings is still under question. For this reason, face image quality assessment (FIQA) techniques aim to support FR systems, by providing them with sample quality information that can be used to reject poor quality data unsuitable for recognition purposes. Several groups of FIQA methods relying on different concepts have been proposed in the literature, all of which can be used for generating quality scores of facial images that can serve as pseudo ground-truth (quality) labels and can be exploited for training (regression-based) quality estimation models. Several FIQA appro\-aches show that a significant amount of sample-quality information can be extracted from mated similarity-score distributions generated with some face matcher. Based on this insight, we propose in this paper a quality label optimization approach, which incorporates sample-quality information from mated-pair similarities into quality predictions of existing off-the-shelf FIQA techniques. We evaluate the proposed approach using three state-of-the-art FIQA methods over three diverse datasets. The results of our experiments show that the proposed optimization procedure heavily depends on the number of executed optimization iterations. At ten iterations, the approach seems to perform the best, consistently outperforming the base quality scores of the three FIQA methods, chosen for the experiments. | ['Vitomir Štruc', 'Žiga Babnik'] | 2022-08-31 | null | null | null | null | ['face-image-quality', 'face-image-quality-assessment'] | ['computer-vision', 'computer-vision'] | [ 1.68347523e-01 -1.17031559e-01 -8.40295628e-02 -1.00289035e+00
-9.82917130e-01 -2.65561551e-01 5.36203265e-01 -7.19331726e-02
-1.33001477e-01 5.90339601e-01 -5.74377999e-02 2.60378808e-01
-4.43143398e-01 -6.71775699e-01 -4.43832874e-01 -7.66805828e-01
-5.32455444e-02 7.10760891e-01 -7.53259733e-02 -3.74052614e-01
3.42885494e-01 8.63901317e-01 -2.32082033e+00 5.93979836e-01
9.91965890e-01 1.51866949e+00 -2.57266968e-01 4.13852453e-01
-1.96206614e-01 4.44288999e-01 -7.82106817e-01 -8.89354229e-01
5.31892776e-01 -5.09976625e-01 -6.80259228e-01 8.79840776e-02
8.33878517e-01 -1.44291490e-01 5.53161949e-02 1.01329494e+00
5.56440234e-01 -1.95298538e-01 7.20767140e-01 -1.41776037e+00
-2.22422674e-01 8.40706006e-02 -2.02634528e-01 1.27319396e-01
6.20649338e-01 3.38762879e-01 1.08189690e+00 -1.12684155e+00
5.24083972e-01 1.56335032e+00 4.62442875e-01 6.70756698e-01
-1.27137887e+00 -7.84928262e-01 -4.16642368e-01 5.24344504e-01
-1.41102993e+00 -1.02243567e+00 7.36483157e-01 -3.43984038e-01
5.59117794e-01 3.56175005e-01 4.45323616e-01 7.39900529e-01
1.70429014e-02 6.53825760e-01 1.37403083e+00 -4.42399502e-01
3.48186880e-01 2.75352299e-01 -9.42024663e-02 9.33864236e-01
-6.81934739e-03 2.56928772e-01 -8.69948566e-01 -3.47840220e-01
3.31435949e-01 -5.12325823e-01 -1.93871260e-01 -4.67349976e-01
-6.05895340e-01 8.43789577e-01 4.43707198e-01 4.82292473e-01
-4.45701241e-01 -1.07733585e-01 4.08191197e-02 7.14386106e-01
4.85645473e-01 2.08191574e-01 -1.45969406e-01 9.58091840e-02
-1.18551409e+00 2.28592858e-01 7.40514815e-01 4.73414153e-01
1.00710320e+00 4.43685539e-02 -5.57625830e-01 9.65979993e-01
5.71071744e-01 7.51553476e-01 1.98930249e-01 -9.67014432e-01
2.38316849e-01 6.59731805e-01 8.65996033e-02 -1.06954992e+00
-2.01406747e-01 -3.33242059e-01 -5.54843128e-01 5.86200774e-01
5.82101047e-01 3.55276525e-01 -9.14855599e-01 1.55977488e+00
3.19538981e-01 1.74137250e-01 -3.72988433e-02 8.71082604e-01
6.99188471e-01 4.43557382e-01 3.17891017e-02 -4.78652090e-01
1.17607248e+00 -4.74879980e-01 -6.13004267e-01 2.10522935e-01
2.62613386e-01 -1.03441083e+00 9.82908309e-01 8.09617460e-01
-1.04398096e+00 -8.90918016e-01 -1.03627872e+00 5.74150383e-01
-6.97892979e-02 3.24118733e-01 3.39993834e-01 1.40970278e+00
-1.36664951e+00 7.38170803e-01 -4.10261035e-01 -1.91586047e-01
7.31336176e-01 6.05941892e-01 -6.29644513e-01 -4.05270308e-01
-8.18372905e-01 7.76890934e-01 1.72868408e-02 9.55474749e-02
-1.21846318e+00 -5.45378327e-01 -4.65244800e-01 3.03288363e-02
3.02348644e-01 -4.65059668e-01 9.79660630e-01 -1.48392916e+00
-1.59918237e+00 1.07366192e+00 -1.96163356e-01 -2.01009914e-01
4.04605120e-01 -6.26264289e-02 -6.62503123e-01 4.99659598e-01
-1.70117691e-01 6.62116885e-01 1.26405203e+00 -1.46539044e+00
-5.04901171e-01 -6.68242216e-01 -2.85701565e-02 -7.07771629e-02
-2.94973284e-01 3.28731865e-01 -4.76514369e-01 -3.99896771e-01
-6.46298006e-02 -6.50468409e-01 2.63018131e-01 3.04533780e-01
5.57410009e-02 -4.10261154e-01 4.71338630e-01 -4.90488678e-01
1.12795496e+00 -1.90566158e+00 1.80103660e-01 5.83710134e-01
-1.19702779e-01 6.09041691e-01 -4.94745404e-01 1.25459656e-01
5.23199327e-02 -3.16115282e-02 -1.41748756e-01 -4.90672767e-01
1.43836834e-03 2.03231335e-01 1.46021307e-01 4.85914648e-01
7.23093867e-01 5.02601445e-01 -6.59173906e-01 -6.30103230e-01
2.35378742e-01 5.35463452e-01 -6.43266737e-01 6.75049663e-01
-1.98074505e-01 4.34780598e-01 -3.72585326e-01 8.50593328e-01
8.88036907e-01 1.30903408e-01 5.12848012e-02 -5.02160192e-01
4.15267110e-01 -1.99804112e-01 -1.50596929e+00 1.40827775e+00
-3.82162869e-01 2.30306923e-01 3.75768133e-02 -9.63511705e-01
1.19524026e+00 3.20215762e-01 6.51732326e-01 -8.97755504e-01
2.63316989e-01 3.78858387e-01 1.03773996e-01 -3.94509524e-01
1.75079048e-01 -2.92166620e-01 3.80620986e-01 3.44864249e-01
5.30560553e-01 -9.38385352e-02 3.82443398e-01 -1.82552099e-01
1.00290263e+00 3.07000410e-02 8.23130831e-02 -3.00303370e-01
1.28153765e+00 -3.88146043e-01 4.82645988e-01 4.37837064e-01
-3.21196288e-01 5.78571737e-01 3.34852755e-01 -3.82077336e-01
-1.02774310e+00 -1.02908969e+00 -3.08798075e-01 8.56020689e-01
-2.04569697e-01 -2.35985398e-01 -1.01622701e+00 -7.96993554e-01
-4.65925299e-02 2.27284357e-01 -6.39001667e-01 -8.77129361e-02
-3.66851628e-01 -6.33534193e-01 5.68865657e-01 -2.65528541e-02
3.82394165e-01 -1.21010792e+00 -3.56533885e-01 5.38678914e-02
-8.95381421e-02 -8.75361502e-01 5.05546369e-02 -2.77998716e-01
-7.51419365e-01 -1.26524544e+00 -6.65108860e-01 -2.50256509e-01
5.64221978e-01 1.34293914e-01 1.40475404e+00 5.67866087e-01
-3.13271016e-01 5.03303230e-01 -5.51450312e-01 -1.40892893e-01
-7.33879626e-01 -2.20491141e-01 1.97905049e-01 6.46483183e-01
4.29385602e-01 -3.04467142e-01 -4.52173531e-01 6.99478149e-01
-9.66182351e-01 -6.33549511e-01 7.59608209e-01 9.58161414e-01
6.09742463e-01 -8.97169113e-03 7.19294488e-01 -7.87671506e-01
5.66239953e-01 -2.56399751e-01 -6.58902764e-01 5.88240087e-01
-8.41637731e-01 4.11279172e-01 5.47927737e-01 -5.95426336e-02
-1.09427071e+00 -1.30036667e-01 -5.67268372e-01 -6.06292486e-01
-2.20601201e-01 1.20178275e-01 -4.67168719e-01 -5.53102255e-01
7.64611423e-01 -4.37282436e-02 2.68412024e-01 -4.22320455e-01
1.20579161e-01 6.81035936e-01 3.68377119e-01 -6.44615352e-01
8.09905231e-01 3.40389699e-01 1.56001642e-01 -6.59233212e-01
-7.28493452e-01 -4.33492064e-01 -4.80437040e-01 -6.87159479e-01
4.74102229e-01 -5.90030789e-01 -9.18950975e-01 4.51675862e-01
-7.30132699e-01 1.50900811e-01 -1.35850802e-01 1.23254962e-01
-6.33490682e-01 4.30084944e-01 -1.01042993e-01 -1.16450417e+00
-3.78220141e-01 -1.43911040e+00 1.24536359e+00 2.70677865e-01
3.02116901e-01 -4.95386690e-01 3.33773568e-02 6.39612436e-01
4.48658049e-01 -3.66709158e-02 6.05774581e-01 -3.94628972e-01
-5.54558337e-01 -1.73543766e-01 -1.87835783e-01 7.01884508e-01
-1.08548170e-02 5.92218004e-02 -1.30050910e+00 -3.41693014e-01
-1.41409919e-01 -5.40347099e-01 6.46203697e-01 1.23613611e-01
1.07550967e+00 -1.28680542e-01 1.37481064e-01 4.82937604e-01
1.52261019e+00 -7.59170577e-02 9.14791405e-01 -3.32240537e-02
8.26742351e-02 7.62640059e-01 1.12014830e+00 5.29464483e-01
-7.85236210e-02 1.09818304e+00 6.46954656e-01 2.83361882e-01
-2.15638563e-01 7.20189810e-02 5.87656021e-01 2.47029349e-01
-1.00901105e-01 -1.82453766e-01 -6.96656287e-01 1.85762897e-01
-1.49726260e+00 -1.13726473e+00 1.25022486e-01 2.29428172e+00
6.43468857e-01 1.28448159e-01 3.44836056e-01 5.25818408e-01
4.24434215e-01 -4.86622751e-02 -3.17718506e-01 -4.01549548e-01
-1.04447529e-01 6.58413112e-01 -3.46326195e-02 3.09785396e-01
-8.54795277e-01 6.59689009e-01 6.19002771e+00 9.57950711e-01
-9.37827468e-01 2.21119504e-02 8.70391011e-01 2.35106185e-01
-2.86168754e-01 -2.91203886e-01 -7.63183057e-01 2.91442066e-01
1.08648562e+00 1.56009093e-01 4.81472045e-01 8.15314591e-01
1.34322897e-01 -2.49270201e-01 -1.10486937e+00 1.30614269e+00
3.66205692e-01 -9.37140584e-01 1.47178993e-01 5.33534363e-02
5.96787512e-01 -3.22263718e-01 2.34214813e-01 1.67580266e-02
2.03935076e-02 -1.24627090e+00 5.86560071e-01 8.68744254e-01
8.88899624e-01 -9.09973562e-01 1.10589933e+00 4.80380701e-03
-9.70682561e-01 -2.28712171e-01 -5.64234376e-01 4.41063821e-01
-1.98358193e-01 8.47578466e-01 -8.79544795e-01 8.80497217e-01
7.50894725e-01 2.27899536e-01 -8.06708992e-01 1.05215681e+00
-1.74128953e-02 5.45784771e-01 -2.48286143e-01 1.41653821e-01
-8.45649466e-02 -2.14654237e-01 3.79715979e-01 8.54694128e-01
4.36925799e-01 5.19310571e-02 -2.51173168e-01 7.27754176e-01
-6.72631860e-02 4.50299025e-01 -2.80285537e-01 3.49292196e-02
2.55621374e-01 1.39742005e+00 -4.44407314e-01 -1.30215645e-01
-2.43301332e-01 8.52583170e-01 2.09064230e-01 -2.14306056e-01
-6.56512856e-01 1.06014393e-01 7.31904149e-01 8.32419172e-02
2.64055371e-01 1.08278438e-01 1.56415910e-01 -9.55244184e-01
9.26208720e-02 -1.27337885e+00 5.61085045e-01 -6.61361814e-01
-1.50042248e+00 1.12810075e+00 -1.33039117e-01 -1.20330703e+00
-4.25097674e-01 -8.29580545e-01 -2.27572963e-01 7.39242375e-01
-1.75607681e+00 -1.00510120e+00 -5.93415022e-01 7.90928543e-01
3.95380437e-01 -5.30495048e-01 8.10024142e-01 5.30352175e-01
-2.76398301e-01 9.67903793e-01 -5.53505309e-02 -1.97410613e-01
7.88341045e-01 -9.75652754e-01 -2.21182644e-01 6.56942725e-01
4.83226597e-01 2.07509905e-01 8.20384800e-01 -3.13432872e-01
-1.45661521e+00 -7.39674687e-01 6.00753486e-01 -3.20858628e-01
8.48916844e-02 -7.94634968e-03 -8.57231021e-01 -2.29900345e-01
-1.41602799e-01 2.83646762e-01 7.82479584e-01 4.83129732e-02
-5.20941019e-01 -7.34358609e-01 -1.66179454e+00 7.16285314e-03
9.60251451e-01 -3.62909943e-01 -3.42916727e-01 -3.02678975e-03
-1.57266095e-01 1.61619186e-01 -1.01384485e+00 5.44940233e-01
6.62994742e-01 -1.54662490e+00 9.96425152e-01 -3.01980227e-01
5.05539812e-02 -3.30674827e-01 -2.34234497e-01 -1.25482619e+00
-9.33879763e-02 -2.06672251e-01 -2.09919829e-02 1.26383805e+00
1.72842354e-01 -2.80283213e-01 9.65964675e-01 4.45534140e-01
5.85180558e-02 -5.78820229e-01 -1.23767614e+00 -7.68393338e-01
-4.08728808e-01 -4.64159578e-01 8.30081403e-01 4.57657367e-01
-4.72319692e-01 5.06443717e-02 -3.64647150e-01 4.86574322e-02
8.48395109e-01 -1.06080018e-01 8.85192513e-01 -1.51802397e+00
-2.87606388e-01 -4.94134724e-01 -8.18431973e-01 -4.31524217e-01
2.54060656e-01 -7.26985812e-01 7.42994472e-02 -1.05263495e+00
5.64047247e-02 -7.03080714e-01 -4.33820963e-01 3.84099215e-01
-1.39164537e-01 7.39614487e-01 2.85499930e-01 1.64054409e-01
-5.72125852e-01 7.09478378e-01 1.02437603e+00 -1.39554784e-01
1.21911764e-01 9.05279368e-02 -2.99239576e-01 2.96019852e-01
5.77804863e-01 -3.90663803e-01 -3.22091907e-01 -2.39526033e-02
1.95992425e-01 1.35811895e-01 1.53318182e-01 -1.46474028e+00
-1.63569793e-01 -1.48601606e-01 5.02263963e-01 -2.73075581e-01
4.97606874e-01 -1.11076593e+00 3.42735529e-01 3.99167955e-01
-1.00072786e-01 -1.49380222e-01 -3.24236639e-02 2.29050890e-01
-4.58012849e-01 -4.04002756e-01 1.15980136e+00 1.22415781e-01
-6.13860786e-01 3.93057406e-01 3.52818929e-02 -3.43426824e-01
8.70972097e-01 -3.94887686e-01 2.37466879e-02 -4.52464998e-01
-6.78533852e-01 -1.29248038e-01 4.56756920e-01 3.98886412e-01
8.43517661e-01 -1.45846653e+00 -1.04867148e+00 4.66592014e-01
4.22730327e-01 -7.02200472e-01 3.01061600e-01 6.21147454e-01
-3.06535929e-01 3.23799044e-01 -6.79840982e-01 -7.67907560e-01
-1.62964153e+00 4.91207212e-01 4.34405625e-01 -2.47288331e-01
1.64602801e-01 7.51089633e-01 -2.39746615e-01 -1.99327856e-01
1.46110371e-01 1.87968567e-01 -5.02270103e-01 1.99216947e-01
8.17516923e-01 2.90767699e-01 6.29251659e-01 -1.03417552e+00
-3.71264279e-01 6.78046763e-01 1.72202304e-01 -1.20432347e-01
1.45155752e+00 4.31063548e-02 -2.95068808e-02 5.21507487e-02
1.04125273e+00 -7.38304704e-02 -9.47695255e-01 -2.12558597e-01
1.76699400e-01 -1.00350130e+00 1.38913002e-03 -9.29431975e-01
-1.40767467e+00 9.25794184e-01 1.31067252e+00 1.58065408e-02
1.72506642e+00 -1.16825558e-01 2.22750589e-01 2.42487922e-01
7.17847049e-01 -1.05846691e+00 4.94255155e-01 -1.08451888e-01
1.01684856e+00 -1.32784462e+00 -1.28459886e-01 -5.18643498e-01
-5.10747790e-01 1.28886914e+00 5.25041878e-01 6.59147725e-02
5.69793344e-01 -9.61124972e-02 3.75920609e-02 -2.29001760e-01
-6.87406063e-01 -4.00510550e-01 6.98002398e-01 7.63015270e-01
3.54073375e-01 -1.10208884e-01 -4.29752320e-01 3.26916724e-01
5.13145812e-02 1.55935451e-01 1.20635115e-01 4.81666923e-01
-6.51309192e-01 -1.65768158e+00 -6.40310526e-01 6.66417956e-01
-4.32061076e-01 3.28710973e-01 -2.69491225e-01 3.88719499e-01
3.82030904e-01 1.20287228e+00 -2.47188061e-01 -4.86501932e-01
4.29145306e-01 2.02123821e-01 8.07284355e-01 -3.15239757e-01
-7.96640575e-01 -1.69388115e-01 1.10370591e-01 -8.82448256e-01
-8.15631747e-01 -7.35235929e-01 -7.10642934e-01 -3.83486360e-01
-4.63529468e-01 2.51713961e-01 6.73455596e-01 8.44631732e-01
2.88102269e-01 7.41908550e-02 1.07454467e+00 -6.00551486e-01
-5.51757038e-01 -9.12359595e-01 -4.81913239e-01 8.03195357e-01
1.16895810e-02 -1.02038574e+00 -2.16160670e-01 -1.12021647e-01] | [13.068641662597656, 0.7640601992607117] |
6a04930a-494c-462a-b1ae-755d817fa358 | recognize-human-activities-from-partially | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Cao_Recognize_Human_Activities_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Cao_Recognize_Human_Activities_2013_CVPR_paper.pdf | Recognize Human Activities from Partially Observed Videos | Recognizing human activities in partially observed videos is a challenging problem and has many practical applications. When the unobserved subsequence is at the end of the video, the problem is reduced to activity prediction from unfinished activity streaming, which has been studied by many researchers. However, in the general case, an unobserved subsequence may occur at any time by yielding a temporal gap in the video. In this paper, we propose a new method that can recognize human activities from partially observed videos in the general case. Specifically, we formulate the problem into a probabilistic framework: 1) dividing each activity into multiple ordered temporal segments, 2) using spatiotemporal features of the training video samples in each segment as bases and applying sparse coding (SC) to derive the activity likelihood of the test video sample at each segment, and 3) finally combining the likelihood at each segment to achieve a global posterior for the activities. We further extend the proposed method to include more bases that correspond to a mixture of segments with different temporal lengths (MSSC), which can better represent the activities with large intra-class variations. We evaluate the proposed methods (SC and MSSC) on various real videos. We also evaluate the proposed methods on two special cases: 1) activity prediction where the unobserved subsequence is at the end of the video, and 2) human activity recognition on fully observed videos. Experimental results show that the proposed methods outperform existing state-of-the-art comparison methods. | ['Daniel Barrett', 'Yuewei Lin', 'Yu Cao', 'Haonan Yu', 'Aaron Michaux', 'Sven Dickinson', 'Song Wang', 'Jeffrey Mark Siskind', 'Siddharth Narayanaswamy', 'Andrei Barbu'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [ 6.76018596e-01 -2.96090245e-01 -2.95407832e-01 -3.65249217e-02
-4.77455497e-01 -2.84247071e-01 4.18566644e-01 -2.13028014e-01
-8.55114982e-02 6.29972160e-01 4.59499359e-01 1.79436252e-01
-8.67375061e-02 -3.33446115e-01 -7.46250033e-01 -1.02290320e+00
-3.32943618e-01 1.11496158e-01 6.55503809e-01 5.20305514e-01
1.64592683e-01 2.84853190e-01 -1.48316693e+00 4.10931855e-01
8.05702150e-01 1.10120952e+00 3.56110185e-01 5.61648011e-01
1.16779566e-01 1.25594592e+00 -6.33882582e-01 1.82648450e-01
1.92141280e-01 -8.44699144e-01 -5.60730517e-01 9.01092291e-01
5.56949899e-02 -4.17502254e-01 -4.39261705e-01 9.22173858e-01
-2.71478184e-02 4.37723309e-01 4.54831183e-01 -1.31140137e+00
-3.27389315e-02 2.33810455e-01 -8.95593643e-01 4.58088160e-01
6.81890368e-01 -2.56939568e-02 6.09461904e-01 -5.87731183e-01
3.52325380e-01 1.07915306e+00 2.65586555e-01 1.76538795e-01
-7.46736526e-01 -6.77688360e-01 4.89152104e-01 6.95212722e-01
-1.49078047e+00 -4.58013088e-01 8.53817821e-01 -5.65477431e-01
4.36696768e-01 -6.52809022e-03 8.25488508e-01 1.06729221e+00
2.23741736e-02 1.31040764e+00 7.85801947e-01 -1.22753106e-01
3.53300929e-01 -3.34414691e-01 -2.01582387e-01 5.51697135e-01
1.24022089e-01 -2.18683049e-01 -5.40091097e-01 -3.51912886e-01
7.49779165e-01 4.78723079e-01 -7.50131667e-01 -3.54598433e-01
-1.56185079e+00 4.81257141e-01 -1.83340684e-01 2.42675334e-01
-6.86181486e-01 -1.73709199e-01 2.30592668e-01 -2.82898247e-02
1.97130412e-01 -3.31218392e-01 -1.90118939e-01 -4.52638626e-01
-1.21768141e+00 1.39458075e-01 7.84453094e-01 9.00815129e-01
5.11871457e-01 1.13697939e-01 -2.26482138e-01 6.75237417e-01
2.87601769e-01 2.76071519e-01 5.49446166e-01 -1.09621823e+00
7.36413896e-01 3.08213651e-01 4.79358435e-01 -9.48420942e-01
9.98193622e-02 -4.05907243e-01 -7.60213256e-01 -3.34804535e-01
5.22688746e-01 -2.70823747e-01 -5.46204388e-01 1.49388468e+00
3.71777266e-01 1.09419060e+00 2.13456437e-01 8.90791416e-01
2.25052387e-01 1.12890649e+00 -7.92912245e-02 -8.69708836e-01
1.04559517e+00 -1.19051433e+00 -8.66817474e-01 -1.55634522e-01
3.43867868e-01 -5.80388665e-01 4.32241946e-01 7.62774110e-01
-8.29072654e-01 -6.46191955e-01 -1.00357378e+00 5.21412551e-01
4.01029944e-01 2.60677159e-01 2.09490722e-03 2.15045333e-01
-4.76246178e-01 4.56338018e-01 -1.23595786e+00 -3.12534750e-01
2.84509987e-01 3.55999693e-02 -3.33803594e-01 -3.63590330e-01
-8.55508864e-01 1.59502417e-01 5.08098781e-01 2.07829773e-01
-1.27122688e+00 -1.11170731e-01 -8.81289899e-01 2.03766733e-01
9.76972222e-01 -1.87687919e-01 1.09715259e+00 -1.29656732e+00
-1.17946780e+00 6.42372519e-02 -7.01241076e-01 -3.57545882e-01
4.66524124e-01 -2.97444195e-01 -6.21825874e-01 4.09379393e-01
1.08778283e-01 -1.50625105e-03 8.74776423e-01 -1.03623879e+00
-9.51930821e-01 -2.12157115e-01 -1.26395985e-01 5.00627100e-01
-1.11133881e-01 -4.63882014e-02 -8.76325786e-01 -8.89674842e-01
2.41199851e-01 -1.06761384e+00 7.48291016e-02 -1.14377640e-01
-1.78769484e-01 -2.12127432e-01 1.11685073e+00 -8.60617638e-01
1.25219631e+00 -2.48582888e+00 1.87220439e-01 6.68111742e-02
-2.09496304e-01 7.29366466e-02 7.77932629e-02 4.18820113e-01
1.87865168e-01 -4.32796657e-01 -3.68275791e-01 -3.12300354e-01
-4.15144205e-01 4.57624167e-01 -1.31737873e-01 5.75443923e-01
-5.44616729e-02 3.23223352e-01 -1.03938937e+00 -7.97257364e-01
1.95018381e-01 3.52916062e-01 -4.89646316e-01 4.29234445e-01
4.38568033e-02 7.83919275e-01 -6.34321570e-01 5.89514196e-01
5.79265118e-01 -3.69192839e-01 3.46889496e-01 5.12950793e-02
-1.09073013e-01 -1.29998818e-01 -1.79798377e+00 1.57122242e+00
9.40152705e-02 4.54500943e-01 -7.85835981e-02 -1.26725137e+00
6.73974156e-01 6.47017777e-01 8.87569845e-01 -9.10573378e-02
-1.71483889e-01 -1.08210899e-01 -7.46057630e-02 -7.71787524e-01
2.00896189e-01 -1.58263016e-02 6.87471181e-02 3.05391937e-01
1.01047911e-01 6.44040108e-01 4.99701381e-01 -6.10900261e-02
1.13209558e+00 2.97818810e-01 5.54384410e-01 2.36003935e-01
8.30112457e-01 -3.61078054e-01 1.21377301e+00 7.10903883e-01
-5.35692275e-01 6.84325457e-01 5.64434707e-01 -1.26729473e-01
-6.40902519e-01 -1.05132508e+00 3.61207336e-01 7.49195695e-01
3.80789250e-01 -2.42536917e-01 -6.55925512e-01 -7.42400110e-01
-4.86613542e-01 2.94847965e-01 -3.14885139e-01 1.08411238e-01
-6.47726417e-01 -5.50498843e-01 1.78119883e-01 5.89901626e-01
8.56943905e-01 -1.07371616e+00 -5.77549338e-01 3.82269055e-01
-7.42400408e-01 -1.42878258e+00 -7.46794045e-01 -2.17940420e-01
-9.18753862e-01 -1.17219186e+00 -9.17868614e-01 -8.94441903e-01
5.26912868e-01 5.88126540e-01 5.26432395e-01 -1.24083258e-01
2.03033105e-01 4.12485510e-01 -5.87080777e-01 2.65495926e-02
-5.75657338e-02 -2.73656875e-01 2.67492011e-02 8.79708707e-01
1.96897253e-01 -4.62770343e-01 -6.89859331e-01 7.22359061e-01
-1.11770201e+00 -6.62318692e-02 6.30438566e-01 5.83104670e-01
7.45184541e-01 5.18703163e-01 2.06358030e-01 -5.79260588e-01
1.19027672e-02 -8.09560895e-01 -3.16514254e-01 3.42057377e-01
5.60467653e-02 -1.99719548e-01 7.38365471e-01 -6.99731946e-01
-1.38638461e+00 3.68572712e-01 2.06223831e-01 -6.25529885e-01
-3.83292645e-01 5.36659598e-01 -4.77635890e-01 3.81252050e-01
1.03051469e-01 7.98929095e-01 -7.40431324e-02 -4.99367565e-01
-2.35575587e-01 5.67209542e-01 5.88255167e-01 -4.74215388e-01
4.93491411e-01 5.98819137e-01 -1.37845829e-01 -1.09551072e+00
-6.88169777e-01 -8.01119924e-01 -4.03643399e-01 -2.87347376e-01
8.56947243e-01 -1.29022634e+00 -3.89289260e-01 6.26700521e-01
-9.82526958e-01 -5.59844114e-02 -9.68332663e-02 1.06127131e+00
-6.00288808e-01 9.93308544e-01 -5.69550455e-01 -9.85327482e-01
2.79258966e-01 -1.23361862e+00 9.65609968e-01 3.49050403e-01
-6.58600777e-02 -8.66509020e-01 1.16975009e-01 5.08196771e-01
-3.30934614e-01 3.25679064e-01 3.96954358e-01 -7.13426292e-01
-7.04309881e-01 -1.59571633e-01 2.06407174e-01 4.39682305e-01
3.21386218e-01 -1.06660411e-01 -4.49723572e-01 -2.46742770e-01
3.07291180e-01 1.09776212e-02 4.77811307e-01 5.36272585e-01
1.11785090e+00 -4.16797727e-01 -5.16171336e-01 2.98608989e-01
1.05719900e+00 6.88352704e-01 7.65763760e-01 -2.74163987e-02
5.15118539e-01 5.37806094e-01 9.78725970e-01 8.10923338e-01
6.21279143e-02 7.17973232e-01 3.11112463e-01 3.82382363e-01
3.23917150e-01 -3.00957024e-01 7.32150674e-01 9.51592028e-01
-3.98395717e-01 -4.71253723e-01 -5.24149835e-01 6.53366566e-01
-2.15199065e+00 -1.41220105e+00 2.49984488e-02 2.30971766e+00
3.09393585e-01 8.93563777e-02 2.28420541e-01 3.86176020e-01
9.82918382e-01 5.38695931e-01 -5.28570712e-01 3.03737104e-01
-7.42015196e-03 -2.22328246e-01 9.37453285e-02 1.27175674e-01
-1.13362396e+00 3.21623951e-01 5.67100048e+00 9.23658431e-01
-6.89061582e-01 8.04339722e-02 3.60762864e-01 -2.45284975e-01
3.47787917e-01 1.84061050e-01 -6.88376009e-01 9.73475695e-01
5.33728421e-01 9.35483947e-02 3.63594919e-01 7.89178312e-01
6.15942597e-01 -4.56274539e-01 -9.73754883e-01 1.03935564e+00
2.76102364e-01 -7.93645918e-01 3.14535871e-02 6.40898049e-02
9.07494962e-01 -4.08414900e-01 -2.92906582e-01 2.43487537e-01
-2.13519067e-01 -4.73317355e-01 6.47989154e-01 6.54371798e-01
2.68753856e-01 -7.01153338e-01 7.20990837e-01 8.81054401e-01
-1.59343994e+00 -3.02742809e-01 -7.12016970e-02 -1.66089833e-01
4.93397653e-01 4.17047054e-01 -4.18134391e-01 5.98227739e-01
6.66932940e-01 1.17723632e+00 -2.00493231e-01 1.25334430e+00
-2.34106287e-01 8.98364842e-01 -2.54529417e-01 1.84983522e-01
2.06852973e-01 -4.33629692e-01 6.31286979e-01 9.50289190e-01
6.80664659e-01 3.39386672e-01 5.67202687e-01 3.32968533e-01
1.91261113e-01 -1.28401918e-02 -4.33376789e-01 2.00918652e-02
4.18175012e-01 7.33763814e-01 -6.99392021e-01 -6.31907940e-01
-9.27851617e-01 1.19744706e+00 -6.49724156e-02 5.84516346e-01
-1.00778615e+00 -1.50347337e-01 3.87981117e-01 1.94344357e-01
7.12750435e-01 -2.20046580e-01 5.43387890e-01 -1.48100543e+00
3.93729180e-01 -8.64651144e-01 6.15924478e-01 -8.10299516e-01
-8.88103962e-01 2.17280596e-01 2.35821262e-01 -1.76131785e+00
-4.24227923e-01 1.84593555e-02 -6.00124240e-01 4.01092857e-01
-1.17417538e+00 -7.97559559e-01 -4.84657526e-01 9.00044501e-01
1.01813352e+00 -4.17716652e-02 2.22759500e-01 3.21824342e-01
-6.22917414e-01 6.63617700e-02 3.39719266e-01 2.42862299e-01
4.53172624e-01 -7.16415048e-01 -2.04825491e-01 1.14412606e+00
3.09459865e-01 3.14909190e-01 5.54541051e-01 -7.81411409e-01
-9.91702616e-01 -1.01365733e+00 7.57527649e-01 1.78059012e-01
4.15060759e-01 2.62182727e-02 -1.00243843e+00 9.61795986e-01
3.23208682e-02 1.66127190e-01 8.33387256e-01 -3.40092212e-01
2.01473489e-01 7.01706624e-03 -8.88116658e-01 2.92462170e-01
1.10478795e+00 -3.18277538e-01 -8.27015579e-01 4.79659624e-02
3.18062842e-01 -1.91029206e-01 -7.02384472e-01 5.12208641e-01
6.92940235e-01 -1.00939107e+00 6.65519774e-01 -2.07584083e-01
2.87173569e-01 -7.81819880e-01 -1.31705627e-01 -1.03030765e+00
-2.63012111e-01 -5.61403394e-01 -6.59025609e-01 1.02242649e+00
-1.26161918e-01 -2.50185102e-01 9.61601377e-01 2.15593383e-01
-2.95443162e-02 -6.45218909e-01 -9.83189344e-01 -9.86500144e-01
-7.80666888e-01 -4.36323851e-01 3.34256202e-01 7.30717599e-01
1.83927849e-01 -7.71138594e-02 -8.89856696e-01 3.91559154e-01
6.13553345e-01 8.46315026e-02 6.80272341e-01 -9.76143301e-01
-7.33420193e-01 1.02437943e-01 -5.39712846e-01 -1.77941179e+00
9.24029201e-03 -1.80803552e-01 3.18485945e-01 -1.29412436e+00
5.13873994e-01 2.47419626e-01 -3.47805262e-01 3.27904187e-02
-2.73127854e-01 -9.36583430e-02 1.65090322e-01 5.96001744e-01
-9.64288175e-01 7.04990029e-01 1.10412550e+00 -1.40362203e-01
-2.34618723e-01 2.83936292e-01 -1.92830339e-01 8.76739919e-01
4.37293530e-01 -4.68270063e-01 -6.19181573e-01 -1.14985511e-01
-4.68947828e-01 6.72148585e-01 2.77209014e-01 -1.53370953e+00
2.10014030e-01 -4.20904577e-01 3.05579901e-01 -7.65381634e-01
4.51764762e-01 -9.98545825e-01 5.13970077e-01 4.89203125e-01
-1.16300218e-01 -2.98135906e-01 -2.82791555e-01 1.20839024e+00
-5.83492637e-01 -2.85944223e-01 5.54318845e-01 -8.05837363e-02
-7.74742246e-01 4.67737019e-01 -6.91851735e-01 -5.27952984e-02
1.35521317e+00 -5.01789153e-01 8.03373531e-02 -5.10313869e-01
-1.10854685e+00 3.27661335e-01 3.26360315e-01 3.52708817e-01
6.22121453e-01 -1.38049817e+00 -4.47907120e-01 1.46544755e-01
1.52792722e-01 -1.56457592e-02 6.35841012e-01 1.07294238e+00
-3.32570046e-01 1.50280491e-01 -3.08353603e-02 -8.03208947e-01
-1.35205603e+00 6.90508842e-01 -5.79547882e-02 -3.89793783e-01
-3.57711434e-01 3.19201887e-01 4.37411726e-01 4.60168958e-01
3.05500567e-01 -1.66167215e-01 -5.17583489e-01 -1.91965066e-02
6.31859601e-01 6.05475724e-01 -4.89258975e-01 -9.52899635e-01
-2.31694147e-01 6.05408728e-01 1.53746858e-01 -7.69246593e-02
1.03748047e+00 -2.56155103e-01 2.73058683e-01 6.98428273e-01
1.18441272e+00 -4.19692956e-02 -1.79835880e+00 -4.56674069e-01
-1.26564324e-01 -6.91664934e-01 -4.24195081e-01 -1.07330136e-01
-1.11429703e+00 6.31915390e-01 4.50070381e-01 6.26709834e-02
1.21261823e+00 -4.25242856e-02 9.74613607e-01 1.38480276e-01
5.07429361e-01 -1.00494790e+00 2.92800784e-01 2.33928114e-01
4.57176179e-01 -9.91999567e-01 7.90580437e-02 -4.70974416e-01
-9.61383939e-01 8.85368347e-01 4.23493683e-01 -6.46799579e-02
6.40448749e-01 -1.56444162e-01 -4.49962467e-01 1.87999263e-01
-5.98671556e-01 -1.26061738e-01 1.92481309e-01 4.76705492e-01
9.60660577e-02 -9.61102098e-02 -3.38039637e-01 6.12654030e-01
5.66701055e-01 3.98323923e-01 5.58918417e-01 1.31876981e+00
-5.43483436e-01 -7.48313069e-01 -6.85781300e-01 3.94445866e-01
-5.79155803e-01 3.49080473e-01 5.80753684e-02 4.55404431e-01
4.41741407e-01 1.09639573e+00 6.81183636e-02 -1.83378696e-01
2.81945646e-01 1.74110755e-01 3.49249929e-01 -5.58023155e-01
1.41575605e-01 5.53033233e-01 -3.77702191e-02 -5.78424692e-01
-8.89064193e-01 -1.21286166e+00 -1.04273081e+00 6.22776970e-02
-2.33605430e-01 3.76476675e-01 -5.72540648e-02 1.12862742e+00
2.07422256e-01 3.02093714e-01 6.51361644e-01 -8.54577601e-01
-2.76392162e-01 -9.82152522e-01 -9.90035653e-01 6.87302172e-01
2.91765541e-01 -8.01895738e-01 -5.01818359e-01 6.55257642e-01] | [8.48975944519043, 0.5577136278152466] |
eceb7b48-b1fe-45cb-ade7-2f6ffd49e083 | detecting-robotic-affordances-on-novel | 1909.05770 | null | https://arxiv.org/abs/1909.05770v2 | https://arxiv.org/pdf/1909.05770v2.pdf | Recognizing Object Affordances to Support Scene Reasoning for Manipulation Tasks | Affordance information about a scene provides important clues as to what actions may be executed in pursuit of meeting a specified goal state. Thus, integrating affordance-based reasoning into symbolic action plannning pipelines would enhance the flexibility of robot manipulation. Unfortunately, the top performing affordance recognition methods use object category priors to boost the accuracy of affordance detection and segmentation. Object priors limit generalization to unknown object categories. This paper describes an affordance recognition pipeline based on a category-agnostic region proposal network for proposing instance regions of an image across categories. To guide affordance learning in the absence of category priors, the training process includes the auxiliary task of explicitly inferencing existing affordances within a proposal. Secondly, a self-attention mechanism trained to interpret each proposal learns to capture rich contextual dependencies through the region. Visual benchmarking shows that the trained network, called AffContext, reduces the performance gap between object-agnostic and object-informed affordance recognition. AffContext is linked to the Planning Domain Definition Language (PDDL) with an augmented state keeper for action planning across temporally spaced goal-oriented tasks. Manipulation experiments show that AffContext can successfully parse scene content to seed a symbolic planner problem specification, whose execution completes the target task. Additionally, task-oriented grasping for cutting and pounding actions demonstrate the exploitation of multiple affordances for a given object to complete specified tasks. | ['Fu-Jen Chu', 'Ruinian Xu', 'Patricio A. Vela', 'Chao Tang'] | 2019-09-12 | null | null | null | null | ['affordance-recognition', 'affordance-detection'] | ['computer-vision', 'computer-vision'] | [ 4.49204385e-01 4.70089316e-01 -3.86643946e-01 -4.26508158e-01
-2.85862744e-01 -7.34108567e-01 8.05625677e-01 1.57513887e-01
-2.53394127e-01 2.85259068e-01 6.02791071e-01 -3.69457990e-01
-2.42039993e-01 -5.36820292e-01 -7.18508363e-01 -2.78943360e-01
-2.13261560e-01 5.61386287e-01 4.07332659e-01 -3.41900200e-01
6.22145653e-01 9.37555552e-01 -1.64463568e+00 5.74665010e-01
4.50019151e-01 8.24839056e-01 7.51403153e-01 6.17335677e-01
-4.08980474e-02 7.71154046e-01 -3.19571018e-01 4.84676033e-01
4.42671359e-01 -1.90931454e-01 -1.19787097e+00 3.30195993e-01
3.71589750e-01 -7.72373438e-01 -2.63775110e-01 7.59134471e-01
-2.09384441e-01 5.56134582e-01 7.40240276e-01 -1.33176851e+00
-5.07956386e-01 5.81567347e-01 -1.74067430e-02 2.82569379e-01
5.46298683e-01 9.12097692e-01 1.25620782e+00 -5.27833998e-01
9.49746192e-01 1.59366167e+00 -1.12416334e-01 5.84041476e-01
-1.35465384e+00 -6.97803423e-02 5.67697287e-01 4.78767455e-02
-8.77482533e-01 -4.30840760e-01 6.31948829e-01 -4.93071198e-01
1.71363568e+00 1.25888541e-01 9.38473940e-01 8.49764466e-01
1.51356548e-01 1.05025816e+00 8.87369037e-01 -3.17660600e-01
2.49232844e-01 -7.44038999e-01 1.38373047e-01 7.89575279e-01
-3.03355865e-02 4.41395521e-01 -4.46023256e-01 -4.35228869e-02
1.40945363e+00 3.47058326e-01 -2.88444534e-02 -6.93163991e-01
-1.71760976e+00 4.53596234e-01 1.00523579e+00 3.81763011e-01
-4.41523284e-01 5.36304295e-01 2.70252496e-01 1.49764791e-01
-1.77307487e-01 1.11186194e+00 -6.42819941e-01 3.02365925e-02
-3.07165593e-01 6.11031413e-01 5.73235750e-01 1.44972038e+00
8.09215784e-01 -1.23317361e-01 -4.04471546e-01 2.49319911e-01
4.90601242e-01 4.81247231e-02 1.90802947e-01 -1.44503629e+00
4.26315010e-01 9.65401173e-01 2.68277496e-01 -4.39521968e-01
-7.10803807e-01 2.19028771e-01 2.01859534e-01 5.49445391e-01
5.97160995e-01 4.75826591e-01 -1.27653146e+00 1.54848099e+00
4.35257018e-01 -2.65153140e-01 -1.03822285e-02 1.33314908e+00
4.75234807e-01 3.25992584e-01 6.03904486e-01 3.41803282e-01
1.59044445e+00 -9.99539912e-01 -2.85834849e-01 -7.24717975e-01
7.22420990e-01 -3.43881458e-01 1.42403650e+00 1.98326752e-01
-7.74790227e-01 -4.49786633e-01 -9.99118686e-01 -5.97653806e-01
-1.69757426e-01 -8.00919309e-02 1.32575059e+00 -2.74166971e-01
-6.57931328e-01 7.61744261e-01 -1.22015107e+00 -6.19825244e-01
8.19345057e-01 3.41245651e-01 -6.76032722e-01 -1.80899099e-01
-3.87294918e-01 1.14143109e+00 7.74764001e-01 -2.65661869e-02
-1.48242915e+00 -2.46943980e-01 -1.25545955e+00 4.88103144e-02
7.23754168e-01 -7.29537070e-01 1.47858000e+00 -8.23342621e-01
-1.34393907e+00 6.78601801e-01 -7.95139521e-02 -3.79680574e-01
-6.98552728e-02 -4.12921369e-01 2.10327297e-01 2.97199309e-01
4.80438352e-01 1.20473492e+00 9.49590981e-01 -1.00561440e+00
-5.88894963e-01 -4.59349096e-01 6.59123659e-01 4.65446979e-01
7.08315313e-01 -1.09536657e-02 -1.77277416e-01 -4.58139926e-01
8.75752211e-01 -7.77725399e-01 -4.02302861e-01 1.83791533e-01
-4.95046407e-01 -6.55203640e-01 6.92203522e-01 -2.99639404e-01
5.16202748e-01 -2.17442966e+00 4.07935947e-01 -1.46959201e-01
6.38766633e-03 -5.99462688e-01 -3.94185334e-01 4.13263798e-01
-8.27671215e-02 -3.22623663e-02 -3.23191136e-01 1.23294570e-01
2.47186095e-01 7.49762833e-01 -6.06500506e-01 4.75475132e-01
6.39537096e-01 1.08804941e+00 -1.12986314e+00 -2.15881720e-01
3.49512339e-01 -1.32503748e-01 -9.69467521e-01 5.36813915e-01
-1.03487360e+00 7.84969747e-01 -7.25184858e-01 8.99824679e-01
-4.34322469e-02 3.84279974e-02 2.33995825e-01 -6.64620996e-02
-1.80325136e-01 9.85978127e-01 -1.01645148e+00 2.41715193e+00
-2.39731491e-01 3.12899739e-01 3.22558172e-02 -7.84607947e-01
5.78248441e-01 1.36976570e-01 2.17416093e-01 -2.15455443e-01
9.21761841e-02 2.73230404e-01 5.00384271e-01 -7.41452217e-01
4.01055217e-01 -1.08138107e-01 -3.50594521e-01 6.79344177e-01
4.34556156e-01 -8.38068902e-01 2.11563930e-01 2.20455751e-01
1.19026005e+00 1.29533529e+00 6.22121572e-01 -2.60932833e-01
-1.26488253e-01 6.40703142e-01 3.12826186e-01 7.50358224e-01
-3.47337246e-01 3.89741480e-01 6.13107860e-01 -7.70650089e-01
-1.16686797e+00 -1.11354184e+00 1.37881814e-02 1.56195080e+00
2.97971725e-01 -7.82774389e-02 -4.31243986e-01 -6.25752568e-01
2.06173658e-02 1.04674137e+00 -6.01675153e-01 -2.13190634e-02
-8.56752574e-01 -9.42230374e-02 9.78281200e-02 7.20777154e-01
1.26877651e-01 -1.79008222e+00 -1.62181151e+00 1.62126198e-01
9.73875821e-02 -9.64385867e-01 -2.92341292e-01 5.44806778e-01
-1.00601590e+00 -1.41494751e+00 -1.83679044e-01 -6.30275130e-01
9.87004995e-01 7.61218220e-02 8.59725595e-01 2.04676017e-01
-4.42787737e-01 6.69032037e-01 -2.92645693e-01 -1.97096288e-01
-3.76238525e-01 -1.43212192e-02 -3.15069705e-02 -6.41901493e-01
4.41096872e-01 -4.85149741e-01 -5.37017345e-01 1.34385943e-01
-6.82762086e-01 3.57315302e-01 6.35957360e-01 6.87418997e-01
5.54643273e-01 -5.15917063e-01 8.79101455e-03 -2.09499717e-01
4.06561732e-01 -3.11523259e-01 -5.49588084e-01 -5.99040985e-02
-1.06763408e-01 5.07297337e-01 2.57867835e-02 -5.34053087e-01
-1.07813215e+00 3.76026899e-01 2.66182095e-01 -1.52555719e-01
-7.89365351e-01 2.88535118e-01 -2.62071729e-01 4.83147949e-01
7.93876946e-01 -1.37475476e-01 1.48647008e-02 -4.44464803e-01
7.52509773e-01 2.69602016e-02 6.33545280e-01 -1.09644544e+00
4.31378812e-01 5.23718953e-01 3.16708945e-02 -2.93907762e-01
-9.38342035e-01 -4.77411240e-01 -1.21845400e+00 1.36971608e-01
1.14417934e+00 -6.46954358e-01 -9.90795314e-01 -1.66392609e-01
-1.41703677e+00 -8.02332520e-01 -7.21585155e-01 4.87088442e-01
-1.28266907e+00 -2.19873235e-01 -4.36854541e-01 -6.24340832e-01
3.22426021e-01 -1.51844382e+00 1.46221483e+00 -4.72226366e-02
-6.97096825e-01 -2.79275239e-01 -5.04362881e-01 1.40364412e-02
1.63202896e-03 2.62052923e-01 1.06317055e+00 -7.64337420e-01
-1.11760950e+00 1.81007180e-02 -3.53084832e-01 -3.07737887e-01
1.61873117e-01 -4.94939059e-01 -6.66645050e-01 9.05747637e-02
-9.36780050e-02 -4.40429717e-01 7.01163530e-01 1.70754388e-01
9.81778920e-01 -4.30313945e-01 -5.35984278e-01 3.76834989e-01
1.10108125e+00 1.25619680e-01 4.23053741e-01 5.23199260e-01
7.32637286e-01 9.27413642e-01 9.36062515e-01 1.35377511e-01
3.32286835e-01 5.70566952e-01 9.24925685e-01 4.20211911e-01
-1.00496724e-01 -3.78263086e-01 2.60149568e-01 -3.67674291e-01
-2.55649477e-01 3.25923830e-01 -1.02184629e+00 4.88465667e-01
-1.88500166e+00 -9.49440360e-01 1.12667140e-02 1.71802962e+00
9.30903077e-01 1.45225212e-01 2.15124805e-02 -1.59127221e-01
5.77008799e-02 1.48147941e-01 -1.00541294e+00 -3.79802048e-01
3.69915992e-01 -1.42264534e-02 3.92070651e-01 5.24269760e-01
-1.13562953e+00 1.56307948e+00 5.92652750e+00 1.74748257e-01
-6.52260184e-01 1.19625203e-01 -1.30382985e-01 7.42049068e-02
-1.75511822e-01 5.46428859e-01 -6.42104387e-01 -4.82127577e-01
3.53203326e-01 3.32539052e-01 7.88503587e-01 1.14927435e+00
1.26149967e-01 -6.16697967e-01 -1.76995230e+00 3.42842251e-01
-1.02744028e-01 -1.16841292e+00 -4.51804325e-02 6.00790679e-02
2.31268495e-01 1.92751288e-01 -3.18475395e-01 3.62986684e-01
6.30035400e-01 -1.08556521e+00 1.12381256e+00 6.17579877e-01
5.48784673e-01 -2.14344606e-01 -3.87427509e-02 5.62683225e-01
-1.02450633e+00 -5.28900206e-01 -2.91276574e-01 -5.35565913e-01
2.75859565e-01 -3.37337822e-01 -1.18747544e+00 -6.53476082e-03
4.77807045e-01 6.35155261e-01 -1.79453447e-01 6.36448205e-01
-8.68684411e-01 -3.13731693e-02 -4.46513832e-01 2.87482347e-02
5.25413036e-01 7.70120397e-02 8.72742295e-01 7.07577050e-01
-8.01459029e-02 4.38346624e-01 6.15667522e-01 1.38708377e+00
4.41467285e-01 -3.29443485e-01 -7.60422707e-01 -2.14900568e-01
2.55476773e-01 8.58817816e-01 -8.87702763e-01 -3.89030784e-01
-4.13066521e-02 9.91172791e-01 4.79276806e-01 5.25941789e-01
-3.34011614e-01 -9.66440421e-04 7.80710280e-01 1.50311232e-01
2.38261998e-01 -8.66075873e-01 -5.34268498e-01 -8.79648626e-01
-1.30889013e-01 -7.41463363e-01 2.30388820e-01 -1.23169518e+00
-6.86853766e-01 1.44018009e-01 4.69088107e-01 -1.01578844e+00
-4.11330163e-01 -1.00818932e+00 -3.38619649e-01 9.64197755e-01
-1.31670916e+00 -1.49533355e+00 -1.91237479e-01 3.97922903e-01
1.03566110e+00 1.12183662e-02 1.04997623e+00 -6.56082451e-01
1.04694016e-01 -4.49631125e-01 -1.02139485e+00 1.30043060e-01
2.77133077e-01 -1.29730952e+00 4.08232778e-01 7.82599747e-01
1.66771099e-01 1.13442767e+00 6.87750578e-01 -9.66977656e-01
-1.51072466e+00 -6.45604610e-01 4.00979221e-01 -1.05263674e+00
6.58417583e-01 -3.66426468e-01 -7.92042255e-01 1.15997970e+00
-9.99097526e-02 2.66194165e-01 2.61430889e-01 1.31226271e-01
-6.17696881e-01 5.90783536e-01 -1.08875561e+00 9.10471141e-01
1.62797081e+00 -6.48129106e-01 -1.42849398e+00 4.92107481e-01
1.01795197e+00 -7.30730951e-01 -7.33791232e-01 2.83244371e-01
5.61169028e-01 -6.98097169e-01 8.88979137e-01 -1.09454811e+00
4.97299373e-01 -6.71947658e-01 -3.84154379e-01 -9.23279703e-01
-3.37290525e-01 -3.89181495e-01 -2.34607026e-01 4.62915242e-01
1.10684633e-01 -2.34539941e-01 3.79079521e-01 8.98970723e-01
-5.90573967e-01 -3.38509560e-01 -8.03408086e-01 -5.68196356e-01
-3.43017846e-01 -8.12705219e-01 9.35375452e-01 5.87151647e-01
4.39022154e-01 3.30140650e-01 4.41876650e-01 2.02941105e-01
3.84137005e-01 5.20405531e-01 9.04041708e-01 -1.04107869e+00
-2.64242202e-01 -5.71898878e-01 -2.67585039e-01 -1.45129871e+00
4.42730606e-01 -1.12862694e+00 5.57901978e-01 -1.81831539e+00
-1.35722548e-01 -7.00688064e-01 8.34549218e-02 1.27582228e+00
1.70991179e-02 -2.52276003e-01 5.11413813e-01 5.55148065e-01
-4.11574751e-01 3.05970192e-01 1.74602604e+00 -2.17195809e-01
-5.15260756e-01 -2.27529302e-01 -5.77709377e-01 9.38917816e-01
7.31014132e-01 -3.79161298e-01 -3.35974008e-01 -6.78823173e-01
-4.13315520e-02 -5.70430197e-02 8.10488999e-01 -6.32315576e-01
-2.58435272e-02 -6.75892770e-01 2.59792358e-01 -3.94242734e-01
3.79396170e-01 -1.00307381e+00 -3.83670479e-01 6.48427367e-01
-5.43569922e-01 -6.26929328e-02 1.96963862e-01 5.80597997e-01
2.61069685e-01 -2.30915472e-01 2.83525556e-01 -7.03891993e-01
-1.44289303e+00 2.46363759e-01 -4.15576786e-01 -2.79353112e-01
9.31429625e-01 -3.65378201e-01 -2.10399017e-01 2.27228388e-01
-1.16858041e+00 1.64261535e-01 5.68028152e-01 6.94451094e-01
6.52151346e-01 -9.35160875e-01 -1.52953669e-01 3.13097507e-01
4.51905936e-01 4.86199826e-01 -2.45162740e-01 6.66908920e-01
-3.58156055e-01 3.51752073e-01 -5.22794306e-01 -7.93735981e-01
-5.55281460e-01 7.46626496e-01 3.44158053e-01 1.19805232e-01
-1.02132928e+00 9.37981963e-01 3.66338581e-01 -8.56924653e-01
1.34112284e-01 -1.01366186e+00 -2.46371970e-01 -4.67845291e-01
2.86756247e-01 -4.61588539e-02 -4.98307198e-01 -6.55855656e-01
-2.86423713e-01 7.89548680e-02 3.32075000e-01 -3.29236180e-01
1.35001016e+00 2.71539539e-02 -4.27876472e-01 6.00537360e-01
6.11887038e-01 -6.45494461e-01 -2.01934528e+00 7.78417438e-02
4.11091894e-01 -3.80875021e-01 -2.89851159e-01 -9.33091223e-01
-1.48709625e-01 9.16074932e-01 3.12113911e-02 -1.02462359e-01
6.16456866e-01 5.37339926e-01 1.27754316e-01 8.02830875e-01
8.21927965e-01 -9.23646748e-01 3.13187838e-01 8.71106446e-01
1.53205872e+00 -1.33310795e+00 1.36302605e-01 -3.36341918e-01
-5.76508105e-01 1.24361694e+00 9.69212890e-01 -3.27371299e-01
2.69911230e-01 3.41657852e-03 -3.67834002e-01 -5.63947916e-01
-6.68047726e-01 -4.31499809e-01 4.87816989e-01 7.92823076e-01
1.92516968e-01 2.28756189e-01 3.03553138e-02 3.36901546e-01
-2.21651852e-01 -2.95987487e-01 2.19759271e-01 1.24675131e+00
-8.51467550e-01 -7.75676966e-01 -4.12116498e-01 5.18648624e-01
4.24543619e-02 -1.17429547e-01 -1.21123627e-01 8.53058815e-01
1.72432467e-01 6.19878531e-01 4.26062912e-01 2.30746165e-01
4.39727515e-01 3.06049675e-01 9.45172191e-01 -1.44729674e+00
-5.12651145e-01 2.45421484e-01 7.76343271e-02 -1.22605765e+00
-6.03442788e-01 -9.87637639e-01 -2.06012511e+00 5.68968713e-01
-7.58775845e-02 -3.61426950e-01 6.47054613e-01 1.40671420e+00
1.82705507e-01 6.74087167e-01 -2.06028238e-01 -1.43865085e+00
-5.05921543e-01 -1.02416527e+00 -2.34250247e-01 3.59748900e-01
4.07241881e-01 -8.76443923e-01 -6.27138987e-02 1.22182019e-01] | [4.826569080352783, 0.3650049567222595] |
16187e73-30ea-4c23-8974-1a911417b28d | conflict-based-cross-view-consistency-for | 2303.01276 | null | https://arxiv.org/abs/2303.01276v3 | https://arxiv.org/pdf/2303.01276v3.pdf | Conflict-Based Cross-View Consistency for Semi-Supervised Semantic Segmentation | Semi-supervised semantic segmentation (SSS) has recently gained increasing research interest as it can reduce the requirement for large-scale fully-annotated training data. The current methods often suffer from the confirmation bias from the pseudo-labelling process, which can be alleviated by the co-training framework. The current co-training-based SSS methods rely on hand-crafted perturbations to prevent the different sub-nets from collapsing into each other, but these artificial perturbations cannot lead to the optimal solution. In this work, we propose a new conflict-based cross-view consistency (CCVC) method based on a two-branch co-training framework which aims at enforcing the two sub-nets to learn informative features from irrelevant views. In particular, we first propose a new cross-view consistency (CVC) strategy that encourages the two sub-nets to learn distinct features from the same input by introducing a feature discrepancy loss, while these distinct features are expected to generate consistent prediction scores of the input. The CVC strategy helps to prevent the two sub-nets from stepping into the collapse. In addition, we further propose a conflict-based pseudo-labelling (CPL) method to guarantee the model will learn more useful information from conflicting predictions, which will lead to a stable training process. We validate our new CCVC approach on the SSS benchmark datasets where our method achieves new state-of-the-art performance. Our code is available at https://github.com/xiaoyao3302/CCVC. | ['Xiangyu Kong', 'Xiaoxia Xing', 'Dong Xu', 'Luping Zhou', 'Zhen Zhao', 'Zicheng Wang'] | 2023-03-02 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_Conflict-Based_Cross-View_Consistency_for_Semi-Supervised_Semantic_Segmentation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_Conflict-Based_Cross-View_Consistency_for_Semi-Supervised_Semantic_Segmentation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['semi-supervised-semantic-segmentation'] | ['computer-vision'] | [ 3.57187688e-01 4.05702233e-01 -2.53232062e-01 -6.37325466e-01
-8.35641623e-01 -4.27731186e-01 4.07249987e-01 -1.11462228e-01
-1.96208939e-01 6.30641282e-01 6.12738170e-03 -1.20226435e-01
-8.58555064e-02 -5.39817214e-01 -8.95030320e-01 -8.37603033e-01
3.91885072e-01 4.08360541e-01 6.22647822e-01 2.27377806e-02
1.02170162e-01 2.83102356e-02 -1.58023298e+00 4.55906242e-01
1.21819186e+00 1.00588095e+00 3.70744675e-01 1.24023765e-01
-1.43593147e-01 5.25997102e-01 -3.43171388e-01 -4.16440398e-01
3.85149151e-01 -5.11041164e-01 -8.81350338e-01 1.24219023e-01
1.88098431e-01 -1.26549751e-01 3.77172023e-01 1.20204508e+00
4.12502736e-01 5.72554301e-03 3.01087260e-01 -1.44599760e+00
-2.44886816e-01 5.44171035e-01 -7.62930751e-01 -2.45732799e-01
-5.10541089e-02 1.17769919e-01 1.18152654e+00 -7.22528696e-01
7.26306677e-01 1.12038624e+00 5.59571564e-01 7.93884873e-01
-1.23890090e+00 -7.07674265e-01 4.71811324e-01 1.89494342e-01
-1.12804329e+00 -2.09981829e-01 9.81255829e-01 -1.54706165e-01
4.91681069e-01 3.77463043e-01 6.18116856e-01 1.08083284e+00
-1.15104057e-01 1.21059036e+00 1.24977160e+00 -3.57359767e-01
2.85999507e-01 2.26471812e-01 2.69568171e-02 7.34320998e-01
2.98042208e-01 -5.29161207e-02 -5.43698609e-01 2.56769601e-02
5.18755257e-01 -5.20661995e-02 -4.08487976e-01 -8.11119556e-01
-8.56477499e-01 7.56407320e-01 4.80463237e-01 3.36646020e-01
-1.53606996e-01 -1.85692489e-01 4.81096685e-01 -1.25136226e-02
4.57586259e-01 3.55949312e-01 -8.21450055e-01 1.21819220e-01
-9.37572241e-01 1.84168473e-01 4.14102376e-01 8.05569828e-01
8.03015590e-01 -3.17584157e-01 -3.11122388e-01 9.42389369e-01
4.02331769e-01 4.44697030e-02 3.66056532e-01 -1.01464081e+00
6.52339578e-01 1.00701058e+00 -5.47601767e-02 -7.73317754e-01
-3.00870717e-01 -6.53714418e-01 -7.67718613e-01 1.80769607e-01
3.97274792e-01 1.39505416e-01 -8.72781217e-01 1.96750224e+00
6.56353414e-01 2.63388962e-01 1.24520652e-01 9.80827153e-01
5.69095433e-01 4.48665619e-01 -1.51992604e-01 -1.61187232e-01
1.04334152e+00 -1.24406302e+00 -4.19082403e-01 -2.82114476e-01
7.69594133e-01 -5.39192975e-01 1.14152336e+00 3.82368356e-01
-9.76261973e-01 -5.56554854e-01 -1.13848484e+00 1.28072619e-01
-1.14865653e-01 1.24164149e-01 2.97135592e-01 3.23498636e-01
-7.00277209e-01 6.93036616e-01 -9.82687056e-01 2.12014709e-02
7.09501445e-01 3.22971404e-01 -4.13731784e-01 -2.20663965e-01
-9.73612607e-01 5.88986933e-01 5.77062547e-01 3.34109545e-01
-5.80048978e-01 -5.95394373e-01 -7.92026877e-01 8.61038174e-03
8.58759046e-01 -4.82257217e-01 1.09425163e+00 -1.45825660e+00
-1.28648555e+00 8.03473115e-01 -2.13955626e-01 -1.45887598e-01
7.71946132e-01 -2.19983712e-01 -5.89356013e-02 8.33039079e-03
4.32960510e-01 9.11864281e-01 6.94364429e-01 -1.77851987e+00
-6.95782125e-01 -3.69040221e-01 -1.14828125e-01 2.61644185e-01
-2.41062194e-01 -2.01166689e-01 -8.39343131e-01 -5.82409203e-01
4.54667777e-01 -1.01866066e+00 -1.86507255e-01 3.23942950e-04
-7.64819384e-01 -2.64069974e-01 7.60358512e-01 -3.45950991e-01
1.09843099e+00 -2.06276131e+00 3.15516025e-01 1.69403806e-01
3.27178501e-02 5.63645840e-01 -1.93306878e-01 -3.03220600e-02
-4.11110520e-02 1.44868478e-01 -5.95853329e-01 -6.72102690e-01
-2.27322206e-01 5.48237383e-01 -1.40992433e-01 1.29327238e-01
4.45027024e-01 7.36235440e-01 -9.65938985e-01 -6.25532925e-01
1.27925072e-02 1.57354563e-01 -6.48587584e-01 2.79670715e-01
-3.83961827e-01 6.26131594e-01 -3.68001968e-01 4.76663798e-01
9.57454324e-01 -3.60678434e-01 3.99538219e-01 -1.32137060e-01
2.77738683e-02 1.73382491e-01 -1.19232321e+00 1.69359767e+00
-2.48995408e-01 -6.52950555e-02 -5.41718751e-02 -1.15198565e+00
8.61062169e-01 8.75742733e-02 3.54363889e-01 -7.12249577e-01
7.17490166e-02 3.36535215e-01 -1.57349501e-02 -2.30078250e-01
6.56490326e-02 -9.24266055e-02 4.54727970e-02 2.93257654e-01
-3.38479020e-02 1.82859644e-01 1.78343579e-01 1.82603851e-01
7.21258104e-01 5.51555872e-01 7.77237415e-02 -2.04737961e-01
7.84958959e-01 -9.25209094e-03 1.29767001e+00 4.71132487e-01
-2.16020867e-01 7.76459515e-01 7.35655189e-01 -3.35599095e-01
-7.63796329e-01 -7.98592567e-01 3.74934264e-02 7.21985757e-01
3.64948303e-01 -3.29943269e-01 -8.25117946e-01 -1.46043015e+00
-1.76095024e-01 7.95077145e-01 -5.53456426e-01 -2.97960609e-01
-4.24757570e-01 -5.45360386e-01 2.49951139e-01 6.70459688e-01
6.07212663e-01 -1.00804210e+00 -6.37800872e-01 9.65522900e-02
-4.06662911e-01 -1.02823985e+00 -4.29499030e-01 4.37566966e-01
-9.05378520e-01 -1.20190048e+00 -6.25516891e-01 -6.70995712e-01
1.10837054e+00 2.30578139e-01 8.58414650e-01 2.76702493e-01
6.44358844e-02 -3.06465387e-01 -4.95769173e-01 -1.32482097e-01
-4.65747684e-01 1.96934730e-01 -1.36651844e-01 4.76044528e-02
1.87783211e-01 -4.66150790e-01 -4.97834116e-01 6.56287551e-01
-8.82408798e-01 6.34212613e-01 4.57362205e-01 1.10778463e+00
9.63429332e-01 -1.00255094e-01 5.97388446e-01 -1.08442903e+00
3.36916931e-02 -3.10591578e-01 -7.16080129e-01 4.36514467e-01
-9.29781199e-01 3.41931164e-01 8.94315124e-01 -2.52125114e-01
-1.19263566e+00 3.17436159e-01 -1.27454221e-01 -6.86831772e-01
-1.91084325e-01 2.83803612e-01 -5.35101295e-01 2.72284001e-01
2.32647017e-01 2.65492976e-01 -1.08732611e-01 -4.64240313e-01
2.01680869e-01 4.20801580e-01 3.97594005e-01 -5.73999763e-01
6.27448678e-01 4.06916648e-01 -6.83969632e-02 -3.02807465e-02
-1.23403990e+00 -4.50287640e-01 -7.26822674e-01 -2.17481300e-01
7.51960635e-01 -7.81320870e-01 -3.63260508e-01 5.18218875e-01
-1.08923268e+00 -2.60367244e-01 -1.34579480e-01 1.97898880e-01
-4.49728459e-01 4.43727136e-01 -2.73067832e-01 -7.79074788e-01
-2.00281963e-01 -1.28826022e+00 1.09235251e+00 3.73625249e-01
4.18396182e-02 -6.19522929e-01 5.25349714e-02 6.61331952e-01
-4.49623950e-02 1.92248881e-01 7.79220700e-01 -8.85577023e-01
-5.21914065e-01 1.76534995e-01 -2.73144156e-01 6.32197082e-01
1.45232096e-01 -7.24870199e-03 -9.69599664e-01 -2.64202982e-01
-7.96225369e-02 -4.84248936e-01 9.93591666e-01 2.77631611e-01
1.21797287e+00 -2.23006919e-01 -3.31093997e-01 6.03410363e-01
1.35052705e+00 1.20800018e-01 4.63531554e-01 4.33347106e-01
7.98626304e-01 8.36347163e-01 9.99517798e-01 2.81839430e-01
3.79408866e-01 6.70101643e-01 6.16622627e-01 -1.24888204e-01
3.45263891e-02 -4.22177196e-01 1.10331088e-01 7.33599603e-01
1.34112582e-01 -2.47669607e-01 -6.74057126e-01 4.54786986e-01
-2.23116922e+00 -6.78047836e-01 -2.48623237e-01 2.29697204e+00
9.87111330e-01 4.67355549e-01 -3.81338373e-02 2.66635954e-01
8.12107325e-01 1.46540463e-01 -6.93960190e-01 -2.11202145e-01
7.97502175e-02 -6.35362193e-02 2.02446550e-01 2.28138059e-01
-1.09550321e+00 9.81415927e-01 4.48806477e+00 9.96778071e-01
-1.10625219e+00 2.19176546e-01 9.52084899e-01 -1.13213463e-02
-4.97739196e-01 1.92670986e-01 -9.03861821e-01 6.24621451e-01
3.50589842e-01 3.55911762e-01 2.82972548e-02 9.77791905e-01
5.93736134e-02 -3.19367528e-01 -9.69921649e-01 6.40292108e-01
-1.34832099e-01 -1.21223569e+00 1.65695902e-02 -1.34134963e-01
8.27376664e-01 -1.28155187e-01 -1.74582690e-01 1.84812710e-01
1.92512289e-01 -5.93461037e-01 8.10513854e-01 1.44643933e-01
6.12888277e-01 -8.22503984e-01 8.07532907e-01 5.72897136e-01
-1.14179611e+00 -2.42696870e-02 -2.34592155e-01 3.55529457e-01
1.60658747e-01 7.96138883e-01 -6.31402969e-01 1.03010201e+00
7.07398713e-01 6.04330599e-01 -4.84015942e-01 7.29358971e-01
-5.11548758e-01 6.77438498e-01 -1.75851092e-01 2.45305851e-01
2.71172643e-01 -2.48142675e-01 4.94519919e-01 7.99557924e-01
-2.31690127e-02 -1.94172949e-01 2.05477610e-01 1.02969682e+00
8.11954960e-02 -3.86860482e-02 -2.67168492e-01 2.84790188e-01
3.65542054e-01 1.26417100e+00 -1.00947380e+00 -2.97626287e-01
-4.34184551e-01 1.02530336e+00 5.50737917e-01 4.28971350e-02
-9.61781561e-01 -5.20146750e-02 3.01100254e-01 -4.15395014e-02
4.31167126e-01 2.42080808e-01 -5.08234262e-01 -1.16106725e+00
3.62513542e-01 -7.51823902e-01 3.66472751e-01 -5.72300792e-01
-1.23190796e+00 6.36256039e-01 -1.66311376e-02 -1.39779973e+00
-9.70277190e-02 -2.59289116e-01 -6.33884311e-01 6.75440669e-01
-1.72422659e+00 -1.19113517e+00 -2.10059330e-01 3.41880411e-01
6.33999586e-01 1.13472782e-01 5.19714057e-01 2.59328902e-01
-8.11149716e-01 6.76302254e-01 -1.34144902e-01 -1.30753696e-01
7.87928820e-01 -1.26070356e+00 2.91710258e-01 9.50803459e-01
3.00036203e-02 4.32076126e-01 5.17189503e-01 -7.96362877e-01
-8.43375862e-01 -1.19962263e+00 7.48286605e-01 -3.94232348e-02
1.19415827e-01 -3.02380323e-01 -1.15916061e+00 4.85484928e-01
-1.67463496e-01 3.29560339e-01 5.80691934e-01 3.92318256e-02
-5.27941227e-01 -2.43734911e-01 -1.12549865e+00 4.37744319e-01
1.09375405e+00 -1.15710527e-01 -4.83164161e-01 9.61295590e-02
7.36642480e-01 -2.99829602e-01 -3.84138227e-01 7.32374847e-01
5.25656521e-01 -1.21523690e+00 6.51368141e-01 -5.06678045e-01
7.45682418e-01 -5.04201174e-01 -6.03314899e-02 -1.17728102e+00
-2.10344434e-01 -2.99745679e-01 7.65941367e-02 1.52643287e+00
5.27951419e-01 -5.80244064e-01 8.18595231e-01 5.73093534e-01
-3.36169869e-01 -1.28131628e+00 -1.00127888e+00 -9.98097956e-01
-1.30319059e-01 -3.25034797e-01 4.39960927e-01 9.95709658e-01
-1.68786228e-01 2.54477412e-01 -4.35576856e-01 1.93280980e-01
4.67463821e-01 3.22555602e-01 6.59170389e-01 -1.14488649e+00
-5.38958490e-01 -3.62435162e-01 3.36491759e-03 -9.66679037e-01
1.00896731e-01 -8.72770846e-01 3.56299400e-01 -1.51083064e+00
4.76386160e-01 -8.10377896e-01 -3.31666678e-01 8.53694260e-01
-5.01723647e-01 3.01118139e-02 2.18899518e-01 2.10202187e-01
-7.87277639e-01 7.37851799e-01 1.37011111e+00 2.53681272e-01
-2.36414179e-01 1.29947424e-01 -7.64453590e-01 8.26833427e-01
9.15081739e-01 -8.10007095e-01 -6.00854099e-01 -1.95076793e-01
1.75006866e-01 -8.62996057e-02 2.98304021e-01 -7.89918721e-01
9.88521725e-02 -1.78314984e-01 2.56235957e-01 -5.55212677e-01
8.48239437e-02 -9.17619348e-01 1.22834697e-01 6.64158344e-01
-2.88342386e-01 -1.94812253e-01 1.12233862e-01 6.13940060e-01
-2.12202534e-01 -3.29721212e-01 9.36212838e-01 -1.73007846e-01
-5.08628368e-01 1.06735833e-01 1.71383187e-01 1.37336954e-01
1.12004316e+00 -2.01596707e-01 -2.86021888e-01 1.25729889e-02
-4.46154237e-01 6.54272139e-01 5.19470990e-01 3.92905682e-01
3.49435866e-01 -1.28088045e+00 -4.42681044e-01 2.16843039e-01
1.88725129e-01 5.98816812e-01 3.62509042e-01 8.73421967e-01
-9.03097913e-02 2.14087084e-01 -1.23010442e-01 -7.41937280e-01
-1.43376195e+00 5.09355307e-01 1.04112044e-01 -6.37195647e-01
-4.62961704e-01 9.67192650e-01 2.88046420e-01 -6.56482816e-01
2.12005958e-01 -2.29299203e-01 -4.20298390e-02 -1.37225568e-01
1.37229890e-01 2.41366416e-01 3.71099673e-02 -4.30196255e-01
-5.18098354e-01 4.47398692e-01 -3.21699888e-01 1.26629602e-02
1.30256879e+00 -1.62660182e-02 2.64679417e-02 3.55217695e-01
1.09198630e+00 -8.28979239e-02 -1.62060332e+00 -1.98474094e-01
3.82155962e-02 -4.85223562e-01 -1.17253050e-01 -1.01728737e+00
-1.32797337e+00 7.60342658e-01 3.82861793e-01 2.32089171e-03
1.30595171e+00 9.90124345e-02 8.69903862e-01 -1.03619043e-02
2.64367670e-01 -1.20805335e+00 1.59589648e-01 2.49552041e-01
6.17386043e-01 -1.32044947e+00 -1.28990903e-01 -6.74964249e-01
-1.08074045e+00 8.29867184e-01 1.08013511e+00 5.17593883e-02
2.55794048e-01 1.01268068e-01 -5.60474023e-02 -1.00218914e-01
-7.65310526e-01 -9.52332765e-02 3.07200462e-01 2.46147931e-01
1.07087269e-01 -9.08090994e-02 -3.61682832e-01 1.00411010e+00
2.07591087e-01 -9.43964720e-03 1.42854080e-01 9.95667875e-01
-2.60863423e-01 -1.38683677e+00 -8.70755687e-02 3.51530164e-01
-3.03378075e-01 6.46725222e-02 -3.84151399e-01 5.58112144e-01
5.13903916e-01 6.73789680e-01 -1.37483135e-01 -4.75317448e-01
2.46367201e-01 2.02726290e-01 2.75299460e-01 -5.68133235e-01
-4.36326325e-01 2.86117226e-01 -1.02500334e-01 -6.83932602e-01
-6.04804099e-01 -7.04578519e-01 -1.49829292e+00 1.20158218e-01
-6.72524810e-01 5.05021811e-02 4.23925519e-01 1.09213388e+00
6.02430820e-01 6.25096560e-01 8.38695467e-01 -5.46264946e-01
-5.08370399e-01 -6.71207309e-01 -2.12089732e-01 4.12735075e-01
4.05728929e-02 -8.27294827e-01 -3.44673246e-01 -1.14400782e-01] | [9.4781494140625, 1.315173625946045] |
7eccebb2-0e38-454a-90e3-e551971b7e22 | improve-few-shot-voice-cloning-using-multi | 2203.09708 | null | https://arxiv.org/abs/2203.09708v1 | https://arxiv.org/pdf/2203.09708v1.pdf | Improve few-shot voice cloning using multi-modal learning | Recently, few-shot voice cloning has achieved a significant improvement. However, most models for few-shot voice cloning are single-modal, and multi-modal few-shot voice cloning has been understudied. In this paper, we propose to use multi-modal learning to improve the few-shot voice cloning performance. Inspired by the recent works on unsupervised speech representation, the proposed multi-modal system is built by extending Tacotron2 with an unsupervised speech representation module. We evaluate our proposed system in two few-shot voice cloning scenarios, namely few-shot text-to-speech(TTS) and voice conversion(VC). Experimental results demonstrate that the proposed multi-modal learning can significantly improve the few-shot voice cloning performance over their counterpart single-modal systems. | ['Yue Lin', 'Haitong Zhang'] | 2022-03-18 | null | null | null | null | ['voice-cloning'] | ['speech'] | [ 1.21952102e-01 -2.51205917e-02 -5.02470672e-01 -9.11233425e-02
-1.31102240e+00 -1.93010181e-01 6.63872719e-01 -4.73092556e-01
4.48663682e-02 4.92109954e-01 5.90789437e-01 -4.26880360e-01
3.02743465e-02 -3.42639208e-01 -2.57719368e-01 -6.15496039e-01
5.85061848e-01 3.27796072e-01 2.54214704e-01 -1.59378335e-01
1.85550898e-01 2.02608868e-01 -1.66633856e+00 1.09846190e-01
7.38143265e-01 5.25009394e-01 5.02875924e-01 1.18794465e+00
-6.50077462e-01 8.03157032e-01 -6.40017807e-01 -1.99541226e-01
-3.84735048e-01 -9.04677987e-01 -7.85502791e-01 1.58077389e-01
-1.55708879e-01 -8.44037682e-02 -3.51650387e-01 7.70209014e-01
1.29890406e+00 4.06523615e-01 7.03158021e-01 -1.27648401e+00
-6.96616828e-01 9.34845746e-01 -2.89414316e-01 2.28279635e-01
3.10789406e-01 -5.72367683e-02 9.77622211e-01 -1.01943612e+00
3.72549891e-01 1.51412165e+00 7.07805276e-01 9.75892544e-01
-8.11029315e-01 -6.16789639e-01 -2.16858774e-01 7.03751147e-02
-1.45481932e+00 -1.12139261e+00 7.71549642e-01 -3.08294117e-01
1.01158118e+00 3.83773386e-01 2.37021685e-01 9.93749917e-01
-8.81024599e-02 1.01154923e+00 7.29138374e-01 -7.52411723e-01
2.37510294e-01 2.10576504e-01 -2.43164867e-01 6.68087840e-01
-3.64610881e-01 4.17849451e-01 -7.00032830e-01 -1.15680598e-01
3.41743588e-01 -1.58843130e-01 1.00047514e-01 -6.16182685e-02
-8.85493279e-01 7.70513475e-01 -3.02659571e-01 8.23578358e-01
5.47564030e-03 -3.11277504e-03 4.87860978e-01 1.78296119e-01
4.96803939e-01 1.58836871e-01 -2.85397582e-02 -9.15350914e-01
-1.21988034e+00 -2.88928509e-01 7.42532074e-01 1.42016983e+00
3.71989399e-01 7.47815430e-01 -2.78931320e-01 1.52621996e+00
1.13070488e-01 3.91127676e-01 1.05948985e+00 -5.93996882e-01
5.29860631e-02 -8.79377872e-02 -2.99805582e-01 -1.34661451e-01
-2.17682049e-01 -1.93755046e-01 -7.79145300e-01 -3.53475332e-01
-2.45505184e-01 -3.30763638e-01 -1.00889564e+00 1.40573215e+00
1.10418439e-01 4.80060399e-01 1.95486397e-01 5.13135612e-01
1.16028678e+00 1.03423536e+00 -6.69767112e-02 -5.84959388e-01
1.03998232e+00 -1.43578744e+00 -1.07919300e+00 6.44604191e-02
7.65572369e-01 -1.36570907e+00 1.35431850e+00 3.89381871e-02
-7.72884309e-01 -5.73304117e-01 -8.50137532e-01 1.09689906e-01
-1.93934232e-01 7.62359872e-02 3.86297494e-01 1.37120378e+00
-4.96369302e-01 3.61594975e-01 -5.37874818e-01 -7.07189739e-01
2.58216530e-01 -4.45058057e-03 -2.83868667e-02 1.38256237e-01
-1.02978802e+00 4.64485556e-01 2.39626601e-01 -6.23247743e-01
-1.10682666e+00 -7.06893325e-01 -5.54873109e-01 3.74611616e-02
5.83554864e-01 -4.89623934e-01 1.75284648e+00 -1.94682181e-01
-2.30829167e+00 4.82874811e-01 -4.29440796e-01 3.48287374e-02
-3.35266255e-02 -8.46306831e-02 -8.94244373e-01 -1.34986952e-01
-2.96524763e-02 3.22500408e-01 1.08378756e+00 -1.29025722e+00
-9.00769353e-01 1.06853999e-01 -3.08034807e-01 6.30966499e-02
-5.35910785e-01 3.88305962e-01 -4.19522434e-01 -8.86922717e-01
-1.24362752e-01 -8.05794835e-01 1.81770712e-01 -5.08251548e-01
-6.19838178e-01 -1.52503043e-01 1.13382876e+00 -4.26325500e-01
1.52612531e+00 -2.24625969e+00 2.55692273e-01 -8.39755654e-01
-2.78578758e-01 5.11499286e-01 -2.16376022e-01 6.98594689e-01
-1.57102458e-02 1.93332031e-01 -1.93495899e-01 -8.82664323e-01
1.16785802e-01 2.96733141e-01 -4.24205124e-01 9.42157432e-02
-9.48707610e-02 6.98009610e-01 -5.75620890e-01 -6.47400260e-01
2.27242976e-01 1.31719068e-01 -5.66603661e-01 5.47974885e-01
-2.94360816e-02 4.59907949e-01 -3.38551737e-02 1.05218434e+00
4.27494526e-01 4.11982924e-01 -2.12936789e-01 2.25038692e-01
-2.85439968e-01 1.86479330e-01 -8.09829831e-01 1.82486510e+00
-9.09319222e-01 4.50806081e-01 -1.25259042e-01 -5.85663199e-01
8.61290574e-01 7.70974934e-01 2.55022109e-01 -1.85181797e-01
4.90442097e-01 2.11240843e-01 4.02744114e-02 -6.02353454e-01
8.57628047e-01 -1.15175891e+00 -9.11294594e-02 5.89189172e-01
6.35818481e-01 -5.34232676e-01 -2.92398930e-01 2.10300162e-01
6.34107292e-01 -1.58196673e-01 2.65025437e-01 2.61568904e-01
5.94048023e-01 -4.00331974e-01 6.82284713e-01 6.30224526e-01
-3.29834163e-01 9.80131149e-01 1.56895667e-01 5.27029037e-01
-1.10189760e+00 -9.68490005e-01 1.16183542e-01 1.66061592e+00
1.12024518e-02 -6.18374527e-01 -8.55799556e-01 -2.48631597e-01
-2.69115150e-01 1.30197906e+00 1.18891358e-01 -5.93680024e-01
-3.08572352e-01 -3.94951791e-01 1.40352249e+00 3.80164683e-01
-1.27148647e-02 -9.28927779e-01 3.61707717e-01 3.58905017e-01
-3.47014755e-01 -1.03544748e+00 -7.06718266e-01 2.20233351e-01
-6.70178115e-01 -3.32245320e-01 -1.07903671e+00 -1.25673544e+00
-1.09299466e-01 4.86653328e-01 4.51864898e-01 -3.90206516e-01
-3.79160345e-01 5.22188485e-01 -7.45790124e-01 -4.99450505e-01
-1.01687169e+00 3.99505198e-01 4.92202699e-01 -1.80011950e-02
-5.82160130e-02 -6.32372022e-01 2.64364719e-01 3.52283537e-01
-7.41119862e-01 -2.58316666e-01 4.99689728e-01 1.08231056e+00
3.52026522e-01 1.70949996e-01 1.03013098e+00 -6.59979939e-01
9.30355668e-01 -3.87467146e-01 7.68890828e-02 3.32508624e-01
-5.54162979e-01 -1.02328695e-01 8.19277942e-01 -5.81947625e-01
-1.44118285e+00 -2.62623191e-01 -4.98727560e-01 -7.65057087e-01
-1.30625054e-01 2.21332833e-01 -4.18782532e-01 -8.52776915e-02
3.19228292e-01 7.60200500e-01 -5.19473366e-02 -1.13313937e+00
9.77409363e-01 1.54498029e+00 6.12542987e-01 -6.87487006e-01
7.11317182e-01 -4.95912693e-02 -5.03073692e-01 -1.37838769e+00
-4.82344300e-01 -8.04597080e-01 -6.31717443e-01 -1.94372013e-01
6.44541264e-01 -7.17296898e-01 -3.89233619e-01 7.87350416e-01
-1.00848281e+00 1.49122924e-01 -4.87418085e-01 6.86567307e-01
-1.12465072e+00 5.51974654e-01 -5.46184659e-01 -1.26575792e+00
-3.20534855e-01 -1.08785605e+00 1.13496995e+00 1.41937822e-01
-3.06708869e-02 -7.45714843e-01 3.19575161e-01 4.73243654e-01
5.36216378e-01 -8.18100333e-01 1.28095591e+00 -7.88495541e-01
-1.42914876e-01 -6.90947250e-02 2.06364423e-01 3.26800019e-01
6.59481406e-01 -1.32211044e-01 -1.12059295e+00 -4.15207475e-01
1.05523370e-01 -3.45827669e-01 6.45222485e-01 -3.51767614e-02
6.52662992e-01 5.27912518e-03 -3.02591443e-01 4.97452825e-01
7.31908500e-01 4.53193247e-01 3.91242802e-01 -2.10977748e-01
8.15350711e-01 1.23423114e-01 8.62503171e-01 7.87247479e-01
-6.60708770e-02 6.36658311e-01 1.36538714e-01 3.56755912e-01
-7.23668873e-01 -4.90200996e-01 4.17362601e-01 1.87862706e+00
2.13897988e-01 -5.42904019e-01 -6.44399524e-01 7.80211329e-01
-1.46739364e+00 -9.89324570e-01 3.37799817e-01 1.88652897e+00
6.47516370e-01 4.59890813e-02 2.84294456e-01 2.14412928e-01
1.26518381e+00 6.55581772e-01 -3.76897335e-01 -6.28318310e-01
-1.38137773e-01 2.19211549e-01 -1.81755386e-02 4.82338667e-01
-8.49528313e-01 1.48143995e+00 6.85121536e+00 1.45515943e+00
-1.24672341e+00 5.38370192e-01 2.99878530e-02 -2.73385942e-01
-2.75303900e-01 2.61189550e-01 -9.54392910e-01 3.49180549e-01
1.16002214e+00 -7.63070285e-01 5.97544551e-01 8.97748947e-01
1.01478972e-01 3.78222108e-01 -8.21190178e-01 1.16424716e+00
2.85215855e-01 -1.07696462e+00 3.05903256e-01 -2.52637178e-01
5.28926492e-01 -2.32782722e-01 8.22051838e-02 7.93279529e-01
-7.61101246e-02 -7.54559815e-01 6.59217536e-01 4.96215045e-01
1.22617698e+00 -1.04988253e+00 3.60575378e-01 7.04754472e-01
-1.41242635e+00 -1.81277484e-01 -4.83312935e-01 2.24441528e-01
6.89461052e-01 3.33618313e-01 -1.18477285e+00 6.66351914e-01
8.49010944e-02 4.68749702e-01 -3.40108424e-02 1.04413295e+00
1.32897273e-01 1.05704856e+00 6.37400895e-02 -2.44021609e-01
1.13971591e-01 1.29850045e-01 1.08528614e+00 1.18957078e+00
7.82804489e-01 4.35595773e-02 -4.56519648e-02 5.10854542e-01
-1.77493423e-01 4.78895664e-01 -6.00937247e-01 -9.33538616e-01
7.19882488e-01 9.01315391e-01 -2.28596836e-01 -4.61336434e-01
-4.34947968e-01 1.20642042e+00 -5.00301830e-02 3.84165160e-02
-8.22640598e-01 -8.40766251e-01 8.20988476e-01 -8.66489410e-02
6.84293151e-01 -1.18101098e-01 6.19908515e-03 -1.00156152e+00
-5.42865098e-01 -7.14959621e-01 8.95756930e-02 -9.25873756e-01
-1.29576933e+00 5.40179968e-01 -4.47322190e-01 -1.51558042e+00
-4.98575300e-01 -3.77474166e-02 -8.98775697e-01 5.69736183e-01
-1.19037318e+00 -1.42277110e+00 3.57310683e-01 4.50141340e-01
1.10789680e+00 -1.03163838e+00 1.09695208e+00 5.95649004e-01
-7.45967746e-01 1.05062366e+00 4.69568640e-01 -2.88967848e-01
9.52046692e-01 -7.62731493e-01 3.58304411e-01 5.87720275e-01
1.52397513e-01 5.29099286e-01 5.43081462e-01 -5.48071146e-01
-1.52427793e+00 -1.16130865e+00 9.32216346e-01 1.98975846e-01
6.94867253e-01 -2.35718459e-01 -6.93526387e-01 4.60333526e-01
2.63938904e-01 -2.12260097e-01 1.05755389e+00 1.92843735e-01
-2.55393922e-01 4.27593514e-02 -1.04014301e+00 6.65157914e-01
1.10113716e+00 -1.07041121e+00 -9.80026424e-01 1.31765366e-01
1.36428547e+00 5.66825503e-03 -8.16059709e-01 9.43551436e-02
4.09744829e-01 -6.70356631e-01 6.72045887e-01 -9.68662143e-01
-7.69407314e-04 1.33442311e-02 -5.90233445e-01 -1.53236079e+00
-5.01142323e-01 -1.10824072e+00 2.76908427e-02 2.02178574e+00
1.11204684e-01 -3.01282972e-01 6.14166737e-01 -2.85670400e-01
-7.30496228e-01 -3.45832884e-01 -1.43056715e+00 -1.34422493e+00
2.38571882e-01 -5.05005836e-01 7.15523005e-01 7.97086596e-01
2.53143847e-01 8.25322151e-01 -1.03688025e+00 -2.05051638e-02
2.88988620e-01 1.60519287e-01 7.55638719e-01 -8.68260562e-01
-4.91359890e-01 -3.04246247e-01 -3.20659816e-01 -1.12690163e+00
2.59185135e-01 -1.17123592e+00 3.78662765e-01 -1.07519209e+00
2.95377910e-01 -9.53532979e-02 -3.61701429e-01 2.64084727e-01
-4.44293767e-02 -1.49324968e-01 3.58379424e-01 4.17341664e-02
-4.13319677e-01 1.20846367e+00 1.16115117e+00 -2.66848505e-01
-3.00219446e-01 2.19535083e-01 -6.20170116e-01 6.92954063e-01
6.44194126e-01 -7.47474015e-01 -3.14841211e-01 7.44713917e-02
-7.58212090e-01 3.81819695e-01 -3.50119978e-01 -1.10311031e+00
5.29420137e-01 -9.13368687e-02 -5.79152226e-01 -8.78435433e-01
6.51222408e-01 -4.31141466e-01 2.56678816e-02 3.13223362e-01
-1.73118442e-01 -5.99172592e-01 1.97420821e-01 6.52570069e-01
-3.45595121e-01 -6.48254991e-01 9.59207237e-01 -3.28943133e-02
-7.40959585e-01 1.36628717e-01 -1.01073217e+00 8.94780904e-02
8.08796108e-01 -1.65780813e-01 -1.42388985e-01 -5.94635963e-01
-9.13788795e-01 -3.65043312e-01 4.68981341e-02 8.39636803e-01
7.37667918e-01 -1.59887230e+00 -3.78860295e-01 1.64979726e-01
4.48250711e-01 -7.24233508e-01 3.48183662e-01 6.82943165e-01
1.47422671e-01 5.79709709e-01 1.66132655e-02 -5.34505486e-01
-1.50410295e+00 6.87460124e-01 1.23824656e-01 3.46284539e-01
-4.23099339e-01 1.09061933e+00 -5.86714931e-02 -8.97915065e-01
3.75857204e-01 1.97480589e-01 -4.04733280e-03 1.34328961e-01
3.70276392e-01 7.81597257e-01 -1.62232086e-01 -9.72883224e-01
-2.53839195e-01 6.37093246e-01 -1.07652113e-01 -3.86602104e-01
1.13817573e+00 -4.91687626e-01 3.16622965e-02 1.07273173e+00
1.32678449e+00 3.70331943e-01 -3.99610370e-01 -3.99935603e-01
-2.10957214e-01 -5.40108919e-01 9.70055014e-02 -5.77023685e-01
-8.01022530e-01 1.18162858e+00 5.61197937e-01 -3.95994484e-02
7.67437339e-01 1.11905254e-01 1.56789804e+00 4.00688946e-01
3.66647661e-01 -1.16399062e+00 2.35438436e-01 8.68385494e-01
6.30029321e-01 -9.07102585e-01 -2.93307602e-01 -5.60139537e-01
-7.44219959e-01 9.77899194e-01 4.03164297e-01 5.59578121e-01
7.03954041e-01 3.07456195e-01 1.67618528e-01 2.70041525e-01
-8.97841811e-01 -4.97180849e-01 -2.37618778e-02 6.10077560e-01
4.25323725e-01 1.70221165e-01 -2.56339014e-01 1.19278836e+00
-3.77034813e-01 2.44163070e-02 5.23223758e-01 9.53145146e-01
-9.27925467e-01 -1.40288949e+00 -5.22283852e-01 2.50253469e-01
-2.46383190e-01 -2.88851947e-01 -6.52189434e-01 1.84693381e-01
-1.09625652e-01 1.37454355e+00 -1.08893596e-01 -9.17089283e-01
3.29650402e-01 6.70471251e-01 3.25335443e-01 -8.89347076e-01
-3.88772547e-01 4.71314192e-01 3.60127479e-01 1.10897064e-01
-5.52564971e-02 -4.94634807e-01 -1.26745284e+00 -3.99640173e-01
-8.70970070e-01 1.58848405e-01 7.09578693e-01 9.53768253e-01
4.72498834e-01 9.01080132e-01 1.05569017e+00 -5.92705548e-01
-5.79834104e-01 -1.14039981e+00 -1.08008504e+00 -1.64739758e-01
9.65117887e-02 -7.56183743e-01 -4.87525284e-01 -5.91492541e-02] | [14.830669403076172, 6.676365852355957] |
f141b18f-4af0-4e6a-80b0-f7540e950b22 | sample-efficient-optimisation-with | 2205.13902 | null | https://arxiv.org/abs/2205.13902v2 | https://arxiv.org/pdf/2205.13902v2.pdf | Sample-Efficient Optimisation with Probabilistic Transformer Surrogates | Faced with problems of increasing complexity, recent research in Bayesian Optimisation (BO) has focused on adapting deep probabilistic models as flexible alternatives to Gaussian Processes (GPs). In a similar vein, this paper investigates the feasibility of employing state-of-the-art probabilistic transformers in BO. Upon further investigation, we observe two drawbacks stemming from their training procedure and loss definition, hindering their direct deployment as proxies in black-box optimisation. First, we notice that these models are trained on uniformly distributed inputs, which impairs predictive accuracy on non-uniform data - a setting arising from any typical BO loop due to exploration-exploitation trade-offs. Second, we realise that training losses (e.g., cross-entropy) only asymptotically guarantee accurate posterior approximations, i.e., after arriving at the global optimum, which generally cannot be ensured. At the stationary points of the loss function, however, we observe a degradation in predictive performance especially in exploratory regions of the input space. To tackle these shortcomings we introduce two components: 1) a BO-tailored training prior supporting non-uniformly distributed points, and 2) a novel approximate posterior regulariser trading-off accuracy and input sensitivity to filter favourable stationary points for improved predictive performance. In a large panel of experiments, we demonstrate, for the first time, that one transformer pre-trained on data sampled from random GP priors produces competitive results on 16 benchmark black-boxes compared to GP-based BO. Since our model is only pre-trained once and used in all tasks without any retraining and/or fine-tuning, we report an order of magnitude time-reduction, while matching and sometimes outperforming GPs. | ['Haitham Bou Ammar', 'Jun Wang', 'Rasul Tutunov', 'Antoine Grosnit', 'Matthieu Zimmer', 'Alexandre Maraval'] | 2022-05-27 | null | null | null | null | ['bayesian-optimisation'] | ['methodology'] | [ 3.08415622e-01 3.37627113e-01 2.08276451e-01 -5.33354208e-02
-1.04335904e+00 -4.72901464e-01 8.61287296e-01 4.27910328e-01
-6.51434183e-01 9.34580028e-01 -9.55444500e-02 -2.41551608e-01
-7.37463713e-01 -6.78688407e-01 -8.36201012e-01 -1.10962117e+00
-1.80996522e-01 8.75517607e-01 2.83632487e-01 7.17754737e-02
2.38685876e-01 4.66467977e-01 -1.65707004e+00 -8.37113336e-02
9.46196258e-01 1.20939171e+00 2.35707149e-01 6.54544175e-01
1.77199468e-01 1.55207619e-01 -6.54855609e-01 -6.69563055e-01
2.59563923e-01 -3.80002744e-02 -5.74696124e-01 -1.65172175e-01
-5.07232500e-03 1.71956837e-01 1.21202193e-01 8.89284015e-01
6.95243180e-01 3.21281731e-01 8.32652450e-01 -9.39989746e-01
-1.41803429e-01 3.89680892e-01 -3.18008602e-01 1.18324809e-01
1.08222872e-01 4.41875964e-01 9.36954319e-01 -5.64449310e-01
1.88600808e-01 1.22381175e+00 1.01312578e+00 3.62829685e-01
-1.78216863e+00 -9.44871232e-02 5.03099374e-02 -3.09481502e-01
-1.43350661e+00 -4.61803824e-01 3.42372537e-01 -3.26557636e-01
8.58072817e-01 4.40262198e-01 4.39354837e-01 1.35329640e+00
2.88949788e-01 7.53378332e-01 1.18613517e+00 -3.35696757e-01
6.31024897e-01 3.94852221e-01 -3.14352840e-01 6.98733777e-02
1.89796224e-01 4.34453219e-01 -3.37614685e-01 -4.09596533e-01
5.13503313e-01 -3.82654130e-01 -4.39140081e-01 -6.66655719e-01
-8.93536389e-01 7.82059193e-01 2.49716967e-01 1.81961536e-01
-5.64433038e-01 2.04563677e-01 2.81287253e-01 5.45368306e-02
6.11443758e-01 5.74398637e-01 -6.07308030e-01 -5.94687164e-01
-1.24895167e+00 5.18285930e-01 9.61953104e-01 8.16196084e-01
4.29496348e-01 -2.59176344e-01 -4.17406946e-01 7.62654364e-01
4.16178703e-01 3.66169244e-01 4.41469103e-01 -5.35265982e-01
4.65152472e-01 1.95408270e-01 3.14748883e-01 -8.02357018e-01
-3.88941318e-01 -1.00933671e+00 -7.39603639e-01 2.75009423e-01
7.39904642e-01 -8.12498406e-02 -7.82349348e-01 1.66513288e+00
3.76533896e-01 -9.77192074e-02 2.50150282e-02 7.31639326e-01
-5.45645393e-02 5.14259398e-01 2.56283253e-01 -2.09902942e-01
1.36791480e+00 -5.82657397e-01 -2.62741625e-01 -1.94487035e-01
2.33642265e-01 -6.63434267e-01 1.11831021e+00 8.97137403e-01
-1.18478537e+00 -4.53141451e-01 -9.30408061e-01 3.43925238e-01
-3.40089619e-01 -9.48025957e-02 3.98461401e-01 9.53201771e-01
-1.00967705e+00 1.14957023e+00 -1.07071245e+00 -1.81600198e-01
5.10635614e-01 4.44833487e-01 -1.64832883e-02 1.73904046e-01
-9.89342511e-01 8.92111599e-01 6.21463478e-01 2.38717422e-01
-8.20878327e-01 -8.91834497e-01 -4.69564289e-01 3.65911424e-01
5.17658651e-01 -7.24628210e-01 1.22356987e+00 -8.99161935e-01
-1.95548820e+00 4.62125003e-01 4.04864289e-02 -8.49011600e-01
1.13980544e+00 -4.68682081e-01 -1.19393528e-01 -1.60298333e-01
-2.79328346e-01 4.58277076e-01 1.16175687e+00 -1.34391403e+00
-6.34442091e-01 -2.36095324e-01 -8.09105411e-02 7.34506547e-02
-7.59430677e-02 -1.55179054e-01 -3.83466691e-01 -6.48224354e-01
-9.69596654e-02 -8.11262786e-01 -4.87055749e-01 -4.02637511e-01
-3.58222902e-01 -1.86128423e-01 3.11835855e-01 -5.09377599e-01
1.26459086e+00 -1.99035931e+00 2.68911272e-01 4.34647113e-01
-6.62589893e-02 3.60753596e-01 2.26393342e-01 4.08666909e-01
-8.82264879e-03 1.49659872e-01 -6.80510998e-01 -6.20179772e-01
3.46023709e-01 3.79029781e-01 -4.15173709e-01 5.95436454e-01
5.66990137e-01 7.28478193e-01 -1.08262300e+00 -2.11370975e-01
2.52715915e-01 7.29530692e-01 -4.78290677e-01 5.80925774e-03
-4.63147432e-01 6.45059645e-01 -3.97952467e-01 3.59885633e-01
5.88908851e-01 2.93762144e-02 -1.20422512e-01 2.57619262e-01
4.35824730e-02 3.19928050e-01 -1.34131372e+00 1.50609136e+00
-6.71763361e-01 3.43128264e-01 1.45915849e-03 -1.08972692e+00
8.75530124e-01 2.94340849e-01 1.77536309e-01 -4.49769706e-01
8.56426284e-02 4.06969935e-01 -1.59425408e-01 9.26446989e-02
5.54446280e-01 -2.74378419e-01 1.46304771e-01 4.42292765e-02
1.62774354e-01 -2.07840651e-01 1.62395164e-01 -4.77664471e-01
1.00340652e+00 6.29243910e-01 1.57266945e-01 -4.97171700e-01
5.77019095e-01 -2.79625565e-01 3.45349908e-01 9.85385418e-01
1.11514606e-01 7.92724252e-01 8.08384955e-01 -1.70977205e-01
-9.36045945e-01 -1.03693080e+00 -6.18468583e-01 9.58564460e-01
-2.78863490e-01 -3.62912446e-01 -6.10855579e-01 -6.13541007e-01
7.15334713e-02 1.05624139e+00 -5.71140885e-01 -8.32595453e-02
-3.90515983e-01 -1.17366779e+00 4.29430991e-01 2.49775082e-01
4.93646339e-02 -9.51332510e-01 -1.10494864e+00 5.45239031e-01
2.95753807e-01 -6.20317936e-01 1.65764391e-01 6.78081453e-01
-1.08392227e+00 -6.47729099e-01 -9.81598437e-01 9.54292715e-02
4.27986026e-01 -5.32092690e-01 1.41602635e+00 -3.69987577e-01
1.49935484e-01 2.81093240e-01 -1.37565807e-01 -6.14198804e-01
-4.53085154e-01 2.96191722e-01 -1.91646993e-01 -6.18094616e-02
1.10940292e-01 -7.72371113e-01 -6.67548120e-01 4.31473494e-01
-8.58385861e-01 -3.70543212e-01 8.02004337e-01 9.89327550e-01
5.26178837e-01 2.42146432e-01 2.74185002e-01 -6.09106600e-01
7.36161232e-01 -4.56512362e-01 -9.14202511e-01 1.26010478e-01
-7.02387393e-01 4.98769432e-01 5.19829929e-01 -4.83941287e-01
-1.01583540e+00 -1.79290250e-01 -1.98212594e-01 -2.66782939e-01
-3.13311934e-01 5.42200744e-01 -5.72833791e-02 1.67931110e-01
7.81291485e-01 1.11391209e-01 -5.94200790e-02 -6.43736541e-01
3.24623495e-01 4.98856574e-01 4.92055386e-01 -9.33171213e-01
6.17571235e-01 3.26713145e-01 2.49488220e-01 -5.73408008e-01
-7.03883410e-01 -4.80502516e-01 -4.75790411e-01 6.67261630e-02
4.71850574e-01 -5.97297192e-01 -7.10321069e-01 2.54762828e-01
-9.15167093e-01 -4.36403304e-01 -6.03667438e-01 3.01121801e-01
-8.29261482e-01 2.31865495e-01 -1.62445024e-01 -1.29584575e+00
-1.87085718e-01 -1.11475253e+00 1.22304034e+00 1.22125544e-01
-2.56052613e-01 -1.08497131e+00 1.84922189e-01 -2.56342106e-02
5.32350719e-01 4.21768427e-01 5.29732466e-01 -7.91271448e-01
-3.06336492e-01 -3.64584684e-01 2.75510736e-02 4.41392481e-01
-3.02231014e-01 -7.14476686e-03 -1.33138299e+00 -3.97505641e-01
2.93965220e-01 -7.10685253e-02 8.15632164e-01 4.30131048e-01
1.08430743e+00 -2.25809589e-01 -3.57003629e-01 5.80521405e-01
1.33208036e+00 -5.99944741e-02 6.26597106e-01 6.16748273e-01
1.42271549e-01 7.92084754e-01 5.35735428e-01 4.92328584e-01
-9.01734307e-02 8.04505646e-01 5.12462556e-01 2.36799881e-01
3.17406386e-01 -2.01744765e-01 2.20763430e-01 6.79432750e-02
-2.61932462e-01 -1.79387629e-01 -1.00866830e+00 6.04329050e-01
-1.92569613e+00 -6.09267116e-01 -2.81865243e-03 2.82121515e+00
1.00918233e+00 6.22776031e-01 3.07328939e-01 1.90750569e-01
4.37951326e-01 -5.61369136e-02 -3.25448722e-01 -3.31962734e-01
1.43972039e-01 4.85160440e-01 7.74335206e-01 4.30775523e-01
-1.20742607e+00 4.29881215e-01 5.24141312e+00 1.25295162e+00
-1.00071704e+00 1.95789218e-01 8.10363591e-01 -2.25881308e-01
-1.04270801e-01 9.16676819e-02 -7.43844569e-01 7.83647656e-01
1.32147515e+00 1.35447130e-01 3.80310684e-01 9.10622418e-01
1.18346334e-01 -3.04907084e-01 -1.12152839e+00 7.19077587e-01
-4.55289990e-01 -8.27271044e-01 -3.38152200e-01 3.34887236e-01
5.84332645e-01 9.17731822e-02 1.55691475e-01 4.63348031e-01
3.11589867e-01 -1.30109847e+00 1.06736910e+00 6.89256787e-01
3.72456968e-01 -1.01449025e+00 1.01875520e+00 5.29897988e-01
-6.32224858e-01 -1.03019126e-01 -5.02483606e-01 1.26817569e-01
2.48736739e-01 9.54254448e-01 -6.74769640e-01 9.50546324e-01
7.42400408e-01 1.81950912e-01 -3.98706555e-01 1.39842892e+00
-3.06962460e-01 6.87808752e-01 -8.60067129e-01 -3.59232421e-04
4.75275517e-01 -2.92134315e-01 8.64315927e-01 1.34427619e+00
4.10681665e-01 -5.32479048e-01 -3.34916234e-01 1.09518790e+00
3.99086058e-01 -2.43392497e-01 -2.45689362e-01 2.01449975e-01
1.93872362e-01 1.07573318e+00 -7.21096039e-01 4.69471551e-02
5.56583237e-03 5.78006208e-01 2.21187830e-01 3.88159871e-01
-8.32177758e-01 -2.73481607e-01 3.67355883e-01 7.83065185e-02
7.81333923e-01 -2.77411547e-02 -4.44576353e-01 -6.28853738e-01
2.41857395e-01 -7.37647831e-01 2.22617045e-01 -3.36547434e-01
-1.31885731e+00 6.16729021e-01 3.33584398e-01 -8.75651896e-01
-4.21390355e-01 -6.42959654e-01 -4.85685259e-01 1.17259955e+00
-1.47980154e+00 -8.22831631e-01 1.27920434e-01 5.11626638e-02
1.68102369e-01 2.38811418e-01 6.83261037e-01 5.47660924e-02
-3.51960748e-01 5.50459325e-01 5.06900311e-01 -5.87891102e-01
3.35656285e-01 -1.58631861e+00 4.77236152e-01 6.97429240e-01
1.81999639e-01 5.69935501e-01 1.24521029e+00 -3.28177750e-01
-1.01807058e+00 -8.23288739e-01 6.82223558e-01 -8.31166089e-01
7.21196115e-01 -4.71001953e-01 -1.05972028e+00 2.75741756e-01
-1.70229107e-01 -9.48307663e-02 2.59665221e-01 2.84218967e-01
1.77615657e-01 4.99688834e-03 -1.20676041e+00 5.62918901e-01
6.83155119e-01 -7.82073289e-02 -6.52835667e-01 2.10624397e-01
3.11321765e-01 -5.49076140e-01 -9.69172001e-01 5.37710786e-01
4.31822270e-01 -1.19081676e+00 1.06283939e+00 -3.25673312e-01
2.11154699e-01 -2.02684209e-01 5.56323342e-02 -1.29619324e+00
1.10349208e-01 -1.13890409e+00 -3.13452125e-01 1.34538376e+00
3.75099599e-01 -8.84126127e-01 7.31056452e-01 6.73720837e-01
-3.43117081e-02 -1.18644059e+00 -1.35853159e+00 -8.28505516e-01
2.42947727e-01 -8.22871089e-01 5.43736935e-01 2.67245203e-01
-4.01404589e-01 6.21248484e-02 -1.06711671e-01 1.98116139e-01
5.37608266e-01 -3.08648646e-01 8.08902979e-01 -1.28413785e+00
-7.69698083e-01 -8.21593940e-01 -3.08739811e-01 -1.15340793e+00
-3.21380019e-01 -3.71513128e-01 3.90993565e-01 -1.11978924e+00
-2.59615481e-01 -7.11413264e-01 -3.74577016e-01 1.27648219e-01
-3.33625942e-01 -8.94957557e-02 2.35167816e-02 2.05108345e-01
-3.56639862e-01 7.51370013e-01 8.40352297e-01 2.54163593e-01
-3.23412538e-01 5.65386236e-01 -5.13749838e-01 6.98449671e-01
6.94879353e-01 -6.84647202e-01 -3.84169817e-01 -6.14213124e-02
5.12528300e-01 -1.16829328e-01 6.14250302e-01 -1.17253077e+00
-2.17178520e-02 1.39438212e-01 3.32144350e-01 -4.50928807e-01
3.29027086e-01 -8.46467614e-01 2.92626679e-01 2.98676401e-01
-7.84052759e-02 -1.76731706e-01 4.41474706e-01 8.61536384e-01
-8.61569792e-02 -7.22408354e-01 7.04954803e-01 -6.30109832e-02
-6.50380626e-02 -1.64513230e-01 -3.58249664e-01 -3.73824872e-02
8.11245143e-01 -3.07178408e-01 1.30930459e-02 -1.71804041e-01
-7.48397052e-01 2.54448056e-01 4.55594212e-01 6.26420528e-02
8.26490894e-02 -8.74596655e-01 -6.05852544e-01 8.71068910e-02
-1.41903296e-01 5.89415193e-01 4.54345383e-02 1.15041983e+00
-3.90101135e-01 5.48082054e-01 2.33374298e-01 -8.85489523e-01
-7.80766129e-01 5.05072057e-01 3.60453337e-01 -8.34925830e-01
-5.19954920e-01 9.24858451e-01 -3.80791724e-02 -4.06078815e-01
5.17266572e-01 -6.46303415e-01 6.97947964e-02 3.29110096e-03
1.51671663e-01 3.65778565e-01 3.19680542e-01 -2.14707419e-01
-2.48461992e-01 2.92630881e-01 2.29968905e-01 -2.95688480e-01
1.35840094e+00 1.26367956e-02 1.77303523e-01 5.15935421e-01
6.59760714e-01 -4.74403836e-02 -1.83236563e+00 -4.26054560e-02
5.05185843e-01 -5.09367824e-01 9.75544751e-02 -8.82967472e-01
-5.37913859e-01 9.39995110e-01 4.91661787e-01 4.63696539e-01
1.00518131e+00 -1.96072966e-01 2.76895523e-01 1.85013294e-01
3.51610065e-01 -1.03443801e+00 -4.18414295e-01 3.35265964e-01
9.46454883e-01 -9.75168645e-01 9.58607718e-02 -8.48631095e-03
-6.03983283e-01 1.01835370e+00 -6.58865944e-02 -6.99260607e-02
4.20545995e-01 7.50595555e-02 -3.22541088e-01 -1.11504368e-01
-7.69536495e-01 -1.13301262e-01 5.07341325e-01 6.78612232e-01
1.63665488e-01 -6.97965696e-02 -1.39039800e-01 6.27000034e-01
-4.42377836e-01 -1.42716661e-01 -9.64243412e-02 9.15121675e-01
-1.84951812e-01 -1.04794371e+00 -5.65207839e-01 3.90772194e-01
-7.00250387e-01 -1.71691850e-01 7.70833641e-02 9.83857274e-01
-7.78243179e-03 8.02157879e-01 6.71216100e-02 1.46712869e-01
4.40863818e-01 1.91806823e-01 4.22194391e-01 -3.58167052e-01
-7.59088516e-01 2.22387657e-01 4.65533212e-02 -5.48751533e-01
-2.25244388e-01 -8.38413298e-01 -5.57771325e-01 -7.09857270e-02
-5.21395385e-01 3.04940999e-01 8.42695713e-01 9.89017129e-01
2.66980231e-01 7.03512132e-01 3.63102518e-02 -1.18325901e+00
-1.30525470e+00 -1.16095448e+00 -3.80829662e-01 1.38242021e-01
2.79649198e-01 -8.59714031e-01 -5.80324173e-01 -2.69039035e-01] | [6.456062316894531, 3.799975872039795] |
32113c37-ddfc-4655-89f7-cbe3bba4cf38 | the-single-noun-prior-for-image-clustering | 2104.03952 | null | https://arxiv.org/abs/2104.03952v2 | https://arxiv.org/pdf/2104.03952v2.pdf | Dataset Summarization by K Principal Concepts | We propose the new task of K principal concept identification for dataset summarizarion. The objective is to find a set of K concepts that best explain the variation within the dataset. Concepts are high-level human interpretable terms such as "tiger", "kayaking" or "happy". The K concepts are selected from a (potentially long) input list of candidates, which we denote the concept-bank. The concept-bank may be taken from a generic dictionary or constructed by task-specific prior knowledge. An image-language embedding method (e.g. CLIP) is used to map the images and the concept-bank into a shared feature space. To select the K concepts that best explain the data, we formulate our problem as a K-uncapacitated facility location problem. An efficient optimization technique is used to scale the local search algorithm to very large concept-banks. The output of our method is a set of K principal concepts that summarize the dataset. Our approach provides a more explicit summary in comparison to selecting K representative images, which are often ambiguous. As a further application of our method, the K principal concepts can be used to classify the dataset into K groups. Extensive experiments demonstrate the efficacy of our approach. | ['Yedid Hoshen', 'Niv Cohen'] | 2021-04-08 | null | null | null | null | ['image-clustering'] | ['computer-vision'] | [ 2.29703993e-01 -1.23595051e-01 -5.71328402e-01 -2.75027424e-01
-1.13884544e+00 -7.97524869e-01 3.89778107e-01 4.11891013e-01
-1.31107643e-01 4.28205252e-01 4.52040583e-01 -5.69290146e-02
-5.66393614e-01 -4.62962300e-01 -6.25826657e-01 -9.35259998e-01
1.39902636e-01 4.97977823e-01 -3.25456709e-01 -7.54497806e-03
6.48709476e-01 5.31127155e-01 -1.66447151e+00 4.42588836e-01
7.87017643e-01 9.90882933e-01 5.47642589e-01 5.03468454e-01
-2.88761735e-01 5.03059089e-01 -6.80566907e-01 -6.11265041e-02
4.46970195e-01 -4.46459383e-01 -1.03729177e+00 8.09421778e-01
2.66933560e-01 1.96988717e-01 6.98853582e-02 9.90250051e-01
2.73771197e-01 2.51812816e-01 8.93003047e-01 -1.27751613e+00
-6.06425643e-01 5.26208639e-01 -5.84515274e-01 2.74505883e-01
3.10297757e-01 -4.09535050e-01 1.21362019e+00 -1.28548455e+00
7.16280997e-01 1.28576040e+00 1.41495809e-01 4.64544475e-01
-1.40342069e+00 -4.75984514e-01 4.33719665e-01 2.14433819e-01
-1.81987214e+00 -3.62858117e-01 1.10938084e+00 -4.87509578e-01
6.73392355e-01 6.54031873e-01 5.14290273e-01 4.14294958e-01
1.80468127e-01 1.06868982e+00 7.57007003e-01 -7.46374428e-01
5.20661056e-01 5.92877030e-01 2.11061940e-01 4.78333741e-01
3.20519060e-01 -5.13888538e-01 -7.31175840e-01 -5.44415355e-01
5.18315494e-01 1.74261063e-01 -4.34519172e-01 -8.38130414e-01
-1.36550498e+00 1.02957141e+00 3.87029111e-01 1.11802682e-01
-3.69887948e-01 7.87146762e-02 2.79546916e-01 2.22432077e-01
3.35127175e-01 8.35576296e-01 -4.74327385e-01 3.51468682e-01
-6.64584100e-01 2.74522483e-01 4.67323124e-01 1.06112266e+00
1.12092829e+00 -3.12836021e-01 8.73778537e-02 7.97394812e-01
-5.57246916e-02 2.21705303e-01 5.04611850e-01 -9.12128508e-01
6.02440119e-01 7.75339544e-01 1.26132995e-01 -1.23809981e+00
-3.05578232e-01 -1.61588252e-01 -7.57590353e-01 -2.71628082e-01
-2.94543535e-01 1.01294361e-01 -8.83477628e-01 1.28733778e+00
2.20615551e-01 7.99966753e-02 3.22121799e-01 8.61083686e-01
5.36865473e-01 7.42361963e-01 -2.11618453e-01 -5.04881799e-01
1.28834605e+00 -9.62511361e-01 -4.76582617e-01 -5.77411577e-02
6.22614682e-01 -5.81399918e-01 7.78959334e-01 1.89481005e-01
-9.76799607e-01 -3.67474824e-01 -8.16851139e-01 2.58569956e-01
-6.01308346e-01 4.23072129e-01 6.11465454e-01 5.71914852e-01
-1.19096828e+00 1.20545641e-01 -2.93690503e-01 -5.69310963e-01
3.39875489e-01 5.58911681e-01 -4.42009479e-01 -1.88499033e-01
-7.89604664e-01 5.58429301e-01 7.86381006e-01 -1.68963492e-01
-7.97289848e-01 -5.03816605e-01 -1.05070174e+00 7.27865770e-02
7.97073483e-01 -7.57518530e-01 7.86112368e-01 -1.21728671e+00
-7.55409002e-01 9.76704240e-01 -1.67877689e-01 -1.69369116e-01
-3.36379632e-02 7.84690753e-02 -3.28406453e-01 6.10124707e-01
3.50930274e-01 6.72129214e-01 1.16928351e+00 -1.96374905e+00
-1.03406620e+00 -2.99399316e-01 1.61328271e-01 7.42292821e-01
-5.58760464e-01 1.55247912e-01 -9.45452690e-01 -8.32483888e-01
5.04912555e-01 -1.04424930e+00 -6.20611906e-01 -1.88930988e-01
-5.82142591e-01 -4.25764509e-02 7.32374847e-01 -3.33167791e-01
1.47623527e+00 -2.02509212e+00 4.68011945e-01 7.90508151e-01
3.81006807e-01 -4.22308445e-01 -4.25807275e-02 5.91990113e-01
-2.70679146e-01 2.41726935e-01 -4.00945157e-01 -2.89298445e-01
-3.94813925e-01 4.75146234e-01 -3.76860619e-01 3.92274708e-01
2.04258770e-01 6.37119949e-01 -9.51042831e-01 -5.99275887e-01
1.52667984e-01 6.22593015e-02 -5.61735153e-01 -1.35434061e-01
-2.17367187e-02 2.00843841e-01 -7.35334277e-01 7.68572152e-01
3.53281975e-01 -3.50938261e-01 2.32845142e-01 -9.83742326e-02
9.38611701e-02 -2.92507052e-01 -1.39015102e+00 1.64114928e+00
-2.29932770e-01 2.89418191e-01 -2.34741986e-01 -1.38313866e+00
9.74669456e-01 4.91746254e-02 7.90741682e-01 -3.08900550e-02
-1.13755010e-01 8.71008933e-02 -4.56511647e-01 -4.13367420e-01
9.60097015e-01 6.34983033e-02 -5.70044875e-01 4.58983272e-01
-2.01777935e-01 -1.01745710e-01 -2.31699180e-02 6.58599436e-01
6.25130892e-01 -5.45488894e-01 5.49130499e-01 -6.34686410e-01
7.57881463e-01 3.62605780e-01 5.46353877e-01 9.24148083e-01
-1.54624470e-02 7.75455177e-01 4.33484882e-01 -5.68500400e-01
-9.44708169e-01 -7.14576125e-01 3.48242782e-02 8.49927425e-01
5.04201233e-01 -6.34772778e-01 -5.75023174e-01 -7.47687578e-01
2.42749199e-01 3.60413700e-01 -6.34323776e-01 -2.47529536e-01
-4.38227683e-01 -6.42671049e-01 -7.79085681e-02 4.92955238e-01
1.44381166e-01 -6.84779465e-01 -5.98999143e-01 1.38386786e-01
-2.67662108e-01 -8.57479513e-01 -7.47748792e-01 2.95514435e-01
-6.92754745e-01 -9.32162702e-01 -5.70128441e-01 -1.12009501e+00
1.44884932e+00 7.05720067e-01 1.08480000e+00 1.07446555e-02
-3.56411040e-01 8.98446143e-01 -6.13871455e-01 -4.06338811e-01
-4.50370684e-02 -1.60841227e-01 4.77953166e-01 2.75116175e-01
2.46911556e-01 5.41995205e-02 -4.06489879e-01 1.67234138e-01
-1.02173138e+00 -8.30095112e-02 4.68747437e-01 9.62095678e-01
1.19528091e+00 7.52065957e-01 5.17399132e-01 -9.61494505e-01
7.92886674e-01 -5.14755070e-01 -4.15331364e-01 5.43280602e-01
-5.86698949e-01 2.16862738e-01 5.84930658e-01 -4.23457056e-01
-6.79746568e-01 6.34398341e-01 6.95839047e-01 -7.00572610e-01
-7.33799785e-02 9.41068530e-01 -2.65218228e-01 -9.88335232e-05
4.79124665e-01 5.75533688e-01 -2.50380635e-01 -4.62795854e-01
7.18183517e-01 6.98595643e-01 6.65579140e-01 -6.82977200e-01
7.36008942e-01 5.83742440e-01 -1.10020593e-01 -8.55068862e-01
-7.11331010e-01 -1.05859804e+00 -7.59339035e-01 -1.33426830e-01
6.78160012e-01 -1.06733465e+00 -3.60307515e-01 -5.82581721e-02
-8.73506606e-01 3.36759299e-01 -4.46450472e-01 2.48531997e-01
-7.21672416e-01 3.24371785e-01 -8.65944102e-03 -8.45660985e-01
-1.90985471e-01 -1.27035141e+00 1.20601451e+00 1.29682317e-01
3.05086654e-02 -9.17220414e-01 -2.18965843e-01 3.83812040e-01
-9.04872566e-02 1.75913900e-01 1.17789233e+00 -7.13041306e-01
-7.15970218e-01 -3.07749569e-01 1.09771334e-01 2.77320623e-01
3.95166248e-01 -2.63034105e-01 -6.45735800e-01 -5.94297528e-01
-5.62010743e-02 -1.56544596e-01 1.09814870e+00 5.60900629e-01
1.81182599e+00 -6.18177414e-01 -7.84791172e-01 3.41365486e-01
1.85954309e+00 2.73236275e-01 3.26973170e-01 3.30227077e-01
7.03364015e-01 6.58765912e-01 8.52826834e-01 5.85653484e-01
5.13990402e-01 7.77956665e-01 3.38371783e-01 -7.82071725e-02
4.39596146e-01 -5.85252680e-02 1.58483610e-01 9.04733002e-01
2.63699710e-01 -2.93846726e-01 -1.08293557e+00 9.33615625e-01
-2.05866289e+00 -1.00197554e+00 1.97848827e-01 2.28064156e+00
5.26972651e-01 -9.71852913e-02 2.59685755e-01 2.38366961e-01
8.51737678e-01 -7.51819611e-02 -4.43219721e-01 -3.99884015e-01
-1.46736503e-01 -3.60149294e-01 6.91368341e-01 3.75786155e-01
-1.30983686e+00 7.53353834e-01 6.47590256e+00 1.04245377e+00
-9.10281360e-01 -3.87090683e-01 7.05994427e-01 -1.13124698e-02
-5.44924319e-01 7.70268962e-02 -8.20581436e-01 2.50119627e-01
5.70557475e-01 -7.56816864e-01 2.59468079e-01 9.29418981e-01
1.12343349e-01 -2.80285954e-01 -1.11933196e+00 1.49597371e+00
6.10002756e-01 -1.72342396e+00 5.23615956e-01 8.44216794e-02
1.16975439e+00 -4.77321416e-01 3.42603683e-01 -1.60740018e-01
1.84195682e-01 -8.93869698e-01 6.85487270e-01 4.02398467e-01
1.01145041e+00 -1.14996386e+00 4.55401152e-01 4.01488990e-01
-1.47591329e+00 -5.85854650e-01 -6.22066796e-01 3.73238951e-01
-1.99728206e-01 3.37759852e-01 -7.88239002e-01 9.38986123e-01
7.38314033e-01 6.78131223e-01 -4.45657611e-01 1.01874208e+00
2.81222284e-01 2.24866167e-01 -1.42354786e-01 1.13480754e-01
3.32218289e-01 -2.27621555e-01 5.01507342e-01 1.21350908e+00
2.32486039e-01 4.89657551e-01 7.28904188e-01 3.61018866e-01
-1.81330502e-01 3.73883575e-01 -8.34188044e-01 6.59670308e-02
6.30984247e-01 1.25181282e+00 -9.84921694e-01 -6.90681756e-01
-3.33184183e-01 1.08452177e+00 8.13912302e-02 3.76186758e-01
-2.89877832e-01 -3.25647712e-01 6.23357534e-01 -1.11204572e-01
3.05872768e-01 8.07652902e-03 -2.08251357e-01 -1.08272123e+00
7.01309741e-02 -7.90893495e-01 8.52484763e-01 -7.18662202e-01
-1.19858313e+00 4.01892543e-01 3.85569721e-01 -1.53366077e+00
-9.70957503e-02 -4.31948543e-01 -3.26168865e-01 7.77824104e-01
-1.43648255e+00 -9.44898069e-01 -2.17531160e-01 7.23230720e-01
7.76936114e-01 -5.22136688e-01 9.41910207e-01 -9.04283226e-02
-3.98384124e-01 3.34103137e-01 6.76815152e-01 -1.93773314e-01
3.71912867e-01 -1.53209817e+00 -8.10260624e-02 7.78314054e-01
1.39828727e-01 7.38306046e-01 4.85556304e-01 -4.19659019e-01
-1.30565429e+00 -1.30006349e+00 1.05953085e+00 -7.57507265e-01
4.47186351e-01 -3.56592119e-01 -6.03586137e-01 4.79922026e-01
-1.15288191e-01 -1.93343177e-01 9.84605253e-01 -1.55913457e-01
-1.47348061e-01 -3.04570228e-01 -9.88706708e-01 5.59843779e-01
6.68435752e-01 -5.36229849e-01 -6.21120512e-01 5.60125232e-01
7.46526897e-01 -1.54478461e-01 -5.83992541e-01 -1.78000852e-01
1.70197412e-01 -2.57078260e-01 1.17440975e+00 -7.62203753e-01
4.52125639e-01 -4.91795212e-01 -2.86857635e-01 -1.49015284e+00
-4.79045004e-01 -4.25710231e-01 1.07181706e-01 1.04374766e+00
2.49740109e-01 -1.97648808e-01 8.70780110e-01 7.08614528e-01
-5.85499045e-04 -7.48124599e-01 -7.06257999e-01 -8.72068882e-01
-8.96867886e-02 -1.93494812e-01 5.85960090e-01 1.18054402e+00
3.35185438e-01 2.73996592e-01 -5.55363774e-01 3.51491004e-01
6.38186276e-01 6.01142764e-01 6.16891205e-01 -1.25427651e+00
1.68720663e-01 -2.34354928e-01 -4.40232247e-01 -1.02513647e+00
2.83914864e-01 -1.15821576e+00 1.33724466e-01 -1.61411011e+00
6.71221852e-01 -6.25593424e-01 -4.67787176e-01 6.04169488e-01
-2.52312452e-01 -2.22317755e-01 2.42062107e-01 6.46150470e-01
-7.80256271e-01 3.83547366e-01 8.72330844e-01 -4.23921913e-01
-5.77332258e-01 -2.09725171e-01 -1.20794213e+00 3.98135275e-01
8.34998369e-01 -5.69438279e-01 -6.74059272e-01 -3.76495659e-01
2.71040164e-02 1.10444255e-01 9.88001898e-02 -6.40249312e-01
3.62053603e-01 -5.80505967e-01 2.33124241e-01 -7.48375058e-01
2.42777035e-01 -1.17241120e+00 1.80264682e-01 2.58529186e-01
-6.26819015e-01 1.74114063e-01 -1.37527371e-02 7.41120875e-01
-3.85345876e-01 -1.91472813e-01 4.10805017e-01 -2.75500953e-01
-1.09102106e+00 3.97120982e-01 -2.28998601e-01 -4.82434332e-01
1.18065560e+00 -4.05683815e-01 -1.14956222e-01 -4.88071948e-01
-5.82332790e-01 6.64150894e-01 5.04669428e-01 5.15849233e-01
1.02489603e+00 -1.58789277e+00 -7.18673408e-01 3.47839668e-02
7.57224739e-01 1.96854934e-01 5.57601675e-02 2.47183904e-01
-3.33323598e-01 6.52471721e-01 -6.28735200e-02 -5.30452609e-01
-1.27793050e+00 1.03030646e+00 1.64679121e-02 -4.65531647e-02
-7.02711344e-01 8.88577044e-01 4.58986610e-01 -8.82086083e-02
1.89144731e-01 -5.70066720e-02 -6.14249051e-01 2.08969101e-01
5.84828854e-01 3.84936258e-02 -1.25951543e-01 -1.00310004e+00
-6.74046993e-01 6.71920896e-01 -1.43440247e-01 -1.27125401e-02
1.45394671e+00 -5.27561486e-01 -2.58152455e-01 3.43366355e-01
1.33309829e+00 -5.77760600e-02 -7.97746778e-01 -3.84622693e-01
1.81792378e-01 -7.54405379e-01 -3.68580297e-02 -3.34532768e-01
-9.87515926e-01 3.54850143e-01 4.34808195e-01 1.88743711e-01
1.60067570e+00 4.18442905e-01 1.08808950e-01 5.08240998e-01
4.86952037e-01 -1.37256253e+00 3.13773900e-01 3.20050493e-02
1.05876029e+00 -1.16999578e+00 9.11759883e-02 -5.26826680e-01
-1.13902378e+00 1.05407178e+00 3.41659069e-01 1.03655010e-01
5.35133898e-01 -2.09893703e-01 1.13259628e-01 -4.47300225e-01
-8.06825578e-01 -3.00443918e-01 2.82634914e-01 5.77421010e-01
-2.81539917e-01 2.05352992e-01 -4.43019032e-01 5.63758612e-01
-9.63760242e-02 -4.65591103e-01 8.69393170e-01 9.94658947e-01
-6.42541766e-01 -8.95924330e-01 -7.61673272e-01 7.36920834e-01
-1.81992218e-01 1.54453531e-01 -4.74843532e-01 2.90633738e-01
1.64151192e-01 9.71111298e-01 8.79956186e-02 -3.81139994e-01
3.53310943e-01 -1.37225538e-02 -9.95833278e-02 -8.86780381e-01
-1.94236025e-01 2.26241276e-01 -3.63032669e-01 -4.26726967e-01
-7.00235188e-01 -6.72352135e-01 -1.20588350e+00 -6.66926652e-02
-5.47380805e-01 3.53003800e-01 5.25132239e-01 8.14182222e-01
1.58549860e-01 1.30831301e-01 1.08777106e+00 -6.95381463e-01
-2.37861231e-01 -9.61277783e-02 -7.69927979e-01 5.87560952e-01
4.98391509e-01 -4.55034792e-01 -1.22826219e-01 5.51450789e-01] | [9.334650993347168, 3.0593950748443604] |
c80efe67-c7f0-4de2-b3f9-c930b66d0037 | from-graph-generation-to-graph-classification | 2302.07989 | null | https://arxiv.org/abs/2302.07989v1 | https://arxiv.org/pdf/2302.07989v1.pdf | From Graph Generation to Graph Classification | This note describes a new approach to classifying graphs that leverages graph generative models (GGM). Assuming a GGM that defines a joint probability distribution over graphs and their class labels, I derive classification formulas for the probability of a class label given a graph. A new conditional ELBO can be used to train a generative graph auto-encoder model for discrimination. While leveraging generative models for classification has been well explored for non-relational i.i.d. data, to our knowledge it is a novel approach to graph classification. | ['Oliver Schulte'] | 2023-02-15 | null | null | null | null | ['graph-classification'] | ['graphs'] | [ 2.22495049e-01 6.80222392e-01 -4.58070576e-01 -7.23214090e-01
-6.40975118e-01 -5.65812647e-01 8.51531029e-01 2.09850773e-01
3.79431069e-01 5.69629848e-01 -9.78316739e-02 -7.73047090e-01
-3.21414083e-01 -1.49652243e+00 -4.51529771e-01 -6.98110044e-01
-4.32078868e-01 1.05851257e+00 -9.57077220e-02 3.06418031e-01
-1.04989866e-02 5.62614322e-01 -1.08053923e+00 1.23449020e-01
5.39243221e-01 8.39618564e-01 -3.32066774e-01 9.66696799e-01
-1.66506991e-01 1.08344102e+00 -8.90931368e-01 -7.03285575e-01
-8.92756060e-02 -5.28757155e-01 -1.22423136e+00 4.03363138e-01
4.74874943e-01 -4.16437127e-02 -6.66372716e-01 1.04141164e+00
1.34595796e-01 1.09405614e-01 1.69517446e+00 -1.88553309e+00
-1.07072902e+00 5.75292766e-01 -8.62857774e-02 1.68466836e-01
5.81352293e-01 -4.62116808e-01 1.45136857e+00 -9.84393656e-02
7.10785210e-01 1.20477021e+00 7.20566750e-01 3.01481962e-01
-1.68855870e+00 -5.75651109e-01 -8.52025673e-03 -9.70495269e-02
-1.44987249e+00 3.36355776e-01 7.75654078e-01 -6.78016305e-01
1.04594684e+00 6.15411364e-02 9.65040982e-01 1.27260959e+00
7.35812426e-01 8.11044097e-01 1.12741828e+00 -4.68777657e-01
1.68845192e-01 5.30130928e-03 4.22004461e-01 1.14454305e+00
3.17746282e-01 2.28248555e-02 -2.99617529e-01 -1.07552066e-01
6.66054606e-01 3.02112009e-02 1.88730806e-01 -5.45297205e-01
-3.65913451e-01 1.34396863e+00 5.50667763e-01 2.10709035e-01
6.52216300e-02 4.67342198e-01 2.15363830e-01 5.29820383e-01
6.74795747e-01 2.21703559e-01 1.04599454e-01 2.38799661e-01
-7.57783353e-01 -1.03430912e-01 1.12048817e+00 1.20964706e+00
9.82143044e-01 1.47026300e-01 -2.82764375e-01 4.77814376e-01
6.42025709e-01 3.28401417e-01 -4.95523959e-02 -3.61932069e-01
-6.02731965e-02 8.52903724e-01 -7.93238223e-01 -8.26751292e-01
-2.43615225e-01 -6.90826118e-01 -7.00560570e-01 1.81281909e-01
2.62477607e-01 1.76754400e-01 -1.41307950e+00 1.63535249e+00
-1.27416089e-01 -1.07457172e-02 1.31006315e-01 2.94380963e-01
1.11815751e+00 3.68089586e-01 3.73350650e-01 3.17966729e-01
7.79463828e-01 -5.00869513e-01 -4.88081664e-01 -2.69665979e-02
7.13835835e-01 -4.78661627e-01 4.48939472e-01 1.77691981e-01
-7.01832771e-01 -2.35922083e-01 -1.00345659e+00 -7.82261714e-02
-7.85929382e-01 -2.73434490e-01 1.45482516e+00 1.04094362e+00
-1.29341376e+00 5.38369715e-01 -8.06787491e-01 -6.54601097e-01
6.81386948e-01 4.36638713e-01 -5.12312353e-01 -2.51598179e-01
-9.02474999e-01 7.87051857e-01 4.38390404e-01 -4.51490551e-01
-1.22067571e+00 -9.21343192e-02 -1.02124274e+00 3.16409878e-02
7.53996000e-02 -7.02176690e-01 9.37160790e-01 -8.15559745e-01
-1.33806539e+00 1.19266486e+00 2.95833468e-01 -4.47226912e-01
-1.25853494e-01 3.32098752e-01 -5.68754554e-01 8.27496573e-02
3.20894659e-01 4.82938498e-01 8.35448146e-01 -1.19097114e+00
-1.54631793e-01 -2.46293128e-01 2.29178190e-01 -7.12637678e-02
-1.25859422e-03 -2.68379688e-01 1.71026234e-02 -5.69842994e-01
2.37541184e-01 -1.00196123e+00 4.86765467e-02 -5.97614884e-01
-1.00289118e+00 -5.42982638e-01 6.83803797e-01 -3.51194561e-01
1.10560739e+00 -1.86259460e+00 7.57563338e-02 6.72870040e-01
6.93760037e-01 -1.71575889e-01 -4.90007140e-02 7.61839449e-01
-2.91998893e-01 4.16165143e-01 -5.05593009e-02 -2.22736433e-01
3.16461474e-01 5.27318776e-01 -2.22024187e-01 5.58854461e-01
3.74151736e-01 1.11120737e+00 -7.83504784e-01 -4.79422987e-01
6.48613498e-02 6.90649867e-01 -4.95767474e-01 3.79052848e-01
-1.14376724e-01 2.99259014e-02 -3.16721797e-01 7.33389974e-01
5.56179821e-01 -8.00003529e-01 6.22156680e-01 7.28282779e-02
5.88922679e-01 3.86140674e-01 -7.77187884e-01 1.11455643e+00
-2.01629534e-01 7.23256528e-01 -5.27343154e-01 -1.29583371e+00
1.19843233e+00 1.20860018e-01 2.37182379e-01 -1.99878559e-01
2.49111965e-01 5.47832297e-03 7.35192979e-03 1.75398305e-01
8.97596125e-03 -2.60243148e-01 -3.45117748e-01 6.12259746e-01
9.50016141e-01 -1.81959733e-01 2.29323760e-01 5.64686716e-01
1.45672870e+00 1.66387722e-01 4.15115505e-01 -2.94075698e-01
5.63768717e-03 -8.79625529e-02 -1.14906669e-01 1.03995693e+00
1.83187246e-01 3.18079531e-01 7.86032438e-01 -4.28437114e-01
-6.05634093e-01 -1.69068170e+00 -2.67812043e-01 9.67039466e-01
-1.66593954e-01 -9.00511444e-01 -4.36953127e-01 -9.42215681e-01
1.66532293e-01 6.05367124e-01 -9.39011157e-01 -5.83673120e-01
1.68982267e-01 -8.96079957e-01 4.19579327e-01 6.63219333e-01
6.04586042e-02 -7.29440391e-01 9.42355916e-02 1.22247506e-02
4.84713495e-01 -9.22358572e-01 -1.44768655e-01 8.63164842e-01
-7.26491094e-01 -1.32231545e+00 -1.39357775e-01 -1.01671374e+00
8.62337351e-01 -1.06160017e-02 1.74604666e+00 -3.87409888e-02
-4.00553077e-01 8.89061213e-01 -3.69521409e-01 -2.63352573e-01
-7.98498988e-01 3.22923422e-01 -4.26315457e-01 -2.34047979e-01
5.44880569e-01 -7.05285847e-01 -9.80057940e-02 4.18233089e-02
-8.18692386e-01 -2.02708393e-02 5.76047242e-01 7.02996790e-01
6.22818410e-01 4.02032703e-01 1.75042465e-01 -1.30868065e+00
7.32413471e-01 -8.04107666e-01 -3.85841399e-01 4.84541178e-01
-1.10333014e+00 3.63654524e-01 2.11878330e-01 -3.92012298e-01
-6.08284295e-01 -7.95925483e-02 -4.58346233e-02 -1.53516993e-01
-2.78050929e-01 8.10001075e-01 -1.11221999e-01 -2.44788840e-01
3.40385169e-01 -1.25960231e-01 -1.30830646e-01 -2.51689553e-01
4.33692992e-01 4.01309580e-01 1.44628108e-01 -4.63088632e-01
7.02517211e-01 2.17101768e-01 7.69587815e-01 -5.56180120e-01
-9.69203711e-01 -2.13214502e-01 -7.20005870e-01 -2.28023663e-01
1.34041607e+00 -7.19777584e-01 -6.41259909e-01 1.37889028e-01
-7.75787055e-01 -4.01801288e-01 -1.29911721e-01 4.78368342e-01
-6.42064095e-01 1.25036553e-01 -6.95936561e-01 -8.47054601e-01
4.11866903e-02 -7.95163929e-01 1.06530499e+00 1.01753481e-01
-3.42051983e-01 -1.70386672e+00 2.70793885e-01 -6.56243935e-02
1.29828826e-01 4.66878563e-01 1.25495338e+00 -9.75381255e-01
-7.77371109e-01 -5.70849121e-01 -2.53153741e-01 1.85590759e-01
3.78830619e-02 2.37011448e-01 -7.16206849e-01 -3.41964304e-01
-6.70033991e-01 -3.75362456e-01 8.39589834e-01 4.20339942e-01
9.36035872e-01 -1.24682277e-01 -8.09279084e-01 6.70182168e-01
1.50687897e+00 3.09885949e-01 5.82476497e-01 -1.08002692e-01
8.13099980e-01 9.66019407e-02 -5.33300899e-02 -6.59330636e-02
4.64478344e-01 4.25106436e-01 3.33855450e-01 -1.17337011e-01
-3.96841764e-01 -6.54381871e-01 1.98579296e-01 6.30026639e-01
1.19669594e-01 -8.36273491e-01 -1.09777880e+00 1.75287157e-01
-1.39459503e+00 -7.74431705e-01 -3.22504103e-01 1.79928327e+00
4.00042742e-01 2.54966855e-01 1.38350919e-01 1.00665957e-01
7.48276234e-01 1.45596519e-01 -1.85540140e-01 -5.12614787e-01
-1.29315183e-01 8.87292206e-01 4.38352227e-01 5.96729457e-01
-1.25784159e+00 8.69859338e-01 8.03930378e+00 6.40875041e-01
-8.30405831e-01 4.51163538e-02 6.02029622e-01 7.24910438e-01
-4.70580012e-01 4.33246940e-01 -7.57458210e-01 1.61761209e-01
1.17211497e+00 -1.59078881e-01 3.24304104e-01 8.37671280e-01
-1.06977665e+00 -6.70571718e-03 -1.37561381e+00 7.84841120e-01
2.87199795e-01 -1.02763605e+00 1.76277027e-01 5.36522746e-01
6.87341392e-01 -9.92939714e-03 1.45833388e-01 5.56494296e-01
1.17776799e+00 -1.41911376e+00 3.23382646e-01 7.83741295e-01
8.23540986e-01 -7.43580580e-01 7.09268153e-01 -3.66776176e-02
-1.12061763e+00 3.12624514e-01 -1.96290389e-01 -8.11207294e-03
-3.32385957e-01 6.45518303e-01 -1.25297809e+00 7.15573251e-01
3.86794925e-01 7.93882728e-01 -9.29684162e-01 6.67060018e-01
-5.89071035e-01 9.19094324e-01 -3.73884700e-02 -8.15600306e-02
-4.41338085e-02 -2.30869189e-01 3.65098059e-01 1.21424806e+00
2.08896846e-01 -3.78230989e-01 6.67347610e-01 9.20245528e-01
-2.76481081e-02 -4.12336625e-02 -9.60189402e-01 -6.85973346e-01
6.19411916e-02 1.12054896e+00 -1.18538940e+00 -3.81726921e-01
-3.31975400e-01 9.75164771e-01 7.31031775e-01 4.36100811e-01
-6.45030439e-01 -2.60899574e-01 1.55905202e-01 3.19112465e-02
1.33259833e-01 -3.65189940e-01 3.91985066e-02 -1.10679352e+00
-7.20404685e-01 -2.71692872e-01 9.37235832e-01 -8.59039843e-01
-1.81928670e+00 8.06469083e-01 2.24018767e-01 -8.67465436e-01
-5.23418903e-01 -9.13204908e-01 -5.55751145e-01 8.10933173e-01
-1.03347933e+00 -1.44529486e+00 -2.92735726e-01 4.05532449e-01
-3.57658535e-01 -3.61026794e-01 1.39656866e+00 9.36819240e-03
-3.70785326e-01 4.61421788e-01 -1.26866281e-01 4.71042067e-01
3.76248151e-01 -1.97521245e+00 4.50677574e-01 5.99211931e-01
8.28280926e-01 5.11340916e-01 3.99961501e-01 -8.59023809e-01
-1.16103911e+00 -1.02739894e+00 8.02943587e-01 -6.50887907e-01
5.94862700e-01 -5.99549174e-01 -6.84834540e-01 1.25813353e+00
2.30128631e-01 1.74003944e-01 1.24133325e+00 6.16135359e-01
-6.29394412e-01 2.55164325e-01 -9.32420671e-01 1.82119593e-01
1.12997139e+00 -8.68049622e-01 -2.16778427e-01 6.52264357e-01
2.04142004e-01 -2.31956616e-01 -1.28034520e+00 2.64894158e-01
2.69251436e-01 -8.70766163e-01 7.57472873e-01 -9.10791636e-01
1.77015245e-01 -2.06587419e-01 -3.91173244e-01 -1.43066680e+00
-6.90888584e-01 -4.16770369e-01 -1.77400917e-01 1.12560534e+00
3.37523431e-01 -7.72087216e-01 8.83909762e-01 2.32171625e-01
-3.65707092e-02 -6.83812201e-01 -7.55817890e-01 -9.39443111e-01
6.01081587e-02 -4.68664140e-01 5.71377516e-01 9.69836175e-01
-1.87718436e-01 8.14822257e-01 6.93808775e-03 1.96515843e-01
6.91853285e-01 4.71165717e-01 7.64118552e-01 -1.58311999e+00
-4.72250700e-01 -5.22405982e-01 -1.32358027e+00 -6.83985114e-01
6.15430772e-01 -1.88422728e+00 -4.27182674e-01 -2.06215429e+00
4.51264560e-01 -3.83741260e-01 -1.19971424e-01 5.76906025e-01
1.59246475e-01 4.79932904e-01 -1.89250812e-01 -5.28750382e-02
-6.35834455e-01 2.40338981e-01 9.80135500e-01 -2.65186548e-01
3.90558302e-01 1.29173487e-01 -4.59309220e-01 1.96229428e-01
7.52863169e-01 -5.92069864e-01 -7.00940847e-01 1.49603933e-01
3.95944327e-01 -2.56225606e-03 5.77908874e-01 -9.34383273e-01
-1.39022291e-01 6.14412129e-02 4.32781786e-01 -5.50536811e-01
3.78253055e-03 -5.13184130e-01 6.34136379e-01 1.62667423e-01
-3.20479780e-01 3.09290383e-02 -2.97746927e-01 7.81460524e-01
-3.26082379e-01 -2.78473914e-01 5.97934484e-01 -8.48544240e-02
-3.88211280e-01 4.57621545e-01 -4.79570448e-01 2.01888621e-01
9.24838960e-01 -2.06805900e-01 -2.38191545e-01 -7.98545897e-01
-1.27738786e+00 -2.35210449e-01 3.10140163e-01 1.08805649e-01
4.64972675e-01 -1.55459094e+00 -4.19663042e-01 2.91172385e-01
1.74170360e-01 -2.55686849e-01 -1.96893856e-01 4.36078340e-01
-2.69322813e-01 2.60440946e-01 -1.38706893e-01 -6.14618659e-01
-1.04200995e+00 6.19194031e-01 3.83511186e-01 -5.15103877e-01
-4.78383064e-01 9.23902452e-01 2.46961281e-01 -3.88128817e-01
-1.75883248e-01 -1.01208344e-01 1.02227174e-01 -3.90202627e-02
-2.53143162e-01 1.58773944e-01 2.60116300e-04 -7.03224421e-01
-2.90901333e-01 2.40573078e-01 9.15051177e-02 2.00273216e-01
1.19343972e+00 2.97653675e-01 -2.66077787e-01 7.36559272e-01
1.58899939e+00 1.19764693e-02 -7.77810752e-01 1.25007033e-01
-5.39435744e-02 -2.77882308e-01 5.16285971e-02 -9.35420632e-01
-1.02427936e+00 8.25630426e-01 3.13224375e-01 9.00039136e-01
7.25857794e-01 5.85580230e-01 1.57302752e-01 5.26271202e-02
2.86295921e-01 -5.91527402e-01 3.12343985e-01 4.14831281e-01
5.64724207e-01 -1.10972345e+00 2.04404369e-01 -8.10220420e-01
-4.11158532e-01 1.17316926e+00 2.48768628e-01 -5.02405524e-01
1.09435785e+00 3.09487313e-01 -2.88782775e-01 -7.41763592e-01
-5.62556386e-01 -5.58789194e-01 7.23431110e-01 1.05283618e+00
7.71890759e-01 6.81208432e-01 6.50793016e-02 1.63371235e-01
-4.39652860e-01 -3.89221460e-01 3.47170234e-01 6.85847998e-01
7.26591097e-03 -1.41169477e+00 1.69407070e-01 8.95898640e-01
-3.15888613e-01 -1.20900668e-01 -6.64437234e-01 8.71833444e-01
-7.02015236e-02 8.68843913e-01 2.20297545e-01 -6.39490664e-01
-1.67045802e-01 4.51027364e-01 8.57501388e-01 -1.05443048e+00
-1.06395960e-01 -2.65863329e-01 2.75028229e-01 -1.12625137e-01
-4.74941492e-01 -5.81237674e-01 -1.14566851e+00 -1.98881119e-01
-4.87547278e-01 1.96715295e-01 4.64302868e-01 6.06715798e-01
1.86801210e-01 6.21874869e-01 3.95624578e-01 -2.39957899e-01
-5.52033298e-02 -9.62826908e-01 -9.19017971e-01 5.03362477e-01
-1.12563521e-01 -1.03311980e+00 -4.99652505e-01 -2.07163826e-01] | [6.9996209144592285, 6.184803485870361] |
a418ed31-dbe9-4129-9285-4d25433b9173 | skrl-modular-and-flexible-library-for | 2202.03825 | null | https://arxiv.org/abs/2202.03825v2 | https://arxiv.org/pdf/2202.03825v2.pdf | skrl: Modular and Flexible Library for Reinforcement Learning | skrl is an open-source modular library for reinforcement learning written in Python and designed with a focus on readability, simplicity, and transparency of algorithm implementations. In addition to supporting environments that use the traditional interfaces from OpenAI Gym and DeepMind, it provides the facility to load, configure, and operate NVIDIA Isaac Gym and NVIDIA Omniverse Isaac Gym environments. Furthermore, it enables the simultaneous training of several agents with customizable scopes (subsets of environments among all available ones), which may or may not share resources, in the same run. The library's documentation can be found at https://skrl.readthedocs.io and its source code is available on GitHub at https://github.com/Toni-SM/skrl. | ['Simon Bøgh', 'Dimitris Chrysostomou', 'Nestor Arana-Arexolaleiba', 'Antonio Serrano-Muñoz'] | 2022-02-08 | null | null | null | null | ['omniverse-isaac-gym', 'isaac-gym-preview'] | ['robots', 'robots'] | [-9.02203977e-01 -1.65497273e-01 -2.12080687e-01 -9.43485126e-02
-2.74228215e-01 -8.04167688e-01 2.30462179e-01 -6.50789291e-02
-5.43137670e-01 8.10284197e-01 -6.95126802e-02 -5.26979148e-01
-6.73033893e-02 -7.47856915e-01 -5.20473838e-01 -6.57477677e-01
-1.79744661e-01 3.73122007e-01 1.89645439e-01 -3.22119504e-01
4.49347757e-02 2.91882694e-01 -1.83171272e+00 -7.72730485e-02
8.39188278e-01 4.16324914e-01 6.64885104e-01 1.00929284e+00
5.06520569e-01 9.45854604e-01 -6.37795985e-01 3.88406068e-01
3.38587046e-01 -5.70686068e-03 -6.91196024e-01 -4.93788421e-01
1.69455111e-01 -7.17898369e-01 -5.22632241e-01 7.02239513e-01
7.72588849e-01 3.48091871e-01 -3.85225207e-01 -1.47644436e+00
-2.08570629e-01 2.90082574e-01 4.31324840e-02 4.55145389e-01
6.01435184e-01 7.23832607e-01 7.15011358e-01 -2.68767983e-01
4.41385835e-01 7.73955762e-01 2.64455050e-01 7.60632932e-01
-8.86198938e-01 -8.61619413e-01 3.89198400e-02 1.80823937e-01
-1.27428436e+00 -3.87294888e-01 1.88675910e-01 -2.31877521e-01
1.03163898e+00 3.80091786e-01 1.06543529e+00 1.37005961e+00
9.34386402e-02 8.16283822e-01 1.11970210e+00 -1.60790965e-01
6.57495141e-01 -4.74671125e-02 4.13318090e-02 8.34315002e-01
3.18597734e-01 2.63622731e-01 -5.22259831e-01 -2.03695744e-01
1.02081299e+00 -2.38761947e-01 -2.77646124e-01 -4.81383204e-01
-9.76440549e-01 6.93453610e-01 2.55955249e-01 -1.22295190e-02
-2.37127319e-01 4.30371791e-01 4.92247194e-01 2.36606196e-01
-1.34283975e-01 3.59264225e-01 -6.13241136e-01 -8.91591430e-01
3.75567414e-02 6.26585543e-01 9.30190682e-01 1.02220917e+00
6.84939504e-01 3.13512266e-01 3.76511127e-01 4.97280151e-01
2.55371779e-01 2.79924959e-01 6.80718124e-01 -1.46528649e+00
2.01635249e-02 4.94573921e-01 3.08396906e-01 -3.28723073e-01
-7.83653140e-01 -3.36088121e-01 -1.82869341e-02 9.02228713e-01
3.62040490e-01 -5.85090935e-01 -4.03841376e-01 1.46753979e+00
7.84221590e-01 6.95659444e-02 2.16823649e-02 1.06090331e+00
1.13070881e+00 1.72686577e-01 1.66667312e-01 5.03990650e-01
1.18861520e+00 -1.00116503e+00 -4.29888248e-01 -3.60874593e-01
7.41026044e-01 -5.55530429e-01 1.46549630e+00 3.14214677e-01
-1.03041732e+00 -3.84775013e-01 -1.15652204e+00 -1.26754448e-01
-5.80658853e-01 2.88545117e-02 1.14093292e+00 6.52327657e-01
-1.33015764e+00 6.04608238e-01 -1.41271877e+00 -5.24156332e-01
8.18798393e-02 4.41653699e-01 -1.62640661e-01 2.41675973e-01
-8.96565199e-01 8.48039985e-01 6.30013347e-01 -3.62357408e-01
-1.06085384e+00 -6.08594954e-01 -8.81741405e-01 -1.87807932e-01
4.27232593e-01 -6.20849907e-01 1.65705192e+00 -7.41067827e-01
-1.93638802e+00 5.48158348e-01 3.78288120e-01 -1.40356973e-01
4.98390734e-01 -3.99580568e-01 -9.71981287e-02 8.57967362e-02
8.70534480e-02 5.83798766e-01 9.29669663e-03 -8.02792549e-01
-7.66685009e-01 -3.26966733e-01 6.27292752e-01 7.82847643e-01
-9.77728143e-02 -6.81596249e-02 -2.46987402e-01 -9.29864198e-02
-4.88534957e-01 -8.88711154e-01 -8.51591006e-02 -2.56913644e-03
6.44349232e-02 -1.99357092e-01 9.50358748e-01 -5.14883757e-01
8.13015640e-01 -2.26489902e+00 -4.69514308e-03 -1.74665853e-01
8.93379897e-02 2.80739427e-01 7.70230740e-02 7.43193746e-01
1.77388832e-01 -1.77496076e-01 3.59945238e-01 -3.08471657e-02
3.19832504e-01 4.01733577e-01 3.77712548e-01 4.39017832e-01
-4.68376696e-01 4.76948172e-01 -1.09821534e+00 -1.50553167e-01
7.23687589e-01 4.22697723e-01 -4.52922136e-01 1.22439072e-01
-2.08626375e-01 7.29549170e-01 -4.96088237e-01 3.99729431e-01
6.17929637e-01 -2.15295419e-01 5.79436064e-01 4.46105242e-01
-6.68353856e-01 6.68897331e-01 -1.54345703e+00 1.94326472e+00
-4.13795054e-01 3.86791706e-01 3.49820554e-01 -5.03963292e-01
4.53207433e-01 2.49322698e-01 5.36173820e-01 -6.56477273e-01
2.10029334e-01 1.19614847e-01 -3.30507345e-02 -5.39980292e-01
5.24592936e-01 6.35764956e-01 1.53130190e-02 7.07038224e-01
1.26701787e-01 -3.53890052e-03 5.41346371e-01 1.59674734e-01
1.28392982e+00 8.55043650e-01 2.44146138e-01 -3.94770503e-01
2.48514358e-02 1.92017347e-01 6.07679129e-01 6.26591563e-01
-3.59640956e-01 -1.33504316e-01 1.72281981e-01 -4.63990420e-01
-1.00486875e+00 -1.31204188e+00 -2.86606282e-01 1.43922579e+00
1.10445522e-01 -7.23769307e-01 -6.91215754e-01 -1.94875136e-01
4.88702068e-03 7.91246593e-01 -2.16051087e-01 2.41093919e-01
-2.95710057e-01 -2.75113761e-01 5.85614979e-01 4.19560462e-01
7.66539872e-01 -1.33156300e+00 -1.38966715e+00 -1.37759328e-01
-1.86842352e-01 -9.15801108e-01 -7.27489144e-02 3.52792770e-01
-5.45549035e-01 -1.19083655e+00 4.47008200e-02 -7.38362849e-01
2.39307284e-01 2.34348640e-01 9.64800954e-01 3.81757438e-01
-6.36718512e-01 7.02649117e-01 -4.59703535e-01 -3.24044675e-01
-9.32119042e-02 9.61913913e-02 1.57793805e-01 -9.52470958e-01
2.92288184e-01 -6.50286973e-01 -7.41862714e-01 1.33852847e-02
-4.78093714e-01 4.09860283e-01 -1.10705368e-01 4.38909948e-01
3.43602836e-01 -1.77707896e-01 1.85854524e-01 -3.15563083e-01
4.63472992e-01 -3.72531801e-01 -1.05870807e+00 -3.15731257e-01
-4.22634959e-01 -3.58296514e-01 6.05398417e-01 -6.14959672e-02
-6.99895442e-01 -1.15363784e-01 -2.96242595e-01 2.59481352e-02
-7.17376947e-01 1.99515477e-01 5.03520621e-03 -2.64603466e-01
7.05713511e-01 2.04733044e-01 2.91427970e-01 -3.43852967e-01
1.06306806e-01 6.91721082e-01 2.50444531e-01 -6.70109212e-01
4.68597680e-01 3.19450796e-01 -5.54148078e-01 -7.92154968e-01
-2.24187449e-01 -1.59745052e-01 -2.62531519e-01 -4.76432949e-01
6.96054459e-01 -1.13926566e+00 -1.24716496e+00 5.49481928e-01
-3.05701643e-01 -1.35019100e+00 -3.63812178e-01 5.78520119e-01
-6.14522517e-01 1.82443246e-01 -8.78690481e-01 -4.21896487e-01
-2.96719104e-01 -1.13483799e+00 6.44655108e-01 9.80193436e-01
-3.69402140e-01 -1.10129213e+00 2.06580818e-01 3.85227710e-01
3.93337458e-01 8.28000382e-02 4.04614151e-01 -4.94805366e-01
-4.26727742e-01 1.16676977e-03 3.72322947e-01 1.70457259e-01
7.69905373e-02 2.03321263e-01 -7.71321297e-01 -6.83825731e-01
-4.96028930e-01 -7.04640687e-01 -1.16566211e-01 3.26925069e-01
1.13615763e+00 -1.50395453e-01 -6.18028566e-02 6.64570391e-01
1.50073779e+00 2.95760870e-01 6.25586450e-01 1.02434349e+00
3.45412970e-01 6.69114068e-02 6.44639552e-01 9.69725668e-01
7.11951196e-01 6.76399529e-01 7.51514256e-01 1.43756215e-02
1.21585011e-01 -8.60368088e-02 3.35186571e-01 5.49442291e-01
-1.58973053e-01 1.96522802e-01 -7.85261810e-01 2.24904567e-01
-1.98587239e+00 -9.95198071e-01 -1.20062970e-01 2.38726401e+00
6.75124764e-01 -1.99955016e-01 4.67515767e-01 -3.01293433e-01
2.58808732e-01 5.78847788e-02 -6.43141687e-01 -6.39214039e-01
2.10976824e-01 4.03214276e-01 4.23484653e-01 6.11382306e-01
-9.22690511e-01 9.65376735e-01 5.64265919e+00 4.69171464e-01
-1.07036006e+00 2.75016844e-01 6.07320927e-02 -5.84324598e-01
1.24864027e-01 1.07641071e-01 -6.00541890e-01 5.47013164e-01
9.96736765e-01 -3.90843689e-01 1.06312013e+00 1.13913727e+00
6.11277223e-01 -4.96920615e-01 -6.24672771e-01 5.49373448e-01
-5.48740506e-01 -1.25918198e+00 -7.53556013e-01 1.13935515e-01
3.76165569e-01 5.59519827e-01 -1.40058860e-01 4.51049894e-01
9.43103850e-01 -7.14938581e-01 7.03851461e-01 -7.55955651e-02
3.51677775e-01 -9.58590984e-01 3.91581208e-01 4.34073061e-01
-1.01123643e+00 -2.31894314e-01 -3.08445871e-01 -7.91615844e-01
-3.74366075e-01 -2.10290700e-01 -4.34362918e-01 5.16178548e-01
1.38463223e+00 4.79832113e-01 -4.38022226e-01 1.23703814e+00
-4.36852992e-01 2.99527317e-01 -2.96385169e-01 -2.43224651e-01
1.07725628e-01 -3.31539750e-01 2.52015412e-01 7.15979874e-01
-5.24706878e-02 9.46078524e-02 6.34452760e-01 4.96033430e-01
4.27631527e-01 4.68526781e-02 -6.29411161e-01 2.13770598e-01
8.28241408e-01 1.62241590e+00 -5.15899956e-01 -1.81093752e-01
-3.94085407e-01 7.74765849e-01 4.38155621e-01 2.80224442e-01
-1.04534471e+00 -4.86134946e-01 1.11892951e+00 -5.84793538e-02
4.64922227e-02 -5.57608902e-01 -2.48100996e-01 -8.30230534e-01
-5.04618436e-02 -1.16570866e+00 4.90006536e-01 -9.00825560e-01
-5.57812929e-01 3.26939195e-01 -1.35496100e-02 -1.16385818e+00
-1.53289050e-01 -6.86092496e-01 -7.64905334e-01 6.15285337e-01
-9.83152390e-01 -8.52200866e-01 -7.86198020e-01 7.20183372e-01
4.99336839e-01 -4.66488563e-02 1.32811093e+00 3.01119626e-01
-8.67923260e-01 2.56051779e-01 3.07342798e-01 -4.82734144e-02
4.78804111e-01 -1.47853792e+00 2.89761186e-01 5.94303846e-01
-5.42041659e-01 5.81756234e-01 7.73745358e-01 -5.69834709e-01
-1.49385691e+00 -7.45626986e-01 5.39881401e-02 -1.49124339e-01
7.52896369e-01 -1.88188478e-01 -5.45805752e-01 1.21164334e+00
6.26845419e-01 -1.28945783e-01 8.58763933e-01 2.12774470e-01
8.35790634e-02 5.95346354e-02 -9.29035068e-01 9.81282592e-01
8.29334676e-01 -2.03978702e-01 1.29242122e-01 6.73747540e-01
2.58186638e-01 -1.09331763e+00 -9.37246859e-01 -3.10856495e-02
5.97462833e-01 -1.13169062e+00 7.04781234e-01 -2.10738823e-01
1.35269508e-01 -4.73014832e-01 9.64670330e-02 -1.35813332e+00
-3.37292314e-01 -5.90385973e-01 4.37283814e-02 6.57078207e-01
1.01679422e-01 -1.13167429e+00 7.73310900e-01 4.81554210e-01
-4.77170229e-01 -6.42566502e-01 -9.27715421e-01 -7.81427443e-01
-4.42071557e-02 -1.87534869e-01 5.91444314e-01 8.61597478e-01
6.74700201e-01 -2.86321752e-02 -1.65752858e-01 5.13562202e-01
5.97829819e-01 -9.39876363e-02 1.04016495e+00 -7.55257130e-01
-7.30472744e-01 -2.86856443e-01 -3.19378108e-01 -6.86714888e-01
-1.49311319e-01 -8.25824082e-01 -8.63373131e-02 -1.77696204e+00
-1.14071205e-01 -6.53392673e-01 9.11084935e-02 9.56209600e-01
1.07680708e-01 2.25169081e-02 2.17038661e-01 4.84145805e-02
-7.88088262e-01 4.00097102e-01 1.13168013e+00 5.24626076e-01
-3.11211944e-01 -5.57639115e-02 -5.44819951e-01 6.85321987e-01
1.65217042e+00 -9.26122963e-02 -4.60671782e-01 -5.70752382e-01
-2.04050601e-01 -1.20124482e-01 7.13596165e-01 -1.58586669e+00
8.02372918e-02 -3.28811467e-01 4.80941504e-01 2.22768052e-03
5.41785717e-01 -4.83338445e-01 3.16744685e-01 8.49408031e-01
-5.58341220e-02 4.99616176e-01 6.19829059e-01 -1.12123773e-01
4.49235678e-01 -2.76057124e-01 5.61146379e-01 -5.58374822e-01
-9.75578189e-01 8.19428265e-02 -9.13905203e-01 -5.58107719e-03
1.32388246e+00 -1.31081402e-01 -8.84213686e-01 -4.13380742e-01
-6.35244966e-01 6.19723320e-01 9.15511191e-01 4.68621105e-01
2.59709835e-01 -1.03966796e+00 -2.53853381e-01 9.64507163e-02
-3.12179364e-02 -5.45931943e-02 5.17867148e-01 6.94905937e-01
-1.08954620e+00 7.65841752e-02 -8.57418776e-01 -2.58714408e-01
-1.36028397e+00 7.61831924e-02 6.18619740e-01 1.27715394e-01
-9.79662895e-01 5.43820798e-01 -2.03818098e-01 -9.87488449e-01
3.60353172e-01 6.97924346e-02 -6.43866882e-02 -7.90391207e-01
7.21001446e-01 4.37760949e-01 -4.12119962e-02 -6.19629398e-02
-4.19288963e-01 -6.65939599e-02 1.29176602e-01 -6.94583505e-02
1.28078759e+00 -5.80054298e-02 5.67386560e-02 4.58751291e-01
3.81931812e-01 -1.32504925e-01 -1.47993255e+00 3.40117246e-01
-3.53273839e-01 -4.46555406e-01 -4.81964163e-02 -8.14086735e-01
-8.11529040e-01 3.55418146e-01 1.04155076e+00 -7.93131217e-02
7.80970454e-01 -1.89212933e-01 3.16788524e-01 2.10979939e-01
7.48606026e-01 -1.51745892e+00 6.37359023e-02 5.68057060e-01
6.66317403e-01 -9.05559897e-01 1.45788163e-01 -6.56705424e-02
-7.32802153e-01 1.08685410e+00 1.22605801e+00 -2.32180357e-01
3.57500076e-01 5.70454240e-01 3.64883482e-01 -2.71367192e-01
-9.08744752e-01 -2.88076818e-01 -7.14480221e-01 9.82397676e-01
5.56420743e-01 4.89053249e-01 -3.99577022e-01 2.05041945e-01
-6.25296354e-01 2.10392386e-01 8.28486085e-01 1.50718713e+00
-4.51033473e-01 -1.18042505e+00 -2.39507273e-01 1.78516135e-01
-1.34421676e-01 2.19960719e-01 1.73561677e-01 1.00323474e+00
-8.72512236e-02 8.49942744e-01 1.08946018e-01 -3.29817832e-01
1.03009120e-01 -1.64731070e-01 3.78207982e-01 -4.69296604e-01
-8.12925458e-01 -2.11583138e-01 3.75467688e-01 -8.77659976e-01
2.17613176e-01 -7.35129893e-01 -1.64230454e+00 -8.93352270e-01
2.21332207e-01 1.30574659e-01 7.01563179e-01 4.70448285e-01
5.66719294e-01 7.49417484e-01 2.12713823e-01 -1.12954438e+00
-5.23424923e-01 -7.77920902e-01 -5.99902928e-01 -1.16916552e-01
-1.79127693e-01 -6.39092445e-01 -3.89696397e-02 -5.60690045e-01] | [4.14199686050415, 1.2421175241470337] |
3890cb22-7094-4f23-bf5a-7f01ae342271 | a-unifying-framework-for-causal-explanation | 2205.15462 | null | https://arxiv.org/abs/2205.15462v2 | https://arxiv.org/pdf/2205.15462v2.pdf | Causal Explanations for Sequential Decision Making Under Uncertainty | We introduce a novel framework for causal explanations of stochastic, sequential decision-making systems built on the well-studied structural causal model paradigm for causal reasoning. This single framework can identify multiple, semantically distinct explanations for agent actions -- something not previously possible. In this paper, we establish exact methods and several approximation techniques for causal inference on Markov decision processes using this framework, followed by results on the applicability of the exact methods and some run time bounds. We discuss several scenarios that illustrate the framework's flexibility and the results of experiments with human subjects that confirm the benefits of this approach. | ['Shlomo Zilberstein', 'Claudia V. Goldman', 'Saaduddin Mahmud', 'Samer B. Nashed'] | 2022-05-30 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 3.20483387e-01 7.01964557e-01 -2.33898267e-01 -5.29859841e-01
-3.50135654e-01 -3.87738407e-01 1.11408699e+00 3.50015223e-01
-1.85614638e-02 9.59368587e-01 6.63058043e-01 -9.54931796e-01
-7.05107629e-01 -5.60067773e-01 -2.86555320e-01 -6.48551524e-01
-4.42799062e-01 6.73101902e-01 4.90683317e-01 -1.10923640e-01
4.79291916e-01 3.74377966e-01 -1.37648344e+00 2.11877018e-01
4.27167386e-01 3.46918434e-01 -1.82371750e-01 9.55000818e-01
-1.84885878e-02 1.30644941e+00 -3.36253017e-01 -3.20378423e-01
-4.52045910e-02 -6.03810668e-01 -1.12615168e+00 -2.19208971e-01
-1.93078935e-01 -4.80723172e-01 -3.38689387e-01 7.87726521e-01
-1.17082801e-02 2.65024126e-01 9.43044364e-01 -1.69138694e+00
-5.68253160e-01 1.14759517e+00 -2.25529745e-01 4.02999163e-01
9.54623699e-01 1.07519053e-01 1.03035486e+00 -2.38625094e-01
3.67224932e-01 2.06297207e+00 4.08719659e-01 7.29564965e-01
-1.44283319e+00 -3.13382506e-01 7.03604102e-01 3.80982459e-01
-8.26855600e-01 -2.61848241e-01 2.81526536e-01 -4.75580633e-01
9.16460395e-01 4.57445681e-01 5.13085961e-01 1.17086399e+00
5.18710315e-01 7.47341454e-01 1.51729250e+00 -5.47093034e-01
5.75810194e-01 -2.37318367e-01 5.30513346e-01 6.36132479e-01
5.14782965e-01 8.25932503e-01 -6.99027359e-01 -6.46804750e-01
9.67134655e-01 -3.96236703e-02 1.09408855e-01 -3.63099426e-01
-1.11733294e+00 9.31757152e-01 -6.57200217e-02 1.63115323e-01
-4.07173872e-01 7.71286547e-01 1.40357524e-01 1.76568970e-01
1.37491360e-01 1.19769864e-01 -2.44315490e-01 -2.61881277e-02
-2.62613028e-01 7.46724546e-01 9.76076245e-01 8.50692809e-01
-6.48266694e-04 -1.01539761e-01 -2.91432649e-01 -2.53382981e-01
7.38277256e-01 4.65713531e-01 -1.87428162e-01 -1.16440082e+00
1.22897699e-01 2.07680106e-01 5.95984817e-01 -5.80445707e-01
-4.75005448e-01 2.66718030e-01 -3.95621449e-01 4.24947590e-01
6.25961423e-01 -2.01575592e-01 -6.63877428e-01 1.75880301e+00
3.52207065e-01 2.23947287e-01 3.39976758e-01 6.54363692e-01
1.47579640e-01 3.86666715e-01 5.88248312e-01 -6.10717297e-01
1.10989928e+00 -4.27639514e-01 -9.10207570e-01 6.10508844e-02
1.70662925e-01 -3.37845087e-01 6.62027240e-01 4.48483288e-01
-9.81018722e-01 -8.55394304e-02 -6.76128566e-01 2.48983875e-01
2.42281571e-01 -6.76899374e-01 1.22065687e+00 7.56444871e-01
-8.92192781e-01 6.45052850e-01 -1.01996720e+00 -4.16591972e-01
-5.70274191e-03 2.71800935e-01 1.99465811e-01 1.65220678e-01
-1.21115601e+00 9.56920445e-01 3.04029882e-01 -1.52403638e-01
-1.38242519e+00 -3.78262669e-01 -4.82256800e-01 1.56844184e-01
7.00865507e-01 -9.98270512e-01 1.84949505e+00 -6.51972771e-01
-1.43641174e+00 1.85696021e-01 -4.17080194e-01 -6.69337332e-01
7.45899498e-01 -3.21771324e-01 -4.57806855e-01 2.14206800e-01
2.45726839e-01 1.17206201e-01 4.58179235e-01 -1.34708416e+00
-9.02117610e-01 -3.52941245e-01 4.95664805e-01 1.57340802e-03
6.36873007e-01 3.59663874e-01 5.29548824e-01 -3.28932822e-01
3.14603969e-02 -8.54204237e-01 -9.63652015e-01 -6.17579460e-01
-6.40530944e-01 -5.36398888e-01 4.70246710e-02 -1.08076610e-01
1.06094611e+00 -1.76701880e+00 2.63417279e-03 4.69281852e-01
1.31465420e-01 -5.65153301e-01 2.48325452e-01 7.48118758e-01
-4.23570007e-01 4.44727331e-01 -2.47095078e-01 4.06063721e-02
3.72156978e-01 4.29243088e-01 -8.15116882e-01 3.68044227e-01
8.93730391e-03 8.16488206e-01 -1.11741114e+00 -4.53770995e-01
4.65900153e-01 2.02106480e-02 -2.34193742e-01 2.18903840e-01
-4.86562967e-01 4.19115901e-01 -6.91484571e-01 3.37516189e-01
-4.79906350e-02 -4.06130850e-02 7.27983534e-01 6.72091901e-01
-2.76411027e-01 6.17656648e-01 -1.33879483e+00 9.94203210e-01
-4.96356897e-02 3.58064950e-01 -4.33817655e-01 -5.59125721e-01
1.50536701e-01 7.78456926e-01 1.99426010e-01 -1.36660516e-01
1.66175947e-01 -9.44162533e-02 3.50146666e-02 -6.52149320e-01
1.22208113e-03 -6.95806146e-01 -2.87146807e-01 7.94073761e-01
-4.32569653e-01 -9.84718353e-02 1.42671153e-01 3.57281119e-01
1.25688374e+00 2.26171136e-01 9.01851952e-01 -5.89897692e-01
2.56200850e-01 2.82421887e-01 4.33811635e-01 1.26802635e+00
-2.19775930e-01 -1.75822787e-02 9.37009752e-01 -6.25926018e-01
-7.23668098e-01 -1.23234200e+00 -7.15719536e-03 8.12360168e-01
4.00613457e-01 -4.55319703e-01 -7.86066294e-01 -6.85093164e-01
1.42926678e-01 1.52061725e+00 -9.86284435e-01 1.55545935e-01
-3.87853920e-01 -7.48174012e-01 4.35863465e-01 7.58387804e-01
-1.15347229e-01 -8.84726465e-01 -1.10964608e+00 1.97260872e-01
-3.84107940e-02 -6.87618375e-01 2.83916831e-01 2.63777912e-01
-9.62761581e-01 -1.37686980e+00 5.55194095e-02 1.84452251e-01
3.65816385e-01 3.09723586e-01 9.50392365e-01 1.08629644e-01
4.23519313e-02 6.62293732e-01 -9.93806273e-02 -4.94898528e-01
-8.26485991e-01 -7.37834275e-01 3.00243586e-01 -4.06346977e-01
3.45275939e-01 -4.51987833e-01 -2.94420213e-01 2.74918437e-01
-6.92387462e-01 7.25298794e-03 1.71115294e-01 5.26220620e-01
1.67007610e-01 3.05916041e-01 2.74595767e-01 -9.00794744e-01
8.78465354e-01 -6.13777578e-01 -8.15709412e-01 3.54934305e-01
-8.92737746e-01 4.48082834e-01 1.68159366e-01 -1.74304605e-01
-1.61761665e+00 -9.51485857e-02 4.29716200e-01 2.03009635e-01
-5.00627756e-01 4.34668899e-01 -1.89628154e-01 4.78742272e-01
6.51780069e-01 -3.65667254e-01 -2.99730241e-01 -2.35538453e-01
5.61913550e-01 9.96009540e-03 4.55931813e-01 -9.16333973e-01
6.36675775e-01 6.10243857e-01 5.20673931e-01 -2.60459602e-01
-7.95077622e-01 -5.25791645e-02 -4.12613571e-01 -2.90639460e-01
9.75482643e-01 -5.15854955e-01 -1.35192907e+00 -1.21657044e-01
-1.39061415e+00 -5.48849463e-01 -3.17451239e-01 6.64339900e-01
-1.01893401e+00 2.05700129e-01 -5.69582164e-01 -1.64117432e+00
5.36945105e-01 -1.01503587e+00 7.02530980e-01 5.06245941e-02
-8.47438753e-01 -1.21557987e+00 2.49893382e-01 -6.06789428e-04
-1.99604198e-01 2.70331621e-01 1.20729411e+00 -8.45719218e-01
-7.29213178e-01 2.26768941e-01 -2.11799368e-02 -4.72863913e-01
-2.21095711e-01 1.98503569e-01 -7.39810824e-01 3.23204428e-01
1.28390208e-01 1.02507591e-01 6.96344972e-01 6.00770116e-01
6.69012785e-01 -4.13400173e-01 -7.48644233e-01 -4.32239145e-01
1.36540675e+00 5.82744837e-01 3.86124551e-01 2.52462983e-01
3.47629875e-01 8.79661560e-01 7.98760235e-01 4.60201651e-01
4.18622732e-01 4.92694467e-01 5.26945114e-01 2.91931987e-01
2.63988882e-01 -3.86183709e-01 4.36495185e-01 -1.39569342e-01
-7.19473720e-01 -4.79569644e-01 -9.28650439e-01 5.61602831e-01
-2.41343451e+00 -1.48292184e+00 -8.16207051e-01 2.05255866e+00
5.37013650e-01 3.06979686e-01 2.78959930e-01 8.54859054e-02
7.25935340e-01 -2.90048957e-01 -3.62493515e-01 -7.66533971e-01
1.16623320e-01 -1.21333515e-02 6.20688200e-01 1.08198786e+00
-8.97536039e-01 6.86482787e-01 8.91790485e+00 2.28782386e-01
-5.04481420e-02 2.91951120e-01 3.85976374e-01 -3.08408402e-02
-7.18571901e-01 5.45559824e-01 -6.94761336e-01 -1.28529398e-02
1.29640841e+00 -4.95610833e-01 3.34475577e-01 7.19999671e-01
5.46781898e-01 -5.58406234e-01 -1.49460447e+00 2.62236118e-01
-5.16606808e-01 -1.12467539e+00 -1.10040801e-02 2.13889584e-01
6.62382245e-01 -5.96538067e-01 -2.77339220e-01 -2.34531939e-01
1.62680101e+00 -1.01773930e+00 1.19660926e+00 4.13868725e-01
1.66431978e-01 -7.83154547e-01 6.22316837e-01 1.89631805e-01
-6.92343891e-01 -5.18020689e-01 -7.21948817e-02 -7.08823860e-01
5.79500496e-01 6.34618342e-01 -8.00881684e-01 6.85240805e-01
4.65398312e-01 2.98308879e-01 1.87122021e-02 6.43667936e-01
-1.16720164e+00 7.73397565e-01 -2.34759465e-01 -9.37996730e-02
2.18287155e-01 4.32480350e-02 6.78687155e-01 1.16626263e+00
5.91524467e-02 6.91823423e-01 1.71819460e-02 8.65197420e-01
8.50025654e-01 -4.65921462e-01 -7.42734849e-01 2.28287935e-01
3.96742612e-01 4.86253887e-01 -8.95288289e-01 -5.92424929e-01
-1.59037158e-01 5.12433946e-01 -1.26788214e-01 3.50176156e-01
-1.11775732e+00 5.34602821e-01 6.69605315e-01 -4.36154753e-02
-2.15611070e-01 -2.78259277e-01 -5.89619458e-01 -9.24985647e-01
-4.59378779e-01 -8.09180021e-01 8.25933278e-01 -6.31404877e-01
-1.24196661e+00 2.85338581e-01 9.26035106e-01 -5.77684224e-01
-7.28971303e-01 -4.20802742e-01 -5.35634160e-01 7.75695026e-01
-8.88634384e-01 -7.67567754e-01 3.48875552e-01 4.67994660e-01
4.48230803e-01 1.93637356e-01 7.84220576e-01 -5.86877525e-01
-4.14783180e-01 -3.11307073e-01 -2.89604038e-01 -5.57826877e-01
2.47483298e-01 -1.72578955e+00 5.21031022e-01 1.06004071e+00
3.60379182e-02 8.84442866e-01 1.38008463e+00 -9.16909814e-01
-1.06701386e+00 -4.68760312e-01 1.11569262e+00 -8.02046001e-01
9.01611388e-01 -2.91333813e-02 -5.53585649e-01 1.14113903e+00
4.38155234e-01 -8.79473805e-01 6.52251184e-01 6.09355867e-01
-2.02574104e-01 3.86791617e-01 -9.26985443e-01 1.02327418e+00
1.28207302e+00 -1.63692579e-01 -1.28714716e+00 3.92305344e-01
7.23661244e-01 4.49475273e-02 -4.49308306e-01 2.58497000e-01
7.05323517e-01 -1.25209105e+00 9.45096672e-01 -1.07802939e+00
6.92579865e-01 -5.30528903e-01 -2.51968145e-01 -1.03670132e+00
-5.00784039e-01 -9.89323735e-01 4.82474156e-02 9.55872536e-01
3.55356306e-01 -8.26469660e-01 2.04126656e-01 1.16870999e+00
1.89523175e-01 -7.92055428e-02 -1.05603874e+00 -7.60825872e-01
-9.22693610e-02 -8.30302894e-01 6.81228876e-01 7.52645612e-01
4.89789337e-01 3.74227017e-01 -1.58063650e-01 5.35723686e-01
9.70527709e-01 2.04881504e-01 4.39809144e-01 -1.29922116e+00
-3.54361206e-01 -3.83797973e-01 -2.05856964e-01 -6.45259917e-01
3.31307888e-01 -2.70023793e-01 1.73893616e-01 -1.50136995e+00
3.54096174e-01 -3.33785504e-01 -1.05299659e-01 5.68573892e-01
-3.43840778e-01 -5.86078882e-01 1.07578702e-01 1.44302070e-01
-5.06294787e-01 3.18828911e-01 7.92894542e-01 1.95795357e-01
-1.26329377e-01 1.65661827e-01 -8.31800580e-01 1.06894016e+00
8.44960034e-01 -8.16319466e-01 -7.55320072e-01 -3.34471539e-02
4.68647897e-01 7.00445771e-01 9.56646919e-01 -5.72517514e-01
1.46696970e-01 -9.93607104e-01 1.04104832e-01 -4.16508496e-01
-1.36627913e-01 -8.03946257e-01 5.70005953e-01 8.78587246e-01
-9.11752462e-01 2.36714676e-01 7.67911524e-02 1.03995299e+00
1.79632604e-01 -3.24671775e-01 2.88688630e-01 -1.32295400e-01
-7.16325521e-01 -2.98644274e-01 -9.54042315e-01 -4.77620184e-01
1.26506364e+00 1.47648156e-01 -2.70295292e-01 -5.38227677e-01
-1.01689124e+00 2.20693424e-01 1.15874715e-01 2.78230578e-01
5.18575847e-01 -1.04759395e+00 -6.06808245e-01 -5.18279791e-01
-2.15053394e-01 -4.27250624e-01 8.26734155e-02 8.02266061e-01
-6.51194304e-02 7.34450758e-01 -6.95652813e-02 -2.60356605e-01
-1.17384446e+00 1.13312721e+00 2.41945043e-01 -2.27757692e-01
-5.84933817e-01 4.30611581e-01 6.58629656e-01 1.66579112e-01
-2.21374348e-01 -4.75680709e-01 -2.03009233e-01 -2.55422235e-01
7.60107756e-01 6.48775160e-01 -6.69614315e-01 -6.31086305e-02
-6.29137278e-01 -6.69834996e-03 2.69757092e-01 -9.43937123e-01
9.56080437e-01 -6.18307590e-01 -2.93536216e-01 9.85787034e-01
-1.45050377e-01 -1.15095124e-01 -1.30720520e+00 7.12782294e-02
4.12654936e-01 -3.60939354e-01 -3.33305866e-01 -8.84438157e-01
-2.61351466e-01 6.64066315e-01 2.19956562e-01 7.53030121e-01
7.34946012e-01 2.61307240e-01 -1.70744270e-01 1.88424408e-01
5.93988538e-01 -7.08185852e-01 -4.71275628e-01 1.55176803e-01
9.19881642e-01 -7.30337977e-01 1.31030202e-01 -7.90806890e-01
-5.23463786e-01 1.16988671e+00 1.01956755e-01 -3.24923992e-02
3.76741827e-01 3.18147689e-01 -1.62468374e-01 -4.36345696e-01
-1.34545636e+00 -3.06248605e-01 -2.57189840e-01 6.39885962e-01
2.72916883e-01 7.05112755e-01 -3.88203114e-01 4.69222248e-01
-1.17698543e-01 1.29120365e-01 9.40891504e-01 8.57085407e-01
-3.19640726e-01 -1.05290830e+00 -7.68028915e-01 8.02807603e-03
-5.10073721e-01 -1.02444939e-01 -6.00450575e-01 1.01456797e+00
-3.13461512e-01 1.76652682e+00 -1.76163808e-01 -1.09995179e-01
2.47338459e-01 2.04783037e-01 7.82618940e-01 -4.27784622e-01
-4.04259384e-01 1.46253232e-03 5.46455503e-01 -8.52371633e-01
-7.79456913e-01 -1.28082216e+00 -1.34583175e+00 -6.03252828e-01
2.32590437e-02 2.60648042e-01 2.58145183e-01 1.33399332e+00
-2.43643522e-01 6.78727388e-01 2.42749169e-01 -3.83987099e-01
-7.08882809e-01 -5.20013809e-01 -4.70158249e-01 3.47188562e-02
2.67683834e-01 -8.64924550e-01 -4.93561447e-01 2.57311165e-01] | [8.22803783416748, 5.835142135620117] |
8eafcd60-7878-4924-a579-0880bb1dab03 | machine-learning-based-intrusion-detection-1 | 2307.01570 | null | https://arxiv.org/abs/2307.01570v1 | https://arxiv.org/pdf/2307.01570v1.pdf | Machine Learning-Based Intrusion Detection: Feature Selection versus Feature Extraction | Internet of things (IoT) has been playing an important role in many sectors, such as smart cities, smart agriculture, smart healthcare, and smart manufacturing. However, IoT devices are highly vulnerable to cyber-attacks, which may result in security breaches and data leakages. To effectively prevent these attacks, a variety of machine learning-based network intrusion detection methods for IoT networks have been developed, which often rely on either feature extraction or feature selection techniques for reducing the dimension of input data before being fed into machine learning models. This aims to make the detection complexity low enough for real-time operations, which is particularly vital in any intrusion detection systems. This paper provides a comprehensive comparison between these two feature reduction methods of intrusion detection in terms of various performance metrics, namely, precision rate, recall rate, detection accuracy, as well as runtime complexity, in the presence of the modern UNSW-NB15 dataset as well as both binary and multiclass classification. For example, in general, the feature selection method not only provides better detection performance but also lower training and inference time compared to its feature extraction counterpart, especially when the number of reduced features K increases. However, the feature extraction method is much more reliable than its selection counterpart, particularly when K is very small, such as K = 4. Additionally, feature extraction is less sensitive to changing the number of reduced features K than feature selection, and this holds true for both binary and multiclass classifications. Based on this comparison, we provide a useful guideline for selecting a suitable intrusion detection type for each specific scenario, as detailed in Tab. 14 at the end of Section IV. | ['Hung Tran', 'Thien Van Luong', 'Tuan-Cuong Vuong', 'Vu-Duc Ngo'] | 2023-07-04 | null | null | null | null | ['intrusion-detection', 'network-intrusion-detection'] | ['miscellaneous', 'miscellaneous'] | [ 2.41134971e-01 -3.12910080e-01 -3.03773701e-01 -1.87605977e-01
1.41127497e-01 -4.76529747e-01 3.80074769e-01 5.72170079e-01
-4.42803115e-01 5.86885810e-01 -5.05600393e-01 -5.50350606e-01
-5.45292974e-01 -1.24320745e+00 4.90798205e-02 -8.04982781e-01
-5.18548936e-02 1.32274315e-01 3.59720737e-01 2.34834254e-02
1.89680144e-01 9.06959653e-01 -1.64506209e+00 -1.70845091e-01
6.14558995e-01 1.41929615e+00 -4.49992329e-01 1.09174348e-01
-6.89383373e-02 7.60579854e-02 -5.85276961e-01 -4.35474545e-01
3.52848649e-01 -1.31363645e-01 -5.11433661e-01 -1.42330244e-01
-5.84586680e-01 2.32095718e-02 -1.85593203e-01 1.24884057e+00
2.65781730e-01 1.08944587e-02 4.50553775e-01 -1.58263767e+00
-4.09308262e-03 5.90408146e-01 -2.68289745e-01 6.35066256e-02
1.37181476e-01 2.13635921e-01 5.26124179e-01 -1.43505245e-01
1.66881643e-02 9.79428530e-01 2.63731956e-01 3.64277750e-01
-9.80160058e-01 -1.12497723e+00 -1.58794262e-02 2.24052101e-01
-1.32201421e+00 -1.38162687e-01 7.37707198e-01 -9.66398194e-02
8.43826175e-01 5.56575835e-01 5.56433201e-01 7.00310409e-01
4.55364168e-01 2.96779215e-01 9.37275529e-01 -3.61762106e-01
5.21190882e-01 2.97528327e-01 3.91904503e-01 1.56008467e-01
9.24052358e-01 1.37172341e-01 2.21366957e-01 -3.26770663e-01
2.88774669e-01 5.16505539e-01 -1.26844253e-02 -9.74997431e-02
-9.49630678e-01 8.90434802e-01 3.09329897e-01 5.77707231e-01
-3.95966142e-01 -4.14146453e-01 6.42459512e-01 4.06251907e-01
-7.70629272e-02 4.52232420e-01 -8.50572050e-01 5.89399077e-02
-4.02292013e-01 -3.38890851e-01 9.29250002e-01 5.63790262e-01
8.32692087e-01 5.18409200e-02 2.79843301e-01 3.30050349e-01
1.70047522e-01 5.82557499e-01 5.54503977e-01 -2.99572885e-01
3.82208914e-01 1.00642776e+00 -2.48918355e-01 -1.08268034e+00
-8.02606702e-01 -3.21361214e-01 -1.43764555e+00 3.41145918e-02
1.27034336e-01 -1.05577335e-01 -6.66922748e-01 1.47307587e+00
6.45433962e-01 -1.17197856e-02 3.29205543e-01 3.19110692e-01
4.42924440e-01 4.55819696e-01 7.71472678e-02 -4.85093176e-01
1.48629153e+00 -1.24436453e-01 -7.32879400e-01 -5.05765341e-02
6.31785035e-01 -6.29996181e-01 5.92418075e-01 4.89055485e-01
-3.15625638e-01 -3.92344445e-01 -1.21211755e+00 7.16496944e-01
-9.03777957e-01 -2.96841234e-01 8.27154338e-01 1.27649939e+00
-2.04312205e-01 4.36693490e-01 -7.79634655e-01 -5.11236191e-01
3.57290179e-01 7.86971092e-01 -3.57784986e-01 -9.85203031e-03
-1.35480571e+00 6.81719601e-01 8.18351090e-01 -6.83072433e-02
3.37278284e-02 -2.46645212e-01 -7.68774152e-01 1.60523653e-01
2.78957278e-01 -3.27039719e-01 5.64726591e-01 -5.06806791e-01
-1.33439398e+00 2.06318468e-01 3.07352692e-01 -5.74008882e-01
1.18575148e-01 8.27739835e-02 -1.09615159e+00 -4.30851839e-02
-2.13029489e-01 6.86181411e-02 7.63602853e-01 -6.88252866e-01
-8.01981330e-01 -6.66927159e-01 3.89130376e-02 -5.83749592e-01
-6.76670551e-01 6.64220899e-02 1.47357225e-01 -2.66956031e-01
2.22301379e-01 -9.69934225e-01 -3.73494059e-01 -4.25683767e-01
-5.56359529e-01 -1.97111681e-01 1.44328213e+00 -9.01757181e-02
1.20377362e+00 -2.26499081e+00 -4.96849865e-01 8.94869268e-01
-1.48975402e-01 7.58132041e-01 1.22553624e-01 3.42972130e-01
-1.47169754e-01 2.83800751e-01 -1.25786647e-01 4.01224017e-01
-1.20256804e-01 2.18132839e-01 -1.69078544e-01 4.59382415e-01
2.50786215e-01 5.03701985e-01 -5.72849512e-01 -8.91653076e-02
6.87947154e-01 5.14621019e-01 -2.00039685e-01 -2.15891823e-01
1.98015913e-01 4.91601318e-01 -9.12525356e-01 6.97697937e-01
7.21310794e-01 -7.31902272e-02 2.97635287e-01 -2.88069546e-01
-2.95215216e-03 1.56121790e-01 -1.39318824e+00 6.12939417e-01
-6.70396984e-01 1.03145123e-01 -3.26304793e-01 -1.10354483e+00
1.28892291e+00 4.01404798e-01 7.02321112e-01 -8.34588766e-01
4.41319138e-01 2.40012631e-01 2.29689509e-01 -2.26032317e-01
-1.89608052e-01 5.55526838e-02 -2.78303295e-01 3.92606765e-01
-4.89538074e-01 4.62767966e-02 3.44450712e-01 -2.04495311e-01
1.22966981e+00 -7.09741712e-01 8.00806940e-01 -4.56079133e-02
9.67254162e-01 -4.07418072e-01 6.27167106e-01 4.45665658e-01
-6.16849512e-02 -1.90682575e-01 2.75115579e-01 -6.55334294e-01
-4.66136456e-01 -5.39825559e-01 -3.86040866e-01 1.74909368e-01
3.76671553e-01 -3.99756670e-01 -4.48621988e-01 -9.58465755e-01
6.59924820e-02 7.65689552e-01 -2.29243711e-01 -6.54564083e-01
-4.20811892e-01 -9.87098038e-01 5.58941305e-01 2.52753496e-01
8.64552915e-01 -1.08279097e+00 -1.05950010e+00 1.73274457e-01
2.17516631e-01 -1.29675496e+00 2.41219610e-01 5.25023162e-01
-9.55875218e-01 -1.53020489e+00 2.80691743e-01 -3.10789287e-01
5.40201068e-01 2.87449360e-01 5.22635043e-01 3.12439352e-01
-1.68353379e-01 4.80900556e-02 -6.47594273e-01 -7.24843562e-01
-4.92743582e-01 1.63753480e-01 3.23004723e-01 1.45877600e-01
7.87055969e-01 -5.47403395e-01 -4.98250604e-01 6.21800542e-01
-1.25360000e+00 -6.48620367e-01 7.90852666e-01 5.51314712e-01
3.21474940e-01 9.50554073e-01 7.42847621e-01 -6.60335243e-01
3.46238196e-01 -4.45959896e-01 -7.62521744e-01 1.57216877e-01
-8.78307462e-01 -1.33884177e-01 1.05215263e+00 -4.76384163e-01
-5.74368894e-01 4.29948308e-02 -3.28220159e-01 7.08974525e-02
-4.72290367e-01 2.98555344e-01 -6.98865533e-01 -3.02687943e-01
4.05570209e-01 4.07653153e-02 -5.37934080e-02 -3.10041517e-01
-2.85836965e-01 9.89510000e-01 -1.35163546e-01 1.90145522e-02
1.05401421e+00 3.83559406e-01 4.70375538e-01 -1.09446990e+00
-3.45062494e-01 -4.06005800e-01 -4.96246397e-01 1.58826992e-01
4.17576969e-01 -3.20223033e-01 -9.64839816e-01 4.11647230e-01
-7.88827121e-01 2.88669974e-01 -2.61789203e-01 7.87342310e-01
5.59200486e-03 3.22911620e-01 -2.92162448e-01 -7.57912040e-01
-6.53168082e-01 -1.29665124e+00 5.06410062e-01 3.15402865e-01
-3.01236153e-01 -7.78180718e-01 -4.06790078e-01 -1.09660672e-02
5.09499848e-01 3.36405396e-01 1.27284300e+00 -1.24149835e+00
-1.21945068e-01 -8.52710605e-01 -1.73001692e-01 4.80187684e-01
6.92879975e-01 9.52519570e-03 -6.05481863e-01 -5.26702106e-01
3.53385478e-01 1.47529021e-01 3.87367755e-01 1.28303552e-02
1.13680959e+00 -4.35450375e-01 -5.34325659e-01 4.56817031e-01
1.30739820e+00 6.97665989e-01 8.49264741e-01 6.57270193e-01
3.19847822e-01 4.49257046e-01 8.39347839e-01 4.79180574e-01
-1.91989005e-01 4.92158294e-01 7.66635716e-01 8.98213387e-02
4.67482179e-01 3.33534539e-01 6.99819401e-02 3.93949658e-01
2.27083117e-01 -2.00933084e-01 -7.05064178e-01 -5.24817258e-02
-1.48562002e+00 -7.43415952e-01 -6.27207831e-02 2.58542085e+00
3.44302177e-01 4.27884161e-01 1.29537418e-01 1.07898903e+00
7.98523307e-01 -1.72070831e-01 -7.05168068e-01 -5.61063826e-01
1.56961650e-01 4.16284204e-01 6.34144425e-01 -1.84636906e-01
-1.21630216e+00 4.42636788e-01 5.13240576e+00 6.13843799e-01
-1.46639609e+00 -1.16079420e-01 4.03903782e-01 3.53627801e-01
2.64728874e-01 -5.34243621e-02 -7.59215474e-01 7.04960883e-01
1.03490484e+00 -2.03962680e-02 1.11249022e-01 8.81237924e-01
5.76654486e-02 -7.08372518e-02 -6.72159195e-01 9.71751332e-01
-3.99562210e-01 -8.73391688e-01 1.04964674e-01 2.73699224e-01
1.78611562e-01 -2.41923943e-01 -2.04209581e-01 8.87022018e-02
-1.42187908e-01 -6.13900304e-01 7.31165186e-02 -1.66139930e-01
4.84000683e-01 -1.26674712e+00 1.18909585e+00 3.53119493e-01
-1.18735397e+00 -5.57297349e-01 -2.17878506e-01 -1.29849151e-01
-7.22422004e-02 1.08751118e+00 -4.10462916e-01 8.20221722e-01
7.49627352e-01 2.96661466e-01 -3.39099765e-01 1.12177646e+00
-2.37382188e-01 4.24517423e-01 -6.04989707e-01 -1.64344221e-01
-1.37546569e-01 -4.67089489e-02 4.35958624e-01 8.16832721e-01
3.70032877e-01 1.11522675e-01 3.03590417e-01 1.64454341e-01
2.07027674e-01 1.51414141e-01 -5.21759272e-01 2.75043603e-02
6.76970541e-01 1.32103956e+00 -1.14437068e+00 -9.95241255e-02
-4.41905081e-01 3.91535103e-01 -4.06556666e-01 2.11246032e-02
-7.10554183e-01 -8.12998533e-01 7.97254443e-01 7.82476217e-02
2.23993897e-01 -1.08639568e-01 -2.77653694e-01 -9.33159769e-01
-1.93299934e-01 -8.15617859e-01 5.40084898e-01 2.10995734e-01
-1.12004912e+00 6.37410641e-01 -4.08227071e-02 -1.47066021e+00
-7.63912573e-02 -5.71615636e-01 -6.04884923e-01 3.12866777e-01
-1.37975276e+00 -8.24058712e-01 -2.06362963e-01 7.68985271e-01
8.51662010e-02 -4.36401479e-02 1.04593086e+00 4.18710887e-01
-8.87605786e-01 6.75780118e-01 -3.76310311e-02 1.14633977e-01
3.56110126e-01 -4.43768829e-01 2.39942074e-01 9.31960523e-01
-9.40183252e-02 6.18033350e-01 6.17700219e-01 -5.52234650e-01
-1.38840389e+00 -1.29289162e+00 4.82674420e-01 2.59656250e-01
5.48591256e-01 -7.35208467e-02 -6.38764381e-01 3.79542202e-01
-4.38566893e-01 1.27386376e-01 7.77987301e-01 -1.31755263e-01
-2.69561082e-01 -5.45662761e-01 -1.76659131e+00 4.62738276e-01
5.24593413e-01 -1.54023096e-01 -1.65619731e-01 1.91235468e-01
4.61445659e-01 1.27041698e-01 -1.08330858e+00 8.27507198e-01
5.54098248e-01 -9.56054330e-01 8.79525959e-01 -2.12838963e-01
-4.46843654e-01 -4.61914718e-01 -1.21921226e-01 -8.33115160e-01
-2.27901638e-01 -4.34138119e-01 -2.30954632e-01 1.43915796e+00
2.94796526e-01 -1.35997152e+00 5.87104797e-01 4.87404913e-01
4.43530440e-01 -6.84498310e-01 -1.00362670e+00 -1.15240073e+00
-4.04281288e-01 -4.97246891e-01 1.10536063e+00 8.97896588e-01
-5.60336262e-02 2.75967985e-01 -4.60740505e-03 3.50346148e-01
4.74988371e-01 4.66381796e-02 7.17706501e-01 -1.71129775e+00
1.69165134e-01 -2.92364806e-01 -9.77528512e-01 -1.64031819e-01
-8.55110809e-02 -4.28657472e-01 -4.69436884e-01 -1.08343840e+00
-2.07167879e-01 -7.63479829e-01 -6.14610136e-01 7.28703201e-01
2.48343602e-01 3.73804063e-01 9.02477130e-02 1.41740456e-01
-2.13452071e-01 1.73500553e-01 7.71605074e-01 -2.80921340e-01
-4.99946892e-01 7.22280264e-01 -5.15980303e-01 8.25686753e-01
1.29098535e+00 -5.47776639e-01 -3.67676437e-01 3.25206280e-01
1.35890618e-01 -2.77606964e-01 1.69545427e-01 -1.22728789e+00
7.52838179e-02 -3.04217041e-01 4.16169137e-01 -3.81413370e-01
3.57407294e-02 -1.49870813e+00 3.80992323e-01 1.07319331e+00
3.62490147e-01 -8.06309506e-02 2.45520502e-01 3.89192730e-01
-1.09507933e-01 -1.34652615e-01 8.69659126e-01 2.08446264e-01
-5.51607370e-01 2.71577269e-01 -3.35688919e-01 -5.69723010e-01
1.42246783e+00 -5.35555601e-01 -2.82291204e-01 -3.41767408e-02
-5.01438200e-01 1.46112023e-02 4.09546793e-01 4.45711136e-01
5.25315702e-01 -1.11229849e+00 -1.40783414e-01 7.03818440e-01
1.61897063e-01 -1.34935379e-01 6.52860478e-02 9.29142833e-01
-1.55935556e-01 5.23379445e-01 -2.79822826e-01 -4.33297813e-01
-1.29939699e+00 9.06241596e-01 -1.38653293e-01 -4.22016889e-01
-4.87829000e-01 1.66853398e-01 -2.81167507e-01 -1.30658999e-01
2.29921252e-01 -3.01220447e-01 -3.49393755e-01 7.21943453e-02
8.06333840e-01 6.80225253e-01 5.74737072e-01 -4.24349546e-01
-7.98775852e-01 6.69611454e-01 -6.90290034e-02 5.95078468e-01
1.00577939e+00 -5.37517667e-02 -2.86031932e-01 -3.11337616e-02
1.04955184e+00 -6.53849840e-02 -4.22940493e-01 6.12392351e-02
2.39126906e-01 -3.30367595e-01 -9.54685360e-02 -5.81614554e-01
-1.31019855e+00 5.33460021e-01 7.33060896e-01 9.87816393e-01
1.41141069e+00 -3.15878808e-01 9.83806908e-01 5.58184624e-01
8.27548265e-01 -7.38119304e-01 -3.00913572e-01 4.80033934e-01
1.89633578e-01 -1.03079939e+00 1.51316836e-01 -6.62371874e-01
-1.73979148e-01 1.16041386e+00 3.50282580e-01 7.66270757e-02
1.04222798e+00 3.99640948e-01 -1.62376866e-01 8.05111080e-02
-1.74772322e-01 -1.62612975e-01 -7.99606517e-02 6.63423836e-01
-9.49298963e-02 7.75811151e-02 -4.49174523e-01 5.34035802e-01
-1.54785454e-01 -1.44301578e-01 2.62559444e-01 1.08061302e+00
-3.69542658e-01 -1.37267244e+00 -4.39164281e-01 6.67267501e-01
-5.60107291e-01 2.35839605e-01 -2.15050697e-01 8.76539409e-01
1.88832805e-01 1.45327342e+00 -2.48943064e-02 -8.27576637e-01
5.00834644e-01 -8.17015171e-02 4.92371023e-02 -2.88225859e-01
-6.26037121e-01 -4.29195583e-01 -1.54452145e-01 -6.85109437e-01
-3.94035786e-01 -3.77595752e-01 -1.40064037e+00 -4.54335064e-01
-7.82908320e-01 3.74389231e-01 8.97830486e-01 1.21120417e+00
3.62743258e-01 4.47614253e-01 9.45327163e-01 -5.32333970e-01
-5.81834078e-01 -7.60827780e-01 -6.37040496e-01 1.86019897e-01
2.63069812e-02 -7.11842299e-01 -5.90937972e-01 -6.33749425e-01] | [5.22662353515625, 7.141569137573242] |
9d16eb2f-3172-48bb-9606-c543bffb362f | balancing-profit-risk-and-sustainability-for | 2207.02134 | null | https://arxiv.org/abs/2207.02134v1 | https://arxiv.org/pdf/2207.02134v1.pdf | Balancing Profit, Risk, and Sustainability for Portfolio Management | Stock portfolio optimization is the process of continuous reallocation of funds to a selection of stocks. This is a particularly well-suited problem for reinforcement learning, as daily rewards are compounding and objective functions may include more than just profit, e.g., risk and sustainability. We developed a novel utility function with the Sharpe ratio representing risk and the environmental, social, and governance score (ESG) representing sustainability. We show that a state-of-the-art policy gradient method - multi-agent deep deterministic policy gradients (MADDPG) - fails to find the optimum policy due to flat policy gradients and we therefore replaced gradient descent with a genetic algorithm for parameter optimization. We show that our system outperforms MADDPG while improving on deep Q-learning approaches by allowing for continuous action spaces. Crucially, by incorporating risk and sustainability criteria in the utility function, we improve on the state-of-the-art in reinforcement learning for portfolio optimization; risk and sustainability are essential in any modern trading strategy and we propose a system that does not merely report these metrics, but that actively optimizes the portfolio to improve on them. | ['Christian W. Omlin', 'Charl Maree'] | 2022-06-06 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-6.65141523e-01 1.47125810e-01 -3.59967500e-01 1.50265098e-01
-6.82650805e-01 -6.78953111e-01 6.52853727e-01 1.70448154e-01
-6.88116372e-01 1.24035168e+00 3.81251991e-01 -5.01705885e-01
-4.96113241e-01 -1.08299446e+00 -5.54548442e-01 -6.75091922e-01
-3.86170089e-01 6.07245326e-01 8.50402638e-02 -5.25068581e-01
6.90042198e-01 3.69802922e-01 -1.30312574e+00 -3.84315282e-01
8.58208060e-01 1.23297441e+00 -1.68759748e-01 5.39714456e-01
-1.28960088e-01 1.23480451e+00 -6.15827799e-01 -6.69672668e-01
4.91851896e-01 -4.54488605e-01 -5.54102659e-01 -3.43764782e-01
-1.57378867e-01 -6.15812063e-01 1.19286440e-01 8.82088780e-01
6.12320662e-01 1.07726075e-01 5.92118740e-01 -1.25200236e+00
-6.65496290e-01 7.49216497e-01 -3.19638401e-01 1.25997394e-01
-7.94302076e-02 3.73045295e-01 1.63277137e+00 -3.59006405e-01
4.21811312e-01 1.16191673e+00 5.25870621e-01 4.40022290e-01
-9.08691287e-01 -3.50769520e-01 4.23937261e-01 1.69275552e-01
-3.50151122e-01 -2.90489197e-02 6.11471951e-01 -4.22791690e-01
1.14582253e+00 2.00647011e-01 1.18554556e+00 6.46843553e-01
4.72366661e-01 8.20557237e-01 1.31925046e+00 -2.91211545e-01
7.33248889e-01 7.29786381e-02 -5.02381802e-01 4.48536158e-01
4.22944039e-01 8.14345777e-01 -3.01573038e-01 -1.85950711e-01
7.89706469e-01 -1.44629762e-01 1.40987962e-01 -5.86472332e-01
-9.02457356e-01 1.25603533e+00 1.82820126e-01 1.51477277e-01
-7.39318430e-01 6.72913373e-01 3.02012146e-01 7.03836203e-01
5.02801120e-01 1.10161340e+00 -3.95734966e-01 -4.49828535e-01
-8.70862007e-01 7.10739076e-01 1.12226987e+00 1.90422982e-01
6.40302300e-01 4.11058903e-01 -5.47333479e-01 4.00330842e-01
2.53308147e-01 3.77082705e-01 4.33079928e-01 -1.44063938e+00
3.17198336e-01 4.27055687e-01 8.43116164e-01 -5.34781754e-01
-3.64687145e-01 -6.51654065e-01 -8.18850771e-02 8.83365273e-01
5.05124688e-01 -5.55005193e-01 -3.61234993e-01 1.57198274e+00
1.97210208e-01 -5.52571900e-02 8.97222906e-02 6.07374012e-01
-1.06326036e-01 4.97135848e-01 -9.37565491e-02 -4.98251647e-01
8.00968707e-01 -1.04761314e+00 -7.47344375e-01 -5.37885837e-02
4.69967425e-01 -3.20729256e-01 9.86444652e-01 5.37433326e-01
-1.51964378e+00 1.17801733e-01 -1.03017175e+00 8.08689952e-01
-3.35524529e-01 -4.38381016e-01 6.65630996e-01 8.35009217e-01
-1.31814253e+00 1.32977140e+00 -5.31631589e-01 1.86412349e-01
3.23177248e-01 2.77108908e-01 5.04644096e-01 6.29456997e-01
-1.29874849e+00 1.48694956e+00 2.09192619e-01 -2.93799251e-01
-9.69875455e-01 -7.24596381e-01 -3.25490206e-01 2.34785125e-01
8.54532003e-01 -5.32246649e-01 1.57900572e+00 -8.01795721e-01
-2.16899729e+00 3.92715544e-01 8.77689242e-01 -7.28525698e-01
9.94902551e-01 -3.37783605e-01 -1.12676077e-01 2.60075368e-02
-5.42025454e-02 1.70913994e-01 9.34336662e-01 -9.61985826e-01
-7.82700181e-01 -1.16351597e-01 3.84791851e-01 5.10218441e-01
-5.12871981e-01 1.31031079e-02 5.16042411e-01 -6.05811536e-01
-7.93105543e-01 -6.99545264e-01 -5.02473295e-01 -2.27446645e-01
-1.00213159e-02 -2.95889735e-01 2.04468697e-01 -5.49535096e-01
1.40988719e+00 -1.48761272e+00 -3.32769081e-02 2.05590054e-01
-1.26256377e-01 2.20090464e-01 -1.07218236e-01 6.32005513e-01
1.81422994e-01 2.52935320e-01 -1.96678683e-01 -1.29862010e-01
6.77268922e-01 2.04970822e-01 -3.88803750e-01 4.52280730e-01
2.73077637e-01 1.07279658e+00 -1.26575506e+00 -6.14673458e-02
-3.90925929e-02 -2.27394432e-01 -6.02059782e-01 3.66605282e-01
-6.95484519e-01 1.14059178e-02 -4.60209876e-01 5.66559196e-01
1.89254850e-01 -4.49037142e-02 4.00053374e-02 5.66715717e-01
-6.60841286e-01 3.71418238e-01 -1.34343791e+00 1.02483666e+00
-4.18643326e-01 -1.22580091e-02 -2.83635892e-02 -1.09998548e+00
9.97980773e-01 1.28606588e-01 8.78913820e-01 -9.12385345e-01
-1.05006382e-01 6.37927949e-01 -8.77112374e-02 -2.51990169e-01
5.68846047e-01 -3.33997160e-01 3.19630504e-02 1.08549714e+00
-2.52824098e-01 -2.75557220e-01 3.13994169e-01 -1.52845800e-01
1.14644361e+00 6.27909005e-01 1.87987924e-01 -5.05361319e-01
2.40295500e-01 -1.10968128e-01 5.90673447e-01 7.97991574e-01
-4.06018943e-01 -2.47453898e-02 1.24792349e+00 -2.96547860e-01
-1.05626905e+00 -1.06086874e+00 2.52486318e-01 1.21503007e+00
-2.68359989e-01 1.75119221e-01 -6.25797153e-01 -8.62490773e-01
7.20782578e-01 9.57741976e-01 -7.03451693e-01 -1.30158857e-01
-6.46792829e-01 -1.01577055e+00 1.32148921e-01 3.29723775e-01
2.91268975e-01 -1.40609205e+00 -9.96138513e-01 7.76761532e-01
5.94157994e-01 -2.67689109e-01 -4.94036853e-01 2.93526649e-01
-7.35802710e-01 -1.08722091e+00 -1.05834806e+00 -1.21791981e-01
5.01264818e-02 -4.54187185e-01 1.48850965e+00 -1.64320901e-01
1.04419366e-01 6.31147027e-01 -2.00478807e-01 -6.81770742e-01
-4.80320454e-01 -1.83932468e-01 1.35312798e-02 -1.39352381e-01
-2.11580992e-01 -6.23721600e-01 -9.47076499e-01 1.59704298e-01
-7.29614854e-01 -5.97239494e-01 4.40607816e-01 1.05176997e+00
3.98758113e-01 -8.01282823e-02 1.34576261e+00 -6.69063151e-01
1.13149250e+00 -4.97480869e-01 -1.22727156e+00 3.69966775e-01
-1.38496888e+00 3.87450844e-01 5.82617581e-01 -2.59084135e-01
-8.85445178e-01 -5.40072680e-01 9.31465905e-03 -1.06623344e-01
8.01748037e-01 5.42003870e-01 1.85462922e-01 -6.54417127e-02
4.19411689e-01 -2.40575280e-02 3.10346514e-01 -2.59679526e-01
5.10926902e-01 1.80023029e-01 4.16875668e-02 -6.27331197e-01
5.16408682e-01 2.47662902e-01 2.06023842e-01 -6.42929785e-03
-6.43098354e-01 -6.93612099e-02 -3.76254804e-02 -4.32359785e-01
4.81206477e-01 -5.05054235e-01 -1.07317901e+00 5.28670192e-01
-6.16480291e-01 -6.57076657e-01 -9.29120779e-01 3.95440221e-01
-1.16305006e+00 4.26571481e-02 -5.19690037e-01 -1.26482534e+00
-4.85738397e-01 -8.59462619e-01 3.89599502e-01 1.83369905e-01
1.76482365e-01 -1.32963538e+00 6.33712888e-01 -2.92970836e-01
8.68520975e-01 4.94367629e-01 6.61151528e-01 -4.92042214e-01
-5.40398896e-01 1.90336972e-01 8.60190839e-02 4.83426362e-01
8.36760476e-02 -8.58157352e-02 -6.15974844e-01 -4.47985053e-01
2.22530276e-01 -5.37188709e-01 9.09989476e-01 5.06685913e-01
6.47174716e-01 -7.09550261e-01 3.18549186e-01 4.05807555e-01
1.76209629e+00 3.01108718e-01 4.70274061e-01 1.21991897e+00
2.32770100e-01 7.50946283e-01 8.41624260e-01 1.00626767e+00
3.54476333e-01 4.83336151e-01 8.53666425e-01 2.22636819e-01
3.70112389e-01 -1.48735806e-01 6.54576242e-01 3.62480044e-01
-1.28269851e-01 1.18534312e-01 -6.31310761e-01 6.33790970e-01
-2.13199592e+00 -1.12714636e+00 5.32600641e-01 2.35914087e+00
1.08802640e+00 4.32293266e-01 7.97268152e-01 -1.27421975e-01
3.21495175e-01 3.01111072e-01 -1.16251802e+00 -7.90067732e-01
-9.38515738e-02 3.17750871e-01 1.03602052e+00 4.49322671e-01
-8.43159378e-01 7.75780678e-01 6.80367899e+00 6.32779598e-01
-1.00561142e+00 -1.67591631e-01 7.11643636e-01 -4.04415429e-01
-7.98115492e-01 -3.18443985e-03 -7.32525766e-01 5.84750652e-01
1.01979613e+00 -5.09171724e-01 7.94954181e-01 8.66500914e-01
4.66997474e-01 -1.29163176e-01 -8.09090018e-01 3.16263109e-01
-5.31813562e-01 -1.44987833e+00 -4.45262074e-01 2.56438881e-01
9.76241112e-01 -8.45887065e-02 3.12495410e-01 4.71048862e-01
9.62315977e-01 -9.63815331e-01 1.07170594e+00 7.47044861e-01
2.94654846e-01 -1.19377732e+00 4.52008516e-01 1.30141526e-01
-7.30531037e-01 -6.17979348e-01 -2.56859958e-01 -1.46449255e-02
1.46471903e-01 5.73473752e-01 -4.95269269e-01 3.49537134e-01
4.59357888e-01 4.34749544e-01 -1.73791602e-01 1.04513597e+00
-3.91911447e-01 4.53027070e-01 -1.83356017e-01 -5.91103733e-01
8.02713037e-01 -5.63816011e-01 4.98492777e-01 8.11554193e-01
5.00444472e-01 -4.15791154e-01 4.04807106e-02 9.92972970e-01
-5.75586129e-03 1.40697211e-01 -3.41258585e-01 -3.11891943e-01
2.34086812e-01 9.70203280e-01 -3.63773584e-01 -9.68132168e-02
-2.47299656e-01 4.84393865e-01 4.07449931e-01 2.18617857e-01
-6.10796869e-01 -3.27630937e-01 9.20926273e-01 6.00264175e-03
6.38357937e-01 -3.17347087e-02 -3.26362491e-01 -9.12243187e-01
-2.01577265e-02 -8.28538299e-01 4.40994740e-01 -1.84937641e-01
-1.40841281e+00 8.59967917e-02 -1.44839317e-01 -1.12942851e+00
-7.99567938e-01 -7.16371536e-01 -8.86922359e-01 8.56752217e-01
-2.26168418e+00 -5.48348308e-01 5.92037857e-01 2.98131496e-01
-1.81363225e-02 -4.74961996e-01 5.44900954e-01 -1.58732831e-01
-3.76610696e-01 3.56495440e-01 8.57994676e-01 -3.55124027e-01
5.02879620e-01 -2.00893521e+00 4.29350168e-01 4.46860164e-01
-3.19266647e-01 1.39033878e-02 6.92442417e-01 -7.23602474e-01
-1.33032298e+00 -7.89946258e-01 4.58122462e-01 -1.07927278e-01
1.35185051e+00 1.02830321e-01 -4.79167908e-01 1.42950550e-01
4.11165893e-01 -3.82216513e-01 3.53267252e-01 -6.70379028e-02
1.49598926e-01 -2.86102980e-01 -1.11292434e+00 7.65953600e-01
7.14770317e-01 -1.40470386e-01 -3.26253265e-01 3.00589800e-01
7.11378753e-01 -1.35427371e-01 -9.72023606e-01 5.63503355e-02
5.15567899e-01 -1.18081248e+00 8.36636245e-01 -7.75845349e-01
3.84491056e-01 8.28815103e-02 9.16737765e-02 -1.76073980e+00
-3.70669544e-01 -1.15484273e+00 -6.11864328e-01 1.01232934e+00
3.76506984e-01 -9.79022086e-01 1.01931345e+00 5.50146401e-01
4.24599163e-02 -1.25494921e+00 -1.07682955e+00 -1.27103114e+00
6.26560688e-01 9.14815217e-02 9.45217431e-01 7.07451105e-01
5.61307259e-02 -2.08364442e-01 -5.30715525e-01 -5.06365418e-01
9.16284502e-01 2.93521285e-01 2.25421727e-01 -8.65463376e-01
-6.57579899e-01 -1.21004927e+00 3.07202637e-01 -3.66733670e-01
-5.87497391e-02 -4.81104046e-01 -2.05710664e-01 -1.60431945e+00
-3.88953894e-01 -4.95710284e-01 -7.75885463e-01 2.11758286e-01
-2.67462730e-01 -4.41622764e-01 4.47880954e-01 4.28238325e-02
-5.24748325e-01 9.69961405e-01 1.47455466e+00 -2.35278998e-02
-4.03513968e-01 3.09981346e-01 -1.01451373e+00 3.67344409e-01
1.05380380e+00 -4.95233208e-01 -2.56969333e-01 1.21757790e-01
8.18834603e-01 3.29686582e-01 4.74323444e-02 -6.63370430e-01
-7.56217316e-02 -7.19932497e-01 2.80237291e-03 -3.00372243e-01
2.60457862e-02 -3.10486764e-01 -1.27532840e-01 9.12007213e-01
-3.76405954e-01 3.14292371e-01 -2.08385751e-01 3.95323336e-01
-3.03884353e-02 -6.46791220e-01 6.77218378e-01 -6.26050889e-01
-4.96944398e-01 2.56211758e-01 -3.44368845e-01 3.05136800e-01
9.80465293e-01 -3.60398330e-02 -4.14666474e-01 -6.01473451e-01
-4.01123315e-01 7.46624649e-01 4.42119777e-01 9.50321853e-02
4.43334043e-01 -1.36919427e+00 -1.03677118e+00 -4.77871239e-01
-3.36730987e-01 -5.01999259e-01 -2.44709983e-01 5.13433039e-01
-5.25074542e-01 2.73581386e-01 -3.88459474e-01 1.90418735e-01
-3.19481760e-01 3.66628170e-01 8.58177304e-01 -1.14007068e+00
-2.74408609e-01 6.83960617e-01 -2.14249492e-01 -3.09140682e-01
3.13659191e-01 -1.86121196e-01 -3.68428141e-01 6.00967526e-01
4.72768843e-01 7.73640513e-01 -5.12651764e-02 1.90856174e-01
-4.22807448e-02 2.49728188e-01 1.25130460e-01 -6.57226145e-01
1.72954357e+00 8.54025930e-02 -4.76846471e-02 2.94144541e-01
7.23053396e-01 -1.73314512e-01 -1.82374036e+00 -1.37838097e-02
4.72424805e-01 -3.77615750e-01 2.14807987e-01 -1.12098014e+00
-1.03375280e+00 4.94707972e-01 5.26244879e-01 7.19859779e-01
8.93828154e-01 -6.03362620e-01 6.64234340e-01 4.69577372e-01
3.65155846e-01 -1.93679500e+00 4.42644149e-01 6.51706636e-01
1.01558065e+00 -1.03570652e+00 2.47716650e-01 6.16373360e-01
-9.12988007e-01 1.08471704e+00 3.17072898e-01 -5.85785031e-01
8.21934462e-01 2.21991599e-01 4.33064252e-02 -3.88619266e-02
-9.55110490e-01 -4.79364425e-01 -9.68670696e-02 4.25620168e-01
3.51981074e-02 1.50138423e-01 -5.21470845e-01 4.74140644e-01
-2.27570593e-01 -1.88381523e-02 6.53420568e-01 1.17001688e+00
-6.92938805e-01 -1.49814951e+00 -3.84200722e-01 6.69543445e-01
-7.69365549e-01 -3.57891954e-02 -2.03629702e-01 6.41016304e-01
-2.72932142e-01 7.05881953e-01 1.64742433e-02 2.72846017e-02
4.50433940e-01 -2.81836122e-01 4.77209717e-01 -3.43239546e-01
-1.08074689e+00 5.04591204e-02 1.69070333e-01 -5.94121635e-01
-3.24260473e-01 -9.96250927e-01 -9.82224941e-01 -3.46107841e-01
-1.13825426e-01 3.58579069e-01 7.69072711e-01 7.75959551e-01
-1.93930902e-02 5.43349504e-01 1.27854180e+00 -6.97680414e-01
-1.62561727e+00 -6.30225658e-01 -1.10157633e+00 1.02918083e-02
4.51704413e-01 -7.47297943e-01 -4.04515356e-01 -7.64868379e-01] | [4.351058006286621, 3.6806254386901855] |
12d82b46-3a75-4b53-945d-3f56f6dc4840 | easy-and-efficient-transformer-scalable-1 | null | null | https://aclanthology.org/2022.naacl-industry.8 | https://aclanthology.org/2022.naacl-industry.8.pdf | Easy and Efficient Transformer: Scalable Inference Solution For Large NLP Model | Recently, large-scale transformer-based models have been proven to be effective over various tasks across many domains. Nevertheless, applying them in industrial production requires tedious and heavy works to reduce inference costs. To fill such a gap, we introduce a scalable inference solution: Easy and Efficient Transformer (EET), including a series of transformer inference optimization at the algorithm and implementation levels. First, we design highly optimized kernels for long inputs and large hidden sizes. Second, we propose a flexible CUDA memory manager to reduce the memory footprint when deploying a large model. Compared with the state-of-the-art transformer inference library (Faster Transformer v4.0), EET can achieve an average of 1.40-4.20x speedup on the transformer decoder layer with an A100 GPU. | ['Zeng Zhao', 'Xiaoxi Mao', 'Changjie Fan', 'Bai Liu', 'Rongsheng Zhang', 'Ziyang Luo', 'Duan Wang', 'Jingzhen Ding', 'Yadong Xi', 'Gongzheng li'] | null | null | null | null | naacl-acl-2022-7 | ['inference-optimization'] | ['audio'] | [-1.23699918e-01 -4.19730507e-02 -8.66938531e-02 -3.49977821e-01
-9.49500203e-01 -4.56409454e-01 2.10737780e-01 -3.97763789e-01
-7.06192181e-02 5.51653624e-01 -4.90518436e-02 -8.35985243e-01
3.23477864e-01 -1.11654985e+00 -9.14674163e-01 -4.66255009e-01
5.68808556e-01 6.16089880e-01 4.67401564e-01 -5.03640361e-02
4.40276004e-02 1.00858755e-01 -1.42484176e+00 3.13507557e-01
8.89235914e-01 1.09398377e+00 5.69968045e-01 6.41326904e-01
-1.49302617e-01 8.00704002e-01 -5.32382369e-01 -9.15861726e-01
-9.49598923e-02 -2.11008862e-01 -6.63875997e-01 -4.15254444e-01
4.47738916e-01 -7.02473879e-01 -3.61670971e-01 1.03686786e+00
7.96503365e-01 -4.42593127e-01 6.76996857e-02 -1.19713533e+00
-3.66426528e-01 1.11423242e+00 -3.47076654e-01 5.09853959e-02
3.70605811e-02 1.66237175e-01 1.08967984e+00 -8.41228306e-01
1.79743886e-01 1.21203744e+00 7.93470800e-01 1.82608053e-01
-9.14039731e-01 -1.14643228e+00 -3.98174599e-02 3.37831438e-01
-1.42002797e+00 -6.60341978e-01 3.27470690e-01 -1.39519125e-01
1.69262922e+00 2.72598386e-01 8.35229933e-01 1.05181372e+00
4.44262892e-01 9.96812880e-01 6.81886435e-01 -1.94966849e-02
2.93673631e-02 -9.11182761e-02 9.51478258e-02 1.20373762e+00
-3.31353769e-03 -2.40063265e-01 -4.76260960e-01 -1.38679937e-01
7.04167485e-01 -1.68978363e-01 2.62874514e-01 4.00787085e-01
-1.17007220e+00 7.22675323e-01 1.89370558e-01 -4.63519152e-03
-3.42183560e-02 6.03616178e-01 8.44772160e-01 2.28821546e-01
4.55716401e-01 9.01570171e-02 -5.59282362e-01 -6.22372329e-01
-9.45294917e-01 2.98868001e-01 7.74374723e-01 1.48607302e+00
7.33656883e-01 2.67864615e-01 -4.82861996e-01 8.68276656e-01
2.70655483e-01 6.49110675e-01 1.84889421e-01 -8.32659066e-01
7.45072842e-01 5.50807118e-01 -4.75550205e-01 -6.49683535e-01
-7.93075636e-02 -5.82345426e-01 -1.05188811e+00 -3.95244122e-01
1.94185123e-01 -2.70238006e-03 -6.01236045e-01 1.56022215e+00
2.61465162e-01 1.57459021e-01 -2.25165620e-01 5.73803604e-01
6.30424619e-01 9.20413435e-01 -3.04532219e-02 6.97659776e-02
1.61551130e+00 -1.49493182e+00 -9.37623918e-01 -4.50326413e-01
7.26030231e-01 -1.02815247e+00 1.18854141e+00 4.24930483e-01
-1.35375011e+00 -5.69909155e-01 -1.10130954e+00 -6.89286351e-01
-2.58416116e-01 5.72113812e-01 1.09660542e+00 6.20907962e-01
-1.00968242e+00 4.02521938e-01 -1.17946434e+00 -7.10140588e-03
3.49518925e-01 4.86012429e-01 1.62677631e-01 3.71320061e-02
-1.12389851e+00 7.80982971e-01 4.27947342e-01 3.53029728e-01
-1.02725196e+00 -9.54027891e-01 -8.28982115e-01 3.41251105e-01
2.85454482e-01 -9.80258226e-01 1.53242433e+00 -3.99441987e-01
-1.95334065e+00 6.14889443e-01 -4.90329087e-01 -4.44500536e-01
3.71186078e-01 -4.52041030e-01 -2.92224973e-01 -4.42621827e-01
-1.44010365e-01 3.67501676e-01 8.34637880e-01 -2.41270468e-01
-7.29825318e-01 -2.09855944e-01 8.74238387e-02 -1.43357992e-01
-3.52751881e-01 1.82308033e-01 -1.02181780e+00 -6.26055062e-01
-2.52284944e-01 -9.38826382e-01 1.60311796e-02 -1.05915070e-01
-6.88233137e-01 -3.05655122e-01 8.00547540e-01 -8.72376144e-01
1.63423824e+00 -1.94997430e+00 1.82364747e-01 -1.63031876e-01
4.88853961e-01 4.13555831e-01 3.22894782e-01 1.38474509e-01
4.34307158e-01 -2.53243327e-01 -2.19651163e-02 -4.60566640e-01
4.74039882e-01 3.37359667e-01 -4.11422402e-01 4.90402319e-02
-8.97391737e-02 1.21569419e+00 -7.33259976e-01 -7.03742325e-01
2.94412374e-01 4.34449404e-01 -9.84297335e-01 4.12446558e-01
-2.81348616e-01 -7.62402117e-02 -5.84039390e-01 8.32720518e-01
8.00736785e-01 -4.54914927e-01 3.94917190e-01 -5.50355136e-01
-1.49516553e-01 7.89015055e-01 -7.64082909e-01 1.83562791e+00
-8.12032163e-01 7.55265474e-01 7.11891875e-02 -7.68775225e-01
5.91567397e-01 2.93459624e-01 -5.05240373e-02 -7.15207338e-01
3.25689226e-01 4.07420754e-01 -9.47606191e-02 -1.95571721e-01
5.68466723e-01 1.94835901e-01 -2.48644531e-01 4.84260470e-01
3.68077159e-01 2.26822793e-02 1.59800217e-01 7.74749294e-02
1.10030639e+00 7.30134249e-02 8.38729665e-02 -3.79963100e-01
2.98119485e-01 -2.74632990e-01 6.24658525e-01 6.07909620e-01
2.77672648e-01 2.09888041e-01 6.25978351e-01 -4.37601447e-01
-1.07583857e+00 -1.10028446e+00 -1.08671121e-01 1.32108414e+00
-2.83007950e-01 -9.54646587e-01 -1.10151184e+00 -2.39223719e-01
-2.47845158e-01 7.65326738e-01 -5.33914901e-02 -7.15807006e-02
-7.17354298e-01 -9.06934023e-01 1.03270447e+00 8.77376854e-01
1.19753540e+00 -6.46406233e-01 -5.84544957e-01 4.13545132e-01
-2.37281933e-01 -1.33387232e+00 -6.76270962e-01 9.67505872e-02
-8.49776089e-01 -5.43232381e-01 -2.02050418e-01 -6.86328828e-01
2.69720227e-01 -2.26495191e-01 1.52751458e+00 -2.68397797e-02
-2.12001711e-01 -4.50572789e-01 1.05549537e-01 -2.51927465e-01
-3.04455787e-01 6.56667411e-01 -2.04480261e-01 -5.08257866e-01
3.26853365e-01 -6.74074709e-01 -3.23217750e-01 2.66357571e-01
-4.39813882e-01 8.62057924e-01 7.57746220e-01 7.92352080e-01
6.13448560e-01 -6.50258437e-02 7.54093705e-03 -9.56594169e-01
6.42008722e-01 -2.50070542e-01 -1.08004487e+00 5.68477452e-01
-6.85408533e-01 2.76355475e-01 8.35581899e-01 -2.89521158e-01
-1.18577611e+00 -1.05351381e-01 -7.77136326e-01 -3.65423888e-01
6.19918823e-01 2.65601546e-01 -3.02033722e-01 1.60117492e-01
3.59354503e-02 4.03434604e-01 -2.36409858e-01 -4.44007546e-01
1.59427926e-01 6.41676068e-01 3.19312036e-01 -8.96114469e-01
5.11808097e-01 -1.47473887e-01 -2.42299199e-01 -3.35132003e-01
-1.02253067e+00 1.21264376e-01 -2.10063517e-01 1.17987990e-01
7.88429379e-01 -1.24429262e+00 -1.08613145e+00 7.74451256e-01
-1.30874395e+00 -6.54805243e-01 3.18967164e-01 2.46686891e-01
-3.33723545e-01 -1.41733885e-01 -1.04620528e+00 -3.46679956e-01
-1.04375851e+00 -1.79786420e+00 1.53184152e+00 -7.17988312e-02
-8.82718712e-02 -9.35488462e-01 -1.12691097e-01 4.60659534e-01
6.97685838e-01 -3.66527468e-01 9.11724091e-01 -3.95209081e-02
-1.13467109e+00 7.45736882e-02 -4.78149593e-01 3.33644390e-01
-1.82350904e-01 6.98439330e-02 -1.08184290e+00 -2.40705371e-01
-1.69273585e-01 -2.80339450e-01 5.61447918e-01 1.22691751e-01
1.71112418e+00 -3.85112673e-01 -5.94222069e-01 1.18242776e+00
1.26867342e+00 -7.94700384e-02 7.18504190e-01 -1.35583788e-01
9.10290420e-01 -3.78863752e-01 4.17989165e-01 3.84853840e-01
5.54434180e-01 9.21286821e-01 -1.09184003e-02 -9.17814150e-02
-6.04053475e-02 -5.72477043e-01 4.54694331e-01 1.55318117e+00
-6.70308471e-02 -2.19260186e-01 -6.93171144e-01 2.88353026e-01
-1.81145716e+00 -6.70783341e-01 -3.16686481e-02 1.89545584e+00
1.07688344e+00 3.61898839e-01 -1.66049883e-01 2.67582107e-03
3.38926554e-01 1.24434985e-01 -5.97463548e-01 -7.26063371e-01
2.32805386e-01 6.32665992e-01 5.63117623e-01 4.68547910e-01
-7.26588666e-01 1.28085160e+00 7.22055721e+00 1.40325487e+00
-1.11161566e+00 4.42557454e-01 4.36906487e-01 -1.17035806e-01
-2.47508347e-01 5.19219562e-02 -1.30846262e+00 6.72784090e-01
1.28422523e+00 -1.23939238e-01 6.35208666e-01 1.08931589e+00
-2.40027189e-01 1.24269277e-01 -1.14547527e+00 1.05899751e+00
-2.50153691e-01 -1.56204557e+00 8.94170031e-02 4.96404171e-02
2.55941570e-01 1.42351806e-01 6.40291199e-02 7.19157338e-01
4.56078589e-01 -8.66390347e-01 6.87804818e-01 1.14834987e-01
9.10722136e-01 -8.73647988e-01 6.63564265e-01 1.62413627e-01
-1.55641913e+00 1.88550338e-01 -6.11809075e-01 -2.77640730e-01
2.99282223e-01 9.28457677e-01 -7.29690492e-01 2.34727308e-01
8.04502606e-01 6.80452406e-01 -4.33308542e-01 3.39080662e-01
-4.26404715e-01 7.96794534e-01 -4.68850911e-01 -9.48188826e-02
2.06773788e-01 -2.32852802e-01 -6.91185668e-02 1.38461602e+00
4.78672624e-01 -1.09166771e-01 -1.68168485e-01 9.92766798e-01
-2.34886214e-01 -3.10258955e-01 -2.72479743e-01 -1.93856284e-02
7.47214973e-01 1.33174050e+00 -3.16699713e-01 -6.35612607e-01
-4.44807112e-01 1.13557553e+00 6.74170792e-01 -5.75139076e-02
-1.63622010e+00 -2.96673447e-01 9.89324450e-01 -8.38642195e-02
2.57147193e-01 -1.96512178e-01 -3.13352734e-01 -1.31486809e+00
9.42455530e-02 -9.52402711e-01 1.64286509e-01 -8.83873999e-01
-7.27406859e-01 5.78401148e-01 -1.97040141e-01 -7.95347333e-01
-2.60768056e-01 -6.40669107e-01 -3.26856047e-01 7.48157442e-01
-1.33519042e+00 -1.32070684e+00 -2.16494471e-01 5.07895470e-01
6.44020021e-01 -1.34754509e-01 9.56218839e-01 9.09256339e-01
-9.31813598e-01 1.17830956e+00 -1.25588698e-03 2.26441011e-01
3.50636810e-01 -1.09666348e+00 1.00667214e+00 7.62460828e-01
-1.68674007e-01 9.53422189e-01 4.82004225e-01 -3.08944404e-01
-2.20015097e+00 -1.23857379e+00 1.03180337e+00 -3.74815047e-01
8.39911878e-01 -8.67831707e-01 -5.94461381e-01 1.28062332e+00
3.86901915e-01 -7.61674643e-02 3.80613893e-01 3.18404645e-01
-4.84251440e-01 -4.12858725e-01 -8.59130859e-01 7.06551075e-01
1.37274861e+00 -8.29498291e-01 -5.80622405e-02 5.09159923e-01
9.04397011e-01 -1.03409076e+00 -1.26487625e+00 2.00128347e-01
8.02273452e-01 -8.31947148e-01 1.09602475e+00 -1.22373223e-01
5.49327970e-01 -1.27227709e-01 -2.00451434e-01 -1.00654244e+00
-3.06577176e-01 -7.73659468e-01 -5.21991432e-01 1.23248649e+00
4.82624561e-01 -7.70877242e-01 5.40282249e-01 3.42673570e-01
-4.58995253e-01 -9.14075553e-01 -8.34521234e-01 -4.75085378e-01
1.45776160e-02 -6.76987946e-01 8.30712676e-01 3.87196243e-01
-7.16755539e-02 7.68936872e-01 -5.68865240e-01 7.42850080e-02
6.53955996e-01 3.52093011e-01 9.47761714e-01 -6.90213084e-01
-6.09255314e-01 -3.74175370e-01 -1.47479758e-01 -1.47537255e+00
3.21102381e-01 -8.69922519e-01 -1.11767441e-01 -1.24759269e+00
4.63787258e-01 -4.82738018e-01 6.06787950e-02 6.65770888e-01
-1.59568161e-01 8.60880911e-02 -3.51371691e-02 -1.94439515e-01
-6.09569430e-01 4.84159559e-01 1.18683851e+00 -2.53510743e-01
3.95868719e-01 -9.54487473e-02 -4.54422176e-01 5.01273811e-01
7.05407143e-01 -3.44522208e-01 -5.67008317e-01 -1.27233255e+00
5.39877236e-01 -3.39751244e-02 1.51839137e-01 -1.00281262e+00
2.81639695e-01 2.54412621e-01 1.03373177e-01 -8.25008094e-01
5.17917633e-01 -5.86768150e-01 4.89827216e-01 4.04451519e-01
-7.40065202e-02 4.41873550e-01 4.26349044e-01 7.41205923e-03
-1.81072056e-01 1.13806352e-01 6.47648394e-01 -1.11532517e-01
-6.15586519e-01 3.33213240e-01 -3.26476455e-01 1.32844150e-01
6.96965218e-01 1.86095536e-01 -3.31434578e-01 9.36773419e-02
-1.92076504e-01 1.92761689e-01 8.90288278e-02 3.19695324e-01
4.13507819e-01 -1.28189123e+00 -5.01203239e-01 4.28445011e-01
-1.59106508e-01 4.83929366e-02 2.54293770e-01 8.46375227e-01
-7.62705386e-01 7.03871071e-01 -2.25304794e-02 -7.29762375e-01
-1.25309324e+00 3.97176981e-01 3.44834775e-01 -5.90638101e-01
-7.40177155e-01 8.46131504e-01 1.22294568e-01 -6.01119459e-01
3.88520993e-02 -7.92375326e-01 6.33697808e-01 -3.14567208e-01
6.50123715e-01 4.61423635e-01 4.34027225e-01 4.09741234e-03
-2.71755785e-01 5.08740723e-01 -8.47953856e-02 2.04411209e-01
1.13296068e+00 2.75497109e-01 -4.61488128e-01 1.84575453e-01
1.34297621e+00 -1.38752759e-01 -9.71193433e-01 -1.36456206e-01
-3.92928928e-01 -2.83560485e-01 3.78304154e-01 -5.98120332e-01
-1.40380287e+00 1.02538788e+00 3.59713078e-01 -2.58150637e-01
1.20395684e+00 -1.57329053e-01 1.41216147e+00 5.76290309e-01
5.94554901e-01 -1.07263708e+00 -4.88903701e-01 7.95539677e-01
4.24091667e-01 -8.30301046e-01 7.89396465e-02 -5.55161655e-01
-2.51951307e-01 9.35806990e-01 5.95119476e-01 1.32264271e-01
6.43071055e-01 1.15103054e+00 -4.31852579e-01 -1.06168605e-01
-1.15279520e+00 1.40477359e-01 -2.23140772e-02 1.24090984e-01
5.96324384e-01 3.69987756e-01 -7.67209232e-02 5.49408436e-01
-5.64523101e-01 2.37059876e-01 -2.09055454e-01 6.90416753e-01
-1.51318997e-01 -9.99155283e-01 3.86680551e-02 5.26966095e-01
-4.94042784e-01 -5.91144323e-01 1.18747085e-01 6.25694811e-01
-1.22756682e-01 5.95297754e-01 3.31846505e-01 -5.40761054e-01
1.63027912e-01 -6.98689148e-02 8.02116930e-01 -2.99723804e-01
-8.17131579e-01 3.49223837e-02 3.13627839e-01 -8.10680091e-01
1.15408465e-01 -1.87442273e-01 -1.00731385e+00 -9.49407995e-01
-2.16384009e-01 -1.78695127e-01 5.49685299e-01 1.08493614e+00
6.29178703e-01 9.00098503e-01 7.06790423e-04 -5.49960017e-01
-4.60710675e-01 -1.08852708e+00 -3.57121348e-01 -2.44701058e-01
-2.01902345e-01 -5.94417691e-01 4.62622456e-02 -8.82381797e-02] | [8.73317813873291, 3.604215621948242] |
267abd60-b489-4a08-87cd-7bdbcdc4ac67 | light-field-for-rf | 1901.03953 | null | http://arxiv.org/abs/1901.03953v1 | http://arxiv.org/pdf/1901.03953v1.pdf | Light-Field for RF | Most computer vision systems and computational photography systems are
visible light based which is a small fraction of the electromagnetic (EM)
spectrum. In recent years radio frequency (RF) hardware has become more widely
available, for example, many cars are equipped with a RADAR, and almost every
home has a WiFi device. In the context of imaging, RF spectrum holds many
advantages compared to visible light systems. In particular, in this regime, EM
energy effectively interacts in different ways with matter. This property
allows for many novel applications such as privacy preserving computer vision
and imaging through absorbing and scattering materials in visible light such as
walls. Here, we expand many of the concepts in computational photography in
visible light to RF cameras. The main limitation of imaging with RF is the
large wavelength that limits the imaging resolution when compared to visible
light. However, the output of RF cameras is usually processed by computer
vision and perception algorithms which would benefit from multi-modal sensing
of the environment, and from sensing in situations in which visible light
systems fail. To bridge the gap between computational photography and RF
imaging, we expand the concept of light-field to RF. This work paves the way to
novel computational sensing systems with RF. | ['Ramesh Raskar', 'Manikanta Kotaru', 'Guy Satat', 'Sachin Katti'] | 2019-01-13 | null | null | null | null | ['rf-based-pose-estimation'] | ['computer-vision'] | [ 7.52741218e-01 -1.59262478e-01 2.28583351e-01 -1.87798455e-01
1.99783891e-01 -5.81119120e-01 5.72807074e-01 -7.42471039e-01
-5.85125148e-01 8.10017347e-01 -6.88784719e-02 -3.56506348e-01
-4.48512249e-02 -1.20854676e+00 -3.74913007e-01 -1.14143503e+00
5.86312711e-01 -2.17846483e-01 2.01536380e-02 -2.11906433e-01
1.21709608e-01 5.57307959e-01 -1.67918408e+00 6.21622503e-02
4.66551065e-01 1.08771336e+00 6.26324534e-01 4.25026655e-01
-1.65660039e-01 5.06382525e-01 -4.71249163e-01 -1.16447225e-01
6.71823442e-01 -2.49785781e-01 -3.13471317e-01 -2.80093461e-01
3.31077456e-01 -3.21175605e-01 -4.37596679e-01 1.19517052e+00
3.03487480e-01 1.36545703e-01 3.71920109e-01 -8.75414371e-01
-7.89284348e-01 1.56948745e-01 -6.97529972e-01 -7.51767009e-02
6.78324699e-01 1.12650342e-01 2.03448087e-01 -5.20734549e-01
6.12542927e-01 8.09581935e-01 2.87180156e-01 6.40588939e-01
-7.84238994e-01 -8.15182269e-01 -3.04178625e-01 3.03916186e-01
-1.00809515e+00 -5.45205593e-01 8.17737162e-01 -2.21193597e-01
7.11069345e-01 3.32408428e-01 6.66324139e-01 9.35419738e-01
5.87779880e-01 2.81980038e-02 1.42489064e+00 -6.79180920e-01
3.45526457e-01 5.90715647e-01 4.81499806e-02 4.73929942e-01
2.50410914e-01 5.61377704e-01 -4.16034549e-01 -3.80946770e-02
4.59887117e-01 5.29474020e-01 -9.12058055e-01 -2.06988856e-01
-9.89387274e-01 4.49973047e-01 2.29564980e-01 6.47191703e-01
-7.85459280e-02 1.06978267e-01 -4.25898045e-01 4.68236506e-01
2.76317388e-01 5.31763375e-01 1.40069067e-01 3.25521946e-01
-7.00766563e-01 -6.94610119e-01 6.85336649e-01 4.76251304e-01
8.68489683e-01 -1.61364973e-01 4.03234005e-01 6.57840014e-01
4.39357668e-01 1.39583790e+00 8.41172934e-02 -8.11323762e-01
1.85299933e-01 3.95210809e-04 1.62407205e-01 -8.08424473e-01
-5.11170447e-01 -5.74964046e-01 -9.01098907e-01 8.06427121e-01
3.16746801e-01 -3.13535601e-01 -8.69652450e-01 1.48230731e+00
3.46621335e-01 1.60824671e-01 3.40837151e-01 1.04009771e+00
8.70115221e-01 5.69614887e-01 -4.26028877e-01 -5.06955802e-01
1.40314090e+00 -3.80766332e-01 -6.63062990e-01 -2.55332202e-01
8.61573517e-02 -8.91096115e-01 5.05518019e-01 5.51623642e-01
-8.38419378e-01 -5.47409713e-01 -1.07048845e+00 1.88192636e-01
-3.00108314e-01 -1.49190798e-01 6.68213665e-01 1.05649245e+00
-8.68335962e-01 1.24483325e-01 -3.45060617e-01 -3.95609319e-01
2.16997162e-01 2.85116613e-01 -2.24268064e-01 -4.66216475e-01
-1.20313287e+00 8.83117020e-01 -2.82158762e-01 2.53732670e-02
5.39309271e-02 -8.29517663e-01 -3.94069463e-01 -4.94678058e-02
3.39336306e-01 -1.07351434e+00 5.44379175e-01 -7.78580427e-01
-1.88463950e+00 8.26228321e-01 -4.27068435e-02 -3.87646377e-01
3.69694173e-01 -7.47489780e-02 -7.90271580e-01 4.76142704e-01
-2.76907444e-01 3.07033416e-02 1.07438791e+00 -1.13577855e+00
-5.25063932e-01 -5.41688919e-01 3.18612128e-01 -1.55638546e-01
3.37195694e-02 -1.37777254e-01 1.36797994e-01 1.37967631e-01
9.88761485e-02 -1.17743337e+00 -1.41714558e-01 6.19879067e-02
-1.41182795e-01 3.36309701e-01 1.21222782e+00 1.70209214e-01
4.30154055e-01 -2.36518145e+00 -4.70384717e-01 1.44121408e-01
1.99831992e-01 3.32200021e-01 1.98077455e-01 5.77349842e-01
1.02521569e-01 -3.32788080e-01 4.23616692e-02 7.58333504e-02
-4.92636889e-01 -2.77579099e-01 -5.63124359e-01 8.17530990e-01
-6.82502031e-01 4.89324719e-01 -6.49613202e-01 -2.27663033e-02
6.59341991e-01 8.07002425e-01 -9.49491709e-02 -3.55367988e-01
1.45434409e-01 8.64669800e-01 -6.42438233e-01 4.71942723e-01
1.28650153e+00 -8.99923667e-02 -5.13082594e-02 -2.72214204e-01
-4.73280102e-01 -3.78167421e-01 -9.44454432e-01 1.53720617e+00
-9.10044491e-01 8.42822254e-01 3.36203337e-01 -8.36154401e-01
9.58711505e-01 2.49159962e-01 4.62849170e-01 -1.05858755e+00
5.26976027e-02 -1.12548165e-01 -4.23309386e-01 -7.20095575e-01
2.58960426e-01 -5.06543458e-01 3.06926191e-01 6.91530466e-01
-5.71653008e-01 -2.59552717e-01 -3.73057455e-01 -7.64732212e-02
1.24736845e+00 -1.33168712e-01 1.31101295e-01 2.80746110e-02
6.46234810e-01 -3.36608917e-01 1.21050261e-01 8.43987584e-01
3.45122129e-01 3.84284526e-01 -6.37754679e-01 -4.02181804e-01
-2.91246742e-01 -1.12738097e+00 -4.94083434e-01 5.06406248e-01
7.93081641e-01 8.77556652e-02 -2.77448356e-01 -9.33756232e-02
-5.05162142e-02 7.51255333e-01 -3.09763942e-02 -1.60120517e-01
-2.61236697e-01 -6.13657653e-01 3.72171253e-02 2.15670336e-02
1.06827414e+00 -8.80101085e-01 -1.11082661e+00 -9.45873708e-02
1.61495525e-02 -1.44096231e+00 2.89135695e-01 -9.85720158e-02
-5.05737305e-01 -1.04395211e+00 -6.34337902e-01 -2.74071068e-01
6.10306680e-01 1.04608333e+00 9.04987693e-01 5.92857786e-02
-7.78523386e-01 1.11193645e+00 -2.74432719e-01 -6.85373724e-01
-1.85154840e-01 -6.54423058e-01 1.68313637e-01 2.66090631e-01
4.44903284e-01 -8.76686394e-01 -9.66046572e-01 1.98001489e-01
-5.93912780e-01 -7.52285272e-02 8.00954521e-01 4.19815063e-01
3.44965786e-01 3.69575799e-01 2.56364554e-01 -8.98723006e-01
-6.44003078e-02 -3.22759211e-01 -9.43974257e-01 1.63730428e-01
-4.43868995e-01 -3.40474725e-01 8.20041120e-01 -1.88210040e-01
-1.43253219e+00 1.13322221e-01 1.76488072e-01 -7.54902959e-02
-3.71418148e-01 -1.33409221e-02 -1.35770813e-01 -5.25720417e-01
6.37342513e-01 1.87706441e-01 -1.74732044e-01 -1.33160397e-01
3.01120758e-01 8.97417426e-01 3.48065704e-01 1.17235459e-01
1.04205334e+00 1.32350218e+00 6.21630967e-01 -1.60745740e+00
-5.17242670e-01 -5.15011072e-01 -1.54238388e-01 -4.60782886e-01
1.09063065e+00 -7.45672047e-01 -1.07206738e+00 1.86031997e-01
-8.73195887e-01 2.00751722e-01 -3.89809430e-01 1.13191962e+00
-4.31929022e-01 4.00430650e-01 -2.59343117e-01 -1.19830728e+00
-2.81308562e-01 -7.11697280e-01 8.12017322e-01 5.45892477e-01
1.77720755e-01 -7.25023746e-01 1.36859581e-01 6.13396287e-01
5.58213949e-01 7.04564303e-02 6.53771639e-01 5.95352530e-01
-9.28197205e-01 -1.01064369e-01 -5.25339991e-02 2.78935358e-02
1.02387413e-01 -4.06914860e-01 -1.41030478e+00 -2.82637805e-01
6.22479260e-01 1.20519370e-01 1.21913218e+00 7.19673574e-01
9.13409412e-01 3.23234409e-01 -8.81880045e-01 9.78799462e-01
1.79864633e+00 5.63594520e-01 9.02798712e-01 2.23870426e-01
3.20182234e-01 5.21068275e-01 5.37453294e-01 2.21218109e-01
-3.84711027e-01 5.81470251e-01 4.91691679e-01 -4.71504867e-01
-8.09161961e-02 4.61121768e-01 3.33662331e-01 8.64979550e-02
-4.29423928e-01 -4.18653816e-01 -2.80725092e-01 -1.43494472e-01
-1.27870691e+00 -1.51528144e+00 -2.98687935e-01 2.43547702e+00
1.36123419e-01 -3.74340773e-01 -5.09963036e-01 -1.84361190e-02
7.09065378e-01 -6.35699183e-03 -4.27237719e-01 -1.32082090e-01
-1.81516975e-01 4.40858692e-01 8.46218526e-01 4.55279231e-01
-7.86449969e-01 2.52648383e-01 5.87520790e+00 3.51504982e-01
-1.49936473e+00 7.94655979e-02 -1.17705084e-01 -1.48397267e-01
-5.85951507e-01 1.37788892e-01 -6.28610253e-01 5.47716498e-01
3.68548810e-01 2.00469777e-01 4.44200873e-01 5.21608829e-01
1.24825694e-01 -7.35808849e-01 -9.13798511e-01 1.54597938e+00
-9.01042894e-02 -1.09731829e+00 -3.71458918e-01 4.53667372e-01
3.98773640e-01 1.39773712e-01 4.83021498e-01 -3.87402564e-01
-6.89232443e-03 -8.16296637e-01 1.51644588e-01 5.99644184e-01
1.09886432e+00 -5.90770602e-01 4.98017043e-01 4.39441144e-01
-8.43821287e-01 -1.79131910e-01 -7.44891226e-01 -1.49648592e-01
2.67613679e-01 1.31895661e+00 -4.28348124e-01 6.49080992e-01
8.96708012e-01 4.20425296e-01 1.36083454e-01 6.82799935e-01
-1.00238606e-01 2.72406638e-01 -4.14497435e-01 -2.58579180e-02
-2.51953423e-01 -8.85464728e-01 8.03039670e-01 8.50798190e-01
6.86802864e-01 4.46681499e-01 -2.87327007e-03 1.08349252e+00
2.04841137e-01 -4.22320545e-01 -1.22298658e+00 2.67024696e-01
2.94631183e-01 1.44783592e+00 -8.02346230e-01 6.25128448e-02
-8.64364982e-01 8.46090138e-01 -6.72281086e-01 6.71157658e-01
-6.32080078e-01 -5.56011975e-01 6.21018231e-01 3.03134948e-01
1.87762409e-01 -4.99160677e-01 2.01344192e-01 -1.08738446e+00
-2.16438826e-02 -2.91797426e-02 -8.93400833e-02 -1.10197425e+00
-1.10134387e+00 3.83393466e-01 -3.11903387e-01 -1.42008877e+00
2.12096259e-01 -4.84561741e-01 -3.78216445e-01 8.43898535e-01
-1.91574824e+00 -9.26798820e-01 -6.00528181e-01 9.47890341e-01
1.06796317e-01 3.31426896e-02 7.93985426e-01 1.16127871e-01
3.03897381e-01 7.88429826e-02 4.80257899e-01 -1.63664758e-01
8.30576897e-01 -7.05340743e-01 -1.96082562e-01 7.02836156e-01
3.52732629e-01 7.87951052e-01 5.10544538e-01 -3.87807548e-01
-2.00342131e+00 -6.43445969e-01 6.51318878e-02 -1.83739960e-01
1.06632374e-01 -3.56043547e-01 -1.89324886e-01 3.28695118e-01
3.80064487e-01 1.27550140e-01 9.45280790e-01 -1.56661883e-01
-2.39660650e-01 -3.96131784e-01 -1.60824156e+00 2.61677921e-01
1.20858896e+00 -6.48874879e-01 -3.58062804e-01 5.16798019e-01
3.07045668e-01 -9.31364670e-02 -3.37564379e-01 1.99264765e-01
8.37605059e-01 -1.24003792e+00 1.16101170e+00 2.61002749e-01
-1.05057687e-01 -4.98040080e-01 -2.98604310e-01 -1.15088904e+00
-3.19858640e-01 -6.80122495e-01 4.66046870e-01 8.84359300e-01
9.14355516e-02 -1.43836033e+00 7.35891521e-01 3.25099975e-01
2.43635461e-01 7.01479893e-03 -8.45131814e-01 -8.80941451e-01
-5.57823181e-01 -4.87995833e-01 3.87883067e-01 9.27923381e-01
-2.00737640e-01 3.97069365e-01 -3.39942545e-01 4.13509607e-01
1.05530083e+00 5.66240549e-01 5.88423193e-01 -1.47786951e+00
-5.84653676e-01 2.35367075e-01 -4.69608158e-01 -1.06685138e+00
8.12281147e-02 -9.80510235e-01 -2.91420132e-01 -1.51473331e+00
8.58840123e-02 -5.28036296e-01 -1.40560186e-02 -1.72685191e-01
6.04653358e-01 5.56590438e-01 3.24446619e-01 2.76671648e-01
-9.37489122e-02 1.15787745e-01 1.22884202e+00 -2.59986907e-01
-2.42468998e-01 4.45331007e-01 -3.92286062e-01 7.66177475e-01
5.94285011e-01 -2.50959784e-01 -5.88964343e-01 -3.82144332e-01
6.38974428e-01 8.13030303e-02 5.14641702e-01 -1.31844497e+00
5.27807176e-01 -1.98931798e-01 6.99109614e-01 -2.94021368e-01
6.85032487e-01 -1.38291419e+00 5.64817309e-01 4.86879319e-01
3.12271506e-01 -9.00359511e-01 -3.58869761e-01 8.90478253e-01
1.25125408e-01 -2.69918919e-01 1.04060078e+00 -5.89505672e-01
-6.65677547e-01 1.58871152e-03 -6.51380479e-01 -3.93454820e-01
1.14734185e+00 -7.02990890e-01 -9.20645475e-01 -5.52315354e-01
-3.93319428e-01 -3.17383766e-01 6.03537798e-01 -2.57417224e-02
6.02748752e-01 -7.09415019e-01 -1.44825026e-01 3.60537440e-01
-2.05042526e-01 -4.35363829e-01 4.65316325e-01 8.29175532e-01
-3.02642852e-01 4.77913290e-01 -1.21000260e-01 -6.03848636e-01
-1.62712967e+00 6.57176554e-01 2.63306350e-01 4.90239352e-01
-1.17017591e+00 6.20487034e-01 6.80373192e-01 1.55500337e-01
-2.18932465e-01 -7.13041127e-02 -3.38713020e-01 -4.76272196e-01
9.50555921e-01 3.76213223e-01 -1.00816628e-02 -2.73910344e-01
-4.86879915e-01 1.46802950e+00 4.91686732e-01 -5.59520870e-02
1.35848665e+00 -4.85552967e-01 -1.85992286e-01 7.09924698e-02
8.60098362e-01 7.35782325e-01 -6.15079701e-01 -7.19635561e-02
-7.85317779e-01 -6.91305101e-01 4.19501066e-01 -6.75701320e-01
-1.17720366e+00 8.95161390e-01 8.57561171e-01 4.49774176e-01
1.47332907e+00 1.76649481e-01 8.15885186e-01 7.54444480e-01
1.08112013e+00 -1.15733290e+00 -3.98949504e-01 7.21912906e-02
2.55850405e-01 -9.48117077e-01 2.63034821e-01 -6.46663487e-01
-3.09299538e-03 1.42075241e+00 -5.24823107e-02 1.14197373e-01
9.05925274e-01 6.78829551e-01 2.70181596e-01 -2.06650808e-01
-2.56550431e-01 -2.02042490e-01 -2.75575757e-01 1.00674570e+00
2.07114499e-02 -1.13214985e-01 -6.97543249e-02 3.62500660e-02
-1.32456943e-01 1.70445532e-01 8.45663905e-01 9.38396215e-01
-6.82626486e-01 -9.33080137e-01 -7.18553126e-01 2.58459985e-01
-6.18561804e-01 2.09806100e-01 -2.43471667e-01 4.69382286e-01
2.47969493e-01 1.44345820e+00 -1.01888865e-01 -2.43246511e-01
9.84790642e-03 -2.53493637e-01 8.15462470e-01 -4.95693624e-01
-4.66267243e-02 -1.84569195e-01 -2.66182005e-01 -6.53381586e-01
-9.99533355e-01 -2.75269032e-01 -1.36080599e+00 -2.73553461e-01
-7.11926222e-01 1.54171064e-02 1.14906120e+00 8.36026371e-01
3.94116402e-01 1.59177884e-01 8.07595253e-01 -5.04146695e-01
6.37189066e-03 -3.16921473e-01 -1.08935583e+00 -3.66285853e-02
5.21581233e-01 -7.58064568e-01 -7.98713326e-01 -2.19329044e-01] | [10.118849754333496, -2.643252372741699] |
82b2a0aa-5795-47ee-bee2-a6f8e7dbc389 | inverse-path-tracing-for-joint-material-and-1 | null | null | http://openaccess.thecvf.com/content_CVPR_2019/html/Azinovic_Inverse_Path_Tracing_for_Joint_Material_and_Lighting_Estimation_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Azinovic_Inverse_Path_Tracing_for_Joint_Material_and_Lighting_Estimation_CVPR_2019_paper.pdf | Inverse Path Tracing for Joint Material and Lighting Estimation | Modern computer vision algorithms have brought significant advancement to 3D geometry reconstruction. However, illumination and material reconstruction remain less studied, with current approaches assuming very simplified models for materials and illumination. We introduce Inverse Path Tracing, a novel approach to jointly estimate the material properties of objects and light sources in indoor scenes by using an invertible light transport simulation. We assume a coarse geometry scan, along with corresponding images and camera poses. The key contribution of this work is an accurate and simultaneous retrieval of light sources and physically based material properties (e.g., diffuse reflectance, specular reflectance, roughness, etc.) for the purpose of editing and re-rendering the scene under new conditions. To this end, we introduce a novel optimization method using a differentiable Monte Carlo renderer that computes derivatives with respect to the estimated unknown illumination and material properties. This enables joint optimization for physically correct light transport and material models using a tailored stochastic gradient descent.
| [' Matthias Niessner', ' Anton Kaplanyan', ' Tzu-Mao Li', 'Dejan Azinovic'] | 2019-06-01 | null | null | null | cvpr-2019-6 | ['lighting-estimation'] | ['computer-vision'] | [ 6.27840757e-01 -6.54314280e-01 8.01988542e-01 -3.89941752e-01
-4.45283502e-01 -4.66484487e-01 6.56253040e-01 1.05846375e-02
-1.88175544e-01 8.11500072e-01 -2.04455405e-01 1.94080211e-02
-2.14209035e-01 -8.84941876e-01 -6.39916003e-01 -8.21315825e-01
4.27099109e-01 6.78011894e-01 -7.37121701e-02 1.12557001e-02
3.43924791e-01 1.04418182e+00 -1.78167558e+00 -1.55924857e-01
8.89795542e-01 6.18032098e-01 2.60587662e-01 9.58205283e-01
-2.33255565e-01 4.42200363e-01 5.08146873e-03 -2.09710836e-01
2.87058383e-01 -2.56885946e-01 -4.43320036e-01 4.26046103e-01
5.35077512e-01 -5.71801782e-01 -5.97140454e-02 9.34623659e-01
3.42366427e-01 4.55490023e-01 8.70978236e-01 -7.01795995e-01
-3.82243067e-01 -3.13974977e-01 -6.75555348e-01 -4.38836843e-01
7.38472581e-01 2.00435266e-01 5.34786582e-01 -9.67486799e-01
5.28032124e-01 1.05100131e+00 6.65307581e-01 2.93275893e-01
-1.44832301e+00 -6.24327436e-02 2.72417217e-02 1.10213026e-01
-1.52777350e+00 -4.10075784e-01 1.12671912e+00 -4.51350480e-01
7.27182806e-01 4.06928033e-01 9.75642264e-01 6.07880950e-01
2.40756661e-01 8.87496024e-02 1.57480800e+00 -7.06225634e-01
3.97848934e-01 5.20538807e-01 -2.35877875e-02 8.43565941e-01
1.38060868e-01 2.13924751e-01 -4.99587208e-01 -2.53269881e-01
9.30000842e-01 5.99200875e-02 -3.64700645e-01 -5.70145071e-01
-1.13263798e+00 2.47976616e-01 1.64208189e-01 -1.86417639e-01
-7.52325237e-01 1.91976696e-01 -3.73553395e-01 3.90232056e-02
8.52574825e-01 1.57546267e-01 -2.78853565e-01 2.44754389e-01
-5.69409311e-01 2.57738739e-01 8.94944608e-01 6.94247067e-01
1.29059267e+00 2.00434029e-02 9.75905806e-02 7.51617908e-01
7.90831864e-01 1.30518591e+00 -6.26967490e-01 -1.34068751e+00
-5.03985360e-02 1.59417257e-01 6.36171699e-01 -9.09310341e-01
-1.71947017e-01 -3.85749727e-01 -4.63583559e-01 6.00084245e-01
3.67823094e-01 1.08665526e-01 -6.84468806e-01 1.34649253e+00
9.34372842e-01 4.19404715e-01 1.42802659e-03 8.25218320e-01
5.13612807e-01 6.53056085e-01 -2.94399798e-01 -3.42670053e-01
9.74335551e-01 -6.11263156e-01 -5.01223862e-01 8.49368423e-02
5.29398024e-02 -1.13708222e+00 7.94762611e-01 6.36882782e-01
-1.29219759e+00 -1.08921267e-01 -7.02818334e-01 -2.29891166e-01
-5.32820746e-02 1.96507171e-01 5.68293631e-01 6.57022595e-01
-8.35721433e-01 6.99534178e-01 -7.15560019e-01 -1.17362201e-01
1.15134805e-01 -3.86462696e-02 7.75834545e-02 -4.33634132e-01
-5.27336597e-01 7.90004611e-01 -4.78947639e-01 2.84726977e-01
-8.83517981e-01 -9.61663246e-01 -5.48519254e-01 -3.93513113e-01
2.23538563e-01 -1.11707747e+00 8.68005097e-01 -6.71254754e-01
-2.25254560e+00 9.42153871e-01 -3.25746477e-01 1.45414442e-01
6.43388987e-01 -3.54776591e-01 4.37631160e-02 9.00172666e-02
-1.85149550e-01 -5.07598510e-03 1.05230236e+00 -1.94568086e+00
-1.16945073e-01 -3.34724844e-01 4.05336395e-02 5.12043655e-01
3.57129663e-01 -2.54025817e-01 -3.74375850e-01 -3.14423703e-02
3.49095553e-01 -6.70242727e-01 -3.15212965e-01 6.44392073e-01
-4.17485863e-01 6.05526805e-01 2.88877934e-01 -7.78491318e-01
2.86176741e-01 -1.84900641e+00 1.87623993e-01 4.19673890e-01
-1.26113920e-02 -1.86741084e-01 -9.87143815e-02 4.73415285e-01
2.10041836e-01 -5.82981765e-01 -5.00111997e-01 -8.55491340e-01
-2.41896808e-01 1.09410077e-01 -2.59067059e-01 8.10019433e-01
-2.01287493e-01 5.34316063e-01 -9.45578754e-01 -1.99956909e-01
8.85933638e-01 1.29463756e+00 -4.43496794e-01 1.57312617e-01
-4.22729701e-01 1.08315241e+00 -5.17670512e-01 5.58703244e-01
1.19280648e+00 1.67136818e-01 -1.41483381e-01 -5.15933752e-01
-6.78756893e-01 -6.35560304e-02 -1.53384650e+00 1.70084119e+00
-1.15234327e+00 3.77659351e-01 7.24142313e-01 -3.67908001e-01
8.57347786e-01 9.29230303e-02 6.27029896e-01 -4.83871907e-01
2.79631317e-01 3.14478278e-01 -7.75661469e-01 -3.28527480e-01
4.45879042e-01 -3.90587628e-01 8.44007194e-01 6.07011139e-01
-6.69144511e-01 -1.14661837e+00 -4.22766328e-01 -3.62641141e-02
5.86505175e-01 7.10695148e-01 -7.75884762e-02 -1.61604553e-01
7.71754146e-01 1.80009287e-02 9.10235792e-02 5.88078082e-01
4.84761149e-01 5.77001989e-01 -3.82047445e-01 -3.44980896e-01
-1.09431791e+00 -1.21844673e+00 -2.23357394e-01 6.23528063e-01
2.76305079e-01 3.31138879e-01 -8.62302721e-01 1.68264925e-01
9.87309217e-02 1.06692851e+00 -2.21034542e-01 1.41623244e-01
-5.66097021e-01 -8.08553040e-01 -3.39032471e-01 -2.49957934e-01
5.15259027e-01 -5.97256899e-01 -6.52364492e-01 1.87128723e-01
-3.50564010e-02 -1.20107913e+00 2.93308403e-04 -4.48371768e-01
-8.77778888e-01 -1.11595118e+00 -6.88384831e-01 -1.07424095e-01
9.95676219e-01 6.15324974e-01 1.26831198e+00 1.35471389e-01
-7.81282961e-01 1.23261118e+00 -6.91633970e-02 -3.41345996e-01
-4.43837404e-01 -6.81209922e-01 -1.59001708e-01 6.86520815e-01
-3.97038698e-01 -6.90285921e-01 -7.43295610e-01 2.97644138e-01
-6.55669868e-01 6.35709167e-01 1.12287790e-01 1.26324892e-01
1.10348284e+00 2.04264577e-02 -4.33583170e-01 -9.03081298e-01
3.06893736e-01 -1.65205210e-01 -1.04862309e+00 2.78599709e-01
-5.15543342e-01 -1.67400375e-01 3.90466779e-01 -1.07412316e-01
-1.79298079e+00 5.44669293e-02 -7.68235028e-02 -2.30268270e-01
-3.23126256e-01 -9.60639492e-02 -1.66036218e-01 -5.74571967e-01
5.18861830e-01 2.38225520e-01 -3.72765839e-01 -6.82441890e-01
5.23424506e-01 1.73355177e-01 2.98498362e-01 -8.11143994e-01
9.39129531e-01 1.34096503e+00 7.32940495e-01 -1.29142344e+00
-8.02882731e-01 -4.40109611e-01 -5.76918185e-01 -5.13589680e-01
5.59333742e-01 -6.59423172e-01 -7.71814525e-01 7.38501966e-01
-1.16738319e+00 -4.92340863e-01 -5.45804620e-01 7.29603648e-01
-6.80610359e-01 4.99307245e-01 -2.44111151e-01 -1.12662482e+00
-2.28288502e-01 -1.01969326e+00 1.34177148e+00 3.89449671e-02
2.59749919e-01 -1.24392664e+00 3.30700725e-01 4.46527809e-01
3.95995975e-01 3.84135693e-01 6.13931596e-01 9.37948406e-01
-1.29185104e+00 1.86590292e-02 -4.39158976e-01 3.76542836e-01
3.24502438e-01 3.97554785e-01 -1.22225046e+00 -1.51435539e-01
3.82714331e-01 2.47941211e-01 5.91582298e-01 6.03899956e-01
8.94395769e-01 1.76651731e-01 -2.12692872e-01 1.10725284e+00
2.01447749e+00 -1.50696844e-01 3.94362748e-01 -2.08232906e-02
9.77772832e-01 7.99803376e-01 5.15348017e-01 7.60643601e-01
2.71172225e-01 7.31384337e-01 6.85899854e-01 -5.45070916e-02
-5.74074328e-01 2.18403205e-01 -1.51027907e-02 6.03353679e-01
-6.91582561e-01 -2.44923636e-01 -6.47579908e-01 2.03894511e-01
-1.17059910e+00 -6.59431875e-01 -7.85006106e-01 2.60071778e+00
5.81042767e-01 -4.88504916e-01 -6.10392570e-01 -1.38507470e-01
3.56097937e-01 -1.35599479e-01 -5.46728790e-01 -1.67677790e-01
-1.54458433e-01 3.60459685e-01 7.61635542e-01 1.22773778e+00
-4.33097869e-01 5.49868882e-01 6.04721165e+00 3.85710806e-01
-9.19637740e-01 1.96388707e-01 -1.42008914e-02 8.11088681e-02
-1.03819406e+00 2.69319922e-01 -6.49010837e-01 6.70803636e-02
4.41573173e-01 2.73933828e-01 1.14923525e+00 1.63737655e-01
6.13192320e-01 -5.74999332e-01 -8.66601944e-01 1.07996273e+00
1.30005538e-01 -9.97411370e-01 4.08382863e-02 -4.61707413e-02
1.05195498e+00 -2.89201178e-02 -2.95969122e-03 -5.95696092e-01
2.49293581e-01 -3.12630922e-01 7.21500576e-01 1.32994199e+00
7.21912622e-01 -3.98745477e-01 1.67705283e-01 2.91403055e-01
-7.82470703e-01 4.93045002e-01 -1.74440354e-01 7.97374099e-02
6.07271135e-01 1.17483366e+00 -5.23967445e-01 6.62503481e-01
3.39145213e-01 6.15776360e-01 -3.46963406e-02 1.03542793e+00
-4.77268040e-01 3.10161620e-01 -6.26763284e-01 2.37609163e-01
-3.59035462e-01 -1.02520978e+00 9.03512239e-01 8.55306208e-01
3.66161317e-01 1.49596378e-01 2.79935487e-02 1.27388489e+00
2.26664215e-01 2.03036487e-01 -3.48207146e-01 4.99860257e-01
-7.27728233e-02 1.35417736e+00 -7.49169469e-01 -2.57760976e-02
-3.16862941e-01 1.12156510e+00 -5.46058752e-02 8.71360719e-01
-6.23050272e-01 1.13821067e-01 5.58630645e-01 3.10983032e-01
-3.26583952e-01 -6.21973991e-01 -3.14001918e-01 -1.17503726e+00
1.06360791e-02 -7.77966008e-02 -4.89907980e-01 -1.15379655e+00
-1.18600190e+00 -1.01325482e-01 1.41213620e-02 -8.25511754e-01
2.89269179e-01 -6.78693533e-01 -4.04541820e-01 1.24075818e+00
-2.03058600e+00 -1.23142815e+00 -5.58916092e-01 7.24211335e-01
3.44173610e-01 4.74986345e-01 7.42375672e-01 2.12495327e-01
-1.94060177e-01 -3.46206814e-01 7.28706717e-01 -6.55939221e-01
3.90081853e-01 -1.07106543e+00 1.74008086e-01 5.88463128e-01
-1.73202798e-01 4.37535375e-01 9.08196270e-01 -5.94111145e-01
-1.87036228e+00 -8.46026599e-01 3.01485896e-01 -4.08490598e-01
2.71200895e-01 -1.49332121e-01 -6.28976583e-01 3.48101854e-01
-1.22185282e-01 5.38105592e-02 2.85423428e-01 -3.15132141e-01
8.03962201e-02 -2.61175632e-01 -1.39201117e+00 5.07546425e-01
1.08278322e+00 -6.20665193e-01 9.45901126e-02 7.07991242e-01
2.45411262e-01 -5.56398094e-01 -5.16607642e-01 2.35712171e-01
6.17247701e-01 -1.17269444e+00 1.42707181e+00 1.71747297e-01
-6.64497241e-02 -4.80761111e-01 -2.99434066e-01 -1.27905953e+00
-6.59556165e-02 -6.51539743e-01 -1.82085298e-02 1.06060004e+00
9.23518091e-03 -8.73850107e-01 6.41210616e-01 9.38980997e-01
-2.54579306e-01 -2.76058346e-01 -5.75516284e-01 -5.07372499e-01
-4.43312138e-01 -7.22237408e-01 5.01759768e-01 7.17663765e-01
-1.16536379e+00 -6.55555874e-02 -4.05816913e-01 5.70525825e-01
1.47037637e+00 4.60758120e-01 8.08379471e-01 -1.38523650e+00
-3.42622966e-01 -2.02226453e-02 3.51516932e-01 -9.68244255e-01
1.52984500e-01 -6.47267938e-01 1.73439190e-01 -1.76516449e+00
6.93767294e-02 -8.48462462e-01 2.95329630e-01 -3.52048576e-01
1.52568728e-01 2.64274806e-01 -3.13508421e-01 1.75139681e-01
-6.10655956e-02 6.34631991e-01 1.57707822e+00 3.98772024e-02
-4.67824548e-01 3.00732017e-01 -5.72285280e-02 9.21055257e-01
6.92289650e-01 -2.94411927e-01 -4.00022179e-01 -9.47561324e-01
7.33568132e-01 -1.70669779e-01 8.55524182e-01 -8.40837538e-01
-5.88314943e-02 -4.51487094e-01 1.56455100e-01 -5.59150577e-01
7.91808248e-01 -1.17386937e+00 6.98166132e-01 1.43703058e-01
-4.36892882e-02 -6.68613553e-01 9.96436551e-02 6.47203982e-01
4.51521188e-01 -4.92336303e-01 8.15031469e-01 -3.82882565e-01
-4.31020111e-01 3.91829699e-01 -1.97858065e-01 -3.06357175e-01
7.09048331e-01 -3.99239689e-01 6.49500787e-02 -1.83922023e-01
-5.93376994e-01 -3.41550887e-01 9.14347291e-01 -3.23797315e-01
7.97863483e-01 -1.06219411e+00 -7.86241889e-01 2.91515976e-01
-2.30978832e-01 7.65812993e-02 4.53751951e-01 7.31524944e-01
-1.17613685e+00 -7.89861307e-02 2.28135303e-01 -5.68518162e-01
-1.25361538e+00 1.53921142e-01 5.86331427e-01 2.04007477e-01
-6.15648091e-01 7.67691791e-01 1.90294415e-01 -5.83525360e-01
-3.20989430e-01 -2.37087354e-01 2.30235949e-01 -3.66661429e-01
2.78073221e-01 9.46048439e-01 1.92822590e-01 -7.74612188e-01
-1.84557706e-01 1.42564094e+00 5.20913422e-01 -2.06470221e-01
1.35030270e+00 -7.16569364e-01 -4.59855109e-01 5.48435569e-01
9.55400586e-01 5.28617918e-01 -1.41494119e+00 -3.45045120e-01
-9.58486617e-01 -9.64239359e-01 6.00141823e-01 -6.03476405e-01
-8.27419043e-01 8.35201025e-01 2.83368021e-01 -2.32618541e-01
1.00221264e+00 -2.35540912e-01 5.61398983e-01 3.62835437e-01
7.46579587e-01 -1.03516424e+00 -4.52910304e-01 3.71321976e-01
7.43588030e-01 -1.02078438e+00 5.65074503e-01 -9.75247085e-01
1.30603582e-01 1.06009936e+00 -9.70803052e-02 -6.29476160e-02
8.35379481e-01 6.77830502e-02 3.91687602e-02 -3.50265503e-01
-1.45513266e-01 -6.00885004e-02 3.01600605e-01 6.77886188e-01
2.04144359e-01 8.05287361e-02 9.02868286e-02 -7.26152003e-01
2.19708756e-01 -4.74516153e-02 3.35630864e-01 7.44714260e-01
-2.84792483e-01 -9.26464736e-01 -7.13175893e-01 9.40840170e-02
1.10591454e-02 -1.51725575e-01 -3.00804973e-02 1.24868169e-01
1.46254554e-01 8.18208575e-01 -8.09842497e-02 4.45722252e-01
4.19334531e-01 -3.62370461e-01 9.64066863e-01 -7.39459932e-01
-6.87214062e-02 1.16811149e-01 -5.45908622e-02 -5.46064258e-01
-8.19265664e-01 -9.17519987e-01 -1.03146374e+00 -1.54704198e-01
-4.51571405e-01 -1.62342072e-01 1.36871338e+00 6.85925424e-01
-2.53977384e-02 3.51355582e-01 7.66878188e-01 -1.23084474e+00
-3.93161736e-02 -1.98524460e-01 -8.33716571e-01 3.33729625e-01
4.75172877e-01 -7.91073263e-01 -5.54332137e-01 1.93066567e-01] | [9.734335899353027, -3.0745701789855957] |
7694a274-1759-4db4-b359-a307889bb72f | spatial-reasoning-for-few-shot-object | 2211.01080 | null | https://arxiv.org/abs/2211.01080v1 | https://arxiv.org/pdf/2211.01080v1.pdf | Spatial Reasoning for Few-Shot Object Detection | Although modern object detectors rely heavily on a significant amount of training data, humans can easily detect novel objects using a few training examples. The mechanism of the human visual system is to interpret spatial relationships among various objects and this process enables us to exploit contextual information by considering the co-occurrence of objects. Thus, we propose a spatial reasoning framework that detects novel objects with only a few training examples in a context. We infer geometric relatedness between novel and base RoIs (Region-of-Interests) to enhance the feature representation of novel categories using an object detector well trained on base categories. We employ a graph convolutional network as the RoIs and their relatedness are defined as nodes and edges, respectively. Furthermore, we present spatial data augmentation to overcome the few-shot environment where all objects and bounding boxes in an image are resized randomly. Using the PASCAL VOC and MS COCO datasets, we demonstrate that the proposed method significantly outperforms the state-of-the-art methods and verify its efficacy through extensive ablation studies. | ['Seong-Whan Lee', 'Hong-Gyu Jung', 'Geonuk Kim'] | 2022-11-02 | null | null | null | null | ['few-shot-object-detection'] | ['computer-vision'] | [ 1.87926769e-01 4.02711285e-03 1.99895844e-01 -5.62252462e-01
-1.10173680e-01 -6.44944429e-01 8.62824857e-01 4.74456489e-01
-7.35101104e-01 3.76214027e-01 -1.28432885e-01 1.14235006e-01
-8.31531435e-02 -8.10114086e-01 -9.17100191e-01 -3.59519690e-01
-1.82854369e-01 -3.24630775e-02 8.57430279e-01 -1.89733610e-01
4.18213785e-01 9.07364368e-01 -1.75254214e+00 4.06465113e-01
8.97216499e-01 8.71081114e-01 5.56687355e-01 5.03218293e-01
-3.88179440e-03 5.79974949e-01 -6.52404308e-01 -1.48716807e-01
3.39101017e-01 -3.35169911e-01 -5.71266115e-01 1.25856265e-01
5.67805946e-01 -2.47464567e-01 -6.09869063e-01 1.14896131e+00
2.27118209e-01 6.99728727e-01 7.09690809e-01 -1.17283237e+00
-9.72477078e-01 2.97078401e-01 -7.31012225e-01 7.01941907e-01
2.97434419e-01 2.81024098e-01 8.06712627e-01 -1.35360897e+00
7.89794326e-01 1.07331955e+00 4.13557619e-01 3.06415439e-01
-1.18572640e+00 -3.98547530e-01 4.74156946e-01 5.84969521e-01
-1.71919608e+00 -2.71129221e-01 9.49286342e-01 -2.79057175e-01
9.95744228e-01 1.98302999e-01 7.05660820e-01 8.52877915e-01
-8.66875518e-03 7.64584541e-01 8.50378335e-01 -5.89125514e-01
4.77280349e-01 1.56446770e-01 1.12482466e-01 6.43311441e-01
2.10498154e-01 2.40294710e-01 -4.91408557e-01 1.83945522e-01
6.84022665e-01 5.30457795e-01 -5.15217632e-02 -6.52660966e-01
-1.17783010e+00 5.92379391e-01 1.36382437e+00 4.80917871e-01
-2.88061202e-01 2.06556097e-01 6.77471235e-02 -3.07551354e-01
9.04499888e-02 5.34303844e-01 3.41091007e-02 4.82850373e-01
-8.14265668e-01 1.43495694e-01 6.32755309e-02 1.28229928e+00
9.12012875e-01 -1.32739723e-01 -4.57857072e-01 6.43960893e-01
3.16387750e-02 2.64228135e-01 3.36033046e-01 -5.72124362e-01
2.93487549e-01 9.69655097e-01 -1.48865670e-01 -1.31996429e+00
-4.11697239e-01 -6.46614850e-01 -6.03209078e-01 2.74072528e-01
2.44814575e-01 3.10873508e-01 -1.18336213e+00 1.67519665e+00
5.08474171e-01 3.36127907e-01 -1.74245164e-01 9.77248490e-01
8.02891970e-01 2.97786236e-01 2.32177079e-01 4.47858125e-02
1.17232978e+00 -1.09561563e+00 -3.91524523e-01 -6.69624686e-01
4.12347347e-01 -3.36419731e-01 1.15621555e+00 -5.63261844e-02
-8.69319201e-01 -8.22488427e-01 -9.88272846e-01 -2.97964245e-01
-8.43471110e-01 1.54047921e-01 6.55856490e-01 1.88809022e-01
-7.46947169e-01 5.61718702e-01 -6.19392037e-01 -6.21972322e-01
1.01320541e+00 7.08571449e-02 -4.29727912e-01 -2.77164131e-01
-7.28446484e-01 8.36859643e-01 8.17545176e-01 2.62226164e-01
-1.10579646e+00 -5.07431328e-01 -9.36857641e-01 1.43873915e-01
4.42951262e-01 -6.51155949e-01 6.93998754e-01 -8.51477861e-01
-6.29424989e-01 8.60607266e-01 -2.74036050e-01 -3.49181890e-01
2.61977196e-01 -1.82591900e-01 -4.59864676e-01 3.18539411e-01
3.37115794e-01 9.21137750e-01 9.62172091e-01 -1.34435284e+00
-8.76801908e-01 -4.96007234e-01 2.68511325e-01 1.85781673e-01
-2.08300412e-01 -1.46189436e-01 -4.09123987e-01 -7.91396916e-01
5.37517428e-01 -7.94088304e-01 -4.01866972e-01 2.96339691e-01
-4.69161540e-01 -2.59720743e-01 1.00276136e+00 -3.02955747e-01
1.00143218e+00 -2.38454247e+00 -6.81767017e-02 1.89112559e-01
4.55821007e-01 1.26730725e-01 -2.39803225e-01 2.91259568e-02
-1.47018090e-01 2.44342145e-02 -2.23109022e-01 -5.59226833e-02
-1.49231955e-01 1.30144283e-01 -4.21738952e-01 3.83178174e-01
4.41187561e-01 1.13817704e+00 -1.24081290e+00 -4.16520506e-01
5.03933609e-01 2.76118636e-01 -4.91957515e-01 2.02646583e-01
-1.83026582e-01 5.10896087e-01 -4.77723658e-01 6.85546517e-01
8.25485229e-01 -2.16952294e-01 -1.17918558e-01 -3.22972476e-01
5.46394363e-02 -4.80834618e-02 -1.04451227e+00 1.98717034e+00
-3.00799254e-02 6.61208630e-01 -5.27356207e-01 -1.00301385e+00
8.62525582e-01 -3.64358127e-01 -1.37472346e-01 -8.68047476e-01
1.04664825e-01 9.47303046e-03 1.46167874e-01 -6.48684800e-01
4.20905799e-01 1.25874743e-01 5.90341985e-02 2.58116096e-01
9.82076898e-02 5.45804426e-02 3.15814912e-01 5.16660333e-01
1.24053264e+00 7.52673894e-02 4.97685969e-01 -1.58119693e-01
3.73265117e-01 -7.48941228e-02 5.23132086e-01 1.12557340e+00
-4.32621092e-01 6.76238537e-01 2.73693025e-01 -4.93563265e-01
-8.45579028e-01 -1.50429678e+00 -6.28105029e-02 1.34230232e+00
6.05829716e-01 -3.09954882e-01 -5.66244841e-01 -1.00899851e+00
4.05174233e-02 9.84311700e-01 -1.14069879e+00 -2.89129823e-01
-5.23380995e-01 -2.67482877e-01 1.97246119e-01 1.05358088e+00
6.79385722e-01 -1.37145877e+00 -9.93004858e-01 -5.04017957e-02
3.31092566e-01 -1.24441051e+00 -3.95179480e-01 1.35502487e-01
-7.19003975e-01 -1.20922327e+00 -4.16658342e-01 -8.81917119e-01
1.15814817e+00 6.63884521e-01 1.02736211e+00 8.06843638e-02
-7.94604242e-01 4.14868295e-01 -4.06913191e-01 -4.15424466e-01
7.26621002e-02 -1.80570930e-01 -5.08543774e-02 -3.65205184e-02
4.74008292e-01 -6.02453232e-01 -9.21833813e-01 2.31306612e-01
-9.89320517e-01 2.12505117e-01 7.25080311e-01 6.51649058e-01
6.60785913e-01 -4.88096811e-02 5.38522482e-01 -7.32922435e-01
1.72009468e-01 -4.02908474e-01 -3.55048448e-01 4.48328108e-01
-1.12415373e-01 6.22099340e-02 5.56357324e-01 -5.19084275e-01
-1.06775403e+00 2.17620701e-01 5.28028309e-01 -4.33929026e-01
-3.77805054e-01 3.71958911e-01 -8.15401971e-02 -1.27958864e-01
1.06745815e+00 1.76381364e-01 -6.19849503e-01 -2.59476930e-01
8.47376943e-01 2.32355878e-01 7.01037943e-01 -3.90209466e-01
9.07770038e-01 8.09730589e-01 3.76642644e-02 -7.40876973e-01
-1.06726062e+00 -5.50643504e-01 -1.28072882e+00 -3.61377269e-01
7.70177245e-01 -6.83172524e-01 -3.99659306e-01 -4.10210574e-03
-1.19359481e+00 -2.88514085e-02 -5.84542990e-01 4.27897602e-01
-3.71674627e-01 9.07011554e-02 -4.83498983e-02 -6.99404120e-01
1.03267051e-01 -7.11748242e-01 9.68805611e-01 5.76208472e-01
-1.59951032e-03 -7.45554149e-01 -3.67746323e-01 -1.38618112e-01
2.17592254e-01 4.85374838e-01 8.34569335e-01 -6.11408472e-01
-1.05460584e+00 -1.14530407e-01 -5.68817377e-01 -2.32645851e-02
1.78262159e-01 -2.69295484e-01 -1.08331800e+00 -3.36135447e-01
-2.78632134e-01 -1.10109389e-01 1.17792201e+00 1.27101019e-01
1.55733228e+00 -5.15064001e-02 -8.28531861e-01 5.20548880e-01
1.34432173e+00 1.36240423e-01 5.38257122e-01 1.54236794e-01
8.14569652e-01 6.07623041e-01 7.00264633e-01 3.14893484e-01
1.22957148e-01 4.97951239e-01 5.75194895e-01 -1.37478158e-01
-2.50774652e-01 -3.20126653e-01 -3.07087123e-01 5.33856638e-02
-4.72866260e-02 1.97063992e-03 -9.58667994e-01 8.85367990e-01
-1.88833404e+00 -8.72566998e-01 1.48511082e-01 2.13174963e+00
4.71066117e-01 2.16189265e-01 -1.97168320e-01 -1.40460491e-01
8.75508726e-01 1.01826012e-01 -8.25127721e-01 4.95557189e-02
-1.82109833e-01 1.84748441e-01 2.01356828e-01 -8.99620429e-02
-1.12456942e+00 1.13018703e+00 6.22714472e+00 4.98034686e-01
-8.80740166e-01 6.04252741e-02 3.81868571e-01 -2.08221436e-01
-5.20429201e-02 4.69309697e-03 -5.50759733e-01 4.76209223e-02
1.86127275e-01 -2.30888143e-01 4.55412418e-01 1.04214382e+00
-8.59662443e-02 -3.73755991e-01 -1.44161880e+00 9.97988820e-01
3.41263980e-01 -1.25040281e+00 -5.09092398e-02 -3.68784040e-01
8.73536944e-01 -6.97671100e-02 4.28718239e-01 4.08893377e-01
3.76963913e-02 -1.09279919e+00 7.60833800e-01 6.58389390e-01
6.67779088e-01 -4.92218018e-01 2.10543364e-01 3.54601711e-01
-1.40176463e+00 -4.28964734e-01 -7.65369415e-01 -8.90893042e-02
-1.71298757e-01 4.24846739e-01 -1.00290906e+00 3.03656250e-01
9.37503695e-01 8.83679628e-01 -1.35794115e+00 1.22946668e+00
-5.70255578e-01 4.47473563e-02 -2.30814964e-01 -1.01832218e-01
1.79195404e-01 9.39731002e-02 5.33427477e-01 1.00574231e+00
2.80775458e-01 3.88743520e-01 1.93964085e-03 1.49154735e+00
-1.26371890e-01 -1.90108959e-02 -7.20706761e-01 2.31558904e-01
5.43721497e-01 1.27993333e+00 -1.07870853e+00 -3.85042161e-01
-4.20999616e-01 1.10009956e+00 8.65346193e-01 6.72177255e-01
-7.94214547e-01 -5.07703483e-01 2.99483925e-01 1.74302533e-01
5.17451704e-01 -2.37603664e-01 -3.30603302e-01 -9.33117211e-01
2.08547577e-01 -3.76467317e-01 5.28022766e-01 -1.15864062e+00
-1.21967053e+00 5.43908000e-01 7.76977912e-02 -1.18719542e+00
7.62267113e-02 -5.67294776e-01 -6.98875248e-01 6.69545412e-01
-1.36611545e+00 -1.22611809e+00 -8.27047706e-01 4.86922830e-01
5.53364873e-01 -1.57046705e-01 5.26953578e-01 -6.50696233e-02
-3.22229236e-01 5.15533447e-01 -3.75360213e-02 2.62688160e-01
4.44147199e-01 -1.03799951e+00 5.91633856e-01 1.19553268e+00
5.22298992e-01 9.95403171e-01 4.33387876e-01 -7.20506787e-01
-8.81632388e-01 -1.37369084e+00 4.66307580e-01 -6.16090655e-01
3.40251833e-01 -6.63661182e-01 -1.13872015e+00 7.50957608e-01
-9.95795354e-02 7.68363893e-01 2.64439464e-01 4.95807678e-02
-5.93614757e-01 1.28798392e-02 -1.08200788e+00 9.23076689e-01
1.64399099e+00 -5.79136968e-01 -9.83500957e-01 3.17394167e-01
6.73119485e-01 -2.80178964e-01 -1.60055771e-01 6.03232920e-01
3.22441518e-01 -1.15960932e+00 1.13967419e+00 -8.22699785e-01
8.31464678e-02 -7.72084117e-01 -1.55548602e-01 -1.32005119e+00
-4.99514580e-01 -1.18968040e-01 -1.44749850e-01 9.54499841e-01
2.30298176e-01 -2.99745262e-01 6.40437305e-01 5.42170942e-01
-1.51061743e-01 -4.65882331e-01 -8.51583421e-01 -8.99264693e-01
-2.82208711e-01 -4.18182641e-01 5.50534606e-01 9.23170805e-01
-1.32732347e-01 3.46695274e-01 2.16391951e-01 3.09507221e-01
5.79160571e-01 2.24036872e-01 6.87176347e-01 -1.01116645e+00
-4.38745357e-02 -2.88585931e-01 -8.06134284e-01 -9.37722921e-01
1.56777978e-01 -9.41258848e-01 9.47229043e-02 -1.51752758e+00
4.00082529e-01 -4.74051118e-01 -6.09424293e-01 5.60819268e-01
-4.51225907e-01 3.38869900e-01 2.96044230e-01 -8.62609223e-02
-1.07677281e+00 6.15893781e-01 1.24095750e+00 -3.33914548e-01
-3.82314950e-01 -3.39063495e-01 -5.35745323e-01 8.08710337e-01
4.90173578e-01 -5.51617682e-01 -3.47822607e-01 -3.78264189e-01
1.14515767e-01 -5.68260312e-01 8.52774203e-01 -1.31509650e+00
5.30118406e-01 -1.63913101e-01 9.48311508e-01 -8.78073275e-01
1.51625678e-01 -1.01500833e+00 -3.81973423e-02 3.09991509e-01
-4.73738283e-01 -1.80653870e-01 3.90851855e-01 9.35285926e-01
3.21103893e-02 -2.75028020e-01 8.70830297e-01 -1.34788990e-01
-1.42360973e+00 2.78385222e-01 9.02651697e-02 1.52200460e-01
1.56019843e+00 -4.39820021e-01 -4.93209749e-01 5.11562489e-02
-9.37986553e-01 4.72935252e-02 4.40093964e-01 6.41792834e-01
1.08957005e+00 -1.48368979e+00 -3.67572963e-01 5.03764510e-01
6.58701479e-01 1.77066803e-01 5.15859842e-01 4.78930920e-01
-4.64502841e-01 1.92240819e-01 -3.11481297e-01 -8.63298833e-01
-1.15769315e+00 1.21420217e+00 3.69243860e-01 4.90652531e-01
-7.79633164e-01 1.03519320e+00 8.55286598e-01 -3.14669907e-02
1.93019882e-01 -4.17458683e-01 -3.06853235e-01 -2.84184098e-01
6.79423690e-01 2.41910756e-01 -2.06104591e-01 -5.99389374e-01
-6.13584220e-01 5.29153705e-01 -1.58871636e-01 1.39652520e-01
1.13088536e+00 -1.18167832e-01 1.40013974e-02 4.68627363e-01
1.10927987e+00 -1.78098336e-01 -1.45965886e+00 -5.50730944e-01
1.16620474e-01 -9.21158075e-01 -1.15133800e-01 -7.62593448e-01
-7.30927408e-01 9.40957963e-01 7.30057538e-01 -1.19086675e-01
1.18532944e+00 3.58133763e-01 2.91040331e-01 5.09594798e-01
3.44719708e-01 -1.14520669e+00 6.63102031e-01 3.63773972e-01
1.08915710e+00 -1.31004202e+00 1.17153645e-01 -6.97195590e-01
-6.10087335e-01 9.42058444e-01 8.44907641e-01 -2.24442393e-01
5.79350531e-01 -2.12800279e-01 -2.80082822e-01 -3.06600928e-01
-3.08832109e-01 -6.12477720e-01 6.73415244e-01 9.80460644e-01
-1.60066020e-02 -5.44535741e-02 6.49346504e-03 5.63092291e-01
4.13275994e-02 -2.76628762e-01 3.08125913e-01 9.79865015e-01
-5.89974344e-01 -4.23816651e-01 -2.66209662e-01 4.57099348e-01
3.21427099e-02 -2.14551508e-01 -3.18791986e-01 7.91463792e-01
3.11220437e-01 8.21296155e-01 4.14703667e-01 -2.33989581e-01
5.73172629e-01 -2.24929765e-01 6.04532719e-01 -9.42961156e-01
-2.60627389e-01 -3.29269499e-01 -4.74913001e-01 -5.93696773e-01
-3.52583796e-01 -4.95576590e-01 -1.41827989e+00 7.15766624e-02
-3.60502660e-01 -2.49879867e-01 4.45761621e-01 9.14040864e-01
4.84050274e-01 8.90819252e-01 5.80024183e-01 -7.50519872e-01
-6.27385750e-02 -8.87302935e-01 -6.99095309e-01 7.15024292e-01
2.10825920e-01 -1.05264676e+00 -2.88111359e-01 1.81506768e-01] | [9.884894371032715, 1.627812147140503] |
40522563-7b83-484f-a115-9f5471a85bad | visualization-of-contributions-to-open-source | 2010.08874 | null | https://arxiv.org/abs/2010.08874v1 | https://arxiv.org/pdf/2010.08874v1.pdf | Visualization of Contributions to Open-Source Projects | We want to analyze visually, to what extend team members and external developers contribute to open-source projects. This gives a high-level impression about collaboration in that projects. We achieve this by recording provenance of the development process and use graph drawing on the resulting provenance graph. Our graph drawings show, which developers are jointly changed the same files -- and to what extent -- which we show at Germany's COVID-19 exposure notification app 'Corona-Warn-App'. | ['Andreas Schreiber'] | 2020-10-17 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [-3.34362566e-01 2.90862948e-01 5.06769896e-01 -3.21155079e-02
-2.04317629e-01 -8.91860068e-01 4.89685774e-01 4.89907384e-01
1.77920461e-01 4.14496243e-01 4.34712231e-01 -4.02236640e-01
-6.76736534e-02 -5.80802679e-01 -4.04663473e-01 1.39329180e-01
-2.13946223e-01 -3.81751865e-01 4.43247110e-01 -2.41618380e-01
7.17937946e-01 -7.76298866e-02 -9.84841168e-01 5.27319849e-01
4.18087959e-01 -2.39633843e-02 -3.83951627e-02 9.14618552e-01
-1.04856834e-01 1.20240808e+00 -8.82648468e-01 -1.03461993e+00
1.53777197e-01 -5.33166289e-01 -8.57096016e-01 -4.80315596e-01
2.34998241e-01 -1.23519763e-01 2.51459956e-01 1.16237378e+00
1.03835851e-01 -7.10571826e-01 1.52634054e-01 -1.65142822e+00
-1.05499589e+00 8.43600929e-01 -7.52769768e-01 5.05671382e-01
8.49394202e-01 1.78141952e-01 6.83534145e-01 -8.59209359e-01
1.72697210e+00 7.01255977e-01 1.36301434e+00 1.36094928e-01
-1.03444827e+00 -6.16457403e-01 -3.64294440e-01 -1.75969169e-01
-1.12425840e+00 -1.76093295e-01 4.53907281e-01 -1.53134704e+00
8.40406358e-01 2.81452417e-01 8.06705296e-01 9.15522933e-01
6.34727240e-01 -2.46796817e-01 1.16311872e+00 -4.36998278e-01
1.70747370e-01 2.55635083e-01 3.55845153e-01 5.60857534e-01
6.87202990e-01 -3.70657206e-01 -7.89976358e-01 -7.03242481e-01
5.94951987e-01 3.62511650e-02 -4.42479014e-01 1.29286617e-01
-1.20076573e+00 2.25403141e-02 -1.82881296e-01 5.20873845e-01
-3.00864905e-01 7.28592217e-01 5.59373856e-01 6.23165309e-01
6.40258968e-01 3.32931370e-01 -3.22328478e-01 -8.37951422e-01
-8.44531536e-01 8.47442774e-04 1.06576204e+00 1.12299883e+00
8.03700686e-01 -3.62049252e-01 2.01552317e-01 2.44557083e-01
7.82009065e-01 3.58276851e-02 -3.50194097e-01 -8.94538879e-01
5.26841521e-01 6.79048240e-01 2.17025682e-01 -9.05896008e-01
2.58471370e-01 5.91099123e-03 6.04547784e-02 1.09377825e+00
3.69517028e-01 -3.22342992e-01 -2.03091606e-01 1.11081815e+00
7.71096349e-02 -8.75534341e-02 -3.85938734e-01 2.70003021e-01
4.32016551e-01 -2.15367842e-02 -1.18125498e-01 -1.80007383e-01
1.22732127e+00 -3.83967370e-01 -1.07778203e+00 3.20444912e-01
7.59507358e-01 -1.06720805e+00 7.41751611e-01 5.01511753e-01
-9.48812962e-01 -2.55140603e-01 -1.19951546e+00 3.38211596e-01
-2.23538384e-01 -2.13618338e-01 1.30927145e-01 1.03044629e+00
-1.26728320e+00 9.48348522e-01 -8.66567671e-01 -6.24534011e-01
6.87194824e-01 -4.55331624e-01 -6.20063543e-01 1.54536679e-01
-7.35204339e-01 8.04678559e-01 -5.67013443e-01 -9.01810080e-02
-8.49019289e-01 -1.12968493e+00 -2.05469817e-01 -2.53170848e-01
-5.03726751e-02 -3.16788316e-01 1.29409552e+00 -6.64218307e-01
-7.97733366e-01 9.83298361e-01 3.88444513e-01 3.28067780e-01
7.00211942e-01 -3.80221397e-01 -7.94122636e-01 -1.61833495e-01
4.12457347e-01 -5.16807139e-01 5.37037134e-01 -1.36888397e+00
-5.77931345e-01 -4.00118232e-01 3.21177915e-02 -7.36487508e-01
-2.66033322e-01 1.33034408e+00 -2.15288118e-01 -5.47131360e-01
-4.32676315e-01 -6.93110943e-01 -2.59630322e-01 3.96430820e-01
-1.80566788e-01 3.31791416e-02 7.09787309e-01 -1.08748591e+00
1.90278208e+00 -2.13285160e+00 -1.68214336e-01 1.74271598e-01
8.49925816e-01 -1.18763469e-01 1.46923006e-01 1.40584874e+00
-1.18436053e-01 9.94425356e-01 5.72298048e-03 -1.43324941e-01
3.77083793e-02 -3.24232101e-01 3.60877515e-04 7.22110033e-01
-1.00981578e-01 4.38697577e-01 -1.09806752e+00 -2.29019687e-01
-6.97064877e-01 4.16656375e-01 -1.21290594e-01 6.56253621e-02
1.29521415e-01 1.09978773e-01 -3.19892436e-01 3.60068887e-01
8.29946756e-01 -3.13621074e-01 4.52531040e-01 5.61524153e-01
-7.80096114e-01 2.50022933e-02 -9.76678908e-01 1.96490872e+00
-2.93656252e-02 1.31698656e+00 3.12413335e-01 1.97805107e-01
1.14740598e+00 7.71916986e-01 -9.80248675e-02 -1.42233208e-01
-3.06523323e-01 7.12504014e-02 -1.07623845e-01 -8.05161715e-01
5.28927624e-01 2.53421247e-01 -8.30721483e-02 1.30674696e+00
-2.45477200e-01 1.75120950e-01 1.62541598e-01 7.04081118e-01
1.87568986e+00 5.76492369e-01 2.72669256e-01 -4.05930638e-01
-1.80990741e-01 -2.64155772e-02 5.25184989e-01 3.77702475e-01
-1.16806298e-01 6.80233181e-01 1.44689059e+00 -4.27087963e-01
-1.10122418e+00 -8.59848619e-01 -5.20306304e-02 6.81446850e-01
-1.65955886e-01 -1.40357876e+00 -9.07137215e-01 -8.24374139e-01
-2.41179079e-01 5.03716111e-01 -1.00682604e+00 4.22364742e-01
-3.83448333e-01 1.73259079e-01 6.61015749e-01 4.30004537e-01
-3.46547395e-01 -9.58853900e-01 -1.18132472e+00 -8.12937915e-02
2.68807650e-01 -5.01842380e-01 -2.52039045e-01 -4.01042074e-01
-5.69585741e-01 -1.53806913e+00 -5.93024373e-01 -1.54501528e-01
7.76555181e-01 -2.07254604e-01 1.20207500e+00 6.35612845e-01
-6.05714738e-01 5.04129708e-01 -5.89802325e-01 -4.08623040e-01
-8.29617321e-01 -6.07778728e-01 -5.41408479e-01 -4.50788110e-01
1.75310329e-01 -8.77689064e-01 -2.36889720e-01 1.38041168e-01
-5.52377701e-01 -2.76950777e-01 -3.05969477e-01 -6.71691634e-03
-3.35184813e-01 -2.55662143e-01 1.55885473e-01 -1.31352878e+00
9.49674308e-01 -9.77865696e-01 -5.00926375e-01 6.19451463e-01
-7.27057397e-01 -6.43034577e-01 2.22184405e-01 1.48984447e-01
-8.75279605e-01 -6.09955013e-01 5.47660351e-01 -3.03863883e-01
4.62686531e-02 7.92667627e-01 2.78408289e-01 -4.87517789e-02
1.00331378e+00 -6.96220875e-01 -5.16512953e-02 -6.69274688e-01
1.98173910e-01 8.22724402e-01 3.22802722e-01 -3.96064669e-01
7.87150741e-01 4.44823623e-01 -5.19800723e-01 -3.54707420e-01
2.29594111e-01 -2.12540716e-01 -6.57089770e-01 -6.68076634e-01
7.53816545e-01 -7.63660252e-01 -4.20796126e-01 4.29029495e-01
-1.94542623e+00 -4.48954135e-01 -3.88771892e-01 1.74491361e-01
3.14937718e-02 2.94802129e-01 -5.51354766e-01 -9.85231936e-01
3.86693776e-02 -6.78704679e-01 2.98249125e-01 1.44212022e-01
-7.42459238e-01 -8.64348888e-01 1.17855549e+00 -1.73823506e-01
6.63775325e-01 8.94517541e-01 4.00141060e-01 -2.82904506e-01
-6.00071311e-01 -3.49835167e-03 -2.85651654e-01 2.06592262e-01
8.20465535e-02 1.07862711e+00 -8.55489016e-01 -1.88309059e-01
-2.66066223e-01 4.76883739e-01 -3.44381332e-01 -3.46975714e-01
2.99529761e-01 -8.54325518e-02 -5.76750040e-01 1.31115571e-01
1.71583676e+00 8.83467216e-03 9.12609339e-01 4.68930215e-01
7.50215232e-01 9.23878908e-01 4.35578912e-01 8.96406233e-01
6.13966733e-02 4.82669383e-01 4.30516064e-01 2.76982546e-01
-4.68620002e-01 -6.62032291e-02 4.65306640e-01 9.62382138e-01
-6.49433851e-01 -2.67891049e-01 -1.58128178e+00 8.17069292e-01
-1.87300646e+00 -9.26391602e-01 -1.38306582e+00 1.84348881e+00
5.70677340e-01 -1.65135115e-01 -1.46334440e-01 -2.43916512e-01
1.17080295e+00 1.23297289e-01 4.04590577e-01 -4.66850668e-01
5.26524305e-01 8.55231732e-02 5.83276451e-02 1.68606699e-01
-9.70291719e-02 1.05831251e-01 7.32422829e+00 2.92819571e-02
-5.58545649e-01 7.94980407e-01 -2.72170454e-01 4.43610996e-02
-8.63919377e-01 9.32203829e-01 -2.00364873e-01 8.13256621e-01
1.12554872e+00 -5.41710019e-01 4.82613146e-02 7.13876426e-01
2.79602092e-02 -2.61404186e-01 -1.28649807e+00 5.64581394e-01
-8.30101594e-02 -1.70630765e+00 -7.33734369e-01 1.83008239e-01
8.70829105e-01 7.00784400e-02 -6.50006652e-01 -4.52512443e-01
8.08924854e-01 -7.44882643e-01 1.23037052e+00 1.01670778e+00
1.18122649e+00 -2.97490656e-01 3.60922515e-01 -1.54941842e-01
-9.47224379e-01 1.89537451e-01 -2.10890859e-01 -4.12798703e-01
5.84605873e-01 9.55224156e-01 -5.20161688e-01 6.90825939e-01
1.14722788e+00 6.78093135e-01 -7.43723452e-01 1.17654157e+00
-7.17189193e-01 6.00028992e-01 2.42909983e-01 4.14006822e-02
-5.69893479e-01 -3.20948571e-01 5.94274282e-01 1.67651486e+00
6.25530064e-01 -5.93899786e-02 -7.06044137e-01 1.09934235e+00
2.77522784e-02 -2.85449415e-01 -8.64518404e-01 -4.56932902e-01
7.68054247e-01 1.39722133e+00 -8.28693390e-01 -7.80431628e-02
-4.37075585e-01 7.17295051e-01 1.60133421e-01 2.72715062e-01
-4.87633973e-01 -1.03659713e+00 7.25900114e-01 6.16562903e-01
1.34144366e-01 -2.89294541e-01 -2.57773966e-01 -6.86816692e-01
4.10643041e-01 -6.15190029e-01 -1.41499266e-01 -1.34490657e+00
-1.11026335e+00 6.95883632e-01 -2.43447736e-01 -1.20210922e+00
1.89939156e-01 6.77521974e-02 -1.21531630e+00 1.01142299e+00
-4.62713391e-01 -1.15065289e+00 -6.12293303e-01 2.18791008e-01
-3.00648123e-01 -1.99799120e-01 6.62912607e-01 2.75010258e-01
-4.51008767e-01 1.69538572e-01 -3.07192449e-02 1.37774691e-01
8.83258343e-01 -1.24086416e+00 9.36429441e-01 1.27367878e+00
5.17017655e-02 1.08707714e+00 6.97645187e-01 -1.22298920e+00
-1.05434465e+00 -6.03372514e-01 1.12954271e+00 -1.14155149e+00
1.33952940e+00 -5.56509554e-01 -8.00764561e-01 9.13296759e-01
1.15594721e+00 5.40102184e-01 9.08772290e-01 2.49946371e-01
-7.09342778e-01 3.46238256e-01 -9.01219428e-01 2.53800690e-01
1.29431796e+00 -1.13904679e+00 -6.77251577e-01 3.68386924e-01
5.63759685e-01 -1.76497802e-01 -1.48673630e+00 -5.85429907e-01
6.08262241e-01 -1.47608912e+00 1.53341621e-01 -5.50923884e-01
1.08320951e+00 -6.10637605e-01 3.35015655e-01 -1.25603056e+00
-1.62682593e-01 -1.42369199e+00 3.73325050e-01 2.03691459e+00
3.09751481e-01 -7.78664112e-01 2.11271018e-01 7.44010627e-01
-1.50663391e-01 -1.84513614e-01 -5.96353590e-01 -7.85459876e-01
-1.65324837e-01 -5.08723974e-01 8.13049316e-01 1.71458745e+00
8.86336148e-01 -1.85053304e-01 1.33182947e-02 1.24146387e-01
3.09013993e-01 -5.60878813e-01 9.18666363e-01 -1.40142858e+00
-3.79395902e-01 -5.04507162e-02 -4.99199152e-01 2.43377224e-01
-5.98021805e-01 -7.80265987e-01 -2.63745785e-01 -1.70959532e+00
4.12197471e-01 -6.58907175e-01 5.97782806e-02 5.19811273e-01
8.08124915e-02 1.18308604e-01 2.99751103e-01 6.21177435e-01
-6.57440662e-01 -5.48456252e-01 6.98106527e-01 5.53671479e-01
2.28400305e-01 -5.83788216e-01 -8.58597338e-01 3.39832991e-01
3.14013779e-01 -1.29521298e+00 -1.08923778e-01 -5.29390514e-01
1.38228500e+00 2.45218471e-01 5.69053590e-01 -8.82967055e-01
3.87386024e-01 8.33076760e-02 -1.29815117e-01 -2.37226367e-01
-5.98425329e-01 -5.22903502e-01 1.46223485e+00 6.03023469e-01
-1.10208042e-01 4.40182537e-01 5.31088300e-02 5.93086720e-01
-2.72043571e-02 -8.51610124e-01 -1.20470226e-02 -7.59944692e-02
-1.32276818e-01 -8.73935968e-02 -5.26343942e-01 1.75909296e-01
1.19546163e+00 -4.60471690e-01 -1.24139094e+00 1.38122022e-01
-4.61588144e-01 -5.86170964e-02 1.40073943e+00 3.34810555e-01
1.38178334e-01 -1.16651773e+00 -7.94614255e-01 -3.13106090e-01
2.14629695e-01 -4.86189812e-01 7.13994876e-02 9.87874627e-01
-1.03060162e+00 -5.61253130e-01 -6.45916700e-01 -1.77244574e-01
-1.40419924e+00 2.86343694e-01 -1.47233263e-01 -4.89653181e-03
-8.71206999e-01 9.52162981e-01 -2.83449531e-01 -1.27554774e-01
-3.76767129e-01 1.03173703e-01 -2.90009156e-02 2.59871602e-01
7.83674419e-01 1.14338398e+00 1.65946692e-01 -5.38993403e-02
-7.97892332e-01 6.33337498e-01 2.01143190e-01 -3.51992249e-01
1.60454345e+00 -8.03457722e-02 -8.14681113e-01 8.53919983e-01
1.06530559e+00 8.28198075e-01 -1.41342282e+00 3.82418722e-01
4.04686064e-01 -8.96205902e-01 -4.10338581e-01 -8.12814653e-01
-8.74734282e-01 9.93103504e-01 3.33483517e-01 8.07814240e-01
1.66795656e-01 9.47194695e-02 -1.01766378e-01 -4.11968142e-01
5.94434679e-01 -8.52308452e-01 -3.91484499e-02 3.78466658e-02
1.27709150e+00 -4.34893489e-01 5.99187613e-01 -5.34824371e-01
-3.81817847e-01 1.50118780e+00 2.36868322e-01 -2.53671482e-02
7.90922463e-01 8.07185590e-01 1.98928952e-01 -8.61206055e-01
-8.36808026e-01 6.73237741e-01 -5.19610763e-01 1.00958073e+00
6.52200341e-01 -1.07933424e-01 -4.94852334e-01 6.71547711e-01
1.34056553e-01 5.79374373e-01 1.56379306e+00 1.36240375e+00
9.53755006e-02 -1.34826815e+00 -5.19560635e-01 1.60834774e-01
-6.87586069e-01 -1.12965748e-01 -6.23566210e-01 8.41702878e-01
3.19779843e-01 7.81010151e-01 -1.71479937e-02 -3.32278758e-01
2.48132974e-01 -2.06499457e-01 4.27395135e-01 -9.16874111e-01
-1.46759081e+00 -3.98250967e-01 7.03362644e-01 -8.12241673e-01
-1.87205851e-01 -1.01090765e+00 -9.46057141e-01 -4.30865377e-01
-3.01459074e-01 1.68961450e-01 1.04706943e+00 4.55331028e-01
7.00434923e-01 6.90019906e-01 2.49792084e-01 -4.75936532e-01
-6.28028810e-02 -8.95580590e-01 -8.24896097e-01 1.64603174e-01
3.65380757e-02 -2.73079515e-01 -6.80633068e-01 5.29859543e-01] | [7.847101211547852, 7.661530494689941] |
e54002f2-310e-4021-b252-f5d4137887f8 | convolutional-neural-network-achieves-human | 1802.09697 | null | http://arxiv.org/abs/1802.09697v1 | http://arxiv.org/pdf/1802.09697v1.pdf | Convolutional Neural Network Achieves Human-level Accuracy in Music Genre Classification | Music genre classification is one example of content-based analysis of music
signals. Traditionally, human-engineered features were used to automatize this
task and 61% accuracy has been achieved in the 10-genre classification.
However, it's still below the 70% accuracy that humans could achieve in the
same task. Here, we propose a new method that combines knowledge of human
perception study in music genre classification and the neurophysiology of the
auditory system. The method works by training a simple convolutional neural
network (CNN) to classify a short segment of the music signal. Then, the genre
of a music is determined by splitting it into short segments and then combining
CNN's predictions from all short segments. After training, this method achieves
human-level (70%) accuracy and the filters learned in the CNN resemble the
spectrotemporal receptive field (STRF) in the auditory system. | ['Mingwen Dong'] | 2018-02-27 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [ 3.75435442e-01 -4.56041694e-01 1.86999619e-01 -1.80656239e-02
-5.36227286e-01 -6.79624498e-01 1.59344837e-01 5.43947816e-02
-4.32083160e-01 2.42664978e-01 2.00794324e-01 6.52828291e-02
-2.04144016e-01 -6.65637374e-01 -4.11703616e-01 -4.98737872e-01
9.39377174e-02 -9.18241125e-03 1.43832117e-01 -1.35669664e-01
6.78318024e-01 2.71734387e-01 -1.86084068e+00 9.11277413e-01
5.84140897e-01 1.64359522e+00 3.17520320e-01 8.65013957e-01
-9.35304537e-02 6.16961181e-01 -7.36457527e-01 -3.53482254e-02
1.83343172e-01 -7.86112964e-01 -1.00445724e+00 -3.49413902e-01
3.72922063e-01 1.46133333e-01 -2.53489077e-01 1.07534063e+00
6.21023655e-01 2.27246851e-01 6.56147718e-01 -7.92116880e-01
-5.11197031e-01 9.88170505e-01 -1.66974217e-01 3.54259610e-01
5.27467191e-01 -2.60345265e-02 1.24315977e+00 -6.31905794e-01
2.00991064e-01 8.60881209e-01 9.00985122e-01 4.41948563e-01
-1.21227169e+00 -1.05640113e+00 -2.63236403e-01 4.61661816e-01
-1.51438105e+00 -2.81269342e-01 9.28835392e-01 -8.14290643e-01
1.03233325e+00 2.88735330e-01 1.26663005e+00 1.02830553e+00
3.91119242e-01 5.37378311e-01 9.97560799e-01 -5.07233024e-01
1.94113016e-01 -2.36023754e-01 -2.36784071e-02 9.18716565e-03
-3.43631923e-01 1.47196963e-01 -8.13799739e-01 2.64750626e-02
7.81695306e-01 -2.79710472e-01 -1.92435369e-01 4.57174808e-01
-1.40641534e+00 5.38012087e-01 7.01197267e-01 7.41203249e-01
-3.95244628e-01 1.53209835e-01 6.26662552e-01 4.48147774e-01
1.09017804e-01 1.00952971e+00 -3.26460510e-01 -2.27295652e-01
-1.22418106e+00 2.81692505e-01 6.15123391e-01 4.41375047e-01
1.38221368e-01 3.57014239e-01 -1.99696615e-01 1.17780924e+00
-2.95004360e-02 1.62002414e-01 8.83757353e-01 -1.04281342e+00
-1.47817403e-01 4.03914720e-01 -3.03116113e-01 -8.81077111e-01
-5.45871615e-01 -8.90819550e-01 -8.96884501e-01 2.70842612e-01
6.09112978e-01 2.84170985e-01 -5.78034222e-01 1.61015546e+00
-2.51384407e-01 1.76284790e-01 -3.41614306e-01 1.20784414e+00
7.75107265e-01 4.51975912e-01 4.07028198e-03 -9.54204127e-02
1.63990211e+00 -5.32468975e-01 -4.36353892e-01 1.09658003e-01
-8.72225538e-02 -1.03733218e+00 1.13688910e+00 1.07487440e+00
-8.34577143e-01 -1.13085103e+00 -1.18733227e+00 1.19937919e-01
-2.37411886e-01 1.69255584e-01 4.80886906e-01 3.61396104e-01
-6.66516721e-01 9.67924237e-01 -1.96814090e-01 -2.64297962e-01
5.44584155e-01 4.67469424e-01 -1.29852235e-01 5.99050462e-01
-1.16802633e+00 6.17861331e-01 5.33598661e-01 -1.18409991e-01
-1.09659457e+00 -6.92899466e-01 -2.27371767e-01 2.97605127e-01
-1.90397978e-01 -6.91294909e-01 1.55252743e+00 -1.31005621e+00
-1.59405529e+00 9.01179671e-01 1.42199039e-01 -4.79146183e-01
-1.35584801e-01 -3.16180922e-02 -6.34500504e-01 7.79766813e-02
-3.29252072e-02 6.27939999e-01 9.95986342e-01 -8.22801828e-01
-8.67733777e-01 -2.46908948e-01 3.09126023e-02 -7.61165544e-02
-2.98506796e-01 1.76671371e-01 1.62828133e-01 -1.04058886e+00
3.43930155e-01 -8.04465592e-01 1.65586755e-01 -3.76110613e-01
-3.49758387e-01 -4.04441327e-01 1.82171598e-01 -5.67848742e-01
1.41654742e+00 -2.55839658e+00 -3.45512256e-02 1.18915163e-01
1.98082715e-01 1.86447594e-02 -2.88795501e-01 3.61637563e-01
-2.45918065e-01 -1.09261312e-02 -4.72602583e-02 2.41877168e-01
1.58039734e-01 -4.57865030e-01 -6.68168187e-01 1.52469829e-01
-2.01581627e-01 6.40590966e-01 -7.40909338e-01 -9.45211127e-02
-2.11584102e-02 3.24680865e-01 -7.18902588e-01 4.84122112e-02
-2.44632229e-01 7.32241690e-01 -9.92387831e-02 6.02917373e-01
3.82361144e-01 1.42402336e-01 7.74034038e-02 -3.77751619e-01
-2.04889908e-01 7.30788231e-01 -9.30834949e-01 1.98375201e+00
-4.52626050e-01 1.00181842e+00 -2.48733267e-01 -1.01482487e+00
8.93731058e-01 5.10064602e-01 5.56634128e-01 -8.37870598e-01
3.99439394e-01 3.68841887e-01 8.97424281e-01 -4.28471386e-01
1.42674565e-01 -5.58801830e-01 -2.57691681e-01 5.27171612e-01
3.75176102e-01 -2.01289818e-01 -1.03277050e-01 -4.85336035e-01
9.59719718e-01 -4.01178226e-02 3.10485691e-01 -1.16320394e-01
4.60460782e-01 -1.27386823e-01 4.49374914e-01 8.03235173e-01
-1.57253310e-01 6.77956164e-01 2.44475514e-01 -6.59487903e-01
-9.54423428e-01 -1.05141497e+00 -3.80350977e-01 1.40375376e+00
-2.29503348e-01 -5.43499172e-01 -8.97092760e-01 -9.03606340e-02
-7.72341341e-02 4.84535784e-01 -6.06409252e-01 -3.33462358e-01
-3.10383618e-01 -3.36007029e-01 9.86576200e-01 4.83183026e-01
4.44014490e-01 -1.73250747e+00 -5.20212531e-01 4.12499636e-01
-3.17208916e-01 -8.55578482e-01 -2.59029537e-01 3.46519977e-01
-6.79722309e-01 -8.94831121e-01 -3.49623710e-01 -9.21681523e-01
-1.49899885e-01 -1.12526387e-01 9.06548500e-01 -2.59052157e-01
-4.64659959e-01 -2.13023394e-01 -3.53562474e-01 -6.65766120e-01
-9.88727733e-02 1.96552217e-01 1.83339074e-01 1.02534153e-01
4.29517448e-01 -1.05064726e+00 -7.88857341e-01 1.68387398e-01
-4.55414593e-01 -1.10015720e-01 4.53910857e-01 6.68685019e-01
7.02065766e-01 2.04785943e-01 9.03045893e-01 -3.01338732e-01
7.38987267e-01 -9.21709314e-02 -5.41301444e-02 -2.81500012e-01
-2.59182960e-01 -5.75641394e-01 8.99545312e-01 -9.24921453e-01
-3.59430850e-01 2.19844967e-01 -3.90542656e-01 -4.09488797e-01
-5.04429936e-01 4.64593560e-01 2.28792474e-01 -8.00216496e-02
8.93538177e-01 3.25704724e-01 -2.80457765e-01 -6.14749432e-01
-8.98333490e-02 1.09801579e+00 6.69850349e-01 -3.92290592e-01
3.39404702e-01 1.31865039e-01 -3.41213256e-01 -7.62836993e-01
-1.21565068e+00 -3.80382359e-01 -5.49276292e-01 -5.34336686e-01
9.51721847e-01 -7.27359176e-01 -1.33864617e+00 5.46084464e-01
-1.03675354e+00 -2.14538291e-01 -4.69610929e-01 9.60823178e-01
-9.21817183e-01 -7.51761049e-02 -6.74688160e-01 -8.06517541e-01
-3.81292254e-01 -7.63914168e-01 9.58256066e-01 1.58211112e-01
-7.58479476e-01 -2.90596783e-01 2.23497018e-01 1.68603063e-01
3.75986069e-01 -6.92064017e-02 1.13129425e+00 -7.76668251e-01
-9.56202969e-02 -2.07427174e-01 -2.14865562e-02 3.97157550e-01
-6.95192963e-02 -3.46785367e-01 -1.73469210e+00 7.73812309e-02
2.34650388e-01 -2.71911949e-01 9.65791523e-01 6.04529381e-01
1.73825872e+00 8.65399931e-03 2.25551710e-01 6.98014140e-01
1.12998903e+00 5.98505020e-01 6.68526351e-01 3.32190454e-01
3.81789982e-01 3.67868006e-01 1.97640657e-01 3.07962477e-01
-1.23128325e-01 7.89712071e-01 3.28579128e-01 2.29166418e-01
-2.82896042e-01 -2.43277699e-01 2.82526076e-01 1.02456522e+00
-4.10035878e-01 1.58827260e-01 -6.94384634e-01 2.88218707e-01
-1.58333445e+00 -1.30094457e+00 9.44406539e-02 2.21690798e+00
8.01660180e-01 2.59939015e-01 4.94400740e-01 9.88579929e-01
5.99462152e-01 -2.27298588e-01 -4.91215765e-01 -7.40111291e-01
-9.17629898e-02 6.82603121e-01 -2.65135348e-01 -5.22471825e-03
-1.27717090e+00 8.25662136e-01 7.28103447e+00 1.12113106e+00
-1.42866766e+00 4.75146286e-02 5.49050001e-03 -1.85665235e-01
2.92376310e-01 -2.44752005e-01 -2.67927080e-01 3.06094021e-01
1.09540141e+00 -6.63615018e-02 1.03256464e+00 5.67735553e-01
2.16783836e-01 8.22443962e-02 -1.05226529e+00 1.25627863e+00
7.60364607e-02 -1.20446348e+00 1.31206274e-01 -8.11308902e-03
4.67858762e-01 -1.34493396e-01 1.73773840e-01 3.92495781e-01
-4.54787821e-01 -1.38246107e+00 1.10195935e+00 8.13620269e-01
7.95113564e-01 -8.39524269e-01 5.94696641e-01 4.20345277e-01
-1.28452957e+00 -6.18340492e-01 -5.00465274e-01 -5.57238042e-01
-3.06389749e-01 4.40410346e-01 -6.52746379e-01 6.32682070e-02
9.62471902e-01 8.41371596e-01 -4.12920952e-01 1.46061897e+00
4.67119068e-02 8.78786206e-01 6.74897507e-02 3.03856134e-02
-1.03719145e-01 1.99052263e-02 6.18049562e-01 1.31240356e+00
4.93863165e-01 5.42350635e-02 2.30267532e-02 9.04051363e-01
-9.58079547e-02 3.39260429e-01 -4.29940850e-01 -4.05913532e-01
2.40221709e-01 1.19321775e+00 -8.01733255e-01 -2.09255829e-01
-4.53826180e-03 7.87634969e-01 3.99307571e-02 2.56359577e-01
-5.91266572e-01 -6.90496087e-01 6.11657262e-01 2.06324980e-02
2.82264680e-01 1.14521580e-02 -6.71143234e-01 -8.69547009e-01
-3.49073172e-01 -1.00595367e+00 2.55622268e-01 -1.08089244e+00
-1.41445541e+00 7.67501593e-01 -6.74479961e-01 -1.74582469e+00
-7.68176019e-02 -6.96088254e-01 -5.86367130e-01 8.33043814e-01
-8.26958179e-01 -9.44293797e-01 -1.09542780e-01 6.12976313e-01
6.70373678e-01 -4.45830405e-01 1.38731015e+00 4.35421467e-01
-8.47795382e-02 6.15285933e-01 -1.01959608e-01 4.85334933e-01
6.64282262e-01 -1.21261132e+00 -4.67730425e-02 5.10247312e-02
5.82545877e-01 5.56109905e-01 4.17392880e-01 -1.64764494e-01
-8.49575937e-01 -7.47118473e-01 9.44641292e-01 -1.65508986e-01
7.85123885e-01 -4.94098991e-01 -6.80112481e-01 1.73838228e-01
1.73615187e-01 -3.66781831e-01 1.23608029e+00 3.89116526e-01
-7.12818921e-01 -3.89088720e-01 -7.55376697e-01 3.07016015e-01
1.04630792e+00 -1.05487573e+00 -9.01787639e-01 1.20209217e-01
3.79698038e-01 -1.79574937e-01 -1.06630874e+00 2.93301553e-01
1.33348012e+00 -8.77667665e-01 9.73465383e-01 -6.59010351e-01
5.72879136e-01 -5.26611984e-01 -3.95184726e-01 -1.35957038e+00
-7.01059639e-01 -1.84263602e-01 1.21001288e-01 7.12560534e-01
3.03704172e-01 -2.16652513e-01 1.10831872e-01 -5.72883964e-01
-3.86566669e-01 -4.90533918e-01 -1.00212216e+00 -7.41174459e-01
1.05521083e-01 -9.27503943e-01 5.54786026e-01 7.89738476e-01
3.86365443e-01 4.86361444e-01 -1.02036834e-01 -2.48675510e-01
2.42734030e-01 4.97205436e-01 3.92752439e-01 -1.59663332e+00
-3.26361090e-01 -1.02275753e+00 -8.15954447e-01 -5.56718767e-01
1.68649226e-01 -1.23968339e+00 5.20534925e-02 -1.14840984e+00
2.12709546e-01 -1.88984256e-02 -9.20834005e-01 3.76313180e-01
4.11339790e-01 9.29510713e-01 3.18411976e-01 3.75001520e-01
-4.28564906e-01 9.94377062e-02 1.43813920e+00 -4.44781780e-01
-1.42549500e-01 2.74532646e-01 -8.07161808e-01 8.98753047e-01
8.60489845e-01 -5.02518892e-01 -8.82069170e-02 2.03314871e-02
3.28424811e-01 4.44781482e-02 5.30045688e-01 -1.55089700e+00
1.97929084e-01 8.38625990e-03 8.82920504e-01 -4.63597000e-01
3.14380199e-01 -4.75464284e-01 3.30880821e-01 5.26910722e-01
-7.44665325e-01 -4.09741402e-01 2.71062315e-01 3.25765640e-01
-5.48049390e-01 -1.36230960e-01 8.94414783e-01 -5.12665808e-02
-5.02654791e-01 -6.45339713e-02 -7.15788960e-01 -1.98512152e-01
3.47224563e-01 -3.12393457e-01 2.88906437e-03 -1.99350879e-01
-1.35566223e+00 -6.44980669e-01 1.61718708e-02 3.67832541e-01
4.51626092e-01 -1.53951859e+00 -6.06762350e-01 3.72797996e-01
1.68117657e-01 -7.16287017e-01 4.46269453e-01 9.37243938e-01
-2.75046468e-01 3.65605175e-01 -6.01642132e-01 -7.67026424e-01
-9.43353593e-01 4.04633582e-01 6.13759458e-01 2.89376020e-01
-6.83354497e-01 8.64248872e-01 1.91947833e-01 -1.39396459e-01
3.67803395e-01 -4.13216144e-01 -7.62059629e-01 3.62418711e-01
7.51618803e-01 -7.17340223e-03 1.64614633e-01 -6.68868482e-01
-2.73888260e-01 9.17406797e-01 3.12634349e-01 -2.60633975e-01
1.31286538e+00 2.47721180e-01 -1.17066599e-01 1.09504557e+00
9.29205120e-01 -7.40933372e-03 -5.17662704e-01 5.86992987e-02
-2.04576880e-01 -3.13116789e-01 -1.04319461e-01 -1.05172896e+00
-9.04471099e-01 1.16531682e+00 7.87884414e-01 6.88740969e-01
1.49560356e+00 -1.64014459e-01 8.83128166e-01 3.39909017e-01
2.47300342e-01 -1.29830265e+00 1.09115943e-01 6.49870157e-01
1.10817420e+00 -6.91012323e-01 -4.47830826e-01 3.25557292e-02
-4.52395141e-01 1.39889717e+00 3.73679131e-01 -6.61811531e-01
9.58316386e-01 2.48856489e-02 5.59986271e-02 -1.05866440e-01
-6.26533210e-01 -3.64037782e-01 7.39044011e-01 5.06420374e-01
7.58449554e-01 3.03506613e-01 -2.89824635e-01 1.35480499e+00
-1.10473585e+00 1.50718540e-01 1.22574702e-01 3.41429740e-01
-6.83888376e-01 -5.49398661e-01 -3.54176193e-01 6.19203091e-01
-7.64623463e-01 -1.74556732e-01 -5.46229839e-01 2.90735394e-01
7.01919496e-01 1.08679867e+00 2.49030128e-01 -1.08198738e+00
4.01790589e-01 3.90707344e-01 7.25330353e-01 -7.42690802e-01
-1.15531003e+00 5.93675137e-01 -2.60085940e-01 -4.71504331e-01
-3.59521955e-01 -5.64739108e-01 -1.11203754e+00 -8.28051120e-02
-1.47080421e-01 1.25979736e-01 6.25798762e-01 8.65555227e-01
1.45539427e-02 9.71586168e-01 6.87041461e-01 -1.02319133e+00
-2.00147837e-01 -1.28485811e+00 -1.08149135e+00 3.56975585e-01
2.57010609e-01 -6.13727212e-01 -4.47323322e-01 2.55291522e-01] | [15.78613567352295, 5.173326015472412] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.